Regulatory T cells (Tregs) are not terminally differentiated but can acquire effector properties. Here we report an increased expression of human endogenous retrovirus 1 (HERV1-env) proteins in Tregs of patients with de novo autoimmune hepatitis and autoimmune hepatitis, which induces endoplasmic reticulum (ER) stress. HERV1-env-triggered ER stress activates all three branches (IRE1, ATF6, and PERK) of the unfolded protein response (UPR). Our coimmunoprecipitation studies show an interaction between HERV1-env proteins and the ATF6 branch of the UPR. The activated form of ATF6α activates the expression of RORC and STAT3 by binding to promoter sequences and induces IL-17A production. Silencing of HERV1-env results in recovery of Treg suppressive function. These findings identify ER stress and UPR activation as key factors driving Treg plasticity (species: human).

Autoimmune hepatitis (AIH) is a chronic necroinflammatory liver disease of unknown etiology featuring interface hepatitis by T cells, macrophages, and plasma cells that invade the periportal parenchyma (1, 2). The incidence and prevalence of AIH are 0.4 and 3.0 cases per 100,000 children, respectively (3), and ∼17 per 100,000 population in adults (4–9). Characteristic features of AIH include elevation of liver enzymes, detectable autoantibodies, elevation of serum IgG levels, and interface hepatitis with plasma cell infiltration in liver histology (10). The trigger that activates the innate and adaptive immune responses leading to tissue injury in AIH remains unknown (11). Putative mechanisms posited for autoimmune liver injury include presentation of self-antigens to naive CD4+ and CD8+ effector T cells (12–14). Depending on the prevailing cytokine milieu, differentiation of naive CD4+ T cells into Th1 cells occurs and leads to production of IL-2 and IFN-γ and concomitant activation of cytotoxic CD8+ T lymphocytes that produce IFN-γ and TNF. Exposure of hepatocytes to IFN-γ results in upregulation of MHC class I molecules and aberrant expression of MHC class II molecules, leading to further T cell activation and perpetuation of liver damage (12–14). IFN-γ also induces monocyte differentiation, promotes macrophage activation, and contributes to increased NK cell activity (15, 16). Differentiation of naive CD4+ T cells into Th2 cells leads to secretion of cytokines essential for B cell maturation to plasma cells that secrete autoantibodies. Autoantibodies can induce damage through Ab-mediated cellular cytotoxicity and complement activation (17). Although Th17 cells have been reported in AIH, their role in the pathogenesis of AIH is under investigation (18). Additionally, a possible role of T follicular helper cells in the pathogenesis of autoimmune disease is increasingly being reported (19–21). Other cell types implicated in AIH pathogenesis include γδT cells present in the liver and responsible for granzyme B and IFN-γ secretion (22) and macrophages (23). Loss of self-tolerance has also been implicated in AIH. Thus, defective immunoregulation in AIH results not only from reduced number and function of regulatory T cells (Tregs) but also from increased conversion of Tregs into effector cells (24–26). Current therapies in AIH interrupt the adaptive immune process globally but have a nonspecific target (27); moreover, 10–40% of patients have to discontinue therapy because of side effects (27). This highlights the need for improved understanding of disease pathogenesis because this may allow immune-targeted treatments with greater efficacy and fewer associated adverse effects.

The term “de novo autoimmune hepatitis” (dnAIH) in the context of liver transplantation was first used in 1998 to describe a distinct form of graft dysfunction in pediatric liver transplant recipients that clinically and histologically resembled AIH but occurred in children who underwent transplant for indications other than AIH (28). Thus, the classic features of AIH with elevation of liver enzymes, detectable autoantibodies, elevation of serum IgG levels, and interface hepatitis with plasma cell infiltration in liver histology are observed in patients with dnAIH. The frequency of dnAIH has been estimated at 1.7–11% (28–38). It can be an aggressive disease in children because chronic rejection, cirrhosis, portal hypertension and subsequent graft loss, and a need for retransplant occurred in long-term follow-up despite standard treatment (30, 34, 36, 38). The pathogenesis of post–liver transplant de novo autoimmunity, and in particular dnAIH, remains to be elucidated. The pathogenic mechanisms responsible for AIH development before liver transplant are probably similar to those that promote dnAIH, but the introduction of a donor graft with different APCs and immune reactive cells may alter the recipients’ immune response (39). Further studies are therefore needed to determine if the dnAIH form of graft dysfunction represents an autoimmune or an alloimmune reaction. A better understanding of the immunological pathogenic mechanisms will further improve the diagnosis, management, and patient outcomes of this challenging disorder.

Multiple human endogenous retroviruses (HERVs) are distributed throughout the genome as DNA fossils from prior germline infection (40). Although they form ∼8% of the human genome, HERVs usually lack function and expression because of restrictive epigenetic modifications or genetic mutations. However, HERVs may be expressed in pathological processes such as autoimmune diseases and cancers, and some viral infections activate HERV expression. Notably, infection with CMV (41), EBV infection (42–44), herpes simplex virus (45, 46), human herpesvirus 6A (47), and human herpesvirus 6B (48) all activate HERV expression. These observations have led to the hypothesis that HERV may trigger autoimmune diseases through various effects of HERV-mediated superantigen expression and immune dysregulation (40, 49); however, the exact role of HERV in triggering autoimmunity remains unclear.

Intracellular pathogens, particularly viruses, can frequently hijack organelles such as the endoplasmic reticulum (ER) to facilitate survival and replication (50). Cellular defense mechanisms have consequently evolved to detect biological perturbations frequently associated with infection, such as a rapid increase in glycoprotein production, activating the unfolded protein response (UPR) (51). The UPR is a comprehensive, three-pronged system that corrects pathological protein misfolding and aggregation in the ER (50). To combat excessive accumulation of misfolded proteins, multicellular organisms mount a coordinated response driven primarily by three ER-localized transmembrane proteins: the dual kinase/endoribonuclease inositol-requiring enzyme 1α (IRE1α/Ern1), the kinase protein kinase R-like ER-resident kinase (PERK/Eif2ak3), and activating transcription factor 6α (ATF6α/Atf6) (50). Virus-mediated UPR activation varies by strain. For instance, the African swine fever virus and lymphochoriomeningitis virus solely activate the ATF6α branch (52, 53), whereas Kaposi’s sarcoma-associated herpesvirus and herpes simplex virus 1 impact all three UPR branches (54). Unremitting UPR activation may directly contribute to failures in immunological tolerance observed in inflammatory bowel disease and rheumatoid arthritis by dangerously lowering the threshold for proinflammatory signaling and cell death (50).

The objective of this study was to assess the role of HERV in altering the immune response and Treg function in autoimmune liver disease using blood and liver tissue obtained from children with dnAIH and AIH compared with children with normal graft function and healthy children. We observed increased protein levels of HERV1-env and HERV1-env antigenemia in Tregs and plasma, respectively, of patients with AIH and dnAIH (55). We therefore hypothesized that loss of immunological tolerance in dnAIH and AIH is initiated by HERV1-mediated UPR activation (56). Because of their role in maintaining tolerance to self-antigens and curbing autoimmunity, we chose to focus on Tregs in this study. Here we show that HERV1-env, through its interaction with ATF6, induces Th17 Tregs and loss of Treg suppressor function. We also show that HERV1-env-expressing Tregs exhibit an enhanced expression of PD1 and LAG3. Thus, unremitting UPR activation induced by HERV1-env contributes to loss of immunological tolerance in autoimmune liver disease.

The study protocol was reviewed and approved by the institutional review board of Yale University. This was a cross-sectional study involving pediatric liver transplant recipients with a diagnosis of dnAIH (n = 15) and pediatric liver transplant recipients without a diagnosis of dnAIH who had normal allograft function at the time of enrollment and blood draw. These latter subjects served as the liver transplant control group (LTC) (n = 24). Other study cohorts included pediatric liver transplant recipients with acute cellular rejection (n = 15) and healthy children who had no underlying immune-mediated disorders and were not receiving any immunomodulatory agents. This latter group served as the healthy control group (HC) (n = 17). Last, nontransplanted children with a diagnosis of AIH were also included (n = 10). The transplanted subjects were enrolled when they attended their routine post-transplant clinics, the nontransplanted children with AIH were enrolled when they attended their routine pediatric hepatology clinics, and the healthy nontransplanted subjects were enrolled in response to flyers posted in New Haven, CT, as well as using the help us discover database, a database of healthy children registered as research control subjects maintained by the Yale Center for Clinical Investigation. The definition of dnAIH was as previously described: a liver transplant recipient without a history of autoimmune liver disease presenting with unknown etiology of late graft dysfunction. Late graft dysfunction was characterized by elevated aminotransferases, and graft dysfunction not due to any of the following causes: acute and chronic rejection, hepatitis B and C infection, EBV and CMV infections, vascular problems, biliary complications, drug toxicity, sepsis, recurrence of primary disease, or post-transplant lymphoproliferative disease; elevated serum IgG, positive autoantibody titers, including antinuclear Abs, anti–smooth muscle Abs, anti–liver-kidney microsomal Ab; characteristic biopsy findings of dense lymphocytic portal tract infiltrate with plasma cells; and interface hepatitis (38). The definition of AIH was as previously described (57, 58). The definition of acute cellular rejection was as previously described (59, 60).

Inclusion/exclusion criteria were established prospectively. The inclusion criteria for the liver transplant control group were age ≤20 y, serum aminotransferases and γ-glutamyl transpeptidase less than or equal to the upper limit of normal at blood draw, no past diagnosis of dnAIH, no biliary or vascular complication at the time of blood draw, no past or present history of chronic rejection, and no recent acute rejection (>1 y). The exclusion criteria included age ≥21 y, biliary/vascular complication at time of blood draw, serum aminotransferases and γ-glutamyl transpeptidase greater than the upper limit of normal at blood draw, recent acute rejection (≤1 y), and history of chronic rejection. We enrolled patients aged 0 through 20 y from January 2017 to December 2019. The rule for stopping patient enrollment and sample collection was determined by the dates as stated above. At enrollment and every 3 months over the duration of the study, 30 ml of peripheral blood was obtained from each participant. If the study participant underwent a clinically indicated liver biopsy, liver tissue was obtained for isolation of RNA from formalin-fixed, paraffin-embedded (FFPE) tissue. For the transplanted subjects with acute rejection, 30 ml of peripheral blood was obtained once a biopsy-proven diagnosis of acute rejection was made prior to commencement of treatment. Each experiment had two technical replicates with substantiation of the results. Informed consent was obtained from parents or guardians, and assent was obtained as necessary per institutional review board guidelines.

Sample size

No sample size calculation was performed; however, we believe the sample size is adequate to measure the effect size, given the highly significant differences demonstrated between the study groups.

Randomization

Because these were in vitro experiments, randomization was not performed.

Blinding

Laboratory personnel were blinded to the diagnostic category of the enrolled subjects for all experiments and data analyses in which subjects from the various diagnostic groups were used.

Replication

Biological replicates were used in all experiments. Additionally, each experiment had two technical replicates.

PBMCs were isolated from whole blood obtained from study participants by Ficoll Hypaque gradient centrifugation for cell isolation and FACS sorting as previously described (61, 62). Tregs were defined as CD4+CD25hiCD127. Th1 cells were defined as CD4+CXCR3+CCR4CCR6 and Th17 cells defined as CD4+CXCR3CCR4CCR6+.

