Downregulation of Long Non-Coding RNA Kcnq1ot1: An Important Mechanism of Arsenic Trioxide-Induced Long QT Syndrome

Arsenic trioxide (ATO) is a well-known traditional medicine with the chemical formula As₂O₃. In the 1970s, Dr. Zhang Tingdong, a Chinese researcher, uncovered the extraordinary therapeutic potential of ATO against acute promyelocytic leukemia (APL). Since then, ATO (Trisenox^TM) has become a choice for the treatment of relapsed or refractory APL worldwide. Thus far, it is the most active and effective single agent in the treatment of APL. ATO-based salvage therapy offers long-lasting remissions and cures up to 90% of patients with relapsed APL [1-3].

However, the high frequency and severity of complex arrhythmias as a consequence of the compound’s cardiotoxicity limited the clinical application of ATO. Long QT syndrome (LQTS) is the most common cardiac rhythm abnormality observed during ATO therapy [3, 4]. Approximately, two thirds of patients receiving ATO develop significant but rapidly reversible QTc prolongation, which can predispose patients to temporary discontinuation, torsade de pointes and even sudden cardiac death unless stringent precautions are taken [5]. Due to these adverse cardiac effects, some patients have to be precluded from ATO therapy for APL. Therefore, clarifying the molecular toxicity mechanisms of ATO and minimizing the risk of LQTS are very important.

LQTS is characterized by delayed cardiac repolarization and increased risk of developing potentially fatal ventricular arrhythmias. It can be divided into congenital and acquired LQTS. Congenital LQTS is identified as LQT1 to LQT13 according to mutations in different genes [6, 7]. KCNQ1/KCNE1-encoded slow delayed rectifying K⁺ channel (I_Ks) are expressed in diverse tissues and serve a variety of functions [8-10]. Mutation of KCNQ1 is a cause of LQT1, making up approximately 40% to 55% of all LQTS cases [11]. In addition, Kcnq1-deficient mice show an LQTS phenotype [12]. I_Ks blockade significantly contributes to acquired LQTS, especially when repolarization reserve is reduced [13]. Most studies have shown that the QT prolongation induced by ATO can be attributed to cardiac repolarization abnormalities [14]. The chaos of ion channel stability has been well characterized in ATO-related cardiotoxicity. The alteration of Kcnq1 is involved in ATO-induced QT interval prolongation [14]. However, the regulatory mechanisms for this involvement are still not fully understood.

Long non-coding RNAs (lncRNAs) are functional RNA molecules larger than 200 nucleotides without protein-coding potential. Recently, increasing evidence has suggested that lncRNAs play important roles in cardiac development and diseases, including acute myocardial infarction, hypertrophy and heart failure [15-18]. KCNQ1 overlapping transcript1 (Kcnq1ot1) is an lncRNA transcribed from the KCNQ1 locus. It plays pivotal regulatory role in the expression of both ubiquitously and tissue-specific imprinted genes within the Kcnq1 domain and is thus involved in the pathogenesis of cancer, acute myocardial infarction and Beckwith-Wiedemann syndrome [19-22]. Therefore, it is likely that alteration of the lncRNA Kcnq1ot1 may affect cardiac rhythm through the regulation of Kcnq1.

In preliminary experiments, we found that ATO inhibited the expression of both Kcnq1ot1 and Kcnq1 in primary cultured neonatal mouse cardiomyocytes. In addition, ATO-induced LQTS was associated with the blockage of I_Ks [14]. Based on these findings, we hypothesized that the alteration of lncRNA Kcnq1ot1 may be an important mechanism of ATO-induced LQTS.

Mice were obtained from the Experimental Animal Center of Harbin Medical University. The methods were performed in accordance with the National Guidelines for Experimental Animal Welfare (the Ministry of Science and Technology, People’s Republic of China, 2006) and the Guide for the Care and Use of Laboratory Animals published by the US National Institutes of Health (NIH Publication No. 85-23, revised 1996). All experimental protocols were pre-approved by the Experimental Animal Ethic Committee of Harbin Medical University, China (No. HMUIRB 20150034).

