J Oastroenterol 1997; 32:414423

Iournal of

Gastroenterology

9

Springer-Verlag 1997

Review

The role of apoptosis in intestinal disease ALASTAIR J.M. WATSON Department of Medicine, Hope Hospital, University of Manchester, Salford, M6 8HD, UK

Introduction Apoptosis is a stereotypic form of programmed cell death,--defined on morphological criteria,--that is under tight genetic control. The concept of a programmed cell death has been familiar for over 100 years to embryologists who observed cell death of certain cell types occurring at specific sites and at particular times during embryogenesis. However, this concept was not formalized until 1965, when Richard Lockshin coined the term "programmed cell death"/ Following observations of cell death in a variety of pathological states, Kerr, Wyllie, and Currie 2 coined the term "apoptosis" in 1972 to describe a "programmed cell death" characterized by some specific morphological features found on electron microscopy. The word "apoptosis", which is derived from ancient Greek and means "leaves falling in autumn," invokes the idea of death that is required for life. The two concepts of "programmed cell death" and "apoptosis" have now come together because of the discovery of a number of genes which determine whether or not a cell undergoes apoptosis after application of an appropriate stimulus. While all examples of "apoptosis" are "programmed" or genetically regulated, there are examples of "programmed cell death", particularly in invertebrates and other lower organisms, which do not display the morphological features of apoptosis. 3 Unlike necrosis, which occurs only in pathological situations, apoptosis is a normal and essential part of growth and development that acts to counterbalance cell division and remove unwanted or damaged cells. 4 It is therefore not surprising that inappropriate regulation of apoptosis can lead to disease. For example, excessive apoptosis is part of the pathogenesis of

Offprint requests to: A.J.M, Watson (Received Aug. 14, 1996; accepted Oct. 25, 1996)

neurodegenerative disease and AIDS, whereas failure of apoptosis can allow the development of cancerr Apoptosis in the gastrointestinal tract is a new area of investigation. Fundamental questions remain unanswered, and there are also significant methodological difficulties that inhibit the investigation of apoptosis in gastrointestinal disease. However, it is already clear that abnormal regulation of apoptosis contributes to a number of gastrointestinal diseases, most notably cancer. 6 Moreover, increasing understanding of the regulation of apoptosis in the gastrointestinal tract is allowing the development of novel therapeutic strategies for cancer and other disease s t a t e s / I n this article the role of apoptosis in the homeostasis of normal intestinal epithelium will be described, after which the role of apoptosis in the pathogenesis of intestinal disease and its treatment will be discussed. Only passing references will be made to apoptosis in the upper gastrointestinal tract, as this has not yet been studied systematically. On the other halad, there is a great deal of information on the relationship between apoptosis and liver disease. For reasons of space this will not be included (for review seeS). However, before apoptosis in the gastrointestinal tract is discussed it is important that the general morphological features of apoptosis are described.

General feature of apoptosis In apoptosis the cell first detaches from its neighbors and shrinks. 9 The nucleus shrinks as well and, under electron microscopy, the nuclear chromatin can be seen to condense around the edge of the nucleus. Cell surface features, such as microvilli, are lost, and blebs of plasma membrane develop? This is a very dynamic process which, under time-lapse video microscopy, gives the cell the appearance of "boiling".1~ Finally the cell breaks apart into apoptotic bodies which undergo phagocytosis by macrophages or neighboring epithelial cells. II This

A.J.M. Watson: Apoptosis and intestinal disease can be a rapid process, with a single cell proceeding from normal morphology to fragmentation in 15 min. Apoptosis presents a pattern of cell death distinct from that of necrosis. Necrotic cells swell then burst because of loss of integrity of the plasma membrane early in the necrotic process, with consequent failure of intracellular ionic homeostasis, t2 The escape of intracellular contents in necrosis invariably provokes an inflammatory response, which results in further damage to neighboring healthy cells. Plasma membrane integrity is maintained until very late in the apoptotic process. This is a critical feature of apoptosis, since it prevents the escape of intracellular contents. As inflammation is rarely seen as a consequence of apoptosis it is likely that plasma membrane integrity is maintained until after phagocytosis. Prevention of the premature rupture of the apoptotic cell is assisted by the activation of yglutamyltransferase, which forms cross-links between membrane proteins, forming protein shells that are insoluble in detergent? 3 The mechanisms that attract a n d facilitate the action of phagocytes are not fully understood, but redistribution of plasma membrane phosphatidylserine to the outer leaflet of the plasma membrane bilayer may be an indicator of apoptosis to surrounding cells. This redistribution of lipid within the leaflets of the plasma membrane appears to be widespread in apoptosis and can be regulated by two genes which are known to inhibit apoptosis, bcl-2 and abl. ~4 Phagocytes, such as bone marrow and blood-derived macrophages, express specialized cell surface molecules, such as thrombospondin/ctv~3/CD36 complex, which aids in the recognition of apoptotic neutrophils. ~5 Another feature differentiating apoptosis from necrosis is that apoptosis occurs asynchronously, whereas necrosis occurs in large populations of cell simultaneously. These features, together with its rapid time course, explain why apoptosis went unnoticed for so tong.

415 acridine orange or Hoechst 33342, and counting the number of cells with condensed, brightly fluorescent nuclei. Another feature found in many examples of apoptosis is DNA fragmentation, in which DNA is cleaved between the nucleosomes into fragments that are multiples of 200 basepairs in size. This is now known to be a late feature of apoptosis. Initially DNA is digested, yielding 300-kb rosettes or 30- to 50-kb loops, before cleavage into smaller fragments.17 This oligonucleosomal cleavage of DNA (often referred to as DNA "ladders") can be easily demonstrated by gel electrophoresis of DNA extracted from cell culture models of apoptosis. 18 However, it is often difficult to obtain DNA "ladders" from tissue samples, as the apoptotic cells form only a tiny fraction of total cells. To circumvent this problem a number of investigators have used an end-labelling technique in which strand breaks of fragmented DNA are labelled with a colored marker such as dUTP-biotin. 19 This type of technique (often called TUNEL or ISEL) can be used in histological sections and can, in principle, identify single apoptotic cells. However, it has the important disadvantage of not giving information on the size of the DNA fragments identified. Unfortunately, it has also proved to be unreliable in gastrointestinal tissue.2~ Finally, flow cytometry can be a useful method of quantifying apoptosis in cell suspensions. One method is to stain DNA with a fluorescent DNA dye such as Hoechst 33342 and perform a cell cycle analysis. Apoptotic cells appear as a subdiploid peak to the left of cells in G0/G1 phase of the cell cycle. Other more sophisticated flow cytometric methods have been developed which do not require cell fixation prior to analysis.2~

