Nailfold capillaries and myositis-specific antibodies in anti–melanoma differentiation–associated gene 5 antibody-positive dermatomyositis

Abstract

Objectives

This study aimed to quantify nailfold capillary (NFC) abnormalities in anti–melanoma differentiation–associated gene 5 (MDA5) -positive DM patients and to evaluate the association with clinical parameters, including serum biomarkers. In addition, we aimed to clarify the period leading to remission of NFC abnormalities during immunosuppressive treatment in patients with DM.

Methods

A prospective observational study was conducted including patients (n = 10) who first visited Hiroshima University Hospital and were diagnosed with DM or clinically amyopathic DM with anti-MDA5 antibodies. We compared the NFC abnormalities detected by nailfold-video capillaroscopy (NVC), physical findings, blood tests, respiratory function tests, and vascular-related growth factors measured using a LEGENDplexTM Multi-Analyte Flow Assay Kit.

Results

NFC abnormalities improved in all patients from 2 to 17 weeks after the initiation of immunosuppressive treatment. The NVC scores were inversely correlated with anti-MDA5 antibody titres at baseline. NVC scores and forced vital capacity were positively correlated. Baseline values of M-CSF and stem cell factor were correlated with anti-MDA-5 titres.

Conclusion

Our study suggested that NVC scores and disease activity were inversely correlated before treatment. Vascular-related growth factors, such as M-CSF and stem cell factor, may be associated with the disease mechanism in patients with anti-MDA5 antibody-positive DM.

Introduction

DM is characterized by cutaneous symptoms and skeletal muscle weakness due to the typical pattern of autoimmune myositis and heterogeneous clinical symptoms. Many myositis-specific autoantibodies have been reported, each of which is strongly associated with particular clinical phenotypes [1]. Anti–melanoma differentiation–associated gene 5 (MDA5) antibody-positive clinically amyopathic DM (CADM) is strongly associated with rapidly progressive respiratory diseases and poor prognoses [2–5]. It has been reported that combination therapy using high-dose steroids, calcineurin inhibitors, and i.v. CYC improved the prognosis of survival in these patients [6]. In real-world practice, the clinical course is heterogeneous, with some cases requiring plasmapheresis, and others improving with monotherapy or dual therapy [7]. To date, few methods have been established to predict the intensity of immunosuppressive therapy that is required in patients with anti-MDA5 antibody-positive DM before treatment. We focused on the importance of assessing vascular abnormalities, as cluster analysis in patients with anti-MDA5 antibody-positive DM extracted a group of severe skin vasculopathy patients with a better prognosis than those with rapidly progressive interstitial lung disease (RP-ILD) [8].

Nailfold video-capillaroscopy (NVC) is a non-invasive technique widely used in autoimmune diseases such as SSc [9]. Abnormalities in nailfold capillaries (NFCs) are also observed in patients with DM and 87.5% of patients with anti-MDA5 antibodies [10]. It has been reported that NFC abnormalities in patients with anti-MDA5 antibody-positive DM are improved by immunosuppressive treatment [11]. Recently, the association of NFC findings with vascular-related growth factors has been reported. Multiple reports demonstrated that NFC findings in SSc patients were correlated with VEGF, suggesting the critical role of these growth factors in the formation of NFC [12, 13]. However, Miossi et al. reported that VEGF levels were higher in patients with DM than in healthy controls, but were not correlated with NFC findings [14]. On the other hand, it has also been reported that DM showed the upregulation of markers related to interferon, endothelial activation, inhibition of angiogenesis, inflammation and leucocyte chemo-attraction, and that many of the disease-specific analytes, including vasculopathy-associated markers, are related to disease activity and decrease to near-normal levels during treatment [15]. In patients with DM, the association between NFC findings and vascular-related growth factors is important but remains unclear.

