The death of small intestinal epithelial cells has been characterized and quantitated after irradiation of mice rendered homozygously null for the p53 gene.
Oncogene (1997) 14, 2759 ± 2766 1997 Stockton Press All rights reserved 0950 ± 9232/97 $12.00
Apoptosis in small intestinal epithelia from p53-null mice: evidence for a delayed, p53-indepdendent G2/M-associated cell death after g-irradiation Anita J Merritt1,3,4, Terence D Allen2, Christopher S Potten1 and John A Hickman3 Cancer Research Campaign Departments of 1Epithelial Biology and 2Ultrastructure, The Paterson Institute for Cancer Research, Christie Hospital Trust, Wilmslow Road, Manchester and 3Molecular and Cellular Pharmacology Group, School of Biological Sciences, The University of Manchester, Manchester M13 9PT, UK
The death of small intestinal epithelial cells has been characterized and quantitated after irradiation of mice rendered homozygously null for the p53 gene. In wild-type animals homozygous for p53 a rapid (4.5 h) elevation of p53 protein was observed in the proliferative compartment of the crypts after 8 Gy of irradiation. Cells underwent cell death by apoptosis in this region. We had reported previously a total repression of apoptosis in small intestinal crypt epithelia 4.5 h after the g-irradiation (8 Gy) of p53 homozygously null animals. Thus, while 400 apoptotic cells were observed in 200 half crypts taken from wild-type animals at 4.5 h, this fell to background levels (10 ± 30) in the p53 null animals (Merritt et al., 1994) and did not increase by 12 h. However, we have now found a delayed initiation of a p53-independent apoptosis after 8 Gy of g-radiation: at 24 h, approximately 100 apoptotic cells were observed in 200 half crypts. This late wave of apoptosis was not observed after 1 Gy of g-radiation. The morphological appearance of this p53-independent apoptosis suggested that death may have arisen as the result of aberrant mitosis. Analysis of the regeneration of crypts 3 days after irradiation of mice with between 11 and 17 Gy showed that there was no signi®cant increase (P=0.135) in the potential of clonogenic cells from the p53 null animals to repopulate the crypts. The data support the idea that a p53independent apoptotic mechanism permits the engagement of apoptosis, probably by a mitotic catastrophe, after 8 Gy of g-irradiation in vivo and that a loss of p53 does not make these epithelial cells radioresistant in vivo to doses of 8 Gy and above. In contrast, irradiation with 1 Gy failed to induce a p53-independent apoptosis in vivo, suggesting that the p53 `sensor' of damage was more sensitive than that engaging the p53-independent mechanism of cell death. Keywords: p53; apoptosis; p53-independent apoptosis; radiation; small intestine; G2/M checkpoint
Introduction The initiation of death by apoptosis in cells which have sustained DNA damage is a critical process which can limit oncogenesis. The survival of cells with
Correspondence: JA Hickman 4 Current address: McArdle Laboratory for Cancer Research, 1400 University Avenue, Madison, Wisconsin 53706 ± 1599, USA Received 16 August 1996; revised 27 February 1997; accepted 4 March 1997
DNA damage is permissive for subsequent malignant transformation. The tumor suppressor p53 has been suggested to play a key role in determining the fate of cells which have received potentially carcinogenic DNA damage (Lane, 1992). The amount of p53 protein is increased in DNA-damaged cells and it may act either as a transcriptional activator or repressor. According to cell type, the nature of changes in gene expression and whether the cells are proliferating or not, DNA damaged cells may either undergo a G1 and/or, in some cases, a G2 cell cycle `checkpoint' (Kastan et al., 1991; Kuerbitz et al., 1992; Stewart et al., 1995; Agarwal et al., 1995; Guillouf et al., 1995). This may allow repair of a damaged DNA template or surveillance of damaged spindles (Cross et al., 1995). Alternatively, p53 may induce cell death by apoptosis (Yonish-Rouach et al., 1993). In non-proliferating thymocytes from mice which have been rendered genetically homozygously null for p53 (7/7), radiation and other types of DNA damage failed to initiate apoptosis in short term assays of cell integrity (Clarke et al., 1993; Lowe et al., 1993a). However, DNA damage-induced cell death in mitogen stimulated, proliferating T lymphocytes from the same p53 7/7 mice was at the same levels as in p53+/+ cells (Strasser et al., 1994) and the clonogenic survival of irradiated primary fibroblasts from p537/7 mice also showed no dierence from the p53+/+ cells in their sensitivity to DNA damage (Slichenmeyer et al., 1993). These latter studies are important because they reported clonogenic assays of survival, removing the uncertainty that a loss of p53 may have only a temporal eect on the initiation of cell death, delaying it rather than inhibiting it. They also provide a measure of the maintenance of cellular functionality (replicative potential). Taken together these results suggest that, in some cell types, p53 independent mechanisms of cell death initiated by DNA damage may be as important as p53-dependent mechanisms. We are interested in the determinants of apoptotic death in the rapidly proliferating epithelia of the gastrointestinal tract (reviewed by Potten, 1992). In particular, we wish to determine what factors promote the so-called `altruistic' apoptosis of damaged stem cells in the small intestine, an eect which we have proposed may limit cancer incidence at this site, essentially by deleting cells with DNA damage so as to prevent tumour initiation (Potten, 1992). The contribution that expression of p53 made to the induction of apoptosis in the epithelia of the small intestine was investigated using p53 homozygously null mice (7/7). The normal wave of
p53 independent cell death AJ Merritt et al
2760
apoptotic death observed between 4 h and 8 h after g radiation (8 Gy) was completely abrogated in the absence of p53 (Merritt et al., 1994; Clarke et al., 1994). This suggested that the genotypically p53 null small intestinal epithelia were radioresistant, in the way that thymocytes appeared to be. In the present study we have investigated this further by analysis of the long-term fate of the p53 null epithelial cells in the small intestinal crypts both by a morphological analysis and by assays of the clonogenic potential of cells which remain viable shortly after irradiation.
Results Induction of apoptosis in the small intestinal epithelia Figure 1 shows the numbers of apoptotic cells at various times post irradiation, scored at positions 1 to 7 in 200 half crypts from p53 wild type (+/+), heterozygotes (+/7) and null (7/7) mice. After 8 Gy of g radiation, the small intestinal crypt epithelia from the p53+/+ animals showed a rapid induction of apoptosis which was declining by 24 h. The results at the early timepoints (512 h) for both the p53+/+ and p537/7 animals essentially recapitulate those reported by us (Merritt et al., 1994) and by Clarke et al. (1994), with a highly signi®cant reduction of radiation-induced apoptosis in the p537/7 animals. At the lower dose of 1 Gy, there was a reduction of apoptosis in the +/7 genotype in comparision to wild-type +/+ animals, as he had been reported previously for 5 Gy or irradiation (Clarke et al., 1994). We were surprised to observe a late wave of apoptosis in the p537/7 null animals at 24 h after 8 Gy. This late wave of apoptotic cell death was still apparent at 40 h (Figure 1). Irradiation of animals with 1 Gy (Figure 1) shows that the p53 null genotype completely inhibited the apoptosis otherwise observed in both the +/+ and +/7 mice at 4.5, with no late apoptoses at 24 h or 40 h. Figure 2a shows the morphology of a longitudinal section of the crypts of murine small intestine from a p537/7 animal 24 h after 8 Gy of radiation, stained with haematoxylin and eosin (H&E) and incubated with terminal deoxynucleotidyl transferase for the detection of apoptosis-associated DNA strand breaks. Some cells with an apoptotic morphology had the appearance of giant cells, with distinct fragments of condensed chromatin in their nucleus (arrows). The cleavage of DNA to form signi®cant numbers of strand breaks in these large cells was con®rmed by the immunohistochemistry of DNA fragmentation using terminal deoxynucleotidyl transferase and biotinylated dUTP (ISEL) (Figure 2a and b). Further examples of these apoptotic cells are shown in Figure 2c to 2i, where is it clear that not all apoptotic cells or apoptotic bodies were ISEL positive. Electron microscopy con®rmed the unusual morphology of some of the late, p53-independent cell deaths, in comparison to classically apoptotic cells (Figure 3). Many of these giant cells (Figure 3A) contained marginated and condensed chromatin in nuclei which were large and, in some cases appeared to be attempting mitosis.
