Can We Stop One Heart from Breaking: Triumphs and Challenges in Cardiac Reprogramming

Abstract

Ischemic cardiac injury causes irreversible muscle loss and scarring, but recent years have seen dramatic advances in cardiac reprogramming, the field focused on regenerating cardiac muscle. With SARS-CoV2 increasing the age-adjusted cardiovascular disease mortality rate, it is worth evaluating the state of this field. Here, we summarize novel innovations in reprogramming strategies, insights into their mechanisms, and technologies for factor delivery. We also propose a broad model of reprogramming to suggest directions for future research. Poet Emily Dickinson wrote, “If I can stop one heart from breaking, I shall not live in vain.” Today, researchers studying cardiac reprogramming view this line as a call to action to translate this revolutionary approach into life-saving treatments for patients with cardiovascular diseases.

Introduction

A 2023 update to the American Heart Association’s statistics on cardiovascular disease revealed that 2020, the first year of the SARS-CoV2 pandemic in the United States, also saw a record high in CVD-related deaths, topping the previous high in 2003 by nearly 19,000 deaths. What’s more, this increase has hit Asian, Black, and Hispanic communities disproportionately. (1) Ischemic injury, such as a myocardial infarction, accounts for a large chunk of this mortality. (1,2) In mammalian hearts, non-muscle cells stop contributing to the pool of heart muscle cells, cardiomyocytes (CMs), early in fetal development, and CMs lose nearly all proliferative capacity shortly after birth. Beating CMs lost during a heart attack are therefore largely replaced by less functional scar tissue, and this loss of function often leads to chronic heart failure. Whole organ transplant is the gold standard of treatment because there do not currently exist therapies to reduce scar size and improve cardiac function after ischemic injury. (3)

If medication or gene therapy could cause a patient's own CMs to reenter the cell cycle or cause the cardiac fibroblasts (CFs) already populating the scar area to change into beating CMs, then we could improve cardiac function after ischemic injury reducing loss of life and the need for invasive therapies like organ transplant. (3) The first hints that such an intervention might be possible came in 1987 when researchers converted fibroblasts to myoblasts by expressing the cardiac lineage regulator MYOCD. (2) However, it is only in the decade immediately preceding the pandemic that research into direct cardiac reprogramming—reprogramming of non-myocytes, typically fibroblasts, into induced cardiomyocytes (iCMs) without passing through a pluripotent intermediate—as a potential therapeutic for ischemic injury graduated from a curiosity to an inevitability. (3) However, low reprogramming efficiency and incomplete maturation of iCMs remain significant barriers to clinical application.

Transcription factors, micro-RNAs, and small molecules are the three broad categories of factors in reprogramming cocktails. Table 1 presents a survey of cocktails used in literature published since 2019. Whichever factors are included, these cocktails are designed to repress starting cell identity and promote CM identity. The most common TF reprogramming strategy for CMs uses three transcription factors: GATA4 (G), MEF2C (M) and TBX5 (T). Often included with GMT is HAND2, (3) which may lead to more mature iCMs. (4) The most common microRNA strategy is “miRCombo”, miR-1, miR-133, miR-208 and miR-499. (5) Strategies using small molecules typically include inhibitors of fibrosis, inflammation and DNA and histone methylation and promoters of chromatin opening and cAMP production. (2)

Table 1:

Survey of recently published literature on cardiac reprogramming and maturation strategies. The factors listed form the optimal cocktail for each study except where otherwise noted. The efficiency of each strategy given represents the most specific assessment assay reported (e.g., for flow cytometry experiments with multiple CM markers, only the double-positive percentage is listed). Where possible, the assay is listed first followed by the effieciency, and when studies use multiple cocktails or models they are given in parentheses. Abbreviations: CM-cardiomyocyte; pm-pacemaker; CPC-cardiac progenitor cell; EC-endothelial cell; Fb-fibroblast; CF-cardiac fibroblast; MEF-mouse embryonic fibroblast; DF-dental fibroblast; H9-human embryonic stem cell line; ADMSC-adipose-derived mesenchymal stromal/stem cell; UC-urine cell; TTF-tail tip fibroblast; VM-ventricular myocytes; AM-atrial myocytes; EHM-engineered human myocardium; HFF-human foreskin fibroblast; m-mouse; r-rat; h-human; a-adult; n-neonatal; i-induced; d-dividing; miR-microRNA; TF-transcription factor; SM-small molecule; DN-dominant negative; Np-nanoparticle; AA-ascorbic acid; ElecStim-electrical stimulation; RA-retanoic acid; Mat-Matrigel; RCCS-rotary cell culture system; FC-flow cytometry; IF-immunofluorescence.

