Research into Cancer Stem Cells

Research into Cancer Stem Cells Cancers are composed of a heterogeneous mix of cells with varying differentiation, proliferation and tumourigenic properties. In vivo studies have demonstrated that within a cancer population, only percentage of cells are able to initiate tumour development [1]. It is widely believed that the heterogeneous groups of cells include a small population of cancer cells with stem cell properties: the cancer stem cell (CSC). These cells have the capacity to self-renew and differentiate asymmetrically and give rise to bulk populations of nontumourigenic cancer cells. Current cancer treatments may eradicate the tumour bulk but spare the populations of stem cells which are able to restore tumour tissue causing recurrence of the cancer. This may explain why initial tumour regression does not necessarily translate to improved patient survival in many clinical trials. Identification and characterisation of these stem cells may offer means of targeting cancer at its root. Cancer Stem Cell Definition The AACr workshop in 2006 defined a cancer stem cell as: A cell within a tumour that possesses the capacity to self-renew and to cause the heterogeneous lineages of cancer cells that comprise the tumour. Cancer stem cells can thus only be defined experimentally by their ability to recapitulate the generation of a continuously growing tumour.[2] Therefore the stem cell definition requires that cell possess 2 fundamental properties. Self renewal, the process whereby at least one daughter cell of a dividing stem cell retains stem cell properties Potency, the ability of cells to differentiate into diverse cells that comprise the tumour. [3]. It was agreed that defined CSCs may not necessarily derive from normal tissue stem cells, indeed one important and unanswered question is whether tumours derive from organ stem cells that retain self renewal properties or whether tumour stem cells are proliferative progenitors that acquire self-renewal capacity [2]. Normal Tissue Heterogeneity The continuous replacement of differentiated, functional cells by proliferation of more primitive cells in human tissue is a normal homeostatic process. Organs are composed of collections of differentiated cells that perform discrete functions [4]. The total cell population is regarded as constituting a cell division hierarchy [5]. The stem cell is central in the renewal hierarchy and has two functions within this model. It can act as the initiating cell in a cell division and differentiation process, producing a large family of differentiated descendants, a process known clonal expansion. Another function is for the cells to undergo division to produce two stem cell daughters identical to the initial stem cell and to replace the stem cells used in clonal expansion. This process is called self-renewal [6] and is shown diagrammatically in Figure 1. As cells move down the hierarchy they acquire the differentiated features associated with tissue function and the proportion of differenti ated cells increases. In this way the stem cell has the ability to maintain organ life [4]. This concept predicts the existence of three categories of cell within the population: Proliferating, self renewing stem cells; Proliferating non-renewing transitional cells (transit amplifying); Non-proliferating, differentiated end cells. Following division the stem cell can give rise to a transit amplifying cell that will undergo further rapid proliferation to produce offspring which expand the populations of cells arising from the initial division and progressively commit irreversibly to differentiation along one or several lineages[4]. An important feature of a stem cell is their ability to undergo asymmetric cell division giving rise to a progenitor cell and to a new stem cell. Somatic SCs reside in confined tissue compartments referred to as the niche. Here the microenvironment suppresses SC proliferation, resulting in a quiescent SC population. This population maybe triggered to proli ferate and differentiate in response to injury (Ghotra, 2009). Seven common and distinguishing features of stem cells have been described [4]: Stem cells comprise a small subpopulation of a given tissue. Stem cells are ultra-structurally unspecialized, with a large nuclear-to-cytoplasmic ratio and few organelles Stem cells can be pluripotent Stem cells are slow cycling but may be induced to proliferate more rapidly in response to certain stimuli Stem cells have a proliferatve reserve that exceeds an individuals lifetime An intermendiate group of transit amplifying cells exists The microenvironment plays a critical role in the homeostasis of the stem cell and the differentiation of its progeny. The stem cell is capable of division and clonal expansion. As cells differentiate they lose their proliferative potential. The stem cell can self renewal or divide to produce proliferative transitional cells. Tumour Heterogeneity It has been recognised for many years that tumours exhibit morphologic heterogeneity but they are also functionally heterogeneous in terms of cell proliferation and tumour forming capacity based on transplantation assays [7]. Heterogeneity within tumours is seen within individual tumours in terms of morphology, cell