REVIEW

Schizophrenia as a Disorder of Molecular Pathways

Szatmár Horváth and Károly Mirnics

Over the last decade, transcriptome studies of postmortem tissue from subjects with schizophrenia revealed that synaptic,mitochondrial, immune system, gamma-aminobutyric acidergic, and oligodendrocytic changes are all integral parts of the diseaseprocess. The combined genetic and transcriptomic studies argue that the molecular underpinnings of the disease are even more variedthan the symptomatic diversity of schizophrenia. Ultimately, to decipher the pathophysiology of human disorders in general, we willneed to understand the function of hundreds of genes and regulatory elements in our genome and the consequences of theiroverexpression and reduced expression in a developmental context. Furthermore, integration of knowledge from various data sourcesremains a monumental challenge that has to be systematically addressed in the upcoming decades. In the end, our success ininterpreting the molecular changes in schizophrenia will depend on our ability to understand the biology using innovative ideas andcannot depend on the hope of developing novel, more powerful technologies.

Key Words: convergence, gene expression, gene network,genetics, postmortem, schizophrenia

Gene expression changes in the postmortem brains ofsubjects with schizophrenia have been studied for manydecades. These studies initially included low throughput

methodologies, encompassing northern blotting and in situhybridization and were expanded more recently to quantitativepolymerase chain reaction, DNA microarrays, and RNA sequenc-ing (RNAseq) (1). Over the last half century, a tremendous amountof data have been generated, yet our understanding of geneexpression changes in schizophrenia is still limited. Reasons forthis incomplete knowledge are complex and include bothdisease-related factors and technical limitation. Schizophrenia isa spectrum disorder rather than a single diagnosis (2) and theetiology of the disease is complex, encompassing both environ-mental and genetic factors (3). Substance abuse is quite commonin the patient population (4), and various comorbidities andlifestyle differences also have a strong effect on the findings. Thepicture is further complicated by limited availability of postmor-tem material, postmortem interval, medication history, circum-stances of death, and the disease progression between diseaseonset and the brain harvest (5). In addition, considering thepotential cohort biases and very small sample sizes, differences inexperimental methodology, and the diverse data analyticalmethods used, it is perhaps not surprising that transcriptomefindings often do not replicate across the investigated cohorts (6).

Still, while most of the single-gene expression changes inschizophrenia have poor reproducibility across studies, data-driven approaches have been able to provide us with a morereproducible list of gene expression network disruptions that arerelated to schizophrenia. While the cascade of causality remainsuncertain, it appears that synaptic (7,8), mitochondrial (9,10),immune system (6,11), gamma-aminobutyric acid (GABA)ergic(12,13), and oligodendrocytic (14,15) messenger RNA (mRNA)changes are all integral parts of the disease process (1,16).However, it is important to point out that not all patients show

From the Department of Psychiatry (SH, KM), and Vanderbilt KennedyCenter for Research on Human Development (KM), Vanderbilt Uni-versity, Nashville, Tennessee; and Department of Psychiatry (SH, KM),University of Szeged, Szeged, Hungary.

Address correspondence to Károly Mirnics, M.D., Ph.D., Vanderbilt Uni-versity, Department of Psychiatry, 8128 MRB III, 465 21st AvenueSouth, Nashville, TN 37232; E-mail: karoly.mirnics@vanderbilt.edu.

Received Oct 3, 2013; revised Jan 2, 2014; accepted Jan 5, 2014.

0006-3223/$36.00http://dx.doi.org/10.1016/j.biopsych.2014.01.001

deficits in all molecular domains—there is a clear molecularsubstratification of patients (6,11)—and that synaptic, immune,oligodendrocytic, GABAergic, or other, etiologically diverse proc-esses might give rise to the same behavioral disturbance. Thus,schizophrenia is not a disease of a single molecular pathway—rather, transcriptome changes argue for the existence of predom-inantly synaptic, oligodendrocytic, and multiple other molecularsubtypes of schizophrenia (17) that sort along a continuum in acomplex, partially overlapping pattern: each subject with schizo-phrenia might show a dominant deficit in one of the moleculardomains, yet the overall molecular deficit might also encompasselements from other molecular pathways.

The current review is focused on mRNA changes; however, it isclear that other, noncoding RNA species also play a critical role inregulating gene expression and appear to significantly contributeto the disease process of schizophrenia (18,19).

