LncRNA ZFAS1 as a SERCA2a Inhibitor to Cause Intracellular Ca2+ Overload and Contractile Dysfunction in a Mouse Model of Myocardial Infarction.

RATIONALE: Ca2+ homeostasis-a critical determinant of cardiac contractile function-is critically regulated by SERCA2a (sarcoplasmic reticulum Ca2+-ATPase 2a). Our previous study has identified ZFAS1 as a new lncRNA biomarker of acute myocardial infarction (MI).
OBJECTIVE: To evaluate the effects of ZFAS1 on SERCA2a and the associated Ca2+ homeostasis and cardiac contractile function in the setting of MI. METHODS AND
RESULTS: ZFAS1 expression was robustly increased in cytoplasm and sarcoplasmic reticulum in a mouse model of MI and a cellular model of hypoxia. Knockdown of endogenous ZFAS1 by virus-mediated silencing shRNA partially abrogated the ischemia-induced contractile dysfunction. Overexpression of ZFAS1 in otherwise normal mice created similar impairment of cardiac function as that observed in MI mice. Moreover, at the cellular level, ZFAS1 overexpression weakened the contractility of cardiac muscles. At the subcellular level, ZFAS1 deleteriously altered the Ca2+ transient leading to intracellular Ca2+ overload in cardiomyocytes. At the molecular level, ZFAS1 was found to directly bind SERCA2a protein and to limit its activity, as well as to repress its expression. The effects of ZFAS1 were readily reversible on knockdown of this lncRNA. Notably, a sequence domain of ZFAS1 gene that is conserved across species mimicked the effects of the full-length ZFAS1. Mutation of this domain or application of an antisense fragment to this conserved region efficiently canceled out the deleterious actions of ZFAS1. ZFAS1 had no significant effects on other Ca2+-handling regulatory proteins.
CONCLUSIONS: ZFAS1 is an endogenous SERCA2a inhibitor, acting by binding to SERCA2a protein to limit its intracellular level and inhibit its activity, and a contributor to the impairment of cardiac contractile function in MI. Therefore, anti-ZFAS1 might be considered as a new therapeutic strategy for preserving SERCA2a activity and cardiac function under pathological conditions of the heart.

The heart beats rhythmically to drive blood circulating throughout the body, and each single heart beat begins with the initial phase with Ca2+ influx into cells and Ca2+ release from sarcoplasmic reticulum (SR) through RyR2 (ryanodine receptor 2) to trigger contraction of cardiac muscles, followed by the second phase with Ca2+ reuptake or load to SR to render relaxation of cardiac muscles. Thus, intracellular Ca2+ homeostasis is critical to cardiac contractile function and is delicately regulated by a large body of proteins in a highly coordinated manner.[1,2] Among these regulatory proteins, SERCA2a (SR Ca2+-ATPase 2a) is the key protein that governs the normal intracellular Ca2+ handling process and thereby cardiac contractile function by mediating Ca2+ reuptake into SR in cardiac muscles.[3] SERCA2a activity is susceptible to various cellular environmental cues and is under the tight regulation by an array of factors. SERCA2a function can be impaired as a result of either expression deregulation (mostly downregulation) or activity depression leading to deficiency of Ca2+ reuptake back to SR leading to intracellular Ca2+ overload. The consequence of this is manifested by impairment of cardiac contractile function and by cardiomyocyte apoptosis as well.[4,5] Indeed, SERCA2a dysfunction has been associated with a wide spectrum of cardiovascular diseases, such as myocardial infarction (MI), cardiac hypertrophy, arrhythmias, ischemia/reperfusion, and heart failure especially.[6-10] Both basic scientists and clinical practitioners have shown their enthusiasm on the potential of SERCA2a overexpression/replacement as a therapeutic strategy.[11-13] However, a recent randomized, double-blind, placebo-controlled clinical trial named CUPID2 (Calcium Upregulation by Percutaneous Administration of Gene Therapy in Patients With Cardiac Disease) presented negative results on the benefit of SERCA2a overexpression in patients with advanced heart failure with the lack of improvement of reduced ejection fraction (EF).[14] The authors proposed the inadequate dose of SERCA2a as one of the explanations for the lack of benefit of SERCA2a in their study. Clearly, the benefits of SERCA2a replacement have not been well established in the clinical settings, and our understanding of the precise mechanisms for the pathological expression downregulation and functional depression of SERCA2a and the multilayer complex regulatory network for Ca2+ handling is still far from being complete. This constitutes an obstacle for the potential clinical applications of SERCA2a replacement to the treatment of heart disease.

Editorial, see p

Meet the First Author, see p

MI is the worst threat to human lives, and it affects an increasing number of individuals worldwide. This is because MI is not only the main cause of sudden cardiac death but also the primary process leading to heart failure.[15-17] MI is a metabolic catastrophe characterized by contractile dysfunction, arrhythmia, and cell death because of reduced supply of O2 and nutrients to the myocardium. Hypoxia—a condition of insufficient O2 supply to support metabolism—is a frequently encountered problem of the cellular microenvironment and a major factor in the pathology of MI.[18] The contractile dysfunction occurring in MI or hypoxia is primarily caused by impairment of intracellular Ca2+ homeostasis because of the chaos of Ca2+ handling.[19-21] SERCA2a dysfunction as a causal factor for intracellular Ca2+ overload and cardiac contractile dysfunction in the setting of MI has been well documented by many research groups.[22-24] Yet, how SERCA2a expression and function are regulated during MI remained poorly understood.

Human genome sequencing and GENCODE (human genome encyclopedia of DNA elements) project have revealed that only ≈2% of the human genome can be translated into proteins and the rest of it is primarily transcribed into ncRNAs.[25,26] Recently, several lncRNAs have been reported to be involved in heart disease.[27-35] In addition, lncRNAs are also emerging as biomarkers for cardiovascular diseases.[29,35] In our previous study, we identified ZFAS1—an antisense lncRNA to the 5′ end of the protein-coding gene ZNFX1—as an independent predictor of acute MI[29]. Intriguingly, ZFAS1 level was found to be markedly decreased in the bloodstream but elevated in the myocardium. We have, therefore, proposed that in addition to its potential as a biomarker of MI, ZFAS1 might also contribute to the development of MI[29].

This thought prompted us to carry out the present study to exploit the role of ZFAS1 in the pathological process associated with MI and to decipher the underlying molecular and signaling mechanisms. Specifically, we investigated the effects of ZFAS1 on cardiac contractile function and intracellular Ca2+ handling with both gain- and loss-of-function approaches in a mouse model of MI and a cellular model of hypoxia. Our experimental results indicate that ZFAS1 is a detrimental factor to the heart in the setting of MI, and knockdown of this lncRNA is able to mitigate the ischemic contractile dysfunction.

