Myositis‐specific autoantibodies and their clinical associations in idiopathic inflammatory myopathies

Myositis‐specific autoantibodies (MSAs) have been found to be present predominantly in patients with idiopathic inflammatory myopathies (IIMs). This study aimed to investigate the prevalence of MSAs and their associated complications in a cohort of patients with IIMs.

hands, arthritis, and fever. 3 Examples of the anti-synthetase autoantibodies include anti-Jo-1, anti-PL-7, anti-PL-12, anti-OJ, and anti-EJ antibodies. Around 90% patients with anti-synthetase autoantibodies were found to have interstitial lung disease (ILD), and nearly 50% mortality in these patients was attributable to ILD. 4 The anti-melanoma differentiation-associated gene 5 antibody (anti-MDA5 Ab), or formerly known as the anti-CADM-140 autoantibody, was first described in Japan in 2005. 5 It has become one of the most important MSAs especially in Asia, and was found to be specifically expressed in patients with clinically amyopathic dermatomyositis (CADM).
These patients may have a tendency to develop rapidly progressive interstitial lung disease (RP-ILD), which is associated with high mortality. 6 A recent study in Hong Kong showed that anti-MDA5 Ab was present exclusively in 30% of patients with dermatomyositis (DM), which were all clinically amyopathic and was significantly associated with RP-ILD, suggesting the clinical usefulness of examination of this autoantibody. 7 On the other hand, the anti-transcriptional intermediary factor 1-gamma antibody (anti-TIF1-γ Ab) was found to be significantly associated with malignancies. 8,9 Despite the suggestion of association with potential life-threatening complications in IIMs, currently testing of MSA is not universally performed. Up to this date, the MSA profile has not been well established in different populations, and apparently, there are discrepancies in the prevalence and phenotypic presentation of MSAs in different cohorts. 10 The primary objective of the study was to establish the prevalence of MSAs in existing adult Chinese patients with IIMs in Hong Kong. The secondary objective was to examine the association between different MSAs and the clinical features, as well as complications of these patients.

| Patient recruitment
Consecutive patients with IIMs seen in the rheumatology clinics and wards of the participating major regional hospitals were recruited in this multicentered prospective study. Diagnosis of IIMs was based on Bohan and Peter's criteria with probable or definite cases included. 11 Patients with CADM were also recruited, and they must have the typical Gottron's or heliotrope rash as determined by rheumatologists or dermatologists, and with no symptoms or signs of muscle involvement. 12

| Identification of antibody
The line blot technique was used in this study. It is based on an immunoblotting procedure with highly purified antigens coated on protein-binding membrane without the need of gel electrophore-

| Statistical analysis
The SPSS statistical package version 23.0 was used. Descriptive statistics were presented as frequencies, means with standard deviation or medians with ranges as appropriate. Comparisons between clinical variables were done using the chi-square test or Fisher's exact test for categorical variables, independent-samples t test for continuous variables with normal distribution, or Mann-Whitney U test for nonparametric continuous variables. In each MSA subset, demographic features and clinical characteristics were studied, and antibody-positive cases were compared with antibody-negative cases. Logistic regression was used to determine the independent risk factors for RP-ILD, while Cox regression was used to investigate the independent risk factors for malignancy. Potential explanatory variables of the outcomes were selected from the univariate analyses. They should have a P-value < .1 and were biologically plausible.
Results were considered statistically significant if the P-value was <.05.

| Association of MSAs with clinical features
For the major MSA subsets, antibody-positive cases were compared with negative cases for different clinical characteristics, as shown in Table 3.
Two out of 5 anti-SAE1 Ab-positive patients were found to have malignancy, and 4 of them had CADM. One of the patients died 5 months after the diagnosis of myositis of uncertain cause.
One of the 4 patients with anti-NXP2 Ab had biopsy-proven calcinosis cutis, and one patient died 2 months after diagnosis because of infection. None of the anti-NXP2 Ab was found to have malignancy. No statistical tests were performed due to the small sample sizes.

| Risk factors for malignancy
There were 34 malignancies diagnosed after and 3 before the onset of myositis. After adjusting for age, gender, smoking history, alcohol history, current use of immunosuppressants, and the presence of refractory rash, it was found that anti-TIF1γ Ab was an independent risk factor for malignancy (HR 3.42). Other independent risk factors included the DM subtype (HR 3.87) and family history of cancer (HR 3.67). History of immunosuppressant use was a negative predictor of malignancy (HR 0.355). Details are shown in Table 5.

