Cyclooxygenases Inhibitors Efficiently Induce Cardiomyogenesis in Human Pluripotent Stem Cells

Application of human pluripotent stem cell-derived cardiomyocytes (hPSC-CMs) is limited by the challenges in their efficient differentiation. Recently, the Wingless (Wnt) signaling pathway has emerged as the key regulator of cardiomyogenesis. In this study, we evaluated the effects of cyclooxygenase inhibitors on cardiac differentiation of hPSCs. Cardiac differentiation was performed by adherent monolayer based method using 4 hPSC lines (HES3, H9, IMR90, and ES4SKIN). The efficiency of cardiac differentiation was evaluated by flow cytometry and RT-qPCR. Generated hPSC-CMs were characterised using immunocytochemistry, electrophysiology, electron microscopy, and calcium transient measurements. Our data show that the COX inhibitors Sulindac and Diclofenac in combination with CHIR99021 (GSK-3 inhibitor) efficiently induce cardiac differentiation of hPSCs. In addition, inhibition of COX using siRNAs targeted towards COX-1 and/or COX-2 showed that inhibition of COX-2 alone or COX-1 and COX-2 in combination induce cardiomyogenesis in hPSCs within 12 days. Using IMR90-Wnt reporter line, we showed that inhibition of COX-2 led to downregulation of Wnt signalling activity in hPSCs. In conclusion, this study demonstrates that COX inhibition efficiently induced cardiogenesis via modulation of COX and Wnt pathway and the generated cardiomyocytes express cardiac-specific structural markers as well as exhibit typical calcium transients and action potentials. These cardiomyocytes also responded to cardiotoxicants and can be relevant as an in vitro cardiotoxicity screening model.


Introduction
Generation of functional cardiomyocytes (CMs) from human pluripotent stem cells (hPSCs) including, human embryonic stem cells (hESCs) and human-induced pluripotent stem cells (hiPSCs), offer a relevant tool not only to study the cardiovascular development, but also for disease modelling, toxicity studies, and drug screening [1]. However, for an in vitro cardiotoxicity screening, it is controversially discussed whether both hESCs and hiPSC-derived CMs are equally suitable [2][3][4][5][6][7].

Cardiac Induction in Monolayer Culture
Undifferentiated HES3 cells were dissociated and seeded on matrigel-coated 60 mm plates at 3 × 10 5 cells/plate and maintained in iPS-Brew XF media with media changed on every alternate day. When cells achieved desired confluence (≥80%), cardiac differentiation was induced by adding CHIR99021 (10 µM) in RPMI/B-27 -ins media (day 0 to day 1). The medium was then changed to basal RPMI/B-27 -ins medium and cells were kept for further 24 h. At day 2, RPMI/B-27 -ins medium with small molecule WNT inhibitor (2.5 µM, 5 µM and 10 µM) was added and cells were kept for 48 h (day 2 to day 4). Afterwards, cells were maintained in basal RPMI/B-27 -ins media and spontaneously beating clusters were visible by day 9 onwards. To enrich the HES3-CM population the beating clusters were kept in DMEM (no glucose) media (Gibco) supplemented with 4 mM sodium DL-lactate up to day 12. Generated HES3-CMs maintained either in RPMI/B-27 -ins media or in iCell cardiomyocyte maintenance media (Cellular Dynamics, Madison, WI, USA).

RNA Isolation and Quantitative RT-PCR
To analyse the mRNA expression, cells were homogenised with QIAzol lysis reagent (QIAGEN, Hilden, Germany), and the total RNA was extracted using the miRNeasy Mini Kit (QIAGEN, Hilden, Germany) according to the manufacturer's instructions (for more details see Supplementary Materials). All of the 27 target genes are listed in Figure S3A.

Action Potential Measurements
For action potential (AP) recordings, day 30 HES3-CMs were dissociated into single cells and plated on sterile 0.1% gelatin-coated glass coverslip. 48 h post-plating, APs were recorded by the whole-cell current-clamp technique using PULSE program and EPC 9 amplifier (HEKA) as we described previously [22] (for brief description see Supplementary Materials).

