Close
About
FAQ
Home
Collections
Login
USC Login
Register
0
Selected
Invert selection
Deselect all
Deselect all
Click here to refresh results
Click here to refresh results
USC
/
Digital Library
/
University of Southern California Dissertations and Theses
/
Perimenopausal transition increases blood brain permeability: implications for neurodegenerative diseases
(USC Thesis Other)
Perimenopausal transition increases blood brain permeability: implications for neurodegenerative diseases
PDF
Download
Share
Open document
Flip pages
Contact Us
Contact Us
Copy asset link
Request this asset
Transcript (if available)
Content
i
PERIMENOPAUSAL TRANSITION INCREASES BLOOD BRAIN PERMEABILITY:
IMPLICATIONS FOR NEURODEGENERATIVE DISEASES
by
Maunil Kandarp Desai
A Dissertation Presented to the
FACULTY OF THE USC GRADUATE SCHOOL
UNIVERSITY OF SOUTHERN CALIFORNIA
In Partial Fulfillment of the
Requirements for the Degree
DOCTOR OF PHILOSOPHY
(CLINICAL AND EXPERIMENTAL THERAPEUTICS)
May 2019
Copyright 2019 Maunil Kandarp Desai
ii
Dedication
ॐ
सर्व े भर्व न्त ु स ु ख ि न ः ।
सर्व े सन्त ु नन रामय ाः ।
सर्व े भद्रानि पश्यन्त ु ।
मा क नित ् द ु ःि भाग्भर्व े त ् ॥
ॐ शाखन्तः शाखन्तः शाख न्त ः ॥
May all be happy
May all be free from illness
May all see what is spiritually uplifting
May no one suffer
May peace reign supreme
To all women and men; researchers, scientists, healthcare providers, patients
and participants; to each and every one who has contributed to moving the
needle to understand, prevent, detect, diagnose and cure every disease we
know.
iii
Acknowledgements
To my parents, for all the sacrifices they made to give me the opportunities that I
have; for the unconditional love and support that they always and
unquestioningly bestowed upon me and for the blessings they shower upon me.
To the Bhavsar family, for taking me under their wings, treating me as a family
member and making me feel at home away from home.
To my wife and my in-laws who were ever so supportive in my quest for
knowledge and whose contributions are too numerous to list.
To my advisor, Dr. Roberta Diaz Brinton, thank you for your mentorship; for
encouraging me to grow as a researcher and for providing me opportunities to
learn and gain experience, both in pre-clinical and translational science projects
as well as the Allopregnanolone clinical trial. I very much value and appreciate
your faith, support and guidance.
To my committee chair, Dr. Stan Louie who has always been supportive,
understanding and cared about students with the deepest affection, and Dr.
Enrique Cadenas for honoring me by being on my committee and being an
amazingly helpful, kind and wonderful role model. Special thanks to Dr. Curtis
iv
Okamoto, Dr. Ronald Irwin, Dr. Annie Wong-Beringer, Wade Thompson-Harper
and Rosie Soltero.
To the members of the Brinton Lab, and Dr. Yibu Chen and Meng Li from the
Bioinformatics Center, thank you so much for all the help and assistance you all
have provided through the years. I appreciate everything each one of you taught
me and consider myself fortunate to have you as colleagues and friends.
Last but not the least, to the participants of the Allopregnanolone clinical trial,
who were some of the most amazing women and men I have had to opportunity
to converse with and whose selfless and anonymous contribution to science is
immortalized.
v
Table of Contents
DEDICATION II
ACKNOWLEDGEMENTS III
TABLE OF CONTENTS V
LIST OF FIGURES VII
LIST OF TABLES IX
ABSTRACT X
BACKGROUND 1
1.1 WOMEN AND ALZHEIMER ’S DISEASE 1
1.2 WOMEN AND IMMUNE SYSTEM 2
1.3 COGNITION AND PERIMENOPAUSE 3
1.4 ESTROGEN, BLOOD BRAIN BARRIER AND AD 4
1.5 THE RODENT PERIMENOPAUSAL MODEL 7
1.6 CONCLUSION/HYPOTHESIS 13
TRANSCRIPTOMIC ANALYSIS 14
2.1 ABSTRACT 14
2.2 INTRODUCTION 14
2.3 MATERIALS & METHODS 15
2.4 RESULTS 21
2.5 DISCUSSION 48
2.6 CONCLUSION 54
BLOOD BRAIN BARRIER (BBB) PERMEABILITY IN PERIMENOPAUSAL FEMALE RATS 55
3.1 ABSTRACT 55
3.2 INTRODUCTION 55
3.3 MATERIALS & METHODS 56
3.4 RESULTS 59
3.5 DISCUSSION 66
3.6 CONCLUSION 68
ALLOPREGNANOLONE: A POTENTIAL NEUROREGENERATIVE THERAPEUTIC 70
4.1 ABSTRACT 70
4.2 INTRODUCTION 71
4.3 METHODS: 72
vi
4.4 RESULTS 83
4.5 DISCUSSION 111
4.6 CONCLUSION 117
CONCLUSION 119
5.1 TRANSCRIPTOMIC AND BIOINFORMATIC APPROACH TO THE SYSTEMS BIOLOGY OF ENDOCRINE AGING 120
5.2 FUNCTIONAL CONFIRMATION 122
5.3 ROLE OF ALLOPREGNANOLONE 124
BIBLIOGRAPHY 127
vii
List of Figures
FIGURE 1.1 ROLE OF ESTROGEN IN MAINTAINING A HEALTHY BLOOD BRAIN BARRIER. .......... 5
FIGURE 2.1 FLOW CHART OF DATA ANALYSIS FOR RNA-SEQ BIOINFORMATICS PIPELINE. ...21
FIGURE 2.2 DEVELOPING RNA-SEQ DATA ANALYSIS PIPELINE: FIRST ITERATION. ..............22
FIGURE 2.3 DEVELOPING RNA-SEQ DATA ANALYSIS PIPELINE: SECOND ITERATION. ..........23
FIGURE 2.4 FUNCTIONAL ANALYSES RESULTS OF THE SECOND ITERATION WITH FIVE DEG
LISTS. .....................................................................................................................24
FIGURE 2.5 FUNCTIONAL DATA USED TO DEVELOP BIOINFORMATICS PIPELINE. ...................25
FIGURE 2.6 RNA-SEQ DATA ANALYSIS: CROSS VALIDATION WITH PATHWAYS PUBLISHED
PREVIOUSLY. ...........................................................................................................26
FIGURE 2.7 NUMBER OF GENES SIGNIFICANTLY UP- AND DOWN-REGULATED BETWEEN
DIFFERENT PERIMENOPAUSAL GROUPS. ....................................................................28
FIGURE 2.8 NUMBER OF HIPPOCAMPAL GENES UP- AND DOWN-REGULATED DUE TO
OVARIECTOMY. ........................................................................................................29
FIGURE 2.9 FUNCTIONAL CLASSIFICATION OF GENES. .......................................................30
FIGURE 2.10A-C UCP2 EXPRESSION IN THE HIPPOCAMPAL TRANSCRIPTOME OF
PERIMENOPAUSAL RATS. .........................................................................................32
FIGURE 2.11A-B PANTHER CLASSIFICATION SYSTEM: REGULAR CYCLERS 6M VS. REGULAR
CYCLERS 9-10M. .....................................................................................................35
FIGURE 2.12A-B PANTHER CLASSIFICATION SYSTEM: REGULAR CYCLERS 9-10M VS.
IRREGULAR CYCLERS 9-10M. ...................................................................................37
FIGURE 2.13A-B PANTHER CLASSIFICATION SYSTEM: REGULAR OVX 10-10.5M VS.
REGULAR CYCLERS 9-10M. ......................................................................................39
FIGURE 2.14A-B UP-REGULATION OF T LYMPHOCYTE RELATED PROCESS: IRREGULAR 9-10M
VS. REGULAR 9-10M. ...............................................................................................41
FIGURE 2.15 DIFFERENTIAL REGULATION OF INTERLEUKIN (IL)-27 BETWEEN DIFFERENT
GROUPS OF FEMALE SD RATS. .................................................................................42
FIGURE 2.16 INSULIN SIGNALING PATHWAY IN THE HIPPOCAMPAL TRANSCRIPTOME DURING
THE REGULAR TO IRREGULAR TRANSITION AT 9-10 MONTHS OF AGE IN FEMALE SD
RATS. .....................................................................................................................43
FIGURE 2.17 DOWNREGULATION OF MITOCHONDRIAL COMPLEXES OF THE ELECTRON
TRANSPORT CHIAN AS WELL AS CARNITINE PALMITOYLTRANSFERASE I IN THE
HIPPOCAMPAL TRANSCRIPTOME DURING THE REGULAR TO IRREGULAR TRANSITION AT
9-10 MONTHS OF AGE IN FEMALE SD RATS. ..............................................................44
FIGURE 2.18 DOWNREGULATION OF PDGF SIGNALING PATHWAY DURING THE
PERIMENOPAUSAL TRANSITION. ................................................................................45
FIGURE 2.19 TGFB1 GENE NETWORK SHOWING DOWNREGULATION OF PDGF, SOLUTE
CARRIERS AND INTEGRIN BINDING PROTEINS IN THE HIPPOCAMPAL TRANSCRIPTOME
DURING THE PERIMENOPAUSAL TRANSITION. .............................................................46
FIGURE 3.1 CALCULATION OF BLOOD BRAIN BARRIER PERMEABILITY INDEX (BBB-PI). .....59
FIGURE 3.2 BLOOD BRAIN BARRIER PERMEABILITY INDEX FOR HYPOTHALAMUS IN FEMALE
SD RATS. ................................................................................................................60
FIGURE 3.3 BLOOD BRAIN BARRIER PERMEABILITY INDEX OF HIPPOCAMPUS IN FEMALE SD
RATS. .....................................................................................................................61
viii
FIGURE 3.4 BLOOD BRAIN BARRIER PERMEABILITY INDEX OF CORTEX IN FEMALE SD RATS.
..............................................................................................................................62
FIGURE 3.5 BLOOD BRAIN BARRIER PERMEABILITY INDEX (BBB-PI) OF HYPOTHALAMUS
WITH STATISTICAL SIGNIFICANCE. .............................................................................63
FIGURE 3.6 BLOOD BRAIN BARRIER PERMEABILITY INDEX OF HIPPOCAMPUS AND CORTEX
WITH STATISTICAL SIGNIFICANCE. .............................................................................64
FIGURE 3.7 VARIABILITY IN BLOOD BRAIN BARRIER PERMEABILITY INDEX OF
HYPOTHALAMUS, HIPPOCAMPUS AND CORTEX. ..........................................................65
FIGURE 4.1 ALLO ‘PULSE’ PROTOCOL. .............................................................................78
FIGURE 4.2 BODY WEIGHTS OF DIFFERENT GROUPS OF APOE MICE BY SEX, GENOTYPE AND
TREATMENT. ...........................................................................................................83
FIGURE 4.3 DIFFERENCE IN EXPLORATION TIME BETWEEN DIFFERENT GROUPS. ..................84
FIGURE 4.4 DISCRIMINATION INDEX IN DIFFERENT GROUPS. ...............................................85
FIGURE 4.5A-B DIFFERENCE IN EXPLORATION TIME AND DISCRIMINATION INDEX FOR ALLO
AND SALINE TREATED APOE 3/4 HETEROZYGOUS FEMALE MICE. ................................86
FIGURE 4.6 UP- AND DOWN-REGULATED CANONICAL PATHWAYS IN APOE 4/4 FEMALES IN
RESPONSE TO ALLO. ...............................................................................................88
FIGURE 4.7 CANONICAL PATHWAYS UP- AND DOWN-REGULATED BY ALLO TREATMENT IN ALL
THREE FEMALE GENOTYPES. ....................................................................................89
FIGURE 4.8 UPREGULATION OF PDGF-BB IN APOE 4/4 ALLO TREATED FEMALES COMPARED
TO SALINE TREATED FEMALES OF IDENTICAL GENOTYPE. ...........................................90
FIGURE 4.9 GENOTYPE AND GENDER DIFFERENCES IN TREATMENT EFFECT OF ALLO ON
HIPPOCAMPAL TRANSCRIPTOME IN APOE MICE. ........................................................91
FIGURE 4.10 DIFFERENCES IN CORTICAL METABOLITES IN ALLO VS. SALINE TREATED APOE
4/4 FEMALES. ..........................................................................................................94
FIGURE 4.11 CORTICAL METABOLITE DIFFERENCES BETWEEN SALINE TREATED APOE 3/3
AND 3/4 FEMALES. ...................................................................................................95
FIGURE 4.12 PLASMA METABOLITE DIFFERENCES BETWEEN SALINE TREATED APOE 3/3 VS
3/4 FEMALE MICE. ....................................................................................................96
FIGURE 4.13A-C PLASMA METABOLITE DIFFERENCES BETWEEN SALINE TREATED APOE 4/4
VS 3/4 FEMALE MICE. ...............................................................................................97
FIGURE 4.14A-B CORTICAL AND PLASMA DIFFERENCES IN METABOLITES BETWEEN SALINE
TREATED APOE 3/3 AND 4/4 FEMALE MICE. ...............................................................99
FIGURE 4.15A-D METABOLITE DIFFERENCES IN CORTEX AND PLASMA BETWEEN SALINE
TREATED APOE 3/3 AND 4/4 MALES AND FEMALES. ................................................. 102
FIGURE 4.16 ALLO DOSE RESPONSE CURVE IN HNSCS WITH RESPECT TO MITOCHONDRIAL
POTENTIATION. ...................................................................................................... 105
FIGURE 4.17A-B ALLO AND ITS ANALOGUES DIFFERENTIALLY POTENTIATE MITOCHONDRIAL
FUNCTION AND INCREASE SPARE RESPIRATORY CAPACITY OF HNSCS. ..................... 106
FIGURE 4.18 ALLO AND ITS ANALOGUES POTENTIATE MITOCHONDRIAL FUNCTION. ........... 106
FIGURE 4.19A-B EXPOSURE OF HNSCS TO 30-MIN. PULSE OF ALLO AND ITS EFFECTS ON
MICHODNRIAL FUNCTION. ....................................................................................... 108
FIGURE 4.20 EVALUATION OF ALLO AS AN AGONIST OR ANTAGONIST OF TSPO IN
MITOCHONDRIA. .................................................................................................... 109
FIGURE 4.21 ALLO STATISTICALLY SIGNIFICANTLY INCREASED SRC OF MITOCHONDRIA IN
HNSC AFTER 24-HOUR EXPOSURE COMPARED TO TSPO AGONIST AND ANTAGONIST.
............................................................................................................................ 110
ix
List of Tables
TABLE 1.1 COMPARISONS OF THE DIFFERENT CHRONOLOGICAL AND ENDOCRINOLOGICAL
GROUPS OF FEMALE RATS. .......................................................................................11
x
Abstract
Introduction. The onset of reproductive senescence occurs at the perimenopause, which
is characterized by major physiological changes in the endocrinological and reproductive
systems which can be associated with altered energy metabolism, cognition, bone-mineral
density, cardiovascular function and immune system responses. Our rat model of the
perimenopause to menopause to post menopause transitions captures both chronological
and endocrinological conversions. In the current study, differential regulation of genes in
the hippocampi of the endocrine characterized female rats by sequencing hippocampal
total RNA was investigated, based on which it was hypothesized that perimenopausal
transition increases blood brain permeability with grave implications for
neurodegenerative diseases in women. However, based on rodent model, a novel
neuroregenerative therapeutic, Allopregnanolone, may help ameliorate the increase in
permeability and improve cognitive function.
Methods. Paired end sequencing of hippocampal total RNA from 36 Sprague Dawley
(SD) female rat hippocampi, belonging to six different groups was conducted. A list of
differentially expressed genes (DEG) for each comparison was obtained using TopHat
and Cufflinks employed in the Partek Flow environment (http://www.partek.com/).
Differentially expressed genes were analyzed using Ingenuity Pathway Analysis and
PANTHER to identify gene pathways altered during the perimenopause. Blood brain
barrier (BBB) permeability was measured in different groups of female SD rats using
sodium fluorescein. Both male and female mice were treated with Allopregnanolone
intramuscularly. Behavior was assessed using Novel Object Recognition. Plasma and
xi
cortex from mice were queried for 185 metabolites using ultra-performance-LCMS. Mice
hippocampi were used for RNA-Seq.
Results. RNA-Seq detected a decrease in platelet-derived growth factor (PDGF) activity,
associated with decreased brain estrogen, indicating potential for increased blood brain
barrier permeability during the perimenopausal hippocampus. Analyses of significantly
differentially expressed genes (DEG) using PANTHER (http://pantherdb.org/index.jsp)
revealed enrichment of leukocytes, B and T lymphocyte genes. Curated literature based
bioinformatic analysis indicated activation of pathways related to lymphocytic proliferation
and differentiation during the transition from regular to irregular cycling, and inactivation
of pathways responsible for T cell apoptosis. Interestingly, interleukin-27, a pan-T
lymphocyte regulator was down-regulated in the hippocampal transcriptome during the
perimenopausal transition, hinting at potential T lymphocyte dysregulation associated with
the perimenopausal transition. Increased permeability of BBB during the perimenopausal
transition was confirmed using sodium fluorescein. A separate experiment suggested that
Allopregnanolone improved cognitive function in human ApoE ε4 allele containing
Targeted Replacement mice, potentially upregulated PDGF signaling pathway in the
hippocampal transcriptome which suggests protection of blood brain barrier integrity,
increased lipid metabolism to generate acetyl-CoA to feed into the TCA cycle and
promoted ATP generation in the mitochondria. Further, Allo treatment increased indicators
of protein metabolism and potentiated mitochondria in human neural stem cells.
Conclusions. These RNA-Seq and bioinformatic analyses suggested a functional
disruption of the blood brain barrier during the perimenopausal transition state that was
confirmed. Both chronological and endocrinological aging in female rats resulted in
divergence into sub-populations based on blood brain barrier permeability corroborating
xii
evidence of inter-individual variation in neurodegenerative diseases. Allopregnanolone not
only ameliorated cognitive function but based on analysis of hippocampal transcriptome
suggested amelioration of blood brain barrier permeability, which must be functionally
confirmed, and its mechanism elicited in future studies.
1
Background
1.1 Women and Alzheimer’s Disease
Alzheimer’s disease (AD) is a major cause of morbidity and mortality in the United States.
(Association, 2018). Women bear and unequal burden of AD by being more likely to be
AD patients as well as caregivers (Beydoun et al., 2013). At age 65, a woman’s risk of
being diagnosed with AD is almost double that of a man of the same age (Association,
2014).
Recent publications have explored the extent of contribution of sex biology to vulnerability
to AD. Even when healthy, adult men suffer from faster reduction in brain volume
compared to women (Pfefferbaum et al., 2013), but this is reversed in AD where brain
volume declines faster in women than men (Skup et al., 2011). More clarity is needed
regarding the contribution of X and Y chromosomes in AD (Snyder et al., 2016).
ApoE ε4 gene, known to increase risk of AD in general, (Corder et al., 1993; Liu et al.,
2013), elevates women’s risk of developing AD more than men (Farrer et al., 1997b) and
causes higher dysregulation of brain functions in women (Altmann et al., 2014;
Damoiseaux et al., 2012; Fleisher et al., 2008; Sampedro et al., 2015).
Even though research conducted in vitro and in animal models suggested neuroprotective
effects of sex steroids (estrogens and progestins) in women, such evidence has been
lacking in the clinic and trials in large number of women have not shown benefit of
hormonal supplements (Carcaillon et al., 2014; Dubal et al., 2012; Espeland et al., 2004;
Gleason et al., 2015; Henderson et al., 2016; Laughlin et al., 2010) and need further
investigation. In the future, studies could focus on the mechanisms by which hormones
affect the ability of the brain to respond to injury and ameliorate effects of noxious stimuli
2
in men and women in mid- and late-life thereby clarifying the protective role of sex
hormones in AD.
1.2 Women and Immune System
Postmenopausal females are at greater risk for not only AD (Christensen and Pike, 2015)
but also autoimmune diseases (Fairweather et al., 2008). Incidence of autoimmune
diseases such as Rheumatoid Arthritis and Multiple Sclerosis rise in women through
perimenopausal age, peak around menopause and decrease thereafter (Grytten et al.,
2015; Peppercorn, 2014). Additionally, evolutionary biologists have hypothesized that
earlier age at menopause would shift the onset of RA to the left i.e. the disease incidence
would peak earlier in younger age groups (Straub and Schradin, 2016).
Differences in the immune system have been known to exist among males and females
across their lifespan, across different mammalian species and in both innate and adaptive
immune system (Griesbeck et al., 2015; Klein and Flanagan, 2016; Pisitkun et al., 2006)
with an important role for sex steroids (Griesbeck et al., 2015) which suggests that sex
steroids, such as androgens and estrogens may be directly responsible for differences in
innate immune responses (Hannah et al., 2008).
Males and females differ in the production of cytokines and chemokines (Aomatsu et al.,
2013; Rettew et al., 2008; Torcia et al., 2012) and females across different ethnicities have
higher counts of helper T cells compared to males (Abdullah et al., 2012; Lee et al., 1996;
Lisse et al., 1997; Uppal et al., 2003), and sexes also show differences in the way these
helper T cells set up an immune response (Giron-Gonzalez et al., 2000; Hewagama et al.,
2009; Roberts et al., 2001; Sankaran-Walters et al., 2013; Zhang et al., 2012). Studies in
mice exploring sex-based differences in regulatory T cells describe contrasting results with
3
respect to organ-specific regulatory T cell counts in multiple diseases; on the other hand,
human studies show more regulatory T cells in men versus women (Afshan et al., 2012).
Studies have shown that not only B cells are higher in females among older (Teixeira et
al., 2011) and younger adults (Abdullah et al., 2012), but PBMCs from both sexes when
stimulated lead to a significantly differential alteration in Natural Killer cell count (Abdullah
et al., 2012). Another study found sex-based differences in expression of genes in B cells,
and authors of the study suspected gonadal hormonal influence (Fan et al., 2014).
1.3 Cognition and Perimenopause
Perimenopause is the period that precedes menopause when women begin to experience
menstrual cycle changes (McKinlay et al., 1992; Nelson, 2008) and is associated with
hormonal fluctuations (Brinton et al., 2015; Burger et al., 1998). Perimenopausal women
not only undergo changes in the reproductive system but neurological alterations are also
observed (Genazzani et al., 2005; Greendale et al., 2010; Maki et al., 2008).
Decline in levels of estrogen around menopausal transition are associated with declines
in cognitive function as well as receptor expression alterations (Arimoto et al., 2013; Paris
et al., 2011; Sherwin, 1994, 2003; Woods et al., 2000). It has also been shown that this
cognitive decline due to change in the hormonal environment is age independent (Berent-
Spillson et al., 2012) but dependent on endocrine status (Weber et al., 2014). A clinical
review published in 2016 concluded that while cognitive decline is attributable to
menopause, HRT is not recommended as treatment (Maki and Henderson, 2016).
A large number of women will be undergoing menopausal transition in the next decade,
with estimates of up to 1.1 billion women (Shifren et al., 2014) and these women will be
vulnerable to neurodegenerative diseases such as AD. Therefore, it is important to
4
understand perimenopause-induced changes in the brain using a translationally valid
model, such as in rodents (Diaz Brinton, 2012; Finch, 2014).
1.4 Estrogen, Blood Brain Barrier and AD
Brain has always been believed to be an immune-privileged site protected by the BBB
from circulating immune cells, until recently when lymphatic vessels were discovered in
the walls of dural sinuses (Louveau et al., 2015). The BBB consists of a physical barrier
of endothelial cells, astrocytes, and pericytes, all of which regulate transport of substances
from general circulation to the brain. Transcellular passage of proteins requires ligand-
specific receptors and transporters, while paracellular passage of cells is strictly limited by
junctional proteins located in the clefts between endothelial cells (Raub et al., 1992;
Wolburg and Lippoldt, 2002).
1.4.1 SEX STEROIDS AND BBB
Sex-steroid dependent changes in BBB have been observed (Bake and Sohrabji, 2004)
which may be due to control of gap junction protein, connexin 43, being under the control
of steroid hormones (Gulinello and Etgen, 2005). 17β-estradiol (or E2) affects tight junction
(TJ) protein, such as occludin, and endothelial cells of blood vessels (Ye et al., 2003)
along with another TJ protein called zonula occludens-1 (ZO-1) (Khan-Dawood et al.,
1996). Recently, a study proved the protective role of estrogen in specifically guarding the
BBB against lymphocytic invasion in a proinflammatory environment (Maggioli et al.,
2016), which would explain previous published observation that repeated estradiol
treatment prevents BBB disruption due to antigenic stimulation (Tomás ‐Camardiel et al.,
2005).
5
Mechanistically estrogens play a major role in maintaining healthy BBB (Fig. 1.1), via
upregulation of PDGF, VWF and TGFBR1, and downregulation of SLC38a3, Aqp4,
SLC13a3 and VCAM1 (Humphreys et al., 2014).
Evidence exists that implies that female gender is at a higher risk for increased BBB
permeability, based on findings from an acute sepsis model (Minami et al., 2002), and
extravasation of Evan’s blue dye in a hypoosmolality model (Oztaş et al., 2000). One study
also found that in CSF expression of IgG increases significantly with increase in age
(Pakulski et al., 2000).
Figure 1.1 Role of estrogen in maintaining a healthy blood brain barrier.
Estradiol (E2) upregulates Pdgf, Vwf, Tgfbr1 and downregulates Slc38a3, Aqp4,
Slc13a3, and Vcam1 to promote pericyte proliferation and maintenance, enhance BBB
adaptability to hypoxia, maintain fluid balance in brain and prevent or reduce
neuroinflammation.
Pdgf=Platelet Derived Growth Factor, Vwf=von Willebrand Factor,
Tgfbr1=Transforming growth factor beta 1, Aqp4=Aquaporin-4, VCAM1=vascular
adhesion molecule 1, SLC38a3 SLC13a3 are both solute carrier proteins.
6
Although the family of estrogens are grouped together, they have specific effect on the
BBB. Artificial estrogen such as ethinyl estradiol was found to elevate BBB permeability
to albumin (Gammal and Zuk, 1980), water (Reid et al., 1983), inulin and sucrose (Ziylan
et al., 1990), whereas endogenously produced estrogen, 17β-estradiol, was found to be
protective in cases of neural injury models, such as injury caused by ischemia (Chi et al.,
2005; Chi et al., 2002), VEGF- (Chi et al., 2004) or 3-nitropropionic acid (Nishino et al.,
1998) in which estradiol treatment resulted in a less leaky BBB. Literature also suggests
that estrogenic neuroprotection is likely mediated via both ERα and ERβ (Naderi et al.,
2015). Thus, the effect of estrogens on the BBB integrity may be compound specific and
dependent on the type of cells that it may affect.
1.4.2 AD AND BBB
In AD, it is found that capillary density (Fischer et al., 1990) and capacity (Bell and Ball,
1986) both change. Also, nerve plexuses adjacent to blood vessels are lost (Scheibel and
Duong, 1988) with an increase in nitric oxide synthase activity (Dorheim et al., 1994) and
lipid peroxidation (Andorn et al., 1998). Post-mortem studies have found blood-borne
proteins from the periphery (pro-thrombin, albumin, immunoglobulins, fibrinogen, and
thrombin) accumulated in the brain of AD patients implying BBB disruption (Fiala et al.,
2002; Hultman et al., 2013; Ryu and McLarnon, 2009; Salloway et al., 2002; Zipser et al.,
2007). D’Andrea has suggested that AD exhibits signs of being an autoimmune disease
where neurons may be targeted after the BBB is disrupted (D’Andrea, 2005) based on
findings that show increase in IgG-positive neurons in the hippocampus and entorhinal
regions of AD brains which was 8 times when compared to brains of non-AD individuals
in similar age range (D’Andrea, 2003). Aβ overproduction in AD due to APP gene mutation
results in amyloid accumulation in the brain and cerebrovasculature, concurrently with
7
BBB disruption (Iadecola, 2004; Jellinger, 2002; Zlokovic, 2005). A pilot study found that
amyloid deposition in cerebral vessels correlates with recognized criteria for AD diagnosis
(Attems and Jellinger, 2004). In the triple transgenic AD mouse model, both the amyloid
deposition and memory deficits are preceded by a disruption of the BBB (Ujiie et al., 2003).
It is evident that estrogen is an important regulator of BBB integrity, thus the decline in sex
steroids, specifically estradiol, in the perimenopausal rats, is likely to negatively affect BBB
integrity. A compromised BBB would increase women’s risk of cognitive decline as seen
in AD as well as susceptibility to autoimmune diseases.
1.5 The Rodent Perimenopausal Model
While both men and women undergo puberty, menopause is unique to women.
Menopausal transition takes years to complete. It begins with subtle changes in
experience which can last for up to 5 years, then proceeds to irregular periods and this
may last for 3 years before the Last Menstrual Period (LMP) and cessation of menses,
ushering in menopause (Nelson, 2008). The years of transition before menopause are
termed as the perimenopausal period. Even though symptoms of menopause do not begin
till mid to late forties, estrogen levels, after peaking between late twenties and early thirties,
begin to decline years before the last menstrual period (Sherman et al., 1976).
1.5.1 DIFFERENT RODENT MODELS OF MENOPAUSE
Studies of reproductive aging in humans are complex (Henderson and Brinton, 2010;
Soares and Maki, 2010; Soules et al., 2001). This complexity can be better understood
using animal models that offer a glimpse into the biology of human reproductive
senescence. Animal models make it easier to explore and understand the multiple parallel
events that occur in organs and organ systems, as well as changes in cellular, molecular,
8
and genomic processes during perimenopausal transition (Bellino, 2000; Bellino and
Wise, 2003; Van Kempen et al., 2011; Walker and Gore, 2011; Walker and Herndon,
2008). Accordingly, the translational validity of animal models is of extreme importance in
order to understand etiopathogenesis of diseases suspected to begin during this period,
identify and develop new drugs, and predict results of therapeutic interventions (Brown,
2012; Jucker, 2010; Nakao et al., 2009; Shineman et al., 2011). Comprehending the
validity of translational animal models of human menopause would allow the scientific
community to better integrate in vitro data, in vivo observations in animal models, evidence
and observations from population-based epidemiological studies, as well as results of
intervention trials (Diaz Brinton, 2012). Three of the most commonly used animal models
to study human reproductive senescence in women are (1) natural reproductive
senescence or ovary-intact model, (2) ovariectomy or surgical menopause model and (3)
using ovotoxins to induce accelerated ovarian failure (Koebele and Bimonte-Nelson,
2016).
Female rodents undergo estrus cycles, akin to humans’ menstrual cycle and these estrus
cycles become irregular in mid-to-old aged rodents before ceasing to operate and these
changes in estrus cyclicity are associated with neuroendocrine changes (Downs and
Wise, 2009; Kermath and Gore, 2012; Wise et al., 2002). Hence, the natural reproductive
senescence model can be used to understand age-related alterations in the brain at
cellular and molecular levels that are part of normal reproductive aging which in turn would
help us develop better therapies to address symptoms associated with aging and
menopause (Koebele and Bimonte-Nelson, 2016), despite some drawbacks of the ovary-
intact model such as differences in hormonal levels and ovarian follicle reserve compared
to humans and inter-animal variability (Burger, 2006; Lu et al., 1979; Timiras et al., 1995).
9
The ovariectomy (OVX) model of surgical menopause entails bilateral excision of the
ovaries (Olson and Bruce, 1986). The OVX model helps explore and understand the
effects of lack of sex hormones on various organs and organ systems as well as delineate
the effect of externally administered steroids on brain and periphery (Koebele and
Bimonte-Nelson, 2016). As far as translational research is concerned, OVX model is quite
different from how most women undergo perimenopausal transition. Also, the age at which
ovaries are excised in rodents has bearing on the results (Chakraborty and Gore, 2004;
Diz-Chaves et al., 2012; Foster et al., 2003). The hormonal profile in the OVX model is
different from the one found in human females (Mayer et al., 2004; Mayer et al., 2002).
