Inhibition of miR-25 improves cardiac contractility in the failing heart

Abstract

Heart failure is characterized by a debilitating decline in cardiac function, and recent clinical trial results indicate that improving the contractility of heart muscle cells by boosting intracellular calcium handling might be an effective therapy. MicroRNAs (miRNAs) are dysregulated in heart failure but whether they control contractility or constitute therapeutic targets remains speculative. Using high-throughput functional screening of the human microRNAome, here we identify miRNAs that suppress intracellular calcium handling in heart muscle by interacting with messenger RNA encoding the sarcoplasmic reticulum calcium uptake pump SERCA2a (also known as ATP2A2). Of 875 miRNAs tested, miR-25 potently delayed calcium uptake kinetics in cardiomyocytes in vitro and was upregulated in heart failure, both in mice and humans. Whereas adeno-associated virus 9 (AAV9)-mediated overexpression of miR-25 in vivo resulted in a significant loss of contractile function, injection of an antisense oligonucleotide (antagomiR) against miR-25 markedly halted established heart failure in a mouse model, improving cardiac function and survival relative to a control antagomiR oligonucleotide. These data reveal that increased expression of endogenous miR-25 contributes to declining cardiac function during heart failure and suggest that it might be targeted therapeutically to restore function.

Extended Data Fig. 1. Effect of miR-25 on IP3R1 and cardiomyocyte calcium transients in vitro

Effect of miR-25, anti-miR-25 and siIP3R1 transfection on IP3R1 protein levels in HL-1 cells (a). Sequence of the putative miR-25 target recognition element in IP3R1 mRNA and the corresponding alteration by site-directed mutagenesis (b). Mutation abolished inhibition of luciferase signal by miR-25 (n=10). Representative Ca²⁺ transient of HL-1 cells transfected as indicated (c). Note that miR-25 slowed the repolarization phase kinetics, whereas anti-miR-25 quickened the kinetics. Co-transfection normalized kinetics. Kinetic imaging cytometry analysis of Ca²⁺ transient kinetics during the decay phase (Ca²⁺ transient duration time from 75% to 25% maximal value, CaTD_75-25) of transfected NRVC (d) (n>550 cells). In panels a,b data is represented as mean ± s.e.m. In panel d, box defines interquartile range; whiskers = ±5^th and 95^th percentile; dots = outliers

Extended Data Fig. 2. Effect of miR-25 on calcium handling proteins

Expression of candidate targets of miR-25 in transfected HL-1 cells (lysates collected 5 days post-transfection). No effect of miR-25 or anti-miR-25 was noted for these proteins. Data represented as mean ± s.e.m.

Extended Data Fig. 3. Calcium transient effects of miR-25 compared to that of siRNAs against SERCA2a and IP3R isoforms

Cardiomyocyte-like HL-1 cells were transfected with miR-25, siRNA to SERCA2a (siSERCA2a) (left panels) or siRNAs to IP3R1 or IP3R2 (siIP3R1, siIP3R2) (right panels), and analyzed 72 hours later by kinetic imaging cytometry. Kinetic parameters are CaTD₅₀ (Ca²⁺ transient duration 50, which is the time from maximal value to 50% maximal value), CaTD_75-25 (Ca²⁺ transient duration 75-25, which is the time from 75% maximal value to 25% maximal value), and V_max upstroke (the maximal velocity of the upstroke phase of the Ca²⁺ transient). Data are represented as whisker plots, with the box denoting the 25^th and 75^th percentiles, the whiskers the 5^th and 95^th percentiles, the middle bar is the median, and outliers are indicated as individual dots. Note that siSERCA2a and miR-25 elicited comparable effects, both markedly delaying the Ca²⁺ uptake phase parameters CaTD₅₀ and CaTD_75-25 without appreciably altering V_max upstroke kinetics. (n>550 cells per group). Box defines interquartile range; whiskers = ±5^th and 95^th percentile; dots = outliers. Also note that siIP3R1 only minimally affected the Ca²⁺ kinetic parameters. siIP3R2 slowed V_max upstroke and broadened the distribution of uptake phase kinetic parameters CaTD₅₀ and CaTD_75-25. miR-25 in contrast slowed the uptake phase parameters but did not appreciably affect V_max upstroke. (n>550 cells per group). Box defines interquartile range; whiskers = ±5^th and 95^th percentile; dots = outliers. Although IP3R1 might be a direct target of miR-25, several lines of evidence suggest that it is unlikely to mediate miR-25’s effect in heart failure. IP3Rs are intracellular ligand-gated Ca²⁺ release channels that in the sarcoplasmic reticulum are associated with excitation-contraction coupling or spontaneous Ca²⁺ release and enhanced Ca²⁺ transients, whereas in the nuclear envelope promote nuclear Ca²⁺ signaling^-, but a specific role for IP3R1 in heart failure has not been identified. Nonetheless, miR-25 control of IP3R1 might play a critical role under conditions that sensitize cardiomyocytes to IP3, such in response to endothelin-1, angiotensin and phenylephrine or in local Ca²⁺ control.

