Biomarker profiles in idiopathic and progressive pulmonary fibrosis after LPA₁ antagonism: exploratory analysis from a phase 2 trial of admilparant

Abstract

Background

In a phase 2 trial (NCT04308681), treatment with admilparant (BMS-986278), an oral lysophosphatidic acid receptor 1 (LPA₁) antagonist, reduced lung function decline in patients with idiopathic pulmonary fibrosis (IPF) or progressive pulmonary fibrosis (PPF). In this exploratory analysis, we evaluated post-treatment changes in circulating biomarkers of lung fibrosis to elucidate mechanisms of admilparant action.

Methods

Patients with IPF or PPF were randomized 1:1:1 to receive twice-daily admilparant (30 or 60 mg) or placebo for 26 weeks; background antifibrotics were allowed. The IPF and PPF cohorts were analyzed separately. Changes from baseline (CfB) in serum proteins associated with epithelial injury, inflammation, and fibrosis were measured by quantitative immunoassays at 4, 12, and 26 weeks, and compared between patients who received admilparant versus placebo. Pharmacodynamic biomarker changes were evaluated by clinical response status at week 26. Plasma samples from patients with IPF were assessed post hoc by SomaScan v4.1 proteomics assay. Statistical evaluations used linear mixed-effects models.

Results

In the IPF cohort (n=276), nine serum proteins showed significant CfB (p<0.05) at week 26 in patients treated with 60-mg admilparant versus placebo, including increased adiponectin, MMIF, CD163, CEA, and ENRAGE, and decreased markers of epithelial injury and fibrosis (CA-125/MUC16, MMP-7, TN-C, PRO-FIB). In the PPF cohort (n=116), significant differences at week 26 (p<0.05) were observed for 11 serum proteins, including increased CEA and decreased periostin, IL6Rβ, CD163, KIM-1, multiple inflammatory markers (YKL-40, VCAM-1, PECAM-1, ferritin), and collagen degradation markers (C3M and C4M). Two serum proteins, CA-125 and TN-C, showed significantly greater CfB (p<0.05) in responders compared with non-responders in the IPF cohort. Plasma proteomic analysis identified differential expression (adjusted p<0.1) of adiponectin, CKMT1A, ANGPTL3, PDCD1LG2, and IGFBP6 at week 26 in patients with IPF treated with 60-mg admilparant versus placebo.

Conclusions

Treatment with 60-mg admilparant improved circulating biomarkers associated with epithelial injury and fibrosis in IPF, and with inflammation, fibrosis, and collagen degradation in PPF. Proteomic analysis identified additional biomarkers associated with mitochondrial and metabolic pathways in IPF. These findings expand our understanding of potential mechanisms of LPA₁ antagonism with admilparant and identify biomarkers that may help evaluate treatment response and disease activity.

Supplementary Information

The online version contains supplementary material available at https://doi.org/10.1186/s12931-026-03809-w.

Background

Interstitial lung diseases (ILDs) are a diverse group of chronic restrictive disorders marked by inflammation and/or fibrosis of the lung that share some clinical and pathophysiological features. Idiopathic pulmonary fibrosis (IPF) has an unknown etiology and is characterized by progressive scarring of the lung tissue leading to irreversible loss of lung function and poor prognosis [1–3]. Progressive pulmonary fibrosis (PPF) refers to the progressive worsening of a preexisting ILD and includes idiopathic interstitial pneumonia, autoimmune ILDs (e.g. rheumatoid arthritis, systemic sclerosis), exposure-related ILDs (e.g. hypersensitivity pneumonitis), ILDs with cysts and/or airspace filling (e.g. Langerhans cell histiocytosis), and sarcoidosis [3, 4].

Treatment of IPF and PPF is focused on targeting various pathways of disease pathogenesis. Fibrotic lung diseases are caused by an aberrant wound-healing response to repetitive epithelial injury (such as viral infection, gastroesophageal reflux, cigarette smoke, and other environmental exposures) leading to the activation and migration of fibroblasts to the lungs and excessive extracellular matrix (ECM) deposition and stiffening [4]. PPF often begins with an inflammatory phase triggered by either an endogenous autoantigen or an exogenous antigen, such as an environmental trigger [5]. Upon repetitive exposure or in a chronic setting, persistent inflammation can activate fibrogenic pathways. Ultimately, this pathogenic process leads to irreversible destruction of the lung architecture and loss of lung function. IPF and PPF share some common downstream mechanistic pathways (e.g. activation of latent transforming growth factor β [TGF-β], platelet-derived growth factor, connective tissue growth factor, macrophage-stimulating growth factor) that result in progressive fibrosis, irrespective of the initial lung injury or trigger [5–7].