Calcium Imaging
Calcium imaging was performed in HES3-CMs loaded with 5µM Cal-520 AM (AAT Bioquest, Sunnyvale, CA, USA) according to the manufacture protocol with modifications as described previously [23]. Briefly, cells were cultured in p35 plates with 0.08 mm thin glass coverslip bottom and incubated with 1 ml of loading dye solution containing Cal-520 AM at 37 • C for 15 min in dark. Videos and line scans were captured through the Olympus FV1000 Microscope (Olympus, Tokyo, Japan).

Generation of WNT Reporter Human iPSC Line
The IMR90 (hiPS) cells were expanded on Matrigel in iPS-Brew XF medium for a few days. After reaching desired confluency, the cells were transfected with 5 µg of 7TGP plasmid (7TGP was a gift from Roel Nusse (Addgene plasmid # 24305)) consisting 7×Tcf-eGFP cassette using ViaFect TM Transfection Reagent (E498A, Promega, Madison, WI, USA). Positive clones were then selected by adding 1 µg/ml Puromycin for 2 weeks. After the selection, GFP positive clones were picked up for expansion and WNT reporter assay.

Short Interfering RNA-Targeted Gene Silencing
Pool of COX-1 (M-004556-00-0005) and COX-2 (M-004557-01-0005) siRNAs (Dharmacon, Freiburg, Germany) were used to interfere with human COX-1 and COX-2 expression in IMR90 cells. An equivalent amount of scrambled siRNA was used as a negative control. The IMR90 cells were seeded into 60-mm culture dishes and treated with CHIR99021 from day 0 to day 1. On day 2 cells were transfected with COX-1 and COX-2 siRNA and scrambled siRNA using Magnetofectamine™ (OZBiosciences, Marseille, France) according to the manufacturer's protocol.