While the surgical model of menopause has its own merits in assessing the effects of
specific hormones and drugs on brain and body systems in the absence of steroid
hormones produced endogenously, it is also equally important to understand the
processes of natural reproductive senescence as well as the manner in which exogenous
hormones interact with the organ systems in presence of intact ovaries.
The ovotoxic model of human menopause in rodents is dependent on using a chemical
called 4-vinylcyclohexene diepoxide, or VCD, a metabolite of 4-vinylcyclohexene (Hoyer
et al., 2001; National Toxicology, 1989). Researchers have used this chemical to develop
a rodent model of menopause since the chemical reduced the pool of ovarian follicles that
are not growing but leaves growing follicles unaffected, (Borman et al., 1999; Flaws et al.,
1994; Hirshfield, 1991; Hoyer et al., 2001; Hu et al., 2001a; Hu et al., 2001b; Kao et al.,
1999; Mayer et al., 2004; Mayer et al., 2005; Mayer et al., 2002; Springer et al., 1996a;
Springer et al., 1996b; Springer et al., 1996c) which is the result of the chemical
upregulating pro-apoptotic proteins in affected follicles (Hu et al., 2001a; Hu et al., 2001b;
Springer et al., 1996a; Springer et al., 1996b; Springer et al., 1996c; Van Kempen et al.,
10
2011). Follicular depletion caused by VCD in female rodents results in hormone profiles
similar to ovary-intact menopausal women (Acosta et al., 2009; Acosta et al., 2010; Mayer
et al., 2004; Mayer et al., 2002; Timiras et al., 1995).
Toxic and carcinogenic effects of this compound as well its potential for accumulation in
the brain due to its lipophilic nature are major drawbacks of this model (Diaz Brinton, 2012;
Van Kempen et al., 2011).
While each animal model has its own strengths and weaknesses with regards to human
menopause, natural reproductive senescence in rodents closely mimics many features of
human menopausal transition (Morrison and Baxter, 2012; Walker and Herndon, 2008).
This perimenopausal rodent model may help better understand the biological processes
occurring in humans and subsequently assist in developing effective therapies (Bethea et
al., 2000; Choi et al., 2003; Kaplan et al., 2010). While the perimenopausal transition in
the rodents takes place over a much shorter time period, it shows many features found in
the human natural reproductive senescence such as decline in ovarian follicles,
lengthening of estrus cycles that results in irregular cycling due to steroid hormone
fluctuations, and irregular fertility (Finch et al., 1984; Van Kempen et al., 2011). Cycle
length irregularity is a good proxy for irregular fertility and advancing reproductive
senescence, which begins at approximately 8 months of age (Finch et al., 1984).
1.5.2 THE NATURAL REPRODUCTIVE SENESCENCE OR OVARY-INTACT MODEL
IN PRACTICE
Previously, Brinton lab has conducted studies using the ovary-intact or natural
reproductive senescence model of human menopause in female Sprague-Dawley (SD)
rats (Yin et al., 2015). Young female SD rats have 4-5-day estrus cycle, and these rats
are called regular cyclers (Finch, 2014). 5-8 day long irregular cycles observed at the
11
beginning of reproductive senescence are found in a small percentage of rats around 6
months of age, but the rest of female rats at that age were found to be regular cyclers (Yin
et al., 2015). The percentage of irregular cyclers increase in number until 9-10 months of
age when irregularly cycling rats constitute the majority of female rats in the group at that
age (Yin et al., 2015). At the same time, some female rats no longer undergo estrus cycling
and transition to constant estrus phase of reproductive senescence that continues up to
at least 16 months of age (maximum period under study). Hence, we have five groups of
female rats, 6 month old female rats that cycle regularly (Regular cyclers 6m), 9-10 month
old female rats that cycle regularly (Regular cyclers 9-10m), 9-10 month old female rats
that cycle irregularly (Irregular cyclers 9-10m), 9-10 month old female rats that no longer
cycle (Acyclic 9-10m) and 16 month old female rats that no longer cycle (Acyclic 16m).
Table 1.1 summarizes the different aspects of menopausal transition that can be studied
using the rat model described, while keeping either the age or the endocrinological status
as a constant. Regular cyclers 6m and Regular cyclers 9-10m groups of female rats differ
in age but have similar estrus cycles (regular 4-5-day cycles) and therefore allow the
Table 1.1 Comparisons of the different chronological and endocrinological
groups of female rats.
The first comparison aims to detect the effect of chronological aging before
endocrinological transition (pre-transition). The second and the third comparisons aim
to detect the effect of endocrinological transition to menopause while controlling the
age variable, and the last comparison aims to detect the effect of chronological aging
post-transition.
12
detection of effects of aging in female rats before the occurrence of menopausal transition.
The three groups of 9-10-month-old animals (Regular cyclers 9-10m, Irregular cyclers 9-
10m, Acyclic 9-10m) allow the comparison of the perimenopausal transition stages while
keeping age constant. The last group of animals (Acyclic 16m) helps detect age related
changes in acyclic animals when compared to Acyclic 9-10m group.
13
1.6 Conclusion/Hypothesis
Thus, it can be concluded that women are more susceptible to cognitive decline and AD,
immune dysregulation and even autoimmune diseases later in life to which
perimenopausal transition maybe a contributing factor, and sex-specific steroids, mainly
estrogen, play a major role in maintaining both cognition and healthy immune system. A
systems biology approach to understanding the hippocampal transcriptomic changes
during the perimenopausal transition in female SD rats could elucidate the pathways in
the hippocampus that influence physiologic alterations preceding actual pathologies. A
breached blood brain barrier could make women increasingly more susceptible to noxious
substances from the peripheral circulation and I hypothesize that this breach happens
during the perimenopausal transition. A compromised BBB by itself may not result in
outward manifestation of pathogenesis specific to a particular disease. However,
interaction between a compromised BBB and other injurious stimuli such as aggregates
of A or -synuclein, low-grade chronic inflammation seen in aging or environmental
factors in the presence of genetic susceptibility could further accelerate the erosion of BBB
integrity potentially subjecting females to increased vulnerability to AD or autoimmune
disease states. The breached BBB could potentially be ameliorated by a neurosteroid
such as Allopregnanolone which in pre-clinical studies has resulted in improvement of
cognitive function.
14
Transcriptomic Analysis
2.1 Abstract
In this study I applied a systems biology approach to the hippocampal transcriptome
during the perimenopausal transition in order to determine the underlying mechanisms
that make women susceptible to AD and other degenerative disease during and after the
menopausal transition. Hippocampal transcriptome from six different groups of female rats
was queried using high-throughput RNA sequencing and the data was analyzed using two
different bioinformatic tools. Findings suggested immune dysregulation, compromised
blood brain barrier and downregulation of metabolic and bioenergetic pathways in the
hippocampus during the perimenopausal transition and these findings need to be
examined in greater detail to understand cell-type specific and pathway dependent
alterations in the hippocampus during perimenopause.
2.2 Introduction
Aging hippocampus (Colangelo et al., 2002; Terao et al., 2002), and other parts of the
brain (Lee et al., 2000) exhibit differential expression of genes that are known to have
immune system functions. These changes were known to be sexually dimorphic with
proportionally greater activation of immune system in the female brain (Berchtold et al.,
2008). Also genes that were highly expressed with age appeared to be involved in
activation of T cells and microglia and also included proinflammatory cytokines and
chemokines (Wyss-Coray, 2006). In adult human brain it was found that among the genes
that are most differentially expressed between the two sexes, the most significant genes
seem to have immune functions (Trabzuni et al., 2013). This would explain why immune-
15
related diseases have a gender dependent bias (Fung et al., 2012; McCombe et al., 2009)
such as the bias seen in multiple sclerosis and AD (Amor et al., 2010). Innate immune
system found in the aging hippocampus of female rats was detected to be responsive to
the hormonal environment (Sárvári et al., 2012; Sárvári et al., 2014). As aging proceeds
immune system becomes senescent along with development of age-related long-term
low-grade inflammation which in turn could repeatedly prime microglia, which likely lose
their neuroprotective function and further activate innate and adaptive immune pathways
in the brain (Deleidi et al., 2015). This profile of high-risk of abnormal immune response
associated with aging results in atypical inflammatory cascades which in turn could
increase susceptibility to diseases.
Based on these observations in literature as well as the results of the customized gene
array data obtained and published earlier (Yin et al., 2015) by Brinton lab, led to a more
bioinformatics intensive approach using transcriptomic analysis for an in-depth
understanding of the systems biology of the aging female hippocampus during the
perimenopausal transition in female SD rats.
2.3 Materials & Methods
2.3.1 ANIMALS
The process by which animals were classified into groups for this experiment is outlined
in detail in the published work by Dr. Fei Yin from Brinton lab (Yin et al., 2015). Briefly, 5
and 8-month-old female Sprague-Dawley (SD) were cycled and their estrus cycle status
was documented every day in the morning using vaginal cytology of smears obtained via
lavage with PBS. Based on age and regularity of estrus cycling animals were classified
into Regular and Irregular groups at 6 and 9-10-months of age. Those female rats that did
not cycle and exhibited a constant estrus stage every day for at least nine consecutive
16
days were labeled as Acyclic, which were divided into two groups based on age: 9-10-
months old and 16 months old. Rats that did not meet the predetermined criteria for
classification into groups were not included in the study. An OVX group of female rats was
added which consisted of regularly cycling female rats at 9-10 months of age that were
ovariectomized at 9-10-months of age and euthanized at 10-10.5-months. Thus, I had six
groups in the study based on age and endocrine status: Regular cyclers 6m, Regular
cyclers 9-10 months, Irregular cyclers 9-10 months, Acyclic 9-10 months, Regular OVX
10-10.5 months and Acyclic 16 months.
After anesthesia, we rapidly dissected the brains on ice, dissected out cerebellum and
brain stem and separated the two hemispheres, which were peeled laterally to obtain the
hippocampus from each brain. All dissected tissues were frozen in −80 °C for planned
experiments. PureLink RNA Mini Kit (Invitrogen) was used to isolate RNA from rat
hippocampal tissues. Some of the data obtained from these rats were published in Yin et
al., 2015.
2.3.2 SEQUENCING OF HIPPOCAMPAL RNA
The RNA extracted from the hippocampus of female rats was sent to Active Motif (Active
Motif Inc., Carlsbad) for sequencing and raw reads in the form of FASTQ files were
obtained. Before sequencing Active Motif analyzed RNA quantity and quality using Qubit
RNA IQ Assay (Thermo Fisher, MA) and all samples exhibited high RNA quality. RNA
from 36 female rats that belonged to the six groups mentioned earlier were sequenced,
with six rats per group. Paired end sequencing of hippocampal RNA was carried out with
read length of 50 base pairs and a read depth of about 50 million reads per sample using
Illumina HiSeq 2500.
17
2.3.2.1 The RNA-Seq process
The comprehensive set of transcripts in a cell i.e. the transcriptome, sheds light on the
processes occurring in the cell and can help understand the physiology and pathology of
various cellular function and disease states. (Wang et al., 2009). Transcriptomic studies
via microarrays and Sanger sequencing both have drawbacks. (Boguski et al., 1994;
Gerhard et al., 2004; Kodzius et al., 2006; Okoniewski and Miller, 2006; Royce et al., 2007;
Wang et al., 2009). RNA-Seq can detect previously unknown transcripts (Vera et al.,
2008), provide data on Single Nucleotide Polymorphisms (SNPs) (Cloonan et al., 2008;
Morin et al., 2008), shows very low background signal (Wang et al., 2009) and possesses
high level of accuracy and reproducibility (Cloonan et al., 2008; Mortazavi et al., 2008;
Nagalakshmi et al., 2008) and 1% error rate (Dohm et al., 2008). Isolated total RNA is first
selected or enriched for mRNA or other RNAs such as long non-coding RNA, and then
converted to cDNA, which is then sequenced through repeated rounds of pre-determined
steps such as incorporation of nucleotide base, washing, imaging and severance, and
fluorescent imaging is used for detection of the incorporated nucleotide (Bentley et al.,
2008; Guo et al., 2008; Wang et al., 2009). The Illumina HiSeq 2500 was used to sequence
RNA for this study.
2.3.3 ANALYSIS OF SEQUENCED HIPPOCAMPAL RNA
I first trimmed raw reads (files with extension .fastq) from both ends using a Phred score
threshold of 20 and read length of 25 in Partek Flow version 5.0. The Partek Flow software
suite is available through the USC Norris Medical Library Bioinformatics Services
(http://norris.usc.libguides.com/nml-bioinfo). I mapped the reads using TopHat2 to the
Rn6 rat reference genome (Ensemble 80) and quantified to transcriptome using Cufflinks.
Only annotated transcripts were quantified, and I used bias correction and multi-read
18
correction along with upper quartile normalization to generate deferentially expressed
genes (DEG) list for each comparison group. The development of this bioinformatics
pipeline is discussed in detail in the results section. The DEG lists for each comparison
group was then uploaded into Ingenuity Pathway Analysis (IPA) software and Protein
ANalysis THrough Evolutionary Relationships (PANTHER) to explore differential
regulation of critical pathways during the perimenopausal transition in the hippocampus.
2.3.4 STATISTICAL CONSIDERATIONS
Differential gene expression data obtained through gene arrays and RNA-Seq methods
necessitate a rethink of statistical considerations. A simple statistical cut-off of p<0.05 for
significantly differentially expressed genes does not suffice due to the errors introduced
by multiple comparisons which essentially includes thousands or even tens of thousands
of genes. For this purpose, I used a cut-off of FDR <0.05 for individual genes of statistical
significance. FDR (also termed ‘q’) is defined as the expected proportion of type I errors
or false positives (Colquhoun, 2014). At the same time, I used FDR<0.2 while carrying out
analyses of biological pathways. (FDR=False Discovery Rate=q)
2.3.5 CHOICE OF SOFTWARE FOR PATHWAY & NETWORK ANALYSIS
At the end of many experiments that result in large amount of data of protein expression,
gene expression or metabolic data, the next question is how to obtain meaningful and
biologically relevant analyses from these datasets. The abundance of transcriptomic data
that resulted from the RNA-Seq required a way to analyze and interpret it. Annotation
enrichment, also called pathway analysis (Curtis et al., 2005), allows us to understand the
biological underpinnings of differences in gene expression related to experimental
conditions. Researchers have mapped genes and proteins to their associated expression,
19
structure and functions and have divided each of these parameters into three domains:
cellular component, molecular function and biological process and then provided a name
(term) and a unique numeric identifier for each such term called Gene Ontology (GO) term
(Ashburner et al., 2000). The presence of these GO terms in the experimental data with
higher than background frequency is called “enrichment” (Huang et al., 2008a). It is
important to note that GO is not complete and it’s a work currently in progress
(Baumgartner Jr et al., 2007; Lewis, 2017). Therefore, its annotations could be biased
towards well-studied genes and well-studied diseases (Alterovitz et al., 2006; Young et
al., 2010).
Functional tools for enrichment such as Database for Annotation, Visualization and
Integrated Discovery i.e. DAVID (Dennis et al., 2003; Huang et al., 2008b), look at
relationships between different terms and integrate annotation terms from various different
sources and lower the redundancy which provides a more biology centric rather than gene
centric approach (Huang et al., 2008a; Khatri and Drăghici, 2005).
From a list of various excellent functional enrichment tools available (Ramanan et al.,
2012), the pathway annotation tools I considered for my RNA-Seq data were IPA and
PANTHER. Using the gene-expression dataset from my RNA-Seq study, I wished to
explore the biological pathways and functions represented by the transcriptomic data as
well as observe the predicted upstream regulators and diseases processes.
The causal network underlying IPA algorithms is based on the Ingenuity Knowledge Base
(Krämer et al., 2013), with extensive findings manually curated as well as a huge number
of nodes and molecules.
I also considered Cytoscape and MetaCore. However, there are both advantages and
disadvantages of using Cytoscape versus IPA. While IPA is easier to learn than
20
Cytoscape, and learning curve is subjective; it provides less information about the
‘networks’ (Thomas and Bonchev, 2010). Cytoscape lets researchers carry out the
statistical assessment of network properties based on graph theory, while IPA only
provides graphical representation of biological pathways as networks of interacting genes
and emphasizes predominantly the biological nature of the pathways and not their graph
theoretical properties (Thomas and Bonchev, 2010). Therefore, I considered IPA a good
resource for specific molecular/cellular pathways based on known biological interactions,
as opposed to the more theoretical approach to network analysis provided by Cytoscape.
MetaCore requires a paid subscription and unfortunately not procured by University of
Southern California, unlike IPA. A study that compared 10 different pathway databases,
including IPA and MetaCore, concluded that different pathway analysis tools and software
programs will provide slightly different results and therefore biological relevance and
cross-validation with existing data are prerequisites for choosing a pathway database or
tool (Shmelkov et al., 2011). In order to retain a more unbiased and holistic view of my
dataset, I also utilized PANTHER. IPA functional enrichment assesses for pathways based
on IPA’s proprietary ‘knowledge base’. This knowledge base is based upon manually-
curated descriptions of gene-gene (or protein-protein) interactions from the literature and
IPA is therefore able to provide gene interaction pathways and upstream regulators.
PANTHER on the other hand assesses for functional enrichment into known GO
categories which are publicly accessible (Mi et al., 2013). Using both IPA and PANTHER
thus provided curated knowledge from two completely different unrelated sources.
According to authors of a published study that evaluated multiple different analysis tools,
researchers should use at least two different analysis tools or software and then base their
21
conclusions on the common findings from the two different tools (Thomas and Bonchev,
2010).
In brief, I chose the analysis pipeline for RNA-Seq data that delivered results most
consistent with our previously published data (Yin et al., 2015), which is in agreement with
a recent review on multiple sources of bias that could confound functional enrichment
analysis of -omics data (Timmons et al., 2015).
2.4 Results
2.4.1 DEVELOPING RNA-SEQ ANALYISIS PIPELINE
It was initially challenging to develop an analysis pipeline for the RNA-Seq data as we
possessed previously published data from the perimenopausal rat model and our goal
Figure 2.1 Flow chart of data analysis for RNA-Seq bioinformatics pipeline.
RNA-Seq was carried out by Active Motif but their bioinformatics pipeline did not meet
our needs. I developed a customized data analysis pipeline for analyzing the data from
the hippocampal RNA-Seq based on two factors: cross-validation with published
biologically releavnt data and using latest versions of well-referenced bioinformatic
tools.
22
was data consistency and cross-validation. The sequencing of RNA was carried out by
Active Motif (AM) and the company provided us with a list of differentially expressed genes
(DEG) for each comparison between the six groups. However, the analysis pipeline
carried out by the company to obtain the DEG was standardized, detected novel
transcripts by default and provided 26 statistically significant genes (FDR<0.05). I
therefore developed a customized fine-tuned workflow with more updated gene annotation
to produce biologically relevant results compared to those from AM’s generic pipeline.
I followed a method of trial and error in analyzing the RNA-Seq data from raw reads stage
(.fastq files) to establish an analysis pipeline that provided biologically consistent results.
With that goal in mind I attempted multiple permutations and combinations to construct an
analysis pipeline that delivered results consistent with published data.
Each of my analysis focused on a single DEG list obtained when comparing two groups:
Irregular 9-10m and
Regular 9-10m in order
to evaluate multiple
pipelines for the
identical comparison.
For the process of
alignment of raw reads
I used three different
options available to
me: using aligned files
(.bam) provided by AM
using TopHat, aligning
Figure 2.2 Developing RNA-Seq data analysis pipeline:
First iteration.
In the first iteration, multiple different permutations and
combinations of two aligner tools, two sets of rat reference
genomes and two different tools for transcript quantification
were conducted to select the most appropriate bioinformatic
pipeline.
23
raw reads myself using an aligner
called TopHat2 (Kim et al., 2013a)
and aligning reads using a
different aligner called Star2. At
the next step, aligned raw reads
need to be quantified and
statistical analysis must be carried
out to generate DEG, for which I decided to use both Cufflinks (Trapnell et al., 2013;
Trapnell et al., 2010) and Gene Specific Analysis (GSA). Cufflinks has its own tools for
statistical analysis called Cuffmerge and Cuffdiff (Trapnell et al., 2012) to generate lists of
DEG and differentially expressed transcripts (DET) after transcript quantification.
However, GSA is a statistical analysis tool and must be used in conjunction with Partek
E/M (Partek Flow's optimization of the expectation-maximization algorithm) for transcript
quantification. I also found that using the right annotation for the rat reference genome
could vary results. I considered two annotation tools, RefSeq and Ensembl, but ultimately
decided in favor of Ensembl since Ensembl annotates many more genes than RefSeq
(Zhao and Zhang, 2015). Based on the initial evaluation of different pipelines (Figure 2.2
– screenshot of all pipelines from IPA) I determined that for RNA quantification and DEG
analyses, Cufflinks + Cuffdiff and Partek E/M + GSA were the most appropriate methods,
but I was still considering both aligners Star2 and TopHat2. When all other parameters
were the same, Partek E/M + GSA produced 30% fewer DEG than Cufflinks. About 60%
of the former overlapped with the latter, but less than 35% of the latter overlapped with
the former. When all other parameters were the same, the choice of different alignment
methods (TopHat2 vs Star2) had limited impact on the DEG results as over 80% of DEG
Figure 2.3 Developing RNA-Seq data analysis
pipeline: Second iteration.
After narrowing down the number of analysis
pipelines to five, in the second iteration results
from all five pipelines were cross-validated with
published data from the perimenopausal model.
24
overlapped. Based on these results, I narrowed down the bioinformatics pipeline to five
(Figure 2.3 – screenshot of all pipelines from IPA) and used the newest available version
of the annotation tool Ensembl (Ensembl 80). Next, I looked at functional data from all five
pipelines in IPA which included canonical pathways, upstream regulators and disease and
functions (Figures 2.4 and 2.5 – images exported from IPA).
Figure 2.4 Functional analyses results of the second iteration with five DEG lists.
As seen in this image exported from IPA, the final five bioinformatic analysis pipelines were
tested for functional data in IPA cross-validated with published literature on
perimenopausal model. Overall, they exhibit similarity, especially for the top ranked items
in canonical pathways and upstream regulators (ranked by p-value of overlap).
25
Figure 2.5 Functional data used to develop bioinformatics pipeline.
As seen in this image exported from IPA, diease and biofunctions, and upstream
regulators obtained using five DEG lists of second iteration used to develop
bioinformatics analysis pipeline. Overall, they exhibit similarity.
26
To help determine which DEG list makes the most sense, I created 20 gene lists for the
key functional concepts mentioned in Yin et al., 2015 and conducted the functional
analyses considering these gene lists (Fig 2.6 – image exported from IPA). Based on this
evaluation, the TP2_rn6Ensembl80_Cufflinks_nonovel was found to be the best method
with biologically relevant results and was chosen for all subsequent analysis. The novel
Figure 2.6 RNA-Seq data analysis: Cross validation with pathways published
previously.
As seen in this image exported from IPA, final five bioinformatics pipelines under
consideration were cross-validated with previously published findings from Yin et al.,
2015.
27
transcript option does not require any annotation reference, it will carry out de novo
assembly to reconstruct transcripts and estimate their abundance. I instead chose the
non-novel (or annotated transcripts) option for transcript quantification when running
Cufflinks to focus on mapping known genes and transcripts.
For additional cross-validation, the 8 down-regulated genes (Atpaf2, Esrra, Nfkb2, Il1rl1,
Mapk3, Plcb3, Thop1, Map2) reported in Yin et al., 2015 were checked against DEG
results from the chosen method and except Map2, all other genes were found to be
consistently down-regulated, similar to published findings. Map2 has three different
isoforms, only one of which was down-regulated. This bioinformatic pipeline was used to
analyze the hypothalamic RNA-Seq data published this year (Bacon et al., 2018).
2.4.2 MAPPED TRANSCRIPTS FOR EACH COMPARISON GROUP
The Differentially Expressed Gene (DEG) list generated by Cufflinks enumerated more
than 32,000 transcripts out of which 20,493 were mapped to genes in the rat genome
within IPA. Only a tiny fraction (ranging from 0.27% - 1.93%) of the total number of
hippocampal genes mapped by IPA to the rat reference genome met the FDR<0.05 cut-
off for significantly up- or down regulated genes in all of the one-on-one comparison
between the six groups (Figure 2.7). Very few genes were significantly up- or down-
regulated when female SD rats transitioned from Irregular cyclers 9-10m to Acyclic 9-10m
(31 upregulated and 25 downregulated genes), whereas more than 650 genes were
significantly changed between Regular cyclers 9-10m and Acyclic 9-10m (493 upregulated
and 176 downregulated). Almost 400 genes were significantly differentially expressed
between Acyclic 16m and Regular cyclers 6m (285 upregulated and 103 downregulated)
and Irregular 9-10m and Acyclic 16m (219 upregulated and 177 downregulated). Upwards
of 250 genes were differentially expressed between Regular 9-10m and Regular 6m (140
28
upregulated and 116 downregulated) as well as between Regular 9-10m and Irregular 9-
10m (153 upregulated and 113 downregulated).
When hippocampal gene expression was compared between Regular cyclers 9-10m and
Regular Ovariectomized 10-10.5m, the highest number of significantly differentially
expressed genes (more than 900 genes total) were observed (Fig. 2.8). Abrupt removal
of ovaries and cessation of a functional HPA axis in female rats has demonstrable impact
on the hippocampus as seen from the differential regulation of hippocampal gene
expression.
Figure 2.7 Number of genes significantly up- and down-regulated between
different perimenopausal groups.
Excluding the surgical menopausal group of Regular OVX 10-10.5m, the highest number
of genes changed significantly between groups involved in intital chronological and
endocrinological aging i.e. Regular cyclers 9-10m vs. Regular cyclers 6m and Irregular
cyclers 9-10m vs. Regular cyclers 9-10m (cut-off: q=FDR= False Discovery Rate<0.05).
29
2.4.3 FUNCTIONAL CLASSIFICATION OF DIFFERENTIALLY EXPRESSED GENES
OF FIRST ENDOCRINE TRANSITION
Regular 9-10m vs. RegOVX 10-10.5m
Mapped transcripts
Unmapped transcripts
Genes downregulated
(q<0.05)
Genes upregulated
(q<0.05)
516
386
11968
20493
Figure 2.8
Figure 2.8 Number of hippocampal genes up- and down-regulated due to ovariectomy.
Female SD rats that exhibited regular estrus cycles had their ovaries removed at 9 months and
were sacrificed at 10-10.5 months to determine the effect ovariectomy had on hippocampal
gene expression.
Figure 2.9a Number of genes in each category
30
Previously our lab has established that an important period during the entire
perimenopausal transition was the progression from regular to irregular estrus cycling at
9-10 months of age in
female SD rats (Yin et al.,
2015). The sequencing of
hippocampal genes found
that 266 genes that were
significantly differentially
regulated during this
transition period
(FDR<0.05) out of which
150 genes had known
functions in the NCBI
database and were
classified into functional groups (metabolic, neurological, cell adhesion and membrane
Figure 2.9 Functional classification of genes.
This figure illustrates the functional classification of significantly up- and down-regulated
genes (number of genes in each category), along with their fold change, for the first
endocrinological transition from Regular cyclers 9-10m to Irregular cyclers 9-10m
(FDR<0.05).
Figure 2.9b
Figure
2.10a
31
trafficking, inflammation and Alzheimer’s disease). The genes in these functional groups
are shown in Figure 2.9a-b along with fold change for each group. From the above data
(Fig. 2.9) it is evident that at 9-10 months of age, the female rat hippocampus experiences
significant changes in expression of certain genes during two transitions. However,
expression of most numbers of genes is significantly altered during the first endocrine
transition from Regular cyclers 9-10m to Irregular cyclers 9-10m (total 266 genes
significantly differentially expressed) compared to the second endocrine transition from
Irregular 9-10m to Acyclic 9-10m (total 56 genes significantly differentially expressed),
whereas expression of
a total of 669 genes is
significantly altered
when Regular cyclers
9-10m are compared to
Acyclic 9-10m. These
observations and data
presented by our lab
previously (Yin et al.,
2015) suggests that the
Regular to Irregular
endocrine transition
likely has significant impact on hippocampal gene expression in perimenopausal female
rats.
Figure
2.10b
32
As an example of a neurological gene
significantly differentially expressed
during Regular 9-10 to Irregular 9-10m
transition, consider uncoupling protein 2
(UCP2). UCP2 is found in the membrane of
mitochondria that serves to disconnect
oxidative phosphorylation from ATP
synthesis furnishing neuroprotection (Sans
et al., 2000). In perimenopausal female rats,
hippocampal transcriptome shows
statistically significant (FDR<0.05)
upregulation of UCP2 during the Regular 6
to Regular 9-month transition but its
expression is downregulated during the
Regular to Irregular transition (first endocrine
transition) and from Regular to Acyclic
transition (Figs. 2.10a-c).
Figure 2.10a-c UCP2 expression in
the hippocampal transcriptome of
perimenopausal rats.
A neuroprotective gene, UCP2, is
significantly upregulated during the
Regular 6 to Regular 9-10m transition
but downregulated during subsequent
endocrine transition from Regular 9-10
to Irregular 9-10m, as well as from
Regular 9-10 to Acyclic 9-10m.
Figure
2.10c
33
2.4.4 PATHWAY ANALYSIS USING IPA AND PANTHER
Pathway analysis was performed using IPA and PANTHER. For pathway analysis all the
genes that were differentially expressed at FDR<0.2 were included in order to obtain
meaningful data using the pathway analysis tools. In figure 2.11a-b, the three different
colors represent the number of genes in a pathway (in blue), percent of genes out of the
total number of significant genes in a pathway (in orange) and percent of genes to the total
number of pathways (in green). Comparing the hippocampal transcriptome of Regular
cyclers 6m to Regular cyclers 9-10m suggested that the genes that were significantly
altered between these two groups were associated with PDGF signaling pathway which
is important for BBB permeability, AD (presenilin and amyloid secretase pathways) and
proinflammatory cytokines pathways.
Comparing significantly differentially regulated genes between Irregular cyclers 9-10m
and Regular cyclers 9-10m, using PANTHER it was discovered that this set of genes were
associated with T cell activation and AD amyloid secretase pathways (Figures 2.12a-b),
whereas comparing significantly differentially regulated genes between Regular OVX 10-
10.5m and Regular cyclers 9-10m (Fig. 2.13a-b) found these set of genes to be associated
with T cell activation and two AD pathways as well. While surgical menopause (OVX
group) and transition group (Irregular 9-10m) to natural reproductive senescence both
show altered expression of genes, the absolute quantity of genes significantly differentially
expressed in the OVX group support the abrupt nature of menopause achieved by surgical
removal of ovaries. At the same time similar pathways are altered in the hippocampus
between both Regular to Irregular perimenopausal transition at 9-10m as well as the
Regular to Regular OVX surgically achieved abrupt menopause: T cell activation and AD
pathways.
34
Percent of gene hit against total #
pathway hits
Percent of gene hit against total #
genes
Number of genes
Figure 2.11a
35
Figure 2.11a-b PANTHER classification system: Regular cyclers 6m vs. Regular cyclers 9-10m.
The significance cut-off used for genes was FDR<0.2. When genes that were significantly differentially expressed in hippocampi of
Regular cyclers 6m and Regular cyclers 9-10m female rats were compared, genes that were significantly altered between these two
groups were associated with Platelet Derived Growth Factor (PDGF) signaling pathway which is important for BBB permeability,
Alzheimer’s disease (presenelin and amyloid secretase pathways) and proinflammatory cytokines pathways.
BBB permeability
AD pathways:
• Amyloid-secretase
pathway
• Presenilin pathway
Pro-inflammatory cytokines
Percent of gene hit against total #
pathway hits
Percent of gene hit against total #
genes
Number of genes
Figure 2.11b
36
Percent of gene hit against total #
pathway hits
Percent of gene hit against total #
genes
Number of genes
Figure 2.12a
37
Percent of gene hit against total #
pathway hits
Percent of gene hit against total #
genes
Number of genes
T cell activation
AD Amyloid-secretase pathway
(300+ genes)
Figure 2.12b
Figure 2.12a-b PANTHER classification system: Regular cyclers 9-10m vs. Irregular cyclers 9-10m.