Tregs (50,000 cells) were stimulated with PMA at 50 ng/ml and ionomycin (250 ng/ml), and intracellular staining of cytokines and FOXP3 was performed as previously described (61).

FACS-sorted memory Tregs from patients with dnAIH or AIH from and the LTC group were resuspended in PBS at a concentration of 1 × 106 cells/ml and subjected to culture in the presence of rIL-2 over 24–48 h. Tregs were then washed twice with cold PBS and resuspended in 1× binding buffer (BD Pharmingen, 556547) at a concentration of 1 × 106 cells/ml. Resuspended cells (100 μl, 1 × 105 cells) were transferred to a 5-ml V-bottomed tube, and 5 μl of annexin V and 1 μl of SYTOX Blue (Thermo Fisher Scientific, S34857) (or 5 μl of propidium iodide; BD Pharmingen, 556547) and 1 μl each of anti-human LAG3 (BioLegend, 369309) and PD1 Ab (BioLegend, 367427) were added and incubated in the dark for 15 min at room temperature. Following incubation, 400 μl of 1× binding buffer were added, and samples were analyzed by flow cytometry.

FACS-sorted Tregs were first transduced with either scramble control particle or shRNA-HERV1-env particle in the presence of rIL-2 for 24 h, then they were washed with media. Transduced Tregs were cocultured with CellTrace Violet 450–labeled effector (responder) CD4+CD25CD127+ T cells at a Treg/responder ratio of 0:1, 0.5:1, or 1:1 (i.e., 0 Tregs/50,000 responders, 25,000 Tregs/50,000 responders, 50,000 Tregs/50,000 responders), and proliferation of viable responder T cells was measured as previously described (61).

Total RNA quality was determined by estimating the A260/A280 and A260/A230 ratios using NanoDrop technology. RNA integrity was determined by running an Agilent Bioanalyzer gel, which measured the percentage of RNA fragments greater than 200 nt for each sample (also called DV200). For RNA-seq library preparation, following the Illumina TruSeq RNA Exome protocol (RS-301-2001), samples were binned on the basis of DV200 scores. High-quality samples had a DV200 >70%, medium-quality samples had a DV200 of 50–70%, low-quality samples had a DV200 of 30–50%, and samples with a DV200 <30% were not used. The RNA was fragmented using divalent cations under elevated temperature at various time periods based on DV200. cDNA was generated from the cleaved RNA fragments using random priming during first- and second-strand synthesis. Then, sequencing adapters were ligated to the resulting double-stranded cDNA fragments. The coding regions of the transcriptome were captured from this library using sequence-specific probes to create the final library. Indexed libraries that met appropriate cutoffs for both were quantified by quantitative real-time PCR (qRT-PCR) using a commercially available kit (KAPA Biosystems), and insert size distribution was determined with the Perkin Elmer LabChip GX or Agilent Bioanalyzer. Samples with a yield ≥0.5 ng/μl were used for sequencing.

Sample concentrations were normalized to 10 nM and loaded onto Illumina Rapid or high-output flow cells at a concentration that yielded 150 million–250 million passing filter clusters per lane. Samples were sequenced using 75-bp paired-end sequencing on an Illumina HiSeq 2500 according to Illumina protocols. The 6-bp index was read during an additional sequencing read that automatically followed the completion of read 1. Data generated during sequencing runs were simultaneously transferred to the YCGA high-performance computing cluster. A positive control (prepared bacteriophage PhiX library) provided by Illumina was spiked into every lane at a concentration of 0.3% to monitor sequencing quality in real time. For data analysis and storage, signal intensities were converted to individual base calls during a run using the system’s Real Time Analysis software. Base calls were transferred from the machine’s dedicated personal computer to the Yale High Performance Computing cluster via a 1-Gigabit network mount for downstream analysis. In the primary analysis, sample demultiplexing and alignment to the human genome were performed using Illumina’s CASAVA 1.8.2 software suite. The data were returned to the user if the sample error rates were less than 2% and the distribution of reads per sample in a lane was within reasonable tolerance. Data were retained on the cluster for at least 6 mo and then transferred to a tape backup system.

The RNA-seq sequencing reads were first trimmed for quality, trimming to the last base with quality ≥20 and dropping any read pair where either of the trimmed reads was shorter than 45 bp. For mRNA analysis, the trimmed reads were aligned to the hg19 reference genome by TopHat2 (63–65) using the UCSC RefSeq GTF file for gene/transcript annotations. Transcript abundance calculation and differential gene expression analysis were performed using cuffdiff (66), identifying significantly differentially expressed genes between each pair of conditions. Those genes were given to the Ingenuity Pathway Analysis software (QIAGEN, https://www.qiagenbioinformatics.com/products/ingenuitypathway-analysis) for pathway enrichment analysis.

For HERV RNA analysis, the trimmed reads were aligned to the hg38 reference genome by TopHat2 using the gEVE GTF file of 33,966 human HERV locations (67). Transcript abundance calculation and differential gene expression analysis were performed using cuffdiff, again identifying significant HERV expression differences between conditions.

RNA was isolated using the Qiagen RNeasy Micro Kit (QIAGEN), following the manufacturer’s guidelines, and converted to cDNA by reverse transcription as previously described (61, 62). For gene expression assays, Fast SYBR Green Master Mix (Applied Biosystems, 4309155) was used, and the reactions were set up following the manufacturer’s guidelines. Values are presented as the variations in cycle threshold values normalized to the 18S rRNA gene or β-actin for each sample by using the following formula: Relative RNA expression = (2-dCt) × 1000 (see Supplemental Table I for primers used).

For isolation of RNA from FFPE tissue, RNA was isolated from one to three FFPE curls using the Beckman Coulter FormaPure Total kit (C16675), following the RNA isolation only method. RNA was eluted in 40 μl of nuclease-free water. SpectraMax NanoDrop technology and Agilent Bioanalyzer 2100s DV200 score were used to assess RNA quality and quantity.

Protein was harvested using MPER solution with PhosSTOP according to the manufacturer’s instructions, and concentrations were measured using the Bradford method as previously described (62).

The protein concentration of ENV,HERV1_I was measured by sandwich ELISA using recombinant protein of ENV,HERV1_I (RKNKRNVFTQISTVENISTNIKKDIEIGSWKDKEWLPERIIKYYGPATWAQDRSWGYHTPIYMLNQIIRLQAVLEIIVNETA; Thermo Fisher custom design). Frozen plasma samples were thawed completely and centrifuged for 5 min prior to addition to the assay. Samples were run in triplicates. Briefly, each well was coated with 50 μl of ENV,HERV1_I Ab (15 μg/ml). Blocking of each well was then done using 150 μl of blocking buffer (1% BSA, 1% sucrose in 1× PBS). Plasma (200 μl) from each study subject was then added and incubated for 2 h. The second Ab of biotin-labeled anti-HERV1_I was added to bind with the protein using the biotin protein labeling kit from Roche (11418165001). Substrate with 1:10,000 HRP-streptavidin (Invitrogen, 43-4323) was then added to the well, and the plate was read at 450/595 nm.

FACS-sorted Tregs obtained from PBMCs were stimulated for 4 h with a mixture of PMA and ionomycin (Invitrogen, 00-4970-93). The cells were fixed with 1% formaldehyde, then washed and lysed in the presence of protease and phosphatase inhibitors. Chromatin fragments were prepared by MNase digestion prior to immunoprecipitation using the Pierce Agarose ChIP Kit (Thermo Fisher Scientific, 26156) according to the manufacturer’s protocol. Briefly, chromatin aliquots were incubated with 5 µg anti-ATF6α Ab, negative control IgG, or positive control anti–RNA polymerase II Ab and incubated overnight at 4°C under constant agitation. The next day, 20 μl of protein A/G plus agarose was added to each aliquot and incubated under constant agitation for 1 h. After extensive washing, the DNA-protein-Ab complexes were eluted from the agarose beads with 150 μl of provided immunoprecipitation elution buffer (1×). All samples, including a total input, were incubated for 40 min at 65°C with shaking to revert cross-linking. After treatment with NaCl and proteinase K, the immunoprecipitated DNA was recovered using a DNA Clean-Up column. The following ChIP-grade Abs were used: anti-ATF6 (Santa Cruz Biotechnology, sc-166659), anti–RNA polymerase II Ab (Thermo Fisher Scientific, 1862243), and negative control mouse IgG (Thermo Fisher Scientific, 10400C). Immunoprecipitated DNA was quantified for RORC and STAT3 promoter binding by qPCR (3 min at 95°C, 45 cycles of 95°C for 15 s, followed by 60°C for 1 min) using the Applied Biosystems 7500 Real-Time PCR System and the iQ SYBR Green SuperMix (Bio-Rad Laboratories, 170-8880). Promoter binding sites that were upstream were identified. A melting curve analysis was performed to discriminate between specific and nonspecific PCR products. The relative amount of RORC and STAT3 promoter DNA was determined using the following primers: RORC forward: 5′-AGAACACCATCTCCAGCCTCA -3′; RORC reverse: 5′-GAAGTCCTTAAATCCCAGCCAC-3′; and STAT3 forward: 5-′GCTCTCAACCTCGCCACC-3′; STAT3 reverse: 5′-CGGAATGTCCTGCTGAAAAC-3′. Data were normalized by input control DNA and expressed with respect to those of control IgG (used as a calibrator).

FACS-sorted Tregs obtained from PBMCs were stimulated for 4 h with a mixture of PMA and ionomycin (Invitrogen, 00-4970-93). Cells were lysed in Pierce IP Lysis Buffer (Thermo Fisher Scientific, 87787) containing protease and phosphatase inhibitors following the manufacturer’s protocol. Lysates were centrifuged and incubated with Protein G Sepharose 4 Fast Flow (GE Healthcare, 17-0618-01) and anti-ATF6 (Santa Cruz Biotechnology, sc-22799), anti-HERV1 (Santa Cruz Biotechnology, sc-365885), negative control mouse IgG (Thermo Fisher Scientific, 10400C), or negative control rabbit IgG (Thermo Fisher Scientific, 1862244) overnight at 4°C shaking. The beads were washed with cold PBS four times and eluted with 2-ME containing SDS sample buffer by boiling for 10 min. Protein samples were then loaded and subjected to Western blot analysis. The following Abs were used: anti-ATF6 (Santa Cruz Biotechnology, sc-22799), anti-HERV1 (Santa Cruz Biotechnology, sc-365885), and anti-GAPDH (Santa Cruz Biotechnology, sc-25778).

Lentiviral particles expressing shRNA and overexpression particles were obtained from OriGene Technologies (Rockville, MD) for env,HERV1_I. FACS-sorted CD4+CD25hiCD127lo Tregs were subjected to transduction of nonadherent primary cells following the manufacturer’s protocol. In brief, 5 × 104 Tregs were transduced with viral particles containing a vector expressing specific shRNA or a vector expressing unspecific shRNA. Transduction was carried out at a multiplicity of infection of 5 by centrifugation at 2250 rpm for 30 min at room temperature in the presence of 3 μg/ml polybrene (EMD Millipore, Billerica, MA). Cells were subsequently cultured in X-VIVO 15 media with anti-CD3, anti-CD28, both at 1 μg/ml, and IL-2 (50 U/ml) for 48 h at 37°C with 5% CO2. Cells were subsequently gently resuspended, and RNA and protein were harvested for qRT-PCR and Western blot analysis, respectively.

