Elevated Viral Restriction Factor Levels in Cortical Blood Vessels in Schizophrenia

Abstract

Background

Higher tissue transcript levels of immune-related markers, including the recently discovered viral restriction factor interferon-induced transmembrane protein (IFITM) which inhibits viral entry and replication, have been reported in the prefrontal cortex in schizophrenia. Interestingly, mouse models of neuroinflammation have higher IFITM levels and deficits in GABA-related markers that are similar to findings in schizophrenia, suggesting that a shared pathogenetic process may underlie diverse cortical pathology in the disorder. However, the cell types that overexpress IFITM mRNA in schizophrenia are unknown, and it is unclear whether higher IFITM mRNA levels are associated with lower GABA-related marker levels in the same schizophrenia subjects.

Methods

We used quantitative PCR and in situ hybridization with film and grain counting analyses to quantify IFITM mRNA levels in prefrontal cortex area 9 of 57 schizophrenia and 57 healthy comparison subjects and in antipsychotic-exposed monkeys.

Results

Quantitative PCR and in situ hybridization film analysis revealed markedly elevated IFITM mRNA levels (+114% and +117%, respectively) in prefrontal gray matter in schizophrenia. Interestingly, emulsion-dipped, Nissl-stained sections from schizophrenia and comparison subjects revealed IFITM mRNA expression in pia mater and blood vessels. IFITM grain density over blood vessels was 71% higher in schizophrenia. IFITM mRNA levels were negatively correlated with GABA-related mRNAs in the same schizophrenia subjects.

Conclusions

The finding that schizophrenia subjects with higher IFITM mRNA levels in cortical blood vessels have greater disturbances in cortical GABA neurons suggests that these cell-type distinct pathological disturbances may be influenced by a shared upstream insult that involves immune activation.

Introduction

Immune- and inflammation-related abnormalities have been implicated in the pathophysiology of schizophrenia. For example, genome-wide association studies have identified associations between single-nucleotide polymorphisms in genes involved in immune and inflammatory signaling pathways and schizophrenia (1-5). Furthermore, maternal exposure to infection (6,7) and elevated serum cytokine levels during pregnancy (8) have been associated with an increased risk of schizophrenia in offspring (9). In adult schizophrenia subjects, elevated levels of proinflammatory cytokines such as IL-6 have been consistently reported in the peripheral serum (10,11), including subjects with first-episode psychosis (12), and have also been found in the prefrontal cortex (13).

Additional immune-related markers have been reported to be abnormally expressed in schizophrenia. For example, two microarray studies reported higher mRNA levels for interferon-induced transmembrane protein (IFITM) in the prefrontal cortex gray matter homogenates in schizophrenia (14,15). However, the cellular localization of IFITM in human prefrontal cortex has not been determined, and the cell types that overexpress IFITM mRNA in the prefrontal cortex in schizophrenia remain unknown. In addition, it is unclear whether higher IFITM mRNA levels in schizophrenia are related to long-term treatment with antipsychotic medications. Interestingly, more recent studies have found that IFITM plays a unique role as a viral restriction factor in preventing viral replication prior to fusion of the viral and cellular membranes (16) of many different viruses (17-22). Furthermore, IFITM expression is induced by interferons and cytokines including IL-6 (16,19,23). A mouse model of mild chronic neuroinflammation (i.e. mice with deficits in a nuclear factor-κB site-binding protein which normally suppresses production of cytokines such as IL-6) has been reported to have higher IFITM levels in cortex (24). Of particular interest, mice with chronic neuroinflammation have also been reported to have deficits in GABA neuron-related markers, including the calcium-binding protein parvalbumin and the GABA synthesizing enzyme GAD67 (24). Similar disturbances in prefrontal GABA neurons have been commonly reported in schizophrenia (25-38). However, it is unclear whether the same schizophrenia subjects that have higher IFITM mRNA levels also have deficits in GABA neuron-related markers which would suggest a shared underlying pathogenetic process that is related to immune activation.