The significance cut-off used for genes was FDR<0.2. When genes that were significantly differentially expressed in hippocampi
of Irregular cyclers 9-10m and Regular cyclers 9-10m female rats were compared, genes that were significantly altered between
these two groups were associated with were T cell activation and AD amyloid secretase pathways.
38
Figure 2.13a
Percent of gene hit against total #
pathway hits
Percent of gene hit against total #
genes
Number of genes
39
Percent of gene hit against total #
Pathway hits
Percent of gene hit against total #
genes
Number of genes
T cell activation
AD pathways:
• Amyloid-secretase
pathway
• Presenilin pathway
BBB permeability
Figure 2.13b
Figure 2.13a-b PANTHER classification system: Regular OVX 10-10.5m vs. Regular cyclers 9-10m.
The significance cut-off used for genes was FDR<0.2. When genes that were significantly differentially expressed in
hippocampi of Regular OVX 10-10.5m and Regular cyclers 9-10m female rats were compared, genes that were significantly
altered between these two groups were associated with were T cell activation and two AD pathways.
40
shown).
Figure 2.14a
41
Additionally, IPA analysis also showed upregulation of functions such as T cell
development and development of lymphocytes, and downregulation of apoptosis of
Figure 2.14b
Figure 2.14a-b Up-regulation of T lymphocyte related process: Irregular 9-10m vs.
Regular 9-10m.
When genes that were significantly differentially expressed (FDR<0.2) in hippocampi of
Irregular cyclers 9-10m and Regular cyclers 9-10m female rats were analyzed by IPA it was
found that this set of genes signaled an upregulation of functions such as T cell development
and development of lymphocytes, and downregulation of apoptosis of lymphocytes and
leukocytes.
42
lymphocytes and leukocytes in the hippocampus (Fig. 2.14a-b) during the Regular to
Irregular transition at 9-10 months of age.
Reg9-10m vs. Reg6m Acy9m vs. Irreg9m Irreg9-10m vs. Reg9-10m
Irreg9m vs. Reg9m Acy16m vs. Acy9-10m Acy9-10m vs. Irreg9-10m
Figure 2.15
Figure 2.15 Differential regulation of interleukin (IL)-27 between different
groups of female SD rats.
Analysis of hippocampal RNA-Seq using IPA suggests that IL-27 level increases
from Regular cyclers 6m to Regular cyclers 9-10m and decreases thereafter
through Irregular and Acyclic 9-10m groups, and then shows an increase in the
Acyclic 16m group. IL-27 is an immunomodulatory cytokine with wide-ranging
effects on all subsets of T lymphocytes
Upregulation Downregulation
43
IL-27 is an immunomodulatory cytokine with wide-ranging effects on all subsets of T
Figure 2.16 Insulin signaling pathway in the hippocampal transcriptome
during the Regular to Irregular transition at 9-10 months of age in female SD
rats.
Insulin signaling pathway is downregulated in the hippocampus during
perimenopausal transition from Regular to Irregular stage in 9-10 months old female
SD rats.
Figure 2.16
Upregulation
Downregulation
44
lymphocytes (Yoshida and Hunter, 2015). IL-27 appears to be downregulated in the
hippocampus during the two endocrinological transition stages and is elevated before and
after the perimenopausal transition (Fig. 2.15); its downregulation may explain
dysregulated T cell functions predicted in the hippocampus during perimenopausal
transition shown by analysis of hippocampal transcriptome using IPA and PANTHER.
Besides the immune system and blood brain barrier permeability, it was also discovered
that the perimenopausal transition affected the bioenergetic and metabolic pathways in
the hippocampus, implying reduced utilization of glucose in the brain (Figures 2.16 and
2.17) evidence by downregulation of AKT and ERK.
Figure 2.17
Upregulation
Downregulation
Figure 2.17 Downregulation of mitochondrial complexes of the electron transport
chian as well as Carnitine palmitoyltransferase I in the hippocampal
transcriptome during the Regular to Irregular transition at 9-10 months of age in
female SD rats.
Analysis of the hippocampal transcriptome using IPA suggests downregulation of
almost all the proteins and complexes of the electron transport chain, with only complex
II unaffected and complex IV mildly downregulated. Additionally, carnitine
palmitoyltransferase I, essential in the beta-oxidation of long chain fatty acids, too
appears to be downregulated. These findings together suggest reduced bioenergetic
function in the hippocampus during the transition from Regular to Irregular stage at 9-
10m of age in female SD rats.
45
Increased ROS, which possibly could be a result of downregulated UCP2 (Fig. 2.10a-c),
could potentially explain downregulation of mitochondrial complexes in the hippocampus
seen in Figure 2.17 and ROS can also inhibit Carnitine palmitoyltransferase I (CPT1)
enzymatic activity (Setoyama et al., 2013). All these confirm the deleterious effect of ROS,
which could be the result of statistically significantly reduced expression of UCP2, in the
downregulation of fatty acid oxidation via inhibition of CPT1. In the mitochondria (Fig. 2.17)
Figure 2.18 Downregulation of PDGF signaling pathway during the
perimenopausal transition.
Transcriptomic expression obtained through RNA-Seq demonstrates
downregulation of PDGF signaling pathway during the endocrinological transition
from Regular 9-10-month to Irregular 9-10-month.
Upregulation
Downregulation
Figure 2.18
46
almost all the proteins and complexes of the electron transport chain are predicted to be
downregulated with only complex II unaffected and complex IV mildly downregulated.
Figure 2.19 TGFB1 gene network showing downregulation of PDGF, solute
carriers and integrin binding proteins in the hippocampal transcriptome during
the perimenopausal transition.
Transcriptomic expression obtained through RNA-Seq demonstrates downregulation
of TGFB1 gene, which is a central node in the network accompanied by downregulation
of PDGF, solute carriers, integrin binding proteins and pro-apoptotic proteins during
the endocrinological transition from Regular 9-10-month to Irregular 9-10-month in the
hippocampal transcriptome.
Upregulation
Downregulation
Figure 2.19
47
Moreover, the insulin signaling pathway, which is an important metabolic pathway is
downregulated with predicted reduction in transcription during Regular to Irregular
transition (Fig. 2.16), which could be due to chronic oxidative stress induced by decreased
UCP2 that impedes the function of ROS in insulin secretion (Diano and Horvath, 2012).
PDGF signaling pathway (Fig. 2.18), an important pathway with respect to blood brain
barrier, appears to be downregulated in the hippocampal transcriptome during the first
endocrine transition, but PDGFRα is upregulated. According to IPA, PDGFB is part of a
network of genes with Transforming growth factor beta-1 (TGFB1) as a central node (Fig.
2.19). During transition from Regular to Irregular stage, TGFB1 is downregulated in the
hippocampal transcriptome which seems to downregulate PDGF, transport proteins such
as SLC7A5, AQP9, SLC4A2 (via LOXL1), integrin binding proteins such as SEMA7A,
SEMA4A and SOD3, member of the superoxide dismutase family. Downregulation of
TGFB1 and other associated proteins during the Regular to Irregular transition in the
hippocampus indicates reduced cell growth and proliferation which may directly impact
the blood brain barrier integrity. Integrins, seen downregulated in Fig. 2.19, serve to keep
the extra-cellular matrix adherent and their downregulation suggests reduced attachment
of cells to the ECM which in the context of BBB coupled with reduced PDGF and TGFB1
can be interpreted to indicate downregulation of pericytes and endothelial cells leading to
lower cell growth and survival and cellular differentiation. On the other hand,
downregulation of BCL2L1 and BAD indicate reduced cell apoptosis since BAD
downregulation reduces proapoptotic activity. SOD3 downregulation impairs the ability to
neutralize superoxide free radicals. Taken together these results suggest reduced growth
and cell proliferation, reduced apoptosis and decreased integrity of ECM in the
hippocampus which could result in compromised BBB.
48
2.5 Discussion
Perimenopausal transition is known to increase risk of cognitive decline, AD and immune
dysregulation as discussed earlier. A systems biology study of the perimenopausal
hippocampal transcriptome was carried out to understand transcript-level changes in the
hippocampi of female SD rats during chronological and endocrinological aging. The data
in this study were obtained from the sequencing of messenger RNA (protein-coding RNA)
extracted from homogenized hippocampi of female rats in various stages of chronological
and endocrinological transition. These data represent transcriptomic changes in the
hippocampus as a whole, and consequently it is challenging to attribute transcriptomic
alterations to specific cell types found in the hippocampus. The complexity of the data
allows only a broad overview of the alterations in hippocampal transcriptome. In this
section I discuss the general findings of the RNA-Seq and their interpretations in light of
published literature.
Differentially expressed genes obtained through RNA-Seq must be analyzed with respect
to specific statistical cut-offs. A standard cut-off of p<0.05 will result in a large number of
false positives due to the tens of thousands of genes being involved in the study.
Therefore, the most stringent cut-off of FDR<0.05 was used that guarantees less than five
percent false positive genes among the list of genes deemed to be significant. For each
comparison, only those genes that met this stringent cut-off were termed as significantly
differentially expressed genes.
From the results it is evident that both age and endocrine changes play a significant role
on gene expression in the hippocampus of female rats. The first aging and endocrine
transitions - Regular cyclers 6m to Regular cyclers 9-10m and Regular cyclers 9-10m to
49
Irregular cyclers 9-10m respectively show similar number of genes significantly altered
(FDR<0.05). The second endocrine transition from Irregular 9-10m to Acyclic 9-10m has
very few genes significantly altered. This phenomenon demonstrates that hippocampal
transcriptomic expression alters the most from the age of 6 to 9 months in female rats and
again when they transition from Regular to Irregular estrus cycling status at 9-10m, but
very few genes are significantly differentially expressed during the transition from Irregular
to Acyclic 9-10m, during which the perimenopausal transition is almost over and
menopause begins.
However, the greatest number of genes significantly differentially expressed in the
hippocampi are observed when Regular 9-10m and Regular OVX 10-10.5m are
compared. This reinforces the belief that sudden loss of ovarian steroids (such as during
surgical menopause) could result in dramatic alterations in the hippocampus based on
changes observed in the hippocampal transcriptome of perimenopausal female rats.
Surgical menopause is not representative of the majority of human females that undergo
natural reproductive senescence and abrupt cessation of ovarian hormones in rodent
females is dissimilar to the reproductive senescence experienced by the postmenopausal
ovary in humans (Mayer et al., 2004; Mayer et al., 2002). Data from the RNA-Seq also
supports natural reproductive senescence model as the least disruptive model of
perimenopause. At the same time similar pathways are altered between both Regular to
Irregular perimenopausal transition at 9-10m as well as the Regular 9-10m to Regular
OVX 10-10.5m surgically achieved abrupt menopause: T cell activation and AD pathways
as predicted by PANTHER, which supports the idea that changes happening during
menopause could influence the immune system and risk of AD.
50
Based on the concurrence data published previously by our lab (Yin et al., 2015) and
currently presented data analysis of the hippocampal transcriptome, it is likely that the
Regular to Irregular endocrine transition at 9-10 months of age is a critical period in the
natural rodent perimenopause. Functional classification of genes significantly differentially
expressed during the Regular to Irregular transition at 9-10 months of age in female rat
hippocampus revealed that majority of genes (30%) performed functions that I could
classify as the Neurological genes category, 22% were in Alzheimer’s disease category
and the rest are in Metabolic, Inflammation and Cell adhesion & membrane trafficking
category, in a descending order of the number of genes. These findings uncover the extent
to which hippocampal transcriptome diverges during the Regular to Irregular transition in
female rats with significant implications for cognitive decline, reduction in bioenergetic
gene expression and suboptimal mitochondrial function demonstrated earlier (Yin et al.,
2015). An example to support this is the expression of UCP2 during the perimenopausal
transition. Absence of UCP2 results in deficits in procedural memory and spatial cognition
(Hermes et al., 2016). UCP2 downregulation has been found to be associated with
increased oxidative stress, atherosclerosis, vascular damage and shorter lifespan in mice
(Andrews and Horvath, 2009; Ma et al., 2010; Ma et al., 2014; Moukdar et al., 2009).
Increased UCP2 reduced ROS and protected endothelial cells against damage from ROS
(Ma et al., 2013; Tian et al., 2012). Neuroprotective function of UCP2 has been
demonstrated both in vitro and in vivo in multiple different studies (Haines et al., 2010;
Hass and Barnstable, 2016; Mattiasson et al., 2003; Normoyle et al., 2015). One group
also demonstrated that UCP2 conferred protection against stroke (Rubattu et al., 2017).
It is therefore worrisome to observe UCP2 downregulation in the hippocampal
transcriptome during endocrine transition in 9-10m old female SD rats, after its
51
upregulation from Regular 6m to Regular 9-10m old rats. However, direct increase in ROS
due to decrease in UCP2 expression was not measured in this study and therefore a
clearer correlation of UCP2 expression with ROS production in the hippocampus in the
perimenopausal model is required. If extrapolatable to humans, this data from hippocampi
of female rats would mean that during perimenopause women may lose the
neuroprotective effect of UCP2 which in turn would increase their risk for
neurodegenerative diseases.
Additionally, IPA analysis of differentially expressed genes between Regular cyclers 9-
10m and Irregular cyclers 9-10m shows a differential regulation of IL-27 transcript, which
is a cytokine (Hunter and Kastelein, 2012b). IL-27 has exhibited both pro- and anti-
inflammatory properties, not only based on cell type it affects but also based on the context
of the entire immune system milieu and type of infection or disease under study (Yoshida
and Hunter, 2015). IL-27 has been shown to both reduce (Huber et al., 2007; Neufert et
al., 2007; Stumhofer et al., 2007) and promote (Hall et al., 2012; Kim et al., 2013b; Moon
et al., 2013) population of regulatory T cells, while it controls proinflammatory T cell
responses of Th2 and Th17 cells (Hunter and Kastelein, 2012a). A neuroprotective role of
IL-27 has been described but it is unclear whether these effects of IL-27 are mediated
locally within the CNS or are CNS manifestations of events in the periphery or both
(Yoshida and Hunter, 2015). Considering the complex role of IL-27 in regulating various
aspects of the immune system, predicted downregulation of IL-27 in the hippocampal
transcriptome during perimenopause in female SD rats suggests dysregulation of the
immune system. IL-27 has been found to be associated with estrogen and progesterone
levels during pregnancy in women and increases throughout gestation, before dropping
dramatically postpartum (Enninga et al., 2015), the significance of which is unclear but this
52
observation hints at a role of sex steroids in IL-27 regulation. Therefore, it is difficult to
ascertain the role of IL-27 during perimenopause in the hippocampus but its
downregulation in the hippocampal transcriptome during the endocrine transition at 9-10m
suggests a role of both innate and adaptive immune system in the hippocampus during
endocrinological aging. Based on these findings in our model of natural reproductive
senescence, future studies exploring immune system interactions during perimenopause
may wish to delineate the role of IL-27 in this endocrine transition.
Insulin signaling pathway and bioenergetic gene expression are both downregulated
during the Regular to Irregular transition in 9-10-month-old female SD rats. These
observations are corroborated by results of the FDG-PET and mitochondrial function our
group published previously in the same model (Yin et al., 2015).
Both IPA and PANTHER predicted downregulation of PDGF signaling (Fig. 2.18) in the
hippocampal transcriptome of female SD rats during the Regular to Irregular transition at
9-10m, which could be in response to reduced cortical estrogen level observed during
perimenopause (Yin et al., 2015) consistent with findings in the literature outlined in
Chapter 1. The isoform of PDGF called PDGF-BB has been shown to be responsible for
maintaining the health of BBB (Zhao et al., 2015; Zlokovic, 2008) and is regulated by
estrogen (Bake and Sohrabji, 2004; Jiang et al., 2018). Additionally, upregulation of
PDGFRα signaling seen concomitantly with PDGF-BB downregulation may contribute to
BBB impairment via p38 MAPK mediated pathway (Ma et al., 2011). Besides PDGF,
estrogen also affects tight junction proteins as mentioned in Chapter 1. Additionally, in the
RNA-Seq data the TGFB1 gene network shows downregulation of solute carriers, integrin
binding proteins, SOD3 and PDGF, as well as reduced expression of pro-apoptotic
proteins, suggesting reduced cell proliferation and reduced apoptosis but at the same time
53
decreased cell survival and maintenance. Since the RNA-Seq results reflect the
transcriptome of homogenized hippocampal tissue, cell-specific pathways and functions
are difficult to elucidate, and it is possible that these conflicting findings reflect pathways
in two or more differing cell types. Ligand–receptor systems such as TGFB and PDGF
have been implicated in maintenance of vascular stability (von Tell et al., 2006) and TGFB
signaling plays a role in differentiation, maturation, proliferation, migration and attachment
of endothelial cells and pericytes (Darland and D'amore, 2001; Maddaluno et al., 2013;
Reyahi et al., 2015; Van Geest et al., 2010), confirmed in vivo in murine models
(Maddaluno et al., 2013). In fact, in an in vitro BBB model it was shown that continuous
TGFB production by pericytes increased the barrier function of BBB and treatment with
anti-TGFB1 antibody inhibited this barrier function (Dohgu et al., 2005). In PDGF deficient
mice lack of pericytes is the main cause of phenotype since endothelial cells in the
capillaries of these mutant mice seem to be unable to attract pericyte progenitor cells
(Crosby et al., 1998; Lindahl et al., 1997) and these mice exhibited morphological sings of
increased vascular permeability (Hellström et al., 2001). Most recently it was discovered
that PDGF-BB signaling in pericytes results in secretion of microvesicles and growth
factors that may have neuroprotective and neurorestorative role (Gaceb et al., 2018).
However, pericytes in a proinflammatory environment produce proinflammatory cytokines
resulting in capillary leakage (Edelman et al., 2007a, b). This can further cause
detachment of pericyte from the basal lamina and demonstrate structural disorganization
(Nishioku et al., 2009). Thus, the proinflammatory environment of perimenopausal
hippocampus (Malutan et al., 2014; Russu and Antonescu, 2018; Yin et al., 2015) may
compromise BBB integrity independently from reduced cortical level of estrogen.
54
Therefore, these findings from the hippocampal transcriptome suggest that during the first
endocrine transition from Regular to Irregular cycling status at 9-10m of age the reduced
cortical levels of estrogen and a proinflammatory environment may result in increased
permeability of BBB likely facilitated by PDGF and TGFB-mediated pathways.
2.6 Conclusion
Perimenopausal transition in the hippocampi of female SD rats is a period of dramatic up-
and down-regulation of genes and pathways involved in blood brain barrier permeability,
immune system regulation, mitochondrial bioenergetics and metabolomic changes. The
most dramatic gene level changes are seen in response to abrupt cessation of ovarian
hormones achieved by surgical menopause. Compromised BBB during perimenopause
could potentially be a major risk factor for women to develop neurodegenerative diseases
and autoimmune states and a functional confirmation of breached BBB during
perimenopause in a rat model would help validate this finding from the systems biology
study.
55
Blood Brain Barrier (BBB) Permeability in Perimenopausal
Female Rats
3.1 Abstract
In this experiment I tested the BBB permeability in female SD rats using sodium
fluorescein. I found that perimenopause is indeed a stage of increased BBB permeability
in female rats and is dependent on endocrine status more than age. I also found that intra-
group variability in blood brain barrier permeability increased with gradual endocrine aging
rather than with chronological aging, and abrupt endocrine aging (surgical menopause)
surprisingly did not lead to appreciable increase in intra-group variability of blood brain
barrier permeability when compared to natural reproductive senescence group of female
rats.
3.2 Introduction
To reiterate, blood brain barrier permeability is increased during aging, neurodegenerative
diseases and injury such as traumatic brain injury. Mechanistically a compromised BBB
requires either altered tight junction proteins, damaged endothelial cells or pericytes or
astrocyte foot processes or a compromised and exposed basement membrane or a
combination of all these processes (Hawkins and Davis, 2005). In females, sex steroids,
mainly estrogen, play a major role in directly and indirectly regulating both tight junction
proteins as well as cellular components of BBB. On the other hand, in males, testosterone
maintains healthy BBB and increased BBB permeability is seen with testosterone-
depletion (Atallah et al., 2017) Supplementation with testosterone ameliorated the effects
of lack of testosterone on BBB permeability (Atallah et al., 2017). Women begin to lose
56
the protective effect of estrogen during the perimenopausal transition and this chapter
explores the effect of loss of estrogenic protection on BBB integrity by measuring the
amount of a fluorescent dye, sodium fluorescein, that extravasates across the BBB in a
rat model of perimenopause.
3.3 Materials & Methods
3.3.1 CHOICE OF MARKER FOR BBB DISRUPTION
The use of Evans blue (EB) dye to measure BBB permeability is quite common. However,
there are certain drawbacks to using EB (Saunders et al., 2015). After EB is injected, not
all of the injected dye binds exclusively to albumin and free dye could enter extravascular
spaces and bind to tissues, thus distorting results. It may be difficult to measure quantity
of EB in some experiments since its excitation and emission wavelengths change in the
presence of proteins (Saunders et al., 2015). Another marker, Horseradish Peroxidase
(HRP), causes an allergic reaction and affects permeability of blood vessels (Majno et al.,
1961) in some commonly used strains of rats (Cotran and Karnovsky, 1967) but not in
others (COTRAN et al., 1968). HRP itself could cause membrane damage (Mazariegos et
al., 1984) which may lead to less accurate measurement of BBB permeability. The third
dye under consideration, sodium fluorescein (NaF) binds to proteins weakly and can serve
as an effective tracer of BBB compromise compared to protein-binding EB (Wolman et al.,
1981). Due to its smaller molecular size compared to EB, measurement of NaF
extravasation across BBB allows detection of subtler changes in BBB integrity versus
radioactive tracers (Kaya and Ahishali, 2011). No NaF induced deleterious effects on BBB
have been reported so far and NaF is considerably less toxic than Evans blue or
Horseradish Peroxidase (Saunders et al., 2015). Based on these observations in
57
literature, I chose NaF as a marker for BBB disruption from among three markers
considered namely EB, NaF and HRP.
3.3.2 DYE INJECTION AND SACRIFICE
I used sodium fluorescein (Sigma-Aldrich) for detecting suspected BBB compromise in
perimenopausal female rats. Utilizing daily ovarian cytology I cycled five and eight months
old female SD rats and classified them into groups based on age and estrus cycles. These
groups were: Regular cyclers 6m, Regular cyclers 9-10 months, Irregular cyclers 9-10
months and Acyclic 9-10 months. Additionally, I ovariectomized some rats in the Regular
cyclers 6m group while some others underwent sham ovariectomy procedure where I
carried out the surgical procedure without dissecting the ovaries, in order to explore the
effect of ovariectomy on BBB permeability. I labeled these two groups Regular OVX 7m
and Regular Sham OVX 7m, based on the fact that female SD rats in both these groups
of rats were sacrificed at 7 months of age. I decided to have a one-month lag between
ovariectomy and euthanasia for the effect of surgical menopause to manifest itself. After
female rats were classified into specific groups and reached the necessary age, I
sacrificed each rat on the day of estrus (E) as determined by microscopic analysis of daily
vaginal smears.
On the day of sacrifice, I injected the animals with 10 mg of sodium fluorescein in 0.1 ml
of sterile saline, administered via intraperitoneal route, adapted from the protocol
published by Li et al (Li et al., 2015). Thirty minutes later (Kozler and Pokorny, 2003; Yen
et al., 2013), I anesthetized the animals with ketamine-HCl (100–200 mg/kg). I collected
cardiac blood and perfused the rats with PBS to drain blood out. I centrifuged cardiac
blood at 10,000 x g for 10 min at 4°C. I removed the serum (supernatant) and stored it at
-80 °C in 50 μL aliquots and discarded the cell pellet. I dissected the brain of each female
58
rat on ice and collected hypothalamus, cortex and hippocampus. All the brain tissue that
I harvested were frozen at −80 °C for future experiments.
3.3.3 ANALYSIS OF SERUM AND BRAIN TISSUE TO DETERMINE BBB
PERMEABILITY (LI ET AL., 2015)
I thawed the serum (one 50 μL aliquot) and mixed with an equal (50 μL) volume of 15%
trichloroacetic acid (TCA). After centrifugation for 10 min at 10,000 × g at 4°C, I recovered
the supernatant and made it up to 150 μl by adding 30 μl of 5 M sodium hydroxide (NaOH)
and rest 7.5% TCA. I removed from the freezer the brain tissue dissected and stored
earlier (hypothalamus, hippocampus, cortex) and weight each tissue individually. I then
homogenized the brain tissues in cold 7.5% TCA and centrifuged them for 10 min at
10,000 × g at 4°C to remove insoluble precipitates. After the addition of 30 μl of 5 M NaOH
to 120 μl of supernatant, I used Synergy H1 Hybrid Multi-Mode Reader (Bio-Tek
Instruments, Wonooski, VT) to determine the fluorescence of each sample with excitation
at 485 nm and emission at 530 nm. I used standards of NaF (125 to 4,000 μg/mL) in
identical solvent to calculate the NaF content of the samples. NaF uptake into tissue is
expressed as microgram (μg) of fluorescence in tissue/mg of tissue divided by μg of
fluorescence in serum/mL of blood to normalize values of fluorescence detected in brain
tissue of each animals to the value of fluorescence detected in the blood (Fig. 3.1).
59
3.3.4 STATISTICS
Statistical analysis was performed in Prism 6.0. For each tissue type (hypothalamus,
hippocampus and cortex), statistical analysis was carried out using Mann-Whitney test
and each group of female rats was compared to every other group. The difference
between groups was considered significant if the p-value was <0.05.
3.4 Results
The results of dye extravasation are described by tissue type. Highest BBB-PI was
observed in hypothalamus from all animals, whereas hippocampus had lower BBB-PI and
cortex had the lowest BBB-PI.
Figure 3.1 Calculation of Blood Brain Barrier Permeability Index (BBB-PI).
60
Regular cyclers 6m group had the lowest BBB-PI among all groups in case of
hypothalamic tissue, closely followed by Regular Sham OVX 7m (Fig. 3.2). BBB-PI
climbed higher in the Regular cyclers 9-10m group and further increased in Irregular
cyclers 9-10m. Highest BBB-PI was observed in Acyclic 9m followed by Regular OVX 7m.
Figure 3.2 Blood Brain Barrier Permeability Index for hypothalamus in female
SD rats.
In the hypothalamus, blood brain barrier permeability increases with increase in
age and progression of endocrine status towards menopause. The magnitude of
BBB permeability produced by OVX is less variable but equal in magnitude to
Acyclic endocrine status.
61
It is also notable that the individual animals are more widely distributed in Regular cyclers
9-10m, Irregular cyclers 9-10m and Acyclic 9-10m with higher standard deviation than
other groups, implying higher variability in these groups.
Figure 3.3 Blood Brain Barrier Permeability Index of hippocampus in female
SD rats.
In the hippocampus, similar to hypothalamus, blood brain barrier permeability
increases with increase in age and progression of endocrine status towards
menopause. The variation in the magnitude of BBB permeability produced
becomes apparent and animals group together into sub-populations, reflecting
variability in the endocrine transition process; highest variability is exhibited in the
Acyclic group.
62
In case of hippocampus (Fig. 3.3), a similar pattern is followed. Lowest permeability is
seen in Regular cyclers 6m and Regular Sham OVX 7m, and highest permeability is seen
in Acyclic 9-10m and Regular OVX 7m (Fig 3.3). It is also notable that the individual
animals are more widely distributed in Regular cyclers 9-10m, Irregular cyclers 9-10m and
Figure 3.4 Blood Brain Barrier Permeability Index of cortex in female SD rats.
The BBB-PI, a measurement of degree of compromise of BBB is lowest in cortical
tissue. Similar to hippocampus, the variability of BBB compromise is higher in
older groups progressing towards menopause.
63
Acyclic 9-10m with much higher standard deviation than other groups, implying higher
variability in these groups. Highest intra-group variability for hippocampus is seen in the
Acyclic 9m animals.
In case of cortex (Fig. 3.4) similar pattern as seen in hippocampus is followed with some
differences. All groups show higher intra-group variability compared to hippocampus, and
in some groups (Regular Sham OVX 7m, Regular cyclers 9-10m and Irregular cyclers 9-
10m) certain animals have a BBB-PI similar to certain Acyclic 9-10m animals.
Figure 3.5 Blood Brain Barrier Permeability Index (BBB-PI) of hypothalamus with
statistical significance.
Regular cyclers 6m and Regular Sham OVX 7m are similar and significantly different
from other groups. BBB-PI in the Acyclic group is the highest and statistically
significantly different from all other groups.
0
50
100
150
200
250
Hypothalamus
BBB Permeability Index
Regular 6m Regular OVX 7m Regular Sham OVX 7m
Regular 9-10m Irregular 9-10m Acyclic 9-10m
#
#, α, β, γ
#,*, α, β
#, *, α
*
# p< 0.05 vs. Regular cyclers 6m
* p< 0.05 vs. Regular OVX 7m
α p< 0.05 vs. Regular Sham OVX 7m
β p< 0.05 vs. Regular 9-10m
γ p< 0.05 vs Irregular 9-10m
64
Statistically, in case of hypothalamic tissue (Fig. 3.5), Regular cyclers 6m and Regular
Sham OVX 7m are similar to each other but significantly different from other groups. Also,
Regular OVX 7m and Acyclic 9m are statistically similar. Acyclic 9-10m is statistically
significantly different from all other groups except Regular OVX 7m. In case of
hypothalamus Irregular cyclers 9-10m and Regular cyclers 9-10m are statistically different
Figure 3.6 Blood Brain Barrier Permeability Index of hippocampus and cortex
with statistical significance.
Hippocampus and cortical tissue show BBB-PI in Regular cyclers 6m and the Sham
OVX groups to be similar and significantly different from the other groups. Surgical
and natural reproductive senescence groups have similar BBB-PI, but the Acyclic
group exhibits higher variability.
0
10
20
30
40
50
60
70
80
Hippocampus Cortex
BBB Permeability Index
Regular 6m Regular OVX 7m Regular Sham OVX 7m
Regular 9-10m Irregular 9-10m Acyclic 9-10m
#
#,*
#, α, β, γ
*
#
#,*
#,*
#, α, β, γ
# p< 0.05 vs. Regular cyclers 6m
* p< 0.05 vs. Regular OVX 7m
α p< 0.05 vs. Regular Sham OVX 7m
β p< 0.05 vs. Regular 9-10m
γ p< 0.05 vs Irregular 9-10m
65
but this is surprisingly not seen in case of hippocampus and cortex (Fig. 3.6) although the
p-values are less than 0.1 and could reach a level of significance either with more animals
or reduced intra-group variability. Even though Regular cyclers 6m and 9m in case of
hippocampus are not statistically significantly different, the p-value is 0.06 and trending
towards statistical significance is seen in both hypothalamus and cortex.
Figure 3.7 depicts the observed intra-group variability, in the form of standard deviation,
in blood brain barrier permeability index which is a measure of how variable is the
compromise of BBB among these groups. Lower variability suggests little difference in
BBB permeability among the animals in the group and higher variability suggests larger
difference in BBB permeability between animals of the same group. Variability varies by
Figure 3.7 Variability in Blood Brain Barrier Permeability Index of
hypothalamus, hippocampus and cortex.
Variability in the form of standard deviation was calculated for each group for each
tissue type and suggests presence of sub-populations in each group, especially in
older animals as well as endocrinologically advanced animals. The only exception is
the OVX group which has lower variability compared to 9-10m groups.
0
5
10
15
20
25
30
Hypothalamus Hippocampus Cortex
Standard Deviation
Variability in Blood Brain Barrier Permeability Index
Regular 6m Regular OVX 7m Regular Sham OVX 7m
Regular 9-10m Irregular 9-10m Acyclic 9-10m
66
tissue type, but highest variability is seen in 9-10m groups while the younger groups have
lower variability.
