The Clonal Evolution Model

Published Date: 02 Nov 2017

Abstract

The similarities and differences between normal stem cells and cancer stem cells (CSCs) have recently received much consideration. The capabilities of CSCs mirror that of normal stem cells, particularly in their abilities to undergo self-renewal and extensive proliferation. CSCs also have the ability to initiate and maintain tumours. Recently there has been a misconception that CSCs must arise from normal stem cells, this however does not seem to be the case. Research is beginning to shift towards understanding the cell of origin .and how it gains these tumour-initiating qualities to better design targeted therapies.

Introduction

The field of cancer stem cells (CSCs) has received much attention in recent years. The term CSC seems straightforward but the reality is far from that, with the views expressed on this topic being extremely broad. On one side, it is believed that the discovery of CSCs is a huge step towards ultimately curing cancer and understanding them will be key in achieving this. On the other hand, many question the mere existence of CSCs and their importance in tumour formation. Recent hypotheses suggest that recurrence and tumour initiation is caused by a subset of tumor cells with stem cell qualities including self-renewal, ability to differentiate and reconstitute a tumor, and resistance to chemotherapeutic drugs and radiation1. Studies have confirmed the existence of this tumorigenic subset, however their origin is unclear. Regardless of ones take on CSCs it is clear that the stem cell biology and cancer biology fields are intertwined. The capabilities of normal stem cells and tumorigenic cancer cells are similar in many ways, especially the fact that both cell types have a considerable proliferative capacity2. This review will look at the models of cancer progression, the heterogeneity of tumors, and evaluate the evidence and controversy surrounding CSCs.

Models of Cancer

One concept that is generally agreed upon is that tumours contain phenotypically and functionally heterogeneous populations of cancer cells3,4. Different models have been generated in an attempt to explain how cancers arise and progress. These models include the clonal evolution model (or stochastic model), the CSC model, and the interconversion model 5.

Clonal Evolution Model

The clonal evolution model of cancer propagation is based on genetic instability of cancer cells. Peter Nowell defined cancer as an evolutionary process that is driven by stepwise, somatic-cell mutations with sequential, subclonal selection4 (Figure 1). Clones evolve through the interaction of selectively advantageous 'driver' mutations, which inevitably promotes further 'passenger' mutations. Accumulation of these mutations leads to intratumoural genetic heterogeneity that has been observed in multiple cancers6, 7, 8. Different clones can therefore independently retain their malignant potential allowing for expansion of the tumour population. The clonal evolution model also provides a mechanistic understanding of how therapy resistance can be acquired through selectively advantageous mutations9.

CSC Model

The CSC model refers to a mechanism by which cancer cells propagate and not the cellular origin of these cancer cells10. In this model there are distinct populations of cancer cells within a given tumour, cells with the potential to initiate tumour formation and those that arise from these tumourgenic cells but are unable to initiate tumours11. One key difference between the CSC model and the clonal evolution model is that the cells following the CSC model lose their tumorigenic potential through a hierarchical mechanism of differentiation (Figure 2). That is, non-tumourgenic cells arise from tumourgenic cells that have lost their self-renewal capacities. The differentiation that occurs is irreversible which results in the non-tumourgenic cells making up the bulk of the tumour. In order to validate the CSC model these tumourgenic cells need to be different than the non-tumourgenic cells allowing for their separation, if this were not possible there would be no way of testing the theory.

Interconversion Model

The concept of interconversion between tumorigenic and non-tumorigenic cells was recently proposed. In this model a cell, which is non-tumourigenic in one context, can become tumourigenic in another (Figure 3)12. Although a new model, Pinner et at. have reported evidence in support of this interconversion 13. They showed melanoma metastasis using a novel intravital imaging method to directly analyze cells escaping from the primary tumor and entering the blood. The results reveal that disseminating cells are poorly pigmented and express high levels of a Bm2-GFP reporter. However, this phenotype can be reversed at metastatic sites providing evidence for a phenotypic change. Another example is Epithelial-mesenchymal transition (EMT), which has been widely studied in the context of embryonic morphogenesis. It also seems to have a key role in the acquisition of invasive and migratory traits by many types of carcinoma cells 14. Reversibility in EMT shows that interconversion between tumorigenic and non-tumorigenic cells could be significant in tumour progression.