surface markers, cell proliferation kinetics and response to therapy. In vitro and in vivo observations suggest that most cancer cells do not proliferate and that loss of capacity to divide is a feature of the tumour. Only a small proportion of cells have the ability to form tumours in vivo, referred to as tumourgenicity. The cancer stem cell theory posits that neoplasms, like physiological tissue can be hierarchically organised, and that CSCs at the apex of this of this cellular hierarchy and seem to comprise of only a subpopulation of tumour cells are essential for its initiation [8, 9]. Two models have been proposed to explain tumours heterogeneity Stochastic and Hiera rchy, summarised in Figure 2. Both models account for the existence of a cell with stem cell properties, but only the hierarchy model predicts the existence of a stem cell at the top of a hierarchy, which the potential to produce all other cell types within the tumour. Stochastic Model The stochastic model predicts that a tumour is biologically homogeneous and the behaviour of the cancer cells is influenced by intrinsic (eg signalling pathways, levels of transcription factors) or extrinsic factors (eg host factors, immune response, and microenvironment). It is suggested that the randomness and unpredictability of these factors result in heterogeneity in many aspects of marker expression and tumours initiation capacity [10]. A key requirement of the stochastic model is that all cells are equally sensitive to such influences and that the cells can revert from one state to another. For this model to be functional all tumour cells are not permanently affected and all have equal capacity to be induced to one state or another and the changes upon the cell are not permanent [11]. A growth fraction of Hierarchy Model The second model is the hierarchy model which predicts that the tumour is a caricature of normal tissue development and a hierarchy where the stem cell is at the tops is maintained (Pierce) [7]. The cancer stem cell maintains itself and its clones by self-renewal. The cells also mature to produce differentiated offspring which form the bulk of the tumour and lack stem cell properties. As in normal tissue only a small percentage of the tumour population maintain the capacity for long term proliferation while most cells proceed forward down the differentiation pathway resulting in aberrant terminal differentiation [4]. Due to differences in characteristics, stem cells can be selected and enriched for. Variations in tumour growth rates may be due the effects of normal homeostatic mechanisms that regulate stem cells and transit amplifying cell reproduction or alterations of the stem cell niche microenvironment [4]. Much of the evidence for this comes from clonogenic and tumourgenic assay s, which will be discussed further. Hierarchy model contains cells that are composed of biologically distinct cells including cancer stem cells which are all have different functional properties. The stochastic model predicts that all cells are equal the cell heterogeneity is due to intrinsic and extrinsic influences upon the cells which result in heterogeneity of cell function. Experimental Evidence Early Work The first evidence for the existence of cancer stem cells came from functional cell proliferation studies in the1940s 1960s. Radiolabelling cells and autoradiography enabled measurements into the proliferation, lifespan and hierarchical relationships in normal and neoplastic tissues [10, 12]. From these studies came the proposal that tumours are caricatures of normal development including the existence of stem cells [7]. Much early work was on the cancer of the haematopoietic system. In the 1970s Clarkson and other groups carried out pioneering studies that established cancers exhibited functional heterogeneity [10, 13]. These include cytokinetic studies carried out in cell lines, murine models of the acute leukaemias and in vivo examination of leukaemia blast proliferation kinetics in human AML and ALL patients. The data showed that the majority of leukemic blasts were post mitotic and needed to be continuously replenished from a relatively small proliferative fraction. Only a smal l number of leukemic blast cells were cycling in vivo and of these two proliferative fractions were observed: a larger, fast cycling subset with a 24 hour cell cycle time and a smaller, slow cycling, with a dormancy of weeks to months. From this data it was suggested that the slow cycling fraction was generating the fast cycling fraction thought to be the leukemic stem cell population because they had similar kinetic properties to those observed for normal haematopoietic stem cells. This was a clear suggestion that tumours exhibit functional heterogeneity in terms of proliferative potential. Following the identification of these slow cycling cells it was predicted the inability to kill the leukaemic stem cells (LSCs) was the cause of relapse and failure of chemotherapeutic therapies. Whilst combining treatment with in vivo cytokinetic studies, investigators observed that LSCs respond to the depletion of the of the leukemic cell mass by go into cycle after chemotherapy. It was sugges ted the way to eliminate dormant LSCs was to find the window when they are cycling. Identifying