Automated Immunohistochemical Analysis of the Orbitofrontal Cortex in Patients with Schizophrenia, Bipolar Disorder and Major Depressive Disorder

Kathleen Trought

A thesis submitted in conformity with the requirements for the degree of Master of Science Institute of Medical Science University of Toronto

Kathleen Trought

Master of Science

Institute of Medical Science University of Toronto

2017 Abstract

Previous studies have found evidence for orbitofrontal cortex (OFC) pathology in major depressive disorder (MDD), bipolar disorder (BP) and to a lesser extent schizophrenia (SCZ).

However, given that the OFC is a large heterogeneous area, it is difficult to assess how findings from small subareas translate to the entire region. The aim of this thesis is to analyze the entire

OFC in patients with MDD, BP and SCZ. Using a novel approach with layer-specific immunohistochemical markers and an automated counting protocol, we were able to analyze the cortical width, cell density, cell area and distance from pia in the entire OFC of 60 post-mortem brain samples (15 control, 15 MDD, 15 BP, 15 SCZ). We did not find strongly significant differences between patients and control subjects. Our findings suggest that inconsistencies in the literature may arise from sampling only small areas of the cortex in a limited number of subjects.

Acknowledgments

I would like to thank my supervisor and mentor Dr. Albert Wong for his continuous support, guidance and encouragement throughout the completion of this degree. I greatly appreciate his mentorship over the past two years and his constant support in helping me to achieve my future goals. I would also like to thank my committee members Dr. Sheena Josselyn and Dr. Jeff Daskalakis for supporting me throughout my degree and providing me with guidance.

There are several individuals who have helped me a tremendous amount throughout this project and I want to express my sincerest gratitude. Firstly, thank you to both the present and former members of the Wong lab for all of their support and assistance: Mohamad Abbass,

Donald Wang, Frankie Lee, James Samson, John Zawadzki, Jialun Chen and Meng Xi Yu. Thank you to Paul Paroutis from the Imaging Facility at the Hospital for Sick Children. I would also like to extend my gratitude to Dr. Maree Webster, the Stanley Research Laboratory and Brain

Collection and all of the individuals who generously donated their organs for research. Finally, I am extremely thankful for my family and close friends who have continuously supported and encouraged me.

iii

Contributions

Mohamad Abbass and Dr. Albert Wong conceived and planned the study when Mohamad analyzed the anterior cingulate cortex (Abbass, 2014). Myself and Dr. Wong planned the OFC study and made changes to the protocol when needed. Myself, Mohamad Abbass and Dr. Albert

Wong conceived and planned the method validation. Dr. Albert Wong contributed to the interpretation of the results.

Jialun Chen assisted with the staining of the tissue. Mohamad Abbass developed the protocol for ImageJ, which was adapted for the purpose of the OFC study. Anton Semechko developed the MATLAB algorithm, which was used to calculate cortex width, cell density, cell area and distance from pia. Dr. Maree Webster and the Stanley Medical Research Institute provided us with the cortical samples.

Table of Contents

Title Page i

Abstract ii

Acknowledgments iii

Contributions iv

Table of Contents v

List of Tables viii

List of Figures ix

List of Appendices xii

List of Abbreviations xiii

Chapter 1: Introduction and Literature Review 1 1.1 Cerebral Cortex 1 1.1.1 Overview 1 1.1.2 Neurons in the Neocortex 2 1.1.2.1 Pyramidal Neurons 2 1.1.2.2 Non-Pyramidal Neurons 4 1.1.3 Cytoarchitecture 5 1.2 Cortical Development 6 1.2.1 Development of the Human Centreal Nervous System 6 1.2.2 Corticogenesis 7 1.2.3 Cell-Fate Determination 9 1.2.4 Neuron Migration 15 1.2.4.1 Radial Migration 15 1.2.4.2 Tangential Migration 17 1.3 Orbitofrontal Cortex 18 1.3.1 Orbitofrontal Cortex Anatomy 18 1.3.2 Orbitofrontal Cortex Function 22 1.4 Psychiatric Disorders 23

1.4.1 Schizophrenia 23 1.4.1.1 Neurodevelopmental Hypothesis of Schizophrenia 24 1.4.1.2 Neurochemical Pathologies 25 1.4.1.3 Gross Anatomical Pathologies 26 1.4.1.4 Histological Pathologies 28 1.4.1.5 Genetic and Molecular Pathologies 29 1.4.1.6 Orbitofrontal Cortex in Schizophrenia 31 1.4.2 Bipolar Disorder 33 1.4.2.1 Orbitofrontal Cortex in Bipolar Disorder 33 1.4.3 Major Depressive Disorder 34 1.4.3.1 Orbitofrontal Cortex in Major Depressive Disorder 35 1.4.4 Cytoarchitecture of the Orbitofronal Cortex in Psychiatric Disorders 36

Chapter 2: Research Aims and Hypotheses 38

Chapter 3: Methods 41 3.1 Tissue Samples 41 3.2 Immunohistochemistry 42 3.3 Image Analysis 43 3.3.1 Microscopy - Zeiss Epifluorescence Microscope 43 3.3.2 Regional and Laminar Delineation 43 3.3.3 Automatic Cell Segmentation 46 3.3.4 Automatic Data Generation 47 3.3.4.1 Cortex Width 49 3.3.4.2 Cell Density 49 3.3.4.3 Cell Area 51 3.3.4.4 Distance From Pia 51 3.4 Statistical Analysis 52 3.5 Method Validation 53 3.5.1 Tissue Samples 53 3.5.2 Staining 53 3.5.2.1 Cresyl Violet (Nissl) Staining 53 3.5.2.2 Anti-CUX2 and Anti-NeuN 54

3.5.2.3 Anti-ZNF312 and Anti-NeuN 54 3.5.3 Image Analysis 55 3.5.4 Statistical Analysis 56

Chapter 4: Results 57 4.1 Method Validation 57 4.1.1 Automated Versus Manual Counts 57 4.1.2 Nissl Stain 58 4.1.3 Neuronal Nuclear Antigen (NeuN) 59 4.2 Pearson's Correlation 61 4.3 Brodmann Area 47l 62 4.4 Brodmann Area 47m 68 4.5 Entire Orbitofrontal Cortex 73

Chapter 5: Discussion 78 5.1 Labeled Cell Population 78 5.2 Method Validation 81 5.3 Summary of Findings 83 5.4 Schizophrenia 85 5.5 Bipolar Disorder 88 5.6 Major Depressive Disorder 91 5.7 Significance 95 5.8 Limitations 96 5.9 Conclusion 98

Chapter 6: Future Directions 101

Chapter 7: References 104

Chapter 8: Appendix 126

vii

List of Tables

Table 1. Embryonic zones of the human cerebral cortex. 8

Table 2. Demographic information of the Neuropathology Consortium 41 of the Stanley Medical Research Institute.

Table 3. Automated counts of DAPI, CUX2, ZNF312, CUX2+ve/ 57 DAPI+ve and ZNF312+ve/DAPI+ve cells are similar to manual counts.

Table 4. Cortical thickness measurements are similar between Nissl 58 and immunohistochemically stained slides.

Table 5. Summary of findings. 81

viii

List of Figures

Figure 1. Corticogenesis in the human brain. 10

Figure 2. Neocortical projection neurons are generated in an “inside-out” 12 fashion by progenitor cells.

Figure 3. Three sulcogyral patterns in the orbitofrontal cortex of the 20 human brain.

Figure 4. BAs 47 and 11 in the OFC based on gross anatomical landmarks. 21

Figure 5. A model of the orbitofrontal cortex function. 23

Figure 6. Gray-scale images of CUX2, ZNF312 and DAPI and an 43 overlapped artificially coloured image.

Figure 7. Regional and laminar delineation of the orbitofrontal cortex. 44

Figure 8. Cortical layers delineated based on cytoarchitectonic criteria. 45

Figure 9. Automatic cell segmentation using ImageJ for (A) CUX2, 47 (B) ZNF312, and (C) DAPI.

Figure 10. The five images that are input into MATLAB and the resulting 48 image. Figure 11. Delaunay Triangulation. 52

Figure 12. Nissl stain of the orbitofrontal cortex. 56

Figure 13. The percentage of CUX2 cells that are co-stained with NeuN 59 is 82.82%.

Figure 14. The percentage of ZNF312 cells that are co-stained with NeuN 60 is 72.34%

Figure 15. Trend for an increased thickness of layer V in SCZ in BA47l. 63

Figure 16. Decreased relative density of ZNF312 cells in layer V in BP, 64 MDD and SCZ in BA47l.

Figure 17. Decreased relative density of CUX2 cells in layer V in BP, 65 and MDD in BA47l.

Figure 18. Trend for an increase in the relative density of CUX2+ve/ 66 ZNF312-ve cells in layer I and decrease in layer IV in MDD in BA47l.

Figure 19. No significant differences in ZNF312 cell size in BA47l. 67

Figure 20. No significant differences in relative distance from pia in BA47l. 67

Figure 21. Trend for an increased thickness of layer IV in MDD in BA47m. 68

Figure 22. Trend for a decrease in the relative density of ZNF312 cells 69 in layer VI in MDD in BA47m.

Figure 23. Trend for an increase in the absolute density of CUX2 cells in 70 layer I and II and decrease in the relative density in layer V and VI in MDD in BA47m.

Figure 24. Trend for an increase in the absolute and relative density of 71 CUX2+ve/ZNF312-ve cells in layer I in MDD in BA47m.

Figure 25. No significant differences in ZNF312 cell size in BA47m. 72

Figure 26. No significant differences in relative distance from pia in BA47m. 72

Figure 27. Trend for an increased thickness of layer IV in SCZ in the 73 entire OFC.

Figure 28. Trend for a decrease in the relative density of ZNF312 cells in 74 layer IV in SCZ, layer V and VI in MDD and layer VI in BP in the entire OFC.

Figure 29. Decreased relative density of CUX2 cells in layer VI in BP in the 75 entire OFC.

Figure 30. Trend for an increase in the absolute and relative density of 76 CUX2+ve/ZNF312-ve cells in layer I in MDD in the entire OFC.

Figure 31. No significant differences in ZNF312 cell size in the entire OFC. 77

Figure 32. No significant differences in relative distance from pia in the 77 entire OFC.

List of Appendices

Appendix 1. Derivation of Equation 3. 126

xii

List of Abbreviations

A: Area

BA: Brodmann Area

BA47l: Brodmann Area 47 lateral

BA47m: Brodmann Area 47 medial

BAD: Bcl-2 Associated Death Promoter

BAX: Bcl-2-Associated X Protein

BP: Bipolar Disorder

CAL: Calretinin

CB: Calbindin

CFPN: Corticofugal cOFC: Caudal Orbitofrontal Cortex

CP: Cortical Plate

CPN: Callosal Projection Neurons

CR: Cajal-Retzius Cells

CThPN: Corticothalmic Projection Neurons

CUX2: Cut-like homeobox 2

D1: Dopamine Type I Receptors

D2: Dopamine Type II Receptors

DAPI: 4’,6’-Diamidino-2-Phenylindole

DISC1: Disrupted-in Schizophrenia 1 dM: Measured Diameter

DM: Measured Density

DSM: Diagnostic and Statistical Manual of Mental Disorders xiii

dT: True Diameter

DT: True Density

ERBB4: Erb-B2 Receptor Tyrosine Kinase 4

F: Abercrombie’s Correction Factor

FBS: Fetal Bovine Serum

FEZF2: Forebrain Embryonic Zinc Finger Protein 2 (also known as ZNF312)

FFT: Fast Fourier Transformation fMRI: Functional Magnetic Resonance Imaging

GABA: Gamma-Aminobutyric Acid

GAD: Glutamic Acid Decarboxylase

GFAP: Glial Fibrillary Acidic Protein

GN: Granular Neurons

GW: Gestational Weeks

H: Height

IOS: Intermediate Orbital Sulcus

IP: Intermediate Progenitors

IZ: Intermediate Zone

LOS: Lateral Orbital Sulcus

LOSc: Caudal Portion of the Lateral Orbital Sulcus

LOSr: Rostral Portion of the Lateral Orbital Sulcus

MDD: Major Depressive Disorder

MOS: Medial Orbital Sulcus

MOSc: Caudal Portion of the Medial Orbital Sulcus

MOSr: Rostral Portion of the Medial Orbital Sulcus

xiv

MRI: Magnetic Resonance Imaging

MZ: Marginal Zone

NE: Neuroepithelial Cells

NeuN: Neuronal-Specific Nuclear Protein

NRG1: Neuregulin1

OFC: Orbitofrontal Cortex

OLF: Olfactory Sulcus

PBS: Phosphate Buffered Saline

PBS-Triton: Phosphate Buffered Saline with 0.2% Triton-X

PMI: Post-Mortem Interval

PV: Parvalbumin r: Radius

RELN: Reelin

RG: Radial Glial Cells

SCPN: Subcerebral Projection Neurons

SCZ: Schizophrenia

SP: Subplate

SVZ: Subventricular Zone

T: Thickness

TOS: Transverse Orbital Sulcus

WM: White Matter

VZ: Ventricular Zone

ZNF312: Zinc Finger Protein 312 (also known as FEZF2)

Chapter 1: Introduction and Literature Review 1.1 Cerebral Cortex

1.1.1 Overview

A detailed knowledge of the cerebral cortex is required in order to comprehend the biological bases of cognition, emotion and behavior. The cerebral cortex is the outer layer of neural tissue of the cerebrum and consists of two hemispheres, with four lobes in each hemisphere (Baars and Gage, 2010). In brief, the anterior part of the cortex is the frontal lobe, which includes the motor regions of the cortex. Superiorly and posteriorly to the frontal lobe is the parietal lobe, which contains the somatosensory regions of the cortex. The temporal lobe, which is inferior to the parietal lobe and adjacent to the frontal lobe, contains the auditory, olfactory and gustatory regions of the cortex. Lastly, the occipital lobe, which is the posterior region of the cortex, contains the visual cortex.

The cerebral cortex is composed of the archicortex, paleocortex and neocortex (Pandya et al., 2015; Sanides, 1969). The archicortex (hippocampus) and paleocortex (olfactory cortex proper) together known as the allocortex, are phylogenetically older and show regionally highly variable appearances, from a hardly visible single cell band to 10 layers of cells (Zilles, 2004).

This differs from the other major part of the cortex, the neocortex, which is evolutionarily the newest region of the cortex and consists of six fully developed cortical layers (Pandya et al.,

2015). The neocortex, also known as the isocortex, comprises sensory, motor and association areas (Zilles, 2004). Sandines (1969) further sub-categorized the cortex based on two structural steps: the first going from the peri/archicortex to isocortex, referred to as periallocortex, and the second between the latter and the mature isocortex, known as the proisocortex. The

periallocortex can be further subdivided into the peripaleocortex (claustral region) and the periarchicortex (entorhinal, presubicular, retrosplenial and parts of the cingulate gyrus), which transition from the paelocortex and archicortex to the neocortex, respectively. The proisocortex is mostly found in limbic areas and forms the bulk of the cingulate gyrus and insula.

1.1.2 Neurons in the Neocortex

Neurons in the neocortex are morphologically divided into two groups: the pyramidal cells and the non-pyramidal cells. Pyramidal cells are the primary neurons of the neocortex and comprise approximately 80% of the total neuron population (Kageyama and Yamamori, 2013).

The remaining neocortical neurons include a variety of morphological types that share several common features. Kageyama and Yamamori (2013) and Nieuwenhuys et al. (2008) review the neocortical neurons.

1.1.2.1 Pyramidal Neurons

Pyramidal cells are excitatory neurons of the neocortex with a triangular shaped soma, hence the name pyramidal. These cells use glutamate as an excitatory neurotransmitter and project to various subcortical and cortical areas. Additionally, pyramidal cells emit axon collaterals locally and act as local circuit neurons. Typically, a pyramidal cell will extend a single apical dendrite form the apex of the soma towards the cortical surface and have basal dendrites around the soma. The dendrites of pyramidal neurons are densely covered with spines, which are small protrusions of the plasma membrane. It is here where the neuron receives excitatory synaptic input and compartmentalizes postsynaptic response (Hering and Sheng,

2001).

Subpopulations of pyramidal cells are identifiable based on firing characteristics, morphologies and molecular composition (Arlotta et al., 2005; Hevner et al., 2003; Molyneaux et al., 2007; van Aerde and Feldmeyer, 2015). For example, an in vitro intracellular recording study in the rat prefrontal cortex found three main classes of pyramidal cells based on their firing patterns in response to depolarizing current pulses and the characteristics of their action potentials: regular spiking, intrinsic bursting and non-inactivating bursting cells (Dégenètais et al., 2002). Further, the regular spiking cells were subdivided into slow-adapting and fast- adapting type depending on their firing frequency adaptation.

The different pyramidal cells are also commonly characterized based on their cortical projections. In broad terms, associative projection neurons are those whose axons extend within one cerebral hemisphere, commissural projection neurons are those whose axons cross the midline to the contralateral hemisphere and corticofugal projection neurons (CFPN) have axons directed away from the cortex (Greig et al., 2013). Associative projection neurons are present in all cortical layers and consist of short-distance intrahemispheric axons. Commissural projection neurons cross the midline typically through the corpus callosum and are hence called callosal projection neurons (CPN). CPN reside primarily in layer II/III and extend axons in the same functional area of the contralateral hemisphere, allowing for bilateral integration of information.

Corticofugal projection neurons can be subdivided into corticothalmic projection neurons

(CThPN) and subcerebral projection neurons (SCPN), which reside in layer VI and V, respectively. CThPN extend axons to specific thalamic nuclei, whereas SCPN extend to different targets in the brainstem and spinal cord. There exists evidence to suggest that different projection cell types exhibit characteristic molecular expression profiles and that neurons in the same layer may express different transcription factors (Hevner et al., 2003).

1.1.2.2 Non-Pyramidal Neurons

Non-pyramidal cells in the neocortex differ from pyramidal cells in that they are morphologically diverse in both axons and dendrites. Additionally, these cells differ from pyramidal cells in that they lack a conical soma and dominant apical dendrite, and their dendrites bear only few spines or are spine-free. There are two subsets of non-pyramidal neurons: glutamatergic, also known as spiny stellate cells, and GABAergic aspiney cells. Stellate neurons are found in all cortical layers however they are predominantly found in layer IV (Schubert et al.,

2003). The dendrites of these cells radiate from the soma in all directions and branch infrequently, whereas their axon arborization forms a local plexus occupying the same territory as the dendrites.

The vast majority of non-pyramidal neurons use GABA as their primary neurotransmitter, suggesting that most of these cells have an inhibitory function. Approximately a quarter of the GABAergic cortical neurons express one or several neuropeptides, including substance P, vasoactive intestinal polypeptide, cholecystokinin, neuropeptide Y, somatotropin- release-inhibiting factor, corticotropin-releasing factor and tachykinin (Nieuwenhuys et al.,

2008). Several subpopulations of GABAergic cortical neurons appear to be definable based on immunoreactivity for specific neuropeptides. It has also been shown that differential immunoreactivity for the calcium-binding proteins parvalbumin (PV), calbindin (CB) and calretinin (CAL) can be used as a marker for different subpopulations of non-pyramidal neurons

(Fujise et al., 1995; Mikkonen et al., 1997). In brief, the most characteristic neocortical neurons immunoreactive for PV are chandelier cells and large basket cells, for CB are double bouquet cells and for CAL are double bouquet and bipolar cells (DeFelipe et al., 1989; del Rio and

DeFelipe, 1997; Mikkonen et al., 1997).

1.1.3 Cytoarchitecture

Architectonics is the formal study of how the cells and fibers in the cortex are arranged into layers and columns (Pandya et al., 2015). Although several investigators studied the cellular architecture of the cerebral cortex in the 1800s, it was not until the 20th century that the field reached maturity. Some of the most comprehensive maps were those of Korbinian Brodmann. In his early 1900s work, he presented a method of architectonic analysis and proposed the principle of the six-layer cortex. Brodmann divided the human cerebral cortex into more than 40 areas based on differences in the organization of cells. Since function reflects architecture, the different Brodmann areas have to a considerable extent different functions (Rolls, 2016).

Although the cellular organization varies between the areas, the layers in each area may be divided based on the distribution, density and size of the neurons after staining (Creutzfeldt,

1995). According to Creutzfeldt (1995), the principal layers can be characterized by the following features:

- Layer I (the plexiform or molecular layer): a few scattered neurons, many extensions of apical dendrites and horizontally oriented axons, Cajal-Retzius cells in the outer zone and spiny stellate cells in the inner zone

- Layer II (the outer granular layer): predominantly small pyramidal and stellate cells

- Layer III (the outer pyramidal layer): predominantly small and medium sized pyramidal cells, as well as non-pyramidal cells with ascending or descending axons

- Layer IV (the granular layer): different types of stellate and pyramidal cells

- Layer V (the inner pyramidal or ganglionic layer): large pyramidal cells and interneurons

- Layer VI (the multiform or spindle cell layer): a few large pyramidal neurons, many small spindle-like pyramidal and multiform cells.

1.2 Cortical Development

1.2.1 Development of the Human Central Nervous System

The first 8 weeks of human development (embryonic period) can be divided into 23 stages known as the Carnegie stages (O’Rahilly and Müller, 1987). Donkelaar et al (2014) provide an overview of the development of the human central nervous system. During week 1

(stages 2-4) the blastocyst is formed, which contains two distinct cell types: the trophoblast

(precursor of placenta) and the inner cell mass (precursor of the embryo proper) (Flynn, 1991).

During week 2 (stages 5 and 6), implantation of the embryo occurs and during weeks 2 and 3 the three embryonic layers are formed: ectoderm, mesoderm and endoderm (Flynn, 1991).

Stages 7-12 are depicted by the formation of the notochordal process, the beginning of neurulation and finally the closure of the neuronal tube. First, the notochord induces the overlying ectoderm to become the neuroectoderm and eventually the v-shaped neuroplate

(Atkinson, 2013). The first indication of the neural plate in human embryos is around 23 days of development (week 3). At approximately 25 days (stage 9), this v-shaped neural plate becomes deeper and longer, forming the neural groove. The rostral half of the neural groove represents the forebrain and the caudal half represents the hindbrain. The brain first subdivides into the forebrain, midbrain and hindbrain while the neural folds are still unfused. At stage 10, the two subdivisions of the forebrain, the telencephalon and diencephalon, become evident. Around this time, the first indication of the internal ears and developing eye can be recognized. The closure of the neuronal tube begins at the level of the future cervical regions and proceeds both rostrally and caudally (stages 11 and 12). The fusing neuro-ectodermal cells of the neural folds give rise to neural crest cells, which migrate extensively to generate a large diversity of differentiated cell types.

The final 11 Carnegie stages occur over the course of gestational weeks 5-8. The stages include the development of the retinal and lens discs (stage 13), cerebral hemispheres (stage 14 and 15), adenohypophyseal pouch (stage 14), neurohypophyseal pouch (stage 16), cerebellar swellings (stage 17), semicircular ducts (stage 18), olfactory bulbs and choroid plexus of the fourth ventricle (stage 19), choroid plexus of lateral ventricle (stage 20), cortical plate and optic tract (stage 21), olfactory tract (stage 22) and insula, caudate nucleus and putamen (stage 23).

This time period is extremely important, as it is the initial formation of all organ systems and developmental insults at this stage will result in major congenital defects (Flynn, 1991).

1.2.2 Corticogenesis

The developing cerebral wall contains several transient embryonic zones, beginning with the ventricular zone (Table 1). After fusing, the wall of the neural tube consists of a single layer of neuroepithelial cells (Donkelaar et al., 2014). Over time, the layer thickens and acquires the appearance of a pseudostratified epithelium. This germinal neuroepithelium is known as the ventricular zone and this is where cortical neurons are generated. Between gestational weeks 4 and 5, or pre-neurogenesis, neuroepithelial cells divide symmetrically to create new progenitor cells (Budday et al., 2015). Gestational week 5 marks the onset of neurogenesis, where progenitor cells in the ventricular zone switch from symmetric to asymmetric cell division.

During asymmetric division, one daughter cell remains in the ventricular zone as a radial glial cell (see section 1.2.3) and the other becomes a postmitotic neuron or an intermediate progenitor cell (Pontious et al., 2008).

Postmitotic neurons reach the other embryonic zones by migrating from the ventricular zone to their destined location, which is further discussed in section 1.2.2 and 1.2.3. The first postmitotic cells migrate radially out of the neuroepithelium and form the first recognizable

Embryonic Zones Characteristics

Ventricular Zone Composed of dividing neural progenitor cells.

Preplate Transient layer with predecessor neurons and Cajal-Retzius cells. Forms before the appearance of the cortical plate and is later subdivided into the marginal zone and subplate.

Subventricular Zone Acts early in corticogenesis as a secondary progenitor compartment. Later acts as the major source of glial cells. Appears before the emergence of the cortical plate.

Intermediate Zone Layer between the proliferative layers and postmigratory cells above. Contains radially and tangentially migrating cells and a thickening layer of axons that eventually constitutes white matter.

Subplate Transient layer directly below the cortical plate. Essential in orchestrating thalmocortical connectivity.

Cortical Plate Initial condensation of postmitotic cells that will become layers II-VI of the cortex.

Marginal Zone Superficial, cell sparse layer composed largely of Cajal-Retzius neurons. Residual superficial part of the preplate after the appearance of the cortical plate. Later becomes layer I in the mature cortex.

Table 1. Embryonic zones of the human cerebral cortex. Characteristics of the ventricular zone, preplate, subventricular zone, intermediate zone, subplate, cortical plate and marginal zone.

cortical layer, the preplate (de Graaf-Peters and Hadders-Algra, 2006). Next, the intermediate

zone is formed and becomes progressively thicker as more cells are added to it from the germinal

neuroepithelium (Donkelaar et al., 2014). This layer will eventually contain the afferent and

efferent axons of the cortex that make up the white matter (Nadarajah and Parnavelas, 2002).

During later proliferative stages, progenitor cells accumulate in another proliferative layer known

as the subventricular zone (Flynn, 1991). Around gestational week 7, neurons from the

ventricular and subventricular zone begin to migrate up into the true cortical plate, which is now

between the outer part of the preplate (marginal zone) and the lower part (subplate) (Budday et

al., 2015). The earliest-generated neurons of the cerebral cortex are found in the marginal zone and subplate (Nadarajah and Parnavelas, 2002). Cells in the marginal zone, which later becomes layer I in the mature cortex, differentiate into Cajal-Retzius cells and other types of neurons.

Autoradiographic studies have shown that layers II-VI of the cortical plate are formed in an

“inside-out” manner, such that early neurons reside in the deepest layers and later-born neurons must past the existing layers (Angevine and Sidman, 1961; Berry and Rogers, 1965).

From week 9 onwards, neurons reach their destined layers and the neocortex becomes distinguishable. Between gestational weeks 9 and 12, the entire cortex thickens as the neurons reach their final positions (Budday et al., 2015). Between weeks 13 and 15, the ventricular zone becomes thinner and at week 18 the six distinct layers of the neocortex are clearly distinguishable. By week 25-27, the ventricular zone is only a one-cell-thick ependymal layer and the subventricular zone becomes the major source of cortical neurons (Zecevic et al., 2005).

