Abstract Objective Diagnosis of SLE relies on the detection of autoantibodies. We aimed to assess the diagnostic potential of histone H4 and H2A variant antibodies in SLE. Methods IgG-autoantibodies to histones H4 (HIST1H4A), H2A type 2-A (HIST2H2AA3) and H2A type 2-C (HIST2H2AC) were measured along with a standard antibody (SA) set including SSA, SSB, Sm, U1-RNP and RPLP2 in a multiplex magnetic microsphere-based assay in 153 SLE patients [85% female, 41 (13.5) years] and 81 healthy controls [77% female, 43.3 (12.4) years]. Receiver operating characteristic analysis was performed to assess diagnostic performance of individual markers. Logistic regression analysis was performed on a random split of samples to determine the additional value of histone antibodies in comparison with SA by likelihood ratio test and determination of diagnostic accuracy in the remaining validation samples. Results Microsphere-based assay showed good interclass correlation (mean 0.85, range 0.73–0.99) and diagnostic performance in receiver operating characteristic analysis (area under the curve (AUC) range 84.8–93.2) compared with routine assay for SA parameters. HIST1H4A-IgG was the marker with the best individual diagnostic performance for SLE vs healthy (AUC 0.97, sensitivity 95% at 90% specificity). HIST1H4A-IgG was an independent significant predictor for the diagnosis of SLE in multivariate modelling (P < 0.0001), and significantly improved prediction of SLE over SA parameters alone (residual deviance 45.9 vs 97.1, P = 4.3 × 10−11). Diagnostic accuracy in the training and validation samples was 89 and 86% for SA, and 95 and 89% with the addition of HIST1H4A-IgG. Conclusion HIST1H4A-IgG antibodies improve diagnostic accuracy for SLE vs healthy. systemic lupus erythematosus, antinuclear antibodies, histone, histone H4, multiplex assay Rheumatology key message Histone H4-IgG antibodies accurately discern systemic lupus patients and healthy individuals. Introduction The clinical heterogeneity of SLE often constitutes an obstacle to ready diagnosis, especially in a primary care setting . The correct interpretation of confirmatory laboratory investigations requires knowledge about a number of pitfalls . For instance, ANAs are detectable in virtually all SLE patients, but they lack specificity and may even be detected in healthy individuals at comparatively low titres . Determination of ANA binding specificities such as SSA, SSB, U1-RNP and Sm antibodies included in most standard assays increases specificity for the diagnosis of SLE. However, the prevalence of these distinct antibodies in SLE is usually low . Thus, markers with superior performance or a marker combination may lead to a more confident diagnosis with faster referral to specialized rheumatological care. Histones form the protein component of the nucleosome, around which DNA is wound [5, 6]. The core histones comprise dimers of H2A, H2B, H3 and H4 while H1 weakly associates with linker DNA between nucleosomes [5, 7]. Histone genes are organized in clusters resulting in multiple copies of each gene and different variant histone proteins whose functional distinctiveness is not well understood [5, 6]. Antibodies to nucleosomes and deposition of resulting immune complexes are implicated in the pathogenesis of SLE . Thus, testing for nucleosome antibodies is a promising way of improving the diagnostic potential of laboratory tests for SLE. However, it has been stressed that nucleosomes contain a wide variety of individual components such as different histones (or variants thereof) or dsDNA, and that efforts should be undertaken to more clearly define the target of potentially diagnostic antibodies . We therefore explored the diagnostic potential of antibodies to clearly defined histone core components H4 (HIST1H4A) and H2A (HIST2H2AA3 and HIST2H2AC) in comparison with routine antigens in an ELISA-based multiplex assay. Methods Patients and controls One