Uncoupling of Proliferative Capacity from Developmental Stage During Directed Cardiac Differentiation of Pluripotent Stem Cells

Abstract

Lineage-specific differentiation of human-induced pluripotent stem cells (hiPSCs) into cardiomyocytes (CMs) offers a patient-specific model to dissect development and disease pathogenesis in a dish. However, challenges exist with this model system, such as the relative immaturity of iPSC-derived CMs, which evoke the question of whether this model faithfully recapitulates in vivo cardiac development. As in vivo cardiac developmental stage is intimately linked with the proliferative capacity (or maturation is inversely correlated to proliferative capacity), we sought to understand how proliferation is regulated during hiPSC CM differentiation and how it compares with in vivo mouse cardiac development. Using standard Chemically Defined Media 3 differentiation, gene expression profiles demonstrate that hiPSC-derived cardiomyocytes (hiPSC-CMs) do not progress past the equivalent of embryonic day 14.5 of murine cardiac development. Throughout differentiation, overall DNA synthesis rapidly declines with <5% of hiPSC-CMs actively synthesizing DNA at the end of the differentiation period despite their immaturity. Bivariate cell cycle analysis demonstrated that hiPSC-CMs have a cell cycle profile distinct from their non-cardiac counterparts from the same differentiation, with significantly fewer cells within G1 and a marked accumulation of cells in G2/M than their non-cardiac counterparts throughout differentiation. Pulse-chase analysis demonstrated that non-cardiac cells progressed completely through the cell cycle within a 24-h period, whereas hiPSC-CMs had restricted progression with only a small proportion of cells undergoing cytokinesis with the remainder stalling in late S-phase or G2/M. This cell cycle arrest phenotype is associated with abbreviated expression of cell cycle promoting genes compared with expression throughout murine embryonic cardiac development. In summary, directed differentiation of hiPSCs into CMs uncouples the developmental stage from cell cycle regulation compared with in vivo mouse cardiac development, leading to a premature exit of hiPSC-CMs from the cell cycle despite their relative immaturity.

Introduction

Lineage-specific differentiation of human-induced pluripotent stem cells (hiPSCs) into cardiomyocytes (hiPSC-CMs) offers an in vitro model of human cardiac development and disease pathophysiology, and yields an unlimited supply of patient-specific CMs that can be utilized to screen for novel therapeutics and ultimately as a source of cells for cardiac regenerative therapies. A number of protocols have been developed to differentiate hiPSCs into spontaneously contracting CMs, recapitulating features of the native myocardium, including cardiac gene expression and electrophysiological profiles [1,2]. Although these protocols generate hiPSC-CMs with high efficiency, the cells remain relatively immature, thus compromising their utility for modeling health, disease, and drug response in physiologically relevant, mature CMs [3,4].

In most developmental contexts, it is believed that maturation is inversely correlated with proliferation, such that progenitor cells continue to proliferate until terminal differentiation, which is associated with permanent cell cycle arrest [5]. Indeed, in vivo cardiac development and growth is intimately linked to proliferation of both multipotent cardiac progenitors and differentiated CMs, which is directly correlated to the developmental stage and tightly regulated by the expression of intracellular transcription factors and extrinsic signals from the microenvironment. Differentiation of progenitors into an initial pool of CMs occurs early during cardiac development and is largely completed by embryonic (E) day E10 [6]. Subsequent heart growth is accomplished through rapid proliferation of CMs predominantly in the ventricular wall, which declines after E14 to low levels observed during the perinatal period [7]. During the week after birth, rodent CMs undergo DNA replication but do not complete the cell cycle and undergo cytokinesis, resulting in the development of multinucleated/polyploid cells that are believed to have permanently exited the cell cycle [7–9]. In humans, endoreplication appears to begin during late gestation and continues after birth, with CM cytokinesis observed in human hearts up to age 20 [10], with limited CM turnover observed later in life [11].

Given that cellular specialization/maturation appears to be inversely correlated with proliferative capacity during cardiac development, we investigated the temporal changes that occur in proliferation and cell cycle activity during directed hiPSC-CM differentiation to determine whether proliferative capacity correlates with developmental stage as compared with in vivo murine cardiac development.

