The cellular origin of activated fibroblasts in the infarcted and remodeling myocardium

Inflammation is intricately associated with fibrosis. Macrophages, lymphocytes and mast cells infiltrating injured tissues can acquire a fibrogenic phenotype, activating local fibroblast populations, and leading to increased deposition of extracellular matrix (ECM) proteins. In addition, some experimental studies have indicated a more direct link between inflammation and fibrosis, suggesting that leukocyte subpopulations may be capable of fibroblast conversion. The notion that fibroblasts infiltrating injured tissues may derive from circulating cells was proposed more than 150 years ago¹ and has remained an area of controversy ever since. In 1994, Bucala and co-workers² coined the term “fibrocyte” to describe a circulating, bone marrow (BM)-derived cell with features of both fibroblasts and monocytes, which has the ability to adopt a mesenchymal phenotype and contributes to scar formation. Fibrocytes are characterized through the combined expression of hematopoietic and progenitor cell markers (such as CD45 and CD34), and the production of structural ECM proteins, such as collagen. Fibrocytes represent a very small fraction of circulating leukocytes in normal subjects, but are mobilized in response to inflammation, and have been reported to infiltrate tissues in patients with fibrotic diseases.

Several studies have identified fibroblasts of BM origin in injured and failing hearts³. However, recent investigations using lineage tracing approaches suggested that populations of intracardiac cells (resident fibroblasts, pericytes, or epicardial-derived interstitial cells) are the predominant cellular source of activated fibroblasts in infarcted and remodeling myocardium^4,5,6,7,8 (Figure 1). Considering the conflicting experimental evidence, the relative contribution of circulating progenitors in the marked expansion of fibroblasts observed following myocardial injury remains unclear.

FIGURE.

Schematic cartoon illustrating the origin of activated fibroblasts in the infarcted myocardium. Although 2 independent studies using BM transplantation strategies suggested that up to a quarter of infarct fibroblasts may be derived from the BM^3,10, the bulk of the evidence suggests very low numbers of BM-derived fibroblasts in the infarcted myocardium^5,9,4,11 Lineage tracing strategies documented that intracardiac interstitial cells (Tcf21+ fibroblasts, or epicardial-derived interstitial cells) are the predominant source of infarct fibroblasts^5,9,4. Although a study investigating the contribution of Gli-1+ cells in tissue fibrosis suggested that 60% of infarct myofibroblasts are derived from perivascular cells⁸, the role of the pericytes (and the potential overlap between fibroblast and pericyte populations) remains understudied. In some studies, endothelial cells (EC) were found to be a major source of infarct fibroblasts¹⁵, whereas other investigations showed very low numbers of endothelial-derived fibroblasts in the infarct^9,4. Conflicting findings reflect the use of different approaches to trace the cells, and the absence of reliable and specific markers to identify fibroblasts.

In the current issue of the journal, Moore-Morris and co-workers⁹ used a combination of lineage tracing and BM transplantation strategies to explore the contribution of BM lineages to the population of activated fibroblasts in non-reperfused mouse infarcts. Lineage tracing studies showed that practically no infarct fibroblasts were derived from Vav+ hematopoietic cells. Experiments using chimeric mice with collagen1α1-GFP-expressing BM cells, confirmed the absence of BM-derived fibroblasts in the healing infarct. 96% of infarct fibroblasts were derived from WT1+ epicardial cells, whereas endothelial cells contributed a very small proportion (~4%). The findings provide strong support to a growing consensus on the origin of fibroblasts in myocardial disease, suggesting that the majority of fibroblasts in the remodeling heart originate from intracardiac populations, and not from leukocytes, or vascular endothelial cells.

The hard evidence on the contribution of BM-derived fibroblasts in cardiac remodeling

