Circulating biomarkers in patients with progressive fibrosing interstitial lung disease treated with nintedanib: a pilot study

Abstract

Progressive fibrosing interstitial lung diseases (PF-ILDs) are characterized by persistent progression and have limited treatment options. The identification of reliable biomarkers to monitor fibrosis and therapeutic response remains a clinical challenge. This study investigated circulating plasma biomarkers associated with PF-ILDs and their potential role in monitoring disease evolution during nintedanib treatment. From 127 putative fibrosis biomarkers, seven candidates were identified with high diagnostic value (area under the curve [AUC] > 0.7), of which five (IGFBP2, PTX3, LGALS1, LGALS9, and MMP2) showed significant dynamic changes (assessed by longitudinal plasma proteomic analysis) in PF-ILD patients treated with 12-months nintedanib, correlating with improvements in forced vital capacity and diffusing capacity of the lung for carbon monoxide. Principal component analysis identified a shift in molecular profiles over time, suggesting nintedanib-induced modulation of these biomarkers. Receiver operating characteristic analysis demonstrated that while LGALS9 maintained a stable predictive value during nintedanib treatment, LGALS1, IGFBP2, PTX3, and MMP2 exhibited increasing AUC scores, indicating their potential role in monitoring fibrosis progression. We also identified optimal biomarker cut-off values at 12 months, which may provide reliable thresholds for fibrosis assessment. In conclusion, our exploratory analysis identified five biomarkers whose plasma concentrations changed during antifibrotic treatment, highlighting their potential prognostic value. Further validation in larger cohorts is needed to confirm their clinical utility.

Introduction

Interstitial lung diseases (ILDs) represent a heterogeneous group of disorders characterized by inflammation and fibrosis of the lung, leading to impaired gas exchange and a progressive decline in lung function¹. Among ILDs, idiopathic pulmonary fibrosis (IPF) represents a particularly severe and rapidly progressive subtype, with a median survival of only 4.5 years after diagnosis². However, beyond IPF, there is a subset of patients with other chronic fibrosing ILDs who also exhibit a progressive phenotype, marked by worsening fibrosis, declining lung function, and reduced quality of life^3,4. The recognition of this progressive fibrosing phenotype across various ILDs has generated interest in elucidating common pathobiological mechanisms underlying disease progression^{5, 6–7}. Despite differences in etiology and clinical manifestations, these ILDs share key features such as dysregulated tissue repair and aberrant fibroblast activation, ultimately leading to excessive extracellular matrix (ECM) deposition and irreversible fibrotic remodeling of the lung parenchyma^{8, 9, 10, 11–12}. Understanding the molecular pathways driving fibrosis progression in these diverse ILDs holds promise for identifying novel therapeutic targets and developing more effective treatment strategies.

Nintedanib, an intracellular tyrosine kinase inhibitor initially developed for anticancer therapy, has emerged as a promising treatment for managing fibrotic lung diseases^{13, 14, 15–16}. Preclinical studies have highlighted its ability to attenuate fibrosis progression by targeting key signaling pathways involved in fibroblast activation, ECM deposition, and vascular remodeling¹⁷. Clinical trials in patients with IPF and other progressive fibrosing ILDs (PF-ILDs) have demonstrated that nintedanib treatment significantly slows the decline in lung function, reduces the frequency of acute exacerbations, and improves overall clinical outcomes^15,16. However, despite its efficacy in slowing disease progression, the identification of reliable biomarkers to monitor treatment response and predict long-term outcomes remains a critical need in the field of ILD research. A recent study by Jenkins and colleagues¹⁸ investigated the effects of nintedanib on circulating biomarkers in IPF, highlighting its impact on specific molecular pathways involved in disease progression, and demonstrated that nintedanib modulates key fibrotic biomarkers, providing valuable insights into its mechanism of action and potential for monitoring therapeutic responses. However, further investigations are needed to identify additional biomarkers that can effectively track fibrosis progression and therapeutic efficacy across different ILDs, including those with a progressive fibrosing phenotype. Traditional clinical parameters, such as forced vital capacity (FVC), diffusing capacity of the lung for carbon monoxide (DLCO), and high-resolution computed tomography scores, have limitations in predicting disease progression and guiding therapeutic decisions^{19, 20–21}. On the other hand, specific biomarkers may offer the potential for a non-invasive and objective assessment of disease activity, enabling the early detection of disease progression and the development of personalized treatment approaches.

An increasing number of genomic approaches are being developed for the diagnosis and prognosis of ILDs^22,23, and a better understanding of the underlying mechanisms involved, as well as the discovery of new biomarkers, currently represent the major clinical challenge. Proteomic investigations have also emerged as a powerful tool for identifying biomarkers associated with disease pathogenesis and treatment response in ILDs^{24, 25–26}. By analyzing the expression profiles of proteins involved in fibrotic pathways, proteomic studies can uncover novel molecular signatures that reflect disease severity, predict treatment response, and guide therapeutic decisions²⁷. Moreover, proteomic profiling of patient samples offers the opportunity to discover biomarkers that overcome diagnostic confines and could apply to different ILD subtypes, allowing patient stratification^26,27.

In this study, we aimed to identify protein biomarkers of pulmonary fibrosis in patients with PF-ILDs treated with nintedanib. Our central hypothesis was that plasma biomarkers associated with fibrosis progression can exhibit robust associations with disease severity across diverse ILD subtypes. The present study aimed to investigate the predictive role of protein plasma biomarker levels in the context of PF-ILDs, in which disease progression is slowed when patients are treated with nintedanib. We hypothesized that plasma biomarkers would show a strong association with the PF-ILD phenotype, making them valuable for clinical diagnosis, prognosis, and disease monitoring.

Methods

Study population and sample collection

Nineteen patients (8 males, 64.15 ± 12.20 years; 11 females, 72.23 ± 7.80 years) were enrolled in the study (Table 1). All patients had a definitive diagnosis of an ILD according to the ATS/ERS/JRS/ALAT Guidelines²⁸. They were selected for nintedanib treatment according to the Italian national drug inclusion criteria and had not been previously treated with other antifibrotic drugs. All patients tolerated nintedanib therapy at a stable full dose of 150 mg twice a day for a minimum of 3 months. Following this period, five patients discontinued the study due to adverse drug reactions. Lung function tests were performed to obtain FVC and DLCO values. The assessment of clinical progression was performed by measuring FVC and DLCO on a regular basis (pre-treatment and 3, 6, and 12 months of treatment). The study was approved by the IRCCS ISMETT’s Institutional Research Review Board supervisors and the relevant local ethics committee (project number: IRRB/08/22), and all patients gave written informed consent to participate in the study. Additionally, six healthy individuals (3 males, 50.66 ± 2.52 years; 3 females, 46.66 ± 5.03 years) were included as controls (Table 1).

