Patient characteristics

Table 1 presents the patient characteristics. We evaluated 18 and 19 patients showing DM-ILD and IPF, respectively. The DM-ILD group included younger patients (p = 0.001) and more females (p = 0.001) than the IPF group. In the DM-ILD group (n = 18), six, one, and eleven cases involved autopsy, lung transplant, and SLB, respectively. Nine (50%) of the 18 DM-ILD cases showed seropositivity for the anti-MDA5 antibody, while six and three cases involved autopsy and SLB, respectively. Thus, all six autopsy cases showed seropositivity for anti-MDA5 antibodies. Seven of the nine anti-MDA5 antibody-seropositive DM-ILD cases showed a clinical course of RP-ILD. Moreover, the lung tissues of all nine seropositive DM-ILD cases showed a diffuse alveolar damage (DAD) pattern. Among the nine cases of anti-MDA5 antibody-seronegative DM-ILD were positive for anti-Aminoacyl-tRNA Synthetase antibodies. Additionally, two, one, and six cases showed DAD, usual interstitial pneumonia (UIP), or nonspecific interstitial pneumonia (NSIP) patterns, respectively.

Table 1 Patient characteristics

In the IPF group (control, n = 19), nine (47%) patients experienced acute exacerbations (AEs) and were histopathologically diagnosed as showing DAD patterns at autopsy. The remaining 10 (53%) patients showed a chronic course and were diagnosed as showing a UIP pattern through SLB.

All seven cases (100%, six autopsies and one lung transplant) and eight of nine cases (89%) in the DM-ILD and IPF groups, respectively, were treated with corticosteroids. Additionally, calcineurin inhibitors and cyclophosphamide were used in six and five DM-ILD cases and in two and three IPF autopsies, respectively (Table 1). Tests for significance for these differences were not calculated as the usage of steroids and immunosuppressive agents in DM-ILD and IPF groups were different.

Among the 15 autopsies (6 DM-ILD and 9 IPF cases), 3 were of patients in the IPF group who died from lung cancer. The remaining 12 patients died due to progressive respiratory failure, including acute exacerbation of interstitial lung disease. Additionally, one patient in the DM-ILD group underwent lung transplantation because of progressive respiratory failure.

MDA5 overexpression in lungs of patients with DM-ILD and IPF

Immunohistochemistry analysis revealed moderate (n = 15) or strong (n = 2) MDA5 expression in the lungs of 17/18 DM-ILD patients and weak expression in one patient. No correlations were observed among the histopathological pattern, tissue collection method, and MDA5 expression. Notably, all 19 IPF cases showed moderate (n = 14) or strong (n = 5) MDA5 expression in the lungs. Strong MDA5 expression was observed in the lungs of both DM-ILD and IPF groups (Fig. 2b, d). However, no significant different between two group was observed in the expression intensity (p = 0.156) (Table 2).

Fig. 2

Histopathological findings in the lungs of patients with DM-ILD and idiopathic pulmonary fibrosis (IPF). A Hematoxylin & eosin (H&E) staining of a patient with DM-ILD. This patient showed an exudative diffuse alveolar damage (DAD) pattern. B IHC staining with anti-MDA5 monoclonal antibody (mAb) (clone H27) in a patient with DM-ILD. MDA5 is extensively and strongly expressed in the alveolar epithelium. C H&E staining in a patient with IPF. This patient showed an exudative DAD pattern, as well as a usual interstitial pneumonia (UIP) pattern in the background. D IHC staining with anti-MDA5 mAb (clone H27) in a patient with IPF. There was strong MDA expression in the alveolar epithelium

Table 2 Evaluation of immunohistochemical stainingStrong expression of complement and ig in the lungs of DM-ILD but not IPF patients

Figure 3 shows strong expression of complement C3c and IgG in the lungs of DM-ILD patients. Immunohistochemistry staining for complement C3c showed positive results in 15/18 cases, of which eight cases showed moderate or greater intensity. All 18 cases showed positive results for IgG, while 11 cases showed positive results for IgM. Interestingly, 13/18 cases showed positive results for IgA; however, only two cases showed strongly positive results, while 11 cases showed weakly positive results (Table 2).

Fig. 3

IHC staining with complement protein and immunoglobulins (Ig) in the lungs of a patient with DM-ILD. This case shows strong expression of complement C3c and IgG, as well as weak expression of IgM and IgA, in alveolar cells

Figure 4 shows the expression of complement C3c and Ig in the lungs of patients with IPF. Quantitative immunohistochemical analysis for IgG and IgA showed double-positive results in 15/19 IPF cases. IgG expression was weak (n = 10) or moderate (n = 5) in the lungs of 15 IPF cases. IgA expression was weak (n = 9) or moderate (n = 6) in the lungs of 15 IPF cases. Complement C3c and IgM were absent in 17/19 IPF cases (Table 2, Additional file 3: Fig. S3).

Fig. 4

IHC staining with complement protein and Ig in the lungs of a patient with IPF. This patient was negative for complement C3c and Ig

In comparison with the IPF group, the DM-ILD group showed a significantly higher expression of complement C3c (p < 0.001), IgG (p < 0.001), and IgM (p = 0.001) (Table 2, Figs. 3 and 4). Further, we performed separate statistical analyses for acute and chronic clinical courses (Table 3). We compared 9 DM-ILD patients who met the criteria for RP-ILD (acute course) and 9 IPF patients showing AEs; the DM-ILD group showed significantly higher expression of complement C3c (p < 0.001) and IgG (p = 0.011). Compared with IPF with a chronic clinical course (n = 10), DM-ILD with a chronic clinical course (n = 9) showed significantly higher expression of C3c (p = 0.011), IgG (p = 0.037) and IgM (p = 0.001).