Target	Starting	Factors	Type	Delivery	Maturation	Max Efficiency	Ref
iCM	mnCF; in vivo	miRCombo (miR-1, miR-133, miR-208, miR-499)	miR	Carbon dots	AA + Se	100x cells/field vs cnt (cell morphology); 1/3 Fibrosis area post-MI	(36)
	mnCF	miRCombo	miR	H pep-mod neutrophil-mimicking membrane-mesoporous Si Np		FC cTnT+ 15.2%	(72)
	haCF	miRCombo	miR	lipoplexes		FC cTnT+ 15%	(73)
	haCF	miRCombo	miR	DharmaFect 1		FC cTnT+ 11%	(5)
	MEF, maCF	PTC-209 (BMIi) -> CHIR990 21 + RepSox (TGFβR1i) + Forskolin + VPA (HDACi) + Parnate (LSD1i) + TTNPB (RA analogue)	SM	SM		FC aMHC+ 41% (MEF); 27.2% (maCF)	(39)
	MEF	5-azac + A83-01	SM	SM		FC cTnT+ 48.7%; FC ActA+ 57.3% FC aMHC-GFP+ 22.2% (MEF); IF aMHC+ 40% (HDFs); 1/8 scar area post-MI FC cTnT+ 17% (rCF); 7.5% (hCF) FC cTnT+ 13% (EC); 10.3% (CF) IF ActA+ 10.5% (MEF AM); 9.5% (MEF AMGT); FC cTnT+ActA+ 9.5% (mnCF AM); 14% (mnCF AMGT); 6.5% (hESC-Fb AM); 11.9% (hESC-Fb AMGT)	(42)
	MEF; hnDF; in vivo	GMT	TF	Au Np			(74)
	MEF; rnCF; hnCF	GM + Tead1	TF	Lenti			(46)
	rECs; rnCF	EC: GMT CF: GMT + ETV2	TF	Lenti			(28)
	MEF; mnCF; H9-Fb	Ascl1 + M(133); AMGT(13 3)	TF	Lenti + Retro-Polycistronic			(24)
	H9c2 cardiomyoblast	MH	TF	Lenti tet-on Mef2c + recombinant Hand2		FC aMHC-GFP+ 30%	(76)
	hADMSC	GMT or MESP1	TF	mRNA	Cardiac DIff Media	10-100x expression of actn2, tbx5, gja1, tnni3	(75)
	hnDF	GMTH + MYOCD	TF	Neon Transfection	Coculture w/hECs in layers + ElecStim	3x cTnT expression vs monoculture	(4)
	MEF; in vivo	MGTH	TF	Retro		IF TnI+ 39%, IF Lin+ 2% iVM, 12% iAMs (MEF); 100 iCM/section, 100% iVM (in vivo)	(21)
	hUC	MT + Mesp1 + MyoCD PHF7 + GMTH + Akt1	TF	Retro		FC cTnT+ 22.4%	(19)
	TTF		TF	Retro		FC cTnT+aMHC-GFP+ 18.8%	(32)
	Fb	MGT	TF	Retro-polycistronic		NA	(43)
	MEF; TTF; mnCF	Mef2c isoform 2 + GT	TF	Retro-polycistronic		FC aMHC-GFP+ 5.5% (MEF); 11.5% (mnCF); 12.8% (TTF) IF Titin-eGFP+ActA+ 25%	(30)
	Fb	MGTH	TF	Retro-polycistronic			(78)
	MEF	GMTH	TF	Sendai	8kPa Mat	IF ActA+ 20%	(84)
	in vivo	GMT	TF	Sendai		IF cTnT+ 2.5%; 1/3 fibrosis area post-MI	(70)
	Mouse non-CM; in vivo	GMTH + DN-TGFb + DN-Wnt8a + acid ceramidase	TF + DN + SM	modRNA		IF aMHC-mCherry+ ActA+ 42% (nonCM); 1/2 scar area post-MI	(68)