Table 1. Demographic and clinical characteristics of the our patient cohort at baseline and after 3, 6, and 12 months (M) of nintedanib treatment.

Characteristics	Baseline (n = 19)	Nintedanib
		3 M (n = 19)	6 M (n = 14)	12 M (n = 14)
Healthy controls (n = 6)
Age, years	48.66 ± 4.18	-	-	-
Gender, n. (%)
Female	3 (50)	-	-	-
Age, years	46.66 ± 5.03	-	-	-
Male	3 (50)	-	-	-
Age, years	50.66 ± 2.52	-	-	-
Patients (n = 19)
Age, years	68.99 ± 10.32	-	-	-
Gender, n. (%)
Female	11 (58)	-	-	-
Age, years	72.23 ± 7.80	-	-	-
Male	8 (42)	-	-	-
Age, years	64.15 ± 12.20	-	-	-
Body mass index (kg/m2)	30.01 ± 4.94	-	-	-
Smoking status, n. (%)
Yes	9 (47)	-	-	-
No	10 (53)	-	-	-
Primary disease, n. (%)
SSc-ILD	1 (5.3)	-	-	-
NSIP	3 (15.6)	-	-	-
RA-ILD	4 (21)	-	-	-
ASyS-ILD	1 (5.3)	-	-	-
CTD-ILD	1 (5.3)	-	-	-
HP-ILD	5 (26.3)	-	-	-
DIP-ILD	1 (5.3)	-	-	-
Other ILD	3 (15.9)	-	-	-
Most frequent comorbidities, n. (%)
Chronic atrial fibrillation	1 (5.3)	-	-	-
Chronic obstr. bronchopathy	3 (15.6)	-	-	-
Arterial hypertension	9 (47)	-	-	-
Obstr. Sleep Apnea Syndrome	1 (5.3)	-	-	-
Chronic renal failure	1 (5.3)	-	-	-
Asthma	1 (5.3)	-	-	-
Oncology	1 (5.3)	-	-	-
FVC
mL	1910 ± 155	1923 ± 641	1918 ± 799	2031 ± 180
% of predicted	73.21 ± 11.8	69.16 ± 10.45	77.21 ± 16.87	82.64 ± 13.11
DLCO
mL	8.43 ± 1.92	8.58 ± 2.13	6.95 ± 2.16	9.13 ± 1.55
% of predicted	40.63 ± 3.59	38.33 ± 3.25	33.71 ± 8.43	38.00 ± 5.89

Unless otherwise stated, values are n (%) or mean ± standard deviation (SD).

SSc-ILD, interstitial lung disease associated with systemic sclerosis; NSIP, non-specific interstitial pneumonia; RA-ILD, rheumatoid arthritis-associated interstitial lung disease; ASyS-ILD, antisynthetase syndrome-associated interstitial lung disease; CTD-ILD, connective tissue disease-associated interstitial lung disease; HP-ILD, chronic hypersensitivity pneumonitis-associated interstitial lung disease; DIP-ILD, desquamative interstitial pneumonia-associated interstitial lung disease; DLCO, diffusing capacity of the lung for carbon monoxide; FVC, forced vital capacity; ILD, interstitial lung disease.

Peripheral blood was collected from selected patients at time 0, before starting nintedanib, and after 3, 6, and 12 months of therapy. After isolation, the plasma was aliquoted and stored at − 80 °C until analysis.

Data acquisition and biomarker screening

To identify the most suitable putative serum biomarkers for lung fibrosis, we retrieved data from the Gene Expression Omnibus repository (https://www.ncbi.nlm.nih.gov/gds) for both ILD and healthy samples. Two datasets containing information on both healthy individuals and ILD patients were used: gene expression data of five healthy subjects and 23 ILD patients (with variable disease severity status based on FVC values) taken from the GSE21411 dataset, and gene expression data of five healthy subjects and 10 ILD patients (with variable disease severity status based on DLCO values) taken from the GSE24206 dataset. The clinical features of these cohorts are summarized in Supplementary Table S1. We identified dysregulated biomarkers in ILD subjects with variable disease severity status using volcano plots (p < 0.05 and fold change > 1.5) and further filtered the biomarkers using receiver operating characteristic (ROC) analysis (area under the curve [AUC] > 0.7).

Biomarker analysis

The plasma concentration of selected biomarkers (CCL18, IGFBP2, PTX3, LGALS1, LGALS9, MMP2, and MMP7) was measured in patients treated for progressive pulmonary fibrosis using magnetic bead technology from Luminex with the ProcartaPlex Human Magnetic Luminex Kits (Affymetrix, Austria) at different time points from the first dose administration (0, 3, 6, and 12 months). Assays were conducted according to the manufacturer’s instructions, and the results were obtained from Luminex 200 Systems (Luminex Corporation, USA). Fluorescence intensity data from the assays were used for the analysis. Hierarchical cluster analysis and principal component analysis (PCA) (ClustVis web tool)²⁹ of protein expression were used to group patients based on treatment time (0, 3, 6, and 12 months).

Statistical analysis

Statistical analysis was performed using GraphPad Prism 6.0 (GraphPad Software, USA). A one-way ANOVA with the Tukey multiple-comparison test was used for statistical comparison, and p-values < 0.05 were considered statistically significant. The DATAtab web tool (https://datatab.net) was used to analyze the data and draw the ROC curve. We used Epitools Epidemiological Calculators (http://epitools.ausvet.com.au.) to determine cut-off values, sensitivity, and specificity.

Results

Screening of potential biomarkers of progressive fibrosing ILDs

The flow chart illustrating the first study phase is depicted in Fig. 1. We selected 127 potential biomarkers of pulmonary fibrosis (Supplementary Table S2) from the scientific literature, which have been found to be dysregulated in samples from patients with ILDs, and for which strong scientific evidence supports their clinical role^{30, 31, 32, 33, 34, 35, 36, 37–38}.

