Muscle glycome in idiopathic inflammatory myopathies: Impact in IL-6 production and disease prognosis

Summary Idiopathic inflammatory myopathies (IIM) are a group of chronic autoimmune diseases mainly affecting proximal muscles. Absence of meaningful prognostic factors in IIM has hindered new therapies development. Glycans are essential molecules that regulate immunological tolerance and consequently the onset of autoreactive immune response. We showed that muscle biopsies from patients with IIM revealed a deficiency in the glycosylation pathway resulting in loss of branched N-glycans. At diagnosis, this glycosignature predicted disease relapse and treatment refractoriness. Peripheral CD4+ T cells from active-disease patients shown a deficiency in branched N-glycans, linked to increased IL-6 production. Glycan supplementation, restoring homeostatic glycosylation profile, led to a decrease in IL-6 levels. This study highlights the biological and clinical importance of glycosylation in IIM immunopathogenesis, providing a potential mechanism for IL-6 production. This pinpoints muscle glycome as promising biomarker for personalized follow-up and a potential target for new therapies in a patients’ subgroup with an ominous evolution.


INTRODUCTION
Immune-mediated or idiopathic inflammatory myopathies (IIM) are a group of rare systemic autoimmune diseases characterized by proximal skeletal muscle weakness leading to long-term disability, decreased quality of life, and reduced life expectancy. IIM are generally characterized by an increase in muscle enzymes, with or without skin involvement. In addition, other organs may also be affected in a large proportion of patients, namely the gastrointestinal tract, lungs, and heart, often resulting in a poor prognosis. 1 Notwithstanding recent advances in the classification of the disease, 2 clinically the most important subtypes of IIM are dermatomyositis, namely the subgroup of amyopathic dermatomyositis, inclusion body myositis, immune-mediated necrotizing myopathy, and overlap myositis (which includes the subgroup of anti-synthetase syndrome). Polymyositis appears to be an increasingly dying entity. 3 The etiology of IIM is still unknown, but a relationship among environmental triggers, genetic risk factors, and defective immunoregulatory mechanisms appears to play an important role in the pathogenesis of the disease, particularly with an important, albeit unexplored, role for the interleukin (IL)-6 cytokine. 4,5 IIM share many features with other systemic autoimmune diseases such as rheumatoid arthritis and systemic lupus erythematosus. It is a chronic, lifelong, and recurrent disease that primarily affects occupationally active people and is associated with significant morbidity, economic burden, and mortality. 6,7 Although survival in these patients has improved since the widespread use of corticosteroids and immunosuppressants, mortality remains elevated in patients with IIM. The 5-year survival rate has been reported to be 60%, especially in the first year after diagnosis. 8 The increased morbidity is primarily due to severe muscle weakness and visceral involvement. Recent reports suggest that only 20%-40% of treated patients achieve remission, while 60%-80% experience a polycyclic or chronic continuous disease progression, which has a major impact on quality of life at mid-and long-term follow-up, as up to 80% of treated patients remain disabled. 6 One of the notable features of IIM is the presence of myositis-specific antibodies, which have been used to predict the manifestations of IIM. At our center, we have found that the presence of Ro52 antibodies and the overlap with other autoimmune diseases appear to be independent variables associated with the development of damage and refractoriness in IIM and act as predictors of the worst prognosis (data not published). However, there has long been debate about the extent to which the presence of the various autoantibodies commonly associated with IIM may represent accurate biomarkers for the detection of IIM subcategories or behave as epiphenomena of the disease process. 9 On the other hand, about 15%-20% of patients are seronegative, i.e. they do not display significant levels of autoantibodies. Therefore, there is an urgent need in the clinical setting for accurate and reliable biomarkers to improve the classification and diagnosis of IIM subgroups and to obtain significant reliable prognostic information about disease progression, as this lack of diagnostic and prognostic biomarkers hinders major advances in the treatment of these diseases. [10][11][12] Furthermore, the therapeutic tools available for the treatment of IIM are limited. The choice of different immunotherapeutic drugs is mainly empirical, and standard treatment guidelines are not conservative and not disease/subgrouping-directed. The reasons for this include the rarity of IIM, its heterogeneous clinical phenotype and lack of consensual classification criteria, as well as the small number of randomized, double-blind controlled clinical trials. 13 Given these major problems in the clinical and therapeutic management of IIM and the significant medical and societal impact of the disease, there is an urgent need in clinics for the identification and characterization of key molecular parameters/biomarkers that could contribute to the correct diagnosis of the disease subgroup at an early stage of disease progression and, more importantly, to the stratification of patients at increased risk of developing aggressive disease and premature mortality. This knowledge will certainly improve the course of the disease and the rate of therapeutic success, paving the way for the development of novel targeted therapeutic strategies that will ultimately have a positive impact on patient survival and quality of life.
