University of Southern California Dissertations and Theses

Patch aRNA in vitro amplification (PAIA): single cell RNA-seq to expand the understanding of the developing and developed nervous system

(USC Thesis Other)

Patch aRNA in vitro amplification (PAIA): single cell RNA-seq to expand the understanding of the developing and developed nervous system

PDF

Download

Open document

Flip pages

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content i PATCH aRNA IN VITRO AMPLIFICATION (PAIA): SINGLE CELL RNA-SEQ TO EXPAND THE UNDERSTANDING OF THE DEVELOPING AND DEVELOPED NERVOUS SYSTEM by Jae Mun Kim ______________________________________________________________ A Dissertation Presented to the FACULTY OF THE GRADUATE SCHOOL UNIVERSITY OF SOUTHERN CALIFORNIA In Partial Fulfillment of the Requirements for the Degree DOCTOR OF PHILOSOPHY NEUROSCIENCE December 2017 ii Epigraph “Freedom of thought is best promoted by the gradual illumination of men’s minds which follows from the advance of science.” ― Charles Darwin iii Acknowledgements Firstly, I would like to express my sincere gratitude to my advisor Prof. James A. Knowles for the continuous support of my PhD study: for his patience, motivation, and immense knowledge. His guidance was a constant source of help, particularly because I had to change labs in my second year of PhD training. When we encountered challenges developing an effective protocol for single cell RNAseq, his guidance helped me select a new protocol and adapt it into its current iteration. I could not have made it here without his mentorship. Besides my advisor, I would like to thank the rest of my thesis committee: Prof. Robert Chow, Prof. Jeannie Chen, and Prof. Scott Fraser, for their insightful comments and encouragement, but also for the challenges they gave me, which my research from various perspectives. My sincere thanks also go to Prof. Oleg Evgrafov, who educated and encouraged me when I had a difficult time figuring out the NuGEN protocol and encountered other problems with single cell RNAseq. I still remember the joy when my first qPCR worked using my own primers with Oleg’s support. Without his essential support, it would not have been possible for me to conduct this research project. iv I thank my fellow lab mates, Adrian Camarena, Edder Lopez, Chris Walker and everyone else for the stimulating discussions, for the sleepless nights when we were working toward NIH deadlines, and for all the fun we have had in the last five years. I also would like to thank Sonia Ming Yi Lin and Victory Woesly for their awesome contributions on the electrophysiological recordings and cell collections. Last but not the least, I would like to thank my family, specifically my lovely wife Suhn Kyong Rhie. She has provided moral and emotional support throughout my life, including during the course of writing this thesis. v Table of Contents Epigraph .......................................................................................... ii Acknowledgements ............................................................................ iii List of Tables ................................................................................... xi List of Figures ………………………………………………………………….….xiv Abstract ......................................................................................... xvi Chapter 1: Single Cell RNAseq Method Development ................... 1 1.1 Introduction ......................................................................................... 1 1.1.1 Overview .......................................................................................... 1 1.1.2 qPCR, Microarray and RNAseq........................................................ 2 1.1.3 Single Cell RNAseq .......................................................................... 3 1.2 Materials and Methods ............................................................................ 7 1.2.1 Overview of Initial Comparison of Three Methods ............................ 7 1.2.2 SMARTer Ultra Low input RNA V4 with NEXTera ............................ 7 1.2.3 NuGEN Ovation RNA-Seq System V2 ........................................... 11 1.2.4 aRNA amplification ......................................................................... 12 1.2.5 Method Modification from the NuGEN kit ....................................... 16 1.2.6 Method Modification from the aRNA amplification method ............. 20 vi 1.2.7 Method PAIA (Patch-aRNA in vitro transcription amplification) Extraction of cytoplasm using Patch Clamp technique ............................ 21 1.2.8 Alignment, Quality Control and Data analysis ................................ 23 1.3. Results ................................................................................................. 24 1.3.1 NuGEN RNAseqV2 Method Modification to process 10pg of RNA 24 1.3.2 aRNA Method and Modification to process 10pg ........................... 26 1.3.3 Comparison of Methods (SMARTer, NuGEN, aRNA) with UHR 10pg ................................................................................................................ 29 1.3.5 PAIA application ............................................................................. 31 1.3.6 Single cell data from PAIA comparing with public dataset .............. 32 1.4 Discussion ............................................................................................ 33 1.4.1 Additional Modifications .................................................................. 34 1.4.2 What caused the improvement? ..................................................... 35 1.4.3 Future direction of PAIA ................................................................. 36 1.5 Figures .................................................................................................. 38 1.6 Tables ................................................................................................... 49 Chapter 2. Identification and Characterization of Cajal Retzius Neurons ............................................................................. 55 2.1 Introduction ........................................................................................... 55 vii 2.1.1 Overview ........................................................................................ 55 2.1.2 Role of Cajal-Retzius cells ............................................................. 56 2.1.3 Approaches .................................................................................... 58 2.2 Methods & Materials ............................................................................. 59 2.2.1 Brain slice preparation .................................................................... 59 2.2.2 Electrophysiological recording ........................................................ 60 2.2.3 Data Analysis of electrophysiological recording ............................. 61 2.2.4 Single Cell RNAseq ........................................................................ 61 2.2.5 Data Analysis of single cell RNAseq .............................................. 62 2.3 Results .................................................................................................. 63 2.3.1 Identification of human Cajal-Retzius neurons ............................... 63 2.3.2 Spontaneous synaptic activity in human Cajal-Retzius neurons .... 65 2.3.3 Human Cajal-Retzius neurons fires repetitively .............................. 65 2.3.4 Single-cell transcriptomic analysis of human Cajal-Retzius neurons ................................................................................................................ 67 2.4 Discussion ............................................................................................ 69 2.4.1 Spontaneous synaptic activity in human Cajal-Retzius neurons .... 70 2.4.2 Human Cajal-Retzius neurons express secreting and contacting cues ......................................................................................................... 71 viii 2.4.3 Cell-adhesion molecules are enriched in human Cajal -Retzius neurons ................................................................................................... 71 2.4.4 Wnt signaling molecules in human Cajal-Retzius neurons ............. 73 2.4.5 Calcium signaling in human Cajal-Retzius neurons .................... 74 2.4.6 Differences between human and rodent Cajal-Retzius neurons .... 75 2.5 Figures .................................................................................................. 77 2.6 Tables ................................................................................................... 87 Chapter 3. Subtypes and characterization of Embryonic Spinal Cord Neurons ............................................................................ 113 3.1 Introduction ......................................................................................... 113 3.2 Methods & Materials ........................................................................... 114 3.2.1 Spinal Cord slice preparation ....................................................... 114 3.2.2 Electrophysiological recording ...................................................... 115 3.2.3 Data Analysis of electrophysiological recording ........................... 115 3.2.4 Single Cell RNAseq ...................................................................... 115 3.2.5 Data Analysis of single cell RNAseq ............................................ 116 3.3 Result .................................................................................................. 117 3.3.1 Region specific gene expression differences ............................... 117 3.3.2 Interneuron Subtypes ................................................................... 120 ix 3.3.3 Motor Neuron Subtypes with Firing Pattern .................................. 121 3.4 Discussion .......................................................................................... 124 3.4.1 Sensory and Motor Neuron Markers ............................................ 124 3.4.2 Interneuron subtypes .................................................................... 125 3.4.3 Motor Neuron: Electrophysiological profile and subtypes ............. 126 3.5 Figures ................................................................................................ 128 3.6 Tables ................................................................................................. 135 Chapter 4. Electrophysiological pattern and Gene expression pattern from Human Temporal Lobe Studies ............................. 147 4.1 Introduction ......................................................................................... 147 4.1.1 Overview ...................................................................................... 147 4.1.2 Layer Specific Expression Profile of Neurons .............................. 149 4.1.3 Electrophysiological Profile and Subtypes .................................... 151 4.2 Materials and Methods ........................................................................ 151 4.2.1 Tissue preparation ........................................................................ 151 4.2.2 Collection, Ephys and RNA sequencing ....................................... 152 4.2.3 Data processing and analysis ...................................................... 152 4.3 Result .................................................................................................. 154 4.3.1 Layer Specific Expression on temporal lobe pyramidal neurons .. 154 x 4.3.2 Biomarkers and known marker analysis ....................................... 157 4.3.3 Electrophysiological pattern of temporal lobe cells ....................... 158 4.4 Discussion .......................................................................................... 159 4.4.1 Pyramidal Neurons and layers, Biomarkers ................................. 159 4.4.2 Electrophysiological pattern ......................................................... 160 4.5 Figures ................................................................................................ 163 4.6 Tables ................................................................................................. 171 Summary & Conclusion ................................................................... 188 References ................................................................................... 193 xi List of Tables Table 1-1. Comparison between qPCR, Microarray and RNAseq 49 Table 1-2. Comparison of original aRNA and modified aRNA with 4 metrics. 50 Table 1-3. Comparison of Commercial methods with the modified aRNA with Four metrics. 51 Table 1-4. Number of Palindromic unmapped reads in samples. 52 Table 2-1. List of All embryinic brain(EB) cells Used for analysis 87 Table 2-2. Distributions of the morphology of Cajal Retzius neurons 108 Table 2-3. Top 100 differentially expressed genes between Cajal Retzius and other EB cells 109 Table 2-4. Gene Ontology (GO) analysis 112 Table 3-1. Top 30 Genes differentially expressed between Motor Neuron and Dorsal Horn. 135 Table 3-2. Gene Set Enrichment Analysis of genes differentially expressed between motor and dorsal horn interneurons. 136 Table 3-3. Top 30 Genes differentially expressed between Motor neuron and sensory neuron from dorsal root ganglia (DRG). 137 Table 3-4. GSEA of the differential expressed genes between motor neurons from ventral horn and sensory neurons from DRG. 138 Table 3-5. Top 30 genes differentially expressed between DRG sensory neurons and DH interneurons. 139 xii Table 3-6. GSEA on the differentially expressed gene between DRG sensory neurons and DH interneurons. 140 Table 3-7. Top 30 genes differentially expressed between Intermediate zone interneurons and dorsal horn interneurons. 141 Table 3-8, GSEA of the differentially expressed genes between interneurons from intermediate area and the dorsal horn. 142 Table 3-9. Repetitive adaptive VS repetitive non adaptive action potential firing patterned motor neurons. 143 Table 3-10. Repetitivs adaptive VS single action potential patterned motor neurons. 144 Table 3-11. Repetitive non adaptive VS single action potential patterned motor neurons. 145 Table 3-12. Potential Biomarkers of motor neurons with 3 different electrophysiological patterns. 146 Table 4-1. Temporal lobe cells collected by layers. 171 Table 4-2. Layer I VS Layer II DESeq analysis 172 Table 4-3. Gene Set Enrichment Analysis (GSEA) of differentially expressed genes between Layer I and II. (top) 173 Table 4-4. Layer II VS Layer III Differentially expressed genes 174 Table 4-5. GSEA of differentially expressed genes between Layer II and III. 175 Table 4-6. Layer III VS Layer IV Differentially expressed genes 176 xiii Table 4-7. GSEA of differentially expressed genes between Layer III and IV. 177 Table 4-8. Layer IV VS Layer V Differentially expressed genes 178 Table 4-9. GSEA of differentially expressed genes between Layer IV and V. 179 Table 4-10. Layer V VS Layer VI Differentially expressed genes 180 Table 4-11. GSEA of differentially expressed genes between Layer V and VI. 181 Table 4-12. Number of Differentially expressed genes 182 Table 4-13. Number of up and downregulated potential biomarkers. 183 Table 4-14. GSEA on potential biomarkers on each layer. 184 Table 4-15. Repetitive Adaptive VS Non-Adaptive differentially expressed genes. 186 Table 4-16. GSEA on Repetitive Adaptive VS Non-Adaptive differentially expressed genes. (Gene sets enriched for the both up and downregulated gene lists from comparison Adaptive vs. Non-adaptive neurons.) 187 xiv List of Figures Figure 1-1. Schematic of NuGEN RNAseq V2 protocol. 38 Figure 1-2. Basic Schematic of the aRNA amplification method to amplify RNA. 39 Figure 1-3. Schematics of PAIA protocol 40 Figure 1-4. NuGEN library input comparisons. 41 Figure 1-5. NuGEN Ultralow vs Rapid Library 42 Figure 1-6. Limitation of NuGEN RNAseq V2 Kit 43 Figure 1-7. aRNA Method Modification steps and changes in correlation 44 Figure 1-8. aRNA yield measured by Tapestation 45 Figure 1-9. Correlation and quality of UHR data. 46 Figure 1-10. Tissue Collected from embryonic brain 47 Figure 1-11. Comparison of Single cell data 48 Figure 2-1. Norepinephrine’s regulation of cortical layer formation. 77 Figure 2-2. Microscopic image of the human embryonic brain tissue 78 Figure 2-3. Morphological variation in CR neurons 79 Figure 2-4. Spontaneous Activity observation of Cajal-Retzius Neurons 80 Figure 2-5. Firing pattern with current injection in CR, CP and SP neurons 81 Figure 2-6. Sources of technical and biological noise 83 Figure 2-7. PCA plot of covariates 84 Figure 2-8. Boxplot of normalized expression of known marker genes for human Cajal Retzius Cell. 85 xv Figure 2-9. Heatmap of the differentially expressed genes 86 Figure 3-1 Embryonic Spinal Cord development 128 Figure 3-2. Morphologies of spinal cord neurons 129 Figure 3-3. Distribution of collected samples from Embryonic Spinal Cord (ES) 130 Figure 3-4. Motor Neuron with Lucifer Yellow Staining 131 Figure 3-5. Volcano plot of Differentially expressed genes 132 Figure 3-6. Three Firing patterns of motor neurons. 133 Figure 3-7. Heatmaps of some differential expressed gene in repetitive adaptive, non-adaptive and single action potential firing neurons 134 Figure 4-1. Six layers of cerebral cortex. 163 Figure 4-2. Diversity of layer VI pyramidal neurons and projections 164 Figure 4-3. Distribution of number of cell types collected. 165 Figure 4-4. Morphology of pyramidal neurons from different layers 166 Figure 4-5. Normalized gene expression of the potential biomarkers. 167 Figure 4-6. Two electrophysiological patterns in Temporal lobe. 168 Figure 4-7. Boxplot of differentially expressed genes between repetitive adaptive vs non adaptive neurons. 169 Figure 4-8. Expression of markers for neurons based on neurotransmitter 170 xvi Abstract The genome encodes information for all the different cell types in the body. For any cell type, only a fraction of the genome is transcribed to messenger RNA (mRNA). Recently, it has become possible to sequence RNA in the minute ( ≤ 10 pg) amounts found in single cells (scRNAseq). Preliminary analysis of single-cell RNA shows considerable cell-to-cell heterogeneity, even among ostensibly identical cells, raising the question: how much of the variability is true biological variability, and how much is technical noise. The first part of my thesis describes how I analyzed scRNAseq methodology and implemented improvements. Subsequent chapters detail the application of the improved scRNAseq methodology to neurons of human adult and embryonic brain. Currently, most single cell sequencing techniques entail separating tissue into individual cells and the lysis of individual, isolated cells, followed by the immediate collection of released nucleotides. Automation, using microfluidics platforms, has enabled high throughput; however, the separation of tissues into individual cells leads to loss of information about cell morphology (particularly among neurons, which typically have complex neuritic extensions that are lost in cell separation) and may alter the cellular transcriptome, owing to the traumatic treatment that cells undergo. xvii An alternate approach for collecting RNA from single cells is to re- purpose the electrophysiological technique of patch clamp recording. A hollow patch pipette is attached to individual cells, enabling the recording of cellular electrical activity, after which the cytoplasm may be extracted for scRNAseq. Since the tissue is not disaggregated before patch clamp is used, the morphology of the cells is maintained. Thus, using this approach, we can correlate single-cell transcriptomes with the cell morphology and electrophysiology. Other groups have previously published on what has been called “Patch-seq.” The drawback of the published approach is that the collected RNA (which averages 10 pg) is subjected to substantial, exponential amplification by PCR, which distorts the relative amounts of RNA species (it leads to under-representation of rarer RNA species). Amplification of RNA is necessary for single cell RNAseq, as current RNA sequencing approaches require nanogram amounts of RNA. We chose to avoid PCR amplification and to use the linear amplication approach called “aRNA.” First introduced in the 1990’s by the Eberwine group, this approach uses T7 enzyme for in vitro transcription. As detailed in chapter 1, we introduced a number of changes to the originally published protocol to reduce the time required and to improve reproducibility. We then compared our in-house method, which we call Patch-aRNA in vitro transcription amplification (PAIA), with other approaches (the original aRNA-based and xviii PCR-based protocols). PAIA yields the highest transcriptome complexity (as defined by number of genes discovered) among the approaches tested. In order to investigate the unique niches occupied by different cell types in the central nervous system, we applied PAIA to human embryonic brain and spinal cord, and human adult brain. As described in Chapter 2, in embryonic brain we investigated Cajal-Retzius (CR) neurons, which are believed to play a major role in orchestrating the layering of the cortex. CR neurons have been relatively understudied in human samples. We compared their transcriptome profiles with those of other developing neurons. We found distinct subtypes of CR neurons, and we identified changes according to development stage. In Chapter 3, we used PAIA to help identify biomarkers of embryonic spinal cord neurons within specific areas: dorsal horn, ventral horn, dorsal root ganglia, and middle area. Moreover, we compared the subtypes of interneurons based on the area they were found. Additionally, using electrophysiological patterns, we defined the subtypes of motor neurons and explored differences in expression profiles between them. In Chapter 4, we characterized the difference in adult temporal lobe pyramidal neuron projections based on their locations (layers) and electrophysiological activities. Because pyramidal neurons project to different areas of the brain, their classification helps answer numerous biological questions, including target-specific expression pattern differences. xix scRNAseq and the new PAIA approach open the door to many address questions that were not possible to address previously. The comprehensive characterization of neurons in these three regionshelps expand the understanding of the roles individual neurons play in neuronal development. 1 Chapter 1: Single Cell RNAseq Method Development 1.1 Introduction 1.1.1 Overview In the field of neuroscience, neuronal cells are typically characterized by three properties: morphology, electrophysiology, and expression (RNA, protein) profiles. First, almost every neuron possesses very distinct morphology, with many variations within the axon, dendrite, and soma. The variance in morphology helps define its function and connectivity (Cadwell et al., 2015; Luciano da F. Costa, Campos, Estrozi, Rios-Filho, & Bosco, 2000; Luciano da Fontoura Costa et al., 2002). Second, neurons have specific electrical activity, which can be measured by electrophysiological recording. In a neuronal organ such as the retina, M1 retinal ganglion cell subtype can be classified based on differences in the electrical activity differences (Schmidt, Chen, & Hattar, 2011). The last property is molecular identity. Techniques such as Fluorescent In Situ Hybridization (FISH) and immunohistochemistry can measure localization and expression levels of marker genes that are specifically transcribed in particular neurons. As technology has advanced, improved methods that can measure expression of multiple to all genes at once have been developed. 2 1.1.2 qPCR, Microarray and RNAseq The following three techniques are useful for the elucidation of neuronal gene expression: qPCR (quantitative Polymerase Chain Reaction), microarray, and RNA sequencing. In all three methods, cDNA synthesis with primers (either oligo dT or random hexamers) should be performed in advance (Wang, Gerstein, & Snyder, 2009a). qPCR can detect the expression of specific genes. The Primers or fluorescently attached probes are bound to designated exons of a specific gene in order to measure quantity of RNA during the amplification process. The SYBR green assay can only detect one gene per reaction while the Taqman assay, which utilizes probes with fluorescence, can target up to three genes using a tri-fluorescence measurement device. Recently, the Fluidigm qPCR device (PN 68000088 K1) was used to characterize more than 48 gene targets simultaneously within a single cell (Citri, Pang, Südhof, Wernig, & Malenka, 2011). However, measuring the entire transcriptome from a tissue or a single cell is not possible with this technique. Microarray technology enables an increased number of known transcripts to be assessed. By using array of probe sets, all known transcripts can be measured in the array simultaneously (Hoheisel, 2006). By designing the locations of probes across genes, it is also possible to limit the genes which can be measured at once. The use of different types of probes alongside specialized arrays enable different aspects of gene expression, 3 such as exons and splicing, to be characterized. However, because of high background noise and limitations on novel transcripts and splice variants, it has almost been entirely replaced with the newer RNAseq technology. RNA sequencing (RNA seq) has greatly expanded the molecular characterization of gene expression to the whole-transcriptome level (Wang, Gerstein, & Snyder, 2009b). It detects all signals of known and unknown transcripts with much less background noise. Unlike the other two methods, which are based on primer hybridization, RNA seq is based on sequencing of the whole transcriptome. Thus, RNA seq provides expression levels of almost every gene expressed in the sample. The comparison between qPCR, RNAseq, and microarray is described in Table 1-1. 1.1.3 Single Cell RNAseq Most RNAseq protocols are not able to profile expression levels in a single cell, limiting our ability to investigate cellular heterogeneity (Qiu et al., 2012). We, as well as other researchers (Wu et al., 2014), have been developing more sensitive protocols, which can characterize different types of neuronal cells by profiling expression at the single-cell level. There are currently three methods which are commonly used to perform RNA-sequencing with a single cell: the NuGEN Ovation RNA-Seq System V2 (part number 7102; NuGEN Technologies Incorporated), the SMARTer Ultra Low input RNA (part number 634935) and the aRNA amplification kit (Moll, Duschl, & Richter, 2004; Morris, Singh, & Eberwine, 2011; Phillips & Eberwine, 4 1996). The NuGEN RNA-seq V2 method uses DNA linear amplification, the SMARTer method uses PCR amplification, and the aRNA amplification uses linear RNA in vitro transcription (IVT), also known as linear RNA amplification. However, these methods require one to follow a complex protocol thatis labor- intensive and materially inefficient. Because these currently available methods produce RNAseq data that is hindered by amplification bias, there is a great demand to improve these protocols and develop a quick and easy single cell RNAseq method. Furthermore, the methods used to collect a single neuronal cell for single cell RNAseq can be improved. In most of such studies, single cells are isolated by dissociation and sorting such as via fluorescence-activated cell sorting (FACS) (Jaitin et al., 2014; Morrison, Box, McKinney, McLennan, & Kulesa, 2015) or individual tagging (Hashimshony, Wagner, Sher, & Yanai, 2012; Klein et al., 2015). For example, the Drop-seq and Fluidigm methods were performed using the dissociation method. In Drop-seq, separated cells are labeled with a specific barcode and separated as microfluidic bubbles in oil (Klein et al., 2015). The Fluidigm method used a microfluidic chamber to separate cells and then create cDNA and RNA libraries of the lysate collected from each chamber. However, the dissociation method has several limitations for neurological research. Dissociation destroys cell connectivity and does not allow for electrophysiological and morphological data to be collected, which provide indispensable information for the study of neuronal heterogeneity. 5 Moreover, cell transcription profiles are altered during the dissociation process, introducing a concerning source of confounding (Cadwell et al., 2015; Zeisel & Linnarsson, 2014). Unlike methods to isolate a single cell with dissociation, sorting, or tagging, collection using a patch clamp affords several benefits to neuronal studies. Using a patch clamp preserves cellular morphology and electrophysiology and does not alter the transcription profile of the cell. A transcriptome protocol based on patch clamp collection of single cell material (Patch-seq) has been recently developed (Cadwell et al., 2015). However, the Patch-seq approach amplifies cDNA using the SMARTer method based on PCR amplification. The exponential amplifications that occur from PCR are prone to bias, especially when the number of cycles is high (Aird et al., 2011). Therefore, development of a better single cell RNAseq method that can reduce amplification bias and while obtaining the morphology, electrophysiology, and expression profile of single cell is in great demand. In order to develop a better single cell RNAseq method, we first compared three currently available single cell RNAseq methods and demonstrated that in our experimental conditions, aRNA amplification was the most sensitive and robust technique. By modifying the aRNA amplification approach, we developed a new method called patch-aRNA in vitro transcription amplification (PAIA). 6 PAIA uses the aRNA method with a reduced number of linear amplification rounds. This modified aRNA method permits 1) higher mapping rate, 2) greater transcription profile complexity as measured by the number of genes discovered in the same number of exonic reads, 3) higher correlation with bulk RNA, and 4) better reproducibility. When we performed direct comparison of cells processed from PAIA with the same type of cells processed using other methods, PAIA showed higher transcriptomic complexity of amplified material, when compared with the original aRNA method. In addition, cellular contents are collected using a patch clamp, which provides electrophysiological recordings from an intact single cell. We applied PAIA to thousands of neurons and other nervous system cells to characterize their distinct gene expression patterns, including Cajal- Retzius neurons, sensory and motor neurons, interneurons, and neurons in the spinal cord. For example, we found the subtypes of Cajal-Retzius neurons possessed unique action potential patterns that were associated with expression changes in genes involved in the Reelin pathway. PAIA method enables neuronal transcription profiles of single cells to be studied alongside their morphological and electrophysiological features, which will benefit researchers attempting to characterize heterogeneous populations of cells. 7 1.2 Materials and Methods 1.2.1 Overview of Initial Comparison of Three Methods The collected materials from a single neuron are around 10pg of RNA on average, which is far less than the amount of nucleic acids required for sequencing. Therefore, an amplification step is required prior to the sequencing step. The ideal amplification will avoid artificial bias coming from a massive PCR amplification as well as minimize the loss of materials. To minimize the bias caused by exponential growth, linear amplification of DNA or RNA is utilized (Hafner, Yang, Wolter, Stafford, & Giffard, 2001; Liu, Schreiber, & Bernstein, 2003). For initial comparison, we used three methods, 1) SMARTer Ultra Low input RNA (part number 634935), which uses PCR amplification, 2) NuGEN Ovation RNA-Seq System V2 (part number 7102; NuGEN Technologies Incorporated), which uses DNA linear amplification, and 3) the aRNA amplification kit (Moll et al., 2004; Morris et al., 2011; Phillips & Eberwine, 1996), which uses RNA amplification. 1.2.2 SMARTer Ultra Low input RNA V4 with NEXTera The SMARTer Ultralow RNA method was performed by following the protocols given from the Clontech with a few optimizations. The protocols we used are briefly described below. For example, after testing the 10pg method, we found that the cDNA generated from 18 cycles of PCR was not enough for library prep, so 3 more cycles were added. 8 1. On cold block, the Universal Human Reference RNA sample was thawed and diluted into 10pg/μl concentration. (For bulk control, 1ng/μl) 1 μl of the diluted sample was added to 4 μl of nuclease free water to make 5 μl of initial volume. 2. All of the first-strand cDNA synthesis reagents (5X Ultra Low First- Strand Buffer, 3’ SMART-Seq CDS Primer II A (12 μM), 5X Ultra Low First- Strand Buffer, SMART-Seq v4 Oligonucleotide (48 µM)) were thawed on ice. Because the 5X Ultra Low First-Strand Buffer has precipitates, it was vortexed several times in room temperature to suspend them then placed it on ice. 2. The stock reaction buffer was made by mixing 19 μl 10X Lysis Buffer and 1 μl RNase Inhibitor, the stock reaction buffer was made. Then 1 μl of reaction buffer, 4.5 μl of nuclease free water, and 5 μl of sample were added to make 10.5 µl total volume. 3. 2 μl of 3’ SMART-Seq CDS Primer II A (12 μM) was added to this mixture, and tubes were then incubated at 72°C for 3 minutes. 4. During incubation, a master mix was made by mixing, per reaction, 4 μl of 5X Ultra Low First-Strand Buffer, 1 μl of SMART-Seq v4 Oligonucleotide (48 µM) and 0.5 μl of RNase Inhibitor (40 U/µl). When the 3 minute-incubation was finished, sample tubes were placed on ice for 2 minutes. 2 μl of the SMARTScribe Reverse Transcriptase per reaction was added to the master mix at this time. Then, 7.5 μl of the master mix was added to the sample mix. 9 After several mixings by pipetting up and down, the SMARTER1 program was used on thermocycler (42°C for 90 min, 70°C for 10 min, 4°C forever). 5. When the program neared completion, another master mix was generated with 25 μl of 2X SeqAmp PCR Buffer, 1 μl of PCR Primer II A (12 μM), 1 μl of SeqAmp DNA Polymerase, and 3 μl of Nuclease-Free water (per sample) to make 30 μl total volume. When the program finished, 30 μl of master mix was added to the samples and SMARTER2 program was started on thermocycler (95°C for 1 min, 21 cycles of: (98°C for 10 sec, 65°C for 30 sec, 68°C for 3 min), 72°C 10 min, 4°C forever). For 1ng sample, 14-cycle PCR was performed. 6. While waiting, AMPure XP beads were removed from 4°C storage , vortexed until evenly mixed, then left in room temperature so that the beads became room temperature. When the PCR program finished, 1 µl of 10X Lysis Buffer was added to each PCR product. Then, 50 μl of the AMPure XP beads were added to each sample and mixed the by pipetting the entire volume up and down at 10 times. 7. For 8 minutes, the sample-bead mix was incubated at room temperature to let the cDNA bind to the beads. After briefly spinning, the samples were placed on the magnet until the liquid became transparent and there were no beads left in the supernatant. 10 8. While the samples remained on the magnet, the supernatant was discarded by pipetting and 200 μl of freshly made 80% ethanol was added to each sample for washing. Two washing steps were performed at 30 seconds each. 9. After the washing step, the sample tube/plate was briefly spun down to discard any remaining ethanol on in the tube and samples were dried at room temperature for 2.5 minutes until the pellet became opaque. When the beads were dry, 17 μl of Elution Buffer was added to cover the bead pellet, and samples were removed from the magnet and mixed thoroughly. The Elution buffer-bead mixes were incubated at room temperature for 2 minutes for rehydration. 10. Once samples were rehydrated, they were briefly spun down and placed back on the magnet until the solution became transparent. 17 μl of each sample was then carefully removed and placed in a new nuclease-free, low-adhesion tube. 11. The quality and quantity of 1 μl of each sample was measured using Agilent 2200 Tapestation device. 12. 35 μl of nuclease free was then added to the cDNA and approximately 50 μl of the sample was transferred into a 100μl Covaris 50- tube(ube AFA Fiber & Cap 12x12mm (100)). Using the Covaris S2(S220) device, DNA shearing was performed: (Peak Power: 175, Duty:10%, Burst 11 Cycle:200, Time:5 min). After shearing, Tapestation 2200 was used once more to check the quantity and quality of shearing. Library prep was performed based on this quantity. 13. Nextera XT DNA Library Preparation Kit from Illumina was used for performing libraries preparation for these samples. 1.2.3 NuGEN Ovation RNA-Seq System V2 The NuGEN Ovation RNA-Seq System V2 method was performed by following the protocolsgiven by NuGEN. The original NuGEN RNAseq V2 protocol is composed of three different reactions: first strand, second strand and SPIA linear amplification (Figure 1-1). Initially, 10-100pg of UHR was tested with the original NuGEN ovation protocol to generate cDNA. Two NuGEN protocols (i.e. NuGen Ultralow and NuGen Rapid) were used: 1. For NuGEN-NuGEN Ultralow protocol was following “NuGEN RNAseq V2” protocol combined with Covaris shearing and NuGEN Ultralow library protocol. Both protocols are available at www.nugen.com. 2. For NuGEN-NuGEN Rapid protocol was following “NuGEN RNAseq V2” protocol combined with Covaris shearing and NuGEN Rapid library protocol. Both protocols are available at www.nugen.com. 12 Because the original NuGEN RNAseqV2 method result was not stable with a 10pg of initial material, the modified method explained below in 1.2.5 was used for the comparison. 1.2.4 aRNA amplification The aRNA protocol was used as previous studies described (Eberwine et al., 1992; Gentleman et al., 2004; Morris et al., 2011)e. The T7 promotor linked with Oligo dT and one anchoring base was used on the first strand synthesis for poly A selection. The second strand synthesis, which primes with nicking enzyme RNase H was followed. After bead purification, the material loss was minimal. With courtesy of Dr. Jim Eberwine from the University of Pennsylvania, the most recent version of the protocol was provided (Figure 1- 2). 1. On cold block, The Universal Human Reference RNA sample was thawed and diluted into 10pg /μl concentration. Take 1 μl of the diluted sample and add 4.9 μl of nuclease free water to make 5.9 μl of initial volume. 2. After thawing all reagents, the first strand master mix, per sample, composed of 2.4 μl of 5x First Strand buffer, 1.2 μl of dNTP’s (2.5 mM each), 0.3 μl of dT-T7 oligo (10ng/μl) and 0.45 μl of DTT (100mM) was made. After adding the master mix to the sample, Program JEPR1 (5 minutes at 70°C) was ran for removing secondary structures of RNA. When the program is done, samples were immediately placed on ice (cold block). 13 3. The first strand enzyme master mix composed of, per sample, 0.3 μl of Rnasin (RNAse inhibitor from Promega), 0.45 μl of Superscript III (Life technology), 1 μl of DEPC water, was made. Then, 1.75 μl of master mix was added to each sample, mixed by pipetting and span briefly. Then the program JEPR2 (Incubate 30 minutes at 42°C, Incubate 15 minutes at 70°C) was run. 4. At this stage, the protocol allowed samples either to be stored in -20 or -80°C or to be continued to next steps. For our samples, we always have gone through this part without stopping. 5. When the JEPR2 program was finished, second strand master mix was made on ice. Per sample, it is composed of 8.25 μl of RNase Free Water, 5.56 μl of 5X Second strand buffer (Life technology), 0.75 μl of dNTP mix (2.5mM each/10mM, Thermo Scientific), 1 μl of DNA polymerase I (10U/μl, Life technology) and 0.25 μl RNase H (2U/μl, Life Technology). After adding the mastermix to the sample, Program JEPR3 (Incubate 2 hours at 16°C) was performed (second strand synthesis). This two hour should be exact as the second strand buffer also fragments DNA when the delay become longer. 6. After exactly 2 hours, 1 μl T4 DNA polymerase (5 U/μl) was added to each sample, mixed by pipetting and spin briefly. Then the Program JEPR3.5 (Incubate 10 more minutes at 16°C) was used. This step is for end repair. 