[See PDF for image]

Fig. 1

Flow chart illustrating the steps for biomarker screening.

We retrospectively evaluated these biomarkers in two independent cohorts of ILD patients at different stages of disease severity (mild, moderate, and severe pulmonary fibrosis), classified based on functional FVC or DLCO values (Supplementary Table S1). Using a volcano plot analysis, we identified dysregulated biomarkers by comparing severity groups to controls (Supplementary Fig. S1). In cohorts 1 and 2, we identified 32 and 35 dysregulated markers, respectively (Fig. 1). Of these, nine exhibited consistent dysregulation and similar expression changes across severity groups in both cohorts (Supplementary Fig. S2). Given this reproducible expression pattern, we considered these nine markers as potential indicators of progressive fibrosis. To evaluate the diagnostic performance of the selected biomarkers in distinguishing ILD patients from healthy controls, we conducted ROC curve analysis. Of the nine biomarkers initially analyzed, seven (CCL18, IGFBP2, PTX3, LGALS1, LGALS9, MMP2, and MMP7) showed an AUC > 0.7, indicating good discriminatory power (Fig. 2). These seven biomarkers were selected for further validation within our patient cohort. A detailed description of biomarker screening in both cohorts is provided in Appendix A of the Supplementary Materials.

[See PDF for image]

Fig. 2

Receiver operating characteristic (ROC) curves of the biomarkers showing similar trends of expression in cohort 1 and cohort 2. The analysis was conducted on 10 healthy subjects and 33 ILD patients (GSE24206 and GSE21411 datasets).

Characteristics of our patient cohort

The demographic and clinical characteristics of our patient cohort with ILDs treated with nintedanib are summarized in Table 1. A total of 19 patients received nintedanib. Regarding the baseline characteristics of the patients, the mean age was 68.99 ± 10.32 years, and the body mass index was 30.01 ± 4.94. The most frequent diagnoses of ILD were chronic hypersensitivity pneumonitis-associated ILD (in 26.3% of patients), rheumatoid arthritis-associated ILD (in 21% of patients), and nonspecific interstitial pneumonia (in 15.6% of patients). Moreover, the most frequent comorbidities were arterial hypertension (47%) and chronic obstructive bronchopathy (15.6%). At baseline, the FVC was 73.21 ± 11.8%, and DLCO was 40.63 ± 3.59%.

Analysis of the seven selected protein biomarkers in plasma samples of our patient cohort

As explained above, to identify candidate biomarkers for fibrosis progression, we performed a stepwise data filtering of 127 putative fibrosis biomarkers based on differential gene expression in two independent cohorts of ILD patients, which included individuals with varying disease severity. We selected seven biomarkers (CCL18, IGFBP2, PTX3, LGALS1, LGALS9, MMP2, and MMP7) associated with progressive fibrosis for analysis in our patient cohort. In these patients, treatment with nintedanib appeared to slow the progression of ILD. In particular, Fig. 3 illustrates the changes in pulmonary function parameters, FVC and DLCO, before and after nintedanib treatment. Panels A and B display the mean changes of percentage and absolute values of FVC, respectively, while panels C and D show the corresponding DLCO variations. In panel A, FVC (%) shows a decline from 6 months before treatment (6 M Pre) to 3 months post-treatment initiation (3 M Post), followed by a progressive improvement after 6 and 12 months, with a statistically significant change at 12 months compared with baseline (0 M). Similarly, panel B shows a significant initial reduction in FVC (mL) from 6 months before treatment to baseline, followed by stabilization at 3 and 6 months post-treatment, and a significant recovery at 12 months. Panels C and D demonstrate the trends in DLCO, indicating a gradual decline in DLCO (%) and DLCO (mL) before treatment, with the most pronounced decrease occurring between 6 months before and 6 months post-treatment, and with a statistically significant change at 6 months post-treatment compared with baseline. After treatment initiation, DLCO appears to stabilize with a slight improvement at 12 months post-treatment (Table 1; Fig. 3).

[See PDF for image]

Fig. 3

Decline from baseline in forced vital capacity (FVC) and diffusing capacity of the lung for carbon monoxide (DLCO). (A) and (B) mean changes from 6 M Pre in FVC over the 12-month trial period in the overall population. (C) and (D) mean changes from 6 M Pre in DLCO over the 12-month trial period in the overall population. The analysis included 19 ILD patients assessed at 6 months before treatment (6 M Pre), baseline (0 M), and 3 months post-treatment initiation (3 M Post), and 14 ILD patients evaluated at 6 months (6 M Post) and 12 months (12 M Post) after starting nintedanib. *p < 0.05 vs. baseline (0 M). Bars indicate the standard deviation.

Among the selected biomarkers, five proteins (IGFBP2, PTX3, LGALS1, LGALS9, and MMP2) showed significant changes in expression during the 12-month nintedanib treatment period (Fig. 4A and B). The box plots of these proteins highlight their pattern of expression in the ILD patients and healthy subjects (Fig. 4C). Specifically, all proteins were more abundant in ILD samples compared with healthy subjects (p < 0.05). LGALS1, LGALS9, and PTX3 decreased progressively from month 0 to month 12 of nintedanib treatment, whereas IGFBP2 and MMP2 increased gradually (Fig. 4C). Hierarchical clustering analysis and PCA of these five markers showed a clear separation between healthy subjects and ILD patients. After 12 months of nintedanib treatment, the biomarker spatial distribution of ILD patients in PCA returned to a non-fibrotic pattern, similar to that of healthy subjects (Fig. 5). These preliminary results, if confirmed, suggest that these five biomarkers may play a role in predicting the response to antifibrotic treatment. Therefore, we performed ROC analysis to further evaluate the predictive value of these biomarkers. Specifically, we evaluated the predictive power of each biomarker at each treatment time point (3, 6, and 12 months). Interestingly, while we detected a good and stable AUC score for LGALS9 after 3, 6, and 12 months (0.711, 0.707, and 0.658, respectively), a gradual increase of AUC score was observed for LGALS1 (0.503, 0.722, and 0.883, respectively), IGFBP2 (0.695, 0.771, and 0.842, respectively), PTX3 (0.709, 0.870, and 0.842, respectively), and MMP2 (0.571, 0.718, and 0.759, respectively), showing an association with treatment duration, during which a slowing of fibrosis progression was observed (Fig. 6).