Statistical Analysis
All the experiments were performed in triplicates and data are presented as mean ± SEM or ± SD. Unpaired two-tailed Student's t test was used to calculate statistical significance and p values ≤ 0.05 were considered as statistically significant.
Despite their differences in inhibition mechanisms ( Figure 1a) these small molecules efficiently inhibit Wnt signaling and transcription of Wnt target genes. For example, IWP2 and Wnt-C59 inhibit PORCN involved in Wnt synthesis and Sulindac and 3289-8625 inhibit Wnt signaling through blockade of the Dishevelled (Dvl-PDZ domain) whereas XAV939 and FH535 block Wnt signaling by interacting with TNKS and TCF/LEF, respectively. Cardiac differentiation was induced using the protocol outlined in Figure 1b. In brief, the undifferentiated HES3 cells were treated with CHIR99021 (day 0-1) followed by Wnt inhibitor (day 2-4) and maintained in drug-free media from day 4 onwards with media change on every alternate day. At day 12, eGFP + cells were quantified using flow cytometry. Nine out of 10 tested Wnt inhibitors induced cardiogenesis in HES3 cells with eGFP + cell populations ranging from 60-95% and Sulindac with 88-92% eGFP + cells was found to be the most effective small molecule in our model ( Figure 1c). Phase-contrast images taken over the 20 day period of Sulindac induced differentiation show pluripotent HES3 cells and cell in cardiac transition (Figure 1d), spontaneously beating clusters of Sulindac-derived HES3-CMs were observed from day 8 onwards (Video-1). To further enrich this CMs population, the beating cells were cultured in glucose-free DMEM supplemented with sodium lactate from day 12-20, sequential flow cytometry analysis confirmed the enrichment in CMs with 98-99% eGFP + cells (Figure 1e) (Video-2). These CMs can be maintained in synchronously beating cardiac sheets for >120 days (Video-3). To further confirm the purity of the CM cultures and terminal differentiation in CMs, we injected undifferentiated HES3 cells and Sulindac-derived HES3-CMs into immunosuppressed mice. As shown in Figure S1A, we observed tumour formation in mice injected with undifferentiated HES3 cells whereas mice injected with purified Sulindac-derived HES3-CMs did not develop tumours. Lack of efficacy across multiple hESC and hiPSC lines is the major drawback associated with several established protocols [1]. To explore the applicability of Sulindac-based cardiac differentiation protocol, we tested our protocol with three more hPSC lines (IMR90: hiPSC line; ES4SKIN: a patient-derived hiPSC line, and H9: hESC line) at passages between 10 and 25. We observed that Sulindac treatment efficiently induced cardiac differentiation with 50-70% cardiac troponin T positive (cTnT + ) cells ( Figure S1B). Taken together data confirm that Sulindac promotes cardiomyogenesis in hESCs and hiPSCs. (a) Schematic representation of the Wnt signaling transduction pathway. When activation occurs via binding of Wnt ligand to Frizzled receptor, beta catenin translocates into the nucleus and activates transcription of Wnt target genes. Different small molecule Wnt inhibitors interact with different target proteins in Wnt pathway leading to phosphorylation and degradation of beta catenin and inhibition of transcription of Wnt target genes; (b) Protocol for in vitro cardiac differentiation using Wnt signaling modulators. Undifferentiated hPSC colonies were maintained in iPS-brew XF media from days −4 to 0. Mesendodermal differentiation was initiated in the first phase by changing the culture media to RPMI-B27 -/media containing CHIR99021 (10 µM). In the second phase, cells were cultured in the presence of Wnt inhibitors (2.5 µM, 5 µM or 10 µM) from day 2 to 4. Spontaneously beating cardiac clusters were observed from day 9 onwards; (c) Proportion of eGFP + cells observed with different Wnt inhibitors at different concentrations from 3 independent experiments. Cardiac differentiation was carried out according to (b) and the proportion of eGFP + cells was measured on day 12 by flow cytometry (Error bars, ±SEM; n = 3 independent biological replicates, Student's t test, * p ≤ 0.05, ** p ≤ 0.01, *** p ≤ 0.001); (d) Representative phase-contrast images of HES3 cells taken over the 20 day period of Sulindac-induced differentiation show pluripotent HES3 cells and cell in cardiac transition. Scale bar, 50 µm. Videos of day 10 and day 20 CMs are provided in Video-1 and Video-2 respectively; (e) Representative FACS analyses for percent eGFP + cells before lactate treatment (day 12) and after lactate treatment (day 20). Control (Gray) indicates undifferentiated HES3 cells.

Sulindac Treatment was Sufficient to Accelerate Cardiac Differentiation from hESCs and Maturation of Cardiomyocytes
Using custom designed RT 2 profiler PCR arrays, we performed step-by-step molecular analysis of Sulindac-induced cardiac differentiation to unravel the rapid changes in gene expressions leading to cardiogenesis. Figure S2A shows the heat map representing the expression profile changes in the 27 reference pluripotency, mesoderm, and cardiac development marker genes during the differentiation protocol. We observed significant downregulation of pluripotency markers POU5F1 and NANOG with simultaneous upregulation of early mesodermal and primitive streak markers like T-bra, EOMES and MIXL1 from day 1-4, transforming hESCs into mesodermal progenitor cells (MPCs) [25][26][27][28][29] (Figure 2a).
Noticeable, upregulation in cardiac markers like, Myosin heavy chain (MYH6), Troponin T2 (TNNT2) and Troponin I3 (TNNI3) was observed from day 6 onwards (Figure 2d). The mRNA expression pattern of two major isoforms of myosin light-chain 2, MYL2 (MLC2v) and MYL7 (MLC2a) which mark ventricular-and atrial-like cells respectively [36,37] and HCN4 which marks pacemaker-/nodal-like cells [38], confirmed the presence of three subtypes of CMs (Figure 2e). Consistent with a previous report [8], we observed that the maximum numbers of the eGFP + cells were primarily expressing MYL7 by day 10, representing atrial-like and immature ventricular CMs. However, by day 30, the expression of MYL7 decreased with a simultaneous rapid increase in MYL2, suggesting a possible increase in ventricular-like cells and maturation of CMs [39]. In addition, the upregulation of SIRPA and VCAM1 mRNA levels from day 10 onwards suggest that Sulindac treatment leads to transformation of CPCs (NKX2-5 + ) to myocardial committed cells (NKX2-5 + /SIRPA + ) and then to functional CMs   Figure 1d. All Ct values are normalised using GAPDH and relative fold change was calculated using Day 0 HES3 cells as control. Error bars, ±SEM; n = 3 independent biological replicates; (f) Schematic representation of transition of HES3 cells from pluripotent stem cells to cardiomyocytes represented by stage-specific expressed marker genes.

