University of Southern California Dissertations and Theses

Identification of therapeutic targets for neurons and microglia in amyotrophic lateral sclerosis

(USC Thesis Other)

Identification of therapeutic targets for neurons and microglia in amyotrophic lateral sclerosis

PDF

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content Copyright 2022 Shu-Ting Hung Identification of Therapeutic Targets for Neurons and Microglia in Amyotrophic Lateral Sclerosis by Shu-Ting Hung A Dissertation Presented to the FACULTY OF THE USC GRADUATE SCHOOL UNIVERSITY OF SOUTHERN CALIFORNIA In Partial Fulfillment of the Requirements for the Degree DOCTOR OF PHILOSOPHY (DEVELOPMENT, STEM CELLS AND REGERATIVE MEDICINE) December 2022 ii Acknowledgements I began this PhD journey in the summer of 2016 with a single flight ticket and two luggage. With all the love and support from many people along the way, I am grateful to fulfill this six-year journey. First, I want to express my huge thankfulness to my mentor, Dr. Justin Ichida, for his guidance, support and help throughout this journey. There are a lot of up and down for me during the past six years, but he always encourages me and makes me stay positive and confident no matter what happened. Dr. Ichida is always there when I come up with any questions and inspires me how to think independently and how to tackle problems with different angles. I think it is impossible for me to have a better mentor than Dr. Ichida. I want to thank all the current and previous Ichida lab members for their advice and assistance. I want to thank my committee members Dr. Berislav Zlokovic, Dr. Qilong Ying and Dr. Michael Bonaguidi for their advice and encouragement throughout the years. I want to thank all our collaborators for the three projects, flow cytometry core, imaging core and Choi family therapeutic screening core for all their help. I would like to thank all my friends for their support and love throughout this journey. Pursuing a PhD as an international student is not easy especially during the pandemic. Thanks for celebrating all the important moments of my life and going through homesick together. Importantly, I would like to thank my husband, Chih-Ting (Tim) Chen, for his unconditional love, caring and support. He celebrates every happy moment and goes through tough times with me. He has supported me immensely to reach the end of my PhD. I am grateful for having so much support and love from my family and Tim’s family through this journey. I especially want to thank my brother, Shu- Fan Hung, the Hsu family, and Tim’s parents, Chin-Hui Chen and Tusi-Hui Hsiao, for their forever love and encouragement. Lastly and most importantly, I would like to thank my parents, Pei-Lin Hung and Tze-Ti Kao, for their unconditional love and support. Thanks for providing me the best environment and education and thanks for supporting me whatever I decide. Without their love and support I would never have made it this far. iii Table of Contents Acknowledgements ································································································· ii List of Figures ········································································································ v Abstract ··············································································································· vii Chapter 1 : Introduction ···························································································· 1 1.1 Amyotrophic lateral sclerosis and frontotemporal dementia ······································· 1 1.2 C9ORF72 mechanism and pathology in ALS························································· 3 1.3 Genetic etiologies and non-cell autonomous effect in ALS ········································ 7 1.4 Thesis Organization ······················································································ 10 Chapter 2 : Identification and Therapeutic Rescue of Autophagosome and Glutamate Receptor Defects in C9ORF72 and Sporadic ALS Neurons ············································ 12 2.1 Abstract ····································································································· 12 2.2 Introduction ································································································· 13 2.3 Results ······································································································ 14 2.4 Discussion ·································································································· 26 2.5 Figures ······································································································ 28 Chapter 3 : PIKFYVE Inhibition Mitigates Disease in Models of Diverse Forms of ALS ·········· 40 3.1 Abstract ····································································································· 40 3.2 Introduction ································································································· 40 3.3 Results ······································································································ 42 3.4 Discussion ·································································································· 63 3.5 Figures ······································································································ 66 Chapter 4 : Harnessing Neuroprotective Microglia for Therapeutic Benefit in C9ORF72 ALS/FTD ············································································································ 83 4.1 Abstract ····································································································· 83 4.2 Introduction ································································································· 84 4.3 Results ······································································································ 87 4.4 Discussion ·································································································100 4.5 Figures ·····································································································107 Chapter 5 : Conclusions ························································································122 iv 5.1 Discussion I: Antisense oligonucleotides as therapeutics for ALS ····························122 5.2 Discussion II: Astrocytes in ALS ······································································123 5.3 Discussion III: Microglia in sporadic ALS ···························································123 5.4 Conclusion ·································································································124 References ········································································································126 Appendices ········································································································147 Appendix A : Methods and Materials ·····································································147 A.1 Methods ································································································147 A.2 Experimental Models ················································································161 A.3 Statistical Analysis ···················································································162 Appendix B : Supplementary Figures for Chapter 2 ··················································163 Appendix C : Supplementary Figures for Chapter 3 ··················································176 Appendix D : Supplementary Figures for Chapter 4 ··················································197 v List of Figures Figure 1.1 Pathological overlap of ALS and FTD. ··························································· 3 Figure 1.2 ALS/FTD-linked mutations on TDP-43 and FUS domains. ·································· 3 Figure 1.3 Three proposed pathological mechanisms for C9off72 mutation. ·························· 5 Figure 1.4 The major human cell types in the CNS. ························································ 9 Figure 2.1 Identification of neurodegenerative phenotypes in iMNs from sporadic ALS patients. ············································································································· 28 Figure 2.2 C9ORF72 and sporadic ALS iMNs share autophagosome formation abnormalities that are rescued by 3K3A-APC. ··············································································· 30 Figure 2.3 Rescue of autophagosome formation abnormalities by 3K3A-APC improves proteostasis. ········································································································ 32 Figure 2.4 C9ORF72 and sporadic ALS iMNs have elevated glutamate receptor levels that are normalized by 3K3A-APC. ··········································································· 34 Figure 2.5 3K3A-APC rescues the survival of C9ORF72 and sporadic ALS iMNs in a PAR1- dependent manner. ······························································································· 36 Figure 2.6 3K3A-APC rescues C9ORF72 ALS proteostasis and glutamate receptor phenotypes in vivo. ······························································································· 38 Figure 3.1 PIKFYVE inhibition ameliorates C9ORF72 ALS/FTD neurodegeneration. ············ 66 Figure 3.2 PIKFYVE inhibition ameliorates disease pathology in C9ORF72 ALS/FTD iMNs. ··· 68 Figure 3.3 PIKFYVE inhibition increases exocytosis of aggregation-prone proteins from iMNs. 70 Figure 3.4 PIKFYVE inhibition clears pTDP-43 through amphisome and multivesicular body exocytosis ··········································································································· 73 Figure 3.5 PIKFYVE inhibition improves iMN proteostasis and survival for diverse forms of ALS. ··············································································································· 76 Figure 3.6 Pikfyve suppression improves motor function and extends survival of TDP-43 and C9ORF72 mice. ··································································································· 78 Figure 3.7 Pikfyve suppression reduces TDP-43 and C9ORF72 pathology and neurodegeneration in vivo. ······················································································ 80 Figure 4.1 Transcription factor-mediated conversion of iPSCs into induced microglia. ··········107 Figure 4.2 C9ORF72 ALS/FTD microglia display different transcriptomic profiles and alter endolysosome phenotypes. ····················································································109 Figure 4.3 C9ORF72 ALS/FTD microglia exert neuroprotection on C9ORF72 ALS/FTD neurons. ············································································································111 Figure 4.4 A novel CSF1R-low neuroprotective state of C9ORF72 ALS/FTD microglia ·········113 Figure 4.5 Gain-of-function mechanism of C9ORF72 ALS/FTD neurons drives the conversion of C9ORF72 ALS/FTD microglia to neuroprotective state ·······························115 Figure 4.6 C9ORF72 ALS/FTD microglia shift from CSF1R-high to CSF1R-low neuroprotective state in vitro. ··················································································116 Figure 4.7 Validation of the presence of CSF1R-low microglia in vivo. ······························118 Figure 4.8 CSF1R ASO as therapeutics in vivo. ··························································120 Figure S2. 1 Identification of neurodegenerative phenotypes in sporadic ALS patient iMNs. ··163 Figure S2. 2 C9ORF72 and sporadic ALS iMNs share autophagosome formation abnormalities that are rescued by 3K3A-APC. ····························································165 Figure S2. 3 Rescue of autophagosome formation abnormalities by 3K3A-APC improves proteostasis. ·······································································································167 Figure S2. 4 C9ORF72 and sporadic ALS iMNs have elevated glutamate receptor levels that are normalized by 3K3A-APC. ··········································································169 vi Figure S2. 5 3K3A-APC rescues the survival of C9ORF72 ALS iMNs in a PAR1-dependent manner. ·············································································································171 Figure S2. 6 3K3A-APC rescues the survival of sporadic ALS iMNs in a PAR1-dependent manner. ·············································································································173 Figure S2. 7 Rapamycin rescues the survival of iMNs from some sporadic ALS lines. ··········175 Figure S3. 1 PIKFYVE inhibition ameliorates C9ORF72 ALS/FTD neurodegeneration. ·········176 Figure S3. 2 PIKFYVE inhibition ameliorates disease pathology in C9ORF72 ALS/FTD iMNs. ················································································································179 Figure S3. 3 PIKFYVE inhibition increases exocytosis of aggregation-prone proteins from iMNs. ················································································································181 Figure S3. 4 PIKFYVE inhibition clears pTDP-43 through amphisome and multivesicular body exocytosis. ··································································································184 Figure S3. 5 PIKFYVE inhibition improves iMN proteostasis and survival for diverse forms of ALS. ·················································································································187 Figure S3. 6 Pikfyve suppression improves motor function and extends survival of TDP-43 and C9ORF72 mice. ·····························································································191 Figure S3. 7 Pikfyve suppression reduces TDP-43 and C9ORF72 pathology and neurodegeneration in vivo. ·····················································································193 Figure S4. 1 Transcription factor-mediated conversion of iPSCs into induced microglia. ·······197 Figure S4. 2 C9ORF72 ALS/FTD microglia display different transcriptomic profiles and alter endolysosome phenotypes. ····················································································198 Figure S4. 3 C9ORF72 ALS/FTD microglia exert neuroprotection on C9ORF72 ALS/FTD neurons. ············································································································200 Figure S4. 4 A novel CSF1R-low neuroprotective state of C9ORF72 ALS/FTD microglia. ·····201 Figure S4. 5 Gain-of-function mechanism of C9ORF72 ALS/FTD neurons drives the conversion of C9ORF72 ALS/FTD microglia to neuroprotective state ·······························202 Figure S4. 6 C9ORF72 ALS/FTD microglia shift from CSF1R-high to CSF1R-low neuroprotective state in vitro. ··················································································203 Figure S4. 7 Validation of the presence of CSF1R-low microglia in vivo. ···························205 vii Abstract Amyotrophic lateral sclerosis (ALS) is a detrimental neurogenerative diseases characterized by severe loss of motor neurons in the brain and spinal cord. In ALS patients, the motor neurons degenerate rapidly after disease onset, which leads to lose the ability to swallow, speak, breathe, and eventually causes death. However, there is no current cure for ALS. One reason is that ALS has diverse genetic etiology with up to 90% of cases are sporadic, caused by unknown mutations. Another reason is that more studies have shown that non-neuronal cells in the brain, such as microglia, play an important role in regulating neurodegeneration. To identify efficacious therapeutics for ALS, it is in a pressing need to determine the shared disease mechanisms between familial and sporadic ALS and to investigate the role of non-neuronal cells in ALS. In this dissertation, I present our efforts to identify three therapeutic targets for motor neurons and microglia in ALS, which provides effective therapeutic approach for multiple forms of ALS. 1 Chapter 1 : Introduction 1.1 Amyotrophic lateral sclerosis and frontotemporal dementia Amyotrophic lateral sclerosis (ALS), also known as Charcot’s disease or Lou Gehrig’s disease, is a fatal nervous system disorder featured by progressive degeneration of motor neurons in the brain and spinal cord. Symptoms begin with gradual muscle weakness and progress to the paralysis. ALS is adult-onset with the average age of 55, which patients die within the average of 2-5 years after diagnosis from respiratory failure (Cleveland and Rothstein, 2001; Taylor et al., 2016). Clinically, ALS correlates with frontotemporal dementia (FTD), with the fact that more than 50% of ALS patients detect with FTD symptoms, and approximately 30% FTD patients experience motor neuron defects (Ferrari et al., 2011). FTD, known as Pick’s disease, is one of the common types of dementia, comprised by the progressive atrophy of the frontal and temporal lobes of the brain. The onset of FTD also begins in mid-age between the range of ages 40 to 65 (Burrell et al., 2011; Abramzon et al., 2020). FTD patients undergo dramatic changes in personality and behaviors, and loss the ability to comprehend languages. Due to the trend of population ageing, the global ALS cases are estimated to increase by 69% in the next two decades according to epidemiologic study (Arthur et al., 2016). Unfortunately, there are no effective therapies for ALS and FTD (Tsai and Boxer, 2014). Besides clinical correlations, ALS and FTD share genetic overlaps. Over the past two decades, a plethora of genetic factors have been uncovered to genetically linked ALS and FTD and each genetic factors contribute to ALS and FTD in different extents, supporting the concept that ALS and FTD form a continuous genetic spectrum (Ling et al., 2013; Abramzon et al., 2020). These known genetic factors could be categorized into various pathways, further linking ALS and FTD mechanistically and pathologically. 2 Mutations related to RNA homeostasis includes TARDBP and FUS (Fig. 1.1) (Ferrari et al., 2011; Ling et al., 2013). TARDBP encodes TAR DNA-binding protein 43 (TDP-43), which normally presents in the nucleus and functions as RNA splicing and transportation regulator. One of the hallmarks for ALS and FTD is the cytoplasmic inclusions of the TDP-43 in the neurons. In the postmodern brain tissues of ALS and FTD patients, the cytoplasmic inclusions mainly composed of the C-terminus fragments of TDP-43, which is called glycine-rich domain and contains various ALS-associated mutations (Fig. 1.2) (Igaz et al., 2008; Cohen et al., 2011; Berning and Walker, 2019). The mis-localization of TDP-43 in the cytoplasm depletes the functional TDP-43 in the nucleus and causes aberrant splicing of stathmin-2 (STMN2) with STMN2 levels reduction and cryptic exon inclusions, another hallmark for ALS and FTD (Melamed et al., 2019; Ma et al., 2022). In addition, mutations in FUS (fused in sarcoma gene) occur in less than 1% of ALS and 10% of FTD, mostly located in prion-like domain and C-terminal nuclear localization signals (NLS) domain (Fig. 1.2) (Kwiatkowski et al., 2009). The pathology of FUS mutation is the accumulation of FUS aggregates in the cytoplasm, which constrains FUS from the nucleus and impairs its normal functions in RNA splicing and transport (Vance et al., 2009). Several genetic mutations shared between ALS and FTD are involved in autophagy and protein degradation pathways, such as OPTN, SQSTM1, TBK1 and UBQLN2 (van Rheenen et al., 2016; Abramzon et al., 2020). Optineurin and p62, encoded by OPTN and SQSTM1, bind target polyubiquitinated protein cargos to LC3 and direct the cargos for autophagy while Ubiquilin-2, translated from UBQLN2, associated polyubiquitinated protein cargos for proteasome degradation (Moscat and Diaz-Meco, 2012; Williams et al., 2012; Ying and Yue, 2016).TANK- binding kinase 1, encoded by TBK1, phosphorylates optineurin to enhance its binding ability to LC3 and promote autophagy (Richter et al., 2016; Oakes et al., 2017). These ALS and FTD-linked proteins lead to disrupted protein homeostasis and impaired protein clearance in the neurons, which ultimately contributes to neurodegeneration. 3 The percentage of cases with major pathological protein inclusions in ALS and FTD. SOD = superoxide dismutase 1; TAU = microtubule-associated protein tau; UPS = ubiquitin- proteasome system. (Modified from Ling et al., 2013) Figure 1.2 ALS/FTD-linked mutations on TDP-43 and FUS domains. Schematic of domain structures of TARBP (TDP-43) and FUS. TDP-43 and FUS have similar domain structures. The majority of ALS/FTD-linked mutations are in the domain marked with red star. NLS = Nuclear localization signal; RRM = RNA recognition motif; NES = Nuclear export signal; Glycine-rich = Glycine-rich region; QGSY = Glutamate, Glycine, Serine, Tyrosine-rich region; RGG = Arginine, Glycine-rich region; ZnF = Zinc finger domain. (Modified from Ling et al., 2013) 1.2 C9ORF72 mechanism and pathology in ALS In 2011, two groups have identified the massive hexanucleotide (GGGGCC) repeat expansion in the first intronic region of C9orf72 (chromosome 9 open reading frame 72) as the most common genetic cause for both ALS and FTD, tightening the linkage between ALS and FTD (DeJesus- Figure 1.1 Pathological overlap of ALS and FTD. 4 Hernandez et al., 2011; Renton et al., 2011). Healthy individuals often have less than 10 repeats, while patients carry hundreds or up to thousands of repeats. The C9orf72 mutation accounts for 40-50% of patients with familial ALS and is also responsible for 25% of familial FTD. This highlights the importance to elucidate the underlaying disease mechanisms in C9ORF72 ALS/FTD (Taylor et al., 2016). The C9orf72 gene produces three coding variants, including one short variant (Variant 1) and two long variants (Variant 2 and 3). Through alternative splicing, these three variants produce two protein isoforms: variant 1 translates to short isoform (222 amino acids), while variant 2 and 3 gives rise to long isoforms with 481 amino acids (DeJesus-Hernandez et al., 2011; Renton et al., 2011; Balendra and Isaacs, 2018; Smeyers et al., 2021). The C9ORF72 protein has been reported to be structurally like the DENN (differentially expressed in normal and neoplastic cells) domain protein family, which function as guanosine diphosphate (GTP)–guanosine diphosphate (GDP) exchange factors (GEF) and mediate membrane vesicle trafficking (Marat et al., 2011; Levine et al., 2013). Other evidence related to the C9ORF72 protein function is that C9ORF72 interacts or binds with RAB1A-ULK1 and SMCR8-WDR41, indicating its role in regulating autophagy and lysosomal degradation (Webster et al., 2016; Yang et al., 2016; Zhang et al., 2018). Three mechanisms have been proposed for C9orf72 hexanucleotide repeat expansion to induce neurodegeneration in ALS and FTD, including one loss-of-function and two gain-of-function mechanisms (Fig. 1.3). The loss-of-function mechanism is mediated by the haploinsufficiency of C9ORF72 functional activity. The repeat expansion suppresses the production of C9ORF72 proteins by inhibiting the transcription, which leads to dysregulation of endogenous C9ORF72 pathways (Shi et al., 2018). Clinical study has shown that a reduction of C9orf72 levels was detected in the frontal cortex and cerebellum of C9ofr72 repeat carriers (van Blitterswijk et al., 2015). The two gain-of-function mechanisms are resulted from transcription and translation of the C9orf72 expanded repeats. The transcripts from the expanded repeats form RNA foci, which 5 sequester RNA-binding proteins (RBPs) and affect RNA processing (Cooper-Knock et al., 2015; Conlon et al., 2016). The repeat associated non-ATG (RAN) translation from both sense and antisense directions generate five different dipeptide repeat proteins (DPRs), which form aggregates and accumulate in the cells (Zu et al., 2011; Kwon et al., 2014). More evidence has raised the possibility that three mechanisms are non-mutually exclusive to mediate toxicity. Figure 1.3 Three proposed pathological mechanisms for C9off72 mutation. (Top) The long hexanucleotide (GGGGCC) repeats locates in the non-coding region between exon 1a and exon 1b in C9orf72 locus. (1) Loss-of-function mechanism: The GGGGCC repeat expansion interferes the transcription of C9orf72 gene, leading to a decrease in C9ORF72 protein and loss its function. (2) RNA gain-of-function mechanism: Sense and antisense of RNA transcripts of the C9orf72 repeat expansion are generated, leading to accumulation of toxic RNA foci. (3) DPR gain-of-function mechanism: Sense and antisense of RNA are translated by RAN translation to synthesize dipeptide repeat proteins (DPRs), which accumulated in the neurons. (Modified from Gitler and Tsuiji, 2016) A variety of animal models, ranging from zebrafish, C. elegans, Drosophila to mice, have been used to dissect the pathogenic mechanisms of C9ORF72 repeat expansion. Zebrafish serve as a good genetic model for neurological diseases due to the development of complex central nerve system. Blockage the translation of zebrafish C9orf72 orthologue by antisense morpholino oligonucleotides suppresses C9orf72 levels and shortens axons of motor neurons in zebrafish (Ciura et al., 2013), while expression of C9orf72 hexanucleotide repeat expansion in zebrafish 6 cause both RNA and DPR toxicities and development of motor impairment and cognitive abnormalities (Shaw et al., 2018). The nematode C. elegans is a critical tool for modeling neurodegenerative diseases with well characterized nervous system and locomotion. C. elegans contains a C9orf72 homolog, alfa-1 (ALS/FTD associated gene homolog), sharing 58% protein sequence similarity. Therrien et al. found that deletion of alfa-1 leads to an age-dependent paralysis along with specific degeneration in motor neurons; on the other hand, Rudich et al. recapitulated DPR toxicity in C. elegans with the observation of the nuclear localized arginine- containing DPRs and age-dependent motor defects alone with severe neurodegeneration after expressing codon-optimized transgenes for DPRs (Therrien et al., 2013; Rudich et al., 2017). Drosophila is a good genetic model for identifying functional mutants and screening therapeutic compounds for neurodegenerative diseases. Unlike zebrafish and C. elegans, Drosophila lacks the C9orf72 homology. However, Drosophila is widely used for studying G4C2 repeat RNA and DPR toxicity (Xu et al., 2013; Mizielinska et al., 2014; Freibaum et al., 2015; Tran et al., 2015). One example of utilization the Drosophila model to identify the therapeutic targets is to perform RNAi screening and verified several nucleocytoplasmic transport genes as modifier of DPR toxicity (Boeynaems et al., 2016). A lot of efforts were made to develop mammalian system, the most common is mouse models, to pathological mechanisms of C9orf72. Two human C9ORF72 BAC (bacterial artificial chromosome)-transgenic mouse models were established in 2015, which displayed nuclear RNA foci and produced DPR proteins without developing neurodegenerative or behavioral features (O’Rourke et al., 2015; Liu et al., 2016). The mouse C9orf72 ortholog, 3110043O21Rik, is highly expressed in neurons which enhances the establish of mice models to mimic loss-of-function mechanism (Suzuki et al., 2013). Neither neuron-specific nor full ablation of C9orf72 in the mice is sufficient to drive motor deficits, but C9orf72-deficient mice triggers aberrant immune responses and phenotypes (Koppers et al., 2015; Sudria-Lopez et al., 2016; Atanasio et al., 2016; Burberry et al., 2016; O’Rourke et al., 2016). 7 To recapitulate human neurodegenerative diseases in a dish, the direct cellular reprogramming approach has been wisely utilized to convert neurons from human induced pluripotent stem cell (iPSC) through forced expression of neuronal transcription factors (Takahashi and Yamanaka, 2006; Vierbuchen et al., 2010; Son et al., 2011). Recent work from our team has generated induced motor neurons (iMNs) from C9ORF72 ALS/FTD patients iPSCs using seven neuronal transcription factors, Ascl1, Brn2, Myt1l, Isl1, Lhx3, NeuroD1 and Ngn2 (Shi et al., 2018). We demonstrated that iMNs derived from C9ORF72 ALS/FTD patients degenerate faster than healthy neurons, recapitulating ALS/FTD disease process. We also showed that restoring each isoform of C9ORF72 proteins in C9ORF72 patient iMNs is sufficient to rescue the neurodegeneration, and deletion of C9ORF72 in iMNs causes severe degeneration, supporting the notion that C9ORF72 is haploinsufficientent in ALS/FTD. In the C9ORF72 ALS/FTD iMNs, we also noticed the disrupted lysosomal biogenesis, emphasizing the function of C9ORF72 in vesicle trafficking (Webster et al., 2016; Yang et al., 2016; Zhang et al., 2018). In addition, we also demonstrated that C9ORF72 ALS/FTD iMNs present ALS-like pathologies such as accumulation of DPRs in the nucleus and elevated levels of glutamate receptors on the neurites. These findings reveal that loss-of-function and gain-of-function mechanisms work cooperatively to drive neurodegeneration in C9ORF72 ALS/FTD patients. 1.3 Genetic etiologies and non-cell autonomous effect in ALS Although lots of efforts has been devoted to seeking cure for ALS, currently, only five modest disease-modifying drugs are approved by FDA (Chiò et al., 2020). Riluzole along with its two various formulae, Tiglutik and Exservan, prolong the survival of ALS patients by the average of two months (Miller et al., 2002). In 2017, Edaravone was approved 22 years after Riluzole, which slows the disease progression possibly through reducing oxidative stress in the neurons (Rothstein, 2017). Nuedexta, the combination of dextromethorphan and quinidine, was first 8 approved by FDA for treatment of pseudobulbar affect and has been shown to improve the function of bulbar muscles (Smith et al., 2017). The obstacles of developing therapeutics for ALS are due to the diverse genetic etiologies and non-cell autonomy. Approximately 10% of ALS is familial cases, mostly inherited in an autosomal dominant fashion. The remaining 90% of ALS cases are due to complex genetic and environmental factor, termed as sporadic ALS. The TDP-43 pathology is the most known feature shared in over 90% of ALS patients, leading to both loss of nuclear TDP-43 functions and gain of toxic cytoplasmic TDP-43 aggregates (Neumann et al., 2006; Sreedharan et al., 2008). The unknown genetic and mechanistic cause of sporadic ALS impedes the progress of finding effective therapeutics. Thus, there is a pressing need to decipher the shared pathological mechanisms between various forms of ALS. For years, the studies of neurodegenerative diseases have been centered on neurons themselves, investigating the cell-autonomous effect on neurodegeneration. However, brain is composed of three main types of neurons, motor neurons, interneurons, and sensory neurons, and a various of non-neuronal cells, including microglia, astrocytes, oligodendrocytes, vesicular endothelial cells and pericytes (Hodge et al., 2019). These non-neuronal cells exert their normal functions and work together to maintain brain homeostasis. Emerging evidence have indicated that non- neuronal cells play critical roles in neurodegeneration, emphasizing the importance of non-cell autonomous effects (Garden and La Spada, 2012; Srinivasan et al., 2016; Chen et al., 2018; Beers and Appel, 2019; Clarke and Patani, 2020; Saez-Atienzar et al., 2021). 9 Figure 1.4 The major human cell types in the CNS. A summary of major cell types in the brain. There are three major types of neurons: interneurons, sensory neurons, and motor neurons; and three types of glia cells: astrocytes, microglia, and oligodendrocytes. Other non-neuronal cells include pericytes and vesicular endothelial cells. Non-neuronal cells contribute to the neurodegeneration process via the non- cell autonomous manners. (Modified from Penney et al., 2020) Neuroinflammation is a hallmark for ALS and most of the neurodegenerative disorders. Microglia, the resident immune cells in the brain, are responsible for immune surveillance and preventing neurons from injuries. Under the stress of neurodegeneration, microglia have been reported to possess both beneficial and detrimental roles: engulfing the amyloid plaques or triggering excessive cytokine releases and synaptic pruning (Salter and Stevens, 2017). In the postmodern tissues from ALS patients, microglia are over-activated and possess inflammatory transcriptomic signatures, correlating with disease progression in the motor cortex; while the positron emission tomography (PET) imaging further provides the evidence of microglial activation in the brains of living ALS patients (Brettschneider et al., 2012; Corcia et al., 2012; Dols-Icardo et al., 2020). Genome-wide association studies (GWAS) have identified genetic variants in microglial genes as risk factors for several neurodegenerative diseases. For instance, TREM2 (Triggering Receptor Expressed on Myeloid cells 2) variants R47H is associated with Alzheimer’s disease and sporadic ALS (Jonsson et al., 2013, Cady et al., 2014), and other microglial genes with risk variants include CR1, CD33, ABCA7 and APOE (Belbin et al., 2007, Brouwers et al., 2011; Malik et al., 2013 De Roeck et al., 2019). 10 The development of single-cell RNA sequencing serves as a powerful tool for dissecting the cell diversities and identifying rare cell types in different organs (Deng et al., 2014; Grün et al., 2015; Paul et al., 2015; Zeisel et al., 2015). The single-cell RNA-seq analysis uncovered a novel microglial subpopulation, disease-associated microglia (DAM), in Alzheimer’s disease, regulating by the presence of Aβ-plaques and TREM2-APOE pathways (Krasemann et al., 2017; Keren- Shaul et al., 2017). The DAM state of microglia is also detected in the SOD ALS mice model, yet the field lacks studies examining the role microglia in the C9ORF72 context where both gain- and loss-of-function disease mechanisms of the repeat expansion are present (Keren-Shaul et al., 2017). In addition, it remains unclear in which contexts neuron-microglia interactions drive or rescue neurodegeneration in C9ORF72 ALS/FTD. Thus, it is a pressing need to understand how microglia modulate neurodegeneration in the presence of both gain- and loss-of-function mechanisms of C9ORF72 and leverage these findings for therapeutic benefit. 1.4 Thesis Organization In this thesis, we aim to identify therapeutic targets for neurons and microglia in C9ORF72 ALS/FTD. Given that high percentage of ALS cases are sporadic with the diverse genetic etiology, we begin with deciphering the pathological mechanisms shared between C9ORF72 ALS/FTD and sporadic ALS. We utilized cellular reprogramming to generate induced motor neurons from C9ORF72 ALS/FTD and sporadic ALS. We found that impaired autophagosome formation and abnormal accumulation of glutamate receptors were two common pathologies shared between C9ORF72 ALS/FTD and a subset of sporadic ALS. Our findings suggest that restoring the autophagosome deficiency with anticoagulation-deficient form of activated protein C (3K3A-APC) significantly ameliorates the pathologies and promotes homeostasis in the neurons. Next, we showed that inhibition of PIKFYVE kinase is broadly efficacious on extending the survival and mitigating ALS pathology of induced motor neurons derived from C9ORF72 ALS/FTD, a subset of sporadic and other familial forms of ALS, as well as several ALS animal models. Harnessing 11 these finding, we identify PIKFYVE inhibition in the neurons as a therapeutic approach for multiple forms of ALS. In the last chapter, we moved our focus from neurons to microglia. To investigate the role of microglia in neurodegeneration in the C9ORF72 disease context, we established cellular reprogramming method to generate induced microglia from C9ORF72 ALS/FTD patients and develop neuron-microglia co-culture system. In the presence of C9ORF72 ALS/FTD motor neurons, we performed single-cell RNA-seq and revealed a distinct subpopulation of microglia in C9ORF72 ALS/FTD, featured by low levels of CSF1R expression and neuroprotective effects. We further show that suppression of CSF1R converts the microglia from a neurotoxic to a neuroprotective state both in vitro and in vivo. In summary, inspired by the diversity of different forms of ALS and non-cell autonomous effect, we identified APC and PIKFYVE inhibition as therapeutic targets for neurons and CSF1R suppression as target for microglia in C9ORF72 ALS/FTD. 12 Chapter 2 : Identification and Therapeutic Rescue of Autophagosome and Glutamate Receptor Defects in C9ORF72 and Sporadic ALS Neurons 2.1 Abstract Amyotrophic lateral sclerosis (ALS) is a fatal motor neuron disease with diverse etiologies. Therefore, the identification of common disease mechanisms and therapeutics targeting these mechanisms could dramatically improve clinical outcomes. To this end, we developed induced motor neuron (iMN) models from C9ORF72 and sporadic ALS (sALS) patients to identify targets that are effective against these types of cases, which together comprise ~90% of patients. We find that iMNs from C9ORF72 and several sporadic ALS patients share two common defects – impaired autophagosome formation and the aberrant accumulation of glutamate receptors. Moreover, we show that an anticoagulation-deficient form of activated protein C, 3K3A-APC, rescues these defects in both C9ORF72 and sporadic ALS iMNs. As a result, 3K3A-APC treatment lowers C9ORF72 dipeptide repeat protein (DPR) levels, restores nuclear TDP-43 localization, and rescues the survival of both C9ORF72 and sporadic ALS iMNs. Importantly, 3K3A-APC also lowers glutamate receptor levels and rescues proteostasis in vivo in C9ORF72 gain- and loss-of-function mouse models. Thus, motor neurons from C9ORF72 and at least a subset of sporadic ALS patients share early defects in autophagosome formation and glutamate receptor homeostasis, and a single therapeutic approach may be efficacious against these disease processes. 13 2.2 Introduction The genetic etiologies that underlie ALS are diverse (Pihlstrøm et al., 2018). This underscores the need to identify therapeutic targets that are broadly efficacious. The identification of pathophysiological processes that are shared between sporadic ALS patients could enable the development of broadly efficacious therapeutics. Genetic, neuropathological, and pharmacological approaches implicate aberrant proteostasis and neuronal excitability in ALS and other neurodegenerative diseases (Wainger et al., 2014; Wen et al., 2014; Menzies et al., 2015; Ichida and Kiskinis, 2015; Shi et al., 2018), but the mechanisms that drive defective proteostasis and hyper- or hypoexcitability in patient neurons are unclear, and it is unknown if there are therapeutic targets that will be capable of mitigating these processes in a multiple different ALS subtype. The lack of models for sporadic ALS has hampered efforts to identify disease mechanisms and therapeutic targets that apply broadly to multiple sporadic forms of ALS (Shi et al., 2018; Van Damme et al., 2017). To address this, we derived induced pluripotent stem cells (iPSCs) from a cohort of sporadic ALS patients and generated induced motor neurons (iMNs) from these iPSCs. We show that similarly to C9ORF72 ALS iMNs, the sporadic ALS iMNs from all six donors degenerate significantly faster than controls and show pronounced TDP-43 mis-localization. Thus, these sporadic ALS iMNs display relevant disease processes. We hypothesized that phenotypic analysis of iMNs from these sporadic ALS patients and C9ORF72 ALS/FTD patients would enable the identification of degenerative mechanisms that are shared by C9ORF72 and at least a subset of sporadic ALS motor neurons. We further hypothesized that pharmacologic agents capable of normalizing these phenotypes in C9ORF72 and sporadic ALS iMNs might also rescue neurodegeneration and lead to the identification of therapeutic targets with broad efficacy in the ALS population. 14 Indeed, we find that C9ORF72 and these sporadic ALS iMNs share defects in autophagosome formation and the aberrant accumulation of glutamate receptors. We find that an engineered, anticoagulation-deficient form of activated protein C called 3K3A-APC, but not heat-inactivated 3K3A-APC (inactive 3K3A-APC), can restore autophagosome formation in both C9ORF72 ALS/FTD and sporadic ALS iMNs. Interestingly, 3K3A-APC also normalizes glutamate receptor levels in C9ORF72 and sporadic ALS iMNs. As a result of these activities, 3K3A-APC lowers dipeptide repeat protein (DPR) levels in C9ORF72 ALS/FTD iMNs, and potently restores nuclear TDP-43 localization and the normal survival of both C9ORF72 and sporadic ALS iMNs. We show that the ability of 3K3A-APC to rescue ALS iMN survival is dependent on its ability to activate Protease-Activated Receptor-1 (PAR1), identifying PAR1 as a therapeutic target for both C9ORF72 and sporadic ALS. Importantly, 3K3A-APC also rescues glutamate receptor levels and proteostasis impairments in vivo. Thus, motor neurons from C9ORF72 and at least a subset of sporadic ALS patients share common defects in autophagosome formation and glutamate receptor homeostasis, and pharmacologic rescue of these defects by PAR1 activation may slow or prevent neurodegeneration in a substantial fraction of ALS cases. 2.3 Results Identification of neurodegenerative phenotypes in iMNs from sporadic ALS patients We and others have shown that iPSC-based models can recapitulate ALS disease processes provided the iPSC donors harbor ALS-causing genetic variants (Wainger et al., 2014; Wen et al., 2014; Shi et al., 2018; Kramer et al., 2018; Donnelly et al., 2013; Zhang et al., 2015; Kiskinis et al., 2014; Moore et al., 2019). As a key example, we have shown that iMNs from control lines consistently survive longer than iMNs from C9ORF72 ALS lines (n=4 controls and n=6 C9ORF72 ALS patients) (Wen et al., 2014; Shi et al., 2018; Kramer et al., 2018). Because all environmental factors are held constant amongst the different lines in the iMN survival assay, we reasoned that 15 sporadic ALS patients whose iMNs degenerate significantly faster than control iMNs likely harbored genetic mutations that promoted their observed motor neuron disease. Therefore, these sporadic ALS iMNs could serve as models of sporadic ALS. To identify sporadic ALS iPSC lines whose iMNs degenerated more rapidly than controls, we generated iPSCs from six sporadic ALS patients. We did not bias our selection of sporadic ALS samples based on genetic or clinical information. Whole genome or exome sequencing and repeat-primed PCR for the C9ORF72 locus showed that the six sporadic ALS patients that did not contain rare variants in known ALS genes, nor a C9ORF72 repeat expansion (Figure S2.1A- B). To confirm our previous findings showing that iMNs derived from patients harboring ALS-causing genetic mutations display rapid neurodegeneration in vitro (Wen et al., 2014; Shi et al., 2018; Kramer et al., 2018), we used transcription factor-mediated conversion (Shi et al., 2018, Son et al., 2011) to generate iMNs from fibroblast-like cells derived from three control and three C9ORF72 ALS patient iPSC lines we previously established and characterized (Figure 2.1A) (Shi et al., 2018). To track individual iMNs, we labeled them using an Hb9::RFP lentivirus that accurately labels HB9+ motor neurons derived in this manner (Wen et al., 2014; Shi et al., 2018; Kramer et al., 2018) (Figure 2.1A and Figure S2.1C) and used robotic microscopy and longitudinal tracking of iMNs to measure their survival (Shi et al., 2018) (Figure 2.1B). The density of Hb9::RFP+ iMNs corresponded to about a 10-25% conversion rate and was similar across the different control and C9ORF72 ALS lines, and iMNs showed consistent survival across the range of iMN densities used in our experiments (Figure S2.1C-E). The vast majority of neurons in iMN cultures were motor neurons, with only rare Tyrosine Hydroxylase+ (dopaminergic neuron marker) or CTIP2+ (cortical neuron marker) neurons detected (Figure S2.1F-I). iMN cultures were supplemented with primary mouse mixed glia to facilitate iMN maturation. Consistent with our previous studies (Wen et al., 2014; Shi et al., 2018), a 12-hour pulse treatment of 10 μ M glutamate induced significantly faster degeneration of C9ORF72 patient iMNs compared 16 to controls (Figure 2.1B (individual lines shown), C (iMNs from all control or patient lines combined into a single trace for clarity) and Figure S2.1J, K). Glutamate treatment was not strictly required to elicit this effect, as withdrawal of neurotrophic factor supplementation also selectively induced the rapid degeneration of C9ORF72 patient iMNs without glutamate treatment (Figure 2.1D and Figure S2.1L). Thus, ALS-causing mutations can cause iMNs to degenerate significantly faster than those from controls. With either pulsatile glutamate treatment or neurotrophic factor withdrawal, iMNs from all six sporadic ALS patients degenerated significantly faster than control iMNs (Figure 2.1B-F and Figure S2.1M, N). iMNs derived from additional iPSC clones from the same sporadic ALS patients also degenerated faster than control iMNs (Figure S2.1O-Q). Thus, iMNs derived from these sporadic ALS patients degenerate significantly faster than controls and at a similar rate to C9ORF72 ALS iMNs. These findings strongly suggest that the donor sporadic ALS patients from which these iMNs were derived harbored genetic variants that contributed to their motor neuron disease. Therefore, iMNs from these patients have the potential to provide insights into mechanisms of neurodegeneration in sporadic ALS cases. A key molecular hallmark of ALS is the nuclear depletion and cytoplasmic accumulation of TDP- 43 in motor neurons (Neumann et al., 2006). The cytosolic accumulation of TDP-43 is thought to drive motor neuron degeneration (Neumann et al., 2006; Barmada et al., 2010). Immunofluoresence verified that both C9ORF72 and sporadic ALS iMNs possessed a lower nuclear-to-cytoplasmic ratio of TDP-43 than controls (Figure 2.1G-I, two controls, two C9ORF72 ALS, and six sporadic ALS patients). Thus, these sporadic ALS iMNs exhibit neurodegenerative processes. C9ORF72 and sporadic ALS iMNs share autophagosome abnormalities that are rescued by 3K3A-APC 17 To identify shared disease mechanisms between C9ORF72 and sporadic ALS motor neurons, we performed phenotypic analysis of patient iMNs. Previous studies have shown that reducing C9ORF72 protein levels using siRNAs impairs autophagosome formation in neurons (Webster et al., 2016; Sellier et al., 2016). To determine if the reduction in C9ORF72 protein levels caused by the C9ORF72 repeat expansion (Shi et al., 2018) is sufficient to block autophagosome formation in motor neurons, we expressed an mRFP-GFP-LC3 construct in iMNs to enable quantification of autophagosomes, which are GFP+/mRFP+, whereas autolysosomes are only mRFP+ due to the lability of GFP at the acidic pH of lysosomes (Kimura et al., 2007). Treatment with 50 nM Bafilomycin to block autophagosome-lysosome fusion significantly increased the number of GFP+/mRFP+ vesicles, indicating that this system could accurately monitor autophagosome number in iMNs (Figure 2.2A (CTRL, –Bafilomycin vs CTRL, +Bafilomycin) and Figure S2.2A, three controls). C9ORF72 ALS iMNs contained significantly fewer GFP+/mRFP+ vesicles than controls (Figure 2.2A, B (CTRL, inactive 3K3A-APC, +Bafilomycin vs C9-ALS, inactive 3K3A-APC, +Bafilomycin), three controls and three C9ORF72 ALS patients), indicating that autophagosome formation was impaired in C9ORF72 ALS iMNs. Interestingly, in the presence of 50 nM Bafilomycin, sporadic ALS iMNs also contained significantly fewer GFP+/mRFP+ vesicles than controls (Figure 2.2A, C and Figure S2.2B, three controls and five sporadic ALS patients), indicating that autophagosome formation was impaired in sporadic ALS iMNs. Thus, both the C9ORF72 and sporadic ALS iMNs display decreased rates of autophagosome formation. C9ORF72 regulates autophagy by promoting the RAB1A-dependent trafficking of the ULK1 autophagy initiation complex to the phagophore (Webster et al., 2016). Consequently, mTOR inhibitors were shown to be ineffective at rescuing autophagosome formation in C9ORF72- deficient neurons and it is unclear if pharmacological agents can restore normal autophagy initiation (Webster et al., 2016). 18 In surveying pharmacological agents that might be capable of promoting autophagosome formation in C9ORF72 ALS iMNs, we identified activated protein C as a possible candidate. Activated protein C is an endogenous blood serine-protease with both anticoagulant and cytoprotective activities, that latter of which is mediated by the activation of Protease-activated Receptor 1 (PAR1) through non-canonical cleavage (Griffin et al., 2018). Zhong and colleagues previously showed that activated protein C treatment can rescue neurodegeneration in SOD1G93A mice (Zhong et al., 2009) and promote autophagosome formation in murine lung epithelial cells (Yen et al., 2013). However, it remained unclear if activated protein C could stimulate autophagosome formation in C9ORF72 ALS iMNs, rescue neurodegeneration in non- SOD1 forms of ALS, or rescue neurodegenerative processes in patient-derived motor neurons. To test these possibilities, we used 3K3A-APC, an engineered form of activated protein C in which three active-site lysine residues required for degrading factor Va and VIIIa are changed to alanine, resulting in a 90% reduction in anticoagulant activity but full retention of PAR1 activation ability (Mosnier et al., 2004). 3K3A-APC is well-tolerated in mice under chronic dosing at levels that activate PAR1 within the central nervous system, and short-term peripheral administration is safe in monkeys (Zhong et al., 2009; Williams et al., 2012b; Lyden et al., 2013). As a negative control for these studies, we used heat-inactivated 3K3A-APC. 3K3A-APC potently rescued autophagosome formation in C9ORF72 ALS iMNs, as determined by its ability to increase the number of GFP+/mRFP+ vesicles in Bafilomycin-treated iMNs (Figure 2.2A, B, (C9-ALS, inactive 3K3A-APC, +Bafilomycin vs C9-ALS, 3K3A-APC, +Bafilomycin), three controls and three C9ORF72 ALS patients). Interestingly, 3K3A-APC also increased autophagosome formation in sporadic ALS iMNs (Figure 2.2A, C, (sALS, inactive 3K3A-APC, +Bafilomycin vs sALS, 3K3A-APC, +Bafilomycin), three controls and five sporadic ALS patients and Figure S2.2B), indicating that is also capable of rescuing C9ORF72-independent autophagy impairments. Western blot analysis of C9ORF72 ALS motor neurons showed that 3K3A-APC increased the ratio of LC3-II:LC3-I in the presence of Bafiloymicin, verifying the results of the 19 mRFP-GFP-LC3 assay (Figure S2.2C, D). Therefore, 3K3A-APC treatment rescues autophagosome formation in C9ORF72 and sporadic ALS iMNs. To determine how 3K3A-APC might stimulate autophagosome formation, we performed RNA sequencing on C9ORF72 ALS iMNs treated with 3K3A-APC or inactive 3K3A-APC for three days (two C9ORF72 ALS patients). Consistent with our findings in the mRFP-GFP-LC3 and immunoblot assays, both Ingenuity Pathway Analysis and KEGG pathway analysis revealed “autophagy” as one of the most significantly enriched gene sets (Figure S2.2E-H). RNA-seq and qRT-PCR showed that 3K3A-APC significantly increased the expression of ATG5 and ATG10, which encode proteins with key roles in driving autophagosome formation (Yamaguchi et al., 2012) (Figure 2.2D (qRT-PCR), Figure S2.2I (RNA-Seq)). Thus, 3K3A-APC increases the expression of genes encoding autophagy-initiating factors in ALS iMNs. Although mTOR inhibitors were ineffective at rescuing autophagosome formation in C9ORF72- deficient cells (Webster et al., 2016), we wondered whether this approach might be effective in sporadic ALS iMNs. Indeed, treatment with 10 μM rapamycin significantly increased autophagosome formation in sporadic ALS iMNs (Figure S2.2J, K). C9ORF72 ALS iMNs have fewer lysosomes than control iMNs (Shi et al., 2018), which could affect autophagic flux downstream of autophagosome formation. To determine if 3K3A-APC affects lysosome number in C9ORF72 ALS iMNs, we examined the number of LAMP2+ vesicles in iMNs (Shi et al., 2018). 3K3A-APC restored normal lysosome numbers in C9ORF72 ALS iMNs (Figure 2.2E, two controls and two C9ORF72 ALS patients, and Figure S2.2L, M). 3K3A-APC did not alter C9ORF72 protein (Figure S2.2N, O, three C9ORF72 ALS patients) or mRNA levels (Figure S2.2P) in C9ORF72 ALS iMNs. Therefore, 3K3A-APC rescues lysosome numbers in C9ORF72 ALS iMNs without altering C9ORF72 levels. iMNs from 3/6 sporadic lines also had significantly fewer LAMP2+ vesicles than control iMNs, and 3K3A-APC treatment rescued the number of lysosomes in these lines (Figure 2.2F and Figure 20 S2.2M, Q). Thus, iMNs from some, but not all sporadic ALS lines display low lysosome numbers. However, for those that do show low lysosome numbers, 3K3A-APC can rescue this phenotype. Consistent with the notion that rescuing lysosome numbers in C9ORF72 ALS iMNs might enable autophagic flux even with increased autophagosome formation, the increase in the ratio of GFP+/mRFP+ vesicles to GFP-/mRFP+ vesicles upon 3K3A-APC treatment was modest compared to the overall increase in GFP+/mRFP+ vesicles in both C9ORF72 and sporadic ALS iMNs (Figure S2.2R, S, 3 C9ORF72 ALS patients and 5 sporadic ALS patients). Thus, 3K3A-APC treatment increases autophagy in C9ORF72 and sporadic ALS iMNs by rescuing both autophagosome formation and lysosome numbers. Rescue of autophagosome abnormalities by 3K3A-APC improves proteostasis To determine if the effects of 3K3A-APC on stimulating autophagy in C9ORF72 and sporadic ALS iMNs were sufficient to lower DPR and cytoplasmic TDP-43 levels, we first measured levels of glycine-arginine (poly(GR)) and proline-arginine (poly(PR)) DPRs in C9ORF72 ALS iMNs with and without 3K3A-APC treatment. Treatment with 3K3A-APC for 6 days reduced the number of nuclear poly(GR)+ and poly(PR)+ punctae in C9ORF72 ALS iMNs to control levels (Figure 2.3A- C and Figure S2.3A-E, two controls and two C9ORF72 ALS patients). Dot blot analysis for poly(GR) confirmed the decrease in iMNs after 3K3A-APC treatment (Figure 2.3D, E). Moreover, 3K3A-APC reduced total nuclear and cytoplasmic levels of poly(GR)+ and poly(PR)+ punctae in C9ORF72 ALS iMNs (Figure S2.3F-H, J-L) and did not significantly alter the ratio of nuclear:cytoplasmic poly(PR)+ punctae, although it slightly lowered the ratio of nuclear:cytoplasmic poly(GR)+ punctae intensity (Figure S2.3I, M). These data indicate that 3K3A-APC treatment decreased both nuclear and cytoplasmic levels of DPRs. qRT-PCR using primers specific for C9ORF72 variants 1 and 3 showed that 3K3A-APC did not reduce RNA levels of the repeat expansion (Figure S2.3N), excluding the possibility that the reduced DPR levels were caused by transcriptional repression. 3K3A-APC also did not alter the expression of genes 21 associated with the integrated stress response (Figure S2.3O), indicating that its ability to lower DPR levels was independent of this pathway. Since 3K3A-APC treatment lowered DPR levels, we wondered if this effect was significant enough to prevent DPR-induced neurodegeneration. To test this, we measured the ability of 3K3A-APC to rescue the degeneration caused by the overexpression of a GR(50)-GFP fusion protein in control iMNs. 3K3A-APC significantly reduced the total intensity of GR(50)-GFP and GR(50)-GFP punctae size indicating that it reduced GR(50)-GFP levels (Figure S2.3P, Q). Moreover, 3K3A- APC increased the survival of GR(50)-GFP-expressing iMNs, indicating that it prevents DPR- induced neurodegeneration (Figure 2.3F). Because autophagy induction can enhance TDP-43 turnover in motor neurons (Barmada et al., 2014), we wondered whether 3K3A-APC treatment could prevent the cytoplasmic accumulation of TDP-43 in ALS iMNs. 3K3A-APC treatment for 6 days significantly increased the nuclear-to- cytoplasmic ratio of TDP-43 in C9ORF72 and sporadic ALS iMNs to control levels (Figure 2.3G- J, two controls, two C9ORF72 ALS, and six sporadic ALS patients). Thus, 3K3A-APC treatment can rescue autophagosome formation, lower DPR levels in C9ORF72 ALS iMNs and rescue DPR-mediated neurotoxicity, and reverse the cytosolic accumulation of TDP-43 in C9ORF72 and sporadic ALS patient iMNs. C9ORF72 and sporadic ALS iMNs have elevated glutamate receptor levels that are normalized by 3K3A-APC Studies in patients and animal and cell models have linked neuronal hyperexcitability to ALS (Wainger et al., 2014; Devlin et al., 2015; Kuo et al., 2004; van Zundert et al., 2008; Vucic and Kiernan, 2006; Vucic and Kiernan, 2006b; Rothstein et al., 1996), but it remains unclear if there are common cell autonomous mechanisms that drive hyperexcitability in different ALS patients. We previously found that reduced C9ORF72 levels lead to elevated surface-bound levels of glutamate receptors in C9ORF72 ALS iMNs, C9ORF72-deficient mice, and C9ORF72 patients, 22 and that these differences were not caused by differential gene expression (Shi et al., 2018). We also showed that the elevated glutamate receptor levels increased sensitivity to glutamate- induced excitotoxicity in vitro and in vivo (Shi et al., 2018). To determine if increased glutamate receptor levels might be a conserved mechanism that drives hyperexcitability in multiple forms of ALS, we examined levels of the NR1 NMDA receptor subunit in neurites of control, C9ORF72 ALS, and sporadic ALS iMNs by immunostaining. Similar to our previous study, C9ORF72 iMNs possessed significantly more NR1+ punctae on neurites than control iMNs (Figure 2.4A, B, two controls and two C9ORF72 ALS patients) (Shi et al., 2018). Co- labeling with a MAP2-specific antibody verified that the NR1+ punctae that were increased in abundance were localized on dendrites (Figure S2.4A, B). Interestingly, sporadic ALS iMNs also displayed more NR1+ punctae on neurites than control iMNs (Figure 2.4A, C and Figure S2.4C- H, two controls and six sporadic ALS patients). Calcium imaging confirmed that C9ORF72 and sporadic ALS iMNs experienced more calcium transients than controls in response to glutamate, indicating that they may be more sensitive to excitotoxicity (Figure 2.4D, E, three controls, three C9ORF72 ALS patients, one sporadic ALS patient). Thus, both C9ORF72 and sporadic ALS iMNs display increased NR1 levels, which could reflect a shared mechanism of increased susceptibility to excitotoxicity. Because glutamate receptor homeostasis is maintained in part through vesicle trafficking (Chen et al., 2007; Seebohm et al., 2012) and we had observed that 3K3A-APC exerted potent effects on autophagosomal and lysosomal pathways in iMNs, we determined if 3K3A-APC could normalize glutamate receptor levels in ALS patient-derived motor neurons. Indeed, 3K3A-APC reduced the number of NR1+ punctae on C9ORF72 and sporadic ALS iMN neurites to control iMN levels (Figure 2.4C and Figure S2.4A, B, two controls, two C9ORF72 ALS, and six sporadic ALS patients). Using surface protein biotinylation, we were able to purify surface-bound proteins from iMNs (Figure 2.4F-I and Figure S2.4I). Immunoblottting confirmed that 3K3A-APC reduced 23 membrane-bound NR1 levels on C9ORF72 and sporadic ALS iMNs (Figure 2.4F-I and Figure S2.4J-M), but not control iMNs (Figure S2.4N-Q). 3K3A-APC treatment did not alter total NR1 levels, but specifically reduced the amount of surface-bound NR1 on ALS iMNs (Figure S2.4J-M, R-W). 3K3A-APC also did not alter total NR1 levels in control iMNs (Figure S2.4X-Z). Thus, 3K3A- APC normalizes NR1 levels on C9ORF72 and sporadic ALS iMNs. 3K3A-APC rescues the survival of C9ORF72 and sporadic ALS iMNs in a PAR1-dependent manner We next determined if 3K3A-APC’s ability to rescue autophagosome production, increase lysosomal numbers, lower levels of DPRs and mislocalized TDP-43, and lower glutamate receptor levels could rescue the degeneration of ALS iMNs. We first tested the ability of the iMN survival assay to replicate the neuroprotective effects previously shown for activated protein C in SOD1G93A mice (Zhong et al., 2009). iMNs derived from an ALS patient carrying an SOD1A4V mutation degenerated significantly faster than control iMNs (Figure S2.5A and Figure S2.5A, three controls and one SOD1A4V ALS patient) and 3K3A-APC rescued their survival (Figure S2.5A and Figure S2.5A, one SOD1A4V ALS patient). Thus, the iMN model mimics disease biology observed in vivo. 3K3A-APC potently rescued the survival of iMNs from all three C9ORF72 ALS lines in a dose- dependent manner (Figure 2.5B, C and Figure S2.5B, C, three C9ORF72 ALS patients). In contrast, 3K3A-APC did not improve the survival of control iMNs (Figure 2.5D and Figure S2.5D, three controls). Activated protein C can mediate neuroprotective effects by activating PAR1 receptor signaling in neurons (Zhong et al., 2009; Zhang et al., 2014) and RNA-seq data confirmed that iMNs express high levels of the F2R gene, which encodes PAR1 (Figure S2.5E). Thus, we measured C9ORF72 ALS iMN survival in the presence or absence of PAR1, PAR2, or PAR3 antisense oligonucleotides or small molecule antagonists. Chemical inhibition of PAR1 or PAR2 alone did not affect C9ORF72 ALS iMN survival (Figure S2.5F, G, two C9ORF72 ALS 24 patients). However, PAR1 antagonist treatment blocked the ability of 3K3A-APC to rescue C9ORF72 ALS iMN survival (Figure 2.5E and Figure S2.5H, two C9ORF72 ALS patients), while a PAR2 antagonist did not (Figure 2.5F and Figure S2.5I, two C9ORF72 ALS patients). Consistent with these findings, ASO-mediated suppression of PAR1 significantly reduced the neuroprotective activity of 3K3A-APC on C9ORF72 ALS iMN survival (Figure 2.5G and Figure S2.5J, K, two C9ORF72 ALS patients). In contrast, ASO-mediated suppression of PAR2 or PAR3 had no effect on the ability of 3K3A-APC to rescue C9ORF72 ALS iMN survival (Figure S2.5L-Q, two C9ORF72 ALS patients). Therefore, 3K3A-APC rescues C9ORF72 ALS iMN survival through activation of PAR1. 3K3A-APC also rescued the survival of iMNs from all six sporadic ALS lines (Figure 2.5H and Figure S2.6A-F, six sporadic ALS patients). When we tested iMNs from separate iPSC clones of the sporadic patient lines, 3K3A-APC also significantly improved their survival, indicating that the rescue effects were not specific to iMNs derived from certain iPSC clones (Figure S2.6G-J, three sporadic ALS patients). Reinforcing the notion that activating autophagosome formation in ALS iMNs could increase survival, rapamycin treatment significantly improved the survival of iMNs from 4/5 sporadic ALS lines tested (Figure S2.7A-J, six sporadic ALS patients). In contrast, rapamycin slightly decreased the survival of control iMNs (Figure S2.7K, L). These results suggest that rescuing autophagosome formation in sporadic ALS iMNs can increase their survival. Treatment with the small molecule PAR1 antagonist alone did not affect the survival of sporadic ALS iMNs (Figure S2.6K-P, six sporadic ALS patients), but the PAR1 antagonist blocked the ability of 3K3A-APC to rescue sporadic ALS iMN survival (Figure 2.5I and Figure S2.6Q-V, six sporadic ALS patients). Thus, 3K3A-APC can rescue C9ORF72 and sporadic ALS iMN survival through activation of PAR1. 25 3K3A-APC rescues C9ORF72 ALS proteostasis and glutamate receptor phenotypes in vivo We next wondered if 3K3A-APC could lower DPR levels in vivo. Baloh and colleagues previously generated C9-BAC mice that harbor a human C9ORF72 gene containing 100-1000 GGGGCC repeats and produce DPRs that aggregate in neurons (O’Rourke et al., 2015). 48 hours after direct injection into the hippocampus, 3K3A-APC-injected hippocampi showed a significant reduction in poly(GR)+, poly(PR)+, and poly(GP)+ punctae (Figure 2.6A-D). Thus, in vivo, 3K3A- APC can reduce levels of both nuclear and cytoplasmically localized DPRs, as well as sense- and antisense transcript-derived DPRs. To determine if 3K3A-APC can normalize glutamate receptor levels on neurons in vivo, we examined its ability to lower NR1 levels and NMDA-induced excitotoxicity in C9orf72+/- mice. We previously showed that hippocampal, cortical, and spinal cord neurons of C9orf72+/- mice have increased NR1 levels compared to C9orf72+/+ mice (Shi et al., 2018). We previously developed an acute, in vivo NMDA-induced excitotoxicity assay in the hippocampus (Figure 2.6E) and verified using cresyl violet staining as well as Fluoro-Jade B–positive dead neuron counting in sham control mice that the operation procedure itself does not result in significant injury (Shi et al., 2018). Similar to iMN cultures, reduced C9ORF72 levels in C9orf72+/- and C9orf72-/- mice cause hippocampal neurons to undergo greater NMDA-induced excitotoxicity than in C9orf72+/+ mice, resulting in a significantly larger loss of neurons in C9ORF72-deficient mice that could be detected by a reduction in cresyl violet staining (Figure 2.6F, G) (Shi et al., 2018). A single injection of 3K3A-APC into the hippocampus of C9orf72+/- mice significantly lowered NR1 levels after 48 hours (Figure 2.6H, I) (n=3 mice). Moreover, when co-injected with NMDA, 3K3A-APC significantly reduced the amount of NMDA-induced hippocampal neurodegeneration in C9orf72+/- mice (Figure 2.6F, G). Thus, in vivo, 3K3A-APC improves proteostasis and normalizes glutamate receptors, thereby rescuing critical C9ORF72 ALS/FTD gain- and loss-of-function disease processes, respectively. 26 2.4 Discussion Our data indicate that motor neurons derived from C9ORF72 and at least a subset of sporadic ALS patients share similar defects in autophagosome formation and glutamate receptor accumulation. Importantly, PAR1 activation by 3K3A-APC can reverse these C9ORF72 and sporadic ALS disease processes in vitro and in vivo (Figure 2.6). We draw several key conclusions from our findings. First, motor neurons derived from six sporadic ALS patients selected without bias toward certain genetic mutations or clinical symptoms display defects in autophagosome formation and accumulate NMDA receptors on their neurites. Both C9ORF72 and sporadic ALS iMNs exhibit these phenotypes, revealing an important area of mechanistic convergence between C9ORF72 ALS and at least some forms of sporadic ALS. A previous study found that in postmortem tissue from sporadic ALS patients, motor neurons may have had lower levels of NMDA receptors (Krieger et al., 1993). However, this may be because the motor neurons containing the highest amounts of NMDA receptors were most vulnerable to degeneration. Alternatively, the patients examined post-mortem in this study may have differed from the patients in our cohort. This highlights the importance of expanding the number of sporadic ALS patient iPSC lines. There are at least three potential reasons for the shared abnormalities in autophagosome formation and glutamate receptor levels we have observed. First, some sporadic ALS patients may harbor genetic variants that directly impair autophagosome formation. If so, this could induce the autophagosome formation defects that in turn could cause excess glutamate receptor accumulation. Alternatively, genetic variants affecting other aspects of proteostasis could lead to the accumulation of unfolded proteins, which might impair autophagosome formation through some unknown mechanism. The accumulation of unfolded protein could also impair overall protein turnover and lead to glutamate receptor accumulation. Third, the sporadic ALS patients 27 could harbor genetic variants that affect vesicular signaling molecules, such as phosphotidylinositols, that regulate both endocytosis of surface proteins and autophagosome formation (Vanhauwaert et al., 2017). It will also be important to determine how conserved these neuronal phenotypes are in other neurodegenerative diseases. One study has clearly shown that a Parkinson’s disease mutation in Synaptojanin impairs autophagosome formation in iPSC- neurons (Vanhauwaert et al., 2017). A second key conclusion is that a single therapeutic approach that rescues the autophagosome and glutamate receptor abnormalities we discovered can slow the degeneration of ALS patient motor neurons. Importantly, we show that this approach is effective for both C9ORF72 and sporadic ALS patients. A third conclusion from our findings is that a single therapeutic approach can rescue both gain- and loss-of-function C9ORF72 disease processes. 