	H9-Fb; hCF	MGT133 + TBX20	TF + miR	Lenti-polycistronic		IF cTnI+ 40% (H9-Fb); 70% (HCF)	(27)
	hCF	GMT or HMyocd + shp63	TF + shRNA	Lenti		FC cTnT+ 4.3% (GMT shP); 14.9% (HMyshP)	(69)
	H9-Fb	MGT133 + shEZH2	TF + shRNA	Retro-polycistronic		FC cTnT+ 32.7%	(38)
	mnCF; maCF; MEF: H9-Fb	MGT + Atg5 or shBeclin1	TF + shRNA	Retro/Lenti-polycistronic		FC cTnT+ MGTA 9.4% (mnCF); MGTshB 21% (mnCF); 5% (maCF); 6% (MEF); 20% (H9-Fb)	(51)
	MEF; raCF; haCF	rat: GMT; human: GMTH + MYOCD + miR-590 GHMT + DMSO GMT + CREB/CB Pi MGT + IGF1 + Mll1i (MM589) + A83–1 + PTC-209 shox2 + hcn2 + tbx5 or tbx18	TF	Lenti	HDACi (NaBut) + WNTi (ICG-001) + RA	FC cTnT+ 23.2% (raCF); 25% (haCF)	(44)
	MEF		TF + SM	Retro		IF cTnT+ 13%	(41)
	MEF		TF + SM	Retro		FC aMHC-GFP+ 20%	(40)
	MEF; mnCF		TF + SM	Retro-polycistronic		FC aMHC-GFP+ 25% (MEF); 12.9% (mnCF)	(49)
	hADMSC-iCPC		TF	Lenti		CX30.2-mCherry+ 5.9%	(17)
ipmCM	MEF	GHMT + Isl1; GMT + d1-32HAND2	TF	Retro		IF Lin+ GHMTI 0.6% iPM, 0.5% iAM, 0.4% iVM; GMTdH 1% iPM, 0.1% iAM, 0.1% iVM	(22)
Purkinje Cells	hCM; hiPSC-CM	Rolipram + Forskolin + CHIR990 21 + SB43154 2 + VPA + RG108 + Parnate + Resveratrol	SM	SM	RA + NRG1 + Epi	Contactin2-mCherry+ 15.5%	(18)
dCM	EHM; in vivo	mouse epicardial ECVs	ECVs	Media/Injection	Removal of factors	FUCCI SG2M CM 4%; 15% D7 after P1MI; 4% D7 after P7MI; IF PH3+ 0.5% (EHM)	(10)
	in vivo rat	SubQ W8B2+ hCSC	ECVs	TheraCyte device		2/3 scar area post-MI	(11)
	In vivo	WNTi (CDMG1/2)	SM	SM	Removal of factors	IF PH3+ 0.14%; 2/3 scar area post-MI	(9)
	H9-CM	miR-10b	miR	Lipo-fectamine RNAiMAX MYH6-Cre-caErbB2 aMHC-Cre-YAP5SA		IF Edu+ 5.5%	(6)
	maCM	caErbB2	TF			IF AurKB 2%	(58)
	in vivo	caYap	TF			IF PH3+ 0.95% IF PH3+ 0.7% (ex vivo maCM); 2/3 scar area post-MI, IF EdU+ 5-11% at injury border	(7)
	maCM; in vivo	transient OSKM	TF	i4F-Cre-heart	Removal of factors		(8)
CPC	MEF; HFF	CHIR990 21 + A83-01 + GSK126 + Forskolin + CTPB + AM580 GT + Nkx2.5 + CHIR990 21	SM	SM		FC Flk1+PdgfRα + 75.2%	(15)
	Fb		TF + SM	Lenti CRISPRa		Nkx2.5-eGFP+ 85.8%	(12)
	HFF	GMTH + MEIS1	TF	Lenti CRISPRa	ActA + BMP4 -> Epi -> RCCS	FC cTnT+ 8.7%; TPM11+ 62.95%	(13)
	hADMSC	MESP1 + ETS2	TF	Recombinant Protein	3D Spheres	FC KDR+PDGFR a+ 71.4%	(14)