Material and Methods

Cell culture

Human iPSC lines were generated from skin fibroblasts from healthy individuals (Mayo Clinic IRB 10-006845, 13-007298 and 18-010099) using a CytoTune™-iPS 2.0 Sendai Reprogramming Kit (Invitrogen) by ReGen Theranostics (Rochester, MN, USA). The iPSC lines were cultured in mTeSR™ (StemCell Technologies) on Geltrex™ (Gibco) with daily media changes and were passaged every 4 or 5 days using ReLeSR™ (StemCell Technologies).

Cardiac differentiation was induced in Chemically Defined Media 3 (CDM3—Roswell Park Memorial Institute, recombinant Human Albumin and Ascorbic Acid) containing 8 μM CHIR-99021 and 10 nM Activin A for 20 h, followed by replacement with CDM3 for 48 h [12]. On Day 3, cells were treated with 5 μM IWP2 and 10 nM BMP4 for 48 h, after which CDM3 was changed every 48 h.

Quantitative real-time polymerase chain reaction

RNA was isolated using a RNeasy^® Plus Mini Kit (Qiagen) and cDNA generated using iScript™ Select cDNA Synthesis Kit (Bio-Rad). Real-time polymerase chain reactions (RT-PCR) were run using iTaq™Universal SYBR^® Green Supermix (Table 1) on a CFX384 Touch™ RT-PCR Detection System (Bio-Rad). Values were normalized to the housekeeping gene RSP29 [13], and relative expression was calculated using the delta Cq. Differential gene expression analysis of murine cardiac development was previously quantified using Affymetrix Mouse Genome 430 2.0 arrays (raw data files accession No. GSE43197 in the NCBI GEO database) [14,15]. Heatmaps were generated using ClustVis: a web tool for visualizing clustering of multivariate data (https://biit.cs.ut.ee/clustvis/) [16].

Table 1.

Quantitative Real-Time Polymerase Chain Reaction Primers

Gene name	Primer 1	Primer 2
CCNB1	5′-TGTAGTGAATATGTGAAAGATATTTATGCT-3′	5′-TGAACCTGTACTAGCCAGTCA-3′
CCND2	5′-GACATCCAACCCTACATGCG-3′	5′-CCAAGAAACGGTCCAGGTAA-3′
CDK4	5′-TACCGAGCTCCCGAAGT-3′	5′-TTCAGAGTTTCCACAGAAGAGAG-3′
CDK6	5′-CGAAGTCTTGCTCCAGTCC-3′	5′-TCAACATCTGAACTTCCACGA-3′
CCNG1	5′-GCCTCTCGGATCTGATATCGT-3′	5′-GAGAGTCAGTTGTTGTCAGTACC-3′
CCNG2	5′-TGTATTAGCCTTGTGCCTTCTC-3′	5′-GCTAGGCATTTAGAAACCAACTC-3′
MKI67	5′-CGCCTGGTTACTATCAAAAGGA-3′	5′-GAAGCTGGATACGGATGTCA-3′
MDM2	5′-AGAAGGACAAGAACTCTCAGATG-3′	5′-GTGCATTTCCAATAGTCAGCTAA-3′
GADD45A	5′-TGTACGAAGCGGCCAAG-3′	5′-GGAGATTAATCACTGGAACCCA-3′
RPS29	5′-AAT ATG TGC CGC CAG TGT TT-3′	5′-CCC GGA TAA TCC TCT GAA GG-3′
OCT4	5′-AGT TTG TGC CAG GGT TTT TG-3′	5′-ACT TCA CCT TCC CTC CAA CC-3′
SOX2	5′-CTTGACCACCGAACCCAT-3′	5′-GTACAACTCCATGACCAGCTC-3′
NANOG	5′-CCTTCTGCGTCACACCATT-3′	5′-AACTCTCCAACATCCTGAACC-3′
MIXL1	5′-GAAGGATTTCCCACTCTGACG-3′	5′-GTACCCCGACATCCACTTG-3′
MESP1	5′-CGGTGCTCACAGAGACG-3′	5′-CAGGCGATGGAGCCAAG-3′
GATA4	5′-TTGCTGGAGTTGCTGGAA-3′	5′-GGAAGCCCAAGAACCTGAA-3′
NKX2-5	5′-CACTCAGCATTTGTAGAAAGTCAG-3′	5′-ACCCTAGAGCCGAAAAGAAAG-3′
MEF2C	5′-CTTTCTCTTTCCTGTTTCCTCCA-3′	5′-CCCAAGGACTAATCTGATCGG-3′
TNNT2	5′-TCTTCGTCCTCTCTCCAGTC-3′	5′AGAAGAGGTGGTGGAAGAGTA-3′
TNNI1	5′-CCAGCATTCCTTGGCCTT-3′	5′-GAACAAGGTGCTGTCTCACT-3′
MYL7	5′-GAACATCTGCTCCACCTCAG-3′	5′-GCCTTCCGCATGTTTGAC-3′
CDKN1A	5′-GAGACTAAGGCAGAAGATGTAGAG-3′	5′-GCAGACCAGCATGACAGAT-3′
CDKN2A	5′-CGCTACCTGATTCCAATTCCC-3′	5′-CCAACGCACCGAATAGTTACG-3′
p14ARF	5′-CCC TCG TGC TGA TGC TAC TG-3′	5′-CAT CAT GAC CTG GTC TTC TAG GAA-3′
p15INK4b	5′-GGG AAA GAA GGG AAG AGT GTC GTT-3′	5′-GCA TGC CCT TGT TCT CCT CG-3′
p16INK4a	5′-GGG GGC ACC AGA GGC AGT-3′	5′-GGT TGT GGC GGG GGC AGT T-3′

