Background/Purpose: Myositis specific autoantibodies (MSA) produced in patients with polymyositis/dermatomyositis (PM/DM) are clinically useful biomarkers in diagnosis and management. Anti-Mi-2 antibodies that recognize nucleosome remodeling deacetylase complex are classic marker of DM. Anti-p155/140 (transcription intermediary factor 1 gamma and alpha, respectively) antibodies are one of the new MSA that have been studied actively because of their tight link to cancer-associated DM. A well known but poorly explained characteristic in MSA production is the rare coexistence of more than one MSA in each individual patient. We here report frequent coexistence of anti-Mi-2 and anti-TIF1 alpha antibodies as an exception for this concept.

Methods: Sera of patients with PM/DM from United States, Mexico, Italy, and Japan were screened for their autoantibody specificities by immunoprecipitation (IP) of ³⁵S-labeled K562 cells extract. Sera with anti-Mi-2 and anti-p155/140 were further characterized by IP-western blot (IP-WB) using monoclonal antibodies to TIF1alpha and TIF1gamma. Antibodies to TIF1alpha, gamma, and Mi-2 were also tested by ELISA using recombinant proteins. Clinical information was from the database and chart review.

Results: Forty-one anti-Mi-2 and 18 anti-p155/140 positive sera were identified and their characteristics are summarized in Table. Anti-Mi-2 positive sera immunoprecipitated a ~140kD protein that comigrates exactly with TIF1alpha. This protein was determined as TIF1alpha that was immunoprecipitated by coexisting autoantibodies to TIF1alpha in majority of human anti-Mi-2 sera based on 1) disappearance of the 140kD protein from human anti-Mi-2 IP by preincubation of cell extract with anti-TIF1alpha mAb, 2) mAb to Mi-2 and rare human sera with anti-Mi-2 do not IP the 140kD protein, 3) human anti-Mi-2 sera were positive for TIF1alpha by IP-WB, and 4) human anti-Mi-2 sera were positive in anti-TIF1alpha ELISA. Autoantibodies to TIF1gamma in human anti-Mi-2 sera appeared uncommon based on 1) the 155kD protein that comigrates with TIF1gamma is not seen in human anti-Mi-2 IP, 2) lack of anti-TIF1gamma reactivity in ELISA, 3) although many human anti-Mi-2 sera were also positive for TIF1gamma, the levels of TIF1gamma immunoprecipitated by human anti-Mi-2 sera are very little compared with those by human anti-p155/140 sera, suggesting that they are via co-IP by TIF1alpha.

Conclusion: Majority of human sera with anti-Mi-2 also have antibodies to TIF1alpha but not TIF1gamma. This is a new linked set of autoantibodies in DM and an exception of the rare coexistence of MSA in individual patient. This finding will affect identification and classification of PM/DM based on serology and may require reevaluation of clinical significance of anti-TIF1 alpha antibodies as anti-TIF1alpha coexisting with anti-Mi-2 has not been considered in the past.

Table 1. Reactivity of anti-Mi-2 and p155/140 sera