Stem Cells, Vol. 15, No. 2, 82-93,
March 1997
© 1997 AlphaMed Press
Regulation and Significance of Apoptosis in the Stem Cells of the Gastrointestinal Epithelium
C.S. Potten,
J.W. Wilson,
C. Booth
CRC Department of Epithelial Biology, Paterson Institute for Cancer Research, Christie Hospital NHS Trust, Withington, Manchester, United Kingdom
Key Words. Apoptosis • Stem cell • Intestine • Epithelium
Dr. C.S. Potten, CRC Department of Epithelial Biology, Paterson Institute for Cancer Research, Christie Hospital NHS Trust, Wilmslow Road, Withington, Manchester M20 9BX, United Kingdom.
 |
Abstract
|
|---|
In rapidly proliferating tissues the stringent control of cell proliferation and cell death by apoptosis is central to the maintenance of tissue homeostasis. In the gastrointestinal tract most work studying the control of tissue cell number has traditionally focused on the growth factor control of proliferation, and the changes that occur during carcinogenesis. However, in recent years it has become increasingly apparent that the control of apoptosis is also crucial. Apoptosis is an important mechanism for eliminating both excess normal cells and those cells which have sustained damage; therefore maintaining a tissue, i.e., stem cells with preserved DNA integrity.
In this review the incidence of apoptosis in the stem cells of both the small and large intestine will be discussed in relation to the expression of a number of apoptosis regulating genes (e.g. p53, Bcl-2, bax) within these cells. The importance of apoptosis as a means of controlling stem cell number (and therefore cellular output) will be addressed, as will the mechanisms by which any alterations to this process may contribute to malignancy.
|
Introduction
|
|---|
Apoptosis
Apoptosis is a form of programmed cell death (i.e., genetically controlled self-destruction or cell suicide) that is central to tissue morphogenesis in the developing animal and tissue homeostasis in the adult. It can be used by nature to remove unwanted but otherwise healthy cells (e.g., tissue restructuring during development), or to remove damaged single cell(s) in order to protect the embryo, an adult animal or tissue. It has also become apparent over recent years that cellular inability to undergo apoptosis may have important consequences during neoplastic transformation. Thus damaged, and possibly mutated cells, normally eliminated by apoptosis, are prevented from entering this pathway, and instead survive and may go on to divide and perpetuate or expand the mutated cells thereby increasing carcinogenic risk.
A classical apoptotic cell has a very distinctive morphology throughout the death process which is well described in the literature. Initially the cell cytoplasm shrinks and the cell becomes detached from its neighbors. The nucleolus disappears and the chromatin becomes condensed around the nuclear membrane (marginated) and consequently frequently appears as a crescent shape in histological cross section. The cell membrane takes on a "blebbed" appearance. The cell and its nucleus fragment into smaller membrane-bound vesicles which become phagocytosed by neighboring cells. This type of deliberate fragmentation and "bite size" packaging of the cell therefore eliminates the possibility of any cellular "leakage," subsequent inflammatory response and major tissue disturbance (unlike necrosis, in which physical injury destroys both organelle and cytoplasmic membrane integrity). Apoptosis is, therefore, a deliberate stepwise event that appears to be genetically controlled. It is an active sequence of events requiring the activation of specific "cell death" proteins, such as the ICE/CED3-related proteases [1]. One of the most distinctive features of apoptosis is the internucleosomal cleavage of DNA into specific-sized fragments by calcium/magnesium dependent endonucleases. These fragments of approximately 200 base pair multiples can be observed as a DNA ladder when separated by gel electrophoresis [2]. Some characteristically apoptotic cells, however, do not demonstrate internucleosomal cleavage but do show specific cleavage of DNA into larger fragments of 50 and/or 300 kbp in size [3]. The presence of such DNA fragments can be utilized in certain techniques to identify apoptotic cells. The widely used ISEL/TUNEL labeling techniques attach tagged molecules to the exposed ends of the DNA fragments in situ, and can therefore be used to support microscopic morphological identification [4]. Other cellular targets which are cleaved during apoptosis include the DNA repair enzyme poly(ADP-ribose) polymerase; cleavage of the enzyme can be demonstrated using Western blotting [5].
|
The Intestinal Epithelium
|
|---|
Tissue homeostasis depends on both cell proliferation and cell death. Hence, in a rapidly renewing tissue, such as the small intestinal epithelium, cells are lost from the villus into the gut lumen and are generally replaced at an equal rate by the proliferation of cells in the crypts of Lieberkühn. There is therefore a linear migratory pathway from crypt base to villus tip. Most studies on this renewal process have focused on the control of cell proliferation and the growth factor/signal transduction changes that may result in hyperproliferation and ultimately transformation. However, it is now becoming increasingly apparent that the control of cell death is an equally, if not more, important regulator of cell number and susceptibility to neoplastic transformation. Thus, both decreased proliferation and/or increased cell death may reduce cell number, whereas increased proliferation and/or decreased death may increase cell number.
