Glial changes in schizophrenia: Genetic and epigenetic approach

Abstract

Background:

Schizophrenia (SCZ) is a severe mental illness that affects one percent of the population, affecting how people think, feel, and behave. Evidence suggests glial cell alteration and some researchers have found genetic risk loci and epigenetic marks that may regulate glia-related genes implicated in SCZ.

Aim:

The aim of this study is to identify genetic and epigenetic changes that have been reported in glial cells or glial-associated genes in SCZ.

Materials and Methods:

We searched the articles from PubMed, PubMed Central, Medline, Medscape, and Embase databases up to December 2020 to identify relevant peer-reviewed articles in English. The titles and abstracts were screened to eliminate irrelevant citations.

Results:

Twenty-four original articles were included in the review. Studies were categorized into the following four thematic via: (1) oligodendrocytes, (2) microglia, (3) astrocytes, and (4) perspectives.

Conclusion:

This study is the first of its kind to review research on genetic variants and epigenetic modifications associated with glia-related genes implicated in SCZ. Epigenetic evidence is considerably less than genetic evidence in this field. Understanding the pathways of some risk genes and their genetic and epigenetic regulation allows us to understand and find potential targets for future interventions in this mental illness.

INTRODUCTION

Schizophrenia (SCZ) is a chronic, life-long, and debilitating disorder characterized by the presence of psychosis, social withdrawal, unusual and uncharacteristic behavior, and cognitive impairment,[1,2,3] which could have a genetic and neurobiological background. The worldwide prevalence of this disease is about 1%. SCZ strikes men and women equally, and can affect different aspects of life, including self-care, family and social relationships, education, employment, and housing.[4] SCZ has three main symptoms: positive symptoms (i.e., auditory hallucinations and delusion), negative symptoms (i.e., impaired motivation, social withdrawal, and reduction in spontaneous speech), and cognitive impairment.[5]

There is evidence suggesting that severe alterations in the glia may contribute to the pathophysiology of SCZ. Some variations include reduced number of oligodendrocytes and low expression of myelin/oligodendrocyte-related genes, which can be linked to white matter abnormalities leading to abnormalities in the hemispheric connectivity previously reported in SCZ, as well as changes in astrocyte/oligodendrocyte populations in the white matter that may affect the structural or metabolic support of axons.[6] In addition, astrocyte and oligodendrocyte gene sets have been associated with an increased risk of SCZ.[7] It has been proposed that neuroinflammatory factors may be relevant to the pathophysiology of psychosis. The microglia (immune cells resident in the brain) are mainly associated with neuroinflammatory processes in the several brain regions of SCZ patients, and there is evidence of increased density and activation of microglia in this illness.[8]

It has been suggested that SCZ has a substantial genetic component, such that the concordance of this pathology is higher in monozygotic twins than dizygotic twins.[9] Several candidate genes are associated with this pathology, such as DISC1, DTNBP1, NRG1, and catechol-O-methyltransferase (COMT).[10] However, studies of gene variants do not necessarily indicate causal pathways of the disease and the associations found are not specific to it.[11]

In addition to the genetic factors, epigenetic mechanisms are relevant because they can regulate the expression pattern of genes involved in various pathologies. It has been seen that gene-environment interaction can be mediated by epigenetic tags and environmental risk factors can play a crucial role in developing severe psychiatric disorders.[12] Epigenetic marks linked to environmental risk factors have been suggested to explain the increased risk of SCZ.[11] Epigenetic changes are external modifications to DNA, the most common being DNA methylation, histone posttranslational modification, and transcriptional regulation by miRNAs. Interestingly, epigenetic marks are dynamic and often reversible, making them attractive therapeutic targets for treating complex diseases.[13]

