Thesis Template

Molecular and Synaptic Alterations in the Striatum of GluN1-Deficient Mice

Yuxiao Chen

A thesis submitted in conformity with the requirements for the degree of Doctor of Philosophy Department of Pharmacology and Toxicology University of Toronto

Yuxiao Chen

Doctor of Philosophy

Department of Pharmacology and Toxicology University of Toronto

2018 Abstract

N-methyl-D-aspartate receptors (NMDARs) are essential for proper neurodevelopment and cognitive function. NMDAR dysfunction contributes to many neurological disorders, with striatal changes linked to psychosis and cognitive impairment. GluN1 knockdown (GluN1KD) mice have low NMDAR expression, striatal synaptic deficits, and cognitive behavioural abnormalities. How these phenotypes functionally relate to one another is unclear. We studied molecular and synaptic changes in the GluN1KD striatum to understand the molecular underpinnings of GluN1KD behaviours. From a screen of Rho GTPase pathway proteins, which regulate synapses and cognition, Ras-related C3 botulinum toxin substrate 1 (Rac1) signaling proteins were specifically identified. GluN1KD mice at 6 and 12 weeks of age consistently had less Rac1 and its downstream effector Wiskott-Aldrich syndrome protein-family verprolin homologous protein 1 (WAVE-1). These deficits were developmentally aligned with medium spiny neuron (MSN) spine density deficits and previously published behavioural abnormalities.

To assess the biological and behavioural significance of these findings, we restored WAVE-1 levels by intercross-breeding GluN1KD mice and WAVE-1 overexpressing (WAVE-Tg) mice.

GluN1KD-WAVE-Tg F1 hybrids had improved WAVE-1 levels and spine density at the striatum, as well as better Y-maze and 8-arm radial maze performance. These phenotypic recoveries were not mirrored by WAVE-1 levels in the hippocampus nor by neuronal dendritic ii spines in cornu ammonis 1. From a broader RNAseq analysis of the GluN1KD striatal transcriptome, both neurodegeneration-related and cell-growth signaling pathways were distinguished as affected by NMDAR knockdown. These transcript changes were greatly sex- dependent, potentially mediated by aberrations of estrogen receptor α signaling. Our data suggest that different molecular changes in male and female GluN1KD mice contribute to similar behavioural abnormalities. Restoring WAVE-1 at the striatum improved some of these phenotypes, pointing to a nuanced relationship between its function, striatal synaptic plasticity, and cognition. Specifically, striatal WAVE-1 and MSN spine density may be important for goal- directed maze exploration in mice.

iii

Acknowledgments

I would like to thank all my friends and family who have supported me in various ways, professors who have guided me throughout my PhD program, and labmates who worked and commiserated with me side-by-side through all the highs and lows. For all my friends and family outside the lab, I have already (I hope, by the time this is submitted) said my thanks for all of your emotional and moral support and for putting up with me. That must not have been easy.

Of the professors that I must thank for their guidance over these years, my supervisor Dr. Amy J. Ramsey is first and foremost, of course. I am very thankful to you for the opportunity to conduct research in your lab. Thank you for all of your lessons and advice. You are a wonderful teacher and mentor and I attribute a lot of my development as a scientist and academic to your efforts. Similarly, I would like to thank Dr. Ali Salahpour for day-to-day support and for my overall amazing working experience in the Ramsey-Salahpour lab. I also want to express gratitude to all of my PhD supervisory committee members, past and present: Drs. José N. Nobrega, David S. Riddick, Zhengping Jia, and Jason Matthews. Thank you all for your constructive feedback throughout my PhD program. I know that your comments have helped me tremendously in progressing my research. While you all offered various forms of guidance, I particularly wanted to thank Dr. Nobrega for statistical help, Dr. Jia for in-depth insight into my research topics, and Drs. Riddick and Matthews for grounding my perspective outside of neuroscience research. Finally, I wanted to double-thank Dr. Riddick for being my early thesis reader. I swear I tried to get the best possible draft to you.

Finally, I want to thank the labmates who have worked with me in the trenches (and benches) all these years. In particular, I want to show appreciation for all the work done by Marija Milenkovic and Wendy Horsfall in support of my research. Thank you Marija for all of your help with behavioural experiments presented in this thesis, not to mention your previous Rho GTPase work that laid the foundations for my research. Thank you, Wendy, for all the animal husbandry work you do every day, as well as the various other bits you do to support and teach all the students. I also want to recognize all the previous Ramsey-Salahpour graduate students who have been mentors to me in their own right: Kasia Mielnik, Kristel Bermejo, Shababa T. Masoud, Laura M. Vecchio, Vincent Lam, and Pieter Beerepoot. You guys were the best colleagues I could have asked for.

Table of Contents

Acknowledgments...... iv

Table of Contents ...... v

List of Tables ...... ix

List of Figures ...... x

List of Appendices ...... xii

List of Abbreviations ...... xiii

Chapter 1 Introduction ...... 1

Introduction ...... 1

1.1 The N-methyl-D-aspartate Receptor ...... 1

1.1.1 N-methyl-D-aspartate Receptor Characterization and Organization ...... 1

1.1.2 NMDAR Activation and Subsequent Downstream Signaling ...... 6

1.1.3 Physiological Roles of NMDARs ...... 13

1.1.4 NMDARs in Health and Disease; Links to Neurological Disorders ...... 18

1.2 Mouse Model of NMDAR Deficiency by GluN1 Knockdown ...... 25

1.2.1 Background and Overview of GluN1 Knockdown Mice ...... 25

1.2.2 GluN1KD Behavioural Abnormalities and Their Pharmacological Treatments ...28

1.2.3 Cellular and Molecular Changes in GluN1KD Mice ...... 40

1.3 Study Overview ...... 49

1.3.1 GluN1KD Model Validity, Utility and Research Impact ...... 49

1.3.2 Rationale and Purpose of the Study ...... 51

1.3.3 Specific Research Objectives ...... 56

Chapter 2 Methods and Materials ...... 58

Methods and Materials ...... 58

2.1 Animals Used in the Study ...... 58

2.1.1 GluN1KD Mice ...... 58 v

2.1.2 WAVE-Tg Mice...... 58

2.1.3 Animal Handling and Ethics ...... 59

2.2 Dendritic Spine Analyses ...... 59

2.2.1 Dendritic Spine Density Analysis of WT and GluN1KD Mice at Three, Six, and Twelve Weeks of Age ...... 59

2.2.2 Dendritic Spine Density and Morphology Analyses of WT, WAVE-Tg, GluN1KD, and GluN1KD-WAVE Mice ...... 60

2.3 Western Blotting ...... 60

2.4 Fluorescent in situ hybridization via RNAscope ...... 61

2.5 Behavioural Tests of Cognition in Mice ...... 62

2.5.1 Y-Maze Spontaneous Alternation Test ...... 62

2.5.2 Eight-Arm Radial Maze Test ...... 63

2.5.3 Puzzle Box Test ...... 65

2.5.4 Open Field Locomotion and Stereotypy Test ...... 67

2.5.5 Social Approach Behaviour Test ...... 67

2.6 RNAseq Analyses ...... 68

2.6.1 Sample Preparation and Differential Gene Expression Analysis ...... 68

2.6.2 Gene List Analyses ...... 69

2.7 Statistical Analysis ...... 70

Chapter 3 Results ...... 72

Results ...... 72

3.1 GluN1KD Mice Show MSN Spine Density Deficits and Rac1 Signaling Protein Aberrations that are Age-Dependent ...... 72

3.1.1 MSN Dendritic Spine Density is Decreased in Six- and Twelve-Week-Old GluN1KD Mice ...... 72

3.1.2 Rac1 and WAVE-1 Protein Levels are Decreased in Six- and Twelve-Week Old Mice but Differentially Altered in Three Week Olds ...... 74

3.2 WAVE-Tg and GluN1KD-WAVE Mice Express Wasf1 and WAVE-1 Reliably, Predictably, and Abundantly ...... 75 vi

3.2.1 WASF1 Transgene Expression is Abundant and Similar to Endogenous Wasf1 at the STR and HPC of WAVE-Tg and GluN1KD-WAVE Mice ...... 75

3.2.2 WAVE-1 Protein Levels are Increased in WAVE-Tg Mice and Improved in GluN1KD-WAVE Littermates at the STR ...... 79

3.3 Spine Density Deficits and Maze Exploration Abnormalities are Improved in GluN1KD-WAVE Mice ...... 81

3.3.1 Striatal MSN Dendritic Spine Density is Improved in GluN1KD-WAVE Hybrids ...... 81

3.3.2 Y-Maze Spontaneous Alternation Performance is Normalized in GluN1KD- WAVE Hybrids ...... 81

3.3.3 Eight-Arm Radial Maze WME is Improved in GluN1KD-WAVE Hybrids ...... 82

3.4 Other Behavioural Abnormalities Persist in GluN1KD-WAVE Hybrids ...... 84

3.4.1 Puzzle Box Solution and Memory Deficits Persist in GluN1KD-WAVE Mice ...84

3.4.2 GluN1KD Horizontal Activity and Stereotypy Abnormalities Persist in GluN1KD-WAVE Mice ...... 85

3.4.3 GluN1KD Social Cognition Deficits Persist in GluN1KD-WAVE Mice ...... 85

3.5 Improvements in GluN1KD-WAVE Hybrids are not Reflected by WAVE-1 Levels at the HPC nor CA1 Spines ...... 87

3.5.1 Hippocampal WAVE-1 Levels are not Different Between WT, WAVE-Tg, GluN1KD, and GluN1KD-WAVE Mice ...... 87

3.5.2 CA1 Pyramidal Neuron Spine Density is not Different Between WT, WAVE- Tg, GluN1KD, and GluN1KD-WAVE Mice ...... 88

3.6 RNAseq Analyses Find Sex-Dependent GluN1KD Gene Expression Changes, Likely Involving Sex Hormone Signaling...... 90

3.6.1 GluN1KD Gene Expression Changes from RNAseq Show Internal Validity and Depend on the Sex of Mice ...... 90

3.6.2 Gene List Analyses Predict that Sex-Dependent GluN1KD Transcriptomic Changes are Related to Sex Hormone Signaling and Lead to Convergent Outcomes ...... 95

Chapter 4 Discussion ...... 101

Discussion ...... 101

4.1 Developmentally Correlated Synaptic and Molecular Deficits in GluN1KD Mice ...... 101

vii

4.2 Effect of Restoring WAVE-1 on GluN1KD Phenotypes ...... 107

4.3 Striatal and Hippocampal Differences in GluN1KD and GluN1KD-WAVE mice ...... 111

4.4 Sex-Dependency of Transcriptomic Changes in the GluN1KD STR ...... 114

4.5 Concluding Statements ...... 121

4.5.1 Strengths of the Study ...... 121

4.5.2 Limitations of the Study...... 122

4.5.3 Future Directions ...... 124

4.5.4 Conclusion ...... 125

Bibliography ...... 127

Appendices ...... 193

viii

List of Tables

Introduction Table I-I Summary of NMDAR subunit distribution patterns in the brain Page 6 and their typical characteristics Table I-II Representative list of reported endogenous NMDAR ligands Page 8 Table I-III Major behavioural abnormalities found in previous GluN1KD Page 29 studies

Methods Table II-I Puzzle box configuration over all trials Page 67

Results Table III-I RNAseq-quantified expression of Grin1, Xist, and Y-linked Page 91 genes demonstrate internal validity of results Table III-II Breakdown of GluN1KD gene expression changes by sex and Page 93 directionality of changes Table III-III Breakdown of GluN1KD gene expression changes for “Male Page 94 Genes” by sex and directionality of changes Table III-IV Breakdown of GluN1KD gene expression changes for “Female Page 94 Genes” by sex and directionality of changes Table III-V Representative results from IPA causal network analysis of Page 96 GluN1KD gene expression changes in male mice Table III-VI Representative results from IPA causal network analysis of Page 96 GluN1KD gene expression changes in female mice Table III-VII Enrichr gene list analysis results associate similar disease-related Page 98 records with multiple input lists Table III-VIII Enrichr gene list analysis results associate similar chemical- Page 100 perturbation records with multiple input lists

Appendix 1: Supplemental Figures and Tables Supplemental Protein expression changes in adult male GluN1KD mice Page 193 Table SI relative to WT littermates from previous proteomic studies Supplemental Full differentially-expressed gene lists showing Page 204 Table SII demasculinization or feminization in GluN1KD males and masculinization or defeminization in GluN1KD females

List of Figures

Introduction Figure 1.1 Representative schematic of NMDARs Page 3 Figure 1.2 Representative schematic of downstream molecular signaling Page 11 after NMDAR activation Figure 1.3 Summary of NMDAR-mediated biological processes and their Page 19 dysregulation in key representative neurological disorders Figure 1.4 Major cellular and molecular changes reported in previous Page 41 GluN1KD studies Figure 1.5 Figure 1.5. Representative schematic of WAVE-1 molecular Page 54 mechanisms in relation to NMDARs

Methods Figure 2.1 Schematic of Y-maze test (a) and potential exploration pattern Page 63 results (b-d) Figure 2.2 Schematic of 8-arm radial maze Page 65 Figure 2.3 Schematic of puzzle box test Page 66 Figure 2.4 Schematic of modified social approach behaviour test Page 68

Results Figure 3.1 Representative images (left) and graphs (right) showing MSN Page 73 dendritic spine density of WT and GluN1KD mice aged 3, 6, and 12 weeks Figure 3.2 Representative western blots (top) and their quantifications Page 75 (bottom) measuring Rac1 and WAVE-1 at the STR of 3-, 6-, and 12-week-old WT and GluN1KD mice Figure 3.3 RNAscope visualization of Wasf1 expression at the STR of WT, Page 77 WAVE-Tg, GluN1KD, and GluN1KD-WAVE mice Figure 3.4 RNAscope visualization of Wasf1 expression at the HPC of WT, Page 78 WAVE-Tg, GluN1KD, and GluN1KD-WAVE mice Figure 3.5 Representative western blots (top) and their quantifications Page 80 (bottom) measuring Rac1, WAVE-1, and pS397-WAVE-1 in the STR of WT, WAVE-Tg, GluN1KD, and GluN1KD-WAVE mice Figure 3.6 GluN1KD-WAVE Hybrids have improved MSN spine density Page 83 (a), Y-maze spontaneous 3-arm alternation (b), and 8-arm radial maze WME (c) compared to GluN1KD littermates Figure 3.7 GluN1KD-WAVE hybrids did not show improvements in puzzle Page 86 box (a), open field (b) or social approach behavior (c) test performance Figure 3.8 Assessments of Hippocampal WAVE-1 protein levels (a) and Page 89 CA1 pyramidal neuron spine density (b) in WT, WAVE-Tg, GluN1KD and GluN1KD-WAVE mice

Figure 3.9 Sex-factored RNAseq identified more GluN1KD gene Page 92 expression changes, including most non-sex-factored RNAseq results Figure 3.10 IPA-constructed, Esr1 (implied ERα)-centric, networks of male Page 97 (left) and female (right) GluN1KD gene expression changes have the same structures but are largely opposite in regulation

Discussion Figure 4.1 Representative schematic of the Rho GTPase signaling pathways Page 103 related to cytoskeletal rearrangement as part of synaptic plasticity Figure 4.2 Schematic of NMDAR signaling pathways commonly targeted Page 120 by drugs from Enrichr analysis

Appendix 1: Supplemental Figures and Tables Supplemental Dendritic spine morphology measurements of striatal MSNs in Page 195 Figure S1 WT, WAVE-Tg, GluN1KD and GluN1KD-WAVE mice Supplemental Dendritic spine categorization of striatal MSNs in WT, WAVE- Page 196 Figure S2 Tg, GluN1KD and GluN1KD-WAVE mice Supplemental Apical dendritic spine morphology measurements of CA1 Page 197 Figure S3 pyramidal neurons in WT, WAVE-Tg, GluN1KD and GluN1KD-WAVE mice Supplemental Apical dendritic spine categorization of CA1 pyramidal neurons Page 199 Figure S4 in WT, WAVE-Tg, GluN1KD and GluN1KD-WAVE mice Supplemental Basolateral dendritic spine morphology measurements of CA1 Page 201 Figure S5 pyramidal neurons in WT, WAVE-Tg, GluN1KD and GluN1KD-WAVE mice Supplemental Basolateral dendritic spine categorization of CA1 pyramidal Page 203 Figure S6 neurons in WT, WAVE-Tg, GluN1KD and GluN1KD-WAVE mice

List of Appendices

Appendix 1: Supplemental Figures and Tables Page 193

xii

List of Abbreviations

2-DG : 2-deoxyglucose AD : Alzheimer's disease AKT : protein kinase B AMPA : α-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid AMPAR : AMPA receptor ARC : activity-regulated cytoskeleton-associated protein Arp2/3 : actin related protein 2/3 ASD : autism spectrum disorder ATD : amino-terminal domain BH : Benjamini-Hochberg CA1 : cornu ammonis 1 Cacnb2 : Ca2+ voltage-gated channel auxiliary subunit β2 CaMK : Ca2+/calmodulin-dependent protein kinase cAMP : cyclic adenosine monophosphate CBP : CREB-binding protein Cdc42 : cell division control protein 42 homolog CDK5 : cyclin-dependent kinase 5 CNS : central nervous system CREB : cAMP response element-binding protein CTD : carboxy-terminal domain CTX : cortex CytC : cytochrome c D1R : DA receptor D1 D2R : DA receptor D2 DA : dopamine DAT : DA transporter Ddx3y : DEAD-Box Helicase 3, Y-linked DISC1 : disrupted in schizophrenia 1 Drd2 : DA receptor D2 (gene) Drp1 : dynamin related protein 1 EEG : electroencephalography Eif2s3y : eukaryotic translation initiation factor 2 subunit 3, Y-linked EPSP : excitatory postsynaptic potential ERK : extracellular signal–regulated kinases ERP : event-related potentials Erα : estrogen receptor α Esr1 : estrogen receptor α gene FoxO : forkhead box, class O GAP : GTPase-activating protein GEF : guanine nucleotide exchange factor GluN : glutamate NMDAR subunit GluN1KD : GluN1 knockdown Grin/GRIN : glutamate receptor ionotropic, NMDA GTP : guanosine triphosphate HD : Huntington’s disease HPC : hippocampus xiii

HTT : huntingtin KAL7 : kalirin-7 Kdm5d : lysine demethylase 5D KS : Kolmogorov–Smirnov LBD : ligand-binding domain LIMK1 : LIM domain kinase 1 LIS1 : lissencephaly-1 Lrp1 : lipoprotein receptor-related protein 1 LSN : lateral septal nuclei LTP : long-term potentiation MHCIIB : myosin heavy chain IIB MLC : myosin light chain MSN : medium spiny neurons mTOR : mammalian target of rapamycin NAc : nucleus accumbens Ncam2 : neural cell adhesion molecule 2 NMDA : N-methyl-D-aspartate NMDAR : NMDA receptor nNOS : neuronal nitric oxide synthase NO : nitric oxide NUDEL : NudE-like N-WASP : neural Wiskott-Aldrich syndrome protein Pak1 : p21-activated kinase 1 PCP : phencyclidine PHLPP1β : pleckstrin homology domain and leucine rich repeat protein phosphatase 1β PI3K : phosphoinositide 3-kinase PKC : protein kinase C PPI : prepulse inhibition Rac1 : Ras-related C3 botulinum toxin substrate 1 RAF : rapidly accelerated fibrosarcoma RHEB : Ras homolog enriched in brain RhoA : Ras homolog gene family, member A ROCK : Rho-associated protein kinase RSK2 : ribosomal S6 kinase 2 Shank2 : SH3 and multiple ankyrin repeat domains 2 STEP : striatal-enriched protein tyrosine phosphatase STR : striatum TH : tyrosine hydroxylase Tiam1 : T-lymphoma invasion and metastasis-inducing protein 1 TMD : transmembrane domain TRIO : triple functional domain protein TZG : (tetrazol-5-yl)glycine Uty : ubiquitously transcribed tetratricopeptide repeat containing, Y-linked VCA : Verprolin–Cofilin–Acidic domain (in WAVE-1 protein) WAVE-1 : Wiskott-Aldrich syndrome protein-family verprolin homologous protein 1 WRP : WAVE-associated Rac GTPase-activating protein WT : wild-type Xist : X-inactive specific transcript xiv

YLD : years lived with disability

Chapter 1 Introduction Introduction 1.1 The N-methyl-D-aspartate Receptor

1.1.1 N-methyl-D-aspartate Receptor Characterization and Organization

The N-methyl-D-aspartate (NMDA) receptor (NMDAR) is an ionotropic glutamate receptor and a nonspecific cation channel (Pachernegg et al., 2012; Paoletti et al., 2013; Scheefhals and MacGillavry, 2018). It is named after an early-identified potent agonist that was also a mimic of glutamate (Curtis and Watkins, 1961; Sucher et al., 1996). NMDARs are expressed in organisms throughout the animal kingdom (Sakarya et al., 2007; Alsaloum et al., 2016), and occupy a special place in neuroscience due to their essential roles in learning and memory (Newcomer et al., 2000; Whitlock et al., 2006; Baez et al., 2018).

NMDARs are heterotetrameric receptors. NMDAR subunit proteins are called glutamate NMDAR subunits (GluNs), which are encoded by genes named glutamate receptor ionotropic, NMDA (GRIN in humans; Grin in mice, which will be the major focus in this thesis). NMDARs are typically composed of two GluN1 subunits and some combination of GluN2 or GluN3 subunits (Köhr, 2006; Paoletti et al., 2013). There are four isoforms of GluN2 (A-D) and two isoforms of GluN3 (A-B), each encoded by their own gene (Grin2a-d and Grin3a-b, respectively) (Andersson et al., 2001; Teng et al., 2010). In contrast, GluN1 subunit variation is predominantly through alternative splicing of Grin1, resulting in eight different isoforms named GluN1-1a to GluN1-4a and GluN1-1b to GluN1-4b (Vrajová et al., 2010; Paoletti et al., 2013). For a complete NMDAR, subunits are likely assembled first into GluN1-GluN2/3 heterodimers before the two dimers are subsequently combined (Schüler et al., 2008). Variation in subunit combinations is a common mechanism by which NMDARs differ in signaling or function between different brain regions or cell types (Paoletti et al., 2013; Swanger et al., 2015; Alsaad et al., 2018), different locations in neurons (Köhr, 2006; Bouvier et al., 2015), and between biological samples of different ages (Dumas, 2005; Xing et al., 2006; Henson et al., 2008) and sex (Dong et al., 2007; Ridge et al., 2009; Gray et al., 2015).

1 2

Recently, crystal structures of NMDARs containing GluN1-GluN2B subunits have been solved (Karakas and Furukawa, 2014; Lee et al., 2014), though they required stabilizing mutations. One of these studies described the NMDAR as mushroom-shaped and being approximately 150 Å in height and 125 x 115 Å in width (Lee et al., 2014). In general, NMDARs have four main domains on top of one another: an amino-terminal domain (ATD) furthest away from the cell membrane; an extracellular ligand-binding domain (LBD) containing orthosteric binding sites right “below” the ATD; a transmembrane domain (TMD) that anchors the receptor and regulates cation permeability via Mg2+ blockade; and an intracellular carboxy-terminal domain (CTD) for downstream signaling interactions (Burnashev et al., 1992b; Zhu and Gouaux, 2017; Gu and Wang, 2018). A representative schematic of NMDARs is shown in Figure 1.1. Beyond individual receptors, NMDARs have been reported to be partitioned into ~0.8MDa complexes or ~1.5 MDa supercomplexes at mammalian forebrain synapses, a synapse being the primary signaling connection between neurons (Jessell and Kandel, 1993; Rozental et al., 2000; Frank et al., 2016).

Figure 1.1. Representative schematic of NMDARs. NMDARs are heterotetrameric non- selective cation channels whose subunits determine much of their characteristics. In general, NMDA receptors are composed of two GluN1 and two GluN2 subunits (left). Incorporation of GluN3 subunits generally decreases Ca2+ influx and strength of Mg2+ block (right). Different roles and functions of NMDARs in different tissues, cell positions, and other contexts can largely be facilitated by varying subunit compositions. The CTDs of GluN subunits can bind to various scaffolding and downstream signaling proteins, not detailed in specifics in this image. Binding sites in LBDs are orthosteric sites for endogenous glutamate, glycine, and D-serine while various drugs can bind to NMDARs at either the ATD or LBD. Abbreviations: ATD - amino-terminal domain, LBD - ligand-binding domain, TMD - transmembrane domain, CTD - carboxy-terminal domain. Figure was constructed with information from current NMDAR literature (Cull-Candy and Leszkiewicz, 2004; Paoletti et al., 2013; Sanz-Clemente et al., 2013; Pérez-Otaño et al., 2016; Alam et al., 2017; Zhu and Gouaux, 2017).

NMDARs are not limited to synapses on neurons, nor to only neurons (Murugan et al., 2013; Bouvier et al., 2015; Hogan-Cann and Anderson, 2016). Indeed, current knowledge of NMDARs points to a high level of complexity and nuance – such as subunit composition in crystallography and brain region specificity in protein-interaction studies exemplified above. In addition to the complexity presented by different subunits (Paoletti et al., 2013; Sanz-Clemente et al., 2013) and brain region differences (Newcomer et al., 2000; Frank and Grant, 2017), NMDARs have multi- faceted activation mechanisms (Traynelis et al., 2010; Cummings and Popescu, 2015), location- dependent functions and signaling in neurons (Bouvier et al., 2015; Scheefhals and MacGillavry, 2018) and non-neuronal cells of the brain and periphery (Du et al., 2016; Hogan-Cann and Anderson, 2016). The research presented in this thesis investigates the molecular, cellular, and functional consequences of a whole-body NMDAR deficit in mice as well as how such consequences may be attenuated. Thus, while some NMDAR complexities are not specifically addressed in this experimental body of work, it is still important to appreciate how NMDARs can vary in structure, organization, and function in order to establish context.

While a vast number of investigations relate NMDARs of various subunit compositions to different experimental outcomes in specific biological contexts, some general trends have emerged. The LBD of GluN2 subunits accept glutamate at its orthosteric binding site while GluN1 and GluN3 bind to glycine (Yao et al., 2013). GluN1 isoforms are expressed throughout the central nervous system (CNS) and primarily arise from differential splicing of exons 5, 21, and 22 (Okabe et al., 1999; Hansen et al., 2017; Yi et al., 2018). GluN1 isoforms that differ in exon 5 are particularly important, as those without exon 5-coded motifs in their ATDs (GluN1- 4a) are more sensitive to pH changes and have slower deactivation speed (Regan et al., 2018). GluN2 subunits have been much more heavily studied. In general, higher conductance, higher Ca2+ permeability and faster deactivation is seen with GluN2A- and GluN2B-containing NMDARs than GluN2C or GluN2D NMDARs (Wyllie et al., 2013). GluN3 subunits are reported to negatively regulate NMDARs by decreasing Ca2+ permeability and may play nuanced roles in complex biological processes like synaptogenesis, the formation of new synapses (Matsuda et al., 2002; Henson et al., 2010; Pachernegg et al., 2012; Nelson et al., 2015). These are the general characteristic differences between GluN subunits and their different isoforms. There are also spatial and temporal variations in GluN expression, which when combined with their distinctive characteristics, may imply much about specific NMDAR functions.

There are established general patterns of where and when NMDARs with certain subunits are found, as summarized in Table I-I. As the obligatory subunit, GluN1 is found throughout the brain, though the significance of specific isoform distributions, both spatial and developmental, is not currently clear (Laurie and Seeburg, 1994; Sheng et al., 1994; Yi et al., 2018). GluN2A and GluN2B are expressed throughout the brain, especially at the cortex (CTX) and hippocampus (HPC) (Naassila and Pierrefiche, 2018; Sun et al., 2018) while GluN2C and GluN2D are more limited to the midbrain and cerebellum or brainstem, respectively (Paoletti et al., 2013; Tovar and Westbrook, 2017). There appears to be no unifying broad theme for the peripheral distribution of NMDAR subunits, as GluN2 (as well as GluN3) subunits have been reported in various organs and cell types (Chen et al., 2005; Sproul et al., 2011; Hogan-Cann and Anderson, 2016). In terms of GluN2-NMDAR expression differences by age, GluN2B-GluN2A switching is a well-established phenomenon for developing brains, where primarily GluN2B- containing NMDARs that are abundant in early development are replaced with primarily GluN2A-containing NMDARs (Sanz-Clemente et al., 2013; Baez et al., 2018). GluN2D and GluN2C are similar to GluN2B and GluN2A, respectively, as GluN2D is expressed much more in embryonic brains while GluN2C expression starts after birth (Watanabe et al., 1992; Monyer et al., 1994; Wenzel et al., 1997). Like GluN2A and GluN2B, GluN3 subunits have been investigated and found primarily in the CTX and HPC, though reports of expression in the hind- and midbrain exist as well (Matsuda et al., 2002; Wong et al., 2002; Wee et al., 2008). GluN3A is generally considered to be highly expressed in early development while GluN3B is more highly expressed in later developmental stages (Wong et al., 2002; Prithviraj and Inglis, 2008; Pachernegg et al., 2012). Sex differences in the context of NMDAR research have been investigated often (Dong et al., 2007; McRoberts et al., 2007; Ikeda et al., 2010), though relatively few consider the role of different NMDAR subunits. Two studies have specifically investigated NMDAR expression differences between males and females to some degree, studying postmortem human samples in the context of alcoholism (Ridge et al., 2009) or major depression (Gray et al., 2015). Neither reported significant differences between male and female control samples. Instead, they reported a unique trend of increase for Grin2b in the primary motor CTX of male alcoholics and that only female patients with depression had increased Grin1 and Grin2a-d expression (Ridge et al., 2009; Gray et al., 2015). Taken together, the different NMDAR subunits and their combinations offer a robust mechanism of variation and regulation in vivo. This variation can greatly influence what signals an active NMDAR sends.

Table I-I. Summary of NMDAR subunit distribution patterns in the brain and their typical characteristics. There are some reports of the 8 GluN1 isoforms varying in brain region expression, though the biological significance is unknown. There does not currently appear to be any unifying broad theme or pattern for subunit distributions between sexes or amongst peripheral organs and cell types.

Orthosteric site Other general Subunit Brain region expresison Developmental expression ligand characteristics Obligatory subunit, 8 Both early and late in GluN1 glycine or D-serine Throughout isoforms by alternative development splicing High conductance, fast Throughout; particularly at GluN2A glutamate Late in development deactivation (relative to cortex and hippocampus other GluN2 subunits) High conductance, fast Throughout; particularly at GluN2B glutamate Early in development deactivation (relative to cortex and hippocampus other GluN2 subunits) Low conductance, slow GluN2C glutamate Midbrain and cerebellum Late in development deactivation (relative to other GluN2 subunits) Low conductance, slow GluN2D glutamate Midbrain and brainstem Early in development deactivation (relative to other GluN2 subunits) Incorporation in NMDARs generally GluN3A glycine or D-serine Throughout Early in development decreases Ca2+ conductance Incorporation in NMDARs generally GluN3B glycine or D-serine Throughout Late in development decreases Ca2+ conductance

1.1.2 NMDAR Activation and Subsequent Downstream Signaling

The activation of NMDARs and subsequent downstream effects are also complex, with general trends reported in the literature. In terms of ligands, as mentioned in the discussion of subunit compositions above, NMDARs have binding pockets for both glutamate and glycine (Johnson and Ascher, 1990; Yao et al., 2013). From their initial discovery and characterization in the 1980s, NMDARs have been considered glutamate receptors with unique characteristics, such as the Mg2+ block (McLennan et al., 1981; Coyle, 2012). There have since been many studies of the interplay between glutamate and NMDARs in neurons, astrocytes, and microglia at the brain (Murugan et al., 2013; Pál, 2018) as well as in peripheral tissues, particularly in the context of disease and injury (Seidlitz et al., 2009; Du et al., 2016; Mahieu et al., 2016). However, it was

7 also discovered early on that glycine binding is necessary for NMDAR activation (Kleckner and Dingledine, 1988; Hirai et al., 1996; Cummings and Popescu, 2015). In fact, GluN1-GluN3 heterotetramers, formed when Grin1 and Grin3a/b are co-expressed in Xenopus oocytes or human embryonic kidney cells, are functional excitatory glycine receptors that are unresponsive to glutamate (Chatterton et al., 2002; Smothers and Woodward, 2007; Madry et al., 2010). This is particularly intriguing as canonical glycine receptors are chloride channels whose activation would lead to hyperpolarization of cells (Lynch et al., 2017). The specific biological significance of pure GluN1-GluN3 NMDARs is unclear and may be minimal, but such an atypical combination has been found in vivo in rodent optic nerve myelin (Pina-Crespo et al., 2010). The glycine binding site is also activated endogenously by D-serine and both transmitters may be released by glia or neurons, though glycine may be more likely released by glia onto extrasynaptic NMDARs and D-serine by neurons onto synaptic NMDARs (Papouin et al., 2012; Li et al., 2013; Balu and Coyle, 2015). Evidence points to glycine and D-serine both being important for NMDAR signaling with each potentially coding for different information depending on the biological context (Li et al., 2009, 2013; Acton and Miles, 2017). Finally, it is important to mention that glutamate, glycine, and D-serine are considered the major endogenous NMDAR ligands, but many other endogenous molecules may bind to and regulate NMDARs as well, such as Zn2+ (Rachline, 2005; Madry et al., 2008), cholesterol (Paul et al., 2013; Korinek et al., 2015), and glutathione (Sucher and Lipton, 1991; Varga et al., 1997). A representative list of reported endogenous ligands is shown in Table I-II.

Table I-II. Representative list of reported endogenous NMDAR ligands. The list is a representative highlight of commonly studied and biologically relevant ligands. Specific isomers are not listed, though both can function as ligands in most cases. Related metabolites of listed ligands, such as those processed by redox mechanisms, may also bind to and affect NMDAR activity. Some of the listed ligands are structurally and/or metabolically related while other isoforms and metabolites are not listed for clarity and brevity. Designations of ligands as agonists vs positive allosteric modulators (or antagonists vs negative allosteric modulators) are primarily based on how they are treated in the literature, as competition binding studies have not been completed for all. Similarly, whether ligands increase or decrease NMDAR activity (designation as agonist vs antagonist) is based on consensus in the literature – context-specific, conflicting reports may exist. Abbreviations: PAM - positive allosteric modulator, NAM - negative allosteric modulator.

Ligand Function Sample Reference (Mayer et al., 1989; Chatterton et al., 2002; Mitani et al., Alanine Agonist 2006) Aspartate Agonist (Schell et al., 1997; Morland et al., 2012, 2013) Glutamate Agonist (McLennan et al., 1981; Traynelis et al., 2010; Coyle, 2012) (Kleckner and Dingledine, 1988; Li et al., 2009; Cummings Glycine Agonist and Popescu, 2015) (Olney et al., 1987; Yuzaki and Connor, 1999; Lewerenz and Homocysteate Agonist Maher, 2015) (Monaghan and Beaton, 1991; Guidetti et al., 2004; Barton Quinolinate Agonist et al., 2015) Sarcosine Agonist (Zhang et al., 2009a; Lee et al., 2017) Serine Agonist (Papouin et al., 2012; Li et al., 2013; Balu and Coyle, 2015) (Levy et al., 1991; Sucher and Lipton, 1991; Varga et al., Glutathione Mixed 1997; Ogita et al., 2002) Kynurenate Antagonist (Birch et al., 1988; Barton et al., 2015) Antagonist/ (Yang and Reis, 1999; Gibson et al., 2002; Wang et al., Agmatine NAM 2006) Cholesterol PAM (Paul et al., 2013; Korinek et al., 2015) (Compagnone and Mellon, 2000; Nicolas et al., 2001; Mellon Dehydroepiandrosterone PAM and Griffin, 2002) Pregnenolone sulfate PAM (Wu et al., 1991; Bowlby, 1993; Emnett et al., 2015) (Mony et al., 2011; Carvalho et al., 2012; Zhu and Paoletti, Spermidine PAM 2015) (Traynelis and Cull-Candy, 1990; Low et al., 2000; Chizh et H+ NAM al., 2001; Mony et al., 2011) (Hatton and Paoletti, 2005; Rachline, 2005; Takeda and 2+ NAM Zn Tamano, 2009; McAllister and Dyck, 2017)

Beyond ligands, NMDARs are often termed “coincidence detectors” (Seeburg et al., 1995; Banerjee et al., 2016; Baez et al., 2018) because they also require concurrent local depolarization

9 for pore opening (Clarke et al., 2013; Sweatt, 2016). This concurrent depolarization may be facilitated by nearby α-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid (AMPA) receptors (AMPARs) at the synapse, which are also glutamate-gated cation channels (Burnashev et al., 1992a; Rao and Finkbeiner, 2007; Rogawski, 2011). The requirement for concurrent depolarization is due to the presence of a Mg2+ block in NMDARs (Mayer et al., 1984; Nowak et al., 1984; Johnson and Ascher, 1990). NMDAR pores have binding sites for Mg2+ and the strength of Mg2+ blockade is greatly affected by GluN2 subtype (Clarke et al., 2013; Lee et al., 2014). A key exchange between leucine and serine at an amino acid position between 632 and 657 (specific locations dependent on the specific GluN2 subunit), termed the “GluN2 S/L site”, greatly influences blockade strength. GluN2A- and GluN2B-containing NMDARs, which have serine at the site, demonstrate stronger Mg2+ block while the opposite is true in GluN2C- and GluN2D-contraining NMDARs (Retchless et al., 2012; Clarke et al., 2013). An NMDAR bound with activating ligands without concurrent depolarization is not necessarily inert, however. Such a phenomenon is generally studied in the context of biologically relevant “silent synapses”, which express NMDARs without AMPARs (Isaac et al., 1995; Malinow et al., 2000; Yoon and Choi, 2017). Stimulation of silent synapses generate different types of action potentials (Huang et al., 2015; Li et al., 2017) and trigger their own downstream signaling cascades (Kerchner and Nicoll, 2008; Lee and Dong, 2011). Therefore, the presence or absence of concurrent depolarization is a key factor in eliciting activity from NMDARs.

NMDAR activity and subsequent downstream signaling cascades manifest in two main ways: cation, primarily Ca2+, influx (Xia et al., 1996; Tao et al., 1998; Hardingham and Bading, 2003) and interaction with signaling proteins, particularly through their CTDs (Niethammer et al., 1996; Gardoni et al., 1999; Krapivinsky et al., 2003). These two mechanisms of activity are not necessarily exclusive to one-another, as CTD binding of signaling proteins could help facilitate Ca2+-regulated effects (Strack and Colbran, 1998; Gardoni et al., 1999). However, metabotropic effects of NMDARs independent of channel opening have also been reported and may serve important biological functions (Dore et al., 2016; Gray et al., 2016; Weilinger et al., 2016). Many signaling networks have been associated with NMDAR activation and specific pathways may vary greatly depending on the physiological context (Hardingham and Bading, 2003; Luo et al., 2011). One striking example is the linking of NMDAR activation to potassium channels in neurons and in pancreatic islet cells to produce meaningful inhibitory K+ efflux – opposite to the

10 canonical excitatory Ca2+ influx of NMDARs (Isaac et al., 1995; Yu et al., 1999; Marquard et al., 2015). It is important to understand downstream signaling mechanisms as they may have a drastic impact on the ultimate physiological consequences of NMDAR activation.

Various molecular consequences have been reported downstream of NMDAR stimulation. Multiple studies have reported an important role for the extracellular signal–regulated kinase (ERK) pathway in NMDAR-mediated synapse development and maintenance (Thomas and Huganir, 2004; Tian et al., 2004; Schwechter et al., 2013). As part of this process, Rho GTPases are downstream effectors of NMDAR-regulated actin dynamics in dendritic spines – the physical location of most excitatory synapses on neurons (Nakayama et al., 2000; Dillon and Goda, 2005; Saneyoshi and Hayashi, 2012). Another broad type of signaling network that operates downstream of NMDARs is caspase signaling (Girling et al., 2018; Szychowski et al., 2018), which is part of regulated cell death (Allen et al., 1999; McIlwain et al., 2013). For both cell growth- and death-related physiological contexts, NMDAR activation also regulates phosphoinositide 3-kinase (PI3K) – protein kinase B (AKT) – mammalian target of rapamycin (mTOR) signaling (Gong et al., 2006; Li et al., 2010a; Heras-Sandoval et al., 2014; Wang Yansong et al., 2014). In line with these downstream signals, NMDAR signaling is also interconnected with the signaling of neurotrophins (Ortega et al., 2010; Fukuchi et al., 2015) and cytokines (Jander et al., 2000; Wang et al., 2009). Furthermore, downstream NMDAR signaling can involve Ca2+/calmodulin-dependent protein kinases (CaMKs) and cAMP response element- binding protein (CREB) that regulate gene transcription (Penzes et al., 2008; Hardingham and Bading, 2010; Bustos et al., 2017). It may also involve proteins that couple NMDARs with traditional metabotropic glutamate receptor signaling (Naisbitt et al., 1999; Tu et al., 1999; Kaizuka and Takumi, 2018), nitric oxide (NO) destined for nearby cells’ cyclic-guanosine- monophosphate signaling pathways (Belsham et al., 1996; Araújo et al., 2018), and more (Ko et al., 2002; Banko et al., 2004; Rodríguez-Durán and Escobar, 2014). Finally, NMDAR signaling can also regulate, and be regulated by, sex hormones (Hsu et al., 2000; Bi et al., 2001; Hojo et al., 2008; van den Buuse et al., 2017) and the activity of other receptors (Mao, 1999; Pei et al., 2004; Kruse et al., 2009; Chen et al., 2018). A representative schematic of commonly studied NMDAR signaling pathways is shown in Figure 1.2. Overall, NMDAR signaling spans a multitude of molecular mechanisms dependent on cellular context and physiological function.

Figure 1.2. Representative schematic of downstream molecular signaling after NMDAR activation. NMDAR activation plays an integral role in generating, developing, maintaining, and eliminating cells and synapses. In general, synaptic NMDAR signaling represents pro-survival, pro-growth, and synaptic strengthening signals while extrasynaptic NMDAR signals largely serve opposing biological processes. The former is primarily mediated by GluN2A-containing NMDARs while the latter uses GluN2B-NMDARs. Chronic or excessive NMDAR stimulation often overlaps with extrasynaptic NMDAR signaling mechanisms. These pathways mainly represent neuronal NMDAR signaling; nonneuronal signaling may involve different molecules or molecular interactions and have been omitted for clarity. Similarly, other NMDAR signaling

12 contexts have been omitted for clarity, such as molecular mechanisms involved with LTD, reactive oxygen species, and crosstalk pathways with other receptors. Green pointed arrows indicate a positive relationship of some form, such as enhanced activation or release from a sequestering factor. Different shades of green only serve to clarify overlapping arrows and do not represent any meaningful difference. Red blunted lines indicate a negative relationship, causing some form of inhibition, decreased activity, or sequestering of the target. Black arrows indicate a more complex regulatory relationship from mixed or limited experimental results. Blue proteins indicate those that are affected by both synaptic and extrasynaptic NMDAR signaling. Abbreviations are the same as those used in text, in addition to the following: RSK2 – ribosomal S6 kinase 2; PHLPP1β – pleckstrin homology domain and leucine rich repeat protein phosphatase 1β; STEP – striatal-enriched protein tyrosine phosphatase; PKC – protein kinase C; RAF – rapidly accelerated fibrosarcoma; ARC – activity-regulated cytoskeleton-associated protein; RHEB – Ras homolog enriched in brain; CBP – CREB-binding protein; FoxO – forkhead box, class O; RhoA – Ras homolog gene family, member A; DISC1 – disrupted in schizophrenia 1; KAL7 – kalirin-7; Rac1 – Ras-related C3 botulinum toxin substrate 1; WAVE-1 – Wiskott-Aldrich syndrome protein-family verprolin homologous protein 1; Arp2/3 – actin related protein 2/3; Pak1 – p21-activated kinase 1; LIMK1 – LIM domain kinase 1; CytC – cytochrome c; nNOS – neuronal nitric oxide synthase; CDK5 – cyclin-dependent kinase 5; Drp1 – dynamin related protein 1; LTP – long-term potentiation. This figure was constructed with literature cited within text as well as additional articles (Vallano, 1998; Snyder and Ferris, 2000; Korhonen et al., 2004; Jocoy et al., 2011; Uribe and Wix, 2011; Baudry and Bi, 2016; Haiying et al., 2017; Mackay et al., 2018).

To sum, the NMDAR is a heterotetrameric cation channel that is canonically activated with three concurrent events: glutamate binding to pockets involving GluN2 subunits, glycine binding to GluN1/GluN3 subunits, and local membrane depolarization to relieve a Mg2+ block (Kleckner and Dingledine, 1988; Clarke et al., 2013; Paoletti et al., 2013). NMDARs are composed from 14 different subunits that have different spatial and developmental distributions and determine the unique characteristics of each receptor (Paoletti et al., 2013; Wyllie et al., 2013; Hogan-Cann and Anderson, 2016). Upon activation, NMDARs canonically allow Ca2+ entry into the cell and trigger downstream signaling cascades that regulate gene expression and cell survival (Xia et al.,

1996; Hardingham and Bading, 2003; Luo et al., 2011). These fundamental characteristics lend themselves to the various physiological roles of NMDARs.

1.1.3 Physiological Roles of NMDARs

NMDARs play numerous roles depending on the context of their signaling (Shouval et al., 2002; Luo et al., 2011; Hogan-Cann and Anderson, 2016), as shown in Figure 1.2. As mentioned above, NMDAR signaling is involved with the maintenance of synapses (Tian et al., 2004; Schwechter et al., 2013) and with regulated cell death (Allen et al., 1999; Szychowski et al., 2018). NMDARs, particularly at the synapses of cortical and hippocampal neurons, are important for learning and memory behaviour (Newcomer et al., 2000; Whitlock et al., 2006; Baez et al., 2018; Roelfsema and Holtmaat, 2018). As such, they also play an important role in molecular and cellular correlates of learning and memory, such as the commonly-studied long- term potentiation (LTP) and synaptic plasticity (Morris et al., 1986; Calabresi et al., 1992; Shouval et al., 2002; Bouvier et al., 2015). The role of NMDARs in signaling cell death has also been well established, particularly in the context of excitotoxicity where overactivation of neurons leads to neuronal death (Hardingham and Bading, 2003; Weilinger et al., 2016; Girling et al., 2018). These two major functions of NMDARs may seem discordant and perhaps unrelated to one another. However, a simple reframing of NMDARs’ role in learning and memory as a role in selecting which neurons and synaptic connections should survive and thrive may help connect it to cell death signaling. Put another way, the core underlying principle of NMDARs’ physiological roles is determining when a cell (or a part of a cell, like a synapse) should grow and when it should die. This paradigm is further supported by the fact that NMDARs also have roles beyond the synapse that affect learning and memory, such as regulation of neurogenesis – the proliferation and development of new neurons (Gould et al., 1997; Nacher and McEwen, 2006; Lai et al., 2016). Commonly proposed physiological roles for NMDARs outside the brain also often involve cell growth or death, though research on peripheral NMDARs is relatively limited (Du et al., 2016; Hogan-Cann and Anderson, 2016). It is therefore not surprising that many researchers discuss NMDAR signaling and function as a contrasting combination of cell survival and cell death (Hardingham and Bading, 2003; Zhang et al., 2007; Luo et al., 2011).

The gating and activation properties of NMDARs particularly lend themselves to determining the fates of cells and synapses. Whether a cell or synapse would survive and grow or weaken and die may be determined by the context of NMDAR activation (Hardingham and Bading, 2003; Zhang et al., 2007). As mentioned above, NMDARs are coincidence detectors that are opened with concurrent local membrane depolarization but can also signal without allowing a Ca2+ influx (Seeburg et al., 1995; Dore et al., 2016; Weilinger et al., 2016; Baez et al., 2018). This ability for NMDARs to detect and respond differentially to distinct levels of stimulation may be a key mechanism by which they signal synaptic strengthening or cell death. LTP, the enhancement of synaptic transmission efficiency as determined by changes in excitatory postsynaptic potentials (EPSPs), is induced at synapses with high frequency stimulation. LTD (long-term depression, the opposite weakening of synaptic connections) is induced with low or moderate frequency stimulation (Herron et al., 1986; Bliss and Collingridge, 1993; O’Dell and Kandel, 1994; Shouval et al., 2002). Coupled with cell death signaling being triggered by NMDARs when there is overstimulation (Shan et al., 1997; Carvajal et al., 2016; Girling et al., 2018), there is ample evidence of NMDARs facilitating synaptic strengthening or cell death as a result of the magnitude of stimulation. However, the gating properties of NMDARs are not the only mechanisms by which cell growth or cell death signaling can be chosen. The location of NMDARs may activate specific signaling pathways: synaptic NMDARs promote cell survival while extrasynaptic NMDARs, potentially responding to glutamate spillage from overactive synapses, promote cell death (Hardingham et al., 2002; Papadia et al., 2008; Okamoto et al., 2009; Xu et al., 2009). The selectivity of these signaling pathways may be facilitated by predominantly GluN2A-NMDARs at the synapse and predominantly GluN2B-NMDARs outside of the synapse, though this has not been established definitively (Hardingham and Bading, 2010; Luo et al., 2011). Finally, the fact that NMDARs are non-selective cation channels may also allow them to signal cell death, as apoptosis has been reported with K+ efflux mediated by NMDARs (Yu et al., 1999). There is an overall abundance of experimental evidence that links cell survival or cell death signaling to different, specific contexts of NMDAR activation.

Moving away from the general concepts of cell survival and death, as well as synaptic strengthening and weakening, towards specific physiological contexts will reveal how NMDAR signaling affects living, multi-cellular organisms. As mentioned before, proper NMDAR signaling in the brain is important for cognition, particularly learning and memory (Malhotra et

15 al., 1996; Newcomer et al., 2000; Barria and Malinow, 2005; Gilmour et al., 2012). Many studies reporting a significant role for NMDARs in cognition operate with outcomes related to synaptic plasticity (Gilbert and Mack, 1990; Barria and Malinow, 2005), brain development and maintenance (Papadia et al., 2008; Komuro et al., 2015; Lai et al., 2016), and in vivo cognitive behaviour (McHugh et al., 2007; Morris et al., 2013).

Synaptic plasticity is a general term for the ability of synapses to change their transmissive strength and efficiency in response to local activity (Shouval et al., 2002; De Pittà et al., 2016). LTP is a form of synaptic plasticity and is commonly studied as a cellular correlate of learning and memory (Abbott and Nelson, 2000; De Pittà et al., 2016). As mentioned above, one method of LTP induction is through high frequency stimulation. This stimulation increases, and often depends on, NMDAR activity to induce post- and pre-synaptic changes that facilitate the more efficient synaptic transmission that is characteristic of LTP (Bliss and Collingridge, 1993; Shouval et al., 2002; Lisman and Raghavachari, 2006). These NMDA-mediated changes involve, but are not limited to, colocalized CaMK-II activity (Barria and Malinow, 2005), presynaptic actin and scaffolding protein rearrangement (Antonova, 2001; Colicos et al., 2001), trafficking of AMPARs to the synapses to enhance transmission (Malinow and Malenka, 2002; Bredt and Nicoll, 2003; Kim et al., 2005), activation of the Rho GTPases RhoA and Cdc42 to expand hippocampal dendritic spines (Murakoshi et al., 2011), and CaMKII- and protein-synthesis- dependent stabilization of dendritic spines (De Roo et al., 2008; Hill and Zito, 2013). Pharmacologically antagonizing NMDARs, fittingly, has been reported to block the LTP between rat perforant path and dentate gyrus (Gilbert and Mack, 1990), in mature and newborn mouse dentate granule cells (Ge et al., 2007), and in rat hippocampal neurons (Barria and Malinow, 2005). As to the propriety of studying LTP and synaptic plasticity as cellular correlates of learning and memory, recent experiments with optogenetics provide strong support for this paradigm. Association of a neutral optogenetics-based conditioned stimulus with a foot shock in rats was NMDAR-dependent, extinguishable by an LTD protocol, and restorable with an LTP protocol (Nabavi et al., 2014). Furthermore, LTP induction in dentate gyrus cells attenuated spine density deficits and memory impairments in the mutant β-amyloid precursor/presenilin-1 double-transgenic mouse model of Alzheimer's disease (AD) (Roy et al., 2016). NMDARs play an important role in facilitating and regulating LTP and synaptic plasticity, and thus learning and memory.

There are numerous studies that link NMDARs to cognition and brain homeostasis beyond LTP and the synapse. NMDAR activation can suppress apoptotic neurodegeneration in perinatal developing mouse brains (Ikonomidou et al., 1999; Heck et al., 2008), promote neurogenesis in neonatal rats after hypoxic-ischemic injury (Lai et al., 2016), act as a signal for mouse cerebellar granule cell migration during brain development (Komuro and Rakic, 1993; Komuro et al., 2015), and protect rat cortical neurons from oxidative insult by increasing thioredoxin activity (Papadia et al., 2008). NMDAR activation may also regulate sleep-wake cycles (Manfridi et al., 1999; Kawai et al., 2015) as well as facilitate proper sleep architecture (Armstrong-James and Fox, 1988) and sleep-dependent memory consolidation (Feld et al., 2013).

In terms of direct assessments of learning and memory, there is an abundance of behavioural evidence that points to an important role for NMDARs. Many of these studies involve investigating the consequences of perturbing NMDAR signaling. Acute NMDAR antagonism in mice resulted in diminished in vivo perforant path – to – dentate gyrus LTP, impaired spatial memory formation, and impaired avoidance learning (Walker and Gold, 1991). Chronic NMDAR blockade in rats resulted in both suppressed in vivo LTP at the dentate gyrus and impaired spatial memory formation in multiple configurations of the Morris water maze (Morris et al., 1986, 2013). Distinct from a pharmacological approach, knocking out GluN1 subunits in the granule cells of the mouse dentate gyrus produces mice with impaired contextual pattern learning and discrimination (McHugh et al., 2007). Taken together, numerous studies operating with a wide range of experimental approaches and outcomes indicate that NMDARs are important for synaptic strengthening, neuronal maintenance, and learning and memory.

Another major role of NMDARs is to propagate signals related to cell death and synaptic weakening (Shouval et al., 2002; Hardingham and Bading, 2003; Luo et al., 2011; Girling et al., 2018). NMDAR-mediated cell death signaling cascades are often studied and reported in the context of injury and disease, such as excitotoxic processes during ischemia (Seubert et al., 1989; Gao et al., 2005) and other neuropathological states (Weilinger et al., 2016; Wong et al., 2016). While regulated cell death is certainly an important part of normal brain development (Oppenheim, 1991; Vaux and Korsmeyer, 1999; Kuan et al., 2000), the role of NMDARs in this developmental context is primarily to promote neuronal survival after activation (Fiske and Brunjes, 2001; Martins et al., 2005; Heck et al., 2008). In the context of adult neurogenesis, NMDAR activation may be considered as indirectly supporting cell-death signaling by

17 suppressing growth-related mechanisms (Cameron et al., 1995; Gould and Cameron, 1996). Generally, nonpathological NMDAR-mediated cell death requires further characterization and study as current evidence is limited.

Synaptic weakening by synapse pruning or LTD are normal physiological processes mediated by NMDARs. Synaptic pruning is an important part of brain development whereby redundant connections between neurons, overabundant in youth, are reduced to establish efficient, mature neuronal networks (Uesaka et al., 2014; Lieberman et al., 2018; Millán et al., 2018). Many experiments have shown that NMDAR blockade disrupts synaptic pruning and leaves more aberrant synaptic connections intact (Cline and Constantine-Paton, 1989; Rabacchi et al., 1992; Hickmott and Constantine-Paton, 1997; Lüthi et al., 2001). It is difficult to explicitly and specifically state what purpose synaptic pruning serves on a macroscopic level, but disruptions in synaptic pruning have been linked to autism spectrum disorders (ASDs) (Tang et al., 2014; Kim et al., 2017), schizophrenia (Sekar et al., 2016; Johnson and Stevens, 2018), and other neurological disorders (Johnston et al., 2001; Stertz et al., 2013). NMDAR-mediated synapse elimination can involve molecular signals related to apoptosis without fully prompting cell death, such as caspase-3 activation, and shares similar molecular mechanisms with NMDAR-mediated LTD (Li et al., 2010b p.3; Jiao and Li, 2011; Erturk et al., 2014; Piochon et al., 2016).

