CXCL10 predicts autoimmune features and a favorable clinical course in patients with IIP: post hoc analysis of a prospective and multicenter cohort study

Abstract

Background: Interstitial pneumonia with autoimmune features (IPAF), which does not meet any of the criteria for connective tissue diseases (CTD), has been attracting an attention in patients with idiopathic interstitial pneumonia (IIP). However, the biomarkers that reflect the clinical course of these patients have not been fully elucidated.

Objective: To identify useful serum biomarkers reflecting CTD-related features and favorable prognoses in patients with IIP.

Methods: This was a post hoc analysis of a prospective and multicenter cohort study between 2015 and 2020. Newly diagnosed patients with IIP were consecutively enrolled, and 74 autoimmune features and autoantibodies were comprehensively checked during IIP diagnosis. Serum levels of CXCL10, CXCL1, CCL2, BAFF, angiopoietin-2, and leptin were evaluated at the time of IIP diagnosis.

Results: Two hundred twenty-two patients (159 men and 63 women) with IIP were enrolled. The median observation duration was 36 months. The median age was 71 years old, and median %forced vital capacity (FVC) was 84.1% at the time of IIP diagnosis. The proportion of patients who met the classification criteria for IPAF was 11.7%. In patients with high serum CXCL10, changes in both %FVC and %diffusion lung capacity for carbon monoxide at one year were significantly higher than those in patients with low CXCL10 (p = 0.014 and p = 0.009, respectively), whereas these changes were not significant for other chemokines and cytokines. High CXCL10 levels were associated with acute/subacute onset (p < 0.001) and the diagnosis of nonspecific interstitial pneumonia with organizing pneumonia overlap (p = 0.003). High CXCL10 levels were related to a higher classification of IPAF (relative risk for IPAF was 3.320, 95%CI: 1.571-7.019, p = 0.003) and lower classification of progressive pulmonary fibrosis (PPF; relative risk for PPF was 0.309, 95%CI: 0.100-0.953, p = 0.027) compared to those with low CXCL10. Finally, survival was higher in patients with IPF and high CXCL10 (p = 0.044), and high CXCL10 was a significant prognostic factor in multivariate Cox proportional hazards models (hazard ratio 0.368, p = 0.005).

Conclusions: High serum levels of CXCL10 are associated with CTD-related features, the favorable clinical course, and survival in patients with IIP, especially IPF.

Clinical trial number: Not applicable.

Keywords: C-X-C motif chemokine 10; Idiopathic interstitial pneumonia; Interstitial pneumonia with autoimmune features; Nonspecific interstitial pneumonia; Organizing pneumonia.

Fig. 1

Serum chemokines/cytokines concentrations and changes in lung physiological parameters. Serum chemokines and cytokines including CXCL1, CXCL10, CCL2, ANGP-2, BAFF/BLyS, and leptin, which are related to the pathogenesis of CTDs, were measured at the time of IIP diagnosis. The volcano plots show Wilcoxon rank-sum test p-values of chemokines/cytokines for the changes in %FVC (A) or %DL_CO (B). Serum CXCL10 level is a sole significant chemokine, indicating that high CXCL10 level (divided by the mean value of 73.12 pg/mL) can predict the increase of %FVC (p = 0.014, A) and %DL_CO (p = 0.009, B) one year after IIP diagnosis. CTD, connective tissue disease; IIP, idiopathic interstitial pneumonia; CXCL, C-X-C motif chemokine ligand; CCL, C-C motif chemokine ligand; BAFF: B-cell activating factor; BLyS, B-lymphocyte stimulator; ANGP, angiopoietin; FVC, forced vital capacity; DL_CO, the diffusion lung capacity for carbon monoxide

Fig. 2

Treatments and lung physiological parameters in patients with high serum CXCL10. Serum CXCL10 levels were divided into two groups based on the mean value of 73.12 pg/mL. Baseline %FVC at the time of IIP diagnosis were not different between patients with high and low serum CXCL10 levels (A). Patients with high CXCL10 level were more frequently treated with immunosuppressive therapies including corticosteroids and immunosuppressants than those with low CXCL10 level (B). Changes in %FVC (C) and %DL_CO (D) one year after IIP diagnosis are shown. Changes in %FVC and %DL_CO significantly increased in patients with high CXCL10 (p = 0.014 and p = 0.009, respectively). CXCL, C-X-C motif chemokine ligand; FVC, forced vital capacity; DL_CO, the diffusion lung capacity for carbon monoxide; IIP, idiopathic interstitial pneumonia

