Bringing doctors together

or

A free, private network for verified physicians

Leading voices

Dr. Bruno Bezerra Lima

Advanced Cardiac Imaging · VUMC

View

Dr. Sara Ersozlu

Advanced Cardiac Imaging · MGH

View

Dr. Christopher Kramer

Advanced Cardiac Imaging · University of Virginia

View

Dr. Filipe Moura

Advanced Cardiac Imaging · Yale College

View

Dr. Robert Hendel

Advanced Cardiac Imaging · Tulane University

View

Dr. Venkatesh Murthy

Advanced Cardiac Imaging · UMHS

View

Dr. Dorairaj Prabhakaran

General Cardiology

View

Dr. Pugazhendhi Vijayaraman

EP · Geisinger

View

Dr. Robert Schaller

EP · HUP

View

Dr. Trey Woods

General Cardiology · Emory Healthcare

View
  • Discover your community

  • Stay current

  • A profile that represents you

  • Find your people

About RoonPrivacyTerms

    Overview

    Proteomic Analysis of Valsartan for Attenuating Disease Evolution in Early Sarcomeric Hypertrophic Cardiomyopathy (VANISH) Clinical Trial.

    May 9, 2025DOI: 10.1161/circheartfailure.124.012393Journal ArticleAHA

    Written by: Constantin-Cristian TopriceanuChristoffer Rasmus VissingAnna Axelsson RajaSharlene M DayMark W RussellKenneth ZahkaAlexandre C PereiraSteven D ColanAnne M MurphyCharles CanterRichard G BachMatthew T WheelerJoseph W RossanoAnjali T OwensLuisa MestroniMatthew R G TaylorJames C MoonGabriella CapturAmit R PatelIvan WilmotJonathan H SoslowJason R BeckerChristine E SeidmanNeal K LakdawalaHenning BundgaardUsman A TahirCarolyn Y Ho

    Find these authors on Roon

    CT
    Dr. Constantin-Cristian Topriceanu
    Advanced Cardiac Imaging · Mayo Clinic College of Medicine

    INTRODUCTION

    Hypertrophic cardiomyopathy (HCM) is often caused by damaging variants in sarcomere genes. However, not all individuals who carry pathogenic or likely pathogenic variants (G+) develop clinically overt HCM, as traditionally defined by unexplained left ventricular hypertrophy (LVH) (maximal LV wall thickness [MWT] ≥13mm)1. The mechanisms through which sarcomere variants lead to clinically overt HCM (G+LVH+), as well as the factors which render some variant carriers with subclinical HCM (G+LVH-) more resistant or more susceptible to developing penetrant disease, have not been fully elucidated. Robust biomarkers that track phenotypic progression have not been identified, but are critically needed to reliably predict phenotypic conversion and to assess the impact of emerging therapies intended to intercept disease progression2.Given the absence of extensive, potentially irreversible, remodeling in patients with established disease, individuals with subclinical and early-stage HCM may be ideal targets for disease-modifying therapies intended to slow, or even halt, disease progression. Understanding circulating protein profiles associated with phenotypic progression could provide valuable insights about the biological pathways driving disease evolution. They may also serve as biomarkers to track these processes or monitor treatment effect. Previous proteomics studies suggested that pathways involved in myocardial fibrosis, energetics, and growth may be activated in overt HCM3–5. However, studies on the proteomics of early or subclinical HCM are scarce.Using proteomics data from the Valsartan for Attenuating Disease Evolution in Early Sarcomeric HCM (VANISH) clinical trial, we aimed to gain insights into potential circulating biomarkers that track disease progression or response to therapy. Since valsartan was previously shown to stabilize or improve cardiac remodeling in early HCM in the primary trial6, we also aimed to identify proteins associated with adverse remodeling.