Small Signals in Genetics Versus Strong Signals inTranscriptome

Postmortem gene expression studies are typically performedon dozens of brains, while genome-wide association studies(GWAS) include thousands of patient samples. To date, GWASidentified a number of genetic elements that predispose toschizophrenia (20–23). It appears that two different but interre-lated mechanisms are at work: common alleles conferring small,cumulative risk to the disease through single nucleotide poly-morphisms (SNPs) and low-frequency large effect structuralchromosomal abnormalities known as copy-number variants(CNVs). Yet, common SNPs with relatively small effect sizes thatcan only partially explain the strong heritability of the disease(20–24) and defining a CNV as causal to the disease is even morechallenging. In contrast, postmortem gene expression studiesreveal much stronger disease-associated signals: even with smallsample sizes, there are well-replicated gene expression distur-bances that are present in a significant subpopulation of subjectswith schizophrenia. For example, glutamate decarboxylase1 (GAD1)/glutamate decarboxylase 67 (GAD67) underexpressionappears to be a hallmark of the illness (25), and present in themajority of the subjects with the disease—yet, this cannot beexplained by genetic susceptibility in the GAD1 gene itself—lifestyle, medication history, or other confounds. Similarly,immune system disturbances can be identified in �20% of thepostmortem brains of diseased subjects (6,11,26), but thesechanges cannot be traced back to a specific genetic predisposi-tion. Thus, this strongly suggests that gene expression changes

BIOL PSYCHIATRY 2014;]:]]]–]]]& 2014 Society of Biological Psychiatry

2 BIOL PSYCHIATRY 2014;]:]]]–]]] S. Horváth and K. Mirnics

are a cumulative readout of different genetic susceptibilities andenvironmental insults, which converge onto common molecular(and ultimately functional cellular) pathways (5,27).

Expression quantitative trait loci (eQTL) studies of schizophre-nia also support the notion of common, converging expressionreadout of genetic vulnerabilities. Expression quantitative trait lociare genomic loci that regulate expression levels of mRNAs orproteins. In the context of schizophrenia, one would predict thatschizophrenia susceptibility alleles are enriched for those that canaffect gene expression and that eQTLs should carry more trueassociation signals. Recent studies provide strong support for thisprediction: schizophrenia susceptibility alleles are enriched forSNPs that affect gene expression in adult human brain. Further-more, higher probability eQTLs predict schizophrenia better thanthose with a lower probability for being an eQTL (21,28).

Further evidence for genome-transcriptome convergencecomes from an interesting relationship between genetic suscept-ibility found in patient DNA and gene expression changes seen inpostmortem brains of subjects. The vast majority of the schizo-phrenia susceptibility genes also show altered transcript expres-sion in the postmortem brain, even in subjects that do not harborthese particular disease-predisposing variants (5). This is true forboth the previously identified candidate gene studies and theGWAS uncovered susceptibility genes. For example, the majorhistocompatibility complex locus confers genetic susceptibility toschizophrenia (29), and transcripts originating from this cytoge-netic region also show dysregulation in postmortem transcrip-tome studies (30,31)—even in the brains with no apparentgenetic susceptibility in the same locus. Similarly, risk-associatedSNP of the alpha-1C subunit of the L-type voltage-gated calciumchannel shows mRNA expression changes in the brain of subjectswith schizophrenia (32), and a similar relationship is observed fortranscription factor 4 (33) and possibly other GWAS-uncoveredschizophrenia-predisposing genes. This can be explained by theabove-discussed pathway view of schizophrenia: there might bevarious, patient-specific disease-predisposing genetic elements inthe pathways upstream of the changed transcript. Yet, thedifferent genetic susceptibilities are likely to give rise to acommon expression change at points of molecular convergences.

Environmental Influences and Genetic VulnerabilityConverge on the Transcriptome

As mentioned above, genetic susceptibility can be stronglypotentiated by environmental factors. Increased incidence ofschizophrenia has been associated with urban lifestyle, prenatalinfections, malnutrition, adolescent cannabis abuse, perinatalhypoxia, and other factors (34). These adverse events act inconcert with genetic predisposition, and the transcriptomechanges represent a sum of gene � environment interactionsthat jointly tip the balance of the transcriptome. Ultimately, thetranscriptome changes will affect growth, axonal pathfinding,neuronal arborization, synapse formation and pruning, energymetabolism, and many other developmental and cellular proc-esses. Once the compensatory mechanisms are exceeded, thesechanges manifest themselves as a specific behavioral phenotype,which we define as the symptoms of the disease.

Many human and transgenic mouse studies demonstrate thisvery eloquently. However, while human evidence comes fromepidemiologic studies (34) and can be considered only indirectproof, the animal studies provide clear evidence that the effect ofputative schizophrenia susceptibility genes is greatly influenced

www.sobp.org/journal

by environment. Disc1 mutant mice, when exposed to earlyimmune activation or social paradigms, show more pronouncedbehavioral and molecular deficits not observed in the unchal-lenged mutants (35). In addition, a mild isolation stress affects themesocortical projection of dopaminergic neurons, but only whencombined with a relevant genetic risk for neuropsychiatricdisorders (36). Furthermore, Ifitm3�/� mice do not develop thecharacteristic cellular-molecular-behavioral phenotype aftermaternal immune activation (e.g., impaired neurite outgrowthand dendritic spine formation, diminished microtubule-associatedprotein 2 expression, altered object recognition, and exploratorybehavior), underscoring that the gene � environment interactioncan act both as detrimental and protective factors (37).