Fig. 3

Diagnoses of IIP, classification of IPAF, and clinical characteristics of patients with high CXCL10. In patients with high CXCL10, the proportions of NSIP with OP overlap, COP, and NSIP increased, while that of IPF decreased (p = 0.003, A and B). In patients with high CXCL10, hospital visit due to any symptoms tended to increase (p = 0.051, C), and increased acute/subacute onset was found (p < 0.001, D). Higher systemic autoimmune disease-specific autoantibody positivity was found (p = 0.002, E) and more patients with IPAF (p = 0.003, F) had high CXCL10 levels. IIP, idiopathic interstitial pneumonia; IPAF, interstitial pneumonia with autoimmune features; CXCL, C-X-C motif chemokine ligand; NSIP, nonspecific interstitial pneumonia; COP, cryptogenic organizing pneumonia; SAID, systemic autoimmune disease

Fig. 4

HRCT findings and lung histopathological specimens of patients with high CXCL10. Representative HRCT pattern of NSIP with OP overlap is shown (A). A 76 year-old man visited our hospital because of dry cough. The patient was found positive for rheumatoid factor (49.9 IU/mL), anti-Scl-70 antibody, and MPO-ANCA. This patient was classified as having IPAF and had a high serum CXCL10 level (688 pg/mL). In all patients with IIP, consolidation opacity on HRCT was more frequently found in patients with high CXCL10 level (B). According to the guideline for IPF [31], a higher proportion of alternative pattern for IPF and a lower proportion of UIP/probable UIP pattern on HRCT were found, but not significant, in patients with high CXCL10 level compared to those with low CXCL10 level (p = 0.073, C). Regarding surgical lung biopsy specimens, no significant differences in systemic autoimmune disease-like findings (interstitial lymphoid aggregates with germinal centers, prominent plasmacytic infiltration, dense perivascular collagen, and extensive pleuritis) were observed between the two groups (p = 0.302, D). HRCT, high-resolution computed tomography; MPO-ANCA, myeloperoxidase-anti-neutrophil cytoplasmic antibody; IPAF, interstitial pneumonia with autoimmune features; CXCL, C-X-C motif chemokine ligand; NSIP, nonspecific interstitial pneumonia; OP, organizing pneumonia; SAID, systemic autoimmune disease

Fig. 5

Clinical course, classification of PPF, occurrence of AE-IIP, and survival in patients with high CXCL10 level. Patients with high CXCL10 level showed significant improvements in respiratory symptoms (p < 0.001, A) and opacities on HRCT (p = 0.002, B) at one year. Less number of patients with high CXCL10 met the criteria for PPF one year after IIP diagnosis (p = 0.012, C), and the relative risk of high CXCL10 level for PPF was 0.309 (95%CI: 0.100-0.953). Cumulative incidence of AE-IIP tended to be lower in patients with high CXCL10 level than in those with low CXCL10 level (IPF: Gray’s test, p = 0.126, D; non-IPF: p = 0.384, E). IPF patients with high CXCL10 levels showed a significantly better prognosis (log-rank test, p = 0.044, F), while non-IPF patients with high CXCL10 levels did not (p = 0.927, G) compared to those with low CXCL10 levels. When patients were limited to those who received immunosuppressive treatments or no treatment, patients with IPF with high CXCL10 level showed significantly better survival than those with low CXCL10 level (p = 0.037, H), whereas non-IPF patients showed no differences (p = 0.856, I). PPF, progressive pulmonary fibrosis; AE-IIP, acute exacerbation of idiopathic interstitial pneumonia; CXCL, C-X-C motif chemokine ligand; HRCT, high-resolution computed tomography; CI, confidence interval

Fig. 6

Serum CXCL10 in patients with IPAF. Patients with IPAF (11.7% of all IIP, 26 patients) showed significantly higher serum CXCL10 level than those without IPAF (median 51 vs. 0 pg/mL and mean 240 vs. 53 pg/mL, p = 0.003; A). At the time of IIP diagnosis, baseline %FVC was significantly lower in patients with high CXCL10 level than in those with low CXCL10 level (p = 0.021, B). Patients with high CXCL10 level tended to receive more immunosuppressive therapy compared to those with low CXCL10 level (p = 0.099, C). Change in %FVC at one year was significantly higher in patients with high CXCL10 level than in those with low CXCL10 level (p = 0.029, D). Change in %DL_CO at one year tended to be higher in patients with high CXCL10 level than in those with low CXCL10 level (p = 0.245, E). CXCL, C-X-C motif chemokine ligand; IPAF, interstitial pneumonia with autoimmune features; FVC, forced vital capacity; DLCO, the diffusion lung capacity for carbon monoxide