Patient characteristics were summarized using median and range for continuous variables and frequency and percentage for categorical variables. The Wilcoxon rank-sum test was conducted to test group differences for continuous variables. The χ2 test or Fisher’s exact test was conducted to test for associations between patient groups and categorical variables. Patients’ test results between scramble control and shRNA-HERV1-env or between control and overexpression samples were tested using the Wilcoxon signed-rank test. A two-sided p value <0.05 considered statistically significant. Analyses were carried out using SAS 9.4.3 (SAS Institute, Cary, NC).

The clinical characteristics of patients with dnAIH or AIH and control populations are presented in Table I. The duration from transplant and age at time of blood collection was not significantly different between the patients with dnAIH and the LTC group. Likewise, the age at the time of blood collection was not significantly different between the patients with AIH and the HC group. The male/female ratio for each cohort is shown in Table I.

Table I.
Patient characteristics
dnAIH (n = 15)LTC (n = 24)AR (n = 15)AIH (n = 10)HC (n = 17)p Value
Sex 5:10 17:7 5:10 2:8 8:9 0.029a 
 M:F      0.81b 
      0.27c 
Age at blood draw, mo 178.8 159.1 22.8 131.1 117.0 0.47a 
 Median (range) (30.6–299.3) (58.6–215.4) (11.4–43.4) (78.6–195.0) (66.7–207.2) 0.5c 
      0.006b 
Duration from transplant at blood draw, mo 110.7 131.3 4.0 n/a n/a 0.43a 
 Median (range) (21.1–226.4) (41.2–195.1) (1.2–36.2)   0.004b 
Age at transplant, mo 23.0 11.9 11.3 n/a n/a 0.15a 
 Median (range) (7.0–209.8) (4.3–185.1) (6.5–195.0)   0.41b 
FK level at time of blood draw, ng/ml 5.2 2.5 4.5 n/a n/a <0.001a 
 Median (range) (3.8–11.6) (1.1–4.5) (0.6–8.7)   0.67b 
Pretransplant diagnosis    n/a n/a 0.017a 
 Biliary atresia 16 10   0.014b 
 Malignancy    
 Metabolic    
 Other    
dnAIH (n = 15)LTC (n = 24)AR (n = 15)AIH (n = 10)HC (n = 17)p Value
Sex 5:10 17:7 5:10 2:8 8:9 0.029a 
 M:F      0.81b 
      0.27c 
Age at blood draw, mo 178.8 159.1 22.8 131.1 117.0 0.47a 
 Median (range) (30.6–299.3) (58.6–215.4) (11.4–43.4) (78.6–195.0) (66.7–207.2) 0.5c 
      0.006b 
Duration from transplant at blood draw, mo 110.7 131.3 4.0 n/a n/a 0.43a 
 Median (range) (21.1–226.4) (41.2–195.1) (1.2–36.2)   0.004b 
Age at transplant, mo 23.0 11.9 11.3 n/a n/a 0.15a 
 Median (range) (7.0–209.8) (4.3–185.1) (6.5–195.0)   0.41b 
FK level at time of blood draw, ng/ml 5.2 2.5 4.5 n/a n/a <0.001a 
 Median (range) (3.8–11.6) (1.1–4.5) (0.6–8.7)   0.67b 
Pretransplant diagnosis    n/a n/a 0.017a 
 Biliary atresia 16 10   0.014b 
 Malignancy    
 Metabolic    
 Other    

AR, acute rejection.

a

dnAIH versus LTC comparison.

b

dnAIH versus AR comparison.

c

AIH versus HC comparison.

To evaluate the presence of and characterize the HERV loci in patients with autoimmune liver disease, we subjected FACS-sorted Tregs from the blood and FFPE liver tissue of patients with dnAIH and AIH to RNA-seq, confirmatory qRT-PCR, and Western blot analysis. Plasma of patients with dnAIH and AIH was subjected to ELISA. RNA-seq results showed increased expression of HERV in Tregs of patients with dnAIH as compared with patients with normal liver graft function (Fig. 1A). Furthermore, there were significant differences in the expression of HERV in the liver between these two groups of patients (Fig. 1B). The RNA-seq result was confirmed by qRT-PCR of both Tregs and liver (Fig. 1C, 1D). On the basis of RNA data, we examined whether the increased expression of HERV1-env was leading to protein. Tregs of patients with dnAIH and AIH were therefore subjected to Western blot analysis. We observed increased HERV1-env levels in Tregs of patients with dnAIH and AIH compared with patients with normal liver graft function and healthy children (Fig. 1E, 1F). We also detected HERV1-env antigenemia in the plasma of patients with dnAIH and AIH (Fig. 1G).

FIGURE 1.

Upregulation of human endogenous retrovirus in patients with autoimmune liver disease. RNA-seq depicting upregulation of HERV in peripheral blood memory Tregs (A) and liver (B) of patients with dnAIH (n = 6) compared with patients with normal graft function (n = 9). HERV expression fragments per kilobase of exon model per million reads mapped values in Tregs (A) and liver (B) of patients with dnAIH for HERV locations found to be significantly differentially expressed between patients with dnAIH and patients with normal graft function by a TopHat/CuffDiff analysis (using the gEVE database of human ERV elements [67]). At all locations differentially expressed between the sets of samples from patients with dnAIH and patients with normal graft function (LTC), no expression was found in the LTC sample, so N-fold were –inf or inf in all cases. Significant HERV1-env expression in peripheral blood memory Tregs of patients with dnAIH versus patients with normal graft function (LTC) and patients with acute rejection was confirmed by qRT-PCR and shown in (C). Significant HERV1-env expression in peripheral blood memory Tregs of patients with AIH versus the HC group is shown in (C). Significant HERV1-env expression in the liver of patients with dnAIH versus patients with normal graft function (LTC) and patients with acute rejection was confirmed by qRT-PCR and is shown in (D). Significantly increased HERV1-env levels in peripheral blood memory Tregs of patients with dnAIH versus patients with normal graft function (LTC) and patients with acute rejection and of patients with AIH versus HC is shown in (E and F). Significant HERV1-env antigenemia in patients with dnAIH versus patients with normal graft function (LTC) and in patients with AIH versus HC is shown in (G). GAPDH was used as an internal loading control in Western blot experiments. PCR results are normalized to the housekeeping gene β-actin. Results are shown as individual dot plots with a line at the median (C, D, F and G). ****p < 0.0001, ***p < 0.001, **p < 0.01, *p < 0.05. AR, acute rejection.

FIGURE 1.

Upregulation of human endogenous retrovirus in patients with autoimmune liver disease. RNA-seq depicting upregulation of HERV in peripheral blood memory Tregs (A) and liver (B) of patients with dnAIH (n = 6) compared with patients with normal graft function (n = 9). HERV expression fragments per kilobase of exon model per million reads mapped values in Tregs (A) and liver (B) of patients with dnAIH for HERV locations found to be significantly differentially expressed between patients with dnAIH and patients with normal graft function by a TopHat/CuffDiff analysis (using the gEVE database of human ERV elements [67]). At all locations differentially expressed between the sets of samples from patients with dnAIH and patients with normal graft function (LTC), no expression was found in the LTC sample, so N-fold were –inf or inf in all cases. Significant HERV1-env expression in peripheral blood memory Tregs of patients with dnAIH versus patients with normal graft function (LTC) and patients with acute rejection was confirmed by qRT-PCR and shown in (C). Significant HERV1-env expression in peripheral blood memory Tregs of patients with AIH versus the HC group is shown in (C). Significant HERV1-env expression in the liver of patients with dnAIH versus patients with normal graft function (LTC) and patients with acute rejection was confirmed by qRT-PCR and is shown in (D). Significantly increased HERV1-env levels in peripheral blood memory Tregs of patients with dnAIH versus patients with normal graft function (LTC) and patients with acute rejection and of patients with AIH versus HC is shown in (E and F). Significant HERV1-env antigenemia in patients with dnAIH versus patients with normal graft function (LTC) and in patients with AIH versus HC is shown in (G). GAPDH was used as an internal loading control in Western blot experiments. PCR results are normalized to the housekeeping gene β-actin. Results are shown as individual dot plots with a line at the median (C, D, F and G). ****p < 0.0001, ***p < 0.001, **p < 0.01, *p < 0.05. AR, acute rejection.

Close modal

Importantly, there was an absence of a similar increase of HERV1-env in patients with acute rejection. Together, these results show that HERV expression is increased in the liver, Tregs, and plasma of patients with dnAIH and AIH, and this increase is not a result of ongoing inflammation, because we were unable to observe a similar increase of HERVs in patients with acute rejection, a disease process well known to be associated with inflammation both histologically and biochemically (68). Furthermore, CD69+ and CD69+HLA-DR+ expression was not different in Tregs of patients with AIH versus HCs (data not shown), suggesting that HERV upregulation in Tregs of patients with dnAIH and AIH is not a consequence of inflammation in the liver microenvironment.

Our RNA-seq results showed enrichment of ATF6 gene in peripheral blood memory Tregs of patients with dnAIH compared with patients with normal graft function (Supplemental Fig. 1A). We examined FACS sorted memory Tregs from the blood and FFPE liver tissue of patients with dnAIH and AIH for evidence of ER stress and UPR activation. Our results showed increased ATF6α, spliced X box protein (XBP1s), phosphorylated translation initiation factor (eIF2αp), and C/EBP homologous protein in the Tregs and liver of patients with dnAIH and AIH compared with patients with normal graft function and healthy children (Fig. 2A2D). Under homeostatic conditions, the luminal domains of each of the ER-localized transmembrane proteins are physically bound and held inactive by the chaperone protein, binding Ig protein (BiP) (69, 70). However, BiP has a higher affinity for misfolded proteins than for UPR mediators; thus, accumulation of misfolded substrates titrates away this steady-state inhibition. Release from BiP leads to activation of all three UPR branches (50). The activation of UPR branches leads to production of transcription factors, which upregulate ER chaperones, including BiP. Our results showed upregulation of BiP in patients with dnAIH and AIH (more so in patients with dnAIH than in patients with AIH) (Fig. 2E, 2F). Together, these results confirm that ER stress and UPR activation occur in autoimmune liver disease.

FIGURE 2.

Effect of stressed memory Tregs and liver. Activation of UPR mediators in peripheral blood memory Tregs and liver of patients with autoimmune liver disease. Significant UPR activation (as evidenced by ATF6A, XBP-1S, and CHOP upregulation) in peripheral blood memory Tregs of patients with dnAIH versus patients with normal graft function (LTC group) and patients with acute rejection is confirmed by qRT-PCR and shown in (A). Similarly, significant UPR activation in peripheral blood memory Tregs of patients with AIH versus HC, also confirmed by qRT-PCR, is shown in (A). Significant UPR activation in liver of patients with dnAIH versus patients with normal graft function (LTC group) and patients with acute rejection is confirmed by qRT-PCR and shown in (B). In separate experiments, protein was extracted from peripheral blood memory Tregs and used to study the three major UPR branches. Significantly increased levels of UPR mediators, specifically cleavage of ATF6α, phosphorylation of phospho-eIF2α, splicing of xbp1, and CHOP, in peripheral blood memory Tregs of patients with dnAIH versus patients with normal graft function (LTC group) and patients with acute rejection (AR) and of patients with AIH versus HC are shown in (C and D). Representative blot of three separate experiments is shown (C). Dissociation of BiP from the ER localized transmembrane proteins is shown in (E and F), further confirming ER stress with UPR activation in patients with autoimmune liver disease. GAPDH was used as an internal loading control in Western blot experiments. PCR results are normalized to housekeeping gene β-actin. Results are shown as individual dot plots with a line at the median (A, B, D and E). ****p < 0.0001, ***p < 0.001, **p < 0.01, *p < 0.05.