The intestinal epithelium is an ideal model for investigating the relationship between these two processes and their effects or controls within a tissue environment. For example, the small intestine has very high levels of proliferation and cell loss, which occur in a well-defined and polarized topographical organization in which the hierarchy or cellular "age" can be assessed by the position of that cell in the tissue [6]. Due to the unique structure of the crypt and villus, it is easy to record the ability, or inability, of any cell in the hierarchy to either proliferate or die (Fig. 1
). For example, data from numerous cell kinetic and cell migration experiments have indicated that the epithelial stem cells in the small intestine are located four to five cell positions up from the crypt base, immediately above the differentiated Paneth cells, whereas in the regions of the large intestine they are located at the very base of the crypt. Thus, by assessing the incidence of cell death or DNA synthesis at these cell positions we can, by inference, determine the ability of the stem cells to undertake these options in a number of different scenarios. Similarly, frequency distributions can be generated in which the number of "events" at each position up the crypt (as seen in histological cross section, with position 1 referring to the cell at the very base of the crypt) can be measured (Fig. 1).
View larger version (14K):
[in this window]
[in a new window]
|
Figure 1. Frequency of S-phase and apoptotic cells along the length of an intestinal crypt. Top: Small Intestine; Bottom: Large Intestine. ( ______ ): Frequency of cells labeled with tritiated thymidine at each cell position (unirradiated adult mouse); (__ __ __ ): frequency of spontaneously occurring apoptosis (unirradiated adult mouse); (- - - - -): frequency of radiation-induced apoptosis (4 1/2 hours after 1 Gy  irradiation). Apoptosis can be seen to be most frequent in the small intestine at positions 4-6—thought to be the location of the stem cells (the stem cell position is indicated by the arrowhead).
|
|
|
Spontaneous Apoptosis in the Normal Intestine
|
|---|
In the normal adult small intestine a persistent low frequency of apoptosis can be seen in the crypts of healthy animals and man (Fig. 2). This we will refer to as spontaneous apoptosis. The apoptotic cells are easy to identify, although occasionally they can be difficult to distinguish from mitotic cells (which are also rounded cells containing condensed chromatin). It has been estimated that about one apoptotic cell is seen in every fifth histological longitudinal crypt section (equivalent to less than 1% of the crypt cell number), although this level varies slightly throughout the day [7, 8]. This spontaneous apoptosis occurs predominantly in the lower regions of the crypt. When measured on a positional basis, these cells appear to be restricted to the stem cell region (positions 4-5), and may reflect an incidence of about up to 10% in these cells (Fig. 1
). Similar results are seen in mice where the p53 damage response gene is deleted (see below). This suggests apoptosis may be an inherent part of the regulatory mechanism determining stem cell numbers in normal adult tissue [8, 9]. This tight control of stem cell number at the epithelial cell origin will be fundamental in maintaining a steady level of cellular output from the crypt and onto the villus, and maintaining a stable crypt size. A single extra stem cell produced by an occasional symmetrical division, by virtue of subsequent divisions of its daughter cells, has the potential to seriously perturb the crypt and dramatically increase cell production. An additional stem cell will produce 60-120 extra cells per crypt. Circumstantial evidence for apoptotic control of excess stem cell number also comes from observations of murine small intestine following irradiation [10]. After the initial damage there is a regenerative burst in proliferation which overshoots the normal steady-state level of proliferation in order to initiate mucosal repair. This also leads to an increased crypt cell number, i.e., crypt size. In order to correct this the epithelium appears to subsequently reduce the level of proliferation. This is coincidental with an observed increase in apoptotic incidence—possibly reducing cellular output by reducing stem cell numbers back to steady state levels (unpublished observation).
View larger version (130K):
[in this window]
[in a new window]
|
Figure 2. Spontaneous apoptosis in the small intestine. A: Human, formalin fixed. B: Mouse, Carnoy's fixed. The arrows indicate apoptotic cells, whereas the arrowheads indicate mitotic cells. The difficulties in distinguishing these two types of cell are clearly apparent. Pink secretory granules within the Paneth cells at the crypt base are also clearly visible in the human tissue.