One- to three-day-old neonatal mice were rinsed quickly in 75% ethanol solution for surface sterilization. Hearts were extracted from the body and were transferred immediately into the bacterial dish containing cold DMEM (HyClone, UT, USA). The ventricles were washed and minced into small pieces. The cells were dissociated at 37°C with trypsin-EDTA solution (Beyotime Institute of Biotechnology, Jiangsu, China). Cells from subsequent digestion were added to an equal volume of DMEM containing 10% fetal bovine serum (BI, Kibbutz Beit Haemek, Israel) and were kept at 4°C. After final collection, the resulting mixture was filtered and centrifuged for 3 min at 1500 rpm to obtain pellet cells. The cells were resuspended in culture medium and incubated for 1.5 h under a water-saturated atmosphere of 5% CO₂-95% air to allow the attachment of non-cardiomyocytes. The suspended cells were then collected, plated into a new culture plate and incubated under the same conditions as mentioned above.

In vitro and in vivo model of ATO-induced cardiotoxicity

Neonatal mouse cardiomyocytes were incubated with or without ATO (2.5 or 5 μM) for 48 h. Subsequently, the total RNA and protein were extracted for the following experiments. Male adult Kunming mice weighting 25-30 g were used in this experiment. ATO (0.5, 1.5 or 4.5 mg·kg^-1) or an equivalent volume of saline was administered to mice through the tail vein on alternate days. The electrocardiograms (ECGs) were monitored before and 7 days after treatment. The cardiac tissues were collected and frozen at -80°C until further analysis.

Neonatal mouse cardiomyocytes were transfected with siRNA targeting Kcnq1ot1 or scrambled negative control siRNA (GenePharma, Shanghai, China) using X-treme GENE siRNA transfection reagent (Roche Applied Science, Prague, Czech Republic) according to the manufacturer’s instructions. The siRNA sequences were as follows: si-Kcnq1ot1-1, GGGAAUCUGGUCUAAUGAATT, UUCAUUAGACCAGAUUCCCTT; si-Kcnq1ot1-2, CCUGGUGAAGGUACUAAAUTT, AUUUAGUACCUUCACCAGGTT.

Lentivirus-shRNAs were designed and synthesized by GenePharma (GenePharma, Shanghai, China). Initially, the effects of lentivirus-shRNA were verified in vitro. Lentivirus-shRNAs were added to culture medium to yield a final multiplicity of infection (MOI) of 1. For in vivo experiments, lentivirus-shRNA (1×10⁹ TU) diluted in 50 μL of saline was administered to mice through the tail vein.

Action potential duration (APD) was measured by the whole-cell patch-clamp method using a MultiClamp 700B amplifier (Axon Instruments, CA, USA). All experiments were performed using pCLAMP 10.2 software (Axon Instruments, Foster City, CA, USA). Boltzman’s fits were performed.

The standard limb lead II ECG was continuously recorded using a BL420s multichannel recorder (TME Technology, Chengdu, China). Arrythmias in mouse hearts were measured by ECG for a continuous period of 10 min. The rate-corrected QT interval (QTc) was calculated using Bazett’s formula: QTc = QT/(RR/100)^1/2 [23].