Methods

The data that support the findings of this study are available from the corresponding author on reasonable request.

Human Samples

Human cardiac preparations used in this study were collected from the Tissue bank of the Heilongjiang Academy of Medical Sciences (Harbin, China). The use of human cardiac tissues for the present study was approved from the Ethics Committee of the Harbin Medical University (No. HMUIRB20170034). Our study protocols complied with the guidelines that govern the use of human tissues outlined in the Declaration of Helsinki.

Mouse Model of MI

C57BL/6 mice ranging from 8 to 10 weeks in age and weighing between 22 and 25 g each were used for animal studies, and pregnant C57BL/6 mice were used for neonatal myocyte isolation (Animal Experimental Ethical Inspection Protocol No. HMUIRB20170034). Use of animals was approved by the Ethic Committees of Harbin Medical University and conformed to the Guide for the Care and Use of Laboratory Animals published by the US National Institutes of Health (Publication No. 85-23, revised 1996).

Data Analysis

Data are expressed as mean±SEM and were analyzed with SPSS 13.0 software. Statistical comparisons among multiple groups were performed using ANOVA followed by Dunnett test. Student t test was performed for comparisons between 2 groups. A 2-tailed P<0.05 was taken to indicate a statistically significant difference. Nonlinear least square curve fitting of raw data points was performed with GraphPad Prism.

Detailed descriptions on the materials and methods used in this study are provided in the Online Data Supplement.

Results

Upregulation of ZFAS1 Expression in Ischemic Hearts and Hypoxic Cardiomyocytes

We have demonstrated in our recently published study that ZFAS1 expression is markedly elevated in the myocardium of mice with acute MI (from 1 to 24 hours post-MI).[29] Here, we verified this result in MI mice: compared with the sham group, the expression level of ZFAS1 was increased by 1.8-fold in mouse myocardium 12 hours after MI (Figure 1A, left). Similar increase in ZFAS1 level was also observed in human myocardium from patients with MI (Figure 1A, right).

Figure 1.

Upregulation of , Elevation of cardiac ZFAS1 level in the peri-infarct areas of a mouse model of myocardial infarction (MI; left) and of myocardium from patients with MI (right). Mouse MI was created by ligation of the left descending coronary artery, and the measurements were conducted 12 h after MI. ZFAS1 levels were determined by real-time polymerase chain reaction (same below). *P<0.05, MI mice vs Sham mice (n=6) or patients with MI vs without MI (n=3). B, Increase in ZFAS1 expression in cultured neonatal mouse cardiomyocytes (NMCMs; left) and in AC16 cells (adult human ventricular cardiomyocyte cell line; right) after hypoxia treatment for 12 h, relative to the cells kept under normoxic conditions. **P<0.01, hypoxia vs control; n=≈4–6. Above data are presented as mean±SEM. C, Representative images of In Situ Hybridization (ISH) in NMCMs showing an increase in ZFAS1 expression after hypoxia treatment for 12 h. Note that ZFAS1 was distributed evenly in both cytosol and nucleus. Red arrows pointing to the nucleus stained in light blue and green ones pointing to ZFAS1 stained in brown. Images are presented with a magnification of ×100 for the (top) and ×500 for the (bottom). Similar results were consistently observed in another 3 batches of cells.

Hypoxia is a crucial event in the setting of MI, and it alone can cause many of the detrimental alterations of ischemic myocardium.[36,37] We thought that hypoxia might be sufficient to cause upregulation of ZFAS1 as MI did. Our experiments conducted with cultured neonatal mouse cardiomyocytes (NMCMs) indeed generated the data showing a significant increase in ZFAS1 level after exposure to hypoxic environment for 12 hours (2.6-fold up; P<0.01; Figure 1B, left). Qualitatively the same results were reproduced in adult human ventricular cardiomyocyte cell line AC16 cells cultured under hypoxic condition (ZFAS1 upregulated by 2-fold; P<0.01; Figure 1B, right), relative to that in cells kept under normoxic condition.

Moreover, in situ hybridization in NMCMs showed markedly increased abundance of ZFAS1, as indicated by enhanced staining in brown, after exposure to hypoxic conditions, relative to the normoxic conditions (Figure 1C). Also noted was that ZFAS1 is distributed in both cytoplasm and nucleus.

Impairment of Cardiac Contractile Function by ZFAS1 in MI Mice

A question we asked was whether the upregulation of ZFAS1 in MI is a contributor to ischemic cardiac injuries or is merely a bystander or a consequence of MI. To clarify this issue, we investigated the effects of ZFAS1 on cardiac function in vivo using echocardiography. We first used a loss-of-function strategy with adeno-associated virus serotype 9 vector carrying a ZFAS1-shRNA fragment (shZFAS1-V) and investigated whether ZFAS1 knockdown could alter cardiac function. Adeno-associated virus serotype 9 was chosen because it is one of the most promising gene transfer tools for gene therapy, and it has been demonstrated to have impressively high infection efficiency and safety profile in cardiac tissue.[38] As illustrated in Figure 2A, both EF and FS (fractional shortening) were substantially lower in the MI mice, compared with the sham-operated control counterparts. Moreover, left ventricular (LV) internal dimension at end diastole and LV internal dimension at systole were both enlarged. These changes indicated severe cardiac dysfunction in our MI model. Strikingly, in the MI mice that had received shZFAS1-V for 4 weeks, the abovementioned parameters of cardiac function were all improved with EF and fractional shortening nearly recovered back to normal levels (Figure 2A). In addition, there was a decrease in LV anterior wall thickness and dilatation, as reflected by the diminished LV internal dimension at end diastole and LV internal dimension at systole (Figure 2A). The negative control construct shNC-V (the negative control shRNA engineered into the AAV9 vector) failed to affect the ischemia-induced cardiac dysfunction.

Figure 2.

