AUTOIMMUNITY

Abstract Anti‐Ro52 autoantibodies (Ro52‐autoAbs) appear in the sera of connective tissue disease (CTD) patients with interstitial lung disease (ILD). Studies using patient sera have shown a correlation between the generation of Ro52‐autoAbs and the clinical morbidity and severity of CTD with ILD. In this study, we used a single B‐cell manipulating technology and obtained 12 different monoclonal Ro52‐autoAbs (mRo52‐autoAbs) from the selected four patients suffering from severe ILD with a high titer of Ro52‐autoAbs in their sera. Western blot analysis revealed that 11 of 12 mRo52‐autoAbs bound to the coiled‐coil domain of Ro52. Competitive ELISA demonstrated that mRo52‐autoAbs competed with each other to bind to Ro52. Epitope mapping showed that two of them specifically bound to a peptide (PEP08) in the coiled‐coil domain. We then examined the titer of Ro52‐autoAbs in the sera of 192 CTD patients and assessed the relationship between the serum levels of Ro52‐autoAbs that were reactive to PEP08 peptide and the clinical morbidity and severity of ILD. Statistical analysis revealed that the production of PEP08‐reactive Ro52‐autoAbs correlated with the morbidity and severity of ILD in CTD. Assessment of the production of PEP08‐reactive Ro52‐autoAbs in autoimmune diseases is useful for predicting the clinical morbidity of ILD. Introduction Ro52/TRIM21 is known as an E3 ligase involved in the process of ubiquitination and plays a prominent regulatory role in inflammation 1-7. Ro52 is the common target of autoantibodies in connective tissue diseases (CTDs) 8, 9, such as systemic lupus erythematosus (SLE) 10-12, Sjögren's syndrome (SS) 13-17, polymyositis/dermatomyositis (PM/DM) 18, 19, and systemic sclerosis (SSc) 20. Ro52‐specific autoantibodies (Ro52‐autoAbs) have been detected in the sera of many patients with various CTDs 9, 21. Some of these patients develop severe interstitial lung disease (ILD), and several reports have indicated that the production of Ro52‐autoAbs in patient sera is related to the incidence of ILD in CTD 22, 23. In SSc patients, the generation of Ro52‐autoAbs has been correlated to the morbidity and severity of ILD 23. In mixed connective tissue disease (MCTD) patients, Ro52‐autoAbs have been detected in 50% of patients with lung fibrosis 22. The Ro52 protein consists of multi‐domains, including an N‐terminal RING finger domain with E3 ubiquitin ligase activity, a B‐box domain, a coiled‐coil (C‐C) domain involved in oligomerization, and the C‐terminal PRYSPRY domain that commonly determines functions by acting as a target 24, 25. Some epitopes of Ro52‐autoAbs in SLE and SS patients have been reported 26, most of which reside in the C‐C domain of Ro52 27-34. The ring and PRYSPRY domains are also targets of Ro52‐autoAbs 29, 35. These epitopes were determined in patient sera using polyclonal antibodies. Concerning monoclonal Ro52‐autoAbs (mRo52‐autoAbs), Mietzner et al. generated monoclonal antibodies (mAbs) from IgG memory B cells of SLE patients and showed that some of them reacted to Ro52 36. However, there have been no reports of epitopes for human mRo52‐autoAbs produced in patients with ILD. In this study, we used our single cell analysis method denoted the immunospot array assay on a chip (ISAAC) 37, 38 to obtain mRo52‐autoAbs from the PBMCs of four DM patients suffering from severe ILD. We obtained twelve human mRo52‐autoAbs and showed that eleven of them bound to the C‐C domain of Ro52. We further determined an epitope for two of these mRo52‐autoAbs. Finally, we examined the relationship between the production of Ro52‐autoAbs that were reactive to the determined epitope in patient sera as well as the clinical morbidity and severity of ILD in CTD. Results Isolation of mRo52‐autoAbs from DM patients with severe ILD using the ISAAC system First, we examined 192 CTD patients to determine the titer of Ro52‐autoAbs in their sera by EUROLINE scan. We also examined them by scanning high‐resolution computed tomography (HRCT) and found that 57% (109/192) of