Apoptosis in normal intestinal epithelium Methods for studying apoptosis Apoptosis can be identified by examination of cellular morphology, detection of non-random DNA fragmentation, or by flow cytometric techniques. The best method remains electron microscopy, as this is the only way to unequivocally identify the defining morphological changes of apoptosis. However, in standard histological sections stained with hematoxylin and eosin, apoptotic bodies can be identified and distinguished from mitotic cells as round, densely staining eosinophilic bodies, separated from adjacent cells. This technique has proved very useful in the study of apoptosis in the gastrointestinal tract. 16If the cells under study are in suspension, apoptosis can be identified from their morphology by staining with fluorescent DNA dyes, such as

When studying apoptosis in mouse intestinal epithelium it is helpful to describe the cell position of the apoptotic cell along the crypt/villus axis (Fig. 1). A numbering system has been developed in which the epithelial cell at the base of the crypt is designated "1" and the cells are then numbered up the sides of the crypt. 22 In the small intestine, spontaneous apoptosis has a maximum incidence at cell positions 4-6, which are the same positions as those of the stem cells. This implies that approximately 10% of the stem cells are undergoing apoptosis at any one time. However, as a percentage of all crypt cells, apoptosis is rare (<1%); being seen in only one of every five to ten crypt sections. In the colon, the stem cells are located lower down in the crypt, at cell positions 1 and 2, while apoptosis is at the higher cell positions 4-6. 23The function of apoptosis at the base of the

416

A.J.M. Watson: Apoptosis and intestinal disease Tip of villus/colonic surface

Normal morphology

Initiation

~

Apoptotic morphology

Detachment from neighbours ~

Shedding into lumen

~[- . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Non-random DNA fragmentation?

Small intestinal crypt

p53 > colon bax++

11 o

Colonic crypt

bcl-2 p53
7

Stem cells

67 Stem cells

Maximumm u t a g e n inducedapoptosis

J

/

3 1

2

11

3 1

2

crypt is, presumably, to regulate the number of cells entering the crypt/villus axis and to ensure the integrity of the stem cells by deleting those with mutations (see below). 24 Crypt cells migrate up to the tip of the villus where, in the mouse, they are shed at a rate of 1400 cells/villus per/day. The process by which cells are shed has been a subject of controversy. The rapid loss of cells, together with the exquisite control of the dimensions of the villus, suggest that some form of programmed cell death accounts for the cell-shedding process. Careful examination of the tip of the villus by conventional electron microscopy reveals that apoptosis is a very rare event; certainly not enough to account for the rapid cell shedding observed. More recently, studies in which cells with fragmented DNA were identified by the T U N E L technique claimed that the majority of cells of the upper portions of the villus were apoptotic. 19 This result was almost certainly erroneous, as careful studies using the T U N E L technique in which false positive staining was minimized demonstrated single cells at the villus tip, with fragmented D N A (and therefore presumably apoptotic). 25New electron microscopy studies appear to have resolved the argument. By careful fixation so that loosely attached cells are not lost during preparation of the specimen, it has been demonstrated in the human small intestine that, at the villus tip, epithelial cells first shrink to form dome-like profiles, leaving large vacuoles underneath. A lymphocyte often migrates into this

Maximummutagen inducedapoptosis

Fig. 1. Comparison of apoptosis at various sites in intestinal epithelium

vacuole. Then chromatin condensation develops in the epithelial cell just before it is shed into the lumen. The barrier function of the epithelium is maintained by neighboring cells extending processes under the shed cell and forming tight junctions with its neighbors. The authors claim that this process has not been observed before, as the dome-like epithelial cells are lost during preparation. 26 The resolution of the argument appears to lie in the order of events (Fig. 1). Cells shrink and loosen from their neighbors while still histologically normal and do not develop chromatin condensation until the point of cell shedding, at which time they are lost during standard histological preparation. The T U N E L results imply that D N A fragmentation occurs while the cell is still histologically normal. It will be helpful if the T U N E L results can be confirmed by some other technique which can identify the early stages of apoptosis unambiguously. The factors regulating spontaneous apoptosis along the crypt/villus axis are poorly understood. Bcl-2, a gene which functions to suppress apoptosis, is expressed only at the base of the colonic crypt and is barely expressed in the small intestine. 27Bax, another member of the bcl2 gene family, accelerates apoptosis and is expressed most strongly at the surface of colonic epithelium and base of small intestinal crypt epithelium. 2sRecently bak, another member of the bcl-2 family which is proapoptotic, has been found to be expressed strongly by epithelial cells of the villus but only weakly in the

A.J.M. Watson: Apoptosis and intestinal disease crypt. 29 The apoptotic potential of individual cells is likely to be determined by the relative expression of all family members. Other factors contribute to the regulation of apoptosis. For example, many investigators of the gastrointestinal tract have noticed that when intestinal epithelial cells are isolated they tend to die. All epithelial cells in human colonic crypts isolated by calcium chelation undergo apoptosis within 4h. 3~ This process can be delayed by culturing isolated crypts in a collagen matrix containing [31-integrin, highlighting the importance of integrin/matrix interactions in the prevention of apoptosis. Cells at the base and top of the crypt are far more sensitive to loss of cell/ matrix interactions, and apoptosis at these sites cannot be blocked by [31-integrin. However, it is not known whether cell/matrix interactions actually regulate spontaneous apoptosis in intact tissue, though it is interesting that expression of ~7B[31 integrin is absent at the base of crypts and the upper third of the villus. 31Studies in transgenic mice have shown that cell/cell contacts via cadherin-catenin complexes are also required to prevent apoptosis in epithelial cells of the villus and crypt. 32 In summary, there are important differences between apoptosis in the crypt and the villus tip. In the crypt the apoptotic cells loosen from their neighbors and undergo phagocytosis in situ. At the villus tip, although the apoptotic cells loosen from their neighbors the barrier function of the epithelium is maintained. 33 Rather than undergoing phagocytosis, the cells are shed. This suggests that apoptosis at the villus tip is a specialized process and not typical of the generality of apoptosis. Further work is required in this area.