MATERIALS AND METHODS

Search strategy and study selection

This study is in concordance with the five-stage scoping review framework described previously by Arksey and O’Malley.[14] This review was guided by the research question, “What genetic and epigenetic changes have been previously reported in glial cells or glial-associated genes in SCZ?” We searched articles from PubMed, PubMed Central, Medline, Medscape, and Embase databases up to December 2020. We considered relevant peer-reviewed articles in English for this review. The keywords employed for the search strategy consisted of terms associated with SCZ (psychotic or positive symptoms, negative symptoms, cognitive symptoms, first-episode SCZ, major psychosis, and paranoid SCZ), and epigenetic alterations such as DNA methylation, DNA hydroxymethylation, histone modification, and interference RNA. All articles were imported to the EndNote software (Clarivate Analytics), and duplicate reports were removed. All studies included in the analysis were case − control studies: (a) involved in a human cohort (b) medicated and drug-naïve patients with SCZ (c) monozygotic twins and concordance for SCZ (d) major psychosis and, (e) first-episode SCZ. The titles and abstracts were screened to eliminate irrelevant citations for the research question of this review. The search on review resulted in 24 records. The data were ordered in four thematic vias: (1) oligodendrocytes, (2) microglia, (3) astrocytes, and (4) perspectives.

Approved gene symbols and nomenclature were obtained from the HUGO Gene Nomenclature Committee.

A description of the studies of genetic variants associated with glia-related genes is provided in Table 1. Table 2 summarizes the main findings on epigenetic modifications. The glial-genes differential methylated sites described in epigenome-wide association studies (EWAS) are summarized in Table 3.

Table 1.

Summary of genetic variants associated in glia-related genes implicated in schizophrenia

Author	Symbol	Gene name	Gene annotation	Polymorphism	Regulation	SCZ	HC	Subgroup	Region	Ethnic group/race
Sims et al., 2009[15]	OLIG2	Oligodendrocyte transcription factor 2	Intergenic	rs762237 and rs2834072A	Putative↓	1088	1286	LOAD with psychotic symptoms	United Kingdom	Caucasian
Maeno et al., 2007[16]	SOX10	SRY-box transcription factor 10	Intron 3	rs139887B	↓	915	927	SCZ	Japan	Japanese
Tao et al., 2014[17]	ZNF804AE3E4	Zinc finger protein 804A splice variant	Intron 2	rs1344706C	↓	53	90	SCZ	United States	Caucasian (43%) African Americans (51%) Hispanic (6%)
O’Donovan et al., 2008[18]	ZNF804A	Zinc finger protein 804A	Intron 2	rs1344706D	-	479	2937	SCZ	United Kingdom	Presumably Caucasian
Kadasah et al., 2017[19]	TNF	TNF	Intergenic	rs1800629B	-	84	138	SCZ	Saudi Arabia	Saudi Arabs
Suchanek-Raif et al., 2017[20]	TNFR1	TNF receptor superfamily member 1A	Intron 1	rs4149577 and rs1860545C	Putative↑	388	657	SCZ	Poland	Polish
Pinheiro et al., 2017[21]	GSTM1/GSTT1	Glutathione S-transferase mu 1/glutathione S-transferase theta 1		Double-null genotypeE	-	54	78	Treatment-resistant schizophrenia	Brazil	Brazilian
Raffa et al., 2013[22]	GSTT1	Glutathione S-transferase theta 1		Null genotypeF	-	138	123	SCZ	Tunisia	Tunisian
Kim et al., 2008[23]	HSPA1A	Heat shock protein family A (Hsp70) member 1A	Exon 2	rs2075799G	-	294	157	SCZ	South Korea	Korean
Kowalczyk et al., 2014[24]	HSPA1A	Heat shock protein family A (Hsp70) member 1A	Exon 1	rs1043618B	-	203	243	SCZ	Poland	Polish
Bozidis et al., 2014[25]	HSPA8	Heat shock protein family A (Hsp70) member 8	Intron 1	rs1136141H	-	50	50	First-episode SCZ	Greece	Greeks
Ding et al., 2016[26]	RGS4	Regulator of G protein signaling 4	Intergenic	rs10917670, rs2661347, rs951436, and rs2661319C	↓	35	35	SCZ	United States	Caucasian (97.14%) Native Americans (1.42%) Hispanic (1.42%)

^AMassarray; ^BPolymerase chain reaction-restriction fragment length polymorphism; ^CTaqMan assay; ^DAffymetrix GeneChip 500K mapping array; ^EReal-time polymerase chain reaction (SYBR green); ^FPolymerase chain reaction; ^GPyrosequencing; ^HAmplification refractory mutation system; SCZ – Schizophrenia; HC – Healthy controls; LOAD – Late-onset Alzheimer’s disease; ↑ – Increased; ↓ – Decreased; - – No reported; TNF – Tumor necrosis factor, SYBR: SYBR green (green fluorescent cyanine dye)

Table 2.