By gestational week 28, the marginal zone has now fully developed into layer I and the intermediate zone transforms into white matter tissue (Budday et al., 2015). Figure 1 depicts the stages of development of the cerebral cortex.

1.2.3 Cell-Fate Determination

Molyneaux et al. (2007) review the neuronal subtype specification in the cerebral cortex

(Figure 2). Within the mature cortex, there exist distinct populations of neurons that are located in different cortical layers and areas, and have unique morphological features and functions.

Over the past few decades, studies have identified basic mechanisms that control general neuronal specification, migration and connectivity during development. Additionally, molecular programmes have been identified that instruct early steps of progenitor specification and define neuronal subtype and layer identity.

Figure 1. Corticogenesis in the human brain. The simplified schematic depicts the development of the Ventricular Zone (VZ), Subventricular Zone (SVZ), Intermediate Zone (IZ), Subplate (SP), Cortical Plate (CP), Marginal Zone (MZ) and White Matter (WM) over gestational weeks (GW) 4 to 28 (Image adapted from Molyneaux et al., 2007).

As specified in section 1.2.2, progenitors residing in the ventricular zone and subventricular zone produce the projection neurons in a tightly controlled manner. At early stages, progenitors are able to give rise to pyramidal neurons across layers II-VI (Frantz and

McConnell, 1996). However, these progenitors become more restricted with time and at the close of neurogenesis, they produce neurons mostly destined for the upper layers (Frantz and

McConnell, 1996). Studies have shown that early cortical progenitors fated to form deep-layer

neurons can generate later born neurons of the upper layer when transplanted into the niche of late progenitors, indicating that environmental factors are important for laminar fate (McConnell and Kaznowski, 1991). In contrast, progenitors of upper layer neurons are less plastic and are restricted to producing upper layer neurons, even when transplanted into younger hosts (Frantz and McConnell, 1996).

There are three types of neurogenic progenitor cells in the developing neocortex: neuroepithelial cells, radial glia and intermediate progenitors (Götz and Huttner, 2005). Initially, there is a single layer of pseudostratified neuroepithelial cells, which undergo both symmetric cell divisions to create more multipotent progenitors and to a lesser extent asymmetric cell division to generate the earliest born neurons (Götz and Huttner, 2005; Molyneaux et al., 2007).

As neurogenesis progresses, these early progenitor cells transform into distinct but related radial glial cells, which exhibit residual neuroepithelial and astroglial properties (Anthony et al., 2004;

Malatesta et al., 2003). A study using time-lapse imaging found that radial glia play a large role in the generation of pyramidal neurons, either directly through mitoses or indirectly through the production of proliferating intermediate progenitors (Noctor et al., 2004). Intermediate progenitors (also known as basal progenitors) are located in the basal ventricular zone in early neurogenesis and later in the subventricular zone upon its formation (Molyneaux et al., 2007).

Mammalian progenitor cells produce different cell types during different stages of development under the influence of multiple signaling pathways (Wen et al., 2009). Although the exact mechanisms of neural stem cell commitment are unknown, it is thought that both extrinsic and intrinsic mechanisms affect the process, including changes in the environment, epigenetic modifications and transcription factor expression patterns. Further evidence suggests that cell- and layer-specific transcription factors play a role in migration, the final laminar

Figure 2. Neocortical projection neurons are generated in an “inside-out” fashion by progenitor cells. Neuroepithelial cells (NE) and radial glia (RG) produce projection neurons in the ventricular zone. RG also generate intermediate progenitors (IP), which reside in the subventricular zone. Cajal-Retzius cells (CR) migrate into layer I and the other projection neurons reside in layers II-VI. Projection neurons include layer IV granular neurons (GN), callosal projection neurons (CPN), corticothalmic projection neurons (CThPN), subcerebral projection neurons (SCPN) and subplate projection neurons (SPN). For more information on the cortical layers, see Figure 1. (Image adapted from Molyneaux et al., 2007 and Greig et al., 2013).

position of cortical neurons, molecular identity and axonal connectivity (Kwan et al., 2012).

Additionally, several studies have identified genes encoding for transcription factors that are selectively expressed in particular layers or subtypes of cortical projection neurons. Two transcription factors thought to be of particular importance in layer differentiation are FEFZ2 and

CUX2.

The zinc-finger transcription factor FEZF2 (also known as ZNF312) is thought to control, in part, the laminar position of early-born neurons (Greig et al., 2013; Kwan et al., 2012). As mentioned in section 1.1.3.1, SCPN and CThPN are subtypes of CFPN that are generated early in neurogenesis and reside in deep layers of the neocortex (V and VI, respectively) (Greig et al.,

2013). FEZF2 functions in the specification of CFPN identity and, more specifically, is crucial for specification of SCPN. Fezf2 is expressed at high levels by SCPN and in Fezf2-null mice, the large pyramidal cells that normally define layer V are absent. Additionally, in these mutant mice, corticospinal motor neurons and other subcerebral projection neurons are absent whereas in overexpression models, there is an excess production of these neurons (Chen et al., 2005;

Molyneaux et al., 2005). These studies suggest that Fezf2 plays a role in the differentiation of layer V SCPN. In regards to CThPN, these cells express Fezf2 at lower levels. However, even low levels of expression appear to be essential for differentiation. The same studies using Fezf2- mutant mice found that CThPN were disorganized and a number of CThPN-specific genes failed to be expressed (Chen et al., 2005; Molyneaux et al., 2005).

In contrast, cut-like homeobox 2 (Cux2) is expressed in the ventricular zone and subventricular zone when neurons destined for upper layers are being formed (Nieto et al., 2004;

Zimmer et al., 2004). Additionally, a later study found that CUX2 specifically is expressed by

CPN within the superficial layers, suggesting that this transcription factor may play a role in the demarcation of this subtype of projection neurons (Molyneaux et al., 2009). There exists

evidence that CUX2 controls the proliferation of intermediate precursors in the subventricular zone and studies in CUX2 deficient mice have found upper layer neurons in excess (Cubelos et al., 2008). Additionally, CUX2 is thought to regulate dendritic branching, spine morphology and synapses in upper layer neurons (Cubelos et al., 2010). In CUX2 deficient mice, abnormal dendrites and synapses correlated with reduced synaptic function and defects in working memory.

Over the past few years, there has been controversy in regards to whether progenitor cells are lineage-committed or whether they generate diverse subtypes. More specifically, recent studies are interested in the fate of CUX2+ radial glial cells and FEZF2+ radial glial cells. A study based upon genetic lineage-tracing experiments found that radial glial cells labeled by

CUX2 exclusively produce CPN that reside in the upper layers and that this was independent of niche or birthdate (Franco et al., 2012). In contrast, a later study utilizing in vivo genetic fate mapping demonstrated that Fefz2-expressing radial glial cells existed throughout corticogenesis and that they generated different projection neurons residing in layers II-VI as well as glial cells

(Guo et al., 2013). In regards to CUX2+ radial glial cells, the authors found that these progenitors were able to generate both deep- and upper- layer neurons, which differs from the findings by Franco et al. (2012). However, they concluded that these results do not necessarily refute the findings by Franco et al. (2012), but that laminar-fate restricted RGCs cannot be identified by CUX2 expression alone. This same group of researchers expanded the work and found that most, if not all, CUX2+ and FEZF2+ radial glial cells generated diverse projection neuron subtypes, further suggesting that these progenitors are multipotent (Eckler et al., 2015).

Further research is required in order to expand the knowledge of progenitor cell heterogeneity and fate, which will aid in uncovering the mechanisms that generate diversity in projection neurons.

CUX2 and ZNF312 have been largely characterized in rodents, however few studies have assessed their expression in human tissue. To our knowledge, there has been one study that analyzed CUX2 and ZNF312 mRNA expression in the human prefrontal cortex (Arion et al.,

2007). Using DNA microarray transcriptome profiling, they found CUX2 and ZNF312 to be the genes with the most changed expression in either direction. They then performed in situ hybridization and found that CUX2 showed a robust signal in the supragranular cortical layers, whereas ZNF312 showed a strong and uniform layering in the infragranular layers. These findings are similar to those from animal studies, which suggest that CUX2 and ZNF312 label upper and lower layer neurons, respectively. No other layer-specific markers, to our knowledge, have been extensively studied in both the rodent and human prefrontal cortex.

1.2.4 Neuron Migration

Neurons utilize two distinct modes of migration to reach their destinations in the cortex: radial and tangential migration (Nadarajah and Parnavelas, 2002). Radial migration is the principal mode of migration in the developing cerebral cortex and involves the movement of neurons orthogonal to the surface of the brain. In contrast, tangential migration is when neurons move parallel to the surface of the brain and often transgress regions of the brain. In general, neurons that become pyramidal or glutamatergic neurons migrate radially and GABA-containing interneurons migrate tangentially (Ghashghaei et al., 2007).

1.2.4.1 Radial Migration

Neurons destined to migrate radially adopt two distinct modes of movement depending on the stage of corticogenesis. Somal translocation, which occurs early in development, is when neurons extend a long, basal process from the ventricular zone to the pial surface and then

shorten the basal process by rapid nucleokinesis (Ghashghaei et al., 2007). In contrast, cells that adopt glial-guided locomotion have shorter radial processes that are not attached to the pial surface (Nadarajah and Parnavelas, 2002). As a result, these cells have a slower saltatory pattern of locomotion, in which short bursts of movements are interspersed with stationary phases. A study using time-lapse imaging found evidence for these two distinct forms of cell movement

(Nadarajah et al., 2001). The authors also found that the two subtypes of radial migration were not cell-type specific, such that some cells may move by translocation only, whereas other locomoting cells will translocate when their leading processes reach the marginal zone.

Three steps are required for the movement of any cell: extension of the leading process, nucleokinesis (movement of the nucleus into the leading process) and the retraction of the trailing process (Nadarajah and Parnavelas, 2002). During somal translocation, the soma advances outwards, the pia-directed radial process becomes thicker and shorter while the terminal end remains attached to the pial surface (Nadarajah et al., 2001). In contrast, the cell movement of neurons guided by radial glial is not predetermined, given that there is no pia- connected basal process. These neurons explore the microenvironment to recognize and rely on migratory substrates as well as stop signals to cue the end of migration (Nadarajah and

Parnavelas, 2002). Additionally, neurons that require glial guidance depend largely on microfilament networks and microtubule-dependent mechanisms. Studies have found that disruption of actin filaments inhibits migration completely, suggesting an important role of actin subunit assembly in neuronal migration (Rivas and Hatten, 1995). Both types of cell movements, somal translocation and glial-guided movement, are essential in the early development of the cerebral cortex (Nadarajah et al., 2001).

1.2.4.2 Tangential Migration

Although most cortical neurons migrate predominantly by radial migration, there exists evidence for non-radial routes of migration. Previous studies found tangentially dispersed

GABA-cells in the cortex and at the time, it was assumed that these cells arose in the ventricular zone along with the cortical pyramidal neurons (Nadarajah and Parnavelas, 2002; Tan et al.,

1998). Anderson et al. (1997) were the first to show that a subpopulation of neocortical interneurons originates within the subcortical telencephalon. They were able to show that the number of GABA-expressing cells in the neocortex is significantly reduced when the neocortex is separated from the subcortical telencephalon. During the formation of the forebrain, the primordium is divided into a dorsoventral domain and an anteroposterior domain (Nadarajah and

Parnavelas, 2002). The cerebral cortex arises from the dorsal telencephalic proliferative zone, whereas the ventral telencephalon contains two proliferating cell masses, the lateral and medical ganglionic eminences. Studies have shown that most of the tangentially migrating cells arise from the medial ganglionic eminence, the primordium of the globus pallidus. Additionally,

GABA neurons derived from the subpallial telencephalon populate all regions of the cortex, including the neocortex, piriform cortex and hippocampus (Lavdas et al., 1999; Marin et al.,

2001; Pleasure et al., 2000). There also exists evidence that the subpallial telencephalon generates oligodendrocytes that tangentially migrate to the cortex (Olivier et al., 2001).

The characteristics of the tangentially migrating cells appears to be related to the place and timing of their production (Marín and Rubenstein, 2001). During the early stages of telencephalic development, the medial ganglionic eminence and the anterior entopeduncular area are the primary sources of tangentially migrating cells (Anderson et al., 2001; Lavdas et al.,

1999; Marin et al., 2001). These cells migrate superficially to the developing striatum and invade

the cortical marginal zone and subplate. At mid-embryonic stages, the medial ganglionic eminence is the principal source of cells tangentially migrating into the cortex (Marín and

Rubenstein, 2001). These cells migrate either deep or superficially to the developing striatum and populate the subventricular zone, the lower intermediate zone, the subplate and extend into the cortical plate (Anderson et al., 2001; Lavdas et al., 1999; Marin et al., 2001). During the later stages of telencephalic development, tangentially migrating cells derive from both the lateral ganglionic eminence and medial ganglionic eminence (Anderson et al., 2001).

1.3 Orbitofrontal Cortex

1.3.1 Orbitofrontal Cortex Anatomy

The orbitofrontal cortex (OFC) is a prefrontal cortex region situated on the ventral surface of the frontal lobes of the brain (Kringelbach, 2005). It is defined as the part of the prefrontal cortex that receives projections from the magnocellular medial nucleus of the mediodorsal thalamus, differing from the other areas of the prefrontal cortex that receive projections from other parts of the mediodorsal thalamus (Fuster, 1997). Over the years, several researchers have mapped the OFC region, which has resulted in differences in the position, extent and nomenclature of the areas (Beck, 1949; Brodmann, 1914; Hof et al., 1995; Öngür et al., 2003; Petrides and Pandya, 2002; von Economo and Koskinas, 1925). To overcome this issue, Uylings et al. (2010) defined a set of cytoarchitectonic criteria for the delineation of the individual OFC areas. To do so, they obtained 32 human post-mortem brain tissues and stained them with a modification of Gallyas’ method in order to visualize the neuronal cell bodies

(Uylings et al., 1999).

Similar to previous studies, Uylings et al. (2010) distinguished the following main sulci: the olfactory sulcus (OLF), medial orbital sulcus (MOS), transverse orbital sulcus (TOS) and

lateral orbital sulcus (LOS). In many cases, the MOS, LOS and TOS form an H-like pattern on the basal surface of the brain. The gyri in the OFC include the lateral orbital gyrus (lateral to the

LOS), the medial orbital gryus (medial to the MOS), and the anterior and posterior orbital gyrus, which are located anterior and posterior to the TOS, respectively. Some cases may have an intermediate orbital sulcus (IOS) situated between the medial orbital sulcus and lateral orbital sulcus, anterior to the transverse orbital sulcus. The IOS is variable between individuals, such that most cases have one whereas others may have two or none (Chiavaras and Petrides, 2000;

Sarkar et al., 2016).

Three major types of sulcal patterns have been identified in the OFC based on the arrangement of the main sulci (Figure 3) (Chiavaras and Petrides, 2000). Type I is the most common pattern formed, where the rostral portion of the lateral orbital sulcus (LOSr) is connected with the caudal portion of the lateral orbital sulcus (LOSc) (Figure 3A). Additionally, the portions of the medial orbital sulcus are distinct, such that there is a clearly distinguishable rostral portion of the medial orbital sulcus (MOSr) and caudal portion of the medial orbital sulcus (MOSc). There exist slight variations in each type, whereby in some cases the MOSr is a free lying sulcus whereas in other cases it emerges from the IOS anteromedially. Type II is considered the classic “H-pattern”, formed by the joining of the MOS, LOS and TOS. This pattern is observed in 30% of the cases, making it the second most common type of orbitofrontal sulcal pattern in the human brain (Figure 3B). In Type III, the MOSr and LOSr are separated from the TOS and their caudal portions (Figure 3C). Type III is the most rare, found in approximately 14% of cases. When the intermediate orbital sulcus (or sulci) is present, in any of the three types, it is highly variable, such that in some cases it branches anteriorly into a medial and lateral branch and in other cases it is situated anterior to the TOS as a free sulcus. It may also emerge from either the MOSr or the LOSr to form a Y-shape.

Figure 3. Three sulcogyral patterns in the orbitofrontal cortex of the human brain. Main sulci of the orbitofrontal cortex include the olfactory sulcus (OLF), rostral and caudal portion of the medial orbital sulcus (MOSr and MOSc, respectively), rostral and caudal portion of the lateral orbital sulcus (LOSr and LOSc, respectively) and the transverse orbital sulcus (TOS) (Image adapted from Chiavaras and Petrides, 2000).

Uylings et al. (2010) adopted a nomenclature in which each cortical area and subarea was

indicated by a Brodmann area (BA) number and number with a suffix, respectively. The two

main cytoarchitectonic areas include BA47, located on the orbital surface of the hemisphere, and

BA11, located on the ventromedial edge of the orbital surface. BA47 is subdivided into medial

(47m) and lateral (47l) areas, with the former being located on the medial, anterior and posterior

orbital gyri and the latter located on the lateral orbital gyrus (Figure 4). 47m can be further

subdivided into 47m1, 47m2, and 47m3 and 47l is subdivided into 47l1 and 47l2, giving rise to 5

distinct subareas (Uylings et al., 2010).

Figure 4. BAs 47 and 11 in the OFC based on gross anatomical landmarks. LOS = lateral orbital sulcus, MOS = medial orbital sulcus, OLF = olfactory sulcus. (Image adapted from Uylings et al., 2010).

The cytoarchitectonic criteria were based on the following three categories: (1) granularity of layer IV, (2) cell density and relative soma sizes in the different layers, and (3) absolute and relative thickness of cortical layers. BA47 and BA11 possess common cytoarchitectonic features that distinguish them from the surrounding areas, including a narrow cortex, a relatively thin and discernible layer IV, a narrow layer III, wide infragranular layers and overall small size of neurons. In comparison to BA11, BA47 has larger neuronal cell bodies, more differentiated layers III and V and a clearly delineated granular layer IV. The medial subdivision of BA47 differs from the lateral area due the presence of two darkly stained horizontal bands in infragranular layer V and VI.

1.3.2 Orbitofrontal Cortex Function

The OFC plays a role in sensory integration, the modulation of autonomic reactions, participation in learning and prediction, and decision making for emotional and reward-related behaviours (Kringelbach, 2005). A large meta-analysis of 87 neuroimaging studies found two distinct trends of neural distinction (Kringelbach and Rolls, 2004). The first is a medial-lateral distinction, where the medial OFC is related to monitoring, learning and memory of the reward value of reinforcers and the lateral OFC is related to the evaluation of punishers, leading to changes in ongoing behavior. The second distinction is a posterior-anterior one, with simple reinforcers, such as taste or pain, represented in the posterior OFC and more complex or abstract reinforcers, such as monetary gain or loss, represented in the anterior OFC. The link between the orbitofrontal cortex with the sensory and rewarding properties of reinforcers suggests that this area may play a role in emotional disorders, such as depression.

To summarize these findings, Kringelbach (2005) proposed a bottom up model of OFC function (Figure 5). In brief, sensory information is conveyed to brain structures in the posterior parts of the orbitofrontal cortex for multimodal integration. The orbitofrontal cortex receives input from the five classic sensory modalities, as well as visceral sensory information, making this area one of the most polymodal regions in the entire cortical mantle (Carmichael and Price,

1995). Reward value is assigned in the anterior OFC and then the information can be used to influence subsequent behavior (lateral anterior OFC, with connections to anterior cingulate cortex), stored for learning and memory (medial anterior OFC) or made available for subjective hedonic experience (mid-anterior OFC).

Figure 5. A model of the orbitofrontal cortex function. Two distinct trends of neural distinction: a posterior-anterior distinction and a mediolateral distinction (Image adapted from Kringelbach, 2005).

1.4 Psychiatric Disorders

1.4.1 Schizophrenia

Schizophrenia is a chronic psychiatric disorder affecting individual’s thoughts, feelings and behavior. Patients usually experience a combination of positive symptoms, such as hallucinations and delusions, negative symptoms, such as flat affect and social withdrawal, and cognitive symptoms, such as disorganized speech and problems with concentration (American

Psychiatric Association, 2013). The lifetime prevalence of schizophrenia is approximately 0.3%-

0.7%, although this can vary by race, gender, country, and geographic origin (McGrath et al.,

2008). Psychotic features of schizophrenia typically emerge between the late teens and mid- thirties, and the incidence of schizophrenia is generally higher in men (Ochoa et al., 2012). There is a strong genetic risk factor for the development of schizophrenia and twin studies indicate that the heritability is estimated to be 0.81 (Sullivan et al., 2012). A number of environmental risk factors have also been identified, including complications of pregnancy and birth, season of birth, use of illicit drugs, migrant status, growing up in an urban setting and advanced paternal age (McDonald and Murray, 2000; McGrath and Murray, 2011).

Individuals diagnosed with schizophrenia typically have a complex clinical presentation, which is accompanied by an equally complex etiology and pathology (Green and Glausier,

2016). As previously mentioned, it is likely a combination of genetic, environmental and developmental risk factors that produce the wide range of symptoms seen in schizophrenia. In some cases, it may be that different etiologies can result in common clinical representations. A recent model to explain this phenomenon, where different upstream etiologies can lead to similar core pathologies, has been called equifinal (Green and Glausier, 2016). However, given the clinical heterogeneity of the disorder, it is even more likely that there are heterogeneous neurobiological presentations. This poses challenges when studying the neuropathology of schizophrenia.

1.4.1.1 Neurodevelopmental Hypothesis of Schizophrenia

The etiology of schizophrenia has long been debated and it has been unclear whether it is a neurodegenerative disorder, neurodevelopmental disorder or a combination of both.

Schizophrenia was originally regarded as a neurodegenerative disorder involving the loss of

structure or function of nerve cells (Przedborski et al., 2003). However, the emphasis has shifted away from this conception, with evidence for usual rates of neurodegenerative pathological features in schizophrenia, including neurofibrillary tangles, plaques, astrocytes and microglia

(Arnold et al., 1998). In addition, a meta-analysis found that progressive dementia is not more common in schizophrenia than in age-matched controls (Baldessarini et al., 1997). The neurodevelopmental theory was first proposed by Thomas Clouston who at the time called it

“developmental insanity” (O’Connell et al., 1997). Advancing technology allowed for more research in the area of neurobiology of schizophrenia and emerging evidence was in support of the neurodevelopmental theory (Gupta and Kulhara, 2010).

The neurodevelopmental model suggests that early insults as late as first trimester or early second trimester can lead to the activation of pathologic neural circuits during adolescence or young adulthood, leading to the emergence of schizophrenia (Fatemi and Folsom, 2009). The four main areas of evidence include: 1) brain pathology, including enlargement of the ventricles, changes in gray and white matter and abnormal laminar organization, 2) genetics, such as changes in the expression of proteins involved in migration, cell proliferation, axonal outgrowth, synaptogenesis and apoptosis, 3) environmental factors, including increased frequency of obstetric complications and increased rates of schizophrenic births due to prenatal viral or bacterial infections, and 4) gene-environmental interactions, such as an overrepresentation of pathogen-related genes among schizophrenia candidate genes.

1.4.1.2 Neurochemical Pathologies

Several neurotransmitters have been proposed to be key neurochemical factors associated with schizophrenia. The hypotheses that have generated the most interest have been changes in dopamine, serotonin, noradrenaline, GABA and neuropeptides (Owen and Crow, 1987).

Evidence comes from studies that have found amelioration of symptoms in schizophrenia by chemical means as well as symptoms provoked in healthy subjects. However, many of these findings have been inconsistent or have failed to be pathognomonic (Kleinman et al., 1988).

The dopamine hypothesis of schizophrenia has been one of the most enduring (Howes and Kapur, 2009). The original dopamine hypothesis suggested that schizophrenic symptoms were due to hyperactive dopamine transmission (Brisch et al., 2014). However, this hypothesis has been revised in recent years to propose hyperactive dopamine transmission in the mesolimbic areas and hypoactive transmission in the prefrontal cortex in schizophrenia (da Silva Alves et al.,

2008). One of the most reproducible neuropathological findings in schizophrenia is an increased number of dopamine type II (D2) receptors in the caudate and putamen, which are structures of the striatum (Kleinman et al., 1988). It has been further hypothesized that positive symptoms of schizophrenia are associated with increased D2 receptor activation, whereas the negative and cognitive symptoms are associated with decreased D1 stimulation (Brisch et al., 2014; Shen et al., 2012). The primary target of many typical and atypical antipsychotic drugs is antagonism at striatal D2 receptors (Shen et al., 2012).

1.4.1.3 Gross Anatomical Pathologies

There exist several lines of evidence to suggest that schizophrenia is associated with brain structural abnormalities, including decreased brain volume (Jaaro-Peled et al., 2010). A large meta-analysis with over 18,000 subjects found that intracranial and total brain volume was significantly decreased in medicated schizophrenia patients compared to controls (Haijma et al.,

2013). The study concluded that brain loss in schizophrenia is likely related to neurodevelopmental processes as well as illness progression. In addition to decreased volume of the

whole brain, there also exists evidence for decreased volume in the temporal lobe, hippocampus, amygdala, thalamus, anterior cingulate and basal ganglia (Ellison-Wright et al., 2008; Steen et al., 2006; Vita et al., 2006; Wright et al., 2000). However, other studies have found no differences, including a meta-analysis that found no changes in the volume of the temporal lobe and amygdala (Vita et al., 2006).

Other studies in schizophrenia have found reductions in the volume of the whole brain associated with increases in the volume of the lateral and third ventricular volume (Steen et al.,

2006; Vita et al., 2006). Similar studies have also found progressive ventricular enlargement after illness onset greater than in controls (Kempton et al., 2010). Additionally, ventricular dimensions in first-episode schizophrenia were found to be similar to those obtained from meta- analyses of chronic patients (Lawrie and Abukmeil, 1998; Van Horn and McManus, 1992;

Wright et al., 2000). These findings strongly suggest that ventricular enlargement is associated with schizophrenia.

There also exists evidence to suggest that schizophrenia is associated with changed asymmetry and decreased white matter. In regards to anatomical asymmetry, a meta-analysis found that asymmetry of the planum temporale and the Sylvian fissure was significantly decreased in schizophrenia patients compared to controls (Sommer et al., 2001). Not surprisingly, patients with schizophrenia also have a decrease in functional asymmetry (Ribolsi et al., 2014). Finally, reductions in white matter have been observed as well as abnormalities in white matter tracts in the prefrontal and temporal cortices, cingulum and corpus callosum

(KubickiI et al., 2007; Wright et al., 2000). Additionally, decreases in the volume of white matter were in line with reductions in the absolute whole brain volume (Wright et al., 2000). Evidently, there are a large number of anatomical pathologies associated with schizophrenia; however it

remains unclear whether they are as a result of the disorder or due to secondary effects, such as drug treatment.