It is important to note that the environmental modification ofgenetic disease predisposition is not only related to schizophreniabut appears to be a universal theme across many neurologicaland psychiatric disorders. For example, when coupled withadverse life experiences, individuals with one or two copies ofthe short allele of the serotonin transporter promoter polymor-phism show more depressive symptoms, diagnosable depression,and suicidality than individuals homozygous for the long allele(38,39). In contrast, Alzheimer’s disease mutant mice models,when exposed to enriched environment, show significantlyreduced amyloid deposition in the brain and remarkable sparingof cognitive functions (40).

Environment Predisposes, Genetic SusceptibilitySpecifies Disease

It is well established that environmental influences protect orpredispose to disease (34). Yet, environmental factors appear tobe quite nondisease-specific in their actions. For example,physical exercise slows the progression of Alzheimer’s disease(41), Parkinson’s disease (42), Huntington’s disease (43), and manyother brain disorders. Similarly, prenatal immune activation byvarious agents appears to predispose both to schizophrenia andautism (3). Thus, most of the environmental influences could beviewed as common predisposers/protectors across various braindisorders. As a result, we hypothesize that the disease specificityand phenotypic symptoms are defined by the genetic predis-position: the same environmental insult could have no effect atall, have only a minor effect, or result in well-defined behavioralend-phenotype, often giving rise to a full-blown clinical diagnosis(Figure 1). Thus, the specific behavioral end results are likely to bedefined by the genetic susceptibility. The recent findings of theCross Disorders Group of the Psychiatric Genomics Consortium,showing that some disorders have considerable genetic covaria-tion between disorders (21), actually strengthen our hypothesis.This study basically states that there are common building blocks(genes) of multiple mental disorders. Knowing that the GWASuncovered disease-predisposing genetic elements that are quitedistinct for schizophrenia and major depression (22,44), it is likelythat specific disease predisposition is a result of a combination ofvarious genetic elements within the individual’s genome. Thus,the same genetic elements, in different arrangement, will predis-pose to different disorders.

Furthermore, it is also noteworthy that the exact timing,duration, frequency, and magnitude of the deleterious environ-mental influences have a huge influence on the ultimate out-come, as genetic predisposition and environmental influencesinteract on a developmental timeline (45,46). Although weknow that developmental insults during prenatal development

Figure 1. Environmental influences predispose to disease, genetic susceptibility specifies phenotypic manifestations of brain disorders. Variousenvironmental influences might predispose to brain disease, but they do this is a nondisease-specific fashion. The disease specificity arises from thegenetic makeup of the individual (predisposing mutations, single nucleotide polymorphisms [SNPs], or copy-number variants [CNVs]), and the gene �environment interactions together give rise to diverse symptoms. The combination of various phenotypic manifestations we classify as a disease andgroup them together into diagnoses.

S. Horváth and K. Mirnics BIOL PSYCHIATRY 2014;]:]]]–]]] 3

predispose to schizophrenia (e.g., prenatal maternal infection), itappears that there are multiple vulnerability periods that extendall the way to adolescence, as cannabis abuse during adolescencehas been associated with a greater incidence of the disease (34).Unfortunately, at the current time, we do not understand how thebehavioral phenotypes are related to the transcriptome changesproduced by the gene � environment interaction in the humanbrain—these processes can only be partially modeled in animals.

Genetic Diversity and Phenotypic Similarity

The diagnosis of schizophrenia is established based onphenotypic-symptomatic manifestations of the disease (2). Whilethere is a tremendous genetic diversity across the genome ofpatients, the manifestations of the disease (established by theDSM-IV diagnostic criteria) are relatively common. The treatmentof schizophrenia today is also quite similar, regardless of theunderlying genetic diversity (47). Personalized treatment basedon genetic makeup of patients, although showing considerablepromise, remains elusive (48). However, the phenotypic similarityand use of similar therapy across patients with schizophreniasuggests that there is a strong molecular convergence at work:while the genetic � environmental pathophysiology might beunique for each patient, they feed into common disease-relatedpathways, disrupting fundamental processes that give rise toaltered information processing, emotion, and behavior. Thisconcept is strongly supported by the microanatomical disturban-ces and transcriptome signature in the postmortem brain fromsubjects with schizophrenia, where the disturbances are charac-teristic of a relatively large subgroup of the diseased subjects(6)—even when the studied cohort is too small for the detectionof the genetic predisposition to the disease. In this view,schizophrenia can be considered a disease of converging molec-ular pathways.

Hub Genes and Converging Pathways

From the perspective of disease pathophysiology, not allgenes and transcripts are created equal. Some genes encodeproteins that have hundreds of interacting partners, while othersserve a single purpose. Nonsense mutations or deletion of some

genes lead to death or a well-defined disease, while others can beremoved from the genome with no apparent deleterious effectson overall health or behavior. Thus, gene transcripts can becompared with a set of complexly arranged domino tiles withmany converging lines, where a fall of a single domino can havean enormous effect or quite small impact on the overall network.