Gastrointestinal diseases with excessive apoptosis Apoptosis plays an important role in a wide range of gastrointestinal diseases. Broadly speaking, these can be divided into diseases where there is excessive or inappropriate apoptosis and diseases where apoptosis is defective. Induction of apoptosis in target cells occurs by a variety of mechanisms and is an intrinsic part of the inflammatory response. 34 Activated neutrophils kill microorganisms by the generation of oxygen free radicals. These reactive species induce apoptosis in intestinal epithelial cells at low concentration, where cellular damage is mildY At high concentration, cellular membranes break down rapidly through lipid peroxidation reactions, and necrosis results. 36 Secondly, cytotoxic T cells induce apoptosis in target cells by two distinct pathways. 37 First, by the secretion of perforin, a poreforming protein, which is inserted into the plasma membrane of the target cell and allows granzyme B to pass

417 through it into the target cell. 3s The mechanism by which granzyme B induces apoptosis is not fully understood. However, it has been demonstrated that it activates CPP32, one of the interleukin-l[3-converting enzyme (ICE)-like cysteine proteases, 39 and a novel 55-kDa protein, designated FLICE, which has homology to both FADD and the ICE/CED-3 family of cysteine proteases, which are part of the effector mechanisms of apoptosis. 4~Secondly, ligation of CD95 ligand on T cells to CD95 (Fas/APO1) on target cells will induce apoptosis. 41 Like granzyme B, CD95/ CD95L ligation activates ICE-like proteases, but through a different pathway which is initiated by the generation of ceramide through the activation of an acidic sphingomyelinase. 42 Thus, it is expected that apoptosis of both target epithelial cells and lymphocytes will be found in inflammatory reactions in the gastrointestinal tract. This is indeed the case, and apoptosis has been found in conditions as diverse as Helicobacter pylori-induced gastritis, 43,44 inflammatory bowel disease, 45 celiac disease 46 and graftversus-host disease. 47 It should be emphasised that much of this data is preliminary and requires confirmation. There is, however, little evidence as yet that the regulation of apoptosis is defective or that apoptosis is playing an important role in initiation of these conditions. However, given the importance of apoptosis in normal immunoregulation, one can speculate that abnormalities in apoptosis could give rise to inflammatory bowel disease. Lamina propria T lymphocytes (T-LPL) are constantly exposed to antigen stimulation, but this does not usually lead to an inflammatory response. Regulation of the population size of T-LPL and the termination of the immune responses by CD95mediated apoptosis is part of normal immunoregulation. 4s One tantalising piece of evidence linking apoptosis and Crohn's disease comes from the study of mice which express a mutant non-functional form of the cell adhesion molecule N-cadherin. Examination of their intestinal mucosa reveals histological changes reminiscent of Crohn's disease, with a substantial increase in the number of apoptotic cells in both villi and crypts. 49Further examination of these issues is required. However, there are a number of conditions where the inappropriate or abnormal induction of apoptosis plays an important role in pathogenesis. Shigella flexneri can evade immune attack by inducing apoptosis in host macrophages. This is the only bacterium currently known to employ such a strategy. 5~Clostridium difficile causes detachment of intestinal epithelial cells, resulting in their apoptosis. 51 Non-steroidal anti-inflammatory drugs (NSAIDs) induce apoptosis in intestinal epithelium. The presence of apoptotic bodies can help the histopathologist in the diagnosis of NSAID-induced colitis/enteropathy. 52

418 Another interesting example of abnormal and excessive apoptosis has come from the study of the SW620 colorectal cancer cell line23 This line expresses both CD95 (Fas) and CD95 ligand (FasL), but does not undergo apoptosis through ligation of its own CD95 receptor as would be expected. 34 Surprisingly however, SW620 are able to induce apoptosis in Jurkat cells through ligation of SW620 CD95 ligand to Jurkat CD95. These results suggest that at least some colorectal cancers can evade immune attack through mounting a counterattack against activated T cells. This phenomenon, which its investigators have Called the "Fas counterattack" (or perhaps, more correctly, CD95 counterattack) requires further investigation. The induction of apoptosis through the ligation of CD95 can be turned to therapeutic advantage. Testicular cells constitutively express CD95 ligand, which induces apoptosis in mature T cells (which express CD95) entering the testis. 54 This probably explains the mechanism of immune-privilege in the testis. This observation has been applied to the problem of transplant rejection. Pancreatic islets were wrapped in myoblasts which has been transfected with CD95 ligand. When transplanted under the renal capsule of mice, the transfected myoblasts substantially prolonged the life of the transplanted islets by inducing apoptosis in invading T cells before the T cells were able to establish a rejection reaction. 5s These observations may be of profound importance for transplantation medicine.

Gastrointestinal diseases with deficient apoptosis Cells are constantly exposed to mutagens and divide billions of times in the life span of a human being. Therefore the chances of a cell developing a malignant mutation and becoming a fatal carcinoma should be high. However, only one in three humans ever develops cancer. This low cancer rate compared to the high number of opportunities for a cell to become malignant suggest that sophisticated mechanisms have evolved to prevent the development of cancer. Apoptosis is one of these anti-cancer mechanisms. Recently it has become clear that for a cell to become malignant, not only must it acquire mutations to give the cell unregulated growth but also mutations to prevent the destruction of the mutant cell by apoptosis before it can undergo clonal expansion. 56 In the gastrointestinal tract the evidence in favor of the hypothesis that failure of appropriate apoptosis is a necessary step in the development of cancer is strong. Examination of the site of apoptosis within the crypt provides suggestive evidence for the importance of apoptosis in the prevention of cancer. Small intestinal cancer is at least 100% less common than colorectal