Novel reprogramming strategies aim to improve the reprogramming efficiency of these base strategies and/or the maturity of the resulting CMs. The gold standard for assessing novel reprogramming strategies in vitro is to evaluate iCM morphology and expression of CM markers (typically, sarcomeric proteins such as cardiac troponin T and α-actinin evaluated with rt-qPCR, flow cytometry, and immunocytochemistry). In vivo assessments include lineage tracing and evaluating cardiac function post-MI. Assays for assessing maturity of iCMs include evaluating for expression of chamber-specific markers, regularity of sarcomere organization, proportion of cells exhibiting calcium cycling, and frequency of spontaneously beating loci. (2,3)

The purpose of the present literature review is to summarize the current state of the field, with particular focus on the advances occurring since the pandemic brought new levels of urgency to the issue. We will lay out the key factors that have been used for reprogramming in the heart with an eye toward common mechanisms. We will also look at modulation of signaling pathways to improve reprogramming cocktails. Finally, we will explore recent research into safe and efficacious delivery of factors and remaining barriers to clinical translation. Our hope is that this review will synthesize work from a broad range of researchers in various fields to help direct future research to realize that goal.

1. Advances in Cardiac Reprogramming Factors

1.1. Cardiac Reprogramming Strategies to Specific Target Cell Types

One attractive target for cardiac regenerative medicine is causing mature CMs to reenter the cell cycle like developing, proliferative CMs. Strategies to induce proliferation in mature CMs are varied. A fairly straightforward strategy involves exogenous expression of cell cycle regulators or treatment with their agonists could induce proliferation in post-mitotic rat and mouse CMs. (6) Some investigators have found that promotion of metabolic reprogramming to a more glycolytic state (7) is sufficient to promote cell cycle entry for mature CMs. Others have accomplished this by transiently promoting expression of some stem cell regulators (8) or inhibition of others. (9) Still others have shown that ECVs from cultured epicardial cells (10) or signaling from subcutaneously implanted cardiac stem cells (11) can also induce mature CM proliferation among other cardioprotective effects.

Going back even further in the development timeline, many recent papers have explored strategies to reprogram various fibroblasts into multipotent cardiac progenitor cells. These cells, when transplanted, could differentiate into CMs, smooth muscle cells, and endothelial cells at the site of an injury. The strategies include variations of GMT: GT with NKX2.5 and inhibition of GSK3 (12) and GMTH with vascular patterning and circadian regulator MEIS1. (13) One group formed multipotent CPCs using MESP1 and ETS2, a transcription factor involved in a broad range of cellular processes. They then activated β-adrenergic signaling and cultured cells in cardiac spheroids with electrical stimulation to derive highly conductive ventricular CM-like cells. (14) Lastly, one group designed a small molecule cocktail that reprogrammed fibroblasts into CPCs that could be cultured long term and differentiated both in vitro and in vivo. The cocktail included inhibitors for GSK3, TGFβ and EZH2 methyltransferase, with activators of adenylyl cyclase and histone acetyltransferase and a retinoic acid derivative. (15)

Healthy hearts contain not only contractile CMs, but also conductive pacemaker and Purkinje cells that transmit electrical impulses to regulate beating. Recent work has shown it is possible to generate these conductive cells from their contractile cousins. One group used only TBX18, the transcription factor regulating formation of the sinoatrial node, to induce pacemaker cells from neonatal rat ventricular CMs. This was improved by coculture with CFs which helped drive aerobic glycolysis and regular rhythm in the nascent pacemakers. (16) Another showed that expression of three factors, pattern formation regulator SHOX2, the pacemaker channel HCN2, and TBX5, in cardiac progenitor cells induced pacemaker formation. (17) Lastly, Prodan, et al., recently developed a strategy using a small molecule cocktail to convert human CMs into Purkinje cells, which allow rapid proliferation of action potentials from the sinoatrial node to the rest of the heart. (18)

The bulk of this review, however, concerns the direct reprogramming of non-myocytes into mature iCMs. Recent years have seen significant advances in both the efficiency of direct cardiac reprogramming strategies and the maturity of the resulting iCMs. Starting cells for iCM reprogramming have included multiple fibroblast types and non-myocyte cardiac cells from mice, rats and humans as well as human urine cells (19) and adipose-derived mesenchymal stem cells. (14) Different starting cells require variations on reprogramming factors, but broadly speaking reprogramming is currently most efficient in CFs and endothelial cells isolated from neonatal (specifically, P0-P3) mice and rats. (2) Not all the mechanisms of this variation are known, but it is likely that expression of cardiac lineage factors has a priming effect for CFs compared with, for example, MEFs or tail-tip fibroblasts. Additionally, reprogramming efficiency declines rapidly with organismal age. It was recently shown that knockdown of the TF EPAS1 could improve reprogramming specifically for adult CFs, suggesting a greater role for hypoxia and metabolic reprogramming in promoting plasticity of the adult cells. (20)