Extended Data Fig. 4. Effect of miR-25 and miR-92a overexpression in vivo

Effect of AAV9 cardiac gene transfer of miR-25 and miR-92a on target protein levels (a). Cardiac gene transfer of miR-25 and miR-92a increased levels of their respective microRNAs but miR-92a was selective against confirmed target integrin subunit α5 (ITGA5). Effects of AAV9 cardiac gene transfer of miR-25 and miR-92a on cardiac function (b-c). Tendency towards decreased cardiac function in both ESPVR (b) and dP/dtmax (c) in AAV9-miR-25 animals relative to control (AAV9-VLP) and AAV9-miR-92a treated animals. n=5 (AAV9-VLP); 4 (AAV9-miR-25); 5 (AAV9-miR-92a) for panels b-c . In all panels data is represented as mean ± s.e.m.

Extended Data Fig. 5. AAV9-mediated gene transfer integration in ventricular myocardium

The number of integrated miR-25 and miR-92a copies in ventricular myocardium 6 weeks after AAV gene transfer, determined by quantitative PCR (see Detailed Methods). n = 5 (AAV9-VLP), n = 4 (AAV9-miR-25) and n = 5 (AAV9-miR-92a). Data is represented as mean ± s.e.m.

Extended Data Fig. 6. Effects of anti-miR-25 on endogenous miR-25 and SERCA2a in WT and Serca2a-null hearts

Anti-miR-25 or control (scrambled sequence) anti-miR was administered intravenously to sham-operated WT mice (Sham) or to unoperated Serca2a-cardiomyocyte null (S2a KO) mice, as for the experiments in Fig. 4 (see Detailed Methods). Sham-operated animals were injected with anti-miRs 1 week after surgery and analyzed at 4 weeks. S2a KO animals were injected with 4-OH tamoxifen i.p. for 4 days to delete Serca2a (see Detailed Methods), injected with anti-miRs 1 week later and analyzed 4 weeks later (5 weeks from initial 4-OH-tamoxifen injection). Note that anti-miR-25 decreased endogenous miR-25 levels in sham-operated WT and S2a KO mice relative to control-treated animals (n = 3) (a). Furthermore, SERCA2a protein levels increased in the anti-miR-25 treated WT animals relative to control (n = 3) (b). SERCA2a is absent in S2a KO hearts (c). In all panels data is represented as mean ± s.e.m.

Extended Data Fig. 7. Selectivity of anti-miR-25 on miR-25 family

Expression levels of miR-25 and family members (miR-32, miR-92a, and miR-92b) in sham-operated mice that had been injected with anti-miR-25 or control (scrambled sequence) anti-miR as in Fig. 4 (see Detailed Methods) (n = 3). Note that control anti-miR in sham-operated animals did not alter expression of any of the miRs tested (blue bars). In contrast, anti-miR-25 significantly reduced levels of miR-25 but not other family members in sham-operated animals (red bars). Data is represented as mean ± s.e.m.

