Stem cell plasticity and tumour formation

Abstract

Stem cell plasticity refers to the ability of certain stem cells to switch lineage determination and generate unexpected cell types. This review applies largely to bone marrow cells (BMCs), which appear to contribute positively to the regeneration of several damaged non-haematopoietic tissues. This beneficial effect on regeneration may be a direct result of BMCs giving rise to organ parenchymal cells. Alternatively, it could be due to BMCs fusing with existing parenchymal cells, or providing paracrine growth factor support, or contributing to neovascularisation. In the context of oncology, BMC derivation of the tumour stroma and vasculature has profound biological and therapeutic implications, and there are several examples of carcinomas seemingly being derived from BMCs.

Introduction

Morbidity and mortality as a result of failing vital organs plagues even the most technologically advanced societies. Because of a dearth of transplantable organs there is a growing hope that stem cells may lead to the possibility of replacing tissues affected by age or disease. Indeed, it is almost impossible to open a newspaper today without seeing yet another apparent ‘breakthrough’ in stem cell research. Most adult tissues have multipotential stem cells; cells capable of producing a limited range of differentiated cell lineages appropriate to their location, e.g. small intestinal stem cells can produce all four indigenous lineages (Paneth, goblet, absorptive columnar and enteroendocrine), central nervous system (CNS) stem cells have trilineage potential generating neurones, oligodendrocytes and astrocytes,¹ whereas the recently discovered stem cells of the heart can give rise to cardiomyocytes, endothelial cells and smooth muscle.² However, describing tissue-based stem cells as ‘multipotential’ may be incorrect if, as it appears, some adult stem cells, when removed from their usual location can transdifferentiate into cells that arise from any of the three germ layers (so-called plasticity).

All tissues have stem cells, though in some tissues, notably brain and heart, they do not appear to be activated sufficiently adequately to replace damaged cells.

The resurgence of interest in stem cells has reaped dividends in terms of how we understand other diseases. Metaplastic and heterotopic changes from one recognisable tissue phenotype to another are well known in histopathology and are mostly seen in tissues with a high turnover of cells; such changes may result from genetic or epigenetic changes that affect expression of transcription factors, presumably in stem cells. For example, overexpression of the transcription factor Cdx2 targeted to the gastric epithelium, which does not normally express Cdx2, results in islands of intestinal metaplasia,³ conversely the absence of Cdx2 expression in cdx2 null: wild type chimaeric mice results in patches of Cdx2 null gastric phenotype within wild type colonic mucosa⁴; importantly the junctional epithelium had the phenotype of small intestinal mucosa, despite being of wild-type heritage, and so their local stem cell units had adopted a specific relevant program of differentiation appropriate to their location.

Myofibroblasts are a distinguishing feature of pathological fibrosis, historically regarded as having originated by the activation of local parenchymal fibroblasts, and being the primary collagen-producing cells. However, such fundamental concepts will have to be reconsidered in the light of recent findings that bone marrow-derived cells contribute to fibrogenesis in both pulmonary⁵ and hepatic scarring.⁶ Moreover, bone marrow-derived cells are at least in part responsible for the tumour desmoplastic response⁷ (see Fig. 6). Thus, bone marrow may provide a platform for the delivery of anticancer agents.

Since the classic ‘initiation–promotion’ experiments involving painting carcinogens on mouse skin,⁸ it has been apparent that many cancers, particularly those of continually renewing tissues (blood, gut, skin), are in fact a disease of stem cells. These are the only cells that persist in the tissues for a sufficient length of time to acquire the requisite number of genetic changes for malignant development.⁹ In fact, tumours are heterogeneous populations in which many cells are terminally differentiated (reproductively sterile) or transit amplifying cells with limited division potential, and so it seems that only tumour stem cells are capable of ‘transferring the disease’. For example, in human acute myeloid leukaemia, only the CD34⁺ CD38⁻ cells are capable of propagating the disease in immunodeficient NOD/SCID mice,¹⁰ while in human breast cancer, the CD44⁺ ESA⁺CD24^−/low fraction has a similar potential.¹¹ In the central nervous system (CNS), CD133 appears to be expressed on those cells with the greatest clonogenic potential in vitro,¹² and these CD133-positive cells are the ones that give rise to further medulloblastomas in NOD/SCID mice.¹³ Therefore, there is a growing conviction that successful cancer chemotherapy depends upon eradicating all the stem cells within a cancer. This review focuses on bone marrow stem cell plasticity, how bone marrow could be the origin of some carcinomas, and how other cells of bone marrow origin, notably myofibroblasts and endothelial progenitor cells (EPCs) could be exploited therapeutically.

Stem cells feature prominently in disease processes; metaplasia illustrates stem cell plasticity, the bone marrow is a source of fibrogenic and endothelial progenitor cells, normal stem cells are the likely carcinogen targets, and cancers themselves probably all have stem cells.