3K3A-APC treatment reduces levels of NMDA receptors on iMNs and hippocampal neurons in vivo and rescues NMDA-induced excitotoxicity in C9orf72-deficient mice. Our observation that 3K3A-APC stimulates autophagosome formation indicates that 3K3A-APC can also rescue this aspect of the loss of C9ORF72 function. 3K3A- APC significantly lowers DPR levels in C9ORF72 ALS iMNs and C9ORF72 BAC transgenic mice, thereby mitigating this gain-of-function disease process. A fourth implication of our study is that a single therapeutic agent can reduce levels of DPRs produced from both the sense and antisense C9ORF72 transcripts as well as nuclear and cytoplasmic DPRs, suggesting that it is effective against all DPR species. Our findings provide pre-clinical evidence suggesting that 3K3A-APC and PAR1 activation could rescue neurodegeneration in C9ORF72 and sporadic ALS cases. In complex diseases such as ALS, the identification therapeutic strategies efficacious across a wide range of cases is critical for clinical success. We have developed new tools, including C9ORF72 and sporadic ALS iMN disease models, that enable the identification of these rare, but critical therapeutic strategies. 28 2.5 Figures Figure 2.1 Identification of neurodegenerative phenotypes in iMNs from sporadic ALS patients. (A) Production of Hb9::RFP+ iMNs and survival tracking by time-lapse microscopy. (B, C) Survival of control (CTRL) and C9ORF72 ALS patient (C9-ALS) iMNs with a 12-hour pulse treatment of excess glutamate (shown for each individual line separately (B), or for iMNs from all lines in aggregate (C). For (B, C), n=90 iMNs per line for 3 control and 2 C9-ALS lines, iMNs quantified from 3 biologically independent iMN conversions per line. 29 (D) Survival of control and C9ORF72 ALS patient iMNs after withdrawal of neurotrophic factor supplementation. iMNs from all control or C9ORF72 patient lines shown in aggregate. n=90 iMNs per line for 3 control and 3 C9-ALS lines, iMNs quantified from 3 biologically independent iMN conversions per line. (E, F) Survival of control and sporadic ALS (sALS) patient lines after glutamate treatment (E) or withdrawal of neurotrophic factor supplementation (F). iMNs from all control or C9ORF72 patient lines shown in aggregate. n=90 iMNs per line for 3 control and 6 (E) or 5 (F) sporadic ALS lines, except sALS6 which had 60 (E) or 40 (F) iMNs counted. iMNs quantified from 3 biologically independent iMN conversions per line. (G-I) Immunofluorescence analysis of total TDP-43 (G) and quantification of the ratio of nuclear to cytoplasmic total TDP-43 in control, C9-ALS (H), or sporadic ALS (I) iMNs. Ratio of nuclear to cytoplasmic total TDP-43 in individual iMNs treated with 10 nM inactive 3K3A-APC or 3K3A-APC for 6 days. iMNs from two controls and two C9-ALS patients (H) or four sporadic ALS patients (I) were quantified. n=30 (controls), 30 (C9-ALS), or 36 (sporadic) iMNs per line per condition from two biologically independent iMN conversions of two control, two C9-ALS, or four sporadic ALS lines were quantified. Each grey circle represents a single iMN. Median +/- interquartile range. Unpaired Mann-Whitney test. Scale bars = 5 μm. Dotted lines outline the nucleus and cell body. For iMN survival experiments, significance was measure by two-sided log-rank test using the entire survival time course. The day of differentiation stated on each panel indicates the day of differentiation on which the experimental treatment or time course was initiated. 30 Figure 2.2 C9ORF72 and sporadic ALS iMNs share autophagosome formation abnormalities that are rescued by 3K3A-APC. (A) mRFP-GFP-LC3 fluorescence in control, C9-ALS, or sporadic ALS iMNs treated with or without 50 nM Bafilomycin and 10 nM inactive 3K3A-APC or 3K3A-APC. Bafilo = Bafilomycin. Scale bars = 5 μm. sALS = sporadic ALS. Solid and dotted lines outline the cell body and nucleus, respectively. Cell bodies were visualized using mRFP-GFP-LC3 fluorescence using a longer exposure and increased gain. 31 (B, C) Number of RFP+/GFP+ vesicles per iMN in control, C9-ALS (B) or sporadic ALS (C) iMNs treated with 10 nM inactive 3K3A-APC or 3K3A-APC and 50 nM Bafilomycin for 24 hours. iMNs from 3 controls, 3 C9-ALS (B), and 5 sporadic ALS (C) patients were quantified. n=12 iMNs per line per condition across 2 independent iMN conversions were quantified. Each grey circle represents a single iMN. Median +/- interquartile range. Kruskal-Wallis testing. (D) qRT-PCR analysis of mRNA levels of ATG5 and ATG10 in C9-ALS iMNs treated with 10 nM inactive 3K3A-APC or 3K3A-APC for 3 days. n=4 independent iMN conversions and treatments per condition. Each grey circle represents a single RNA sample. Mean +/- s.d. Two-tailed t-test, unpaired. (E) Number of LAMP2+ vesicles in control or C9-ALS iMNs treated with 10 nM inactive 3K3A- APC or 3K3A-APC for 24 hours. iMNs from 2 controls and 2 C9-ALS patients were quantified. n=21 iMNs per line per condition across 2 independent iMN conversions were quantified. Each grey circle represents a single iMN. Mean +/- s.e.m. One-way ANOVA. (F) Number of LAMP2+ vesicles in control or sporadic ALS iMNs treated with 10 nM inactive 3K3A-APC or 3K3A-APC for 24 hours. Each grey circle represents a single iMN. iMNs from 2 controls and 6 sporadic ALS patients were quantified. n=33 iMNs per line per condition across 2 independent iMN conversions were quantified. Median +/- interquartile range. Kruskal-Wallis testing. The day of differentiation stated on each panel indicates the day of differentiation on which the experimental treatment or time course was initiated. 32 Figure 2.3 Rescue of autophagosome formation abnormalities by 3K3A-APC improves proteostasis. (A-C) Immunostaining (A) and quantification (B, C) to determine endogenous poly(GR)+ punctae in control or C9-ALS iMNs with 10 nM inactive 3K3A-APC or 3K3A-APC treatment for 6 days. Quantified values represent the average number of nuclear poly(GR)+ punctae in n=30 iMNs (controls) or 40-44 iMNs (C9-ALS) per line per condition from two control or two C9-ALS patient 33 lines. For each line, iMNs were quantified from two independent iMN conversions per line per condition. Median +/- interquartile range. Each grey circle represents the number of poly(GR)+ punctae/unit area in a single iMN. Mann-Whitney testing. Solid and dotted lines in (A) outline the cell body and nucleus, respectively. Scale bar = 5 μm. (D, E) Dot blot (D) and quantification (E) of poly(GR)+ levels in iMNs from two C9-ALS patient lines with 10 nM inactive 3K3A-APC or 3K3A-APC treatment for 6 days. Each grey circle represents one dot blot sample. Mean +/- s.d. n=3 independent iMN conversions per line per condition. One-way ANOVA with Tukey correction across all comparisons. (F) Survival of control iMNs without excess glutamate with overexpression of eGFP or GR(50)- eGFP and 10 nM inactive or active 3K3A-APC. n=90 iMNs per condition, iMNs quantified from 3 biologically independent iMN conversions. Two-sided log-rank test, corrected for multiple comparisons, statistical significance was calculated using the entire survival time course. n=90 iMNs per condition. (G-J) Immunofluorescence analysis of total TDP-43 (G) and quantification of the ratio of nuclear to cytoplasmic total TDP-43 in control, C9-ALS (H), or sporadic ALS iMNs (I, J). Ratio of nuclear to cytoplasmic total TDP-43 in individual C9-ALS iMNs treated with 10 nM inactive 3K3A-APC or 3K3A-APC for 6 days. iMNs from two controls, two C9-ALS, and four sporadic ALS patients were quantified. n=30 iMNs per line (control and C9-ALS) per condition or n=26 iMNs (I), 30 iMNs (J) (inactive 3K3A-APC) or n=35 iMNs (J) (3K3A-APC)(sporadic ALS) per condition per line from two biologically independent iMN conversions were quantified. Each grey circle represents a single iMN. For (H), median +/- interquartile range. Kruskal-Wallis testing. For (I, J), mean +/- s.e.m. Unpaired t-test with Welch’s correction. Scale bars = 5 μm. Dotted lines outline the nucleus and cell body. The day of differentiation stated on each panel indicates the day of differentiation on which the experimental treatment or time course was initiated. 34 Figure 2.4 C9ORF72 and sporadic ALS iMNs have elevated glutamate receptor levels that are normalized by 3K3A-APC. (A) Immunofluorescence images showing NR1+ punctae on neurites of iMNs treated with 10 nM inactive 3K3A-APC or 3K3A-APC for 6 days. Scale bar: 2 µm. This experiment was repeated 3 times with similar results. (B, C) NR1+ punctae per unit area in control, C9-ALS (B) or sporadic ALS (C) iMNs. Each grey circle represents the number of NR1+ punctae per area unit on a single neurite (one neurite 35 quantified per iMN). n=33 (controls, C9-ALS) or 13 (sporadic) iMNs quantified per line per condition from two biologically independent iMN conversions of two CTRL, two C9-ALS, or six sporadic ALS lines. Median +/- interquartile range. Kruskal-Wallis testing. (D) Number of calcium transients per 30 seconds in control or C9-ALS iMNs treated with 10 nM inactive 3K3A-APC or 3K3A-APC. n=21 iMNs per line per condition from three biologically independent iMN conversions of three CTRL and three C9-ALS lines. For the C9-ALS + 3K3A- APC condition, n=19 iMNs per line. Median +/- interquartile range. Kruskal-Wallis testing. (E) Number of calcium transients per 30 seconds in control or sporadic ALS iMNs treated with inactive 3K3A-APC or 3K3A-APC. n=20 iMNs per line per condition from three biologically independent iMN conversions of three CTRL and one sporadic line. Median +/- interquartile range. Kruskal-Wallis testing. (F) Immunoblotting of surface NR1 after surface protein biotinylation in C9-ALS iMNs generated with NGN2, ISL1, and LHX3 and treated with 10 nM inactive 3K3A-APC or 3K3A-APC for 6 days. TF = Transferrin. (G) Quantification of NR1 immunoblotting from (F). n=4 biologically independent iMN conversions. Each grey circle represents an individual sample. The ratio of surface to total Transferrin Receptor was used to normalize for the membrane protein extraction efficiency and TUJ1 was used to normalize for neuron number. (H) Immunoblotting of surface NR1 after surface protein biotinylation in sporadic ALS iMNs (one patient) generated with NGN2, ISL1, and LHX3 and treated with 10 nM inactive 3K3A-APC or 3K3A-APC for 6 days. TF = Transferrin. The full blot for total TUJ1 is shown. (I) Quantification of NR1 immunoblotting from (H). n=4 biologically independent iMN conversions. Each grey circle represents an individual sample. The ratio of surface to total Transferrin Receptor was used to normalize for the membrane protein extraction efficiency and TUJ1 was used to normalize for neuron number. The day of differentiation stated on each panel indicates the day of differentiation on which the experimental treatment or time course was initiated. 36 Figure 2.5 3K3A-APC rescues the survival of C9ORF72 and sporadic ALS iMNs in a PAR1-dependent manner. (A) Survival of iMNs from 3 control or 1 SOD1A4V ALS patient lines in excess glutamate with 10 nM inactive 3K3A-APC or 3K3A-APC. n=90 iMNs per line per condition, iMNs from all control lines shown in aggregate for clarity. iMNs quantified from 3 biologically independent iMN conversions per line. (B) Survival of iMNs from 2 C9-ALS lines in excess glutamate with inactive 3K3A-APC or different concentrations of 3K3A-APC. n=90 iMNs per line per condition, iMNs from both lines shown in aggregate for clarity. iMNs quantified from 3 biologically independent iMN conversions per line. (C) C9-ALS iMNs at day 12 of survival in excess glutamate with inactive 3K3A-APC or 3K3A-APC treatment. This experiment was repeated three times with similar results. Scale bar = 100 μm. 37 (D) Survival of control iMNs in excess glutamate with 10 nM inactive 3K3A-APC or 3K3A-APC, n=90 iMNs per line per condition for 3 control and 3 C9-ALS lines, iMNs quantified from 3 biologically independent iMN conversions per line. (E-G) Survival of iMNs from 2 C9-ALS lines in excess glutamate with 3K3A-APC with or without 3 μM PAR1 antagonist treatment (E) or PAR2 antagonist treatment (F). n=90 iMNs per line per condition, iMNs from both lines shown in aggregate for clarity. iMNs quantified from 3 biologically independent iMN conversions per line. Survival of iMNs from 2 C9-ALS lines in excess glutamate with 3K3A-APC with or without 9 μM PAR1 ASO treatment (G). n=90 iMNs per line per condition, iMNs from both lines shown in aggregate for clarity. iMNs quantified from 3 biologically independent iMN conversions per line. Each trace includes neurons from 2 donors with the specified genotype. All iMN survival experiments were analyzed by two-sided log-rank test and corrected for multiple comparisons if applicable. Statistical significance was calculated using the entire survival time course. (H, I) Survival of iMNs from sALS clone 1 in excess glutamate with 10 nM inactive 3K3A-APC or 3K3A-APC (H), or with 3K3A-APC and DMSO or a PAR1 antagonist (I). n=90 iMNs per condition. iMNs quantified from 3 biologically independent iMN conversions. For all iMN survival experiments, significance was measure by two-sided log-rank test using the entire survival time course. The day of differentiation stated on each panel indicates the day of differentiation on which the experimental treatment or time course was initiated. 38 Figure 2.6 3K3A-APC rescues C9ORF72 ALS proteostasis and glutamate receptor phenotypes in vivo. (A) Overview of the experimental procedure for testing the ability of 3K3A-APC to reduce DPR levels in the hippocampus of C9-BAC mice. (B-D) The effect of 10 nM inactive 3K3A-APC or 3K3A-APC on the level of poly(GR)+ punctae in the dentate gyrus of C9-BAC mice. Mean +/- s.d. of the number of poly(GR)+ (B), poly(GP)+ (C), and poly(PR)+ (D) punctae per cell, each data point represents a single cell. Cells quantified from 3 mice per condition, one-way ANOVA with Tukey correction for all comparisons, Scale bars = 10 39 μm, dotted lines outline cell bodies. Neuronal area was determined by manual outlining in ImageJ on basis of the staining pattern provided by Tuj1 or Map2. (E) Overview of the experimental procedure for inducing NMDA-injury in the hippocampus and testing the ability of 3K3A-APC to mitigate this injury. (F, G) The effect of 0.2 µg of 3K3A-APC delivered in a volume of 0.3 µl on NMDA-induced hippocampal injury in C9orf72+/+, and C9orf72+/- mice. (Mean +/- s.e.m. of n=3 mice per condition, one-way ANOVA with Tukey correction across all comparisons, red dashed lines outline the injury sites (F)). Vehicle control conditions were published in a previous study (4). (H, I) Immunostaining (H) and quantification (I) of NR1 levels in C9orf72+/- mice treated with vehicle or 0.2 µg of 3K3A-APC delivered in a volume of 0.3 µl (n=3 mice per condition, 72 cells quantified per condition). Each grey data point represents a single cell. Mean +/- interquartile range. Mann-Whitney test. 40 Chapter 3 : PIKFYVE Inhibition Mitigates Disease in Models of Diverse Forms of ALS 3.1 Abstract Amyotrophic lateral sclerosis (ALS) is a fatal neurodegenerative disease that results from many diverse genetic causes. Although therapeutics specifically targeting known causal mutations may rescue individual types of ALS, these approaches cannot treat most cases since they have unknown genetic etiology. Thus, there is a pressing need for new therapeutic strategies that rescue multiple forms of ALS, particularly those with unknown etiologies. Misfolded protein accumulation induces neurodegeneration and is a common feature of ALS, but key aspects of canonical proteostasis pathways such as lysosomal function decline with age and may be difficult to stimulate effectively. Here, we show that pharmacological inhibition of PIKFYVE kinase activates an unconventional proteostasis mechanism involving exocytosis of aggregation-prone proteins. PIKFYVE inhibition ameliorates ALS pathology and motor dysfunction and extends survival of animal models and patient-derived motor neurons representing diverse forms of ALS. Thus, reducing PIKFYVE activity may be a broadly effective therapeutic approach for ALS. 3.2 Introduction ALS is a fast-progressing neurodegenerative disease characterized by motor neuron loss, paralysis, and death within 2-5 years of onset (Mejzini et al., 2019). Similar to other prominent neurodegenerative diseases such as Alzheimer’s disease and frontotemporal dementia (FTD), ALS has many diverse genetic etiologies (Pihlstrøm et al., 2017). Since each genetic form is relatively rare and the etiology is unknown for most cases, a central goal in treating ALS is the identification of common pathways that can rescue multiple forms of ALS, particularly those with unknown genetic etiologies. 41 The accumulation of misfolded proteins can induce neuron death and is a common feature of neurodegenerative diseases (Barmada et al., 2010; Mejzini et al., 2019; Sweeney et al., 2017). Canonical proteostasis mechanisms including the ubiquitin-proteosome system and autophagy may be difficult to stimulate effectively since key aspects of these pathways, such as lysosomal function, decline in neurons with age and are sometimes compromised by neurodegenerative disease-causing mutations (Barmada et al., 2014; Kaushik and Cuervo, 2015; Mejzini et al., 2019). Intriguingly, recent studies suggest that neurons can sometimes utilize a third, unconventional protein clearance approach involving exocytosis of aggregation-prone proteins (Iguchi et al., 2016; Melentijevic et al., 2017). By bypassing lysosomal degradation in neurons, this pathway could present an effective alternative to macroautophagy in neurodegenerative disease settings. However, it remains unclear if stimulating exocytosis of aggregation-prone proteins could be used to treat neurodegenerative diseases. Here, we show that PIKFYVE kinase inhibition is broadly efficacious and extends the survival of induced motor neurons (iMNs) derived from C9ORF72, TARDBP, FUS, and sporadic ALS patients. Suppressing PIKFYVE activity also increases neuromuscular junction (NMJ) and motor function in C9ORF72 and TDP-43 D. melanogaster and C. elegans ALS models. In TDP-43- overexpressing mice, deleting one copy of Pikfyve or suppressing Pikfyve using an antisense oligonucleotide (ASO) improves motor function, ameliorates TDP-43 pathology, and extends life span. PIKFYVE inhibition stimulates the exocytosis of neurotoxic proteins including C9ORF72 dipeptide repeat proteins (DPRs) and phosphorylated TDP-43 (pTDP-43) from ALS patient induced motor neurons (iMNs), reduces intracellular DPR and pTDP-43 pathology, and increases TDP-43 splicing activity. Pharmacologically blocking exosome generation or release largely neutralizes the therapeutic effect of PIKFYVE inhibition in human iMNs and TDP-43 mice, confirming that exocytosis is required for efficacy. Thus, PIKFYVE inhibition and activating the exocytosis of pathological proteins from neurons may mitigate disease processes in diverse forms of ALS. 42 3.3 Results PIKFYVE inhibition ameliorates C9ORF72 ALS/FTD neurodegeneration We previously showed that C9ORF72 ALS/FTD iMNs possess fewer lysosomes than controls and that this accelerates DPR aggregation, TDP-43 mis-localization, and neurodegeneration (Shi et al., 2018). In a phenotypic chemical screen to identify targets that extend C9ORF72 ALS/FTD iMN survival, we found that PIKFYVE kinase inhibitors such as apilimod were among the most efficacious compounds (Shi et al., 2018). PIKFYVE is a lipid kinase that converts phosphatidylinositol-3-phosphate (PI3P) to phosphatidylinositol-3,5-bisphosphate (PI(3,5)P2) (Zolov et al., 2012). A portion of PI(3,5)P2 resides on autophagosomal, endosomal, and lysosomal membranes and regulates vesicle fusion and trafficking (de Lartigue et al., 2009). PIKFYVE inhibition lowered DPR levels in C9ORF72 ALS/FTD iMNs and a 48 hour treatment of apilimod reduced poly(GR)+ punctae in C9ORF72-BAC mice (Shi et al., 2018). However, how PIKFYVE inhibition ameliorates C9ORF72 ALS/FTD iMN proteinopathy and degeneration, if it can mitigate neurodegeneration in other forms of ALS, and if this approach is effective in chronic in vivo models remains unknown. To confirm that PIKFYVE inhibition improves C9ORF72 ALS/FTD iMN survival, we repeated our initial in vitro efficacy studies and extended our analyses to additional DPR proteins and TDP-43 loss-of-function phenotypes. To this end, we expressed Ngn2, Isl1, Lhx3, Ascl1, Brn2, Myt1l, NeuroD1, and a lentiviral Hb9::RFP reporter to convert control and C9ORF72 ALS/FTD induced pluripotent stem cells (iPSCs) into Hb9::RFP+ iMNs as we published previously (Figure 3.1A; Figure S3.1A) (Shi et al., 2018, 2019). We previously characterized all iPSC lines for pluripotency and the presence/absence of the C9ORF72 repeat expansion (Shi et al., 2019). iPSC lines were karyotypically normal, and control and patient lines generated iMNs with similar efficiency (Figure S3.1B). Since iMNs fire repetitive action potentials, respond appropriately to neurotransmitters, and form functional NMJs at day 17 of differentiation, we used this time point for experiments (Shi 43 et al., 2018). C9ORF72 ALS/FTD iMNs displayed prominent poly(glycine-arginine (GR))+ and poly(proline-arginine (PR))+ punctae, whereas iMNs from control lines did not (Figure S3.1C, D). Removal of the repeat expansion from a C9ORF72 ALS/FTD line using CRISPR-Cas9 editing eliminated the detection of poly(GR)+ and poly(PR)+ punctae and confirmed the specificity of the dipeptide repeat protein (DPR) antibodies (Figure S3.1E-H). C9ORF72 ALS/FTD iMNs also displayed TDP-43 mis-localization into the cytoplasm, further indicating that the repeat expansion caused protein pathology known to drive neurodegeneration (Figure S3.1I) (Barmada et al., 2010; Shi et al., 2018, 2019; Wen et al., 2014). Similar to our previous studies, we used robotic microscopy and longitudinal tracking of individual iMNs to determine their survival (Kramer et al., 2018; Maor-Nof et al., 2021; Shi et al., 2018, 2019; Yamada et al., 2019). This approach prevents overestimation of iPSC-neuron survival by distinguishing between existing neurons and new neurons that form during the survival experiment. In an effort to preserve the biological relevance of this assay, we did not add exogenous stressors to the cultures that would have induced larger differences between patient and control iMNs but would have also introduced additional neurotoxic stimuli. Instead, we opted to simply withdraw neurotrophic factor supplementation. When stressed by neurotrophic factor withdrawal, C9ORF72 ALS/FTD iMNs degenerated faster than controls (Figure 3.1B, C; Figure S3.1J, K, n=3 control, 3 C9ORF72 donors). To facilitate quantitative comparisons between multiple groups, we used the Mantel-Haenszel method to assess the iMN survival data over the entire time course and calculate it as a hazard rate (Muenz et al., 1977). In the Mantel-Haenszel method, the hazard rate is defined as the likelihood of death for a neuron during a specified time interval (Muenz et al., 1977). By specifying the time interval as the entire survival time course, hazard rate analysis allowed detection of viability differences at the beginning, middle, or end of the survival assay and, in comparison to Kaplan-Meier survival plots, provided easier visualization of differences between multiple groups. To facilitate comparisons, we normalized the hazard rate for each group to a control condition to generate a hazard ratio. For C9ORF72 ALS/FTD iMNs, 44 this resulted in a hazard ratio, or the hazard rate of C9ORF72 ALS/FTD iMNs relative to the hazard rate of the control group, or control iMNs in this case, >1 (Figure 3.1C). In this analysis, a hazard ratio of 2 would indicate that C9ORF72 iMNs died twice as fast as controls during the survival assay. Similar to results we obtained previously in cultures stressed by treatment with excess glutamate, the small molecule PIKFYVE inhibitor apilimod improved the survival of C9ORF72 ALS/FTD iMNs and decreased their hazard rates in a dose-dependent manner, while not affecting control iMN survival (Figure 3.1D, E; Figure S3.1L-N, n=3 control, 3 C9ORF72 donors) (Shi et al., 2018). Multiple ASOs that suppressed PIKFYVE also increased ALS iMN survival in a dose-dependent manner, confirming that reducing PIKFYVE activity slowed C9ORF72 ALS/FTD iMN degeneration (Figure 3.1F; Figure S3.1O-R). Indeed, PIKFYVE inhibition or suppression improved C9ORF72 ALS/FTD iMN survival to or near the level of control iMNs in multiple cases (Figure 3.1D-F; Figure S3.1P-R). To examine whether PIKFYVE inhibition improves C9ORF72 ALS/FTD motor neuron function, we tested apilimod in Drosophila larvae overexpressing poly(GR), which display reduced NMJ active zones and impaired locomotion as measured by larval turning (Figure 3.1G; Figure S3.1S, T) (Perry et al., 2017). Apilimod treatment for 5 days throughout larval development ameliorated the locomotor defect and increased the number of NMJ active zones in poly(GR)-expressing Drosophila larvae, indicating that PIKFYVE inhibition improves motor neuron function in this 3. ALS/FTD model (Figure 3.1G; Figure S3.1S, T). PIKFYVE inhibition ameliorates disease pathology in iMNs Although we had previously found that apilimod could reduce the glutamate-induced hyperactivity we observed in C9ORF72 ALS/FTD iMNs, we had only observed hyperactivity in the presence of excess glutamate, which we did not use in the current study (Shi et al., 2018). Thus, we hypothesized that PIKFYVE inhibition might act through another mechanism to mitigate neurodegeneration in this context. Strikingly, gene ontology analysis of bulk RNA-seq data 45 comparing C9ORF72 ALS/FTD iMNs treated with DMSO or apilimod indicated that 4 of the top most differentially-expressed gene categories were autophagy- or vesicle-related (Figure S3.2A). In addition, amongst all publicly-available transcription factor perturbation gene sets, the gene expression changes induced by apilimod were most similar to overexpression of TFEB, a key transcriptional regulator of autophagy (Figure S3.2A) (Di Malta et al., 2019). These data suggested that PIKFYVE inhibition might affect proteostasis in C9ORF72 ALS/FTD iMNs. At autopsy, ~97% of ALS patients harbor TDP-43 pathology in neurons of affected regions, which is characterized by nuclear depletion of TDP-43, increased cytoplasmic pTDP-43, and cytoplasmic inclusions comprised largely of pTDP-43 (Neumann et al., 2009; Suk and Rousseaux, 2020). Accumulation of cytoplasmic pTDP-43 is neurotoxic and accelerates the loss of nuclear TDP-43, which further impairs neuronal function (Gasset-Rosa et al., 2019; Melamed et al., 2019). Transcripts of several TDP-43 target genes such as STMN2 and UNC13A show increased cryptic exon incorporation and consequently, lower mRNA levels, in neurons exhibiting loss of TDP-43 nuclear function due to aggregation in the cytoplasm (Brown et al., 2022; Klim et al., 2019; Ma et al., 2022; Melamed et al., 2019). Consistent with our immunocytochemical analysis showing that C9ORF72 ALS/FTD iMNs exhibited mislocalization of TDP-43 into the cytoplasm, qRT-PCR analysis indicated that C9ORF72 ALS/FTD iMNs displayed a higher proportion of truncated STMN2 transcripts containing the TDP-43-regulated cryptic exon compared to control iMNs (Figure 3.2A). Importantly, apilimod treatment reduced STMN2 cryptic exon inclusion and increased total transcript levels for STMN2 and UNC13A (Figure 3.2A; Figure S3.2B). These results suggested that PIKFYVE inhibition might ameliorate TDP-43 pathology and improve TDP- 43 function in C9ORF72 ALS/FTD iMNs. Based on these data, we hypothesized that PIKFYVE inhibition slowed neurodegeneration by improving proteostasis and ameliorating ALS pathology in iMNs. To test this hypothesis, we examined TDP-43 and DPR pathology in C9ORF72 ALS/FTD iMNs. A 24-hr apilimod treatment significantly lowered cytoplasmic TDP-43 and restored the 46 nuclear:cytoplasmic TDP-43 ratio in C9ORF72 ALS/FTD iMNs to control iMN levels (Figure 3.2B- D, n=2 C9ORF72 ALS/FTD donors). Although the differences in nuclear:cytoplasmic TDP-43 ratio between control and C9ORF72 ALS/FTD iMNs or C9ORF72 ALS/FTD iMNs with and without apilimod were modest, they were similar to the shift in iMN TDP-43 localization induced by ALS- associated mutations in VCP (Figure 3.2B-D, n=2 C9ORF72 ALS/FTD donors) (Harley et al., 2021). Moreover, these shifts in TDP-43 mis-localization were correlated with a significant reduction in STMN2 cryptic exon inclusion and increases in total STMN2 and UNC13A transcript levels, suggesting they had functional consequences (Figure 3.2A; Figure S3.2B). Poly(GR) and poly(PR) are highly neurotoxic C9ORF72 DPRs detectable in the nuclei and cytoplasm of spinal motor neurons in postmortem tissue and iMNs (Wen et al., 2014) (Figure S3.1C, D, H). We confirmed the specificity of our poly(GR) and poly(PR) antibodies in iMNs using an isogenic control line in which the repeat expansion was removed from a C9ORF72 ALS/FTD line by CRISPR/Cas9-editing (Figure S1C-H). A 24-hr apilimod treatment significantly reduced the number of poly(GR)+ and poly(PR)+ punctae in C9ORF72 ALS/FTD iMNs (Figure 3.2E, F; Figure S3.2C, n=2 C9ORF72 donors). Poly(GA) is a highly abundant DPR that localizes predominantly to the cytoplasm and although it is less neurotoxic than poly(GR) and poly(PR) on its own, it may modify the toxicity of other DPRs in vivo (Nguyen et al., 2020; Wen et al., 2014). C9ORF72 ALS/FTD iMNs possessed elevated levels of poly(GA) and apilimod treatment significantly reduced poly(GA)+ punctae (Figure S3.2D). Thus, PIKFYVE inhibition can ameliorate TDP-43 and DPR pathology in C9ORF72 ALS/FTD iMNs. PIKFYVE inhibition stimulates exocytosis of aggregation-prone proteins from iMNs Since autophagy-related genes were enriched in the gene sets differentially-expressed between vehicle- and apilimod-treated iMNs, we determined if PIKFYVE inhibition increased autophagy (Figure S3.2A). We transduced C9ORF72 ALS/FTD iMNs with a lentiviral mRFP-GFP-LC3 construct that labels autophagosomes as GFP+ and mRFP+, whereas autolysosomes only 47 fluoresce mRFP+ due to the lability of GFP at the acidic pH of lysosomes (Figure S3.3A) (Kimura et al., 2007; Shi et al., 2019). In the presence of bafilomycin, which blocks autophagic flux, apilimod treatment increased the number of mRFP+/GFP+ autophagosomes by about 3-fold, indicating that PIKFYVE inhibition stimulated autophagosome production (Figure S3.3A, B) (Mauvezin and Neufeld, 2015). Although our RNA-seq analysis indicated that gene expression changes resulting from apilimod treatment were similar to those induced by TFEB overexpression, we did not observe increased TFEB levels in the nucleus or elevated phosphorylated TFEB in apilimod-treated iMNs (Figure S3.3C, D). In addition, we did not see clear activation of lysosomal TFEB target genes when we inhibited or suppressed PIKFYVE in C9ORF72 ALS/FTD iMNs. Thus, mechanisms leading to increased autophagosome production upon PIKFYVE inhibition most likely occurred downstream of TFEB. In the absence of bafilomycin, apilimod increased the number of mRFP+/GFP+ autophagosomes by the same magnitude as it did with bafilomycin treatment, about 3-fold, suggesting that PIKFYVE inhibition did not increase autophagic flux (Figure S3.3A, B). Consistent with this notion, in the absence of bafilomycin, apilimod increased the ratio of mRFP+/GFP+ autophagosomes to total mRFP+ (mRFP+/GFP+ and mRFP+/GFP-) vesicles and the ratio of LC3-II to LC3-I by a slightly greater amount than it increased the number of autophagosomes with bafilomycin, suggesting that it modestly decreased autophagic flux (Figure S3.3A, B, E, F). Supporting this notion, apilimod significantly decreased the number of mRFP+/GFP- autolysosomes in the absence of bafilomycin (Figure S3.3G). This is reminiscent of other studies that found that PIKFYVE inhibition can decrease autophagic flux (Hessvik et al., 2016; Martin et al., 2013). The decreased autolysosome production in apilimod-treated C9ORF72 ALS/FTD iMNs did not result from a lack of acidified lysosomes since lysosensor green, which fluoresces in an acidic environment, positively labeled more lysotracker+ vesicles after PIKFYVE inhibition (Figure S3.3H). Moreover, co-treating with bafilomycin during apilimod treatment failed to prevent PIKFYVE inhibition from increasing C9ORF72 ALS/FTD iMN survival (Figure S3.3I). Thus, 48 PIKFYVE inhibition did not increase macroautophagy, nor was macroautophagy required apilimod’s ability to increase C9ORF72 ALS/FTD iMN survival. Examining the LC3-mRFP-labeled vesicles showed that the total number of LC3-mRFP+ vesicles (mRFP+/GFP+ autophagosomes and mRFP+/GFP- autolysosomes) did not increase in C9ORF72 ALS/FTD iMNs upon apilimod treatment (Figure S3.3J). This was unexpected given that PIKFYVE inhibition increased autophagosome production and lowered autophagic flux, and it suggested that autophagosomes might leave the cell or be degraded through a route other than macroautophagy (Figure S3.3A-G). A previous study showed that in an immortalized cell line, PIKFYVE inhibition can activate an unconventional form of secretion in which autophagosome-associated proteins, cytosolic aggregation-prone proteins, and leaderless proteins are preferentially exocytosed (Hessvik and Llorente, 2018; Hessvik et al., 2016; Ponpuak et al., 2015). The authors’ data suggested that PIKFYVE inhibition-induced exocytosis was possibly due to a reduction in the fraction of autophagosomes and multivesicular bodies that fused with lysosomes, reminiscent of the ability of bafilomycin and ammonium chloride to increase exosomal secretion of ⍺-synuclein (Alvarez- Erviti et al., 2011; Hessvik et al., 2016). Thus, we hypothesized that PIKFYVE inhibition might mitigate C9ORF72 ALS/FTD neurodegeneration by activating exocytosis to clear pathological aggregation-prone proteins (Figure 3.3A). To determine whether apilimod treatment activates exocytosis in iMNs, we assessed extracellular vesicle release in iMN supernatants using a polyethylene glycol/ultracentrifugation method that eliminates most non-vesicular material and is scalable enough to accommodate multiple conditions and cell lines (Ludwig et al., 2018). Proteins associated with the endosomal sorting complex required for transport (ESCRT), such as TSG101, facilitate cargo loading into exosomes and are therefore commonly used as exosomal markers (Hessvik et al., 2016; Lee et al., 2019; Willms et al., 2016). After normalizing extracellular vesicle fractions by total protein content and controlling for differences in neuron death between samples using TUJ1 levels in the cell pellet 49 fraction, immunoblotting showed that a 24-hour apilimod treatment increased TSG101 levels in the extracellular vesicle fractions isolated from control and C9ORF72 ALS/FTD iMN supernatants (Figure 3.3B, C, quantification of n=3 control, 3 C9ORF72 donors; Figure S3.3K). Supporting the notion that the elevated extracellular TSG101 levels reflected increased exocytosis, co-treatment with GW4869, a small molecule Neutral Sphingomyelinase 2 inhibitor that blocks exosome release by restricting the ceramide-dependent intraluminal vesicle formation within multivesicular bodies and restricting their fusion with the plasma membrane, abolished the ability of apilimod to increase TSG101 levels in the supernatant (Figure 3.3B, C, quantification of n=3 control, 3 C9ORF72 donors) (Iguchi et al., 2016; Kosaka et al., 2010). These findings were consistent with our hypothesis that inhibiting PIKFYVE might stimulate exosome release from iMNs. To directly visualize exosome release, we overexpressed GFP in iMNs, isolated the exosome fraction after treatment with DMSO or apilimod, labeled this fraction with carboxyfluorescein succinimidyl ester (CFSE)-far red dye which crosses lipid membranes and labels free amines, and performed flow cytometry analysis (Lyons, 2000). A 24-hour apilimod treatment elicited a significant increase in the number of GFP+/CSFE+ particles in the iMN supernatant (Figure S3.3L). Electron microscopy confirmed that the particles secreted by C9ORF72 ALS/FTD iMNs upon apilimod treatment were exosomes of ~100 nm in diameter (Figure 3.3D) (Zaborowski et al., 2015). Together, these data indicate that apilimod increases the release of TSG101+ exosomes from iMNs. We next performed mass spectrometry on exosomal samples that contained the same total protein levels for each sample and were controlled for cell death differences using TUJ1 levels in the cell pellet fraction (Figure S3.3K). This analysis revealed that compared to exosomes from vehicle-treated iMNs, exosomes released from control and C9ORF72 ALS/FTD iMNs upon apilimod treatment showed enrichment of autophagosome-associated proteins such as LC3, OPTN, and p62, aggregation-prone proteins linked to neurodegenerative disease such as ⍺- synuclein, Huntingtin, SOD1, and TDP-43, and leaderless proteins (Figure 3.3E; Figure S3.3M). In contrast, exosomes released upon apilimod treatment showed de-enrichment of proteins 50 associated with Ras signaling and cell-cell junctions compared to exosomes from vehicle-treated iMNs (Figure 3.3E). Since we loaded the same amount of total protein for each sample in the mass spectrometry analyses, these findings reflected true enrichment/de-enrichment of these proteins in the exosomal fraction. Moreover, for proteins that were enriched in the exosomal fraction upon apilimod treatment, these results are not likely to be explained by elevated cell death and release of intracellular contents upon apilimod treatment since we controlled for cell death by monitoring TUJ1 levels in the cell pellet fraction and PIKFYVE inhibition improved C9ORF72 ALS/FTD iMN survival and did not greatly affect control iMNs (Figure 3.1D-F; Figure S3.3K). Immunoblotting of exosomes released from control and C9ORF72 ALS/FTD iMNs confirmed these results, showing that apilimod increased the secretion of LC3-II, OPTN, p62, ⍺-synuclein and Huntingtin (Figure 3.3F-I, Figure S3.3N, n=3 control, 3 C9ORF72 donors). Comparing immunoblots from exosomal and cell pellet fractions showed that apilimod treatment resulted in an exosomal:cell pellet ratio of ~2:1 and ~0.4:1 for TSG101 and OPTN, respectively, indicating that PIKFYVE inhibition induced exocytosis of large fractions of these proteins (Figure 3.3B, H; Figure S3.3O). Thus, PIKFYVE inhibition stimulates the exocytosis of autophagy-associated and aggregation-prone proteins from C9ORF72 ALS/FTD iMNs. To determine if proteins that drive neurodegeneration in ALS are substrates for exocytosis induced by PIKFYVE inhibition, we examined their secretion from ALS/FTD iMNs upon apilimod treatment. Immunoblotting showed that apilimod treatment significantly increased the secretion of phosphorylated TDP-43 (Ser409/410) (pTDP-43) from C9ORF72 ALS/FTD iMNs, but not controls (Figure 3.3J, K). Apilimod also increased the release of total TDP-43 from C9ORF72 ALS/FTD iMNs to a much greater extent than from control iMNs (Figure S3.3P). Co-administration of GW4869 severely reduced pTDP-43 and total TDP-43 release, confirming that pTDP-43 secretion was mediated through exocytosis (Figure 3.3J, K, quantification from n=3 control, 3 C9ORF72 donors; Figure S3.3P). Comparing immunoblots from exosomal and cell pellet fractions showed that apilimod treatment resulted in an exosomal:pellet ratio of ~0.4:1 of pTDP-43, about 51 a 10-fold increase over the DMSO-treated condition (Figure 3.3L). In addition, immunoblotting of the cytoplasmic fraction that was isolated from cell pellets and verified for purity using Fibrillarin and HSP90 showed that apilimod treatment reduced cytoplasmic pTDP-43 in C9ORF72 ALS/FTD iMNs to control iMN levels (Figure 3.3M, N, Figure S3.3Q). Thus, PIKFYVE inhibition triggers the secretion and clearing of pTDP-43 from C9ORF72 ALS/FTD iMNs. To determine if PIKFYVE inhibition induces secretion of C9ORF72 DPRs from iMNs, we examined poly(GR) levels in iMN-derived exosomes using a previously-validated ELISA assay (Choi et al., 2019; Krishnan et al., 2022). Apilimod treatment significantly increased poly(GR) levels in exosomes harvested from C9ORF72 ALS/FTD, but not control iMNs (Figure 3.3O). Comparing poly(GR) levels in exosomal and cell pellet fractions showed that apilimod treatment resulted in an exosomal:pellet ratio of ~2:1 of poly(GR) and reduced poly(GR) in C9ORF72 ALS/FTD iMNs to background levels (Figure 3.3O-Q). Thus, proteins that drive ALS pathology are substrates for PIKFYVE inhibition-induced exocytosis in iMNs. PIKFYVE inhibition clears pTDP-43 through amphisome and multivesicular body exocytosis To determine if PIKFYVE inhibition extends ALS iMN survival by stimulating exocytosis, we tested whether apilimod could improve C9ORF72 ALS/FTD iMN survival if exocytosis was blocked with GW4869. Indeed, while GW4869 had no effect on control or C9ORF72 ALS/FTD iMN survival on its own, it blocked the ability of apilimod to increase C9ORF72 ALS/FTD iMN survival (Figure 3.4A; Figure S3.4A). To further investigate the mechanism of secretion activated by PIKFYVE inhibition, we used ASOs to selectively suppress different secretory and protein clearance pathways in C9ORF72 ALS/FTD iMNs and determined which were required for the therapeutic efficacy of apilimod. We suppressed the genes encoding VAMP7 and RAB27A, which are critical for plasma membrane fusion of multivesicular bodies and amphisomes (products of autophagosome and multivesicular body fusion), ATG7 and RAB8A, which promote amphisome exocytosis by stimulating autophagosome biogenesis and fusion with the plasma membrane, respectively, HSPA8, which is required for microautophagy and chaperone-mediated autophagy, 52 MCOLN1, a protein critical for lysosomal exocytosis, GORASP1, an important component of the secretory autophagy pathway, and NSMAF, which is required for LC3-dependent extracellular vesicle loading and secretion (Figure S3.4B) (Chen et al., 2017; Chiang et al., 1989; Gee et al., 2011; Kim et al., 2019; Komatsu et al., 2005; Leidal et al., 2020; Ostrowski et al., 2010; Padmanabhan and Manjithaya, 2020; Sahu et al., 2011; Son et al., 2016; Wojnacki et al., 2021). While none of the ASOs affected the survival of control iMNs, suppression of VAMP7, RAB27A, ATG7, and RAB8A significantly reduce the ability of apilimod to increase C9ORF72 ALS/FTD iMN survival (Figure 3.4B; Figure S3.4C). In contrast, suppression of HSPA8, MCOLN1, GORASP1, and NSMAF did not significantly reduce the efficacy of apilimod on C9ORF72 ALS/FTD iMNs (Figure 3.4B). Since we had observed that exosomal release was critical for the therapeutic efficacy of PIKFYVE inhibition, we also examined the ability of ASOs targeting each gene to suppress the release of TSG101+ exosomes upon PIKFYVE inhibition. To control for any differences in cell death we normalized to total protein levels in the cell pellet fraction, which were similar between samples. Mirroring our iMN survival results, ASOs targeting VAMP7, RAB27A, ATG7, and RAB8A significantly reduced the ability of apilimod to induce the secretion of TSG101+ exosomes from C9ORF72 ALS/FTD iMNs, whereas suppressing most of the other genes did not (Figure 3.4C, D; Figure S3.4D). Of the genes whose suppression did not affect the efficacy of apilimod in the iMN survival assay, only MCOLN1 downregulation significantly lowered TSG101+ levels in the exosome fraction upon PIKFYVE inhibition (Figure 3.4C, D). This suggested that while PIKFYVE inhibition may induce lysosomal exocytosis, this mechanism does not seem to contribute strongly to its ability to increase C9ORF72 ALS/FTD iMN survival. Finally, we assessed the effects of VAMP7, RAB27A, ATG7, and RAB8A ASO treatment on exosomal secretion of pTDP-43. We found that suppression of all 4 genes reduced the ability of apilimod to induce exosomal secretion of pTDP-43 from C9ORF72 ALS/FTD iMNs (Figure 3.4E, F). These results suggest that amphisome and multivesicular body exocytosis are most critical for the efficacy of PIKFYVE inhibition, while microautophagy, chaperone-mediated autophagy, lysosomal 53 exocytosis, secretory autophagy, and LC3-dependent extracellular vesicle loading and secretion are either less important or do not contribute to the therapeutic effect. It is important to note that RAB8A can also promote secretory autophagy and GORASP1 suppression caused a noticeable though not significant reduction of apilimod’s efficacy in the iMN survival assay (Figure 3.4B (Dupont et al., 2011). These data suggest that secretory autophagy may also contribute to the therapeutic effect of PIKFYVE inhibition, although to a lesser extent than amphisome and multivesicular body exocytosis. In addition, although VAMP7 can also facilitate lysosomal exocytosis, the fact that MCOLN1 suppression did not affect the efficacy of apilimod in the C9ORF72 ALS/FTD iMN survival assay suggests that lysosomal exocytosis does not play a role in the therapeutic mechanism of PIKFYVE inhibition (Figure 3.4B) (Verderio et al., 2012). To directly examine the effects of PIKFYVE inhibition on vesicle trafficking and secretion in iMNs, we performed immunofluorescence and electron microscopy analysis. Similar to our previous results in which apilimod treatment did not increase the total number of LC3-RFP+ vesicles (LC3- RFP+/GFP+ and LC3-RFP+/GFP-), PIKFYVE inhibition did not significantly increase the total number of LC3+ vesicles in C9ORF72 ALS/FTD iMNs (Figure S3.3J and S3.4E). However, this was possibly because PIKFYVE inhibition also increased secretion of LC3+ vesicles since apilimod significantly increased LC3+ vesicles when we simultaneously suppressed RAB27A (Figure S3.4E). To assess the rate at which the LC3+ autophagosomes fused with late endosomes/multivesicular bodies to form amphisomes, we quantified the number of vesicles positive for autophagosome and multivesicular body markers LC3+ and LAMP1+ or CD63+, respectively. Apilimod significantly increased the number of LC3+/LAMP1+ and LC3+/CD63+ vesicles in C9ORF72 ALS/FTD iMNs (Figure S3.4F, G). In addition, RAB27A suppression further increased the number of LC3+/CD63+ vesicles within apilimod-treated C9ORF72 ALS/FTD iMNs, suggesting that PIKFYVE inhibition normally caused secretion of these vesicles (Figure S4G). Notably, this increase in intracellular LC3+/CD63+ vesicles upon RAB27A suppression only occurred in the presence of apilimod, suggesting that the secretion of LC3+/CD63+ amphisomes 54 required PIKFYVE inhibition (Figure S3.4G). One potential mechanism contributing to the elevated amphisome formation is that apilimod increased the number of RAB7+ late endosomes in C9ORF72 ALS/FTD iMNs (Figure S3.4H). Increased amphisome formation did not seem to result from fewer available primary lysosomes since PIKFYVE inhibition in C9ORF72 ALS/FTD iMNs slightly increased the number of small, primary lysosomes and decreased the abundance of large, secondary lysosomes that had likely fused with endosomes (Figure S3.4I). It is important to note that lysosomes also possess high LAMP1 levels and the LC3/LAMP1 co-labeling results could, in principle, reflect increased autophagosome-lysosome fusion. However, our experiments with the dual color LC3 construct showed that apilimod decreased the number of LC3-RFP+/GFP- autolysosomes (Figure S3.3G). Thus, these new results suggest that PIKFYVE inhibition increased the fusion of autophagosomes with late endosomes/multivesicular bodies and promoted their subsequent secretion. We next determined if the increased amphisome formation we observed upon PIKFYVE inhibition mediated pTDP-43 clearance. Indeed, immunofluorescence analysis showed that apilimod treatment significantly increased the number of pTDP-43+/LC3+/CD63+ vesicles in C9ORF72 ALS/FTD iMNs (Figure 3.4G, H). Overlaying immunofluorescence and electron microscopy imaging enabled direct visualization of the presence and secretion of pTDP-43+/CD63+ vesicles from apilimod-treated C9ORF72 ALS/FTD iMNs (Figure 3.4I). In addition, blocking exocytosis with RAB27A suppression significantly increased the number of pTDP-43+/LC3+/CD63+ vesicles within apilimod-treated C9ORF72 ALS/FTD iMNs, providing further evidence that PIKFYVE inhibition induced the secretion of pTDP-43+ exosomes (Figure 3.4G, H). Consistent with PIKFYVE inhibition mitigating pTDP-43 pathology in through through secretion of pTDP-43, apilimod treatment reduced the number and size of cytoplasmic pTDP-43+ punctae and did not affect the localization or intensity of pTDP-43+ punctae in C9ORF72 ALS/FTD iMNs (Figure S3.4J). Together, these results suggest that PIKFYVE inhibition mitigates pTDP-43 pathology in 55 C9ORF72 ALS/FTD iMNs by increasing the formation and secretion of pTDP-43+ amphisomes and multvesicular bodies. PIKFYVE inhibition improves iMN proteostasis and survival for diverse forms of ALS Although PIKFYVE inhibition triggered exocytosis of C9ORF72 DPRs and pTDP-43 from C9ORF72 ALS/FTD iMNs and extended their survival, it remained unclear if this approach could ameliorate the pathology and degeneration of ALS iMNs with non-C9ORF72 etiologies. Since the accumulation of misfolded proteins is a central disease process common to all ALS patients, we hypothesized that activating exocytosis might increase iMN survival for diverse forms of ALS if it was capable of improving proteostasis. We previously identified sporadic ALS iPSC lines whose iMNs displayed ALS disease phenotypes (Shi et al., 2019). Importantly, next-generation sequencing and repeat-primed PCR showed these lines did not harbor rare variants in known ALS genes (Shi et al., 2019). We performed survival and TDP-43 localization assays on iMNs from these 6 lines plus 2 additional sporadic ALS lines with no known ALS mutations. iMNs from all sporadic ALS lines displayed TDP-43 mislocalization into the cytoplasm and degenerated faster than controls upon neurotrophic factor withdrawal (Figure 3.5A, B, E; Figure S3.5A, B). Apilimod restored the normal nuclear:cytoplasmic TDP-43 ratio in sporadic ALS iMNs and extended iMN survival for all sporadic ALS lines (Figure 3.5C-E; Figure S3.5A, B). Apilimod increased sporadic ALS iMN survival in a dose-dependent manner and ASO-mediated suppression of PIKFYVE also slowed sporadic ALS iMN degeneration, confirming that reducing PIKFYVE activity was responsible for the therapeutic effect (Figure S3.5C-E). In contrast to most ALS cases in which neurons develop cytosolic inclusions of normal TDP-43, rare patients harbor ALS-causing mutations in the glycine-rich region of TDP-43 that alter its biophysical properties. The G298S mutation promotes TDP-43 aggregation and increases the viscosity of neuronal TDP-43 ribonucleoprotein granules while impairing their axonal transport (Gopal et al., 2017; Mann et al., 2019). To determine if activating neuronal exocytosis can 56 ameliorate neurodegeneration driven by the G298S mutation in TDP-43, we tested apilimod on iMNs from an ALS patient harboring a TDP-43 G298S mutation. TDP-43 G298S iMNs degenerated significantly faster than controls and apilimod potently increased their survival (Figure 3.5F; Figure S3.5F). Thus, PIKFYVE inhibition increases TDP-43 G298S iMN survival. A small percentage of ALS patients harbor disease-causing mutations in Fused in sarcoma (FUS) and develop neurotoxic FUS pathology instead of TDP-43 pathology (Lai et al., 2011). FUS and TDP-43 differ in their propensity to translocate to the cytoplasm and aggregate in response to stress and thus proteostasis mechanisms may differ in their ability to clear misfolded FUS and TDP-43 (Tischbein et al., 2019). To determine if PIKFYVE inhibition can extend the survival of FUS ALS iMNs, we generated iMNs from two ALS patients who carried ALS-associated FUS mutations (H517Q, R522R) that cause mislocalization of FUS to the cytoplasm in iMNs (Liu et al., 2016). Indeed, we found that FUS ALS iMNs displayed decreased nuclear:cytoplasmic ratios of FUS compared to control iMNs (Figure S3.5G). iMNs from both FUS-ALS lines degenerated faster than controls, and apilimod increased both the nuclear:cytoplasmic ratio of FUS and the survival of FUS ALS iMNs (Figure 3.5F; Figures S3.5F, S3.5G). Thus, PIKFYVE inhibition extends the survival of iMNs derived from diverse forms of ALS including C9ORF72, TARDBP, FUS, and sporadic ALS without known causal mutations. To determine if PIKFYVE inhibition increased the survival of non-C9ORF72 ALS iMNs by stimulating exocytosis, we examined iMN survival in the presence of apilimod and ASOs that suppressed RAB27A, VAMP7, or ATG7. Suppression of RAB27A, VAMP7, or ATG7 blocked the ability of apilimod to increase the expression of sporadic and FUS ALS iMNs, suggesting that exocytosis and autophagosome formation were required for the therapeutic effect of PIKFYVE inhibition in these lines (Figure 3.5G; Figure S3.5H). In addition, GW4869 greatly reduced the therapeutic efficacy of apilimod on TARDBP, FUS, and sporadic ALS iMNs (Figure S3.5A, S3.5I- M). Although GW4869 did not fully block the efficacy of apilimod on iMNs from some FUS and sporadic ALS lines, this may be due to variability in Neutral Sphingomyelinase 2 or ceramide 57 levels between lines since RAB27A, VAMP7, or ATG7 suppression completely neutralized the efficacy of apilimod on FUS and sporadic ALS iMNs (Figure 3.5G; Figures S3.5A, S3.5I-M). To rule out the ATG7 was required for efficacy due to its ability to promote macroautophagy, we examined FUS ALS iMN survival in the presence of apilimod and bafilomycin. Similar to how bafilomycin failed to prevent apilimod from increasing the survival of iMNs with TDP-43-mediated disease processes, it also had little effect on the ability of PIKFYVE inhibition to increase FUS ALS iMN survival (Figure S3.3I, S3.5N). Thus, autophagic flux was not required for the efficacy of apilimod in iMNs with TDP-43- or FUS-mediated disease processes. Together, these results suggest that PIKFYVE inhibition extends the survival of iMNs derived from diverse forms of ALS including C9ORF72, TARDBP, FUS, and sporadic ALS, and exocytosis is required for its therapeutic effect. PIKFYVE inhibition increases motor function in fly and worm TDP-43 ALS models To begin to assess if blocking PIKFYVE activity preserves motor neuron function in small animal models of TDP-43-driven or sporadic ALS, we first assessed the effect of lowering PIKFYVE activity in Drosophila larvae overexpressing human TDP-43 G298S in motor neurons, which causes neurodegeneration and reduced motor function in a larval turning time assay (Coyne et al., 2014). Chronic apilimod treatment and Pikfyve RNAi both ameliorated motor defects in TDP-43 G298S larvae (Figure 3.5H, I). We next tested apilimod in Caenorhabditis elegans overexpressing TDP-43 A315T . This strain displays neurodegeneration and motility defects leading to progressive paralysis (Figure S3.5O- Q) (Aggad et al., 2014). Apilimod significantly reduced the percentage of TDP-43 A315T worms displaying paralysis and neurodegeneration (Figure S3.5O-Q). Therefore, reducing PIKFYVE activity increases motor neuron function in Drosophila and C. elegans models of TDP-43-driven or sporadic ALS. 58 Pikfyve suppression improves motor function and extends survival of TDP-43 and C9ORF72 mice To determine if activating neuronal exocytosis increases motor function and extends the survival of ALS mice, we utilized the TAR4/4 Thy1::TDP-43 mouse model (Wils et al., 2010). These mice display motor impairment at 14 days, loss of spinal motor neurons by day 21, and paralysis by day 23 (Becker et al., 2017; Wils et al., 2010). Importantly, other groups have used these mice to test novel ASO-based treatments for ALS, including Atxn2 suppression (Becker et al., 2017). Intrathecal injection of apilimod in adult mice increased cerebrospinal fluid (CSF) levels of markers enriched in PIKFYVE inhibition-induced exosomes such as OPTN, LC3, and WIPI2 (autophagosome markers), Huntingtin and SOD1 (aggregation-prone proteins), and ANXA2 (leaderless protein) within 4 hours, suggesting that PIKFYVE inhibition can induce exocytosis in vivo and excretion of these proteins into the CSF (Figure S3.6A). To chronically and stably activate exocytosis for efficacy studies, we genetically suppressed Pikfyve by two approaches - an ASO targeting Pikfyve or genetic deletion of exon 6 in one copy of Pikfyve using a CMV-Cre driver line crossed with a published Pikfyve-flox line (Figure S3.6B-G) (Ikonomov et al., 2011; Schwenk et al., 1995). Deletion of exon 6 in one copy of Pikfyve or intracerebroventricular Pikfyve ASO administration reduced PIKFYVE mRNA and protein levels by about 50% (Figure S3.6B-G). Immunoblotting showed that Pikfyve ASO treatment increased levels of the autophagy receptor OPTN in the cerebrospinal fluid (CSF), suggesting that suppressing Pikfyve by ~50% activated exocytosis in vivo (Figure 3.6A). Genetic deletion of one copy of Pikfyve or Pikfyve ASO treatment ameliorated motor impairment in TDP-43 mice and extended their survival by 25-30% (Figure 3.6B-E; Figure S3.6H-K). The life span extension induced by Pikfyve ASO treatment was similar to the largest effects previously observed with a pharmacological intervention in this aggressive TDP-43 model, including Atxn2 ASO treatment (Becker et al., 2017). Importantly, exocytosis was integral to the efficacy of the Pikfyve ASO treatment as co-administration with GW4869 blocked stimulation of exocytosis by the ASO and impaired the ASO’s ability to increase motor function in TDP-43 mice (Figure 3.6F-G; Figure S3.6L-M). To estimate the therapeutic index of Pikfyve ASO 59 treatment, we tested the efficacy of a 5-fold lower dose (5 ug instead of 25 ug ASO). The 5 ug ASO dose significantly improved motor function in TDP-43 mice, suggesting this approach has a reasonable therapeutic index of at least 5x (Figure S3.5N-P). Thus, Pikfyve suppression improves motor function and extends the survival of TDP-43 mice in an exocytosis-dependent manner. To assess the efficacy of Pikfyve suppression in a slightly less aggressive animal model and one that mimics C9ORF72 disease processes, we tested Pikfyve ASO treatment in the AAV-GR100- GFP mouse model developed by the Petrucelli group (Zhang et al., 2018). These mice are transduced intracerebroventricularly with an AAV encoding 100 repeats of poly(GR) at P1 and develop poly(GR)+ aggregates, neurodegeneration, and impaired motor function (Zhang et al., 2018). Given that poly(GR) is one of the most neurotoxic C9ORF72 DPRs, this model mimics a critical aspect of the C9ORF72 ALS/FTD pathophysiology (Wen et al., 2014). Similar to published work, we found that mice transduced with AAV-GR100-GFP displayed GFP+/poly(GR)+ cells in the cortex, whereas AAV-GFP-infected mice only possessed GFP+/poly(GR)- cells (Figure S3.6Q) (Zhang et al., 2018). AAV-GR100-GFP mice developed a hindlimb clasping phenotype at about 3 weeks of age and about 50% of the AAV-GR100-GFP mice treated with a negative control ASO at P4 and P30 died by day 35 (Figure 3.6H, I). In contrast, Pikfyve ASO treatment at P4 and P30 significantly reduced hindlimb clasping and extended median survival to beyond day 57, which was the last time point tested (Figure 3.6H, I). This represents at least a 60% increase in median survival. These results suggest that Pikfyve suppression can mitigate C9ORF72 poly(GR) disease processes in vivo. Pikfyve suppression reduces TDP-43 and C9ORF72 pathology and neurodegeneration in vivo We next examined the ability of Pikfyve suppression to ameliorate neurodegeneration and TDP- 43 pathology. The most pronounced neurodegeneration in TAR4/4 TDP-43 mice occurs in the spinal cord (Becker et al., 2017; Wils et al., 2010). Indeed, TDP-43 mice displayed a marked loss 60 of motor neurons in the ventral horn (Figure 3.7A, B). Pikfyve ASO treatment fully preserved the number of motor neurons in the lumbar spinal cords of TDP-43 mice at day 21 (Figure 3.7A, B). In contrast, Pikfyve ASO treatment did not decrease neurodegeneration when co-administered with the exocytosis inhibitor GW4869 (Figure 3.7A, B). Thus, Pikfyve suppression reduces neurodegeneration in TDP-43 mice and this effect requires exocytosis. In ALS patient neurons with TDP-43 pathology, cytosolic inclusions are predominantly composed of pTDP-43, making pTDP-43 a marker of aggregated TDP-43 (Neumann et al., 2009). To determine whether Pikfyve suppression ameliorated TDP-43 pathology in vivo, we quantified pTDP-43+ (Ser403/404) punctae in the cytoplasm of spinal motor neurons. TDP-43 mice displayed more cytosolic pTDP-43+ punctae in spinal motor neurons than controls, and Pikfyve ASO treatment ameliorated this pathology and lowered cytoplasmic pTDP-43 levels in brain lysates without affecting total TDP-43 levels in the central nervous system (CNS) (Figure 3.7C, D; Figure S3.7A, B). Importantly, GW4869 blocked the ability of Pikfyve ASO treatment to reduce pTDP-43+ punctae in spinal motor neurons, indicating that exocytosis was required to clear this pathology (Figure 3.7E, F). In addition, Pikfyve ASO treatment reduced TDP-43 mislocalization and increased the TDP-43 nuclear:cytoplasmic ratio (Figure 3.7G, H). Thus, Pikfyve suppression ameliorates TDP-43 pathology in vivo and these effects require exocytosis. To determine if activating exocytosis spreads TDP-43 pathology from the ventral spinal cord to other regions of the CNS, we examined other neuronal populations. Similar to its effect in spinal motor neurons, Pikfyve ASO administration reduced neuronal pTDP-43+ punctae in dorsal spinal cord neurons and CTIP2+ layer V neurons in the motor cortex in TDP-43 mice (Figure S3.7C, D). In the caudal putamen, in which neurons did not display significantly elevated pTDP-43+ punctae in TDP-43 mice, Pikfyve ASO treatment did not alter neuronal pTDP-43+ punctae levels (Figure S3.7E). Therefore, Pikfyve suppression ameliorates TDP-43 pathology in multiple regions of the CNS without spreading pathology to neurons in surrounding areas. 61 In contrast to neurons, microglia displayed more pTDP-43+ punctae after Pikfyve ASO treatment (Figure 3.7I, J). Since microglia in TDP-43 mice did not display pTDP-43 pathology without Pikfyve ASO treatment, this suggests microglia received the pTDP-43 exocytosed from neighboring neurons (Figure 3.7I, J). To investigate this possibility, we generated iPSC-derived microglia according to a protocol recently published by Cowley and colleagues and administered exosomes harvested from C9ORF72 ALS/FTD iMNs treated with apilimod (Haenseler et al., 2017). Within 24 hours of exposure to the exosomal material, iPSC-microglia treated with C9ORF72 ALS/FTD iMN exosomes displayed a large increase in pTDP-43+ (Ser409/410) punctae and developed a more amoeboid morphology (Figure S3.7F-H). After 96 hours, the number of pTDP-43+ punctae and the fraction of microglia with amoeboid morphology had decreased significantly (Figure S3.7F-H). These results suggested that microglia can take up the pTDP-43 secreted from ALS motor neurons upon PIKFYVE inhibition, that they can clear pTDP- 43 over time, and that any changes in microglial morphology are transient and reversible. Pikfyve ASO treatment did not increase the number of microglia or astrocytes in control or TDP-43 mice, indicating that activation of exocytosis did not stimulate gliosis (Figure S3.7I, J). Instead, Pikfyve ASO treatment partially reduced the microgliosis in TDP-43 mice (Figure S3.7I). Together, these data suggest that microglia help to clear the pathological proteins released from neurons by exocytosis, leading to an overall reduction in TDP-43 pathology in neurons. To examine the efficacy of reducing Pikfyve expression in a less aggressive TDP-43 model with a longer disease course, we employed TAR4 mice, which contain half the expression of human TDP-43 (0.6x of endogenous TDP-43) compared to TAR4/4 mice and develop neurodegeneration at 10 months of age instead of 1 month (Wils et al., 2010). To reduce Pikfyve expression by 50%, we crossed TAR4 mice with Pikfyve-flox mice harboring a ubiquitous Cre driver (Ikonomov et al., 2011). At 10 months, we examined the number of spinal motor neurons and pTDP-43 pathology and found that Pikfyve+/-;TAR4 mice possessed significantly more spinal motor neurons and fewer pTDP-43+ (Ser409/410) punctae in their motor neurons than Pikfyve+/+;TAR4 mice (Figure 62 3.7K-N). Thus, reducing PIKFYVE levels mitigates neurodegeneration and TDP-43 proteinopathy over at least a 10-month period in mice displaying a longer disease course. To determine if Pikfyve suppression also ameliorated neurodegeneration caused by C9ORF72 DPRs, we examined day 26 AAV-GR100-GFP mice that had been treated with a negative control or Pikfyve ASO at P4. AAV-GR100-GFP displayed significantly fewer spinal motor neurons than AAV-GFP mice, and Pikfyve ASO treatment greatly reduced this deficit in motor neurons in AAV- GR100-GFP mice (Figure 3.7O, P). Therefore, Pikfyve suppression can mitigate neurodegeneration caused by C9ORF72 disease processes in vivo. Although at least a 50% loss of PIKFYVE activity is well-tolerated in mice and humans, more complete loss of PIKFYVE can cause cellular toxicity in certain conditions (Gee et al., 2015; Lakkaraju et al., 2021; Zolov et al., 2012). Severe reduction in PIKFYVE activity can result in the formation of enlarged LAMP1+ vacuoles that may drive cytotoxicity (Lakkaraju et al., 2021; Zolov et al., 2012). However, it was unclear if reducing PIKFYVE levels by ~50%, which was sufficient for efficacy in vivo, was severe enough to perturb intracellular vesicle populations. In cortical neurons at day 20, when Pikfyve ASO-treated mice still had about 50% lower PIKFYVE levels than negative control ASO-treated mice, we did not observe any differences in the number of RAB7+ late endosomes, percentage of large LAMP1+ lysosomes, or number of LC3+/LAMP1+ autolysosomes between negative control or Pikfyve ASO-treated mice in either the control or TDP-43 (TAR4/4) backgrounds (Figure S3.6G; S3.7K-M). In addition, immunoblotting on fractionated brain tissue did not reveal any differences in the nuclear:cytoplasmic ratio of TFEB between negative control and Pikfyve ASO-treated mice (Figure S3.7N). To extend this analysis to mice with a 50% reduction in PIKFYVE levels over a longer period of time, we also examined neurons in Pikfyve+/- (Cre-flox); TAR4 mice at 10 months of age and did not observe any differences in the percentage of large LAMP1+ vesicles between Pikfyve+/+ and Pikfyve+/- mice in either the control or TAR4 backgrounds (Figure S3.7O). In addition, examination of control and C9ORF72 ALS/FTD iMNs in the middle of the in vitro survival assay (day 8) and several days 63 after the last bolus addition of apilimod showed that drug-treated iMNs did not display differences in the number of LAMP1+ punctae (Figure S3.7P). Therefore, moderate reductions in PIKFYVE levels that are sufficient for therapeutic efficacy in ALS models do not significantly alter the late endosome, lysosome, and autolysosome vesicle populations in neurons. 