Impairment of cardiac contractile function by , Amelioration of impairment of contractile function by ZFAS1 overexpression produced by the recombinant adeno-associated virus (serotype 9; AAV9) vector carrying the shRNA (shZFAS1-V) to knock down endogenous ZFAS1 in MI mice. The viral constructs were intravenously injected into mice. Note that shZFAS1-V abrogated the ischemia-induced decreases in ejection fraction (EF) and fractional shortening (FS), and enlargement of left ventricular internal dimension at end diastole (LVIDd) and left ventricular internal dimension at systole (LVIDs). shNC-V: the negative control shRNA engineered into the AAV9 vector. **P<0.01 vs Sham, #P<0.05 vs MI, §P<0.05 vs shZFAS1-V; n=≈12–18. B, Impairment of contractile function induced by ZFAS1 overexpression generated by AAV9 vector carrying the full-length ZFAS1 gene (AAV9 vector carrying the ZFAS1 gene [ZFAS1-V]) in healthy mice. Note that ZFAS1-V significantly decreased EF and FS, and increased LVIDd and LVIDs, similar to those seen in MI mice, and these effects were essentially reversed by coinjection of shZFAS1-V. NC-V: the empty AAV9 vector as a negative control for ZFAS1-V. **P<0.01 vs control or NC-V, ##P<0.01 vs ZFAS1-V; n=≈7–10. C, Decreased maximum rate of rise of left ventricular pressure during contraction (+dp/dtmax) and the maximum rate of drop of left ventricular pressure during relaxation (−dp/dtmax) in MI mice. **P<0.01 vs Sham. ZFAS1 overexpressed by ZFAS1-V delivery showing reduced ±dp/dtmax. *P<0.05 vs control, #P<0.05 vs ZFAS1-V; n=3. D, Raw traces (left) showing the changes of sarcomere shortening (SS) as an index of contractility of cardiac muscles isolated from MI hearts, and mean values of SS in the presence of shZFAS1-V (middle) or ZFAS1-V (right). **P<0.01 vs Sham or control or NC-V, #P<0.05 vs MI, ##P<0.01 vs ZFAS1-V, §§P<0.01 vs shZFAS1-V; n=≈20–43. E, Representative cardiac sections showing the successful delivery of shZFAS1-V and ZFAS1-V into mouse myocardium in vivo the significant presence of fluorescence elicited by GFP (green fluorescent protein) attached to the viral vectors. F, Verification of knockdown of endogenous ZFAS1 by shZFAS1-V in MI myocardium determined by quantitative real-time-polymerase chain reaction. *P<0.05 vs Sham, #P<0.05 vs MI, §P<0.05 vs shZFAS1-V; n=≈12–20. G, Verification of overexpression of ZFAS1 elicited by ZFAS1-V in normal mice 4 and 7 wk after infection. *P<0.05 vs control, #P<0.05 vs ZFAS1-V; n=≈4–10. Data are all expressed as mean±SEM.

The data presented above suggest that upregulation of ZFAS1 in MI produces deleterious effects on the heart. If this is true, then overexpression of ZFAS1 in otherwise normal mice should be able to reproduce the phenotypes of MI-induced cardiac contractile dysfunction. To examine this notion, we went on to use the gain-of-function approach for our subsequent experiments using the adeno-associated virus serotype 9 vector carrying the ZFAS1 gene (ZFAS1-V) for its overexpression under in vivo conditions. As depicted in Figure 2B, ZFAS1-V, but not the negative control NC-V (the empty AAV9 vector as a negative control for ZFAS1-V), significantly reduced EF and fractional shortening and enlarged LV internal dimension at end diastole and LV internal dimension at systole, the effects resembling those induced by MI. These deleterious effects were nearly abolished by shZFAS1-V to knock down ZFAS1, indicating the specificity of actions by ZFAS1.

The maximum rate of rise of LV pressure during contraction (+dp/dtmax) and the maximum rate of drop of LV pressure during relaxation (−dp/dtmax) have been considered the better manifestations of cardiac contraction and relaxation functions.[39,40] Our results showed marked decreases in ±dp/dtmax in MI mice compared with the sham control animals (Figure 2C). Similarly, the healthy mice pretreated with ZFAS1-V for overexpression also exhibited substantially reduced ±dp/dtmax values relative to control mice. Notably, the decreases in ±dp/dtmax induced by ZFAS1 overexpression were nearly completely reversed by coapplication of shZFAS1-V (Figure 2C).

The effects of ZFAS1 on cardiac contractile function revealed by the results described above indicate that ZFAS1 produces a negative impact on cardiac mechanical function. To test this notion, we compared the contractility of cardiac muscles isolated from MI mice pretreated with shZFAS1-V or shNC-V and sham-operated control mice. It is well established that sarcomere shortening (SS) is an index of contraction force of cardiac muscles: the longer the SS, the greater the contraction force.[41] This is because an increase in SS can result in shortening of interfilament lattice spacing bringing cross-bridges in closer proximity to actin monomers thereby increasing the number of force generating cross-bridges (Figure 2D, left). We, therefore, used SS to reflect contractility of cardiac muscles. As shown in Figure 2D (middle), SS was pronouncedly shortened in MI mice compared with the sham control animals, indicating the mechanical deficiency under ischemic insults. Strikingly, in the cardiac cells isolated from the MI mice infected with shZFAS1-V, but not with shNC-V, SS was restored back toward the normal level as that observed in the sham group. Similar to those occurring in MI, in the cardiac fibers isolated from the healthy mice pretreated with ZFAS1-V, SS was substantially reduced (Figure 2D, right), whereas SS was unaltered by NC-V. Moreover, the decrease of SS induced by ZFAS1 overexpression was considerably mitigated by coapplication of shZFAS1-V.

Successful delivery of shZFAS1-V and ZFAS1-V into the mouse myocardium in vivo was verified by the appreciable presence in myocardial sections of fluorescence signal elicited by GFP (green fluorescent protein) attached to the shZFAS1-V and ZFAS1-V vectors and in mouse myocardium determined by quantitative real-time polymerase chain reaction (Figure 2E; Online Figure IA). Furthermore, the efficiency of shZFAS1-V to knockdown endogenous ZFAS1 in MI mice was also confirmed by quantitative real-time polymerase chain reaction showing significant diminishment of the ZFAS1 transcript level in the myocardium (Figure 2F). Overexpression of ZFAS1 on administration of ZFAS1-V for 4 and 7 weeks to healthy mice was also affirmed (Figure 2G; Online Figure IB). As expected, the negative controls NC-V did not affect the ZFAS1 level, whereas coapplication of shZFAS1-V and ZFAS1-V abolished the overexpression of ZFAS1.

