Identification of MMP28 as a biomarker for the differential diagnosis of idiopathic pulmonary fibrosis

Abstract

Background and objective

Idiopathic Pulmonary Fibrosis (IPF) is a progressive disease of unknown etiology. The diagnosis is based on the identification of a pattern of usual interstitial pneumonia either by high resolution computed tomography and/or histology. However, a similar pattern can be observed in other fibrotic lung disorders, and precise diagnosis remains challenging. Studies on biomarkers contributing to the differential diagnosis are scanty, and still in an exploratory phase. Our aim was to evaluate matrix metalloproteinase (MMP)-28, which has been implicated in abnormal wound healing, as a biomarker for distinguishing IPF from fibrotic non-IPF patients.

Methods

The cell localization of MMP28 in lungs was examined by immunohistochemistry and its serum concentration was measured by ELISA in two different populations. The derivation cohort included 82 IPF and 69 fibrotic non-IPF patients. The validation cohort involved 42 IPF and 41 fibrotic non-IPF patients.

Results

MMP28 was detected mainly in IPF lungs and localized in epithelial cells. In both cohorts, serum concentrations of MMP28 were significantly higher in IPF versus non-IPF (mostly with lung fibrosis associated to autoimmune diseases and chronic hypersensitivity pneumonitis) and healthy controls (ANOVA, p<0.0001). The AUC of the derivation cohort was 0.718 (95%CI, 0.635–0.800). With a cutoff point of 4.5 ng/mL, OR was 5.32 (95%CI, 2.55–11.46), and sensitivity and specificity of 70.9% and 69% respectively. The AUC of the validation cohort was 0.690 (95%CI, 0.581–0.798), OR 4.57 (95%CI, 1.76–12.04), and sensitivity and specificity of 69.6% and 66.7%. Interestingly, we found that IPF patients with definite UIP pattern on HRCT showed higher serum concentrations of MMP28 than non-IPF patients with the same pattern (7.8±4.4 versus 4.9±4.4; p = 0.04). By contrast, no differences were observed when IPF with possible UIP-pattern were compared (4.7±3.2 versus 3.9±3.0; p = 0.43).

Conclusion

These findings indicate that MMP28 might be a useful biomarker to improve the diagnostic certainty of IPF.