A preliminary study of frequency and clinical relevance of cytotoxic peripheral CD4 and CD8 T cells in patients with anti-MDA5 positive dermatomyositis

Study Subjects

Peripheral blood samples were collected from 19 MDA5⁺DM-ILD patients who were admitted to the Department of Rheumatology in Renji Hospital from September 2021 to March 2022, and 19 age- and sex-matched HDs from the Renji Hospital Biobank. All the patients fulfilled the 2017 EULAR/ACR Classification Criteria for Idiopathic Inflammatory Myopathies. Detailed information about the patients is listed in Table 1.

Clinical Parameters

Serum, lactate dehydrogenase (LDH), ferritin, anti-MDA5, and anti-Ro-52, as well as neutrophil-to-lymphocyte ratio (NLR), were obtained from the clinical laboratory. Anti-MDA5 and anti-Ro-52 autoantibodies were detected by immunoblot testing following the manufacturer's instructions (Euroimmun, Lubeck, Germany). The gray-scale value of the antibody band was scanned to obtain semi-quantitative results, and gray-scale values of 0–10 were defined in the following way: 11–25 as +, 26–50 as ++, and >50 as +++. Myositis Intention-to-Treat Index (MITAX)^[12] is a disease activity score described by the International Myositis Assessment & Clinical Studies Group with 7 domains (cutaneous, muscle, constitutional, skeletal, gastrointestinal, pulmonary, and cardiovascular). The MITAX score was the sum score divided by the total score of 63 points, and the details of clinical information required for the calculation were obtained from the medical records. We defined a MITAX score of 0.14 (9/63) or higher as disease “active” and all 19 MDA5⁺DM patients recruited in this study were classified as “active” patients (Table 1).

Flow Cytometry

Peripheral blood mononuclear cells (PBMCs) were isolated by density gradient centrifugation on Lymphoprep (Axis-shield). The following antibodies were used in flow cytometry: BUV395 anti-CD3 (clone SK7, BD, San Diego, CA), Brilliant Violet 570 anti-CD4 (clone RPA-T4, Biolegend, San Diego, CA), Alexa Fluor 700 anti-CD8a (clone HIT8a, Biolegend, San Diego, CA), PE anti-Granzyme B (clone GB11, BD, San Diego, CA), Alexa Fluor 647 anti-Granzyme K (clone G3H69, BD, San Diego, CA), and FITC anti-Perforin (clone B-D48, Biolegend, San Diego, CA).

PBMCs were first stained with Live/Dead dye FVS780 (BD, San Diego, CA) in PBS at room temperature for 30 min. Then, cells were stained with surface markers CD3, CD4, and CD8a in FACS buffer (1× PBS containing 1%FBS, 2.5 mM EDTA, and 0.1% NaN₃) at room temperature for 15 min, followed by washing twice with FACS buffer. The cells were then fixed and subject to permeabilization with FOXP3 Fix/Perm buffer and FOXP3 perm buffer (eBioscience, San Diego, CA) at 4°C for 60 min. After this, they were washed twice with Foxp3 Perm buffer (eBioscience, San Diego, CA), and the intercellular Granzyme B, Granzyme K, and Perforin were stained with corresponding fluorochrome-conjugated antibodies at 4°C for 30 min in Foxp3 Perm buffer. Finally, the cells were washed twice with Foxp3 Perm buffer, filtered, and resuspended in FACS buffer before acquisition. Data were collected using a flow cytometry analyzer (LSRFortessa, BD, San Diego, CA) and analyzed using FlowJo Software (v.10.4, TreeStar, Ashland, OR).

Statistical Analysis

Statistical analysis was performed using Prism 8.0.2 (GraphPad, San Diego, CA). Student's t test or Mann–Whitney test was used to compare the two groups as needed. Similarly, Spearman or Pearson's correlation coefficient was used to evaluate the correlation between the two groups as appropriate. A P value <0.05 was considered statistically significant.

Characteristics of the Study Population

A total of 19 patients diagnosed with MDA5⁺DM and 19 healthy controls (HDs) were included in the study. All of these MDA5⁺DM patients were combined with ILD, and treated with 1–2 mg/kg methylprednisolone and CNIs. Five of them died within 6 months because of RP-ILD. There was no significant difference in age, serum ferritin levels, and NLR between the Alive and Dead groups (Table 1). However, the serum levels of LDH in the Dead group were significantly higher than those in the Alive group (Dead vs. Alive: 423 ± 112 U/L vs. 299 ± 10 U/L, P = 0.0339). Similarly, the MITAX scores were also significantly increased in the Dead group (Dead vs. Alive: 0.46 ± 0.03 vs. 0.29 ± 0.12, P = 0.0079) (Table 1). These results indicate that the Dead group shows a more severe disease status compared to the Alive group in MDA5⁺DM patients.

Altered Frequency of Cytotoxic CD4 and CD8 T Cell Subsets in MDA5⁺DM Patients

A decrease of T cell counts^[13,14] and a strong activation of CD8 T cells^[10] have been reported in MDA5⁺DM patients. However, to our knowledge, the expression of cytotoxic molecules in T cells has not been explored in these patients. To this end, we developed a flow cytometric panel to measure key cytotoxic molecules Perforin, Granzyme K (GZMK), and Granzyme B (GZMB) in peripheral CD4 and CD8 T cells (Figure 1). We first observed that the frequency of CD4 T cells was significantly increased in MDA5⁺DM patients compared with that in HDs (MDA5⁺DM vs. HDs: mean±SD [same for the following], 65.59 ± 14.31% vs. 53.85 ± 12.09%, P = 0.0215) (Figure 2). The frequency of CD8 T cells showed a reverse tendency, but did not achieve significance (Figure 2). These results are largely consistent with the early reports indicating a decrease of CD8 T cells in MDA5⁺DM patients.^[13,14]

Figure 1

Gating strategy for flow cytometry analysis of peripheral cytotoxic CD4 and CD8 T cell subsets from HDs and MDA5⁺DM patients. DM, dermatomyositis; HDs, healthy donors; MDA5, melanoma differentiation-associated gene 5.

