Deep Pathological Phenotyping of Dermatomyositis with Different Autoantibodies

Objective Dermatomyositis with different myositis-specic autoantibodies has distinctive clinical presentations. Pathological variation of patients with different antibodies has not been fully understood. Methods A retrospective review of muscle pathological features was performed in dermatomyositis patients with known myositis-specic antibodies. Results A total of 46 dermatomyositis patients with one myositis-specic autoantibody (anti-MDA5 11, anti-Mi-2 10, anti-NXP2 13, anti-TIF1γ 8, anti-SAE 4) were included and the pathological severity score was evaluated. Patients with anti-Mi-2 demonstrated higher pathological severity scores and apparent sarcolemmal complement deposition, which was in consistency of more severe muscle weakness and higher level of muscle enzymes. In contrast, patients with anti-MDA5 generally had minimal pathological changes in muscle with less inammatory cell inltration, fewer membrane attack complex deposition, and milder myxovirus resistance protein A upregulation. Patients with anti-SAE had more inammatory cell inltration and MAC deposition compared to anti-MDA5 group. Muscle pathological scores varied largely in patients with anti-NXP2 and anti-TIF1γ. The muscle pathological features varies among dermatomyositis with different autoantibodies, which further indicates the heterogeneity of dermatomyositis.


Introduction
Dermatomyositis (DM) is an idiopathic in ammatory myopathy with a range of muscle and skin manifestations. Muscle involvement usually manifests as proximal muscle weakness, with or without myalgias or tenderness.
Typical skin lesions in DM include heliotrope rash, periorbital edema, Gottron papules, Gottron sign, V-sign and shawl sign. Two special subtypes of DM, the amyopathic variant and dermatomyositis sine dermatitis, have been described in the 2017 UELAR/ACR DM criteria [1]. In addition to muscle and skin manifestations, interstitial lung disease, joint disease and malignancy are may also present in DM patients. The inconsistent clinical ndings result in challenges in diagnosis and treatment of DM patients.
In the last decades, about 20 myositis-speci c antibodies (MSA) and myositis-associated antibodies (MAA) were discovered [2]. DM is associated with ve MSAs including anti-melanoma differentiation associated gene 5 (MDA5), anti-mitochondrial M2-associated protein (Mi-2), anti-nuclear matrix protein 2 (NXP2), anti-transcription intermediary factor 1 (TIF1γ) and anti-small ubiquitin-like modi er activating enzyme (SAE) antibodies. These autoantibodies are associated with distinct clinical features [3]. For example, anti-MDA5 dermatomyositis is more associated with mucocutaneous ulceration, mild muscle weakness and progressive interstitial lung disease [4]. The anti-Mi-2 patients is reported to be associated with typical dermatomyositis skin lesion, high levels of muscle enzymes and good response to therapy [3,5]. The anti-NXP2 and anti-TIF1γ dermatomyositis is associated with a higher risk of malignancy [6,7]. The anti-SAE antibody was found exclusively in adult DM patients, and the common symptoms include a diffuse dark red skin rash and mild muscular weakness, but dysphagia occurred more frequently [8]. All these serological ndings induce a new clinic-serological classi cations of IIMs based on MSA and MAA [2].
Although different MSAs are associated with distinct clinical features, the pathological phenotype of different MSA subgroups remains unclear. Recently a surrogate marker of induce type I interferon (IFN1) pathway, myxovirus resistance protein A (MxA), has been recognized as a speci c marker for DM with much higher sensitivity than perifascicular atrophy (PFA), the classic pathological feature of DM. [9] Anti-MDA5 antibodies were reported to be associated with relatively normal muscle biopsy without PFA, while anti-Mi-2 patients demonstrated classic PFA with perifascicular necrosis and MxA upregulation [10]. In this study we aim to study the detailed muscle pathology pro les of different MSA subgroups, which may help revealing different underlying pathogenesis of muscle damage.