7. When end repair is finished. The sample reactions were cleaned with MinElute Reaction Cleanup Kit (Qiagen, 28204). The clean up was performed 14 following the protocol in the mini guide provided with Qiagen (www.qiagen.com). On last elution step, 11 μl of nucleus free water instead of Buffer EB was added to each sample and samples were transferred the samples to a 0.2-mL PCR tube. 8. Then sample were concentrated to ~ 8 μl using Speedvac. During concentration, a master mix was made, per sample by mixing, 8 μl of NTP mix (18.75 mM each), 2μl of 10X reaction buffer-RT and mixed it until there is no white speckles on solution. Then 2μl 10X enzyme mix was added to the master mix per sample. When the concentration is completed, samples volumes were adjusted to 8 μl by adding nuclease free water. Then,12 μl of master mix was added and program JEPR4 was used (37 ˚C 14 hours and 4 ˚C hold). It was processed overnight and came back within at most 2hrs after the 14hr cycle is over. 9. On next day, the samples were cleaned with MEGAclear kit following MEGAclear protocol with ethanol precipitation. 10. To each pellet, 4 ul with MEGAclear elution buffer and 1 μl of Random primers (0.05 mg/ml) was added. To remove any secondary structure and anneal the primer, program JEPR5 (Heat at 70 ˚C for 10 minutes) was used and samples were mmediately placed sample tubes on ice for at least 2 minutes. 15 11. While the JEPR5 program is running 2 nd round 1 st strand master mix per sample was generated by mixing 2 μl of 5x First Strand buffer, 0.5 μl of dNTP’s (2.5 mM each/10mM), 0.5 μl of RNasin(Promega), 1 μl of DTT (100mM) and 1 μl of Superscript III. After adding 5 μl of master mix to the sample, program JEPR6 (Incubate 10 minuntes at 25 ˚C, Incubate 30 minutes at 42 ˚C, Heat 5 minutes at 95 ˚C to denature) was run for first strand synthesis. 12. When program was finished, samples were placed on ice for at least 2 minutes (water ice bath) and 2 μl of dT-T7 oligo (10 ng/μl) was added. For alignment, program JEPR7 (Heat for 6 minutes at 70˚C) was used. 13. While waiting, master mix was made by mixing, per sample, 43.5 μl of DEPC water, 15 μl of 5X Second strand buffer, 1.5 μl of dNTP mix (2.5 mM each/10mM), 2 μl of DNA polymerase I. After mixing thoroughly by pipetting and spining briefly, 62 μl of master mix was added to the sample and program JEPR3(Incubate 2 hours at 16˚C) was run on thermocycler. 14. After exactly 2 hrs. 2 μl T4 DNA polymerase (5U/μl) was added to each sample for end repair and Program JEPR3.5 (Incubate 10 more minutes at 16˚C) was used. 15. The step 7-14 (3 rd round) was repeated. 16 16. The step 7-9 was repeated and samples were eluted in 50 μl of nucleus free water. 17. With 1 μl of each sample the quality and quantitiy of the sample was measured using Agilent 2200 Tapestation device (RNA tape). 18. Continued to Illumina Truseq mRNA protocol after frag-prime part 1.2.5 Method Modification from the NuGEN kit In order to increase the sensitivity of the NuGEN RNAseq V2 protocol, first, the transcriptase enzyme for the first strand step was replaced with more reliable enzymes such as qscript (95048-025, Quanta bio) and Superscript III (Life technology). Next, the reaction volume was reduced using 0.01ul syringe to process microvolume reactions to increase sensitivity. The original NuGEN RNAseq V2 protocol recommends to use the NuGEN Ultralow library kit, which utilizes PCR amplification. In order to develop a method, which can minimize PCR biases, after the NuGEN RNAseq V2, three different library kits (the NuGEN Rapid or the Lucigen Nxseq library kit) were used. These protocols do not over-amplify the library unlike the NuGEN Ultra library kit. For testing, the protocol below was followed: 1. On cold block, the Universal Human Reference RNA(UHR) sample was dilute into 10pg /μl concentration. Take 1 μl of the diluted sample and add 17 4 μl of nuclease free water to make 5 μl of initial volume. (extra low volume protocol: 1 μl) 2. 2 μl (extra low: 0.4μl) of A1 (first strand primer mix) was added to each sample. Then, the tubes were placed in a pre-warmed thermal cycler programmed to run Program Nu1: 65°C – 2 min, hold at 4°C. 3. When programming is finished, the first strand master mix was prepared by combining, per sample, 2.5 μl of A2 and 0.5 μl of A3. Then, 3 µL (extra low: 0.6 μl) of the first strand master mix was added to each tube and mixed the sample with mix by pipetting 10 times. The samples were spun down and placed in a pre-cooled thermal cycler programmed to run Program Nu2: 4°C – 1 min, 25°C – 10 min, 42°C – 10 min, 70°C – 15 min, hold at 4°C). 4. When the first strand synthesis is finished, the second strand master mix using, per sample was prepared. 9.7 μl of the Second Strand Buffer Mix (yellow: B1) and 0.3 μl of the Second Strand Enzyme Mix (yellow: B2). The 10 μl (extra low 2 μl) of master mix was added to the sample and mixed by pipetting entire volume 10 times. The sample tubes were placed in a pre- cooled thermal cycler programmed to run Program Nu3: 4°C – 1 min, 25°C – 10 min, 50°C – 30 min, 80°C – 20 min, hold at 4°C. When program was started, Agencourt RNA Clean XP beads was removed from the 4°C fridge. 18 5. When the program is finished, 32 μl (6.4 μl for extra low protocol) of the Agencourt RNA Clean XP beads was added to each sample. For 10 minutes, the sample-bead mix was incubated at room temperature to let the cDNA bind to the beads. After briefly spinning, the samples were placed on the magnet until the liquid becomes completely transparent. 6. While the samples are remaining on the magnet, the supernatant was discarded by pipetting and 200μl of freshly made 70% ethanol was added to each sample to wash them. Each washing protocol was 30 seconds and another two washes were performed afterward. 7. After the washing step, samples were spun down to discard any remaining ethanol on in the tube and dried the samples at room temperature for 2.5 minutes until the pellet until a small crack is visible. While waiting by mixing, the SPIA master mix was made by mixing 5 μl SPIA Primer Mix (red: C1), 10 μl of SPIA Buffer Mix (red: C2) and 5 μl SPIA Enzyme Mix (red: C3). 8. When the beads were dry, 20 μl of SPIA master mix was added directly to the beads and mixed by several pipetting until it gets rehydrated. When samples are rehydrated, they were spun down the samples and place the tubes in a pre-cooled thermal cycler programmed to run Program Nu4: 4°C – 1 min, 47°C – 1hr 10 min, 80°C – 20 min, hold at 4°C (on 47°C, several different lengths were tested). 19 9. When program is finished, 36 μl of Agencourt RNA Clean XP beads was added to each sample. For 10 minutes, the sample-bead mix was incubated at room temperature to let the cDNA bind to the beads. After briefly spinning, the samples were placed on the magnet until the liquid becomes completely transparent. 10. While the samples are remaining on the magnet, the supernatant was discarded by pipetting and 200μl of freshly made 70% ethanol was added to each sample to wash them. Each washing protocol was 30 seconds and another wash was performed afterward. 11. After the washing step, samples were spun down to discard any remaining ethanol on in the tube and dried the samples at room temperature for 2.5 minutes until the pellet until a small crack is visible. 50μl of RNase-free water was added and mixed thoroughly. After placing the sample tube back on magnet, supernatant was placed in new nucleus free, low binding tube. 12. With 1μl of sample, the quantity and quality of the sample was measured using Agilent 2200 Tapestation. The samples with enough quantity were placed in Covaris AFA tube. 13. Using the Covaris S2 device, samples were sheared into 213bp. 14. Follow Library prep (Lucigen NxSeq, NuGEN Rapid) listed in each protocol. 20 1.2.6 Method Modification from the aRNA amplification method The original aRNA protocol is composed of 3 rounds of amplification. In order to increase the yield, bead purification was used for the purification steps instead of the column purification. Moreover, there are several changes in reaction volumes. For these experiments, either 5pg or 10pg of the Universal Human RNA Reference (UHR, Agilent, 740000) was used. 1. On step 7 of the aRNA amplification purification protocol, instead of following the QIAgen protocol, two variations were made. 1) QIA cube purification (it is mostly same as the original method, but QIAcube protocol instead was used) 2) Bead purification. <Bead Purification: Replacing Step 7> A. When program is finished, 52 μl of the Agencourt RNA Clean XP beads was added to each sample. For 10 minutes, the sample-bead mix was incubated at room temperature to let the cDNA bind to the beads. After briefly spinning, the samples were placed on the magnet until the liquid becomes completely transparent. B. While the samples are remaining on the magnet, the supernatant was discarded by pipetting and 200μl of freshly made 70% ethanol was added to each sample to wash them. Each washing protocol was 30 seconds and another two washes was performed separately afterward. 21 C. After the washing step, samples were briefly spun down the to discard any remaining ethanol on in the tube and dried at room temperature. When a small first crack is observed in pellet, 4 μl of nuclease free water was added to cover the bead pellet. Samples were removed from magnet and mixed thoroughly. When all the beads were liquidified, the sample-bead mixs were moved back to the magnet and transfered 4 μl to a new PCR tube 2. On step 8, there are few changes made. First, the volume of the reaction become halved and as the elution volume from step 7 was 4 μl, the concentration step was skipped. 3. On step 9-10, the Megaclear kit and ethanol precipitation was replaced with bead purification. The initial volume of bead added to the samples was 19 μl and the final elution volume was 4 μl. 4. On step 12-14, the total reaction volume became halved and the amount of water was decreased. 5. On step 15, the Minelute kit was replaced with bead purification and the T7 became half reaction. Moreover, using this modified protocol, the 3 rd round of amplification is no longer required. 1.2.7 Method PAIA (Patch-aRNA in vitro transcription amplification) Extraction of cytoplasm using Patch Clamp technique Using above modifications, the PAIA method was developed for a single cell (Figure 1-3A). This technique also allows to record 22 electrophysiological properties of the same cell prior to extraction of cytoplasm for RNAseq. Briefly, the PAIA has 3 steps. The first step is using the optical microscopy with an infrared Dodt gradient contrast system. The morphological information was collected with these criteria: (1) location of the cell (layer); (2) the cell body size; and (3) shape. This information was later used for classification of neurons. From temporal lobe, the shape and size information allowed us to define pyramidal neurons and interneurons. In embryonic spinal cord, the information was used to define motor neurons from others. In embryonic brain, combined with the tissue information such as embryonic age, the morphological information was used to define Cajal Retzius neurons from others. The second step is recording the electrophysiology by patch clamp (EPC-10 amplifier, HEKA). Glass patch pipettes was used (6-10 MOhm; 1.2 mm O.D.) filled with intracellular solution composed of K-gluconate130 mM), potassium chloride (KCl, 2 mM), calcium chloride (CaCl2, 1 mM), MgATP (4 mM), GTP (0.3 mM), phosphocreatine (8 mM), 4-(2-hydroxyethyl)-1- piperazineethanesulfonic acid (HEPES, 10 mM), ethylene glycol-bis(β- aminoethyl ether)-N,N,N',N'-tetraacetic acid (EGTA,11 mM), RNase inhibitor (0.4 U/μl; RNase inhibitor, Clontech). The solution was titrated to pH 7.25 and 300 mOsm. After reaching cell with positive pressure, the giga-seal was formed by gentle suction (15 to -30 mmHg). By applying a brief pulse of suction, whole-cell configuration was made by membrane rupture (-90 to -150 23 mmHg for 400 ms). Only on few cells, the current-clamp mode measurement was performed to monitor spontaneous membrane potential change. Because of RNA degradation, spontaneous activity was measured in only few selected cells. In most of the cells we measured activity, a series of current were injected to measure the triggered action potentials. After electrophysiological recording, the cytoplasm of the cell was extracted using negative pressure (- 200 to -250 mmHg). The content was released into PCR tube containing 5 μl of lysis buffer (5g/L NaCl; 1% Triton X-100; 1% NP-40; 5% sodium deoxycholate; 5% Tris-HCl; 20mM HEPES; Proteinase and phosphatase inhibitors) by breaking tip of the pipette on bottom of the low binding tube filled with lysis buffer and applying positive pressure (25-50 mmHg). The cellular material was centrifuged and completely lysed by 1~2 freeze thaw cycles. The third part is RNA sequencing. By using the modified aRNA method (1.2.6), the single cell RNAseq for 2,815 cells was performed. 1.2.8 Alignment, Quality Control and Data analysis A single cell RNAseq data was sequenced with the Illumina Hiseq 2500, producing fastq files. Fastq files were mapped with GT-FAR (https://genomics.isi.edu/gtfar) program, which includes Perm (Y. Chen, Souaiaia, & Chen, 2009) as the aligner. The Genecode 22 was used as the annotation. The samples with read less than 100,000 exonic reads were dropped. The total mapping rate and exonic mapping rate was measured. Most of the analysis was performed with all the samples, which are 24 downsampled to 100,000 reads. The number of genes was measured and PCA plots of methods were generated with downsampled data. The hierarchical clustering (Euclidean) was used. 1.3. Results 1.3.1 NuGEN RNAseqV2 Method Modification to process 10pg of RNA The recommended amounts of materials required for the library prep on the NuGEN library kits were 100ng and 50ng. To find the minimum amount of the input required for these library preps, a large amount of cDNA was generated from 1ng of starting material by using the NuGEN RNAseqV2. Using the different amount of sheared cDNA (20, 30, 40, 50ng), libraries were prepared by either the NuGEN Rapid library kit or NuGen Ultralow library kit. Each library was sequenced for 10 million reads in the Illumina Hiseq sequencer. The correlation between technical replicates was 0.96 up to 10ng of initial material of cDNA when either the NuGEN Rapid library kit or NuGen Ultralow library kit was used (Figure 1-4). With the 100pg of UHRs, which were processed with the NuGEN RNAseq, libraries prepared by the NuGEN Rapid library kit and the one from the NuGEN Ultralow library systems were compared (Figure 1-5). The correlation between the NuGEN Rapid technical replicates was much higher (0.59) than the Ultralow protocol library method (0.35). The average mapping 25 rate was 35% for the NuGEN Rapid and 27% for the NuGEN Ultra low library prep. These results indicated that the data quality increased with the NuGEN Rapid protocol, which uses non-PCR based library preparation. Therefore, the NuGEN RNAseq2 was further performed with the Rapid library kit using a 10pg of UHRs. However, the original NuGEN RNAseqV2 was not sensitive enough to process a 10pg of UHRs. So, the NuGEN RNAseq V2 protocol was modified as we described following. When we tested whether the reaction volume affects the yield of cDNA, we found that, by reducing the reaction volume, the template concentration could be increased and the yield of cDNA in each step. Therefore, as the first modification, the first strand synthesis and second strand synthesis volume were reduced in ½ and the SPIA amplification volume was reduced into ¼. After testing the yield of amplified cDNA with the time of SPIA amplification, we found that by increasing time on the linear amplification, higher yield was made. Therefore, additional 20 minutes were added to the SPIA amplification time. The above modified NuGEN RNAseq V2 protocol was applied to 10pg of UHR with the NuGEN Rapid library prep. The correlation was 0.7 (R) between replicates. Therefore, we decided to use the NuGEN rapid protocol for library prep and processed a real single cell. However, unlike UHR data, single cell data showed much lower mapping rate even with the NuGEN rapid protocol. The lower mapping was due to the large amount of spurious repetitive sequences. In summary, the NuGEN RNAseq V2 with NuGEN Rapid 26 library prep provided a robust result on a 100pg of initial input, however, it was not well optimized for a 10pg even after modifying several steps (Figure 1-6). 1.3.2 aRNA Method and Modification to process 10pg The original aRNA amplification method was performed as described in previous studies (Morris et al., 2011; Spaethling et al., 2017) using a 10pg of Universal Human RNA Reference (UHR, Agilent, 740000). The correlation between technical replicates was 0.58 (R). However, we realized that the original protocol uses three rounds of amplification, and the resulted amount was far enough. Therefore, we modified the original aRNA amplification protocol in several steps (Figure 1-3B): replacement of purification, volume alterations and reduction of amplification cycles. The most recent iteration (Spaethling et al., 2017) of the aRNA amplification described was using column purification steps such as Megaclear purification (Ambion, AM1908) for RNA and MinElute kit (Qiagen, 28004) for DNA. However, it was commonly accepted that the column-based purification methods are prone to loss of materials (Flagstad, Røed, Stacy, & Jakobsen, 1999). Moreover, the Megaclear purification needs to be followed by ethanol precipitation, and the MinElute kit needs to be followed by a drying step. These steps are time consuming and require RNase-free environment. Therefore, these purification steps could potentially compromise the yield and complexity of the transcriptome data. 27 Compared to the column-based methods, purification using magnetic beads was found to increase yields of nucleic acids (Ribeiro-Silva, Zhang, & Jeffrey, 2007). Therefore, we utilized the magnetic bead purification to replace both the Megaclear RNA purification step and the MinElute DNA purification step. Because bead purification does not require following ethanol precipitation, this replacement removed both volume adjustment processes and resulted in a significantly reduction in the procedure time. Additionally, this method gives an advantage on flexibility of elution volume. This flexibility allowed us for the second part of modification (Figure 1-7). Because the initial volume for the next reaction is depending on the previous elution volume from purification, the reaction volume could be reduced by half with the bead purification step. In a half reaction, the concentration of initial template became higher and the efficiency of the reaction was increased. Moreover, this could reduce the cost of the RNA sequencing substantially. The substitution of both column purification steps with bead purification and reduction of volumes in some reactions resulted in sufficient improvements in yield (Figure 1-8), and it allowed us to reduce number of rounds of amplification from three to two (see details in Methods). In addition, several intermediate versions were tested for addressing the difficulties with a large number of data production. To overcome these issues, some high-throughput applications were implemented. The QiaCube allowed us to partially automate the process 12 samples of the DNA 28 purification with the Qiagen MinElute Kit at once. Furthermore, a single sample Megaclear column RNA purification was replaced with a 96-well based Megaclear kit (Ambion, AM1909). Implementing these steps shortened the protocol thus reducing potential RNA degradation. Though these modifications were replaced with bead purification in final modified protocol, concept using high throughput was passed down to usage of multi-channel protocol on the modified method. To compare the data from intermediate and modified methods with the original aRNA method, each modified method was performed by using the same amount (10pg) of Universal Human Reference RNA (UHR). For each sample, 2~4 million reads were sequenced, then the data qualities were evaluated using following four metrics; mapping rate to the transcriptome, the number of genes, correlation between replicates, correlation to bulk RNA (Figure 1-9). The first metric was the mapping rate to the transcriptome. The percent of reads mapped to the transcriptome excluding rRNA and mtRNA (GENCODE v22) was measured (Figure 1-9B). We also separately measured a mapping rate to transcriptome calculated from total number of reads and those calculated from reads mapped to a human genome and transcriptome, including rRNA and mtDNA. The second metric was calculating number of genes detected. In order to adjust the bias from different number of reads, for this analysis, each sample was downsampled (100,000 reads) from total 29 mapped reads, and calculated number of detected genes, which excluded rRNA and mtRNA as a proxy of library complexity. The third metric was measuring the Pearson correlation (R) between technical replicates as the proxy of reproducibility. The last metric was to compare the gene expression levels in each sample of each method with the one generated from a bulk RNA (100ng) processed using the Truseq Total RNA and Truseq PolyA RNA protocols. Similar to the second metric, the samples were downsampled to 100,000 reads for this metric.1.3.4 aRNA Method Modification Comparison with Single Cells. Interestingly, the modified aRNA method showed better quality in all of four metrics than the original aRNA method (Table 1-2). The mapping rate and the number of genes was increased in the modified aRNA method. The correlation was increased between technical replicates in modified aRNA method and the correlation to the bulk was also increased. 1.3.3 Comparison of Methods (SMARTer, NuGEN, aRNA) with UHR 10pg We compared some of the commercially available popular RNAseq methods: Clontech SMARTer-NEXTEra (polyA RNA, PCR amplification), NuGEN RNAseq V2 (total RNA, based on linear DNA amplification) with our modified aRNA amplification. For the NuGEN method, the modified NuGEN method (see above) with the Rapid library kit was used. All three methods were performed with Universal Human Reference RNA (UHR) and compared 30 the reproducibility and accuracy between them. Then, expression measurements generated by three single cell RNAseq methods were characterized in terms of sensitivity, precision and accuracy. For initial assessment, 10pg of the Universal Human Reference RNA (UHR) was used. To measure the quality, the above four metrics (mapping rate to the transcriptome, the number of genes, correlation between replicates, correlation to bulk RNA) were used for comparison (Table 1-2,Table 1-3). The Clontech SMARTer-Nextera method had the lowest transcriptome mapping rate calculated in relation to the total number of raw reads, but the highest percentage of mapped reads out of the total number of reads mapped to genome and transcriptome. It meant that most reads did not map to the genome or transcriptome. We found that these unmapped reads are mostly derivatives of the SMARTer IIA primer. This method had also the lowest number of genes detected and the lowest correlation to the bulk RNA. The modified NuGEN Ovation RNAseq V2 method had a low mapping rate in general but had the highest complexity and accuracy. The modified aRNA protocol was the second highest on most of these metrics (0.61, 0.60 correlation) but had the highest mapping rate to the transcriptome. Therefore, we decided to use two better methods (the NuGEN and aRNA method) for further comparison. Next, we applied the two methods, modified NuGEN Ovation RNAseq V2 and modified aRNA method using the 5 pg of RNA, which is an average 31 amount of RNA can be collected from a single neuron with a patch clamp pipette. When 5pg of RNA was used, most times, the NuGEN RNAseq V2 method did not produce enough library for sequencing. Only 16 libraries out of 333 had sufficient material for sequencing, and the average mapping rate for these libraries was 10%, even lower than for 10 pg of UHR. Even if the amount of library material was sufficient for sequencing, the mapping rate was low. Unmapped reads were often including homologous or repeating sequences, suggesting that they are products of spurious amplification (Table 1-4). On the other hand, modified aRNA protocol consistently produced libraries even at RNA concentration of 5 pg. All the libraries had sufficient material for sequencing, and average mapping rate was 70% (Figure 1-9B). This comparison indicated that modified aRNA is working robustly on a small amount of RNA materials and is potentially suitable for single cell transcriptome studies. 1.3.5 PAIA application The comparison with UHR indicated a clear improvement in aRNA method with the modification. However, it was not clear that the improved method could be applied to an actual single cell, specifically, for the cells collected with patch clamp recording. Therefore, the same comparison was performed using single cells. The cytoplasm of 680 cells from human embryonic brains (EB) was extracted with patch clamp pipette. In order to 32 define the cell types of each collected cell, cell morphology and location of the cell in the tissue section were recorded. Among these cells, 8 cells were processed with original aRNA method, 204 were processed with an intermediate method and 1,279 were processed with the modified method. The reads were aligned to transcriptome and genome and assigned to genes (GENCODE V22) by GT-FAR pipeline (https://genomics.isi.edu/gtfar). The average mapping rate, Pearson correlation to the bulk embryonic brain data from the BrainSpan (http://www.brainspan.org) project and the number of gene discovered when downsampled to 100,000 reads were used as a proxy of quality of the data. On some of EB cells, Pearson correlation of combined single cells data with the age-matched data from BrainSpan samples was calculated as a proxy for accuracy of the data. The average mapping rate was 25% in original method, 45% in intermediate method, 75% in modified method. The average mapping rate was higher in cells processed with modified protocol by 30%. The number of gene discovered in downsampled data was also increased in the modified protocol (2 fold). 1.3.6 Single cell data from PAIA comparing with public dataset The Patch clamp-based collection combined with modified aRNA protocol, Patch-aRNA in vitro transcription amplification (PAIA), was applied to non-dissociated embryonic brain (EB) cells. we have collected 680 cells from different regions of embryonic brain (Figure 1-10). 33 Then, our EB data was compared with publicly available data generated from 466 human brain single cells (Darmanis et al., 2015), collected from fetal and adult brain samples’ dissociated tissues. In contrast to our study, Darmanis et al. dissociated brain cells and used Fluidigm C1 instrument, facilitating the SMARTer protocol followed by the NEXTEra library preparation. All original raw data from these studies (GSE67835) were processed with the same data analysis pipeline, mapped to GENCODE 22 human or mouse transcriptome (depending on origin of cells). To assess quality of single-cell RNA-seq data, we compared Darmanis et al study with our study using QC metrics, which included total number of genes discovered per 100,000 sampled reads as a metric of library complexity, and correlation between replicates and with bulk RNA (also based on 100,000 sampled reads) to evaluate accuracy (Figure 1-11). 1.4 Discussion The PAIA protocol provides the electrophysiological recording and improved quality of single cell RNA expression data. Moreover, the single cell RNAseq data processed with this method had less amplification resulting, increased number of gene discovered and mapping rate, compared to the other methods (SMARTER, NuGEN). Therefore, the modified aRNA method could be used as a replacement for the commercially available, PCR-based amplification. 34 For single cell data, the read breakdown was analyzed by counting where each read goes such as exon, junction, intron, intergenic and unknown. The random priming is prone to increase the ribosomal RNA, and exon rate is substantially lower (10~20%). The modified aRNA method had the highest exon rate, compare to the other 2 methods. 1.4.1 Additional Modifications In addition to current modifications, the NEB had recently developed a new T7 kit, which increases the yield of amplification. From the RNA amount measurement after the second round amplification, we have found that the yield becomes almost double of the modified protocol. It indicated that we could possibly use the amplified RNA after first round amplification. However, because the Truseq mRNA library construction kit requires at least 100ng of initial RNA, it was almost impossible to start with amplified RNA after 1 st round amplification. Moreover, we have investigated additional modifications to reduce the PCR amplification in the library prep step to avoid overamplification and bias. The Swift 2S Library kit is one of the options to address both issues. It requires no PCR for 100ng cDNA and only requires 9 cycles of PCR with 1ng of RNA. Therefore, we can either amplify 2 rounds of aRNA and use no PCR or 1 round of aRNA and use less or equal number of PCR. Both options had been tested and sequenced. The method with one round of aRNA using the NEB Hiscribe T7 kit and 15 cycles of PCR showed much higher Pearson correlation with technical 35 replicates and the bulk RNA, indicating high reproducibility and accuracy. However, because this improvement was much greater change than others, most of the cells were processed with the previous method (aRNA modified). Moreover, the amount of amplified materials, measured by the Agilent 2200 Tape station method, were compared. Because current modified method requires only 2 rounds of amplification, only the samples with 2 rounds of amplification were compared. It indicated that there was about 100 times increase of the yield for the one amplified with RNA purification with beads, compared to column based, and the one with DNA purification with beads, compared to column purification. 1.4.2 What caused the improvement? The average mapping rate and the number of gene discovered were both significantly increased with the modified method. In both UHR and single cell data, the improvement in single cell data quality was observed. There are 3 factors could have caused or affected the result: reaction volume, purification yield and number of amplification cycles. It indicated that the reduction of amplification cycle improved the transcriptome complexity, number of genes at same number of mapped reads, and the quality defined by mapping rate. In most of single cell methods that are publically available, overamplification had compromised the data quality. The switching method of 36 amplification and reduction number of amplification (e.g. exponential amplification by PCR) could have improved the quality. 1.4.3 Future direction of PAIA The electrophysiological data is extremely important for neuronal cells, which have electrical activity. However, obtaining both expression profiles and matched electrophysiological data for a single cell has been difficult due to a small amount of materials available. The previous study, the Patch-seq has profiled both expression and electrophysiology of a single cell, but this method uses the PCR based amplification which utilizes method which had far less efficient then aRNA protocol. PAIA, however, by combining the patch clamp measurement method with the linear amplification, could provide higher quality information of single cell RNAseq data. Because of its higher quality RNAseq data, the resolution of the analysis on the data could be increased. It allows us to investigate RNA sequencing in even smaller unit, such as mitochondria, a specific dendrite and an axon of a projection neuron. These applications will be used for investigation of function and changes in neuronal activities. One of the other possible addition on the PAIA protocol could be an application to non-messenger RNA such as microRNA and piRNA. Instead of using poly A tails, the specific adaptor attachment protocol provided by Illumina protocol could be used. Though the additional modification will provide even more applications of PAIA, the current PAIA protocol could already have enough potential to 37 have many application in neuroscience. In next chapters, I will demonstrate how PAIA could be used as a tool to investigate neuronal development and function. 38 1.5 Figures Figure 1-1. Schematic of NuGEN RNAseq V2 protocol. From the collected extract, rest of gDNA will be removed by application of DNAse. Next, cDNA is synthesized with the Nugen RNAseq V2 protocol: First, reverse transcription, 2nd strand synthesis and SPIA (Single Primer Isothermal Amplification) linear amplification. The key point of this protocol is the linear amplification. 39 Figure 1-2. Basic Schematic of the aRNA amplification method to amplify RNA. The Method was originally developed by Eberwine group at University of Pennsylvania in 1992 and the above shows one of the improved versions (modified Figure from Jacqueline Morris et al. 2014 CSL). 40 Figure 1-3. Schematics of PAIA protocol (A) Basic Schematic of PAIA protocol. (B) Modification of aRNA protocol 41 Figure 1-4. NuGEN library input comparisons. Technical replicates of NuGEN rapid libraries generated from 20, 30, 40 and 50ng of input show similar quality libraries. 42 Figure 1-5. NuGEN Ultralow vs Rapid Library The correlation between technical replicates for NuGEN ultralow and rapid libraries when cDNA was generated with NuGEN RNAseq V2 kit. 43 Figure 1-6. Limitation of NuGEN RNAseq V2 Kit (A) Reduction of reaction volume and SPIA time increased the quality of 100pg UHR data (B) The mapping rate of single cell samples processed with NuGEN RNAseq v2 method and examples of repeating sequences found from unmapped read 44 Figure 1-7. aRNA Method Modification steps and changes in correlation The aRNA method was improved in several steps. After these steps of modification, both 100pg and 10pg UHR had improved correlation values, R 2 . 45 Figure 1-8. aRNA yield measured by Tapestation (A) 10pg RNA original protocol after 2 rounds of amplification (high Sensitivity Tape), (B) 10pg RNA modified protocol after 2 rounds of amplification (C) 10pg RNA original protocol after 3 rounds of amplification (D) 5pg RNA modified protocol after 2 rounds of amplification (Regular Sensitivity Tape used for B,C,D). 46 Figure 1-9. Correlation and quality of UHR data. (A) Correlation between technical replicates (10pg) with Bulk UHR RNA (10ng) processed with Truseq protocol. (B) Distribution of mapping rate of original protocol, intermediate modification and final modification samples (C) Average Mapping rated and number of gene discovered when downsampled. 47 Figure 1-10. Tissue Collected from embryonic brain The Single cells used for this analysis were collected from different regions of embryonic brain tissues from several origins. 48 Figure 1-11. Comparison of Single cell data (A) Number of gene distributions in 10,000 downsampled data (B) Number of average genes found in the single cell data from different methods. 49 1.6 Tables Table 1-1. Comparison between qPCR, Microarray and RNAseq Category qPCR MicroArray RNAseq Throughput 1-3 High High Price ~6.5 cents/ ul $175~$500/array $1850/lane Resolution Primer specific 100bp Single Base Methods Hybridization/Amplification Hybridization Sequencing Step after reverse transcription 1 5 >7 Background Noise None High Low Minimum input Very low(~10pg) Very high (>100ng~ug) High (100ng~ug) Isoform detection Limited Limited Unlimited SNP detection Limited Limited Unlimited 50 Table 1-2. Comparison of original aRNA and modified aRNA with 4 metrics. Transcrptomic Mapping rate % # of detected genes* Lowest Input required Correlation between Technical Replicates Correlation to Bulk RNA From Total From Mapped Total RNA Poly A Original aRNA 47.25 48.15 5085.25 10pg 0.52 0.55 0.57 Modified aRNA* 48.85 60.193 8242.40 5pg 0.67 0.61 0.61 51 Table 1-3. Comparison of Commercial Methods with the modified aRNA with Four metrics. Transcrptomic Mapping rate % # of detected genes* Lowest Input required Correlation between Technical Replicates Correlation to Bulk RNA From Total From Mapped Total RNA Clontech 3.19 65.08 4053.00 10pg -** 0.43 NuGEN Ovation V2 19.03 24.43 8779.33 >10pg 0.64 0.74 modified aRNA* 48.85 60.193 8242.40 5pg 0.53 0.61 52 Table 1-4. Number of Palindromic unmapped reads in samples. Sample # of Palindromic Sequence LG10A0-1-NGS2-R10-2_GTGGCC_L006_R1.fastq.gz 23408032 LG10A0-1-NGS2-R10-2_GTGGCC_L007_R1.fastq.gz 22845795 LG10A0-1-NGS2-R10-2_GTGGCC_L005_R1.fastq.gz 21926995 LG50_NOW5_t39_CGTACG_L002_R1.fastq.gz 21869167 LG10A0-1-NGS3-R10-6_TTAGGC_L001_R1.fastq.gz 21687462 LG-NDS1-R10-1_TAGCTT_L006_R1.fastq.gz 21459122 LG50_NOW5_t39_CGTACG_L001_R1.fastq.gz 21319604 LG10A0-1-NGS2-R10-2_GTGGCC_L008_R1.fastq.gz 20847086 LG10A0-1-NGS3-R10-6_TTAGGC_L001_R1_001.fastq.gz 20466201 LG10A0-1-NGSO1-R10-2_TAGCTT_L002_R1.fastq.gz 20418499 LG10A0-1-NGS3-R10-6_TTAGGC_L002_R1.fastq.gz 20417980 LG10A0-1-NGSO1-R10-2_TAGCTT_L001_R1.fastq.gz 20274266 LG10A0-1-NGS2-R10-7_CGTACG_L006_R1.fastq.gz 19541468 LG10A0-1-NGSO1-R10-1_ACTTGA_L001_R1.fastq.gz 19488373 LG10A0-1-NGSO1-R10-1_ACTTGA_L002_R1.fastq.gz 19388373 LG10A0-1-NGSO1-R10-2_TAGCTT_L003_R1.fastq.gz 19191702 LG10A0-1-NGS3-R10-6_TTAGGC_L002_R1_001.fastq.gz 19002696 LG10A0-1-NGSO1-R10-2_TAGCTT_L001_R1_001.fastq.gz 18992804 53 LG10A0-1-NGSO1-R10-2_TAGCTT_L002_R1_001.fastq.gz 18957114 LG10A0-1-NGS2-R10-7_CGTACG_L007_R1.fastq.gz 18836702 LG10A0-1-NGS3-R10-6_TTAGGC_L003_R1.fastq.gz 18644771 LG10A0-1-NGS2-R10-7_CGTACG_L005_R1.fastq.gz 18436716 LG10A0-1-NGSO1-R10-1_ACTTGA_L001_R1_001.fastq.gz 18299986 LG10A0-1-NGSO1-R10-1_ACTTGA_L003_R1.fastq.gz 18030217 LG10A0-1-NGSO1-R10-1_ACTTGA_L002_R1_001.fastq.gz 17974152 LG10A0-1-NGSO1-R10-2_TAGCTT_L003_R1_001.fastq.gz 17672443 LG10A0-1-NGSO1-R10-2_TAGCTT_L004_R1.fastq.gz 17550970 LG10A0-1-NGS4-R10-5_TTAGGC_L007_R1.fastq.gz 17236273 LG10A0-1-NGS3-R10-6_TTAGGC_L003_R1_001.fastq.gz 17186959 LG10A0-1-NGS2-R10-7_CGTACG_L008_R1.fastq.gz 17052701 LG10A0-1-NGS4-R10-5_TTAGGC_L006_R1.fastq.gz 16980665 LG10A0-1-NGSO1-R10-1_ACTTGA_L003_R1_001.fastq.gz 16556354 LG10A0-1-NGSO1-R10-1_ACTTGA_L004_R1.fastq.gz 16386152 LG10A0-1-NGS4-R10-5_TTAGGC_L005_R1.fastq.gz 15970098 LG10A0-1-NGS4-R10-5_TTAGGC_L008_R1.fastq.gz 15952857 LG10A0-1-NGSO1-R10-2_TAGCTT_L004_R1_001.fastq.gz 15770047 LG-NLL1-R10-2_GATCAG_L006_R1.fastq.gz 14848459 LG10A0-1-NGSO1-R10-1_ACTTGA_L004_R1_001.fastq.gz 14627859 54 LG10A0-1-NGSOO1-R10-2_CGTACG_L002_R1.fastq.gz 13197808 LG10A0-1-NGSOO1-R10-2_CGTACG_L001_R1.fastq.gz 12965681 LG10A0-1-NGS3-R10-6_TTAGGC_L004_R1.fastq.gz 12578476 LG10A0-1-NGSOO1-R10-2_CGTACG_L003_R1.fastq.gz 12470453 LG10A0-05-NGS16-Mouse2_AGTTCC_L002_R1_001.fastq.gz 12284230 LG10A0-05-NGS16-Mouse2_AGTTCC_L001_R1_001.fastq.gz 12203867 LG10A0-1-NGSOO1-R10-2_CGTACG_L001_R1_001.fastq.gz 12022020 LG10A0-1-NGSOO1-R10-2_CGTACG_L002_R1_001.fastq.gz 11968567 LG10A0-1-NGSOO1-R10-2_CGTACG_L004_R1.fastq.gz 11498872 LG10A0-1-NGS3-R10-6_TTAGGC_L004_R1_001.fastq.gz 11189373 LG10A0-1-NGSOO1-R10-2_CGTACG_L003_R1_001.fastq.gz 11182619 Samples less than 10pg processed with NuGEN RNAseq V2 has lower mapping rate due to some palindromic sequences. 