Figure 1 Kinetic analysis of the appearance of apoptotic cells at positions 1 ± 7 from the base of the crypt in the small intestine of mice with the p53 genotypes shown. The upper panel shows the response to 8 Gy of g-irradiation and the lower panel to 1 Gy. Results showing signi®cant dierences are denoted on the x-axis
DNA synthesis in small intestinal crypt epithelia after irradiation The incorporation of radiolabelled thymidine into small intestinal crypt epithelial cells undergoing DNA synthesis is shown in Figure 4. The labelling index for p53 wild type (+/+) and homozygously null (7/7) unirradiated animals was signi®cantly dierent with respect to the cell positions in the crypt at which proliferative cells were found (positions 10 ± 17). These are not the cell positions at which p53 is optimally induced after irradiation (positions 4 ± 8) (Merritt et al., 1994). Four h after irradiation with 8 Gy a signi®cantly greater percentage of cells in small intestinal crypts from p537/7 mice incorporated radiolabelled thymidine than those from the p53+/+ mice. This increase in labelling index in the p537/7 crypts was at cell positions which showed greatest expression of p53 after irradiation (Merritt et al., 1994). At 4 h, the percentage of proliferative cells, in both wild-type and heterozygotes, fell only very marginally despite the imposition of considerable amounts of cell damage and loss in this region of the crypt (see Figures 1 and 2). Minimum labelling was observed at 14 h and there was evidence of regenerative recovery at 24 h. Changes in mitotic index after irradiation Table 1 shows the mitotic indices of the small intestinal crypt epithelia from p53 wild type (+/+), heterozygotic (+/7) and homozygously null (7/7) animals various times after irradiation with 8 Gy of g-radiation.
p53 independent cell death AJ Merritt et al
In all genotypes 8 Gy inhibited mitosis for at least 6 h after irradiation, i.e. irrespective of p53 genotype, but by 9 and 12 h the crypt cells from the p53 7/7 mice had signicantly recovered. A second reduction of mitotic index was observed at 32 h and was presumed to be part of an oscillatory recovery (Potten, 1990). Clonogenic survival of small intestinal crypt from p53+/+ and p537/7 mice Figure 5 shows the results from an experiment designed to assess the survival of stem cells in the small
intestinal crypts after radiation. The regrowth of crypts was assessed in mice from the dierent p53 backgrounds. The data show that there was no signi®cant dierence in the survival of clonogenic stem cells according to p53 status (P=0.135 by an ftest). Estimates of clonogenic cell numbers by the `split dose' method (see Materials and methods) were made after 8+8 Gy and 12+12 Gy of radiation. After 8+8 Gy the numbers of clonogenic cells were estimated to be 4+2 in the p53+/+ animals and 6+2 in the p53 nulls. After 12+12 Gy these ®gures were 13+8 and 16+15 respectively.
Figure 2 The morphology of apoptotic cells from the small intestinal crypts of a p537/7 mouse 24 h after 8 Gy of g-irradiation. The apoptotic cells (arrows) stained by the use of terminal deoxynucleotidyl transferase (ISEL) (see Materials and methods) are stained brown. (a) and (b) show longitudinal crypt sections with both apoptotic cells (arrows) and apoptotic fragments (arrowheads). (M) = mitotic cells; (P) = Paneth cells. (c ± i) show a range of individual cell morphologies. In (c), the events observed represent a possible sequence of events as apoptosis takes place in a cell attempting mitosis: A1 resembles a mitosis at a very early stage of apoptosis where chromatin is beginning to condense and, A2, starts to fragment (ISEL+). A3 shows a fully apoptotic morphology which is ISEL positive. (d ± i) show further examples. In these it is shown that although the majority of morphologically apoptotic cells (arrows) or apoptotic bodies (arrowheas) were ISEL positive, there were examples with condensed chromatin which were negative. (a,61400; b ± i,62500)
2761
p53 independent cell death AJ Merritt et al
2762
Figure 4 The positional distribution of 3H-thymidine labelled cells, to give a % labelling index (LI) in the small intestinal crypts of mice with p53+/+, +/7 or 7/7 genotypes at various times after 8 Gy of g-irradiation (cp = cell position)
Table 1 Mitotic indices of small intestinal crypt epithelia from animals of different p53 genotypes various times after 1 or 8 Gy of girradiation Figure 3 Electron micrographs showing: (A) the ultrastructure of p53-independent apoptosis of small intestinal epithelial cells 24 h after 8 Gy of g-irradiation (64300). (B), the morphology of an apoptotic cell from a small intestinal crypt of a p53+/+ mice at 8 h. The micrograph (B) shows that there is (1) an apoptotic cell which appears to have previously phagocytosed an apoptotic body (2) and that both of these have been phagocytosed by a healthy cell, whose nucleus is marked N. P = Paneth granules (64300)
