Undeniable evidence that the adult mammalian heart lacks an endogenous regenerative stem cell

Background

Over the past two decades, reports of adult stem cells serving as mediators of cardiac regeneration of infarcted tissue have been a source of great hope in cardiovascular medicine. One of the most highly touted adult stem cell types was defined by expression of the cell surface receptor c-Kit, which in bone marrow marks the known hematopoietic progenitors that give rise to all immune and blood cells. With respect to the heart, publications in the early 2000s showed that injection of c-Kit⁺ cells, either bone marrow-derived or derived from within the heart itself, were able to regenerate up to ~70% of the infarcted rodent myocardium when injected into the injured area^1,2. While independent laboratories were unable to replicate these early findings^3–5, the field nonetheless moved quickly to the clinic with an array of human trials, starting with bone marrow mononuclear cells and followed thereafter with expanded autologous cardiac stem cells or mesenchymal stem cells from various sources. Collective evidence from these trials has demonstrated that “adult stem cell” injection is largely safe in patients, but that efficacy remains unclear with some trials failing to meet primary endpoints⁵. Importantly, none of these injected cell-types have demonstrated an ability to engraft and directly generate new myocardium in clinical applications, and today the preponderance of recent evidence in rodent models fails to demonstrate physiologically relevant new cardiomyocyte generation with injection of any sort of presumed adult stem cell^6,7.

The assertion that the heart contains its own dedicated cardiomyocyte producing stem cell is absolutely central to the entire cardiac regeneration - stem cell debate. If the heart truly lacks such an endogenous adult stem cell, it calls into question previous studies in which cardiac regeneration was reported, such as from c-Kit⁺ cells isolated from the heart, selected, expanded, and reinjected back into the circulation where it then homed to the heart and created new myocardium⁶. Indeed, the FDA recently summarized the state of affairs surrounding stem cell therapy for indications such as heart failure as; “Outside the setting of hematopoietic reconstitution and a few other well-established indications, the assertion that stem cells are intrinsically able to sense the environment into which they are introduced and address whatever functions require replacement or repair — whether injured knee cartilage or a neurologic deficit — is not based on scientific evidence”⁸.

In our opinion, the more recent and rigorous genetic approaches in mouse models^9–12, such as the current paper in this issue of Circulation by Bin Zhou and colleagues¹³, strongly demonstrate that the adult heart lacks a predetermined cardiomyocyte producing stem cell. For example, to directly assess if c-Kit⁺ cells are endogenous cardiac stem cells, multiple independent laboratories generated Kit locus genetically targeted mice in which both constitutive and inducible Cre recombinase-dependent lineage tracing was performed in the adult heart at baseline and with injury^9–11. The results of these studies showed that cardiomyocyte generation from a c-Kit⁺ cell lineage was an extremely rare event (<0.005%) that was several orders of magnitude below the known physiological turnover rate for cardiomyocytes¹⁴. However, advocates of the c-Kit stem cell hypothesis raised concerns over the technical caveats of lineage tracing, specifically whether Kit allele heterozygosity used in these approaches might hinder myogenic potential, or whether lineage tracing mouse models are unable to mark very low c-Kit⁺ expressing cells that have more “stemness”¹⁵. In this issue of Circulation, Bin Zhou and colleagues provide definitive evidence with genetic lineage tracing that is not dependent on the Kit allele, or any other allele that presumably marks a stem cell population. Instead, they employed a dual Cre-Dre lineage tracing genetic approach in mice that conclusively shows that the adult heart does not generate new cardiomyocytes from non-myocyte sources, whether c-Kit expressing or otherwise¹³.

Summary of the results

More specifically, the laboratory of Bin Zhou at the Shanghai Institute of Biochemistry and Cell Biology previously showed that traditional Cre-recombinase driven lineage tracing systems can lead to misidentification of Kit derived cardiomyocytes due to ectopic expression in pre-existing cardiomyocytes¹¹. A similar concern arises with testing any putative cardiac stem cell candidate that relies on a specific marker gene for locus-dependent lineage tracing (e.g. Sca1, Abcg2, Isl1, BMi1, etc.). To address this limitation, Bin Zhou’s group employed an elegant dual genetic recombinase system showing that Kit-derived cardiomyocytes are never observed when the possibility of ectopic Kit allele expression is eliminated¹². In the group’s most recent work published in this issue of Circulation¹³, this finding is taken a step further, utilizing the same dual recombinase lineage tracing system with a series of “interleaved genetic reporters.” Using this more refined system the authors first label all cardiomyocytes permanently, followed by a second wave of recombination induced in all other cell-types, which showed that in no case were cardiomyocytes derived from non-cardiomyocytes in the adult mouse heart at baseline or with injury¹³.