55 Chapter 2. Identification and Characterization of Cajal Retzius Neurons 2.1 Introduction 2.1.1 Overview The mammalian cerebral cortex is composed of multiple layers. The development of these layers is initiated by the Cajal-Retzius cells (Marín- Padilla, 1998). The Cajal-Retzius neurons have a key role in orchestrating the formation of the mammalian neocortex. The Cajal Retzius cells are located in the outermost layer of the developing brain, the marginal zone (G Meyer, Goffinet, Fairén, & Namur, 1999; Ramo´n y Cajal, 1891; Retzius, 1893), above the cortical plate and they are the major source of Reelin, a secreted extracellular glycoprotein critical for guiding neuronal migration and establishing the cortical laminar structure (D’Arcangelo, 2014; Sekine, Kubo, & Nakajima, 2014; Tissir & Goffinet, 2003). As the chemical signals draw Cajal- Retzius cells towards layer I, they express Reelin protein, which allows the rest of interneurons migrate to form layers (Ogawa et al., 1995). Loss of Reelin or disruption of the Reelin signaling pathway causes major defects in neuronal migration, leading to completely inverted or entangled cellular layers in the neocortex (D’Arcangelo, 1995; Howell, Hawkes, Soriano, & Cooper, 1997; Ogawa et al., 1995; Trommsdorff et al., 1999). In addition to secreting Reelin, Cajal-Retzius neurons direct neuronal 56 migration via cell adhesion molecules such as NECTIN 1/NECTIN 3 and CADHERIN 2 (Gil-Sanz et al., 2013). Cajal-Retzius neurons also help scaffolding of radial glial cells (Marín-Padilla, 1998), a cell type critical in radial migration during development (Soriano, Alvarado-Mallart, Dumesnil, Del Río, & Sotelo, 1997; Supèr, Del Río, Martínez, Pérez-Sust, & Soriano, 2000). 2.1.2 Role of Cajal-Retzius cells Cajal-Retzius cells are the first neurons found in the developing cortex and form the earliest synaptic networks as evidenced by both morphological and functional studies (Anstötz et al., 2014; del Río, Martínez, Fonseca, Auladell, & Soriano, 1995; Janusonis, 2004; Myakhar, Unichenko, & Kirischuk, 2011; Radnikow, Feldmeyer, & Lübke, 2002). Cajal-Retzius neurons have long axonal projections spanning up to millimeters in the marginal zone and sometimes reaching the upper cortical layers in the developing cortex (Anstötz et al., 2014; Radnikow et al., 2002). The Cajal-Retzius synaptic network is formed with both cortical plate and subplate neurons, often using their apical dendrites (Kirischuk, Luhmann, & Kilb, 2014; O’Leary & Koester, 1993). Layer formation concludes with the removal of Cajal-Retzius cells by the neurotransmitter norepinephrine. The first known role of norepinephrine is the regulation of the Cajal-Retzius cell removal postnatally as evidenced by the reduced rate of Cajal-Retzius removal upon the application of 6- hydroxydopamine (6-OHDA), a selective toxin for noradrenergic neurons (Figure 2-1), thus delaying layer formation in the cortex (Naqui, Harris, 57 Thomaidou, & Parnavelas, 1999). There is a debate on the method of Cajal- Retzius population depletion after the migration. Studies in mouse and cat neocortex suggest that the Cajal-Retzius cells underwent apoptosis (Derer & Derer, 1990; Noback & Purpura, 1961). Alternatively, dog and rat studies support the theory that Cajal-Retzius cells become pyramidal cells by transformation after migration (Fox & Inman, 1966; Parnavelas & Edmunds, 1983). It is still unknown what molecular mechanism is employed in Cajal- Retzius cell removal in human brain development. Moreover, little is known about the development of electrophysiological and neurotransmission properties in human Cajal-Retzius neurons. Specifically, the timing in become electrically active during development us unknown. In the developing mouse brain, Cajal-Retzius neurons have been shown to fire action potentials with a s low time course, exhibit spontaneous synaptic activities, and respond to multiple neurotransmitters such as glutamate, GABA, glycine, serotonin, and noradrenalin. Cajal-Retzius neurons form glutamatergic synapses using elongated axons that connect to different regions of the central nervous system, such as the presubicular cortex (Achilles et al., 2007; Anstötz et al., 2014; Chameau et al., 2009; Chowdhury, 2010; del Río et al., 1995; Janusonis, 2004; Kilb et al., 2002; Kilb & Luhmann, 2001; H. G. Kim, Fox, & Connors, n.d.; H J Luhmann, Reiprich, Hanganu, & Kilb, 2000; Myakhar et al., 2011; Radnikow et al., 2002; Sava et al., 2010; Schwartz et al., 1998). This early active synaptic network is known to have a 58 significant role in instructing cortical neuron migration (Heiko J. Luhmann, Fukuda, & Kilb, 2015; Manent & Represa, 2007) and stabilizing established synapses of other developing neurons (McKinney, Capogna, Dürr, Gähwiler, & Thompson, 1999). 2.1.3 Approaches We took a multidisciplinary approach to studying the developing human neocortex at the single cell level. We performed electrophysiological measurements and investigated the transcriptomes of human Cajal-Retzius (CR) neurons in the second trimester of gestation. We also examined neurons from the cortical plate (CP) and subplate (SP) for comparison. We observed that human CR neurons in the second trimester have spontaneous synaptic activity and fire action potentials. Furthermore, we have investigated the expression of the transcriptomes of hundreds of non-disrupted CR, CP and SP cells in relation to location and morphology, and found clear differences in global gene expression with numerous significantly differentially expressed genes by cell-type. The electrophysiology and transcriptome data of neurons characterized here are useful resources as biomarkers and reference transcription profiles of different cell types in humain brain development. 59 2.2 Methods & Materials 2.2.1 Brain slice preparation Postmortem fetal brain tissues, in the second trimester of gestation were used in this study. Tissue was donated with informed consent, using a procurement procedure that was approved by the institutional review board of USC. Specimens were transported to the laboratory in Hypothermosol solution (Sigma, USA; Biolife Solutions, USA) on ice. Tissue transport time was held under two hours. Brain tissue was then visually examined for structural integrity and the anterior or posterior face of the tissue was glued on a specimen plate. The tissue was then sliced into 400μm slices in cold (4˚C), oxygenated (95%O2, 5% CO2) N-Methyl-D-glucamin-artificial cerebral spinal fluid solution (NMDG-ACSF; N-Methyl-D-glucamin (NMDG) 93mM; KCl 2.5mM; NaH2PO4 1.2mM; NaHCO3 30mM; HEPES 20mM; glucose 25mM; sodium ascorbate 5mM; thiourea 2mM; sodium pyruvate 3mM; MgSO4 10mM; CaCl2 0.5mM) using a vibratome (Leica VT1200S, Germany). Vertical deflection of the blade was minimized with Vibrocheck technology and slicing parameters were: speed 0.1-0.15 mm/s and vibration amplitude 1.5 mm. Slices were transferred to a recovery chamber (32˚C, 95%O2, 5% CO2) and allowed to recover in artificial cerebral spinal fluid (ACSF; NaCl 124 mM; KCl 4 mM; NaHCO3 26 mM; glucose 10 mM; CaCl2 2mM; MgCl2 2mM) for 30-60 minutes before recording and cell collection was performed. 60 2.2.2 Electrophysiological recording Cajal-Retzius neurons were visualized by optical microscopy using an infrared Dodt gradient contrast system. They are readily identified by several criteria: (1) location in the marginal zone; (2) large cell body; and (3) mostly horizontal orientation with long, easily-resolved processes. Patch pipettes (6- 10 MOhm; 1.2 mm O.D.) were filled with intracellular solution (K-gluconate 130 mM; KCl 2 mM; CaCl2 1 mM; MgATP 4 mM; GTP 0.3 mM; phosphocreatine 8 mM; HEPES 10 mM; egtazic acid (EGTA) 11 mM; pH 7.25 and 300 mOsm) containing RNase inhibitor (0.4 U/μl; RNase inhibitor, Clontech). Pressure control of the patch pipette was performed with an automatic pressure control unit (ez-gSEAL 100b, Neobiosystem, USA). Before reaching the cell, positive pressure (25-50 mmHg) was applied. Once the pipette touched the cell membrane, a gentle suction (-15 to -30 mmHg) was applied to form a giga-seal. While the seal was stabilized for 1 minute, the membrane rupture was achieved by applying a brief pulse of suction (-90 to -150 mmHg for 400 ms) to attain the whole-cell configuration. Standard whole-cell patch clamp (EPC-10 amplifier, HEKA) was performed in current-clamp mode first to monitor spontaneous membrane potential change; then a series of current injections was used to trigger action potentials. A steady DC current (typically <-50pA) was applied when necessary to hold the membrane potential at -60 mV. 61 Following electrophysiological recording, the somatic cytoplasm of the cell, from which had been recorded was collected into the patch pipette with a pulse of strong negative pressure (-200 to -250 mmHg). The content was expelled into a PCR tube containing 5 μl of lysis buffer (5g/L NaCl; 1% Triton X-100; 1% NP-40; 5% sodium deoxycholate; 5% Tris-HCl; 20mM HEPES; Proteinase and phosphatase inhibitors) by breaking the pipette tip and applying positive pressure (25-50 mmHg). The cytoplasmic content was then flash frozen in liquid nitrogen before being stored in a -80˚C freezer. 2.2.3 Data Analysis of electrophysiological recording Electrophysiological recordings were processed and examined using customized macros in Igor Pro (Wavemetrics, USA) and Matlab (Mathworks, USA). 2.2.4 Single Cell RNAseq Cytoplasmic content from individual cells was amplified using the aRNA method (Eberwine et al., 1992; Phillips & Eberwine, 1996) with some modifications. The aRNA method was composed of 2 cycles of four procedures: first strand cDNA synthesis, second strand cDNA synthesis, DNA purification, in vitro transcription (IVT) and RNA purification. The first strand cDNA was synthesized with Superscript III (Life Technology) and Oligo-dT T7 primers (Custom, Life Technology) followed by second strand synthesis using DNA polymerase I (Life Technology) and RNase H (Life Technology). The DNA was then purified using RNAClean XP beads (Beckman, CA, USA) and 62 used as a template for in vitro transcription (IVT). IVT was performed with T7 RNA polymerase and Megascript T7 kit (Life Technology). The reaction volume was reduced to half for IVT and enzymes were replaced with thermostable alternatives and typically performed only two rounds of IVT. With this protocol, 100~500ng of amplified aRNA was produced reliably from single cell samples. Libraries were then generated from 150-400ng of amplified RNA using TruSeq Stranded mRNA Sample Preparation Kit (Illumina, CA, USA). Briefly, single strand cDNA was synthesized using SuperScript II (Life Technology) and random primers. The cDNA was converted into double stranded cDNA, blunt-ended and adenylated at 3’-end. The resulting library dsDNA fragments were ligated to adapters and enriched using Illumina PCR primers followed by purification with AMPure XP beads (Beckman, CA, USA). DNA fragment size and library concentration were examined using 2200 TapeStation (Agilent Technologies) before sequencing with hiseq 2500 rapid mode. 2.2.5 Data Analysis of single cell RNAseq Prior to mapping, reads containing more than 50% of adapter sequences, monomers or other low entropy reads (metric entropy below 1%) were removed. The remaining reads from each individual channel were trimmed (if adapters constitute less than 50% of the read) and sequentially aligned to rRNA, mtDNA, the rest of human transcriptome (GENCODE v22; 60,483 gene models) and genome (GRCh38) using our custom RNA-Seq 63 alignment pipeline, GT-FAR (https://genomics.isi.edu/gtfar). Reads mapped to rRNA and mtDNA were excluded from following analysis. Genes with ≤10 reads in ≤10% of cells were discarded. To adjust for differences in library complexity, the data from each cell was downsampled, without replacement, to 100,000 mapped exonic reads and log-normalized. Current single cell RNA- Seq technologies fail to reliably detect the presence of transcripts that are in the individual cell, particularly at low expression levels. These dropouts create a sparse data matrix that presents challenges for dimensional reduction. Moreover, the variation caused by batch effect, source of tissue and age of the cells need to be corrected. To address this problem, the covariates for flowcell, biopsy ID and gestational week of each cell was matched. The regression from groups of single cells of each known subtype to create multiple synthetic pooled samples was condensed. For each subtype, random groups of 16 cells were repeatedly selected and condensed into pools, which were dimensionally reduced using Principal Components Analysis (PCA). Expression values from each single cell were then projected onto the first (PC1) and second (PC2) principle components of the condensed data. 2.3 Results 2.3.1 Identification of human Cajal-Retzius neurons During the second trimester, the human cerebral cortex exhibits distinct cellular layers with different cell densities (Figure 2-2A), which allows one to identify and collect neurons subsequently from three major upper laminations: 64 marginal zone, cortical plate and subplate. The Cajal-Retzius neurons are distinctly located in the marginal zone with a large and ovoid shaped cell body (Figure 2-2B). These neurons typically have at least one identifiable long, tapered dendrite stemming from the cell body and occasionally an axon is also be observed (Kirischuk et al., 2014). CP (Figure 2-2C) and SP neurons (Figure 2-2D) are readily identifiable by location and are useful for an age-matched comparison of the neuronal cells (Chameau et al., 2009; Ogawa et al., 1995). To extract cytoplasm, standard glass micropipettes were used and the electrical activity of a cell in situ, from intact tissue slices was recorded with a whole-cell patch clamp (Figure 2-2E&F). The extracted cytoplasm of the same cell was used for a single cell RNA-seq to profile the global transcriptome (Eberwine et al., 1992; Moll et al., 2004; Phillips & Eberwine, 1996; Qiu et al., 2012). In total, we collected cytoplasm and performed single cell RNA-Seq from 247 Cajal-Retzius, 290 cortical plate and 223 subplate neurons (n=760), from samples of human cortex (n=52), aged 15-20 gestation weeks (Table 2- 1). The majority of the Cajal-Retzius neurons exhibited prototypical horizontal orientation (64%), but angled (19%) or vertically (15%) orientated Cajal- Retzius neurons were also observed as previously described (MartÃnez- Cerdeno & Noctor, 2014) (Figure 2-3, Table 2-2) 65 2.3.2 Spontaneous synaptic activity in human Cajal-Retzius neurons We successfully achieved whole-cell patch clamp recordings from 77 Cajal-Retzius, 62 cortical plate and 86 subplate neurons. Interestingly, spontaneous synaptic activity from a portion of recorded neurons was monitored and all three types of neurons; 65% of the Cajal-Retzius neurons and 93-100% of the cortical plate neurons and subplate neurons showed spontaneous synaptic activity. Spontaneous electrical activity was detected as early as 15.2 GW and was observed in all ages examined (15.2-19.6 GW). The used internal pipette solution was low in chloride, enabling us to identify both glutamate-mediated excitatory and GABA/glycine-mediated inhibitory responses. All Cajal-Retzius neurons with spontaneous activity exhibited excitatory postsynaptic potentials (EPSPs), 35% of these also displayed inhibitory postsynaptic potentials (IPSPs), and 30% of them fired spontaneous action potentials (APs) (Figure 2-4). 2.3.3 Human Cajal-Retzius neurons fires repetitively Next, neurons were subjected to current injections and membrane potential changes were monitored to characterize their responses. Most fetal neurons were capable of firing action potentials upon being subjected to depolarizing stimuli, but the majority of them were still immature and capable of firing only single spikes (Figure 2-5A). Of the Cajal-Retzius neurons, about 15% could fire repetitively with spike amplitude decreasing progressively. The 66 repetitive firing was observed as early as 15.2-16 GW in human Cajal-Retzius neurons. Subplate neurons were the most mature among the three types of neurons in the study, with 31% showing repetitive firing (Figure 2-5B). We also found that hyperpolarizing current injections induced a hyperpolarization- activated cation current (Ih), causing a prominent membrane voltage sag in a significant population of all three types of fetal neurons (68% in Cajal-Retzius neurons, Figure 2-5C). Immediately following the termination of a hyperpolarization step, some neurons fired rebound action potentials. This was observed in 20% of the Cajal-Retzius neurons and more frequently seen in cortical plate and subplate neurons (Figure 2-5D). These findings provide the first evidence that human Cajal-Retzius neurons can generate action potentials and may communicate with other neurons using electrical impulses during development. From the sets of cells with which were able to make all three observations (Repetitiveness, voltage sag and rebound current), the same number of cells from three cell types (n=100) were selected. These cells were analyzed to study dependencies between the observations. In these neurons, voltage sag showed significant correlation with a repetitive adaptive firing pattern (Figure 2-5E) on all three neurons. Interestingly, the Cajal-Retzius neurons have less correlation between two (voltage sag and rebound action potential) or three properties (repetitiveness, voltage sag and rebound AP), 67 whereas the subplate neurons had higher correlation between all three properties compared to cortical plate and Cajal-Retzius cells. 2.3.4 Single-cell transcriptomic analysis of human Cajal-Retzius neurons scRNA-Seq was performed for 717 cells (235 CR, 275 CP and 207 SP) using the extracted cytoplasm by amplifying the poly-A RNA using a modification of the aRNA protocol (Eberwine et al., 1992; Moll et al., 2004; Morris et al., 2011; Phillips & Eberwine, 1996), followed by library construction using Illumina TruSeq kits and sequencing on the Illumina HiSeq DNA sequencers (see Methods). The sequence reads were trimmed and sequentially aligned to rRNA, mtDNA, the human transcriptome (GENCODE v22) and genome (GRCh38), allowing up to 6 mis-matches, using PerM (Y. Chen et al., 2009), as implemented in GT-FAR (https://genomics.isi.edu/gtfar). Reads that uniquely aligned to a single gene model were assigned to the gene and used as a proxy of gene expression. On average, each cell had 1,098,218 reads, 19.95% of which mapped to exons and detected 6,659 genes with one or more reads. The majority of the cells had exonic mapped reads between 100,000 – 10,000,000 and 1,000 – 10,000 genes expressed. To normalize gene expression in each cell to enable cell to cell comparisons, all the cells were downsampled, without replacement, to 100,000 reads followed by log-normalization, to account for differences in library size. 68 One of the biggest challenges in the single cell transcriptome analysis is the general sparseness of single-cell transcriptome data (van Dijk et al., 2017). Because of the sparseness, dimensional analysis such as PCA cannot be performed with the raw data for cell classification. The sparseness was due to the biological noise and technical noise, which was generated by process of amplification (Brennecke et al., 2013). By separating covariates correlated to each biological noise (Biopsy ID: Tissue source) and technical noise (Batch effect), the cell types were clearly separated in PCA coordinates. The covariates reflecting batch effect (characterized by flowcell), tissue variation (characterized by biopsy ID) (Figure 2-6) and gestational weeks were matched for the PCA plot. In the corrected PCA coordinates, Cajal Retzius cells were clustered separately from cortical plate and subplate neurons (Figure 2-7). The cluster of subplate neurons were located closer to the Cajal Retzius cell cluster than cortical plate cluster. Furthermore, differential gene expression analysis revealed that Cajal- Retzius neurons had significantly higher expression of known marker genes (Gil, Nocentini, & Del Río, 2014; Kirischuk et al., 2014; Windrem et al., 2014; Yamazaki, Sekiguchi, Takamatsu, Tanabe, & Nakanishi, 2004): RELN, CXCR4, CALB2, PCP4, OLFM1 and LHX1 (Figure 2-8). Differential gene expression analysis discovered 376 up-regulated genes (p<0.001 and fold change >2) and 54 down-regulated genes (fold change<2) in Cajal-Retzius neurons compared to cortical plate and subplate neurons (Table 2-3). The 69 heat map of normalized covariates with differentially expressed genes in each neuronal type showed the molecular distinction between the three groups of neurons (Figure 2-9A). Upon further examination of the genes enriched in Cajal-Retzius neurons, we found in addition to known marker genes, we identified many new Cajal-Retzius neuron-specific genes critical for neural development, including transcription factors, cell-adhesion, calcium-regulating, and secreting signaling molecules. Many of the genes are significant signaling molecules for neural development, such as AMIGO2, CD9, CLU, CNTNAP2, CTNNB1, DPP6, NCS1, NDRG4, NR2F2, NRBP2, OLFM1, PTN, SEZ6L, SPINT2, SPOCK2, TENM1, TXN, VCP, ZIC2, ZNF503 (See Table 2-3). When we subjected the Cajal-Retzius neuron up-regulated genes to gene ontology analysis, we found these genes to be highly enriched in major biological processes regulating nervous system development (Table 2-4). In summary, our data identified novel Cajal-Retzius neuron-specific genes significant in their role in cortex development. 2.4 Discussion Our study shows that human Cajal-Retzius neurons are specialized neurons regulating critical aspects of cortical development. They are electrically active, expressing genes pivotal for human brain formation. I report here for the first time that human Cajal-Retzius neurons exhibit spontaneous synaptic activity and action potential firing in early to late second trimester (15- 20 GW). By performing single-cell transcriptome analysis, I show that Cajal- 70 Retzius neurons are molecularly distinct from CP and SP neurons, and I identify new Cajal-Retzius neuron-specific genes such as GRN and TAGLN2. 2.4.1 Spontaneous synaptic activity in human Cajal-Retzius neurons Our data also suggest that Cajal-Retzius neurons receive synaptic inputs and may be active participants in the early cortical network. Human Cajal-Retzius neurons are born at around 5 GW and an electron microscopy study identified synaptic structures in the marginal zone as early as 8.5 GW (Gundela Meyer, Schaaps, Moreau, & Goffinet, 2000; Molliver, Kostovic, & van der Loos, 1973). Our electrophysiological recordings show that there is spontaneous activity in human Cajal-Retzius neurons, mediated through both glutamatergic and GABAergic receptors (Figure 2-4). The transcriptome data supports our electrophysiological findings and previous work on human Cajal- Retzius neurons (Lu, Zecevic, & Yeh, 2001). Electrical activity and neurotransmitter release have been known to regulate neuron migration and establish early synaptic networks (Hua & Smith, 2004; Spitzer, 2006; Zhang & Poo, 2001). Our findings here provide the first evidence that Cajal-Retzius neuron play a role in integrating synaptic information in early human development. 71 2.4.2 Human Cajal-Retzius neurons express secreting and contacting cues Neuronal migration and lamination of the cortex are highly complex and tightly regulated processes. Mutations or defects in the genes coordinating migration can cause severe brain malformation and often lead to psychiatric disorders (Kwan, Sestan, & Anton, 2012). Because of their strategic location, Reelin secretion and expression of contact guidance cues (Gil-Sanz et al., 2013), Cajal-Retzius neurons have always been hypothesized to function as guideposts for neuron migration (Gil-Sanz et al., 2013). Our data strongly support this hypothesis, identifying enrichment of secretion and contact signaling molecules in human Cajal-Retzius neurons. These molecules are not only important for guiding neuronal migration but also critical in regulating neurogenesis, differentiation, cortical lamination, neurite outgrowth and synaptogenesis, the disregulation of which have been implicated in neurological disorders (Table 2-3,Table 2-4). 2.4.3 Cell-adhesion molecules are enriched in human Cajal - Retzius neurons Cell adhesion molecules serve to instruct target recognition, axon guidance and synapse formation in development (Sakurai, 2016). Several cell adhesion molecules were found to be highly expressed in Cajal-Retzius cells in our study. AMIGO2 is a member of cell surface transmembrane proteins expressed in the brain and is implicated in axon tract development, regulation 72 of neural circuit development, dendritic growth and neuronal survival (Yanan Chen, Hor, & Tang, 2012; Kuja-Panula, Kiiltomki, Yamashiro, Rouhiainen, & Rauvala, 2003; Zhao et al., 2014). Another gene, CD9 is a member of the tetraspanin family, composed of cell surface transmembrane glycoproteins. It is expressed in the developing and adult central nervous system (Perron & Bixby, 1999; Tole & Patterson, 1993). In immune and cancer cells, CD9 has been shown to bind to β1 integrins and modulate cell migration (Schmid & Anton, 2003). Furthermore, mutations of CNTNAP2 (contactin associated protein-like 2) have been linked to cognitive disorders (Rodenas-Cuadrado, Ho, & Vernes, 2014). In both human and mice, lack of CNTNAP2 leads to ectopic neurons in the white matter (Peñagarikano et al., 2011; Strauss et al., 2006). Another cell adhesion molecule, TENM1 (teneurin transmembrane protein 1) is known to regulate synapse formation and control target selection (Mosca, 2015; Mosca, Hong, Dani, Favaloro, & Luo, 2012). In contrast to rodents, human Cajal-Retzius neurons express Cadherin 4 (CDH4) rather than Cadherin 2 (Gil-Sanz et al., 2013). Cadherins are calcium-dependent cell adhesion molecules, and are involved in almost all aspects of cortical organization (Hirano & Takeichi, 2012). The relatively high expression of cell- adhesion molecules in human Cajal-Retzius neurons suggests that cell-cell interactions play a major role in human brain development. 73 2.4.4 Wnt signaling molecules in human Cajal-Retzius neurons The Wnt family is involved in many aspects of neural development, from the morphogenesis of the neural tube, neurogenesis to differentiation, axon guidance and synapse formation (Mulligan & Cheyette, 2012). Human Cajal-Retzius neurons preferentially express multiple Wnt-related genes. OLFM1 (noelin) is a secreted glycoprotein structurally similar to Wnt, essential for cell-cell signaling and involved in neuronal differentiation and axon guidance (Moreno & Bronner-Fraser, 2005; Nakaya, Sultana, Lee, & Tomarev, 2012). RSPO3 (R-spondin 3) is a secreted protein that binds to LGR4-6 receptors and activates the canonical Wnt signaling pathway (Carmon, Gong, Lin, Thomas, & Liu, 2011; Glinka et al., 2011; Gong et al., 2012) CTNNB1 (β- catenin 1), forms adherens junctions with cadherins that function to provide cell-cell adhesion and modulate cellular re-organization using actin (Hartsock & Nelson, 2008). Activation of Wnt signaling is known to increase β-catenin expression and cadherins may sequester β-catenin, reducing its availability (Nusse, 2004). β-catenin signaling may determine the laminar position of neural progenitors (Christopher A. Mutch, Nobuo Funatsu, Edwin S. Monuki, 2009), and elevated β-catenin in cultured hippocampal neurons exhibited expansion in dendritic arborization (Yu & Malenka, 2003). Recently, an elegant genetic study investigated the function of adhesion and transcriptional activity of β-catenin independently and showed that neural precursor fate was determined by transcriptional control rather than through cell adhesion 74 (Draganova et al., 2015). Interestingly, they also found an expansion in Cajal- Retzius neuron population when β-catenin transcriptional activity was disrupted. Our data suggest that Cajal-Retzius neurons may control the fate of target cells through secretion of these Wnt-related proteins. 2.4.5 Calcium signaling in human Cajal-Retzius neurons Calcium signaling molecules are preferentially overexpressed in human Cajal-Retzius neurons, and calcium transients are known to regulate neuronal migration, neurotransmitter release, axon guidance, and induce transcriptional change (Komuro & Kumada, 2005; Sutherland, Pujic, & Goodhill, 2014; D. B. Wheeler, Randall, & Tsien, 1994; D. G. Wheeler, Barrett, Groth, Safa, & Tsien, 2008). Calbindin, or calretinin (CALB2) is a calcium binding protein known to be expressed by Cajal-Retzius neurons (Vogt Weisenhorn, Weruaga Prieto, & Celio, 1994). It functions as a calcium sensor, and knocking it out in Purkinje neurons alters excitability (Schwaller, 2014). CAMK2D is a member of the calmodulin-kinases that modulates synaptic plasticity (Wayman, Lee, Tokumitsu, Silva, & Soderling, 2008). SPOCK2 or TESTICAN2 is a calcium- binding proteoglycan expressed in the brain (Yamamoto et al., 2014), and SPOCK2 inhibits neurite outgrowth in culture (Schnepp et al., 2005). THSD7B is a protein-coding gene for thrombospondin. Thrombospondin 1 binds to APOER2 and VLDLR, inducing Dab1 phosphorylation albeit activating a separate signaling cascade rather than the REELIN signaling cascade (Blake et al., 2008). This calcium-binding extracellular glycoprotein has been known 75 to interact with extracellular matrix components and growth factors and supports neurite outgrowth (Adams & Lawler, 2011; Arber & Caroni, 1995; O’Shea, Liu, & Dixita, 1991). Mice lacking either thrombospondin 1 or 2 have lower synaptic density and has showed poor recovery after stroke (Christopherson et al., 2005; Liauw et al., 2008). Neuronal calcium sensor 1 (NCS1) is another calcium-binding protein believed to regulate vesicle trafficking and synaptic plasticity, and its mutation has been associated with autism disorder in humans (Burgoyne & Haynes, 2010; Handley, Lian, Haynes, & Burgoyne, 2010). Together our findings suggest that calcium signaling is a significant contributor in Cajal-Retzius neuron-mediated action. 2.4.6 Differences between human and rodent Cajal-Retzius neurons Most of our current understanding about Cajal-Retzius neurons has been built on rodent studies and limited human histological studies. My data and previous studies (Kirischuk et al., 2014; Ogawa et al., 1995), however, suggest that there are species dependent differences for Cajal-Retzius neurons. As previously discussed, while Cadherin 4 is expressed in human Cajal-Retzius neurons (Table 2-3), a different isoform, Cadherin 2, is expressed in mice. (Gil-Sanz et al., 2013). Receptor-mediated synaptic response also differs between species. In human Cajal-Retzius neurons, glutamatergic responses were mediated through both NMDA and AMPA receptors (Lu et al., 2001) but in mice, only NMDA responses were observed 76 (Lu et al., 2001) and later shown to be strain dependent in mice (Chan & Yeh, 2003; Dvorzhak, Unichenko, & Kirischuk, 2012). Embryonically, rodent Cajal- Retzius neurons exhibit sodium and potassium currents (Albrieux, 2004; Mienville, 1998). Cajal-Retzius neurons in neonatal rats showed repetitive action potential firing between postnatal days 1-5 (Hestrin & Armstrong, 1996; Mienville, Maric, Maric, & Clay, 1999; Radnikow et al., 2002). Our data show that human Cajal-Retzuis neurons are functionally more mature than rodent ones as they exhibit repetitive action potential firing as early as 15.2-16 GW. This study provides electrophysiological and molecular profiling of human Cajal-Retzius neurons from acute brain slices, which best mimics in vivo conditions. Formation of cortical circuitry is a highly sophisticated process requiring precise control in timing and positioning. In this study, we show that human Cajal-Retzius neurons are electrically active prenatally and identify multiple novel Cajal-Retzius neuron specific molecules that may contribute to the proper assembly of the brain architecture. 77 2.5 Figures Figure 2-1. Norepinephrine’s regulation of cortical layer formation. (A) The number of Cajal Retzius cells on postnatal days with 6-OHDA and control (Naqui, et al., 1999). (B) Migration distance in saline, medetomidine and wash treatment (Riccio, et al., 2012). The speed of migration is decreased with application of norepinephrine receptor agonist. 78 Figure 2-2. Microscopic image of the human embryonic brain tissue (A) A low magnitude bright field image shows distinct layers in a 21 GW cerebral cortex. Three major upper layers in the developing cortex are shown: marginal zone, cortical plate and subplate. The cell-dense layer in the middle of the marginal zone is the subpial granular layer. (B), (C) and (D) Sample image of a Cajal-Retzius neuron, a cortical plate neuron and a subplate neuron. (E) and (F) demonstrate a Cajal-Retzius neuron being collected by the micropipette. Scale bars are 10μm. 79 Figure 2-3. Morphological variation in CR neurons (A),(D) Vertical shape, (B),(D) Horizontal shape, (C),(F) Angled morphology of CR neurons. 80 Figure 2-4. Spontaneous Activity observation of Cajal-Retzius Neurons (A) Current clamp recording shows spontaneous excitatory postsynaptic potentials (EPSPs) and action potential (AP) firing from a Cajal-Retzius neuron (top). In another Cajal-Retzius neuron (bottom), both EPSPs and inhibitory postsynaptic potentials (IPSPs) were observed. There is a large EPSP (~40mV) in the bottom trace. Both neurons were clamped at ~-75mV. Scale bar: 10mV, 2s. (B) Quantification of percentage of cells exhibiting spontaneous activity. (C) Of the neurons with spontaneous activity, percentage of neurons exhibiting IPSPs, EPSPs and spontaneous action potential firing are shown. 81 (n=26 Cajal-Retzius neurons, n=15 cortical plate neurons, n=17 subplate neurons). Figure 2-5. Firing pattern with current injection in CR, CP and SP neurons (A) Sample recording traces from a 16.6GW Cajal-Retzius neuron in response to hyperpolarizing (black) and depolarizing (red) current injections. The neuron 82 displays a prominent voltage sag (arrow) with hyperpolarization and rebound action-potential firing (asterisk) following the cessation of a hyperpolarizing step. Depolarization triggered repetitive action potential firing with decreasing amplitude. The membrane potential was held at -65mV. Scale bar: 25mV, 200ms. (B) Action potential firing of each neuron is categorized as no action potential firing, single and repetitive action potential firing. Subplate neurons had the highest percentage of neurons with repetitive firing. (C) Percentage of cells with a voltage sag is shown. (D) Percentage of cells with a rebound action potential is quantified. (n=87 Cajal-Retzius neurons, n=56 cortical plate neurons, n=74 subplate neurons) Subplate neurons had the highest percentage of neurons with repetitive firing. (E) Donut plot indicating the correlation between patterns of electrophysiological properties. The % cell metric indicates which properties were correlated more in each cell type. 83 Figure 2-6. Sources of technical and biological noise (A) PCA plot based on the biopsy numbers (Same tissue). Each color represents different tissue source. Each shape indicates the cell types. (B) PCA plot based on batch effect. Each color represents the batch of RNAseq protocol and each shape indicates the cell types. 84 Figure 2-7. PCA plot of covariates The Covariates matched with gestational week (gw), flowcell and biopsy ID. 85 Figure 2-8. Boxplot of normalized expression of known marker genes for human Cajal Retzius Cell. CALB1, CXCR4, LHX1, PCP4, OLFM1, RELN were highly expressed in Cajal- Retzius cells compared to other cells in embryonic brain. 