hundred and fifty-three consecutive SLE patients [85% female, age 41 (13.5) years] who met 1997 updated 1982 ACR classification criteria for SLE (reference standard) were included from the rheumatology outpatient department of Heinrich-Heine-University between 1999 and 2000. Mean disease duration was 9.4 (7.3) years, and patients were treated with various compounds [e.g. 61.6% glucocorticosteroids (median dose 7.5 mg), 48.7% antimalarials]—further details are given in supplementary Table S1, available at Rheumatology online). Blood samples were drawn at the time of SLE classification, frozen at −80 °C and used for further antibody determination. Additional information gathered consisted of dsDNA antibodies as measured by radioimmunoassay, C3c and C4 values, CRP, total IgG, routine ENA testing including SSA, SSB, U1-RNP and Sm (Anti-ENA ProfilPlus 1, Euroimmune, Lübeck, Germany). Eighty-one healthy controls [76.8% female, 43.3 (12.4) years] who had an unremarkable examination by a trained rheumatologist during a rheumatology community screening programme were randomly selected to frequency-match our SLE group. The study complies with the Declaration of Helsinki. All patients gave their full informed consent. The study was approved by the Ethics Committee of the Medical Faculty of Heinrich-Heine-University Düsseldorf (study no. 2850). Multiplex antibody assay Recombinant antigens known to be associated to SLE and histones H4 or H2A were produced in Escherichia coli, purified as detailed elsewhere [9, 10], and covalently coupled to magnetic microspheres according to the manufacturer’s instructions (Luminex Corp., Austin, TX, USA). According to official GeneID and abbreviation (www.ncbi.nlm.nih.gov/gene/) the used antigens are named TROVE2 [SSA (60 kDa)], TRIM21 [SSA (52 kDa)], SSB (SSB), SNRNP70 (U1-RNP), SNRPB (Sm), RPLP2 (RPLP2), HIST2H2AC (HIST2H2AC), HIST2H2AA3 (HIST2H2AA3) and HIST1H4A (HIST1H4A). The coupled antigens were then incubated with probands’ sera, and after appropriate washing procedures incubated with a secondary PE-labelled anti-human-IgG antibody. A Magpix instrument (Luminex Corp.) was used for the detection and quantification of fluorescence in analogy to previous descriptions [9, 10] (index test). Statistical analysis Interclass correlations (ICCs) were computed comparing routine ENA on a standard assay to antibodies against the same parameters within the multiplex assay (i.e. SSA, SSB, U1-RNP and Sm) to estimate retest reliability. The standard assay and the multiplex were compared by receiver operating characteristic (ROC) analysis. ROC analysis was also performed to estimate the individual diagnostic performance for all parameters to distinguish SLE from healthy. Next, the cohort was randomly split into a training cohort (75%) and a validation cohort (25%). Binary logistic regression analysis was performed for the routine autoantibodies vs models including the histone antibodies (comparator models) and compared by likelihood ratio test. The accuracy of allocation to healthy and SLE was assessed in the validation cohort for all models. Clinical information and results of reference standard (i.e. SLE classification) were unavailable to the reader of the multiplex assay and results of the latter were unavailable to assessors of the reference standard. There were no missing data on either reference standard or index test within the current cohort. Generalized linear modelling was applied including all multiplex parameters and routine clinical information to assess prediction of disease activity measured by Systemic Lupus Activity Measure (SLAM). There was no prior estimation of sample size due to the exploratory nature of the study. R version 3.2.2 was used for statistical analysis (The R Foundation for Statistical Computing, Vienna, Austria). Results Comparison of the multiplex assay with a routine assay First, we were interested in the retest reliability of the multiplex