Impairment of Intracellular Ca2+ Homeostasis by ZFAS1 in Cardiomyocytes

Ca2+ homeostasis is the key determinant of cardiac contractility, and abnormal intracellular Ca2+ handling might account at least partially for the observed impairment of contractile function caused by ZFAS1 in the setting of MI. We, therefore, went on to study the effects of ZFAS1 on dynamic Ca2+ transient and resting intracellular Ca2+ concentration ([Ca2+]i), reflecting the intracellular Ca2+ handling process and the consequence of alteration of such a process, respectively.[42,43] To this end, the adult ventricular myocytes isolated from MI hearts were loaded with fluo-3 and electrically stimulated at 1 Hz to initiate intracellular Ca2+ transients. The amplitude of peak systolic Ca2+ transient was calculated according to the equation (F−F0)/F0, where F represents the maximum value of a Ca2+ transient and F0 the diastolic/resting Ca2+ level immediately before the onset of a Ca2+ transient, and the time courses of rising (an indication of the rate of SR Ca2+ release) and decaying (an indication of the rate of SR Ca2+ reuptake) phases were obtained by exponential curve fitting to the raw traces of Ca2+ transient. As shown in Figure 3A (left), the amplitude of Ca2+ transient was decreased by 34.5% in MI relative to that in sham controls, which was partially but significantly restored toward the normal values in the MI mice pretreated with shZFAS1-V. More strikingly, the MI-induced prolongation of time course for the decay phase of Ca2+ transient was remarkably accelerated by shZFAS1-V, as reflected by the reduced decay time constant (τd; Figure 3A, right). The rate of rising phase of Ca2+ transient (τr) exhibited a trend of decelerating in MI, and such a slowing was also effectively reversed by knockdown of endogenous ZFAS1 (Figure 3A, middle right). In all cases, the negative control construct shNC-V did not affect the deleterious alterations in MI mice.

Figure 3.

Impairment of intracellular Ca, Restoration of the decreased amplitude of Ca2+ transient by adeno-associated virus serotype 9 vector carrying a ZFAS1-shRNA fragment (shZFAS1-V; left) and acceleration of the slowed time courses of the rising and decaying phases of Ca2+ transient by shZFAS1-V (right) in adult cardiac cells isolated from myocardial infarction (MI) hearts. **P<0.01 vs Sham, #P<0.05 vs MI, §P<0.05 vs shZFAS1-V; n=≈17–25. Middle left, Typical examples of Ca2+ transient traces recorded in Fluo-3-loaded cardiomyocytes isolated from MI mice with or without shZFAS1-V treatment. Right, Averaged data of time constants for sarcoplasmic reticulum (SR) Ca2+ release from (τr for the rising phase) and Ca2+ reuptake back to SR (τd for the decaying phase). The time constants were acquired by single exponential curve fitting to the data points of the rising phase and decaying phase, respectively. *P<0.05 vs Sham, #P<0.05 vs MI, §P<0.05 vs shZFAS1-V; n=≈15–17. B, Decline of the amplitude of Ca2+ transient induced by adeno-associated virus 9 vector carrying the ZFAS1 gene (ZFAS1-V; left) and increases in the time constants for the rising and decaying phases of Ca2+ transient by ZFAS1-V (right) in adult cardiac cells isolated from healthy mice. Middle left, Typical examples of Ca2+ transient traces recorded in Fluo-3-loaded cardiomyocytes isolated from healthy mice with or without ZFAS1-V or NC-V treatment. Right, Mean values of time constants for SR Ca2+ release from (τr) and Ca2+ reuptake (τd). *P<0.05 vs control or NC-V, #P<0.05 vs ZFAS1-V; n=≈27–32. C, Left, Mitigation of increased intracellular Ca2+ concentration ([Ca2+]i) by shZFAS1-V in MI mice. *P<0.05 vs Sham, #P<0.05 vs MI, §P<0.05 vs shZFAS1-V; n=≈15–25. Right, Elevation of [Ca2+]i induced by forced expression of ZFAS1 generated by ZFAS1-V infection. **P<0.01 vs control and NC-V, #P<0.05 vs ZFAS1-V; n=≈10–20. D, Mitigation of increased intracellular Ca2+ concentration ([Ca2+]i) by ZFAS1 siRNA (siZFAS1; n=≈13–19; left) and restoration of the decreased rate of Ca2+ reuptake into SR by siZFAS1 (n=≈9–13; right) in neonatal mouse cardiomyocytes (NMCMs) exposed to hypoxic environment. siRNAs were transfected into cardiomyocytes using X-tremeGENE siRNA transfection reagent. *P<0.05, **P<0.01 vs control, #P<0.05 vs hypoxia, §P<0.05 vs siZFAS1. E, Left, Elevation of [Ca2+]i induced by forced expression of ZFAS1 generated by pCDNA-ZFAS1 vector (ZFAS1-P) in nonhypoxia cardiomyocytes, and abrogation of the effects by cotransfection with siZFAS1. **P<0.01 vs control, ##P<0.01 vs ZFAS1-P, §§P<0.01 vs siZFAS1; n=≈11–16. Right, Slowing of Ca2+ reuptake into SR by forced expression of ZFAS1 generated by ZFAS1-P in nonhypoxia cardiomyocytes, and abrogation of the effects by cotransfection with siZFAS1. *P<0.05 vs control, #P<0.05 vs ZFAS1-P, §P<0.05 vs siZFAS1; n=≈8–12. F, Verification of the efficacy of siZFAS1 to knockdown endogenous ZFAS1 transcripts in cultured NMCMs. **P<0.01 vs control, ##P<0.01 vs siZFAS1; n=4. G, Verification of overexpression of ZFAS1 produced by ZFAS1-P. **P<0.01 vs control, #P<0.05 vs ZFAS1-P; n=4. siNC indicates scrambled negative control siRNA.

As anticipated, overexpression of ZFAS1 by ZFAS1-V in healthy mice produced the phenotypic changes of Ca2+ transient, which were qualitatively the same as those seen in MI mice (Figure 3B). Specifically, ZFAS1-V slowed the kinetics of decaying phases, as the manifested by the remarkable increase of the time constants for Ca2+ reuptake (τd; Figure 3B, right), and such an effect was completely abrogated by shZFAS1-V. By comparison, the kinetics of rising phases (τr) exhibited a trend to slow, but the alterations did not reach statistical significance (Figure 3B, middle right). Along with the changes of the time constants was a decrease in the amplitude of Ca2+ transient with ZFAS1 overexpression (Figure 3B, left).

The opposite changes of Ca2+ transient seen between ischemia or ZFAS1-V and shZFAS1-V predicted a loss of Ca2+ homeostasis leading to intracellular Ca2+ overload in MI and a relief of such an overload by shZFAS1-V. The results acquired from the measurements of [Ca2+]i in adult cardiomyocytes isolated from MI mice and in cultured neonatal cardiomyocytes were indeed in line with this note. As illustrated in Figure 3C (left), the resting [Ca2+]i in isolated cardiomyocytes from MI mice was increased, and such an increase was abolished by shZFAS1-V to knockdown endogenous ZFAS1. On the contrary, forced expression of ZFAS1 with ZFAS1-V infection directly caused Ca2+ overload in healthy mice (Figure 3C, right). These effects of ZFAS1-V were effectively reversed by coinjection of shZFAS1-V, and the negative control NC-V was unable to exert any appreciable impact on [Ca2+]i.

