Anti-melanoma-associated Antigen A1 Autoantibodies Predict the Outcome of Patients With Idiopathic Inammatory Myopathies Related Interstitial Lung Disease

The discovery of novel autoantibody in patients with idiopathic inammatory myopathies (IIMs) is of great signicance for clinical subtype stratication and potential pathogenesis. Previous literatures have found that tumor-associated antigens were involved in the pathogenesis of IIM. Therefore, in this study, we aimed to explore the prevalence and clinical association of autoantibodies against melanoma-associated antigen A1 (MAGE-A1) in patients with IIM. ELISA was performed to detect anti-MAGE-A1 autoantibodies in patients with IIM, systemic lupus erythematosus, rheumatoid arthritis, primary Sjögren syndrome, systemic sclerosis, and healthy controls, and the results were conrmed using the dot-immunoblotting assay. The association between anti-MAGE-A1 autoantibody and clinical characteristics was analyzed in IIM patients. T test, Mann-Whitney U test, double-sided Pearson's Chi-square, Fisher exact test, Spearman correlation analysis as well as the generalized estimating equation were applied in the statistical analyses. growth


Background
Idiopathic in ammatory myopathies (IIMs) are a group of heterogeneous systemic autoimmune diseases characterized by in ammatory cell in ltration in skeletal muscles, potentially resulting in multiple organ damage and a high degree of morbidity and mortality. Dermatomyositis (DM), polymyositis (PM), and immune-mediated necrotizing myopathies (IMNM) with distinct clinical, serological, and histopathological features are common subtypes in adult IIM patients.
Interstitial lung disease (ILD) is a common feature of IIM patients and a critical factor associated with disease severity and mortality. Myositis speci c autoantibodies (MSAs) have been recognized as one of the most important biomarkers for clinical subtype strati cation and potential factors involved in IIM pathogenesis (1)(2)(3). Some MSAs such as anti-aminoacyl-tRNA synthetase (ARS) and anti-melanoma differentiation-associated gene-5 (MDA5) autoantibodies are strongly associated with ILD and poor prognosis of IIM patients. Measurement of serum MSA levels could serve as a useful tool to predict the outcome of IIM. However, IIM patients with one MSAs can present different clinical process and prognosis. For example, patients with anti-MDA5 autoantibodies are at the highest risk of rapidly progressive ILD (RP-ILD) and have the worst prognosis among all IIM patients (4)(5)(6)(7). However, anti-MDA5-positive patients with poor prognoses have signi cant increased serum ferritin level than those with favorable outcome (8)(9)(10). Moreover, previous study reported that IIM patients carrying antithreonyl tRNA-synthetase (PL-7) and anti-Ro-52 autoantibodies have worse outcomes than those with anti-PL-7 autoantibody (11). These ndings suggest the need to identify clinical biomarkers for IIM.
Melanoma-associated antigen-A1 (MAGE-A1), rst reported as one of the cancer-testis antigen (CTAs) by Van Der Bruggen et al. in 1991(12), is a useful biomarker for tumor imaging and immunotherapy (13)(14)(15). MEGE-A1 belongs to the MAGE family located on the X chromosome, which can be classi ed as type I and type II MAGEs (18). Under normal circumstances, MEGE-A1 is expressed only in normal testes. However, the MEGE-A1 promoter in humans is highly methylated. In cases of unstable demethylation, MEGE-A1 is upregulated in various tumor tissues and is associated with tumor progression and a poor prognosis (15,17,18).
A subsequent study investigated the role of MAGE genes in autoimmune diseases. McCurdy et al. reported that MAGE-A1 is overexpressed in synovial uid cells and peripheral blood mononuclear cells of juvenile rheumatic arthritis (JRA) both at the mRNA and the protein levels, suggesting that MAGE-A1 dysregulation is potentially associated with autoimmune disease and contributes to the pathogenesis of autoimmune diseases (19). Tumor-associated antigens, including transcription intermediary factor 1-γ (TIF1-γ), are potentially overexpressed in regenerating muscle bers of IIM, and anti-TIF1-γ autoantibodies can be detected in the serum of IIM patients (20,21). These results suggest that some tumor antigens may be converted to muscle antigens under unknown factors to stimulate the immune response for autoantibody production, thus mediating muscle ber and organ damage associated with the clinical disease phenotype (22)(23)(24). However, no studies to date have investigated the role of MAGE-A1 in IIM.
This study aimed to investigate the association between anti-MAGE-A1 autoantibodies and the clinical features of IIM.

