Immune responses to Mycobacterial heat shock protein 70 accompany self-reactivity to human BiP in rheumatoid arthritis

Rheumatoid arthritis (RA) is an autoimmune disease, and a member of human heat shock protein (HSP) 70 protein family, Binding Immunoglobulin Protein (BiP), has been identified as an important autoantigen for T and B cells. We herein focused on Mycobacterial (Myc) HSPs and immune responses to MycHSPs in RA patients. Serum titers of antibodies against MycHSP70 were significantly elevated in RA patients and correlated with serum anti-BiP antibody titers. A MycHSP70-derived HLA-DR4 major epitope was identified using the proliferative capacity of RA PBMCs as an indicator. The major epitope, MycHSP70287–306, was located at the corresponding position in the major epitope for human BiP336–355, and a strong correlation was found between the proliferation of PBMCs in response to MycHSP70287–306 and BiP336–355. The immunization of HLA-DR4 transgenic mice with MycHSP70 induced the proliferation of T cells and development of anti-BiP antibodies. In contrast, the oral administration of MycHSP70287–306 resulted in the amelioration of collagen-induced arthritis, serum antibody responses, and T cell proliferation. In conclusion, immune responses to MycHSP70 were associated with adaptive immunity against BiP in RA, and could be an important mechanism underlying the development of autoimmunity.

Other studies have established several MycHSP70-specific T cell clones with proliferative capacities and IFN-γ production potentials 11 . Therefore, the precise features of MycHSP70-specific T cells in RA remain unclear. The results of the present study revealed a close relationship between immune responses to MycHSP70 and human BiP in RA patients, which could support the importance of Myc and human HSPs in RA immunity.

Results
Serum anti-bacterial and human HSP antibodies in RA. Serum antibody titers for human and the corresponding bacterial HSPs were measured in RA patients and healthy donors (HDs) (Fig. 1). Cardiovascular disease patients were excluded because of the presence of serum anti-human HSP70 antibodies in these patients 12 .
Anti-human BiP antibody titers were significantly higher in RA patients than in HDs (Fig. 1A), whereas serum anti-human HSP60 antibody titers were similar (Fig. 1B). Anti-MycHSP70 antibody titers were also increased in RA sera (Fig. 1A), whereas anti-MycHSP65 antibody titers were not (Fig. 1B). The results obtained for anti-human HSP60 and anti-MycHSP65 antibodies were consistent with previous findings 13,14 . Anti-human HSP40 antibody titers were significantly higher in RA patients than in HDs (Fig. 1C), whereas no significant difference was observed in serum anti-human Cpn10 antibody titers (Fig. 1D). As a model of microbial mucosal exposure, we selected E. coli HSPs as a control. Although sequence similarity between MycHSPs and E.coli HSPs was up to 60%, no significant differences were observed in antibody titers against E. coli HSPs between RA patients and HDs (Fig. 1). We then found a correlation between anti-human HSP antibody titers and anti-MycHSP antibody titers ( Fig. 2A,B). Anti-MycHSP70 antibody titers and anti-citrullinated BiP antibody titers, in particular, showed a clear positive correlation (Fig. 2A). Anti-MycHSP70 antibody titers were significantly increased in RA patients and were associated with anti-human BiP and citrullinated BiP antibody titers.
A binding assay showed that MycHSP70 287-306 bound to the HLA-DRB1*0401 and 0405 molecule (Fig. 4A). Alanine scanning of this peptide showed that the substitution of F293 (P1), V296 (P4), and D299 (P7) significantly decreased its binding ability (Fig. 4B). These results demonstrated that MycHSP70 287-306 is an immunogenic HLA-DR4 epitope in RA patients.  Immunization of MycHSP70 broke self-tolerance in mice. We next addressed adaptive immunity induced by MycHSP70 287-306 in vivo. The immunization of HLA-DR4 transgenic mice with the MycHSP70 protein led to an elevation in serum antibody titers to MycHSP70 and BiP (Fig. 5A). Furthermore, significant CD4 + T cell proliferation was induced in response to the MycHSP70 287-306 peptide, providing additional evidence that the MycHSP70 287-306 peptide was an HLA-DR4 epitope (Fig. 5B). The significant proliferation of CD4 + T cells in response to the BiP 336-355 peptide was also observed (Fig. 5B). These results indicate that immune responses to MycHSP70 induce the loss of self-tolerance to BiP.
