Mesenchymal stem cells: the fibroblasts’ new clothes?

This review article presents the evidence that mesenchymal stem cells and fibroblasts share much more in common than previously recognized.

Abstract

Mesenchymal stem cells are adherent stromal cells, initially isolated from the bone marrow, characterized by their ability to differentiate into mesenchymal tissues such as bone, cartilage and fat. They have also been shown to suppress immune responses in vitro. Because of these properties, mesenchymal stem cells have recently received a very high profile. Despite the dramatic benefits reported in early phase clinical trials, their functions remain poorly understood. Particularly, several questions remain concerning the origin of mesenchymal stem cells and their relationship to other stromal cells such as fibroblasts. Whereas clear gene expression signatures are imprinted in stromal cells of different anatomical origins, the anti-proliferative effects of mesenchymal stem cells and fibroblasts and their potential to differentiate appear to be common features between these two cell types. In this review, we summarize recent studies in the context of historical and often neglected stromal cell literature, and present the evidence that mesenchymal stem cells and fibroblasts share much more in common than previously recognized.

Introduction

The stem cell properties of bone marrow stroma were first described by Friedenstein in 1968¹ and subsequent experiments demonstrated their multipotent differentiation potential and immunosuppressive activity in the late nineties.^2,3 The apparently surprising immunosuppressive functions were further substantiated by reports of their activity when transfused intravenously into animal models of graft-versus-host disease (GVHD), arthritis and encephalitis^4,5,5–9 although some concerns have been raised about their immunogenicity and susceptibility to malignant transformation.^10,11 The results of early phase clinical trials with mesenchymal stem cells (MSC) in humans have been dramatic. In the first report, a nine year old boy with steroid-resistant GVHD, an invariably fatal condition, responded to intravenous infusions of haploidentical ex vivo expanded MSC¹² and in subsequent Phase I and II trials 6 out of 8 and 39 out of 55 patients with steroid-resistant GVHD responded to MSC treatment.^13,14 Although GVHD prevention in humans has been reported to be at the expense of the desirable graft versus leukemia (GVL) effect,¹⁵ this was not observed in other clinical studies in which MSC infusions were exploited to reduce stem cell graft failure and GVHD.^14,16 The potency of MSC immunotherapy in humans is certainly encouraging. However, many important scientific questions remain unanswered, especially regarding the identity of these cells in relation to fibroblasts and the physiological relevance of their immunoregulatory properties.

Mesenchymal stem cells: the fibroblasts’ new clothes?

MSC are currently defined as plastic adherent, multipotential fibroblast-like cells expressing CD73, CD105 and negative for the hematopoietic markers CD14, CD34 and CD45^17,18 but these properties and markers are also shared by fibroblasts (Table 1). Osteoblastic, chondrogenic, adipogenic differentiation from fibroblasts has also been described.^19–21 More recently, hepatocyte differentiation potential of adult human dermal fibroblasts was demonstrated in an in vivo model of liver-injured immunodeficient mice.²¹ The current definition suggested by the International Society of Cellular Therapy (ISCT) is thus incapable of distinguishing MSC from generic fibroblasts.^17,18 More recent studies have involved markers such as SSEA-1, SSEA-4 and GD2.^22–24 These studies have established a hierarchy of mesenchymal differentiation and appear encouraging. Despite these limitations, there has been widespread speculation that MSC constitute a unique cell type distinct from fibroblasts.²⁵

Table 1.

Characteristics of fibroblasts and mesenchymal stem cells.

There is also a wealth of historical data on the immunosuppressive properties of fibroblasts. In fact, it had been comprehensively demonstrated some ten years earlier that fibroblasts from various tissue sites inhibit mitogen and allo-antigen stimulated T-cell proliferation^26–29 and IFNγ production³⁰ in exactly the same vein as more recent reports using MSC.^3,31,32