LTD of synapses can be induced with low frequency stimulation of presynaptic neurons, different than LTP that needs high frequency stimulation and both pre- and post-synaptic neuronal activity (Herron et al., 1986; Shouval et al., 2002; Luscher and Malenka, 2012). As mentioned before, optogenetically-induced LTD can extinguish associative memory in rats (Nabavi et al., 2014). Furthermore, specific blockade of GluN2B-NMDARs prevents LTD induction and conditioned fear memory extinction in rats (Dalton et al., 2012). LTD is not only important for erasing a memory, however, but also memory formation de novo. Hippocampal LTD, for example, may work together with LTP changes to encode memories of objects in new environments (Kemp and Manahan-Vaughan, 2004) or social stress (Artola et al., 2006) in rats. Combined with NMDARs’ robust involvement with LTP and neuronal health, NMDAR- mediated synaptic pruning and LTD greatly affect cognition in vivo. It is therefore not surprising that there is a rich scientific literature regarding NMDARs in physiological responses to diseases and insults as well as NMDAR dysfunction in neurological disorders.

1.1.4 NMDARs in Health and Disease; Links to Neurological Disorders

Disruptions of NMDARs can lead to various neurological disorders and associated symptoms (Luby et al., 1959; Irani et al., 2010; Burnashev and Szepetowski, 2015). A summative diagram of how NMDAR disruptions relate to pathological processes and states is shown in Figure 1.3. Some of the earliest evidence supporting a significant role of NMDARs in mental illness come from pharmacological studies. Pharmacological blockade of NMDARs with ketamine or phencyclidine (PCP) produces positive, negative, and cognitive symptoms similar to those seen in schizophrenia (Luby et al., 1959; Allen and Young, 1978; Krystal et al., 1994; Malhotra et al., 1996). NMDAR agonism can mediate the effects of glutamatergic excitotoxins (Giulian et al., 1990; Berman and Murray, 1997) and can induce seizures and convulsions (Schoepp et al., 1991; Toscano et al., 2008). More chronic and permanent NMDAR disruptions, such as by genetic mutations of NMDAR subunit genes, can also serve as a causal factor in neurological disorders. Many studies have linked human GRIN mutations to epilepsy, schizophrenia, ASDs, attention deficit hyperactivity disorder, and other forms of cognitive impairment (Soto et al., 2014; Burnashev and Szepetowski, 2015; Yuan et al., 2015). Anti-NMDAR autoimmune encephalitis, where patients produce antibodies targeting their own NMDARs, also produce neuropsychiatric symptoms, cognitive impairments, and seizures (Dalmau et al., 2007; Florance-Ryan and Dalmau, 2010; Irani et al., 2010). Overall, abundant evidence suggests that NMDAR disruption plays a very important role in many neurological disorders. Both too much and too little NMDAR signaling can be detrimental and may even lead to similar outcomes (Hardingham and Bading, 2003; Parsons et al., 2007; Schmeisser et al., 2012; Won et al., 2012), but each direction of dysfunction is more closely associated with certain disorders than others. It is therefore useful to categorize the health-related contexts and consequences of NMDAR hyperfunction and hypofunction.

Figure 1.3. Summary of NMDAR-mediated biological processes and their dysregulation in key representative neurological disorders. Mentioned thus far, proper NMDAR signaling is important for the genesis, development, maintenance, and elimination of cells and synapses. Altered synaptic plasticity can have multiple manifestations, such as an altered excitation- inhibition balance that is epileptogenic or an impairment of synaptic strengthening that weakens learning and memory. Cell growth and survival mechanisms are especially important for neurodevelopment while cell death contributes to late-onset neurodegenerative disorder. Green and red arrows denote positive and negative relationships, as denoted in Figure 1.2. Black, purple, and grey arrows only serve to establish a relationship between generalized NMDAR signaling pathways, the biological processes they serve, and the disorders both contribute to

20 when dysregulated. These assorted colours are for visual clarity and do not represent any meaningful differences. Abbreviations follow those used in text and previous figures.

NMDAR hyperactivity and subsequent excitotoxicity is a major component of the body’s responses to physiological insults. NMDAR-mediated excitotoxicity is considered part of physiological responses to ischemia (Seubert et al., 1989; Gao et al., 2005), stroke (Weilinger et al., 2016; Wong et al., 2016), infections (Haughey et al., 2001; Costa et al., 2017), and trauma (Lynch and Dawson, 1994; Aiba and Shuttleworth, 2012; Estrada-Rojo et al., 2018). It has been demonstrated in rat lung tissue (Said et al., 1996) and postulated as a potential pathological mechanism in other peripheral tissues (Gill et al., 2000; Ortiz et al., 2006). Excitotoxicity may very well be the most prominent NMDAR-related pathological process in peripheral tissues, but this may be because peripheral NMDAR studies are relatively scarce (Du et al., 2016; Hogan- Cann and Anderson, 2016). The main commonality between different physiological insults, occurring anywhere in the body, that can cause NMDARs to respond similarly is very likely the excessive glutamate release from tissue damage or stress (Seubert et al., 1989; Weilinger et al., 2016; Costa et al., 2017; Estrada-Rojo et al., 2018). Indeed, many clinical trials were conducted to evaluate NMDAR antagonists for preventing cell death, particularly after ischemia, though these trials generally produced negative results (Morris et al., 1999; Davis et al., 2000; Albers et al., 2001; Sacco et al., 2001; Yurkewicz et al., 2005). It is thought that these NMDAR antagonists failed clinical trials because they also disrupted important nonpathological NMDAR functions, including neuroprotection (Ikonomidou and Turski, 2002; Muir, 2006; Shohami and Biegon, 2014). This is consistent with previously mentioned reports of NMDAR activity promoting neurogenesis after hypoxic ischemia (Lai et al., 2016) and enhancing antioxidant mechanisms (Papadia et al., 2008), as well as other reports of NMDAR-dependent mechanisms for adapting to acute ischemia (Grabb and Choi, 1999; Gonzalez-Zulueta et al., 2000). Fittingly, the only NMDAR antagonist used to treat AD, memantine, is largely thought to be effective because it has a rapid off-rate and preferentially blocks excessive NMDAR activity (Tariot et al., 2004; Lipton, 2006; Parsons et al., 2007). NMDARs therefore clearly play a prominent role in pathological, preservation, and recovery processes (Petrovic et al., 2005; Bading, 2017; Olloquequi et al., 2018). However, it is also clear that simply blocking NMDARs during or after injury may not be effective in improving health outcomes (Ikonomidou and Turski, 2002; Lipton, 2006; Shohami and Biegon, 2014). Manipulating NMDAR activity to prevent

21 excitotoxicity and promote survival may be the ideal therapy, but more research is required to understand when, where, and with what drug NMDARs need to be regulated.

Excessive and/or inappropriate NMDAR activation has also been linked to neuropathic pain (Alles and Smith, 2018; Inquimbert et al., 2018) and neurodegenerative diseases such as AD (Danysz and Parsons, 2003; Okamoto et al., 2009; Tönnies and Trushina, 2017). In the case of neuropathic pain, the involvement of NMDAR signaling in its etiology is well supported (Zhuo, 2008; Alles and Smith, 2018). Rather than excitotoxicity, NMDAR-mediated synaptic changes such as LTP are key facilitators of enhancing pain signaling (Svendsen et al., 1998; Ren and Dubner, 2007; Alles and Smith, 2018). Treatment of neuropathic pain using the NMDAR antagonist ketamine is generally effective, but clinically limited due to undesirable side effects such as hallucinations and cognitive impairment (Hewitt, 2000; Aiyer et al., 2018). Inappropriate NMDAR hyperactivity may contribute to AD via excitotoxic mechanisms and neurodegeneration (Hynd et al., 2004; Zhang et al., 2016; Tönnies and Trushina, 2017). Amyloid-β protein, the aggregates of which are a hallmark of AD, can directly activate NMDARs (Texidó et al., 2011), increase glutamate release from microglia (Noda et al., 1999), disrupt Ca2+ homeostasis to enhance NMDAR-mediated excitotoxicity (Mattson et al., 1992; Bezprozvanny and Mattson, 2008), and decrease the ratio of synaptic-to-extrasynaptic NMDAR activity (Snyder et al., 2005; Li et al., 2011; Talantova et al., 2013). A relative increase in extrasynaptic NMDAR activity is particularly meaningful for signaling excitotoxicity and cell death (Hardingham et al., 2002; Parsons and Raymond, 2014) and is consistent with reports of synaptic NMDAR inhibition by amyloid-β (Snyder et al., 2005; Shankar et al., 2007; Zhang et al., 2009b). As mentioned before, memantine may be useful for treating AD by suppressing excessive NMDAR activity (Tariot et al., 2004; Parsons et al., 2007). Memantine may also preferentially block active extrasynaptic NMDARs, making it particularly suitable for preventing NMDAR-mediated neurodegeneration (Lipton, 2006; Xia et al., 2010). The differential status of synaptic and extrasynaptic NMDARs in AD reinforces the idea of multi-faceted NMDAR changes in neurological disorders, as both increased and decreased NMDAR activity is reported. Much of the evidence supports a strong role of enhanced NMDAR activation leading to excitotoxic neurodegeneration (Hynd et al., 2004; Parsons and Raymond, 2014; Tönnies and Trushina, 2017). However, decreased synaptic NMDAR signaling may also simultaneously contribute to LTD and memory loss in AD (Snyder et al., 2005; Shankar et al., 2007; Zhang et al., 2009b). NMDAR hyperactivity, particularly

22 through excitotoxic mechanisms, can overall cause great damage to surrounding tissue in many contexts, but other forms of NMDAR disruption can also contribute to disease symptoms.

Epilepsy and ASDs are two illustrative examples of neurological disorders with mixed- direction NMDAR disruptions. Despite epilepsy being a neurological disorder of hyperexcitability in the brain (Devinsky et al., 2013), most of the functionally-assessed Grin mutations related to epilepsy resulted in unchanged or decreased NMDAR currents (Xu and Luo, 2018). Similarly, most animal models of epilepsy had unchanged or decreased levels of Grin mRNA when assessed (Ghasemi and Schachter, 2011). It is not completely clear how loss-of- function NMDAR mutations contribute to a hyperexcitable brain (Oyrer et al., 2018), but NMDAR hypofunction can lead to unstrengthened inhibitory synaptic connections that allow for excessive non-NMDAR glutamatergic signaling (Moscato et al., 2014; Chiu et al., 2018). That is not to say that only NMDAR deficits have been associated with epilepsy, however. Firstly, NMDAR-mediated excitotoxicity and cell death is, expectedly, a concerning mechanism in epilepsy (Quincozes-Santos et al., 2014; Salazar and Tapia, 2015). NMDAR activation has also been reported to be necessary for developing epilepsy in rats (Rice and DeLorenzo, 1998). Endogenous NMDAR activity is increased in resected tissue from mesial temporal lobe epilepsy patients (Banerjee et al., 2016), NMDA-mediated EPSPs were prolonged in epileptic hippocampal tissue from humans (Isokawa and Levesque, 1991), and NMDAR density was increased in epileptic human cortical tissue when assessed by radioligand binding (Steffens et al., 2005). Furthermore, use of memantine to block excess NMDAR has been suggested as a personalized therapy option for some patients (Pierson et al., 2014) while NMDAR enhancers may have a very narrow therapeutic window (Oyrer et al., 2018). However, D-cycloserine, a glycine-site partial agonist of NMDARs, has been reported to have anticonvulsant effects in preclinical models (Peterson and Schwade, 1993; De Sarro et al., 2000).

Both memantine (Chez et al., 2007; Erickson et al., 2007; Ghaleiha et al., 2013) and D- cycloserine have also been shown to be beneficial for patients with autism in clinical trials (Posey et al., 2004; Urbano et al., 2014, 2015; Minshawi et al., 2016). Increased and decreased NMDAR function has been linked to animal models of ASDs. Mice with a SH3 and multiple ankyrin repeat domains 2 (Shank2) mutation that models ASDs showed impaired NMDAR- mediated LTP and LTD in the HPC along with a decreased NMDAR/AMPAR signal ratio (Won et al., 2012). However, Shank2 knockout mice had increased GluN subunits at the HPC and

23 striatum (STR) as well as increased hippocampal NMDAR/AMPAR signaling ratio (Schmeisser et al., 2012). Knockout of Grin1 in parvalbumin interneurons of mice resulted in autism-relevant social behaviour deficits (Saunders et al., 2013) while prenatal valproic acid exposure in rats, a model of teratogenic autism, resulted in increased cortical protein levels of GluN2A and GluN2B as well as improved LTP (Rinaldi et al., 2007). In the case of ASDs, some of the variation in NMDAR disruption may be due to age, as early brain development aberrations can differ from consequent mature brain abnormalities (Contestabile, 2000; Rinaldi et al., 2007; Spooren et al., 2012). Increased and decreased NMDAR activity may therefore contribute to the same neurological disorder via altering different physiological functions at different sites and timepoints.

Schizophrenia is the prime example of a neurological disorder associated with NMDAR hypofunction. To start, the glutamatergic theory of schizophrenia, first proposed by Arvid Carlsson (Carlsson and Carlsson, 1990; Carlsson et al., 1999), is centered on a hypothesis of NMDAR hypofunction. NMDAR hypofunction is a current, major paradigm in schizophrenia research (Gilmour et al., 2012; Moghaddam and Javitt, 2012; Nakazawa et al., 2017). The core findings underlying the NMDAR hypofunction hypothesis are the reports of schizophrenia-like positive, negative, and cognitive symptoms seen in humans after pharmacological NMDAR blockade (Luby et al., 1959; Krystal et al., 1994; Gilmour et al., 2012; Moghaddam and Javitt, 2012). Furthermore, when given to patients with schizophrenia, ketamine exacerbated their symptoms (Lahti et al., 1995, 2001; Malhotra et al., 1997). The role of genetic NMDAR dysfunction in schizophrenia is likely more complex and could include many disrupted physiological functions. Reported GRIN mutations of schizophrenia patients are fairly evenly distributed amongst non-obligatory GluN subunits, though GRIN2A and GRIN2B mutations may be good general representations of NMDAR dysfunction in schizophrenia (Paoletti et al., 2013; Yuan et al., 2015). GRIN2A mutations are particularly common in epilepsy and, consistent with the NMDAR hypofunction hypothesis, are often loss-of-function mutations (Xu and Luo, 2018). Thus, underactive NMDARs may also facilitate a hyperglutamatergic state in schizophrenia, which is supported by evidence like the attenuation of ketamine-induced psychosis via inhibition of glutamate release (Anand et al., 2000). An increased glutamatergic tone may act as noise in the brain to disrupt neuronal networks, manifesting as psychosis (Corlett et al., 2011; Moghaddam and Javitt, 2012). GRIN2B mutations have been considered particularly common to

24 neurodevelopmental disorders and cognitive impairment, potentially pointing to disrupted brain development and synaptic pruning in schizophrenia (Ohno et al., 2010; Soto et al., 2014; Yuan et al., 2015). This would be consistent with perinatal NMDAR blockade in animals producing long- lasting, schizophrenia-relevant phenotypes (du Bois and Huang, 2007) and with the disrupted synaptic connectivity found in post-mortem brains of schizophrenia patients (Garey et al., 1998; Glantz and Lewis, 2000; Black et al., 2004; Weickert et al., 2013). Clinical studies of D- cycloserine treatment for schizophrenia, often as adjunct therapy, reported efficacy in some cases (Evins et al., 2002; Heresco-Levy et al., 2002; Goff et al., 2008), no effect in others (Heresco- Levy et al., 1998; Duncan et al., 2004a; Goff et al., 2005), and even symptom worsening in still other cases (Cascella et al., 1994; Berckel et al., 1999; Goff et al., 1999). While these studies varied greatly in design, the complex etiology of schizophrenia may ultimately mean that D- cycloserine is best suited as a personalized therapy for some patients.

Overall, many studies strongly link NMDAR function to proper brain health and its disruption to neurological disorders (Newcomer et al., 2000; Paoletti et al., 2013; Burnashev and Szepetowski, 2015). The illnesses listed above are just some of the more well-studied, illustrative examples of how NMDAR dysfunction can contribute to neuropsychiatric pathology and symptoms. Briefly, other examples of NMDAR involvement in mental illnesses include the rapid antidepressant effects of NMDAR blockade in major depression (Li et al., 2010a), investigations of NMDAR-mediated synaptic plasticity to understand and treat post-traumatic stress disorder (Yamamoto et al., 2008; Mikics et al., 2017), and evidence of dysregulated glutamate transmission in obsessive-compulsive disorder (Alonso et al., 2012). Disrupted NMDAR function can have many consequences and manifest as various symptoms in humans. Current evidence suggests that moderating NMDAR activity, rather than net activation or inhibition, with drugs like memantine and D-cycloserine, may improve mental health outcomes. However, the reported benefits of these drugs are often only moderate for some patients, if at all (Herrmann et al., 2011; Schade and Paulus, 2016). Further research is therefore warranted to understand the molecular and cellular mechanisms underlying neurological disorders related to NMDAR dysfunction; this in turn will help the development of future treatments that are more specific and efficacious. My research uses a genetic mouse model of NMDAR deficiency to study such mechanisms.

1.2 Mouse Model of NMDAR Deficiency by GluN1 Knockdown

1.2.1 Background and Overview of GluN1 Knockdown Mice

The GluN1 knockdown (GluN1KD) mouse model was generated by Dr. Amy J. Ramsey in the lab of Dr. Beverly H. Koller and was first published in 1999 (Mohn et al., 1999). As the name implies, GluN1KD mice have decreased levels of the obligatory GluN1 subunit of NMDARs. This is a global decrease that has been confirmed by assessments of RNA (Mohn et al., 1999; Bodarky et al., 2009), protein (Mohn et al., 1999; Bodarky et al., 2009; Ramsey et al., 2011), and radioligand binding (Mohn et al., 1999; Duncan et al., 2002, 2004b). GluN1KD heterozygotes are less characterized, but have demonstrated WT-like levels of NMDARs in all brain regions assessed by radioligand binding (Duncan et al., 2002). Functionally, NMDAR deficits have been less characterized in GluN1KD mice. When measured as evoked EPSPs in striatal medium spiny neurons (MSNs), NMDAR-specific currents were reduced (Jocoy et al., 2011). Injection of the direct and potent NMDAR agonist (tetrazol-5-yl)glycine (TZG) in GluN1KD mice produced muted Fos protein expression, a marker of increased intracellular Ca2+, in many brain regions (Inada et al., 2007; Duncan et al., 2008). NMDAR function was thus demonstrated to be generally deficient in GluN1KD brains. However, Fos induction by TZG and other methods in the GluN1KD STR was often similar to wild-type (WT) littermates (Miyamoto et al., 2004; Inada et al., 2007; Duncan et al., 2008). Such an absence of NMDAR hypofunction contrasts with decreased NMDAR currents in GluN1KD MSNs (Jocoy et al., 2011), which represent more than 90% of neurons in the STR (Rymar et al., 2004). These conflicting results may point to limitations of using Fos to assess NMDAR activity, the remaining NMDARs being sufficient for striatal Fos induction, or effective compensatory mechanisms. Further studies are needed to more directly assess NMDAR activity and functionality in GluN1KD mice. As a whole, however, the current evidence confirms a global reduction of NMDAR quantity and function in the GluN1KD brain.

The GluN1KD hypomorphic mutation is an insertion of a neomycin resistance cassette into intron 19 of the Grin1 gene. GluN1KD mice were initially of mixed background, containing alleles from the 129/Ola, C57BL/6, and DBA/2 mouse lines. This background heterogeneity was controlled by tracking 11 microsatellite markers on 10 chromosomes (Mohn et al., 1999). In contrast, later studies used a defined genetic background to compare GluN1KD mice and WT

26 littermates (Fradley et al., 2005; Duncan et al., 2006a; Milenkovic et al., 2014). The defined genetic backgrounds of experimental animals have been F1 hybrids of C57BL/6J and either 129/Ola (Moy et al., 2008), 129/SvEv (Duncan et al., 2006a; Wesseling et al., 2014), or 129X1/SvJ (Mielnik et al., 2014; Milenkovic et al., 2014). Despite the differences in genetic background between some studies, findings have been largely consistent, particularly for commonly-studied GluN1KD behavioural deficits (Gandal et al., 2012; Milenkovic et al., 2014; Islam et al., 2017). To date, more than 30 studies have been published using GluN1KD mice; these include characterization studies (Mohn et al., 1999; Halene et al., 2009a; Gandal et al., 2012), drug discovery studies (Gregory et al., 2013; Grannan et al., 2016), and studies addressing other basic science questions (Moy et al., 2008; Jocoy et al., 2011). Our lab has worked on characterizing GluN1KD mice and using them to answer basic science questions related to NMDAR hypofunction (Mielnik et al., 2014; Milenkovic et al., 2014; Islam et al., 2017).

On a gross, macroscopic level, GluN1KD mice do not present with very overt physical or neurological abnormalities (Mohn et al., 1999; Miyamoto et al., 2004). They survive to adulthood (Mohn et al., 1999; Miyamoto et al., 2004; Ramsey et al., 2008) and have WT-like body temperature (Fradley et al., 2005), but are not considered fertile – GluN1KD males have seemingly normal testicular histology but severe deficits in reproduction-related behaviour (Mohn et al., 1999; Halene et al., 2009a; Gandal et al., 2012). The hypomorphic mutation can take a toll on the survivability and health of mice, however, as generation of GluN1KD mice using only one parental strain can result in less homozygous knockdown mice, stunted growth, and higher mortality (Duncan et al., 2006a; Ramsey et al., 2011; Wesseling et al., 2014). Studies using GluN1KD mice on a predominantly C57BL/6 background have reported reduced bodyweights, particularly in males, compared to WT littermates (Fradley et al., 2005; Moy et al., 2012) while others have noted reduced body size or weight even when using F1 GluN1KD hybrids (Mohn et al., 1999; Islam et al., 2017). Other signs of poorer health in GluN1KD mice include reports of self-injury by excessive grooming and scratching (Moy et al., 2008; Gandal et al., 2012), increased incidence of seizures and seizure-like electrical activity (Dzirasa et al., 2009; Halene et al., 2009a), poor pre-weaning survival (Fradley et al., 2005), and increased mortality in males when fed an ω-3 fatty acid deficient diet (Islam et al., 2017). Overall, however, the GluN1KD mice would generally appear healthy if small, unless the experimenter comes upon one of the minority that is coincidentally seizing or damaged from excessive

27 grooming. They have no major anatomical or neuro-anatomical abnormalities (Mohn et al., 1999; Miyamoto et al., 2004).

Several assessments of the sensory and motor systems of GluN1KD mice have been conducted, generally pointing to functional but abnormal systems (Bickel et al., 2007; Moy et al., 2008; Halene et al., 2009a). The main direct piece of evidence pointing to a functional visual system in GluN1KD mice comes from electrophysiological experiments. Investigators have detected and reported timely event-related potentials (ERP) after visual stimulation from a flash box. However, these ERPs had higher amplitudes compared to WT littermates, suggesting abnormal sensory processing (Halene et al., 2009a). Secondary evidence of a functioning visual system in GluN1KD mice comes from their ability to identify visual components of behavioural tests such as an open door in puzzle box tests (Milenkovic et al., 2014) and a visible platform in swim tests (Fradley et al., 2005). Auditory ERPs are much more commonly studied than visual ERPs in GluN1KD mice (Bickel et al., 2007, 2008; Bodarky et al., 2009; Halene et al., 2009a, 2009b). They are again present and timely but also generally greater in amplitude or longer lasting and less modulated by endogenous mechanisms compared to WT littermates (Bickel et al., 2007; Bodarky et al., 2009; Halene et al., 2009a; Gandal et al., 2012). In one of these experiments, GluN1KD test mice can even produce appropriate reductions in auditory ERP amplitude when presented with follow-up secondary auditory stimuli (Halene et al., 2009a). Behavioural secondary evidence also points to a functional auditory system in GluN1KD mice, particularly via the presence of timely, and even enhanced, acoustic startle (Fradley et al., 2005; Duncan et al., 2006b; Islam et al., 2017). Olfaction has been less directly assessed in GluN1KD mice, with the most direct experiment reporting appropriate auditory ERP modulation in the presence of fox urine, a predator odorant (Halene et al., 2009b). Acoustic startle is altered in GluN1KD mice exposed to soiled rat bedding, another predator odorant (Duncan et al., 2004b), while a study using an exploratory hole-board reported an abnormal preference for the scent of novel cage bedding in a particular GluN1KD male (Moy et al., 2008). Multiple studies have also reported normal social sniffing behaviour frequency in GluN1KD mice (Duncan et al., 2004b, 2009), but this does not hold for time spent on social investigation or sniffing (Mohn et al., 1999; Halene et al., 2009a; Moy et al., 2012). In general, these behavioural studies convey a working olfactory system whose signals are not always handled appropriately (Duncan et al., 2004b; Moy et al., 2008; Halene et al., 2009b). The senses of taste and touch, to our knowledge, have not

28 been directly assessed in GluN1KD mice. GluN1KD mice may show some preference for chocolate chips in the exploratory board task, but the data were not statistically significant and odor is a confounding factor (Moy et al., 2008). GluN1KD mice were quick to move away from a hot plate and displayed altered ERPs and motor responses during fear conditioning with foot shocks (Bickel et al., 2007), but pain is also not the same as general tactile sensation. Finally, balance and motor-coordination in GluN1KD mice have been reported to be WT-like when assessed by swimming and rotarod tests (Duncan et al., 2004b; Fradley et al., 2005). GluN1KD are therefore generally healthy, functional, and, as already implied by the above examples, suitable for behavioural testing.

1.2.2 GluN1KD Behavioural Abnormalities and Their Pharmacological Treatments

GluN1KD mice present with behavioural abnormalities that span multiple domains of cognition and are considered relevant to neuropsychiatric disorders (Mohn et al., 1999; Halene et al., 2009a; Gandal et al., 2012; Milenkovic et al., 2014). A summary of these abnormalities is listed in Table I-III. Our lab has previously reported deficits in locomotor activity, habituation, sensorimotor gating, stereotypy, social behaviour, and higher cognitive functions like learning and memory (Mielnik et al., 2014; Milenkovic et al., 2014; Islam et al., 2017). Others have reported the same or comparable deficits using similar behavioural testing methodology (Dzirasa et al., 2009; Halene et al., 2009a; Gandal et al., 2012; Grannan et al., 2016) as well as other abnormalities in behavioural correlates of anxiety and motivation (Duncan et al., 2004b; Moy et al., 2008; Halene et al., 2009a). The only behavioural study that tested GluN1KD-heterozygotes generally pointed to the Grin1 gene being haplosufficient (Moy et al., 2012). The GluN1KD mouse model and its reported behavioural deficits in homozygous knockdown mice have been considered relevant in particular to schizophrenia and ASDs (Mohn et al., 1999; Halene et al., 2009a; Gandal et al., 2012). Locomotor activity and sensorimotor gating, for example, may be considered behavioural correlates for the positive symptoms of schizophrenia (Fradley et al., 2005; Hazlett et al., 2007). Furthermore, the relatively rarer studies of sex and age differences in GluN1KD behavioural phenotypes also support a connection to schizophrenia and autism (Moy et al., 2012; Milenkovic et al., 2014). GluN1KD behavioural deficits can be treated by currently approved antipsychotics (Mohn et al., 1999; Duncan et al., 2006a) as well as experimental therapies for schizophrenia that target related molecular signaling mechanisms (Gregory et al.,

2013; Grannan et al., 2016). Overall, GluN1KD mice present a model of NMDAR hypofunction that is relevant and useful in a variety of behavioural and neuropsychiatric contexts.

Table I-III. Major behavioural abnormalities found in previous GluN1KD studies. GluN1KD mice present with behavioural abnormalities across various domains of behaviour and cognition. Such a diversity of irregular behaviours resemble the multiple types of schizophrenia symptoms (positive, negative, cognitive).

Behaviour domain or type Behavioural test paradigm Test measurement GluN1KD abnormality

Elevated plus maze; open Anxiety-like behaviour Time in open spaces Increased field locomotion Memory; cue/context learning Fear conditioning Conditioned freezing Decreased 8-arm radial, Y-, and T- Maze-arm exploration Memory; exploratory behaviour Decreased mazes pattern Memory; exploratory behaviour 8-arm radial maze 8-arm entries-to-repeat Decreased Memory; exploratory behaviour Novel object recognition Investigation time Decreased Memory; puzzle solving Puzzle box Puzzle box solution time Decreased

Conditioned place Motivation-based conditioning Place-preference Mixed abnormalities preference Horizontal locomotor Increased activity; Motor activity; habituation Open field locomotion activity Decreased habituation Stereotypic fine Motor activity; stereotypy Open field locomotion Increased movements Sensorimotor gating Prepulse inhibition Prepulse inhibition Decreased Sensorimotor response Prepulse inhibition Acoustic startle Increased Social approach; resident- Social cognition Social investigation time Mixed abnormalities intruder Aggressive behaviour Social cognition Resident-intruder Decreased count

Elevated locomotor activity and deficits in sensorimotor gating are considered particularly relevant to the positive symptoms of schizophrenia (Fradley et al., 2005; Hazlett et al., 2007). Locomotor activity and its decrease over time as a result of habituation are assessed by open field tests where horizontal mouse movements are tracked by motion sensors or video recording (Duncan et al., 2006a; Halene et al., 2009a; Milenkovic et al., 2014). The most consistent finding regarding GluN1KD locomotor abnormalities is the delayed habituation of GluN1KD mice in novel environments. In other words, GluN1KD mice move around exploring new cages or arenas longer than WT-littermates (Mohn et al., 1999; Duncan et al., 2006a, 2010; Gregory et al., 2013). This often means a net increase in overall travelled distance over the course of the experiment (Mohn et al., 1999; Duncan et al., 2006a; Milenkovic et al., 2014). To note, closer examination

30 of the tracked time-courses do not always show GluN1KD mice moving more at the beginning of open field tests, reinforcing the idea of a specific GluN1KD deficit of habituation (Duncan et al., 2002, 2006a, 2012; Miyamoto et al., 2004; Moy et al., 2012). The GluN1KD mice do habituate to WT levels of movement after 4 hours (Mohn et al., 1999), but they have also been reported to travel more in home cages during light cycle hours (Gandal et al., 2012). Finally, there are still other reports of WT-like movement by GluN1KD mice in home cages and arenas of other behavioural tests, which the authors believe may be due to the size of experimental spaces (Halene et al., 2009a; Gandal et al., 2012). In general, GluN1KD mice display hyperlocomotion and decreased habituation on a consistent basis.

Impaired sensorimotor gating is another behavoural change that is consistently observed in GluN1KD mice across multiple labs. Sensorimotor gating is assessed in the prepulse inhibition (PPI) paradigm wherein mice are placed in chambers that measure the auditory startle response (how much a mouse jumps when exposed to loud auditory stimuli). The startle-inducing stimuli may vary in intensity and are sometimes randomly preceded by a less intense auditory stimulus (the prepulse). How much each test mouse jumps after the main auditory stimuli is recorded as acoustic startle responses while how much this startle is decreased in the presence of a prepulse is recorded as PPI % (Moy et al., 2006; Bickel et al., 2008; Islam et al., 2017). To date, GluN1KD mice have consistently presented with increased acoustic startle and decreased PPI (Duncan et al., 2004b, 2010; Fradley et al., 2005; Bickel et al., 2008; Gandal et al., 2012; Islam et al., 2017). This finding has been replicated electrophysiologically as the same GluN1KD mice with PPI impairments also demonstrated impaired attenuation of second-tone ERPs when pairs of auditory stimuli were presented. However, the link between sensory gating of ERPs and PPI is not robust in the literature (Bickel et al., 2008). This specific electro-physiological finding has also not been specifically replicated, though other such experiments in GluN1KD mice still reported auditory ERP abnormalities. The authors of these studies attributed the discrepancies in findings to differences in the interval between auditory stimuli (Halene et al., 2009a; Gandal et al., 2012). Interestingly, habituation was also impaired in GluN1KD mice when PPI experiments were repeated with the same test mice (Bickel et al., 2008). GluN1KD deficits in PPI, locomotor activity, and the habituation of both thus paint an overall picture of neuronal dysregulation and information misprocessing (Halene et al., 2009a; Gandal et al., 2012; Grannan et al., 2016). Fittingly, neuronal disinhibition and signal processing deficits are thought to be neurological

31 underpinnings of the positive symptoms of schizophrenia (Gilani et al., 2014; Litman et al., 2016; Stevenson et al., 2017). This is again consistent with both PPI deficits and hyperlocomotion being behavioural correlates of positive schizophrenia symptoms (Fradley et al., 2005; Hazlett et al., 2007).

Studies on how antipsychotics affect GluN1KD mouse behaviour has focused on the open field and PPI paradigms. Enhanced startle response in GluN1KD mice has been shown to be attenuated by risperidone, olanzapine, and quetiapine while impaired PPI was improved by haloperidol, risperidone, olanzapine and quetiapine (Duncan et al., 2006a, 2006b). Acoustic startle and PPI abnormalities were also improved with administration of 3 mg/kg clozapine to GluN1KD mice that were C57BL/6 x 129/SvEv hybrids (Duncan et al., 2006b) but not with 1 mg/kg clozapine to mice on a mostly C57BL/6 genetic background (Fradley et al., 2005). GluN1KD hyperlocomotion was also improved with administration of haloperidol, olanzapine, and clozapine (Mohn et al., 1999; Duncan et al., 2006a). It is important to note, however, that these antipsychotics sometimes also decrease locomotor activity, decrease startle response, and increase PPI in WT littermates (Duncan et al., 2006a, 2006b). Thus, while GluN1KD locomotor and PPI abnormalities, related electrophysiology, and their improvement with antipsychotics present a consistent picture for modeling positive symptoms, more research is needed to investigate how these correlates specifically relate to GluN1KD mice. For example, blocking the kainate subtype of glutamate receptors in GluN1KD mice can also improve their locomotor, acoustic startle, and PPI behaviours (Duncan et al., 2010, 2012). However, only PPI is increased in WT littermates when LY382884 was used for kainate receptor antagonism (Duncan et al., 2010). It would therefore be interesting to further investigate glutamate or kainate receptor signaling in GluN1KD mice, along with other instances where responses differ between GluN1KD mice and WT littermates.

GluN1KD mice also consistently show increased stereotypic movements and impaired social behaviour (Mohn et al., 1999; Moy et al., 2008; Gandal et al., 2012; Milenkovic et al., 2014). Both behavioural abnormalities have been considered relevant to the core symptoms of autism (Moy et al., 2008; Gandal et al., 2012) and schizophrenia, with stereotypic behaviours being a positive symptom (Fradley et al., 2005) and reduced sociability being a negative symptom (Cox et al., 2016). Increased stereotypy in GluN1KD mice has been demonstrated in three experimental methodologies: concurrent tracking of fine repetitive movements during the open

32 field test (Mohn et al., 1999; Miyamoto et al., 2004; Milenkovic et al., 2014), video recording of mouse behaviour while suspended by their tails (Gandal et al., 2012), and manual tracking of grooming behaviour across experiments (Moy et al., 2008). When reported, GluN1KD mice consistently show increased total repetitive fine movements in open field tests (Mohn et al., 1999; Milenkovic et al., 2014; Islam et al., 2017). One study also found this specifically with rearing movements (Duncan et al., 2010). Further observation of time-courses again reveals decreased habituation of stereotypic movements in GluN1KD mice (Mohn et al., 1999; Miyamoto et al., 2004; Milenkovic et al., 2014; Islam et al., 2017). An instance of WT-like stereotypic movement levels in GluN1KD mice at the beginning of open field tests has been shown (Miyamoto et al., 2004), but such instances are rarer than what is seen for horizontal locomotor activity. This may be due to stereotypic movement time-courses being less reported. Also unlike horizontal locomotor activity, increased repetitive behaviour has been more consistently replicated in other behavioural tests. More GluN1KD mice demonstrated handclasping behaviour, where mice repeatedly clasped and shook their forepaws in front of their faces, during tail suspension tests than WT littermates (Gandal et al., 2012). Furthermore, self-injurious repetitive grooming has been documented in GluN1KD mice with an incidence of 25-30%, abnormally high compared to WT littermates (Moy et al., 2008; Gandal et al., 2012). While open field stereotypic behaviour has been shown to be treatable with haloperidol and clozapine (Mohn et al., 1999), consistent with increased stereotypy being a correlate of positive schizophrenia symptoms, GluN1KD handclasping and excessive grooming have not been assessed pharmacologically to date. Non-antipsychotic medications for autism, such as methylphenidate or selective serotonin reuptake inhibitors (LeClerc and Easley, 2015), have not been tested on GluN1KD mice. The kainate receptor antagonist LY382884, which improves GluN1KD horizontal locomotor activity, also attenuates GluN1KD stereotypic movements.

Multiple forms of social behaviour abnormalities have been demonstrated in GluN1KD mice (Mohn et al., 1999; Duncan et al., 2009; Mielnik et al., 2014). From the beginning, the first studied and reported GluN1KD mice were noted as having increased social withdrawal and impaired mating (Mohn et al., 1999). Mating behaviour, as a subset of social behaviour, has been primarily demonstrated in GluN1KD males via decreased mating frequency (Mohn et al., 1999) and ultrasonic vocalizations in the presence of a receptive female (Gandal et al., 2012). However, the majority of investigations into GluN1KD social behaviour did not focus on mating behaviour

33 but instead focused on social approach tests and resident-intruder assays (Mohn et al., 1999; Duncan et al., 2009; Mielnik et al., 2014; Milenkovic et al., 2014). The social approach test measures how long mice spend investigating novel mice, which are often matched for age and sex or are gonadectomized, relative to a control setup like an empty wire cage (Halene et al., 2009a; Milenkovic et al., 2014). When assessed by this method, GluN1KD mice consistently spent less time investigating the presented social stimulus mouse compared to WT littermates (Duncan et al., 2004b; Halene et al., 2009a; Mielnik et al., 2014). Less time spent on social investigation has also been demonstrated for male GluN1KD mice in the resident-intruder paradigm, where mouse behaviour is monitored after a genotype- and sex-matched intruder mouse is introduced to the home cage of a resident mouse (Mohn et al., 1999). Another way to perform the resident-intruder assay is to more specifically score mouse social interactions. When assessed this way, investigators reported that male and female GluN1KD mice perform less acts of aggression and dominance compared to WT littermates; male mice fight less while female mice mount less, but neither show decreased frequency of investigative social sniffing (Duncan et al., 2004b, 2009). Concurrent Fos induction studies have associated social approach deficits with decreased Fos at the cingulate CTX and lateral septal nuclei (LSN) (Mielnik et al., 2014) and resident-intruder deficits with decreased Fos at the same regions along with others like the hypothalamus and medial preoptic area (Duncan et al., 2009). Both the cingulate CTX and LSN have fittingly been associated with social cognitive function in animals and humans (Apps et al., 2016; Lockwood and Wittmann, 2018; Rilling et al., 2018; Weger et al., 2018), supporting the use of GluN1KD mice to model neuropsychiatric social impairments. Social behaviour pharmacology has been mixed in GluN1KD mice: 0.5 mg/kg clozapine improved resident- intruder behaviour and mating frequency in initially-generated mixed-background GluN1KD mice (Mohn et al., 1999) but 1 mg/kg clozapine did not improve social approach behaviour in C57BL/6J x 129X1Sv/J GluN1KD mice (Mielnik et al., 2014). While this is an inconsistency, it also reflects clinical trials of clozapine in patients with schizophrenia; clozapine had little or no benefits for treating the negative symptoms of schizophrenia (Breier et al., 1994; Buchanan et al., 1998). Non-antipsychotic medications for ASDs have not been tested on GluN1KD social behaviours. Overall, the evidence of social behaviour deficits in GluN1KD mice is robust and may be particularly useful as a model of the negative symptoms of schizophrenia.

There is also convincing evidence of cognitive impairments in GluN1KD mice, but the experiments involved are more varied compared to previously discussed behavioural abnormalities. Through these various assessment methods, GluN1KD mice consistently demonstrated impairments in learning and memory, puzzle solving, and attention (Dzirasa et al., 2009; Milenkovic et al., 2014; Grannan et al., 2016). In the Y-maze and T-maze tests, where mice must remember most-recently visited maze arms to alternately visit each of three arms consecutively (Deacon and Rawlins, 2006; Mandillo et al., 2008), GluN1KD mice consistently had impaired alternation performance (Gandal et al., 2012; Gregory et al., 2013; Milenkovic et al., 2014). In the similar but more complicated 8-arm radial maze, where mice learn to efficiently find food rewards at the end of each arm by strategically and quickly visiting each arm once, GluN1KD mice failed to learn over weeks of repeated trials. Instead, they readily and persistently explore already-visited maze arms (Dzirasa et al., 2009). GluN1KD mice also demonstrated impaired learning of obstacle solutions in the puzzle box test, as well as an impaired ability to bypass obstacles in the first place (Milenkovic et al., 2014; Islam et al., 2017). Exposure to the puzzle box test as a juvenile 3-week old can improve GluN1KD performance 4 weeks later when compared to naïve 6-week old GluN1KD mice, but this was not a complete behavioural rescue (Milenkovic et al., 2014). Finally, in the fear conditioning paradigm, where mice are expected to decrease their movements when exposed to a context or cue they have learned to associate with a foot shock, GluN1KD mice showed a lack of freezing behaviour (Grannan et al., 2016). As mentioned before, altered electrophysiological responses to fear conditioning cues have been reported, but these responses were abnormal and considered an impairment of attention (Bickel et al., 2007). Abnormal modulation of ERPs in GluN1KD brains when exposed to various types of stimuli (Bodarky et al., 2009; Halene et al., 2009b) can similarly be considered forms of attentional deficits, as can previously mentioned sensorimotor gating (Moy et al., 2006), social approach (Moy et al., 2012), and habituation (Bickel et al., 2008). The common concept between these different behavioural and electrophysiological measures is the inability of a nervous system to distinguish between important or novel stimuli from background environmental noise (Bickel et al., 2007, 2008; Bodarky et al., 2009; Halene et al., 2009b; Moy et al., 2012). To date, the main report of pharmacological rescue of GluN1KD cognitive deficits, aside from PPI and habituation improvements with antipsychotics, is the improvement of cue-conditioned freezing with VU6004256. VU6004256 is a potent positive allosteric modulator of the M1 muscarinic acetylcholine receptor and a drug being evaluated for

35 treating positive and cognitive symptoms of schizophrenia (Grannan et al., 2016). Cholinesterase inhibitors, the foremost class of drugs for treating cognitive impairments and dementia (Gauthier et al., 2006; Deardorff et al., 2015), have not been tested in GluN1KD mice. Memantine, the main approved non-cholinesterase-inhibitor drug for AD (Tariot et al., 2004; Deardorff et al., 2015), may not be suitable to try in GluN1KD mice due to its function as a NMDAR antagonist (Parsons et al., 2007). The NMDAR partial agonist D-cycloserine may be interesting to test in GluN1KD mice, however, as it has been reported, though rarely, to improve cognitive outcomes in human patients (Tsai et al., 1999; Goff et al., 2008). The data overall point to a rescuable cognitive deficit in GluN1KD mice, but more experiments are needed to study it in detail.

The behavioural deficits mentioned above are the major groups of findings from behavioural studies of GluN1KD mice, but there are a few other odd reports as well. GluN1KD mice have been reported to spend more time in the open arms of an elevated zero maze and the center of open field arenas, indicative of decreased anxiety-related inhibition (Halene et al., 2009a). In behavioural test paradigms involving olfactory appetitive stimuli, mainly chocolate chips, GluN1KD mice can detect and obtain them but do not demonstrate a strong preference for them when compared to WT littermates (Duncan et al., 2004b; Moy et al., 2008). For one of these studies, the authors offered an interpretation of possible motivational deficits in GluN1KD mice, at least in the context of desire for a chocolate chip (Duncan et al., 2004b). In contrast, GluN1KD mice can perform like WT littermates in the conditioned place preference behavioural paradigm; in response to repeated injections of amphetamine, GluN1KD mice demonstrate a preference for being in chambers were the injections took place when given the choice. When more closely examined with cocaine conditioning, however, GluN1KD mice developed the place preference more slowly compared to WT mice. This may be another indication of cognitive or motivational deficits, but cocaine and amphetamine also elicit multiple different responses in GluN1KD mice (Ramsey et al., 2008). Amphetamine generally produces WT-like behavioural responses in GluN1KD mice at lower doses and enhanced or altered responses at higher doses while cocaine is less often studied (Miyamoto et al., 2004; Moy et al., 2006; Ramsey et al., 2008). How these findings correlate to motivation is unclear, but it will be useful to keep them in mind when considering dopamine (DA) system changes in GluN1KD mice later. Finally, normal rotarod performance in GluN1KD mice was mentioned previously, indicative of a functional motor nervous system (Duncan et al., 2004b; Fradley et al., 2005). One of the studies employing

36 the rotarod test conducted multiple trials and demonstrated that GluN1KD mice, like their WT littermates, were able to stay on rotarods longer with repeat trials. The study did not report a statistically significant effect of trial repetition on rotarod performance for WT or GluN1KD mice, but it also did not explicitly state whether such an effect was significantly detected and the experiment design was susceptible to ceiling effects. If both WT and GluN1KD mice significantly improved their rotarod performance with repeated trials, such a result may indicate that GluN1KD mice are not impaired in learning primarily motor tasks (Fradley et al., 2005). No pharmacological interventions have been tested or reported for these behavioural paradigms in GluN1KD mice. Even more so than the established behavioural abnormalities mentioned earlier, GluN1KD deficits related to anxiety, motivation, and motor skill learning need to be replicated and investigated further before any robust interpretations or conclusions can be made.

Gene dose-, sex-, and age-based differences in GluN1KD behavioural abnormalities also need further investigation, as studies that account for these factors are relatively rare. The studies that have been completed report Grin1 haplosufficiency, worse abnormalities in males, and nuanced results on the effects of age (Moy et al., 2006, 2012; Milenkovic et al., 2014). To date, only one published study includes behavioural assessments of GluN1KD-heterozygotes. These heterozygotes largely did not behave differently compared to WT littermates in all the behavioural tests assessed: social approach, PPI, and open field. The only reported difference was that heterozygous males spent more time sniffing a novel mouse than an empty cage at 34- 38 days of age while the time-spent difference in WT males was in the same direction but not statistically significant (Moy et al., 2012). This is consistent with the main molecular characterization of GluN1KD-heterozygotes: NMDAR binding was unchanged in their brains relative to WT littermates (Duncan et al., 2002). In all, there is no evidence of a gene dose effect for the neomycin insertion mutation into Grin1.

A few more GluN1KD behavioural studies have considered sex as a factor in their study design or at least their statistical analyses. The results from these studies often pointed to greater abnormalities in males (Moy et al., 2006, 2012; Milenkovic et al., 2014). Such a pattern may not be surprising, as GluN1KD males may be more likely to have smaller body sizes (Moy et al., 2012) and die on an ω -3 deficient diet (Islam et al., 2017). For example, deprivation of ω-3 fatty acids in GluN1KD mice also produces a male-specific exacerbation of puzzle box deficits (Islam et al., 2017). In the PPI test paradigm study, GluN1KD males have been shown to have more

37 enhanced acoustic startle responses than females (Moy et al., 2006). They also have different acoustic startle and PPI changes in response to exposure to predatory odor compared to GluN1KD females and WT mice of both sexes (Duncan et al., 2004b). Only male GluN1KD mice displayed significant hyperlocomotion at 12 days of age, increased open-field center time at 18-20 days of age, and impaired PPI at 13 and 17 days of age in one wide-ranging study of how GluN1KD behaviours change with sex and age (Moy et al., 2012). In a similarly thorough study from our lab, we found that GluN1KD males had greater total horizontal locomotor activity and stereotypy measures as well as lower social-approach investigation time and Y-maze alternation percentage than GluN1KD females at 6 weeks of age. Furthermore, we found that 12-week old GluN1KD males had a significantly worse Y-maze alternation score than WT males while the deficit in same-aged GluN1KD females was not significant (Milenkovic et al., 2014). One piece of potentially conflicting evidence is the finding that GluN1KD males preferentially investigated novel social stimulus mice at 34-38 days of age when GluN1KD females and WT mice of both sexes did not (Moy et al., 2012). However, this can be interpreted as a GluN1KD abnormality only manifesting in males rather than the GluN1KD males demonstrating an advantageous behaviour compared to GluN1KD females. Overall, most if not all reports of sex-GluN1KD interaction effects point to the males presenting with a more severe phenotype. This increased severity is highly age-dependent, as most sex differences presented above disappear by 12 weeks of age (Moy et al., 2012; Milenkovic et al., 2014).

The overall effect of sex and age on GluN1KD behavioural phenotypes can be summarized as abnormality manifestation or exacerbation with age along with earlier and more severe phenotypes in GluN1KD males (Moy et al., 2012; Milenkovic et al., 2014). Particularly in the study from our lab, the results seem to reflect earlier symptom presentation and worse outcomes in male schizophrenia patients (Eranti et al., 2013; Milenkovic et al., 2014; Thara and Kamath, 2015). However, there are some nuances presented by the two comprehensive studies discussed. In particular, while Y-maze and social approach data from our lab reflect schizophrenia quite well, other behaviours do not (Moy et al., 2012; Milenkovic et al., 2014). Multiple GluN1KD behavioural abnormalities are already robustly present in the youngest tested GluN1KD males and females: acoustic startle in 13-day old mice (Moy et al., 2012) and stereotypic movement count and puzzle box performance in 3-week old mice (Milenkovic et al., 2014). However, PPI was only impaired in 13-day old GluN1KD males with certain prepulse decibels and the amount

38 of time spent on stereotypic movements was not significantly different in 3-week old GluN1KD mice compared to WT littermates (Moy et al., 2012; Milenkovic et al., 2014). Such nuanced abnormalities need further investigations for meaningful interpretations, but the enhanced startle and stereotypy counts could still be indicative of abnormal behaviours manifesting in the early infancy of GluN1KD mice. Indeed, infant GluN1KD mice of at least one sex also demonstrate significantly increased horizontal locomotor activity (Moy et al., 2012; Milenkovic et al., 2014), open field center time, and social approach behaviour (Moy et al., 2012). This is inconsistent with the early-adulthood or later onset of schizophrenia in humans (Sham et al., 1994; Thorup et al., 2007; Eranti et al., 2013), but the behavioural abnormalities may be more representative of the childhood onset seen in autism (Baxter et al., 2015). This is also not a complete resolution, however, as the age-of-onset and age trajectories in autism patients are similar between the sexes. There are also reports of greater symptom severity in male autism patients, but this is not firmly established due to a potential underdiagnosis of female patients and their subsequent underrepresentation in research studies (Baxter et al., 2015; Halladay et al., 2015). There is a higher prevalence of autism in males, however (Baxter et al., 2015). The results paint a complex picture overall, but the early-onset and late-onset groups of GluN1KD behavioural phenotypes may support the idea of using GluN1KD mice to model autism and schizophrenia, respectively.

Similarly complex are the specific effects of age on GluN1KD social approach time and horizontal locomotor activity. Recall for social approach behaviour that 3-week old GluN1KD mice behaved like WT littermates, 5-week old males were more social, 6-week old males were less social, and GluN1KD females were only significantly different (less social) after adulthood. For horizontal locomotor activity, 2-week old GluN1KD males were hyperactive, 3-week old mice of both sexes were hyperactive, and this hyperactivity was then persistent but decreased significantly by 6- and 12-weeks of age (Moy et al., 2012; Milenkovic et al., 2014). The decreased horizontal locomotor activity with age may be due to the mice spending more time engaged in stereotypic behaviour, but more research is needed to investigate this explicitly (Milenkovic et al., 2014). Overall, such patterns need to be replicated in GluN1KD mice of one genetic background using one consistent methodology. And if replicated, future research into the relationship between NMDAR deficiency and behaviour may benefit from targeting the neurodevelopmental windows presented by these results. Male GluN1KD brains between 5 and 6 weeks of age, for example, may be undergoing neurodevelopmental changes very important for

39 social behaviour. Unfortunately, no other study to date has assessed the effects of age on GluN1KD behaviour. This also means no studies to date have investigated how drug effects might vary in GluN1KD mice of different ages. The current results of age studies using GluN1KD mice point to most behavioural abnormalities being present by adulthood and persisting or worsening with age. Some of these phenotypes start manifesting during adolescence (5-6 weeks of age) while other are already present before weaning (3 weeks or younger). The results also currently point to a strong interaction between sex and age effects on GluN1KD behaviours.

There is a robust body of evidence demonstrating a variety of behavioural abnormalities in GluN1KD mice. These abnormalities have been considered relevant to positive, negative, and cognitive symptoms of schizophrenia as well as symptoms seen in ASDs (Mohn et al., 1999; Halene et al., 2009a; Gandal et al., 2012; Moy et al., 2012; Milenkovic et al., 2014). The relatively rarer GluN1KD studies that consider the effects of age and sex on behavioural phenotypes offer some trends that may reflect both schizophrenia (Eranti et al., 2013; Milenkovic et al., 2014; Thara and Kamath, 2015) and autism (Moy et al., 2012; Baxter et al., 2015). Many of the GluN1KD behavioural phenotypes have been improved with antipsychotic treatment (Mohn et al., 1999; Duncan et al., 2006a, 2006b) while studies with other approved medications for neurological disorders are few (Duncan et al., 2006b; Mielnik et al., 2014). Studies with drugs that act on molecular systems known to interact with NMDAR functions have also reported improvements of GluN1KD phenotypes (Duncan et al., 2010; Gregory et al., 2013; Grannan et al., 2016). However, there is still a lack of molecular experiments confirming these drugs’ mechanisms of action in GluN1KD mice (Duncan et al., 2010; Gregory et al., 2013). Investigating the effects of other classes of drugs on GluN1KD mice will be important in understanding the relationship between GluN1KD phenotypes and the human symptoms they likely model, as well as how the test drug’s mechanism of action is involved. To facilitate an improved understanding of this last relationship, studies characterizing GluN1KD cellular and molecular phenotypes are important. Such studies will also help inform the decision and prioritization of what drugs to test next in GluN1KD mice.

1.2.3 Cellular and Molecular Changes in GluN1KD Mice

Investigations into the cellular and molecular alterations of GluN1KD mice mainly focus on the brain, as illustrated in Figure 1.4. They are fewer than behavioural studies, though there are overlaps; two such overlaps were mentioned with GluN1KD behavioural abnormalities. Impaired secondary auditory ERPs were detected in GluN1KD mice PPI deficits (Bickel et al., 2008) and decreased Fos induction at distinct brain regions was detected during social behaviour test paradigms (Duncan et al., 2009; Mielnik et al., 2014). Other studies focus purely on cellular and molecular changes; they are primarily characterization studies (Duncan et al., 2002; Wesseling et al., 2014; Cox et al., 2016). Studies that account for sex, and age differences in GluN1KD mice are again few but very useful, especially when considered in the context of results from behavioural studies (Duncan et al., 2009; Ramsey et al., 2011; Islam et al., 2017). Other than WT-like NMDAR binding in the brains of GluN1KD-heterozygotes (Duncan et al., 2002), no other cellular or molecular study published to date has investigated gene dose effects. Overall, the results from these non-behaviour-focused GluN1KD studies are disconnected and focused on their individual research contexts, but a few general trends can be discerned. While there is evidence of hypofunctional GluN1KD brains, particularly and expectedly with NMDAR- specific measures (Jocoy et al., 2011), there are also striking reports of neuronal hyperexcitability (Jocoy et al., 2011; Ferris et al., 2014). These two conflicting types of findings are likely related via inappropriate disinhibition and altered neuronal circuits in GluN1KD mice (Ferris et al., 2014; Grannan et al., 2016). Fittingly, of all the biological processes regulated by NMDARs, synaptic connectivity and plasticity may be particularly and importantly disrupted in GluN1KD mice (Ramsey et al., 2011; Wesseling et al., 2014). These trends may be unsurprising due to previously discussed behavioural abnormalities and may indeed also benefit from being considered with behavioural test results in mind (Halene et al., 2009b; Ramsey et al., 2011).

Figure 1.4. Major cellular and molecular changes reported in previous GluN1KD studies. There is evidence for both increased and decreased neuronal activity in the GluN1KD brain. Such conflicting or complex alterations may be due to more general changes of impaired synaptic plasticity. The most studied brain region in GluN1KD mice to date is the STR, followed by the CTX and HPC.

