MMP expression and abnormal lung permeability are important determinants of outcome in IPF

Subjects

In total, 20 patients with IPF diagnosed according to current American Thoracic Society (ATS) criteria were recruited from the specialist Interstitial Lung Disease clinic at the University Hospital Birmingham (Birmingham, UK). Diagnosis was supported by open lung biopsy in five cases, histology post lung transplant in one case and high-resolution computed tomography evidence in all cases. Where open lung biopsy was performed, tissue was obtained from at least two sites, the upper and lower lobes of the same lung. Baseline bronchoscopy was performed during the investigation stage after referral, before the start of definitive treatment. Eight healthy individuals free from respiratory disease were recruited as normal controls. The study was approved by the local ethical committee and participants gave written informed consent.

Measurements

Patients underwent bronchoscopy and bronchoalveolar lavage (BAL) as described previously 14. TIMP-1, TIMP-2 and VEGF were measured in BALF using ELISA kits (R&D Systems, Abingdon, UK). BALF MMP-1, -2, -3, -7, -8, -9, -12 and -13 were measured using Luminex array (R&D Systems). Luminex array values for normal MMP levels have been included in an as-yet unpublished study on acute respiratory distress syndrome (ARDS; C. O'Kane, personal communication). IPF MMP-3 levels have been quoted in a further unpublished study concerning endostatin cleavage (A.G. Richter, personal communication). Protein was measured using the Bio-Rad DC Protein Assay Kit II (Bio-Rad Laboratories, Hemel Hempstead, UK). The protein permeability index (PPI) was calculated as the ratio of BALF to plasma protein, as described previously 15.

Gelatin zymography

Gelatin zymography was carried out as previously described 16. In brief, BALF was diluted 1/10, and 20-μL aliquots were mixed with 5 μL loading buffer (0.25 M Tris pH 6.8, 50% glycerol, 5% SDS, bromophenol blue) and were run on 11% acrylamide gels impregnated with 0.1% gelatin at 180 V for 3.5 h (buffer 25 mM Tris, 190 mM glycine, 0.1% SDS). After incubation in 2.5% Triton X for 1 h with agitation and two brief washes in collagenase buffer (55 mM Tris base, 200 mM sodium chloride, 5 mM calcium chloride, 0.02% Brij, pH 7.6), gels were incubated for 16 h in fresh collagenase buffer at 37°C. Gelatinolytic activity was detected by a single step stain/de-stain method using 0.02% Coomasie blue in 1/3/6 acetic acid/methanol/water. All experimental samples were run in parallel with 2 ng recombinant MMP-9 (Oncogene, Cambridge, UK) to standardise between gels.

MMP-9 activity assay

BALF was diluted 1/25 or 1/50 and analysed using the MMP-9 fluorokine activity assay (R&D Systems) according to the manufacturer's instructions. A standard curve of purified MMP-9 was run on the same plates (and pro-forms of MMP-9 standard were activated by 4-aminophenylmercuric acetate (APMA) for 2 h before the addition of the fluorogenic substrate) to allow quantification. The minimal detectable concentration of MMP-9 using this assay is reported as 5 pg·mL⁻¹ (manufacturer's data). This assay takes into account the effect of TIMPs in the sample and gives a measure of net MMP-9 activity.

Pulmonary function testing

Forced vital capacity (FVC) was measured using the Jaeger Compact system (Viasys Healthcare, Basingstoke, UK). Total lung diffusing capacity for carbon monoxide (T_L,CO) was measured using a single-breath technique (Jaeger Compact system). Results were expressed as % of predicted values. Where possible, sequential lung function was obtained 3 monthly as part of clinical follow-up. Rapid lung decline was defined as a consistent reduction (>10%) in FVC, total lung capacity or T_L,CO 1 yr after presentation 1.

Statistics

IPF BALF protein data were non-parametric and are thus presented as median (interquartile range (IQR)). Between-group comparisons were performed using the Mann–Whitney U-test. A p-value of ≤0.05 for all data was considered significant. Correlations were made using the Pearson's test after log transformation of non-parametric data. The present study was considered hypothesis-generating so a power calculation was not performed.