p53 genotype
+/+
+/7
7/7
0 Gy
2.7 n=4 0.1a n=4 0a n=4 0.3a n=5 2.1b n=4 2.0c n=4 1.7c n=4 1.4b n=4 2.3 n=4
3.6 n=4 0.1a n=4 0a n=3 0.5b n=3 1.9 n=4 ND
2.4 n=4 0.2b n=4 0.1b n=4 1.5 n=4 3.5 n=3 2.9 n=4 1.8# n=4 1.9# n=4 1.8 n=4
8 Gy 4.5 h 8 Gy 6 h 8 Gy 9 h 8 Gy 12 h 8 Gy 24 h
Discussion It has been suggested that in some cell types the loss of a wild-type p53 genotype renders cells resistant to the lethal eects of DNA damage (Lowe et al., 1993a; Clarke et al., 1993; 1994; Merritt et al., 1994). The loss of p53 as the `guardian of the genome' (Lane, 1992) might therefore be important in promoting the survival of cells which had undergone the initiating events of carcinogenesis, in the progression of tumours because of increased genetic instability (reviewed by Almasan et al., 1995), and in resistance to radiation therapy and genotoxic chemotherapy (Lowe et al., 1993b, 1994). Indeed, in our own previous study, which was the ®rst to investigate the eects of the speci®c loss of p53 on the response to DNA damage in vivo, utilising p53 null mice, we suggested that epithelial cells of the gastrointestinal tract were essentially rendered radioresistant by the deletion of the p53 gene (Merritt et al., 1994), results con®rmed by Clarke et al. (1994) using dierent p537/7 mice. Clarke et al. (1994) also showed that, in comparison to the +/+ animals,
8 Gy 32 h 8 Gy 36 h 8 Gy 40 h
1.9b n=4 1.3a n=4 3.3 n=4
n=number of experiments. aP40.001, bP40.02, cP40.05 compared to controls (t-test). #, 32, 36 and 40 h signi®cantly lower than 12 h (P40.05)
DNA synthesis, measured by bromodeoxyuridine incorporation, continued in small intestinal crypt epithelia after irradiation of the p537/7 animals. However, observations of short-term radioresistance in the absence of p53 contrast with in vitro data using cells from p537/7 mice, where, using a clonogenic assay, there was no dierence in the long-term survival of cells derived from p53+/+ or 7/7 mice (Slichenmeyer et al., 1993, Strasser et al., 1994). Furthermore, with respect to the intestinal tract, in vivo studies by Clarke et al. (1995) of p537/7 mice
p53 independent cell death AJ Merritt et al
Figure 5 Crypt survival curves for p53+/+ and 7/7 mice. Groups of four or more mice were irradiated with 11 to 17 Gy of g-radiation and sacri®ced 3 days later. The numbers of surviving crypts/regenerating crypts was estimated as described (Materials and methods). Values from individual mice were plotted. The survival curves were not signi®cantly dierent by an f-test (P=0.135)
crossed with APC+/7 mice showed that there was no increase in the incidence in the rate of adenoma formation; it was argued cogently that the prevention of mutation in this tissue, which might have been the result of the engagement of a p53-dependent cell death or the engagement of a cell cycle checkpoint, was achieved instead by some p53 independent route. Similar conclusions were reached with respect to the loss of p53 on the events initiating skin papilloma formation (Kemp et al., 1993) and a recent study of sporadic human colon cancers has suggested a limited role for p53 in the maintenance of genetic stability (Kahlenberg et al., 1996). In the light of these questions about the importance of the role of p53 in the initiation of some epithelial malignancies, we investigated the long term fate of epithelial cells from the small intestinal crypts of the p53 null mice which had not undergone apoptosis but had sustained considerable DNA damage after 8 Gy of g-radiation (Merritt et al., 1994). It was considered that their continued viability might permit DNA repair, although possibly not with high ®delity since Clarke et al. (1994) reported that they continued to progress into S-phase, suggesting that this was due to the lack of any G1 checkpoint. It is this scenario, of continued DNA synthesis and subsequent division in the presence of DNA damage, which is considered to be able to ®x clastogenic damage so as to promote the emergence of malignant cells. However, analysis of cell viability beyond 12 h shows that there was a wave of apoptosis in the small intestine at 24 and 40 h in the p53 null animals irradiated with 8 Gy (Figures 1 and 2). The initial peak of radiation-induced apoptosis induced by 8 Gy was absent from the crypts of the p53 null animals,