1.4.1.4 Histological Pathologies

One of the more consistent findings in terms of brain pathology in schizophrenia is a modest and likely widespread reduction in neuronal size (Bakhshi and Chance, 2015). Early studies found reductions in neuronal size in the hippocampus and prefrontal cortex as well as the cerebellar Purjinke cells (Arnold et al., 1995; Benes et al., 1991; Rajkowska et al., 1998; Zaidel et al., 1997). A later study, which used design-based stereology, found that somal size of deep layer III pyramidal neurons in the prefrontal cortex in subjects with schizophrenia was significantly decreased compared to controls (Pierri et al., 2001). Similarly, another study found significant reductions in somal volume of pyramidal cells in deep layer III in BA41 in schizophrenia compared to controls, however not in layer V of BA42 (Sweet et al., 2004). In contrast, other studies have been unable to detect differences in the size of hippocampal pyramidal neurons in schizophrenia (Christison et al., 1989; Highley et al., 2003). These studies suggest that somal size reduction may not be as widespread as anticipated. Although there are inconsistencies in the literature, decreased neuronal size is considered a somewhat robust cytoarchitectural feature of schizophrenia (Harrison, 2000).

A reduced neuropil hypothesis has been suggested in schizophrenia, such that the amount of neuropil is decreased and as a result neuronal density is increased (Selemon and Goldman-

Rakic, 1999). Decreases in dendritic spines have been observed in the temporal and frontal lobes of patients with schizophrenia (Garey et al., 1998; Glantz and Lewis, 2000). However, studies of cell density in schizophrenia appear to be more inconsistent than cell size, such that there have been findings of increases, decreases as well as no significant differences in cell density

compared to controls (Bakhshi and Chance, 2015). A study in the prefrontal cortex found a decrease in the density of large pyramidal neurons and an increase in small neurons (Rajkowska et al., 1998). Two other studies found elevated neuronal density in prefrontal areas 9 and 46 in schizophrenia patients (Selemon et al., 1998, 1995). Other studies in the prefrontal cortex have found increased density of interstitial white matter neurons, suggesting that these neurons may have failed to migrate properly during development (Eastwood and Harrison, 2005; Joshi et al.,

2012). In contrast, a later study found no significant differences in the neuronal density in area 9 of the dorsolateral prefrontal cortex in schizophrenia compared to controls (Cotter et al., 2002).

The cell density findings in schizophrenia are inconsistent and it remains unclear as to whether neuronal density is increased, decreased or unchanged in the prefrontal cortex in schizophrenia. Similar to the prefrontal cortex, these findings are inconsistent in other areas of the brain, including the hippocampus, anterior cingulate cortex, temporal lobe, thalamus, basal ganglia and visual cortex (Bakhshi and Chance, 2015; Fornito et al., 2009; Harrison, 2004). It may be possible that inconsistencies in the literature are due to the substantial heterogeneity in the methodology of histological studies.

1.4.1.5 Genetic and Molecular Pathologies

Genetic studies have identified several genes involved in neuronal migration as candidate susceptibility genes for schizophrenia, including DISC1 (Muraki and Tanigaki, 2015). The disrupted-in schizophrenia 1 (DISC1) gene was first discovered in a large Scottish family suffering from major depressive disorder (MDD), bipolar disorder (BD) and schizophrenia

(Millar et al., 2000). DISC1 is a scaffolding protein that interacts with a number of other proteins involved in neuronal migration, neurite outgrowth, spine development and signal transduction

(Brandon et al., 2009; Camargo et al., 2007). Several mouse models expressing DISC1 with an

altered sequence have shown schizophrenia-like phenotype. One study investigated neuron migration using a mouse model with an L100P point mutation (Lee et al., 2011). They performed immunohistochemistry with two separate upper layer protein markers and found that labeled neurons were further away from the pia and extended into the lower layers, as compared to wild- type littermates. L100P mutant mice have also shown reduced brain volume, reduced number of cortical neurons and schizophrenia-like behavioral abnormalities (Clapcote et al., 2007; Lee et al., 2011). These findings provide evidence that Disc1 is involved in neurogenesis and neuron migration.

Reelin (RELN), another migration-related risk gene, is a large glycoprotein synthesized in gamma-aminobutyric acid (GABA) neurons that controls neuronal cell migration and the lamination of the corticolimbic structures during embryonic development (D’Arcangelo et al.,

1995). RELN expression has been shown to be reduced by approximately 50% in the prefrontal cortex, temporal cortex, hippocampus and caudate nucleus of patients with schizophrenia

(Fatemi et al., 2000; Guidotti et al., 2000; Impagnatiello et al., 1998). In addition, the reelin promoter, which is de-methylated in reelin expressing cells, has been found to be hypermethylated in the post-mortem brain tissue of patients with schizophrenia (Abdolmaleky et al., 2005; Grayson et al., 2005; Tochigi et al., 2008). The heterozygote reeler mouse exhibits a

50% downregulation of reelin expression and mimics schizophrenia more accurately than complete knockout mice (Costa et al., 2002). A study using haploinsufficient heterozygous reeler mice found decreased dendritic spine density of pyramidal neurons and increased density of neurons in layers III-IV, but not II (Liu et al., 2001). It may be possible that neurons destined for layer II did not fully reach their destination and were found in lower layers.

Neuregulins are epidermal growth factor-like proteins that play a role in neuronal development, cell migration and activity in the mature central nervous system (Muraki and

Tanigaki, 2015). Specifically, Neuregulin1 (NRG1) regulates migration of excitatory glutamatergic neurons and GABA-producing interneurons in the embryonic cortex (Muraki and

Tanigaki, 2015). NRG1 and its receptor Erb-B2 Receptor Tyrosine Kinase 4 (ERBB4) are strong candidate genes for schizophrenia (Corfas et al., 2004). Additionally, post-mortem studies have reported both increased and decreased NRG1/ERBB4 signaling in schizophrenia (Joshi et al.,

2014; Law et al., 2006; Silberberg et al., 2006; Weickert et al., 2012). Other studies have shown that NRG1-induced cell migration is significantly decreased in patients with schizophrenia as compared to controls (Sei et al., 2007). These findings suggest that NRG1-mediated neural migration may be abnormal during development in schizophrenia.

1.4.1.6 Orbitofrontal Cortex in Schizophrenia

Evidence for OFC pathology in schizophrenia has been inconsistent and appears to be more debated than the affective disorders. A review article on 15 neuroimaging studies found that the majority of studies reported volume reductions in the OFC in schizophrenia (Bellani et al., 2010). However, at least five neuroimaging studies found no difference in the OFC volumes as compared to controls or even an increase (Baaré et al., 1999; Rupp et al., 2005; Shad et al.,

2006; Szeszko et al., 2000; Yamasue et al., 2004).

Other studies have found alterations in the sulcogyral pattern in schizophrenia. One of the first studies to evaluate the sulcogyral pattern in schizophrenia found that the distribution of types differed between schizophrenics and controls (Nakamura et al., 2007). Specifically, they found fewer cases of Type I, whereas Type III, which is rare in the healthy population, was increased in schizophrenia. Furthermore, they found correlations between Type III expression

and poorer socioeconomic status, poorer cognitive function, severity of symptoms and impulsivity as compared to patients without Type III expression. Another more recent study in a large cohort of individuals at ultra-high risk for psychosis found that individuals who later transitioned to psychosis had a decreased expression of the Type I OFC in the right hemisphere compared to controls (Lavoie et al., 2014). These findings offer further evidence for the neurodevelopmental hypothesis, seeing as they are likely not due to secondary effects of the illness (Nakamura et al., 2007). Additionally, these findings suggest that there may be a protective effect of possessing a Type I OFC (Lavoie et al., 2014).

There exist few cytoarchitectural studies in the orbitofrontal cortex in schizophrenia patients. Garey et al. (2006) utilized 17 post-mortem brain tissues from schizophrenic and control patients, and immunostained for the presence of the kainate receptor (GluR5/6/7) in the

OFC (BA11). Both pyramidal and non-pyramidal cells were labeled, and kainate-positive neurons were found in all cortical layers in both control and schizophrenia cases. The authors found a significant reduction (21%) in the density of kainate receptor-positive neurons in schizophrenia (488 cells/mm2) in comparison to controls (618 cells/mm2) (p=0.033) (Garey et al., 2006). However, Nissl stained tissue showed no significant difference in the total neuronal density between the two groups. These findings suggest that the glutamatergic activity in the

OFC in schizophrenia may be reduced. Another study in the orbitofrontal cortex in schizophrenia found higher GABA neuron density in the white matter, which may be related to GABAergic deficits in the overlying gray matter (Joshi et al., 2012). An increased interstitial white matter neuron density may suggest inappropriate migration of cortical inhibitory interneurons in schizophrenia.

Given these results, as well as the data from neuroimaging studies, there is reason to believe that the OFC may be abnormal in schizophrenia. It may be possible that inconsistencies exist due to different delineations of the OFC, different methods and diverse clinical characteristics of the subjects. This is particularly important in schizophrenia, given the heterogeneity of the disorder. In addition, whenever possible, the OFC areas should be separately analyzed, as it may be possible that only certain sub-regions are affected in schizophrenia.

1.4.2 Bipolar Disorder

Bipolar disorder is an affective disorder characterized by at least one manic or hypomanic episode, which includes symptoms such as inflated self-esteem, pressured speech, fleeting thoughts or ideas and distractibility (American Psychiatric Association, 2013). For a diagnosis of bipolar I disorder, patients must meet the criteria for a manic episode, which may have been preceded or followed by a hypomanic or major depressive episode. For a diagnosis of bipolar II disorder, it is necessary to meet the criteria for a hypomanic episode and a current or past major depressive episode, without having had a manic episode. The pathophysiology of bipolar disorder includes, but is not limited to, neurotransmitter dysregulation (norepinephrine, dopamine and serotonin), brain volume reductions (prefrontal cortex, basal ganglia, hippocampus, anterior cingulate) and functional deficits (amygdala) (Miklowitz and Johnson,

2006).

1.4.2.1 Orbitofrontal Cortex in Bipolar Disorder

Studies have reported reduced OFC volume in bipolar disorder. A study in 2004 performed a whole-brain voxel-based morphometry study on 10 adolescent BP patients and found gray matter deficits in the OFC (Wilke et al., 2004). More recently, a voxel-based

magnetic resonance imaging (MRI) study of patients with bipolar disorder and their healthy siblings found that the left OFC (BA11) was smaller in patients with BP in comparison to controls (Eker et al., 2014). Another study assessed OFC volume using MRI in adolescent patients and found that male BP patients had smaller gray matter volumes in medial, right medial and right lateral OFC in comparison to male controls, whereas female patients had larger gray matter volumes in left, lateral and left lateral OFC as compared to female controls (Najt et al.,

2007). These findings suggested that gender differences may exist in the OFC and this may be involved in the pathophysiology of the illness. However, several investigators have failed to find differences in the total gray matter volume in the OFC or have even noted increases (Konarski et al., 2008)

There also exists evidence to suggest that the activation of the OFC is altered in patients with BP (Blumberg et al., 1999; Drevets, 1999; López-Larson et al., 2002). A functional magnetic resonance imaging (fMRI) study on 18 healthy controls, 18 euthymic, 12 depressed and 12 manic BP patients found that BP patients in all mood states showed reduced activity in the bilateral OFC in comparison to controls (Van der Schot et al., 2010). This suggests that the activity of the OFC may be independent of mood state. However, in contrast, another longitudinal fMRI study found that the OFC was over-activated in patients with BP (Chen et al.,

2010). Finally, a study in 2008 found that BP is associated with attenuated bilateral OFC activation (BA47) and heightened left OFC activation (BA10) (Altshuler et al., 2008). These findings further suggest that the OFC is not a uniform region, both anatomically or functionally.

1.4.3 Major Depressive Disorder

Major depressive disorder is an affective disorder characterized by the presence of sadness, irritable mood, and somatic and cognitive changes that significantly alter the

individual’s capacity to function (American Psychiatric Association, 2013). Risk factors for

MDD include both environmental and genetic factors, with heritability being approximately

40%. Globally, there are more than 350 million people suffering from depression and it is predicted to be the second largest cause of disability by 2020 (Organization, 2008). There are several clinically relevant neurobiological hypotheses of MDD. These include, but are not limited to, deficiency of monoamines (serotonin, norepinephrine), altered hypothalamic- pituitary-adrenal axis (increased corticotropin-releasing hormone), abnormal brain activity

(lateral frontal and temporal cortices, insula, cerebellum), brain volume reductions (prefrontal cortex, anterior cingulate, hippocampus, striatum), reduced GABAergic activity, dysregulation of glutamate system and impaired circadian rhythms (Hasler, 2010).

1.4.3.1 Orbitofrontal Cortex in Major Depressive Disorder

Several studies have found reduced volume of the OFC in major depressive disorder. A study in 2002 used magnetic resonance imagining to measure OFC volume in 15 MDD patients in remission and 20 controls, and found that patients with depression had a statistically smaller medial orbitofrontal cortical volume, without changes in other frontal regions or the whole brain

(Bremner et al., 2002). Another similar study found that MDD patients had smaller gray matter volumes in both right medial and left lateral OFC, in comparison to controls (Lacerda et al.,

2004). A more recent study conducted in 2014 employed whole-brain voxel-based morphometry in a sample of 54 participants and found that the severity of mild depressive symptoms was associated with reduced gray matter volume in the orbitofrontal cortex, anterior cingulate, thalamus, superior temporal gyrus and superior frontal gyrus (Webb et al., 2014). However, according to a review article on 140 published magnetic resonance imaging investigations in mood disorders, several studies have failed to find differences or even found an increase in the

total gray matter volume in the OFC in MDD (Konarski et al., 2008). In terms of neurophysiology, several studies have shown altered activation and decreased blood flow in the

OFC of MDD patients (Drevets, 2007; Liotti et al., 2002).

In regards to morphometric evidence, Rajkowska et al. (1999) were the first to find neuronal and glial cell pathology in the OFC in MDD. Using computer-assisted three- dimensional cell counting, they found a decrease in cortical thickness, neuronal sizes and neuronal and glial densities in the upper cortical layers of the rostral OFC (Rajkowska et al.,

1999). In 2005, the same group studied the OFC of elderly MDD patients and found that the overall packing density was decreased by 30% as compared to controls (Rajkowska et al., 2005).

They also found that pyramidal neuron density was significantly reduced in layers IIIc and V.

However, a more recent study found no changes in glial cell, pyramidal or non-pyramidal neuron density, or in pyramidal or non-pyramidal neuron volume in the OFC in patients with late-life depression (Khundakar et al., 2011). These differential findings may support the notion that the

OFC is not an anatomically uniform region.

1.4.4 Cytoarchitecture of the Orbitofrontal Cortex in Psychiatric Disorders

In 2005, Cotter and colleagues investigated the OFC in 60 post-mortem brains from patients with BP, MDD and schizophrenia, as well as controls. The authors used stereological probes in order to examine the glial and neuronal cell size and density in all cortical layers in the caudal OFC (cOFC). Since stereological cell counting tends to be labor-intensive and time consuming (Schmitz et al., 2014), the authors chose a small region of the OFC that covers the lateral wall of the caudal orbitofrontal sulcus. In terms of results, they found a significant neuronal size reduction in BP in layer I (21%, p = 0.007) and a trend for a reduction in layer V

(20%, p = 0.05) (Cotter et al., 2005). In addition, they observed right hemisphere neuronal size

reductions in MDD in layer III (30%, p < 0.001). No evidence for neuronal or glial pathology was found in schizophrenia. The authors concluded that the cOFC may be specifically altered in affective disorders and not schizophrenia.

Given that the OFC is heterogeneous and varies anatomically and functionally both posteriorly to anteriorly and medially to laterally, it is difficult to assess how these findings translate to the other subdivisions of the OFC. More importantly, it is possible that histological abnormalities may arise from sampling only small sections of the cortex in a limited number of subjects. In the present study, we will be analyzing larger areas of the OFC using post-mortem brain tissue and automated cell counting methods. We hope that novel methods such as ours will standardize cortical analysis and help discover subtle changes in the cytoarchitecture, as well as resolve existing inconsistencies in the literature.

Chapter 2: Research Aims and Hypotheses

Multiple lines of evidence suggest that schizophrenia is a neurodevelopmental disorder associated with impaired neuron migration. Genetic studies have identified several genes involved in neuronal migration as candidate susceptibility genes for schizophrenia, including

DISC1, RELN and NRG1. Knockout mice have shown increased neuron densities in lower layers of the cortex as compared to upper layers, which may be explained by defects in neuronal migration. Other studies have found higher GABA neuron density in the white matter in the

OFC, which may be related to GABAergic deficits in the overlying gray matter. These studies suggest that schizophrenia is associated with neuronal migration defects and that the effects can be seen in brain tissue sections.

Additionally, several studies have found evidence for orbitofrontal pathology in schizophrenia. Several neuroimaging studies have found volume reductions in schizophrenia compared to controls. Other studies have found alterations in the sulcogyral pattern in schizophrenia, including an increased expression of the OFC Type III pattern compared to controls. Finally, although few cytoarchitectural studies have been done in the OFC in schizophrenia, reductions in the density of kainate receptor-positive neurons has been reported.

Another cytoarchitectural study found no significant differences in glial or neuronal cell size or density in schizophrenia compared to controls. However, it is possible that these findings arose from sampling only small sections of the caudal orbitofrontal cortex. The orbitofrontal cortex is an extremely heterogeneous area and it is difficult to assess how these findings would relate to other subareas or the region as a whole.

Given the existing literature, we hypothesize that schizophrenia is associated with neuronal migration defects in the orbitofrontal cortex. More specifically, we hypothesize that neurons in the upper cortical layer will be most affected by deficits in migration, since they have the longest distance to travel. We presume that these neurons will not reach their destination and will therefore be found ectopically in lower cortical layers. We therefore expect to find greater densities of neurons destined for the upper layers in lower layers, as compared to healthy controls. We also expect the findings to be subtle in schizophrenia as compared to a neuronal migration disorder, such as lissencephaly.

In order to test this hypothesis, we will be using two different neuronal molecular markers that are biased towards cells in different layers. Specifically, we will be using antibodies against CUX2 to identify upper layer neurons (II-III) and antibodies against ZNF312 to identify lower layer neurons (V-VI). Animal studies have found CUX2 and ZNF312 to be primarily expressed in callosal projection neurons and subcerebral projection neurons, respectively. Additionally, in situ hybridization studies have found CUX2 and ZNF312 to be differentially expressed in the supragranular and infragranular layers of the human prefrontal cortex. We will also exclude cells labeled with both markers and will examine a population of

CUX2+ve/ZNF312-ve neurons, which are predominately found in the upper cortical layers. In order to test our specific hypothesis, we will be examining the distance of each cell from the pia and we predict that in schizophrenia, the neurons will be farther away from the pia as compared to healthy controls. Additionally, we expect to find a higher density of CUX2+ve and

CUX2+ve/ZNF312-ve neurons in lower layers, since they did not reach their final destination.

We will be examining post-mortem tissue samples of the orbitofrontal cortex from patients with schizophrenia, bipolar disorder and major depressive disorder as well as healthy

controls. Specifically, we will be looking at Brodmann area 11 and 47, with the latter being subdivided into a medial and lateral component. In order to examine this large area of the cortex, we will be using semi-automated methods for segmenting and counting the cells. This includes the manual delineation of the cortical layers using Photoshop CS6, an ImageJ protocol that automatically segments cells and a MATLAB algorithm that calculates cell density and distance from pia. The algorithm also allows us to analyze cortex width and cell area. Although these measures do not directly test the hypothesis, they may enhance the understanding of the involvement of the orbitofrontal cortex in psychiatric disorders. This semi-automated method is advantageous because it provides objective data and allows for the analysis of a greater area of the cortex. This differs from studies using manual methods, where only a small area of the cortex is typically analyzed. We hope that by examining larger areas of the cortex, we will be able to identify subtle changes in the cytoarchitecture and clarify inconsistencies in the literature. Additionally, it will be interesting to assess how the findings from this study compare to other studies on the pathophysiology of the OFC, such as those who have found changes in the affective disorders but not schizophrenia.

Lastly, in order to improve the accuracy of our findings, we will be comparing our automated counting methods to older traditional methods. The delineation of cortical layers will be performed on both Nissl stained tissue and tissue stained with anti-CUX2, anti-ZNF312 and

DAPI, in order to compare measures of cortical thickness. Slides will also be stained with an antibody against NeuN, a commonly used neuronal nuclear antigen, in order to determine the overlap between CUX2 and ZNF312 cells with NeuN. Lastly, automated counts will be compared to manual counts, in order to determine the accuracy of our automated counting method.

Chapter 3: Methods 3.1 Tissue Samples

Brain tissue was obtained from the Neuropathology Consortium of the Stanley Medical

Research Institute. The sample consisted of 60 subjects, with 15 controls, 15 schizophrenia patients, 15 BP patients and 15 MDD patients. Diagnoses were made according to Diagnostic and Statistical Manual of Mental Disorders (DSM) IV criteria (American Psychiatric

Association, 1994). The samples were coronal cross-sections in which Brodmann area 11 and 47 could be identified. Although 60 samples were provided, 6 were eliminated due to damage. The demographic information for the remaining 54 samples can be seen in Table 2. Some samples had damage to specific areas, however other areas could be used in the analyses.

Control MDD BP SCZ Number of Cases 15 11 14 14 Male:Female 9:6 7:4 9:5 8:6 Age 48.1 (29-68) 47.5 (32-65) 43.6 (30-61) 43.4 (25-62) PMI (hours) 23.7 (8-42) 25.6 (7-47) 33.1 (12-62) 33.9 (12-61) pH 6.27 (5.8-6.6) 6.26 (6.0-6.5) 6.16 (5.8-6.5) 6.16 (5.8-6.6) Days in Freezer 338 (31-773) 513 (85-931) 605 (224-834) 617 (68-937) Side of Brain 8 L 7 R 6 L 5 R 7 L 7 R 9 L 5 R Weight of Brain (g) 1501 1442 1434 1481 Substance Abusea 14:0:1:0:0:0 7:1:0:0:1:1 7:0:0:4:1:2 9:0:1:1:2:1 Alcohol Abuseb 5:6:2:2:0:0 4:4:0:0:2:1 1:4:1:2:3:2 4:3:2:2:2:1 Suicide 0 5 8 4 Number of 11 2 2 4 6 Number of 47l 12 11 10 12 Number of 47m 15 11 14 14 Number of OLF 12 11 11 9 Number of MOS 13 10 13 13 Number of IOS 9 8 11 9

Table 2. Demographic information of the Neuropathology Consortium of the Stanley Medical Research Institute. Ratios a/b : little/none : social : moderate use (past) : moderate use (present) : heavy use (past) : heavy use (present).

3.2 Immunohistochemistry

The protocol used to stain the slides was published in Nature Protocols (Waldvogel et al., 2006). Staining was performed using mounted slides provided by the Neuropathology

Consortium. Frozen sections were fixed by being submersed in 4% paraformaldehyde for 10 minutes at room temperature. Sections were then placed in Phosphate Buffered Saline (PBS) with 0.2% Triton-X (PBS-triton) overnight at 4ºC. The following day, sections were placed in a solution of 0.1M sodium citrate buffer at a pH of 4.5 and kept overnight at 4ºC. Ten slides at a time were placed in a new 200 mL solution of 0.1M sodium citrate buffer (pH 4.5) and then placed in the center of a 650W microwave for 90 seconds. The solution cooled to room temperature and then the slides were washed with PBS-triton three times for 15 minutes each.

The slides were first incubated with a mouse anti-CUX2 monoclonal antibody (1:600,

Abnova), diluted in solution of PBS-triton with 1:100 Fetal Bovine Serum (FBS), for three days at 4ºC. The following day, the slides were washed with PBS-triton three times for 15 minutes and then were incubated with an Alexa Fluor® 488 Goat Anti-Mouse IgG antibody (1:200,

ThermoFisher), diluted in solution of PBS-triton with 1:100 Fetal Bovine Serum and kept overnight at 4ºC. On day seven, the slides were washed with PBS-triton three times for 15 minutes each at room temperature and then were incubated with a rabbit anti-ZNF312B polyclonal antibody (1:600, abcam), diluted in solution of PBS-triton with 1:100 Fetal Bovine

Serum (FBS), for three days at 4ºC. On the tenth day, after being washed with PBS-triton three times for 15 minutes each, the slides were incubated with an Alexa Fluor® 594 Goat Anti-Rabbit

IgG antibody (1:200, ThermoFisher) overnight at 4ºC. On the final day, slides were once again washed with PBS-triton three times for 15 minutes and then coverslips were placed on the slides with the addition of ProLong Gold Antifade Mountant with DAPI.

3.3 Image Analysis

3.3.1 Microscopy – Zeiss Epifluorescence Microscope

Images were obtained using a Zeiss Epifluorescence Microscope at 10x magnification and 10% overlap. Three images were produced (CUX2, FEZF2 and DAPI) and background corrections were applied to each of the images in order to correct for shading effects (Figure 6). The images were then stitched together using Volocity® 3D Image Analysis Software.

Figure 6. Gray-scale images of CUX2, ZNF312 and DAPI and an overlapped artificially coloured image. DAPI was coloured in blue, CUX2 in green and ZNF312 in red.

3.3.2 Regional and Laminar Delineation

Photoshop CS6 was used to crop the OFC into BA11, medial BA47 and lateral BA47, and then trace the cortical layers, white matter and pia. Determination of the BAs was based on gross anatomical landmarks, as well as cytoarchitectural differences. The anterior portion of the

OFC was divided into BA11, BA47l and BA47m, as defined by Uylings et al. (2010) (Figure 7).

As previously mentioned, BA11 is located on the ventromedial edge, covering the crown of the gyrus rectus. BA47 is located on the orbital surface of the hemisphere, with BA47m located on the medial and anterior orbital gyrus and BA47l located on the lateral orbital gyrus. We chose to keep BA47m and BA47l as uniform regions and did not further subdivide the areas. Given that the sections are frozen and only 14 μm thick, it would be extremely difficult to accurately distinguish the borders between the subdivisions. In terms of cytoarchitecture, BA47 can be distinguished from BA11 by larger neuronal cell bodies, more differentiated layers III and V, and a clearly delineated granular layer IV. The medial subdivision of BA47 can be distinguished from the lateral area by the presence of two darkly stained horizontal bands in infragranular layer V and VI. These bands are composed of densely packed cells, located in the sublayer Va

(upper band) and in the upper part of layer VI (lower band).

Figure 7. Regional and laminar delineation of the orbitofrontal cortex. OFC consists of BAs 11 and 47, with 47 being divided into a lateral (47l) and medial (47m) component. Each BA has six cortical layers.