Demographics

In total, 20 patients with IPF and eight healthy controls were recruited. A summary of patient characteristics is shown in table 1⇓. Eight patients agreed to a sequential BAL 3–6 months after starting treatment with prednisolone 0.5 mg·kg⁻¹ and azathioprine 1–1.5 mg·kg⁻¹ with or without N-acetylcysteine 600 mg three times daily. Patients had to be free from clinical suggestion of pulmonary infection in the 4 weeks prior to bronchoscopy.

Of the 20 IPF patients enrolled in the present study, 12 died during the 3-yr period of follow-up. The cause of death was deterioration of lung function causing respiratory failure (n = 5), lung carcinoma (n = 3), pneumonic infection (n = 2), pulmonary embolus (n = 1) and myocardial infarction (n = 1). One patient underwent lung transplantation during the follow-up period. Median time to death was 18 months from time of first bronchoscopy.

Elevated BALF macrophage and neutrophil counts in patients with IPF

IPF BALF cell count (median 1,260,000 cells·mL⁻¹ (IQR 560,000–3,220,000 cells·mL⁻¹)) was significantly increased compared with normal BALF cell count (412,000 cells·mL⁻¹ (253,000–470,000 cells·mL⁻¹); p = 0.0034). A significantly elevated total alveolar macrophage (AM) count was noted (IPF: 1,030,000 AMs·mL⁻¹ BALF (446,000–2,740,000 AMs·mL⁻¹ BALF); normal: 391,000 AMs·mL⁻¹ (245,000–450,000 AMs·mL⁻¹); p = 0.0051). BALF % macrophage count was reduced (IPF: % AM 88.3 (73.2–94.1); normal BAL: % AM 96 (95.5–97.7); p = 0.031).

There was an increase in IPF BALF total and % neutrophil count compared with normal BALF counts. Total IPF neutrophil count was 120,000 cells·mL⁻¹ (56,500–279,000 cells·mL⁻¹) and normal total neutrophil count was 400 cells·mL⁻¹ (52.5–956 cells·mL⁻¹); p = 0.0373. IPF BALF % neutrophil count was 11.7% (5.9–26.8%) and normal was 2% (0.75–2.75%); p = 0.016. There was an increase in % neutrophils (but not total neutrophils or macrophages) in patients who died during the follow-up phase (in those who died: 25.0% neutrophils (13–62% neutrophils); in those who survived: 7.9% neutrophils; p = 0.05), as has recently been reported 3. There was no significant difference in BALF lymphocyte or eosinophil count between normal and IPF BALF in the cohort of the present study.

Elevated MMPs in patients with IPF

MMP-7, -8 and -9 represent the majority of MMP protein in IPF BALF (fig. 1⇓). MMP-2, -3,- 7,- 8 and -9 levels were significantly elevated compared with those in normal control patients (table 2⇓). In the eight patients who underwent sequential bronchoscopy after treatment, levels of MMP-2, -3, -7, -8 and -9 remained elevated (table 2⇓). BALF levels of MMP-12 and -13 were largely undetectable in both normal and IPF BALF, and were not further demonstrated. MMP levels did not correlate with age, sex, smoking status or pack-yr history (data not shown).

• Download figure
• Open in new tab
• Download powerpoint

Fig. 1—

Bronchoalveolar lavage fluid (BALF) matrix metalloproteinase (MMP) levels expressed as a scatter plot. BALF data were combined from baseline and follow-up bronchoscopy (n = 28). The scatter plot demonstrates that the majority of BALF MMP protein is MMP-7, -8 and -9.

Elevated BALF MMP levels in those who have rapid progression of lung function 1 yr after presentation

BALF levels of MMP- 3, -7, -8 and -9 were elevated in those patients who had a rapid decline in lung function (n = 8) compared with non-rapid progressing cases (n = 12). However, only the differences between MMP-8 and -9 were significant (table 3⇓). There were no important relationships of IPF BALF MMP-1, -2, -8, -12 or -13 with FVC, T_L,CO or transfer coefficient of the lung for carbon monoxide. BALF MMP-7 correlated weakly with T_L,CO (r = 0.28, p = 0.008) and MMP-3 correlated with FVC (r = 0.20, p = 0.048).