Summary of epigenetic modifications in glia-related genes implicated in schizophrenia

Author	Symbol	Gene name	Gene annotation	Epigenetic modification	Regulation	SCZ	HC	Subgroup	Tissue	Region	Ethnic group/race
Kordi-Tamandani et al., 2014[27]	GSTP1	GSTP1: Glutathione S-transferase pi 1	Promoter	Significant DNA hypermethylationA	Putative↓	80	71	SCZ	PBC	Iran	Presumably Persians
	GSTT1	Glutathione S-transferase theta 1	Promoter	Significant DNA hypermethylationA
Venugopal et al., 2018[28]	IL6	Interleukin 6	Putative promoter	Significant DNA hypomethylationB	↑	47	47	Drug-naïve SCZ patients	PBC	India	Presumably Kannadigas
Ding et al., 2016[26]	RGS4-1	Regulator of G protein signaling 4 splice variant 1	Putative promoter	No significant DNA methylationC	↓	35	35	SCZ	DLPFC	United States	Caucasian (97.14%) Native Americans (1.42%) Hispanic (1.42%)
Iwamoto et al., 2005[29]	SOX10	SRY-box transcription factor 10	Body	Increased percentage of the methylated alleleB	↓	13	15	SCZ	PFC	United States	Presumably Caucasian
Bönsch et al., 2012[30]			Promoter	Significant relative DNA hypermethylation (promoter)D	-	20	8	Discordant twins for SCZ	PBC	Germany	Presumably Caucasian
Wu et al., 2020[31]	ZNF804A	Zinc finger protein 804A	3′-UTR	MiR-148b-3p directly targeted ZNF804A mRNA	↓	44	44	SCZ	PBC	China	Han Chinese

^AMSP; ^BBisulfite sequencing; ^CBSP; ^DMSRE. SCZ – Schizophrenia; HC – Healthy controls; DLPFC – Dorsolateral prefrontal cortex; PBC – Peripheral blood cells; ↑ – Increased; ↓ – Decreased; PCR – Polymerase chain reaction; MSP – Methylation-specific PCR; BSP – Bisulfite sequencing PCR; MSRE – Methylation-sensitive restriction enzyme

Table 3.

Summary of glia-related genes differentially methylated sites in epigenome-wide association studies in schizophrenia

Author	Symbol	Gene name	Gene annotation	CpG site/position	SCZ	HC	Subgroup	Tissue	Region	Ethnic group/race
Wockner et al., 2014[32]	SOX10	SRY-box transcription factor 10	5’UTR; Exon 1	cg23109891 and cg06614002	23	24	SCZ	FCA	United States	Undeterminate
	ZNF804A	Zinc finger protein 804A	Body	ch. 2.3731798R
Lee and Huang, 2016[33]	GNA13	G protein subunit alpha 13	Promoter	-	191	335	SCZ	DLPFC (BA46/9)A	United States	Presumably Caucasian
van Eijk et al., 2015[34]	HLA-C	Major histocompatibility complex, class I, C	Intron 4	cg16097079	260	250	SCZ	PBCB	United States	Undeterminate
Kinoshita et al., 2013[35]	GFRA2	GDNF family receptor alpha 2	Intron 1	cg22276619	24	23	Medication-free SCZ	PBCA	Japan	Japanese
	HDAC4	Histone deacetylase 4	Promoter	cg15142485
	PCM1	Pericentriolar material 1		cg25884094
	CD244	CD244 molecule	cg11939496
	FAM173A	Adenine nucleotide translocase lysine methyltransferase	Intron 2	cg09830866
	MPG	N-methylpurine DNA glycosylase	Exon 1	cg16003913
Hannon et al., 2016[36]	DDO	D-aspartate oxidase	Intron 2	cg14038731	414	433	SCZ	PBCA	United Kingdom	Presumably Caucasian
	FAM126A	Family with sequence similarity 126 member A	Promoter region	cg10311104
	IL15	Interleukin 15	Intron 3	cg11418177
	SIK3	SIK family kinase 3	Intron 1	cg23684410
	TNFAIP8	TNF alpha induced protein 8	Intron 1	cg03665078

