The all-chemical approach: A solution for converting fibroblasts into myocytes

Abstract

Converting cardiac fibroblasts to cardiomyocytes has been considered as a regenerative strategy for myocardial infarction and other disorders. Recently, Cao et al. 2016 defined a cocktail of 9 chemical compounds with this capability, increasing the likelihood of clinical success.

It has taken a while, but the concept that the terminally differentiated state is immutably stable no longer dominates modern biology. In 1938, Hans Spemann contemplated a “fantastical experiment” in which transfer of an egg nucleus could redirect a recipient somatic cell to become pluripotent, and John Gurdon made this a reality in 1958 by converting gut epithelial cells into whole frogs ¹. Inspired by Gurdon and others, Shinya Yamanaka developed the iPSC technology where four transcription factors sufficed in generating pluripotent stem cells ². The last decade has witnessed an explosion in cell fate manipulations. Many somatic cell types were interconverted by the means of transcription factors, epigenetic and cellular pathway modulators. As one of the most difficult cells to regenerate, cardiomyocyte (CM) was successfully generated from fibroblasts by introducing cocktails of transcriptions factors ^3-6. In a recent paper in Science, Cao et al. report the successful reprograming of functional cardiomyocytes by small molecules and growth factors ⁷.

Cao et al. devised a three-step protocol. Human foreskin fibroblasts (HFFs) or human fetal lung fibroblasts (HLFs) were first treated with nine compounds (9C: CHIR99021, A83-01, BIX01294, AS8351, SC1, Y27632, OAC2, SU16F, and JNJ10198409) that were the outcome of multiple rounds of focused screening. Next, the cells were incubated with a cardiac induction medium (CIM) containing previously known cardiogenic factors Activin A, BMP4, VEGF, and CHIR99021. Finally, human cardiomyocyte-conditioned medium was used to aid maturation. At day 30, 6.6 ± 0.4% cells in the culture were cardiac troponin T (cTnT) positive, a remarkable improvement in efficiency over transcription factor-mediated reprogramming. The chemically converted cells (ciCMs) in many ways resembled CMs derived from human pluripotent stem cells (hPSC-CMs). They displayed sarcomeric structures, epigenetic signatures, transcriptomes, and electrophysiological properties comparable to early stage cardiomyocytes.

This all-chemical approach represents a major advance for cardiac regeneration. From a cell therapy perspective, it is appealing since it circumvents inadvertent genomic modifications that can occur with genetic methods, and the chemical compounds themselves are non-immunogenic. Other potential advantages include less variability, tighter dose control and greater cost-effectiveness if translated clinically. From a chemical biology perspective, a deeper understanding of the protein targets of the compounds, and the signals they evoke constitute a new angle on understanding cardiac differentiation and reprogramming mechanisms.

Mechanism of cardiac specification

9C appeared to operate exclusively in the first step of reprogramming. Treatment led to decondensation of closed chromatin regions, which was apparent by a decrease in H3K9me3, HP1γ, H3K27me3 marks, and an increase in H3K4me3 and H3K27ac marks. These changes correlated with induction of genes that mark embryonic mesoderm, including EOMES, T, MESP1 and KDR that, in the context of an embryo, are expressed prior to the appearance of cardiac cells. A more detailed characterization of the cells induced by 9C will answer whether the cocktail directs fibroblasts to a progenitor state, such as the transient extraembryonic endoderm (XEN)-like cells that appear during chemical reprogramming toward iPSCs ⁸. Another important question is whether 9C narrowly specifies cardiac fate, or makes cells capable of following multiple lineages, such as mesendoderm progenitors created by genetic reprogramming with MESP1 ⁹. Such breadth would make 9C broadly important for tissue regeneration.

The efficiency of reprogramming was low (about 7%), raising the question of whether 9C acts deterministically, but only a few cells in the starting population are competent to respond, versus whether directed differentiation is stochastic and carries a low probability of overall success. In either case, defining the cell types that are competent to respond will also be important for translation.

Mechanisms of chemical reprogramming

All-chemical cocktails have been reported to convert somatic cell into other cell types, including pluripotent stem cells, neurons, and pancreatic islet cells ^10-13. Unlike TF-mediated reprogramming, the components of chemical cocktails often overlap substantially, though the destination cell types differ greatly. Accordingly, the molecules used by Cao et al. have been used before. CHIR99021 is a GSK3β inhibitor that activates the canonical Wnt pathway and is present in most cell fate conversion cocktails leading to iPSCs, neurons, and now, cardiomyocytes ^{7, 10-12}. A83-01 is an ALK4/5/7 inhibitor that blocks the Activin/Nodal branch of the TGFβ pathway. Blocking A83-01 and other ALK inhibitors (RepSox, SB431542) have also been used in iPSC, neuron and pancreatic islet cell reprogramming ^{10, 12, 13} and likely lead to chromatin remodeling conducive for the cardiac muscle program of gene expression ¹⁴. Similarly, BIX01294 (a G9a histone methyltransferase inhibitor) and other epigenetic modulators are common components of reprogramming cocktails ^{8, 10, 12, 13}. A direct hypothesis is that these common pathway/epigenetic modulators lead fibroblasts into a highly plastic state, which is poised to receive further instructions to activate the cardiomyogenic program. For practical applications, it is immensely important to define the conditions to exclude unwanted cell conversions, i.e. teratomas, especially for therapeutic reprogramming in vivo where dose, timing, and stoichiometry will be hard to control.

The lingering challenge of cardiomyocyte maturation

After exposing HFFs sequentially to 9C and CIM, Cao et al. injected the cells into infarcted hearts of immunodeficient mice, where the cells expressed CM markers and showed organized sarcomeric structures after two weeks, demonstrating that 9C-treated cells are capable of forming cardiomyocytes in vivo. The ciCMs lacked an organized sarcoplasmic reticulum (SR) and Transverse (T)-tubule structures typical of adult cardiomyocytes, suggesting electrical and mechanical immaturity. Thus, the ciCMs appear to be similar to hPSC-CMs, which resemble fetal far more than adult myocytes. In the absence of CASQ2, KCNAB1, CAV3, KCNJ2, KCNJ8, S100A1, KCND3, SLN and CACNA1, the cells are likely to have poor calcium handling. Although physiological function was not fully characterized, their immature morphology suggests far less force generation than adult myocytes, poor calcium-induced calcium release, immature adrenergic responsiveness, automaticity and pro-arrhythmic potential. Thus, we anticipate a shift of emphasis to establishing the chemical conditions for cardiomyocyte maturation in future endeavors, possibly by inducing or mimicking the function of cardiomyogenic transcription factors.

Figure 1. Cardiac reprogramming.

In a major advance, Cai et al. discovered a cocktail of 9 small molecules (1) that reprogrammed fibroblasts to cardiovascular progenitors in a process involving chromatin remodeling. Subsequent differentiation (2) to immature cardiomyocytes required factors from human cardiomyocytes. How to coax these cells form other clinically useful cells (e.g. endothelial cells) and to fully functional cardiomyocytes remains a major challenge (4).