    METHODS

    The corresponding author will consider requests for collaborative study and analysis of de-identified individual participant data to further understanding of the trial results or genetic cardiomyopathies. Interested researchers should submit proposals and analytic plans to the corresponding author, and approval will be granted by consensus of the executive committee. A data use agreement will be issued and will be compliant with relevant patient confidentiality and privacy regulations.
    The trial was approved by ethics committees at all sites and was overseen by an independent data safety monitoring board. All participants provided written consent; parental consent and child assent were obtained for youths.
    VANISH was a multi-center, double-blind, placebo-controlled, randomized clinical trial conducted between April 2014 to July 2019 in 17 HCM specialist centers across four countries (Brazil, Canada, Denmark, and the US). The full-details, including study design, inclusion and exclusion criteria, baseline characteristics, and main trial results were published elsewhere6–8. Briefly, all participants were required to carry a pathogenic/likely pathogenic (P/LP) sarcomere gene variant, classified according to standard criteria across all centers, accounting for segregation, conservation, public databases, and low frequency in control populations9,10. The study consisted of a primary cohort of participants with early non-obstructive HCM (age 8–45 years, LV MWT 12–25 mm in adults or body surface area (BSA)-adjusted LV MWT z-score 3–18 in children, New York Heart Association [NYHA] functional class I or II, absence of resting or provoked LV outflow gradients >30mmHg, LV ejection fraction≥55%, no secondary prevention implantable cardioverter-defibrillator [ICD] or history of appropriate ICD therapy, and no history of septal reduction therapy). In addition, an exploratory cohort of subclinical HCM was enrolled. These participants carried P/LP sarcomere variants, had normal LV wall thickness but exhibited minimal and non-diagnostic phenotypic manifestations of sarcomere variants (e.g., electrocardiographic [ECG] abnormalities other than non-specific ST-T wave change, an LV WT to LV internal diameter ratio ≥0.19 on structural cardiac imaging, or E’ velocity z-score ≤−1.5 on Doppler echocardiography). Eligible participants who successfully completed the active run-in phase were randomly assigned in a 1:1 ratio to receive valsartan or placebo for 2 years. Randomization was stratified based on age, pubertal status, NYHA I vs II, and LV MWT. VANISH participants were assessed at baseline, year 1, and year 2 via serum or plasma biomarkers, echocardiography, ECG, cardiovascular magnetic resonance imaging (CMR), cardiopulmonary testing, and health questionaries6.
    The VANISH trial aimed to explore whether valsartan could attenuate disease progression in early HCM. The effects of valsartan were assessed using a composite outcome representing the change from baseline of 9 components that reflect cardiac structure and function, including serum biomarkers quantified using a standard clinical assay (decrease in troponin T and N-terminal pro B-type natriuretic peptide [NT-proBNP] levels representing improvement), echocardiography metrics (decrease in LV MWT; increase in E’ velocity; and increase in S’ velocity) and CMR metrics (decrease in left atrial [LA] volume; increase in LV end-systolic (ESV) and end-diastolic volumes (EDV), and decrease in LV mass)6. These variables were either indexed to BSA (LA volume, LV mass, LV MWT, LV ESV, and LV EDV), age-adjusted (E’ and S’ velocities), or log-transformed (troponin and NT-proBNP). For each participant, the value of each metric at year 2 was subtracted from its baseline value. For each metric, these differences were converted to z-scores by subtracting the mean change and dividing by the standard deviation calculated across all participants. Lastly, for each participant, the nine z-scores were averaged to produce the composite z-score. A decrease (negative change in the composite z-score) reflects relative worsening and an increase (positive change) reflects relative improvement. This composite z-score is more advantageous than solely relying on LV MWT to assess for phenotypic progression, as it captures remodeling changes that extend beyond simple alterations in wall thickness and minimizes the impact of single measurement variability.
    Blood samples were collected at baseline and after 2 years of follow-up using K3 ethylenediaminetetraacetic acid tubes. At each center, the serum and plasma were separated within 60 minutes of phlebotomy, stored at −70 °C, and sent to Brigham and Women’s Hospital, where the proteomics analysis was conducted. In total, 276 proteins from the Cardiovascular Disease I & II and Cardiometabolic panels were quantified both at baseline and after 2 years of follow-up using a proximity extension assay (Target series, Olink, Sweden). The full methodological details have been published elsewhere11. Both the baseline and year 2 proteomics assays were run on the same plates. Proteomics data are often affected by non-biological systematic biases stemming from sample preparation, experimental conditions, and signal post-processing. To account for these biases and facilitate comparability across samples, the protein signal intensities were normalized via log2 transformation12. For each assay, the limit of detection (LOD) was calculated by the manufacturer as the average signal intensity of 12 healthy volunteers which acted as negative controls plus three standard deviations. In line with Olink’s recommendations13, assays in which more than 50% of the participants had a signal below the LOD were removed. We also removed data points with a quality control (QC) warning label, as assigned by Olink during protein quantification. Coefficients of variations in twelve healthy controls are presented in Supplemental Table S1. Of note, NT-proBNP levels used to derive the trial outcome were quantified using a standard clinical assay, while those included in the proteomics analysis were quantified using the Olink assay.