Consequently, to further investigate the role of IFITM in the pathophysiology of schizophrenia, we first replicated the finding of higher IFITM mRNA levels in the prefrontal cortex in schizophrenia using quantitative PCR in a large new cohort (n=57) of subjects. We next used in situ hybridization to identify the cell types that overexpress IFITM mRNA in schizophrenia. We then determined whether IFITM mRNA levels are affected by antipsychotic medications by quantifying IFITM mRNA levels in the prefrontal cortex of monkeys chronically exposed to olanzapine, haloperidol, or placebo. Finally, we determined whether elevated IFITM mRNA levels are associated with larger deficits in GABA neuron-related markers in the same schizophrenia subjects which may indicate a shared pathogenetic origin in the disorder.

Methods and Materials

Human subjects

Brain specimens were obtained during routine autopsies conducted at the Allegheny County Medical Examiner's Office after consent was obtained from next-of-kin. An independent committee of experienced research clinicians made consensus DSMIV (39) diagnoses for each subject using structured interviews with family members and review of medical records (40). The absence of a psychiatric diagnosis was similarly confirmed in healthy comparison subjects. To control for experimental variance, subjects with schizophrenia or schizoaffective disorder (n=57) were matched individually to one healthy comparison subject for sex and as closely as possible for age (Supplement: Table S1) as previously described (33,40,41). Samples from subjects in a pair were processed together throughout all stages of the study. Fourteen subject pairs had previously been studied for IFITM mRNA levels by microarray (14). The mean age, postmortem interval, freezer storage time, and RNA integrity number (RIN; Agilent Bioanalyzer) did not differ between subject groups (Table 1; t₍₁₁₂₎ ≤ 1.06, p ≥ .29), and each subject had a RIN ≥ 7.0. Mean brain pH was slightly lower in schizophrenia (6.6 ± 0.3) relative to healthy comparison subjects (6.7 ± 0.2; t₍₁₁₂₎=2.09, p=.04); each subject had a brain pH ≥ 5.9. All procedures were approved by the University of Pittsburgh's Committee for the Oversight of Research Involving the Dead and Institutional Review Board.

Table 1.

Summary of demographic and postmortem characteristics of human subjects

Parameter	Healthy Comparison	Schizophrenia
N	57	57
Sex	42M / 15F	42M / 15F
Race	47W / 10B	43W / 14B
Age (years)	47.8 ± 13.9	46.9 ± 12.9
Postmortem Interval (hours)	18.4 ± 5.6	18.9 ± 8.5
Freezer Storage Time (months)	111.0 ± 51.8	109.5 ± 53.4
Brain pH	6.7 ± 0.2	6.6 ± 0.3
RNA Integrity Number	8.2 ± 0.6	8.1 ± 0.6
Medications At Time of Death
Antipsychotic	-	48/57
Antidepressant	-	25/57
Benzodiazepine/Anticonvulsant	-	21/57
NSAID	11/57	13/57
Cause of Death
Cardiopulmonary-Related	42/57	19/57
Infection/Inflammation-Related	1/57	5/57
Suicide	-	16/57
Other	14/57	17/57

For age, postmortem interval, freezer storage time, brain pH, and RNA integrity number, values are group means ± standard deviation. For medications at time of death and cause of death, number of subjects in each applicable category are provided. NSAID – non-steroidal anti-inflammatory drug.

Quantitative PCR

RNA was isolated from prefrontal cortex area 9 from each subject and used to synthesize cDNA, as previously described (26,42) (Supplemental Methods). Primer sets were designed to quantify the three relevant variants of IFITM mRNA (IFITM1, IFITM2, IFITM3; IFITM4 is a pseudogene and IFITM5 is only found in osteoblasts (16)) (Supplement: Table S2). Due to the very high sequence similarity between IFITM2 and IFITM3 in humans, one primer set targeted a homologous region in IFITM2 and IFITM3 (termed IFITM2/3; Supplement: Table S2). All primer pairs (Supplement: Table S2) demonstrated high amplification efficiency (>96%) across a wide range of cDNA dilutions, and dissociation curve analysis[notdef]of amplified products revealed melting temperatures nearly identical to that predicted by online oligonucleotide calculator software programs. Quantitative PCR was performed using the comparative cycle threshold (CT) method with Power SYBR Green dye and the ViiA-7 Real-Time PCR System (Applied Biosystems), as previously described (26,42) (Supplemental Methods). Based on their stable relative expression levels between schizophrenia and comparison subjects (42), three reference genes (beta actin, cyclophilin A, and glyceraldehyde-3-phosphate dehydrogenase [GAPDH]) were used to normalize target mRNA levels. The difference in CT (dCT) for each target transcript was calculated by subtracting the geometric mean CT for the three reference genes from the CT of the target transcript (mean of four replicate measures). Because dCT represents the log2-transformed expression ratio of each target transcript to the reference genes, the relative level of the target transcript for each subject is reported as 2^−dCT (41,43,44).

