When the money is not in the bank

This editorial refers to ‘Cardiac mesenchymal stromal cells are a source of adipocytes in arrhythmogenic cardiomyopathy’†, by E. Sommariva et al., on page 1835.

There is an apocryphal story about the famous American bank robber ‘Slick’ Willie Sutton. When asked by a reporter, ‘Why do you rob banks?’ he replied, ‘Because that's where the money is.’ Willie's wise-crack led to Sutton's Law, or the principle that when searching for something one is well advised to look for it in the obvious place first. Thus, in developing an understanding of cardiomyopathies it stands to reason that cardiomyocytes should be the focus of disease modelling. Following Sutton's Law has indeed led investigators to multiple insights, including successful disease modelling using induced pleuripotent stem cells (iPSCs), and development of potential diagnostic tests for arrhythmogenic cardiomyopathy (ACM).^1,2 Genetic diagnosis in inherited cardiomyopathies also favours myocyte-specific disease modelling as the protein products of genes with pathogenic mutations identified in families with inherited cardiomyopathy are nearly universally important in myocyte function. Chief among these are proteins responsible for force generation and transmision, cytoskeletal structure, cell–cell adhesion, and transmembrane ion traffic.

However, as in all things, context matters. Myocytes function within a complex environment of extracellular matrix, non-contractile cardiac cells, and vascular networks. On top of this, a multitude of immune, paracrine, autocrine, and nervous system mechanisms impact structure and function on a small and large scale. It stands to reason that investigation of myocyte context can and does yield important insights into disease development even when genetic factors suggest a primary role for the myocyte.

Notwithstanding recent interest in the role of ion channels in pre-clinical arrhythmia in ACM,³ fibrofatty infiltration has long been presumed to be the primary substrate for disease expression—as regards both arrhythmic potential and functional deterioration of the ventricular myocardium. Current thinking on the biological pathways from desmosomal gene abnormality to adipogenesis is largely focused on activation of canonical WNT/β-catenin signaling through nuclear plakoglobin.¹ However, the source of adipocytes in the myocardium remains an area of intense curiosity due to its presumed role in the pathogenesis of ACM.

In a paper in the current issue of the journal, Elena Sommariva and colleagues have bypassed Sutton's Law with an important investigation of the role of cardiac mesenchymal stromal cells (C-MSCs) in the pathogenesis of ACM.⁴ C-MSCs are primitive cells originating from the mesodermal germ layer and are known to give rise to connective tissues, skeletal muscle cells, and cells of the vascular system.⁵ The exact proportion of myocardium made up by MSCs is unknown, but they are clearly present in important numbers. Compared with cadiomyocytes they are relatively easy to harvest and study, making them good targets for disease modelling. By staining adipocytes for myocyte- and MSC-specific markers, the authors convincingly demonstrate that C-MSCs are an important source of excess apidocytes in cardiac tissue of ACM patients. In order to flesh out this hypothesis, they then demonstrate that C-MSC cells from ACM hearts are more likely to progress down an adipocyte lineage pathway than those from normal subjects. Ultimately, investigating the necessary linkage between desmosome mutation and C-MSC-based adipogenesis, the authors identify expression of desmosomal genes in C-MSCs.

These data provide compelling evidence that apidogenesis in ACM is not critically myocyte dependent. This finding may prove helpful for iPSC-based modelling of ACM. In part because ex vivo human cardiomyocytes are inherently challenging to harvest and study, attention has focused on iPSC technology as a strategy for discovery of new biology, and ideally new therapies. Although, as previously noted, iPSC technology has led to profound insights into disease development, modelling with iPSCs continues to be somewhat limited by challenges in recapitulating a mature organ environment. One may suspect that incorporating C-MSCs into iPSC models by means of engineered heart tissue would be more likely to recapitulate the complex interactions that lead to disease development.

On a related note, looking at it from the reverse side, can we say that isolated C-MSCs constitute a suitable cellular substrate for mechanistic and therapeutic studies in ACM? It is worth noting that intramyocardial fat is not specific to ACM. Theoretically pathological amounts of intramyocardial fat have been observed both in non-ACM myopathies and in normal individuals with benign prognoses.⁶ Additionally, the link between development of intramyocardial fat, fibrofatty replacement, and ultimately organ dysfunction/heart failure has yet to be pinned down. With regard to this, current diagnostic criteria for ACM do not include any imaging detection of myocardial fat, and intramyocardial fat without concurrent fibrosis is not considered diagnostically useful on myocardial biopsy.⁷ Furthermore, studies have shown that arrhythmia may precede structural and histological changes.³ While fat does not appear to be part of the early arrhythmic phenotype in ACM, it may play a significant role when disease has progressed to the myopathic phase. In this sense, C-MSCs are quite likely to best serve as a component of effective modelling for ACM when combined with other factors in a complex environment, similar to iPSCs.

Shifting gears to the interaction between genetic mutation and disease expression, what are we to make of the fact that genotype-negative and PKP2 mutation-positive ACM patients are biologically indistinguishable in this study? Have the authors identified a final common pathway to development of all ACM, or have they identified the pathogenic mechanisms of a specific subtype of disease? Replication of these C-MSC findings in other genetically characterized ACM subtypes (arrhythmogenic left ventricular cardiomyopathy, non-desmosomal-based ACM) would be a useful next step to tease this clinically important question apart.

Taking these questions and thoughts into account, is there ultimately hope in translation? The pathobiological insights presented by Sommariva and colleagues broaden our understanding of the pathogenesis of ACM, and identify an additional target for screening of novel therapeutics. They also further inform our knowledge of the potential limitations and benefits of cellular disease modelling. In these ways we may expect that this study's identification of non-myocyte pathobiology will add to existing translational models that aim to shed light on diagnosis, prognosis, classification, and treatment of ACM (Figure 1).

Figure 1.

Idealized relationship between phenotype, human tissue disease modeling, and translational clinical impact. ACM, arrhythmogenic cardiomyopathy; ALVC, arrhythmogenic right ventricular cardiomyopathy; ARVC, arrhythmogenic right ventricular cardiomyopathy; C-MSC; cardiac mesenchymal stromal cell; iPS, induced pleuripotent stem; LV, left ventricle; RV, right ventricle.

This hope in translation is an exciting notion that may yet bear clinical fruit, and into which our patients count on our investment. Ultimately, paradigm-shifting disease therapy depends on our continued willingness to investigate all avenues, aiming away from the mark when necessary rather than restricting ourselves to ‘where the money is’.

Funding

This work was supported by NIH (I.D. #R21 HL126025-01).

Conflict of interest: none declared.