To address this, the main foci of brain transcriptome studiestoday are the molecular pathways, rather than single genes. Thisapproach may group genes together based on known functions(e.g., synaptic release genes, transcription factors), gene-centeredpathways (e.g., tyrosine kinase B, AKT, brain-derived neurotrophicfactor, or disrupted-in-schizophrenia 1 signaling) (49), diseasepathophysiology (e.g., apolipoprotein E or presenilin 1 function),metabolic pathways (e.g., oxidative phosphorylation, pentosecycle), substrate activity (e.g., receptor tyrosine kinases,G protein-coupled receptors), intracellular location of theirencoded protein (cell membrane, cytosol, endoplasmic reticu-lum), chromosomal location (22q11-13 genes), or any otherprinciple that is appropriate for a particular study. Then, afterthe pathway analysis is performed, the investigator can makeconclusions about the preferential involvement or enrichment ofa particular group of transcripts.

These molecular pathway analyses can be performed by awide variety of commercial or custom made tools. Over the last10 years, several customizable, complementary, and free softwarepackages and approaches were developed. They were instru-mental for shifting our attention from single-gene analysis topathways assessments. In Supplement 1, we will review only afew of the commonly used bioinformatics tool sets for compre-hensive assessment of molecular differences. However, it shouldbe pointed out that our goal was only to highlight the variousstrategies and the remaining challenges and not a comprehen-sive review of the various analytical approaches and methods.

While these analytical tools contribute greatly to our under-standing of the disturbed transcriptome and altered cellularfunction, several significant challenges remain. First, as alreadymentioned, the relative weight of any gene in a particularnetwork is poorly understood. Should the so called hub genes(points of molecular convergence between various pathways)receive special consideration in these analyses, based on theirparticipation in a variety of molecular networks? Should

www.sobp.org/journal

4 BIOL PSYCHIATRY 2014;]:]]]–]]] S. Horváth and K. Mirnics

transcription factors be more heavily weighted, as they oftencontrol the expression of multiple genes and extensive down-stream molecular cascades?

Second, including many members of a single gene family in apathway can overwhelm the studied pathway, leading to bothfalse negative and false positive findings. For example, includingthe �15 synaptotagmin genes (50) in a synaptic release pathwaycould potentially mask the effect on the molecular cascade,especially if the synaptic release pathway is narrowly definedand includes relatively few gene transcripts beyond the synapto-tagmin family. Assessing the expression of such a group of geneswould represent a synaptotagmin family analysis rather than asynaptic pathway assessment, potentially leading to false con-clusions about the involvement of a synaptic pathway.

Third, how inclusive should any molecular pathway be? Tran-scripts encoding multifunctional proteins (e.g., early-immediategenes, 14-3-3 gene family) (51) could potentially influencehundreds of molecular pathways, with various effect sizes oneach of those. Including them into all the pathways they mightparticipate in might compromise pathway specificity, yet how canwe decide when they should be included or left out?

Fourth, most of the pathway analyses do not respect themicroanatomical compartmentalization and functional diversity ofthe brain cells. Yet, the different cell types often use differenttranscript networks. Most transcriptome experiments are per-formed on RNA originating from diverse cell populations (e.g.,brain region or cortical layer), which contains multiple cell-typespecific transcriptome profiles (glia, principal neurons, interneur-ons) and restricted, cell-type specific changes are washed out andblended into an artificial average for the entire sample. Forexample, while GABA receptors and glutamate decarboxylase 1transcripts can be both classified as part of the GABAergicpathway, they might not be expressed on the same cell type.Furthermore, underexpression of the same receptor can havevastly different effects on the neural network and behavior if itoccurs in the inhibitory interneurons or in principal, glutamatergicexcitatory neurons. Our analysis tools are not well suited torespect this anatomical compartmentalization. By finding bio-logically meaningful correlations, weighted correlation networkanalysis is perhaps the best suited analytical method to alleviatesome of these concerns (52,53); yet, even this approach cannotassign transcriptome changes to specific cell types of anatomicalsubstructures. As a result, cell-type specific enrichment of startingmaterial using laser dissection microscopy (54–57) and flowsorting (58,59) can take transcriptome profiling to a whole newlevel of resolution. However, these harvesting methods are laborintensive, time and resource consuming, and low throughput andhave seen only limited use in transcriptome profiling experimentsto date.

The above-mentioned four challenges still significantly limitour understanding of transcriptome data (60). We must be awareof them in performing our analyses, and as a result, we shouldalways tailor the analytical strategy to the experimental designand the scientific question pursued (1).

Methods Evolve, the Main Challenges Remain

Transcriptome profiling of human brain disorders is stillevolving, and we have an ever-expanding arsenal of toolsavailable to us (61). Northern hybridizations gave way to in situhybridization and quantitative polymerase chain reaction, whilegene expression microarrays and serial analysis of gene

www.sobp.org/journal

expression are rapidly losing ground to RNA sequencing. Whileall of these methods generate valuable, interesting, and techni-cally correct data, significant challenges remain.