Determination of the cortical layers was based on the typical cytoarchitecture of the neocortex, as described by Creutzfeldt (1995), and the cytoarchitecture of the OFC, as described by Uylings et al. (2010). According to Creutzfeldt (1995), the mature brain shows major or minor differences in the final design, which allow a distinction between cortical layers based on cytoarchitectonic criteria. These criteria are based on differences in the distribution and size of neuronal cell bodies revealed by cellular staining, such as Nissl stain. The following criteria, which were described in section 1.1.3, were used to delineate the cortical layers (Figure 8):

Figure 8. Cortical layers delineated based on cytoarchitectonic criteria. Layer I is neuron sparse, layer II and IV consist of small densely packed pyramidal cells, layer III is large, layer V contains large pyramidal cells and layer VI contains a multitude of cells.

- Layer I (molecular layer): a few scattered neurons, Cajal-Retzius cells, spiny stellate cells and extensions of apical dendrites and horizontally oriented axons.

- Layer II (outer granular layer): small pyramidal cells and numerous stellate cells.

- Layer III (outer pyramidal layer): predominantly small and medium sized pyramidal cells, and non-pyramidal cells with ascending or descending intracortical axons.

- Layer IV (granular layer): mostly spiny stellate cells and also pyramidal cells.

- Layer V (inner pyramidal layer): large pyramidal cells as well as interneurons.

- Layer VI (multiform layer): few large pyramidal cells, many small spindle-like pyramidal and multiform cells.

3.3.3 Automatic Cell Segmentation

ImageJ was used to segment outlines for cells (http://rsb.info.nih.gov/ij/). The protocol was created by a previous graduate student (Abbass, 2014), which was adapted from Woeffler-

Maucler et al. (2014). In summary, images are converted to 8-bit grayscale images and pixels with intensities greater than 200 are removed as outliers. Next, a fast-Fourier transformation

(FFT) was performed, with a bandpass filter applied to filter out small and large objects (smaller than 4.6 μm and greater than 61.2 μm). The FFT-bandpass filter also enhances the relative intensity of the cells and removes fluctuations over large and small distances. Images were autoscaled, putting the intensities to 0 and 255, and saturated to conserve differences in intensity. Next, the images were dichromatized using an intermodes automatic threshold, which places an intensity threshold on the image, where pixels above the threshold are made black and pixels below are made white. This creates a binary image with black filled cells on a white background. Next, a watershed function is applied, which separates two adjacent cells that may be connected. Lastly, outlines are automatically drawn over the segmented objects (Figure 9).

Figure 9. Automatic cell segmentation using ImageJ for (A) CUX2, (B) ZNF312, and (C) DAPI. From left to right, the first image depicts the original grayscale images, the next image is after removing outliers and performing the FFT function with a bandpass filter, the next image represents the application of the intermodes threshold and the last image is obtained after applying the watershed function and outlining the cells.

3.3.4 Automatic Data Generation

The three images (DAPI, CUX2, FEZF2) containing outlines of the segmented cells, along with the two images of the drawn cortical layers and the white matter and pia boundaries, are processed with algorithms produced in MATLAB (Figure 10). First, the area is enclosed by joining two perpendicular lines at the ends of the white matter and pia lines, creating a space

A B

D C

E F

Figure 10. The five images that are input into MATLAB and the resulting image. The three images produced by ImageJ containing the outlined CUX2 cells (A), ZNF312 cells (B) and DAPI cells (C), and the images created using Photoshop delineating the cortical layers (D) and pia and white matter boundaries (E) are input into MATLAB. The algorithm creates a composite image (F) where the layers are enclosed in the boundaries.

containing all of the layers. The algorithm differentiates the pia from the white matter as having the highest curvature and this was manually confirmed for each image. The algorithm then calculates cortex width, cell density, cell area and distance from pia.

3.3.4.1 Cortex Width

First, the area of each of the layers as well as the total cortex area was computed. Then, the width of each cortical layer was computed by dividing the area of each layer by the mean length of the lines enclosing the layer.

3.3.4.2 Cell Density

Cell density was measured by calculating the number of CUX2, ZNF312 and CUX2+ve/

ZNF312-ve cells per area per layer. Cells were only analyzed if they overlapped with a DAPI signal in order to increase the probability that the outline is a cell. The density measured is biased because cell centroids that are found outside the thickness of the slide could still extend into the slide and be counted as a unit. This leads to an overestimation of density, which also increases with cell size. We corrected for this problem, which is inherent to two-dimensional histological analyses, by multiplying the density measured by the Abercrombie correction factor

(F) (Abercrombie, 1946). F takes into account both the thickness of the slide and the average height of each cell when calculating the true density:

F = T DT = F X DM T + H

Equation 1: Abercrombie’s Correction Factor (F). F is calculated by dividing the thickness of the section (T) by the sum of T and the mean height of the cells (H). True density (DT) is equal to the measured density (DM) multiplied by F.

In brief, if the centroid of the cell is located at radius length or less from the slice, part of the cell will be within the thickness and it will be counted as a cell. If we consider the radius length on both sides of the slice of tissue, this equals to 2r, or the diameter (height) of the cell.

Thus, if the centroid of the cell is located within the thickness of the slice or within 1r on either side of the slice, it will be counted as a unit. When only a small portion of the cell is extended into the thickness it is counted as a unit, thus increasing density. In order to correct for this problem, we used an Abercrombie correction factor (equation 1). Abercrombie correction factor creates a ratio, where the true density over the thickness of the slice is equal to the measured density over the thickness plus the height of the cell (2r). When rearranging the equation, true density equals the measured density multiplied by thickness over thickness plus height. It is apparent that a larger T and smaller H would result in a more accurate DM. This is partly why we only analyzed cells that overlapped with DAPI, as DAPI only stains nuclei and results in a lower

It is impossible to measure H, given that it is perpendicular to the plane of the slide being investigated (Abbass, 2014). An assumption was made that cell nuclei are spherical and that two-dimensional representations of DAPI would be circular. The height of the nucleus was therefore equal to the diameter of the nucleus in the plane. The average measured diameter (dM) was calculated using equation 2.

dM =2 √(A/π)

Equation 2: Measured diameter (dM) for each cell is calculated from the area measured (A).

However, dM underestimates the true diameter (dT), given that some DAPI nuclei may be cut at the edges and would result in smaller cell diameters being measured (Andersen and Gundersen,

1999). Abbass (2014) created a formula to correct for this effect, which can be seen in Equation

3. The derivation of this formula can be seen in Abbass (2014) and is also provided in Appendix

1. The dT is substituted for H in equation 1.

2 dT =2(dM - T + √((T - dM) + π dM T)) π

Equation 3: True diameter (dT) calculated from measured diameter (dM) and thickness of the section (T).

3.3.4.3 Cell Area

Cell area was measured in the OFC in BA47l, BA47m and the entire cortex. Cell area was measured for ZNF312 cells only, since ZNF stains the entire cell body to a greater degree than

CUX2. Cell area was measured in layers III, IV and V because this is where ZNF312 is predominantly expressed.

3.3.4.4 Distance from Pia

The distance of each cell to the pial surface relative to the cortex width was measured for

CUX2, ZNF312 and CUX2+ve/ZNF-ve cells. In order to measure distance from pia, the cortex was segmented into triangles through Delaunay triangulation (Abbass, 2014). The triangles constituting the cortex are then mapped onto a trapezoid using area preserving parameterization.

That is, the trapezoid conserves the relative distances of the pia (y-axis 1) and white-matter lines

(y-axis 2) as well as cortex area and the area of each triangle (Figure 11). The centroid of each cell is mapped onto a triangle and its relative position is maintained when being mapped onto

the trapezoid. A given cell’s relative distance across the cortex is measured as its position on the x-axis and is given as a number between 0 and 1, with zero being located at the pia boundary and one being located at the white matter boundary.

A B

Figure 11. Delaunay triangulation. A. Delaunay triangulation on the cortex. B. Delaunay triangulation transformed onto a trapezoid with an area preserving parameterization. Colours are used to show the areas of the cortex transformed onto the trapezoid. Pia line (blue) and white matter line (green) (Image from Abbass, 2014).

3.4 Statistical Analysis

We performed three planned comparisons between controls and patients for all of our measures in order to test our hypothesis. First, Pearson’s r correlation was used to determine correlations between PMI, pH, days in freezer, age and our measures. For any significant differences found, these variables were included as covariates in the general linear model for each planned comparison. Additionally, we assessed whether gender and cerebral hemisphere affected the results. Any significant differences were included in the general linear model as a fixed variable.

Next, in order to counteract the problem of multiple comparisons, a Bonferroni post hoc test was applied. For measures of layer width and density, the p-value for significance (p=0.05) was divided by six to account for six layers investigated. Since only 3 layers were analyzed for cell

area, the p-value for significance was p=0.0167. All statistical analyses were performed on IBM

SPSS Statistics 24 and graphs were created using Microsoft Excel 2011.

3.5 Method Validation

3.5.1 Tissue Samples

Extra samples were provided from the Neuropathology Consortium of the Stanley Medical

Research Institute in order to validate the methods. We received 40 samples from 10 individuals and from each individual a pair of consecutive slides was provided from both the anterior cingulate cortex and the orbitofrontal cortex. A previous thesis project by Abbass (2014) used these methods to analyze the anterior cingulate cortex, thus the sample was expanded to include both the ACC and OFC for validation. For the purpose of validating the methods, the demographic information and diagnoses were not disclosed.

3.5.2. Staining

Of the 40 samples, 10 samples (5 ACC and 5 OFC) were Nissl stained and their consecutive pairs (5 ACC and 5 OFC) were immuohistochemically stained with the protocol described in section 3.2. The remaining 20 samples were stained with the protocol described below. In brief,

10 samples (5 ACC and 5 OFC) were stained with anti-CUX2 and anti-NeuN and their consecutive pairs (5 ACC and 5 OFC) were stained with anti-ZNF312 and anti-NeuN.

3.5.2.1 Cresyl Violet (Nissl) Staining

The cresyl violet solution was prepared by combining 0.3 g of cresyl violet acetate and 300 mL of double distilled water. The solution was mixed with a magnetic stir bar for one hour.

Immediately before use, 750 μl of glacial acetic acid was added to the solution and then it was

filtered. The frozen sections were first fixed by being submersed in 4% paraformaldehyde in

PBS (pH 7.4) for 10 minutes at room temperature. The sections were then rinsed twice for 5 minutes in PBS. Following that, the slides were rinsed for 1 minute in double distilled water.

Then, the slides were dipped in the cresyl violet solution for 20 minutes. Following staining, the slides were rinsed for 1 minute in double distilled water. Then the slides were dehydrated through an alcohol gradient: 3 minutes in 70% ethanol, 2 minutes in 95% ethanol and 2 minutes in 100% ethanol. Finally, the slides were incubated in two changes of xylene for 5 minutes each.

The slides were removed from xylene one at a time and coverslips were mounted with Cytoseal

Mounting Medium.

3.5.2.2 Anti-CUX2 and Anti-NeuN

Immunohistochemistry was performed as described in section 3.2 with a minor modification. On day seven, after the slides were washed with PBS-triton three times for 15 minutes at room temperature, the slides were then incubated with a rabbit anti-NeuN monoclonal antibody

(1:600, abcam) as opposed to a rabbit anti-ZNF312B polyclonal antibody. All of the subsequent steps were kept the same.

3.5.2.3 Anti-ZNF312 and Anti-NeuN

Immunohistochemistry was performed as described in section 3.2 with a minor modification. On day three, after the slides were washed with PBS-triton three times for 15 minutes at room temperature, the slides were then incubated with a mouse anti-NeuN monoclonal antibody

(1:600, Millipore) as opposed to a mouse anti-CUX2 monoclonal antibody. All of the subsequent steps were kept the same.

3.5.3 Image Analysis

Immunohistochemistry images were obtained using a Zeiss Epifluorescence Microscope as described above in section 3.3.1. Images of Nissl stained slides were captured using a Nikon

Eclipse TE200 Microscope at 10x magnification. Individual tiles were stitched together using the Grid/Collection Stitching Plugin on ImageJ (Preibisch et al., 2009). An example of a Nissl stained slide after it has been stitched can be seen in Figure 12. Nissl stained slides and slides stained with anti-CUX2, anti-ZNF312B and DAPI were used to compare the delineation of cortical layers. For both sets of slides, determination of the cortical layers was based on the typical cytoarchitecture of the neocortex (Creutzfeldt, 1995). As mentioned above, Photoshop

CS6 was used to delineate the layers and MATLAB automatically calculated the width of each layer and the entire cortex. Consecutive slides were used and the same subareas were analyzed to determine the similarity between the cortical width measured from Nissl stained slides versus slides stained with anti-CUX2, anti-ZNF312B and DAPI.

The remaining 20 samples were used to determine the overlap between CUX2 and

ZNF312 cells with NeuN. The representative images were randomly selected and have a surface area of approximately 160,000 μm2. A similar study selected images with a surface area of approximately 66,000 μm2 and had manual counts around 40 (Woeffler-Maucler et al., 2014).

We chose 160,000 μm2 in order to have approximately 100 cells per image. The Multi-point tool on ImageJ was used to manually count the number of CUX2 cells, the number of NeuN cells and the number of cells that expressed both CUX2 and NeuN. The same was done on consecutive slides with ZNF312B cells and NeuN cells. Manual counts were also compared to automatic counts generated by the ImageJ protocol mentioned above (section 3.3.3 and 3.3.4) in order to validate the soundness of our automatic counting protocol.

Figure 12. Nissl stain of the orbitofrontal cortex. Nissl stain clearly delineates the different layers of the OFC.

3.5.4 Statistical Analysis

Paired t-tests were used to compare the number of DAPI cells, CUX2+ cells and ZNF312+ cells manually counted versus automatically counted. Paired t-tests were also used to compare the thickness of the cortical layers and total cortex width measured from the Nissl stained slides to those immunohistochemically stained with anti-CUX2, anti-ZNF312 and DAPI. Bonferroni correction was used, dividing the p value 0.05 by 6 layers. Data is expressed as mean percentage

+ standard error of the mean. All statistical analyses were performed on IBM SPSS Statistics 24.

Chapter 4: Results 4.1 Method Validation

4.1.1 Automated Versus Manual Counts

Of the 20 slides stained, 9 anti-CUX2/anti-NeuN/DAPI slides and 7 anti-ZNF312/anti-

NeuN/Dapi slides were used in the analysis. The other 4 slides were eliminated due to damage.

Table 3 presents the manual and automated counts for CUX2 cells, ZNF312 cells and DAPI.

Cells Mean Manual Mean Automated Person's R Paired T-Test Sig. Count Count Correlation (Sig.) DAPI 118.06 + 4.55 117.81 + 3.49 0.890 (>0.001) 0.909 CUX2 83.78 + 4.25 81.00 + 6.10 0.839 (0.005) 0.441 ZNF312 85.14 + 9.21 83.57 + 8.31 0.979 (>0.001) 0.464 CUX2+Dapi+ 69.33 + 5.04 67.56 + 5.33 0.903 (0.001) 0.462 ZNF312+Dapi+ 64.00 + 5.21 62.00 + 5.48 0.927 (0.003) 0.369

Table 3. Automated counts of DAPI, CUX2, ZNF312, CUX2+ve/DAPI+ve and ZNF312+ve/DAPI+ve cells are similar to manual counts.

In general, the manual counts were significantly similar to the counts obtained with the

automated ImageJ protocol. The mean manual count for DAPI cells was 118.06, which was

similar to the automated count of 117.81 (p=0.909). Additionally, Pearson’s correlation test

indicated a relatively strong correlation (r=0.890, p>0.001). The mean manual count for CUX2

cells (83.78) was not significantly different from the mean automated count for CUX2 (81)

(p=0.441). Again, Pearson’s correlation test indicated a relatively strong correlation (r=0.839,

p=0.005). For ZNF312 cells, the mean manual count was 85.14, which was significantly similar

to the mean automated count (83.57) (p=0.464). Pearson’s correlation test indicated a strong

correlation between the two counting protocols (r=0.979, p>0.001). Similar results were seen

with the CUX2 and ZNF312 cells that were intersected with DAPI. When manually counting,

there was an average of 69.33 CUX cells, which was not significantly different than 67.56 cells

found with the automated ImageJ protocol (p=0.462). There was a relatively strong Pearson’s r

correlation between the counts found with the manual and automated protocol (r=0.903,

p=0.001). Finally, 64 ZNF312+ve/DAPI+ cells were manually counted and 62 were

automatically counted, which was not significantly different (p=0.369). Additionally, Pearson’s

correlation test indicated a relatively strong correlation (r=0.927, p=0.003).

4.1.2 Nissl Stain

Of the 20 samples stained with either Nissl or our staining protocol, two samples were

damaged. These two samples, along with their consecutive pairs, were not included in the

analysis. In total, there were 4 pairs from the OFC and 4 pairs for the ACC analyzed. The results

found no significant differences between the cortical thickness when measured from Nissl

stained slides and our immunohistochemistry protocol (Table 4). In layer II, there was a trend

for an increase in the width measured from Nissl stained slides, however this finding was not

significant after Bonferroni correction (p=0.03). All other findings were not statistically

different and there were no other trends observed.

Layer Nissl (μm) IHC (μm) Sig. Nissl Relative (%) IHC Relative (%) Sig. I 374.7 + 23.5 355.4 + 19.5 0.19 13.74 + 0.77 13.22 + 0.66 0.32 II 188.4 + 9.3 * 173.9 + 6.7 0.03 6.93 + 0.35 6.52 + 0.30 0.07 III 1062.1 + 67.7 1060.44 + 74.4 0.94 38.52 + 1.64 38.89 + 1.72 0.73 IV 169.3 + 5.6 175.3 + 7.4 0.35 6.23 + 0.21 6.61 + 0.39 0.21 V 458.7 + 32.4 466.2 + 42.9 0.77 16.64 + 0.79 16.91 + 0.95 0.69 VI 511.1 +44.4 508.7 + 56.6 0.93 18.37 + 0.97 18.25 + 1.18 0.86 Overall 2752.7 + 125.3 2728.7 + 152.4 0.68

Table 4. Cortical thickness measurements are similar between Nissl and immunohistochemically stained slides. Mean cortical thickness and standard error of the mean. * p < 0.05, ** p < 0.0083 (p=0.0083 is the cutoff for significance after Bonferroni correction).

4.1.3 Neuronal Nuclear Antigen (NeuN)

Sixteen slides (Section 4.1.1) were used to determine the number of CUX2 and ZNF312 cells that overlapped with NeuN. The percentage of CUX2 cells that were NeuN+ was 82.82% +

3.02. This was similar to the percentage of CUX2+NeuN+ cells when only analyzing CUX2 cells that intersected with DAPI (82.64% + 3.23). Figure 13 depicts the overlap between CUX2 cells, NeuN cells and DAPI cells.

Figure 13. The percentage of CUX2 cells that are co-stained with NeuN is 82.82%. Top row, from left to right, depicts DAPI stained cells, CUX2 stained cells and NeuN stained cells. Bottom row, from left to right, depicts the overlap between DAPI and CUX2 cells, the overlap of DAPI, CUX2 and NeuN and finally the overlap between CUX2 and NeuN. DAPI in blue, CUX2 in green, NeuN in red.

Similarly, the percentage of ZNF312 cells that were NeuN+ was 72.34% + 4.02. When only analyzing ZNF312+ cells that were intersected with DAPI, the percentage that were NeuN+ was

76.6% + 3.03. Figure 14 depicts the overlap between ZNF312 cells, NeuN cells, and DAPI cells.

Overall, these findings suggest that our antibody markers do not completely overlap with NeuN.

Figure 14. The percentage of ZNF312 cells that are co-stained with NeuN is 72.34%. Top row, from left to right, depicts DAPI stained cells, ZNF312 stained cells and NeuN stained cells. Bottom row, from left to right, depicts the overlap between DAPI and ZNF312 cells, the overlap of DAPI, ZNF312 and NeuN and finally the overlap between ZNF312 and NeuN. DAPI in blue, ZNF312 in red, NeuN in green.

4.2 Pearson’s Correlation

Of the 54 samples analyzed, only 14 samples had a visible area 11. Due to a small sample number, BA11 was not included in the analysis. Of the 52 measures, only 5 were significantly different (p<0.05) between 47m and 47l after performing ANOVAs. The significant differences were ZNF312 density (layer I), overall distance from pia (ZNF312 and

CUX2) and relative density of CUX2 (layer II and V). None of these findings were significant after Bonferroni correction. Since the areas were not as cytoarchitecturally different as expected, we also analyzed the orbitofrontal cortex area as a whole.

We used Pearson’s correlation and found that PMI, pH, days in freezer and age were correlated with several measures (p<0.05). PMI was positively correlated with total cortex width, width of layer III (%), ZNF312 density (layer II), relative density of CUX2 (layer II), relative density of CUX2+ve/ZNF312-ve (layer I) and negatively correlated with relative density of CUX2 (layer VI), density of CUX2+ve/ZNF312-ve (layer IV-VI), relative density of

CUX2+ve/ZNF312-ve (layer VI) and distance from pia (CUX2+ve/ZNF312-ve). The pH was positively correlated with ZNF312 density (layer III, V and VI), relative density of ZNF312

(layer III and V), overall density (CUX2 and CUX2+ve/ZNF312-ve), CUX2 density (layer II-

VI), relative density of CUX2 (layer VI), CUX2+ve/ZNF312-ve density (layer I, II and VI), relative density of CUX2+ve/ZNF312-ve (layer I, layer II) and negatively correlated with relative density of ZNF312 (layer I and II), cell area (layer III and V), relative density of CUX2

(layer IV), and relative density of CUX2+ve/ZNF312-ve (layer III-V). Days in freezer was positively correlated with relative density of ZNF312 (layer IV), relative density of CUX2 (layer

II and IV), density of CUX2+ve/ZNF312-ve (layer II), relative density of CUX2+ve/ZNF312-ve

(layer II and IV) and negatively correlated with overall density (ZNF312), ZNF312 density

(layer I, III and V), density of CUX2+ve/ZNF312-ve (layer VI), relative density of

CUX2+ve/ZNF312-ve (layer VI) and distance from pia (CUX2+ve/ZNF312-ve). Age was negatively correlated with relative density of ZNF312 (layer II). Given that the results are inconsistent, these findings are likely due to chance. However, since PMI, pH, days in freezer and age did correlate with multiple measures, they were includes as covariates in our model.

The p-values obtained are reported with and without covariates and in almost all cases, the inclusion of covariates did not affect the results.

We also found that gender and hemisphere correlated with some measures (p<0.05), and these were therefore included in the model as fixed variables. Gender was positively correlated with relative density of ZNF312 (layer II and IV), CUX2+ve/ZNF312-ve density (layer I) and negatively correlated with ZNF312 density (overall, layer I, III, V and VI). Hemisphere was positively correlated with width of layer II (%), width of layer IV (%), relative density of CUX2

(layer VI), and negatively correlated with width of layer V (%), relative density of ZNF312

(layer V) and relative density of CUX2 (layer III). Since gender and hemisphere correlated with multiple measures, they were included as fixed variables in our model.

4.3 Brodmann Area 47l

There were few significant findings in regards to cortex width or cortical width by layer in BA47l. As seen in Figure 15, patients with schizophrenia had a wider layer V (p=0.009) compared to controls. This trend remains without including covariates (p=0.006), however was not significant after post-hoc correction.

A Cortex Width B Cortical Width by Layer CNTRL 3500 45 MDD BP 3000 40 SCZ 35 ) 2500

μm 30 2000 25 20 * 1500 Width(%) 15 1000 Cortex Cortex Width( 10 500 5 0 0 CNTRL MDD BP SCZ 1 2 3 Layer 4 5 6

Figure 15. Trend for an increased thickness of layer V in SCZ in BA47l. A. Mean cortical width and standard error of the mean. B. Mean percentage of cortical width by layer and standard error of the mean. * p<0.05 with covariates, **p<0.0083 with covariates as compared to controls. p=0.0083 is the cutoff for significance after Bonferroni correction.

Absolute and relative density of ZNF312 is shown in Figure 16. There were no

significant differences in the absolute density of ZNF312 cells in patients as compared to control

subjects. In terms of relative density, there was a significant decrease in layer V in bipolar

disorder (p=0.001), major depressive disorder (p=0.001) and schizophrenia (p<0.001). All of

these findings were significant after Bonferroni correction and remained significant without

covariates in the model (p=0.001 for bipolar disorder, p=0.003 for major depressive disorder

and p=0.002 schizophrenia). There was also a trend for a decrease in the relative density of

ZNF312 in layer VI in bipolar disorder (p=0.04), however this finding was not significant after

post-hoc correction. This finding was robust without covariates being included in the model

(p=0.028).

A CNTRL ZNF312 Density B 400 1.8 ZNF312 Relative Density MDD BP 350 1.6 * SCZ 2 300 1.4 1.2 ** ** 250 ** 1 * 200 0.8 150 Relative Density 0.6 100 NumberCells/mm of 0.4 50 0.2

0 0 Overall 1 2 3 4 5 6 1 2 3 4 5 6 Layer Layer

Figure 16. Decreased relative density of ZNF312 cells in layer V in BP, MDD and SCZ in BA47l. A. Mean absolute density of ZNF312 cells and standard error of the mean. B. Mean density of ZNF312 cells relative to overall density and standard error of the mean. * p<0.05 with covariates, **p<0.0083 with covariates as compared to controls. p=0.0083 is the cutoff for significance after Bonferroni correction.

Similar results were found in regards to CUX2 density in BA47l (Figure 17). In terms of

absolute density, there was a trend for an increase in CUX2 density in layer I in major

depressive disorder (p=0.019), however this finding was not significant after post-hoc

correction. This finding was robust when covariates were not included in the model (p=0.034).

In terms of relative density of layer I, there was a trend for an increase in bipolar disorder

(p=0.015) and major depressive disorder (p=0.027), however neither of these findings were

significant after post-hoc correction. Additionally, only the finding in bipolar disorder was

robust without including covariates (p=0.021 for bipolar disorder). There was a significant

decrease in the relative density of CUX2 in layer V in bipolar disorder (p<0.001) and major

depressive disorder (p=0.001). Both of these findings were significant after Bonferroni

correction and were robust without the inclusion of covariates in the model (p<0.001 in bipolar

disorder and p=0.001 major depressive disorder). There was also a trend for a decrease in

schizophrenia (p=0.011), although this finding was not significant after post-hoc testing. The

finding remained without the inclusion of covariates in our model (p=0.019). Lastly, there was a

trend for a reduction in the relative density of CUX2 in layer VI in bipolar disorder (p=0.008)

and schizophrenia (p=0.046); however, these findings were not significant after post-hoc

correction. The findings were robust without the inclusion of covariates in our model (p=0.003

for bipolar disorder and p=0.013 for schizophrenia).

CNTRL CUX2 Density B CUX2 Relative Density A 450 1.8 MDD BP 400 1.6 SCZ

350 1.4 2 300 1.2 ** * ** 250 1 * * 200 * 0.8 * *

Relative Density 150 0.6

100 0.4 NumberCells/mm of 50 0.2 0 0 Overall 1 2 3 4 5 6 1 2 3 4 5 6 Layer Layer

Figure 17. Decreased relative density of CUX2 cells in layer V in BP and MDD in BA47l. A. Mean absolute density of CUX2 cells and standard error of the mean. B. Mean density of CUX2 cells relative to overall density and standard error of the mean. * p<0.05 with covariates, **p<0.0083 with covariates as compared to controls. p=0.0083 is the cutoff for significance after Bonferroni correction.