    All statistical analyses were performed in R version 4.2.2. The distribution of each variable was visually assessed using histograms. Continuous variables are expressed as mean ± 1 standard deviation (SD), or median (interquartile range) as appropriate, and categorical variables as counts and percentage. Approximately normally distributed continuous variables were compared using the t-test, while skewed continuous variables using the Mann-Whitney U-test. A two-tailed p-value <0.05 was considered statistically significant.
    Results were corrected for multiple testing at a false discovery rate (FDR) of 0.05, using the Benjamini Hochberg method14. The number of comparisons correspond to the number of genes, meaning that if two different proteins originate from the same gene, they are not considered separate analyses when correcting for multiple testing. The corrected p-values, accounting for multiple testing, are referred to as adjusted p-values.
    We assessed the correlations between normalized protein signal intensities and age across all participants by calculating the Pearson correlation coefficient (ρ). This analysis was conducted to ascertain whether the observed differences at baseline or temporal changes in protein levels could be attributed to age differences rather than disease progression. Proteins which were at least moderately correlated with age (i.e., |ρ| ≥ 0.4) were assumed to have changed during follow-up as a result of ageing.
    We compared protein abundance at baseline between early and subclinical HCM using the t-test. A protein was considered relatively differentially abundant if there was at least a 5% statistically significant difference in mean signal intensities. Lastly, we evaluated whether these differences persisted after adjusting for age, sex, and body mass index (BMI) using generalized linear models with Gaussian distribution and identity link Among proteins that remained relatively differentially abundant after adjustment and passed the correction for multiple testing (n=8), we aimed to identify the multi-protein panels that best distinguish subclinical from early HCM. Since for binary classification tasks it is recommended to have at least 10 observations per predictor15, and given that we have 32 individuals with subclinical HCM, we analyzed panels of up to three proteins, constructing logistic regression models for all possible combinations. For all logistic regression models, receiver operating characteristic curve (ROC) analyses were conducted and multi-protein panels were ranked based on the area under the ROC curve (AUC). In addition, logistic regression scores were calculated as score = intercept coefficient + Σ protein regression coefficient * measured normalized protein signal intensity. We also explored if these logistic regression scores based on these multi-protein panels that best distinguish subclinical from early HCM can also effectively differentiate convertors from non-convertors Mann–Whitney U test.
    We compared using the Mann–Whitney U test the abundance of proteins at baseline and change in mean signal intensities from baseline to year 2 between subclinical HCM participants who developed clinically overt HCM during follow-up versus those who did not. Only proteins that underwent at least a 5% increase or decrease in mean signal intensity in year 2 compared to baseline were considered. Given the small number of participants who had phenotypic conversion (n=5), further models adjusted for age, sex, and BMI were not pursued. However, correlations of proteins with age were considered when discussing whether differences in protein intensities may be due to disease progression rather than age differences. We also explored whether the change in protein signal intensities can track the change in absolute LV MWT in sarcomere variant carriers using an unadjusted generalized linear model.
    We aimed to identify proteins whose signal intensities changed between baseline and year 2, based on the phenotypic progression z-score in early HCM participants allocated to placebo. Generalized linear models with a Gaussian distribution and identity link were used to test for the association between the phenotypic progression z-scores and the percentage change in mean signal intensities between baseline and year 2. Models were adjusted for age, sex, and BMI. Proteins were ranked for their ability to track phenotypic progression based on the absolute percentage changes per one-unit worsening in the z-score. Although proteomic-derived NT-proBNP levels appeared to track phenotypic progression, the composite z-score outcome includes NT-proBNP quantified using a standard clinical assay. Therefore, sensitivity analyses were also performed, removing NT-proBNP from the z-score.
    Among proteins which passed the correction for multiple testing, we sought to identify multi-protein panels that best track disease progression. Multivariable generalized linear models with a Gaussian distribution, using the composite z-score as the dependent variable and the proteins in the panel as the independent variables, were compared using the likelihood ratio test (to determine model fit improvement), Adjusted McFadden’s R2 (to quantify variance explained), Akaike Information Criterion (to measure model parsimony), root mean square error (to assess prediction accuracy), and 5-fold cross-validation error (to evaluate generalizability). Models were adjusted for age, sex, and BMI.
    Valsartan was shown to improve cardiac structure and function in early HCM6, but no significant benefit was detected in the small exploratory cohort with subclinical HCM8. Therefore, we analyzed only the early HCM cohort to identify proteins whose absolute signal intensities changed between baseline and year 2, comparing those allocated to valsartan or placebo. We used a linear mixed model, accounting for subject-specific variability and the non-independence of the repeated measures. The model was adjusted for age, sex, and BMI.
    We also aimed to understand whether the efficacy of valsartan on the phenotypic progression composite z-score is influenced by the baseline levels of proteins. To calculate the effect of valsartan on the composite z-score, we fitted a generalized linear model with a Gaussian distribution and identity link, adjusted for age, sex, and BMI. Building upon this model, we tested the interaction of each protein with treatment allocation across all early HCM participants, to identify potential effect modifiers.