Glycosylation is a major post-translational modification that occurs in essentially all cells. It is characterized by the enzymatic attachment of glycans (sugar chains) to other saccharides, proteins, and lipids, resulting in a diverse and abundant repertoire of glycans on the cell surface collectively known as glycome. Glycans have important biological functions within a cell 14 and are extremely important as master regulators of the immune system. Changes in cellular glycome, including immune cell glycome, can occur in response to environmental and genetic stimuli and are often associated with the acquisition of altered cellular phenotypes such as malignancy and chronic inflammation. [14][15][16][17] In particular, we and others have demonstrated the importance of glycans as regulators of various immunological mechanisms, including adaptive immunity, especially involving T cells. 18,19 Glycans have been reported to control the rate of T cell receptor (TCR) endocytosis and modulate cellular responses through interaction with lectins (such as galectins) and their ligands. 20 A reduction in the branching of N-glycans (in the context of a malfunction/deficiency of N-acetlyglycosaminil transferase V (GnT-V) glycosyltransferase) on T cells has been shown to lead to an increased clustering of TCRs and consequently a lower threshold for T cell activation. 16 This was demonstrated by our group in patients with ulcerative colitis, in whom a deficiency of branched glycans on intestinal T cells was associated with T cell hyperactivity and increased disease severity. 19 In addition, GnT-V-deficient mice were shown to have an enhanced delayed-type hypersensitivity response and increased susceptibility to experimental autoimmune encephalitis (EAE) 18 and inflammatory bowel disease (IBD). 21 Treatment of these mice with high concentrations of N-acetylglucosamine (GlcNAc) increased GnT-V-mediated N-glycan branching and inhibited TCR activation and autoimmune responses in mouse models of EAE, IBD, and type 1 diabetes. Furthermore, there is evidence for a role of glycosylation in immune cell differentiation, e.g. in the generation and phenotypic profiles of T helper (Th) cell subsets, particularly activated CD4 + effector T cells (which differentiate into TH1, TH2, TH17, and regulatory T cells), in part through its effect on the production of ligands of various lectins. In particular, N-glycan branching has been shown to play an important role in autoimmunity by promoting the differentiation of anti-inflammatory TH2 cells over pro-inflammatory TH17 cells. 22 In addition to the role of glycans as direct regulators of immune cell function, we have also shown that alterations in epithelial cell glycosylation have an impact on the loss of self-tolerance. We have shown that  In the present work, we aimed to characterize the clinical and biological impact of muscle glycome in the immunopathogenesis of IIM, in particular how glycosylation changes at peripheral immune cells regulate the systemic immune response and IIM prognosis. We demonstrated for the first time a clear truncation of the branching N-glycosylation pathway in muscle tissue of patients with IIM, which was associated with a poor prognosis. Importantly, we observed that this glycosylation alteration could be detected in the periphery, where circulating CD4 + T cells exhibit deficient N-glycosylation associated with IL-6 production. This pathogenic immunophenotype was reversed by metabolic supplementation with glycans from fresh muscle biopsies.
This study contributes to the pathophysiological knowledge of IIM, focusing on epigenetic alterations that may contribute to the pathogenesis evolution and prognosis of this group of rare and heterogeneous diseases. Targeted and more effective therapy relies on a deeper knowledge of the pathogenesis implied in etiology, evolution, and expected prognosis.

RESULTS
IIM muscle biopsies display an altered N-glycosylation profile associated with poor disease course and non-response to therapy Taking in consideration the natural progression of N-glycans from high-mannose precursors toward more complex and branched structures ( Figure 1A), we aimed to obtain a simple glycosignature from the IIM skeletal muscle based on the levels of these two distant N-glycan traits. To do so, histochemistry of GNA and L-PHA lectins were used, to assess the relative levels of high-mannose and b1,6-branched N-glycans (respectively) in a cohort of 22 IIM muscle tissue (IIM) and 11 healthy muscle tissue (HC). The demographic and clinical features of patients with IIM are included in the Table S1. The results showed an overall increased reactivity of GNA in muscle biopsies from patients with IIM, compared with HC, together with a decreased L-PHA staining ( Figure 1B), despite not statistically significant for L-PHA (Figure 1C). Interestingly, we observed that the major GNA reactivity was detected in the stromal component of the tissue, with less reactivity in the muscle fibers ( Figures 1D and 1E). This tissue glycoprofile was further confirmed at transcriptional level, by determining the expression of glycogenes involved in the N-glycosylation pathway, such as MGAT1 and MAN2A1 ( Figure 1A). IIM biopsy cells seem to be significantly deficient in the expression of MGAT1 ( Figure 1F), a gene that encodes for N-acetylglucosaminyltransferase-I (GnT-I), a key enzyme that adds the first GlcNAc antenna, converting high-mannose into hybrid N-glycan structure. Surprisingly, MAN2A1, the gene that encodes for a-mannosidase (an enzyme that acts immediately into the GnT-I product) was increased ( Figure 1F), which may be explained by a compensatory mechanism to recover the progression in the N-glycosylation pathway and maintain the homeostatic glycan repertoire.