Extended Data Fig. 8. Effects of anti-miR-25 on cardiac function of WT and Serca2a-null hearts

Anti-miR-25 or control (scrambled sequence) anti-miR was administered intravenously to sham-operated WT mice (Sham) or to unoperated Serca2a-cardiomyocyte null (S2a KO) mice, as for the experiments in Fig. 4 (see Detailed Methods). Sham-operated animals were injected with anti-miRs 1 week after surgery and followed by echocardiography. S2a KO animals were generated by injection with 4-OH tamoxifen i.p. for 4 days (see Detailed Methods) to delete Serca2a, then injected with anti-miRs 1 week later and followed by echocardiography. Representative two-dimensional guided M-mode images of the left ventricles from WT and S2a KO mice (a). Echocardiographic indices of left ventricular inner dimension during diastole, LVIDd (b) and systole, LVIDs (c) (n = 3 (sham+control); 8, (sham+anti-miR-25); 7 (S2a KO+control); 11 (S2a-KO+anti-miR-25) 4 weeks after control or anti-miR injection. Echocardiographic measurement of fractional shortening (FS) expressed as a percentage at time points after control or anti-miR injection (n = 4 (sham+control); 3 (sham+anti-miR-25); 3 (S2a KO+control); 3 (S2a KO+anti-miR-25)) (d). S2a KO mice show characteristic dilation and decline in cardiac function following 4-OH tamoxifen-induced excision of Serca2a; ***, P < 0.001 (Student’s t-test for difference between S2a KO and Sham-operated control anti-miR groups at week 4 following injection). Hemodynamic effect of anti-miR-25 and control anti-miR injection represented by pressure-volume plots of treatment cohorts as indicated 4 weeks after injection of control or anti-miR (e). Note that specific anti-miR-25 and control anti-miR acted similarly in sham-operated WT animals (n = 3 (sham+control); 3, (sham+anti-miR-25); 2 (S2a KO+control); 3 (S2a-KO+anti-miR-25). Moreover, anti-miR-25 did not increase cardiac function of S2a KO mice, unlike TAC-operated WT mice (Fig. 4), suggesting that the beneficial effect on cardiac function depends on SERCA2a. In all panels data is represented as mean ± s.e.m.

Extended Data Fig. 9. Kaplan Meier survival curve for anti-miR-25 treatment

Survival probability is plotted over time, showing cumulative protective effect of anti-miR-25 relative to control (scrambled sequence) anti-miR injections following trans-aortic constriction (TAC). The summary of two experiments is shown plotting time from injection. Groups were sham-operated (n = 8), TAC + anti-miR-25 (n = 8), TAC + control (scrambled sequence) anti-miR (n = 22). Note that injection with specific anti-miR-25 increased survival (P = 0.0131, Log-rank test) relative to TAC + control miR.

Extended Data Fig. 10. Effect of anti-miR-25 on accumulation of SUMOylated SERCA2a

Immunoblots of lysates from heart tissue at termination of the in vivo study of Fig. 4 (5.5 months post-TAC, corresponding to 3.5 months post-injection of anti-miR-25 or control (scrambled sequence) anti-miR). Lysates were immunoprecipitated with anti-SUMO1 followed by Western blotting with anti-SERCA2a, as in. Note reduction in SUMOylated SERCA2a upon TAC in the control anti-miR treated hearts, but restoration of expression following specific anti-miR-25 injection. Total levels of SUMO1 and actin are shown.

Figure 1. High content screening identifies miRs that control SERCA2a

a,b. Target sensor construct (a) and screening workflow (b). c. Screening summary. d,e. Primary (d) and confirmatory (e) screen data plotted as % inhibition relative to siRNA against Serca2a (100% inhibition) and scrambled sequence control (0% inhibition) (X-axis) and P-value from Student’s t-test (Y-axis). f. eGFP-SERCA2a inhibition by miR-25 and inactive miR-766 relative to scrambled sequence control n = 10. g. Ca²⁺ transient kinetic analysis of HL-1 cells transfected with miRs that inhibited eGFP-SERCA2a (panel e) and are evolutionarily conserved and upregulated in human heart failure. CaTD_75-25 is decay phase duration from 75% to 25% maximal value (n > 550 cells per group). Box = 25^th to 75^th percentiles; whiskers = 5^th and 95^th percentiles; dots = outliers. *, **, *** = P < 0.05, < 0.01, < 0.001. ns = not significant (one-tailed ANOVA). h. Frequency distribution and log-normal curve fits for CaTD_75-25 values from panel f, normalized to sample size. Both siSERCA2a and miR-25 increased CaTD_75-25 values. i. miR-25 is upregulated in human heart failure samples, by Q-PCR. Error bars = s.e.m. ** indicates P < 0.01 (n = 5), Student’s t-test. All replicates (n) are biological.