A.J.M. Watson: Apoptosis and intestinal disease cancer. In the small intestine, chemical carcinogeninduced apoptosis occurs with maximum frequency at the level of stem cells within the crypt (cell positions 46). This suggests that a stem cell with a potentially malignant mutation will be deleted immediately.57 By contrast, in the colon, the maximum incidence of carcinogen-induced apoptosis is at cell positions 5-10, above the positions of the stem cells, which are at positions 1-2. These data suggest that mutant stem cells in the colon have a greater chance of avoiding deletion by apoptosis than mutant stem cells of the small intestineY Further evidence comes from study of the genes that are either mutant or abnormally expressed in colorectal cancer. A number of studies have shown that bcl-2 is expressed in approximately 75% of colonic adenomas and in a lower proportion of colorectal carcinomas.59At present it is not known whether this expression merely reflects the phenotype of the cell which originally gave rise to the tumor or whether the expression is abnormal. Curiously, to date there have been no studies to indicate whether the bcl-2 gene is mutant or wild-type. The reason for bcl-2 expression disappearing during progression from adenoma to carcinoma is not fully understood. Truncation or deletion of the bcl-2 gene is a likely explanation in a proportion of cases, as deletions of chromosome 18q, the locus of the bcl-2 gene, occur in 69% of colorectal cancers. 6~ In addition there is an exquisite inverse relationship between bcl-2 and p53 expression) 9 The explanation for this inverse relationship is not clear, as in the majority of cases p53 will be mutant and non-functional. Nevertheless, wild-type p53 and some mutants (e.g., mut 175) 62 down-regulate bcl-2 by binding to a transcriptional silencer element within the bcl-2 promoter. 63 Overall, these studies suggest that bcl-2 acts as a survival factor during the adenoma phase of the development of colorectal neoplasia, preventing deletion of mutant clones by apoptosis. However, it should be emphasized that our understanding of the role of bcl-2 in colorectal cancer development is at a very early stage. Bcl-2 is a member of a gene family of which some members inhibit apoptosis (e.g., bcl-xL 64 and bag-16~) (not a family member but an associated protein), while bax, 66 bcl-xs 64 and bak 67 promote apoptosis. These proteins heterodimerize with each other (e.g., bcl-2/bax) with the functional results depending on the ratio of the two proteins. 68 Recent results indicate that there is a progressive fall of bak and rise in bcl-xL during colorectal neoplasia, with bax remaining unchanged.69 Together these changes put the tissue into an anti-apoptotic state. A paradox of bcl-2 expression in colorectal cancer is that bcl-2-positive colorectal cancers have a better prognosis than bcl-2negative ones. 7~This is counterintuitive, as it might be predicted that bcl-2-positive tumors would be resistant

A.J.M. Watson: Apoptosis and intestinal disease to apoptosis and therefore more likely to survive and progress. One possible explanation is that bcl-2negative tumors express other survival factors while have more aggressive characteristics than bcl-2. Mutation of p53 occurs in approximately 86% of colorectal cancers and occurs most frequently at the transition between adenoma and carcinoma. 71 Mutant p53 protein is readily detectable by immunohistochemistry as it has a far longer half-life than wild-type protein. 72 The function of p53 is to prevent the accumulation of genetic damage within the cell. 73 In response to DNA damage, oncogene activation, infection with certain viruses, or tissue hypoxia p53 levels rise. In general, if the damage has been mild, the cyclindependent-kinase inhibitor p21 TM is activated; this halts the cell cycle and allows repair to take place. 75If, on the other hand, genomic damage is severe, the cell is induced to undergo apoptosis via a signal pathway which is not currently understood. Severity of DNA damage is not the only determinant of p53 action. Other, as yet poorly defined, cellular characteristics influence the action of p53 following DNA damage. In the intestine, p53 plays an important role in the regulation of apoptosis. Intestinal cells which are null for p53 undergo apoptosis when transfected with wildtype p53. 76 In irradiated wild-type mice,p53 is expressed at the same cell positions in the small intestine as those which undergo apoptosis, whereas in the colon the correlation is much less precise, corroborating the idea that stem cells from the colon are much less susceptible to apoptosis than those from the small intestine. Functional p53 is essential for the induction of apoptosis at the base of small and large intestinal crypt 4.5 h after irradiation, since it does not occur in p53-null mice. 77Of interest, p53 has no influence on spontaneous apoptosis, as this is unaltered in p53-null mice. Tumors, for example colorectal adenomas, often contain areas of hypoxia because of inadequate blood supply. Defective apoptosis plays a particularly interesting role in hypoxic tumors, which can lead to the selection ofp53-mutant clones. In normal tissue, hypoxia induces p53, causing the death of hypoxic cells through apoptosis. However, if a few cells have mutant p53 they will be given a selective advantage over p53 wild-type cells, since they will not undergo apoptosis during hypoxia. TM Thus, in areas of hypoxia within a tumor, p53 mutant cells will tend to replace p53 wild-type cells and lead to the progression of the tumor's malignant phenotype. The simplest experiment to evaluate whether defective apoptosis is important in the development of colorectal cancer is to count the number of apoptotic cells within a tumor compared to normal tissue. However, such experiments are fraught with methodological difficulties and the results to date have been very vari-