Oral administration of the MycHSP70 epitope ameliorated collagen-induced arthritis. We also examined the therapeutic effects of the oral administration of the MycHSP70 287-306 peptide on collagen-induced arthritis (CIA) in DBA/1 J mice. The oral administration of the MycHSP70 287-306 peptide significantly improved the development of arthritis and histological scores in CIA (Fig. 6A,B). The proliferation of CD4 + T cells in response to the MycHSP70 287-306 and BiP 336-355 peptides was reduced by the orally administered MycHSP70 287-306 peptide (Fig. 6C). The oral administration of the MycHSP70 287-306 peptide also suppressed bovine type II collagen-specific T cell proliferation (Fig. 6C) and pro-inflammatory cytokine secretion from CD4 + T cells (Supplementary Figure S2). Serum anti-BiP and anti-BC II antibody titers were significantly decreased in MycHSP70 287-306 -treated mice (Fig. 6D). Furthermore, the secretion of IL-10 and TGF-β 1 from CD4 + T cells was induced by MycHSP70 287-306 and BiP 336-355 recall stimulations in MycHSP70 287-306 -treated mice (Fig. 6E,F). The amount of cytokines released in response to HSP70 epitopes was previously reported to be lower than that from orally tolerized TCR-transgenic CD4 + T cells 15,16 . This finding suggests that the frequencies of MycHSP70 287-306 and BiP 336-355 -specific T cells were limited in this model. The frequency of CD4 + CD25 + Foxp3 + regulatory T cells was increased by the oral administration of the MycHSP70 287-306 peptide to CIA mice (Supplementary Figure  S2). These results indicate that the induction of tolerance to MycHSP70 ameliorated inflammatory arthritis and tolerance to autoantigens, including BiP.

Discussion
Environmental factors account for half of the causes of seropositive RA 17 . With the exception of a smoking habit, most environmental risks remain unknown because difficulties are associated with following environmental exposure from birth to the age of arthritis onset. Since a T cell clone specific to the Mycobacterial-derived epitope, MycHSP65 180-188 , has arthritogenic capacity, Mycobacterial exposure was presented as a potential candidate environmental risk factor for RA 6 . RA has been shown to increase the risk of Mycobacterial infection independent of immunosuppressive therapies 18 . Nevertheless, evidence for the relationship between Mycobacterial infection with the pathogenesis and exacerbation of RA is limited. We herein demonstrated that T and B cell responses to MycHSP70 were significantly enhanced in RA, which suggests Mycobacterial exposure in RA patients. Moreover, PBMC proliferative responses to MycHSP70 287-306 and BiP 336-355 were closely associated. T cell repertoire analysis demonstrated the clonal expansion of T cells in RA, and the increase observed in proliferation against these epitopes could be partially explained by RA PBMCs containing more antigen-experienced clones 19 . Moreover, we previously reported that BiP 336-355 -induced PBMC proliferation correlated with serum anti-BiP antibody titers. The immunological basis for this correlation may be important for understanding the pathogenesis of RA.
MycHSP70 and BiP share 50% of their amino acid sequences, and MycHSP70 287-306 is located at a position corresponding to that of the BiP-derived HLA-DR4 effector epitope. We previously identified two major BiP-derived HLA-DR4 epitopes in RA. One (BiP 336-355 ) was recognized by effector T cells and the other (BiP 456-475 ) by IL-10-secreting regulatory T cells 8 . MycHSP70 287-306 strongly induces proliferation and is regarded as an effector epitope. MycHSP70 287-306 and BiP 336-355 share 50% of the amino acids in their core sequences, which bind HLA-DR pockets. They also share the P4V residue, which is regarded as the most important amino acid for HLA-DR4 binding 20 . This sequence similarity could contribute to the breakdown of T cell tolerance to self BiP in RA, and also suggests the role of molecular mimicry; however, we were unable to exclude the contribution of epitope spreading. One possibility is that the same clone recognizes MycHSP70 287-306 and BiP 336-355 . A vinculin-specific TCR has been shown to recognize DERAA-containing epitopes from three bacterial species, Campylobacter coli, Lactobacillus curvatus, and L. sakei 4 . A second possibility is epitope spreading from the Myc antigen to a self-antigen that shares similarities via recognition by different T cell clones. Epitope spreading is described as the diversification of T cell activation from one epitope to non-cross-reactive autoantigen-derived epitopes and is induced during chronic inflammation. T cell epitope spreading has mainly been examined in mouse models of multiple sclerosis 21,22 . The findings obtained suggest that intermolecular and intramolecular epitope spreading occurs as a result of inflammation and tissue damage. According to this hypothesis, inflammation induced by MycHSP70-specific T cells could up-regulate the expression of BiP and BiP-specific T cell activation. These possibilities warrant further study.