To the best of our knowledge, there are no reports regarding the short-term evaluation of NFC abnormalities in multiple untreated patients with DM. Therefore, the period leading to remission of NFC abnormalities in patients with anti-MDA5 antibody-positive DM remains unknown. Additionally, there are no reports quantifying NFC abnormalities in patients with anti-MDA5-positive DM and evaluating their association with clinical parameters, including vascular-related growth factors. Therefore, we evaluated the period leading to the remission of NFC abnormalities in patients with anti-MDA5 antibody-positive DM. In addition, we investigated whether NFC abnormalities could predict disease activity in patients with anti-MDA5 antibody-positive DM. Third, in patients with anti-MDA5 antibody-positive DM, we investigated the relationship between NFC abnormalities and vascular-related growth factors and the changes before and after treatment.

Methods

Study design and patients

A prospective observational study was conducted regarding patients who first visited Hiroshima University Hospital from April 2018 to August 2020. All patients were followed up for at least 6 months until February 2021. This study was approved by the clinical ethics committee of Hiroshima University Hospital (approval number: E-1393; approval date: 18 October 2018). We provided adequate information to patients and healthy volunteers in accordance with the Declaration of Helsinki and its subsequent modifications. A signed informed consent was obtained from all participants in this study. All patients were diagnosed with CADM or DM by at least two rheumatologists at the time of hospitalization. Patients who tested positive for anti-MDA5 antibodies diagnosed with DM according to Bohan and Peter’s criteria [16] or diagnosed with CADM based on Sontheimer’s criteria [17] were included in this study. Patients with severe respiratory failure or requiring mechanical ventilation on the day of the first visit were excluded. All patients had interstitial lung disease (ILD), which was diagnosed based on clinical symptoms, physical examination findings, and CT imaging. Respiratory function tests were performed for all patients.

Measurement of autoimmune antibodies and serum biomarkers

Anti-MDA5 antibodies were measured using the MESACUP anti-MDA5 test (Medical & Biological Laboratories, Nagoya, Japan). RF was measured using the LZ test (Eiken Chemical, Tokyo, Japan). Anti-CCP and anti-SS-A antibodies were measured using the STACIA MEBLux test (Medical & Biological Laboratories, Nagoya, Japan). Krebs von den Lungen (KL-6) levels were measured using the Lumipulse test (Fujirebio, Tokyo, Japan). The serum levels of vascular-related growth factors, such as angiopoietin-2, epidermal growth factor (EGF), erythropoietin (EPO), fibroblast growth factor-basic (FGF-basic), G-CSF, GM-CSF, hepatocyte growth factor (HGF), M-CSF, platelet-derived growth factor-AA (PDGF-AA), platelet-derived growth factor-BB (PDGF-BB), stem cell factor (SCF), TGF-α and VEGF were measured using a LEGENDplexTM Multi-Analyte Flow Assay Kit Human Growth Factor Panel (Biolegend, San Diego, USA). All serums were diluted 2-fold, and the assay was performed in a V-bottom plate according to the manufacturer’s protocol. Data acquisition was performed using a Cytoflex flow cytometer (Beckman Coulter, Krefeld, Germany) and analysed using the LEGENDplexTM Data Analysis Software provided as an online cloud-based program (Biolegend). The sensitivity of the assay is provided in parentheses: angiopoietin-2 (25.7 pg/ml), EGF (1.2 pg/ml), EPO (18.0 pg/ml), FGF-basic (39.4 pg/ml), G-CSF (31.6 pg/ml), GM-CSF (9.2 pg/ml), HGF (36.3 pg/ml), M-CSF (7.2 pg/ml), PDGF-AA (25.2 pg/ml), PDGF-BB (6.9 pg/ml), SCF (6.7 pg/ml), TGF-α (36.7 pg/ml) and VEGF (48.0 pg/ml).These tests were conducted on the 10 patients and 10 healthy controls. Fluorescence-encoded beads were analysed using CytoFLEX (Beckman Coulter, Tokyo, Japan).