these patients had ILD. The results agreed with previous studies 22, 23. We then tried to analyze the correlation of Ro52‐autoAb‐production and ILD. To achieve this goal, we tried to obtain mRo52‐autoAbs from DM patients because they are frequently complicated with ILD 39. We randomly selected 4 DM patients with severe ILD who also suffered from anti‐synthetase syndrome and had anti‐SS‐A, anti‐Jo‐1, or anti‐PL‐7 autoantibodies as well as Ro52‐autoAbs in their sera (Supporting Information Table 1). One hundred six antibody‐secreting cells (ASCs) that produced Ro52‐autoAbs were detected in the PBMCs of these four DM patients using ISAAC and retrieved from the chip (Fig. 1A). The antibody cDNAs were cloned from the single cells, and mAbs were produced. Fifteen recombinant mAbs specific to recombinant Ro52 (rRo52) protein were obtained. Sequence analysis revealed that three of 15 mRo52‐autoAbs had the same amino acid sequence as another clone, which indicated that these clones were derived from the same B‐cell clone (Table 1). We eventually obtained 12 different mRo52‐autoAbs from four DM patients with ILD (Fig. 1B). Heavy‐chain Light‐chain Pt. No. Name of mRo52‐autoAb Subclass V D J CDR3‐IMGT Family V J CDR3‐IMGT 100,112 110,112‐20 IgG1 IGHV3 IGHD6 IGHJ4 AKDWAQFVLRYFFDH κ IGKV1 IGKJ1 QQYDSYPWT 100,112 110,112‐19a IgG1 IGHV3 IGHD6 IGHJ4 AKDWAQFVLRYFFDH κ IGKV1 IGKJ1 QQYDSYPWT 100,112 110,112‐26 IgG1 IGHV7 IGHD6 IGHJ4 ARQGMFSTGWYLAH κ IGKV1 IGKJ1 QNYKSYSRT 100,112 110,112‐29 IgG1 IGHV7 IGHD4 IGHJ5 ARGALTMPTRVDP κ IGKV1 IGKJ5 QQTYSTLPIT 55 55‐37 IgG1 IGHV1 IGHD3 IGHJ4 ARSFYFGSGSYPY κ IGKV3 IGKJ2 QQYGSSPYT 55 55‐34a IgG1 IGHV1 IGHD3 IGHJ4 ARSFYFGSGSYPY κ IGKV3 IGKJ2 QQYGSSPYT 150 150‐08 IgG1 IGHV3 IGHD1 IGHJ4 ARGKSWLDY κ IGKV3 IGKJ4 QQYGSSLT 150 150‐14 IgG1 IGHV1 IGHD2 IGHJ5 ARGVYRRSGGNSWGFDP κ IGKV2 IGKJ4 MQALQTPPT 150 150‐42 IgG1 IGHV3 IGHD1 IGHJ4 ARDLGGGRRIYYFDY κ IGKV3 IGKJ4 QQYGSSPL 150 150‐44 IgG1 IGHV3 IGHD3 IGHJ4 AKVIGITFGGVIFPFDY κ IGKV3 IGKJ1 QQYGTAPT 150 150‐45 IgG1 IGHV3 IGHD3 IGHJ4 AKGFGTTFGGVIFPFDY κ IGKV3 IGKJ1 QQYSTSPT 150 150‐46 IgG1 IGHV3 IGHD5 IGHJ4 ARGKYWLDY κ IGKV3 IGKJ4 QQYGSSLT 150 150‐60 IgG1 IGHV3 IGHD4 IGHJ3 ARARFRTTVTTYAFDI κ IGKV2 IGKJ1 MQALQTPPWT 150 150‐61 IgG1 IGHV3 IGHD2 IGHJ4 ARERGGFCSGGSCYSCDY κ IGKV2 IGKJ1 MQALQTPRT 150 150‐58a IgG1 IGHV3 IGHD2 IGHJ4 ARERGGFCSGGSCYSCDY κ IGKV2 IGKJ1 MQALQTPRT a same as the above‐mentioned mRo52‐autoAbs were obtained PBMCs from 4 patients (#55, #100, #112, and #150). PBMCs of 2 patients (#100, #112) were mixed since the number of PBMCs of these patients were too small for ISAAC. 15 mAbs were cloned and found to bind rRo52 protein in which 3 mAbs had the same amino acid sequence. Final number of mRo52‐autoAbs became 12. V, D, J gene segments and CDR3‐IMGT in heavy chain as well as V, J gene segments and CDR3‐IMGT in light chain are shown. Epitope mapping To determine the epitopes of mRo52‐autoAbs, we first synthesized the various truncated fragments of rRo52 and then examined the reactivity of each mRo52‐autoAb to the fragments (Fig. 2A and B). Eleven of 12 mRo52‐autoAbs reacted with the C‐C domain as well as the full‐length rRo52 protein. One mRo52‐autoAb (Ab100, 112–20) reacted with the full‐length protein but not with any of the other fragments. We then examined whether these eleven mRo52‐autoAbs competed for binding to a specific site of the C‐C domain using competitive ELISA as previously reported 40. mRo52‐autoAbs together with one of the peroxidase (POD)‐conjugated mRo52‐autoAbs were added to wells that were coated with rRo52 proteins. The bound enzyme activity was then measured after washing the samples. As shown in Fig. 3A, all eleven mRo52‐autoAbs, but not the control mAb, inhibited the binding of POD‐conjugated Ab55‐37 mAb. Similarly, all 11 mRo52‐autoAbs inhibited the binding of other POD‐conjugated mRo52‐autoAbs (Ab150‐61, Ab150‐42, and Ab150‐45). These results suggest that the mRo52‐autoAbs that were obtained from DM patients with severe ILD (DM‐ILD patients) bind to the same or juxtaposed epitopes