As expected, the resting [Ca2+]i in NMCMs exposed to hypoxic environment underwent similar changes as in MI: it was increased in hypoxia, and such an increase was abolished by siZFAS1 to silence ZFAS1 expression (Figure 3D, left). Consistent with the increase in [Ca2+]i, the rate of Ca2+ reuptake into SR was tremendously slowed (τd increased) in hypoxic cells compared with the control group, and this anomaly was essentially rescued by siZFAS1 (Figure 3D, right). In sharp contrast to ZFAS1 silencing, forced expression of ZFAS1 with transfection of the pCDNA-ZFAS1 vector directly caused Ca2+ overload in nonhypoxic cardiomyocytes (Figure 3E, left). These effects of ZFAS1 were effectively reversed by cotransfection of siZFAS1, and the negative control construct of siZFAS1 was unable to exert any appreciable impact on [Ca2+]i. Expectedly, overexpression of ZFAS1 was able to induce slowing of τd—an effect abrogated by siZFAS1—whereas negative control construct of siZFAS1 did not elicit any significant changes (Figure 3E, right).

The effects of shZFAS1-V and ZFAS1-V on Ca2+ transient were also assessed by optical mapping techniques. As depicted in Online Figure IIA, τd was significantly delayed in the heart of MI mice, and this deceleration was countered by shZFAS1-V. Moreover, overexpression of ZFAS1 by ZFAS1-V was sufficient to reproduce the slowing of Ca2+ reuptake in normal mice as that caused by MI (Online Figure IIB).

The efficacy of siZFAS1 to knock down endogenous ZFAS1 in cultured cardiomyocytes was verified by ≈50% reduction of the expression level of ZFAS1 in siZFAS1-treated cells relative to negative control construct of siZFAS1-treated ones (Figure 3F). Conversely, transfection of pCDNA-ZFAS1 vector resulted in a robust elevation of ZFAS1 level, which was not seen with NC-P (the empty pCDNA vector as a negative control for ZFAS1-P; Figure 3G).

Downregulation and Dysfunction of SERCA2a Induced by ZFAS1

It has been reported that SERCA2a—the primary cardiac isoform and the key protein involved in sequestration of Ca2+ into SR during diastole—plays a major role in the regulation of Ca2+ homeostasis in cardiomyocytes.[3] Moreover, SERCA2a has been reported to be abnormally downregulated in a variety of cardiac conditions, including MI.[22,44] The delayed time course of the decaying phase of Ca2+ transient by ZFAS1 in our study suggests an impairment of Ca2+ reuptake into SR. A question raised in our mind was whether ZFAS1 is involved in the regulation of SERCA2a expression and function or whether alteration of SERCA2a mediates the effects of ZFAS1. As shown in Figure 4A, the cardiac expression of SERCA2a at the protein level was prominently decreased in MI mice relative to the sham animals. Similarly, SERCA2a level was also significantly downregulated on overexpression of ZFAS1 in normal mice (Figure 4B). In both of these situations, shZFAS1-V rescued the loss of SERCA2a. The expression of SERCA2a mRNA demonstrated qualitatively the same changes as its protein levels (Figure 4A and 4B).

Figure 4.

Downregulation of SERCA2a (sarcoplasmic reticulum Ca2+-ATPase 2a) expression induced by , Downregulation of SERCA2a expression at both protein (top) and mRNA (bottom) levels in myocardial infarction (MI) hearts relative to the sham animals and recovery of SERCA2a expression by adeno-associated virus serotype 9 vector carrying a ZFAS1-shRNA fragment (shZFAS1-V). *P<0.05, **P<0.01 vs Sham, #P<0.05 vs MI, §P<0.05 vs shZFAS1-V; n=≈6–8. B, Downregulation of SERCA2a expression at both protein (top) and mRNA (bottom) levels in healthy mice pretreated with adeno-associated virus 9 vector carrying the ZFAS1 gene (ZFAS1-V) and recovery of SERCA2a expression by shZFAS1-V. *P<0.05 vs control, #P<0.05 vs ZFAS1-V; n=≈4–6. C, Downregulation of SERCA2a expression at both protein (top) and mRNA (bottom) levels in neonatal mouse cardiomyocytes (NMCMs) cultured under hypoxic conditions relative to the cells in normoxic environment. Note that silence of ZFAS1 by siZFAS1 normalized the SERCA2a expression. *P<0.05, **P<0.01 vs control, #P<0.05 vs hypoxia, §P<0.05 vs siZFAS1; n=5. D, Downregulation of SERCA2a expression at both protein (top) and mRNA (bottom) levels by ZFAS1-P in NMCMs cultured under normoxic conditions and reversal of SERCA2a downregulation by siZFAS1. *P<0.05 vs control, #P<0.05 vs ZFAS1-P, §P<0.05 vs siZFAS1; n=≈4–6. Data are expressed as mean±SEM.

Similar patterns of expression alterations of SERCA2a mRNA and protein were consistently observed in primary NMCMs cultured under hypoxic insult and with treatments with varying constructs (Figure 4C and 4D).

With the above results indicating the significant role of ZFAS1 in regulating SERCA2a expression, we came up with a further question: Does ZFAS1 have any direct effects on SERCA2a activities in addition to expression regulation? To get insight into this issue, the first step we took was to perform theoretical analysis for RNA:protein binding using the RNA–Protein Interaction Prediction database. Our analysis revealed a high probability of ZFAS1:SERCA2a interaction (Online Figure III). This initial analysis encouraged us to verify the functional relationship between the 2 molecules. We, therefore, switched to an experimental approach using RNA-binding protein immunoprecipitation to see whether ZFAS1 could physically bind SERCA2a. The results depicted in Figure 5A clearly indicate the presence of such an interaction: immunoprecipitation of SERCA2a carried an appreciable amount of ZFAS1; conversely, RNA pulldown of ZFAS1 also dragged down an appreciable quantity of SERCA2a (Figure 5B, right), suggesting that ZFAS1 has a strong affinity to SERCA2a.

Figure 5.