Cancer Stem Cell

With all the recent information and hype surrounding CSCs many misconceptions have developed from this field. Most significantly is the idea that a CSC must arise from normal stem cells15. This idea is not expressed in any of the cancer propagation models or experimental data. The term 'cancer stem cell' was chosen because it was the most accurate description of a cell with the ability to undergo self-renewal and differentiation16. Therefore; even though normal stem cells may give rise to CSCs, the cell of origin could be any cell that develops the CSC characteristics. These tumor-forming cells could originate from stem, progenitor, or differentiated cells17. Although there is strong evidence that many tumours contain cells that display stem cell-like features, the identity of the normal cells that acquire the first genetic 'hits' that lead to tumour initiation has remained elusive18. The current evidence indicates that most cancers arise from a single cell that has undergone malignant transformation driven by frequent genetic mutations similar to the clonal evolution model19. The cells that lead to tumour initiation and development seem coordinated by cells that have stem cell-like features.

Given the similarities between tumor-initiating cells and stem cells, there has been a great interest in determining if CSCs arise from stem cells, progenitor cells, or differentiated cells. If CSCs arise from normal stem cells present in the adult tissue, de-differentiation as presented in the interconversion model would not be necessary for tumor formation. Most adult tissues are maintained over the lifetime of an organism by the self-renewal capacity of stem cells, which gives them a long time to develop tumourigenic mutations in comparison to shorter-lived differentiated cells20. The self-renewal of normal stem cells is highly regulated and therefore limits the expansion of these cells in normal tissues. Disruption in this regulation could be a key event in the development of cancer. The possibility that CSCs can originate from a population of more differentiated transit-amplifying/progenitor cells is also wide spread21. These progenitors live for a much shorter time before terminal differentiation. Thus, progenitors would likely need to gain the self-renewal capacity to have the chance to accrue the mutations that would lead to their transformation22. Similar to progenitors differentiated cells could gain the same self-renewal capabilities and de-differentiate to become more stem cell-like. Oncogenic mutations would need to drive the de-differentiation process as well as the subsequent self-renewal of the proliferating cells. De-differentiation, as shown in the interconversion model, allows for the possibility that a large population of cells could be tumorigenic. This could be the case in a mouse model of B cell acute lymphoblastic leukemia (ALL) where at least 50% of cancer cells are capable of transferring disease into wild-type recipient mice23. It seems as though tumour propagation can follow many paths and therefore there is a need to better understand the tumourinitiating cell of origin.

Cancer Cell of Origin

Identifying and understanding the cell of origin in cancer is critical to further understand cancer and ultimately design better therapies.

Hematopoietic Cancers

In different leukaemias, both normal stem cells and committed progenitor cells have been associated with cellular transformation and cancer development. Lapidot et al. provided some of the first evidence to back the CSC hypothesis when they used cell-surface protein markers to identifY a rather rare population of stem-like cells in acute myeloid leukemia (AML)24. They identified an AML-initiating cell by transplantation it into severe combined immune-deficient (SCID) mice. The cells located to the bone marrow and proliferated significantly. Limiting dilution experiments showed that approximately 1:250,000 cells, could initiate this human AML in the SCID mice. Further studies showed that these tumour-initiating cells possessed the differentiative and proliferative capacities and the potential for self-renewal expected of a leukemic stem cell (LSC)25. In a mouse model of CML, BCR-ABL expression in HSCs, but not to committed progenitor cells, induced myeloproliferative disease26. Even though the HSCs maintain the chronic phase of CML, data have shown that patients in blast crisis, the acute and advanced stage of disease, have genetic events arising in downstream progenitor cells that give rise to a population LSCs27. Huntly et al. also showed that known leukemia oncogenes could transform committed myeloid progenitor cells lacking the capacity for self-renewal. When the MOZ-TIF2 was introduced into common myeloid progenitors they resulted in AML in vivo that could be serially transplanted. In another report it was shown that LSCs couId be generated from committed progenitors through introduction of the MLL-AF9 fusion protein28. It seems as though both HSCs and committed progenitors can be transformed to develop a state of self-renewal.