Total RNA was extracted using TRIzol reagent (Invitrogen, CA, USA) from neonatal mouse cardiomyocytes or mouse myocardium. First-strand cDNA was synthesized using a reverse transcriptase kit (TaKaRa, Shiga, Japan) according to the manufacturer’s instructions. Real-time RT-PCR analysis was performed on an ABI 7500 fast Real Time PCR system (Applied Biosystems, CA, USA) using SYBR Green I (Toyobo, Osaka, Japan). The thermal cycling conditions consisted of one cycle for 60s at 95°C and 40 cycles of 15 s at 95°C, 15 s at 60°C and 45 s at 72°C. GAPDH served as an internal control. Relative expression levels of target genes were calculated using the 2^-ΔΔCT method. The primers sequences were as follows: Gapdh, 5’-AAGAAGGTG-GTGAAGCAGGC-3’ (forward), 5’-TCCACCACCCTGTTGCTGTA-3’ (reverse); Kcnq1ot1, 5’-GCACTCTGGGTCCT-GTTCTC-3’ (forward) 5’-CACTTCCCTGCCTCCTACAC-3’ (reverse); Kcnq1, 5’-CAAAGACCGTGGCAGTAAC-3’ (forward), 5’-CCTTCATTGCTGGCTACAAC-3’ (reverse); Kcnh2, 5’-TGGGGAGAAAAGTGACACCTG -3’ (forward), 5’-CAGGGCTAGACAAGGGGATG -3’ (reverse); Kcne2, 5’-GGTCTCAAGCTGAAAGTGCC -3’ (forward), 5’-TGCT-GTGTGGTATGTGAGCA -3’ (reverse).

Total protein extractions were obtained from neonatal mouse cardiomyocytes or cardiac tissues. The protein samples were separated by electrophoresis in 10% sodium dodecyl sulfate polyacrylamide gels, and the protein was transferred to polyvinylidene difluoride membranes (Millipore, MA, USA). Membranes were blocked with 5% non-fat dry milk in phosphate-buffered saline (PBS) for 2 h. The blots were incubated overnight at 4°C with primary antibodies. The next day, membranes were washed three times for 10 min each time with PBS containing 0.5% Tween 20 (PBS-T) and were then incubated with secondary antibodies for 1 h at room temperature. Finally, membranes were rinsed with PBS before scanning. Images were captured on the GelDox XR System (Bio-Rad, CA, USA). Western blotting bands were quantified using Quantity one software. Gapdh served as an internal control.

Data are expressed as the means ± SEM and were analyzed with SPSS 13.0 software. Statistical comparisons between two groups were performed using Student’s t-test. Statistical comparisons among multiple groups were examined using analysis of variance (ANOVA). A two-tailed P < 0.05 was considered statistically significant. Graphs were generated using GraphPad Prism 5.0.

To evaluate the effect of ATO on QTc interval, mice were administered different doses of ATO. No significant difference was observed between the 0.5 mg·kg^-1 group and the control group after 7 days of treatment. However, the QTc interval was prolonged in the 1.5 and 4.5 mg·kg^-1 groups compared with the control group (Fig. 1).

The cardiac tissues were harvested after ECG recording in mice. The expression levels of lncRNA Kcnq1ot1 and Kcnq1 were determined using real-time PCR and western blot analysis. ATO at the doses of 1.5 and 4.5 mg·kg^-1 significantly decreased Kcnq1ot1 expression (Fig. 2A). Meanwhile, the Kcnq1 mRNA and protein expression levels were both downregulated (Fig. 2B & C). Similarly, Kcnq1ot1 and Kcnq1 expression levels were also downregulated in neonatal mouse cardiomyocytes incubated with ATO (Fig. 2D-F).

To assess the effect of lncRNA Kcnq1ot1 knockdown on Kcnq1 expression, two different siRNAs were used. lncRNA Kcnq1ot1 expression was significantly downregulated after transfection in neonatal mouse cardiomyocytes. Both of the siRNAs targeting lncRNA Kcnq1ot1 exerted similar effects on Kcnq1 expression. The mRNA and protein expression levels of Kcnq1 were both downregulated after the silencing of Kcnq1ot1 (Fig. 3). The siRNA with better efficiency was selected for the subsequent experiments.

APD in cardiac cells is the major determinant of the length of the QT interval. APD is governed by the balance between membrane depolarization and repolarization. The delayed rectifier K⁺ current carried by the I_Ks channel is responsible for the terminal phase of cardiac repolarization, and the latter is encoded by Kcnq1, which is regulated by Kcnq1ot1. Our results showed that after Kcnq1ot1 knockdown, the APD was significantly prolonged in primary cultured neonatal mouse cardiomyocytes (Fig. 4A & B). However, Kcnh2 and Kcne2 mRNA expression levels were not changed (Fig. 4C & D).