419 able, with some investigators finding apoptosis to be decreased, 79while others find it increased. 8~At least two methodological problems face the investigator. First concerns the appropriate normal cells to use as control. As stated previously, the incidence of apoptosis varies widely along the crypt/villus axis. Which point along this axis should be used to obtain control values for apoptosis? Presumably the most appropriate control would be cells at the cell position which gave rise to the tumor, but for any given tumor this will be unknown. A second problem concerns the half-life of apoptotic cells within a tissue or tumor before they are removed by phagocytosis. An absolute requirement for the comparison of apoptosis counts between two tissues is that the half-life of the apoptotic cells within the two tissues is identical. 2~This assumption may well not be true because of the altered phenotype of tumor cells and the abnormal blood supply which might impede the entry of phagocytes into a tumor. One of the most compelling new developments in gastroenterology is the discovery that NSAIDs can prevent or reduce the risk of developing colorectal cancer. Evidence for this comes from a variety of independent sources. A number o f epidemiological studies have shown that individuals taking one aspirin per day have a reduced risk of developing colorectal cancer of up to 50%. 81,82 Moreover, a randomized controlled trial has demonstrated that sulindac can reduce the size and number of polyps in patients with familial adenomatous polyposis (FAP). 83 Furthermore, a number of animal studies have shown that a variety of NSAIDs can prevent the development of carcinogen-induced colorectal neoplasia. 84.85 It is likely these drugs act by inducing apoptosis in colorectal tumors, thereby slowing or halting their development. FAP patients treated with sulindac for 3 months show increased levels of apoptosis in adenomas, without changes in proliferation rates. 86 Qualitatively similar results have been obtained with the colon cancer cell line HT-29. 87 The mechanism of induction of apoptosis by NSAIDs is controversial. NSAIDs inhibit an enzyme called cyclooxygenase (COX), which is sometimes also referred to as prostaglandin endoperoxide synthase (Fig. 2). This enzyme is involved in the synthesis of prostaglandins and catalyses the conversion of free arachidonic acid to prostaglandin G/H2, which is then converted to other prostaglandins depending on the cell type involved. 88There are two isoforms of COX. COX1 is constitutively expressed in the gastrointestinal tract and is responsible for the maintenance of tissue integrity.89COX-2, on the other hand, is usually undetectable in resting conditions but can be induced by growth factors or cytokines.9~ Study of human colorectal tumors has shown that 43% of adenomas and 90% of carcinomas express COX-2 at both the mRNA and protein

420

A.J.M. Watson: Apoptosis and intestinal disease

Aspirin-like drugs (NSAIDs)

Normalepithelium Fig. 2.

~-lCox-2 --~tApoptosis

Action of cyclooxygenase

~- Cancer

(COX)-2 and non-steroidal

antiinflammatory drugs (NSAIDs) in colorectal neoplasia

level. 91 There is no such change in the expression of COX-1. These data suggest that the expression of COX2 may be important in the development of colorectal cancer from adenomas. Overexpression of COX-2 in the non-transformed intestinal epithelial cell line RIE-1 causes a number of important phenotypic changes. 92 First, unlike control cells, they do not undergo apoptosis in response to exposure to sodium butyrate. 93 Significantly, this inhibition of apoptosis was associated with an increase in bcl-2 protein expression. The NSAID sulindac can restore apoptosis in a dose-dependent manner, suggesting that NSAIDs may regulate bcl-2 expression; a result of great importance. Secondly, COX-2-overexpressing cells have increased adhesion to extracellular matrix proteins compared to controis. Thirdly, COX-2 expression activates matrix metalloproteinase-2, which enhances the invasiveness of cells and their potential to form metastases. 94 These data suggest that NSAIDs prevent colorectal cancer through the inhibition of COX-2. Although most of the NSAIDs currently available are more potent inhibitors of COX-1 than COX-2, they do retain some activity against COX-2. 95 This implies that new specific COX-2 inhibitors will be far more potent than the current COX-1 inhibitors in the prevention of colorectal cancer. As NSAID-induced gastropathy is the result of inhibition of COX-1 it would be predicted that specific COX-2 inhibitors will not have this side-effect. However, it should be noted that, although COX-2 inhibitors will probably cause less ulceration in the upper gastrointestinal tract, COX-2 activity is required for ulcer healing, so that COX-2 inhibitors may cause preexisting ulcers to bleed or perforate. Sulindac may induce apoptosis by mechanisms other than inhibition of COX-2. Sulindac sulfone, which has no activity against COX-2, induces apoptosis in HT-29 cells, though at higher concentrations than the active metabolite of sulindac, sulindac sulfide. 96 Preliminary evidence suggests this second mechanism may involve modification of the composition of intestinal bile acids. 97 Both irradiation and the anti-cancer chemotherapeutic agents induce apoptosis in intestinal epithelium. In detailed studies by Ijiri and Potten, 98,99 the maximum incidence of apoptosis along the crypt/villus axis of

BDF1 mice was documented during a 12-h period following administration of 15 different cytotoxic drugs. These agents could be separated into three classes depending on the site of apoptosis; isopropylmethane-sulfonate, Neomycin, and adriamycin affected cells at positions 4-6 (the probable stem cells); bischloroethylnitrosourea, actinomycin D, cyclophosphamide, and cycloheximide affected cells at positions 6-8 (the early transit cells); while mechloroethane, triethylenethiophosphoramide, vincristine, 5fluorouracil, hydroxyurea, and methotrexate affected cell positions 8-11 (the late transit cells). The mechanisms and relevance of this differential sensitivity are not fully understood. Variations in tissue access, pharmacokinetics, and drug target availability are likely to make a contribution, though it is possible the ability to engage apoptosis varies as well. These results suggest that apoptosis may account for the enteropathy which sometimes occurs after chemotherapy. Studies are underway in our laboratory to evaluate this hypothesis. Mutations of tumor suppressor and abnormal expression of bcl-2 family members not only drive the evolution of normal epithelium into cancer but also prevent response to the treatment. As mentioned above, radiation induces less apoptosis in p53-null cells77 or cells expressing bcl-2Y The same is also true for apoptosis at the crypt base following treatment with 5-fluorouracil, where apoptosis is attenuated inp53-null transgenic mice and increased in bcl-2 null transgenic mice.100 The purpose of this review was to summarize some of the recent information on the role of apoptosis in intestinal disease. This is a new area of investigation, though important methodological and theoretical challenges remain. However, the advances made so far justify the prediction that major therapeutic advances in the fields of inflammatory disease, oncoiogy, and transplantation will be made through apoptosis research.

Acknowledgments. The

author acknowledges helpful discussions with Prof Chris Potten, Prof John Hickman, Dr Mark Pritchard, Dr Rod Benson, Dr Quan Chen, and all members of the Molecular Pharmacology Group at the University of Manchester.

References 1. Lockshin RA, Williams CM. Programmed cell death: 1. Cytology of degeneration in the intersegmental muscles of the pernyi silk moth. J Insect Physiol 1965;11:123-133. 2. Kerr JFR, Wyllie AH, Currie AR. Apoptosis: A basic biological phenomenon with wide ranging implications in tissue kinetics. Br J Cancer 1972;26:239-257. 3. Schwartz LM, Smith SW, Jones MEE, et al. Do all programmed cell deaths occur via apoptosis? PNAS 1993;90:980-984. 4. Arends MJ, Wyllie AH. Apoptosis: Mechanisms and roles in pathology. Int Rev Exp Pathol 1991;32:223-254.