The lysophosphatidic acid 1 receptor (LPA₁) pathway is a potential target for the treatment of fibrotic lung disease. While the bioactive lipid LPA promotes normal wound healing and collagen deposition, LPA₁ activation has been associated with processes that drive the progression of pulmonary fibrosis, including proinflammatory responses, epithelial cell apoptosis, fibroblast recruitment to the site of injury, increased activation/proliferation and survival of fibroblasts, and collagen secretion [8–10]. Proof-of-concept studies demonstrated that BMS-986020, a first-generation LPA₁ antagonist, inhibited LPA₁-induced fibrogenesis in scar-in-a-jar in vitro experiments [8]. This was further supported by phase 2 clinical trial data showing that administration of BMS-986020 reduced serum ECM neopeptide biomarkers in patients with IPF [8]. BMS-986020 was efficacious in slowing the rate of decline in forced vital capacity (FVC); however, the trial was discontinued due to compound-specific off-target effects that were unrelated to LPA₁ antagonism [11–13]. Building on these insights, admilparant (BMS-986278), a second-generation oral LPA₁ antagonist, was developed and evaluated in a phase 2 trial with parallel IPF and PPF cohorts. The results showed that 60-mg admilparant twice daily reduced the rate of decline in percent of predicted FVC (ppFVC) over 26 weeks compared with placebo, while demonstrating a generally safe and well-tolerated profile in both cohorts [14].

To further understand the mechanisms of admilparant action and identify biomarkers of treatment response and disease activity, we evaluated changes in circulating biomarkers of lung fibrosis using targeted quantitative immunoassays in the IPF and PPF cohorts in the phase 2 trial after treatment with 30-mg admilparant, 60-mg admilparant, or placebo. Pharmacodynamic biomarker changes were also evaluated by clinical response status at week 26. Additionally, a post hoc plasma proteomic analysis evaluated more than 7000 protein analytes from samples obtained from patients in the IPF cohort.

Methods

Primary study design

A double-blind, placebo-controlled phase 2 trial (ClinicalTrials.gov identifier NCT04308681) was conducted at 79 sites in 17 countries (July 29, 2020, through September 22, 2023). Parallel cohorts of eligible patients with IPF or PPF were randomized (1:1:1) to receive 30-mg admilparant, 60-mg admilparant, or placebo orally twice daily for 26 weeks. The primary endpoint was the rate of change in ppFVC (%) from baseline to week 26 in the IPF cohort. Key efficacy and safety results, as well as the study design, have been previously reported in detail [14, 15]. Briefly, eligible patients with IPF were aged ≥40 years with a diagnosis of IPF within 7 years; eligible patients with PPF were aged ≥21 years with high-resolution computed tomography obtained at screening that demonstrated evidence of >10% parenchymal fibrosis within the whole lung and evidence of progression within 2 years. Background antifibrotic treatment (nintedanib or pirfenidone) was allowed in both cohorts if dosing was stable for ≥3 months before randomization. Background ILD-targeted immunosuppressant therapy (mycophenolate mofetil, mycophenolic acid, azathioprine, or tacrolimus) was allowed in the PPF cohort if dosing was stable for ≥6 months before randomization. The objective of this exploratory analysis was to describe the pharmacodynamic effects of LPA₁ antagonism with admilparant and to gain mechanistic insight.

Targeted quantitative assay biomarker analyses

Predose serum samples for circulating biomarkers were collected at baseline, and weeks 4, 12, and 26. Serum samples were stored at −80°C and shipped to the following laboratories for testing. All longitudinal samples from the same patient were assayed together on the same plate in the same batch, ensuring that within‑patient changes were not influenced by plate‑ or batch‑level effects. Assays for a panel of epithelial damage, lung fibrosis, inflammation, and thrombosis markers in the PROLIFIC (Prognostic Lung Fibrosis Consortium) multiplex biomarker assay and additional IPF disease biomarkers and potential pharmacodynamic biomarkers of LPA₁ inhibition were conducted at Rules-Based Medicine (formerly Myriad RBM), an IQVIA business (Austin, TX, USA). A panel of ECM biomarkers associated with collagen formation/degradation, inflammation, angiogenesis, and wound healing were measured at Nordic Bioscience A/S (Herlev, Denmark). Supplementary Table 1 lists the 60 unique biomarkers examined in this analysis.

The analyses were exploratory and performed separately in the IPF and PPF cohorts. Data were analyzed for all randomized patients who received at least one dose of study treatment (placebo, 30-mg admilparant, or 60-mg admilparant) and had a valid serum sample at any time point through week 26. Biomarkers for which more than 33% of samples were below the lower limit of quantification (LLOQ) were excluded from the analysis. Measurements of the remaining biomarkers that were below the LLOQ were imputed as half the LLOQ value, whereas those that exceeded the upper limit of quantification were assigned the upper limit of quantification value. Biomarkers were log₂-transformed before modeling.