Structural and Functional Characterisation of Sulindac-Derived HES3-CMs
To determine the success of our Sulindac-derived cardiac differentiation protocol, we performed detailed characterisation of the resulting CMs. Immunofluorescence staining of day 30 HES3-CMs with anti-sarcomeric α-Actinin and anti-cTnT antibodies demonstrated a typical striation pattern and organised myofilaments similar to those of adult heart tissue (Figure 3a). Transmission electron microscopy also revealed structural features of Sulindac-derived HES3-CMs such as bundles of myofilaments and sarcomere with Z-line, A-band and abundant mitochondria (Figure 3b). To assess the electrophysiological (EP) properties we performed single cell patch-clamp recordings (Table S1). Based on AP properties like action potential amplitude (APA), action potential duration (APD) at 10%, 50%, or 90% of repolarisation and dV/dtmax, three subtypes of CMs were confirmed (Figure 3c) [41]. Final data analysis revealed CM population comprised of 60-70% ventricular-like, 20-25% atrial-like and 10-15% nodal-like cells (Table S2). We also investigated intracellular Ca 2+ dynamics in Sulindac-derived HES3-CMs by recording Ca 2+ transients using Cal-520 AM calcium indicator. Line-scan images and data analysis showed Ca 2+ oscillations and Ca 2+ transient parameters comparable to that of native CMs (Figure 3d) (Table S3). Furthermore, we also observed spontaneous localised intracellular Ca 2+ release events "calcium sparks" (Figure S2C) (Video-4) released through ryanodine receptor 2 (RYR2) that controls calcium-induced calcium release (CICR) in CMs [42]. RT-PCR and Immunofluorescence staining of ryanodine receptor 2 (RYR2) confirmed gradual increase in RYR2 mRNA levels ( Figure S2D) and the presence of functional RyR2 channels in Sulindac-derived HES3-CMs (Figure 3e). Collectively, these results show that Sulindac-derived HES3-CMs are structurally and electrically mature and functional.