The most common cardiac reprogramming strategies largely induce non-chamber-specific CMs or ventricular CM-like cells, particularly the latter in in vivo reprogramming, with pacemaker-like cells coming in a distant third. (21) One group showed that inclusion of HAND2 in GMTH led to an increase in the proportion of pacemaker-like cells induced from CFs compared to GMT alone. (22) Much less work has targeted generation of significant proportions of atrial CM-like cells. Recently, Zhao, et al., developed Biowire II, a platform to push cells toward either ventricular or atrial fates through directed differentiation of pluripotent cells and electrical conditioning, (23) but this strategy has yet to be attempted for direct reprogramming.

1.2. Genetic regulators as key cardiac reprogramming factors

Historically, strategies for reprogramming have relied on an understanding of the factors that bring about the target cell type during organismal development. Indeed, most published strategies for direct cardiac reprogramming have relied on an understanding of CM differentiation when selecting factors to screen for transdifferentiation potential. (24) The development rationale broadly underpins all TF-based reprogramming. Introduction of just a few TFs can lead to systemic changes in gene expression, which means that introduction of TFs crucial for the normal development of CMs can drive the suppression of starter cell identity and induction of a CM gene program. The factors making up the major TF-based cardiac reprogramming strategy GMT(H) are all involved in normal CM differentiation at various developmental phases. GATA4 and MEF2C are involved in directing cardiac development from nearly the earliest differentiation of the 3 germ layers. TBX5 regulates chamber formation and atrioventricular septation, (25) and HAND2 is important in ventricular development. (26)

Many recent attempts to modify this core strategy to make cardiac reprogramming more efficient have explored a variety of TFs important to cardiac development. These include adding MYOCD (4) to GMTH, the master-regulator of CM development and the cardiac conductive system TBX20 to MGT133 (27) and the endothelial lineage regulator ETV2 to GMT. (28) The latter pushed CFs through a transient endothelial cell-like state, which is in keeping with the recent observations that CMs can derive from progenitors expressing ETV2 at low levels. (29) Additionally, it has been recently shown that specific isoforms of reprogramming factors, in this case isoform 2 of MEF2C, can lead to higher reprogramming efficiencies. (30)

At the same time, it is becoming more appreciated that reprogramming strategies need not hew so closely to development as to exclude all factors not associated with the target lineage. Some reprogramming mechanisms are common to multiple lineages. For instance, reprogramming strategies typically include a pioneer factor, which locally opens repressive chromatin to make cells more amenable to changes in gene expression. (31) GATA4 fills this role in MGT(+) reprogramming strategies, (24) but recent work has shown that ASCL1, a neuronal reprogramming factor, could replace GATA4 and TBX5, (24) and PHF7, a factor associated with spermatogenesis, replaced GATA4. (32) These factors do not even have to occur naturally. Directed evolution through site saturation mutagenesis and chimeragenesis of Oct4 and other POU family proteins led to the development of an enhanced POU protein which could more effectively reprogram cells. (33) This strategy could be applied to pioneer factors of many lineages, but Oct4 and other pluripotency reprogramming factors have also been used to create more permissive genomes for cardiac reprogramming. (8)

1.3. Epigenetic and Post-Transcriptional Regulation as a Crucial Reprogramming Mechanisms

Lineage specific chromatin patterns represent a major barrier to reprogramming. Missinato, et al., identified four TFs (ATF7IP, JUNB, SP7, and ZNF207) that maintain closed chromatin around TFs important for reprogramming toward multiple lineages. (34) Pioneer factors begin the work of undoing this repression by promoting chromatin opening at their various recognition sites. Additionally, our lab recently showed through integrated scATAC-seq and scRNA-seq that MGT operates at cis-regulatory elements to cooperatively directly open chromatin in regions related to the cardiovascular lineage and indirectly facilitate closing chromatin regions related to fibroblast lineage. (35) Ascorbic acid has also been used alongside miR-Combo to promote chromatin opening in its capacity as a cofactor for TET and H3K36 demethylase. (36,37)