86 Figure 2-9. Heatmap of the differentially expressed genes (A) Rank of covariates. (B) Rank of raw expression normalized. Each row indicates one gene. 87 2.6 Tables Table 2-1. List of All embryinic brain(EB) cells Used for analysis Cell Biopsy ID AGE(GW) CELLTYPE SPONTANEOUS ACTIVE FIRING PATTERN EB1001 BID103 17.1 CP UNKNOWN UNKNOWN EB1005 BID103 17.1 CP UNKNOWN SINGLE EB1007 BID103 17.1 CP EPSP UNKNOWN EB1013 BID104 10.3 CP UNKNOWN SINGLE EB1015 BID104 10.3 CP EPSP REPETITIVE_ADAPTIVE EB1017 BID104 10.3 CP UNKNOWN UNKNOWN EB1025 BID104 10.3 CP EPSP UNKNOWN EB1027 BID104 10.3 CP EPSP UNKNOWN EB1029 BID104 10.3 CP EPSP SINGLE EB1045 BID104 10.3 CP EPSP REPETITIVE EB1047 BID104 10.3 CP UNKNOWN UNKNOWN EB1049 BID104 10.3 CP UNKNOWN UNKNOWN EB1051 BID104 10.3 CP UNKNOWN UNKNOWN EB1053 BID104 10.3 CP UNKNOWN UNKNOWN EB1097 BID107 18.4 CP UNKNOWN UNKNOWN EB1099 BID107 18.4 CP UNKNOWN UNKNOWN EB1101 BID107 18.4 CP UNKNOWN UNKNOWN EB1121 BID107 18.4 CP UNKNOWN UNKNOWN EB1123 BID107 18.4 CP UNKNOWN UNKNOWN EB1125 BID107 18.4 CP UNKNOWN UNKNOWN EB114 No BID 18.3 CP UNKNOWN UNKNOWN EB1177 BID109 17.1 CP UNKNOWN SINGLE EB1179 BID109 17.1 CP UNKNOWN UNKNOWN EB1181 BID109 17.1 CP UNKNOWN UNKNOWN EB1183 BID109 17.1 CP UNKNOWN UNKNOWN EB1185 BID109 17.1 CP UNKNOWN UNKNOWN EB1187 BID109 17.1 CP UNKNOWN UNKNOWN EB1189 BID109 17.1 CP UNKNOWN UNKNOWN EB1191 BID109 17.1 CP UNKNOWN UNKNOWN 88 EB1193 BID109 17.1 CP UNKNOWN UNKNOWN EB1195 BID109 17.1 CP UNKNOWN SINGLE EB1197 BID109 17.1 CP UNKNOWN SINGLE EB1199 BID109 17.1 CP UNKNOWN SINGLE EB1201 BID109 17.1 CP UNKNOWN SINGLE EB1203 BID109 17.1 CP UNKNOWN UNKNOWN EB1245 BID124 16 CP UNKNOWN UNKNOWN EB1249 BID124 16 CP UNKNOWN UNKNOWN EB1253 BID124 16 CP UNKNOWN SINGLE EB1281 BID128 14.1 CP UNKNOWN UNKNOWN EB1288 BID126 18.5 CP UNKNOWN UNKNOWN EB1389 BID142 14.4 CP UNKNOWN UNKNOWN EB142 No BID 16.0 CP UNKNOWN REPETITIVE_ADAPTIVE EB144 No BID 16.0 CP UNKNOWN REPETITIVE_ADAPTIVE EB1442 BID147 16 CP UNKNOWN SINGLE EB1443 BID147 16 CP UNKNOWN UNKNOWN EB1444 BID147 16 CP UNKNOWN UNKNOWN EB1582 BID153 15.4 CP UNKNOWN UNKNOWN EB253 BID037 11.2 CP UNKNOWN UNKNOWN EB254 BID015 19.1 CP UNKNOWN UNKNOWN EB260 BID015 19.1 CP UNKNOWN UNKNOWN EB261 BID015 19.1 CP UNKNOWN UNKNOWN EB269 BID015 19.1 CP UNKNOWN UNKNOWN EB273 BID015 19.1 CP UNKNOWN UNKNOWN EB274 BID015 19.1 CP EPSP_AP SINGLE EB275 BID015 19.1 CP UNKNOWN NO EB286 BID018 16.3 CP UNKNOWN UNKNOWN EB289 BID018 16.3 CP UNKNOWN UNKNOWN EB290 BID018 16.3 CP UNKNOWN UNKNOWN EB291 BID018 16.3 CP UNKNOWN UNKNOWN EB292 BID018 16.3 CP UNKNOWN UNKNOWN EB293 BID018 16.3 CP UNKNOWN UNKNOWN EB294 BID018 16.3 CP UNKNOWN UNKNOWN EB297 BID018 16.3 CP UNKNOWN UNKNOWN EB299 BID018 16.3 CP UNKNOWN UNKNOWN EB310 BID026 17.5 CP UNKNOWN UNKNOWN EB313 BID026 17.5 CP UNKNOWN UNKNOWN EB315 BID026 17.5 CP UNKNOWN UNKNOWN 89 EB318 BID026 17.5 CP UNKNOWN UNKNOWN EB320 BID026 17.5 CP UNKNOWN UNKNOWN EB321 BID026 17.5 CP UNKNOWN UNKNOWN EB322 BID026 17.5 CP UNKNOWN UNKNOWN EB323 BID026 17.5 CP UNKNOWN UNKNOWN EB324 BID026 17.5 CP UNKNOWN UNKNOWN EB329 BID030 17.0 CP UNKNOWN UNKNOWN EB339 BID044 19.3 CP UNKNOWN UNKNOWN EB353 BID045 17.5 CP UNKNOWN UNKNOWN EB354 BID045 17.5 CP UNKNOWN UNKNOWN EB355 BID045 17.5 CP UNKNOWN UNKNOWN EB359 BID046 19.2 CP UNKNOWN UNKNOWN EB36 BID004 17.6 CP UNKNOWN SINGLE EB360 BID046 19.2 CP UNKNOWN UNKNOWN EB362 BID046 19.2 CP UNKNOWN UNKNOWN EB363 BID046 19.2 CP UNKNOWN UNKNOWN EB365 BID046 19.2 CP UNKNOWN UNKNOWN EB366 BID046 19.2 CP UNKNOWN UNKNOWN EB373 BID046 19.2 CP UNKNOWN UNKNOWN EB375 BID046 19.2 CP UNKNOWN UNKNOWN EB376 BID046 19.2 CP UNKNOWN UNKNOWN EB379 BID047 19.5 CP UNKNOWN UNKNOWN EB380 BID047 19.5 CP UNKNOWN UNKNOWN EB381 BID047 19.5 CP UNKNOWN UNKNOWN EB382 BID047 19.5 CP UNKNOWN UNKNOWN EB383 BID047 19.5 CP UNKNOWN UNKNOWN EB384 BID047 19.5 CP UNKNOWN UNKNOWN EB385 BID047 19.5 CP UNKNOWN UNKNOWN EB387 BID047 19.5 CP UNKNOWN UNKNOWN EB405 BID063 19.6 CP EPSP_AP SINGLE EB406 BID063 19.6 CP UNKNOWN UNKNOWN EB408 BID063 19.6 CP UNKNOWN SINGLE EB410 BID063 19.6 CP UNKNOWN UNKNOWN EB411 BID063 19.6 CP UNKNOWN UNKNOWN EB412 BID063 19.6 CP UNKNOWN UNKNOWN EB414 BID063 19.6 CP EPSP SINGLE EB415 BID063 19.6 CP UNKNOWN UNKNOWN EB416 BID063 19.6 CP EPSP SINGLE 90 EB418 BID063 19.6 CP UNKNOWN UNKNOWN EB420 BID063 19.6 CP UNKNOWN UNKNOWN EB424 BID063 19.6 CP UNKNOWN UNKNOWN EB44 BID052 17.6 CP EPSP_AP SINGLE EB456 BID073 17.6 CP UNKNOWN UNKNOWN EB459 BID074 17.5 CP UNKNOWN UNKNOWN EB461 BID074 17.5 CP UNKNOWN UNKNOWN EB462 BID074 17.5 CP UNKNOWN UNKNOWN EB463 BID081 17.5 CP UNKNOWN UNKNOWN EB464 BID081 17.5 CP UNKNOWN UNKNOWN EB465 BID081 17.5 CP UNKNOWN UNKNOWN EB467 BID081 17.5 CP UNKNOWN UNKNOWN EB469 BID081 17.5 CP UNKNOWN UNKNOWN EB47 BID052 17.6 CP IPSP_EPSP SINGLE EB470 BID081 17.5 CP UNKNOWN UNKNOWN EB471 BID081 17.5 CP UNKNOWN UNKNOWN EB472 BID081 17.5 CP UNKNOWN UNKNOWN EB473 BID081 17.5 CP UNKNOWN UNKNOWN EB474 BID081 17.5 CP UNKNOWN UNKNOWN EB475 BID081 17.5 CP UNKNOWN UNKNOWN EB476 BID081 17.5 CP UNKNOWN UNKNOWN EB477 BID081 17.5 CP UNKNOWN UNKNOWN EB478 BID081 17.5 CP UNKNOWN UNKNOWN EB479 BID081 17.5 CP UNKNOWN UNKNOWN EB480 BID081 17.5 CP UNKNOWN UNKNOWN EB481 BID081 17.5 CP UNKNOWN UNKNOWN EB482 BID081 17.5 CP UNKNOWN UNKNOWN EB483 BID081 17.5 CP UNKNOWN UNKNOWN EB485 BID081 17.5 CP UNKNOWN UNKNOWN EB486 BID081 17.5 CP UNKNOWN UNKNOWN EB487 BID081 17.5 CP UNKNOWN UNKNOWN EB488 BID081 17.5 CP UNKNOWN SINGLE EB489 BID081 17.5 CP UNKNOWN UNKNOWN EB490 BID081 17.5 CP UNKNOWN UNKNOWN EB491 BID081 17.5 CP UNKNOWN UNKNOWN EB492 BID081 17.5 CP UNKNOWN UNKNOWN EB493 BID081 17.5 CP UNKNOWN REPETITIVE_ADAPTIVE EB494 BID081 17.5 CP UNKNOWN UNKNOWN 91 EB495 BID081 17.5 CP UNKNOWN UNKNOWN EB496 BID081 17.5 CP UNKNOWN SINGLE EB497 BID081 17.5 CP UNKNOWN UNKNOWN EB498 BID081 17.5 CP UNKNOWN UNKNOWN EB499 BID081 17.5 CP UNKNOWN UNKNOWN EB501 BID081 17.5 CP UNKNOWN UNKNOWN EB503 BID081 17.5 CP UNKNOWN UNKNOWN EB563 BID084 17.2 CP UNKNOWN UNKNOWN EB565 BID084 17.2 CP UNKNOWN UNKNOWN EB567 BID084 17.2 CP UNKNOWN UNKNOWN EB569 BID084 17.2 CP UNKNOWN UNKNOWN EB571 BID084 17.2 CP UNKNOWN UNKNOWN EB573 BID084 17.2 CP UNKNOWN UNKNOWN EB584 BID084 17.2 CP UNKNOWN UNKNOWN EB587 BID084 17.2 CP UNKNOWN UNKNOWN EB588 BID084 17.2 CP UNKNOWN UNKNOWN EB589 BID083 18.0 CP UNKNOWN UNKNOWN EB591 BID083 18.0 CP UNKNOWN UNKNOWN EB593 BID083 18.0 CP UNKNOWN UNKNOWN EB595 BID083 18.0 CP UNKNOWN UNKNOWN EB606 BID085 19.0 CP UNKNOWN UNKNOWN EB608 BID085 19.0 CP UNKNOWN UNKNOWN EB610 BID085 19.0 CP UNKNOWN SINGLE EB612 BID085 19.0 CP UNKNOWN UNKNOWN EB614 BID085 19.0 CP UNKNOWN UNKNOWN EB616 BID085 19.0 CP UNKNOWN UNKNOWN EB618 BID085 19.0 CP UNKNOWN UNKNOWN EB622 BID085 19.0 CP UNKNOWN UNKNOWN EB624 BID085 19.0 CP UNKNOWN UNKNOWN EB625 BID085 19.0 CP UNKNOWN SINGLE EB626 BID085 19.0 CP UNKNOWN UNKNOWN EB627 BID085 19.0 CP UNKNOWN UNKNOWN EB628 BID085 19.0 CP UNKNOWN UNKNOWN EB650 BID086 15.2 CP UNKNOWN UNKNOWN EB652 BID086 15.2 CP UNKNOWN UNKNOWN EB654 BID086 15.2 CP UNKNOWN UNKNOWN EB656 BID086 15.2 CP UNKNOWN UNKNOWN EB658 BID086 15.2 CP UNKNOWN UNKNOWN 92 EB659 BID086 15.2 CP UNKNOWN SINGLE EB661 BID086 15.2 CP UNKNOWN UNKNOWN EB663 BID086 15.2 CP UNKNOWN UNKNOWN EB664 BID086 15.2 CP UNKNOWN UNKNOWN EB665 BID086 15.2 CP UNKNOWN UNKNOWN EB666 BID086 15.2 CP UNKNOWN UNKNOWN EB667 BID086 15.2 CP UNKNOWN UNKNOWN EB668 BID086 15.2 CP UNKNOWN UNKNOWN EB669 BID086 15.2 CP UNKNOWN UNKNOWN EB670 BID086 15.2 CP UNKNOWN UNKNOWN EB671 BID086 15.2 CP UNKNOWN UNKNOWN EB673 BID086 15.2 CP UNKNOWN UNKNOWN EB674 BID086 15.2 CP UNKNOWN UNKNOWN EB676 BID086 15.2 CP UNKNOWN UNKNOWN EB677 BID086 15.2 CP UNKNOWN UNKNOWN EB678 BID086 15.2 CP EPSP SINGLE EB679 BID086 15.2 CP UNKNOWN UNKNOWN EB690 BID090 17.2 CP UNKNOWN UNKNOWN EB692 BID090 17.2 CP UNKNOWN UNKNOWN EB694 BID090 17.2 CP UNKNOWN UNKNOWN EB696 BID090 17.2 CP UNKNOWN UNKNOWN EB697 BID090 17.2 CP UNKNOWN UNKNOWN EB698 BID090 17.2 CP UNKNOWN UNKNOWN EB699 BID090 17.2 CP UNKNOWN UNKNOWN EB700 BID090 17.2 CP UNKNOWN UNKNOWN EB701 BID090 17.2 CP UNKNOWN UNKNOWN EB702 BID090 17.2 CP UNKNOWN UNKNOWN EB703 BID090 17.2 CP UNKNOWN UNKNOWN EB704 BID090 17.2 CP UNKNOWN UNKNOWN EB705 BID090 17.2 CP UNKNOWN UNKNOWN EB706 BID090 17.2 CP UNKNOWN UNKNOWN EB707 BID090 17.2 CP NO SINGLE EB715 BID090 17.2 CP UNKNOWN UNKNOWN EB717 BID090 17.2 CP UNKNOWN UNKNOWN EB721 BID090 17.2 CP UNKNOWN UNKNOWN EB723 BID090 17.2 CP UNKNOWN UNKNOWN EB725 BID090 17.2 CP UNKNOWN UNKNOWN EB727 BID090 17.2 CP UNKNOWN UNKNOWN 93 EB738 BID094 16.2 CP UNKNOWN UNKNOWN EB740 BID094 16.2 CP UNKNOWN UNKNOWN EB742 BID094 16.2 CP UNKNOWN UNKNOWN EB744 BID094 16.2 CP UNKNOWN UNKNOWN EB746 BID094 16.2 CP UNKNOWN UNKNOWN EB747 BID094 16.2 CP UNKNOWN UNKNOWN EB748 BID094 16.2 CP UNKNOWN UNKNOWN EB750 BID094 16.2 CP UNKNOWN UNKNOWN EB752 BID094 16.2 CP UNKNOWN UNKNOWN EB753 BID094 16.2 CP UNKNOWN UNKNOWN EB754 BID094 16.2 CP UNKNOWN UNKNOWN EB755 BID094 16.2 CP UNKNOWN UNKNOWN EB756 BID094 16.2 CP UNKNOWN SINGLE EB757 BID094 16.2 CP UNKNOWN UNKNOWN EB758 BID094 16.2 CP UNKNOWN UNKNOWN EB759 BID094 16.2 CP UNKNOWN UNKNOWN EB760 BID094 16.2 CP UNKNOWN UNKNOWN EB761 BID094 16.2 CP UNKNOWN UNKNOWN EB762 BID094 16.2 CP UNKNOWN UNKNOWN EB763 BID094 16.2 CP UNKNOWN UNKNOWN EB768 BID096 18.3 CP EPSP SINGLE EB769 BID096 18.3 CP UNKNOWN UNKNOWN EB771 BID096 18.3 CP UNKNOWN SINGLE EB773 BID096 18.3 CP UNKNOWN UNKNOWN EB774 BID096 18.3 CP UNKNOWN SINGLE EB775 BID096 18.3 CP UNKNOWN SINGLE EB777 BID096 18.3 CP UNKNOWN UNKNOWN EB780 BID096 18.3 CP UNKNOWN SINGLE EB782 BID096 18.3 CP UNKNOWN UNKNOWN EB784 BID096 18.3 CP UNKNOWN UNKNOWN EB786 BID096 18.3 CP UNKNOWN UNKNOWN EB788 BID096 18.3 CP UNKNOWN UNKNOWN EB790 BID096 18.3 CP UNKNOWN UNKNOWN EB792 BID096 18.3 CP UNKNOWN UNKNOWN EB794 BID096 18.3 CP UNKNOWN SINGLE EB796 BID096 18.3 CP UNKNOWN SINGLE EB798 BID096 18.3 CP UNKNOWN SINGLE EB800 BID096 18.3 CP UNKNOWN REPETITIVE_ADAPTIVE 94 EB802 BID096 18.3 CP UNKNOWN SINGLE EB804 BID096 18.3 CP UNKNOWN UNKNOWN EB806 BID096 18.3 CP UNKNOWN SINGLE EB808 BID096 18.3 CP UNKNOWN UNKNOWN EB81 BID006 17.4 CP EPSP_AP SINGLE EB810 BID096 18.3 CP UNKNOWN SINGLE EB825 BID099 20.4 CP UNKNOWN UNKNOWN EB827 BID099 20.4 CP UNKNOWN UNKNOWN EB83 BID006 17.4 CP UNKNOWN UNKNOWN EB839 BID100 17 CP UNKNOWN UNKNOWN EB841 BID100 17 CP UNKNOWN UNKNOWN EB842 BID100 17 CP UNKNOWN SINGLE EB843 BID100 17 CP UNKNOWN REPETITIVE_ADAPTIVE EB844 BID100 17 CP UNKNOWN NO EB846 BID100 17 CP UNKNOWN UNKNOWN EB848 BID100 17 CP UNKNOWN UNKNOWN EB877 BID101 18.1 CP UNKNOWN SINGLE EB879 BID101 18.1 CP UNKNOWN UNKNOWN EB881 BID101 18.1 CP UNKNOWN UNKNOWN EB89 BID006 17.4 CP EPSP SINGLE EB935 BID102 16.6 CP UNKNOWN UNKNOWN EB937 BID102 16.6 CP UNKNOWN SINGLE EB939 BID102 16.6 CP UNKNOWN UNKNOWN EB94 BID006 17.4 CP EPSP SINGLE EB945 BID102 16.6 CP UNKNOWN UNKNOWN EB96 BID006 17.4 CP UNKNOWN SINGLE EB976 BID103 17.1 CP UNKNOWN UNKNOWN EB978 BID103 17.1 CP UNKNOWN UNKNOWN EB991 BID103 17.1 CP UNKNOWN UNKNOWN EB993 BID103 17.1 CP UNKNOWN UNKNOWN EB995 BID103 17.1 CP UNKNOWN UNKNOWN EB997 BID103 17.1 CP UNKNOWN UNKNOWN EB999 BID103 17.1 CP UNKNOWN SINGLE EB1020 BID104 10.3 CR UNKNOWN SINGLE EB1028 BID104 10.3 CR UNKNOWN SINGLE EB1034 BID104 10.3 CR UNKNOWN UNKNOWN EB1036 BID104 10.3 CR UNKNOWN UNKNOWN EB1038 BID104 10.3 CR UNKNOWN UNKNOWN 95 EB1054 BID105 16.5 CR UNKNOWN UNKNOWN EB1176 BID109 17.1 CR UNKNOWN SINGLE EB277 BID015 19.1 CR UNKNOWN UNKNOWN EB278 BID015 19.1 CR UNKNOWN UNKNOWN EB306 BID060 17.5 CR UNKNOWN SINGLE EB311 BID026 17.5 CR EPSP SINGLE EB312 BID026 17.5 CR UNKNOWN SINGLE EB314 BID026 17.5 CR EPSP SINGLE EB316 BID026 17.5 CR UNKNOWN UNKNOWN EB317 BID026 17.5 CR UNKNOWN UNKNOWN EB319 BID026 17.5 CR UNKNOWN UNKNOWN EB340 BID045 17.5 CR UNKNOWN SINGLE EB341 BID045 17.5 CR UNKNOWN UNKNOWN EB342 BID045 17.5 CR UNKNOWN UNKNOWN EB343 BID045 17.5 CR UNKNOWN UNKNOWN EB344 BID045 17.5 CR UNKNOWN UNKNOWN EB345 BID045 17.5 CR UNKNOWN UNKNOWN EB346 BID045 17.5 CR UNKNOWN UNKNOWN EB348 BID045 17.5 CR UNKNOWN UNKNOWN EB349 BID045 17.5 CR UNKNOWN UNKNOWN EB352 BID045 17.5 CR UNKNOWN UNKNOWN EB378 BID046 19.2 CR UNKNOWN SINGLE EB400 BID063 19.6 CR UNKNOWN UNKNOWN EB401 BID063 19.6 CR UNKNOWN UNKNOWN EB402 BID063 19.6 CR UNKNOWN UNKNOWN EB404 BID063 19.6 CR UNKNOWN UNKNOWN EB409 BID063 19.6 CR UNKNOWN UNKNOWN EB421 BID063 19.6 CR UNKNOWN UNKNOWN EB422 BID063 19.6 CR UNKNOWN UNKNOWN EB423 BID063 19.6 CR UNKNOWN UNKNOWN EB425 BID063 19.6 CR UNKNOWN UNKNOWN EB426 BID063 19.6 CR UNKNOWN UNKNOWN EB427 BID063 19.6 CR NO SINGLE EB428 BID063 19.6 CR UNKNOWN UNKNOWN EB429 BID063 19.6 CR NO SINGLE EB430 BID063 19.6 CR EPSP_AP REPETITIVE_ADAPTIVE EB431 BID067 16.2 CR UNKNOWN SINGLE EB434 BID067 16.2 CR UNKNOWN UNKNOWN 96 EB435 BID067 16.2 CR UNKNOWN UNKNOWN EB436 BID067 16.2 CR UNKNOWN UNKNOWN EB437 BID067 16.2 CR UNKNOWN UNKNOWN EB438 BID067 16.2 CR NO SINGLE EB439 BID067 16.2 CR UNKNOWN SINGLE EB440 BID067 16.2 CR UNKNOWN UNKNOWN EB441 BID067 16.2 CR UNKNOWN UNKNOWN EB442 BID067 16.2 CR EPSP SINGLE EB443 BID067 16.2 CR UNKNOWN SINGLE EB444 BID067 16.2 CR UNKNOWN UNKNOWN EB445 BID067 16.2 CR UNKNOWN UNKNOWN EB446 BID067 16.2 CR UNKNOWN UNKNOWN EB447 BID067 16.2 CR UNKNOWN UNKNOWN EB448 BID067 16.2 CR UNKNOWN UNKNOWN EB449 BID072 18.6 CR UNKNOWN UNKNOWN EB450 BID072 18.6 CR UNKNOWN UNKNOWN EB451 BID072 18.6 CR UNKNOWN UNKNOWN EB452 BID072 18.6 CR UNKNOWN UNKNOWN EB453 BID072 18.6 CR UNKNOWN SINGLE EB454 BID072 18.6 CR UNKNOWN UNKNOWN EB455 BID072 18.6 CR UNKNOWN UNKNOWN EB505 BID082 16.0 CR UNKNOWN UNKNOWN EB507 BID082 16.0 CR UNKNOWN UNKNOWN EB508 BID082 16.0 CR UNKNOWN UNKNOWN EB511 BID082 16.0 CR UNKNOWN UNKNOWN EB512 BID082 16.0 CR UNKNOWN UNKNOWN EB513 BID082 16.0 CR UNKNOWN UNKNOWN EB514 BID082 16.0 CR UNKNOWN UNKNOWN EB516 BID082 16.0 CR UNKNOWN UNKNOWN EB521 BID082 16.0 CR UNKNOWN UNKNOWN EB522 BID082 16.0 CR UNKNOWN UNKNOWN EB523 BID082 16.0 CR UNKNOWN SINGLE EB524 BID082 16.0 CR UNKNOWN UNKNOWN EB525 BID082 16.0 CR UNKNOWN UNKNOWN EB526 BID082 16.0 CR UNKNOWN UNKNOWN EB528 BID082 16.0 CR UNKNOWN UNKNOWN EB531 BID082 16.0 CR UNKNOWN SINGLE EB533 BID082 16.0 CR UNKNOWN UNKNOWN 97 EB535 BID082 16.0 CR UNKNOWN UNKNOWN EB536 BID082 16.0 CR UNKNOWN SINGLE EB537 BID082 16.0 CR UNKNOWN REPETITIVE_ADAPTIVE EB541 BID082 16.0 CR UNKNOWN SINGLE EB542 BID082 16.0 CR UNKNOWN UNKNOWN EB544 BID082 16.0 CR UNKNOWN SINGLE EB545 BID082 16.0 CR EPSP SINGLE EB547 BID082 16.0 CR UNKNOWN SINGLE EB548 BID082 16.0 CR UNKNOWN UNKNOWN EB549 BID082 16.0 CR UNKNOWN SINGLE EB550 BID082 16.0 CR UNKNOWN UNKNOWN EB551 BID082 16.0 CR UNKNOWN SINGLE EB552 BID082 16.0 CR EPSP SINGLE EB553 BID082 16.0 CR UNKNOWN UNKNOWN EB554 BID082 16.0 CR UNKNOWN UNKNOWN EB555 BID082 16.0 CR UNKNOWN UNKNOWN EB556 BID082 16.0 CR UNKNOWN UNKNOWN EB557 BID082 16.0 CR UNKNOWN UNKNOWN EB558 BID082 16.0 CR UNKNOWN UNKNOWN EB559 BID082 16.0 CR UNKNOWN UNKNOWN EB560 BID082 16.0 CR UNKNOWN SINGLE EB561 BID082 16.0 CR UNKNOWN SINGLE EB562 BID084 17.2 CR UNKNOWN UNKNOWN EB564 BID084 17.2 CR UNKNOWN UNKNOWN EB566 BID084 17.2 CR UNKNOWN UNKNOWN EB568 BID084 17.2 CR UNKNOWN UNKNOWN EB570 BID084 17.2 CR UNKNOWN UNKNOWN EB572 BID084 17.2 CR UNKNOWN UNKNOWN EB574 BID084 17.2 CR UNKNOWN UNKNOWN EB576 BID084 17.2 CR UNKNOWN UNKNOWN EB578 BID084 17.2 CR UNKNOWN UNKNOWN EB580 BID084 17.2 CR UNKNOWN UNKNOWN EB582 BID084 17.2 CR UNKNOWN UNKNOWN EB590 BID083 18.0 CR UNKNOWN UNKNOWN EB592 BID083 18.0 CR NO SINGLE EB594 BID083 18.0 CR UNKNOWN UNKNOWN EB598 BID083 18.0 CR UNKNOWN UNKNOWN EB599 BID085 19.0 CR UNKNOWN UNKNOWN 98 EB600 BID085 19.0 CR UNKNOWN UNKNOWN EB601 BID085 19.0 CR UNKNOWN UNKNOWN EB602 BID085 19.0 CR UNKNOWN UNKNOWN EB603 BID085 19.0 CR UNKNOWN SINGLE EB604 BID085 19.0 CR UNKNOWN UNKNOWN EB605 BID085 19.0 CR UNKNOWN UNKNOWN EB607 BID085 19.0 CR UNKNOWN UNKNOWN EB609 BID085 19.0 CR UNKNOWN UNKNOWN EB611 BID085 19.0 CR UNKNOWN UNKNOWN EB620 BID085 19.0 CR UNKNOWN UNKNOWN EB621 BID085 19.0 CR UNKNOWN UNKNOWN EB623 BID085 19.0 CR UNKNOWN UNKNOWN EB629 BID085 19.0 CR UNKNOWN UNKNOWN EB631 BID085 19.0 CR UNKNOWN UNKNOWN EB632 BID085 19.0 CR UNKNOWN UNKNOWN EB633 BID086 15.2 CR UNKNOWN UNKNOWN EB634 BID086 15.2 CR UNKNOWN UNKNOWN EB635 BID086 15.2 CR UNKNOWN UNKNOWN EB637 BID086 15.2 CR UNKNOWN UNKNOWN EB638 BID086 15.2 CR UNKNOWN UNKNOWN EB639 BID086 15.2 CR UNKNOWN SINGLE EB640 BID086 15.2 CR UNKNOWN SINGLE EB641 BID086 15.2 CR UNKNOWN UNKNOWN EB642 BID086 15.2 CR UNKNOWN UNKNOWN EB643 BID086 15.2 CR UNKNOWN SINGLE EB645 BID086 15.2 CR EPSP SINGLE EB646 BID086 15.2 CR UNKNOWN UNKNOWN EB647 BID086 15.2 CR UNKNOWN UNKNOWN EB648 BID086 15.2 CR EPSP_AP SINGLE EB649 BID086 15.2 CR UNKNOWN SINGLE EB651 BID086 15.2 CR UNKNOWN UNKNOWN EB653 BID086 15.2 CR UNKNOWN UNKNOWN EB660 BID086 15.2 CR UNKNOWN UNKNOWN EB662 BID086 15.2 CR UNKNOWN UNKNOWN EB680 BID090 17.2 CR UNKNOWN UNKNOWN EB681 BID090 17.2 CR EPSP_IPSP SINGLE EB682 BID090 17.2 CR UNKNOWN UNKNOWN EB683 BID090 17.2 CR UNKNOWN UNKNOWN 99 EB684 BID090 17.2 CR UNKNOWN UNKNOWN EB685 BID090 17.2 CR UNKNOWN UNKNOWN EB687 BID090 17.2 CR UNKNOWN UNKNOWN EB689 BID090 17.2 CR UNKNOWN UNKNOWN EB691 BID090 17.2 CR UNKNOWN UNKNOWN EB693 BID090 17.2 CR UNKNOWN UNKNOWN EB728 BID094 16.2 CR NO REPETITIVE_ADAPTIVE EB729 BID094 16.2 CR UNKNOWN UNKNOWN EB730 BID094 16.2 CR UNKNOWN UNKNOWN EB731 BID094 16.2 CR UNKNOWN SINGLE EB732 BID094 16.2 CR EPSP SINGLE EB733 BID094 16.2 CR NO SINGLE EB735 BID094 16.2 CR NO SINGLE EB736 BID094 16.2 CR UNKNOWN UNKNOWN EB737 BID094 16.2 CR UNKNOWN UNKNOWN EB739 BID094 16.2 CR EPSP_IPSP SINGLE EB741 BID094 16.2 CR UNKNOWN REPETITIVE_ADAPTIVE EB743 BID094 16.2 CR EPSP_AP REPETITIVE_ADAPTIVE EB745 BID094 16.2 CR UNKNOWN UNKNOWN EB749 BID094 16.2 CR UNKNOWN REPETITIVE_ADAPTIVE EB751 BID094 16.2 CR EPSP REPETITIVE_ADAPTIVE EB764 BID096 18.3 CR UNKNOWN UNKNOWN EB765 BID096 18.3 CR UNKNOWN UNKNOWN EB766 BID096 18.3 CR UNKNOWN UNKNOWN EB767 BID096 18.3 CR UNKNOWN UNKNOWN EB779 BID096 18.3 CR UNKNOWN UNKNOWN EB783 BID096 18.3 CR UNKNOWN UNKNOWN EB785 BID096 18.3 CR UNKNOWN SINGLE EB787 BID096 18.3 CR NO SINGLE EB789 BID096 18.3 CR UNKNOWN UNKNOWN EB791 BID096 18.3 CR UNKNOWN UNKNOWN EB793 BID096 18.3 CR UNKNOWN SINGLE EB795 BID096 18.3 CR UNKNOWN SINGLE EB811 BID099 20.4 CR UNKNOWN UNKNOWN EB813 BID099 20.4 CR UNKNOWN UNKNOWN EB829 BID099 20.4 CR UNKNOWN UNKNOWN EB831 BID099 20.4 CR UNKNOWN UNKNOWN EB832 BID099 20.4 CR UNKNOWN UNKNOWN 100 EB916 BID102 16.6 CR UNKNOWN UNKNOWN EB917 BID102 16.6 CR UNKNOWN UNKNOWN EB918 BID102 16.6 CR UNKNOWN UNKNOWN EB919 BID102 16.6 CR UNKNOWN UNKNOWN EB920 BID102 16.6 CR UNKNOWN SINGLE EB921 BID102 16.6 CR UNKNOWN REPETITIVE_ADAPTIVE EB922 BID102 16.6 CR UNKNOWN UNKNOWN EB923 BID102 16.6 CR UNKNOWN UNKNOWN EB924 BID102 16.6 CR UNKNOWN SINGLE EB925 BID102 16.6 CR UNKNOWN UNKNOWN EB926 BID102 16.6 CR UNKNOWN UNKNOWN EB927 BID102 16.6 CR UNKNOWN SINGLE EB928 BID102 16.6 CR UNKNOWN UNKNOWN EB929 BID102 16.6 CR UNKNOWN UNKNOWN EB930 BID102 16.6 CR UNKNOWN NO EB931 BID102 16.6 CR UNKNOWN SINGLE EB932 BID102 16.6 CR UNKNOWN SINGLE EB933 BID102 16.6 CR UNKNOWN SINGLE EB934 BID102 16.6 CR UNKNOWN SINGLE EB936 BID102 16.6 CR NO SINGLE EB944 BID102 16.6 CR UNKNOWN SINGLE EB946 BID102 16.6 CR UNKNOWN SINGLE EB947 BID102 16.6 CR UNKNOWN UNKNOWN EB948 BID102 16.6 CR UNKNOWN UNKNOWN EB950 BID102 16.6 CR UNKNOWN SINGLE EB953 BID102 16.6 CR EPSP_AP REPETITIVE_ADAPTIVE EB955 BID102 16.6 CR UNKNOWN SINGLE EB956 BID102 16.6 CR NO SINGLE EB957 BID102 16.6 CR UNKNOWN REPETITIVE_ADAPTIVE EB959 BID102 16.6 CR EPSP_AP SINGLE EB961 BID102 16.6 CR NO SINGLE EB962 BID102 16.6 CR UNKNOWN SINGLE EB963 BID102 16.6 CR NO SINGLE EB964 BID102 16.6 CR UNKNOWN UNKNOWN EB965 BID102 16.6 CR UNKNOWN UNKNOWN EB966 BID102 16.6 CR UNKNOWN UNKNOWN EB967 BID102 16.6 CR UNKNOWN UNKNOWN EB968 BID102 16.6 CR UNKNOWN UNKNOWN 101 EB969 BID102 16.6 CR UNKNOWN UNKNOWN EB970 BID102 16.6 CR UNKNOWN UNKNOWN EB971 BID102 16.6 CR UNKNOWN UNKNOWN EB972 BID102 16.6 CR UNKNOWN UNKNOWN EB973 BID102 16.6 CR UNKNOWN UNKNOWN EB974 BID102 16.6 CR UNKNOWN UNKNOWN EB975 BID102 16.6 CR UNKNOWN UNKNOWN EB977 BID103 17.1 CR UNKNOWN SINGLE EB979 BID103 17.1 CR UNKNOWN UNKNOWN EB981 BID103 17.1 CR UNKNOWN UNKNOWN EB983 BID103 17.1 CR UNKNOWN UNKNOWN EB985 BID103 17.1 CR UNKNOWN UNKNOWN EB987 BID103 17.1 CR UNKNOWN UNKNOWN EB989 BID103 17.1 CR UNKNOWN UNKNOWN EB1009 BID103 17.1 SP UNKNOWN UNKNOWN EB103 No BID 18.4 SP EPSP_AP SINGLE EB1031 BID104 10.3 SP UNKNOWN REPETITIVE_ADAPTIVE EB1033 BID104 10.3 SP UNKNOWN UNKNOWN EB1035 BID104 10.3 SP UNKNOWN UNKNOWN EB1037 BID104 10.3 SP UNKNOWN UNKNOWN EB1039 BID104 10.3 SP UNKNOWN UNKNOWN EB1041 BID104 10.3 SP UNKNOWN UNKNOWN EB1043 BID104 10.3 SP UNKNOWN UNKNOWN EB1055 BID106 18.1 SP UNKNOWN UNKNOWN EB1057 BID106 18.1 SP UNKNOWN UNKNOWN EB1059 BID106 18.1 SP UNKNOWN UNKNOWN EB1088 BID107 18.4 SP UNKNOWN SINGLE EB1090 BID107 18.4 SP UNKNOWN UNKNOWN EB1092 BID107 18.4 SP UNKNOWN UNKNOWN EB1094 BID107 18.4 SP UNKNOWN UNKNOWN EB1096 BID107 18.4 SP UNKNOWN SINGLE EB1098 BID107 18.4 SP UNKNOWN UNKNOWN EB1116 BID107 18.4 SP UNKNOWN REPETITIVE_ADAPTIVE EB1127 BID107 18.4 SP UNKNOWN UNKNOWN EB1129 BID107 18.4 SP UNKNOWN UNKNOWN EB1131 BID107 18.4 SP UNKNOWN UNKNOWN EB1133 BID107 18.4 SP UNKNOWN UNKNOWN EB1135 BID107 18.4 SP UNKNOWN UNKNOWN 102 EB1138 BID108 16.2 SP UNKNOWN UNKNOWN EB115 No BID 18.3 SP EPSP_AP SINGLE EB1194 BID109 17.1 SP UNKNOWN UNKNOWN EB1205 BID109 17.1 SP UNKNOWN SINGLE EB1207 BID109 17.1 SP UNKNOWN UNKNOWN EB1209 BID109 17.1 SP UNKNOWN SINGLE EB1211 BID109 17.1 SP UNKNOWN UNKNOWN EB128 No BID 18.3 SP IPSP_EPSP SINGLE EB132 No BID 18.3 SP IPSP_EPSP SINGLE EB14 BID001 15.3 SP UNKNOWN SINGLE EB1498 BID149 16.3 SP UNKNOWN UNKNOWN EB150 No BID 16.0 SP UNKNOWN SINGLE EB164 No BID 9.2 SP UNKNOWN SINGLE EB165 No BID 9.2 SP UNKNOWN NO EB166 No BID 9.2 SP UNKNOWN NO EB167 No BID 9.2 SP UNKNOWN NO EB168 No BID 9.2 SP UNKNOWN UNKNOWN EB169 No BID 9.2 SP UNKNOWN UNKNOWN EB172 No BID 9.2 SP UNKNOWN UNKNOWN EB173 No BID 9.2 SP UNKNOWN NO EB175 BID007 18.0 SP UNKNOWN REPETITIVE_ADAPTIVE EB180 BID007 18.0 SP UNKNOWN UNKNOWN EB19 BID001 15.3 SP EPSP_AP UNKNOWN EB210 BID008 18.6 SP UNKNOWN SINGLE EB215 BID009 19.0 SP EPSP REPETITIVE_ADAPTIVE EB217 BID009 19.0 SP UNKNOWN UNKNOWN EB235 BID011 11.6 SP UNKNOWN UNKNOWN EB236 BID011 11.6 SP UNKNOWN UNKNOWN EB246 BID012 17.6 SP UNKNOWN UNKNOWN EB248 BID012 17.6 SP UNKNOWN UNKNOWN EB255 BID015 19.1 SP UNKNOWN UNKNOWN EB256 BID015 19.1 SP UNKNOWN SINGLE EB257 BID015 19.1 SP UNKNOWN UNKNOWN EB258 BID015 19.1 SP UNKNOWN UNKNOWN EB259 BID015 19.1 SP UNKNOWN UNKNOWN EB263 BID015 19.1 SP UNKNOWN UNKNOWN EB264 BID015 19.1 SP UNKNOWN UNKNOWN EB265 BID015 19.1 SP UNKNOWN UNKNOWN 103 EB266 BID015 19.1 SP UNKNOWN UNKNOWN EB267 BID015 19.1 SP UNKNOWN SINGLE EB268 BID015 19.1 SP UNKNOWN UNKNOWN EB27 BID001 15.3 SP EPSP_AP REPETITIVE_ADAPTIVE EB270 BID015 19.1 SP UNKNOWN SINGLE EB271 BID015 19.1 SP UNKNOWN SINGLE EB272 BID015 19.1 SP UNKNOWN UNKNOWN EB276 BID015 19.1 SP UNKNOWN UNKNOWN EB281 BID015 19.1 SP UNKNOWN UNKNOWN EB282 BID018 16.3 SP UNKNOWN SINGLE EB283 BID018 16.3 SP UNKNOWN SINGLE EB284 BID018 16.3 SP UNKNOWN UNKNOWN EB285 BID018 16.3 SP UNKNOWN UNKNOWN EB287 BID018 16.3 SP UNKNOWN SINGLE EB288 BID018 16.3 SP UNKNOWN SINGLE EB298 BID018 16.3 SP UNKNOWN SINGLE EB300 BID018 16.3 SP UNKNOWN SINGLE EB301 BID018 16.3 SP UNKNOWN UNKNOWN EB302 BID018 16.3 SP UNKNOWN UNKNOWN EB303 BID018 16.3 SP UNKNOWN UNKNOWN EB304 BID018 16.3 SP UNKNOWN SINGLE EB305 BID018 16.3 SP UNKNOWN UNKNOWN EB331 BID030 17.0 SP UNKNOWN UNKNOWN EB336 BID044 19.3 SP UNKNOWN UNKNOWN EB337 BID044 19.3 SP UNKNOWN SINGLE EB338 BID044 19.3 SP UNKNOWN UNKNOWN EB347 BID045 17.5 SP UNKNOWN UNKNOWN EB350 BID045 17.5 SP UNKNOWN UNKNOWN EB351 BID045 17.5 SP UNKNOWN UNKNOWN EB356 BID045 17.5 SP UNKNOWN UNKNOWN EB357 BID045 17.5 SP UNKNOWN UNKNOWN EB358 BID045 17.5 SP UNKNOWN UNKNOWN EB364 BID046 19.2 SP UNKNOWN UNKNOWN EB368 BID046 19.2 SP UNKNOWN SINGLE EB370 BID046 19.2 SP UNKNOWN SINGLE EB371 BID046 19.2 SP UNKNOWN SINGLE EB372 BID046 19.2 SP UNKNOWN UNKNOWN EB374 BID046 19.2 SP UNKNOWN NO 104 EB377 BID046 19.2 SP UNKNOWN UNKNOWN EB377 BID046 19.2 SP UNKNOWN UNKNOWN EB386 BID047 19.5 SP UNKNOWN UNKNOWN EB388 BID047 19.5 SP UNKNOWN UNKNOWN EB390 BID047 19.5 SP UNKNOWN UNKNOWN EB391 BID047 19.5 SP UNKNOWN UNKNOWN EB392 BID047 19.5 SP UNKNOWN UNKNOWN EB393 BID047 19.5 SP UNKNOWN UNKNOWN EB395 BID047 19.5 SP NO SINGLE EB396 BID047 19.5 SP UNKNOWN UNKNOWN EB397 BID047 19.5 SP UNKNOWN UNKNOWN EB398 BID047 19.5 SP UNKNOWN UNKNOWN EB399 BID063 19.6 SP UNKNOWN UNKNOWN EB413 BID063 19.6 SP NO REPETITIVE_ADAPTIVE EB417 BID063 19.6 SP UNKNOWN UNKNOWN EB419 BID063 19.6 SP UNKNOWN UNKNOWN EB45 BID052 17.6 SP UNKNOWN REPETITIVE_ADAPTIVE EB46 BID052 17.6 SP UNKNOWN UNKNOWN EB460 BID074 17.5 SP UNKNOWN UNKNOWN EB50 No BID 9.4 SP UNKNOWN REPETITIVE_ADAPTIVE EB53 No BID 9.4 SP IPSP_EPSP_AP SINGLE EB54 No BID 9.4 SP UNKNOWN REPETITIVE_ADAPTIVE EB575 BID084 17.2 SP UNKNOWN UNKNOWN EB577 BID084 17.2 SP UNKNOWN UNKNOWN EB579 BID084 17.2 SP UNKNOWN UNKNOWN EB58 No BID 9.4 SP UNKNOWN SINGLE EB581 BID084 17.2 SP UNKNOWN UNKNOWN EB583 BID084 17.2 SP UNKNOWN UNKNOWN EB585 BID084 17.2 SP UNKNOWN UNKNOWN EB59 No BID 9.4 SP UNKNOWN SINGLE EB60 No BID 9.4 SP EPSP_AP SINGLE EB613 BID085 19.0 SP UNKNOWN UNKNOWN EB615 BID085 19.0 SP UNKNOWN REPETITIVE_ADAPTIVE EB617 BID085 19.0 SP UNKNOWN SINGLE EB619 BID085 19.0 SP UNKNOWN SINGLE EB62 No BID 9.4 SP UNKNOWN REPETITIVE_ADAPTIVE EB630 BID085 19.0 SP UNKNOWN SINGLE EB64 No BID 9.4 SP AP REPETITIVE_ADAPTIVE 105 EB655 BID086 15.2 SP UNKNOWN UNKNOWN EB686 BID090 17.2 SP UNKNOWN UNKNOWN EB688 BID090 17.2 SP UNKNOWN UNKNOWN EB708 BID090 17.2 SP UNKNOWN UNKNOWN EB709 BID090 17.2 SP UNKNOWN UNKNOWN EB71 BID006 17.4 SP UNKNOWN UNKNOWN EB710 BID090 17.2 SP UNKNOWN SINGLE EB711 BID090 17.2 SP UNKNOWN UNKNOWN EB712 BID090 17.2 SP UNKNOWN UNKNOWN EB713 BID090 17.2 SP UNKNOWN SINGLE EB714 BID090 17.2 SP UNKNOWN UNKNOWN EB716 BID090 17.2 SP UNKNOWN UNKNOWN EB718 BID090 17.2 SP UNKNOWN SINGLE EB720 BID090 17.2 SP UNKNOWN UNKNOWN EB722 BID090 17.2 SP UNKNOWN UNKNOWN EB724 BID090 17.2 SP UNKNOWN UNKNOWN EB726 BID090 17.2 SP UNKNOWN UNKNOWN EB772 BID096 18.3 SP UNKNOWN UNKNOWN EB776 BID096 18.3 SP UNKNOWN UNKNOWN EB778 BID096 18.3 SP UNKNOWN UNKNOWN EB797 BID096 18.3 SP UNKNOWN UNKNOWN EB799 BID096 18.3 SP UNKNOWN UNKNOWN EB801 BID096 18.3 SP UNKNOWN UNKNOWN EB803 BID096 18.3 SP UNKNOWN UNKNOWN EB805 BID096 18.3 SP UNKNOWN SINGLE EB807 BID096 18.3 SP UNKNOWN UNKNOWN EB809 BID096 18.3 SP UNKNOWN UNKNOWN EB815 BID099 20.4 SP UNKNOWN UNKNOWN EB817 BID099 20.4 SP UNKNOWN SINGLE EB819 BID099 20.4 SP UNKNOWN SINGLE EB821 BID099 20.4 SP UNKNOWN UNKNOWN EB823 BID099 20.4 SP UNKNOWN SINGLE EB833 BID100 17 SP UNKNOWN REPETITIVE_ADAPTIVE EB835 BID100 17 SP UNKNOWN SINGLE EB837 BID100 17 SP UNKNOWN UNKNOWN EB840 BID100 17 SP UNKNOWN UNKNOWN EB845 BID100 17 SP UNKNOWN SINGLE EB847 BID100 17 SP UNKNOWN UNKNOWN 106 EB849 BID100 17 SP UNKNOWN REPETITIVE_ADAPTIVE EB851 BID100 17 SP UNKNOWN SINGLE EB853 BID100 17 SP UNKNOWN REPETITIVE_ADAPTIVE EB855 BID100 17 SP UNKNOWN REPETITIVE_ADAPTIVE EB857 BID100 17 SP UNKNOWN SINGLE EB859 BID100 17 SP UNKNOWN REPETITIVE_ADAPTIVE EB86 BID006 17.4 SP UNKNOWN SINGLE EB861 BID100 17 SP UNKNOWN UNKNOWN EB863 BID100 17 SP UNKNOWN REPETITIVE_NONADAPTIVE EB865 BID100 17 SP UNKNOWN REPETITIVE_ADAPTIVE EB867 BID100 17 SP UNKNOWN SINGLE EB869 BID100 17 SP UNKNOWN REPETITIVE_ADAPTIVE EB871 BID101 18.1 SP UNKNOWN SINGLE EB873 BID101 18.1 SP UNKNOWN UNKNOWN EB875 BID101 18.1 SP UNKNOWN SINGLE EB88 BID006 17.4 SP IPSP_EPSP_AP REPETITIVE_ADAPTIVE EB883 BID101 18.1 SP UNKNOWN UNKNOWN EB885 BID101 18.1 SP UNKNOWN UNKNOWN EB887 BID101 18.1 SP UNKNOWN SINGLE EB889 BID101 18.1 SP UNKNOWN UNKNOWN EB891 BID101 18.1 SP UNKNOWN UNKNOWN EB893 BID101 18.1 SP UNKNOWN SINGLE EB895 BID101 18.1 SP UNKNOWN SINGLE EB897 BID101 18.1 SP UNKNOWN UNKNOWN EB899 BID101 18.1 SP UNKNOWN UNKNOWN EB901 BID101 18.1 SP UNKNOWN UNKNOWN EB903 BID101 18.1 SP UNKNOWN UNKNOWN EB905 BID101 18.1 SP UNKNOWN UNKNOWN EB907 BID101 18.1 SP UNKNOWN UNKNOWN EB909 BID101 18.1 SP UNKNOWN UNKNOWN EB911 BID101 18.1 SP UNKNOWN UNKNOWN EB913 BID101 18.1 SP UNKNOWN UNKNOWN EB915 BID101 18.1 SP UNKNOWN UNKNOWN EB938 BID102 16.6 SP UNKNOWN UNKNOWN EB98 No BID 18.4 SP EPSP_AP UNKNOWN EB980 BID103 17.1 SP UNKNOWN UNKNOWN EB982 BID103 17.1 SP UNKNOWN UNKNOWN EB99 No BID 18.4 SP UNKNOWN NO 107 EB812 BID099 20.4 SP UNKNOWN UNKNOWN EB814 BID099 20.4 SP UNKNOWN SINGLE EB816 BID099 20.4 SP UNKNOWN UNKNOWN EB818 BID099 20.4 SP UNKNOWN UNKNOWN EB824 BID099 20.4 SP UNKNOWN UNKNOWN EB826 BID099 20.4 SP UNKNOWN UNKNOWN EB828 BID099 20.4 SP UNKNOWN UNKNOWN EB830 BID099 20.4 SP UNKNOWN UNKNOWN EB870 BID100 17 SP UNKNOWN SINGLE 108 Table 2-2. Distributions of the morphology of Cajal Retzius neurons Numbers Percentages Horizontal 159 64.37247 Angled 47 19.02834 Vertical 39 15.78947 No photo 2 0.809717 109 Table 2-3. Top 100 differentially expressed genes between Cajal Retzius and other EB cells Gene Name Fold Change p-value RELN 3.52 4.10E-57 CALB2 7.28 2.80E-32 OLFM1 2.63 7.30E-31 CXCR4 5.2 4.10E-26 ISOC1 6.08 1.70E-23 AMIGO2 6.89 5.20E-22 TAGLN2 6.92 3.40E-19 PCP4 4.27 5.40E-19 ZIC2 6.01 1.40E-18 EBF3 4.86 5.20E-18 CLSTN1 2.71 1.90E-17 ZNF503 5.81 1.10E-16 CBLN1 4.05 1.70E-16 DIABLO/B3GNT4 3.31 4.50E-16 EMX2 6.64 4.60E-16 OPCML 3.18 4.70E-16 DDAH1 2.72 1.40E-15 SEZ6L 2.51 2.60E-15 PTN 3.02 2.80E-15 LHX1 4.37 4.50E-15 DBI 2.62 5.50E-15 TMEM163 4.2 1.70E-14 GRN 7.92 2.60E-14 THSD7A 2.61 2.80E-14 CNTNAP2 2.04 3.10E-14 CLU 2.82 3.50E-14 THSD7B 4.19 5.20E-14 FSTL5 3.49 9.20E-14 CD9 4.09 1.10E-13 GPR153 4.43 4.20E-13 DUSP23 4.56 4.20E-13 IGFBP5 2.77 4.40E-13 NHLH2 3.27 4.80E-13 SPINT2 2.59 5.50E-13 110 CACNA2D3 4.51 5.90E-13 C1QTNF3 6.51 8.10E-13 B3GAT2 3.51 9.00E-13 ADARB1 2.71 1.80E-12 NR2F2 2.63 3.40E-12 RSPO3 3.32 3.90E-12 TRH 4.68 4.10E-12 CTNNB1 2.02 4.30E-12 CCK 5.44 4.40E-12 TXNP6/TXNP4 2.64 5.40E-12 PDE4B 2.47 5.10E-11 TXNP6/TXNP1/TXNP5 2.7 6.40E-11 BCAR3 3.71 1.10E-10 TXNP1 4.33 1.60E-10 TENM1 2.51 2.30E-10 COX5B 2.12 2.40E-10 NTNG2 3.8 2.50E-10 IGSF9 4.54 4.00E-10 SUSD4 3.31 4.30E-10 GADD45A 4.01 4.80E-10 GOT1 2.57 5.90E-10 DBIP1 4.79 6.80E-10 CDK14 2.53 8.40E-10 SIPA1L2 2.63 8.60E-10 ACKR1 5.09 8.70E-10 RRAGB 3.31 9.30E-10 GNAL 2.02 1.10E-09 PAM 2.91 1.20E-09 MAB21L1 5.29 1.20E-09 CAMK2D 2.21 1.80E-09 LAPTM4A 2.25 1.90E-09 PLSCR4 4.03 2.50E-09 HMGCS1 2.07 2.70E-09 RAB6B 2.28 2.80E-09 ELFN2 3.62 3.40E-09 ZCCHC17 2.81 4.10E-09 SYNGR3 2.92 4.40E-09 CD81 2.21 5.20E-09 RAB15 2.96 5.30E-09 SEC13 2.7 7.10E-09 111 LINC01133 5.38 7.40E-09 LOC124685 2.74 7.60E-09 ELAVL2 2.02 7.90E-09 POMGNT2 3.87 1.30E-08 FBLN1 2.34 1.40E-08 NDUFB8P2 2.41 1.50E-08 VAT1 2.37 1.80E-08 NDNF 5.11 1.80E-08 TPGS2/FHOD3 3.5 2.40E-08 SNCA 2.09 2.60E-08 MANEAL 3.6 2.90E-08 PGRMC2 2.24 3.50E-08 KCNC2 2.8 3.70E-08 LY6H 3.45 4.00E-08 PTPRN2 2.44 4.10E-08 PRDX5 2.08 4.20E-08 AKAP1 2.6 4.20E-08 NDUFB8 2.04 4.70E-08 CADM3 2.66 4.80E-08 CNN3 2.17 5.00E-08 NCS1 2.24 5.40E-08 NISCH 2.1 6.70E-08 PITPNA 2.39 6.70E-08 ABLIM3 3.64 7.10E-08 PFKP 2.37 7.30E-08 YWHABP2 2.35 9.10E-08 . 112 Table 2-4. Gene Ontology (GO) analysis GO biological process complete Homo sapiens (REF) uploaded expected Fold Enrichment P value neuron migration 114 10 0.85 11.79 1.63E-04 single-organism cellular process 9830 101 73.12 1.38 4.47E-02 single-organism process 12677 124 94.3 1.31 2.59E-03 generation of neurons 1389 28 10.33 2.71 1.18E-02 neurogenesis 1487 30 11.06 2.71 4.53E-03 nervous system development 2203 44 16.39 2.69 5.91E-06 system development 4157 57 30.92 1.84 7.85E-03 anatomical structure development 5073 64 37.74 1.7 2.21E-02 developmental process 5433 68 40.41 1.68 1.10E-02 single-organism developmental process 5349 68 39.79 1.71 6.02E-03 single-multicellular organism process 5527 69 41.11 1.68 9.42E-03 cell differentiation 3457 50 25.71 1.94 1.11E-02 cellular developmental process 3529 50 26.25 1.9 2.04E-02 regulation of synapse organization 120 9 0.89 10.08 3.12E-03 regulation of synapse structure or activity 124 9 0.92 9.76 4.09E-03 regulation of biological quality 3538 51 26.32 1.94 9.08E-03 113 Chapter 3. Subtypes and characterization of Embryonic Spinal Cord Neurons 3.1 Introduction Human embryonic spinal cord can be divided into distinct anatomical areas, where different population of motor, sensory, and commissural neurons reside. Motor neurons are mostly located in the ventral horn, sensory neurons are located in dorsal root ganglion (DRG), and commissural neurons are located in dorsal horn. During early embryonic development, cells in a neuronal tube undergo dorsal-ventral formation through processes of cell fate specification. These processes are orchestrated by expressing gradient of the Sonic Hedge Hog (Shh) proteins. The notochord, mesodermal originated tissue adjacent to the ventral region of this structure, serves as a source of the Shh expression and also induces development of neural tube cells (T. Yamada, Placzek, Tanaka, Dodd, & Jessell, 1991); The cells lateral to the floorplate becomes ventral horn that motor neurons exist (Toshiya Yamada, Pfaff, Edlund, & Jessell, 1993). Moreover, the ventral-caudal migratory movement of neurons determines cell fate in chick and mouse embryo (Doetsch & Alvarez-Buylla, 1996; Leberl & Sane&, 1995). In human embryo, the formation of dorsal-ventral becomes more pronounced at ~6 gestational week (6 GW). The dorsal area forms alar plate, which will later become interneurons. On the other hand, the ventral area forms basal plate, which will 114 become motor neurons. After formation of dorsal-ventral coordination, the Dorsal Root Ganglion (DRG) is formed on the dorsal side of the spinal cord. These cells become the primary sensory neuron and they will form connections with several adjacent neurons. Though the spinal cord development process was well characterized, expression profile differences between individual neurons from different origins had not yet been studied well. For example, motor neurons in ventral horn can be from neural crest or neural tube (Figure 3-1). Here, in this chapter, subtypes of embryonic spinal cord neurons were characterized by applying the PAIA. 3.2 Methods & Materials 3.2.1 Spinal Cord slice preparation Postmortem fetal spinal cord tissues, in the second trimester of gestation were used in this study. Tissue was donated with informed consent, using a procurement procedure that was approved by institutional review board of the University of Southern California. Specimens were transported to the laboratory in Hypothermosol solution (Sigma, USA; Biolife Solutions, USA) on ice. Tissue transport time was always within two hours. The slice of spinal cord tissue was then visually examined for structural integrity by dorsal-ventral horn formation and the caudal face of the tissue was glued on a specimen plate (horizontal sections). The tissue was then sliced as horizontal section into 400μm slices in the same method as embryonic brain tissue. 115 3.2.2 Electrophysiological recording Three different regions (Ventral horn, Dorsal horn and Dorsal Root Ganglia) of the embryonic spinal cord were visualized by optical microscopy using an infrared Dodt gradient contrast system. The electrophysiological recording was performed using the same protocol as embryonic brain tissue was recorded (see details in section 2.2.2). The cytoplasmic content was flash-frozen in liquid nitrogen before storing in a -80˚C freezer. 3.2.3 Data Analysis of electrophysiological recording Electrophysiological recordings were processed and examined using customized macros in Igor Pro (Wavemetrics, USA) and Matlab (Mathworks, USA). 3.2.4 Single Cell RNAseq Cytoplasmic content from individual cells was amplified using the in- house developed method, PAIA (Patch-aRNA in vitro transcription amplification), introduced in Chapter 1. The aRNA method was composed of 2 cycles of four procedures: first strand cDNA synthesis, second strand cDNA synthesis, DNA purification, in vitro transcription (IVT) and RNA purification. The first strand cDNA was synthesized with Superscript III (Life Technology) and Oligo-dT T7 primers (Custom, Life Technology) followed by second strand synthesis using DNA polymerase I (Life Technology, 18010025) and RNase H (Life Technology, 18021071). The DNA was then purified using the RNAClean 116 XP beads (Beckman, CA, USA) and used as a template for in vitro transcription (IVT). IVT was performed with T7 RNA polymerase and Megascript T7 kit (Life Technology, AM1334). We reduced the reaction volume to half for IVT, replaced enzymes to thermostable ones and typically performed only two rounds of IVT. With this protocol, 100~500ng of amplified RNA was reliably produced from single cell samples. Libraries were then generated from 150-400ng of amplified RNA using TruSeq Stranded mRNA Sample Preparation Kit (Illumina, CA, USA). Briefly, single strand cDNA was synthesized using the SuperScript II (Life Technology) and random primers. The cDNA was converted into double stranded cDNA, blunt-ended and adenylated at 3’-end. The resulting library dsDNA fragments were ligated to adapters and enriched using Illumina PCR primers followed by purification with AMPure XP beads (Beckman, CA, USA). DNA fragment size and library concentration were examined using 2200 TapeStation (Agilent Technologies) before sequencing with Illumina hiseq 2500 rapid mode. 3.2.5 Data Analysis of single cell RNAseq Prior to mapping, reads containing more than 50% of adapter sequences, monomers or other low entropy reads (metric entropy below 1%) were removed. The rest of reads from each individual channel were trimmed (if adapters constitute less than 50% of the read) and sequentially aligned to rRNA, mtDNA, the rest of human transcriptome (GENCODE v22; 60,483 gene models) and genome (GRCh38) using the in-house custom RNA-Seq 117 alignment pipeline, GT-FAR (https://genomics.isi.edu/gtfar). Reads mapped to rRNA and mtDNA were excluded from following analysis. Genes with ≤10 reads in ≤10% of cells were discarded. To adjust differences in library complexity, the data from each cell was downsampled, without replacement, to 100,000 mapped exonic reads and log-normalized. These normalized data were processed with regular DESeq2 analysis. The list of differentially expressed genes was analyzed with Gene Ontology (http://www.geneontology.org/) and Gene set enrichment analysis (GSEA) (http://software.broadinstitute.org/gsea/index.jsp) for functional characterization. 3.3 Result 3.3.1 Region specific gene expression differences To characterize individual neuron in the embryonic brain, cells from three different regions of embryonic brain (Ventral horn, Dorsal horn, DRG and intermediate area) were collected with their morphological information (Figure 3-2). With electrophysiological pattern, three potential subtypes of motor neurons from ventral horn of embryonic spinal cord in 10~16 pcw (post conceptual week) were found. In this stage, most of the neuronal nuclei and ganglions are formed and the dorsal-ventral separation has become very clearly distinguishable. Using the Single Cell RNAseq method, PAIA in which developed in first chapter, in total, 324 embryonic spinal cord cells were processed (Figure 3-3). 118 The morphology of each neuron was examined and cell location was detected by the microscopy. Some of the motor neurons were injected with Lucifer Yellow (LY) to confirm the shape and size of the neurons (Figure 3-4). Among 324 cells processed, 251 cells were motor neurons from ventral horn. 35 cells were interneurons from dorsal horn, and 11 cells were interneurons from the middle of the spinal cord. 7 DRG and 1 Rwnshaw neurons were also processed. 17 samples were either missing photos or difficult to identify the cell-type, so these samples were labeled “Unknown” and not included in the further analysis. With the single cell RNAseq data, differential expression analysis (DEseq) was performed using motor neurons from ventral horn (n=252) with sensory neurons from DRG (n=7) and interneurons from dorsal horn (n=36) and middle of the spinal cord (n=11). Because of the sparsity of the single cell RNAseq data, all the samples were downsampled into 100,000 reads. Between dorsal horn interneurons and the ventral horn motor neurons, 600 differentially expressed genes were found (adjusted p<0.05, Table 3-1). 351 genes were upregulated in motor neurons and 249 were upregulated in dorsal horn interneurons. The most differentially expressed genes by p value were Vacuolar-Sorting Protein SNF8 (p=9.74E-13) and RSRC2, Arginine/Serine-Rich Coiled-Coil 2 (p=2.48E-11). Both genes were upregulated in dorsal horn interneurons (8 fold and 7 fold). Gene Set Enrichment Analysis (GSEA) analysis found that these differentially enriched 119 genes were involved in the pathways such as cellular transportation (FDR q=4.31E-12) and RNA binding (FDR q=7.63E-12) (Table 3-2). Between sensory neurons from dorsal root ganglia (DRG) and the ventral horn motor neurons, 522 differentially expressed genes were found (adjusted p<0.05, Table 3-3). In sensory neurons, 80 genes were upregulated and in motor neurons 442 genes were upregulated. The most differentially expressed genes by p value were SMIM14, Small Integral Membrane Protein 14 (p=4.90E-05) and NNAT, Neuronatin (p=4.90E-05). SWIM14 (fold change>4) was overexpressed in sensory neurons and NNAT (fold change>5) was overexpressed in motor neurons. In GSEA analysis, Alzheimer disease related pathway (FDR q= 2.28E-104) and biopolar disorder pathway (FDR q= 1.93E-100) (Table 3-4) were found as the most relevant pathway that these differentially genes were involved in. The sensory neurons from DRG and the dorsal horn interneurons were analyzed with the differential expression analysis. 402 genes were found to be differentially expressed (adjusted p<0.05, Table 3-5). 70 genes were upregulated in sensory neurons and 332 genes were upregulated in dorsal horn interneurons. By p-value, the most differentially expressed genes were MARCKSL1 (Macrophage Myristoylated Alanine-Rich C Kinase Substrate 1) (p=5.41E-06) and OCIAD1 (OCIA Domain Containing 1) (p=7.16E-06). Both genes were overexpressed in dorsal horn interneurons (fold change >8, >6). In GSEA analysis, similar to previous comparison, Alzheimer disease related 120 pathway (FDR q= 7.76E-77) and bipolar disorder pathway (FDR q=3.52E-71) were found as the most relevant pathway by q value and cellular transportation pathway (FDR=1.1E-32) also showed significance (Table 3-6). The expression profiles of differentially expressed genes of the largest groups (Motor neuron and Dorsal Horn Interneurons) were plotted into a heatmap (Figure 3-5). 3.3.2 Interneuron Subtypes The interneurons were found from dorsal horn (commissural neurons) and middle of the spinal cord. They are known to have different migratory origins (Junge, Yung, Goodrich, & Chen, 2016). Therefore, the differential expression analysis between dorsal horn interneurons and intermediate zone interneuron could identify key genes, which could describe their origin. With this analysis, 137 genes were differentially expressed (adjusted p<0.05, Table 3-7). In intermediate interneurons, 64 genes were upregulated and 73 genes were downregulated, compared to dorsal horn interneurons. By p-value, the most differentially expressed genes were HSH2D (p= 2.31E-5) and RP11- 252A24.2 (p= 2.61E-05). Both genes were overexpressed in intermediate interneurons than dorsal horn interneurons (fold change >9, >5). With GSEA analysis, differentially expressed genes were enriched in signaling pathway regulation (FDR q=1.49E-4) and transportation pathways (FDR q= 3.99E-4) (Table 3-8). 121 3.3.3 Motor Neuron Subtypes with Firing Pattern From 252 motor neurons collected, the firing patterns of 69 cells were measured by electrophysiological recording. The firing pattern with current injection was measured from these cells. First, with current clamp recording by current injection, the changes in action potential firing pattern were observed. The frequency of action potential changes was measured with increment of current injection. Out of these 69 cells, three different types of action potential firing patterns were observed: single, repetitive non-adaptive, repetitive adaptive (Figure 3-6). 23 cells showed a single action potential firing similar to embryonic brain cells (Single AP). 12 cells increased firing frequency as the amount of injected current increases, and reached to a plateau (Repetitive Non-adaptive). The most cells (n=29), increased action potential frequency as the current (~100pA) was injected, and instead of reaching plateau, the frequency decreased slowly after the stimulus (Repetitive Adaptive). By categorizing neurons based on these three different electrophysiological patterns, differential gene expression analysis was performed using DESeq2. 130 genes were differentially expressed between Repetitive adaptive and non-adaptive motor neurons (Table 3-9). 