HERV-W ENV Induces Innate Immune Activation and Neuronal Apoptosis via linc01930/cGAS Axis in Recent-Onset Schizophrenia

Schizophrenia is a severe neuropsychiatric disorder affecting about 1% of individuals worldwide. Increased innate immune activation and neuronal apoptosis are common findings in schizophrenia. Interferon beta (IFN-β), an essential cytokine in promoting and regulating innate immune responses, causes neuronal apoptosis in vitro. However, the precise pathogenesis of schizophrenia is unknown. Recent studies indicate that a domesticated endogenous retroviral envelope glycoprotein of the W family (HERV-W ENV, also called ERVWE1 or syncytin 1), derived from the endogenous retrovirus group W member 1 (ERVWE1) locus on chromosome 7q21.2, has a high level in schizophrenia. Here, we found an increased serum IFN-β level in schizophrenia and showed a positive correlation with HERV-W ENV. In addition, serum long intergenic non-protein coding RNA 1930 (linc01930), decreased in schizophrenia, was negatively correlated with HERV-W ENV and IFN-β. In vitro experiments showed that linc01930, mainly in the nucleus and with noncoding functions, was repressed by HERV-W ENV through promoter activity suppression. Further studies indicated that HERV-W ENV increased IFN-β expression and neuronal apoptosis by restraining the expression of linc01930. Furthermore, HERV-W ENV enhanced cyclic GMP-AMP synthase (cGAS) and stimulator of interferon genes protein (STING) expression and interferon regulatory factor 3 (IRF3) phosphorylation in neuronal cells. Notably, cGAS interacted with HERV-W ENV and triggered IFN-β expression and neuronal apoptosis caused by HERV-W ENV. Moreover, Linc01930 participated in the increased neuronal apoptosis and expression level of cGAS and IFN-β induced by HERV-W ENV. To summarize, our results suggested that linc01930 and IFN-β might be novel potential blood-based biomarkers in schizophrenia. The totality of these results also showed that HERV-W ENV facilitated antiviral innate immune response, resulting in neuronal apoptosis through the linc01930/cGAS/STING pathway in schizophrenia. Due to its monoclonal antibody GNbAC1 application in clinical trials, we considered HERV-W ENV might be a reliable therapeutic choice for schizophrenia.


Introduction
Human endogenous retroviruses (HERVs), discovered in 1981 [1], are remnants of retroviral infection to human germline cells million years ago [2], constituting about 8% of the whole genome [3]. HERVs are regularly composed of gag, pro, pol, and env, with two long terminal repeats aside [4]. Most HERVs remain inactive due to mutation accumulation [5]. However, some HERVs still have open reading frames to encode functional transcripts and participate in various normal physiological processes, such as embryogenesis [6]. Recent studies show that HERVs derive nucleic acids or proteins involved in antiviral responses [7]. ERV-derived long noncoding RNA (lncRNA) enhances innate immune responses [8]. Furthermore, HERVs constitute a dynamic reservoir of interferon study identifies linc01930 as a susceptibility locus to schizophrenia [36]. However, there is no report on linc01930. Here, we first detected the expressions of linc01930 in the serum of 21 schizophrenia patients and 26 healthy controls. There were no significant differences in age, education level, gender, smoking status and BMI between control subjects and schizophrenia patients (Supplementary Table S1). We discovered that serum linc01930 level was decreased in schizophrenia patients compared with healthy controls (Figure 1a), with a median of 0.0466 and 0.2699, respectively (Table 1). Additionally, we found that IFN-β was increased in the blood sample of schizophrenia patients compared with healthy controls by enzyme-linked immunosorbent assay (ELISA) (Figure 1b), with a median of 52.1293 ng/L and 31.0150 ng/L, respectively (Table 2). Moreover, we also found increased HERV-W ENV at mRNA level in schizophrenia patients compared with healthy controls (Figure 1c), with a median of 1.6501 and 0.2272, respectively (Table 3). Spearman correlation analyses indicated that HERV-W ENV had a negative correlation to linc01930 ( Figure 1d) and a positive correlation to IFN-β (Figure 1e), while linc01930 had a negative correlation to IFN-β (Figure 1f). In schizophrenia patients, our further analyses revealed that the consistency ratio of HERV-W ENV and linc01930 (Table 4), HERV-W ENV and IFN-β (Table 5), linc01930 and IFN-β ( Table 6) was 57.1%, 66.7% and 42.8%, respectively. Thus, HERV-W ENV, linc01930, and IFN-β might be potential risk factors in schizophrenia.