    RESULTS

    Study flow is depicted in Figure 1. Baseline proteomics data were available in 192 (91%) out of the 212 randomized participants (160 early-stage HCM and 32 subclinical HCM). In 20 participants, proteomics analysis was not performed either because a blood sample was not available or due to early study termination. The baseline characteristics of the 192 participants included in this study are presented in Table 1. The characteristics of the participants with (n=192) and without proteomics (n=20) data were similar (Supplemental Table S2). Of the 192 included study participants, a subset of proteomic data were not available in 6 participants due to QC warnings. Because a significant treatment effect of valsartan was not detected in subclinical HCM8, this exploratory cohort is not stratified by treatment allocation.
    The 276 proteins assessed and myocardial expression (based on RNA normalized transcripts per million [nTPM] according to the Human Protein Atlas [https://www.proteinatlas.org/]16) are summarized in Supplemental Table S1. Of these, 12 proteins (4%) were removed because more than 50% of the participants had protein signal intensities below the LOD.
    Twenty-eight proteins showed a relative difference in abundance in participants with early-stage versus subclinical HCM (i.e., ≥5% statistically significant difference in mean signal intensities), of which 12 passed correction for multiple testing at a FDR of 5% (Supplemental Table S3; shown in green in Figure 2). Of these, 9 had higher relative abundance in early HCM. The genes encoding all these proteins are expressed in the myocardium, except hydroxyacid oxidase and matrix extracellular phosphoglycoprotein [MEPE] (Supplemental Table S1). Proteins involved in immune regulation (e.g., chitinase-3-like protein 1 [CHI3L1] and lipocalin 2 [LCN2]), extracellular matrix (ECM) remodeling (e.g., tissue plasminogen activator [tPA]), cellular growth (e.g., fibroblast growth factor 21 [FGF21]), and markers of hemodynamic stress (NT-proBNP and angiotensin converting enzyme 2 [ACE2]) had a higher relative abundance in early HCM. In contrast, MEPE and osteopontin, which are ECM remodeling proteins, had lower relative abundance in early HCM compared with subclinical HCM.
    Of note, individuals with early-stage HCM were older (mean age 24 vs 16 years, p<0.001). ACE2 (ρ = 0.43) levels were positively correlated with age (ρ = 0.43, p<0.001), while MEPE (ρ= −0.71) and osteopontin (ρ = −0.67) were negatively correlated with age (all p<0.001) (Supplemental Table S1). After adjusting for age, sex, and BMI, individuals with early HCM still had a higher relative abundance of neutrophil gelatinase-associated lipocalin 2, BNP, NT-proBNP, FGF21, tPA, CHI3L1, lectin-like oxidized low-density lipoprotein receptor 1 (LOX1), and a lower abundance of interleukin-17 receptor A (IL-17 RA) (Supplemental Table S3).
    Panels consisting of up to 3 proteins were tested to differentiate early and subclinical HCM (logistic regression models are presented in Supplemental Table S4). A panel consisting of NT-proBNP, tPA, and IL-17 RA achieved a sensitivity of 91%, specificity of 66%, and an AUC of 0.85, when distinguishing subclinical from early HCM (p-value<0.001). This was followed by tPA, LOX1, and IL-17 RA (AUC 0.83), and LCN2, NT-proBNP, and tPA (AUC 0.82).
    During the study period, six of thirty-four participants (18%) with subclinical HCM underwent phenotypic conversion with an increase in LVWT that qualified for a diagnosis of HCM. Baseline and follow up proteomics data were available in five of these individuals. Age at baseline was similar in convertors and non-convertors (15 vs 16 years, p=0.26). At baseline, participants who underwent phenotypic conversion had a higher abundance of four proteins (ACE2, IL-1 receptor-like 2, renin, and matrix metalloprotease 7) and a lower abundance in another four (fatty acid binding protein 6, insulin-like growth factor [IGF] binding protein 1, growth hormone, and granulysin) compared to those who did not undergo phenotypic conversion (n=27) (Figure 3). However, none remained