Several pieces of evidence point to hypofunctionality in GluN1KD brains (Duncan et al., 2002, 2008; Jocoy et al., 2011). NMDAR hypofunction, as mentioned before, has been demonstrated directly in striatal MSNs (Jocoy et al., 2011). Uptake of 2-deoxyglucose (2-DG) and induction of Fos expression, both commonly used markers of brain activity (Duncan et al., 1993; Lundgaard et al., 2015; Renier et al., 2016), are generally muted in GluN1KD mice (Duncan et al., 2002, 2008; Inada et al., 2007; Mielnik et al., 2014). Basally, multiple layers and nuclei in the GluN1KD CTX, STR, and other brain regions show reduced 2-DG uptake compared to WT littermates. No assessed region displayed increased 2-DG uptake in GluN1KD mice (Duncan et al., 2002). Compared to WT littermates, GluN1KD mice have demonstrated decreased Fos induction in various brain regions after pharmacological NMDAR agonism with TZG (Inada et al., 2007; Duncan et al., 2008), amphetamine administration (Miyamoto et al., 2004), cocaine administration (Ramsey et al., 2008), and social behaviour testing (Duncan et al., 2009; Mielnik et al., 2014). There were also no assessed brain regions in any of these studies that showed increased Fos induction in GluN1KD mice compared to WT littermates. As discussed

42 before, these differences in Fos induction might be of vague biological significance, such as when WT-like levels of Fos at the GluN1KD STR is considered with impaired 2-DG uptake and MSN EPSPs (Duncan et al., 2002, 2009; Miyamoto et al., 2004; Jocoy et al., 2011). This may be in part due to a brain region resolution issue, as there is one report of differential GluN1KD Fos- induction responses between dorsal and ventral STR (Ramsey et al., 2008). But Fos induction in GluN1KD LSN may serve as a robust counter-example of when the indirect neuronal activity marker works as intended: two independent studies confirmed GluN1KD Fos induction deficits after two different social behaviour test paradigms (Duncan et al., 2009; Mielnik et al., 2014) and an increased Fos induction coincided with improved social behaviour in GluN1KD mice (Mielnik et al., 2014). Overall, these studies present many indicators of decreased activity in the GluN1KD brain. However, one study did report increased Fos induction in GluN1KD mice after kainic acid administration. The authors also reported GluN1KD behavioural improvements in the open field and PPI tests with systemic administration of a kainate receptor antagonist (Duncan et al., 2010). This is one example of evidence for a hyperglutamatergic state in GluN1KD mice.

There are several pieces of evidence for a hyperexcitable or hyperglutamatergic brain in GluN1KD mice. One of the most overt pieces is the noticeable incidence of seizures and seizure- like brain activity in GluN1KD mice (Dzirasa et al., 2009; Halene et al., 2009a). Furthermore, 20 mg/kg kainic acid was reported to reliably induce lethal seizures in GluN1KD mice but not WT littermates (Duncan et al., 2010). But still consistent with the idea of NMDAR hypofunction in GluN1KD mice, they were not more susceptible than WT mice to seizures induced by 1.25-1.5 mg/kg TZG (Duncan et al., 2008). Another blatant sign of hyperexcitability in GluN1KD brains is the enhanced AMPAR-mediated EPSPs in MSNs – the same type of neurons that demonstrated inhibited NMDAR-mediated responses (Jocoy et al., 2011). To note, no changes of AMPA or kainate receptors were detected in any brain region of GluN1KD mice by quantitative autoradiography (Duncan et al., 2002). AMPAR subunit protein levels were also unchanged when assessed by western blotting while GluN2A and GluN2B were decreased in the GluN1KD STR, HPC, and prefrontal CTX (Ramsey et al., 2008). This would imply that glutamatergic activity or excitability increases in GluN1KD mice may well be mediated by alterations in function or trafficking of AMPA or kainate receptors – the two main non-NMDA ionotropic glutamate receptors (Duncan et al., 2002; Fradley et al., 2005). There are also the ERP studies that have been mentioned and discussed multiple times now. GluN1KD ERPs to both visual and

43 auditory stimuli are enhanced and less inhibited (Bickel et al., 2008; Halene et al., 2009b, 2009a; Gandal et al., 2012), consistent with increased excitability and disinhibition at the brain and somewhat reinforced by the reports of GluN1KD PPI deficits (Bickel et al., 2008). Also of behavioural relevance, GluN1KD mice have an absence of muscarinic-receptor induced LTD and an increase of prefrontal CTX pyramidal neuron firing. Normalization of these measures was related to GluN1KD cognitive improvements after administration of VU6004256 (Grannan et al., 2016). Overall, studies using electroencephalography (EEG) and electrophysiology consistently demonstrate increased brain activity in GluN1KD mice, which is also reflected by seizures and behavioural changes.

Aside from general or glutamatergic electrical activity in the brain, the DA system in GluN1KD mice may serve as another example of hyperactivity. Like AMPA and kainate receptors in GluN1KD mice, many measures of DA system homeostasis were reported as unchanged compared to WT littermates (Mohn et al., 1999; Duncan et al., 2002; Ramsey et al., 2008). From their initial characterization, GluN1KD mice had WT-like levels of DA and its metabolites 3,4-dihydroxyphenylacetic acid and homovanillic acid at the STR (Mohn et al., 1999). This was true for both tissue content and extracellular levels and was later reconfirmed in a stimulant challenge study where DA release was measured in vivo by microdialysis (Mohn et al., 1999; Ramsey et al., 2008). GluN1KD mice also had WT-like numbers of MSNs along with unperturbed protein levels for DA transporters (DAT), D1 receptors (D1Rs), and D2 receptors

(D2Rs). The DA receptors were also determined to be functional in vivo when direct agonism resulted in horizontal locomotor activity changes similar to what is seen in WT mice (Ramsey et al., 2008). The first report of a dopaminergic disruption in GluN1KD mice was the finding that

DA enhancement of NMDAR currents in MSNs via D1Rs was weaker, but inhibition via D2Rs was not (Jocoy et al., 2011). Furthermore, when electrically-evoked DA release was measured in brain slices by fast-scan cyclic voltammetry, GluN1KD mice demonstrated deficits at both the nucleus accumbens (NAc) and caudate. Such a finding was reported alongside decreased synaptic tyrosine hydroxylase (TH) protein levels and function, meaning less DA synthesis, as well as increased synaptic DAT protein and function, meaning more clearance of extracellular DA (Ferris et al., 2014). These results point to decreased DA signaling, but further experiments also revealed impaired D2R-mediated autoinhibition of DA release and higher tonic firing rate of ventral tegmental area (VTA) neurons, which release DA (Taber et al., 2002; Ferris et al., 2014).

VTA neurons in GluN1KD mice were also more excitable but less likely to produce burst-firing patterns, similar to previous results of pharmacological NMDAR blockade on in vivo rat cortical neurons (Jackson et al., 2004; Ferris et al., 2014). The authors of this study ultimately interpreted these findings as GluN1 knockdown leading to chronically hyperactive VTA neurons, desensitized D2 autoreceptors, and an ultimately altered DA system with decreased transmission (Ferris et al., 2014). Overall, studies of the GluN1KD DA system found subtle changes pointing to a slightly hypodopaminergic state resulting from chronic hyperactivity.

Behavioural changes in GluN1KD mice after administration of DA-system-targeting and related drugs can further help understand their dopaminergic state. When treated with amphetamine, which blocks DAT and increases DA release, GluN1KD open field and PPI responses were WT-like at low doses but enhanced or altered at higher doses (Miyamoto et al., 2004; Moy et al., 2006; Ramsey et al., 2008). Cocaine, which affects the DA system primarily by blocking DAT, produced more muted locomotor activity changes and delayed place-preference conditioning when compared to amphetamine (Ramsey et al., 2008). Finally, all instances of differential benefits between haloperidol, which potently and selectively blocks D2Rs, and atypical antipsychotics, which affect multiple neurotransmitter systems, point to the typical antipsychotic being less effective (Duncan et al., 2006b, 2006a; Wesseling et al., 2014). These results help interpret the findings from DA system studies of GluN1KD mice. The dopaminergic deficits in GluN1KD mice do not seem to be enough to diminish amphetamine responses up to the highest dose tested of 10 mg/kg (Moy et al., 2006) but enough to affect cocaine (Ramsey et al., 2008). This is consistent with the findings of increased DAT in GluN1KD mice and suggests that the decreased TH was not enough to affect amphetamine-mediated DA release (Ferris et al., 2014). The combined results from DA-related GluN1KD studies also confirm an important role of DA in open field and PPI behavioural abnormalities, as haloperidol produces robust improvements, but DA is also not the only player as atypical antipsychotics seem to generally do better (Mohn et al., 1999; Duncan et al., 2006b, 2006a). This may be a reflection of human schizophrenia patients, but the advantage of atypical antipsychotics over haloperidol in humans may be more due to less side effects and/or better efficacy on non-positive symptoms (Leucht et al., 2009; Nandra and Agius, 2012). More specific conclusions will need further research. Currently, DA may be the best characterized neurotransmitter system in GluN1KD mice due to many studies investigating the mice for modelling schizophrenia (Mohn et al., 1999; Moy et al.,

2006; Ramsey et al., 2008; Dzirasa et al., 2009; Ferris et al., 2014). Other reports of cellular or molecular changes in GluN1KD mice are more disparate in research context.

The altered DA transmission mentioned above can also serve as an example of disrupted synaptic and neuronal connectivity in GluN1KD mice, which is a common theme in several reports of GluN1KD brain changes (Duncan et al., 2002; Dzirasa et al., 2009; Gandal et al., 2012). The differential 2-DG uptake and Fos-induction deficits between brain regions in GluN1KD mice can be interpreted as a disruption of neuronal networks as well (Duncan et al., 2002, 2008; Inada et al., 2007). For example, this was suggested for 2-DG uptake deficits in the GluN1KD CTX: uptake is impaired in layer 6 output neurons but not in layers 3 and 4 that receive thalamic input (Duncan et al., 2002). On a network-activity level, auditory-evoked synchronization of γ brain waves was deficient in GluN1KD mice, likely reflecting excitation- inhibition imbalances (Gandal et al., 2012). Brain wave synchronization patterns were also abnormal in GluN1KD mice during exploration of an 8-arm radial maze. Specifically, intra- structure synchrony at the HPC and prefrontal CTX was disrupted along with an abnormally increased synchronization between hippocampal θ waves and cortical γ waves (Dzirasa et al., 2009). Coincidental impairments of maze performance and neuronal synchrony are consistent with previous reports connecting synchronous brain activity to memory formation and rodent spatial navigation (Fell et al., 2001; Jones and Wilson, 2005; Herrmann et al., 2010). Disrupted neuronal synchrony in human patients has been reported in both ASDs and schizophrenia (Dzirasa et al., 2009; Gandal et al., 2012). Finally, direct measurements of MSN dendritic spine density revealed a deficit in GluN1KD mice at 6 weeks of age but not 2 weeks (Ramsey et al., 2011). Dendritic spines, to reiterate, are the physical locations of most excitatory synapses – they are therefore also representations of specific neuronal connections with behavioural relevance (Moser et al., 1994; Sanders et al., 2012; Saneyoshi and Hayashi, 2012). The reported deficit was primarily in mature, mushroom-shaped spines (Ramsey et al., 2011) and was coincidental with the onset of some GluN1KD behavioural abnormalities: stereotypy time, social approach time, and Y-maze alternation (Milenkovic et al., 2014). It was also found with concurrent striatal decreases of the proteins 14-3-3ε and disrupted in schizophrenia 1 (DISC1), which are related to Rho GTPase regulation, neurite growth, cognitive function, and risk for schizophrenia (Hayashi- Takagi et al., 2010; Brandon and Sawa, 2011; Ramsey et al., 2011).

This investigation into striatal DISC1 and MSN spine density of 2- and 6-week old mice is also the primary molecular study considering the effects of age in GluN1KD mice. Along with spine, 14-3-3ε and DISC1 deficits, 6-week old mice also had increased synaptic levels of the lissencephaly-1 (LIS1) protein (Ramsey et al., 2011). However, total STR level of LIS1 was decreased, pointing to dysregulation and disruption of its work with 14-3-3ε and DISC1 for neurodevelopmental processes like axon elongation and neuronal migration (Toyo-oka et al., 2003; Taya et al., 2007; Ramsey et al., 2011). The juvenile 2-week old GluN1KD STR had more modest decreases in DISC1 – only significant for one isoform at the synapse – with no other deficits of DISC1 interactor proteins (Ramsey et al., 2011). Total tissue and synaptic 14-3-3ε decreases were also confirmed in the STR of adult GluN1KD mice later, though the study focused on 14-3-3ε’s ability to increase TH activity and affect DA synthesis (Ferris et al., 2014). These molecular changes, like the spine deficits, generally mirror GluN1KD behaviour abnormality onset and progression – signs of disruption appear at 2 weeks of age and worsen by 6 weeks (Ramsey et al., 2011; Moy et al., 2012; Milenkovic et al., 2014). Overall, several independent studies using various distinct methodologies have demonstrated disrupted synaptic connectivity in some form in GluN1KD mice.

The last main group of molecular studies in GluN1KD mice are proteomic studies (Wesseling et al., 2014; Cox et al., 2016). The first is a characterization study of GluN1KD mice investigating protein level differences in blood serum, frontal CTX, and HPC (Wesseling et al., 2014). The second included GluN1KD mice in a proteomic comparison of multiple animal models of schizophrenia with the profiles of human patients, focusing on the anterior prefrontal CTX (Cox et al., 2016). The identified protein differences from both studies are compiled in Supplemental Table SI, using raw data from their supplementary materials. In the characterization study, dozens of proteins were differentially expressed between GluN1KD mice and WT littermates. The authors emphasized that various neurotrophic factors are increased in the serum of GluN1KD mice and that computational pathways analyses pointed to proteomic changes in both brain regions being related to synaptic LTP, myelination, and ERK-signaling (Wesseling et al., 2014). Combined with the fact that ERK signaling is often related to NMDAR- mediated development and maintenance of synapses, the proteomic characterization study offers consistent and robust evidence for synaptic disruption in GluN1KD mice (Thomas and Huganir, 2004; Tian et al., 2004; Wesseling et al., 2014). The second proteomic study did not investigate

GluN1KD mice quite as specifically or deeply. Instead, the authors considered changes in schizophrenia patients and animal models together to find consensus pathways and functional domains that they then used to evaluate the similarity between each animal model and human patients. The five consensus functional domains of protein changes were generally vague, but two bared some resemblance to the characterization study results: “development and differentiation” and “nucleic acid metabolism and ATP/GTPase activity”. Closer examination of specific pathways altered in GluN1KD mice reinforced this echoing of results: “cell projection organization”, “regulation of cell projection organization”, and “small GTPase mediated signal transduction” (Cox et al., 2016). Note that small GTPases include Rho GTPases as a subfamily, which were previously mentioned as part of the mechanism relating DISC1 to synaptic plasticity and cognition (Luo, 2000; Hayashi-Takagi et al., 2010; Brandon and Sawa, 2011). Taken together, the patterns of proteomic changes reported in GluN1KD mice offer some consistent support for the idea of synaptic disruptions in the GluN1KD brain.

Cellular and molecular studies of GluN1KD mice that assess sex differences are limited, but a few sex differences have been published (Duncan et al., 2009; Islam et al., 2017). Basal activation of Fos was only decreased in the male cingulate CTX and basolateral nucleus while female GluN1KD mice were not different from female WT littermates. This difference was found in control mice of the investigation into differential Fos induction after resident-intruder mice. It is important to point out then that these differences may not be completely reflective of a basal state; they may instead be differential responses to a week of social isolation that is part of the protocol for control mice (Duncan et al., 2009). Such an interpretation is possible as both brain regions have been reported to be altered with social isolation or rejection (Somerville et al., 2006; Wang et al., 2012). To note, no sex differences in Fos induction after resident-intruder testing was reported (Duncan et al., 2009). The other reported molecular sex difference in GluN1KD mice relates to the increased mortality of males on an ω-3 deficient diet. Screening of more than a dozen fatty acids in the brain and serum of GluN1KD mice in this study revealed significant increases of α-linolenic acid and linoleic acid in the brain as well as decreased gondoic acid in serum. Supplementation or deficiency of ω-3 fatty acids resulted in many changes to the GluN1KD lipid profile, but supplementation did not affect behaviours tested. GluN1KD mice on an ω-3 deficient diet, to reiterate, had worse puzzle box performance along with their increased mortality. Changes in GluN1KD fatty acids after an ω-3 deficient diet were

48 mostly the same between males and females however: increases of various ω-6 fatty acids in the brain and serum, decreases of α-linolenic and docosahexaenoic acids (both ω-3) in the brain and serum, and decrease of oleic acid (ω-9) in the serum (Islam et al., 2017). While these changes may have various specific meanings, increased ω-6 fatty acids or ω-6/ω-3 ratios are generally associated with poorer health measures (Simopoulos, 2002). Complementing this idea in GluN1KD mice, ω-3 supplementation has been reported to improve mental health outcomes in patients with depression or schizophrenia (Peet and Stokes, 2005). Despite all these changes however, the main sex-dependent change was an increase of serum eicosadienoic acid in GluN1KD males after ω-3 deficiency. A specific role for eicosadienoic acid may be to serve as negative feedback in prolonged inflammation (Huang et al., 2011), potentially pointing to a worse inflammatory state in ω-3 deficient GluN1KD males. Further research is needed to understand the presented molecular sex differences in GluN1KD mice, as well as answer other questions like why ω-3 supplementation did not improve GluN1KD phenotypes.

Cellular and molecular studies of GluN1KD mice present evidence of both hypofunction and hyperactivity in their brains (Duncan et al., 2002; Jocoy et al., 2011; Ferris et al., 2014; Grannan et al., 2016). Evidence also suggests that the net effect of developmental and chronic NMDAR deficiency is a disinhibited brain with disrupted neuronal signaling connections (Dzirasa et al., 2009; Ramsey et al., 2011; Wesseling et al., 2014). It is important to note here that these general trends are based on current, limited evidence, and may be affected by factors such as research interest bias. As an example, other molecular pathways and consensus functional domains were represented by GluN1KD proteomic changes aside from those discussed above (Wesseling et al., 2014; Cox et al., 2016). These pathways or domains can be vague and broad, such as the “biosynthetic processes and energy metabolism” consensus domain (Cox et al., 2016). It is possible that this domain is related to 2-DG uptake deficiencies in the CTX of GluN1KD mice, but more research is needed than one uptake and two proteomic studies (Duncan et al., 2002; Wesseling et al., 2014; Cox et al., 2016). There is also a report of the miR-219 being decreased in the prefrontal CTX and HPC of GluN1KD mice, which the authors related to NMDAR- antagonist-induced hyperlocomotion, but the whole field of microRNAs has been largely unexplored so far in the GluN1KD model (Kocerha et al., 2009). It is reassuring, however, that the currently available evidence points to consistent themes that can direct future research.

1.3 Study Overview

1.3.1 GluN1KD Model Validity, Utility and Research Impact

Results of many studies support the validity of using GluN1KD mice to model neurological disorders, particularly schizophrenia and ASDs (Mohn et al., 1999; Halene et al., 2009a; Gandal et al., 2012). There is strong support for the construct validity of GluN1KD mice for these disorders, as NMDAR dysfunction has been often related to both (Moghaddam and Javitt, 2012; Won et al., 2012; Burnashev and Szepetowski, 2015). Molecular evidence also supports the validity of the GluN1KD model, such as when proteins or biological processes altered in GluN1KD mice were also reported as changed in human patients (Ramsey et al., 2011; Wesseling et al., 2014; Cox et al., 2016). GluN1KD PPI deficits and hyperactivity, social behaviour deficits, and cognitive impairments may represent the positive, negative, and cognitive symptoms of schizophrenia, respectively (Mohn et al., 1999; Halene et al., 2009a; Milenkovic et al., 2014; Grannan et al., 2016). Social and cognitive impairments in GluN1KD mice can also be relevant to ASDs along with multiple forms of stereotypic behaviour, a hallmark of autism (Moy et al., 2008; Gandal et al., 2012; Milenkovic et al., 2014). To the extent that mouse behaviour can seem like human behaviour, the behavioural phenotypes seen in GluN1KD mice presents a compelling degree of face validity. Fittingly, antipsychotics improve multiple GluN1KD behavioural abnormalities (Mohn et al., 1999; Duncan et al., 2006b, 2006a) and are used to treat both schizophrenia and ASDs (Anderson et al., 1984; Miyamoto et al., 2005; Posey et al., 2008). The GluN1KD model thus also presents a degree of pharmacological predictive validity. Taken together, there are many pieces of evidence supporting the use of GluN1KD mice to model schizophrenia and ASDs.

However, the utility of GluN1KD mice is not necessarily limited to studying schizophrenia and autism. The reports of seizures and seizure-like activity in GluN1KD mice (Dzirasa et al., 2009; Halene et al., 2009a), which might reflect a higher susceptibility to seizures in autism patients (Volkmar and Nelson, 1990; Qadir et al., 2017), may also indicate a relevance to epilepsy. GluN1KD mice may be especially useful to study how NMDAR hypofunction can lead to a hyperactive or hyperexcitable brain. This would be important for understanding the many epileptogenic GRIN mutations that result in a loss of function (Xu and Luo, 2018). If GluN1KD brains are indeed in a generally hyperglutamatergic or hyperexcitable state, which certainly

50 needs more experimental confirmation, then GluN1KD mice may also be used to study the excitotoxic processes related to various pathologies (Parsons and Raymond, 2014; Weilinger et al., 2016; Tönnies and Trushina, 2017; Estrada-Rojo et al., 2018). While GluN1KD brains may generally have decreased NMDARs, about 10% still remain and can respond (Mohn et al., 1999; Duncan et al., 2008; Jocoy et al., 2011). If the distribution of remaining NMDARs in GluN1KD mice is WT-like, which is another interesting hypothesis to test, a concurrent hyperglutamatergic state may translate to a considerable level of excitotoxic signaling via relatively increased extrasynaptic NMDAR signaling (Hardingham et al., 2002; Parsons and Raymond, 2014). Furthermore, GluN1KD cognitive impairments may be useful to model similar impairments in AD patients, as NMDAR hypofunction can contribute to cognitive decline (Snyder et al., 2005; Shankar et al., 2007; Zhang et al., 2009b). Overall, the utility of GluN1KD mice likely extends beyond schizophrenia and ASDs. The various GluN1KD behavioural, cellular, and molecular phenotypes may, individually and in combinations, be useful as animal correlates of many human disorders.

Many people suffer from the neurological disorders that GluN1KD mice can model, even if just considering schizophrenia and ASDs (Kessler et al., 2005b; Melville et al., 2008; Simeone et al., 2015; McKenzie et al., 2016). Estimates of lifetime prevalence for schizophrenia and schizophreniform disorder range between 0.3 to 1.6% (Kessler et al., 2005a), with 1% being generally accepted as a good approximation (Vanasse et al., 2012). More recent global lifetime prevalence estimates for schizophrenia range from 0.48 to 0.72% of the population, though there is much variation between epidemiological studies from all over the world (McGrath et al., 2008; Simeone et al., 2015). In Canada, a lifetime prevalence of up to 1.46% was estimated in a 2012 study based on a sample size of 6 million people from Quebec (Vanasse et al., 2012). Prevalence estimates of ASDs have been in a similar range, from about 0.58 to 1.16% (Melville et al., 2008; Elsabbagh et al., 2012). Moreover, the prevalence of autism has been increasing in recent decades (Chien et al., 2011; Idring et al., 2015). Such prevalence increases have been reported in Canada and the USA (Newschaffer, 2005; Ouellette-Kuntz et al., 2014), and may not be completely explained by increased diagnoses or reporting (Ouellette-Kuntz et al., 2014; Hansen et al., 2015). From the 2010 Global Burden of Disease Study, schizophrenia was estimated to result in 13 million years lived with disability (YLDs) and 0.6 million years of life lost to premature mortality for its patients, the most severe after depressive and anxiety disorders.

Pervasive developmental disorders, which includes ASDs and Asperger's syndrome, were estimated to have a burden of 7.7 million YLDs (Whiteford et al., 2013). Beyond the patients, caregivers of patients with neurological disorders also bear significant emotional, social, and economical burdens (Kessler et al., 2008; Zahid and Ohaeri, 2010; Mak and Cheung, 2012). On a societal level, annual costs of 94 million to 102 billion US$, or 0.02% to 1.65% of gross domestic product, have been estimated for schizophrenia depending on country and study (Chong et al., 2016). Studies publishing such large-scale estimates are less common for ASDs, but they similarly report cost estimates on the scale of billions of US$ (Järbrink and Knapp, 2001; Ganz, 2007; Knapp et al., 2009; Horlin et al., 2014). Given the scale of the problem, treatment options are relatively limited for schizophrenia and autism patients, particularly for cognitive impairments (Green, 2016; Kishi et al., 2017; Masi et al., 2017). Impaired cognition greatly contributes to patients’ decreased quality of life and worse long term prognoses (Carpenter and Strauss, 1991; Howlin et al., 2004; Mitchell et al., 2010; Tolman and Kurtz, 2012). Thus, there is a great need for further research into the etiology, pathophysiology, and treatments of these disorders. Again, this is not to mention other disorders that GluN1KD mice can model or the more recently reported cases of autoimmune NMDAR encephalitis (Dalmau et al., 2007; Florance-Ryan and Dalmau, 2010; Irani et al., 2010). Using GluN1KD mice to study schizophrenia, ASDs, and other disease contexts related to NMDAR hypofunction can help address the current need for further mental health research.

1.3.2 Rationale and Purpose of the Study

A major goal for research with GluN1KD mice is to facilitate new treatment development. As part of this process, new research should build on and complement what is known and published previously. The most studied drug class in GluN1KD mice are approved antipsychotics, which improves some behavioural abnormalities (Mohn et al., 1999; Duncan et al., 2006b, 2006a). Given that these drugs are already used in patients with limited effects on non-positive schizophrenia symptoms (van Os and Kapur, 2009; Green, 2016; Masi et al., 2017), other classes of drugs and mechanisms of action should be explored. Other drugs tested on GluN1KD mice to date have targeted kainate receptors (Duncan et al., 2010, 2012), metabotropic glutamate receptors (Gregory et al., 2013), muscarinic receptors (Grannan et al., 2016), and γ-aminobutyric acid receptors (Mielnik et al., 2014). These drugs only resulted in limited improvement for a few behaviours (Duncan et al., 2012; Gregory et al., 2013) or were only tested in a few behavioural

52 paradigms (Mielnik et al., 2014; Grannan et al., 2016) and most had limited cellular or molecular GluN1KD data informing their drug use or explaining drugs’ mechanisms of action. To increase the chances of finding effective treatments for GluN1KD phenotypes and the interpretability of experimental findings regardless of efficacy, more studies of cellular and molecular changes in GluN1KD mice are needed. Preferably, these studies would also consider the behavioural relevance of molecular changes. The is part of my first research aim.

As discussed before, current cellular and molecular GluN1KD data point to a few general themes such as hyperexcitability, disinhibition, dysregulated DA signaling, and synaptic dysconnectivity (Dzirasa et al., 2009; Jocoy et al., 2011; Ramsey et al., 2011; Ferris et al., 2014). All of these can be, at least partially, subsumed under the last theme of synaptic dysconnectivity since excitatory, inhibitory, and DA signaling disruptions may well contribute to and manifest from dysfunctional synapses. The most in-depth knowledge about GluN1KD neuronal changes and synaptic disruptions relate to the STR: 2-DG uptake is inhibited (Duncan et al., 2002); Fos- induction is WT-like in almost all experimental paradigms (Miyamoto et al., 2004; Inada et al., 2007; Duncan et al., 2008; Ramsey et al., 2008); DA system abnormalities were found in the STR, where it also receives input from overactive VTA neurons (Voorn et al., 2004; Ferris et al., 2014); NMDAR currents are decreased and less enhanced by DA but AMPAR currents are increased in MSNs (Jocoy et al., 2011); and age-dependent DISC1 and spine density deficits were also found in the STR and its MSNs, respectively (Ramsey et al., 2011). Furthermore, striatal Fos-induction was associated with open field behaviour changes (Miyamoto et al., 2004; Ramsey et al., 2008) while MSN spine deficits were developmentally consistent with the onset of multiple GluN1KD behavioural abnormalities (Ramsey et al., 2011; Milenkovic et al., 2014). Thus, we chose synaptic plasticity as a major concept of focus for the research in this thesis and also focused our studies on the STR to build on previous knowledge.

To build on and complement previous knowledge of GluN1KD molecular, cellular, and behavioural phenotypes, my research includes a hypothesis-driven research aim and an unbiased bioinformatics-based aim. These two types of approaches reflect those of previous GluN1KD studies: those that focused on more specific research questions (Duncan et al., 2008; Mielnik et al., 2014) and those that were broader characterizations of GluN1KD phenotypes (Duncan et al., 2002; Wesseling et al., 2014). Firstly, we aimed to further investigate the relationship between DISC1, synaptic connectivity, and behaviour in GluN1KD mice. As noted previously, DISC1

53 can regulate Rho GTPases that develop and shape dendritic spines; all steps and components of this process are important for proper cognition (Dillon and Goda, 2005; Hayashi-Takagi et al., 2010; Brandon and Sawa, 2011). Our lab therefore asked whether levels of Rho GTPase signaling proteins were altered in the STR of GluN1KD mice concurrently with MSN spine deficits. This work was largely completed by Marija Milenkovic, a previous student of the Ramsey lab, who found general decreases of the proteins Ras-related C3 botulinum toxin substrate 1 (Rac1) and Wiskott-Aldrich syndrome protein-family verprolin homologous protein 1 (WAVE-1; data shown in results section 3.1). This was expected given the close functional relationship between DISC1 and Rac1 – DISC1 regulates guanine nucleotide exchange factors (GEFs) that activate Rac1 (Hayashi-Takagi et al., 2010; Brandon and Sawa, 2011; Chen et al., 2011). WAVE-1 changes were also consistent with DISC1 and Rac1 changes since WAVE-1 acts downstream of Rac1 (Eden et al., 2002; Soderling et al., 2003). WAVE-1 acting downstream of Rac1is also a major reason for why we further focused our attentions on studying WAVE-1.

WAVE-1 is a scaffolding protein with a molecular weight of approximately 85 kDA (Kim et al., 2006a; Soderling et al., 2007). Some of its known regulators and interactions with NMDARs and Rac1 are outlined in Figure 1.5. It is a regulator of actin dynamics in mammalian cells, and is more enriched in the CNS and neurons compared to the isoforms WAVE-2 and WAVE-3 (Kim et al., 2006b, 2006a). WAVE proteins contain three main domains: an amino-terminal Scar homology domain, a central polyproline domain, and a carboxy-terminal Verprolin–Cofilin– Acidic (VCA) domain (Kim et al., 2006a; Soderling et al., 2007). The ATD and polyproline regions of WAVE-1 bind to phospholipids for localization to the cell membrane as well as other proteins for actin and GTPase regulation (Soderling et al., 2007; Pollitt and Insall, 2009). The VCA domain is considered the main executor of WAVE-1 functions; it binds to and activates the actin-related protein (Arp) 2/3 complex, serving together as a scaffold for actin polymerization (Kim et al., 2006a; Soderling et al., 2007). Rac1 activation leads to the dissociation of cytoplasmic FMRP-interacting protein 1 from WAVE-1 complexes and a subsequent disinhibition of WAVE-1 (Pathania et al., 2014). Besides Rac1, there are also other regulators of WAVE-1 function, such as CDK5, WAVE-associated Rac GTPase-activating protein (WRP), and more (Kim et al., 2006b; Sung et al., 2008; Ceglia et al., 2010; Huang et al., 2017). Many of these regulators act by changing the phosphorylation state of WAVE-1, with phosphorylated

WAVE-1 generally considered as inhibited (Kim et al., 2006b; Sung et al., 2008). WAVE-1 and Arp2/3 activation induces actin polymerization and ultimately plays a role in dendritic spine growth, trafficking of mitochondria into spines, and other neurodevelopmental processes (Kim et al., 2006a, 2006b; Sung et al., 2008; Ito et al., 2018). It is therefore unsurprising for WAVE-1 dysregulation to contribute to neurological disorders.

Figure 1.5. Representative schematic of WAVE-1 molecular mechanisms in relation to NMDARs. Multiple reports have related WAVE-1 function to NMDAR signaling and synaptic plasticity. Proteins disrupted in GluN1KD mice are denoted in red text. Green pointed arrows indicate a positive relationship of some form, such as enhanced activation or release from a

55 sequestering factor. Red blunted lines indicate a negative relationship, causing some form of inhibition, decreased activity, or sequestering of the target. Black double-ended arrows indicate limited evidence that suggest some sort of relationship – such as bioinformatics-based associations. Abbreviations are the same as those used in text. This figure was constructed with literature cited within text as well as additional articles (Soderling et al., 2002; Camargo et al., 2007).

While WAVE-1 alterations do not cause Wiskott-Aldrich syndrome, an x-linked disease with cytoskeletal abnormalities and immunodeficiency (Soderling et al., 2003; Takenawa and Suetsugu, 2007), they do result in clinically relevant pathophysiological changes. In vitro experiments with cultured neurons have demonstrated that decreased WAVE-1 function, via phosphorylation or RNA interference, results in less mature dendritic spines (Kim et al., 2006b; Sung et al., 2008; Sanchez et al., 2009). WAVE-1 knockout mice, similar to GluN1 knockout mice, have been reported to die postnatally (Forrest et al., 1994; Li et al., 1994; Dahl et al., 2003) while other in vivo experiments have associated disruption of WAVE-1 function with decreased dendritic spines (Kim et al., 2006b; Soderling et al., 2007). Targeted WAVE-1 disruption can also result in mice with sensorimotor abnormalities and cognitive deficits (Soderling et al., 2003). Overall, the reports of spine density deficits and mouse behavioural abnormalities after WAVE-1 alterations echo previous studies of GluN1KD mice (Dzirasa et al., 2009; Halene et al., 2009a; Ramsey et al., 2011; Milenkovic et al., 2014). In humans, genetic or transcriptional disruption of the WAVE-1 encoding gene, WASF1, has been associated with schizophrenia and autism-like symptoms (Maycox et al., 2009; Ito et al., 2018), also similar to NMDAR dysfunction and GluN1KD mice (Mohn et al., 1999; Halene et al., 2009a; Gandal et al., 2012). We therefore asked whether restoring WAVE-1 improves GluN1KD spine density and behavioural abnormalities. In other words, were the discovered Rho GTPase signaling alterations in GluN1KD mice biologically significant and/or feasible targets for behavioural improvement?

We also wanted to more broadly investigate the molecular changes of the GluN1KD STR using an unbiased, bioinformatics approach. More specifically, we aimed to characterize the GluN1KD transcriptome using RNAseq, which is more sensitive compared to proteomic approaches and traditional microarrays (Hegde et al., 2003; Seyednasrollah et al., 2015). Studying the GluN1KD STR this way fills an important knowledge gap because bioinformatics- based assessments so far only investigated non-striatal brain regions (Wesseling et al., 2014; Cox

56 et al., 2016) despite many other studies including or focusing on the STR (Miyamoto et al., 2004; Ramsey et al., 2011; Ferris et al., 2014). For example, previous proteomic studies offered evidence that protein changes in GluN1KD mice likely represented changes related to synaptic plasticity at the frontal CTX and HPC (Wesseling et al., 2014; Cox et al., 2016). It would be interesting to see if similar outcomes appear for the STR when RNAseq results are analyzed with gene list analyses (Chen et al., 2013; Krämer et al., 2014; Kuleshov et al., 2016). We wanted to determine the molecular changes of the GluN1KD STR, the biological processes and pathways those changes most likely represent, and how the overall results align with what we currently know about GluN1KD mice.

Sex was included as a factor for statistical analyses in both studies when sufficiently powered because both schizophrenia and ASDs show elements of sexual dimorphism (Eranti et al., 2013; Baxter et al., 2015; Halladay et al., 2015). Again, schizophrenia has an earlier onset in males, who may also have more severe symptoms and worse outcomes (Häfner, 2003; Thorup et al., 2007; Eranti et al., 2013). ASDs are more prevalent in males (Baxter et al., 2015). Considering sex differences is also important because previous GluN1KD studies infrequently accounted for sex differences, forming another gap of knowledge. Bioinformatics-scale data are notably absent for female GluN1KD mice as previous proteomic studies only assessed males (Wesseling et al., 2014; Cox et al., 2016).

To sum, this thesis aims to investigate the relationship between Rho GTPase signaling proteins, synaptic connectivity, and behaviour in GluN1KD mice as well as to characterize the GluN1KD striatal transcriptome.

1.3.3 Specific Research Objectives

This section is meant to reiterate explicitly and specifically the research goals of the work in this thesis. There are two major research aims, both are directed at understanding the cellular and molecular changes in the GluN1KD STR. The first aim also has a specific focus on changes related to synaptic plasticity while the second involves a more unbiased bioinformatics approach.

Aim 1: Investigation into the effect of restoring of WAVE-1 on MSN spine density and behaviour in GluN1KD mice. We hypothesize that decreased striatal WAVE-1 in GluN1KD mice is biologically significant and that restoring WAVE-1 protein will improve spine density

57 and GluN1KD behaviours: Y-maze, 8-arm radial maze, puzzle box, open field, and social approach tests will be employed.

Aim 2: Characterization of the striatal transcriptome in GluN1KD mice. As this was an unbiased RNAseq characterization study, no specific hypothesis was established a priori. Pathways analyses will be run on the resulting transcriptomic differences between GluN1KD mice and WT littermates. In general, we expect both the initial RNAseq results and the subsequent pathways analyses results to be consistent with current knowledge about GluN1KD mice.

Chapter 2 Methods and Materials Methods and Materials 2.1 Animals Used in the Study

2.1.1 GluN1KD Mice

GluN1KD mice and WT littermates aged 3, 6, or 12-14 weeks were used in this study. Male and female mice were used for all experiments comparing only WT and GluN1KD mice. GluN1KD mice generation was previously described (Mohn et al., 1999). Briefly, a neomycin cassette was inserted into intron 19 of Grin1, which encodes GluN1 subunits, resulting in a reduction of full-length Grin1 mRNA by more than 90%. The resulting mouse line was backcrossed for more than 20 generations onto two genetic backgrounds, C57B1/6J and 129X1/SvJ. Specifically, mice used for comparisons between GluN1KD and WT littermates in this study were F1 progeny from the intercross breeding of C57Bl/6J Grin1+/− and 129X1/Sv Grin1+/− heterozygotes. This breeding strategy was chosen to minimize the potential confound of homozygous mutations in parent strains, in accordance with Banbury Conference recommendations (Silva et al., 1997). Experiment mice were genotyped by polymerase chain reaction analysis of tail sample DNA: the primers used to amplify GluN1KD Grin1 were 5′ AAG CGA TTA GAC AAC TAA GGG T 3′ and 5′ GCT TCC TCG TGC TTT ACG GTA T 3′ while those used for WT Grin1 were 5′ TGA GGG GAA GCT CTT CCT GT 3′ and 5′ AAG CGA TTA GAC AAC TAA GGG T 3′.

2.1.2 WAVE-Tg Mice

WAVE-Tg mice were generated by the Soderling lab from Duke University. They were constructed by the pronuclear injection of BAC DNA (id: RP11-1072O1) containing the human WASF1 locus, with tens of kbp upstream and downstream, into fertilized eggs of C57B1/6J mice. Mice used to compare between the WT, WAVE-Tg, GluN1KD and GluN1KD-WAVE genotypes were F1 progeny generated by intercross breeding between C57B1/6J Grin1+/− WAVE-Tg mice and 129X1/Sv Grin1+/− non-WAVE-Tg mice. Male and female mice were used for all experiments involving all four genotypes except for RNAscope and dendritic spine experiments, which used only males. For genotyping, the Wasf1-targeting primers 5′ CAA CTC

ATT GCA AGA ACG TGT GGA C 3′ and 5′ AAT AAA AAA ATT AGC CAG GCG TGG TG 3′ were used along with the Grin1 targeting primers mentioned above.

2.1.3 Animal Handling and Ethics

All animal housing conditions and experimentation protocols were in line with institutional (University of Toronto Faculty of Medicine and Pharmacy Animal Care Committee, approved protocol #20011988) and federal (Canadian Council on Animal Care) guidelines. Behavioural tests were performed during the light cycle (between 9am and 6pm local time) using mice that were previously naïve to all behavioral tests. To minimize animal suffering, all mouse brains were either collected after quick and precise cervical dislocation for biochemical measures or after terminal anesthesia with 250 mg/kg tribromoethanol for dendritic spine measures.

2.2 Dendritic Spine Analyses

2.2.1 Dendritic Spine Density Analysis of WT and GluN1KD Mice at Three, Six, and Twelve Weeks of Age

Dendritic spine density was assessed using DiOLISTIC labeling and confocal microscopy described previously (Seabold et al., 2010; Ruddy et al., 2015). After deep tribromoethanol anesthesia, mice were perfused transcardially with phosphate buffered saline (PBS) followed by 4% paraformaldehyde (PFA) in PBS over 10 minutes. Brains were then collected and kept in 4% PFA for an hour on ice before being stored in PBS at 4oC. Brains were then sliced coronally at a thickness of 100 µm within two days. Slices focusing on the STR (lateral 1.2 mm) were collected and their neurons were randomly labeled with DiI (1-1'-Dioctadecyl-3,3,3',3'- tetramethylindocarbocyanine perchlorate). Specifically, slices were impregnated with DiI coated gold beads using a Helios gene gun (Bio-Rad, Hercules, CA, USA). Z-stack images with steps of 0.5 μm at 60X magnification were collected using an IX81 confocal microscope and Fluoview FV 1000 (Olympus, Tokyo, Japan). Dendritic spine counting was performed on these images using NIS-Elements Basic Research (Version 3.10, Nikon, Tokyo, Japan). Striatal MSN dendrite segments were chosen mostly randomly for analysis, only systematically avoiding branching points. Three to seven z-stack image compositions were used for each mouse assessed. Spine density analysis was blinded.

2.2.2 Dendritic Spine Density and Morphology Analyses of WT, WAVE-Tg, GluN1KD, and GluN1KD-WAVE Mice

Dendritic spine density and morphology for WT, WAVE-Tg, GluN1KD, and GluN1KD- WAVE mice were also assessed via DiOLISTIC labeling and confocal microscopy (Seabold et al., 2010; Ruddy et al., 2015). Anesthetized mice were perfused transcardially with PBS followed by 4% PFA over 10 minutes before brain collection. Brains were submerged in 4% PFA for one hour before storage in PBS at 4oC and sectioned coronally at a thickness of 150 µm within two days. Slices centering on the STR and cornu ammonis 1 (CA1) region of the HPC were collected and their neurons were randomly labeled via impregnation with DiI coated tungsten beads delivered by Helios gene gun (Bio-Rad). Z-stack images with steps of 0.9 μm at 60X magnification were collected via an IX81 confocal microscope and Fluoview FV 1000 (Olympus). Z-stack composite images were analyzed with ImageJ [Version 1.50d; (Schneider et al., 2012) from the National Institutes of Health (NIH, Bethesda, MD, USA)] using the NeuronJ (Version 1.4.3) plug-in (Meijering et al., 2004). Striatal spine density and morphology assessments comparing WT, WAVE-Tg, GluN1KD and GluN1KD-WAVE mice were focused on 20 µm dendrite segments starting approximately 20 µm away from the soma. For hippocampal slices, 20 µm segments of secondary or tertiary dendrites from the basolateral and apical sides of CA1 pyramidal neurons were analyzed. Basolateral dendrites were assessed using segments starting at 30-45 µm away from the soma while apical dendrites were assessed near the initial branching points of secondary or tertiary dendrites closest to the soma – which varied in distance from soma. All dendritic segments selected for analysis avoided including actual branching points as much as possible. Three to six z-stack image compositions were used for each mouse assessed. Both spine density and spine morphology measurements were blinded.

2.3 Western Blotting

Western blot analysis was performed to assess the levels of proteins of interest. Dissected mouse STR and HPC samples were homogenized in 400 µl of PHEM buffer (0.5% TritonX 100, 60 mM PIPES, 25 mM HEPES, 10 mM EGTA, and 2 mM MgCl2) with added protease and phosphatase inhibitors (1.5 µg/ml aprotinin, 10 µg/ml leupeptin, 10 µg/ml pepstatin A, 0.1 mg/ml benzamidine, 0.25 mM PMSF, 5 mM Na orthovanadate, 10 mM NaF, 2.5 mM Na

61 pyrophosphate, 1 mM β-glycerophosphate). Loading samples were then prepared from these protein homogenates using standard SDS protein sample buffer with 5% β‑mercaptoethanol, followed by heating at 100oC for 5 minutes. Depending on the specific protein being measured, 15-25 μg of loading sample protein were electrophoresed in 7.5-12% bis-acrylamide gels at 80- 100 V for 1-3 hours and then transferred onto polyvinylidene difluoride membranes at 100 V for 1 hour. The resulting protein-bearing membranes were blocked with 5% milk or bovine serum albumin in tris-buffered saline and Tween 20, according to primary antibody supplier instructions, before overnight incubation with appropriate primary antibodies at 4oC. Membranes were incubated with fluorescent secondary antibodies the following day and imaged with the Odessey Infrared Imaging System (LI-COR, Lincoln, NE, USA). Densitometry was performed using Odessey (Version 3, LI-COR), Image Studio (Version 5, LI-COR), or ImageJ (Version 1.45, NIH). Primary antibodies used were as follows: α-WAVE-1(1:1000-2000, catalog number 75-048, NeuroMab, Davis, CA, USA), α-pS397-WAVE-1(1:2000, catalog number W2768, Sigma, St. Louis, MO, USA), α-Rac1 (1:500, catalog number ARC03, Cytoskeleton, Denver, CO, USA), and α-GAPDH (1:4000, catalog number G8795, Sigma, St. Louis, MO, USA). The secondary antibodies used were as follows: donkey α-mouse IgG-800 (1:5000, catalog number 610-731-002, Rockland, Pottstown, PA, USA), goat α-rabbit IgG-680 (1:5000, catalog number A21076, Invitrogen, Carlsbad, CA, USA), and goat α-rabbit IgG-800 (1:15000, catalog number 926-32211, LI-COR, Lincoln, NE, USA). Blots were normalized to control loading protein (GAPDH) bands, provided in the results figures, or to REVERT total protein stain (LI-COR, Lincoln, NE, USA) before analysis.

2.4 Fluorescent in situ hybridization via RNAscope

Mouse and human WASF1 expression in WT, WAVE-Tg, GluN1KD, and GluN1KD-WAVE mice was visualized by a chromogenic in situ hybridization technique called RNAscope. The experiment largely followed the RNAscope Multiplex Fluorescent Reagent Kit v2 protocol from ACD Bio (Newark, CA, USA). Fresh mouse brains were frozen by submersion in cold 2- methylbutane and sectioned sagittally (lateral 1.2 mm Bregma) at a thickness of 20 µm. Selected sections were mounted onto charged slides, fixed by submersion in 4% PFA (in PBS), and dehydrated via serial submersion in 50%, 70%, and 100% ethanol. Afterwards, the slices were

62 treated with hydrogen peroxide, followed by RNAscope Protease IV. The multiplex fluorescent assay was then performed: sections were hybridized to both human- (product reference ID 533691, ACD Bio) and mouse- (ref ID 533701) specific Wasf1 probes or to one of two control probe mixtures (positive control probes ref ID 320881, negative control probes ref ID 320871); amplification reagents 1-3 were then used on the slides to boost signal strength; finally, fluorescent signals were developed using TSA Plus cyanine 3 and 5 (PerkinElmer, Llantrisant, UK) and channel-specific horseradish peroxidases for channels 1 and 2, respectively. Vectashield with DAPI (Vector Laboratories, Burlingame, CA, USA) was then applied to the slices before glass coverslip placement. Slides were dried overnight and then imaged with an Axio Scan.Z1 slide scanner (Zeiss, Oberkochen, DE) at 20X magnification.

2.5 Behavioural Tests of Cognition in Mice

2.5.1 Y-Maze Spontaneous Alternation Test

The Y-maze spontaneous alternation test was used to assess the cognitive function of WT, WAVE-Tg, GluN1KD, and GluN1KD-WAVE mice. Its procedure and equipment have been described in previous studies (Mandillo et al., 2008; Milenkovic et al., 2014). As shown in Figure 2.1, the Y-maze consisted of three identical arms measuring 38 × 7.6 × 12.7 cm that were separated by 120o between each pair (San Diego Instruments, San Diego, CA, USA). Test mice were placed at the end of one arm and tracked with Biobserve Viewer (Version 2; St. Augustin, DE) for 8 minutes as they explored the maze. Digital zones were defined for each arm in Viewer 2 starting at approximately 5 cm away from the center of the maze. Spontaneous 3-arm alternations were defined as instances where the mouse enters each of the three zones consecutively – that is, without a repeat entry into any one zone. After the first two zone entries were excluded, the percentage of spontaneous 3-arm alternation zone entries relative to all zone entries was taken as the Y-maze performance score.

Figure 2.1. Schematic of Y-maze test (a) and potential exploration pattern results (b-d). Test mice explored the maze for 8 minutes. b-d) Red paths follow black paths. b) When the mouse enters each arm consecutively without a repeat, it is an instance of a spontaneous 3-arm alternation. c-d) When a mouse re-enters a recently-explored arm, it does not count as a 3-arm alternation, potentially indicating an error of working memory or exploration strategy.

2.5.2 Eight-Arm Radial Maze Test

The cognitive function of test mice was also assessed using the 8-arm radial maze, particularly to complement and contrast the Y-maze test. Previously, GluN1KD mice have also been reported to perform poorly in the 8-arm radial maze test (Dzirasa et al., 2009). The test maze consisted of

8 identical arms measuring 22.86 × 7.62 × 15.24 cm that met together at the center of the maze as equal sides of an octagon (San Diego Instruments). The experiment lasted 6 weeks for each test mouse cohort, which contained at least 2 mice of each genotype. Each Monday, mice were started on a restricted diet of 2-3g of food per mouse per day and no preparation sessions or test trials were performed. On subsequent Tuesdays to Fridays, food was given only after the day’s habituation session or test trial. Mice were allowed to feed ad libitum on weekends and monitored to ensure no mouse dropped below 90% of their free-feeding bodyweight through the week. The first week consisted of habituating test mice to the maze center with or without a cut- up piece of cereal for 5-10 minutes. Weeks 2-6 were 5 blocks of 4 daily maze trials. Each trial consisted of 1 minute of habituation to the maze center followed by the maze arms opening up and free maze exploration for 5 minutes or until all 8 cereal piece-baited arms were visited. A schematic of the set-up is shown in Figure 2.2. Measures of cognition were termed working memory error (WME, the number of mouse entries into already-visited arms) and entries-to- repeat (ETR, the number of maze arms visited before the test mouse makes a repeat entry into any arm) and were recorded by Biobserve Viewer (Version 3).

Figure 2.2. Schematic of 8-arm radial maze. A piece of cereal was placed at the end of each arm, offering a goal for diet-restricted test mice. Mice explored the maze for up to 5 minutes. Every entry into an already-visited (within the trial) maze arm counts as a working memory error. The number of maze arm entries before a repeat entry is taken as a score termed “entries- to-repeat”.

2.5.3 Puzzle Box Test

The puzzle box test was used to assess problem solving, short-term memory, and long-term memory in mice, as described previously (Ben Abdallah et al., 2011; Milenkovic et al., 2014). As shown in Figure 2.3, the puzzle box was an arena with two adjacent compartments measuring

58 × 28 × 27.5 cm and 14 × 28 × 27.5 cm. The two compartments could be separated by a divider with an opening that allowed free passage for mice or by one that only leaves a mouse- sized underpass connection between the compartments. For each trial, test mice were placed in the larger compartment under bright-light conditions facing away from the divider. As the smaller compartment has a roof providing dim-light conditions, mice are motivated to enter it. Test performance was defined by the amount of time it took each mouse to enter the dim-light compartment, which was manually recorded. There were 9 trials spread to 3 per day with new obstacles introduced between the bright- and dim-light compartments during the second trial of each day. The open-door divider was replaced with the door-less divider for trial 2 and all subsequent trials. The underpass was blocked with cage bedding in trials 5-7 or with a cardboard plug in trials 8-9. A summary of the obstacle arrangements is shown in Table II-I. A maximum of 5 minutes was given for each trial and 2-minute breaks were given between trials on the same day. Mice that did not reach the dim-light compartment during the first trial with an open-door divider were excluded from the experiment.

Figure 2.3. Schematic of puzzle box test. Test mice are set in the brightly-lit and larger arena compartment (depicted on the right), face away from the dimmer goal compartment (left). One of two dividers separate the compartments – with or without a door. An underpass also connects both arena compartments, below the middle of the divider. Divider and underpass states for each puzzle box trial are listed in Table II-I. The amount of time it takes for test mice to enter the goal compartment (up to 5 min) is taken as test performance scores.

Table II-I. Puzzle box configuration over all trials. Increasingly difficult challenges are presented to test mice: dividers are switched from open-door to doorless from trial 2 onwards, the underpass is filled with cage bedding for trials 5-7, and the underpass is blocked with a cardboard plug for trials 8-9. The introduction of new challenges tests the puzzle-solving ability of mice while the repeat-trials are assessments of short- and long-term memory.

Trial 1 Trial 2 Trial 3 Trial 4 Trial 5 Trial 6 Trial 7 Trial 8 Trial 9 (Day 1) (Day 1) (Day 1) (Day 2) (Day 2) (Day 2) (Day 3) (Day 3) (Day 3) Divider Open-door Doorless Doorless Doorless Doorless Doorless Doorless Doorless Doorless Filled with Filled with Filled with Filled with Filled with Underpass Empty Empty Empty Empty cage cage cage cardboard cardboard bedding bedding bedding plug plug

2.5.4 Open Field Locomotion and Stereotypy Test

The open field test was used to assess horizontal motor activity, habituation, and stereotypy of mice, as described previously (Choleris et al., 2001; Milenkovic et al., 2014). Test mice were placed in clear Plexiglas chambers with the dimensions 20 × 20 × 45 cm, a novel environment to them, for 2 hours. Mouse locomotor activity and measures of stereotypy were tracked and recorded by the infrared beam sensors of digital activity monitors from Omnitech Electronics (Columbus, OH, USA). The 2-hour time-course of horizontal distance travelled and stereotypy event count were analyzed to assess the overall activity of test mice.

2.5.5 Social Approach Behaviour Test

WT, WAVE-Tg, GluN1KD, and GluN1KD-WAVE mice were also assessed for their social cognition via the social approach behaviour test. Specifically, a modified social approach behaviour test that was previously described (Moy et al., 2004; Mielnik et al., 2014; Milenkovic et al., 2014) was used. The test uses a white Plexiglas arena (62 × 40.5 × 23 cm) containing two inverted wire cups that were placed symmetrically. One cup was empty while the other contained novel age- and sex- matched mice, which served as social stimuli. The set-up is diagrammed in Figure 2.4. Test mice were placed in the arena at an equal distance away from both cups and their subsequent interactions with the wire cups were recorded with a video camera for 10 minutes. Biobserve Viewer 2 was used to define circular zones (5 cm radius) centered on each cup, to track each test mouse’s center of body mass, and to quantify the amount of time spent by

68 test mice in each zone. The amount of time spent in the social stimulus and nonsocial zones were taken as measures of social cognition and object-novelty control, respectively.

Figure 2.4. Schematic of modified social approach behaviour test. Mice were placed equidistant and facing away from both wire cups. Afterwards, test mice explored the arena for 10 minutes. The position of social stimulus mice alternated between the two cups for different test mice. The amount of time test mice spent investigating each cup is recorded as a measure of social cognition (mouse-containing cup) or as a control for object novelty (empty cup).

2.6 RNAseq Analyses

2.6.1 Sample Preparation and Differential Gene Expression Analysis

Total RNA was isolated from dissected whole-striatal tissue from 6 WT and 6 GluN1KD mice, 3 of each sex for each genotype and at 12 weeks of age. RNA isolation was performed with Tri Reagent (Molecular Research Center, Cincinnati, OH, USA) and sample RNA purity was verified with OD260/OD280 ratios taken with an Epoch Microplate Spectrophotometer (BioTek, Winooski, VT, USA). For each sample, 2-6 µg of RNA was sent to The Centre for

Applied Genomics (TCAG; Toronto, Canada) for poly-A isolation, library generation, amplification, and sequencing. TCAG generated raw sequence data using the Illumina HiSeq 2500 protocol (Illumina, San Diego, CA, USA).