HERV-W ENV Activated Antiviral Innate Immune Responses and Caused Neuronal Apoptosis
Our clinical data showed a positive correlation between HERV-W ENV and IFN-β in schizophrenia. The human neuroblastoma SH-SY5Y cells, which are from neuroblasts and have the potential to differentiate into neuronal cells [45], and rat primary neuronal cells, have been widely used as neuronal models of schizophrenia [20,22]. Therefore, we used SH-SY5Y and rat primary neurons to study the causal relationship between HERV-W ENV and IFN-β in neurons. Successful expression of HERV-W ENV in SH-SY5Y cells and primary neurons were shown (Supplementary Figure S1a-d).
We found that HERV-W ENV statistically significantly increased IFN-β expression levels at the mRNA (Figure 2a,b) and protein (Figure 2c,d) in neuronal cells. Luciferase assays showed that HERV-W ENV enhanced IFN-β promoter activity in SH-SY5Y cells (Figure 2e). The production of type I interferon, including IFN-β, is the hallmark of antiviral innate immune responses [46]. So the above results indicated that HERV-W ENV activated antiviral innate immune responses in neuronal cells.
In a word, HERV-W ENV evoked antiviral innate immune responses in neurons and inflated neuronal apoptosis.

HERV-W ENV Dowregulated the Expression of linc01930 in Neuronal Cells
Our clinical data suggested that HERV-W ENV was negatively correlated to linc01930 in schizophrenia patients. LncRNAs act as key regulators in brain disorders, including schizophrenia [51]. Our results from in vitro and in vivo studies showed that HERV-W ENV prominently impaired linc01930 expression in neuronal cells (Figure 3a,b). Promoters serve as a kind of "On" switch to initiate the biological process of transcription for the genes [52]. Luciferase assays indicated that HERV-W ENV markedly reduced linc01930 promoter activity in SH-SY5Y cells (Figure 3c), suggesting that HERV-ENV repressed linc01930 expression through its promoter. Typically, apoptotic vulnerability is increased in schizophrenia patients [47]. The type I interferon IFN-β has been reported to influence cell apoptosis [48][49][50]. The CCK8 assay results demonstrated that HERV-W ENV reduced neuronal cell proliferation ( Figure 2f). Furthermore, the flow cytometry assays revealed that HERV-W ENV accelerated neuronal cell apoptosis (Figure 2g).
In a word, HERV-W ENV evoked antiviral innate immune responses in neurons and inflated neuronal apoptosis.

HERV-W ENV Dowregulated the Expression of linc01930 in Neuronal Cells
Our clinical data suggested that HERV-W ENV was negatively correlated to linc01930 in schizophrenia patients. LncRNAs act as key regulators in brain disorders, including schizophrenia [51]. Our results from in vitro and in vivo studies showed that HERV-W ENV prominently impaired linc01930 expression in neuronal cells (Figure 3a,b). Promoters serve as a kind of "On" switch to initiate the biological process of transcription for the genes [52]. Luciferase assays indicated that HERV-W ENV markedly reduced linc01930  western blotting. (f-h) Cellular distribution of linc01930 (F, DFn, Dfd, 2.703, 2, 2) was mainly located at the nucleus in the SH-SY5Y cell. Nuclear and cytoplasmic separation effects were quantified to RPS14 (F, DFn, Dfd, 1.316, 2, 2) in the cytoplasmic part and U6 (F, DFn, Dfd, 2.941, 2, 2) in the nuclear part. Statistical analysis was performed by one-way analysis of variance (ANOVA). * p < 0.05; ** p < 0.01.
As for the lack of a functional Open Reading Frame (ORF), LncRNAs can not encode proteins. However, several recent reports indicate that some lncRNAs take part in the pathogenesis of disease with their encoded peptides [53]. We found three open-reading frame fragments in NCBI ORF Finder ( Figure 3d) and constructed the fragment separately in the pEGFP-N3 plasmid. The western blot analyses indicated that linc01930 did not encode peptides (Figure 3e).
LncRNAs have diverse functions depending on their cellular localization [54]. Our results indicated that linc01930 was mainly located in the nucleus, implying that linc01930 could regulate underlying target expression at the transcriptional level (Figure 3f-h).
Together, linc01930, suppressed by HERV-W ENV through the promoter activity, was mainly located at the nucleus and did not code peptide.