Flow cytometry

Cells were treated with 10 μM 5-ethynyl-2′-deoxyuridine (EdU) for 1 h (pulse) followed by replacement with growth media and collection of cells at 0, 8, and 24 h (chase). Cells were dissociated on differentiation day ≤5 using TrypLE™, and on day ≥7 by Liberase™ (Roche) plus DNase I followed by TrypLE. Cells were stained with LIVE/DEAD™ Fixable Near-IR Dead Cell Stain Kit (Invitrogen), fixed with 3% paraformaldehyde, permeabilized with ice-cold methanol, and stained with a Click-It™ EdU 555 Flow Cytometry kit (Invitrogen), and primary antibodies [cTnI Alexa Fluor 647 (BD Bioscience) and cTnT FITC (Abcam)] at 4°C overnight. Cells were resuspended in FACS buffer (0.5% bovine serum albumin in Dulbecco's phosphate-buffered saline, with 20 μM ethylenediaminetetraacetic acid) with 1 μg/mL 4′,6-diamidino-2-phenylindole dihydrochloride (DAPI) and were analyzed on either a Celesta or Fortessa Flow Cytometer (BD Biosciences), and data were analyzed using Kaluza Analysis 2.1 (Beckman Coulter).

Confocal microscopy

Cells were permeabilized with 0.5% Triton X-100, and incubated in conjugated primary antibodies at 37°C for 1 h. Coverslips were mounted onto slides with ProLong^® Diamond Antifade Mountant (Molecular Probes) and imaged on an LSM 800 inverted confocal microscope (Zeiss).

siRNA

Two Silencer Select validated siRNA per gene and a Negative Control siRNA (10 nM) were transfected using Lipofectamine RNAiMAX (Invitrogen) at days 14 and 21 of differentiation. Cells were collected at 48 h post-transfection for RT-PCR validation of gene knockdown and at 96 h post-transfection for flow cytometry analysis.

Statistical analysis

All data were analyzed using GraphPad Prism, version 7.03. For normal distributions, parametric tests were utilized, namely one-way analysis of variance, corrected for multiple comparisons using Tukey test. Non-parametric data were analyzed by Kruskal–Wallis test and corrected for multiple comparisons using Dunn's test.

Results

The emergence of cardiac lineages during in vitro differentiation in comparison to developing mouse heart

hiPSC cardiac induction with CHIR-99021/Activin A leads to rapid downregulation of pluripotency genes (NANOG, OCT4 and SOX2) and increased expression of mesoderm genes (MESP1, MIXL1 and TBXT; Fig. 1A), which peak at differentiation day 1 and decline with the addition of BMP4/IWP2. The onset of cardiac specification occurred between day 3 and 9 with the expression of cardiac genes NKX2.5, MYL7, TNNI1, and TNNT2 increasing two to three orders of magnitude and spontaneous contraction starting day 7. This coincides with a rapid increase in cardiac-specific TNNT1 (cTnI) and TNNT2 (cTnT) protein expression between day 5 and 9 (Fig. 1B), when the majority of iPSC-CMs emerge, with peak CM proportion at day 15.