Resulting FASTQ files from TCAG were analyzed using the tuxedo protocol (Trapnell et al., 2012) and the public Galaxy platform (http://usegalaxy.org) (Afgan et al., 2018). After uploading, FASTQ files were prepared for use on Galaxy with FASTQ Groomer (Version 1.0.4) and each file’s raw sequence reads were aligned to the GRCm38/mm10 mouse genome assembly with TopHat (Version 2.1.0). Aligned reads were then used to generate and predict transcripts for each sample via Cufflinks (Version 2.2.1.0) while referencing a RefSeq annotation from the National Center for Biotechnology Information (NCBI; Bethesda, MD, USA). Resulting cufflinks assemblies for each sample were then combined together with Cuffmerge (Version 2.2.1.0). Gene expression quantifications and comparisons were performed with Cuffdiff (Version 2.2.1.3) using the aligned sequence reads for every sample, combined transcript index from Cuffmerge, and GRCm38/mm10 genome assembly to correct for organism expression bias. Two types of comparisons were performed: WT vs GluN1KD and WT male vs WT female vs GluN1KD male vs GluN1KD female. P and q (false discovery rate-adjusted p) value cut-offs were set to 0.05 and 0.1 respectively.

Genes that were expressed differentially between WT and GluN1KD mice for at least one sex were identified and annotated for their specific significant differences between the four genotype X sex groups. Unrecognized entries that were significantly different between WT and GluN1KD mice were manually double-checked for an identifier via cross-referencing their genomic positions with data from the University of California, Santa Cruz Table Browser (https://genome.ucsc.edu/cgi-bin/hgTables) (Karolchik et al., 2004), Ensembl BioMart (accessed through http://www.ensembl.org/index.html) (Kinsella et al., 2011), and NCBI gene database (https://www.ncbi.nlm.nih.gov/gene) (O’Leary et al., 2016).

2.6.2 Gene List Analyses

All identified genes based on sex-matched genotype comparisons were further analyzed as gene lists with Ingenuity Pathway Analysis (IPA; QIAGEN Bioinformatics, Aarhus, Denmark) and Enrichr (http://amp.pharm.mssm.edu/Enrichr/) (Chen et al., 2013; Kuleshov et al., 2016). IPA was run to predict key molecules and pathways connected to and represented by the

70 transcriptomic differences between mice of different GluN1KD genotype and sex. Four gene lists served as inputs for IPA – for each sex-matched genotype comparison and each genotype- matched sex comparison, a list of genes with significant differences was included. Each list included gene names, fold changes and q-values. IPA was run including all information from the Ingenuity Knowledge Base (including all node types, mutation studies, and data sources) that was experimentally observed and considering endogenous chemical effects and relationships. Analyses were however constrained to knowledge specific to mice, nervous-system-related tissues and cell lines, and other cell types that could be in the original sample of dissected mouse STR (such as endothelial cells). Gene lists were also analyzed with Enrichr to identify transcriptome signatures of curated databases or specific study subjects that are the most similar. Four gene lists, consisting simply of gene names, were entered to Enrichr: genes that had increased expression in GluN1KD males compared to male WT littermates, those that were decreased in GluN1KD males, those that were increased in GluN1KD females relative to female WT littermates, and those that were decreased in GluN1KD females. Done this way, the input gene lists could be matched with Enrichr entries that specify the direction of transcriptomic changes.

2.7 Statistical Analysis

Prism (Version 6.01, GraphPad Software, La Jolla, CA, USA) was used to graph and analyze most experimental data. Exceptions include repeated measures data that were analyzed by two- way ANOVAs with repeated measures using SPSS (Version 20.0.0, IBM, Armonk, NY, USA) and RNAseq data, which were analyzed with derivatives of the Jensen-Shannon divergence (Trapnell et al., 2010) or the Fisher exact test (Chen et al., 2013; Krämer et al., 2014; Kuleshov et al., 2016) followed by Benjamini-Hochberg (BH) procedural corrections for multiple testing. Assumptions of equal variance and sphericity were tested where appropriate. WT and GluN1KD comparisons were mostly assessed with independent, two-tailed t-tests. Comparisons between WT, WAVE-Tg, GluN1KD, and GluN1KD-WAVE mice were mostly assessed with one-way ANOVAs, Kruskal-Wallis tests when variance was not equal, or two-way ANOVAs with repeated measures. Comparisons of repeated-measures data included locomotor activity, stereotypy, and 8-arm radial maze performance (WME and ETR). Bonferroni corrections were

71 used after ANOVAs for post hoc comparisons between all groups. Dunn’s post hoc multiple comparisons tests were used after Kruskal-Wallis tests. Kolmogorov–Smirnov (KS) Tests were also used to assess spine morphology differences as distribution comparisons. As KS tests could only be performed for pairs of groups, an a priori testing strategy was decided to decrease the chance of false positives: WT vs GluN1KD mice would be compared first; if significantly different, GluN1KD vs GluN1KD-WAVE and WT vs GluN1KD-WAVE would be compared to assess rescue. A cut-off p-value of 0.05 was chosen for statistical significance while a cut-off q- value of 0.1 was chosen for RNAseq results. Where appropriate, sample sizes are indicated within results figures. Data are expressed as mean ± standard error of the mean. Post hoc power analyses were performed with G*Power [Version 3.1.9.2; (Faul et al., 2007, 2009) from the Heinrich Heine University Düsseldorf].

Chapter 3 Results Results 3.1 GluN1KD Mice Show MSN Spine Density Deficits and Rac1 Signaling Protein Aberrations that are Age-Dependent

Note: Data presented in all of section 3.1.1 and most of 3.1.2 were produced by and part of the thesis of Marija Milenkovic, a previous student of the Ramsey Lab.

3.1.1 MSN Dendritic Spine Density is Decreased in Six- and Twelve-Week- Old GluN1KD Mice

We first compared the MSN dendritic spine densities of 3-, 6-, and 12-week-old WT and GluN1KD mice to establish a time course of synaptic deficits. As shown in Figure 3.1, GluN1KD mice show spine density deficits by 6 weeks of age, which persists when they are 12 weeks old. Specifically, 6-week-old WT mice had 138.2 ± 1.5 spines/100 μm while GluN1KD littermates had a spine density of 123.8 ± 2.1 spines/100 μm, a 11% reduction (independent, two- tailed t(4) = 5.55, p < 0.01). At 12 weeks, spine densities were 135.5 ± 1.7 spines/100 μm and 114.3 ± 1.4 spines/100 μm for WT and GluN1KD mice, respectively (16% reduction in

GluN1KD mice, t(4) = 9.74, p < 0.01). Our results point to MSN spine loss manifesting between 3 and 6 weeks of age for GluN1KD mice and persisting at 12 weeks of age.

Figure 3.1. Representative images (left) and graphs (right) showing MSN dendritic spine density of WT and GluN1KD mice aged 3, 6, and 12 weeks. GluN1KD mice have significant spine density deficits at 6 and 12 weeks of age compared to WT littermates (11% and 16% decreases, respectively). 3-7 dendrite images were quantified for each mouse. Mouse sample sizes are indicated in each bar. The scale bar for representative images represents 20 µm. Data in this figure were analyzed by two-tailed, independent t-tests. ** p < 0.01.

3.1.2 Rac1 and WAVE-1 Protein Levels are Decreased in Six- and Twelve- Week Old Mice but Differentially Altered in Three Week Olds

Striatal Rho GTPase pathway proteins in 3-, 6-, and 12-week-old WT and GluN1KD mice were measured by western blotting. Rac1, RhoA, WAVE-1, LIMK1, actin, cortactin, cofilin, pS3-cofilin, cell division control protein 42 homolog (Cdc42), and neural Wiskott-Aldrich syndrome protein (N-WASP) were investigated. Of these proteins of interest, most were not changed in GluN1KD mice compared to WT littermates. However, significant differences were found consistently for Rac1 and WAVE-1, as shown in Figure 3.2. Rac1 was expressed at 125.8

± 2.8 % (independent, two-tailed t(6) = 6.00, p < 0.01), 69.7 ± 10.1 % (t(6) = 2.45, p = 0.05), and

76.3 ± 1.5 % (t(6) = 5.07, p < 0.01) relative to WT levels at 3, 6, and 12 weeks of age, respectively. WAVE-1 was expressed at 74.5 ± 4.9 % (independent, two-tailed t(6) = 3.04, p =

0.02), 63.3 ± 9.0 % (t(6) = 2.44, p = 0.05), and 79.5 ± 2.4 % (t(6) = 2.94, p = 0.03) of WT levels at 3, 6, and 12 weeks of age, respectively. From these results, we saw that Rac1 signaling was particularly altered relative to all the Rho GTPase signaling proteins of interest, as WAVE-1 is a downstream effector of Rac1 (Eden et al., 2002; Bompard and Caron, 2004).

Figure 3.2. Representative western blots (top) and their quantifications (bottom) measuring Rac1 and WAVE-1 at the STR of 3-, 6-, and 12-week-old WT and GluN1KD mice. Rac1 was significantly higher in 3-week-old GluN1KD mice before decreasing to levels significantly lower than WT by 6 and 12 weeks of age. WAVE-1, a downstream effector of Rac1, was consistently lower in GluN1KD mice at all ages tested. Rac1 and WAVE-1 bands were first normalized to GAPDH loading control bands. As indicated in the graph legends, a sample size of 4 mice was used for every group. Data in this figure were analyzed by two-tailed, independent t- tests. * p ≤ 0.05, ** p < 0.01.

3.2 WAVE-Tg and GluN1KD-WAVE Mice Express Wasf1 and WAVE-1 Reliably, Predictably, and Abundantly

3.2.1 WASF1 Transgene Expression is Abundant and Similar to Endogenous Wasf1 at the STR and HPC of WAVE-Tg and GluN1KD- WAVE Mice

Previous studies have reported a causal relationship between Rac1 and WAVE-1 reductions and subsequent dendritic spine loss (Soderling et al., 2007; Haditsch et al., 2009), both observed

76 in older GluN1KD mice. Chronic Rac1 activity increase, which may be present in GluN1KD mice between 3 and 6 weeks of age, can however cause different spine abnormalities, including increased or unchanged spine densities but deteriorated spine morphology (Luo et al., 1996; Tashiro et al., 2000; Hayashi-Takagi et al., 2010). We therefore considered that Rac1 and WAVE-1 decreases could contribute to GluN1KD spine loss or be compensatory for spine loss prevention. To assess these possibilities, we aimed to restore WAVE-1 in GluN1KD mice by intercross breeding with WAVE-1-overexpressing, WASF1 transgenic mice.

WAVE-1-overexpressing, WASF1 transgenic mice (WAVE-Tg) were generated by pronuclear injection of the entire the human WASF1 genomic sequence (including 86.5 kb of upstream and 70.3 kb of downstream sequence) via a bacterial artificial chromosome. Progeny of the GluN1KD/WAVE-Tg intercross breeding were viable and survived to adulthood. As we observed GluN1KD spine density deficits in older mice, similar to previous reports of GluN1KD behavioural abnormalities (Ramsey et al., 2011; Milenkovic et al., 2014), we focused on adult GluN1KD/WAVE-Tg intercross progeny for further experiments. WASF1 transgenic mice (both WAVE-Tg and GluN1KD-WAVE-Tg hybrids, hereafter termed GluN1KD-WAVE) have increased WASF1 mRNA compared to non-transgenic (WT and GluN1KD) littermates at the STR (Figure 3.3) and HPC (Figure 3.4). The WASF1 mRNA increases of transgenic mice are specific to the human transgene (green signal) and similar to endogenous mouse Wasf1 expression (red signal) in pattern.

Figure 3.3. RNAscope visualization of Wasf1 expression at the STR of WT, WAVE-Tg,

GluN1KD, and GluN1KD-WAVE mice. Endogenous mouse Wasf1 mRNA is visualized in red

(left) while transgenic human WASF1 is in green (middle). Transgenic WASF1 was expressed abundantly and specifically in WAVE-Tg and GluN1KD-WAVE mice with an expression pattern similar to endogenous Wasf1.

Figure 3.4. RNAscope visualization of Wasf1 expression at the HPC of WT, WAVE-Tg,

GluN1KD, and GluN1KD-WAVE mice. Endogenous mouse Wasf1 mRNA is visualized in red

3.2.2 WAVE-1 Protein Levels are Increased in WAVE-Tg Mice and Improved in GluN1KD-WAVE Littermates at the STR

We also measured striatal Rac1 and WAVE-1 protein levels in WT, WAVE-Tg, GluN1KD, and GluN1KD-WAVE mice by western blots. To get a clearer sense of WAVE-1 signaling in test mice, we also measured pS397-WAVE-1, which is considered an inactive form of WAVE-1 (Kim et al., 2006b). The results are shown in Figure 3.5. One-way ANOVA revealed a significant effect of genotype on levels of WAVE-1 and pS397-WAVE-1 (F3,20 = 33.59, p <

0.01; F3,20 = 10.37, p < 0.01; respectively) but not Rac1 (F3,20 = 1.51, p = 0.24, β = 0.66). Rac1 was expressed less in GluN1KD genotypes (85.6 ± 4.5 % and 92.0 ± 6.4 % for GluN1KD and GluN1KD-WAVE mice, respectively) compared to non-GluN1KD genotypes (100 ± 6.0 % and 99.6 ± 5.4 % for WT and WAVE-Tg mice, respectively) but, again, these differences were not statistically significant. Intercrossing the GluN1KD and WAVE-Tg mouse lines produced transgenic progeny with increased expression of WAVE-1. Compared to WT littermates, WAVE-1 was expressed at 119.7 ± 4.6 % in WAVE-Tg mice (Bonferroni adjusted p < 0.01), 68.6 ± 2.3 % in GluN1KD mice (p < 0.01), and 87.1 ± 3.5 % in GluN1KD-WAVE hybrids (p = 0.14). pS397-WAVE-1 was expressed at 116.4 ± 6.1 % of WT levels in WAVE-Tg mice (Bonferroni adjusted p = 0.32 vs WT), 143.3 ± 8.1 % in GluN1KD mice (p < 0.01 vs WT), and 125.9 ± 2.2 % in GluN1KD-WAVE hybrids (p = 0.02 vs WT) after normalization to total WAVE-1. pS397-WAVE-1 was not significantly different between GluN1KD and GluN1KD- WAVE mice (p = 0.24). These results point to significantly reduced WAVE-1 function in GluN1KD mice as total WAVE-1 is reduced and what is left is highly phosphorylated. However, our results also indicate that the Wasf1 transgene improved total WAVE-1 levels in GluN1KD- WAVE hybrids.

Figure 3.5. Representative western blots (top) and their quantifications (bottom) measuring Rac1, WAVE-1, and pS397-WAVE-1 in the STR of WT, WAVE-Tg, GluN1KD, and GluN1KD-WAVE mice. No significant effect of genotype was detected statistically for Rac1. Total WAVE-1 was confirmed as decreased in GluN1KD mice, increased in WAVE-Tg mice, and improved towards WT levels in GluN1KD-WAVE hybrids. The percentage of pS397- WAVE-1, normalized to total WAVE-1, was higher in GluN1KD and GluN1KD-WAVE mice. All blots were normalized to REVERT total protein stain as loading control. Indicated in bars, a mouse sample size of 6 was used for every genotype. All data in the figure were analyzed by one-way ANOVAs followed by Bonferroni post hoc tests for all pairings. * p < 0.05, ** p < 0.01.

3.3 Spine Density Deficits and Maze Exploration Abnormalities are Improved in GluN1KD-WAVE Mice

3.3.1 Striatal MSN Dendritic Spine Density is Improved in GluN1KD-WAVE Hybrids

As total WAVE-1 protein levels were improved in GluN1KD-WAVE hybrids, we also asked whether their spine density was improved. Furthermore, we assessed MSN spine morphology, which could also reflect alterations in Rho GTPases and behavioural abnormalities (Nakayama et al., 2000; Schubert et al., 2006; Arellano et al., 2007). Dendritic spine density results are shown in Figure 3.6a. One-way ANOVA yielded a genotype effect on spine density trending towards significance (F3,32 = 2.79, p = 0.06). GluN1KD mice had a spine density of 144.9 ± 8.7 spines/100 μm, lower than WT levels of 184.8 ± 11.0 spines/100 μm (Bonferroni adjusted p = 0.05). GluN1KD-WAVE hybrids, in contrast, had a spine density of 172.3 ± 6.3 spines/100 μm (p > 0.99 vs WT). No morphological differences were detected between MSN spines of WT, WAVE-Tg, GluN1KD and GluN1KD-WAVE mice (Supplemental Figures S1 and S2). Thus, the improved WAVE-1 levels of GluN1KD-WAVE mice may be contributing to MSN spine density recovery. These results further led us to ask whether GluN1KD-WAVE mice perform better than GluN1KD littermates in cognitive behaviour tests – a more definitive assessment of the biological significance of increasing WAVE-1 in NMDAR-deficient mice.

3.3.2 Y-Maze Spontaneous Alternation Performance is Normalized in GluN1KD-WAVE Hybrids

The Y-maze test was used to evaluate the cognitive function of WT, WAVE-Tg, GluN1KD, and GluN1KD-WAVE mice. One-way ANOVA showed a significant effect of genotype on Y- maze performance scores (F3,88 = 5.70, p < 0.01). As shown in Figure 3.6b, GluN1KD mice had a lower percentage of spontaneous 3-arm alternations (43.2 ± 2.4 %) relative to WT (51.3 ± 1.5 %, Bonferroni adjusted p = 0.04) and WAVE-Tg littermates (51.3 ± 1.8 %, p = 0.03). GluN1KD- WAVE hybrids performed better than GluN1KD mice (54.1 ± 2.0 %, p = 0.01 vs GluN1KD) and were not significantly different from WT or WAVE-Tg mice (p > 0.99 when compared to either). No sex differences were observed in this behaviour test. Thus, in the context of Y-maze exploration, intercrossing GluN1KD and WAVE-Tg mice resulted in GluN1KD-WAVE hybrids that have improved cognitive function.

3.3.3 Eight-Arm Radial Maze WME is Improved in GluN1KD-WAVE Hybrids

To confirm and complement our Y-Maze results, we used the 8-arm radial maze to investigate the cognition of WT, WAVE-Tg, GluN1KD, and GluN1KD-WAVE mice. The results for both WME and ETR are shown in Figure 3.6c. Two-way repeated measures ANOVAs reported significant genotype effects on both WME and ETR (F3,40 = 53.03, p < 0.01; F3,40 = 66.80, p < 0.01; respectively). WME results indicated that, over the course of 5 trial blocks, GluN1KD mice entered already-explored arms more often than WT (Bonferroni adjusted p < 0.01), WAVE-Tg (p < 0.01), and GluN1KD-WAVE littermates (p = 0.02). In contrast to their Y-maze performance however, GluN1KD-WAVE hybrids also performed worse than WT (p < 0.01) and WAVE-Tg mice (p < 0.01). Furthermore, lower ETR scores indicated that GluN1KD and GluN1KD- WAVE mice were both quick to re-enter an arm that was already explored (p < 0.01 for each pairwise comparison between a non-GluN1KD and a GluN1KD genotype). GluN1KD and GluN1KD-WAVE mice did not have significantly different mean ETR scores. When tested by the 8-arm radial maze, GluN1KD-WAVE hybrids displayed a selective and partial improvement of WME scores.

Figure 3.6. GluN1KD-WAVE Hybrids have improved MSN spine density (a), Y-maze spontaneous 3-arm alternation (b), and 8-arm radial maze WME (c) compared to GluN1KD littermates. a) Representative MSN dendrite images (left) and spine density quantifications (right) show GluN1KD deficits, relative to WT littermates, that are improved in GluN1KD-WAVE hybrids. 3-6 dendrite images were analyzed per mouse. Mouse sample sizes for each genotype are indicated within the bars. Scale bar represents 20µm for dendrite images. MSN Spine density data was analyzed by one-way ANOVA with Bonferroni post hoc comparisons for all pairings. * p = 0.05. b) GluN1KD mice had significantly lower 3-arm alternation scores compared to WT, WAVE-Tg, and GluN1KD-WAVE littermates. GluN1KD- WAVE hybrids were not significantly different from WT and WAVE-Tg mice. Y-maze scores

84 were analyzed by one-way ANOVA with Bonferroni post hoc comparisons for all pairings. * p < 0.05, ** p < 0.01. Group sample sizes indicated in each bar. No sex differences were found for Y-maze performance. c) 8-arm radial maze WME (top) and ETR (bottom) over time show different effects of overexpressing WAVE-1 in GluN1KD mice. WME and ETR were worse in both GluN1KD genotypes compared to either non-GluN1KD genotype. However, GluN1KD- WAVE hybrids also had less WMEs over the course of 5 trial blocks compared to GluN1KD littermates. Both WME and ETR were analyzed by two-way repeated measures ANOVAs. # Bonferroni adjusted p < 0.01 for each pairwise comparison between a non-GluN1KD and a GluN1KD genotype. * p = 0.02 for GluN1KD-WAVE vs GluN1KD WME. Sample sizes are denoted next to genotypes in the graph legend. No sex differences were found in 8-arm radial maze WME or ETR.

3.4 Other Behavioural Abnormalities Persist in GluN1KD-WAVE Hybrids

3.4.1 Puzzle Box Solution and Memory Deficits Persist in GluN1KD-WAVE Mice

The puzzle box test was also used to assess the cognitive function of test mice. Different trials evaluated puzzle solving, short-term memory, and long-term memory to varying degrees. Therefore, the solution times for each trial were analyzed by Kruskal-Wallis tests, which 2 consistently reported significant genotype effects (휒3 = 25.87-69.29, N = 94, and p < 0.01 for all puzzle box trials). Puzzle box completion times for all trials are shown in Figure 3.7a. GluN1KD and GluN1KD-WAVE mice were slower to complete all trials relative to WT and WAVE-Tg littermates. Furthermore, GluN1KD and GluN1KD-WAVE mice regularly failed to complete trials 5-9 within the 5-minute maximum exploration time. Dunn’s post hoc tests reported significantly different rankings for each pairwise comparison between a non-GluN1KD genotype and a GluN1KD genotype, for all trials (p ≤ 0.015). GluN1KD-WAVE hybrids were not significantly different from GluN1KD littermates, indicating no improvement of cognition in the puzzle box context.

3.4.2 GluN1KD Horizontal Activity and Stereotypy Abnormalities Persist in GluN1KD-WAVE Mice

We performed the open field test to investigate the effects WAVE-1 rescue might have on other aspects of cognition – locomotor activity, habituation, and stereotypic behaviour. Two-way repeated measures ANOVAs revealed significant genotype effects on locomotor activity and stereotypic incidence number over time (F3,83 = 53.59, p < 0.01; F3,83 = 55.42, p < 0.01; respectively). Shown in Figure 3.7b, GluN1KD and GluN1KD-WAVE mice had more horizontal locomotor activity, decreased habituation (slower activity decrease over time), and more stereotypic behaviour. All post hoc pairwise comparisons between mice of a non-GluN1KD genotype and those of a GluN1KD genotype were significant (Bonferroni adjusted p < 0.01). GluN1KD and GluN1KD-WAVE mice were not significantly different, indicating no open field test improvement was associated with increased striatal WAVE-1.

3.4.3 GluN1KD Social Cognition Deficits Persist in GluN1KD-WAVE Mice

The social approach behavior test was used to investigate WAVE-1 rescue effects on social cognition, represented by how long test mice spend in zones containing social stimulus mice (Mielnik et al., 2014). One-way ANOVA revealed a significant effect of genotype on social approach behavior (F3,92 = 5.69, p < 0.01). GluN1KD mice (202.2 ± 12.4 s) and GluN1KD- WAVE hybrids (213.1 ± 15.9 s) spent significantly less time in social zones compared to WAVE-Tg littermates (269.9 ± 14.0 s, Bonferroni adjusted p ≤ 0.01 for both comparisons). WT mice also spent more time in social zones (258.7 ± 13.7 s) relative to GluN1KD mice (p = 0.04), though it was a trend compared to GluN1KD-WAVE mice (p = 0.15). Nonsocial zone time, the time spent by test mice in zones with empty wire cups, was also measured and reported as a control for general object novelty. Nonsocial zone time also had a significant genotype effect reported by one-way ANOVA (F3,92 = 16.84, p < 0.01). WT mice also spent more time investigating empty cups compared to GluN1KD (141.6 ± 9.3 s vs 63.2 ± 7.7 s respectively, p < 0.01) and GluN1KD-WAVE littermates (83.7 ± 7.7 s, p < 0.01). No sex differences were observed in either measures of the social approach test. Social cognition was not improved in GluN1KD-WAVE mice.

Figure 3.7. GluN1KD-WAVE hybrids did not show improvements in puzzle box (a), open field (b) or social approach behavior (c) test performance. a) GluN1KD and GluN1KD-

WAVE mice were significantly slower in completing puzzle box trials when compared to WT and WAVE-Tg littermates. Data for each trial, representing different cognitive aspects, was analyzed by Kruskal-Wallis tests with Dunn’s post hoc multiple comparisons. # indicates p ≤ 0.015 for each pairwise comparison between a non-GluN1KD genotype and a GluN1KD genotype for each trial’s completion time. GluN1KD and GluN1KD-WAVE mice did not differ. Sample sizes are indicated in the graph legend next to their respective genotypes. No sex differences were found for any puzzle box trials. b) Distance traveled (top) and stereotypy number (bottom) were recorded over time in the open field test. GluN1KD and GluN1KD- WAVE mice travelled further and had more stereotypic incidents, not showing as much habituation over time, when compared to non-GluN1KD littermates. Data were analyzed by two- way repeated measures ANOVAs. # Bonferroni post hoc p < 0.01 for each pairwise comparison between a non-GluN1KD and a GluN1KD genotype. GluN1KD and GluN1KD-WAVE mice did not differ. Sample sizes are indicated in the graph legend next to their respective genotypes. No sex differences were found for the open field test. c) GluN1KD and GluN1KD-WAVE mice spent less time in social zones compared to non-GluN1KD littermates. GluN1KD-WAVE social zone time was significantly lower than WAVE-Tg social zone time and trended towards significance when compared to WT results († p = 0.15). WT nonsocial zone time was also higher compared to both GluN1KD and GluN1KD-WAVE mice. GluN1KD and GluN1KD-WAVE mice did not differ. Data were analyzed by one-way ANOVAs with Bonferroni post hoc comparisons for all pairings. * p < 0.05, ** p < 0.01. Sample sizes are indicated in the bars of their respective genotypes. No sex differences were observed in the social approach behaviour test.

3.5 Improvements in GluN1KD-WAVE Hybrids are not Reflected by WAVE-1 Levels at the HPC nor CA1 Spines

3.5.1 Hippocampal WAVE-1 Levels are not Different Between WT, WAVE- Tg, GluN1KD, and GluN1KD-WAVE Mice

Finally, since Y-maze and 8-arm radial maze performance were improved in GluN1KD- WAVE mice, we further asked whether WAVE-1 and spine density improvements were seen at the HPC, a key brain region for learning and memory formation (McHugh et al., 2007; Neves et

88 al., 2008). Western blots measuring total WAVE-1 in the HPC of test mice are shown in Figure

3.8a. One-way ANOVA found no significant genotype effect (F3,20 = 2.33, p = 0.11, β = 0.50). WAVE-1 was expressed at 102.5 ± 13.2 % of WT levels in WAVE-Tg mice, 102.7 ± 11.4 % in GluN1KD mice, and 84.2 ± 19.9 % in GluN1KD-WAVE hybrids.

3.5.2 CA1 Pyramidal Neuron Spine Density is not Different Between WT, WAVE-Tg, GluN1KD, and GluN1KD-WAVE Mice

One-way ANOVAs found no significant genotype effect on CA1 pyramidal neuron dendritic spines, whether they be apical or basolateral (F3,28 = 0.93, p = 0.44, β = 0.77; F3,28 = 2.48, p = 0.08, β = 0.45; respectively). Hippocampal spine density data are shown in Figure 3.8b. GluN1KD mice had an apical spine density of 130.0 ± 10.2 spines/100 μm while GluN1KD- WAVE hybrids had 131.3 ± 9.2 spines/100 μm. These measures were lower, but not significant, compared to WT and WAVE-Tg littermates, which had 150.4 ± 10.7 spines/100 μm and 141.9 ± 9.7 spines/100 μm, respectively. Basolateral spine density measurements were similar: GluN1KD mice had 106.0 ± 8.4 spines/100 μm and GluN1KD-WAVE hybrids had 105.0 ± 8.8 spines/100 μm, neither significantly lower than the 124.8 ± 5.3 spines/100 μm of WT mice or the 124.8 ± 4.6 spines/100 μm of WAVE-Tg mice. No major or consistent morphological differences were detected between test mice as well (Supplemental Figures S3-6).

Overall, we did not find GluN1KD deficits for hippocampal WAVE-1 or CA1 spine density. Intercrossing the GluN1KD and WAVE-Tg mouse lines also did not significantly affect either measure. However, our experiments investigating the HPC of test mice were found to be underpowered post hoc.

Figure 3.8. Assessments of Hippocampal WAVE-1 protein levels (a) and CA1 pyramidal neuron spine density (b) in WT, WAVE-Tg, GluN1KD and GluN1KD-WAVE mice. a) Representative western blots of WAVE-1 (top) and their quantifications (bottom) show no significant difference between WT, WAVE-Tg, GluN1KD, and GluN1KD-WAVE mice. Blots were normalized to GAPDH loading control. Indicated in bars, a mouse sample size of 6 was used for each genotype. Quantified WAVE-1 was analyzed by one-way ANOVA. b) Apical (top) and basolateral (bottom) dendritic spine analyses are shown with representative images (left) and

90 quantifications (right). Mice with a GluN1KD genotype had less apical and basolateral spines than non-GluN1KD littermates, but no significant genotype effect was found by one-way ANOVAs. Sample sizes are indicated within each bar. For each mouse, 3-6 dendrite images were analyzed. Scale bars represent 20µm for dendrite images.

3.6 RNAseq Analyses Find Sex-Dependent GluN1KD Gene Expression Changes, Likely Involving Sex Hormone Signaling

3.6.1 GluN1KD Gene Expression Changes from RNAseq Show Internal Validity and Depend on the Sex of Mice

To characterize the striatal transcriptome of GluN1KD mice, we used RNAseq analyses following a modified tuxedo protocol (Trapnell et al., 2012). When comparing WT and GluN1KD mice with sexes combined (sample size of 6 for each group), expression of 235 genes was significantly different. We also performed sex-factored RNAseq analyses to consider sex differences in GluN1KD gene expression changes, comparing WT males, WT females, GluN1KD males, and GluN1KD females (sample size of 3 for each group). Sex-factored RNAseq identified 784 genes with changed expression in GluN1KD mice for at least one sex. As an internal validation of these results, we checked the expression of Grin1 and sex-linked genes in test mice. Shown in Table III-I, the expression of checked genes was consistent with expectations: Grin1 was low in both male and female GluN1KD mice, the X-inactivation effector gene Xist was near-silent in male test mice, and Y-linked genes were near-silent in female test mice. Cross-referencing both sets of RNAseq results found an overlap of 195 genes, as illustrated in Figure 3.9. Thus, most differentially-expressed genes identified by non-sex- factored RNAseq were also identified via sex-factored RNAseq, another good marker of consistency for our results.

Table III-I. RNAseq-quantified expression of Grin1, Xist, and Y-linked genes demonstrate internal validity of results. Grin1 is decreased in GluN1KD mice of both sexes, Xist is near- absent in male mice of both genotypes, and Y-linked genes (Eif2s3y, Ddx3y, Kdm5d, and Uty) are near-absent in female mice of both genotypes. Quantifications are in units of FPKM – number of Fragments Per Kilobase of transcript per Million mapped reads. Gene full names: Xist – X-inactive specific transcript; Eif2s3y – eukaryotic translation initiation factor 2 subunit 3, Y- linked; Ddx3y – DEAD-Box Helicase 3, Y-linked; Kdm5d – lysine demethylase 5D; Uty – ubiquitously transcribed tetratricopeptide repeat containing, Y-linked.

Gene WT Male GluN1KD Male WT Female GluN1KD Female Grin1 139.373 60.896 130.084 66.430

Eif2s3y 23.106 22.082 0.002 0.007 Ddx3y 15.048 16.518 0.004 0.004 Kdm5d 5.265 5.912 0.005 0.003 Uty 4.376 4.348 0.006 0.004

Xist 0.052 0.041 87.014 88.523

Figure 3.9. Sex-factored RNAseq identified more GluN1KD gene expression changes, including most non-sex-factored RNAseq results. 195 differentially-expressed genes were identified by both analyses, 589 were uniquely identified by sex-factored analysis, and 40 were unique to non-sex-factored analysis.

Our RNAseq results point to sex being a key factor for transcriptomic comparisons, at least in the context of GluN1KD gene expression changes. Firstly, also illustrated in Figure 3.9, sex- factored RNAseq uniquely identified 589 differentially-expressed genes along with most of the non-sex-factored RNAseq results. Most of the genes identified from sex-factored RNAseq also varied in how they were changed by GluN1 knock-down depending on the sex of the mice. As shown in Table III-II, only 35 of the 784 GluN1KD gene expression changes were in the same direction when comparing male and female mice. The rest were different between mice of the

93 opposite sex, with gene expression changes being present only in one sex or oppositely affected by GluN1 knock-down. GluN1KD transcriptomic changes at the STR were greatly sex- dependent.

Table III-II. Breakdown of GluN1KD gene expression changes by sex and directionality of changes. Totaled differentially-expressed genes in GluN1KD mice were relative to sex-matched WT littermates. 749 of 784 genes identified by sex-factored RNAseq showed sex-dependency. 35 GluN1KD gene expression changes were in the same direction for both sexes.

Increased in Male Not Changed in Male Decreased in Male Genes of Interest: 784 GluN1KD Mice GluN1KD Mice GluN1KD Mice Increased in Female 2 161 48 GluN1KD Mice Not Changed in Female 192 25767 119 GluN1KD Mice Decreased in Female 28 201 33 GluN1KD Mice

To help us further understand sex-dependency of GluN1KD gene expression changes, we filtered our results to genes that were also significantly different between WT males and females by sex-factored RNAseq analysis. Out of 784 genes, expression of 293 were significantly different between male and female WT mice – 156 “Male Genes” were higher in WT males while 137 “Female Genes” were higher in females. As shown in Tables III-III and III-IV, GluN1KD transcription changes (relative to sex-matched WT littermates) of Male Genes, were only decreases in GluN1KD males, increases in GluN1KD females, or both. Similarly, expression of Female Genes was only decreased in GluN1KD females, increased in GluN1KD males, or both. In other words, when focusing on Male and Female Genes, male GluN1KD mice are demasculinized or feminized while female GluN1KD mice are masculinized or defeminized. The gene changes showing demasculinization or feminization patterns in GluN1KD males and those showing masculinization or defeminization patterns in GluN1KD females are listed in Supplemental Table SII. Over all, we have strong evidence of sex-dependency in GluN1KD gene expression changes. We therefore focused our post-RNAseq gene list and pathways analyses on sex-factored RNAseq results.

Table III-III. Breakdown of GluN1KD gene expression changes for “Male Genes” by sex and directionality of changes. Male Genes refer to those out of the 784 sex-factored RNAseq results that were also significantly higher in WT males than WT females. Totaled differentially- expressed genes in GluN1KD mice were relative to sex-matched WT littermates. Expression of Male Genes changed by GluN1 knockdown was limited to downregulation in male mice and upregulation in female mice.

Increased in Male Not Changed in Male Decreased in Male Male Genes: 156 GluN1KD Mice GluN1KD Mice GluN1KD Mice Increased in Female 0 58 45 GluN1KD Mice Not Changed in Female 0 54 53 GluN1KD Mice Decreased in Female 0 0 0 GluN1KD Mice

Table III-IV. Breakdown of GluN1KD gene expression changes for “Female Genes” by sex and directionality of changes. Female Genes refer to those out of the 784 sex-factored RNAseq results that were also significantly higher in WT females than WT males. Totaled differentially- expressed genes in GluN1KD mice were relative to sex-matched WT littermates. Expression of Female Genes changed by GluN1 knockdown was limited to upregulation in male mice and downregulation in female mice.

Increased in Male Not Changed in Male Decreased in Male Female Genes: 137 GluN1KD Mice GluN1KD Mice GluN1KD Mice Increased in Female 0 0 0 GluN1KD Mice Not Changed in Female 62 104 0 GluN1KD Mice Decreased in Female 26 49 0 GluN1KD Mice

3.6.2 Gene List Analyses Predict that Sex-Dependent GluN1KD Transcriptomic Changes are Related to Sex Hormone Signaling and Lead to Convergent Outcomes

To help us further understand and interpret the RNAseq-identified gene expression changes, we used IPA (Krämer et al., 2014) and Enrichr (Chen et al., 2013; Kuleshov et al., 2016). Pathways analysis with IPA reported multiple curated canonical pathways, upstream regulators, diseases, biological functions, and predicted networks that were significantly over-enriched in our results, some of which were associated with predicted activation scores. In the context of our investigation into sex-differences between male and female GluN1KD transcriptomic changes, many IPA results were different between males and females. Representative sex-dependent IPA results are shown with causal network analysis results in Tables III-V and III-VI and constructed networks in Figure 3.10.

Causal network analyses report upstream regulators most associated with input genes and predict the activation states of these regulators based on input gene expression changes. As shown in Tables III-V and III-VI, two of the top three causal network analysis results for male and female GluN1KD gene expression changes (with significant activation scores for at least one sex) are sex dependent. Neural cell adhesion molecule 2 (Ncam2) regulates 27 genes from our inputs and, based on the specific expression changes of these 27 genes relative to sex-matched WT mice, was predicted to be activated in GluN1KD males (activation z-score = 2.18) but inhibited in females (z = -2.50). Estrogen receptor α (ERα; Esr1 is its encoding gene, which IPA continues to use in its outputs since genes were initially entered) was also identified as a significant regulator of GluN1KD-changed gene expression (p = 5.94 x 10-12 for differentially- expressed genes in both male and female GluN1KD mice), particularly interesting as it is an effector of estrogen signaling. Based on the sex-dependent expression changes of 17 genes it regulates, Esr1 is predicted to be activated in GluN1KD males (z = 4.12) and neither activated or inhibited in females (z = -1.70). Low density lipoprotein receptor-related protein 1 (Lrp1) was predicted to be inhibited in both GluN1KD males and females, but with different activation scores (z = -2.58 and -2.32, respectively) based on different gene expression changes like Wnt9a and Wnt5a.

Figure 3.10 further illustrates the sex-dependency of GluN1KD transcriptomic changes via IPA-constructed networks that include Esr1 as a hub based on connectivity. In male GluN1KD

96 mice, genes connected to Esr1 have increased expression while a connected hub centered on histone H3 contains primarily downregulated genes on the right and primarily upregulated genes on the left. These patterns are largely flipped in female GluN1KD mice: most genes connected to Esr1 are downregulated, the right half of the H3 hub largely consists of upregulated genes, and the left half of the H3 hub mainly contains downregulated genes. Taken together, our IPA results also point to a strong effect of sex on GluN1KD gene expression changes. Esr1, in particular, may be a strong intermediator between mouse sex and GluN1 knockdown.

Table III-V. Representative results from IPA causal network analysis of GluN1KD gene expression changes in male mice. Listed are the top three upstream regulators that are related to male GluN1KD gene expression changes and that have significant activation z-scores.

Upstream Activation Gene list over- Predicted upstream GluN1KD gene expression changes that GluN1KD gene expression changes that regulator z-score representation p-value regulator state predict activation predict inhibition

↑ Ttr, ↑ Tmem72, ↑ Sulf1, ↑ Sostdc1, ↓ Ptgs2, ↓ Klf2, ↓ Ido1, ↓ Fos, ↓ Egr4, ↓ Egr2, ↑ Slc13a4, ↑ Rdh5, ↑ Otx2, ↑Malat1, ↑ Lbp, Ncam2 2.117 5.23E-12 Activated ↓ Egr1, ↓ Cdh9 ↑ Kl, ↑ Igf2, ↑ Gal, ↑ Folr1, ↑ Enpp2, ↑ Col8a1, ↑ Clic6, ↑ Cldn2, ↑ Cldn1, ↑ Ace ↑ Ttr, ↑ Tmem72, ↑ Sulf1, ↑Sostdc1, ↑ Slc13a4, ↑ Rdh5, ↑ Otx2, ↑ Lbp, ↑ Kl, ↑ Igf2, Esr1 4.123 5.94E-12 Activated None ↑ Gal, ↑ Folr1, ↑ Enpp2, ↑ Col8a1, ↑ Clic6, ↑ Cldn2, ↑ Ace ↓ Wnt9a, ↑ Trp73, ↑ Tgfb2, ↓ Slc8a2, ↑ Serpine1, ↑ Postn, ↓ Per1, ↑ Pcolce, ↑ Wnt5a, ↑ Sfrp2, ↓ Serpinb1a, ↓ Ptgs2, ↓ Nr4a2, ↓ Nr4a1, ↓ Neurl1b, ↓ Ndfip2, ↑ Plin3, ↓ Pdp1, ↑ Mki67, ↓ Lsp1, ↓ Klf2, ↓ Nab2, ↓ Myl4, ↓ Me2, ↑ Malat1, ↑ Mag, Lrp1 -2.582 2.59E-11 Inhibited ↓ Ier5, ↑ Hsp90b1, ↓ Hba-a2, ↑ Gfap, ↓ E2f1, ↓ Junb, ↓ Islr2, ↑ Igf2, ↓ Ido1, ↓ Hsd11b1, ↓ Dusp1, ↓ Dkk3, ↓ Col6a2, ↓ Bok, ↓ ↓ Grin1, ↓ Egr4, ↓ Egr2, ↓ Egr1, ↓ Dusp14, Adcyap1, ↑ Ace ↓ Dlgap3, ↑ Col5a2, ↑ Clic4, ↓ Cck, ↓ Camk2n1, ↑ Cab39l, ↓ Btg2, ↓ Bhlhe40, ↓ Arc, ↑ Aqp1, ↓ Apold1, ↑ Amotl2, ↓ Adcy1

Table III-VI. Representative results from IPA causal network analysis of GluN1KD gene expression changes in female mice. Listed are the top three upstream regulators that are related to male GluN1KD gene expression changes and that have significant activation z-scores.

Upstream Activation Overlap Predicted upstream GluN1KD gene expression changes that GluN1KD gene expression changes that regulator z-score p-value regulator state predict activation predict inhibition ↓ Ttr, ↓ Tmem72, ↓Sostdc1, ↓ Otx2, ↓ Klf2, ↑ Sulf1, ↑ Slc13a4, ↑ Rdh5, ↑ Ptgs2, ↑Malat1, ↓ Igf2, ↓ Ido1, ↓ Gal, ↓ Fos, ↓ Folr1, ↓ Enpp2, Ncam2 -2.502 5.23E-12 Inhibited ↑ Lbp, ↑ Kl ↓ Egr4, ↓ Egr2, ↓ Egr1, ↓ Col8a1, ↓Clic6, ↓ Cldn2, ↓ Cldn1, ↓ Cdh9, ↓ Ace ↓ Ttr, ↓ Tmem72, ↓Sostdc1, ↓ Otx2, ↓ Igf2, Neither activated nor Esr1 -1.698 5.94E-12 ↑ Sulf1, ↑ Slc13a4, ↑ Rdh5, ↑ Lbp, ↑ Kl ↓ Gal, ↓ Folr1, ↓ Enpp2, ↓ Col8a1, ↓Clic6, inhibited ↓ Cldn2, ↓ Ace ↓ Wnt5a, ↑ Trp73, ↑ Tgfb2, ↓ Sfrp2, ↑ Serpine1, ↑ Ptgs2, ↑ Postn, ↓ Plin3, ↓ Per1, ↑ Wnt9a, ↑ Slc8a2, ↓ Serpinb1a, ↓ Pdp1, ↑ Pcolce, ↓ Nr4a1, ↓ Ndfip2, ↓ Me2, ↑ Malat1, ↑ Nr4a2, ↑ Neurl1b, ↑ Nab2, ↑ Myl4, ↑ Mki67, ↑ Lsp1, ↓ Junb, ↑ Ier5, ↓ Ido1, ↓ Hsp90b1, Lrp1 -2.324 2.59E-11 Inhibited ↓ Mag, ↓ Klf2, ↑ Islr2, ↓ Igf2, ↓ Hba-a2, ↓ Hsd11b1, ↓ Grin1, ↓ Gfap, ↓ Egr4, ↓ Egr2, ↓ Dusp1, ↓ Col6a2, ↓ Clic4, ↑ Cck, ↓ Cab39l, ↓ Egr1, ↑ E2f1, ↓ Dusp14, ↓ Dlgap3, ↑ Dkk3, ↓ Amotl2, ↑ Adcy1 ↑ Col5a2, ↓ Camk2n1, ↓ Btg2, ↑ Bok, ↓ Bhlhe40, ↓ Arc, ↑ Aqp1, ↓ Apold1, ↑ Adcyap1, ↓ Ace

Figure 3.10. IPA-constructed, Esr1 (implied ERα)-centric, networks of male (left) and female (right) GluN1KD gene expression changes have the same structures but are largely opposite in regulation. The legend for network components is shown on the bottom. Double- lined shapes represent complexes. Red shapes indicate upregulation while green shapes indicate downregulation. Shading reflects magnitude of change. Solid lines represent experimentally- determined direct relationships while dotted lines represent indirect relationships. Arrows denote direction of relationship – the component at the beginning of the arrow affects the component at the end.

Pathways analyses and gene list analyses are also useful for determining the biological function and relevance of sets of genes and gene changes. Gene list analyses performed with Enrichr are shown in Tables III-VII and III-VIII and show similar results despite four different input lists – differentially-expressed genes separated by sex and direction of change (increased or decreased, in GluN1KD males or females). In the context of diseases in Table III-VII, Huntington’s disease (HD) records show up frequently amongst the top 5 results for each list’s analysis. Table III-VIII contains Enrichr results based on chemical-induced gene expression changes, particularly focusing on what chemical-induced changes could reverse those in our input lists. Similar to Table III-VII, Table III-VIII also has top results that are common to

98 multiple gene lists’ analyses – records involving withaferin A. Furthermore, though the other records of Table III-VIII largely involve different drugs, many are used for similar purposes, particularly for cancer treatment. These gene list analysis results are an illustration of different gene expression change patterns converging to similar biological significance and outcomes, at least in the context of NMDAR deficits in mice.

Table III-VII. Enrichr gene list analysis results associate similar disease-related records with multiple input lists. Differentially-expressed genes from sex-factored RNAseq were separated into four input lists based on mouse sex and direction of change. Analyses referenced “Disease Perturbations from GEO up” and “Disease Perturbations from GEO down” datasets for upregulated and downregulated genes in GluN1KD mice, respectively. Gene-set records related to HD appeared often in the results of all four lists’ analyses. For each gene list analysis, the top 5 results based on (and ordered by) Enrichr combined scores are shown. Enrichr reported p- values were adjusted by BH method for multiple comparisons. Denominators of gene overlap ratios represent the number of genes in each matched Enrichr record.

Adjusted Gene Input gene list Matched record from Enrichr Overlapping genes (between input gene list and Enrichr record) p-value overlap Acss3, Pde1c, Myof, Ecel1, Wnt5a, Igfbp2, Ebf3, Zic4, Dlk1, Increased in Males Schizophrenia DOID-5419 human 1.5E-09 21/263 Baiap3, Gabrg1, Calcr, Zic2, Peg10, Smoc1, Syt10, Zic1, (222 genes) GSE25673 sample 892 Pip5k1b, Plch1, Slc16A9, Spata18 Clic6, Kl, Slc13a4, 1500015O10Rik, Ecel1, Igfbp2, Pcolce, F5, Increased in Males Huntington's disease DOID-12858 9.6E-08 20/306 Ttr, Enpp2, Col8a2, Slc29a4, Slc17a6, Sostdc1, Otx2, Lbp, (222 genes) mouse GSE19780 sample 527 Cbln4, Gprasp2, Folr1, Vat1l Clic6, 1500015O10Rik, Abca4, Igfbp2, Foxj1, Wdr86, Cldn2, Increased in Males Depression DOID-1595 mouse 4.6E-06 18/327 Aqp1, F5, Cldn11, Ttr, Enpp2, Col8a2, Sostdc1, Col9a3, Otx2, (222 genes) GSE26025 sample 463 Lbp, Folr1 Zcchc12, Trpc5, Gad2, Ebf3, Slc6A11, Tmem130, Usp29, Increased in Males Spinal Muscular Atrophy C0026847 1.6E-06 20/368 Gfap, Gabrg1, Cldn11, Gja1, Tppp3, Gjb6, Resp18, Slc17A6, (222 genes) mouse GSE10599 sample 235 Hap1, A2M, Dgkk, Scg2, Vat1L Itih3, Postn, Sparc, Sdc4, Myof, Wnt5A, Lamb1, Pcolce, Increased in Males pancreatic ductal adenocarcinoma DOID- 1.5E-05 21/487 Cldn2, Cldn1, Agt, Thbs4, F5, Sfrp2, Rnf128, Gjb6, Col8A2, (222 genes) 3498 human GSE15471 sample 604 Col6A3, Col8A1, Gabre, Steap2 Nrn1, Hpca, 3110035E14Rik, Cplx2, Camkk2, Dusp14, Nrgn, Jph4, Rph3a, Rgs2, Sv2b, Camk2n1, Basp1, Efhd2, Slc17a7, Decreased in Males Huntington's disease DOID-12858 1.5E-24 38/371 Nptx1, Fam81a, Egr1, Lingo1, Mef2c, Egr3, Rbfox3, Dusp1, (200 genes) mouse GSE3621 sample 703 Cck, Fos, Dgkz, Sept5, Dusp6, Olfm1, Nr4A1, Arc, Cacnb3, Rab15, Itpka, Bhlhe40, Vip, Ctxn1, Stx1a Cpne9, Nrn1, Hpca, 3110035E14Rik, Cplx2, Camkk2, Dusp14, Nrgn, Jph4, Rph3a, Sv2b, Camk2n1, Efhd2, Slc17a7, Nptx1, Decreased in Males Huntington's disease DOID-12858 3.2E-23 36/356 Fam81a, Prkcg, Egr1, Lingo1, Mef2c, Egr3, Rbfox3, Dusp1, (200 genes) mouse GSE3621 sample 704 Stard10, Cck, Fos, Dgkz, Sept5, Olfm1, Nr4a1, Arc, Itpka, Bhlhe40, Vip, Ctxn1, Stx1a Nrn1, Hpca, 3110035E14Rik, Cplx2, Camkk2, Dusp14, Nrgn, Decreased in Males Huntington's disease DOID-12858 Sv2b, Camk2n1, Basp1, Efhd2, Slc17a7, Nptx1, Fam81a, 8.3E-22 32/289 (200 genes) mouse GSE3621 sample 702 Prkcg, Egr1, Lingo1, Mef2c, Egr3, Dusp1, Cck, Fos, Dgkz, Sept5, Olfm1, Nr4a1, Arc, Itpka, Bhlhe40, Vip, Ctxn1, Stx1a Cpne9, Hpca, Dusp14, Rph3A, Neurod6, Dnajb5, Rspo2, Decreased in Males Alzheimer's disease DOID-10652 human Nptx1, Dlgap3, Prkcg, Klf10, Lingo1, Egr3, Bdnf, Rtn4Rl2, 2.6E-13 27/380 (200 genes) GSE36980 sample 519 Dusp6, Sytl2, Nr4A2, Cacnb3, Efna3, Nr4A3, Nov, Fam163B, Rab15, Rimbp2, Fxyd7, Vip glaucoma associated with systemic Cpne9, Nrn1, Rbfox3, Satb2, Hpca, Islr2, Tbr1, Gpr3, 43348, Decreased in Males syndromes DOID-1686 mouse 3E-09 21/341 Sytl2, Jph4, Golga7B, Rph3A, Olfm1, Cacnb3, Pcsk1N, (200 genes) GSE26299 sample 491 Rab15, Cyp2S1, Adora1, Fxyd7, Rprm Rtn4r, Prdm8, Nrn1, Fhl2, 3110035E14Rik, Adcy1, Ak5, Crhbp, Increased in Females Huntington's disease DOID-12858 Neurod2, Neurod6, Creg2, Rspo2, Efhd2, Slc17a7, Nptx1, 1.3E-20 32/288 (211 genes) mouse GSE9038 sample 717 Igfbp6, Fezf2, Rprm, Lingo1, Slc30a3, Bdnf, Plekhg5, Islr2, Cck, Prss12, C1ql3, Slc7a4, Nov, Stac2, Stx1a, Cbln2, Tuba8 Increased in Females Huntington's disease DOID-12858 Rtn4r, Nxph3, Plekhg5, Islr2, Adcy1, Prss12, Slc7a4, Cyp4f15, 0.00018 16/328 (211 genes) mouse GSE19780 sample 528 Nov, Neurod6, Basp1, Stac2, Cpne2, Fezf2, Tuba8, Rprm Increased in Females multiple sclerosis DOID-2377 human Rap1gap2, Mef2c, Nrn1, Rpl10, Gda, Cck, Miat, Ak5, Dkk3, 0.0003 15/310 (211 genes) GSE38010 sample 737 Olfm1, Sv2b, Basp1, Slc17a7, Nptx1, Serpini1 Increased in Females multiple sclerosis DOID-2377 human Rap1Gap2, Mef2C, Nrn1, Cck, Adcy1, Miat, Ak5, Dkk3, Olfm1, 0.00111 14/309 (211 genes) GSE38010 sample 738 Sv2B, Basp1, Slc17A7, Nptx1, Serpini1 Increased in Females endometriosis DOID-289 human Col11A1, Serpine1, Fhl2, Sulf1, Ctgf, Nr4A2, Smoc2, Mafb, 0.00111 14/315 (211 genes) GSE23339 sample 883 Col5A1, Rbp1, Col5A2, Col6A1, Rplp2, Ccdc3 Pmch, Gpr88, Hpca, Hba-a2, Rasgrp2, Rerg, Klk6, Rasd2, Wolfram syndrome UMLS CUI- Decreased in Females Rgs4, Adamts4, Drd2, Gpr155, Rxrg, Egr1, Egr2, Gpx6, C0043207 mouse GSE15293 sample 4.2E-15 32/350 (262 genes) Kcnab1, Ano3, Sh3rf2, Pcp4, Mag, Inf2, Pde10a, Adora2a, 695 Cyp2s1, Aldh1a1, Gm9989, Rgs7bp, Scn4b, Kcnk2, Ido1, Lhx8 Tnfaip6, Pmch, Lpar1, Ndrg1, Hsp90b1, Adamts4, Gpr155, Decreased in Females Niemann-Pick disease Type C DOID- Plekhh1, Bcas1, Ppp1r16b, Egr1, Car2, Hspa8, Egr3, Gpr37, 7.5E-15 29/290 (262 genes) 14504 mouse GSE39621 sample 469 Hspa5, Kcnab1, Zdhhc14, Ermn, Pdp1, Qdpr, Arc, Pcp4, Mag, Inf2, Rgs7bp, Calr, Cryab, Scn4b Pmch, Hpca, Hba-a2, Camkk2, Hsp90b1, Dusp14, Rgs4, Decreased in Females Huntington's disease DOID-12858 Rgs2, Ai593442, Camk2n1, Nefm, Cacng3, Egr1, Car2, Xbp1, 7.7E-15 32/371 (262 genes) mouse GSE3621 sample 703 Egr3, Hspa5, Kcnab1, Sepp1, Agt, Pdp1, Qdpr, Nr4a1, Arc, Pcp4, Pde10a, Vamp1, Rgs7bp, Calr, Scn4b, Hspa1b, Pvalb

Pmch, Hpca, Hba-A2, Camkk2, Dusp14, Rgs4, Ttr, Camk2N1, Decreased in Females Huntington's disease DOID-12858 Cnr1, Nefm, Egr1, Il33, Car2, Egr3, Kcnab1, Sepp1, Agt, Pdp1, 8.9E-13 29/356 (262 genes) mouse GSE3621 sample 704 Qdpr, Nr4A1, Arc, Pcp4, Aldh1A1, Vamp1, Rgs7Bp, Calr, Scn4B, Pvalb, Cbln1 Stc1, Magt1, Ndrg1, Apold1, Hsp90B1, Dusp14, Akap12, Decreased in Females Papillary Carcinoma of the Thyroid Rgs2, Ubb, Ubc, Ptgds, Plekhh1, Egr1, Il33, Egr2, Hspa5, 2.5E-12 26/292 (262 genes) C0238463 human GSE3678 sample 306 Kcnab1, Pdia6, Pdia4, Arc, Pcp4, Aldh1A1, Fosb, Amotl2, Calr, Cryab

100

Table III-VIII. Enrichr gene list analysis results associate similar chemical-perturbation records with multiple input lists. Differentially-expressed genes from sex-factored RNAseq were separated into four input lists based on sex and change direction. Analyses referenced “LINCS L1000 Chem Pert down” and “LINCS L1000 Chem Pert up” datasets for upregulated and downregulated genes in GluN1KD mice, respectively. Opposite-directionality was chosen to identify drugs (in bold) that may reverse GluN1KD transcriptomic changes. Enrichr gene-set records related to withaferin A were identified in multiple gene list analyses. For each gene list analysis, the top 5 results based on (and ordered by) Enrichr combined scores are shown. Enrichr reported p-values were adjusted by BH method for multiple comparisons. Denominators of gene overlap ratios represent the number of genes in each matched Enrichr record.