as we reported previously (Merritt et al., 1994). These cells appeared viable and, in agreement with the studies of Clarke et al. (1994) incorporated deoxyntucleotides into DNA (Figure 4). We presume that these, the `undead' cells, were able to complete DNA synthesis but then reach a G2/M checkpoint, as described in other p537/7 cells (Agarwal et al., 1995; Strasser et al., 1994). This is supported by our observation that mitotic ®gures were initially absent from all epithelial cells irrespective of p53 genotype up to 9 h after 8 Gy of irradiation, a result similar to that reported by Clarke et al. (1994) after 5 Gy of radiation. After accumulation in what we assume to be the G2 phase of the cell cycle, a p53-independent mechanism of cell death was then initiated in cells which presumably retained signi®cant amounts of DNA damage. G2 has been suggested to be the common radiation-induced checkpoint for these cells (Lesher and Bauman, 1969). A puzzling aspect of this delayed cell death concerns those cells which were irradiated when in late S-phase, which would shortly reach G2. Why was death by a p53-independent pathway not initiated within 12 h, the time it took for mitosis to be reestablished (Table 1)? In this way it would be expected that a continuous and steady number of apoptotic cells would be observed from 12 h onwards, as cells move to G2, undergo a checkpoint and die. The delay of a further 12 h, and their presumed movement into mitosis, may be caused by the requirement for the synthesis of some trigger for progression into M-phase. The morphology of some of the apoptotic cells undergoing a p53-independent cell death was quite unlike that of those seen 3 ± 4 h after radiation of p53+/+ mice. These were approximately twice the volume, and super®cially, appeared to be either multinucleate or to possess many condensed chromatin masses (Figures 2 and 3). In addition to their positive staining for DNA breaks using the TdT reaction (Figure 2) electron microscopy con®rmed that they were apoptotic, although with an unusual morphology (Figure 3A). Electron micrographs con®rmed that these were large cells, presumably which had reached G2, and had a pattern of condensed chromatin which was marginated but in lobe-like structures. This was strongly reminiscent of the morphology of vincristine-induced apoptosis in the small intestine (Harmon et al., 1992) suggesting some aberration during mitosis. Indeed, we observed the occasional apoptotic cell that appeared to be in mitosis which was also TdT positive (Figure 2). When animals were irradiated with 1 Gy, there was again an early onset and peak of apoptosis between 4.5 h and 9 h in animals with wild-type p53 although, as expected, the levels were reduced in comparison to those given 8 Gy and was observed to be gene dosage dependent (Figure 1). Loss of p53 completely abrogated this early wave of cell deletion. Moreover, although a continuing pattern of apoptosis at 24 h was observed in the p53 +/+ animals, this was not observed in the epithelial cells in the null animals after 1 Gy. Measurement of the mitotic indices in all genotypes showed that 1 Gy had no signi®cant eect on the percentage of cells in mitosis at early times (data not shown), despite the fact that considerable numbers of cells were undergoing apoptosis (Figure 1). We have
2763
p53 independent cell death AJ Merritt et al
2764
previously suggested that after a low dose of radiation, such as 1 Gy, the very radiosensitive stem cells are normally deleted but these are replaced by their immediate daughters from within the crypt, thus reestablishing crypt hierarchy without the necessity for regenerative or compensatory cell divisions (Potten, 1990). The incorporation of [3H]thymidine suggests that after radiation damage of small intestinal epithelial crypts of wild-type (+/+) animals, p53 induced either a checkpoint which inhibited entry to S phase or generally slowed progress through S phase (Figure 4). Early studies by Lesher and Bauman (1969) had shown that irradiation of the small intestine produced a G2 block. There was no evidence for a G1 block. This is dicult to reconcile with the observation of a fall in DNA synthesis. Alternatively, the higher labelling of cells from the null animals at 4 h may be because surviving cells enter S-phase, whilst in the wild-type animals they have been rapidly deleted from the crypt. We are currently investigating the expression of the cyclin-dependent kinase inhibitor waf-1 under these conditions. Preliminary data shows expression of protein to be largely p53 dependent, supporting the idea that a p53-regulated cell cycle block may occur. The general pattern observed here is similar to the changes reported in a comprehensive study on BDF1 mice (Potten et al., 1990). In that study, an initial regression of labelled cells and labelling index re¯ected the shrinkage of crypt size due to apoptosis, mitotic delay (G2 block), reproductive sterilisation