|
|
Deletion of excess, but otherwise normal, cells by apoptosis may also be responsible for removing effete cells from the villus tip. The maturation, differentiation and senescence of enterocytes on the villus is a genetically programmed event which is therefore "programmed cell death." There has been some debate as to whether this is true apoptosis. However, recent evidence appears to support this idea [11, 12]. For example, one argument against apoptotic cell loss from the villus has been the apparently very rare occurrence of obviously apoptotic morphologies in this epithelium [13]. However, recent techniques which label the DNA strand breaks in situ (on tissue sections) have revealed approximately one labeled cell per villus in a 3 µm tissue section, with the frequency of incidence increasing with increasing distance from the proliferative crypt compartment [11]. When these data are corrected for villus size and the estimated rate of cell loss per villus, this incidence appears to correlate with apoptosis being responsible for controlling the majority of intestinal epithelial cell loss. Although such labeling techniques are fraught with problems, the data have now been supported by the observation that high levels of the pro-apoptotic protein, BAX, have also been detected in these terminally differentiated cells (see below) and that the expression of a possible apoptotic endonuclease (DNase I) also increases towards the villus tip [14]. Possible increased expression of transforming growth factor-ß and evidence for reduced adhesion may also lend support to this argument [15-17]. The presence of large numbers of macrophages and lymphocytes at the villus tip may also be consistent with apoptosis of terminally differentiated cells [12]. Thus, cells at the end of their functional life may activate at least part of the apoptotic process. This may be because the action of survival signals/genes may be weaker in these aged cells, permitting the expression of the "fall back" status which is apoptosis, i.e., cells may constitutively possess "death" genes and activate these when they receive, or fail to receive, the appropriate contextual signals [18].
Cell loss from the villus can also be termed anoikis ("homelessness") in which cell loss occurs via lost adhesion, which may in turn induce apoptosis. Altered adhesion signals have been shown to initiate apoptosis in epithelial tissues, either by loss of integrin function [19-21], or altered extracellular matrix composition [22-24]. It may be relevant that there is increased tenascin expression up the villus axis, an extracellular matrix component with anti-adhesive properties [16, 17]. Furthermore, the epithelial adhesion molecule E-cadherin, which is present in the enterocyte basement membrane, has been shown to moderate their death [25] (see discussion below).
If apoptosis is confirmed as the general method of cell removal, it may also explain how downwardly migrating epithelial cells (such as the small intestinal Paneth cells or the cells at the base of the stomach glands) are removed after fulfilling their differentiated function.
In the normal intestine, therefore, it appears that the ability of a healthy cell to undergo apoptosis is determined by the cell's positional (i.e., hierarchical) status and the possible changes that the cells encounter in the extracellular environment as they migrate to the villus tip. Positional influence on apoptosis can also be observed in the response to certain cytotoxics, but other studies with a range of cytotoxics have shown that all the epithelial cells are capable of undergoing apoptosis (see below).
Unlike the small intestine, spontaneous apoptosis in the proliferative zone of the large intestine is a rare event (about 10 times lower incidence). Furthermore, rather than being most prevalent in the stem cell zone, the apoptotic cells are observed scattered throughout the crypt. Thus, in the large bowel spontaneous apoptosis is unlikely to be an effective means of regulating stem cell numbers. However, like the small intestine it may be an effective means of removing senescent cells from the top of the crypts (plateau zones) at the lumenal face of the mucosa [11].
Measurements of proliferation kinetics have revealed few major differences in organization between the small and large intestine, and may indicate that proliferation per se is less relevant to cancer incidence than apoptosis (although the cell cycle in the large bowel is about two to four times longer [6]). The differences observed in positional apoptotic incidence in the small and large intestine, however, have direct implications on cancer incidence. If the small intestine is capable of removing excess stem cells, whereas the large intestine is not, the increased number of stem cells may result in hyperplastic crypts susceptible to transformation (the excess stem cells being the carcinogenesis target cells). This may therefore be one contributing factor to the observation that human colonic cancer is much more common than small intestinal cancer.
|
Damage-Induced Apoptosis
|
|---|
Cytotoxic agents can induce apoptosis in many tissues. For example, low and high linear energy transfer ionizing radiation causes a dose-dependent increase in apoptosis in the small intestinal crypt within three to six hours of exposure [7, 26]. Radiation-induced apoptosis occurs predominantly within the stem cell region—again indicating the specific sensitivity of cells at the stem cell position to apoptosis. Indeed, doses as low as 1-5 cGy (roughly equivalent to just one DNA-damaging event per cell) can induce stem cell apoptosis [7]. As the dose increases, however, more cells die, up to a dose of about 1 Gy in the small intestine where about six cells per crypt are killed. These highly sensitive cells do not appear capable of repair, as evidenced by the lack of a dose rate effect [26]. Higher doses do not induce more cells to enter apoptosis even though they are heavily damaged. Many of the damaged cells appear to prematurely differentiate and emigrate onto the villus [27]. A more extensive investigation of this phenomenon carried out by Ijiri and Potten [28, 29] involved 18 different cytotoxic agents (later supplemented by four mutagens [30, 31]) and examined the resultant apoptosis at each crypt position. It was found that each agent tended to target a specific crypt position or region but when examined collectively, the study showed that apoptosis could be induced throughout the crypt [28, 29]. Further data supporting the hypothesis that all the epithelial cells are capable of apoptosis come from crypts isolated and placed in vitro which undergo apoptosis throughout the structure [32], and the fact that villus cells can be induced to apoptose by bacterial toxins [33] or by serial elution [34]. Similar interesting data have been produced in lymphocytes lacking p53 where radiation will not induce apoptosis, whereas other cytotoxics are potent apoptosis inducers [35, 36]. These findings support the hypothesis that normally it is the in vivo ability to suppress apoptosis that is important, and distinguishes the survival ability of the different epithelial populations. This may be due to selective gene expression and also changes in the growth signals received by the cells.