[See PDF for image]

Fig. 4

(A) Cluster analysis to compare the plasma protein expression of the seven selected biomarkers during the treatment period with nintedanib (12 months). (B) Expression trend of the biomarkers during the treatment period. (C) Box plots of the abundance of biomarkers that showed variations in patients’ plasma during nintedanib treatment. The analysis included 6 healthy subjects (Healthy), 19 ILD patients assessed at baseline (M0), and 3 months post-treatment initiation (M3), and 14 ILD patients evaluated at 6 months (M6) and 12 months (M12) after starting nintedanib. ^#p < 0.05 vs. healthy subjects. *p < 0.05 vs. M0.

[See PDF for image]

Fig. 5

Cluster and principal component analysis (PCA) of selected biomarkers which showed significant variation in expression during the nintedanib treatment. The analysis included 6 healthy subjects (NO FIB), 19 ILD patients assessed at baseline (M0), and 3 months post-treatment initiation (M3), and 14 ILD patients evaluated at 6 months (M6) and 12 months (M12) after starting nintedanib.

[See PDF for image]

Fig. 6

Receiver operating characteristic (ROC) curves of the biomarkers which showed predictive values during treatment with nintedanib. The analysis was conducted on 19 ILD patients at 3-month nintedanib treatment vs. 19 ILD patients at baseline (3 Months), 14 ILD patients at 6-month nintedanib treatment vs. 19 ILD patients at baseline (6 Months), and 14 ILD patients at 12-month nintedanib treatment vs. 19 ILD patients at baseline (12 Months).

We then determined the optimal cut-off values, sensitivity, and specificity for these biomarkers at 12 months of treatment, a time point at which functional parameters such as FVC and DLCO showed significant improvement compared with baseline (month 0). Specifically, for LGALS1, the optimal cut-off value was 354 pg/mL, yielding a sensitivity of 0.74 and a specificity of 0.86. For LGALS9, the optimal cut-off value was 1191 pg/mL, with a sensitivity of 0.53 and a specificity of 0.86. For IGFBP2, the optimal cut-off value was 16,003 pg/mL, yielding a sensitivity of 0.71 and a specificity of 0.95. For PTX3, the optimal cut-off value was 126.75 pg/mL, with a sensitivity of 0.58 and a specificity of 1. For MMP2, the optimal cut-off value was 11,189.76 pg/mL, yielding a sensitivity of 0.5 and a specificity of 1.

Discussion

ILD-related injury can initiate a fibrotic process. A chain of events can be triggered in this case, leading to disease progression³⁹. The onset and evolution of these diseases involve intricate interactions among various pathogenic factors, with fibrosis assuming a pivotal role in its later stages⁴⁰. Chronic fibrosing ILDs with a progressive phenotype show similarities in clinical behavior and pathogenic mechanisms that drive disease progression⁴. These conditions pose a significant challenge, marked by their persistent advance in afflicted individuals and the lack of conventional treatment approaches^41,42. Identifying potent noninvasive biomarkers holds promise in providing crucial diagnostic insights and unraveling the trajectory of fibrotic processes. Exploiting such biomarkers could be crucial in identifying individuals at high risk of pulmonary progressive fibrosis, guiding clinicians in selecting appropriate candidates for pertinent clinical trials where new treatments could be tested. In this regard, nintedanib, an approved treatment for patients with IPF^13,14, has demonstrated antifibrotic effects in several ILD animal models^43,44 and has had significant efficacy in reducing ILD progression in patients with ILD^15,16.

In this work, we identified five circulating plasma biomarkers (IGFBP2, PTX3, LGALS1, LGALS9, and MMP2) potentially linked to PF-ILD in a heterogeneous patient cohort. To do this, we screened 127 putative fibrosis biomarkers by integrating genomic data (from the Gene Expression Omnibus database) with patients of different disease severity status from two independent cohorts. Specifically, we analyzed differential gene expressions related to diverse severity statuses (based on FVC or DLCO values) in both datasets, allowing us to perform an extensive filtration of molecular markers related to different stages of fibrosis progression. From this analysis, we identified nine biomarkers with similar variations in expression across different severity statuses in both cohorts and selected seven biomarkers (CCL18, IGFBP2, PTX3, LGALS1, LGALS9, MMP2, and MMP7) using ROC analysis (AUC > 0.7). To assess the ability of these biomarkers to monitor the progression of fibrosis, we evaluated their diagnostic value (in terms of plasma protein abundance) in ILD patients undergoing nintedanib treatment, in whom fibrosis progression had slowed, by analyzing the proteomic levels of the seven biomarkers both before and after one year of treatment. Our results highlighted these markers as indicative of modifications of molecular pathways induced during treatment. Notably, we found that the plasma levels of five biomarkers (IGFBP2, PTX3, LGALS1, LGALS9, and MMP2) had a specific trend of variation during the various treatment time points.

Previous studies have also demonstrated that plasma concentrations of both LGALS1 and LGALS9 were higher in ILD patients than in healthy controls⁴⁵. In a mouse model of lung fibrosis, LGALS1 was shown to physically interact with FAK1 (a crucial regulator of TGF-β signaling) with consequent activation in lung epithelial cells, contributing to tissue injury and fibrosis⁴⁶. Using the same study model, it was also demonstrated that LGALS9 could increase TGF-β signaling, playing a critical role in fibroblast activation and ECM deposition⁴⁷. Nintedanib is a potent molecule that inhibits receptor tyrosine kinase activities, including TGF-β signaling and focal adhesion kinase proteins^48,49, thereby counteracting the progression of fibrosis. Our data, which showed decreased levels of LGALS1/9, are in line with these findings, at least from a functional perspective.