A . J . M . Watson: A p o p t o s i s a n d intestinal disease 5. Thompson CB. Apoptosis in the pathogenesis and treatment of disease. Science 1995;267:1456-1462. 6. Pritchard DM, Watson AJM. Apoptosis and gastrointestinal pharmacology. Pharmacol Ther 1996; 72:149-169. 7. Watson AJM. Manipulation of cell death--the development of novel strategies for the treatment of gastrointestinal disease. Aliment Pharmacol Ther 1995;9:215-226. 8. Que FG, Gores GJ. Cell death by apoptosis: Basic concepts and disease relevance for the gastroenterologist. Gastroenterology 1996;110:1238-1243. 9. Benson RSP, Heer S, Dive C, Watson AJM. Characterisation of cell volume loss in CEM-C7A cells during dexamethasoneinduced apoptosis. A m J Physiol 1996;270 (Cell Physiol 39):Cl190-1203. 10. Evan GI, Wyllie AH, Gilbert CS, et al. Induction of apoptosis in fibroblasts by c-myc protein. Cell 1992;63:119-128. 11. Savill J. Apoptosis in disease. Eur J Clin Invest 1994;24:715723. 12. Watson AJM. Apoptosis and necrosis in the gastrointestinal tract. Gut 1995;37:165-167. 13. Fesus L, Thom~izy V, Falus L. Induction and activation of tissue transglutaminase during programmed cell death. FEBS Lett 1991;224:104-108. 14. Martin SJ, Reutelingsperger CPM, McGahon AJ, et al. Early redistribution of plasma membrane phosphatidylserine is a general feature of apoptosis regardless of the initiating stimulus: Inhibition by overexpression of Bcl-2 and Abl. J Exp Med 1995;182:1545-1556. 15. Savill J, Hogg N, Ren Y, et aL Thrombospondin cooperates with CD36 and the vitronectin receptor in macrophage recognition of neutrophils undergoing apoptosis. J Clin Invest 1992;90:15131522. 16. Merritt AJ, Jones LS, Potten CS. Apoptosis in murine small intestinal crypts. In: Cotter TG, Martin SJ (eds) Techniques in apoptosis, a user's guide. 1996; London: Portland Press. 17. Oberhammer F, Wilson JW, Dive C, et al. Apoptotic death in epithelial cells: Cleavage of D N A to 300 and/or 50 kb fragments prior to or in the absence of internucleosomal fragmentation. EMBO J 1993;12:3679-3684. 18. Wyllie AH. Glucocorticoid-induced thymocyte apoptosis is assodated with endogenous endonuclease activation. Nature 1980;284:555-556. 19. Gavrieli Y, Sherman Y, Ben-Sasson SA. Identification of programmed cell death in situ via specific labelling of nuclear DNA fragmentation. J Cell Biol 1992;119:493-501. 20. Potten CS. What is an apoptotic index measuring? Br J Cancer 1996;74:1743-1748. 2i. Darzynkiewicz Z, Li X. Measurements of cell death by flow cytometry. In: Cotter TG, Martin SJ (eds) Techniques in apoptosis, a user's guide. 1996, London: Portland Press. 22. Potten CS. Structure, function and proliferative organisation of mammalian gut. In: Potten CS, Hendry JH (eds) Radiation and gut, Amsterdam: Elsevier: Science, 1995;1-31. 23. Potten CS. Effects of radiation on murine gastrointestinal cell proliferation. In: Potten CS, Hendry JH (eds) Radiation and gut, Amsterdam: Elsevier Science, 1995;61-84. 24. Potten CS. The significance of spontaneous and induced apoptosis in the gastrointestinal tract of mice. Cancer Metastas Rev 1992;11:179-195. 25. Hall PA, Coates P J, Ansari B, et al. Regulation of cell number in the mammalian gastrointestinal tract: The importance of apoptosis. J Cell Science 1994;107:3569-3577. 26. Shibahara T, Sato N, Waguri S, et al. The fate of effete epithelial cells at the villus tips of the human small intestine. Arch Histol 1995;58:205-219. 27. Merritt A J, Potten CS, Watson AJM, et al. Differential expression of bcl-2 in intestinal epithelia--correlation with attenuation of apoptosis in colonic crypts and the incidence of colonic neoplasia. J Cell Science 1995;108:2261-2271.