FIGURE 2.

Effect of stressed memory Tregs and liver. Activation of UPR mediators in peripheral blood memory Tregs and liver of patients with autoimmune liver disease. Significant UPR activation (as evidenced by ATF6A, XBP-1S, and CHOP upregulation) in peripheral blood memory Tregs of patients with dnAIH versus patients with normal graft function (LTC group) and patients with acute rejection is confirmed by qRT-PCR and shown in (A). Similarly, significant UPR activation in peripheral blood memory Tregs of patients with AIH versus HC, also confirmed by qRT-PCR, is shown in (A). Significant UPR activation in liver of patients with dnAIH versus patients with normal graft function (LTC group) and patients with acute rejection is confirmed by qRT-PCR and shown in (B). In separate experiments, protein was extracted from peripheral blood memory Tregs and used to study the three major UPR branches. Significantly increased levels of UPR mediators, specifically cleavage of ATF6α, phosphorylation of phospho-eIF2α, splicing of xbp1, and CHOP, in peripheral blood memory Tregs of patients with dnAIH versus patients with normal graft function (LTC group) and patients with acute rejection (AR) and of patients with AIH versus HC are shown in (C and D). Representative blot of three separate experiments is shown (C). Dissociation of BiP from the ER localized transmembrane proteins is shown in (E and F), further confirming ER stress with UPR activation in patients with autoimmune liver disease. GAPDH was used as an internal loading control in Western blot experiments. PCR results are normalized to housekeeping gene β-actin. Results are shown as individual dot plots with a line at the median (A, B, D and E). ****p < 0.0001, ***p < 0.001, **p < 0.01, *p < 0.05.

Close modal

Previous studies have shown that activation of ATF6 can be triggered by overexpression of a membrane protein (71); thus, we asked whether ATF6 can be activated by HERV1-env. To test this, we first examined the interaction between HERV1-env and ATF6 in FACS-sorted memory Tregs (from the blood of patients with dnAIH and AIH) using coimmunoprecipitation. Our results showed a protein interaction between ATF6 and HERV1-env (Fig. 3A) in patients with dnAIH and AIH and absence of a similar interaction in patients with normal graft function and in healthy children, suggesting that ATF6 activation and hence ER stress and UPR activation are triggered by the envelope protein of HERV1. To confirm that the envelope protein of HERV1 is inducing ER stress, we used a targeted approach and overexpressed HERV1-env on FACS-sorted memory Tregs of healthy children with a lentiviral transduction system. We also silenced HERV1-env on FACS-sorted memory Tregs of patients with dnAIH by shRNA. The transduction efficiency of the overexpression and silencing experiments was determined at both mRNA (Supplemental Fig. 1B) and protein levels (Supplemental Fig. 1C), achieving >50% knockdown of HERV1-env with shRNA-HERV1-env. Silencing of HERV1-env downregulated ATF6A expression and UPR activation. HERV1-env overexpression upregulated ATF6A expression and UPR activation (Fig. 3B3F).

FIGURE 3.

HERV1-env activates memory Treg ATF6α to induce UPR activation. (A) HERV1-env has a stronger interaction with ATF6α (and vice versa) in peripheral blood memory Tregs from patients with dnAIH and AIH than in memory Tregs of patients with normal graft function (LTC group), patients with acute rejection (AR), and HC. (B) The effect of HERV1-env silencing and overexpression in peripheral blood memory Tregs of patients with dnAIH (n = 10) and HC (n = 10), respectively, at the mRNA level. Downregulation of ATF6A and the mediators of UPR activation occurs in memory Tregs of patients with dnAIH following HERV1-env silencing and is shown in (B). Upregulation of ATF6A and the mediators of UPR activation occurs in memory Tregs of HC following lentiviral transduction of HERV1-env (multiplicity of infection [MOI] 7.5) and is shown in (B). (C and D) The effect of HERV1-env overexpression in peripheral blood memory Tregs of HC at the protein level. Increased levels of ATF6α and the mediators of UPR activation occur following lentiviral transduction of HERV1-env (MOI 7.5) and are shown in (C and D). (E and F) The effect of HERV1-env silencing in memory Tregs of patients with dnAIH at the protein level. Decreased levels of ATF6α and the mediators of UPR activation occurs following HERV1-env silencing in memory Tregs of patients with dnAIH and is shown in (E and F). Representative immunoblot of three separate experiments is shown. GAPDH was used as an internal loading control for Western blot experiments. PCR results are normalized to housekeeping gene β-actin. Results are shown as individual dot plots with a line at the median. ****p < 0.0001, ***p < 0.001, **p < 0.01, *p < 0.05.

FIGURE 3.

HERV1-env activates memory Treg ATF6α to induce UPR activation. (A) HERV1-env has a stronger interaction with ATF6α (and vice versa) in peripheral blood memory Tregs from patients with dnAIH and AIH than in memory Tregs of patients with normal graft function (LTC group), patients with acute rejection (AR), and HC. (B) The effect of HERV1-env silencing and overexpression in peripheral blood memory Tregs of patients with dnAIH (n = 10) and HC (n = 10), respectively, at the mRNA level. Downregulation of ATF6A and the mediators of UPR activation occurs in memory Tregs of patients with dnAIH following HERV1-env silencing and is shown in (B). Upregulation of ATF6A and the mediators of UPR activation occurs in memory Tregs of HC following lentiviral transduction of HERV1-env (multiplicity of infection [MOI] 7.5) and is shown in (B). (C and D) The effect of HERV1-env overexpression in peripheral blood memory Tregs of HC at the protein level. Increased levels of ATF6α and the mediators of UPR activation occur following lentiviral transduction of HERV1-env (MOI 7.5) and are shown in (C and D). (E and F) The effect of HERV1-env silencing in memory Tregs of patients with dnAIH at the protein level. Decreased levels of ATF6α and the mediators of UPR activation occurs following HERV1-env silencing in memory Tregs of patients with dnAIH and is shown in (E and F). Representative immunoblot of three separate experiments is shown. GAPDH was used as an internal loading control for Western blot experiments. PCR results are normalized to housekeeping gene β-actin. Results are shown as individual dot plots with a line at the median. ****p < 0.0001, ***p < 0.001, **p < 0.01, *p < 0.05.

Close modal

STAT3 and RORC are transcriptional targets of ATF6A, and our RNA-seq analysis showed that they were significantly enriched in memory Tregs of patients with dnAIH (Supplemental Fig. 1D, 1E). Because STAT3 and RORC are transcription factors required for differentiation of naive CD4+ T cells to Th17 cells (72, 73), we asked whether ATF6 binds to STAT3 and RORC. To address this question, FACS-sorted memory Tregs from patients with dnAIH and AIH were subjected to ChIP DNA qPCR. qPCR of ChIP DNA suggests strong binding of ATF6 to the transcription promoter regions of both RORC and STAT3 in patients with dnAIH and AIH (Fig. 4A, 4B). Silencing of HERV1-env in memory Tregs of patients with dnAIH downregulated IL-17A mRNA (Fig. 4C). RORC and STAT3 mRNA were similarly downregulated. Together, these results suggest ATF6, by binding to the transcription promoter regions of known Th17 transcription factors, may drive Treg differentiation to Th17-like Tregs in patients with autoimmune liver disease.

FIGURE 4.

Effect of memory Treg ER stress on adaptive immunity. ATF6 drives IL-17A transcription in peripheral blood memory Tregs. (A and B). ChIP DNA qPCR confirms strong upstream binding of ATF6 to the transcription promoter regions of RORC and STAT3 in patients with dnAIH and AIH and is shown in (A and B). Silencing of HERV1-env (which downregulates ATF6α) abrogates IL-17A transcription in memory Tregs of patients with dnAIH (n = 10) and is shown in (C). PCR result is normalized to the housekeeping gene 18S. Results are shown as individual dot plots with a line at the median. ***p < 0.001, **p < 0.01, *p < 0.05. AR, acute rejection.

FIGURE 4.

Effect of memory Treg ER stress on adaptive immunity. ATF6 drives IL-17A transcription in peripheral blood memory Tregs. (A and B). ChIP DNA qPCR confirms strong upstream binding of ATF6 to the transcription promoter regions of RORC and STAT3 in patients with dnAIH and AIH and is shown in (A and B). Silencing of HERV1-env (which downregulates ATF6α) abrogates IL-17A transcription in memory Tregs of patients with dnAIH (n = 10) and is shown in (C). PCR result is normalized to the housekeeping gene 18S. Results are shown as individual dot plots with a line at the median. ***p < 0.001, **p < 0.01, *p < 0.05. AR, acute rejection.

Close modal

The two responses to ER stress are adaptation or apoptosis (50). To examine the effect of HERV1-env on the viability of Tregs, FACS-sorted memory Tregs from patients with dnAIH or AIH and LTC were subjected to culture in the presence of rIL-2 for 24 and 48 h. Although memory Tregs of patients with dnAIH or AIH and LTC similarly underwent early apoptosis (i.e., increased annexin V uptake, without SYTOX Blue uptake; Fig. 5, upper panel), dnAIH and AIH Tregs undergoing early apoptosis displayed enhanced expression of PD1 and LAG3 when compared with LTC Tregs undergoing early apoptosis (Fig. 5, lower panel). The gating strategy is shown in Supplemental Fig. 2A. HERV1-env silencing on Tregs of patients with dnAIH and AIH did not decrease the frequency of Treg cells undergoing early apoptosis, and, although Treg cell number showed a minimal increase, it did not achieve statistical significance (Supplemental Fig. 2B). Together, these results indirectly suggest a possible relationship between HERV1-env and expression of immune checkpoints.

FIGURE 5.

Memory Tregs of patients with dnAIH and AIH undergoing early apoptosis display enhanced expression of PD1 and LAG3. Peripheral blood memory Tregs from patients with dnAIH, AIH, and LTC cultured in the presence of rIL-2 for 24 and 48 h. Cell viability was monitored using SYTOX Blue and annexin V. Upper panel, frequency of Tregs undergoing early apoptosis (i.e., percentage of annexin V+ SYTOX Blue cells in FACS analysis) is similar between the three patient groups (LTC 5.31%; dnAIH 7.28%; AIH 4.32%). Summary data are shown for 24 and 48 h. Results are shown as individual dot plots. Lower panel, dnAIH and AIH Tregs undergoing early apoptosis display enhanced PD1+LAG3+ expression on FACS analysis compared with LTC Tregs (LTC 0%; dnAIH 76.7%; AIH 35.7%), suggesting a relationship between HERV1-env and expression of immune checkpoints. Summary data are shown for 24 h. Results are shown as individual dot plots. dnAIH n = 5; AIH n = 3; LTC n = 10. ****p < 0.0001.

FIGURE 5.

Memory Tregs of patients with dnAIH and AIH undergoing early apoptosis display enhanced expression of PD1 and LAG3. Peripheral blood memory Tregs from patients with dnAIH, AIH, and LTC cultured in the presence of rIL-2 for 24 and 48 h. Cell viability was monitored using SYTOX Blue and annexin V. Upper panel, frequency of Tregs undergoing early apoptosis (i.e., percentage of annexin V+ SYTOX Blue cells in FACS analysis) is similar between the three patient groups (LTC 5.31%; dnAIH 7.28%; AIH 4.32%). Summary data are shown for 24 and 48 h. Results are shown as individual dot plots. Lower panel, dnAIH and AIH Tregs undergoing early apoptosis display enhanced PD1+LAG3+ expression on FACS analysis compared with LTC Tregs (LTC 0%; dnAIH 76.7%; AIH 35.7%), suggesting a relationship between HERV1-env and expression of immune checkpoints. Summary data are shown for 24 h. Results are shown as individual dot plots. dnAIH n = 5; AIH n = 3; LTC n = 10. ****p < 0.0001.