Given our previous knowledge that alterations in the tissues glycome might be associated with pathogenesis and prognosis of the disease, 23,24 we explored, at a longitudinal level, the predictive capacity of GNA reactivity iScience Article at diagnosis to the disease outcome. We have observed that high GNA reactivity in the stromal compartment (staining for more than 50% of the cells) at diagnosis was able to stratify patients according to the risk of having more than 1 flare (poor prognosis), with a specificity of 90% and sensitivity of 83.3% ( Figure 1G). The univariate analysis also demonstrated that high GNA reactivity at diagnosis increased 45-times the odds (p value = 0.004) of patients with IIM to have more than 1 flare at 3 years of follow-up. Interestingly, multivariate analysis revealed that among other clinical parameters used to monitor the disease progression, high GNA reactivity detected at muscle biopsy was the only variable that predicts the development of more than 1 flare in an independent way ( Figure 1H). Moreover, high GNA reactivity was also able to identify refractory patients (defined as those that do not respond to 2 immunosuppressive drugs given, concomitantly or subsequently, for at least 3 months) with a specificity of 71.4% and sensitivity of 87.5% ( Figure 1I). The univariate analysis showed that high GNA reactivity at diagnosis increased 18-times the odds (p value = 0.004) of patients with IIM to be non-responders at 3 years of follow-up. Multivariate analysis revealed that among other clinical parameters used to monitor the disease, high GNA reactivity at muscle biopsy was the only independent variable that predicts the non-response to treatment ( Figure 1J).

Deficiency in complex type N-glycans at the surface of peripheral immune cells is associated with pro-inflammatory systemic profile in patients with IIM
Taking in consideration that stromal glycoprofile has shown to be more significantly affected with glycosylation alterations ( Figure 1E), we further analyzed the glycoprofile of immune cell population at the periphery, exploring a possible less invasive proxy of the in situ glycobiomarker. We have found that CD4 + T cells display significantly lower levels of surface b1,6-branched N-glycans, concomitantly with significantly increased levels of high-mannose N-glycans, detected by L-PHA and GNA reactivity, respectively (Figures 2A and S1A). The transcriptional profile of isolated peripheral CD3 + T cells showed no differences in MGAT1 gene expression ( Figure 2B), but revealed a deficient expression of the gene MAN2A1, which may result in the lower hydrolysis of one of the mannose branches precluding the addition of a second GlcNAc antennae and thus leading to the accumulation of less complex N-glycans (Figure 2C). Given the previous evidence showing that loss of b1,6-branched N-glycans in CD4 + T cells has an impact on the TCR signaling and T cell function, 16,18,21 we analyzed the impact of this glycosylation alteration in CD4 + T cell-producing cytokines, namely IL-6 (one of the key cytokines in the etiopathogenesis of IIM). To do so, we stratified CD4 + T cells into L-PHA high vs. L-PHA low. Interestingly, we have observed that L-PHA low subpopulation showed a significant increase in these IL-6-producing cells in IIM ( Figures 2D and S1B). Overall, and as expected, CD4 + T cells from patients with IIM are more prone to produce IL-6, but no differences were found for other cytokines, such as IL-4 ( Figures 2E and 2F). Accordingly, serum IL-6 levels from our cohort were shown to be significantly increased in active patients with IIM compared to controls ( Figure 2G). Moreover, and to show the dependency of TCR glycosylation in defining the threshold for activation, we have cultured IIM and HC peripheral blood mononuclear cells in the presence of anti-CD3 antibody and checked for the activation and cytokine production of CD4 + T cells. As expected, CD4 + T cells from patients with IIM, displaying lower levels of b1,6-branched N-glycans ( Figure 2H), revealed increased levels of TCR-dependent activation ( Figure 2I) and increased IL-6 production ( Figure 2J). iScience Article Furthermore, we also observed that non-T cells (CD3 À ) from patients with IIM may also contribute for the elevated levels of IL-6 in these patients ( Figure 2K), which may include dendritic cells and monocytes, among others. Interestingly, dendritic cells from patients with IIM seem to express higher levels of DC-SIGN, a glycan receptor that recognizes high-mannose N-glycans ( Figure 2L), implicating a role for innate immunity in the hyperactive environment of IIM.