70 genes were overexpressed in repetitive adaptive neurons and 61 genes were overexpressed in repetitive non-adaptive neurons. The examples of most differentially expressed genes by adjusted p value were MANSC1 (p=6.5E-7), a DNA binding protein and AP1G1 (p=4.7E-5), a Clathrin assembly Protein 122 Complex 1 Gamma-1 Large Chain. The MANSC1 gene was highly expressed in repetitive non-adaptive neurons compared to repetitive adaptive neurons (log fold change>7). The AP1G1 was highly expressed in repetitive adaptive neurons (log fold change>4.7). This gene is also known as a regulating component of clathrin complex for the transportation coated vesicles from the trans-Golgi network (Johnson, Gagnon, & Chang, 2016). The differential gene expression analysis between Repetitive adaptive and Single action potential firing revealed 125 genes (Table 3-10). The Gene Ontology pathway analysis showed that the genes overexpressed in cells with single action potential firing were involved in the ubiquitin-dependent protein catabolic process pathway. GSEA also found that these genes were involved in macromolecule catabolic pathway (FDR=7.65E-8). The most differentially expressed genes by p value were SORL1 (p=8.67E-10) and ARIH1 (p=2.24E- 6). SORL1 (Sortilin Related Receptor 1) was highly expressed in neurons with single action potential (fold change>7). This gene was known to be related to LDL receptor. ARIH1 is highly differentially expressed in repetitive adaptive neurons (fold change>4) and known to act as a ubiquitin protein ligase. When we compared gene expression profiles between repetitive non- adaptive neurons and single action potential, 75 genes were differentially expressed (Table 3-11). In GSEA analysis, genes overexpressed in cells with single action potential were enriched in the gene set, which was reported to be downregulated in brain from patients with Alzheimer's disease (FDR= 1.72E-3) 123 and another gene set, related to cellular development (FDR q=2.82E-3). Some of these overexpressed genes were also found from the previous lists of the differentially expressed genes (Table 3-11). For example, MANSC1 (p=6.89E- 7), USP34 (p=7.68E-6) and SORL1 (p=1.63E-5) were the most differentially expressed genes between repetitive non-adaptive neurons and single action potential neurons. The MANSC1 was overexpressed in non-adaptive neurons, but SORL1 was overexpressed in single action potential neurons. The USP34 was a protein-coding gene for Ubiquitin Specific Peptidase 34, which was also related to the catabolic function as described previously. By combining gene lists expressed in each cell type with an action potential firing pattern, the expression pattern of differentially expressed gene (fold change>5, p<0.05) was plotted as a heatmap (Figure 3-7). Moreover, the potential biomarkers were defined based on each firing pattern. To identify potential biomarkers,, genes that are highly expressed in each comparison (e.g. repetitive adaptive vs repetitive non-adaptive) were first selected. Genes overexpressed in a group from all comparisons were defined as the potential biomarkers linked to the specific firing pattern (Table 3-12). 124 3.4 Discussion 3.4.1 Sensory and Motor Neuron Markers The dorsal root ganglion neurons were originated from neural crest whereas most of the motor neurons and interneurons in spinal cord were originated from neural tube (Gans & Northcutt, 1983; Moreno & Bronner- Fraser, 2005; Wakamatsu, Maynard, & Weston, 2000). By measuring gene expression differences between these neurons, it is possible to study how the process of neurodevelopment occurs. For example, the Neuronatin (NNAT), which was overexpressed in motor neuron compared to the DRG neurons in spinal cord (fold change>5, p=4.90E-5), is a gene related to the segment formation (Wijnholds, Chowdhury, Wehr, & Gruss, 1995). This supports that the segmentation is limited to neural tube, which is made from motor neurons. Recent study reported that the expression of RPLP2 (Ribosomal Protein Lateral Stalk Subunit P2), which was found to be differentially expressed between sensory and motor neurons (fold change>7, p=4.90E-5) in our study, has relation to Tar DNA binding protein (TDP43) deposition in Amyotrophic Lateral Sclerosis (Koyama et al., 2016). It is consistent with previous finding that the TDP43 deposits the inclusion bodies only in motor neuron but not in the other areas (Bodansky et al., 2010). In dorsal horn interneurons, BEX1 was one of highly overexpressed genes compared to DRG (p=1.03E-4, fold change>5). BEX2 was also one of 125 the overexpressed genes in dorsal horn interneurons, compared to motor neurons (p=3.7E-3, fold change>2). These genes are related to the cell cycle, p75 pathway and nerve growing (Vilar et al., 2006; Yoon et al., 1998). Unlike motor and sensory neurons, dorsal horn interneurons can be re-differentiated during spinal cord injury and allow other neurons to regrow. The BEX genes are one of factors regulating this process (Khazaei et al., 2010). 3.4.2 Interneuron subtypes Similar to embryonic brain neuronal migration, the migration of inter neurons from ventral zone (VZ) in embryonic spinal cord was previously studied, and several signaling molecules such as Slit/Robo, Netrin-1, and Netrin-2 were discovered to be involved in this process (Caspary & Anderson, 2003; Junge et al., 2016; Kennedy, Serafini, de la Torre, & Tessier-Lavigne, 1994; M. Kim et al., 2015; Leberl & Sane&, 1995). The gene expression profiles of interneurons located in different areas allowed us to characterize the origins of interneurons migrated. Interestingly, SEC61A2 gene was overexpressed in interneurons from intermediate area compared to the dorsal horn interneurons (p=2.61E-5, fold change>7). This gene is one of the sigma-1 receptor modulators and it is reported as a cellular pluripotency modulator (Su, Su, Nakamura, & Tsai, 2016). This suggests that the change in differentiation stage of the neurons may be associated with interneuron types in embryonic spinal cord. 126 3.4.3 Motor Neuron: Electrophysiological profile and subtypes . The electrophysiological patterns of motor neuron in animal models affect the patterns of muscle contraction (Brezina, Orekhova, & Weiss, 2000). Migratory origin (Leberl & Sane&, 1995) and different stages of development including synaptic pruning (Ren & Greer, 2003) may explain why motor neurons have distinct electrophysiological profiles. The single action potential firing pattern found in motor neurons was similar to the firing pattern observed in most of cortical plate (CP) neurons in Chapter 2. It was the only pattern observed in the newborn neurons. However, on Cajal-Retzius neurons, which were relatively old neurons described from the previous chapter, were mostly expressing repetitive adaptive and non- adaptive firing patterns. It is consistent with the finding from GSEA results: differentially expressed genes between the repetitive non-adaptive neurons and the single action potential firing neurons were found to be involved with Alzheimer related pathway and cellular development pathway. One of the potential biomarkers for single action potential firing motor neurons was SORL1 (Sortilin Related Receptor 1). This gene is reported as a key differentially expressed gene linked to human Alzheimer disease (Young et al., 2015). Even though the Alzheimer has an old age onset, some of recent studies indicated that Alzheimer could be a developmental disorder (Arendt, Stieler, & Ueberham, 2017). 127 The repetitive adaptive firing pattern neuron’s potential biomarkers were involved in the macromolecule catabolic pathway. One of the biomarkers, AP1G1 is related to Immune System, Vesicle-mediated transport and Hypoactive Sexual Desire Disorder (Johnson et al., 2016). The finding of catabolic pathway may support that some of these neurons could go through synaptic pruning. Unlike other two firing neurons, most neurons have the repetitive non- adaptive pattern. The potential biomarkers for these neurons include genes involved in neuronal functions such as actin filament (TNS2, WHAMML1, DCBLD2, CSNK1G3) and adhesion molecules (CD226, NCKAP1, DOCK5). This finding supports that repetitive non-adaptive neurons could be less developed neurons possibly with more neuronal functions. 128 3.5 Figures Figure 3-1 Embryonic Spinal Cord development Until Stage c, brain and spinal cord development is similar. In stage d, the commissural neurons (C) and motor neuron (M) projections are formed and staring migration of DRG. Figure is adapted from the image from Jessell et al (Jessell, 2000). 129 Figure 3-2. Morphologies of spinal cord neurons (A) Interneuron from dorsal horn of embryonic spinal cord (ES) (B) Motor neuron from ventral horn of ES (C) Sensory Neuron from dorsal root ganglia (DRG) (D) Interneuron from intermediate area 130 Figure 3-3. Distribution of collected samples from Embryonic Spinal Cord (ES) 131 Figure 3-4. Motor Neuron with Lucifer Yellow Staining Microscope image of (A) Before and (B) After injection of Lucifer yellow dye. 132 Figure 3-5. Volcano plot of Differentially expressed genes In each volcano plot in A VS B format, genes highly expressed in A are on the right side and highly expressed in B are on the left side. 133 Figure 3-6. Three Firing patterns of motor neurons. Within Motor neurons collected, there were 3 types of pattern in action potential (AP) firing measured by electrophysiological recording. (A) Repetitive Adaptive: increased frequency of AP until 100pA injection and decrease afterward. (B) Repetitive Non-adaptive: ncreased frequency of AP and reached the plateau as more current was injected. (C) Single: No matter how much current was injected, there was only single Action potential. 134 Figure 3-7. Heatmaps of some differential expressed gene in repetitive adaptive, non-adaptive and single action potential firing neurons (A) The left set shows the genes overexpressed in Repetitive adaptive (fold change>5), middle set shows the genes overexpressed in Repetitive non- adaptive neurons (fold change>5), right set shows genes overexpressed in single action potential firing neurons (fold change>5). (B) expression levels of example potential biomarkers linked to each firing pattern. 135 3.6 Tables Table 3-1. Top 30 Genes differentially expressed between Motor Neuron and Dorsal Horn. Gene baseMean log2FoldChange pvalue Padj SNF8 15.58333 -3.96735 5.87E-17 9.74E-13 RSRC2 65.84568 -4.63924 2.99E-15 2.48E-11 MDM4 28.58333 -3.48988 6.78E-15 3.75E-11 FAM101B 9.42284 -4.8153 4.04E-13 1.34E-09 ZFYVE26 11.56481 -4.3509 3.30E-13 1.34E-09 PKHD1 18.39506 -6.36103 1.48E-12 3.70E-09 UBXN2A 18.77469 -4.26238 1.56E-12 3.70E-09 PTOV1-AS1 8.104938 -3.99631 2.55E-12 5.28E-09 NCBP1 9.935185 -4.05094 3.00E-12 5.52E-09 GADD45A 41.16667 -6.70861 7.56E-12 1.25E-08 OTUB1 14.05556 -3.77375 1.30E-11 1.97E-08 MSMO1 20.35802 -3.23027 1.98E-11 2.74E-08 SYT17 28.53086 -4.36815 3.20E-11 4.09E-08 TPST1 40.41358 -6.67737 6.05E-11 7.16E-08 LRRC49 12.84877 -3.47967 1.41E-10 1.56E-07 PAFAH2 17 -6.23251 1.80E-10 1.87E-07 GPR137C 189.5926 5.375293 2.40E-10 2.34E-07 RRP15 13.65432 -3.08678 3.74E-10 3.44E-07 OCIAD1 40.75926 -3.26534 5.20E-10 4.54E-07 SNX18 8.009259 -5.54746 6.22E-10 5.16E-07 ETNK1 26.55556 -4.0361 6.75E-10 5.33E-07 USP24 26.98457 -4.38344 7.17E-10 5.40E-07 RP11-299H22.1 4.719136 -4.05718 3.85E-09 2.77E-06 CTD-2012J19.3 17.94136 -6.71789 5.01E-09 3.46E-06 AP006285.2 3.62037 -5.32716 6.48E-09 4.30E-06 RNF144A 12.04321 -3.93554 1.42E-08 9.04E-06 BTAF1 11.11111 -4.17046 2.12E-08 1.30E-05 ARHGAP21 26.90432 -3.01684 2.41E-08 1.43E-05 136 Table 3-2. Gene Set Enrichment Analysis of genes differentially expressed between motor and dorsal horn interneurons. Gene Set Name [# Genes (K)] # Genes in Overlap (k) p-value FDR q- value GO_POSITIVE_REGULATION_OF_MOLECULAR_FUNCTION [1791] 66 1.55 e-19 2.76 e-15 PUJANA_BRCA1_PCC_NETWORK [1652] 61 3.89 e-18 3.46 e-14 GO_POSITIVE_REGULATION_OF_CATALYTIC_ACTIVITY [1518] 57 2.59 e-17 1.54 e-13 PILON_KLF1_TARGETS_DN [1972] 65 7.51 e-17 3.02 e-13 GO_PHOSPHATE_CONTAINING_COMPOUND_METABTABOLIC_PROCESS [1977] 65 8.49 e-17 3.02 e-13 GO_RIBONUCLEOTIDE_BINDING [1860] 61 9.42 e-16 2.79 e-12 GO_ADENYL_NUCLEOTIDE_BINDING [1514] 54 1.63 e-15 4.15 e-12 GO_ESTABLISHMENT_OF_LOCALIZATION_IN_CELL [1676] 57 1.94 e-15 4.31 e-12 GO_RNA_BINDING [1598] 55 3.86 e-15 7.63 e-12 GO_REGULATION_OF_PROTEIN_MODIFICATIONON_PROCESS [1710] 57 4.57 e-15 8.12 e-12 137 Table 3-3. Top 30 Genes differentially expressed between Motor neuron and sensory neuron from dorsal root ganglia (DRG). Gene baseMean log2FoldChange pvalue padj SMIM14 28.58642 4.482584 2.68E-08 4.90E-05 RPLP2 46.83025 -5.85821 2.29E-08 4.90E-05 NNAT 53.53086 -6.40935 3.06E-08 4.90E-05 MARCKSL1 44.89198 -7.07258 2.46E-08 4.90E-05 ZNF69 4.537037 7.704651 4.10E-08 5.26E-05 TUBB2B 164.5309 -4.5627 1.51E-07 0.000138 TUBB2A 127.5401 -4.89613 1.43E-07 0.000138 LDHB 47.43827 -5.89238 1.83E-07 0.000146 YWHAB 106.8302 -4.78859 2.38E-07 0.00017 LINC00840 5.978395 9.025331 2.99E-07 0.000192 ACTG1 230.8951 -3.9394 3.86E-07 0.000225 MCL1 28.48765 6.32474 4.57E-07 0.000229 ACTB 97.56481 -4.07251 4.64E-07 0.000229 KCNE4 17.28086 4.499058 5.94E-07 0.000256 BASP1 92.49074 -5.08112 5.98E-07 0.000256 UCHL1 61.70679 -4.66536 8.29E-07 0.00028 YWHAQ 123.0247 -4.79458 8.25E-07 0.00028 TUBB 73.80247 -5.07055 7.84E-07 0.00028 GDI1 23.51852 -6.66079 8.17E-07 0.00028 BEX1 85.0463 -5.02035 8.78E-07 0.000282 INA 56.9321 -5.28581 1.17E-06 0.000356 FXR1 18.8179 5.151206 1.27E-06 0.000371 ATP5EP2 22.07099 -6.20484 1.69E-06 0.00047 RGS4 24.44753 6.458116 1.85E-06 0.000493 ATP6V1A 38.75926 -4.7984 2.26E-06 0.000556 RPS23P8 13.64506 -6.38061 2.19E-06 0.000556 GHITM 30.66049 -5.85658 3.27E-06 0.000777 RPS11 17.40432 -6.34431 3.46E-06 0.000793 AZIN1 79.83333 -5.73922 3.81E-06 0.000814 CKB 16.39506 -6.2054 3.76E-06 0.000814 138 Table 3-4. GSEA of the differential expressed genes between motor neurons from ventral horn and sensory neurons from DRG. Gene Set Name [# Genes (K)] # Genes in Overlap (k) p-value FDR q-value BLALOCK_ALZHEIMERS_DISEASE_DN [1237] 144 1.28 e-108 2.28 e-104 KIM_BIPOLAR_DISORDER_OLIGODENDROCYTE_DENSITY_CORR_UP [682] 115 2.17 e-104 1.93 e-100 KIM_ALL_DISORDERS_OLIGODENDROCYTE_NUMBER_CORR_UP [756] 113 2.36 e-96 1.4 e-92 KIM_ALL_DISORDERS_CALB1_CORR_UP [548] 90 6.16 e-80 2.74 e-76 HSIAO_HOUSEKEEPING_GENES [389] 72 1.27 e-67 4.51 e-64 MODULE_83 [320] 62 1.7 e-59 5.04 e-56 MODULE_114 [338] 63 2.47 e-59 6.29 e-56 MODULE_151 [318] 58 4.7 e-54 1.05 e-50 GO_RNA_BINDING [1598] 106 7.23 e-54 1.43 e-50 MORF_RAN [268] 54 7.96 e-53 1.42 e-49 139 Table 3-5. Top 30 genes differentially expressed between DRG sensory neurons and DH interneurons. Gene baseMean log2FoldChange pvalue padj MARCKSL1 44.89198 -8.18674 9.72E-10 5.41E-06 OCIAD1 40.75926 -6.86647 2.98E-09 7.16E-06 ATP5EP2 22.07099 -8.00762 3.86E-09 7.16E-06 LINC00840 5.978395 10.69359 4.34E-08 6.04E-05 BEX1 85.0463 -5.88049 9.30E-08 0.000103 PCMTD1 29.72531 6.259391 1.71E-07 0.000159 RSRC2 65.84568 -6.30119 3.12E-07 0.000248 DNAJA1 53.58025 -5.6485 5.11E-07 0.000294 MSMO1 20.35802 -5.38093 5.13E-07 0.000294 CPNE3 12.43827 6.764333 5.69E-07 0.000294 B3GALT5 7.527778 8.446982 5.81E-07 0.000294 SLC25A5 80.80864 5.854678 7.75E-07 0.00035 NNAT 53.53086 -6.08082 8.18E-07 0.00035 TUBB2A 127.5401 -4.90235 1.06E-06 0.000423 MKRN1 37.52778 -5.5635 1.59E-06 0.000548 RPLP2 46.83025 -5.35216 1.67E-06 0.000548 SNN 28.58642 -6.26671 1.67E-06 0.000548 PEG10 33.59568 -5.92667 2.12E-06 0.000657 ARHGAP21 26.90432 -5.47684 2.70E-06 0.000753 INA 56.9321 -5.48898 2.71E-06 0.000753 ACTB 97.56481 -4.07344 2.96E-06 0.000784 ZNF69 4.537037 7.101289 4.27E-06 0.001081 BEX2 36.01235 -5.74515 4.73E-06 0.001144 BASP1 92.49074 -5.01155 5.04E-06 0.001169 NME7 10.06481 4.99378 5.83E-06 0.001297 OTUB1 14.05556 -5.38314 7.77E-06 0.001602 STMN2 350.1204 -3.8839 7.77E-06 0.001602 LINC01503 6.462963 6.709956 8.78E-06 0.001707 MYEF2 15.2963 4.38055 8.90E-06 0.001707 UCHL1 61.70679 -4.49521 1.03E-05 0.001908 140 Table 3-6. GSEA on the differentially expressed gene between DRG sensory neurons and DH interneurons. Gene Set Name [# Genes (K)] # Genes in Overlap (k) p-value FDR q- value BLALOCK_ALZHEIMERS_DISEASE_DN [1237] 109 4.36 e-81 7.76 e-77 KIM_BIPOLAR_DISORDER_OLIGODENDROCYTE_DENSITY_CORR_UP [682] 85 3.96 e-75 3.52 e-71 KIM_ALL_DISORDERS_CALB1_CORR_UP [548] 72 5.16 e-65 3.06 e-61 KIM_ALL_DISORDERS_OLIGODENDROCYTE_NUMBUMBER_CORR_UP [756] 72 6.24 e-55 2.77 e-51 HSIAO_HOUSEKEEPING_GENES [389] 53 1.44 e-48 5.13 e-45 GO_RNA_BINDING [1598] 78 2.44 e-38 7.22 e-35 GO_POLY_A_RNA_BINDING [1170] 66 4.78 e-36 1.1 e-32 GO_ESTABLISHMENT_OF_LOCALIZATION_IN_CELL [1676] 77 4.93 e-36 1.1 e-32 PUJANA_BRCA1_PCC_NETWORK [1652] 76 1.37 e-35 2.7 e-32 MODULE_83 [320] 40 3.03 e-35 5.39 e-32 141 Table 3-7. Top 30 genes differentially expressed between Intermediate zone interneurons and dorsal horn interneurons. Genes baseMean log2FoldChange pvalue padj HSH2D 8.598765 9.279784 1.65E-09 2.31E-05 RP11- 252A24.2 8.132716 5.516325 4.95E-09 2.61E-05 SEC61A2 76.83333 7.232301 5.58E-09 2.61E-05 RPL38 56.12654 4.958558 1.64E-08 5.74E-05 LIPH 43.81173 5.782686 8.82E-08 0.000247 AC012358.7 1.771605 6.667214 1.19E-07 0.000278 DGKH 40.02469 3.722986 1.80E-07 0.000361 ACAT2 5.858025 -8.47164 3.29E-07 0.000577 RP11- 416H1.2 9.614198 4.995903 3.99E-07 0.000621 SYNE3 6.324074 5.019338 4.52E-07 0.000626 MAPK1 37.66667 4.947924 4.91E-07 0.000626 SGTB 11.42284 4.258404 6.76E-07 0.00079 COL6A3 13.23765 4.930762 8.64E-07 0.000932 COA3 20.50617 -6.69432 1.08E-06 0.001082 RP11- 848P1.2 5 4.778854 1.82E-06 0.001701 OCIAD1 40.75926 -4.33604 2.35E-06 0.002061 ETNK1 26.55556 -5.28753 2.56E-06 0.002062 RPL27A 37.98148 3.66542 2.65E-06 0.002062 C19orf66 17.60185 3.631289 3.34E-06 0.002464 MDM4 28.58333 -3.66989 3.53E-06 0.002476 IAPP 5.04321 4.895722 4.77E-06 0.003182 C17orf77 39.0679 8.139219 6.17E-06 0.003717 PKHD1 18.39506 -6.63878 6.52E-06 0.003717 PABPC1L 3.45679 5.026498 6.58E-06 0.003717 RSRC2 65.84568 -4.56904 6.63E-06 0.003717 FAM101B 9.42284 -5.10988 8.70E-06 0.004691 AP000347.2 3.37037 -6.40896 9.28E-06 0.004818 EIF2B3 6.141975 -6.78689 9.68E-06 0.004849 JPX 6.524691 3.584338 1.19E-05 0.00573 METTL25 3.981481 -7.35265 1.28E-05 0.005895 142 Table 3-8, GSEA of the differentially expressed genes between interneurons from intermediate area and the dorsal horn. Gene Set Name [# Genes (K)] # Genes in Overlap (k) p-value FDR q- value PILON_KLF1_TARGETS_DN [1972] 23 7.51 e- 11 1.34 e-6 YRTCANNRCGC_UNKNOWN [2940] 25 6.71 e-9 5.96 e-5 PUJANA_BRCA1_PCC_NETWORK [1652] 18 3.23 e-8 1.49 e-4 GO_REGULATION_OF_INTRACELLULAR_SIGNAL_TRANSDUCTION [1656] 18 3.35 e-8 1.49 e-4 GO_ORGANOPHOSPHATE_BIOSYNTHETIC_PROCESS [450] 10 8.43 e-8 2.37 e-4 TTGCWCAAY_CEBPB_02 [1972] 19 8.81 e-8 2.37 e-4 GO_CATABOLIC_PROCESS [1773] 18 9.32 e-8 2.37 e-4 PUJANA_ATM_PCC_NETWORK [1442] 16 1.61 e-7 3.36 e-4 JOHNSTONE_PARVB_TARGETS_3_DN [918] 13 1.7 e-7 3.36 e-4 GO_ESTABLISHMENT_OF_LOCALIZATION_IN_CELL [1676] 17 2.25 e-7 3.99 e-4 143 Table 3-9. Repetitive adaptive VS repetitive non adaptive action potential firing patterned motor neurons. Genes baseMean log2FoldChange pvalue padj MANSC1 32.98406 -7.49857 6.34E-11 6.58E-07 AP1G1 31.47012 5.316074 9.14E-09 4.74E-05 PSPH 7.844622 -6.38006 4.57E-08 0.000119 ARIH1 65.21912 4.709931 4.26E-08 0.000119 HTN3 8.904382 -5.87156 1.31E-07 0.000227 GPR75-ASB3 22.53386 -4.29028 1.25E-07 0.000227 MTX2 37.15538 -6.32705 8.47E-07 0.000879 HACE1 10.55777 -5.58917 8.46E-07 0.000879 KDM2A 10.15936 -5.11611 5.95E-07 0.000879 UBE3B 11.56972 6.284258 6.99E-07 0.000879 DNAJC15 7.581673 6.170005 1.41E-06 0.001333 RP11-250B2.5 5.035857 -8.26686 1.96E-06 0.001501 LSM3 30.83665 -5.26988 1.95E-06 0.001501 C11orf88 16.3506 6.033402 2.03E-06 0.001501 EXOC1 18.87251 -5.52131 3.17E-06 0.001957 RNF41 19.59363 -4.57174 3.19E-06 0.001957 SLC12A2 10.94422 5.439511 3.21E-06 0.001957 CD226 20.60558 -4.71607 4.26E-06 0.002453 MYT1 14.80876 6.095384 4.54E-06 0.002477 USP48 11.52988 -4.72188 4.98E-06 0.002572 USP24 31.18327 5.126584 5.34E-06 0.002572 ZADH2 38.75697 5.659401 5.64E-06 0.002572 EPHA3 10.19124 6.378317 5.70E-06 0.002572 CCDC13-AS1 7.390438 -7.53586 6.59E-06 0.002849 SYT11 23.77291 -4.10183 7.13E-06 0.002959 VKORC1 9.944223 6.431295 7.98E-06 0.003183 SMIM19 5.207171 6.332527 1.04E-05 0.003993 LINC00937 9.864542 -4.75437 1.18E-05 0.004315 LRRC28 9.091633 4.764551 1.21E-05 0.004315 EIF2S3 11.13944 -4.87029 1.30E-05 0.004487 Top 30 genes differentially expressed between motor neurons from the comparison. 144 Table 3-10. Repetitivs adaptive VS single action potential patterned motor neurons. Genes baseMean log2FoldChange pvalue padj SORL1 38.14343 -5.64403 6.36E-14 8.67E-10 ARIH1 65.21912 4.506947 3.28E-10 2.24E-06 CYB561 5.087649 -6.14422 1.55E-09 5.28E-06 C11orf88 16.3506 6.722545 1.38E-09 5.28E-06 SYT17 35.59363 -5.60983 7.22E-09 1.97E-05 AC022819.3 20.10757 -4.79123 1.96E-07 0.000333 EIF3K 39.88446 -4.64184 1.95E-07 0.000333 PSPN 5.454183 4.771981 2.20E-07 0.000333 RP11-179K3.2 29.64143 8.009249 1.65E-07 0.000333 FAM168B 36.88446 -4.85118 3.42E-07 0.000446 UBXN2A 21.65737 -4.51978 3.80E-07 0.000446 UBE3B 11.56972 5.456193 4.09E-07 0.000446 PFKFB3 30.91633 5.483443 4.25E-07 0.000446 USP24 31.18327 4.806099 5.56E-07 0.000542 FAM101B 11.15139 -4.72058 1.23E-06 0.001051 MID1IP1 6.434263 4.992873 1.18E-06 0.001051 HSPG2 17.23108 6.165726 1.43E-06 0.001144 RPL13A 64.38247 3.295339 1.99E-06 0.001507 ACAT2 6.784861 -6.04732 2.18E-06 0.001509 DDX19A 7.239044 -3.86206 2.21E-06 0.001509 ANKRD36B 8.111554 4.35057 2.65E-06 0.001719 MGST3 19.21912 -3.13857 4.22E-06 0.002617 POMGNT1 7.824701 -3.94309 4.83E-06 0.002865 DFFA 8.055777 3.29149 6.74E-06 0.003829 C2 2.079681 -5.20093 9.16E-06 0.004194 POLR2M 19.88446 -4.6692 8.28E-06 0.004194 DDB1 31.25498 -3.77824 9.53E-06 0.004194 VDAC3 26.49402 -3.75942 8.20E-06 0.004194 ZNF461 3.458167 3.512021 9.38E-06 0.004194 SLC4A1 4.988048 4.274008 8.82E-06 0.004194 145 Table 3-11. Repetitive non adaptive VS single action potential patterned motor neurons. Gene baseMean log2FoldChange pvalue padj MANSC1 32.98406 7.74083 7.20E-11 6.89E-07 USP34 16.76096 -7.76252 1.61E-09 7.68E-06 SORL1 38.14343 -5.4385 5.12E-09 1.63E-05 DOCK5 6.645418 5.839023 3.44E-08 5.48E-05 CD226 20.60558 5.950336 2.37E-08 5.48E-05 UBXN2A 21.65737 -5.73211 1.64E-07 0.000224 FAM101B 11.15139 -6.11691 3.23E-07 0.000386 CTB-129P6.7 9.820717 5.181985 8.60E-07 0.000914 NCKAP1 21.96414 3.698649 1.72E-06 0.001553 CCDC13-AS1 7.390438 8.189785 1.79E-06 0.001553 POLR2M 19.88446 -5.90993 2.81E-06 0.002241 GPR75-ASB3 22.53386 3.856017 4.37E-06 0.002544 DHTKD1 7.167331 5.011426 4.52E-06 0.002544 TNS2 6.996016 5.264421 3.53E-06 0.002544 DCBLD2 25.15538 5.347619 4.12E-06 0.002544 ZNF563 3.649402 7.4434 4.50E-06 0.002544 ELAVL4 63.23108 -4.74868 5.73E-06 0.003043 RP11-177H13.2 11.64542 -7.3444 7.84E-06 0.00357 SYT17 35.59363 -5.2386 7.68E-06 0.00357 HTN3 8.904382 5.115552 7.82E-06 0.00357 NCMAP 3.677291 5.559622 1.06E-05 0.004608 DDB1 31.25498 -4.56777 1.32E-05 0.00526 HACE1 10.55777 5.105336 1.27E-05 0.00526 EIF3K 39.88446 -4.6739 1.74E-05 0.006318 RNF41 19.59363 4.352775 1.78E-05 0.006318 LSM3 30.83665 4.919099 1.69E-05 0.006318 ABLIM1 43.86853 -4.47145 1.88E-05 0.006431 PSD3 24.70518 -4.46534 2.86E-05 0.009431 PCNXL4 24.79681 3.170657 3.31E-05 0.010559 WHAMML1 22.71315 6.699088 3.72E-05 0.011463 146 Table 3-12. Potential Biomarkers of motor neurons with 3 different electrophysiological patterns. Gene baseMean pvalue padj Marker MANSC1 32.98406 6.34E-11 6.58E-07 Repetitive Non-adaptive CD226 20.60558 2.37E-08 5.48E-05 Repetitive Non-adaptive DOCK5 6.645418 3.44E-08 5.48E-05 Repetitive Non-adaptive PSPH 7.844622 4.57E-08 0.000119 Repetitive Non-adaptive GPR75-ASB3 22.53386 1.25E-07 0.000227 Repetitive Non-adaptive ARIH1 65.21912 3.28E-10 2.24E-06 Repetitive Adaptive C11orf88 16.3506 1.38E-09 5.28E-06 Repetitive Adaptive AP1G1 31.47012 9.14E-09 4.74E-05 Repetitive Adaptive RP11- 179K3.2 29.64143 1.65E-07 0.000333 Repetitive Adaptive UBE3B 11.56972 4.09E-07 0.000446 Repetitive Adaptive SORL1 38.14343 6.36E-14 8.67E-10 Single USP34 16.76096 1.61E-09 7.68E-06 Single SYT17 35.59363 7.22E-09 1.97E-05 Single UBXN2A 21.65737 1.64E-07 0.000224 Single EIF3K 39.88446 1.95E-07 0.000333 Single 147 Chapter 4. Electrophysiological pattern and Gene expression pattern from Human Temporal Lobe Studies 4.1 Introduction 4.1.1 Overview The adult human cerebral cortex is composed of several different neuronal tissues such as frontal lobe, temporal lobe, parietal lobe and occipital lobe. In each area, immunohistochemistry is able to show the separation of cortical layers: layer I-VI (Figure 4-1). The human brain transcriptome has been studied at the tissue level. However, because of the limited access to tissue along with limitations in single cell technology, few studies have investigated individual cells in brain cortical layers (Christophe et al., 2005; Peng, Barreda Tomás, Klisch, Vida, & Geiger, 2017; Vogt Weisenhorn et al., 1994). Each layer of human brain is mainly composed of Pyramidal neurons and Interneurons. The pyramidal neurons typically project to different areas of the central nervous system (CNS) such as thalamus and spinal cord to regulate homeostatic functions or to perform motor activity through cortico- thalamic and cortico-spinal tract (Judaš & Pletikos, 2010; Levitt, Lewis, Yoshioka, & Lund, 1993; Salimi, Friel, & Martin, 2008). The interneurons form 148 the neuronal network in the brain. These networks regulate the major functions of the brain such as learning and memory. In order to study functions of neurons in the human cerebral cortex, three different approaches have generally been used: morphological studies based on electron microscopy, electrophysiological recordings, and the genetics approach (i.e. expression profile studies). Morphological studies were first initiated following the discovery of neurons by Santiago Ramon y Cajal (Ringstedt et al., 1998). The original Nissl stain allowed them to observe the shape of the neurons and the connections between them. Microscopy and staining techniques were also used in order to understand the different layers of the cortical structure (Valverde, 1977). The development of immunohistochemistry allowed scientists to study the expression and localization of specific proteins in neurons. For example, the Allen Brain Institute has performed immunohistochemistry on 256 different proteins on coronal sections of the human and mouse brain. The electrophysiological approach was first developed by Hodgkin and Huxley (Hodgkin & Huxley, 1952). Technique such as voltage clamp and current clamp were used for understanding electrical activity of specific neurons. The patch clamp technique was developed to make this observation available for small neurons (Sakmann & Neher, 1984). For instance, in many epilepsy and behavioral studies, activity of neurons in specific nuclei or ganglions were investigated to understand how the electrophysiological 149 properties of neurons are affected by changes in behavior(Sakmann & Neher, 1984). Recently, gene expression profiling has become one of a necessity in the study of neurons. Unlike the Immunohistochemistry, RNA-seq allows us to study the whole transcription profile, which represents the genome-wide expression of neurons. Moreover, the expression profiling of different neuronal types in cortical layers could be used to identify cell-type specific markers, which can be used for selecting antibodies in Immunohistochemical studies. The connectivity of pyramidal neurons and a potential function of interneurons based on their morphology varies depending on their location, electrical activity and projections. Therefore, the results from all three methods could be beneficial to understand cortical functions. 4.1.2 Layer Specific Expression Profile of Neurons The pyramidal neurons are the neurons known as the “output” of the cortex. The expression profile of single pyramidal neurons in different layers could reveal general differences in the projections and the function of each pyramidal neuron. It was known that pyramidal neurons from different layers have different functions. In several animal models, the functions of these neurons have been defined. For example, layer II/III neurons in occipital lobe, which form specific structure called columnar organization, are related to function of color sensory functions, whereas layer IV neurons are controlling motor control (Christopher A. Mutch, Nobuo Funatsu, Edwin S. Monuki, 2009; 150 Gilbert & Kelly, 1975; Rakic, 1974). The layer V neurons are known to have subtypes distinguished by their electrophysiological pattern and projection target (Molnár & Cheung, 2006). Previous studies investigated expression profiling of pyramidal neuron projections using microarray and RNAseq in both mouse and rat models (J.-G. Chen, Rasin, Kwan, & Sestan, 2005). However, in humans, the accessibility of brain tissue is limited and there have only been a few studies done with pyramidal neurons in human temporal lobe. The layer VI neurons have been studied on their diverse projecting pattern to several different areas such as spinal cord and thalamus (Thomson, 2010) but expression profile studies are still limited (Figure 4-2). The RNAseq of each layer revealed that there is a layer specific expression pattern in human brain (Belgard et al., 2011; Miller et al., 2014) indicating that there are possible key genes linked to specific neuronal function. However, so far, most of the previous studies have used a cell population in each layer, not individual cells. Aside from pyramidal neurons, each layer of the brain contains different types of cells, including glial cells. Therefore, to understand the neuronal function of each neuron in a layer, studies carried out at the single cell resolution are preferred. By using PAIA protocol, 273 cells were collected and sequenced from different layers (I~VI) of human temporal lobe. 151 4.1.3 Electrophysiological Profile and Subtypes Electrophysiological profiles of the pyramidal neurons are known to have repetitive firing patterns (Shao, Halvorsrud, BorgGraham.Lyle, & Storm, 1999). Within repetitive firing patterns, based on the adaptiveness differences, layer V pyramidal neurons in sensorimotor cortex were classified to RSna and RSad groups in rat model (P12). It was also found that there were projection target differences between these two groups (Christophe et al., 2005; Franceschetti, Sancini, Panzica, Radici, & Avanzini, 1998; Gao & Zheng, 2004). However, the expression profile differences between each subtype of pyramidal neurons and interneurons were understudied. Single cell qPCR on selected genes was used for classifying neurons with distinct electrophysiological patterns (Christophe et al., 2005), but only a limited number of genes were characterized. To profile genome-wide gene expression in single neurons and determine brain cortex layer specific genes, the electrophysiological pattern was collected and expression levels were profiled using PAIA protocol from 129 pyramidal neurons. 4.2 Materials and Methods 4.2.1 Tissue preparation For tissue preparation, tissue chucks were provided by the epilepsy clinic (USC hospital). Specimens were transported to the laboratory in 152 Hypothermosol solution (Sigma, USA; Biolife Solutions, USA) on ice. Tissue transport time was always within two hours. Brain chunks were then visually examined for structural integrity and then glued on a specimen plate. The tissue was then sliced into 400m slices in cold (4˚C), oxygenated (95%O2, 5% CO2) NMDG- artificial cerebral spinal fluid (NMDG-ACSF; NMDG 93mM; KCl 2.5mM; NaH2PO4 1.2mM; NaHCO3 30mM; HEPES 20mM; glucose 25mM; sodium ascorbate 5mM; thiourea 2mM; sodium pyruvate 3mM; MgSO4 10mM; CaCl2 0.5mM) using a vibratome (Leica VT1200S, Germany). Vertical deflection of the blade was minimized with the Vibrocheck technology and slicing parameters were: speed 0.1-0.15 mm/s and vibration amplitude 1.5 mm. Slices were transferred to a recovery chamber (32˚C, 95%O2, 5% CO2) and allowed to recover in ACSF (NaCl 124 mM; KCl 4 mM; NaHCO3 26 mM; glucose 10 mM; CaCl2 2mM; MgCl2 2mM) for 30-60 minutes before recording and cell collection was performed. 4.2.2 Collection, Ephys and RNA sequencing PAIA protocol was performed as described in Chapter 1,2 and 3. During this step, phenotypic information such as location and morphology was collected. 685 neurons from human temporal lobe were processed with this method (Figure 4-3). 4.2.3 Data processing and analysis The RNA-seq data collected from sequencer was demuliplexed with CASAVA and converted into fastq format. Using the GT-FAR pipeline 153 (https://genomics.isi.edu/gtfar), PerM (Y. Chen et al., 2009) was used to align the data and generate a table for gene expression. In this step, the mapping rate to genome and transcriptome was calculated and PCR duplicates were removed. Due to the sparseness of the single cell data, it was inefficient to use either quartile or quantile normalization methods when processing single cell data. Therefore, the processed data tables were downsampled to 100,000 for normalization. Differential expression analysis was performed using DESeq2 (http://bioconductor.org/packages/release/bioc/html/DESeq2.html). The data with layer information (235 temporal lobe neurons) were analyzed for result 4.3.1 and the data with electrophysiological pattern information (190 temporal lobe neurons) were analyzed for result 4.3.2. For layer analysis, differentially expressed genes between the sequential layers were examined. Based genes found to be differentially expressed between cortical layers, the potential biomarkers were found for each layer by comparing up and downregulated gene sets. 154 4.3 Result 4.3.1 Layer Specific Expression on temporal lobe pyramidal neurons Pyramidal neurons from Layer I~VI were collected (Table 4-1). Few pyramidal cells were found between layers. These were labeled as LayerA_B and were excluded from the analysis. There was a small morphological difference between neurons from different layers (Figure 4-4). The differential expression analysis was performed between pyramidal neurons collected from different layers. With DESeq2 method, between Layer I and Layer II, 318 differentially expressed genes were identified (146 downregulated, 172 upregulated in layer 1) (Table 4-2). By significance, the most differentially expressed genes were GGPS1 (Geranylgeranyl Diphosphate Synthase 1, p=2.07E-7, log fold change=-5.9) and TSEN2 (TRNA Splicing Endonuclease Subunit 2, p=2.07E- 7, log fold change=8.2). When differentially expressed gene sets were analyzed with Gene Set Enrichment analysis (GSEA), 19.8% of downregulated genes were known to be involved in genes downregulated in Alzheimer disease (FDR q=8.75E-15) and 15.7% of upregulated genes were involved in cellular response pathways (FDR q=2.37E-6) (Table 4-3). Between Layer II and III neurons, 396 genes were differentially expressed (266 downregulated, 130 upregulated in layer II,Table 4-4). The most 155 differentially expressed genes defined by lowest p-value were DNAJA1 (p=1.70E-9) and RPH3AL (p=3.43E-6). Both were upregulated in layer II (log fold change>7). In downregulated genes in layer II, GSEA found that 10% of these genes are related to RNA binding function (FDR q=2.74 E-7) and in upregulated genes, 18% of the genes were related to RNA binding function (FDR q =1.32E-7) (Table 4-5). Comparing layer III and layer IV, 1,020 genes were differentially expressed (294 downregulated, 726 upregulated with respect to layer II, Table 4-6). Based on highest level of significance (-log p), the most differentially expressed genes were GTDC1 (p=2.29E-15), PTAFR (p=1.93E-14) and PARK7 (p= 2.39E-10). All three genes were downregulated in layer III (log fold change >4). PARK7 gene is also known to be related to Parkinson’s disease (Nunome, Miyazaki, Nakano, Iguchi-Ariga, & Ariga, 2008). GSEA on downregulated genes was able to show enrichment in several neuronal disease related gene sets such as Alzheimer’s disease (FDR q=3.0E-22), Oligodendrocyte related disease (FDR q= 3.02E-18) and bipolar disorders (FDR q= 3.36E-13). The upregulated genes were found to be enriched in the cytoskeleton gene set (FDR q= 2.61E-13) and endogenous responses gene set (FDR q= 2.64 e-12)(Table 4-7). The layer IV and layer V had 318 differentially expressed genes (211 downregulated and 107 upregulated in layer IV, Table 4-8). By p-values, the most differentially expressed genes were PTAFR (p=9.38E-9), GTDC1 156 (p=1.08E-8) and TXK (p=1.76E-08). PTAFR and GTDC1 were found to be upregulated in layer IV (log fold change >6,4) and TXK (Tyrosine Kinase) was downregulated in layer IV (log fold change>7). GTDC1 which was also found to be upregulated in layer IV from previous analysis with layer III, also found upregulated in layer IV in this analysis, showing that this gene could be a biomarker for layer IV. Genes downregulated in layer IV were involved in the positive response pathway (FDR q=7.24E-5) and Neuromedin U pathway (FDR q=1.89E-5). With the upregulated gene lists, Alzheimer’s disease related genes (FDR q=6.81E-7) and genes involved in Enzyme binding pathway were found to be enriched (FDR q=8.8E-6) (Table 4-9). In between layer V and VI, 108 differentially expressed genes were found to be differentially expressed (54 genes each upregulated/downregulated in layer V, Table 4-10). The most differentially expressed genes by p-value were NCOR1 (p=2.71E-08), RP11-573D15.8 (p=1.90E-05), and TXK (p=1.90E-05). The first two genes were downregulated (fold change >5) and TXK was upregulated (fold change>7) in layer V compared to layer VI. One of the pathways found to be enriched with downregulated genes in layer V was dynein pathway (FDR q=1.47E-2). GSEA found that genes involved in Oligodendrocyte functions (FDR q= 4.1E-3) and Microglia in brain compared to other organs (FDR q=2.57E-5) were upregulated genes in layer V (Table 4-11). 157 DESeq2 analysis was also used for non-consecutive combination such as I vs III, I vs IV and II vs V (Table 4-12). This information was used for identify potential biomarkers of each layer of brain cortex. 4.3.2 Biomarkers and known marker analysis In order to identify the biomarkers in each layer, the list of upregulated and downregulated genes in each layer from the above consecutive comparisons were combined (i.e. I vs II, I vs III, I vs IV, I vs V, I vs VI, II vs III, II vs IV, II vs V, II vs VI, III vs IV, III vs V, III vs VI, IV vs V, IV vs VI, V vs VI) (Table 4-12). Then, based on the occurrences (between 5 comparisons each) in the combined table, the genes with 4-5 occurrences were selected as the potential biomarkers. Layer I had 120 upregulated biomarkers and 27 downregulated biomarkers (Table 4-13). The upregulated biomarkers indicated the RNA binding (FDR q value=1.1E-4) as the most enriched pathway in GSEA. The downregulated genes were enriched with CALB1 related disease pathway (FDR q=7.33E-11). When both biomarker lists were combined, the Alzheimer pathway found to be the pathway with most genes were enriched (FDR q=9.69E-7) (Table 4-14). The layer II had 19 upregulated biomarkers and no downregulated biomarkers. Because the number of differentially expressed genes was low, GSEA could not find statistically significantly enriched pathways. Layer III had 138 upregulated biomarkers and only 1 downregulated biomarker, TAF13. In GSEA, the cytoskeleton related pathway was mostly enriched (FDR q=3.02E- 158 3). TAF13 is TATA box binding protein which is related to RNA polymerase II binding site. layer IV had 45 upregulated markers and 2 downregulated markers. GSEA found enzyme binding pathway as the most enriched pathway (FDR q=3.26E-3). Layer V had 47 upregulated biomarkers with no downregulated biomarkers. Many of these upregulated genes were involved in microglia pathway (FDR q=1.35E-2) and cellular communication (FDR q=2.17E-2). Layer VI had 33 upregulated biomarkers and no downregulated biomarker. GSEA found that genes in dendritic cells (DC) compared to neurophils were enriched in these biomarkers (FDR q=1.41E-2) (Table 4-14). The expression levels of the potential biomarkers were plotted into sets of boxplots (Figure 4-5). 4.3.3 Electrophysiological pattern of temporal lobe cells The electrophysiological patterns of temporal lobe pyramidal neurons were composed of two different patterns, repetitive adaptive and non-adaptive, similar to what was found from neurons in the embryonic spinal cord (Figure 4- 6). All the cells with action potential firing had repetitive action potential firing: repetitive adaptive or non-adaptive. Unlike the embryonic spinal cord, most of the neuron populations, specifically the pyramidal cell population, are not including any immature neurons. Therefore, though there is some similarity in general electrophysiological pattern, expression profiles between repetitive adaptive and non-adaptive should be different in pyramidal cells and in embryonic spinal cord. 159 DESeq2 found 588 differentially expressed (360 upregulated in repetitive, 228 downregulated in repetitive adaptive, Table 4-15) between repetitive adaptive and repetitive non-adaptive neurons. GSEA found enrichment in Alzheimer’s disease related genes (FDR q=1.29E-18) and Oligodendrocyte differentiation functions (FDR q= 4.27E-12) in upregulated genes in repetitive adaptive. Neuroblastoma pathway (FDR q=1.16 E-6) was found to be enriched in downregulated genes (Table 4-16). The expression profiles of differentially expressed genes in each type of cells were plotted within a heatmap (Figure 4-7). 