| Survival of patients with different MSAs
The survival curves of patients with different MSAs are shown in

| D ISCUSS I ON
The prevalence of anti-MDA5 Ab was found to differ among different ethnicities. In a study performed in the United States in 2013, anti-MDA5 Ab was only found in 11 out of 160 (6.9%) DM patients. 14 In this study, it was found to be much more prevalent, and was identified in 23.0% of DM/CADM patients. In one study comparing MSAs in 145 Chinese and 165 Japanese patients with IIMs, the prevalence of anti-MDA5 Ab was found to be 36.6% and 15.8%, respectively, in all IIM patients. 15 The prevalence of anti-MDA5 Ab in all IIM patients in the current study was 13.9%, which was comparable to the Japanese cohort but significantly lower than that of the Chinese cohort. The discrepancies between the prevalence of anti-MDA5 Ab in Chinese and Japanese population may be explained by genetic and environmental factors. The association of anti-MDA5 Ab with DRB1*0101/*0405 was reported in Japanese patients, 16 while the combined allele frequencies of DRB1*0101 and DRB1*0405 were different between Japanese and Chinese. 17,18 However, the reason behind the even greater discrepancy between Hong Kong and Chinese patients was uncertain.
One possibility could be the different MSA detection methods being used. In the Chinese and Japanese cohort, ELISA technique was used, while the current study employed the line blot assay.
A study comparing the line blot technique with immunoprecipitation for the detection of MSAs showed good concordance rates, justifying the use of the line blot testing in our study. 19 Meanwhile, the potential environmental influences on the frequency of MSAs should be further investigated.
Another striking feature of the pattern of MSAs in our cohort was the prevalence of anti-TIF1γ (13.9%), which was much higher than that of the China (5.5%) and Japanese (8.5%) cohorts, as well as a large Caucasian (7%) cohort. 15 spectively. This is also apparently lower than those reported in the Caucasian cohorts. 10 Again, the lower prevalence could be explained F I G U R E 1 Survival curves of patients with different myositis-specific antibodies by the difference in detection method, as well as genetic and environmental factors.

TA B L E 5 Multivariate analysis of risk factors for malignancy after adjustment, by Cox proportional hazards regression
In this study, 7.0% of the patients were found to have more than one positive MSA. MSAs are supposed to be mutually exclusive, and previous studies have shown that MSAs rarely coexist with each other. 23 However, there were reports of sporadic cases in which the MSAs coexist and were associated with more complex disease expression. 24 ILD. 33 One postulation is that the up-regulation of the type I IFN system and modification of host antigen after pulmonary TB infection could induce the dreadful disease in genetically predisposed individuals. Unfortunately, molecular analyses were not available in the current study.
Anti-TIF1γ Ab was found to be significantly associated with refractory skin rash and malignancy, and negatively associated with ILD. The association between anti-TIF1γ Ab and malignancy has been well established. 9 The findings here were in agreement with the previously described characteristics in anti-TIF1γ Ab-positive patients.
The use of immunosuppressants appeared to have a "protective" effect on malignancy in the regression analysis. It is likely because immunosuppressants were more used in patients who suffered from complications such as ILD, which were negatively associated with malignancy. Moreover, use of these agents might be avoided in patients at a high risk of malignancy, which further contributes to the possible confounding by indication. On the other hand, a recent Chinese study showed that apart from anti-TIF1γ Ab, anti-NXP2, and anti-SAE1 Ab were both associated with increased risk of malignancy in patients with IIMs. 34 In our study, out of the 5 anti-SAE1 Ab-positive patients, 2 of them had malignancy (CA colon and NPC).
However, the sample size was not large enough to give a statistically significant result. Larger studies will be necessary to confirm the association of malignancy with these antibodies. Meanwhile for anti-TIF1γ Ab-positive patients, vigilant malignant screening would be necessary.
There were a few limitations in this study. Firstly, despite the increasing popularity, the line blot immunoassay has not been fully validated. Some antibodies, for example anti-OJ antibody, could not be measured accurately by the line blot assay. 35 The assay also did not test for anti-HMG-Coenzyme A reductase (HMGCR) antibody, which was estimated to be present in around 5% of patients with IIMs. 36

ACK N OWLED G EM ENTS
None.

CO N FLI C T O F I NTE R E S T
All authors report no disclosures relevant to the manuscript.

DATA AVA I L A B I L I T Y S TAT E M E N T
Data can be shared upon request.