Recently, Guiot and colleagues demonstrated that the serum level of IGFBP2 was significantly higher in patients with pulmonary fibrosis than in healthy subjects⁵⁰. In contrast to our data, the Guiot study found that treatment with nintedanib reduced the serum level of IGFBP2 after 14.3 months of treatment. While previous studies, such as that by Habeb et al.⁵¹, reported a significant decrease in serum and bronchoalveolar fluid levels of IGFBP2 following 12 months of pirfenidone therapy in usual interstitial pneumonia patients, our findings revealed a progressive increase in circulating IGFBP2 during nintedanib treatment. This discrepancy may be due to differences in the mechanism of action between the two antifibrotic drugs or in patient selection and disease stages. Interestingly, beyond being a marker of disease activity, IGFBP2 may also exert protective roles in pulmonary fibrosis. Recent evidence suggests that IGFBP2 counteracts alveolar epithelial cell senescence and modulates the fibrotic secretome, thereby contributing to epithelial repair and limiting fibrosis progression⁵². Thus, the rise in IGFBP2 levels that we observed might reflect a compensatory or adaptive response induced by nintedanib, rather than a signal of worsening disease. Further mechanistic studies are needed to elucidate the biological function of IGFBP2 in treated IPF.

PTX3 is upregulated in the plasma of fibrosing ILD patients and the lungs of bleomycin-exposed mice and is correlated with both disease severity and adverse outcomes⁵³. Interestingly, PTX3 has been shown to activate lung fibroblasts to differentiate towards collagen‐expressing myofibroblasts, and in PTX3‐deficient mice, it has been shown to decrease fibrosis activity and improve lung function⁵³. These findings agree with our data, which showed that PTX3 abundance was high before nintedanib treatment but decreased after 3–12 months of treatment.

Although the role of MMP2 in pulmonary fibrosis remains unclear⁵⁴, the level of MMP2 is higher in patients with pulmonary fibrosis than in control subjects⁵⁵. Moreover, in a mouse model of lung fibrosis, lung homogenates showed a significant increase in the level of MMP2 compared with controls⁵⁶. MMP2 might possess antifibrotic effects, mainly influencing the degree of collagen deposition and ECM remodeling during fibrosis⁵⁷. Interestingly, this finding is supported by in vitro data from Hostettler and colleagues⁵⁸, who demonstrated that nintedanib significantly enhances pro-MMP2 expression and activity in fibroblasts derived from IPF patients, while concurrently reducing TIMP-2 levels. These molecular effects were interpreted as part of the drug’s anti-fibrotic action, facilitating ECM degradation and turnover. Accordingly, the increase in circulating MMP2 observed in our study may represent a biomarker of treatment-induced ECM remodeling processes. Moreover, MMP-2 activity has been associated with attenuation of fibrosis in preclinical models⁵⁹, suggesting a dual, context-dependent role that may be beneficial under the influence of antifibrotic therapies, such as nintedanib. We also observed an increased level of MMP2 during nintedanib treatment, which, similar to the behavior of IGFBP2, could counteract disease progression.

The antifibrotic and anti-inflammatory properties of nintedanib underscore its value in treating ILDs^{13, 14, 15–16}. Although some of its targets, such as fibroblast growth factor receptor, vascular endothelial growth factor receptor, and platelet-derived growth factor receptor, are well-known, other molecular mechanisms remain to be discovered. Here, our approach identified five proteins whose abundance changed as the progression of fibrosis slowed over the course of one year of nintedanib treatment. Interestingly, in Fig. 5, which visually represents the distribution of patient samples across different time points of nintedanib treatment and healthy controls, the first principal component explains 45% of the variance, while the second principal component accounts for 25.4%, highlighting the key differences in molecular profiles among these groups. Notably, the healthy control group is well separated from the fibrotic patient samples, confirming a distinct molecular signature between healthy individuals and those with fibrosis. At baseline, the distribution was more dispersed, which may indicate greater heterogeneity in molecular profiles before treatment. Over time, with the progression of therapy, the patient samples tended to cluster more closely, indicating a shift in their molecular characteristics, which may potentially reflect treatment-induced changes, while the overlap among M3, M6, and M12 suggests a gradual alignment of molecular profiles, possibly associated with the stabilization of fibrosis progression. This PCA analysis supports the findings that fibrosis-related biomarkers change dynamically in response to treatment, reinforcing their potential role in monitoring disease progression and therapeutic response.

We also observed that the diagnostic values of these five biomarkers were high during nintedanib treatment. Our findings suggest that the predictive power of these biomarkers evolved during nintedanib treatment, with a gradual increase in AUC scores for LGALS1, IGFBP2, PTX3, and MMP2. At the same time, LGALS9 maintained a relatively stable diagnostic performance. This trend suggests that these biomarkers may reflect molecular changes associated with the therapeutic response and the slowing of fibrosis progression. Furthermore, after 12 months of treatment, when significant improvements in FVC and DLCO were observed, we established specific plasma concentration cut-off values for these biomarkers, providing preliminary thresholds that may help stratify patients based on fibrosis progression. Their dynamic changes over time, together with improvements in FVC and DLCO, suggest that nintedanib may influence ECM remodeling and fibrosis-related pathways in a complex manner. Further investigations are needed to clarify the precise mechanistic role of these biomarkers in response to treatment and to establish their potential utility as prognostic indicators for ILD progression. A hypothetical model describing the role of the biomarkers identified in this study in pulmonary fibrosis is depicted in Fig. 7.

[See PDF for image]

Fig. 7

Molecular model of pulmonary fibrosis illustrating the potential involvement of the biomarkers identified in our study (black boxes). The image was created using Microsoft PowerPoint 2013 (Microsoft Corporation, https://www.microsoft.com/powerpoint).

Our exploratory analysis identified a subgroup of biomarkers with the potential to distinguish between fibrotic patients and healthy controls. We also demonstrated how their plasma concentrations changed during antifibrotic treatment, highlighting their potential prognostic value. These biomarkers may offer advantages over current clinical and imaging methods for evaluating the therapeutic effects of nintedanib. Traditional approaches, such as pulmonary function tests (FVC and DLCO) and high-resolution computed tomography, provide critical insights into disease status but can be influenced by patient effort and may lack sensitivity in detecting subtle or early changes in fibrotic activity. In contrast, plasma biomarkers may reflect ongoing molecular and pathophysiological processes in real time, thereby capturing early treatment responses that may precede measurable changes in lung function or radiological appearance. Their noninvasive nature also facilitates more frequent assessments, enabling closer monitoring of disease trajectory and treatment efficacy.