NVC

NFC abnormalities were evaluated using NVC (Optipix capillaroscopy, Optilia Instruments AB, Sollentuna, Sweden), performed and evaluated by a single rheumatologist. NFC was assessed in eight fingers, except for the thumbs, and a 2 mm range was assessed in each finger. In accordance with previous reports, the following capillaroscopic parameters were recorded: enlarged capillaries (defined as an increase in capillary diameter to >20 µm) , giant capillaries (defined as homogeneously enlarged loops with a diameter of >50 µm) and haemorrhages (characterized as dark masses caused by hemosiderin deposits). No abnormality within the range of observation was scored as 0 points, with 1 point being assigned for up to one-third of the field, 2 points for between one-third and two-thirds, and 3 points for greater than one-third of the field [10, 18–20]. The total NVC scores are the sum of the scores for the eight fingers, which were calculated on a nine-point scale, including the sum of the enlarged capillaries, giant capillaries, and haemorrhages [20].

NVC testing was initially performed at the first visit and subsequently conducted every few weeks. An NVC score of ≤8 points was defined as almost normal. These tests were continued until the NFC abnormalities were almost normalized.

Statistical analysis

We used JMP Pro 15 (SAS Institute Inc., San Diego, CA, USA) and GraphPad Prism 9 (GraphPad Software, San Diego, CA, USA) for statistical analysis. Two-variable data such as clinical data, serum markers, and NVC scores were evaluated using simple linear regression analysis. Changes in serum markers before and after treatment were analysed using the Wilcoxon signed-rank test. The level of significance was set at P < 0.05.

Results

Baseline characteristics of the study participants

Twelve patients with anti-MDA5 antibody-positive DM were identified during the entry period, but two were excluded due to ventilator management at the time of admission. Table 1 summarizes the clinical characteristics of the 10 remaining patients. Five patients were women. The median age was 49 (interquartile range: 36.8–52) years. The time from the onset of clinical symptoms (including skin lesions) to the initiation of treatment was 3.8 (range: 1–7) months. All patients exhibited Gottron’s sign and periungual erythema. Palmar papules, the so-called inverse Gottron’s sign, were found in eight cases. Eight patients exhibited arthritic symptoms. Three cases were positive for RF, and three were positive for anti-CCP antibodies. No patient tested positive for anti-SS-A antibody. No deaths occurred during the observational period. Chest CT revealed characteristic signs of ILD in all 10 patients. All patients were initiated on prednisolone 1.0 mg/kg. All 10 patients were treated with i.v. CYC (IVCY) and concomitantly received tacrolimus, a calcineurin inhibitor. As an additional therapy, plasmapheresis was performed in two patients evaluated as severe in the clinical course (No. 9 and No. 10).

NFC abnormalities were evaluated by NVC and observed in 9 of the 10 cases. NFC abnormalities were almost normalized after immunosuppressive treatment (Fig. 1). Table 1 shows that the NFC abnormalities were normalized within 2–17 weeks of the initiation of treatment. The ninth case exhibited no abnormality at the first visit, but changes in NFC were observed during the clinical course. There were no haemorrhage findings 6 months later, but ramification and loss of capillaries were observed. The NVC score for enlarged capillaries and haemorrhages never exceeded 8 points (Supplementary Fig. S1, available at Rheumatology online).

Comparison between NVC score and clinical parameters

The average NVC score at the first visit was 24.2 (0–43) points. We analysed the relationship between NVC scores and the following parameters associated with the severity of anti-MDA5 antibody-positive DM: serum anti-MDA5 antibody titres, serum ferritin level, forced vital capacity (FVC), and serum KL-6. NVC scores were inversely correlated with anti-MDA 5 antibody titres (P < 0.001) and ferritin (P = 0.021), and positively correlated with FVC (P = 0.015) (Fig. 2).

Comparison between NVC scores and vascular-related growth factors

We measured the following vascular-related growth factors in the above 10 cases and healthy controls: angiopoietin-2, EGF, EPO, FGF-basic, G-CSF, GM-CSF, HGF, M-CSF, PDGF-AA, PDGF-BB, SCF, TGF-α and VEGF. Baseline values of angiopoietin-2, EGF, EPO, G-CSF, GM-CSF, HGF, M-CSF, SCF and VEGF were significantly higher than those of the healthy controls. We then investigated the alterations in these nine biomarkers before and after treatment (Table 2, Fig. 3). Both angiopoietin-2 and EGF levels decreased during the treatment periods.