significant after correcting for multiple testing using an FDR of 5% (Supplemental Table S5). In addition, no proteins had a statistically significant change in mean signal intensity from baseline and year 2 in convertors versus non-convertors. NT-proBNP increased by 10% on average during follow-up in convertors, compared to 1% in non-convertors, albeit the difference was not statistically significant. Lastly, no proteins were identified whose change in signal intensity from baseline to year 2 tracked with the absolute change in LV MWT during the same period. In contrast to analysis of single proteins, the logistic regression scores that were able to distinguish subclinical and early HCM based on the multi-protein panels LCN2, LOX1, and NT-proBNP (p-value=0.024), and LCN2, LOX1, and IL-17RA (p-value=0.025), were also able to distinguish phenotypic convertors vs non-convertors (Supplemental Table S4).
    Protein associations with phenotypic progression and the natural history of early HCM were assessed by studying the 79 participants with early HCM allocated to placebo. Of these, 42 (53%) had phenotypic worsening (negative change in the composite z-score) during the 2 years of follow-up. The change in protein levels between baseline and year 2 blood samples was compared to changes in the natural history of disease over the study period. A one-unit worsening of the phenotypic progression z-score during follow-up was associated with an increase in 16 proteins and a decrease in one protein from baseline to year 2, after adjusting for age, sex, and BMI (Supplemental Table S6; Figure 4). After correcting for multiple testing, a one-unit worsening of the phenotypic progression z-score remained significantly associated with an increase in three proteins: NT-proBNP by 27% (95% CI: 17–37%); ACE2 by 10% (95% CI: 5–15%); decorin (an ECM protein highly expressed [nTPM 1254] in the myocardium) by 6% (95% CI: 3–9%) (Supplemental Table S1). In sensitivity analyses removing NT-proBNP from the composite z-score, circulating NT-proBNP levels remained significantly associated with the other metrics of phenotypic progression, increasing by 18% (95% CI: 7–29%) per one-unit worsening of the revised z-score. All four multi-protein panels based on combinations of NT-proBNP, ACE2, and decorin were able to track the phenotypic progression z-score (all p-values <0.001), after adjusting for age, sex, and BMI (Supplemental Table S7). The combination of NT-proBNP with either ACE2 or decorin was better at tracking phenotypic progression compared to NT-proBNP alone (both likelihood ratio test p-values ≤0.021). However, the multi-protein panel consisting of NT-proBNP, ACE2, and decorin demonstrated the best performance, explaining the most variance (37%), achieving the highest prediction accuracy (root mean square error =0.33), and demonstrating the best generalizability in 5-fold cross-validation (error =0.14).
    Baseline and end of study proteomics data were available in 81 participants with early HCM allocated to valsartan and 79 allocated to placebo. As expected, renin levels increased from baseline to year 2 in participants treated with valsartan17. In addition, there was a relative decrease in six proteins (IgG Fc receptor II-b, α-L-iduronidase, galectin 9, interleukin-1 receptor antagonist, NT-proBNP, and nuclear factor-κB [NF-κB] essential modulator) in those treated with valsartan, after adjusting for age, sex, and BMI (Figure 5; Supplemental Table S8). All of these proteins are expressed in the myocardium (Supplemental Table S1). However, only renin and remained significant after correction for multiple testing using a FDR of 5%.
    Valsartan had a beneficial effect on cardiac remodeling compared to placebo, as reflected by an increase in the composite z-score of 0.23 (95% CI: 0.10–0.36)6. Valsartan treatment was associated with lower serum levels of tyrosine-protein kinase receptor Tie-1, poly [ADP-ribose] polymerase 1, NT-proBNP, and BNP (Supplemental Table S9). However, changes in protein levels did not remain significantly associated with treatment after correction for multiple testing.