Patients And Methods
Patients, clinical data and muscle imaging DM patients diagnosed in Huashan Hospital, Fudan University and Tongji Hospital, Huazhong University of Science and Technology from 2016 to 2019 were included in this study. The diagnosis of DM was according to the European Neuromuscular Centre (ENMC) criteria in 2004 [11]. Clinical data, laboratory tests and treatment status at time of biopsy were recorded. Muscle strength was evaluated according to MMT-8 scores (scale 0-80) at the time of diagnosis [12]. Magnetic resonance imaging (MRI; T1-and T2-weighted and short T1 inversion recovery) was used to scan the muscles in the bilateral lower extremities before a biopsy [13]. The study was performed in accordance with the Declaration of Helsinki and was approved by the Ethics Committees of Huashan Hospital (2018 − 153) and Tongji Hospital (TJ-C20121221). Informed consents were obtained from all the patients.

Autoantibody testing
Serum from patients with dermatomyositis were sent to Oumeng V Medical Laboratory for screening autoantibodies using ELISA test as described previously [14]. Brie y, the assay of the MSAs and MAAs subsets was performed with EUROLINE autoimmune in ammatory myopathy Ag (IgG) test kit (EUROIMMUN, Germany). The positive control was provided by the kit while the negative control was the sample buffer. EUROBlotOne (EUROIMMUN, Germany) was used to detect the signal intensity, and the cutoff threshold was above 25.

Muscle pathology
Muscles with moderate muscle weakness (MRC 3/5 or 4/5), or with edema on MRI, or with myopathic changes on electromyography were chosen for biopsy. Muscle biopsies were obtained using standard techniques and were then snap-frozen by immersion in liquid nitrogen-cooled isopentane. Pathological scoring A semiquantitative scoring system called the MIC Pathological Scale (Table 1) was adopted from Lucy Wedderburn [15]. Brie y, severity of pathology was assessed in 3 domains: muscle ber, in ammatory and connective tissue domain. Components of the muscle ber domain used to determine abnormalities were PFA, sarcoplasmic MxA overexpression, assessment of the number of necrotic and regenerating bers within fascicles and in peri-fascicular regions, and presence of internal nuclei. PFA was de ned by the presence of at least one muscle fascicle possessing a cluster of small bers that occupied more than 60% of the bers along the edge of the fascicle [9]. The components of in ammatory domain included the numbers of in ammatory cells in ltration on CD4, CD8, CD20 and CD68 staining, overexpression of MHC class I and the deposition of MAC. Excess brosis was assessed on H&E and MGT stains in perimysial and endomysial regions. The total score includes assessment in 3 domains and the values range from 0-24, with higher scores indicating more severe pathology. All biopsy assessments were performed by two independent observers (Zeng L, Ma X).

Statistical analysis
All statistical analyses were carried out using Prism (Graphpad, La Jolla, CA). Qualitative variables were expressed as percentages and frequencies while quantitative variables were expressed as median, rst and third quartiles. A factorial analysis of variance (ANOVA) was performed using the nonparametric Kruskal-Wallis test to evaluate the main effects of MSA subgroups on clinical and pathological scores. Post-hoc comparisions were conducted using Dunn's Multiple Comparision test to identify pairs of MSA subgroups whose clinical or pathological scores were signi cantly differed from each other. The Chi square test were used to compare categorical ndings between different autoantibodies. For all analyses, values of p < 0.05 were considered signi cant.

