Clinical characteristics of anti-Jo-1 autoantibody-positive patients

The study included 115 patients who were positive for anti-Jo-1 on first admission. Anti-Jo-1 antibody positivity was determined by both ELISA and line blot. The enrolled patients were 85 women (73.9%) and 30 men (26.1%), with an age of onset of [52 (41,61; IQR)] years and a disease duration of [9(3,36); IQR] months. At baseline, 70 patients (60.8%) initially presented with ILD, 46 patients (40.0%) with muscle weakness, 46 (40%) with skin involvement, and 58 (50.4%) with arthritis. Based on lung and muscle involvement at baseline, 115 patients included 52 patients (45.2%) with isolated ILD, 28 (24.3%) patients with isolated myositis, 18 (15.7%) patients with both and 17 (14.8%) patients with neither. Anti-Jo-1 antibody levels were as 135 (98,185) U/ml at baseline for all patients. During the course of disease, 112 (97.4%) patients developed ILD, 52 (45.2%) with muscle weakness, 75 (65.2%) with skin involvement, and 64 (55.7%) with arthritis (Table 1).

Table 1 Characteristics of anti-Jo-1-positive patientsAssociation of anti-Jo-1 antibody levels with different organ involvement

When comparing the serum levels of anti-Jo-1 antibody levels at baseline between patients with different symptoms on admission, patients with ILD at baseline had lower levels of anti-Jo-1 antibodies than those without ILD at baseline [122 (94, 167) vs. 158 (113, 200) U/ml, p = 0.012] (Fig. 1A). When comparing anti-Jo-1 levels in the four groups of patients with both ILD and myositis, isolated ILD, isolated myositis, and neither, a statistically significant difference was found only in the isolated myositis and isolated ILD groups [177(124,200) vs. 121(95,155), p < 0.001] (Fig. 1B). Patients with isolated myositis had higher baseline anti-Jo-1 levels than those with isolated ILD [177 (124,200) vs. 121(95,155) U/ml, p = 0.004] (Fig. 1B). Baseline anti-Jo-1 antibody levels were higher in patients with muscle weakness [168(110, 200) vs. 122 (96, 165) U/ml, p = 0.018] (Fig. 1C), skin involvement [153(118, 200) vs. 122 (86, 168) U/ml, p = 0.009] (Fig. 1D), and arthritis [164 (111, 200) vs. 120 (85, 157) U/ml, p = 0.006] (Fig. 1E) compared to those without these manifestations. We also compared anti-Jo-1 levels in ASS with and without anti-Ro52, as well as in treated and treatment naïve patients. There was no difference in anti-Jo-1 antibody levels in patients with and without anti-Ro-52 [139(110,192) vs. 120(82,181), p = 0.112], and treated and treatment naive patients [129(96,182) vs. 136(105,199), p = 0.527] (Supplementary Tables 1 and 2).

Fig. 1

The serum levels of anti-Jo-1 in patients with anti-synthetase syndrome. A Anti-Jo-1 levels in patients with and without ILD at baseline [122 (94, 167) vs. 158 (113, 200) U/ml, p = 0.012]; B Anti-Jo-1 levels in patients with isolated ILD [121(95,155)], both ILD and myositis [144(86,195)], isolated myositis [177(124,200)], and neither ILD nor myositis [143(109,180)]; C Anti-Jo-1 levels in patients with and without muscle weakness at baseline[168(110, 200) vs. 122 (96, 165) U/ml, p = 0.018]; D Anti-Jo-1 levels in patients with and without skin involvement on the first admission [153(118, 200) vs. 122 (86, 168) U/ml, p = 0.009]; E Anti-Jo-1 levels in patients with and without arthritis at baseline [164 (111, 200) vs. 120 (85, 157) U/ml, p = 0.006]. *p < 0.05; **p < 0.01

Anti-Jo-1 levels in patients with new development of organ involvement were also analyzed. Anti-Jo-1 levels were higher in patients with new development of skin involvement than those without new development of skin involvement [136 (101, 191) vs. 107 (66, 155) U/ml, p = 0.032], but no statistical difference was found in new development of other organ involvement (Supplementary Figure 1).

Correlation between anti-Jo-1 autoantibody levels and disease activity

In the overall cohort, the anti-Jo-1 levels were negatively correlated with the age of onset (r =-0.30, p = 0.001). Baseline anti-Jo-1 antibody levels of patients were positively correlated with skin VAS score (r =0.25, p = 0.006), but were not correlated with PGA, muscle, pulmonary, or joint VAS scores. In addition, the baseline anti-Jo-1 antibody levels were positively associated with levels of CK (r = 0.24, p = 0.011), IgG (r = 0.25, p = 0.008), IgM (r = 0.22, p = 0.019), and erythrocyte sedimentation rate (ESR; r = 0.20, p = 0.035). No correlation was found between anti-Jo-1 antibody levels and lung function (Fig. 2).