A related mechanism is inhibiting factors that promote repressive chromatin or chromatin modifications leading to differentiation. One potent strategy in this vein is inhibition of polycomb repressive conplex subunits like EZH2 and BMI1 either with shRNAs (38) or small molecules. (39) Inhibition of CREB/CBP chromatin binding and/or remodeling using specific inhibitors or culture in 1% DMSO has also been shown to improve cardiac reprogramming efficiency. (40,41) Finally, KO/shRNA or small molecule inhibitors of histone (e.g., MM589, Forskolin) (15,42) or DNA methyltransferases (e.g., DMAP1-KO, 5-azacytidine, Parnate) have also been shown in recent years to promote cardiac reprogramming. (39)

MicroRNAs represent another avenue for altering gene expression by directing cofactors to specific genes or mRNAs, modulating transcription for the former and translation or degradation dynamics for the latter. The miRs that make up the miRcombo reprogramming cocktail each regulate targets related to silencing fibroblast identity and promoding CM identity. Specifically, miR-1 promotes a litany of cardioprotective phenotypes, miR-133 represses TGFβ, which is important for fibroblast activation and function, miR-208 regulates expression of the sarcomeric myosin heavy chain, and miR-499 regulates calcium signaling. (5)

2. Pathways to Modulate Cardiac Reprogramming

2.1. Classic Signaling Pathways in Cardiac Reprogramming

Epigenetic remodeling is also one of many consequences of altering cell signaling, others including metabolic transition, modulation of transcription landscape, and cell cycle regulation. Various reprogramming strategies take advantage of small molecule inhibitors of signaling pathways such as TGFβ, (15,42,43) WNT, (44) GSK3, (43) BMI1, (39) etc. This section will summarize the recent advances in the understanding of the cell signaling events involved in promoting or inhibiting cardiac reprogramming.

For cardiac reprogramming, suppression of Hippo signaling has long been known to be advantageous by interfering with YAP/TAZ interactions with both the activating SWI/SNF and repressive NuRD complexes. (45) Additionally, inhibition of Hippo signaling frees TEAD1, itself a cardiac reprogramming factor, for binding with MEF2C to promote activation of cardiac lineage transcription patterns. (19,46) Interestingly, expression of a constitutively active YAP in mature CMs promoted CM proliferation in vivo, suggestive of a less differentiated state. (7) This is in keeping with YAP/TAZ as a barrier to induction of mature CMs but may suggest a role for Hippo signaling in cardiac regenerative medicine.

Inhibition of downstream cAMP or p38/MAPK signaling (40) were shown to improve cardiac reprogramming efficiency at least in part by inhibiting CREB-CBP chromatin remodeling. Similarly, inhibition of pro-inflammatory and pro-fibrotic signaling, which prevents activation of fibroblasts to myofibroblasts, improved reprogramming efficiency. (47) Myofibroblast-specific actin isoforms have been shown to be a potent barrier to reprogramming. (48) C-C chemokine signaling was shown to inhibit cardiac reprogramming while inhibitors of CCR1, CCR4, and CCR5 improved reprogramming efficiency. (49) In contrast, agonists of the xeno-RNA sensor Rig1 improved miRCombo reprogramming efficiency. (50) Inhibition of WNT signaling has also been shown to enhance cardiac reprogramming efficiency (44) as well as to induce proliferation in postmitotic CMs. (9) Complicating this, we have recently shown that early activation of WNT signaling by knockdown of Beclin1 before day 4 of MGT reprogramming increases reprogramming efficiency. (51)

This effect was autophagy-independent, related to Beclin1’s role in endosome maturation. (52) Beclin1 has also been shown to inhibit STAT3/JAK2 interactions. (53) Additionally, Beclin1 knockdown has been shown to increase signaling of AKT1, which is used in some cardiac reprogramming cocktails. (32) AKT is a broad-specificity kinase and operates through various mechanisms in reprogramming. One intriguing possibility is its capacity to phosphorylate and inactivate translational repressor YBX1, thereby promoting translation of its target mRNAs. (54) Indeed, we have seen that YBX1 knockdown at early time points enhances cardiac reprogramming efficiency (manuscript in revision). This may represent a generalizable post-transcriptional regulatory pathway for reprogramming.