Sulindac-Derived HES3-CMs Offer a Human-Relevant in vitro Cardiotoxicity Screening Model
Functional CMs derived from hPSCs offer a human-relevant in vitro cardiotoxicity screening platform. In this context, the cardiotoxic effects of Doxorubicin (DOXO) were assessed using recently reported xCELLigence Real-Time Cell Analyser (RTCA) based screening platform [43]. DOXO is a anthracycline class anticancer drug, well-known to induce acute cardiotoxicity leading to heart failure [44]. We treated commercially available hiPSC-CMs (iCell Cardiomyocytes) and Sulindac-derived HES3-CMs (day 30) with a single dose of 300 nM, 150 nM and 75 nM DOXO for 48 h followed by 48 h of drug washout. Raw data for cell viability, beating rate, beating amplitude and beating profile were acquired using the xCELLigence RTCA Cardio system and analysed using RTCA Cardio software version 1.0. Figure 4 represents data obtained from Sulindac-derived HES3-CMs (Figure 4a-d) and hiPSC-CMs (Figure 4e-h).
In brief, we observed significant drop in cell index when CMs were treated with 300 nM DOXO (Figure 4a,e) whereas CMs treated with 150 nM DOXO showed a striking irreversible increase in beating rate with a significant decrease in beating amplitude (Figure 4b,f, and c,g). Further analysis of beating profiles of CMs from control and DOXO treated group confirmed the arrhythmic beating in DOXO treated CMs (Figure 4d,h). Our data clearly shows that Sulindac-derived HES3-CMs are more sensitive towards known cardiotoxicants like DOXO compared to hiPSC-CMs. Though factors like differentiation protocol used, culture conditions, batch-to-batch variations, maturity and age of CMs could explain these differences, our findings demand further in-depth analysis of hESC-CMs and hiPSC-CMs based screening models. . Functional studies of Doxorubicin-exposed Sulindac-derived HES3-CMs and hiPSC-CMs using the xCELLigence RTCA Cardio system. The HES3-CMs and hiPSC-CMs were cultured in fibronectin-coated E-cardio plates and treated with Doxorubicin for 48 h followed by 48 h drug washout (W/O). Cells were monitored using xCELLIgence RTCA Cardio system for real time changes; (a,e) The representative graphs displays DOXO exposure induces cytotoxicity and drop in Cell Index values in HES3-CMs and hiPSC-CMs respectively. Error bars, ±SEM; n = 3 independent biological replicates; (b,f) The representative graphs displays DOXO exposure induces changes in % beating rates in HES3-CMs and hiPSC-CMs respectively. Error bars, ±SEM; n = 3 independent biological replicates; (Student's t test, * p ≤ 0.05, ** p ≤ 0.01, *** p ≤ 0.001) (c,g) The representative graphs displays DOXO exposure induces changes in beating amplitude in HES3-CMs and hiPSC-CMs respectively. Error bars, ±SEM; n = 3 independent biological replicates; (Student's t test, * p ≤ 0.05, ** p ≤ 0.01, *** p ≤ 0.001) (d,h) Representative 12 s beating traces of HES3-CMs and hiPSC-CMs after DOXO exposures and during drug washout. Y-axis represents Normalised Cell Index. Beating activity demonstrates the development of arrhythmic beating upon DOXO exposure. Error bars, ±SEM; n = 3 independent biological replicates.
Our initial FACS data also showed that, in comparison to other tested Wnt inhibitors including 3289-8625, which also binds and inhibits Dvl-PDZ domain, Sulindac yields a significantly higher number of eGFP + cells (Figure 1a,c). These data suggest that, Sulindac promotes cardiogenesis not only by inhibiting Wnt signaling but also by a yet unexplored secondary mechanism. Sulindac is a NSAID and non-selectively inhibits both isoforms of cyclooxygenase (COX) namely, cyclooxygenase-1 (COX-1) and cyclooxygenase-2 (COX-2). We hypothesised that this inhibition of COX by Sulindac could also play a role in cardiogenesis. To validate our hypothesis, we knocked down the COX-1 and COX-2 expression in hPSCs either by using small molecule COX-inhibitors like Piroxicam, Nimesulide and Diclofenac or by transfecting pool of siRNAs directed against COX-1 and COX-2. Quantification of eGFP + cells on day 12 confirmed the cardiac differentiation in COX-inhibitor treated HES3 cells with 85-90% eGFP + cells (Figure 5c) similar results were observed in IMR90 cells (data not shown). We next transfected IMR90 and HES3 cells with COX-1 and COX-2 directed siRNA pools and measured the cTnT + and eGFP + cells on day 12 using FACS. Our data show that inhibition of COX-1 alone does not improve cardiac differentiation significantly however inhibition of COX-2 alone or inhibition of COX-1 and COX-2 together significantly improves cardiac differentiation and increase the yield of CMs (Figure 5d) ( Figure S3A). These findings suggest that, controlled and stage-specific inhibition of COX promotes cardiomyogenesis from hPSCs in vitro.
To further explore the molecular mechanism involved in cardiac differentiation and COX inhibition, we transfected IMR90-WNT reporter line with COX-1 and COX-2 siRNAs. The fluorescence images were captured and analysed using ImageJ to obtain relative GFP intensity values. Our data show that inhibition of COX-1 had no significant effect on WNT pathway activity whereas inhibition of COX-2 alone or COX-1 and COX-2 together showed significant downregulation in Wnt pathway activity when compared with scramble siRNA control (Figure 5e,f). This finding explains higher yields of CMs in Sulindac and Diclofenac as they inhibit both COX isoforms. These data also indicate that COX pathway acts upstream to Wnt pathway and both pathways collectively contribute in cardiac differentiation.