The longitudinal analyses were performed with linear mixed-effects models. The model included fixed effects of timepoints, treatment, and the interaction between timepoint and treatment. Random effect was accounted for by random intercepts for the patients. The model covariates were age, sex, MUC5B genotype, and the baseline stratification factors for each cohort. Randomization stratification factors for IPF were country (Japan vs. non-Japan) and background use of antifibrotic treatment (pirfenidone vs. nintedanib vs. none). Randomization stratification factors for PPF were presence of usual interstitial pneumonia pattern (present vs. absent) and background therapy category (antifibrotics ± ILD-targeted immunosuppressants vs. ILD-targeted immunosuppressants vs. none). Differences in percent change from baseline for the 30-mg and 60-mg admilparant groups versus placebo group were assessed.

To identify biomarkers that may predict treatment response, we tested whether baseline biomarker levels were associated with lung function change (relative ppFVC at week 26) differently across treatment arms. A linear regression model was applied with the covariates mentioned above. P values were not adjusted for multiple testing due to the exploratory nature of the analysis, and significance was based on p<0.05. Corresponding false discovery rate–adjusted q values (Benjamini–Hochberg method) are provided in the Supplementary Tables.

In the IPF cohort, pharmacodynamic biomarker changes were also evaluated by clinical response status at week 26, defined based on ppFVC: responders (no decline), partial responders (0–5% decline), and non-responders (>5% decline). Longitudinal changes from baseline were analyzed using linear mixed-effects models including fixed effects for baseline, treatment, visit, response group, and their interactions, with patient-level random intercepts. Selected biomarkers with nominal significance (responders vs. non-responders, p<0.05) at week 26 were highlighted and visualized. Pharmacodynamic analyses by response were not performed in the PPF cohort due to limited sample size within subgroups.

SomaScan proteomic analysis

The proteomic analysis was conducted post hoc in the IPF cohort. Plasma collected from patients in the trial was frozen and sent to SomaLogic (Boulder, CO, USA) for analysis of 7288 protein analytes via the SomaScan v4.1 slow off-rate modified aptamer assay. Data from a subset of randomized patients in the IPF cohort who received at least one dose of 60-mg admilparant or placebo and had a valid plasma sample at baseline, week 4, and week 26 were analyzed.

Significant changes (adjusted p<0.1) in differentially expressed proteins from baseline and between treatment groups were identified by linear mixed-effects models using the limma approach. SomaScan measurements underwent adaptive normalization by maximum likelihood and were then log₂-transformed and adjusted by age, sex, MUC5B genotype, and region. Treatment effects in the 60-mg admilparant group at weeks 4 and 26 were compared with baseline within each arm, respectively, and versus placebo at weeks 4 and 26. P values were adjusted for multiple testing with Benjamini–Hochberg method for each comparison across all proteins.

Results

Targeted quantitative assay biomarker analyses in the IPF and PPF cohorts

The biomarker analysis included 276 patients with IPF and 116 patients with PPF. Baseline characteristics were generally balanced across treatment arms in each cohort (Supplementary Table 2). Most patients were male (IPF 80%; PPF 53%) and White (IPF 70%; PPF 78%), and had a mean age of 69 years (IPF) and 70 years (PPF), respectively. The mean ppFVC at baseline ranged from 76.0% to 76.9% across treatment groups in the IPF cohort, and 63.9% to 68.5% in the PPF cohort.

In the IPF cohort, after adjustment for covariates in the modeled analysis, nine serum proteins showed significant change (p<0.05) from baseline at week 26 in patients treated with 60-mg admilparant compared with placebo: adiponectin, carcinoembryonic antigen (CEA), and cancer antigen 125 (CA-125, also known as mucin 16), matrix metalloproteinase (MMP)−7, tenascin C (TN-C), thrombin-mediated fibrinogen degradation (PRO-FIB), macrophage migration inhibitory factor (MMIF), cluster of differentiation 163 (CD163), and extracellular newly identified receptor for advanced glycation end-products binding protein (EN-RAGE) (Fig. 1, Supplementary Table 3). Increases in adiponectin, a potential biomarker of anti-inflammatory and antifibrotic effects [16], were dose- and time-dependent, with estimated mean percent change from baseline of 23.7% and 34.7% at weeks 4 and 26, respectively, in the 60-mg group (Fig. 1; Supplementary Table 3). Significant reductions (p<0.05) in CA-125 (–6.2% vs. 8.5%), MMP-7 (0.6% vs. 12.5%), TN-C (1.7% vs. 13.9%), and PRO-FIB (–5.5% vs. 13.7%), which are biomarkers of epithelial injury [17–19], were observed at week 26 in the 60-mg admilparant group versus placebo, respectively. The full biomarker expression profile after treatment with 30-mg admilparant, 60-mg admilparant, or placebo at weeks 4, 12, and 26 in the IPF cohort is shown in Supplementary Fig. 1.