4.4 Discussion 4.4.1 Pyramidal Neurons and layers, Biomarkers Differential gene expression analysis between pyramidal cells from different layers of the temporal lobe showed the specific role and projections of each pyramidal neuron in an each layer (J.-G. Chen et al., 2005). In recent studies, it was found that distinct layer-specific excitatory and inhibitory synapses were organized in layers in rat presubiculum (Peng et al., 2017). In the results, single cell RNA-seq data presented above, genes such as EPHB2 and NEURL1, which are uniquely involved in axonal guidance in notch pathway(D’Souza, Miyamoto, & Weinmaster, 2008), were found to be biomarkers of the pyramidal neurons in layer I. This finding suggests that the neurons in layer I could potentially determine the projection of other neurons. 160 Genes found to significantly downregulated in layer IV compared to layer III or V, were found to be enriched in genes related to Alzheimer’s disease. GSEA also found that the genes downregulated in Alzheimer’s disease were enriched in genes differentially expressed in layer IV. Previous studies showed that neurons in layer IV were also linked to disease (Arnold, Hyman, Flory, Damasio, & Van Hoesen, 1991; Kowall & Kosik, 1987). The potential biomarkers of layer VI may have relation to the Parkinson disease (Kahle, Waak, & Gasser, 2009; Nuytemans, Theuns, Cruts, & Van Broeckhoven, 2010). For example, PARK7 was upregulated in layer VI. Another interesting finding was that both CEBPZOS and CSRP1 genes were found to be markers for both layer I and layer II. No other biomarkers showed overlaps among layers other than these two genes. CEBPZOS is the opposite strand of CEBPZ gene which encodes an enhancer binding protein that alters expression of the heat shock protein, HSP70 (Lum, Sultzman, Kaufman, Linzer, & Wu, 1990). The CSRP1, cysteine and glycine rich protein 1, is related to differentiation of neurons and oligodendrocytes (Dugas, Tai, Speed, Ngai, & Barres, 2006; Park et al., 2012). 4.4.2 Electrophysiological pattern When gene expression levels of repetitive adaptive neurons were compared with repetitive non-adaptive neurons, the GSEA result indicated that these two neuron populations have upregulated genes involved in Alzheimer’s disease and downregulated genes involved in the neuroblastoma pathway. 161 Previous studies have reported that Alzheimer’s disease and neuroblastomas have similarities with respect to neuronal activities (Koriyama, Furukawa, Muramatsu, Takino, & Takeuchi, 2015; Wozniak, Hutchison, Morris, & Hutchison, 1998). Findings from my study may reveal specific subgroups of neurons linked to these neurological diseases. One of the most significant findings from the list of differentially expressed genes was that there were many Zinc Finger Domain proteins (ZNF197, ZNF221, ZNF229, ZNF346, ZNF532, ZNF682, ZNF707, ZNF738, ZNF776, ZNF84) in gene sets overexpressed in repetitive non-adaptive neurons. On the other hand, few zinc finger domain proteins (ZNF280D, ZNF362, ZNF415, ZNF555, ZNF562) were highly expressed in repetitive adaptive neurons. The Zinc Finger proteins are one of the DNA binding transcription factors, which can regulate of the expression of their target genes (Pavletich & Pabo, 1991). Interestingly, a change in zinc finger proteins’ expression was only associated with electrophysiological pattern change not changed between different layers. Neuron specific markers such as TUBA4A were highly expressed in repetitive neurons compared to non-adaptive neurons. Although most of the neurons in temporal lobe are mature neurons, unlike spinal cord, differential expression of neuronal markers could mean that there still are some developing cells within a tissue. To test this, some of neuronal and pre- neuronal markers, provided by R&D systems, were compared 162 (https://www.rndsystems.com/research-area/neural-stem-cell-markers) in the collected neurons. The neural stem cell markers and neural progenitor markers were not present in pyramidal neurons from the human brain cortex layers. The repetitive non-adaptive neurons had lower expression levels of immature neuron markers than adaptive neurons. The repetitive non-adaptive neurons expressed more glutamatergic markers, but adaptive neurons expressed more of mixture of GABAergic and serotonergic markers. Both did not express cholinergic markers (Figure 4-8). 163 4.5 Figures Figure 4-1. Six layers of cerebral cortex. The picture was adapted from the image generated by Santiago Ramon y Cajal. from Kandel et al. Principles of Neural Science. 164 Figure 4-2. Diversity of layer VI pyramidal neurons and projections Figure illustrates the diversity of neurons in layer VI projecting to the different organs Picture from (Thomson, 2010). 165 Figure 4-3. Distribution of number of cell types collected. Most cells were pyramidal neurons (pyramidal shape) from the temporal lobe. Some cells from the hippocampus, which were used for the epilepsy project, were also collected. 166 Figure 4-4. Morphology of pyramidal neurons from different layers (A) Layer I (B) Layer II (C) Layer III (D) Layer IV (E) Layer V (F) Layer VI 167 Figure 4-5. Normalized gene expression of the potential biomarkers. (A) Potential biomarker for Layer I (C2orf168), (B) Potential biomarker of Layer II (CEBPZOS), (C) Potential biomarker for Layer III (SPTLC3), (D) Potential biomarker for Layer IV (UCHL1), (E) Potential biomarker for Layer V (PGAP1), (F) Potential biomarker for Layer VI (MTPAP). 168 Figure 4-6. Two electrophysiological patterns in Temporal lobe. (A) Repetitive Adaptive (B) Repetitive Non-Adaptive 169 Figure 4-7. Boxplot of differentially expressed genes between repetitive adaptive vs non adaptive neurons. (A)ZNF562,(C)BLOC1S6 ,(E)PLEKHA are upregulated in Repetitive Adaptive neurons. (B)PPM1B,(D)NTRK3,(F)MATR3 are upregulated in Repetitive Non-adaptive neurons. 170 Figure 4-8. Expression of markers for neurons based on neurotransmitter Purple arrow indicates the glutamatergic markers, Red indicates GABAergic markers and genes with no arrow are Cholinergic markers. 171 4.6 Tables Table 4-1. Temporal lobe cells collected by layers. Layer Number LAYER1 31 LAYER2 13 LAYER3 52 LAYER4 21 LAYER5 20 LAYER6 14 LAYER1_2 17 LAYER2_3 12 LAYER3_4 24 LAYER4_5 58 LAYER5_6 11 LayerA_B indicates neurons that were found between two layers. These neurons are not used for analysis in this study. 172 Table 4-2. Layer I VS Layer II DESeq analysis Gene baseMean log2FoldChange pvalue padj GGPS1 13.00733 -5.90226 4.32E-11 2.07E-07 RP11-119D9.1 36.08791 5.730516 4.48E-11 2.07E-07 TSEN2 17.51648 8.056596 3.11E-11 2.07E-07 TTC39C 12.28938 -4.73361 1.17E-09 4.06E-06 ARAP2 11.05495 -6.85017 5.25E-09 1.22E-05 C12orf76 14.23077 -4.20792 5.09E-09 1.22E-05 RPH3AL 10.92674 -7.20953 7.77E-09 1.54E-05 NUCKS1 35.91941 -5.58037 3.48E-08 5.06E-05 EXTL3 6.791209 6.40141 3.45E-08 5.06E-05 REEP4 16.75092 9.357725 3.64E-08 5.06E-05 RPS27L 26.20513 4.860692 7.88E-08 8.42E-05 HN1 48.72161 6.041024 7.25E-08 8.42E-05 PLXNA2 18.45055 6.78308 7.36E-08 8.42E-05 RGS5 24.2967 -7.42371 9.80E-08 9.73E-05 DGCR2 8.380952 -7.62869 1.14E-07 0.000106 DNAJA1 166.3663 -5.36336 1.39E-07 0.000121 EIF3M 29.59707 4.787517 1.75E-07 0.000143 NEFL 36.64469 -5.37601 3.08E-07 0.000238 GNRHR2 9.684982 4.595291 4.10E-07 0.000271 LINC00641 40.26007 6.218743 3.86E-07 0.000271 USP38 13.30037 7.277031 4.08E-07 0.000271 CCDC6 16.86813 -5.99744 4.32E-07 0.000273 FAM179A 37.3663 3.591471 5.36E-07 0.000324 TMEM39A 26.70696 7.488721 5.64E-07 0.000327 EPT1 6.978022 -5.37242 7.49E-07 0.000416 C5orf22 6.681319 -6.62785 8.08E-07 0.000432 NDUFA5 34.38828 -4.45698 1.55E-06 0.000797 B4GALT5 61.16117 5.236865 1.94E-06 0.000963 PAQR5 7.545788 6.792889 2.08E-06 0.000963 ZNF793-AS1 6.208791 7.690744 2.03E-06 0.000963 173 Table 4-3. Gene Set Enrichment Analysis (GSEA) of differentially expressed genes between Layer I and II. (top) Table shows GSEA results using upregulated genes in layer I. (bottom) Table shows GSEA results using downregulated genes in layer I. Name [# Genes (K)](Upregulated) Overlap (k) p-value FDR q- value AAAYWAACM_HFH4_01 [1890] 29 6.54 e-13 1.16 e-8 DACOSTA_UV_RESPONSE_VIA_ERCC3_DN [855] 19 2.06 e-11 1.83 e-7 GO_CELLULAR_RESPONSE_TO_STRESS [1565] 23 4.96 e-10 2.37 e-6 JOHNSTONE_PARVB_TARGETS_3_DN [918] 18 5.33 e-10 2.37 e-6 YRTCANNRCGC_UNKNOWN [2940] 31 1.17 e-9 4.16 e-6 CHARAFE_BREAST_CANCER_LUMINAL_VS_MESENCHYMAL_DN [460] 13 2.12 e-9 6.28 e-6 GO_ENDOPLASMIC_RETICULUM_PART [1163] 19 3.41 e-9 8.66 e-6 GO_ENDOPLASMIC_RETICULUM [1631] 22 5.87 e-9 1.3 e-5 PILON_KLF1_TARGETS_DN [1972] 24 8.03 e-9 1.59 e-5 DODD_NASOPHARYNGEAL_CARCINOMA_DN [1375] 20 8.96 e-9 1.59 e-5 Gene Set Name [# Genes (K)] (Downregulated) Overlap (k) p-value FDR q- value KIM_ALL_DISORDERS_CALB1_CORR_UP [548] 22 3.5 e-19 6.22 e-15 BLALOCK_ALZHEIMERS_DISEASE_DN [1237] 29 9.85 e-19 8.75 e-15 YRTCANNRCGC_UNKNOWN [2940] 37 4.55 e-15 2.7 e-11 GO_PHOSPHATE_CONTAINING_COMPOUND_METABOLIC_PROCESS [1977] 27 7.31 e-12 3.25 e-8 GNF2_AF1Q [25] 6 7.37 e-11 2.62 e-7 GO_MITOCHONDRION [1633] 23 1.8 e-10 5.34 e-7 MODULE_11 [540] 14 4.6 e-10 1.05 e-6 MODULE_100 [544] 14 5.05 e-10 1.05 e-6 MODULE_137 [546] 14 5.3 e-10 1.05 e-6 MODULE_66 [552] 14 6.1 e-10 1.06 e-6 174 Table 4-4. Layer II VS Layer III Differentially expressed genes Gene baseMean log2FoldChange pvalue padj DNAJA1 166.3663 7.078704 1.39E-13 1.70E-09 RPH3AL 10.92674 7.250084 5.62E-10 3.43E-06 LYRM2 18.94872 -5.20124 4.40E-09 1.79E-05 TNRC18 32.16117 -7.76069 6.28E-09 1.84E-05 TTC39C 12.28938 4.195663 7.51E-09 1.84E-05 LINC00954 42.38462 -6.08156 1.09E-08 2.23E-05 SEC14L1 45.24176 -5.51326 2.01E-08 3.19E-05 GGPS1 13.00733 4.658627 2.09E-08 3.19E-05 DCTN4 24.08059 -6.07963 2.74E-08 3.72E-05 PSMB6 12.57143 4.729273 6.60E-08 8.06E-05 SPTLC3 43.57509 -3.46259 1.10E-07 0.000123 NRG1 13.30403 -6.86736 1.39E-07 0.000131 CRELD2 10.52015 -6.63387 1.30E-07 0.000131 FRMD6 5.043956 -7.10941 1.63E-07 0.000142 C12orf76 14.23077 3.433728 3.14E-07 0.000256 ADARB1 10.12454 5.842947 3.35E-07 0.000256 UHMK1 42.73993 5.56756 3.57E-07 0.000257 CEBPZOS 7.886447 3.883163 3.80E-07 0.000258 RASA4 2.505495 5.604143 5.40E-07 0.000347 SOAT1 10.99634 -4.85543 8.50E-07 0.000519 CARF 17.87546 -4.66965 1.13E-06 0.000627 IDH2 11.96337 -3.88502 1.08E-06 0.000627 RGS5 24.2967 6.294957 1.36E-06 0.000724 APPL2 10.72527 -6.78151 1.44E-06 0.00073 PDXDC1 15.63736 4.4739 1.49E-06 0.00073 RSRC2 26.98901 -5.17359 1.72E-06 0.000807 GEMIN2 16.75458 -6.67346 1.80E-06 0.000812 CD44 9.941392 -6.00305 1.86E-06 0.000812 TEX41 19.34066 -6.30166 1.97E-06 0.000832 KLC4 21.97802 -4.95029 2.06E-06 0.00084 175 Table 4-5. GSEA of differentially expressed genes between Layer II and III. (Table on the top shows upregulated genes in Layer II and on the bottom shows downregulated genes.) Genee Set Name [# Genes (K)] (Up) # Genes in Overlap (k) p-value FDR q- value YRTCANNRCGC_UNKNOWN [2940] 34 3.76 e-14 6.69 e-10 GO_RNA_BINDING [1598] 23 1.48 e-11 1.32 e-7 PILON_KLF1_TARGETS_DN [1972] 24 1.55 e-10 9.21 e-7 BLALOCK_ALZHEIMERS_DISEASE_DN [1237] 19 3.7 e-10 1.64 e-6 GRAESSMANN_APOPTOSIS_BY_DOXORUBICIN_UP [1142] 18 7.51 e-10 2.67 e-6 WEI_MYCN_TARGETS_WITH_E_BOX [795] 15 2.01 e-9 5.97 e-6 GO_NEURON_PROJECTION [942] 16 2.5 e-9 6.36 e-6 DODD_NASOPHARYNGEAL_CARCINOMA_DN [1375] 18 1.34 e-8 2.7 e-5 JOHNSTONE_PARVB_TARGETS_3_DN [918] 15 1.37 e-8 2.7 e-5 KIM_ALL_DISORDERS_CALB1_CORR_UP [548] 12 1.81 e-8 3.21 e-5 Gene Set Name [# Genes (K)] (Down) # Genes in Overlap (k) p-value FDR q- value GRAESSMANN_APOPTOSIS_BY_DOXORUBICIN_DN [1781] 37 2.33 e-14 4.14 e-10 GO_POLY_A_RNA_BINDING [1170] 26 5.65 e-11 2.74 e-7 CUI_TCF21_TARGETS_2_DN [830] 22 7.06 e-11 2.74 e-7 JOHNSTONE_PARVB_TARGETS_2_DN [336] 15 7.54 e-11 2.74 e-7 JOHNSTONE_PARVB_TARGETS_3_DN [918] 23 7.7 e-11 2.74 e-7 GSE14908_ATOPIC_VS_NONATOPIC_PATIENT_RESTING_CD4_TCELL_UP [200] 12 2.16 e-10 6.41 e-7 BUYTAERT_PHOTODYNAMIC_THERAPY_STRESS_UP [811] 21 2.98 e-10 7.56 e-7 GO_RNA_BINDING [1598] 29 4.66 e-10 1.04 e-6 DACOSTA_UV_RESPONSE_VIA_ERCC3_DN [855] 21 7.65 e-10 1.38 e-6 GO_CATALYTIC_COMPLEX [1038] 23 8.26 e-10 1.38 e-6 176 Table 4-6. Layer III VS Layer IV Differentially expressed genes Gene baseMean log2FoldChange pvalue padj GTDC1 43.31868 -5.22648 1.55E-19 2.29E-15 PTAFR 30.31868 -6.45943 2.63E-18 1.93E-14 PARK7 45.79487 -4.69494 4.88E-14 2.39E-10 FAM178A 37.49817 -6.43997 3.25E-13 1.20E-09 TRIM69 18.2967 -5.3352 5.15E-13 1.52E-09 RREB1 32.31136 -5.30059 9.84E-13 2.41E-09 FITM2 27.37363 -5.35413 1.66E-12 3.49E-09 KLC4 21.97802 6.326936 1.05E-11 1.93E-08 PRKCE 17.33333 -6.37662 1.11E-10 1.81E-07 RSPRY1 23.63736 -6.04847 1.23E-10 1.81E-07 SPTLC3 43.57509 3.433078 5.23E-10 7.00E-07 RP11-345J4.5 7.230769 6.519069 6.03E-10 7.40E-07 ARHGAP28 31.76923 5.990531 6.84E-10 7.74E-07 TNRC18 32.16117 7.380692 9.73E-10 1.02E-06 GABRG3 13.20879 -4.68069 1.19E-09 1.17E-06 IPO7 41.11722 -4.82193 1.27E-09 1.17E-06 PHACTR4 22.35897 3.318925 1.87E-09 1.62E-06 NCK2 10.23443 -7.65592 2.72E-09 2.22E-06 CD226 12.92308 -4.0178 4.69E-09 3.64E-06 IPO5 32.44689 -4.05234 5.83E-09 4.29E-06 ING1 11.69597 -4.59094 6.42E-09 4.50E-06 RSU1 16.24908 -5.28085 7.27E-09 4.86E-06 XRN1 21.62637 5.027308 1.04E-08 6.65E-06 SDCCAG8 8.593407 5.120573 1.10E-08 6.73E-06 XPNPEP3 17.96337 -3.27214 1.33E-08 7.75E-06 INO80C 8.556777 4.789821 1.37E-08 7.75E-06 STX4 45.06593 -6.7417 1.49E-08 8.12E-06 ERCC4 13.45788 5.007054 1.68E-08 8.84E-06 LYRM2 18.94872 4.194451 2.00E-08 1.02E-05 CA5A 11.79487 3.700214 3.06E-08 1.50E-05 177 Table 4-7. GSEA of differentially expressed genes between Layer III and IV. (Table on the top shows upregulated gene in Layer III and bottom shows downregulated genes.) Gene Set Name [# Genes (K)] UP # Genes in Overlap (k) p-value FDR q- value GRAESSMANN_APOPTOSIS_BY_DOXORUBICIN_DN [1781] 75 2.76 e-22 4.9 e-18 TGGAAA_NFAT_Q4_01 [2061] 80 1.42 e-21 9.39 e-18 PILON_KLF1_TARGETS_DN [1972] 78 1.58 e-21 9.39 e-18 GO_CYTOSKELETON [1967] 70 5.88 e-17 2.61 e-13 TTGCWCAAY_CEBPB_02 [1972] 69 2.36 e-16 8.38 e-13 GO_RESPONSE_TO_ENDOGENOUS_STIMULUS [1450] 57 8.91 e-16 2.64 e-12 YRTCANNRCGC_UNKNOWN [2940] 86 1.32 e-15 3.32 e-12 GO_CELLULAR_RESPONSE_TO_ORGANIC_SUBSTANCE [1848] 65 1.49 e-15 3.32 e-12 GO_PHOSPHATE_CONTAINING_COMPOUND_METABOLIC_PROCESS [1977] 67 3.13 e-15 6.18 e-12 DIAZ_CHRONIC_MEYLOGENOUS_LEUKEMIA_UP [1382] 54 6.9 e-15 1.23 e-11 Gene Set Name [# Genes (K)]Down # Genes in Overlap (k) p-value FDR q- value BLALOCK_ALZHEIMERS_DISEASE_DN [1237] 49 1.69 e-26 3 e-22 KIM_ALL_DISORDERS_OLIGODENDROCYTE_NUMBER_CORR_UP [756] 36 3.4 e-22 3.02 e-18 DIAZ_CHRONIC_MEYLOGENOUS_LEUKEMIA_UP [1382] 44 3.92 e-20 2.32 e-16 YRTCANNRCGC_UNKNOWN [2940] 61 1.61 e-18 7.16 e-15 KIM_BIPOLAR_DISORDER_OLIGODENDROCYTE_DENSITY_CORR_UP [682] 29 9.45 e-17 3.36 e-13 PUJANA_BRCA1_PCC_NETWORK [1652] 43 1.65 e-16 4.88 e-13 JOHNSTONE_PARVB_TARGETS_3_DN [918] 32 6.25 e-16 1.59 e-12 GO_ESTABLISHMENT_OF_LOCALIZATION_IN_CELL [1676] 42 1.43 e-15 3.03 e-12 GO_GOLGI_APPARATUS [1445] 39 1.53 e-15 3.03 e-12 GO_POLY_A_RNA_BINDING [1170] 35 2.44 e-15 4.34 e-12 178 Table 4-8. Layer IV VS Layer V Differentially expressed genes Gene baseMean log2FoldChange pvalue padj PTAFR 30.31868132 6.36318897 6.75E-13 9.38E-09 GTDC1 43.31868132 4.904130277 1.56E-12 1.08E-08 TXK 35.05494505 -7.89209955 3.80E-12 1.76E-08 FITM2 27.37362637 6.169530644 1.28E-11 4.44E-08 KIAA1257 9.406593407 -6.665315028 6.72E-11 1.87E-07 MDFIC 19.08424908 -7.681648674 2.98E-10 6.89E-07 RREB1 32.31135531 5.532594185 4.67E-10 8.44E-07 SPTLC2 23.44688645 -6.514065745 4.86E-10 8.44E-07 RBM23 17.07692308 -4.706882964 6.06E-10 9.36E-07 FAM178A 37.4981685 6.216507415 2.91E-09 3.89E-06 SERAC1 24.88644689 -5.659727933 3.08E-09 3.89E-06 GABPB1-AS1 44.56410256 4.140273937 6.34E-09 6.78E-06 ZNF26 28.00732601 -3.808226212 5.98E-09 6.78E-06 TANGO2 6.465201465 -4.849995537 8.46E-09 8.40E-06 AZI2 7.32967033 8.781780026 1.08E-08 9.12E-06 RSPRY1 23.63736264 6.369397258 1.06E-08 9.12E-06 GABRG3 13.20879121 5.286078849 1.12E-08 9.12E-06 RP1-283E3.8 7.710622711 -6.599352233 1.96E-08 1.51E-05 FASN 20.47252747 -6.432264128 2.90E-08 2.12E-05 IPO7 41.11721612 5.195849554 3.93E-08 2.73E-05 IPO5 32.44688645 4.477213595 7.66E-08 5.07E-05 GINS4 4.098901099 -5.192544715 8.32E-08 5.26E-05 PAXBP1-AS1 7.864468864 -5.037603504 8.70E-08 5.26E-05 MRPL22 19.36630037 7.101048824 1.56E-07 9.02E-05 PLEKHA3 21.95970696 5.113362739 2.13E-07 0.000116039 STK25 9.194139194 -4.535638957 2.17E-07 0.000116039 SCIN 15.39194139 -4.705455761 2.29E-07 0.000117794 PTPRQ 8.256410256 -6.832559852 2.77E-07 0.000137241 DNAJA2 47.88278388 6.444133585 3.04E-07 0.000145725 TRIM69 18.2967033 4.46844835 3.66E-07 0.000169706 179 Table 4-9. GSEA of differentially expressed genes between Layer IV and V. (Table on the top shows upregulated gene in Layer IV and bottom shows downregulated genes.) Gene Set Name [# Genes (K)] UP # Genes in Overla p (k) p-value FDR q- value GO_RESPONSE_TO_OXYGEN_CONTAINING_COMPOUND [1381] 19 5.49 e- 11 6.81 e-7 BLALOCK_ALZHEIMERS_DISEASE_DN [1237] 18 7.66 e- 11 6.81 e-7 BENPORATH_MYC_MAX_TARGETS [775] 14 7.77 e- 10 4.61 e-6 PUJANA_BRCA1_PCC_NETWORK [1652] 19 1.09 e-9 4.84 e-6 GO_ENZYME_BINDING [1737] 19 2.48 e-9 8.8 e-6 BYSTRYKH_HEMATOPOIESIS_STEM_CELL_QTL_TRANS [882] 14 4.03 e-9 1.2 e-5 GO_MITOCHONDRION [1633] 18 6.15 e-9 1.43 e-5 JOHNSTONE_PARVB_TARGETS_3_DN [918] 14 6.68 e-9 1.43 e-5 GO_ENVELOPE [1090] 15 7.24 e-9 1.43 e-5 PILON_KLF1_TARGETS_DN [1972] 19 1.91 e-8 3.26 e-5 Gene Set Name [# Genes (K)] Down # Genes in Overla p (k) p-value FDR q- value GO_POSITIVE_REGULATION_OF_MOLECULAR_FUNCTION [1791] 27 2.56 e- 10 4.55 e-6 GO_REGULATION_OF_PROTEIN_MODIFICATION_PROCESS [1710] 25 2.37 e-9 1.89 e-5 GSE1791_CTRL_VS_NEUROMEDINU_IN_T_CELL_LINE_6H_DN [200] 10 3.19 e-9 1.89 e-5 GO_PHOSPHATE_CONTAINING_COMPOUND_METABOLIC_PROCESS [1977 ] 26 9.62 e-9 4.27 e-5 GO_POSITIVE_REGULATION_OF_RESPONSE_TO_STIMULUS [1929] 25 2.55 e-8 7.24 e-5 DIAZ_CHRONIC_MEYLOGENOUS_LEUKEMIA_UP [1382] 21 2.65 e-8 7.24 e-5 ZHENG_BOUND_BY_FOXP3 [491] 13 2.86 e-8 7.24 e-5 GO_POSITIVE_REGULATION_OF_CELL_COMMUNICATION [1532] 22 3.26 e-8 7.24 e-5 GO_REGULATION_OF_PHOSPHORUS_METABOLIC_PROCESS [1618] 22 8.41 e-8 1.55 e-4 GO_SMALL_MOLECULE_METABOLIC_PROCESS [1767] 23 9.07 e-8 1.55 e-4 180 Table 4-10. Layer V VS Layer VI Differentially expressed genes Gene baseMean log2FoldChange pvalue padj NCOR1 44.0989011 -5.998406068 1.74E-12 2.71E-08 RP11-573D15.8 17.75457875 -5.446859577 2.48E-09 1.90E-05 TXK 35.05494505 7.208805354 3.67E-09 1.90E-05 EMC3 21.58241758 -5.402644676 3.53E-08 0.000137261 NRD1 17.26739927 -6.014677328 5.43E-08 0.000169048 DNAH11 10.66666667 -7.821972195 8.76E-08 0.000227126 RAB21 35.4029304 4.676787599 2.03E-07 0.000450064 MKRN9P 17.73260073 8.621131681 2.99E-07 0.000581439 NUAK2 7.677655678 -4.976608887 5.41E-07 0.000841135 CACHD1 23.84249084 7.703648537 4.92E-07 0.000841135 TANGO2 6.465201465 4.482682623 1.12E-06 0.001446637 KIAA1257 9.406593407 5.284936128 1.03E-06 0.001446637 LINC00476 12.53113553 -3.922730272 1.78E-06 0.001989668 SERAC1 24.88644689 4.948049401 1.92E-06 0.001989668 GOT2 25.73260073 6.28585795 1.85E-06 0.001989668 CLDN19 3.146520147 -6.372095544 2.49E-06 0.002419549 RBM23 17.07692308 3.892399971 2.87E-06 0.002629886 SLC36A4 13.08058608 6.07877095 3.93E-06 0.003395598 ZCCHC11 9.230769231 4.398931199 4.56E-06 0.003734937 RP11-365F18.3 4.769230769 6.052353344 5.59E-06 0.003782212 LIN7B 12.81318681 6.135272781 5.41E-06 0.003782212 WDR43 5.663003663 6.316618004 5.46E-06 0.003782212 HRNR 3.644688645 7.759341398 5.02E-06 0.003782212 WWP2 4.351648352 -6.687333369 5.88E-06 0.003813465 FASN 20.47252747 5.565012724 8.07E-06 0.005023151 MDFIC 19.08424908 5.66150581 9.58E-06 0.005517404 ZNF622 3.919413919 7.204065992 9.39E-06 0.005517404 KDELR3 4.293040293 7.545494522 1.09E-05 0.006046058 MSL3 10.86080586 6.330891003 1.52E-05 0.008156452 EIF2B4 5.868131868 -5.999173127 1.85E-05 0.009168186 181 Table 4-11. GSEA of differentially expressed genes between Layer V and VI. (Table on the top shows upregulated gene in Layer IV and bottom shows downregulated genes.) Gene Set Name [# Genes (K)] UP # Genes in Overlap (k) p-value FDR q- value GSE29949_MICROGLIA_BRAIN_VS_MONOCYTE_BONE_MARROW_DN [200] 7 1.44 e-9 2.57 e-5 LINDVALL_IMMORTALIZED_BY_TERT_UP [78] 5 1.79 e-8 1.59 e-4 GSE6674_CPG_VS_PL2_3_STIM_BCELL_UP [200] 6 5.83 e-8 3.46 e-4 KIM_ALL_DISORDERS_OLIGODENDROCYTE_NUMBER_CORR_UP [756] 8 9.23 e-7 4.1 e-3 GO_GOLGI_APPARATUS [1445] 10 1.64 e-6 4.36 e-3 GSE10325_LUPUS_CD4_TCELL_VS_LUPUS_MYELOID_UP [200] 5 1.96 e-6 4.36 e-3 GSE29949_MICROGLIA_BRAIN_VS_CD8_POS_DC_SPLEEN_UP [200] 5 1.96 e-6 4.36 e-3 GSE39556_CD8A_DC_VS_NK_CELL_UP [200] 5 1.96 e-6 4.36 e-3 GO_ESTABLISHMENT_OF_LOCALIZATION_IN_CELL [1676] 10 6.11 e-6 1.21 e-2 JOHNSTONE_PARVB_TARGETS_2_DN [336] 5 2.42 e-5 3.29 e-2 Gene Set Name [# Genes (K)] Down # Genes in Overlap (k) p-value FDR q- value SPIELMAN_LYMPHOBLAST_EUROPEAN_VS_ASIAN_UP [479] 7 1.01 e-7 1.79 e-3 CAIRO_HEPATOBLASTOMA_CLASSES_UP [605] 7 4.84 e-7 4.3 e-3 RAO_BOUND_BY_SALL4_ISOFORM_B [517] 6 3.41 e-6 1.47 e-2 IVANOVA_HEMATOPOIESIS_LATE_PROGENITOR [544] 6 4.56 e-6 1.47 e-2 MCAATNNNNNGCG_UNKNOWN [545] 6 4.6 e-6 1.47 e-2 TIEN_INTESTINE_PROBIOTICS_24HR_UP [557] 6 5.21 e-6 1.47 e-2 GO_DYNEIN_COMPLEX [43] 3 5.8 e-6 1.47 e-2 GSE45365_NK_CELL_VS_CD8_TCELL_UP [154] 4 7.35 e-6 1.63 e-2 GO_CATALYTIC_COMPLEX [1038] 7 1.67 e-5 2.61 e-2 GSE20366_EX_VIVO_VS_DEC205_CONVERSION_NAIVE_CD4_TCELL_UP [200] 4 2.05 e-5 2.61 e-2 182 Table 4-12. Number of Differentially expressed genes Diff genes I- II- III- IV- V- VI- I+ 0 172 544 633 289 223 II+ 146 0 130 71 39 21 III+ 759 266 0 726 377 265 IV+ 396 67 294 0 107 80 V+ 275 73 241 211 0 54 VI+ 225 35 166 125 54 0 When + and - are compared, the number indicates the number of genes upregulated in + and downregulated in -. 183 Table 4-13. Number of up and downregulated potential biomarkers. Upregulated Downregulated Layer I 120 27 Layer II 19 0 Layer III 138 1 Layer IV 45 2 Layer V 47 0 Layer VI 33 0 Number of potential biomarkers found by combining the differential expression gene lists and counting the number of occurrences. 184 Table 4-14. GSEA on potential biomarkers on each layer. (Gene sets enriched for the both up and downregulated gene lists from each layer.) Set Name [# Genes (K)] Layer I Upregulated Overlapped p-value FDR q- value GO_POLY_A_RNA_BINDING [1170] 15 8.63 e-9 1.1 e-4 GO_RNA_BINDING [1598] 17 1.24 e-8 1.1 e-4 GO_CELLULAR_CARBOHYDRATE_BIOSYNTHETIC_PROCESS [51] 5 5.63 e-8 2.62 e-4 GO_ORGANONITROGEN_COMPOUND_METABOLIC_PROCESS [1796] 17 6.69 e-8 2.62 e-4 MILI_PSEUDOPODIA_HAPTOTAXIS_DN [668] 11 8.3 e-8 2.62 e-4 GO_ORGANONITROGEN_COMPOUND_BIOSYNTHETIC_PROCESS [1024] 13 1.04 e-7 2.62 e-4 GSE10239_NAIVE_VS_KLRG1INT_EFF_CD8_TCELL_DN [200] 7 1.46 e-7 2.62 e-4 GSE13484_12H_VS_3H_YF17D_VACCINE_STIM_PBMC_DN [200] 7 1.46 e-7 2.62 e-4 GSE3982_EOSINOPHIL_VS_TH1_DN [200] 7 1.46 e-7 2.62 e-4 GO_CARBOHYDRATE_BIOSYNTHETIC_PROCESS [121] 6 1.53 e-7 2.62 e-4 Gene Set Name [# Genes (K)] Layer I Downregulated Overlapped p-value FDR q- value KIM_ALL_DISORDERS_CALB1_CORR_UP [548] 11 4.12 e-15 7.33 e-11 MODULE_11 [540] 10 2.07 e-13 9.16 e-10 MODULE_100 [544] 10 2.23 e-13 9.16 e-10 MODULE_137 [546] 10 2.31 e-13 9.16 e-10 MODULE_66 [552] 10 2.58 e-13 9.16 e-10 MODULE_12 [360] 6 4.47 e-8 1.32 e-4 GNF2_AF1Q [25] 3 3.67 e-7 9.31 e-4 RCGCANGCGY_NRF1_Q6 [952] 7 7.47 e-7 1.66 e-3 GO_CELL_PART_MORPHOGENESIS [633] 6 1.21 e-6 2.16 e-3 RODRIGUES_THYROID_CARCINOMA_POORLY_DIFFERENTIATED_UP [633] 6 1.21 e-6 2.16 e-3 Gene Set Name [# Genes (K)] Layer I Combined Overlapped p-value FDR q- value KIM_ALL_DISORDERS_CALB1_CORR_UP [548] 15 9.31 e-12 1.65 e-7 MODULE_100 [544] 14 1.07 e-10 7.69 e-7 MODULE_66 [552] 14 1.3 e-10 7.69 e-7 DODD_NASOPHARYNGEAL_CARCINOMA_DN [1375] 20 2.26 e-10 9.69 e-7 BLALOCK_ALZHEIMERS_DISEASE_DN [1237] 19 2.72 e-10 9.69 e-7 YRTCANNRCGC_UNKNOWN [2940] 28 6.23 e-10 1.84 e-6 MODULE_11 [540] 13 1.16 e-9 2.94 e-6 MODULE_137 [546] 13 1.32 e-9 2.94 e-6 GSE13484_12H_VS_3H_YF17D_VACCINE_STIM_PBMC_DN [200] 9 2.21 e-9 4.36 e-6 GO_POSITIVE_REGULATION_OF_CATALYTIC_ACTIVITY [1518] 19 7.73 e-9 1.37 e-5 ene Set Name [# Genes (K)] Layer III Upregulated Overlapped p-value FDR q- value GO_CYTOSKELETAL_PROTEIN_BINDING [819] 12 2.65 e-7 3.02 e-3 GRAESSMANN_APOPTOSIS_BY_DOXORUBICIN_DN [1781] 17 3.4 e-7 3.02 e-3 ATCMNTCCGY_UNKNOWN [102] 5 3.26 e-6 1.04 e-2 GO_CELLULAR_COMPONENT_MORPHOGENESIS [900] 11 4.9 e-6 1.04 e-2 GARY_CD5_TARGETS_DN [431] 8 5.3 e-6 1.04 e-2 GO_HYDROLASE_ACTIVITY_ACTING_ON_ESTER_BONDS [739] 10 5.61 e-6 1.04 e-2 GSE20366_EX_VIVO_VS_HOMEOSTATIC_CONVERSION_TREG_UP [200] 6 5.66 e-6 1.04 e-2 185 GSE22601_IMMATURE_CD4_SINGLE_POSITIVE_VS_DOUBLE_POSITIVE_THYMOCYTE_DN [200] 6 5.66 e-6 1.04 e-2 TGGNNNNNNKCCAR_UNKNOWN [919] 11 5.97 e-6 1.04 e-2 GO_CYTOSKELETON [1967] 16 6.1 e-6 1.04 e-2 Gene Set Name [# Genes (K)] Layer IV Upregulated Overlapped p-value FDR q- value GO_ENZYME_BINDING [1737] 11 1.83 e-7 3.26 e-3 GSE39820_TGFBETA3_IL6_VS_TGFBETA3_IL6_IL23A_TREATED_CD4_TCELL_UP [200] 5 8.63 e-7 4.77 e-3 GSE3982_MAST_CELL_VS_CENT_MEMORY_CD4_TCELL_UP [200] 5 8.63 e-7 4.77 e-3 JOHNSTONE_PARVB_TARGETS_3_DN [918] 8 1.07 e-6 4.77 e-3 BENPORATH_NANOG_TARGETS [988] 8 1.85 e-6 5.14 e-3 SOX9_B1 [237] 5 1.99 e-6 5.14 e-3 ROME_INSULIN_TARGETS_IN_MUSCLE_UP [442] 6 2.23 e-6 5.14 e-3 DIAZ_CHRONIC_MEYLOGENOUS_LEUKEMIA_UP [1382] 9 2.31 e-6 5.14 e-3 NUYTTEN_NIPP1_TARGETS_UP [769] 7 4.12 e-6 7.83 e-3 GO_REGULATION_OF_CELL_PROLIFERATION [1496] 9 4.4 e-6 7.83 e-3 Gene Set Name [# Genes (K)] Layer V Upregulated Overlapped p-value FDR q- value GSE29949_MICROGLIA_BRAIN_VS_CD8_POS_DC_SPLEEN_UP [200] 5 7.58 e-7 1.35 e-2 KRIGE_RESPONSE_TO_TOSEDOSTAT_24HR_DN [1011] 8 1.79 e-6 1.59 e-2 GO_POSITIVE_REGULATION_OF_CELL_COMMUNICATION [1532] 9 4.26 e-6 2.17 e-2 GO_LIPID_METABOLIC_PROCESS [1158] 8 4.88 e-6 2.17 e-2 GO_CELLULAR_LIPID_METABOLIC_PROCESS [913] 7 1.05 e-5 3.75 e-2 GO_PHOSPHOLIPID_METABOLIC_PROCESS [364] 5 1.4 e-5 3.98 e-2 GSE19825_NAIVE_VS_DAY3_EFF_CD8_TCELL_DN [200] 4 2.54 e-5 3.98 e-2 GSE20727_H2O2_VS_ROS_INHIBITOR_TREATED_DC_DN [200] 4 2.54 e-5 3.98 e-2 GSE22025_TGFB1_VS_TGFB1_AND_PROGESTERONE_TREATED_CD4_TCELL_DN [200] 4 2.54 e-5 3.98 e-2 GSE26351_UNSTIM_VS_BMP_PATHWAY_STIM_HEMATOPOIETIC_PROGENITORS_UP [200] 4 2.54 e-5 3.98 e-2 Gene Set Name [# Genes (K)] Layer VI Upregulated Overlapped p-value FDR q- value GSE3982_DC_VS_NEUTROPHIL_UP [200] 4 1.6 e-6 1.41 e-2 GSE41176_UNSTIM_VS_ANTI_IGM_STIM_BCELL_6H_UP [200] 4 1.6 e-6 1.41 e-2 RAO_BOUND_BY_SALL4_ISOFORM_B [517] 5 2.38 e-6 1.41 e-2 GO_RNA_3_END_PROCESSING [98] 3 1.04 e-5 4.64 e-2 186 Table 4-15. Repetitive Adaptive VS Non-Adaptive differentially expressed genes. Gene baseMean log2FoldChange pvalue padj PTAFR 60.21588232 -6.447678787 2.90E-16 2.57E-12 CARF 15.77603036 -5.375137513 6.45E-11 1.66E-07 MATR3 94.08456025 -4.099789846 4.30E-11 1.66E-07 GABPB1- AS1 45.06894676 -3.516925873 7.51E-11 1.66E-07 C17orf75 27.85199224 -4.164661647 2.00E-09 3.55E-06 WIPF2 25.63507357 -4.904906647 3.78E-09 5.58E-06 RNF4 17.59484039 -5.438732936 1.09E-08 1.38E-05 ARIH1 38.26526512 -3.322899362 2.09E-08 2.32E-05 FAM135A 13.63717221 -4.187712928 3.66E-08 3.61E-05 TANC1 32.65271399 -5.214972359 6.11E-08 5.41E-05 TTC9C 17.61045249 -4.517925552 8.11E-08 6.53E-05 RUNDC3A 31.53644861 4.117511436 9.48E-08 7.01E-05 PPM1B 19.16385677 -3.361244275 1.05E-07 7.19E-05 KCNMB4 18.14907005 -4.694632372 1.38E-07 8.77E-05 IPO5 52.07097507 -3.830835644 3.03E-07 0.000179035 COQ5 27.74270751 -7.446351451 3.26E-07 0.000180716 MDM2 51.97269179 3.812711077 3.97E-07 0.000206911 NPTN 74.95327882 5.25662435 4.68E-07 0.000230628 WSCD2 13.87915981 -6.180955271 5.82E-07 0.00026077 DCN 20.22689506 -4.944061041 5.88E-07 0.00026077 STARD7 54.48304503 -4.412249768 9.57E-07 0.000371949 KIAA1257 16.13191334 -4.102305083 9.65E-07 0.000371949 LIMCH1 19.30897636 4.278286611 9.49E-07 0.000371949 C9orf106 2.942881467 -4.601940625 1.41E-06 0.000519988 C9orf72 27.9066346 -5.616651241 1.79E-06 0.000586154 LMBR1 7.345494589 -3.512313563 1.68E-06 0.000586154 ANK3 25.48675859 3.782773176 1.74E-06 0.000586154 HEATR1 3.837382017 -5.071663705 1.99E-06 0.000615851 TMEM170A 7.92314241 -3.131403024 2.15E-06 0.000615851 STMN1 126.4750639 -2.828671592 2.03E-06 0.000615851 187 Table 4-16. GSEA on Repetitive Adaptive VS Non-Adaptive differentially expressed genes. (Gene sets enriched for the both up and downregulated gene lists from comparison Adaptive vs. Non- adaptive neurons.) Gene Set Name [# Genes (K)] Up in Adaptive # Genes in Overlap (k) p-value FDR q- value BLALOCK_ALZHEIMERS_DISEASE_DN [1237] 49 7.25 e-23 1.29 e-18 PILON_KLF1_TARGETS_DN [1972] 58 1.09 e-20 9.68 e-17 SCHLOSSER_SERUM_RESPONSE_DN [712] 35 1.93 e-19 1.15 e-15 TTGCACT_MIR130A_MIR301_MIR130B [403] 27 1.39 e-18 6.19 e-15 TGGAAA_NFAT_Q4_01 [2061] 53 1.87 e-16 6.65 e-13 GO_NEURON_PART [1265] 41 3.4 e-16 1.01 e-12 GO_ESTABLISHMENT_OF_LOCALIZATION_IN_CELL [1676] 47 4.74 e-16 1.2 e-12 GOBERT_OLIGODENDROCYTE_DIFFERENTIATION_DN [1080] 37 1.92 e-15 4.27 e-12 AAAYWAACM_HFH4_01 [1890] 49 2.19 e-15 4.32 e-12 GO_CELL_PROJECTION [1786] 47 5.06 e-15 9 e-12 Gene Set Name [# Genes (K)] # Genes in Overlap (k) p-value FDR q- value PILON_KLF1_TARGETS_DN [1972] 36 6.89 e-14 1.23 e-9 DIAZ_CHRONIC_MEYLOGENOUS_LEUKEMIA_UP [1382] 26 1.57 e-10 1.16 e-6 LASTOWSKA_NEUROBLASTOMA_COPY_NUMBER_DN [800] 20 1.95 e-10 1.16 e-6 YRTCANNRCGC_UNKNOWN [2940] 37 1.35 e-9 4.85 e-6 WWTAAGGC_UNKNOWN [1896] 29 1.5 e-9 4.85 e-6 GRAESSMANN_APOPTOSIS_BY_DOXORUBICIN_DN [1781] 28 1.64 e-9 4.85 e-6 ZHANG_BREAST_CANCER_PROGENITORS_UP [425] 14 3.72 e-9 9.44 e-6 GRAESSMANN_RESPONSE_TO_MC_AND_DOXORUBICIN_DN [770] 18 4.69 e-9 1.04 e-5 WEI_MYCN_TARGETS_WITH_E_BOX [795] 18 7.66 e-9 1.51 e-5 SCHLOSSER_SERUM_RESPONSE_DN [712] 17 9.4 e-9 1.63 e-5 188 Summary & Conclusion The PAIA technique can be applied to the gene expression analysis of single-cells in various fields. Although we have given three examples of these applications, an unlimited number of other applications exist which could give rise to revolutionary scientific achievements. Three major areas of interest within the neuroscience field were this technique could be useful are developmental studies, circadian rhythm studies, and pathway analysis. In Chapter 2, by using the PAIA on embryonic brain neurons of humans, we showed that one of applications of PAIA could expand the understanding of embryonic development. However, Cajal-Retzius neurons are only one part of the story in brain development. Most previous developmental studies solely depend on the analysis of cell populations in groups. For example, the effect of neurotransmitters such as catecholamine during neuronal migration and differentiation was studied in groups of neurons where they showed many different transcriptional changes in interneurons, which migrate and generate different layers of the neocortex. (Herlenius & Lagercrantz, 2001). In embryonic stage 16-17, norepinephrine is known to alter the speed of migration (Riccio et al., 2012) and post-migrational differentiation of interneurons (Hevner et al., 2001). However, the underling mechanisms remain unknown. 189 The inherent limitation of studying a group of neurons could be resolved using our PAIA technique. In each stage of development (E16, 17, 18, and after), the migrating interneurons can be collected with a patch-clamp pipette and sequenced in both wild type and norepinephrine knockout mice, such as DBH KO (Thomas, Matsumoto, & Palmiter, 1995). The comparison between these groups in gene expression profile would allow us determine the effect of norepinephrine on genes from fetal interneurons at different developmental stages. Based on the change in gene expression, such as migration and differentiation specific markers (e.g. reeler, bag1) (Caviness & Sidman, 1973; Kermer et al., 2002), a mechanism of neuronal migration and differentiation patterning at each embryonic stage could be identified. Another possible application of the PAIA is for detecting new cell types and potential biomarkers. In many studies, cell types were only defined by physiological characterization such as firing pattern and spike threshold. The gene expression data of these cells could allow the detection of the different molecular pathways affected by the electrophysiological characteristics. The analysis could also reveal specific biomarkers of a given cell type. Alternatively, PAIA could be applied to circadian rhythm studies. The melanopsin ganglion cells, named after their photoreceptive pigment, alternate the circadian rhythm by the perception of light. A large number of melanopsin ganglion cells’ axons are mostly projecting to the superchiasmic nucleus (SCN), which serves as the central clock of the brain. However, several 190 groups of melanopsin ganglion cells have been shown immunihistochemically to be connected to different targets (Schmidt et al., 2011). These cell types were vaguely defined based on their morphology and electrophysiology. The PAIA application could provide a gene expression profile of these cell types. By finding differently expressed gene in these ganglion neurons, the biomarker of each cell type could be determined. In Chapter 3, we detailed how our PAIA technique could be used to determine the cell types and biomarkers of embryonic spinal motor neurons. Some studies indicate that there are differences in the origins of the subgroups of these neurons (Doetsch & Alvarez-Buylla, 1996). Moreover, there are indications that the electrophysiological patterns of these neurons were different from others. The PAIA application could detect the biomarkers of these groups of motor neurons. Similarly, in Chapter 4, we have shown some specific biomarkers related to the electrophysiological patterns in the adult brain. Furthermore, the PAIA technique could be used to study pathways of biological processes such as the unfolded protein response (UPR). The UPR is a specific reaction induced by stress in the endoplasmic reticulum (ER). The primary action of UPR is the removal of the stress by inducing chaperon activities, which expedite protein folding, or by increasing ER capacity. Eventually, UPR induces cell death (Chakrabarti, et al., 2011). Though there are many studies on how UPR affects the general ER stress reduction, the 191 effects of UPR on neuronal function prior to apoptosis remain unknown. Prior to apoptosis, UPR affects multiple biological functions to reduce ER stress. For example, reduction of calcium deposits in ER occurs by releasing calcium from the ER (Ron & Walter, 2007). In this case, neuronal activities of signal transmission, which are heavily dependent on cytoplasmic calcium, should be affected. Moreover, calcium can alter the gene expression of transcription factors such as calcium-calmodulin (CAM) kinase. To detect both calcium- mediated physiological changes and the expression level changes, the single cell RNA-seq combined with voltage clamp recording can be used. For example, in retinal ganglion cells (RGCs), PAIA could be applied to measure both expression pattern and the electrophysiological recording. When there is UPR with changes in calcium levels, changes in gene expression levels of knockout mice and wild type mice can be measured. This comparison may inform that what molecules and what neuronal activities were affected by UPR. This pathway analysis study could be applied in other studies such as epileptic temporal lobe studies. In conclusion, the Patch-aRNA protocol allows us to learn about human transcriptome profile from an intact tissue without dissociation. It also allows us to possibly get the phenotypic information about the cell, including morphological properties of the cell and physiological recording. With PAIA application, understanding in many fields of neuroscience could be expanded. In this case, neuronal activities of signal transmission, which are heavily 192 dependent on cytoplasmic calcium, should be affected. Moreover, calcium can alter the gene expression of transcription factors such as calcium-calmodulin (CAM) kinase. To detect both calcium-mediated physiological changes and the expression level changes, single cell RNA-seq combined with voltage clamp recording can be used. For example, in retinal ganglion cells (RGCs), PAIA could be used to measure both expression pattern and record the electrophysiology. When there is UPR with changes in calcium levels, changes in gene expression levels of knockout mice and wild type mice can be measured. This comparison may inform us what molecules and what neuronal activities were affected by UPR. This pathway analysis study could also be applied in other areas such as epileptic temporal lobe studies. In conclusion, the Patch-aRNA protocol allows us to learn about human transcriptome profile from an intact tissue without dissociation. It also allows us to possibly get the phenotypic information about the cell, including morphological properties of the cell and physiological recording. With the use of PAIA, understanding in many fields of neuroscience could be expanded. 193 References Achilles, K., Okabe, A., Ikeda, M., Shimizu-Okabe, C., Yamada, J., Fukuda, A., … Kilb, W. (2007). Kinetic properties of Cl uptake mediated by Na+- dependent K+-2Cl cotransport in immature rat neocortical neurons. The Journal of Neuroscience : The Official Journal of the Society for Neuroscience, 27(32), 8616–8627. https://doi.org/10.1523/JNEUROSCI.5041-06.2007 Adams, J. C., & Lawler, J. (2011). The thrombospondins. Cold Spring Harb Perspect Biol, 3(10), a009712. https://doi.org/10.1101/cshperspect.a009712 Aird, D., Ross, M. G., Chen, W.-S., Danielsson, M., Fennell, T., Russ, C., … Gnirke, A. (2011). Analyzing and minimizing PCR amplification bias in Illumina sequencing libraries. Genome Biology, 12(R18). https://doi.org/10.1186/gb-2011-12-2-r18 Albrieux, M. (2004). Early Expression of Sodium Channel Transcripts and Sodium Current by Cajal-Retzius Cells in the Preplate of the Embryonic Mouse Neocortex. Journal of Neuroscience, 24(7), 1719–1725. https://doi.org/10.1523/jneurosci.3548-02.2004 194 Anstötz, M., Cosgrove, K. E., Hack, I., Mugnaini, E., Maccaferri, G., & Lübke, J. H. R. (2014). Morphology, input–output relations and synaptic connectivity of Cajal–Retzius cells in layer 1 of the developing neocortex of CXCR4-EGFP mice. Brain Structure and Function, 219(6), 2119–2139. https://doi.org/10.1007/s00429-013-0627-2 Arber, S., & Caroni, P. (1995). Thrombospondin-4, an extracellular matrix protein expressed in the developing and adult nervous system promotes neurite outgrowth. J.Cell Biol., 131(4), 1083–1094. Arendt, T., Stieler, J., & Ueberham, U. (2017). Is sporadic Alzheimer′s disease a developmental disorder? Journal of Neurochemistry. https://doi.org/10.1111/jnc.14036 Arnold, S. E., Hyman, B. T., Flory, J., Damasio, A. R., & Van Hoesen, G. W. (1991). The topographical and neuroanatomical distribution of neurofibrillary tangles and neuritic plaques in the cerebral cortex of patients with Alzheimer’s disease. Cerebral Cortex (New York, N.Y. : 1991), 1(1), 103–16. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/1822725 Belgard, T. G., Marques, A. C., Oliver, P. L., Abaan, H. O., Sirey, T. M., Hoerder-Suabedissen, A., … Ponting, C. P. (2011). A Transcriptomic Atlas of Mouse Neocortical Layers. Neuron, 71(4), 605–616. https://doi.org/10.1016/j.neuron.2011.06.039 195 Blake, S. M., Strasser, V., Andrade, N., Duit, S., Hofbauer, R., Schneider, W. J., & Nimpf, J. (2008). Thrombospondin-1 binds to ApoER2 and VLDL receptor and functions in postnatal neuronal migration. The EMBO Journal, 27(22), 3069–80. https://doi.org/10.1038/emboj.2008.223 Bodansky, A., Kim, J. M. H., Tempest, L., Velagapudi, A., Libby, R., & Ravits, J. (2010). TDP-43 and ubiquitinated cytoplasmic aggregates in sporadic ALS are low frequency and widely distributed in the lower motor neuron columns independent of disease spread. Amyotrophic Lateral Sclerosis : Official Publication of the World Federation of Neurology Research Group on Motor Neuron Diseases, 11(3), 321–327. Brennecke, P., Anders, S., Kim, J. K., Kołodziejczyk, A. A., Zhang, X., Proserpio, V., … Heisler, M. G. (2013). Accounting for technical noise in single-cell RNA-seq experiments. Nature Methods, 10(11), 1093–1095. https://doi.org/10.1038/nmeth.2645 Brezina, V., Orekhova, I. V, & Weiss, K. R. (2000). The neuromuscular transform: the dynamic, nonlinear link between motor neuron firing patterns and muscle contraction in rhythmic behaviors. Journal of Neurophysiology, 83(1), 207–31. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/10634868 196 Burgoyne, R. D., & Haynes, L. P. (2010). Neuronal calcium sensor proteins: emerging roles in membrane traffic and synaptic plasticity. F1000 Biology Reports, 2(January), 1–5. https://doi.org/10.3410/B2-5 Cadwell, C. R., Palasantza, A., Jiang, X., Berens, P., Deng, Q., Yilmaz, M., … Tolias, A. S. (2015). Electrophysiological, transcriptomic and morphologic profiling of single neurons using Patch-seq. Nature Biotechnology, 34(December), 1–8. https://doi.org/10.1038/nbt.3445 Carmon, K. S., Gong, X., Lin, Q., Thomas, A., & Liu, Q. (2011). R-spondins function as ligands of the orphan receptors LGR4 and LGR5 to regulate Wnt/beta-catenin signaling. Proceedings of the National Academy of Sciences of the United States of America, 108(28), 11452–7. https://doi.org/10.1073/pnas.1106083108 Caspary, T., & Anderson, K. V. (2003). Patterning cell types in the dorsal spinal cord: what the mouse mutants say. Nature Reviews Neuroscience, 4(4), 289–297. https://doi.org/10.1038/nrn1073 Caviness, V. S., & Sidman, R. L. (1973). Time of origin of corresponding cell classes in the cerebral cortex of normal and reeler mutant mice: An autoradiographic analysis. The Journal of Comparative Neurology, 148(2), 141–151. https://doi.org/10.1002/cne.901480202 197 Chameau, P., Inta, D., Vitalis, T., Monyer, H., Wadman, W. J., & van Hooft, J. a. (2009). The N-terminal region of reelin regulates postnatal dendritic maturation of cortical pyramidal neurons. Proceedings of the National Academy of Sciences of the United States of America, 106(17), 7227– 7232. https://doi.org/10.1073/pnas.0810764106 Chan, C.-H., & Yeh, H. H. (2003). Enhanced GABAA receptor-mediated activity following activation of NMDA receptors in Cajal-Retzius cells in the developing mouse neocortex. The Journal of Physiology, 550(Pt 1), 103– 111. https://doi.org/10.1113/jphysiol.2003.042556 Chen, J.-G., Rasin, M.-R., Kwan, K. Y., & Sestan, N. (2005). Zfp312 is required for subcortical axonal projections and dendritic morphology of deep-layer pyramidal neurons of the cerebral cortex. Proceedings of the National Academy of Sciences of the United States of America, 102(49), 17792–7. https://doi.org/10.1073/pnas.0509032102 Chen, Y., Hor, H. H., & Tang, B. L. (2012). AMIGO is expressed in multiple brain cell types and may regulate dendritic growth and neuronal survival. Journal of Cellular Physiology, 227(5), 2217–2229. https://doi.org/10.1002/jcp.22958 198 Chen, Y., Souaiaia, T., & Chen, T. (2009). PerM: efficient mapping of short sequencing reads with periodic full sensitive spaced seeds. Bioinformatics, 25(19), 2514–2521. https://doi.org/10.1093/bioinformatics/btp486 Chowdhury. (2010). Fate of Cajal-Retzius neurons in the postnatal mouse neocortex. Frontiers in Neuroanatomy, 4(March), 1–8. https://doi.org/10.3389/neuro.05.010.2010 Christophe, E., Doerflinger, N., Lavery, D. J., Molnár, Z., Charpak, S., & Audinat, E. (2005). Two Populations of Layer V Pyramidal Cells of the Mouse Neocortex: Development and Sensitivity to Anesthetics. Journal of Neurophysiology, 94(5). Retrieved from http://jn.physiology.org/content/94/5/3357.short Christopher A. Mutch, Nobuo Funatsu, Edwin S. Monuki, and A. C. (2009). Laminar, Beta-catenin signaling levels in progenitors influence the cell fates of projection neurons. The Journal of Neuroscience, 4(164), 13710– 13719. https://doi.org/10.1126/scisignal.2001449.Engineering Christopherson, K. S., Ullian, E. M., Stokes, C. C. A., Mullowney, C. E., Hell, J. W., Agah, A., … Barres, B. A. (2005). Thrombospondins are astrocyte- secreted proteins that promote CNS synaptogenesis. Cell, 120(3), 421– 433. https://doi.org/10.1016/j.cell.2004.12.020 199 Citri, A., Pang, Z. P., Südhof, T. C., Wernig, M., & Malenka, R. C. (2011). Comprehensive qPCR profiling of gene expression in single neuronal cells. Nature Protocols, 7(1), 118–127. https://doi.org/10.1038/nprot.2011.430 Costa, L. da F., Campos, A. G., Estrozi, L. F., Rios-Filho, L. G., & Bosco, A. (2000). A Biologically-Motivated Approach to Image Representation and Its Application to Neuromorphology (pp. 407–416). Springer, Berlin, Heidelberg. https://doi.org/10.1007/3-540-45482-9_41 Costa, L. da F., Manoel, E. T. M., Faucereau, F., Chelly, J., van Pelt, J., & Ramakers, G. (2002). A shape analysis framework for neuromorphometry. Network: Computation in Neural Systems, 13(3), 283–310. https://doi.org/10.1088/0954-898X_13_3_303 D’Arcangelo, G. (1995). A protein related to extracellular matrix proteins deleted in the mouse mutant reeler. Nature, 374, 719–723. D’Arcangelo, G. (2014). Reelin in the Years: Controlling Neuronal Migration and Maturation in the Mammalian Brain. Advances in Neuroscience, 2014, 1–19. https://doi.org/10.1155/2014/597395 D’Souza, B., Miyamoto, A., & Weinmaster, G. (2008). The many facets of Notch ligands. Oncogene, 27(38), 5148–5167. https://doi.org/10.1038/onc.2008.229 200 del Río, J. a, Martínez, A., Fonseca, M., Auladell, C., & Soriano, E. (1995). Glutamate-like immunoreactivity and fate of Cajal-Retzius cells in the murine cortex as identified with calretinin antibody. Cerebral Cortex, 5(1), 13–21. https://doi.org/10.1093/cercor/5.1.13 Derer, P., & Derer, M. (1990). Cajal-Retzius cell ontogenesis and death in mouse brain visualized with horseradish peroxidase and electron microscopy. Neuroscience, 36(3), 839–56. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/2234416 Doetsch, F., & Alvarez-Buylla, A. (1996). Network of tangential pathways for neuronal migration in adult mammalian brain. Neurobiology Communicated by Fernando Nottebohm, 93, 14895–14900. Retrieved from http://www.pnas.org/content/93/25/14895.full.pdf Draganova, K., Zemke, M., Zurkirchen, L., Valenta, T., Cantù, C., Okoniewski, M., … Sommer, L. (2015). Wnt/β-Catenin Signaling Regulates Sequential Fate Decisions of Murine Cortical Precursor Cells. STEM CELLS, 33(1), 170–182. https://doi.org/10.1002/stem.1820 Dugas, J. C., Tai, Y. C., Speed, T. P., Ngai, J., & Barres, B. A. (2006). Functional Genomic Analysis of Oligodendrocyte Differentiation. Journal of Neuroscience, 26(43). Retrieved from http://www.jneurosci.org/content/26/43/10967.short 201 Dvorzhak, A., Unichenko, P., & Kirischuk, S. (2012). Glutamate transporters and presynaptic metabotropic glutamate receptors protect neocortical Cajal-Retzius cells against over-excitation. Pflugers Archiv European Journal of Physiology, 464(2), 217–225. https://doi.org/10.1007/s00424- 012-1109-8 Eberwine, J., Yeh, H., Miyashiro, K., Cao, Y., Nair, S., Finnell, R., … Coleman, P. (1992). Analysis of gene expression in single live neurons. Neurobiology, 89, 3010–3014. Retrieved from https://www.ncbi.nlm.nih.gov/pmc/articles/PMC48793/pdf/pnas01081- 0511.pdf Flagstad, Røed, Stacy, & Jakobsen. (1999). Reliable noninvasive genotyping based on excremental PCR of nuclear DNA purified with a magnetic bead protocol. Molecular Ecology, 8(5), 879–883. https://doi.org/10.1046/j.1365-294X.1999.00623.x Fox, M. W., & Inman, O. (1966). Persistence of Retzius-Cajal cells in developing dog brain. Brain Research, 3(2), 192–4. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/4165894 202 Franceschetti, S., Sancini, G., Panzica, F., Radici, C., & Avanzini, G. (1998). Postnatal differentiation of firing properties and morphological characteristics in layer V pyramidal neurons of the sensorimotor cortex. Neuroscience, 83(4), 1013–1024. https://doi.org/10.1016/S0306- 4522(97)00463-6 Gans, C., & Northcutt, R. G. (1983). Neural Crest and the Origin of Vertebrates: A New Head. Science, 220(4594). Retrieved from http://science.sciencemag.org/content/220/4594/268 Gao, W.-J., & Zheng, Z.-H. (2004). Target-specific differences in somatodendritic morphology of layer V pyramidal neurons in rat motor cortex. The Journal of Comparative Neurology, 476(2), 174–185. https://doi.org/10.1002/cne.20224 Gentleman, R. C., Carey, V. J., Bates, D. M., Bolstad, B., Dettling, M., Dudoit, S., … Kim, J. (2004). Assessing characteristics of RNA amplification methods for single cell RNA sequencing. Genome Biology, 5(10), R80. https://doi.org/10.1186/gb-2004-5-10-r80 Gil-Sanz, C., Franco, S. J., Martinez-Garay, I., Espinosa, A., Harkins-Perry, S., & Müller, U. (2013). Cajal-Retzius cells instruct neuronal migration by coincidence signaling between secreted and contact-dependent guidance cues. Neuron, 79(3), 461–477. https://doi.org/10.1016/j.neuron.2013.06.040 203 Gil, V., Nocentini, S., & Del Río, J. a. (2014). Historical first descriptions of Cajal-Retzius cells: from pioneer studies to current knowledge. Frontiers in Neuroanatomy, 8(MAY), 32. https://doi.org/10.3389/fnana.2014.00032 Gilbert, C. D., & Kelly, J. P. (1975). The projections of cells in different layers of the cat’s visual cortex. The Journal of Comparative Neurology, 163(1), 81–105. https://doi.org/10.1002/cne.901630106 Glinka, A., Dolde, C., Kirsch, N., Huang, Y.-L., Kazanskaya, O., Ingelfinger, D., … Niehrs, C. (2011). LGR4 and LGR5 are R-spondin receptors mediating Wnt/β-catenin and Wnt/PCP signalling. EMBO Reports, 12(10), 1055– 1061. https://doi.org/10.1038/embor.2011.175 Gong, X., Carmon, K. S., Lin, Q., Thomas, A., Yi, J., & Liu, Q. (2012). LGR6 is a high affinity receptor of R-Spondins and potentially functions as a tumor suppressor. PLoS ONE, 7(5). https://doi.org/10.1371/journal.pone.0037137 Hafner, G. J., Yang, I. C., Wolter, L. C., Stafford, M. R., & Giffard, P. M. (2001). Isothermal Amplification and Multimerization of DNA by Bst DNA Polymerase. BioTechniques , 30, 852–867. Retrieved from https://www.researchgate.net/profile/Greg_Hafner/publication/12021080_I sothermal_amplification_and_multimerization_of_DNA_by_Bst_DNA_poly merase/links/00b7d5237d270cf68f000000.pdf 204 Handley, M. T. W., Lian, L. Y., Haynes, L. P., & Burgoyne, R. D. (2010). Structural and functional deficits in a neuronal calcium sensor-1 mutant identified in a case of autistic spectrum disorder. PLoS ONE, 5(5). https://doi.org/10.1371/journal.pone.0010534 Hartsock, A., & Nelson, W. J. (2008). Adherens and tight junctions: Structure, function and connections to the actin cytoskeleton. Biochimica et Biophysica Acta - Biomembranes, 1778(3), 660–669. https://doi.org/10.1016/j.bbamem.2007.07.012 Hashimshony, T., Wagner, F., Sher, N., & Yanai, I. (2012). CEL-Seq: Single- Cell RNA-Seq by Multiplexed Linear Amplification. Cell Reports, 2(3), 666–673. https://doi.org/10.1016/j.celrep.2012.08.003 Herlenius, E., & Lagercrantz, H. (2001). Neurotransmitters and neuromodulators during early human development. Early Human Development, 65(1), 21–37. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/11520626 Hestrin, S., & Armstrong, W. E. (1996). Morphology and physiology of cortical neurons in layer I. The Journal of Neuroscience : The Official Journal of the Society for Neuroscience, 16(17), 5290–5300. 205 Hevner, R. F., Shi, L., Justice, N., Hsueh, Y.-P., Sheng, M., Smiga, S., … Rubenstein, J. L. . (2001). Tbr1 Regulates Differentiation of the Preplate and Layer 6. Neuron, 29(2), 353–366. https://doi.org/10.1016/S0896- 6273(01)00211-2 Hirano, S., & Takeichi, M. (2012). Cadherins in Brain Morphogenesis and Wiring. Physiological Reviews, 92(2), 597–634. https://doi.org/10.1152/physrev.00014.2011 Hodgkin, A. L., & Huxley, A. F. (1952). A quantitative description of membrane current and its application to conduction and excitation in nerve. The Journal of Physiology, 117(4), 500–44. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/12991237 Hoheisel, J. D. (2006). Microarray technology: beyond transcript profiling and genotype analysis. Nature Reviews Microbiology, 7(3), 200–210. https://doi.org/10.1038/nrg1809 Howell, B. W., Hawkes, R., Soriano, P., & Cooper, J. A. (1997). Neuronal position in the developing brain is regulated by mouse disabled-1. Nature, 389(6652), 733–7. https://doi.org/10.1038/39607 Hua, J. Y., & Smith, S. J. (2004). Neural activity and the dynamics of central nervous system development. Nature Neuroscience, 7(4), 327–332. https://doi.org/10.1038/nn1218 206 Jaitin, D. A., Kenigsberg, E., Keren-Shaul, H., Elefant, N., Paul, F., Zaretsky, I., … Amit, I. (2014). Massively Parallel Single-Cell RNA-Seq for Marker- Free Decomposition of Tissues into Cell Types. Science, 343(6172). Retrieved from http://science.sciencemag.org/content/343/6172/776 Janusonis, S. (2004). Early Serotonergic Projections to Cajal-Retzius Cells: Relevance for Cortical Development. Journal of Neuroscience, 24(7), 1652–1659. https://doi.org/10.1523/JNEUROSCI.4651-03.2004 Jessell, T. M. (2000). Neuronal specification in the spinal cord: inductive signals and transcriptional codes. Nature Reviews Genetics, 1(1), 20–29. https://doi.org/10.1038/35049541 Johnson, K. R., Gagnon, L. H., & Chang, B. (2016). A hypomorphic mutation of the gamma-1 adaptin gene (Ap1g1) causes inner ear, retina, thyroid, and testes abnormalities in mice. Mammalian Genome, 27(5–6), 200– 212. https://doi.org/10.1007/s00335-016-9632-0 Judaš, M., & Pletikos, M. (2010). The discovery of the subpial granular layer in the human cerebral cortex. Translational Neuroscience, 1(3), 255–260. https://doi.org/10.2478/v10134-010-0037-4 Junge, H. J., Yung, A. R., Goodrich, L. V, & Chen, Z. (2016). Netrin1/DCC signaling promotes neuronal migration in the dorsal spinal cord. Developmental Biology, 11(19). https://doi.org/10.1186/s13064-016-0074- x 207 Kahle, P. J., Waak, J., & Gasser, T. (2009). DJ-1 and prevention of oxidative stress in Parkinson’s disease and other age-related disorders. Free Radical Biology and Medicine, 47(10), 1354–1361. https://doi.org/10.1016/j.freeradbiomed.2009.08.003 Kennedy, T. E., Serafini, T., de la Torre, J., & Tessier-Lavigne, M. (1994). Netrins are diffusible chemotropic factors for commissural axons in the embryonic spinal cord. Cell, 78(3), 425–435. https://doi.org/10.1016/0092- 8674(94)90421-9 Kermer, P., Krajewska, M., Zapata, J. M., Takayama, S., Mai, J., Krajewski, S., & Reed, J. C. (2002). Bag1 is a regulator and marker of neuronal differentiation. Cell Death and Differentiation, 9(4), 405–413. https://doi.org/10.1038/sj.cdd.4400972 Khazaei, M. R., Halfter, H., Karimzadeh, F., Koo, J. H., Margolis, F. L., & Young, P. (2010). Bex1 is involved in the regeneration of axons after injury. Journal of Neurochemistry, 115(4), 910–920. https://doi.org/10.1111/j.1471-4159.2010.06960.x Kilb, W., Ikeda, M., Uchida, K., Okabe, A., Fukuda, A., & Luhmann, H. J. (2002). Depolarizing glycine responses in Cajal-Retzius cells of neonatal rat cerebral cortex. Neuroscience, 112(2), 299–307. https://doi.org/10.1016/S0306-4522(02)00071-4 208 Kilb, W., & Luhmann, H. J. (2001). Spontaneous GABAergic postsynaptic currents in Cajal-Retzius cells in neonatal rat cerebral cortex. Eur J Neurosci, 13(7), 1387–1390. https://doi.org/ejn1514 [pii] Kim, H. G., Fox, K., & Connors, B. W. (n.d.). Properties of excitatory synaptic events in neurons of primary somatosensory cortex of neonatal rats. Cerebral Cortex (New York, N.Y. : 1991), 5(2), 148–57. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/7620291 Kim, M., Fontelonga, T., Roesener, A. P., Lee, H., Gurung, S., Mendonca, P. R. F., & Mastick, G. S. (2015). Motor neuron cell bodies are actively positioned by Slit/Robo repulsion and Netrin/DCC attraction. Developmental Biology, 399(1), 68–79. https://doi.org/10.1016/j.ydbio.2014.12.014 Kirischuk, S., Luhmann, H. J., & Kilb, W. (2014). Cajal-Retzius cells: Update on structural and functional properties of these mystic neurons that bridged the 20th century. Neuroscience, 275, 33–46. https://doi.org/10.1016/j.neuroscience.2014.06.009 Klein, A. M., Mazutis, L., Akartuna, I., Tallapragada, N., Veres, A., Li, V., … Kirschner, M. W. (2015). Droplet Barcoding for Single-Cell Transcriptomics Applied to Embryonic Stem Cells. Cell, 161(5), 1187– 1201. https://doi.org/10.1016/j.cell.2015.04.044 209 Komuro, H., & Kumada, T. (2005). Ca2+ transients control CNS neuronal migration. Cell Calcium, 37(5 SPEC. ISS.), 387–393. https://doi.org/10.1016/j.ceca.2005.01.006 Koriyama, Y., Furukawa, A., Muramatsu, M., Takino, J., & Takeuchi, M. (2015). Glyceraldehyde caused Alzheimer’s disease-like alterations in diagnostic marker levels in SH-SY5Y human neuroblastoma cells. Scientific Reports, 5(1), 13313. https://doi.org/10.1038/srep13313 Kowall, N. W., & Kosik, K. S. (1987). Axonal disruption and aberrant localization of tau protein characterize the neuropil pathology of Alzheimer’s disease. Annals of Neurology, 22(5), 639–643. https://doi.org/10.1002/ana.410220514 Koyama, A., Sugai, A., Kato, T., Ishihara, T., Shiga, A., Toyoshima, Y., … Onodera, O. (2016). Increased cytoplasmic TARDBP mRNA in affected spinal motor neurons in ALS caused by abnormal autoregulation of TDP- 43. Nucleic Acids Research, 44(12), 5820–5836. https://doi.org/10.1093/nar/gkw499 Kuja-Panula, J., Kiiltom??ki, M., Yamashiro, T., Rouhiainen, A., & Rauvala, H. (2003). AMIGO, a transmembrane protein implicated in axon tract development, defines a novel protein family with leucine-rich repeats. Journal of Cell Biology, 160(6), 963–973. https://doi.org/10.1083/jcb.200209074 210 Kwan, K. Y., Sestan, N., & Anton, E. S. (2012). Transcriptional co-regulation of neuronal migration and laminar identity in the neocortex. Development, 139(9), 1535–1546. https://doi.org/10.1242/dev.069963 Leberl, S. M., & Sane&, J. R. (1995). Migratory Paths of Neurons and Glia in the Embryonic Chick Spinal Cord. The Journal of Neuroscience, 75(2), 1236–246. Retrieved from http://www.jneurosci.org/content/jneuro/15/2/1236.full.pdf Levitt, J. B., Lewis, D. A., Yoshioka, T., & Lund, J. S. (1993). Topography of pyramidal neuron intrinsic connections in macaque monkey prefrontal cortex (areas 9 and 46). The Journal of Comparative Neurology, 338(3), 360–376. https://doi.org/10.1002/cne.903380304 Liauw, J., Hoang, S., Choi, M. M., Eroglu, C., Choi, M. M., Sun, G., … Steinberg, G. K. (2008). Thrombospondins 1 and 2 are necessary for synaptic plasticity and functional recovery after stroke. Journal of Cerebral Blood Flow and Metabolism : Official Journal of the International Society of Cerebral Blood Flow and Metabolism, 28(10), 1722–1732. https://doi.org/10.1038/jcbfm.2008.65 211 Liu, C. L., Schreiber, S. L., & Bernstein, B. E. (2003). Development and validation of a T7 based linear amplification for genomic DNA. BMC Genomics, 4(19). Lu, S. M., Zecevic, N., & Yeh, H. H. (2001). Distinct NMDA and AMPA receptor-mediated responses in mouse and human Cajal-Retzius cells. Journal of Neurophysiology, 86(5), 2642–6. Luhmann, H. J., Fukuda, A., & Kilb, W. (2015). Control of cortical neuronal migration by glutamate and GABA. Frontiers in Cellular Neuroscience, 9(January), 4. https://doi.org/10.3389/fncel.2015.00004 Luhmann, H. J., Reiprich, R. A., Hanganu, I. L., & Kilb, W. (2000). Cellular physiology of the neonatal rat cerebral cortex: Intrinsic membrane properties, sodium and calcium currents. Journal of Neuroscience Research, 62(July), 574–584. Lum, L. S., Sultzman, L. A., Kaufman, R. J., Linzer, D. I., & Wu, B. J. (1990). A cloned human CCAAT-box-binding factor stimulates transcription from the human hsp70 promoter. Molecular and Cellular Biology, 10(12), 6709–17. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/2247079 212 Manent, J.-B., & Represa, A. (2007). Neurotransmitters and brain maturation: early paracrine actions of GABA and glutamate modulate neuronal migration. The Neuroscientist : A Review Journal Bringing Neurobiology, Neurology and Psychiatry, 13(3), 268–279. https://doi.org/10.1177/1073858406298918 Marín-Padilla, M. (1998). Cajal–Retzius cells and the development of the neocortex. Trends in Neurosciences, 21(2), 64–71. https://doi.org/10.1016/S0166-2236(97)01164-8 MartÃnez-Cerdeno, V., & Noctor, S. C. (2014). Cajal, Retzius, and Cajal- Retzius cells. Frontiers in Neuroanatomy, 8(June), 48. https://doi.org/10.3389/fnana.2014.00048 McKinney, R. a, Capogna, M., Dürr, R., Gähwiler, B. H., & Thompson, S. M. (1999). Miniature synaptic events maintain dendritic spines via AMPA receptor activation. Nature Neuroscience, 2(1), 44–49. https://doi.org/10.1038/4548 Meyer, G., Goffinet, A., Fairén, a, & Namur, B. (1999). What is a Cajal – Retzius Cell ? A Reassessment of a Classical Cell Type Based on Recent Observations in the Developing Neocortex. Cerebral Cortex, 9(8), 765– 775. https://doi.org/10.1093/cercor/9.8.765 213 Meyer, G., Schaaps, J. P., Moreau, L., & Goffinet, A. M. (2000). Embryonic and Early Fetal Development of the Human Neocortex. The Journal of Neuroscience, 20(5), 1858–1868. Mienville, J. M. (1998). Persistent depolarizing action of GABA in rat Cajal- Retzius cells. Journal of Physiology, (512.3), 809–817. Mienville, J. M., Maric, I., Maric, D., & Clay, J. R. (1999). Loss of IA expression and increased excitability in postnatal rat Cajal-Retzius cells. Journal of Neurophysiology, 82(3), 1303–1310. Miller, J. A., Ding, S.-L., Sunkin, S. M., Smith, K. A., Ng, L., Szafer, A., … Lein, E. S. (2014). Transcriptional landscape of the prenatal human brain. Nature, 508(7495), 199–206. https://doi.org/10.1038/nature13185 Moll, P. R., Duschl, J., & Richter, K. (2004). Optimized RNA amplification using T7-RNA-polymerase based in vitro transcription. Analytical Biochemistry, 334(1), 164–174. https://doi.org/10.1016/j.ab.2004.07.013 Molliver, M. E., Kostovic, I., & van der Loos, H. (1973). The development of synapses in cerebral cortex of the human fetus. Brain Res, 50(2), 403– 407. https://doi.org/0006-8993(73)90741-5 [pii] ET - 1973/02/28 Molnár, Z., & Cheung, A. F. P. (2006). Towards the classification of subpopulations of layer V pyramidal projection neurons. Neuroscience Research, 55(2), 105–115. https://doi.org/10.1016/j.neures.2006.02.008 214 Moreno, T. A., & Bronner-Fraser, M. (2005). Noelins modulate the timing of neuronal differentiation during development. Developmental Biology, 288(2), 434–447. https://doi.org/10.1016/j.ydbio.2005.09.050 Morris, J., Singh, J. M., & Eberwine, J. H. (2011). Transcriptome Analysis of Single Cells. Journal of Visualized Experiments, (50), e2634–e2634. https://doi.org/10.3791/2634 Morrison, J. A., Box, A. C., McKinney, M. C., McLennan, R., & Kulesa, P. M. (2015). Quantitative single cell gene expression profiling in the avian embryo. Developmental Dynamics, 244(6), 774–784. https://doi.org/10.1002/dvdy.24274 Mosca, T. J. (2015). On the Teneurin track: a new synaptic organization molecule emerges. Front Cell Neurosci, 9(May), 204. https://doi.org/10.3389/fncel.2015.00204 Mosca, T. J., Hong, W., Dani, V. S., Favaloro, V., & Luo, L. (2012). Trans- synaptic Teneurin signalling in neuromuscular synapse organization and target choice. Nature, 484(7393), 237–241. https://doi.org/10.1038/nature10923 Mulligan, K. A., & Cheyette, B. N. R. (2012). Wnt signaling in vertebrate neural development and function. Journal of Neuroimmune Pharmacology, 7(4), 774–787. https://doi.org/10.1007/s11481-012-9404-x 215 Myakhar, O., Unichenko, P., & Kirischuk, S. (2011). GABAergic projections from the subplate to Cajal-Retzius cells in the neocortex. Neuroreport, 22(11), 525–9. https://doi.org/10.1097/WNR.0b013e32834888a4 Nakaya, N., Sultana, A., Lee, H. S., & Tomarev, S. I. (2012). Olfactomedin 1 interacts with the Nogo A receptor complex to regulate axon growth. Journal of Biological Chemistry, 287(44), 37171–37184. https://doi.org/10.1074/jbc.M112.389916 Naqui, S. Z. ., Harris, B. ., Thomaidou, D., & Parnavelas, J. . (1999). The noradrenergic system influences the fate of Cajal–Retzius cells in the developing cerebral cortex. Developmental Brain Research, 113(1), 75– 82. https://doi.org/10.1016/S0165-3806(99)00003-6 Noback, C. R., & Purpura, D. P. (1961). Postnatal ontogenesis of neurons in cat neocortex. The Journal of Comparative Neurology, 117(3), 291–307. https://doi.org/10.1002/cne.901170303 Nunome, K., Miyazaki, S., Nakano, M., Iguchi-Ariga, S., & Ariga, H. (2008). Pyrroloquinoline Quinone Prevents Oxidative Stress-Induced Neuronal Death Probably through Changes in Oxidative Status of DJ-1. Biological & Pharmaceutical Bulletin, 31(7), 1321–1326. https://doi.org/10.1248/bpb.31.1321 216 Nusse, W. J. N. and R. (2004). Convergence of Wnt, b -Catenin, and Cadherin Pathways. Science, 303(March), 1483–1488. https://doi.org/10.1126/science.1094291.Convergence Nuytemans, K., Theuns, J., Cruts, M., & Van Broeckhoven, C. (2010). Genetic etiology of Parkinson disease associated with mutations in the SNCA, PARK2, PINK1, PARK7, and LRRK2 genes: a mutation update. Human Mutation, 31(7), 763–780. https://doi.org/10.1002/humu.21277 O’Leary, D. D. M., & Koester, S. E. (1993). Development of projection neuron types, axon pathways, and patterned connections of the mammalian cortex. Neuron, 10(6), 991–1006. https://doi.org/10.1016/0896- 6273(93)90049-W O’Shea, K. S., Liu, L. H. J., & Dixita, V. M. (1991). Thrombospondin and a 140 kd fragment promote adhesion and neurite outgrowth from embryonic central and peripheral neurons and from PC12 cells. Neuron, 7(2), 231– 237. https://doi.org/10.1016/0896-6273(91)90261-W Ogawa, M., Miyata, T., Nakajimat, K., Yagyu, K., Seike, M., Ikenaka, K., … Mikoshibat, K. (1995). The reeler gene-associated antigen on cajal-retzius neurons is a crucial molecule for laminar organization of cortical neurons. Neuron, 14(5), 899–912. https://doi.org/10.1016/0896-6273(95)90329-1 217 Park, J. H., Choi, M. R., Park, K. S., Kim, S. H., Jung, K. H., & Chai, Y. G. (2012). The characterization of gene expression during mouse neural stem cell differentiation in vitro. Neuroscience Letters (Vol. 506). https://doi.org/10.1016/j.neulet.2011.10.046 Parnavelas, J. G., & Edmunds, S. M. (1983). Further evidence that Retzius- Cajal cells transform to nonpyramidal neurons in the developing rat visual cortex. Journal of Neurocytology, 12(5), 863–871. https://doi.org/10.1007/BF01258156 Pavletich, N., & Pabo, C. (1991). Zinc finger-DNA recognition: crystal structure of a Zif268-DNA complex at 2.1 A. Science, 252(5007). Retrieved from http://science.sciencemag.org/content/252/5007/809 Peñagarikano, O., Abrahams, B. S., Herman, E. I., Winden, K. D., Gdalyahu, A., Dong, H., … Geschwind, D. H. (2011). Absence of CNTNAP2 leads to epilepsy, neuronal migration abnormalities, and core autism-related deficits. Cell, 147(1), 235–246. https://doi.org/10.1016/j.cell.2011.08.040 Peng, Y., Barreda Tomás, F. J., Klisch, C., Vida, I., & Geiger, J. R. P. (2017). Layer-Specific Organization of Local Excitatory and Inhibitory Synaptic Connectivity in the Rat Presubiculum. Cerebral Cortex (New York, N.Y. : 1991), 27(4), 2435–2452. https://doi.org/10.1093/cercor/bhx049 218 Perron, J. C., & Bixby, J. L. (1999). Tetraspanins expressed in the embryonic chick nervous system. FEBS Letters, 461(1–2), 86–90. https://doi.org/10.1016/S0014-5793(99)01429-5 Phillips, J., & Eberwine, J. H. (1996). Antisense RNA Amplification : A Linear Amplification Method for Analyzing the mRNA Population from Single Living Cells, 288(104), 283–288. Qiu, S., Luo, S., Evgrafov, O., Li, R., Schroth, G. P., Levitt, P., … Wang, K. (2012). Single-neuron RNA-Seq: technical feasibility and reproducibility. Frontiers in Genetics, 3, 124. https://doi.org/10.3389/fgene.2012.00124 Radnikow, G., Feldmeyer, D., & Lübke, J. (2002). Axonal projection, input and output synapses, and synaptic physiology of Cajal-Retzius cells in the developing rat neocortex. The Journal of Neuroscience : The Official Journal of the Society for Neuroscience, 22(16), 6908–6919. https://doi.org/20026725 Rakic, P. (1974). Neurons in Rhesus Monkey Visual Cortex: Systematic Relation between Time of Origin and Eventual Disposition. Science, 183(4123). Retrieved from http://science.sciencemag.org/content/183/4123/425 Ramo´n y Cajal, S. (1891). Sur la structure de l’e´corce ce´rebrale de quelques mammife`res. Cellule (Vol. 7). 219 Ren, J., & Greer, J. J. (2003). Ontogeny of Rhythmic Motor Patterns Generated in the Embryonic Rat Spinal Cord. Journal of Neurophysiology, 89(3). Retrieved from http://jn.physiology.org/content/89/3/1187.short Retzius, G. (1893). Die Cajal’schen Zellen der Grosshirnrinde beim Menschen und bei Sugetieren. Biologische Untersuchungen, 5, 1–8. Ribeiro-Silva, A., Zhang, H., & Jeffrey, S. S. (2007). RNA extraction from ten year old formalin-fixed paraffin-embedded breast cancer samples: a comparison of column purification and magnetic bead-based technologies. BMC Molecular Biology, 8(8). https://doi.org/10.1186/1471- 2199-8-118 Riccio, O., Hurni, N., Murthy, S., Vutskits, L., Hein, L., & Dayer, A. (2012). Alpha2-adrenergic receptor activation regulates cortical interneuron migration. European Journal of Neuroscience, 36(7), 2879–2887. https://doi.org/10.1111/j.1460-9568.2012.08231.x Ringstedt, T., Linnarsson, S., Wagner, J., Lendahl, U., Kokaia, Z., Arenas, E., … Ibáñez, C. F. (1998). BDNF regulates reelin expression and Cajal- Retzius cell development in the cerebral cortex. Neuron, 21(2), 305–315. https://doi.org/10.1016/S0896-6273(00)80540-1 220 Rodenas-Cuadrado, P., Ho, J., & Vernes, S. C. (2014). Shining a light on CNTNAP2: complex functions to complex disorders. European Journal of Human Genetics : EJHG, 22(2), 171–8. https://doi.org/10.1038/ejhg.2013.100 Sakmann, B., & Neher, E. (1984). Patch Clamp Techniques for Studying Ionic Channels in Excitable Membranes. Annual Review of Physiology, 46(1), 455–472. https://doi.org/10.1146/annurev.ph.46.030184.002323 Sakurai, T. (2016). The role of cell adhesion molecules in brain wiring and neuropsychiatric disorders. Molecular and Cellular Neuroscience. https://doi.org/10.1016/j.mcn.2016.08.005 Salimi, I., Friel, K. M., & Martin, J. H. (2008). Pyramidal tract stimulation restores normal corticospinal tract connections and visuomotor skill after early postnatal motor cortex activity blockade. The Journal of Neuroscience : The Official Journal of the Society for Neuroscience, 28(29), 7426–34. https://doi.org/10.1523/JNEUROSCI.1078-08.2008 Sava, B. A., Dávid, C. S., Teissier, A., Pierani, A., Staiger, J. F., Luhmann, H. J., & Kilb, W. (2010). Electrophysiological and morphological properties of Cajal-Retzius cells with different ontogenetic origins. Neuroscience, 167(3), 724–734. https://doi.org/10.1016/j.neuroscience.2010.02.043 221 Schmid, R. S., & Anton, E. S. (2003). Role of integrins in the development of the cerebral cortex 2. Cerebral Cortex, 13(1047–3211), 219–224. Schmidt, T. M., Chen, S.-K., & Hattar, S. (2011). Intrinsically photosensitive retinal ganglion cells: many subtypes, diverse functions. Trends in Neurosciences, 34(11), 572–80. https://doi.org/10.1016/j.tins.2011.07.001 Schnepp, A., Lindgren, P. K., Hülsmann, H., Kröger, S., Paulsson, M., & Hartmann, U. (2005). Mouse testican-2: Expression, glycosylation, and effects on neurite outgrowth. Journal of Biological Chemistry, 280(12), 11274–11280. https://doi.org/10.1074/jbc.M414276200 Schwaller, B. (2014). Calretinin: from a “simple” Ca(2+) buffer to a multifunctional protein implicated in many biological processes. Front Neuroanat, 8(February), 3. https://doi.org/10.3389/fnana.2014.00003 Schwartz, T. H., Rabinowitz, D., Unni, V., Kumar, V. S., Smetters, D. K., Tsiola, A., & Yuste, R. (1998). Networks of coactive neurons in developing layer 1. Neuron, 20(3), 541–552. https://doi.org/10.1016/S0896- 6273(00)80993-9 Sekine, K., Kubo, K. I., & Nakajima, K. (2014). How does Reelin control neuronal migration and layer formation in the developing mammalian neocortex? Neuroscience Research, 86, 50–58. https://doi.org/10.1016/j.neures.2014.06.004 222 Shao, L., Halvorsrud, R., BorgGraham.Lyle, & Storm, J. (1999). The role of BK-type Ca+dependent K+ channels in spike broadening during repetitive firing in rat hippocampal pyramidal cells. Journal of Physiology, 521(1), 135–146. https://doi.org/10.1111/j.1469-7793.1999.00135.x Soriano, E., Alvarado-Mallart, R. M., Dumesnil, N., Del Río, J. A., & Sotelo, C. (1997). Cajal-Retzius cells regulate the radial glia phenotype in the adult and developing cerebellum and alter granule cell migration. Neuron, 18, 563–577. https://doi.org/10.1016/S0896-6273(00)80298-6 Spaethling, J. M., Na, Y.-J., Lee, J., Ulyanova, A. V., Baltuch, G. H., Bell, T. J., … Eberwine, J. H. (2017). Primary Cell Culture of Live Neurosurgically Resected Aged Adult Human Brain Cells and Single Cell Transcriptomics. Cell Reports, 18(3), 791–803. https://doi.org/10.1016/j.celrep.2016.12.066 Spitzer, N. C. (2006). Electrical activity in early neuronal development. Nature, 444(7120), 707–712. https://doi.org/10.1038/nature05300 Strauss, K. A., Puffenberger, E. G., Huentelman, M. J., Gottlieb, S., Dobrin, S. E., Parod, J. M., … Morton, D. H. (2006). Recessive symptomatic focal epilepsy and mutant contactin-associated protein-like 2. The New England Journal of Medicine, 354(13), 1370–1377. https://doi.org/354/13/1370 [pii] 10.1056/NEJMoa052773 223 Su, T.-P., Su, T.-C., Nakamura, Y., & Tsai, S.-Y. (2016). The Sigma-1 Receptor as a Pluripotent Modulator in Living Systems. Trends in Pharmacological Sciences, 37(4), 262–278. https://doi.org/10.1016/j.tips.2016.01.003 Supèr, H., Del Río, J. A., Martínez, A., Pérez-Sust, P., & Soriano, E. (2000). Disruption of neuronal migration and radial glia in the developing cerebral cortex following ablation of Cajal-Retzius cells. Cerebral Cortex (New York, N.Y. : 1991), 10(6), 602–613. https://doi.org/10.1093/cercor/10.6.602 Sutherland, D. J., Pujic, Z., & Goodhill, G. J. (2014). Calcium signaling in axon guidance. Trends in Neurosciences, 37(8), 424–432. https://doi.org/10.1016/j.tins.2014.05.008; 10.1016/j.tins.2014.05.008 Thomas, S. A., Matsumoto, A. M., & Palmiter, R. D. (1995). Noradrenaline is essential for mouse fetal development. Nature, 374(6523), 643–646. https://doi.org/10.1038/374643a0 Thomson, A. M. (2010). Neocortical layer 6, a review. Frontiers in Neuroanatomy, 4, 13. https://doi.org/10.3389/fnana.2010.00013 Tissir, F., & Goffinet, A. M. (2003). Reelin and brain development. Nature Reviews. Neuroscience, 4(6), 496–505. https://doi.org/10.1038/nrn1113 224 Tole, S., & Patterson, P. H. (1993). Distribution of CD9 in the developing and mature rat nervous system. Developmental Dynamics : An Official Publication of the American Association of Anatomists, 197(2), 94–106. https://doi.org/10.1002/aja.1001970203 Trommsdorff, M., Gotthardt, M., Hiesberger, T., Shelton, J., Stockinger, W., Nimpf, J., … Herz, J. (1999). Reeler/disabled-like disruption of neuronal migration in knockout mice lacking the VLDL receptor and ApoE receptor 2. Cell, 97(6), 689–701. https://doi.org/10.1016/S0092-8674(00)80782-5 Valverde, F. (1977). Lamination of the striate cortex. Journal of Neurocytology, 6(4), 483–484. https://doi.org/10.1007/BF01178230 van Dijk, D., Nainys, J., Sharma, R., Kathail, P., Carr, A. J., Moon, K. R., … Pe’er, D. (2017). MAGIC: A diffusion-based imputation method reveals gene-gene interactions in single-cell RNA-sequencing data. bioRxiv. Retrieved from http://biorxiv.org/content/early/2017/02/25/111591 Vilar, M., Murillo-Carretero, M., Mira, H., Magnusson, K., Besset, V., & Ibáñez, C. F. (2006). Bex1, a novel interactor of the p75 neurotrophin receptor, links neurotrophin signaling to the cell cycle. The EMBO Journal, 25(6), 1219–30. https://doi.org/10.1038/sj.emboj.7601017 225 Vogt Weisenhorn, D. M., Weruaga Prieto, E., & Celio, M. R. (1994). Localization of calretinin in cells of layer I (Cajal-Retzius cells) of the developing cortex of the rat. Developmental Brain Research, 82(1–2), 293–297. https://doi.org/10.1016/0165-3806(94)90171-6 Wakamatsu, Y., Maynard, T. M., & Weston, J. A. (2000). Fate determination of neural crest cells by NOTCH-mediated lateral inhibition and asymmetrical cell division during gangliogenesis. Development, 127(13). Retrieved from http://dev.biologists.org/content/127/13/2811.short Wang, Z., Gerstein, M., & Snyder, M. (2009a). RNA-Seq: a revolutionary tool for transcriptomics. Nature Reviews Genetics, 10(1), 57–63. https://doi.org/10.1038/nrg2484 Wang, Z., Gerstein, M., & Snyder, M. (2009b). RNA-Seq: a revolutionary tool for transcriptomics. Nature Reviews Genetics, 10(1), 57–63. https://doi.org/10.1038/nrg2484 Wayman, G. A., Lee, Y. S., Tokumitsu, H., Silva, A., & Soderling, T. R. (2008). Calmodulin-Kinases: Modulators of Neuronal Development and Plasticity. Neuron, 59(6), 914–931. https://doi.org/10.1016/j.neuron.2008.08.021 Wheeler, D. B., Randall, A., & Tsien, R. W. (1994). Roles of N-Type and Q- Type Ca2 + Channels in Supporting Hippocampal Synaptic Transmission. Science, 264(5155), 107–111. 226 Wheeler, D. G., Barrett, C. F., Groth, R. D., Safa, P., & Tsien, R. W. (2008). CaMKII locally encodes L-type channel activity to signal to nuclear CREB in excitation - transcription coupling. Journal of Cell Biology, 183(5), 849– 863. https://doi.org/10.1083/jcb.200805048 Wijnholds, J., Chowdhury, K., Wehr, R., & Gruss, P. (1995). Segment-Specific Expression of the neuronatin Gene during Early Hindbrain Development. Developmental Biology, 171(1), 73–84. https://doi.org/10.1006/dbio.1995.1261 Windrem, M. S., Schanz, S. J., Morrow, C., Munir, J., Chandler-Militello, D., Wang, S., & Goldman, S. A. (2014). A Competitive Advantage by Neonatally Engrafted Human Glial Progenitors Yields Mice Whose Brains Are Chimeric for Human Glia. Journal of Neuroscience, 34(48), 16153– 16161. https://doi.org/10.1523/JNEUROSCI.1510-14.2014 Wozniak, A., Hutchison, R. E., Morris, C. M., & Hutchison, J. B. (1998). Neuroblastoma and Alzheimer’s Disease Brain Cells Contain Aromatase Activity. Steroids, 63(5), 263–267. https://doi.org/10.1016/S0039- 128X(98)00029-4 Wu, A. R., Neff, N. F., Kalisky, T., Dalerba, P., Treutlein, B., Rothenberg, M. E., … Quake, S. R. (2014). Quantitative assessment of single-cell RNA- sequencing methods. Nature Methods, 11(1), 41–6. https://doi.org/10.1038/nmeth.2694 227 Yamada, T., Pfaff, S. L., Edlund, T., & Jessell, T. M. (1993). Control of cell pattern in the neural tube: Motor neuron induction by diffusible factors from notochord and floor plate. Cell, 73(4), 673–686. https://doi.org/10.1016/0092-8674(93)90248-O Yamada, T., Placzek, M., Tanaka, H., Dodd, J., & Jessell, T. M. (1991). Control of cell pattern in the developing nervous system: Polarizing activity of the floor plate and notochord. Cell, 64(3), 635–647. https://doi.org/10.1016/0092-8674(91)90247-V Yamamoto, A., Uchiyama, K., Nara, T., Nishimura, N., Hayasaka, M., Hanaoka, K., & Yamamoto, T. (2014). Structural abnormalities of corpus callosum and cortical axonal tracts accompanied by decreased anxiety- like behavior and lowered sociability in spock3-mutant mice. Developmental Neuroscience, 36(5), 381–395. https://doi.org/10.1159/000363101 Yamazaki, H., Sekiguchi, M., Takamatsu, M., Tanabe, Y., & Nakanishi, S. (2004). Distinct ontogenic and regional expressions of newly identified Cajal-Retzius cell-specific genes during neocorticogenesis. Proceedings of the National Academy of Sciences of the United States of America, 101(40), 14509–14. https://doi.org/10.1073/pnas.0406295101 228 Yoon, S. O., Casaccia-Bonnefil, P., Carter, B., Chao, M. V, Zhang, H., Shin, H., … Bredesen, D. (1998). Competitive signaling between TrkA and p75 nerve growth factor receptors determines cell survival. The Journal of Neuroscience : The Official Journal of the Society for Neuroscience, 18(9), 3273–81. https://doi.org/10.1074/JBC.274.42.30202 Young, J. E., Boulanger-Weill, J., Williams, D. A., Woodruff, G., Buen, F., Revilla, A. C., … Goldstein, L. S. B. (2015). Elucidating Molecular Phenotypes Caused by the SORL1 Alzheimer’s Disease Genetic Risk Factor Using Human Induced Pluripotent Stem Cells. Cell Stem Cell, 16(4), 373–385. https://doi.org/10.1016/j.stem.2015.02.004 Yu, X., & Malenka, R. C. (2003). Beta-catenin is critical for dendritic morphogenesis. Nature Neuroscience, 6(11), 1169–77. https://doi.org/10.1038/nn1132 Zeisel, A., & Linnarsson, S. (2014). Cell types in the mouse cortex and hippocampus revealed by single-cell RNA-seq. Science, 25, 279–284. Retrieved from http://www.sciencemag.org/content/347/6226/1138.long Zhang, L. I., & Poo, M. M. (2001). Electrical activity and development of neural circuits. Nature Neuroscience, 4 Suppl, 1207–1214. https://doi.org/10.1038/nn753 229 Zhao, X., Kuja-Panula, J., Sundvik, M., Chen, Y.-C., Aho, V., Peltola, M. a, … Rauvala, H. (2014). Amigo adhesion protein regulates development of neural circuits in zebrafish brain. The Journal of Biological Chemistry, 289(29), 19958–75. https://doi.org/10.1074/jbc.M113.545582