Elevated MMP levels in those who died during follow-up

The present cohort of patients had a mortality level of 40% during the 3 yrs of follow-up. BALF levels of MMP-3, -8 and -9 were elevated in those who subsequently died during the follow-up phase (table 4⇓). MMP levels in those who died did not predict the time to death.

Elevated BALF TIMP-1 and -2 in patients with IPF

TIMP-1 and -2 levels were elevated in first and sequential BALF compared to normal. Median normal BALF TIMP-1 was 505 pg·mL⁻¹ (IQR 290–1,062 pg·mL⁻¹) versus 7,343 pg·mL⁻¹ (1,830–18,668 pg·mL⁻¹) for IPF BALF TIMP-1 (p = 0.0044). BALF TIMP-1 remained elevated in those IPF patients who had treatment (n = 8; 9,725 pg·mL⁻¹ (2,500–20,230 pg·mL⁻¹); p = 0.0022) compared to normal (fig. 2⇓).

• Download figure
• Open in new tab
• Download powerpoint

Fig. 2—

Boxplots of bronchoalveolar lavage fluid (BALF) levels of a) tissue inhibitor of metalloproteinase (TIMP)-1 and b) TIMP-2 are persistently elevated in patients with idiopathic pulmonary fibrosis (IPF), both at baseline (BAL1) and at follow-up bronchoscopy (BAL2). The horizontal bars intersecting the boxes represent the median and the boxes represent the interquartile range; whiskers show the minimum–maximum range. ns: nonsignificant. ^#: p = 0.0044; ^¶: p = 0.0022; ⁺: p = 0.0046; ^§: p = 0.024; ^ƒ: p = 0.49.

Initial IPF BALF TIMP-2 levels (898 pg·mL⁻¹ (460–1,836 pg·mL⁻¹); p = 0.0046) and IPF repeat BALF TIMP-2 levels (1,623 pg·mL⁻¹ (606–2,296 pg·mL⁻¹); p = 0.0022) were elevated compared to normal (276 pg·mL⁻¹ (43–764 pg·mL⁻¹); fig. 2b⇑). There were no relationships between TIMP-1 or -2 levels and static lung function at diagnosis. BALF TIMP-1 and -2 levels were elevated in those who died during follow-up compared to survivors (table 4⇑).

Molar ratios of MMP: altered TIMP in IPF

Calculation of the MMP/TIMP-1 molar ratio suggested that there was no difference between MMP-1 and -2 and normal. MMP-3 molar ratio was significantly reduced (IPF 0.00304 versus normal 0.02; p = 0.0066). In contrast, molar ratios were increased for MMP-7 (IPF 0.89 versus normal 0.25; p = 0.0374), MMP-8 (IPF 0.417 versus normal 0.092; p = 0.0078) and MMP-9 (IPF 0.973 versus normal 0.171; p = 0.018). MMP/TIMP-2 ratios were not elevated compared to normal for MMP-1, -2 and -3, but were significantly elevated for MMP-7, -8 and -9 (data not shown).

In order to confirm that protein levels reflected activity, BALF MMP activity assays were performed for MMP-9. MMP-9 was chosen as it is the best validated assay for MMP activity.

Elevated MMP activity in patients with IPF

Using zymography, MMP-9 was detected in 19 out of 20 patients with IPF and one normal control patient. A representative gel is shown in figure 3⇓. Fluorimetric MMP-9 activity was 23,791 pg·mL⁻¹ (5,145–62,787 pg·mL⁻¹) in IPF BALF compared to those below the detection limit in seven out of eight samples (median allocated value 125 pg·mL⁻¹; p = 0.000). The addition of APMA increased the activity of MMP-9 in IPF BALF from 23,791 pg·mL⁻¹ to 52,529 pg·mL⁻¹ (11,728–134,631 pg·mL⁻¹; p = 0.14; fig. 4⇓).