2.2. Metabolic and Prosurvival Pathways in Cardiac Reprogramming

CMs show high metabolic dependence on oxidative phosphorylation. Consequently, cardiac reprogramming involves a metabolic shift away from glycolysis and towards oxidative phosphorylation. (55) Strategies to alter metabolism and metabolic signaling remain an untapped potential resource in cardiac reprogramming (55) This section will cover new insights into metabolic interventions which should be investigated for their capacity to facilitate cardiac reprogramming.

Terminal differentiation and cell cycle exit of CMs was found to be associated with loss of PKM2 and a shift towards mitochondrial oxidative phosphorylation. (56) Indeed, several groups have recently reported that shifting metabolism towards glycolysis after cardiac injury, through malonate inhibition of SDH, (57) Neuregulin 1 promotion of ErbB2 signaling (58) or metformin treatment at higher doses (59) is sufficient to promote cell cycle entry and proliferation of terminally differentiated CMs. Among the factors that have been shown to promote cardiac reprogramming are agents known to either inhibit pro-glycolytic pathways (e.g. IGF2, (49) GSK3 and YAP/TAZ inhibitors) or promote mitochondrial oxidative phosphorylation (e.g. AKT1 (32,60) retinoic acid (61)). Whether these activities are causative during cardiac reprogramming remains to be explored.

CMs are typically characterized by higher mitochondrial fusion activity, leading to more energetic efficiency. (62) MiR-499, one fourth of miRCombo, has been shown to promote mitochondrial fusion by binding and silencing the mRNA of the master fission regulator DRP1. (63) Likewise, GSK3 inhibition, widely used in cardiac reprogramming, (15,43) has been shown to promote fusion. (60) Indeed, AKT1 is an upstream inhibitor of GSK3 and a potent promoter of fusion and mitochondrial biogenesis. (32) In our own unpublished work, we have seen that inhibition of mitochondrial fission and promotion of mitochondrial biogenesis were sufficient to improve reprogramming efficiency.

Mitochondrial clearance through mitophagy is crucial for normal mitochondrial organization during both differentiation of CPCs (64) and perinatal CM maturation. (65) The same holds for other prosurvival pathways. Cardiac stem cell differentiation is characterized in part by increases in autophagy, mediated in part by an intracellular pool of matrix metalloprotease 9. (66) We showed that knocking down the autophagy regulator ATG5 dramatically reduced reprogramming efficiency in an autophagy-dependent manner. (51) Additionally, a major function of AKT1 is mTORC1 activation, which promotes autophagy, cell growth and metabolic reprogramming. (67) On the other hand, inhibiting proapoptotic pathways by eliminating stress-induced ceramides with acid ceramidase (68) or knocking down p63 (69) has been shown to improve cardiac reprogramming.

3. Advances in Factor Delivery and In Vivo Cardiac Reprogramming

Much work has recently been done exploring alternative delivery systems, both viral and nonviral, that will be both more efficient and safer for clinical translation than lenti- and retroviruses. In the former category are Sendai, adeno and adeno-associated viruses, which do not integrate into the genome. Sendai viruses are attractive for clinical applications because of their ability to replicate in the cytoplasm, leading to higher transgene expression with a lower viral load. (25) In cardiac reprogramming, Sendai virus vectors have improved efficiency and effectively reduced infarct size and cardiac dysfunction in MI models. (70) Adenoviral vectors are comparatively resistant to mutagenesis and allow for tissue specificity. (71) They have also been used clinically as vectors for a subset of SARS-CoV2 vaccines with minimal safety concerns. Finally, adeno-associated viruses have much the same advantages of adenoviruses but much lower immunogenicity. However, they have about half the packaging capacity of Sendai and adenoviruses (~4kbp vs ~8kbp). (71)