FIG. 1.

Emergence of cardiac lineages during directed hiPSC differentiation. (A) Overview of directed hiPSC cardiac differentiation. (B) Lineage identity was monitored over the course of directed differentiation of human iPSCs by qRT-PCR for loss of pluripotency genes (black) and expression of mesoderm (blue) and cardiac genes (red). (C) Cardiac lineage was quantified over the course of differentiation using flow cytometry analysis of complementary antibodies for cTnT and cTnI (no statistically significant differences were observed between antibodies). Values represent means ± SEM of a minimum of four independent cell lines. hiPSC, human-induced pluripotent stem cells; qRT-PCR, quantitative real-time polymerase chain reaction; SEM, standard error of the mean.

To determine how well hiPSC cardiac differentiation recapitulates natural cardiogenesis, we quantified a broader set of lineage-specific genes and compared them with gene expression data generated from in vivo murine cardiogenesis [14,16]. During differentiation, pluripotency genes are rapidly downregulated and mesoderm genes are upregulated, peaking 24 h post-induction (Fig. 2A). Expression of cardiac progenitor (MEF2C, GATA4, and NKX2.5) and contractile machinery genes (MYL7, TNNI1, and TNNT2) starts to rise on day 7, with robust expression being observed by day 15.

FIG. 2.

Lineage specification during directed hiPSC cardiac differentiation in comparison to mouse cardiogenesis. (A) Pluripotent, mesoderm, cardiac progenitor, and cardiac lineage-specific genes were quantified by RT-PCR throughout hiPSC cardiac differentiation. Values are generated from three hiPSC cell lines in triplicate. The data set was then processed and displayed using the “ClustVis” heatmap function. (B) Data for the murine homologs of the genes examined in A were assembled from [14], processed and displayed using the “ClustVis” heatmap function.

In contrast, during murine heart development (Fig. 2B), both pluripotent and mesoderm genes maintain expression until E7.5, when they rapidly decline, and cardiac specific genes are expressed. Another key difference between the two models is that of cardiac maturation. In murine cardiac development, there is a prolonged period of expression of MEF2C, GATA4, and NKX2.5 ranging from E7.5 into adulthood, whereas these genes are downregulated by differentiation day 25. In addition, there is a marked decrease in expression of the fetal isoforms of cardiac troponin I (TNNI1) and myosin light chain (MYL7) in newborn and adult samples [14], which is not recapitulated in hiPSC-CMs. Based upon these observations at day 25, hiPSC-CMs reach a developmental age similar to E14.5 of murine development.

Decline in DNA replication during cardiac lineage commitment.

CM proliferation is highly dependent on the developmental stage; however, this relationship has not been well characterized during hiPSC-CM differentiation. Quantification of EdU incorporation demonstrated a significant decline in DNA replication during differentiation, from ∼60% of cells replicating DNA at day 0/1 to <5% at day 25 (Fig. 3 and Supplementary Fig. S1). Both the transition between pluripotent and mesoderm lineages, as well as from mesoderm to cardiac are marked by significant reductions in DNA replication.

FIG. 3.

EdU incorporation during directed hiPSC cardiac differentiation. Percent of cells that stained positive by flow cytometry for EdU incorporation (yellow squares), cardiac markers (red), or double positive for cardiac markers and EdU (black). Values represent mean ± SEM of a minimum of five independent cell lines. P-values calculated for percent of EdU+ cells (Kruskal–Wallis, non-parametric analysis of variance, and repeat measures corrected by Dunn's tests). EdU, 5-ethynyl-2′-deoxyuridine.

We specifically examined DNA replication in hiPSC-CMs beginning on day 7, where a few cells were positive for both cardiac markers and EdU; however, between day 9 and 15, the majority of EdU+ cells were also positive for cardiac markers. This period was accompanied by a significant rise in hiPSC-CMs, suggesting that CMs may have a growth or survival advantage over non-cardiac cells in the differentiation. After day 15, the proportion of hiPSC-CMs incorporating EdU declines until <5% of CMs are incorporating DNA at day 25.