Mycobacterial species, particularly non-tuberculous mycobacteria, live in the soil and water, and a large number of individuals are continuously exposed to several mycobacteria in daily life 23 . Most of these bacteria are considered to be non-pathogenic for humans. However, difficulties are associated with detecting and identifying these bacteria due to their slow growth rates. Next generation sequencing of 16S rRNA in microbiota has PBMCs (9 patients with SE + /+ and 14 patients with SE + /− ). (D) Blockade of the MycHSP 287-306 -induced proliferation of PBMCs from HLA-DR4-positive RA patients (n = 6) using an anti-HLA-DR antibody. (E) The secretion of MycHSP 287-306 -induced IFN-γ from the PBMCs of HLA-DR4-positive HDs (n = 6) and RA patients (n = 8). (F) Blockade of MycHSP 287-306 -induced cytokine secretion from the PBMCs of HLA-DR4-positive RA patients (n = 6) using an anti-HLA-DR antibody. *p < 0.05, **p < 0.01, ***p < 0.001. provided new evidence for the "silent" exposure of humans to Mycobacterial species 24 . The clear correlation between serum anti-MycHSP70 antibody titers and anti-citrullinated BiP antibody titers indicates that a relationship exists between immune responses to mycobacteria with ACPA formation. Mycobacteria in the lungs have been shown to play a role in the pathogenesis of RA because the lungs are a candidate initiator organ in which the triggering of specific immunity occurs in RA 24 . However, most RA patients have no history of Mycobacterial infection. Since immune responses to mycobacteria were activated in RA patients, our results imply the clinical importance of subclinical exposure to mycobacteria and its association with the pathogenesis of RA. If it is possible to determine the route of exposure, vaccination or tolerance induction to mycobacteria may become a new strategy for RA prevention and therapy. Since the oral administration of the MycHSP70 287-306 peptide improved experimental arthritis and suppressed autoreactive T cell reactions, the presence of mycobacteria in the gut, but not in the lungs, may regulate autoimmunity against self-HSPs.
As a limitation of this study, the patients who participated took steroids and/or immunosuppressive drugs, which may have affected their T and B cell functions. Even though these potentially immunosuppressive drugs were used, increased immune responses were observed in RA patients.  In conclusion, we herein demonstrated the presence of T and B cell responses to MycHSP70 in RA. We identified a new effector T cell epitope derived from MycHSP70, and PBMC proliferation induced by the MycHSP70-derived epitope correlated with that induced by the BiP-derived epitope in RA. The sequence similarity between MycHSP70 and human BiP at the T cell epitope level supports a link between Mycobacterial exposure and the breakdown of tolerance to BiP in RA patients. We also demonstrated the potential of the therapeutic application of these epitopes. The induction of tolerance to the MycHSP70-derived epitope could regulate  A,B) The MycHSP70 287-306 peptide or control (PBS alone) was orally administered to bovine type II collagen-immunized DBA/1 J mice after the occurrence of CIA for 5 continuous days (n = 12). Mice were sacrificed on day 35. Arthritis scores were measured by finger and paw swelling, and histological scores were measured by cell infiltration (C), pannus formation (P), and bone destruction (B). (C) Splenic CD4 + T cell proliferation from CIA-induced DBA/1 J mice in response to a recall antigen (Ag) stimulation (MycHSP70 287-306 and BiP 336-355 epitopes, and heat-denatured bovine type II collagen (BC II)). Proliferation was determined by 3 -H thymidine incorporation. (D) Serum anti-BiP and anti-BC II antibody titers were measured by ELISA. (E,F) Mouse IL-10 and TGF-beta1 concentrations in the splenic CD4 + T cell culture were measured by ELISA. Data were representative of at least three independent trials. *p < 0.05, **p < 0.01, ***p < 0.001.