Abstract (if available)

Abstract The genome encodes information for all the different cell types in the body. For any cell type, only a fraction of the genome is transcribed to messenger RNA (mRNA). Recently, it has become possible to sequence RNA in the minute (≤ 10 pg) amounts found in single cells (scRNAseq). Preliminary analysis of single-cell RNA shows considerable cell-to-cell heterogeneity, even among ostensibly identical cells, raising the question: how much of the variability is true biological variability, and how much is technical noise. The first part of my thesis describes how I analyzed scRNAseq methodology and implemented improvements. Subsequent chapters detail the application of the improved scRNAseq methodology to neurons of human adult and embryonic brain. ❧ Currently, most single cell sequencing techniques entail separating tissue into individual cells and the lysis of individual, isolated cells, followed by the immediate collection of released nucleotides. Automation, using microfluidics platforms, has enabled high throughput

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

Understanding ancestry-specific disease allelic effect sizes by leveraging multi-ancestry single-cell RNA-seq data

PDF

Model selection methods for genome wide association studies and statistical analysis of RNA seq data

PDF

Best practice development for RNA-Seq analysis of complex disorders, with applications in schizophrenia

$Malignant cell fraction prediction using deep learning: from point estimate to uncertainty quantification$

PDF

Malignant cell fraction prediction using deep learning: from point estimate to uncertainty quantification

PDF

Single cell gene expression analysis of KSHV infection in three-dimensional oral epithelial tissue cultures

PDF

LINC00261 induces a G2/M cell cycle arrest and activation of the DNA damage response in lung adenocarcinoma

PDF

Cultured neuronal cells derived from the olfactory neuroepithelium growing in three dimensions as a model system for schizophrenia

PDF

The role of miRNA and its regulation in pulmonary hypertension in sickle cell disease

PDF

Nonlinear dynamical modeling of single neurons and its application to analysis of long-term potentiation (LTP)

PDF

Metabolic profiling of single hematopoietic stem cells for developing novel ex vivo culture strategies

PDF

The role of adipocytes in acute lymphoblastic leukemia cell migration and survival against daunorubicin

PDF

Generation and long-term culture of human cerebellar organoids to study development and disease.

PDF

Rational selection of CRISPR/Cas9 guide RNAs for homology directed genome editing and its utility in the development of gene therapies

PDF

The development of targeted transcription factor transposition and understanding chromatin dynamics in hypertrophic cardiomyopathy

PDF

Axon guidance cues in development of the mammalian auditory circuit

PDF

Characterization of black carbon: from source to evolution of physical and optical properties in the atmosphere

PDF

Detecting joint interactions between sets of variables in the context of studies with a dichotomous phenotype, with applications to asthma susceptibility involving epigenetics and epistasis

Asset Metadata

Creator Kim, Jae Mun (author)

Core Title Patch aRNA in vitro amplification (PAIA): single cell RNA-seq to expand the understanding of the developing and developed nervous system

Contributor Electronically uploaded by the author (provenance)

School College of Letters, Arts and Sciences

Degree Doctor of Philosophy

Degree Program Neuroscience

Publication Date 10/09/2017

Defense Date 05/26/2017

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

Tag Development,embryonic,neuron,neuronal migration,OAI-PMH Harvest,Patch-seq,RNA-seq,single cell,single cell RNA-seq

Language English

Advisor Chen, Jeannie (committee chair), Chow, Robert Hsiu-Ping (committee member), Fraiser, Scott E. (committee member), Knowles, James Arnold (committee member)

Creator Email jaemunki@usc.edu,neurohugo@gmail.com

Permanent Link (DOI) https://doi.org/10.25549/usctheses-c40-444862

Unique identifier UC11264238

Identifier etd-KimJaeMun-5827.pdf (filename),usctheses-c40-444862 (legacy record id)

Legacy Identifier etd-KimJaeMun-5827.pdf

Dmrecord 444862

Document Type Dissertation

Rights Kim, Jae Mun

Type texts

Source 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 a...

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