In Situ Hybridization

A 266–base pair fragment corresponding to bases 145-410 of the human IFITM2 gene (NM_006435), a region with high homology in all three human IFITM variants, was PCR-amplified, and nucleotide sequencing confirmed 100% homology for the amplified fragment to the reported IFITM2 sequence. Sense and antisense ³⁵S-labeled riboprobes were generated by in vitro transcription, and hybridization procedures were performed using 3 tissue sections from each subject. Tissue sections were then exposed to Biomax MR film (Kodak), coated with nuclear emulsion, developed, and counterstained with cresyl violet, as previously described (26,29,45) (Supplemental Methods). A Microcomputer Imaging Device (MCID) system (Imaging Research Inc, London, Ontario, Canada) was employed to measure optical density in the gray matter and white matter using film autoradiographs, as previously described (45,46) (Supplemental Methods). IFITM mRNA expression was then evaluated at the cellular level using the MCID system coupled to a microscope equipped with a motor-driven stage. Silver grains generated by the ³⁵S-labeled riboprobe in emulsion-dipped sections were counted over traced outlines of blood vessels excluding the lumen when present. Blood vessels were identified as distinct clusters of packed cells with at least two elongated nuclei and the occasional presence of a lumen or branching point, in two 500 μm-wide gray matter traverses in each of three sections per subject. The mean (SD) number of blood vessels sampled in each subject was 70 (10) for control subjects and 73 (10) for schizophrenia subjects.

Antipsychotic-exposed Monkeys

Young adult, male, long-tailed monkeys (Macaca fascicularis) received oral doses of haloperidol, olanzapine or placebo (n=6 monkeys per group) twice daily for 17–27 months, as previously described (47). RNA was isolated from prefrontal cortical area 9, and qPCR was conducted for the same three reference genes and IFITM (Supplement: Table S2) with all monkeys from a triad processed together on the same plate. All animal studies followed the NIH Guide for the Care and Use of Laboratory Animals and were approved by the Institutional Animal Care and Use Committee.

Statistical Analysis

The ANCOVA model we report includes mRNA level (i.e. normalized expression level, optical density, or grain density) as the dependent variable, diagnostic group as the main effect, subject pair as a blocking factor, and postmortem interval, brain pH, RIN, and freezer storage time as covariates. Subject pairing may be considered an attempt to account for the parallel processing of tissue samples from a pair and to balance diagnostic groups for sex and age, and not a true statistical paired design. Therefore, a second ANCOVA model without subject pair as a blocking factor and including sex and age as covariates was also used, and both models produced similar results. Subsequent analyses of differences in mRNA levels between schizophrenia subjects grouped by predictors and indicators of disease severity, psychotropic medications, nicotine use, and cause of death were conducted using the unpaired ANCOVA models with α=.05. For the antipsychotic-exposed monkey study, an ANOVA model with mRNA level as the dependent variable, treatment group as the main effect, and triad as a blocking factor was employed.