Interaction between lncRNA , RNA-binding protein immunoprecipitation (RIP) analysis for ZFAS1:SERCA2a interaction. Note that immunoprecipitation (IP) of SERCA2a retrieved a robust amount of ZFAS1. **P<0.01 vs anti-IgG; n=4. B, RNA pulldown of ZFAS1 dragged down an appreciable quantity of SERCA2a. The band for the binding between ZFAS1 and SERCA2a protein disappeared when treated with an antisense fragment to ZFAS1 (AsZFAS1). Additionally, an unrelated lncRNA PLSCR4 (phospholipid scramblase 4) as a negative control, could not drag down SERCA2a, indicating the specific of the ZFAS1:SERCA2a interaction. **P<0.01, ##P<0.01, §§P<0.01 vs ZFAS1; n=3. C, Verification of the purity of isolated sarcoplasmic reticulum (SR) by the enhanced activity of SR-specific NADPH (nicotinamide adenine dinucleotide phosphate) cytochrome C reductase determined by colorimetry assay. **P<0.01 vs total protein samples; n=4. D, Upregulation of ZFAS1 expression in SR isolated from myocardial infarction (MI) hearts relative to Sham hearts, determined by quantitative real-time-polymerase chain reaction. *P<0.05 vs Sham; n=3. E, Downregulation of SERCA2a protein in SR of MI myocardium relative to sham control, determined by Western blot analysis. *P<0.05 vs Sham; n=3.

If ZFAS1 indeed binds to SERCA2a, then it should be present in the SR and colocalize with the latter. The purity of SR was verified by the exceptionally enhanced activities of SR-specific NADPH (nicotinamide adenine dinucleotide phosphate) cytochrome C reductase (Figure 5C). The data depicted in Figure 5D show that ZFAS1 level was significantly increased in the RNA samples isolated from purified SR. Meanwhile, our data exhibited that SERCA2a protein level in SR was significantly decreased (Figure 5E). Moreover, following the same approach described by other groups,[45,46] our immunofluorescent staining of SERCA2a protein in conjunction with GFP-labeling of lncRNA ZFAS1 demonstrated that ZFAS1 and SERCA2a proteins had essentially the same subcellular distribution pattern primarily around the nucleus and the 2 molecules colocalized (Online Figure IV).

As already mentioned, in addition to SERCA2a, other proteins, such as RyR2, PLN (phospholamban), and Cav1.2 (the pore-forming α-subunit of L-type Ca2+ channels), are also crucially involved in the regulation of intracellular Ca2+ homeostasis.[47-49] We, therefore, evaluated the effects of ZFAS1 on the expression of these proteins. As shown in Online Figure V, no significant alterations of PLN, RyR2, and Cav1.2 protein levels are observed with ZFAS1 overexpression. Moreover, the lack of effect of ZFAS1 on L-type Ca2+ channel current density further excluded the contribution of this channel to the intracellular Ca2+ overload induced by ZFAS1 (Online Figure VI).

Identification of the Functional Domain of ZFAS1 Key to the Modulation of SERCA2a

By sequence alignment, we identified a region of high-degree conservation across varying species, including man and mouse (Online Figure VII). We contemplated that this sequence domain might be the functional region for ZFAS1 (Functional Domain of ZFAS1 [ZFAS1-FD]) to interact with SERCA2a. Our analysis using the computational docking software Hex 8.0[50,51] indeed provided a piece of theoretical evidence for our conjecture by showing the ability of ZFAS1-FD to bind SERCA. The docking results obtained by Pymol software revealed that the nucleotides U19, G20, G23, G25, and G26 encompassed by ZFAS1-FD are likely the core motif for specific binding to SERCA amino acids Ala241, Glu243, Glu689, Ser693, Arg667, Arg671, and Arg672 (Online Figure VIII). Further intriguing was that the SERCA domain that binds ZFAS1 identified in this study is known to be the phosphorylation domain of SERCA, which contains several potential phosphorylation sites (Ser693, Arg667, and Arg671).[52,53] This information prompted us to further examine our assumption by conducting the following experiments with gain-of-function and loss-of-function approaches.

First, an oligonucleotide fragment corresponding to the conserved region of ZFAS1 gene (ZFAS1-FD; Figure 6A, top) was synthesized. ZFAS1-FD was transfected into NMCMs by using X-tremeGENE Transfection Reagent (No. 10810500; Roche) according to the manufacturer’s instructions. The ability of ZFAS1-FD to bind SERCA2a and its effects on SERCA2a and Ca2+ handling were examined. As illustrated in Figure 6B through 6D, application of ZFAS1-FD reproduced the effects of full-length ZFAS1 on SERCA2a expression and intracellular Ca2+ status in NMCMs: downregulation of SERCA2a at both protein and mRNA levels (Figure 6B), deceleration of decaying kinetics of Ca2+ transient (Figure 6C), and increase in resting Ca2+ concentration (Figure 6D). The negative control oligonucleotides fragment (NC) did not produce any of these effects.

Figure 6.

Dysfunction of SERCA2a (sarcoplasmic reticulum Ca, Top, An oligonucleotide fragment corresponding to the conserved region of ZFAS1 gene (functional domain of ZFAS1 [ZFAS1-FD]). Bottom, An oligonucleotide fragment antisense to ZFAS1-FD (AsZFAS1-FD). B, Downregulation of SERCA2a expression at both protein (left) and mRNA (right) levels in neonatal mouse cardiomyocytes (NMCMs) transfected with ZFAS1-FD. *P<0.05 vs control, #P<0.05 vs ZFAS1-FD; n=≈5–7. C, Deceleration of decaying kinetics of Ca2+ transient in NMCMs transfected with ZFAS1-FD. *P<0.05 vs control, #P<0.05 vs ZFAS1-FD; n=≈17–22. D, Increase in resting Ca2+ concentration ([Ca2+]i) in NMCMs transfected with ZFAS1-FD. **P<0.01 vs control, ##P<0.01 vs ZFAS1-FD; n=≈14–30. E, Effects of Mut-ZFAS1-FD and ZFAS1-FD on the expression of SERCA2a at both protein (left) and mRNA (right) levels in NMCMs. *P<0.05 vs control or NC; n=≈6–8. F, Resting Ca2+ concentration ([Ca2+]i) in NMCMs transfected with Mut-ZFAS1-FD. **P<0.01 vs control or NC; n=≈25–36. G, The decaying kinetics of Ca2+ transient in NMCMs transfected with Mut-ZFAS1-FD. *P<0.05 vs control or NC; n=≈9–12. H, Upregulation of SERCA2a expression at both protein (left) and mRNA (right) levels by AsZFAS1-FD in NMCMs pretreated with ZFAS1-P for ZFAS1 overexpression. *P<0.05 vs ZFAS1-P, #P<0.05 vs AsZFAS1-FD; n=≈5–9. I, Reversal of ZFAS1-induced reduction of Ca2+ transient amplitude by AsZFAS1-FD. *P<0.05 vs ZFAS1-P, #P<0.05 vs AsZFAS1-FD; n=≈10–18. J, Reversal of the time delay of Ca2+ release and reuptake by AsZFAS1-FD in NMCMs pretreated with ZFAS1-P for ZFAS1 overexpression. *P<0.05 vs AsZFAS1-FD (left) or ZFAS1-P (right), #P<0.05 vs AsZFAS1-FD; n=≈12–18. K, Verification of the efficacy of AsZFAS1-FD in reducing ZFAS1 level in NMCMs pretreated with ZFAS1-P for ZFAS1 overexpression. **P<0.01 vs control, #P<0.05 vs ZFAS1-P, §P<0.05 vs AsZFAS1-FD; n=6.