Nonviral delivery mechanisms include techniques to deliver plasmid, mRNA, protein or RNPs to the starting cells. Groups have recently used carbon dots, (36) H peptide-modified neutrophil-mimicking membranes on mesoporous silicon nanoparticles (72) and lipoplexes (73) to improve delivery of miRCombo to the starting cells both in vitro and in vivo. Additionally, GMT has been effectively delivered nonvirally using gold nanoparticles. (74) An mRNA delivery system, which allows for more precise control of the timing of factor expression, was used to efficiently transdifferentiate human adipose-derived MSCs to iCMs using GMT or MESP1. (75) Further, using modified mRNA, such as that used in the Pfizer/BioNTech SARS-CoV2 vaccine, can further fine-tune expression level and timing. (68) Similarly, directly delivering protein to starting cells has been used to efficiently differentiate human cardiomyoblasts to beating CMs and to induce hCPCs from hADMSCs. (76)(14) Finally, advances have been made in the use of CRISPR activation (CRISPRa) to increase the expression of endogenous genes to facilitate reprogramming. The current CRISPRa cardiac reprogramming literature uses lentiviral dCas9 and guide RNAs, (12,13) but the possibility remains of using assembled RNPs allowing for nonviral delivery. (77) This strategy is particularly attractive because the actual reprogramming is performed by endogenous gene expression, but endogenously generating the necessary expression levels of reprogramming factors remains a significant barrier.

Controlling the stoichiometry of reprogramming factors can also have a profound effect on efficiency. One group recently showed that a ratio of 1/8:1:1:1:1 of viruses for GMT + MESP1 + MYOCD was optimal for reprogramming. Another way researchers have accomplished this is through development of polycistronic vectors wherein the order of genes after the promoter affects their relative expression. For GMT and GMTH reprogramming polycistronic vectors have been optimized with the order MGT and MGTH. (51,78) In keeping with this, one group recently showed that increasing MEF2C expression 30-fold by fusion with a powerful transactivating domain dramatically improved MGT reprogramming efficiency. (79) Stoichiometry is also important with nonviral delivery. The group that used modRNA for GMTH along with DN-TGFβ, DN-WNT8a, and acid ceramidase tried several stoichiometric ratios with equal or higher proportion of MEF2c or HAND2 required for optimal reprogramming. (68)

4. Remaining Barriers, Future Directions and Concluding Remarks

Despite the rapid progress the field of cardiac reprogramming has seen over the past few years, significant barriers remain for clinical translation. The efficiency of converting starting cells to mature iCMs remains low. Encouragingly, strides have already been made to improve both reprogramming efficiency and maturation. Interestingly, recent work demonstrates that soft, three-dimensional matrices, promotion of mitochondrial energetics, coculture with endothelial cells, and electrical and mechanical stimulation are effective combinatory strategies to produce CMs with mature phenotypes. (4,14,27) The field has benefited in this respect from work on maturation of PSC-derived CMs where recent advances include overexpression of SIRT3 (80) and TRPA1 (81) promoting mitochondrial oxidative phosphorylation, regular cycling of mechanical stimulation (82) and modulation of WNT, MAPK, and AKT signaling. (83) Even accounting for the reduced efficiency of reprogramming in vivo, these techniques have already shown tremendous benefits to cardiac function in animal models of myocardial infarction. (70,84) Along with recent advances in factor delivery, this suggests transition to the clinic is more a matter of time than of needing paradigm-shifting breakthroughs.

Nonetheless, we have barely scratched the surface of what is possible in regenerative medicine. One group recently identified 290 transcription factors which alone are sufficient to change cell fate. (85) Most of them had not been previously reported as reprogramming factors for any lineage, and several were likely able to induce lineages that do not match those in which the factors play a role in normal development. In cardiac reprogramming specifically, there are still promising avenues that remain un- or under-explored. Designing strategies to induce metabolic reprogramming toward oxidative phosphorylation and promote mitochondrial fusion and motility is likely to benefit reprogramming efficiency. Likewise, pro-survival and cell cycle interventions may further improve reprogramming strategies.

In conclusion, while tremendous progress has been made in the study of direct cardiac reprogramming, much work remains before it can be applied to putting a dent in the pandemic-associated rise in CVD-related mortality. Exploring both novel and well-characterized mechanisms of reprogramming and cardiac development will make our strategies smarter and more efficient, and advances in bioengineering will help us deploy these techniques more safely and on a wider scale. When it comes to the future of cardiac reprogramming, the world is our bi(cuspid)valve.

Highlights.

• Innovations to cardiac reprogramming improve efficiency and increase iCM maturation

• Strategies have been designed to induce proliferative CMs and cardiac progenitor cells

• Epigenetics, cell signaling, and metabolism are crucial reprogramming mechanisms

• Advances in factor delivery promising for cardiac reprogramming move to clinic