Scientific RepoRts | 6:22486 | DOI: 10.1038/srep22486 immune responses to self BiP and arthritis, and, in the present study, we provided evidence to suggest that the induction of tolerance to microbial epitopes induces tolerance to related self-antigens. This system may be a suitable model for understanding how autoimmune responses develop in the pathogenesis of RA, and may lead to new therapeutic strategies for RA.

Materials and Methods
Patients and HLA-DRB1 typing. Blood samples were obtained from HDs (n = 20) and RA patients without cardiovascular diseases (n = 48). Demographic data are shown in Table S2. We included RA patients who fulfilled the classification criteria of the 2010 American College of Rheumatology (ACR)/European League Against Rheumatism 25 . Although some established RA patients had been diagnosed by the 1987 ACR criteria, all of them also met the 2010 criteria. No donor had a previous history of active tuberculosis infection. HLA-DRB1 genotyping was performed in accordance with a previous study 26 .Written informed consent was obtained from all subjects before samples were taken. This study was approved by the Ethical Committee of the University of Tokyo Hospital. The methods were performed in accordance with the approved guidelines.  Table S1) (Operon Biotechnology). The purity of each peptide was greater than 70%. Alanine-substituted peptides, which were derived from MycHSP 287-306 and had one amino acid substituted to alanine, were prepared for the binding assay.

Proteins. Human recombinant
Human PBMC culture. PBMCs (1 × 10 6 /mL) isolated with Ficoll-Paque (GE Healthcare) were cultured in the presence or absence of each peptide (10 μg/mL) with RPMI 1640 medium supplemented with 10% human serum, 2 mM L-glutamine, 100 U/ml penicillin, and 100 μg/ml streptomycin in a humidified incubator at 37 °C, 5% CO 2 , for 96 hours. Cells were incubated for 18 h following a 3 H-thymidine pulse into each well and harvested using a semiautomated sample harvester. Counts per minute were measured with a scintillation counter. In some experiments, the culture supernatants were collected after a 72-hour incubation and cytokine concentrations were measured in the culture supernatants by a human IFN-γ high sensitivity ELISA kit and human IL-17 high sensitivity ELISA kit (eBioscience). In blocking experiments, 10 μg/mL of an anti-HLA DR monoclonal antibody (L243) (Biolegend, San Diego, CA, USA) was added.
Peptide binding assay. The procedure to assess peptide binding to HLA-DRB1*0401 and 0405 was described in a previous study 8 . In brief, HLA-DRA1B1*0401 and 0405 was cloned and transfected into S2 Drosophila cells. Soluble DR4 molecules in the supernatant were collected and purified through an affinity column coupled with the anti-DR antibody LB3.1. In the peptide binding assay, a 40 nM solution of purified DR4 was incubated at 37 °C for 4 h with the HA 307-319 peptide (0.5 nM), which had been labeled at the N terminus with biotin. The indicated concentrations of the peptides were added as competitors to HA peptide binding. Bound peptides were separated from free peptides by immobilizing the DR molecules on microtiter plates coated with LB3.1. The antibody was coated to the plate by the overnight incubation of a 5 μg/mL solution at 4 °C. Bound biotinylated peptides were detected by an incubation with streptavidin-alkaline phosphatase. An automatic microplate reader (Bio-rad 550, Bio-rad) was used to measure optical density.
CIA. CIA in the background of DBA/1 J female mice (SLC Japan) was performed and arthritis scores were acquired as previously described 7 . In the therapeutic experiments, mice were administered 200 μg of MycHSP 287-306 peptides orally for 5 continuous days following the onset of arthritis. Mice were sacrificed on day 35, and the CD4 + T cell culture was performed as described above. Mouse cytokine concentrations in the supernatants were measured using a mouse IL-10 ELISA kit and mouse/human TGF-beta 1 ELISA kit (eBioscience) according to the manufacturer's protocol. Tissue samples were embedded in paraffin wax after 10% formaldehyde fixation and decalcification. Sections were stained with hematoxylin and eosin. Synovial tissues were graded by mononuclear cell infiltration, pannus formation, and cartilage erosion, as described previously 7 . All animal experiments were conducted in accordance with Institutional and National Guidelines, and the protocols were approved by the Ethical Committee on Animal Experiments of the University of Tokyo.