OP0118 DECIPHERING THE ANTI-PROTEIN-ARGININE DEIMINASE (PAD) RESPONSE IDENTIFIES PAD1 AND PAD6 AS NOVEL AUTOANTIGENS IN RHEUMATOID ARTHRITIS

Background:

Protein-arginine deiminase (PAD) 4 enzymes play a central role in the pathogenesis of rheumatoid arthritis (RA) and represents an antigenic target. Among the five known family members (PAD1, PAD2, PAD3, PAD4 and PAD6), only PAD2, PAD3 and PAD4 have been described to have autoantigenic properties. Furtheremore, very little is known on the the isotype usage of these autoantibodies. Understanding the molecular basis of the anti-PAD antibody reponse has the potential to open novel approaches for precision medicine in RA.

Objectives:

The objectives of this study were to screen for the presence of antibodies to the five PAD family members and to evaluate the isotype usage of the anti-PAD4 response in RA.

Methods:

First, we developed a panel for the detection of anti-PAD IgG based on a particle-based multi-analyte technology (PMAT), that utilized paramagnetic particles coupled with the different human recombinant PAD proteins (PAD1, PAD2, PAD3, PAD4 and PAD6) and anti-human IgG conjugate. This panel was used to test sera from RA patients (n=33) and non-RA controls (n=36). The controls were comprised of apparently healthy individuals (n=10), and patients with infectious diseases (n=10), systemic lupus erythematosus (n=7), systemic sclerosis (n=9) and Sjogren’s syndrome (n=1). Next, the PAD4-coupled beads were tested with anti-human IgM, IgA and IgG conjugates on an extended cohort of RA patients (n=62) and the same non-RA controls.

Results:

All five anti-PAD IgG (Figure 1) demonstrated the ability to discriminate between RA patients and controls. At greater than 90% specificity, anti-PAD4 IgG, followed by anti-PAD3 IgG, showed the best diagnostic performance. Significantly higher levels of the five antibodies were observed in RAvs.controls (p-values of 0.0041, <0.0001, 0.0014, 0.0039, and 0.0140 for anti-PAD1, 2, 3, 4 and 6, respectively). Significant correlation was observed between all the antibodies, with the highest between anti-PAD1 and anti-PAD4 (Spearman´srho=0.87,p<0.0001) and the lowest between anti-PAD4 and anti-PAD2 (Spearman’srho=0.38,p=0.0015) and anti-PAD4 and anti-PAD6 (Spearman’srho=0.38,p=0.0011). While principal component analysis (PCA) (Figure 2) showed an association between all anti-PAD antibodies, there was further discrimination that displayed closer association between anti-PAD1, 3 and 4 on one hand, and between anti-PAD2 and 6. For the extended testing of anti-PAD4 with IgG, IgA and IgM, all three isotypes were identified in the sera of RA patients. Higher levels of the three isotypes were observed in RA patients with erosive disease when compared with the patients without erosion, but this association was only significant for anti-PAD4 IgA (p=0.0086).

Figure 1.

Receiver operating characteristics (ROC) analysis of the discrimination between rheumatoid arthritis (RA) and controls of IgG to protein-arginine deiminase (PAD) 1, PAD2, PAD3, PAD4 and PAD6. The area under the curve (AUC) values are shown in brackets for each biomarker.Abbreviations:TPF: true positive fraction; FPF: false positive fraction

Figure 2.

Two dimensional principal component analysis (PCA) plot of the anti-PAD levels in RA patients (n=33) and controls (n=36). Anti-PAD1, 3 and 4 have the main contribution to PC1, which explains 51.7% of the variance, and anti-PAD2 and 6 to PC2, that represents 20.8% of it.Abbreviations:PC: principal component

Conclusion:

Our study is the first to describe PAD1 and PAD6 as novel antigenic targets in RA and to demostrate that the anti-PAD4 B-cell immune response uses all three isotypes (IgG, IgA and IgM). The strong and significant association between anti-PAD4 IgA and joint erosion is of particular clinical relevance.

Disclosure of Interests:

Laura Martinez-Prat Employee of: I am an employee of Inova Diagnostics, an in vitro diagnostics company., Mary Ann Aure Employee of: I am an employee of Inova Diagnostics, an in vitro diagnostics company., Chelsea Bentow Employee of: I am an employee of Inova Diagnostics, an in vitro diagnostics company., David Lucia Employee of: I am an employee of Inova Diagnostics, an in vitro diagnostics company., Marcos Lopez-Hoyos Consultant of: Inova Diagnostics, an in vitro diagnostics company., Michael Mahler Employee of: I am an employee of Inova Diagnostics, an in vitro diagnostics company.

Development and multi-center evaluation of a novel immunoadsorption method for anti-DFS70 antibodies

OBJECTIVE

Antinuclear antibodies (ANA) represent a hallmark in the diagnosis of ANA-associated rheumatic diseases (AARD). However, anti-DFS70 antibodies are present in a higher portion of the healthy individuals (HI) than in patients with AARD. Consequently, we developed a novel, highly specific indirect immunofluorescence (IIF) method that blocks anti-DFS70 antibodies from binding to HEp-2 cells and to evaluate the method in a multi-center study.

