The goal of this study was to use multiplex autoantibody detection technology to determine associations between the presence of specific autoantibodies and classification criteria within a large, ethnically diverse cohort of SLE patients. We screened a panel of 10 autoantibody specificities that are often detected in SLE (dsDNA, chromatin, ribosomal P, 60
kD Ro, 52
kD Ro, La, Sm, Sm/RNP, RNP 68, and RNP A). Our study, as well as those of others, has described the BioPlex 2200 assay as a highly sensitive method for the detection of these autoantibody specificities [27
]. Understanding associations between specific autoantibodies and SLE criteria as detected by this new methodology which is in widespread clinical use provides key insights into prognostic relevance for clinical application and may improve screening tests for diagnostic purposes.
Our study participants demonstrate a diverse, representative SLE patient population. While HI patients had a statistically lower age at participation (37.9 ± 12.8) than the other groups (EA 43.2 ± 13.6, AA 40.7 ± 12.4, and other 39.9 ± 12), this is consistent with the earlier age of SLE onset in the overall HI population [33
]. The most prevalent criteria in our cohort were ANA, arthritis, immunologic, and hematologic criteria, similar to those observed in previous studies [34
Ethnic differences in autoantibody prevalence and association with ACR SLE classification criteria are observed in our study. We report a detailed association analysis between multiple autoantibodies and hematological disorder in our large Hispanic cohort, which has not been previously reported. We observed a significant difference between AA BioPlex 2200 ANA-positive and ANA-negative SLE patients. Hematological disorder (P
= 0.001), lymphopenia (P
= 0.030), and immunological disorder (P
= 0.0068) were significantly enriched in ANA-positive AA patients. However, when a similar analysis was performed using indirect immunofluorescence, these associations disappeared. This is most likely due to the difference in specificity and sensitivity between the two assays. It is important to note that our EA patient group did have the lowest autoantibody prevalence for all tested specificities. Thus, the use of the BioPlex 2200 ANA may require the use of traditional autoantibody assays to confirm absence of ANA in this population subgroup. In AA patients, multiple autoantibodies associate with hematologic involvement in SLE. The associations between SLE autoantibody specificities and ACR criterion observed in our study confirm those observed in previous work [9
Our initial univariate conditional analysis demonstrates association between multiple autoantibodies. However, our multivariate adjusted conditional logistic regression analysis shows that antibodies to 60
kD Ro and RNP 68 are significantly independently enriched in EA patients with hematological disorders. In AA patients, anti-SM/RNP and anti-La antibodies are correlated with hematological disorder, while anti-RNP A antibodies alone are highly associated with hematological disorder in HI patients. Our results show that antibodies to ribonucleoproteins are highly prevalent in patients with hematological disorder. These results differ than those of Agmon-Levin et al. [2
] and To and Petri [37
]. Here, Sm/RNP antibodies were underrepresented in SLE patients with hematologic criteria.
In our multivariate analysis, a significant association between anti-52
kD Ro, anti-ribosomal P, anti-RNP antibodies and the hematological ACR criterion is observed in EA and AA patients. Previous studies have identified correlations between anti-Ro and both lymphopenia and leukopenia or with lymphopenia alone and suggested a moderate association between anti-dsDNA and lymphopenia mostly in EA patient cohorts [12
]. A significant enrichment of anti-RNP 68 antibodies was observed in EA patients with lymphopenia. HI patients showed unique antibody associations with hematological ACR criterion. Interestingly, while antibodies to dsDNA were inversely associated with lymphopenia, anti-chromatin antibodies were directly associated with lymphopenia in HI patients. Our analysis has not only replicated the association between antibody to Ro and lymphopenia, but also revealed the inverse correlation between anti-dsDNA antibodies and lymphopenia in HI patients. However, differences in hematological criteria, especially luekopenia, may be due to benign ethnic neutropenia [38
], a well-described characteristic that is not a manifestation of lupus, in some cases.
The association of dsDNA antibody specificity with renal disease has been widely demonstrated [8
] and is also confirmed in our study. Interestingly, we have observed different distinct autoantibody associations with renal disease in EA and AA patients. The overall autoantibody associations with renal disease in EA are anti-chromatin and anti-Sm antibodies. Anti-dsDNA antibodies and female sex are more common in patients with renal disease; while anti-La antibodies are underrepresented in SLE patients with renal disease. These autoantibody association differences were maintained when examining renal disease subcriteria (proteinuria and cellular casts). In AA patients, female sex is associated with proteinuria; while anti-dsDNA and anti-Sm/RNP antibodies are correlated with cellular casts. The association between anti-Sm/RNP antibodies and renal criteria has not been previously described. In EA SLE patients, anti-Sm is mildly associated with proteinuria with odds ratio approaching 1.00, further studies are necessary to confirm this effect. No significant association between autoantibodies and renal involvement was observed in HI patients. Thus, our study suggests that anti-chromatin and absence of anti-La antibodies are the main predictors for renal involvement driven by prevalence of proteinuria in EA patients. However, the lack of autoantibody associations with cellular casts in EA SLE patients might be due to low numbers of EA patients being tested for cellular casts. Additionally, our study suggests that anti-dsDNA and anti-Sm/RNP antibodies are the strong correlates for renal disease as measured by cellular casts in AA patients.
Ethnic differences in mucocutaneous manifestations of SLE are observed in our patient cohort. In our study population, EA patients show an increased prevalence of malar rash and photosensitivity; while AA patients exhibit increased frequency for discoid rash which is consistent with previous studies [43
]. Our autoantibody results indicate that in EA, malar rash is positively correlated with female sex; anti-La antibodies are negatively associated with malar rash in EA patients, while anti-La and anti-52
kD Ro antibodies are enriched in AA patients with photosensitivity. Several previously reported studies have observed an association between anti-RNP antibodies and photosensitivity [45
], which is not seen in our study. However, it is important to note that these two previous studies utilized a primarily Asian cohort [46
]. It is possible that this association also exists in our population, but the relatively small number of Asian study participants prevents this observation.
The goal of this study was to examine associations between the prevalence of autoantibodies as detected by a Luminex bead-based assay (BioRad BioPlex 2200) and the occurrence of various clinical criteria in a large, ethnically diverse SLE patient cohort. The major findings of this paper identify associations between several antibody specificities and hematological disease, more specifically leukopenia and lymphopenia, and identify ethnic differences in autoantibody associations with renal subcriteria. Work is currently underway to better understand the mechanisms underlying these associations, particularly between leukopenia, lymphopenia, and autoantibody production. Autoantibodies play a crucial role in the diagnosis of SLE and a better understanding of the relationships between antibody prevalence and the presentation of other clinical criteria will further strengthen their prognostic implications.