Few significant differences were found when analyzing CUX2+ve/ZNF312-ve cells in

BA47l (Figure 18). There were no significant differences in regards to the absolute density in

patient groups as compared to control subjects. There was a trend for an increase in the relative

density of CUX2+ve/ZNF312-ve cells in major depressive disorder in layer 1 (p=0.047),

however this finding was not significant after Bonferroni correction and was not observed

without the inclusion of covariates. There was also a trend for a reduction in the relative density

of CUX2+ve/ZNF312-ve in major depressive disorder in layer IV (p=0.028), but similarly this

finding was not significant after post-hoc correction and was not observed without the use of

covariates.

CUX2-ZNF312 Relative Density CNTRL A CUX2-ZNF312 Density B 2.5 90 MDD BP 80 SCZ 2

2 70 60 1.5 * 50 *

40 1 Relative Relative Density 30

NumberCells/mm of 20 0.5 10 0 0 Overall 1 2 3 4 5 6 1 2 3 4 5 6 Layer Layer

Figure 18. Trend for an increase in the relative density of CUX2+ve/ZNF312-ve cells in layer I and decrease in layer IV in MDD in BA47l. A. Mean absolute density of CUX2+ve/ ZNF312-ve cells and standard error of the mean. B. Mean density of CUX2+ve/ZNF312-ve cells relative to overall density and standard error of the mean. * p<0.05 with covariates, **p<0.0083 with covariates as compared to controls. p=0.0083 is the cutoff for significance after Bonferroni correction.

Results from ZNF312 cell areas are reported in Figure 19. No significant differences or trends

were observed in the cell area in BA47l in patient groups as compared to controls. Relative

distance of cells from the pia in relation to the white matter is shown in Figure 20. No

significant differences or trends were observed in the relative distance of cells in patient groups

as compared to controls.

ZNF312 Cell Area CNTRL MDD 210 BP SCZ

200

)

2 m

μ 190

180

Cell Area ( 170

160

150 3 5 6 Layer

Figure 19. No significant differences in ZNF312 cell size in BA47l. Mean ZNF312 cell size and standard error of the mean.

CNTRL Relative Distance MDD 0.6 BP 0.58 SCZ 0.56 0.54 0.52

0.5 0.48

0.46 Relative Relative Distance from Pia 0.44 0.42 0.4 ZNF312 CUX2 CUX2-ZNF312

Figure 20. No significant differences in relative distance from pia in BA47l. Relative mean distances from pia and standard error of the mean.

4.4 Brodmann Area 47m

Cortex width and cortical width by layer for BA47m can be seen in Figure 21. There was a

trend for an increase in the thickness of layer IV in major depressive disorder (p=0.048)

compared to controls, however this finding was not significant after Bonferroni correction and

the trend was not observed without the inclusion of covariates. Similar to BA47l, there were no

significant differences or trends in total cortical thickness. The absolute and relative density of

ZNF312 cells can be seen in Figure 22. There were few significant differences or trends in

ZNF312 cell density in BA47m. There were no significant differences or trends in the absolute

density of ZNF312 cells. There was a trend for a reduction in the relative density of ZNF312

cells in layer VI in major depressive disorder (p=0.03) compared to controls. This finding was

not significant after post-hoc correction and was robust without the inclusion of covariates

(p=0.033).

Cortex Width A B Cortical Width by Layer CNTRL 3000 45 MDD 40 BP 2500 SCZ

35 )

μm 2000 30 25 1500

20 Width(%) 1000 15 Cortex Cortex Width( 10 * 500 5

0 0 1 2 3 4 5 6 CNTRL MDD BP SCZ Layer

Figure 21. Trend for an increased thickness of layer IV in MDD in BA47m. A. Mean cortical width and standard error of the mean. B. Mean percentage of cortical width by layer and

standard error of the mean. * p<0.05 with covariates, **p<0.0083 with covariates as compared to controls. p=0.0083 is the cutoff for significance after Bonferroni correction.

CNTRL A B ZNF312 Relative Density 400 ZNF312 Density 1.8 MDD BP 350 1.6 SCZ

2 300 1.4 1.2 250 1 * 200 0.8

150 Relative Density 0.6 100 NumberCells/mm of 0.4

50 0.2 0 0 Overall 1 2 3 4 5 6 1 2 3 4 5 6 Layer Layer

Figure 22. Trend for a decrease in the relative density of ZNF312 cells in layer VI in MDD in BA47m. A. Mean absolute density of ZNF312 cells and standard error of the mean. B. Mean density of ZNF312 cells relative to overall density and standard error of the mean. * p<0.05 with covariates, **p<0.0083 with covariates as compared to controls. p=0.0083 is the cutoff for significance after Bonferroni correction.

There were several trends in CUX2 absolute and relative density in BA47m (Figure 23).

In regards to absolute density, there was a trend for an increase in layer I (p=0.019) and layer II

(p=0.018) in major depressive disorder. These findings were not significant after post-hoc

correction and the trends remained without the use of covariates (p=0.027 and p=0.009,

respectively). In regards to the relative density, there were several trends found only without the

use of covariates. There was a trend for a decrease in the relative density in layer V in bipolar

disorder (p=0.04), however this findings did not remain with the inclusion of covariates. There

was also a trend for a decrease in the relative density of CUX2 cells in layer VI in schizophrenia

(p=0.03) and major depressive disorder (p=0.027) without the use of covariates. Once again,

these findings did not remain with the inclusion of covariates.

In terms of trends with the inclusion of covariates in the model, there was a decrease in

the relative density of cells in layer V in major depressive disorder (p=0.028). This finding was

not significant after post-hoc correction and was robust with the inclusion of covariates

(p=0.021). There was also a decrease in the relative density of CUX2 in layer VI in bipolar

disorder (p=0.015). This find was significant after post-hoc correction and the trend remained

without the inclusion of covariates (p=0.002).

B CUX2 Relative Density CNTRL A CUX2 Density MDD 450 1.8 * BP 400 1.6 SCZ

1.4

2 350 300 1.2 * * 250 1 200 0.8

* Relative Density 150 0.6 100 0.4 NumberCells/mm of 50 0.2 0 0 Overall 1 2 3 4 5 6 1 2 3 4 5 6 Layer Layer

Figure 23. Trend for an increase in the absolute density of CUX2 cells in layer I and II and decrease in the relative density in layer V and VI in MDD in BA47m. A. Mean absolute density of CUX2 cells and standard error of the mean. B. Mean density of CUX2 cells relative to overall density and standard error of the mean. * p<0.05 with covariates, **p<0.0083 with covariates as compared to controls. p=0.0083 is the cutoff for significance after Bonferroni correction.

In regards to CUX2+ve/ZNF-ve cells, there were a few trends that were observed only

without the inclusion of covariates (Figure 24). There was a trend for an increase in the relative

density of cells in layer IV in bipolar disorder (p=0.017) without the inclusion of covariates.

Additionally, there was a trend for a reduction in the relative density in layer VI in bipolar

disorder (p=0.009) and schizophrenia (p=0.012) without the inclusion of covariates. In regards

to absolute density, there was a trend for an increase in layer I in major depressive disorder

(p=0.035). This finding was not significant after post-hoc correction and was robust without the

inclusion of covariates (p=0.035). Similarly, there was a trend for an increase in the relative

density of cells in layer I in major depressive disorder (p=0.031). This finding was not

significant after Bonferroni correction and the trend remained without the inclusion of

covariates (p=0.023).

CNTRL A CUX2-ZNF312 Density B CUX2-ZNF312 Relative Density 80 2 MDD BP 70 1.8 *

2 SCZ 1.6 * 60 1.4 50 1.2 40 1

30 0.8 Relative Relative Density 0.6

NumberCells/mm of 20 0.4 10 0.2 0 0 Overall 1 2 3 4 5 6 1 2 3 4 5 6 Layer Layer

Figure 24. Trend for an increase in the absolute and relative density of CUX2+ve/ZNF312-ve cells in layer I in MDD in BA47m. A. Mean absolute density of CUX2+ve/ZNF312-ve cells and standard error of the mean. B. Mean density of CUX2+ve/ZNF312-ve cells relative to overall density and standard error of the mean. * p<0.05 with covariates, **p<0.0083 with covariates as compared to controls. p=0.0083 is the cutoff for significance after Bonferroni correction.

Mean cell areas in BA47m are seen in Figure 25. No significant differences or trends

were observed in cell area in patient groups as compared to controls. Mean relative distance of

cells from pia is shown in Figure 26. No significant differences or trends were observed in the

relative distance in patient groups as compared to controls.

ZNF312 Cell Area CNTRL MDD 210 BP SCZ 200

)

2 190

m μ 180

Cell Cell Area ( 170

160

150 3 5 6 Layer

Figure 25. No significant differences in ZNF312 cell size in BA47m. Mean ZNF312 cell size and standard error of the mean.

CNTRL Relative Distance MDD 0.6 BP 0.58 SCZ 0.56 0.54 0.52 0.5 0.48 0.46 Relative Relative Distance from Pia 0.44 0.42

0.4 ZNF312 CUX2 CUX2-ZNF312

Figure 26. No significant differences in relative distance from pia in BA47m. Relative mean distances from pia and standard error of the mean.

4.5 Entire Orbitofrontal Cortex

There were similar significant findings and trends when analyzing the orbitofrontal

cortex as a whole. There were no significant differences or trends in total cortex width (Figure

27). There was a trend for an increase in the width of layer IV in schizophrenia (p=0.012).

However, this finding was not significant after Bonferroni correction and the trend did not

remain without the inclusion of covariates.

A B CNTRL Cortex Width Cortical Width by Layer MDD 3000 45 BP 40 SCZ 2500

) 35

μm 2000 30 25 1500 Width(%) 20 1000 15

Cortex Cortex Width( 10 * 500 5

0 0 CNTRL MDD BP SCZ 1 2 3 Layer 4 5 6

Figure 27. Trend for an increased thickness of layer IV in SCZ in the entire OFC. A. Mean cortical width and standard error of the mean. B. Mean percentage of cortical width by layer and standard error of the mean. * p<0.05 with covariates, **p<0.0083 with covariates as compared to controls. p=0.0083 is the cutoff for significance after Bonferroni correction.

In regards to absolute density of ZNF312 cells, there was a trend for a reduction in

bipolar disorder in layer VI (p=0.04) without the use of covariates (Figure 28). However, this

trend did not remain after the inclusion of covariates. For relative density, there was a trend for a

reduction in layer IV in schizophrenia (p=0.036). This finding was not significant after post-hoc

correction and the trend did not remain without the inclusion of covariates. There was also a

trend for a reduction in layer V in major depressive disorder (p=0.027). Once again, this finding

was not significant after Bonferroni correction and the trend did not remain without the

inclusion of covariates. There was a decrease in the relative density of ZNF312 cells in layer VI

in bipolar disorder (p=0.022) and major depressive disorder (p=0.017). These findings were not

significant after post-hoc correction. The trends remained for major depressive disorder without

the inclusion of covariates (p=0.035).

CNTRL A B ZNF312 Relative Density 400 ZNF312 Density 1.8 MDD BP 1.6 350 * SCZ 1.4 2 300 1.2 250 * 1 * * 200 0.8

150 Relative Desnity 0.6

100 0.4 NumberCells/mm of 50 0.2 0 0 Overall 1 2 3 4 5 6 1 2 3 4 5 6 Layer Layer

Figure 28. Trend for a decrease in the relative density of ZNF312 cells in layer IV in SCZ, layer V and VI in MDD and layer VI in BP in the entire OFC. A. Mean absolute density of ZNF312 cells and standard error of the mean. B. Mean density of ZNF312 cells relative to overall density and standard error of the mean. * p<0.05 with covariates, **p<0.0083 with covariates as compared to controls. p=0.0083 is the cutoff for significance after Bonferroni correction.

In regards to CUX2 absolute and relative density, there were several significant findings

and trends (Figure 29). There was a trend for an increase in the absolute density in layer I in

major depressive disorder (p=0.022). This finding was not significant after post-hoc correction

and was robust without the inclusion of covariates (p=0.032). There was also a trend for an

increase in absolute density in layer II in major depressive disorder (p=0.037) without the

inclusion of covariates. In regards to relative density, there was a trend for an increase in layer I

in major depressive disorder (p=0.026). However, this finding was not significant after

Bonferroni correction and the trend did not remain without the inclusion of covariates. In layer

II, there was a trend for an increase in schizophrenia (p=0.04) without the inclusion of

covariates. In layer V, there was a trend for a decrease in major depressive disorder (p=0.027)

and the trend remained without the inclusion of covariates (p=0.014). However, the finding was

not significant after Bonferroni correction. There was also a trend for a decrease in layer V in

bipolar disorder (p=0.023) without the inclusion of covariates. In layer VI, there was a

significant decrease in the relative density of CUX2 cells in bipolar disorder (p=0.004). This

finding was significant after post-hoc correction and was robust without the inclusion of

covariates (p=0.001). There was a trend for a reduction in layer VI in major depressive disorder

(p=0.026) and schizophrenia (p=0.042). Both of these findings were not significant after post-

hoc correction and were robust without the inclusion of covariates (p=0.022 and p=0.012,

respectively).

CNTRL A CUX2 Density B CUX2 Relative Density 450 1.8 MDD BP 400 1.6 SCZ

2 350 1.4 300 1.2 * * * 250 1 ** 200 0.8 * * Relative Density 150 0.6

NumberCells/mm of 100 0.4

50 0.2 0 0 Overall 1 2 3 4 5 6 1 2 3 4 5 6 Layer Layer Figure 29. Decreased relative density of CUX2 cells in layer VI in BP in the entire OFC. A. Mean absolute density of CUX2 cells and standard error of the mean. B. Mean density of CUX2 cells relative to overall density and standard error of the mean. * p<0.05 with covariates, **p<0.0083 with covariates as compared to controls. p=0.0083 is the cutoff for significance after Bonferroni correction.

Similar trends were seen in the CUX2+ve/ZNF312-ve cells (Figure 30). In regards to

absolute density, there was a trend for an increase in layer I in major depressive disorder

(p=0.043). This finding was not significant after post-hoc correction and the trend remained

without the inclusion of covariates (p=0.041). Similarly, there was an increase in the relative

density of cells in layer I in major depressive disorder (p=0.02) and this trend was robust

without the inclusion of covariates (p=0.02). However, this finding was not significant after

Bonferroni correction. Other trends included an increase in layer IV in bipolar disorder

(p=0.037) and a decrease in layer VI in bipolar disorder (p=0.012) and schizophrenia (p=0.012).

CNTRL A CUX2-ZNF312 Density B CUX2-ZNF312 Relative Density 80 2 MDD BP 70 * 1.8 SCZ 1.6 * 2 60 1.4 50 1.2 40 1

Relative Relative Density 0.8 30 0.6 20 NumberCells/mm of 0.4 10 0.2 0 0 Overall 1 2 3 4 5 6 1 2 3 4 5 6 Layer Layer

Figure 30. Trend for an increase in the absolute and relative density of CUX2+ve/ZNF312-ve cells in layer I in MDD in the entire OFC. A. Mean absolute density of CUX2+ve/ZNF312-ve cells and standard error of the mean. B. Mean density of CUX2+ve/ZNF312-ve cells relative to overall density and standard error of the mean. * p<0.05 with covariates, **p<0.0083 with covariates as compared to controls. p=0.0083 is the cutoff for significance after Bonferroni correction.

Mean ZNF312 cell areas in layers III, V and VI in the entire OFC can be seen in Figure

31. Mean distances from pia in the three cell populations (CUX2, ZNF312, CUX2+ve/ZNF312-

ve) are shown in Figure 32. Similar to the subareas of the OFC, there were no significant

differences or trends in cell area or distance from pia when analyzing the entire OFC.

CNTRL ZNF312 Cell Area MDD 210 BP SCZ

200 )

190 m μ 180

Cell Cell Area ( 170

160

150 3 5 6 Layer

Figure 31. No significant differences in ZNF312 cell size in the entire OFC. Mean ZNF312 cell size and standard error of the mean.

Relative Distance CNTRL MDD 0.6 BP SCZ 0.58 0.56 0.54 0.52 0.5 0.48 0.46

Relative Distance from Pia 0.44 0.42 0.4 ZNF312 CUX2 CUX2-ZNF312

Figure 32. No significant differences in relative distance from pia in the entire OFC. Relative mean distances from pia and standard error of the mean.

Chapter 5: Discussion 5.1 Labeled Cell Population

CUX2 and ZNF312 cells were found in all of the layers of the cortex. This finding was surprising, given that CUX2 and ZNF312 are thought to control laminar position of upper and lower layer neurons, respectively (Greig et al., 2013; Kwan et al., 2012; Nieto et al., 2004;

Zimmer et al., 2004). Additionally, CUX2 is expressed at high levels in CPN and ZNF312 in

CFPN (Greig et al., 2013; Molyneaux et al., 2009). In the present study, CUX2 and ZNF312 cells were found in all layers, with greater densities in layer II and IV. The findings were similar to overall neuron densities, where layers II and IV are the densest layers in the neocortex. It appears as though cells throughout the cortex express CUX2 and ZNF312, however the level of expression may differ. A population of CUX2+ve/ZNF312-ve cells was also investigated and these cells were found throughout layers I-VI, however with a greater proportion in the upper layers. This population of cells is difficult to characterize, however it appears as though they are preferentially found in layer II. The relative distance to pia findings suggest that CUX2 cells and

CUX2+ve/ZNF312-ve cells are closer to the pia than ZNF312, however these findings are subtle and were not significant.

There may be several possible explanations for why the markers were not layer-specific, including non-specific binding. Non-specific binding occurs when the antibody non-specifically binds to an endogenous receptor or a combination of ionic and hydrophobic interactions

(Buchwalow et al., 2011). This may create unwanted background staining, which then gets selected by the automated counting protocol and counted as a cell expressing the known marker.

Fetal bovine serum was used as a blocking reagent, which decreases the likelihood of non- specific binding.

Another possibility is that CUX2 and ZNF312 are expressed in cells in all of the cortical layers. CUX2 was expressed in a greater number of layer II neurons, but also was expressed in a large number of lower layer neurons. Similarly, ZNF312 was largely expressed in lower layers, however it was also expressed to a substantial degree in upper layers. A previous graduate student assessed the intensity of the stain and found that layer II cells were more intensely stained with CUX2 in comparison to the other layers, whereas layer V cells were more intensely stained with ZNF312 (Abbass, 2014). However, both markers were expressed in all of the layers. It may be possible that CUX2 and ZNF312 are expressed to a higher degree in upper and lower layers, respectively, however they are also expressed throughout the six layers. Further research is required in order to determine the protein expression of CUX2 and ZNF312 in the different layers of the neocortex (see Chapter 6). The cells that were less intensely stained are still counted by the automated counting protocol, since they are still visible cells. A threshold could be applied to only include the intensely stained cells, however this would increase the subjectivity of the methods.

It is also possible that CUX2 and ZNF312 stain non-neuronal cells, such as glial cells.

More specifically, it may be possible that non-neuronal cells expressing either of these markers are found throughout the cortical layers and are not as layer-specific as neurons. In other words, it may be possible that neurons expressing these markers are confined to certain cortical layers, however non-neuronal cells are found throughout the cortex. However, if this were the case, we would still expect a larger proportion of CUX2 cells in upper layers and ZNF312 cells in lower layers. Further studies are required in order to characterize the CUX2 and ZNF312 cell populations (see Chapter 6). For example, by co-staining our markers with glial and interneuron markers, we would be able to better understand and characterize which cells express CUX2 and

ZNF312.

Finally, these markers are well characterized in rodent cortical development; however there have been few studies in human tissue (Greig et al., 2013). Studies that found ZNF312 and

CUX2 expressed at high levels by SCPN and CPN, respectively, were conducted in mice (Chen et al., 2005; Molyneaux et al., 2009, 2005). Additionally, strong evidence comes from knockout mice, such as the absence of large pyramidal cells in layer V in Fezf2-null mice (Greig et al.,

2013). Although these markers have been well characterized in rodents, their expression in humans is unknown. A study by Arion et al. (2007) used in situ hybridization and found CUX2 and ZNF312 to be preferentially expressed in the upper and lower layers of the human prefrontal cortex, respectively. In our study, we used immunohistochemistry and found that

CUX2 and ZNF312 were expressed throughout all of the layers of the OFC. It may be possible that the levels of mRNA and protein do not correlate, which may be due to differences in the regulation of mRNA and protein synthesis and degradation. It is also possible that the expression of these transcription factors differs between areas of the cortex. This may be likely given that the OFC has a unique set of connections in comparison to the rest of the prefrontal cortex (Kringelbach, 2005).

The Allen Brain Atlas is a useful tool, as it provides gene expression data for mouse brains as well as human brains (Allen Institute for Brain Science, 2017). There is no specific data on the orbitofrontal cortex, however in the dorsolateral prefrontal cortex and frontal cortex,

CUX2 expression is enhanced in the upper layers. However, it also appears as though CUX2 is expressed throughout the other cortical layers to a lesser degree. This may explain why we observe CUX2 cells in all of the layers, especially given that our protocol does not differentiate between the intensity of the signals. The data on ZNF312 expression is limited and further research is required in order to characterize the expression in human tissue.

5.2 Method Validation

The mean cell counts were significantly similar between the manual and automated counting methods. When assessing the percentage of mean automated cells to those counted manually for all cell populations the percentage was approximately 95-100%. More specifically, for DAPI, the mean manual count was 118.06 and the mean automated count was 117.81

(99.78%). For CUX2 cells, the mean manual count was 83.78 and the mean automated count was 81 (96.67%). Similarly, for ZNF312 cells, the mean manual and automated counts were

85.14 and 83.57 (97.69%), respectively. There may exist slight discrepancies simply due to the subjectivity of manually counting. Overall, these findings suggest that our methods are valid and are able to measure similar cell counts as labor-intensive manual methods. Additionally, by using automated counting protocols, we were able to increase the objectivity of the methods and analyze larger areas of the cortex than previously investigated.

Overall, we found that cortical layer thickness measurements were similar between

Nissl-stained slides and slides immunohistochemically stained with anti-CUX2, anti-ZNF312 and DAPI. When looking at both absolute and relative measurements for cortical thickness, there were no significant differences between Nissl stained slides and slides stained using our protocol. There was a trend for an increase in the thickness of layer II when measured using

Nissl stained slides (188.42 μm) in comparison to the immunohistochemically stained slides

(173.89 μm). However, this finding was not significant after Bonferroni correction (p=0.03).

Although we were using consecutive slices, it may be possible that the areas of the cortex selected for analysis differed slightly between the two groups, accounting for the subtle differences in measurements. Overall, these findings suggest that our immunohistochemistry protocol can be used as an accurate way to delineate the cortical layers. Additionally, we found

the delineation of cortical layers to be more straightforward and objective when using our markers in comparison to Nissl stains. Given that we are labeling three different markers with three different colours, we were better able to visualize the border between the different cortical layers. We believe that our methods will allow for more accurate delineation and analysis of the cortical layers and could be used in the future by other researchers.

The cells were co-stained with NeuN in order to determine the number of cells expressing our markers that were neurons. Approximately 82% of CUX2 cells were co-stained with NeuN when analyzing both CUX2 cells as well as CUX2 cells that overlap with DAPI. In regards to ZNF312, approximately 72% of cells were co-stained with NeuN and when analyzing

ZNF312 cells that overlap with DAPI the percentage co-stained with NeuN was approximately

76%. These findings were somewhat unexpected, given that we expected that all of the cells expressing CUX2 or ZNF312 were neurons. However, there may be several possible reasons for these results. It may be possible that the remaining 20-30% of cells are non-neuronal cells, such as glial cells. Increasingly, this may be possible given that recent studies have shown that

CUX2- and ZNF312-expressing radial glial cells generate different projection neurons residing in layers II-VI as well as glial cells (Eckler et al., 2015; Guo et al., 2013). However, it remains unclear whether these mature glial cells express CUX2 and ZNF312. Further, there is a lack of evidence on the expression of CUX2 and ZNF312 in human brain cells, given that most of the research comes from rodent cortical development.

Another possibility is that the co-staining is in reality closer to 100%, however part of the cell is outside of the slice or not in the plane analyzed, and therefore is not counted as a unit.

This could be possible, given that in some cases where the cell was CUX2+ and NeuN-, a faint

NeuN signal was observable in the background however it was not counted as a cell. It may be

possible that the entire cell was not in the thickness of the slice. This could be possible for

ZNF312, given that ZNF312 primarily stains the cell body whereas NeuN stains the nucleus.

Therefore, if the center of the nucleus was found outside the thickness of the slice, there would be a ZNF312 signal but not a NeuN signal. However, this is unlikely given that we found similar results when analyzing ZNF312 cells that intersected with DAPI, and both DAPI and

NeuN stain the nucleus. Additionally, this is unlikely the case for CUX2 cells, given that both

CUX2 and NeuN primarily stain the nucleus.

Finally, it may be possible that the 20-30% of cells were simply artifacts and not actually cells. However, this is unlikely for several reasons. Firstly, manual-counting methods, which allow for the inspection of cells, found similar cells counts to those obtained with automated counts. Additionally, nearly 100% of cells were co-stained with either DAPI or NeuN, increasing the likelihood that the CUX2 or ZNF312 cells are in fact cells (data not shown). It is likely that either the 20-30% of cells are non-neuronal cells or that a portion of the cells are extended outside the thickness of the slice; however, future studies are required in order to confirm these hypotheses (see Chapter 6).

5.3 Summary of Findings

There were minimal differences found in schizophrenia, BP and MDD compared to controls (Table 5). These results are consistent with the literature, which have found subtle abnormalities in the cytoarchitecture of the cerebral cortex in these psychiatric disorders. The strongest findings were in the lateral OFC and included a decreased relative density of ZNF312 cells in layer V in BP, MDD and schizophrenia and a decreased relative density of CUX2 cells in layer V in BP and MDD. Other trends were observed, however, they were not statistically significant. There were several trends in the medial OFC, however no significant findings.

Trends were somewhat consistent between the medial and lateral OFC, such that there was a

decrease in the relative density of ZNF312 and CUX2 cells in lower layers. When analyzing the

entire OFC, the only statistically significant finding was a decrease in the relative density of

CUX2 cells in layer VI in bipolar disorder.