of Ro52. We then synthesized the peptide library of the C‐C domain of Ro52 (Fig. 3B and Supporting Information Table 2) to determine the precise epitope for mRo52‐autoAbs and examined the binding of mRo52‐autoAbs to these peptides. Two of 11 mRo52‐autoAbs (Ab55‐37 and Ab150‐61) were used for the binding assay because they showed clear binding activity to the C‐C domain (Fig. 2B). Those two mRo52‐autoAbs reacted with a peptide (PEP08, RIHAEFVQQKNFLVEE) but not with the other peptides of the C‐C domain or the control HIV‐env peptide (Fig. 3C). It is notable that these two mRo52‐autoAbs were obtained from PBMCs of different patients (Table 1). These data show that PEP08 contains an epitope of the mRo52‐autoAbs. These data, together with the competitive‐binding data (Fig. 3) indicate that all the mRo52‐autoAbs obtained by the ISAAC system bind to this epitope or a juxtaposed epitope. Relationship between the production of PEP08‐reactive Ro52‐autoAb and the morbidity of ILD in CTD As described above, the production of Ro52‐autoAb was reported to be clinically associated with ILD in SSc and MCTD 22, 23. To investigate the relationship between the production of the Ro52‐autoAbs that bind to PEP08 in the sera of CTD patients and the incidence and severity of ILD, 192 CTD patients were selected from our hospital whose Ro52‐autoAb titer was examined by EUROLINE scan (Fig. 4A). Overall, 43% (82/192) of them were positive for Ro52‐autoAb, while the rest (110/192) were negative. The reactivity of Ro52‐autoAb sero‐positive sera to PEP08 was analyzed. The results showed that the Ro52‐autoAb from 45% (37/82) of the patients bound to PEP08, whereas that from 55% (45/82) of the patients did not (Fig. 4A and B). All the profiles of PEP08‐Ab‐positive, PEP08‐Ab‐negative, and Ro52‐autoAb‐negative patients with CTD are shown in Supporting Information Table 3. We then compared the morbidity and severity of ILD in 38 patients in the PEP08‐Ab‐positive group and 154 patients in the PEP08‐Ab‐negative group using an odds ratio and Fisher's exact test. In these analyses, we assumed that 110 patients who were Ro52‐autoAb‐negative were negative for PEP08‐Ab (Fig. 4A). The assumption was confirmed by analyzing 14 Ro52‐autoAb‐negative patients in Supporting Information Table 3. They were also negative for PEP08‐Ab (Fig. 4C). For the morbidity of ILD, 30/38 (79%) patients in the PEP08‐Ab‐positive group had ILD. In contrast, 80/154 (52%) patients in the PEP08‐Ab‐negative group had ILD (Fig. 4A and D, odds ratio = 3.47, P = 0.0031, CI: 1.50–8.05). We also compared the severity of ILD in the PEP08‐Ab‐positive group and the PEP08‐Ab‐negative group using the KL‐6 value that had been considered to be an indication of severity of ILD 41, 42.Overall, 26/33 (79%) patients in the PEP08‐Ab‐positive group had a high KL‐6 value (KL‐6 ≥ 500). In contrast, 82/143 (57%) patients in the PEP08‐Ab‐negative group had a high KL‐6 value (KL‐6 ≥ 500) (Fig. 4A and E, odds ratio = 2.65, P = 0.0285, CI: 1.13–6.78). There was no correlation between the titer of PEP08‐Ab and the KL‐6 value (Supporting Information Fig. 1). The results indicate that PEP08‐Ab‐positive CTD patients had a higher indication of morbidity and severity of ILD than PEP08‐Ab‐negative CTD patients. Discussion Ro52 is the common target of autoAbs in CTD, and the production of Ro52‐autoAbs in patient sera has been reported to be related to the incidence of ILD with CTD 22, 23; however, there have been no reports concerning the epitopes of Ro52 protein related to ILD. To gain an understanding of the relationship of the Ro52‐autoAb epitopes and ILD, we obtained 12 human mRo52‐autoAbs from peripheral blood B cells of DM‐ILD patients (Table 1 and Fig. 1) and determined their antigenic epitopes. We found that 11 of them reacted with the C‐C domain of the Ro52 protein (Fig. 2). Previous studies have also demonstrated that Ro52‐autoAbs reacts with the C‐C domain using serum of SS or SLE patients 27-34. The C‐C domain plays an important role in protein–protein interactions