Proteomics profiling reveals distinct molecular subtypes of hypertrophic cardiomyopathy with differential risks of major adverse cardiovascular events

Abstract

Abstract Background Hypertrophic cardiomyopathy (HCM) is a heterogeneous disease, and a subset (10–15%) of patients per year experience major adverse cardiovascular events (MACE). Current risk prediction models based solely on clinical data have yielded limited power in predicting MACE. Although the diagnosis of HCM is made based on clinical features, the heterogeneity of HCM suggests that “HCM” may actually encompass several subtypes of disease with distinct biological mechanisms that can only be elucidated via molecular-level investigation. Proteomics profiling measures concentrations of thousands of proteins simultaneously and has the potential to reveal underlying molecular mechanisms of disease and distinct subtypes of HCM. Purpose Our aim was to derive distinct subtypes of HCM in our multi-center biorepository of patients with HCM using proteomics profiling, and to examine their associations with MACE. Methods We applied unsupervised machine learning methods to derive HCM molecular subtypes using plasma proteomics profiling data from 258 patients in our multi-center HCM prospective cohort study. The primary outcome was MACE, defined as a composite of arrhythmias, progressive heart failure, stroke, and sudden cardiac death. We performed time-to-event analysis with a Cox proportional hazards model, using the subtype with the lowest risk as the reference group. We also performed pathway analysis of proteins differentially regulated in patients in the highest-risk subtype. Results We identified 4 distinct HCM subtypes. Patients in subtype D (n=74) were older on average compared to the other groups and more likely to be female. 50% of patients in subtype D had NYHA class II or greater heart failure symptoms compared to 37% in the rest of the groups combined. There were no significant differences in echocardiographic parameters or rates of genotype positivity between the four subtypes. Compared to the reference group (i.e., subtype A), patients in subtype D had a higher rate of MACE with a hazard ratio of 2.59 (95% CI 1.21–5.57, p=0.01 compared to subtype A; Figure). Pathway analysis identified significantly upregulated signaling pathways in the high-risk subtype D compared to the other subtypes (Table), with a number of upstream and downstream pathways associated with the Ras-MAPK pathway (FDR 0.0005). Conclusions Our study exhibits not only the presence of HCM molecular subtypes but also their distinct pathobiological mechanisms associated with different MACE risks. The four subtypes identified bore similar clinical characteristics and would not have been distinguishable based on the conventional analysis of clinical and echocardiographic parameters alone, but rather required deeper investigation at the molecular level. Furthermore, pathway analysis revealed an upregulation of Ras-MAPK pathways in the subtype with the highest risk of MACE, implicating this pathway as a potential underlying mechanism of more severe disease. Funding Acknowledgement Type of funding sources: Public grant(s) – National budget only. Main funding source(s): NIH