Close modal

Tregs are not a terminally differentiated population but show some degree of plasticity and can, under specific environmental conditions, acquire the phenotype of effector T cells (74). Knowing that a subset of human Tregs produce IFN-γ (75) or IL-17 (61) and thus share features of Th1 and Th17 effector cells, specifically, given our published observation of IL-17A and IFN-γ expression and production from Tregs of patients with dnAIH (61), we examined whether HERV1-env contributes to loss of immune tolerance by lowering the threshold for proinflammatory signaling in Tregs. FACS-sorted memory Tregs from healthy children were transduced with HERV1-env overexpression construct, and IFN-γ and IL-17A gene and protein expression were measured by qRT-PCR and FACS, respectively, following stimulation with PMA/ionomycin. Successful transduction was confirmed by qRT-PCR and Western blot analysis (Supplemental Fig. 1B, 1c). HERV1-env-expressing Tregs have increased IFN-γ and IL-17A expression as compared with control particles (Fig. 6A), and this was confirmed at the protein level for IL-17A and IFN-γ (Fig. 6B). The gating strategy is shown in Supplemental Fig. 2C). Importantly, these HERV1-env-transduced Tregs maintained Foxp3 expression (data not shown). Moreover, HERV1-env-transduced Tregs were indeed shown to demonstrate UPR activation (Fig. 3B3D) similar to the observation in Tregs of patients with dnAIH and AIH (Fig. 2A, 2C2F).

FIGURE 6.

Alteration of proinflammatory signaling and function in memory Tregs. (A) Upregulation of IL-17A and IFN-γ mRNA in memory Tregs of healthy children (n = 10) following HERV1-env lentiviral transduction. (B) Increased IL-17A and IFN-γ production from memory Tregs of healthy children following HERV1-env lentiviral transduction. IL-17A, control particle 1.19%; overexpression particle, 8.66%. IFN-γ, control particle 1.38%; overexpression particle 6.17%. PCR result is normalized to the housekeeping gene 18S. Results are shown as individual dot plots with a line at the median. (C) Tregs from patients with dnAIH (transduced with either scramble control particle or shRNA-HERV1-env particle) were cocultured in 1:1 and 1:0.5 ratios with CellTrace Violet 450–labeled effector (responder) CD4+CD25CD127+ T cells, and proliferation of viable responder T cells was measured. The stimulus used was Dynabeads anti-CD3, anti-CD28 beads. Three representative histograms are shown. Top panel: scramble control particle Treg/responder cell ratio 0.5:1, 74.4%; shRNA-HERV1-env particle Treg/responder cell ratio 0.5:1; 54.1%. Middle panel: scramble control particle Treg/responder cell ratio 0.5:1; 47.6%. shRNA-HERV1-env particle Treg/responder cell ratio 0.5:1; 19.4%. Lower panel: scramble control particle Treg/responder cell ratio 0.5:1; 67%. shRNA-HERV1-env particle Treg/responder cell ratio 1:1; 6.73%. Summary data for suppression assay shown (n = 7). ****p < 0.0001, **p < 0.01, *p < 0.05. (D) Schematic representation showing HERV1-env-induced UPR activation leading to differentiation of memory Tregs to proinflammatory cytokine–producing Tregs. FMO, fluorescence minus one; SSC-A, side scatter area.

FIGURE 6.

Alteration of proinflammatory signaling and function in memory Tregs. (A) Upregulation of IL-17A and IFN-γ mRNA in memory Tregs of healthy children (n = 10) following HERV1-env lentiviral transduction. (B) Increased IL-17A and IFN-γ production from memory Tregs of healthy children following HERV1-env lentiviral transduction. IL-17A, control particle 1.19%; overexpression particle, 8.66%. IFN-γ, control particle 1.38%; overexpression particle 6.17%. PCR result is normalized to the housekeeping gene 18S. Results are shown as individual dot plots with a line at the median. (C) Tregs from patients with dnAIH (transduced with either scramble control particle or shRNA-HERV1-env particle) were cocultured in 1:1 and 1:0.5 ratios with CellTrace Violet 450–labeled effector (responder) CD4+CD25CD127+ T cells, and proliferation of viable responder T cells was measured. The stimulus used was Dynabeads anti-CD3, anti-CD28 beads. Three representative histograms are shown. Top panel: scramble control particle Treg/responder cell ratio 0.5:1, 74.4%; shRNA-HERV1-env particle Treg/responder cell ratio 0.5:1; 54.1%. Middle panel: scramble control particle Treg/responder cell ratio 0.5:1; 47.6%. shRNA-HERV1-env particle Treg/responder cell ratio 0.5:1; 19.4%. Lower panel: scramble control particle Treg/responder cell ratio 0.5:1; 67%. shRNA-HERV1-env particle Treg/responder cell ratio 1:1; 6.73%. Summary data for suppression assay shown (n = 7). ****p < 0.0001, **p < 0.01, *p < 0.05. (D) Schematic representation showing HERV1-env-induced UPR activation leading to differentiation of memory Tregs to proinflammatory cytokine–producing Tregs. FMO, fluorescence minus one; SSC-A, side scatter area.

Close modal

Silencing of HERV1-env on Tregs of patients with dnAIH led to recovery of suppressive function of Tregs (Fig. 6C). Together, these data suggest that HERV1-env, by inducing unremitting UPR activation, may contribute to failures in immunological tolerance by lowering the threshold for proinflammatory signaling in Tregs (Fig. 6D).

Last, given the participation of Th1 and Th17 cells in liver injury in both AIH and dnAIH, we questioned if HERV1-env upregulation and UPR activation are specific to Tregs. To address this, FACS-sorted Th1 and Th17 cells from blood of children with dnAIH, AIH, or acute rejection, children with liver transplants without dnAIH, and healthy children were subjected to qRT-PCR for HERV1-env expression and downstream mediators of UPR activation. Compared with HERV1-env expression in Tregs of patients with dnAIH, Th1 and Th17 cells of patients with dnAIH and AIH did not demonstrate significant HERV1-env expression (Supplemental Fig. 3A), nor did they demonstrate significant ATF6A, CHOP, or XBP-1s (Supplemental Fig. 3B) expression, supporting HERV1 upregulation being specific to Tregs in autoimmune liver disease.

Here we confirm the previously reported involvement of HERVs in the pathogenesis of several autoimmune diseases (76–86) and extend it to the autoimmune liver diseases dnAIH and AIH. Our data further suggest that ER stress occurs in Tregs and livers of patients with dnAIH and AIH. Here we show that memory Tregs of patients with dnAIH produce IL-17A and IFN-γ under ER stress induced by HERV1-env reactivation. Thus, HERV1-env-induced ER stress, mediated through ATF6α, drives Tregs to differentiate and acquire a Th1-like and Th17-like effector phenotype. Furthermore, silencing of HERV1-env in Tregs led to recovery of the suppressive function of Tregs. Our data thus reveal another mechanism by which Tregs demonstrate plasticity and acquire effector-like properties.

We have previously shown that peripheral blood memory Tregs of patients with dnAIH produce IFN-γ and IL-17 (61). Although the production of IFN-γ is driven in part by inflammasome activation of CD14+ monocytes (62), it does not explain the production of IL-17. Th17 cells contribute to autoimmunity by producing proinflammatory cytokines (including but not limited to IL-17) and inducing hepatocytes to secrete IL-6 (18), which further enhances Th17 cell activation. Our current work suggests that ATF6α, by binding to the transcription promoter regions of RORC and STAT3 in peripheral blood memory Tregs, drives differentiation of these Tregs to a Th17-like phenotype. Importantly, HERV1-env silencing downregulates ATF6A, as well as IL-17A expression and production. Similarly, HERV1-env overexpression drives differentiation of Tregs to Th1-like Tregs that produce IFN-γ. Exposure of hepatocytes to IFN-γ results in upregulation of MHC class I molecules and aberrant expression of MHC class II molecules, leading to further T cell activation and perpetuation of liver damage (13, 14). Moreover, IFN-γ also induces monocyte differentiation, promotes macrophage and immature dendritic cell activation (15), and contributes to increased NK cell activity (16). To the best of our knowledge, this is the first report of ATF6α driving Tregs to acquire a proinflammatory phenotype.

The development of autoimmune disease is enhanced by breakdown of self-tolerance mechanisms. Some published data indicate a functional defect in Tregs in AIH (87–89). Defective immune regulation in AIH also results from increased conversion of Tregs into effector cells (25). Our observation that, upon silencing of HERV1-env, the Treg suppressive function is restored is important because it confirms that unremitting UPR activation may alter tolerance thresholds by promoting excessive proinflammatory signaling or stress-mediated exposure of novel immunogenic or misfolded autoantigens (90). Unraveling how ER stress alters the threshold for proinflammatory signaling may help explain how genetic susceptibilities in patients with autoimmune disease collaborate with environmental stimuli to corrupt homeostatic tolerance mechanisms (50).

Env-HERV1-expressing memory Tregs of patients with dnAIH and AIH also exhibited enhanced expression of PD1 and LAG3. Ma et al. have reported on ER stress induced by cholesterol through XBP1, resulting in PD1 expression and exhaustion in CD8 tumor-infiltrating T cells (91). Our results show activation of the three major UPR branches, including XBP1 in dnAIH and AIH; moreover, env-HERV1 overexpression upregulated XBP1s, whereas env-HERV1 knockdown downregulated XBP1s. It is thus possible that env-HERV1 induces XBP1 expression, which then regulates the expression of PD-1 on memory Tregs. This would have to be further investigated and exceeds the scope of this paper.

Although EBV has been reported to activate HERVs (42–44), the amount of EBV DNA was not increased in plasma of these patients, suggesting to us that EBV was not an issue here. Interestingly, although a similar level of env-HERV1 antigenemia was observed in subjects with dnAIH and AIH (Fig. 1G), env-HERV1 expression level in Tregs was higher in subjects with dnAIH than in those with AIH (Fig. 1C, 1F). One could argue that if env-HERV1 activation is a result of an exogenous viral infection, then the effects of such a viral infection may be enhanced in subjects on immunosuppression, although this would not explain the higher env-HERV1 expression level in Tregs of subjects with dnAIH than in those with AIH.

A critical question that arises is why env-HERV1 expression and the subsequent ATF6 interaction, UPR activation, and all the downstream molecular and functional consequences are restricted to Tregs and exclude conventional T cells. We have documented the absence of env-HERV1 expression in Th1 and Th17 cells from patients with AIH and dnAIH (Supplemental Fig. 3A), and we have shown diminished UPR responses in Th1 and Th17 cells (Supplemental Fig. 3B). Although the trigger(s) of HERV expression in AIH and dnAIH are not known, others have reported that activation of HERV may lead to differential responses in Tregs versus conventional T cells during chronic viral herpes virus infection (92).

Last, in genetically modified models and immunodeficiency syndromes linked with autoimmune disease, loss of Treg homeostasis may result in prolonged inflammatory responses (93); although env-HERV1-induced UPR activation may well contribute to both neoantigen formation and apoptosis in Tregs, this was not examined in our present study.