Adjusted Gene Input gene list Matched record from Enrichr Overlapping genes (between input gene list and Enrichr record) p-value overlap Increased in Males LJP008_HT29_24H-KU-55933-10 0.00272 6/28 (222 genes) Postn, Gja1, Sparc, Peg10, Pdgfd, Wnt5A Increased in Males LJP006_LNCAP_3H-withaferin-a-3.33 0.00272 7/39 (222 genes) Cpm, Postn, Sparc, Peg10, Igfbp2, Scg2, Folr1 Increased in Males LJP008_HEPG2_24H-IOX2-10 0.0036 6/31 (222 genes) Tcf7L2, Igfbp2, Myof, Enpp2, Gabre, Esr1 Increased in Males LJP006_BT20_24H-KIN001-043-0.04 0.00272 8/64 (222 genes) Tppp3, Dab2, Sparc, Peg10, Pdgfd, Tsc22D3, Sgk1, Dlk1 Increased in Males LJP005_MCF10A_24H-PHA-665752-10 0.01044 6/43 (222 genes) Cpm, Postn, Enpp2, Lamb1, Msx1, Sepp1 Btg2, Satb2, Fhl2, Ctgf, Dusp14, Rgs2, Junb, Ier5, Ier2, Klf10, Decreased in Males LJP006_SKBR3_24H-withaferin-a-3.33 8.6E-17 26/213 Egr1, Egr3, Egr4, Dusp1, Fos, Coq10B, Klf2, Fosl2, Nr4A2, (200 genes) Nr4A1, Per1, Arc, Nr4A3, Bhlhe40, Fosb, Trib1 Klf10, Egr1, Btg2, Egr2, Nrn1, Egr3, Dusp1, Apold1, Klf2, Decreased in Males LJP006_HME1_3H-QL-XII-47-1.11 8.6E-17 22/135 Dusp6, Fosl2, Nr4A2, Nr4A1, Per1, Rgs2, Arc, Nr4A3, (200 genes) Bhlhe40, Fosb, Junb, Ier5, Ier2 Klf10, Egr1, Btg2, Egr2, Egr3, Dusp1, Fos, Klf2, Dusp6, Ctgf, Decreased in Males LJP006_HME1_3H-QL-XII-47-3.33 8.6E-17 22/135 Fosl2, Nr4A2, Nr4A1, Per1, Rgs2, Nr4A3, Bhlhe40, Fosb, (200 genes) Junb, Ier5, Stx1A, Ier2 Egr1, Ighm, Btg2, Mef2C, Egr3, Dusp1, Cck, Fos, Dusp6, Decreased in Males LJP005_HA1E_24H-withaferin-a-3.33 4.9E-16 21/130 Nr4A2, Nr4A1, Rgs2, Arc, Nr4A3, Frzb, Camk2N1, Fosb, Trib1, (200 genes) Junb, Ier5, Ier2 Klf10, Egr1, Btg2, Egr2, Egr3, Egr4, Dusp1, Fhl2, Fos, Klf2, Decreased in Males LJP006_MCF7_24H-withaferin-a-10 4.9E-16 22/149 Ctgf, Nr4A2, Nr4A1, Rgs2, Arc, Nr4A3, Bhlhe40, Fosb, Junb, (200 genes) Ier5, Stx1A, Ier2 Increased in Females Col5A1, Col11A1, Serpine1, Col5A2, Rbp1, Fhl2, Mical2, Sulf1, LJP006_BT20_24H-dasatinib-0.04 3E-05 10/66 (211 genes) Ctgf, Dkk3 Increased in Females LJP008_MCF7_24H-(+)-JQ1-0.04 0.02173 6/47 (211 genes) Nr4A2, Col5A1, Col5A2, Col6A1, Trbc1, Fhl2 Increased in Females LJP009_HA1E_24H-BI-2536-10 0.01063 8/75 (211 genes) Nrn1, Lgals1, Col5A1, Col11A1, Serpine1, Mical2, Prss23, Ctgf Increased in Females LJP007_HA1E_24H-HG-9-91-01-1.11 0.02173 5/27 (211 genes) Nrn1, Serpine1, Mical2, Prss23, Ctgf Increased in Females LJP005_MCF10A_3H-PD-0325901-0.37 0.02173 7/68 (211 genes) Lgals1, Col11A1, Fhl2, Mical2, Sox11, Stx1A, Mylk Decreased in Females Xbp1, Sdf2L1, Hspa5, Amigo2, Cdk2Ap2, Calr, Magt1, Creld2, LJP005_MCF10A_3H-NVP-AUY922-10 5.8E-06 11/59 (262 genes) Pdia6, Pdia4, Hsp90B1 Decreased in Females LJP005_MDAMB231_3H-NVP-AUY922- Hspa8, Xbp1, Sdf2L1, Hspa5, Aldh1A1, Cdk2Ap2, Calr, Creld2, 1.2E-05 11/67 (262 genes) 10 Coch, Pdia4, Hsp90B1 Decreased in Females LJP005_MCF10A_3H-geldanamycin- Hspa8, Xbp1, Sdf2L1, Hspa5, Magt1, Pdia6, Pdia4, Hsp90B1, 2.1E-05 13/117 (262 genes) 3.33 Pdzd2, Tnnt1, Cdk2Ap2, Calr, Creld2 Decreased in Females LJP005_HCC515_24H-NVP-AUY922- Hsd11B1, Hspa8, Rgs2, Rcn1, Xbp1, Ddc, Mme, Hspb8, 4.9E-05 11/89 (262 genes) 0.04 Cryab, Sepp1, Hsp90B1 Decreased in Females Hspa8, Egr1, Egr2, Pfkfb3, Egr3, Hspb8, Rnd3, Fosl2, Akap12, LJP005_SKBR3_3H-withaferin-a-3.33 1.8E-05 14/133 (262 genes) Nr4A1, Arc, Fosb, Nefm, Cryab

101

Chapter 4 Discussion Discussion 4.1 Developmentally Correlated Synaptic and Molecular Deficits in GluN1KD Mice

As the overall aim of this study is to investigate the relationship between Rho GTPase signaling pathways, synaptic connectivity, and behaviour in GluN1KD mice, we first sought to more firmly establish the link between dendritic spine density and behaviour. Previously, GluN1KD studies have reported MSN spine density deficits at 6 weeks of age but not at 2 weeks (Ramsey et al., 2011). This deficit was developmentally aligned with the onset of some GluN1KD behavioural abnormalities at 6 weeks of age (Milenkovic et al., 2014). However, while spine density studies were conducted on two-week old mice, the behavioural studies were conducted on three-week old mice (Ramsey et al., 2011; Milenkovic et al., 2014). A one-week- age difference for the youngest mice in these studies may seem minimal, but this peri-weaning age gap may be developmentally important; female open field behaviour and male PPI have been shown to change between 2- and 3-weeks of age (Moy et al., 2012). We therefore sought to more definitively align MSN spine density and GluN1KD behavioural abnormalities by counting MSN spines at 3-, 6-, and 12-weeks of age.

As shown in Figure 3.1, GluN1KD mice started having significantly lower spine densities compared to WT littermates at 6-weeks of age, which persisted by 12-weeks of age. These results strongly mirror the developmental trajectories of GluN1KD behavioural abnormalities that manifest at 6 weeks of age (Milenkovic et al., 2014). Such a correlation is likely biologically relevant, as loss of synapses has also been associated with fear conditioning (Lai et al., 2012; Sanders et al., 2012) and cognitive impairment in AD (Tönnies and Trushina, 2017). More research will be needed to specifically understand how MSN spine changes contribute to GluN1KD behavioural phenotypes. The results presented in Figure 3.1 also narrow the possible developmental window for MSN spine loss in GluN1KD mice from the previous 2-6 weeks of age (Ramsey et al., 2011) to 3-6 weeks. Taken together, both age-related spine studies further suggest that synaptic deficits in GluN1KD mice are more likely due to problems with synaptic elimination than synaptogenesis. This could guide future research into the neurodevelopment of

102

GluN1KD brains. Overall, we have more fully characterized and established the course of age- dependent spine loss in GluN1KD MSNs.

We also report evidence for dysregulated Rho GTPase signaling in GluN1KD mice. Specifically, two proteins were consistently altered in GluN1KD mice at all three ages: Rac1 and WAVE-1. As mentioned before, changes in these two specific proteins are consistent with the previous reports of DISC1 deficits in GluN1KD mice (Ramsey et al., 2011). This is because DISC1 regulates the GEFs that activate Rac1 (Hayashi-Takagi et al., 2010; Chen et al., 2011), which in turn acts on WAVE-1 (Eden et al., 2002; Soderling et al., 2003; Kim et al., 2006b). Antagonism of NMDARs with PCP has also been reported to decrease cortical WAVE-1 in rats (Andrews et al., 2015) consistent with our findings and suggesting that WAVE-1 decreases in GluN1KD mice are indeed related to NMDAR hypofunction. The fact that primarily Rac1- related signaling is aberrant in GluN1KD mice is consistent with current, albeit limited, evidence for the relationship between DISC1 and Rho GTPases. RhoA, Cdc42, and Rac1 are the most characterized and emphasized Rho GTPases when it comes to cytoskeleton rearrangement during neuronal and synaptic development (Hedrick and Yasuda, 2017; Huang et al., 2017). Of the three, Rac1 has generally been considered the main interactor with DISC1 through Kalirin-7 (Dillon and Goda, 2005; Hayashi-Takagi et al., 2010; Saneyoshi and Hayashi, 2012), as illustrated with Figure 4.1. Proteomic screening of NMDAR-associated complexes in the mouse brain has also specifically reported the presence of Rac1 but not RhoA or Cdc42 (Husi et al., 2000). In general, Rac1 may be the most relevant or affected of the three Rho GTPases in a context of NMDAR and DISC1 deficiency. More specific experiments and focused studies also reinforce this idea.

103

Figure 4.1. Representative schematic of the Rho GTPase signaling pathways related to cytoskeletal rearrangement as part of synaptic plasticity. Green pointed arrows indicate a positive relationship of some form, such as enhanced activation or release from a sequestering factor. Red blunted lines indicate the opposite, causing some form of inhibition or decreased activity. Black arrows indicate a more complex regulatory relationship from mixed or limited experimental results. Abbreviations follow those used in text and previous figures along with newly mentioned proteins: GAP – GTPase-activating proteins; Tiam1 – T-lymphoma invasion and metastasis-inducing protein 1; ROCK – Rho-associated protein kinase; MLC – myosin light

104 chain; MHCIIB – myosin heavy chain IIB. Figure was constructed from previously cited work (Nakayama et al., 2000; Dillon and Goda, 2005; Hayashi-Takagi et al., 2010; Saneyoshi and Hayashi, 2012).

There is evidence of connections between DISC1 and non-Rac1 Rho GTPases, though the evidence is more robust for DISC1 and Rac1. RhoA interacts with dynamin-related protein 1 (Drp1) and its protein level was decreased with knockdown of Drp1 in human glioma cells (Yin et al., 2016). DISC1 may therefore also be indirectly connected to RhoA since its knockdown in glioblastoma cells decreases Drp1, cell proliferation and self-renewal, and cell migration and invasion. The activity of DISC1 in this context was believed to be via mitochondrial mechanisms (Gao et al., 2016). DISC1 has also been shown to regulate the transport of NudE-like (NUDEL) and LIS1 in axons (Taya et al., 2007), which in turn can indirectly activate Cdc42 as part of cell migration mechanisms (Kholmanskikh et al., 2006; Shen et al., 2008). These connections between DISC1 and RhoA or Cdc42 are indirect, experimentally and molecularly, as well as more in the context of cell motility and proliferation than synaptic plasticity (Kholmanskikh et al., 2006; Shen et al., 2008; Gao et al., 2016). In contrast, the connection between DISC1 and Rac1 has been demonstrated more directly in single studies (Hayashi-Takagi et al., 2010; Chen et al., 2011). In both in vitro and in vivo experiments, DISC1 was shown to act at synapses by sequestering Kalirin-7 and inhibiting its activation of Rac1. Upon NMDAR activation, DISC1 and Kalirin-7 dissociate and Rac1 activity increases, resulting in more and larger spines as well as strengthened synapses (Hayashi-Takagi et al., 2010). DISC1 has also been demonstrated to increase Rac1 activity by releasing the GEF domain of triple functional domain protein (TRIO) in the axons of Caenorhabditis elegans and COS cells (Chen et al., 2011). These two studies thus demonstrated within one set of experiments a functional connection between DISC1 and Rac1 (Hayashi-Takagi et al., 2010; Chen et al., 2011). The second study in C. elegans and COS cells focused on the role of DISC1 and Rac1 in axon guidance and may therefore be less relevant to synaptic plasticity in a context of altered NMDAR signaling (Chen et al., 2011). In this sense, the DISC1/Kalirin-7/Rac1 pathway associated with NMDAR activity in dendritic spines may be particularly relevant to GluN1KD mice. It is also interesting to note that RhoA or Cdc42 was not activated by Kalirin-7 when Rac1 was (Hayashi-Takagi et al., 2010); this is also consistent with our findings. Our results could also be improved by looking more specifically at synaptic levels of Rho GTPase signaling proteins. This is because synaptic TH and DAT were previously

105 reported to be decreased in GluN1KD mice while striatal total-tissue levels of these proteins were not (Ferris et al., 2014). Even DISC1 deficits were more pronounced at the synapse (Ramsey et al., 2011). Furthermore, synaptic LIS1 was increased while tissue LIS1 was decreased in 6-week old GluN1KD mice (Ramsey et al., 2011), inviting the question of whether synaptic Cdc42 may be altered. As a whole, however, the current data point to Rac1 signaling deficits in the STR of GluN1KD mice that are consistent with the NMDAR/DISC1/Rac1/ WAVE-1 literature (Soderling et al., 2003; Hayashi-Takagi et al., 2010; Ramsey et al., 2011).

Striatal Rac1 and WAVE-1 abnormalities in GluN1KD mice were also age-dependent, as shown in Figure 3.2. These proteins were both decreased in 6-week old adolescent mice and 12- week old adult mice but not in 3-week old juvenile mice. Specifically, Rac1 was increased while WAVE-1 was decreased in these younger GluN1KD mice. The fact that Rac1 signaling proteins were less overtly deficient in 3-week old mice may be consistent with how DISC1 was also more moderately decreased in 2-week old GluN1KD mice than 6-week olds (Ramsey et al., 2011). In other words, the mixed changes of Rac1 and WAVE-1 in 3-week old GluN1KD mice may be representative of a milder Rac1 signaling abnormality; the increase in Rac1 may be in some way compensatory for decreased DISC1 and WAVE-1. On the other hand, prolonged knockdown of DISC1 for 6 days resulted in sustained increases of Rac1 bound to guanosine triphosphate (GTP) in cultured rat cortical neurons, though total Rac1 level was not changed (Hayashi-Takagi et al., 2010). It would be interesting to study the time-course of Rac1 and WAVE-1 changes in response to chronic DISC1 deficiency in vivo. The transient increase of Rac1 at the GluN1KD STR also brings to mind the altered DA system in GluN1KD mice. Just as experiments with 12- to 15-week old brains found signs of previous hyperdopaminergia (Ferris et al., 2014), the current results present a period of increased Rac1. It is possible that the two findings are related. For example, as overactive VTA neurons project to the NAc in GluN1KD mice (Berton et al., 2006; Ferris et al., 2014), they may also increase the co-release of glutamate (Dal Bo et al., 2004; Stuber et al., 2010; Tecuapetla et al., 2010). This would produce an increased glutamatergic tone, activate the NMDAR/DISC1/Kalirin-7/Rac1 pathway discussed previously (Hayashi-Takagi et al., 2010) and potentially contribute to long-term protein-level changes for Rac1 and WAVE-1.

The mixed-direction changes of Rac1 and WAVE-1 in 3-week old GluN1KD mice, as well as milder DISC1 deficits in 2-week old GluN1KD mice, coincided with WT-like MSN spine density (Ramsey et al., 2011). Chronic Rac1 activation has been repeatedly demonstrated to

106 result in deficient dendritic spines (Luo et al., 1996; Tashiro et al., 2000; Hayashi-Takagi et al., 2010). As we did not see this in 3-week old GluN1KD mice, it is possible that the increase of Rac1, along with decreased WAVE-1, was starting to have a deleterious effect on GluN1KD MSN spines. WAVE-1 deficits are also consistent with dendritic spine disruptions, as it promotes actin polymerization at synapses (Bompard and Caron, 2004; Kim et al., 2006b; Soderling et al., 2007). Both in vitro neuronal studies and in vivo mouse studies have previously demonstrated decreased dendritic spines with WAVE-1 disruption (Kim et al., 2006b; Soderling et al., 2007). Mice with disrupted Arp2/3 complexes, which act with WAVE-1 (Bompard and Caron, 2004; Soderling et al., 2007), similarly demonstrated decreased dendritic spines (Kim et al., 2013). The changes of Rac1 and WAVE-1 are therefore generally consistent with a state of spine loss. However, it is also important to point out that the measures discussed so far are protein levels and spine numbers – not measures of functionality. It is possible that the effects of GluN1KD on the activity of Rac1 and WAVE-1 are not necessarily reflected in protein level changes. Indeed, both Rac1-GTP and phosphorylated WAVE-1, activity-dependent forms of their respective proteins, may represent only a portion of total Rac1 and WAVE-1 (Kim et al., 2006b; Hayashi-Takagi et al., 2010). Functional analyses of Rac1, WAVE-1, and MSN dendritic spines would improve the interpretability of the results in our study.

Together, the changes of Rac1, WAVE-1, and MSN spine density are generally consistent with GluN1KD behavioural abnormalities. The spine density deficits were absent in 3-week old GluN1KD mice, manifested by 6 weeks of age, and persisted into adulthood. Again, these spine density deficits mirrored the trajectories of some GluN1KD behavioral abnormalities – stereotypy time, social approach time, and Y-maze alternation performance (Milenkovic et al., 2014). If we interpret the mixed Rac1 and WAVE-1 changes in 3-week old GluN1KD mice as a moderate phenotype where increased Rac1 is compensatory for WAVE-1 deficits, then these molecular changes align with how MSN spines change with age. However, the increased Rac1 may instead represent a protracted increase in Rac1 activity that leads to atrophied spines (Luo et al., 1996; Tashiro et al., 2000; Hayashi-Takagi et al., 2010). If this is true, and the increased Rac1 compounds on the effects of a WAVE-1 decrease, then Rac1 signaling may explain some of the discrepancies between MSN spine density and GluN1KD behaviours. A worse or equally- disturbed Rac1 signaling status in 3-week old GluN1KD mice may be reflective of their demonstrated abnormalities in open field and puzzle box test performance (Milenkovic et al.,

107

2014). The altered Rac1 signaling in juvenile mice in this case may be signs of synaptic anomalies more subtle than complete spine loss. Assessing the function of GluN1KD MSN spines would be especially helpful for determining whether this is true. If so, then there would be a more coherent picture: GluN1KD MSN synapses are abnormal at the molecular level at 3 weeks of age and these abnormalities are exacerbated to the point of spine loss in later ages.

The neurological disorder literature also offers some consistency with the Rac1, WAVE-1, and dendritic spine abnormalities we report here. Rac1 mRNA was reduced in the dorsolateral prefrontal CTX of schizophrenia patients, though the decrease was not significant after correction for multiple comparisons. As an interesting aside, Cdc42 decreases did remain significant and correlated with spine density deficits (Hill et al., 2006), again encouraging a second look at synaptic Cdc42 levels in GluN1KD mice. Results from an in silico bioinformatics study have also strongly pointed to the RAC1 gene as a key node associated with autism related biological processes, cellular components, and molecular functions (Zeidán-Chuliá et al., 2013). The relationship of these two disorders to WAVE-1 is more robust and direct: WASF1 was found to be commonly decreased in two large schizophrenia cohorts (Maycox et al., 2009) and de novo truncation mutations of WASF1 were found in humans with intellectual disability, autistic features, and seizures (Ito et al., 2018). Finally, just as GluN1KD behaviour onset was evocative of symptom onset in schizophrenia previously (Eranti et al., 2013; Milenkovic et al., 2014; Hardingham and Do, 2016), our observations of concurrently decreased Rac1 and WAVE-1 at later ages are also developmentally consistent. Overall, there is strong and consistent evidence of Rac1 signaling disruption leading to dendritic spine deficits and behavioural abnormalities in GluN1KD mice. These phenotypes are likely relevant to neurological disorders, thus supporting further investigations into treating them.

4.2 Effect of Restoring WAVE-1 on GluN1KD Phenotypes

Since WAVE-1 is generally acknowledged to act downstream of Rac1 (Eden et al., 2002; Soderling et al., 2003), we targeted its restoration for the improvement of GluN1KD molecular, synaptic, and behavioural phenotypes. Progeny from intercrossing the GluN1KD and WAVE-Tg mouse lines were viable and survived to adulthood. As shown in Figures 3.3 and 3.4, the expression of the transgenic human WASF1 was as expected in the STR and HPC of F1 mice:

108 transgene expression was abundant and similar to endogenous Wasf1 in WAVE-Tg and GluN1KD-WAVE mice and near absent in non-WAVE littermates. It would be useful in the future to further study these expression patterns via an additional co-labeling with Grin1 probes. Such labling would allow us to more carefully discern interactions between knocking down Grin1 and WASF1/Wasf1 expression in our test mice. The protein levels of striatal WAVE-1 in F1 mice are shown in Figure 3.5 and were also consistent with expectations. WAVE-1 was confirmed as decreased in GluN1KD mice compared to WT littermates, higher than WT levels in WAVE-Tg mice, and recovered towards WT levels in GluN1KD-WAVE mice. However, the high proportion of pS397-WAVE-1 in GluN1KD-WAVE mice relative to WT littermates may represent a moderation of any effects from total WAVE-1 increases. This is because phosphorylated WAVE-1 is generally considered inactive, since neuronal stimulation via depolarization, D1R agonism, and cyclic adenosine monophosphate (cAMP) increases results in WAVE-1 dephosphorylation and spine growth (Kim et al., 2006b; Sung et al., 2008). Some of these experiments have also demonstrated WAVE-1 dephosphorylation after NMDAR stimulation, consistent with an idea of NMDAR hypofunction leading to more phosphorylated WAVE-1 (Sung et al., 2008; Ceglia et al., 2010). However, 17β-estradiol activation of ERα in rat cortical neurons was also associated with WAVE-1 phosphorylation, Arp2/3 colocalization to sites of spine formation, and increased dendritic spines (Sanchez et al., 2009). While our observations are likely more related to depolarization and D1R mechanisms, particularly since many of those experiments used mouse striatal MSNs (Kim et al., 2006b), we cannot rule out that established striatal ERα functions (Küppers and Beyer, 1999; Dubal et al., 2001; Schultz et al., 2009) are changed in GluN1KD mice. Further investigating how WAVE-1 in GluN1KD MSNs respond to different stimuli would help us understand the biological relevance of increased pS397-WAVE-1. GluN1KD striatal Rac1 was lower than WT levels, but the difference was unexpectedly not statistically significant. This could be due to insufficient statistical power (post hoc β = 0.66). Overall, the most important molecular characterization of our GluN1KD- WAVE mice was an intended and expected primary outcome: total striatal WAVE-1 was improved by intercrossing GluN1KD with WAVE-Tg mice.

The rescue of striatal WAVE-1 in GluN1KD-WAVE mice was also reflected by MSN spine density improvement, as shown in Figure 3.6a. GluN1KD-WAVE mice had a MSN dendritic spine density closer to WT mice than GluN1KD littermates, which again displayed a deficit. This

109 result is consistent with our hypothesis and supports the idea of targeting WAVE-1 specifically, and Rho GTPase or Rac1 signaling generally, to treat GluN1KD phenotypes. Achieving dendritic spine rescue is also consistent with previous experiments where cotransfection of knockdown-resistant Wasf1 improved the spine density of cultured neurons with endogenous Wasf1 knockdown (Kim et al., 2006b; Sung et al., 2008; Sanchez et al., 2009). Our results have now recapitulated these findings in vivo. The spine loss reported here is similar to the dendritic spine deficits reported in neuron-specific Rac1 knockout mice (Corbetta et al., 2009) and WAVE-1 knockout mice (Soderling et al., 2007). The spine density changes we observed are also similar in magnitude or proportion to those reported after fear conditioning and other behavioural tests (Roberts et al., 2009; Lai et al., 2012; Sanders et al., 2012). Thus it was possible that typical GluN1KD behavioral phenotypes were also altered in GluN1KD-WAVE mice.

GluN1KD behavioural abnormalities were retained in GluN1KD mice born from the intercross breeding of C57B1/6J Grin1+/− WAVE-Tg mice and 129X1/SvJ Grin1+/− non- WAVE-Tg mice. Our assessments of Y-maze, 8-arm radial maze, puzzle box, open field, and social approach test performance were consistent with previous GluN1KD studies (Mohn et al., 1999; Dzirasa et al., 2009; Milenkovic et al., 2014). These results, along with what happens with WAVE-1 rescue, are shown in Figures 3.6 and 3.7. GluN1KD mice displayed decreased Y-maze spontaneous alternations, 8-arm radial maze ETR, and social approach time compared to WT littermates. They also had increased 8-arm radial maze WME, puzzle box completion time, horizontal locomotor activity, and stereotypy incident number.

GluN1KD-WAVE mice demonstrated a selective improvement in maze exploration performance, consistent with the idea that striatal WAVE-1 and MSN spine improvements are behaviourally relevant. Specifically, GluN1KD-WAVE mice behaved like WT littermates in the Y-maze and had a significant improvement of 8-arm radial maze WME but not ETA compared to GluN1KD littermates. The GluN1KD-WAVE mice still had significantly more WMEs compared to WT littermates, indicating a selective and partial rescue in 8-arm radial maze performance. Such a combination of results is not unusual since the 8-arm maze is more complex than the 3-armed Y-maze. Furthermore, 8-arm WME is similar to 3-arm alternations as both measures are based on counts of incorrect or correct, respectively, arm entries (Mandillo et al., 2008; Dzirasa et al., 2009). ETR, on the other hand, is a measure of how quickly mice make their

110 first erroneous entry into an already-visited maze arm (Levin et al., 2009). An improvement in Y-maze alternations or 8-arm WME would therefore not necessarily predict better ETA scores. Our results are also limited due to potential confounds between hyperactivity and maze exploration performance – more errors may be due to more overall activity in GluN1KD mice, for example. However, these limitations are somewhat controlled in a few ways: Y-maze performance is a % of total arm entries rather than absolute counts of arm entries; 8-arm radial maze errors are counted until all maze arms are entered regardless of how quickly the entries were made. Furthermore, based on trends of hyperactivity in our open field test, we would expect GluN1KD-WAVE mice to perform as badly or worse than GluN1KD mice if our maze results were largely driven by horizontal movement differences. A limited behavioural rescue in GluN1KD-WAVE mice was also not unusual because more than WAVE-1 decreases could contribute to abnormal behaviours, such as the previously discussed GluN1KD-altered DA system (Ferris et al., 2014). The high proportion of pS397-WAVE-1, if truly inactive in our experimental context, would further predict a more modest behavioural improvement in GluN1KD-WAVE mice. The discovery of GluN1KD-WAVE behavioural improvements is consistent with previous reports of the STR being important for cognition and maze exploration (McDonald and White, 1993; Floresco et al., 1997; Simpson et al., 2010; Shu et al., 2015), though such evidence is fairly limited. Overall, it makes sense for striatal WAVE-1 and MSN spine increases in GluN1KD-WAVE mice to co-present with a modest improvement of maze exploration behaviour.

In contrast, the current literature does not elucidate why only maze exploration was improved in GluN1KD-WAVE mice. Striatal activity has also been linked to social (Báez-Mendoza and Schultz, 2013; Gunaydin et al., 2014) and locomotor behaviour (Burns et al., 1994; Patterson et al., 2015), though again the evidence is overall limited. The combination of GluN1KD-WAVE behaviours has not been replicated in any reports of GluN1KD mice, though the specific behavioural test battery we employed has also not been previously used in one GluN1KD study. There has been one prior instance of Y-maze behaviour being distinct in GluN1KD mice: it was the only behaviour that was not significantly worse in female adult GluN1KD mice (Milenkovic et al., 2014). This may mean that Y-maze deficits are later-onset than other GluN1KD behavioural abnormalities and, consequently, beneficial WAVE-1 increases further delayed or prevented its manifestation. However, we did not detect a significant effect of sex in our

111 behavioural tests. More Y-maze tests of male and female GluN1KD mice with greater age resolution around 12 weeks may help us understand its precise onset and trajectory. These and further experiments are needed to clarify why only maze exploration behaviours are improved in GluN1KD-WAVE mice. As part of this initiative, we further asked whether WAVE-1 and dendritic spine changes in the GluN1KD and GluN1KD-WAVE STR were mirrored at the HPC, which, along with the CTX, is very important for cognition and maze exploration (Choy et al., 2008; Morris et al., 2013).

4.3 Striatal and Hippocampal Differences in GluN1KD and GluN1KD-WAVE mice

We assessed total hippocampal tissue WAVE-1 and dendritic spines of CA1 pyramidal neurons, whose function has been previously associated with Y-maze and 8-arm radial maze performance (Markus et al., 1994; Li et al., 2005). Unchanged WAVE-1 expression in the HPC of our test mice was consistent with the previous report of WT-like DISC1 levels in the GluN1KD HPC (Ramsey et al., 2011). In the context of maze exploration improvement in GluN1KD-WAVE mice, however, a lack of WAVE-1 and spine density deficits in the HPC of GluN1KD mice (and its theoretical rescue with incorporation of a WASF1 transgene) is unexpected. This is because performance improvements in maze tests are cannonically considered improvements of learning and memory (Maurice et al., 1994; Dean et al., 2009; Kim et al., 2013), which is traditionally considered a main function of the HPC (McHugh et al., 2007; Neves et al., 2008; Buzsáki and Moser, 2013). To note, CA1 dendritic spine density experiments were found to be underpowered post hoc (β = 0.45-77). This does not resolve the inconsistency between our hippocampal and behavioural results, however: if CA1 spine density comparison results were false negatives, we would predict spine deficits in GluN1KD mice but not rescue in GluN1KD-WAVE littermates when considering our current data. GluN1KD-WAVE improvements of maze exploration thus remain incongruent with our WAVE-1 and spine data in the HPC.

Our synaptic and molecular data from the STR may still offer important clues about the relationship between WAVE-1, MSN spines, and mouse behaviour. In previous studies of striatal function and maze exploration, the dorsal STR was suggested to facilitate the reinforcement of

112 stimulus-response associations (McDonald and White, 1993), the NAc was associated with goal- directed exploration (Floresco et al., 1997), and the marginal division was associated with consolidation processes for long-term memory (Shu et al., 2015). As we did not employ any overt stimuli or cues in our maze tests, stimulus-response reinforcement may be a more minor factor in our study. In contrast, goal directed exploration may be especially applicable for food- restricted mice and long-term memory can form from learning by repetition. Both food restriction and repeat-testing were part of our 8-arm radial maze tests. Based on this information, the striatal improvements in GluN1KD-WAVE mice may be related to, for example, a recovered capacity for learning food-foraging strategies. The improved Y-maze alternation scores of GluN1KD-WAVE mice may similarly be due to the test’s similarity to nondelayed random foraging tasks, which depended on the NAc (Floresco et al., 1997). It is important to note, however, that neither the marginal division of the STR nor the NAc were previously linked to cognitive functioning independently of the HPC (Floresco et al., 1997; Shu et al., 2015). Therefore, what is happening in the HPC of GluN1KD-WAVE mice is still of great future research interest.

There are multiple leads that we can further investigate to elucidate the relationship between the HPC and behaviours in GluN1KD-WAVE mice. One simple, straightforward approach is to repeat our experiments with targeted WASF1 overexpression (location-specific virus injections or promoter-based specific transgene expression) in the STR or HPC of GluN1KD mice. As mentioned before for our Rac1, WAVE-1, and spine density results in the GluN1KD STR, our measures thus far are not measures of function. Another approach may thus be to consider hippocampal WAVE-1 activity along with CA1 neuron spine functionality in GluN1KD and GluN1KD-WAVE mice. Previous studies demonstrated that WAVE-1 activation by dephosphorylation occurs upon NMDAR stimulation while total WAVE-1 levels are unchanged (Sung et al., 2008; Ceglia et al., 2010). Though this was demonstrated in both the STR and HPC, it remains to be seen if these brain regions might differ with chronic NMDAR deficiency in vivo. In other words, one possible explanation for the observed differences between the two brain regions is that chronic NMDAR hypofunction may have more effects on the activity of WAVE-1 than its total protein level in the HPC. The clearest next step investigating this line of reasoning is to assess pS397-WAVE-1 in the HPC of WT, WAVE-Tg, GluN1KD, and GluN1KD-WAVE mice. It may also be beneficial to assess other WAVE-1 phosphorylation sites such as serine

113 residues 310 and 441 in both the STR and HPC (Kim et al., 2006b). Similarly, it would be important to assess dendritic spine functionality in the STR and HPC of our test mice – did the previously reported NMDAR current deficit in MSNs (Jocoy et al., 2011) improve in GluN1KD- WAVE mice? Such information would provide insight along with an appraisal of CA1 dendrite functionality in all four genotypes of our test mice. CA1 pyramidal neurons with eliminated AMPAR and NMDAR signaling, but no physically altered dendritic spines, have been reported in mice before. The authors instead observed electrical and functional alterations (Lu et al., 2013). Finally, it is also possible that non-CA1 subdivisions of the HPC could be more affected by WAVE-1 rescue. Region 3 of the cornu ammonis or the dentate gyrus could be potential candidates as they are also important for cognition (Conrad et al., 1996; Manahan-Vaughan et al., 2008). Overall, many measures of WAVE-1 and synapse status in the HPC of our test mice can still be investigated.

Another way to reconcile the differences between our striatal and hippocampal data is to ask why WAVE-1 rescue worked in the STR rather than why it did not (by the measures we’ve assessed so far) in the HPC. Previous whole-brain assessments of functionality with Fos, though not 2-DG uptake (Duncan et al., 2002), pointed to the GluN1KD STR being a responsive, WT- like region (Miyamoto et al., 2004; Inada et al., 2007; Duncan et al., 2008). In contrast, Fos induction in hippocampal brain regions was consistently muted in GluN1KD mice along with 2- DG uptake abnormalities (Duncan et al., 2002, 2008; Inada et al., 2007). The STR may therefore be more functional and rescuable than the HPC in GluN1KD mice. On the other hand, DISC1, which was the changed protein that guided us to investigate Rho GTPases, was not altered in the HPC or CTX of 6-week old GluN1KD mice (Ramsey et al., 2011). This may simply point to Rac1 signaling being a somewhat STR-specific abnormality in GluN1KD mice. WAVE-1 has also recently been proposed to be part of a negative feedback mechanism for glutamatergic synapses in MSNs expressing D1Rs (Ceglia et al., 2017). Thus, WAVE-1 rescue in the STR may be attenuating a hyperglutamatergic state in GluN1KD mice, which was previously suggested by increased AMPAR currents in MSNs (Jocoy et al., 2011) and VTA neuron firing rate (Ferris et al., 2014). One way to address these questions, to various extents, is to attempt cell-specific WAVE-1 rescue in GluN1KD mice. For example, assessing GluN1KD phenotypes after WAVE-

1 rescue only in D1R-expressing neurons may help us interpret our current results. This and other studies are needed to more fully understand GluN1KD phenotypes. In relation to our research

114 questions, more investigations into the GluN1KD HPC, STR, and other brain regions important for rodent cognition and behaviour would be especially helpful.

4.4 Sex-Dependency of Transcriptomic Changes in the GluN1KD STR

To further our understanding of the GluN1KD STR, we employed an unbiased bioinformatic approach – RNAseq analysis of mRNA changes. The number of genes that had significantly different expression between WT and GluN1KD mice was higher than previously reported proteomic differences (Wesseling et al., 2014; Cox et al., 2016), but within the range of expectations for RNAseq transcriptomic studies (Darlington et al., 2013; Hodes et al., 2015; Seyednasrollah et al., 2015). Such a high number of gene expression changes is consistent with our modest GluN1KD-WAVE behavioural rescue results; hundreds of gene expression changes in the GluN1KD STR likely means that WAVE-1 deficits are only part of the aberrations in GluN1KD mice. Table III-I provided initial support for us to trust our RNAseq results, as differential expressions between GluN1KD and WT mice and between male and female mice were as expected. However, the decrease of Grin1 by only about 50% in GluN1KD mice was unexpected quantitatively. Closer examination of our RNAseq data later explained this discrepancy as there was a substantial expression of incomplete Grin1 transcripts upstream of intron 19 in GluN1KD mice. This would be more wholly accounted for by RNAseq compared to specific primer-based polymerase-chain reaction experiments. Some of our detected genes were previously reported as mutated in schizophrenia patient genetic studies and are consistent with etiological evidence, such as Cacnb2 and Drd2 (Maycox et al., 2009; Ripke et al., 2014). Cacnb2 stands for Ca2+ voltage-gated channel auxiliary subunit β2 and encodes a subunit of voltage 2+ gated Ca channels while Drd2 stands for DA receptor D2, encoding D2Rs. Mutations in both of these genes have been associated with schizophrenia previously (Berger and Bartsch, 2014; Vink et al., 2016). There were also general class-overlaps between our results and previous GluN1KD proteomic studies, such as alterations of growth factors and interleukins (Wesseling et al., 2014). This again offers a sense of consistency between results, but we would not necessarily expect greatly similar results because previous studies investigated protein changes in the HPC and prefrontal CTX of male mice (Wesseling et al., 2014; Cox et al., 2016) while we studied mRNA

115 changes in the STR of male and female mice. And as shown in Figure 3.9, sex was indeed an important factor in our transcriptomic assessments of the GluN1KD STR.

The great extent of sex-dependency observed in GluN1KD transcriptional changes at the STR is not surprising. Although less sexually dimorphic than other organs and tissues, hundreds of sex differences in gene expression have been reported for the brain (Rinn and Snyder, 2005; Yang et al., 2006; Hodes et al., 2015). Gonadal hormones can affect many changes through transcriptional regulation, alone or in conjunction with growth factors, while genes on sex chromosomes can likewise influence the CNS and behaviour (Reisert and Pilgrim, 1991; Ngun et al., 2011; Fester and Rune, 2015; McCarthy et al., 2015). NMDAR function, specifically, is also modulated by sex hormones (Bi et al., 2001; van den Buuse et al., 2017). For example, gonadectomy has been demonstrated to reduce NMDAR-antagonism-induced hyperactivity in male but not female mice (van den Buuse et al., 2017). The rat estrus cycle has also been demonstrated to regulate NMDAR phosphorylation and CA1 LTP (Bi et al., 2001). For an example with methodology similar to our work, a study of how stress affects transcription in the NAc of male and female mice reported hundreds of genes with sex-dependent expression changes (Hodes et al., 2015). This is akin to our reported results in Table III-II; apart from the expression of 35 genes that was changed in the same direction between male and female GluN1KD mice, hundreds of changes were sex-dependent. Finally, our results may be reflective of the sex differences seen in GluN1KD behavioural abnormalities, which have been reported multiple times (Duncan et al., 2004b; Moy et al., 2006, 2012; Milenkovic et al., 2014). The most robust behavioural sex differences demonstrated for GluN1KD mice were at younger ages (Moy et al., 2012; Milenkovic et al., 2014), but some differences remain in adulthood (Duncan et al., 2004b; Moy et al., 2006; Milenkovic et al., 2014). As we assessed the striatal transcriptome of adult GluN1KD mice, our findings may represent the molecular underpinnings of sex-dependent behavioural differences found in adulthood. It is also possible that our results include leftover signatures from neurodevelopmental processes that, at its peak, manifested as more severe behavioural differences between young GluN1KD males and females. Characterizing the striatal transcriptome of 3- and 6-week old, male and female, GluN1KD mice may help us distinguish these two interpretations. Over all, a strong effect of sex on the GluN1KD striatal transcriptome is consistent with previous evidence.

116

Our results displayed in Table III-III, Table III-VI and Figure 3.10 present further clues regarding the sex-dependency of GluN1KD gene expression changes in the STR. As shown in Tables III-III and III-IV, the directionality of transcriptomic changes becomes striking when considering “male” and “female” genes that are altered in GluN1KD mice – those genes that were more highly expressed in WT males or females, respectively, when compared to WT littermates of the opposite sex. To sum, all gene expression changes were opposite of what may be expected by biological sex: male genes were downregulated in male GluN1KD mice and/or upregulated in female GluN1KD mice while female genes were downregulated in female GluN1KD mice and/or upregulated in male GluN1KD mice. Put another way, male GluN1KD transcriptomes were more female-like and less male-like while female GluN1KD transcriptomes were more male-like and less female-like. Such a pattern may be reflective of multiple reports associating gender dysphoria and transsexualism with schizophrenia and ASDs (Gruzelier, 1994; Borras et al., 2007; de Vries et al., 2010; Saleem and Rizvi, 2017). However, clinical gender- factored studies are limited in number (Lewine et al., 2017; Saleem and Rizvi, 2017) and competing theories exist as well, such as hyper-masculinization in autistic brains (Baron-Cohen, 2005). It would be interesting to see how GluN1KD sex differences compare with findings from future studies that consider gender in a clinical setting. For now, this striking pattern of change represents to us a high likelihood that sex-dependent differences in GluN1KD striatal transcriptomes were due to sex-centered mechanisms.

Pathways analysis of our RNAseq results with IPA yielded further evidence of altered sex- related mechanisms in GluN1KD mice. IPA identified Esr1 as a key node that was connected to many other gene expression changes found in our GluN1KD STR samples, as shown in Tables III-V and III-VI. Specifically, it predicted ERα activity changes based on how genes it regulates were changed in our samples and based on how Esr1 expression itself was changed. Consistent with our results in Tables III-III and III-IV, GluN1KD Esr1 expression and ERα predicted activity changes were opposite of expectations based on sex: Esr1 was increased and ERα was predicted to be more active in GluN1KD males while Esr1 was decreased in GluN1KD females and their ERα activation z-score was negative, though statistically insignificant. This pattern was similar to algorithmically-determined Esr1 association networks shown in Figure 3.10. These networks, also generated by IPA, demonstrate visually a group of sex-dependent, Esr1/ERα- centered, gene expression changes in GluN1KD mice. Interestingly, a similar IPA network has

117 been previously published with one of the GluN1KD proteomic studies. Although the authors did not focus on Esr1, they too reported an Esr1-centered network that was largely downregulated or inhibited in the frontal CTX of male mice (Wesseling et al., 2014). Their network was associated with “developmental disorder, hematological disease, and molecular transport”, potentially reflecting the neurodevelopmental nature of schizophrenia and ASDs (Wesseling et al., 2014; Goldstein et al., 2018). The current striatal bioinformatics data have revealed ERα as a key target for future research to understand the molecular changes and sex differences in GluN1KD mice. Investigations into ERα signaling may also shed more light on our WAVE-1 rescue experiments due to its unique role in promoting both WAVE-1 phosphorylation and dendritic spine formation (Sanchez et al., 2009).

Lastly, we analyzed our RNAseq results with Enrichr to assess the diseases our transcriptomic changes most likely represented and the drugs most likely to reverse them. This was especially interesting to us because male and female GluN1KD mice presented with mostly the same behavioural phenotypes, if to different degrees (Moy et al., 2012; Milenkovic et al., 2014). The degree of sex differences in GluN1KD behaviours also often decreased by adulthood (Milenkovic et al., 2014). The diseases and drugs resulting from our analysis are shown in Tables III-VII and III-VIII, respectively. Generally, results in both tables present a fair degree of consistency despite being generated from 4 different gene list inputs. The consistency for Enrichr disease results was readily apparent as most came from mouse models of HD. On the other hand, most of the drug records looked to involve different drugs. However, closer examination of the drugs’ mechanisms of action pointed to some consistent molecular targets and clinical purposes: growth signaling and anti-cancer treatment, respectively. The results therefore suggest that various sex-dependent gene expression changes in GluN1KD mice generally contribute to similar larger-scale phenotypes.

Enrichr disease records associated with our RNAseq results are mostly consistent with the literature. The instances of schizophrenia and AD records are consistent with previously discussed associations between these disorders and GluN1KD mice or NMDARs (Mohn et al., 1999; Shankar et al., 2007; Halene et al., 2009a; Tönnies and Trushina, 2017). How NMDARs and GluN1KD mice relate to HD, which was our most common Enrichr result, needs further discussion. HD is a neurodegenerative disease caused by increased CAG repeats in the huntingtin (HTT) gene that primarily targets striatal MSNs and cortical neurons (Shehadeh et al.,

118

2006; Indersmitten et al., 2015; Mehrabi et al., 2016). An association between HD and GluN1KD mice is therefore feasible given the many alterations in MSNs we and previous GluN1KD studies have observed (Jocoy et al., 2011; Ramsey et al., 2011). Though GluN1KD mice that were 10- to 16-weeks old did not demonstrate reduced MSNs (Ramsey et al., 2008), we cannot rule out that they do with age or that surviving GluN1KD MSNs are simply less fit than WT MSNs. Patients with HD have been reported to have less NMDARs in the putamen subregion of the STR (Young et al., 1988). However, NMDAR hyperactivity and excitotoxicity are often considered associated with the neurodegeneration seen in HD (Cepeda et al., 2001; Fan and Raymond, 2007). This may seem incongruent with the decreased NMDARs of GluN1KD mice, but experiments with varying doses of memantine in cultured neurons demonstrated that increased extrasynaptic NMDAR activity resulted in greater vulnerability to mutant HTT protein. Furthermore, high-dose memantine that blocked both synaptic and extrasynaptic NMDARs worsened outcomes in a mouse model of HD (Okamoto et al., 2009). Our Enrichr disease results may therefore be considered consistent with the evidence from HD research, particularly with previous signs of a hyperexcitable brain in GluN1KD mice (Jocoy et al., 2011; Ferris et al., 2014). These results also demonstrate how different gene expression changes in male and female GluN1KD mice can still be related to one macroscopic phenotype.

The drug-related records identified from Enrichr databases also demonstrated how different molecular changes can affect similar processes to contribute to similar endpoints. In general, most of the drugs detected by Enrichr analysis were antineoplastic agents (Oh and Kwon, 2009; Ueno et al., 2012; Valianou et al., 2015; Bollard et al., 2017) that acted on signaling mechanisms related to cellular and vesicular growth (Mohan et al., 2004; Li and Yang, 2010; Valianou et al., 2015; Qiu et al., 2016). A general summary of these pathways and where our Enrichr-identified drugs act is illustrated in Figure 4.2. These drugs often act on ERK- (Lorusso et al., 2005; Fukuyo et al., 2010; Khalil et al., 2011) or AKT- (Li and Yang, 2010; Qiu et al., 2016; Suman et al., 2016) signaling pathways, which are commonly dysregulated in cancer (Steelman et al., 2011; De Luca et al., 2012). This is consistent with previous GluN1KD proteomic findings, as ERK aberrations have been reported for the GluN1KD HPC and CTX too (Wesseling et al., 2014). Furthermore, ERK and AKT pathways are regulated by NMDARs (Thomas and Huganir, 2004; Gong et al., 2006; Wang Yansong et al., 2014) and interact with estrogen signaling pathways (Hisamoto et al., 2001; Rocca et al., 2014; Filardo et al., 2018). A disruption of

119

NMDARs, which was strongly demonstrated in GluN1KD mice (Mohn et al., 1999; Duncan et al., 2002; Bodarky et al., 2009; Jocoy et al., 2011), or estrogen signaling, which was implied by our RNAseq data for the GluN1KD STR, could therefore result in dysregulated ERK and ATK pathways. These two pathways can interact with each other (Steelman et al., 2011; De Luca et al., 2012), potentially compensating for or exacerbating each other’s dysregulation. Both may also be affected by the same drug from our Enrichr results, such as KU-55933. KU-55933 administration has been shown to result in AKT phosphorylation and ERK dephosphorylation, ultimately presenting an anti-survival signal (Khalil et al., 2011). Beyond just AKT and ERK, drugs from our Enrichr results can have other targets and effects (Oh et al., 2008; Clark et al., 2012; Liang et al., 2017). For example, the frequently-reported Withaferin A has been demonstrated to have both antitumor and antiangiogenic effects through nitric oxide and other molecular mediators (Mohan et al., 2004; Oh and Kwon, 2009; Khan and Ghosh, 2010; Suman et al., 2016). However, the fact our identified drugs often have wide-ranging effects also means that more targeted studies on ERK and AKT pathways will be needed to confirm their specific dysregulation and biological significance in GluN1KD mice. Confirmatory experiments in the future could include quantitative polymerase-chain reaction experiments to study RNA and western blotting experiments to study protein changes.

120

Figure 4.2. Schematic of NMDAR signaling pathways commonly targeted by drugs from Enrichr analysis. ERK and AKT signaling, strongly associated with cell growth and survival, were distinctly common targets of Enrichr-result drugs. These signaling pathways were distinctly targeted out of the many that act downstream of NMDARs, such as those previously represented in Figure 1.2. Green and red arrows again indicate generally positive and negative relationships, respectively. Dashed lines with blunted ends for drugs represent indirect inhibition or currently unspecific mechanism of action. Solid lines with blunted ends (for drugs) denote more direct inhibition. Abbreviations follow those used in text and previous figures. This figure was constructed with literature cited in-text as well as other studies (Christensen et al., 2003; Ueno et al., 2012; Li et al., 2014; Hirabayashi and Cagan, 2015).

121

Our RNAseq and gene list analyses revealed a divergence of transcriptomic changes in male and female GluN1KD mice followed by a reconvergence of these changes into similar biological processes. Such results mirror GluN1KD behavioural phenotypes that are sex-dependent but largely very similar (Moy et al., 2012; Milenkovic et al., 2014). Estrogen, neurodegeneration and growth signaling pathways are some of the major themes that stand out from our assessment of the GluN1KD striatal transcriptome. Estrogen, due to its role as a sex hormone, may be a key mediator of GluN1KD sex differences. Neurodegenerative signaling may be related to imbalanced extrasynaptic-to-synaptic NMDAR function (Okamoto et al., 2009; Xu et al., 2009; Pál, 2018), which could be present in GluN1KD mice that showed signs of a hyperexcitable brain previously (Dzirasa et al., 2009; Jocoy et al., 2011; Ferris et al., 2014). Alterations in growth factor signaling is consistent with previous GluN1KD proteomic results (Wesseling et al., 2014), but need confirmation of specific pathway changes. GluN1KD growth factor signaling at the STR also need to be carefully considered and investigated in the context of potentially concurrent neurodegeneration mechanism changes. Overall, our unbiased bioinformatics investigation into the GluN1KD STR has provided several new clues regarding the molecular changes that occur with NMDAR deficiency and their biological significance.

4.5 Concluding Statements

4.5.1 Strengths of the Study

The present study investigates the effects of NMDAR hypofunction on the STR with multiple experimental approaches and perspectives. The initial assessment of MSN dendritic spines and Rho GTPase signaling proteins in GluN1KD mice covered three key ages: 3-week old juveniles, 6-week old adolescents, and 12-week old adults. Not only did this provide an age-trajectory of molecular and synaptic changes, it also allowed an assessment of developmental alignment with our previous behavioural study of similarly-aged mice (Milenkovic et al., 2014). This was an advantage because the comparisons allowed us a rough gauge of the biological and behavioural relevance of our findings. The age-trajectories of GluN1KD phenotypes also offered clues regarding the validity of using these mice to model early-onset neurological disorders like ASDs (Baxter et al., 2015; Halladay et al., 2015) or later-onset ones like schizophrenia (Sham et al., 1994; Thorup et al., 2007; Eranti et al., 2013). Our MSN spine density results seem to point to

122 the latter while our Rho GTPase results will require further study into the significance of mixed- direction changes for Rac1 and WAVE-1.

For our experiments with WAVE-1 rescue, a good range of neurological and behavioural measures were employed. In terms of different levels of biology, we surveyed molecular changes, spine density changes, and in vivo behavioural changes. Such a range of measures allowed us to associate phenotypes on all three levels and investigate a well-rounded and relatively holistic hypothesis – that restoring WAVE-1, a protein known to regulate dendritic spine formation (Kim et al., 2006b; Soderling et al., 2007; Ceglia et al., 2010), in GluN1KD mice would improve their MSN spine density and behavioural phenotypes. The range of behavioural tests employed was also a strength since it covered multiple domains of cognitive function. Such considerations are particularly important for disorders like schizophrenia where some symptoms are currently better treated than others (Green, 2016; Kishi et al., 2017). Finally, our WAVE-1 experiments looked at multiple brain regions, the STR and the HPC. Results from multiple brain regions are often useful for complementing each other and advancing the understanding of both regions’ function and status.

A bioinformatics approach often offers the advantage of yielding, analyzing, and interpreting large amounts of data. RNAseq experiments generally possess great sensitivity, often reporting hundreds if not thousands of significant changes that investigators can analyze by leveraging high computational power (Darlington et al., 2013; Hodes et al., 2015; Seyednasrollah et al., 2015). With our discovery of hundreds of gene expression changes, we surpassed the quantity of significant findings from previous GluN1KD proteomic studies (Wesseling et al., 2014; Cox et al., 2016). To be clear, this is not better per se. Having more potential targets to look at individually, for example, may instead be more time-consuming and distracting. However, it is often beneficial to have more data – more significant hits – to analyze computationally. In such cases, more data may result in better differentiation of results and more accurate identification of significant associations, clusters, or pathways. Results from analyses with large data-sets may better indicate the most biologically relevant changes occurring in tested samples.

4.5.2 Limitations of the Study

There are a few interpretive limitations on the results of our study, some of which were mentioned with previous discussion senctions. As mentioned previously when discussing Rho

123

GTPases, WAVE-1, and dendritic spine changes in GluN1KD mice, our experimental measures did not explicitly indicate functional changes. There were more or less protein or spines, but such changes did not necessarily reflect how active or functional these CNS components were. As alterations in function have been reported independently of quantity differences before (Kim et al., 2006b; Lu et al., 2013), it stands to reason that functional changes in our mice could suggest different conclusions than our current results. While we did assess one form of WAVE-1 phosphorylation, more could be done. The functionality of dendritic spines also offers a different level of depth to explore experimentally: circuit and network-level phenomena that may represent more than the sum of individual neuronal activity. Network-level changes have been demonstrated in the past for neurological disorders and GluN1KD mice, specifically between the CTX and the HPC (Dzirasa et al., 2009). It would be interesting to investigate which synaptic connections or networks are different between WT, WAVE-Tg, GluN1KD, and GluN1KD- WAVE mice. Furthermore, it may be beneficial to study EEG changes in our mice during behavioural test performance and contrast those findings with resting brain activity.

The interpretative limitations of our RNAseq analyses are similar to those in other bioinformatics studies (Wesseling et al., 2014; Cox et al., 2016). In general, the interpretability of our data was limited to what curated databases were present to serve as references in IPA and Enrichr. For example, without records of HD mouse models or withaferin A in Enrichr, our results would be drastically different. Our results and interpretations were largely limited to the set of possible outcomes defined in IPA and Enrichr. To go one step further, the databases and possible outcomes in IPA and Enrichr are limited to the current scientific literature. Connecting our list of transcriptomic changes to a record labeled “withaferin A” would be hard to interpret without previous studies into its mechanism of action and effects. Without previous studies (Oh and Kwon, 2009; Ueno et al., 2012; Valianou et al., 2015; Bollard et al., 2017), we would not have been able to connect the various drugs identified in our output to growth factor signaling and antineoplastic effects. This is the reliance of bioinformatics approaches on other types of experiments. The fact that we studied transcriptomic changes further compounds the limitations of our data’s interpretability. Despite the greater sensitivity we had with our approach, proteomic data are likely more indicative of functional changes in sample tissues (Cox et al., 2016). However, transcript levels have been reported to be more strongly correlated with clinically relevant mouse phenotypes than protein levels before (Ghazalpour et al., 2011). Our

124 transcriptomic data are therefore still very important for understanding GluN1KD mice. Overall, it is clear where we should take the results from our bioinformatics analyses: back to more hypothesis-based testing. To fully interpret our data, more experiments will be needed in the future to confirm and follow-up ERα-, neurodegeneration-, and growth factor-related signaling in the GluN1KD STR. This would include basic assessments like signaling component levels as well as more functional measures: how striatal proteins, dendritic spines, and neurons change after various estrogenic, damaging, or growth-promoting stimuli.