Ex vivo glycosylation reprograming of muscle tissue with glycans supplementation recovers branching N-glycans composition and hampers IL-6 production in situ In order to repair the deficiency in the N-glycosylation pathway observed in Figure 1, we have promoted the progression toward branching N-glycosylation and thus restoring the levels of branching N-glycans. Fresh muscle biopsies from patients with IIM were collected and supplemented ex vivo with GlcNAc ( Figure 3A). This increases the substrate availability to GnTs and consequently promote the expression of b1,6-branching N-glycans on tissue, as previously done by us. 21,25 We have observed a significant increase in the overall levels of b1,6-branching N-glycans in tissue glycoproteins ( Figures 3B and S2A) which was not significant for high-mannose N-glycans ( Figures 3C and S2A). We showed that in fact, muscle T cells are a target of glycosylation modification, such as complex branched N-glycans ( Figures S2B and S2C). Importantly, the increase in b1,6-branching N-glycans resulted in a significant reduction of IL-6 production together with an increase of IL-4 cytokine ( Figure 3C).

DISCUSSION
Muscle cells surface is enriched in glycans and glycoproteins (the glycocalyx) which role in muscle homeostasis and function remains largely unexplored. In fact, changes in glycosylation are a hallmark among many diseases including cancer, chronic inflammation, and autoimmune diseases. 15,17,23,26 However, whether and how the muscle glycome has a clinical and biological impact in the immunopathogenesis of IIM remains completely unknown. Here, we demonstrated that in fact muscle biopsies from patients with IIM exhibit an abnormal glycosylation profile characterized by a decreased expression of iScience Article complex branching N-glycans along with an increased expression of mannose-enriched glycoproteins, in comparison with normal controls. These alterations in glycosylation signature appear to be, predominantly observed, in the stroma compartment and inflammatory infiltrate. However, and in order to precisely analyze the structural glycosylation profile of infiltrating muscle T cells, mass spectrometry analysis should be performed in future allowing a better structural insight of the muscle glycome and correlation with the pathogenesis of muscle diseases. Importantly from the clinical standpoint, we showed that this specific glycosignature was found to have prognostic value in patients with IIM. We demonstrated that levels of high-mannose N-glycan content in stroma of muscle biopsies, detected at diagnosis, are correlated with a poor disease course (more flares) and non-response to therapy. High GNA reactivity in muscle biopsy, at diagnosis, proved to be an independent variable in predicting the development of more than 1 flare, as well as in identification of refractory patients, with a good sensitivity and specificity, when compared with currently available tools/markers. These tissue-specific glycosylation alterations have been previously reported in other autoimmune diseases, namely in kidney from lupus nephritis, 23,25 synovial fibroblasts from rheumatoid arthritis, 27 as well as fucosylation in IBDs 28 and sialylation in GNE myopathy. 29 At the periphery, our findings in IIM revealed that CD4 + T cells from peripheral blood of patients with IIM display the same truncation of the N-glycosylation pathway as observed in the tissue, suggesting that CD4 + T glycoprofile can be a proxy for what is happening at the muscle tissue, potentially constituting a promising non-invasive biomarker in IIM. Interestingly, we found that peripheral CD4 + T cells with the lower levels of L-PHA (from patients with IIM) seem to present lower TCR thresholds, with increased activation markers and IL-6 production. Accordingly, in 2009, Okayama and colleagues published an elegant study refereeing to the critical role of IL-6 in a murine model of myositis. 5 They found increased IL-6 mRNA levels in the mice muscles (particularly in macrophages) which correlated with the histological severity of myositis. IL-6 was proved to be essential for the development of myositis, in this model, and its blockage suppressed the incidence and severity of myositis. No further studies developing this mechanism were published yet, but IL-6 blockade has been effectively used in some patients with IIM. 30,31 Further studies, using larger cohorts, are needed to understand if this peripheral CD4 + T cellspecific glycosignature may be associated with clinical variables that allow a better monitorization of the patients with less invasive techniques.