and assaying the potential LSCs was a major stumbling block and characterising them was impossible. This was when attention focused on the clonogenic assay was adapted by several groups to assay AML which identified phenotype of AML cultures in vitro with differing proliferative potential, providing the further proof for hierarchy in AML [14-16]. Clonogenicity Definition of a clone A clone is an operationally defined as a group of cells derived from a single ancestor cell. Clonogenicity is the ability of a given cell population, when plated as single cells, to produce one or more clones. This can be measured by the clonogenic assay which can quantify the proportion of colony forming cells, as a percentage of plated population, referred to as colony forming efficiency (CFE). It has been suggested that colony-forming cells possess two fundamental properties of progenitor cells: the ability to give rise to differentiated descendents and the capacity for self-perpetuation [17]. Therefore the ability to measure the capacity of cells to form clones is a useful tool in the study of the cancer stem cell concept. Quantitative measurement of clonogenicity Development of the clonogenic assay. Puck and Marcus The first clonogenic assay In 1956 Puck and Marcus published a paper describing a cell culture technique for assessment of colony forming ability of single mammalian cells [18]. Plated in culture dishes with a suitable medium human cervical carcinoma cells (HeLa) were supplemented with a large number of irradiated feeder cells and the number of colonies formed was counted. Their technique was a simple rapid method for growing single mammalian cells into macroscopic colonies with a colony forming efficiency of 80 100% . The authors developed this assay further to enable quantification of the effects of high energy radiation on cell populations in vitro [18-20]. They plated HeLa cells and measured their response to x-rays, producing the first in vitro radiation cell survival curve [21]. This assay has since been used for a wide variety of studies with many cell types using improved culture conditions, and for the testing of many potential chemotherapeutic agents. Till and McCulloch Following the work of Puck and Marcus, Till and McCulloch generated the first in vivo survival curves [22, 23]. They showed that when mouse bone marrow cells were injected into recipient mice that had been given total body irradiation to suppress endogenous haematopoiesis, visible colonies developed in the spleens that derived from cells in the graft. This work demonstrated that the cells injected into the mice were capable of self-renewal and it was speculated that these cells were stem cells. The evidence for this conclusion was that the curve from the number of marrow cells transplanted proportional to the number of colonies developed within the spleen. In addition, the radiation survival curve of cells that form colonies closely resembled survival curves developed by Puck and Marcus for in vitro cells [21]. This, however, was only indirect evidence and did not prove that the colonies originated from single cells, so the group carried out further experiments to determine the singl e cell origin on the colonies within the spleens [24]. Heavily irradiated bone marrow was transplanted into heavily irradiated recipient mice. The idea was that some cells containing genetic abnormalities caused by irradiation in the donor bone marrow cells would retain the ability to proliferate and produce clones containing this abnormality [24]. This worked to some extent, with a small number of colonies containing cells which all showed the same chromosome abnormality within that colony. It was hypothesised that if the capacity to form colonies is to be considered as a criterion to identify stem cells, then cells must lose this capacity upon undergoing differentiation. This hypothesis was tested by applying hypoxia as a differentiating pressure to mouse bone marrow, which resulted in a reduction in colony formation in the spleens of hypoxic mice [17]. They described how the number of colonies form in the spleens of mice in hypoxic conditions is reduced. This was thought to be du e to hypoxia stimulating erythropoiesis which stimulates erythropoietin, indicating that erythropoietin reducing colony forming production in the spleen. This data suggested that an increased demand for differentiated cells reduces the number of stem cells, resulting in the reduction of colony forming ability. Later Developments Since its development, the in vitro clonogenic assay has become a valuable tool in the study of cell growth and differentiation. [25]. Several adaptations to the original method have been made including immobilising cells in a top layer of 0.3% agar to avoid formation of tumor cell aggregates by random movement which might be confused with colony growth [26]. Agar has also been replaced by some groups with agarose, which is easier to handle (Laboise 1981) or methylcellulose which allows better recovery of the colony for replating. Others have simplified the culture medium and omitted the need for feeder cells. The exact protocol depends largely on cell type, but the basic system remains the same. The development of a protocol for secondary plating