^AInfinium humanmethylation450 BeadChip Kit (Illumina); ^BInfinium humanmethylation27 BeadChip Kit (Illumina). SCZ – Schizophrenia; HC – Healthy controls; DLPFC – Dorsolateral prefrontal cortex; FC – Frontal cortex; PBC – Peripheral blood cells

RESULTS

Oligodendrocytes

There is evidence of reported oligodendrocyte and myelin dysfunction and impaired myelination in SCZ.[37] Microarray and qPCR experiments allowed to observe a reduction of crucial oligodendrocyte-related and myelin-related genes in SCZ, such as the oligodendrocyte transcription factors OLIG1, OLIG2, and SOX10, the platelet-derived growth factor receptor alpha, myelin basic protein, myelin-associated glycoprotein, proteolipid protein 1, claudin 11, myelin oligodendrocyte glycoprotein, Erb-B2 receptor tyrosine kinase 3 and transferrin.[38] There is evidence of a significant association of psychotic symptoms in Alzheimer’s patients and two SNPs of the OLIG2 gene (rs762237 and rs2834072) in the Caucasian population.[15] Table 1 summarizes the main findings of genetic variants associated with glia-related genes. A direct correlation between the DNA methylation status of the SOX10 CpG Islands in the Y-box and their lower gene expression levels in postmortem prefrontal cortices (BA10) in SCZ participants has been reported in comparison with controls. However, these were not observed for OLIG2 or MOBP (oligodendrocytic basic protein) genes.[29] It has been proposed that Olig2 may activate the Sox10 distal enhancer in mice models.[39] Also, relative hypermethylation of the SOX10 promoter in the peripheral blood of monozygotic discordant twins for SCZ has been shown.[30] A genome-wide DNA methylation analysis performed on the frontal cortex of postmortem human brain tissue from individuals with SCZ found that cg23109891 and cg06614002 sites on SOX10 were differentially methylated in SCZ patients.[32] However, this study did not screen for the use of antipsychotic medications in study subjects. This limitation is notable because haloperidol can be associated with higher global DNA methylation in SCZ.[40] The SNP rs139887, located at Intron 3 of SOX10, was found to have a significant association with SCZ in the Japanese population, especially in male patients.[16] However, the effect of DNA methylation status and polymorphism on the expression of SOX10 in SCZ remains unclear.

The GNA13 gene that encodes the G protein subunit alpha 13 may be considered a potential biomarker of SCZ, according to genome-wide association studies studies.[41] It has been described that GNA13 may play an essential role for g-protein signaling in the development and maintenance of white matter microstructure. Transcript levels of GNA13 from untransformed lymphocytes were significantly correlated with global fractional anisotropy which reflects a combination of myelin thickness, fiber coherence, and axon integrity.[42] In addition, a DNA methylation profile study from the prefrontal cortex from schizophrenic patients and healthy controls described GNA13 as one of the ten SCZ candidate genes differentially methylated genes.[33]

Another notable feature in SCZ is the decreased density in cortical layer III oligodendrocytes and a decrease in the white matter in schizophrenic human brains;[43] this progressive reduction of white matter has been associated with greater negative SCZ symptom severity.[44] The Zinc Finger Protein 804A (ZNF804A) has been associated with brain white matter microstructure. In gestational brains, ZNF804A is highly expressed in radial glial cells and is critical for embryonic neurodevelopment.[45] It has been suggested that ZNF804A plays a role in DNA binding and transcription, and their expression regulates transcription levels of the SCZ-associated genes COMT, dopamine receptor D2, phosphodiesterase 4B, and serine protease 16 (PRSS16).[46] Variations in the ZNF804A intronic single-nucleotide polymorphism rs1344706 lead to white matter microstructural abnormalities in SCZ.[47] It has been reported that truncated ZNF804A transcript (ZNF804A^E3E4) expression was decreased in the dlPFC (dorsolateral prefrontal cortex) from patients with SCZ. Furthermore, a significant effect of rs1344706 on the ZNF804A^E3E4 splice variant was found.[17] A genome-wide association study of SCZ provided strong evidence for the association of ZNF804A as an SCZ susceptibility gene.[18] Besides, a positive association of rs1344706 with SCZ was observed in the Northern Chinese population.[48] Notably, the position ch. 2.3731798R of the ZNF804A gene was found to be differentially methylated when postmortem brains of SCZ patients and controls were compared.[32] On the other hand, a study performed in peripheral blood of patients with first-onset SCZ showed a significantly reversed expression profile of ZNF804A and miR-148b-3p. This work reveals that miR-148b-3p can regulate COMT and PRSS16 genes by targeting sites in the 3’- UTR region of ZNF804A messenger RNA (mRNA).[31]