The origin for this glycan alteration in the immune compartment of affected muscle in patients with IIM still needs further investigation. However, some mechanisms have been proposed by others and us to explain the link between glycosylation changes and T cells hyperactivity and a pro-inflammatory immune response. 16 In fact, b1,6-branched N-glycans were shown to be essential regulator of TCR thresholds in CD4 + T cells in both IBD 19 and multiple sclerosis. 18 In IBD, we showed that a deficiency in branched N-glycans in mucosa T cells from patients with ulcerative colitis was associated with disease severity and T cell hyperactivity. 21 Additionally, we also found that innate immune cells from patients with IIM display an increased capacity to recognize tissue glycosylation alterations as we have observed an increased expression of DC-SIGN in innate immune cells that specifically recognize mannose structures. This observation may underlie the impact of muscle glycosylation alteration in innate immune recognition and consequent potentiation of the pro-inflammatory immune response, as we have observed in lupus, 25 an issue that deserves further investigation.
Therapeutic strategies in immune-mediated systemic diseases, historically and presently, rely on immunosuppression, to reduce the hyperreactivity underlying clinical symptoms and organ dysfunction. Side effects, particularly infections (the main cause of flares in IIM), are the dark face of this approach. Theoretically, restoring immune homeostasis is much more appealing and a more natural way to treat the fundamental problem of these patients, but much harder to achieve. Based in our findings, we propose the glycosylation reprogramming of the muscle as an attractive strategy to recover a homeostatic glycome toward the natural progression of branching N-glycosylation with consequences in immune homeostasis by hampering IL-6 production. Overall, glycosylation remodeling by metabolic supplementation with glycans may constitute an appealing therapeutic strategy to attenuate the pro-inflammatory immune response in IIM and might represent a novel target for directed therapy, eventually associated with immunosuppression, in a subset of patients with a hazardous evolution and altered glycosignature. Despite the promising impact of this study in identifying a mechanism that underlies IIM immunopathogenesis, one of the limitations is the retrospective nature of the analysis of the clinical parameters defining poor prognosis. A prospective cohort to validate the findings will be worth. The number of muscle samples from controls was low (comparing to patients), although enough for statistical analysis. The limited sample size of patients with IIM and subgroups precluded some statistical significance supporting the need to validate the mechanism in a larger and prospective cohort of IIM samples and controls.

STAR+METHODS
Detailed methods are provided in the online version of this paper and include the following: Human IIM muscle explant supplementation A fragment of each IIM muscle biopsy was cut longitudinally into two similar sections. One of the sections was incubated in 200 mL of complete RPMI supplemented with 150 mM GlcNAc at 37 C and 5% CO 2 for 72 h. The other part of the biopsy was incubated in the same medium. The other section of the biopsy was incubated under the same conditions but without the addition of GlcNAc. At the end of the incubation period, the supernatants were collected for cytokine analysis and the biopsy parts were weighed. Protein lysates were obtained by dissociating the tissue with a pestle in lysis buffer (50 mM TrisHCl, pH 7.4, 150 mM NaCl, 1 mM EDTA, 1% Nonidet P-40) supplemented with protease and phosphatase inhibitors, 100 mm PMSF, 100 mM Sodium ortovanadate and Complete Protease Inhibitor Cocktail (Roche).

ELISA and CBA
IL-6 concentration in sera from IIM and healthy donors (mouse ELISA Ready-SET -Go! kits from eBioscience) was measured by ELISA according to the manufacturer's protocol. TMB (3,3',5,5'-tetramethylbenzidine) chromogen solution (eBioscience) was used as substrate and 2N H 2 SO 4 as stop solution. Absorbance was measured at 450 nm and 570 nm using a microplate reader (Biotek Instruments).
For the supplemented human muscle biopsies, cytokine concentrations were analysed by flow cytometry using cytometric bead arrays: the BD Cytometric Bead Array Human Th1/Th2/Th17 Kit (BD Bioscience), according to the manufacturer's instructions. Samples were measured on the BD Accuri C6 instrument (BD Biosciences, US) using a specific template provided by BD Biosciences.

QUANTIFICATION AND STATISTICAL ANALYSIS
Data visualisation and statistical analyses (non-parametric Mann-Whitney t-test) were performed using GraphPad Prism 9 software.
The predictive power of GNA reactivity to discriminate patients with poor disease progression from those with good disease progression was determined by plotting receiver operating characteristic curves (ROC) and calculating the area under the curve (AUC). Adaptability was assessed using the Hosmer-Lemeshow statistic and test. Results are presented in the form of odds ratios (ORs) for each category compared to a predefined reference category and their respective 95% confidence intervals (CIs). Odds ratios above one or below one are indicative of a higher or lower probability, respectively, of developing a poor disease outcome compared to a reference category.
Datapoints were tested for outliers using ROUT testing and excluded when identified.