efficiency has proved a useful tool for the measurement of self-renewal and has the advantage of being able to identify cells that are able to undergo a large number of cell divisions [26]. This involves selecting specific colonies to deter mine their proliferative potential over a number of passages. Clonogenicity and Cell Renewal Hierarchy Clonogenic assays have been used to identify and morphologically characterise the three cell types above. Barrandon and Greens [27] work identified the clonal types of keratinocytes and linked this to their capacity for multiplication. They defined colonies as Holoclone, Meroclone or Paraclone. The Holoclone was described as a colony with a larger smooth nearly circular perimeter containing many small cells, which it has been suggested that these cells represent the proliferating self renewing stem cells. Paraclones were described as differentiated end cells which are more elongated and flattened in appearance, however paraclones can divide quite rapidly therefore classification of clonal type cannot be deduced form the study of growth rates alone or morphology alone. Meroclones were described as a combination of holoclones and paraclones. Relating morphology and colony size to clonogenicity can be used to further identify potential stem cells within the clonogenic assay and give mor e detail to the fate of their descendents. The differences in growth unit size may reflect several properties including different proliferative capacities and clonogenic cell kinetics. However, clonogenicity in vitro alone, does not define a stem cell, and other subpopulations, such as transit amplifying cells may also be able to produce a colony size of 32 or more cells. Although ability of a cell to form a colony implies substantial proliferative capacity, this does not unambiguously identify a stem cell [28]. Tumor Cell Heterogeneity and Hierarchy Certain characteristics have emerged from clonogenic studies on cells derived from human tumors. It was noticed that a few cells in each tumor were able to give rise to colonies in culture, whilst some colonies contained transit amplyifing cells undergoing a limited number of terminal divisions. Other cells (usually the majority) were non-proliferating stem cells. Looking at CFE and colony size of human tumors and replating experiments has demonstrated the heterogeneity of a wide range of tumor types including neoplastic human urothelium [29], melanoma [30, 31] and squamous carcinoma [32]. This supports the idea that cells within solid tumors consist of cellular hierarchies, which will be discussed further. The cancer stem cell model accounts for heterogeneity within a primary cancer by proposing that each cancer consists of a small population of cancer stem cells and a much larger population of cells which have lost their self-renewal capacity [5]. The clonogenic assay has been used explore this cellular heterogeneity present in human tumors, lending support to the stem cell model of tumor growth. Multiple myeloma has served as a valuable model in early clonogenic assay development. This was studied by Hamburger and Salmon in 1977 [33], who created an essentially selective system which restrict proliferation to cells capable of anchorage independent growth, thought to be a characteristic of stem cells [34]. They described an in vitro bioassay for human myeloma colony-forming units in culture which was applied to the study of patients with multiple myeloma and related monoclonal bone marrow derived B cell neoplasm. Bone marrow samples from patients with multiple myeloma and normal volun teers were cultured in the presence of an agar feeder layer prepared by either human type O+ washed erythrocytes or adherent spleen cells of BALB/c mice. They found a linear relationship between colony formation and the number of nucleated bone marrow cells plated. Multiple myeloma patients exhibited much higher numbers of colonies formed compared to normal volunteers. It was shown that the number of colonies was proportional to the number of colonies plated, suggesting that colonies were derived from single myeloma stem cells. This was the development of the human tumor stem cells assay. The Human Tumor Stem Cell Assay clonogenicity and cancer stem cells The ability to grow human solid tumors in two-layer soft agar culture was developed for the clinical application of testing in vitro tumor sensitivity or resistance to chemotherapeutic agents. It is a possible means by which anticancer drugs can be selected for activity against tumor cells from a patient [35] as a way of tailoring chemotherapeutic regimes to individual patients and of testing new cytostatic agents [36]. The assay assesses treatment effects of stem cells by a testing their ability to reproduce and form a colonies of cells. Using semi-solid agar with enriched medium supports colony growth from cell suspensions from a variety of human tumors. A semi-solid medium suppresses the growth of most normal cells and there is evidence of the malignant nature of these colonies [33] . An important consideration is the relationship between the response of clonogenic cells to drugs in vitro and the response of the tumor to the same drug in the patient [10]. The stem cell model of human cancer suggests that cure or