Plasmatic and myocardial microRNA profiles in patients with Hypertrophic Cardiomyopathy

Dear Editor, MicroRNAs (miRs) have emerged as cardiovascular biomarkers and myocardial regulators with diagnostic and therapeutic potential.1,2 However, the miR profile of patients with hypertrophic cardiomyopathy (HCM) and miRs role in this genetic disease with heterogeneous phenotype3 are incompletely determined. Here amulti-step strategy provided evidence of differentially expressed miRs (DEmiRs) in plasma and myocardial tissue fromHCM(SupplementaryMaterials:Materials and Methods). In the first step, a profile of 1128 expressed mature miRs was identified (Figure 1) in 25 plasma samples from a cohort of 36 HCM (Figure S1, Tables S1 and S2) and healthy individuals (CTRL, n = 11) by next generation sequencing (NGS). Principal component analysis of NGS data did not clearly define a separation betweenHCMandCTRL, while a significant differential expression was found in a subset of 139 plasma miRs. Twenty-eight of them were DEmiRs significant both for p value and false discovery rate (Figures 1B and 1C, Tables S3–S4), and demonstrated 1626 putative targets by in silico search with miRTargetLink Human engine. Among them, 50 target genes were interconnected with 13 DEmiRs by “strong” experimental methods like reporter gene assay (Figure S2A), corresponded to a complex protein network (Figure S2B) with associated functions/pathways (Supplemental_Enrichment). To validate NGS performance and exclude false DEmiRs due to high specificity but moderate sensitivity of the assay,4 37 miRs (16 DEmiRs, 21 identified plasma miRs involved in cardiovascular diseases/proposed as biomarkers for cardiomyopathy or heart failure [HF] by other studies) were determined in 22 HCM, 10 CTRL by quantitative real-time polymerase chain reaction (RT-qPCR). Eight plasma DEmiRs were confirmed by RT-qPCR: hsa-miR19a-3p, hsa-miR-20b -5p, hsa-miR-29b-3p, hsa-miR-1265p, hsa-miR-144-3p, hsa-miR-454-3p and hsa-miR-47325p were up-regulated, and hsa-miR-182-5p was down-

