Myofiber HLA-DR expression is a distinctive biomarker for antisynthetase-associated myopathy

Objectives To assess the value of major histocompatibility complex (MHC) class II antigen (HLA-DR) expression to distinguish anti-synthetase myopathy (ASM) from dermatomyositis (DM). Methods Muscle biopsies from patients with ASM (n = 33), DM without anti-synthetase antibodies (ASAb) (n = 17), and normal muscle biopsy (n = 10) were first reviewed. ASAb included anti-Jo1 (26/33), anti-PL12 (4/33), anti-PL7 (2/33), and anti-EJ (1/33). Immunohistochemistry was performed for MHC-I/HLA-ABC, MHC-II/HLA-DR, membrane attack complex (C5b-9), neural cell adhesion molecule (NCAM)/CD56 expression, and inflammatory cell subsets. Twenty-four ASM and 12 DM patients from another center were added for HLA-DR evaluation. Results Ubiquitous myofiber HLA-ABC expression was equally observed in ASM and DM (93.9% vs 100%, NS). In contrast, myofiber HLA-DR expression was found in 27/33 (81.8%) ASM (anti-Jo1: 23/26, 88.5%; others: 5/7, 71.4%) vs 4/17 (23.5%) DM patients (p < 0.001). HLA-DR was perifascicular in ASM, a pattern not observed in DM. In addition, C5b-9 deposition was observed on sarcolemma of non-necrotic perifascicular fibers in ASM, while, in DM, C5b-9was mainly detected in endomysial capillaries. CD8 cells were more abundant in ASM than in DM (p < 0.05), and electively located in perimysium or in perifascular endomysium. HLA-DR expression correlated positively with the CD8+ cells infiltrates. Strictly similar observations were made in the confirmatory study. Conclusion ASM is characterized by strong myofiber MHC-II/HLA-DR expression with a unique perifascicular pattern, not described so far. HLA-DR detection must be included for routine myopathological diagnosis of inflammatory/dysimmune myopathies. HLA-DR expression in ASM may indicate a specific immune mechanism, possibly involving IFNγ. Electronic supplementary material The online version of this article (doi:10.1186/s40478-014-0154-2) contains supplementary material, which is available to authorized users.


Introduction
Idiopathic inflammatory and dysimmune myopathies (IDM) form a group of acquired conditions characterized by immune-mediated injuries of muscle tissue. Most of them favorably respond to immunomodulatory therapies making necessary their accurate identification at early stages of evolution. On the grounds of histopathological criteria, IDM are usually divided in three main groups [1,2]: (i) polymyositis (PM), (ii) dermatomyositis (DM), and (iii) autoimmune necrotizing myopathies (AINM). From an immunological view, PM are characterized by a cell-mediated autoimmune response directed towards myofibers, as assessed by abnormal ubiquitous MHC class I/HLA-ABC myofiber re-expression and endomysial CD8 T-cells surrounding and invading nonnecrotic fibers. This feature is also observed in inclusions body myositis (IBM), which in turn differs from PM by the presence of chronic myopathic changes, rimmed vacuoles, aggregates and mitochondrial abnormalities. DM is a vasculopathic myopathy with perimysial vascular inflammation, myofiber ischemic lesions, and endomysial microangiopathy with complement activation [3]. AINMs are defined by random myofiber necrosis with complement activation and minimal inflammation [2,4].
Current diagnostic approaches of IDM require the histopathological evaluation of immunological parameters including lymphocyte phenotype, myofiber MHC-I expression and complement activation (C5b9 formation). Myofiber MHC-I/HLA-ABC expression has been extensively studied and was shown to be a very sensitive but poorly sensitive marker of inflammatory myopathy [15][16][17]. In contrast, only a few works evaluated the possible significance of MHC-II/HLA-DR myofiber expression [17][18][19][20][21]. Unlike MHC-I, there is no constitutive expression of MHC-II/HLA-DR molecules by myogenic cells, in vivo or in vitro [22,23]. Myofiber HLA-DR expression was ascribed to the presence of interferon (IFN)γ in the microenvironment of myofibers [19,23,24]. Before delineation of ASM, HLA-DR immunostaining was proposed as a marker of inflammatory myopathy [18,25], but further studies failed to attribute a diagnostic value to the finding [26]. In the present study, we investigated muscle HLA-DR expression in ASM and ASAb-negative DM, and compared it to the expression of other histopathological biomarkers. As opposed to ASAb-negative DM, ASM was characterized by strong myofiber HLA-DR expression, electively localized in perifascicular areas, and typically associated with C5b9 deposition on sarcolemma of non necrotic fibers, and not on capillaries. This characteristic pattern was confirmed in a second series of patients from another center pointing out ASM as a distinct type of IDM.