We also analysed the relationship between these biomarkers and NVC scores or serum anti-MDA5 antibody titres. It was suggested that M-CSF and SCF levels tended to be inversely correlated with NVC scores, but no significant difference was observed (P = 0.053 and P = 0.068, respectively) (Fig. 4A–I). Additionally, there was a positive correlation between anti-MDA5 antibody titres and M-CSF or SCF levels (Fig. 4J–K).

Discussion

In this study, we observed three previously unreported facts in patients with anti-MDA5 antibody-positive DM. First, immunosuppressive therapy showed that NFC abnormalities in patients with anti-MDA5 antibody-positive DM were ameliorated within 2–17 weeks. Second, we found that anti-MDA5 antibody titres and NVC scores were inversely correlated, the FVC and NVC scores were positively correlated, and ferritin levels and NVC scores were inversely correlated. Finally, measurements of vascular-related growth factors before and after treatment revealed that M-CSF and SCF baseline levels were positively correlated with anti-MDA5 antibody titres.

Kubo et al. previously reported that 87.5% of patients with anti MDA5 antibody-positive DM had NFC abnormalities [10], which is compatible with the 90% of those patients (9 out of 10 patients) in our study. It was previously reported that there was an improvement with treatment, but this did not occur in the short term [11]. We have previously reported that in patients with anti-TIF-1γ antibody-positive DM, NFC abnormalities changed in a short period in association with disease activity [20]. Therefore, in our study, we observed and evaluated these changes every few weeks. We confirmed that NFC abnormalities were normalized by treatment within 2–17 weeks in all nine patients. We demonstrated that NFC abnormalities can change within a short period and should be evaluated from the early stages of diagnosis.

Anti-MDA5 antibody titres have been reported to be correlated with the disease activity of RP-ILD associated with DM patients and to predict disease outcomes [21]. Another report highlighted that serum ferritin levels also reflect disease activity [22]. Our study revealed an inverse correlation between anti-MDA5 antibodies and NFC abnormalities at the time of initial diagnosis. We also showed a positive correlation between FVC and NVC scores. Ferritin showed an inverse correlation with the NVC score. These results suggest that NFC abnormalities and disease activity may be inversely correlated. At the same time, low NVC scores may suggest that more appropriate and aggressive therapeutic interventions are needed.

In our study, there was only one case in which oxygenation worsened during treatment, and oxygen administration had to be continued 6 months later. This case is consistent with our findings that NVC scores are inversely correlated with disease activity. Wakura et al. reported that anti-MDA5 antibody-positive DM with severe NFC abnormalities tended to be associated with higher mortality [23]. However, it is not clear from that report whether immunosuppressive treatment in each case was adequately performed. Although that study may be superior to ours in relation to using mortality as an outcome, we believe that there is bias in the results, because it is currently reported that the triple‐drug regimen significantly improves the prognosis [6].

The two cases excluded from our study received ICU treatment such as ventilator management immediately after their first visit, and the NVC examination could not be performed. No obvious NFC abnormalities were observed with the naked eye and dermoscopy at ×10 magnification. A cluster analysis report of 121 patients with MDA5 antibody-positive DM showed that these patients could be divided into three groups: RP-ILD, severe skin vasculopathy, and dermato-rheumatologic patterns (mainly arthritis). The prognosis was worst in the RP-ILD group, best in the dermato-rheumatologic group, and moderate in the severe skin vasculopathy group [8]. Since NFC abnormalities were not evaluated in the study mentioned above, the patients’ prognoses were determined according to the occurrence of RP. We think that it is possible to classify the severe skin vasculopathy group by evaluating NFC abnormalities. We found it useful to perform objectively assessable NVC to separate this group from those patients with poor prognoses.