and continued emigration of cells onto the villus. This is followed by an oscillating regenerative proliferation which originates in the surviving clonogenic cells and becomes evident in the labelling frequency plots by 24 h. Damaged cells which progress into S-phase after 8 Gy then appear to reach a G2 checkpoint and attempt to undergo mitosis, where they undergo a mitotic catastrophe (Figures 2 and 3). This is by a p53-independent mode of apoptosis. This second default mechanism for the engagement of apoptosis occurred only after 8 Gy; after 1 Gy very little apoptosis was observed at 24 h in the p53 null animals. This suggests that either DNA repair had been sucient to allow escape from the second checkpoint or that the threshold of damage being `sensed' by this p53-independent mechanism was greater than that of p53. Although we have discovered that small intestinal crypt epithelial cells which have received g-radiationinduced DNA damage may be deleted by both p53dependent and -independent mechanisms the dierence in the temporal onset of these two processes could be hypothesised to be critical in determining the longterm fate of damaged cells. Failure to rapidly delete DNA damaged cells or to impose a G1 checkpoint, with the opportunity for repair, particularly after lower levels of damage (1 Gy), allows damaged cells to progress through S-phase. The survival of these cells for a further period in G2 may also allow repair where again a decision whether to progress through the cell cycle or to die is made. In G2 the threshold for the detection of, or response to damage appears to be greater (Figure 1). If repair is pro®cient in those cells which escape from the p53-mediated deletion, it would be predicted that they could repopulate the
crypt. Experiments which attempted to establish whether more clonogenic cells survived in p53 null animals suggested that no signi®cant survival advantage was provided by a deletion of both alleles of the p53 gene (Figure 5). Thus, in the epithelia of the murine small intestine, the loss of p53 appears to provide no long-term survival advantage after DNA damage imposed by high doses of g-irradiation. Although the methods used here to establish this may be limited, because only high doses of radiation can be used to destroy crypt integrity for the purpose of measuring the emergence of viable stem cells, our ®ndings that p53 provides no clonogenic survival in intestinal epithelia are concordant with published studies of primary ®broblasts and lymphoid cells (Slichenmeyer et al., 1993; Strasser et al., 1994). That p53 may, in some settings, have a limited role as a guardian of the genome with respect to the prevention of events initiating intestinal carcinogenesis and in the prevention of subsequent genetic instability is a conclusion recently drawn by others (Fishel, 1996; Kahlenberg, 1996). Materials and methods Unless otherwise stated, all general purpose chemicals were obtained from BDH/MERCK (Poole, Dorset, UK). p53 null mice The p53 homozygously null mice were from a colony derived from those generated by Donehower et al. (1992). p53 null (7/7), heterozygous (+/7) and wildtype (+/+) mice were generated by the appropriate crosses and the p53 null strain used had a 129/C57B16 background. All mice used were 10 ± 12 week old males and they were maintained and genotyped as described by us previously (Merritt et al., 1994, 1996). Irradiation and preparation of mouse small intestinal tissue Groups of 4 ± 6 mice were whole body g irradiated with the appropriate doses using a Caesium137 source at a dose rate of 3.8 Gy/min. The mice were sacri®ced at various times. The entire small intestine was removed, ®xed in Carnoy's ®xative and `gut-bundled' as described previously (Merritt et al., 1994). This gave a minimum of 12 cross sections per mouse. The tissue was processed for histology and 3 ± 5 mm sections were cut onto uncoated microscope slides and stained routinely with haematoxylin and eosin or processed for staining with streptavidin-biotin-labelled dUTP after its incorporation by terminal deoxynucleotidyltransferase (ISEL), as described in detail by us previously (Merritt et al., 1995, 1996). Estimation of apoptosis This was performed essentially as described in detail previously (Merritt et al., 1996). Brie¯y, ®xed and stained transverse sections of small intestine were examined and, from haematoxylin and eosin-stained sections apoptotic fragments and mitotic cells were scored according to their cell position, numbering from the base of the crypt. Between 150 and 200 half crypts were counted for each group and in some cases apoptosis was also scored based on the ISEL technique as described previously (Merritt et al., 1995, 1996).