Fig. 1.

Mean percentage change from baseline (95% CI) in circulating biomarkers significantly affected by admilparant treatment vs. placebo in the IPF cohort at 26 weeks. ****p < 0.0001, ***p < 0.001, **p < 0.01, *p < 0.05 vs. placebo; modeled analysis. P values were not adjusted for multiple testing due to the exploratory nature of the analysis, and significance was based on p < 0.05. Multiplicity adjusted FDR q values are reported in Supplementary Table 3

In the PPF cohort, after adjustment for covariates in the modeled analysis, 11 serum proteins showed significant change (p<0.05) from baseline at week 26 in patients treated with 60-mg admilparant versus placebo (Fig. 2A and B, Supplementary Table 4). Admilparant treatment over 26 weeks significantly increased serum levels of CEA (estimated mean change from baseline, 42.4% vs. 5.7%) and decreased serum levels of periostin, a biomarker of fibrosis via the TGF-β pathway [17, 20, 21] (–0.1% vs. 12.0%), interleukin-6 receptor subunit beta (IL-6Rβ, –2.8% vs. 8.9%), and kidney injury molecule-1 (KIM-1; –15.8% vs. 14.7%) compared with placebo (Fig. 2A). In addition, multiple inflammatory biomarkers, including YKL-40 (chitinase-3-like protein 1, –1.5% vs. 23.3%), vascular cell adhesion molecule 1 (VCAM-1; –2.1% vs. 14%), platelet endothelial cell adhesion molecule 1 (PECAM-1; –2.4% vs. 9.0%), CD163 (–0.3% vs. 9.1%), and ferritin (–23.0% vs. 7.3%) were significantly decreased relative to placebo at week 26 (Fig. 2A) [17, 22, 23], and significant reductions in the collagen degradation markers C1M, C3M, and C4M (C×M, neoepitope of MMP-2,9,13–mediated degradation of type × collagen) were observed at various timepoints after treatment with 60-mg admilparant (Fig. 2B) [24, 25]. The full biomarker expression profile after treatment with 30-mg admilparant, 60-mg admilparant, or placebo at weeks 4, 12, and 26 in the PPF cohort is shown in Supplementary Fig. 2.

Fig. 2.

A Mean percentage change from baseline (95% CI) in circulating biomarkers significantly affected by admilparant treatment vs. placebo in the PPF cohort at 26 weeks. B Mean percentage change from baseline (95% CI) in collagen degradation markers in the PPF cohort at various timepoints.****p < 0.0001, ***p < 0.001, **p < 0.01, *p < 0.05 vs. placebo; modeled analysis. P values were not adjusted for multiple testing due to the exploratory nature of the analysis, and significance was based on p < 0.05. Multiplicity adjusted FDR q values are reported in Supplementary Table 4

Baseline concentrations of the 18 circulating biomarkers that showed a significant difference by treatment group at week 26 in either the IPF or PPF cohorts are shown in Supplementary Table 5.

In addition, to explore biomarker predictors of treatment response, we evaluated whether baseline biomarker levels were differentially associated with change in ppFVC at week 26 across treatment arms. Ten biomarkers showed a nominally significant differing association (p<0.05) between the 60-mg admilparant group versus the placebo group: seven in the IPF cohort (epidermal growth factor [EGF], factor VII, macrophage inflammatory protein-3-beta [MIP-3β], MMP-7, PECAM-1, N-terminal type V collagen propeptide [PRO-C5], periostin), and three in the PPF cohort (CD163, eotaxin-1, MMP-1) (Supplementary Table 6).

In the responder versus non-responder analysis in the IPF cohort, two biomarkers, CA-125 and TN-C, showed significantly greater change from baseline (nominal p<0.05) at week 26 in responders compared with non-responders (Fig. 3), indicating that these pharmacodynamic changes track with clinical benefit.

Fig. 3.