Figure 2

Comparison of the frequency of CD4 and CD8 T cells between HD and MDA5⁺DM groups. Statistical difference between HD and MDA5⁺DM groups was analyzed using Mann–Whitney test. ^*P < 0.05. DM, dermatomyositis; HD, healthy donor; MDA5, melanoma differentiation-associated gene 5; ns: not significant.

Next we compared the frequency of Perforin⁺, GZMK⁺GZMB⁻, GZMK⁺GZMB⁺, and GZMK⁻GZMB⁺ CD4 T cells or CD8 T cells between HD and MDA5⁺DM groups. We found that compared to HDs, the frequency of GZMK⁺GZMB⁻ CD4 T cells was significantly increased in MDA5⁺DM patients (HDs vs. MDA5⁺DM: 0.45 ± 0.28% vs. 1.34 ± 0.94%, P = 0.0202), and the frequency of GZMK⁻GZMB⁺ CD4 T cells was significantly decreased in MDA5⁺DM patients (HDs vs. MDA5⁺DM: 6.30 ± 3.93% vs. 3.12 ± 3.35%, P = 0.0093) (Figure 3A). Other CD4 T cell subsets showed no statistical difference between the two groups (Figure 3A). More differences were detected in CD8 T cells. In contrast to CD4 T cells, the frequency of GZMK⁺GZMB⁻ CD8 T cells was significantly lower in MDA5⁺DM patients than that in HDs (HDs vs. MDA5⁺DM: 6.53 ± 5.92% vs. 1.63 ± 1.13%, P = 0.0038) (Figure 3B). We also performed the correlation analyses among cytotoxic T cell subsets in MDA5⁺DM patients and HDs, and in general, we saw high correlations between cytotoxic CD4 and CD8 T cell subsets for both groups (Figures 3C and 3D). Collectively, the above results provide a comprehensive picture of cytotoxic molecule expression in CD4 and CD8 T cells from active MDA5⁺DM patients with reference to the data from HDs.

Figure 3

Comparison of the frequency of cytotoxic CD4 and CD8 T cell subsets between HD and MDA5⁺DM groups. The frequency of Perforin⁺, GZMK⁺GZMB⁻, GZMK⁺GZMB⁺, and GZMK⁻GZMB⁺ CD4 T cells (A) or CD8 T cells (B) was compared between HD (n = 19) and MDA5⁺DM (n = 19) groups. Statistical difference between the two groups was analyzed using Mann–Whitney test. (C and D) Heatmaps showing the correlations among cytotoxic CD4 and CD8 T cell subsets from MDA5⁺DM patients (C) and HDs (D). The numbers inside the heatmaps denote correlation coefficients. Spearman correlation analysis was applied. ^*P < 0.05; ^**P < 0.01. DM, dermatomyositis; GZMB, Granzyme B; GZMK, Granzyme K; HDs, healthy donors; MDA5, melanoma differentiation-associated gene 5; ns: not significant.

Association of Cytotoxic T Cell Subsets with Clinical Features in Active MDA5⁺DM Patients

Finally, we evaluated the potential relationship between the frequency of cytotoxic T cell subsets and clinical parameters in MDA5⁺DM patients, including LDH, ferritin, NLR, MITAX, and prognosis. The data showed that the frequency of GZMK⁺GZMB⁻ CD8 T cells had a significantly positive correlation with serum ferritin levels in MDA5⁺DM patients (R = 0.5143, P = 0.0243) (Figure 4A). Notably, when the MDA5⁺DM patients were divided into Alive and Dead groups, the frequency of both GZMK⁺GZMB⁻ CD4 and GZMK+GZMB⁻CD8 T cells was significantly higher in the Dead group than in the Alive group (Dead vs. Alive: 1.8 ± 0.7% vs. 0.8 ± 0.7%; 2.4 ± 1.3% vs. 1.2 ± 0.85%, respectively) (Figure 4B). Altogether, our results suggest that both GZMK⁺GZMB⁻ CD4 and GZMK⁺GZMB⁻ CD8 T cells may have prognostic roles in active MDA5⁺DM patients.

Figure 4

Association of cytotoxic CD4 and CD8 T cell subsets with clinical parameters in MDA5⁺DM patients. (A) Correlation analysis between serum ferritin levels and the frequency of GZMK⁺DM patients (n = 19). Pearson correlation analysis was applied. (B) Comparison of the frequency of cytotoxic CD4 and CD8 T cells between the Alive (n = 14) and Dead (n = 5) groups in MDA5⁺DM patients (n = 19). Statistical difference between the two groups was analyzed using Student's t test. ^*P < 0.05. DM, dermatomyositis; GZMB, Granzyme B; GZMK, Granzyme K; MDA5, melanoma differentiation-associated gene 5; ns: not significant.

Limitations

We are aware that our current study has several limitations. First, the number of samples was relatively small and we only collected the blood samples at baseline in active MDA5⁺DM patients. Second, we did not include a cohort of “stable” MDA5⁺DM patients at remission. Third, we only detected the phenotypes of cytotoxic T cells without functional assays. We plan to address these limitations in our future exploration by recruiting a prospective cohort with follow-up samples.