p53 independent cell death AJ Merritt et al
Estimation of DNA synthesis and mitotic indices in small intestinal crypts After irradiation groups of four mice were given 25 mCi of tritiated thymidine (6.7 Ci/mmol), intraperitoneally, 40 min prior to the time of sacri®ce. Histological samples prepared as slides (Merritt et al., 1995), were dewaxed overnight in xylene and immersed in four changes of absolute alcohol for 10 min each. After washing, the slides were then rinsed in two changes of 0.5% Photo¯ow in distilled water (Kodak, Hemel Hempstead, Herts, UK). The autoradiographic emulsion was prepared using a 1 : 1 solution of K-5 nuclear track emultion in distilled water (Ilford, Cheshire, UK) and slides dipped rapidly into this prior to air drying for 2 ± 3 h. Autoradiographs were developed under red light conditions using Kodak-D19 developer, according to manufacturers instructions. Slides were washed for 2 mins with 1% glacial acetic acid then ®xed with Hypan ®xative (Ilford, Cheshire, UK). After washing with distilled water, slides were stained with haematoxylin and eosin, dehydrated and mounted in Xam (BDH, Poole, UK). Cells deemed to be in S-phase contained 43 silver grains over their nuclei. The number and cell position of labelled cells was scored in the same way as apoptotic cells (see above).
®tted to the Standard Multitarget Model and curves were ®tted using the DRFIT programme (Roberts, 1990). The curves were plotted as surviving fractions against dose (Gy). In addition to generating crypt survival curves, ftests were performed to determine the signi®cance of the dierence between two survival curves. To estimate the number of clonogenic cells per crypt, the `split dose' method was used as described (Hendry, 1978; Potten and Hendry, 1995). Groups of animals were given either a single fraction of 8 Gy or two identical fractions of 8 Gy plus 8 Gy separated by 4 h. A second experiment involved using doses of 12 Gy single and 12 Gy plus 12 Gy 4 h apart. Animals were sacri®ced 3 days later. The number of surviving crypts were recorded in ten cross sections for each mouse per group. The numbers were then size corrected, as described above (Potten et al., 1981). A mean value for the number of surviving crypts (after size correction) was obtained for each group. Assuming that the fractions each kill the same proportion of clonogens, and that all clonogens have the same radiosensitivity, the original number of clonogens, m, can be computed from the equations: and
S1
1 ÿ
1 ÿ Sc m m
The Crypt microcolony assay The crypt microcolony assay was performed as described by Potten and Hendry 1995. For each mouse, ten crosssections or circumferences were chosen, excluding those containing Peyer's patches, those badly orientated or those damaged during sectioning. The number of surviving crypts around the whole of each circumference was recorded, using the criteria that a survivor consisted of at least ten adjacent, darkly stained, healthy, non-Paneth epithelial cells. Size correction for the regrowing crypts was performed as described by Potten et al. (1981). This is necessary because the size of the surviving crypts can vary quite considerably with the time of sampling and the dose of radiation. Brie¯y, the widths of 15 randomly chosen crypts per mouse were measured at right angles to the line of the lumen using an eye piece graticule and a light microscope. The corrected number of surviving crypts for a given treatment was then calculated using the following equation:
Corrected Number of Survivors crypt width Uncorrected Number 2 Control Treated crypt width The size corrected numbers were then converted to surviving fractions of the control number. The data were
S2 1 ÿ
1 ÿ Sc 2 Where S1 and S2 are the crypt surviving fractions after one and two fractions, Sc the clonogen surviving fraction and m the original number of clonogens. Electron microscopy Mice were irradiated as described above and sacri®ced either 4.5 h or 24 h later. The small intestines were dissected out, and tissue was cut into approximately 1 mm cubes. These were instantly ®xed in 2.5% glutaraldehyde in 0.15 M Sorenson's phosphate buer pH 7.35, followed by 1% osmium tetroxide in the same buer. Specimens were dehydrated through the graded alcohol series and embedded in Agar 100 epoxy resin (Agar Scienti®c, Stanstead, Essex, UK). Sections were stained with uranyl acetate and lead citrate and examined at 80 kV in a Philips EM400 TEM (Philips Analytical, Cambridge, UK). Acknowledgements We thank the Cancer Research Campaign for a studentship (to AJM) and for continued funding. We are grateful to Caroline Chadwick for assistance with assessment of much of the data, and Drs Mark Pritchard, Christine Chresta and Caroline Dive for their comments on the manuscript.