Cell cycle distribution during directed CM differentiation

At a whole cell population, hiPSC-CM differentiation was accompanied by significant shifts in cell cycle distribution as defined by bivariate analysis of EdU incorporation [actively replicating DNA (S-Phase)] versus DAPI (total DNA content) to quantify cell cycle states (G1 or G2/M; Fig. 4 and Supplementary Fig. S2). More than 60% of pluripotent cells at days 1/2 were actively replicating DNA, with non-S-Phase cells evenly distributed between G1 and G2/M. On mesoderm induction at day 3, the fraction of cells in G1 increased significantly from 21% on day 1 to 61% by day 3 and continued throughout differentiation with 81% of cells in G1 on day 25. The reciprocal pattern was observed for replicating cells, with S-Phase comprising 61% of the population on day 1, followed by a sustained decrease with only 2% of differentiating cells in S-Phase on day 25.

FIG. 4.

Quantitative cell cycle analysis of directed hiPSC cardiac differentiation. Quantification of cell cycle distributions from bivariate analysis of DNA content versus EdU in all live cells (A), cTnT+ cardiac cells (B), and cTnT- non-cardiac cells (C) from hiPSC directed cardiac differentiation. Values represent means of a minimum of five hiPSC lines per time point; statistical significance was determined using a one-way ANOVA and a Holm–Sidak multiple-comparison test. ANOVA, analysis of variance.

Stratification of the data by cTnT status demonstrated markedly different cell cycle distribution between hiPSC-CMs (Fig. 4B) and non-cardiac cells from the same differentiation (Fig. 4C). Most non-cardiac cells (71%–89%) resided in G1, with the remaining cells divided between S-Phase and G2/M. At earlier time points (days 5 and 7), non-cardiac cells in S-Phase outnumber cells in G2/M at ∼18% and 9%, respectively, as would be expected with the higher replicative rates. On days 9 and 15, the proportions of non-cardiac cells in S-Phase and G2/M were comparable. By days 20 and 25, this distribution had inverted with more cells residing in G2/M (∼11%) and only 2%–5% of non-cardiac cells actively replicating DNA. In contrast, there were significantly fewer G1 and more G2/M hiPSC-CMs throughout differentiation. Although significantly more hiPSC-CMs were observed in the S-Phase during early differentiation, these numbers progressively declined, resulting in no difference in S-Phase populations between iPSC-CMs and non-cardiac cells at day 25.

To investigate cell cycle progression of EdU+ cells through the S-Phase and G2/M, day 20 cells were pulsed with EdU for an hour and then chased in fresh media with samples collected at 0, 8, and 24 h after EdU removal (Fig. 5 and Supplementary Fig. S3). If EdU+ cells traversed the S-Phase and successfully underwent cytokinesis, they would be detected in the early S-Phase gate at the 8 or 24 h time points (F1 cells). Non-cardiac cells (Fig. 5B) showed progression of EdU+ cells through the cell cycle, with the transition of early S-Phase to mid S-Phase cells at 8 h consistent with continued DNA synthesis after the EdU pulse, and subsequent completion of the cell cycle at 24 h with the majority of EdU+ cells returning to early S-Phase (F1 cells). In contrast, a proportion of EdU+ CMs (Fig. 5A) progressed from early to mid S-Phase at 8 h, however few of them returned to early S-Phase at 24 h. Indeed, more than 50% of the CMs that had incorporated EdU during the pulse had not successfully undergone cytokinesis by 24 h. Although these cells were labeled “Late S-Phase,” flow cytometry cannot discriminate between S-Phase and G2/M arrested cells.

FIG. 5.

EdU pulse chase quantitation of cell division during directed hiPSC cardiac differentiation. To examine the progression of EdU+ cTnT+ cardiac (A) and cTNT-non-cardiac cells (B) through the cell cycle, cells on Day 20 of differentiation were pulsed with EdU for 1 h, returned to standard growth media and then cells were collected at 0, 8, and 24 h of recovery. Values represent mean ± SEM, n = 7. *P < 0.01 between cardiac and non-cardiac, early S-Phase cells and F1 cells. ^#P < 0.01 between cardiac and non-cardiac, late S-Phase cells. Statistical significance was determined using a one-way ANOVA and a Holm-Sidak multiple-comparison test. E, early; L, late; M, mid S-Phase.