duration of remission after clinical treatment should correlate only with killing of stem cells. Assessment of treatment effects on an unselected cell population (eg on the basis or morphological criteria) is therefore likely to be misleading since the effects on a small population of stem cells will be masked by those on the large population of stem cells. Human tumors of a single histological type appear to have a pattern of response in vitro that is similar to their clinical behaviour. Within a histological type, tumours are heterogeneous in response both in vitro and in vivo. Studies directly comparing the response in vitro with the subsequent clinical response have shown important correlations. The proportion of human tumors that grow with a plating efficiency sufficient for assessment of drug activity (à ¢Ã¢â‚¬ °Ã‚ ¥30 colonies per 500,000 cells plated is frequently less that 50%. Usually only a proportion of these tumors will manifest in vitro sensitivity [37]. There have been a wide range of predictive value positives reported for the human clonogenic tumour cell assay when applied to a patient population with an expected clinical response rate of 15-49% [38]. This value could be misleading and in practice may only be workable for cytotoxicity testing for only one third of specimens tested. The limitation exits that not all sam ples will produce clones in vitro so those that do may exhibit a treatment bias [35]. Other problems with the use and interpretation of human tumor clonogenic assays include low plating efficiency and small proportion of tumors available for testing; difficulty in preparing single cell suspensions, production of only small quantities of data, and problems defining drug sensitivity and response criteria [35]. Factors influencing size of sub-populations It has been proposed that as in normal cell populations, human tumor cell populations are also heterogeneous and comprise stem cells, non-stem transitional cells with limited proliferative capacity and end cells [6]. MacKillop suggested that four factors may influence the relative size of these subpopulations: The probability of self-renewal (Psr) of stem cells (producing two daughter stem cells). The distribution of cells within the system can be treated mathematically by assuming probability functions. The potential of the transitional cells for further cell division, as defined by clonal expansion number (n=number of generations between the first generation non-stem cells and the end cells.) The relative effect of cell loss on each subpopulation (Stem cells, transit amplifying, end cells) as described by cell loss factors (ÃŽÂ ¦s, ÃŽÂ ¦t ÃŽÂ ¦ ec). The number of generations of cell proliferation following initiation of the tumor cell population for individual stem cell. Stem cell division in normal tissue must provide a supply of differentiated functional cells to compensate for physiological losses and at the same time maintain a constant stem cell population. A probability of self-renewal in which two stem cells daughters Psr =0.5, would yield a steady state [28]. If no cell loss occurs, it has been modelled that the number of stem cells will increase exponentially with Psr > 0.5 [6]. For the simplest case in which all non-stem cells are end cells (n=0) the proportion of stem cells increases linearly with increasing Psr. and the proportion of stem cells in a tumor decreases as the extent of multiplication of the transitional cell compartment. This results in the stem cell being the less common cell type numerically than transit amplifying and differentiated end cells. These scenarios are affected by cell loss which may occur through necrosis, migration or differentiation, of which only differentiation is selective of cell type. A selective loss th rough differentiation increases the population of stem cells. The modelling of tumor cell growth has implications for the use of clonogenic assays as predictors of the stem cell fraction on human tumors, especially in regards to cut-off points in terms of colony size and determining which cells represent the stem cell fraction [6]. Between studies there are differences between how colonies are scored morphologically and numerically and how long cells are allowed to grow [31] and considering this evidence may be an important issue when comparing data between different studies. Clonogenicity in cell lines and stem cells in cell lines Clonogenicity has recently been used to identify stem cell properties of cells in long term culture cancer cell lines. The colony forming efficiency and secondary plating efficiency of carcinoma derived cell lines including head and neck squamous, breast [39] and prostate [39-42] were investigated and considered to contain potential stem cells. These studies show that cell lines show clear differences between clonal types (holoclone, meroclone, paraclone) and have similar properties in this respect to normal epithelial cells [39]. The proportions of clonal types between the carcinoma cell lines vary greatly. DU145 colonies were evenly spread in number between the clonal types, whereas PC3 cells produced mainly meroclones and LNCaP cells produced mainly paraclones [41], all based on colony morphology. These studies have also looked at the relationship between