The lentivirus carrying shRNA targeting lncRNA Kcnq1ot1 was constructed based on the siRNA sequence. After incubation with the lentivirus-shRNA, Kcnq1ot1 expression was significantly downregulated in neonatal mouse cardiomyocytes. Meanwhile, Kcnq1 mRNA and protein expression levels were also inhibited (Fig. 5).

To determine whether lncRNA Kcnq1ot1 plays a role in cardiac repolarization, ECG recordings from Kcnq1ot1 knockdown and negative control mice were evaluated. The lentivirus-shRNA or the negative control shRNA was administered to mice through the tail vein. The QTc interval was prolonged from day 7 after Kcnq1ot1 knockdown (Fig. 6).

To determine the effects of lncRNA Kcnq1ot1 on Kcnq1 expression in vivo, cardiac tissues were harvested 7 days after administration of the lentivirus-shRNA. In accordance with in vitro experiments, lncRNA Kcnq1ot1 expression was downregulated. Meanwhile, Kcnq1 mRNA and protein expression levels were also decreased (Fig. 7A-C). However, downregulation of Kcnq1 exerted no effect on lncRNA Kcnq1ot1 expression (Fig. 7D-F).

ATO is a highly effective treatment for APL. However, it can cause sudden cardiac death due to LQTS. Although dozens of studies have proposed several possible mechanisms, including disturbance of ion channel balance, apoptosis stimulation and promotion of oxidative injury, the exact regulatory network of ATO-induced cardiac toxicity has not been fully clarified [24-26].

ATO at a median dose of 0.15 mg·kg^-1·d^-1, ranging from 0.06 to 0.2 mg·kg^-1·d^-1, is recommended for the clinical treatment of APL. The QT changes that occur are usually observed within 7 days after treatment [5]. To mimic clinical drug administration, mice were treated with ATO at dose of 0.5, 1.5 or 4.5 mg·kg^-1, with the median dose of ATO in mice being equal to that in humans. The ECGs of mice were then monitored during the experiment. Seven days after treatment, the QTc was prolonged in mice treated with median and high dose of ATO (Fig. 1). This result was in accordance with the research exploring the effect of ATO on QTc interval prolongation in mice [27]. Based on this model, we explored the molecular mechanisms of ATO-induced LQTS.

lncRNAs constitute a novel class of non-coding RNAs that regulate gene expression. Although there is a growing interest in lncRNA-mediated biological and pathophysiological functions, little is known about the involvement of lncRNAs in the regulation of cardiac rhythm. Kcnq1ot1 is a ubiquitously expressed lncRNA that is considered to have regulatory effects on cardiac development and is aberrantly expressed in the setting of myocardial ischemia [28, 29]. To explore the molecular mechanisms of ATO-induced LQTS, we focused on Kcnq1ot1. Our results showed that ATO inhibited the expression of Kcnq1ot1 both in vitro and in vivo (Fig. 2A & D). Moreover, Kcnq1 expression was also downregulated (Fig. 2B, C, E & F).

To investigate the effects of lncRNA Kcnq1ot1 on Kcnq1, siRNAs targeting lncRNA Kcnq1ot1 were synthesized to knock down Kcnq1ot1 in neonatal mouse cardiomyocytes. Both of the siRNAs could efficiently and specifically down regulate Kcnq1ot1 expression, resulting in significant inhibition of Kcnq1 (Fig. 3), and the APD was prolonged due to lncRNA Kcnq1ot1 knockdown (Fig. 4A & B). Although I_Kr is also considered as an important regulator in LQTS pathogenesis, lncRNA Kcnq1ot1 knockdown exerts no significant effects on the mRNA expression of Kcnh2 and Kcne2, which encode I_Kr (Fig. 4C & D).