Next, we mutated the core motif of the ZFAS1-FD by nucleotide substitution to disrupt the binding site for SERCA (Mut-ZFAS1-FD; Online Table I) and examined the effects of this construct on SERCA2a function. The computational docking results confirmed the loss of the original binding sites for SERCA in Mut-ZFAS1-FD (Online Figure IX). Compared with wild-type ZFAS1-FD, Mut-ZFAS1-FD failed to affect the expression of SERCA2a at both mRNA and protein levels in NMCMs (Figure 6E), and as anticipated, Mut-ZFAS1-FD also lost the ability to alter [Ca2+]i (Figure 6F) and Ca2+ transient (Figure 6G).

We then subsequently constructed an oligonucleotide fragment antisense to ZFAS1-FD (AsZFAS1-FD; Figure 6A, bottom) with methylation modification to enhance stability. Introduction of AsZFAS1-FD into the NMCMs effectively abolished the repressive effect of ZFAS1 on SERCA2a expression at both mRNA and protein levels, presumably by annealing to ZFAS1-FD (Figure 6H). Consistently, AsZFAS1-FD also robustly reversed the ZFAS1-induced reduction of the amplitude of Ca2+ transient (Figure 6I) and delay of the time courses of Ca2+ release and reuptake (Figure 6J). By comparison, the negative control antisense fragment failed to affect the effects of ZFAS1.

Finally, to verify that the effects of AsZFAS1-FD observed in our experiments were indeed attributable to the antisense action, we looked at the changes of ZFAS1 level in cells treated with the antisense construct. As shown in Figure 6K, the ZFAS1 level was markedly decreased on transfection with AsZFAS1, but not with negative control antisense fragment, in NMCMs pretreated with pCDNA-ZFAS1 vector.

NFATc2 as a Transactivator of ZFAS1 Expression

Although the above data have indicated the role of ZFAS1 upregulation in causing intracellular Ca2+ overload and the underlying downstream mechanisms in the setting of MI, it remained unknown how ZFAS1 was upregulated in MI. To shed light on this issue, we first performed computational analysis and identified several binding sites for NFATc2 (nuclear factor of activated T cells C2) that has been reported to be abundantly expressed in the cardiac tissue and participated in cardiac hypertrophy and heart failure.[54,55] We, therefore, investigated the possible role of NFATc2 in the regulation of ZFAS1 expression. Our results first demonstrated that the protein level of NFATc2 was robustly upregulated in MI and in NMCMs exposed to hypoxic conditions (Online Figure XA). Application of NFATc2 inhibitor FK506 (10 nmol/L) to NMCMs caused a significant reduction of ZFAS1 expression relative to nontreated cells (Online Figure XB). Moreover, NFATc2 siRNA (siNFATc2) to silence endogenous NFATc2 also downregulated the expression of ZFAS1. (Online Figure XC and XD). In addition to NFATc2, NFATc3 (nuclear factor of activated T cells C3)—another NFAT isoform known to be involved in the regulation of Ca2+ homeostasis in cardiomyocytes—was also investigated for its possible role in regulation ZFAS1 expression. Our data exhibited that although NFATc3 expression was also increased in MI and hypoxia, its siRNA did not alter ZFAS1 expression, consistent with the lack of NFATc3 binding sites in the promoter region of ZFAS1 gene (Online Figure XE through XG).

Discussion

The aims of this study were to elucidate the pathophysiological role of lncRNA ZFAS1 and to delineate the underlying mechanisms in the setting of MI. Our data demonstrated that ZFAS1 was robustly increased in its expression levels in both SR and cytoplasm of MI hearts, and such an upregulation produced significant negative impacts on the heart. This is supported by the finding that knockdown of endogenous ZFAS1 partially abrogated the ischemia-induced contractile dysfunction. Artificial overexpression of ZFAS1 in otherwise normal mice created similar impairments of cardiac function as that observed in MI mice. Moreover, at the cellular level, ZFAS1 overexpression weakened the contractility of cardiac muscles. Furthermore, at the subcellular level, ZFAS1 deleteriously altered the Ca2+ transient leading to intracellular Ca2+ overload in cardiomyocytes. At the molecular level, ZFAS1 was found to directly bind to SERCA2a protein on one hand and to repress its expression on the other hand, the dual actions limiting the activities of this critical Ca2+ handling protein. Probably, the most prominent finding is the identification of the ZFAS1-FD responsible at least partly for the effects of this lncRNA on SERCA2a and its associated intracellular Ca2+ handling process, and the demonstration of AsZFAS1-FD was able to mitigate the effects of full-length ZFAS1. Finally, evidence was obtained for NFATc2 to account at least partially for the upregulation of ZFAS1 in MI. These findings allowed us to propose the following paradigm for the regulation of cardiac contractile function by ZFAS1 in the setting of MI: MI→NFATc2↑→ZFAS1↑→SERCA2a↓→[Ca2+]i↑/Ca2+ overload→contractility↓→contractile function↓, as schematically illustrated in Online Figure XI. Notably, the effects of ZFAS1 were found fully reversible on knockdown of this lncRNA. Based on these findings, we concluded that ZFAS1 is a detrimental lncRNA contributing to cardiac contractile dysfunction through targeting SERCA2a and the associated Ca2+ handling in MI.

Comparison With Published Studies on ZFAS1

ZFAS1—a new member of lncRNAs—was initially identified in patients with breast tumors and subsequently characterized in other cancers.[56-59] Our recently published study suggested circulating level of ZFAS1 as a new biomarker for MI because it was found to be considerably decreased of patients with MI.[29] Intriguingly, expression of ZFAS1 was found to be increased in the myocardium of MI mice—a finding that triggered the present study. During the course of this study, a research article on ZFAS1 in the heart was published.[60] In agreement with our finding with MI mice, the authors showed that ZFAS1 is upregulated in a rat model of acute MI, and such an increase induces cardiomyocyte apoptosis. They further demonstrated that ZFAS1 promotes apoptosis by acting as a ceRNA to reduce the functional availability of miR-150 to increase the level of C-reactive protein.[60] The finding in the present study that ZFAS1 caused intracellular Ca2+ overload might be an alternative mechanism for its proapoptotic property. Yet, future studies are required to verify this point.