In the large intestine differences are again observed when damage-induced apoptosis is considered. The apoptosis occurs at lower levels when specific low doses are studied; it is not specifically located at the stem cell position at the crypt base, and the "plateau" in apoptotic yield seen at 1 Gy in the small bowel occurs at 6-8 Gy in the mid-colon.
|
Genetic Regulation of Apoptosis in the Intestinal Epithelium
|
|---|
The reason why there is a differential susceptibility to stem cell apoptosis along the gastrointestinal (GI) tract is central to our understanding (and hence future ability to clinically manipulate) cancer incidence. This has led, therefore, to an enormous amount of research into the genes that control apoptosis. The mouse has proved to be an invaluable model in these studies and has allowed us and others to examine the effect of specific gene deletions/mutations in the whole animal.
This review will concentrate on the following gene families/groups: A) p53, which is involved in the response of cells to DNA damage; B) Bcl-2 family genes whose products are capable of both suppressing and promoting apoptosis; C) genes coding for cellular adhesion molecules, e.g., cadherin and adenomatous polyposis coli (APC) genes which are involved in controlling cell migration and position-dependent differentiation and D) the cyclooxygenase-2 (cox-2) gene which is responsible for prostaglandin synthesis.
p53 and the Bcl-2 family proteins are intimately linked as p53 has been shown to stimulate transcription of Bax [37]. We have investigated the expression of p53 and Bcl-2 family proteins in the GI tract in the mouse. Our studies show important regional differences in the expression of p53 and Bcl-2 family proteins between the large and the small intestine, which directly relate to tumor incidence in these regions of the GI tract [38, 39].
|
p53
|
|---|
p53 plays an important role in damage surveillance, and as such has been dubbed the "guardian of the genome" [40]. It is normally expressed at very low levels, except during DNA synthesis [41]. Upon DNA damage, expression is upregulated and p53 is translocated to the nucleus where it binds to DNA, acting to regulate the transcription of a number of genes including p21/waf1 [42], mdm2 [43] and Bax [37].
p53 expression has been demonstrated to be controlled at the translational level which permits a rapid response to DNA damage (due to the nondependence on transcription) [44]. It may act to cause G1 arrest, a process mediated by p21/waf-1 inhibition of cyclin dependent kinases (CDK) and proliferating cell nuclear antigen-dependent DNA replication. This allows time for the cell to make a decision on its fate (repair/apoptosis) prior to committing itself to S-phase. Entry into S-phase with damaged DNA can lead to daughter cells having unstable chromosomal structure allowing possible recombination events to occur. Alternatively stable mutations with the potential to promote malignancy may result. p53 may also act directly to initiate apoptosis. These various options may differ depending on the cell type. The response has recently been investigated by Polyak [45], who showed that cell lines which underwent p53-mediated apoptosis expressed dominant trans-acting factors which were capable of overcoming p21-dependent growth arrest. Induction of p21 by p53 has recently been shown to be codependent upon the action of another transcription factor, interferon regulatory factor 1, in embryonic fibroblasts [46]. The authors also showed the interferon regulatory factor 1 dependence of p53-mediated apoptosis in Ha-ras transformed embryonic fibroblasts. These are clearly important results and add an additional layer of complexity to p53 function.
|
p53 and Spontaneous Apoptosis in the Intestinal Epithelium
|
|---|
The p53 gene is frequently mutated in cancers, including cancers of the GI tract [47-49]. Hence, one of the initial experiments we performed was to compare the levels of spontaneous apoptosis in normal and p53 knockout (transgenic) mice. Interestingly, in both cases the levels of spontaneous apoptosis were similar, indicating that this gene has little involvement in controlling the removal of excess, but undamaged, cells [38]. This is supported by the observation that there is very little wild type p53 expressed in the normal intestinal epithelium. Also consistent with this hypothesis that spontaneous apoptosis is p53-independent is the observation that p53 null mouse embryos develop normally ([50, 51], see later discussion). Furthermore, the continued terminal differentiation and loss of cells on the villus is also p53-independent, indicating that this too is a form of spontaneous apoptosis.
