Serological and clinical associations of autoantibodies in Chinese patients with new-onset systemic lupus erythematosus

To study the clinical significance of autoantibodies in Chinese patients with new-onset systemic lupus erythematosus (SLE), we enrolled 526 new-onset patients who met the 1997 Updated American College of Rheumatology SLE Classification Criteria for a retrospective cohort study. Chi-square test and Wilcoxon rank-sum test were used to detect the relationship of autoantibodies with clinical manifestations and serological results respectively. Our results demonstrated that the positive rate of anti-ribosomal P protein (anti-P) antibody in female patients was higher than that in male patients (41.2% vs. 22%, P = 0.008). Patients with anti-SSB (43.95 ± 73.12 vs. 40.92 ± 75.75, P = 0.004; 63.93 ± 103.56 vs. 55.06 ± 120.84, P = 0.008 respectively) antibodies had higher levels of alanine aminotransferase (ALT) and aspartate transaminase (AST), whereas those with anti-P antibody (28.90 ± 25.70 vs. 50.08 ± 93.00, P = 0.014; 38.51 ± 48.19 vs. 69.95 ± 142.67, P = 0.047, respectively) had lower levels of them. Anti-dsDNA antibody (P = 0.021) was associated with pulmonary arterial hypertension (PAH). The patients with anti-Ro60 (P = 0.044), anti-P (P = 0.012) and anti-dsDNA (P = 0.013) antibodies were less likely to develop Interstitial lung disease. Anti-SmRNP antibody was correlated to lower prevalence of neuropsychiatric symptoms (P = 0.037), and patients with anti-centromere antibody (ACA) were more likely to develop serositis (P = 0.016).We identified five clusters of SLE-related autoantibodies, confirmed previously reported associations of autoantibodies, and discovered new associations.


The association of autoantibodies with clinical manifestations. The correlation between clinical
manifestations and the different autoantibodies are summarized in Table 5. Fever was significantly associated with the presence of anti-dsDNA (P = 0.009), anti-P (P = 0.031) and anti-SmRNP (P = 0.041) antibodies. Facial rash was more frequent in patients positive for anti-dsDNA (P = 0.013), anti-P (P < 0.001) and anti-SmRNP (P = 0.031) antibodies, whereas patients with anti-R52 antibody (P = 0.039) were less likely to develop facial rash. Alopecia was associated with anti-dsDNA (P = 0.020), anti-Sm (P = 0.020) and anti-SmRNP (P = 0.034) antibodies, Reynold's phenomenon with anti-Sm (P = 0.007) and anti-SmRNP (P < 0.001) antibodies, and serositis with www.nature.com/scientificreports/ anti-dsDNA (P = 0.014), anti-Ro52 (P = 0.047) and anti-centromere (P = 0.016) antibodies. While the presence of anti-P (P = 0.003) and anti-SmRNP (P = 0.001) antibodies correlated with a higher risk of appendicular rash, patients with anti-SSB antibody (P = 0.038) were less likely to develop the same. PAH was associated significantly with the presence of anti-dsDNA antibody (P = 0.021), whereas the anti-P antibody (P = 0.009; P = 0.012, respectively) was correlated with a lower risk of PAH and ILD. In addition, the patients with anti-dsDNA (P = 0.013) and anti-R60 (P = 0.044) antibodies were less likely to develop ILD. Renal disorder was associated with the presence of anti-dsDNA (P < 0.001) and anti-SSB (P = 0.006) antibodies. Anti-SmRNP antibody was also correlated with oral ulcerations (P = 0.026), whereas patients with anti-SmRNP antibody (P = 0.037) were less likely to develop NPSLE. There was no relationship between the different autoantibodies and arthritis.
The relationship between autoantibodies. The relationship between autoantibodies were detected by Cluster analysis. As shown in Fig. 1, the autoantibodies were classified into 5 clusters. Cluster 1 included the anti-Sm and anti-SmRNP antibodies, and Cluster 2 was composed of anti-R60 and anti-R52 antibodies. Both clusters were formed early in the disease course. Cluster 3 comprised of anti-dsDNA and anti-chromatin antibodies, Cluster 4 included the anti-SSB and anti-centromere antibodies, and Cluster 5 included the anti-P antibody.

