Matrix Metalloproteinase-10: A Novel Biomarker for Idiopathic Pulmonary Fibrosis

Akihiko Sokai; Tomohiro Handa; Kiminobu Tanizawa; Toru Oga; Kazuko Uno; Tatsuaki Tsuruyama; Takeshi Kubo; Kohei Ikezoe; Yoshinari Nakatsuka; Kazuya Tanimura; Shigeo Muro; Toyohiro Hirai; Sonoko Nagai; Kazuo Chin; Michiaki Mishima


Respiratory Research. 2015;16(120) 

In This Article


Patient Characteristics

Fifty-seven patients with IPF were recruited; all patient characteristics are shown in Table 1. Ten (17.5 %) patients were diagnosed by histological confirmation of the usual interstitial pneumonia (UIP) pattern, whereas the other patients were diagnosed based on a definite UIP pattern on high-resolution computed tomography. The mean age of the patients with IPF was 69.4 years, and 51 (89.5 %) patients were male, whereas the mean age of patients with COPD or healthy controls was 70.8 and 65.9 years, respectively. All but one patient with COPD (95.0 %) and all healthy controls were male. No significant difference was observed in age and gender ratios between IPF and COPD nor between IPF and healthy controls (Table 1). All patients with COPD and seven control subjects (46.7 %) had current or former history of smoking. Twenty-five (43.9 %) patients underwent BAL. Three patients had a history of malignant disease: 1 patient had breast cancer, 1 patient had prostate cancer, and another patient had colorectal cancer. No patients suffered disease recurrence over the more than 3 years following their last treatments. Thirteen patients were treated at baseline with prednisolone, an immunosuppressant, or pirfenidone (Additional file 1: Table S1).

Serum and BALF MMPs in IPF

The serum and BALF concentrations are presented in Table 2. MMP-13 was excluded from the analysis because its mean concentration was under the limit of determination in both patient groups and controls. Serum MMP-1 (p < 0.01), MMP-2 (p = 0.03), MMP-7 (p < 0.01), MMP-8 (p = 0.01), MMP-10 (p < 0.01), and MMP-12 (p < 0.01) concentrations were significantly elevated in patients with IPF compared with controls, whereas serum MMP-2 (p < 0.01), MMP-7 (p < 0.01), MMP-9 (p = 0.03), MMP-10 (p < 0.01), MMP-12 (p = 0.01) concentrations were significantly elevated in patients with IPF compared with COPD patients.

Correlation Between Serum MMPs and Disease Severity

The correlation between pulmonary function and MMPs is shown in Table 3. There were significant correlations between MMP-7 and the PFT indices, including %FVC and %DLCO (ρ = −0.31, p = 0.02 and ρ = −0.32, p = 0.02, respectively). There were also significant correlations between MMP-10 and %FVC and %DLCO (ρ = −0.34, p < 0.01 and ρ = −0.43, p < 0.01, respectively). There were significant correlations noted between MMP-10 and the 6-minute walk distance (ρ = −0.38, p < 0.01), as well as between MMP-10 and minimum oxygen saturation (SpO2) during the 6MWT (ρ = −0.42, p < 0.01). Only MMP-10 correlated significantly with the partial pressure of oxygen (ρ = −0.32, p = 0.02).

Correlation Between Serum and BALF MMPs

Based on the MMP results and their relationship with disease severity, the BALF concentrations of MMP-7 and -10 were measured. Among the 25 patients who underwent BAL, serum and BALF MMP measurements were simultaneously performed in 19 patients to investigate potential correlations. Significant correlations were observed between serum and BALF concentrations of MMP-7 and MMP-10 (ρ = 0.64, p < 0.01 and ρ = 0.47, p = 0.04, respectively). We also examined 25 BALF samples from the patients who underwent both BAL and PFT within a 1-month interval. BALF MMP-10 correlated significantly with %FVC and %DLCO (ρ = −0.46, p = 0.02 and ρ = -0.44, p = 0.03, respectively), whereas BALF MMP-7 correlated significantly with only %DLCO (ρ = −0.54, p = < 0.01).

Relationship Between MMPs and Clinical Deterioration or Survival

The median follow-up time from baseline was 459 days (range, 12–1853 days). Fourteen patients deteriorated clinically within 6 months; 2 died due to an acute exacerbation (AE), and 9 exhibited declines in pulmonary function. Furthermore, 3 patients were admitted due to respiratory failure: 2 suffered from chronic respiratory failure, and 1 suffered from a respiratory infection. The causes of death at more than 6 months following baseline included 3 cases of AE, 2 cases of chronic respiratory failure, 1 case of acute lung injury due to pulmonary infection, 1 case of lung cancer, 1 case of sepsis, and 1 case without a clear diagnosis. Logistic analyses for clinical deterioration within 6 months revealed that serum MMP-10 was a significant predictor of clinical deterioration, as were %FVC and %DLCO (Table 4A). A survival analysis performed using a Cox proportional hazard model demonstrated that serum MMP-10, as well as %FVC and %DLCO, was a significant predictor of mortality among patients with IPF (Table 4B). Serum MMP-10 was a significant predictor of clinical deterioration and mortality, even when the patients treated at baseline were excluded (Additional file 1: Table S2). In contrast, serum and BALF MMP-7 were not predictors of either clinical deterioration within 6 months or mortality.

We performed an ROC analysis to determine the optimal cut-off value of serum MMP-10 for predicting clinical deterioration. The curve had an AUC of 0.741 and a cut-off value of 0.986 ng/μL. When 1.0 ng/μL was set as the threshold, patients with higher values had a significantly higher frequency of clinical deterioration within 6 months compared with patients with lowers values (p = 0.01). Furthermore, the mortality of the patients with higher serum MMP-10 values was significantly higher than patients with lower values (log rank test; p = 0.049).


Immunohistochemical staining for MMP-10 was performed in IPF lung tissue. The expression of MMP-10 was localized primarily to the alveolar macrophages, alveolar epithelial cells, and peripheral bronchiolar epithelial cells in IPF lung tissue (Fig. 1). Positive immunostaining for MMP-10 was observed in both macrophages and alveolar epithelial cells in control lung tissue. However, the number of positive cells was reduced, and the signal intensity was weaker in control lung tissue compared with IPF lung tissue.

Figure 1.

The immunohistochemical expression of matrix metalloproteinase (MMP)-10 in idiopathic pulmonary fibrosis (IPF) lung tissue and control lung tissue. (a-c) The expression of MMP-10 is weakly positive in alveolar epithelial cells and macrophages in control lung tissue. (d-f) In IPF lung tissue, MMP-10 is expressed predominantly in alveolar macrophages, alveolar epithelial cells, and peripheral bronchiolar epithelial cells. The staining intensity of MMP-10 in IPF lung tissue is stronger compared with control lung tissue