Results

Quantitative PCR Analysis of IFITM mRNA Levels in the Prefrontal Cortex in Schizophrenia

Mean mRNA levels in schizophrenia subjects were markedly higher for IFITM1 (+71%; F_(1,52)=19.0, p<.0001) and IFITM 2/3 (+114%; F_(1,52)=25.4, p<.0001) relative to healthy comparison subjects (Figure 1). Furthermore, IFITM2/3 mRNA levels were higher in the schizophrenia subject relative to the healthy comparison subject in 86% of the subject pairs, and IFITM2/3 mRNA levels were at least two-fold higher in 40% of schizophrenia subjects relative to healthy comparison subjects (Figure 1). IFITM mRNA levels were previously studied in 14 of the 57 schizophrenia subjects included in the present study (14). In the newly studied 43 schizophrenia subjects alone, we also found higher mRNA levels for IFITM1 (+75%; F_(1,38)=13.2, p=.001) and IFITM 2/3 (+116%; F_(1,38)=15.8, p<.001). Because 1) mRNA levels for IFITM1 and IFITM2/3 are highly correlated across all subjects (r=0.74, p<.0001) and are similarly over-expressed in schizophrenia and 2) the different IFITM variants have largely similar roles as restriction factors for the same viruses (17,19,20), we focused on IFITM2/3 mRNA levels for subsequent data analyses.

Figure 1. Quantitative PCR analysis of IFITM mRNA levels in the prefrontal cortex in schizophrenia.

Transcript levels for each schizophrenia subject relative to the matched healthy comparison subject are indicated by open circles. Data points to the left of the unity line indicate higher mRNA levels in the schizophrenia subject relative to the healthy comparison subject. Mean mRNA levels (± SD) in schizophrenia subjects were statistically significantly higher for IFITM1 (0.027 ± 0.017) and IFITM2/3 (0.078 ± 0.058) relative to healthy comparison subjects (IFITM1: 0.016 ± 0.005; IFITM2/3: 0.036 ± 0.011).

Among schizophrenia subjects, IFITM mRNA levels did not differ as a function of factors that predict a more severe course of illness (male sex, a diagnosis of schizophrenia rather than schizoaffective disorder, first-degree relative with schizophrenia, early age at illness onset [≤18 years of age]) or measures of illness severity (suicide, no history of marriage, low socioeconomic status as measured by the Hollingshead Index of Social Position, living dependently at the time of death) (all F≤.37, p≥.55; Supplement: Figure S1). We also found no relationship between use of antipsychotic, antidepressant, benzodiazepine, or smoking at time of death and IFITM mRNA levels in schizophrenia subjects (all F≤0.56, p≥0.46; Supplement: Figure S1). IFITM mRNA levels also did not differ in the prefrontal cortex of monkeys chronically exposed to haloperidol, olanzapine, or placebo (F_(2,10)=1.3, p=.31; Figure 2). These data suggest that elevated IFITM mRNA levels in schizophrenia are not associated with disease severity and are not attributable to the effects of psychotropic medication.

Figure 2. IFITM mRNA levels in prefrontal area 9 of antipsychotic-exposed monkeys.

Quantitative PCR analysis revealed no statistically significant differences in IFITM mRNA levels (mean ± standard deviation) in monkeys chronically exposed to either olanzapine (0.0120 ± 0.0016; dark gray triangles) or haloperidol (0.0107 ± 0.0011; black triangles) compared to placebo (0.0115 ± 0.0017; light gray triangles). Mean values are shown as horizontal black bars.

We next sought to determine whether higher IFITM mRNA levels in schizophrenia may be related to infection/inflammation-related cause of death. The most common cause of death in our community-based, middle-aged cohort collected from autopsies is cardiopulmonary-related (~54%; Table 1), and only ~5% of subjects died of infection/inflammation-related causes (i.e. peritonitis, myocarditis, pneumonia, and anaphylaxis; Supplement: Table S1). Only including subjects with cardiopulmonary-related deaths revealed that IFITM mRNA levels were still much higher in schizophrenia (+76%; 0.061± 0.030; F_(1,54)=32.9, p<.0001) relative to comparison subjects (0.035 ± 0.010), suggesting that higher IFITM mRNA levels in schizophrenia cannot be explained by infection/inflammation-related cause of death. Furthermore, excluding all subjects exposed to non-steroidal anti-inflammatory drugs at time of death (either prescribed or over-the-counter use indicated by presence of NSAIDs in toxicology screens of blood and/or urine; Table 1, Supplement: Table S1) revealed that IFITM mRNA levels remained much higher in subjects with schizophrenia (+110%; 0.078 ± 0.054; F_(1,83)=22.9, p<.0001) relative to healthy subjects (0.037 ± 0.011). Finally, excluding subjects with overdose as cause of death (Supplement: Table S1) also revealed that IFITM mRNA levels continued to be much higher in subjects with schizophrenia (+115%; 0.078 ± 0.057; F_(1,95)=26.8, p<.0001) relative to healthy subjects (0.036 ± 0.011).

