Proteomics profiling allows us to measure thousands of plasma proteins at one specific timepoint. Such analysis can be used for two different purposes:  to find novel biomarkers that can help us in this risk stratification or (2) to find clusters of increased proteins that could indicate activity of specific pathophysiological pathways and thus provide mechanistic insights. The use of proteomics for these purposes in different cardiovascular diseases has exponentially increased over the last years, mainly due to improved methodology and lower costs. Therefore, proteomics profiling could provide answers to the outstanding questions in the risk stratification of HCM patients.

Lumish et al applied the use op proteomics profiling in HCM patients in a multicentre, prospective cohort study to determine whether proteomics can predict worsening of HF in HCM patients and to identify pathophysiological pathways associated with worsening of HF. In this well-designed study, 389 patients with HCM were included, and plasma proteomics profiling of 4986 different proteins was performed. Results were validated in an external cohort of 121 HCM patients. Using random forest modelling, a 11-protein proteomics-based model was derived from the initial cohort that predicted worsening of HF defined as an increase in NYHA functional class by at least one class. Secondary endpoints were the progression to either NYHA class III or IV and hospitalisation for HF. The 11-protein proteomics based model could predict worsening of HF in the validation cohort with an area under the curve (AUC) of 0.87. In total, there were 1273 differentially regulated proteins between patients with and without worsening of HF. In those patients with worsening of HF, there was an increase of proteins involved in Ras-MAPK signalling, and its upstream PI3K-Akt pathway, among other pathways.

The study adds valuable information to the field of plasma proteomics and HCM but should also be interpreted in their corresponding context. The model was developed for the specific endpoint of an increase in NYHA functional class. For any other outcome, such as heart transplantation, sudden cardiac death or need for a septal myectomy, the model could significantly differ. The current study was not designed and powered to investigate the translation of their findings to these endpoints. The model was applied to predict HF hospitalisation, that occurred in 15% of the cohort and had a notable AUC of 0.8. The translation of proteomics profiling towards clinical use has been challenging, as we are not used to measuring a panel of proteins that results in a single result. Although the 11-protein proteomics based model did not contain NT-proBNP and troponin T, these biomarkers were significantly associated with worsening of HF. Validation of the small panel of 11 proteins in multiple cohorts, settings in relation to different endpoints will be necessary to determine the value in a clinical context.

Another important consideration in the interpretation of the results is that we are analysing the plasma of these patients. This is a systemic approach in which we also detect proteins that are associated with (dys)function of other organs and diseases. Therefore, it would be highly interesting if such efforts could be repeated using cardiac tissue from HCM patients (eg, after a septal myectomy) for proteomics profiling. Noteworthy, the stage of disease is likely to be different when using tissue from patients undergoing myectomy. The combination of plasma and cardiac proteomics could therefore provide answers to the potential cardiac origin of identified plasma proteins.

Using a different statistical approach, the authors were able to identify clusters of proteins that resembled specific signalling pathways. For example, in patients with worsening of HF, there was an increase of proteins involved in Ras-MAPK signalling. Ras signalling has been associated with HCM, and interestingly, germline mutations in Ras signalling lead to a group of syndromes in which HCM is a common phenotype. The fact that proteins of the Ras-MAPK signalling was increased in those with worsening with HF shows that the pathway is not just associated with HCM as a phenotype but also is associated with progression of HF. These are promising pilot results and could pave the way towards novel treatment strategies involving Ras inhibition. Mitogen-activated protein kinase (MEK) is one of the kinases involved in the Ras signalling, for which multiple inhibitors have been developed, for example, trametinib and selumetinib. These MEK inhibitors are used to treat cancers with a somatic BRAF mutation and inoperable plexiform neurofibromas in children with neurofibromatosis type 1. However, several reports show the benefit of off-label treatment with trametinib in the treatment of cardiac hypertrophy in children with RASopathies, for example, Noonan and Costello syndrome. In this specific population, inhibition of the Ras pathway was able to reverse HCM and valvular obstruction; however, in adult non-syndromic forms of HCM, the therapy effect remains unknown. Besides the Ras-MAPK pathway, other clusters such as ‘focal adhesion’ and ‘PI3K-Akt signalling’ were also enriched in the plasma. There is still a lot of potential for future studies to dive further into the molecular pathophysiology of left ventricular hypertrophy and HCM.

One important opportunity for future research would be the use of circulating markers to improve the risk stratification for sudden cardiac death (SCD) in HCM patients. SCD is the most devastating complication of HCM. Identifying those at increased risk remains challenging but could be life-saving by the use of an implantable cardioverter-defibrillator (ICD). Different risk stratification and criteria are used by the European and American guidelines in this setting, illustrating the challenges faced with risk stratification for SCD. The authors provided the ESC HCM risk SCD, which recommends stratifying patients by calculating the 5-year SCD risk score according to a mathematical model. The authors demonstrated no significant difference in HCM risk-SCD scores between patients with no worsening and worsening HF. The scores were relatively low, indicating a tolerable risk of SCD in the majority of patients. Of note, in May 2019, a new American College of Cardiology/American Heart Association (ACC/AHA) strategy for the prevention of SCD was published, also including extensive fibrosis visualised by late gadolinium enhancement (LGE) on cardiac MRI. Herein, fibrosis was a key driver of improving the former risk stratification for SCD. Similar observations are seen in dilated cardiomyopathy, where presence of LGE is currently also applied to lower the threshold of ICD implantation. The current authors identified mainly proteins related to inflammation and cell cycle, which could lead to the development of fibrosis. Unfortunately, the low SCD rate refrained the authors from further analysis. Identifying markers that correlate with myocardial fibrosis would be highly interesting and potentially alter future risk stratification in HCM patients.

Patients with HCM are phenotypically heterogeneous: there are multiple subgroups to distinguish. For example, about 50% of HCM patients have an underlying genetic aetiology, mainly pathogenic variants in MYBPC3 or MYH7. A pathogenic variant in a key element of the sarcomere will have profound effects on the pathophysiology process that is driving hypertrophy in a patient, for example, transcriptomic analysis in biopsies from dilated cardiomyopathy patients with different genetic etiologies revealed distinct pathophysiological processes in the heart. In the current study, there was no distinction between aetiologies, and the plasma proteomics was averaged among the whole cohort. Future efforts should focus on in-depth phenotyping of HCM cohorts followed by omics analysis to compare profiles between subgroups. This could be a sequential approach: phenotyping followed by -omics analysis, or an integrated approach where -omics data are combined with clinical data to define subgroups of HCM. The identification of phenotypical subgroups with specific underlying pathophysiological processes driving the disease will allow us to better tailor medication to the right patients and potentially identify novel pathways that could function as targets for treatment.

https://heart.bmj.com/content/109/24/1808?utm_source=alert&utm_medium=email&utm_campaign=heart&utm_content=toc&utm_term=27112023&utm_campaign=heartjnl-toc%20%28live%29&utm_medium=email&utm_source=adestra