Taken together, our results shed light on the events that trigger proinflammatory cytokine production by memory Tregs and uncover HERV1-env-induced ER stress as a central player in the initiation of this process. Upon ER stress, ATF6α through its binding affinity to Th17 transcription factors induces IL-17A upregulation and production from Tregs, coupled with loss of Treg suppressor function. Furthermore, env-HERV1-expressing Tregs undergo apoptosis and exhibit enhanced expression of PD1 and LAG3. These results highlight the key role of HERV1-env in contributing to autoimmunity and identify its envelope protein as a target to treat dnAIH and AIH by acting at the early steps in the differentiation and destruction of Tregs.

Although our findings demonstrate that HERV1-env induced Treg dysfunction, this work does have some limitations. ER stress was induced by HERV1-env, but what triggers HERV1-env activation is unclear. HERV1-env-expressing Tregs had upregulated immune checkpoint expression, but how HERV1-env induces enhanced PD1 and LAG3 expression is unclear. The detailed transcriptional program of XBP-1 related to Treg exhaustion in autoimmune liver disease should be investigated but is beyond the scope of this article.

The authors have no financial conflicts of interest.

This work was supported by the Clinical and Translational Science Award Grant UL1 TR000142 from the National Center for Advancing Translational Sciences, a component of the National Institutes of Health, and by National Institute of Diabetes and Digestive and Kidney Diseases of the National Institutes of Health under Award P30KD034989.

The online version of this article contains supplemental material.

