Int. J. Cancer: 50,683-688 (1992) Publication of the International Union Against Cancer Publication de I'Union Internationale Contre le Cancer

0 1992 Wiley-Liss, Inc.

OVER-EXPRESSION OF p53 NUCLEAR ONCOPROTEIN IN COLORECTAL ADENOMAS Massimo PIGNATELL11'2'5 , Gordon W.H. STAMP'.*, Georgia KAFIRI', David LANE' and Walter F. BODMER4 'Department of Histopathology, Royal Postgraduate Medical School, London; 2RCSIICRF Histopathology Unit, Lincoln's Inn Fields, London; 'Cancer Research Campaign Laboratories, Department of Biochemistry, Dundee University, Dundee; and 41mperial Cancer Research Fund, Lincoln's Inn Fields, London, UK.

p53 is a nuclear phosphoprotein which controls normal cell growth. Normal p53 protein is undetectable by standard immu- nohistochemical staining and the over-expression found in neoplastic cells correlates with the presence of point mutations of evolutionary conserved regions of the p53 gene. We exam- ined the expression of p53 protein in a series of 36 colorectal adenomas (I 3 tubular, I7 tubulovillous, 6 villous) showing different degrees of dysplasia (I I mild, I9 moderate, 6 severe), I I moderately differentiated adenocarcinomas (6 Duke's A, 4 Duke's B, I Duke's C) and 5 metaplastic polyps using the polyclonal antibody CM I which recognises p53 protein in conventionally fixed and processed histological material. We found that 15 out of 36 colorectal adenomas showed p53 immunoreactivity, although in 4 positive cases (26%) the stain- ing was very focal (<0.1% positive cells). More than 80% of severely dysplastic adenomas showed strong p53 immunoreac- tivity and this over-expression was correlated with increased cell proliferative rate as detected by the proliferating-cell- nuclear-antigen (PCNA) staining. p53 nuclear staining was also seen in 8 out of I I (65%) colorectal adenocarcinomas as previously shown. Our data suggest that the p53 gene mutation, with the subsequent over-expression of the protein, occurs in colorectal adenomas and may therefore be a fundamental genetic event underlying the dysplasia and loss of proliferative control that are characteristic of adenomas with malignant potential.

p53 is a 393-aminoacid nuclear phosphoprotein which regu- lates normal cell growth (Lane and Benchimol, 1990). The levels of p53 protein in normal cells and tissues are extremely low because of the protein's short half-life, and are undetect- able by standard immunohistochemical staining (Rodrigues et al., 1990). In contrast, high levels of the protein can be detected by immunohistochemistry in primary lung tumours (Iggo et al., 1990) as well as in mammary (Bartek et al., 1990a, b ) and colorectal adenocarcinomas (Van Den Berg et al., 1989; Rodrigues et al., 1990; Scott et al., 1991). High levels of expression of the p53 protein in tumours and tumour cell lines correlate with the presence of point mis-sense mutations of evolutionary conserved regions of the p53 gene, which is located on the human chromosome 17p (Iggo et al., 1990; Rodrigues et al., 1990; Bartek et al., 1990b; Davidoff et al., 1991). Mutated p53 appears to function as a dominantly-acting oncogene in transformation assay (Parada et al., 1984), al- though wild-type p53 gene has a tumour suppressor function on the growth of human colorectal carcinoma cells in vitro and the in vivo point mutation abrogates this suppressive effect (Baker et al., 1990).

Most colorectal carcinomas arise from adenomas which can be defined as an abnormal focus of epithelial proliferation in which the immature proliferative cells are not limited to the base of the crypt, as in the normal colonic mucosa, but occupy the entire crypt length and appear within the surface epithe- lium (Jass, 1989). Adenomas can be classified according to the histological structure as tubular, tubulovillous and villous adenomas. The cells of these tumours show different degrees of dysplasia, including pleomorphism, nuclear hyperchromasia and variability, loss of polarity and an increased mitotic index. Dysplasia, large size and villous histology are the major characteristics of the adenomas with malignant potential (Muto et al., 1975).

Loss of a portion of chromosome 17p has been found in more than 70% of colorectal carcinomas often associated with point mutation of the remaining allele, but is relatively infrequent in adenomas ( < 10%) (Vogelstein et al., 1988). Previous immunohistochemical studies using cryostat sections have shown that over-expression of the nuclear oncoprotein p53 occurs infrequently (8%) in tubular adenomas (Van Den Berg et al., 1989; Rodrigues et al., 1990). It has therefore been proposed that a mutated p53 gene may play a role in the late stage of colon carcinogenesis, determining the progression from large and dysplastic adenomas to invasive carcinomas, rather than an initial neoplastic event (Fearon and Vogelstein, 1990).

Here we describe the results of an immunohistochemical study for the expression of p53 nuclear protein in a series of 36 adenomas (13 tubular, 17 tubulo-villous, 6 villous) showing different degrees of dysplasia and in 11 moderately differenti- ated adenocarcinomas using the polyclonal antibody CM1 on formalin-fixed paraffin-embedded tissues. These data were correlated with a proliferative index for the tumours as detected by proliferating-cell-nuclear-antigen (PCNA) stain- ing (Mathews et al., 1984) using the PClO monoclonal antibody (MAb) (Hall et al., 1990; Waseem and Lane, 1990).