Patients
We retrospectively studied 50 adult patients (age > 18 years) who underwent muscle biopsy for diagnostic purposes in our institution (Henri Mondor University Hospital) between January 2004 and January 2013. All patients were suspected to have inflammatory myopathy and had comprehensive ASAb testing including anti-Jo1, anti-PL7, anti-PL12, anti-EJ, and anti-OJ autoantibodies. Patients were divided in two groups: Group 1: ASM patients characterized by positive detection of ASAb (ASM, n = 33); Group 2: ASAb-negative DM patients (n = 17). DM diagnosis was based on ENMC 2003 criteria [1]. Ten patients with fibromyalgia and normal muscle biopsy were used as controls. In accordance with current French legislation; all patients gave written consent (Approval #12-009 at CPP Ile-de-France IX).

Quantitative assessment of immunopathological parameters
The severity of myonecrosis process was estimated by determining the percentage of NCAM-positive regenerating fibers. Myofiber expression of HLA-ABC and HLA-DR was defined by sarcolemmal staining associated or not with sarcoplasmic staining. The percentage of HLA-ABC-positive fibers was determined in three representative fascicles. Quantitative evaluation of HLA-DR expression was performed in the most affected fascicle and included (i) the percentage of positive fibers in the whole fascicle, and (ii) the percentage of contiguous positive fibers among the external rim of perifascicular fibers. C5b-9 decoration of myofibers was considered in case of sarcolemmal labeling. In contrast, fibers with marked C5b-9 intracellular immunoreactivities were considered necrotic and not taken into account. Quantitative evaluation of myofibers C5b-9 labeling was performed in the most affected fascicle and included (i) the percentage of positive fibers in the whole fascicle, and (ii) the percentage of positive fibers among the external rim of perifascicular fibers. The amounts of CD3+, CD4+, CD8+, CD20+ and CD68+ cells were scored following the score recently assessed for juvenile DM [29].

Statistical analysis
Statistical analysis was done using R software (www.Rproject.org). Qualitative criteria were compared using Fisher's exact test. In case of non normal distribution, quantitative variables were compared using Mann Whitney U test. To compare more than two groups, the Kruskal-Wallis test was used. Dunn post-hoc test was used for pairwise comparisons. All p-values were twotailed and statistical significance is defined as p < 0.05.