assay. Therefore, ICCs were calculated for the parameters included in the standard assay used in clinical routine [i.e. SSA (60 kDa), SSA (52 kDa), SSB, U1-RNP and Sm] in five samples that were measured 12 times on different trays. Mean ICC was 0.85, ranging from 0.73 to 0.99 (supplementary Table S2, available at Rheumatology online). To determine the diagnostic performance, we compared the measurements of routine parameters SSA (60 kDa), SSA (52 kDa), SSB, U1-RNP and Sm in the multiplex assay with the routine assay by ROC analysis. Area under the curve (AUC) range was 0.85 (for U1-RNP) to 0.9 (for SSA) (further details in supplementary Table S3, available at Rheumatology online). When applying optimal thresholds (according to the Youden method) for the multiplex assay, only three (3.7%) of the specimens in the healthy control group tested positive [two SSA (60 kDa), one SSB]. Diagnostic performance of individual antibodies Next, we were interested in determining the performance of each individual antibody in the multiplex assay to diagnose SLE by ROC analysis. As can be seen in Table 1, AUC ranged from 0.72 to 0.97, with HIST1H4A antibodies performing best at a sensitivity of 95% for preset specificity of 90%. Routine dsDNA antibody measurements were not available for controls. We were nevertheless interested to ascertain that HIST1H4A-IgG levels did not merely reflect dsDNA antibody levels. First, correlation was assessed according to Spearman resulting in a weak association with a coefficient of 0.23 (P = 0.0094). Next, dsDNA antibody levels were dichotomized [positive (i.e. >7 IU/ml) or negative] and compared with HIST1H4A-IgG positivity at the optimal threshold (according to Youden with sensitivity 95.4%, specificity 92.6%). While dsDNA antibodies were false-negative in 47 cases, HIST1H4A-IgG failed to recognize only four cases. The chi-square test as a measure of association between the two markers was insignificant (P = 0.62, further details in supplementary Table S4, available at Rheumatology online). The participant flow diagram for the diagnostic performance of HIST1H4A-IgG is detailed in supplementary Fig. S1, available at Rheumatology online. Table 1 Diagnostic performance of individual antibodies within the multiplex assay to diagnose SLE vs healthy Antibody AUC (95% CI) Sens. at 90% spec. (95%CI) SSA (60 kDa) 0.72 (0.66, 0.79) 0.41 (0.33, 0.50) SSA (52 kDa) 0.87 (0.82, 0.91) 0.63 (0.50, 0.75) SSB 0.90 (0.86, 0.95) 0.76 (0.62, 0.92) U1RNP 0.85 (0.80, 0.90) 0.67 (0.54, 0.80) Sm 0.90 (0.86, 0.94) 0.65 (0.52, 0.80) RPLP2 0.90 (0.86, 0.94) 0.75 (0.56, 0.86) HIST2H2AA3 0.75 (0.68, 0.81) 0.33 (0.20, 0.58) HIST2H2AC 0.81 (0.75, 0.87) 0.50 (0.28, 0.65) HIST1H4A 0.97 (0.96, 0.99) 0.95 (0.88, 0.98) Antibody AUC (95% CI) Sens. at 90% spec. (95%CI) SSA (60 kDa) 0.72 (0.66, 0.79) 0.41 (0.33, 0.50) SSA (52 kDa) 0.87 (0.82, 0.91) 0.63 (0.50, 0.75) SSB 0.90 (0.86, 0.95) 0.76 (0.62, 0.92) U1RNP 0.85 (0.80, 0.90) 0.67 (0.54, 0.80) Sm 0.90 (0.86, 0.94) 0.65 (0.52, 0.80) RPLP2 0.90 (0.86, 0.94) 0.75 (0.56, 0.86) HIST2H2AA3 0.75 (0.68, 0.81) 0.33 (0.20, 0.58) HIST2H2AC 0.81 (0.75, 0.87) 0.50 (0.28, 0.65) HIST1H4A 0.97 (0.96, 0.99) 0.95 (0.88, 0.98) Area under the curve (AUC) and 95% CI according to receiver operating characteristic analysis with determination of sensitivity (sens.) and specificity (spec.) for diagnosis of systemic lupus erythematosus vs healthy control. Diagnostic performance and comparison of autoantibody combinations including histone antibodies In the next step, we hypothesized that addition of histone antibodies to our standard autoantibody set would increase the diagnostic capability. For this purpose, the data set was randomly split into a training cohort (75% of cases or controls) and a validation cohort (25%). Binary logistic regression models for the diagnosis of SLE vs healthy were then calculated for the standard antibodies alone or together with each of the histone