Psychiatric Disorders BA47l BA47m Entire OFC Schizophrenia - Trend for an increase in - No significant - No significant the relative width of differences or trends differences layer V - Trend for an increase - Decrease in the relative in the relative width of density of ZNF312 cells in layer IV layer V Major Depressive - Decrease in the relative - No significant - No significant Disorder density of ZNF312 and differences differences CUX2 cells in layer V - Trend for a decrease - Trend for a decrease in - Trend for an increase in in the relative density of the relative density of the relative density of ZNF312 and CUX2 cells ZNF312 cells and CUX2 CUX2 and CUX2+/ZNF- in layer VI and V, cells in layers V and VI cells in layer I respectively Bipolar Disorder - Decrease in the relative - No significant - Decrease in the relative density of ZNF312 and differences density of CUX2 cells in CUX2 cells in layer V - Trend for a decrease layer VI - Trend for a decrease in in the relative density of - Trend for a decrease in the relative density of CUX2 cells in layer VI the relative density of ZNF312 and CUX2 cells in ZNF312 cells in layer VI layer VI

Table 5. Summary of findings. We found subtle abnormalities of the cytoarchitecture of the OFC in patients with schizophrenia, major depressive disorder and bipolar disorder.

Overall, these findings suggest that there may be evidence for orbitofrontal pathology in

schizophrenia, BP and MDD. In schizophrenia, the strongest finding was a decrease in relative

density of ZNF312 in layer V in BA47l. There were several other trends, including an increased

layer V width in BA47l, an increased layer IV width in the entire cortex and decreased relative

density of CUX2 cells in lower layers V and VI. There were more findings in major depressive

disorder and bipolar disorder, including trends for a decrease in the relative density of cells in

lower layers and an increase in upper layers. These findings were consistent across the medial and lateral OFC and were also observed when analyzing the entire OFC. Overall, the effects were small, suggesting that there may only be subtle abnormalities in the cytoarchitecture of the

OFC in psychiatric disorders.

5.4 Schizophrenia

We hypothesized that schizophrenia is associated with neuronal migration defects in the orbitofrontal cortex and as a result, neurons destined for the upper layers would be found ectopically in lower cortical layers. This would be evidenced by greater densities of neurons in lower layers as well as greater distances from pial surface in schizophrenia compared to control.

Overall, we did not find evidence to support this hypothesis. In contrast, we found lower relative densities of CUX2 and ZNF312 in lower layers compared to controls. Additionally, all populations of cells appeared closer to the pial surface compared to controls, however this finding was not significant.

There are several possible explanations for why the results do not support the hypothesis.

Firstly, it may be possible that there are abnormalities in neuronal migration, however the effects are subtle. Schizophrenia differs from neuronal migration disorders such as lissencephaly, where an impaired cortical lamination of neurons is clearly observable (Moon and Wynshaw-Boris, 2013). Another possibility is that neuronal migration defects affect the areas of the cortex differently and that the effects may be subtle in the OFC. This is unlikely, however, given that one would expect the genes involved in the regulation of neuronal migration to be expressed throughout the different areas of the cortex. Finally, it is likely that schizophrenia has different etiologies, given the heterogeneity of the disorder. Many different combinations of gene variants can cause genetic risk of schizophrenia, suggesting that not every

case may be affected by neuronal migration defects (Muraki and Tanigaki, 2015). It may be possible that the analysis of a heterogeneous population leads to a lack of significant findings.

One of the strongest findings in schizophrenia was a decrease in the relative density of

ZNF312 cells in layer V. This was found when analyzing BA47l and was not consistent in the other areas. However, there were several other trends that were somewhat consistent. In BA47l, there was also a decrease in the relative density of CUX2 cells in layers V and VI. When analyzing the entire cortex, there was a decrease in the relative density of ZNF312 in layer IV and of CUX2 in layer VI. However, these findings were not significant after correcting for multiple comparisons. Another finding was an increased width of layer V in BA47l. However, this finding was not consistent between the lateral and medial area or when analyzing the entire cortex. There was also a trend for an increase in layer IV in the entire OFC, however this finding was not statistically significant. In general, it appears as though there is a decrease in the relative density of cells in the lower layers of the cortex compared to controls. Given that the findings are not robustly consistent, it may be possible that these were chance findings as a result of multiple comparisons as opposed to real findings. Further support for this comes from the fact that the findings are subtle.

It is difficult to compare our results to other studies, given that we used a different and novel methodology. We are analyzing a unique subset of cells that express the CUX2 and

ZNF312 markers, which differs from previous studies that typically analyze whole glial and neuronal cell populations. Additionally, histological studies will often use manual cell counting, which allows for only a small subset of the region to be analyzed. Given that we used a high- throughput approach, we were able to analyze larger areas of the OFC than previously investigated. This is the first cytoarchitectural study, to our knowledge, that separated the OFC

into a lateral and medial region and analyzed both subareas separately as well as together. Other studies have analyzed a small subsection of the OFC, with most studies investigating the lateral wall of the caudal orbitofrontal sulcus. Given that the methodology of our study differs from previous studies, it is difficult to directly compare results.

Other studies have observed abnormalities in cell density in schizophrenia; however, very few studies have analyzed cell density in the orbitofrontal cortex. In the prefrontal cortex, neuronal density has been significantly decreased in some subareas and increased in others

(Selemon et al., 1998, 1995). Other studies have found no significant differences in neuronal density in the prefrontal cortex area (Cotter et al., 2002; Cullen et al., 2006). It is difficult to assess how our findings compare to those in the prefrontal cortex, given the inconsistency in the literature as well as differences within the individual subareas of the prefrontal cortex. A study in the orbitofrontal cortex found no significant differences in neuron or glial density in any of the cortical layers in schizophrenia (Cotter et al., 2005). Another study analyzed kainate-positive neurons and found a significant reduction in the density in schizophrenia compared to controls

(Garey et al., 2006). However, given that this was a specific neuron subpopulation, it is difficult to compare these findings to our results. Overall, we found that there was a significant decrease in the relative density in ZNF312 cells in layer VI in BA47l. However, given that this finding was not consistent between the layers and is not supported by the literature, it may have arisen by chance.

Increased thickness of layer V in schizophrenia has been minimally reported in the cortex and not in the OFC. Abbass (2014) also found a larger layer Va in area 24 of the anterior cingulate cortex in patients with schizophrenia. Other studies have found a decrease in the absolute thickness of layer V in the subgenual cingulate cortex (Williams et al., 2013).

Similarly, Selemon et al. (1995) found a reduction in layer V in BA9 of the prefrontal cortex. To our knowledge, very few studies have analyzed the mean layer thickness by cortical layer in the

OFC in schizophrenia. Cotter et al. (2005) found no significant differences in the mean layer thickness in layers I-VI in schizophrenia compared to controls. Similarly, Williams et al. (2013) found no significant difference in the thickness of cortical pyramidal layers III and V in the

OFC. Given that this finding was not consistent between the medial and lateral OFC and is not supported by the existing literature, it is likely that this was a chance finding.

There is a lack of consistency in the results, such that the findings are not consistent between the subareas and are not supported by the existing literature. It may be possible that these abnormalities only affect one subarea of the OFC. This is reasonable given that the OFC is a large area that is heterogeneous and varies anatomically and functionally both medially to laterally and posteriorly to anteriorly. However, it is also possible that these findings were simply due to chance, seeing as the study involves multiple comparisons in each subarea, for each layer. It is likely that there are few abnormalities in the cytoarchitecture of the OFC in schizophrenia and that this is only evident when analyzing the entire area.

5.5 Bipolar Disorder

There were several robust findings in bipolar disorder, particularly in BA47l and the entire cortex. There was a significant decrease in the relative density of ZNF312 cells and

CUX2 cells in layer V in BA47l. Additionally, when analyzing the entire cortex, there was a significant decrease in CUX2 cells in layer VI, which is somewhat consistent with the findings in BA47l. Overall, these findings suggest that the relative density of cells may be decreased in lower layers in bipolar disorder compared to control. There were several trends that were consistent with these findings. In BA47l, a consistent trend was a decrease in the relative density

of ZNF312 cells and CUX2 cells in layer VI. Similarly, in BA47m, there was a trend for a decrease in the relative density of CUX2 cells in layer VI. Finally, when analyzing the entire cortex, the relative density of ZNF312 cells had a trend of being decreased in layer VI. These trends are not significant after correcting for multiple comparisons. However, given that they are consistent between the areas, it increases the likelihood that these are real findings rather than statistical anomalies. These consistent trends further support the notion that cell density may be decreased in lower layers in bipolar disorder.

An inconsistent finding was an increase in the relative density of CUX2 cells in layer I in

BA47l. Given that this result was not found in any other cell populations or subarea, it is likely that this is a chance finding. However, in major depressive disorder, there were decreases in the relative density of cells throughout the subareas in lower layers as well as increases in the upper layers (Section 5.6). Similar abnormalities may be in bipolar disorder, however the effects were too subtle to detect in the other cell populations and areas. Additionally, the finding in BP was subtle and was not significant after Bonferroni correction. Whether the increased relative density of CUX2 cells in layer I was a chance finding or real effect, the difference in bipolar disorder is subtle.

Few studies have assessed the cytoarchitecture of the OFC in patients with bipolar disorder. A study by Cotter et al. (2005) found no significant differences in cell density in the

OFC in BP, however they found neuronal size reductions in layer I and V. These findings do not support are results, given that we found no differences in cell size in BP as well as lower density of cells in the lower layers. Studies in other areas of the cortex have also found reduced neuronal density. Rajkowska et al. (2001) studied dorsolateral prefrontal cortex area 9 and found reduced neuronal density in layer III and reduced pyramidal cell density in layer III and V. In the

anterior cingulate cortex, studies have found decreased neuronal density in layer II and increased in layer VI in BP (Benes et al., 2001; Chana et al., 2003). This greatly differs from our findings, as we found decreased neuronal densities in lower layers. Overall, the results on neuronal density are inconsistent and vary based on the cortical area studied.

It may be possible that decreased densities were observed due to impaired neuronal migration. Bipolar disorder, like schizophrenia, is considered a neurodevelopmental disorder and genetic studies have identified genes involved in neuronal migration as candidate susceptibility genes. For instance, variations of RELN have been identified as a risk factor for developing bipolar disorder in women (Goes et al., 2010). Animal studies have found a relationship between alterations in Reelin gene expression and glutamic acid decarboxylase

(GAD) positive interneuron deficit (Nullmeier et al., 2011). Additionally, a study by Heckers et al. (2002) studied hippocampal sections and found the density of GAD mRNA-positive neurons was decreased by approximately 40% in patients with bipolar disorder. It is unclear whether our markers are also expressed in interneurons; however, it may be possible that decreases in the relative density were due to migration deficits in interneurons. Studies have also found variations of Nrg1, an epidermal growth-like protein that plays a role in cell migration, associated with the development of bipolar disorder (Prata et al., 2009). It may be possible that deficits in neuronal migration are partially responsible for the decreased relative density of cells observed in the lower layers in our study.

Apoptosis has also been suggested as a model for explaining neuron deficits in bipolar disorder (Uribe and Wix, 2012). Studies in patients with bipolar disorder have found a higher expression of pro-apoptotic molecules, such as Bcl-2 associated death promoter (BAD), Bcl-2- associated x protein (BAX), caspase-3 and caspase-9, as well as a reduced expression of anti-

apoptotic molecules, such as Bcl-2 (Benes et al., 2006). Additionally, Reelin and Nrg1 play a role in neuronal survival and as mentioned above, variations of the genes have been found in bipolar disorder (Uribe and Wix, 2012). Based on the previous literature, it may be possible that lower densities of cells were observed in bipolar disorder due to apoptosis of neurons. However, it is unclear as to why reduced neuron densities were found only in lower layers of the cortex and not throughout all of the layers. Additionally, only relative density was decreased and not absolute cell density, increasing the unlikeliness of apoptosis. Reduced neuronal density in lower layers was also found in major depressive disorder and will be discussed further in

Section 5.6.

5.6 Major Depressive Disorder

The strongest findings in major depressive disorder were a significant decrease in the relative density of ZNF312 cells and CUX2 cells in layer V in BA47l. Both of these findings were also found in bipolar disorder and the reduction of ZNF312 cell density was found in schizophrenia. There were no other statistically significant findings; however there were several consistent trends, including a decrease in the relative densities of cells in lower layers. In

BA47m, there was a trend for a decrease in the relative density of ZNF312 cells in layer VI and of CUX2 cells in layer V. This was not consistent in BA47l, however it was observed when analyzing the entire cortex. More specifically, in the entire OFC, there was a decrease in the relative density of ZNF312 cells and CUX2 cells in layer V and VI, however this finding was not significant after post-hoc correction. These findings suggest that there may be a decrease in the relative density of cells in the lower layers in MDD.

Other consistent findings included a trend for an increase in the density of CUX2 cells and CUX2+ve/ZNF312-ve cells in upper layers. There was an increase in the absolute density of

CUX2 cells and CUX2+ve/ZNF312-ve cells in layer I in BA47m and the entire OFC. There was also an increase in the absolute density of CUX2 cells in layer II in BA47m. These findings were not significant after post-hoc correction. In general, there also appears to be a higher overall density of cells in MDD; however this was not significant. Although it is possible that there is a greater cell density in MDD, it may also due to differences in staining intensity.

Therefore, we also calculated relative density, which was the ratio of the density of each layer over the overall density. When analyzing relative density, there was an increase in the relative density of CUX2 cells in layer I in BA47l and the entire OFC. There was also an increase in the relative density of CUX2+ve/ZNF312-ve cells in layer I in BA47l, BA47m and the entire OFC.

These findings suggest that there may be an increase in the density of CUX2 labeled cells in the upper layers of the cortex in major depressive disorder.

There were a few findings that were inconsistent between the subareas. There was a trend for an increase in the cortical thickness of layer IV in BA47m. However, this effect was subtle and it was not significant after correcting for multiple comparisons. Additionally, there was a trend for a decrease in the relative density of CUX2+ve/ZNF312-ve cells in layer IV in

BA47l. Once again, this finding was not significant after Bonferroni correction. Given that these findings were not consistent in the other subareas, cell populations or even the other psychiatric disorders, it is likely that these were chance findings rather than real effects.

Other studies in the OFC in major depressive disorder have found somewhat similar findings to ours. One study found that the overall pyramidal neuron density was significantly decreased in elderly depressed individuals compared to controls, and specifically in layers IIIc and V (Rajkowska et al., 2005). This finding is similar to ours, where we found a significant decrease in the relative density of CUX2 and ZNF312 cells in layer V in BA47l. The authors

also found no significant differences in the density of non-pyramidal neurons and glial cells. In contrast to our findings, another study by the same group of researchers found a decrease in neuronal density in upper cortical layers of the rostral OFC in patients with MDD (Rajkowska et al., 1999). Other studies found no significant differences in neuronal density in MDD compared to controls (Cotter et al., 2005; Khundakar et al., 2011). There clearly exist inconsistencies in the literature, which is likely due to the different methods employed as well as different subareas of the OFC analyzed. Our findings are neither supported nor rejected by the existing literature; however they contribute to the uncertainty of the orbitofrontal pathology in major depressive disorder.

Similar to bipolar disorder, there are several possible explanations for these results, including impaired neuron migration. Linkage analysis and association studies have found

DISC1 to be significantly associated with schizophrenia, bipolar disorder and major depressive disorder (Muraki and Tanigaki, 2015). Some studies have found an association between NRG1 polymorphisms and MDD, whereas others have failed to find any significant findings (Schosser et al., 2010; Wen et al., 2016). It may be possible that there are slight defects in neuronal migration, accounting for decreases in the relative density in lower layers and increases in the upper layers. Additionally, the increased density in upper layers and decreased density in lower layers may be explained by an over migration of neurons. In other words, neurons destined for lower layers continue to migrate into upper layers. Over migration has been described in neuronal migration disorders such as type II lissencephaly, including Walker-Warburg syndrome (Pang et al., 2008). However, to our knowledge, there is little to no research on over migration of neurons in MDD. Overall, MDD is not typically characterized as a neuronal migration disorder, which may explain why the effects were subtle.

There also exists evidence for apoptotic cell death in MDD, which may account for decreases in the relative density of cells. In the orbitofrontal cortex, a study found an increase in the levels of apoptosis-promoting protein caspase-8 and an increased ratio of apoptotic protein direct IAP binding protein with low pI to anti-apoptotic protein X-linked inhibitor of apoptosis

(Miguel-Hidalgo et al., 2014). These findings suggest that markers of cell vulnerability to degeneration are increased in MDD. However, anti-apoptotic protein levels were normal in

MDD compared to control, suggesting that morphometric evidence of cell death may not be clearly visible. This may explain why in our study the effect was small even when the findings were significant. Additionally, postmortem tissue studies have found increased expression of genes involved in apoptosis in MDD in the frontal cortex (Shelton et al., 2011). Finally, candidate-gene studies have found evidence for an association between apoptosis signaling and depression. For instance, polymorphisms of the adaptor protein Apaf1, which activates caspase-

9, have been associated with MDD (Harlan et al., 2006). It may be possible that decreases in the relative density of cells were due to apoptosis of neurons. However, it seems unlikely that apoptotic genes would only affect cells in the lower layers. Additionally, as previously mentioned, no decreases in absolute density were observed.

A possible explanation for the increased relative density of CUX2 and CUX2+/

ZNF312-ve cells in layer I and II is reduced neuropil. The reduced neuropil hypothesis, which was originally described in schizophrenia, suggests that psychiatric disorders may be associated with atrophy of neuronal processes (Selemon and Goldman-Rakic, 1999). Since this is not accompanied by an actual loss of neurons, an increase in neuronal density is observed. This may explain the increase in the relative density of cells in the upper layers of the cortex. The reduced neuropil hypothesis has been extensively studied and is supported in the hippocampus in MDD

(Stockmeier et al., 2004). However, there exists limited literature on dendritic size and

arborization in the OFC in MDD. Overall, it seems unlikely that reduced neuropil would only affect cells in layer I and II, and not in the other layers. Given that the effect was small, it may be possible that this was a chance finding as a result of multiple testing rather than a real effect.

Overall, it is unclear why there was a reduced relative density of cells observed in the lower layers and increased density in the upper layers. Of the several possibilities given above, over migration is the only one that explains both the reduced relative density in lower layers and the increased relative density in upper layers. It may be possible that neurons migrated beyond their destined layer, causing neurons destined for lower layers to be found superficially in upper layers. In general, there is little to no evidence of over migration of neurons in major depressive disorder and further research is required in order to test this hypothesis.

5.7 Significance of Methods

To our knowledge, this was the first histological study to analyze the entire area of the

OFC. The OFC is a large area that is anatomically and functionally different medially to laterally and anteriorly to posteriorly. Previous studies typically analyze a portion of the OFC, such as the lateral wall of the caudal orbitofrontal sulcus. In contrast, we used semi-automated methods, which allowed us to analyze the entire area of the OFC. For instance, previous studies have analyzed approximately 600 neurons per subject (Cotter et al., 2005). In our study, we were able to analyze on average 30,000 cells in one subject, increasing the number of cells analyzed by a factor of 50. It may be possible that inconsistencies in the literature arise when looking at small areas of the OFC in a limited number of subjects. When analyzing the entire

OFC, we found few abnormalities in schizophrenia, bipolar disorder and major depressive disorder. Our findings further suggest that there may be subtle abnormalities in the

cytoarchitecture in these psychiatric disorders and that this is only detected when analyzing the entire cortex.

The use of automated counting protocols allowed us to objectively analyze the OFC. By using a high-throughput approach, we were able to eliminate limitations of manual counting methods, such as subjectivity and labor-intensiveness. Cellular pathologies for a given disease likely vary across individuals; however there may be observable patterns between individuals.

With the use of automated methods such as ours, we would be able to quickly scan post-mortem brain slices from groups of heterogeneous individuals and determine whether such patterns or subgroups exist. This would ultimately enhance the understanding of complex, heterogeneous psychiatric disorders. Additionally, we believe that the markers used allowed us to better visualize the layers of the cortex, in contrast to traditional methods such as Nissl. We hope that other researchers will use these automated counting protocols in order to objectively analyze large areas of the cortex and resolve inconsistencies in the literature.

5.8 Limitations

There are several limitations to this study that are found in most histological studies.

Firstly, we studied two-dimensional sections because it allowed us to increase the sample size and analyze a greater area of the cortex. However, there are several limitations to studying two- dimensional slides. One problem is that density measurements are typically overestimated, since cells that are primarily outside the thickness of the section, yet partially extend in it are counted as a unit. In order to correct for this problem, we multiplied the density measured by the

Abercrombie correction factor (Abercrombie, 1946). This calculation takes into account both the thickness of the slice and the average height of each cell. However, the height of DAPI cells

was unknown and we therefore made an assumption that they were spherical in order to approximate the height. This increases the likelihood that our measurements of density may have been biased.

Additionally, the Abercrombie correction factor does not correct for a problem known as lost-caps (Hedreen, 1998). Close to both surfaces, where the section was cut, there is a pronounced reduction in the number of neuronal nucleoli (Andersen and Gundersen, 1999). It is thought that this loss of nucleoli is simply due to the cells being opened up when the tissue is cut. As a result, the true cell density may actually be higher than what was reported in our study.

A problem may arise when one group has larger cells, given that this group would likely be more affected by lost-caps and a lower cell density would be reported. Given that the DAPI size was similar across the groups, this effect would not have a large impact on our findings.

Another possible limitation to this study was that the sections might not be sliced completely perpendicular to the cortical surface. This is a potential limitation that is beyond our control, given that the samples were sent to us as slides with mounted brain tissue. If the density of cells in the tangential direction differs from the perpendicular direction, than any slices that were not cut completely perpendicular would have untrue densities reported. For instance, if neurons were more densely packed perpendicularly, than a slanted cut would decrease the density measured. It is reasonable to infer that any slight deviations would affect all of the patient samples and that overall this would not affect the results. However, we also chose to measure relative density in case of differences between the tissue cuts.

A limitation that is somewhat unique to our study is that our methods do not allow us to detect changes in the expression of our markers. Our study places a threshold on the image, where pixels greater that a certain level are considered a cell. As mentioned earlier, a limitation

to this was that it was unclear whether CUX2 and ZNF312 were expressed at different levels in the different cortical layers. A possible solution to this would be to make the threshold stricter, such that only certain cells are counted as a unit. However, this would increase the subjectivity of the methods and it would be difficult to establish a threshold that could be used across all of the different samples. Another similar limitation was seen in cases where overall density was decreased, as this finding may reflect poorer staining as opposed to reduced cell counts.

Although we manually inspected each slide to ensure proper staining, it is difficult to determine minor differences between the samples. This is another reason why we chose to measure relative density, in case there were differences in the overall staining intensity between the samples.

Finally, there were several confounds that we were unable to control for. Using

Pearson’s r correlation, we found correlations between PMI, pH, days in freezer, age and our measures. These variables were therefore included as covariates in the general linear model.

Additionally, gender and cerebral hemisphere also correlated with several measures and were included in the general model as fixed variables. However, there are potentially other factors that affect cortical cytoarchitecture that could not be assessed. For instance, use of medication, especially antipsychotics, may cause reported neuropsychiatric findings in psychiatric disorders

(Galila et al., 2002). Other potential confounds that were not controlled for include education, alcohol and substance abuse and duration of illness.

5.9 Conclusion

In this study, we analyzed the orbitofrontal cortex in 60 post-mortem brain tissues from patients with schizophrenia, bipolar disorder and major depressive disorder using a novel semi- automated approach. We used antibodies against CUX2 and ZNF312, two transcription factors involved in the regulation of cortical development. After capturing the images, we manually

delineated the cortical layers and then used an automated protocol for the analysis. ImageJ was used to automatically segment the cells, which added to the objectivity of the methods. Finally, a MATLAB algorithm was used to automatically calculate cell density, cell area, cortical thickness and distance from pia. The methods were validated using neuron-specific nuclear antigen NeuN, in order to determine the number of CUX2 and ZNF312 cells that overlapped with NeuN. Additionally, delineation of cortical layers was also performed on Nissl stained slides, in order to compare measures of cortical thickness using our markers versus traditional methods.

Overall, we found few significant differences in schizophrenia, major depressive disorder and bipolar disorder compared to controls. We hypothesized that schizophrenia is associated with neuronal migration deficits and that this would be observed in post-mortem brain tissue. We did not find any evidence to support this hypothesis in schizophrenia. However, we did find other significant findings, including a few that were consistent among the orbitofrontal cortex subareas. In schizophrenia, there was a significant decrease in the relative density of ZNF312 cells in layer V in BA47l. In bipolar disorder, there was also a significant decrease in the relative density of ZNF312 cells and CUX2 cells in layer V in BA47l. There was also a significant decrease in the relative density of CUX2 cells in layer VI when analyzing the entire orbitofrontal cortex. Finally, in major depressive disorder, there was also a significant decrease in the relative density of ZNF312 cells and CUX2 cells in layer V in BA47l. These findings suggest that in psychiatric disorders, there may be a decrease in the relative density of cells in lower layers compared to controls.

Overall, this study was significant as it was the first to analyze the cytoarchitecture of the entire area of the OFC. Firstly, the cell markers allowed for the clear delineation of cortical

layers. Next, using an automated counting protocol, we were able to investigate larger areas of the OFC than previously investigated. We believe that our findings suggest that there are subtle abnormalities in the cytoarchitecture in these psychiatric disorders, however inconsistencies arise in the literature by sampling only a small area of the cortex in a limited number of subjects.

We believe that other researchers could benefit from the automated analysis method, which would ultimately improve the comparability of results between different studies and resolve inconsistencies in the literature.

100

Chapter 6: Future Directions

Firstly, we plan on using these methods to analyze other areas of the cortex in psychiatric disorders as well as other disorders. The Neuropathology Consortium of the Stanley

Medical Research Institute has several different brain sections widely available for research purposes, including areas of the forebrain, hindbrain and midbrain. One area we plan on investigating is the superior temporal cortex, since abnormalities in this area have been previously associated with psychiatric disorders. For example, a recent study found a series of genes differentially expressed in the superior temporal cortex of schizophrenia patients, including changes to cell functions, apoptosis and cell motility (Sellmann et al., 2014). These findings suggest that differences may also be observed at the cytoarchitectural level.

Future studies could also use the automated counting protocols to analyze disorders with known cytoarchitectural abnormalities. For example, lissencephaly is a neuronal-migration disorder and abnormalities would likely be more prominent than those seen in psychiatric disorders (Moon and Wynshaw-Boris, 2013). By analyzing entire regions of the cortex, we would be able to further characterize these disorders and assess which regions of the cortex are most affected. Additionally, these disorders could be used as a positive-control to test our automated counting protocol and ensure that we are able to detect differences when they are present.

We could also improve our methods by automating the delineation of the cortical layers.

Currently, we are manually demarcating the cortical layers using Photoshop CS6. However, this increases the subjectivity of the methods, especially given that different researchers may divide the cortical layers slightly differently. In our study, to decrease the subjectiveness, we had only one student delineate the cortical layers for all of the subjects (KT). However, this is the most

101

labor-intensive step and thus if it was automated, we would increase the efficiency of our protocol while also enhancing the objectivity. This may be difficult, given that the each cortical sample tends to have slightly different cytoarchitectural features. However, this may be an interesting project for a bioinformatics student and would increase the overall objectivity of the methods.