Linc01930 Suppressed the Antiviral Innate Immune and Neural Apoptosis Caused by HERV-W ENV
Several studies suggest that lncRNAs regulate innate immune response [55]. Our clinical data indicated a negative correlation between linc01930 and IFN-β in schizophrenia. However, there is no report about the effect of linc01930 on IFN-β. Efficient expression of linc01930 in neuronal cells was confirmed at the mRNA level (Supplementary Figure S2a,b). We found that linc01930 led to noticeable reductions in the mRNA (Figure 4a As for the lack of a functional Open Reading Frame (ORF), LncRNAs can not encode proteins. However, several recent reports indicate that some lncRNAs take part in the pathogenesis of disease with their encoded peptides [53]. We found three open-reading frame fragments in NCBI ORF Finder ( Figure 3d) and constructed the fragment separately in the pEGFP-N3 plasmid. The western blot analyses indicated that linc01930 did not encode peptides (Figure 3e).
LncRNAs have diverse functions depending on their cellular localization [54]. Our results indicated that linc01930 was mainly located in the nucleus, implying that linc01930 could regulate underlying target expression at the transcriptional level (Figure 3f-h).
Together, linc01930, suppressed by HERV-W ENV through the promoter activity, was mainly located at the nucleus and did not code peptide.

Linc01930 Suppressed the Antiviral Innate Immune and Neural Apoptosis Caused by HERV-W ENV
Several studies suggest that lncRNAs regulate innate immune response [55]. Our clinical data indicated a negative correlation between linc01930 and IFN-β in schizophrenia. However, there is no report about the effect of linc01930 on IFN-β. Efficient expression of linc01930 in neuronal cells was confirmed at the mRNA level (Supplementary Figure S2a,b). We found that linc01930 led to noticeable reductions in the mRNA (Figure 4a   Some lncRNAs regulate cell apoptosis and influence disease pathogenesis [56]. The biological function of linc01930 has been ambiguous till up to now. In this article, we first reported that linc01930 increased cell proliferation (Figure 4f  Some lncRNAs regulate cell apoptosis and influence disease pathogenesis [56]. The biological function of linc01930 has been ambiguous till up to now. In this article, we first reported that linc01930 increased cell proliferation (Figure 4f) and decreased apoptosis (Figure 4g) in SH-SY5Y cells. These findings denoted that linc01930 attenuated neuronal apoptosis by suppressing IFN-β.
Western blotting (Figure 5a,b) and ELISA (Figure 5c,d) indicated that linc01930 could deteriorate the increased IFN-β production stimulated by HERV-W ENV in neuronal cells. The efficient transfection of HERV-W ENV and linc01930 was shown (Supplementary Figure S5a-d). Furthermore, we found that linc01930 reversed the decreased cell prolif-eration caused by HERV-W ENV ( Figure 5e) and markedly lessened cell apoptosis rate increased by HERV-W ENV (Figure 5f,g) in SH-SY5Y cells. Together, these results suggested that linc01930 impaired antiviral innate immune responses and neuronal apoptosis mediated by HERV-W ENV.
Western blotting (Figure 5a,b) and ELISA (Figure 5c,d) indicated that linc01930 could deteriorate the increased IFN-β production stimulated by HERV-W ENV in neuronal cells. The efficient transfection of HERV-W ENV and linc01930 was shown (Supplementary Figure S5a-d). Furthermore, we found that linc01930 reversed the decreased cell proliferation caused by HERV-W ENV ( Figure 5e) and markedly lessened cell apoptosis rate increased by HERV-W ENV (Figure 5f,g) in SH-SY5Y cells. Together, these results suggested that linc01930 impaired antiviral innate immune responses and neuronal apoptosis mediated by HERV-W ENV.