Model-adjusted mean percent change from baseline (95% CI) in CA-125 and TN-C in responders, partial responders, and nonresponders in the IPF cohort. *P < 0.05, **p < 0.01, *** p < 0.001 versus nonresponders. Error bars denote the 95% confidence interval

SomaScan plasma proteomic analysis in the IPF cohort

In total, 446 samples were analyzed for 150 of 185 (81.1%) patients with IPF randomized to 60-mg admilparant or placebo; patients with missing baseline or week 26 samples were excluded. Baseline demographic and disease characteristics appeared broadly comparable between patients in the SomaScan analysis subset versus the overall trial cohort (Supplementary Table 7). The analysis identified 64 differentially expressed proteins in the admilparant group and five in the placebo group at week 26 versus baseline (Fig. 4A). Of the 64 differentially expressed proteins in the admilparant group, 53 were upregulated (12 of which were also upregulated at week 4) and 11 were downregulated (four of which were also downregulated at week 4) versus baseline (Fig. 4A). Upregulated biomarkers were involved in mitochondrial and metabolic pathways, while downregulated biomarkers were enriched in fibroblasts/myofibroblasts and collagen triple helix repeat containing 1 (CTHRC1+) fibroblasts [26].

Fig. 4.

SomaScan analysis of differentially expressed proteins in the IPF cohort. A Differentially expressed proteins relative to baseline in the admilparant and placebo groups at weeks 4 and 26 (linear-mixed effects model). B Differentially expressed proteins in the admilparant group relative to placebo at weeks 4 and 26 (linear mixed-effects model). C Model-estimated fold change from baseline (95% CI) in differentially expressed proteins at weeks 4 and 26 (pre-normalized by adaptive normalization by maximum likelihood).Adjusted p values: ****p < 0.0001, ***p < 0.001, **p < 0.01, *p < 0.1. Blue asterisks indicate adjusted p values for post-treatment vs. baseline within each treatment arm. Red asterisks indicate adjusted p values for the treated arm vs. placebo. P values were adjusted for multiple testing with Benjamini–Hochberg method for each comparison across all proteins.

In the linear mixed-effects models, the levels of five serum proteins were significantly differentially expressed (adjusted p<0.1) at week 26 in the 60-mg admilparant group versus the placebo group (Fig. 4B). Adiponectin, mitochondrial creatine kinase (CKMT1A), angiopoietin-like 3 (ANGPTL3), and programmed death ligand 2 (PDCD1LG2) were significantly upregulated and insulin-like growth factor-binding protein 6 (IGFBP6), a biomarker of injury/fibrotic response, was significantly downregulated at week 26 (adjusted p<0.1) (Fig. 4B). Of these, adiponectin, CKMT1A, and IGFBP6 were also significantly differentially expressed at week 4 in the admilparant group versus placebo (adjusted p<0.1).

The model-estimated fold changes in SomaScan measurements (pre-normalized by adaptive normalization by maximum likelihood) are shown in Fig. 4C. The model showed significant changes in adiponectin, CKMT1A, IGFBP6, PDCD1LG2, and ANGPTL3 at week 26 in the admilparant group versus placebo. Significant changes at week 26 compared with baseline were also observed in coiled-coil domain-containing protein 80 (CCDC80) and single-stranded DNA binding protein 1 (SSBP1).

Discussion

Elucidating the biomarkers associated with response to treatment and the underlying complex pathways of ILD, such as fibrogenesis, ECM remodeling, epithelial cell injury, and inflammation, is an ongoing field of research. Biomarker signatures for progression in IPF have been studied extensively in various cohorts including the Pulmonary Fibrosis Biomarker (PFBIO) cohort, Australian Idiopathic Pulmonary Fibrosis Registry, Trent Lung Fibrosis, and the Prospective Observation of Fibrosis in the Lung Clinical Endpoints (PROFILE) study datasets [24, 27, 28]. Here, an exploratory analysis of a phase 2 trial showed divergent biomarker expression in patients with IPF or PPF treated with the LPA₁ antagonist, admilparant, compared with those receiving placebo. Treatment with 60-mg admilparant versus placebo significantly reduced circulating biomarkers of epithelial injury and fibrosis in patients with IPF and biomarkers of inflammation, fibrosis, and collagen degradation in patients with PPF. Additionally, a proteomics analysis identified an increase in circulating biomarkers associated with mitochondrial and metabolic pathways in patients with IPF after treatment with admilparant. The divergent biomarker profiles observed in the IPF and PPF cohorts may reflect the multifaceted role of LPA₁ signaling in disease pathogenesis and the heterogeneity among patients with fibrotic lung diseases.

Although the main objective of this analysis was to explore the pharmacodynamic effects of LPA₁ antagonism with admilparant and to gain mechanistic insight, the analysis in the IPF and PPF cohorts also identified several biomarkers as potential baseline predictors associated with clinical response. These biomarkers showed prognostic value for beneficial ppFVC outcomes at week 26 with potential utility for advancing mechanistic understanding of treatment effects. The predictive biomarkers with significance span the categories of inflammation, fibrosis, epithelial cell damage, and vascular leak. Several of these markers, including MMP-7, have established associations with IPF disease activity, lending biological plausibility to these findings; however, given the exploratory nature and sample size of this analysis, these predictive relationships require validation in larger, prospective cohorts or studies. Furthermore, in an analysis of responders versus nonresponders, changes in two biomarkers, CA-125 and TN-C, were found to be predictive of a positive treatment response in patients with IPF.