Figure 2. Endogenous miR-25 expression in the heart

In situ hybridization revealing endogenous miR-25 expression in failing left ventricular myocardium (red) compared to periostin (a,a’), αSMA (b,b’) and lectin BS-1 (c,c’) staining (each in green). Hoechst 33342 staining marks nuclei (blue). Arrows indicate examples of non-cardiomyocytes. Scale bar equals 20 µm. Data are representative of 2 biological replicates.

Figure 3. miR-25 directly targets SERCA2a and regulates contractile Ca²⁺ kinetics

a. Effects of miR-25, anti-miR-25 and controls (scrambled sequence and SERCA2a siRNA) on SERCA2a protein levels. b. Mutagenesis of the putative mir-25 recognition element in the SERCA2a mRNA 3’UTR abolished inhibition by miR-25. *, ** indicate P < 0.05 and 0.01 (n = 10) (Student’s t-test), compared to control. c. Anti-miR-25 diminished the effect of miR-25 on CaTD_75-25 in transfected HL-1 cells. siRNA to IP3R1 had a minimal effect. Extended Data Fig. 3 shows additional parameters and similar results using NRVCs. Box plots as in Fig. 1g. *,# indicate significant difference (* or # = P < 0.05, ** or ## = P < 0.01, *** or ### = P < 0.001, one-tailed ANOVA) from negative control (*) or miR-25 (#) (n > 550 cells per group). d. Protocol for AAV9 cardiac gene transfer. e. AAV-miR-25 (n = 4) and AAV-miR-92a (n = 5) increased levels of their respective microRNAs relative to AAV-VLP control (n = 3). f–h. miR-25, but not miR-92a diminished SERCA2a protein levels (f) and fractional shortening (FS; percentage) following injection (n = 5 for each cohort; g). LV M-mode images (h). I–k. Pronounced effect of miR-25 on pressure-volume relationship (i) and P_max (j) showing decreased function relative to miR-92a and control AAV9; and the effect on Tau, the time constant for LV relaxation, suggestive of diastolic dysfunction (n = 5, AAV9-VLP; n = 4, AAV9-miR-25; n = 5, AAV9-miR-92a). Data are represented as mean ± s.e.m in all panels except c. All replicates (n) are biological.

Figure 4. Inhibition of miR-25 normalizes TAC-induced cardiac dysfunction

a. Protocol for anti-miR-25 therapy in the mouse TAC heart failure model. b. LV M-mode images showing dilation with control anti-miR in contrast to anti-miR-25 injected mice. c. Effect on echocardiographic indices of left ventricular function: ejection fraction (EF) and fractional shortening (FS) expressed as percentages. The number of animals initiated is n = 4 (sham operated), 10 (TAC + control anti-miR), and 4 (TAC + anti-miR-25); the numbers analyzed per time point are indicated and reflect deaths in the TAC + control anti-miR group. d–f. Hemodynamic effects. Pressure-volume plots of treatment cohorts as indicated (d). Note anti-miR-25 trend towards normalization of hemodynamic indices of end systolic pressure volume relationship (ESPVR, slope of lines in d) (e), and EF (f). n = 4 (sham operated), 4 (TAC + control anti-miR), and 3 (TAC + anti-miR-25). g. Heart weight to body weight ratio. n = 5 (sham), 8 (TAC + control anti-miR), and 4 (TAC + anti-miR-25). h,i. Effect of treatment on endogenous miR-25 levels (h, Q-PCR, n = 4) and SERCA2a protein (I, immunoblotting, n = 3). j–w. Masson’s trichrome (j,k,n,o,r,s), hematoxylin/eosin (l,p,t), and wheat germ agglutinin (WGA) (m,q,u) stained sections of hearts and LV wall. Quantified fibrotic area (v) [n = 3 (sham operated), 4 (TAC + control anti-miR), and 5 (TAC + anti-miR-25]. Average cardiomyocyte cross sectional area (w) (n = 3 for all cohorts). For all panels, mean ± s.e.m. *, # indicate significant difference (* or #, ** or ##, *** or ### = P <0.05, <0.01, <0.001) (Student’s t-test) between sham and control-anti-miR (*) or control anti-miR and anti-miR-25 (#). All replicates (n) are biological.