and is usually protruding from the main body of proteins. Thus, it tends to be recognized by B cells using their antigen receptors on the cell surface. We could not determine the epitopes of one mRo52‐autoAb, Ab100/112‐20, which reacted only with the full‐length protein, indicating that it may recognize the discontinuous conformational epitope formed by the two domains. In contrast, the remaining mRo52‐autoAbs may bind to linear epitopes since they recognized the antigen in immunoblotting with SDS‐denatured and linearized proteins. However, we could not exclude the possibility that they recognized refolded proteins during immunoblotting. Regarding the epitopes of Ro52‐autoAbs in various CTDs, their epitopes have been reported to be distributed in various rRo52 fragments (Supporting Information Table 4). In this study, by competitive ELISA (Fig. 3A) and epitope mapping using a peptide library (Fig. 3B and C), we found that two mRo52‐autoAbs from DM‐ILD patients bound to a Ro52 epitope (PEP08), and the remaining nine Abs bound to its juxtaposed epitopes. Interestingly, those two mRo52‐autoAbs were obtained from different patients. This outcome posed a strong possibility that PEP08 is one of the important epitopes for ILD in CTD, although we must consider that mRo52‐autoAbs obtained from DM patients may have biased the results. Wahlen et al. reported an epitope of Ro52‐autoAbs that was found to show production correlated with the onset of congenital heart block (CHB) named “P200” (aa200–239) 43. The maternal mRo52‐autoAbs of CHB patients did not react to the peptide named “P176” (aa176‐196). In contrast, our mRo52‐autoAbs from DM‐ILD patients reacted with the PEP08 (aa177‐192) involved in P176. Although all these peptides were included in the C‐C domain (aa133‐255), each epitope may play a different role in oligomerization and might be involved in the onset of different CTD. Concerning the functional role of Ro52‐autoAbs in the diseases, we investigated how the generation of PEP08‐specific Ro52‐autoAbs in the CTD patients affected the incidence and severity of ILD (Fig. 4). We showed that the PEP08‐Ab‐positive group had a higher morbidity of ILD (odds ratio = 3.47, p = 0.0031, CI: 1.50–8.05) (Fig. 4D). The group also exhibited a significant relationship to the severity of ILD (odds ratio = 2.76, p = 0.0285, CI: 1.13–6.78) (Fig. 4E). The results showed that the production of PEP08‐reactive Ro52‐autoAbs influenced the morbidity of ILD in CTD. Although the molecular mechanisms in which Ro52‐autoAbs induce a high risk of ILD are not known, further analysis of the in vivo effect of mRo52‐autoAbs using an animal model may facilitate further understanding of involvement of PEP08‐reactive Ro52‐autoAbs to the onset of ILD in CTD. This study has the following limitations: (1) The titer of KL‐6 was not assessed for some patients who did not develop ILD, as shown in Supporting Information Table 3). The epitopes of Ro52‐autoAbs produced in various CTDs were distributed on various Ro52 fragments (Supporting Information Table 4). Here, we determined the PEP08 epitope using mRo52‐autoAbs from DM‐ILD patients. Thus, the selection of DM patients may have biased the results. (3) In this study, we focused on anti‐Ro52 autoantibodies of CTD patients and investigated the clinical correlation of a mRo52‐autoAbs epitope to ILD. Although PEP08‐Ab‐positive CTD patients had a higher indication of morbidity and severity of ILD than PEP08‐Ab‐negative CTD patients, it remains unknown why many PEP08‐negative patients still had ILD and a high KL‐6 value. For the onset of ILD, multiple risk factors 41 should be considered to be involved. One of the risk factors is autoantibodies; however, it has been reported that ILD‐related autoantibodies are not only anti‐Ro52 autoantibodies but also anti‐PL‐7 and anti‐Jo‐1 autoantibodies 39, 44, among others. Thus, analysis of the involvement of anti‐PL‐7 and anti‐Jo‐1 autoantibodies in ILD may contribute to understanding the morbidity and severity of many PEP08‐negative