We also plan to better characterize the CUX2 and ZNF312 cells investigated. These markers have been well characterized in rodent cortical development, however there is a lack of characterization in the human adult cortex. Firstly, it is important to determine why only 70-

80% of the cells overlapped with NeuN. As previously mentioned in Section 5.2, it may be possible that CUX2 and ZNF312 are not solely expressed neurons. In order to better characterize the cell populations, the markers should also be co-stained with a glial cell marker, such as glial fibrillary acidic protein (GFAP). If the remaining 20% of cells co-stained with a glial cell marker, it would be confirmed that our cells are not exclusively neurons.

In order to further characterize these cells, it would also be necessary to co-stain with an interneuron marker, such as glutamic acid decarboxylase 1 (Gad67) (Taniguchi, 2014).

Additionally, calcium-binding proteins, such as parvalbumin, calbindin and calretinin can be used as marker for different subpopulations of interneurons (Fujise et al., 1995; Mikkonen et al.,

1997). These markers would allow us to better characterize the types of cells that express CUX2 and ZNF312 and would ultimately increase the understanding of our automated counting protocol. Finally, in order to better understand how CUX2 and ZNF312 are expressed throughout the cortex, it would be useful to quantify layer-specific protein expression. These findings would allow us to test our prediction that CUX2 and ZNF312 are expressed throughout all of the cortical layers, however at different levels.

102

Finally, in order to eliminate the limitations of analyzing two-dimensional sections, it would be beneficial to develop an automated counting protocol that could be used on three- dimensional subsections. One problem inherent to two-dimensional histological analyses is that density measurements are typically biased, since cell centroids that are primarily outside the thickness of the slide are still counted as a unit if part of the cell extends into the slide. Although we did correct for this problem by using an Abercrombie correction factor, our estimates of density were likely still biased given that the cells were assumed to be spherical for the purpose of the calculation. When using three-dimensional sections and analyses, there would no longer be a need for the Abercrombie correction factor and measurements of density would be more accurate.

Overall, there are several benefits to analyzing three-dimensional subsections. We would be able to analyze large areas of the cortex using automated counting methods, while also removing the major limitations of analyzing thin sections of the cortex. However, it is important to note that three-dimensional subsections are not as readily available as thin sections and the project may not be as applicable. However, if possible, the creation of an automated counting protocol for three-dimensional subsections would be an extremely novel and interesting thesis project for future graduate students.

103

Chapter 7: References

Abbass, M.H., 2014. Cytoarchitecture of the Anterior Cingulate Cortex in Patients with Schizophrenia, Bipolar Disorder and Major Depression. University of Toronto .

Abdolmaleky, H.M., Cheng, K., Russo, A., Smith, C.L., Faraone, S. V., Wilcox, M., Shafa, R., Glatt, S.J., Nguyen, G., Ponte, J.F., Thiagalingam, S., Tsuang, M.T., 2005. Hypermethylation of the reelin (RELN) promoter in the brain of schizophrenic patients: A preliminary report. Am. J. Med. Genet. Part B Neuropsychiatr. Genet. 134B, 60–66.

Abercrombie, M., 1946. Estimation of nuclear population from microtome sections. Anat. Rec. 94, 239–247.

Allen Institute for Brain Science, 2017. Allen Brain Atlas [WWW Document]. URL http://www.brain-map.org/

Altshuler, L., Bookheimer, S., Townsend, J., Proenza, M.A., Sabb, F., Mintz, J., Cohen, M.S., 2008. Regional brain changes in bipolar I depression: a functional magnetic resonance imaging study. Bipolar Disord. 10, 708–717.

American Psychiatric Association, 2013. Diagnostic and Statistical Manual of Mental Disorders, Fifth Edition, American Psychiatric Association. Arlington, VA.

American Psychiatric Association, 1994. Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition, American Psychiatric Association. Washington, DC.

Andersen, B.B., Gundersen, H.J.G., 1999. Pronounced loss of cell nuclei and anisotropic deformation of thick sections. J. Microsc. 196, 69–73.

Anderson, S.A., Eisenstat, D.D., Shi, L., Rubenstein, J.L., 1997. Interneuron migration from basal forebrain to neocortex: dependence on Dlx genes. Science 278, 474–6.

Anderson, S.A., Marín, O., Horn, C., Jennings, K., Rubenstein, J.L., 2001. Distinct cortical migrations from the medial and lateral ganglionic eminences. Development 128, 353–63.

Angevine, J.B., Sidman, R.L., 1961. Autoradiographic study of cell migration during

104

histogenesis of cerebral cortex in the mouse. Nature 192, 766–768.

Anthony, T.E., Klein, C., Fishell, G., Heintz, N., 2004. Radial glia serve as neuronal progenitors in all regions of the central nervous system. Neuron 41, 881–890.

Arion, D., Unger, T., Lewis, D.A., Mirnics, K., 2007. Molecular markers distinguishing supragranular and infragranular layers in the human prefrontal cortex. Eur. J. Neurosci. 25, 1843–1854.

Arlotta, P., Molyneaux, B.J., Chen, J., Inoue, J., Kominami, R., Macklis, J.D., 2005. Neuronal Subtype-Specific Genes that Control Corticospinal Motor Neuron Development In Vivo. Neuron 45, 207–221.

Arnold, S.E., Franz, B.R., Gur, R.C., Gur, R.E., Shapiro, R.M., Moberg, P.J., Trojanowski, J.Q., 1995. Smaller neuron size in schizophrenia in hippocampal subfields that mediate cortical- hippocampal interactions. Am. J. Psychiatry 152, 738–748.

Arnold, S.E., Trojanowski, J.Q., Gur, R.E., Blackwell, P., Han, L.Y., Choi, C., 1998. Absence of neurodegeneration and neural injury in the cerebral cortex in a sample of elderly patients with schizophrenia. Arch. Gen. Psychiatry 55, 225–232.

Atkinson, M.E., 2013. Anatomy for Dental Students, 4th ed. Oxford University Press.

Baaré, W.F., Hulshoff Pol, H.E., Hijman, R., Mali, W.P., Viergever, M.A., Kahn, R.S., 1999. Volumetric analysis of frontal lobe regions in schizophrenia: relation to cognitive function and symptomatology. Biol. Psychiatry 45, 1597–605.

Baars, B.J., Gage, N.M., 2010. Cognition, Brain, and Consciousness: Introduction to Cognitive Neuroscience, 2nd ed. Academic Press.

Bakhshi, K., Chance, S.A., 2015. The neuropathology of schizophrenia: A selective review of past studies and emerging themes in brain structure and cytoarchitecture. Neuroscience 303, 82–102.

Baldessarini, R.J., Hegarty, J.D., Bird, E.D., Benes, F.M., 1997. Meta-analysis of postmortem studies of Alzheimer’s disease-like neuropathology in schizophrenia. Am. J. Psychiatry 154,

105

861–863.

Beck, E., 1949. A cytoarchitectural investigation into the boundaries of cortical areas 13 and 14 in the human brain. J. Anat. 83, 147–57.

Bellani, M., Cerruti, S., Brambilla, P., 2010. Orbitofrontal cortex abnormalities in schizophrenia. Epidemiol. Psichiatr. Soc. 19, 23–25.

Benes, F.M., Matzilevich, D., Burke, R.E., Walsh, J., 2006. The expression of proapoptosis genes is increased in bipolar disorder, but not in schizophrenia. Mol. Psychiatry 11, 241– 251.

Benes, F.M., Sorensen, I., Bird, E.D., 1991. Reduced neuronal size in posterior hippocampus of schizophrenic patients. Schizophr. Bull. 17, 597–608.

Benes, F.M., Vincent, S.L., Todtenkopf, M., 2001. The density of pyramidal and nonpyramidal neurons in anterior cingulate cortex of schizophrenic and bipolar subjects. Biol. Psychiatry 50, 395–406.

Berry, M., Rogers, A.W., 1965. The migration of neuroblasts in the developing cerebral cortex. J. Anat. 99, 691–709.

Blumberg, H.P., Stern, E., Ricketts, S., Martinez, D., de Asis, J., White, T., Epstein, J., Isenberg, N., McBride, P.A., Kemperman, I., Emmerich, S., Dhawan, V., Eidelberg, D., Kocsis, J.H., Silbersweig, D.A., 1999. Rostral and orbital prefrontal cortex dysfunction in the manic state of bipolar disorder. Am. J. Psychiatry 156, 1986–8.

Brandon, N.J., Millar, J.K., Korth, C., Sive, H., Singh, K.K., Sawa, A., 2009. Understanding the role of DISC1 in psychiatric disease and during normal development. J. Neurosci. 29, 12768–12775.

Bremner, J.D., Vythilingam, M., Vermetten, E., Nazeer, A., Adil, J., Khan, S., Staib, L.H., Charney, D.S., 2002. Reduced volume of orbitofrontal cortex in major depression. Biol. Psychiatry 51, 273–279.

Brisch, R., Saniotis, A., Wolf, R., Bielau, H., Bernstein, H.G., Steiner, J., Bogerts, B., Braun, K.,

106

Braun, A.K., Jankowski, Z., Kumaratilake, J., Kumaritlake, J., Henneberg, M., Gos, T., 2014. The role of dopamine in schizophrenia from a neurobiological and evolutionary perspective: old fashioned, but still in vogue. Front. psychiatry 5, 47.

Brodmann, K., 1914. Physiologie des Gehirns. Druck Deutsche Verlagsgesellschaft, Stuttgart.

Buchwalow, I., Samoilova, V., Boecker, W., Tiemann, M., 2011. Non-specific binding of antibodies in immunohistochemistry: fallacies and facts. Sci. Rep. 1, 28.

Budday, S., Steinmann, P., Kuhl, E., 2015. Physical biology of human brain development. Front. Cell. Neurosci. 9.

Camargo, L.M., Collura, V., Rain, J.C., Mizuguchi, K., Hermjakob, H., Kerrien, S., Bonnert, T.P., Whiting, P.J., Brandon, N.J., 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.

Carmichael, S.T., Price, J.L., 1995. Sensory and premotor connections of the orbital and medial prefrontal cortex of macaque monkeys. J. Comp. Neurol. 363, 642–664.

Chana, G., Landau, S., Beasley, C., Everall, I.P., Cotter, D., 2003. Two-dimensional assessment of cytoarchitecture in the anterior cingulate cortex in major depressive disorder, bipolar disorder, and schizophrenia: evidence for decreased neuronal somal size and increased neuronal density. Biol. Psychiatry 53, 1086–98.

Chen, B., Schaevitz, L.R., McConnell, S.K., 2005. Fezl regulates the differentiation and axon targeting of layer 5 subcortical projection neurons in cerebral cortex. Proc. Natl. Acad. Sci. 102, 17184–17189.

Chen, C.H., Suckling, J., Ooi, C., Jacob, R., Lupson, V., Bullmore, E.T., Lennox, B.R., 2010. A longitudinal fMRI study of the manic and euthymic states of bipolar disorder. Bipolar Disord. 12, 344–347.

Chiavaras, M.M., Petrides, M., 2000. Orbitofrontal sulci of the human and macaque monkey brain. J. Comp. Neurol. 422, 35–54.

107

Christison, G.W., Casanova, M.F., Weinberger, D.R., Rawlings, R., Kleinman, J.E., 1989. A quantitative investigation of hippocampal pyramidal cell size, shape, and variability of orientation in schizophrenia. Arch. Gen. Psychiatry 46, 1027–32.

Clapcote, S.J., Lipina, T. V., Millar, J.K., Mackie, S., Christie, S., Ogawa, F., Lerch, J.P., Trimble, K., Uchiyama, M., Sakuraba, Y., Kaneda, H., Shiroishi, T., Houslay, M.D., Henkelman, R.M., Sled, J.G., Gondo, Y., Porteous, D.J., Roder, J.C., 2007. Behavioral phenotypes of Disc1 missense mutations in mice. Neuron 54, 387–402.

Corfas, G., Roy, K., Buxbaum, J.D., 2004. Neuregulin 1-erbB signaling and the molecular/cellular basis of schizophrenia. Nat. Neurosci. 7, 575–580.

Costa, E., Davis, J., Pesold, C., Tueting, P., Guidotti, A., 2002. The heterozygote reeler mouse as a model for the development of a new generation of antipsychotics. Curr. Opin. Pharmacol. 2, 56–62.

Cotter, D., Hudson, L., Landau, S., 2005. Evidence for orbitofrontal pathology in bipolar disorder and major depression, but not in schizophrenia. Bipolar Disord. 7, 358–369.

Cotter, D., Mackay, D., Chana, G., Beasley, C., Landau, S., Everall, I.P., 2002. Reduced neuronal size and glial cell density in area 9 of the dorsolateral prefrontal cortex in subjects with major depressive disorder. Cereb. Cortex 12, 386–94.

Creutzfeldt, O.D., 1995. Cortex Cerebri. Oxford University Press, New York, NY.

Cubelos, B., Sebastián-Serrano, A., Beccari, L., Calcagnotto, M.E., Cisneros, E., Kim, S., Dopazo, A., Alvarez-Dolado, M., Redondo, J.M., Bovolenta, P., Walsh, C.A., Nieto, M., 2010. Cux1 and Cux2 regulate dendritic branching, spine morphology, and synapses of the upper layer neurons of the cortex. Neuron 66, 523–535.

Cubelos, B., Sebastian-Serrano, A., Kim, S., Moreno-Ortiz, C., Redondo, J.M., Walsh, C.A., Nieto, M., 2008. Cux-2 controls the proliferation of neuronal intermediate precursors of the cortical subventricular zone. Cereb. Cortex 18, 1758–1770.

Cullen, T.J., Walker, M.A., Eastwood, S.L., Esiri, M.M., Harrison, P.J., Crow, T.J., 2006. Anomalies of asymmetry of pyramidal cell density and structure in dorsolateral prefrontal

108

cortex in schizophrenia. Br. J. Psychiatry 188, 26–31.

D’Arcangelo, G., Miao, G.G., Chen, S.C., Soares, H.D., Morgan, J.I., Curran, T., 1995. A protein related to extracellular matrix proteins deleted in the mouse mutant reeler. Nature 374, 719–723. da Silva Alves, F., Figee, M., van Amelsvoort, T., Veltman, D., de Haan, L., 2008. The revised dopamine hypothesis of schizophrenia: evidence from pharmacological MRI studies with atypical antipsychotic medication. Psychopharmacol. Bull. 41, 121–32. de Graaf-Peters, V.B., Hadders-Algra, M., 2006. Ontogeny of the human central nervous system: What is happening when? Early Hum. Dev. 82, 257–266.

DeFelipe, J., Hendry, S.H.C., Jones, E.G., 1989. Synapses of double bouquet cells in monkey cerebral cortex visualized by calbindin immunoreactivity. Brain Res. 503, 49–54.

Dégenètais, E., Thierry, A.M., Glowinski, J., Gioanni, Y., 2002. Electrophysiological properties of pyramidal neurons in the rat prefrontal cortex: an in vivo intracellular recording study. Cereb. Cortex 12, 1–16. del Rio, M.R., DeFelipe, J., 1997. Double bouquet cell axons in the human temporal neocortex: relationship to bundles of myelinated axons and colocalization of calretinin and calbindin D-28k immunoreactivities. J. Chem. Neuroanat. 13, 243–251.

Donkelaar, H.J. ten, Lammens, M., Hori, A., 2014. Clinical neuroembryology : development and developmental disorders of the human central nervous system. Springer.

Drevets, W.C., 2007. Orbitofrontal cortex function and structure in depression. Ann. N. Y. Acad. Sci. 1121, 499–527.

Drevets, W.C., 1999. Prefrontal Cortical-Amygdalar Metabolism in Major Depression. Ann. N. Y. Acad. Sci. 877, 614–637.

Eastwood, S.L., Harrison, P.J., 2005. Interstitial white matter neuron density in the dorsolateral prefrontal cortex and parahippocampal gyrus in schizophrenia. Schizophr. Res. 79, 181–8.

Eckler, M.J., Nguyen, T.D., McKenna, W.L., Fastow, B.L., Guo, C., Rubenstein, J.L.R., Chen,

109

B., 2015. Cux2-positive radial glial cells generate diverse subtypes of neocortical projection neurons and macroglia. Neuron 86, 1100–1108.

Eker, C., Simsek, F., Yılmazer, E.E., Kitis, O., Cinar, C., Eker, O.D., Coburn, K., Gonul, A.S., 2014. Brain regions associated with risk and resistance for bipolar I disorder: a voxel-based MRI study of patients with bipolar disorder and their healthy siblings. Bipolar Disord. 16, 249–261.

Ellison-Wright, I., Glahn, D.C., Laird, A.R., Thelen, S.M., Bullmore, E., 2008. The anatomy of first-episode and chronic schizophrenia: an anatomical likelihood estimation meta-analysis. Am. J. Psychiatry 165, 1015–23.

Fatemi, S.H., Earle, J.A., McMenomy, T., 2000. Reduction in Reelin immunoreactivity in hippocampus of subjects with schizophrenia, bipolar disorder and major depression. Mol. Psychiatry 5, 654–63, 571.

Fatemi, S.H., Folsom, T.D., 2009. The neurodevelopmental hypothesis of schizophrenia, revisited. Schizophr. Bull. 35, 528–548.

Flynn, J.T., 1991. Strabismus A Neurodevelopmental Approach. Springer, New York, NY.

Fornito, A., Yücel, M., Pantelis, C., 2009. Reconciling neuroimaging and neuropathological findings in schizophrenia and bipolar disorder. Curr. Opin. Psychiatry 22, 312–319.

Franco, S.J., Gil-Sanz, C., Martinez-Garay, I., Espinosa, A., Harkins-Perry, S.R., Ramos, C., Muller, U., 2012. Fate-restricted neural progenitors in the mammalian cerebral cortex. Science 337, 746–749.

Frantz, G.D., McConnell, S.K., 1996. Restriction of late cerebral cortical progenitors to an upper-layer fate. Neuron 17, 55–61.

Fujise, N., Hunziker, W., Heizmann, C.W., Kosaka, T., 1995. Distribution of the calcium binding proteins, calbindin D-28K and parvalbumin, in the subicular complex of the adult mouse. Neurosci. Res. 22, 89–107.

Fuster, J.M., 1997. The prefrontal cortex : anatomy, physiology, and neuropsychology of the

110

frontal lobe, 3rd ed. Lippincott Williams and Wilkins.

Galila, A., Everall, I.P., Belmaker, R.H., 2002. The postmortem brain in psychiatric research. Springer Science & Business Media.

Garey, L.J., Ong, W.Y., Patel, T.S., Kanani, M., Davis, A., Mortimer, A.M., Barnes, T.R., Hirsch, S.R., 1998. Reduced dendritic spine density on cerebral cortical pyramidal neurons in schizophrenia. J. Neurol. Neurosurg. Psychiatry 65, 446–53.

Garey, L.J., Von Bussmann, K.A., Hirsch, S.R., 2006. Decreased numerical density of kainate receptor-positive neurons in the orbitofrontal cortex of chronic schizophrenics. Exp. brain Res. 173, 234–242.

Ghashghaei, H.T., Lai, C., Anton, E.S., 2007. Neuronal migration in the adult brain: are we there yet? Nat. Rev. Neurosci. 8, 141–151.

Glantz, L.A., Lewis, D.A., 2000. Decreased dendritic spine density on prefrontal cortical pyramidal neurons in schizophrenia. Arch. Gen. Psychiatry 57, 65–73.

Goes, F.S., Willour, V.L., Zandi, P.P., Belmonte, P.L., MacKinnon, D.F., Mondimore, F.M., Schweizer, B., National Institute of Mental Health Genetics Initiative Bipolar Disorder Consortium, N.I. of M.H.G.I.B.D., DePaulo, J.R., Gershon, E.S., McMahon, F.J., Potash, J.B., Potash, J.B., 2010. Sex-specific association of the Reelin gene with bipolar disorder. Am. J. Med. Genet. B. Neuropsychiatr. Genet. 153B, 549–53.

Götz, M., Huttner, W.B., 2005. The cell biology of neurogenesis. Nat. Rev. Mol. Cell Biol. 6, 777–788.

Grayson, D.R., Jia, X., Chen, Y., Sharma, R.P., Mitchell, C.P., Guidotti, A., Costa, E., 2005. Reelin promoter hypermethylation in schizophrenia. Proc. Natl. Acad. Sci. 102, 9341–9346.

Green, I.W., Glausier, J.R., 2016. Different Paths to Core Pathology: The Equifinal Model of the Schizophrenia Syndrome. Schizophr. Bull. 42, 542–549.

Greig, L.C., Woodworth, M.B., Galazo, M.J., Padmanabhan, H., Macklis, J.D., 2013. Molecular logic of neocortical projection neuron specification, development and diversity. Nat. Rev.

111

Neurosci. 14, 755–769.

Guidotti, A., Auta, J., Davis, J.M., Di-Giorgi-Gerevini, V., Dwivedi, Y., Grayson, D.R., Impagnatiello, F., Pandey, G., Pesold, C., Sharma, R., Uzunov, D., Costa, E., DiGiorgi Gerevini, V., 2000. Decrease in reelin and glutamic acid decarboxylase67 (GAD67) expression in schizophrenia and bipolar disorder: a postmortem brain study. Arch. Gen. Psychiatry 57, 1061–1069.

Guo, C., Eckler, M.J., McKenna, W.L., McKinsey, G.L., Rubenstein, J.L.R., Chen, B., 2013. Fezf2 expression identifies a multipotent progenitor for neocortical projection neurons, astrocytes, and oligodendrocytes. Neuron 80, 1167–1174.

Gupta, S., Kulhara, P., 2010. What is schizophrenia: A neurodevelopmental or neurodegenerative disorder or a combination of both? A critical analysis. Indian J. Psychiatry 52, 21–27.

Haijma, S. V., Van Haren, N., Cahn, W., Koolschijn, P.C.M.P., Hulshoff Pol, H.E., Kahn, R.S., 2013. Brain Volumes in Schizophrenia: A Meta-Analysis in Over 18 000 Subjects. Schizophr. Bull. 39, 1129–1138.

Harlan, J., Chen, Y., Gubbins, E., Mueller, R., Roch, J.M., Walter, K., Lake, M., Olsen, T., Metzger, P., Dorwin, S., Ladror, U., Egan, D.A., Severin, J., Johnson, R.W., Holzman, T.F., Voelp, K., Davenport, C., Beck, A., Potter, J., Gopalakrishnan, M., Hahn, A., Spear, B.B., Halbert, D.N., Sullivan, J.P., Abkevich, V., Neff, C.D., Skolnick, M.H., Shattuck, D., Katz, D.A., 2006. Variants in Apaf-1 segregating with major depression promote apoptosome function. Mol. Psychiatry 11, 76–85.

Harrison, P., 2004. The hippocampus in schizophrenia: a review of the neuropathological evidence and its pathophysiological implications. Psychopharmacology (Berl). 174, 151–62.

Harrison, P.J., 2000. Postmortem studies in schizophrenia. Dialogues Clin. Neurosci. 2, 349–57.

Hasler, G., 2010. Pathophysiology of depression: do we have any solid evidence of interest to clinicians? World psychiatry 9, 155–61.

Heckers, S., Stone, D., Walsh, J., Shick, J., Koul, P., Benes, F.M., 2002. Differential

112

hippocampal expression of glutamic acid decarboxylase 65 and 67 messenger RNA in bipolar disorder and schizophrenia. Arch. Gen. Psychiatry 59, 521–9.

Hedreen, J.C., 1998. Lost caps in histological counting methods. Anat. Rec. 250, 366–72.

Hering, H., Sheng, M., 2001. Dendritic spines: structure, dynamics and regulation. Nat. Rev. Neurosci. 2, 880–888.

Hevner, R.F., Daza, R.A.M., Rubenstein, J.L.R., Stunnenberg, H., Olavarria, J.F., Englund, C., 2003. Beyond laminar fate: toward a molecular classification of cortical projection/pyramidal neurons. Dev. Neurosci. 25, 139–151.

Highley, J.R., Walker, M.A., McDonald, B., Crow, T.J., Esiri, M.M., 2003. Size of hippocampal pyramidal neurons in schizophrenia. Br. J. Psychiatry 183.

Hof, P.R., Mufson, E.J., Morrison, J.H., 1995. Human orbitofrontal cortex: Cytoarchitecture and quantitative immunohistochemical parcellation. J. Comp. Neurol. 359, 48–68.

Howes, O.D., Kapur, S., 2009. The dopamine hypothesis of schizophrenia: version III--the final common pathway. Schizophr. Bull. 35, 549–62.

Impagnatiello, F., Guidotti, A.R., Pesold, C., Dwivedi, Y., Caruncho, H., Pisu, M.G., Uzunov, D.P., Smalheiser, N.R., Davis, J.M., Pandey, G.N., Pappas, G.D., Tueting, P., Sharma, R.P., Costa, E., 1998. A decrease of reelin expression as a putative vulnerability factor in schizophrenia. Proc. Natl. Acad. Sci. U. S. A. 95, 15718–15723.

Jaaro-Peled, H., Ayhan, Y., Pletnikov, M. V, Sawa, A., 2010. Review of pathological hallmarks of schizophrenia: comparison of genetic models with patients and nongenetic models. Schizophr. Bull. 36, 301–13.

Joshi, D., Fullerton, J.M., Weickert, C.S., 2014. Elevated ErbB4 mRNA is related to interneuron deficit in prefrontal cortex in schizophrenia. J. Psychiatr. Res. 53, 125–132.

Joshi, D., Fung, S.J., Rothwell, A., Weickert, C.S., 2012. Higher gamma-aminobutyric acid neuron density in the white matter of orbital frontal cortex in schizophrenia. Biol. Psychiatry 72, 725–733.

113

Kageyama, R., Yamamori, T., 2013. Cortical development: Neural diversity and neocortical organization. Springer Science & Business Media.

Kempton, M.J., Stahl, D., Williams, S.C.R., DeLisi, L.E., 2010. Progressive lateral ventricular enlargement in schizophrenia: A meta-analysis of longitudinal MRI studies. Schizophr. Res. 120, 54–62.

Khundakar, A., Morris, C., Thomas, A.J., 2011. The immunohistochemical examination of GABAergic interneuron markers in the dorsolateral prefrontal cortex of patients with late- life depression. Int. Psychogeriatr. 23, 644–653.

Kleinman, J.E., Casanova, M.F., Jaskiw, G.E., 1988. The Neuropathology of Schizophrenia 209 14.

Konarski, J.Z., McIntyre, R.S., Kennedy, S.H., Rafi-Tari, S., Soczynska, J.K., Ketter, T.A., 2008. Volumetric neuroimaging investigations in mood disorders: bipolar disorder versus major depressive disorder. Bipolar Disord. 10, 1–37.

Kringelbach, M.L., 2005. The human orbitofrontal cortex: linking reward to hedonic experience. Nat. Rev. Neurosci. 6, 691–702.

Kringelbach, M.L., Rolls, E.T., 2004. The functional neuroanatomy of the human orbitofrontal cortex: Evidence from neuroimaging and neuropsychology. Prog. Neurobiol. 72, 341–372.