4.5.3 Future Directions

Many of the future experiments that could build upon our results have been mentioned previously. An assessment of functional outcomes will greatly improve our understanding of changes in the GluN1KD and GluN1KD-WAVE STR. Specifically, future investigators can execute something like the following in WT, WAVE-Tg, GluN1KD, and GluN1KD-WAVE: measure Rac1 and WAVE-1 function, such as via quantification of Rac1-GTP and all forms of phosphorylated WAVE-1 (Kim et al., 2006b; Hayashi-Takagi et al., 2010); culture MSNs and assess their ability to form new spines or larger spines after stimulation with glutamate or DA, as an assessment of ultimate Rac1 signaling outcomes; assess electrophysiological properties of these same cultured MSNs to see if synaptic signaling is directly improved. To help confirm that any observed differences are indeed due to WAVE-1 changes, cultured neuron experiments can be repeated with Wasf1 knockdown and WASF1 transfection, similar to previous studies (Kim et al., 2006b; Sung et al., 2008; Sanchez et al., 2009). Doing these same experiments in hippocampal neurons will also help clarify the state, function, and relevance of WAVE-1 in the GluN1KD HPC. If a similar pattern remains where striatal measures are improved but hippocampal ones are not, it may be helpful to change the direction of future experimental approach – look for signs of restored GluN1KD-WAVE brain function first. Like in previous GluN1KD studies, this can be done indirectly by using Fos-induction or 2-DG uptake as surrogates (Duncan et al., 2002, 2008; Miyamoto et al., 2004; Inada et al., 2007). A more direct method may be to investigate the coupling of local field potentials between different brain regions and the STR of test mice. This can be done in mice exploring a maze, like previously performed (Dzirasa et al., 2009), as well as at rest. Any brain region that has improved function in GluN1KD-WAVE mice compared to GluN1KD mice can then serve as a target for further assessing molecular and synaptic rescue in GluN1KD-WAVE mice. Any GluN1KD-WAVE

125 brain regions identified and confirmed this way will likely be working with the rescued STR to improve maze exploration behaviour. Overall, these experiments can provide many new clues about the relationship between WAVE-1, synaptic plasticity, NMDAR hypofunction, and mouse behaviour.

Future experiments related to our RNAseq, IPA, and Enrichr analyses have also been mentioned above, particularly at the end of the study limitations section. On top of what was suggested there, more RNAseq analyses of the GluN1KD striatal transcriptome in 6-, 3-, 5-, and 2-week old (in order of priority) male and female mice may have excellent value. Combining results from these experiments with our current results, we would have striatal transcriptomic profiles in alignment with the two main GluN1KD behavioural studies that included age and sex as factors (Moy et al., 2012; Milenkovic et al., 2014). Transcriptomic changes that align with striking age- and sex-dependent GluN1KD behavioural differences, such as increased sociability in 5-week old GluN1KD males (Moy et al., 2012) or generally worse phenotypes in males at 6- weeks of age (Milenkovic et al., 2014), may reveal signaling pathways particularly relevant for aligned behaviours. Completing a set of RNAseq experiments for WT, WAVE-Tg, GluN1KD, and GluN1KD-WAVE mice can also help us further understand the scope of molecular changes that can come from increasing total WAVE-1 levels. As part of this, however, it may be better to employ single-cell RNAseq methods (Tang et al., 2010; Hashimshony et al., 2012) to more specifically study MSNs. Results from these experiments may also shed more light on the altered DA system (Ferris et al., 2014) and MSN electrophysiology in GluN1KD mice (Jocoy et al., 2011). But like our current RNAseq results, however, more RNAseq analyses will likely result in more hypotheses that need non-bioinformatics approaches to test. In which case, more characterizations to confirm experimental targets would be needed. This would be followed by yet further studies, perhaps like our work with WAVE-1 rescue. And with every step, we would work towards a greater understanding of NMDAR deficiency and synaptic plasticity in the STR.

4.5.4 Conclusion

Our work has identified Rac1 signaling, particularly through WAVE-1, as a viable target for improving GluN1KD mouse phenotypes. This key finding is supportive of and consistent with the idea of targeting Rac1 signaling to improve schizophrenia-related outcomes, such as by inhibiting p21-activated kinases (Hayashi-Takagi et al., 2014). The behavioural improvements

126 we observed from WAVE-1 rescue were modest, specific to maze exploration, and potentially representative of better long-term memory or goal-directed exploration (Floresco et al., 1997; Shu et al., 2015). Concurrently, GluN1KD-WAVE mice showed an improvement of MSN spine density, which supported WAVE-1’s mechanism of action as regulating synaptic plasticity (Kim et al., 2006b; Soderling et al., 2007; Sung et al., 2008). In contrast, RNAseq, IPA, and Enrichr analyses emphasized growth factor- and neurodegeneration-related signaling alterations in the GluN1KD STR. These signaling alterations were based on drastically sex-dependent gene expression changes, which may be mediated by ERα signaling changes. More studies are needed to confirm and expand upon the results from our work, particularly the new leads coming from our unbiased bioinformatics approach. The primary strength of our study was the association of changes at three different levels of biology: molecular signaling, dendritic spines, and in vivo behaviour. This aided in the interpretation of our results, particularly regarding the behavioural significance of our findings at microscopic levels. Interpreting our behavioural data as a whole was a challenge, since maze exploration has been associated with spatial working memory, curiosity, and other aspects of cognition (Hughes, 2004). Our results of STR-specific improvements in WAVE-1 levels and spine density with better maze exploration provide more evidence for the importance of striatal function in proper cognition.

Bibliography

Abbott LF, Nelson SB (2000) Synaptic plasticity: taming the beast. Nat Neurosci 3:1178.

Acton D, Miles GB (2017) Differential regulation of NMDA receptors by d-serine and glycine in mammalian spinal locomotor networks. J Neurophysiol 117:1877–1893.

Afgan E et al. (2018) The Galaxy platform for accessible, reproducible and collaborative biomedical analyses: 2018 update. Nucleic Acids Res 46:W537–W544.

Aiba I, Shuttleworth CW (2012) Sustained NMDA receptor activation by spreading depolarizations can initiate excitotoxic injury in metabolically compromised neurons: Neuronal injury following spreading depolarization. J Physiol 590:5877–5893.

Aiyer R, Mehta N, Gungor S, Gulati A (2018) A Systematic Review of NMDA Receptor Antagonists for Treatment of Neuropathic Pain in Clinical Practice. Clin J Pain 34:450–467.

Alam A, Suen KC, Hana Z, Sanders RD, Maze M, Ma D (2017) Neuroprotection and neurotoxicity in the developing brain: an update on the effects of dexmedetomidine and xenon. Neurotoxicol Teratol 60:102–116.

Albers GW, Goldstein LB, Hall D, Lesko LM, for the Aptiganel Acute Stroke Investigators (2001) Aptiganel Hydrochloride in Acute Ischemic Stroke: A Randomized Controlled Trial. JAMA 286:2673.

Allen JW, Knoblach SM, Faden AI (1999) Combined mechanical trauma and metabolic impairment in vitro induces NMDA receptor-dependent neuronal cell death and caspase-3- dependent apoptosis. FASEB J 13:1875–1882.

Allen R, Young S (1978) Phencyclidine-Induced Psychosis. Am J Psychiatry 135:1081–1084.

Alles SRA, Smith PA (2018) Etiology and Pharmacology of Neuropathic Pain Isom LL, ed. Pharmacol Rev 70:315–347.

127 128

Alonso P, Gratacós M, Segalàs C, Escaramís G, Real E, Bayés M, Labad J, López-Solà C, Estivill X, Menchón J (2012) Association between the NMDA glutamate receptor GRIN2B gene and obsessive–compulsive disorder. J Psychiatry Neurosci 37:273–281.

Alsaad HA, DeKorver NW, Mao Z, Dravid SM, Arikkath J, Monaghan DT (2018) In the Telencephalon, GluN2C NMDA Receptor Subunit mRNA is Predominately Expressed in Glial Cells and GluN2D mRNA in Interneurons. Neurochem Res Available at: https://doi.org/10.1007/s11064-018-2526-7.

Alsaloum M, Kazi R, Gan Q, Amin J, Wollmuth LP (2016) A Molecular Determinant of Subtype-Specific Desensitization in Ionotropic Glutamate Receptors. J Neurosci 36:2617–2622.

Anand A, Charney DS, Oren DA, Berman RM, Hu XS, Cappiello A, Krystal JH (2000) Attenuation of the Neuropsychiatric Effects of Ketamine With Lamotrigine: Support for Hyperglutamatergic Effects of N-methyl-D-aspartate Receptor Antagonists. Arch Gen Psychiatry 57:270.

Anderson LT, Campbell M, Grega DM, Perry R, Small AM, Green WH (1984) Haloperidol in the treatment of infantile autism: effects on learning and behavioral symptoms. Am J Psychiatry 141:1195–1202.

Andersson O, Stenqvist A, Attersand A, von Euler G (2001) Nucleotide Sequence, Genomic Organization, and Chromosomal Localization of Genes Encoding the Human NMDA Receptor Subunits NR3A and NR3B. Genomics 78:178–184.

Andrews JL, Newell KA, Matosin N, Huang X-F, Fernandez-Enright F (2015) NMDA receptor antagonism by phencyclidine reduces NWASP and WAVE1 protein expression and reduces levels of myelination markers in the prefrontal cortex of rats. Available at: http://ro.uow.edu.au/ihmri/515/ [Accessed June 2, 2016].

Antonova I (2001) Rapid Increase in Clusters of Presynaptic Proteins at Onset of Long-Lasting Potentiation. Science 294:1547–1550.

Apps MAJ, Rushworth MFS, Chang SWC (2016) The Anterior Cingulate Gyrus and Social Cognition: Tracking the Motivation of Others. Neuron 90:692–707.

129

Araújo JRC, Júnior J de MA de M, Damasceno M de BMV, Santos SAAR, Vieira-Neto AE, Lobo MDP, Campos AR, Moreira R de A, Monteiro-Moreira AC de O (2018) Neuropharmacological characterization of frutalin in mice: Evidence of an antidepressant-like effect mediated by the NMDA receptor/NO/cGMP pathway. Int J Biol Macromol 112:548–554.

Arellano JI, Benavides-Piccione R, DeFelipe J, Yuste R (2007) Ultrastructure of dendritic spines: correlation between synaptic and spine morphologies. Front Neurosci 1:131.

Armstrong-James M, Fox K (1988) Evidence for a specific role for cortical NMDA receptors in slow-wave sleep. Brain Res 451:189–196.

Artola A, von Frijtag JC, Fermont PCJ, Gispen WH, Schrama LH, Kamal A, Spruijt BM (2006) Long-lasting modulation of the induction of LTD and LTP in rat hippocampal CA1 by behavioural stress and environmental enrichment. Eur J Neurosci 23:261–272.

Bading H (2017) Therapeutic targeting of the pathological triad of extrasynaptic NMDA receptor signaling in neurodegenerations. J Exp Med 214:569.

Baez MV, Cercato MC, Jerusalinsky DA (2018) NMDA Receptor Subunits Change after Synaptic Plasticity Induction and Learning and Memory Acquisition. Neural Plast 2018:1–11.

Báez-Mendoza R, Schultz W (2013) The role of the striatum in social behavior. Front Neurosci 7:233.

Balu DT, Coyle JT (2015) The NMDA receptor ‘glycine modulatory site’ in schizophrenia: d- serine, glycine, and beyond. Curr Opin Pharmacol 20:109–115.

Banerjee A, Larsen RS, Philpot BD, Paulsen O (2016) Roles of Presynaptic NMDA Receptors in Neurotransmission and Plasticity. Trends Neurosci 39:26–39.

Banko JL, Hou L, Klann E (2004) NMDA receptor activation results in PKA- and ERK- dependent Mnk1 activation and increased eIF4E phosphorylation in hippocampal area CA1. J Neurochem 91:462–470.

Baron-Cohen S (2005) Sex Differences in the Brain: Implications for Explaining Autism. Science 310:819–823.

130

Barria A, Malinow R (2005) NMDA Receptor Subunit Composition Controls Synaptic Plasticity by Regulating Binding to CaMKII. Neuron 48:289–301.

Barton S, Navarro SL, Buas MF, Schwarz Y, Gu H, Djukovic D, Raftery D, Kratz M, Neuhouser ML, Lampe JW (2015) Targeted plasma metabolome response to variations in dietary glycemic load in a randomized, controlled, crossover feeding trial in healthy adults. Food Funct 6:2949– 2956.

Baudry M, Bi X (2016) Calpain-1 and Calpain-2: The Yin and Yang of Synaptic Plasticity and Neurodegeneration. Trends Neurosci 39:235–245.

Baxter AJ, Brugha TS, Erskine HE, Scheurer RW, Vos T, Scott JG (2015) The epidemiology and global burden of autism spectrum disorders. Psychol Med 45:601–613.

Belsham DD, Wetsel WC, Mellon PL (1996) NMDA and nitric oxide act through the cGMP signal transduction pathway to repress hypothalamic gonadotropin-releasing hormone gene expression. EMBO J 15:538–547.

Ben Abdallah NM-B, Fuss J, Trusel M, Galsworthy MJ, Bobsin K, Colacicco G, Deacon RMJ, Riva MA, Kellendonk C, Sprengel R, Lipp H-P, Gass P (2011) The puzzle box as a simple and efficient behavioral test for exploring impairments of general cognition and executive functions in mouse models of schizophrenia. Exp Neurol 227:42–52.

Berckel BNM van, Evenblij CN, Loon BJAM van, Maas MF, Geld MAM van der, Wynne HJ, Ree JM van, Kahn RS (1999) D-Cycloserine Increases Positive Symptoms in Chronic Schizophrenic Patients When Administered in Addition to Antipsychotics: A Double-Blind, Parallel, Placebo-Controlled Study. Neuropsychopharmacology 21:203–210.

Berger SM, Bartsch D (2014) The role of L-type voltage-gated calcium channels Cav1.2 and Cav1.3 in normal and pathological brain function. Cell Tissue Res 357:463–476.

Berman FW, Murray TF (1997) Domoic acid neurotoxicity in cultured cerebellar granule neurons is mediated predominantly by NMDA receptors that are activated as a consequence of excitatory amino acid release. J Neurochem 69:693–703.

131

Berton O, McClung CA, DiLeone RJ, Krishnan V, Renthal W, Russo SJ, Graham D, Tsankova NM, Bolanos CA, Rios M, Monteggia LM, Self DW, Nestler EJ (2006) Essential Role of BDNF in the Mesolimbic Dopamine Pathway in Social Defeat Stress. 311:6.

Bezprozvanny I, Mattson MP (2008) Neuronal calcium mishandling and the pathogenesis of Alzheimer’s disease. Trends Neurosci 31:454–463.

Bi R, Foy MR, Vouimba R-M, Thompson RF, Baudry M (2001) Cyclic changes in estradiol regulate synaptic plasticity through the MAP kinase pathway. Proc Natl Acad Sci 98:13391– 13395.

Bickel S, Lipp HP, Umbricht D (2007) Impaired attentional modulation of auditory evoked potentials in N-methyl-d-aspartate NR1 hypomorphic mice. Genes Brain Behav 6:558–568.

Bickel S, Lipp H-P, Umbricht D (2008) Early auditory sensory processing deficits in mouse mutants with reduced NMDA receptor function. Neuropsychopharmacol Off Publ Am Coll Neuropsychopharmacol 33:1680–1689.

Birch PJ, Grossman CJ, Hayes AG (1988) Kynurenate and FG9041 have both competitive and non-competitive antagonist actions at excitatory amino acid receptors. Eur J Pharmacol 151:313– 315.

Black JE, Kodish IM, Grossman AW, Klintsova AY, Orlovskaya D, Vostrikov V, Uranova N, Greenough WT (2004) Pathology of Layer V Pyramidal Neurons in the Prefrontal Cortex of Patients With Schizophrenia. Am J Psychiatry 161:742–744.

Bliss TVP, Collingridge GL (1993) A synaptic model of memory: long-term potentiation in the hippocampus. Nature 361:31–39.

Bodarky CL, Halene TB, Ehrlichman RS, Banerjee A, Ray R, Hahn C-G, Jonak G, Siegel SJ (2009) Novel Environment and GABA Agonists Alter Event-Related Potentials in N-Methyl-D- aspartate NR1 Hypomorphic and Wild-Type Mice. J Pharmacol Exp Ther 331:308–318.

Bollard J, Miguela V, Ruiz de Galarreta M, Venkatesh A, Bian CB, Roberto MP, Tovar V, Sia D, Molina-Sánchez P, Nguyen CB, Nakagawa S, Llovet JM, Hoshida Y, Lujambio A (2017)

132

Palbociclib (PD-0332991), a selective CDK4/6 inhibitor, restricts tumour growth in preclinical models of hepatocellular carcinoma. Gut 66:1286–1296.

Bompard G, Caron E (2004) Regulation of WASP/WAVE proteins: making a long story short. J Cell Biol 166:957–962.

Borras L, Huguelet P, Eytan A (2007) Delusional “Pseudotranssexualism” in Schizophrenia. Psychiatry Interpers Biol Process 70:175–179.

Bouvier G, Bidoret C, Casado M, Paoletti P (2015) Presynaptic NMDA receptors: Roles and rules. Neuroscience 311:322–340.

Bowlby MR (1993) Pregnenolone sulfate potentiation of N-methyl-D-aspartate receptor channels in hippocampal neurons. Mol Pharmacol 43:813–819.

Brandon NJ, Sawa A (2011) Linking neurodevelopmental and synaptic theories of mental illness through DISC1. Nat Rev Neurosci 12:707–722.

Bredt DS, Nicoll RA (2003) AMPA Receptor Trafficking at Excitatory Synapses. Neuron 40:361–379.

Breier A, Buchanan RW, Kirkpatrick B, Davis OR, Irish D, Summerfelt A, Carpenter WTJ (1994) Effects of clozapine on positive and negative symptoms in outpatients with schizophrenia. Am J Psychiatry 151:20–26.

Buchanan RW, Breier A, Kirkpatrick B, Ball P, Carpenter WT (1998) Positive and Negative Symptom Response to Clozapine in Schizophrenic Patients With and Without the Deficit Syndrome. Am J Psychiatry 155:751–760.

Burnashev N, Monyer H, Seeburg PH, Sakmann B (1992a) Divalent ion permeability of AMPA receptor channels is dominated by the edited form of a single subunit. Neuron 8:189–198.

Burnashev N, Schoepfer R, Monyer H, Ruppersberg J, Gunther W, Seeburg P, Sakmann B (1992b) Control by Asparagine Residues of Calcium Permeability and Magnesium Blockade in the Nmda Receptor. SCIENCE 257:1415–1419.

133

Burnashev N, Szepetowski P (2015) NMDA receptor subunit mutations in neurodevelopmental disorders. Curr Opin Pharmacol 20:73–82.

Burns LH, Everitt BJ, Kelley AE, Robbins TW (1994) Glutamate-dopamine interactions in the ventral striatum: role in locomotor activity and responding with conditioned reinforcement. Psychopharmacology (Berl) 115:516–528.

Bustos FJ, Jury N, Martinez P, Ampuero E, Campos M, Abarzúa S, Jaramillo K, Ibing S, Mardones MD, Haensgen H, Kzhyshkowska J, Tevy MF, Neve R, Sanhueza M, Varela-Nallar L, Montecino M, van Zundert B (2017) NMDA receptor subunit composition controls dendritogenesis of hippocampal neurons through CAMKII, CREB-P, and H3K27ac. J Cell Physiol 232:3677–3692.

Buzsáki G, Moser EI (2013) Memory, navigation and theta rhythm in the hippocampal- entorhinal system. Nat Neurosci 16:130–138.

Calabresi P, Pisani A, Mercuri NB, Bernardi G (1992) Long-term potentiation in the striatum is unmasked by removing the voltage-dependent magnesium block of NMDA receptor channels. Eur J Neurosci 4:929–935.

Camargo LM, Collura V, Rain J-C, Mizuguchi K, Hermjakob H, Kerrien S, Bonnert TP, Whiting PJ, Brandon NJ (2007) Disrupted in Schizophrenia 1 Interactome: evidence for the close connectivity of risk genes and a potential synaptic basis for schizophrenia. Mol Psychiatry 12:74–86.

Cameron HA, McEwen BS, Gould E (1995) Regulation of adult neurogenesis by excitatory input and NMDA receptor activation in the dentate gyrus. J Neurosci Off J Soc Neurosci 15:4687– 4692.

Carlsson A, Hansson LO, Waters N, Carlsson ML (1999) A glutamatergic deficiency model of schizophrenia. Br J Psychiatry 174:2–6.

Carlsson M, Carlsson A (1990) Interactions between glutamatergic and monoaminergic systems within the basal ganglia-implications for schizophrenia and Parkinson’s disease. Trends Neurosci 13:272–276.

134

Carpenter WTJ, Strauss JS (1991) The prediction of outcome in schizophrenia. IV: Eleven-year follow-up of the Washington IPSS cohort. J Nerv Ment Dis 179:517–525.

Carvajal FJ, Mattison HA, Cerpa W (2016) Role of NMDA Receptor-Mediated Glutamatergic Signaling in Chronic and Acute Neuropathologies. Neural Plast 2016:1–20.

Carvalho FB, Mello CF, Marisco PC, Tonello R, Girardi BA, Ferreira J, Oliveira MS, Rubin MA (2012) Spermidine decreases Na+,K+-ATPase activity through NMDA receptor and protein kinase G activation in the hippocampus of rats. Eur J Pharmacol 684:79–86.

Cascella NG, Macciardi F, Cavallini C, Smeraldi E (1994) d-Cycloserine adjuvant therapy to conventional neuroleptic treatment in schizophrenia: an open-label study. J Neural Transm 95:105–111.

Ceglia I, Kim Y, Nairn AC, Greengard P (2010) Signaling pathways controlling the phosphorylation state of WAVE1, a regulator of actin polymerization. J Neurochem 114:182– 190.

Ceglia I, Lee K-W, Cahill ME, Graves SM, Dietz D, Surmeier DJ, Nestler EJ, Nairn AC, Greengard P, Kim Y (2017) WAVE1 in neurons expressing the D1 dopamine receptor regulates cellular and behavioral actions of cocaine. Proc Natl Acad Sci 114:1395–1400.

Cepeda C, Ariano MA, Calvert CR, Flores-Hernández J, Chandler SH, Leavitt BR, Hayden MR, Levine MS (2001) NMDA receptor function in mouse models of Huntington disease: NMDA Receptors in HD Models. J Neurosci Res 66:525–539.

Chatterton JE, Awobuluyi M, Premkumar LS, Takahashi H, Talantova M, Shin Y, Cui J, Tu S, Sevarino KA, Nakanishi N, Tong G, Lipton SA, Zhang D (2002) Excitatory glycine receptors containing the NR3 family of NMDA receptor subunits. Nature 415:793–798.

Chen EY, Tan CM, Kou Y, Duan Q, Wang Z, Meirelles G, Clark NR, Ma’ayan A (2013) Enrichr: interactive and collaborative HTML5 gene list enrichment analysis tool. BMC Bioinformatics 14:128.

135

Chen H, Fitzgerald R, Brown AT, Qureshi I, Breckenridge J, Kazi R, Wang Y, Wu Y, Zhang X, Mukunyadzi P, Eidt J, Moursi MM (2005) Identification of a homocysteine receptor in the peripheral endothelium and its role in proliferation. J Vasc Surg 41:853–860.

Chen S-Y, Huang P-H, Cheng H-J (2011) Disrupted-in-Schizophrenia 1-mediated axon guidance involves TRIO-RAC-PAK small GTPase pathway signaling. Proc Natl Acad Sci 108:5861– 5866.

Chen W, Ennes HS, McRoberts JA, Marvizón JC (2018) Mechanisms of μ-opioid receptor inhibition of NMDA receptor-induced substance P release in the rat spinal cord. Neuropharmacology 128:255–268.

Chez MG, Burton Q, Dowling T, Chang M, Khanna P, Kramer C (2007) Memantine as Adjunctive Therapy in Children Diagnosed With Autistic Spectrum Disorders: An Observation of Initial Clinical Response and Maintenance Tolerability. J Child Neurol 22:574–579.

Chien I-C, Lin C-H, Chou Y-J, Chou P (2011) Prevalence and Incidence of Autism Spectrum Disorders Among National Health Insurance Enrollees in Taiwan from 1996 to 2005. J Child Neurol 26:830–834.

Chiu CQ, Martenson JS, Yamazaki M, Natsume R, Sakimura K, Tomita S, Tavalin SJ, Higley MJ (2018) Input-Specific NMDAR-Dependent Potentiation of Dendritic GABAergic Inhibition. Neuron 97:368-377.e3.

Chizh BA, Headley PM, Tzschentke TM (2001) NMDA receptor antagonists as analgesics: focus on the NR2B subtype. Trends Pharmacol Sci 22:636–642.

Choleris E, Thomas AW, Kavaliers M, Prato FS (2001) A detailed ethological analysis of the mouse open field test: effects of diazepam, chlordiazepoxide and an extremely low frequency pulsed magnetic field. Neurosci Biobehav Rev 25:235–260.

Chong HY, Teoh SL, Wu DB-C, Kotirum S, Chiou C-F, Chaiyakunapruk N (2016) Global economic burden of schizophrenia: a systematic review. Neuropsychiatr Dis Treat 12:357–373.

136

Choy KHC, de Visser Y, Nichols NR, van den Buuse M (2008) Combined neonatal stress and young‐adult glucocorticoid stimulation in rats reduce BDNF expression in hippocampus: Effects on learning and memory. Hippocampus 18:655–667.

Christensen JG, Schreck R, Burrows J, Kuruganti P, Chan E, Le P, Chen J, Wang X, Ruslim L, Blake R, Lipson KE, Ramphal J, Do S, Cui JJ, Cherrington JM, Mendel DB (2003) A selective small molecule inhibitor of c-Met kinase inhibits c-Met-dependent phenotypes in vitro and exhibits cytoreductive antitumor activity in vivo. Cancer Res 63:7345–7355.

Clark K, MacKenzie KF, Petkevicius K, Kristariyanto Y, Zhang J, Choi HG, Peggie M, Plater L, Pedrioli PGA, McIver E, Gray NS, Arthur JSC, Cohen P (2012) Phosphorylation of CRTC3 by the salt-inducible kinases controls the interconversion of classically activated and regulatory macrophages. Proc Natl Acad Sci 109:16986–16991.

Clarke RJ, Glasgow NG, Johnson JW (2013) Mechanistic and Structural Determinants of NMDA Receptor Voltage-Dependent Gating and Slow Mg2+ Unblock. J Neurosci 33:4140– 4150.

Cline HT, Constantine-Paton M (1989) NMDA receptor antagonists disrupt the retinotectal topographic map. Neuron 3:413–426.

Colicos MA, Collins BE, Sailor MJ, Goda Y (2001) Remodeling of Synaptic Actin Induced by Photoconductive Stimulation. Cell 107:605–616.

Compagnone NA, Mellon SH (2000) Neurosteroids: Biosynthesis and Function of These Novel Neuromodulators. Front Neuroendocrinol 21:1–56.

Conrad CD, Galea LA, Kuroda Y, McEwen BS (1996) Chronic stress impairs rat spatial memory on the Y maze, and this effect is blocked by tianeptine treatment. Behav Neurosci 110:1321.

Contestabile A (2000) Roles of NMDA receptor activity and nitric oxide production in brain development. Brain Res Rev 32:476–509.

Corbetta S, Gualdoni S, Ciceri G, Monari M, Zuccaro E, Tybulewicz VLJ, de Curtis I (2009) Essential role of Rac1 and Rac3 GTPases in neuronal development. FASEB J 23:1347–1357.

137

Corlett PR, Honey GD, Krystal JH, Fletcher PC (2011) Glutamatergic Model Psychoses: Prediction Error, Learning and Inference. Neuropsychopharmacology 36:294–315.

Costa VV et al. (2017) N-Methyl-d-Aspartate (NMDA) Receptor Blockade Prevents Neuronal Death Induced by Zika Virus Infection Buchmeier MJ, ed. mBio 8:e00350-17.

Cox DA, Gottschalk MG, Wesseling H, Ernst A, Cooper JD, Bahn S (2016) Proteomic systems evaluation of the molecular validity of preclinical psychosis models compared to schizophrenia brain pathology. Schizophr Res 177:98–107.

Coyle JT (2012) NMDA Receptor and Schizophrenia: A Brief History. Schizophr Bull 38:920– 926.

Cull-Candy SG, Leszkiewicz DN (2004) Role of Distinct NMDA Receptor Subtypes at Central Synapses. Sci STKE 2004:re16.

Cummings KA, Popescu GK (2015) Glycine-dependent activation of NMDA receptors. J Gen Physiol 145:513–527.

Curtis DR, Watkins JC (1961) Analogues of glutamic and gamma-amino-n-butyric acids having potent actions on mammalian neurones. Nature 191:1010–1011.

Dahl JP, Wang-Dunlop J, Gonzales C, Goad ME, Mark RJ, Kwak SP (2003) Characterization of the WAVE1 knock-out mouse: implications for CNS development. J Neurosci 23:3343–3352.

Dal Bo G, St-Gelais F, Danik M, Williams S, Cotton M, Trudeau L-E (2004) Dopamine neurons in culture express VGLUT2 explaining their capacity to release glutamate at synapses in addition to dopamine: Dopamine neurons in culture express VGLUT2. J Neurochem 88:1398–1405.

Dalmau J, Tüzün E, Wu H, Masjuan J, Rossi JE, Voloschin A, Baehring JM, Shimazaki H, Koide R, King D, Mason W, Sansing LH, Dichter MA, Rosenfeld MR, Lynch DR (2007) Paraneoplastic anti-N-methyl-D-aspartate receptor encephalitis associated with ovarian teratoma. Ann Neurol 61:25–36.

138

Dalton GL, Wu DC, Wang YT, Floresco SB, Phillips AG (2012) NMDA GluN2A and GluN2B receptors play separate roles in the induction of LTP and LTD in the amygdala and in the acquisition and extinction of conditioned fear. Neuropharmacology 62:797–806.

Danysz W, Parsons CG (2003) The NMDA receptor antagonist memantine as a symptomatological and neuroprotective treatment for Alzheimer’s disease: preclinical evidence. Int J Geriatr Psychiatry 18:S23–S32.

Darlington TM, Ehringer MA, Larson C, Phang TL, Radcliffe RA (2013) Transcriptome analysis of Inbred Long Sleep and Inbred Short Sleep mice: Transcriptome analysis of ILS and ISS mice. Genes Brain Behav 12:263–274.

Davis SM, Lees KR, Albers GW, Diener HC, Markabi S, Karlsson G, Norris J (2000) Selfotel in Acute Ischemic Stroke : Possible Neurotoxic Effects of an NMDA Antagonist. Stroke 31:347– 354.

De Luca A, Maiello MR, D’Alessio A, Pergameno M, Normanno N (2012) The RAS/RAF/MEK/ERK and the PI3K/AKT signalling pathways: role in cancer pathogenesis and implications for therapeutic approaches. Expert Opin Ther Targets 16:S17–S27.

De Pittà M, Brunel N, Volterra A (2016) Astrocytes: Orchestrating synaptic plasticity? Neuroscience 323:43–61.

De Roo M, Klauser P, Muller D (2008) LTP Promotes a Selective Long-Term Stabilization and Clustering of Dendritic Spines Sheng M, ed. PLoS Biol 6:e219.

De Sarro G, Gratteri S, Naccari F, Pasculli MP, De Sarro A (2000) Influence of d-cycloserine on the anticonvulsant activity of some antiepileptic drugs against audiogenic seizures in DBA/2 mice. Epilepsy Res 40:109–121. de Vries ALC, Noens ILJ, Cohen-Kettenis PT, van Berckelaer-Onnes IA, Doreleijers TA (2010) Autism Spectrum Disorders in Gender Dysphoric Children and Adolescents. J Autism Dev Disord 40:930–936.

Deacon RMJ, Rawlins JNP (2006) T-maze alternation in the rodent. Nat Protoc 1:7–12.

139

Dean O, Bush AI, Berk M, Copolov DL, van den Buuse M (2009) Glutathione depletion in the brain disrupts short-term spatial memory in the Y-maze in rats and mice. Behav Brain Res 198:258–262.

Deardorff WJ, Feen E, Grossberg GT (2015) The Use of Cholinesterase Inhibitors Across All Stages of Alzheimer’s Disease. Drugs Aging 32:537–547.

Devinsky O, Vezzani A, Najjar S, De Lanerolle NC, Rogawski MA (2013) Glia and epilepsy: excitability and inflammation. Trends Neurosci 36:174–184.

Dillon C, Goda Y (2005) The actin cytoskeleton: integrating form and function at the synapse. Annu Rev Neurosci 28:25–55.

Dong X-D, Mann MK, Kumar U, Svensson P, Arendt-Nielsen L, Hu JW, Sessle BJ, Cairns BE (2007) Sex-related differences in NMDA-evoked rat masseter muscle afferent discharge result from estrogen-mediated modulation of peripheral NMDA receptor activity. Neuroscience 146:822–832.

Dore K, Aow J, Malinow R (2016) The Emergence of NMDA Receptor Metabotropic Function: Insights from Imaging. Front Synaptic Neurosci 8:20. du Bois TM, Huang X-F (2007) Early brain development disruption from NMDA receptor hypofunction: Relevance to schizophrenia. Brain Res Rev 53:260–270.

Du J, Li X-H, Li Y-J (2016) Glutamate in peripheral organs: Biology and pharmacology. Eur J Pharmacol 784:42–48.

Dubal DB, Zhu H, Yu J, Rau SW, Shughrue PJ, Merchenthaler I, Kindy MS, Wise PM (2001) Estrogen receptor α, not β, is a critical link in estradiol-mediated protection against brain injury. Proc Natl Acad Sci 98:1952.

Dumas TC (2005) Developmental regulation of cognitive abilities: Modified composition of a molecular switch turns on associative learning. Prog Neurobiol 76:189–211.

140

Duncan EJ, Szilagyi S, Schwartz MP, Bugarski-Kirola D, Kunzova A, Negi S, Stephanides M, Efferen TR, Angrist B, Peselow E, Corwin J, Gonzenbach S, Rotrosen JP (2004a) Effects of d- cycloserine on negative symptoms in schizophrenia. Schizophr Res 71:239–248.

Duncan G, Johnson K, Breese G (1993) Topographic patterns of brain activity in response to swim stress: assessment by 2-deoxyglucose uptake and expression of Fos-like immunoreactivity. J Neurosci 13:3932.

Duncan G, Moy S, Lieberman J, Koller B (2006a) Typical and atypical antipsychotic drug effects on locomotor hyperactivity and deficits in sensorimotor gating in a genetic model of NMDA receptor hypofunction. Pharmacol Biochem Behav 85:481–491.

Duncan GE, Inada K, Farrington JS, Koller BH (2008) Seizure responses and induction of Fos by the NMDA Agonist (tetrazol-5-yl)glycine in a genetic model of NMDA receptor hypofunction. Brain Res 1221:41–48.

Duncan GE, Inada K, Farrington JS, Koller BH, Moy SS (2009) Neural activation deficits in a mouse genetic model of NMDA receptor hypofunction in tests of social aggression and swim stress. Brain Res 1265:186–195.

Duncan GE, Inada K, Koller BH, Moy SS (2010) Increased sensitivity to kainic acid in a genetic model of reduced NMDA receptor function. Brain Res 1307:166–176.

Duncan GE, Koller BH, Moy SS (2012) Effects of the selective kainate receptor antagonist ACET on altered sensorimotor gating in a genetic model of reduced NMDA receptor function. Brain Res 1443:98–105.

Duncan GE, Miyamoto S, Gu H, Lieberman JA, Koller BH, Snouwaert JN (2002) Alterations in regional brain metabolism in genetic and pharmacological models of reduced NMDA receptor function. Brain Res 951:166–176.

Duncan GE, Moy SS, Lieberman JA, Koller BH (2006b) Effects of haloperidol, clozapine, and quetiapine on sensorimotor gating in a genetic model of reduced NMDA receptor function. Psychopharmacology (Berl) 184:190–200.

141

Duncan GE, Moy SS, Perez A, Eddy DM, Zinzow WM, Lieberman JA, Snouwaert JN, Koller BH (2004b) Deficits in sensorimotor gating and tests of social behavior in a genetic model of reduced NMDA receptor function. Behav Brain Res 153:507–519.

Dzirasa K, Ramsey AJ, Takahashi DY, Stapleton J, Potes JM, Williams JK, Gainetdinov RR, Sameshima K, Caron MG, Nicolelis MAL (2009) Hyperdopaminergia and NMDA Receptor Hypofunction Disrupt Neural Phase Signaling. J Neurosci 29:8215–8224.

Eden S, Rohatgi R, Podtelejnikov AV, Mann M, Kirschner MW (2002) Mechanism of regulation of WAVE1-induced actin nucleation by Rac1 and Nck. Nature 418:790–793.

Elsabbagh M, Divan G, Koh Y-J, Kim YS, Kauchali S, Marcín C, Montiel-Nava C, Patel V, Paula CS, Wang C, Yasamy MT, Fombonne E (2012) Global Prevalence of Autism and Other Pervasive Developmental Disorders: Global epidemiology of autism. Autism Res 5:160–179.

Emnett CM, Eisenman LN, Mohan J, Taylor AA, Doherty JJ, Paul SM, Zorumski CF, Mennerick S (2015) Interaction between positive allosteric modulators and trapping blockers of the NMDA receptor channel: Positive modulators and clinical NMDA receptor blockers. Br J Pharmacol 172:1333–1347.

Eranti SV, MacCabe JH, Bundy H, Murray RM (2013) Gender difference in age at onset of schizophrenia: a meta-analysis. Psychol Med 43:155–167.

Erickson CA, Posey DJ, Stigler KA, Mullett J, Katschke AR, McDougle CJ (2007) A retrospective study of memantine in children and adolescents with pervasive developmental disorders. Psychopharmacology (Berl) 191:141–147.

Erturk A, Wang Y, Sheng M (2014) Local Pruning of Dendrites and Spines by Caspase-3- Dependent and Proteasome-Limited Mechanisms. J Neurosci 34:1672–1688.

Estrada-Rojo F, Morales-Gomez J, Coballase-Urrutia E, Martinez-Vargas M, Navarro L (2018) Diurnal variation of NMDA receptor expression in the rat cerebral cortex is associated with traumatic brain injury damage. BMC Res Notes 11:150.

Evins AE, Amico E, Posever TA, Toker R, Goff DC (2002) d-Cycloserine added to risperidone in patients with primary negative symptoms of schizophrenia. Schizophr Res 56:19–23.

142

Fan M, Raymond L (2007) N-Methyl-d-aspartate (NMDA) receptor function and excitotoxicity in Huntington’s disease. Prog Neurobiol 81:272–293.

Faul F, Erdfelder E, Buchner A, Lang A-G (2009) Statistical power analyses using G*Power 3.1: Tests for correlation and regression analyses. Behav Res Methods 41:1149–1160.

Faul F, Erdfelder E, Lang A-G, Buchner A (2007) G* Power 3: A flexible statistical power analysis program for the social, behavioral, and biomedical sciences. Behav Res Methods 39:175–191.

Feld GB, Lange T, Gais S, Born J (2013) Sleep-Dependent Declarative Memory Consolidation- Unaffected after Blocking NMDA or AMPA Receptors but Enhanced by NMDA Coagonist D- Cycloserine. Neuropsychopharmacology 38:2688–2697.

Fell J, Klaver P, Lehnertz K, Grunwald T, Schaller C, Elger CE, Fernández G (2001) Human memory formation is accompanied by rhinal–hippocampal coupling and decoupling. Nat Neurosci 4:1259–1264.

Ferris MJ, Milenkovic M, Liu S, Mielnik CA, Beerepoot P, John CE, España RA, Sotnikova TD, Gainetdinov RR, Borgland SL, Jones SR, Ramsey AJ (2014) Sustained N-methyl-d-aspartate receptor hypofunction remodels the dopamine system and impairs phasic signaling. Eur J Neurosci 40:2255–2263.

Fester L, Rune GM (2015) Sexual neurosteroids and synaptic plasticity in the hippocampus. Brain Res 1621:162–169.

Filardo EJ, Quinn JA, Bland KI, Frackelton AR (2018) Estrogen-Induced Activation of Erk- 1 and Erk-2 Requires the G Protein- Coupled Receptor Homolog, GPR30, and Occurs via Trans- Activation of the Epidermal Growth Factor Receptor through Release of HB-EGF. 14:12.

Fiske BK, Brunjes PC (2001) NMDA receptor regulation of cell death in the rat olfactory bulb. J Neurobiol 47:223–232.

Florance-Ryan N, Dalmau J (2010) Update on anti-N-methyl-D-aspartate receptor encephalitis in children and adolescents: Curr Opin Pediatr 22:739–744.

143

Floresco SB, Seamans JK, Phillips AG (1997) Selective roles for hippocampal, prefrontal cortical, and ventral striatal circuits in radial-arm maze tasks with or without a delay. J Neurosci 17:1880–1890.

Forrest D, Yuzaki M, Soares HD, Ng L, Luk DC, Sheng M, Stewart CL, Morgan JI, Connor JA, Curran T (1994) Targeted disruption of NMDA receptor 1 gene abolishes NMDA response and results in neonatal death. Neuron 13:325–338.

Fradley RL, O’Meara GF, Newman RJ, Andrieux A, Job D, Reynolds DS (2005) STOP knockout and NMDA NR1 hypomorphic mice exhibit deficits in sensorimotor gating. Behav Brain Res 163:257–264.

Frank RA, Grant SG (2017) Supramolecular organization of NMDA receptors and the postsynaptic density. Curr Opin Neurobiol 45:139–147.

Frank RAW, Komiyama NH, Ryan TJ, Zhu F, O’Dell TJ, Grant SGN (2016) NMDA receptors are selectively partitioned into complexes and supercomplexes during synapse maturation. Nat Commun 7:11264.

Fukuchi M, Tabuchi A, Kuwana Y, Watanabe S, Inoue M, Takasaki I, Izumi H, Tanaka A, Inoue R, Mori H, Komatsu H, Takemori H, Okuno H, Bito H, Tsuda M (2015) Neuromodulatory Effect of G s- or G q-Coupled G-Protein-Coupled Receptor on NMDA Receptor Selectively Activates the NMDA Receptor/Ca2+/Calcineurin/cAMP Response Element-Binding Protein- Regulated Transcriptional Coactivator 1 Pathway to Effectively Induce Brain-Derived Neurotrophic Factor Expression in Neurons. J Neurosci 35:5606–5624.

Fukuyo Y, Hunt CR, Horikoshi N (2010) Geldanamycin and its anti-cancer activities. Cancer Lett 290:24–35.

Gandal MJ, Anderson RL, Billingslea EN, Carlson GC, Roberts TPL, Siegel SJ (2012) Mice with reduced NMDA receptor expression: more consistent with autism than schizophrenia? Genes Brain Behav 11:740–750.

Ganz ML (2007) The Lifetime Distribution of the Incremental Societal Costs of Autism. Arch Pediatr Adolesc Med 161:343.

144

Gao J, Duan B, Wang D-G, Deng X-H, Zhang G-Y, Xu L, Xu T-L (2005) Coupling between NMDA Receptor and Acid-Sensing Ion Channel Contributes to Ischemic Neuronal Death. Neuron 48:635–646.

Gao X, Mi Y, Guo N, Hu Z, Hu F, Liu D, Gao L, Gou X, Jin W (2016) Disrupted in schizophrenia 1 (DISC1) inhibits glioblastoma development by regulating mitochondria dynamics. Oncotarget 7:85963–85974.

Gardoni F, Schrama LH, van Dalen JJW, Gispen WH, Cattabeni F, Di Luca M (1999) αCaMKII binding to the C-terminal tail of NMDA receptor subunit NR2A and its modulation by autophosphorylation. FEBS Lett 456:394–398.

Garey LJ, Ong WY, Patel TS, Kanani M, Davis A, Mortimer AM, Barnes TRE, Hirsch SR (1998) Reduced dendritic spine density on cerebral cortical pyramidal neurons in schizophrenia. J Neurol Neurosurg Psychiatry 65:446–453.

Gauthier S, Reisberg B, Zaudig M, Petersen RC, Ritchie K, Broich K, Belleville S, Brodaty H, Bennett D, Chertkow H, Cummings JL, de Leon M, Feldman H, Ganguli M, Hampel H, Scheltens P, Tierney MC, Whitehouse P, Winblad B (2006) Mild cognitive impairment. The Lancet 367:1262–1270.

Ge S, Yang C, Hsu K, Ming G, Song H (2007) A Critical Period for Enhanced Synaptic Plasticity in Newly Generated Neurons of the Adult Brain. Neuron 54:559–566.

Ghaleiha A, Asadabadi M, Mohammadi M-R, Shahei M, Tabrizi M, Hajiaghaee R, Hassanzadeh E, Akhondzadeh S (2013) Memantine as adjunctive treatment to risperidone in children with autistic disorder: a randomized, double-blind, placebo-controlled trial. Int J Neuropsychopharmacol 16:783–789.

Ghasemi M, Schachter SC (2011) The NMDA receptor complex as a therapeutic target in epilepsy: a review. Epilepsy Behav 22:617–640.

Ghazalpour A et al. (2011) Comparative Analysis of Proteome and Transcriptome Variation in Mouse Snyder M, ed. PLoS Genet 7:e1001393.

145

Gibson DA, Harris BR, Rogers DT, Littleton JM (2002) Radioligand binding studies reveal agmatine is a more selective antagonist for a polyamine-site on the NMDA receptor than arcaine or ifenprodil. Brain Res 952:71–77.

Gilani AI, Chohan MO, Inan M, Schobel SA, Chaudhury NH, Paskewitz S, Chuhma N, Glickstein S, Merker RJ, Xu Q, Small SA, Anderson SA, Ross ME, Moore H (2014) Interneuron precursor transplants in adult hippocampus reverse psychosis-relevant features in a mouse model of hippocampal disinhibition. Proc Natl Acad Sci 111:7450–7455.

Gilbert ME, Mack CM (1990) The NMDA antagonist, MK-801, suppresses long-term potentiation, kindling, and kindling-induced potentiation in the perforant path of the unanesthetized rat. Brain Res 519:89–96.

Gill SS, Mueller RW, Mcguire PF, Pulido OM (2000) Potential Target Sites in Peripheral Tissues for Excitatory Neurotransmission and Excitotoxicity. Toxicol Pathol 28:277–284.

Gilmour G, Dix S, Fellini L, Gastambide F, Plath N, Steckler T, Talpos J, Tricklebank M (2012) NMDA receptors, cognition and schizophrenia – Testing the validity of the NMDA receptor hypofunction hypothesis. Neuropharmacology 62:1401–1412.

Girling KD, Demers M-J, Laine J, Zhang S, Wang YT, Graham RK (2018) Activation of caspase-6 and cleavage of caspase-6 substrates is an early event in NMDA receptor-mediated excitotoxicity. J Neurosci Res 96:391–406.

Giulian D, Vaca K, Noonan CA (1990) Secretion of neurotoxins by mononuclear phagocytes infected with HIV-1. Science 250:1593–1596.

Glantz LA, Lewis DA (2000) Decreased Dendritic Spine Density on Prefrontal Cortical Pyramidal Neurons in Schizophrenia. Arch Gen Psychiatry 57:65.

Goff D, Cather C, Gottlieb J, Evins A, Walsh J, Raeke L, Otto M, Schoenfeld D, Green M (2008) Once-weekly d-cycloserine effects on negative symptoms and cognition in schizophrenia: An exploratory study. Schizophr Res 106:320–327.

Goff DC, Henderson DC, Evins AE, Amico E (1999) A placebo-controlled crossover trial of d- cycloserine added to clozapine in patients with schizophrenia. Biol Psychiatry 45:512–514.

146

Goff DC, Herz L, Posever T, Shih V, Tsai G, Henderson DC, Freudenreich O, Evins AE, Yovel I, Zhang H, Schoenfeld D (2005) A six-month, placebo-controlled trial of d-cycloserine co- administered with conventional antipsychotics in schizophrenia patients. Psychopharmacology (Berl) 179:144–150.

Goldstein G, Minshew NJ, Allen DN, Seaton BE (2018) High-functioning autism and schizophrenia A comparison of an early and late onset neurodevelopmental disorder. Arch Clin Neuropsychol:15.

Gong R, Park CS, Abbassi NR, Tang S-J (2006) Roles of Glutamate Receptors and the Mammalian Target of Rapamycin (mTOR) Signaling Pathway in Activity-dependent Dendritic Protein Synthesis in Hippocampal Neurons. J Biol Chem 281:18802–18815.

Gonzalez-Zulueta M, Feldman AB, Klesse LJ, Kalb RG, Dillman JF, Parada LF, Dawson TM, Dawson VL (2000) Requirement for nitric oxide activation of p21ras/extracellular regulated kinase in neuronal ischemic preconditioning. Proc Natl Acad Sci 97:436–441.

Gould E, Cameron HA (1996) Regulation of Neuronal Birth, Migration and Death in the Rat Dentate Gyrus. Dev Neurosci 18:22–35.

Gould E, McEwen BS, Tanapat P, Galea LAM, Fuchs E (1997) Neurogenesis in the Dentate Gyrus of the Adult Tree Shrew Is Regulated by Psychosocial Stress and NMDA Receptor Activation. J Neurosci 17:2492.

Grabb MC, Choi DW (1999) Ischemic Tolerance in Murine Cortical Cell Culture: Critical Role for NMDA Receptors. J Neurosci 19:1657.

Grannan MD, Mielnik CA, Moran SP, Gould RW, Ball J, Lu Z, Bubser M, Ramsey AJ, Abe M, Cho HP, Nance KD, Blobaum AL, Niswender CM, Conn PJ, Lindsley CW, Jones CK (2016) Prefrontal Cortex-Mediated Impairments in a Genetic Model of NMDA Receptor Hypofunction

Are Reversed by the Novel M 1 PAM VU6004256. ACS Chem Neurosci 7:1706–1716.

Gray AL, Hyde TM, Deep-Soboslay A, Kleinman JE, Sodhi MS (2015) Sex differences in glutamate receptor gene expression in major depression and suicide. Mol Psychiatry 20:1057– 1068.

147

Gray JA, Zito K, Hell JW (2016) Non-ionotropic signaling by the NMDA receptor: controversy and opportunity. F1000Research 5.

Green MF (2016) Impact of Cognitive and Social Cognitive Impairment on Functional Outcomes in Patients With Schizophrenia. J Clin Psychiatry:8–11.

Gregory KJ et al. (2013) N-Aryl Piperazine Metabotropic Glutamate Receptor 5 Positive Allosteric Modulators Possess Efficacy in Preclinical Models of NMDA Hypofunction and Cognitive Enhancement. J Pharmacol Exp Ther 347:438–457.

Gruzelier JH (1994) Syndromes of schizophrenia and schizotypy, hemispheric imbalance and sex differences: implications for developmental psychopathology. Int J Psychophysiol 18:167–178.

Gu Q, Wang C (2018) Chapter 18 - The NMDA Receptors: Physiology and Neurotoxicity in the Developing Brain. In: Handbook of Developmental Neurotoxicology (Second Edition), Second Edition. (Slikker W, Paule MG, Wang C, eds), pp 207–214. Academic Press. Available at: https://www.sciencedirect.com/science/article/pii/B9780128094051000183.

Guidetti P, Luthicarter R, Augood S, Schwarcz R (2004) Neostriatal and cortical quinolinate levels are increased in early grade Huntington’s disease. Neurobiol Dis 17:455–461.

Gunaydin LA, Grosenick L, Finkelstein JC, Kauvar IV, Fenno LE, Adhikari A, Lammel S, Mirzabekov JJ, Airan RD, Zalocusky KA, Tye KM, Anikeeva P, Malenka RC, Deisseroth K (2014) Natural Neural Projection Dynamics Underlying Social Behavior. Cell 157:1535–1551.

Haditsch U, Leone DP, Farinelli M, Chrostek-Grashoff A, Brakebusch C, Mansuy IM, McConnell SK, Palmer TD (2009) A central role for the small GTPase Rac1 in hippocampal plasticity and spatial learning and memory. Mol Cell Neurosci 41:409–419.

Häfner H (2003) Gender differences in schizophrenia. Psychoneuroendocrinology 28:17–54.

Haiying G, Mingjie H, Lingyu Z, Qingxiang W, Haisong W, Bingxi Z (2017) Anesthetics inhibit extracellular signal-regulated Kinase1/2 phosphorylation via NMDA receptor, phospholipase C and protein kinase C in mouse hippocampal slices. Neurochem Int 103:36–44.

148

Halene TB, Ehrlichman RS, Liang Y, Christian EP, Jonak GJ, Gur TL, Blendy JA, Dow HC, Brodkin ES, Schneider F, Gur RC, Siegel SJ (2009a) Assessment of NMDA receptor NR1 subunit hypofunction in mice as a model for schizophrenia. Genes Brain Behav 8:661–675.

Halene TB, Talmud J, Jonak GJ, Schneider F, Siegel SJ (2009b) Predator odor modulates auditory event-related potentials in mice: NeuroReport 20:1260–1264.

Halladay AK, Bishop S, Constantino JN, Daniels AM, Koenig K, Palmer K, Messinger D, Pelphrey K, Sanders SJ, Singer AT, Taylor JL, Szatmari P (2015) Sex and gender differences in autism spectrum disorder: summarizing evidence gaps and identifying emerging areas of priority. Mol Autism 6:36.

Hansen KB, Yi F, Perszyk RE, Menniti FS, Traynelis SF (2017) NMDA Receptors in the Central Nervous System. In: NMDA Receptors: Methods and Protocols (Burnashev N, Szepetowski P, eds), pp 1–80. New York, NY: Springer New York. Available at: https://doi.org/10.1007/978-1- 4939-7321-7_1.

Hansen SN, Schendel DE, Parner ET (2015) Explaining the Increase in the Prevalence of Autism Spectrum Disorders: The Proportion Attributable to Changes in Reporting Practices. JAMA Pediatr 169:56.

Hardingham GE, Bading H (2003) The Yin and Yang of NMDA receptor signalling. Trends Neurosci 26:81–89.

Hardingham GE, Bading H (2010) Synaptic versus extrasynaptic NMDA receptor signalling: implications for neurodegenerative disorders. Nat Rev Neurosci 11:682–696.

Hardingham GE, Do KQ (2016) Linking early-life NMDAR hypofunction and oxidative stress in schizophrenia pathogenesis. Nat Rev Neurosci 17:125.

Hardingham GE, Fukunaga Y, Bading H (2002) Extrasynaptic NMDARs oppose synaptic NMDARs by triggering CREB shut-off and cell death pathways. Nat Neurosci 5:405–414.

Hashimshony T, Wagner F, Sher N, Yanai I (2012) CEL-Seq: Single-Cell RNA-Seq by Multiplexed Linear Amplification. Cell Rep 2:666–673.

149

Hatton CJ, Paoletti P (2005) Modulation of Triheteromeric NMDA Receptors by N-Terminal Domain Ligands. Neuron 46:261–274.

Haughey NJ, Nath A, Mattson MP, Slevin JT, Geiger JD (2001) HIV-1 Tat through phosphorylation of NMDA receptors potentiates glutamate excitotoxicity: Tat potentiates glutamate excitotoxicity. J Neurochem 78:457–467.

Hayashi-Takagi A, Araki Y, Nakamura M, Vollrath B, Duron SG, Yan Z, Kasai H, Huganir RL, Campbell DA, Sawa A (2014) PAKs inhibitors ameliorate schizophrenia-associated dendritic spine deterioration in vitro and in vivo during late adolescence. Proc Natl Acad Sci 111:6461– 6466.

Hayashi-Takagi A, Takaki M, Graziane N, Seshadri S, Murdoch H, Dunlop AJ, Makino Y, Seshadri AJ, Ishizuka K, Srivastava DP, Xie Z, Baraban JM, Houslay MD, Tomoda T, Brandon NJ, Kamiya A, Yan Z, Penzes P, Sawa A (2010) Disrupted-in-Schizophrenia 1 (DISC1) regulates spines of the glutamate synapse via Rac1. Nat Neurosci 13:327–332.