2’,3’-Cyclic Nucleotide 3’ Phosphodiesterase (CNP) is an oligodendroglial transmembrane protein that may play a role in the oligodendrocyte function and myelination. Previous studies have shown significantly-reduced CNP mRNA levels in post-mortem tissue from the PFC (prefrontal cortex) of SCZ patients.[49] On the other hand, it has been shown that the loss of MeCP2 in the MeCP2 null mice model displayed reduced expression of CNP and other myelin-related proteins.[50] The lower-expressing A allele of the rs2070106 was significantly associated with SCZ in Caucasian populations.[51,52] However, in the Chinese Han population studies, the exonic SNP rs2070106 showed no significant association.[53] The observation may be due to the differences in the allele frequency between ethnic groups.

Microglia

Microglia are the primary innate immune cells of the brain and may play a role in the cortical circuitry disturbances reported in SCZ, such as decreased dendritic spine density on PFC pyramidal neurons in SCZ.[54] There are reports about microglial dysfunction in SCZ, suggesting a relation with the pathology of grey and white matter.[8] It has been suggested that neuroinflammation processes may be associated with white matter pathology in people with SCZ, contributing to structural and functional abnormalities observed in psychosis.[55] Furthermore, the pro-inflammatory cytokine interleukin-6 (IL-6), tumor necrosis factor-alpha (TNF-a), and interferon-gamma (IFN-g) are produced by microglial cells, which play roles in cytotoxicity and have significant effects on dopaminergic and glutamatergic pathways and cognitive processes that are implicated in SCZ.[56] Interestingly, the rs1800629 (-G308A) TNF-a gene polymorphism has been associated with SCZ.[19,57] There is evidence that the TNF receptor 1 (TNF-R1) is increased in the brain (anterior cingulate and frontal cortex) and serum of patients with SCZ.[58] Similarly, rs4149577 and rs1860545 of TNFR1 have been associated with the intensity of the excitement symptoms of paranoid SCZ in Caucasian Polish men.[20]

Increases of IL-6 mRNA are found in peripheral blood in people with SCZ compared to healthy controls[59] and in dlPFC (BA46);[60] these findings support the pro-inflammatory pathophysiological role of cytokine IL-6 in patients with SCZ. Besides, the association of IL6 rs1800795 (-174G/C) polymorphism and higher serum IL-6 with SCZ was found in the Armenian population, which might be an SCZ risk factor.[61] A study carried out in patients diagnosed with SCZ analyzed the methylation status of the IL6 promoter in antipsychotic-naïve/free patients compared with matched healthy controls from peripheral blood by bisulfite sequencing method, and a hypomethylation pattern in the IL6 promoter in SCZ was shown. Interestingly, this methylation pattern was reversed by antipsychotic administration.[28] Contrastingly, associations between several human leukocyte antigen (HLA) polymorphisms and SCZ risk were observed.[62] Another study examined differential methylation levels and genome-wide genotype data from peripheral blood of SCZ patients versus healthy controls; this research showed CpG sites of HLA-C gene differentially methylated and expressed in patients. HLA-C is located within the MHC region on chromosome 6 and belongs to the HLA class 1 molecules.[34]