AIH

autoimmune hepatitis

BiP

binding Ig protein

ChIP

chromatin immunoprecipitation

ER

endoplasmic reticulum

HERV1-env

human endogenous retrovirus 1

qRT-PCR

quantitative real-time PCR

RNA-seq

RNA sequencing

Treg

regulatory T cell

UPR

unfolded protein response

1
Manns
,
M. P.
,
A. W.
Lohse
,
D.
Vergani
.
2015
.
Autoimmune hepatitis—update 2015
.
J. Hepatol.
62
(
1 Suppl
):
S100
S111
.
2
Golbabapour
,
S.
,
K.
Bagheri-Lankarani
,
S.
Ghavami
,
B.
Geramizadeh
.
2020
.
Autoimmune hepatitis and stellate cells: an insight into the role of autophagy
.
Curr. Med. Chem.
27
:
6073
6095
.
3
Deneau
,
M.
,
M. K.
Jensen
,
J.
Holmen
,
M. S.
Williams
,
L. S.
Book
,
S. L.
Guthery
.
2013
.
Primary sclerosing cholangitis, autoimmune hepatitis, and overlap in Utah children: epidemiology and natural history
.
Hepatology
58
:
1392
1400
.
4
Boberg
,
K. M.
,
E.
Aadland
,
J.
Jahnsen
,
N.
Raknerud
,
M.
Stiris
,
H.
Bell
.
1998
.
Incidence and prevalence of primary biliary cirrhosis, primary sclerosing cholangitis, and autoimmune hepatitis in a Norwegian population
.
Scand. J. Gastroenterol.
33
:
99
103
.
5
Delgado
,
J. S.
,
A.
Vodonos
,
S.
Malnick
,
O.
Kriger
,
R.
Wilkof-Segev
,
B.
Delgado
,
V.
Novack
,
A.
Rosenthal
,
Y.
Menachem
,
E.
Melzer
,
A.
Fich
.
2013
.
Autoimmune hepatitis in southern Israel: a 15-year multicenter study
.
J. Dig. Dis.
14
:
611
618
.
6
Feld
,
J. J.
,
E. J.
Heathcote
.
2003
.
Epidemiology of autoimmune liver disease
.
J. Gastroenterol. Hepatol.
18
:
1118
1128
.
7
Ngu
,
J. H.
,
K.
Bechly
,
B. A.
Chapman
,
M. J.
Burt
,
M. L.
Barclay
,
R. B.
Gearry
,
C. A.
Stedman
.
2010
.
Population-based epidemiology study of autoimmune hepatitis: a disease of older women?
J. Gastroenterol. Hepatol.
25
:
1681
1686
.
8
van Gerven
,
N. M.
,
B. J.
Verwer
,
B. I.
Witte
,
K. J.
van Erpecum
,
H. R.
van Buuren
,
I.
Maijers
,
A. P.
Visscher
,
E. C.
Verschuren
,
B.
van Hoek
,
M. J.
Coenraad
, et al
Dutch Autoimmune hepatitis STUDY group
.
2014
.
Epidemiology and clinical characteristics of autoimmune hepatitis in the Netherlands
.
Scand. J. Gastroenterol.
49
:
1245
1254
.
9
Werner
,
M.
,
H.
Prytz
,
B.
Ohlsson
,
S.
Almer
,
E.
Björnsson
,
A.
Bergquist
,
S.
Wallerstedt
,
H.
Sandberg-Gertzén
,
R.
Hultcrantz
,
P.
Sangfelt
, et al
.
2008
.
Epidemiology and the initial presentation of autoimmune hepatitis in Sweden: a nationwide study
.
Scand. J. Gastroenterol.
43
:
1232
1240
.
10
Harada
,
K.
,
N. C.
Hiep
,
H.
Ohira
.
2017
.
Challenges and difficulties in pathological diagnosis of autoimmune hepatitis
.
Hepatol. Res.
47
:
963
971
.
11
Dyson
,
J. K.
,
E.
De Martin
,
G. N.
Dalekos
,
J. P. H.
Drenth
,
J.
Herkel
,
S. G.
Hubscher
,
D.
Kelly
,
M.
Lenzi
,
P.
Milkiewicz
,
Y. H.
Oo
, et al
IAIHG Consortium
.
2019
.
Review article: unanswered clinical and research questions in autoimmune hepatitis—conclusions of the International Autoimmune Hepatitis Group Research Workshop
.
Aliment. Pharmacol. Ther.
49
:
528
536
.
12
Ichiki
,
Y.
,
C. A.
Aoki
,
C. L.
Bowlus
,
S.
Shimoda
,
H.
Ishibashi
,
M. E.
Gershwin
.
2005
.
T cell immunity in autoimmune hepatitis
.
Autoimmun. Rev.
4
:
315
321
.
13
Lobo-Yeo
,
A.
,
G.
Senaldi
,
B.
Portmann
,
A. P.
Mowat
,
G.
Mieli-Vergani
,
D.
Vergani
.
1990
.
Class I and class II major histocompatibility complex antigen expression on hepatocytes: a study in children with liver disease
.
Hepatology
12
:
224
232
.
14
Senaldi
,
G.
,
A.
Lobo-Yeo
,
A. P.
Mowat
,
G.
Mieli-Vergani
,
D.
Vergani
.
1991
.
Class I and class II major histocompatibility complex antigens on hepatocytes: importance of the method of detection and expression in histologically normal and diseased livers
.
J. Clin. Pathol.
44
:
107
114
.
15
Delneste
,
Y.
,
P.
Charbonnier
,
N.
Herbault
,
G.
Magistrelli
,
G.
Caron
,
J. Y.
Bonnefoy
,
P.
Jeannin
.
2003
.
Interferon-gamma switches monocyte differentiation from dendritic cells to macrophages
.
Blood
101
:
143
150
.
16
Schroder
,
K.
,
P. J.
Hertzog
,
T.
Ravasi
,
D. A.
Hume
.
2004
.
Interferon-gamma: an overview of signals, mechanisms and functions
.
J. Leukoc. Biol.
75
:
163
189
.
17
Liberal
,
R.
,
E. L.
Krawitt
,
J. M.
Vierling
,
M. P.
Manns
,
G.
Mieli-Vergani
,
D.
Vergani
.
2016
.
Cutting edge issues in autoimmune hepatitis
.
J. Autoimmun.
75
:
6
19
.
18
Zhao
,
L.
,
Y.
Tang
,
Z.
You
,
Q.
Wang
,
S.
Liang
,
X.
Han
,
D.
Qiu
,
J.
Wei
,
Y.
Liu
,
L.
Shen
, et al
.
2011
.
Interleukin-17 contributes to the pathogenesis of autoimmune hepatitis through inducing hepatic interleukin-6 expression
.
PLoS One
6
:
e18909
.
19
Ma
,
C. S.
,
E. K.
Deenick
.
2014
.
Human T follicular helper (Tfh) cells and disease
.
Immunol. Cell Biol.
92
:
64
71
.
20
Abe
,
K.
,
A.
Takahashi
,
H.
Imaizumi
,
M.
Hayashi
,
K.
Okai
,
Y.
Kanno
,
H.
Watanabe
,
H.
Ohira
.
2016
.
Interleukin-21 plays a critical role in the pathogenesis and severity of type I autoimmune hepatitis
.
Springerplus
5
:
777
.
21
Kimura
,
N.
,
S.
Yamagiwa
,
T.
Sugano
,
T.
Setsu
,
K.
Tominaga
,
H.
Kamimura
,
M.
Takamura
,
S.
Terai
.
2018
.
Possible involvement of chemokine C-C receptor 7- programmed cell death-1+ follicular helper T-cell subset in the pathogenesis of autoimmune hepatitis
.
J. Gastroenterol. Hepatol.
33
:
298
306
.
22
Wen
,
L.
,
M.
Peakman
,
G.
Mieli-Vergani
,
D.
Vergani
.
1992
.
Elevation of activated gamma delta T cell receptor bearing T lymphocytes in patients with autoimmune chronic liver disease
.
Clin. Exp. Immunol.
89
:
78
82
.
23
Grønbaek
,
H.
,
M.
Kreutzfeldt
,
K.
Kazankov
,
N.
Jessen
,
T.
Sandahl
,
S.
Hamilton-Dutoit
,
H.
Vilstrup
,
H. J.
Møller
.
2016
.
Single-centre experience of the macrophage activation marker soluble (s)CD163 — associations with disease activity and treatment response in patients with autoimmune hepatitis
.
Aliment. Pharmacol. Ther.
44
:
1062
1070
.
24
Liberal
,
R.
,
C. R.
Grant
,
B. S.
Holder
,
J.
Cardone
,
M.
Martinez-Llordella
,
Y.
Ma
,
M. A.
Heneghan
,
G.
Mieli-Vergani
,
D.
Vergani
,
M. S.
Longhi
.
2015
.
In autoimmune hepatitis type 1 or the autoimmune hepatitis–sclerosing cholangitis variant defective regulatory T-cell responsiveness to IL-2 results in low IL-10 production and impaired suppression
.
Hepatology
62
:
863
875
.
25
Grant
,
C. R.
,
R.
Liberal
,
B. S.
Holder
,
J.
Cardone
,
Y.
Ma
,
S. C.
Robson
,
G.
Mieli-Vergani
,
D.
Vergani
,
M. S.
Longhi
.
2014
.
Dysfunctional CD39POS regulatory T cells and aberrant control of T-helper type 17 cells in autoimmune hepatitis
.
Hepatology
59
:
1007
1015
.
26
Mieli-Vergani
,
G.
,
D.
Vergani
,
A. J.
Czaja
,
M. P.
Manns
,
E. L.
Krawitt
,
J. M.
Vierling
,
A. W.
Lohse
,
A. J.
Montano-Loza
.
2018
.
Autoimmune hepatitis
.
Nat. Rev. Dis. Primers
4
:
18017
.
27
Johnson
,
P. J.
,
I. G.
McFarlane
,
R.
Williams
.
1995
.
Azathioprine for long-term maintenance of remission in autoimmune hepatitis
.
N. Engl. J. Med.
333
:
958
963
.
28
Kerkar
,
N.
,
N.
Hadzić
,
E. T.
Davies
,
B.
Portmann
,
P. T.
Donaldson
,
M.
Rela
,
N. D.
Heaton
,
D.
Vergani
,
G.
Mieli-Vergani
.
1998
.
De-novo autoimmune hepatitis after liver transplantation
.
Lancet
351
:
409
413
.
29
Hernandez
,
H. M.
,
P.
Kovarik
,
P. F.
Whitington
,
E. M.
Alonso
.
2001
.
Autoimmune hepatitis as a late complication of liver transplantation
.
J. Pediatr. Gastroenterol. Nutr.
32
:
131
136
.
30
Gupta
,
P.
,
J.
Hart
,
J. M.
Millis
,
D.
Cronin
,
L.
Brady
.
2001
.
De novo hepatitis with autoimmune antibodies and atypical histology: a rare cause of late graft dysfunction after pediatric liver transplantation
.
Transplantation
71
:
664
668
.
31
Andries
,
S.
,
L.
Casamayou
,
C.
Sempoux
,
M.
Burlet
,
R.
Reding
,
J.
Bernard Otte
,
J. P.
Buts
,
E.
Sokal
.
2001
.
Posttransplant immune hepatitis in pediatric liver transplant recipients: incidence and maintenance therapy with azathioprine
.
Transplantation
72
:
267
272
.
32
Spada
,
M.
,
A.
Bertani
,
A.
Sonzogni
,
W.
Petz
,
S.
Riva
,
G.
Torre
,
M. L.
Melzi
,
D.
Alberti
,
M.
Colledan
,
A.
Segalin
, et al
.
2001
.
A cause of late graft dysfunction after liver transplantation in children: de-novo autoimmune hepatitis
.
Transplant. Proc.
33
:
1747
1748
.
33
Petz
,
W.
,
A.
Sonzogni
,
A.
Bertani
,
M.
Spada
,
A.
Lucianetti
,
M.
Colledan
,
B.
Gridelli
.
2002
.
A cause of late graft dysfunction after pediatric liver transplantation: de novo autoimmune hepatitis
.
Transplant. Proc.
34
:
1958
1959
.
34
Miyagawa-Hayashino
,
A.
,
H.
Haga
,
H.
Egawa
,
Y.
Hayashino
,
T.
Sakurai
,
S.
Minamiguchi
,
K.
Tanaka
,
T.
Manabe
.
2004
.
Outcome and risk factors of de novo autoimmune hepatitis in living-donor liver transplantation
.
Transplantation
78
:
128
135
.
35
Gibelli
,
N. E.
,
U.
Tannuri
,
E. S.
Mello
,
E. R.
Cançado
,
M. M.
Santos
,
A. A.
Ayoub
,
J. G.
Maksoud-Filho
,
M. C.
Velhote
,
M. M.
Silva
,
M. L.
Pinho-Apezzato
,
J. G.
Maksoud
.
2006
.
Successful treatment of de novo autoimmune hepatitis and cirrhosis after pediatric liver transplantation
.
Pediatr. Transplant.
10
:
371
376
.
36
Venick
,
R. S.
,
S. V.
McDiarmid
,
D. G.
Farmer
,
J.
Gornbein
,
M. G.
Martin
,
J. H.
Vargas
,
M. E.
Ament
,
R. W.
Busuttil
.
2007
.
Rejection and steroid dependence: unique risk factors in the development of pediatric posttransplant de novo autoimmune hepatitis
.
Am. J. Transplant.
7
:
955
963
.
37
Cho
,
J. M.
,
K. M.
Kim
,
S. H.
Oh
,
Y. J.
Lee
,
K. W.
Rhee
,
E.
Yu
.
2011
.
De novo autoimmune hepatitis in Korean children after liver transplantation: a single institution’s experience
.
Transplant. Proc.
43
:
2394
2396
.
38
Ekong
,
U. D.
,
P.
McKiernan
,
M.
Martinez
,
S.
Lobritto
,
D.
Kelly
,
V. L.
Ng
,
E. M.
Alonso
,
Y.
Avitzur
.
2017
.
Long-term outcomes of de novo autoimmune hepatitis in pediatric liver transplant recipients
.
Pediatr. Transplant.
21
:
e12945
.
39
Stirnimann
,
G.
,
M.
Ebadi
,
A. J.
Czaja
,
A. J.
Montano-Loza
.
2019
.
Recurrent and de novo autoimmune hepatitis
.
Liver Transpl.
25
:
152
166
.
40
Gröger
,
V.
,
H.
Cynis
.
2018
.
Human endogenous retroviruses and their putative role in the development of autoimmune disorders such as multiple sclerosis
.
Front. Microbiol.
9
:
265
.
41
Assinger
,
A.
,
K. C.
Yaiw
,
I.
Göttesdorfer
,
C.
Leib-Mösch
,
C.
Söderberg-Nauclér
.
2013
.
Human cytomegalovirus (HCMV) induces human endogenous retrovirus (HERV) transcription
.
Retrovirology
10
:
132
.
42
Hsiao
,
F. C.
,
M.
Lin
,
A.
Tai
,
G.
Chen
,
B. T.
Huber
.
2006
.
Cutting edge: Epstein-Barr virus transactivates the HERV-K18 superantigen by docking to the human complement receptor 2 (CD21) on primary B cells
.
J. Immunol.
177
:
2056
2060
.
43
Sutkowski
,
N.
,
G.
Chen
,
G.
Calderon
,
B. T.
Huber
.
2004
.
Epstein-Barr virus latent membrane protein LMP-2A is sufficient for transactivation of the human endogenous retrovirus HERV-K18 superantigen
.
J. Virol.
78
:
7852
7860
.
44
Sutkowski
,
N.
,
B.
Conrad
,
D. A.
Thorley-Lawson
,
B. T.
Huber
.
2001
.
Epstein-Barr virus transactivates the human endogenous retrovirus HERV-K18 that encodes a superantigen
.
Immunity
15
:
579
589
.
45
Ruprecht
,
K.
,
K.
Obojes
,
V.
Wengel
,
F.
Gronen
,
K. S.
Kim
,
H.
Perron
,
J.
Schneider-Schaulies
,
P.
Rieckmann
.
2006
.