Cell cycle regulators during CM differentiation versus the developing mouse heart

The hiPSC-CMs display an immature cardiac phenotype yet appear to prematurely exit the cell cycle in comparison to normal cardiac development; therefore, we investigated the upstream cell cycle regulators that may contribute to this observation. As expected, the expression of positive regulatory genes associated with S-Phase and G2/M progression [CCNE1 (Cyclin E1), CCNA2 (Cyclin A2), AURKB (Aurora Kinase B), and CCNB1 (Cyclin B1)] correlates with EdU incorporation, with high levels observed at day 1 that rapidly decline thereafter (Fig. 6A). These genes are also highly expressed in early in vivo cardiac development, but they display a prolonged decline during natural development, with expression still detectable into the newborn period. Likewise, G1 promoting genes CCND2 and CDK4/6 are expressed throughout in vivo cardiac development, peaking during early differentiation and progressively declining into the newborn period, with significant downregulation occurring in adulthood. Again, temporal expression patterns of these genes differ during in vitro differentiation, with expression being restricted to days 18–25.

FIG. 6.

Gene expression of cell cycle regulatory genes during directed hiPSC cardiac differentiation in comparison to mouse development. Positive (A) and negative (B) cell cycle regulatory gene expression. Data were generated from three hiPSC cell lines in triplicate. Data for the murine homologs of the genes examined in hiPSC were assembled from [14]. The data set was then processed and displayed using the “ClustVis” heatmap function.

Cell cycle inhibitors, CDKN1A (p21), 1B (p27), 2B (p15INK4b), and RB1 (Retinoblastoma 1) are also expressed in the later stages of differentiation (days 18–25), with CDKN1B/2B peaking at day 18 and CDKN1A and RB progressively increasing through the end of differentiation (Fig. 6B). Although these expression patterns are similar to those observed for the mouse orthologs during late gestational and early newborn development, they appear to be prematurely activated during hiPSC-CM differentiation. Independent siRNA knockdown of each of these genes (Supplementary Fig. S4A) at days 18 and 25 of differentiation was insufficient to increase CM EdU incorporation (Supplementary Fig. S4B, C) or cell cycle distribution (Supplementary Fig. S4D, E).

The expression of CDKN2A is very distinct between in vitro differentiation and in vivo cardiac development, with the caveat that primers were utilized to specifically distinguish between p14 and p16 expression for in vitro samples [17], whereas the gene array data cannot distinguish between p16 and p19arf (the mouse ortholog of p14arf). CDKN2A is highly expressed in mESCs at E7.5, but it remains low throughout the rest of murine cardiac development. In contrast, p14 and p16 are low during the early stages of differentiation, peaking at day 18 and remaining expressed at the end of differentiation.

In the context of CM proliferation, the observed increases in CDKN2A/B align with the decrease not only in cell cycle activity of in vitro generated CMs (Fig. 4B), but also in the notable decrease in S-Phase associated genes E2F1, cyclins E and A. Thus, although the mRNA expression levels of known cell cycle inhibitory proteins can be reconciled with changes in proliferative capacity in both systems, the timing of the apparent cell cycle withdrawal in vitro is considerably earlier in hiPSC-CM differentiation than in murine cardiogenesis.

Discussion

hiPSC-CM differentiation is a promising tool to study human CM development and disease progression, as well as a method for large-scale production of patient-specific CMs. Although CM development and proliferation are tightly coupled during in vivo development, this relationship and the impact on hiPSC-CM maturation has not been investigated during directed differentiation. Our data demonstrate that although hiPSC-CM differentiation reproduces overall developmental trends observed in murine cardiogenesis, there is an uncoupling of developmental stage and proliferative capacity with hiPSC-CMs exiting the cell cycle despite having a gene expression profile that does not progress beyond E14.5 of mouse development. The hiPSC-CMs display a distinct cell cycle profile with a reduced proportion of cells in G1, and a significantly increased proportion in G2/M when compared with non-cardiac cells derived from the same differentiation. It is unclear in the literature as to whether this is true of in vivo developing CMs; however, the substantial G2/M arrest of cardiac cells during in vitro differentiation is a limiting factor for the production of CMs. Thus, identifying the underlying causes that restrict CM cell cycle progression would not only improve the production capacity of the system, but may also give valuable insights into an aspect of CM cell biology that is central to the regenerative potential of these cells.