3.4 Discussion Here, we show that PIKFYVE inhibition is broadly efficacious in in vitro and in vivo models of diverse forms of ALS. PIKFYVE inhibition stimulates neuronal exocytosis, an unconventional protein clearance pathway, and this potently reduces pathology, neurodegeneration, and motor dysfunction in models of diverse forms of ALS. Given that ALS is caused by many different rare genetic etiologies, broadly effective therapeutic strategies are desperately needed. Surprisingly, one of the most potent and broadly effective neuroprotective mechanisms we identified by phenotypic screening was to activate an unconventional form of secretion that has not been extensively explored for treating neurodegenerative diseases. Exocytosis improved proteostasis and survival of iMNs from an extremely diverse set of ALS patients including C9ORF72, TARDBP, FUS, and sporadic ALS. We also showed that PIKFYVE inhibition increased motor neuron function in vivo, as evidenced by improved voluntary movement in fly, worm, and mouse models of ALS. Although NMJ phenotypes in the Thy1::TDP-43 mouse model were too variable to assess, we were able to measure NMJ function in GR.100 Drosophila larvae, which displayed a severe NMJ phenotype that apilimod could ameliorate. A potential advantage of the unconventional secretory proteostasis pathway is that it reduces the need to rescue lysosomal dysfunction in neurons, which occurs in C9ORF72 and other forms of ALS and during aging (Root et al., 2021; Shi et al., 2018). Although prior studies suggested that neurons engage exocytosis at a low level when proteolytically-stressed, we show for the first time that pharmacologically activating 64 exocytosis to superphysiological levels can potently improve proteostasis and neuronal survival in vitro and in vivo (Iguchi et al., 2016). Although exocytosis can spread pathology in some contexts (Maphis et al., 2015), we did not observe spreading and instead found reduced neuronal pTDP-43+ punctae in several regions of the CNS. One mechanism for mitigating spread appeared to be the ability of microglia to take up secreted pTDP-43. Interestingly, the increased secretion of exosomes did not cause gliosis, but instead reduced the number of microglia in TDP-43 mice, potentially by improving neuronal viability. In addition, PIKFYVE inhibition can block the seeding activity of aggregation-prone proteins taken up by neurons, and this may help stem the spread of disease pathology in vivo (See et al., 2021; Soares et al., 2021). Since GW4869 and RAB27A and VAMP7 suppression blocked the efficacy of PIKFYVE inhibition, we conclude that exocytosis is required for most of the therapeutic benefit. In addition, the fact that ATG7 and RAB8A suppression greatly reduced the efficacy of apilimod suggests that amphisome secretion in particular plays a large role in the therapeutic effect of PIKFYVE inhibition. However, we cannot rule out the possibility that other unexpected effects of PIKFYVE inhibition that RAB27A, VAMP7, ATG7, and RAB8A also modulate may help mitigate neurodegeneration. Further investigation into the mechanisms underlying PIKFYVE inhibition-induced exocytosis and cargo selection will increase the therapeutic potential of this pathway. PIKFYVE inhibition provides one means to stimulate exocytosis of pathological proteins, but other means of stimulating exosome formation and release may provide unforeseen advantages. Our data show that PIKFYVE inhibition can clear C9ORF72 DPRs and pTDP-43 from neurons, and others have observed the removal of ⍺-synuclein (Ejlerskov et al., 2013). It will be important to determine if PIKFYVE inhibition-induced exocytosis can eliminate other disease-associated proteins and if perturbing cargo selection mechanisms increases efficacy. Lastly, one must assess the long-term effects of chronic exocytosis stimulation in humans and the optimal frequency of activation in neurodegenerative indications. 65 With regard to safety, mice lacking one copy of Pikfyve are healthy and reduction of PIKFYVE activity does not become cytotoxic until it is more severely reduced (Lakkaraju et al., 2021; Zolov et al., 2012). In vitro studies using cells derived from mice with different combinations of loss-of- function and hypomorphic alleles of Pikfyve observed cytotoxicity at 80-90% reduction of PIKFYVE activity (Zolov et al., 2012). In addition, a recent study suggests that prions induce vacuolization in neurons by reducing PIKFYVE activity (Lakkaraju et al., 2021). Although vacuolization emerged with greater than 50% reduction of PIKFYVE and the contribution of vacuolization to neuron viability was not explicitly determined, it will be important to incorporate this information and carefully monitor levels of PIKFYVE inhibition during therapeutic development (Lakkaraju et al., 2021). Since a 50% reduction in PIKFYVE levels was sufficient to stimulate exocytosis in our experiments, it suggests that long-term activation of exocytosis is tolerated, although one must verify that genetic deletion of Pikfyve continues to simulate neuronal exocytosis at older ages. The fact that we observed efficacy in vivo with a dose of Pikfyve ASO 5-fold lower than the dose that achieved 50% Pikfyve suppression suggests that lowering Pikfyve expression by less than 50% is able to activate exocytosis. The unusually broad efficacy we observed across models of different forms of ALS and the novelty of this unconventional secretory mechanism for rescuing proteostasis merit further investigation into this therapeutic approach. 66 3.5 Figures Figure 3.1 PIKFYVE inhibition ameliorates C9ORF72 ALS/FTD neurodegeneration. (A) Production of Hb9::RFP+ iMNs and survival tracking by longitudinal microscopy. 67 (B) Kaplan–Meier survival curves of control (CTRL) and C9ORF72 ALS/FTD patient (C9- ALS/FTD) iMNs after withdrawal of neurotrophic factor supplementation (shown for each line separately). n=90 iMNs/line for three control and three C9-ALS/FTD lines. iMNs quantified from three biologically independent iMN conversions per line. Two-sided log-rank test. Statistical significance calculated using the entire survival course (days 1-20). (C) The hazard ratio (Mantel–Haenszel method) for (B). The hazard ratio = the hazard rate of each group relative to the hazard rate of the three CTRL lines in aggregate, which was set as 1 (red dotted line). The hazard ratio was derived from days 1-20 of iMN survival. Mean of biological replicates (independent conversions) ± s.e.m. n=3 independent conversions/line/condition. Statistical significance (one-way ANOVA) was calculated comparing three CTRL lines in aggregate to each C9-ALS/FTD line individually. P-values were corrected for multiple comparisons using statistical hypothesis testing (Šídák method). (D) Kaplan–Meier survival curves of C9-ALS/FTD iMNs in DMSO or 3 μM apilimod (AP) after withdrawal of neurotrophic factor supplementation (shown each line in separate graph). n = 90 iMNs per line per condition for three C9-ALS/FTD lines (C9-ALS/FTD1, C9-ALS/FTD2 and C9- ALS/FTD3). iMNs quantified from three independent iMN conversions per line per condition. Two- sided log-rank test. Statistical significance calculated using the entire survival course (days 1-20). (E) The hazard ratio (Mantel–Haenszel method) of iMNs from CTRL (three lines in aggregate) and three C9-ALS/FTD lines (n=90 iMNs per line per condition) treated with DMSO and 3 μM apilimod (AP) after withdrawal of neurotrophic factor supplementation. The hazard ratio = the hazard rate of each group relative to the hazard rate of the three CTRL lines in aggregate treated with DMSO, which was set as 1 (red dotted line). The hazard ratio was derived from days 1-20 of iMN survival. Mean of biological replicates (independent conversions) ± s.e.m. n=3 independent conversions/line/condition. Statistical significance (one-way ANOVA) was calculated comparing the DMSO conditions to the AP condition for CTRL iMNs and comparing the DMSO condition to the AP condition and to the CTRL + DMSO group for each C9-ALS/FTD line. P-values were corrected for multiple comparisons using statistical hypothesis testing (Šídák method). (F) The hazard ratio (Mantel–Haenszel method) of iMNs from CTRL (two lines in aggregate) and three C9-ALS/FTD lines (n=120 iMNs per line per condition, n=70 iMNs for C9-ALS3) treated with 10 μM negative control ASO (NC ASO), PIKFYVE ASO1, or PIKFYVE ASO2 after withdrawal of neurotrophic factor supplementation. The hazard ratio = the hazard rate of each group divided by the hazard rate of CTRL iMNs treated with the negative control (NC) ASO, which was set as 1 (red dotted line). The hazard ratio was derived from days 1-14 of iMN survival. Mean of biological replicates (independent conversions) ± s.e.m. n=3 independent conversions/line/condition. Statistical significance (one-way ANOVA) was calculated comparing the PIKFYVE ASO1 or PIKFYVE ASO2 groups to the NC ASO group for each line and comparing the NC ASO group to CTRL NC ASO group for each C9-ALS/FTD line. P-values were corrected for multiple comparisons using statistical hypothesis testing (Šídák test). (G) A chronic 5-day apilimod treatment (10 µM) throughout larval development ameliorated locomotor deficits in Drosophila larvae overexpressing C9ORF72 GR.100 (100 repeat-Glycine- Arginine) DPRs in motor neurons. Genotypes and treatments as indicated. N=30 larvae per group. Kruskal-Wallis test. Mean ± s.e.m. P-values were corrected for multiple comparisons using statistical hypothesis testing (Dunn’s test). 68 Figure 3.2 PIKFYVE inhibition ameliorates disease pathology in C9ORF72 ALS/FTD iMNs. (A) RNA levels (relative to MAP2) of truncated Stathmin2 relative to full-length Stathmin2 in CTRL and C9-ALS/FTD iPSC-derived motor neurons treated with DMSO or 3 μM apilimod (AP) for 24 hours. n=5 biological replicates (independent conversions) per condition for CTRL and n=4 biological replicates (independent conversions) for C9-ALS/FTD patients from two CTRL and two C9-ALS/FTD donors. Ordinary one-way ANOVA. Mean ± s.e.m. P-values were corrected for multiple comparisons using statistical hypothesis testing (Šídák method). (B) Immunostaining of total TDP-43 in CTRL and C9-ALS/FTD iMNs with DMSO or 3 μM apilimod (AP) treatment for 24 hours. Solid and dotted lines outline the cell body and nucleus, respectively. Scale bar = 2 μm. (C) Quantification of (B). Quantified values represent the mean intensity of cytoplasmic TDP-43 in n=30 CTRL and n=30 C9-ALS/FTD iMNs from two CTRL and two C9-ALS/FTD patient lines. iMNs were quantified from two independent conversions/group. Each gray circle represents the mean intensity from a single iMN. Kruskal-Wallis test. Median ± interquartile range. P-values were corrected for multiple comparisons using statistical hypothesis testing (Dunn’s test). (D) Quantification of (B). Quantified values represent the average ratio of nuclear to cytoplasmic TDP-43 in n=30 CTRL and n=30 C9-ALS/FTD iMNs from two CTRL and two C9-ALS/FTD patient lines. iMNs were quantified from two independent conversion/group. Each gray circle represents the ratio in a single iMN. Kruskal-Wallis test. Median ± interquartile range. P-values were corrected for multiple comparisons using statistical hypothesis testing (Dunn’s test). 69 (E) Immunostaining of endogenous poly(GR)+ punctae in CTRL or C9-ALS/FTD iMNs treated with DMSO or 3 μM apilimod (AP) for 24 hours. Solid and dotted lines outline the cell body and nucleus, respectively. Scale bar = 2 μm. (F) Quantification of (E) to determine endogenous poly(GR)+ punctae in CTRL or C9-ALS/FTD iMNs with DMSO or 3 μM apilimod (AP) treatment for 24 hours. Quantified values represent the average number of nuclear poly(GR)+ punctae/μm 2 from two CTRL and two C9-ALS/FTD patient lines (n = 40 iMNs for CTRL+DMSO, n = 38 iMNs for CTRL+AP, and n=40 iMNs for C9-ALS/FTD iMNs for both conditions). iMNs were quantified from two independent conversions per group. Each gray circle represents the average number of punctae in a single iMN. Kruskal-Wallis test. Median ± interquartile range. P-values were corrected for multiple comparisons using statistical hypothesis testing (Dunn’s test). 70 Figure 3.3 PIKFYVE inhibition increases exocytosis of aggregation-prone proteins from iMNs. (A) Schematic depicting PIKFYVE’s putative role in the exocytosis of aggregation-prone proteins. DPRs = C9ORF72 dipeptide repeat proteins. pTDP-43 = phosphorylated TDP-43. 71 (B) Immunoblots of exosomal or pellet fractions from CTRL and C9-ALS/FTD iPSC-derived motor neurons against TSG101 (46kDa). Cells were treated with DMSO, 3 μM apilimod (AP), 3 μM apilimod with 10 μM GW4869 (AP+GW) or 10 μM GW4869 (GW) for 24 hours. Samples were normalized by total protein content. Pellet TUJ1 (55kDa) served as a control for cell death. (C) Quantification of exosomal TSG101 for (B). n=9 independent conversions/condition (from three CTRL and three C9-ALS/FTD patients, n=3 independent conversions/line). Samples were normalized by total protein content. Pellet TUJ1 (55kDa) served as a control for cell death. The values were calculated as the relative intensity of exosomal TSG101 normalized to total protein. One-way ANOVA. Mean ± s.e.m. P-values were corrected for multiple comparisons using statistical hypothesis testing (Šídák method). (D) Electron microscopic images of secreted vesicles in the exosomal fraction harvested from C9- ALS/FTD iPSC-derived motor neurons treated with 3 μM apilimod (AP). Scale bar = 100 nm (Left) and 200 nm (Right). (E) Fold-change of selected protein levels in exosomal fractions after 3 μM apilimod (AP) treatment versus DMSO from CTRL (blue bars) and C9-ALS/FTD (red bars) iPSC-derived motor neurons. Each value is the average of three independent differentiations/condition from one CTRL and one C9-ALS/FTD line. Values >1 reflect greater abundance in apilimod-treated cultures compared to DMSO treatment. Samples were normalized by total protein content. Pellet TUJ1 (55kDa) served as a control for cell death. (F) Immunoblots of the exosomal fraction from CTRL and C9-ALS/FTD iPSC-derived motor neurons for LC3 (LC3-I above 15kDa and LC3-II below 15kDa) and p62 (62 kDa). Cells were treated with DMSO and 3 μM apilimod (AP) for 24 hours. Samples were normalized by total protein content. Pellet TUJ1 (55kDa) served as a control for cell death. (G) Quantification of exosomal p62 and LC3-II for (F). n=9 independent conversions/condition (from three CTRL and three C9-ALS/FTD patients, n=3 independent conversions/line). Samples were normalized by total protein content. Pellet TUJ1 (55kDa) served as a control for cell death. The values were calculated as the relative intensity of exosomal p62 or LC3-II normalized to total protein. One-way ANOVA. Mean ± s.e.m. P-values were corrected for multiple comparisons using statistical hypothesis testing (Šídák method). (H) Immunoblots of exosomal and pellet fractions from C9-ALS/FTD iPSC-derived motor neurons for Optineurin (70 kDa). Cells were treated with DMSO or 3 μM apilimod (AP) for 24 hours. (I) Quantification of exosomal Optineurin (OPTN)(70 kDa) for (H). n=9 independent conversions/condition (from three CTRL and three C9-ALS/FTD patients, n=3 independent conversions per line). Samples were normalized by total protein content. Cell pellet TUJ1 was used to control for any differences in cell death between samples (not shown). The values were calculated as the relative intensity of exosomal OPTN normalized to total protein. Unpaired t-test. Mean ± s.e.m. Statistical significance was calculated comparing DMSO to AP for CTRL and C9- ALS/FTD lines individually. (J) Immunoblots of exosomal and pellet fractions from CTRL and C9-ALS/FTD iPSC-derived motor neurons for phosphorylated (Ser409/410) TDP-43 (50 kDa) (p-TDP43). Cells were treated with DMSO, 3 μM apilimod (AP), 3 μM apilimod with 10 μM GW4869 (AP+GW), or 10 μM GW4869 (GW) for 24 hours. Samples were normalized by total protein content. Cell pellet TUJ1 (55kDa) served as a control for cell death. The same blot from Figure 3B was used in order to facilitate comparisons between different markers. Exosome TSG101 is the same as in Figure 3B. (K) Quantification of exosomal phosphorylated (Ser409/410) TDP-43 (p-TDP-43) from (J). n=9 independent conversions/condition (from three CTRL and three C9-ALS/FTD patients, n=3 independent conversions/line). Samples were normalized by total protein content. Pellet TUJ1 (55kDa) served as a control for cell death. Values were calculated as the relative intensity of exosomal p-TDP43 normalized to total protein. One-way ANOVA. Mean ± s.e.m. P-values were corrected for multiple comparisons using statistical hypothesis testing (Šídák method). 72 (L) Ratio of phosphorylated (Ser409/410) TDP-43 (p-TDP-43) in the exosomal vs. cell pellet fractions of CTRL and C9-ALS/FTD iPSC-derived motor neurons. n=6 independent conversions/condition from two CTRL and two C9-ALS/FTD patients, n=3 independent conversions/line). Samples were normalized by total protein content. Pellet TUJ1 (55kDa) served as a control for cell death. One-way ANOVA. Mean ± s.e.m. P-values were corrected for multiple comparisons using statistical hypothesis testing (Šídák method). (M) Immunoblots of cytoplasmic fractions from CTRL and C9-ALS/FTD iPSC-derived motor neurons against phosphorylated (Ser409/410) TDP-43 (50 kDa) (p-TDP43) or total TDP-43 (43 kDa). Cells were treated with DMSO or 3 μM apilimod (AP) for 24 hours. Samples were normalized by total protein content. Cell pellet TUJ1 (55kDa) served as a control for cell death (not shown). (N) Quantification of cytoplasmic phosphorylated (Ser409/410) TDP-43 (p-TDP-43) from (M). n=6 independent conversions/condition from two CTRL (n=3 independent conversions/line) and n=8 from independent conversions/condition from three C9-ALS/FTD patients (n=2 independent conversions for C9-ALS/FTD1 and n=3 independent conversions for C9-ALS/FTD2 and C9- ALS/FTD3). Samples were normalized by total protein content. Values were calculated as the relative intensity of cytoplasmic phosphorylated (Ser409/410) p-TDP43 normalized to total protein. One-way ANOVA. Mean ± s.e.m. P-values were corrected for multiple comparisons using statistical hypothesis testing (Šídák method). (O) poly(GR) levels in exosomal fractions of CTRL and C9-ALS/FTD iPSC-derived motor neurons as measured by immunoassay. n=11 biological replicates (independent conversions)/condition from three CTRL lines (n=3 independent conversions for CTRL2 and n=4 independent conversions for CTRL1 and CTRL3) and n=7 independent conversions/condition from two C9- ALS/FTD lines (n=3 independent conversions for C9-ALS/FTD1 and n=4 independent conversions for C9-ALS/FTD2). Each gray circle represents one biological replicate. Samples were normalized by total protein content. One-way ANOVA. Mean ± s.e.m. P-values were corrected for multiple comparisons using statistical hypothesis testing (Šídák method). (P) poly(GR) levels in cell pellet fractions of CTRL and C9-ALS/FTD iPSC-derived motor neurons as measured by immunoassay. n=6 biological replicates (independent conversions)/condition from two CTRL lines (n=3 independent conversions for CTRL2 and CTRL3) and n=9 independent conversions/condition from three C9-ALS/FTD lines (n=3 independent conversions for C9- ALS/FTD1, C9-ALS/FTD2 and C9-ALS/FTD3). Samples were normalized by total protein content. Each gray circle represents one biological replicate. One-way ANOVA. Mean ± s.e.m. P-values were corrected for multiple comparisons using statistical hypothesis testing (Šídák method). (Q) Ratio of poly(GR) levels in the exosomal vs. cell pellet fractions of CTRL and C9-ALS/FTD iPSC-derived motor neurons. n=7 independent conversions/condition from two C9-ALS/FTD patients, n=3 independent conversions for C9-ALS/FTD1 and n=4 independent conversions for C9-ALS/FTD2). Mann-Whitney test. Mean ± s.e.m. 73 Figure 3.4 PIKFYVE inhibition clears pTDP-43 through amphisome and multivesicular body exocytosis (A) The hazard ratio (Mantel–Haenszel method) of CTRL and C9-ALS/FTD iMNs (3 lines in aggregate/genotype, n=90 iMNs/line/condition) treated with DMSO, 3 μM apilimod (AP), 3 μM 74 apilimod+100 nM GW4869 (AP+GW), or 100 nM GW4869 (GW) after withdrawal of neurotrophic factor supplementation. The hazard ratio = the hazard rate of each group divided by the hazard rate of the three CTRL lines in aggregate treated with DMSO, which was set as 1 (red dotted line). The hazard ratio was derived from days 1-20 of iMN survival. Mean of biological replicates (independent conversions) ± s.e.m. n=3 independent conversions/line/condition. Statistical significance (one-way ANOVA) was calculated comparing CTRL+DMSO to C9-ALS/FTD+DMSO and comparing the DMSO condition to the AP, AP+GW, and GW conditions for CTRL and C9- ALS/FTD lines. P-values were corrected for multiple comparisons using statistical hypothesis testing (Šídák method). (B) The hazard ratios (Mantel–Haenszel method) of CTRL iMNs (2 lines in aggregate) treated with 9 μM of negative control (NC) ASO plus DMSO or 3 μM apilimod (AP) or C9-ALS/FTD iMNs (2 lines in aggregate) treated with 9 μM negative control (NC) ASO, RAB27A ASO, VAMP7 ASO, ATG7 ASO, RAB8A ASO, HSPA8 ASO, MCOLN1 ASO, GORASP1 ASO, or NSMAF ASO plus DMSO or 3 μM apilimod (AP) after withdrawal of neurotrophic factor supplementation. The hazard ratio = the hazard rate of each group divided by the hazard rate of the three CTRL lines in aggregate treated with NC ASO and DMSO, which was set as 1 (red dotted line). For CTRL iMNs, n=429 iMNs for NC ASO+DMSO and n=326 iMNs for NC ASO+AP. For C9-ALS/FTD iMNs, n=291 iMNs for NC ASO+DMSO, n=174 iMNs for NC ASO+AP, n=211 iMNs for RAB27A ASO+AP, n=207 iMNs for VAMP7 ASO+AP, n=213 iMNs for ATG7 ASO+AP, n=225 iMNs for RAB8A ASO+AP, n=201 iMNs for HSPA8 ASO+AP, n=215 iMNs for MCOLN1 ASO+AP, n=219 iMNs for GORASP1 ASO+AP, and n=229 iMNs for NSMAF ASO+AP. The hazard ratio was derived from days 1-12 of iMN survival. Mean of biological replicates (independent conversions) ± s.e.m. n=3 independent conversions/line/condition. Statistical significance (one-way ANOVA) was calculated comparing each condition to C9-ALS/FTD NC ASO+AP. P-values were corrected for multiple comparisons using statistical hypothesis testing (Šídák test). Each ASO is color-coded and categorized by different forms of secretion: red (RAB27A and VAMP7, amphisome and multivesicular body exocytosis), brown (ATG7 and RAB8A, amphisome formation and exocytosis, respectively), yellow (HSPA8, microautophagy and chaperone-mediated autophagy), green (MCOLN1, lysosomal exocytosis), blue (GORASP1, secretory autophagy), and purple (NSMAF, LC3-dependent extracellular vesicle loading and secretion). (C) Immunoblots of exosomal or pellet fractions from CTRL and C9-ALS/FTD iPSC-derived motor neurons for TSG101 (46kDa). Cells were treated with 9 μM of negative control (NC), RAB27A ASO, VAMP7 ASO, ATG7 ASO, RAB8A ASO, HSPA8 ASO, MCOLN1 ASO, GORASP1 ASO, or NSMAF ASO for 48 hours and then treated with DMSO or 3 μM apilimod (AP) for 24 hours. (D) Quantification of exosomal TSG101 for (C). n=4 independent iMN conversions/condition from two CTRL (n=2 independent conversions/line for CTRL2 and CTRL3) and n=7 independent conversions/condition from three C9-ALS/FTD lines (n=3 independent conversions/line for C9- ALS/FTD1 and C9-ALS/FTD2, and n=1 independent conversion for C9-ALS/FTD3). The values were calculated as the relative intensity of exosomal TSG101 normalized to total protein. One- way ANOVA. Mean ± s.e.m. P-values were corrected for multiple comparisons using statistical hypothesis testing (Šídák method). Each ASO is color-coded and categorized by different forms of secretion: red (RAB27A and VAMP7, amphisome and multivesicular body exocytosis), brown (ATG7 and RAB8A, amphisome production and exocytosis, respectively), yellow (HSPA8, microautophagy and chaperone-mediated autophagy), green (MCOLN1, lysosomal exocytosis), blue (GORASP1, secretory autophagy), and purple (NSMAF, LC3-dependent extracellular vesicle loading and secretion). (E) Immunoblots of exosomal fractions from C9-ALS/FTD iPSC-derived motor neurons against phosphorylated (Ser409/410) TDP-43 (50 kDa) (p-TDP43). Cells were treated with 9 μM of negative control (NC), VAMP7 ASO, ATG7 ASO or RAB8A ASO for 24 hours and then treated with DMSO or 3 μM apilimod (AP) for 24 hours. TSG101 (46 kDa) from the exosomal fraction was used as a control for exosome secretion and collection. 75 (F) Quantification of exosomal phosphorylated (Ser409/410) TDP-43 (p-TDP-43) from (E). n=7 independent conversions/condition from three C9-ALS/FTD patients (n=3 independent conversions/line for C9-ALS/FTD1 and C9-ALS/FTD2, and n=1 independent conversion for C9- ALS/FTD3). Values calculated as the relative intensity of exosomal p-TDP43 normalized to total protein. One-way ANOVA. Mean ± s.e.m. P-values were corrected for multiple comparisons using statistical hypothesis testing (Šídák method). Each ASO is color-coded and categorized by different forms of secretion: red (VAMP7, amphisome and multivesicular body exocytosis) and brown (ATG7 and RAB8A, amphisome exocytosis). (G) Immunostaining of phosphorylated (Ser409/410) TDP-43 (p-TDP-43)+/CD63+/LC3B+ punctae in C9-ALS/FTD iPSC-derived motor neurons. Cells were treated with 9 μM of negative control (NC) or RAB27A ASO for 24 hours and then treated with DMSO or 3 μM apilimod (AP) for 24 hours. Confocal Z-axis scanning was performed to identify colocalization of punctae in MAP2+ neurons (white). The yellow arrows denote p-TDP-43 (green) colocalized with CD63 (red) and LC3B (blue) in three-dimensional space. The white arrows denote other colocalized p-TDP- 43+/CD63+/LC3B+ punctae in the cell. Scale bar = 5 μm. (H) Quantification of (G) to determine the number of colocalized p-TDP-43+/CD63+/LC3B+ punctae in C9-ALS/FTD iPSC-derived MAP2+ motor neurons. Quantified values represent the average number of p-TDP-43+/CD63+/LC3B+ punctae/μm 2 from one C9-ALS/FTD patient line (n =15 iMNs per condition). iMNs were quantified from two independent conversions per group. Each gray circle represents the average number in a single MAP2+ neuron. Ordinary one-way ANOVA. Mean ± s.e.m. p-values were corrected for multiple comparisons using statistical hypothesis testing (Šídák method). Each ASO is color-coded and categorized by different forms of secretion: red (RAB27A, amphisome and multivesicular body exocytosis). (I) Correlative light electron microscopic (CLEM) images of secreted CD63+ vesicles containing TDP-43 on the membrane of a C9-ALS/FTD iMN treated with 3 μM apilimod (AP). Top row: light microscopy images of DAPI (blue, nucleus), HB9::RFP (white, motor neuron reporter), TDP-43 (green) and CD63 (red). The yellow arrow indicates a CD63+ vesicle containing TDP-43 that appears to be in the middle of being secreted. Bottom row: electron microscopy images of the same cell. The arrow shown in the top and bottom rows marks the CD63+/TDP-43+ vesicle in the process of being secreted. Scale bar = 2 μm (top row and bottom row left) and 500 nm (bottom row right). 76 Figure 3.5 PIKFYVE inhibition improves iMN proteostasis and survival for diverse forms of ALS. (A) Immunostaining of total TDP-43 in CTRL and sporadic ALS (sALS) iMNs. Solid and dotted lines outline the cell body and nucleus, respectively. Scale bar = 2 μm. 77 (B) Quantification of (A). Average ratio of nuclear to cytoplasmic TDP-43 in n=15 iMNs/line from two CTRL and six sALS patient lines. iMNs were quantified from two independent conversions per group. Each gray circle represents the ratio from a single iMN. Mann-Whitney test. Median ± interquartile range. (C) Immunostaining of total TDP-43 in sporadic ALS (sALS) iMNs treated with DMSO or 3 μM apilimod (AP) for 24 hours. Solid and dotted lines outline the cell body and nucleus, respectively. Scale bar = 2 μm. (D) Quantification of (C). The average ratio of nuclear to cytoplasmic TDP-43 in n=90 sALS iMNs from six sALS patient lines (n=15/line/condition). iMNs were quantified from two independent conversions per group. Each gray circle represents the ratio in a single iMN. Mann-Whitney test. Median ± interquartile range. (E) The hazard ratios (Mantel–Haenszel method) of iMNs from two control (CTRL) and eight sporadic ALS (sALS) donors in DMSO or 3 μM apilimod (AP) after withdrawal of neurotrophic factor supplementation. n=90 iMNs/line/condition. The hazard ratio = the hazard rate of each group divided by the hazard rate of the two CTRL lines in aggregate treated with DMSO, which was set as 1 (red dotted line). The hazard ratio was derived from days 1-20 of iMN survival. Mean of biological replicates (independent conversions) ± s.e.m. n = 3 independent conversions/line/condition. Statistical significance (one-way ANOVA) was calculated comparing the DMSO condition for each sALS line to the AP condition and the CTRL iMN +DMSO condition. P-values were corrected for multiple comparisons using statistical hypothesis testing (Šídák method). (F) The hazard ratios (Mantel–Haenszel method) of iMNs from CTRL, two FUS ALS (H517Q, R522R mutation), and TARDBP ALS (G298S mutation) donors treated with DMSO or 3 μM apilimod (AP). N=50 iMNs/line/condition. The hazard ratio = the hazard rate of each group divided by the hazard rate of the CTRL iMNs treated with DMSO, which was set as 1 (red dotted line). The hazard ratio was derived from days 1-15 of iMN survival. Mean of biological replicates (independent conversions) ± s.e.m. N=3 independent conversions/line/condition. Statistical significance (one-way ANOVA) was calculated comparing the DMSO condition for each ALS line to the AP condition and the CTRL+DMSO group. P-values were corrected for multiple comparisons using statistical hypothesis testing (Šídák method). (G) The hazard ratios (Mantel–Haenszel method) of iMNs from CTRL (two lines in aggregate), two sporadic ALS (sALS3 and sALS5), and one FUS ALS (R522R mutation) line treated with 9 μM negative control (NC) ASO or RAB27A ASO plus DMSO or 3 μM apilimod (AP) after withdrawal of neurotrophic factor supplementation. The hazard ratio = the hazard rate of each group divided by the hazard rate of the CTRL iMNs treated with NC ASO+DMSO, which was set as 1 (red dotted line). The hazard ratio was derived from days 1-12 of iMN survival. For CTRL, n=439 iMNs for NC ASO+DMSO. For sALS3, n=214 iMNs for NC ASO+DMSO, n=164 iMNs for NC ASO+AP, n=121 iMNs for RAB27A ASO+DMSO and n=111 iMNs for RAB27A ASO+AP. For sALS5, n=217 iMNs for NC ASO+DMSO, n=103 iMNs for NC ASO+AP, n=106 iMNs for RAB27A ASO+DMSO and n=104 iMNs for RAB27A ASO+AP. For FUS R522R, n=133 iMNs for NC ASO+DMSO, n=111 iMNs for NC ASO+AP, n=92 iMNs for RAB27A ASO+DMSO and n=104 iMNs for RAB27A ASO+AP. Mean of biological replicates (independent conversions) ± s.e.m. n=3 independent conversions/line/condition. Statistical significance (one-way ANOVA) was calculated comparing NC ASO+AP to NC ASO+DMSO or RAB27A ASO+AP for each sALS and FUS line. P-values were corrected for multiple comparisons using statistical hypothesis testing (Šídák method). (H-I) Chronic PIKFYVE knock-down by RNAi (H) and a chronic 5-day apilimod treatment (10 µM) throughout larval development (I) ameliorated locomotor deficits in Drosophila larvae overexpressing TDP-43 G298S in motor neurons. Genotypes and treatments, as indicated. N=33 larvae/group. Ordinary one-way ANOVA. Mean ± s.e.m. P-values were corrected for multiple comparisons using statistical hypothesis testing (Šídák method). 78 Figure 3.6 Pikfyve suppression improves motor function and extends survival of TDP-43 and C9ORF72 mice. (A) Immunoblot and quantification of Optineurin (OPTN) in the CSF 6 days after 25 μg of NC (n=14 mice) or Pikfyve ASO (n=28 mice) intracerebroventricular administration in neonatal mice. The values were calculated as the relative intensity of OPTN to total protein. Unpaired t-test. Mean± s.e.m. Each data point represents one mouse. 79 (B) Gait impairment scores of Pikfyve +/+ ;WT (n=9 mice), Pikfyve +/- ;WT (n=10 mice), Pikfyve +/+ ;TDP-43 Tg/Tg (n=12 mice) and Pikfyve +/- ; TDP-43 Tg/Tg (n=7 mice) mice. A score of 0 indicates no phenotype and 4 indicates the most severe phenotype. Statistical significance was calculated by comparing Pikfyve +/+ ;TDP-43 Tg/Tg and Pikfyve +/- ; TDP-43 Tg/Tg at each time point. Unpaired t-test. Mean ± s.e.m. (C) Kaplan–Meier survival curves comparing the survival of Pikfyve +/+ ;TDP-43 Tg/Tg (n=12 mice) and Pikfyve +/- ;TDP-43 Tg/Tg (n=7 mice) mice. Log-rank test. (D) Gait impairment scores of WT mice treated with vehicle and negative control (NC) ASO (n=10 mice), WT mice treated with vehicle and Pikfyve ASO (n=13 mice), TDP-43 Tg/Tg mice treated with vehicle and NC ASO (n=12 mice), TDP-43 Tg/Tg mice treated with vehicle and Pikfyve ASO (n=16 mice). Postnatal day 1 (P1) mouse pups received 25 μg of NC or Pikfyve ASO by intracerebroventricular injection. Vehicle were administered by intraperitoneal injection every 48 hrs starting from P5. Statistical significance was calculated by comparing the TDP-43 Tg/Tg +vehicle +NC ASO group to the TDP-43 Tg/Tg +vehicle +Pikfyve ASO group at each time point. Unpaired t- test. Mean ± s.e.m. (E) Kaplan–Meier survival curves comparing the survival of TDP-43 Tg/Tg mice treated with vehicle and NC ASO (n=12 mice) or vehicle and Pikfyve ASO (n=21 mice). Postnatal day 1 (P1) mouse pups received 25 μg of NC or Pikfyve ASO by intracerebroventricular injection. Vehicle were administered by intraperitoneal injection every 48 hrs starting from P5. Statistical significance was calculated by comparing the +vehicle +NC ASO group to the +vehicle +Pikfyve ASO group. Log- rank test. (F) Immunoblot and quantification of optineurin (OPTN) in the CSF 6 days after administration of GW4869 and 25 μg of NC ASO (n=10 mice) or GW4869 and 25 μg of Pikfyve ASO (n=22 mice) in WT mice. Values were calculated as the relative intensity of OPTN normalized to total protein. Two-tailed unpaired t-test. Mean± s.e.m. Each data point represents one mouse. (G) Gait impairment scores of TDP-43 Tg/Tg mice treated with vehicle and NC ASO (n=12 mice), vehicle and Pikfyve ASO (n=16 mice), GW4869 and NC ASO (n=10 mice), and GW4869 and Pikfyve ASO (n=11 mice). Postnatal day 1 (P1) mouse pups received 25 μg of NC or Pikfyve ASO by intracerebroventricular injection. Vehicle or GW4869 were administered by intraperitoneal injection every 48 hrs starting from P5. Statistical significance was calculated by comparing the TDP-43 Tg/Tg mice +vehicle +Pikfyve ASO group to the TDP-43 Tg/Tg mice +GW4869 +Pikfyve ASO group at each time point. Unpaired t-test. Mean ± s.e.m. (H) Hindlimb clasping score of AAV-eGFP-(GR) 100 mice treated with NC ASO (n=5 mice) or Pikfyve ASO (n=10 mice). Intracerebroventricular injection of 2 μl (1 × 10 10 genomes/μl) of AAV- eGFP-(GR) 100 virus was performed in postnatal day1 (P1) C57BL/6J pups. Postnatal day 4 (P4) mouse pups received 25 μg of NC or Pikfyve ASO by intracerebroventricular injection. Statistical significance was calculated by comparing AAV-(GR) 100+NC ASO to AAV-(GR) 100 +Pikfyve ASO at P22, P24 and P26 . Unpaired t-test. Mean ± s.e.m. (I) Kaplan–Meier survival curves comparing the survival of AAV-(GR) 100 mice treated with NC ASO (n=5 mice) and Pikfyve ASO (n=11 mice). Intracerebroventricular injection of 2 μl (1 × 10 10 genomes/μl) of AAV-eGFP-(GR) 100 mice virus was performed in postnatal day1 (P1) C57BL/6J pups. Postnatal day 4 (P4) mouse pups received 25 μg of NC or Pikfyve ASO by intracerebroventricular injection. Statistical significance was calculated by comparing the AAV- eGFP-(GR)100+NC ASO and AAV-eGFP-(GR) 100 +Pikfyve ASO groups. Log-rank test. 80 Figure 3.7 Pikfyve suppression reduces TDP-43 and C9ORF72 pathology and neurodegeneration in vivo. (A) Representative images of NeuroTrace (Nissl)-positive lateral motor column (LMC) spinal motor neurons in the lumbar spinal cord in WT or TDP-43 Tg/Tg mice treated with vehicle or GW4869 and NC or Pikfyve ASO. The LMC motor neuron region in the spinal cord ventral horn is marked with a red dotted-line. Scale bar = 100 μm. 81 (B) Quantification of (A). Each data point represents the average number of NeuroTrace-positive LMC motor neurons/ventral horn (hemicord) section for one mouse. In vehicle treatment groups, n=10 mice for WT treated with NC, n=9 mice for WT treated with Pikfyve ASO, n=10 mice for TDP-43 Tg/Tg mice treated with NC ASO and n=7 mice for TDP-43 Tg/Tg treated with Pikfyve ASO. In GW4869 treatment groups, n=7 mice for WT treated with NC or Pikfyve ASO, n=9 mice for TDP- 43 Tg/Tg treated with NC ASO, and n=6 mice for TDP-43 Tg/Tg treated with Pikfyve ASO. 4-6 ventral horn sections were quantified and averaged per mouse. One-way ANOVA. Mean ± s.e.m. P- values were corrected for multiple comparisons using statistical hypothesis testing (Šídák method). (C) Immunostaining of pTDP-43+ (Ser403/404) punctae in TUJ1+ motor neurons in the spinal cord ventral horn from WT or TDP-43 Tg/Tg mice treated with NC or Pikfyve ASO. Solid and dotted lines outline the cell body and nucleus, respectively. Neuron cell bodies were identified using TUJ1 immunostaining. Orange arrows marked pTDP-43+ punctae. Scale bar = 10 μm. (D) Quantification of (C). Each data point represents the average number of pTDP-43+ (Ser403/404) punctae in TUJ1+ motor neurons per μm 2 for one mouse. n=19 mice for WT treated with NC or Pikfyve ASO, n=18 mice for TDP-43 Tg/Tg treated with NC ASO, and n=10 mice for TDP- 43 Tg/Tg treated with Pikfyve ASO. 40-50 cells were analyzed per mouse. One-way ANOVA. Mean ± s.e.m. Pp-values were corrected for multiple comparisons using statistical hypothesis testing (Šídák method). (E) Immunostaining of pTDP-43+ (Ser403/404) punctae in TUJ1+ motor neurons in the spinal cord ventral horn from WT or TDP-43 Tg/Tg mice treated with GW4869 and NC or Pikfyve ASO. Solid and dotted lines outline the cell body and nucleus, respectively. Neuron cell bodies were identified using TUJ1 immunostaining. Orange arrows mark pTDP-43+ punctae. Scale bar = 10 μm. (F) Quantification of (E). Each data point represents the average number of pTDP-43+ (Ser403/404) punctae in TUJ1+ motor neurons per μm 2 for one mouse. n=5 mice for WT treated with GW4869 and NC ASO, n=6 mice for WT treated with GW4869 and Pikfyve ASO, and n=6 mice for TDP-43 Tg/Tg treated with GW4869 and NC or Pikfyve ASO. 40-50 cells were analyzed per mouse. One-way ANOVA. Mean ± s.e.m. P-values were corrected for multiple comparisons using statistical hypothesis testing (Šídák method). (G) Immunostaining of total TDP-43 in TUJ1+ spinal motor neurons in the ventral horn of WT or TDP-43 Tg/Tg mice treated with NC or Pikfyve ASO. Solid and dotted lines outline the cell body and nucleus, respectively. Neuron cell bodies were identified using TUJ1 immunostaining. Scale bar = 10 μm. (H) Quantification of (G). Each data point represents the average ratio of nuclear to cytoplasmic TDP-43 in TUJ1+ motor neurons for one mouse. n=7 mice for WT treated with NC or Pikfyve ASO, n=11 mice for TDP-43 Tg/Tg treated with NC ASO, and n=10 mice for TDP-43 Tg/Tg treated with Pikfyve ASO. 20-30 cells were analyzed per mouse. One-way ANOVA. Mean ± s.e.m. P-values were corrected for multiple comparisons using statistical hypothesis testing (Šídák method). (I) Immunostaining of pTDP-43+ (Ser403/404) punctae in IBA1+ microglia in the spinal cord ventral horn from WT or TDP-43 Tg/Tg mice treated with NC or Pikfyve ASO. Solid lines outline the cell body. Microglial cell bodies were identified using IBA1 immunostaining. Orange arrows marked pTDP-43+ punctae. Scale bar = 10 μm. (J) Quantification of (I). Each data point represents the average number of pTDP-43+ (Ser403/404) punctae in IBA1+ microglia per μm 2 for one mouse. n=11 mice for WT treated with NC ASO, n=13 mice for WT treated with Pikfyve ASO, and n=12 mice for TDP-43 Tg/Tg treated with NC ASO or Pikfyve ASO. 20-30 cells were analyzed per mouse. One-way ANOVA. Mean ± s.e.m. One-way ANOVA. Mean ± s.e.m. P-values were corrected for multiple comparisons using statistical hypothesis testing (Šídák method). (K) Representative images of NeuroTrace (Nissl)-positive lateral motor column (LMC) spinal motor neurons in the lumbar spinal cord in 10-month-old Pikfyve +/+ ;WT, Pikfyve +/- ;WT, 82 Pikfyve +/+ ;TDP-43 Tg/+ and Pikfyve +/- ; TDP-43 Tg/+ mice. The LMC motor neuron region in the spinal cord ventral horn is marked with a red dotted-line. Scale bar = 50 μm. (L) Quantification of (K). Each data point represents the average number of NeuroTrace-positive LMC motor neurons/ventral horn (hemicord) section for one mouse (10-month-old). n=5 mice for Pikfyve +/+ ;WT, n=3 mice for Pikfyve +/- ;WT, n=4 mice for Pikfyve +/+ ;TDP-43 Tg/+ , and n=4 mice for Pikfyve +/- ; TDP-43 Tg/+ . 4-6 ventral horn sections were quantified and averaged per mouse. One- way ANOVA. Mean ± s.e.m. P-values were corrected for multiple comparisons using statistical hypothesis testing (Šídák method). (M) Immunostaining of pTDP-43+ (Ser409/410) punctae in TUJ1+ motor neurons in the spinal cord ventral horn from 10-month-old Pikfyve +/+ ;WT, Pikfyve +/- ;WT, Pikfyve +/+ ;TDP-43 Tg/+ and Pikfyve +/- ; TDP-43 Tg/+ mice. Solid and dotted lines outline the cell body and nucleus, respectively. Neuron cell bodies were identified using TUJ1 immunostaining. Orange arrows mark pTDP-43+ punctae. Scale bar = 5 μm. (N) Quantification of (M). Each data point represents the average number of pTDP-43+ (Ser403/404) punctae in TUJ1+ motor neurons per μm 2 for one mouse. n=5 mice for Pikfyve +/+ ;WT, n=3 mice for Pikfyve +/- ;WT, n=4 mice for Pikfyve +/+ ;TDP-43 Tg/+ , and n=4 mice for Pikfyve +/- ; TDP- 43 Tg/+ . 40-50 cells were analyzed per mouse. One-way ANOVA. Mean ± s.e.m. P-values were corrected for multiple comparisons using statistical hypothesis testing (Šídák method). (O) Representative images of NeuroTrace (Nissl)-positive lateral motor column (LMC) spinal motor neurons in the lumbar spinal cord in day 26 AAV-eGFP or AAV-eGFP-(GR) 100 mice treated with NC or Pikfyve ASO. The LMC motor neuron region in the spinal cord ventral horn is marked with a red dotted-line. Scale bar = 50 μm. (P) Quantification of (O). Each data point represents the average number of NeuroTrace-positive LMC motor neurons/ventral horn (hemicord) section for one mouse. n=7 mice for AAV-eGFP mice treated with NC ASO, n=7 mice for AAV-eGFP mice treated with Pikfyve ASO, n=11 mice for AAV-eGFP-(GR) 100 mice treated with NC ASO, and n=11 mice for AAV-(GR) 100 mice treated with Pikfyve ASO. 4-6 ventral horn sections were quantified and averaged per mouse. One-way ANOVA. Mean ± s.e.m. P-values were corrected for multiple comparisons using statistical hypothesis testing (Šídák method). 83 Chapter 4 : Harnessing Neuroprotective Microglia for Therapeutic Benefit in C9ORF72 ALS/FTD 4.1 Abstract Several genes expressed specifically in microglia harbor strongest risk alleles for dementia. Microglia display altered function in models of Alzheimer’s disease, frontotemporal dementia (FTD), and amyotrophic lateral sclerosis (ALS), and ablating them can profoundly affect neurodegeneration. Neuron-glia interaction is a key regulator for disease onset and progression. However, it remains unclear in which contexts neuron-microglial interactions drive or rescue neurodegeneration, and studies suggest this may depend on the form of stage of dementia. Thus, there is a pressing need to understand how neuron-microglia interactions modulate neurodegeneration and identify ways to leverage this knowledge for therapeutic benefit in dementia and related neurodegenerative diseases. The repeat expansion in C9ORF72 is the most common genetic cause of ALS and FTD, and it causes neurodegeneration through both loss- and gain-of-function processes. The prevailing view is that loss of C9ORF72 function activates inflammatory programs in microglia that promote neurodegeneration. However, no studies examined the effects of neuron-microglia interactions on neurodegeneration in the C9ORF72 ALS/FTD patient context. To study microglia in a C9ORF72 ALS/FTD patient context, we established an induced microglia (iMG) model by using two microglial transcription factors and IL-34 to convert induced pluripotent stem cells (iPSCs) into microglia. We verified that iMG possess phagocytotic ability, have similar transcriptomic signatures to primary human microglia, and respond appropriately to known activators such as lipopolysaccharide. To determine how microglia affect neurodegeneration in C9ORF72 ALS/FTD, we developed co-culture models using C9ORF72 ALS/FTD patient iMG and induced motor neurons (iMNs). We found that C9ORF72 iMG significantly reduced the survival of healthy control iMNs, indicating that the C9ORF72 repeat expansion induces a default neurotoxic state in microglia. Surprisingly, however, C9ORF72 iMG 84 significantly increased C9ORF72 ALS/FTD iMNs survival. We also found that C9ORF72 ALS/FTD iMG significantly lowered levels of neurotoxic, repeat expansion-derived dipeptide repeat proteins (DPRs) in neighboring neurons. In addition, we found that co-culture with iMNs from healthy individuals overexpressing DPRs, but not C9ORF72-deficient iMNs or sporadic ALS iMNs triggered microglia to become neuroprotective. Thus, C9ORF72 gain-of-function processes in neurons induce microglia switching into the neuroprotective state. From single-cell RNA seq, we found that co-culture with C9ORF72 ALS/FTD iMNs converted a subset of C9ORF72 ALS/FTD iMG into a neuroprotective state marked by reduced CSF1R expression and high IGF-1 signaling. We verified that flow-purified CSF1R-low C9ORF72 ALS/FTD iMG rescued survival of C9ORF72 ALS/FTD iMNs. Since CSF1R-low C9ORF72 ALS/FTD iMG were neuroprotective while the CSF1R-high C9ORF72 ALS/FTD iMG were neurotoxic, we found that suppressing CSF1R with an antisense oligonucleotide (ASO) shifted C9ORF72 ALS/FTD iMG gene expression into neuroprotective states in vitro. Moreover, GFP-(GR) 100 mice injected with Csf1r ASO showed improved behavior function and reduced neuronal loss in the lumbar region of the spinal cord. In contrast to the prevailing view, our findings suggest that a subset of microglia switch from neurotoxic to neuroprotective in C9ORF72 ALS/FTD. Converting microglia into the neuroprotective state via CSF1R ASO could be a therapeutic benefit for C9ORF72 ALS/FTD. 4.2 Introduction Microglia are resident immune cells in the brain that operate brain surveillance and maintain brain homeostasis. Originated from yolk-sac, myeloid progenitors, expressing Pu.1 and IRF8, migrate to the central nervous system (CNS) and develop into microglia (Kierdorf et al., 2013). In the brain, the major roles of microglia are to monitor the surrounding environments, to respond to stimulations, and to maintain neuronal plasticity by their functions such as phagocytosis, cytokine release and synaptic pruning (Salter and Stevens, 2017). 85 The critical roles of non-neuronal cells contributing to neurodegeneration have arisen over the past decade. Neuroinflammation triggered by microglia as well as neurons have been demonstrated to influence the disease progression in various neurodegenerative diseases. Several genes expressed selectively in microglia harbor strong risk alleles for Alzheimer’s disease and frontotemporal dementia (FTD) according to genome-wide association studies (McQuade and Blurton-Jones, 2019). Microglia display altered functions in models of Alzheimer’s disease, FTD, and amyotrophic lateral sclerosis (ALS) and ablating microglia in neurodegenerative diseases could profoundly affect disease progression (Zhang et al., 2020; Spiller et al., 2018). In Alzheimer’s disease, microglia cause excessive inflammation or synaptic pruning which accelerate neurodegeneration, in contrast, microglia get rid of amyloid plaques through their activation of phagocytic activity in the TREM2 regulated mechanisms (Jay et al., 2015; Wang et al., 2015). In FTD, progranulin deficient microglia shift to disease state and increase TDP-43 pathology and neuron loss in the mice (Zhang et al., 2020). However, after a pulse of TDP-43 overexpression in the mice model, microglia exert neuroprotection by lowering TDP-43 pathology and preventing neuronal loss (Spiller et al., 2018). However, it remains unclear if neuron-glial interactions drive or rescue neurodegeneration, how microglia mediate their effects on neurons, and how specific genetic mutations alter neuron-glial interactions. The hexanucleotide repeat expansion in the C9ORF72 locus is the most common genetic mutation in ALS and FTD, and it causes neurodegeneration through both loss- and gain-of- function processes (Shi et al., 2018). In the CNS, the gene expression of C9ORF72 varies across different cell types which myeloid cells and microglia highly express C9ORF72 (Zhang et al., 2016; Rizzu et al., 2016). For the loss-of-function mechanism, the expression of C9ORF72 proteins is reduced, affecting the normal function of C9ORF72 proteins. Sharing similar homology with the Differentially Expressed in Normal and Neoplasia (DENN) family, C9ORF72 has been considered as GDP/GTP exchange factor (GEF) and associated with regulating membrane and vesicle 86 trafficking in the cells (Levine et al., 2013; Zhang et al., 2018a). The C9orf72 depleted mice developed severe autoimmunity, displayed excessive accumulation of myeloid cells and showed enlarged spleen and lymph nodes (O’Rourke et al., 2016; Burberry et al., 2016). The prevailing view in the field is that the C9ORF72 deficiency causes microglia to become inflammation active and trigger neurotoxic programs. Recent studies have shown that the C9ORF72 -/- microglia upregulated type I interferon signalings, suppressed the degradation of STING (stimulator of interferon genes) proteins, and caused synaptic loss in the Alzheimer’s disease mice model (McCauley et al., 2020; Lall et al., 2021). The gain-of-function mechanism is featured by long repeat-containing RNA transcription and repeat-associated non-ATG (RAN) translation, leading to toxicity of the accumulation of RNA foci and dipeptide-repeat proteins (DPRs) (Cleary and Ranum, 2013; Sareen et al., 2013). Overexpression of C9orf72 repeat expansion in microglia cell line, BV-2, caused accumulations of DPRs in microglia, while did not lead to accumulation of RNA foci and did not affect microglia functions, such as phagocytic activity (Rostalski et al., 2020). However, no studies have examined the effects of neuron-microglia interactions on neurodegeneration in the C9ORF72 patient context where both loss-and gain-of function mechanisms. In this study, we used cellular reprogramming strategy to generate induced motor neurons (iMNs) and induced microglia (iMG) from C9ORF72 ALS/FTD patients and developed co-culture models of microglia with motor neurons to study the effect of microglia in C9ORF72 ALS/FTD neurodegeneration. Our data suggested that C9ORF72 ALS/FTD microglia possessed distinct transcriptomic signatures from healthy control microglia with altered endolysosomal phenotypes. Using the co-culture model, we revealed that C9ORF72 FTD/ALS microglia significantly increased C9ORF72 FTD/ALS motor neuronal survival and decreased the accumulation of DRPs in motor neurons. We further uncovered a unique subset of C9ORF72 ALS/FTD microglia, characterized by low CSF1R expression, exerted neuroprotective effects and presented specifically when co- 87 culturing with C9ORF72 ALS/FTD motor neurons and were thought to convert from the default CSF1R-high and neurotoxic state. Utilizing small molecular CSF1R inhibitors or antisense oligonucleotides targeting CSF1R efficiently converted CSF1R-high C9ORF72 ALS/FTD microglia into CSF1R-low neuroprotective state both in vitro and in vivo. Overall, our study defined a novel phenotypic switch in C9ORF72 ALS/FTD microglia which transformed them from neurotoxic to neuroprotective and leveraging this switch could be served as therapeutic benefits in C9ORF72 ALS/FTD and other neurodegenerative disorders. 4.3 Results Transcription factor-mediated conversion of iPSCs into induced microglia To establish an efficient approach to generate human microglia, we used transcription factor- mediated reprogramming to generate human microglia from induced pluripotent stem cells (iPSCs). In order to identify the transcription factors that promote microglial fate, we selected twenty transcription factors (TFs) involved in microglial development and differentiation, utilized retrovirus transduction system, and screened the ability of each transcription factor to activate the expression of TMEM119, a microglial specific marker, in human iPSCs (Fig S4.1A-B). By dropping out each transcription factor, we found that Pu.1, a myeloid lineage determining transcription factor, is required to activate the Tmem119-T2A-tomato reporter, and the other five transcription factors also play crucial roles: IRF8 (interferon regulatory factor 8), IRF5 (Interferon regulatory factor 5), MAFB (MAF BZIP Transcription Factor B), CEBPA (CCAAT/enhancer-binding protein alpha), and ATF3 (activating Transcription Factor 3) (Fig S4.1C) (Dräger et al., 2022). To increase the conversion of microglia rather than macrophages, we supplemented the cultures with colony stimulating factor 1 (CSF1) and Interleukin 34 (IL-34). We further optimized the cocktail via transducing Pu.1 with different combinations of other five factors and found that the forced expression of Pu.1 and CEBPA more robustly generated more TMEM119+ cells (Fig 4.1A-B). To 88 validate the microglial identity of Tmem119+ cells, we FACS-purified Tmem119+ cells and performed immunostaining to confirm the expression of canonical microglial markers including MAFB, IBA1, CD11B and CX3CR1 in these cells (Fig 4.1C). Bulk RNA sequencing (RNA-seq) confirmed the global transcriptomic signatures of Tmem119+ cells were distinct from iPSCs and fibroblasts (PC1, 63% variance) and similar to primary human microglia (PC2, 19% variance) (Fig 4.1D, S4.1D). The analysis of differential expressions genes between Tmem119+ cells and iPSCs also suggested that our Tmem119+ cells expressed a majority of microglial markers such as CSF1R, GPR34, TREM2, P2RY12, C1QA, AIF1, CTSD and IRF8, while down-regulated the pluripotent marker, such as KIF4A and MYC (Fig S4.1E) (Muffat et al., 2016; Abud et al., 2017). These results indicate that we can generate induced microglia-like cells (iMG) by forced expression of two microglial transcription factors. To examine whether the iMG were functional microglia, we tested the phagocytic activity and inflammatory response of our iMG. iMG robustly phagocytosed fluorescent pHrodo beads with 60% of iMG and 96% of iMG possessing phagocytic activity after 2 and 24 hours respectively (Fig 4.1E). iMG also responded appropriately to 24 hours of lipopolysaccharide (LPS) stimulation via secreting inflammatory cytokines such as IL-1B, IL-6 and IL-18 (Fig 4.1G). Furthermore, iMG efficiently integrated into human cortical brain organoids and interacted with neurons within seven days (Fig 4.1H). Thus, we verified that iMG possessed molecular and functional attributes of microglia. C9ORF72 ALS/FTD microglia display different transcriptomic profiles and alter endolysosome phenotypes To study the intrinsic phenotypes of patient-derived microglia, we utilized the overexpression of Pu.1 and CEBPA to convert healthy control and C9ORF72 ALS/FTD patient iPSCs into iMG, isolated iMG via FACS sorting with CD11b, a microglial surface marker highly expressed microglia, and performed bulk RNA-sequencing (Fig S4.1F, S4.2A). To characterize whether C9ORF72 89 ALS/FTD microglia possessed unique features, we generated four genotypes of iMG: (1) control iMG (n=2), (2) C9ORF72 ALS/FTD iMG (n=2), (3) C9ORF72 homozygous knockout (C9ORF72 - /- ) iMG (n=1) and (4) C9ORF72 isogenic iMG (n=2). C9ORF72 -/- line was generated through introducing a frame-shift mutation into both alleles of the control line to mimic the loss-of-function mechanism, while C9ORF72 isogenic lines were generated from correcting the repeat expansion to normal length in C9ORF72 ALS patient lines to study gain-of-function mechanism. Principal component analysis revealed C9ORF72 ALS/FTD iMG were clearly separated from control iMG in the first principal components (PC1, 32% variance) (Fig 4.2A). C9ORF72 -/- iMG were in proximity to control iMG, indicating the loss of C9ORF72 proteins did not change the transcriptomic signatures of microglia. C9ORF72 isogenic iMG were distinct from C9ORF72 ALS/FTD iMG in PC2 (16% variance), suggesting the long repeat expansion of C9ORF72 affected the transcriptomic profiles of microglia. Using unsupervised hierarchical clustering, we identified 1,537 differentially expressed genes (DEGs, adjusted p value < 0.01) in C9ORF72 ALS/FTD iMG compared to control iMG, of which 582 genes were up-regulated and 955 genes were down- regulated (Fig 4.2B, S4.2B). C9ORF72 ALS/FTD iMG reduced the expression of homeostatic markers such as TGFBR1, P2RY12 and MEF2C and increased APOE expression (Fig 4.2C), showing a similar shift from TGFβ1 to APOE signaling in microglia after phagocytosing β-amyloid (Ab)-plaques (Keren-Shaul et al., 2017). Gene set enrichment analysis of the 582 up-regulated DEGs in C9ORF72 ALS/FTD iMG showed enrichment in various microglial activation pathways including TYROBP signaling, microglial pathogen phagocytosis, complement cascades and cytokine response pathways (Fig 4.2D, S4.2C). We observed that deletion of long repeat expansion in C9ORF72 ALS/FTD background (C9ORF72 isogenic iMG) resulted in significantly downregulation of these pathways, indicating that the C9ORF72 gain-of-function in microglia might drive the activation of microglia (Fig S4.2D-F). In the comparison of C9ORF72 -/- iMG to CTRL iMG, there were 524 DEGs (adjusted p value < 0.01) with 165 up-regulated genes including 90 SPP1, C3, MMP2 and MMP9, and 359 down-regulated genes related to p53 and interferon signalings (Fig S4.2B, S4.2G). We previously showed that C9ORF72 is required for normal vesicle trafficking and lysosomal biogenesis in motor neurons (Shi et al., 2018). To assess the impact of C9ORF72 repeat expansion in microglia, we examined the expression of early endosomal marker EEA1 and lysosomal marker LAMP2. C9ORF72 microglia possessed enlarged EEA1+ endosomes and large clusters of LAMP2+ lysosomes, demonstrating the altered endocytosis and lysosomal phenotypes (Fig 4.2E-F). Next, we sought to assess whether C9ORF72 microglia had impaired ability to degrade phagocytic materials by measuring the percentage of microglia with fluorescent signals after 24, 48 and 72 hours of adding pHrodo beads. After 72 hours, the percentage of C9ORF72 microglia with fluorescent signals were higher than control microglia, suggesting that the C9ORF72 microglia have slower speed to degrade phagocytosed materials (Fig 4.2G). Together, these results provide support for naturally distinct functional features of C9ORF72 microglia from healthy control microglia. C9ORF72 ALS/FTD microglia exert neuroprotection on C9ORF72 ALS/FTD neurons To investigate the effect of microglia in neurodegeneration, we established a motor neurons- microglia co-culture system using transcription factor-mediated reprogramming to generate both cell types. We used seven neuronal transcription factors, Ascl1, Brn2, Myt1l, Isl1, Lhx3, Ngn2, NeuroD1, and a lentiviral Hb9::RFP reporter to convert control and C9ORF72 ALS/FTD iPSCs into Hb9::RFP+ induced motor neurons (iMNs) as we previously reported 8 . In the meantime, we generated induced microglia (iMG) with Pu.1 and CEBPA overexpression in control and C9ORF72 ALS/FTD iPSCs. On day 18 of iMNs differentiation, we harvested CD11b+ iMG and seeded into the monolayer culture of iMNs with the ratio of three motor neurons to one microglia. We previously characterized that C9ORF72 ALS/FTD iMNs degenerated significantly faster than control iMNs and developed dipeptide repeat proteins (DPR) aggregates in the cell body which 91 recapitulate neurodegenerative processes in C9ORF72 ALS/FTD (Fig S4.3A-C) (Shi et al., 2018). To access the impact of microglia in C9ORF72 ALS/FTD disease context, we set up four different combinations of co-cultures: control iMNs with control iMG, control iMNs with C9ORF72 ALS/FTD iMG, C9ORF72 ALS/FTD iMNs with control iMG and C9ORF72 ALS/FTD iMNs with C9ORF72 ALS/FTD iMG (Fig 4.3A). We performed longitudinal tracking of Hb9::RFP+ iMNs under these four co-cultures to study the effect of microglia on neuronal survival. The survival results were presented as hazard ratio. The hazard rate is defined as the likelihood of death for a motor neuron during a specified time interval, and we normalized the hazard rate for each group to a control condition (hazard ratio =1) to generate a hazard ratio. Higher hazard ratio indicated the motor neurons degenerated faster than control motor neurons and had poor neuronal survival. In accordance with the notion that C9ORF72 repeat expansion shifts microglia into a neurotoxic state, C9ORF72 ALS/FTD iMG, but not control iMG, reduced the survival of control iMNs, as shown by hazard rate being >1 (Fig 4.3B, S4.3D). Control iMG improved the survival of C9ORF72 ALS/FTD iMNs, indicating the healthy microglia exert the normal function to maintain brain homeostasis. Interestingly however, C9ORF72 ALS/FTD iMG robustly increased C9ORF72 ALS/FTD iMNs survival, shown as the reduction of hazard ratio compared to no microglia condition in C9ORF72 ALS/FTD iMNs. Loss of C9ORF72 in microglia was sufficient to induce the neuroprotective effect, as C9ORF72 -/- iMG also reduced control iMNs survival but increased the survival of C9ORF72 ALS/FTD iMNs (Fig 4.3C, S4.3E). We further tested the effect of microglia on DPR pathology in C9ORF72 ALS/FTD iMNs. Immunostaining of endogenous poly(GR) detected a significantly reduced punctae in C9ORF72 ALS/FTD iMNs when co-culturing with both control and C9ORF72 ALS/FTD iMG (Fig 4.3D-E). Considering the low density of iMNs and iMG in the co-cultures, we tested whether secreted factors mediated the C9ORF72 ALS/FTD microglia effects. Cytokine profiling of all combinations of control and C9ORF72 iMNs-iMG co-cultures identified IL-1β and GRO- ⍺ as neurotoxic cytokines which specifically increased in control iMN- C9ORF72 ALS/FTD iMG co-cultures, while IL-10 and CX3CL1 as neuroprotective cytokines 92 which most significantly elevated in C9ORF72 ALS/FTD iMN-C9ORF72 iMG co-cultures (Fig 4.3F). To validate the effect of selected cytokines on neurons, we directly treated control iMNs with IL-1β and GRO- ⍺ and C9ORF72 ALS/FTD iMNs with IL-10 and CX3CL1 and monitored the neuronal survival. In agreement with the cytokine profiling results, recombinant IL-1β and GRO- ⍺ decreased the survival of control iMNs survival while recombinant IL-10 and CX3CL1 improved the survival of C9ORF72 ALS/FTD iMNs (Fig 4.3G-H). Combined, these results suggested that C9ORF72 ALS/FTD microglia affected neuronal survival through cytokine secretion. A novel CSF1R-low neuroprotective state of C9ORF72 ALS/FTD microglia To gain deeper understanding of the molecular mechanisms defining neurotoxic and neuroprotective states in C9ORF72 ALS/FTD microglia, we isolated Hb9::RFP+ iMNs and CD11b+ microglia after 3 days of co-cultures and performed single cell RNA-sequencing (sc-RNA seq) on all four combinations of control and C9ORF72 ALS/FTD iMNs-iMG co-cultures (Fig S4.4A). We performed principal component analysis and identified microglial clusters in each condition with the expression of AIF1, also known as IBA1 (Fig 4.4A-D, S4.4B-E). One cluster of AIF1+ microglia with a modest level of CSF1R expression was identified in control iMNs-control iMG co-cultures, suggesting the fact that microglia utilize CSF1R signaling for differentiation and maturation (Fig S4.4B-C, S4.4F) (Elmore et al., 2014). In co-culturing with C9ORF72 ALS/FTD iMNs, AIF1+ control iMG concentrated within two close clusters expressing high CD68 and low CSF1R, which we thought to be neuroprotective (Fig 4.4SD-E, S4.4F). A single cluster of AIF1+ C9ORF72 ALS/FTD microglia was detected with high expression of CSF1R and CD68 when co- culturing with control iMNs, as previously described as neurotoxic microglia (Fig 4.4A-B, S4.4F). Strikingly, our analysis revealed two separated clusters (cluster 6 and cluster 7) of AIF1+ microglia in the C9ORF72 ALS/FTD iMNs-C9ORF72 ALS/FTD iMG co-cultures (Fig 4.4C-D, S4.4F). Cluster 6 microglia displayed the high CSF1R expression and was termed as CSF1R-high state. CSF1R-high microglia also highly expressed CD68 and TYROBP and had similar transcriptomic 93 signatures with C9ORF72 ALS/FTD microglia co-culturing with control neurons as expected to be neurotoxic. Surprisingly, we identified a distinct microglia population (cluster 7) of CSF1R-low C9ORF72 ALS/FTD microglia that only appeared when co-culturing with C9ORF72 ALS/FTD neurons. The CSF1R-low microglia (cluster 7) did not express the activation marker CD68 and TYROBP, while expressed high levels of TGFβ1, IGF1, the transcription factors MEF2C, MAF, and PBX1 (Fig 4.4E, S4.4F). To further understand the difference between CSF1R-high and -low state, we performed differential gene expression analysis across the clusters and conducted enrichment pathway analysis on both microglia populations. CSF1R-high microglia were highly enriched in TYROBP signaling, interferon signaling and complement cascades pathways, indicating that CSF1R-high microglia were in active and inflammatory state. In contrast, the pathways enriched in CSF1R-low microglia were proteasome, endocytosis, TGFβ signaling and IGF-1 signaling, suggesting that CSF1R-low microglia were in a distinct state from CSF1R-high microglia (Fig 4.4F). Pseudotime analysis of single cell RNA-seq data showed that C9ORF72 FTD/ALS iMG started in the CSF1R-high state, but some iMG converted to the CSF1R-low state over time if cultured with C9ORF72 FTD/ALS iMNs (Fig. 4.4G). Together, we found control iMG cultured with control iMNs has medium levels of CSF1R expression. C9ORF72 ALS/FTD iMG cultured with control iMNs had high CSF1R expression and was neurotoxic. Control iMG and a subset of C9ORF72 ALS/FTD iMG became CSF1R-low when cultured with C9ORF72 ALS/FTD iMNs and were neuroprotective. Since C9ORF72 ALS/FTD iMG were only protective toward C9ORF72 ALS/FTD iMNs and CSF1R-low C9ORF72 ALS/FTD only presented in co-culturing with C9ORF72 ALS/FTD iMNs, we hypothesized that the CSF1R-low microglia were the neuroprotective population. To examine the neuroprotective role of CSF1R-low microglia, the C9ORF72 ALS/FTD iMG were stained with fluorescent-tagged CSF1R and CD11b antibodies and subjected to FACS to isolated CSF1R- low/CD11b+ and CSF1R-high/CD11b+ populations. CSF1R-high and CSF1R-low iMG were co- cultured individually with control or C9ORF72 ALS/FTD iMNs to track the neuronal survival. 