patients. It was recently reported that the production of Ro52‐autoAbs in patient sera were related to undifferentiated connective tissue diseases (UCTD) 45, 46. Measurement of PEP08‐Ab in sera of UCTD patients may contribute to the prediction of ILD. In summary, we obtained twelve mRo52‐autoAbs from four DM‐ILD patients. Eleven of these mRo52‐autoAbs bound to the C‐C domain in immunoblotting, and they competed with each other to bind to the C‐C domain. We determined the binding site of mRo52‐autoAb using the peptide library and analyzed the relationship of the production of PEP08‐Ab and the clinical morbidity and severity of ILD. We showed that the PEP08‐Ab‐positive patients exhibited a higher incidence risk of ILD than PEP08‐Ab‐negative patients. Future studies using the mRo52‐autoAbs in vivo system may be required to understand the pathological function of Ro52‐Abs in ILD with CTD. Materials and methods Patients and their characterization Patients who visited the First Department of Internal Medicine at Toyama University Hospital and were diagnosed with CTD were included in the study (Fig. 4). Informed consent was obtained from all patients in accordance with a protocol approved by the Ethical Review Board of the University of Toyama. The characteristics of the patients at the time of enrollment are summarized in Supporting Information Table 3. The patients included 145 females (76%) and 47 males (24%). Their median age was 63 years. Sixty patients (31%) had PM/DM, 44 (23%) had SSc, 40 (21%) had rheumatoid arthritis, 17 (9%) had SS, 13 (7%) had MCTD, 10 (5%) had SLE, 6 (3%) had amyopathic dermatomyositis, and 2 (1%) had vasculitis syndrome. The Ro52‐autoAb‐titer in their sera was examined by a EUROLINE scan (Cosmic Corporation, Japan). In the EULOLINE scan, recombinant Ro‐52 (52 kDa) produced with the baculovirus system in insect cells is used to detect Ro52‐autoAbs. We defined the Ro52‐autoAb titer as “high” when the EUROLINE scan determined the titer as 2± or 3±. Diagnosis of ILD was performed by examining the interstitial markings in the lung in HRCT. The severity of ILD was examined by measuring the KL‐6 value (SRL, Tokyo, Japan). In this study, KL‐6 was used as a potential biomarker of ILD in CTD 41, 47, 48. KL‐6 was measured using electrochemiluminescence immunoassay as described previously 49. The sensitivity, specificity, and diagnostic accuracy were 93.9%, 96.3%, and 95.7% for KL‐6; 81.8%, 86.6%, and 85.2% for SP‐A; 69.7%, and 95.1%, and 87.8% for SP‐D; and 51.5%, 92.7%, and 80.9% for MCP‐1; respectively 41, 47, 48, indicating that KL‐6 is a superior serum biomarker for ILD compared with SP‐A or SP‐D. It has also been reported that KL‐6 is negatively well correlated with DLco and FVC 50 and positively correlated with the HRCT score 51 in cases of ILD with CTD. Severe ILD was defined when the KL‐6 value was greater than 500 U/mL. Detection of ASCs using the immunospot array assay on a chip (ISAAC) PBMCs were prepared from 10 mL of heparinized peripheral blood from patients with a high serum titer of Ro52‐autoAbs after centrifugation at 400 × g using a Ficoll‐Hypaque gradient. CD138+ cells were enriched with human CD138 MicroBeads (Miltenyi Biotec) using an autoMACS Pro‐separator (Miltenyi Biotec) according to the manufacturer's instructions. The ISAAC method is covered by patents that have been exclusively licensed to Valneva (Nantes, France). Details and instructions regarding the microwell array chip and the ISAAC method have been previously described 37, 38. Briefly, we coated the surface of the chip with 10 μg/mL recombinant Ro52 (rRo52) protein (ProSpec) in PBS by incubating the chip overnight at 4°C. After removing the antibody solution, we blocked the chip with 0.01% Biolipidure (NOF Corporation, Japan) for 30 min at room temperature (r.t.) and subsequently washed the chip with the culture medium (RPMI 1640 containing 10% fetal calf serum, 50 μM 2‐mercaptoethanol, 100 μg/mL streptomycin, and 100 U/mL penicillin). We then arrayed the CD138+ cells in culture