The results are exceptionally rigorous not only because of the sophisticated genetic experimental design, but also because of the high redundancy of experiments and convincing positive controls in both developmental and adult settings. To achieve multiple independent approaches to pre-label all cardiomyocytes, the authors utilized two different cardiomyocyte promoters (Tnnt2-Dre or Tnnt2-Cre, and Tnni3-Dre) in various combinations with their ubiquitous non-myocyte drivers that were based on targeting of the ubiquitous Rosa26 locus or the β-actin promoter (R26-iCre, R26-DreER, and Actb-Cre), which showed remarkable reproducibility with both interleaved reporters. More importantly, the differentiation of non-cardiomyocytes into cardiomyocytes was readily observed during early heart development, and in adult skeletal muscle myogenic differentiation from non-myocytes. Even with less than 100% efficiency of Rosa26 reporter allele recombination driven by the various Cre or Dre lines, there is no biologic reason why the minor unlabeled population would selectively exclude all potential cardiomyogenic stem cells as the entire system is essentially “blind” to the labeling of non-myocytes. Lastly, nine different inducible Cre lines were employed, several of which were generated for this manuscript, again demonstrating that cardiac cell-types including endocardial cells, mesenchymal stromal cells, resident and activated fibroblasts, endothelial cells, pericytes, smooth muscle cells and epicardial cells are incapable of producing new cardiomyocytes in the adult heart¹³.

Ramifications and conclusions

The primary biologic ramification of the genetic evidence provided by Bin Zhou and colleagues¹³ is that the adult heart lacks an endogenous cardiomyocyte producing stem cell. This conclusion may not be a surprise to many cardiac biologists and cardiologists, as it is fairly well-known that the heart is one of the least regenerative tissues in the adult mammal, with estimates of approximately 1% cardiomyocyte nuclear DNA synthesis events per year, with true cellular division likely occurring at an even smaller frequency¹⁴. However, the even greater ramification is that we should now reinterpret a portion of the previous cardiac literature in this area, such as those reports claiming that cardiac-derived c-Kit⁺ cells broadly regenerate the heart with new cardiomyocytes^2,15,16. Moreover, it is highly doubtful that another putative adult stem cell-type, such as mesenchymal stromal cells (MSCs), adipocyte progenitors, fractionated bone marrow cells, etc., would be inherently more cardiogenic or reparative to the heart.

The final and perhaps more controversial ramification of the work of Bin Zhou and colleagues¹³ is that the field needs to potentially re-evaluate the current cadre of ongoing and planned clinical trials with cardiac cell transplantation, as it is clear that we no longer understand the true mechanistic basis for this approach. It is also difficult to optimize such clinical trials for success if we lack a comprehensive biologic understanding of how cell therapy works, which is especially germane in light of recent negative results from ongoing trials⁵. For example, the results of Bin Zhou and colleagues suggest that harvesting autologous c-Kit⁺ cells from the human heart, or harvesting any other presumed progenitor cell-type for clinical trials is of no greater mechanistic benefit in promoting cardiac rejuvenation compared with using cells derived from non-cardiac sources. Indeed, animal studies and some clinical trials have shown a reparative benefit from cellular transplantation approaches from an array of seemingly unrelated cell-types from many different sources^6,7,8. Thus, the potential “rejuvenating” benefit underlying cell therapy needs to be more rigorously investigated so that clinical trials can be better designed and optimized for efficacy in moving forward. Recent emerging hypotheses are that the injected/infused progenitor cells serve as a temporary source of paracrine factors that can have a reconditioning effect on the newly injured heart, or that the injected/infused cells induce an acute inflammatory response as these cells die and are cleared, resulting in a secondary acute healing response. However, the results from Bin Zhou and colleagues, in conjunction with other recent genetic approaches such as Kit allele lineage tracing^9–13, unequivocally demonstrate that cell therapy is not creating new cardiomyocytes, suggesting that we need to expand scientific efforts in the laboratory to continue elucidating the underlying mechanistic basis for cardiac cell therapy.