Solid Cancers

Evidence from solid cancers shows that stem, progenitor, and differentiated cells may be able to initiate tumors (Figure 5). Some cancers can arise from normal stem cells through mutations that over activate the cells self-renewal mechanisms. Barker et al. showed that mutations in the adenomatous polyposis coli (APC) gene in stem cells lead to a continuously growing cancer29. When the APC gene is mutated in transit-amplifying cells or committed progenitors the growth of the tumour rapidly stops. Another group showed that the combined activation of Ras and Akt induces high-grade gliomas that appear to arise after gene transfer to neural progenitors, but not after transfer to differentiated astrocytes30. Whereas Bachoo et al. demonstrated that combined loss of p116INK4a and p19ARF with epidermal growth factor receptor (EGFR) activation could enable both astrocyte de-differentiation and neural stem cells (NSCs) to develop into high-grade giiomas31. These findings demonstrate that the combination of EGFR pathway activation and loss of both p16INK4a and p19ARF tumor suppressor function provokes a common high-grade glioma phenotype regardless of the specific cell-of-origin.

Tumour Heterogeneity

Another layer of complexity is added when one considers the vast heterogeneity within a tumour alongside the various cancer propagation models. It is well established that many tumours contain phenotypically and functionally heterogeneous cells32. Heterogeneity of cancer cells within a given tumour can arise in several ways. First, the clonal evolution model helps explain how genetic and epigenetic changes arise4,33. The cancer cell is genetically unstable and mutations accumulate within each cell that can activate various self-renewal pathways. Second, heterogeneity can arise through a cancer cells interaction with the external environments34. This tumour microenvironment can have varying effects on the tumor depending on where the cells are located (Figure 4). The spatial differences in mutated cells can have an affect on the cells properties. For example, medulloblastomas can develop either from Hedgehog (Hh) pathway activation in granule neuron precursors of the cerebellum, or from Wnt pathway signaling in dorsal brainstem progenitors35. Recent work has also shown that inflammatory cells can also promote tumours through the production of growth-stimulating proteins and DNA-damaging chemicals that can trigger cancer-causing mutation depending on the spatial location of the

cells36. A third mechanism for tumour heterogeneity is explained in the CSC model37. Tumourigenic cells undergo differentiation into non-tumourigenic cells creating a hierarchical organization that results in different pools of cells that could gain further tumourgenic mutations. Differences in mutations and cell-of-origin among different cancers and different patients will undoubtedly give rise to heterogeneous cancers. This highlights the importance of determining the cell of origin in cancers. Although there have been several markers to identify CSCs their reliability is questionable.

Identifying CSCs

The CSC hypothesis suggests that the malignancies associated with cancer originate from a small population of stem-like, tumor-initiating cells. There have been a variety of cellular markers discovered to distinguish tumourgenic from non-tumourgenic cells. However, it has proven difficult to definitively say that these markers distinguish tumorigenic from nontumorigenic cells. CD 133 appeared to be a robust marker of brain tumor stem cells in initial studies38. However, recent studies have found that this marker does not distinguish tumorigenic from non-tumorigenic cells in many brain tumors39. It has also been reported that tumorigenic ovarian cancer cells are enhanced with the CD44+ and CD44+CD117+ markers40. However, Stewart et al. were unable to find an enhanced CD44+CD1l7+ phenotype on cells in most ovarian cancers and when they did find CD44+CD117+ cells they had reduced tumorigenic activity41. Trying to identify tumourgenic cells within a tumour is likely to be an ongoing issue in many cancers because the markers used to differentiate tumorigenic from non-tumorigenic cancer cells may turn out to work in some settings but not in others, likely due to tumour heterogeneity. With uncertainties in marker strength, markers alone are likely not enough to select tumorigenic from non-tumorigenic cells.