The substantial increase in adiponectin levels observed after treatment with 60-mg admilparant in patients with IPF was consistent with results from a previous trial in patients with IPF of a first-generation LPA₁ antagonist, BMS-986020 [12]. Adiponectin was also confirmed to be upregulated in the SomaScan proteomic analysis. Adiponectin is a cytokine, primarily secreted from adipose tissue, involved in inflammation atherogenesis [29]. Typically, elevated adiponectin levels are associated with improved insulin sensitivity [30], cardio-protection, and improved kidney and liver function [29]. In a preclinical model of systemic sclerosis, an adiponectin receptor agonist was able to reduce matrix deposition and inflammatory gene expression in the skin [31]. An additional study demonstrated that adiponectin suppressed stiffness-dependent, profibrotic activation of lung by signaling through the cadherin 13 receptor [16]. These data suggest that adiponectin may reflect part of the biological mechanism underlying LPA₁ inhibition. Patients with fibrotic ILDs including IPF have shown reduced circulating levels of serum adiponectin [32].

Previous work using comprehensive single-cell RNA sequencing analysis identified major changes in the cellular populations of the lungs of patients with IPF including aberrant basaloid cells, myofibroblasts and invasive fibroblasts, vascular endothelial cells, and profibrotic macrophages [33]. This IPF cell atlas provides a resource toward understanding biomarkers on a cellular level. Two of the biomarkers significantly reduced by admilparant, MMP-7 [33–35] and TN-C [36], are highly expressed by aberrant basaloid cells in IPF. MMPs are endopeptidases that degrade ECM proteins, playing crucial roles in tissue repair and remodeling. MMP-7 is a highly validated peripheral marker of mortality in IPF [37, 38]. MMP-7 levels have been found to be significantly elevated in patients with IPF compared with healthy controls and patients with non-IPF ILDs [39]. Elevated MMP-7 was shown to be an early indicator of a progressive phenotype in clinical trials [17, 19] and was associated with poor outcomes in the PROFILE dataset [27, 40]. TN-C, found in the IPF lung in aberrant basaloid cells and myofibroblasts [33, 41], is an ECM glycoprotein sparsely expressed in normal tissue, but transiently and locally upregulated during inflammation, wound healing, and cancer invasion. Accumulating evidence indicates proinflammatory and profibrotic properties of TN-C, and it may be an early marker of inflammation and progressing fibrosis in various tissues [42]. Animal and in vitro studies suggest an involvement of TN-C in hypertrophic scar formation, highlighting its potential as a biomarker for lung fibrosis [42–44]. Normally secreted in small amounts by the pulmonary epithelium, CA-125 (also known as MUC16) is a protein that plays an important role in forming a protective mucous barrier and likely reflects proximalization of the distal lung [33]. It is highly expressed in patients with IPF [41], and elevated levels are considered a marker for epithelial damage [19] and progressive disease [28]. The significant decrease in CA-125 levels after admilparant treatment is consistent with reports after treatment with nintedanib, a tyrosine kinase inhibitor that targets multiple receptors involved in fibrogenesis, including vascular endothelial growth factor, fibroblast growth factor, and platelet-derived growth factor receptors [45]. In this exploratory analysis, reductions in CA-125 and TN-C were associated with a positive treatment response to admilparant at week 26. Taken together, these data suggest that admilparant treatment may reduce epithelial cell injury and fibroblast and basaloid cell activation in IPF.

The post hoc SomaScan analysis in the IPF cohort identified expression of biomarkers indicative of reduced fibrosis and improved mitochondrial and metabolic health upon treatment with admilparant. CKMT1A is responsible for the transfer of high-energy phosphate from mitochondria to the cytosolic carrier creatinine and is primarily associated with expression in brown adipose tissue as part of the uncoupled protein pathway [46–48]. One study has demonstrated that fibroblast conversion to brown adipose tissue was associated with a simultaneous increase in CKMT1A and adiponectin [49], which was also significantly increased in admilparant-treated patients. Increased plasma concentrations of SSBP1 may be indicative of improved mitochondrial function as it is an important protein for mitochondrial DNA replication and oxidative phosphorylation [50]. Neither CKMT1A nor SSBP1 are highly expressed in the lung [33]. CCDC80 is broadly involved in metabolism, as it improves insulin sensitivity in preclinical models of diabetes [51], and it is expressed in lung fibroblast subsets [33]. ANGPTL3 also has a metabolic function enabling the binding of fatty acids in peripheral blood [52]. Finally, a decrease in IGFBP6 may be linked to metabolic function as expression of this molecule is linked to lactate exposure in vitro [53]. IGFBP6 is increased with TGF-β signaling and expressed in lung fibroblasts [54].