|
The Bcl-2 Gene Family
|
|---|
One gene family that has been intensively studied over the last few years is the Bcl-2 family. Bcl-2 was first identified in B cell lymphoma patients, where a t14:18 chromosome translocation occurs, juxtaposing the Bcl-2 gene with the heavy chain immunoglobulin gene enhancer sequences [52, 53]. The resulting disregulated Bcl-2 expression led to enhanced B cell survival and development of lymphoma/leukemia [54, 55]. Since this initial discovery the number of Bcl-2-related proteins has increased dramatically. The family members are characterized by specific conserved domains; Bcl-2 homology regions are essential for the protein interactions with other members of this family [56]. The family can be split into two groups: those members which are involved in suppressing apoptosis, i.e., Bcl-2, BclxL [57], mcl1 [58], bag-1 [59], bfl-1 [60], brag-1 [61] and Bcl-w [62]; and those which promote apoptosis, i.e., bax [63], bad [64], BclxS [57], bak [65-67] and bik [68]. Two models have been proposed. In the first, homodimers of the pro-apoptotic protein BAX provide a dominant signal for apoptosis which can be suppressed by the competitive interaction of apoptosis-suppressing members to form heterodimers with BAX [69]. Thus, in the normal "healthy" cell the majority of the pro-apoptotic BAX would be in heterodimeric complexes with suppressors such as Bcl-2 or BclxL. In the second model, the apoptosis-suppressing proteins provide a dominant survival signal, independent of their binding to the apoptosis-promoting members of the family [70].
|
Bcl-2 and Spontaneous Apoptosis
|
|---|
Bcl-2 is commonly overexpressed in colonic adenocarcinomas. Interestingly, although there is minimal expression in the small intestine (in both mouse and humans), there is expression at the base of the colonic crypts in both species, indicating this may be involved in overriding the apoptotic homeostatic mechanism in these cells [39] (Fig. 3). This hypothesis was supported by the finding that in Bcl-2 knockout mice the incidence of spontaneous apoptosis is dramatically increased in the stem cell region of the colon but unchanged in the stem cell region of the small intestine [39].
The pro-apoptotic protein BAX, on the other hand, is strongly expressed in the Paneth cells and stem cell region of the small intestine (Fig. 4, unpublished data). We have also observed BAX expression in cells at the tip of the villus which are on the verge of being lost into the gut lumen (possibly via an apoptotic process, discussed above). BAX shows polarized expression in the colon with little or no BAX at the base of the crypts, and moderate to high levels in the cells of the table region. Overall the observations from our experiments clearly fit in with the hypothesis that the ratio of death-promoting and death-suppressing members of the Bcl-2 protein family in a given cell is important in controlling its sensitivity to apoptosis.
View larger version (62K):
[in this window]
[in a new window]
|
Figure 4. Pattern of BAX expression in murine small intestine (A), which shows greatest immunoreactivity at the base of the crypts, and colon (B), which shows greatest immunoreactivity at the lumenal surface (magnification x 100).
|
|
|
Genetic Regulation of Radiation-Induced Apoptosis
|
|---|
p53 and Radiation-Induced Apoptosis
We have examined the changes that occur in p53 expression following exposure to ionizing radiation in the small and large bowel. In the small intestine there is an increase in nuclear immunoreactivity, which occurs most strongly in the stem cell zone with weaker staining being observed in some other cells. When the strong p53 immunoreactivity and apoptosis are recorded on a positional basis, they overlay each other precisely [38], consistent with p53 either initiating apoptosis (thereby protecting the tissue) or holding the cell in G1 or G2 whilst damage is repaired. This observation also confirms the particular radiosensitivity of this cell population. However, p53 is not expressed in all apoptotic cells but is expressed strongly in neighboring cells. It is postulated that these are the more resistant clonogenic cells [71]. In contrast, after radiation damage in the large intestine the increased levels of p53 and apoptosis occur along the whole crypt in both the proliferative daughter (transit) cells and in the stem cells at the crypt base [38]. This again illustrates the limited ability of colonic stem cells to apoptose or repair damage.
Experiments conducted in this laboratory and by others [38, 72], have shown that mice lacking functional p53 are insensitive to ionizing radiation-induced apoptosis in the GI tract (measured at 4 h postirradiation). This appears to confirm that radiation-induced apoptosis is p53-dependent. However, there are apparently conflicting reports showing that radiation can induce apoptosis in colorectal tumor cell lines containing mutated p53 [73]. This paper shows that cells do not apoptose in G1, but appear to accumulate at G2/M. We have now obtained data that agree with this observation showing that delayed damage-induced apoptosis can occur in the intestinal epithelium of null mice. Following irradiation, p53 null intestinal epithelial cells seem to undergo arrest at G2/M and ultimately undergo apoptosis approximately 24 h postirradiation (unpublished data). These data are consistent with those of Hooper [74] who showed that the epithelial cells in the crypts of p53 null mice, although resistant to radiation-induced apoptosis, accumulate in G2 and exhibit reduced DNA synthesis in the first few hours after exposure. Similar events have been described in vitro in cells that lack p21/WAF1. p21/WAF1-deficient cells failed to undergo G1 arrest following DNA damage but arrested in G2. They then carried out further rounds of DNA synthesis (S-phase) without undergoing any mitoses; the resulting polyploid cells eventually died by apoptosis [75].