In addition, we measured 13 vascular-related growth factors. Among them, the patients’ M-CSF and SCF levels before treatment were higher than those of healthy controls, and these markers were positively correlated with anti-MDA5 antibody titres, which correlated with the severity of anti-MDA5 antibody-positive DM. Both M-CSF and SCF tended to be inversely correlated with the NVC score, and these markers were speculated to reflect disease activity. It has been reported that high blood M-CSF levels are observed in anti-MDA-5 antibody-positive DM [24]. It has also been reported that M-CSF may be associated with macrophage activation, and that macrophages present in the interstitium of the lung are strongly associated with M-CSF [25]. To our knowledge, there are no reports to date on the relationship between SCF and anti-MDA-5 antibody-positive DM. SCF is a receptor tyrosine kinase that binds to the c-Kit receptor (CD117) [26]. Activation of c-Kit leads to activation of multiple signalling cascades, including mitogen-activated protein kinase/extracellular signal-related kinases, phosphoinositide 3′-kinase, Src kinase, and Janus kinase/signal transducers and activators of transcription pathways [26]. These results may aid in understanding the link between vascular abnormalities and disease activity.

In addition, our data indicate that both angiopoietin-2 and EGF levels were reduced by treatment and could be valuable biomarkers for DM. It has already been reported that angiopoietin-2 is reduced by treatment in patients with anti MDA5 antibody-positive DM [15]. Hamaguchi et al. noted that serum FGF23 levels were significantly decreased before treatment and improved after treatment, and that NFC abnormalities were therapeutically reversible [11]. They also suggested that serum FGF23 levels reflect the degree of microvascular damage in patients with anti-MDA-5 antibody-positive DM. There were no reports of EGF alterations during the treatment period for DM. In a study of interstitial pneumonia associated with PM/DM, it was reported that IL-15 is a useful marker for finding RP-ILD in addition to anti-MDA5 antibodies and ferritin [27]. Ono et al. reported that type 1 IFN was elevated in patients with anti-MDA5 antibody-positive DM and that this was closely related to skin angiopathy [28]. It has been previously reported that the side effects of IFN treatment include RP and NFC abnormalities [29, 30]. Retinoic acid–inducible gene I (RIG-I) and MDA5, collectively known as RIG-I-like receptors, are key protein sensors of pathogen-associated molecular patterns that induce the expression of type 1 IFN and other pro-inflammatory cytokines during the early stage of viral infection [31]. The phenotype may change depending on the balance between Type-1 IFN (which may cause NFC abnormalities) and GM-CSF and SCF, which are less likely to cause NFC abnormalities. The phenotype, including angiopathy, may change depending on the expression level of each cytokine determined in the signal transduction centred on MDA5.

This study has some limitations. For example, this was a single-centre study involving a small number of patients. It is necessary to increase the size of the patient cohort and reanalyse the data. The observed population does not include patients who died, and thus we may not have been able to accurately assess critically ill patients. Of the 10 treatments examined, high-dose steroids, tacrolimus, and CYC were administered in all cases. Plasmapheresis was performed in two cases. Since this study was not an intervention trial, the treatment regimens did not vary according to the NVC scores. Further verification of these findings will be required in the future.

Conclusion

Our study suggests that NVC scores and disease activity are inversely correlated before treatment. Vascular-related growth factors, such as M-CSF and SCF, may be associated with the disease mechanism in patients with anti-MDA5 antibody–positive DM.

Acknowledgements

We thank all of the patients who participated in our clinical study and the clinicians and staff who contributed to the treatment and care of the patients. We would like to thank Editage (www.editage.com) for English language editing.

Funding: This work was supported in part by the Japan Society for the Promotion of Science (JSPS) KAKENHI (Grant Number 19K08908 to E.S., 19K18499 to S.M., and 19K07940 to S.H.), the Mitsubishi Foundation (S.M.), the Takeda Science Foundation (S.M.), the Mochida Memorial Foundation for Medical and Pharmaceutical Research (S.M.), a Japanese Respiratory Foundation Grant (S.M.), and the Japan Rheumatism Foundation (S.M.).

Disclosure statement: The authors have declared no conflicts of interest.

Data availability statement

The data that support the findings of this study are available from the corresponding author, T.S., upon reasonable request.

Supplementary data

Supplementary data are available at Rheumatology online.

References