Xenotransplantation of human cancer cells into immune compromised mice has become a standard functional assay to determine tumourgenicity. NOD/SCID mice, which are lacking functional B and T cells, are commonly used to asses tumourgenicity. In this model tumor cells are injected into NOD/SCID mice and their tumor-initiating properties are evaluated by the subsequent tumor formation. These mice however still retain natural killer cells that may be capable of rejecting some human cells42. NOD/SCID mice that are also deficient in the interleukin-2 receptor Æ´ chain (IL2RÆ´null mice) have recently been used to assess tumourgenicity. With this model Quintana et al. showed that approximately 25% of melanoma cells formed tumours. Compared with only one in a million melanoma cells forming turnours NOD/SCID mice43. This evidence suggests that tumorigenic cells could be much more frequent then

predicted by the CSC model.

Cancer Stem Cells and Therapeutics

The concept of CSCs has lead to the idea that current anti-cancer treatments can often briefly shrink tumours by targeting the tumour bulk, but these therapies fail to target and kill CSCs resulting in recurrence. The exact cell of origin in cancer is highly debated but if one ignores the debate arid defines a CSC as a cell that can undergo self-renewal and maintains tumourgenic properties, than its implications in therapy become important. If CSCs do follow the CSC model and have a hierarchical organization then it will be essential to eliminate this population of cells. However; if a cancer does not follow this model but rather follows more of an interconversion model then targeting CSCs may be unsuccessful.

Jordan states that 'the most central feature of stem cells is self-renewal, which also appears to be an intrinsic property of "successful" tumours.' With this understanding it may be essential to target the self-renewal mechanism of CSCs. It has been shown that the activation of the Wnt/β-catenin self-renewal pathway may to be the driving force behind human blast crisis LSC propagation27. As well, evidence suggests that aberrant Hedgehog (Hh) signaling is a feature that causes an expanded multiple myeloma stem cell population44. Inhibition of both the Wnt/β-catenin and Hh pathways has shown good results in eliminating this cell population. Successful future therapies may target these and other self-renewal pathways such as Notch45 to eliminate the tumourgenic populations in cancer. Therapies may also be combined with standard therapies to help de-bulk the tumor by killing of the non-tumourogenic cells. Self-renewing cells may also have an inherent resistance to therapy. For example, it was shown that the polycomb group protein BMII plays a role in radioresistance in neural stem cells through recruitment of the DNA damage response machinery46. Inhibition of BMII and other resistance inducing proteins combined with conventional therapy may provide an effective mean to target these cells. Regardless if a cell follows the hierarchical CSC model of the interconversion model it seems as though continued self-renewal is a key component of tumorgenesis.

Conclusion

The CSC concept has generated a great deal of interest because of the potential clinical implications of these cells. However, the exact origin of these cells still remains to be solved. The data indicate that CSC can arise from normal stem cells as seen with BCR-ABL expression in HSCs leading to CML26. Cancers can also follow the interconversion and clonal evolution models such as when mutations can lead to the loss of pl6INK4a and p19ARF, which enables astrocyte de-differentiation, and acquisition of self-renewal31. The tumour environment can influence cancer progression by casing damage or proliferation in response to inflammatory cells for example36. The idea that cancers arise from normal stem cells does not seem to occur in all cancers and trying to just target this stem cell population may work in some cancers but not all. Preventing self-renewal and de-differentiation of cancer cells could potentially provide benefit in some cancers. Many more years of work are required to understand the cell of origin in cancer, decipher how cancer progresses, and ultimately find better ways to treat this disease.