To further detect the effect of lncRNA Kcnq1ot1 knockdown in vivo, lentivirus-shRNA targeting lncRNA Kcnq1ot1 was constructed. The effects of lentivirus-shRNA were verified in vitro. After transfection, Kcnq1ot1 expression was inhibited in neonatal mouse cardiomyocytes (Fig. 5A). Meanwhile, Kcnq1 mRNA and protein expression levels were also decreased (Fig. 5B & C). We then performed an in vivo experiment. The mice treated with lentivirus-shRNA targeting lncRNA Kcnq1ot1 exhibited a long QT phenotype in 7 days after lentivirus-shRNA administration (Fig. 6). In addition, Kcnq1 expression was also inhibited. These observations indicate that lncRNA Kcnq1ot1 knockdown could inhibit Kcnq1 expression (Fig. 7A-C). These results disagreed with the research conducted by Korostowski et al., who purposed that the absence of Kcnq1ot1 leads to Kcnq1 overexpression [29]. This discrepancy could be explained by the variability of lncRNA effects depending on development stage and pathologic status. We also investigated whether a positive feedback loop existed between lncRNA Kcnq1ot1 and Kcnq1 by knocking down Kcnq1 to observe the changes in lncRNA Kcnq1ot1 in vivo. Silencing of Kcnq1 exerted no significant effect on lncRNA Kcnq1ot1 (Fig. 7D-F). These findings indicated that ATO-induced LQTS was at least partially mediated by the regulation of the lncRNA Kcnq1ot1/Kcnq1 axis.

Theoretically, a rescue experiment aiming to observe whether Kcnq1ot1 upregulation could attenuate ATO-induced LQTS is indispensable. Unfortunately, due to technical difficulties, lncRNA Kcnq1ot1 overexpression could not be achieved. Therefore, RNA interference was performed to mimic the effect of ATO on Kcnq1ot1 in cardiomyocytes. Considering the characteristics of lncRNAs, it was difficult to proceed from mouse to human. Compared with protein-coding sequences, most lncRNAs are poorly conserved among vertebrates [30]. A recent study identified only 18 conserved lncRNAs in humans from 3133 mouse lncRNAs, confirming the fact that only a small minority of lncRNAs in mice or humans have transcribed homologous sequences across different species [31, 32]. In contrast, highly conserved lncRNAs could play critical and conserved roles across different species [33]. Fortunately, Kcnq1ot1 is a relatively conserved lncRNA and likely exerts similar effects in different species. For instance, the dysfunction of Kcnq1ot1 is responsible to Beckwith-Wiedemann syndrome in humans, similar to what has been observed for large offspring syndrome in bovines [34, 35]. In the present study, we found that the aberrant expression of lncRNA Kcnq1ot1 was involved in ATO-induced LQTS in mice and could likely have a similar effect in humans. However, it remains to be further investigated whether homologous lncRNA KCNQ1OT1 is aberrantly expressed in APL patients treated with ATO or merely in those with LQTS.

Overall, this study is the first to demonstrate that lncRNAs have regulatory potential for cardiac rhythm. The discovery of lncRNA Kcnq1ot1-mediated effects on cardiac rhythm in ATO-induced LQTS not only proposes a therapeutic target for ameliorating the cardiac toxicity of ATO but also provides a novel strategy for the treatment of arrhythmias and other related diseases.

The authors declare no conflicts of interest.

This project was supported by the National Natural Science Foundation of China (Grant No. 81673426 and 81603105), the Foundation for Innovative Research Groups of the National Natural Science Foundation of China (Grant No. 81421063), National Natural Science Foundation of Heilongjiang Province (Grant No. LC2015034). B.Y., Y.B., P.V. and Y.J. participated in research design. Y.J., Y.B., Z.W. and E.Y. wrote or contributed to the writing of the manuscript. Q.C., G.T., L.C. and X.S. performed the western blotting. Q.C. and Y.Q. performed the real-time PCR. Y.J. and A.L. performed primary cell culture and transfection. W.D. and G.L. performed experiments in the mouse model. Y.L. performed patch clamp analysis. H.W. performed data analysis. All authors reviewed the manuscript.

Y. Jiang, W. Du and Q. Chu contributed equally to this work.