94 Indeed, CSF1R-low C9ORF72 ALS/FTD microglia rescued the survival of C9ORF72 ALS/FTD iMNs. In contrast, CSF1R-high C9ORF72 ALS/FTD microglia did not rescue C9ORF72 ALS/FTD iMNs and were toxic to control iMNs (Fig 4.4H, S4.4G). Taken together, we uncovered a unique CSF1R-low microglia population in C9ORF72 ALS/FTD and revealed its potential of reverting neurodegeneration. Gain-of-function mechanism of C9ORF72 ALS/FTD neurons drives the conversion of C9ORF72 ALS/FTD microglia to neuroprotective state As previously mentioned, neurons and microglia might communicate through secreted factors due to the low density of both populations in our co-culture system. It is possible that C9ORF72 ALS/FTD neurons contributed to the conversion of C9ORF72 ALS/FTD microglia via secreted factors. To elucidate this, we collected the conditioned media from C9ORF72 ALS/FTD iMNs cultures and added directly to control iMNs-C9ORF72 ALS/FTD iMG co-cultures which the microglia were originally in CSF1R-high neurotoxic state. We found that C9ORF72 ALS/FTD iMNs conditioned media was sufficient to shift C9ORF72 ALS/FTD microglia from neurotoxic into neuroprotective state by inhibiting neurodegeneration (Fig 4.5A, S4.5A-B). This reinforced the idea that secreted factors play a critical role in crosstalk between neurons and microglia. Two pathogenic mechanisms of the C9ORF72 repeat expansion in neurons could result to the shift of microglial state: One is loss-of-function mechanism characterized by loss of C9ORF72 proteins; the other is gain-of-function mechanism featured by translating the repeat expansions into dipeptide repeat proteins. To mimic loss-of-function mechanisms in neurons, we co-cultured C9ORF72 ALS/FTD microglia with C9ORF72 -/- iMNs. C9ORF72 -/- iMNs were known to undergo rapid neurodegeneration compared to healthy control iMNs; however, C9ORF72 ALS/FTD microglia were unable to rescue the degeneration (Fig 4.5B, S4.5C). In order to mimic the gain- of-function mechanism, we overexpressed poly(GR) 50-GFP in control iMNs and showed that DPR-overexpressed control iMNs degenerate much faster than control iMNs expressing GFP 95 alone. Significantly, C9ORF72 ALS/FTD microglia rescue the control iMNs with DPR expression (Fig 4.5C, S4.5D). In conclusion, C9ORF72 gain-of-function processes in neurons was responsible for driving the microglial into neuroprotective state. C9ORF72 ALS/FTD microglia shift from CSF1R-high to CSF1R-low neuroprotective state in vitro Since CSF1R-low microglia were neuroprotective while CSF1R-high microglia were neurotoxic in C9ORF72 ALS/FTD, we hypothesized that reduction of CSF1R signaling might promote the neuroprotective signatures in C9ORF72 ALS/FTD microglia. To determine if CSF1R inhibition shift C9ORF72 ALS/FTD microglia into CSF1R-low state, we examined marker genes that were highly differentially expressed between CSF1R-low and CSF1R-high C9ORF72 ALS/FTD microglia in our single cell RNA-seq data. CSF1R-low C9ORF72 ALS/FTD microglia displayed reduced CSF1R expression and elevated MEF2C, IGF1 and TGFβ1 levels (Fig 4.4E). We utilized two approaches to suppress CSF1R expression: (1) small molecule CSF1R inhibitors JNJ- 40346527 or PLX3397, and (2) antisense oligonucleotides (ASOs) targeting CSF1R. qRT-PCR analysis on FACS-purified CD11B+ C9ORF72 ALS/FTD microglia showed that a 24 hour transient treatment with CSF1R inhibitor JNJ-40346527 decreased CSF1R and increased IGF1 and TGFβ1 expression (Fig 4.6A). In addition, transient treatment of PLX3397 on C9ORF72 ALS/FTD microglia showed similar changes in marker gene expression in a dose-dependent manner (Fig S4.6A). In support of this notion, multiple ASOs that targeted human CSF1R shifts C9ORF72 ALS/FTD microglia toward the CSF1R-low states as evidenced in part by reduced CSF1R and increased IGF1 expression (Fig 4.6B). Thus, these results provide support for CSF1R inhibition as therapeutic potential to keep C9ORF72 ALS/FTD microglia in the neuroprotective state. In order to delineate the regulatory network the switch from CSF1R-high to CSF1R-low C9ORF72 ALS/FTD microglia, we reasoned that knocking down transcription factors specifically upregulated 96 in CSF1R-high state such as NR1H3 (Liver X receptor alpha) or overexpressing transcription factors uniquely expressed in CSF1R-low state such as MEF2C (MADS box transcription enhancer factor 2) could be efficient to trigger the conversion from CSF1R-high to CSF1R-low state (Fig S4.6B). We used small interfering RNA (siRNA) targeting each gene to treat C9ORF72 ALS/FTD microglia and checked if suppression of each gene is sufficient to trigger the CSF1R- low state. Among them, qRT-PCR analysis of siRNA-treated C9ORF72 ALS/FTD microglia showed that suppression of NR1H3 lead to inhibition of CSF1R expression and promotion of MEF2C and TGFβ1 expression (Fig 4.6C). Moreover, we validated that overexpression of MEF2C in C9ORF72 ALS/FTD microglia not only suppressed CSF1R expression but increased IGF1 expression, indicating the presence of CSF1R-low microglia (Fig S4.6C). Thus, in addition to CSF1R inhibition, suppression of NR1H3 or overexpression of MEF2C also contributed to the conversion to CSF1R-low state. To deeply understand the regulatory networks governing CSF1R-high and CSF1R-low states in C9ORF72 ALS/FTD microglia, we further analyzed our single cell RNA-seq data with SCENIC (single-cell regulatory network inference and clustering) which simultaneously reconstructing gene regulatory network and identifying cell states (Aibar et al., 2017; Van de Sande et al., 2020). When co-cultured with C9ORF72 ALS/FTD iMNs, SCENIC revealed three states of C9ORF72 ALS/FTD microglia driven by different transcription factors and regulons (Fig 4.6SD). CSF1R-high C9ORF72 ALS/FTD microglia fall into two closely related states regulated by transcription factors such as NR2F6, MYC and ETV6 (Fig 4.6D). On the other hand, CSF1R-low C9ORF72 ALS/FTD microglia formed a distinct state driven by transcription factors including MAF, PBX1 and YY1, highlighting CSF1R-low C9ORF72 ALS/FTD microglia had distinct gene regulatory network from CSF1R-high microglia (Fig 4.6D). Intercellular communication plays a pivotal role in modulating the phenotypic changes between different cell types. In our co-culture systems, we previously demonstrated that neurons and microglia signaled through secreted factors. To decipher the crosstalk between neurons and 97 microglia in C9ORF72 ALS/FTD, we analyzed our single cell RNA-seq data with iTALK package (identifying and illustrating alterations in intercellular signaling network) (Wang et al.). To characterize ligand-receptor pairs specifically appeared in C9ORF72 ALS/FTD neurons and microglia, we also performed analysis on the single cell data from control neurons-C9ORF72 ALS/FTD microglia co-cultures and eliminated the ligand-receptor pairs that overlapped. We expected to identify ligand-receptor pairs from both directions. One is ligands expressed on C9ORF72 ALS/FTD neurons and receptors expressed on CSF1R-high C9ORF72 ALS/FTD microglia, which indicates the signals from C9ORF72 ALS/FTD neurons to CSF1R-high C9ORF72 ALS/FTD microglia to modulate the switch to CSF1R-low neuroprotective state. The other direction is ligands expressed on CSF1R-low C9ORF72 ALS/FTD microglia and receptors on C9ORF72 ALS/FTD neurons, which delivering neuroprotective substrates to C9ORF72 ALS/FTD neurons to promote neuron survival. For the signals from C9ORF72 ALS/FTD neurons to drive the switch of CSF1R-high C9ORF72 ALS/FTD microglia, we identified four ligand- receptor pairs: JAG1-NOTCH2, INHBA-BAMBI, IGF2-IGF2R, SLIT3-ROBO1 (Fig S4.6E). We also identified three ligand-receptor pairs including TIMP2-ITGA3, IGF1-IGF1R and TGFB1- TGFB1R that CSF1R-low C9ORF72 ALS/FTD microglia signals C9ORF72 ALS/FTD neurons to inhibit neurodegeneration (Fig S4.6E). To further validate these ligand-receptor pairs, we blocked the signals by adding neutralizing antibodies of each ligand to the C9ORF72 ALS/FTD iMNs- C9ORF72 ALS/FTD microglia co-cultures and monitored the neuronal survival of C9ORF72 ALS/FTD neurons. We found that JAG1 and IGF2 neutralizing antibodies efficiently blocked the rescue effect in these co-cultures, indicating that C9ORF72 ALS/FTD neurons might mediate the microglial shift via secreting JAG1 and IGF2 (Fig 4.6F, S4.6F). In addition, TIMP2, IGF1 and TGFβ1 neutralizing antibodies had significant effect on blocking the neuroprotection, providing evidence that CSF1R-low C9ORF72 ALS/FTD microglia triggered the neuroprotective effect via secreting these three factors (Fig 4.6G, S4.6G). In conclusion, crosstalk between neuron and 98 microglia in C9ORF72 ALS/FTD provided critical signals to modulate microglial state switch and neuroprotection. Validation of the presence of CSF1R-low microglia in vivo Various mice models have been generated to understand the loss-of-function and gain-of-function effects of C9ORF72 repeat expansion in vivo. For the gain-of-function mechanism, C9orf72 bacterial artificial chromosome (C9BAC) transgenic contains the full length of C9orf72 gene with the expression of both sense and antisense transcripts. C9BAC mice showed survival deficacy, disability in motor functions and displayed pathological features such as accumulation of RNA foci, DPR proteins and TDP inclusions (Liu et al., 2016; O’Rourke et al., 2015). On the other hand, C9orf72 null mice (C9orf72 -/- ) displayed enlarged spleens and lymph nodes, triggered neuroinflammatory by recruitment of macrophages and microglia, while remaining normal survival and motor functions 13 . Moreover, Chen’s group examined the combined effects of loss-and gain- of-function by producing C9BAC mice with loss of one allele of C9orf72 (C9orf72 -/+ ;C9BAC) (Shao et al., 2019). Consistent with their study, we found that deleting one copy of C9orf72 in C9BAC mice consistently accelerated the onset of motor deficits in the hanging wire test by 9 months of age whereas C9BAC mice normally did not show phenotypes at this age (Fig S4.7A). We also confirmed the prominent accumulation of DPRs in the TUJ1+ neurons in the cortex of C9orf72 - /+ ;C9BAC mice (Fig S4.7B). Therefore, both loss-of-function and gain-of-function mechanisms work together to cause disease progression in C9ORF72 ALS in vivo. Consistent with the notion emerging from our in vitro data that CSF1R-high C9ORF72 ALS/FTD microglia were neurotoxic, we utilized PLX3397, a brain-penetrant small molecule CSF1R inhibitor that can be formulated in the chow and administered for an extended duration, to deplete CSF1R-high microglia in 9-month old C9orf72 -/+ ;C9BAC (Fig 74.SC) (Sosna et al., 2018). After 35 days of treatment, C9orf72 -/+ ;C9BAC with PLX3397 significantly improved the motor function in hanging wire test, showed lower levels of poly(GP) accumulation in the TUJ1+ neurons in the 99 cortex, and increased the number of surviving neurons in the lumbar spinal cord, comparing to control chow treated mice (Fig 4.7A-C). We also observed that the number of IBA1+ microglia in the brain reduced in a time-dependent manner after PLX3397 treatment, however, a subset of microglia was resistant to CSF1R inhibitor in the end of 35 days of treatment, indicating the presence of CSF1R-low microglia and their protective roles (Fig 4.7SD-E). Next, we tested if transplantation of CSF1R-low C9ORF72 ALS/FTD microglia into the cortex of 10-month old C9orf72 -/+ ;C9BAC exerted neuroprotective effects. We FACS-purified CSF1R-low or CSF1R-high iPSC-derived C9ORF72 ALS/FTD microglia and performed intracranial injections of microglia (40,000 cells per mice) into one lateral of the cortex in the C9orf72 -/+ ;C9BAC mice (Hasselmann et al., 2019). We left one lateral of the cortex uninjected as to compare the effect of injected regions of the same mice (Fig 4.7D). After 3 days of injection, immunostaining with human specific STEM121 antibody showed a robust engraftment of human microglia surrounding the injected site in the cortex. Since the engrafted microglia did not spread throughout the brain, we did not observe the improvement of rotor rod motor function or neuronal number of lumbar spinal cord in the C9orf72 -/+ ;C9BAC mice compared to the CSF1R-high microglia injected mice (Fig S4.7G-H). Intriguingly, we detected a dramatically reduction of poly(GP) in TUJ1+ neurons in the cortex of C9orf72 -/+ ;C9BAC near the injected site of CSF1R-low C9ORF72 ALS/FTD microglia, compared to the noninjected site in the same mice or CSF1R-high C9ORF72 ALS/FTD microglia injected mice (Fig 4.7E, S4.7I-J). Moreover, we also performed the same engraftment into the C9BAC mice and found similar results with reduction of poly(GP) in cortex neurons in the injected sites and no effect on motor function (Fig 4.7E, S4.7I-J). In summary, we concluded that CSF1R- low C9ORF72 ALS/FTD microglia profoundly mitigated the accumulation of DPRs in the neurons in the C9BAC mice with or without deleting C9orf72 gene. CSF1R ASO as therapeutics in vivo 100 Because inhibiting CSF1R levels in C9ORF72 ALS/FTD microglia drives the transition to CSF1R- low state in vitro, we developed antisense oligonucleotides (ASOs) targeting mice Csf1r and tested the efficacy of Csf1r ASOs in C9ORF72 mice model. To mimic both loss- and gain-of- function mechanisms of C9ORF72 pathology, C9orf72 -/+ mice were transduced intracerebroventricularly at P1 with adeno-associated virus (AAV)-GFP or GFP-(GR) 100 to overexpress 100 repeats of poly(GR) (Fig 4.8A) (Zhang et al., 2018b). As previously reported, we found that AAV-GFP-(GR) 100 mice displayed poly(GR) aggregates in the cortex, severe neuronal loss in lumbar spinal cord, and developed hindlimb clasping phenotypes at about 3 weeks of age; while (AAV)-GFP mice remain intact (Zhang et al., 2018b). To study the efficacy of Csf1r ASOs, we intracerebroventricularly administered negative control (NC) or Csf1r ASOs at P4 to (AAV)- GFP or GFP-(GR) 100 mice, which lead to ~50% of CSF1R mRNA and protein levels (Fig 4.8A-C). We monitored the hindlimb clasping test starting from P16 to P28 and performed rotor rod test at P28. Csf1r ASOs significantly reduced the hindlimb severity in C9orf72 -/+ ;GFP-(GR) 100 mice compared to NC ASO and improved the latency to fall in the rotor rod test (Fig 4.8D-E). In addition, AAV-GFP-(GR)100 mice had significantly fewer motor neurons than AAV-GFP mice in the lumbar spinal cord, and Csf1r ASO treatment greatly reduced the loss of motor neurons in AAV-GFP- (GR) 100 mice (Fig 4.8F). Based on our results, Csf1r ASOs can ameliorate neurodegeneration and motor deficits caused by disease process in C9ORF72 mice model. Suppression of Csf1r had the potential to convert the microglia into neuroprotective CSF1R-low state in C9ORF72 ALS/FTD and ASOs targeting CSF1R could serve as an effective therapeutic strategy. 4.4 Discussion Accumulating evidence has emerged that neuroinflammation is tightly linked with the progression and outcome of a variety of neurodegenerative disorders. In the CNS, microglia are the main type of immune cells and interact closely with neurons to support their function and survival. In the 101 past decade, tremendous studies have been dedicated to understanding the role of microglia in neurodegeneration. It is challenging to study microglia in the disease context due to the limited access of primary microglia from patients and the influence of the in vitro culture environment on microglia transcriptome signatures (Gosselin et al., 2017). A majority of protocols generate patient-derived microglia by mimicking embryonic development and by driving cells to microglial lineage in chemically defined media (Muffat et al., 2016; Abud et al., 2017), however these approaches require more than 30 days to harvest mature and functional microglia. To overcome these obstacles and inspired by our established way to generate induced motor neurons, we sought to generate patient-derived microglia by forced expression of microglia specific transcription factors in iPSCs. This conversion strategy preserves all genetic information in the induced cells and the reprogramming acts on cells intrinsically, not relying on culture media that act extrinsically (Son et al., 2011; Shi et al., 2018). Recently, other groups have performed screening on microglia transcription factors and developed reprogramming approaches to generate functional human microglia (Dräger et al., 2022; Chen et al., 2021; Liu et al.). In our screening, we identified Pu.1 and CEBPA as the most effective combination to generate functional microglia from human iPSCs (Fig 4.1A-B). Pu.1, also known as SPI1, regulates microglial identity through interacting with other core factors and has been utilized in human cortical brain organoids to generate functional microglia (Gosselin et al., 2017; Cakir et al., 2022). CEBPA binds to Pu.1 in the distal enhancer region and acts as a cofactor to drive the monocyte lineage (Yeamans et al., 2007). Previous studies have shown that overexpression of Pu.1 and CEBPA converted skin- derived fibroblasts into macrophage-like cells with phagocytic activity (Feng et al., 2008). In consistent with previous studies, Chen et al. also established a microglia conversion protocol with the utilization of Pu.1 and CEBPA in an independent screen, and Aubert et al. analyzed transcriptomic data from different microglial conversion protocols and revealed that Pu.1 and 102 CEBPA were the core regulators in the network of microglial transcription factors (Chen et al., 2021; Aubert et al., 2021). The development of single-cell or -nuclei RNA sequencing speeded up the discovery of disease- associated microglia (DAM) (Song and Colonna, 2018), which have been identified in multiple disease mice models including Aβ-transgenic (5XFAD) (Krasemann et al., 2017; Liu et al., 2016; Keren-Shaul et al., 2017) and tauopathy (P301S) (Friedman et al., 2018; Leyns et al., 2017) mice models for Alzheimer’s disease and SOD1 G93A mutant mice for ALS (Chiu et al., 2013; Liu et al., 2016). The DAM present in different neuropathologies share similar transcriptomic profiles, featured by elevated type I and type II interferon signalings and regulated by TREM2 (triggering receptor expressed on myeloid cells 2)-APOE (apolipoprotein E) dependent pathways (Chiu et al., 2013; Keren-Shaul et al., 2017; Leyns et al., 2017). Another common feature across various types of neurodegenerative diseases is that DAM are triggered by neurodegeneration-associated molecular patterns (NAMPs) such as apoptotic cells, protein aggregates and myelin debris. The homeostatic microglia express a set of receptors that can sense the signals of NAMPs from degenerating neurons and respond to stimulation via converting to DAM state (Krasemann et al., 2017; Wlodarczyk et al., 2017). This raised our interest to investigate whether there is a unique population of microglia in C9ORF72 ALS/FTD patients that could be triggered by the degenerating neurons or other pathologies related to the long repeat expansion. Indeed, we uncovered a novel subset of C9ORF72 ALS/FTD microglia (CSF1R-low state) that only appeared when co-culturing with degenerating C9ORF72 ALS/FTD motor neurons. This CSF1R-low C9ORF72 ALS/FTD microglia have distinct transcriptomic signatures from DAM and homeostatic microglia. We reasoned that this unique state of microglia is highly related to the pathologies resulting from the C9ORF72 long repeat expansions in both microglia and motor neurons. To this end, we found that overexpression dipeptide repeat proteins in the motor neurons triggered the conversion of microglial switch and C9ORF72-deficient microglia possess the neuroprotective ability (Fig 4.3C, 103 4.5C). Therefore, we proposed that the gain-of-function mechanism in the motor neurons induce the conversion of neuroprotective microglia; while loss-of-function mechanisms in microglia are sufficient to exert the neuroprotective effects. However, more studies need to be done to identify the detailed mechanism of the crosstalk between motor neurons and microglia in this scenario and to understand the difference between DAM and CSF1R-low C9ORF72 ALS/FTD microglia. CSF1R (colony stimulating factor 1 receptor), a receptor tyrosine kinase, is highly expressed in microglia and is potent for the development, proliferation and maintenance of microglia (Chitu et al., 2016). CSF1R is regulated by two cognate ligands, CSF1 (colony stimulating factor 1) and IL- 34 (Interleukin-34), and CSF1R signaling leads to the activation of downstream MEK-ERK and PI3K-AKT cascades (Chitu et al., 2016). CSF1 and IL-34 distributed differently in the brain with CSF1 controlling the white matter microglia and Il-34 responsible for gray matter microglia (Chitu et al., 2016; Easley-Neal et al., 2019). Studies have shown that TREM2-DAP12 (DNAX-activating protein of 12 kDa) signaling also interacts with the CSF1-CSF1R axis and controls microglial survival, supported by the evidence that DAP2 deficiency leads to impaired CSF1R-induced proliferation in microglia (Otero et al., 2009; Konishi and Kiyama, 2018). In the aspect of neurodegenerative diseases, CSF1R signaling is altered and modulates microglial proliferation. ALSP, a dementia called adult-onset leukoencephalopathy with axonal spheroids and pigmented glia, is featured by CSF1R deficiency and CSF2 overproduction, which balancing CSF1R-CSF2 signaling could serve as therapeutic targets (Chitu et al., 2020). In the multiple sclerosis (MS), CSF1R signaling was upregulated and led to excessive microglia activation, and suppression of CSF1R with small molecule inhibitor ameliorate neuroinflammation, reduced proliferation of microglia and prohibited neurodegeneration (Hagan et al., 2020). Additionally, in the model of Alzheimer’s disease (AD), DAM is converted from proliferative microglia and entered the senescence stage of the cell cycle to cause pathology. Inhibiting CSF1R could effectively prevent microglia from senescence and revert the amyloid pathology (Mathys et al., 2017; Hu et al., 2021). 104 In this study, we found the expression levels of CSF1R in the microglia correlated with their effect on neurodegeneration in C9ORF72 ALS/FTD, with lower CSF1R expression in microglia possessing neuroprotective effects (Fig 4.4B). C9ORF72 ALS/FTD microglia highly expressed CSF1R and induced neurodegeneration when culturing with control motor neurons; however, with the presence of C9ORF72 ALS/FTD motor neurons, C9ORF72 ALS/FTD microglia converted to a new state characterized by low CSF1R expression and the role of neuroprotection. We reported the importance of the novel CSF1R-low microglia population in C9ORF72 ALS/FTD and showed inhibition CSF1R signaling improved the neuronal survival and pathologies both in vitro and in vivo. It has been shown that long-term treatment of PLX3397, a CSF1R inhibitor, eliminated the microglia widely in the CNS, however, after removal of PLX3397, microglia derived from nestin+CD45+KI67+ cells repopulated in the brain rapidly (Elmore et al., 2014). We noticed that CSF1R-low C9ORF72 ALS/FTD microglia did not express high levels of Ki67 (data not shown), suggesting that they did not arise from proliferation of rare CSF1R-low microglia and supporting the rational that they converted from CSF1R-high states. Another evidence is that pseudotime analysis of our single-cell RNA-seq data suggested that C9ORF72 ALS/FTD microglia started in the CSF1R-high state, but some microglia convert to CSF1R-low state over the time of co- culturing with C9ORF72 ALS/FTD motor neurons. To fully understand how to convert C9ORF72 ALS/FTD microglia into CSF1R-low state, we identified PBX1, MEF2C and MAF as core transcription factors in CSF1R-low C9ORF72 ALS/FTD microglia. PBX1 activates IL-10 secretion in microglia, which our data show is increased in cultures with neuroprotective iMG and rescues C9ORF72 iMN survival, and MEF2C and MAF promote homeostatic programs in microglia through suppressing inflammatory responses (Chung et al., 2007; Deczkowska et al., 2017; Hunt et al., 2012). We suggested that promoting the expression of core transcription factors (PBX1/MEF2C/MAF) in CSF1R-high C9ORF72 ALS/FTD 105 microglia can induce the phenotypic switch to CSF1R-low states. In addition, we directly injected iPSCs-derived CSF1R-low C9ORF72 ALS/FTD microglia into the cortex of C9ORF72 mice model and found a significant reduction of DPRs in the neurons near the injected site after a short term of CSF1R-low C9ORF72 ALS/FTD microglia engraftment (Fig 4.7D-E). Since we did not utilize immune deficient mice in this experiment, it is hard to monitor the effect of microglia engraftment over a longer period of time without triggering immune response in the mice. The effect of CSF1R- low C9ORF72 ALS/FTD microglia was also limited to the injected regions since microglia did not have enough time to migrate and spread to different regions in the brain. Further studies using a better mice model would be required to investigate the widely-spread and long-term effects of CSF1R-low C9ORF72 ALS/FTD microglia in the C9ORF72 mice model. To develop therapeutic benefits and clinical applications, we sought to keep C9ORF72 ALS/FTD microglia in the CSF1R-low protective state by suppressing CSF1R expression with small molecule inhibitors. Currently, there are several existing CSF1R inhibitors, including PLX3397, PLX5622, GW2580 and JNJ-527. PLX3397, also known as Pexidartinibalso, has been approved by the FDA to treat tenosynovial giant cell tumor (TGCT) (Palmerini et al., 2020). PLX3397 could deplete over 95% of microglia in the wide-type adult mice brain without affecting the behavior functions (Elmore et al., 2014). We showed that after long-term treatment of PLX3397, C9orf72 - /+ ;C9BAC mice not only improved motor functions, also restrained neuronal loss and DPRs accumulation (Fig 4.7A-C). However, PLX3397 is also potent inhibitors for c-KIT and FLT3 and has been reported to cause hepatotoxicity (Palmerini et al., 2020). PLX5622, structurally modified from PLX3397 to selectively target CSF1R over c-KIT and FLT3, was applied to the 5XFAD Alzheimer’s disease model to eliminate microglia for preventing plaques formation (Spangenberg et al., 2019). GW2580, another selective CSF1R inhibitor, has been used in the SOD1 G93A mice model to reduce neuroinflammation and ameliorate disease progression (Martínez-Muriana et al., 2016). Blockage CSF1R with JNJ-527, brain-penetrant CSF1R inhibitor, efficiently limited microglial proliferation and slowed tau pathology in the P301S mice model (Mancuso et al., 2019). 106 In the C9ORF72 ALS/FTD case, we revealed that transient treatment of iPSCs-derived microglia with modest doses of JNJ-527 successfully elevated the expression of CSF1R-low microglia markers (Fig 4.6A). In the future, we plan to apply the JNJ-527 to C9orf72 -/+ ;C9BAC mice model and investigate the effect on motor function and C9ORF72 pathologies. Another approach to inhibit CSF1R is to use antisense oligonucleotides (ASOs) specifically targeting CSF1R mRNA. Inspired by the application of ASOs technology, we designed ASOs targeting CSF1R in humans or mice and tested the conversion efficiency of CSF1R-low in C9ORF72 ALS/FTD microglia. To this end, we validated that CSF1R-targeting ASOs suppressed CSF1R and shifted C9ORF72 ALS/FTD microglia toward CSF1R-low state as shown by elevated IGF-1 expression (Fig 4.6B) in vitro. For in vivo efficacy studies, mice Csf1r ASOs, which reduced Csf1r expression in the brain when injected intracerebroventricularly, dramatically improved motor disability and prevented neuronal loss in the spinal cord. In summary, we have identified CSF1R as a critical therapeutic target for C9ORF72 ALS/FTD and this could be broadly effective for other dementias or neurodegenerative diseases. 107 4.5 Figures Figure 4.1 Transcription factor-mediated conversion of iPSCs into induced microglia. (A) Production of human induced microglia (iMG). (B) Number of tdTomato + iMG produced by forced expression of transcription factors. Kruskal- Wallis test. Median ± interquartile range. n=18 biological replicates for each condition. (C) Immunostaining of microglial markers, MAFB, IBA1, CD11B and CX3CR1, in tdTomato + iMG. Scale bar = 5 μm. 108 (D) Unsupervised clustering heatmap of consensus microglial gene expression from RNA-seq analysis of iPSCs, human fibroblasts, iMG and primary human microglia. n=2 biological replicates per cell type. iMG were FACS-purified by CD11B antibody. (E)Percentage of iMG showing pHrodo bead phagocytosis and fluorescence after 2 and 24 hours of treatment. One-way ANOVA. Mean ± s.e.m. n=2 biological replicates for each condition. (F) Concentration of IL-1B, IL-6 and IL-18 secreted from iMG or iPCSc upon stimulation of Lipopolysaccharides (LPS). One-way ANOVA. Mean ± s.e.m. n=2 biological replicates for each condition. (G) Engraftment of tdTomato + iMG into an iPSC-cortical organoid for 1 and 7 days. 30,000 cells of iMG per organoid. Scale bar = 250 μm. 109 Figure 4.2 C9ORF72 ALS/FTD microglia display different transcriptomic profiles and alter endolysosome phenotypes. (A) Principal component analysis (PCA) of control iMG (green), C9ORF72 ALS/FTD iMG (red), C9ORF72 -/- iMG (blue) and C9ORF72 isogenic iMG (purple). PC1 (32% variance) reflects the difference between C9ORF72 ALS/FTD iMG or C9ORF72 isogenic iMG and control iMG. PC2 110 (16% variance) reflects the influence of long repeat expansions in C9ORF72 ALS/FTD. n=2 biological replicates per line. (B) Unsupervised clustering heatmap of bulk RNA-seq data from iMG from 2 control and 2 C9ORF72 ALS/FTD. n=2-3 biological replicates per line. iMG were FACS-purified by CD11B antibody. (C) Normalized gene counts of MEF2C, P2RY12, TGFBR1 and APOE from control and C9ORF72 ALS/FTD iMG. Mann-Whitney test. Median ± interquartile range. n=2-3 biological replicates per group. (D) Gene ontology of significantly up-regulated genes in C9ORF72 ALS/FTD iMG over control iMG. (E) Representative image and quantification of EEA1+ endosomal size in IBA1+ control and C9ORF72 ALS/FTD iMG. n=52 control and n=59 C9ORF72 ALS/FTD iMG from two control and two C9ORF72 ALS/FTD patient lines. Each gray circle represents a single iMG. Mann-Whitney test. Median ± interquartile range. scale bar = 10 μm. (F) Representative image and quantification of LAMP2+ lysosomal clusters. (F) EEA1+ endosomal size in IBA1+ control and C9ORF72 ALS/FTD iMG. n=3 control and n=3 C9ORF72 ALS/FTD patient lines. Each gray circle represents the ratio in one line of iMG. Unpaired t-test. Mean ± s.e.m. scale bar = 10 μm. (G) Percentage of control and C9ORF72 ALS/FTD iMG showing pHrodo bead phagocytosis and fluorescence after 24, 48 and 72 hours of treatment. Two-way ANOVA. Mean ± s.e.m. n=3 control and n=2 C9ORF72 ALS/FTD patient lines. 111 Figure 4.3 C9ORF72 ALS/FTD microglia exert neuroprotection on C9ORF72 ALS/FTD neurons. (A) Schematic of production and co-culture of induced microglia (iMG) and induced motor neurons (iMN) from control and C9ORF72 ALS/FTD patients. (B) The hazard ratio (Mantel–Haenszel method) of control and C9ORF72 ALS/FTD (C9-ALS/FTD) iMNs with no iMG or co-cultured with control and C9-ALS/FTD iMG. The hazard ratio = the hazard 112 rate of each group relative to the hazard rate of CTRL iMNs with no iMG, which was set as 1 (red dotted line). n=2 lines/genotype for iMNs and iMG. Mean± s.e.m. n=3 independent conversions/line/condition. Statistical significance (one-way ANOVA) was calculated comparing control iMNs with no iMG to control iMNs+control iMG, control iMNs+C9-ALS/FTD iMG and C9- ALS/FTD iMNs with no iMG, as well as C9-ALS/FTD iMNs with no iMG to C9-ALS/FTD iMNs+control iMG, C9-ALS/FTD iMNs+C9-ALS/FTD iMG. (C)The hazard ratio (Mantel–Haenszel method) of control and C9ORF72 ALS/FTD (C9-ALS/FTD) iMNs with no iMG or co-cultured with C9ORF72 -/- (C9 -/- ) iMG. The hazard ratio = the hazard rate of each group relative to the hazard rate of CTRL iMNs with no iMG, which was set as 1 (red dotted line). n=2 lines/genotype for iMNs and n=1 line/genotype for iMG. Mean± s.e.m. n=3 independent conversions/line/condition. Statistical significance (one-way ANOVA) was calculated comparing control iMNs with no iMG to control iMNs+C9 -/- iMG, control iMNs+C9-ALS/FTD iMG and C9-ALS/FTD iMNs with no iMG to C9-ALS/FTD iMNs+C9 -/- iMG. (D) Immunostaining of endogenous poly(GR)+ punctae in C9-ALS/FTD iMNs with no iMG or co- cultured with control and C9-ALS/FTD iMG. Solid and dotted lines outline the cell body and nucleus, respectively. Scale bar = 5 μm. (E) Quantification of (D) to determine endogenous poly(GR)+ punctae in C9-ALS/FTD iMNs with no iMG or co-cultured with control and C9-ALS/FTD iMG. Quantified values represent the average number of nuclear poly(GR)+ punctae per μm 2 in n=34 C9-ALS/FTD iMNs with no iMG, n=60 C9- ALS/FTD iMNs+CTRL iMG and n=68 C9-ALS/FTD iMNs+ C9-ALS/FTD iMG. n=2 lines/genotype for iMNs and n=2 line/genotype for iMG. (n=15-30 iMNs per line). iMNs were quantified from two independent conversions per group. Each gray circle represents the average number in a single iMN. Kruskal-Wallis test. Median ± interquartile range. (F) Cytokine profiling of four co-cultures conditions.Mean of three biological replicates. Log-rank test. Candidate neurotoxic cytokines in red and neuroprotective cytokines in blue. (G) Kaplan–Meier survival curves of control (CTRL) iMNs treated with vehicle or 10 ng/mL GRO- α or IL-1B during the entire time course. n=600 iMNs for vehicle, n=500 for GRO-α and IL-1B treatment from two CTRL lines. iMNs quantified from three biologically independent iMN conversions per line. Two-sided log-rank test. Statistical significance calculated using the entire survival course (days 1-18). (H) Kaplan–Meier survival curves of C9ORF72 ALS/FTD (C9-ALS/FTD) iMNs treated with vehicle or 10 ng/mL IL-10 or CX3CL1 during the entire time course. n=400 iMNs for vehicle, n=300 for IL-10 and CX3CL1 treatment from two C9-ALS/FTD lines. iMNs quantified from three biologically independent iMN conversions per line. Two-sided log-rank test. Statistical significance calculated using the entire survival course (days 1-18). 113 Figure 4.4 A novel CSF1R-low neuroprotective state of C9ORF72 ALS/FTD microglia (A) UMAP plot depicting single-cell RNA-seq data for co-cultures of control (CTRL) iMNs with C9ORF72 ALS/FTD (C9) iMG. Cluster 0-7 are neuron populations and cluster 8 (black dotted line) is AIF1+ microglia population. 114 (B) Violin plots of AIF1 and CSF1R in all the clusters of (A). (C) UMAP plot depicting single-cell RNA-seq data for co-cultures of C9ORF72 ALS/FTD (C9) iMNs with C9ORF72 ALS/FTD (C9) iMG. Cluster 0-5 are neuron populations. Cluster 6 and cluster 7 (black dotted lines) are distinct AIF1+ microglia populations. (D) Violin plots of AIF1 and CSF1R in all the clusters of (C). (E) Violin plots of gene differentially expressed between CSF1R-high iMG (cluster 6 from C9 iMNs+C9iMG) and CSF1R-low iMG (C9 iMNs+CTRL iMG and cluster 7 from C9 iMNs+C9iMG). (F) Enriched pathways of CSF1R-high iMG (cluster 6 from C9 iMNs+C9iMG) and CSF1R-low iMG (cluster 7 from C9 iMNs+C9iMG). (G) Pseudotime analysis of single-cell RNA-seq data from C9 iMNs+C9iMG cluster 6 and cluster 7 (CSF1R-high and CSF1R-low iMG). (H) The hazard ratio (Mantel–Haenszel method) of control and C9ORF72 ALS/FTD (C9) iMNs with no iMG or co-cultured with FACS-purified CSF1R-high or CSF1R-low C9-ALS/FTD iMG. The hazard ratio = the hazard rate of each group relative to the hazard rate of CTRL iMNs with no iMG, which was set as 1 (red dotted line). n=2 lines/genotype for iMNs and iMG. Mean± s.e.m. n=3 independent conversions/line/condition. Statistical significance (one-way ANOVA) was calculated comparing control iMNs with no iMG to control iMNs+CSF1R-low C9 iMG and C9 iMNs with no iMG, as well as C9 iMNs with no iMG to C9 iMNs+CSF1R-high C9 iMG, C9 iMNs+CSF1R- low C9 iMG. 115 Figure 4.5 Gain-of-function mechanism of C9ORF72 ALS/FTD neurons drives the conversion of C9ORF72 ALS/FTD microglia to neuroprotective state (A) Schematic of addition conditioned media from C9ORF72 ALS/FTD patient (C9) iMNs to the control (CTRL) iMNs-C9 iMG co-cultures. (B) The hazard ratio (Mantel–Haenszel method) of control iMNs with no iMG or co-cultured with C9ORF72 ALS/FTD (C9) MG with or without conditioned media (CM) from C9 iMNs. The hazard ratio = the hazard rate of each group relative to the hazard rate of CTRL iMNs with no iMG, which was set as 1 (red dotted line). n=2 lines/genotype for iMNs and iMG. Mean± s.e.m. n=3 independent conversions/line/condition. Statistical significance (one-way ANOVA) was calculated comparing control iMNs with no iMG to control iMNs+ C9 iMG with or without conditioned media (CM). (C) The hazard ratio (Mantel–Haenszel method) of control (CTRL) iMNs with no iMG and C9ORF72 -/- (C9-KO) iMNs with no iMG or co-cultured with CTRL or C9ORF72 ALS/FTD (C9) iMG. The hazard ratio = the hazard rate of each group relative to the hazard rate of CTRL iMNs with no iMG, which was set as 1 (red dotted line). n=2 lines for control iMNs, n=1 line for C9-KO iMNs, and n=2 line/genotype for iMG. Mean± s.e.m. n=3 independent conversions/line/condition. Statistical significance (one-way ANOVA) was calculated comparing control iMNs with no iMG to C9-KO iMNs with no iMG, as well as C9-KO iMNs with no iMG to C9-KO iMNs+CTRL or C9 iMG. (D) The hazard ratio (Mantel–Haenszel method) of control (CTRL) iMNs expression GFP with no iMG and CTRL iMNs expressing GFP-GR50 with no iMG or co-cultured with CTRL or C9ORF72 ALS/FTD (C9) iMG. The hazard ratio = the hazard rate of each group relative to the hazard rate of CTRL-GFP iMNs with no iMG, which was set as 1 (red dotted line). n=2 lines/genotype for iMNs and iMG. Mean± s.e.m. n=3 independent conversions/line/condition. Statistical significance (one-way ANOVA) was calculated comparing CTRL-GFP iMNs with no iMG to CTRL-GFP-GR 50 116 iMNs with no iMG, as well as CTRL-GFP-GR50 iMNs with no iMG to CTRL-GFP-GR 50 iMNs+CTRL or C9 iMG. Figure 4.6 C9ORF72 ALS/FTD microglia shift from CSF1R-high to CSF1R-low neuroprotective state in vitro. 117 (A) mRNA levels (relative to GAPDH) of CSF1R, IGF1 and TGFB1 in control (CTRL) or C9ORF72 ALS/FTD (C9-ALS) iMG treated with DMSO or 1 μM JNJ-527 (JNJ) for 24 hours. n=6 biological replicates (independent conversions) per condition for CTRL and C9-ALS/FTD from two CTRL and C9-ALS/FTD patients. Ordinary one-way ANOVA. Mean ± s.e.m. iMG were FACS-purified by CD11B antibody. (B) mRNA levels (relative to GAPDH) of CSF1R and IGF1 in control (CTRL) or C9ORF72 ALS/FTD (C9-ALS) iMG treated with 10 μM of negative control (NC) or CSF1R ASO for 3 days. n=6 biological replicates (independent conversions) per condition for CTRL and C9-ALS/FTD from two CTRL and C9-ALS/FTD patients. Ordinary one-way ANOVA. Mean ± s.e.m. iMG were FACS-purified by CD11B antibody. (C) mRNA levels (relative to GAPDH) of NR1H3, CSF1R, IGF1 and MEF2C in control (CTRL) or C9ORF72 ALS/FTD (C9-ALS) iMG treated with 30 nM of negative control (NC) or NR1H3 siRNA for 3 days. n=4 biological replicates (independent conversions) per condition for CTRL and C9- ALS/FTD from two CTRL and C9-ALS/FTD patients. Ordinary one-way ANOVA. Mean ± s.e.m. iMG were FACS-purified by CD11B antibody. (D) Binary regulon activity matrix of CSF1R-high iMG (cluster 6 from C9 iMNs+C9iMG) and CSF1R-low iMG (cluster 7 from C9 iMNs+C9iMG) after applying SCENIC analysis. Select regulons are labeled. (E) The hazard ratio (Mantel–Haenszel method) of C9ORF72 ALS/FTD (C9-ALS/FTD) iMNs with no iMG or co-cultured with C9-ALS/FTD iMG treated without or with 4 µg/mL of neutralizing antibodies (NuAb) (JAG1, INHBA, IGF2 and SLIT3). The hazard ratio = the hazard rate of each group relative to the hazard rate of CTRL iMNs with no iMG, which was set as 1 (shown as red dotted line). n=1 lines/genotype for iMNs and iMG. Mean± s.e.m. n=3 independent conversions/line/condition. Statistical significance (one-way ANOVA) was calculated comparing control iMNs with no iMG to control iMNs+control iMG, control iMNs+C9-ALS/FTD iMG to each condition. (F) The hazard ratio (Mantel–Haenszel method) of C9ORF72 ALS/FTD (C9-ALS/FTD) iMNs with no iMG or co-cultured with C9-ALS/FTD iMG treated without or with 4 µg/mL of neutralizing antibodies (NuAb) (TIMP2, IGF1 and TGFB1). The hazard ratio = the hazard rate of each group relative to the hazard rate of CTRL iMNs with no iMG, which was set as 1 (shown as red dotted line). n=1 lines/genotype for iMNs and iMG. Mean± s.e.m. n=3 independent conversions/line/condition. Statistical significance (one-way ANOVA) was calculated comparing control iMNs with no iMG to control iMNs+control iMG, control iMNs+C9-ALS/FTD iMG to each condition. 118 Figure 4.7 Validation of the presence of CSF1R-low microglia in vivo. (A) Lanency to fail from hanging wire test for 9-month-old C9orf72 -/+ ;C9BAC mice before treatment (n=14), after control (CTRL) chow treatment (n=6) or PLX3397 (PLX) chow treatment (n=8) for 35 days. Kruskal-Wallis test. Median ± interquartile range. (B) Representative images and quantification of NeuroTrace (Nissl)-positive lateral motor column (LMC) spinal motor neurons in the lumbar spinal cord in 9-month-old C9orf72 -/+ ;C9BAC mice treated with control (CTRL) chow (n=6) or PLX3397 (PLX) chow (n=8) for 35 days. The LMC motor neuron region in the spinal cord ventral horn is marked with a white dotted-line. Scale bar 119 = 100 μm. 4-6 ventral horn sections were quantified and averaged per mouse. Mann-Whitney test. Median ± interquartile range. (C) Representative images and quantification of poly(GP) punctae in TUJ1+ motor neurons in the motor cortex from the brain of 9-month-old C9orf72 -/+ ;C9BAC mice treated with control (CTRL) chow (n=6) or PLX3397 (PLX) chow (n=8) for 35 days. Unpaired t-test. Mean ± s.e.m. Dotted lines outline the cell body. Neuron cell bodies were identified using TUJ1 immunostaining. Scale bar = 10 μm. (D) Schematic of transplantation of FACS-purified CSF1R-high or CSF1R-low C9ORF72 ALS/FTD microglia into the cortex of WT or C9orf72 -/+ ;C9BAC through intracranial injections of microglia into one lateral of the cortex. (E) Representative images and quantification of poly(GP) punctae in TUJ1+ motor neurons in the motor cortex from the brain of 10-month-old WT mice (shown in gray) injected with CSF1R-high (n=6) or CSF1R-low (n=6) control microglia, C9BAC mice (shown in red) injected with CSF1R- high (n=4) or CSF1R-low (n=6) C9ORF72 ALS/FTD microglia, or C9orf72 -/+ ;C9BAC mice (shown in blue) injected with CSF1R-high (n=5) or CSF1R-low (n=6) C9ORF72 ALS/FTD microglia for 3 days. Ordinary one-way ANOVA. Mean ± s.e.m. Dotted lines outline the cell body. Neuron cell bodies were identified using TUJ1 immunostaining. Scale bar = 10 μm. 120 Figure 4.8 CSF1R ASO as therapeutics in vivo. (A) Schematic of AAV-GFP or AAV-GFP-(GR) 100 model. Intracerebroventricular injection of AAV- GFP or AAV-GFP-(GR) 100 was performed in P1 in C9orf72 -/+ (C9HET) mice. P4 mice received 12.5 μg of NC or Pikfyve ASO by intracerebroventricular injection. The monitoring of hindlimb clasping scores begins at P16 to P28. (B) mRNA levels of Csf1r (relative to beta-actin) in whole brains of C9orf72 -/+ ;AAV-GFP-(GR) 100 mice treated with NC ASO (n=3 mice) or Csf1r ASO (n=3 mice) at P28 after 12.5 μg of ASO injection at P4. Unpaired t-test. Mean± s.e.m. 121 (C) Immunoblots of the whole brain lysates from C9orf72 -/+ ;AAV-GFP-(GR) 100 mice treated with NC ASO (n=3 mice) or Csf1r ASO (n=3 mice) at P28 against CSF1R (180 kDa) and IBA1 (17 kDa). Mice were injected with 12.5 μg of ASO injection at P4. (D) Representative images and quantification of hindlimb clasping scores of C9orf72 -/+ ;AAV-GFP mice or C9orf72 -/+ ;AAV-GFP-(GR) 100 mice treated with negative control (NC) ASO or Csf1r ASO. n=16 for C9orf72 -/+ ;AAV-GFP+NC ASO, n=16 for C9orf72 -/+ ;AAV-GFP+Csf1r ASO, n=19 for AAV- GFP-(GR) 100+NC ASO, n=24 for C9orf72 -/+ ;AAV-GFP-(GR) 100+Csf1r ASO. Statistical significance was calculated by comparing AAV-GFP-(GR)100+NC ASO to AAV-GFP-(GR)100+Csf1r ASO at P21 to P28. Unpaired t-test. Mean ± s.e.m. (E) Latency to fail (seconds) in the rotor rod test of C9orf72 -/+ ;AAV-GFP mice or C9orf72 -/+ ;AAV- GFP-(GR) 100 mice treated with negative control (NC) ASO or Csf1r ASO. n=16 for C9orf72 -/+ ;AAV- GFP+NC ASO, n=16 for C9orf72 -/+ ;AAV-GFP+Csf1r ASO, n=19 for AAV-GFP-(GR) 100+NC ASO, n=24 for C9orf72 -/+ ;AAV-GFP-(GR)100+Csf1r ASO. Ordinary one-way ANOVA. Mean ± s.e.m. (F) Representative images and quantification of NeuroTrace (Nissl)-positive lateral motor column (LMC) spinal motor neurons in the lumbar spinal cord in C9orf72 -/+ ;AAV-GFP mice or C9orf72 - /+ ;AAV-GFP-(GR) 100 mice treated with negative control (NC) ASO or Csf1r ASO. n=16 for C9orf72 - /+ ;AAV-GFP+NC ASO, n=16 for C9orf72 -/+ ;AAV-GFP+Csf1r ASO, n=19 for AAV-GFP-(GR)100+NC ASO, n=24 for C9orf72 -/+ ;AAV-GFP-(GR) 100+Csf1r ASO. Scale bar = 50 μm. 4-6 ventral horn sections were quantified and averaged per mouse. Ordinary one-way ANOVA. Mean ± s.e.m. 122 Chapter 5 : Conclusions 5.1 Discussion I: Antisense oligonucleotides as therapeutics for ALS The development of antisense oligonucleotides (ASOs) as clinical therapeutics grows rapidly over the past few years due to their advantages in specificity and delivery. ASOs bind to its target mRNA and activates RNase H to cleave the DNA-RNA hybrid and diminishes the amount of target mRNA. The modifications on the sequence of ASOs is important for the efficiency, stability, and binding affinity: phosphorothioate linkage improves nuclease stability and uptake, while adding 2’-O-Methyl (2’-OMe) modification increases binding affinity and nuclease stability (Khvorova and Watts, 2017). The advantage of ASOs is the direct delivery to the CNS through intrathecal administration, which can improve safety when targeting ubiquitously expressed genes. Several studies have developed ASOs therapies for ALS. One example of ASOs in clinical trials is SOD1 ASOs which have been shown to modify disease activities and to extend lifespan of SOD1 G93A mice (McCampbell et al., 2018; Miller et al., 2020). Besides the SOD1 ALS model, ASOs targeting C9ORF72 long repeat expansions have been developed (Tran et al., 2022). ASOs targeting ataxin-2 ameliorate the TDP-43 mediated toxicity and improved the survival of the ALS mice model (Becker et al., 2017). The C9ORF72 ASOs significantly reduced the repeat expansion transcripts and DPRs in the neurons from brain cortex and spinal cord (Tran et al., 2022). Here, we identified PIKFYVE inhibition in the neurons as therapeutic targets for multiple forms of ALS and designed ASOs targeting PIKFYVE both in vitro and in vivo. We found that PIKFYVE ASO significantly improves the survival of induced motor neurons derived from C9ORF72 ALS patients. Moreover, Pikfyve ASOs mitigate the motor impairment and extend the survival of TDP- 43 mice. Thus, PIKFYVE ASOs have the potential to develop clinically as a broadly effective therapeutic approach for ALS. 123 5.2 Discussion II: Astrocytes in ALS In this thesis, we focused on the studying the neuron-microglia interaction in C9ORF72 ALS. In the CNS, astrocytes deliver energy to neurons and react to the stimulations and stress for maintaining brain homeostasis (Sofroniew and Vinters, 2009). Reactive astrogliosis is the hallmark for neurodegeneration and neuronal injuries, suggesting that astrocytes are key regulators of neurodegeneration (Siracusa et al., 2019). In Alzheimer disease, astrocytes participate in clearance of plaques and neurofibrillary tangles and secret pro-inflammatory cytokines, indicating their beneficial roles in disease progression; however, astrocytes altered their functions and impaired neuronal support (Garwood et al., 2017). It remains unknow whether astrocytes are protective or toxic toward neurodegeneration, especially in C9ORF72 ALS. Microglia secret TNF, IL-1A and C1q to induce the appearance of a subpopulation of reactive astrocytes in the neurodegenerative diseases, highlighting the importance of microglia-astrocyte interactions in the disease progression (Liddelow et al., 2017). Canals et al. have developed cellular reprogramming strategy to generate induced astrocytes by force expression of transcription factors, SOX9 and NFIB, from human iPSCs (Canals et al., 2018). Combining the patient-derived astrocytes into our neuron-microglia co-culture system will extend our understanding of how glia cells and neurons interact with each other and affect neurodegeneration. 5.3 Discussion III: Microglia in sporadic ALS In the context of C9ORF72 ALS/FTD, our results indicate that motor neurons drive the conversion of microglia into the CSF1R-low states and CSF1R-low microglia exert protective effects on motor neurons, emphasizing the importance of neuron-microglia crosstalk. The results also suggests that the long hexanucleotide repeat expansions in the C9ORF72 ALS/FTD motor neurons is the key regulator for the microglia phenotypic switch. Given the fact that some pathological 124 mechanisms are similar in C9ORF72 and sporadic ALS, it is possible that there is a distinct subpopulation of microglia presence in sporadic ALS and the sporadic ALS motor neurons are responsible for the microglial phenotypic switch. Limited studies have investigated the microglia in the sporadic ALS cases. Microglia have been shown to lose postsynaptic transcripts and acquire inflammatory signatures in the motor cortex of sporadic ALS tissue (Dols-Icardo et al., 2020). Quek et al. demonstrated that monocyte-derived microglia from sporadic ALS patients develop TDP-43 pathology, possess altered microglial functions and activated inflammasome (Quek et al., 2022). We have established a neuron-microglial co-culture method using the induce motor neurons and induced microglia derived from patient-derived iPSCs. This method has enabled us to elucidate the role of microglia in C9ORF72 ALS neurodegeneration. Future work will be required to expand this method to sporadic ALS and determine what is the role of microglia in sporadic ALS and whether there is a similar subset of microglia responsible for the neuroprotective effect. 5.4 Conclusion The complexity of cause and process in neurodegeneration is the main impediment for developing effective therapeutics for ALS. The diverse genetic etiologies and the participation of other non- neuronal cells drive this complexity. In this thesis, we dedicated to elucidating pathological mechanisms and identifying therapeutic targets from the aspect of both neurons and microglia in C9ORF72 ALS/FTD. Impaired protein homeostasis in the neurons is hallmark shared between C9ORF72 ALS/FTD and sporadic ALS. Improving protein homeostasis and clearance of accumulated proteins inside the neurons with small molecules or ASOs serve as a promising therapeutic approach across multiple forms of ALS. We demonstrated that activated protein C (3K3A-APC) promotes the autophagosome pathways in ALS neurons and efficiently mitigates the accumulation of misfolded proteins, leading to rescuing the neuronal survival. In addition, we 125 show that pharmacological inhibition of PIKFYVE activates an unconventional proteostasis mechanism involving exocytosis of aggregation-prone proteins, which significantly ameliorates pathology and neurodegeneration in ALS. Finally, we uncover a novel subset of neuroprotective microglia in C9ORF72 ALS/FTD and utilize ASOs to suppress CSF1R expression in microglia, which promotes phenotypic switch of microglia into a beneficial state. In conclusion, we link the pathological mechanisms with the identification of therapeutic targets, which will greatly improve our ability to develop new therapeutic for C9ORF72 ALS/FTD as well as other neurodegenerative disorders. 126 References Abramzon, Y.A., Fratta, P., Traynor, B.J., and Chia, R. (2020). The Overlapping Genetics of Amyotrophic Lateral Sclerosis and Frontotemporal Dementia. Frontiers in Neuroscience 14. Abud, E.M., Ramirez, R.N., Martinez, E.S., Healy, L.M., Nguyen, C.H.H., Newman, S.A., Yeromin, A.V., Scarfone, V.M., Marsh, S.E., Fimbres, C., et al. (2017). iPSC-Derived Human Microglia-like Cells to Study Neurological Diseases. Neuron 94, 278-293.e9. Aggad, D., Veriepe, J., Tauffenberger, A., and Parker, J.A. (2014). TDP-43 Toxicity Proceeds via Calcium Dysregulation and Necrosis in Aging Caenorhabditis elegans Motor Neurons. Journal of Neuroscience 34, 12093–12103. Aibar, S., González-Blas, C.B., Moerman, T., Huynh-Thu, V.A., Imrichova, H., Hulselmans, G., Rambow, F., Marine, J.-C., Geurts, P., Aerts, J., et al. (2017). SCENIC: single-cell regulatory network inference and clustering. Nature Methods 14, 1083–1086. Alvarez-Erviti, L., Seow, Y., Schapira, A.H., Gardiner, C., Sargent, I.L., Wood, M.J.A., and Cooper, J.M. (2011). Lysosomal dysfunction increases exosome-mediated alpha-synuclein release and transmission. Neurobiology of Disease 42, 360–367. Arthur, K.C., Calvo, A., Price, T.R., Geiger, J.T., Chiò, A., and Traynor, B.J. (2016). Projected increase in amyotrophic lateral sclerosis from 2015 to 2040. Nature Communications 7. Atanasio, A., Decman, V., White, D., Ramos, M., Ikiz, B., Lee, H.-C., Siao, C.-J., Brydges, S., LaRosa, E., Bai, Y., et al. (2016). C9orf72 ablation causes immune dysregulation characterized by leukocyte expansion, autoantibody production and glomerulonephropathy in mice. Scientific Reports 6. Aubert, A., Stüder, F., Colombo, B.M., and Mendoza-Parra, M.A. (2021). A Core Transcription Regulatory Circuitry Defining Microglia Cell Identity Inferred from the Reanalysis of Multiple Human Microglia Differentiation Protocols. Brain Sciences 11, 1338. Balendra, R., and Isaacs, A.M. (2018). C9orf72-mediated ALS and FTD: multiple pathways to disease. Nature Reviews Neurology 14, 544–558. Barmada, S.J., Skibinski, G., Korb, E., Rao, E.J., Wu, J.Y., and Finkbeiner, S. (2010). Cytoplasmic Mislocalization of TDP-43 Is Toxic to Neurons and Enhanced by a Mutation Associated with Familial Amyotrophic Lateral Sclerosis. Journal of Neuroscience 30, 639–649. Barmada, S.J., Serio, A., Arjun, A., Bilican, B., Daub, A., Ando, D.M., Tsvetkov, A., Pleiss, M., Li, X., Peisach, D., et al. (2014). Autophagy induction enhances TDP43 turnover and survival in neuronal ALS models. Nature Chemical Biology 10, 677–685. Becker, L.A., Huang, B., Bieri, G., Ma, R., Knowles, D.A., Jafar-Nejad, P., Messing, J., Kim, H.J., Soriano, A., Auburger, G., et al. (2017). Therapeutic reduction of ataxin-2 extends lifespan and reduces pathology in TDP-43 mice. Nature 544, 367–371. Beers, D.R., and Appel, S.H. (2019). Immune dysregulation in amyotrophic lateral sclerosis: mechanisms and emerging therapies. The Lancet Neurology 18, 211–220. 127 Belbin, O., Dunn, J.L., Ling, Y., Morgan, L., Chappell, S., Beaumont, H., Warden, D., Smith, D.A., Kalsheker, N., and Morgan, K. (2007). Regulatory region single nucleotide polymorphisms of the apolipoprotein E gene and the rate of cognitive decline in Alzheimer’s disease. Human Molecular Genetics 16, 2199–2208. Berning, B.A., and Walker, A.K. (2019). The Pathobiology of TDP-43 C-Terminal Fragments in ALS and FTLD. Frontiers in Neuroscience 13. Boeynaems, S., Bogaert, E., Michiels, E., Gijselinck, I., Sieben, A., Jovičić, A., De Baets, G., Scheveneels, W., Steyaert, J., Cuijt, I., et al. (2016). Drosophila screen connects nuclear transport genes to DPR pathology in c9ALS/FTD. Scientific Reports 6. Brettschneider, J., Toledo, J.B., Van Deerlin, V.M., Elman, L., McCluskey, L., Lee, V.M.-Y. ., and Trojanowski, J.Q. (2012). Microglial Activation Correlates with Disease Progression and Upper Motor Neuron Clinical Symptoms in Amyotrophic Lateral Sclerosis. PLoS ONE 7, e39216. Brouwers, N., Van Cauwenberghe, C., Engelborghs, S., Lambert, J-C., Bettens, K., Le Bastard, N., Pasquier, F., Montoya, A.G., Peeters, K., Mattheijssens, M., et al. (2011). Alzheimer risk associated with a copy number variation in the complement receptor 1 increasing C3b/C4b binding sites. Molecular Psychiatry 17, 223–233. Brown, A.-L., Wilkins, O.G., Keuss, M.J., Hill, S.E., Zanovello, M., Lee, W.C., Bampton, A., Lee, F.C.Y., Masino, L., Qi, Y.A., et al. (2022). TDP-43 loss and ALS-risk SNPs drive mis-splicing and depletion of UNC13A. Nature 603, 131–137. Burberry, A., Suzuki, N., Wang, J.-Y., Moccia, R., Mordes, D.A., Stewart, M.H., Suzuki-Uematsu, S., Ghosh, S., Singh, A., Merkle, F.T., et al. (2016a). Loss-of-function mutations in the   C9ORF72   mouse ortholog cause fatal autoimmune disease. Science Translational Medicine 8, 347ra93–347ra93. Burberry, A., Suzuki, N., Wang, J.-Y., Moccia, R., Mordes, D.A., Stewart, M.H., Suzuki-Uematsu, S., Ghosh, S., Singh, A., Merkle, F.T., et al. (2016b). Loss-of-function mutations in the   C9ORF72   mouse ortholog cause fatal autoimmune disease. Science Translational Medicine 8, 347ra93–347ra93. Burrell, J.R., Kiernan, M.C., Vucic, S., and Hodges, J.R. (2011). Motor Neuron dysfunction in frontotemporal dementia. Brain 134, 2582–2594. Cady, J., Koval, E.D., Benitez, B.A., Zaidman, C., Jockel-Balsarotti, J., Allred, P., Baloh, R.H., Ravits, J., Simpson, E., Appel, S.H., et al. (2014). TREM2 Variant p.R47H as a Risk Factor for Sporadic Amyotrophic Lateral Sclerosis. JAMA Neurology 71, 449–453. Cakir, B., Tanaka, Y., Kiral, F.R., Xiang, Y., Dagliyan, O., Wang, J., Lee, M., Greaney, A.M., Yang, W.S., duBoulay, C., et al. (2022). Expression of the transcription factor PU.1 induces the generation of microglia-like cells in human cortical organoids. Nature Communications 13, 430. Canals, I., Ginisty, A., Quist, E., Timmerman, R., Fritze, J., Miskinyte, G., Monni, E., Hansen, M.G., Hidalgo, I., Bryder, D., et al. (2018). Rapid and efficient induction of functional astrocytes from human pluripotent stem cells. Nature Methods 15, 693–696. 128 Chen, H., Kankel, M.W., Su, S.C., Han, S.W.S., and Ofengeim, D. (2018). Exploring the genetics and non-cell autonomous mechanisms underlying ALS/FTLD. Cell Death & Differentiation 25, 648–662. Chen, P., Gu, Z., Liu, W., and Yan, Z. (2007). Glycogen synthase kinase 3 regulates N-methyl- D-aspartate receptor channel trafficking and function in cortical neurons. Molecular Pharmacology 72, 40–51. Chen, S.-W., Hung, Y.-S., Fuh, J.-L., Chen, N.-J., Chu, Y.-S., Chen, S.-C., Fann, M.-J., and Wong, Y.-H. (2021). Efficient conversion of human induced pluripotent stem cells into microglia by defined transcription factors. Stem Cell Reports 16, 1363–1380. Chiò, A., Mazzini, L., and Mora, G. (2020). Disease-modifying therapies in amyotrophic lateral sclerosis. Neuropharmacology 167, 107986. Chitu, V., Gokhan, Ş., Nandi, S., Mehler, M.F., and Stanley, E.R. (2016). Emerging Roles for CSF-1 Receptor and its Ligands in the Nervous System. Trends in Neurosciences 39, 378–393. Chitu, V., Biundo, F., Shlager, G.G.L., Park, E.S., Wang, P., Gulinello, M.E., Gokhan, Ş., Ketchum, H.C., Saha, K., DeTure, M.A., et al. (2020). Microglial Homeostasis Requires Balanced CSF- 1/CSF-2 Receptor Signaling. Cell Reports 30, 3004-3019.e5. Chiu, Isaac M., Morimoto, Emiko T.A., Goodarzi, H., Liao, Jennifer T., O’Keeffe, S., Phatnani, Hemali P., Muratet, M., Carroll, Michael C., Levy, S., Tavazoie, S., et al. (2013). A Neurodegeneration-Specific Gene-Expression Signature of Acutely Isolated Microglia from an Amyotrophic Lateral Sclerosis Mouse Model. Cell Reports 4, 385–401. Choi, S.Y., Lopez-Gonzalez, R., Krishnan, G., Phillips, H.L., Li, A.N., Seeley, W.W., Yao, W.-D., Almeida, S., and Gao, F.-B. (2019). C9ORF72-ALS/FTD-associated poly(GR) binds Atp5a1 and compromises mitochondrial function in vivo. Nature Neuroscience 22, 851–862. Chung, E.Y., Liu, J., Homma, Y., Zhang, Y., Brendolan, A., Saggese, M., Han, J., Silverstein, R., Selleri, L., and Ma, X. (2007). Interleukin-10 Expression in Macrophages during Phagocytosis of Apoptotic Cells Is Mediated by Homeodomain Proteins Pbx1 and Prep-1. Immunity 27, 952–964. Ciura, S., Lattante, S., Le Ber, I., Latouche, M., Tostivint, H., Brice, A., and Kabashi, E. (2013). Loss of function of C9orf72 causes motor deficits in a zebrafish model of Amyotrophic Lateral Sclerosis. Annals of Neurology 74, 180–187. Clarke, B.E., and Patani, R. (2020). The microglial component of amyotrophic lateral sclerosis. Brain 143, 3526–3539. Cleary, J.D., and Ranum, L.P.W. (2013). Repeat-associated non-ATG (RAN) translation in neurological disease. Human Molecular Genetics 22, R45–R51. Cleveland, D.W., and Rothstein, J.D. (2001). From charcot to lou gehrig: deciphering selective motor neuron death in als. Nature Reviews Neuroscience 2, 806–819. Cohen, T.J., Lee, V.M.Y., and Trojanowski, J.Q. (2011). TDP-43 functions and pathogenic mechanisms implicated in TDP-43 proteinopathies. Trends in Molecular Medicine 17, 659–667. 129 Conlon, E.G., Lu, L., Sharma, A., Yamazaki, T., Tang, T., Shneider, N.A., and Manley, J.L. (2016). The C9ORF72 GGGGCC expansion forms RNA G-quadruplex inclusions and sequesters hnRNP H to disrupt splicing in ALS brains. ELife 5. Cooper-Knock, J., Bury, J.J., Heath, P.R., Wyles, M., Higginbottom, A., Gelsthorpe, C., Highley, J.R., Hautbergue, G., Rattray, M., Kirby, J., et al. (2015). C9ORF72 GGGGCC Expanded Repeats Produce Splicing Dysregulation which Correlates with Disease Severity in Amyotrophic Lateral Sclerosis. PLOS ONE 10, e0127376. Corcia, P., Tauber, C., Vercoullie, J., Arlicot, N., Prunier, C., Praline, J., Nicolas, G., Venel, Y., Hommet, C., Baulieu, J.-L., et al. (2012). Molecular Imaging of Microglial Activation in Amyotrophic Lateral Sclerosis. PLoS ONE 7, e52941. Coyne, A.N., Siddegowda, B.B., Estes, P.S., Johannesmeyer, J., Kovalik, T., Daniel, S.G., Pearson, A., Bowser, R., and Zarnescu, D.C. (2014). Futsch/MAP1B mRNA Is a Translational Target of TDP-43 and Is Neuroprotective in a Drosophila Model of Amyotrophic Lateral Sclerosis. Journal of Neuroscience 34, 15962–15974. de Lartigue, J., Polson, H., Feldman, M., Shokat, K., Tooze, S.A., Urbé, S., and Clague, M.J. (2009). PIKfyve Regulation of Endosome-Linked Pathways. Traffic (Copenhagen, Denmark) 10, 883–893. De Roeck, A., Van Broeckhoven, C., and Sleegers, K. (2019). The role of ABCA7 in Alzheimer’s disease: evidence from genomics, transcriptomics and methylomics. Acta Neuropathologica 138, 201–220. Deczkowska, A., Matcovitch-Natan, O., Tsitsou-Kampeli, A., Ben-Hamo, S., Dvir-Szternfeld, R., Spinrad, A., Singer, O., David, E., Winter, D.R., Smith, L.K., et al. (2017). Mef2C restrains microglial inflammatory response and is lost in brain ageing in an IFN-I-dependent manner. Nature Communications 8, 717. DeJesus-Hernandez, M., Mackenzie, Ian R., Boeve, Bradley F., Boxer, Adam L., Baker, M., Rutherford, Nicola J., Nicholson, Alexandra M., Finch, NiCole A., Flynn, H., Adamson, J., et al. (2011). Expanded GGGGCC Hexanucleotide Repeat in Noncoding Region of C9ORF72 Causes Chromosome 9p-Linked FTD and ALS. Neuron 72, 245–256. Deng, Q., Ramskold, D., Reinius, B., and Sandberg, R. (2014). Single-Cell RNA-Seq Reveals Dynamic, Random Monoallelic Gene Expression in Mammalian Cells. Science 343, 193–196. Devlin, A.-C., Burr, K., Borooah, S., Foster, J.D., Cleary, E.M., Geti, I., Vallier, L., Shaw, C.E., Chandran, S., and Miles, G.B. (2015). Human iPSC-derived motoneurons harbouring TARDBP or C9ORF72 ALS mutations are dysfunctional despite maintaining viability. Nature Communications 6. Di Malta, C., Cinque, L., and Settembre, C. (2019). Transcriptional Regulation of Autophagy: Mechanisms and Diseases. Frontiers in Cell and Developmental Biology 7. Dobin, A., Davis, C.A., Schlesinger, F., Drenkow, J., Zaleski, C., Jha, S., Batut, P., Chaisson, M., and Gingeras, T.R. (2012). STAR: ultrafast universal RNA-seq aligner. Bioinformatics 29, 15–21. 130 Dols-Icardo, O., Montal, V., Sirisi, S., López-Pernas, G., Cervera-Carles, L., Querol-Vilaseca, M., Muñoz, L., Belbin, O., Alcolea, D., Molina-Porcel, L., et al. (2020). Motor cortex transcriptome reveals microglial key events in amyotrophic lateral sclerosis. Neurology - Neuroimmunology Neuroinflammation 7, e829. Donnelly, Christopher J., Zhang, P.-W., Pham, Jacqueline T., Haeusler, Aaron R., Mistry, Nipun A., Vidensky, S., Daley, Elizabeth L., Poth, Erin M., Hoover, B., Fines, Daniel M., et al. (2013). RNA Toxicity from the ALS/FTD C9ORF72 Expansion Is Mitigated by Antisense Intervention. Neuron 80, 415–428. Dräger, N.M., Sattler, S.M., Huang, C.T.-L., Teter, O.M., Leng, K., Hashemi, S.H., Hong, J., Aviles, G., Clelland, C.D., Zhan, L., et al. (2022). A CRISPRi/a platform in human iPSC-derived microglia uncovers regulators of disease states. Nature Neuroscience 25, 1149–1162. Du, Z.-W., Chen, H., Liu, H., Lu, J., Qian, K., Huang, C.-L., Zhong, X., Fan, F., and Zhang, S.-C. (2015). Generation and expansion of highly pure motor neuron progenitors from human pluripotent stem cells. Nature Communications 6. Dupont, N., Jiang, S., Pilli, M., Ornatowski, W., Bhattacharya, D., and Deretic, V. (2011). Autophagy-based unconventional secretory pathway for extracellular delivery of IL-1β. The EMBO Journal 30, 4701–4711. Easley-Neal, C., Foreman, O., Sharma, N., Zarrin, A.A., and Weimer, R.M. (2019). CSF1R Ligands IL-34 and CSF1 Are Differentially Required for Microglia Development and Maintenance in White and Gray Matter Brain Regions. Frontiers in Immunology 10. Ejlerskov, P., Rasmussen, I., Nielsen, T.T., Bergström, A.-L., Tohyama, Y., Jensen, P.H., and Vilhardt, F. (2013). Tubulin Polymerization-promoting Protein (TPPP/p25α) Promotes Unconventional Secretion of α-Synuclein through Exophagy by Impairing Autophagosome- Lysosome Fusion. Journal of Biological Chemistry 288, 17313–17335. Elmore, Monica R.P., Najafi, Allison R., Koike, Maya A., Dagher, Nabil N., Spangenberg, Elizabeth E., Rice, Rachel A., Kitazawa, M., Matusow, B., Nguyen, H., West, Brian L., et al. (2014). Colony-Stimulating Factor 1 Receptor Signaling Is Necessary for Microglia Viability, Unmasking a Microglia Progenitor Cell in the Adult Brain. Neuron 82, 380–397. Estes, P.S., Boehringer, A., Zwick, R., Tang, J.E., Grigsby, B., and Zarnescu, D.C. (2011). Wild- type and A315T mutant TDP-43 exert differential neurotoxicity in a Drosophila model of ALS. Human Molecular Genetics 20, 2308–2321. Estes, P.S., Daniel, S.G., McCallum, A.P., Boehringer, A.V., Sukhina, A.S., Zwick, R.A., and Zarnescu, D.C. (2013). Motor neurons and glia exhibit specific individualized responses to TDP- 43 expression in a Drosophila model of amyotrophic lateral sclerosis. Disease Models & Mechanisms 6, 721–733. Feng, R., Desbordes, S.C., Xie, H., Tillo, E.S., Pixley, F., Stanley, E.R., and Graf, T. (2008). PU.1 and C/EBPα/β convert fibroblasts into macrophage-like cells. Proceedings of the National Academy of Sciences 105, 6057–6062. Ferrari, R., Kapogiannis, D., D. Huey, E., and Momeni, P. (2011). FTD and ALS: A Tale of Two Diseases. Current Alzheimer Research 8, 273–294. 131 Freibaum, B.D., Lu, Y., Lopez-Gonzalez, R., Kim, N.C., Almeida, S., Lee, K.-H., Badders, N., Valentine, M., Miller, B.L., Wong, P.C., et al. (2015). GGGGCC repeat expansion in C9orf72 compromises nucleocytoplasmic transport. Nature 525, 129–133. Friedman, B., Srinivasan, K., Ayalon, G., Hansen, D., van der Brug, M., and Kaminker, J. (2018). Diverse Brain Myeloid Expression Profiles Reveal Distinct Microglial Activation States and Aspects of Alzheimer’s Disease Not Evident in Mouse Models. Cell Reports 22, 832–847. Garden, Gwenn A., and La Spada, Albert R. (2012). Intercellular (Mis)communication in Neurodegenerative Disease. Neuron 73, 886–901. Garwood, C.J., Ratcliffe, L.E., Simpson, J.E., Heath, P.R., Ince, P.G., and Wharton, S.B. (2017). Review: Astrocytes in Alzheimer’s disease and other age-associated dementias: a supporting player with a central role. Neuropathology and Applied Neurobiology 43, 281–298. Gasset-Rosa, F., Lu, S., Yu, H., Chen, C., Melamed, Z., Guo, L., Shorter, J., Da Cruz, S., and Cleveland, D.W. (2019). Cytoplasmic TDP-43 De-mixing Independent of Stress Granules Drives Inhibition of Nuclear Import, Loss of Nuclear TDP-43, and Cell Death. Neuron 102, 339-357.e7. Gee, J.A., Frausto, R.F., Chung, D.