MMP1 and MMP7 as potential peripheral blood biomarkers in idiopathic pulmonary fibrosis

Abstract

Background: Idiopathic pulmonary fibrosis (IPF) is a chronic progressive fibrotic lung disease associated with substantial morbidity and mortality. The objective of this study was to determine whether there is a peripheral blood protein signature in IPF and whether components of this signature may serve as biomarkers for disease presence and progression.

Methods and findings: We analyzed the concentrations of 49 proteins in the plasma of 74 patients with IPF and in the plasma of 53 control individuals. We identified a combinatorial signature of five proteins-MMP7, MMP1, MMP8, IGFBP1, and TNFRSF1A-that was sufficient to distinguish patients from controls with a sensitivity of 98.6% (95% confidence interval [CI] 92.7%-100%) and specificity of 98.1% (95% CI 89.9%-100%). Increases in MMP1 and MMP7 were also observed in lung tissue and bronchoalveolar lavage fluid obtained from IPF patients. MMP7 and MMP1 plasma concentrations were not increased in patients with chronic obstructive pulmonary disease or sarcoidosis and distinguished IPF compared to subacute/chronic hypersensitivity pneumonitis, a disease that may mimic IPF, with a sensitivity of 96.3% (95% CI 81.0%-100%) and specificity of 87.2% (95% CI 72.6%-95.7%). We verified our results in an independent validation cohort composed of patients with IPF, familial pulmonary fibrosis, subclinical interstitial lung disease (ILD), as well as with control individuals. MMP7 and MMP1 concentrations were significantly higher in IPF patients compared to controls in this cohort. Furthermore, MMP7 concentrations were elevated in patients with subclinical ILD and negatively correlated with percent predicted forced vital capacity (FVC%) and percent predicted carbon monoxide diffusing capacity (DLCO%).

Conclusions: Our experiments provide the first evidence for a peripheral blood protein signature in IPF to our knowledge. The two main components of this signature, MMP7 and MMP1, are overexpressed in the lung microenvironment and distinguish IPF from other chronic lung diseases. Additionally, increased MMP7 concentration may be indicative of asymptomatic ILD and reflect disease progression.

Figure 1. Peripheral Blood Proteins Distinguish IPF Patients from Controls

(A) Heatmap of proteins measured in the plasma of IPF and control patients. Columns, individual patients; rows, proteins. Every protein level was divided by the geometric mean of values for the same proteins for all patients and log based 2 transformed. Increasing shades of yellow, increased; increased shades of purple, decreased; gray, unchanged. Proteins were clustered using Genomica. Red vertical line, cluster of proteins increased in IPF; green vertical line, cluster of proteins decreased in IPF. (B) Classification tree obtained by CART applied to plasma protein concentration data from IPF patients and controls. A blue box identifies a terminal node as control; a red box as IPF. All counts are listed as control/IPF. Concentrations are in ng/ml. In the subgroup with high MMP7 concentration but low MMP1 concentration (14 IPF samples, five control samples), splitting on IGFBP1 and TNFRSF1A improves classification, while in the subgroup with low MMP7, MMP8 improves classification. (C) ROC curves for using each of five markers, or their combination, to classify samples as IPF or control. Sensitivity, or true positive rate, is plotted on the y-axis, and false positive rate, or 1 − specificity, on the x-axis. The area under each ROC curve is equivalent to the numerator of the Mann-Whitney U-statistic comparing the marker distributions between IPF and control samples. The magenta line labeled “Combined” is for the combinatorial classifier using all five markers. The identity line at 45 ° represents a marker that performed no better than classifying samples as IPF or control by flipping a fair coin.

Figure 2. MMP7 and MMP1 Gene and Protein Levels Are Significantly Increased in the Lungs of Patients with IPF

(A) Average gene expression levels (log scale) measured using gene expression microarrays of genes that encode the 49 protein markers in IPF lungs (y-axis) compared to control lungs (x-axis). Colored squares (black or red) are genes that encode proteins that changed significantly in plasma. Red squares are genes that changed significantly (SAM Q value <5%) in gene expression data and that encode proteins measured in peripheral blood. Green oblique lines denote 2-fold change. (B and C) MMP7 (B) and MMP1 (C) concentrations (ng/ml) are significantly (p < 0.00001 and p = 0.018, respectively) higher in BAL fluid of patients with IPF (n = 22) compared to control individuals (n = 10).

Figure 3. MMP7 and MMP1 Plasma Concentrations Are High in IPF, but Not Sarcoidosis or COPD

Concentrations (ng/ml) of MMP7 (A) and MMP1 (B) are significantly higher in patients with IPF (n = 74; p < 0.00001 and p = 0.018, respectively), compared to controls (n = 53), but not sarcoidosis (n = 47; p = 0.78 and p = 0.28, respectively), compared to controls (n = 53) or COPD (n = 73; p = 0.21 and 0.85, respectively, stratified by GOLD class, as 0–I, II, and III–IV).

Figure 4. MMP7 and MMP1 Serum Concentrations Are Higher in IPF, Compared to HP

(A and B) Concentrations (ng/ml) of MMP7 (A) and MMP1 (B) in the blood are significantly higher in patients with IPF (n = 34) than in patients with HP (n = 41). (C) Average gene expression levels (log scale) in IPF samples (y-axis) compared to HP (x-axis) measured by gene expression microarrays. Gray circles, all genes on the array; red circles, MMP1 and MMP7. Green oblique lines denote 2-fold change. (D) Combinations of serum MMP7 (y-axis) and MMP1 concentrations (x-axis) in IPF (closed circles) and HP patients (open circles). Corners represent points in which the trade-off between positive predictive value (PPV) and negative predictive value (NPV) are optimal for ruling out IPF (blue) or concluding IPF (red) on the basis of MMP1 and MMP7 concentrations. (E) ROC curves for using MMP1 or MMP7, or their combination, to classify samples as IPF or HP. Sensitivity, or true positive rate, is plotted on the y-axis, and false positive rate, or 1 − specificity, on the x-axis. The identity line at 45 ° represents a marker that performed no better than classifying samples as IPF or HP by flipping a fair coin.

Figure 5. MMP7 Concentrations Significantly Distinguish Control from Subclinical ILD, Familial, or Sporadic IPF

(A) Dark solid lines show median concentrations in each group. The interquartile range (IR) or middle 50% of concentrations is delimited by a box. Data are expressed on a log base 2 scale. (B) ROC curves for using MMP1 or MMP7, or their combination, to classify samples as IPF (sporadic or familial) or control in validation cohort. (C and D) Serum MMP7 concentrations moderately correlate with decreases in FVC% (C) and DL_CO% (D). Linear regressions and 95% CI inversely relate MMP7 concentration (ng/ml) to FVC% and DLCO%. *p < 0.05, **p < 0.01, and ***p < 0.001.