421 28. Krajewski S, Krajewski M, Shabaik A, et al. Immunohistochemical determination of in vivo distribution of Bax, a dominant inhibitor of bcl-2. Am J Pathol 1994;115:1323-1336. 29. Krajewski S, Krajewski M, Reed JC. Immunohistochemical analysis of in vivo patterns of Bak expression, a proapoptotic me-mber of the Bcl-2 protein family. Cancer Res 1966;56:28492855. 30. Strater J, Wedding U, Barth TFE, et al. Rapid onset of apoptosis in vitro follows disruption of 131-intergrin/matrix interactions in human colonic crypt cells. Gastroenterology 1996;110:17761784. 31. Basora N, Perreault N, Vachon PH, et al. Restricted expression of the ct7BI31 integrin, a laminin receptor in human intestinal epithelial cells. Gastroenterology 1996;110:A791. 32. Hermiston ML, Gordon JI. In vivo analysis of cadherin function in the mouse intestinal epithelium: Essential roles in adhesion, maintenance of differentiation, and regulation of programmed cell death. J Cell Biol 1995;129:489-506. 33. Madara JL. Maintenance of the macromolecular barrier at cell extrusion sites in intestinal epithelium: Physiological rearrangement of tight junctions. J Memb Biol 1990;116:177-184. 34. Osborne B. Apoptosis and the maintenance of homeostasis in the immune system. Curr Opin Immunol 1996;8:245-254. 35. Watson AJM, Askew JN, Benson RSP. Poly(Adensosine diphosphate ribose) polymerase inhibition prevents necrosis induced by H202 but not apoptosis. Gastroenterology 1995;109: 472~482. 36. Watson AJM. Askew JN, Sandle GI. Characterisation of oxidative injury to an intestinal cell line (HT-29) by hydrogen peroxide. Gut 1995;35:1575-1581. 37. Lowin B, Hahne M, Tschopp J. Cytolytic T-cell cytotoxicity is mediated through perforin and Fas lytic pathways. Nature 1994: 370:650-652. 38. Berke G. The CTLs kiss of death. Cell 1995;81:9-12. 39. Damon A J, Nicholson DW, Bleakley RC. Activation of the apoptotic protease CPP32 by cytotoxic T-cell-derived granzyme B. Nature 1995;377:446--448. 40. Muzio M, Arul M, Chinnaiyan FC, et al. FLICE, A Novel FADD homologous ICE/CED-3-1ike protease, is recruited to the CD95 (Fas/Apo-1) death-inducing signalling complex. Cell 1996;85:817-827. 41. Rouvier E, Luciani M-F, Golstein P. Fas involvement in Ca 2+independent T cell-mediated cytotoxicity. J Exp Med 1993; 177:195-200. 42. De Maria R~ Boirivant M, Cifone MG, et al. Functional expression of Fas and Fas ligand on human gut lamina propria T lymphocytes. J Clin Invest 1996;97:316-322. 43. Jones NL, Yeger H, Cutz E, et al. Helicobacter pylori induces apoptosis of gastric antral epithelial cells in vivo. Gastroenterology 1996;110:A933. 44. Peek RM, Moss SF, Tham KT, et al. Infection with H. pylori CAGA + strains dissociates gastric epithelial proliferation from apoptosis. Gastroenterology 1996;110:A575. 45. Reich K, Lingnau F, Williams RM, et al. Corticosteroids downregulate bcl-2 and induce apoptosis in CD4+ LPL in Crohn's disease. Gastroenterology 1996;110:A999. 46. Attila AL, Moss SF, Walters JRF, et al. Increased small bowel epithelial apoptosis reflects celiac sprue activity. Gastroenterology 1995;108:A271. 47. Gallucci BB; Sale GE, McDonald GB, et al. The fine structure of human rectal epithelium in acute graft-versus-host disease. Am J Surg Pathol 1982;6:293-305. 48. De Maria R, Fais S, Silvestri L, et al. Continuous in vivo activation and transient hyporesponsiveness to TCR/CD3 triggering of human gut lamina propria T lymphocytes. Eur J Immunol 1993;23:3104-3108. 49. Hermiston ML, Gordon JI. Inflammatory bowel disease and adenomas in mice expressing a dominant negative N-cadherin. Science 1995;270:1203-1207.

422 50. Zychlinsky A, Preuost MC, Sansonetti PJ. Shigella flexneri induces apoptosis in infected macrophages. Nature 1992;358:167169. 51. Mahida YR, Makh S. Hyde S, et aL Human intestinal epithelial cells produce interleukin-8 (IL-8) and undergo apoptosis following cell detachment induced by Clostridium difficile toxin A. Gastroenterology 1995;108:A868. 52. Lee FD. Drug-related pathological lesions of the intestinal tract. Histopathology 1994;25:303-308. 53. O'Connell J, O'Sullivan G, Collins JK, et al. The fas counteratt a c k - a n o t h e r mode of lymphoepithelial communication. Gastroenterology 1996;110:A981. 54. Bellgrau D, Gold D, Selawry H, et al. A role for CD95 ligand in preventing graft rejection. Nature 1995;377:630-632. 55. Lau HT, Yu M, Fontana A, et al. Prevention of islet allograft rejection with engineered myoblasts expressing FasL in mice. Science 1996;273:10%112. 56. Harrington E, Fanidi A, Evan GI. Oncogenes and cell death. Curt Opin Genet Dev 1994;4:120-129. 57. Li YQ, Fan CY, O'Connor PJ, et al. Target cells for the cytotoxic effects of carcinogens in the murine small bowel. Carcinogenesis 1992;13:361-368. 58. Potten CS, Li YQ, O'Connor PJ, et al. A possible explanation for the differential cancer incidence in the intestine, based on distribution of the cytotoxic effects of carcinogens in the murine large bowel. Carcinogenesis 1992;13:2305-2312. 59. Watson AJM, Merritt A J, Jones LS, et aL Evidence for reciprocity of bct-2 and p53 expression in human colorectal adenomas and carcinomas. Br J Cancer 1996;76:889-994. 60. Ayhan A, Yasui W, Yokozaki H, el al. Loss of heterozygosity at the bcl-2 gene locus and expession of bcl-2 in human gastric and colorectal carcinomas. Jpn J Cancer Res 1994;85:584-591. 61. Jen J, Kim H, Piantadosi S, et al. Allelic loss of chromosome 18q and prognosis in colorectal cancer. N Engl J Med 1994;331:213221. 62. Haldar S, Negrini M, Monne M, et al. Down-regulation of bcl-2 by p53 in breast cancer cells. Caner Res 1994;54:2095-2097. 63. Miyashita T, Harigal M, Hanada M, et al. Identification ofa p53dependent negative response element in the bcl-2 gene. Cancer Res 1994;54(12):3131-3135. 64. Boise LH, Gonzalez-Garcia M, Postema CE, et al. bel-x, a bcl-2related gene that functions as a dominant regulator of apoptotic cell death. Cell 1994;74:597-608. 65. Takayama S, Sato T, Krajewski S, et al. Cloning and functional analysis of BAG-l: A novel Bcl-2 binding protein with anti-cell death activity. Cell 1995;80:279-284. 66. Oltvai ZN, Hillman CL, Korsmeyer SJ. Bcl-2 heterodimerises in vivo with a conserved homolog, Bax, that accelerates programmed cell death. Cell 1993;74:609-619. 67. Farrow SN, White JHM, Martinou I, et al. Cloning of a bcl-2 homologue by interaction with adenovirus EIB 19K. Nature 1995;374:731-733. 68. Yin XM, Oltvai ZN, Korsmeyer SJ. BH1 and BH2 domains of Bcl-2 are required for inhibition of apoptosis and heterodimerization with bax. Nature 1994;369:321-323. 69. Krajweski M, Moss SF, Krajewski S, et al. Elevated expression of Bcl-X and reduced Bak in primary coloreetal adenocarcinomas. Cancer Res 1996;56:2422-2427. 70. Ofner D, Riehemann K, Maier H, et al. Immunohistochemically detectable bcl-2 expression in colorectal carcinoma: Correlation with tumour stage and patient survival. Br J Cancer 1995;72:981985. 71. Baker SJ, Preisinger AC, Jessup JM, et al. p53 gene mutations occur in combination with 17p allelic deletions as late events in colorectal tumorigenesis. Cancer Res 1990;50:7717-7722. 72. Lane DP, Midgley CA, Hupp TR, et al. On the regulation of the p53 tumour suppressor, and its role in the cellular response to DNA damage. Philos Trans R Soc Lond Biol 1995;347:8387.