Linc01930 is Involved in the cGAS-Mediated Antiviral Signaling Pathway Activated by HERV-W ENV
Exogenous retroviruses trigger cGAS-dependent IFN-β production and innate immune response [57]. There is no report about the impact of endogenous retroviruses (ERVs) on cGAS. Here we found that HERV-W ENV substantially elevated mRNA expression of cGAS (Figure 6a Interferon regulatory factor 3 (IRF3) phosphorylation at Ser 386 sit is essential to cGASinduced IFN-β expression [58]. The western blotting indicated that HERV-W ENV enhanced the phosphorylation of IRF3 (Figure 6e), suggesting that HERV-W ENV triggered the cGAS signaling pathway. Co-IP analyses indicated that HERV-W ENV interacted with cGAS ( Figure 6f). Together, we found that HERV-W ENV interacted with cGAS and stimulated cGAS-STING axis through IRF3 phosphorylation.

Linc01930 is Involved in the cGAS-Mediated Antiviral Signaling Pathway Activated by HERV-W ENV
Exogenous retroviruses trigger cGAS-dependent IFN-β production and innate immune response [57]. There is no report about the impact of endogenous retroviruses (ERVs) on cGAS. Here we found that HERV-W ENV substantially elevated mRNA expression of cGAS (Figure 6a,b) and STING (Supplementary Figure S3a,b) in neuronal cells. Consistently, western blot analyses showed HERV-W ENV mediated higher levels of cGAS (Figure 6c,d) and STING (Supplementary Figure S3c,d) in neuronal cells. Interferon regulatory factor 3 (IRF3) phosphorylation at Ser 386 sit is essential to cGAS-induced IFN-β expression [58]. The western blotting indicated that HERV-W ENV enhanced the phosphorylation of IRF3 (Figure 6e), suggesting that HERV-W ENV triggered the cGAS signaling pathway. Co-IP analyses indicated that HERV-W ENV interacted with cGAS ( Figure 6f). Together, we found that HERV-W ENV interacted with cGAS and stimulated cGAS-STING axis through IRF3 phosphorylation.

cGAS-Mediated Antiviral Signaling Pathway is Necessary for the Antiviral Innate Immune Responses and Neuronal Apoptosis Caused by HERV-W ENV
cGAS promotes IFN-β production and mediates innate immune response [46]. Our results also found that the knockdown of cGAS decreased IFN-β expression at the protein level (Figure 8a