Fig. 2

Correlation matrix of baseline anti-Jo-1 levels, laboratory test results, and disease activity. Spearman’s correlation analysis was used to analyze the correlation between the variables. The numbers in the boxes represent the correlation coefficient between the two variables (red negative, blue positive) when p < 0.05. PGA, Physician’s Global Assessment; VAS, Visual Analog Scale; ALT, alanine aminotransferase; AST, aspartate aminotransferase; LDH, lactic dehydrogenase; CK, creatine kinase; CRP, C reactive protein; ESR, erythrocyte sedimentation rate; %FVC, percent predicted forced vital capacity; %FEV1, percent predicted forced expiratory volume in one second; %DLco, percent predicted carbon monoxide diffusion capacity

The correlation between anti-Jo-1 levels and disease activity in different subgroups were shown in Supplementary Table 3 (with and without ILD and baseline), Supplementary Table 4 (with and without anti-Ro-52) and Supplementary Table 5 (treated and treatment naive), and no significant correlation was found.

Longitudinal correlation between anti-Jo-1 autoantibody levels and disease activity

Ninety-eight of the 115 patients were followed up to survival status, with 1–9 visits. The follow-up disease activity and anti-Jo-1 antibody levels were examined in 43 patients with two or more follow-up visits. The longitudinal correlation between anti-Jo-1 antibody levels and disease activity was further estimated using GEE models. Overall, changes in anti-Jo-1 antibody levels were significantly positively correlated with the changes in PGA (β = 0.002, p = 0.001), muscle (β = 0.003, p < 0.0001), and pulmonary (β = 0.002, p = 0.013) VAS scores, but not with skin and joint VAS scores. Meanwhile, changes in anti-Jo-1 antibody levels were also positively correlated with changes in the levels of AST, CK, IgG, IgM (all p < 0.05) and negatively correlated with changes in the percent predicted forced vital capacity (%FVC; β = -0.025, p < 0.0001), percent predicted forced expiratory volume in one second (%FEV1; β = -0.021, p < 0.0001) (Fig. 3 and Supplemental Table 6).

Fig. 3

Longitudinal relationship between anti-Jo-1 levels and the disease activity in all anti-Jo-1 patients (A), anti-Jo-1 patients with ILD at baseline (B), and anti-Jo-1 patients with non-ILD at baseline (C). LD interstitial lung disease, PGA Physician’s Global Assessment, VAS Visual Analog Scale

In patients with ILD at baseline, changes in anti-Jo-1 antibody levels were correlated with changes in PGA, pulmonary, and joint VAS scores; IgG, IgA, IgM, and CRP levels; %FVC; %FEV1; and percent predicted carbon monoxide diffusion capacity (%DLco) (all p < 0.05). Variations in anti-Jo-1 antibody levels were significantly correlated with changes in PGA, muscle, and pulmonary VAS scores; ALT, CK, IgG, and IgA levels; %FVC, and %FEV1 in patients without ILD at baseline (all p < 0.05) (Table 2).

Table 2 Univariate and multivariate Cox proportional hazard models of death in anti-Jo-1-positive patients

Among 43 patients with two or more follow-up visits, 8 who were anti-Jo-1-positive had more than 3 visits. The variability of anti-Jo-1 antibody and serum CK levels, %FVC, and %DLco over time in these patients showed comparable trends over time in the majority of patients (Fig. 4).

Fig. 4

Longitudinal changes of anti-Jo-1 levels in and serum CK levels, %FVC pred, and %DLco pred over time in 8 anti-Jo-1-positive patients with more than 3 times visits. The black dashed line indicates the cut-off line for the normal level of anti-Jo-1 antibodies(15U/ml). Abbreviation: ILD, interstitial lung diseases; CK, creatine kinase; %FVC, percent predicted forced vital capacity; %DLco, percent predicted carbon monoxide diffusion capacity; GCs, corticosteroids steroids; DMARDS, disease-modifying anti-rheumatic drugs; CsA, ciclosporin A; CTX, cyclophosphamide; MMF, mycophenolate mofetil; IVIG, intravenous. immunoglobulin; tac, tacrolimus

Association between anti-Jo-1 antibody levels and prognosis in patients with ASS

A total of 98 anti-Jo-1 positive patients were followed up for 1 month to 118.87 months, and the median follow-up time was 41.89 months. Eleven deaths were observed, including nine from respiratory failure, one from endometrial cancer, and one from malignancies.

Risk factors for mortality in anti-Jo-1 positive patients were analyzed. Univariate analyses showed that older age of onset (HR 1.074, 95% CI: 1.021–1.129, p = 0.006), arthritis at baseline(HR 0.184, 95% CI: 0.039–0.855, p = 0.031), and higher levels of IgA (HR 1.006, 95% CI: 1.001–1.010, p = 0.018) and CRP (HR 1.338, 95% CI: 1.037–1.726, p = 0.025) were risk factors for death in anti-Jo-1-positive patients. Further multivariate Cox regression analysis showed that a higher age of onset (HR 1.069, 95% CI: 1.010–1.133, p = 0.022), and higher CRP levels (HR 1.333, 95% CI: 1.035–1.717, p = 0.026) were risk factors for death in anti-Jo-1-positive patients. Patients who died had an older age of onset [61 (48, 79) vs. 50 (39, 61) years old, p = 0.009] and higher CRP levels [1.55 (0.41, 2.90) vs. 0.43 (0.21, 1.30) mg/dl, p = 0.039] than those who survived. Baseline anti-Jo-1 levels were not risk factors for death in anti-Jo-1-positive patients (p = 0.997) (Table 2). The multivariate cox model, which included age of onset and CRP, showed the best efficacy for predicting death, with the largest area under the ROC curve [0.770(0.625–0.915)] (Supplementary Table 6).