Deep Phenotyping of Cardiomyopathies Identifies Gremlin-1 as a Disease-Specific, Prognostic Biomarker Linked to Cardiac Remodelling and Poor Outcomes of Heart Failure Patients

Background
Cardiomyopathies are a leading cause of progressive heart failure (HF), and despite significant advancements in diagnostics and treatment, mortality rates remain high. The current classification of non-ischemic cardiomyopathies (NICM) is based on a phenotype-driven clinical workflow that includes endomyocardial biopsy (EMB). Inflammation and fibrosis are key factors in the progression of HF and contribute to adverse cardiovascular (CV) events. Although the diagnostic algorithm aids in risk stratification of HF patients, many risk factors are not apparent before disease onset.

Objective
This study aims to analyze EMB patterns in NICM patients and explore fibroinflammatory changes in the myocardium that are linked to phenotypic diversity and HF progression.

Methods
We prospectively enrolled 703 HF patients with NICM in a large-scale, all-comers cohort. All participants underwent guideline-based phenotypic classification, including EMB and cardiac imaging. RNA sequencing, histological, and immunohistochemical analyses were performed on cardiac tissue samples. A ten-year follow-up was conducted to monitor adverse CV events associated with HF progression.

Results
Guideline-based classification of cardiomyopathy patients identified distinct etiological risk groups (Figure 1A). Patients with symptomatic cardiomyopathy exhibited increased expression of pro-inflammatory and mediators in the myocardium leading to enhanced adverse fibrotic remodeling (Figure 1B&C). Notably, elevated Gremlin-1 expression, a key downstream mediator of the TGF-beta pathway, was associated with pro-fibrotic cardiac remodeling linked to poor prognosis (Figure 1D). Further, Gremlin-1 was associated with enriched expression of pro-inflammatory mediators within the myocardium (Figure 1E). Strikingly, Gremlin-1 was highly expressed in patients with NICM when compared to non-failing heart controls underscoring its role in the pathophysiology in progressive HF (Figure 1F). Higher levels of Gremlin-1 correlated with reduced LVEF and showed an inverse relationship with LGE (Figure 2A&B). The transcriptomic landscape was significantly altered between NICM phenotypes, hinting at adverse cardiac remodeling (Figure 2C). Additionally, fibrotic and inflammatory RNA signaling pathways were enriched by Gremlin-1 expression, indicating potential pathophysiological cascades (Figure 2D&E).

Strikingly, Gremlin-1 expression was independently linked to increased CV risk over the ten-year follow-up period (Figure 3A&B). Thus, patients with Gremlin-1+ EMB had a higher risk of cardiac and all-cause mortality. The incidence of ICD implantation and appropriate ICD discharges was significantly higher in patients with Gremlin-1+ EMB during follow-up (Figure 3C&D). Machine learning models incorporating Gremlin-1 expression enhanced ten-year risk stratification among patients with symptomatic cardiomyopathy (Figure 3E). Thus, risk prediction including cardiac Gremlin-1 was strongly associated with reliable risk factors of disease activity in NICM patients (Figure 3F).

Conclusion
Our findings indicate that Gremlin-1 is associated with inflammation and cardiac remodeling in NICM patients. Patients with Gremlin-1⁺ EMB have a higher risk of adverse CV events, making Gremlin-1 a potential biomarker for NICM. Thus, histological evaluation of Gremlin-1 could help identify underlying pathophysiological mechanisms and improve early risk stratification and management of HF patients.