medium on the chip and removed the residual cells outside the wells with gentle washing. We cultured the cells on the chip for 3 h at 37°C. After gentle washing, we added 1 μg/mL Cy3‐conjugated anti‐human IgG‐Fc antibody (Millipore) for 30 min. Finally, we stained the cells with 1 μM Oregon Green (Molecular Probes) for 5 min at r.t. After gentle washing, antigen‐specific antibody‐secreting cells were retrieved under a fluorescence microscope (BX51WI, Olympus) using a capillary with a micromanipulator. Amplification of antibody cDNA, production of antibody, and purification of antibody We amplified antibody cDNAs encoding the VH and VL fragments using a single‐cell 5’‐RACE method with primers for the γ, κ, and λ chains. We then inserted these cDNAs into expression vectors that contained cDNA encoding the whole immunoglobulin constant region for IgG1 or the light chains, as previously described 38, 52. Then, we co‐transfected Expi293FTM cells with both the heavy and light chain expression vectors using the Expi293FTM Expression System (Life Technologies). We then collected the culture supernatant and purified the antibody using a protein G column as previously described 38. The immunoglobulin gene repertoire was analyzed with the IMGT/V‐Quest tool (http://www.imgt.org/). Identification of mRo52‐autoAbs We coated the wells of 96‐well plates (Nunc) with 50 μL of 0.5 μg/mL rRo52 protein in PBS overnight at 4°C. The next day, we blocked the wells with 3% BSA in PBS containing 0.1% Tween 20 (PBST). After washing the wells, we added to the wells monoclonal antibodies (1 μg/mL) and patient sera in PBST and incubated the plate at r.t. for 1 h. After washing the wells, we added alkaline phosphatase‐conjugated anti‐human IgG Fc‐antibody (Sigma) and incubated the plate at r.t for 1 h. We washed the wells and detected the amount of bound antibodies using p−nitrophenylphosphate (Sigma). Subsequently, we measured the optical absorbance at 405 nm with an ELISA reader according to the manufacturer's instructions. Western blotting of mutant recombinant Ro52 (rRo52) proteins with mRo52‐autoAbs We conjugated peroxidase to mRo52‐autoAbs or an anti‐influenza nucleoprotein antibody 38, 53 (control antibody) using the Peroxidase Labeling Kit‐NH2 (Dojindo, Japan) according to the manufacturer's instructions. Full‐length or truncated rRo52 proteins were produced in vitro with the cell‐free protein synthesis system using purified wheat embryos as previously described 54. Briefly, we amplified cDNAs encoding full‐length or truncated rRo52 proteins by PCR and inserted them into a pEU‐E01 expression vector (CellFree Sciences, Ehime, Japan). All proteins were fused to a D1CE epitope tag at the C‐terminus 55. We then produced recombinant proteins using the WEPRO7240 Expression Kit (CellFree Sciences) according to the manufacturer's instructions. Next, we resolved the rRo52 proteins using SDS‐PAGE with 5–20% gradient acrylamide gel (BIO CRAFT, Tokyo, Japan) under reducing conditions and transferred them to an Immobilon‐P nylon membrane (Millipore). We treated the membrane with Blocking One (Nacalai Tesque, Kyoto, Japan) and probed it with peroxidase‐conjugated antibodies. Visualization was performed using the ECL detection system (GE Healthcare). To confirm the loading of recombinant proteins on the blotting membrane, we washed the membrane with PBST for a few minutes and inactivated the peroxidase of the peroxidase‐conjugated mRo52‐autoAb 0.01% sodium azide for 30 min followed by a wash with PBST for a few minutes. We then treated the membrane with Blocking One and probed it with peroxidase‐conjugated anti‐D1CE tag antibodies. Competitive ELISA assay with mRo52‐autoAbs We coated the wells of 96‐well Maxisorp plates (Nunc) with 50 μL of 5 μg/mL rRo52 in PBS and incubated them at 4°C overnight. The next day, we blocked the wells with 3% BSA in PBS at r.t. for 1 h. We prepared the solution containing 10 clones of mRo52‐autoAbs or control mAb at 10 μg/mL and one of the clones of mRo52‐autoAbs conjugated with