The Bcl-2 Family and Radiation-Induced Apoptosis
The radiosensitivity of the small intestinal and colonic stem cells correlates well with the regional expression of different Bcl-2 protein family members (Figs. 3 and 4). As described above, Bcl-2 is expressed, albeit at low levels, in the colonic stem cell region, and mice lacking functional Bcl-2 have a higher incidence of apoptosis in the stem cell region. This is significantly increased following irradiation [39]. This increased apoptosis was concentrated at the base of the crypt, supporting the theory that Bcl-2 is normally responsible for preventing apoptosis in these cells. These experiments show that Bcl-2 expression is essential for the normal survival of colonic stem cells and can inhibit apoptosis in response to DNA damage. It has been reported that BAX expression is increased in the small intestinal epithelium following exposure to ionizing radiation [76], an observation which we have recently been able to confirm (unpublished observation). BAX protein levels were found to be elevated in the small bowel within 2 h of exposure to low dose ionizing radiation [76]. This is prior to the peak levels of apoptosis which occur at about 4 h postirradiation. Again, these results support the model of Oltavi and Korsmeyer [69]. We are currently examining the interaction of Bcl-2, p53 and now Bax in a series of doubly mutant animals.
|
p53, Bcl-2 and Tumor Incidence
|
|---|
During our studies we noticed that the differential expression of Bcl-2 in the stem cell populations of the small and large bowel correlates well with the frequency of tumors observed in these regions [39]. This has now also been found to be true with regard to the expression of BAX, i.e., the sensitivity to apoptosis can govern the frequency of malignancy. We and others have also investigated the expression of p53 and the Bcl-2 family proteins in human colon tumor samples. Bcl-2 expression is associated with the onset of early stage disease. This is based on the evidence that Bcl-2 was found to be most frequently expressed in early stage, well-differentiated tumors, and increased levels of expression were observed in adjacent normal tissue. Late stage tumors that were moderately or poorly differentiated showed a greatly reduced frequency of Bcl-2 expression. Bcl-2 expression is probably reduced as tumor cells acquire other mutations which provide more dominant signals for cell survival or mutations which repress dominant signals that promote apoptosis. p53 was found to have a reciprocal pattern of expression as compared to Bcl-2, with the highest levels of expression being found in late stage disease [49]. p53 immunoreactivity observed in these studies is thought to probably be due to the expression of mutant rather than wild type protein.
Li-Fraumeni syndrome is an inherited disorder that is the result of germ line mutations in p53 [77, 78]. Sufferers show a greatly enhanced frequency of tumors (e.g., lymphomas, sarcomas and breast carcinoma) but levels of colonic tumors are not greatly increased [78]. However, it may be that Li-Fraumeni sufferers die of these other cancers before they have time to develop colonic cancers.
|
p53 Function—Protection against Teratogenesis
|
|---|
The importance of p53 in eliminating damaged cells cannot be overemphasized. Its key role in both adult and embryonic tissue protection is becoming increasingly apparent. For example, it has recently been shown that although p53 null mice develop normally (again possibly indicating this gene does not have a role in spontaneous apoptosis), p53 does appear to have a role in controlling the viability of the embryo [51]. When pregnant, normal (wild type) and p53 null mice were irradiated in order to damage the embryos. The wild type mice aborted 60% of the embryos, compared to only 10% in the null mice. However, although more embryos survive in the null mice the majority have malformations. It therefore appears that p53 normally induces the abortion of damaged embryos, probably via a cellular "proofreading" mechanism operating using apoptosis. Thus as an extension of its role as "guardian of the genome" p53 also appears to be the "guardian of the embryo." In fact, it may have evolved initially to perform this function.
|
Adhesion Molecules and Apoptosis
|
|---|
As mentioned previously, the genes which have a profound influence on the turnover of GI epithelial cells are those which code for cell surface adhesion molecules and the cytoplasmic proteins with which they interact to influence cell function. The cadherins are such proteins and are involved in cell-cell adhesion [79]. Their function has been demonstrated in an elegant series of experiments in chimeric mice expressing both wild type and mutant forms of cadherin in the GI epithelium [80, 25]. These studies showed that overexpression of E-cadherin greatly reduced cell proliferation and migration rates in small intestinal crypts and resulted in a small increase in the frequency of apoptosis in the transit cell population at the very top of the crypts. The architectural integrity of the tissue remained intact. In contrast, expression of a mutant N-cadherin, which lacks an extracellular domain, results in a complete breakdown of crypt/villus organization. This is the result of increased cell proliferation and apoptosis, failure of cells to differentiate, and increased and chaotic migration of cells from the crypts up the villi. These effects resulted in an inflammatory condition in the animals resembling Crohn's disease in humans and ultimately resulted in an increased frequency of tumors.
