Cross-reactive anti-CENP-A autoantibodies induce analytic interference in anti-TIF1γ detection using line-dot immunoassay

Dear Editor, Idiopathic inflammatory myopathies (IIM) are a heterogeneous group of autoimmune diseases, characterized by inflammatory cell infiltrates in muscle tissue. They usually manifest as muscle weakness but can also be associated with extra-muscular manifestations [1]. Myositis-specific autoantibodies (MSA) play a key role in diagnosis and are increasingly recognized as important tools for the classification, management and prognosis of IIM. At present, different techniques exist to detect the presence of MSA in the serum of patients. Immunoprecipitation is considered as the gold standard for most autoantibodies but is commonly not available in routine laboratories. Line-dot immunoassays, which use recombinant antigens, allow multiple autoantibodies testing at the same time and represent an easy to handle and fast option to detect these autoantibodies in daily practice. Different commercial assays are available, but the lack of standardization leads to variable performances in terms of sensitivity, specificity and predictive values [2–4]. Among MSA, discordances between tests have been especially described for anti-transcriptional intermediary factor 1 gamma (TIF1γ) antibodies [5] (except with one ELISA [6]), with moderate to low overall concordance rate between different assays [7] and a detection that has been considered as not fully reliable by some authors [8]. However, identification of antibodies directed against TIF1γ is of particular interest as these autoantibodies are especially associated with severe skin manifestations, and are strongly associated with malignancy. In a recent work, we retrospectively analysed the clinical characteristics of patients carrying anti-TIF1γ, anti-nuclear matrix protein 2 (anti-NXP2) or anti-SUMO-activating enzyme subunit 1/2 (anti-SAE1/2) autoantibodies detected using a line-dot immunoassay from Alphadia^® (Myositis Profile 12 A IgG Dot for BlueDiver, Mons, Belgium). Among them, 23 patients were positive for anti-TIF1γ antibodies but did not present any clinical phenotype of DM. Surprisingly, 8 of those 23 patients had SSc with positive ACA [anti-centromere B (CENP-B)] detected by chemiluminescence (Bioflash, Inova, Werfen, France) [9].

Prospectively, we identified three new sera of patients with anti-centromere positive SSc who also presented a reactivity against TIF1γ on the immunoassay from Alphadia^® (Myositis Profile 12 A IgG Dot for BlueDiver). In order to assess whether this reactivity could be due to antibodies cross-reactivity, we performed inhibition assays for the three patients that we identified prospectively (patients A, B and C in Fig. 1). Due to lack of remaining serum afterwards however, we could not replicate the data that are presented in Fig. 1. Recombinant full-length CENP-B protein produced in insect cells was a kind gift from ThermoFisher/Phadia Company. Synthetic CENP-A peptide was provided by INOVA Diagnostics (San Diego, USA). The purity of the peptide was 95% as assessed by mass spectrometry analysis and HPLC chromatogram. Synthetic CENP-A peptide encompassed the N-terminal sequence of human CENP-A, which contains the main immunodominant linear epitopes targeted by anti-CENP-A autoantibody as described in [10–13]. It also contains the G/A-P-R/S-R-R auto-antigenic motif present as cryptic mimotopes in several autoantigens [14]. Because of a limited sample volume, one serum (serum B) was incubated with recombinant CENP-B and synthetic peptide CENP-A together. Sera reactivity against TIF1γ protein on the Alphadia^® immunodot was assessed before and after inhibition assay. As a positive control we used anti-TIF1γ antibodies positive serum from a DM patient. As seen in Fig. 1A, pre-incubation with either CENP-B recombinant protein or CENP-A synthetic peptide did not modify anti-TIF1γ signal (Fig. 1A). In contrast, in the three SSc patients’ sera pre-incubated with CENP-A peptide, the anti-TIF1γ signal was significantly reduced, while no modification was observed with recombinant CENP-B (Fig. 1B). Interestingly, none of the sera tested positive using another available commercial kit, the EUROLINE^® Myositis Profile (Autoimmune Inflammatory Myopathies 16 Ag, Euroimmun, Lübeck, Germany), suggesting that the cross-reactivity was restricted to the Alphadia^® assay (data not shown). We did not perform immunoprecipitation, as this technique is not available in our laboratory, and as patients’ serum samples were in short supply.

Fig. 1

(A) Detection of anti-TIF1γ autoantibodies in one DM patient (positive control) in Alphadia^® immunoassay before or after incubation with 70 µg/ml of recombinant CENP-B protein or CENP-A synthetic peptide. (B) Detection of anti-TIF1γ autoantibodies in three sera of two SSc patients (A–C) in Alphadia^® immunoassay before or after incubation with 70 µg/ml of recombinant CENP-B protein or CENP-A synthetic peptide. For serum B, CENP-B and CENP-A synthetic peptide were incubated together due to limited serum quantity. Arrowheads indicate anti-TIF1γ reactivity in line-dot. TIF1γ: transcriptional intermediary factor 1 gamma; CENP-A/B: centromere A/B peptides.

Open in new tabDownload slide

Together, these results unveil a cross-reactivity between anti-CENP-A autoantibodies and TIF1γ protein, even though the prevalence of false-positive TIF1γ results is still to be defined in the global anti-CENP-A-positive SSc patient population. The absence of inhibition in the DM patients’ serum suggests that the epitopes recognized by anti-CENP-A on TIF1γ are different from the ones recognized by anti-TIF1γ antibodies associated with clinically proven DM. In addition, the restriction of the cross-reactivity to the Alphadia^® immunoassay is possibly related to the specificity of the structure of the antigen absorbed in this particular assay.

Anti-CENP-A autoantibodies have been already described to cross-react with numerous antigens such as transcription factor forkhead box E3 protein [15], Epstein–Barr nuclear antigen 1 [14] or Bicaudal D cargo adaptor 2 (BICD2) [16]. Those cross-reactivities involve linear but also cryptic epitopes. Interestingly, the linear mimotrops SRKPEAPRRRSPSP and SPSPTPTPGPSRRG, reacting with multiple antigen peptides [10], were not found in TIF1γ sequence. Hence, whether antigenic alteration in the Alphadia^® assay allows the recognition of a conformational or cryptic epitope exposed to anti-CENP-A, or whether this cross-reaction might happen in patients and might be involved in clinical specific features, is still to be determined.

In conclusion, we describe here a specific analytic interference in anti-TIF1γ detection associated to a cross-reactivity with anti-CENP-A autoantibodies in SSc patients. As MSA gain importance in the management of IIM and as line-dot immunoassays appear to be interesting tools in this field, we believe it is important that both clinicians and laboratory professionals are fully aware of the possible pitfalls associated with these techniques. Interpretation of the results must always be done with considerable caution.

Disclosure statement: The authors declare no conflicts of interest.

References

Author notes