3.5 Discussion
This set of experiments furnish functional proof of blood brain barrier permeability in three
brain regions including hippocampus hypothesized by hippocampal transcriptomic
analysis in perimenopausal female SD rats described in Chapter 2. This set of data taken
together further support our hypothesis. Additionally, literature supports the thought that
decreased level of estrogen plays a critical role in decreased integrity of BBB due to
regulation of PDGF. PDGF levels decrease as a consequence of estrogen decline and
subsequently reduced PDGF results in sub-optimal function of endothelial cells in
cerebrovasculature (Bake and Sohrabji, 2004). Endothelial cells along with astrocyte foot
processes, pericytes and basement membrane are responsible for an optimally
functioning BBB (Armulik et al., 2010).
Our previous publication showed that all perimenopausal groups had varied levels of E2
in cortex and serum and that the serum and cortical levels of E2 were not correlated (Yin
et al., 2015). Our group also found that cortical levels of E2 saw a statistically significant
decline from 6 months to 9- to 10-month-old female SD rats, while the amount of E2
measured in acyclic rats was very small and statistically significantly lower than Regular
6 and Irregular 9-10 month groups but not dissimilar to Regular 9-10 month group (Yin et
al., 2015). These previously published findings of cortical levels of E2 in the
perimenopausal model of female rats correlate well with the observed BBB-PI, which is a
measurement of increased extravasation of sodium fluorescein marker across the BBB
and into the brain parenchyma, carried out in an identical rodent model of perimenopause.
67
The decrease of cortical E2 levels seen from 6-month old to 9- to 10-month old rats
correspond to the increased permeability of BBB in the same groups. Cortical estrogen
levels were found to rise in the Irregular 9-10m group which is reflected in a reduced
increase in BBB-PI from Regular to Irregular female rats at 9-10-months of age. Highest
BBB permeability is observed in Regular OVX 7m and Acyclic 9m old female rats implying
that endocrine aging and cortical estrogen levels may determine the optimal functioning
of BBB rather than age.
Additionally, in the Regular 6m group the female SD rats cluster closely around the mean.
All the rats in this group have very similar measurements of BBB-PI with little variation.
Two other groups of female rats that chronologically follow Regular 6m group, namely,
Regular OVX 7m and Regular Sham OVX 7m show higher variation, as measured by
standard deviation, than Regular 6m but lower than the remaining three groups. Rats in
the Regular 9-10m show a much higher standard deviation and the highest standard
deviation is exhibited by Irregular 9-10m female rats in case of hypothalamus. The
variation among the Acyclic 9-10m group is reduced compared to the previous two groups
(Fig. 3.7).
Comparatively less intra-group variability among the 6m and Sham OVX 7m old female
rats suggests that the process of aging results in divergence of blood brain barrier
permeability. While ovariectomy does increase permeability of the BBB in female SD rats,
this increase is relatively uniform across the group due to abrupt cessation of ovarian
hormone production achieved by surgical menopause. Gradual ageing of all the 9-10m
groups, as opposed to sudden surgically induced menopause seen in Regular OVX 7m,
results in separation of the group into distinct subpopulations. In the Regular Sham OVX
7m group in cortex and to some extent in hippocampus we can identify at least two distinct
68
sub-groups: those animals that have maintained BBB permeability similar 6m old animals
and those with elevated BBB permeability. Such sub-populations are visible in Regular,
Irregular and Acyclic 9-10m groups as well.
Since all the female rats are genetically identical and have very similar environmental
exposure and handling, an alternate explanation for the variability in BBB-PI among
animals in the study maybe due to differences in exposure to light in the vivarium (Emmer
et al., 2018; Heywood, 1980), microbiota in the gut (Clarke et al., 2013) and potentially
epigenetic differences (Bacon et al., 2018). However, these factors would likely uniformly
affect all female SD rats in this study but the observed difference in variability between the
different groups suggests a more group-specific cause of intra-group variability in BBB-PI.
Future studies in this research area may focus on the mechanisms of increased BBB
permeability in perimenopausal female rats which I suspect involves the PDGF signaling
pathway, pericytes and tight junction proteins, which are all indirectly influenced by
estrogen according to published literature.
3.6 Conclusion
In our model of natural reproductive senescence, BBB permeability is increased in female
SD rats undergoing perimenopause. Both endocrine and chronological aging predispose
to this increase, but data suggests that the increase in permeability is more dependent on
gradual endocrine aging rather than pure chronological aging or abrupt endocrine aging.
The divergence of BBB permeability among female SD rats as a result of chronological
and endocrinological aging hints that there may be certain protective or deleterious factors
besides estradiol that exert influence in modifying the permeability of BBB during the aging
process in females. If found extrapolatable to humans these findings could explain the
69
variability seen among human females in the symptoms, onset and progression of
diseases such as Multiple Sclerosis that exhibit high inter-individual variability (Kappos et
al., 2015) where BBB permeability may even serve as a pathological marker of
neuroinflammation (Cramer et al., 2015).
70
Allopregnanolone: A Potential Neuroregenerative
Therapeutic
4.1 Abstract
This study is a translational research project that uses ApoE-Targeted Replacement (TR)
mouse model to examine a promising neuroregenerative therapeutic in a genotype
susceptible to AD. In this experiment, I injected mice with the human form of ApoE gene
with a novel neuroregenerative therapeutic, Allopregnanolone (Allo), which is currently
under clinical development in humans. I found that both males and females with ApoE 4/4
genotype demonstrated significant cognitive improvement as measured by the novel
object recognition test. RNA-Seq study of the hippocampi of the Allo-treated female ApoE
4/4 mice suggested upregulation of PDGF-BB compared to saline treated counterparts,
suggesting that Allo could potentially ameliorate age related increase in blood brain barrier
permeability. The RNA-Seq data also suggested Allo’s effect on multiple immune system,
cell signaling and metabolomic pathways. Study of 185 plasma and cortex metabolites
suggest major differences in glycerophospholipids, protein and lipid metabolism based on
gender and genotype. In vitro Allo and its analogues have been shown to potentiate
mitochondrial function in human neural stem cells and 30 minutes of exposure to Allo
confers similar mitochondrial potentiation as 24-hour continuous exposure in these stem
cells. Allo was also evaluated as a potential agonist of mitochondrial Transmembrane
protein (TSPO) but found to have antagonistic-like activity in human neural stem cells.
71
4.2 Introduction
Apolipoprotein (Apo) E4 allele is the strongest genetic risk factor for AD (Corder et al.,
1993; Farrer et al., 1997a; Genin et al., 2011; Huang and Mucke, 2012). Females bear an
unequal burden of ApoE4 as a genetic risk factor with advancing age (Damoiseaux et al.,
2012; Farrer et al., 1997a; Leung et al., 2012). Heterozygosity for ApoE4 allele increased
AD risk by a factor of three, while homozygosity for the same allele (ApoE4/4 genotype)
increased AD risk by a factor of ten (Holtzman et al., 2012; Loy et al., 2014). Those with
the E4 allele are also likely to manifest AD at an earlier age compared to those with E2 or
E3 allele (Spinney, 2014).
Mice whose ApoE locus contains human ApoE gene are termed as ApoE Targeted
Replacement (ApoE-TR) mice. In aging ApoE4-TR mice learning and memory is
significantly impaired due to loss of certain specific neurons (Andrews-Zwilling et al., 2010;
Leung et al., 2012; Li et al., 2009). Tong et al (Tong et al., 2016) demonstrated that
treatment of these mice with a molecule that acts as a GABAA receptor potentiator reduced
cognitive decline and improved functional outcome in mice.
There have been multiple publications from our lab on Allopregnanolone (Allo). Allo (3α-
hydroxy-5α-pregnan-20-one) activates GABAA receptor channels (Irwin and Brinton,
2014). Previously, we have established that Allo significantly increases proliferation of rat
neural progenitor cells and human neural stem cells (Wang et al., 2005), reverses deficits
in neurogenesis and cognition in the 3xTg mouse model of AD (Wang and Brinton, 2008;
Wang et al., 2010), and improves survival of neural progenitor cells (Singh et al., 2012).
In fact, Allo is currently being studied as a first-in-class neuroregenerative therapeutic for
AD and mild cognitive impairment in a Phase Ib/IIa clinical trial (ClinicalTrials.gov
Identifier: NCT02221622). While it is known that activation of TSPO by its ligands such as
72
FGIN 1-27 and PK 11195 leads to Allo production (Daugherty et al., 2013), producing
anxiolytic effects in animal studies (Bitran et al., 2000), our aim was to detect if Allo itself
acted as an agonist or antagonist of TSPO with respect to mitochondrial potentiation,
which then would have suggested a feedback loop.
4.3 Methods:
4.3.1 ANIMALS
Animal studies were performed following National Institutes of Health guidelines on use of
laboratory animals; protocols were approved by the University of Southern California
Institutional Animal Care and Use Committee. Experimental and control animals had
identical housing conditions from birth through euthanasia (12h light/dark cycle and
PicoLab Rodent Diet 20). Female and male ApoE3-TR and ApoE4-TR homozygous mice
and heterozygous female ApoE3/4-TR mice on a C57BL/6 background strain were
obtained from Taconic Inc. and housed and bred at University of Southern California
animal facility. I weighed the mice every week from the start of the study till euthanasia.
All mice were fed identical chow and I monitored their food intake along with body weight
to control for loss of body weight due to reduced consumption of food. We attempted to
obtain our own ApoE 3/4 mice, however, the breeding did not succeed. I divided the mice
into groups based on their genotype (ApoE3, ApoE4, ApoE 3/4), gender (male and female)
as well as treatment group (Allo and Saline). Thus, the ten groups were ApoE3 Female
Allo, ApoE3 Female Saline, ApoE3 Male Allo, ApoE3 Male Saline, ApoE4 Female Allo,
ApoE4 Female Saline, ApoE4 Male Allo, ApoE4 Male Saline, ApoE 3/4 Female Allo and
ApoE 3/4 Female Saline.
73
4.3.2 ALLOPREGNANOLONE INJECTIONS AND SACRIFICE
I injected the mice weekly with intramuscular injection of Allopregnanolone (1.5 mg/mL)
or saline (0.9%). Allo and saline I used to treat mice were identical to the Allo and saline
used in the clinical trial (ClinicalTrials.gov Identifier: NCT02221622). The amount of
Allopregnanolone to be injected was calculated based on a dose of 2 mg which was
identical (extrapolated to mice) to the dose that human participants received in the first
cohort of the phase Ib/IIa clinical trial. I administered these weekly injections on the same
day and in the same time window of nine to eleven AM for 26 weeks, starting from 10
months of age to 16 months of age. I ensured that any behavior study or blood collection
procedure were separated in time from injections by at least four hours.
I sacrificed the animals after the last Allo injection. I first anaesthetized the mice using
Ketamine-Xylazine cocktail (87.5 mg/kg Ketamine + 12.5 mg/kg Xylazine) 0.1mL/20g
mouse weight injected intraperitoneally and collected cardiac blood in EDTA-coated tubes.
I then centrifuged the tubes for 10 minutes at 2,000 x g using a refrigerated centrifuge
previously adjusted to be at 4 °C. After centrifugation I immediately transferred the plasma
to 1.5mL microcentrifuge tube and stored all plasma at -80 °C. The mice were not fasted
before collecting this plasma and this plasma was not appropriate to carry out
metabolomic analysis. Separate plasma was collected for metabolomic study as outlined
later. The mice were then perfused with phosphate buffered saline (PBS) for at least 15
minutes. I then dissected the brain of each mouse on ice and separated brain stem and
cerebellum first and then each cortical hemisphere laterally to dissect the hippocampus. I
stored all brain tissues in -80 °C.
74
4.3.3 NOVEL OBJECT RECOGNITION TEST
I tested all mice for the Novel Object Recognition (NOR) test in accordance with published
literature (Antunes and Biala, 2012; Ennaceur, 2010; Leger et al., 2013; Piterkin et al.,
2008; Taglialatela et al., 2009). NOR was carried out after 24 once a week Allo injections.
In short, I first acclimatized the mice to the NOR test room for five days during the
habituation phase, taking care that the mice entered and exited the behavior room at the
same time each day. I also mimicked the regular opening and closing of the room door
during the habituation phase by opening and closing the room door every 15 minutes.
During the second week, I allowed the mice to acclimatize to the wooden boxes in which
the NOR test was to be carried out (one mouse/box) for 15 minutes, on the first three
consecutive days. On the next day (fourth day of second week) I allowed the mice to
explore for 10 minutes two identical objects (either two identical blue Lego pieces or two
identical glass beakers) while in the wooden box and noted the specific object explored
by each mouse. On the last day of NOR testing (day five of second week), I switched one
of the two objects for the other and each mouse had a choice between a blue Lego piece
and a glass beaker. One of the two objects had been explored by the mouse the previous
day (familiar object) and the other was the novel object. Using video recording, the number
of seconds spent by each mouse exploring each of the two (familiar (Tf) and novel (Tn))
objects was measured and two quantities were calculated. First, the difference in
exploration time (in seconds) was calculated by subtracting the time taken to explore
familiar object from that of the novel object (Tn – Tf) and denoted by Df. Second, a ratio
of the difference in exploration time to the total time spent exploring was calculated (Tn-
Tf/Tn+Tf) and denoted by Discrimination Index (DI).
75
4.3.4 SEQUENCING OF HIPPOCAMPAL RNA
The RNA I extracted from the hippocampi of ApoE mice was shipped to VANTAGE
(Vanderbilt Technologies for Advanced Genomics) at Vanderbilt University and raw reads
in the form of .fastq files were obtained. Before sequencing, VANTAGE analyzed RNA
quantity and quality Qubit (Thermo Fisher, MA) and Bioanalysis Pico (Agilent
Technologies, CA) and found the RNA to meet quality standards. VANTAGE utilized
polyA-selected mRNA library preparation process and used Illumina HiSeq3000 to carry
out paired end sequencing with a read length of 75 base pairs and a read depth of about
50 million reads. RNA from eight different groups of ApoE mice was sequenced with five
mice per group (a total of 40 RNA samples) and initial raw data processing was carried
out by Dr. Shang at UA and I analyzed the DEG list using IPA.
4.3.5 PLASMA COLLECTION AND METABOLOMIC STUDY
I collected whole blood in EDTA-coated tubes from each mouse through the eye as per
the approved protocol, after subjecting the mice to an overnight fast of 14-16 hours, after
24 weeks of once a week treatment with Allo and before beginning NOR. I centrifuged the
tubes for 10 minutes at 10,000 x g using a refrigerated centrifuge previously adjusted to
be at 4 °C. After centrifugation I immediately transferred the plasma to 1.5mL
microcentrifuge tube and stored all plasma at -80 °C. The plasma and cortex from each of
the experimental mice groups were queried for 185 metabolites using ultra-performance
liquid chromatography-tandem mass spectrometry [(UP)LC-MS/MS] by TGen (Phoenix,
AZ).
4.3.6 IN VITRO STUDY IN HUMAN NEURAL STEM CELLS (HNSC)
Human embryonic brain cortical stem cells were provided as cryopreserved neurospheres
by Dr. Clive Svendsen in vials and stored in -80 C. I thawed them and transferred them to
76
T25 flask which contained Thaw media (section 4.3.6.4). The cells were passaged (section
4.3.6.5) at least twice before being used for experiments in order to ensure cell viability
after long-term storage.
4.3.6.1 Determining optimal cell density:
In preliminary experiments, optimum cell density and Allo concentration to be used were
determined using identical methods as the study below. Optimal density was found to be
40,000 cells per well and optimal Allo concentration was 100 nM. Allo concentrations of
1nM, 10nM and 1000nM were used in certain experiments to obtain dose response curve
of mitochondrial potentiation in hNSCs in response to Allo exposure.
4.3.6.2 Mitochondrial function measurement using Seahorse XF96 Extracellular
Flux Analyzer in response to Allo and analogues:
I cultivated freshly passaged (section 4.3.6.5) Svendsen’s human Neural Stem Cells
(hNSC) in T25 flasks for 7 days in 100% growth factor (EGF, bFGF, Heparin) containing
‘Thaw’ media (section 4.3.6.4). I changed the media on day 4 post-seeding. On day 7, I
passaged the cells again (section 4.3.6.5), and seeded them by columns at 40,000 cells
per well in 100% growth factor containing Thaw media in Seahorse XF96 FluxPak (Agilent
Technologies, Santa Clara, CA) plate coated with Matrigel hESC-qualified matrix (354277,
Corning, Corning, NY) and incubated overnight (section 4.3.6.5). On day 2, I replaced the
media with 0% growth factors containing Thaw media (EGF, bFGF, Heparin absent) in the
morning and incubated for 4 hours. This was the starvation step. I performed a second
wash step in order to minimize the residual amount of growth factors left behind in the
wells of the plate, which had in earlier experiments confounded our results. After 4 hours,
cells were treated with either vehicle (ethanol 0.001%) or Allopregnanolone (1nM, 10nM,
100nM or 1000nM) or an Allo Analogue (100nM). I used ethanol as a solvent for Allo and
77
its analogues. I measured Oxygen Consumption Rate (OCR) using Seahorse XF96
Extracellular Flux Analyzer 24 hrs. post-treatment and assessed mitochondrial respiration
(section 4.3.6.6). To correct for any plating differences, I used protein readings of all wells
(5000006, Bio-Rad Laboratories, Hercules, CA) to normalize measurements from all wells
within the Seahorse XF96 Wave software.
I obtained the aggregated differences in mitochondrial potentiation in response to Allo and
analogues treatment of hNSCs from six different experiments. All the data were
normalized to protein content as described earlier and statistics were calculated by
normalizing values of each plate to the average value of OCR in wells with vehicle treated
hNSCs. Therefore, values of OCR for all non-vehicle treated hNSC-containing wells were
calculated as a ratio of the OCR generated by the specific treatment to the OCR measured
in the same cells in the same experiment by treatment with vehicle during basal
respiration; an average of all six readings from all wells at basal respiration was calculated
for this purpose. I only used hNSCs with passage 14 or less in these experiments. No
cells in the experiment exceeded 5 freeze-thaw cycles.
4.3.6.3 Measurement of mitochondrial function at various time points in response
to 30-min pulse treatment with Allo
I conducted a pulse-based experiment to determine activation time required to activate
hNSCs and potentiate mitochondrial respiration by Allo. The cells were exposed to Allo
only for 30-min (a 30-min Allo ‘pulse’) and mitochondrial function of hNSCs determined at
multiple time points post-treatment. I chose the 30-min pulse treatment because that was
the duration of exposure of the clinical trial participants of Allo phase Ib/IIa clinical trial
(ClinicalTrials.gov Identifier: NCT02221622) to intravenous infusion of Allo. The pulse
treatments were compared to a 24-hour continuous Allo treatment, as positive control, for
78
which I already had data. Pulse protocol is shown in figure. Briefly, I seeded cells in
Seahorse plate as mentioned earlier and incubated them overnight. I then ‘starved’ the
cells by removing the growth factor containing media using two washes and then after four
hours of ‘starvation’ specific wells on the plate were treated with Allo as shown in Figure
4.1. I removed Allo after 30-min. by two washes. Only for hNSCs in one set of cells I
allowed 24-hour exposure to Allo rather than 30-min and this was my positive control
group. Each treatment with Allo or vehicle had at least 12 wells in each plate and this
experiment was repeated six times.
Figure 4.1 Allo ‘pulse’ protocol.
hNSCs were treated for 30-min at different time points preceding specific number
of hours before measurement of mitochondrial function to understand the
activation time required for and duration of, Allo’s effect on mitochondria of hNSCs.
The timeline of each set of wells is outlined above where grey denotes the
absence of both Allo and growth factors, blue denotes the presence of Allo and
pink denotes the presence of growth factors. The larger blue arrow in the last set
of wells denotes the larger dose of Allo used (1000nM), while all other Allo doses
were 100nM. Ethanol was used as a solvent for Allo and vehicle-treated cells were
treated with 0.001% ethanol, which is the same concertation of ethanol as that of
the solvent in Allo preparation.
79
Evaluation of Allo as an agonist or antagonist of TSPO
I carried out passaging, seeding and incubation of hNSC cells as described earlier. I
treated the cells seeded into wells of Seahorse plates with either Allo, vehicle, a known
agonist of TSPO (FGIN 1-27), a known antagonist of TSPO (PK 11195), a combination of
Allo and antagonist and a combination of agonist and antagonist. Based on literature the
I chose a concentration of 1µL for both FGIN 1-27 and PK 11195 and Allo concentration
was 100nM as per earlier experiments. I used ethanol as vehicle and its concentration in
the wells was 0.001%. Six independent experiments were carried out and cell passage
numbers did not exceed 14. No cells in the experiment exceeded 5 freeze-thaw cycles.
4.3.6.4 Preparation of media for cell culture and Seahorse XF96 procedure
4.3.6.4.1 Thaw Media (100 percent (%) growth factor containing thaw media – 100GF):
Thaw Media with specific amount of growth factors (defined as 100% for experimental
purposes) can be prepared by combining 33.9ml DMEM (Gibco/Invitrogen, Cat. # 21063-
029), 14.5ml Ham’s F-12 (Gibco/Invitrogen, Cat. # 11765-054), 0.5ml
Penicillin/Streptomycin (Gibco/Invitrogen, Cat. # 15140-122), 1ml B27 (50X)
(Gibco/Invitrogen, Cat. # 17504-044), 100µl of 10μg/ml stock bFGF (Gibco/Invitrogen, Cat.
# 13256-029), 100μl of 10μg/ml stock EGF (Gibco/Invitrogen, Cat. # 13247-051) and 125μl
of 2mg/ml stock Heparin
4.3.6.4.2 Thaw Media (without growth factors – zero % growth factor containing thaw
media):
Thaw Media without any amount of growth factors (defined as zero % for experimental
purposes) can be prepared by combining 33.9ml DMEM (Gibco/Invitrogen, Cat. # 21063-
029), 14.5ml Ham’s F-12 (Gibco/Invitrogen, Cat. # 11765-054), 0.5ml
80
Penicillin/Streptomycin (Gibco/Invitrogen, Cat. # 15140-122) and 1ml B27 (50X)
(Gibco/Invitrogen, Cat. # 17504-044). No growth factors are added to this media.
4.3.6.5 Passaging hNSCs for in vitro experiments
First, I made sure to aspirate any Thaw medium already in the flask. I then washed the
flask twice with 10mL of Hank’s Balance Salt Solution (BSS) to remove any remaining
serum, as this will inhibit the action of the dissociation enzyme. I then aspirated out the
Hank’s BSS and added 1 mL of Accutase (07920, STEMCELL Technologies, Vancouver,
BC, Canada). Then I placed the flask in the incubator (37 °C) for 3 minutes. After that I
neutralized the Accutase enzyme with 7 mL of 100GF Thaw Medium. I then transferred
contents of the flask to a sterile 15mL conical tube. After washing the flask with another
4mL of 100GF Thaw Medium to collect any remaining cells, I added this wash to the 15mL
conical tube. I then centrifuged the suspension at 300 x g (gravitational constant) for 5
minutes. After centrifugation I aspirated the supernatant, resuspend the cell pellet by
gently tapping the tube, then add 5mL 100GF Thaw Medium to fully resuspend cells. I
then counted the cells using a hemocytometer.
4.3.6.6 Using Seahorse XF96 Extracellular Flux Analyzer
4.3.6.6.1 Seahorse Medium
Seahorse medium can be prepared by combining 1 vial of Dulbecco′s Modified Eagle′s
Medium (DMEM) powder (without glucose, L-glutamine, phenol red, sodium pyruvate and
sodium bicarbonate) (D5030-1L, Sigma-Aldrich, St. Louis, MO) with 10mL of Glutamax
(35050-061, Thermo Fisher Scientific, Waltham, MA), 1.85g of NaCl (0241, Amresco LLC,
Solon, OH) in 1L of sterile diH2O. I measured 900mL sterile diH2O into large beaker,
added DMEM powder and dissolved it with stirring. Then I added glutamax and NaCl and
adjusted the pH to 7.4 at 37 °C with NaOH/HCl. I made it up to 1000mL with sterile diH2O
81
while stirring briefly. I filtered it with sterile filter (431096, Corning, Corning, NY), and stored
it at 4 °C. On day of use, I warmed it to 37
°C and added 25mM glucose (BDH9230, VWR
International, Radnor, PA).
4.3.6.6.2 Seahorse XF96 procedure:
I assessed mitochondrial function using the Seahorse XF96 Analyzer (Agilent
Technologies, Santa Clara, CA). On Day 1, I plated hNSCs at a density of 40,000 cells/well
in 100% growth factors containing Thaw medium on 96- well cell culture microplate from
the Seahorse XF96 FluxPak previously coated with Matrigel (354277, Corning, Corning,
NY). I left the plate at room temperature for 30-minutes to allow cells to settle and then I
incubated the plate at 37 °C with 5% CO2. On Day 2, I replaced the media in the wells with
zero % growth factors containing Thaw medium in the morning and incubated it for 4 hours
(starvation step). I performed a second wash step to minimize the residual amount of
growth factors left behind in the wells of the plate which could interfere with the results.
After 4 hours, I treated cells in each well with either vehicle (ethanol 0.001%) or
Allopregnanolone (1nM, 10nM, 100nM or 1000nM) or an Allo Analogue (100nM). Ethanol
was used as a solvent for Allo and Allo analogues. I then left the plate to incubate at 37
°C. On day 2, I also rehydrated a XF96 cartridge overnight from the Seahorse XF96 Flux
Pak (102416, Agilent Technology, Santa Clara, CA) as per manufacturer’s instructions.
On Day 3, 24-hours after the treatment step, I changed the media in the wells to unbuffered
Seahorse Medium, with 25mM glucose, warmed to 37
°C. I then incubated the plate at 37
°C without CO2 for 60 minutes. I loaded the XF96 cartridge with the following inhibitors:
Port A: Sodium Pyruvate (50mM); Port B: Oligomycin (32μM; to inhibit ATP synthase);
Port C: FCCP (carbonylcyanide p-trifluoromethoxyphenylhydrazone, 9μM; to uncouple
mitochondria); Port D: Rotenone (10μM; to inhibit electron transfer from complex I to
82
ubiquinone). Injection volumes for all ports were 25μl. The inhibitor concentrations at
injection are: Sodium Pyruvate 7mM (S8636, Sigma-Aldrich, St. Louis, MO), Oligomycin
4μM (151786, MP Biomedicals, Santa Ana, CA), FCCP 1μM (0453, Tocris Cookson,
Bristol, UK), Rotenone 1μM (0215015410, MP Biomedicals, Santa Ana, CA). I set the
machine to carry out six baseline measurements of oxygen consumption rate (OCR) and
extracellular acidification rate (ECAR) before sequential injection of mitochondrial
inhibitors. I also set the machine to carry out three measurements following inhibitor
addition and prior to automated addition of subsequent inhibitors. OCR and ECAR were
automatically calculated and recorded by Seahorse XF96 Wave software. To correct for
any plating differences, I used protein readings of all wells (5000006, Bio-Rad
Laboratories, Hercules, CA) to normalize values of OCR and ECAR from all the wells
within the Seahorse XF96 Wave software. I calculated the parameters of mitochondrial
function as follows: Basal Respiration (Average first six OCR readings); Maximal Capacity
(First reading after FCCP addition – Non-Mito Respiration); Reserve Capacity (First
reading after FCCP addition – Basal Respiration). I calculated outliers with the ROUT
method using Prism 6.0 and keeping FDR of 1%.
83
4.3.7 IMAGING
MRI study was carried out on brains of certain Allo and saline treated mice that I shipped
to University of Arizona. At least two mice were selected per treatment group based on
their performance in the NOR test to obtain the best visual evidence of differences due to
Allo treatment compared to saline in the brain.
4.4 Results
4.4.1 BODY WEIGHTS OF MICE
Body weights of mice showed a general trend of increase during the study. No
statistically significant differences were observed between Allo and saline treated
mice of identical sex and genotype. All male mice had statistically significantly
Figure 4.2 Body weights of different groups of ApoE mice by sex, genotype
and treatment.
Body weights of mice showed a general trend of increase during the study. No
statistically signficant differences were observed between Allo and saline treated
mice of identical sex and genotype.
84
higher weight than corresponding female mice with two exceptions, male vs.
female ApoE 3/3 Allo (end of study weights p-value=0.07) and male vs. female Allo
ApoE 4/4 (baseline body weight p-value=0.09). By genotype, I found no statistically
significant difference in body weights except between Allo treated ApoE 3/3 and
ApoE 4/4 males at baseline and between vehicle treated ApoE 3/3 and ApoE 4/4
females at end of study.
4.4.2 COGNTIVE FUNCTION OF MICE
Figure 4.3 Difference in exploration time between different groups.
Both Allo-treated ApoE 4/4 males and females explored novel object for longer
period of time compared to their saline-treated counterparts. However, only in
female ApoE 4/4 mice this difference reached significance.
85
NOR test results indicate that ApoE 4/4 female mice spent significantly more time
exploring the novel object compared to the object seen previously suggesting that ApoE
4/4 female mice benefit significantly from Allo treatment and Allo prevents decline in
cognitive function in ApoE 4/4 female mice (Figure 4.1). ApoE 4/4 male mice spend more
time exploring the novel object compared to the object seen previously, however, this
difference in time does not reach the level of significance (p = 0.08). Discrimination Index
Figure 4.4 Discrimination index in different groups.
Both Allo-treated ApoE 4/4 males and females have statistcally significantly higher
discrimination index compared to their saline-treated counterparts.
86
(DI) is the ratio of
the difference in
exploration time to
the total time spent
exploring both
objects and it
controls for the
variation in the
amount of time
mice spend
exploring both
objects. Both ApoE
4/4 female and male mice have
a significantly higher DI than
ApoE 3/3 female and male,
Figure 4.5a-b Difference in
exploration time and
discrimination index for
Allo and saline treated
ApoE 3/4 heterozygous
female mice.
Allo treated ApoE 3/4 mice
showed considerable
improvement in cognitive
function but it did not reach
statistical significance. I
believe increasing the
number of mice in each
group could facilitate
detection of statistically
significant difference on Allo
treatment, if present.
Figure 4.5a
Figure 4.5b
87
respectively (Figure 4.2). ApoE 3/3 female and male mice do not show cognitive benefit
from Allo treatment.
In ApoE 3/4 female mice, Allo treatment improves cognitive function when tested by the
novel object recognition test. However due to low number of animals in each group (n=5
per treatment group), the results are not statically significant (p-values are slightly higher
than 0.05).
4.4.3 RNA-SEQ
Canonical pathways in ApoE 4/4 Allo treated females
In ApoE 4/4 Allo treated mice the canonical pathways found through the RNA-Seq are
shown in Figure 4.6. Out of the 23 pathways shown almost half (11/23) are connected to
the immune system and involve innate and adaptive immune system, Th1 and Th2
activation, cytokine production and antibody-mediated immunity. Other pathways of note
include pathways related to estrogen, glucocorticoids and mineralocorticoids, pathways
that suggest increased metabolism and energy utilization such as mannose degradation
and acyl-CoA hydrolysis as well as pathways related to gene and transcription regulation.
These pathways suggest that Allo plays a role in cell cycle regulation and signaling as well
as repression of DNA transcription via methylation.
Comparison of Allo treated female mice of all three genotypes
Figure 4.7 shows the comparison of all canonical pathways up- or downregulated when
all Allo treated female mice are compared to their vehicle treated counterparts.
Interestingly the top pathways are consistently similar between ApoE 4/4 Allo vs. saline
treated females, while ApoE 3/3 are different. Based on hippocampal transcriptome ApoE
3/3 Allo treated females are a distinct phenotype while ApoE 3/4 and ApoE 4/4 Allo treated
88
females are similar. IL-3 and IL-7 signaling as well as leptin and prolactin pathways which
Figure 4.6 Up- and down-regulated canonical pathways in ApoE 4/4 females in
response to Allo.
Canonical pathways up- and down-regulated when ApoE 4/4 Allo treated females are
compared to saline treated females of identical genotype. A number of immune system
related pathways, along with metabolomic and signaling pathways are seen.
89
are considered pro-inflammatory are seen
upregulated in ApoE 3/3 Allo treated female
mice while the same canonical pathways are
downregulated in both ApoE 4/4 and ApoE 3/4
female mice due to Allo treatment suggesting
Allo’s effect on immune regulation. This
corroborates the results of NOR in which both
ApoE 3/4 and ApoE 4/4 females showed
cognitive improvement while ApoE 3/3 females
did not. The mechanism of Allo-induced
amelioration of cognitive function in ApoE 4/4
and 3/4 females is likely due to differential
hippocampal transcription.