MSC-mediated immunomodulation is promoted by close contact but ultimately mediated by a number of soluble factors including hepatocyte growth factor-1 (HGF-1), transforming growth factor-β (TGF-β), indoleamine 2,3-dioxygenase (IDO), prostaglandin-E2 (PGE₂) nitric oxide and insulin-like growth factor (IGF) binding proteins.^20,33–38 Similarly, PGE2 and IDO have also been implicated in fibroblast-mediated T-cell suppression.^20,26,27 Furthermore, both MSC and fibroblast suppressive effects are enhanced in the presence of inflammatory cytokines such as IFNγ and TNFα.^{27,28,30–39} Pre-treatment of human fibroblasts and MSC with IFNγ and TNFα up-regulates MHC Class II molecule expression but both cell types have poor capacity to activate allo-responses.^27,40 Different culture conditions, experimental kinetics, species and cell populations used in the in vitro assays may account for the variety of soluble factors identified as responsible for fibroblast and MSC-mediated suppression but may also reflect a redundancy or pleiotropy in the mechanisms employed by these cells. However, nearly all studies suggest that an inflammatory microenvironment is a prerequisite for observing stromal-mediated suppressive effects.⁴¹

MSC-mediated inhibition of monocyte differentiation into dendritic cells^42,43 has also been previously documented using fibroblasts.⁴⁴ This effect is dependent on interleukin 6 (IL-6)^44,45 and involves cell cycle arrest.⁴⁶ More recently, direct comparison between adult fibroblasts from various tissues and bone marrow MSC showed similar in vitro immunosuppressive potency.^20,41,47 Both MSC and fibroblasts induce cell cycle arrest, prevent apoptosis and support the survival of T cells.^41,48 Although this could be a fundamental process to maintain memory T cells, it may have a negative effect when MSC are used in the clinical setting leading to the preservation of pathogenic memory T cells with future adverse consequences.

Both fibroblasts and MSC may be isolated using tissue culture adherence from many tissue sites including adipose tissue, placenta, skin, thymus, periosteum, muscle, synovium, synovial fluid, fetal liver and blood, and cord blood.^49–51 Bone marrow-derived MSC and fibroblasts from various anatomical sites have been shown to have distinct gene expression profiles⁵² (Collin M, unpublished data). However, it is also well recognized that fibroblasts from different tissues possess site-specific molecular identity and topographical memory due to differential expression of homeobox (HOX) genes.⁵³ Although subtle and interesting niche-specific differences may exist between stromal cells, there is no evidence that these alter the general immunosuppressive and differentiation properties that have been described for these cells.

The naked fibroblast

Fibroblasts exist in virtually every organ in the human body. They are defined as adherent cells, which are not endothelium, epithelium or hematopoietic in origin, and which have the capacity to synthesize and remodel the extracellular matrix. In addition to their presumed role as scaffolding support, fibroblasts have been directly shown to play roles in regulating self-tolerance, organ development, wound healing, inflammation and fibrosis (Figure 1).^54–57

Figure 1.

Stromal cells modulate diverse biological processes. Stromal cells are actively involved in all of the above processes although subsets with specialized functions within the heterogeneous site-specific population remain to be defined

Central and peripheral immunological tolerance

Fibroblasts have at least two recognized supportive roles in central tolerance. Firstly, thymic fibroblasts support the proliferation of thymic epithelial cells through the release of FGF-1, FGF-7 and FGF-10.^58,59 Secondly, they are directly involved in the recruitment of early T-cell precursors⁶⁰ and migration of developing T cells through the thymic medulla and cortex.⁵⁹ Furthermore, a recent study has elegantly described a startling role for lymph node stroma in maintaining peripheral tolerance. Antigen presentation by lymph node stromal cells was shown to be functionally similar to medullary thymic epithelial cells leading to active deletion of self-reactive peripheral T cells.⁶¹ Marrow stromal cells are crucial for B-cell development^62,63 and more recent studies have shown that MSC under particular circumstances, can promote the survival of B cells⁶⁴ and stimulate B-cell antibody production.^65,66

Wound healing and tissue repair

Tissue injury and wounding are accompanied by changes in the extracellular matrix, mechanical stress and inflammation in the surrounding microenvironment. These changes result in the activation of fibroblasts which express contractile bundles and α-smooth muscle actin and differentiate into myofibroblasts.