cGAS-Mediated Antiviral Signaling Pathway is Necessary for the Antiviral Innate Immune Responses and Neuronal Apoptosis Caused by HERV-W ENV
cGAS promotes IFN-β production and mediates innate immune response [46]. Our results also found that the knockdown of cGAS decreased IFN-β expression at the protein level (Figure 8a
The current diagnosis of schizophrenia relies on the experience of the doctor and can lead to misdiagnosed results [83]. Therefore, the efficient and early detection of biomarkers is necessary to offer a reliable way for a schizophrenia diagnosis. To our knowledge, there is no blood marker available for schizophrenia because of the bloodbrain barrier [84]. Considering the fact that lncRNAs participate in neuropsychiatric disorders and easily pass through the blood -brain barrier [85], they may be suitable blood markers for neuropsychiatric disorders, including schizophrenia [83]. Some lncRNAs, such as Gomafu and AK096174, have been supposed to be potential blood biomarkers in cancers [86,87]. Nevertheless, no clinical trials of lncRNAs have been documented in schizophrenia. Bioinformatic data indicates that linc01930 is a novel susceptible locus for schizophrenia [36]. There are only a few reports that disclose the abnormal expression of linc01930 in pheochromocytoma and paraganglioma [88], and neuroblastoma [89]. The role of linc01930 in the etiology of schizophrenia remains unclear.
In this paper, we first reported that linc01930 was decreased in schizophrenia, suggesting that serum linc01930 might be a novel potential blood marker and risk factor for schizophrenia. The type I interferon IFN-β is the essential mediator of innate immunity [90].
Our clinical data showed that IFN-β was increased in the blood samples of schizophrenia. This is consistent with the reports of Volk et al. [31] and Hidese et al. [32] on brain tissue. These findings displayed IFN-β as a potential blood biomarker. Together, linc01930 and IFN-β might be new potential biomarkers for a schizophrenia diagnosis. The cut point between schizophrenia patients and healthy controls might not be significantly obvious, largely attributed to the small sample size. In addition, healthy controls possibly had a low level of linc01930 and a high level of IFN-β to show false positive results, for example, the clinical use of alpha-fetoprotein in live cancer [91]. Although the correlations among HERV-W ENV, linc01930 and IFN-β were moderately relevant, the consistency ratio of HERV-W ENV to linc01930 and IFN-β was 57.1% and 66.7%, respectively, indicating more samples possibly improved cut point of the linc01930 and IFN-β between schizophrenia patients and healthy controls, which was our aim in the further study.
Further analyses suggested that linc01930 was negatively correlated with HERV-W ENV in the serum of schizophrenia. In vitro experiments indicated that HERV-W ENV suppressed linc01930 expression in neuronal cells via promoter activity. Subcellular localization of lncRNAs has valuable clues for their molecular functions [54]. Our data demonstrated that linc01930 was mainly located in the nucleus and unable to encode function peptide, indicating it might play a role as a transcriptional regulator. As far as we know, the biological function of linc01930 remains unclear. We found that linc01930 exerted opposite effects on IFN-β expression through repressing promoter activity. Further studies manifested that linc01930 restrains neuronal apoptosis and exerts a cell proliferation role via inactivating IFN-β. From these, we could conclude that linc01930 might restrain innate immune activation and facilitate neural cell proliferation.
Our clinical data also suggested that IFN-β, increased in the blood sample of schizophrenia patients, had a positive correlation with HERV-W ENV. IFN-β, the type I interferon, is a vital mediator in innate immune activation, which functions to modulate cell growth and influence the activation of various immune cells [9]. Quite a few reports describe innate immune imbalances in schizophrenia [28,31]. In addition, several studies, including GWAS [92], support the role of innate immune activation in schizophrenia [93]. Notably, HERVs and their transcripts actively participate in innate immunity [94] and regulate the antiviral interferon network integrating into or near immune-related genes [95]. In addition, HERV insertions may lead to the amplification of IFN transcription [9]. Our cellular experiments revealed that HERV-W ENV stimulated IFN-β expression via promoter activity, suggesting that HERV-W ENV may induce antiviral innate immune responses in schizophrenia.
A recent article reports that IFN-β exerts apoptotic activity by increasing p38 MAPK activity, MK2 impulse, and HSP27 phosphorylation in SH-SY5Y cells [48]. In addition, IFN-β aggravates neuronal damage by inhibiting neuronal survival and neurite outgrowth through BDNF/TrkB axis [50]. Furthermore, IFN-β provokes the neurotoxicity directly via JAK/STAT and PI3K/AKT pathway in SH-SY5Y cell and rat primary neurons, causing cytochrome C release and intrinsic apoptotic pathway activation [49]. There is an increased susceptibility to apoptosis in Schizophrenia. The anti-apoptotic membrane-bound protein Bcl2 is decreased in the cortical of schizophrenia [96], and Bax/Bcl2 ratio is significantly higher in schizophrenia patients [97]. All these reports indicate that cell apoptosis is dysregulated in schizophrenia, which possibly leads to neuronal damage [47]. In this paper, we found that HERV-W ENV stimulated neuronal apoptosis through IFN-β. In a word, HERV-W ENV mediated neuronal apoptosis, which possibly functions in the pathogenesis of schizophrenia. An additional study demonstrated that Linc01930 repressed innate antiviral immunity and neuronal apoptosis mediated by HERV-W ENV. Together, HERV-W ENV led to neuronal damage through IFN-β via inhibiting linc01930.
Several signaling pathways, including cGAS/STING pathway, regulate the expression of IFN-β and induce innate antiviral immunity [46]. cGAS/STING induces IFN-β expression through IRF3 phosphorylation [98]. As a cytosolic DNA sensor, cGAS also mediates immune activation by HIV and other retroviruses [57]. A present study unveils that HERV-K (HML-2) stimulates interferon via cGAS/STING in COVID-19 patients [99]. Our previous work notices that HERV-W ENV triggers immune response activation through TLRs [20,77]. However, there is no report about the effect of HERV-W ENV on cGAS. In this paper, we found that HERV-W interacted with cGAS and triggered the activation of cGAS and STING in neuronal cells. Linc01930 suppressed the increased cGAS mediated by HERV-W ENV. Our in-depth study reveals that cGAS is involved in innate antiviral immunity and neuronal apoptosis induced by HERV-W ENV.
GNbAC1, a humanized IgG4 monoclonal antibody specifically interacting with HERV-W ENV [100], has been used in a one-year phase 2b clinical trial for multiple sclerosis [101]. Additionally, GNbAC1 also has favorable prospects in clinical trials for immune-related patients, such as type 1 diabetes (T1D) [102]. Our results promulgated that HERV-W ENV might be a potential target for clinical treatment in schizophrenia. Thus, a monoclonal antibody to HERV-W ENV may be significant as a novel therapy for schizophrenia treatment.