antibodies. The goodness of fit of the different models to the actual data was compared in the training cohort. The accuracy of all models to correctly predict disease status was assessed in the validation cohort. As outlined in Table 2, addition of HIST2H2AC or HIST2H2AA3 antibodies resulted in a slightly better fit, while addition of HIST1H4A antibodies considerably improved model fit and predictive accuracy at a highly significant level. The logistic model including HIST1H4A is detailed in supplementary Table S5, available at Rheumatology online. Table 2 Comparison of binary logistic regression models to predict SLE vs healthy Model Ab included Resid. dev. P-value Acc. (train) (%) Acc. (valid.) (%) Standard SSA (60 kDa), SSA (52 kDa), SSB, U1-RNP, Sm, RPLP2 97.1 NA 89.1 86.4 Model 1 Standard + HIST2H2AA3 87.5 0.02 90.3 86.4 Model 2 Standard + HIST2H2AC 87.4 0.02 89.1 86.4 Model 3 Standard + HIST1H4A 45.9 4.3 × 10−11 95.4 89.8 Model Ab included Resid. dev. P-value Acc. (train) (%) Acc. (valid.) (%) Standard SSA (60 kDa), SSA (52 kDa), SSB, U1-RNP, Sm, RPLP2 97.1 NA 89.1 86.4 Model 1 Standard + HIST2H2AA3 87.5 0.02 90.3 86.4 Model 2 Standard + HIST2H2AC 87.4 0.02 89.1 86.4 Model 3 Standard + HIST1H4A 45.9 4.3 × 10−11 95.4 89.8 Binary logistic regression models for prediction of SLE vs healthy. A standard model incorporating the designated antibodies (Ab) was compared with models with the addition of histone antibodies (models 1–3). Model fit to data of the training cohort (resid. dev., residual deviance) was compared between standard and each model by likelihood ratio test (P). Acc.: accuracy of allocation to SLE or healthy in the training (train) and validation cohort (valid.). Assessment of the contribution of histone antibodies to predict disease activity Finally, we evaluated which parameters of the routine clinical and laboratory assessment were associated to disease activity represented by the SLAM and if the level of any individual antibodies could improve prediction of the SLAM. We therefore included all gathered routine parameters as outlined, as well as antibodies from the multiplex assay and performed stepwise linear modelling with backward elimination in order to identify significant predictors of disease activity. Significant predictors in the order of their relative impact consisted in CRP, damage index, C3c, dsDNA and HIST1H4A-IgG. Higher levels of HIST1H4A-IgG were associated with lower SLAM (details in supplementary Table S6, available at Rheumatology online). Discussion HIST1H4A-IgG antibodies demonstrated superior specificity and sensitivity for the diagnosis of SLE when compared with conventional diagnostic markers. The diagnostic utility was maintained in multivariate analysis and in a validation sample set in the current study. Previously, histone H2A and B antibodies were reported to be of superior utility in SLE diagnostics, while H4 antibodies performed less favourable in this regard [11–14]. Differences in the antibody detection technology, usage of non-recombinant H4 with contamination of other histones , or usage of non-human proteins  may account for the differences. Currently, we cannot exclude that immunosuppressive treatments may influence antibody formation to HIST1H4A. The heterogeneous treatment of our patients and a lack of treatment naïve SLE patients is therefore a limitation to the study. Additionally, besides SLE, other rheumatic disorders may be associated with histone antibodies . The diagnostic performance of HIST1H4A-IgG reported herein may prove to be lower when disease control groups and/or treatment naïve patients are included in future studies. Interestingly, lower concentrations of HIST1H4A-IgG were weakly associated to a more active SLE disease status. HIST1H4A-IgG may therefore be characteristically detectable in SLE patients, in whom they exert anti-inflammatory properties This notion is supported by the finding that histones act in a proinflammatory