First, it appears that our main challenge is biological and nottechnical. Schizophrenia is a very broad diagnosis, as it describesa spectrum rather than a well-defined disease entity (2). Thecombined genetic and transcriptomics studies argue thatthe molecular underpinnings might be as heterogeneous as theclinical presentations among patients, but this will have to befurther validated on large-scale studies. Thus, it is not surprisingthat gene expression findings of schizophrenia often do notreplicate across different cohorts of subjects (6). Categorizingschizophrenia into more homogenous biological and geneticconstructs has been a major challenge: the clinical subclassifica-tion efforts have not yet resulted in the discovery of clearlypredictive (central or peripheral) biomarkers of the disease (62).

The second challenge of molecular profiling is access tomeaningful biological material, and postmortem studies aretypically performed on dozens of samples, potentially biasingthe outcome of such studies. Postmortem brain tissue is a limited,nonrenewable resource, obtained several decades after the onsetof the disease. The obtained transcriptome profiles reflect acombination of the disease process, progress of illness, post-mortem interval, lifestyle changes, comorbidity, treatment ofdisease, individual variability, and many other factors (5). Incontrast, peripheral biomarkers, while readily accessible andbetter controlled, do not clearly capture the disease process thatprimarily occurs in the brain. Induced pluripotent stem cells andolfactory cells from patients are much better suited for under-standing the disease mechanisms than assessing biochemicalanalyses, and the initial studies suggest that they hold a greatpromise for deciphering disease pathophysiology (63,64). Yet, atthe current time, this is a low throughput technology withconsiderable limitations: schizophrenia is also about connectivitybetween cells, various cell types, lamination patterns, and manyother factors that are extremely challenging to study usingin vitro systems.

What Does the Future Hold?

In the end, understanding the biology and integration of dataare as critical as developing novel, more powerful technologies.While novel technologies clearly open the door to new, excitingdiscoveries, diseases are about disturbed function, superimposedon individual variability. We can find enrichment in CNVs inpatients with schizophrenia, identify predisposing SNPs, sequencethe DNA of all the individuals on the planet, and describetranscriptome disturbances associated with the disease, but allthis knowledge might not be sufficient: such knowledge alonewill provide us very limited understanding of the underlyingpathophysiology. Our ultimate understanding of the disease willpartially come from novel analytical approaches [such as mappingde novo mutations on human brain transcriptome networksthroughout development (65)], combined with targeted, func-tional, and mechanistic experiments. Ultimately, to decipher thedisease process of schizophrenia, we will need to understand thefunction of hundreds of genes and regulatory elements in ourgenome and the consequences of their overexpression andunderexpression in a developmental context. Furthermore, wewill need to integrate the genetic-transcriptomic-epigenetic-proteomic knowledge with each other and with all other datasources, including (but not limited to) imaging, anatomic,

S. Horváth and K. Mirnics BIOL PSYCHIATRY 2014;]:]]]–]]] 5

electrophysiologic, and epidemiologic information. This is theonly way that we can understand the phenotypic diversity thatcharacterizes schizophrenia and only this approach will lead totrue personalized diagnosis and treatment.

These are monumental tasks and they will take many years tocomplete, yet there is reason for optimism. The CommonMindConsortium is a multi-institutional academic, industrial nonprofitpartnership that has a goal to generate and analyze large-scaledata from human subjects with neuropsychiatric and neurodeve-lopmental disorders. The consortium will make these data and theanalytical results broadly available to the public as a free resource,and this joint endeavor has a potential to revolutionize ourunderstanding of the molecular pathophysiology of schizophrenia(www.commonmind.org). Importantly, due to a common analysisplatform, standardized methodology, uniform processing, andlarge sample size, this approach will eliminate many of the false-positive findings and inconsistencies that are found in the literatureto date. Thus, the findings might allow a true molecular character-ization of postmortem brains into different subgroups of patients,ultimately relating them to the underlying endophenotypesassociated with the illness. Furthermore, the National Institute ofMental Health funded Brain Research through Advancing Innova-tive Neurotechnologies Initiative and Human Connectome Project(mapping the neural pathways that underlie human brain function)(66–68), the Neurotherapeutics Network (http://www.neuroscienceblueprint.nih.gov/bpdrugs/—a pipeline between academic andindustry drug development research), The Neuroimaging Infor-matics Tools and Resources Clearinghouse (http://www.nitrc.org/),and other large-scale initiatives from the Allen Institute (http://www.alleninstitute.org) have a potential to be even more trans-formative than the Human Genome Project (69), but only if theknowledge from the various projects is properly integrated. Wherewill this all lead? We do not know for certain, and this is the beautyof research. Three decades ago, it was inconceivable that we wouldever be able to sequence the whole human genome. To quoteAgent K from the Men in Black movie (1997), “1,500 years ago,everybody knew that the Earth was the center of the universe. 500years ago, everybody knew that the Earth was flat. And 15 minutesago, you knew that humans were alone on this planet. Imaginewhat you’ll know tomorrow.” So, we just have to keep on workingand the new discoveries and breakthroughs will come. We are in itfor the long haul.