KubickiI, M., McCarley, R., Westin, C., Park, H., Maier, S., Kikinis, R., Jolesz, F., Shenton, M., 2007. A review of diffusion tensor imaging studies in schizophrenia. J. Psychiatr. Res. 41, 15–30.

Kwan, K.Y., Sestan, N., Anton, E.S., 2012. Transcriptional co-regulation of neuronal migration and laminar identity in the neocortex. Development 139, 1535–1546.

Lacerda, A.L.., Keshavan, M.S., Hardan, A.Y., Yorbik, O., Brambilla, P., Sassi, R.B., Nicoletti, M., Mallinger, A.G., Frank, E., Kupfer, D.J., Soares, J.C., 2004. Anatomic evaluation of the orbitofrontal cortex in major depressive disorder. Biol. Psychiatry 55, 353–358.

Lavdas, A.A., Grigoriou, M., Pachnis, V., Parnavelas, J.G., 1999. The medial ganglionic

114

eminence gives rise to a population of early neurons in the developing cerebral cortex. J. Neurosci. 19, 7881–8.

Lavoie, S., Bartholomeuz, C.F., Nelson, B., Lin, A., McGorry, P.D., Velakoulis, D., Whittle, S.L., Yung, A.R., Pantelis, C., Wood, S.J., 2014. Sulcogyral pattern and sulcal count of the orbitofrontal cortex in individuals at ultra high risk for psychosis. Schizophr. Res. 154, 93– 99.

Law, A.J., Kleinman, J.E., Weinberger, D.R., Weickert, C.S., 2006. Disease-associated intronic variants in the ErbB4 gene are related to altered ErbB4 splice-variant expression in the brain in schizophrenia. Hum. Mol. Genet. 16, 129–141.

Lawrie, S.M., Abukmeil, S.S., 1998. Brain abnormality in schizophrenia. A systematic and quantitative review of volumetric magnetic resonance imaging studies. Br. J. Psychiatry 172, 110–20.

Lee, F.H.F., Fadel, M.P., Preston-Maher, K., Cordes, S.P., Clapcote, S.J., Price, D.J., Roder, J.C., Wong, A.H.C., 2011. Disc1 point mutations in mice affect development of the cerebral cortex. J. Neurosci. 31, 3197–3206.

Liotti, M., Mayberg, H.S., McGinnis, S., Brannan, S.L., Jerabek, P., 2002. Unmasking disease- specific cerebral blood flow abnormalities: mood challenge in patients with remitted unipolar depression. Am. J. Psychiatry 159, 1830–1840.

Liu, W.S., Pesold, C., Rodriguez, M.A., Carboni, G., Auta, J., Lacor, P., Larson, J., Condie, B.G., Guidotti, A., Costa, E., 2001. Down-regulation of dendritic spine and glutamic acid decarboxylase 67 expressions in the reelin haploinsufficient heterozygous reeler mouse. Proc. Natl. Acad. Sci. U. S. A. 98, 3477–3482.

López-Larson, M.P., DelBello, M.P., Zimmerman, M.E., Schwiers, M.L., Strakowski, S.M., 2002. Regional prefrontal gray and white matter abnormalities in bipolar disorder. Biol. Psychiatry 52, 93–100.

Malatesta, P., Hack, M.A., Hartfuss, E., Kettenmann, H., Klinkert, W., Kirchhoff, F., Götz, M., 2003. Neuronal or glial progeny: regional differences in radial glia fate. Neuron 37, 751–

115

764.

Marín, O., Rubenstein, J.L.R., 2001. A long, remarkable journey: Tangential migration in the telencephalon. Nat. Rev. Neurosci. 2, 780–790.

Marin, O., Yaron, A., Bagri, A., Tessier-Lavigne, M., Rubenstein, J.L., 2001. Sorting of Striatal and Cortical Interneurons Regulated by Semaphorin-Neuropilin Interactions. Science 293, 872–875.

McConnell, S.K., Kaznowski, C.E., 1991. Cell cycle dependence of laminar determination in developing neocortex. Science 254, 282–285.

McDonald, C., Murray, R.M., 2000. Early and late environmental risk factors for schizophrenia. Brain Res. Brain Res. Rev. 31, 130–137.

McGrath, J., Saha, S., Chant, D., Welham, J., 2008. Schizophrenia: a concise overview of incidence, prevalence, and mortality. Epidemiol. Rev. 30, 67–76.

McGrath, J.J., Murray, R.M., 2011. Environmental risk factors for schizophrenia, in: Schizophrenia, Third Edition. Wiley-Blackwell, Oxford, UK, pp. 226–244.

Miguel-Hidalgo, J.J., Whittom, A., Villarreal, A., Soni, M., Meshram, A., Pickett, J.C., Rajkowska, G., Stockmeier, C.A., 2014. Apoptosis-related proteins and proliferation markers in the orbitofrontal cortex in major depressive disorder. J. Affect. Disord. 158, 62– 70.

Mikkonen, M., Soininen, H., Pitkänen, A., 1997. Distribution of parvalbumin-, calretinin-, and calbindin-D28k-immunoreactive neurons and fibers in the human entorhinal cortex. J. Comp. Neurol. 388, 64–88.

Miklowitz, D.J., Johnson, S.L., 2006. The Psychopathology and Treatment of Bipolar Disorder. Annu. Rev. Clin. Psychol. 2, 199–235.

Millar, J.K., Wilson-Annan, J.C., Anderson, S., Christie, S., Taylor, M.S., Semple, C.A., Devon, R.S., St Clair, D.M., Muir, W.J., Blackwood, D.H., Porteous, D.J., 2000. Disruption of two novel genes by a translocation co-segregating with schizophrenia. Hum. Mol. Genet. 9,

116

1415–1423.

Molyneaux, B.J., Arlotta, P., Fame, R.M., MacDonald, J.L., MacQuarrie, K.L., Macklis, J.D., 2009. Novel subtype-specific genes identify distinct subpopulations of callosal projection neurons. J. Neurosci. 29, 12343–12354.

Molyneaux, B.J., Arlotta, P., Hirata, T., Hibi, M., Macklis, J.D., 2005. Fezl Is required for the birth and specification of corticospinal motor neurons. Neuron 47, 817–831.

Molyneaux, B.J., Arlotta, P., Menezes, J.R.L., Macklis, J.D., 2007. Neuronal subtype specification in the cerebral cortex. Nat. Rev. Neurosci. 8, 427–437.

Moon, H.M., Wynshaw-Boris, A., 2013. Cytoskeleton in action: lissencephaly, a neuronal migration disorder. Wiley Interdiscip. Rev. Dev. Biol. 2, 229–245.

Muraki, K., Tanigaki, K., 2015. Neuronal migration abnormalities and its possible implications for schizophrenia. Front. Neurosci. 9.

Nadarajah, B., Brunstrom, J.E., Grutzendler, J., Wong, R.O.L., Pearlman, A.L., 2001. Two modes of radial migration in early development of the cerebral cortex. Nat. Neurosci. 4, 143–150.

Nadarajah, B., Parnavelas, J.G., 2002. Modes of neuronal migration in the developing cerebral cortex. Nat. Rev. Neurosci. 3, 423–432.

Najt, P., Nicoletti, M., Chen, H.H., Hatch, J.P., Caetano, S.C., Sassi, R.B., Axelson, D., Brambilla, P., Keshavan, M.S., Ryan, N.D., Birmaher, B., Soares, J.C., 2007. Anatomical measurements of the orbitofrontal cortex in child and adolescent patients with bipolar disorder. Neurosci. Lett. 413, 183–186.

Nakamura, M., Nestor, P.G., McCarley, R.W., Levitt, J.J., Hsu, L., Kawashima, T., Niznikiewicz, M., Shenton, M.E., 2007. Altered orbitofrontal sulcogyral pattern in schizophrenia. Brain 130, 693–707.

Nieto, M., Monuki, E.S., Tang, H., Imitola, J., Haubst, N., Khoury, S.J., Cunningham, J., Gotz, M., Walsh, C.A., 2004. Expression of Cux-1 and Cux-2 in the subventricular zone and

117

upper layers II-IV of the cerebral cortex. J. Comp. Neurol. 479, 168–180.

Nieuwenhuys, R., Voogd, J., Huijzen, C. van., 2008. The human central nervous system, 4th ed. Springer.

Noctor, S.C., Martínez-Cerdeño, V., Ivic, L., Kriegstein, A.R., 2004. Cortical neurons arise in symmetric and asymmetric division zones and migrate through specific phases. Nat. Neurosci. 7, 136–144.

Nullmeier, S., Panther, P., Dobrowolny, H., Frotscher, M., Zhao, S., Schwegler, H., Wolf, R., 2011. Region-specific alteration of GABAergic markers in the brain of heterozygous reeler mice. Eur. J. Neurosci. 33, 689–698.

O’Connell, P., Woodruff, P.W.R., Wright, I., Jones, P., Murray, R.M., 1997. Developmental insanity or dementia praecox: was the wrong concept adopted? Schizophr. Res. 23, 97–106.

O’Rahilly, R., Müller, F., 1987. Developmental stages in human embryos: including a revision of Streeter’s “Horizons” and a survey of the Carnegie Collection. Carnegie Institution of Washington.

Ochoa, S., Usall, J., Cobo, J., Labad, X., Kulkarni, J., 2012. Gender differences in schizophrenia and first-episode psychosis: a comprehensive literature review. Schizophr. Res. Treatment 2012.

Olivier, C., Cobos, I., Perez Villegas, E.M., Spassky, N., Zalc, B., Martinez, S., Thomas, J.L., 2001. Monofocal origin of telencephalic oligodendrocytes in the anterior entopeduncular area of the chick embryo. Development 128, 1757–69.

Öngür, D., Ferry, A.T., Price, J.L., 2003. Architectonic subdivision of the human orbital and medial prefrontal cortex. J. Comp. Neurol. 460, 425–449.

Organization, W.H., 2008. The Global Burden of Disease: 2004 Update. WHO Press.

Owen, F., Crow, T.J., 1987. Neurotransmitters and psychosis. Brituh Med. BulUtw 43, 651–671.

Pandya, D., Petrides, M., Cipolloni, P.B., 2015. Cerebral cortex : architecture, connections, and the dual origin concept. Oxford University Press.

118

Pang, T., Atefy, R., Sheen, V., 2008. Malformations of cortical development. Neurologist 14, 181–191.

Petrides, M., Pandya, D.N., 2002. Comparative cytoarchitectonic analysis of the human and the macaque ventrolateral prefrontal cortex and corticocortical connection patterns in the monkey. Eur. J. Neurosci. 16, 291–310.

Pierri, J.N., Volk, C.L., Auh, S., Sampson, A., Lewis, D.A., 2001. Decreased somal size of deep layer 3 pyramidal neurons in the prefrontal cortex of subjects with schizophrenia. Arch. Gen. Psychiatry 58, 466–73.

Pleasure, S.J., Anderson, S., Hevner, R., Bagri, A., Marin, O., Lowenstein, D.H., Rubenstein, J.L., 2000. Cell migration from the ganglionic eminences is required for the development of hippocampal GABAergic interneurons. Neuron 28, 727–40.

Pontious, A., Kowalczyk, T., Englund, C., Hevner, R.F., 2008. Role of intermediate progenitor cells in cerebral cortex development. Dev. Neurosci. 30, 24–32.

Prata, D.P., Breen, G., Osborne, S., Munro, J., St Clair, D., Collier, D.A., 2009. An association study of the neuregulin 1 gene, bipolar affective disorder and psychosis. Psychiatr. Genet. 19, 113–6.

Preibisch, S., Saalfeld, S., Tomancak, P., 2009. Globally optimal stitching of tiled 3D microscopic image acquisitions. Bioinformatics 25, 1463–5.

Przedborski, S., Vila, M., Jackson-Lewis, V., 2003. Neurodegeneration: what is it and where are we? J. Clin. Invest. 111, 3–10.

Rajkowska, G., Halaris, A., Selemon, L.D., 2001. Reductions in neuronal and glial density characterize the dorsolateral prefrontal cortex in bipolar disorder. Biol. Psychiatry 49, 741– 52.

Rajkowska, G., Miguel-Hidalgo, J.J., Dubey, P., Stockmeier, C.A., Krishnan, K.R.R., 2005. Prominent reduction in pyramidal neurons density in the orbitofrontal cortex of elderly depressed patients. Biol. Psychiatry 58, 297–306.

119

Rajkowska, G., Miguel-Hidalgo, J.J., Wei, J., Dilley, G., Pittman, S.D., Meltzer, H.Y., Overholser, J.C., Roth, B.L., Stockmeier, C.A., 1999. Morphometric evidence for neuronal and glial prefrontal cell pathology in major depression. Biol. Psychiatry 45, 1085–1098.

Rajkowska, G., Selemon, L.D., Goldman-Rakic, P.S., 1998. Neuronal and glial somal size in the prefrontal cortex: a postmortem morphometric study of schizophrenia and Huntington disease. Arch. Gen. Psychiatry 55, 215–24.

Ribolsi, M., Daskalakis, Z.J., Siracusano, A., Koch, G., 2014. Abnormal asymmetry of brain connectivity in schizophrenia. Front. Hum. Neurosci. 8, 1010.

Rivas, R.J., Hatten, M.E., 1995. Motility and cytoskeletal organization of migrating cerebellar granule neurons. J. Neurosci. 15, 981–9.

Rolls, E.T., 2016. Cerebral cortex : principles of operation, First. ed. Oxford University Press.

Rupp, C.I., Fleischhacker, W.W., Kemmler, G., Kremser, C., Bilder, R.M., Mechtcheriakov, S., Szeszko, P.R., Walch, T., Scholtz, A.W., Klimbacher, M., Maier, C., Albrecht, G., Lechner- Schoner, T., Felber, S., Hinterhuber, H., 2005. Olfactory functions and volumetric measures of orbitofrontal and limbic regions in schizophrenia. Schizophr. Res. 74, 149–161.

Sanides, F., 1969. Comparative Architectonics of the Neocortex of Mammals and Their Evolutionary Interpretation. Ann. N. Y. Acad. Sci. 167, 404–423.

Sarkar, A., Chunder, R., Sen, S., Sunandita, N., 2016. Sulcogyral pattern of orbitofrontal cortex in human brain and its clinical implications. Indian J. Basic Appl. Med. Res. 5, 433–442.

Schmitz, C., Eastwood, B.S., Tappan, S.J., Glaser, J.R., Peterson, D.A., Hof, P.R., 2014. Current automated 3D cell detection methods are not a suitable replacement for manual stereologic cell counting. Front. Neuroanat. 8, 27.

Schosser, A., Cohen-Woods, S., Gaysina, D., Chow, P.C., Martucci, L., Farmer, A., Korszun, A., Gunashinghe, C., Gray, J., Jones, L., Craddock, N., Owen, M.J., Craig, I.W., McGuffin, P., 2010. NRG1 gene in recurrent major depression: No association in a large-scale case- control association study. Am. J. Med. Genet. Part B Neuropsychiatr. Genet. 153B, 141–7.

120

Schubert, D., Kötter, R., Zilles, K., Luhmann, H.J., Staiger, J.F., 2003. Cell type-specific circuits of cortical layer IV spiny neurons. J. Neurosci. 23, 2961–2970.

Sei, Y., Ren-Patterson, R., Li, Z., Tunbridge, E.M., Egan, M.F., Kolachana, B.S., Weinberger, D.R., 2007. Neuregulin1-induced cell migration is impaired in schizophrenia: association with neuregulin1 and catechol-o-methyltransferase gene polymorphisms. Mol. Psychiatry 12, 946–957.

Selemon, L.D., Goldman-Rakic, P.S., 1999. The reduced neuropil hypothesis: a circuit based model of schizophrenia. Biol. Psychiatry 45, 17–25.

Selemon, L.D., Rajkowska, G., Goldman-Rakic, P.S., 1998. Elevated neuronal density in prefrontal area 46 in brains from schizophrenic patients: application of a three-dimensional, stereologic counting method. J. Comp. Neurol. 392, 402–12.

Selemon, L.D., Rajkowska, G., Goldman-Rakic, P.S., 1995. Abnormally high neuronal density in the schizophrenic cortex. A morphometric analysis of prefrontal area 9 and occipital area 17. Arch. Gen. Psychiatry 52, 805–18; discussion 819–20.

Sellmann, C., Pildaín, L.V., Schmitt, A., Leonardi-Essmann, F., Durrenberger, P.F., Spanagel, R., Arzberger, T., Kretzschmar, H., Zink, M., Gruber, O., Herrera-Marschitz, M., Reynolds, R., Falkai, P., Gebicke-Haerter, P.J., Matthäus, F., 2014. Gene expression in superior temporal cortex of schizophrenia patients. Eur. Arch. Psychiatry Clin. Neurosci. 264, 297– 309.

Shad, M.U., Muddasani, S., Keshavan, M.S., 2006. Prefrontal subregions and dimensions of insight in first-episode schizophrenia — A pilot study. Psychiatry Res. Neuroimaging 146, 35–42.

Shelton, R.C., Claiborne, J., Sidoryk-Wegrzynowicz, M., Reddy, R., Aschner, M., Lewis, D.A., Mirnics, K., 2011. Altered expression of genes involved in inflammation and apoptosis in frontal cortex in major depression. Mol. Psychiatry 16, 751–762.

Shen, L.H., Liao, M.H., Tseng, Y.C., 2012. Recent Advances in Imaging of Dopaminergic Neurons for Evaluation of Neuropsychiatric Disorders. J. Biomed. Biotechnol. 2012, 1–14.

121

Silberberg, G., Darvasi, A., Pinkas-Kramarski, R., Navon, R., 2006. The involvement ofErbB4 with schizophrenia: Association and expression studies. Am. J. Med. Genet. Part B Neuropsychiatr. Genet. 141B, 142–148.

Sommer, I., Ramsey, N., Kahn, R., Aleman, A., Bouma, A., 2001. Handedness, language lateralisation and anatomical asymmetry in schizophrenia: meta-analysis. Br. J. Psychiatry 178, 344–51.

Steen, R.G., Mull, C., McClure, R., Hamer, R.M., Lieberman, J.A., 2006. Brain volume in first- episode schizophrenia: Systematic review and meta-analysis of magnetic resonance imaging studies. Br. J. Psychiatry 188, 510–518.

Stockmeier, C.A., Mahajan, G.J., Konick, L.C., Overholser, J.C., Jurjus, G.J., Meltzer, H.Y., Uylings, H.B.M., Friedman, L., Rajkowska, G., 2004. Cellular changes in the postmortem hippocampus in major depression. Biol. Psychiatry 56, 640–50.

Sullivan, P.F., Daly, M.J., O’Donovan, M., 2012. Genetic architectures of psychiatric disorders: the emerging picture and its implications. Nat. Rev. Genet. 13, 537–551.

Sweet, R.A., Bergen, S.E., Sun, Z., Sampson, A.R., Pierri, J.N., Lewis, D.A., 2004. Pyramidal cell size reduction in schizophrenia: evidence for involvement of auditory feedforward circuits. Biol. Psychiatry 55, 1128–1137.

Szeszko, P.R., Bilder, R.M., Lencz, T., Ashtari, M., Goldman, R.S., Reiter, G., Wu, H., Lieberman, J.A., 2000. Reduced anterior cingulate gyrus volume correlates with executive dysfunction in men with first-episode schizophrenia. Schizophr. Res. 43, 97–108.

Tan, S.S., Kalloniatis, M., Sturm, K., Tam, P.P.., Reese, B.E., Faulkner-Jones, B., 1998. Separate Progenitors for Radial and Tangential Cell Dispersion during Development of the Cerebral Neocortex. Neuron 21, 295–304.

Taniguchi, H., 2014. Genetic dissection of GABAergic neural circuits in mouse neocortex. Front. Cell. Neurosci. 8, 8.

Tochigi, M., Iwamoto, K., Bundo, M., Komori, A., Sasaki, T., Kato, N., Kato, T., 2008. Methylation Status of the Reelin Promoter Region in the Brain of Schizophrenic Patients.

122

Biol. Psychiatry 63, 530–533.

Uribe, E., Wix, R., 2012. Migración neuronal, apoptosis y trastorno bipolar. Rev. Psiquiatr. Salud Ment. 5, 127–133.

Uylings, H.B.M., Sanz-Arigita, E.J., de Vos, K., Pool, C.W., Evers, P., Rajkowska, G., 2010. 3- D Cytoarchitectonic parcellation of human orbitofrontal cortex. Correlation with postmortem MRI. Psychiatry Res. - Neuroimaging 183, 1–20.

Uylings, H.B.M., Zilles, K., Rajkowska, G., 1999. Optimal Staining Methods for Delineation of Cortical Areas and Neuron Counts in Human Brains. Neuroimage 9, 439–445. van Aerde, K.I., Feldmeyer, D., 2015. Morphological and physiological characterization of pyramidal neuron subtypes in rat medial prefrontal cortex. Cereb. Cortex 25, 788–805.

Van der Schot, A., Kahn, R., Ramsey, N., Nolen, W., Vink, M., 2010. Trait and state dependent functional impairments in bipolar disorder. Psychiatry Res. Neuroimaging 184, 135–142.

Van Horn, J.D., McManus, I.C., 1992. Ventricular enlargement in schizophrenia. A meta- analysis of studies of the ventricle:brain ratio (VBR). Br. J. Psychiatry 160, 687–97.

Vita, A., De Peri, L., Silenzi, C., Dieci, M., 2006. Brain morphology in first-episode schizophrenia: A meta-analysis of quantitative magnetic resonance imaging studies. Schizophr. Res. 82, 75–88. von Economo, C.F., Koskinas, G.N., 1925. Die Cytoarchitektonik der Hirnrinde des erwachsenen Menschen. Springer.

Waldvogel, H.J., Curtis, M.A., Baer, K., Rees, M.I., Faull, R.L.M., 2006. Immunohistochemical staining of post-mortem adult human brain sections. Nat. Protoc. 1, 2719–2732.

Webb, C.A., Weber, M., Mundy, E.A., Killgore, W.D.S., 2014. Reduced gray matter volume in the anterior cingulate, orbitofrontal cortex and thalamus as a function of mild depressive symptoms: a voxel-based morphometric analysis. Psychol. Med. 44, 2833–2843.

Weickert, C.S., Tiwari, Y., Schofield, P.R., Mowry, B.J., Fullerton, J.M., 2012. Schizophrenia- associated HapICE haplotype is associated with increased NRG1 type III expression and

123

high nucleotide diversity. Transl. Psychiatry 2, e104.

Wen, S., Li, H., Liu, J., 2009. Dynamic signaling for neural stem cell fate determination. Cell Adh. Migr. 3, 107–117.

Wen, Z., Chen, J., Khan, R.A.W., Song, Z., Wang, M., Li, Z., Shen, J., Li, W., Shi, Y., 2016. Genetic association between NRG1 and schizophrenia, major depressive disorder, bipolar disorder in Han Chinese population. Am. J. Med. Genet. Part B Neuropsychiatr. Genet. 171B, 468–478.

Wilke, M., Kowatch, R.A., DelBello, M.P., Mills, N.P., Holland, S.K., 2004. Voxel-based morphometry in adolescents with bipolar disorder: first results. Psychiatry Res. Neuroimaging 131, 57–69.

Williams, M.R., Chaudhry, R., Perera, S., Pearce, R.K.B., Hirsch, S.R., Ansorge, O., Thom, M., Maier, M., 2013. Changes in cortical thickness in the frontal lobes in schizophrenia are a result of thinning of pyramidal cell layers. Eur. Arch. Psychiatry Clin. Neurosci. 263, 25– 39.

Woeffler-Maucler, C., Beghin, A., Ressnikoff, D., Bezin, L., Marinesco, S., 2014. Automated immunohistochemical method to quantify neuronal density in brain sections: Application to neuronal loss after status epilepticus. J. Neurosci. Methods 225, 32–41.

Wright, I.C., Rabe-Hesketh, S., Woodruff, P.W.R., David, A.S., Murray, R.M., Bullmore, E.T., 2000. Meta-Analysis of Regional Brain Volumes in Schizophrenia. Am. J. Psychiatry 157, 16–25.

Yamasue, H., Iwanami, A., Hirayasu, Y., Yamada, H., Abe, O., Kuroki, N., Fukuda, R., Tsujii, K., Aoki, S., Ohtomo, K., Kato, N., Kasai, K., 2004. Localized volume reduction in prefrontal, temporolimbic, and paralimbic regions in schizophrenia: an MRI parcellation study. Psychiatry Res. Neuroimaging 131, 195–207.

Zaidel, D.W., Esiri, M.M., Harrison, P.J., 1997. Size, shape, and orientation of neurons in the left and right hippocampus: investigation of normal asymmetries and alterations in schizophrenia. Am. J. Psychiatry 154, 812–818.

124

Zecevic, N., Chen, Y., Filipovic, R., 2005. Contributions of cortical subventricular zone to the development of the human cerebral cortex. J. Comp. Neurol. 491, 109–122.

Zilles, K., 2004. Architecture of the Human Cerebral Cortex: Regional and Laminar Organization, in: The Human Nervous System. pp. 997–1055.

Zimmer, C., Tiveron, M.C., Bodmer, R., Cremer, H., 2004. Dynamics of Cux2 expression suggests that an early pool of SVZ precursors is fated to become upper cortical layer neurons. Cereb. Cortex 14, 1408–1420.

125

Chapter 8: Appendix

Appendix 1: Derivation of Equation 3.

Abbass (2014) derived Equation 3 when using the automated counting protocol to analyze the

ACC. In brief, cells that are not fully within the section but extend into the slice will still be counted as a unit. However, given that a portion of the cell is outside of the slice, it may be cut and therefore the measured diameter would be minimized. Equation 3 was created in order to counteract for this problem. If a cell is within the section, the measured diameter (dM) equals the true diameter (dT). However, if a cell is outside the section, then one needs to estimate what the measured diameter would be. The mean cross-sectional diameter of a circle sectioned randomly is equal to the true diameter multiplied by π/4. In the case of cells that extend into the section, the dM would equal approximately dT (π/4). Therefore, the overall measured diameter equals the true diameter of cells within the thickness of the section (T/(T+H)) plus the true diameter of the cells outside of the section (H/(T+H)) multiplied by π/4.

dM = T (dT ) + dT (dT) π T + dT T + dT 4

dM (T + dT ) = (dT) (T + dT π) 4

2 T dM + dM dT = TdT + π dT 4

2 π dT + T dT = dM dT - T dM = 0 4

2 π dT + dT (T + dM) - T dM = 0 4

2 dT = - (T - dM) + √((T - dM) – 4(π/4)(-T dM)) 2(π/4)

126

2 dT = dM - T + √((T - dM) – 4(π/4)(-T dM)) π/2

We only take the positive square root, given that the negative square root gives a nonsensical answer. Therefore, the final derived equation is:

2 dT = 2( dM – T) + √((T - dM) – πT dM)) π

127