Regulation of human endogenous retrovirus W protein expression by herpes simplex virus type 1: implications for multiple sclerosis
.
J. Neurovirol.
12
:
65
71
.
46
Kwun
,
H. J.
,
H. J.
Han
,
W. J.
Lee
,
H. S.
Kim
,
K. L.
Jang
.
2002
.
Transactivation of the human endogenous retrovirus K long terminal repeat by herpes simplex virus type 1 immediate early protein 0
.
Virus Res.
86
:
93
100
.
47
Tai
,
A. K.
,
J.
Luka
,
D.
Ablashi
,
B. T.
Huber
.
2009
.
HHV-6A infection induces expression of HERV-K18-encoded superantigen
.
J. Clin. Virol.
46
:
47
48
.
48
Turcanova
,
V. L.
,
B.
Bundgaard
,
P.
Höllsberg
.
2009
.
Human herpesvirus-6B induces expression of the human endogenous retrovirus K18-encoded superantigen
.
J. Clin. Virol.
46
:
15
19
.
49
Grandi
,
N.
,
E.
Tramontano
.
2018
.
HERV envelope proteins: physiological role and pathogenic potential in cancer and autoimmunity
.
Front. Microbiol.
9
:
462
.
50
Bettigole
,
S. E.
,
L. H.
Glimcher
.
2015
.
Endoplasmic reticulum stress in immunity
.
Annu. Rev. Immunol.
33
:
107
138
.
51
Liberman
,
E.
,
Y. L.
Fong
,
M. J.
Selby
,
Q. L.
Choo
,
L.
Cousens
,
M.
Houghton
,
T. S.
Yen
.
1999
.
Activation of the grp78 and grp94 promoters by hepatitis C virus E2 envelope protein
.
J. Virol.
73
:
3718
3722
.
52
Galindo
,
I.
,
B.
Hernáez
,
R.
Muñoz-Moreno
,
M. A.
Cuesta-Geijo
,
I.
Dalmau-Mena
,
C.
Alonso
.
2012
.
The ATF6 branch of unfolded protein response and apoptosis are activated to promote African swine fever virus infection
.
Cell Death Dis.
3
:
e341
.
53
Pasqual
,
G.
,
D. J.
Burri
,
A.
Pasquato
,
J. C.
de la Torre
,
S.
Kunz
.
2011
.
Role of the host cell’s unfolded protein response in arenavirus infection
.
J. Virol.
85
:
1662
1670
.
54
Johnston
,
B. P.
,
C.
McCormick
.
2019
.
Herpesviruses and the unfolded protein response
.
Viruses
12
:
17
.
55
Yao
,
J.
,
A.
Ling
,
A.
Arterbery
,
Y.
Avitzur
,
S.
Lobritto
,
M.
Martinez
,
U.
Ekong
.
2019
.
Treg plasticity is induced by ENV,HERV1_I and UPR activation in de novo autoimmune hepatitis [abstract]
.
Am. J. Transplant.
19
(
Suppl 3
):
A38
.
56
Li
,
Y.
,
Y.
Xia
,
X.
Cheng
,
D. E.
Kleiner
,
S. M.
Hewitt
,
J.
Sproch
,
T.
Li
,
H.
Zhuang
,
T. J.
Liang
.
2019
.
Hepatitis B surface antigen activates unfolded protein response in forming ground glass hepatocytes of chronic hepatitis B
.
Viruses
11
:
386
.
57
Hennes
,
E. M.
,
M.
Zeniya
,
A. J.
Czaja
,
A.
Parés
,
G. N.
Dalekos
,
E. L.
Krawitt
,
P. L.
Bittencourt
,
G.
Porta
,
K. M.
Boberg
,
H.
Hofer
, et al
International Autoimmune Hepatitis Group
.
2008
.
Simplified criteria for the diagnosis of autoimmune hepatitis
.
Hepatology
48
:
169
176
.
58
Mieli-Vergani
,
G.
,
S.
Heller
,
P.
Jara
,
D.
Vergani
,
M. H.
Chang
,
T.
Fujisawa
,
R. P.
González-Peralta
,
D.
Kelly
,
N.
Mohan
,
U.
Shah
,
K. F.
Murray
.
2009
.
Autoimmune hepatitis
.
J. Pediatr. Gastroenterol. Nutr.
49
:
158
164
.
59
Demetris
,
A. J.
,
K. P.
Batts
,
A. P.
Dhillon
,
L.
Ferrell
,
J.
Fung
,
S. A.
Geller
,
J.
Hart
,
P.
Hayry
,
W. J.
Hofmann
,
S.
Hubscher
, et al
.
1997
.
Banff schema for grading liver allograft rejection: an international consensus document
.
Hepatology
25
:
658
663
.
60
Demetris
,
A. J.
,
C.
Bellamy
,
S. G.
Hübscher
,
J.
O’Leary
,
P. S.
Randhawa
,
S.
Feng
,
D.
Neil
,
R. B.
Colvin
,
G.
McCaughan
,
J. J.
Fung
, et al
.
2016
.
2016 Comprehensive update of the Banff Working Group on Liver Allograft Pathology: introduction of antibody-mediated rejection
.
Am. J. Transplant.
16
:
2816
2835
.
61
Arterbery
,
A. S.
,
A.
Osafo-Addo
,
Y.
Avitzur
,
M.
Ciarleglio
,
Y.
Deng
,
S. J.
Lobritto
,
M.
Martinez
,
D. A.
Hafler
,
M.
Kleinewietfeld
,
U. D.
Ekong
.
2016
.
Production of proinflammatory cytokines by monocytes in liver-transplanted recipients with de novo autoimmune hepatitis is enhanced and induces TH1-like regulatory T cells
.
J. Immunol.
196
:
4040
4051
.
62
Arterbery
,
A. S.
,
J.
Yao
,
A.
Ling
,
Y.
Avitzur
,
M.
Martinez
,
S.
Lobritto
,
Y.
Deng
,
G.
Geliang
,
S.
Mehta
,
G.
Wang
, et al
.
2018
.
Inflammasome priming mediated via Toll-like receptors 2 and 4, induces Th1-like regulatory T cells in de novo autoimmune hepatitis
.
Front. Immunol.
9
:
1612
.
63
Trapnell
,
C.
,
L.
Pachter
,
S. L.
Salzberg
.
2009
.
TopHat: discovering splice junctions with RNA-Seq
.
Bioinformatics
25
:
1105
1111
.
64
Trapnell
,
C.
,
A.
Roberts
,
L.
Goff
,
G.
Pertea
,
D.
Kim
,
D. R.
Kelley
,
H.
Pimentel
,
S. L.
Salzberg
,
J. L.
Rinn
,
L.
Pachter
.
2012
.
Differential gene and transcript expression analysis of RNA-seq experiments with TopHat and Cufflinks. [Published erratum appears in 2014 Nat. Protoc. 9: 2513.]
Nat. Protoc.
7
:
562
578
.
65
Trapnell
,
C.
,
L.
Pachter
,
S. L.
Salzberg
.
2009
.
TopHat: discovering splice junctions with RNA-seq
.
Bioinformatics
25
:
1105
1111
.
66
Trapnell
,
C.
,
D. G.
Hendrickson
,
M.
Sauvageau
,
L.
Goff
,
J. L.
Rinn
,
L.
Pachter
.
2013
.
Differential analysis of gene regulation at transcript resolution with RNA-seq
.
Nat. Biotechnol.
31
:
46
53
.
67
Nakagawa
,
S.
,
M. U.
Takahashi
.
2016
.
gEVE: a genome-based endogenous viral element database provides comprehensive viral protein-coding sequences in mammalian genomes
.
Database (Oxford)
2016
:
baw087
.
68
Demetris
,
A.
,
D.
Adams
,
C.
Bellamy
,
K.
Blakolmer
,
A.
Clouston
,
A. P.
Dhillon
,
J.
Fung
,
A.
Gouw
,
B.
Gustafsson
,
H.
Haga
, et al
.
2000
.
Update of the International Banff Schema for Liver Allograft Rejection: working recommendations for the histopathologic staging and reporting of chronic rejection
.
Hepatology
31
:
792
799
.
69
Ye
,
J.
,
R. B.
Rawson
,
R.
Komuro
,
X.
Chen
,
U. P.
Davé
,
R.
Prywes
,
M. S.
Brown
,
J. L.
Goldstein
.
2000
.
ER stress induces cleavage of membrane-bound ATF6 by the same proteases that process SREBPs
.
Mol. Cell
6
:
1355
1364
.
70
Bertolotti
,
A.
,
Y.
Zhang
,
L. M.
Hendershot
,
H. P.
Harding
,
D.
Ron
.
2000
.
Dynamic interaction of BiP and ER stress transducers in the unfolded-protein response
.
Nat. Cell Biol.
2
:
326
332
.
71
Maiuolo
,
J.
,
S.
Bulotta
,
C.
Verderio
,
R.
Benfante
,
N.
Borgese
.
2011
.
Selective activation of the transcription factor ATF6 mediates endoplasmic reticulum proliferation triggered by a membrane protein
.
Proc. Natl. Acad. Sci. USA
108
:
7832
7837
.
72
Yang
,
X. O.
,
A. D.
Panopoulos
,
R.
Nurieva
,
S. H.
Chang
,
D.
Wang
,
S. S.
Watowich
,
C.
Dong
.
2007
.
STAT3 regulates cytokine-mediated generation of inflammatory helper T cells
.
J. Biol. Chem.
282
:
9358
9363
.
73
Yang
,
X. O.
,
B. P.
Pappu
,
R.
Nurieva
,
A.
Akimzhanov
,
H. S.
Kang
,
Y.
Chung
,
L.
Ma
,
B.
Shah
,
A. D.
Panopoulos
,
K. S.
Schluns
, et al
.
2008
.
T helper 17 lineage differentiation is programmed by orphan nuclear receptors ROR alpha and ROR gamma
.
Immunity
28
:
29
39
.
74
Kitz
,
A.
,
M.
Dominguez-Villar
.
2017
.
Molecular mechanisms underlying Th1-like Treg generation and function
.
Cell. Mol. Life Sci.
74
:
4059
4075
.
75
Dominguez-Villar
,
M.
,
C. M.
Baecher-Allan
,
D. A.
Hafler
.
2011
.
Identification of T helper type 1-like, Foxp3+ regulatory T cells in human autoimmune disease
.
Nat. Med.
17
:
673
675
.
76
Tugnet
,
N.
,
P.
Rylance
,
D.
Roden
,
M.
Trela
,
P.
Nelson
.
2013
.
Human endogenous retroviruses (HERVs) and autoimmune rheumatic disease: is there a link?
Open Rheumatol. J.
7
:
13
21
.
77
Rucheton
,
M.
,
H.
Graafland
,
H.
Fanton
,
L.
Ursule
,
P.
Ferrier
,
C. J.
Larsen
.
1985
.
Presence of circulating antibodies against gag-gene MuLV proteins in patients with autoimmune connective tissue disorders
.
Virology
144
:
468
480
.
78
Laska
,
M. J.
,
T.
Brudek
,
K. K.
Nissen
,
T.
Christensen
,
A.
Møller-Larsen
,
T.
Petersen
,
B. A.
Nexø
.
2012
.
Expression of HERV-Fc1, a human endogenous retrovirus, is increased in patients with active multiple sclerosis
.
J. Virol.
86
:
3713
3722
.
79
Curtin
,
F.
,
C.
Bernard
,
S.
Levet
,
H.
Perron
,
H.
Porchet
,
J.
Médina
,
S.
Malpass
,
D.
Lloyd
,
R.
Simpson
;
RAINBOW-T1D investigators
.
2018
.
A new therapeutic approach for type 1 diabetes: rationale for GNbAC1, an anti-HERV-W-Env monoclonal antibody
.
Diabetes Obes. Metab.
20
:
2075
2084
.
80
Perron
,
H.
,
R.
Germi
,
C.
Bernard
,
M.
Garcia-Montojo
,
C.
Deluen
,
L.
Farinelli
,
R.
Faucard
,
F.
Veas
,
I.
Stefas
,
B. O.
Fabriek
, et al
.
2012
.
Human endogenous retrovirus type W envelope expression in blood and brain cells provides new insights into multiple sclerosis disease
.
Mult. Scler.
18
:
1721
1736
.
81
Faucard
,
R.
,
A.
Madeira
,
N.
Gehin
,
F. J.
Authier
,
P. A.
Panaite
,
C.
Lesage
,
I.
Burgelin
,
M.
Bertel
,
C.
Bernard
,
F.
Curtin
, et al
.
2016
.
Human endogenous retrovirus and neuroinflammation in chronic inflammatory demyelinating polyradiculoneuropathy
.
EBioMedicine
6
:
190
198
.
82
Feschotte
,
C.
,
C.
Gilbert
.
2012
.
Endogenous viruses: insights into viral evolution and impact on host biology
.
Nat. Rev. Genet.
13
:
283
296
.
83
Volkman
,
H. E.
,
D. B.
Stetson
.
2014
.
The enemy within: endogenous retroelements and autoimmune disease
.
Nat. Immunol.
15
:
415
422
.
84
Dolei
,
A.
2006
.
Endogenous retroviruses and human disease
.
Expert Rev. Clin. Immunol.
2
:
149
167
.
85
Kremer
,
D.
,
T.
Schichel
,
M.
Förster
,
N.
Tzekova
,
C.
Bernard
,
P.
van der Valk
,
J.
van Horssen
,
H. P.
Hartung
,
H.
Perron
,
P.
Küry
.
2013
.
Human endogenous retrovirus type W envelope protein inhibits oligodendroglial precursor cell differentiation
.
Ann. Neurol.
74
:
721
732
.
86
van Horssen
,
J.
,
S.
van der Pol
,
P.
Nijland
,
S.
Amor
,
H.
Perron
.
2016
.
Human endogenous retrovirus W in brain lesions: rationale for targeted therapy in multiple sclerosis
.
Mult. Scler. Relat. Disord.
8
:
11
18
.
87
Longhi
,
M. S.
,
Y.
Ma
,
G.
Mieli-Vergani
,
D.
Vergani
.
2012
.
Regulatory T cells in autoimmune hepatitis
.
J. Hepatol.
57
:
932
934
.
88
Ferri
,
S.
,
M. S.
Longhi
,
C.
De Molo
,
C.
Lalanne
,
P.
Muratori
,
A.
Granito
,
M. J.
Hussain
,
Y.
Ma
,
M.
Lenzi
,
G.
Mieli-Vergani
, et al
.
2010
.
A multifaceted imbalance of T cells with regulatory function characterizes type 1 autoimmune hepatitis
.
Hepatology
52
:
999
1007
.
89
Longhi
,
M. S.
,
Y.
Ma
,
R. R.
Mitry
,
D. P.
Bogdanos
,
M.
Heneghan
,
P.
Cheeseman
,
G.
Mieli-Vergani
,
D.
Vergani
.
2005
.
Effect of CD4+ CD25+ regulatory T-cells on CD8 T-cell function in patients with autoimmune hepatitis
.
J. Autoimmun.
25
:
63
71
.
90
Todd
,
D. J.
,
A. H.
Lee
,
L. H.
Glimcher
.
2008
.
The endoplasmic reticulum stress response in immunity and autoimmunity
.
Nat. Rev. Immunol.
8
:
663
674
.
91
Ma
,
X.
,
E.
Bi
,
Y.
Lu
,
P.
Su
,
C.
Huang
,
L.
Liu
,
Q.
Wang
,
M.
Yang
,
M. F.
Kalady
,
J.
Qian
, et al
.
2019
.
Cholesterol induces CD8+ T cell exhaustion in the tumor microenvironment
.
Cell Metab.
30
:
143
156.e5
.
92
Punkosdy
,
G. A.
,
M.
Blain
,
D. D.
Glass
,
M. M.
Lozano
,
L.
O’Mara
,
J. P.
Dudley
,
R.
Ahmed
,
E. M.
Shevach
.
2011
.
Regulatory T-cell expansion during chronic viral infection is dependent on endogenous retroviral superantigens
.
Proc. Natl. Acad. Sci. USA
108
:
3677
3682
.
93
Marré
,
M. L.
,
J. D.
Piganelli
.
2017
.
Environmental factors contribute to β cell endoplasmic reticulum stress and neo-antigen formation in type 1 diabetes
.
Front. Endocrinol. (Lausanne)
8
:
262
.

Supplementary data