In Situ Hybridization for IFITM mRNA in Schizophrenia

The cellular localization of IFITM in human prefrontal cortex has not been previously reported, and the cell types that overexpress IFITM mRNA in the prefrontal cortex in schizophrenia remain unknown. Therefore, we validated the finding of higher IFITM mRNA levels in schizophrenia using in situ hybridization in a limited number of subjects (i.e. 20 subject pairs that had a sufficient number of available tissue sections; Supplement: Table S1). Optical density analysis of film autoradiographs revealed that mean IFITM mRNA levels were higher in the gray matter (+117%; F_(1,15)=16.2, p=.001) and white matter (+91%; F_(1,15)=11.1, p=.005) in schizophrenia (Figure 3A-D). Furthermore, gray matter IFITM mRNA levels quantified by in situ hybridization and by qPCR were correlated in the same subjects (r=.72, p<.0001), which indicates the reliability of the mRNA quantification techniques and the reproducibility of the results. Interestingly, pseudocolor film autoradiographs from schizophrenia subjects (Figure 3B) revealed small structures with intense signal that were linear, round, and/or had branch points, which is consistent with labeling of blood vessels. Consequently, we exposed the same tissue sections to nuclear emulsion and developed and stained the sections for Nissl substance to conduct a grain counting analysis. Dense silver grain clusters indicating IFITM mRNA expression were present over distinct clusters of packed cells with elongated nuclei that often contained a lumen or branching point. These cell clusters had an either linear or round orientation, consistent with longitudinal or transverse profiles of blood vessels, respectively (Figure 3E-F). Grain clusters were also consistently observed over pia mater (not shown). In contrast, qualitative inspection of all schizophrenia and healthy comparison subjects revealed an absence of distinct grain clusters over neurons or glia (Supplement: Figure S2). In schizophrenia subjects, the mean grain density over blood vessels was higher in schizophrenia subjects (+71%; F_(1,15)=18.8, p<.001; Figure 3G) relative to comparison subjects and was highly correlated with gray matter optical density measures (r=.73, p<.0001). However, we were not able to quantify grain density over pia mater due to the limited presence of pia mater in the tissue sections.

Figure 3. In situ hybridization for IFITM mRNA in the prefrontal cortex of schizophrenia and healthy comparison subjects with film and grain counting analyses.

A-B. Pseudocolored film autoradiographs of tissue sections from a matched pair of healthy comparison (A) and schizophrenia (B) subjects processed by in situ hybridization with an ³⁵S-riboprobe for IFITM mRNA are shown. Solid white line indicates the gray matter/white matter border; calibration bar = 1 mm. While the comparison subject has relatively low levels (colder colors; scale) of IFITM mRNA throughout the gray and white matter, the schizophrenia subject has distinct small structures (white arrows) with intense signal (warmer colors) that are linear, round, and/or have branch points, which is consistent with the anatomical appearance of blood vessels. IFITM mRNA levels in the gray matter (C) and white matter (D) of prefrontal cortical area 9 for schizophrenia subjects and matched control subjects in a pair are indicated by open circles. Data points to the left of the unity line indicate higher mRNA levels in the schizophrenia subject relative to the healthy comparison subject. Mean (± SD) IFITM mRNA levels measured by optical density in schizophrenia subjects were statistically significantly higher in the gray matter (0.197 ± 0.85) and white matter (0.172 ± 0.091) relative to healthy comparison subjects (0.091 ± 0.010 and 0.090 ± 0.016, respectively). E-F. Photomicrograph of a prefrontal cortex section processed by in situ hybridization with an ³⁵S-riboprobe for IFITM mRNA that was subsequently exposed to nuclear emulsion, developed, and Nissl-stained. E. A representative brightfield image from the same tissue section shown in frame B of Figure 5 illustrates a cluster of packed cells with elongated nuclei comprising a round structure with a lumen that is consistent with the transverse profile of a blood vessel. The darkfield image (F) reveals the accumulation of silver grains representing IFITM mRNA labeling over the blood vessel. Mean grain density per profile area (μm²) of blood vessel (i.e. traced outline of blood vessel excluding the lumen when present) was significantly higher in schizophrenia subjects (0.175 ± 0.064) relative to healthy comparison subjects (0.103 ± 0.020) (G).