Hazlett EA, Romero MJ, Haznedar MM, New AS, Goldstein KE, Newmark RE, Siever LJ, Buchsbaum MS (2007) Deficient attentional modulation of startle eyeblink is associated with symptom severity in the schizophrenia spectrum. Schizophr Res 93:288–295.

Heck N, Golbs A, Riedemann T, Sun J-J, Lessmann V, Luhmann HJ (2008) Activity-Dependent Regulation of Neuronal Apoptosis in Neonatal Mouse Cerebral Cortex. Cereb Cortex 18:1335– 1349.

Hedrick NG, Yasuda R (2017) Regulation of Rho GTPase proteins during spine structural plasticity for the control of local dendritic plasticity. Curr Opin Neurobiol 45:193–201.

Hegde PS, White IR, Debouck C (2003) Interplay of transcriptomics and proteomics. Curr Opin Biotechnol 14:647–651.

Henson MA, Roberts AC, Pérez-Otaño I, Philpot BD (2010) Influence of the NR3A subunit on NMDA receptor functions. Prog Neurobiol 91:23–37.

150

Henson MA, Roberts AC, Salimi K, Vadlamudi S, Hamer RM, Gilmore JH, Jarskog LF, Philpot BD (2008) Developmental Regulation of the NMDA Receptor Subunits, NR3A and NR1, in Human Prefrontal Cortex. Cereb Cortex 18:2560–2573.

Heras-Sandoval D, Pérez-Rojas JM, Hernández-Damián J, Pedraza-Chaverri J (2014) The role of PI3K/AKT/mTOR pathway in the modulation of autophagy and the clearance of protein aggregates in neurodegeneration. Cell Signal 26:2694–2701.

Heresco-Levy U, Ermilov M, Shimoni J, Shapira B, Silipo G, Javitt DC (2002) Placebo- Controlled Trial of d-Cycloserine Added to Conventional Neuroleptics, Olanzapine, or Risperidone in Schizophrenia. Am J Psychiatry 159:480–482.

Heresco-Levy U, Javitt DC, Ermilov M, Silipo G, Shimoni J (1998) Double-blind, placebo- controlled, crossover trial of D-cycloserine adjuvant therapy for treatment-resistant schizophrenia. Int J Neuropsychopharmacol 1:131–135.

Herrmann CS, Fründ I, Lenz D (2010) Human gamma-band activity: A review on cognitive and behavioral correlates and network models. Neurosci Biobehav Rev 34:981–992.

Herrmann N, Li A, Lanctôt K (2011) Memantine in dementia: a review of the current evidence. Expert Opin Pharmacother 12:787–800.

Herron CE, Lester RA, Coan EJ, Collingridge GL (1986) Frequency-dependent involvement of NMDA receptors in the hippocampus: a novel synaptic mechanism. Nature 322:265–268.

Hewitt D (2000) The use of NMDA-receptor antagonists in the treatment of chronic pain. Clin J Pain 16:S73—9.

Hickmott PW, Constantine-Paton M (1997) Experimental Down-Regulation of the NMDA Channel Associated With Synapse Pruning. J Neurophysiol 78:1096–1107.

Hill JJ, Hashimoto T, Lewis DA (2006) Molecular mechanisms contributing to dendritic spine alterations in the prefrontal cortex of subjects with schizophrenia. Mol Psychiatry 11:557–566.

Hill TC, Zito K (2013) LTP-Induced Long-Term Stabilization of Individual Nascent Dendritic Spines. J Neurosci 33:678–686.

151

Hirabayashi S, Cagan RL (2015) Salt-inducible kinases mediate nutrient-sensing to link dietary sugar and tumorigenesis in Drosophila Pan D, ed. eLife 4:e08501.

Hirai H, Kirsch J, Laube B, Betz H, Kuhse J (1996) The glycine binding site of the N-methyl-D- aspartate receptor subunit NR1: identification of novel determinants of co-agonist potentiation in the extracellular M3-M4 loop region. Proc Natl Acad Sci 93:6031–6036.

Hisamoto K, Ohmichi M, Kurachi H, Hayakawa J, Kanda Y, Nishio Y, Adachi K, Tasaka K, Miyoshi E, Fujiwara N, Taniguchi N, Murata Y (2001) Estrogen Induces the Akt-dependent Activation of Endothelial Nitric-oxide Synthase in Vascular Endothelial Cells. J Biol Chem 276:3459–3467.

Hodes GE et al. (2015) Sex Differences in Nucleus Accumbens Transcriptome Profiles Associated with Susceptibility versus Resilience to Subchronic Variable Stress. J Neurosci 35:16362–16376.

Hogan-Cann AD, Anderson CM (2016) Physiological Roles of Non-Neuronal NMDA Receptors. Trends Pharmacol Sci 37:750–767.

Hojo Y, Murakami G, Mukai H, Higo S, Hatanaka Y, Ogiue-Ikeda M, Ishii H, Kimoto T, Kawato S (2008) Estrogen synthesis in the brain—Role in synaptic plasticity and memory. Mol Cell Endocrinol 290:31–43.

Horlin C, Falkmer M, Parsons R, Albrecht MA, Falkmer T (2014) The Cost of Autism Spectrum Disorders Mulle JG, ed. PLoS ONE 9:e106552.

Howlin P, Goode S, Hutton J, Rutter M (2004) Adult outcome for children with autism. J Child Psychol Psychiatry 45:212–229.

Hsu C, Hsieh Y-L, Yang R-C, Hsu H-K (2000) Blockage of N-Methyl-D-Aspartate Receptors Decreases Testosterone Levels and Enhances Postnatal Neuronal Apoptosis in the Preoptic Area of Male Rats. Neuroendocrinology 71:301–307.

Huang G-H, Sun Z-L, Li H-J, Feng D-F (2017) Rho GTPase-activating proteins: Regulators of Rho GTPase activity in neuronal development and CNS diseases. Mol Cell Neurosci 80:18–31.

152

Huang YH, Schlüter OM, Dong Y (2015) Silent Synapses Speak Up: Updates of the Neural Rejuvenation Hypothesis of Drug Addiction. The Neuroscientist 21:451–459.

Huang Y-S, Huang W-C, Li C-W, Chuang L-T (2011) Eicosadienoic acid differentially modulates production of pro-inflammatory modulators in murine macrophages. Mol Cell Biochem 358:85–94.

Hughes RN (2004) The value of spontaneous alternation behavior (SAB) as a test of retention in pharmacological investigations of memory. Neurosci Biobehav Rev 28:497–505.

Husi H, Ward MA, Choudhary JS, Blackstock WP, Grant SGN (2000) Proteomic analysis of NMDA receptor–adhesion protein signaling complexes. Nat Neurosci 3:661–669.

Hynd MR, Scott HL, Dodd PR (2004) Glutamate-mediated excitotoxicity and neurodegeneration in Alzheimer’s disease. Neurochem Int 45:583–595.

Idring S, Lundberg M, Sturm H, Dalman C, Gumpert C, Rai D, Lee BK, Magnusson C (2015) Changes in Prevalence of Autism Spectrum Disorders in 2001–2011: Findings from the Stockholm Youth Cohort. J Autism Dev Disord 45:1766–1773.

Ikeda K, Fukushima T, Ogura H, Tsukui T, Mishina M, Muramatsu M, Inoue S (2010) Estrogen regulates the expression of N -methyl- D -aspartate (NMDA) receptor subunit epsilon 4 (Grin2d), that is essential for the normal sexual behavior in female mice. FEBS Lett 584:806–810.

Ikonomidou C, Bosch F, Miksa M, Bittigau P, Vöckler J, Dikranian K, Tenkova TI, Stefovska V, Turski L, Olney JW (1999) Blockade of NMDA Receptors and Apoptotic Neurodegeneration in the Developing Brain. Science 283:70.

Ikonomidou C, Turski L (2002) Why did NMDA receptor antagonists fail clinical trials for stroke and traumatic brain injury? Lancet Neurol 1:383–386.

Inada K, Farrington JS, Moy SS, Koller BH, Duncan GE (2007) Assessment of NMDA receptor activation in vivo by Fos induction after challenge with the direct NMDA agonist (tetrazol-5- yl)glycine: effects of clozapine and haloperidol. J Neural Transm 114:899–908.

153

Indersmitten T, Tran CH, Cepeda C, Levine MS (2015) Altered excitatory and inhibitory inputs to striatal medium-sized spiny neurons and cortical pyramidal neurons in the Q175 mouse model of Huntington’s disease. J Neurophysiol 113:2953–2966.

Inquimbert P, Moll M, Latremoliere A, Tong C-K, Whang J, Sheehan GF, Smith BM, Korb E, Athié MCP, Babaniyi O, Ghasemlou N, Yanagawa Y, Allis CD, Hof PR, Scholz J (2018) NMDA Receptor Activation Underlies the Loss of Spinal Dorsal Horn Neurons and the Transition to Persistent Pain after Peripheral Nerve Injury. Cell Rep 23:2678–2689.

Irani SR, Bera K, Waters P, Zuliani L, Maxwell S, Zandi MS, Friese MA, Galea I, Kullmann DM, Beeson D, Lang B, Bien CG, Vincent A (2010) N-methyl-d-aspartate antibody encephalitis: temporal progression of clinical and paraclinical observations in a predominantly non- paraneoplastic disorder of both sexes. Brain 133:1655–1667.

Isaac JTR, Nicoll RA, Malenka RC (1995) Evidence for silent synapses: Implications for the expression of LTP. Neuron 15:427–434.

Islam R, Trépanier M-O, Milenkovic M, Horsfall W, Salahpour A, Bazinet RP, Ramsey AJ (2017) Vulnerability to omega-3 deprivation in a mouse model of NMDA receptor hypofunction. Npj Schizophr 3:12.

Isokawa M, Levesque MF (1991) Increased NMDA responses and dendritic degeneration in human epileptic hippocampal neurons in slices. Neurosci Lett 132:212–216.

Ito Y et al. (2018) De Novo Truncating Mutations in WASF1 Cause Intellectual Disability with Seizures. Am J Hum Genet 103:144–153.

Jackson ME, Homayoun H, Moghaddam B (2004) NMDA receptor hypofunction produces concomitant firing rate potentiation and burst activity reduction in the prefrontal cortex. Proc Natl Acad Sci 101:8467–8472.

Jander S, Schroeter M, Stoll G (2000) Role of NMDA receptor signaling in the regulation of inflammatory gene expression after focal brain ischemia. J Neuroimmunol 109:181–187.

Järbrink K, Knapp M (2001) The Economic Impact of Autism in Britain. Autism 5:7–22.

154

Jessell TM, Kandel ER (1993) Synaptic transmission: A bidirectional and self-modifiable form of cell-cell communication. Cell 72:1–30.

Jiao S, Li Z (2011) Nonapoptotic Function of BAD and BAX in Long-Term Depression of Synaptic Transmission. Neuron 70:758–772.

Jocoy EL, Andre VM, Cummings DM, Rao SP, Wu N, Ramsey AJ, Caron MG, Cepeda C, Levine MS (2011) Dissecting the contribution of individual receptor subunits to the enhancement of. Front Syst Neurosci 5:28.

Johnson JW, Ascher P (1990) Voltage-dependent block by intracellular Mg2+ of N-methyl-D- aspartate-activated channels. Biophys J 57:1085–1090.

Johnson MB, Stevens B (2018) Pruning hypothesis comes of age. Nature 554:438–439.

Johnston MV, Jeon O-H, Pevsner J, Blue ME, Naidu S (2001) Neurobiology of Rett syndrome: a genetic disorder of synapse development. Brain Dev 23:S206–S213.

Jones MW, Wilson MA (2005) Theta Rhythms Coordinate Hippocampal–Prefrontal Interactions in a Spatial Memory Task Morris R, ed. PLoS Biol 3:e402.

Kaizuka T, Takumi T (2018) Postsynaptic density proteins and their involvement in neurodevelopmental disorders. J Biochem (Tokyo) 163:447–455.

Karakas E, Furukawa H (2014) Crystal structure of a heterotetrameric NMDA receptor ion channel. Science 344:992–997.

Karolchik D, Hinrichs AS, Furey TS, Roskin KM, Sugnet CW, Haussler D, Kent WJ (2004) The UCSC Table Browser data retrieval tool. Nucleic Acids Res 32:D493–D496.

Kawai N, Sakai N, Okuro M, Karakawa S, Tsuneyoshi Y, Kawasaki N, Takeda T, Bannai M, Nishino S (2015) The Sleep-Promoting and Hypothermic Effects of Glycine are Mediated by NMDA Receptors in the Suprachiasmatic Nucleus. Neuropsychopharmacology 40:1405–1416.

Kemp A, Manahan-Vaughan D (2004) Hippocampal long-term depression and long-term potentiation encode different aspects of novelty acquisition. Proc Natl Acad Sci 101:8192–8197.

155

Kerchner GA, Nicoll RA (2008) Silent synapses and the emergence of a postsynaptic mechanism for LTP. Nat Rev Neurosci 9:813–825.

Kessler RC, Birnbaum H, Demler O, Falloon IRH, Gagnon E, Guyer M, Howes MJ, Kendler KS, Shi L, Walters E, Wu EQ (2005a) The Prevalence and Correlates of Nonaffective Psychosis in the National Comorbidity Survey Replication (NCS-R). Biol Psychiatry 58:668–676.

Kessler RC, Chiu WT, Demler O, Walters EE (2005b) Prevalence, severity, and comorbidity of 12-month DSM-IV disorders in the National Comorbidity Survey Replication. Arch Gen Psychiatry 62:617–627.

Kessler RC, Heeringa S, Lakoma MD, Petukhova M, Rupp AE, Schoenbaum M, Wang PS, Zaslavsky AM (2008) Individual and societal effects of mental disorders on earnings in the United States: results from the national comorbidity survey replication. Am J Psychiatry 165:703–711.

Khalil A, Morgan RN, Adams BR, Golding SE, Dever SM, Rosenberg E, Povirk LF, Valerie K (2011) ATM-dependent ERK signaling via AKT in response to DNA double-strand breaks. Cell Cycle 10:481–491.

Khan ZA, Ghosh AR (2010) Possible nitric oxide modulation in protective effects of Withaferin A against stress induced neurobehavioural changes. J Med Plants Res 4:490–495.

Kholmanskikh SS, Koeller HB, Wynshaw-Boris A, Gomez T, Letourneau PC, Ross ME (2006) Calcium-dependent interaction of Lis1 with IQGAP1 and Cdc42 promotes neuronal motility. Nat Neurosci 9:50–57.

Kim H-J, Cho M-H, Shim WH, Kim JK, Jeon E-Y, Kim D-H, Yoon S-Y (2017) Deficient autophagy in microglia impairs synaptic pruning and causes social behavioral defects. Mol Psychiatry 22:1576–1584.

Kim H-J, DiBernardo AB, Sloane JA, Rasband MN, Solomon D, Kosaras B, Kwak SP, Vartanian TK (2006a) WAVE1 Is Required for Oligodendrocyte Morphogenesis and Normal CNS Myelination. J Neurosci 26:5849.

156

Kim IH, Racz B, Wang H, Burianek L, Weinberg R, Yasuda R, Wetsel WC, Soderling SH (2013) Disruption of Arp2/3 Results in Asymmetric Structural Plasticity of Dendritic Spines and Progressive Synaptic and Behavioral Abnormalities. J Neurosci 33:6081–6092.

Kim MJ, Dunah AW, Wang YT, Sheng M (2005) Differential Roles of NR2A- and NR2B- Containing NMDA Receptors in Ras-ERK Signaling and AMPA Receptor Trafficking. Neuron 46:745–760.

Kim Y, Sung JY, Ceglia I, Lee K-W, Ahn J-H, Halford JM, Kim AM, Kwak SP, Park JB, Ho Ryu S, Schenck A, Bardoni B, Scott JD, Nairn AC, Greengard P (2006b) Phosphorylation of WAVE1 regulates actin polymerization and dendritic spine morphology. Nature 442:814–817.

Kinsella RJ, Kahari A, Haider S, Zamora J, Proctor G, Spudich G, Almeida-King J, Staines D, Derwent P, Kerhornou A, Kersey P, Flicek P (2011) Ensembl BioMarts: a hub for data retrieval across taxonomic space. Database J Biol Databases Curation 2011:bar030.

Kishi T, Matsuda Y, Iwata N (2017) Memantine add-on to antipsychotic treatment for residual negative and cognitive symptoms of schizophrenia: a meta-analysis. Psychopharmacology (Berl) 234:2113–2125.

Kleckner N, Dingledine R (1988) Requirement for glycine in activation of NMDA-receptors expressed in Xenopus oocytes. Science 241:835–837.

Knapp M, Romeo R, Beecham J (2009) Economic cost of autism in the UK. Autism 13:317–336.

Ko HW, Park K-Y, Kim H, Han P-L, Kim YU, Gwag BJ, Choi E-J (2002) Ca2+-Mediated Activation of c-Jun N-Terminal Kinase and Nuclear Factor κB by NMDA in Cortical Cell Cultures. J Neurochem 71:1390–1395.

Kocerha J, Faghihi MA, Lopez-Toledano MA, Huang J, Ramsey AJ, Caron MG, Sales N, Willoughby D, Elmen J, Hansen HF, Orum H, Kauppinen S, Kenny PJ, Wahlestedt C (2009) MicroRNA-219 modulates NMDA receptor-mediated neurobehavioral dysfunction. Proc Natl Acad Sci 106:3507–3512.

Köhr G (2006) NMDA receptor function: subunit composition versus spatial distribution. Cell Tissue Res 326:439–446.

157

Komuro H, Rakic P (1993) Modulation of neuronal migration by NMDA receptors. Science 260:95.

Komuro Y, Galas L, Lebon A, Raoult E, Fahrion JK, Tilot A, Kumada T, Ohno N, Vaudry D, Komuro H (2015) The role of calcium and cyclic nucleotide signaling in cerebellar granule cell migration under normal and pathological conditions: Second Messenger in Neuronal Migration. Dev Neurobiol 75:369–387.

Korhonen L, Näpänkangas U, Steen H, Chen Y, Martinez R, Lindholm D (2004) Differential regulation of X-chromosome-linked inhibitor of apoptosis protein (XIAP) and caspase-3 by NMDA in developing hippocampal neurons; involvement of the mitochondrial pathway in NMDA-mediated neuronal survival. Exp Cell Res 295:290–299.

Korinek M, Vyklicky V, Borovska J, Lichnerova K, Kaniakova M, Krausova B, Krusek J, Balik A, Smejkalova T, Horak M, Vyklicky L (2015) Cholesterol modulates open probability and desensitization of NMDA receptors. J Physiol 593:2279–2293.

Krämer A, Green J, Pollard J, Tugendreich S (2014) Causal analysis approaches in Ingenuity Pathway Analysis. Bioinformatics 30:523–530.

Krapivinsky G, Krapivinsky L, Manasian Y, Ivanov A, Tyzio R, Pellegrino C, Ben-Ari Y, Clapham DE, Medina I (2003) The NMDA Receptor Is Coupled to the ERK Pathway by a Direct Interaction between NR2B and RasGRF1. Neuron 40:775–784.

Kruse MS, Prémont J, Krebs M-O, Jay TM (2009) Interaction of dopamine D1 with NMDA NR1 receptors in rat prefrontal cortex. Eur Neuropsychopharmacol 19:296–304.

Krystal J, Karper L, Seibyl J, Freeman G, Delaney R, Bremner J, Heninger G, Bowers M, Charney D (1994) Subanesthetic Effects of the Noncompetitive Nmda Antagonist, Ketamine, in Humans - Psychotomimetic, Perceptual, Cognitive, and Neuroendocrine Responses. Arch Gen Psychiatry 51:199–214.

Kuan C-Y, Roth KA, Flavell RA, Rakic P (2000) Mechanisms of programmed cell death in the developing brain. Trends Neurosci 23:291–297.

158

Kuleshov MV, Jones MR, Rouillard AD, Fernandez NF, Duan Q, Wang Z, Koplev S, Jenkins SL, Jagodnik KM, Lachmann A, McDermott MG, Monteiro CD, Gundersen GW, Ma’ayan A (2016) Enrichr: a comprehensive gene set enrichment analysis web server 2016 update. Nucleic Acids Res 44:W90–W97.

Küppers E, Beyer C (1999) Expression of estrogen receptor-α and β mRNA in the developing and adult mouse striatum. Neurosci Lett 276:95–98.

Lahti AC, Koffel B, LaPorte D, Tamminga CA (1995) Subanesthetic Doses of Ketamine Stimulate Psychosis in Schizophrenia. Neuropsychopharmacology 13:9.

Lahti AC, Weiler MA, Tamara Michaelidis BA, Parwani A, Tamminga CA (2001) Effects of ketamine in normal and schizophrenic volunteers. Neuropsychopharmacol Off Publ Am Coll Neuropsychopharmacol 25:455–467.

Lai CSW, Franke TF, Gan W-B (2012) Opposite effects of fear conditioning and extinction on dendritic spine remodelling. Nature 483:87–91.

Lai Q, Hu P, Li Q, Li X, Yuan R, Tang X, Wang W, Li X, Fan H, Yin X (2016) NMDA receptors promote neurogenesis in the neonatal rat subventricular zone following hypoxic- ischemic injury. Mol Med Rep 13:206–212.

Laurie D, Seeburg P (1994) Regional and developmental heterogeneity in splicing of the rat brain NMDAR1 mRNA. J Neurosci 14:3180–3194.

LeClerc S, Easley D (2015) Pharmacological Therapies for Autism Spectrum Disorder: A Review. Pharm Ther 40:389–397.

Lee BR, Dong Y (2011) Cocaine-induced metaplasticity in the nucleus accumbens: Silent synapse and beyond. Neuropharmacology 61:1060–1069.

Lee C-H, Lü W, Michel JC, Goehring A, Du J, Song X, Gouaux E (2014) NMDA receptor structures reveal subunit arrangement and pore architecture. Nature 511:191–197.

Lee M-Y, Lin Y-R, Tu Y-S, Tseng YJ, Chan M-H, Chen H-H (2017) Effects of sarcosine and N, N-dimethylglycine on NMDA receptor-mediated excitatory field potentials. J Biomed Sci 24:18.

159

Leucht S, Arbter D, Engel RR, Kissling W, Davis JM (2009) How effective are second- generation antipsychotic drugs? A meta-analysis of placebo-controlled trials. Mol Psychiatry 14:429–447.

Levin ED, Petro A, Rezvani AH, Pollard N, Christopher NC, Strauss M, Avery J, Nicholson J, Rose JE (2009) Nicotinic α7- or β2-containing receptor knockout: Effects on radial-arm maze learning and long-term nicotine consumption in mice. Behav Brain Res 196:207–213.

Levy DI, Sucher NJ, Lipton SA (1991) Glutathione prevents N-methyl-D-aspartate receptor- mediated neurotoxicity. Neuroreport 2:345–347.

Lewerenz J, Maher P (2015) Chronic Glutamate Toxicity in Neurodegenerative Diseases—What is the Evidence? Front Neurosci 9:469.

Lewine R, Martin M, Hart M (2017) Sex versus gender differences in schizophrenia: The case for normal personality differences. Schizophr Res 189:57–60.

Li D-Q, Li Y, Liu Y, Bao Y-M, Hu B, An L-J (2005) Catalpol prevents the loss of CA1 hippocampal neurons and reduces working errors in gerbils after ischemia-reperfusion injury. Toxicon 46:845–851.

Li L, Sultan S, Heigele S, Schmidt-Salzmann C, Toni N, Bischofberger J (2017) Silent synapses generate sparse and orthogonal action potential firing in adult-born hippocampal granule cells. eLife 6.

Li N, Lee B, Liu R-J, Banasr M, Dwyer JM, Iwata M, Li X-Y, Aghajanian G, Duman RS (2010a) mTOR-Dependent Synapse Formation Underlies the Rapid Antidepressant Effects of NMDA Antagonists. Science 329:959–964.

Li S, Jin M, Koeglsperger T, Shepardson NE, Shankar GM, Selkoe DJ (2011) Soluble A Oligomers Inhibit Long-Term Potentiation through a Mechanism Involving Excessive Activation of Extrasynaptic NR2B-Containing NMDA Receptors. J Neurosci 31:6627–6638.

Li Y, Erzurumlu RS, Chen C, Jhaveri S, Tonegawa S (1994) Whisker-related neuronal patterns fail to develop in the trigeminal brainstem nuclei of NMDAR1 knockout mice. Cell 76:427–437.

160

Li Y, Krupa B, Kang J-S, Bolshakov VY, Liu G (2009) Glycine Site of NMDA Receptor Serves as a Spatiotemporal Detector of Synaptic Activity Patterns. J Neurophysiol 102:578–589.

Li Y, Sacchi S, Pollegioni L, Basu AC, Coyle JT, Bolshakov VY (2013) Identity of endogenous NMDAR glycine site agonist in amygdala is determined by synaptic activity level. Nat Commun 4:1760.

Li Y, Yang D-Q (2010) The ATM Inhibitor KU-55933 Suppresses Cell Proliferation and Induces Apoptosis by Blocking Akt In Cancer Cells with Overactivated Akt. Mol Cancer Ther 9:113– 125.

Li Z, Jo J, Jia J-M, Lo S-C, Whitcomb DJ, Jiao S, Cho K, Sheng M (2010b) Caspase-3 Activation via Mitochondria Is Required for Long-Term Depression and AMPA Receptor Internalization. Cell 141:859–871.

Li Z, Li J, Bi P, Lu Y, Burcham G, Elzey BD, Ratliff T, Konieczny SF, Ahmad N, Kuang S, Liu X (2014) Plk1 Phosphorylation of PTEN Causes a Tumor-Promoting Metabolic State. Mol Cell Biol 34:3642–3661.

Liang Y, de Wispelaere M, Carocci M, Liu Q, Wang J, Yang PL, Gray NS (2017) Structure– Activity Relationship Study of QL47: A Broad-Spectrum Antiviral Agent. ACS Med Chem Lett 8:344–349.

Lieberman OJ, McGuirt AF, Tang G, Sulzer D (2018) Roles for neuronal and glial autophagy in synaptic pruning during development. Neurobiol Dis Available at: http://www.sciencedirect.com/science/article/pii/S0969996118301347.

Lipton SA (2006) Paradigm shift in neuroprotection by NMDA receptor blockade: Memantine and beyond. Nat Rev Drug Discov 5:160–170.

Lisman J, Raghavachari S (2006) A Unified Model of the Presynaptic and Postsynaptic Changes During LTP at CA1 Synapses. Sci STKE 2006:re11–re11.

Litman RE, Smith MA, Doherty JJ, Cross A, Raines S, Gertsik L, Zukin SR (2016) AZD8529, a positive allosteric modulator at the mGluR2 receptor, does not improve symptoms in schizophrenia: A proof of principle study. Schizophr Res 172:152–157.

161

Lockwood PL, Wittmann MK (2018) Ventral anterior cingulate cortex and social decision- making. Neurosci Biobehav Rev 92:187–191.

Lorusso P, Krishnamurthi S, Rinehart JR, Nabell L, Croghan G, Varterasian M, Sadis SS, Menon SS, Leopold J, Meyer MB (2005) Phase 1-2 clinical study of a second generation oral MEK inhibitor, PD 0325901 in patients with advanced cancer. J Clin Oncol 23:194S-194S.

Low C-M, Zheng F, Lyuboslavsky P, Traynelis SF (2000) Molecular determinants of coordinated proton and zinc inhibition of N-methyl-D-aspartate NR1/NR2A receptors. Proc Natl Acad Sci 97:11062–11067.

Lu W, Bushong EA, Shih TP, Ellisman MH, Nicoll RA (2013) The Cell-Autonomous Role of Excitatory Synaptic Transmission in the Regulation of Neuronal Structure and Function. Neuron 78:433–439.

Luby E, Cohen B, Rosenbaum G, Gottlieb J, Kelley R (1959) Study of a New Schizophrenomimetic Drug - Sernyl. Arch Neurol Psychiatry 81:363–369.

Lundgaard I, Li B, Xie L, Kang H, Sanggaard S, Haswell JDR, Sun W, Goldman S, Blekot S, Nielsen M, Takano T, Deane R, Nedergaard M (2015) Direct neuronal glucose uptake heralds activity-dependent increases in cerebral metabolism. Nat Commun 6:6807.

Luo LQ (2000) Rho GTPases in neuronal morphogenesis. Nat Rev Neurosci 1:173–180.

Luo LQ, Hensch TK, Ackerman L, Barbel S, Jan LY, Jan YN (1996) Differential effects of the Rac GTPase on Purkinje cell axons and dendritic trunks and spines. Nature 379:837–840.

Luo T, Wu W-H, Chen B-S (2011) NMDA receptor signaling: death or survival? Front Biol 6:468–476.

Luscher C, Malenka RC (2012) NMDA Receptor-Dependent Long-Term Potentiation and Long- Term Depression (LTP/LTD). Cold Spring Harb Perspect Biol 4:a005710–a005710.

Lüthi A, Schwyzer L, Mateos JM, Gähwiler BH, McKinney RA (2001) NMDA receptor activation limits the number of synaptic connections during hippocampal development. Nat Neurosci 4:1102–1107.

162

Lynch DR, Dawson TM (1994) Secondary mechanisms in neuronal trauma. Curr Opin Neurol 7:510–516.

Lynch JW, Zhang Y, Talwar S, Estrada-Mondragon A (2017) Glycine Receptor Drug Discovery. In: Advances in Pharmacology (Geraghty DP, Rash LD, eds), pp 225–253. Academic Press. Available at: http://www.sciencedirect.com/science/article/pii/S1054358917300030.

Mackay JP, Nassrallah WB, Raymond LA (2018) Cause or compensation?-Altered neuronal Ca 2+ handling in Huntington’s disease. CNS Neurosci Ther 24:301–310.

Madry C, Betz H, Geiger J, Laube B (2010) Potentiation of glycine-gated NR1/NR3A NMDA receptors relieves Ca2+-dependent outward rectification. Front Mol Neurosci 3:6.

Madry C, Betz H, Geiger JRP, Laube B (2008) Supralinear potentiation of NR1/NR3A excitatory glycine receptors by Zn2+ and NR1 antagonist. Proc Natl Acad Sci 105:12563.

Mahieu S, Klug M, Millen N, Fabro A, Benmelej A, Contini M del C (2016) Monosodium glutamate intake affect the function of the kidney through NMDA receptor. Life Sci 149:114– 119.

Mak WWS, Cheung RYM (2012) Psychological Distress and Subjective Burden of Caregivers of People with Mental Illness: The Role of Affiliate Stigma and Face Concern. Community Ment Health J 48:270–274.

Malhotra AK, Pinals DA, Adler CM, Elman I, Clifton A, Pickar D, Breier A (1997) Ketamine- induced exacerbation of psychotic symptoms and cognitive impairment in neuroleptic-free schizophrenics. Neuropsychopharmacol Off Publ Am Coll Neuropsychopharmacol 17:141–150.

Malhotra AK, Pinals DA, Weingartner H, Sirocco K, Missar CD, Pickar D, Breier A (1996) NMDA receptor function and human cognition: The effects of ketamine in healthy volunteers. Neuropsychopharmacology 14:301–307.

Malinow R, Mainen ZF, Hayashi Y (2000) LTP mechanisms: from silence to four-lane traffic. Curr Opin Neurobiol 10:352–357.

163

Malinow R, Malenka RC (2002) AMPA Receptor Trafficking and Synaptic Plasticity. Annu Rev Neurosci 25:103–126.

Manahan-Vaughan D, von Haebler D, Winter C, Juckel G, Heinemann U (2008) A single application of MK801 causes symptoms of acute psychosis, deficits in spatial memory, and impairment of synaptic plasticity in rats. Hippocampus 18:125–134.

Mandillo S et al. (2008) Reliability, robustness, and reproducibility in mouse behavioral phenotyping: a cross-laboratory study. Physiol Genomics 34:243–255.

Manfridi A, Brambilla D, Mancia M (1999) Stimulation of NMDA and AMPA receptors in the rat nucleus basalis of Meynert affects sleep. Am J Physiol-Regul Integr Comp Physiol 277:R1488–R1492.

Mao J (1999) NMDA and opioid receptors: their interactions in antinociception, tolerance and neuroplasticity. Brain Res Rev 30:289–304.

Markus EJ, Barnes CA, McNaughton BL, Gladden VL, Skaggs WE (1994) Spatial information content and reliability of hippocampal CA1 neurons: Effects of visual input. Hippocampus 4:410–421.

Marquard J et al. (2015) Characterization of pancreatic NMDA receptors as possible drug targets for diabetes treatment. Nat Med 21:363–372.

Martins RAP, Silveira MS, Curado MR, Police AI, Linden R (2005) NMDA receptor activation modulates programmed cell death during early post-natal retinal development: a BDNF- dependent mechanism: NMDA and programmed cell death in developing retina. J Neurochem 95:244–253.

Masi A, DeMayo MM, Glozier N, Guastella AJ (2017) An Overview of Autism Spectrum Disorder, Heterogeneity and Treatment Options. Neurosci Bull 33:183–193.

Matsuda K, Kamiya Y, Matsuda S, Yuzaki M (2002) Cloning and characterization of a novel NMDA receptor subunit NR3B: a dominant subunit that reduces calcium permeability. Mol Brain Res 100:43–52.

164

Mattson M, Cheng B, Davis D, Bryant K, Lieberburg I, Rydel R (1992) beta-Amyloid peptides destabilize calcium homeostasis and render human cortical neurons vulnerable to excitotoxicity. J Neurosci 12:376.

Maurice T, Hiramatsu M, Itoh J, Kameyama T, Hasegawa T, Nabeshima T (1994) Behavioral evidence for a modulating role of sigma-ligands in memory processes . I. Attenuation of dizocilpine (MK-801)-induced amnesia. Brain Res 647:44–56.

Maycox PR, Kelly F, Taylor A, Bates S, Reid J, Logendra R, Barnes MR, Larminie C, Jones N, Lennon M, others (2009) Analysis of gene expression in two large schizophrenia cohorts identifies multiple changes associated with nerve terminal function. Mol Psychiatry 14:1083– 1094.

Mayer ML, Vyklicky L, Clements J (1989) Regulation of NMDA receptor desensitization in mouse hippocampal neurons by glycine. Nature 338:425–427.

Mayer ML, Westbrook GL, Guthrie PB (1984) Voltage-dependent block by Mg2+ of NMDA responses in spinal cord neurones. Nature 309:261–263.

McAllister BB, Dyck RH (2017) Zinc transporter 3 (ZnT3) and vesicular zinc in central nervous system function. Neurosci Biobehav Rev 80:329–350.

McCarthy MM, Pickett LA, VanRyzin JW, Kight KE (2015) Surprising origins of sex differences in the brain. Horm Behav 76:3–10.

McDonald R, White N (1993) A Triple Dissociation of Memory-Systems - Hippocampus, Amygdala, and Dorsal Striatum. Behav Neurosci 107:3–22.

McGrath J, Saha S, Chant D, Welham J (2008) Schizophrenia: A Concise Overview of Incidence, Prevalence, and Mortality. Epidemiol Rev 30:67–76.

McHugh TJ, Jones MW, Quinn JJ, Balthasar N, Coppari R, Elmquist JK, Lowell BB, Fanselow MS, Wilson MA, Tonegawa S (2007) Dentate gyrus NMDA receptors mediate rapid pattern separation in the hippocampal network. SCIENCE 317:94–99.

165

McIlwain DR, Berger T, Mak TW (2013) Caspase Functions in Cell Death and Disease. Cold Spring Harb Perspect Biol 5:a008656–a008656.

McKenzie K, Milton M, Smith G, Ouellette-Kuntz H (2016) Systematic Review of the Prevalence and Incidence of Intellectual Disabilities: Current Trends and Issues. Curr Dev Disord Rep 3:104–115.

McLennan H, Hicks TP, Hall JG (1981) Receptors for the excitatory amino acids. Adv Biochem Psychopharmacol 29:213–221.

McRoberts JA, Li J, Ennes HS, Mayer EA (2007) Sex-dependent differences in the activity and modulation of N-methyl-d-aspartic acid receptors in rat dorsal root ganglia neurons. Neuroscience 148:1015–1020.

Mehrabi NF, Waldvogel HJ, Tippett LJ, Hogg VM, Synek BJ, Faull RLM (2016) Symptom heterogeneity in Huntington’s disease correlates with neuronal degeneration in the cerebral cortex. Neurobiol Dis 96:67–74.

Meijering E, Jacob M, Sarria J-C, Steiner P, Hirling H, Unser M (2004) Design and validation of a tool for neurite tracing and analysis in fluorescence microscopy images. Cytometry A 58:167– 176.

Mellon SH, Griffin LD (2002) Neurosteroids: biochemistry and clinical significance. Trends Endocrinol Metab 13:35–43.

Melville CA, Cooper S-A, Morrison J, Smiley E, Allan L, Jackson A, Finlayson J, Mantry D (2008) The Prevalence and Incidence of Mental Ill-Health in Adults with Autism and Intellectual Disabilities. J Autism Dev Disord 38:1676–1688.

Mielnik CA, Horsfall W, Ramsey AJ (2014) Diazepam improves aspects of social behaviour and neuron activation in NMDA receptor-deficient mice. Genes Brain Behav 13:592–602.

Mikics E, Toth M, Biro L, Bruzsik B, Nagy B, Haller J (2017) The role of GluN2B-containing NMDA receptors in short- and long-term fear recall. Physiol Behav 177:44–48.

166

Milenkovic M, Mielnik CA, Ramsey AJ (2014) NMDA receptor-deficient mice display sexual dimorphism in the onset and severity of behavioural abnormalities. Genes Brain Behav 13:850– 862.

Millán AP, Torres JJ, Johnson S, Marro J (2018) Concurrence of form and function in developing networks and its role in synaptic pruning. Nat Commun 9:2236.

Minshawi NF, Wink LK, Shaffer R, Plawecki MH, Posey DJ, Liu H, Hurwitz S, McDougle CJ, Swiezy NB, Erickson CA (2016) A randomized, placebo-controlled trial of d-cycloserine for the enhancement of social skills training in autism spectrum disorders. Mol Autism 7:2.

Mitani H, Shirayama Y, Yamada T, Maeda K, Ashby CR, Kawahara R (2006) Correlation between plasma levels of glutamate, alanine and serine with severity of depression. Prog Neuropsychopharmacol Biol Psychiatry 30:1155–1158.

Mitchell AJ, Kemp S, Benito-LeÃ3n J, Reuber M (2010) The influence of cognitive impairment on health-related quality of life in neurological disease. Acta Neuropsychiatr 22:2–13.

Miyamoto S, Duncan GE, Marx CE, Lieberman JA (2005) Treatments for schizophrenia: a critical review of pharmacology and mechanisms of action of antipsychotic drugs. Mol Psychiatry 10:79–104.

Miyamoto S, Snouwaert JN, Koller BH, Moy SS, Lieberman JA, Duncan GE (2004) Amphetamine-Induced Fos is Reduced in Limbic Cortical Regions but not in the Caudate or Accumbens in a Genetic Model of NMDA Receptor Hypofunction. Neuropsychopharmacology 29:2180–2188.

Moghaddam B, Javitt D (2012) From revolution to evolution: the glutamate hypothesis of schizophrenia and its implication for treatment. Neuropsychopharmacology 37:4–15.

Mohan R, Hammers H, Bargagna-mohan P, Zhan X, Herbstritt C, Ruiz A, Zhang L, Hanson A, Conner B, Rougas J, Pribluda V (2004) Withaferin A is a potent inhibitor of angiogenesis. Angiogenesis 7:115–122.

Mohn AR, Gainetdinov RR, Caron MG, Koller BH (1999) Mice with Reduced NMDA Receptor Expression Display Behaviors Related to Schizophrenia. Cell 98:427–436.

167

Monaghan DT, Beaton JA (1991) Quinolinate differentiates between forebrain and cerebellar NMDA receptors. Eur J Pharmacol 194:123–125.

Mony L, Zhu S, Carvalho S, Paoletti P (2011) Molecular basis of positive allosteric modulation of GluN2B NMDA receptors by polyamines. EMBO J 30:3134–3146.

Monyer H, Burnashev N, Laurie DJ, Sakmann B, Seeburg PH (1994) Developmental and regional expression in the rat brain and functional properties of four NMDA receptors. Neuron 12:529–540.

Morland C, Nordengen K, Gundersen V (2012) Valproate causes reduction of the excitatory amino acid aspartate in nerve terminals. Neurosci Lett 527:100–104.

Morland C, Nordengen K, Larsson M, Prolo LM, Farzampour Z, Reimer RJ, Gundersen V (2013) Vesicular uptake and exocytosis of l -aspartate is independent of sialin. FASEB J 27:1264–1274.

Morris GF, Bullock R, Marshall SB, Marmarou A, Maas A, Marshall LF (1999) Failure of the competitive N-methyl-D-aspartate antagonist Selfotel (CGS 19755) in the treatment of severe head injury: results of two phase III clinical trials. J Neurosurg 91:737–743.

Morris R, Anderson E, Lynch G, Baudry M (1986) Selective impairment of learning and blockade of long-term potentiation by an n-methyl-d-aspartate receptor antagonist, ap5. Nature 319:774–776.

Morris RGM, Steele RJ, Bell JE, Martin SJ (2013) N-methyl-D-aspartate receptors, learning and memory: chronic intraventricular infusion of the NMDA receptor antagonist D-AP5 interacts directly with the neural mechanisms of spatial learning. Eur J Neurosci 37:700–717.

Moscato EH, Peng X, Jain A, Parsons TD, Dalmau J, Balice-Gordon RJ (2014) Acute mechanisms underlying antibody effects in anti-N-methyl-D-aspartate receptor encephalitis: Anti-NMDAR Encephalitis. Ann Neurol 76:108–119.

Moser M-B, Trommald M, Andersen P (1994) An increase in dendritic spine density on hippocampal CA1 pyramidal cells following spatial learning in adult rats suggests the formation of new synapses. Proc Natl Acad Sci 91:12673–12675.

168

Moy SS, Nadler JJ, Perez A, Barbaro RP, Johns JM, Magnuson TR, Piven J, Crawley JN (2004) Sociability and preference for social novelty in five inbred strains: an approach to assess autistic- like behavior in mice. Genes Brain Behav 3:287–302.

Moy SS, Nadler JJ, Poe MD, Nonneman RJ, Young NB, Koller BH, Crawley JN, Duncan GE, Bodfish JW (2008) Development of a mouse test for repetitive, restricted behaviors: Relevance to autism. Behav Brain Res 188:178–194.

Moy SS, Nikolova VD, Riddick NV, Baker LK, Koller BH (2012) Preweaning Sensorimotor Deficits and Adolescent Hypersociability in Grin1 Knockdown Mice. Dev Neurosci 34:159– 173.

Moy SS, Perez A, Koller BH, Duncan GE (2006) Amphetamine-induced disruption of prepulse inhibition in mice with reduced NMDA receptor function. Brain Res 1089:186–194.

Muir K (2006) Glutamate-based therapeutic approaches: clinical trials with NMDA antagonists. Curr Opin Pharmacol 6:53–60.

Murakoshi H, Wang H, Yasuda R (2011) Local, persistent activation of Rho GTPases during plasticity of single dendritic spines. Nature 472:100–104.

Murugan M, Ling E-A, Kaur C (2013) Glutamate receptors in microglia. CNS Neurol Disord Drug Targets 12:773–784.

Naassila M, Pierrefiche O (2018) GluN2B Subunit of the NMDA Receptor: The Keystone of the Effects of Alcohol During Neurodevelopment. Neurochem Res Available at: https://doi.org/10.1007/s11064-017-2462-y.

Nabavi S, Fox R, Proulx CD, Lin JY, Tsien RY, Malinow R (2014) Engineering a memory with LTD and LTP. Nature 511:348–352.

Nacher J, McEwen BS (2006) The role ofN-methyl-D-asparate receptors in neurogenesis. Hippocampus 16:267–270.

169

Naisbitt S, Kim E, Tu JC, Xiao B, Sala C, Valtschanoff J, Weinberg RJ, Worley PF, Sheng M (1999) Shank, a Novel Family of Postsynaptic Density Proteins that Binds to the NMDA Receptor/PSD-95/GKAP Complex and Cortactin. Neuron 23:569–582.

Nakayama AY, Harms MB, Luo L (2000) Small GTPases Rac and Rho in the maintenance of dendritic spines and branches in hippocampal pyramidal neurons. J Neurosci 20:5329–5338.

Nakazawa K, Jeevakumar V, Nakao K (2017) Spatial and temporal boundaries of NMDA receptor hypofunction leading to schizophrenia. NPJ Schizophr 3:7.

Nandra KS, Agius M (2012) The differences between typical and atypical antipsychotics: the effects on neurogenesis. Psychiatr Danub 24 Suppl 1:S95-99.

Nelson TJ, Sun M-K, Alkon DL (2015) Signaling Pathways Involved in Cognitive Enhancement. In: Cognitive Enhancement (Knafo S, Venero C, eds), pp 11–42. San Diego: Academic Press. Available at: http://www.sciencedirect.com/science/article/pii/B9780124170421000024.

Neves G, Cooke SF, Bliss TVP (2008) Synaptic plasticity, memory and the hippocampus: a neural network approach to causality. Nat Rev Neurosci 9:65–75.

Newcomer JW, Farber NB, Olney JW (2000) NMDA receptor function, memory, and brain aging. Dialogues Clin Neurosci 2:14.

Newschaffer CJ (2005) National Autism Prevalence Trends From United States Special Education Data. PEDIATRICS 115:e277–e282.

Ngun TC, Ghahramani N, Sánchez FJ, Bocklandt S, Vilain E (2011) The genetics of sex differences in brain and behavior. Front Neuroendocrinol 32:227–246.

Nicolas LB, Pinoteau W, Papot S, Routier S, Guillaumet G, Mortaud S (2001) Aggressive behavior induced by the steroid sulfatase inhibitor COUMATE and by DHEAS in CBA/H mice. Brain Res 922:216–222.

Niethammer M, Kim E, Sheng M (1996) Interaction between the C Terminus of NMDA Receptor Subunits and Multiple Members of the PSD-95 Family of Membrane-Associated Guanylate Kinases. J Neurosci:7.

170

Noda M, Nakanishi H, Akaike N (1999) Glutamate release from microglia via glutamate transporter is enhanced by amyloid-beta peptide. Neuroscience 92:1465–1474.

Nowak L, Bregestovski P, Ascher P, Herbet A, Prochiantz A (1984) Magnesium gates glutamate-activated channels in mouse central neurones. Nature 307:462–465.

O’Dell TJ, Kandel ER (1994) Low-frequency stimulation erases LTP through an NMDA receptor-mediated activation of protein phosphatases. Learn Mem Cold Spring Harb N 1:129– 139.

Ogita K, Enomoto R, Nakahara F, Ishitsubo N, Yoneda Y (2002) A Possible Role of Glutathione as an Endogenous Agonist at the N-Methyl-d-Aspartate Recognition Domain in Rat Brain. J Neurochem 64:1088–1096.

Oh JH, Kwon TK (2009) Withaferin A inhibits tumor necrosis factor-α-induced expression of cell adhesion molecules by inactivation of Akt and NF-κB in human pulmonary epithelial cells. China Regul T Cells 2008 9:614–619.

Oh JH, Lee T-J, Park J-W, Kwon TK (2008) Withaferin A inhibits iNOS expression and nitric oxide production by Akt inactivation and down-regulating LPS-induced activity of NF-κB in RAW 264.7 cells. Eur J Pharmacol 599:11–17.

Ohno T, Maeda H, Murabe N, Kamiyama T, Yoshioka N, Mishina M, Sakurai M (2010) Specific involvement of postsynaptic GluN2B-containing NMDA receptors in the developmental elimination of corticospinal synapses. Proc Natl Acad Sci 107:15252–15257.

Okabe S, Miwa A, Okado H (1999) Alternative Splicing of the C-Terminal Domain Regulates Cell Surface Expression of the NMDA Receptor NR1 Subunit. J Neurosci 19:7781–7792.

Okamoto S, Pouladi MA, Talantova M, Yao D, Xia P, Ehrnhoefer DE, Zaidi R, Clemente A, Kaul M, Graham RK, Zhang D, Vincent Chen H-S, Tong G, Hayden MR, Lipton SA (2009) Balance between synaptic versus extrasynaptic NMDA receptor activity influences inclusions and neurotoxicity of mutant huntingtin. Nat Med 15:1407–1413.

O’Leary NA et al. (2016) Reference sequence (RefSeq) database at NCBI: current status, taxonomic expansion, and functional annotation. Nucleic Acids Res 44:D733–D745.

171

Olloquequi J, Cornejo-Córdova E, Verdaguer E, Soriano FX, Binvignat O, Auladell C, Camins A (2018) Excitotoxicity in the pathogenesis of neurological and psychiatric disorders: Therapeutic implications. J Psychopharmacol (Oxf) 32:265–275.

Olney JW, Price MT, Salles KS, Labruyere J, Ryerson R, Mahan K, Frierdich G, Samson L (1987) L-Homocysteic acid: An endogenous excitotoxic ligand of the NMDA receptor. Brain Res Bull 19:597–602.

Oppenheim RW (1991) Cell death during development of the nervous system. Annu Rev Neurosci 14:453–501.

Ortega F, Pérez-Sen R, Morente V, Delicado EG, Miras-Portugal MT (2010) P2X7, NMDA and BDNF receptors converge on GSK3 phosphorylation and cooperate to promote survival in cerebellar granule neurons. Cell Mol Life Sci 67:1723–1733.

Ortiz GG, Bitzer-Quintero OK, Zárate CB, Rodríguez-Reynoso S, Larios-Arceo F, Velázquez- Brizuela IE, Pacheco-Moisés F, Rosales-Corral SA (2006) Monosodium glutamate-induced damage in liver and kidney: a morphological and biochemical approach. Biomed Pharmacother 60:86–91.

Ouellette-Kuntz H, Coo H, Lam M, Breitenbach MM, Hennessey PE, Jackman PD, Lewis MES, Dewey D, Bernier FP, Chung AM (2014) The changing prevalence of autism in three regions of Canada. J Autism Dev Disord 44:120–136.

Oyrer J, Maljevic S, Scheffer IE, Berkovic SF, Petrou S, Reid CA (2018) Ion Channels in Genetic Epilepsy: From Genes and Mechanisms to Disease-Targeted Therapies Sexton PM, ed. Pharmacol Rev 70:142–173.

Pachernegg S, Strutz-Seebohm N, Hollmann M (2012) GluN3 subunit-containing NMDA receptors: not just one-trick ponies. Trends Neurosci 35:240–249.

Pál B (2018) Involvement of extrasynaptic glutamate in physiological and pathophysiological changes of neuronal excitability. Cell Mol Life Sci 75:2917–2949.

Paoletti P, Bellone C, Zhou Q (2013) NMDA receptor subunit diversity: impact on receptor properties, synaptic plasticity and disease. Nat Rev Neurosci 14:383–400.

172

Papadia S, Soriano FX, Léveillé F, Martel M-A, Dakin KA, Hansen HH, Kaindl A, Sifringer M, Fowler J, Stefovska V, Mckenzie G, Craigon M, Corriveau R, Ghazal P, Horsburgh K, Yankner BA, Wyllie DJA, Ikonomidou C, Hardingham GE (2008) Synaptic NMDA receptor activity boosts intrinsic antioxidant defenses. Nat Neurosci 11:476–487.

Papouin T, Ladépêche L, Ruel J, Sacchi S, Labasque M, Hanini M, Groc L, Pollegioni L, Mothet J-P, Oliet SHR (2012) Synaptic and Extrasynaptic NMDA Receptors Are Gated by Different Endogenous Coagonists. Cell 150:633–646.

Parsons CG, Stöffler A, Danysz W (2007) Memantine: a NMDA receptor antagonist that improves memory by restoration of homeostasis in the glutamatergic system - too little activation is bad, too much is even worse. Neuropharmacology 53:699–723.

Parsons MP, Raymond LA (2014) Extrasynaptic NMDA Receptor Involvement in Central Nervous System Disorders. Neuron 82:279–293.

Pathania M, Davenport EC, Muir J, Sheehan DF, López-Doménech G, Kittler JT (2014) The autism and schizophrenia associated gene CYFIP1 is critical for the maintenance of dendritic complexity and the stabilization of mature spines. Transl Psychiatry 4:e374–e374.

Patterson CM, Wong J-MT, Leinninger GM, Allison MB, Mabrouk OS, Kasper CL, Gonzalez IE, Mackenzie A, Jones JC, Kennedy RT, Myers MG (2015) Ventral Tegmental Area Neurotensin Signaling Links the Lateral Hypothalamus to Locomotor Activity and Striatal Dopamine Efflux in Male Mice. Endocrinology 156:1692–1700.

Paul SM, Doherty JJ, Robichaud AJ, Belfort GM, Chow BY, Hammond RS, Crawford DC, Linsenbardt AJ, Shu H-J, Izumi Y, Mennerick SJ, Zorumski CF (2013) The Major Brain Cholesterol Metabolite 24(S)-Hydroxycholesterol Is a Potent Allosteric Modulator of N-Methyl- D-Aspartate Receptors. J Neurosci 33:17290–17300.

Peet M, Stokes C (2005) Omega-3 Fatty Acids in the Treatment of Psychiatric Disorders: Drugs 65:1051–1059.

Pei L, Lee F, Moszczynska A, Vukusic B, Liu F (2004) Regulation of dopamine D1 receptor function by physical interaction with the NMDA receptors. J Neurosci 24:1149–1158.

173

Penzes P, Cahill ME, Jones KA, Srivastava DP (2008) Convergent CaMK and RacGEF signals control dendritic structure and function. Trends Cell Biol 18:405–413.

Pérez-Otaño I, Larsen RS, Wesseling JF (2016) Emerging roles of GluN3-containing NMDA receptors in the CNS. Nat Rev Neurosci 17:623–635.

Peterson SL, Schwade ND (1993) The anticonvulsant activity of d-cycloserine is specific for tonic convulsions. Epilepsy Res 15:141–148.

Petrovic M, Horak M, Sedlacek M, Vyklicky LJ (2005) Physiology and pathology of NMDA receptors. Prague Med Rep 106:113–136.

Pierson TM, Yuan H, Marsh ED, Fuentes-Fajardo K, Adams DR, Markello T, Golas G, Simeonov DR, Holloman C, Tankovic A, Karamchandani MM, Schreiber JM, Mullikin JC, Tifft CJ, Toro C, Boerkoel CF, Traynelis SF, Gahl WA (2014) GRIN2A mutation and early-onset epileptic encephalopathy: personalized therapy with memantine. Ann Clin Transl Neurol 1:190– 198.

Pina-Crespo JC, Talantova M, Micu I, States B, Chen H-SV, Tu S, Nakanishi N, Tong G, Zhang D, Heinemann SF, Zamponi GW, Stys PK, Lipton SA (2010) Excitatory Glycine Responses of CNS Myelin Mediated by NR1/NR3 “NMDA” Receptor Subunits. J Neurosci 30:11501–11505.

Piochon C, Kano M, Hansel C (2016) LTD-like molecular pathways in developmental synaptic pruning. Nat Neurosci 19:1299–1310.

Pollitt AY, Insall RH (2009) WASP and SCAR/WAVE proteins: the drivers of actin assembly. J Cell Sci 122:2575–2578.

Posey DJ, Kem DL, Swiezy NB, Sweeten TL, Wiegand RE, McDougle CJ (2004) A Pilot Study of d-Cycloserine in Subjects With Autistic Disorder. Am J Psychiatry 161:2115–2117.

Posey DJ, Stigler KA, Erickson CA, McDougle CJ (2008) Antipsychotics in the treatment of autism. J Clin Invest 118:6–14.

Prithviraj R, Inglis FM (2008) Expression of the N-methyl-d-aspartate receptor subunit NR3B regulates dendrite morphogenesis in spinal motor neurons. Neuroscience 155:145–153.

174

Qadir AA, Obringer E, Hageman J, Marcuccilli C (2017) Risk of Second Seizure in Pediatric Patients With Idiopathic Autism. J Child Neurol 32:876–879.

Qiu H, Li J, Clark LH, Jackson AL, Zhang L, Guo H, Kilgore JE, Gehrig PA, Zhou C, Bae-Jump VL (2016) JQ1 suppresses tumor growth via PTEN/PI3K/AKT pathway in endometrial cancer. Oncotarget 7:66809–66821.