way by activating the innate immune receptors Toll-like receptor 2 and 4 , which in turn is ameliorated by inhibitory histone antibodies in experimental inflammatory models . However, the pathophysiological situation is likely to be more complex: combinations of several anti-histone antibodies especially in addition to dsDNA antibodies (but not distinct anti-histone antibodies alone) enhanced complement-dependent phagocytosis of necrotic cell material by leucocytes resulting in an oxidative burst . These findings collectively suggest that more pathophysiological and clinical studies on the association of distinct histone antibodies and their combinations with SLE disease activity are needed. This study also provides insight into a technical aspect of diagnostic marker selection: a multiplex approach using Luminex or similar platforms provides the means to assess potential new diagnostic markers such as HIST1H4A-IgG antibodies along with established markers for direct multivariate comparison. If a marker performs well in comparison to already established and available alternatives in this setting, further research into the development of a certified test seems promising. Conclusion HIST1H4A-IgG antibodies demonstrated high specificity and sensitivity for the diagnosis of SLE compared with healthy controls with superior diagnostic performance to conventional markers. HIST1H4A-IgG antibodies should be further assessed in SLE diagnostics. Acknowledgements M.S. was supported by the Hiller-Foundation, Erkrath, Germany. S.V. conceived the study, analysed and interpreted data, and drafted the manuscript. P.B. carried out antibody measurements, interpreted and analysed data. R.B. analysed and interpreted data. R.F.B. and J.R. followed patients and interpreted data. E.B., P.R. and H.G. carried out antibody measurements, and interpreted data. H.D.Z., P.B., P.S.K. and M.S. conceived the study and interpreted data. All authors critically revised the manuscript for intellectual content and approved the final version. Funding: No specific funding was received from any bodies in the public, commercial or not-for-profit sectors to carry out the work described in this manuscript. Disclosure statement: Pe.B. is an employee of Protagen AG. P.S.-K. is a Board Member and Shareholder of Protagen AG. H.-D.Z. is an employee of Protagen AG. All other authors have declared no conflicts of interest. Supplementary data Supplementary data are available at Rheumatology online. References 1 Shaikh MF, Jordan N, D’Cruz DP. Systemic lupus erythematosus. Clin Med (Lond) 2017; 17: 78– 83. Google Scholar CrossRef Search ADS PubMed 2 Liu C-C, Kao AH, Manzi S, Ahearn JM. Biomarkers in systemic lupus erythematosus: challenges and prospects for the future. Ther Adv Musculoskelet Dis 2013; 5: 210– 33. http://dx.doi.org/10.1177/1759720X13485503 Google Scholar CrossRef Search ADS PubMed 3 Leuchten N, Hoyer A, Brinks R et al. Performance of anti-nuclear antibodies for classifying systemic lupus erythematosus: a systematic literature review and meta-regression of diagnostic data. Arthritis Care Res 2017; Advance Access published 23 May 2017, doi: 10.1002/acr.23292. 4 Thong B, Olsen NJ. Systemic lupus erythematosus diagnosis and management. Rheumatology 2016; 56(Suppl 1): i3– i13. 5 Brown DT. Histone variants: are they functionally heterogeneous? Genome Biol 2001; 2:REVIEWS0006. 