Table 3 Subset analysis of immunohistochemical staining performed for clinical course

Of the 18 patients (9 acute and 9 chronic courses) in the DM-ILD groups showed that the expression of C3c, and not IgG, IgM, IgA, or MDA5, was significantly (p = 0.022) higher in the acute course compared to that in the chronic courses. In contrast, in the 19 patients (9 acute and 10 chronic courses) in the IPF groups, there was no significant difference in the expression of C3c, IgG, IgM, IgA, and MDA5 in the acute course compared to the chronic courses. Additionally, there was no significant difference in the histological patterns in terms of the expression of C3c, IgA, IgG, IgM, or MDA5 between cases with DAD pattern (n = 11) and those with NSIP pattern (n = 6) in the DM-ILD group. Notably, the UIP pattern was only observed in one case (Table 1).

Subset analysis: DM-ILD patients with and without anti-MDA5 antibody

Subsequently, we performed a subset analysis of DM-ILD patients who showed positive (n = 9) and negative (n = 9) results for anti-MDA5 antibodies (Table 4). In comparison with DM-ILD patients without anti-MDA5 antibodies, those showing anti-MDA5 antibody significantly (p < 0.001) showed a histological DAD pattern in tissue samples collected at autopsy (p = 0.002). Notably, in comparison with DM-ILD without anti-MDA5 antibody, cases with the anti-MDA5 antibody showed significantly higher C3c expression in the lungs (p = 0.015). However, no significant between-group difference was observed in the expression of IgG, IgA, IgM, and MDA5.

Table 4 Subset analysis of DM-ILD with and without anti-MDA5 antibodyLung injury in human MDA5 transgenic mice

We generated three lines (lines no. 32, 55, 116) of human MDA5 transgenic mice with overexpression of human MDA5 proteins in the lungs and treated them with anti-MDA5 polyclonal antibodies (antisera) to create a new lung injury model. We also administered normal rabbit sera to transgenic mice and used them as control mice and administered antisera or normal rabbit sera to wild-type (BDF1) mice. Treatment with antisera or rabbit sera did not induce severe lung injury in wild-type BDF1 mice. In contrast, treatment with an anti-MDA5 polyclonal antibodies induced lung injury in all three lines of MDA5 transgenic mice. In the transgenic mice treated with antisera, the alveolar septum was strongly infiltrated with lymphocytes, and many of the alveoli had collapsed (Fig. 5a). Transgenic mice treated with rabbit sera also showed inflammatory cell infiltration of the alveolar septum, but it was more reduced than that observed in the lung injury model (Fig. 5b). Transgenic mice treated with purified rabbit anti-MDA5 polyclonal antibodies also induced lung injury and showed inflammatory cell infiltration of the alveolar septum. In contrast, transgenic mice treated with rabbit IgG did not induce lung injury (data not shown).

Fig. 5

Histopathological findings in the lungs of lung injury model mouse. A H&E staining of lung injury model mouse (transgenic mice treated with antisera) showing severe inflammatory cells infiltration with alveolar collapse. B H&E staining of control mouse (transgenic mice treated with rabbit sera). This shows only very mild inflammatory cells infiltration. C IHC staining with anti-human MDA5 mAb (clone H27) in lung injury model mouse showing extensive and strong expression in the alveolar epithelium. D IHC staining with anti-MDA5 mAb (clone H27) in control mouse showing very low expression in the alveolar epithelium

Human-MDA5 expression in human MDA5 transgenic mice

In the lungs of human MDA5 transgenic mice treated with anti-MDA5 antibodies (lung injury model mouse), human MDA5 protein was strongly expressed in alveolar epithelium and macrophages in the alveoli (Fig. 5c, h, e; staining in Fig. 5a). Very low human MDA5 protein expression was also observed in transgenic mice treated with rabbit sera (control mouse) (Fig. 5d, h, e; staining in Fig. 5b). The kidneys of the lung injury model mice showed glomerular atrophy and lymphocytic infiltration. However, control mice rarely showed glomerular atrophy and lymphocytic infiltration (Additional file 4: Fig. S4a, b). Human MDA5 protein was strongly expressed in the renal tubular epithelial cells and glomeruli in lung injury model mice, while control mice showed very weak MDA5 expression in renal tubular epithelial cells and negative expression in the glomeruli (Additional file 4: Fig. S4c, d).

Strong expression of complement and Ig in human MDA5 transgenic mice

IgG, IgM, IgA, and complement (C3) were all strongly expressed in the lungs of the lung injury model. IgG, IgM, IgA, and C3 were strongly expressed in the alveolar epithelium, where human MDA5 was strongly expressed (Fig. 6). When the human MDA5 lung injury models at 4 and 8 weeks were compared, slightly stronger Ig and C3 expression was observed in the mice grown for 8 weeks (Additional file 5: Fig. S5a, b). Although human MDA5 transgenic mice injected with normal rabbit serum also expressed complement and Ig, the expression was less pronounced than that seen in the lung injury model and stronger than that seen in wild-type mice treated with normal rabbit sera (Additional file 5: Fig. S5c, d). In the kidney of the lung injury model, IgG, IgM, IgA, and C3 were all strongly expressed in the glomeruli, where human MDA5 was strongly expressed, in comparison with the corresponding expression levels in the control mice (Additional file 6: Fig. S6).

Fig. 6

Immunohistochemical (IHC) analysis for complement protein and Ig in the lungs of the lung injury model mice. A Lung injury model mice (transgenic mice treated with antisera) showed moderate to severe expression of C3 and Ig. B Control mice (transgenic mice treated with rabbit sera) showed moderate expression of IgG but showed very mild expression of the other immunoglobulins and C3