Discussion
From present results, anti-synthetase myopathy (ASM) appears characterized by (i) a strong myofiber HLA-DR expression, electively localized in perifascicular areas, and (ii) complement activation with C5b-9 deposition at the surface of perifascicular myofibers. This peculiar immunopathological pattern has not been described so far, and appears strikingly distinct to that observed in DM, in which myofiber HLA-DR expression is only occasional or absent, and C5b-9 deposition mainly observed in endomysial capillaries. Our results reinforce the view that ASM corresponds to a unique entity resulting from a specific pathophysiological process.
Identifying new distinctive biomarkers for ASM will necessarily contribute to improve the diagnostic approach. The diagnosis of AS depends on the positive detection of detection of circulating antibodies directed towards antiaminoacyl-tRNA synthetase molecules. However, only eight different ASAbs have been identified so far, that is few considering the number of amino-acids. Delineating a characteristic myopathological pattern for ASM could help to identify patients with "seronegative" anti synthetase syndrome and therefore prompt the identification of new antisynthetase antibodies.
Our results showed that similarities between ASM and DM also exist at the histopathological level. Indeed, both fragmentation of perimysial connective and perifascicular atrophy were found as frequent in ASM as in DM (79.3% vs 80.0%, NS: 44.4% vs 53.3%, respectively). HLA-ABC and NCAM were equally expressed in both conditions. Focal myofilament loss and microinfarctus were electively observed in DM and can represent distinctive features. However, the irregular presence of these features(40% and 33% in DM, respectively) dims their helpfulness. In this context, HLA-DR and C5b-9 proved to be more determining biomarkers for diagnostic delineation. Myofiber HLA-DR expression is clearly distinctive between ASM and DM. Distinguishing DM and ASM is not trivial since adult DM is significantly associated with cancer, as in our cohort. Interestingly none of our ASM patients had malignancy, a result in contrast with recent data showing an overall 12% frequency of cancer in patients with ASAbs [13]. Whether the immunopathological pattern we described in ASM may have a prognostic impact deserves further studies. Only a few works previously addressed the issue of myofiber HLA-DR expression in muscle pathology [18,19,21,25,30]. Reports of positive myofiber HLA-DR expression mainly concerned patients classified as "polymyositis" (PM), but without indication about the association (or not) with anti-synthetase antibodies [19,25]. In these cases, HLA-DR-positive fibers were usually few and randomly distributed [19,25]. Interestingly, the lack of myofiber HLA-DR expression in DM was repeatedly observed [19,21,25], but never regarded as a potential distinctive biomarker. The marked myofiber HLA-DR expression and its perifascicular pattern we described in ASM have never been reported so far and indubitably constitute a new feature in myopathology.
The presence of extensive C5b-9 deposition at the surface of perifascicular myofibers in AS has not been previously reported to our knowledge. C5b-9, also termed membrane attack complex (MAC), is the cytolytic Comparison was done between patients with a CD8+ score null and patients with score equal to 1 or 2, ASM and DM patients having been pooled. effector of complement system, produced after activation through either 'classical' antibody-dependent or alternative pathway. MAC formation in muscle pathology remains a puzzling phenomenon observed in various acquired and non-acquired conditions [1]. It is generally admitted that it does not reflect antibody-mediated immune attack [31,32]. Therefore, in ASM, sarcolemmal MAC deposits could reflect the presence of myofiber alterations in relation with the pathological process that also leads to HLA-DR expression. Complement activation may also contribute to myofiber necrosis and thus could constitute an amplifying factor of muscle injuries.
Unlike MHC class I expression, the expression of MHC class II molecules is most often restricted to antigen presenting cells (APCs) (macrophages, cells dendritic cells, B lymphocytes), at the notable exception of endothelial cells that constitutively express MHC class II in vivo [14,33]. MHC class II expression can be induced by IFNγ in various non-APCs, including mesenchymal stromal cells, fibroblasts and epithelial cells [19,34], and by hypoxia in cultured endothelial cells [33]. IFNγ is produced predominantly by natural killer (NK) and natural killer T (NKT) cells as part of the innate immune response, and by CD4 Th1 and CD8 cytotoxic T lymphocyte (CTL) effector T cells once antigen-specific immunity develops [35]. In accordance with these data, we found a quantitative correlation between HLA-DR expression andCD8 cells infiltrates ( Figure 4). Moreover, CD8 cells were most often observed in close vicinity to HLA-DR expressing myofibers, in perifascicular endomysium or perimysium ( Figure 4). HLA-DR expression is under the control of CIITA (Class II transactivator), a cytosolic protein itself controlled by the secretion of IFNγ [36]. In PM, HLA-DR expression was shown associated with that of IFNγ, Ii, CIITA and HLA-DM [19], and it seems sound to consider that HLA-DR expression reflects the presence of IFNγ in myofiber microenvironment. This view is supported by experimental data showing that myogenic cells respond to IFNγ by expressing HLA-DR. However, this effect of IFNγ seems restricted to proliferating and undifferentiated myoblasts, myotubes and innervated contracting muscle cells being unresponsive to IFNγ [20,26]. Whether mature myofiber may respond to IFNγ has not been documented to date, and mechanisms leading to myofiber HLA-DR expression still remain undetermined.

Conclusion
In conclusion, our results confirm that ASM constitutes a specific type of primary inflammatory and dysimmune myopathies, in addition to PM/IBM, DM and AINM. In particular, ASM can be distinguished from other conditions by a strong myofiber MHC-II/HLA-DR expression, which should be henceforth considered as a key immunopathological biomarker for the diagnosis of IDM, in addition to MHC-I/HLA-ABC and C5b-9. HLA-DR expression suggests a role for type 2 IFNγ in ASM, in contrast with DM that relates to type 1 IFNα/ β-mediated immune mechanism [31].