-W.D., Tangmonkongvoragul, C., Le, D.J., Wang, C., Han, J., and Aldave, Anthony.J. (2015). Identification of novel PIKFYVE gene mutations associated with Fleck corneal dystrophy. Molecular Vision 21, 1093–1100. Gitler, A.D., and Tsuiji, H. (2016). There has been an awakening: Emerging mechanisms of C9orf72 mutations in FTD/ALS. Brain Research 1647, 19–29. Gopal, P.P., Nirschl, J.J., Klinman, E., and Holzbaur, E.L.F. (2017). Amyotrophic lateral sclerosis- linked mutations increase the viscosity of liquid-like TDP-43 RNP granules in neurons. Proceedings of the National Academy of Sciences 114, E2466–E2475. Gosselin, D., Skola, D., Coufal, N.G., Holtman, I.R., Schlachetzki, J.C.M., Sajti, E., Jaeger, B.N., O’Connor, C., Fitzpatrick, C., Pasillas, M.P., et al. (2017). An environment-dependent transcriptional network specifies human microglia identity. Science 356. Griffin, J.H., Zlokovic, B.V., and Mosnier, L.O. (2018). Activated protein C, protease activated receptor 1, and neuroprotection. Blood 132, 159–169. Grün, D., Lyubimova, A., Kester, L., Wiebrands, K., Basak, O., Sasaki, N., Clevers, H., and van Oudenaarden, A. (2015). Single-cell messenger RNA sequencing reveals rare intestinal cell types. Nature 525, 251–255. Haenseler, W., Sansom, S.N., Buchrieser, J., Newey, S.E., Moore, C.S., Nicholls, F.J., Chintawar, S., Schnell, C., Antel, J.P., Allen, N.D., et al. (2017). A Highly Efficient Human Pluripotent Stem Cell Microglia Model Displays a Neuronal-Co-culture-Specific Expression Profile and Inflammatory Response. Stem Cell Reports 8, 1727–1742. Hagan, N., Kane, J.L., Grover, D., Woodworth, L., Madore, C., Saleh, J., Sancho, J., Liu, J., Li, Y., Proto, J., et al. (2020). CSF1R signaling is a regulator of pathogenesis in progressive MS. Cell Death & Disease 11, 904. 132 Harley, J., Hagemann, C., Serio, A., and Patani, R. (2021). TDP-43 and FUS mislocalization in VCP mutant motor neurons is reversed by pharmacological inhibition of the VCP D2 ATPase domain. Brain Communications 3. Hasselmann, J., Coburn, M.A., England, W., Velez, D.X.F., Shabestari, S.K., Tu, C.H., McQuade, A., Kolahdouzan, M., Echeverria, K., Claes, C., et al. (2019). Development of a Chimeric Model to Study and Manipulate Human Microglia In Vivo. Neuron 103, 1016-1033.e10. Hessvik, N.P., and Llorente, A. (2017). Current knowledge on exosome biogenesis and release. Cellular and Molecular Life Sciences 75, 193–208. Hessvik, N.P., Øverbye, A., Brech, A., Torgersen, M.L., Jakobsen, I.S., Sandvig, K., and Llorente, A. (2016). PIKfyve inhibition increases exosome release and induces secretory autophagy. Cellular and Molecular Life Sciences 73, 4717–4737. Hodge, R.D., Bakken, T.E., Miller, J.A., Smith, K.A., Barkan, E.R., Graybuck, L.T., Close, J.L., Long, B., Johansen, N., Penn, O., et al. (2019). Conserved cell types with divergent features in human versus mouse cortex. Nature 573, 61–68. Hu, Y., Fryatt, G.L., Ghorbani, M., Obst, J., Menassa, D.A., Martin-Estebane, M., Muntslag, T.A.O., Olmos-Alonso, A., Guerrero-Carrasco, M., Thomas, D., et al. (2021). Replicative senescence dictates the emergence of disease-associated microglia and contributes to Aβ pathology. Cell Reports 35, 109228. Hunt, D., Raivich, G., and Anderson, P.N. (2012). Activating Transcription Factor 3 and the Nervous System. Frontiers in Molecular Neuroscience 5. Ichida, J.K., and Kiskinis, E. (2015). Probing disorders of the nervous system using reprogramming approaches. The EMBO Journal 34, 1456–1477. Igaz, L.M., Kwong, L.K., Xu, Y., Truax, A.C., Uryu, K., Neumann, M., Clark, C.M., Elman, L.B., Miller, B.L., Grossman, M., et al. (2008). Enrichment of C-Terminal Fragments in TAR DNA- Binding Protein-43 Cytoplasmic Inclusions in Brain but not in Spinal Cord of Frontotemporal Lobar Degeneration and Amyotrophic Lateral Sclerosis. The American Journal of Pathology 173, 182– 194. Iguchi, Y., Eid, L., Parent, M., Soucy, G., Bareil, C., Riku, Y., Kawai, K., Takagi, S., Yoshida, M., Katsuno, M., et al. (2016). Exosome secretion is a key pathway for clearance of pathological TDP- 43. Brain: A Journal of Neurology 139, 3187–3201. Ikonomov, O.C., Sbrissa, D., Delvecchio, K., Xie, Y., Jin, J.-P., Rappolee, D., and Shisheva, A. (2011). The Phosphoinositide Kinase PIKfyve Is Vital in Early Embryonic Development. Journal of Biological Chemistry 286, 13404–13413. Jay, T.R., Miller, C.M., Cheng, P.J., Graham, L.C., Bemiller, S., Broihier, M.L., Xu, G., Margevicius, D., Karlo, J.C., Sousa, G.L., et al. (2015). TREM2 deficiency eliminates TREM2+ inflammatory macrophages and ameliorates pathology in Alzheimer’s disease mouse models. Journal of Experimental Medicine 212, 287–295. 133 Jonsson, T., Stefansson, H., Steinberg, S., Jonsdottir, I., Jonsson, P.V., Snaedal, J., Bjornsson, S., Huttenlocher, J., Levey, A.I., Lah, J.J., et al. (2013). Variant of TREM2 associated with the risk of Alzheimer’s disease. The New England Journal of Medicine 368, 107–116. Kaushik, S., and Cuervo, A.M. (2015). Proteostasis and aging. Nature Medicine 21, 1406–1415. Keren-Shaul, H., Spinrad, A., Weiner, A., Matcovitch-Natan, O., Dvir-Szternfeld, R., Ulland, T.K., David, E., Baruch, K., Lara-Astaiso, D., Toth, B., et al. (2017a). A Unique Microglia Type Associated with Restricting Development of Alzheimer’s Disease. Cell 169, 1276-1290.e17. Keren-Shaul, H., Spinrad, A., Weiner, A., Matcovitch-Natan, O., Dvir-Szternfeld, R., Ulland, T.K., David, E., Baruch, K., Lara-Astaiso, D., Toth, B., et al. (2017b). A Unique Microglia Type Associated with Restricting Development of Alzheimer’s Disease. Cell 169, 1276-1290.e17. Khvorova, A., and Watts, J.K. (2017a). The chemical evolution of oligonucleotide therapies of clinical utility. Nature Biotechnology 35, 238–248. Khvorova, A., and Watts, J.K. (2017b). The chemical evolution of oligonucleotide therapies of clinical utility. Nature Biotechnology 35, 238–248. Kierdorf, K., Erny, D., Goldmann, T., Sander, V., Schulz, C., Perdiguero, E.G., Wieghofer, P., Heinrich, A., Riemke, P., Hölscher, C., et al. (2013). Microglia emerge from erythromyeloid precursors via Pu.1- and Irf8-dependent pathways. Nature Neuroscience 16, 273–280. Kim, D., Langmead, B., and Salzberg, S.L. (2015). HISAT: a fast spliced aligner with low memory requirements. Nature Methods 12, 357–360. Kimura, S., Noda, T., and Yoshimori, T. (2007). Dissection of the Autophagosome Maturation Process by a Novel Reporter Protein, Tandem Fluorescent-Tagged LC3. Autophagy 3, 452–460. Kiskinis, E., Sandoe, J., Williams, Luis A., Boulting, Gabriella L., Moccia, R., Wainger, Brian J., Han, S., Peng, T., Thams, S., Mikkilineni, S., et al. (2014). Pathways Disrupted in Human ALS Motor Neurons Identified through Genetic Correction of Mutant SOD1. Cell Stem Cell 14, 781– 795. Klim, J.R., Williams, L.A., Limone, F., Guerra San Juan, I., Davis-Dusenbery, B.N., Mordes, D.A., Burberry, A., Steinbaugh, M.J., Gamage, K.K., Kirchner, R., et al. (2019). ALS-implicated protein TDP-43 sustains levels of STMN2, a mediator of motor neuron growth and repair. Nature Neuroscience 22, 167–179. Konishi, H., and Kiyama, H. (2018). Microglial TREM2/DAP12 Signaling: A Double-Edged Sword in Neural Diseases. Frontiers in Cellular Neuroscience 12. Koppers, M., Blokhuis, A.M., Westeneng, H.-J., Terpstra, M.L., Zundel, C.A.C., Vieira de Sá, R., Schellevis, R.D., Waite, A.J., Blake, D.J., Veldink, J.H., et al. (2015). C9orf72 ablation in mice does not cause motor neuron degeneration or motor deficits. Annals of Neurology 78, 426–438. Kosaka, N., Iguchi, H., Yoshioka, Y., Takeshita, F., Matsuki, Y., and Ochiya, T. (2010). Secretory Mechanisms and Intercellular Transfer of MicroRNAs in Living Cells. Journal of Biological Chemistry 285, 17442–17452. 134 Kramer, N.J., Haney, M.S., Morgens, D.W., Jovičić, A., Couthouis, J., Li, A., Ousey, J., Ma, R., Bieri, G., Tsui, C.K., et al. (2018). CRISPR–Cas9 screens in human cells and primary neurons identify modifiers of C9ORF72 dipeptide-repeat-protein toxicity. Nature Genetics 50, 603–612. Krasemann, S., Madore, C., Cialic, R., Baufeld, C., Calcagno, N., El Fatimy, R., Beckers, L., O’Loughlin, E., Xu, Y., Fanek, Z., et al. (2017a). The TREM2-APOE Pathway Drives the Transcriptional Phenotype of Dysfunctional Microglia in Neurodegenerative Diseases. Immunity 47, 566-581.e9. Krasemann, S., Madore, C., Cialic, R., Baufeld, C., Calcagno, N., El Fatimy, R., Beckers, L., O’Loughlin, E., Xu, Y., Fanek, Z., et al. (2017b). The TREM2-APOE Pathway Drives the Transcriptional Phenotype of Dysfunctional Microglia in Neurodegenerative Diseases. Immunity 47, 566-581.e9. Krieger, C., Wagey, R., and Shaw, C. (1993). Amyotrophic lateral sclerosis: quantitative autoradiography of [3H]MK-801/NMDA binding sites in spinal cord. Neuroscience Letters 159, 191–194. Krishnan, G., Raitcheva, D., Bartlett, D., Prudencio, M., McKenna-Yasek, D.M., Douthwright, C., Oskarsson, B.E., Ladha, S., King, O.D., Barmada, S.J., et al. (2022). Poly(GR) and poly(GA) in cerebrospinal fluid as potential biomarkers for C9ORF72-ALS/FTD. Nature Communications 13, 2799. Kuleshov, M.V., Jones, M.R., Rouillard, A.D., Fernandez, N.F., Duan, Q., Wang, Z., Koplev, S., Jenkins, S.L., Jagodnik, K.M., Lachmann, A., et al. (2016). Enrichr: a comprehensive gene set enrichment analysis web server 2016 update. Nucleic Acids Research 44, W90–W97. Kuo, J.J., Schonewille, M., Siddique, T., Schults, A.N.A., Fu, R., Bär, P.R., Anelli, R., Heckman, C.J., and Kroese, A.B.A. (2004). Hyperexcitability of cultured spinal motoneurons from presymptomatic ALS mice. Journal of Neurophysiology 91, 571–575. Kwiatkowski, T.J., Bosco, D.A., Leclerc, A.L., Tamrazian, E., Vanderburg, C.R., Russ, C., Davis, A., Gilchrist, J., Kasarskis, E.J., Munsat, T., et al. (2009). Mutations in the FUS/TLS gene on chromosome 16 cause familial amyotrophic lateral sclerosis. Science (New York, N.Y.) 323, 1205–1208. Kwon, I., Xiang, S., Kato, M., Wu, L., Theodoropoulos, P., Wang, T., Kim, J., Yun, J., Xie, Y., and McKnight, S.L. (2014). Poly-dipeptides encoded by the C9orf72 repeats bind nucleoli, impede RNA biogenesis, and kill cells. Science 345, 1139–1145. Lai, S.-L., Abramzon, Y., Schymick, J.C., Stephan, D.A., Dunckley, T., Dillman, A., Cookson, M., Calvo, A., Battistini, S., Giannini, F., et al. (2011). FUS mutations in sporadic amyotrophic lateral sclerosis. Neurobiology of Aging 32, 550.e1–550.e4. Lakkaraju, A.K.K., Frontzek, K., Lemes, E., Herrmann, U., Losa, M., Marpakwar, R., and Aguzzi, A. (2021). Loss of PIKfyve drives the spongiform degeneration in prion diseases. EMBO Molecular Medicine 13, e14714. Lall, D., Lorenzini, I., Mota, T.A., Bell, S., Mahan, T.E., Ulrich, J.D., Davtyan, H., Rexach, J.E., Muhammad, A.K.M.G., Shelest, O., et al. (2021). C9orf72 deficiency promotes microglial- mediated synaptic loss in aging and amyloid accumulation. Neuron 109, 2275-2291.e8. 135 Lee, S.-S., Won, J.-H., Lim, G.J., Han, J., Lee, J.Y., Cho, K.-O., and Bae, Y.-K. (2019). A novel population of extracellular vesicles smaller than exosomes promotes cell proliferation. Cell Communication and Signaling 17. Levine, T.P., Daniels, R.D., Gatta, A.T., Wong, L.H., and Hayes, M.J. (2013a). The product of C9orf72, a gene strongly implicated in neurodegeneration, is structurally related to DENN Rab- GEFs. Bioinformatics 29, 499–503. Levine, T.P., Daniels, R.D., Gatta, A.T., Wong, L.H., and Hayes, M.J. (2013b). The product of C9orf72, a gene strongly implicated in neurodegeneration, is structurally related to DENN Rab- GEFs. Bioinformatics 29, 499–503. Leyns, C.E.G., Ulrich, J.D., Finn, M.B., Stewart, F.R., Koscal, L.J., Remolina Serrano, J., Robinson, G.O., Anderson, E., Colonna, M., and Holtzman, D.M. (2017). TREM2 deficiency attenuates neuroinflammation and protects against neurodegeneration in a mouse model of tauopathy. Proceedings of the National Academy of Sciences 114, 11524–11529. Liao, Y., Smyth, G.K., and Shi, W. (2019). The R package Rsubread is easier, faster, cheaper and better for alignment and quantification of RNA sequencing reads. Nucleic Acids Research 47, e47–e47. Liddelow, S.A., Guttenplan, K.A., Clarke, L.E., Bennett, F.C., Bohlen, C.J., Schirmer, L., Bennett, M.L., Münch, A.E., Chung, W.-S., Peterson, T.C., et al. (2017). Neurotoxic reactive astrocytes are induced by activated microglia. Nature 541, 481–487. Lieu, C.A., Chinta, S.J., Rane, A., and Andersen, J.K. (2013). Age-Related Behavioral Phenotype of an Astrocytic Monoamine Oxidase-B Transgenic Mouse Model of Parkinson’s Disease. PLoS ONE 8, e54200. Ling, S.-C., Polymenidou, M., and Cleveland, Don W. (2013). Converging Mechanisms in ALS and FTD: Disrupted RNA and Protein Homeostasis. Neuron 79, 416–438. Liu, M.-L., Zang, T., and Zhang, C.-L. (2016a). Direct Lineage Reprogramming Reveals Disease- Specific Phenotypes of Motor Neurons from Human ALS Patients. Cell Reports 14, 115–128. Liu, S., Li, L., Zhang, F., Lim, E., Raychaudhuri, S., and Church, G. Iterative transcription factor screening enables rapid generation of microglia-like cells from human iPSC. Liu, Y., Pattamatta, A., Zu, T., Reid, T., Bardhi, O., Borchelt, David R., Yachnis, Anthony T., and Ranum, Laura P.W. (2016b). C9orf72 BAC Mouse Model with Motor Deficits and Neurodegenerative Features of ALS/FTD. Neuron 90, 521–534. Liu, Y., Pattamatta, A., Zu, T., Reid, T., Bardhi, O., Borchelt, David R., Yachnis, Anthony T., and Ranum, Laura P.W. (2016c). C9orf72 BAC Mouse Model with Motor Deficits and Neurodegenerative Features of ALS/FTD. Neuron 90, 521–534. Love, M.I., Huber, W., and Anders, S. (2014). Moderated estimation of fold change and dispersion for RNA-seq data with DESeq2. Genome Biology 15. Ludwig, A.-K., De Miroschedji, K., Doeppner, T.R., Börger, V., Ruesing, J., Rebmann, V., Durst, S., Jansen, S., Bremer, M., Behrmann, E., et al. (2018). Precipitation with polyethylene glycol 136 followed by washing and pelleting by ultracentrifugation enriches extracellular vesicles from tissue culture supernatants in small and large scales. Journal of Extracellular Vesicles 7. Lyden, P., Levy, H., Weymer, S., Pryor, K., Kramer, W., Griffin, J.H., Davis, T.P., and Zlokovic, B. (2013). Phase 1 Safety, Tolerability and Pharmacokinetics of 3K3A-APC in Healthy Adult Volunteers. Current Pharmaceutical Design 19, 7479–7485. Lyons, A.B. (2000). Analysing cell division in vivo and in vitro using flow cytometric measurement of CFSE dye dilution. Journal of Immunological Methods 243, 147–154. Ma, X.R., Prudencio, M., Koike, Y., Vatsavayai, S.C., Kim, G., Harbinski, F., Briner, A., Rodriguez, C.M., Guo, C., Akiyama, T., et al. (2022). TDP-43 represses cryptic exon inclusion in the FTD– ALS gene UNC13A. Nature 603, 124–130. Malik, M., Simpson, J.F., Parikh, I., Wilfred, B.R., Fardo, D.W., Nelson, P.T., and Estus, S. (2013). CD33 Alzheimer’s Risk-Altering Polymorphism, CD33 Expression, and Exon 2 Splicing. Journal of Neuroscience 33, 13320–13325. Malik, M., Parikh, I., Vasquez, J.B., Smith, C., Tai, L., Bu, G., LaDu, M.J., Fardo, D.W., Rebeck, G.W., and Estus, S. (2015). Genetics ignite focus on microglial inflammation in Alzheimer’s disease. Molecular Neurodegeneration 10. Mancuso, R., Fryatt, G., Cleal, M., Obst, J., Pipi, E., Monzón-Sandoval, J., Ribe, E., Winchester, L., Webber, C., Nevado, A., et al. (2019). CSF1R inhibitor JNJ-40346527 attenuates microglial proliferation and neurodegeneration in P301S mice. Brain 142, 3243–3264. Mann, J.R., Gleixner, A.M., Mauna, J.C., Gomes, E., DeChellis-Marks, M.R., Needham, P.G., Copley, K.E., Hurtle, B., Portz, B., Pyles, N.J., et al. (2019). RNA Binding Antagonizes Neurotoxic Phase Transitions of TDP-43. Neuron 102, 321-338.e8. Maor-Nof, M., Shipony, Z., Lopez-Gonzalez, R., Nakayama, L., Zhang, Y.-J., Couthouis, J., Blum, J.A., Castruita, P.A., Linares, G.R., Ruan, K., et al. (2021). p53 is a central regulator driving neurodegeneration caused by C9orf72 poly(PR). Cell 184, 689-708.e20. Maphis, N., Xu, G., Kokiko-Cochran, O.N., Jiang, S., Cardona, A., Ransohoff, R.M., Lamb, B.T., and Bhaskar, K. (2015). Reactive microglia drive tau pathology and contribute to the spreading of pathological tau in the brain. Brain 138, 1738–1755. Marat, A.L., Dokainish, H., and McPherson, P.S. (2011). DENN Domain Proteins: Regulators of Rab GTPases. Journal of Biological Chemistry 286, 13791–13800. Martin, S., Harper, C.B., May, L.M., Coulson, E.J., Meunier, F.A., and Osborne, S.L. (2013). Inhibition of PIKfyve by YM-201636 dysregulates autophagy and leads to apoptosis-independent neuronal cell death. PloS One 8, e60152. Martínez-Muriana, A., Mancuso, R., Francos-Quijorna, I., Olmos-Alonso, A., Osta, R., Perry, V.H., Navarro, X., Gomez-Nicola, D., and López-Vales, R. (2016). CSF1R blockade slows the progression of amyotrophic lateral sclerosis by reducing microgliosis and invasion of macrophages into peripheral nerves. Scientific Reports 6, 1–13. 137 Mathys, H., Adaikkan, C., Gao, F., Young, J.Z., Manet, E., Hemberg, M., De Jager, P.L., Ransohoff, R.M., Regev, A., and Tsai, L.-H. (2017). Temporal Tracking of Microglia Activation in Neurodegeneration at Single-Cell Resolution. Cell Reports 21, 366–380. Mauvezin, C., and Neufeld, T.P. (2015). Bafilomycin A1 disrupts autophagic flux by inhibiting both V-ATPase-dependent acidification and Ca-P60A/SERCA-dependent autophagosome-lysosome fusion. Autophagy 11, 1437–1438. McCampbell, A., Cole, T., Wegener, A.J., Tomassy, G.S., Setnicka, A., Farley, B.J., Schoch, K.M., Hoye, M.L., Shabsovich, M., Sun, L., et al. (2018a). Antisense oligonucleotides extend survival and reverse decrement in muscle response in ALS models. The Journal of Clinical Investigation 128, 3558–3567. McCampbell, A., Cole, T., Wegener, A.J., Tomassy, G.S., Setnicka, A., Farley, B.J., Schoch, K.M., Hoye, M.L., Shabsovich, M., Sun, L., et al. (2018b). Antisense oligonucleotides extend survival and reverse decrement in muscle response in ALS models. The Journal of Clinical Investigation 128, 3558–3567. McCauley, M.E., O’Rourke, J.G., Yáñez, A., Markman, J.L., Ho, R., Wang, X., Chen, S., Lall, D., Jin, M., Muhammad, A.K.M.G., et al. (2020). C9orf72 in myeloid cells suppresses STING-induced inflammation. Nature 585, 96–101. McQuade, A., and Blurton-Jones, M. (2019). Microglia in Alzheimer’s Disease: Exploring How Genetics and Phenotype Influence Risk. Journal of Molecular Biology 431. Melamed, Z., López-Erauskin, J., Baughn, M.W., Zhang, O., Drenner, K., Sun, Y., Freyermuth, F., McMahon, M.A., Beccari, M.S., Artates, J.W., et al. (2019). Premature polyadenylation- mediated loss of stathmin-2 is a hallmark of TDP-43-dependent neurodegeneration. Nature Neuroscience 22, 180–190. Melentijevic, I., Toth, M.L., Arnold, M.L., Guasp, R.J., Harinath, G., Nguyen, K.C., Taub, D., Parker, J.A., Neri, C., Gabel, C.V., et al. (2017). C. elegans neurons jettison protein aggregates and mitochondria under neurotoxic stress. Nature 542, 367–371. Menzies, F.M., Fleming, A., and Rubinsztein, D.C. (2015). Compromised autophagy and neurodegenerative diseases. Nature Reviews. Neuroscience 16, 345–357. Miller, R.G., Mitchell, J.D., Lyon, M., and Moore, D.H. (2002). Riluzole for amyotrophic lateral sclerosis (ALS)/motor neuron disease (MND). The Cochrane Database of Systematic Reviews CD001447. Miller, T., Cudkowicz, M., Shaw, P.J., Andersen, P.M., Atassi, N., Bucelli, R.C., Genge, A., Glass, J., Ladha, S., Ludolph, A.L., et al. (2020a). Phase 1–2 Trial of Antisense Oligonucleotide Tofersen for SOD1 ALS. New England Journal of Medicine 383, 109–119. Miller, T., Cudkowicz, M., Shaw, P.J., Andersen, P.M., Atassi, N., Bucelli, R.C., Genge, A., Glass, J., Ladha, S., Ludolph, A.L., et al. (2020b). Phase 1–2 Trial of Antisense Oligonucleotide Tofersen for SOD1 ALS. New England Journal of Medicine 383, 109–119. 138 Mizielinska, S., Gronke, S., Niccoli, T., Ridler, C.E., Clayton, E.L., Devoy, A., Moens, T., Norona, F.E., Woollacott, I.O.C., Pietrzyk, J., et al. (2014). C9orf72 repeat expansions cause neurodegeneration in Drosophila through arginine-rich proteins. Science 345, 1192–1194. Moore, S., Alsop, E., Lorenzini, I., Starr, A., Rabichow, B.E., Mendez, E., Levy, J.L., Burciu, C., Reiman, R., Chew, J., et al. (2019). ADAR2 mislocalization and widespread RNA editing aberrations in C9orf72-mediated ALS/FTD. Acta Neuropathologica 138, 49–65. Moscat, J., and Diaz-Meco, M.T. (2012). p62: a versatile multitasker takes on cancer. Trends in Biochemical Sciences 37, 230–236. Mosnier, L.O., Gale, A.J., Yegneswaran, S., and Griffin, J.H. (2004). Activated protein C variants with normal cytoprotective but reduced anticoagulant activity. Blood 104, 1740–1744. Muenz, L.R., Green, S.B., and Byar, D.P. (1977). Applications of the Mantel-Haenszel statistic to the comparison of survival distributions. Biometrics 33, 617–626. Muffat, J., Li, Y., Yuan, B., Mitalipova, M., Omer, A., Corcoran, S., Bakiasi, G., Tsai, L.-H., Aubourg, P., Ransohoff, R.M., et al. (2016). Efficient derivation of microglia-like cells from human pluripotent stem cells. Nature Medicine 22, 1358–1367. Neumann, M., Sampathu, D.M., Kwong, L.K., Truax, A.C., Micsenyi, M.C., Chou, T.T., Bruce, J., Schuck, T., Grossman, M., Clark, C.M., et al. (2006). Ubiquitinated TDP-43 in Frontotemporal Lobar Degeneration and Amyotrophic Lateral Sclerosis. Science 314, 130–133. Neumann, M., Kwong, L.K., Lee, E.B., Kremmer, E., Flatley, A., Xu, Y., Forman, M.S., Troost, D., Kretzschmar, H.A., Trojanowski, J.Q., et al. (2009). Phosphorylation of S409/410 of TDP-43 is a consistent feature in all sporadic and familial forms of TDP-43 proteinopathies. Acta Neuropathologica 117, 137–149. Nguyen, L., Montrasio, F., Pattamatta, A., Tusi, S.K., Bardhi, O., Meyer, K.D., Hayes, L., Nakamura, K., Banez-Coronel, M., Coyne, A., et al. (2020). Antibody therapy targeting RAN proteins rescues C9 ALS/FTD phenotypes in C9orf72 mouse model. Neuron 105, 645-662.e11. O’Rourke, J.G., Bogdanik, L., Muhammad, A.K.M.G., Gendron, T.F., Kim, K.J., Austin, A., Cady, J., Liu, E.Y., Zarrow, J., Grant, S., et al. (2015a). C9orf72 BAC Transgenic Mice Display Typical Pathologic Features of ALS/FTD. Neuron 88, 892–901. O’Rourke, J.G., Bogdanik, L., Muhammad, A.K.M.G., Gendron, T.F., Kim, K.J., Austin, A., Cady, J., Liu, E.Y., Zarrow, J., Grant, S., et al. (2015b). C9orf72 BAC Transgenic Mice Display Typical Pathologic Features of ALS/FTD. Neuron 88, 892–901. O’Rourke, J.G., Bogdanik, L., Yáñez, A., Lall, D., Wolf, A.J., Muhammad, A.K.M.G., Ho, R., Carmona, S., Vit, J.P., Zarrow, J., et al. (2016a). C9orf72   is required for proper macrophage and microglial function in mice. Science 351, 1324–1329. O’Rourke, J.G., Bogdanik, L., Yáñez, A., Lall, D., Wolf, A.J., Muhammad, A.K.M.G., Ho, R., Carmona, S., Vit, J.P., Zarrow, J., et al. (2016b). C9orf72   is required for proper macrophage and microglial function in mice. Science 351, 1324–1329. 139 Oakes, J.A., Davies, M.C., and Collins, M.O. (2017). TBK1: a new player in ALS linking autophagy and neuroinflammation. Molecular Brain 10. Otero, K., Turnbull, I.R., Poliani, P.L., Vermi, W., Cerutti, E., Aoshi, T., Tassi, I., Takai, T., Stanley, S.L., Miller, M., et al. (2009). Macrophage colony-stimulating factor induces the proliferation and survival of macrophages via a pathway involving DAP12 and β-catenin. Nature Immunology 10, 734–743. Palmerini, E., Longhi, A., Donati, D.M., and Staals, E.L. (2020). Pexidartinib for the treatment of adult patients with symptomatic tenosynovial giant cell tumor: safety and efficacy. Expert Review of Anticancer Therapy 20, 441–445. Paul, F., Arkin, Y., Giladi, A., Jaitin, D.A., Kenigsberg, E., Keren-Shaul, H., Winter, D., Lara- Astiaso, D., Gury, M., Weiner, A., et al. (2015). Transcriptional Heterogeneity and Lineage Commitment in Myeloid Progenitors. Cell 163, 1663–1677. Penney, J., Ralvenius, W.T., and Tsai, L.-H. (2020). Modeling Alzheimer’s disease with iPSC- derived brain cells. Molecular Psychiatry 25, 148–167. Perry, S., Han, Y., Das, A., and Dickman, D. (2017). Homeostatic plasticity can be induced and expressed to restore synaptic strength at neuromuscular junctions undergoing ALS-related degeneration. Human Molecular Genetics 26, 4153–4167. Pihlstrøm, L., Wiethoff, S., and Houlden, H. (2018). Chapter 22 - Genetics of neurodegenerative diseases: an overview (Elsevier). Ponpuak, M., Mandell, M., Kimura, T., Chauhan, S., Cleyrat, C., and Deretic, V. (2015). Secretory autophagy. Current Opinion in Cell Biology 35, 106–116. Qiu, X., Hill, A., Packer, J., Lin, D., Ma, Y.-A., and Trapnell, C. (2017). Single-cell mRNA quantification and differential analysis with Census. Nature Methods 14, 309–315. Quek, H., Cuní-López, C., Stewart, R., Colletti, T., Notaro, A., Nguyen, T.H., Sun, Y., Guo, C.C., Lupton, M.K., Roberts, T.L., et al. (2022). ALS monocyte-derived microglia-like cells reveal cytoplasmic TDP-43 accumulation, DNA damage, and cell-specific impairment of phagocytosis associated with disease progression. Journal of Neuroinflammation 19. Renton, Alan E., Majounie, E., Waite, A., Simón-Sánchez, J., Rollinson, S., Gibbs, J. Raphael, Schymick, Jennifer C., Laaksovirta, H., van Swieten, John C., Myllykangas, L., et al. (2011). A Hexanucleotide Repeat Expansion in C9ORF72 Is the Cause of Chromosome 9p21-Linked ALS- FTD. Neuron 72, 257–268. Richter, B., Sliter, D.A., Herhaus, L., Stolz, A., Wang, C., Beli, P., Zaffagnini, G., Wild, P., Martens, S., Wagner, S.A., et al. (2016). Phosphorylation of OPTN by TBK1 enhances its binding to Ub chains and promotes selective autophagy of damaged mitochondria. Proceedings of the National Academy of Sciences 113, 4039–4044. Rizzu, P., Blauwendraat, C., Heetveld, S., Lynes, E.M., Castillo-Lizardo, M., Dhingra, A., Pyz, E., Hobert, M., Synofzik, M., Simón-Sánchez, J., et al. (2016). C9orf72 is differentially expressed in the central nervous system and myeloid cells and consistently reduced in C9orf72, MAPT and GRN mutation carriers. Acta Neuropathologica Communications 4. 140 Root, J., Merino, P., Nuckols, A., Johnson, M., and Kukar, T. (2021). Lysosome dysfunction as a cause of neurodegenerative diseases: Lessons from frontotemporal dementia and amyotrophic lateral sclerosis. Neurobiology of Disease 154, 105360. Rostalski, H., Hietanen, T., Leskelä, S., Behánová, A., Abdollahzadeh, A., Wittrahm, R., Mäkinen, P., Huber, N., Hoffmann, D., Solje, E., et al. (2020). BV-2 Microglial Cells Overexpressing C9orf72 Hexanucleotide Repeat Expansion Produce DPR Proteins and Show Normal Functionality but No RNA Foci. Frontiers in Neurology 11. Rothstein, J.D. (2017). Edaravone: A new drug approved for ALS. Cell 171, 725. Rothstein, J.D., Dykes-Hoberg, M., Pardo, C.A., Bristol, L.A., Jin, L., Kuncl, R.W., Kanai, Y., Hediger, M.A., Wang, Y., Schielke, J.P., et al. (1996). Knockout of Glutamate Transporters Reveals a Major Role for Astroglial Transport in Excitotoxicity and Clearance of Glutamate. Neuron 16, 675–686. Rudich, P., Snoznik, C., Watkins, S.C., Monaghan, J., Pandey, U.B., and Lamitina, S.T. (2017). Nuclear localized C9orf72-associated arginine-containing dipeptides exhibit age-dependent toxicity in C. elegans. Human Molecular Genetics 26, 4916–4928. Saez-Atienzar, S., Bandres-Ciga, S., Langston, R.G., Kim, J.J., Choi, S.W., Reynolds, R.H., Abramzon, Y., Dewan, R., Ahmed, S., Landers, J.E., et al. (2021). Genetic analysis of amyotrophic lateral sclerosis identifies contributing pathways and cell types. Science Advances 7. Salter, M.W., and Stevens, B. (2017a). Microglia emerge as central players in brain disease. Nature Medicine 23, 1018–1027. Salter, M.W., and Stevens, B. (2017b). Microglia emerge as central players in brain disease. Nature Medicine 23, 1018–1027. Sareen, D., O’Rourke, J.G., Meera, P., Muhammad, A.K.M.G., Grant, S., Simpkinson, M., Bell, S., Carmona, S., Ornelas, L., Sahabian, A., et al. (2013). Targeting RNA Foci in iPSC-Derived Motor Neurons from ALS Patients with a C9ORF72 Repeat Expansion. Science Translational Medicine 5, 208ra149–208ra149. Schwenk, F., Baron, U., and Rajewsky, K. (1995). A cre-transgenic mouse strain for the ubiquitous deletion of loxP-flanked gene segments including deletion in germ cells. Nucleic Acids Research 23, 5080–5081. Seebohm, G., Neumann, S., Theiss, C., Novkovic, T., Hill, E.V., Tavaré, J.M., Lang, F., Hollmann, M., Manahan-Vaughan, D., and Strutz-Seebohm, N. (2012). Identification of a Novel Signaling Pathway and Its Relevance for GluA1 Recycling. PLoS ONE 7, e33889. Sellier, C., Campanari, M., Julie Corbier, C., Gaucherot, A., Kolb ‐Cheynel, I., Oulad ‐Abdelghani, M., Ruffenach, F., Page, A., Ciura, S., Kabashi, E., et al. (2016). Loss of C9   ORF   72 impairs autophagy and synergizes with polyQ Ataxin ‐2 to induce motor neuron dysfunction and cell death. The EMBO Journal 35, 1276–1297. 141 Shao, Q., Liang, C., Chang, Q., Zhang, W., Yang, M., and Chen, J.-F. (2019). C9orf72 deficiency promotes motor deficits of a C9ALS/FTD mouse model in a dose-dependent manner. Acta Neuropathologica Communications 7. Shaw, M.P., Higginbottom, A., McGown, A., Castelli, L.M., James, E., Hautbergue, G.M., Shaw, P.J., and Ramesh, T.M. (2018). Stable transgenic C9orf72 zebrafish model key aspects of the ALS/FTD phenotype and reveal novel pathological features. Acta Neuropathologica Communications 6. Shi, Y., Lin, S., Staats, K.A., Li, Y., Chang, W.-H., Hung, S.-T., Hendricks, E., Linares, G.R., Wang, Y., Son, E.Y., et al. (2018a). Haploinsufficiency leads to neurodegeneration in C9ORF72 ALS/FTD human induced motor neurons. Nature Medicine 24, 313–325. Shi, Y., Lin, S., Staats, K.A., Li, Y., Chang, W.-H., Hung, S.-T., Hendricks, E., Linares, G.R., Wang, Y., Son, E.Y., et al. (2018b). Haploinsufficiency leads to neurodegeneration in C9ORF72 ALS/FTD human induced motor neurons. Nature Medicine 24, 313–325. Shi, Y., Hung, S.-T., Rocha, G., Lin, S., Linares, G.R., Staats, K.A., Seah, C., Wang, Y., Chickering, M., Lai, J., et al. (2019). Identification and therapeutic rescue of autophagosome and glutamate receptor defects in C9ORF72 and sporadic ALS neurons. JCI Insight 5, 127736. Siracusa, R., Fusco, R., and Cuzzocrea, S. (2019). Astrocytes: Role and Functions in Brain Pathologies. Frontiers in Pharmacology 10. Smeyers, J., Banchi, E.-G., and Latouche, M. (2021). C9ORF72: What It Is, What It Does, and Why It Matters. Frontiers in Cellular Neuroscience 15. Smith, R., Pioro, E., Myers, K., Sirdofsky, M., Goslin, K., Meekins, G., Yu, H., Wymer, J., Cudkowicz, M., Macklin, E.A., et al. (2017). Enhanced Bulbar Function in Amyotrophic Lateral Sclerosis: The Nuedexta Treatment Trial. Neurotherapeutics 14, 762–772. Soares, A.C., Ferreira, A., Mariën, J., Delay, C., Lee, E., Trojanowski, J.Q., Moechars, D., Annaert, W., and De Muynck, L. (2021). PIKfyve activity is required for lysosomal trafficking of tau aggregates and tau seeding. Journal of Biological Chemistry 100636. Sofroniew, M.V., and Vinters, H.V. (2009). Astrocytes: biology and pathology. Acta Neuropathologica 119, 7–35. Son, Esther Y., Ichida, Justin K., Wainger, Brian J., Toma, Jeremy S., Rafuse, Victor F., Woolf, Clifford J., and Eggan, K. (2011a). Conversion of Mouse and Human Fibroblasts into Functional Spinal Motor Neurons. Cell Stem Cell 9, 205–218. Son, Esther Y., Ichida, Justin K., Wainger, Brian J., Toma, Jeremy S., Rafuse, Victor F., Woolf, Clifford J., and Eggan, K. (2011b). Conversion of Mouse and Human Fibroblasts into Functional Spinal Motor Neurons. Cell Stem Cell 9, 205–218. Song, W.M., and Colonna, M. (2018). The identity and function of microglia in neurodegeneration. Nature Immunology 19, 1048–1058. Sosna, J., Philipp, S., Albay, R., Reyes-Ruiz, J.M., Baglietto-Vargas, D., LaFerla, F.M., and Glabe, C.G. (2018). Early long-term administration of the CSF1R inhibitor PLX3397 ablates microglia 142 and reduces accumulation of intraneuronal amyloid, neuritic plaque deposition and pre-fibrillar oligomers in 5XFAD mouse model of Alzheimer’s disease. Molecular Neurodegeneration 13. Spangenberg, E., Severson, P.L., Hohsfield, L.A., Crapser, J., Zhang, J., Burton, E.A., Zhang, Y., Spevak, W., Lin, J., Phan, N.Y., et al. (2019). Sustained microglial depletion with CSF1R inhibitor impairs parenchymal plaque development in an Alzheimer’s disease model. Nature Communications 10. Spiller, K.J., Restrepo, C.R., Khan, T., Dominique, M.A., Fang, T.C., Canter, R.G., Roberts, C.J., Miller, K.R., Ransohoff, R.M., Trojanowski, J.Q., et al. (2018). Microglia-mediated recovery from ALS-relevant motor neuron degeneration in a mouse model of TDP-43 proteinopathy. Nature Neuroscience 21, 329–340. Sreedharan, J., Blair, I.P., Tripathi, V.B., Hu, X., Vance, C., Rogelj, B., Ackerley, S., Durnall, J.C., Williams, K.L., Buratti, E., et al. (2008). TDP-43 Mutations in Familial and Sporadic Amyotrophic Lateral Sclerosis. Science 319, 1668–1672. Srinivasan, K., Friedman, B.A., Larson, J.L., Lauffer, B.E., Goldstein, L.D., Appling, L.L., Borneo, J., Poon, C., Ho, T., Cai, F., et al. (2016). Untangling the brain’s neuroinflammatory and neurodegenerative transcriptional responses. Nature Communications 7. Stuart, T., Butler, A., Hoffman, P., Hafemeister, C., Papalexi, E., Mauck, W.M., Hao, Y., Stoeckius, M., Smibert, P., and Satija, R. (2019). Comprehensive Integration of Single-Cell Data. Cell 177, 1888-1902.e21. Sudria-Lopez, E., Koppers, M., de Wit, M., van der Meer, C., Westeneng, H.-J., Zundel, C.A.C., Youssef, S.A., Harkema, L., de Bruin, A., Veldink, J.H., et al. (2016). Full ablation of C9orf72 in mice causes immune system-related pathology and neoplastic events but no motor neuron defects. Acta Neuropathologica 132, 145–147. Suk, T.R., and Rousseaux, M.W.C. (2020). The role of TDP-43 mislocalization in amyotrophic lateral sclerosis. Molecular Neurodegeneration 15. Suzuki, N., Maroof, A., Merkle, F.T., Koszka, K., Intoh, A., Armstrong, I., Moccia, R., Davis- Dusenbery, B.N., and Eggan, K. (2013). The mouse C9ORF72 ortholog is enriched in neurons known to degenerate in ALS and FTD. Nature Neuroscience 16, 1725–1727. Sweeney, P., Park, H., Baumann, M., Dunlop, J., Frydman, J., Kopito, R., McCampbell, A., Leblanc, G., Venkateswaran, A., Nurmi, A., et al. (2017). Protein misfolding in neurodegenerative diseases: implications and strategies. Translational Neurodegeneration 6. Takahashi, K., and Yamanaka, S. (2006). Induction of Pluripotent Stem Cells from Mouse Embryonic and Adult Fibroblast Cultures by Defined Factors. Cell 126, 663–676. Taylor, J.P., Brown, R.H., and Cleveland, D.W. (2016). Decoding ALS: from genes to mechanism. Nature 539, 197–206. Therrien, M., Rouleau, G.A., Dion, P.A., and Parker, J.A. (2013). Deletion of C9ORF72 Results in Motor Neuron Degeneration and Stress Sensitivity in C. elegans. PLoS ONE 8, e83450. 143 Tischbein, M., Baron, D.M., Lin, Y.-C., Gall, K.V., Landers, J.E., Fallini, C., and Bosco, D.A. (2019). The RNA-binding protein FUS/TLS undergoes calcium-mediated nuclear egress during excitotoxic stress and is required for GRIA2 mRNA processing. The Journal of Biological Chemistry 294, 10194–10210. Tran, H., Almeida, S., Moore, J., Gendron, Tania F., Chalasani, U., Lu, Y., Du, X., Nickerson, Jeffrey A., Petrucelli, L., Weng, Z., et al. (2015). Differential Toxicity of Nuclear RNA Foci versus Dipeptide Repeat Proteins in a Drosophila Model of C9ORF72 FTD/ALS. Neuron 87, 1207–1214. Tran, H., Moazami, M.P., Yang, H., McKenna-Yasek, D., Douthwright, C.L., Pinto, C., Metterville, J., Shin, M., Sanil, N., Dooley, C., et al. (2022a). Suppression of mutant C9orf72 expression by a potent mixed backbone antisense oligonucleotide. Nature Medicine 28, 117–124. Tran, H., Moazami, M.P., Yang, H., McKenna-Yasek, D., Douthwright, C.L., Pinto, C., Metterville, J., Shin, M., Sanil, N., Dooley, C., et al. (2022b). Suppression of mutant C9orf72 expression by a potent mixed backbone antisense oligonucleotide. Nature Medicine 28, 117–124. Tsai, R.M., and Boxer, A.L. (2014). Treatment of Frontotemporal Dementia. Current Treatment Options in Neurology 16. van Blitterswijk, M., Gendron, T.F., Baker, M.C., DeJesus-Hernandez, M., Finch, N.A., Brown, P.H., Daughrity, L.M., Murray, M.E., Heckman, M.G., Jiang, J., et al. (2015). Novel clinical associations with specific C9ORF72 transcripts in patients with repeat expansions in C9ORF72. Acta Neuropathologica 130, 863–876. Van Damme, P., Robberecht, W., and Van Den Bosch, L. (2017). Modelling amyotrophic lateral sclerosis: progress and possibilities. Disease Models & Mechanisms 10, 537–549. Van de Sande, B., Flerin, C., Davie, K., De Waegeneer, M., Hulselmans, G., Aibar, S., Seurinck, R., Saelens, W., Cannoodt, R., Rouchon, Q., et al. (2020). A scalable SCENIC workflow for single- cell gene regulatory network analysis. Nature Protocols 15, 2247–2276. van Rheenen, W., Shatunov, A., Dekker, A.M., McLaughlin, R.L., Diekstra, F.P., Pulit, S.L., van der Spek, R.A.A., Võsa, U., de Jong, S., Robinson, M.R., et al. (2016). Genome-wide association analyses identify new risk variants and the genetic architecture of amyotrophic lateral sclerosis. Nature Genetics 48, 1043–1048. van Zundert, B., Peuscher, M.H., Hynynen, M., Chen, A., Neve, R.L., Brown, R.H., Constantine- Paton, M., and Bellingham, M.C. (2008). Neonatal Neuronal Circuitry Shows Hyperexcitable Disturbance in a Mouse Model of the Adult-Onset Neurodegenerative Disease Amyotrophic Lateral Sclerosis. Journal of Neuroscience 28, 10864–10874. Vance, C., Rogelj, B., Hortobagyi, T., De Vos, K.J., Nishimura, A.L., Sreedharan, J., Hu, X., Smith, B., Ruddy, D., Wright, P., et al. (2009). Mutations in FUS, an RNA Processing Protein, Cause Familial Amyotrophic Lateral Sclerosis Type 6. Science 323, 1208–1211. Vanhauwaert, R., Kuenen, S., Masius, R., Bademosi, A., Manetsberger, J., Schoovaerts, N., Bounti, L., Gontcharenko, S., Swerts, J., Vilain, S., et al. (2017). The   SAC   1 domain in synaptojanin is required for autophagosome maturation at presynaptic terminals. The EMBO Journal 36, 1392–1411. 144 Verderio, C., Cagnoli, C., Bergami, M., Francolini, M., Schenk, U., Colombo, A., Riganti, L., Frassoni, C., Zuccaro, E., Danglot, L., et al. (2012). TI-VAMP/VAMP7 is the SNARE of secretory lysosomes contributing to ATP secretion from astrocytes. Biology of the Cell 104, 213–228. Vierbuchen, T., Ostermeier, A., Pang, Z.P., Kokubu, Y., Südhof, T.C., and Wernig, M. (2010). Direct conversion of fibroblasts to functional neurons by defined factors. Nature 463, 1035–1041. Vucic, S., and Kiernan, M.C. (2006a). Novel threshold tracking techniques suggest that cortical hyperexcitability is an early feature of motor neuron disease. Brain: A Journal of Neurology 129, 2436–2446. Vucic, S., and Kiernan, M.C. (2006b). Axonal excitability properties in amyotrophic lateral sclerosis. Clinical Neurophysiology 117, 1458–1466. Wainger, Brian J., Kiskinis, E., Mellin, C., Wiskow, O., Han, Steve S.W., Sandoe, J., Perez, Numa P., Williams, Luis A., Lee, S., Boulting, G., et al. (2014). Intrinsic Membrane Hyperexcitability of Amyotrophic Lateral Sclerosis Patient-Derived Motor Neurons. Cell Reports 7, 1–11. Wang, Y., Cella, M., Mallinson, K., Ulrich, Jason D., Young, Katherine L., Robinette, Michelle L., Gilfillan, S., Krishnan, Gokul M., Sudhakar, S., Zinselmeyer, Bernd H., et al. (2015). TREM2 Lipid Sensing Sustains the Microglial Response in an Alzheimer’s Disease Model. Cell 160, 1061–1071. Wang, Y., Wang, R., Zhang, S., Song, S., Jiang, C., and Wang, L. iTALK: an R Package to Characterize and Illustrate Intercellular Communication. Webster, C.P., Smith, E.F., Bauer, C.S., Moller, A., Hautbergue, G.M., Ferraiuolo, L., Myszczynska, M.A., Higginbottom, A., Walsh, M.J., Whitworth, A.J., et al. (2016). The C9orf72 protein interacts with Rab1a and the   ULK   1 complex to regulate initiation of autophagy. The EMBO Journal 35, 1656–1676. Wen, X., Tan, W., Westergard, T., Krishnamurthy, K., Markandaiah, S.S., Shi, Y., Lin, S., Shneider, Neil A., Monaghan, J., Pandey, Udai B., et al. (2014). Antisense Proline-Arginine RAN Dipeptides Linked to C9ORF72-ALS/FTD Form Toxic Nuclear Aggregates that Initiate In Vitro and In Vivo Neuronal Death. Neuron 84, 1213–1225. Williams, K.L., Warraich, S.T., Yang, S., Solski, J.A., Fernando, R., Rouleau, G.A., Nicholson, G.A., and Blair, I.P. (2012a). UBQLN2/ubiquilin 2 mutation and pathology in familial amyotrophic lateral sclerosis. Neurobiology of Aging 33, 2527.e3–2527.e10. Williams, P.D., Zlokovic, B.V., Griffin, J.H., Pryor, K.E., and Davis, T.P. (2012b). Preclinical Safety and Pharmacokinetic Profile of 3K3A-APC, a Novel, Modified Activated Protein C for Ischemic Stroke. Current Pharmaceutical Design 18, 4215–4222. Willms, E., Johansson, H.J., Mäger, I., Lee, Y., Blomberg, K.E.M., Sadik, M., Alaarg, A., Smith, C.I.E., Lehtiö, J., EL Andaloussi, S., et al. (2016). Cells release subpopulations of exosomes with distinct molecular and biological properties. Scientific Reports 6, 1–12. Wils, H., Kleinberger, G., Janssens, J., Pereson, S., Joris, G., Cuijt, I., Smits, V., Ceuterick-de Groote, C., Van Broeckhoven, C., and Kumar-Singh, S. (2010). TDP-43 transgenic mice develop 145 spastic paralysis and neuronal inclusions characteristic of ALS and frontotemporal lobar degeneration. Proceedings of the National Academy of Sciences 107, 3858–3863. Wlodarczyk, A., Holtman, I.R., Krueger, M., Yogev, N., Bruttger, J., Khorooshi, R., Benmamar ‐ Badel, A., Boer ‐Bergsma, J.J., Martin, N.A., Karram, K., et al. (2017). A novel microglial subset plays a key role in myelinogenesis in developing brain. The EMBO Journal 36, 3292–3308. Xu, Z., Poidevin, M., Li, X., Li, Y., Shu, L., Nelson, D.L., Li, H., Hales, C.M., Gearing, M., Wingo, T.S., et al. (2013). Expanded GGGGCC repeat RNA associated with amyotrophic lateral sclerosis and frontotemporal dementia causes neurodegeneration. Proceedings of the National Academy of Sciences 110, 7778–7783. Yamada, S.B., Gendron, T.F., Niccoli, T., Genuth, N.R., Grosely, R., Shi, Y., Glaria, I., Kramer, N.J., Nakayama, L., Fang, S., et al. (2019). RPS25 is required for efficient RAN translation of C9orf72 and other neurodegenerative disease-associated nucleotide repeats. Nature Neuroscience 22, 1383–1388. Yamaguchi, M., Noda, Nobuo N., Yamamoto, H., Shima, T., Kumeta, H., Kobashigawa, Y., Akada, R., Ohsumi, Y., and Inagaki, F. (2012). Structural Insights into Atg10-Mediated Formation of the Autophagy-Essential Atg12-Atg5 Conjugate. Structure 20, 1244–1254. Yang, M., Liang, C., Swaminathan, K., Herrlinger, S., Lai, F., Shiekhattar, R., and Chen, J.-F. (2016). A C9ORF72/SMCR8-containing complex regulates ULK1 and plays a dual role in autophagy. Science Advances 2, e1601167. Yeamans, C., Wang, D., Paz-Priel, I., Torbett, B.E., Tenen, D.G., and Friedman, A.D. (2007). C/EBPα binds and activates the PU.1 distal enhancer to induce monocyte lineage commitment. Blood 110, 3136–3142. Yen, Y.-T., Yang, H.-R., Lo, H.-C., Hsieh, Y.-C., Tsai, S.-C., Hong, C.-W., and Hsieh, C.-H. (2013). Enhancing autophagy with activated protein C and rapamycin protects against sepsis-induced acute lung injury. Surgery 153, 689–698. Ying, H., and Yue, B.Y.J.T. (2016). Optineurin: the Autophagy Connection. Experimental Eye Research 144, 73–80. Zaborowski, M.P., Balaj, L., Breakefield, X.O., and Lai, C.P. (2015). Extracellular Vesicles: Composition, Biological Relevance, and Methods of Study. BioScience 65, 783–797. Zeisel, A., Munoz-Manchado, A.B., Codeluppi, S., Lonnerberg, P., La Manno, G., Jureus, A., Marques, S., Munguba, H., He, L., Betsholtz, C., et al. (2015). Cell types in the mouse cortex and hippocampus revealed by single-cell RNA-seq. Science 347, 1138–1142. Zhang, J., Velmeshev, D., Hashimoto, K., Huang, Y.-H., Hofmann, J.W., Shi, X., Chen, J., Leidal, A.M., Dishart, J.G., Cahill, M.K., et al. (2020). Neurotoxic microglia promote TDP-43 proteinopathy in progranulin deficiency. Nature 588, 459–465. Zhang, K., Donnelly, C.J., Haeusler, A.R., Grima, J.C., Machamer, J.B., Steinwald, P., Daley, E.L., Miller, S.J., Cunningham, K.M., Vidensky, S., et al. (2015). The C9orf72 repeat expansion disrupts nucleocytoplasmic transport. Nature 525, 56–61. 146 Zhang, Y., Chen, K., Sloan, S.A., Bennett, M.L., Scholze, A.R., O’Keeffe, S., Phatnani, H.P., Guarnieri, P., Caneda, C., Ruderisch, N., et al. (2014). An RNA-Sequencing Transcriptome and Splicing Database of Glia, Neurons, and Vascular Cells of the Cerebral Cortex. The Journal of Neuroscience 34, 11929–11947. Zhang, Y., Sloan, Steven A., Clarke, Laura E., Caneda, C., Plaza, Colton A., Blumenthal, Paul D., Vogel, H., Steinberg, Gary K., Edwards, Michael S.B., Li, G., et al. (2016). Purification and Characterization of Progenitor and Mature Human Astrocytes Reveals Transcriptional and Functional Differences with Mouse. Neuron 89, 37–53. Zhang, Y., Burberry, A., Wang, J.-Y., Sandoe, J., Ghosh, S., Udeshi, N.D., Svinkina, T., Mordes, D.A., Mok, J., Charlton, M., et al. (2018a). The C9orf72-interacting protein Smcr8 is a negative regulator of autoimmunity and lysosomal exocytosis. Genes & Development 32, 929–943. Zhang, Y., Burberry, A., Wang, J.-Y., Sandoe, J., Ghosh, S., Udeshi, N.D., Svinkina, T., Mordes, D.A., Mok, J., Charlton, M., et al. (2018b). The C9orf72-interacting protein Smcr8 is a negative regulator of autoimmunity and lysosomal exocytosis. Genes & Development 32, 929–943. Zhang, Y.-J., Gendron, T.F., Ebbert, M.T.W., O’Raw, A.D., Yue, M., Jansen-West, K., Zhang, X., Prudencio, M., Chew, J., Cook, C.N., et al. (2018c). Poly(GR) impairs protein translation and stress granule dynamics in C9orf72-associated frontotemporal dementia and amyotrophic lateral sclerosis. Nature Medicine 24, 1136–1142. Zhang, Y.-J., Gendron, T.F., Ebbert, M.T.W., O’Raw, A.D., Yue, M., Jansen-West, K., Zhang, X., Prudencio, M., Chew, J., Cook, C.N., et al. (2018d). Poly(GR) impairs protein translation and stress granule dynamics in C9orf72-associated frontotemporal dementia and amyotrophic lateral sclerosis. Nature Medicine 24, 1136–1142. Zhong, Z., Ilieva, H., Hallagan, L., Bell, R., Singh, I., Paquette, N., Thiyagarajan, M., Deane, R., Fernandez, J.A., Lane, S., et al. (2009). Activated protein C therapy slows ALS-like disease in mice by transcriptionally inhibiting SOD1 in motor neurons and microglia cells. The Journal of Clinical Investigation 119, 3437–3449. Zolov, S.N., Bridges, D., Zhang, Y., Lee, W.-W., Riehle, E., Verma, R., Lenk, G.M., Converso- Baran, K., Weide, T., Albin, R.L., et al. (2012). In vivo, Pikfyve generates PI(3,5)P2, which serves as both a signaling lipid and the major precursor for PI5P. Proceedings of the National Academy of Sciences of the United States of America 109, 17472–17477. Zu, T., Gibbens, B., Doty, N.S., Gomes-Pereira, M., Huguet, A., Stone, M.D., Margolis, J., Peterson, M., Markowski, T.W., Ingram, M.A.C., et al. (2011). Non-ATG–initiated translation directed by microsatellite expansions. Proceedings of the National Academy of Sciences of the United States of America 108, 260–265. 147 Appendices Appendix A: Methods and Materials A.1 Methods iPSC reprogramming Human lymphocytes from healthy subjects and ALS patients were obtained from the NINDS Biorepository at the Coriell Institute for Medical Research and reprogrammed into iPSCs as previously described (Shi et al., 2018a; Shi et al., 2018b). Molecular cloning and viral production The complementary DNA (cDNA) for each iMN factor (Ngn2, Lhx3, Isl1, NeuroD1, Ascl1, Brn2 and Myt1l), each iMG factor (Pu.1, Cebpa, Mef2c, Atf3, Mafb, Irf5 and Ifr8) and mRFP-GFP-LC3 construct was purchased from Addgene and cloned into the pMXs retroviral expression vector using Gateway cloning technology (Invitrogen). The Hb9::RFP lentiviral vector was also purchased from Addgene (ID: 37081). Viruses were produced as follows. HEK293T cells were transfected in a 10-cm dish at 80-90% confluence with viral vectors containing each iMN factor and viral packaging plasmids (PIK-MLV-gp and pHDM for retrovirus, pPAX2 and VSVG for lentivirus) using polyethylenimine (PEI) (Sigma-Aldrich). The medium was changed 24 hours after transfection. Viruses were harvested at 48 hours and 72 hours after transfection. Viral supernatants were filtered with 0.45-μm filters, incubated with Lenti-X concentrator (Clontech) for 24 hours at 4°C, and centrifuged at 1,500 g at 4°C for 45 minutes. Pellets were resuspended in DMEM plus 10% FBS (200 μL per 10-cm dish of HEK293) and stored at –80°C. Conversion of iPSCs into induced motor neurons iPSCs were first differentiated into fibroblast-like cells to enable efficient retroviral transduction as described (Shi et al., 2019). Reprogramming of fibroblast-like cells was performed in 96-well plates (5 x 10 3 cells/ well) or 13-mm plastic coverslip (3 x 10 4 cells/ coverslip) that had been pre- coated with 0.1% gelatin (1 hour, room temperature) and laminin (4°C, overnight). Seven iMN factors were added in 100 μL of fibroblast medium (DMEM plus 10% FBS) per 96-well or 500 μL per coverslip with 5 μg/ml polybrene. Cultures were transduced with Hb9::RFP lentivirus after 48 hours transduction with seven iMN factors. On day 5, primary mouse cortical glial cells from P1 ICR pups were added to the transduced cultures in glia medium containing MEM (Life Technologies), 10% donor equine serum (HyClone), 20% glucose (Sigma-Aldrich), and 1% penicillin/streptomycin. On day 6, cultures were switched to N3 medium containing DMEM/F12 (Life Technologies), 2% FBS, 1% penicillin/streptomycin, N2 and B27 supplements (Life Technologies), 7.5 μM RepSox (Selleck), and 10 ng/ml each of FGF, GDNF, BDNF, and CNTF (R&D). The cultures were maintained in N3 with neurotrophic factors (RepSox, FGF, GDNF, BDNF, and CNTF) and changed every other day. Conversion of iPSCs into induced microglia and FACS sorting of CD11B+ microglia Reprogramming of fibroblast-like cells was performed in 6-well plates (1 x 10 5 cells/ well) or 13- mm plastic coverslip (3 x 10 4 cells/ coverslip) that had been pre-coated with 0.1% gelatin (1 hour, room temperature) and matrigel (1 hour, 37°C). Twi iMG factors were added in 100 μL of fibroblast medium (DMEM plus 10% FBS) per 96-well or 500 μL per coverslip with 5 μg/ml polybrene. Cultures were switched to microglia medium containing DMEM, 5% FBS and 10 ng/ml of human M-CSF (Peprotech) and IL-34 (Biolegend) after 24 hours transduction with two iMG factors. The cultures were maintained in microglia medium with and changed every three days. On day 14, cells were dissociated with Accutase (Innovative Cell Technologies), washed with wash buffer (3% FBS in PBS), and incubated with CD11b antibodies (1:100, Rat, Bio-Rad) for 1 hour at 4°C 148 in dark with rotating. Cells were washed with wash buffer and then incubated with Alexa Flour 488 secondary antibody (1:200, Goat anti-Rat, Invitrogen) for 30 mins at 4°C in dark with rotating. Cells were washed with wash buffer, pelleted and resuspend with DMEM medium for FACS sorting. Sorting of cells for collection was performed on Aria I or Aria II (BD). Live single cells were identified by SSC and FSC gating. Non-fluorescent controls were used to identify fluorescent populations. The population of labeled cells (488+) were sorted and collected as CD11b+ iMG. FACS-purified iMG were cultured with microglia medium in gelatin- and Matrigel-coated wells. Medium were changed every three days. iMNs-iMG co-culture system On day 17 after transduction of iMN factors, the FACS-sorted iMG were added into the wells with Hb9::RFP + iMNs. The ratio of iMNs to iMG per well is 3 to 1. The co-cultures were switched to N3 medium supplemented with neurotrophic factors (RepSox, FGF, GDNF, BDNF, and CNTF) and microglial supplements (M-CSF and IL-34). Medium were changed every three days. For immunostaining and single-cell RNA seq experiments, iMNs and iMG were co-cultured for 3 days; while for the survival assay, iMNs and iMG were co-cultured for 21 days. Induced motor neuron survival assay Hb9::RFP + iMNs appear between day 13-16 after transduction of iMN factors. The iMN survival assay was initiated on day 17. Starting at Day 17, longitudinal tracking of iMNs was performed using Molecular Devices ImageExpress once every other day for 14-20 days. Tracking of neuronal survival was performed using SVcell 3.0 (DRVision Technologies) or ImageJ. Neurons were scored as dead when their soma was no longer detectable by RFP fluorescence. For neurotrophic factor withdrawal conditions, BDNF, GDNF, and CNTF were removed from the culture medium on day 17. For treatment with inactive or active 3K3A-APC, cells were treated with 10 nM inactive or active 3K3A-APC. For treatment with rapamycin, DMSO or 10 μM rapamycin (Sigma, R8781-200UL) after neurotrophic factor withdrawal or 12-hour glutamate treatment. For PAR1 and PAR2 antagonist treatment, cultures were co-treated with 3 μM PAR1 or PAR2 antagonists (Tocris) after neurotrophic factor withdrawal or 12-hour glutamate treatment. For PAR1, PAR2 and PAR3 ASO treatment, the cultures were pre-treated one time with 9 μM ASOs for 72 hours before withdrawal neurotrophic factor. For treatment with apilimod or ASR- 0149, cultures were treated with DMSO or 3 μM apilimod (Achemblock, O33822) after neurotrophic factor withdrawal. For treatment with GW4869 to block exocytosis, cultures were treated with 10 μM GW4869 (Cayman Chemical, 13127) along with DMSO or 3 μM apilimod after neurotrophic factor withdrawal. For PIKFYVE ASO treatment, the iMN cultures were pretreated one time with 10 μM ASOs for 48 hours before neurotrophic factor withdrawal on day 17. For Rab27a ASO treatment, the iMN cultures were pretreated one time with 9 μM ASOs for 24 hours before neurotrophic factor withdrawal (on day 17) and treated with DMSO or 3 μM apilimod starting at day 17. ASO gapmers were designed and produced by IDT. All treatments were maintained for the whole survival assay and the medium was changed every other day. Preparation of 3K3A-APC 3K3A-APC was produced as described (Williams et al., 2012b) and was a gift from ZZ Biotech. Human 3K3A-Protein C (3K3A-PC) zymogen was stably transfected into Chinese hamster ovary (CHO) cells that were grown in suspension in CD OptiCHO medium (Invitrogen) supplemented with 2 mM CaCl2, 10 μg/ml vitamin K, and 2 mM GlutaMAX (Invitrogen). After the required number of serial expansions of culture volumes using shake flask passages and WAVE culturing, the final stirred-tank bioreactor volume was either 200 l or 2500 l in which culture was performed for ten days at 35 degrees C. After 0.2-μm filtration, the clarified culture supernatant was subjected to purification using chromatography on a Q Sepharose fast flow ion exchange column and hydrophobic affinity column. The purified 3K3A-PC zymogen was then activated with recombinant 149 human thrombin (Recothrom, Zymogenetics) to generate 3K3A-Activated Protein C (3K3A-APC). Thrombin was removed from the reaction mixture using ion exchange chromatography with a UnoSphere S flow through resin (BioRad). The concentration of 3K3A-APC was adjusted using tangential flow filtration (TFF). Following the concentration step, Uno Sphere Q Ion Exchange flow through was used with buffer exchange and concentration adjustment to polish the preparation, prior to Planova filtration and addition of polysorbate-80. 3K3A-APC was characterized by SDS- PAGE under reduced and non-reduced conditions, SE-FPLC, amidolytic activity by a chromogenic assay, and optical density at 280 nm. Inactive 3K3A-APC was generated by heat- denaturing the APC. CRISPR/Cas9 genome editing of iPSCs CRISPR/Cas9-mediated genome editing was performed in human iPSCs as previously described, using Cas9 nuclease (Shi et al., 2018). Single guide RNAs (sgRNAs) targeting both sides of the C9ORF72 intronic hexanucleotide repeat expansion were designed (Key resource table). To generate isogenic control iPSCs by removing the repeat expansion, C9ORF72 ALS/FTD patient iPSCs were transfected with human codon-optimized Cas9 (Addgene ID: 31825), the appropriate gRNA constructs by nucleofection (Lonza) according to the manufacturer’s protocol and the homologous recombination donor vector. The surviving colonies were picked on day 7 after transfection and genotyped by PCR amplification and sequencing the targeted genomic site. Colonies showing removal of the repeat expansion were clonally purified on MEF feeders and the resulting colonies were verified by southern blotting. C9ORF72 Southern blotting A 241-bp digoxigenin (DIG)-labeled probe was generated from 100 ng control genomic DNA (gDNA) by PCR reaction using Q5 High-Fidelity DNA Polymerase (NEB) with primers shown in the Key resource table. Genomic DNA was harvested from iPSCs using cell lysis buffer (100 mM Tris-HCl pH 8.0, 50 mM EDTA, 1% w/v sodium dodecyl sulfate (SDS)) at 55 °C overnight and performing phenol:chloroform extraction. A total of 25 µg of gDNA was digested with AflII at 37 ºC overnight, run on a 0.8% agarose gel, then transferred to a positive charged nylon membrane (Roche) using suction by vacuum and UV-crosslinked at 120 mJ. The membrane was pre- hybridized in 25 ml DIG EasyHyb solution (Roche) for 3 h at 47 ºC then hybridized at 47 ºC overnight in a shaking incubator, followed by two 5-min washes each in 2X Standard Sodium Citrate (SSC) and in 0.1% SDS at room temperature, and two 15-min washes in 0.1x SSC and in 0.1% SDS at 68 ºC. Detection of the hybridized probe DNA was carried out as described in DIG System User’s Guide. CDP-Star® Chemilumnescent Substrate (Sigma-Aldrich) was used for detection and the signal was developed on X-ray film (Genesee Scientific) after 20 to 40 min. Retinoic acid/purmorphamine protocol for iPSC-motor neuron differentiation Directed differentiation of iPSC motor neurons were generated for large-scale exosome preparations as previously described with slight modifications (Du et al., 2015). On day 0, iPSCs were dissociated with Accutase (Life Technologies) into single cell suspension and 300,000 iPSCs were seeded into one well of six-well plate (pre-coated with Matrigel (Corning)) in mTeSR medium (Stem Cell Technologies) with 10 μM Rock Inhibitor (Selleck). On day 1, cultures were switched to Neural Differentiation Medium (NDM) consisting of a 1 to 1 ratio of DMEM/F12 (Corning) and Neurobasal medium (Life Technologies), 0.5x N2 (LifeTechnologies), 0.5x B27 (Life Technologies), 0.1 mM ascorbic acid (Sigma), 1x Glutamax (Life Technologies). 3 μM CHIR99021 (Cayman), 2 μM DMH1 (Selleck) and 2 μM SB431542 (Cayman) were also added. The medium was changed every other day. On day 7, cells were dissociated with Accutase and 4-6 million cells were seeded in a 10-cm dish (pre-coated with Matrigel) in NDM plus 1 μM CHIR99021, 2 μM DMH1, 2 μM SB431542, 0.1 μM Retinoic acid (Sigma), 0.5 μM Purmorphamine (Cayman) and 10 μM Rock Inhibitor. Rock inhibitor was removed on day 9 and the medium was 150 changed every other day. On day 13, cells were dissociated with Accutase and 20 million cells were seeded per well in non-adhesive 6-well plates (Corning) in NDM plus 1 μM Retinoic acid, 1 μM Purmorphamine, 0.1 μM Compound E (Cayman), and 5 ng/ml each of BDNF, GDNF and CNTF (R&D Systems). Cells were used for experiments between days 25–35 of differentiation. The medium was changed every other day. Generation of Dox-NIL iMNs and Biotinylation of Surface-bound Glutamate Receptors A Dox-inducible NGN2, ISL1, LHX3 (NIL) polycistronic construct was previously integrated into the AAVS1 safe harbor locus of the C9-ALS patient iPSC line using CRISPR/Cas9 editin (Shi et al., 2018a). Dox-NIL iMNs were generated by plating at ~25% confluency on matrigel coated plates and adding 1 mg/ml of doxycylin in N3 media +7.5 μM RepSox 1 day after plating. Mouse primary mixed glia were added to the cultures at day 6, and doxycyline was maintained throughout conversion. iMN cultures were harvested at day 26. Biotinylation of plasma membrane localized glutamate receptors was performed using the Piece TM Cell Surface Protein Isolation Kit (Thermo Fisher Scientific). Briefly, Dox-NIL iMNs were incubated with 0.25mg/ml Sulfo-NHS-SS-Biotin in cold room for 1-2 hrs with end-to-end shaking. After quenching, cells were harvested by scraping and lysed with lysis buffer from the Pierce TM Cell Surface Protein Isolation Kit or the M-PER TM mammilian protein extraction buffer (Thermo Fisher Scientific). Cell lysate was incubated with High Capacity NeutrAvidin TM agarose beads (Thermo Fisher Scientific), and the bound protein was eluted in 2X SDS-PAGE sample buffer supplemented with 50mM DTT for 1 hr at room temperature with end-to-end rotation. Isolation of extracellular vesicles with polyethylene glycol iPSC-derived motor neurons were cultured in non-adhesive 6-well plates. The conditioned medium was harvested 24 hours after DMSO, 3 μM apilimod, 3 μM apilimod with 10 μM GW4869, or 10 μM GW4869 treatment. The harvested medium was centrifuged at 200g for 5 minutes at room temperature to remove cell pellets. The supernatants were collected and centrifuged at 2,000g for 20 minutes at 4°C to remove remaining cell debris (Ludwig et al., 2018). The supernatants were collected and centrifuged at 10,000g for 30 minutes at 4°C to remove apoptotic bodies. The obtained supernatants were supplemented with 50% w/v stock solutions of PEG 6000 (Sigma-Aldrich, 81253) and with 1M NaCl to final concentration of 8% PEG and 0.5M NaCl. Samples were mixed gently and incubated at 4°C overnight with rotating. Extracellular vesicles were concentrated by centrifuge at 3,000g at 4°C for an hour. Supernatants were removed and pellets were suspended with PBS. Extracellular vesicle samples were kept in -80°C for long-term storage. Conversion of iPSCs into microglia Microglia were generated from human iPSCs as previously described with slight modifications (Haenseler et al., 2017). In brief, 4.5 x 10 6 iPSCs were seeded into a well of an Aggrewell 800 (STEMCELL Technologies) to form embryoid bodies (EBs) in EB medium consisting of mTeSR1 plus BMP4 (Gibco), 50 ng/mL VEGF (Gibco), 20 ng/mL SCF (Miltenyi Biotec), and 10 uM Rock inhibitor (Selleck). A 75% EB medium change was performed daily for 3 days without Rock inhibitor. On Day 4, EBs were transferred to a low-attachment 6-well plate (Corning) and left undisturbed for 3 days. On Day 7, EBs were transferred to a T75 flask pre-coated with growth factor reduced Matrigel (Corning) in X-VIVO15 (Lonza), supplemented with 100 ng/mL M-CSF (Biolegend), 25 ng/mL IL-3 (Biolegend), 1x Glutamax (Life Technologies), 100 U/mL penicillin and 100 μg/mL streptomycin (Gibco), and 55 uM 2-mercaptoethanol (Gibco) and left undisturbed for one week, after which fresh medium added weekly (without replacement). Macrophage precursors began to emerge about 4 weeks, after which cells were collected weekly and flasks replaced with fresh medium. Harvested cells were plated onto plastic coverslips (Thermo Scientific) at 100,000 per cm 2 and differentiated into microglia-like cells for 10 days or more in 151 differentiation medium consisting of DMEM/F12 (Corning), 1x N2 (Life Technologies), 100 U/mL penicillin and 100 μg/mL streptomycin (Gibco), 50 uM 2-mercaptoethanol (Gibco), 1x Glutamax (Life Technologies), and 100 ng/mL IL-34 (Biolegend). Isolation of nuclear and cytoplasmic cell fractions for immunoblotting Pellet samples treated with DMSO or 3 μM apilimod from control and C9ORF72 ALS/FTD patient MNs were collected. The nuclear and cytoplasmic cell fractions were prepared using the NE- PER™ Nuclear and Cytoplasmic Extraction Reagents (Thermo Scientific™, cat. 78833) according to the manufacturer's instructions. Briefly, the cell pellets were washed once with PBS and centrifuged at 500g for 3 minutes. The cell pellet was suspended with 200 μl of cytoplasmic extraction reagent I (CER-I) by vortexing for 15 seconds. The suspension was incubated for 10 minutes on ice. 11 μl of cytoplasmic extraction reagent II (CER-I) was added to the suspension and the samples were vortexed for 5 seconds, incubated for 1 minute on ice, vortexed for another 5 seconds, and centrifuged at 16,000g for 5 minutes. The supernatant fraction (cytoplasmic extract) was transferred to a new tube and stored. The insoluble (pellet) fraction was re- suspended with 100 μl of nuclear extraction reagent (NER) and subjected to 4 cycles of vortexing for 15 seconds and incubation on ice for 10 minutes. Then samples were centrifuged at 16,000g for 10 minutes. The supernatant fraction (nuclear extract) was transferred to a new tube and stored. Correlative Light Electron Microscopy (CL-EM) C9-ALS1 iMNs were converted on 15 mm Nunc™ Thermanox™ Coverslips (Thermo Scientific™, 174969) pre-coated with 0.1% gelatin (1 hour, room temperature) and laminin (4°C, overnight). iMNs were treated with 3 μM apilimod for 24 hours and fixed with 4% paraformaldehyde (PFA) in 0.25 M HEPES for 1 hour at room temperature, washed with 0.1 M HEPES for three times, permeabilized with 0.1% Triton-X in 0.1 M HEPES for 10 mins at room temperature, blocked with 10% donkey serum in 1% BSA in 0.1 M HEPES for two hours at room temperature, and incubated with primary antibodies (diluted in 1% BSA in 0.1M HEPES) at 4°C overnight. The following antibodies were used: rabbit anti-TDP-43 Polyclonal antibody (Proteintech, 10782-2-AP, 1:200) and mouse anti-CD63 Monoclonal antibody (BD Transduction Laboratories, 556019, 1:100). Cells were then washed with 0.1M HEPES with 0.1% Triton-X for three times and incubated with Alexa Fluor–conjugated secondary antibodies (Life Technologies) (diluted in 1% BSA in 0.1M HEPES) for 2 hours at room temperature. iMNs were stained with DAPI (Life Technologies) for 10 minutes at room temperature and then mounted on slides with Vectashield (Vector Labs). Neuronal area was determined on the basis of the Hb9::RFP signals. High magnification images and z-stack images (fluorescence images) were acquired on an LSM 800 confocal microscope with oil immersion at 63x (Zeiss). Low magnification images (10–20X) were acquired with light microscopy to locate the cell of interest. Cells were fixed with glutaraldehyde and osmium tetroxide, stained with uranyl acetate, dehydrated through graded ethanol, infiltrated with epoxy resin. After polymerizing the block for 18 hours, the cells were separated from the Thermonox coverslip to the block, and sectioned into 500 nm sections. The sections were mounted on TEM grid and examined with a transmission electron microscope. Phagocytosis assay iMG were cultured in gelatin- and Matrigel-coated 96 wells and were incubated in 10 µg of pHrodo® Red E. coli Bioparticles® for Incucyte® (Sartorius Corporation, cat no. 4615) for 15 mins at 37°C. Samples were then washed with PBS and imaged with Molecular Devices ImageExpress 2, 24, 48 and 96 hours after adding pHrodo beads. Cytokine profiling 152 For LPS stimulation conditions, iMG or iPSCs culture media wer collected after 24 hours of vehicle or 100 ng/mL of LPS treatment. For iMNs-iMG co-cultures, conditioned media were collected from each group after 3 days of co-culturing. The Human Cytokine 42-Plex Discovery Assay was conducted by Eve Technologies Corp to access the cytokine secretion in each sample. Negative staining transmission electron microscopy Exosomes were fixed with 4% formaldehyde in PBS. 200-mesh, formvar-filmed copper grids were plasma-treated (Harrick PDC-001; Harrick Plasma, Ithaca, NY) and immediately incubated on a drop of exosomal vesicles for ten minutes. The grid was blotted and immediately placed on a drop of 1% phosphotungstic acid at pH7 for one minute, blotted again and air dried. The samples were imaged in a Zeiss EM10 transmission electron microscope (Zeiss, White Plains, NY) at 80 kV using a Gatan Erlangshen CCD camera (Gatan, Pleasanton, CA). Mass spectrometry analysis of exosomes 10 µg of each sample was reduced for 10 minutes at 50°C with 10 mM DTT and alkylated for 30 minutes at room temperature in the dark with 15 mM iodoacetamide. Proteins were precipitated with 8 volumes of ice-cold acetone and 1 volume of ice-cold methanol 2h at -80°C. The pellets were washed 3 times with 250 µl of cold methanol. Protein pellets were resolubilized in 100 µl of 50 mM Tris pH 8.0 and pre-digested with 0.3 µg of Trypsin/LysC for 3 hours at 37°C with agitation. Another 0.3 µg of Trypsin/LysC was added to the proteins and the digestion was continued overnight at 37°C with agitation. Samples were then acidified with 2% formic acid and the peptides were purified by Strata-X reversed phase SPE (Phenomenex). Acquisition was performed with a ABSciex TripleTOF 6600 (ABSciex, Foster City, CA, USA) equipped with an electrospray interface with a 25 μm iD capillary and coupled to an Eksigent μUHPLC (Eksigent, Redwood City, CA, USA). Analyst TF 1.8 software was used to control the instrument and for data processing and acquisition. The samples were analyzed in SWATH acquisition mode (data independent acquisition). The source voltage was set to 5.5 kV and maintained at 325oC, curtain gas was set at 45 psi, gas one at 25 psi and gas two at 25 psi. Separation was performed on a reversed phase Kinetex XB column 0.3 μm i.d., 2.6 μm particles, 150mm long (Phenomenex) which was maintained at 60oC. Samples were injected by loop overfilling into a 5μL loop. For the 45 minutes LC gradient, the mobile phase consisted of the following solvent A (0.2% v/v formic acid and 3% DMSO v/v in water) and solvent B (0.2% v/v formic acid and 3% DMSO in EtOH) at a flow rate of 3 μL/min. SWATH samples were processed using a publicly available ion library (SWATH Atlas). Retention times in the library were recalibrated using known abundant proteins and keratins. Proteins were quantified using a maximum of 10 peptides per protein and 4 MS/MS transitions per peptide. The quantification for a protein represents the sum of the area under the curve (AUC) of all the integrated peptides (max 10) for this protein. The intensity of each sample was normalized against every other sample using the total signal. GO analyses have been done using the Reactome FI plugin of Cytoscape, with only the significant genes. Confocal imaging Confocal microscopy images were acquired using a Zeiss LSM800 microcope. Quantification of Glutamate Receptors For neurites, Image J was used to automate detection of NR1 + punctae in a given are using a threshold of 30. Repeat primed PCR (RP-PCR) 153 Genomic DNA was isolated with Qiagen DNeasy Blood & Tissue Kit (cat 69504). 