Interestingly, a different set of circulating biomarkers was identified in the PPF cohort after treatment with admilparant, including periostin, several collagen degradation markers, and multiple inflammatory markers. Periostin, an ECM protein induced by TGF-β or IL-4/IL-13 in fibroblasts, is crucial in the pathogenesis of pulmonary fibrosis [20], and contributes to the pathophysiology of fibrosis during systemic sclerosis [55]. Elevated periostin levels may serve as a prognostic biomarker, indicating lung function decline and increased risk for mortality [17, 20, 21]. Significant reductions in the collagen degradation markers C3M and C4M after admilparant treatment are consistent with LPA₁ antagonism in a previous trial of BMS-986020 in patients with IPF [8, 12]. Collagen-producing cells maintain the lung architecture and drive pathologic scarring in pulmonary fibrosis [26]. Collagen turnover markers, including C1M, C3M, C4M, and C6M, have been reported to be elevated in patients with progressive disease compared with those with stable disease and to be associated with an increased risk of overall mortality in patients with IPF including the PROFILE dataset [24, 25, 40, 56].

Inflammation is a critical part of fibrogenesis in patients with PPF, especially those with underlying autoimmune diseases such as rheumatoid arthritis or systemic sclerosis [6, 57]. Inflammatory cytokines and chemokines are known to contribute to the interplay between the lung epithelium and fibroblasts. LPA/LPA₁ signaling may play a central role by modulating proinflammatory gene expression, endothelial cell activation, and leukocyte recruitment [58–60]. The reductions in multiple circulating inflammatory markers, including YKL-40, ferritin, CD163, VCAM-1, and PECAM-1 in patients with PPF after admilparant treatment are indicative of reduced disease activity and reduced inflammation. YKL-40 and CD163 are markers of monocyte-derived macrophages in IPF [33, 61, 62]. Elevated YKL-40 has been detected in lung tissues adjacent to fibrotic lesions, alveolar macrophages, and bronchiolar epithelial cells, and high serum YKL-40 levels have been correlated with poor prognosis [17, 61, 63]. CD163 is increased in patients with acute inflammatory diseases and has been associated with disease severity and mortality in some but not all studies in IPF [62, 64]. Ferritin, a crucial protein in iron metabolism and homeostasis, also plays a role in immune system regulation [22, 23]. These functions are essential for normal pulmonary physiology and biochemistry. Emerging evidence suggests that serum ferritin levels correlate with disease activity and poor prognosis in idiopathic ILD [22, 23]. VCAM-1 mediates leukocyte adhesion to the endothelium and plays a role in vasculopathy and inflammatory response [65]. Elevated VCAM-1 levels have been identified as a potential prognostic marker for lung function decline and survival in patients with IPF [38]. PECAM-1 is a transmembrane adhesion and signaling glycoprotein concentrated at endothelial cell junctions in highly vascularized regions and in non-activated leukocytes, and may play a role in the early stages of lung fibrosis [66].

This analysis has several limitations. The sample size in each cohort was small and the 26-week follow-up was relatively short for assessing biomarker changes. Pharmacodynamic analyses by response were not conducted in the PPF cohort due to limited subgroup sample sizes. The analyses were post hoc and the trial was not powered to detect treatment-related changes in biomarker signatures. The proteomic analysis in the IPF cohort was conducted post hoc and a similar analysis in the PPF cohort is actively ongoing. The proteomic analyses were conducted in the admilparant 60-mg dose group in alignment with the design of the phase 3 trials for both IPF and PPF, which include admilparant doses of 60 and 120 mg [67, 68].

Further studies are required to explore the biological mechanism underlying LPA₁ antagonism, and the treatment effect of admilparant in patients with IPF and PPF. Ongoing phase 3 clinical trials (ALOFT-IPF, NCT06003426; ALOFT-PPF, NCT06025578) [67, 68] aim to evaluate the effects of admilparant in large cohorts of patients for up to 3 years.