The limitations of this study must be acknowledged. First, the relatively small sample size and occurrence of adverse drug events may have introduced a degree of selection bias, thereby limiting the generalizability of our findings. While our data provides preliminary evidence supporting the modulation of these biomarkers during antifibrotic treatment, larger prospective studies are needed to confirm these observations in more heterogeneous and representative patient populations. Additionally, the molecular mechanisms underlying the dynamic changes in these biomarkers remain to be elucidated.

In conclusion, our exploratory analysis identified five biomarkers whose plasma concentrations changed during antifibrotic treatment, highlighting their potential prognostic value. The biomarkers identified in this study may complement clinical functional parameters to help monitor biological response to therapy, differentiate responders from non-responders, and support treatment decisions, potentially improving long-term outcomes for patients with PF-ILDs. Future investigations should aim to clarify the regulatory pathways involved and assess the clinical utility of these biomarkers as pharmacodynamic or predictive biomarkers to support ongoing investigations into PF-ILDs.

Acknowledgements

The authors would like to thank UPMCI Language Services for editing the manuscript and Ray Hill for English-language and technical editing and formatting before submission on behalf of Springer Healthcare. Boehringer Ingelheim, Italy, funded this editorial assistance.

Author contributions

V.M., A.C., E.C., and P.V. conceived the project. V.M. prepared the original draft. V.M., A.C., E.C., and P.V. designed all the experiments. N.L., L.M., and C.C. performed sample collection and protein analysis. S.C. performed the statistical analysis. P.G.C. and M.P. revised the manuscript critically for important intellectual content. Funding for the project was obtained by P.V. All the authors have read and agreed to the submitted version of the manuscript.

Funding

This research was funded by the Italian Ministry of Health, Ricerca Corrente. This medical writing assistance was funded by Boehringer Ingelheim. Boehringer Ingelheim had no role in the design, analysis or interpretation of the results in this study. Boehringer Ingelheim was given the opportunity to review the manuscript for medical and scientific accuracy as well as intellectual property considerations.

Data availability

The authors declare that all the data supporting this study’s findings are available within the article or from the corresponding author upon reasonable request. The GSE21411 and GSE24206 datasets used in the current study are available in the Gene Expression Omnibus repository (http://www.ncbi.nlm.nih.gov/geo/).

Declarations

Competing interests

The authors declare no competing interests.

Ethical approval

The study was conducted in accordance with the Declaration of Helsinki. Samples were obtained with fully informed written consent, and ethics approval was obtained from the IRCCS ISMETT’s Institutional Research Review Board supervisors (project number: IRRB/08/22) and the relevant local ethics committee.

Abbreviations

AUC

Area under the curve

DLCO

Diffusing capacity of the lung for carbon monoxide

ECM

Extracellular matrix

FVC

Forced vital capacity

ILD

Interstitial lung disease

IPF

Idiopathic pulmonary fibrosis

PCA

Principal component analysis

PF-ILD

Progressive fibrosing interstitial lung disease

ROC

Receiver operating characteristic

Publisher’s note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

References

1. Travis, WD et al. An official American thoracic society/european respiratory society statement: update of the international multidisciplinary classification of the idiopathic interstitial pneumonias. Am. J. Respir Crit. Care Med.; 2013; 188, pp. 733-748. [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/24032382][PubMedCentral: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC5803655]

2. Kaunisto, J. et al. Demographics and survival of patients with idiopathic pulmonary fibrosis in the FinnishIPF registry. ERJ Open Res.5 (2019).

3. Cottin, V; Teague, R; Nicholson, L; Langham, S; Baldwin, M. The burden of progressive-fibrosing interstitial lung diseases. Front. Med. (Lausanne); 2022; 9, 799912. [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/35178411]

4. Cottin, V. et al. Fibrosing interstitial lung diseases: Knowns and unknowns. Eur. Respir. Rev.28 (2019).

5. Brown, K. K. et al. The natural history of progressive fibrosing interstitial lung diseases. Eur. Respir. J.55 (2020).

6. Maher, TM; Strek, ME. Antifibrotic therapy for idiopathic pulmonary fibrosis: time to treat. Respir Res.; 2019; 20, 205. [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/31492155][PubMedCentral: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC6731623]

7. Wijsenbeek, M; Cottin, V. Spectrum of fibrotic lung diseases. N Engl. J. Med.; 2020; 383, pp. 958-968.1:CAS:528:DC%2BB3cXhslGksrvK [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/32877584]

8. Wynn, TA. Cellular and molecular mechanisms of fibrosis. J. Pathol.; 2008; 214, pp. 199-210.1:CAS:528:DC%2BD1cXhvFCkt7w%3D [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/18161745][PubMedCentral: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC2693329]

9. King, TE, Jr; Pardo, A; Selman, M. Idiopathic pulmonary fibrosis. Lancet; 2011; 378, pp. 1949-1961. [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/21719092]

10. Rock, JR et al. Multiple stromal populations contribute to pulmonary fibrosis without evidence for epithelial to mesenchymal transition. Proc. Natl. Acad. Sci. U S A; 2011; 108, pp. E1475-E1483.1:CAS:528:DC%2BC38XnsFemuw%3D%3D [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/22123957][PubMedCentral: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3248478]

11. Noble, PW; Barkauskas, CE; Jiang, D. Pulmonary fibrosis: patterns and perpetrators. J. Clin. Invest.; 2012; 122, pp. 2756-2762.1:CAS:528:DC%2BC38Xht1SisrvM [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/22850886][PubMedCentral: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3408732]

12. Selman, M; Pardo, A. The leading role of epithelial cells in the pathogenesis of idiopathic pulmonary fibrosis. Cell. Signal.; 2020; 66, 109482.1:CAS:528:DC%2BC1MXit1Ois7vL [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/31760172]

13. Richeldi, L et al. Efficacy of a tyrosine kinase inhibitor in idiopathic pulmonary fibrosis. N Engl. J. Med.; 2011; 365, pp. 1079-1087.1:CAS:528:DC%2BC3MXht1WqsbrN [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/21992121]

14. Richeldi, L et al. Efficacy and safety of nintedanib in idiopathic pulmonary fibrosis. N Engl. J. Med.; 2014; 370, pp. 2071-2082. [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/24836310]