6 Millar CB. Organizing the genome with H2A histone variants. Biochem J 2013; 449: 567– 79. http://dx.doi.org/10.1042/BJ20121646 Google Scholar CrossRef Search ADS PubMed 7 Pisetsky DS. The complex role of DNA, histones and HMGB1 in the pathogenesis of SLE. Autoimmunity 2014; 47: 487– 93. http://dx.doi.org/10.3109/08916934.2014.921811 Google Scholar CrossRef Search ADS PubMed 8 Rekvig OP, van der Vlag J, Seredkina N. Review: antinucleosome antibodies: a critical reflection on their specificities and diagnostic impact: clinical impact of antinucleosome antibodies. Arthritis Rheumatol 2014; 66: 1061– 9. http://dx.doi.org/10.1002/art.38365 Google Scholar CrossRef Search ADS PubMed 9 Vordenbäumen S, Lueking A, Budde P et al. Sequential high-content profiling of the IgG-autoantibody repertoire reveals novel antigens in rheumatoid arthritis. Arthritis Res Ther 2016; 18: 235. Google Scholar CrossRef Search ADS PubMed 10 Budde P, Zucht H-D, Vordenbäumen S et al. Multiparametric detection of autoantibodies in systemic lupus erythematosus. Lupus 2016; 25: 812– 22. Google Scholar CrossRef Search ADS PubMed 11 Gioud M, Kaci MA, Monier JC. Histone antibodies in systemic lupus erythematosus. A possible diagnostic tool. Arthritis Rheum 1982; 25: 407– 13. Google Scholar CrossRef Search ADS PubMed 12 Rubin RL, Joslin FG, Tan EM. Specificity of anti-histone antibodies in systemic lupus erythematosus. Arthritis Rheum 1982; 25: 779– 82. http://dx.doi.org/10.1002/art.1780250712 Google Scholar CrossRef Search ADS PubMed 13 Ghedira I, Landolsi H, Mankai A, Fabien N, Jeddi M. Anticorps antihistones au cours du lupus érythémateux systémique, comparaison entre trois techniques : Elisa, dot blot et immunotransfert. Pathol Biol 2006; 54: 148– 54. Google Scholar CrossRef Search ADS PubMed 14 Konstantinov K, Russanova V, Russeva V. Antibodies to histones and disease activity in systemic lupus erythematosus: a comparative study with an enzyme-linked immunosorbent assay and immunoblotting. Arch Dermatol Res 1986; 278: 410– 5. http://dx.doi.org/10.1007/BF00418172 Google Scholar CrossRef Search ADS PubMed 15 Zirwas MJ, Kress DW, Deng J-S. The utility of antihistone antibody screening in the diagnosis of drug-induced lupus erythematosus. Arch Dermatol 2004; 140: 494– 5. Google Scholar CrossRef Search ADS PubMed 16 Xu J, Zhang X, Monestier M, Esmon NL, Esmon CT. Extracellular histones are mediators of death through TLR2 and TLR4 in mouse fatal liver injury. J Immunol ; 187: 2626– 31. CrossRef Search ADS PubMed 17 Xu J, Zhang X, Pelayo R et al. Extracellular histones are major mediators of death in sepsis. Nat Med 2009; 15: 1318– 21. http://dx.doi.org/10.1038/nm.2053 Google Scholar CrossRef Search ADS PubMed 18 Gullstrand B, Lefort MH, Tyden H et al. Combination of autoantibodies against different histone proteins influences complement-dependent phagocytosis of necrotic cell material by polymorphonuclear leukocytes in systemic lupus erythematosus. J Rheumatol 2012; 39: 1619– 27. http://dx.doi.org/10.3899/jrheum.111511 Google Scholar CrossRef Search ADS PubMed © The Author(s) 2017. Published by Oxford University Press on behalf of the British Society for Rheumatology. All rights reserved. [br]For Permissions, please email: email@example.com
Rheumatology – Oxford University Press
Published: Mar 1, 2018
It’s your single place to instantly
discover and read the research
that matters to you.
Enjoy affordable access to
over 18 million articles from more than
15,000 peer-reviewed journals.
All for just $49/month
Query the DeepDyve database, plus search all of PubMed and Google Scholar seamlessly
Save any article or search result from DeepDyve, PubMed, and Google Scholar... all in one place.
Get unlimited, online access to over 18 million full-text articles from more than 15,000 scientific journals.
Read from thousands of the leading scholarly journals from SpringerNature, Elsevier, Wiley-Blackwell, Oxford University Press and more.
All the latest content is available, no embargo periods.
“Hi guys, I cannot tell you how much I love this resource. Incredible. I really believe you've hit the nail on the head with this site in regards to solving the research-purchase issue.”Daniel C.
“Whoa! It’s like Spotify but for academic articles.”@Phil_Robichaud
“I must say, @deepdyve is a fabulous solution to the independent researcher's problem of #access to #information.”@deepthiw
“My last article couldn't be possible without the platform @deepdyve that makes journal papers cheaper.”@JoseServera