Discussion
SLE is characterized by the presence of specific autoantibodies 1 . The therapy which focus on autoantibodies has been cared much. For example, blocking anti-dsDNA antibodies in a mouse model of SLE alleviated organ injury 5 . In this study, we retrospectively analyzed the correlation of several autoantibodies with sex, age at onset, disease activity, laboratory data and clinical manifestations in Chinese patients with new-onset SLE, in order to identify novel therapy. In this study, fever and arthritis are common initial symptoms, which was also observed in other study 9,10 . Arthritis were more prevalent in some studies 11,12 , which may be attributed to different definition of it. The prevalence of facial rash from our study was similar to the study by Sebastiani et al. 11 , but lower than that in some previous studies 12,13 . Meanwhile, the ocurring rate of appendicular rash which is part of skin manifestation, alopecia and Reynold's phenomenon lower than that in other studies 10,12,13 . Skin manifestation is influenced by sun-exposing, and discoid rash can induce alopecia 9,11 . Cold stimulation is a cause of Reynold's phenomenon. The difference may be due to geological location, climate, as well as genetic factors. NPSLE with more nonspecific symptoms like headache and mood disorder is difficult to diagnose. Therefore, the prevalence of NPSLE was different in previous cohorts, but still a rare initial presentation 11,13 . The incidence of serositis in our cohort was consistent with the finding of Leuchten et al. 13 , but lower than that in study by Sebastiani et al. 11 . The first incidence rate of serositis which diagnosis is dependent on the imaging examination may be underestimated. The frequency of renal disorder was essentially comparable with that from other researches 11,13 . We did not find the first incidence rate of PAH and ILD in previous studies, which are rare symptoms, but the outcomes were coincide with the incidence of them in the course of SLE 14 .
The anti-dsDNA antibody is a reliable diagnostic biomarker for SLE 16 , and its presence is related to tissue damage in the kidneys, skin and brain 5 . Our results just conformed to the first two. The difference may be due to different course of disease and sample size. Another possible explanation is the difficulty of diagnosis for NPSLE and the rarity of it as initial symptom. We found that the anti-dsDNA antibody was associated with disease activity, leukopenia, anemia, serositis, thrombocytopenia, ESR, complement C 4 and Alb, which is consistent with previous findings [22][23][24] . Higher ESR is the likely result of kidney damage, which in turn lowers Alb levels 25 . Furthermore, presence of the anti-dsDNA antibody was related to more serious kidney injury, lower levels of complement C 3 , fever, alopecia, PAH and less prevalence of ILD. According to the research by Li et al. 26 , the frequency of lupus nephritis, serositis and hypocomplementemia in patients with SLE and PAH were significant higher. Meanwhile, the disease activity of them were higher, and our results of association of anti-dsDNA were coincide with above symptoms. PAH and ILD were rare initial sign. We will collect more data of new-onset SLE paitents with them, and conduct follow-up studies in risk factors of these complications in the future. www.nature.com/scientificreports/ The anti-Sm antibody is highly specific for SLE, and we detected significantly higher positive rate of this autoantibody in patients with early-onset SLE compared to those with late-onset disease. Previous studies have shown that SLE patients with anti-Sm antibody tend to be younger compared to those lacking the antibody 17,27 . In addition, the anti-Sm antibody was associated with Reynold's phenomenon and elevated liver enzymes in our study, as demonstrated by other groups as well 24,28 . Alopecia and higher levels of IgG also correlated with the presence of anti-Sm antibody in our cohort. There is evidence that anti-Sm antibody is associated with disease activity 17,29 , renal disorder 28,30,31 and lower levels of complement proteins 31,32 . The differences in the testing methods for anti-Sm antibody may explain the variations in results 33 .
The anti-P antibody has been previously associated with disease activity 18,34 , lower levels of complement 34 , fever 18,21 and malar rash 18,35,36 in SLE patients, which was also observed in our study. However, our findings contradict the previously reported correlation between anti-P antibody and renal disorder 21,36,37 . In fact, we found that the extent of kidney damage of patients with anti-P antibody was slighter than that without it. Furthermore, the liver function of patients with anti-P antibody was better compared to those lacking the antibody, which is also inconsistent with previous studies 21, 38 . In a previous study 39 , the levels of anti-P antibody increased during the active phase of nephritis and resumed to normal in remission stage. In addition, the patient with anti-P antibody was not diagnosed with liver damage at early stage of SLE. Therefore, these discrepancies can be attributed to differences in course of disease, meanwhile, it also may be associated with ethnicity and study design. More follow-up is warranted to observe the long-term complications. Moreover, the presence of anti-P antibody was correlated to ESR, appendicular rash, and lower prevalence of PAH and ILD in our study. However, we did not observe any correlation between anti-P antibody and NPSLE which is difficult to diagnose. One noteworthy finding was that the prevalence of anti-P antibody was higher in female patients than in male patients.