The E-cadherin cytoplasmic domain binds to the cytoplasmic protein ß-catenin as does the product of the APC gene. The APC gene is commonly mutated in human colorectal tumors [81, 82]. Mice bearing an equivalent mutation (multiple intestinal neoplasia [MIN]) also display an increase in GI tumors. The wild type APC gene product has been suggested to negatively regulate cell cycle progression by inhibiting CDK activity required for G1 to S transition [83]. This inhibition of CDK activity could possibly be mediated by controlling expression of p21/WAF1. In normal intestine the expression of p21/WAF1 is restricted to the nonproliferating cell populations in both the small and large intestinal epithelium [84]. This regional distribution prompted the authors to suggest that p21/WAF1 expression could be modulated by signaling pathways linked to cell surface adhesion molecules [84]. It was found that in colonic adenomas there was greatly reduced and disrupted positional expression of p21/WAF1, and the later stage carcinomas expressed virtually no p21/WAF1. As discussed previously, loss of the G1 checkpoint allows G1 cells with DNA damage to progress into S-phase. Although most should die by apoptosis, there is always the possibility of chromosomal translocations occurring or single point mutations becoming stabilized, both of which could have the potential to result in the progression of malignancy.
In studies carried out by Clarke et al. [85] p53 gene mutation failed to increase the frequency of bowel tumors in MIN mice. This resulted in the authors questioning the role of p53 as an important protective factor against carcinogenesis in the GI tract. However, if the APC protein functions by regulating the activity of proteins that are also regulated by p53, such as p21/WAF1, then the failure to observe cooperation between p53 mutations and APC mutations may not be surprising.
APC and E-cadherin clearly have an important influence on signals required for normal tissue homeostasis with a role in controlling cell proliferation and migration. The studies on these molecules suggest that cell function is position dependent and that a cell's life span is probably programmed.
|
Cyclooxygenase
|
|---|
Another gene that has recently been identified as having a possible role in colonic neoplasia is the cox-2 gene. Cyclooxygenase exists as two isoforms, COX-1 and COX-2, and is responsible for the synthesis of prostaglandins and thromboxanes from arachidonic acid. Early studies have pointed to the possible clinical usefulness of employing drugs that inhibit prostaglandin synthesis in the prevention and treatment of colonic tumors [86-89]. More recent work has shown that both human colorectal adenocarcinomas and carcinogen-induced colonic tumors in rats exhibit elevated levels of COX-2 mRNA and protein [90-93]; COX-1 expression was not altered. Furthermore, animals fed diets rich in corn oil showed elevated levels of prostaglandin and thromboxane synthesis and enhanced tumor promotion following treatment with chemical carcinogens [88]. Artificial overexpression of COX-2 in rat intestinal epithelial cells has been shown to inhibit butyrate-induced apoptosis; this was associated with increased expression of both Bcl-2 and E-cadherin which could account for the anti-apoptotic effect observed [94]. Selective inhibition of COX-2 has now been shown to inhibit tumor promotion in rats [95]. These observations (Fig. 5) indicate that the prostaglandins are functioning as autocrine survival/growth factors for colonic tumor cells. The ability of prostaglandins to protect against cell death is not solely confined to the colon. Prostaglandins have also been shown to partially protect against paracetamol-induced cell death in hepatocytes [96].
View larger version (29K):
[in this window]
[in a new window]
|
Figure 5. Schematic diagram illustrating the effects of COX-2 and the relationship with apoptosis in the large intestine. The possible effects of inhibition of this enzyme by non-steroidal anti-inflammatory drugs are also shown.
|
|
|
Conclusions and Prospects
|
|---|
It is clear that the expression of apoptotic regulatory genes is intimately linked to tumor susceptibility in the GI mucosa. In the steady-state adult tissue the high frequency of stem cell apoptosis in the small intestine (relative to the large intestine) is associated with a high BAX:Bcl-2 ratio in these cells (compared to a low BAX:Bcl-2 ratio in the large intestinal stem cell zone). Thus, these proteins appear to be involved in controlling the normal homeostatic apoptotic mechanism in the stem cells. The ability of high levels of Bcl-2 expression to suppress apoptosis is demonstrated by the fact that mice homozygously null for Bcl-2 show increased apoptosis in the colonic stem cell zone. In addition to their role in steady state equilibrium, BAX and Bcl-2 may also act to influence the susceptibility of stem cells to p53-dependent apoptosis in response to DNA damage.
Hence, in the small bowel the sensitivity of the stem cells to undergo apoptosis allows the effective removal of excess stem cells (produced by the occasional symmetrical division) and also the removal of damaged cells. Therefore, in the small intestine apoptosis has a role in both tissue homeostasis, by maintaining a stable stem cell population, and in protecting the tissue from cancer development (Fig. 6).