Astrocytes

As previously mentioned, there is evidence of glutamate abnormalities in SCZ, the dynamic control of the glutamate uptake achieved by astrocytes regulating the activity of glutamatergic synapses.[63] An increase in the expression profiles of cortical astrocytes and decreased expression profiles of fast-spiking parvalbumin interneurons was reported in subjects with SCZ.[64] Furthermore, there is evidence of over-activation of astrocytes in SCZ.[65] Parvalbumin interneurons coordinate the optimal balance of excitation and inhibition in the PFC, maintaining the efficiency of cortical information processing.[66] Glutathione S-transferases (GSTs) are a superfamily of enzymes that quench reactive molecules with the addition of glutathione and protect the cell from oxidative damage.[67] It has been suggested that GSTs, particularly GSTM1, may mediate the astrocyte and microglia inflammation.[68] There is evidence that GSTM1 levels are significantly decreased in the post-mortem prefrontal cortex of patients with SCZ and major depressive disorder.[69] It has also been suggested that the GSTM1 null genotype (GSTM1*0) might be associated with increased susceptibility to SCZ and early-onset severe psychiatric illness.[70] Additionally, another study indicates that the double-null genotype of GSTM1 and GSTT1, which encodes GST theta 1, confers an increased risk of developing treatment-resistant schizophrenia (TRS) in Brazilian patients. There is the decreased level of antioxidants in TRS, together with an increased generation of reactive oxygen species in the brain that may damage particular areas in the brain.[21] A significant association between SCZ and the GSTT1 active genotype was observed in a Tunisian population.[22] The aim of the study was to investigate the association of the promoter methylation status of GSTT1 and GSTP1 by Methylation-specific PCR analysis of peripheral blood of patients with SCZ and healthy controls. In another study, promoter methylation frequency of GSTT and GSTP in the patients was higher, suggesting that GSTs hypermethylation may modify the risk of SCZ.[27]

Findings suggest that the HSP70 (Heat shock protein 70) mediates neuroinflammation in astrocytes.[71] The single nucleotide polymorphisms rs2075799 and rs1043618 in HSPA1A (heat shock protein family A “HSP70” member 1A), have been associated with SCZ.[23,24] At the moment, no epigenetic mechanisms have been described that may regulate the HSP70 expression in SCZ. Nonetheless, it has been reported that valproic acid (VPA) and other HDAC inhibitors (molecules that selectively alter gene transcription by chromatin remodeling), such as sodium butyrate, trichostatin A, MS-275, and apicidin, which are Class I HDAC inhibitors, may increase levels of H3K4Me2 and H3K4Me3 at the HSP70 promoter in astrocytes. H3K4me2 and H3K4 me3 are associated with transcriptionally active chromatin areas. Interestingly, VPA induced activation of the HSP70 promoter in astrocytes by recruiting histone acetyltransferase p300 in rat cortical astrocytes.[72] Besides, there is evidence of increased HSPA1A expression and other genes related to immune function, such as the pro-inflammatory mediators IFITM1, IFITM2, and IFITM3 in postmortem dlPFC cortex samples from SCZ.[73] The HSPA8 variant (rs1136141) was significantly associated with first-episode psychosis in Greek schizophrenic patients.[25]

Regulator of G-protein–signaling 4 (RGS4) is a GTPase-activating protein that plays a key role in the regulation of G-protein–coupled receptor signaling, modulating receptor-mediated neuronal signaling at the synapse.[74] A variation in the RGS4 polymorphism (rs951436) results in specific reductions in white matter structural volume.[75] There are reports regarding the downregulation of RGS4 transcripts in the dlPFC of SCZ patients compared with healthy controls, which may support this gene as a candidate gene for SCZ.[76] Another work showed that RSG4 expression levels of the longest variant RGS4-1 were decreased in the dlPFC of schizophrenic patients.[77] A study tested the methylation status of CpG islands of the RGS4 regulatory region in the postmortem dlPFC samples obtained from subjects with SCZ and healthy controls. The findings suggested that the lower RGS4-1 mRNA expression levels were not associated with hypermethylation status of its CpG islands in the 5’ region. Interestingly, the lower RGS4-1 expression was associated with SNPs rs10917670, rs2661347, rs951436, and rs2661319 in the 5’ regulatory region.[26] A study carried out by Vrajová et al. evaluated the possible epigenetic mechanism of RGS4 expression through silenced RGS4 gene using siRNA against human RGS4 and studied the effects of differential expression in neuroblastoma cell lines. They observed that downregulated RGS4 mRNA changes the expression of 67 genes, including critical transcription factors such as brain derived neurotrophic factor and DISC1 which are associated with SCZ pathology.[78]