In many published studies, the notion that BM-derived fibroblasts contribute to cardiac fibrosis is based solely on immunohistochemical data showing expression of hematopoietic cell markers by collagen-producing cells. Considering the questionable specificity of many immunohistochemical methods, this is not sufficient. Chimeric mice with genetically labeled BM cells represent a powerful tool to study the relative role of BM-derived fibroblasts in myocardial disease. Lineage tracing studies can provide important complementary information by investigating the role of specific hematopoietic or myeloid cell lineages. Considering the lack of specific and reliable fibroblast markers, documentation of fibroblast conversion is a major challenge and often limits the conclusions of the studies. Vimentin is often used as a fibroblast marker, but lacks specificity, as it labels all mesenchymally-derived cells, Localization of α-smooth muscle actin (α-SMA) in interstitial cells is a useful marker of myofibroblast conversion, but may also label pericytes, or vascular smooth muscle cells. Despite concerns regarding its specificity⁴, expression of collagen is a reliable and functionally relevant marker of the fibroblast phenotype, reflecting the fibrogenic capacity of the cell. Supplemental table I summarizes the findings of published in vivo studies that used robust BM transplantation and/or lineage tracing approaches, coupled with appropriate strategies for fibroblast identification, in order to explore the role of BM-derived cells in myocardial fibrosis. Investigations in models of myocardial infarction have produced conflicting results. Although 2 independent studies using BM chimeras^3,10 suggested that ~24% of infarct fibroblasts are derived from the BM, other investigations using similar BM transplantation protocols found insignificant numbers of BM-derived fibroblasts in the infarct^5,11. Moreover, lineage tracing experiments found no evidence of myeloid cell-derived fibroblasts in healing myocardial infarction⁴. Using 2 independent strategies to trace BM-derived cells, and a well-characterized collagen1α1-GFP reporter line to identify fibroblasts, the current study is sufficiently robust to shape our cell biological paradigm on the origin of fibroblasts in cardiac repair. By comparison, previously published observations only used BM transplantation strategies to trace BM-derived cells, and identified fibroblasts through the expression of less-specific markers, such as α-SMA or vimentin. However, are the findings sufficiently conclusive to exclude any contribution of BM-derived fibroblasts in cardiac repair, remodeling and fibrosis?

The relative contribution of bone marrow-derived fibroblasts may depend on the pathophysiological context

Various types of cardiac injury have profoundly different effects on the cellular composition of the myocardium. In myocardial infarction, the timing of reperfusion may have dramatic effects on the fate of resident myocardial cells and on leukocyte recruitment, thus altering the relative contributions of various cell types to the expansion and activation of fibroblasts. Early reperfusion is associated with marked induction of chemokines and with accentuated and accelerated influx of leukocytes, and may also augment recruitment of BM-derived fibroblasts. On the other hand, prolonged ischemic insults may result in death of vascular and interstitial cells in the infarcted region, thus reducing their relative contribution to the expansion of myofibroblasts. In their absence, recruitment of interstitial cells from non-infarcted areas may be required to expand the fibroblast population.

Other mechanisms of myocardial injury may selectively activate specific cell types, thus altering the profile of fibroblast progenitor cells. For example, in myocarditis, induction of an intense inflammatory response may be associated with chemokine-driven recruitment of abundant BM-derived fibroblasts¹². In the current study, the surprising presence of BM-derived fibroblasts around the surgical instrumentation site is independent of myocardial infarction, and highlights the distinct cell biological responses to different types of myocardial injury. Considering that both infarction and instrumentation injury are driven by cellular necrosis and are associated with inflammation, the selective recruitment of BM-derived fibroblasts in the suture area is surprising and has no obvious mechanistic explanation.

It should also be emphasized that our knowledge on the origin of fibroblasts in cardiac injury is based almost exclusively on studies using mouse models. Information in human patients with heart disease is scarce. Sex-mismatched cardiac transplantation provides a unique opportunity to explore the potential role of circulating cells in myocardial fibrosis. A small study using endomyocardial biopsies of 7 male patients who had received a female allograft suggested that the majority of fibroblast-like cells in fibrotic regions were Y chromosome-negative¹³, providing further support to the intracardiac origin of fibroblasts. However, these findings do not exclude a significant contribution of BM-derived fibroblasts in myocardial infarction or in myocarditis, as immunosuppressive therapy following transplantation may inhibit recruitment of BM-derived cells.

Does the cellular origin of fibroblasts have functional implications?

Experimental studies have suggested that populations of intracardiac interstitial fibroblasts, perivascular cells, and (to a lesser extent) endothelial cells may contribute to the expansion of fibroblasts in infarcted and remodeling hearts. It is tempting to hypothesize that fibroblasts originating from different cellular sources may have distinct functional properties, thus explaining the wide range of functions of fibroblast populations in injury sites. Unfortunately, evidence supporting this notion is lacking. Although classification of fibroblast-like cells into subpopulations with different functional properties seems attractive and may explain the wide range of reparative and detrimental properties of fibroblasts in injured hearts¹⁴, whether distinct fibroblast profiles can be identified in vivo remains unclear.