In patients with IPF, serum biomarkers of type I and III collagen turnover measured at the time of diagnosis is related to disease progression within one year. Moreover, we demonstrated that longitudinal levels of type I and III collagen turnover could distinguish patients with progressive disease in a real-world cohort; however, we did not detect any change in collagen turnover biomarkers in patients treated with antifibrotics.
Type I and III collagen are located in the interstitial matrix and are essential components of a healthy lung by providing tensile strength to the alveolar interstitium.24 IPF results from increased levels of interstitial collagens, changing the architecture of the airspaces in the lungs,25–27 and elevated type I and III collagen turnover has been shown to be related to disease progression in IPF patients.15,16 Recently, we demonstrated that type III collagen degradation (C3M) and formation (PRO-C3) are associated with disease severity at baseline and high PRO-C3 levels were related to survival.28 In the current study, we investigated whether biomarkers of type I and III collagen turnover measured at baseline could be used to predict risk of disease progression within 12 months. When we calculated odds ratios for all IPF patients, we found that patients belonging to the middle and upper tertiles of C1M or PRO-C3 had relatively higher risk of disease progression than patients in the lower baseline tertile. These findings suggest that measuring formation of type III collagen or degradation of type I collagen at baseline could identify patients at risk of progression and help clinicians in treatment choice and timing. Currently, there is no biomarker or clinical score to predict or identify disease progression in IPF patients treated with antifibrotic therapy. Our findings suggest that future studies of type I and III collagen turnover as prognostic biomarkers in IPF are warranted. A previous study examined whether certain biomarkers (CCL13, CCL18, CXCL13, COMP, periostin, and YKL40) measured at baseline in the CAPACITY and ASCEND trials could predict FVC changes in patients receiving pirfenidone.29 There was no association between baseline biomarker levels and change in FVC among patients treated with pirfenidone. A common limitation in biomarker studies is the lack of comparison of the prognostic performance of different biomarkers in the same population. Since our data suggest that patients with a higher baseline type I and III collagen turnover have a higher risk of disease progression, it would be interesting to make a direct comparison of C1M and PRO-C3 with the above-mentioned biomarkers to verify if they could perform better.
Interestingly, we have recently shown that degradation of type III collagen (C3M) was related to disease severity at the time of diagnosis.28 However, in the present analyses, it was not related to increased risk of progression, indicating that C3M is more associated with disease severity rather than being a prognostic biomarker in this population. However, previous studies indicate that C3M may be a prognostic biomarker in patients with more progressive disease.15
Next, we investigated longitudinal biomarker levels in stable and progressive IPF patients and showed that degradation of type I collagen and formation of type III collagen was elevated in patients with progressive IPF compared to stable IPF over a one year period. This is in line with results from a previous study where C1M and PRO-C3 levels were elevated at several time points from baseline to six months.15,16 In the present study, elevated levels of C1M and PRO-C3 were also measured at 12 months and this confirms and expands on previous findings that degradation of type I collagen and formation of type III collagen are associated with disease progression in a real-world cohort of patient with IPF and moderate disease severity. The present analyses used a less strict definition of progression to include patients with marginal physiological progression of disease, which have been shown to predict mortality in IPF patients.30–32 Moreover, we did not find any difference in longitudinal levels of type III collagen degradation (C3M) between progressive and stable IPF patients, also in line with previous findings.16 In contrast, it has been shown that C3M levels were increased in progressive patients.15 The differences between the studies could be due to differences in baseline demographics among the patients recruited to the cohorts. In addition, patients in the British cohort did not receive antifibrotic treatment, while most patients in the PFBIO cohort were treated with pirfenidone or nintedanib after the baseline visit. We find it interesting that a degradation fragment of type I collagen was related to disease progression while a degradation fragment of type III collagen was not. This may be explained by the fact that C1M is generated with different matrix metalloproteinases (MMP-2, -9 and -13) and C3M is only generated with MMP-9, what may clarify the reason why C1M is related to progression and C3M is not.
Finally, we compared percentage change from baseline of type I and III collagen turnover in patients receiving antifibrotic therapy or not. Surprisingly, we did not find any differences in biomarker levels at any time points. Furthermore, we did not find any differences in biomarker levels between patients treated with nintedanib or pirfenidone. This could indicate that type I and III collagen turnover is not influenced by current antifibrotic treatments, yet the results need to be confirmed in other studies with a broader group of untreated patients. Patients without antifibrotic therapy in the present study were a highly selected group with higher age and shorter 6MWT distance compared to patients on antifibrotic therapy and this could potentially influence the results. However, our results are in line with the INMARK study where C1M and C3M were not modulated by nintedanib.33 The patients included in the INMARK study and the PFBIO cohort had mild disease, however, the placebo group in the INMARK study showed a decline in FVC, while most patients from the PFBIO cohort had stable FVC over 12 months. In contrast to pirfenidone and nintedanib, it was shown in a small proof of concept study that treatment with omipalasib in IPF patients reduced the levels of type III collagen formation.34 This potential antifibrotic treatment has a different mode of action in comparison to nintedanib and pirfenidone, and it is tempting to consider combinations of antifibrotics targeting different pathways of the fibrotic cascade.
A clear strength of this study is the prospective, longitudinal design which makes it possible to associate biomarkers with disease progression over time. Another advantage is the large number of participants. Patients were included in PFBIO from two of Denmark's specialized ILD centers, and this makes the study representative of a broad IPF population. In addition, a significant strength is that the collection of blood samples was standardized to minimize variation between samples. On the other hand, a weakness of the current study is that analyses were only performed in a single cohort with no validation cohort. The pulmonary function was stable for over one year in treated and untreated patients, and this may explain why the biomarker was stable over 12 months. In addition, it would be interesting to follow patients for more than a year.
In conclusion, type I and III collagen turnover measured at the time of diagnosis is related to disease progression within one year in patients with IPF. Moreover, longitudinal serum levels of type I and III collagen turnover appear to distinguish IPF patients with progressive disease from patients with stable disease. This could provide valuable information for clinicians to identify patients at high risk of disease progression. In addition, treatment with the antifibrotics nintedanib and pirfenidone does not affect serum biomarkers of type I and III collagen turnover in this population.