View larger version (38K):
[in this window]
[in a new window]
|
Figure 6. Summary of the observations and conclusions regarding apoptosis in the intestinal epithelium and their likely implications. SI: small intestine; LI: large intestine; AI: apoptotic index.
|
|
Unlike the small intestine, cancers of the large bowel are much more common. This might be partially explained by the low incidence of apoptosis in the stem cells of this region of the mucosa (Fig. 6). Any excess stem cells produced by asymmetric division may therefore persist. These may accumulate with age and lead to the formation of hyperplastic crypts with larger numbers of carcinogen target cells and hence greater susceptibility to transformation. Should the large bowel stem cells experience a DNA damaging event, the reduced ability, relative to the small bowel, to remove such cells may lead to their survival as initiated, potentially tumorigenic cells. Furthermore, the persistence of tumor cells expressing high levels of anti-apoptotic proteins (e.g., Bcl-2) will affect the ability to clinically treat such tumors—colonic tumors are notoriously chemoresistant. The current inability to effectively treat such tumors is reflected in the high lethality of the disease. Clearly methods which lead to the manipulation of genes such as p53 or Bcl-2 may have a profound clinical impact. Techniques which either biochemically alter gene expression, protein turnover/function, or which introduce genes such as p53 into the tumor site prior to therapy (transient gene therapy) could all influence the level of functional protein at the tumor site, and may help to increase the efficacy of chemotherapy to improve patient survival.
Continued dissection of the pathways controlling apoptosis will provide opportunities to understand the homeostatic control of the normal tissue and cancer incidence, and reveal novel clinical targets.
|
Acknowledgments
|
|---|
All the authors are funded by the Cancer Research Campaign.
|
References
|
|---|
-
Kumar S, Lavin MF. The ICE family of cysteine proteases as effectors of cell death. Cell Death Differ 1996;3:255-267.
-
Wyllie A. Glucocorticoid-induced thymocyte apoptosis is associated with endogenous endonuclease activation. Nature 1990;284:555-556.
-
Oberhammer F, Wilson JW, Dive C et al. Apoptotic death in epithelial cells: cleavage of DNA to 300 and/or 50 kb fragments prior to, or in the absence of, internucleosomal fragmentation. EMBO J 1993;12:3679-3684.[Medline]
-
Gravrieli Y, Sherman Y, Ben-Sasson A. Identification of programmed cell death in situ via specific labelling of nuclear DNA fragmentation. J Cell Biol 1992;119:493-501.[Abstract/Free Full Text]
-
Kaufmann SH, Desnoyers S, Ottaviano Y et al. Specific proteolytic cleavage of poly(ADP-ribose) polymerase: an early marker of chemotherapy-induced apoptosis. Cancer Res 1993;53:3976-3985.[Abstract/Free Full Text]
-
Potten CS. Structure, function and proliferative organization of the mammalian gut. In: Potten CS, Hendry JH, eds. Radiation and Gut. Amsterdam: Elsevier, 1995:1-31.
-
Potten CS. Extreme sensitivity of some intestinal crypts to X and -irradiation. Nature 1977;269:518-521.[Medline]
-
Potten CS. The significance of spontaneous and induced apoptosis in the gastrointestinal tract of mice. Cancer Metastasis Rev 1992;11:179-195.[Medline]
-
Potten CS, Loeffler M. Stem cells: attributes, cycles, spirals, pitfalls, and uncertainties: lessons for and from the crypt. Development 1990;110:1001-1020.[Abstract/Free Full Text]
-
Potten CS. A comprehensive study of the radiobiological response of the murine (BDF1) small intestine. Int J Radiat Biol 1990;58:925-973.[Medline]
-
Hall P, Coates PJ, Ansari B et al. Regulation of cell number in the mammalian gastrointestinal tract: the importance of apoptosis. J Cell Sci 1994;107:3565-3577.
-
Shibahara T, Sato N, Waguri S et al. The fate of effete epithelial cells at the villus tip of the human small intestine. Arch Histol Cytol 1995;58:205-219.[Medline]
-
Potten CS, Allen TD. Ultrastructure of cell loss in intestinal mucosa. J Ultrastruct Res 1977;60:272-277.[Medline]
-
Polzar B, Zanotti S, Stephan H et al. Distribution of deoxyribonuclease I in rat tissues and its correlation to cellular turnover and apoptosis (programmed cell death). Eur J Cell Biol 1994;64:200-210.[Medline]
-
Barnard JA, Warwick GJ, Gold LI. Localisation of transforming growth factor beta isoforms in the normal murine small intestine and colon. Gastroenterology 1993;105:67-73.[Medline]
-
Probstmeier R, Martini R, Tacke R et al. Expression of the adhesion molecules L1, N-CAM and J1/tenascin during development of the murine small intestine. Differentiation 1990;44:42-55.[Medline]
-
Probstmeier R, Martini R, Schachner M. Expression of J1/tenascin in the crypt-villus unit of adult mouse small intestine: implications for its role in epithelial cell shedding. Development 1990;109:313-321.[Abstract]
-
Raff MC, Barnes BA, Burne JF et al. Programmed cell death and the control of cell survival: lessons from the nervous system. Science 1993;262:695-700.[Abstract/Free
Top
Abstract
Introduction