• Download figure
• Open in new tab
• Download powerpoint

Fig. 3—

Representative gels (using zymography) demonstrating increased matrix metalloproteinase (MMP) activity in a) idiopathic pulmonary fibrosis (IPF) bronchoalveolar lavage fluid (BALF) compared to b) normal.

• Download figure
• Open in new tab
• Download powerpoint

Fig. 4—

Bronchoalveolar lavage fluid (BALF) matrix metalloproteinase (MMP)-9 activity assay. 4-aminophenylmercuric acetate (APMA) maximally activates MMP-9. IPF: idiopathic pulmonary fibrosis. The horizontal bars intersecting the boxes represent the median and the boxes represent the interquartile range; whiskers show the minimum–maximum range. •: individual outliers. +: with APMA activation; -: without APMA activation. ^#: p = 0.0003; ^¶: p = 0.0001.

BALF VEGF levels in IPF

VEGF levels in IPF patients were not significantly different in patients with IPF compared to normal (82 pg·mL⁻¹ (25–222 pg·mL⁻¹) versus 140 pg·mL⁻¹; p = 0.499; fig. 5a⇓). VEGF levels did not correlate with baseline lung function. Initial IPF BALF levels of VEGF correlated significantly with both protein permeability index (r = 0.51, p = 0.026) and with BALF MMP-3 (r = 0.53, p = 0.003), MMP-7 (r = 0.36, p = 0.02), MMP-8 (r = 0.38, p = 0.018) and MMP-9 (r = 0.34, p = 0.029; data not shown).

• Download figure
• Open in new tab
• Download powerpoint

Fig. 5—

Bronchoalveolar lavage fluid (BALF) vascular endothelial growth factor (VEGF) levels in a) normal BALF and idiopathic pulmonary fibrosis (IPF) BALF, b) rapid IPF and non-rapid IPF, and c) IPF subjects who died and subjects who survived. a) BALF VEGF does not differ between normal and IPF patients (all 20 considered at baseline). c) BALF VEGF is greater in IPF patients who died during follow-up rather than survived. The horizontal bars intersecting the boxes represent the median and the boxes represent the interquartile range; whiskers show the minimum–maximum range. ^#: p = 0.499; ^¶: p = 0.04; ⁺: p = 0.03.

Baseline BALF VEGF levels were elevated in patients who had a rapid decline in lung function over the first year of follow-up (188 pg·mL⁻¹ (56–301.7 pg·mL⁻¹)) compared to IPF non-progressors (38.4 pg·mL⁻¹ (16.8–152.1 pg·mL⁻¹); p = 0.04; fig. 5b⇑). Baseline BALF VEGF levels were also elevated in patients who died (159 pg·mL⁻¹ (62–272 pg·mL⁻¹) compared to those who survived (34 pg·mL⁻¹ (13.9–70 pg·mL⁻¹); p = 0.03; fig. 5c⇑).

Elevated PPI in patients with IPF

IPF BALF PPI was elevated (0.00439 (0.000853–0.0148)) compared to normal (0.00196 (0.000714–0.0026); p = 0.044; fig. 6⇓). Levels of PPI were greatest in those who died (0.01594) within the 3 yrs of follow-up (p = 0.04; table 4⇑ and fig. 6b⇓). In contrast, neither plasma protein levels nor BALF total protein had any relationship with outcome (table 4⇑). PPI did not differ significantly in those with a rapid decline in lung function (data not shown).

• Download figure
• Open in new tab
• Download powerpoint

Fig. 6—

Protein permeability index (PPI) in bronchoalveolar lavage fluid (BALF) showed an increase in idiopathic pulmonary fibrosis (IPF) patients compared to normal. IPF patients who died during follow-up showed significantly elevated PPI compared to survivors. The horizontal bars intersecting the boxes represent the median and the boxes represent the interquartile range; whiskers show the minimum–maximum range. •: individual outliers. ^#: p = 0.04.