100 ng of genomic DNA was amplified by PCR using primers listed in the Key resource table in a 20 µl PCR reaction consisting of 0.25 mM each of 7-deaza-2-deoxyguanine triphosphate (deaza-dGTP) (NEB), dATP, dCTP and dTTP, 5% DMSO, 1x Qiagen buffer, 1x Taq polymerase (Roche), 1M betaine and 1 µM each of the three primers. During the PCR, the annealing temperature was gradually decreased from 70 °C and 56 °C in 2 °C increments with a 3-min extension time for each cycle. The PCR products were purified using the QiaQuick PCR purification kit (Qiagen) and analyzed using fragment analysis by Genewiz. Western blotting Cell lysates were collected in RIPA buffer (Sigma-Aldrich) with a protease inhibitor cocktail (Roche). Protein quantity from cell lysates and exosomal samples was measured by the BCA assay (Pierce). Samples were run on a 10% SDS gel. For Chapter 2, The gel was transferred onto an Immobilon membrane (Millipore). The membrane was blocked with 5% milk in 0.1% PBS- Tween 20 (PBS-T) (Sigma-Aldrich), incubated with primary antibodies overnight at 4 °C, washed three times with 0.1% PBS-T, then incubated with horseradish peroxidase (HRP)-conjugated (Santa Cruz). After three washes with 0.1% PBS-T, blots were visualized using an Amersham ECL Western Blotting Detection Kit (GE) or the SuperSignal West Femto Maximum Sensitivity Substrate (Thermo) and developed on X-ray film (Genesee). For Chapter 3 and 4, the gels were transferred onto Immobilon-FL PVDF Membrane (Millipore). The total protein for each sample was stained by Revert™ 700 Total Protein Stain Kits (LI-COR Biosciences) and quantified. The membrane was blocked with Intercept (TBS) Blocking buffer (LI-COR Biosciences), incubated with primary antibodies overnight at 4°C, washed three times with 0.1% TBS-T, and then incubated with IRDye® 680RD Donkey anti-Rabbit IgG Secondary Antibody or IRDye® 800RD Donkey anti-Mouse IgG Secondary Antibody (LI-COR Biosciences). After washing with 0.1% TBS-T for three times and TBS once, blots were scanned using LI-COR Odyssey CLx imaging system. The following primary antibodies were used: mouse anti-GAPDH (Santa Cruz, cat. no. sc-32233, 1:1000), mouse anti-NR1 (Novus, cat. no. NB300118, 1:1000), mouse anti-Transferrin receptor (Thermo, cat. no. 136800, 1:1000), mouse anti-TUJ1 (Biolegend, cat. no. MMS-435P, 1:2000), rabbit anti-poly(GR) (Proteintech, cat. no. 23978-1-AP, 1:1000), anti-mouse HRP (Cell Signaling, cat. no. 7076S, 1:5000), anti-rabbit HRP (Cell Signaling, cat. no. 7074S, 1:5000), anti-LC3 (VWR, cat. no. 101732-348, 1:1000), mouse anti-TSG101 (Biosciences, 612697, 1:500), rabbit anti-LC3 (MBL, PM036, 1:500), mouse anti-p62 (Biosciences, 610832, 1:1000), rabbit anti-OPTN Polyclonal antibody (Proteintech, 10837-1-AP, 1:500), rabbit anti-phospho-TDP43 (Ser409/410) Polyclonal antibody (Proteintech, 22309-1-AP, 1:500), Rabbit anti-PIP5K3 Polyclonal antibody (Proteintech, 13361-1-AP, 1:300), rabbit anti-TDP-43 Polyclonal antibody (Proteintech, 10782-2- AP, 1:500), rabbit anti-TFEB Polyclonal antibody (Bethyl Laboratories, A303-673A, 1:500), rabbit anti-phospho-TFEB (Ser142) Polyclonal antibody (Sigma-Aldrich, ABE1971-I-25UL, 1:500), rabbit anti-Huntingtin Monoclonal antibody (abcam, ab109115, 1:500), rabbit anti-alpha-Synuclein Polyclonal Antibody (Invitrogen, PA5-85791, 1:500), rabbit anti-Fibrillarin Monoclonal antibody (abcam, ab166630, 1:1000) and rabbit anti-HSP90 Polyclonal antibody (Proteintech, 13171-1-AP, 1:1000). Poly(GR) Immunoassay Poly(GR) levels in concentrated exosome and pellet samples from control and C9ORF72 ALS/FTD patient MNs were measured in a blinded manner using Meso Scale Discovery (MSD) platform based Poly(GR) sandwich immunoassay with custom made affinity purified rabbit polyclonal GR antibodies as previously described with minor modifications (Choi et al., 2019; Krishnan et al., 2022). Biotinylated poly(GR) antibodies at 0.5 ug/ml were coated on MSD Gold 96-well single spot streptavidin plates and incubated overnight at 4 oC. After washing and 154 blocking, concentrated exosome samples (40 ul) were loaded on plates in duplicate wells and incubated for 1.5 hrs at room temperature on a shaking platform. After 3 time washes, plates were loaded with 0.5 ug/ml MSD-Gold-Sulfo-tagged poly(GR) detection antibody and incubated for 1 hour at room temperature on a shaking platform followed by three final washes. After adding 1X MSD-Read buffer, plates were immediately read using MSD-QuickPlex SQ 120 reader and data presented as raw electrochemiluminescence (ECL) signals detected from the samples. Calcium Imaging of iMNs Calcium imaging was performed using Fluo-4-AM (Thermo, cat no. F14201) according to the manufacturer’s instructions. Cell cultures were treated with 1 μM cyclopiazonic acid for 30 minutes prior to the start of calcium imaging in order to deplete calcium stores from the endoplasmic reticulum and therefore enable more straightforward detection of calcium influx. At the start of the calcium imaging assay, day 17 iMN/mixed glia co-cultures were placed into N3 medium with an additional 1.5 μM glutamate and at least 3 fields per culture were imaged by time lapse for 30-60 seconds using a Nikon Ti inverted microscope. Calcium transients per iMN were quantified manually. Immunocytochemistry iMNs, iMG or iMN-iMG co-cultures were fixed in 4% paraformaldehyde (PFA) for 1 hour at 4°C, permeabilized with 0.1% Triton-X for 10 mins at room temperature, blocked with 10% donkey serum in 1% BSA in PBS for two hours at room temperature, incubated with primary antibodies at 4°C overnight. Cells were then washed with 0.1% PBS-T for three times and incubated with Alexa Fluor–conjugated secondary antibodies (Life Technologies) for 2 hours at room temperature. iMNs were stained with DAPI (Life Technologies) for 10 minutes at room temperature and then mounted on slides with Vectashield (Vector Labs). Neuronal area was determined by manual outlining in ImageJ on the basis of the Hb9::RFP signals. Images were acquired on an LSM 800 confocal microscope with oil immersion at 63x (Zeiss). The following primary antibodies were used: rabbit anti-GR repeat Polyclonal antibody (Proteintech, 23978-1- AP, 1:50), rabbit anti-PR repeat Polyclonal antibody (Proteintech, 23979-1-AP, 1:50), rabbit anti- TDP-43 Polyclonal antibody (Proteintech, 10782-2-AP, 1:200) and chicken anti-MAP2 antibody (Abcam, ab5392, 1:5000), rabbit anti-GA repeat Polyclonal antibody (Proteintech, 24492-1-AP, 1:50), rabbit anti-TFEB Polyclonal antibody (Bethyl Laboratories, A303-673A, 1:100), rabbit anti- LC3B Monoclonal antibody (Abcam, ab192890, 1:100), mouse anti-LAMP1 Monoclonal antibody (Abcam, ab25630, 1:20), mouse anti-CD63 Monoclonal antibody (BD Transduction Laboratories, 556019, 1:100), rabbit anti-RAB7 Monoclonal antibody (Abcam, ab137029, 1:100), rabbit anti- phospho TDP-43 (pS409/410) Polyclonal antibody (Cosmo Bio, CAC-TIP-PTD-P07, 1:200), goat anti-FUS Polyclonal antibody (Bethyl Laboratories, A303-839A, 1:200) and goat anti-IBA1 Polyclonal antibody (Abcam, ab5076, 1:100), anti-NR1 (Novus, NB300118, 1:50,), rat-anti CD11b Monoclonal antibody (Bio-Rad, MCA711G, 1:200), Rabbit-MAFB Polyclonal antibody (Sigma, HPA005653, 1:250), Mouse-CX3CR1 Polyclonal antibody (Abcam, ab167571, 1:100). The following dyes and conjugated antibodies were used: LysoSensor™ Green DND-189 (Invitrogen™, L7535, 5 µM), LysoTracker™ Deep Red (Invitrogen™, L12492, 100 nM), Alexa Fluor® 647 Anti-LC3B antibody (Abcam, ab225383, 1:100). To quantify the number of punctae in iMNs, the number of punctae per area (µm 2 ) was measured with Image J using the find maxima detection tool with noise tolerance=50-100 (depending on the antibodies). To quantify the TDP43, TFEB and FUS signals in iMNs, the fluorescence intensity of the TDP43, TFEB or FUS staining in the cell nucleus and cytoplasm were measured using Image J and recorded as the nuclear to cytoplasmic ratio. To quantify the number of colocalized punctae in MNs, the number of colocalized punctae per area (µm 2 ) was measured with Image J using the spots colocalization (ComDet) tool. Dipeptide repeat protein expression in iMNs and FACS analysis of GFP+/CFSE+ exosomes 155 iMNs were generated as described (Du et al., 2015). At day 25 of differentiation, we performed lentiviral transduction of GP(50)-GFP or GFP using previously-published constructs (Wen et al., 2014). 3 days after lentiviral transduction of 5-6 embryoid bodies per sample, iMNs were treated with DMSO or 3 µM apilimod for 24-hours and supernatant was harvested. 1 µl of CFSE dye (Thermo, catalog number: C34554) was added to the supernatant and incubated for 20 minutes at room temperature with gentle agitation. GFP+/CFSE+ exosomes were quantified using flow- cytometry gated for particles smaller than cells. mRFP-GFP-LC3 assay Retrovirus encoded by pMXs-mRFP-GFP-LC3 was transduced into iMN cultures 24 hours before transduction with iMN reprogramming factors. On day 17, iMNs were treated with DMSO or 3 μM apilimod for 24 hours and then fixed with 4% paraformaldehyde at 4°C for 1 hour. The cultures were immunostained with anti-MAP2 antibody to detect motor neurons. Coverslips were imaged on a Zeiss LSM 800 confocal microscope. Quantification was performed using ImageJ. Bulk RNA-seq and pathway analysis C9-ALS iMNs were cultured with 10 nM inactive 3K3A-APC or 3K3A-APC for 3 days and then Hb9::RFP+ iMNs were flow-purified. iPSC-derived motor neurons from C9-ALS/FTD patients (C9- ALS/FTD1) were cultured with DMSO or 3 μM apilimod for 24 hours and collected. iPSCs, secondary fibroblast, iMG and human primary microglia were cultured for 2 days and collected. Control, C9-ALS, C9ORF72-/- and C9 isogeneic iMG were generated and FACS-purifed CD11B+ iMG were collected. 3′-Digital gene expression RNA-Seq of all samples was performed by Amaryllis Nucleics. Briefly, mRNA was extracted using the NEBNext Poly(A) mRNA Magnetic Isolation Module according to the manufacturer’s instructions. 3′ RNA-Seq libraries were generated using the 3′-Digital Gene Expression RNAseq Library Kit (Amaryllis Nucleics). Libraries were sequenced on an Illumina NextSeq 500 sequencer. A total of 10 million to 25 million 80-bp, single-end reads were obtained for each sample. Reads were aligned to the Hg38 transcriptome using HISAT2 (Kim et al., 2015). A count table was obtained using FeatureCounts with strand specificity enabled. Differential expression analysis was performed using DESeq2 (Liao et al., 2014; Love et al., 2014). KEGG pathway enrichment analysis, transcription factor enrichment analysis (TFEA) and Bioplanet analysis were performed by providing Enrichr with a list of all genes identified by DESeq2 as having a greater than 95% chance of being differentially expressed (Kuleshov et al., 2016). To determine significant pathways, differentially expression genes for inactive 3K3A-APC vs. 3K3A-APC treatment at 3 days were uploaded an analyzed by using the Ingenuity Pathway Analysis (IPA) tool. We focused on “Top Canonical Pathways”. Only canonical pathways with - log(p-value) > 1.5 were retained. P values were calculated by using Fisher’s exact test. iPSC-derived motor neurons from C9-ALS/FTD patients (C9-ALS/FTD1 and C9-ALS/FTD3) were cultured with 10 μM NC ASO or PIKFYVE ASO for 96 hours and collected. mRNA sequencing was performed by Novogene Co., Ltd. (US). Briefly, total RNA was extracted using RNeasy Plus Mini Kit (Qiagen) according to manufacturer’s instructions. Total RNA was then shipped to Novogene Co., Ltd for poly(A) mRNA enrichment and library preparation. Libraries were sequenced on Illumina NovaSeq 6000. A total of 20 million to 40 million 150-bp, paired-end reads were obtained for samples sent. Reads were aligned to the Hg38 transcriptome using STAR (Dobin et al., 2013). Aligned reads were then quantified using Partek E/M with Partek® Flow® software, v10.0 (Partek Inc.). Gene counts were normalized using median ratio with offset by one before performing differential expression analysis using DESeq2 (Liao et al., 2019; Love et al., 2014). Single-cell RNA-sequencing and data analysis 156 After 3 days of co-culturing, Hb9::RFP+ iMNs and FITC-CD11b+ iMG were FACS-purified form the cultures. The barcoded single-cell cDNA libraries were synthesized using 10X Genomics Chromium Single Cell 3′ Library and Gel Bead Kit v.3.1 per the manufacturer’s instructions. Libraries were sequenced on Illumina HiSeq machines (Illumina, cat no. PE-410-1001 and FC- 410-1002) at a depth of at least 500,000 reads per cell for each library. Cellranger v3.1.0 (10X Genomics) was used for alignment against hg19 and gene-by-cell count matrices were generated with default parameters. The count matrices were analyzed by R package Seurat (v3.1.0) (Stuart et al., 2019). The pseudotime analysis were performed by R package Monocle 2 (Qiu et al., 2017), the ligand-receptor pairs were analyzed with R package iTalk (Wang et al.), the gene regulon network analysis were done with R package SCENIC (Aibar et al., 2017; Van de Sande et al., 2020). C. elegans paralysis assay on solid media Standard methods for culturing and handling the worms were used (Aggad et al., 2014; Therrien et al., 2013). Briefly, 40 age-synchronized L4 worms were transferred to NGM plates and scored daily for paralysis, from day 1 to day 12 of adulthood. Animals were counted as paralyzed if they failed to move upon prodding with a worm pick. Worms were scored as dead if they failed to move their head after being prodded on the nose and showed no pharyngeal pumping. All experiments were conducted at 20 C and in triplicate, three times. Some experiments were conducted by dissolving apilimod (1 M) into the NGM plates. C. elegans neurodegeneration assay For scoring of neuronal processes for gaps or breakage, worms were selected at day 9 of adulthood for visualization of motor neurons in vivo. Animals were immobilized in 5 mM levamisole dissolved in M9 and mounted on slides with 2% agarose pads. Visualization was done using a Zeiss Axio Imager M2 microscope, using a 10X objective and a 1.0 Optovar. The software used was Zen 2.3 Lite. At least one hundred worms were scored per condition, over 8 distinct experiments. Drosophila larval turning assay All stocks and crosses were maintained on standard yeast/cornmeal/ molasses food (25.0 °C using) a 12 hours light and dark cycle unless otherwise noted. The following stocks were used: (i) GAL4 motor neuron driver D42-GAL4 was used to drive the expression of UAS transgenes using the GAL4-UAS bipartite expression system, (ii) w 1118 as used as a genetic background control for TDP-43 transgenes, (iii) w 1118 ; UAS-TDP-43 G298S - YFP used to drive TDP-43 expression in motor neurons was previously described (Estes et al., 2013). (iv) y[1]v[1]; P{Y[+(7.7) = GSYP}attP2 (Bloomington Drosophila Stock Center (BDSC # 8622)) was used as genetic background control for PIKFYVE RNAi. y[1] sc[*] v[1]; P{y[+t7.7] v[+t1.8]=TRiP.GL00246}attP2, used to knock-down PIKFYVE by RNAi was obtained from the Bloomington Drosophila Stock Center (BDSC # 35793). To model C9ORF72 we used w1118; P{UAS-poly-GR.PO-100}attP40 (BDSC#58696) with y1 w1118; attP40 as genetic background control. RNAi experiments were performed with chronic RNAi expression and Pikfyve suppression. For apilimod experiments, larvae were fed yeast mixed with either solvent (DMSO) or apilimod (10 µM) for 5 days throughout larval development in a chronic treatment paradigm. Larval turning assays were performed using third instar larvae as previously described (Estes et al., 2011). Briefly, third instar larvae were placed on a Petri dish filled with a solidified grape juice/agar medium and allowed to accommodate for a few seconds. Larvae were gently turned ventral side up with a clean paintbrush and the time it took for each larva to turn ventral side down and make the first forward movement was recorded. 33 larvae per genotype, per condition were screened for YFP and evaluated for locomotor function. Larval turning data was analyzed with Graphpad Prism9 for statistical significance using Kruskal-Wallis with multiple comparisons. 157 Drosophila C9ORF72 GR.100 NMJ assay Drosophila larvae were reared and immunostained as described (Perry et al., 2017). Wild type (w 1118 ) or GR.100 larvae (w;OK6-Gal4/UAS-100xGR) were reared on individual apple plates at 25°C and fed yeast mixed with either solvent (DMSO) or apilimod (30 µM) for 5 days throughout larval development. Briefly, third-instar larvae were dissected in saline and immunostained with anti-BRP (nc82; Developmental Studies Hybridoma Bank; 1:100) and anti-HRP (Cyanine 3 (Cy3) conjugated; Jackson ImmunoResearch; 1:400). Alexa Fluor 488-conjugated mouse secondary antibody was used at 1:400 (Jackson ImmunoResearch). NMJs were imaged using a Nikon A1R Resonant Scanning Confocal microscope with a 100x APO 1.4 NA oil immersion objective. BRP puncta were quantified using Nikon Elements Software and data was analyzed using Graphpad Prism. NMDA/APC hippocampal administration and NMDA-lesion site analysis Mice were anesthetized with i.p. ketamine (100 mg/kg) and xylazine (10 mg/kg), and body temperature kept at 36.9 ± 0.1 °C with a thermostatic heating pad. Mice were placed in a stereotactic apparatus (ASI Instruments, USA) and the head is fixed accordingly. A burr hole was drilled, and an injection needle (33 gauge) was lowered into the hippocampus between CA1 and the dentate gyrus (AP −2.0, ML +1.5, DV −1.8). 10 nmoles of NMDA and 0.2 μg of 3K3A-APC or vehicle in 0.3 μl of phosphate-buffered saline, pH 7.4 was infused over 2 min using a micro- injection system (World Precision Instruments, Sarasota, FL, USA). For the DPR experiments, inactive 3K3A-APC was used as a control to control for any artifacts introduced into the downstream immunostaining measurements. After injection, the needle was left in place for an additional 8 min after the injection. Animals were euthanized 48h later. For injury size determination brains were quickly removed, frozen on dry ice, and stored at −80 °C until processing. 30-μm-thick coronal sections were prepared using a cryostat, and every fifth section 1 mm anterior and posterior to the site of injection was stained with cresyl violet. The lesion area was identified by the loss of staining, measured by ImageJ and integrated to obtain the volume of injury. TDP-43 mouse studies TDP-43 transgenic mice were purchased from the Jackson Laboratory (stock no: 012836). The mouse Thy1 promoter drives the expression of the human TDP-43 gene in neurons. TDP-43 Tg/Tg homozygous mutant mice were generated by crossing TDP-43 Tg/+ heterozygous mice with TDP- 43 Tg/+ heterozygous mice. Homozygous mutants were identified by a single band at 500 bp, heterozygotes were identified by bands at 303 bp and 500 bp, and wild types were identified by a single band at 303 bp. Homozygous mutants and their corresponding WT littermate controls were used in the experiments. For ASO treatments, 25 ug of negative control or Pikfyve ASO was administered by intracerebroventricular injection at P1. The ASOs were synthesized by Integrated DNA Technologies (IDT) and contained chemical modifications designed to bolster their efficacy for inhibition of gene expression. Phosphorothioate bonds were added to the sequence to provide protection from degradation by external nucleases and the modified base 2’-O-methoxy-ethyl (MOE) was also added to facilitate increased nuclease stability and binding affinity of the ASO to the mRNA target of interest. To enhance the purity of the ASOs for the in vivo studies, the ASOs were HPLC purified followed by a sodium/salt exchange. Finally, ethanol precipitation was performed to remove any residual chemical impurities. Briefly, P1 neonatal mice were anesthetized by placing them on a paper towel lined bucket of ice to induce hypothermia. The injection site targeting the lateral ventricles was defined as the distance of 40% between the lambda (anatomical landmark) and the eye. To perform the intracerebroventricular (ICV) injection, a 900 series Hamilton syringe (10 µL) was adapted with 158 the bottom portion of Neuros model 1701 RN syringe. A 33-gauge needle was inserted 2mM deep perpendicular to the skull surface into the left lateral ventricle. Mice received 3 µl of negative control ASO or Pikfyve ASO in PBS (a total of 25 µg). Within each litter half of the pups received negative control ASO and the other half received Pikfyve ASO. Mice were placed under a heat lamp in a new cage containing soiled bedding from the parental cage and were closely monitored until they had fully recovered. Beginning at P5, mice were administered vehicle (3.75% DMSO in 0.9% normal saline) or 6.7 µg/kg GW4869 by intraperitoneal (IP) injection every other day throughout the remainder of the experiment. For experiments using genetic deletion of Pikfyve, B6.Cg-PIKFYVE<tm1.1Ashi>/J were purchased from the Jackson Laboratory (stock no: 029331). The Pikfyve-flox mice were established by inserting loxP sites flanking Pikfyve exon 6 (encoding the FYVE finger domain). Removal of the floxed sequence creates a null allele. Mice heterozygous for the floxed allele (Pikfyve +/flox ) were bred to germline expressing B6.C-Tg(CMV- Cre)1Cgn/J mice (Jackson lab stock no: 006054) to generate Pikfyve mice with a single allele knockout (Pikfyve +/- ). The Pikfyve +/- mice were crossed with TDP-43 Tg/+ to generate Pikfyve +/- ;TDP-43 Tg/+ . The resulting offspring were then crossed with TDP-43 Tg/+ mice to generate Pikfyve +/- ;TDP-43 Tg/Tg mice. Mice were genotyped using several primer pairs according to a previous study (Iknonomov et al, Journal of Biological Chemistry, 2011) and listed in the Key resource table: S0 and A1 generating products of 1232 bp, 1476 bp or 574 bp, for the wild type allele, the floxed allele or the knockout allele, respectively. S1 and A1, generating products of 1063 bp, 1308 bp or 405 bp, for the wild type, the floxed allele or the knockout alleles, respectively. S2 and A1, generating 663 bp product, 852 bp product or no product, for the wild type, the floxed allele or the knockout alleles, respectively; and S3 and A1, generating 208 bp product, 378 bp product or no product, for the wild type allele, the floxed allele or the knockout allele, respectively. Gait impairment, kyphosis, and tremor were examined starting at 14 days of age (first signs of abnormal movement) and continued until 22 days of age. Phenotype scoring was assessed according to a composite rubric previously reported (Becker et al., 2017). All investigators were blinded to the genotype and treatment of each experimental group. The humane euthanasia endpoint for the study occurred when the mouse was no longer able to right itself. Hydrogel, powdered pellets, and high caloric diet gel were placed on the bottom of each cage containing TDP-43 mutant mice. The animals were housed in cages under a temperature and humidity- controlled environment and subjected to a standard 12 h light/dark cycle with food and water available ad libitum. All procedures were approved by the Institutional Animal Care and Use Committee of the University of Southern California in accordance with guidelines of the National Institutes of Health. Cerebrospinal fluid (CSF) Collection P1 wild type C57BL/6J mice (Jackson laboratory stock number: 000664) were injected by ICV injection with negative control ASO (25 µg) or PIKFYVE ASO (25 µg). Beginning at P3, mice were administered vehicle (3.75% DMSO in 0.9% normal saline) or 6.7 µg/kg GW4869 by intraperitoneal (IP) injection every other day until P7. CSF was collected on P7, 3 hours after the final IP injection. The neonatal mice were placed on a bucket of ice lined with paper towels and were anesthetized by hypothermia. The skull was exposed by a large incision through the flap of skin on top of the scalp. The site of the lateral ventricles was defined as the distance of 40% between the lambda (anatomical landmark) and the eye. The 33-gauge needle was inserted 2 mm deep perpendicular to the skull surface and clear CSF fluid was aspirated. For the analysis of CSF by mass spectrometry after vehicle (5% Tween 80, 10% N-methyl-2 pyrrolidone, 85% 30 mM citrate buffer in water, pH 3.1) or apilimod treatment, 6-8 weeks-old female mice were anesthetized and fur was removed from the dorsal lumbar area, extending from L1 to the sacrum. Ophthalmic ointment was placed on the eyes to prevent drying of the cornea. The injection site was disinfected with 70% ethanol using a sterile swab. Using a fingernail, the groove between L5-L6 was located and the needle was inserted at 90º to the spine on the midline 159 and into the groove. The substance (vehicle or apilimod) was slowly injected and the needle left in place for 5-10 sec prior to withdrawal. 10 µL of vehicle or 50 µM apilimod using a 25 µL Hamilton syringe with a 30-32-gauge needle. For CSF collection, mice were anesthetized with tribromoethanol 4 hours after vehicle or apilimod administration. When the plane of anesthesia was reached, CSF was collected through the cisterna magna (~ 1~3 µL/ mouse). All samples were flash frozen on dry ice and then stored in -80ºC until shipped to the study sponsor. Mouse tissue collection Wildtype or TDP43 Tg/Tg mice (postnatal day1) were treated with 25 ug negative control ASO or Pikfyve ASO via intracerebroventricular injection at postnatal day 1. For tissue collection, P19- P21 mice were anesthetized and intracardiac perfusion was performed with cold PBS. Brains were separated into two hemispheres sagittally. One hemisphere was frozen at -80 ℃ for immunoblotting or RNA analysis. The other brain hemisphere and spinal cords were fixed in cold 4%PFA overnight prior to being cryoprotected in 30% sucrose in PBS+0.01% sodium azide in preparation for immunofluorescence analysis. For CSF collection, ASO injected mice were euthanized at P7. Genotyping To provide a quantitative measure of (GGGGCC) n hexanuceotide expansion in C9ORF72, 100 ng of genomic DNA was amplified by touchdown PCR using primers shown in Supplemental Data Table 2, in a 28-µl PCR reaction consisting of 0.2 mM each of 7-deaza-2-deoxyguanine triphosphate (deaza-dGTP) (NEB), dATP, dCTP and dTTP, 7% DMSO, 1X Q-Solution, 1X Taq PCR buffer (Roche), 0.9 mM MgCl 2, 0.7 µM reverse primer (four GGGGCC repeats with an anchor tail), 1.4 µM 6FAM-fluorescently labeled forward primer, and 1.4 µM anchor primer corresponding to the anchor tail of reverse primer (Supplemental Data Table 2). During the PCR, the annealing temperature was gradually decreased from 70 ºC and 56 ºC in 2 ºC increments with a 3 min extension time for each cycle. The PCR products were purified by QiaQuick PCR purification kit (Qiagen) and analyzed using an ABI3730 DNA Analyzer and Peak Scanner TM Software v1.0 (Life Technologies). To detect the presence of variants in other genes in sporadic ALS lines, whole genome sequencing was performed by New York Genome Center, or whole exome sequencing was performed at the Children’s Hospital Los Angeles Genomics Facility. The genome/exome sequences were aligned to hg19 by Novoalign. We used an in-house tool developed by Anton Valouev (USC) for the removal of duplicate reads. Variant calling occurred by the Genome Analysis Toolkit (Broad Institute). Histology Cryoprotected tissues were embedded in O.C.T compound (Fisher Scientific, catalog number 23- 730-571) and cryosectioned to 16 um slices using Leica CM3050S cryostat. Sections were stored at -80 ℃. For immunofluorescence staining with primary antibodies, slides were treated with citrate buffer (pH 6)/0.05% tween-20 for 20 minutes at 80 ℃ prior to antibody staining. All primary antibody stainings were done with 3 overnight incubations at 4 ℃. Secondary antibody incubation was done at room temperature in the dark for one hour. The list of antibodies were provided in the Key resource table. For NeuroTrace (Fluorescent Nissl Stain, Thermo Fisher Scientific, N21480) staining, the slides were rehydrated in 10x PBS for at least 40 minutes and washed in 1x PBS/0.1% Triton X-100 for 10 minutes. The slides were washed in 1x PBS 2 times for 5 minutes each and then incubated with diluted NeuroTrace (1:50 in PBS) in the dark for 20 minutes. Slides were washed for 10 minutes in PBS/0.1% Triton X-100, 2 x 5 minutes in PBS and 2 hours at room temperature or overnight at 4 ℃ in PBS prior to mounting. To quantify the ventral motor neurons in the lumbar spinal cord, 4-5 sections per animal were stained with NeuroTrace and 160 imaged using a Zeiss LSM 880 at 10x. Both lateral motor column and medial motor column motor neurons in each half spinal cord were quantified and the average total motor neuron counts per hemicord was recorded for individual animals. To quantify the mislocalization of TDP43 into the cytoplasm, the nuclear:cytoplasmic ratio was measured using a total TDP43 antibody (10782-2-AP, Proteintech, 1:400). The fluorescence intensity of the TDP43 staining in the cell nucleus and cytoplasm were measured using Image J and recorded as the nuclear to cytoplasmic ratio. Motor neurons in the ventral spinal cord were imaged using Zeiss LSM 880 with oil immersion at 63x. 15-25 neurons were quantified per animal. To quantify the phosphorylated and aggregated TDP43+ punctae, brain and spinal cord sections were stained with phospho(Ser403/404)-TDP43 antibody (Proteintech, 66079-1-lg, 1:100) or phospho(Ser409/410)-TDP43 antibody (Cosmo Bio, CAC-TIP-PTD-P07, 1:200) and imaged at 40x with either a Zeiss LSM800 (spinal cord) or LSM880 (brain). The number of phosphorylated- TDP43+ punctae per cytosolic area was measured with Image J using the find maxima detection tool with noise tolerance=70. Neurons and its cell area were identified using Tuj1 antibody staining (Neuromics, CH23005, 1:100) and cell nucleus was identified using DAPI. Only the cytosolic area in neurons were quantified. 15-25 neurons were quantified per animal. To quantify the number of colocalized punctaes, spinal cord sections were stained with rabbit anti-LAMP1 (Abcam, ab24170, 1:100), mouse anti-RAB7 (Santa Cruz, sc-376362, 1:100) and mouse anti-LC3B (GeneTex, GT3612, 1:100) and images at 40x or 64x with a Zeiss LSM800. The number of colocalized punctae per area (µm 2 ) was measured with Image J using the spots colocalization (ComDet) tool. For glial analysis in the spinal cord, numbers of GFAP+ astrocytes (Abcam, ab7260, 1:100) or IBA1+ microglia (GeneTex, GT101495, 1:100) in a 150μm x150μm area of the ventral horn were quantified manually in Image J. The cell density in each ventral horn section was recorded and 4-6 ventral horn sections were quantified for each animal. The average cell density from each animal was used for statistical analysis. AAV-(GR)100 mouse studies The AAV9-CMV-eGFP-rBG and AAV9-CMV-eGFP-hGR 100-rBG were prepared for generating viruses as previously described (Zhang et al., 2018). Intracerebroventricular injection of 2 μL (1 × 10 10 genomes μ l−1 ) of AAV-eGFP or AA-VeGFP-(GR)100 virus were performed in postnatal day1 (P1) C57BL/6J pups. For Chapter 3, postnatal day 4 (P4) mouse pups received 25 μg of NC or Pikfyve ASO by intracerebroventricular injection. For Chapter 4, postnatal day 4 (P4) mouse pups received 12.5 μg of NC or Csf1r ASO by intracerebroventricular injection. Mice were placed under a heat lamp in a new cage containing soiled bedding from the parental cage and were closely monitored until they had fully recovered. Hindlimb clasping test was examined stating 14 days of age (first signs of abnormal movement) and continued until 26 days of age. Mice were suspended by the tail and hindlimb clasping was observed for 30 seconds. Phenotypic scoring was assessed according to a composite rubric previously reported (Lieu et al., 2013). All investigators were blinded to the genotype and treatment of each experimental group. The humane euthanasia endpoint for the study occurred when the mouse was no longer able to right itself. The animals were housed in cages under a temperature and humidity-controlled environment and subjected to a standard 12 h light/dark cycle with food and water available ad libitum. All procedures were approved by the Institutional Animal Care and Use Committee of the University of Southern California in accordance with guidelines of the National Institutes of Health. The brain sections were stained with anti-GFP Polyclonal antibody (Aves Lab, GFP-1010, 1:100) and rabbit anti-GR repeat Polyclonal antibody (Proteintech, 23978-1-AP, 1:50). The spinal cord sections were stained with NeuroTrace (Fluorescent Nissl Stain, Thermo Fisher Scientific, N21480). 161 The plasma membrane proteins were extracted from NC or Csf1r ASO treated brain tissue using Plasma Membrane Protein Extraction Kit (Abcam, ab65400). The western blotting was performed on the plasma membrane fraction and cytoplasmic fraction with rabbit-anti-CSF1R Monoclonal antibody (Abcam, ab254357, 1:500) and rabbit-anti-IBA1 Polyclonal antibody (GeneTex, GT101495, 1:500). Adult intracranial transplantation iPSCs-derived microglia were FACS-purified for CSF1R-high and CSF1R-low population. The microglia were suspended in 10 μL of sterile 1x PBS. Direct single-side intracranial injections of microglia into cortex of adult C9BAC or C9orf72 -/+ ;C9BAC mice. Adult mice (10-month-old) were anesthetized under continuous isoflurane and secured to a stereotaxic frame. Before exposing the skull, the local anesthetic was applied to the head. Mice received 40,000 cells per mice at the injection site using a 30-guage needle attached to a 10 μL Hamilton syringe. Transplantation was performed at one side of the cortex at the coordinates of 2.06 mm anteroposterior, ± 1.75 mm mediolateral and -0.95 mm dorsoventral relative to bregma (Hasselmann et al., 2019). Animals were allowed to recover on heating pads before being placed in their home cages. Animals were perfused 3 days after transplantation with 1X PBS followed by 4% paraformaldehyde, entire brains and spinal cord were removed for immunohistochemistry. Brains were sectioned sagittally at the thickness of 40 μm. The brain sections were stained with rabbit anti-GP repeat Polyclonal antibody (Proteintech, 24494-1-AP, 1:50) and Tuj1 antibody staining (Neuromics, CH23005, 1:100). The spinal cord sections were stained with NeuroTrace (Fluorescent Nissl Stain, Thermo Fisher Scientific, N21480). PLX chow treatment in C9orf72-/+;C9BAC mouse model 9-month old C9orf72 -/+ ;C9BAC mice were treated with PLX3397 (400 mg/kg formulated in standard chow) (Sosna et al., 2018). Age- and gender-matched mice were fed with the same standard chow without PLX3397 (control chow). After 35 days of chow administration, the hanging wire test were performed and analyzed before euthanasia. Brains and spinal cord were collected for histology. The average amount of PLX3397 was 29.8 mg/kg for male and 32.6 mg/kg for female mice. The brain sections were stained with rabbit anti-GP repeat Polyclonal antibody (Proteintech, 24494-1-AP, 1:50), rabbit-anti-IBA1 Polyclonal antibody (GeneTex, GT101495, 1:100), and Tuj1 antibody staining (Neuromics, CH23005, 1:100). The spinal cord sections were stained with NeuroTrace (Fluorescent Nissl Stain, Thermo Fisher Scientific, N21480). A.2 Experimental Models Cell lines Human iPSC lines were generated by the Ichida lab from lymphoblastoid lines obtained from the NINDS Biorepository (Table S1). The iPSCs were maintained in mTeSR1 medium using a feeder- free culture protocol in six-well plates coated with growth factor-reduced Matrigel. iPSCs were cultured at 37 C and 5% CO2 with daily feeding of 2mL mTeSR per well. To ensure consistency, cultures were kept below 90% confluency. Passaging of iPSC colonies was carried out using Accutase (Innovative Cell Technologies) to the desired split ratio. C. elegans C. elegans mutants were developed and maintained by the Parker lab as previously described (Aggad et al., 2014). 162 Drosophila The wild-type, GR.100, and TDP-43 strains were obtained from the Bloomington Drosophila Stock Center and maintained in the Dickman and Zarnescu labs as described in the Methods. Mouse models FVB/NJ (Stock No: 001800), Wild-type C57BL6/J (strain: 000664), C9ORF72 KO (C57BL/6J- 3110043021Rikem5Lutzy/J, strain: 027068), C9-BAC (C57BL/6J-Tg(C9orf72_i3)112Lutzy/J, strain: 023099), and TAR4/4 (Stock No: 012836) mice were obtained from Jackson Laboratory. Mice were housed in standard conditions with food and water ad libitum in the conventional vivarium at the University of Southern California. All animal use and care were in accordance with local institution guidelines of the University of Southern California and the IACUC board of the University of Southern California with the protocol numbers 21099 and 11938. A.3 Statistical Analysis Analysis was performed with the statistical software package GraphPad Prism version 9.1.0 (216). For iMN survival assay, statistical analysis was performed using two-sided log-rank test to account for the events that did not occur (meaning that iMNs that did not degenerate at the end of the survival assay). For each line, iMNs were randomly selected and used to generate the Kaplan–Meier survival curves. For all experiments, the normal distribution of data sets was tested by the D’Agostino-Pearson omnibus normality test. Differences between two groups were analyzed using a 2-tailed Student’s t test, unless the data were non–normally distributed, in which case 2-sided Mann-Whitney testing was used. Differences between multiple groups were analyzed using Ordinary one-way ANOVA. with Tukey’s correction for all comparisons, unless the data were non–normally distributed for which nonparametric Kruskal-Wallis testing was used. Mean and standard error of the mean was used for normally distributed data sets, and the median and interquartile range were used for non–normally distributed data sets. Significance was assumed at p < 0.05, and * p<0.05, ** p<0.01, *** p<0.001, **** p<0.001. 163 Appendix B: Supplementary Figures for Chapter 2 Figure S2. 1 Identification of neurodegenerative phenotypes in sporadic ALS patient iMNs. (A) Sporadic ALS iPSC lines expressing markers of pluripotency including NANOG (green) and TRA-1-81 (red). Nuclei (blue) are labeled with Hoechst. Scale bars: 80 μm. (B) Repeat-primed PCR data used to detect the presence of the C9ORF72 repeat expansion. (C-D) Images (C) and quantification (D) of the density of iMNs generated from multiple iPSC lines at the start of the iMN survival assay. iMNs quantified from 3 biologically independent iMN conversions per line. Mean +/- s.d. 164 (E) Survival of iMNs from the same line at when cultured at high or low density. iMNs quantified from 3 biologically independent iMN conversions per line. Scale bars: 80 μm. Two-sided log-rank test using the entire survival time course. (F-G) Images (F) and quantification (G) of the number of TH+ neurons in iMN cultures. iMNs quantified from 3 biologically independent iMN conversions per line. Mean +/- s.d. Scale bars: 10 μm. (H-I) Images (H) and quantification (I) of the number of CTIP2+ neurons in iMN cultures. iMNs quantified from 3 biologically independent iMN conversions per line. Mean +/- s.d. Scale bars: 10 μm. (J) Representative images showing the number of controls, C9-ALS, or sporadic ALS iMNs over time during an iMN survival assay. Scale bars: 80 μm. (K) Hazard ratio (relative to control iMNs) of iMNs from three controls or two C9-ALS patients in excess glutamate. iMNs quantified from 3 biologically independent iMN conversions per line. Mean +/- s.e.m. Unpaired t-test. (L) Hazard ratio (relative to control iMNs) of iMNs from three controls or three C9-ALS patients in the neurotrophic factor withdrawal condition. iMNs quantified from 3 biologically independent iMN conversions per line. Mean +/- s.e.m. Unpaired t-test with Welch’s correction. (M) Hazard ratio (relative to control iMNs) of iMNs from three controls or six sporadic ALS patients in excess glutamate. iMNs quantified from 3 biologically independent iMN conversions per line. Mean +/- s.e.m. Unpaired t-test with Welch’s correction. (N) Hazard ratio (relative to control iMNs) of iMNs from three controls or five sporadic ALS patients in the neurotrophic factor withdrawal condition. iMNs quantified from 3 biologically independent iMN conversions per line. Mean +/- s.e.m. Unpaired t-test with Welch’s correction. (O-Q) Survival of iMNs in excess glutamate from two different iPSC lines derived from sporadic ALS patient 1 (O), 2 (P), or 3 (Q). iMNs quantified from 3 biologically independent iMN conversions per line. Scale bars: 80 μm. Two-sided log-rank test using the entire survival time course. The day of differentiation stated on each panel indicates the day of differentiation on which the experimental treatment or time course was initiated. 165 Figure S2. 2 C9ORF72 and sporadic ALS iMNs share autophagosome formation abnormalities that are rescued by 3K3A-APC. (A) Number of RFP+/GFP+ vesicles per iMN in C9-ALS iMNs treated with DMSO or 50 nM Bafilomycin for 24 hours. iMNs from 3 controls and 3 C9-ALS patients were quantified. n=30 iMNs per line per condition across 2 independent iMN conversions were quantified. Median +/- interquartile range. Two-tailed Mann-Whitney test. Bafilo = Bafilomycin. Grey circles represent individual iMNs. (B) Number of RFP+/GFP+ vesicles per iMN in sporadic ALS iMNs treated with 50 nM Bafilomycin and inactive 3K3A-APC or 3K3A-APC for 24 hours. iMNs from 5 sporadic ALS lines were quantified. n=30 iMNs (controls) or 12 iMNs (sporadic ALS patients) per line per condition across 2 independent iMN conversions were quantified. For control (CTRL) vs sALS1, 4, 5, 6, mean +/- s.e.m. and unpaired t-test. For CTRL vs sALS 3, median +/- interquartile range and two-tailed Mann-Whitney test. Bafilo = Bafilomycin. Grey circles represent individual iMNs. 166 (C, D) Immunoblot analysis (C) and quantification (D) of LC3-I and LC3-II in C9-ALS iPSC-derived motor neurons + 50 nM Bafilomycin + 10 nM inactive 3K3A-APC or 3K3A-APC for 48 hours. Samples from four independent motor neuron treatments are shown for each treatment condition. Mean +/- interquartile range. Two-tailed Mann-Whitney test. Grey circles represent individual samples. Samples for replicate 4 were run on the same gel as the other samples but were noncontiguous. The full unedited gels for (C) are shown. (E) The difference in gene expression between inactive 3K3A-APC- and 3K3A-APC-treated C9- ALS iMNs at day 3 of treatment. Genes are plotted based on their log2 fold change (x-axis) from InAPC- to APC- treated cells and the log10 of the p-value for the significance of this change as determined by DESeq2. RNA-seq data from two samples per condition were used for this analysis. (F) Heatmap showing log2 fold change of differentially expressed genes in 10 nM inactive 3K3A- APC- and 3K3A-APC-treated C9-ALS iMNs relative to the average of both at 3 days. RNA-seq data from two independent iMN conversions and treatments were analyzed. (G) Ingenuity Pathway Analysis of RNA-seq data of C9-ALS iMNs treated with 10 nM inactive or active 3K3A-APC for 3 days. (H) KEGG pathway analysis of RNA-seq data from C9-ALS iMNs treated with 10 nM inactive or active 3K3A-APC for 3 days. Two samples included per condition. (I) mRNA levels of ATG5 or ATG10 in C9-ALS iMNs treated with 10 nM inactive or active 3K3A- APC for 3 days. Data were derived from RNA-seq analyses. (J) mRFP-GFP-LC3 fluorescence in control or sporadic ALS iMNs treated with 50 nM Bafilomycin and 10 μM DMSO or rapamycin. Bafilo = Bafilomycin. Scale bars = 5 μm. sALS = sporadic ALS. Solid and dotted lines outline the cell body and nucleus, respectively. (K) Number of RFP+/GFP+ vesicles per iMN in control or sporadic ALS iMNs treated with 50 nM Bafilomycin and 10 μM DMSO or rapamycin for 24 hours. iMNs from 3 controls and 5 sporadic ALS patients were quantified. n=12 iMNs per line per condition across 2 independent iMN conversions were quantified. Each grey circle represents a single iMN. Median +/- interquartile range. Non-parametric Kruskal-Wallis testing. (L) Images showing the number of LAMP2+ lysosomes in control or C9-ALS iMNs treated with inactive 3K3A-APC or 3K3A-APC. White solid lines outline iMN cell bodies. Scale bars = 5 μm. (M) Images showing the number of LAMP2+ lysosomes in control, C9-ALS, or sporadic ALS MAP2+ iMNs treated with inactive 3K3A-APC or 3K3A-APC. White solid lines outline iMN cell bodies. Scale bars = 5 μm. (N-O) Western blots (N) and quantification (O) showing levels of C9ORF72 in C9-ALS iMNs treated with inactive or active 3K3A-APC for 3 days. iMNs from two independent conversions were tested. Grey circles represent individual samples. Mean +/- s.e.m. Two-tailed t-test. The full unedited gels for (N) are shown. (P) mRNA levels of C9ORF72 in Hb9::RFP+ flow-sorted C9-ALS patient iMNs. n=3 biological replicates of iMNs from one C9-ALS patient line per condition. Mean +/- s.d. One-way ANOVA. (Q) Number of LAMP2+ vesicles in sporadic ALS iMNs treated with 50 nM Bafilomycin and 10 nM inactive 3K3A-APC for 24 hours. iMNs from 3 controls and 6 sporadic ALS patients were quantified. n=33 iMNs per line per condition across 2 independent iMN conversions were quantified. Median +/- interquartile range and non-parametric Mann-Whitney testing for sALS 1-4 and mean +/- s.e.m. and unpaired t-test for sALS 5, 6. (R, S) Ratio of RFP+/GFP+ vesicles to RFP+-only vesicles in control, C9-ALS (R), or sporadic ALS (S) iMNs treated with 10 nM inactive 3K3A-APC or 3K3A-APC for 24 hours. iMNs from 3 controls, 3 C9-ALS, and 6 sporadic ALS patients were quantified. n=30 (controls and C9-ALS) or 12 (sporadic ALS) iMNs per line per condition across 2 independent iMN conversions were quantified. Each grey circle represents a single iMN. Mean +/- s.e.m. Two-tailed unpaired t-test. The day of differentiation stated on each panel indicates the day of differentiation on which the experimental treatment or time course was initiated. 167 Figure S2. 3 Rescue of autophagosome formation abnormalities by 3K3A-APC improves proteostasis. (A-D) Immunostaining (A) and quantification (B-D) to determine endogenous poly(PR)+ punctae in control or C9-ALS iMNs with 10 nM inactive 3K3A-APC or 3K3A-APC treatment for 6 days. Quantified values represent the number of nuclear poly(PR)+ punctae in n=30 iMNs per line per condition from two control or two C9-ALS patient lines. For each line, iMNs were quantified from two independent iMN conversions per line per condition. Median +/- interquartile range. Each grey circle represents the number of poly(PR)+ punctae/unit area in a single iMN. Non- parametric Mann-Whitney testing. Solid and dotted lines in (A) outline the cell body and nucleus, respectively. Scale bar = 5 μm. (E) Number of nuclear poly(GR)+ punctae in control iMNs treated with 10 nM inactive 3K3A-APC or 3K3A-APC for 6 days. Quantified values represent the number of nuclear poly(GR)+ punctae in n=30 iMNs per line per condition from two control lines. For each line, iMNs were quantified 168 from two independent iMN conversions per line per condition. Median +/- interquartile range. Each grey circle represents the number of poly(GR)+ punctae/unit area in a single iMN. Non-parametric Mann-Whitney testing. (F-H) Number of total (nuclear and cytoplasmic) poly(GR)+ punctae in control or C9-ALS iMNs treated with 10 nM inactive 3K3A-APC or 3K3A-APC for 6 days. Quantified values represent the number of total poly(GR)+ punctae in n=30 iMNs (controls) or 41-44 iMNs (C9-ALS) per line per condition from two control or C9-ALS lines. For each line, iMNs were quantified from two independent iMN conversions per line per condition. Median +/- interquartile range. Each grey circle represents the number of poly(GR)+ punctae/unit area in a single iMN. Non-parametric Mann-Whitney testing. (I) Ratio of nuclear:cytoplasmic poly(GR)+ punctae/unit area in C9-ALS iMNs treated with 10 nM inactive 3K3A-APC or 3K3A-APC for 6 days. Quantified values represent the number of nuclear:cytoplasmic poly(GR)+ punctae in n=30 iMNs (controls) or 41-44 iMNs (C9-ALS) per line per condition from two control or C9-ALS lines. For each line, iMNs were quantified from two independent iMN conversions per line per condition. Median +/- interquartile range. Each grey circle represents the ratio of nuclear:cytoplasmic poly(GR)+ punctae/unit area in a single iMN. Non-parametric Mann-Whitney testing. (J-L) Number of total (nuclear and cytoplasmic) poly(PR)+ punctae in control or C9-ALS iMNs treated with 10 nM inactive 3K3A-APC or 3K3A-APC for 6 days. Quantified values represent the number of total poly(GR)+ punctae in n=30 iMNs per line per condition from two control or C9- ALS lines. For each line, iMNs were quantified from two independent iMN conversions per line per condition. Median +/- interquartile range except for (K), which is mean +/- s.e.m. Each grey circle represents the number of poly(PR)+ punctae/unit area in a single iMN. Non-parametric Mann-Whitney testing except for (K), which was tested by unpaired t-test. (M) Ratio of nuclear:cytoplasmic poly(PR)+ punctae/unit area in C9-ALS iMNs treated with 10 nM inactive 3K3A-APC or 3K3A-APC for 6 days. Quantified values represent the number of nuclear:cytoplasmic poly(GR)+ punctae in n=30 iMNs per line per condition from two control or C9-ALS lines. For each line, iMNs were quantified from two independent iMN conversions per line per condition. Median +/- interquartile range. Each grey circle represents the ratio of nuclear:cytoplasmic poly(PR)+ punctae/unit area in a single iMN. Non-parametric Mann-Whitney testing. (N) qRT-PCR analysis of C9ORF72 variants 1 and 3 (repeat-expansion-containing transcripts) mRNA in C9-ALS patient motor neurons treated with 10 nM inactive 3K3A-APC or 3K3A-APC for 3 days. n=3 biological replicates of iMNs from one C9-ALS patient line per condition. Mean +/- s.e.m. Unpaired t-test. (O) RNA-seq data showing expression levels of genes associated with the integrated stress response in Hb9::RFP+ C9-ALS iMNs treated with 10 nM inactive 3K3A-APC or 3K3A-APC for the durations shown. n=3 biological replicates of iMNs from one C9-ALS patient line per condition. Mean +/- s.e.m. One-way ANOVA. (P) Relative total GFP intensity of GR(50)-GFP+ punctae in iMNs 6 days before degeneration when treated with 10 nM inactive 3K3A-APC or 3K3A-APC. n=20 iMNs quantified for each condition from 3 independent conversions. Each grey circle represents the total GFP intensity of GR(50)-GFP+ punctae in one iMN. Median +/- interquartile range. Mann-Whitney test. (Q) Relative size of GR(50)-GFP+ punctae in iMNs 6 days before degeneration when treated with 10 nM inactive 3K3A-APC or 3K3A-APC. n=20 iMNs quantified for each condition from 3 independent conversions. Each grey circle represents the size of GR(50)-GFP+ punctae in one iMN. Median +/- interquartile range. Mann-Whitney test. The day of differentiation stated on each panel indicates the day of differentiation on which the experimental treatment or time course was initiated. 169 Figure S2. 4 C9ORF72 and sporadic ALS iMNs have elevated glutamate receptor levels that are normalized by 3K3A-APC. (A) Confocal microscopy images of immunofluorescence shows NR1+ puncta on MAP2+ dendrites of iMNs treated with 10 nM inactive 3K3A-APC or 3K3A-APC for 6 days. Scale bar: 2 µm. This experiment was repeated 3 times with similar results. (B) Number of NR1+ punctae per unit area in control or C9-ALS iMNs. Each grey circle represents the number of NR1+ punctae per area unit on a single dendrite (one dendrite quantified per iMN). n=15 iMNs quantified per line per condition from two biologically independent iMN conversions from two control and two C9-ALS lines. Median +/- interquartile range. Non-parametric Kruskal- Wallis testing. Each grey circle represents a single iMN. 170 (C-H) Number of NR1+ punctae per unit area in control or sporadic ALS iMNs. Each grey circle represents the number of NR1+ punctae per area unit on a single neurite. n=15 (controls) or 15- 17 (sporadic ALS) iMNs quantified per line per condition from two biologically independent iMN conversions from two control or six individual sporadic ALS lines. Median +/- interquartile range. Non-parametric Kruskal-Wallis testing. Each grey circle represents a single iMN. (I) Western blot of the surface-bound and total fractions of iMNs to determine the amount of intracellular protein (TDP-43) contaminating the surface-bound protein fraction. (I) is the full unedited gel. (J) Western blot analysis of surface NR1 after surface protein biotinylation in C9-ALS iMNs generated with 3 factors (NGN2, ISL1, and LHX3) and treated with 10 nM inactive 3K3A-APC or 3K3A-APC for 6 days. TF = Transferrin. Samples from four independent iMN conversions and treatments were used. (K) Western blot analysis of total NR1 in C9-ALS iMNs generated with 3 factors (NGN2, ISL1, and LHX3) and treated with 10 nM inactive 3K3A-APC or 3K3A-APC for 6 days. Samples from three independent iMN conversions and treatments were used. (L) Western blot analysis of surface NR1 after surface protein biotinylation in sporadic ALS iMNs generated with 3 factors (NGN2, ISL1, and LHX3) and treated with 10 nM inactive 3K3A-APC or 3K3A-APC for 6 days. TF = Transferrin. Samples from four independent iMN conversions and treatments were used. (M) Western blot analysis of total NR1 in sporadic ALS iMNs generated with 3 factors (NGN2, ISL1, and LHX3) and treated with 10 nM inactive 3K3A-APC or 3K3A-APC for 6 days. Samples from four independent iMN conversions and treatments were used. (N) Western blot analysis of surface NR1 after surface protein biotinylation in control iMNs generated with 3 factors (NGN2, ISL1, and LHX3) and treated with 10 nM inactive 3K3A-APC or 3K3A-APC for 6 days. TF = Transferrin. Samples from four independent iMN conversions and treatments were used. (O) Western blot analysis of total NR1 in control iMNs generated with 3 factors (NGN2, ISL1, and LHX3) and treated with 10 nM inactive 3K3A-APC or 3K3A-APC for 6 days. Samples from four independent iMN conversions and treatments were used. (P) Immunoblotting analysis of surface NR1 after surface protein biotinylation in control iMNs (one control) generated with 3 factors (NGN2, ISL1, and LHX3) and treated with 10 nM inactive 3K3A- APC or 3K3A-APC for 6 days. TF = Transferrin. (Q) Quantification of NR1 immunoblotting from (P). n=4 biologically independent iMN conversions. Each grey circle represents an individual sample. Mean +/- s.e.m. Unpaired t-test. The ratio of surface to total Transferrin Receptor was used to normalize for the membrane protein extraction efficiency and TUJ1 was used to normalize for neuron number. (R-T) Immunoblotting analysis (R) and quantification (S, T) of total NR1 (S) and surface:total NR1 (T) in C9-ALS iMNs (one patient) generated with 3 factors (NGN2, ISL1, and LHX3) and treated with 10 nM inactive 3K3A-APC or 3K3A-APC for 6 days. TF = Transferrin. Mean +/- s.e.m. Unpaired t-test. (U-W) Immunoblotting analysis (U) and quantification (V, W) of total NR1 (V) and surface:total NR1 (W) in sporadic ALS iMNs (one patient) generated with 3 factors (NGN2, ISL1, and LHX3) and treated with 10 nM inactive 3K3A-APC or 3K3A-APC for 6 days. TF = Transferrin. Mean +/- s.e.m. Unpaired t-test. (X-Z) Immunoblotting analysis (X) and quantification (Y, Z) of total NR1 (Y) and surface:total NR1 (Z) in control iMNs (one control) generated with 3 factors (NGN2, ISL1, and LHX3) and treated with 10 nM inactive 3K3A-APC or 3K3A-APC for 6 days. TF = Transferrin. Mean +/- s.e.m. Unpaired t-test. The day of differentiation stated on each panel indicates the day of differentiation on which the experimental treatment or time course was initiated. (J-O) are the full unedited gels for Fig. 4F, H and Supplemental Fig. 4P, R, U, X. 171 Figure S2. 5 3K3A-APC rescues the survival of C9ORF72 ALS iMNs in a PAR1-dependent manner. (A) Hazard ratio (relative to control iMNs) of control (three controls) or SOD1A4V ALS (one patient) iMNs treated with inactive 3K3A-APC or 3K3A-APC in the excess glutamate condition. iMNs quantified from 3 biologically independent iMN conversions per line. Mean +/- s.e.m. One-way ANOVA. (B) Hazard ratio (relative to control iMNs) of C9-ALS iMNs from two patients treated with inactive or different concentrations of 3K3A-APC in the excess glutamate condition. iMNs quantified from 3 biologically independent iMN conversions per line. Mean +/- s.e.m. One-way ANOVA. (C) Survival of C9-ALS3 iMNs with excess glutamate with inactive 3K3A-APC or 3K3A-APC. iMNs quantified from 3 biologically independent iMN conversions per line. Two-sided log-rank test using the entire survival time course. (D) Hazard ratio (relative to control iMNs treated with inactive 3K3A-APC) of control iMNs from three controls treated with inactive 3K3A-APC or 3K3A-APC in the excess glutamate condition. 172 iMNs quantified from 3 biologically independent iMN conversions per line. Mean +/- s.e.m. Unpaired t-test. (E) RNA-seq data showing the number of gene counts for the F2R gene in flow-purified, Hb9::RFP+ C9-ALS iMNs. (F, G) Survival of C9-ALS iMNs in excess glutamate with or without 3 μM PAR1 antagonist treatment (F) or PAR2 antagonist treatment (G). n=90 iMNs per line per condition, iMNs from both lines shown in aggregate for clarity. iMNs quantified from 3 biologically independent iMN conversions per line. (H) Hazard ratio (relative to control iMNs) of C9-ALS iMNs from two patients treated with 3K3A- APC and DMSO or a PAR1 antagonist in the excess glutamate condition. iMNs quantified from 3 biologically independent iMN conversions per line. Mean +/- s.e.m. Unpaired t-test. Antag. = antagonist. (I) Hazard ratio (relative to control iMNs) of C9-ALS iMNs from two patients treated with 3K3A- APC and DMSO or a PAR2 antagonist in the excess glutamate condition. iMNs quantified from 3 biologically independent iMN conversions per line. Mean +/- s.e.m. Unpaired t-test. (J) qRT-PCR measuring PAR1 expression in C9-ALS iMNs from two patients treated with a scrambled ASO or a PAR1 ASO. iMNs quantified from 3 biologically independent iMN conversions per line. Mean +/- s.d. Unpaired t-test. Scram. = scrambled. (K) Hazard ratio (relative to control iMNs) of C9-ALS iMNs from two patients treated with 3K3A- APC and a scrambled ASO or a PAR1 ASO in the excess glutamate condition. iMNs quantified from 3 biologically independent iMN conversions per line. Mean +/- s.e.m. Unpaired t-test. Scram. = scrambled. (L) qRT-PCR measuring PAR2 expression in C9-ALS iMNs from two patients treated with a scrambled ASO or a PAR2 ASO. iMNs quantified from 3 biologically independent iMN conversions per line. Mean +/- s.d. Unpaired t-test. Scram. = scrambled. (M) Survival of iMNs from 2 C9-ALS lines in excess glutamate with 3K3-APC with or without 9 μM PAR2 ASO treatment. n=90 iMNs per line per condition, iMNs from both lines shown in aggregate for clarity. iMNs quantified from 3 biologically independent iMN conversions per line. Each trace includes neurons from at 2 donors with the specified genotype. Two-sided log-rank test using the entire survival time course. (N) Hazard ratio (relative to control iMNs) of C9-ALS iMNs from two patients treated with 3K3A- APC and a scrambled ASO or a PAR2 ASO in the excess glutamate condition. iMNs quantified from 3 biologically independent iMN conversions per line. Mean +/- s.e.m. Unpaired t-test. Scram. = scrambled. (O) qRT-PCR measuring PAR3 expression in C9-ALS iMNs from two patients treated with a scrambled ASO or a PAR3 ASO. iMNs quantified from 3 biologically independent iMN conversions per line. Mean +/- s.d. Unpaired t-test. Scram. = scrambled. (P) Survival of iMNs from 2 C9-ALS lines in excess glutamate with 3K3-APC with or without 9 μM PAR3 ASO treatment. n=90 iMNs per line per condition, iMNs from both lines shown in aggregate for clarity. iMNs quantified from 3 biologically independent iMN conversions per line. Each trace includes neurons from at 2 donors with the specified genotype. Two-sided log-rank test using the entire survival time course. (Q) Hazard ratio (relative to control iMNs) of C9-ALS iMNs from two patients treated with 3K3A- APC and a scrambled ASO or a PAR3 ASO in the excess glutamate condition. iMNs quantified from 3 biologically independent iMN conversions per line. Mean +/- s.e.m. Unpaired t-test. Scram. = scrambled. For A, B, D, H-K, L, N, O, and Q, grey circles represent individual samples. The day of differentiation stated on each panel indicates the day of differentiation on which the experimental treatment or time course was initiated. 173 Figure S2. 6 3K3A-APC rescues the survival of sporadic ALS iMNs in a PAR1-dependent manner. (A-E) Survival of sALS2 (A), sALS3 (B), sALS4 (C), sALS5 (D), or sALS6 (E) sporadic ALS iMNs with excess glutamate with inactive 3K3A-APC or 3K3A-APC. iMNs quantified from 3 biologically independent iMN conversions per line. Two-sided log-rank test using the entire survival time course. (F) Hazard ratio (relative to control iMNs treated with inactive 3K3A-APC) of sporadic ALS iMNs from six sporadic ALS patients treated with inactive 3K3A-APC or 3K3A-APC in the excess glutamate condition. iMNs quantified from 3 biologically independent iMN conversions per line. Mean +/- s.e.m. Unpaired t-test with Welch’s correction. Grey circles represent individual samples. (G-I) Survival of iMNs from a second clone of sALS1 (G), sALS2 (H), and sALS3 (I) in excess 174 glutamate with inactive 3K3A-APC or 3K3A-APC. iMNs quantified from 3 biologically independent iMN conversions per line. Two-sided log-rank test using the entire survival time course. (J) Hazard ratio (relative to control iMNs treated with inactive 3K3A-APC) of iMNs from clone 2 of three sporadic ALS patients treated with inactive 3K3A-APC or 3K3A-APC in the excess glutamate condition. iMNs quantified from 3 biologically independent iMN conversions per line. Mean +/- s.e.m. Unpaired t-test. Grey circles represent individual samples. (K-P) Survival of sALS1 (K), sALS2 (L), sALS3 (M), sALS4 (N), sALS5 (O), or sALS6 (P) sporadic ALS iMNs with excess glutamate with inactive 3K3A-APC and DMSO or a PAR1 antagonist. iMNs quantified from 3 biologically independent iMN conversions per line. Two-sided log-rank test using the entire survival time course. (Q-U) Survival of sALS2 (Q), sALS3 (R), sALS4 (S), sALS5 (T), or sALS6 (U) sporadic ALS iMNs with excess glutamate with 3K3A-APC and DMSO or a PAR1 antagonist. iMNs quantified from 3 biologically independent iMN conversions per line. Two-sided log-rank test using the entire survival time course. (V) Hazard ratio (relative to control iMNs) of iMNs from six sporadic ALS patients treated with 3K3A-APC and DMSO or a PAR1 antagonist in the excess glutamate condition. iMNs quantified from 3 biologically independent iMN conversions per line. Mean +/- s.e.m. Unpaired t-test with Welch’s correction. Grey circles represent individual samples. The day of differentiation stated on each panel indicates the day of differentiation on which the experimental treatment or time course was initiated. 175 Figure S2. 7 Rapamycin rescues the survival of iMNs from some sporadic ALS lines. (A-E) Survival of sALS1 (A), sALS2 (B), sALS4 (C), sALS5 (D), or sALS6 (E) sporadic ALS iMNs with excess glutamate with 10 μM DMSO or rapamycin. iMNs quantified from 3 biologically independent iMN conversions per line. Two-sided log-rank test using the entire survival time course. (F-J) Hazard ratio (relative to control iMNs treated with DMSO) of sporadic ALS iMNs from five sporadic ALS patients treated with DMSO or rapamycin in the excess glutamate condition. iMNs quantified from 3 biologically independent iMN conversions per line. Mean +/- s.d. Unpaired t-test. Grey circles represent individual samples. (G-J) Survival of iMNs from a second clone of sALS1 (G), sALS2 (H), and sALS3 (I) in excess glutamate with inactive 3K3A-APC or 3K3A-APC. iMNs quantified from 3 biologically independent iMN conversions per line. Two-sided log-rank test using the entire survival time course. (K) Survival of control iMNs (3 controls) with excess glutamate with 10 μM DMSO or rapamycin. iMNs quantified from 3 biologically independent iMN conversions per line. Two-sided log-rank test using the entire survival time course. n=90 iMNs per line. (L) Hazard ratio (relative to control iMNs treated with DMSO) of control iMNs (three controls) treated with DMSO or rapamycin in the excess glutamate condition. iMNs quantified from 3 biologically independent iMN conversions per line per condition. Mean +/- s.e.m. Unpaired t-test. Grey circles represent individual samples. The day of differentiation stated on each panel indicates the day of differentiation on which the experimental treatment or time course was initiated. 176 Appendix C: Supplementary Figures for Chapter 3 Figure S3. 