Molecular signatures of accelerated ageing in heart failure: a multi-protein analysis
J Lee, W Ouwerkerk, I Kassam, S P Chan, N Barascuk-Michaelsen, J Korbinian Vogt, R Linnemann Nielsen, R Foo, A A Voors, C S P Lam, A M Richards, J TrompAbstract
Background
Age is a significant risk factor for heart failure (HF), but the biological mechanisms underlying this relationship remain poorly understood. Therefore, we evaluated the role of biological age in HF with reduced (HFrEF) and preserved (HFpEF) ejection fraction.
Methods
We assessed biological age in 2,121 participants from the prospective Asian neTwork for Translational Research and Cardiovascular Trials (ATTRaCT) cohort using a proteomic ageing clock developed using 2,814 proteins measured by proximity extension assay. The clock was constructed using 179 ageing proteins identified by an Elastic Net model in 927 controls without HF and subsequently applied to estimate biological age in 1,194 patients with HF. Next, we compared biological versus chronological (calendar) age in patients with HFrEF and HFpEF, and identified pathways associated with biological age. Observational results were independently validated in 2,718 participants from UK-BIOBANK and 577 participants from BIOSTAT-CHF. Lastly, we investigated the causal relationship between biological age and HF progression using Mendelian randomization in 139,533 participants from the HERMES study.
Results
Patients with HFrEF and HFpEF were biologically 5.1 (±18.7) and 3.3 (±15.3) years older, respectively, than their chronological age (Figure 1A). Biologically older patients were more often men, had a greater comorbidity burden, and worse HF signs and symptoms. Biological age was associated with pathophysiological pathways involving inflammation (promoting maladaptive cardiac remodeling and impaired myocardial function) and extracellular matrix organization (related to cardiac structural homeostasis), regardless of HF phenotypes (Figure 1B-C). Biological age outperformed chronological age in predicting 2-year composite outcomes of mortality or HF hospitalization across cohorts, with stronger associations and lower AIC values for biological age in both HF phenotypes (adjusted HR [95% CI]: HFrEF, 1.5 [1.3–1.8] vs. 1.1 [1.0–1.3]; HFpEF, 1.7 [1.3–2.3] vs. 1.2 [1.0–1.5]). Among ageing protein, COL4A1 (implicated in ventricular stiffening and contractile dysfunction via type IV collagen accumulation and basement membrane defects) was causally associated with the progression of HFrEF and HFpEF. WNT9A (associated with hypertrophic cardiomyopathy risk) was uniquely causally associated with HFpEF progression, and TNC (a driver of fibrosis, pathological hypertrophy, and myocardial dysfunction) was uniquely causally associated with HFrEF progression (Figure 2).
Conclusion
Patients with HF are biologically older than their chronological age. This is the first study to show that biological age is independently and causally related to HF progression, suggesting that it might be a modifiable pathophysiological target in HF.Biological age and 179 ageing proteinsFor image description, please refer to the figure legend and surrounding text.Mendelian randomization on heart failureFor image description, please refer to the figure legend and surrounding text.