Myofibroblasts participate in wound healing through migration, proliferation and contraction necessary to restore homeostasis in damaged tissue. Return to normal physiology requires resolution of the inflammation accompanying the injury,⁵⁷ a process traditionally thought to occur passively from the fizzling out of inflammatory signals. However, current evidence clearly demonstrates the importance of the local stromal network in mediating active inflammatory cell clearance.⁶⁷

Tissue fibrosis

Inappropriate tissue repair and continued insult can result in chronic inflammation and eventually lead to fibrosis. At the cellular level, accumulation and persistence of myofibroblasts during tissue repair and healing has been proposed as a leading cause of fibrosis.⁶⁸ This process is associated with the transformation of granulation tissue into a hypertrophic scar with excessive production of ECM and rarification of the microvasculature. Fibrosis is modulated by a dynamic ‘leuco-stromal interaction’, a notion supported by the observation that carbon tetrachloride-mediated liver fibrosis is reduced in immunodeficient rag^−/− mice following liver injury⁶⁹ and after selective macrophage depletion during advanced liver fibrosis.⁷⁰ Recently, myofibroblasts in fibrotic tissue have been shown to acquire resistance to Fas-induced apoptosis by T lymphocytes,⁷¹ a process that normally accompanies tissue repair. In addition, fibrosis-related pathological myofibroblasts promote their own survival by expressing Fas molecules and killing surrounding lymphocytes.⁷¹

Tumor survival and metastases

The protective and facilitative role of stroma in tumor growth was first described by pathologists as desmoplasia, a typical feature of many solid tumors.⁷² Tumor stroma is predominantly comprised of myofibroblasts, often referred to as carcinoma associated fibroblasts (CAF).⁷³ The restrictive role of myofibroblasts in wound healing is taken over by growing tumors. Breast tumors with a wound-response gene signature are associated with an increased risk of progression and metastases.⁷⁴ Injection of a mixture of CAF or MSC with breast cancer cells into immunocompromised mice showed that both stromal cell types were capable of accelerating cancer growth and invasiveness.^75,76 CAF provide nutritional support by the secretion of growth factors, promoting neoangiogenesis and ECM remodeling to facilitate tumor invasion and metastasis.⁷⁷ The distribution of tumor metastases is also not random showing clear organ preferences for the various cancer types for certain receptive stromal environments.⁷⁸ Recently, mutations and loss of heterozygosity in the tumor suppressor gene TP53 in the stromal compartment adjacent to breast carcinoma was found to be associated with lymph node metastases presenting a compelling case for stroma-facilitated cancer progression.⁷⁹ However, other animal studies have reported anti-tumor effects with bone marrow and skin-derived stromal cells and this may be related to the ability of the tumor to recruit and activate different stromal cell functions.^80–82

Parenchymal and stem cell regulation

The functional diversity and positional identity of fibroblasts may act to regulate local parenchymal cells in several ways. Firstly, fibroblasts could act as a source of growth factors such as fibroblast growth factor (FGF), keratinocyte growth factor (KGF) and leukemia inhibitory factor (LIF) for cell survival; a property that has been exploited in the laboratory with the use of fibroblast feeder layers to expand parenchymal and stem cells.⁸³ Secondly, fibroblasts may provide a co-ordinate system of positional reference points for site-specific epithelial-mesenchymal interactions critical for the development, differentiation, patterning and renewal of the adjacent epithelia such as in the skin, lung, gastrointestinal, genitourinary systems and the thymus.⁸⁴ In addition, stromal cells also provide the appropriate niche for stem cell maintenance and differentiation. One of the best-studied examples of stem cell-niche regulation is the orchestration of HSC survival and differentiation by bone marrow stroma.⁸⁵ However, stromal cells are heterogenous with specialized niche functions confined to particular subsets. This was recently demonstrated in vivo where only CD146 expressing bone marrow stromal cells were found to be capable of conferring a hematopoietic microenvironment when transplanted to heterotopic sites.⁸⁶

Future speculation and conclusion

The plethora of recent studies on MSC has to some extent recapitulated what had been previously described over ten years ago for fibroblasts. Present definitions of MSC and fibroblasts emphasize generic properties of these cells and fail to distinguish subsets of stromal cells with specialized niche functions. The lack of appropriate markers means we are currently unable to functionally dissect the important differences within the extended fibroblast family. Are there common mesenchymal progenitors throughout the body or are these progenitors specialized and site specific? What is clear is the ubiquitous presence and functional heterogeneity of fibroblasts.

The physiological significance of stromal cell immunoregulation has also been poorly recognized despite the overwhelming evidence for their varied role in maintaining immune equilibrium and in pathology. The consequences of these findings emphasize the need to recognize the common ground between the fields of fibroblast and MSC biology in order to redefine and dissect the complex family of stromal cells.