References Agarwal ML, Agarwal A, Taylor WR and Stark GR. (1995). Proc. Natl. Acad. Sci. USA, 92, 8493 ± 8497. Almasan A, Linke SP, Paulson TG, Huang LC and Wahl GM. (1995). Cancer Metast. Rev., 14, 59 ± 73. Clarke AR, Purdie CA, Harrison DJ, Morris RG, Bird CC, Hooper ML and Wyllie AH. (1993). Nature, 362, 849 ± 852. Clarke AR, Gledhill S, Hooper ML, Bird CC and Wyllie AH. (1994). Oncogene, 9, 1767 ± 1773. Clarke AR, Cummings MC and Harrison DJ. (1995). Oncogene, 10, 1913 ± 1920. Cross SM, Sanchez CA, Morgan CA, Schimke MK, Ramel S, Idzerda RL, Rakind WH and Reid BJA. (1995). Science, 267, 1353 ± 1356. Donehower LA, Harvey M, Slagle BL, McArthur MJ, Montgomery CA Jr, Butel JS and Bradley A. (1992). Nature, 356, 215 ± 221.
Fishel R. (1996). J. Natl. Cancer Inst., 88, 1608 ± 1609. Guillouf C, Rosselli F, Krishnaraju K, Moustacchi E, Homan B and Liebermann DA. (1995). Oncogene, 10, 2263 ± 2270. Harmon BV, Takano YS, Winterford CM and Potten CS. (1992). Cell Prolif., 25, 523 ± 536. Hendry JH. (1978). Radiation Res., 78, 404 ± 414. Kahlenberg MS, Stoler DL, Basik M, Petrelli NJ, Rodriguez-Bigas M and Anderson GR. (1996). J. Natl. Cancer Inst., 88, 1665 ± 1670. Kastan MB, Onyekwere O, Sidransky D, Vogelstein B and Craig RW. (1991). Cancer Res., 51, 6304 ± 6311. Kemp CJ, Donehower LA, Bradley A and Balmain A. (1993). Cell, 74, 813 ± 822. Keurbitz SJ, Plunkett BS, Walsh WV and Kastan MB. (1992). Proc. Natl. Acad. Sci. USA, 89, 7491 ± 7495. Lane DP. (1992). Nature, 358, 15 ± 16.
2765
p53 independent cell death AJ Merritt et al
2766
Lesher S and Bauman J. (1969). Natl. Cancer Inst. Monographs, 30, 185 ± 198. Lowe SW, Schmitt ES, Smith SW, Osborne BA and Jacks T. (1993a). Nature, 362, 847 ± 849. Lowe SW, Ruley HE, Jacks T and Housman DE. (1993b). Cell, 74, 957 ± 967. Lowe SW, Bodis S, McClatchey LR, Ruley HE, Fisher DE, Housman DE and Jacks T. (1994). Science, 266, 807 ± 810. Merritt AJ, Potten CS, Kemp CJ, Hickman JA, Balmain A, Lane DP and Hall PA. (1994). Cancer Res., 54, 614 ± 617. Merritt AJ, Potten CS, Watson AJM, Loh DY, Nakayama L, Nakayama K and Hickman JA. (1995). J. Cell Sci., 108, 2261 ± 2271. Merritt AJ, Jones LS and Potten CS. (1996). In: Eds Martin SJ and Cotter, TG. Techniques in apoptosis: a users guide. Portland Press: London. 269 ± 300. Potten CS. (1990). Int. J. Radiat. Biol., 58, 925 ± 974.
Potten CS, Owen G and Roberts SA. (1990). Int. J. Radiat. Biol., 57, 185 ± 199. Potten CS. (1992). Cancer Metast. Rev., 11, 179 ± 195. Potten CS and Hendry JH. (1995). In: Eds Potten CS and Hendry JH. Radiation and the gut. Elsevier Science: Oxford pp 45 ± 59. Potten CS, Rezvani M, Hendry JH, Moore JV and Major D. (1981). Int. J. Radiat. Biol., 40, 321 ± 326. Roberts SA. (1990). Int. J. Radiat. Biol., 57, 1243 ± 1246. Slichenmeyer WJ, Nelson WG, Slebos RJ, Kastan MB. (1993). Cancer Res., 53, 4164 ± 4168. Stewart N, Hicks GG, Paraskevas F and Mowat M. (1995). Oncogene, 10, 109 ± 115. Strasser A, Harris AH, Jacks T and Cory S. (1994). Cell, 79, 329 ± 339. Yonish-Rouach E, Resnitzky D, Lotem J, Sachs L, Kimchi A and Oren M. (1991). Nature, 353, 345 ± 347.