PDGF-BB upregulation by Allo
In response to Allo treatment, I found that
PDGF is upregulated. Specifically, PDGF-BB
isoform was found to be upregulated in Allo
treated ApoE 4/4 females when compared to
saline treated ApoE 4/4 females (Fig. 4.8),
Figure 4.7 Canonical pathways up- and
down-regulated by Allo treatment in all
three female genotypes.
When all Allo treated female mice are
compared to their saline treated
counterparts, similar canonical pathways
are seen up- and downregulated in female
ApoE 4/4 and 3/4 mice, while ApoE 3/3
mice are different based on hippocampal
gene expression.
90
while -AA and -CC variants of PDGF were downregulated and the -DD variant remained
unchanged. Also, PDGFRα is downregulated. Downstream, the data suggests that most
molecules are downregulated except JNK1, SRF and PKR.
Figure 4.8 Upregulation of PDGF-BB in ApoE 4/4 Allo treated females compared to
saline treated females of identical genotype.
PDGF signaling pathway in the hippocampal RNA-Seq data shows upregulation of
PDGF-BB and downregulation of PDGFRα.
91
Genotype and gender differences in Allo treatment
Effects of Allo on the hippocampus are
genotype specific, as evinced by the
canonical pathways seen in Figure 4.9.
ApoE 3/3 mice have a different
transcriptomic phenotype from ApoE 4/4,
both males and females. Majority of the
pathways are downregulated in ApoE 4/4
mice, both males and females, after
treatment with Allo, while the same
pathways are upregulated in ApoE 3/3
male and female mice despite treatment
with Allo. Inflammatory pathways like
Leptin signaling and IL-3 signaling as well
as GM-CSF signaling are upregulated in
ApoE 3/3 but downregulated in ApoE 4/4.
Surprisingly oxidative phosphorylation and
gluconeogenesis are upregulated only in ApoE 4/4 males and downregulated in all other
Figure 4.9 Genotype and gender
differences in treatment effect of Allo on
hippocampal transcriptome in ApoE
mice.
Effects of Allo on the hippocampus are
genotype specific, as evinced by the
canonical pathways seen here. Majority
of the pathways are downregulated in
ApoE 4/4 mice, after Allo treatment
while the same pathways are
upregulated in ApoE 3/3 mice despite
treatment with Allo.
92
groups. Allo induced downregulation of ALS signaling in ApoE 4/4 supports its
neuroprotective credentials. This observation hints at genotype specific regulation of
hippocampal transcriptome by Allopregnanolone and is similar to what I observed when
only female groups treated with Allo were compared to saline groups. ApoE genotype is
the primary variable that modifies Allo treatment rather than sex and this could justify
phase IIb trial in ApoE 4/4 enriched cohort.
93
4.4.4 METABOLOMIC PANEL
The metabolomic panel was an extensive study of 185 metabolites in plasma and cortex
of mice from all ten groups. I am discussing selected results from the metabolomic study
below.
94
4.4.4.1 ApoE 4/4 Females Allo vs. Saline treated
Eight metabolites were significantly different in cortex of Allopregnanolone and vehicle
treated Apo E4/4 females. Although the Allo and saline treated groups of ApoE 4/4
females showed a clear separation on PCA plot, no metabolites reached significance
when p-value was corrected for multiple comparisons in either cortex or plasma.
Figure 4.10 Differences in cortical metabolites in Allo vs. saline treated ApoE 4/4
females.
In cortex eight metabolites and four glycerophospholipids were slightly lower and
acylcarnitines were higher upon allopregnanolone treatment. The Allo and saline
treated groups showed clear separation on PCA plot. However, no single metabolite
was significantly different in plasma.
95
4.4.4.2 Saline treated ApoE 3/3 vs. ApoE 3/4 females
lysoPC a
C20:4
µ
Apo
E3/4
Apo
E3/3
C18:1-
OH
Apo
E3/4
Apo
E3/3
Non-essential
AA
Apo
E3/4
Apo
E3/3
Figure 4.11 Cortical metabolite differences between saline treated ApoE 3/3 and 3/4
females.
Three metabolites were significantly different in cortex of saline treated ApoE 3/3 and
ApoE 3/4 female mice. However, none reached a level of significance after adjusting
the p-value.
96
When Saline treated ApoE 3/3 and ApoE 3/4 are compared, cortical metabolites show no
statistically significant difference with corrected p-value, but in plasma several
glycerophospholipids, amino acid, biogenic amines and acylcarnitines were significantly
different between the two groups.
Figure 4.12 Plasma metabolite differences between saline treated ApoE 3/3 vs 3/4
female mice.
Several glycerophospholipids, amino acids, biogenic amines and acylcarnitines were
significantly different in plasma of saline treated ApoE 3/3 and ApoE 3/4 female mice.
Majority of glycerophospholipids and two amino acids reached significance after
adjusting for p-value. ApoE 3/4 had higher glycerophospholipids than ApoE 3/3, which
can be used as metabolic phenotype.
97
4.4.4.3 Saline treated ApoE 4/4 vs. ApoE 3/4 females
In both plasma and cortex, several glycerophospholipid were different between ApoE 3/4
and ApoE 4/4 female mice (uncorrected p-value <0.05). None reached significance with
corrected p-value. All glycerophospholipids were higher in Apo E3/4, similar to comparison
of saline treated ApoE 3/3 with ApoE 3/4.
Figure 4.13a-c Plasma metabolite differences between saline treated ApoE 4/4 vs 3/4
female mice.
Several glycerophospholipid were signficantly different in plasma between saline
treated ApoE 3/4 and ApoE 4/4 female mice based on p-value before correction for
multiple comparisons. However, none reached significance with adjusted p-value. All
glycerophospholipids were higher in ApoE 3/4, similar to the comparison between
saline treated ApoE 3/3 and ApoE 3/4 females. Both groups showed clear separation
on PCA plot. In the cortex only one glycerophospholipid was found to be different,
lysoPC a C14:0.
Figure 4.13a: Plasma
98
In the cortex only one glycerophospholipid, lysoPC a C14:0 was found to be different
between the two groups but did not reach a level of significance.
Figure 4.13b: Plasma
lysoPC a C14:0
µM
Apo E3/4
Apo E4/4
Figure 4.13c: Cortex
99
4.4.4.4 Saline treated ApoE 3/3 vs. ApoE 4/4 females
When saline treated ApoE 3/3 and ApoE 4/4 females were compared, 17 metabolites were
significantly (uncorrected p-value) different in cortex and 26 metabolites were significantly
(uncorrected p-value) different in plasma. Even though none reached significance with
corrected p-value, PCA plot shows a clear separation between the two groups in both
plasma and cortex and a much tighter clustering of saline treated ApoE 3/3 female mice
in data obtained from plasma.
Figure 4.14a-b Cortical and plasma differences in metabolites between saline treated
ApoE 3/3 and 4/4 female mice.
Saline treated ApoE 3/3 and ApoE 4/4 females show a clear separation on the PCA
plot, both in cortex (4.10a) and plasma (4.10b). In plasma the ApoE 3/3 saline treated
females cluster tighter together compared to cortex.
Figure 4.14a: Cortex
100
Figure 4.14b: Plasma
101
4.4.4.5 Saline treated ApoE 3/3 and ApoE 4/4 females and males
In both ApoE 3/3 and ApoE 4/4 models, plasma glycerophospholipids were higher in
males in comparison to females. In ApoE 4/4, in cortex arginine and phenylalanine were
found to be higher in males, and polyamines (as a result of arginine catabolism) were
lower in males compared to females. In ApoE 3/3 plasma, carnosine, acyl-ornithine (Ac-
Orn) and asymmetric dimethylarginine (ADMA) were higher in females and alpha-
aminoadipate (αAA) was lower, whereas in cortex, spermidine was high in females and
putrescine and αAA were lower. Ac-Orn, ADMA and carnosine are the end products of
arginine catabolism through urea cycle, polyamine biosynthesis-oxidative stress and
carnosine anabolism, respectively.
Figure 4.15a
Saline treated
ApoE 4/4
males and
females
(cortex).
102
Figure 4.15a-d Metabolite differences in cortex and plasma between saline treated
ApoE 3/3 and 4/4 males and females.
Saline treated ApoE 3/3 and ApoE 4/4 males and females were compared to
understand differences based on sex in plasma and cortex. In ApoE 4/4 genotype,
males exhibit higher glycerophopholipids, arginine and phenylalanine, while in ApoE
3/3, carnosine, Ac-Orn and ADMA are higher in females and alpha-AA is lower.
Figure 4.15b
Saline treated
ApoE 4/4
males and
female
(plasma).
103
Figure 4.15c
Saline treated
ApoE 3/3
males and
female
(cortex).
104
4.4.5 MITOCHONDRIAL POTENTIATION OF HNSCS BY ALLO AND ALLO
ANALOGUES
Allo dose response curve
Dose response study in hNSCs with respect to Allo treatment is shown in Figure 4.16.
Allo at both doses (10nm and 100nM) potentiates mitochondrial function similarly, even
though Allo 100nM seems to potentiate mitochondrial function more than Allo 10nM,
these results are within margin of error. I chose the dose of 100nM for other experiments
in order to be consistent with previous studies from our lab and also because dose
response data from 100nM dose was less variable.
Figure 4.15d
Saline treated
ApoE 3/3
males and
female
(plasma).
105
Allopregnanolone and its analogues potentiate mitochondrial function
Figure 4.17 represents a graphical representation of data obtained from the extracellular
flux analyzer. The data shows that while no treatment and vehicle treated hNSCs exhibit
much lower level of mitochondrial potentiation compared to hNSCs treated with Allo and
its analogues as measured OCR. Maximal respiration and spare respiratory capacity of
hNSCs were increased by 39% and 53% respectively after treatment with Allo for 24h,
compared to vehicle. It is also interesting to note that the analogues cluster into groups
with UCI 2-261 100nM, 5β-pregnan-3α-ol-20-one 100nM and Progesterone 100nM
highest potentiation followed by rest of the analogues including Allo. The chemical
Figure 4.16 Allo dose response curve in hNSCs with respect to mitochondrial
potentiation.
The optimal dose of Allo for mitochondrial potentiation in hNSCs is 10-100nM. A dose
of 100nM was chosen to maintain consistency with earlier experiments with rat NSCs.
106
structures and names of these analogues were originally not known to me and at the end
Figure 4.18 Allo and its analogues potentiate mitochondrial function.
Treating hNSCs in vitro with Allo and its analogues shows improved mitochondrial
function when compared to vehicle treated controls, as measured by the extracellular
flux analyzer.
Figure 4.17a
Figure 4.17a-b Allo and its analogues differentially potentiate mitochondrial
function and increase spare respiratory capacity of hNSCs.
Based on their potentiation of mitochondrial function Allo and its analogues can be
divided into two groups that are statistically significantly different from each other.
107
of the experiment I discovered that what I had labelled as Analogue C was Allo from the
lab of Dr. Gee at University of Irvine, and therefore I labelled it as UCI Allo in the graphs.
The aggregated differences in mitochondrial potentiation in response to Allo treatment of
hNSCs is shown in Figure 4.18. I obtained these data from six different experiments. All
the data were normalized to protein content as described earlier. From the graph it is
evident that Allo and its analogues can be divided into two groups based on their
potentiation of mitochondrial function in hNSCs. First group consists of Allo, Ganaxolone,
5a-pregnan-3ß-ol-20-one and 5-Pregnene-3ß-ol-20-one which are statistically not
dissimilar in their effect on mitochondrial potentiation in hNSCs. The second group
consists of UCI 2-261 (a proprietary compound whose chemical structure and name I am
not aware of), 5ß-pregnan-3a-ol-20-one (epi-Allo) and Progesterone which potentiate
mitochondrial function to a magnitude which is higher and statistically significantly different
from the first group of compounds including Allo. Allo increased spare respiratory capacity
(SRC) by 53% in hNSCs, while Allo’s analogue, UCI 2-261, almost doubled (96%) SRC
of hNSC mitochondria. Both our Allo and UCI Allo show little difference in increasing OCR
in hNSCs and provide validation for this experiment. Another point of interest I noted was
that the effect of Allo (and analogues) declined with increase in passage numbers of
Figure 4.17b
108
hNSCs and the effect completely disappeared after passage number 21. The gradual
decline started at either passage 15 or 16 in preliminary experiments (data not shown)
and therefore I did not use any hNSCs beyond passage 14 for these experiments.
Allo ‘pulse’ treatment of hNSCs
The time period of the pulse was defined as the number of hours that elapsed after the
completion of the 30-min. exposure of hNSCs to Allo and before measurement of
mitochondrial function using extracellular flux analyzer. There were five such groups
based on number of hours, 24h(100nM), 24h (1000nM), 8h, 4h and 2h. Additional groups
were the positive control (24h continuous Allo treatment) and vehicle treatment. Treatment
with a higher dose of Allo (1000nM) was carried out in order to evaluate whether a larger
dose of Allo 24 hours before measurement had any beneficial effect on mitochondrial
function in hNSCs compared to 100nM dose of Allo, everything else being the same.
Results (Fig. 4.19a-b) showed that while my positive control of 24h continues treatment
appeared to be the best with respect to potentiation of mitochondrial function in hNSCs,
Figure 4.19a-b Exposure of hNSCs to 30-min. pulse of Allo and its effects on
mitochondrial function.
The effect of the 30-min Allo exposure lasts at least up to 8 hours and the magnitude
of this effect is not significantly different from a 24-hour Allo exposure. Higher dose of
Allo (1000nM) for 30-min has significant deleterious effect on mitochondrial function in
hNSCs that persists even after 24 hours.
Figure 4.19b
Figure 4.19a
109
all three pulse treatments (8h, 4h and 2h) were statistically not significantly different from
the positive control for maximal respiration and for basal respiration only 8h treatment was
not dissimilar to vehicle treated cells, statistically speaking. A 30-min pulse exposure to
Allo 8h before measurement of mitochondrial function is closest to a 24-hour continuous
Allo exposure at maximal respiration but not statistically distinguishable (4.18a-b).
Surprisingly 1000nM dose of Allo did more harm than good and it actually decreased the
mitochondrial function of hNSCs and this effect was visible even 24h after exposure to a
30-min. ‘pulse’ of Allo. 24h ‘pulse’ of 100nM Allo did not fare well either but appeared to
be less deleterious compared to 1000nM but was statistically not significantly different
from the 24h 1000nM ‘pulse’.
Evaluation of Allo as an agonist or antagonist of TSPO
Evaluation of Allo as a TSPO agonist or antagonist produced mixes results as shown in
Figure 4.20. While Allo exhibited its characteristic potentiation of mitochondrial function,
the results from agonist and antagonist were different from my expectations. TSPO
Figure 4.20 Evaluation of Allo as an agonist or antagonist of TSPO in
mitochondria.
Based on data from measurement of mitochondrial function it was is not clear if Allo is
an agonist or antagonist of TSPO even though Allo exhibits a more antagonistic
picture when compared to known agonist and antagonist of TSPO.
110
antagonist PK 11195 statistically significantly potentiated mitochondrial function compared
to vehicle while TSPO agonist FGIN 1-27 was not statistically dissimilar from vehicle. A
combination of the agonist and antagonist led to potentiation of mitochondrial function, not
as much as the positive control (Allo 100nM) but statistically the difference was not
significant. Allo and the antagonist (PK 11195) in combination too potentiated
mitochondrial function but visibly less than Allo and the agonist + antagonist combination.
However, these differences were not statistically significant. When I calculated Spare
Respiratory Capacity, I found that Allo statistically significantly increased SRC compared
Figure 4.21 Allo statistically significantly increased SRC of mitochondria in hNSC
after 24-hour exposure compared to TSPO agonist and antagonist.
However, neither TSPO agonist or antagonist increased SRC significantly although the
antagonist, PK11195, exhibits a trend towards statistical significance, suggesting that
Allo’s action may be antagonistic-like at TSPO.
111
to vehicle (Fig. 4.21), but neither TSPO agonist nor TSPO antagonist nor a combination
increased SRC in a statistically significant manner. TSPO agonist FGIN 1-27 was
statistically not dissimilar from vehicle and significantly different from Allo treatment. These
findings suggest Allo’s mitochondrial potentiation in hNSCs to be a TSPO-antagonist-like
action and not TSPO-agonist-like action.
4.4.6 IMAGING
MRI study was carried out on brains of certain Allo and saline treated mice that I shipped
to University of Arizona. While the imaging process is complete, data analysis is underway
and should be available soon.
4.5 Discussion
I discussed earlier that ApoE4 allele is strongly associated with the risk of developing AD
and AD, as a disease, currently lacks any therapeutic intervention to thwart its march. Allo
is a novel neuroregenerative that has shown promising effects in the 3xTg mouse model
of AD and in this experiment was used to treat ApoE-TR mice in order to determine its
effects on cognition, hippocampal transcriptome, brain imaging and metabolism. This
extensive study was an attempt to apply a systems biology approach to translation
research in AD and at the same time target a genotype that is clearly highly susceptible
to development of AD. In fact, the Allo and saline I used in this experiment as well as the
dose (extrapolated to mice) were identical to the clinical trial in human participants. The
duration of the study was 6 months, double of the phase Ib/IIa trial but in line with our
proposed phase IIb clinical trial in human participants.
The novel object recognition test used to measure cognitive function showed considerable
improvement in both Allo treated males and females of ApoE 4/4 genotype while males
112
and females of ApoE 3/3 genotype failed to show any improvement compared to their
saline treated counterparts. Interestingly, while all groups exhibited variability in
measurement of cognitive function, Allo treated groups are more readily divisible into two
subpopulations: responders and non-responders, which is similar to the results of our
phase Ib/IIa clinical trial. While the difference in cognition between Allo treated ApoE 4/4
males was trending towards significance (p=0.08) when difference in exploration time was
measured, statistical significance was achieved when discrimination index (DI) of both
these groups were compared. DI is derived by normalizing the difference in exploration
time to the total time spent exploring by the mouse and thus is a ratio. DI is a better
measure of cognitive improvement since it takes into account the total time spent
exploring.
The hippocampal RNA-Seq resulted in extensive amount of data from comparing ten
separate groups. Allo treated ApoE 4/4 females were the only group that exhibited
upregulation of PDGF-BB when compared to their saline treated counterparts. There are
four known PDGF genes with broad expression patterns and except for PDGF-DD, all
others are expressed in the brain in varying amounts (Fredriksson et al., 2004; LaRochelle
et al., 2001). These literature findings agree with absence of PDGF-DD and the expression
of the other three in hippocampus (Fredriksson et al., 2004) of ApoE 4/4 Allo treated
female mice. While PDGF-BB is necessary for maintenance of healthy BBB, PDGFRα
signaling may contribute to BBB disruption (Ma et al., 2011). Allo’s downregulation of
PDGFRα is therefore in agreement with its suggested action of protecting BBB integrity.
However, further downstream regulation of the PDGF signaling pathway needs additional
elucidation.
113
To reiterate, estrogens play a major role in maintaining BBB integrity via PDGF as
described in Chapter 1. ApoE 4/4 genotype is also risk factor for increased blood brain
barrier permeability and multiple studies have found that magnitude of BBB damage may
be determined by which ApoE allele is present i.e. E2, E3 or E4 (Hultman et al., 2013;
Salloway et al., 2002; Zipser et al., 2007; Zonneveld et al., 2014). ApoE4 carriers are
vulnerable to BBB disruption in the prodromal phase of cognitive decline (Halliday et al.,
2013). Mechanistically, ApoE4 appears to expedite death of pericytes which could
contribute to BBB compromise (Halliday et al., 2016). Literature suggests that in a model
of TBI, Allo inhibits proinflammatory pathways and reduces cerebral edema (Ishrat et al.,
2010). Also, in a different study, Allo reduced the magnitude of BBB compromise as well
as size of infarct in focal ischemia (He et al., 2004). Based on these findings it is likely that
the predicted upregulation of PDGF-BB by Allo could ameliorate the increased blood brain
barrier permeability and could potentially be one of the mechanisms by which Allo
neutralizes the deleterious effects of ApoE 4/4 genotype on BBB.
The canonical pathways that I found when I compared the hippocampal transcriptome of
Allo treated ApoE 4/4 females to their saline treated counterparts suggest that various
immune system pathways along with metabolic and signaling cascade pathways are
regulated by Allo. The presence of immune system markers in a perfused hippocampus
strongly suggests immune system regulation of the brain parenchyma by Allo. Another
pathway of interest is Acyl-CoA hydrolysis pathway which in conjunction with data from
the metabolomic study suggests increase in energy metabolism utilizing lipids.
Effects of Allo on hippocampal transcriptome were found to be genotype specific. Allo
downregulated a large number of pathways in ApoE 4/4 animals, both males and females
while the same pathways were either upregulated or unchanged in ApoE 3/3 mice. This
114
suggests that Allo maybe useful only in certain genotypes of ApoE such as E4 and this is
corroborated by our clinical trial data from phase Ib/IIa trial.
It was unfortunate that only a few metabolites were significantly different between most
groups of the study, however, I had only five animals in each group for the metabolomic
panel study and therefore increasing the number of animals in each group may result in
larger number of metabolites that meet statistical significance after correction for multiple
comparisons. Some effect of treatment, genotype and sex is observed in the data from
the metabolomic panel since the results from the inter-group comparisons discussed
above show clear separation of individual animals into groups on the PCA plot. Data
comparing ApoE 4/4 Allo females to saline treated counterparts suggests that Allo
increases glycerophospholipids and reduces acylcarnitines. While data from the cortex
showed little difference between saline treated ApoE 3/3 females and ApoE 3/4 females,
several glycerophospholipids, amino acids, biogenic amines and acylcarnitines were
significantly different in plasma of saline treated ApoE 3/3 and ApoE 3/4 female mice and
majority of glycerophospholipids and two amino acids reached significance after adjusting
the p-value. ApoE 3/4 had higher glycerophospholipids than ApoE 3/3, which potentially
can be utilized as a metabolic phenotype, if these findings are reproducible. In both plasma
and cortex, several glycerophospholipid were different between ApoE 3/4 and ApoE 4/4
female mice, even though none are statistically significantly different. When saline treated
ApoE 3/3 and ApoE 4/4 females were compared, 17 metabolites different in cortex and 26
metabolites were different in plasma. Even though none reached significance with
corrected p-value, PCA plot shows a clear separation between the two groups in both
plasma and cortex and a much tighter clustering of saline treated ApoE 3/3 female mice
in data obtained from plasma. These results indicate that females of all three genotypes
115
exhibit a unique metabolomic signature and a metabolomic panel such as the one we
have used can potentially be used to classify ApoE genotypes into metabolomic
phenotypes, if this data can be verified in human participants. The presence of human
ApoE genes in these mice suggest a strong possibility that humans too may exhibit a
similar pattern.
In saline treated ApoE 3/3 and 4/4 mice, plasma glycerophospholipids were higher in
males in comparison to females. Other findings suggested higher arginine catabolism in
males compared to females. Also plasma αAA has been shown to be lower in
APP/PS1 Alzheimer's model (Pan et al., 2016). Additionally, increase in spermidine,
spermine, putrescine have been reported in Alzheimer disease pathology (Inoue et al.,
2013). Thus, lower αAA in female mice may indicate greater susceptibility or likelihood of
AD but higher spermidine and lower putrescine and αAA are inconclusive to evaluate risk
of AD based on data from the metabolomic panel alone. Both the metabolomic panel and
RNA-Seq are expensive studies and therefore five animals per group for ten groups was
considered cost-effective. However, the effects of Allo treatment may be too subtle and
much larger number of animals (n=9 to 12) per group may be necessary for the
metabolomic study to resolve any statistically significant differences, if truly present. Also,
a cell-type specific RNA-Seq of hippocampal neurons may reveal greater details regarding
the transcriptomic regulation of observed effects of Allo.
Mechanistically, Allo potentiated mitochondrial function in hNSCs in a dose dependent
manner: 100nM was optimal dose of whereas 1mM suppressed mitochondrial function.
Epi-Allo, UCI 2-261 and progesterone significantly increased mitochondrial respiration to
a greater magnitude than Allo. 30-min of exposure to Allo up to 8 hours before
measurement was as effective as 24h of continuous Allo exposure. Allo 30-min. pulse
116
resulted in maximal mitochondrial respiration 2-8h post-treatment and this further supports
the 30-min. infusion protocol of Allo phase Ib/IIa clinical trial for mild cognitive impairment
(MCI) or early AD (ClinicalTrials.gov Identifier: NCT02221622).
Published studies have assigned a central role to mitochondrial dysfunction in
pathogenesis of neurodegenerative disorders, including AD (Beal, 2002; Blass et al.,
2000; Brinton, 2008), corroborated by a shift in cerebral glucose utilization observed in AD
patients compared to controls (Hoyer, 1991; Ishii et al., 1997), which lead to oxidative
stress and eventual cognitive decline (Atamna and Frey, 2007; Reddy and Beal, 2008).
Aβ binding alcohol dehydrogenase (ABAD) is a mitochondrial Aβ binding enzyme
suspected as a cause of mitochondrial dysfunction (Lustbader et al., 2004). In both AD
patients and the 3xTgAD mouse model of AD, ABAD expression has been found to be
correlated with the magnitude of Aβ load and it has been demonstrated that Allo inhibits
ABAD (Yao et al., 2009).
Based on these results and findings in published literature it is evident that Allo
significantly potentiates mitochondrial function and Allo, either alone or with analogues
could be used to improve mitochondrial function in neurodegenerative diseases.
Mitochondrial spare respiratory capacity (SRC) is considered a crucial measure of
mitochondrial function and is defined as the difference between basal ATP production and
its maximal activity, and higher SRC is believed to assist cells in countering oxidative
stress (Hill et al., 2009). Therefore, Allo induced 53% increase in SRC of hNSCs suggests
a strong neuroprotective effect. Allo’s analogue, UCI 2-261 almost doubled SRC and
deserves evaluation as a novel therapeutic. Potentiation of mitochondrial function of
human neural stem cells by Allo and its analogues is in support of its potential to be the
first-of-its-kind neuroregenerative therapeutic. It was also observed that some analogues
117
of Allo such as epi-Allo, UCI 2-261 and progesterone may be better candidates for
improving mitochondrial function in vivo if they possess favorable safety and toxicity
profiles and satisfactory pharmacokinetic characteristics. The results obtained by ‘pulse’
treatment of hNSCs with Allo strengthens our rational for infusing Allo in human
participants over a 30-min period in phase Ib/IIa clinical trial and suggests that effects of
Allo on neural stem cells in humans likely persist at least up to 8 hours after exposure.
This long-term effect of Allo could be attributable to protein changes triggered by Allo’s
allosteric activation of GABAA receptor channels. Detailed elucidation of the exact
mechanism is warranted.
Evaluation of Allo as a TSPO agonist or antagonist produced unexpected results which
suggests Allo’s mitochondrial potentiation action is not due to agonism of TSPO since
TSPO agonist FGIN 1-27 produced an action not statistically dissimilar to vehicle
treatment. In fact, antagonism of TSPO (PK 11195) results in potentiation of mitochondrial
function of a magnitude less than that produced by Allo, although this difference in
magnitude is not statistically significant. This suggests that TSPO antagonism could
potentially be one of the mechanism of action through which Allo potentiates mitochondrial
function, and Allo likely acts through other mechanisms and signaling pathways as well.
The hypothesis that Allo’s agonistic action on TSPO is responsible for mitochondrial
potentiation in hNSCs was rejected.
4.6 Conclusion
Collectively, the data indicate that Allo improves cognitive function in both female and
male ApoE 4/4 mice with females exhibiting a greater response to Allo treatment.
Metabolomic data are suggestive of an effect that Allo to increases protein metabolism,
118
as well as lipid metabolism to generate acetyl-CoA to feed into the TCA cycle to generate
ATP in the mitochondria. Further, Allo treatment increased indicators of protein
metabolism. Allo treatment was evident in both females and males and modified by ApoE
genotype. Hippocampal RNA-Seq suggested Allo’s action as an anti-inflammatory agent,
and protective of BBB integrity as well as Allo’s role in cell cycle regulation and metabolic
pathways based on observed changes in hippocampal transcriptome. Allo’s regulation of
the hippocampal transcriptome was found to be genotype specific rather than gender
specific. Further therapeutic development of Allo is underway and a repeat of the
metabolomic and RNA-Seq study with larger number of animals may help confirm the
underlying mechanisms that are responsible for improved cognitive function. While its
neuroregenerative therapeutic potential has been repeatedly demonstrated in vivo, this is
the first time Allo is being tested in a targeted genotype, as an ongoing effort to promote
precision medicine. Allo’s action on mitochondrial function of hNSCs in vitro strongly
supports its credentials as a potential neuroregenerative therapeutic. Allo’s mitochondrial
potentiation action is not due to agonism of TSPO and in fact, antagonism of TSPO could
potentially be one of the mechanisms of action through which Allo potentiates
mitochondrial function in hNSCs. In addition, Allo likely acts through other mechanisms
and signaling pathways as well since it seems to produce higher magnitude of
mitochondrial potentiation than attributable to TSPO antagonism alone. Exact
mechanisms of Allo’s actions remain to be explored.
119
Conclusion
The findings presented herein describe a systems biology approach to a critical period
during chronological and endocrinological aging in women. Menopause obviously is not a
new phenomenon, but its current definitions are relatively novel. Since the 1970s,
menopause has received increased attention in medical, academic, social and popular
spheres (Bell, 1987; Chornesky, 1998; Utian, 2004). In 1938, the first synthetic
replacement hormone was developed which resulted in menopause being characterized
simply as a ‘hormone deficiency disease’ (Ballard et al., 2001; Bell, 1987; Meyer, 2003).
Even though menopause is a near universal endocrine transition state in women affecting
their holistic health, most research studies in menopausal aged women neglect to include
control for menopausal status (Newhart, 2013). The physiological processes that end in
menopause i.e. cessation of menses, take many years to complete and ‘perimenopause’
is used to denote this period in women’s life leading to menopause (Utian, 2004).
Perimenopause has been demonstrated to be a neurological transition state (Brinton et
al., 2015) with far reaching effects on glucose metabolism, mitochondrial bioenergetics
and immune system (Yin et al., 2015) in vivo in the perimenopausal rat model of natural
reproductive senescence described in Chapter 1. In order to better understand the
complex hippocampal biological processes that occur during perimenopause in the
hippocampus, transcriptomic analysis was carried out by sequencing the RNA obtained
from hippocampi of female SD rats.
120
5.1 Transcriptomic and bioinformatic approach to the systems biology of endocrine
aging
The transcriptomic analysis described in Chapter 2 verify some of the earlier published
findings (Yin et al., 2015) and demonstrated that perimenopausal transition in female rats
significantly changes the expression of certain genes in the hippocampus. The
hippocampal RNA-Seq also resulted in an unexpected discovery; it predicted upregulation
of functions such as T cell development and development of lymphocytes, and
downregulation of apoptosis of lymphocytes and leukocytes. Coupled with prediction of
increased BBB permeability by IPA based on the RNA-Seq data and support of published
findings in literature, resulted in a systems biology hypothesis of increased BBB
permeability in female rats during endocrine aging and a potential amelioration of this
phenomenon by Allo.
RNA-Seq demonstrated that both age and endocrine changes influence hippocampal
gene expression in female SD rats and that Regular to Irregular transition at 9-10 months
had higher number of genes that were significantly differentially expressed compared to
the Irregular to Acyclic transition at the same age of 9-10 months in female SD rats. The
surgical menopausal group of Regular OVX 10-10.5 months had the largest number of
significantly differentially expressed genes when compared to Regular cyclers 9m group,
making it evident that natural reproductive senescence model described in Chapter 1 is
more representative of the perimenopausal transition experienced by women.