KM’s work is supported by National Institute of Mental HealthGrants R01MH067234 and R01 MH079299.

We are grateful to Martin Schmidt for helpful suggestions andedits of the manuscript. We are also grateful for the thoughtful andconstructive comments of the reviewers.

The authors report no biomedical financial interests or potentialconflicts of interest.

Supplementary material cited in this article is available online athttp://dx.doi.org/10.1016/j.biopsych.2014.01.001.

1. Horvath S, Janka Z, Mirnics K (2011): Analyzing schizophrenia by DNAmicroarrays. Biol Psychiatry 69:157–162.

2. Bhati MT (2013): Defining psychosis: The evolution of DSM-5 schizo-phrenia spectrum disorders. Curr Psychiatry Rep 15:409.

3. Michel M, Schmidt MJ, Mirnics K (2012): Immune system genedysregulation in autism and schizophrenia. Dev Neurobiol 72:1277–1287.

4. Chambers RA, Krystal JH, Self DW (2001): A neurobiological basis forsubstance abuse comorbidity in schizophrenia. Biol Psychiatry 50:71–83.

5. Mirnics K, Levitt P, Lewis DA (2006): Critical appraisal of DNA micro-arrays in psychiatric genomics. Biol Psychiatry 60:163–176.

6. Horvath S, Mirnics K (2013): Immune system disturbances in schizo-phrenia. Biol Psychiatry 75:316–323.

7. Faludi G, Mirnics K (2011): Synaptic changes in the brain of subjectswith schizophrenia. Int J Dev Neurosci 29:305–309.

8. Mirnics K, Middleton FA, Marquez A, Lewis DA, Levitt P (2000):Molecular characterization of schizophrenia viewed by microarrayanalysis of gene expression in prefrontal cortex. Neuron 28:53–67.

9. Middleton FA, Mirnics K, Pierri JN, Lewis DA, Levitt P (2002): Geneexpression profiling reveals alterations of specific metabolic pathwaysin schizophrenia. J Neurosci 22:2718–2729.

10. Clay HB, Sillivan S, Konradi C (2010): Mitochondrial dysfunction andpathology in bipolar disorder and schizophrenia. Int J Dev Neurosci 29:311–324.

11. Arion D, Unger T, Lewis DA, Levitt P, Mirnics K (2007): Molecularevidence for increased expression of genes related to immune andchaperone function in the prefrontal cortex in schizophrenia. BiolPsychiatry 62:711–721.

12. Hashimoto T, Arion D, Unger T, Maldonado-Aviles JG, Morris HM, VolkDW, et al. (2008): Alterations in GABA-related transcriptome in thedorsolateral prefrontal cortex of subjects with schizophrenia. MolPsychiatry 13:147–161.

13. Lewis DA, Hashimoto T, Volk DW (2005): Cortical inhibitory neuronsand schizophrenia. Nat Rev Neurosci 6:312–324.

14. Roussos P, Katsel P, Davis KL, Siever LJ, Haroutunian V (2012): Asystem-level transcriptomic analysis of schizophrenia using postmor-tem brain tissue samples. Arch Gen Psychiatry 69:1205–1213.

15. Hakak Y, Walker JR, Li C, Wong WH, Davis KL, Buxbaum JD, et al.(2001): Genome-wide expression analysis reveals dysregulation ofmyelination-related genes in chronic schizophrenia. Proc Natl Acad SciU S A 98:4746–4751.

16. Gavin DP, Akbarian S (2013): Epigenetic and post-transcriptionaldysregulation of gene expression in schizophrenia and relateddisease. Neurobiol Dis 46:255–262.

17. Mirnics K, Lewis DA (2001): Genes and subtypes of schizophrenia.Trends Mol Med 7:281–283.

18. van Erp TG, Guella I, Vawter MP, Turner J, Brown GG, McCarthy G, et al.(2013): Schizophrenia miR-137 locus risk genotype is associated withdorsolateral prefrontal cortex hyperactivation [published online aheadof print August 1]. Biol Psychiatry. doi:10.1016/j.biopsych.2013.06.016.

19. Beveridge NJ, Cairns MJ (2011): MicroRNA dysregulation in schizo-phrenia. Neurobiol Dis 46:263–271.

20. Gershon ES, Alliey-Rodriguez N, Liu C (2011): After GWAS: Searchingfor genetic risk for schizophrenia and bipolar disorder. Am J Psychiatry168:253–256.

21. Cross-Disorder Group of the Psychiatric Genomics Consortium,Genetic Risk Outcome of Psychosis (GROUP) Consortium (2013):Identification of risk loci with shared effects on five major psychiatricdisorders: A genome-wide analysis. Lancet 381:1371–1379.

22. Ripke S, O’Dushlaine C, Chambert K, Moran JL, Kahler AK, Akterin S,et al. (2013): Genome-wide association analysis identifies 13 new riskloci for schizophrenia. Nat Genet 45:1150–1159.