hiPSC-CM differentiation displays a significantly condensed period of mesoderm gene expression associated with a rapid decrease in S-Phase cell cycle promoting genes (Fig. 2), leading to decreased replicative capacity. However, there is a temporal delay until the robust expression of cardiac lineage genes. The lack of temporal overlap with S-Phase promoting genes contributes to low DNA replication in hiPSC-CMs and would suggest a blunting of pro-proliferative signals. Murine cardiogenesis shows no temporal delay in the induction of cardiac genes, and thus their expression is concomitant with that of cell cycle promoting genes, supporting higher replicative rates in developing CMs. Indeed, in utero labeling of mouse CMs with tritiated thymidine demonstrated labeling indexes of 20%–46% between E10-E16, which drops to 10%–20% by E18 [7]. Previous examination of mESC-CM differentiation also demonstrated a significant reduction in CM proliferation from 10% to <1% between days 11 and 21 of differentiation, although the developmental maturity of these cells was not assessed [18]. From day 15 onward of hiPSC-CM differentiation, the expression of both cardiac genes and cell cycle activators (CDK4, CDK6, and cyclin D2) is elevated; however, this does not lead to an increase in DNA replication or cell number (Figs. 3 and 4).

Cell cycle inhibitors are also elevated, which may override the cycle cell activators given that p15/p16 can bind and restrict D-type cyclins from activating CDKs 4/6, and p21/p27 restricting CDKs 1/2, thus restricting S-Phase entry and activity and G2/M progression. In contrast, the expression of these inhibitory genes is largely restricted to late gestation and adulthood during murine cardiogenesis, again indicating premature expression of these genes based on the developmental stage of hiPSC-CMs. However, knockdown of these negative cell cycle regulators during late differentiation was insufficient to maintain or promote the cell cycle in hiPSC-CMs, suggesting that a more complex interplay between positive and negative cell cycle regulators leads to loss of proliferative potential.

The uncoupling of the developmental stage from proliferative capacity of hiPSC-CMs may be due to a mismatch between developmental stage and the in vitro environment. It is well recognized that when isolated adult CMs are placed into in vitro culture, they rapidly lose their mature phenotype or die, thus demonstrating the limitations of these in vitro conditions [19–21]. Indeed, high levels of glucose and lack of other oxidizable substrates, including fatty acids, have been demonstrated to impair the maturation of iPSC-derived CMs, with reduction in glucose and supplementation with fatty acids leading to more functionally mature CMs [22–25]. These studies establish a critical dependence of CM maturation on metabolic substrate supply; however, they focused on maturing already established CMs late during differentiation where they have already largely exited the cell cycle, therefore it would be interesting to determine the impact of substrate supply when CMs are still cycling.

Conclusions

In summary, our data indicate that hiPSC-CM differentiation produces immature CMs equivalent to ∼E14.5 of mouse cardiac development. Despite this immature phenotype, hiPSC-CMs have largely exited the cell cycle and display premature downregulation of cell cycle promoting genes and upregulation of cell cycle inhibitory genes. This is in stark contrast to the equivalent developmental stage in vivo, where a large population of these cells remain proliferative. This uncoupling of developmental stage and proliferative capacity in hiPSC-CMs compared with in vivo cardiac development suggests that the microenvironment in which we are differentiating our cells may not recapitulate their normal in vivo environment, establishing the opportunity to manipulate their microenvironment to promote their maturation versus proliferative potential.

Author Disclosure Statement

No competing financial interests exist.

Funding Information

This work is funded through grants from the NIH (HL121079), Mayo Clinic Center for Regenerative Medicine, the Mayo Clinic Center for Biomedical Discovery, and Todd and Karen Wanek Family Program for Hypoplastic Left Heart Syndrome at Mayo Clinic.

Supplementary Material

Supplementary Figure S1

Supplementary Figure S2

Supplementary Figure S3

Supplementary Figure S4