Clinical Blood Samples
All 21 schizophrenia patients and 26 healthy controls were recruited from Renmin Hospital, Wuhan University (Wuhan, China). The recent onset patients were diagnosed due to the Diagnostic and Statistical Manual of Mental Disorders (DSM-IV) without psychotropic drug treatment before. The healthy volunteers all passed the physical examination. The blood samples were divided into two-part, one for the RT-PCR test and the other for the ELISA test with the supernatants by centrifugation at 4 • C. Samples were stored at −80 • C before use. All subjects were informed of the notification from the Institutional Review Board of Wuhan University, School of Medicine. There were no significant differences in median age, education, BMI (body mass index), smoking habit, and sex between healthy individuals and patients. Details are listed in Supplementary Table S1.

Cell Culture and Transfection
The neuroblastoma cell line SH-SY5Y was purchased from American Type Culture Collection. The cells were maintained in the culture media of Minimal Essential Medium Eagle(MEM) (2225320, Gibco, Baltimore, MD, USA) and F-12 (2209586, Gibco, Baltimore, MD, USA) at equal percent, with the supplement of 10% fetal bovine serum (2001003, Biological Industries, Beit HaEmek, Israel), 1% sodium pyruvate (2185865, Gibco, MD, USA) and 1% penicillin/streptomycin (2185865, Gibco, Baltimore, MD, USA), under the condition of 5% CO 2 at 37 • C. While HEK-293T cell was stored in liquid nitrogen and maintained in the Dulbecco's modified Eagle's medium (11965092, Gibco, MD, USA), with the supplement of 10% fetal bovine serum and 1% penicillin/streptomycin and storage condition as described before.
Primary neurons were acquired in the cerebral cortex from neonatal Sprague Dawley (SD) rats according to the method previously reported [103]. Neonatal SD rats were purchased from Hubei Center for Disease Control and Prevention. Primary neuron cells were preserved in the Neurobasal medium (21103049, Gibco, MD, USA), supplied with 1% B27 (17504044, Gibco, MD, USA), 1% sodium pyruvate (2185865, Gibco, MD, USA) and 1% penicillin/streptomycin (2185865, Gibco, MD, USA), under the condition of 5% CO 2 at 37 • C. Moreover, these experiments on animals got support from the Animal Ethics Committee of Wuhan University Center for Animal Experiment/A3 Laboratory, Wuhan University.
Cell transfection was performed by Neofect TM DNA Transfection reagent (D210101, Neofect Biotech Co., Ltd., Beijing, China) due to the manufacturer's instructions.