Perspectives

EWAS explore new potential molecular targets. For example, a study carried out by Kinoshita et al. showed altered DNA methylation in peripheral blood samples from patients with SCZ, such as CpG sites located in the CpG islands in the promoter regions of putative SCZ susceptibility genes such as PCM1 (pericentriolar material 1), GFRA2 (GDNF family receptor alpha 2) and HDAC4 (histone deacetylase 4).[35] PCM1 has been reported in the centrosome in glia, and GFRA2 is localized in microglia and astrocytes.[79] Genome-wide DNA methylation analysis is critical because it opens the possibility of finding and deepening new DNA methylation-based biomarkers in SCZ. In this sense, a study showed differentially methylated CpG sites in genes such as CD244 (CD244 molecule), MPG (Methylpurine-DNA glycosylase), and FAM173A (Mitochondrial protein-lysine N-methyltransferase), hypomethylated in SCZ patients compared to controls.[80] The participation of CD244 in cortical microglia as part of the immune system,[81] MPG in cell death in astrocyte cultures in the base excision-repair system,[82] and FAM173A in mitochondrial respiration in astrocytes[83] have all been reported. Besides, another case-control study of methylomic variation associated with SCZ performed in peripheral blood showed differentially methylated positions associated with SCZ in the body of some putative genes identified in glia such as SIK3 (SIK family kinase 3), DDO (D-aspartate oxidase), family with sequence similarity 126 member A, TNF-a induced protein 8, and IL15. Those genes participate in different cellular processes such as metabolic homeostasis, catabolic process, oligodendrocytes formation, apoptosis, and pro-inflammatory response, respectively.[36] Genome-wide DNA methylation analysis is critical because it opens the possibility of finding and deepening new DNA methylation-based biomarkers in SCZ. It is necessary to investigate SNP and new rare variants, such as copy number variants,[84] which may represent an important genetic component of complex diseases, such as SCZ. However, the contribution in glial interactions has not been explored yet.

It is also essential to investigate the presence of less common epigenetic marks in those SCZ-related genes whose expression was not previously associated with promoter DNA methylation, histone modifications, or miRNA expression. The presence of chemical modifications on DNA and 5-hydroxymethylcytosine (5hmC) or N (6)-methyladenine are novel modifications found in mammalian cells. There is evidence that 5hmC levels were elevated in the inferior parietal lobule of psychotic patients. Also, increased 5hmC levels were reported at GAD1 promoter in these patients, decreasing GAD67 mRNA expression.[85]

CircRNA is a novel class of long noncoding RNA. A differential circRNA expression in postmortem dorsolateral prefrontal cortex (BA46) of SCZ patients has been suggested. Some target genes were significantly related to neurogenesis, differentiation, and synapse.[86] Human glial cells reveal that oligodendrocytes express circRNAs more abundantly than the other glial cells.[87] Furthermore, miRNAs dysregulation has the potential to be used as a novel SCZ diagnostic model. Recently, an elevation of miR-223 in peripheral blood samples of patients with first-episode SCZ has been observed. Some targets of miR-223 are cell migration-related genes such as INPP5B, and RHOB 78. It has even been described that some miRNAs can be released by extracellular vesicles of endocytic origin (exosomes), such as miR-497, described in the postmortem prefrontal cortex of SCZ patients.[88] There is evidence of exosomal regulation in astrocytes[89] and oligodendrocytes.[90] In this sense, this class of RNAs should be explored more due to their potential importance in the pathophysiology of SCZ.

DISCUSSION

Till date, the cellular and molecular mechanisms behind SCZ remain poorly understood. However, evidence suggests glial role in SCZ, such as abnormal morphological and functional maturation of oligodendrocytes and astrocytes, contributing to hypomyelination and disrupted white matter integrity. Additionally, an increased density of microglial cells might display aberrant immune responses in psychosis. The above suggests an aberrant expression of glia-related genes in psychosis, which may be defined by genetic predisposition and the influence of epigenetic mechanisms.