Analysis of the functions of significantly differentially expressed genes during the Regular
to Irregular transition at 9-10 months as well as pathway analysis using IPA and PANTHER
illustrated the menopausal BBB transition where reduction in sex hormones, mainly
estrogen, that regulates blood brain barrier permeability via PDGF, results in a
121
compromised BBB, making it easier for peripheral blood components, both cellular and
non-cellular, to cross into the brain parenchyma and potentially cause toxicity,
sensitization to previously unfamiliar antigens and immune response. Additionally,
downregulated mitochondrial bioenergetics and insulin signaling pathway were also
observed along with statistically significantly reduced expression of neuroprotective
UCP2. It is possible that aging, coupled with genetic susceptibility and environmental
injury leads to perimenopause, senescence of immune system (Giménez-Llort et al., 2008;
Michaud et al., 2013) and abnormal T cell reactivity to external and self-antigens (Prelog,
2006). Perimenopausal reduction in circulating sex steroids could result in a compromised
blood-brain barrier as evinced in Chapter 3. A dysregulated immune system, age-related
low-grade inflammation and risk factors such as ApoE ε4 too contribute to a more
permeable BBB and neuroinflammation (Dorey et al., 2017; Maftei et al., 2013; Newcombe
et al., 2018). A neuroinflammatory milieu and a leaky blood brain barrier can potentially
trigger cognitive dysfunction (Esteras et al., 2012; Masdeu et al., 2012). BBB itself has
been implicated as a source of this inflammation (Festoff et al., 2016; Grammas et al.,
2011; Liu et al., 2012; Rochfort and Cummins, 2015).
C-Reactive protein (CRP), a measure of peripheral systemic inflammation (Felger et al.,
2018) was found to be higher in AD patients (Gong et al., 2016). CRP is known to have a
negative effect on BBB permeability leading to a compromised barrier (Hsuchou et al.,
2012). In a 9 year long longitudinal study, Corlier et al found that those individuals who at
baseline had higher serum CRP were more likely to have thinner cortex after 9 years
(Corlier et al., 2018).
AD and associated neuropathology does not manifest in the form of clinical symptoms for
many years which implies a long prodromal or preclinical phase of pathogenesis (Sperling
122
et al., 2013). Hormonal changes in women i.e. perimenopausal transition may therefore
contribute to the elevated risk of AD in women (Mosconi et al., 2017).
Based on the perimenopausal hippocampal transcriptome that shows elevated
inflammatory markers, increased leukocytes and lymphocytes infiltration in the brain
coupled with immune system dysregulation, downregulated mitochondrial bioenergetics
and insulin signaling pathway, reduced expression of neuroprotective UCP2 and
downregulated PDGF signaling pathway, and corroborative evidence from literature, it is
likely that a neuroinflammatory axis emerges during perimenopause with the BBB and its
integrity or lack thereof, playing a central role in pathologic processes.
It must be taken into account that this transcriptomic study was carried out only on
homogenized hippocampal tissue and not on other brain regions of the female rats, and
the transcriptome represents the messenger RNA from protein-coding regions of DNA of
all cell types that are found in the hippocampus. Therefore, the pathways and functions
detected by IPA and PANTHER based on this data are limited in their ability to pinpoint
cell-specific pathways or functions. Additionally, RNA species that do not code for proteins
such as microRNA, small nucleolar RNA, small interfering RNA, are known to play crucial
roles in brain evolution, development, plasticity and neurodegenerative diseases (Fatica
and Bozzoni, 2014; Qureshi and Mehler, 2012; Salta and De Strooper, 2012) but are not
covered in this study, and their role in perimenopausal hippocampal transition remains to
be elicited.
5.2 Functional confirmation
In my study I used sodium fluorescein dye to confirm the predicted increase in BBB
permeability in perimenopausal SD rats. An OVX group was added in order to delineate
123
the effect of age and endocrine status on permeability of BBB in aging perimenopausal
female rats. It was notable that 6-month-old regularly cycling female rats had the lowest
BBB permeability in all three tissues analyzed, whereas Regular OVX 7m group of female
rats that belonged to the same group as the Regular cyclers 6m, but which underwent
ovariectomy at 6 months of age had significantly more permeable BBB. The Sham OVX
group at 7m had BBB permeability only slightly elevated compared to Regular cyclers 6m
group and was statistically not significantly different suggesting a stronger role for
endocrine aging in BBB compromise. Both Regular and Irregular cyclers at 9-10 months
had significantly higher BBB permeability compared to the 6-month Regular cycler group
indicating that processes that compromise BBB are initiated much before the stage of
actual reproductive senescence is reached (Acyclic 9-10m). In fact, only hypothalamus
showed BBB permeability significantly different between Regular and Irregular 9-10-
month-old female rats whereas BBB permeability in hippocampus and cortex was
statistically not dissimilar between these two groups, which indicates increase in BBB
permeability precedes the perimenopausal transition as confirmed by vaginal cytology,
which is corroborated in literature (Bacon et al., 2018) where neuroendocrine aging has
been shown to precede perimenopause via analysis of RNA-Seq data based on identical
bioinformatics pipeline discussed in section 2.4.1.
Another point of interest was the increased variability (Figure 3.7) seen between animals
of the same group in Regular, Irregular, Acyclic 9-10m, whereas Regular 6m, Regular
OVX 7m and Regular Sham OVX 7m display lower variability (comparatively lower values
of standard deviation). This variability suggests that not all female SD rats undergo
perimenopause and emerge from this endocrine transition in an identical manner, resulting
in differences in BBB permeability index. This, if found extrapolatable to humans, supports
124
the assertion that not all women will be at increased risk of developing AD or
neurodegenerative diseases during or after endocrinological transition but, for some
women endocrine transition induced hormonal flux portends much greater probability of
severely compromised BBB, a potentially dysfunctional immune system, defective
bioenergetic regulation, cognitive decline and reduced quality of life. Alternatively, this
variability among the female rats could simply be due to differences in exposure to light in
the vivarium, differences in microbiota acquired due to difference in location of each cage
of female rats in the room or due to, unexplained as yet, epigenetic causes. However, if
external factors were affecting the variability, then it is more likely that the variability would
be uniformly distributed across the six different groups.
The transcriptomic study of the hippocampus during perimenopausal transition, the
discovery of a compromised BBB and its functional proof furthers our knowledge of the
complex interaction between cerebrovasculature and the HPA. However, the exact
mechanisms of these interactions are currently unknown and need further elucidation. The
exact role of estrogens, as well as the chemical nature of the specific estrogen
metabolite(s) involved, its interaction with PDGF-BB during perimenopause and the
involvement of endothelial cells, astrocyte foot processes and basement membrane in a
compromised BBB are all potential areas of research that could elicit further evidence in
support of the functional BBB compromise during perimenopause in female rats outlined
herein.
5.3 Role of Allopregnanolone
Multiple studies have shown that ApoE isoforms have differential effects on Aβ
aggregation, degradation and clearance and may affect synaptic function (Bu, 2009;
125
Castellano et al., 2011; Chalmers et al., 2003; Dorey et al., 2014; Nwabuisi-Heath et al.,
2013) and a recent study demonstrated that ApoE4 promotes AD/Aβ-induced
inflammation while ApoE2 inhibits it (Dorey et al., 2017).
Allopregnanolone (Allo), a novel neuroregenerative therapeutic which is structurally an
endogenous neurosteroid, has been in pre-clinical and clinical development (Singh et al.,
2012; Wang and Brinton, 2008; Wang et al., 2005; Wang et al., 2010) and I conducted a
systems biology study to understand the effect of Allo in a targeted genotype (ApoE4),
findings from which could be extrapolated to the clinic since the mice under study have
human ApoE gene. I confirmed Allo’s potential in improvement of cognitive function but
surprisingly Allo’s effect was genotype dependent but sex independent and only males
and females of ApoE4 genotype showed improvement. These results make a strong case
for a phase IIb study in an enriched cohort of ApoE4 carriers, both males and females to
verify Allo’s efficacy in a human population susceptible to AD. Also, based on NOR study
each Allo and saline treated group indicate the existence of sub-populations, such that
some mice respond better to Allo treatment compared to others. This reinforces our
human data from the phase Ib/IIa clinical trial (not enriched for any particular genotype)
that exhibited human responders and non-responders to Allo treatment.
Transcriptomic data from the hippocampus of saline and Allo treated mice supported by
findings in published literature suggest that Allo may be able to ameliorate the increased
BBB permeability observed during endocrine and chronological aging by upregulating
PDGF-BB and downregulating PDGFRα. If this is verified in future studies, an intact BBB
due to Allo treatment could reduce the risk of AD and other diseases such as multiple
sclerosis, where Allo shows promise (Noorbakhsh et al., 2014).
126
In vitro studies in hNSCs showed that Allo and its analogues increase mitochondrial
function and this effect is likely due to downstream changes in transcription and proteins
since the changes are measurable and statistically significant up to 8 hours after Allo
treatment. Allo is dose dependent effect on hNSCs with optimal dose being 10-100nM; a
dose of 1000nM (=1µM) has deleterious effect measurable even 24 hours after Allo
removal. Allo’s potentiation of mitochondrial function is not due to TSPO agonism but
possibly due to antagonism of TSPO as well as due to other signaling pathways.
Some of the factors affecting women’s risk of developing a dysfunctional immune system,
defective bioenergetic regulation and cognitive decline were analyzed in a healthy cohort
of post-menopausal women at risk for cognitive decline using a panel of clinical metabolic
indicators (Rettberg et al., 2016). Further studies of women in perimenopausal transition
are required to confirm findings presented here in humans and explore potential of Allo as
a preventive therapeutic in at-risk populations.
127
Bibliography
Abdullah, M., Chai, P.-S., Chong, M.-Y., Tohit, E.R.M., Ramasamy, R., Pei, C.P., and
Vidyadaran, S. (2012). Gender effect on in vitro lymphocyte subset levels of healthy
individuals. Cellular Immunology 272, 214-219.
Acosta, J.I., Mayer, L., Talboom, J.S., Tsang, C.W., Smith, C.J., Enders, C.K., and
Bimonte-Nelson, H.A. (2009). Transitional versus surgical menopause in a rodent model:
etiology of ovarian hormone loss impacts memory and the acetylcholine system.
Endocrinology 150, 4248-4259.
Acosta, J.I., Mayer, L.P., Braden, B.B., Nonnenmacher, S., Mennenga, S.E., and Bimonte-
Nelson, H.A. (2010). The cognitive effects of conjugated equine estrogens depend on
whether menopause etiology is transitional or surgical. Endocrinology 151, 3795-3804.
Afshan, G., Afzal, N., and Qureshi, S. (2012). CD4+CD25(hi) regulatory T cells in healthy
males and females mediate gender difference in the prevalence of autoimmune diseases.
Clin Lab 58, 567-571.
Alterovitz, G., Xiang, M., Mohan, M., and Ramoni, M.F. (2006). GO PaD: the gene
ontology partition database. Nucleic acids research 35, D322-D327.
Altmann, A., Tian, L., Henderson, V.W., and Greicius, M.D. (2014). Sex modifies the
APOE-related risk of developing Alzheimer disease. Annals of Neurology 75, 563-573.
Amor, S., Puentes, F., Baker, D., and Van Der Valk, P. (2010). Inflammation in
neurodegenerative diseases. Immunology 129, 154-169.
Andorn, A.C., Britton, R.S., Bacon, B.R., and Kalaria, R.N. (1998). Ascorbate-stimulated
lipid peroxidation and non-heme iron concentrations in Alzheimer disease. Molecular and
chemical neuropathology 33, 15-26.
Andrews-Zwilling, Y., Bien-Ly, N., Xu, Q., Li, G., Bernardo, A., Yoon, S.Y., Zwilling, D.,
Yan, T.X., Chen, L., and Huang, Y. (2010). Apolipoprotein E4 causes age- and Tau-
dependent impairment of GABAergic interneurons, leading to learning and memory
deficits in mice. J Neurosci 30, 13707-13717.
Andrews, Z.B., and Horvath, T.L. (2009). Uncoupling protein-2 regulates lifespan in mice.
American Journal of Physiology-Endocrinology and Metabolism 296, E621-E627.
Antunes, M., and Biala, G. (2012). The novel object recognition memory: neurobiology,
test procedure, and its modifications. Cognitive processing 13, 93-110.
Aomatsu, M., Kato, T., Kasahara, E., and Kitagawa, S. (2013). Gender difference in tumor
necrosis factor-α production in human neutrophils stimulated by lipopolysaccharide and
interferon-γ. Biochemical and biophysical research communications 441, 220-225.
Arimoto, J.M., Wong, A., Rozovsky, I., Lin, S.W., Morgan, T.E., and Finch, C.E. (2013).
Age increase of estrogen receptor-α (ERα) in cortical astrocytes impairs neurotrophic
support in male and female rats. Endocrinology 154, 2101-2113.
Armulik, A., Genové, G., Mäe, M., Nisancioglu, M.H., Wallgard, E., Niaudet, C., He, L.,
Norlin, J., Lindblom, P., and Strittmatter, K. (2010). Pericytes regulate the blood–brain
barrier. Nature 468, 557.
Ashburner, M., Ball, C.A., Blake, J.A., Botstein, D., Butler, H., Cherry, J.M., Davis, A.P.,
Dolinski, K., Dwight, S.S., and Eppig, J.T. (2000). Gene Ontology: tool for the unification
of biology. Nature genetics 25, 25.
Association, A.s. (2014). 2014 Alzheimer's disease facts and figures. Alzheimer's &
Dementia 10, e47-e92.
128
Association, A.s. (2018). 2018 Alzheimer's disease facts and figures. Alzheimer's &
Dementia 14, 367-429.
Atallah, A., Mhaouty-Kodja, S., and Grange-Messent, V. (2017). Chronic depletion of
gonadal testosterone leads to blood–brain barrier dysfunction and inflammation in male
mice. Journal of Cerebral Blood Flow & Metabolism 37, 3161-3175.
Atamna, H., and Frey, W.H. (2007). Mechanisms of mitochondrial dysfunction and energy
deficiency in Alzheimer’s disease. Mitochondrion 7, 297-310.
Attems, J., and Jellinger, K.A. (2004). Only cerebral capillary amyloid angiopathy
correlates with Alzheimer pathology—a pilot study. Acta neuropathologica 107, 83-90.
Bacon, E.R., Mishra, A., Wang, Y., Desai, M.k., Yin, F., and Brinton, R.D. (2018).
Neuroendocrine Aging Precedes Perimenopause and is Regulated by DNA Methylation.
Neurobiology of Aging.
Bake, S., and Sohrabji, F. (2004). 17β-Estradiol differentially regulates blood-brain barrier
permeability in young and aging female rats. Endocrinology 145, 5471-5475.
Ballard, K.D., Kuh, D.J., and Wadsworth, M.E. (2001). The role of the menopause in
women’s experiences of the ‘change of life’. Sociology of Health & Illness 23, 397-424.
Baumgartner Jr, W.A., Cohen, K.B., Fox, L.M., Acquaah-Mensah, G., and Hunter, L.
(2007). Manual curation is not sufficient for annotation of genomic databases.
Bioinformatics 23, i41-i48.
Beal, M.F. (2002). Mitochondria in neurodegeneration. In Mitochondrial Disorders
(Springer), pp. 17-35.
Bell, M., and Ball, M. (1986). The correlation of vascular capacity with the parenchymal
lesions of Alzheimer's disease. Canadian Journal of Neurological Sciences 13, 456-461.
Bell, S.E. (1987). Changing ideas: The medicalization of menopause. Social science &
medicine 24, 535-542.
Bellino, F.L. (2000). Nonprimate animal models of menopause: workshop report.
Menopause 7, 14-24.
Bellino, F.L., and Wise, P.M. (2003). Nonhuman primate models of menopause workshop.
Biol Reprod 68, 10-18.
Bentley, D.R., Balasubramanian, S., Swerdlow, H.P., Smith, G.P., Milton, J., Brown, C.G.,
Hall, K.P., Evers, D.J., Barnes, C.L., Bignell, H.R., et al. (2008). Accurate whole human
genome sequencing using reversible terminator chemistry. Nature 456, 53-59.
Berchtold, N.C., Cribbs, D.H., Coleman, P.D., Rogers, J., Head, E., Kim, R., Beach, T.,
Miller, C., Troncoso, J., and Trojanowski, J.Q. (2008). Gene expression changes in the
course of normal brain aging are sexually dimorphic. Proceedings of the National
Academy of Sciences.
Berent-Spillson, A., Persad, C.C., Love, T., Sowers, M., Randolph, J.F., Zubieta, J.-K.,
and Smith, Y.R. (2012). Hormonal Environment Affects Cognition Independent of Age
during the Menopause Transition. The Journal of Clinical Endocrinology & Metabolism 97,
E1686-E1694.
Bethea, C.L., Mirkes, S.J., Shively, C.A., and Adams, M.R. (2000). Steroid regulation of
tryptophan hydroxylase protein in the dorsal raphe of macaques. Biol Psychiatry 47, 562-
576.
Beydoun, H.A., el-Amin, R., McNeal, M., Perry, C., and Archer, D.F. (2013). Reproductive
history and postmenopausal rheumatoid arthritis among women 60 years or older: Third
National Health and Nutrition Examination Survey. Menopause 20, 930-935.
129
Bitran, D., Foley, M., Audette, D., Leslie, N., and Frye, C.A. (2000). Activation of peripheral
mitochondrial benzodiazepine receptors in the hippocampus stimulates allopregnanolone
synthesis and produces anxiolytic-like effects in the rat. Psychopharmacology 151, 64-71.
Blass, J.P., SHEU, R.K.F., and Gibson, G.E. (2000). Inherent abnormalities in energy
metabolism in Alzheimer disease: interaction with cerebrovascular compromise. Annals
of the New York Academy of Sciences 903, 204-221.
Boguski, M.S., Tolstoshev, C.M., and Bassett, D.E., Jr. (1994). Gene discovery in dbEST.
Science 265, 1993-1994.
Borman, S.M., VanDePol, B.J., Kao, S., Thompson, K.E., Sipes, I.G., and Hoyer, P.B.
(1999). A single dose of the ovotoxicant 4-vinylcyclohexene diepoxide is protective in rat
primary ovarian follicles. Toxicol Appl Pharmacol 158, 244-252.
Brinton, R.D. (2008). The healthy cell bias of estrogen action: mitochondrial bioenergetics
and neurological implications. Trends in neurosciences 31, 529-537.
Brinton, R.D., Yao, J., Yin, F., Mack, W.J., and Cadenas, E. (2015). Perimenopause as a
neurological transition state. Nature Reviews Endocrinology 11, 393-405.
Brown, R.E. (2012). Improving animal models for nervous system disorders. Genes Brain
Behav 11, 753-756.
Bu, G. (2009). Apolipoprotein E and its receptors in Alzheimer's disease: pathways,
pathogenesis and therapy. Nature Reviews Neuroscience 10, 333.
Burger, Cahir, Robertson, Groome, Dudley, Green, and Dennerstein (1998). Serum
Inhibins A and B fall differentially as FSH rises in perimenopausal women. Clinical
Endocrinology 48, 809-813.
Burger, H.G. (2006). Physiology and endocrinology of the menopause. Medicine 34, 27-
30.
Carcaillon, L., Brailly-Tabard, S., Ancelin, M.L., Rouaud, O., Dartigues, J.F., Guiochon-
Mantel, A., and Scarabin, P.Y. (2014). High plasma estradiol interacts with diabetes on
risk of dementia in older postmenopausal women. Neurology 82, 504-511.
Castellano, J.M., Kim, J., Stewart, F.R., Jiang, H., DeMattos, R.B., Patterson, B.W.,
Fagan, A.M., Morris, J.C., Mawuenyega, K.G., and Cruchaga, C. (2011). Human apoE
isoforms differentially regulate brain amyloid-β peptide clearance. Science translational
medicine 3, 89ra57-89ra57.
Chakraborty, T.R., and Gore, A.C. (2004). Aging-related changes in ovarian hormones,
their receptors, and neuroendocrine function. Exp Biol Med (Maywood) 229, 977-987.
Chalmers, K., Wilcock, G., and Love, S. (2003). APOEɛ4 influences the pathological
phenotype of Alzheimer's disease by favouring cerebrovascular over parenchymal
accumulation of Aβ protein. Neuropathology and applied neurobiology 29, 231-238.
Chi, O., Barsoum, S., Wen, Y., Liu, X., and Weiss, H. (2004). 17β-estradiol prevents blood-
brain barrier disruption induced by VEGF. Hormone and metabolic research 36, 272-276.
Chi, O., Hunter, C., Liu, X., and Weiss, H. (2005). Effects of 17β-Estradiol on Blood-brain
Barrier Disruption in Focal Ischemia during GABAA Receptor Inhibition. Hormone and
metabolic research 37, 209-213.
Chi, O., Liu, X., and Weiss, H. (2002). Effects of 17β-estradiol on blood-brain barrier
disruption during focal ischemia in rats. Hormone and metabolic research 34, 530-534.
Choi, J.M., Romeo, R.D., Brake, W.G., Bethea, C.L., Rosenwaks, Z., and McEwen, B.S.
(2003). Estradiol increases pre- and post-synaptic proteins in the CA1 region of the
hippocampus in female rhesus macaques (Macaca mulatta). Endocrinology 144, 4734-
4738.
130
Chornesky, A. (1998). Multicultural perspectives on menopause and the climacteric. Affilia
13, 31-46.
Christensen, A., and Pike, C.J. (2015). Menopause, obesity and inflammation: interactive
risk factors for Alzheimer’s disease. Frontiers in Aging Neuroscience 7.
Clarke, G., Grenham, S., Scully, P., Fitzgerald, P., Moloney, R., Shanahan, F., Dinan, T.,
and Cryan, J. (2013). The microbiome-gut-brain axis during early life regulates the
hippocampal serotonergic system in a sex-dependent manner. Molecular psychiatry 18,
666.
Cloonan, N., Forrest, A.R., Kolle, G., Gardiner, B.B., Faulkner, G.J., Brown, M.K., Taylor,
D.F., Steptoe, A.L., Wani, S., Bethel, G., et al. (2008). Stem cell transcriptome profiling via
massive-scale mRNA sequencing. Nat Methods 5, 613-619.
Colangelo, V., Schurr, J., Ball, M.J., Pelaez, R.P., Bazan, N.G., and Lukiw, W.J. (2002).
Gene expression profiling of 12633 genes in Alzheimer hippocampal CA1: transcription
and neurotrophic factor down ‐regulation and up ‐regulation of apoptotic and pro ‐
inflammatory signaling. Journal of neuroscience research 70, 462-473.
Colquhoun, D. (2014). An investigation of the false discovery rate and the
misinterpretation of p-values. R Soc Open Sci 1, 140216.
Corder, E.H., Saunders, A.M., Strittmatter, W.J., Schmechel, D.E., Gaskell, P.C., Small,
G.W., Roses, A.D., Haines, J.L., and Pericak-Vance, M.A. (1993). Gene dose of
apolipoprotein E type 4 allele and the risk of Alzheimer's disease in late onset families.
Science 261, 921-923.
Corlier, F., Hafzalla, G., Faskowitz, J., Kuller, L.H., Becker, J.T., Lopez, O.L., Thompson,
P.M., and Braskie, M.N. (2018). Systemic inflammation as a predictor of brain aging:
Contributions of physical activity, metabolic risk, and genetic risk. NeuroImage 172, 118-
129.
Cotran, R.S., and Karnovsky, M.J. (1967). Vascular leakage induced by horseradish
peroxidase in the rat. Proceedings of the Society for Experimental Biology and Medicine
126, 557-561.
COTRAN, R.S., KARNOVSKY, M.J., and GOTH, A. (1968). Resistance of Wistar/Furth
rats to the mast cell-damaging effect of horseradish peroxidase. Journal of Histochemistry
& Cytochemistry 16, 382-383.
Cramer, S.P., Modvig, S., Simonsen, H.J., Frederiksen, J.L., and Larsson, H.B. (2015).
Permeability of the blood–brain barrier predicts conversion from optic neuritis to multiple
sclerosis. Brain 138, 2571-2583.
Crosby, J.R., Seifert, R.A., Soriano, P., and Bowen-Pope, D.F. (1998). Chimaeric analysis
reveals role of Pdgf receptors in all muscle lineages. Nature genetics 18, 385.
Curtis, R.K., Orešič, M., and Vidal-Puig, A. (2005). Pathways to the analysis of microarray
data. TRENDS in Biotechnology 23, 429-435.
D’Andrea, M.R. (2003). Evidence linking neuronal cell death to autoimmunity in
Alzheimer’s disease. Brain research 982, 19-30.
D’Andrea, M.R. (2005). Add Alzheimer’s disease to the list of autoimmune diseases.
Medical hypotheses 64, 458-463.
Damoiseaux, J.S., Seeley, W.W., Zhou, J., Shirer, W.R., Coppola, G., Karydas, A., Rosen,
H.J., Miller, B.L., Kramer, J.H., Greicius, M.D., et al. (2012). Gender modulates the APOE
ε4 effect in healthy older adults: convergent evidence from functional brain connectivity
and spinal fluid tau levels. J Neurosci 32, 8254-8262.
131
Darland, D., and D'amore, P. (2001). TGFβ is required for the formation of capillary-like
structures in three-dimensional cocultures of 10T1/2 and endothelial cells. Angiogenesis
4, 11-20.
Daugherty, D.J., Selvaraj, V., Chechneva, O.V., Liu, X.-B., Pleasure, D.E., and Deng, W.
(2013). A TSPO ligand is protective in a mouse model of multiple sclerosis. EMBO
Molecular Medicine 5, 891-903.
Deleidi, M., Jäggle, M., and Rubino, G. (2015). Immune aging, dysmetabolism, and
inflammation in neurological diseases. Frontiers in neuroscience 9, 172.
Dennis, G., Sherman, B.T., Hosack, D.A., Yang, J., Gao, W., Lane, H.C., and Lempicki,
R.A. (2003). DAVID: database for annotation, visualization, and integrated discovery.
Genome biology 4, R60.
Diano, S., and Horvath, T.L. (2012). Mitochondrial uncoupling protein 2 (UCP2) in glucose
and lipid metabolism. Trends in Molecular Medicine 18, 52-58.
Diaz Brinton, R. (2012). Minireview: translational animal models of human menopause:
challenges and emerging opportunities. Endocrinology 153, 3571-3578.
Diz-Chaves, Y., Kwiatkowska-Naqvi, A., Von Hulst, H., Pernia, O., Carrero, P., and
Garcia-Segura, L.M. (2012). Behavioral effects of estradiol therapy in ovariectomized rats
depend on the age when the treatment is initiated. Exp Gerontol 47, 93-99.
Dohgu, S., Takata, F., Yamauchi, A., Nakagawa, S., Egawa, T., Naito, M., Tsuruo, T.,
Sawada, Y., Niwa, M., and Kataoka, Y. (2005). Brain pericytes contribute to the induction
and up-regulation of blood–brain barrier functions through transforming growth factor-β
production. Brain research 1038, 208-215.
Dohm, J.C., Lottaz, C., Borodina, T., and Himmelbauer, H. (2008). Substantial biases in
ultra-short read data sets from high-throughput DNA sequencing. Nucleic Acids Res 36,
e105.
Dorey, E., Bamji-Mirza, M., Najem, D., Li, Y., Liu, H., Callaghan, D., Walker, D., Lue, L.-
F., Stanimirovic, D., and Zhang, W. (2017). Apolipoprotein E isoforms differentially
regulate Alzheimer’s disease and amyloid-β-induced inflammatory response in vivo and
in vitro. Journal of Alzheimer's Disease 57, 1265-1279.
Dorey, E., Chang, N., Liu, Q.Y., Yang, Z., and Zhang, W. (2014). Apolipoprotein E,
amyloid-beta, and neuroinflammation in Alzheimer’s disease. Neuroscience bulletin 30,
317-330.
Dorheim, M.-A., Tracey, W.R., Pollock, J.S., and Grammas, P. (1994). Nitric oxide
synthase activity is elevated in brain microvessels in Alzheimer′ s disease. Biochemical
and biophysical research communications 205, 659-665.
Downs, J.L., and Wise, P.M. (2009). The role of the brain in female reproductive aging.
Mol Cell Endocrinol 299, 32-38.
Dubal, D.B., Broestl, L., and Worden, K. (2012). Sex and gonadal hormones in mouse
models of Alzheimer’s disease: what is relevant to the human condition? Biol Sex Differ
3, 24.
Edelman, D.A., Jiang, Y., Tyburski, J.G., Wilson, R.F., and Steffes, C.P. (2007a). Cytokine
production in lipopolysaccharide-exposed rat lung pericytes. Journal of Trauma and Acute
Care Surgery 62, 89-93.
Edelman, D.A., Jiang, Y., Tyburski, J.G., Wilson, R.F., and Steffes, C.P. (2007b).
Lipopolysaccharide activation of pericyte’s Toll-like receptor-4 regulates co-culture
permeability. The American journal of surgery 193, 730-735.
132
Emmer, K.M., Russart, K.L., Walker, I., William, H., Nelson, R.J., and DeVries, A.C.
(2018). Effects of light at night on laboratory animals and research outcomes. Behavioral
neuroscience 132, 302.
Ennaceur, A. (2010). One-trial object recognition in rats and mice: methodological and
theoretical issues. Behavioural brain research 215, 244-254.
Enninga, E.A.L., Nevala, W.K., Creedon, D.J., Markovic, S.N., and Holtan, S.G. (2015).
Fetal Sex-Based Differences in Maternal Hormones, Angiogenic Factors, and Immune
Mediators During Pregnancy and the Postpartum Period. Am J Reprod Immunol 73, 251-
262.
Espeland, M.A., Rapp, S.R., Shumaker, S.A., Brunner, R., Manson, J.E., Sherwin, B.B.,
Hsia, J., Margolis, K.L., Hogan, P.E., Wallace, R., et al. (2004). Conjugated equine
estrogens and global cognitive funtion in postmenopausal women: Women's Health
Initiative Memory Study. Journal of the American Medical Association 291, 2959-2968.
Esteras, N., Alquézar, C., Bartolomé, F., Antequera, D., Barrios, L., Carro, E., Cerdán, S.,
and Martín-Requero, Á. (2012). Systematic evaluation of magnetic resonance imaging
and spectroscopy techniques for imaging a transgenic model of Alzheimer's disease
(AβPP/PS1). Journal of Alzheimer's Disease 30, 337-353.
Fairweather, D., Frisancho-Kiss, S., and Rose, N.R. (2008). Sex Differences in
Autoimmune Disease from a Pathological Perspective. The American Journal of
Pathology 173, 600-609.
Fan, H., Dong, G., Zhao, G., Liu, F., Yao, G., Zhu, Y., and Hou, Y. (2014). Gender
Differences of B Cell Signature in Healthy Subjects Underlie Disparities in Incidence and
Course of SLE Related to Estrogen. Journal of Immunology Research 2014, 1-17.
Farrer, L.A., Cupples, L.A., Haines, J.L., Hyman, B., Kukull, W.A., Mayeux, R., Myers,
R.H., Pericak-Vance, M.A., Risch, N., and van Duijn, C.M. (1997a). Effects of age, sex,
and ethnicity on the association between apolipoprotein E genotype and Alzheimer
disease. A meta-analysis. APOE and Alzheimer Disease Meta Analysis Consortium.
JAMA 278, 1349-1356.
Farrer, L.A., Cupples, L.A., Haines, J.L., Hyman, B., Kukull, W.A., Mayeux, R., Myers,
R.H., Pericak-Vance, M.A., Risch, N., and Van Duijn, C.M. (1997b). Effects of age, sex,
and ethnicity on the association between apolipoprotein E genotype and Alzheimer
disease: A meta-analysis. Journal of the American Medical Association 278, 1349-1356.
Fatica, A., and Bozzoni, I. (2014). Long non-coding RNAs: new players in cell
differentiation and development. Nature Reviews Genetics 15, 7.
Felger, J.C., Haroon, E., Patel, T.A., Goldsmith, D.R., Wommack, E.C., Woolwine, B.J.,
Le, N.A., Feinberg, R., Tansey, M.G., and Miller, A.H. (2018). What does plasma CRP tell
us about peripheral and central inflammation in depression? Mol Psychiatry.
Festoff, B.W., Sajja, R.K., van Dreden, P., and Cucullo, L. (2016). HMGB1 and thrombin
mediate the blood-brain barrier dysfunction acting as biomarkers of neuroinflammation
and progression to neurodegeneration in Alzheimer’s disease. Journal of
Neuroinflammation 13, 194.