Reverse Transcription and Quantitative Real-Time PCR
According to the manufacturer's instructions, total cellular RNA (after transfected and cultured for 24 h) and blood RNA were isolated from TRIzol reagent (15596018, Invitrogen, California, USA) and TRIzol LS reagent (10296028, Invitrogen, California, USA) separately. Then 0.5 µg RNA was used to obtain cDNA through the ReverTra kit (FSQ-301; Toyobo, Osaka, Japan). The mRNA expression level was detected in the detector (T100, Bio-Rad, California, USA) by utilizing a 2× SYBR Green qPCR Mix (2992239AX, Aidlab Biotechnologies Co. Ltd., Beijing, China). Glyceraldehyde-3-phosphate dehydrogenase (GAPDH) was the internal reference, and the mRNA expression value was calculated through the method of 2 −∆∆Ct . All primers were designed by oligo7 and listed in Supplementary Table S3.

Western Blotting Analysis
After transfected and cultured for 48 h, cells were washed with phosphate-buffered saline (PBS) and lysed by M-PER reagents (78501, Pierce Chemical, IL, USA) containing protein inhibitors (ab201119, Abcam, Cambridge, UK)). Protein quantification was achieved by Pierce TM BCA Protein Assay (UD281372; Thermo Fisher Scientific, Waltham, MA, USA). Samples with loading buffer were loaded onto a 10% SDS-PAGE, then electrotransferred to the PVDF membrane (IPVH00010; Amersham Biosciences, NJ, USA). Then membranes were cut due to molecular weight and incubated with primary antibodies at 4 • C overnight. The membranes were washed with TBST and hybridized with secondary antibodies for one hour at room temperature. Finally, ECL chemiluminescence solution (SW2030, Biosharp, Hefei, China) exposure made the protein band visualized through an automatic chemiluminescence system (5200, Tanon, Shanghai, China). Relative protein expression levels were qualified to GAPDH, and data were obtained from independent triplicate samples. Antibodies used in this study were listed in Supplementary Table S4.

Subcellular Fractionation
The separation of nuclear and cytoplasmic fractions was conducted with the method described [104]. In brief, SH-SY5Y cells were harvested and washed with PBS twice. After resuspending and homogenization, cells were centrifuged at 400× g for 15 min at 4 • C. The cytoplasmic fraction of the supernate was added with 1 mL Trizol agent for cytoplasmic RNA extraction. The nuclear RNA was separated after being washed with the nuclear isolation buffer. The cytoplasmic RNA and nuclear RNA were separated with the Trizol agent manufacturer's instructions. The internal reference of the nuclear and cytoplasmic fraction was U6 and RPS14, respectively.

ELISA
According to the manufacturer's instructions, the human IFN-β expression in serum and culture supernatant was tested by ELISA kit (MM-51652H1, Meiman Industrial Co. Ltd., Yancheng, China). The IFN-β concentration was calculated due to its absorbance at 450 nm wavelength by a spectrophotometer (FC357, Thermo Fisher Scientific, MA, USA).

Luciferase Assay
Luciferase activity was measured through Dual Glo Luciferase Assay System (E1960, Promega, Fitchburg, WI, USA) due to the manufacturer's instructions. SH-SY5Y cells were cultured in the cell culture plate of 24 wells. Co-transfection of the plasmid and target gene in SH-SY5Y cells was performed to test luciferase activity after 24 h under the condition of 5% CO 2 at 37 • C. The Renilla luciferase reporter plasmid (pRL-TK, Promega) was used as the internal control.

