Identification of two miRNAs regulating cardiomyocyte proliferation in an Antarctic icefish

Abstract

The hemoglobinless Antarctic icefish develop large hearts to compensate for reduced oxygen-carrying capacity, which serves as a naturally occurred model to explore the factors regulating cardiogenesis. Through miRNAome and microRNAome comparisons between an icefish (Chionodraco hamatus) and two red-blooded notothenioids, we discovered significant upregulation of factors in the BMP signaling pathways and altered expression of many miRNAs, including downregulation of 14 miRNAs in the icefish heart. Through knocking down of these miRNAs, we identified two of them, miR-458-3p and miR-144-5p, involved in enlarged heart development. The two miRNAs were found to regulate cardiomyocyte proliferation by targeting bone morphogenetic protein-2 (bmp2). We further validated that activation of the miRNA-bmp2 signaling in the fish heart could be triggered by hypoxic exposure. Our study suggested that a few miRNAs play important roles in the hypoxia-induced cardiac remodeling of the icefish which shed new light on the mechanisms regulating cardiomyocyte proliferation in heart.

Keywords: Cell biology; Evolutionary ecology; Molecular biology.

Graphical abstract

Figure 1

Morphology comparison between the white-blooded C. hamatus and red-blooded notothenioids T. bernacchii (A) The heart size comparison between T. bernacchii and C. hamatus. The average heart-to-body mass ratio of ten white-blooded C. hamatus and red-blooded notothenioids T. bernacchii is 0.3854 ± 0.0307 and 0.0955 ± 0.0094, respectively. The detailed biological parameters for the two Antarctic fish species were shown in Table S1. (B and C) Hematoxylin-eosin-stained ventricle paraffin sections of the two Antarctic fish species showed similar morphology. Three different individuals for the T. bernacchii (named as TB1, TB2 and TB3) and C. hamatus (named as CH1, CH2 and CH3) were used and 3 randomly chosen square areas (50 × 50μm²) (marked as 1, 2 and 3) per one individual were used for comparison. (D) Cardiomyocytes measurement comparisons of T. bernacchii and C. hamatus (Error bars indicate ±1 standard error of the mean (SEM)).

Figure 2

A schematic presentation of the regulatory networks involved in vertebrate heart development (A) The factors boxed in red rectangles were shown to be up-regulated in icefish heart compared with the red-blooded species T. bernacchii and G. acuticeps. These genes included bmp2, bmp7, smad1, smad 2, smad 9, mef2a and mef2c (Table S4). Factors boxed by gray rectangles showed similar transcription abundance between the hearts of two types of notothenioid fishes. The two-directional black arrows indicate mutual signaling between two TFs. The one-directional black arrow and dashed arrows indicate transcriptional activation or inhibition between the two factors, respectively. The dotted segments indicate non-definitively determined transcriptional signaling leading to cardiomyocyte’s differentiations. (B) Distribution of significantly enriched GO terms in the DEGs between the hearts of icefish and red-blooded notothenioids. Significant enrichment of 10 GO terms was found (p < 0.05, Number of genes involved in one GO term≥10). The GO terms involved in cell multiplication (cell proliferation, cell division, and cell cycle) are indicated by blue color. The X axis shows the –log10 p value of the GO enrichment. The red dotted segment indicates–log10 p = 0.05. And the Y axis shows the name of enriched GO terms.

Figure 3

Western blot analyses of some cardiac developmental regulators in C. hamatus and T. bernacchii hearts (A and B) Results of western blot analyses of 10 factors involved in cardiac development in the hearts of C. hamatus and T. bernacchii. The range of the used protein molecular weight markers is 10–180 kD. The corresponding molecular marker sizes besides each detected protein are as follows: 75KD for Bmp2, 35KD for Bmp4, 60KD for Bmp7, 60KD for Smad1, 45KD for Ccna2, 45KD for Nkx2.5, 60KD for Mef2a, 75KD for Tbx2, 45KD for Gata4 and 45KD for β-actin. (C) The relative abundance of the examined proteins derived from three biological replicates (Error bars indicate ±1 SEM). Levels with significant difference between C. hamatus and T. bernacchii are denoted with ‘∗’ (p < 0.05, Student’s t test) and ‘∗∗’ (p < 0.01).