Patients
In total, 576 adults IIM patients, including 390 dermatomyositis (DM) patients, 69 amyopathic dermatomyositis (ADM) patients, and 117 polymyositis (PM) or immune-mediated necrotizing myopathy (IMNM) patients, admitted to the department of rheumatology at China-Japan Friendship hospital from 2008 July to 2018 March were enrolled in this study. All the patients ful lled the 2017 EULAR/ACR IIM classi cation criteria (25). Since juvenile dermatomyositis and sporadic inclusion body myositis were rare in our cohort, these two IIM subgroups were excluded to reduce selection bias. The clinical data were retrospectively obtained from hospital medical records. Interstitial lung disease (ILD) was diagnosed in accordance with the features of high-resolution chest computed tomography (HRCT). RP-ILD was de ned as the radiologic aggravation accompanied by progressive dyspnea and/or hypoxemia within 3 months of onset of respiratory symptoms (8,26

Generation Of The Mage-a1 Peptides
Based on the gene name, corresponding human gene sequences were obtained from NCBI and their protein sequences were downloaded (msleqrslhc kpeealeaqq ealglvcvqa atssssplvl gtleevptag stdppqspqg asafpttinf trqrqpsegs ssreeegpst scileslfra vitkkvadlv g llkyrar epvtkaemle sviknykhcf peifgkases lqlvfgidvk eadptghsyv lvtclglsyd gllgdnqimp ktg iivlv miamegghap eeeiweelsv mevydgrehs aygeprkllt qdlvqekyle yrqvpdsdpa rye wgpra laetsyvkvl eyvikvsarv rfffpslrea alreeeegv). Thereafter, using the online prediction website IEBD Analysis Resource (http://tools.immuneepitope.org/main/), B cell epitopes were predicted and peptides with a segment length of 25aa (molecular weight, 2838 g/mol) (fpttinf trqrqpsegs ssreeegp) with the highest score were selected on the basis of the amino acid sequence and Hidden Markov Model. The selected peptide segments were then synthesized at Genscript Biotechnology (https://www.genscript.com.cn/), with a purity of ≥ 85%. Finally, the peptides were dissolved in MES buffer and stored at -30℃. 3. Establishment Of Anti-mage-a1 Elisa System ELISA was performed herein. Synthesized MAGE-A1 polypeptides were incubated in vitro at 200 ng together with 100 µg 1-ethyl-3-(3-dimethylaminopropyl) per well in MES buffer (PH 6.0) at 4 °C overnight. After three washes with double steamed water, ELISA plates were blocked at 37 °C for 2 h with 5% skim milk powder. Next, 3 µl of patient sera was diluted with 2.5% milk (dilution 1:50) and added 100 µl of mixed solution to each well and incubated for 1 hour, followed by probing with goat anti-human IgG (dilution 1:20,000) (Abcam, Cambridge, UK) to detect autoantibodies. Serum and goat anti-human IgG were washed ve times with 0.05% PBST. Finally, the absorbance was measured at 450 nm in a microplate reader. All the serum samples were analyzed more than twice to ensure consistency. Based on serial concentration of serum samples with a high titer of anti-MAGE-A1 antibodies, a standard curve was constructed, 1: 25 dilution was de ned as 64U, 1: 50, 32U; similar to the antibody levels, as U/ml (see Supplementary Fig. 1, Additional File 1). If a particular sample presented as an outlier in the standard curve, further dilution was required. The cutoff level was set at 3.3737 U, which was determined through three standard deviations (SDs) above the average of 165 sera of HCs.