Demographic and clinical features
A total of 46 DM patients (anti-MDA5 11, anti-Mi-2 10, anti-NXP2 13, anti-TIF1γ 8, anti-SAE 4) were included ( Table 2). The median age at onset was 48 years and the median time from onset to biopsy was 4 months. Most of the patients showed preferential involvement of proximal muscle. Total MMT scores varied from 30 (severe) to 80 (normal) with the median of 63. The most frequent biopsy site was quadriceps, followed by biceps brachii, gastroenemius, tibialis anterior and deltoid muscles. More than half of the DM patients (52.2%) were on steroids use at the time of biopsy while the median duration from steroids use to biopsy was 2 months (0.875-5.5 months).
The clinical features of different subgroups were compared (Fig. 1). Anti-SAE patients had a later onset compared with the anti-MDA5 group. Interestingly, patients with anti-Mi-2 autoantibodies were more likely to have severe muscle weakness than patients with anti-MDA5, anti-NXP2 and anti-TIF1γ patients (p < 0.05). No signi cant difference was found between other groups. Patients with anti-Mi-2 autoantibody had signi cantly higher levels of creatine kinase (CK) compared with the anti-MDA5, anti-TIF1γ and anti-SAE groups. The anti-TIF1γ group had intermediate CK levels, while anti-MDA5 and anti-SAE groups had normal or slightly elevated serum CK levels. Patients with anti-NXP2 had a wide range of CK levels. Similar trends were observed for the level of lactate dehydrogenase (LDH) and CK/LDH ratio.
MRI of the thigh muscles was performed in 32 patients (anti-MDA5 5, anti-Mi-2 6, anti-NXP2 11, anti-TIF1γ 7, and anti-SAE 3) using a 1.5-T machine (GE; Signal). Representative MRI imaging for DM with anti-Mi-2, anti-MDA5 and anti-SAE are shown in Fig. 2. Patients with anti-Mi-2 exhibited diffuse edema on STIR images, especially the anterior compartments of the thigh. In contrast, fascial edema or patchy distribution were detected in patients with anti-MDA5 and anti-SAE autoantibodies. In the anti-NXP2 and anti-TIF1γ groups, a wide range from foggy distribution to diffuse involvement were observed ( Supplementary Fig. 1).
In the in ammatory domain, high level of in ammation was demonstrated in anti-Mi-2 and anti-SAE group compared with anti-MDA5 group (p < 0.01). The anti-Mi-2 group showed obvious in ltration of CD8 positive T cells (9/10), CD20 positive B cells (4/10) and CD68 positive macrophages (7/10). In anti-SAE group, 3/4 of the samples displayed CD8 positive T cells and CD20 positive B cells in the endomysial tissue ( Supplementary Fig. 2). In contrast, anti-MDA5 group had little or no in ammatory in ltrates. A few CD4 + T cells (3/11) and CD68 + macrophages (1/11) were seen, while CD8 + T cells and CD20 + B cells were not present in these samples. Almost 90% (39/46) samples showed MHC-I upregulation in cytosol or sarcolemmal membrane. Even in anti-MDA5 group, 10 of 11 cases showed increased MHC-I expression. There were also signi cant differences in the MAC deposition among groups. All anti-Mi-2 samples showed MAC deposition on capillaries, and 2 samples even showed sarcolemmal deposition, while only 3 of 11 cases showed capillary MAC deposition in the anti-MDA5 group (Fig. 4). In the connective tissue domain, no signi cant difference between groups were seen.
Overall, patients with anti-Mi-2 demonstrated higher pathological severity scores, both in the myo ber domain and in ammation domain. In contrast, MDA5 associated myopathies generally had minimal myo ber abnormalities with little or no in ammatory in ltration, scattered MxA upregulation and few complement deposition on capillaries.
Patients with SAE had more in ammatory cell in ltration and MAC deposition compared to MDA5 group. NXP2 and TIF1 γ group showed a wide range in muscle pathological scores.