Figure 4

Experimental verifications of the regulatory functions of miR-458-3p and miR-144-5p on cardiomyocyte proliferation (A) Verification of the expression of miR-458-3p (upper)/miR-144-5p (down) in the embryo fish heart micro-injected by miR-458-3p antagomir (upper)/miR-144-5p antagomir (down). MicroRNA qRT-PCR techniques showed the significant down-regulation of miR-458-3p (or miR-144-5p) in the zebrafish heart micro-injected by miR-458-3p antagomir (or miR-144-5p antagomir). At least three biological replicates were conducted for each measurement. Treatments with significant difference (p < 0.05, p < 0.01, Student’s t test) were denoted with ‘∗’ and ‘∗∗’, respectively. (B) Tg Danio rerio (myl7: EGFP) were used for the microinjection experiment. Images of zebrafish development at 72 hpf after microinjection of 50 nM synthetic miR-458-3p antagomir (miR-144-5p antagomir) or NC, compared with the WT zebrafish showing the different heart sizes. The white arrow and red arrow indicate atrium and ventricle, respectively. (C) Comparison of heart sizes measured from each treatment. Treatments with significant difference (p < 0.05, p < 0.01, p < 0.001, Student’s t test) were denoted with ‘∗’, ‘∗∗’ and ‘∗∗∗”, respectively. N.S. is the abbreviation for none significant. Error bars indicate ±1 SEM.

Figure 5

Verification of the miR-458-3p and miRNA-144-5p antagomirs on cardiomyocytes proliferations (A) Images of equal numbers of cultured H9C2 cardiomyocytes transfected with miR-458-3p antagomir, miR-144-5p, negative control (NC), and empty vector (EV), respectively. (B) Cell number counting and statistical analysis for the transfected cells. Treatments with significant differences (p < 0.001, Student’s t test) are indicated by ‘∗∗∗’ (Error bars indicate ±1 SEM). (C) The apex cordis excisional zebrafish microinjected by DEPC water, NC miRNA, miR-458-3p mimics and miR-144-5p mimics showed no significant proliferation signals. The apex cordis excisional zebrafish microinjected by miR-458-3p antagomir or miR-144-5p antagomir showed a significant proliferation signal (red dot).

Figure 6

Experimental verifications of target genes of miR-458-3p and miR-144-5p (A) Predicted binding site of miR-458-3p at the 3′UTR of bmp2 (The 3′UTR sequence of bmp2 was shown in Table S8) and binding site of miR-144-5p at 3′UTR of bmp2. (B) Verification of the function of miR-458-3p on 3′UTR of bmp2 using EGFP reporter constructs in the H9C2 cardiomyocyte cell line and zebrafish fertilized 1- to 2-cell embryos, respectively. Compared to the cells transfected with vector (V), the NC microRNA, the cells transfected with the miR-458-3p mimics showed significantly reduced EGFP expression by the pEGFP-bmp2-3′UTR construct. The expression of GFP was measured using Western blotting after 48 h post transfection. (C) Verification of the function of miR-144-5p on 3′UTR of bmp2 using EGFP reporter assay. (D) bmp2 expression comparisons between WT, zebrafish embryos microinjected by NC miRNA, miR-458-3p antagomir, and miR-144-5p antagomir by qRT-PCR and western blotting analysis. Treatments with significant differences (p < 0.05, Student’s t test) are indicated by ‘∗’ (Error bars indicate ±1 SEM). The range of the used protein molecular weight markers is 10–180 kD. The corresponding molecular marker sizes besides each detected protein are as follows: 75KD for Bmp2, 29KD for GFP and 45KD for β-actin.

Figure 7

Hypoxia induced miR-458-3p/miR-144-5p downregulation and bmp2 upregulation in zebrafish heart and Antarctic fish heart (A) Quantitative RT-PCR showing up-regulated expression of bmp2, reduced miR-458-3p and miR-144-5p in the hypoxia (DO = 1.0 ± 0.2 mg/L) acclimated zebrafishes’ heart (see Table S14 for details). Western blot analysis of Bmp2 in the DO = 1.0 ± 0.2 mg/L acclimated zebrafishes’ heart compared with that of the normoxia zebrafishes. (B) Quantitative RT-PCR showing upregulated expressions for the cell cycle-related genes cdc2, cdc20, cdc27, ccnd1, ccnd2, ccna2 and ccnb1 in the heart of hypoxia acclimated zebrafish (see Table S15 for details). At least three biological replicates were conducted for each measurement. (C) Quantitative RT-PCR showing reduced expression of bmp2 and miR-458-3p in the conditioned red-blooded T. bernacchiis’ heart (see Table S16 for details). Western blot analysis of Bmp2 in the hypoxia conditioned red-blooded T. bernacchiis’ heart compared with that of the normoxia T. bernacchii. The relative abundance of the examined (bmp2) and the gene of miR-458-3p derived from at least three biological replicates. Levels with significant difference between the hypoxia conditioned red-blooded T. bernacchiis’ heart and that of the normoxia T. bernacchii are denoted with ‘∗’ (p < 0.05, Student’s t test). Error bars indicate ±1 SEM. The range of the used protein molecular weight markers is 10–180 kD. The corresponding molecular marker sizes besides each detected protein are as follows: 75KD for Bmp2 and 45KD for β-actin.

Figure 8

A proposed regulatory network for the adaptive heart enlargement in icefish