Validation Of Anti-mage-a1 Autoantibody
Commercial full-length recombinant MEAG-A1 protein (Origene, Rockville, USA) was subjected to a dotimmunoblotting assay to validate the results of ELISA among anti-MEGE-A1-positive patients and HCs. Recombinant MAGE-A1 protein diluted with PBS (dilution 1:50) was doted on a nitrocellulose membrane (200 ng/dot) at room temperature (24-28℃) for 5 min. Then the membrane was placed in 5% skim milk at room temperature for 2 hours. Afterwards, incubated 2 µl serum sample (dilution 1:500) at 4℃ overnight and washed the membrane for 6 times with 0.025%PBST solution. Thereafter, a secondary antibody against human sera (dilution 1:40000) was added. Finally, after repeating the washing step, ECL was used to visualize reactive dots.
A continuous 10 cm visual analog scale (VAS) for physician global assessment (PGA) of patients harboring the anti-MAGE-A1 autoantibody was used to evaluate disease activity in accordance with myositis core set measures (CSM) established by the International Myositis Assessment and Clinical Studies (IMACS) (27). Disease activity was assessed upon IIM diagnosis and during every follow-up visit. Disease courses were divided into the following four types: (1) monocyclic course, de ned as patients retaining no clinical and biochemical signs of disease activity during 2-year follow-up after initial therapy; (2) polycyclic course, de ned as patients presenting more than at least one relapses during 2-year followup; (3) chronic course, de ned as active diseases on a 2-years visit after IIM diagnosis and no sign of disease remission despite regular treatment; (4) unde ned, indicating that patient were followed-up for l < 2 years after diagnosis (28-30).

Statistical Analysis
Continuous variables were presented as mean ± standard deviation (SD) or median (interquartile range, IQR), and the Mann-Whitney U test was performed to compare non-normal distributed data. While categorical variables were described as numbers or percentages, they were compared using double-sided Pearson's Chi-square or Fisher exact test. Spearman correlation analysis was performed for correlation analysis in the cross-sectional study and the generalized estimating equation (GEE) was applied in the longitudinal study. P-values less than 0.05 were considered statistically signi cant. Data were visualized and analyzed using SPSS (version 25.0) and GraphPad Prism (version 8.0).
The second most common MSA that coexisting with anti-MAGE-A1 autoantibody was anti-MDA5 antibody (See Supplementary Table 1, Additional File 1), however, no signi cant differences in the characteristics of ILD and disease course were observed between anti-MDA5-positive ILD patients with or without anti-MAGE-A1 autoantibody (See Supplementary Table 2, Additional File 1).

Correlation between serum anti-MAGE-A1 autoantibody levels and disease activities
In the cross-sectional study, no correlation was found between serum anti-MAGE-A1 level and PGA-VAS (r=-0.049, p = 0.800) among anti-MAGE-A1-positive patients. However, sera from 11 patients followed up at least more than twice were obtained to monitor changes in anti-MAGE-A1 autoantibody levels. Among these patients, four out of 11 were anti-MDA5-positive, two were anti-ARS-positive, two were MSAnegative, and one was anti-NXP2-, anti-HMGCR-, and anti-SRP-positive. The levels of anti-MAGE-A1 autoantibodies decreased with a reduction in disease activity. Anti-MAGE-A1 autoantibody levels were positively correlated with PGA-VAS on multiple GEE analysis performed in the longitudinal study (β = 1.071, p < 0.0001) (Fig. 3).

Discussion
The present study shows that that the incidence of anti-MAGE-A1 autoantibody is approximately 5% in IIM patients, and this autoantibody can be present in patients with all adult subtypes of IIM including DM, Declarations authors have contributed to the last version of the manuscript. The authors read and approved the nal manuscript.

Funding
The work was supported by National Natural Science Foundation of China (81701615, 81971531), the Science and Technology Commission Foundation of Beijing (Z181100001718063, Z171100001017208).

Availability of data and materials
All data generated or analysed during this study are included in this published article [and its supplementary information les].