1 PIKFYVE inhibition ameliorates C9ORF72 ALS/FTD neurodegeneration. (A) Immunostaining shows iMNs express the Hb9::RFP reporter (red) and microtubule-associated protein 2 (MAP2) (yellow). Nuclei (blue) are labeled with DAPI. Scale bar = 10 μm. (B) Images of the density of Hb9::RFP+ iMNs generated from CTRL, C9-ALS/FTD and sporadic ALS (sALS) lines at the start of the iMN survival assay. Scale bar = 100 μm. 177 (C) Immunostaining and quantification to determine endogenous poly(GR)+ punctae in CTRL or C9-ALS/FTD iMNs. Quantified values represent the average number of nuclear poly(GR)+ punctae per μm 2 in n=40 CTRL and n=40 C9-ALS/FTD iMNs from two CTRL and two C9- ALS/FTD patient lines (n=20 per line). iMNs were quantified from two independent conversions per group. Each gray circle represents the average number in a single iMN. Mann-Whitney test. Mean ± interquartile range. Solid and dotted lines outline the cell body and nucleus, respectively. Scale bar = 2 μm. (D) Immunostaining and quantification to determine endogenous poly(PR)+ punctae in CTRL or C9-ALS/FTD iMNs. Quantified values represent the average number of nuclear poly(PR)+ punctae per μm 2 in n=40 CTRL and n=40 C9-ALS/FTD iMNs from two CTRL and two C9- ALS/FTD patient lines (n=20 per line). iMNs were quantified from two independent conversions per group. Each gray circle represents the average number in a single iMN. Mann-Whitney test. Median ± interquartile range. Solid and dotted lines outline the cell body and nucleus, respectively. Scale bar = 2 μm. (E) The intronic hexanucleotide (GGGGCC) repeat expansion (red diamond) is located between the noncoding exon 1a and 1b (grey rectangles), upstream of the start codon (ATG) of the C9ORF72 gene (green rectangle). The C9ORF72 repeat expansion was removed by CRISPR- Cas9 genome editing using two guide RNAs flanking the expansion (each guide RNA is shown as a pair of scissors) to generate the corrected isogenic control. (F) Southern blot for examining C9ORF72 hexanucleotide repeat expansion in C9-ALS/FTD1 and its corrected isogenic line. The C9-ALS/FTD1 corrected isogenic line was generated by removing the hexanucleotide repeat expansion in the C9ORF72 mutant allele of C9-ALS/FTD1 (ND06769) line. The wild-type C9ORF72 allele forms a single 3.8 kb band on the blot, while the multiple bands with higher molecular weights in the C9-ALS/FTD1 line represent different lengths of repeat expansions. (G) Repeat-primed PCR (RP-PCR) for assessing the presence of the repeat expansion in C9- ALS/FTD1 and its corrected isogenic control line. (H) Quantification to determine endogenous poly(GR)+ and poly(PR)+ punctae in C9-ALS/FTD1 and its corrected isogenic control iMNs. Quantified values represent the average number of nuclear poly(GR)+ (left) or poly(PR)+ (right) punctae per μm 2 in n=20 C9-ALS/FTD1 and n=15 corrected isogenic control iMNs. iMNs were quantified from two independent conversions per group. Each gray circle represents the average number in a single iMN. Unpaired t-test. Mean ± s.e.m. (I) Immunostaining and quantification of total TDP-43 in CTRL and C9-ALS/FTD iMNs. Quantified values represent the average ratio of nuclear to cytoplasmic TDP-43 in n = 30 CTRL and n=30 C9-ALS/FTD iMNs from two CTRL and two C9-ALS/FTD patient lines. iMNs were quantified from two independent conversions per group. Each gray circle represents the ratio from a single iMN. Mann-Whitney test. Median ± interquartile range. Solid and dotted lines outline the cell body and nucleus, respectively. Scale bar = 2 μm. (J) Representative images show the number of CTRL, C9-ALS/FTD and sporadic ALS (sALS) iMNs over time during the iMN survival assay. Scale bar = 100 μm. (K) Kaplan–Meier survival curves of CTRL and C9-ALS/FTD iMNs after withdrawal of neurotrophic factor supplementation (shown for all lines in aggregate). n = 90 iMNs per line for three control and three C9-ALS/FTD lines. iMNs quantified from three biologically independent iMN conversions per line. Two-sided log-rank test. Statistical significance calculated using the entire survival course (days 1-20). (L) Representative images showing the number of C9-ALS/FTD iMNs over time when treated with DMSO, 3 μM apilimod (AP), 3 μM apilimod+100 nM GW4869 (AP+GW), or 100 nM GW4869 (GW) during the iMN survival assay. Scale bar = 50 μm. (M) Kaplan–Meier survival curves of C9-ALS/FTD iMNs in DMSO or 3 μM apilimod (AP) after withdrawal of neurotrophic factor supplementation (iMNs from all lines from each genotype shown 178 in aggregate). n=90 iMNs per line for three C9-ALS/FTD lines. iMNs quantified from three biologically independent iMN conversions per line. Two-sided log-rank test. Statistical significance calculated using the entire survival course (days 1-20). (N) Kaplan–Meier survival curves of iMNs from one C9-ALS donor (C9-ALS4) in DMSO, 3 nM, 300 nM and 3 μM AP after withdrawal of neurotrophic factor supplementation. n=100 iMNs for DMSO, n=49 iMNs for 3 nM AP, n=51 iMNs for 300 nM AP and n=42 iMNs for 3 μM AP. iMNs quantified from three biologically independent iMN conversions per line. Two-sided log-rank test. Statistical significance calculated using the entire survival course (days 1-14). (O) mRNA levels of PIKFYVE (relative to GAPDH) in HeLa cells treated with 50 nM negative control ASO (NC), PIKFYVE ASO1 or ASO2. n=3 independent ASO treatments per condition. Ordinary one-way ANOVA. Mean ± s.e.m. P-values were corrected for multiple comparisons using statistical hypothesis testing (Šídák method). (P) Kaplan–Meier survival curves of control or C9-ALS/FTD iMNs treated with 10 μM negative control (NC), PIKFYVE ASO1, or ASO2 after withdrawal of neurotrophic factor supplementation. n= 120 iMNs per line per condition for two CTRL and two C9-ALS/FTD lines (C9-ALS/FTD1 and C9-ALS/FTD2); n = 70 iMNs per treatment for one C9-ALS/FTD line (C9-ALS/FTD3). iMNs quantified from three biologically independent iMN conversions per condition. Two-sided log-rank test. Statistical significance calculated using the entire survival course (days 1-14). (Q) Kaplan–Meier survival curves of iMNs from one CTRL (CTRL3) and one C9-ALS/FTD (C9- ALS3) donor treated with 10 μM negative control (NC), 1 μM, 3 μM or 10 μM PIKFYVE ASO1 after withdrawal of neurotrophic factor supplementation. n = 70 iMNs per treatment per line. iMNs quantified from three biologically independent iMN conversions per condition. Two-sided log-rank test. Statistical significance calculated using the entire survival course (days 1-14). (R) The hazard ratios (Mantel–Haenszel method) of iMNs from one CTRL (CTRL3) and one C9- ALS/FTD (C9-ALS3) treated with 10 μM negative control (NC) ,1 μM, 3 μM or 10 μM PIKFYVE ASO1 after withdrawal of neurotrophic factor supplementation. The hazard ratio = the hazard rate of each group relative to the hazard rate of the CTRL line treated with NC ASO, which was set as 1 (red dotted line). The hazard ratio was derived from days 1-14 of iMN survival. Mean of biological replicates (independent conversions) ± s.e.m. n=3 independent conversions/line/condition. Statistical significance (one-way ANOVA) calculated comparing the NC ASO condition for the C9- ALS line to the conditions with different doses of PIKFYVE ASO1 and to the CTRL line treated with NC ASO group. P-values were corrected for multiple comparisons using statistical hypothesis testing (Šídák method). (S) Representative images of muscle and NMJs in wild type Drosophila larvae or larvae with C9ORF72 GR.100 overexpression in motor neurons. Larvae were exposed to a chronic 5-day treatment throughout larval development with DMSO (solvent) or 30 µM apilimod and immunostained with anti-BRP (active zone marker) and anti-HRP (neuronal membrane marker). Scale bar = 10 μm. (T) Quantification of active zone number at NMJs (labeled by BRP) in the indicated genotypes and conditions. Note the severe reduction in active zone number in C9ORF72 GR.100 larvae, which is significantly increased following a chronic 5-day application of apilimod throughout larval development. Two-tailed unpaired t-test for each comparison. Mean ± s.e.m. 179 Figure S3. 2 PIKFYVE inhibition ameliorates disease pathology in C9ORF72 ALS/FTD iMNs. (A) Transcription factor enrichment analysis (TFEA) and KEGG pathway analysis of differentially expressed genes from RNA-seq data of C9-ALS/FTD iPSC-derived motor neurons treated with DMSO or 3 μM apilimod (AP) for 24 hours. 180 (B) Quantification of normalized expression (counts from RNA-seq data) of STMN2 and UNC13A from C9-ALS/FTD iPSC-derived motor neurons treated with DMSO or 3 μM apilimod (AP) for 24 hours. Unpaired t-test. Mean ± s.e.m. (C) Immunostaining and quantification of endogenous poly(PR)+ punctae in CTRL or C9- ALS/FTD iMNs treated with DMSO or 3 μM apilimod (AP) for 24 hours. Quantified values represent the average number of nuclear poly(PR)+ punctae per μm 2 from iMNs from two CTRL and two C9-ALS/FTD patient lines (n = 40 iMNs per condition per line). iMNs were quantified from two independent conversions per group. Each gray circle represents the average number of punctae in a single iMN. Kruskal-Wallis test. Median ± interquartile range. P-values were corrected for multiple comparisons using statistical hypothesis testing (Dunn’s test). Solid and dotted lines outline the cell body and nucleus, respectively. Scale bar = 2 μm. (D) Immunostaining and quantification of endogenous poly(GA)+ punctae in CTRL or C9- ALS/FTD iMNs treated with DMSO or 3 μM apilimod (AP) for 24 hours. Quantified values represent the average number of nuclear poly(PR)+ punctae per μm 2 from iMNs from two CTRL and two C9-ALS/FTD patient lines (n = 40 iMNs per condition per line). iMNs were quantified from two independent conversions per group. Each gray circle represents the average number of punctae in a single iMN. Kruskal-Wallis test. Median ± interquartile range. P-values were corrected for multiple comparisons using statistical hypothesis testing (Dunn’s test). Solid and dotted lines outline the cell body and nucleus, respectively. Scale bar = 2 μm. 181 Figure S3. 3 PIKFYVE inhibition increases exocytosis of aggregation-prone proteins from iMNs. (A) Images of LC3-mRFP-GFP3 fluorescence in C9-ALS/FTD iMNs in DMSO or 3 μM apilimod (AP) with 50 nM bafilomycin, or DMSO or 3 μM apilimod (AP) without bafilomycin for 24 hours. Solid and dotted lines outline the cell body and nucleus, respectively. Scale bar = 2 μm. (B) Quantification of (A), DMSO or 3 μM apilimod (AP) with or without 50 nM bafilomycin conditions. Quantified values represent the average number of LC3-GFP+/mRFP+ punctae per μm 2 in n=30 iMNs for DMSO with bafilomycin, n=30 iMNs for AP with bafilomycin, n=45 iMNs for DMSO without bafilomycin and n=50 iMNs for AP without bafilomycin from two C9-ALS/FTD patient lines. iMNs were quantified from two independent conversions per group. Each gray circle 182 represents the ratio in a single iMN. Kruskal-Wallis test. Median ± interquartile range. P-values were corrected for multiple comparisons using statistical hypothesis testing (Dunn’s test). (C) Immunostaining and quantification of TFEB in CTRL and C9-ALS/FTD iMNs after withdrawal of neurotrophic factor supplementation for 96 hours and then treatment with DMSO or 3 μM apilimod (AP) treatment for 96 hours. Quantified values represent the mean intensity of nuclear TFEB (left panel) and the average ratio of nuclear to cytoplasmic TFEB (right panel) in n=30 CTRL and n=30 C9-ALS/FTD iMNs from two CTRL and two C9-ALS/FTD patient lines. iMNs were quantified from two independent conversions/group. Each gray circle represents the ratio from a single iMN. Kruskal-Wallis test. Median ± interquartile range. P-values were corrected for multiple comparisons using statistical hypothesis testing (Dunn’s test). Solid and dotted lines outline the cell body and nucleus, respectively. Scale bar = 5 μm. (D) Immunoblots and quantification of pellet fractions from CTRL and C9-ALS/FTD iPSC-derived motor neurons against phosphorylated (Ser142) TFEB (65kDa). Cells were treated with DMSO or 3 μM apilimod (AP) for 24 hours. n=8 biological replicates (independent conversions) per condition from three CTRL lines (n=3 independent conversions for CTRL1 and CTRL2, n=2 independent conversions for CTRL3) and n=8 biological replicates (independent conversions) per condition from three C9-ALS/FTD patients (n=3 independent conversions for C9-ALS/FTD1 and C9-ALS/FTD3, n=2 independent conversions for C9-ALS/FTD2). Values calculated as the relative intensity of phosphorylated (Ser142) TFEB (65kDa) in the cell pellet normalized to total protein. Ordinary one-way ANOVA. Mean ± s.e.m. P-values were corrected for multiple comparisons using statistical hypothesis testing (Šídák method). (E) Quantification of the no bafilomycin conditions from (A). Quantified values represent the ratio of LC3-mRFP+/GFP+ punctae to total LC3-mRFP+ (LC3-mRFP+/GFP+ and LC3-mRFP+/GFP-) punctae in n=45 iMNs for DMSO and n=50 iMNs for AP from two C9-ALS/FTD patient lines. iMNs were quantified from two independent conversions per group. Each gray circle represents the ratio in a single iMN. Mann-Whitney test. Median ± interquartile range. (F) Immunoblots and quantification of pellet fractions from CTRL and C9-ALS/FTD iPSC-derived motor neurons against LC3 (LC3-I above 15kDa and LC3-II below 15kDa). Cells were treated with DMSO or 3 μM apilimod (AP) for 24 hours. Samples were normalized to total protein in the cell pellet fraction and pellet TUJ1 (55kDa) served as a control for differences in cell death. n=6 biological replicates (independent conversions) per condition from two CTRL lines and n=9 biological replicates (independent conversions) per condition from three C9-ALS/FTD patients (n=3 independent conversions per line). The values were calculated as the ratio of LC3-II to LC3- I. Ordinary one-way ANOVA. Mean ± s.e.m. P-values were corrected for multiple comparisons using statistical hypothesis testing (Šídák method). (G) Quantification of the no bafilomycin conditions from (A). Quantified values represent the ratio of LC3-mRFP+/GFP- punctae per μm 2 in n=45 iMNs for DMSO and n=50 iMNs for AP from two C9-ALS/FTD patient lines. iMNs were quantified from two independent conversions per group. Each gray circle represents the ratio in a single iMN. Mann-Whitney test. Median ± interquartile range. (H) Staining and quantification of LysoSensor (Green DND-189) and LysoTracker in C9-ALS/FTD iMNs with DMSO or 3 μM apilimod (AP) treatment for 24 hours. Quantified values represent percentage of LysoSensor+ punctae to total LysoTracker+ punctae from two C9-ALS/FTD patient lines (n = 15 iMNs per condition per line). The LysoSensor+ punctae appear in acidic environments (pKa ∼5.2) and indicate the presence of acidic lysosomes. iMNs were quantified from two independent conversions per group. Each gray circle represents the average number in a single iMN. Kruskal-Wallis test. Median ± interquartile range. Solid and dotted lines outline the cell body and nucleus, respectively. Scale bar = 5 μm. (I) The hazard ratio (Mantel–Haenszel method) of two CTRL and two C9-ALS/FTD iMNs treated with DMSO, 3 μM apilimod (AP), DMSO+0.5 nM bafilomycin, or 3 μM apilimod (AP)+0.5 nM bafilomycin after withdrawal of neurotrophic factor supplementation. The hazard ratio = the hazard 183 rate of each group relative to the hazard rate of the two CTRL lines in aggregate treated with DMSO, which was set as 1 (red dotted line). The hazard ratio was derived from days 1-20 of iMN survival. For CTRL, n=180 iMNs for DMSO, n=180 iMNs for AP, n=260 iMNs for DMSO+bafilomycin, n=232 iMNs for AP +bafilomycin. For C9-ALS/FTD, n=180 iMNs for DMSO, n=180 iMNs for AP, n=254 iMNs for DMSO+bafilomycin, and n=211 iMNs for AP +bafilomycin. Mean of biological replicates (independent conversions) ± s.e.m. n=3 independent conversions/line/condition. Statistical significance (one-way ANOVA) was calculated by comparing the DMSO condition to the AP, DMSO+bafilomycin, or AP+bafilomycin conditions and also comparing the AP condition to the AP+bafilomycin condition for CTRL and C9-ALS/FTD lines. P-values were corrected for multiple comparisons using statistical hypothesis testing (Šídák method). (J) Quantification of (A), the DMSO or 3 μM apilimod (AP) conditions with or without 50 nM bafilomycin. Quantified values represent the number of total LC3-mRFP+ (LC3-mRFP+/GFP+ and LC3-mRFP+/GFP-) punctae per μm 2 in n=30 iMNs for DMSO with bafilomycin, n=30 iMNs for AP with bafilomycin, n=45 iMNs for DMSO without bafilomycin, and n=50 iMNs for AP without bafilomycin from two C9-ALS/FTD patient lines. iMNs were quantified from two independent conversions per group. Each gray circle represents the ratio in a single iMN. Kruskal-Wallis test. Median ± interquartile range. P-values were corrected for multiple comparisons using statistical hypothesis testing (Dunn’s test). (K) Quantification class III beta-tubulin (TUJ1) in the pellet fractions from CTRL and C9-ALS/FTD iPSC-derived motor neurons. Cells were treated with DMSO, 3 μM apilimod (AP), 3 μM apilimod with 10 μM GW4869 (AP+GW), or 10 μM GW4869 (GW) for 24 hours. Values were calculated as the relative intensity of cell pellet TUJ1 (55kDa) normalized to total protein. One-way ANOVA. Mean ± s.e.m. P-values were corrected for multiple comparisons using statistical hypothesis testing (Šídák method). (L) FACS analysis of GFP+/CFSE+ exosomes from CTRL iPSC-derived motor neurons treated with DMSO (n=3 independent differentiations) or 3 μM apilimod (AP) (n=3 independent differentiations). Each data point represents an independently differentiated culture. Unpaired t- test. Mean ± s.e.m. (M) KEGG pathway of differentially-expressed proteins in the exosomal fraction of C9-ALS/FTD iPSC-derived motor neurons treated with DMSO or 3 μM apilimod (AP). (N) Immunoblots and quantification of Huntingtin (384 kDa) and alpha-synuclein (14 kDa) in the exosomal fractions from CTRL and C9-ALS/FTD iPSC-derived motor neurons. Cells were treated with DMSO or 3 μM apilimod (AP) for 24 hours. n=6 biological replicates (independent conversions) per condition (from two CTRL and two C9-ALS/FTD patients, n=3 independent conversions per line). The values were calculated as the relative intensity of exosomal Huntingtin or alpha-synuclein normalized to total protein. Ordinary one-way ANOVA. Mean ± s.e.m. P-values were corrected for multiple comparisons using statistical hypothesis testing (Šídák method). (O) Ratio of TSG101 and Optineurin (OPTN) in the exosomal vs. cell pellet fraction of CTRL and C9-ALS/FTD iPSC-derived motor neurons. n=9 independent conversions/condition from three CTRL and three C9-ALS/FTD patients, n=3 independent conversions/line). One-way ANOVA. Mean ± s.e.m. P-values were corrected for multiple comparisons using statistical hypothesis testing (Šídák method). (P) Immunoblots of exosomal fractions from CTRL and C9-ALS/FTD iPSC-derived motor neurons against total TDP43 (43kDa) and TSG101 (46kDa). Cells were treated with DMSO, 3 μM apilimod (AP), 3 μM apilimod with 10 μM GW4869 (AP+GW) or 10 μM GW4869 (GW) for 24 hours. (Q) Immunoblots of nuclear or cytoplasmic fractions from CTRL and C9-ALS/FTD iPSC-derived motor neurons against Fibrillarin (37kDa) and HSP90 (90kDa) to assess the purity of the nuclear and cytoplasmic fractions. Fibrillarin (37kDa) served as a marker of nuclear fraction and HSP90 (90kDa) served as a marker of cytoplasmic fraction. Cells were treated with DMSO or 3 μM apilimod (AP) for 24 hours. 184 Figure S3. 4 PIKFYVE inhibition clears pTDP-43 through amphisome and multivesicular body exocytosis. (A) Kaplan–Meier survival curves of C9-ALS/FTD iMNs treated with DMSO, 3 μM apilimod (AP) or 3 μM apilimod + 10 μM GW4869 (AP+GW) after withdrawal of neurotrophic factor supplementation (each line shown in a separate graph). n = 90 iMNs per line for three C9- 185 ALS/FTD lines (C9-ALS/FTD1, C9-ALS/FTD2 and C9-ALS/FTD3). iMNs quantified from three biologically independent iMN conversions per line per condition. Two-sided log-rank test. Statistical significance calculated using the entire survival course (days 1-20). Statistical significance shown on the plot was calculated comparing the conditions shown in the table for each line. P-values shown were not corrected for multiple comparisons. (B) mRNA levels of RAB27A (relative to HPRT), VAMP7, ATG7, RAB8A, HSPA8, MCOLN1, GORASP1, and NSMAF (relative to GAPDH) in iPSC-derived motor neurons after being treated with 9 μM negative control (NC) ASO or an ASO targeting each gene for 24 hours. For the RAB27A ASO, n=5 independent ASO treatments per condition from one CTRL and three C9- ALS/FTD lines. For the other ASOs, n=4 independent ASO treatments per condition from one CTRL and one C9-ALS/FTD lines. Two-tailed unpaired t-test. Mean ± s.e.m. Each ASO is color- coded and categorized by different forms of secretion: red (RAB27A and VAMP7, amphisome and multivesicular body exocytosis), brown (ATG7 and RAB8A, amphisome production and exocytosis, respectively), yellow (HSPA8, microautophagy and chaperone-mediated autophagy), green (MCOLN1, lysosomal exocytosis), blue (GORASP1, secretory autophagy), and purple (NSMAF, LC3-dependent extracellular vesicle loading and secretion). (C) The hazard ratio (Mantel–Haenszel method) of CTRL iMNs (2 lines in aggregate) treated with 9 μM negative control (NC) ASO, RAB27A ASO, VAMP7 ASO, ATG7 ASO, RAB8A ASO, HSPA8 ASO, MCOLN1 ASO, GORASP1 ASO, or NSMAF ASO plus DMSO or 3 μM apilimod (AP) after withdrawal of neurotrophic factor supplementation. The hazard ratio = the hazard rate of each group divided by the hazard rate of the three CTRL lines in aggregate treated with NC ASO and DMSO, which was set as 1 (red dotted line). n=429 iMNs for NC ASO+DMSO and n=326 iMNs for NC ASO+AP, n=275 iMNs for RAB27A ASO+AP, n=248 iMNs for VAMP7 ASO+AP, n=307 iMNs for ATG7 ASO+AP, n=356 iMNs for RAB8A ASO+AP, n=209 iMNs for HSPA8 ASO+AP, n=356 iMNs for MCOLN1 ASO+AP, n=277 iMNs for GORASP1 ASO+AP, and n=342 iMNs for NSMAF ASO+AP. The hazard ratio was derived from days 1-12 of iMN survival. Mean of biological replicates (independent conversions) ± s.e.m. n=3 independent conversions/line/condition. Statistical significance (one-way ANOVA) was calculated comparing each condition to the CTRL iMNs treated with NC ASO + AP group. P-values were corrected for multiple comparisons using statistical hypothesis testing (Šídák test). Each ASO is color-coded and categorized by different forms of secretion: red (RAB27A and VAMP7, amphisome and multivesicular body exocytosis), brown (ATG7 and RAB8A, amphisome exocytosis), yellow (HSPA8, microautophagy and chaperone-mediated autophagy), green (MCOLN1, lysosomal exocytosis), blue (GORASP1, secretory autophagy), and purple (NSMAF, LC3-dependent extracellular vesicle loading and secretion). (D) Immunoblots of exosomal fractions from C9-ALS/FTD iPSC-derived motor neurons against TSG101 (46kDa) and phosphorylated (Ser409/410) TDP-43 (50 kDa) (p-TDP43). Cells were treated with DMSO or 3 μM apilimod (AP) along with 9 μM negative control (NC) or RAB27A ASO1 for 24 hours. Samples were normalized to total protein. Cell pellet TSG101 (46kDa) served as a control for any differences in cell death between samples. (E) Quantification of endogenous LC3B+ punctae in CTRL or C9-ALS/FTD iMNs treated with 9 μM of negative control (NC) or RAB27A ASO for 48 hours and then treated with DMSO or 3 μM apilimod (AP) for 24 hours. Quantified values represent the average number of LC3B+ punctae per μm 2 from iMNs from two CTRL and two C9-ALS/FTD patient lines (n = 30 iMNs per condition per line). iMNs were quantified from two independent conversions per group. Each gray circle represents the average number in a single iMN. Kruskal-Wallis test. Median ± interquartile range. P-values were corrected for multiple comparisons using statistical hypothesis testing (Dunn’s test). (F) Immunostaining and quantification of endogenous LC3B+/LAMP1+ punctae in CTRL or C9- ALS/FTD iMNs treated with DMSO or 3 μM apilimod (AP) for 24 hours. Quantified values represent the average number of LC3+/LAMP1+ punctae per μm 2 from iMNs from two CTRL and two C9-ALS/FTD patient lines (n = 30 iMNs per condition per line). iMNs were quantified from two 186 independent conversions per group. Each gray circle represents the average number in a single iMN. Kruskal-Wallis test. Median ± interquartile range. P-values were corrected for multiple comparisons using statistical hypothesis testing (Dunn’s test). Solid lines outline the cell body. Scale bar = 5 μm. (G) Immunostaining and quantification of endogenous LC3B+/CD63+ punctae in CTRL or C9- ALS/FTD iMNs treated with 9 μM of negative control (NC) or RAB27A ASO for 48 hours and then treated with DMSO or 3 μM apilimod (AP) for 24 hours. Quantified values represent the average number of LC3B+/CD63+ punctae per μm 2 from iMNs from two CTRL and two C9-ALS/FTD patient lines (n = 30 iMNs per condition per line). iMNs were quantified from two independent conversions per group. Each gray circle represents the average number of punctae in a single iMN. Kruskal-Wallis test. Median ± interquartile range. P-values were corrected for multiple comparisons using statistical hypothesis testing (Dunn’s test). Solid lines outline the cell body. Scale bar = 5 μm. (H) Immunostaining and quantification of endogenous RAB7+ punctae in CTRL or C9-ALS/FTD iMNs treated with DMSO or 3 μM apilimod (AP) for 24 hours. Quantified values represent the average number of RAB7+ punctae per μm 2 from iMNs from two CTRL and two C9-ALS/FTD patient lines (n = 30 iMNs per condition per line). iMNs were quantified from two independent conversions per group. Each gray circle represents the average number of punctae in a single iMN. Kruskal-Wallis test. Median ± interquartile range. P-values were corrected for multiple comparisons using statistical hypothesis testing (Dunn’s test). Solid and dotted lines outline the cell body and nucleus, respectively. Scale bar = 5 μm. (I) Immunostaining and quantification of endogenous LAMP1+ punctae in CTRL and C9-ALS/FTD iMNs after withdrawal of neurotrophic factor supplementation for 96 hours and then with DMSO or 3 μM apilimod (AP) treatment for 96 hours. Quantified values represent the percentage of small LAMP1+ vesicles per cell (Left) and large LAMP1+ vesicles per cell (Right) from iMNs from two CTRL and two C9-ALS/FTD patient lines (n = 30 iMNs per condition per line). The size of LAMP1+ vesicles < 0.81 μm 2 were defined as small LAMP1+ vesicles; while vesicles > 0.81 μm 2 were defined as large LAMP1+ vesicles. 75% of LAMP1+ vesicles were < 0.81 μm 2 in C9-ALS /FTD iMNs treated with AP. iMNs were quantified from two independent conversions per group. Each gray circle represents the average number of punctae in a single iMN. Ordinary one-way ANOVA. Mean ± s.e.m. P-values were corrected for multiple comparisons using statistical hypothesis testing (Šídák method). Solid and dotted lines outline the cell body and nucleus, respectively. Scale bar = 5 μm. (J) Quantification of phosphorylated (Ser409/410) TDP-43 (p-TDP43)+ punctae in CTRL and C9- ALS/FTD iPSC-derived motor neurons treated with DMSO or 3 μM apilimod (AP) for 24 hours. Quantified values represent the average number of cytoplasmic p-TDP43+ punctae per μm 2 in the neurons (left), the average size (in μm 2 ) of p-TDP43+ punctae per neuron (middle) and the average intensity of p-TDP43+ punctae per neuron (right) from one CTRL and one C9-ALS/FTD patient line. (n = 15 neurons per condition per line). Neurons were quantified from two independent conversions per group. Each gray circle represents the average number, size, or intensity of punctae in a single neuron. Ordinary one-way ANOVA and p-values were corrected for multiple comparisons using statistical hypothesis testing (Šídák method) (left); Unpaired t-test (middle and right). Mean ± s.e.m. 187 Figure S3. 5 PIKFYVE inhibition improves iMN proteostasis and survival for diverse forms of ALS. (A) Representative images showing the number of sporadic ALS iMNs treated with DMSO, 3 μM apilimod (AP) or 3 μM apilimod+500 nM GW4869 (AP+GW) over time during the iMN survival assay. Scale bar = 50 μm. 188 (B) Kaplan–Meier survival curves of CTRL (two lines in aggregate) iMNs treated with DMSO and eight sporadic ALS (sALS) iMNs treated with DMSO or 3 μM apilimod (AP) after withdrawal of neurotrophic factor supplementation (shown for each sALS line separately, sALS1-sALS8). n=90 iMNs per line for two CTRL, n=180 iMNs per line per condition for six sALS lines (sALS1-sALS6) and n=70 iMNs per line for two sALS lines (sALS7-sALS8). iMNs quantified from three biologically independent iMN conversions per line. Two-sided log-rank test. Statistical significance calculated using the entire survival course (days 1-20). (C) Kaplan–Meier survival curves of iMNs from one sALS donor (sALS9) in DMSO, 0.03 μM, 0.3 μM, 3 μM or 30 μM AP after withdrawal of neurotrophic factor supplementation. n=1530 iMNs for DMSO, n=621 iMNs for 0.03 μM AP, n=507 iMNs for 0.3 μM AP, n=488 iMNs for 3 μM AP and n=410 iMNs for 30 μM AP. iMNs quantified from three biologically independent iMN conversions per line. Two-sided log-rank test. Statistical significance calculated using the entire survival course (days 1-14). (D) The hazard ratio (Mantel–Haenszel method) of iMNs from one sALS donor (sALS9) treated with DMSO, 0.03 μM, 0.3 μM, 3 μM or 30 μM apilimod (AP) after withdrawal of neurotrophic factor supplementation. The hazard ratio = the hazard rate of each group divided by the hazard rate of the sALS9 line treated with DMSO, which was set as 1 (red dotted line). n=1530 iMNs for DMSO, n=621 iMNs for 0.03 μM AP, n=507 iMNs for 0.3 μM AP, n=488 iMNs for 3 μM AP and n=410 iMNs for 30 μM AP. The hazard ratio was derived from days 1-14 of iMN survival. Mean of biological replicates (independent conversions) ± s.e.m. n=3 independent conversions per line per condition. Statistical significance (one-way ANOVA) was calculated comparing different doses of AP treatment to the DMSO condition for sALS9 line. P-values were corrected for multiple comparisons using statistical hypothesis testing (Šídák method). (E) The hazard ratio (Mantel–Haenszel method) of iMNs from three CTRL and six sALS lines in aggregate (sALS1-sALS6) treated with 10 μM negative control (NC), PIKFYVE ASO1, or ASO2 after withdrawal of neurotrophic factor supplementation. n = 100 iMNs per treatment for each line. The hazard ratio = the hazard rate of each group divided by the hazard rate of the three CTRL lines in aggregate treated with NC ASO, which was set as 1 (red dotted line). The hazard ratio was derived from days 1-6 of iMN survival. Mean of biological replicates (independent conversions) ± s.e.m. n = 3 independent conversions per line per condition. Statistical significance (one-way ANOVA) was calculated comparing the sALS+NC ASO group to the CTRL+NC ASO, sALS+PIKFYVE ASO1, and sALS+PIKFYVE ASO2 groups. P-values were corrected for multiple comparisons using statistical hypothesis testing (Šídák method). (F) Kaplan–Meier survival curves of two FUS ALS (H517Q, R522R mutation), and TARDBP ALS (G298S mutation) donors treated with DMSO or 3 μM apilimod (AP) after withdrawal of neurotrophic factor supplementation (each line shown in a separate graph). n = 50 iMNs per line. iMNs quantified from three biologically independent iMN conversions per line per condition. Two- sided log-rank test. Statistical significance calculated using the entire survival course (days 1-15). (G) Immunostaining and quantification of FUS in iMNs from two CTRL and two FUS ALS (H517Q, R522R mutation) lines that were treated with DMSO or 3 μM apilimod (AP) for 24 hours. Quantified values represent the average ratio of nuclear to cytoplasmic FUS in n=30 CTRL and n=30 FUS ALS iMNs from two CTRL and two FUS ALS patient lines. iMNs were quantified from two independent conversion/group. Each gray circle represents the ratio from a single iMN. Kruskal-Wallis test. Median ± interquartile range. P-values were corrected for multiple comparisons using statistical hypothesis testing (Dunn’s test). Solid and dotted lines outline the cell body and nucleus, respectively. Scale bar = 5 μm. (H) The hazard ratio (Mantel–Haenszel method) of CTRL (two lines in aggregate), two sporadic ALS (sALS3 and sALS5) and one FUS-ALS (R522R mutation) iMNs treated with 9 μM negative control (NC) ASO, VAMP7 or ATG7 ASO plus DMSO or 3 μM apilimod (AP) after withdrawal of neurotrophic factor supplementation. The hazard ratio = the hazard rate of each group divided by the hazard rate of the CTRL line treated with NC ASO+DMSO, which was set as 1 (red dotted 189 line). The hazard ratio was derived from day 1 to day 12 of iMN survival. For CTRL, n=439 iMNs for NC ASO+DMSO. For sALS3, n=214 iMNs for NC ASO+DMSO, n=164 iMNs for NC ASO+AP, n=173 iMNs for VAMP7 ASO+DMSO, n=133 iMNs for VAMP7 ASO+AP, n=177 iMNs for ATG7 ASO+DMSO and n=132 iMNs for ATG7 ASO+AP. For sALS5, n=217 iMNs for NC ASO+DMSO, n=103 iMNs for NC ASO+AP, n=137 iMNs for VAMP7 ASO+DMSO, n=104 iMNs for VAMP7 ASO+AP, n=134 iMNs for ATG7 ASO+DMSO and n=120 iMNs for ATG7 ASO+AP. For FUS R522R, n=133 iMNs for NC ASO+DMSO, n=111 iMNs for NC ASO+AP, n=113 iMNs for VAMP7 ASO+DMSO, n=111 iMNs for VAMP7 ASO+AP, n=105 iMNs for ATG7 ASO+DMSO and n=107 iMNs for ATG7 ASO+AP. Mean of biological replicates (independent conversions) ± s.e.m. n=3 independent conversions/line/condition. Statistical significance (one-way ANOVA) was calculated comparing NC ASO+AP to NC ASO+DMSO, VAMP7 ASO+AP or ATG7 ASO+AP for each sALS and FUS line. p-values were corrected for multiple comparisons using statistical hypothesis testing (Šídák method). (I) The hazard ratio (Mantel–Haenszel method) of iMNs from CTRL (two lines in aggregate) and three sporadic ALS (sALS1, sALS2 and sALS4) lines that were treated with DMSO, 3 μM apilimod (AP) or 3 μM apilimod+500 nM GW4869 (AP+GW). n=150 iMNs/line/condition. The hazard ratio = the hazard rate of each group divided by the hazard rate of the two CTRL lines in aggregate treated with DMSO, which was set as 1 (red dotted line). The hazard ratio was derived from days 1-20 of iMN survival. Mean of biological replicates (independent conversions) ± s.e.m. n=3 independent conversions/line/condition. Statistical significance (one-way ANOVA) was calculated for the +AP condition compared to the +DMSO and +AP+GW conditions for each line. P-values were corrected for multiple comparisons using statistical hypothesis testing (Šídák method). (J) Kaplan–Meier survival curves of iMNs from three sporadic ALS (sALS1, sALS2 and sALS4) lines that were treated with DMSO, 3 μM apilimod (AP), or 3 μM apilimod+500 nM GW4869 (AP+GW) after withdrawal of neurotrophic factor supplementation. n = 150 iMNs per condition for sALS1 and sALS2. n=80 iMNs per condition for sALS4. iMNs quantified from three biologically independent iMN conversions per line per condition. Two-sided log-rank test. Statistical significance calculated using the entire survival course (days 1-20). Statistical significance shown on the plot was calculated using +AP and +AP+GW. (K) The hazard ratio (Mantel–Haenszel method) of iMNs from one CTRL and one TARDBP ALS (G298S mutation) donor treated with DMSO, 3 μM apilimod (AP), or 3 μM apilimod+500 nM GW4869 (AP+GW). n=150 iMNs/line/condition. The hazard ratio = the hazard rate of each group divided by the hazard rate of the CTRL line in aggregate treated with DMSO, which was set as 1 (red dotted line). The hazard ratio was derived from days 1-20 of iMN survival. Mean of biological replicates (independent conversions) ± s.e.m. n = 3 independent conversions/line/condition. Statistical significance (one-way ANOVA) was calculated for the comparisons labeled with p- values. P-values were corrected for multiple comparisons using statistical hypothesis testing (Šídák method). (L) The hazard ratio (Mantel–Haenszel method) of iMNs from one CTRL and one FUS ALS (H517Q) line treated with DMSO, 3 μM apilimod (AP), or 3 μM apilimod+500 nM GW4869 (AP+GW). n=150 iMNs/line/condition. The hazard ratio = the hazard rate of each group divided by the hazard rate of the CTRL line in aggregate treated with DMSO, which was set as 1 (red dotted line). The hazard ratio was derived from days 1-20 of iMN survival. Mean of biological replicates (independent conversions) ± s.e.m. n = 3 independent conversions/line/condition. Statistical significance (one-way ANOVA) was calculated for the comparisons labeled with p- values. P-values were corrected for multiple comparisons using statistical hypothesis testing (Šídák method). (M) Kaplan–Meier survival curves of FUS ALS (H517Q) and TARDBP ALS (G298S mutation) iMNs treated with DMSO, 3 μM apilimod (AP), or 3 μM apilimod+500 nM GW4869 (AP+GW) after withdrawal of neurotrophic factor supplementation. n=150 iMNs per condition for FUS H517Q and TARDBP iMNs. iMNs quantified from three biologically independent iMN conversions per line 190 per condition. Two-sided log-rank test. Statistical significance calculated using the entire survival course (days 1-20). Statistical significance shown on the plot was calculated for the comparisons shown in the table. P-values were not corrected for multiple comparisons. (N) The hazard ratio (Mantel–Haenszel method) of iMNs from two CTRL (in aggregate) lines treated with DMSO and one FUS ALS (R522R mutation) line treated with DMSO, 3 μM apilimod (AP), DMSO+0.5 nM bafilomycin, or 3 μM apilimod (AP)+0.5 nM bafilomycin after withdrawal of neurotrophic factor supplementation. The hazard ratio = the hazard rate of each group relative to the hazard rate of iMNs from the two CTRL lines in aggregate treated with DMSO, which was set as 1 (red dotted line). The hazard ratio was derived from days 1-10 of iMN survival. For CTRL lines, n=180 iMNs for DMSO. For the FUS ALS line (R522R mutation), n=82 iMNs for DMSO, n=82 iMNs for AP, n=182 iMNs for DMSO+bafilomycin, n=196 iMNs for AP +bafilomycin. Mean of biological replicates (independent conversions) ± s.e.m. n=3 independent conversions/line/condition. Statistical significance (one-way ANOVA) was calculated for the comparisons labeled with p-values. P-values were corrected for multiple comparisons using statistical hypothesis testing (Šídák method). (O) Representative images of worms: control, untreated (Top) or apilimod-treated (Bottom) TDP- 43 A315T transgenics expressing unc-47::GFP in motor neurons. Arrows indicate gaps or breaks along neuronal processes that are scored as neurodegeneration events. (P) Treatment with 1 μM apilimod (AP) suppressed age-dependent motility defects leading to paralysis in TDP-43 A315T transgenics. P-value <0.001 for all treated strains compared to untreated controls. Log-rank (Mantel Cox) test. n>200 worms for each experiment. (Q) Apilimod suppressed motor neuron degeneration in TDP-43 A315T transgenics. ***P- value<0.001 compared to untreated controls. One-way ANOVA with Dunnett’s multiple comparison test. 191 Figure S3. 6 Pikfyve suppression improves motor function and extends survival of TDP- 43 and C9ORF72 mice. (A) Fold-change of selected protein levels in adult mice in the CSF after intrathecal injection of apilimod compared to vehicle. CSF from n=6 mice was pooled and ran in duplicate for the vehicle- treated group; CSF from n=4 mice was pooled together and ran in duplicate for the apilimod- treated group. Each value is the average of two technical replicates per condition per group. (B) Mating scheme for the generation of a single-allele deletion of Pikfyve in TDP-43 Tg/Tg mice. (C) Schematic diagram of the targeting strategy used to disrupt exon-6 of the Pikfyve gene using the Cre-lox system. Exon 6 was excised by Cre recombinase which targeted the LoxP sites. Genotyping of genomic DNA using four primer sets reveals the appropriate genotypes for Pikfyve +/- and Pikfyve +/+ mice. 192 (D) Pikfyve mRNA expression was significantly decreased in the whole brains of Pikfyve +/- ;TDP- 43 Tg/Tg mice. The mice were 15-18 days of age. Two-tailed unpaired t-test. Data were normalized to Pikfyve +/+ ;TDP-43 Tg/Tg mice and are presented as the mean ± s.e.m.. Each data point represents one mouse. (E) Immunoblot analysis of PIKFYVE in the whole brain of Pikfyve +/+ ; WT and Pikfyve +/- ; WT at 15-18 days of age. The values are calculated as the relative intensity of PIKFYVE normalized to the amount of total protein. Statistical significance was calculated comparing Pikfyve +/+ ; WT and Pikfyve +/- ; WT. Unpaired t-test. Mean ± s.e.m. Each data point represents one mouse. (F) mRNA levels of Pikfyve (relative to Ppia) in whole brains of WT mice treated with NC ASO (n=3 mice) or Pikfyve ASO (n=3 mice) at P7 after 25 μg of ASO injection at P0. Unpaired t-test. Mean± s.e.m. (G) Immunoblot analysis of PIKFYVE levels in the whole brains of WT and TDP-43 Tg/Tg mice at P20 after treatment with 25 μg of NC ASO or Pikfyve ASO at P1. WT mice + NC ASO (n=8 mice), WT mice + Pikfyve ASO (n=6 mice), TDP-43 Tg/Tg mice + NC ASO (n=10 mice) and TDP-43 Tg/Tg mice + Pikfyve ASO (n=7 mice). The values are calculated as the relative intensity of PIKFYVE normalized to the amount of total protein. Mann-Whitney test. Median ± interquartile range. Statistical significance was calculated comparing WT mice + NC ASO to WT mice + Pikfyve ASO and TDP-43 Tg/Tg mice + NC ASO to TDP-43 Tg/Tg mice + Pikfyve ASO. Each data point represents one mouse. (H-I) The kyphosis (H) and tremor scores (I) of Pikfyve +/+ ;WT (n=9 mice), Pikfyve +/- ;WT (n=10 mice), Pikfyve +/+ ;TDP-43 Tg/Tg (n=12 mice) and Pikfyve +/- ; TDP-43 Tg/Tg (n=7 mice) mice. A score of 0 indicates no phenotype and 4 indicates the most severe phenotype. Statistical significance was calculated by comparing Pikfyve +/+ ;TDP-43 Tg/Tg and Pikfyve +/- ; TDP-43 Tg/Tg at each day. Unpaired t-test. Mean ± s.e.m. (J-K) The kyphosis (J) and tremor scores (K) of WT mice treated with vehicle and negative control (NC) ASO (n=10 mice), WT mice treated with vehicle and Pikfyve ASO (n=13 mice), TDP-43 Tg/Tg mice treated with vehicle and NC ASO (n=12 mice), TDP-43 Tg/Tg mice treated with vehicle and Pikfyve ASO (n=16 mice). Postnatal day 1 (P1) mouse pups received 25 μg of NC or Pikfyve ASO by intracerebroventricular injection. Vehicle (for GW4869) was administered by intraperitoneal injection every 48 hrs starting from P5. A score of 0 indicates no phenotype and 4 indicates the most severe phenotype. Statistical significance was calculated by comparing TDP-43 Tg/Tg mice treated with vehicle+NC ASO to TDP-43 Tg/Tg mice treated with vehicle +Pikfyve ASO at each time point. Unpaired t-test. Mean ± s.e.m. (L-M) The kyphosis (L) and tremor scores (M) of TDP-43 Tg/Tg mice treated with vehicle and NC ASO (n=12 mice), TDP-43 Tg/Tg mice treated with vehicle and Pikfyve ASO (n=16 mice), TDP- 43 Tg/Tg mice treated with GW4869 and NC ASO (n=10 mice), and TDP-43 Tg/Tg mice treated with GW4869 and Pikfyve ASO (n=11 mice). Postnatal day 1 (P1) mouse pups received 25 μg of NC or Pikfyve ASO by intracerebroventricular injection. Vehicle or GW4869 were administered by intraperitoneal injection every 48 hrs starting from P5. A score of 0 indicates no phenotype and 4 indicates the most severe phenotype. Statistical significance was calculated by comparing TDP- 43 Tg/Tg mice treated with vehicle+Pikfyve ASO to TDP-43 Tg/Tg mice treated with GW4869+Pikfyve ASO at each time point. Unpaired t-test. Mean ± s.e.m. (N-P) The gait impairment (N), kyphosis (O) and tremor (P) scores of WT mice treated with negative control (NC) ASO (n=7 mice), WT mice treated with Pikfyve ASO (n=6 mice), TDP-43 Tg/Tg mice treated with NC ASO (n=12 mice), TDP-43 Tg/Tg mice treated with Pikfyve ASO (n=12 mice). Postnatal day 1 (P1) mouse pups received 5 μg of NC or Pikfyve ASO by intracerebroventricular injection. A score of 0 indicates no phenotype and 4 indicates the most severe phenotype. Statistical significance was calculated by comparing TDP-43 Tg/Tg mice treated with NC ASO to TDP-43 Tg/Tg mice treated with Pikfyve ASO at each time point. Unpaired t-test. Mean ± s.e.m. (Q) Immunostaining of GFP and poly(GR) in the cortex of mice transduced with AAV-eGFP or AAV-eGFP-(GR) 100. Magnification of the area indicated by a rectangle. 193 Figure S3. 7 Pikfyve suppression reduces TDP-43 and C9ORF72 pathology and neurodegeneration in vivo. (A) Immunoblot analysis of total TDP-43 levels in the whole brains of WT and TDP-43 Tg/Tg mice treated with NC ASO or Pikfyve ASO. Groups include WT mice + NC ASO (n=8 mice), WT mice + Pikfyve ASO (n=6 mice), TDP-43 Tg/Tg mice + NC ASO (n=10 mice) and TDP-43 Tg/Tg mice + Pikfyve ASO (n=7 mice). The values are calculated as the relative intensity of TDP-43 normalized 194 to total protein. Ordinary one-way ANOVA. Mean± s.e.m. Statistical significance was calculated comparing WT mice + NC ASO to WT mice + Pikfyve ASO and TDP-43 Tg/Tg mice + NC ASO to TDP-43 Tg/Tg mice + Pikfyve ASO. P-values were corrected for multiple comparisons using statistical hypothesis testing (Dunn’s test). Each data point represents one mouse. (B) Immunoblot analysis of phosphorylated (Ser409/410) TDP-43 (p-TDP-43) levels in the cytoplasmic fraction of whole brain lysates from TDP-43 Tg/Tg mice treated with NC ASO (n=9 mice) or Pikfyve ASO (n=12 mice). The values are calculated as the relative intensity of p-TDP-43 normalized to total protein. Unpaired t-test. Mean± s.e.m. Statistical significance was calculated comparing TDP-43 Tg/Tg mice + NC ASO to TDP-43 Tg/Tg mice + Pikfyve ASO. Each data point represents one mouse. (C) Immunostaining and quantification of pTDP-43+ (Ser403/404) punctae in TUJ1+ motor neurons in the dorsal spinal cord from WT or TDP-43 Tg/Tg mice treated with NC or Pikfyve ASO. Solid and dotted lines outline the cell body and nucleus, respectively. Neuron cell bodies were identified using TUJ1 immunostaining. Orange arrows mark pTDP-43+ punctae. Scale bar = 10 μm. For quantification, each data point represents the average number of pTDP-43+ (Ser403/404) punctae in TUJ1+ motor neurons per μm 2 for one mouse. Each data point represents one mouse. n=7 mice for WT treated with NC ASO, n=8 mice for WT treated with Pikfyve ASO, n=12 mice for TDP-43 Tg/Tg treated with NC ASO and n=8 mice for TDP-43 Tg/Tg treated with Pikfyve ASO. 30-40 cells were analyzed per mouse. One-way ANOVA. Mean ± s.e.m. P-values were corrected for multiple comparisons using statistical hypothesis testing (Šídák method). (D) Immunostaining and quantification of pTDP-43+ (Ser403/404) punctae in layer V neurons in the motor cortex layer WT or TDP-43 Tg/Tg mice treated with NC or Pikfyve ASO. Solid and dotted lines outline the cell body and nucleus, respectively. Neuron cell bodies were identified using TUJ1 immunostaining. Orange arrows mark pTDP-43+ punctae. Scale bar = 10 μm. Quantified values represent the average number of pTDP-43+ (Ser 403/404) punctae in TUJ1+ neurons in each mouse. Each data point represents one mouse. n=5 mice for WT treated with NC ASO, n=6 mice for WT treated with Pikfyve ASO, n=15 mice for TDP-43 Tg/Tg treated with NC ASO and n=16 mice for TDP-43 Tg/Tg treated with Pikfyve ASO. 20-30 cells were analyzed per mouse. Kruskal- Wallis test. Mean ± s.e.m. P-values were corrected for multiple comparisons using statistical hypothesis testing (Dunn’s test). (E) Immunostaining and quantification of pTDP-43+ (Ser403/404) punctae in neurons in the caudal putamen from WT or TDP-43 Tg/Tg mice treated with NC or Pikfyve ASO. Solid and dotted lines outline the cell body and nucleus, respectively. Neuron cell bodies were identified using TUJ1 immunostaining. Orange arrows mark pTDP-43+ punctae. Scale bar = 10 μm. Quantified values represent the average number of pTDP-43+ (Ser403/404) punctae in TUJ1+ neurons in each mouse. Each data point represents one mouse. n=4 mice for WT treated with NC ASO, n=5 mice for WT treated with Pikfyve ASO, n=6 mice for TDP-43 Tg/Tg treated with NC ASO and n=5 mice for TDP-43 Tg/Tg treated with Pikfyve ASO. 20-30 cells were analyzed per mouse. Kruskal- Wallis test. Mean ± s.e.m. P-values were corrected for multiple comparisons using statistical hypothesis testing (Dunn’s test). (F) Immunostaining of pTDP-43+ (Ser409/410) punctae in CTRL iPSC-derived IBA1+ microglia treated with or without exosomes harvested from C9-ALS/FTD iPSC-derived motor neurons for 24 or 96 hours. Orange arrows marked pTDP-43+ punctae in IBA1+ microglia. Scale bar = 10 μm. (G) Quantification of (F). Quantified values represent the average number of pTDP-43+ (Ser409/410) punctae in IBA1+ microglia per μm2 treated with or without exosomes harvested from C9-ALS/FTD iPSC-derived motor neurons for 24 or 96 hours. n=51 microglia without C9- ALS/FTD exosomes for 24 hours, n=40 microglia without C9- ALS/FTD exosomes for 96 hours, n=40 microglia with C9- ALS/FTD exosomes for 96 hours and n=45 microglia with C9- ALS/FTD exosomes for 96 hours. IBA1+ microglia were quantified from two independent conversions per group. Each gray circle represents the average number of pTDP-43+ punctae in a single IBA1+ 195 microglial cell. Kruskal-Wallis test. Median ± interquartile range. P-values were corrected for multiple comparisons using statistical hypothesis testing (Dunn’s test). (H) Quantification of (F). Quantified values represent the percentage of ramified (brown) or amoeboid (green) IBA1+ microglia treated with or without exosomes harvested from C9-ALS/FTD iPSC-derived motor neurons for 24 or 96 hours. (I) Immunostaining and quantification of IBA1+ microglia in the spinal cord ventral horn of WT or TDP-43 Tg/Tg mice treated with NC or Pikfyve ASO. Scale bar = 200 μm. Quantified values represent the average number of IBA1+ microglia per ventral horn section in each mouse. Each data point represents one mouse. n=6 mice for WT treated with NC ASO, n=7 mice for WT treated with Pikfyve ASO, n=7 mice for TDP-43 Tg/Tg treated with NC ASO and n=8 mice for TDP-43 Tg/Tg treated with Pikfyve ASO. 2-4 ventral horns were quantified and averaged per mouse. One-way ANOVA. Mean ± s.e.m. P-values were corrected for multiple comparisons using statistical hypothesis testing (Šídák method). (J) Immunostaining and quantification of GFAP+ astrocytes in the spinal cord ventral horn of WT or TDP-43 Tg/Tg mice treated with NC or Pikfyve ASO. Scale bar = 200 μm. Quantified values represent the average number of GFAP+ astrocytes per ventral horn section in each mouse. Each data point represents one mouse. n=6 mice per group. 2-4 ventral horns were quantified and averaged per mouse. One-way ANOVA. Mean ± s.e.m. P-values were corrected for multiple comparisons using statistical hypothesis testing (Šídák method). (K) Immunostaining and quantification of RAB7+ punctae in TUJ1+ motor neurons in the spinal cord ventral horn from WT or TDP-43 Tg/Tg mice treated with NC or Pikfyve ASO. Neuron cell bodies (solid lines) were identified using TUJ1 immunostaining. Scale bar = 10 μm. Each data point represents the average number of RAB7+ punctae in TUJ1+ motor neurons per μm 2 for one mouse. Each data point represents one mouse. n=4 mice per condition. 40-50 cells were analyzed per mouse. One-way ANOVA. Mean ± s.e.m. P-values were corrected for multiple comparisons using statistical hypothesis testing (Šídák method). (L) Immunostaining and quantification of the percentage of large LAMP1+ vesicles in TUJ1+ motor neurons in the spinal cord ventral horn from WT or TDP-43 Tg/Tg mice treated with NC or Pikfyve ASO. Neuron cell bodies (solid lines) were identified using TUJ1 immunostaining. Scale bar = 10 μm. Each data point represents the average percentage of large LAMP1+ vesicles in TUJ1+ motor neurons for one mouse. Each data point represents one mouse. n=4 mice per condition. 40-50 cells were analyzed per mouse. One-way ANOVA. Mean ± s.e.m. P-values were corrected for multiple comparisons using statistical hypothesis testing (Šídák method). (M) Immunostaining and quantification of LC3B+/LAMP1+ punctae in TUJ1+ motor neurons in the spinal cord ventral horn from WT or TDP-43 Tg/Tg mice treated with NC or Pikfyve ASO. Neuron cell bodies (Solid lines) were identified using TUJ1 immunostaining. Orange arrows marked LC3B+/LAMP1+ punctae. Scale bar = 10 μm. Each data point represents the average number of LC3B+/LAMP1+ punctae in TUJ1+ motor neurons per μm 2 for one mouse. Each data point represents one mouse. n=4 mice per condition. 40-50 cells were analyzed per mouse. One-way ANOVA. Mean ± s.e.m. P-values were corrected for multiple comparisons using statistical hypothesis testing (Šídák method). (N) Immunoblot analysis of TFEB levels in the nuclear or cytoplasmic fraction of whole brain lysates from TDP-43 Tg/Tg mice treated with NC ASO or Pikfyve ASO. TDP-43 Tg/Tg mice + NC ASO (n=9 mice) and TDP-43 Tg/Tg mice + Pikfyve ASO (n=12 mice). The values are calculated as the average ratio of nuclear to cytoplasmic TFEB per mice. Mann-Whitney test. Mean ± s.e.m. Each data point represents one mouse. Fibrillarin (37kDa) served as a marker of the nuclear compartment. (O) Immunostaining and quantification of the percentage of large LAMP1+ vesicles in TUJ1+ motor neurons in the spinal cord ventral horn from 10-month-old Pikfyve +/+ ;WT, Pikfyve +/- ;WT, Pikfyve +/+ ;TDP-43 Tg/+ and Pikfyve +/- ; TDP-43 Tg/+ mice. Solid and dotted lines outline the cell body and nucleus, respectively. Neuron cell bodies were identified using TUJ1 immunostaining. Orange 196 arrows mark LAMP1+ punctae. Scale bar = 2 μm. Each data point represents the average percentage of large LAMP1+ vesicles in TUJ1+ motor neurons for one mouse. Each data point represents one mouse. n=5 for Pikfyve +/+ ;WT mice, n=3 for Pikfyve +/- ;WT mice, n=4 for Pikfyve +/+ ;TDP-43 Tg/+ mice and n=4 for Pikfyve +/- ; TDP-43 Tg/+ mice. 40-50 cells were analyzed per mouse. One-way ANOVA. Mean ± s.e.m. P-values were corrected for multiple comparisons using statistical hypothesis testing (Šídák method). (P) Quantification of LAMP1+ punctae in CTRL and C9-ALS/FTD iMNs after withdrawal of neurotrophic factor supplementation for 96 hours and then with DMSO or 3 μM apilimod (AP) treatment for 96 hours. Quantified values represent the average number of LAMP1+ punctae per μm in n=30 CTRL and n=30 C9-ALS/FTD iMNs from two CTRL and two C9-ALS/FTD patient lines. iMNs were quantified from two independent conversions/group. Each gray circle represents the average number of LAMP1+ punctae in a single iMN. Kruskal-Wallis test. Median ± interquartile range. P-values were corrected for multiple comparisons using statistical hypothesis testing (Dunn’s test). 197 Appendix D: Supplementary Figures for Chapter 4 Figure S4. 1 Transcription factor-mediated conversion of iPSCs into induced microglia. (A) Schematic of screening microglial transcription factors for production of human induced microglia with TMEM119-T2A-tomato iPSCs. (B) Representative image of TMEM119-T2A-tomato + cells after 6 days of infection with Pu.1 and CEBPA. Scale bar = 200 μm. (C) Number of tdTomato + iMG produced by forced expression of different combinations of transcription factors. Each of the 20 transcription factors was dropped out in each group. The number of tdTomato + iMG produced by 20 transcription factors as the standard (red dotted line). One-way ANOVA. Mean ± s.e.m. n=2 biological replicates for each condition. (D) Principal component analysis (PCA) of iPSCs (green), human fibroblasts (purple), iMG (red) and primary human microglia (blue). PC1 (63% variance) and PC2 (19 % variance) reflects induced microglia are more similar to primary human microglia and distinct from human iPSCs and fibroblasts. n=2 biological replicates per cell type. (E) Heatmap of the expression of 8 microglial genes and 2 iPSCs genes in induced microglia and iPSCs. 198 Figure S4. 2 C9ORF72 ALS/FTD microglia display different transcriptomic profiles and alter endolysosome phenotypes. (A) Flow cytometry analysis of CD11B+ human induced microglia. (B) The number of DEGs by pairwise genotype comparisons: C9ORF72 ALS/FTD iMG vs control iMG; C9ORF72 -/- iMG vs control iMG; C9ORF72 isogenic iMG vs C9ORF72 ALS/FTD iMG. Adjusted p value < 0.01. (C) Volcano plot displaying the log 2-fold change (x axis) against the t test-derived −log 10 statistical P value (y axis) for all differentially expressed genes between C9ORF72 ALS/FTD iMG and control iMG. DEGs significantly increased in C9ORF72 ALS/FTD iMG (log 2-fold change > 1) are shown in red, while DEGs significantly decreased in C9ORF72 ALS/FTD iMG (log 2-fold change < -1) are shown in blue. Select DEGs are labeled. 199 (D) Unsupervised clustering heatmap of bulk RNA-seq data from iMG from 2 C9ORF72 ALS/FTD and 2 C9ORF72 isogenic corrected lines. n=2-3 biological replicates per line. iMG were FACS-purified by CD11B antibody. (E) Top gene ontology of significantly up-regulated genes (yellow) and down-regulated genes (gray) in C9ORF72 isogenic iMG over C9ORF72 ALS/FTD iMG. (F) Volcano plot displaying the log 2-fold change (x axis) against the t test-derived −log 10 statistical P value (y axis) for all differentially expressed genes between isogenic C9ORF72 iMG and C9ORF72 ALS/FTD iMG. DEGs significantly increased in isogenic C9ORF72 iMG (log 2- fold change > 1) are shown in red, while DEGs significantly decreased in isogenic C9ORF72 iMG (log 2-fold change < -1) are shown in blue. Select DEGs are labeled. (G) Volcano plot displaying the log 2-fold change (x axis) against the t test-derived −log 10 statistical P value (y axis) for all differentially expressed genes between C9ORF72 -/- iMG and control iMG. DEGs significantly increased in C9ORF72 -/- iMG (log 2-fold change > 1) are shown in red, while DEGs significantly decreased in C9ORF72 -/- iMG (log 2-fold change < -1) are shown in blue. Select DEGs are labeled. 200 Figure S4. 3 C9ORF72 ALS/FTD microglia exert neuroprotection on C9ORF72 ALS/FTD neurons. (A) Kaplan–Meier survival curves of control (CTRL) and C9ORF72 ALS/FTD patient (C9- ALS/FTD) iMNs after withdrawal of neurotrophic factor supplementation (shown in aggregate). n=90 iMNs/line for three control and three C9-ALS/FTD lines. iMNs quantified from three biologically independent iMN conversions per line. Two-sided log-rank test. Statistical significance calculated using the entire survival course (days 1-21). (B) Immunostaining of endogenous poly(GR)+ punctae in CTRL or C9-ALS/FTD iMNs. Solid and dotted lines outline the cell body and nucleus, respectively. Scale bar = 2 μm. (C) Quantification of (B) to determine endogenous poly(GR)+ punctae in CTRL or C9-ALS/FTD iMNs. Quantified values represent the average number of nuclear poly(GR)+ punctae per μm 2 in n=40 CTRL and n=40 C9-ALS/FTD iMNs from two CTRL and two C9-ALS/FTD patient lines (n=20 per line). iMNs were quantified from two independent conversions per group. Each gray circle represents the average number in a single iMN. Mann-Whitney test. Mean ± interquartile range. (D) Kaplan–Meier survival curves of control and C9ORF72 ALS/FTD patient (C9-ALS/FTD) iMNs with no iMG (black lines), control iMG (blue lines) and C9-ALS/FTD iMG (red lines). n=90 iMNs/line for three control and three C9-ALS/FTD lines. iMNs quantified from three biologically independent iMN conversions per line. Two-sided log-rank test. Statistical significance calculated using the entire survival course (days 1-21). (E) Kaplan–Meier survival curves of control and C9ORF72 ALS/FTD patient (C9-ALS/FTD) iMNs with no iMG or C9 -/- iMG. n=90 iMNs/line for two control and two C9-ALS/FTD lines. iMNs quantified from three biologically independent iMN conversions per line. Two-sided log-rank test. Statistical significance calculated using the entire survival course (days 1-21). 201 Figure S4. 4 A novel CSF1R-low neuroprotective state of C9ORF72 ALS/FTD microglia. (A) Flow cytometry analysis of CD11B+ induced microglia (iMG) and HB9::RFP+ induced motor neurons (iMNs) after 3 days of iMNs-iMG co-cultures. (B) UMAP plot depicting single-cell RNA-seq data for co-cultures of control (CTRL) iMNs with control (CTRL) iMG. Cluster 0-4 are neuron populations and cluster 5 (black dotted line) is AIF1+ microglia population. (C) Violin plots of AIF1 in all the clusters of (B). (D) UMAP plot depicting single-cell RNA-seq data for co-cultures of C9ORF72 ALS/FTD (C9) iMNs with control (CTRL) iMG. Clusters 0-1 and 3-6 are neuron populations and cluster 2 and cluster 7 (black dotted line) are AIF1+ microglia populations. (E) Violin plots of AIF1 in all the clusters of (D). (F) Violin plots of AIF1, CD68 and CSF1R expression in iMG from different co-cultures. For C9 iMNs-C9 iMG co-cultures, the cluster of each iMG population is labeled. (G) Kaplan–Meier survival curves of control and C9ORF72 ALS/FTD patient (C9-ALS/FTD) iMNs with no iMG (black lines), FACS-purified CSF1R-high (red lines) or CSF1R-low (blue lines) C9- ALS/FTD iMG. n=70 iMNs/line for one control and one C9-ALS/FTD line. iMNs quantified from three biologically independent iMN conversions per line. Two-sided log-rank test. Statistical significance calculated using the entire survival course (days 1-21). 202 Figure S4. 5 Gain-of-function mechanism of C9ORF72 ALS/FTD neurons drives the conversion of C9ORF72 ALS/FTD microglia to neuroprotective state (A) Kaplan–Meier survival curves of control (CTRL) iMNs with no iMG (black lines) or co-cultured with C9ORF72 ALS/FTD (C9-ALS) MG with (blue line) or without (red line) conditioned media (CM) from C9 iMNs. n=879 iMNs for CTRL iMNs+ no iMG, n=1319 for CTRL iMNs+C9 iMG and n= 3622 for CTRL iMNs+C9 iMG+C9 iMNs CM. n=2 lines/genotype for iMNs and iMG. iMNs quantified from three biologically independent iMN conversions per line. Two-sided log-rank test. Statistical significance calculated using the entire survival course (days 1-18). (B) Kaplan–Meier survival curves of control (CTRL) iMNs with no iMG and C9ORF72 -/- (C9-KO) iMNs with no iMG, C9-KO iMNs with no iMG or co-cultured with CTRL or C9ORF72 ALS/FTD (C9) iMG. n=879 iMNs for CTRL iMNs+ no iMG, n=236 for C9-KO iMNs+ no iMG, n= 247 for C9-KO iMNs+CTRL iMG, and n= 257 for C9-KO iMNs+C9 iMG. n=2 lines for control iMNs, n=1 line for C9-KO iMNs, and n=2 line/genotype for iMG. iMNs quantified from three biologically independent iMN conversions per line. Two-sided log-rank test. Statistical significance calculated using the entire survival course (days 1-18). (C) Kaplan–Meier survival curves of control (CTRL) iMNs expressing GFP with no iMG and CTRL iMNs expressing GFP-GR 50 with no iMG or co-cultured with CTRL or C9ORF72 ALS/FTD (C9) iMG. n=317 iMNs for CTRL-GFP iMNs, n=155 for CTRL-GFP-GR50 iMNs+ no iMG, n= 245 for CTRL-GFP-GR50 iMNs+CTRL iMG, and n= 242 for CTRL-GFP-GR 50 iMNs+C9 iMG. n=2 line/genotype for iMNs and iMG. iMNs quantified from three biologically independent iMN conversions per line. Two-sided log-rank test. Statistical significance calculated using the entire survival course (days 1-18). 203 Figure S4. 6 C9ORF72 ALS/FTD microglia shift from CSF1R-high to CSF1R-low neuroprotective state in vitro. (A) mRNA levels (relative to GAPDH) of CSF1R and IGF1 in C9ORF72 ALS/FTD (C9-ALS) iMG treated with DMSO, 10 nM, 100 nM or 1 μM of PLX3397 (PLX) for 3 days. n=6 biological replicates 204 (independent conversions) per condition from two C9-ALS/FTD patients. Ordinary one-way ANOVA. Mean ± s.e.m. iMG were FACS-purified by CD11B antibody. (B) Violin plots of CSF1R, NR1H3 and MEF2C between CSF1R-high iMG (cluster 6 from C9 iMNs+C9iMG) and CSF1R-low iMG (cluster 7 from C9 iMNs+C9iMG). (C) mRNA levels (relative to GAPDH) of MEF2C, CSF1R and IGF1 in C9ORF72 ALS/FTD (C9- ALS) iMG generated from 2 transcription factors (2TF, Pu.1 and CEBPA) or 3 transcription factors (3TF, Pu.1, CEBPA, MEF2C). n=4 biological replicates (independent conversions) per condition from two C9-ALS/FTD patients. Ordinary one-way ANOVA. Mean ± s.e.m. iMG were FACS- purified by CD11B antibody. (D) t-SNE plot based on the binary regulon activity matrix of CSF1R-high iMG (cluster 6 from C9 iMNs+C9iMG) and CSF1R-low iMG (cluster 7 from C9 iMNs+C9iMG) after applying SCENIC analysis. (E) Dot plots of relative gene expressions for candidate receptors and ligands between (Top) control (CTRL) and C9ORF72 ALS/FTD (C9-ALS) neurons and (Bottom) CSF1R-high iMG (cluster 6 from C9 iMNs+C9iMG) and CSF1R-low iMG (C9 iMNs+CTRL iMG and cluster 7 from C9 iMNs+C9iMG). (F) Kaplan–Meier survival curves of C9ORF72 ALS/FTD (C9) iMNs with no iMG (blue lines) or co-cultured with C9 iMG treated without or with 4 µg/mL of neutralizing antibodies (NuAb) (JAG1, INHBA, IGF2, SLIT3, TIMP2, IGF1 and TGFB1). n=118 for C9 iMN+no iMG, n=128 for C9 iMN+C9 iMG, n=112 for C9 iMN+C9 iMG+JAG1 NuAb, n=120 for C9 iMN+C9 iMG+INHBA NuAb, n=107 for C9 iMN+C9 iMG+IGF2 NuAb, n=102 for C9 iMN+C9 iMG+SLIT3 NuAb, n=124 for C9 iMN+C9 iMG+TIMP2 NuAb, n=69 for C9 iMN+C9 iMG+IGF1 NuAb and n=69 for C9 iMN+C9 iMG+TGFB1 NuAb. n=1 line/genotype for iMNs and iMG. iMNs quantified from three biologically independent iMN conversions per line. Two-sided log-rank test. Statistical significance calculated using the entire survival course (days 1-11). 205 Figure S4. 7 Validation of the presence of CSF1R-low microglia in vivo. (A) Lanency to fail (second) from hanging wire test for 9-month-old wild-type (WT) and C9orf72 - /+ ;C9BAC mice. Unpaired t-test. Mean ± s.e.m. (B) Representative images of poly(GP) punctae in TUJ1+ motor neurons in the motor cortex from the brain of 9-month-old wild-type (WT) and C9orf72 -/+ ;C9BAC mice without treatment. Dotted lines outline the cell body. Neuron cell bodies were identified using TUJ1 immunostaining. Scale bar = 10 μm. (C) Schematic of treating C9orf72 -/+ ;C9BAC mice with control (CTRL) or PLX3397 (PLX) chow for 35 days. 206 (D) Quantification of the number of IBA1+ microglia per μm 2 in the motor cortex of the brain from 10-month-old WT mice treated with PLX3397 (PLX) chow for 0, 3, 5 and 7 days. n=6 mice per group. Ordinary one-way ANOVA. Mean ± s.e.m. (E) Representative images of IBA1+ microglia in the motor cortex from the brain of 9-month-old C9orf72 -/+ ;C9BAC mice treated with control (CTRL) chow or PLX3397 (PLX) chow for 35 days. Scale bar = 5 μm. Magnification of the area indicated by an orange rectangle. (F) Representative images of human cell-specific marker, STEM121, staining in the brain without or with transplantation of human iPSC-derived microglia. Left: non-injected side of the brain; Right: injected side of the brain.Red arrow points to the injected site. (G) Lanency to fail (second) from hanging wire test (shown in relative ratio before treatment) from 10-month-old wild-type C9orf72 -/+ ;C9BAC mice injected with CSF1R-high (n-5) or CSF1R-low (n- 6) C9ORF72 ALS/FTD microglia for 3 days. Unpaired t-test. Mean ± s.e.m. (H) Representative images and quantification of NeuroTrace (Nissl)-positive lateral motor column (LMC) spinal motor neurons in the lumbar spinal cord in 10-month-old wild-type mice injected with CSF1R-high (n=4) or CSF1R-low (n=3) control microglia, or C9orf72 -/+ ;C9BAC mice injected with CSF1R-high (n=4) or CSF1R-low (n=6) C9ORF72 ALS/FTD microglia for 3 days. The LMC motor neuron region in the spinal cord ventral horn is marked with a white dotted-line. Scale bar = 50 μm. 4-6 ventral horn sections were quantified and averaged per mouse. Ordinary one-way ANOVA. Mean ± s.e.m. (I) Quantification of poly(GP) punctae in TUJ1+ motor neurons in the motor cortex from the brain of 10-month-old WT mice (shown in gray) injected without or with CSF1R-low (n=6) control microglia, C9BAC mice (shown in red) injected without or with CSF1R-low (n=6) C9ORF72 ALS/FTD microglia, or C9orf72 -/+ ;C9BAC mice (shown in blue) injected without or with CSF1R- low (n=6) C9ORF72 ALS/FTD microglia for 3 days. Ordinary one-way ANOVA. Mean ± s.e.m. The injected and non-injected data points are from two different sides of the motor cortex of the same animals. (J) Quantification of poly(GP) punctae in TUJ1+ motor neurons in the motor cortex from the brain of 10-month-old WT mice (shown in gray) injected without or with CSF1R-high (n=7) control microglia, C9BAC mice (shown in red) injected without or with CSF1R-high (n=4) C9ORF72 ALS/FTD microglia, or C9orf72 -/+ ;C9BAC mice (shown in blue) injected without or with CSF1R- high (n=5) C9ORF72 ALS/FTD microglia for 3 days. Ordinary one-way ANOVA. Mean ± s.e.m. The injected and non-injected data points are from two different sides of the motor cortex of the same animals. 207