Discussion
In this study we comprehensively compared the pathological features in DM patients with different autoantibodies.
Certain features were relatively common among most DM patients regardless the autoantibody. PFA was found in 52.2% of all DM patients, while sarcoplasmic MxA expression was found in 67.4% of all DM patients. The sensitivity of MxA in our cohort was lower than previous report (67.4% vs 77%) [9], which might be due to more anti-MDA5 case enrollment in our study (24% vs 17.5%) [9]. A signi cant proportion of DM samples showed MHC-I upregulation and complement deposition, 84.8% and 67.4% respectively which were another important pathological features of DM [10].
Notably, the pathological features also showed certain divergence in different subgroups in our cohort. Anti-Mi-2 DM patients had higher prevalence of PFA, ber necrosis and regeneration while anti-MDA5 groups had minimal myo ber abnormalities without classic PFA. One possible mechanism for PFA is the vascular insu ciency, supported by frequent MAC capillary deposition in the perifascicular area [16]. However, we observed not only perifascicular capillary MAC deposition but also sarcolemmal MAC deposition. In addition, we con rmed that Mi-2 patients also had higher number of CD8, CD20 and CD68-positive cell in ltration than anti-MDA5 group, in consistent with previous report [17]. CD68-positive cells could secret a lot of cytokines such as tumor necrosis factor alpha (TNF-) and interleukin-1 beta (IL-1β) [18], which further exacerbate perimysial in ammation. The above ndings, together with higher level of muscle enzymes and edema of thigh, might explain the more severe muscle weakness in anti-Mi-2 patients. Of note, although patients with anti-Mi-2 had more severe muscle weakness, laboratory testing and pathological scores, some research showed anti-Mi-2 patients have signi cantly low risk of interstitial lung diseases and malignancy, the prognosis is favorable with good response to treatment [19,20].
As a distinctive subtype of DM, anti-MDA5 group has been linked with clinical amyopathic DM and progressive interstitial lung disease [21]. In consistent with the mild muscle weakness, low or nearly normal muscle enzymes, anti-MDA5 DM patients generally had minimal myo ber abnormalities with little or no in ammatory in ltration. The MDA5 molecular is one of the viral RNA sensors that indicate IFN1 pathway activation, thereby suppressing virus replication [22]. Interestingly, accumulating pathological evidence suggests that IFN1 signatures are high in serum and affected skin [23,24], but signi cantly lower in muscles in anti-MDA5 DM patients compared to other DM subgroups [25]. In our research, negative or scattered distribution of myo ber MxA expression in MDA5 subgroup further suggested the low activation of IFN1 pathway in muscle tissue, as MxA is a surrogate marker for IFN1 pathway activation. We speculate that selectively recruitment of IFN1 in different tissues of MDA5 patients might contribute to this particular pathological feature.
Different from anti-Mi-2 and anti-MDA5 subgroups, patients with anti-TIF1γ and anti-NXP2 have a wide range in muscle strength, enzyme levels, muscle imaging and histopathological scores. In one study from adult TIF1γ positive patients indicated that capillary MAC deposition is associated with paraneoplastic myositis [26]. However, our data didn't show signi cant difference on C5b-9 immunohistochemical pro le between anti-TIF1γ group and other groups. Some studies suggest anti-NXP2 dermatomyositis tends to be associated with perimysial angiopathy and shows microinfarction and capillary loss in pathology [27,28]. However, in our cohort microinfarction is present in only 4/13 of the patients (data not shown), this may be due to steroid use before biopsy.
This study has several limitations. First, this study involves a relatively small number of patients, especially for rare antibodies such as anti-SAE. Second, although all pathological staining were performed at Huashan hospital, the protocol of clinical data collection and biopsy site were inevitably variable. Nevertheless, half of the patients had already received immunotherapy before biopsy, therefore, the clinical data do not re ect these patients' conditions prior to treatment. Further prospective multicenter study would be helpful in understanding the characteristics among different MSA subgroups. Informed consents to participate were obtained from all the patients.

Consent for publication
Not applicable.

Availability of data and materials
The datasets used and/or analysed during the current study are available from the corresponding author on reasonable request.

Competing interests
The authors declare that they have no competing interests.        Representative pathological ndings for a patient with Mi-2, MDA5 and SAE autoantibodies. Mi-2 group and SAE group typically had high level of myo ber pathology, while MDA5 group had mild appearance with minimal myo ber abnormality and no in ammation. a, b, c -H&E; d,e,f -MxA; g,h,i -MAC; j, k, l -MHC-I. The original magni cation is X100.

Supplementary Files
This is a list of supplementary les associated with this preprint. Click to download.