15. Flaherty, KR et al. Nintedanib in progressive fibrosing interstitial lung diseases. N Engl. J. Med.; 2019; 381, pp. 1718-1727.1:CAS:528:DC%2BC1MXitFWktrzL [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/31566307]

16. Distler, O et al. Nintedanib for systemic sclerosis-associated interstitial lung disease. N Engl. J. Med.; 2019; 380, pp. 2518-2528.1:CAS:528:DC%2BB3cXnsFKgu7c%3D [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/31112379]

17. Wollin, L; Distler, JH; Denton, CP; Gahlemann, M. Rationale for the evaluation of nintedanib as a treatment for systemic sclerosis-associated interstitial lung disease. J. Scleroderma Relat. Disord; 2019; 4, pp. 212-218. [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/35382502][PubMedCentral: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC8922567]

18. Jenkins, RG et al. Effects of nintedanib on Circulating biomarkers of idiopathic pulmonary fibrosis. ERJ Open. Res.; 2024; 10, pp. 00558-2023. [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/39687396][PubMedCentral: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC11647937]

19. Humphries, S. M. et al. Quantitative high-resolution computed tomography fibrosis score: Performance characteristics in idiopathic pulmonary fibrosis. Eur. Respir J.52 (2018).

20. Wong, AW; Ryerson, CJ; Guler, SA. Progression of fibrosing interstitial lung disease. Respir Res.; 2020; 21, 32. [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/31996266][PubMedCentral: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC6988233]

21. Nathan, S. D., Yang, M., Morgenthien, E. A. & Stauffer, J. L. FVC variability in patients with idiopathic pulmonary fibrosis and role of 6-min walk test to predict further change. Eur. Respir J.55 (2020).

22. Adegunsoye, A; Vij, R; Noth, I. Integrating genomics into management of fibrotic interstitial lung disease. Chest; 2019; 155, pp. 1026-1040. [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/30660786][PubMedCentral: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC6533453]

23. Glenn, LM; Troy, LK; Corte, TJ. Novel diagnostic techniques in interstitial lung disease. Front. Med. (Lausanne); 2023; 10, 1174443. [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/37188089]

24. Mismetti, V et al. Proteomic biomarkers for survival in systemic sclerosis-associated pulmonary hypertension. Respir Res.; 2023; 24, 273.1:CAS:528:DC%2BB3sXitlajtLvK [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/37936223][PubMedCentral: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC10629050]

25. Wu, Y et al. Proteomics: potential techniques for discovering the pathogenesis of connective tissue diseases-interstitial lung disease. Front. Immunol.; 2023; 14, 1146904.1:CAS:528:DC%2BB3sXps1SntL8%3D [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/37063894][PubMedCentral: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC10090492]

26. Landi, C et al. Idiopathic pulmonary fibrosis serum proteomic analysis before and after nintedanib therapy. Sci. Rep.; 2020; 10, 9378.2020NatSR.10.9378L1:CAS:528:DC%2BB3cXhtFGgt7vL [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/32523095][PubMedCentral: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC7287088]

27. Bowman, WS et al. Proteomic biomarkers of progressive fibrosing interstitial lung disease: a multicentre cohort analysis. Lancet Respir Med.; 2022; 10, pp. 593-602.1:CAS:528:DC%2BB38XhtFKksbfP [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/35063079][PubMedCentral: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC9177713]

28. Raghu, G et al. Idiopathic pulmonary fibrosis (an update) and progressive pulmonary fibrosis in adults: an official ATS/ERS/JRS/ALAT clinical practice guideline. Am. J. Respir Crit. Care Med.; 2022; 205, pp. e18-e47. [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/35486072][PubMedCentral: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC9851481]

29. Metsalu, T; Vilo, J. ClustVis: a web tool for visualizing clustering of multivariate data using principal component analysis and heatmap. Nucleic Acids Res.; 2015; 43, pp. W566-W570.1:CAS:528:DC%2BC2sXhtVymtLjK [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/25969447][PubMedCentral: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4489295]

30. Emblom-Callahan, MC et al. Genomic phenotype of non-cultured pulmonary fibroblasts in idiopathic pulmonary fibrosis. Genomics; 2010; 96, pp. 134-145.1:CAS:528:DC%2BC3cXhtVajtLrL [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/20451601]

31. Ley, B; Brown, KK; Collard, HR. Molecular biomarkers in idiopathic pulmonary fibrosis. Am. J. Physiol. Lung Cell. Mol. Physiol.; 2014; 307, pp. L681-L691.1:CAS:528:DC%2BC2cXitVeitLrE [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/25260757][PubMedCentral: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4280147]

32. Hamilton, BC; Kukreja, J; Ware, LB; Matthay, MA. Protein biomarkers associated with primary graft dysfunction following lung transplantation. Am. J. Physiol. Lung Cell. Mol. Physiol.; 2017; 312, pp. L531-L41.1:STN:280:DC%2BC1c3gtVWrsw%3D%3D [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/28130262][PubMedCentral: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC5407092]

33. Mayr, CH et al. Integrative analysis of cell state changes in lung fibrosis with peripheral protein biomarkers. EMBO Mol. Med.; 2021; 13, e12871.1:CAS:528:DC%2BB3MXltlenurk%3D [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/33650774][PubMedCentral: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC8033531]

34. Bowman, WS; Echt, GA; Oldham, JM. Biomarkers in progressive fibrosing interstitial lung disease: optimizing diagnosis, prognosis, and treatment response. Front. Med. (Lausanne); 2021; 8, 680997. [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/34041256]

35. Wang, L et al. Serum proteomics identifies biomarkers associated with the pathogenesis of idiopathic pulmonary fibrosis. Mol. Cell. Proteom.; 2023; 22, 100524.1:CAS:528:DC%2BB3sXntVektbo%3D

36. Oldham, JM et al. Proteomic biomarkers of survival in idiopathic pulmonary fibrosis. Am. J. Respir Crit. Care Med.; 2024; 209, pp. 1111-1120.1:CAS:528:DC%2BB2cXhtV2rs7bF [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/37847691]

37. Hambly, N; Shimbori, C; Kolb, M. Molecular classification of idiopathic pulmonary fibrosis: personalized medicine, genetics and biomarkers. Respirology; 2015; 20, pp. 1010-1022. [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/26109466]