potential cancer stem cell markers and clonogenicity. CD133 enriched DU145 cells were assayed for clonogenicity, but no difference was found between the positive and negative cells [41], but when isolated CD44+ integrin ÃŽÂ ±2ÃŽÂ ²1+ CD133+ sorted cells were compared against CD44+ integrin ÃŽÂ ±2ÃŽÂ ²1low CD133low a higher CFE was observed in conjunction with a marked difference in morphology to CD44+ integrin ÃŽÂ ±2ÃŽÂ ²1-/low CD133- in DU145 MACS sorted cells [40]. Immunocytochemistry demonstrated that different clonal types showed varying levels of expression of CD44, ÃŽÂ ±2ÃŽÂ ²1 integrin and ÃŽÂ ²-catenin in PC-3 [42] and DU145 clones [39]. There is further evidence to suggest the presence of cells with stem cell behaviour such as dye-exclusion and higher clonogenicity, in several human epithelial cell lines [39, 43-45], which further supports the idea that cell lines contain stem cells. The ad vantage of cancer cell lines that contain cells displaying stem cell characteristics would facilitate the study of molecular pathways and the properties that define the cancer stem cells in vitro. Recent Developments Much progress has been made in the modelling of the leukemic diseases, where the level of heterogeneity was first and most thoroughly explored. Human cells fulfilling the properties expected of drug resistant cancer stem cells were initially isolated from blood cancers [2]. Improvements in the genetics of recipient mice have led to the definition SCID-repopulating cell (SRC). Many improvements to the NOD/SCID murine model continue to be made by using recipient mice that are engineered to be deficient in natural killer (NK) and macrophage activity; part of that innate immune system. It has been demonstrated that a small subpopulation of acute myeloid leukaemia cells with an immature immunophenotype possess the ability colonise immune deficient NOD/SCID mice to give rise to more differentiated leukaemia cells and to recapitulate the heterogeneous phenotype of the bulk tumour [46]. The phenotypically more mature cells failed to engraft in mice, suggesting the presence of an identifiable tumour cell hierarchy. These cells are referred to as tumour initiating cells. Cancer Stem Cell Identification CSCs have been defined on the basis of their ability to seed tumours in animal hosts, to self renew and to spawn differentiated progeny (non-CSCs)[47]. Pioneering work in this area originated from studies on leukaemia stem cells and later included demonstrations of CSCs in solid tumours, particularly breast and brain cancers. However, work in solid tumours has proved challenging. The frequency of CSCs in solid tumours is highly variable [48]. Difficulties with tumour CSC identification Evidence for the existence of cancer stem cells in solid tumours has been more difficult than in the haematopoietic system to obtain for several reasons: 1) The cells within the tumour are less accessible. Tissue has to undergo mechanical or enzymatic digestion to obtain a single cell suspension which can be analysed. 2) There is a lack of functional assays suitable for detecting and quantifying normal stem cells from many organs. 3) Only a few cell surface markers have been identified and characterised. Of these there is no one marker which is specific for a stem cell or cancer stem cells and for selection they often have to be used in combination. Cancer Stem Cell Markers Stem cells are most commonly identified by staining for cell surface markers, exclusion of fluorescent dyes or labelling with tritiated thymidine [3] . The technology to develop monoclonal antibodies to specific molecules and flow-cytometery based sorting and analysis has been a big driving force in recent CSC developments. Much work has been done to define cell surface markers. It has been shown that two distinct subpopulations can be separated from a single tumour that differ in their cell surface markers and their ability to seed new tumours in vivo. Most of the currently used markers do not recognise functional stem cell activity. By using combinations of cell surface markers, the homogenous purification of stem cells can be obtained [3]. Table 1 below reviews the current suggested markers for some tumour types. The use of animal models has allowed identification and assessment of markers that are expressed by cancer stem cells. The most convincing demonstration of identity CSC s elected by biomarkers comes from serial transplantation of cellular populations into animal models. The CSC containing fraction should re-establish the phenotypic characteristics of the original tumour [48]. In 1997 Bonnet et al showed that the ability to transfer human leukaemias into NOD/SCID mice was retained by a small proportion of cells with the CD34+, CD38- phenotype [46]. The CD44 and CD133 markers have emerged as potential markers of immature epithelial cells for isolating CSCs in several tissue types including brain and prostate. Cells have been isolated from several tumour types and serially transplanted in xenograft models: Breast CD44+ CD24-/low established tumours in recipient mice. Brain CD133+ enriched cells. Prostate Side population CD44+ enriched. In these experiments small numbers of selected cells produced tumours in recipient mice. In this instance CSCs can on