23. Aberg KA, Liu Y, Bukszar J, McClay JL, Khachane AN, Andreassen OA,et al. (2013): A comprehensive family-based replication study ofschizophrenia genes. JAMA Psychiatry 70:573–581.

24. Duan J, Sanders AR, Gejman PV (2010): Genome-wide approaches toschizophrenia. Brain Res Bull 83:93–102.

25. Curley AA, Arion D, Volk DW, Asafu-Adjei JK, Sampson AR, Fish KN,Lewis DA (2011): Cortical deficits of glutamic acid decarboxylase 67expression in schizophrenia: Clinical, protein, and cell type-specificfeatures. Am J Psychiatry 168:921–929.

26. Saetre P, Emilsson L, Axelsson E, Kreuger J, Lindholm E, Jazin E (2007):Inflammation-related genes up-regulated in schizophrenia brains. BMCPsychiatry 7:46.

27. Gilman SR, Chang J, Xu B, Bawa TS, Gogos JA, Karayiorgou M,Vitkup D (2013): Diverse types of genetic variation converge onfunctional gene networks involved in schizophrenia. Nat Neurosci 15:1723–1728.

28. Richards AL, Jones L, Moskvina V, Kirov G, Gejman PV, Levinson DF,et al. (2012): Schizophrenia susceptibility alleles are enriched for allelesthat affect gene expression in adult human brain. Mol Psychiatry 17:193–201.

www.sobp.org/journal

6 BIOL PSYCHIATRY 2014;]:]]]–]]] S. Horváth and K. Mirnics

29. Corvin A, Morris DW (2013): Genome-wide association studies: Find-ings at the major histocompatibility complex locus in psychosis[published online ahead of print September 29]. Biol Psychiatry.

30. Kano S, Nwulia E, Niwa M, Chen Y, Sawa A, Cascella N (2013): AlteredMHC class I expression in dorsolateral prefrontal cortex of nonsmokerpatients with schizophrenia. Neurosci Res 71:289–293.

31. Sinkus ML, Adams CE, Logel J, Freedman R, Leonard S (2013):Expression of immune genes on chromosome 6p21.3-22.1 in schizo-phrenia. Brain Behav Immun 32:51–62.

32. Bigos KL, Mattay VS, Callicott JH, Straub RE, Vakkalanka R, Kolachana B,et al. (2010): Genetic variation in CACNA1C affects brain circuitriesrelated to mental illness. Arch Gen Psychiatry 67:939–945.

33. Kim S, Cho H, Lee D, Webster MJ (2013): Association between SNPsand gene expression in multiple regions of the human brain. TranslPsychiatry 2:e113.

34. Brown AS (2010): The environment and susceptibility to schizophre-nia. Prog Neurobiol 93:23–58.

35. Cash-Padgett T, Jaaro-Peled H (2013): DISC1 mouse models as a toolto decipher gene-environment interactions in psychiatric disorders.Front Behav Neurosci 7:113.

36. Niwa M, Jaaro-Peled H, Tankou S, Seshadri S, Hikida T, Matsumoto Y,et al. (2013): Adolescent stress-induced epigenetic control of dop-aminergic neurons via glucocorticoids. Science 339:335–339.

37. Ibi D, Nagai T, Nakajima A, Mizoguchi H, Kawase T, Tsuboi D, et al.(2013): Astroglial IFITM3 mediates neuronal impairments followingneonatal immune challenge in mice. Glia 61:679–693.

38. Gonda X, Fountoulakis KN, Harro J, Pompili M, Akiskal HS, Bagdy G,Rihmer Z (2010): The possible contributory role of the S allele of5-HTTLPR in the emergence of suicidality. J Psychopharmacol 25:857–866.

39. Caspi A, Sugden K, Moffitt TE, Taylor A, Craig IW, Harrington H, et al.(2003): Influence of life stress on depression: Moderation by apolymorphism in the 5-HTT gene. Science 301:386–389.

40. Lazarov O, Robinson J, Tang YP, Hairston IS, Korade-Mirnics Z, Lee VM,et al. (2005): Environmental enrichment reduces Abeta levels andamyloid deposition in transgenic mice. Cell 120:701–713.

41. Fang Y (2013): Guiding research and practice: A conceptual model foraerobic exercise training in Alzheimer’s disease. Am J Alzheimers DisOther Demen 26:184–194.

42. Grazina R, Massano J (2013): Physical exercise and Parkinson’s disease:Influence on symptoms, disease course and prevention. Rev Neurosci24:139–152.

43. Kloos AD, Fritz NE, Kostyk SK, Young GS, Kegelmeyer DA (2013): Videogame play (Dance Dance Revolution) as a potential exercise therapy inHuntington’s disease: A controlled clinical trial. Clin Rehabil 27:972–982.

44. Major Depressive Disorder Working Group of the Psychiatric GWASConsortium, Ripke S, Wray NR, Lewis CM, Hamilton SP, Weissman MM,et al. (2012): A mega-analysis of genome-wide association studies formajor depressive disorder. Mol Psychiatry 18:497–511.