Discussion
Landmark discovery to reprogram somatic cells to generate iPSCs have opened up new avenues for developing more physiologically relevant platforms for drug discovery and patient-specific cell therapies [48]. Several recent studies show that hiPSC-CMs can be used in regenerative medicine [49] and an in vitro cardiotoxicity screening thereby offering human-relevant alternative for animal testing [43,50,51]. However, high costs associated with commercially available hiPSC-CMs limit their use for most academic research. In addition, line-to-line and batch-to-batch variations result in alterations in expression of key ion channels in hiPSC-CMs hence demands careful evaluation of the data [6,7,52,53]. Differences in the sensitivity towards several drugs amongst hESC-CMs and hiPSC-CMs also suggest that the screening-assays using hiPSC-CMs must be applied with caution [2,7].
Here, we developed an optimised cardiac differentiation protocol by screening 10 small molecule Wnt inhibitors using HES3 cells [19]. Quantification of NKX2-5 driven eGFP expression on day 12 confirmed cardiac differentiation in nine tested Wnt inhibitors with 30-95% eGFP + cells and spontaneously beating clusters were visible by day 9 onwards. Sulindac induced~92% eGFP + cells and is thereby identified as a potent cardiogenic agent. We further demonstrated that Sulindac-induces cardiac differentiation in another hESC cell line (H9) and 2 different hiPSCs (IMP90 and ES4SKIN) and confirmed the cardiac differentiation with 50-70% cTnT + cells by day 12. Although we observed some variations in the efficiency of cardiac differentiation between hiPSC and hESCs, the pattern of differentiation was similar among all three tested lines.
Consistent with previous reports, we observed that CHIR99021 treatment at an early phase, day 0-1, downregulated pluripotency genes POU5F1, NANOG and generate primitive streak cells expressing marker genes T-bra, MIXL1 and EOMES [1,8]. Concurrent upregulation of MESP1 and PDGFR-α + /KDR + cells confirmed the induction of cardiac mesoderm and generation of CPCs after Sulindac treatment [54]. In addition, RT-PCR data also confirmed upregulation of CPC markers like ISL1, GATA4, NKX2-5, MEF2C, MYOCD, HAND1, and SNAI1 from day 4-8. Combined, these genetic changes then activate cardiac structural genes such as TNNT2, TNNI3, and MYH6 leading to the generation of beating CMs. These data also confirm that Sulindac efficiently promotes cardiac differentiation by inducing genes involved in a highly conserved network that controls initial mesodermal differentiation, proliferation of MPCs and CPCs and maturation of CMs [55]. Using Sulindac we efficiently generated CMs within 20 days in a defined medium supplemented with B27 without insulin which can be maintained in cultures for up to 120 days as synchronously beating cardiac sheets. Moreover, our protocol proved beneficial over other known cardiac differentiation methods as it requires no genetic modification and it could be applicable to many existing hPSC lines.
The hPSC-CM population at an early phase of differentiation is largely composed of MYL7 + /MYL2 -, immature and human fetal-like CMs which then mature and transform into MYL7 -/MYL2 + CMs [1,8,56]. Based on the expression profiles of MYL2 and MYL7 mRNAs, we observed similar transition of HES3-CMs in our study. Electrophysiological characterisation confirmed that Sulindac-derived CMs exhibited AP characteristics similar to that of mature CMs with ventricular-like CMs being predominant, making >65% of total cell population by day 30 and >75% by day 60 (data not shown). These findings suggest that Sulindac-derived HES3-CMs exhibit physiologically relevant EP properties.