Dear Editor, MicroRNAs (miRs) have emerged as cardiovascular biomarkers and myocardial regulators with diagnostic and therapeutic potential. 1,2 However, the miR profile of patients with hypertrophic cardiomyopathy (HCM) and miRs role in this genetic disease with heterogeneous phenotype 3 are incompletely determined.
Here a multi-step strategy provided evidence of differentially expressed miRs (DEmiRs) in plasma and myocardial tissue from HCM (Supplementary Materials: Materials and Methods).
In the first step, a profile of 1128 expressed mature miRs was identified ( Figure 1) in 25 plasma samples from a cohort of 36 HCM ( Figure S1, Tables S1 and S2) and healthy individuals (CTRL, n = 11) by next generation sequencing (NGS). Principal component analysis of NGS data did not clearly define a separation between HCM and CTRL, while a significant differential expression was found in a subset of 139 plasma miRs. Twenty-eight of them were DEmiRs significant both for p value and false discovery rate (Figures 1B and 1C, Tables S3-S4), and demonstrated 1626 putative targets by in silico search with miRTargetLink Human engine. Among them, 50 target genes were interconnected with 13 DEmiRs by "strong" experimental methods like reporter gene assay ( Figure S2A), corresponded to a complex protein network ( Figure S2B) with associated functions/pathways (Supplemental_Enrichment).
The Spearman's rank coefficient calculation differently correlated pairs of miRs highly-expressed by RT-qPCR in HCM and CTRL ( Figure S4A), and a significant linear relationship relating hsa-miR-144-3p to both hsa-miR-20b-5p and hsa-miR-25-3p in HCM was found ( Figure 2E). These three miRs were interconnected by 93 predicted targets, including PTEN and BCL2L11 genes backed-up with strong evidence ( Figure S4B), and leading to network of 17 proteins involved in pathways of cardiac relevance such as the regulation of cardiac muscle cell proliferation and of response to endoplasmic reticulum stress ( Figure S4C). Furthermore, the linear association between plasma levels of hsa-miR-144-3p and hsa-miR-20b-5p and % ejection fraction (EF) found in the HCM population ( Figure 2F, Table  S6) suggested a role for these two miRs in the regulation of myocardial function.
Thereafter, 20 miRs already validated in the plasma (including all up-regulated DEmiRs) were determined in septal myectomy samples from 21 obstructive HCM of our cohort and nine donor hearts (ctrl). RT-qPCR showed three downregulated (hsa-miR-144-3p, hsa-miR-451a and hsa-miR-223-3p, Figure 3A) and two up-regulated (hsa-miR-374b-5p and hsa-miR-4485-3p, Figure 3B Mann-Whitney test is applied and significant differences are shown as *p < 0.05 and **p < 0.01, ***p < 0.001, ****p < 0.0001. Receiver-operator characteristic (ROC) curve analysis of plasma DEmiRs is shown (B). AUC > 0.7 with significant p values was considered as threshold for good discriminant performance. The network obtained by miR TargetLink Human for strong interactions among DEmiRs and their putative target genes is shown (C). Orange nodes show target genes associated with three or more miRs, blue nodes those shared by less than three miRs, brown nodes indicate miRs. The interactions among the proteins encoded by target genes drawn by STRING v11 are presented (D). Nodes corresponding to clustered proteins are presented in the same color. Linear relations between hsa-miR-144-3p and hsa-miR-20b-5p or hsa-miR-25-3p in the plasma samples of HCM population are shown (E). Linear relation between the hsa-miR-144-3p and hsa-miR-20b-5p expression levels determined by RT-qPCR and the left ventricle ejection fraction (EF) assessed by routine echocardiography or cardiac magnetic resonance in patients either with non obstructive (HCM) or obstructive (HOCM) hypertrophic cardiomyopathy is shown (F) The miR144/451 family was associated with extracellular matrix remodeling, 5 negative regulation of hypertrophy and autophagy, 6 and cardiac diseases by other reports (e.g., the HUNT study 7 ). Our data in HCM showed that hsa-miR-144-3p and hsa-miR-451a were linearly related, while hsa-miR-4732-5p undetermined in the majority of samples ( Figure S5), confirming a published report in HCM patients, 6 but also suggesting an opposite trend in plasma and tissue for hsa-miR-144-3p expression, due to increased cardiac release/decreased synthesis. This hints to a role in cardiac remodeling in HCM, deserving further investigations.
Moreover, ROC analysis demonstrated the good performance of 4 out of 5 tissue DEmiRs ( Figure 3C).
The other miRs tested in tissues showed comparable expression levels in HCM and ctrl ( Figure S6), but four of them were related to HCM clinical phenotypes (Figure S7). Specifically, hsa-miR-4451 was linearly associated with interventricular septum thickness, both hsa-miR-382-5p and hsa-miR-25-3p to glomerular filtration rate, and hsa-miR-382-5p negatively and non-linearly to Troponin T.
The calculation of Spearman correlation between miR pairs showed some positive relationships in HCM tissue (>0.60) with significant linear fits ( Figure S8A-SAC).
In silico analysis of myocardial DEmiRs and of miRs associated with clinical parameters predicted networks of targets genes and interacting proteins (Figures 3D 3E, and S8D) partially superimposable to those drawn for validated plasma DEmiRs, and showed PTEN as a shared target in both HCM plasma and myocardial tissue. Notably, a significant up-regulation of PTEN gene expression was determined in myectomies from HCM vs. ctrl (p = 0.035, Figure 3F) and increased expressed protein amount (Figures 3G and 3H, p = 0.047) was also found in HCM samples. PTEN deletion in mice drove variable in vivo and in vitro effects on cardiomyocyte hypertrophy, 8 and a prohypertrophic signaling pathway involving miR-20b and PTEN was proposed in conditions of pressure-overload cardiac hypertrophy. [8][9][10] To the best of our knowledge PTEN-related mechanisms in human HCM have not been elucidated. Based on our results we hypothesize a mechanistic role for DEmiRs in the modulation of PTEN in HCM, F I G U R E 3 Analysis of myocardial tissue miRs by RT-qPCR and expression of PTEN. The expression levels of DEmiRs determined by RT-qPCR in the myocardial tissue samples of HCM vs. ctrl are plotted. The miRs down-regulated in HCM are in A, those upregulated in B. Receiver-operator characteristic (ROC) curve analysis of tissue DEmiRs is shown (C). AUC > 0.7 with significant p values was considered as threshold for good discriminant performance. The network obtained by miR TargetLink Human for strong interactions among hsa-miR-4451, hsa-miR-382-5p, hsa-miR-25-3p, tissue DEmiRs and their putative target genes is presented (D). Orange nodes show target genes associated with three or more miRs, blue nodes those shared by less than three miRs, brown nodes indicate miRs. The corresponding protein network drawn by STRING v11 is shown (E). The upregulation of PTEN gene determined by RT-qPCR in HCM vs. ctrl tissues is plotted, (F). Representative qualitative evaluation of the protein presence into myocardial tissues from two HCMs and one ctrl by western blot (G, up) is shown. Immunofluorescence images of HCM and ctrl cryosections are displayed (G, down) and PTEN quantification by specific ELISA is plotted (H). Values In A, B, F, and H plots are presented as boxes (min to max), and dots indicate single sample values. Mann-Whitney test is applied and significant differences are shown as *p < 0.05 possibly in relation to metabolic alterations, and suggesting the need for dedicated mechanistic studies.

F I G U R E 4 Schematic representation of the study results
in silico analyses hinted to a role for PTEN in HCM pathogenesis.

A C K N O W L E D G M E N T S
We would like to thank Marco Magnoni, Silvia Del Rio, and Carmem Sartorio for their help in patient's enrolment and samples collection.   O R C I D Chiara Foglieni https://orcid.org/0000-0002-5155-6426