Fiala, M., Liu, Q., Sayre, J., Pop, V., Brahmandam, V., Graves, M., and Vinters, H. (2002).
Cyclooxygenase ‐2 ‐positive macrophages infiltrate the Alzheimer’s disease brain and
damage the blood–brain barrier. European journal of clinical investigation 32, 360-371.
Finch, C.E. (2014). The menopause and aging, a comparative perspective. The Journal
of Steroid Biochemistry and Molecular Biology 142, 132-141.
Finch, C.E., Felicio, L.S., Mobbs, C.V., and Nelson, J.F. (1984). Ovarian and steroidal
influences on neuroendocrine aging processes in female rodents. Endocr Rev 5, 467-497.
133
Fischer, V., Siddiqi, A., and Yusufaly, Y. (1990). Altered angioarchitecture in selected
areas of brains with Alzheimer's disease. Acta neuropathologica 79, 672-679.
Flaws, J.A., Doerr, J.K., Sipes, I.G., and Hoyer, P.B. (1994). Destruction of preantral
follicles in adult rats by 4-vinyl-1-cyclohexene diepoxide. Reprod Toxicol 8, 509-514.
Fleisher, A.S., Sun, S., Taylor, C., Ward, C.P., Gamst, A.C., Petersen, R.C., Jack, C.R.,
Aisen, P.S., and Thal, L.J. (2008). Volumetric MRI vs clinical predictors of Alzheimer
disease in mild cognitive impairment. Neurology 70, 191-199.
Foster, T.C., Sharrow, K.M., Kumar, A., and Masse, J. (2003). Interaction of age and
chronic estradiol replacement on memory and markers of brain aging. Neurobiol Aging 24,
839-852.
Fredriksson, L., Li, H., and Eriksson, U. (2004). The PDGF family: four gene products form
five dimeric isoforms. Cytokine & Growth Factor Reviews 15, 197-204.
Fung, A., Vizcaychipi, M., Lloyd, D., Wan, Y., and Ma, D. (2012). Central nervous system
inflammation in disease related conditions: mechanistic prospects. Brain research 1446,
144-155.
Gaceb, A., Özen, I., Padel, T., Barbariga, M., and Paul, G. (2018). Pericytes secrete pro-
regenerative molecules in response to platelet-derived growth factor-BB. Journal of
Cerebral Blood Flow & Metabolism 38, 45-57.
Gammal, E., and Zuk, A. (1980). Effect of ethinyl estradiol on endothelial permeability to
125I-labeled albumin in female rats. Experimental and molecular pathology 32, 91-101.
Genazzani, A.R., Bernardi, F., Pluchino, N., Begliuomini, S., Lenzi, E., Casarosa, E., and
Luisi, M. (2005). Endocrinology of menopausal transition and its brain implications. CNS
spectrums 10, 449-457.
Genin, E., Hannequin, D., Wallon, D., Sleegers, K., Hiltunen, M., Combarros, O., Bullido,
M.J., Engelborghs, S., De Deyn, P., Berr, C., et al. (2011). APOE and Alzheimer disease:
a major gene with semi-dominant inheritance. Mol Psychiatry 16, 903-907.
Gerhard, D.S., Wagner, L., Feingold, E.A., Shenmen, C.M., Grouse, L.H., Schuler, G.,
Klein, S.L., Old, S., Rasooly, R., Good, P., et al. (2004). The status, quality, and expansion
of the NIH full-length cDNA project: the Mammalian Gene Collection (MGC). Genome Res
14, 2121-2127.
Giménez-Llort, L., Arranz, L., Maté, I., and De la Fuente, M. (2008). Gender-specific
neuroimmunoendocrine aging in a triple-transgenic 3× Tg-AD mouse model for
Alzheimer’s disease and its relation with longevity. Neuroimmunomodulation 15, 331-343.
Giron-Gonzalez, J., Moral, F., Elvira, J., Garcia-Gil, D., Guerrero, F., Gavilan, I., and
Escobar, L. (2000). Consistent production of a higher TH1:TH2 cytokine ratio by
stimulated T cells in men compared with women. European Journal of Endocrinology 143,
31-36.
Gleason, C.E., Dowling, N.M., Wharton, W., Manson, J.E., Miller, V.M., Atwood, C.S.,
Brinton, E.A., Cedars, M.I., Lobo, R.A., Merriam, G.R., et al. (2015). Effects of Hormone
Therapy on Cognition and Mood in Recently Postmenopausal Women: Findings from the
Randomized, Controlled KEEPS–Cognitive and Affective Study. PLOS Medicine 12,
e1001833.
Gong, C., Wei, D., Wang, Y., Ma, J., Yuan, C., Zhang, W., Yu, G., and Zhao, Y. (2016). A
meta-analysis of C-reactive protein in patients with Alzheimer’s disease. American Journal
of Alzheimer's Disease & Other Dementias® 31, 194-200.
Grammas, P., Martinez, J., and Miller, B. (2011). Cerebral microvascular endothelium and
the pathogenesis of neurodegenerative diseases. Expert reviews in molecular medicine
13.
134
Greendale, G.A., Wight, R.G., Huang, M.-H., Avis, N., Gold, E.B., Joffe, H., Seeman, T.,
Vuge, M., and Karlamangla, A.S. (2010). Menopause-associated symptoms and cognitive
performance: results from the study of women's health across the nation. American journal
of epidemiology 171, 1214-1224.
Griesbeck, M., Ziegler, S., Laffont, S., Smith, N., Chauveau, L., Tomezsko, P., Sharei, A.,
Kourjian, G., Porichis, F., and Hart, M. (2015). Sex differences in plasmacytoid dendritic
cell levels of IRF5 drive higher IFN-α production in women. The Journal of Immunology,
1501684.
Grytten, N., Aarseth, J.H., Lunde, H.M.B., and Myhr, K.M. (2015). A 60-year follow-up of
the incidence and prevalence of multiple sclerosis in Hordaland County, Western Norway.
Journal of Neurology, Neurosurgery & Psychiatry, jnnp-2014-309906.
Gulinello, M., and Etgen, A.M. (2005). Sexually dimorphic hormonal regulation of the gap
junction protein, CX43, in rats and altered female reproductive function in CX43+/− mice.
Brain research 1045, 107-115.
Guo, J., Xu, N., Li, Z., Zhang, S., Wu, J., Kim, D.H., Sano Marma, M., Meng, Q., Cao, H.,
Li, X., et al. (2008). Four-color DNA sequencing with 3'-O-modified nucleotide reversible
terminators and chemically cleavable fluorescent dideoxynucleotides. Proc Natl Acad Sci
U S A 105, 9145-9150.
Haines, B.A., Mehta, S.L., Pratt, S.M., Warden, C.H., and Li, P.A. (2010). Deletion of
mitochondrial uncoupling protein-2 increases ischemic brain damage after transient focal
ischemia by altering gene expression patterns and enhancing inflammatory cytokines.
Journal of Cerebral Blood Flow & Metabolism 30, 1825-1833.
Hall, A.O.H., Beiting, D.P., Tato, C., John, B., Oldenhove, G., Lombana, C.G., Pritchard,
G.H., Silver, J.S., Bouladoux, N., and Stumhofer, J.S. (2012). The cytokines interleukin 27
and interferon-γ promote distinct Treg cell populations required to limit infection-induced
pathology. immunity 37, 511-523.
Halliday, M.R., Pomara, N., Sagare, A.P., Mack, W.J., Frangione, B., and Zlokovic, B.V.
(2013). Relationship between cyclophilin a levels and matrix metalloproteinase 9 activity
in cerebrospinal fluid of cognitively normal apolipoprotein e4 carriers and blood-brain
barrier breakdown. JAMA Neurology 70, 1198-1200.
Halliday, M.R., Rege, S.V., Ma, Q., Zhao, Z., Miller, C.A., Winkler, E.A., and Zlokovic, B.V.
(2016). Accelerated pericyte degeneration and blood–brain barrier breakdown in
apolipoprotein E4 carriers with Alzheimer’s disease. Journal of Cerebral Blood Flow &
Metabolism 36, 216-227.
Hannah, M.F., Bajic, V.B., and Klein, S.L. (2008). Sex differences in the recognition of and
innate antiviral responses to Seoul virus in Norway rats. Brain, behavior, and immunity 22,
503-516.
Hass, D.T., and Barnstable, C.J. (2016). Uncoupling protein 2 in the glial response to
stress: implications for neuroprotection. Neural regeneration research 11, 1197.
Hawkins, B.T., and Davis, T.P. (2005). The blood-brain barrier/neurovascular unit in health
and disease. Pharmacological reviews 57, 173-185.
He, J., Evans, C.-O., Hoffman, S.W., Oyesiku, N.M., and Stein, D.G. (2004). Progesterone
and allopregnanolone reduce inflammatory cytokines after traumatic brain injury.
Experimental neurology 189, 404-412.
Hellström, M., Gerhardt, H., Kalén, M., Li, X., Eriksson, U., Wolburg, H., and Betsholtz, C.
(2001). Lack of pericytes leads to endothelial hyperplasia and abnormal vascular
morphogenesis. The Journal of cell biology 153, 543-554.
135
Henderson, V.W., and Brinton, R.D. (2010). Menopause and mitochondria: windows into
estrogen effects on Alzheimer's disease risk and therapy. Prog Brain Res 182, 77-96.
Henderson, V.W., St John, J.A., Hodis, H.N., McCleary, C.A., Stanczyk, F.Z., Shoupe, D.,
Kono, N., Dustin, L., Allayee, H., and Mack, W.J. (2016). Cognitive effects of estradiol
after menopause. Neurology 87, 699-708.
Hermes, G., Nagy, D., Waterson, M., Zsarnovszky, A., Varela, L., Hajos, M., and Horvath,
T.L. (2016). Role of mitochondrial uncoupling protein-2 (UCP2) in higher brain functions,
neuronal plasticity and network oscillation. Molecular Metabolism 5, 415-421.
Hewagama, A., Patel, D., Yarlagadda, S., Strickland, F.M., and Richardson, B.C. (2009).
Stronger inflammatory/cytotoxic T-cell response in women identified by microarray
analysis. Genes & Immunity 10, 509-516.
Heywood, L. (1980). Testosterone levels in the male laboratory rat: variation under
experimental conditions. International journal of Andrology 3, 519-529.
Hill, B.G., Dranka, B.P., Zou, L., Chatham, J.C., and Darley-Usmar, V.M. (2009).
Importance of the bioenergetic reserve capacity in response to cardiomyocyte stress
induced by 4-hydroxynonenal. Biochemical Journal 424, 99-107.
Hirshfield, A.N. (1991). Development of follicles in the mammalian ovary. Int Rev Cytol
124, 43-101.
Holtzman, D.M., Herz, J., and Bu, G. (2012). Apolipoprotein E and apolipoprotein E
receptors: normal biology and roles in Alzheimer disease. Cold Spring Harbor
perspectives in medicine, a006312.
Hoyer, P.B., Devine, P.J., Hu, X., Thompson, K.E., and Sipes, I.G. (2001). Ovarian toxicity
of 4-vinylcyclohexene diepoxide: a mechanistic model. Toxicol Pathol 29, 91-99.
Hoyer, S. (1991). Abnormalities of glucose metabolism in Alzheimer's disease. Annals of
the New York Academy of Sciences 640, 53-58.
Hsuchou, H., Kastin, A.J., Mishra, P.K., and Pan, W. (2012). C-reactive protein increases
BBB permeability: implications for obesity and neuroinflammation. Cellular physiology and
biochemistry 30, 1109-1119.
Hu, X., Christian, P., Sipes, I.G., and Hoyer, P.B. (2001a). Expression and redistribution
of cellular Bad, Bax, and Bcl-X(L) protein is associated with VCD-induced ovotoxicity in
rats. Biol Reprod 65, 1489-1495.
Hu, X., Christian, P.J., Thompson, K.E., Sipes, I.G., and Hoyer, P.B. (2001b). Apoptosis
induced in rats by 4-vinylcyclohexene diepoxide is associated with activation of the
caspase cascades. Biol Reprod 65, 87-93.
Huang, D.W., Sherman, B.T., and Lempicki, R.A. (2008a). Bioinformatics enrichment
tools: paths toward the comprehensive functional analysis of large gene lists. Nucleic
acids research 37, 1-13.
Huang, D.W., Sherman, B.T., and Lempicki, R.A. (2008b). Systematic and integrative
analysis of large gene lists using DAVID bioinformatics resources. Nature protocols 4, 44.
Huang, Y., and Mucke, L. (2012). Alzheimer mechanisms and therapeutic strategies. Cell
148, 1204-1222.
Huber, M., Steinwald, V., Guralnik, A., Brüstle, A., Kleemann, P., Rosenplänter, C.,
Decker, T., and Lohoff, M. (2007). IL-27 inhibits the development of regulatory T cells via
STAT3. International immunology 20, 223-234.
Hultman, K., Strickland, S., and Norris, E.H. (2013). The APOE ε4/ε4 genotype potentiates
vascular fibrin (ogen) deposition in amyloid-laden vessels in the brains of Alzheimer's
disease patients. Journal of Cerebral Blood Flow & Metabolism 33, 1251-1258.
136
Humphreys, G.I., Ziegler, Y.S., and Nardulli, A.M. (2014). 17β-estradiol modulates gene
expression in the female mouse cerebral cortex. PloS one 9, e111975.
Hunter, C., and Kastelein, R. (2012a). Fifteen years of interleukin-27-discovery, advances
and translation. Immunity 37, 960.
Hunter, C.A., and Kastelein, R. (2012b). Interleukin-27: balancing protective and
pathological immunity. Immunity 37, 960-969.
Iadecola, C. (2004). Neurovascular regulation in the normal brain and in Alzheimer's
disease. Nature Reviews Neuroscience 5, 347.
Inoue, K., Tsutsui, H., Akatsu, H., Hashizume, Y., Matsukawa, N., Yamamoto, T., and
Toyo'Oka, T. (2013). Metabolic profiling of Alzheimer's disease brains. Scientific reports
3, 2364.
Irwin, R.W., and Brinton, R.D. (2014). Allopregnanolone as regenerative therapeutic for
Alzheimer's disease: translational development and clinical promise. Prog Neurobiol 113,
40-55.
Ishii, K., Sasaki, M., Kitagaki, H., Yamaji, S., Sakamoto, S., Matsuda, K., and Mori, E.
(1997). Reduction of cerebellar glucose metabolism in advanced Alzheimer's disease.
Journal of Nuclear Medicine 38, 925-928.
Ishrat, T., Sayeed, I., Atif, F., Hua, F., and Stein, D.G. (2010). Progesterone and
allopregnanolone attenuate blood–brain barrier dysfunction following permanent focal
ischemia by regulating the expression of matrix metalloproteinases. Experimental
neurology 226, 183-190.
Jellinger, K. (2002). Alzheimer disease and cerebrovascular pathology: an update. Journal
of neural transmission 109, 813-836.
Jiang, X., Andjelkovic, A.V., Zhu, L., Yang, T., Bennett, M.V.L., Chen, J., Keep, R.F., and
Shi, Y. (2018). Blood-brain barrier dysfunction and recovery after ischemic stroke.
Progress in Neurobiology 163-164, 144-171.
Jucker, M. (2010). The benefits and limitations of animal models for translational research
in neurodegenerative diseases. Nat Med 16, 1210-1214.
Kao, S.W., Sipes, I.G., and Hoyer, P.B. (1999). Early effects of ovotoxicity induced by 4-
vinylcyclohexene diepoxide in rats and mice. Reprod Toxicol 13, 67-75.
Kaplan, J.R., Chen, H., Appt, S.E., Lees, C.J., Franke, A.A., Berga, S.L., Wilson, M.E.,
Manuck, S.B., and Clarkson, T.B. (2010). Impairment of ovarian function and associated
health-related abnormalities are attributable to low social status in premenopausal
monkeys and not mitigated by a high-isoflavone soy diet. Hum Reprod 25, 3083-3094.
Kappos, L., Hardmeier, M., Gugleta, K., Ecsedi, M., Kappos, L., and Disanto, G. (2015).
A 10-year follow-up of the European multicenter trial of interferon β-1b in secondary-
progressive multiple sclerosis.
Kaya, M., and Ahishali, B. (2011). Assessment of permeability in barrier type of
endothelium in brain using tracers: Evans blue, sodium fluorescein, and horseradish
peroxidase. In Permeability Barrier (Springer), pp. 369-382.
Kermath, B.A., and Gore, A.C. (2012). Neuroendocrine control of the transition to
reproductive senescence: lessons learned from the female rodent model.
Neuroendocrinology 96, 1-12.
Khan-Dawood, F., Yang, J., and Dawood, M. (1996). Localization and expression of
zonula occludens-1 tight junction-associated protein in baboon (Papio anubis) corpora
lutea. Human reproduction 11, 1262-1267.
Khatri, P., and Drăghici, S. (2005). Ontological analysis of gene expression data: current
tools, limitations, and open problems. Bioinformatics 21, 3587-3595.
137
Kim, D., Pertea, G., Trapnell, C., Pimentel, H., Kelley, R., and Salzberg, S.L. (2013a).
TopHat2: accurate alignment of transcriptomes in the presence of insertions, deletions
and gene fusions. Genome Biol 14, R36.
Kim, G., Shinnakasu, R., Saris, C.J., Cheroutre, H., and Kronenberg, M. (2013b). A novel
role for IL-27 in mediating the survival of activated mouse CD4 T lymphocytes. The Journal
of Immunology, 1201017.
Klein, S.L., and Flanagan, K.L. (2016). Sex differences in immune responses. Nature
Reviews Immunology 16, 626.
Kodzius, R., Kojima, M., Nishiyori, H., Nakamura, M., Fukuda, S., Tagami, M., Sasaki, D.,
Imamura, K., Kai, C., Harbers, M., et al. (2006). CAGE: cap analysis of gene expression.
Nat Methods 3, 211-222.
Koebele, S.V., and Bimonte-Nelson, H.A. (2016). Modeling menopause: The utility of
rodents in translational behavioral endocrinology research. Maturitas 87, 5-17.
Kozler, P., and Pokorny, J. (2003). Altered blood-brain barrier permeability and its effect
on the distribution of Evans blue and sodium fluorescein in the rat brain applied by
intracarotid injection. Physiol Res 52, 607-614.
Krämer, A., Green, J., Pollard Jr, J., and Tugendreich, S. (2013). Causal analysis
approaches in ingenuity pathway analysis. Bioinformatics 30, 523-530.
LaRochelle, W.J., Jeffers, M., McDonald, W.F., Chillakuru, R.A., Giese, N.A., Lokker, N.A.,
Sullivan, C., Boldog, F.L., Yang, M., and Vernet, C. (2001). PDGF-D, a new protease-
activated growth factor. Nature cell biology 3, 517.
Laughlin, G.A., Kritz-Silverstein, D., and Barrett-Connor, E. (2010). Endogenous
oestrogens predict 4-year decline in verbal fluency in postmenopausal women: The
Rancho Bernardo Study. Clinical Endocrinology 72, 99-106.
Lee, B.W., Yap, H.K., Chew, F.T., Quah, T.C., Prabhakaran, K., Chan, G.S., Wong, S.C.,
and Seah, C.C. (1996). Age- and sex-related changes in lymphocyte subpopulations of
healthy Asian subjects: from birth to adulthood. Cytometry 26, 8-15.
Lee, C.-K., Weindruch, R., and Prolla, T.A. (2000). Gene-expression profile of the ageing
brain in mice. Nature genetics 25, 294.
Leger, M., Quiedeville, A., Bouet, V., Haelewyn, B., Boulouard, M., Schumann-Bard, P.,
and Freret, T. (2013). Object recognition test in mice. Nat Protoc 8, 2531-2537.
Leung, L., Andrews-Zwilling, Y., Yoon, S.Y., Jain, S., Ring, K., Dai, J., Wang, M.M., Tong,
L., Walker, D., and Huang, Y. (2012). Apolipoprotein E4 causes age- and sex-dependent
impairments of hilar GABAergic interneurons and learning and memory deficits in mice.
PLoS One 7, e53569.
Lewis, S.E. (2017). The vision and challenges of the gene ontology. In The Gene Ontology
Handbook (Springer), pp. 291-302.
Li, F., Wang, Y., Yu, L., Cao, S., Wang, K., Yuan, J., Wang, C., Wang, K., Cui, M., and
Fu, Z.F. (2015). Viral infection of the Central Nervous System and neuroinflammation
precede Blood Brain Barrier disruption during Japanese Encephalitis. Journal of virology,
JVI. 00143-00115.
Li, G., Bien-Ly, N., Andrews-Zwilling, Y., Xu, Q., Bernardo, A., Ring, K., Halabisky, B.,
Deng, C., Mahley, R.W., and Huang, Y. (2009). GABAergic interneuron dysfunction
impairs hippocampal neurogenesis in adult apolipoprotein E4 knockin mice. Cell Stem Cell
5, 634-645.
Lindahl, P., Johansson, B.R., Levéen, P., and Betsholtz, C. (1997). Pericyte loss and
microaneurysm formation in PDGF-B-deficient mice. Science 277, 242-245.
138
Lisse, I.M., Aaby, P., Whittle, H., Jensen, H., Engelmann, M., and Christensen, L.B.
(1997). T-lymphocyte subsets in West African children: impact of age, sex, and season.
The Journal of Pediatrics 130, 77-85.
Liu, C.C., Kanekiyo, T., Xu, H., and Bu, G. (2013). Apolipoprotein e and Alzheimer
disease: Risk, mechanisms and therapy. Nature Reviews Neurology 9, 106-118.
Liu, S., Liu, Y., Hao, W., Wolf, L., Kiliaan, A.J., Penke, B., Rübe, C.E., Walter, J., Heneka,
M.T., and Hartmann, T. (2012). TLR2 is a primary receptor for Alzheimer’s amyloid β
peptide to trigger neuroinflammatory activation. The Journal of Immunology 188, 1098-
1107.
Louveau, A., Smirnov, I., Keyes, T.J., Eccles, J.D., Rouhani, S.J., Peske, J.D., Derecki,
N.C., Castle, D., Mandell, J.W., and Lee, K.S. (2015). Structural and functional features
of central nervous system lymphatic vessels. Nature 523, 337.
Loy, C.T., Schofield, P.R., Turner, A.M., and Kwok, J.B. (2014). Genetics of dementia.
The Lancet 383, 828-840.
Lu, K.H., Hopper, B.R., Vargo, T.M., and Yen, S.S. (1979). Chronological changes in sex
steroid, gonadotropin and prolactin secretions in aging female rats displaying different
reproductive states. Biol Reprod 21, 193-203.
Lustbader, J.W., Cirilli, M., Lin, C., Xu, H.W., Takuma, K., Wang, N., Caspersen, C., Chen,
X., Pollak, S., and Chaney, M. (2004). ABAD directly links Aß to mitochondrial toxicity in
Alzheimer's Disease. Science 304, 448-452.
Ma, Q., Huang, B., Khatibi, N., Rolland, W., Suzuki, H., Zhang, J.H., and Tang, J. (2011).
PDGFR-α Inhibition Preserves Blood-Brain Barrier after Intracerebral Hemorrhage.
Annals of neurology 70, 920-931.
Ma, S., Ma, L., Yang, D., Luo, Z., Hao, X., Liu, D., and Zhu, Z. (2010). Uncoupling protein
2 ablation exacerbates high-salt intake-induced vascular dysfunction. American journal of
hypertension 23, 822-828.
Ma, S., Wang, Q., Zhang, Y., Yang, D., Li, D., Tang, B., and Yang, Y. (2013). Transgenic
overexpression of uncoupling protein 2 attenuates salt-induced vascular dysfunction by
inhibition of oxidative stress. American journal of hypertension 27, 345-354.
Ma, S., Zhang, Y., Wang, Q., Yang, D., Tang, B., and Yang, Y. (2014). Ablation of
uncoupling protein 2 exacerbates salt-induced cardiovascular and renal remodeling
associated with enhanced oxidative stress. International journal of cardiology 175, 206-
210.
Maddaluno, L., Rudini, N., Cuttano, R., Bravi, L., Giampietro, C., Corada, M., Ferrarini, L.,
Orsenigo, F., Papa, E., and Boulday, G. (2013). EndMT contributes to the onset and
progression of cerebral cavernous malformations. Nature 498, 492.
Maftei, M., Thurm, F., Schnack, C., Tumani, H., Otto, M., Elbert, T., Kolassa, I.-T.,
Przybylski, M., Manea, M., and von Arnim, C.A. (2013). Increased levels of antigen-bound
β-amyloid autoantibodies in serum and cerebrospinal fluid of Alzheimer’s disease patients.
PLoS One 8, e68996.
Maggioli, E., McArthur, S., Mauro, C., Kieswich, J., Kusters, D., Reutelingsperger, C.,
Yaqoob, M., and Solito, E. (2016). Estrogen protects the blood–brain barrier from
inflammation-induced disruption and increased lymphocyte trafficking. Brain, behavior,
and immunity 51, 212-222.
Majno, G., Palade, G., and Schoefl, G.I. (1961). Studies on inflammation: II. The site of
action of histamine and serotonin along the vascular tree: A topographic study. The
Journal of Cell Biology 11, 607-626.
139
Maki, P.M., Drogos, L.L., Rubin, L.H., Banuvar, S., Shulman, L.P., and Geller, S.E. (2008).
Objective hot flashes are negatively related to verbal memory performance in midlife
women. Menopause (New York, NY) 15, 848.
Maki, P.M., and Henderson, V.W. (2016). Cognition and the menopause transition.
Menopause 23, 803-805.
Malutan, A.M., Dan, M., Nicolae, C., and Carmen, M. (2014). Proinflammatory and anti-
inflammatory cytokine changes related to menopause. Przegla ̜ d Menopauzalny =
Menopause Review 13, 162-168.
Masdeu, J.C., Kreisl, W.C., and Berman, K.F. (2012). The neurobiology of Alzheimer
disease defined by neuroimaging. Current opinion in neurology 25, 410-420.
Mattiasson, G., Shamloo, M., Gido, G., Mathi, K., Tomasevic, G., Yi, S., Warden, C.H.,
Castilho, R.F., Melcher, T., and Gonzalez-Zulueta, M. (2003). Uncoupling protein-2
prevents neuronal death and diminishes brain dysfunction after stroke and brain trauma.
Nature medicine 9, 1062.
Mayer, L.P., Devine, P.J., Dyer, C.A., and Hoyer, P.B. (2004). The follicle-deplete mouse
ovary produces androgen. Biol Reprod 71, 130-138.
Mayer, L.P., Dyer, C.A., Eastgard, R.L., Hoyer, P.B., and Banka, C.L. (2005).
Atherosclerotic lesion development in a novel ovary-intact mouse model of
perimenopause. Arterioscler Thromb Vasc Biol 25, 1910-1916.
Mayer, L.P., Pearsall, N.A., Christian, P.J., Devine, P.J., Payne, C.M., McCuskey, M.K.,
Marion, S.L., Sipes, I.G., and Hoyer, P.B. (2002). Long-term effects of ovarian follicular
depletion in rats by 4-vinylcyclohexene diepoxide. Reprod Toxicol 16, 775-781.
Mazariegos, M.R., Tice, L.W., and Hand, A.R. (1984). Alteration of tight junctional
permeability in the rat parotid gland after isoproterenol stimulation. The Journal of cell
biology 98, 1865-1877.
McCombe, P.A., Greer, J.M., and Mackay, I.R. (2009). Sexual dimorphism in autoimmune
disease. Current molecular medicine 9, 1058-1079.
McKinlay, S.M., Brambilla, D.J., and Posner, J.G. (1992). The normal menopause
transition. Maturitas 14, 103-115.
Meyer, V. (2003). Medicalized menopause, US style. Health care for women international
24, 822-830.
Mi, H., Muruganujan, A., Casagrande, J.T., and Thomas, P.D. (2013). Large-scale gene
function analysis with the PANTHER classification system. Nature protocols 8, 1551.
Michaud, M., Balardy, L., Moulis, G., Gaudin, C., Peyrot, C., Vellas, B., Cesari, M., and
Nourhashemi, F. (2013). Proinflammatory cytokines, aging, and age-related diseases.
Journal of the American Medical Directors Association 14, 877-882.
Minami, T., Sakita, Y., Ichida, S., and Dohi, Y. (2002). Gender difference regarding
selenium penetration into the mouse brain. Biological trace element research 89, 85-93.
Moon, S.-J., Park, J.-S., Heo, Y.-J., Kang, C.-M., Kim, E.-K., Lim, M.-A., Ryu, J.-G., Park,
S.J., Park, K.S., and Sung, Y.-C. (2013). In vivo action of IL-27: reciprocal regulation of
Th17 and Treg cells in collagen-induced arthritis. Experimental & molecular medicine 45,
e46.
Morin, R., Bainbridge, M., Fejes, A., Hirst, M., Krzywinski, M., Pugh, T., McDonald, H.,
Varhol, R., Jones, S., and Marra, M. (2008). Profiling the HeLa S3 transcriptome using
randomly primed cDNA and massively parallel short-read sequencing. Biotechniques 45,
81-94.
Morrison, J.H., and Baxter, M.G. (2012). The ageing cortical synapse: hallmarks and
implications for cognitive decline. Nat Rev Neurosci 13, 240-250.
140
Mortazavi, A., Williams, B.A., McCue, K., Schaeffer, L., and Wold, B. (2008). Mapping and
quantifying mammalian transcriptomes by RNA-Seq. Nature methods 5, 621.
Mosconi, L., Berti, V., Guyara-Quinn, C., McHugh, P., Petrongolo, G., Osorio, R.S.,
Connaughty, C., Pupi, A., Vallabhajosula, S., Isaacson, R.S., et al. (2017).
Perimenopause and emergence of an Alzheimer’s bioenergetic phenotype in brain and
periphery. PLOS ONE 12, e0185926.
Moukdar, F., Robidoux, J., Lyght, O., Pi, J., Daniel, K.W., and Collins, S. (2009). Reduced
antioxidant capacity and diet-induced atherosclerosis in uncoupling protein-2-deficient
mice. Journal of lipid research 50, 59-70.
Naderi, V., Khaksari, M., Abbasi, R., and Maghool, F. (2015). Estrogen provides
neuroprotection against brain edema and blood brain barrier disruption through both
estrogen receptors α and β following traumatic brain injury. Iranian Journal of Basic
Medical Sciences 18, 138-144.
Nagalakshmi, U., Wang, Z., Waern, K., Shou, C., Raha, D., Gerstein, M., and Snyder, M.
(2008). The transcriptional landscape of the yeast genome defined by RNA sequencing.
Science 320, 1344-1349.
Nakao, K., Yasoda, A., Ebihara, K., Hosoda, K., and Mukoyama, M. (2009). Translational
research of novel hormones: lessons from animal models and rare human diseases for
common human diseases. J Mol Med (Berl) 87, 1029-1039.
National Toxicology, P. (1989). Toxicology and Carcinogenesis Studies of 4-Vinyl-1-
cyclohexene Diepoxide (CAS No. 106-87-6) in F344/N Rats and B6C3F1 Mice (Dermal
Studies). Natl Toxicol Program Tech Rep Ser 362, 1-249.
Nelson, H.D. (2008). Menopause. Lancet 371, 760-770.
Neufert, C., Becker, C., Wirtz, S., Fantini, M.C., Weigmann, B., Galle, P.R., and Neurath,
M.F. (2007). IL ‐27 controls the development of inducible regulatory T cells and Th17 cells
via differential effects on STAT1. European journal of immunology 37, 1809-1816.
Newcombe, E.A., Camats-Perna, J., Silva, M.L., Valmas, N., Huat, T.J., and Medeiros, R.
(2018). Inflammation: the link between comorbidities, genetics, and Alzheimer’s disease.
Journal of Neuroinflammation 15, 276.
Newhart, M.R. (2013). Menopause matters: The implications of menopause research for
studies of midlife health. Health Sociology Review 22, 365-376.