In CMs, Ca 2+ transients act as a chemical signal that transduces electrical depolarisation to mechanical movement, also known as excitation-contraction (EC) coupling. Cardiac EC coupling is governed by a Ca 2+ -induced Ca 2+ release (CICR) mechanism and short-lived Ca 2+ signals (sparks) were revealed as elementary events during EC coupling [57,58]. Ca 2+ transient recordings performed with Sulindac-derived HES3-CMs revealed Ca 2+ oscillations comparable to that of mature CMs and local, short-lived Ca 2+ sparks suggesting the presence of functional SR-Ca 2+ stores. Additionally, upregulation of RYR2 mRNA levels and immunostaining indicated the presence of functional ryanodine receptors which in combination with SR-Ca 2+ stores contribute to Ca 2+ transients [59,60] and demonstrates mature and functional CMs.
The hPSC-CMs generated using most known protocols consist of contaminating population of undifferentiated or non-cardiac cells which would increase the tumour formation after transplantation hence limit their use in clinical applications [61,62]. Culturing hPSC-CMs in glucose-depleted culture media containing abundant lactate shown to enrich the CMs to up to 99% [63]. Using a similar approach, we enriched Sulindac-derived HES3-CMs to >97% purity, the purified CMs did not form tumours after transplantation into immunosuppressed Rag -/mice confirming pure terminally differentiated population of CMs.
Besides clinical use, hPSC-CMs are an important tool for an in vitro cardiotoxicity studies. Recent studies from our and several other groups have shown that commercially available hiPSC-CMs can be used for an in vitro toxicity studies and offer a human-relevant alternative for animal models [43,[64][65][66].
Here we show that Sulindac-derived HES3-CMs in combination with xCELLigence Real-Time Cell Analyser (RTCA) offer a robust, sensitive and reliable model for an in vitro cardiotoxicity assessment. We also show that Sulindac-derived HES3-CMs respond to known cardiotoxicants like Doxorubicin, demonstrating functionally mature and electrophysiologically active CMs.
The non-steroidal anti-inflammatory drug (NSAID)-Sulindac, inhibits cyclooxygenase-1 (COX-1) and cyclooxygenase-2 (COX-2) and is well studied for its anti-inflammatory and antineoplastic potential [20,21]. To ascertain the role of COX-1 and COX-2 in cardiogenesis, we knocked down COX-1 and COX-2 levels in HES3 and IMR90 cells by treatment with different NSAIDs (e.g., Piroxicam, Nimesulide and Diclofenac) and by using pool of siRNAs directed towards COX-1 and COX-2. Quantification of eGFP + cells on day 12 confirmed the cardiogenesis with 85-90% eGFP + cells showing for the first time that NSAIDs holds cardiomyogenic potential and can differentiate hPSCs into functional CMs, in vitro. Our findings coincide with recent study from Muraoka et.al. in which diclofenac was shown to enhance cardiac reprogramming via suppression of COX-2 [67]. Moreover, our siRNA study shows that inhibition of COX-2 alone or COX-1 and COX-2 together can efficiently induce cardiac differentiation whereas inhibition of COX-1 alone did not show significant improvement in cardiac differentiation. Several studies have shown that NSAIDs can inhibit Wnt signalling in cancer cells by targeting c-Met and 5-lipoxygenase [68]. However, no conclusive data on the interactions of Wnt and COX signalling in stem cells is available. Our study for the first time shows that siRNA directed inhibition of COX-2 alone or COX-1 and COX-2 together significantly downregulate Wnt pathway activity in hPSCs. These findings also suggest that in stem cells COX pathway might be an upstream event into Wnt signalling pathway.