38. Inoue, Y et al. Diagnostic and prognostic biomarkers for chronic fibrosing interstitial lung diseases with a progressive phenotype. Chest; 2020; 158, pp. 646-659.1:CAS:528:DC%2BB3cXhtlWnsLfK [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/32268131]

39. Glasser, SW et al. Mechanisms of lung fibrosis resolution. Am. J. Pathol.; 2016; 186, pp. 1066-1077.1:CAS:528:DC%2BC28XltF2rtb4%3D [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/27021937][PubMedCentral: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4861766]

40. Wuyts, WA et al. The pathogenesis of pulmonary fibrosis: A moving target. Eur. Respir J.; 2013; 41, pp. 1207-1218.1:CAS:528:DC%2BC3sXovFajurY%3D [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/23100500]

41. George, PM et al. Progressive fibrosing interstitial lung disease: clinical uncertainties, consensus recommendations, and research priorities. Lancet Respir Med.; 2020; 8, pp. 925-934. [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/32890499]

42. Kalchiem-Dekel, O., Galvin, J. R., Burke, A. P., Atamas, S. P. & Todd, N. W. Interstitial lung disease and pulmonary fibrosis: A practical approach for general medicine physicians with focus on the medical history. J. Clin. Med.7 (2018).

43. Inomata, M; Nishioka, Y; Azuma, A. Nintedanib: Evidence for its therapeutic potential in idiopathic pulmonary fibrosis. Core Evid.; 2015; 10, pp. 89-98.1:CAS:528:DC%2BC1cXlvVejurs%3D [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/26346347][PubMedCentral: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4555978]

44. Wollin, L. et al. Potential of nintedanib in treatment of progressive fibrosing interstitial lung diseases. Eur. Respir. J.54 (2019).

45. d’Alessandro, M et al. Galactin-1, 3 and 9: potential biomarkers in idiopathic pulmonary fibrosis and other interstitial lung diseases. Respir Physiol. Neurobiol.; 2020; 282, 103546. [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/32920057]

46. Kathiriya, JJ et al. Galectin-1 Inhibition attenuates profibrotic signaling in hypoxia-induced pulmonary fibrosis. Cell. Death Discov; 2017; 3, 17010.1:CAS:528:DC%2BC2sXlvVKnt7c%3D [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/28417017][PubMedCentral: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC5385413]

47. Hsu, YA et al. Amelioration of bleomycin-induced pulmonary fibrosis via TGF-beta-induced Smad and non-Smad signaling pathways in galectin-9-deficient mice and fibroblast cells. J. Biomed. Sci.; 2020; 27, 24.1:CAS:528:DC%2BB3cXls12gur0%3D [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/31937306][PubMedCentral: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC6961390]

48. Wollin, L et al. Mode of action of nintedanib in the treatment of idiopathic pulmonary fibrosis. Eur. Respir J.; 2015; 45, pp. 1434-1445.1:CAS:528:DC%2BC2MXhtVGju7jP [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/25745043][PubMedCentral: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4416110]

49. Rangarajan, S et al. Novel mechanisms for the antifibrotic action of nintedanib. Am. J. Respir Cell. Mol. Biol.; 2016; 54, pp. 51-59.1:CAS:528:DC%2BC28XisF2ksrg%3D [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/26072676][PubMedCentral: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4742925]

50. Guiot, J; Bondue, B; Henket, M; Corhay, JL; Louis, R. Raised serum levels of IGFBP-1 and IGFBP-2 in idiopathic pulmonary fibrosis. BMC Pulm Med.; 2016; 16, 86.1:STN:280:DC%2BC2s%2FivVOhsQ%3D%3D [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/27215343][PubMedCentral: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4877734]

51. Habeb, M; Embarak, S; Fathy, A. Non-invasive IGFBP1, IGFBP2 biomarkers as predictors and outcomes of usual interstitial pneumonia (UIP) therapeutic response. Egypt. J. Bronchol.; 2021; 15, 6.

52. Chin, C et al. Loss of IGFBP2 mediates alveolar type 2 cell senescence and promotes lung fibrosis. Cell. Rep. Med.; 2023; 4, 100945.1:CAS:528:DC%2BB3sXjtFaiu7k%3D [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/36787736][PubMedCentral: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC10040381]

53. Chi, JY et al. Blockade of the pentraxin 3/CD44 interaction attenuates lung injury-induced fibrosis. Clin. Transl Med.; 2022; 12, e1099.1:CAS:528:DC%2BB38XjtVymtLzM [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/36336784][PubMedCentral: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC9637652]

54. Inoue, R. et al. Amelioration of pulmonary fibrosis by matrix metalloproteinase-2 overexpression. Int. J. Mol. Sci.24 (2023).

55. Todd, JL et al. Circulating matrix metalloproteinases and tissue metalloproteinase inhibitors in patients with idiopathic pulmonary fibrosis in the multicenter IPF-PRO registry cohort. BMC Pulm Med.; 2020; 20, 64.1:CAS:528:DC%2BB3cXlslyru7c%3D [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/32171287][PubMedCentral: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC7071646]

56. Bormann, T et al. Role of matrix metalloprotease-2 and MMP-9 in experimental lung fibrosis in mice. Respir Res.; 2022; 23, 180.2022cosm.book...B1:CAS:528:DC%2BB38XhvVOisL3O [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/35804363][PubMedCentral: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC9270768]

57. Radbill, BD et al. Loss of matrix metalloproteinase-2 amplifies murine toxin-induced liver fibrosis by upregulating collagen I expression. Dig. Dis. Sci.; 2011; 56, pp. 406-416.1:CAS:528:DC%2BC3MXht1SrtLg%3D [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/20563750]

58. Hostettler, KE et al. Anti-fibrotic effects of nintedanib in lung fibroblasts derived from patients with idiopathic pulmonary fibrosis. Respir Res.; 2014; 15, 157. [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/25496490][PubMedCentral: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4273482]

59. Inoue, R et al. Amelioration of pulmonary fibrosis by matrix Metalloproteinase-2 overexpression. Int. J. Mol. Sci.; 2023; 24, 6695.1:CAS:528:DC%2BB3sXotFSgsL8%3D [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/37047672][PubMedCentral: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC10095307]