The prevalence of anti-SSA and anti-SSB antibodies was higher in SLE patients with secondary Sjogren's syndrome. Earlier studies have demonstrated that anti-SSA (Ro) antibody is related to hemocytopenia 40 and ILD 41 . Our findings regarding the former were similar, whereas that regarding ILD were contradictory, which can be attributed to differences in sample size and ethnic groups. Moreover, ILD mostly occurred in long-course patients 14 , different course of disease may explain the discrepancy. We also found that anti-SSA antibody was correlated to ESR and higher levels of IgG. In addition, we observed a correlation between anti-Ro60 antibody and lower Alb concentrations. The presence of the anti-Ro52 antibody was related to lower prevalence of malar rash, which contradicts the findings of Harley et al., 40 which can be explained by different definition. Some studies have reported an association between anti-SSA antibody and PAH 26,42 , and the risk factors of PAH are pericarditis and pleurisy 26 , which were not observed in our study. However, we detected a correlation between the anti-Ro52 antibody and serositis, which has been reported previously 24 . There may be a potential link between anti-SSA and PAH, and further studies are needed to verify this hypothesis. Previous studies have shown that anti-SSB antibody is associated with higher levels of IgG 7 , lower levels of Complement C 3 23 and hematological symptoms 7,42 , which was confirmed in our study as well. In addition, we found that anti-SSB antibody correlated to ESR, lower levels of IgA and Complement C 4 , lower Alb concentrations, lower prevalence of appendicular rash, higher prevalence of renal disorder and more serious hepatic damage. We observed that the patients with anti-SSB antibody got lower levels of Hb, WBC and Plt which is initial hematological symptoms. It may be caused by disease activity, renal disorder or liver damage, and this new result need further follow-up observation.
The anti-centromere antibody (ACA) has been detected in subjects with CREST syndrome. The positivity rate of ACA in SLE patients in our study was 3.8%. In addition, ACA was correlated to lower levels of Complement C 4 , lower WBC count and serositis. However, we did not detect any association between ACA and Reynold's phenomenon, most likely due to the few samples positive for ACA. The prevalence of anti-chromatin antibody (also called anti-nucleosome antibody) is high in SLE patients 20 , and is associated with disease activity and renal disorder 20,43,44 . We also detected an association between this antibody and disease activity but without renal disorder, which may be due to genetic, course of disease and ethnic influences. Furthermore, the anti-chromatin antibody was correlated to hypo-complementemia in our study. The anti-SmRNP antibody is derived from anti-RNP antibody 45 , and is associated with higher levels of IgG and Hb, higher Plt count, lower levels of AST, fever, skin manifestations, oral ulcerations, Reynold's phenomenon, alopecia and lower prevalence of NPSLE. There is some overlap between the clinical manifestations of anti-SmRNP and anti-RNP antibodies, such as Reynold's phenomenon 46 . Using cluster analysis, the autoantibodies were classified into 5 clusters. Only cluster 1 and cluster 5 fit well with previous studies 42,47 , which could be explained by the different autoantibodies we detected. In addition, anti-dsDNA antibody was highly relevant with anti-chromatin antibody in previous study, which was observed in our study (cluster 3) 44 .
Our study has certain salient features, such as a large cohort and all patients with new-onset SLE. Thus, our findings are more relevant in terms of identifying targets for delaying the progression of SLE, since because the prevalence of autoantibodies can change during disease course 48 . Nevertheless, we could not examine the changes in the spectrum and levels autoantibodies during the course of SLE due to the cross-sectional design of our study. Second, our cohort consisted of only Chinese patients, and the results may not be applicable to other ethnic populations. Third, samples with rare autoantibodies and clinical manifestations were few, and the findings will have to be validated with further studies.
In conclusion, detection of specific autoantibodies in SLE patients can predict organ injury and other complications, and aid in timely intervention. We recommend that patients with anti-dsDNA antibodies should undergo echocardiography to detect PAH in a timely manner, and liver function tests should be conducted for those with anti-SSB antibody. The blood routine examination need to be tested regularly for patients with anti-SSA, anti-SSB and anti-centromere antibodies. Patients with anti-dsDNA and anti-SSB antibody should pay attention to tests of renal function. Imaging examination of such as echocardiography and CT need to be performed to detect serositis for patients with anti-dsDNA, anti-Ro52 and anti-centromere antibodies.