Nishino, H., Nakajima, K., Kumazaki, M., Fukuda, A., Muramatsu, K., Deshpande, S.B.,
Inubushi, T., Morikawa, S., Borlongan, C.V., and Sanberg, P.R. (1998). Estrogen protects
against while testosterone exacerbates vulnerability of the lateral striatal artery to chemical
hypoxia by 3-nitropropionic acid. Neuroscience research 30, 303-312.
Nishioku, T., Dohgu, S., Takata, F., Eto, T., Ishikawa, N., Kodama, K.B., Nakagawa, S.,
Yamauchi, A., and Kataoka, Y. (2009). Detachment of brain pericytes from the basal
lamina is involved in disruption of the blood–brain barrier caused by lipopolysaccharide-
induced sepsis in mice. Cellular and molecular neurobiology 29, 309-316.
Noorbakhsh, F., Baker, G.B., and Power, C. (2014). Allopregnanolone and
neuroinflammation: a focus on multiple sclerosis. Frontiers in cellular neuroscience 8, 134.
Normoyle, K.P., Kim, M., Farahvar, A., Llano, D., Jackson, K., and Wang, H. (2015). The
emerging neuroprotective role of mitochondrial uncoupling protein-2 in traumatic brain
injury. Translational neuroscience 6.
Nwabuisi-Heath, E., Rebeck, G.W., LaDu, M.J., and Yu, C. (2013). ApoE4 delays dendritic
spine formation during neuron development and accelerates loss of mature spines in vitro.
ASN neuro 6, AN20130043.
141
Okoniewski, M.J., and Miller, C.J. (2006). Hybridization interactions between probesets in
short oligo microarrays lead to spurious correlations. BMC Bioinformatics 7, 276.
Olson, M.E., and Bruce, J. (1986). Ovariectomy, ovariohysterectomy and orchidectomy in
rodents and rabbits. Can Vet J 27, 523-527.
Oztaş, B., Kocak, H., Oner, P., and Küçük, M. (2000). Sex-dependent changes in blood-
brain barrier permeability and brain NA (+), K (+) ATPase activity in rats following acute
water intoxication. Journal of neuroscience research 62, 750-753.
Pakulski, C., Drobnik, L., and Millo, B. (2000). Age and sex as factors modifying the
function of the blood-cerebrospinal fluid barrier. Medical Science Monitor 6, 314-318.
Pan, X., Nasaruddin, M.B., Elliott, C.T., McGuinness, B., Passmore, A.P., Kehoe, P.G.,
Hölscher, C., McClean, P.L., Graham, S.F., and Green, B.D. (2016). Alzheimer's disease–
like pathology has transient effects on the brain and blood metabolome. Neurobiology of
Aging 38, 151-163.
Paris, J.J., Walf, A.A., and Frye, C.A. (2011). II. Cognitive performance of middle-aged
female rats is influenced by capacity to metabolize progesterone in the prefrontal cortex
and hippocampus. Brain research 1379, 149-163.
Peppercorn, M.A. (2014). Definition, Epidemiology, and Risk Factors in Inflammatory
Bowel Disease (UpToDate [Internet].).
Pfefferbaum, A., Rohlfing, T., Rosenbloom, M.J., Chu, W., Colrain, I.M., and Sullivan, E.V.
(2013). Variation in longitudinal trajectories of regional brain volumes of healthy men and
women (ages 10 to 85years) measured with atlas-based parcellation of MRI. NeuroImage
65, 176-193.
Pisitkun, P., Deane, J.A., Difilippantonio, M.J., Tarasenko, T., Satterthwaite, A.B., and
Bolland, S. (2006). Autoreactive B cell responses to RNA-related antigens due to TLR7
gene duplication. Science 312, 1669-1672.
Piterkin, P., Cole, E., Cossette, M.-P., Gaskin, S., and Mumby, D.G. (2008). A limited role
for the hippocampus in the modulation of novel-object preference by contextual cues.
Learning & Memory 15, 785-791.
Prelog, M. (2006). Aging of the immune system: a risk factor for autoimmunity?
Autoimmunity reviews 5, 136-139.
Qureshi, I.A., and Mehler, M.F. (2012). Emerging roles of non-coding RNAs in brain
evolution, development, plasticity and disease. Nature Reviews Neuroscience 13, 528.
Ramanan, V.K., Shen, L., Moore, J.H., and Saykin, A.J. (2012). Pathway analysis of
genomic data: concepts, methods, and prospects for future development. TRENDS in
Genetics 28, 323-332.
Raub, T.J., Kuentzel, S.L., and Sawada, G.A. (1992). Permeability of bovine brain
microvessel endothelial cells in vitro: barrier tightening by a factor released from
astroglioma cells. Experimental cell research 199, 330-340.
Reddy, P.H., and Beal, M.F. (2008). Amyloid beta, mitochondrial dysfunction and synaptic
damage: implications for cognitive decline in aging and Alzheimer's disease. Trends in
molecular medicine 14, 45-53.
Reid, A., Teasdale, G., and McCulloch, J. (1983). Hormonal influence on water
permeability across the blood-brain barrier. Clinical and experimental neurology 19, 50-
53.
Rettberg, J.R., Dang, H., Hodis, H.N., Henderson, V.W., St. John, J.A., Mack, W.J., and
Brinton, R.D. (2016). Identifying postmenopausal women at risk for cognitive decline within
a healthy cohort using a panel of clinical metabolic indicators: Potential for detecting an
at-Alzheimer’s risk metabolic phenotype. Neurobiology of aging 40, 155-163.
142
Rettew, J.A., Huet-Hudson, Y.M., and Marriott, I. (2008). Testosterone reduces
macrophage expression in the mouse of toll-like receptor 4, a trigger for inflammation and
innate immunity. Biology of reproduction 78, 432-437.
Reyahi, A., Nik, A.M., Ghiami, M., Gritli-Linde, A., Pontén, F., Johansson, B.R., and
Carlsson, P. (2015). Foxf2 is required for brain pericyte differentiation and development
and maintenance of the blood-brain barrier. Developmental cell 34, 19-32.
Roberts, C.W., Walker, W., and Alexander, J. (2001). Sex-Associated Hormones and
Immunity to Protozoan Parasites. Clinical Microbiology Reviews 14, 476-488.
Rochfort, K.D., and Cummins, P.M. (2015). The blood–brain barrier endothelium: a target
for pro-inflammatory cytokines. Biochemical Society Transactions 43, 702-706.
Royce, T.E., Rozowsky, J.S., and Gerstein, M.B. (2007). Toward a universal microarray:
prediction of gene expression through nearest-neighbor probe sequence identification.
Nucleic Acids Res 35, e99.
Rubattu, S., Stanzione, R., Bianchi, F., Cotugno, M., Forte, M., Della Ragione, F.,
Fioriniello, S., D'Esposito, M., Marchitti, S., Madonna, M., et al. (2017). Reduced brain
UCP2 expression mediated by microRNA-503 contributes to increased stroke
susceptibility in the high-salt fed stroke-prone spontaneously hypertensive rat. Cell Death
&Amp; Disease 8, e2891.
Russu, M.C., and Antonescu, A.C. (2018). Neuroprotection in Perimenopausal Women.
Ryu, J.K., and McLarnon, J.G. (2009). A leaky blood–brain barrier, fibrinogen infiltration
and microglial reactivity in inflamed Alzheimer’s disease brain. Journal of cellular and
molecular medicine 13, 2911-2925.
Salloway, S., Gur, T., Berzin, T., Zipser, B., Correia, S., Hovanesian, V., Fallon, J., Kuo-
Leblanc, V., Glass, D., and Hulette, C. (2002). Effect of APOE genotype on microvascular
basement membrane in Alzheimer's disease. Journal of the neurological sciences 203,
183-187.
Salta, E., and De Strooper, B. (2012). Non-coding RNAs with essential roles in
neurodegenerative disorders. The Lancet Neurology 11, 189-200.
Sampedro, F., Vilaplana, E., de Leon, M.J., Alcolea, D., Pegueroles, J., Montal, V.,
Carmona-Iragui, M., Sala, I., Sánchez-Saudinos, M.B., Antón-Aguirre, S., et al. (2015).
APOE-by-sex interactions on brain structure and metabolism in healthy elderly controls.
Oncotarget 6, 26663-26674.
Sankaran-Walters, S., Macal, M., Grishina, I., Nagy, L., Goulart, L., Coolidge, K., Li, J.,
Fenton, A., Williams, T., Miller, M.K., et al. (2013). Sex differences matter in the gut: effect
on mucosal immune activation and inflammation. Biol Sex Differ 4, 10.
Sans, N., Petralia, R.S., Wang, Y.X., Blahos Ii, J., Hell, J.W., and Wenthold, R.J. (2000).
A developmental change in NMDA receptor-associated proteins at hippocampal
synapses. Journal of Neuroscience 20, 1260-1271.
Sárvári, M., Hrabovszky, E., Kalló, I., Solymosi, N., Likó, I., Berchtold, N., Cotman, C., and
Liposits, Z. (2012). Menopause leads to elevated expression of macrophage-associated
genes in the aging frontal cortex: rat and human studies identify strikingly similar changes.
Journal of neuroinflammation 9, 264.
Sárvári, M., Kalló, I., Hrabovszky, E., Solymosi, N., and Liposits, Z. (2014). Ovariectomy
and subsequent treatment with estrogen receptor agonists tune the innate immune system
of the hippocampus in middle-aged female rats. PloS one 9, e88540.
Saunders, N.R., Dziegielewska, K.M., Møllgård, K., and Habgood, M.D. (2015). Markers
for blood-brain barrier integrity: how appropriate is Evans blue in the twenty-first century
and what are the alternatives? Frontiers in neuroscience 9, 385.
143
Scheibel, A.B., and Duong, T. (1988). On the possible relationship of cortical
microvascular pathology to blood brain barrier changes in Alzheimer's disease.
Neurobiology of aging 9, 41-42.
Setoyama, D., Fujimura, Y., and Miura, D. (2013). Metabolomics reveals that carnitine
palmitoyltransferase ‐1 is a novel target for oxidative inactivation in human cells. Genes to
Cells 18, 1107-1119.
Sherman, B.M., WEST, J.H., and KORENMAN, S.G. (1976). The menopausal transition:
analysis of LH, FSH, estradiol, and progesterone concentrations during menstrual cycles
of older women. The Journal of Clinical Endocrinology & Metabolism 42, 629-636.
Sherwin, B.B. (1994). Estrogenic Effects on Memory in Women. Annals of the New York
Academy of Sciences 743, 213-230.
Sherwin, B.B. (2003). Estrogen and Cognitive Functioning in Women. Endocrine Reviews
24, 133-151.
Shifren, J.L., Gass, M.L., and Group, N.R.f.C.C.o.M.W.W. (2014). The North American
Menopause Society recommendations for clinical care of midlife women. Menopause 21,
1038-1062.
Shineman, D.W., Basi, G.S., Bizon, J.L., Colton, C.A., Greenberg, B.D., Hollister, B.A.,
Lincecum, J., Leblanc, G.G., Lee, L.B., Luo, F., et al. (2011). Accelerating drug discovery
for Alzheimer's disease: best practices for preclinical animal studies. Alzheimers Res Ther
3, 28.
Shmelkov, E., Tang, Z., Aifantis, I., and Statnikov, A. (2011). Assessing quality and
completeness of human transcriptional regulatory pathways on a genome-wide scale.
Biology direct 6, 15.
Singh, C., Liu, L., Wang, J.M., Irwin, R.W., Yao, J., Chen, S., Henry, S., Thompson, R.F.,
and Brinton, R.D. (2012). Allopregnanolone restores hippocampal-dependent learning
and memory and neural progenitor survival in aging 3xTgAD and nonTg mice. Neurobiol
Aging 33, 1493-1506.
Skup, M., Zhu, H., Wang, Y., Giovanello, K.S., Lin, J.A., Shen, D., Shi, F., Gao, W., Lin,
W., Fan, Y., et al. (2011). Sex differences in grey matter atrophy patterns among AD and
aMCI patients: Results from ADNI. NeuroImage 56, 890-906.
Snyder, H.M., Asthana, S., Bain, L., Brinton, R., Craft, S., Dubal, D.B., Espeland, M.A.,
Gatz, M., Mielke, M.M., and Raber, J. (2016). Sex biology contributions to vulnerability to
Alzheimer's disease: A think tank convened by the Women's Alzheimer's Research
Initiative. Alzheimer's & Dementia 12, 1186-1196.
Soares, C.N., and Maki, P.M. (2010). Introducing a special section on menopause,
cognition, and mental health. Menopause 17, 811.
Soules, M.R., Sherman, S., Parrott, E., Rebar, R., Santoro, N., Utian, W., and Woods, N.
(2001). Stages of Reproductive Aging Workshop (STRAW). J Womens Health Gend
Based Med 10, 843-848.
Sperling, R.A., Karlawish, J., and Johnson, K.A. (2013). Preclinical Alzheimer disease—
the challenges ahead. Nature Reviews Neurology 9, 54.
Spinney, L. (2014). The forgetting gene. Nature 510, 26.
Springer, L.N., Flaws, J.A., Sipes, I.G., and Hoyer, P.B. (1996a). Follicular mechanisms
associated with 4-vinylcyclohexene diepoxide-induced ovotoxicity in rats. Reprod Toxicol
10, 137-143.
Springer, L.N., McAsey, M.E., Flaws, J.A., Tilly, J.L., Sipes, I.G., and Hoyer, P.B. (1996b).
Involvement of apoptosis in 4-vinylcyclohexene diepoxide-induced ovotoxicity in rats.
Toxicol Appl Pharmacol 139, 394-401.
144
Springer, L.N., Tilly, J.L., Sipes, I.G., and Hoyer, P.B. (1996c). Enhanced expression of
bax in small preantral follicles during 4-vinylcyclohexene diepoxide-induced ovotoxicity in
the rat. Toxicol Appl Pharmacol 139, 402-410.
Straub, R.H., and Schradin, C. (2016). Chronic inflammatory systemic diseases – an
evolutionary trade-off between acutely beneficial but chronically harmful programs.
Evolution, Medicine, and Public Health, eow001.
Stumhofer, J.S., Silver, J.S., Laurence, A., Porrett, P.M., Harris, T.H., Turka, L.A., Ernst,
M., Saris, C.J., O'Shea, J.J., and Hunter, C.A. (2007). Interleukins 27 and 6 induce STAT3-
mediated T cell production of interleukin 10. Nature immunology 8, 1363.
Taglialatela, G., Hogan, D., Zhang, W.-R., and Dineley, K.T. (2009). Intermediate-and
long-term recognition memory deficits in Tg2576 mice are reversed with acute calcineurin
inhibition. Behavioural brain research 200, 95-99.
Teixeira, D., Longo-Maugeri, I.M., Santos, J.L.F., Duarte, Y.A.O., Lebrão, M.L., and
Bueno, V. (2011). Evaluation of lymphocyte levels in a random sample of 218 elderly
individuals from São Paulo city. Revista Brasileira de Hematologia e Hemoterapia 33, 367-
371.
Terao, A., Apte-Deshpande, A., Dousman, L., Morairty, S., Eynon, B.P., Kilduff, T.S., and
Freund, Y.R. (2002). Immune response gene expression increases in the aging murine
hippocampus. Journal of neuroimmunology 132, 99-112.
Thomas, S., and Bonchev, D. (2010). A survey of current software for network analysis in
molecular biology. Human genomics 4, 353.
Tian, X.Y., Wong, W.T., Xu, A., Lu, Y., Zhang, Y., Wang, L., Cheang, W.S., Wang, Y.,
Yao, X., and Huang, Y. (2012). Uncoupling protein-2 protects endothelial function in diet-
induced obese mice. Circulation research 110, 1211-1216.
Timiras, P.S., Quay, W.B., and Vernadakis, A. (1995). Hormones and aging (CRC Press).
Timmons, J.A., Szkop, K.J., and Gallagher, I.J. (2015). Multiple sources of bias confound
functional enrichment analysis of global-omics data. Genome biology 16, 186.
Tomás ‐Camardiel, M., Venero, J., Herrera, A., De Pablos, R., Pintor ‐Toro, J.A., Machado,
A., and Cano, J. (2005). Blood–brain barrier disruption highly induces aquaporin ‐4 mRNA
and protein in perivascular and parenchymal astrocytes: protective effect by estradiol
treatment in ovariectomized animals. Journal of neuroscience research 80, 235-246.
Tong, L.M., Yoon, S.Y., Andrews-Zwilling, Y., Yang, A., Lin, V., Lei, H., and Huang, Y.
(2016). Enhancing GABA Signaling during Middle Adulthood Prevents Age-Dependent
GABAergic Interneuron Decline and Learning and Memory Deficits in ApoE4 Mice. J
Neurosci 36, 2316-2322.
Torcia, M.G., Nencioni, L., Clemente, A.M., Civitelli, L., Celestino, I., Limongi, D., Fadigati,
G., Perissi, E., Cozzolino, F., and Garaci, E. (2012). Sex differences in the response to
viral infections: TLR8 and TLR9 ligand stimulation induce higher IL10 production in males.
PloS one 7, e39853.
Trabzuni, D., Ramasamy, A., Imran, S., Walker, R., Smith, C., Weale, M.E., Hardy, J.,
Ryten, M., and Consortium, N.A.B.E. (2013). Widespread sex differences in gene
expression and splicing in the adult human brain. Nature communications 4, 2771.
Trapnell, C., Hendrickson, D.G., Sauvageau, M., Goff, L., Rinn, J.L., and Pachter, L.
(2012). Differential analysis of gene regulation at transcript resolution with RNA-seq.
Nature Biotechnology 31, 46.
Trapnell, C., Hendrickson, D.G., Sauvageau, M., Goff, L., Rinn, J.L., and Pachter, L.
(2013). Differential analysis of gene regulation at transcript resolution with RNA-seq.
Nature biotechnology 31, 46.
145
Trapnell, C., Williams, B.A., Pertea, G., Mortazavi, A., Kwan, G., van Baren, M.J.,
Salzberg, S.L., Wold, B.J., and Pachter, L. (2010). Transcript assembly and quantification
by RNA-Seq reveals unannotated transcripts and isoform switching during cell
differentiation. Nat Biotechnol 28, 511-515.
Ujiie, M., Dickstein, D.L., Carlow, D.A., and Jefferies, W.A. (2003). Blood–brain barrier
permeability precedes senile plaque formation in an Alzheimer disease model.
Microcirculation 10, 463-470.
Uppal, S.S., Verma, S., and Dhot, P.S. (2003). Normal values of CD4 and CD8 lymphocyte
subsets in healthy indian adults and the effects of sex, age, ethnicity, and smoking.
Cytometry B Clin Cytom 52, 32-36.
Utian, W.H. (2004). Menopause-related definitions. Paper presented at: International
Congress Series (Elsevier).
Van Geest, R.J., Klaassen, I., Vogels, I.M., Van Noorden, C.J., and Schlingemann, R.O.
(2010). Differential TGF-β signaling in retinal vascular cells: a role in diabetic retinopathy?
Investigative ophthalmology & visual science 51, 1857-1865.
Van Kempen, T.A., Milner, T.A., and Waters, E.M. (2011). Accelerated ovarian failure: a
novel, chemically induced animal model of menopause. Brain Res 1379, 176-187.
Vera, J.C., Wheat, C.W., Fescemyer, H.W., Frilander, M.J., Crawford, D.L., Hanski, I., and
Marden, J.H. (2008). Rapid transcriptome characterization for a nonmodel organism using
454 pyrosequencing. Mol Ecol 17, 1636-1647.
von Tell, D., Armulik, A., and Betsholtz, C. (2006). Pericytes and vascular stability.
Experimental cell research 312, 623-629.
Walker, D.M., and Gore, A.C. (2011). Transgenerational neuroendocrine disruption of
reproduction. Nat Rev Endocrinol 7, 197-207.
Walker, M.L., and Herndon, J.G. (2008). Menopause in nonhuman primates? Biol Reprod
79, 398-406.
Wang, J.M., and Brinton, R.D. (2008). Allopregnanolone-induced rise in intracellular
calcium in embryonic hippocampal neurons parallels their proliferative potential. BMC
Neurosci 9 Suppl 2, S11.
Wang, J.M., Johnston, P.B., Ball, B.G., and Brinton, R.D. (2005). The neurosteroid
allopregnanolone promotes proliferation of rodent and human neural progenitor cells and
regulates cell-cycle gene and protein expression. J Neurosci 25, 4706-4718.
Wang, J.M., Singh, C., Liu, L., Irwin, R.W., Chen, S., Chung, E.J., Thompson, R.F., and
Brinton, R.D. (2010). Allopregnanolone reverses neurogenic and cognitive deficits in
mouse model of Alzheimer's disease. Proc Natl Acad Sci U S A 107, 6498-6503.
Wang, Z., Gerstein, M., and Snyder, M. (2009). RNA-Seq: a revolutionary tool for
transcriptomics. Nature reviews Genetics 10, 57-63.
Weber, M.T., Maki, P.M., and McDermott, M.P. (2014). Cognition and mood in
perimenopause: A systematic review and meta-analysis. The Journal of Steroid
Biochemistry and Molecular Biology 142, 90-98.
Wise, P.M., Smith, M.J., Dubal, D.B., Wilson, M.E., Rau, S.W., Cashion, A.B., Bottner, M.,
and Rosewell, K.L. (2002). Neuroendocrine modulation and repercussions of female
reproductive aging. Recent Prog Horm Res 57, 235-256.
Wolburg, H., and Lippoldt, A. (2002). Tight junctions of the blood–brain barrier:
development, composition and regulation. Vascular pharmacology 38, 323-337.
Wolman, M., Klatzo, I., Chui, E., Wilmes, F., Nishimoto, K., Fujiwara, K., and Spatz, M.
(1981). Evaluation of the dye-protein tracers in pathophysiology of the blood-brain barrier.
Acta neuropathologica 54, 55-61.
146
Woods, N.F., Mitchell, E.S., and Adams, C. (2000). Memory Functioning Among Midlife
Women: Observations From the Seattle Midlife Womenʼs Health Study. Menopause 7,
257-265.
Wyss-Coray, T. (2006). Inflammation in Alzheimer disease: driving force, bystander or
beneficial response? Nature medicine 12, 1005.
Yao, J., Irwin, R.W., Zhao, L., Nilsen, J., Hamilton, R.T., and Brinton, R.D. (2009).
Mitochondrial bioenergetic deficit precedes Alzheimer's pathology in female mouse model
of Alzheimer's disease. Proceedings of the National Academy of Sciences 106, 14670-
14675.
Ye, L., Martin, T.A., Parr, C., Harrison, G.M., Mansel, R.E., and Jiang, W.G. (2003).
Biphasic effects of 17 ‐β ‐estradiol on expression of occludin and transendothelial
resistance and paracellular permeability in human vascular endothelial cells. Journal of
cellular physiology 196, 362-369.
Yen, L.F., Wei, V.C., Kuo, E.Y., and Lai, T.W. (2013). Distinct Patterns of Cerebral
Extravasation by Evans Blue and Sodium Fluorescein in Rats. PLOS ONE 8, e68595.
Yin, F., Yao, J., Sancheti, H., Feng, T., Melcangi, R.C., Morgan, T.E., Finch, C.E., Pike,
C.J., Mack, W.J., Cadenas, E., et al. (2015). The perimenopausal aging transition in the
female rat brain: decline in bioenergetic systems and synaptic plasticity. Neurobiology of
Aging 36, 2282-2295.
Yoshida, H., and Hunter, C.A. (2015). The Immunobiology of Interleukin-27. Annual review
of immunology 33, 417-443.
Young, M.D., Wakefield, M.J., Smyth, G.K., and Oshlack, A. (2010). Gene ontology
analysis for RNA-seq: accounting for selection bias. Genome biology 11, R14.
Zhang, M.A., Rego, D., Moshkova, M., Kebir, H., Chruscinski, A., Nguyen, H., Akkermann,
R., Stanczyk, F.Z., Prat, A., Steinman, L., et al. (2012). Peroxisome proliferator-activated
receptor (PPAR) and - regulate IFN and IL-17A production by human T cells in a sex-
specific way. Proceedings of the National Academy of Sciences 109, 9505-9510.
Zhao, S., and Zhang, B. (2015). A comprehensive evaluation of ensembl, RefSeq, and
UCSC annotations in the context of RNA-seq read mapping and gene quantification. BMC
Genomics 16, 97.
Zhao, Z., Nelson, A.R., Betsholtz, C., and Zlokovic, B.V. (2015). Establishment and
Dysfunction of the Blood-Brain Barrier. Cell 163, 1064-1078.
Zipser, B., Johanson, C., Gonzalez, L., Berzin, T., Tavares, R., Hulette, C., Vitek, M.,
Hovanesian, V., and Stopa, E. (2007). Microvascular injury and blood–brain barrier
leakage in Alzheimer's disease. Neurobiology of aging 28, 977-986.
Ziylan, Y., Lefauconnier, J., Bernard, G., and Bourre, J. (1990). Blood-brain barrier
permeability: regional alterations after acute and chronic administration of ethinyl estradiol.
Neuroscience letters 118, 181-184.
Zlokovic, B.V. (2005). Neurovascular mechanisms of Alzheimer's neurodegeneration.
Trends in neurosciences 28, 202-208.
Zlokovic, B.V. (2008). The Blood-Brain Barrier in Health and Chronic Neurodegenerative
Disorders. Neuron 57, 178-201.
Zonneveld, H.I., Goos, J.D., Wattjes, M.P., Prins, N.D., Scheltens, P., van der Flier, W.M.,
Kuijer, J.P., Muller, M., and Barkhof, F. (2014). Prevalence of cortical superficial siderosis
in a memory clinic population. Neurology 82, 698-704.
Abstract (if available)
Abstract
Introduction. The onset of reproductive senescence occurs at the perimenopause, which is characterized by major physiological changes in the endocrinological and reproductive systems which can be associated with altered energy metabolism, cognition, bone-mineral density, cardiovascular function and immune system responses. Our rat model of the perimenopause to menopause to post menopause transitions captures both chronological and endocrinological conversions. In the current study, differential regulation of genes in the hippocampi of the endocrine characterized female rats by sequencing hippocampal total RNA was investigated, based on which it was hypothesized that perimenopausal transition increases blood brain permeability with grave implications for neurodegenerative diseases in women. However, based on rodent model, a novel neuroregenerative therapeutic, Allopregnanolone, may help ameliorate the increase in permeability and improve cognitive function. ❧ Methods. Paired end sequencing of hippocampal total RNA from 36 Sprague Dawley (SD) female rat hippocampi, belonging to six different groups was conducted. A list of differentially expressed genes (DEG) for each comparison was obtained using TopHat and Cufflinks employed in the Partek Flow environment (https://www.partek.com/). Differentially expressed genes were analyzed using Ingenuity Pathway Analysis and PANTHER to identify gene pathways altered during the perimenopause. Blood brain barrier (BBB) permeability was measured in different groups of female SD rats using sodium fluorescein. Both male and female mice were treated with Allopregnanolone intramuscularly. Behavior was assessed using Novel Object Recognition. Plasma and cortex from mice were queried for 185 metabolites using ultra-performance-LCMS. Mice hippocampi were used for RNA-Seq. ❧ Results. RNA-Seq detected a decrease in platelet-derived growth factor (PDGF) activity, associated with decreased brain estrogen, indicating potential for increased blood brain barrier permeability during the perimenopausal hippocampus. Analyses of significantly differentially expressed genes (DEG) using PANTHER (https://pantherdb.org/index.jsp) revealed enrichment of leukocytes, B and T lymphocyte genes. Curated literature based bioinformatic analysis indicated activation of pathways related to lymphocytic proliferation and differentiation during the transition from regular to irregular cycling, and inactivation of pathways responsible for T cell apoptosis. Interestingly, interleukin-27, a pan-T lymphocyte regulator was down-regulated in the hippocampal transcriptome during the perimenopausal transition, hinting at potential T lymphocyte dysregulation associated with the perimenopausal transition. Increased permeability of BBB during the perimenopausal transition was confirmed using sodium fluorescein. A separate experiment suggested that Allopregnanolone improved cognitive function in human ApoE ε4 allele containing Targeted Replacement mice, potentially upregulated PDGF signaling pathway in the hippocampal transcriptome which suggests protection of blood brain barrier integrity, increased lipid metabolism to generate acetyl-CoA to feed into the TCA cycle and promoted ATP generation in the mitochondria. Further, Allo treatment increased indicators of protein metabolism and potentiated mitochondria in human neural stem cells. ❧ Conclusions. These RNA-Seq and bioinformatic analyses suggested a functional disruption of the blood brain barrier during the perimenopausal transition state that was confirmed. Both chronological and endocrinological aging in female rats resulted in divergence into sub-populations based on blood brain barrier permeability corroborating evidence of inter-individual variation in neurodegenerative diseases. Allopregnanolone not only ameliorated cognitive function but based on analysis of hippocampal transcriptome suggested amelioration of blood brain barrier permeability, which must be functionally confirmed, and its mechanism elicited in future studies.
Linked assets
University of Southern California Dissertations and Theses
Conceptually similar
PDF
From risk mitochondrial and metabolic phenotype towards a precision medicine approach for Alzheimer's disease
PDF
Exploration and expansion of novel therapeutic strategies targeting brain metabolism in neurodegenerative diseases
PDF
Neuroinflammation and ApoE4 genotype in at-risk female aging: implications for Alzheimer's disease
PDF
An iPSC-based biomarker strategy to identify neuroregenerative responders to allopregnanolone
PDF
Epigenetic regulation of endocrine aging transitions of the perimenopausal and menopausal brain
PDF
Bioenergetic fuel shift of reproductive aging: implication for late-onset Alzheimer's disease
PDF
Estrogen regulation of bioenergetics and mitochondrial function: implications for Alzheimer's disease risk and therapeutics
PDF
RPE secretome for the treatment of retinal degeneration in the RCS rat
PDF
Estradiol regulation of cerebral metabolism: implications for neuroprotection and mitochondrial bioenergetics
PDF
Blood-brain barrier pathophysiology in cognitive impairment and injury
PDF
Allopregnanolone promotes cholesterol and amyloid-beta clearance mechanisms
PDF
Evaluation of P2X4 receptor modulation as a novel approach for treating Parkinson’s disease
PDF
Understanding the role of APP and DYRK1A in human brain pericytes
PDF
C-jun N-terminal Kinase (JNK) mediated inhibition of Pyruvate Dehydrogenase (PDH) activity and its effect on mitochondrial metabolism during brain aging
PDF
Mapping water exchange rate across the blood-brain barrier
PDF
Investigating brain aging and neurodegenerative diseases through omics data
PDF
Therapeutic potential of host defense peptides for the treatment of cystic fibrosis lung disease
PDF
The preclinical evaluation of moxidectin as a platform for drug development for alcohol use disorder
PDF
Maternal inflammation disrupts fetal blood-brain barrier formation via cyclooxygenase activation
Asset Metadata
Creator
Desai, Maunil Kandarp
(author)
Core Title
Perimenopausal transition increases blood brain permeability: implications for neurodegenerative diseases
School
School of Pharmacy
Degree
Doctor of Philosophy
Degree Program
Clinical and Experimental Therapeutics
Publication Date
05/06/2019
Defense Date
10/16/2018
Publisher
University of Southern California
(original),
University of Southern California. Libraries
(digital)
Tag
Alzheimer's disease,blood brain barrier,OAI-PMH Harvest,perimenopause,RNAseq
Format
application/pdf
(imt)
Language
English
Contributor
Electronically uploaded by the author
(provenance)
Advisor
Louie, Stan (
committee chair
), Cadenas, Enrique (
committee member
), Diaz Brinton, Roberta (
committee member
)
Creator Email
maunilkd@gmail.com,maunilkd@usc.edu
Permanent Link (DOI)
https://doi.org/10.25549/usctheses-c89-167449
Unique identifier
UC11660416
Identifier
etd-DesaiMauni-7408.pdf (filename),usctheses-c89-167449 (legacy record id)
Legacy Identifier
etd-DesaiMauni-7408.pdf
Dmrecord
167449
Document Type
Dissertation
Format
application/pdf (imt)
Rights
Desai, Maunil Kandarp
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
Tags
Alzheimer's disease
blood brain barrier
perimenopause
RNAseq