Relationship between IFITM mRNA Levels and GABA Neuron-Related Markers in Schizophrenia

Disturbances in prefrontal inhibitory neurons have been commonly reported in schizophrenia (25-38). In a recent study of 42 of the schizophrenia subjects included in the present study, we reported lower mRNA levels for the GABA synthesizing enzyme GAD67, neuronal subpopulation markers parvalbumin and somatostatin, and the GABA neuron-specific transcription factor Lhx6 (37). Interestingly, in schizophrenia subjects (Figure 4), IFITM mRNA levels were inversely correlated with mRNA levels for GAD67 (r=−.38, p=.01), somatostatin (r=−.56, p<.0001), and Lhx6 (r=−.42, p=.005) but did not reach statistical significance for parvalbumin (r=−.24, p=.13). In contrast, in healthy subjects, IFITM mRNA levels were not correlated with GAD67, parvalbumin, or Lhx6 (for all, r<.23, p>.15) but were inversely correlated with somatostatin (r=−.51, p=.001). Furthermore, we had previously identified a “low-GABA-marker” subset of schizophrenia subjects that had the most severe deficits in GAD67, parvalbumin, somatostatin, and Lhx6 mRNA levels (37). IFITM mRNA levels were even higher (F_(2,76)=15.6, p<.0001) in the low-GABA-marker schizophrenia subjects (n=20) relative to other schizophrenia subjects (+61%, p=.001; n=22) and to healthy subjects (+148%, p<.0001; n=42).

Figure 4. Relationship between IFITM mRNA levels and GABA neuron-related markers in schizophrenia.

A. Quantitative PCR analysis revealed that in schizophrenia subjects (grey circles; black linear regression lines, n=42 subjects), IFITM mRNA levels were inversely correlated with mRNA levels for GAD67 (r=−.38, p=.01), somatostatin (r=−.56, p<.0001), Lhx6 (r=−.42, p=.005), but did not reach statistical significance for parvalbumin (r=−.24, p=.13). In contrast, in healthy subjects (open circles, dashed linear regression lines, n=42 subjects), IFITM mRNA levels were not correlated with GAD67, parvalbumin, or Lhx6 (for all, r<−.23, p>.15), but were inversely correlated with somatostatin (r=−.51, p=.001). B. IFITM mRNA levels were even higher in the low-GABA-marker schizophrenia subjects (dark grey circles; n=20) relative to other schizophrenia subjects (light grey circles; n=22) and to healthy comparison subjects (open circles; n=42). IFITM mRNA levels were also higher in the schizophrenia subjects not included in the low-GABA-marker subset relative to healthy comparison subjects.

Discussion

In this study, we found marked elevations in mRNA levels for the viral restriction factor IFITM in the prefrontal cortex of a large cohort of schizophrenia subjects. The consistent reports of higher IFITM mRNA levels across different schizophrenia subject cohorts (14,15) and in the vast majority of schizophrenia subjects relative to healthy comparison subjects in the presently studied cohort suggests that higher IFITM mRNA levels are a common feature in schizophrenia. Furthermore, higher IFITM mRNA levels in schizophrenia did not appear to be attributable to antipsychotic medications, smoking, or an immune- or inflammation-related cause of death. One particularly unique finding of this study is that in schizophrenia, IFITM mRNA levels were markedly elevated in cortical blood vessels, which may indicate a more systemic immune/inflammatory disturbance in the disorder. Finally, we found an association between higher IFITM mRNA levels in cortical blood vessels and deficits in cortical GABA neuron markers in schizophrenia, suggesting that these cell-type distinct disturbances may be affected by a shared upstream insult that is related to immune activation.