    DISCUSSION

    In this exploratory secondary analysis of the VANISH trial, we investigated circulating proteins in the early stages of HCM, aiming to identify biomarkers which may be able to track phenotypic progression, phenotypic conversion, and response to treatment. Although most proteins had a similar abundance in early-stage and subclinical HCM, 12 proteins had relatively different abundance between these two groups after correcting for multiple testing. These proteins were markers of hemodynamic stress (NT-proBNP, ACE2), involved in cellular growth (FGF21), or associated with ECM remodeling (tPA). A multi-protein panel consisting of NT-proBNP, tPA, and IL-17 RA achieved a sensitivity of 91% and specificity of 66% in differentiating subclinical from early HCM. The combination of NT-proBNP, ACE2, and decorin was better at tracking phenotypic progression than NT-proBNP alone. When assessing protein signatures of participants with subclinical HCM who experienced phenotypic conversion to clinically overt HCM, no individual proteins were able to discriminate convertors from non-convertors or tracked with the absolute change in LV MWT during the same period. However, the logistic regression scores that successfully distinguished subclinical and early HCM using the multi-protein panels LCN2, LOX1, and either NT-proBNP or IL-17RA, also proved effective in differentiating convertors from non-convertors. Finally, placebo-treated participants with early HCM were studied to identify proteins associated with phenotypic progression. Significant associations were identified with markers of hemodynamic stress (NT-proBNP, ACE2) or ECM remodeling (decorin).
    Developing disease-modifying therapies that will slow phenotypic progression or prevent the emergence of HCM in subclinical variant carriers is a high priority. Potential approaches include valsartan6, cardiac myosin inhbitors18, or gene therapy19. Both the subclinical and early stages of overt HCM are attractive target populations for therapies intended to intercept disease progression given the absence of extensive fibrosis and adverse remodeling. To support these efforts, biomarkers that can reliably track disease progression and response to therapy must be identified. In the VANISH trial, NT-proBNP was the most robust circulating biomarker for tracking disease progression, since it had a higher abundance in early compared to subclinical HCM and tracked phenotypic progression in placebo-treated participants with early HCM. Participants treated with valsartan had a decrease in NT-proBNP, but this did not pass correction for multiple testing in this study which accounted for 263 simultaneous comparisons. These findings add to existing literature demonstrating the utility of NT-proBNP as a biomarker in the earliest stages of disease. Prior studies have suggested that individuals with early HCM have higher NT-proBNP compared to subclinical HCM20, although subclinical variant carriers with higher serum NT-proBNP were more likely to develop overt HCM over a median follow-up of 3 years starting from a mean age of 16 years21. Additionally, NT-proBNP correlated with extracellular matrix volume by CMR; a marker of interstitial fibrosis and ECM remodeling in the heart22. The prognostic value of a normal NT-proBNP which is stable during long-term follow-up in sarcomere variant carriers is yet to be established.
    Little is known about the molecular pathways that drive phenotypic progression. Studying changes in circulating proteins in different stages of disease may provide insights into the underlying processes. In this study, we identified several categories of proteins associated with disease progression by demonstrating differential expression in subclinical versus early-stage HCM or by association with adverse remodeling as reflected by worsening of the composite z-score for phenotypic progression. Both NT-proBNP and ACE2, markers of hemodynamic stress, were significantly higher in early versus subclinical HCM and in participants who demonstrated phenotypic progression. Pathways involved in myocardial fibrosis may also be activated in subclinical HCM22. Indeed, certain proteins associated with ECM remodeling had different relative abundance in subclinical compared to early HCM in this study. Although there are age-related changes in many ECM-related proteins (MEPE or osteopontin), we found that tPA had a higher abundance in early versus subclinical HCM after correction for age, sex, and BMI. This suggests that these proteins may be more influenced by disease stage than the age difference between the cohorts. Interestingly, markers of inflammation such as LCN2, CHI3L1, and lectin-like oxidized low-density lipoprotein receptor 1 had a higher abundance in early than subclinical HCM. However, whether inflammation precedes ECM fibrotic remodeling has not yet been established.
    In this study, growth factors and regulators of cellular growth had a higher abundance in early compared to subclinical HCM (e.g., FGF21 and CHI3L123). These may be related to mitogen activated-protein kinase (MAPK) or mammalian target of rapamycin (mTOR) signaling. mTOR inhibitors were shown to attenuate LVH in animal HCM models24–27. However, their role in altering disease progression in HCM in humans has not yet been established.
    The study only included sarcomere variant carriers, and the findings might not be generalizable to non-sarcomeric HCM. The 276 proteins were derived from Olink’s Cardiovascular Disease I & II and Cardiometabolic panels, rather than from a single panel. Although using multiple panels may limit direct comparisons or results integration, we mitigated this issue by adjusting for confounders. The small sample size makes this study prone to type 2 errors. However, we were unable to group these proteins into mechanistic or correlation-based clusters to reduce dimensionality and increase statistical power as proteins were involved in multiple pathways. An inherent limitation is the use of serum samples is that circulating proteins may not reflect the proteomics profile in the myocardium. Including only proteins from Olink’s cardiometabolic panel may have limited our ability to fully investigate all relevant biological pathways. Indeed, we assessed a limited number of proteins involved in pathways of interest for HCM such as myocardial fibrosis, growth, metabolism, and response to microvascular dysfunction. Because the proteomics panel employed had relatively few fibrosis-related proteins, the disease-modifying role of valsartan on fibrosis may not be fully explored.