peroxidase at 0.1 μg/mL. After washing the wells, we added 50 μL of a clone of unconjugated mRo52‐autoAb to each well with 50 μL of a clone of peroxidase‐conjugated mRo52‐autoAb and incubated them at r.t. for 1 h. After washing the samples, we added 50 μL of TMB solution (KPL) and incubated the samples for 30 min at r.t.; then, we added 50 μL of stop solution (KPL). We measured the optical absorbance at 450 nm with an ELISA reader (BMG LABTECH) according to the manufacturer's instructions. ELISA of mAbs with a peptide library We first prepared the peptide library for the C‐C domain of Ro52 (Eurofins Genomics). The amino acid sequences are shown in Supporting Information Table 2. This library had 21 peptides. Each peptide consisted of 16 amino acids and overlapped by eight amino acids with an adjacent peptide. We then coated the wells of 96‐well Maxisorp plates (Nunc) with 50 μL of 5 μg/mL peptides in PBS and incubated them at 4°C overnight. The next day, we blocked the wells with 3% BSA in PBS. After washing the samples, we added the peroxidase‐conjugated mRo52‐autoAb or control mAb to the wells and incubated them at r.t. for 1 h. After washing the cells, we detected the antibodies that were bound to antigens using a TMP detection system (KPL). Subsequently, we measured the optical absorbance at 450 nm with an ELISA reader (BMG LABTECH) according to the manufacturer's instructions. ELISA of patient serum samples with the peptides We coated the wells of 96‐well Maxisorp plates (Nunc) with 50 μL of 10 μg/mL peptide (PEP08: RIHAEFVQQKNFLVEE) or HIV env gp160 peptide (RYLRDQQLL) (MBL) in PBS and incubated the cells overnight at 4°C. The next day, we blocked the wells with 3% BSA in PBS. After washing the wells, we added the patient sera to the wells and incubated the plate at r.t. for 1 h. After washing the wells, we added alkaline phosphatase‐conjugated anti‐human IgG Fc‐antibody (Sigma) and incubated the plate at r.t for 1 h. We washed the wells and detected the amount of bound antibodies using p−nitrophenylphosphate (Sigma). Subsequently, we measured the optical absorbance at 405 nm with an ELISA reader according to the manufacturer's instructions. We performed an ELISA three times per patient. When the mean of the OD(PEP08)/OD(HIV) was higher than 1 + SD, we classified the patients in the “PEP08‐positive group.” When the mean of OD(PEP08)/OD(HIV) was less than 1 + SD, the patients were classified in the “PEP08‐negative group.” Acknowledgments We thank Sanae Hirota for technical assistance and Kaoru Hata for secretarial work. This research was supported by research grants from Chugai Pharmaceutical Co., LTD. (H.T.) and from the Platform Project for Supporting in Drug Discovery and Life Science Research (Platform for Drug Discovery, Informatics, and Structural Life Science) from the Ministry of Education, Culture, Sports, Science (MEXT), Japan Agency for Medical Research and development (AMED) (A.M.); Platform Project for Supporting Drug Discovery and Life Science Research (Basis for Supporting Innovative Drug Discovery and Life Science Research (BINDS)) from AMED under Grant Number JP18am0101077 (T.O.), and JSPS KAKENHI Grant Number 16H06499 (H.K.). All authors designed the experiments. M.O., T.O., Y.N., and H.K. performed the experiments. M.O., R.T., K.S., H.T., and K.T. acquired clinical data. M.O., T.O., H.K., and A.M. analyzed the experiments and wrote the manuscript. Conflicts of interest Okumura, Ozawa, Kishi, Taki, Tobe, and Muraguchi have a patent application pending for the Ro52 autoantibody (JP2016‐220793). The rest of the authors declare no commercial or financial conflict of interest. Supporting Information References Abbreviations ASC antibody‐secreting cell CTD connective tissue disease HRCT high‐resolution computed tomography ILD interstitial lung disease ISAAC immunospot array assay on a chip MCTD mixed connective tissue disease POD peroxidase SLE systemic lupus erythematosus UCTD undifferentiated connective tissue disease

Your credible Sjögren's syndrome resource.