Elevated IFITM mRNA levels may be related to neuroinflammatory processes in schizophrenia. For example, elevated serum levels of proinflammatory cytokines, including cytokines known to induce IFITM expression such as IL-6 (16,19,23), have been commonly reported in schizophrenia (10,11); however, peripheral serum was not available for any subjects in the present study. Interestingly, elevated serum cytokine levels have also been reported in drug-naive and first-episode psychosis subjects (11,12), suggesting that higher cytokine levels are not a consequence of treatment or chronic illness. Interestingly, higher mRNA levels of proinflammatory cytokines such as IL-6 were recently found to be associated with lower mRNA levels for GAD67, parvalbumin, and somatostatin in the prefrontal cortex of another cohort of schizophrenia subjects (13). Furthermore, IL-6 reduces levels of GAD67 and parvalbumin in cell culture (48), and overexpression of IL-6 leads to fewer parvalbumin-immunoreactive neurons in the hippocampus (49-51). In addition, administration of an immune stimulating agent to neonatal mouse pups has been reported to greatly elevate IFITM immunoreactivity in hippocampus (52). Taken together, these findings suggest that neuroinflammation, in part through elevated levels of IL-6, may also contribute to GABA neuron dysfunction and IFITM mRNA overexpression in schizophrenia. Consistent with this hypothesis, a mouse model of mild chronic neuroinflammation (i.e. mice with a complete loss of the nuclear factor-κB site-binding protein Schnurri-2 which normally suppresses production of cytokines such as IL-6) has been reported to have lower parvalbumin and GAD67 levels and higher IFITM levels in the cortex and hippocampus (24).

Several lines of evidence suggest that altered immune/inflammatory processes may even be present at very early stages of life in schizophrenia. For example, prenatal exposure to immune activation has been associated with an increased risk of developing schizophrenia later in life. Furthermore, the population attributable risk for schizophrenia due to maternal infection has been estimated to be ~30% (9), indicating that maternal infection is a relatively common in utero exposure in offspring who eventually develop schizophrenia. Animal models using the viral mimic and immune stimulant poly I:C have shown that maternal immune activation raises levels of multiple cytokines in fetal brain (53-55) and induces higher peripheral cytokine levels that persist into adulthood (56), including cytokines known to induce IFITM expression such as IL-6 (16,19,23). Maternal immune activation has also been shown to lower parvalbumin immunoreactivity and GAD67 mRNA levels in the prefrontal cortex of adult mice (57,58), similar to findings in schizophrenia (25-38). Thus, while further proof-of-principle testing is needed, it is possible that maternal immune activation could contribute to long-lasting elevations in IFITM mRNA levels and deficits in GABA-related mRNAs in adult brain through multiple mechanisms that disrupt early fetal development by elevating cytokine levels in fetal brain (53) and through enduring epigenetic modifications at gene promoter regions (59).

These findings may also provide insight into preventative and treatment strategies for schizophrenia. Preventative strategies might involve maternal prenatal care to reduce infection rates and potentially reduce risks for developing schizophrenia in at-risk offspring. For example, prevention of the maternal infection via vaccinations or personal health practice or rapid treatment of maternal infection may help mitigate the deleterious effects of maternal immune activation (9). In addition, chronic neuroinflammation may be targeted by immune-modulating agents such as non-steroidal anti-inflammatory drugs. Indeed, adjunctive treatment with the cyclooxygenase-2 inhibitor celecoxib has been reported to improve schizophrenia-related symptoms (60,61). Finally, another clinical approach may involve targeting the JAK-STAT (Janus-activated kinase and signal transducer and activator of transcription) pathway which is involved in signal transduction for a broad range of pro-inflammatory cytokines and IFITM (62). JAK-STAT inhibitors, such as ruxolitinib which was recently FDA-approved for the treatment of myelofibrosis (63), may potentially limit neuroinflammation and lower IFITM mRNA expression. Thus, further investigation into the role of immune- and inflammatory-related processes in producing cortical inhibitory circuitry disturbances in schizophrenia may help inform novel preventative and treatment strategies for the disorder.