METHODS

A total of 18 samples from systemic lupus erythematosus patients (SLE, n = 7) and HI (n = 11) were used for the initial development of the immunoadsorption method. For the multi-center evaluation, samples with a dense fine speckled (DFS) pattern (n = 99) were collected at three different sites based on their established IIF screening procedure at the respective laboratories. Additionally, four characterized samples with established clinically relevant IIF patterns (centromere, nucleolar, speckled, homogeneous) were blended in five different ratios (10%, 25%, 50%, 75%, 90%) with a sample positive for anti-DFS70 antibodies, which by itself showed a dense fine speckled (DFS) IIF pattern. All samples were tested by IIF with NOVA Lite HEp-2 ANA and NOVA Lite HEp-2 Select on the NOVA View® instrument, and also tested by QUANTA Flash DFS70 chemiluminescent immunoassay (CIA) for confirmation of anti-DFS70 antibodies (Inova Diagnostics, San Diego, CA, USA).

RESULTS

For the development of the immunoadsorption method, only 1/7 ANA-positive samples from SLE patients, but 8/10 ANA-positive samples from healthy individuals turned negative using the immunoadsorption. Subsequently, 73/99 (73.7%) of the DFS pattern samples were positive by CIA for anti-DFS70 antibodies showing a strong quantitative Spearman's correlation (rho = 0.57 (95% CI, 0.39-0.71, p < 0.0001)) between light intensity units (LIU) measured by NOVA View and CIA. Intensities measured with NOVA Lite HEp-2 and NOVA Lite HEp-2 Select demonstrated significantly lower intensity values after inhibition with DFS70 antigen (p < 0.0001). When samples were processed to mimic samples with mixed patterns (DFS + clinically relevant pattern), the new immunoadsorption method demonstrated that all clinically relevant patterns remained unchanged whereas the LIUs from NOVA View analysis significantly decreased after inhibition (p < 0.0001).

CONCLUSION

The data showed that the NOVA Lite HEp-2 Select kit effectively inhibits anti-DFS70 antibody binding to its cellular target antigen.

International multi-center evaluation of a novel chemiluminescence assay for the detection of anti-dsDNA antibodies

OBJECTIVE

Anti-double stranded desoxyribonucleic acid (anti-dsDNA) antibodies are considered fairly specific for systemic lupus erythematosus (SLE) and their quantification is useful for the clinical management of SLE patients. We assessed the diagnostic performance of the QUANTA Flash dsDNA chemiluminescent immunoassay (CIA) in comparison to an ELISA, using patients from five participating countries. The main focus was to evaluate the correlation between anti-dsDNA antibody results from the CIA and global SLE disease activity, as measured by the SLE Disease Activity Index 2000 (SLEDAI-2K).

PATIENTS AND METHODS

A total of 1431 samples (SLE, n = 843; disease controls, n = 588) from five countries (Canada, USA, Portugal, Sweden and Spain) were tested with QUANTA Flash dsDNA (Inova Diagnostics, San Diego, CA, USA). Data obtained with the QUANTA Lite dsDNA SC ELISA (Inova Diagnostics) were available for samples from three sites (Canada, USA and Sweden, n = 566). The SLEDAI-2K scores were available for 805 SLE patients and a cut-off of > 4 was used to define active disease.

RESULTS

QUANTA Flash dsDNA had a sensitivity of 54.3% for the diagnosis of SLE, combined with 89.8% specificity. Anti-dsDNA antibody levels were significantly higher (p < 0.0001) in active SLE (SLEDAI-2K > 4; n = 232; median value 83.0 IU/mL) versus the inactive patients (n = 573; median value 22.3 IU/mL), and the SLEDAI-2K scoring correlated with their dsDNA antibody levels (Spearman's rho = 0.44, p < 0.0001). Similar but less pronounced findings were also found for the ELISA, in relation to disease activity.

CONCLUSIONS

The QUANTA Flash dsDNA assay showed good clinical performance in a large international multi-center study. Additionally, the strong correlation between anti-dsDNA antibody results and SLEDAI-2K scores supported the potential utility of QUANTA Flash dsDNA for monitoring disease activity.

Highlights on novel technologies for the detection of antibodies to Ro60, Ro52, and SS-B

OBJECTIVE

We aimed to compare a chemiluminescent immunoassay (CIA, QUANTA Flash) on BIO-FLASH with a multiplex flow immunoassay (MFI) on BioPlex 2200 for the detection of antibodies to Ro60, Ro52, and SS-B.

METHODS

The study included 241 samples, from patients suffering from systemic autoimmune diseases (n = 108) as well as disease controls (n = 133). All samples were tested for anti-Ro52, anti-Ro60, and anti-SS-B (La) antibodies on QUANTA Flash (INOVA Diagnostics, San Diego, USA) and BioPlex 2200 (Bio-Rad Laboratories Inc., Hercules, USA). Discrepant samples were tested by two independent methods: BlueDot/ANA and QUANTRIX Microarray (both D-tek, Belgium).

RESULTS

The overall qualitative agreements were 95.4% (95% confidence interval, CI 92.0-97.7%) for anti-Ro52, 98.8% (95% CI 96.4-99.7%) for anti-Ro60, and 91.7% (95% CI 87.5-94.9%) for anti-SS-B antibodies. There were 34 discrepant samples among all assays (20 anti-SS-B, 11 anti-Ro52, 3 anti-Ro60). 30/33 of retested samples (by D-tek dot blot) agreed with the QUANTA Flash results. Similar findings were obtained with QUANTRIX Microarray kit.

CONCLUSION

QUANTA Flash and BioPlex 2200 show good qualitative agreement. The clinical performances were similar for anti-Ro52 and anti-Ro60 autoantibodies while differences were observed for anti-SS-B (La) antibodies.