Here, we presented the experimental data with both an animal model of MI and a cellular model of hypoxia clearly indicating the pathophysiological function of ZFAS1 in the regulation of cardiac contractile function. Specifically, abnormal upregulation of ZFAS1 in MI or artificial overexpression of ZFAS1 weakened the contractility of cardiac muscles. It seems that the detrimental action of ZFAS1 on contractility is likely accounted for by impaired intracellular Ca2+ handling process. ZFAS1 produced a negative effect on SERCA2a—a key protein in the homeostasis of intracellular Ca2+—via dual mechanisms: expression repression of SERCA2a gene and functional restriction of SERCA2a protein. Reduced SERCA2a availability resulted in a decrease in Ca2+ reuptake back into SR leading to intracellular Ca2+ overload, the latter of which is likely the main detrimental factor for cardiac dysfunction. Our study ruled out the significant contribution of other Ca2+ handling-regulatory proteins, such as RyR2, PLN, and L-type Ca2+ channels to ZFAS1-induced deleterious alterations. These results in conjunction with those reported by Wu et al[60] suggest that ZFAS1 affects cardiac function through at least 2 different processes or pathways: intracellular Ca2+ overload and apoptotic cell death.

An interesting finding in the present study is that the sequence domain of the ZFAS1 gene that is conserved across species mimicked the effects of the full-length ZFAS1 in terms of its effects on SERCA2a expression and functionality and on the associated slowing of decaying process of Ca2+ transient and increase in resting Ca2+ concentration. More strikingly, mutation of this domain rendered a loss of its ability to regulate SERCA2a and [Ca2+]i, and an antisense fragment to the conserved region of ZFAS1 gene efficiently canceled out the deleterious actions of ZFAS1. The computational docking analysis further unraveled the potential direct interaction between ZFAS1-FD and SERCA2a. Such analysis also revealed the core motif with specific nucleotides contained in ZFAS1-FD being responsible for binding to a SERCA domain containing amino acids Ala241, Glu243, Glu689, Ser693, Arg667, Arg671, and Arg672. Most strikingly, the SERCA domain that binds ZFAS1 identified in this study is right within the phosphorylation domain of SERCA, which contains several potential phosphorylation sites (Ser693, Arg667, and Arg671).[52,53] It is known that SERCA pumps are P-type ion motive ATPases that in their sequence structures contain 3 major domains on the cytoplasmic face: the phosphorylation domain (or phosphorylation domain), nucleotide-binding domain, and the actuator domain.[52,53] Although PLN phosphorylation to increase the affinity of SERCA2 for Ca2+ is the most important modulation of this protein, direct CaMKII (calmodulin-dependent protein kinase II)-dependent phosphorylation of SERCA2 has also been documented to be an important route to control the enzyme function,[61,62] particularly when considering the fact that PLN was unaffected by ZFAS1 in our models. Direct phosphorylation at residue Ser[38] in SERCA2a activates enzyme function and enhances Ca2+ reuptake into SR and cardiac contractility.[63] It is possible that SERCA2a can also be activated with its phosphorylation domain being phosphorylated, and when bounded by ZFAS1, the phosphorylation sites of phosphorylation domain are masked by the latter. Such masking would prevent SERCA2a from being phosphorylated by CaMKII or other kinases thereby preventing SERCA2a from being activated. At present, such a view is highly speculative and requires rigorous studies to clarify. Nonetheless, these facts together with our findings would suggest that we have identified the ZFAS1-FD responsible for the effects of this lncRNA on SERCA2a and its associated intracellular Ca2+ handling process.

Potential Implications of Our Findings

The findings in the present study, together with the published studies from our own laboratory[29] and from other research group,[60] indicate that ZFAS1 is not only a biomarker for, or predictor of, MI but also a determinant of cardiac function in the setting of MI and might be in other cardiovascular pathological processes as well. In particular, the fact that knockdown of ZFAS1 is able to rescue the deleterious actions of ZFAS1 on cardiac function opens up an opportunity for correcting the functional impairment of heart caused by ischemia/hypoxia. This would imply that ZFAS1 could be considered a novel therapeutic target for maintaining cardiac function in MI. Or in other words, shZFAS1-V or other forms of ZFAS1 inhibitor could be developed into novel therapeutic agent for ameliorating cardiac dysfunction.

In addition, our finding that AsZFAS1-FD was also able to eliminate the detrimental action of ZFAS1 on SERCA2a and cardiac function suggests that antisense is another valid strategy for the management of MI-induced cardiac impairment. Such an antisense strategy might be even superior to the knockdown approach because it targets directly to the functional domain of the lncRNA, and it does not induce breakdown of ZFAS1 but merely reduces the functional availability of it. One of the advantages of these properties is that on withdraw of the antisense, the function of ZFAS1 could readily recover.

Possible Limitations of Our Study

We are well aware that our study contains several limitations. First, despite that ZFAS1 elicited remarkable suppressive effects on SERCA2a and the associated Ca2+ transient decaying kinetics and contractility, it did not cause proportional depression of cardiac function (EF and fractional shortening). One possible explanation for this is that ZFAS1 might also be able to evoke some other actions in addition to SERCA2a inhibition, which could relieve its detrimental effects and partially preserve the cardiac function. Second, although we have identified a sequence motif that is evidently important to the role of ZFAS1 in regulating SERCA2a and is reasonably well conserved between mouse and human, the conclusion drawn from our animal study may not be extrapolated directly to man. And third, the present study does not answer the question how ZFAS1 altered the mRNA level of SERCA2a. There are a couple of possible explanations for the observation. First, ZFAS1 might act as transcriptional repressor by modifying the chromatin structure/methylation or indirectly by regulating relevant transcriptional factors or RNA maturation via epigenetic modifications. In addition, lncRNAs can also interact with mRNAs through direct antisense binding to either stabilize or induce degradation of the targeted mRNAs. We found that ZFAS1 carries several sequence stretches that are complementary to the 3′-end untranslated region of SERCA2a, and such an antisense relationship might impose a direct interaction between ZFAS1 and SERCA2a mRNA to alter the expression levels mutually. Yet, rigorous future studies are required to clarify these issues.

Sources of Funding

This work was supported, in part, by the grants from the Funds for National Key Research and Development Program of China (2017YFC1307403 to Dr Yang), the Key Program of National Natural Science Foundation of China (81730012 to Dr Yang), and the National Natural Science Foundation of China (81470490 to Dr Ying Zhang and 81570399 to Dr Yong Zhang).

Disclosures

None.