A.J.M. Watson: A p o p t o s i s a n d intestinal disease 73. Jacks T, Weinberg RA. Cell cycle control and its watchman. Nature 1996;381:643-644. 74. el-Deiry WS, Tokino T, Velculescu VE, et al. WAF1, a potential mediator of p53 tumor suppression. Cell 1993;75:817825. 75. Marx J. How p53 suppresses cell growth. Science 1993;262:16441645. 76. Shaw P, Bovey R, Tardy S, et al. Induction of apoptosis by wildtype p53 in a human colon tumour-derived cell line. Proc Natl Acad Sci USA 1992;89:4495-4499. 77. Merritt A J, Potten CS, Kemp CJ, et al. The role of p53 in spontaneous and radiation-induced apoptosis in the gastrointestinal tract of normal and p53-deficient mice. Cancer Res 1994;54:614-617. 78. Graeber TG, Osmanian C, Jacks T, et al. Hypoxia-mediated selection of cells with diminished apoptotic potential in solid turnouts. Nature 1996;379:88-91. 79. Garewal HS, Bernstein C, Bernstein R, et al. Quantitation of apoptosis in normal and neoplastic colonic epithelium of rats and humans. Gastroenterology 1996:110:A517. 80. Fryer B, Hehald C, Driggers L, et al. Reduced apoptosis in colorectal mucosa: A potential biologic marker for cancer. Gastroenterology 1996;110:A514. 81. Thun MJ, Namboodiri MM, Heath CW. Aspirin use and reduced risk of fatal colon cancer. N Engl J Med 1991;325:15931596. 82. Oiovannuci E, Egan KM~ Hunter D J, et at. Aspirin and the risk of colorectaI cancer in women. N Engl J Med 1995;333:609614. 83. Giardello FM, Hamilton SR, Krush A J, et al. Treatment of colonic and rectal adenomas with sulindac in familial adenomatous polyposis. New Engl J Med 1993;328:1313-1316. 84. Moorghen M, Ince P, Finney K J, et al. A protective effect of sulindac against chemically induced primary colonic tumours in mice. J Pathol 1988;156:341-347. 85. Craven PA, DeRubertis FR. Effects of aspirin on 1,2dimethylhydrazine-induced colonic carcinogenesis. Carcinogenesis t992;13:541-546. 86. Pasricha PJ, Bedi A, O'Connor K, et al. The effects of sulindac on colorectal proliferation and apoptosis in familial adenomatous polyposis. Gastroenterology 1995;109:994-998. 87. Shift SJ, Qiao L, Tsai L, et al. Sulindac sulfide, an aspirin-like compound, inhibits proliferation, causes cell cycle quiescence and induces apoptosis in HT-29 colon adenocarcinoma cells. J Clin Invest 1995;96:491-503. 88. Williams CS, DuBois RN. Prostaglandin endoperoxide synthase: Why two isoforms? Am J Physiol 1996;270:G393G400. 89. Wallace JL. Prostaglandins, NSAIDs and cytoprotection. Gastroenterol Clin North Am 1992;21:631-641. 90. DuBois RN, Awad J, Morrow J, et al. Regulation of eicosanoid production and mitogenesis in rat intestinal epithelial cells by transforming growth factor-a and phorbol ester. J Clin Invest 1994;93:493-498. 91. Eberhart CE, Coffey RJ, Radhika A, et al. Up-regulation of cyclooxygenase 2 gene expression in human colorectal adenomas and adenocarcinomas. Gastroenterology 1994;107: 1183-1188. 92. Tsuji M, DuBois RN. Alterations in cellular adhesion and apoptosis in epithelial cells overexpressing prostaglandin endoperoxide synthase 2. Cell 1995;83:493-501. 93. Hague A, Manning AM, Hanlon KA, et al. Sodium butyrate induces apoptosis in human colonic tumour cell lines in a p53independent pathway: Implications for the possible role of dietary fibre in the prevention of large bowel cancer. Int J Cancer 1993;55:498-505. 94. Tsuji M, DuBois RN. Cyclooxygenase-2 expression in Caco-2 cells activates matrix metalloproteinase-2. Gastroenterology 1996;110:A605.

A . J . M . Watson: A p o p t o s i s a n d intestinal disease 95. Mitchell JA, Akarasereenot P, Thiermermann C, et al. Selectivity of non-steroidal antiinflammatory drugs as inhibitors of constituitive and inducible cyclooxygenase. Proc Natl Acad Sci USA 1994:90:11693-11697. 96. Piazza GA, Rahm ALK, Krutzch M, et al. Antineoplastic drugs sulindac sulfide and sulfone inhibit cell growth by inducing apoptosis. Cancer Res 1995;55:3110-3116. 97. Batta AK, Salen G, Pamukcu R, et al. Sulindac and its sulphone derivative inhibit colon cancer via modification of intestinal bile acids. Gastroenterology 1996;110:A490.

423 98. Ijiri K, Potten CS. Response of intestinal cells of differing topographical and hierarchical status to ten cytotoxic drugs and five sources of radiation. Br J Cancer 1983;47:175-185. 99. Ijiri K, Potten CS. Further studies on the response of intestinal crypt cells of different hierarchical status to 18 different cytotoxic agents. Br J Cancer 1987;55:113-123. 100. Pritchard DM, Watson AJM, Potten CS, et al. Characterization of 5-fluorouracil-induced apoptosis in murine small and large intestine: Studies in p53-null and bcl-2-null transgenic mice. Gastroenterology 1996;110:A578.