Quincozes-Santos A, Bobermin LD, Tramontina AC, Wartchow KM, Tagliari B, Souza DO, Wyse ATS, Gonçalves C-A (2014) Oxidative stress mediated by NMDA, AMPA/KA channels in acute hippocampal slices: Neuroprotective effect of resveratrol. Toxicol In Vitro 28:544–551.

Rabacchi S, Bailly Y, Delhaye-Bouchaud N, Mariani J (1992) Involvement of the N-methyl D- aspartate (NMDA) receptor in synapse elimination during cerebellar development. Science 256:1823–1825.

Rachline J (2005) The Micromolar Zinc-Binding Domain on the NMDA Receptor Subunit NR2B. J Neurosci 25:308–317.

Ramsey AJ, Laakso A, Cyr M, Sotnikova TD, Salahpour A, Medvedev IO, Dykstra LA, Gainetdinov RR, Caron MG (2008) Genetic NMDA Receptor Deficiency Disrupts Acute and Chronic Effects of Cocaine but not Amphetamine. Neuropsychopharmacology 33:2701–2714.

Ramsey AJ, Milenkovic M, Oliveira AF, Escobedo-Lozoya Y, Seshadri S, Salahpour A, Sawa A, Yasuda R, Caron MG (2011) Impaired NMDA receptor transmission alters striatal synapses and DISC1 protein in an age-dependent manner. Proc Natl Acad Sci 108:5795–5800.

Rao VR, Finkbeiner S (2007) NMDA and AMPA receptors: old channels, new tricks. Trends Neurosci 30:284–291.

Regan MC, Grant T, McDaniel MJ, Karakas E, Zhang J, Traynelis SF, Grigorieff N, Furukawa H (2018) Structural Mechanism of Functional Modulation by Gene Splicing in NMDA Receptors. Neuron 98:521-529.e3.

Reisert I, Pilgrim C (1991) Sexual differentiation of monoaminergic neurons - genetic or epigenetic? Trends Neurosci 14:468–473.

175

Ren K, Dubner R (2007) Pain Facilitation and Activity-Dependent Plasticity in Pain Modulatory Circuitry: Role of BDNF-TrkB Signaling and NMDA Receptors. Mol Neurobiol 35:224–235.

Renier N, Adams EL, Kirst C, Wu Z, Azevedo R, Kohl J, Autry AE, Kadiri L, Umadevi Venkataraju K, Zhou Y, Wang VX, Tang CY, Olsen O, Dulac C, Osten P, Tessier-Lavigne M (2016) Mapping of Brain Activity by Automated Volume Analysis of Immediate Early Genes. Cell 165:1789–1802.

Retchless BS, Gao W, Johnson JW (2012) A single GluN2 subunit residue controls NMDA receptor channel properties via intersubunit interaction. Nat Neurosci 15:406–413.

Rice AC, DeLorenzo RJ (1998) NMDA receptor activation during status epilepticus is required for the development of epilepsy. Brain Res 782:240–247.

Ridge JP, Ho AM-C, Dodd PR (2009) Sex Differences in NMDA Receptor Expression in Human Alcoholics. Alcohol Alcohol 44:594–601.

Rilling JK, Chen X, Chen X, Haroon E (2018) Intranasal oxytocin modulates neural functional connectivity during human social interaction. Am J Primatol:e22740.

Rinaldi T, Kulangara K, Antoniello K, Markram H (2007) Elevated NMDA receptor levels and enhanced postsynaptic long-term potentiation induced by prenatal exposure to valproic acid. Proc Natl Acad Sci 104:13501–13506.

Rinn JL, Snyder M (2005) Sexual dimorphism in mammalian gene expression. Trends Genet 21:298–305.

Ripke S et al. (2014) Biological insights from 108 schizophrenia-associated genetic loci. Nature 511:421–427.

Roberts AC, Díez-García J, Rodriguiz RM, López IP, Luján R, Martínez-Turrillas R, Picó E, Henson MA, Bernardo DR, Jarrett TM, Clendeninn DJ, López-Mascaraque L, Feng G, Lo DC, Wesseling JF, Wetsel WC, Philpot BD, Pérez-Otaño I (2009) Downregulation of NR3A- Containing NMDARs Is Required for Synapse Maturation and Memory Consolidation. Neuron 63:342–356.

176

Rocca A, Farolfi A, Bravaccini S, Schirone A, Amadori D (2014) Palbociclib (PD 0332991): targeting the cell cycle machinery in breast cancer. Expert Opin Pharmacother 15:407–420.

Rodríguez-Durán LF, Escobar ML (2014) NMDA receptor activation and PKC but not PKA lead to the modification of the long-term potentiation in the insular cortex induced by conditioned taste aversion: Differential role of kinases in metaplasticity. Behav Brain Res 266:58–62.

Roelfsema PR, Holtmaat A (2018) Control of synaptic plasticity in deep cortical networks. Nat Rev Neurosci 19:166–180.

Rogawski MA (2011) Revisiting AMPA Receptors as an Antiepileptic Drug Target. Epilepsy Curr 11:56–63.

Roy DS, Arons A, Mitchell TI, Pignatelli M, Ryan TJ, Tonegawa S (2016) Memory retrieval by activating engram cells in mouse models of early Alzheimer’s disease. Nature 531:508–512.

Rozental R, Giaume C, Spray DC (2000) Gap junctions in the nervous system. Brain Res Rev 32:11–15.

Ruddy RM, Chen Y, Milenkovic M, Ramsey AJ (2015) Differential effects of NMDA receptor antagonism on spine density: NMDA Receptor Antagonism on Spine Density. Synapse 69:52– 56.

Rymar VV, Sasseville R, Luk KC, Sadikot AF (2004) Neurogenesis and stereological morphometry of calretinin-immunoreactive GABAergic interneurons of the neostriatum. J Comp Neurol 469:325–339.

Sacco RL, DeRosa JT, Haley ECJ, Levin B, Ordronneau P, Phillips SJ, Rundek T, Snipes RG, Thompson JL (2001) Glycine antagonist in neuroprotection for patients with acute stroke: GAIN Americas: a randomized controlled trial. JAMA 285:1719–1728.

Said SI, Berisha HI, Pakbaz H (1996) Excitotoxicity in the lung: N-methyl-D-aspartate-induced, nitric oxide-dependent, pulmonary edema is attenuated by vasoactive intestinal peptide and by inhibitors of poly(ADP-ribose) polymerase. Proc Natl Acad Sci 93:4688–4692.

177

Sakarya O, Armstrong KA, Adamska M, Adamski M, Wang I-F, Tidor B, Degnan BM, Oakley TH, Kosik KS (2007) A Post-Synaptic Scaffold at the Origin of the Animal Kingdom Vosshall L, ed. PLoS ONE 2:e506.

Salazar P, Tapia R (2015) Epilepsy and hippocampal neurodegeneration induced by glutamate decarboxylase inhibitors in awake rats. Epilepsy Res 116:27–33.

Saleem F, Rizvi SW (2017) Transgender Associations and Possible Etiology: A Literature Review. Cureus 9:e1984.

Sanchez AM, Flamini MI, Fu X-D, Mannella P, Giretti MS, Goglia L, Genazzani AR, Simoncini T (2009) Rapid Signaling of Estrogen to WAVE1 and Moesin Controls Neuronal Spine Formation via the Actin Cytoskeleton. Mol Endocrinol 23:1193–1202.

Sanders J, Cowansage K, Baumgartel K, Mayford M (2012) Elimination of Dendritic Spines with Long-Term Memory Is Specific to Active Circuits. J Neurosci 32:12570–12578.

Saneyoshi T, Hayashi Y (2012) The Ca2+ and Rho GTPase signaling pathways underlying activity-dependent actin remodeling at dendritic spines. Cytoskeleton 69:545–554.

Sanz-Clemente A, Nicoll RA, Roche KW (2013) Diversity in NMDA Receptor Composition: Many Regulators, Many Consequences. The Neuroscientist 19:62–75.

Saunders JA, Tatard-Leitman VM, Suh J, Billingslea EN, Roberts TP, Siegel SJ (2013) Knockout of NMDA Receptors in Parvalbumin Interneurons Recreates Autism-Like Phenotypes: NMDAR1 KO in PV interneurons mimics autism. Autism Res 6:69–77.

Schade S, Paulus W (2016) D-Cycloserine in Neuropsychiatric Diseases: A Systematic Review. Int J Neuropsychopharmacol 19:pyv102.

Scheefhals N, MacGillavry HD (2018) Functional organization of postsynaptic glutamate receptors. Mol Cell Neurosci.

Schell MJ, Cooper OB, Snyder SH (1997) D-aspartate localizations imply neuronal and neuroendocrine roles. Proc Natl Acad Sci 94:2013–2018.

178

Schmeisser MJ et al. (2012) Autistic-like behaviours and hyperactivity in mice lacking ProSAP1/Shank2. Nature 486:256–260.

Schneider CA, Rasband WS, Eliceiri KW (2012) NIH Image to ImageJ: 25 years of image analysis. Nat METHODS 9:671–675.

Schoepp DD, Smith CL, Lodge D, Millar JD, Leander JD, Sacaan AI, Lunn WH (1991) D,L- (tetrazol-5-yl) glycine: a novel and highly potent NMDA receptor agonist. Eur J Pharmacol 203:237–243.

Schubert V, Da Silva JS, Dotti CG (2006) Localized recruitment and activation of RhoA underlies dendritic spine morphology in a glutamate receptor–dependent manner. J Cell Biol 172:453–467.

Schüler T, Mesic I, Madry C, Bartholomäus I, Laube B (2008) Formation of NR1/NR2 and NR1/NR3 Heterodimers Constitutes the Initial Step in N -Methyl-D-aspartate Receptor Assembly. J Biol Chem 283:37–46.

Schultz KN, von Esenwein SA, Hu M, Bennett AL, Kennedy RT, Musatov S, Toran-Allerand CD, Kaplitt MG, Young LJ, Becker JB (2009) Viral Vector-Mediated Overexpression of Estrogen Receptor- in Striatum Enhances the Estradiol-Induced Motor Activity in Female Rats and Estradiol-Modulated GABA Release. J Neurosci 29:1897–1903.

Schwechter B, Rosenmund C, Tolias KF (2013) RasGRF2 Rac-GEF activity couples NMDA receptor calcium flux to enhanced synaptic transmission. Proc Natl Acad Sci 110:14462–14467.

Seabold GK, Daunais JB, Rau A, Grant KA, Alvarez VA (2010) DiOLISTIC Labeling of Neurons from Rodent and Non-human Primate Brain Slices. J Vis Exp:e2081.

Seeburg PH, Burnashev N, Köhr G, Kuner T, Sprengel R, Monyer H (1995) The NMDA Receptor Channel: Molecular Design of a Coincidence Detector. In: Proceedings of the 1993 Laurentian Hormone Conference (Bardin CW, ed), pp 19–34 Recent Progress in Hormone Research. Boston: Academic Press. Available at: https://www.sciencedirect.com/science/article/pii/B9780125711500500068.

179

Seidlitz EP, Sharma MK, Saikali Z, Ghert M, Singh G (2009) Cancer cell lines release glutamate into the extracellular environment. Clin Exp Metastasis 26:781–787.

Sekar A, Bialas AR, de Rivera H, Davis A, Hammond TR, Kamitaki N, Tooley K, Presumey J, Baum M, Van Doren V, Genovese G, Rose SA, Handsaker RE, Daly MJ, Carroll MC, Stevens B, McCarroll SA (2016) Schizophrenia risk from complex variation of complement component 4. Nature 530:177–183.

Seubert P, Lee K, Lynch G (1989) Ischemia triggers NMDA receptor-linked cytoskeletal proteolysis in hippocampus. Brain Res 492:366–370.

Seyednasrollah F, Laiho A, Elo LL (2015) Comparison of software packages for detecting differential expression in RNA-seq studies. Brief Bioinform 16:59–70.

Sham PC, MacLean CJ, Kendler KS (1994) A typological model of schizophrenia based on age at onset, sex and familial morbidity. Acta Psychiatr Scand 89:135–141.

Shan Y, Carlock LR, Walker PD (1997) NMDA Receptor Overstimulation Triggers a Prolonged Wave of Immediate Early Gene Expression: Relationship to Excitotoxicity. Exp Neurol 144:406–415.

Shankar GM, Bloodgood BL, Townsend M, Walsh DM, Selkoe DJ, Sabatini BL (2007) Natural Oligomers of the Alzheimer Amyloid- Protein Induce Reversible Synapse Loss by Modulating an NMDA-Type Glutamate Receptor-Dependent Signaling Pathway. J Neurosci 27:2866–2875.

Shehadeh J, Fernandes HB, Zeron Mullins MM, Graham RK, Leavitt BR, Hayden MR, Raymond LA (2006) Striatal neuronal apoptosis is preferentially enhanced by NMDA receptor activation in YAC transgenic mouse model of Huntington disease. Neurobiol Dis 21:392–403.

Shen Y, Li N, Wu S, Zhou Y, Shan Y, Zhang Q, Ding C, Yuan Q, Zhao F, Zeng R, Zhu X (2008) Nudel Binds Cdc42GAP to Modulate Cdc42 Activity at the Leading Edge of Migrating Cells. Dev Cell 14:342–353.

Sheng M, Cummings J, Roldan L, Jan Y, Jan L (1994) Changing Subunit Composition of Heteromeric Nmda Receptors During Development of Rat Cortex. NATURE 368:144–147.

180

Shohami E, Biegon A (2014) Novel approach to the role of NMDA receptors in traumatic brain injury. CNS Neurol Disord Drug Targets 13:567–573.

Shouval HZ, Bear MF, Cooper LN (2002) A unified model of NMDA receptor-dependent bidirectional synaptic plasticity. Proc Natl Acad Sci 99:10831–10836.

Shu S-Y, Jiang G, Zeng Q-Y, Wang B, Li H, Ma L, Steinbusch H, Song C, Chan W-Y, Chen X- H, Wu Y-M, Bao R, Chen Y, Wu J-Y (2015) The Marginal Division of the Striatum and Hippocampus Has Different Role and Mechanism in Learning and Memory. Mol Neurobiol 51:827–839.

Silva AJ, Simpson EM, Takahashi JS, Lipp H-P, Nakanishi S, Wehner JM, Giese KP, Tully T, Abel T, Chapman PF, others (1997) Mutant mice and neuroscience: recommendations concerning genetic background. Neuron 19:755–759.

Simeone JC, Ward AJ, Rotella P, Collins J, Windisch R (2015) An evaluation of variation in published estimates of schizophrenia prevalence from 1990 horizontal line 2013: a systematic literature review. BMC Psychiatry 15:193.

Simopoulos A. (2002) The importance of the ratio of omega-6/omega-3 essential fatty acids. Biomed Pharmacother 56:365–379.

Simpson EH, Kellendonk C, Kandel E (2010) A Possible Role for the Striatum in the Pathogenesis of the Cognitive Symptoms of Schizophrenia. Neuron 65:585–596.

Smothers CT, Woodward JJ (2007) Pharmacological Characterization of Glycine-Activated Currents in HEK 293 Cells Expressing N-Methyl-D-aspartate NR1 and NR3 Subunits. J Pharmacol Exp Ther 322:739–748.

Snyder EM, Nong Y, Almeida CG, Paul S, Moran T, Choi EY, Nairn AC, Salter MW, Lombroso PJ, Gouras GK, Greengard P (2005) Regulation of NMDA receptor trafficking by amyloid-β. Nat Neurosci 8:1051–1058.

Snyder SH, Ferris CD (2000) Novel Neurotransmitters and Their Neuropsychiatric Relevance. Am J Psychiatry 157:1738–1751.

181

Soderling SH, Binns KL, Wayman GA, Davee SM, Ong SH, Pawson T, Scott JD (2002) The WRP component of the WAVE-1 complex attenuates Rac-mediated signalling. Nat Cell Biol 4:970–975.

Soderling SH, Guire ES, Kaech S, White J, Zhang F, Schutz K, Langeberg LK, Banker G, Raber J, Scott JD (2007) A WAVE-1 and WRP Signaling Complex Regulates Spine Density, Synaptic Plasticity, and Memory. J Neurosci 27:355–365.

Soderling SH, Langeberg LK, Soderling JA, Davee SM, Simerly R, Raber J, Scott JD (2003) Loss of WAVE-1 causes sensorimotor retardation and reduced learning and memory in mice. Proc Natl Acad Sci 100:1723–1728.

Somerville LH, Heatherton TF, Kelley WM (2006) Anterior cingulate cortex responds differentially to expectancy violation and social rejection. Nat Neurosci 9:1007–1008.

Soto D, Altafaj X, Sindreu C, Bayés À (2014) Glutamate receptor mutations in psychiatric and neurodevelopmental disorders. Commun Integr Biol 7:e27887.

Spooren W, Lindemann L, Ghosh A, Santarelli L (2012) Synapse dysfunction in autism: a molecular medicine approach to drug discovery in neurodevelopmental disorders. Trends Pharmacol Sci 33:669–684.

Sproul A, Steele SL, Thai TL, Yu S, Klein JD, Sands JM, Bell PD (2011) N-methyl-d-aspartate receptor subunit NR3a expression and function in principal cells of the collecting duct. Am J Physiol-Ren Physiol 301:F44–F54.

Steelman LS et al. (2011) Roles of the Raf/MEK/ERK and PI3K/PTEN/Akt/mTOR pathways in controlling growth and sensitivity to therapy-implications for cancer and aging. Aging 3:192– 222.

Steffens M, Huppertz H-J, Zentner J, Chauzit E, Feuerstein TJ (2005) Unchanged glutamine synthetase activity and increased NMDA receptor density in epileptic human neocortex: Implications for the pathophysiology of epilepsy. Neurochem Int 47:379–384.

Stertz L, Magalhães PVS, Kapczinski F (2013) Is bipolar disorder an inflammatory condition? The relevance of microglial activation. Curr Opin Psychiatry 26:19–26.

182

Stevenson RA, Park S, Cochran C, McIntosh LG, Noel J-P, Barense MD, Ferber S, Wallace MT (2017) The associations between multisensory temporal processing and symptoms of schizophrenia. Schizophr Res 179:97–103.

Strack S, Colbran RJ (1998) Autophosphorylation-dependent Targeting of Calcium/ Calmodulin- dependent Protein Kinase II by the NR2B Subunit of the N -Methyl- d-aspartate Receptor. J Biol Chem 273:20689–20692.

Stuber GD, Hnasko TS, Britt JP, Edwards RH, Bonci A (2010) Dopaminergic Terminals in the Nucleus Accumbens But Not the Dorsal Striatum Corelease Glutamate. J Neurosci 30:8229– 8233.

Sucher N, Awobuluyi M, Choi Y, Lipton S (1996) NMDA receptors: From genes to channels. TRENDS Pharmacol Sci 17:348–355.

Sucher NJ, Lipton SA (1991) Redox modulatory site of the NMDA receptor-channel complex: Regulation by oxidized glutathione. J Neurosci Res 30:582–591.

Suman S, Das TP, Sirimulla S, Alatassi H, Ankem MK, Damodaran C (2016) Withaferin-A suppress AKT induced tumor growth in colorectal cancer cells. Oncotarget 7:13854–13864.

Sun Y, Xu Y, Cheng X, Chen X, Xie Y, Zhang L, Wang L, Hu J, Gao Z (2018) The differences between GluN2A and GluN2B signaling in the brain. J Neurosci Res 96:1430–1443.

Sung JY, Engmann O, Teylan MA, Nairn AC, Greengard P, Kim Y (2008) WAVE1 controls neuronal activity-induced mitochondrial distribution in dendritic spines. Proc Natl Acad Sci U S A 105:3112–3116.

Svendsen F, Tjølsen A, Hole K (1998) AMPA and NMDA receptor-dependent spinal LTP after nociceptive tetanic stimulation: NeuroReport 9:1185–1190.

Swanger SA, Vance KM, Pare J-F, Sotty F, Fog K, Smith Y, Traynelis SF (2015) NMDA Receptors Containing the GluN2D Subunit Control Neuronal Function in the Subthalamic Nucleus. J Neurosci 35:15971–15983.

183

Sweatt JD (2016) Neural plasticity and behavior - sixty years of conceptual advances. J Neurochem 139:179–199.

Szychowski KA, Wnuk A, Rzemieniec J, Kajta M, Leszczynska T, Wojtowicz AK (2018) Triclosan-Evoked Neurotoxicity Involves NMDAR Subunits with the Specific Role of GluN2A in Caspase-3-Dependent Apoptosis. Mol Neurobiol.

Taber MT, Das S, Fibiger HC (2002) Cortical Regulation of Subcortical Dopamine Release: Mediation via the Ventral Tegmental Area. J Neurochem 65:1407–1410.

Takeda A, Tamano H (2009) Insight into zinc signaling from dietary zinc deficiency. Brain Res Rev 62:33–44.

Takenawa T, Suetsugu S (2007) The WASP–WAVE protein network: connecting the membrane to the cytoskeleton. Nat Rev Mol Cell Biol 8:37–48.

Talantova M et al. (2013) Aβ induces astrocytic glutamate release, extrasynaptic NMDA receptor activation, and synaptic loss. Proc Natl Acad Sci 110:E2518.

Tang F, Barbacioru C, Nordman E, Li B, Xu N, Bashkirov VI, Lao K, Surani MA (2010) RNA- Seq analysis to capture the transcriptome landscape of a single cell. Nat Protoc 5:516–535.

Tang G, Gudsnuk K, Kuo S-H, Cotrina ML, Rosoklija G, Sosunov A, Sonders MS, Kanter E, Castagna C, Yamamoto A, Yue Z, Arancio O, Peterson BS, Champagne F, Dwork AJ, Goldman J, Sulzer D (2014) Loss of mTOR-Dependent Macroautophagy Causes Autistic-like Synaptic Pruning Deficits. Neuron 83:1131–1143.

Tao X, Finkbeiner S, Arnold DB, Shaywitz AJ, Greenberg ME (1998) Ca2+ influx regulates BDNF transcription by a CREB family transcription factor-dependent mechanism. Neuron 20:709–726.

Tariot PN, Farlow MR, Grossberg GT, Graham SM, McDonald S, Gergel I, for the Memantine Study Group (2004) Memantine Treatment in Patients With Moderate to Severe Alzheimer Disease Already Receiving Donepezil: A Randomized Controlled Trial. JAMA 291:317.

184

Tashiro A, Minden A, Yuste R (2000) Regulation of dendritic spine morphology by the Rho family of small GTPases: Antagonistic roles of Rac and Rho. Cereb CORTEX 10:927–938.

Taya S, Shinoda T, Tsuboi D, Asaki J, Nagai K, Hikita T, Kuroda S, Kuroda K, Shimizu M, Hirotsune S, Iwamatsu A, Kaibuchi K (2007) DISC1 Regulates the Transport of the NUDEL/LIS1/14-3-3 Complex through Kinesin-1. J Neurosci 27:15–26.

Tecuapetla F, Patel JC, Xenias H, English D, Tadros I, Shah F, Berlin J, Deisseroth K, Rice ME, Tepper JM, Koos T (2010) Glutamatergic Signaling by Mesolimbic Dopamine Neurons in the Nucleus Accumbens. J Neurosci 30:7105–7110.

Teng H, Cai W, Zhou L, Zhang J, Liu Q, Wang Y, Dai W, Zhao M, Sun Z (2010) Evolutionary Mode and Functional Divergence of Vertebrate NMDA Receptor Subunit 2 Genes DeSalle R, ed. PLoS ONE 5:e13342.

Texidó L, Martín-Satué M, Alberdi E, Solsona C, Matute C (2011) Amyloid β peptide oligomers directly activate NMDA receptors. Cell Calcium 49:184–190.

Thara R, Kamath S (2015) Women and schizophrenia. Indian J Psychiatry 57:246.

Thomas GM, Huganir RL (2004) MAPK cascade signalling and synaptic plasticity. Nat Rev Neurosci 5:173–183.

Thorup A, Waltoft BL, Pedersen CB, Mortensen PB, Nordentoft M (2007) Young males have a higher risk of developing schizophrenia: a Danish register study. Psychol Med 37:479.

Tian X, Gotoh T, Tsuji K, Lo EH, Huang S, Feig LA (2004) Developmentally regulated role for Ras-GRFs in coupling NMDA glutamate receptors to Ras, Erk and CREB. EMBO J 23:1567– 1575.

Tolman AW, Kurtz MM (2012) Neurocognitive Predictors of Objective and Subjective Quality of Life in Individuals With Schizophrenia: A Meta-Analytic Investigation. Schizophr Bull 38:304–315.

Tönnies E, Trushina E (2017) Oxidative Stress, Synaptic Dysfunction, and Alzheimer’s Disease. J Alzheimers Dis 57:1105–1121.

185

Toscano CD, Kingsley PJ, Marnett LJ, Bosetti F (2008) NMDA-induced seizure intensity is enhanced in COX-2 deficient mice. NeuroToxicology 29:1114–1120.

Tovar KR, Westbrook GL (2017) Modulating synaptic NMDA receptors. Neuropharmacology 112:29–33.

Toyo-oka K, Shionoya A, Gambello MJ, Cardoso C, Leventer R, Ward HL, Ayala R, Tsai L-H, Dobyns W, Ledbetter D, Hirotsune S, Wynshaw-Boris A (2003) 14-3-3ε is important for neuronal migration by binding to NUDEL: a molecular explanation for Miller–Dieker syndrome. Nat Genet 34:274–285.

Trapnell C, Roberts A, Goff L, Pertea G, Kim D, Kelley DR, Pimentel H, Salzberg SL, Rinn JL, Pachter L (2012) Differential gene and transcript expression analysis of RNA-seq experiments with TopHat and Cufflinks. Nat Protoc 7:562–578.

Trapnell C, Williams BA, Pertea G, Mortazavi A, Kwan G, van Baren MJ, Salzberg SL, Wold BJ, Pachter L (2010) Transcript assembly and quantification by RNA-Seq reveals unannotated transcripts and isoform switching during cell differentiation. Nat Biotechnol 28:511–515.

Traynelis SF, Cull-Candy SG (1990) Proton inhibition of N-methyl-D-aspartate receptors in cerebellar neurons. Nature 345:347.

Traynelis SF, Wollmuth LP, McBain CJ, Menniti FS, Vance KM, Ogden KK, Hansen KB, Yuan H, Myers SJ, Dingledine R (2010) Glutamate Receptor Ion Channels: Structure, Regulation, and Function. Pharmacol Rev 62:405–496.

Tsai GE, Falk WE, Gunther J, Coyle JT (1999) Improved Cognition in Alzheimer’s Disease With Short-Term D-Cycloserine Treatment. Am J Psychiatry 156:467–469.

Tu JC, Xiao B, Naisbitt S, Yuan JP, Petralia RS, Brakeman P, Doan A, Aakalu VK, Lanahan AA, Sheng M, Worley PF (1999) Coupling of mGluR/Homer and PSD-95 Complexes by the Shank Family of Postsynaptic Density Proteins. Neuron 23:583–592.

Ueno T, Tsukuda K, Toyooka S, Ando M, Takaoka M, Soh J, Asano H, Maki Y, Muraoka T, Tanaka N, Shien K, Furukawa M, Yamatsuji T, Kiura K, Naomoto Y, Miyoshi S (2012) Strong

186 anti-tumor effect of NVP-AUY922, a novel Hsp90 inhibitor, on non-small cell lung cancer. Lung Cancer 76:26–31.

Uesaka N, Uchigashima M, Mikuni T, Nakazawa T, Nakao H, Hirai H, Aiba A, Watanabe M, Kano M (2014) Retrograde semaphorin signaling regulates synapse elimination in the developing mouse brain. Science 344:1020–1023.

Urbano M, Okwara L, Manser P, Hartmann K, Deutsch SI (2015) A Trial of D-Cycloserine to Treat the Social Deficit in Older Adolescents and Young Adults With Autism Spectrum Disorders. J Neuropsychiatry Clin Neurosci 27:133–138.

Urbano M, Okwara L, Manser P, Hartmann K, Herndon A, Deutsch SI (2014) A Trial of D- Cycloserine to Treat Stereotypies in Older Adolescents and Young Adults With Autism Spectrum Disorder: Clin Neuropharmacol 37:69–72.

Uribe E, Wix R (2011) Epigenetic control of the GABAergic interneurons migration and NMDA receptor functioning in schizophrenia. In, pp 13.

Valianou M, Cox AM, Pichette B, Hartley S, Paladhi UR, Astrinidis A (2015) Pharmacological inhibition of Polo-like kinase 1 (PLK1) by BI-2536 decreases the viability and survival of hamartin and tuberin deficient cells via induction of apoptosis and attenuation of autophagy. Cell Cycle 14:399–407.

Vallano ML (1998) Developmental aspects of NMDA receptor function. Crit Rev Neurobiol 12:177–204. van den Buuse M, Low JK, Kwek P, Martin S, Gogos A (2017) Selective enhancement of NMDA receptor-mediated locomotor hyperactivity by male sex hormones in mice. Psychopharmacology (Berl) 234:2727–2735. van Os J, Kapur S (2009) Schizophrenia. LANCET 374:635–645.

Vanasse A, Courteau J, Fleury M-J, Grégoire J-P, Lesage A, Moisan J (2012) Treatment prevalence and incidence of schizophrenia in Quebec using a population health services perspective: different algorithms, different estimates. Soc Psychiatry Psychiatr Epidemiol 47:533–543.

187

Varga V, Jenei Z, Janaky R, Saransaari P, Oja SS (1997) Glutathione is an endogenous ligand of rat brain N-methyl-D-aspartate (NMDA) and. Neurochem Res 22:1165–1171.

Vaux DL, Korsmeyer SJ (1999) Cell Death in Development. Cell 96:245–254.

Vink M, de Leeuw M, Luykx JJ, van Eijk KR, van den Munkhof HE, van Buuren M, Kahn RS (2016) DRD2 Schizophrenia-Risk Allele Is Associated With Impaired Striatal Functioning in Unaffected Siblings of Schizophrenia Patients. Schizophr Bull 42:843–850.

Volkmar FR, Nelson DS (1990) Seizure Disorders in Autism. J Am Acad Child Adolesc Psychiatry 29:127–129.

Voorn P, Vanderschuren LJM., Groenewegen HJ, Robbins TW, Pennartz CM. (2004) Putting a spin on the dorsal–ventral divide of the striatum. Trends Neurosci 27:468–474.

Vrajová M, Šťastný F, Horáček J, Lochman J, Šerý O, Peková S, Klaschka J, Höschl C (2010) Expression of the Hippocampal NMDA Receptor GluN1 Subunit and Its Splicing Isoforms in Schizophrenia: Postmortem Study. Neurochem Res 35:994–1002.

Walker DL, Gold PE (1991) Effects of the novel NMDA antagonist, NPC 12626, on long-term potentiation, learning and memory. Brain Res 549:213–221.

Wang W-P, Iyo AH, Miguel-Hidalgo J, Regunathan S, Zhu M-Y (2006) Agmatine protects against cell damage induced by NMDA and glutamate in cultured hippocampal neurons. Brain Res 1084:210–216.

Wang X-Q, Peng Y-P, Lu J-H, Cao B-B, Qiu Y-H (2009) Neuroprotection of interleukin-6 against NMDA attack and its signal transduction by JAK and MAPK. Neurosci Lett 450:122– 126.

Wang Yansong, Wang Wei, Li Dongguo, Li Mi, Wang Peipei, Wen Jian, Liang Min, Su Bo, Yin Yanling (2014) IGF-1 Alleviates NMDA-Induced Excitotoxicity in Cultured Hippocampal Neurons Against Autophagy via the NR2B/PI3K-AKT-mTOR Pathway. J Cell Physiol 229:1618–1629.

188

Wang Y-C, Ho U-C, Ko M-C, Liao C-C, Lee L-J (2012) Differential neuronal changes in medial prefrontal cortex, basolateral amygdala and nucleus accumbens after postweaning social isolation. Brain Struct Funct 217:337–351.

Watanabe M, Inoue Y, Sakimura K, Mashina M (1992) Developmental-Changes in Distribution of Nmda Receptor Channel Subunit Messenger-RNAs. NeuroReport 3:1138–1140.

Wee KS-L, Zhang Y, Khanna S, Low C-M (2008) Immunolocalization of NMDA receptor subunit NR3B in selected structures in the rat forebrain, cerebellum, and lumbar spinal cord. J Comp Neurol 509:118–135.

Weger M, Sevelinges Y, Grosse J, de Suduiraut IG, Zanoletti O, Sandi C (2018) Increased brain glucocorticoid actions following social defeat in rats facilitates the long-term establishment of social subordination. Physiol Behav 186:31–36.

Weickert CS, Fung SJ, Catts VS, Schofield PR, Allen KM, Moore LT, Newell KA, Pellen D, Huang X-F, Catts SV, others (2013) Molecular evidence of N-methyl-D-aspartate receptor hypofunction in schizophrenia. Mol Psychiatry 18:1185–1192.

Weilinger NL, Lohman AW, Rakai BD, Ma EMM, Bialecki J, Maslieieva V, Rilea T, Bandet MV, Ikuta NT, Scott L, Colicos MA, Teskey GC, Winship IR, Thompson RJ (2016) Metabotropic NMDA receptor signaling couples Src family kinases to pannexin-1 during excitotoxicity. Nat Neurosci 19:432–442.

Wenzel A, Fritschy J, Mohler H, Benke D (1997) NMDA receptor heterogeneity during postnatal development of the rat brain: Differential expression of the NR2A, NR2B, and NR2C subunit proteins. J Neurochem 68:469–478.

Wesseling H, Guest PC, Lee C-M, Wong EH, Rahmoune H, Bahn S (2014) Integrative proteomic analysis of the NMDA NR1 knockdown mouse model reveals effects on central and peripheral pathways associated with schizophrenia and autism spectrum disorders. Mol Autism 5:38.

Whiteford HA, Degenhardt L, Rehm J, Baxter AJ, Ferrari AJ, Erskine HE, Charlson FJ, Norman RE, Flaxman AD, Johns N, Burstein R, Murray CJ, Vos T (2013) Global burden of disease

189 attributable to mental and substance use disorders: findings from the Global Burden of Disease Study 2010. The Lancet 382:1575–1586.

Whitlock JR, Heynen AJ, Shuler MG, Bear MF (2006) Learning induces long-term potentiation in the hippocampus. Science 313:1093–1097.

Won H, Lee H-R, Gee HY, Mah W, Kim J-I, Lee J, Ha S, Chung C, Jung ES, Cho YS, Park S-G, Lee J-S, Lee K, Kim D, Bae YC, Kaang B-K, Lee MG, Kim E (2012) Autistic-like social behaviour in Shank2-mutant mice improved by restoring NMDA receptor function. Nature 486:261–265.

Wong H-K, Liu X-B, Matos MF, Chan SF, Pérez-Otaño I, Boysen M, Cui J, Nakanishi N, Trimmer JS, Jones EG, Lipton SA, Sucher NJ (2002) Temporal and regional expression of NMDA receptor subunit NR3A in the mammalian brain: Expression of NMDA Receptor Subunit NR3A. J Comp Neurol 450:303–317.

Wong S-B, Hung W-C, Min M-Y (2016) The Role of Gastrodin on Hippocampal Neurons after N-Methyl-D-Aspartate Excitotoxicity and Experimental Temporal Lobe Seizures. Chin J Physiol 59:156–164.

Wu FS, Gibbs TT, Farb DH (1991) Pregnenolone sulfate: a positive allosteric modulator at the N-methyl-D-aspartate receptor. Mol Pharmacol 40:333–336.

Wyllie DJA, Livesey MR, Hardingham GE (2013) Influence of GluN2 subunit identity on NMDA receptor function. Neuropharmacology 74:4–17.

Xia P, Chen H -s. V, Zhang D, Lipton SA (2010) Memantine Preferentially Blocks Extrasynaptic over Synaptic NMDA Receptor Currents in Hippocampal Autapses. J Neurosci 30:11246–11250.

Xia Z, Dudek H, Miranti CK, Greenberg ME (1996) Calcium Influx via the NMDA Receptor Induces Immediate Early Gene Transcription by a MAP Kinase/ERK-Dependent Mechanism. J Neurosci 16:5425.

Xing G-G, Wang R, Yang B, Zhang D (2006) Postnatal switching of NMDA receptor subunits from NR2B to NR2A in rat facial motor neurons. Eur J Neurosci 24:2987–2992.

190

Xu J, Kurup P, Zhang Y, Goebel-Goody SM, Wu PH, Hawasli AH, Baum ML, Bibb JA, Lombroso PJ (2009) Extrasynaptic NMDA Receptors Couple Preferentially to Excitotoxicity via Calpain-Mediated Cleavage of STEP. J Neurosci 29:9330–9343.

Xu X-X, Luo J-H (2018) Mutations of N-Methyl-D-Aspartate Receptor Subunits in Epilepsy. Neurosci Bull 34:549–565.

Yamamoto S, Morinobu S, Fuchikami M, Kurata A, Kozuru T, Yamawaki S (2008) Effects of single prolonged stress and D-cycloserine on contextual fear extinction and hippocampal NMDA receptor expression in a rat model of PTSD. Neuropsychopharmacol Off Publ Am Coll Neuropsychopharmacol 33:2108–2116.

Yang X, Schadt EE, Wang S, Wang H, Arnold AP, Ingram-Drake L, Drake TA, Lusis AJ (2006) Tissue-specific expression and regulation of sexually dimorphic genes in mice. Genome Res 16:995–1004.

Yang X-C, Reis DJ (1999) Agmatine Selectively Blocks the N-Methyl-D-Aspartate Subclass of Glutamate Receptor Channels in Rat Hippocampal Neurons. 288:6.

Yao Y, Belcher J, Berger AJ, Mayer ML, Lau AY (2013) Conformational Analysis of NMDA Receptor GluN1, GluN2, and GluN3 Ligand-Binding Domains Reveals Subtype-Specific Characteristics. Structure 21:1788–1799.

Yi F, Zachariassen LG, Dorsett KN, Hansen KB (2018) Properties of Triheteromeric N-Methyl- d-Aspartate Receptors Containing Two Distinct GluN1 Isoforms. Mol Pharmacol 93:453–467.

Yin M, Lu Q, Liu X, Wang T, Liu Y, Chen L (2016) Silencing Drp1 inhibits glioma cells proliferation and invasion by RHOA/ ROCK1 pathway. Biochem Biophys Res Commun 478:663–668.

Yoon JY, Choi S (2017) Evidence for presynaptically silent synapses in the immature hippocampus. Biochem Biophys Res Commun 482:1375–1380.

Young A, Greenamyre J, Hollingsworth Z, Albin R, D’Amato C, Shoulson I, Penney J (1988) NMDA receptor losses in putamen from patients with Huntington’s disease. Science 241:981.

191

Yu SP, Yeh C, Strasser U, Tian M, Choi DW (1999) NMDA receptor-mediated K+ efflux and neuronal apoptosis. Science 284:336–339.

Yuan H, Low C-M, Moody OA, Jenkins A, Traynelis SF (2015) Ionotropic GABA and Glutamate Receptor Mutations and Human Neurologic Diseases. Mol Pharmacol 88:203–217.

Yurkewicz L, Weaver J, Bullock MR, Marshall LF (2005) The Effect of the Selective NMDA Receptor Antagonist Traxoprodil in the Treatment of Traumatic Brain Injury. J Neurotrauma 22:1428–1443.

Yuzaki M, Connor JA (1999) Characterization of l -Homocysteate–Induced Currents in Purkinje Cells From Wild-Type and NMDA Receptor Knockout Mice. J Neurophysiol 82:2820–2826.

Zahid MA, Ohaeri JU (2010) Relationship of family caregiver burden with quality of care and psychopathology in a sample of Arab subjects with schizophrenia. BMC Psychiatry 10:71.

Zeidán-Chuliá F, Rybarczyk-Filho JL, Salmina AB, de Oliveira B-HN, Noda M, Moreira JCF (2013) Exploring the Multifactorial Nature of Autism Through Computational Systems Biology: Calcium and the Rho GTPase RAC1 Under the Spotlight. NeuroMolecular Med 15:364–383.

Zhang HX, Hyrc K, Thio LL (2009a) The glycine transport inhibitor sarcosine is an NMDA receptor co-agonist that differs from glycine: Sarcosine is an NMDAR co-agonist. J Physiol 587:3207–3220.

Zhang J, Hou L, Gao X, Guo F, Jing W, Qi J, Qiao J (2009b) Amyloid β-protein differentially affects NMDA receptor- and GABAA receptor-mediated currents in rat hippocampal CA1 neurons. Prog Nat Sci 19:963–972.

Zhang S-J, Steijaert MN, Lau D, Schütz G, Delucinge-Vivier C, Descombes P, Bading H (2007) Decoding NMDA Receptor Signaling: Identification of Genomic Programs Specifying Neuronal Survival and Death. Neuron 53:549–562.

Zhang Y, Li P, Feng J, Wu M (2016) Dysfunction of NMDA receptors in Alzheimer’s disease. Neurol Sci 37:1039–1047.

192

Zhu S, Gouaux E (2017) Structure and symmetry inform gating principles of ionotropic glutamate receptors. Neuropharmacology 112:11–15.

Zhu S, Paoletti P (2015) Allosteric modulators of NMDA receptors: multiple sites and mechanisms. Curr Opin Pharmacol 20:14–23.

Zhuo M (2008) Cortical excitation and chronic pain. Trends Neurosci 31:199–207.

Appendices

Appendix 1: Supplemental Figures and Tables

Supplemental Table SI. Protein expression changes in adult male GluN1KD mice relative to WT littermates from previous proteomic studies. These results are compiled from both previous studies (Wesseling et al., 2014; Cox et al., 2016). Any differences between cortical, hippocampal, and serum results regarding the number of identified proteins do not necessarily reflect biologically significant differences. Instead, they likely reflect the fact that each physiological location involved different studies and methods. Proteins in bold have been confirmed by both studies or by multiple experimental methods. Proteins in red represent findings that are conflicted between studies or experimental methods. Many of these protein changes have been related to synaptic plasticity.

193 194

Frontal Cortex Hippocampus Serum Increased Decreased Increased Decreased Increased Decreased 4F2 SBP2 14-3-3 ε H2B1C TBB4A 3HIDH ANXA3 APOA1 2AR AP180 SUCB1 14-3-3 θ H31 TBB4B ACES ASNS CSF-1 ACTH AP1B1 SURF6 14-3-3 ζ HMGCL TBB6 ATIF1 BIN1 FGA CCL11 AP2α1 SYN1 2A5G HPCA THY1 C560 CaMKIIβ GSTM1 CLUS CACB2 SYNJ 5NT1B HS90A TLK2 CCL9 CHDH IL-12α CXCL5 CDC6 SYT1 ACON HSP7C TM143 CKAP2 F16P1 OSM IGA CI172 SYUB ACTS K2C72 TNNC2 CO8A2 GluN1 TIMP-1 IGF-1 Coro1A TBAL3 ADM2 KCC2G TRPV6 CP1A2 GSTP1 VEGF-A IL-1β CYC TPM3 ALDOA KCRB TWST1 CPLX1 LNP vWF LEP DC1L1 TYRP2 ARF2 KCRS UCRI DDX46 MBP MCP-3 DCDC2 VGLUT1 AT1A1 KCRU VISL1 DLG4 ODO1 MCP-5 DHB3 AT1A2 KPYM WDR5 ERK1 PHB2 MIP2 DLG4 AT1A4 LDHA XIRP2 ERK2 RAB8A XCL1 EFTU ATPA LDHB ZSC21 GAP43 RFIP1 ERK1 BPNT1 LMTK2 GBB1 ROA1 ERK2 CAH2 LRC34 GFAP RYK FRM4A CaMKIIα MAGIX GOT2 SEM6D GABA-T CaNB1 MAL2 HGPRT SPTB1 GAD2 CAZA2 MBP HHATL SYCP3 GAP43 CDR2L MDHM K2C1B TKTL1 GFAP CETN2 NFL MARCS GHC1 CLH1 NKD1 MetAP-2 GOT2 CN37 NU4M MMAA H14 CNP NUCB1 MMSA H2B1 DPYL2 ODPB MPIP2 HPCL4 DYH12 PICAL NCAM1 KIF15 DYN1 PLP NDUA4 LIAS EF1A1 PRDX6 NDUFS1 MARCS EHD4 PTMS NEFL MARCS ENOA RABE2 NO66 MFHA1 ENOB RL40 NSF NCAM1 G6PI RLA1 PACSIN1 NCDN GBB3 SEPT5 PEA-15 NDUFS1 GBG12 SNAB PKM NSF GluN1 SNAPN PLCB3 PEA-15 GNAI1 SNP25 PLP PGES2 GNAI2 STXB1 PTBP2 PKM GNAI3 T3JAM RAB6A PP2BAβ GNAO TAGL2 SPRL1 PURA1 GPSM1 TBA1B SYT1 RBM14 GSTM1 TBB1 TXND2 RRP15 H2A2A TBB3 UBA1Y

195

Supplemental Figure S1. Dendritic spine morphology measurements of striatal MSNs in WT, WAVE-Tg, GluN1KD and GluN1KD-WAVE mice. Spine morphology measures (left) and associated spine-population cumulative-frequency distributions (right) were not significantly different between WT, WAVE-Tg, GluN1KD, and GluN1KD-WAVE mice. Mouse sample sizes are indicated within each bar for bar graphs, which also apply to associated spine-population frequency data. 3-6 sample dendrite images were analyzed for each mouse. Mean morphology measures were analyzed by one-way ANOVAs. WT vs GluN1KD spine-population cumulative frequencies were compared using Kolmogorov–Smirnov tests. Since no significant WT vs GluN1KD differences were found, further pairwise frequency comparisons were not performed.

196

Supplemental Figure S2. Dendritic spine categorization of striatal MSNs in WT, WAVE- Tg, GluN1KD and GluN1KD-WAVE mice. The number (left) and proportion (right) of mushroom, stubby, and thin spines were derived from dendritic spine density and morphology data. Spines were categorized as stubby if they were shorter than 0.8 µm. Of the longer ones,

197 mushroom spines had a head-to-neck ratio ≥ 1.1. The rest were categorized as thin spines. No significant differences were found between genotypes. Mouse sample sizes are indicated within each bar. 3-6 sample dendrite images were analyzed for each mouse. Data were analyzed by one- way ANOVAs.

Supplemental Figure S3. Apical dendritic spine morphology measurements of CA1 pyramidal neurons in WT, WAVE-Tg, GluN1KD and GluN1KD-WAVE mice. Spine morphology measures are shown on the left while associated spine-population cumulative- frequency distributions are shown on the right. GluN1KD-WAVE hybrids had a significantly higher mean spine length compared to WT littermates while GluN1KD mice did not. By spine-

198 population distribution comparisons, both GluN1KD and GluN1KD-WAVE mice had more long spines than WT littermates. For both spine length analyses, GluN1KD and GluN1KD-WAVE mice were not different from each other. Thus, GluN1KD mice might have longer spines than WT mice, but any such abnormality would not be rescued in GluN1KD-WAVE hybrids. Spine head width and neck width were not significantly different between WT, WAVE-Tg, GluN1KD, and GluN1KD-WAVE mice. Mouse sample sizes are indicated within each bar for bar graphs, which also apply to associated spine-population frequency data. 3-6 sample dendrite images were analyzed for each mouse. Mean morphology measures were analyzed by one-way ANOVAs followed by Bonferroni post hoc comparisons for all pairings. * p = 0.03. Spine- population cumulative frequencies were compared using Kolmogorov–Smirnov tests, which were performed first for WT vs GluN1KD mice. GluN1KD vs GluN1KD-WAVE and WT vs GluN1KD-WAVE comparisons were performed only if a significant difference was found for WT vs GluN1KD, to decrease the chance for false positive results. # p < 0.01 vs WT distribution.

199

Supplemental Figure S4. Apical dendritic spine categorization of CA1 pyramidal neurons in WT, WAVE-Tg, GluN1KD and GluN1KD-WAVE mice. The number (left) and proportion (right) of mushroom, stubby, and thin spines were derived from dendritic spine density and morphology data. Spines were categorized as stubby if they were shorter than 0.8 µm. Of the

200 longer ones, mushroom spines had a head-to-neck ratio ≥ 1.1. The rest were categorized as thin spines. GluN1KD-WAVE hybrids had more mushroom spines by proportion and less stubby spines by number compared to WT mice. GluN1KD mice had less stubby spines by number compared to WT mice. No differences were significant between GluN1KD mice and GluN1KD- WAVE hybrids. Thus, GluN1KD mice might have fewer stubby spines than WT littermates, but any such deficit would not be rescued in GluN1KD-WAVE hybrids. No significant differences were found between genotypes for thin spines. Mouse sample sizes are indicated within each bar. 3-6 sample dendrite images were analyzed for each mouse. Data were analyzed by one-way ANOVAs followed by Bonferroni post hoc comparisons for all pairings. * p ≤ 0.05, ** p < 0.01.

201

Supplemental Figure S5. Basolateral dendritic spine morphology measurements of CA1 pyramidal neurons in WT, WAVE-Tg, GluN1KD and GluN1KD-WAVE mice. Spine morphology measures are shown on the left while associated spine-population cumulative- frequency distributions are shown on the right. Mean spine morphology measurements were not significantly different between WT, WAVE-Tg, GluN1KD, and GluN1KD-WAVE mice. By spine-population distribution comparisons, both GluN1KD and GluN1KD-WAVE mice had more long spines than WT littermates. GluN1KD and GluN1KD-WAVE spine distributions were not different from each other. Thus, GluN1KD mice have more long spines than WT littermates, but this was not rescued in GluN1KD-WAVE hybrids. Mouse sample sizes are indicated within

202 each bar for bar graphs, which also apply to associated spine-population frequency data. 3-6 sample dendrite images were analyzed for each mouse. Mean morphology measures were analyzed by one-way ANOVAs. Spine-population cumulative frequencies were compared using Kolmogorov–Smirnov tests, which were performed first for WT vs GluN1KD mice. GluN1KD vs GluN1KD-WAVE and WT vs GluN1KD-WAVE comparisons were performed only if a significant difference was found for WT vs GluN1KD, to decrease the chance for false positive results. # p < 0.01 vs WT distribution.

203

Supplemental Figure S6. Basolateral dendritic spine categorization of CA1 pyramidal neurons in WT, WAVE-Tg, GluN1KD and GluN1KD-WAVE mice. The number (left) and proportion (right) of mushroom, stubby, and thin spines were derived from dendritic spine density and morphology data. Spines were categorized as stubby if they were shorter than 0.8

204

µm. Of the longer ones, mushroom spines had a head-to-neck ratio ≥ 1.1. The rest were categorized as thin spines. No significant differences were found between genotypes. Mouse sample sizes are indicated within each bar. 3-6 sample dendrite images were analyzed for each mouse. Data were analyzed by one-way ANOVAs.

Supplemental Table SII. Full differentially-expressed gene lists showing demasculinization or feminization in GluN1KD males and masculinization or defeminization in GluN1KD females. These gene lists are the result of filtering the 784 sex-factored RNAseq results for those that were also different between male and female WT mice. Demasculinizing or feminizing gene expression changes did not show up at all in GluN1KD females. Similarly, masculinizing or defeminizing gene expression changes did not show up in GluN1KD males. All noted changes in GluN1KD mice are relative to sex-matched WT littermates unless specified otherwise.

205

Changed in Male GluN1KD Mice Changed in Female GluN1KD Mice Demasculinized (98) Feminized (88) Masculinized (103) Defeminized (75) 2310046A06Rik Kcnmb4 2410004P03Rik Peg10 3110035E14Rik Nxph3 4932411E22Rik Prlr 3110035E14Rik Klf10 Acss3 Plagl1 3110047P20Rik Olfm1 AC124426.2 Prox1 3110047P20Rik Klf16 Agt Plce1 Adcy1 Pde1a AC124426.5 Ptgds AC165261.1 Lamc2 Agxt2l1 Plch1 Basp1 Phf24 Agt Pvalb Arhgap15 Lingo1 Arhgap36 Plin3 Bdnf Plekhg5 Akap12 Rab37 Basp1 Mapk11 Asb4 Pnoc Bmp3 Prdm8 Amigo2 Rassf4 Bdnf Mef2c AW551984 Pnpla7 Bok Prss12 Ankfn1 Sdf2l1 Bhlhe22 Mpped1 Baiap3 Prkcd C130074G19Rik Prss23 Ankrd34b Sepp1 Btg2 Myl4 Calb2 Prlr C1qtnf6 Ptgs2 Arhgap24 Sfxn1 C1qtnf6 Neurl1b Calcr Resp18 Cbln2 Rap1gap2 AW551984 Slc17a6 Cabp1 Neurod2 Cd38 Ret Cck Rapgefl1 Baiap3 Slc6a11 Cacnb3 Neurod6 Cd83 Rxfp3 Cckbr Rasl11b Calb2 Slc6a20a Car10 Nov Cldn1 Scn5a Cnih3 Rbm3-ps Calr Stip1 Cck Nptx1 Cldn11 Sdc4 Col11a1 Rnf39 Cbln1 Synpo2 Cckbr Nr4a2 Clic4 Sepp1 Col5a1 Rprm Cd38 Tac2 Cnih3 Nrn1 Cpm Sfrp2 Col6a1 Rspo2 Cldn1 Tcf7l2 Col5a1 Nxph3 Dgkk Sgk1 Cplx3 Rtn4r Creld2 Tmem47 Col6a1 Olfm1 Dlk1 Slc17a6 Crhbp Rtn4rl2 Cryab Trh Cplx3 Pamr1 Ebf3 Slc6a11 Ctgf Rxfp1 Ddc Tshz2 Ctgf Pcsk1n Ecel1 Sparc Dkk3 Satb2 Dgkk Ttr Dgkz Prr7 Emilin2 Tcf7l2 Dkkl1 Scube1 Entpd1 Unc13c Dkkl1 Prss12 Esr1 Tmem130 Dtl Sel1l3 Fbln5 Xbp1 E130012A19Rik Rapgefl1 Fam70a Tmie E2f1 Serinc2 Fign Zim1 Efhd2 Rara Frem3 Tns3 Efhd2 Serpinb8 Frem3 Egr4 Rasgef1c Gabre Tox2 Emx1 Serpini1 Frmpd1 Fam163b Robo3 Gabrg1 Tppp3 Ephb6 Shank1 Gabrq Fam43b Rprm Gabrq Trh Faah Sidt1 Gal Fam81a Rprml Gal Trpm3 Fezf2 Slc17a7 Gm26825 Fhl2 Rspo2 Gja1 Tsc22d3 Fhl2 Slc30a3 Gm9989 Foxo6 Rtn4r Gjb6 Ttr Gm11549 Smoc2 Gpr165 Frzb Rtn4rl2 Gm42644 Usp29 Hs3st2 Sprn Hspa1b Fxyd5 Satb2 Gpc3 Vwa5b1 Igfbp6 Sstr3 Hspa5 Fxyd7 Scube1 Gpr101 Wdr6 Islr2 St8sia5 Hspa8 Gca Slc17a7 Gpr165 Zfhx4 Krt80 Stac2 Hspb8 Gm11549 Slc30a3 Gprasp2 Zic1 Lgals1 Stx1a Igsf1 Gm22175 Smoc2 Gprin2 Zim1 Lingo1 Sulf1 Inadl Gm24022 Snhg14 Hap1 LOC432842 Sv2b Irs4 Gm25747 Sstr3 Heph Mef2c Tbr1 Itih5 Gm40518 Stx1a Hif3a Mical2 Tmem178 Kank1 Golga7b Sv2b Hspb8 Mpped1 Tnfrsf25 Klhl1 Gprin1 Sytl2 Igf2 Ncald Trbc1 Magt1 Gucy2g Tbr1 Igsf1 Necab3 Trbc2 Nde1 Ier2 Vip Irs4 Neurl1b Trbd2 Ngb Ier5 Vipr1 Itih3 Neurod2 Trbj2-1 Nr2f2 Ighj1 Wnt9a Magel2 Neurod6 Trbj2-2 Oxt Ighj2 Xkr7 Mt2 Nfib Trbj2-3 Pdia4 Ighj3 Nde1 Ninl Trbj2-4 Pdia6 Ighj4 Ngb Nov Trbj2-5 Pdp1 Ighm Nr2f2 Nptx1 Trbj2-7 Pdzd2 Igsf21 Nynrin Nr4a2 Tshz3 Pgr15l Islr2 Oxt Nrn1 Wnt9a Plat Junb Pdgfd Ntng2 Plekhg1

No copyrighted works were used in this thesis as of September 26, 2018.

206