Abstract (if available)

Abstract Amyotrophic lateral sclerosis (ALS) is a detrimental neurogenerative diseases characterized by severe loss of motor neurons in the brain and spinal cord. In ALS patients, the motor neurons degenerate rapidly after disease onset, which leads to lose the ability to swallow, speak, breathe, and eventually causes death. However, there is no current cure for ALS. One reason is that ALS has diverse genetic etiology with up to 90% of cases are sporadic, caused by unknown mutations. Another reason is that more studies have shown that non-neuronal cells in the brain, such as microglia, play an important role in regulating neurodegeneration. To identify efficacious therapeutics for ALS, it is in a pressing need to determine the shared disease mechanisms between familial and sporadic ALS and to investigate the role of non-neuronal cells in ALS. In this dissertation, I present our efforts to identify three therapeutic targets for motor neurons and microglia in ALS, which provides effective therapeutic approach for multiple forms of ALS.

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

Functional study of C9ORF72 and its implication in the pathogenesis of amyotrophic lateral sclerosis

PDF

Identification of therapeutic targets in human cerebral brain organoid models of neurodegeneration

PDF

The C9ORF72 repeat expansion impairs neurodevelopment

PDF

Trichostatin A and dihydrotestosterone as potential multi-systemic drugs for amyotrophic lateral sclerosis

PDF

Neural crest-derived cranial pericytes and their dysfunction in disease

PDF

Understanding human nephrogenesis and scaling synthesis of organoids facilitate modeling of kidney development and disease

PDF

Functional study of Snord118 in microglila

PDF

Signaling networks in complex brain disorders

PDF

Investigating molecular roadblocks to enhance direct cellular reprogramming

PDF

The role of vascular dysfunction in cognitive impairment

PDF

Dissecting the heterogeneity of mouse hematopoietic stem cells in vivo

PDF

Transcriptomic maturation of developing human cone precursors in fetal and 3D hESC-derived tissues

PDF

Targeting Ryk-signaling to attenuate multiple sclerosis by promoting oligodendrocyte differentiation and myelination

PDF

Genetics and the environment: evaluating the role of noncoding RNA in autism spectrum disorder

PDF

The autism-associated gene SYNGAP1 regulates human cortical neurogenesis

PDF

The role of ERK1/2 in mouse embryonic stem cell fate control

PDF

Blood-brain barrier pathophysiology in cognitive impairment and injury

PDF

Enabling Alzheimer’s Aβ phagocytosis by targeting IRAK-M innate immunity

PDF

Molecular basis of mouse epiblast stem cell and human embryonic stem cell self‐renewal

PDF

Molecular signatures underlying intercellular differences in leukemia progression and chemotherapy response

Asset Metadata

Creator Hung, Shu-Ting (author)

Core Title Identification of therapeutic targets for neurons and microglia in amyotrophic lateral sclerosis

School Keck School of Medicine

Degree Doctor of Philosophy

Degree Program Development, Stem Cells and Regenerative Medicine

Degree Conferral Date 2022-12

Publication Date 04/25/2024

Defense Date 10/13/2022

Publisher University of Southern California (original), University of Southern California. Libraries (digital)

Tag amyotrophic lateral sclerosis,C9ORF72,microglia,motor neurons,OAI-PMH Harvest

Format theses (aat)

Language English

Contributor Electronically uploaded by the author (provenance)

Advisor Ying, Qi-Long (committee chair), Bonaguidi, Michael (committee member), Ichida, Justin (committee member), Zlokovic, Berislav (committee member)

Creator Email michellehungsth@gmail.com,shutingh@usc.edu

Permanent Link (DOI) https://doi.org/10.25549/usctheses-oUC112195799

Unique identifier UC112195799

Identifier etd-HungShuTin-11284.pdf (filename)

Legacy Identifier etd-HungShuTin-11284

Document Type Dissertation

Format theses (aat)

Rights Hung, Shu-Ting

Internet Media Type application/pdf

Type texts

Source 20221026-usctheses-batch-988 (batch), University of Southern California (contributing entity), University of Southern California Dissertations and Theses (collection)

Access Conditions The author retains rights to his/her dissertation, thesis or other graduate work according to U.S. copyright law. Electronic access is being provided by the USC Libraries in agreement with the author, as the original true and official version of the work, but does not grant the reader permission to use the work if the desired use is covered by copyright. It is the author, as rights holder, who must provide use permission if such use is covered by copyright. The original signature page accompanying the original submission of the work to the USC Libraries is retained by the USC Libraries and a copy of it may be obtained by authorized requesters contacting the repository e-mail address given.

Repository Name University of Southern California Digital Library

Repository Location USC Digital Library, University of Southern California, University Park Campus MC 2810, 3434 South Grand Avenue, 2nd Floor, Los Angeles, California 90089-2810, USA

Repository Email cisadmin@lib.usc.edu