Co-Immunoprecipitation Assay
Co-immunoprecipitation was carried out as previously described [20]. The plasmids pENTER-N-FLAG-cGAS and pXJ40-HA-ENV, negative control plasmids (pENTER-N-FLAG and pXJ40-HA) were transfected into HEK-293T cells at the ratio of 1:1 (5 µg + 5 µg) in 100 mm cell culture dish and incubated for 48 h under the condition of 5% CO 2 at 37 • C. After washing and lysing, cells were centrifuged at 12,000 rpm for 5 min to get the supernatant. Next, the supernatant was mixed with anti-Flag (L-1011, Bio-linkedin, Shanghai, China), anti-HA magnetic beads (L-1009, Bio-linkedin, Shanghai, China) and negative control mouse IgG antibody (dilution 1:200, AC011, ABclonal Technology, Wuhan, China) separately and maintained at 4 • C overnight. Then, the supernatant containing IgG was mixed with protein A/G magnetic beads (L-1004, Bio-linkedin, Shanghai, China) and warmly rotated for 2 h at room temperature. Finally, magnetic beads were washed with cell lysis buffer (P0013, Beyotime, Shanghai, China) and detected by western blotting.

Cell Proliferation Assay
Cell proliferation was performed with the cell counting kit 8 (CCK-8) (ZP328-1, Zomanbio, Beijing, China) according to the manufacturer's instructions. Cells were transfected with plasmids at 96-well plates and incubated for 48 h. The absorbance value at 450 nm through a micro-plate reader after 10 µL CCK8 agent was added to the medium for 45 min.

Statistical Analyses
GraphPad Prism 5 was mainly used for data analysis through Student's t-tests and one-way analysis of variance, with a significance value of p < 0.05. In addition, HERV-W ENV, linc01930, and IFN-β expression in schizophrenia patients and healthy controls were analyzed via median analyses and Mann-Whitney U analyses, with correlation analyses via Spearman's rank correlation. Data were counted at least from three replicates and displayed as the mean ± SD. * p < 0.05; ** p < 0.01; *** p < 0.0001.

Conclusions
In this paper, we found decreased linc01930 in the serum of schizophrenia, which was negatively correlated with HERV-W ENV, suggesting the promising role of linc01930 as a biomarker. We also found the increased IFN-β in schizophrenia, with a negative correlation to linc01930 and a positive correlation to HERV-W ENV. In vitro experiments demonstrated that HERV-W ENV inhibited linc01930. Additional studies suggested that linc01930, with nuclear location and noncoding ability, counteracted antiviral innate immunity, restrained neuronal apoptosis and exerted cell proliferation in neuron cells. Further studies proclaimed that HERV-W ENV induced innate antiviral immunity and neuronal apoptosis through cGAS/STING/IFN-β signaling pathway (Figure 9). . The potential role of HERV-W ENV to trigger neuronal apoptosis via innate immune activation in schizophrenia. The decreased linc01930 was negatively correlated with increased HERV-W ENV and IFN-β in schizophrenia. HERV-W ENV repressed linc01930 expression via its promoter activity. HERV-W ENV activated cGAS and STING expression and elevated IRF3 phosphorylation, while linc01930 functioned as a negative regulator to HERV-W ENV-induced cGAS and STING expression and IRF3 phosphorylation. In addition, linc01930 was involved in regulating the cGAS/STING signaling pathway induced by HERV-W ENV. Moreover, HERV-W ENV activated IFN-β expression via its promoter activity, while linc01930 inhibited linc01930 expression via its promoter activity. Furthermore, HERV-W ENV mediated the increased cGAS and Figure 9. The potential role of HERV-W ENV to trigger neuronal apoptosis via innate immune activation in schizophrenia. The decreased linc01930 was negatively correlated with increased HERV-W ENV and IFN-β in schizophrenia. HERV-W ENV repressed linc01930 expression via its promoter activity. HERV-W ENV activated cGAS and STING expression and elevated IRF3 phosphorylation, while linc01930 functioned as a negative regulator to HERV-W ENV-induced cGAS and STING expression and IRF3 phosphorylation. In addition, linc01930 was involved in regulating the cGAS/STING signaling pathway induced by HERV-W ENV. Moreover, HERV-W ENV activated IFN-β expression via its promoter activity, while linc01930 inhibited linc01930 expression via its promoter activity. Furthermore, HERV-W ENV mediated the increased cGAS and IFN-β expression and neuronal apoptosis by regulating linc01930 expression. Thus, Innate immune activation might contribute to the etiology of schizophrenia. Data Availability Statement: All data is available from the corresponding author upon request.