Conclusions

The exploratory analysis of biomarker profiles in patients with IPF and PPF over 26 weeks of admilparant treatment reveals a complex and multifaceted response that aligns with the broad effects of LPA₁ antagonism and inhibition of key pathways involved in lung fibrosis. In patients with IPF, changes in biomarkers of epithelial injury, fibrosis, and the potential antifibrotic biomarker adiponectin, suggest a focus on epithelial repair and a shift in metabolic processes. In patients with PPF, changes in markers of collagen degradation and inflammation indicate a more pronounced impact on active fibrosis and inflammatory processes in progressive disease. The divergent responses between IPF and PPF cohorts highlight the differences between these two fibrotic lung diseases and suggest that LPA₁ antagonism may have multiple on-target effects. As LPA₁ is involved in both inflammatory and metabolic processes that contribute to fibrosis, the observed differences in each cohort may reflect the factors contributing to the underlying disease. Furthermore, the biomarkers identified in this analysis may have potential value to monitor treatment response and disease activity in future pulmonary fibrosis clinical trials.

Abbreviations

ANGPTL3

Angiopoietin-like 3

C×M

Neoepitope of MMP-2,9,13–mediated degradation of type × collagen

CA-125

Cancer antigen 125

CCDC80

Coiled-coil domain-containing protein 80

CD163

Cluster of differentiation 163

CEA

Carcinoembryonic antigen

CfB

Change from baseline

CKMT1A

Mitochondrial creatine kinase

CTHRC1+

Collagen triple helix repeat containing 1

ECM

Extracellular matrix

EGF

Epidermal growth factor

EN-RAGE

Extracellular newly identified receptor for advanced glycation end-products binding protein

FVC

Forced vital capacity

IGFBP6

Insulin-like growth factor binding protein 6

IL-6Rβ

Interleukin−6 receptor subunit beta

IL-×

Interleukin ×

ILD

Interstitial lung disease

IPF

Idiopathic pulmonary fibrosis

KIM-1

Kidney injury molecule-1

LLOQ

Lower limit of quantification

LPA

Lysophosphatidic acid

LPA₁

Lysophosphatidic acid receptor 1

MMIF

Macrophage migration inhibitory factor

MIP

Macrophage inflammatory protein

MMP

Matrix metalloproteinase

PD-L2

Programmed death ligand 2

PECAM-1

Platelet endothelial cell adhesion molecule 1

PPF

Progressive pulmonary fibrosis

ppFVC

Percent of predicted forced vital capacity

PRO-C5

N-terminal type V collagen propeptide

PRO-FIB

Thrombin-mediated fibrinogen degradation

PROLIFIC

Prognostic Lung Fibrosis Consortium

SSBP1

Single-stranded DNA binding protein 1

TGF-β

Transforming growth factor β

TN-C

Tenascin C

VCAM-1

Vascular cell adhesion molecule 1

YKL-40

Chitinase-3-like protein 1

Authors’ contributions

The authors assume responsibility for data accuracy and completeness. All authors participated in reviewing and editing the manuscript and provided approval for publication. In addition, TMM contributed to the study design; NK and MF-C participated in data interpretation and biomarker prioritization; GH contributed to study conception, study design and data analysis. All authors had access to the data, which were analyzed and verified by MS, FW, CW, SS, PS, and YH. FW led the biomarker analyses and pharmacodynamic data generation. CW conducted the pharmacodynamic analyses.

Funding

This study was funded by Bristol Myers Squibb.

Data availability

Bristol Myers Squibb is committed to sharing data while maintaining the privacy of patients. View the policy on data sharing and submit a request for data at https://www.bms.com/researchers-and-partners/independent-research/data-sharing-request-process.html.

Declarations

Ethics approval and consent to participate

The study was conducted in accordance with the Declaration of Helsinki, the Council for International Organizations of Medical Sciences, Good Clinical Practice, the International Council for Harmonisation, and applicable local requirements. The protocol and documentation provided to patients were reviewed and approved by an independent review board and ethics committees before study initiation. All patients provided written informed consent.

Competing interests

TMM has received honoraria or speaker’s fees from AstraZeneca, Bayer, Boehringer Ingelheim, Bristol Myers Squibb (BMS), CSL Behring, FibroGen, Galapagos, Galecto, GlaxoSmithKline (GSK), IQVIA, Merck, Pliant, Pfizer, Qureight, Roche, Sanofi-Aventis, Structure Therapeutics, Trevi, and Veracyte. NK reports consulting to Boehringer Ingelheim, Pliant, GSK, Three Lake Partners, Merck, AstraZeneca, BMS, Galapagos, Chiesi, Sofinnova, and FibroGen; equity in Pliant; grants from AstraZeneca and BMS; and intellectual property on novel biomarkers and therapeutics in idiopathic pulmonary fibrosis licensed to biotech. MF-C has received honoraria or speaker’s fees or research support from Boehringer Ingelheim, Roche, MSD, AstraZeneca, Pfizer, GSK, Sanofi, and Novartis. FW, CW, and MS are employees and may own stock or shares in BMS. GH, PS, YH, and SS were employees of BMS at the time the study was conducted and may own stock or shares.