The sample size is an essential aspect of population genetic studies, so much that small sample sizes will often result in the identification of few polymorphisms with large effects.[91] An example of the above is the odds ratio (OR) observed by Bozidis et al., 2014, in the rs1136141 polymorphism (OR = 2.8, IC = 1. 08–9. 06) with a sample size of 50 first-episode psychosis patients and 50 healthy participants.[25] In contrast, a large sample size may improve disease prediction with sufficient statistical power. In this sense, the study performed by Maeno et al., 2007, has a modest OR rs139887 (OR = 0.83, IC = 0.72–0.95) with a sample size of 915 schizophrenic patients and 927 controls.[16] Notably, most genetic cohorts cited here have larger small sizes, except for small cohorts such as those of Pinheiro et al., 2017,[21] and Ding et al., 2016,[26] where the sample size was due to patients with specific treatment-resistant profiles and insufficient material in some samples, respectively. Nonetheless, future studies in larger sample sizes are necessary to confirm the findings.

Another critical factor to consider in genetic studies is the populations evaluated. The SNPs mentioned here were evaluated in populations with different ethnic origins, so the allelic frequency of each SNP can affect the OR for particular populations. For example, the allelic frequency observed for +190G/C polymorphism (rs1043618) in HSPA1A is different to the cohorts used by Kowalczyk et al., 2014, and Kim et al., 2008. Notably, the rs1043618 was significant in the Polish population (P = 0.0172),[24] while in the Korean population (P = 0.88) it was not.[23] Curiously, the rs1344706 was associated with lower expression of ZNF804A splice variant and SCZ, despite the different allelic frequency between the American cohort employed by Tao et al., 2014,[17] and the British subjects recruited by O’Donovan et al., 2008.[18] However, these observations may be influenced by the different sample sizes.

Notably, most of the gene variants described in this review are located in intronic regions. Intronic polymorphism can impact alternative splicing by interfering with splice site recognition.[92] An example of the above is the truncated ZNF804A transcript (ZNF804A^E3E4), which, when influenced by the psychosis risk polymorphism rs1344706 ZNF804A^E3E4, is predicted to encode a protein lacking the zinc finger domain.[17] Nevertheless, the effect of the exonic synonymous variant such as rs2075799 of HSPA1A gene, not to be underestimated due to the mechanism of exonic splicing enhancers in human genetic diseases, should be given more relevance.[93]

Epigenetic changes react to environmental stimuli and are attractive targets due to their reversibility. Notwithstanding, there are still few studies described in human glial cells. Most studies cited here were performed in peripheral blood cells, except the cohorts employed by Ding et al., 2016,[26] and Iwamoto et al., 2005,[29] performed in the prefrontal cortex. It has been described that the DNA methylation profile in the white blood cells was significantly lower in SCZ patients in contrast with controls.[40] However, the epigenetic changes in glial genes in peripheral tissue as possible blood biomarker signatures should be evaluated.

Antipsychotic drugs are known to influence DNA methylation and histone modification.[94] This review includes studies with treatment-resistant SCZ, drug-naïve patients with SCZ, and medication-free subjects with SCZ. Therefore, the psychiatric treatment and the psychiatric adherence may be a source of heterogeneity. Despite the medication-free protocol being the best because it may avoid epigenetic influence, these studies are poorly explored.

Although there are many advantages to case-control studies included here, we detected several limitations. All studies cited here were performed in the frontal cortex and peripheral tissue, but none with a cell-type-specific resolution to tackle key biological questions, maybe due to cell sorting and single-cell techniques’ high costs and technical limitations. Also, the accuracy in different methods of genotyping and epigenetic essays may be causes for such heterogeneity.

CONCLUSION

SCZ is a complex biological disorder and remains poorly understood. There is evidence that suggests a strong connection between SCZ and abnormalities in the glia cells. As shown in this review, genetic and epigenetic factors that influence the expression of glia-related genes may contribute to abnormalities observed in SCZ. Understanding genetic causes and epigenetic changes that influence this mental disorder may provide valuable inputs for a reliable diagnostic system and clinical management of SCZ.

Financial support and sponsorship

This work was supported by the National Council of Science and Technology (CONACyT) under Grant CONACyT-SALUD-S0008-2015-02.

Conflicts of interest

There are no conflicts of interest.