    CONCLUSION

    In this study we examined circulating proteomic profiles in subclinical and early-stage HCM, longitudinally to assess short-term natural history, and response to putative disease-modifying treatment. NT-proBNP was the most informative protein, showing a higher abundance in early-stage HCM compared to subclinical HCM and an association with the phenotypic progression z-score in early-stage HCM. These results suggest a potential role as a biomarker for tracking disease progression. ACE2 was also relatively more abundant in early-stage versus subclinical HCM and in participants that demonstrated phenotypic progression. Proteins involved in extracellular matrix remodeling associated with disease stage (tPA) or were found to be associated with phenotypic progression in early-stage HCM (decorin). Lastly, some growth factors and regulators of cellular growth had a higher abundance in early compared to subclinical HCM (FGF21 and CHI3L1). The multi-protein panels LCN2, LOX1, and either NT-proBNP or IL-17RA, may be able to differentiate both subclinical and early HCM as well as convertors and non-convertors. The use of multi-protein panels may further improve discrimination. Additional studies across the spectrum of HCM, from subclinical to advanced stages, are needed to explore biologically relevant pathways and to identify more robust biomarkers to monitor phenotypic progression and response to treatment.
    CTDr. Constantin-Cristian Topriceanu

    Dr. Constantin-Cristian Topriceanu

    Cardiology · Advanced Cardiac Imaging
    Mayo Clinic
    constantincristian