45. Horvath S, Mirnics K (2009): Breaking the gene barrier in schizophre-nia. Nat Med 15:488–490.

46. Lewis DA, Levitt P (2002): Schizophrenia as a disorder of neuro-development. Annu Rev Neurosci 25:409–432.

47. Meltzer HY (2013): Update on typical and atypical antipsychotic drugs.Annu Rev Med 64:393–406.

48. Zhang JP, Malhotra AK (2013): Pharmacogenetics of antipsychotics:Recent progress and methodological issues. Expert Opin Drug MetabToxicol 9:183–191.

49. Glorioso C, Sabatini M, Unger T, Hashimoto T, Monteggia LM,Lewis DA, Mirnics K (2006): Specificity and timing of neocortical

www.sobp.org/journal

transcriptome changes in response to BDNF gene ablation duringembryogenesis or adulthood. Mol Psychiatry 11:633–648.

50. Chapman ER (2008): How does synaptotagmin trigger neurotransmit-ter release? Annu Rev Biochem 77:615–641.

51. Middleton FA, Peng L, Lewis DA, Levitt P, Mirnics K (2005): Alteredexpression of 14-3-3 genes in the prefrontal cortex of subjects withschizophrenia. Neuropsychopharmacology 30:974–983.

52. Mirnics K (2008): What is in the brain soup? Nat Neurosci 11:1237–1238.

53. Winden KD, Oldham MC, Mirnics K, Ebert PJ, Swan CH, Levitt P, et al.(2009): The organization of the transcriptional network in specificneuronal classes. Mol Syst Biol 5:291.

54. Ruzicka WB, Zhubi A, Veldic M, Grayson DR, Costa E, Guidotti A (2007):Selective epigenetic alteration of layer I GABAergic neurons isolatedfrom prefrontal cortex of schizophrenia patients using laser-assistedmicrodissection. Mol Psychiatry 12:385–397.

55. Sodhi MS, Simmons M, McCullumsmith R, Haroutunian V, Meador-Woodruff JH (2011): Glutamatergic gene expression is specificallyreduced in thalamocortical projecting relay neurons in schizophrenia.Biol Psychiatry 70:646–654.

56. Arion D, Horvath S, Lewis DA, Mirnics K (2009): Infragranular geneexpression disturbances in the prefrontal cortex in schizophrenia:Signature of altered neural development? Neurobiol Dis 37:738–746.

57. Arion D, Unger T, Lewis DA, Mirnics K (2007): Molecular markersdistinguishing supragranular and infragranular layers in the humanprefrontal cortex. Eur J Neurosci 25:1843–1854.

58. Faux C, Rakic S, Andrews W, Yanagawa Y, Obata K, Parnavelas JG(2010): Differential gene expression in migrating cortical interneuronsduring mouse forebrain development. J Comp Neurol 518:1232–1248.

59. Nishioka M, Shimada T, Bundo M, Ukai W, Hashimoto E, Saito T, et al.(2012): Neuronal cell-type specific DNA methylation patterns of theCacna1c gene. Int J Dev Neurosci 31:89–95.

60. Mitchell AC, Mirnics K (2011): Gene expression profiling of the brain:Pondering facts and fiction. Neurobiol Dis 45:3–7.

61. Ginsberg SD, Mirnics K (2006): Functional genomic methodologies.Prog Brain Res 158:15–40.

62. Stober G, Ben-Shachar D, Cardon M, Falkai P, Fonteh AN, Gawlik M,et al. (2009): Schizophrenia: From the brain to peripheral markers. Aconsensus paper of the WFSBP task force on biological markers. WorldJ Biol Psychiatry 10:127–155.

63. Yu DX, Marchetto MC, Gage FH (2013): Therapeutic translation ofiPSCs for treating neurological disease. Cell Stem Cell 12:678–688.

64. Kano S, Colantuoni C, Han F, Zhou Z, Yuan Q, Wilson A, et al. (2013):Genome-wide profiling of multiple histone methylations in olfactorycells: Further implications for cellular susceptibility to oxidative stressin schizophrenia. Mol Psychiatry 18:740–742.

65. Gulsuner S, Walsh T, Watts AC, Lee MK, Thornton AM, Casadei S,et al. (2013): Spatial and temporal mapping of de novo mutations inschizophrenia to a fetal prefrontal cortical network. Cell 154:518–529.

66. Kandel ER, Markram H, Matthews PM, Yuste R, Koch C (2013):Neuroscience thinks big (and collaboratively). Nat Rev Neurosci 14:659–664.

67. Lichtman JW, Sanes JR (2008): Ome sweet ome: What can the genometell us about the connectome? Curr Opin Neurobiol 18:346–353.

68. Van Essen DC, Ugurbil K, Auerbach E, Barch D, Behrens TE, Bucholz R,et al. (2012): The Human Connectome Project: A data acquisitionperspective. Neuroimage 62:2222–2231.

69. Roberts L, Davenport RJ, Pennisi E, Marshall E (2001): A history of theHuman Genome Project. Science 291:1195.