MATERIAL AND METHODS

Archival paraffin-embedded tissues were obtained from the Histopathology Department, Royal Postgraduate Medical School, London. The specimens had been fixed in neutral formalin-buffered saline for at least 18 hr and then processed for histological examination. Sections (3 km) of paraffin- embedded tissue were attached to glass slides by incubation at 37°C overnight. Sections were then dewaxed, hydrated, and washed in phosphate-buffered saline (PBS), pH 7.6.

Thirty-six adenomas (13 tubular, 17 tubulo-villous and 6 villous), 11 moderately differentiated adenocarcinomas of the large bowel (6 Duke's A, 4 Duke's B, 1 Duke's C), 5 hyperplastic polyps and 4 normal colonic biopsies were stud- ied. The clinical and histological data are shown in Tables I and 11. The histological diagnosis and degree of dysplasia were based on assessment of corresponding tissue removed for conventional histological analysis.

Paraffin-embedded sections were stained by the avidin-biotin- complex (ABC) immunoperoxidase technique using the rabbit polyclonal antibody CM1 recognizing the p53 nuclear protein (BBrtkovB et al., 1991). Sections were blocked with normal swine serum for 20 min at room temperature before applying the CM1 polyclonal antibody at the dilution of 1500 to 1:lOOO overnight. After washing 3 times with PBS, a biotin-conjugated swine anti-rabbit immunoglobulin was added for 30 min,

'To whom correspondence and reprint requests should be ad- dressed, at Department of Histopathology, Royal Postgraduate Medi- cal School, Du Cane Road, London W12 OHS, UK. FAX: +44 81 740 7417.

Received: July 10,1991 and in revised form October 2,1991

684 PIGNATELLI ET AL.

TABLE I - 053. PCNA EXPRESSION AND CLINICO-PATHOLOGICAL VARIABLES IN COLORECTAL ADENOMAS ~ ~~ ~

- PCNA (av ) p53 PCNA (p53') PCNA (max ) % S % M % S % M % S % M %

Case Sex1 Site1 Type1 number Age Size DySplaSld

1 2 3 4 5 6 7 8 9

10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36

Fl90 F164 Fl71 F/69 MI 14 MI86 F/60 MI76 Fl20 MI37 MI61 MI56 F/77 F162 Fl69 MI66 Fl77 F180 MI86 F165 Fl74 Fl69 F180 MI77 Fl74 MI82 Fl66 MI62 MI74 MI46 MI78 MI67 F/73 Fl68 Fl61 Mi58

Rl0.8 cm R l l cm Rl0.2 cm SlO.2 cm Sl3.2 cm R l l cm D I l cm Sl2.1 cm Rl0.3 cm Rl1.2 cm SlO.6 cm S10.9 cm S l l cm R10.8 cm SlO.2 cm S11 cm S11 cm Rl1.4 cm Rl0.5 cm Dl7.5 cm S10.5 cm RI0.5 cm R11 cm R/ 1.2 cm SlO.5 cm S/2 cm R10.3 cm S13.2 cm S/1 cm Rl0.4 cm R/1 cm S/1 cm Sl1.5 cm R/1 cm S/1.5 cm R/1.5 cm

TAI + TAI + TAI + TAI + TAI + VA/ + TVAI + TVAI + TVAI + TVAI + TVAI + TAI + + TAI + + TAI + + T A / + + TAI + + TAI + + TAI + + VA/ + + VA/++ VAI++ VAI++ TVAI + + TVA/++ TVAI + + TVAI + + TVAI f + TVAI + + TVAI + + TVAI + + T A / + + + VAI+++ TVAl+++ TVAl+++

+I30 - - - - - - -

+I- <0.1 +/-<0.1

+I10 -

- - - -

++ <0.1 + 15 +I10 +/ - <0.1 - - -

+I <0.1 +/<0 .1 + 140 + 120

+I20 + + / l o ++I50

+1<0.1

- -

-

28 16

8 10 11 9

10 42 25 9

40 17 16 6

27 3

28 17

10 5

46 9

15 25 5 5

27 45 53 19 12

-

-

-

25 37

12 15 23 11 12 62 33 23 27 33 20 12 20 8

38 20

28 12

30 28 17 35 11 10 38 58 40 19 42

-

-

-

15 3

6 8 6 5 5

20 13 7

23 13 10 5

23 1

15 12

6 3

14 6 9

13 2 2

14 23 20 6 7

-

-

-

19 1

10 9

10 8 7

39 26 18 25 19 16 8

19 6

20 13

14 9

20 14 13 21

7 8

29 39 29 8

13

-

-

-

TVAI+++ ++I50 20 30 20 35 9 22 TVA/+++ ++I50 50 50 50 50 35 37

A, ascending colon; D, descending colon; R, rectum; S, sigmoid colon; TA, tubular adenoma; VA, villous adenoma; TVA, tubular-villous adenoma. +, mild dysplasia; + +, moderate dysplasia; + + +, severe dysplasia. S, strong staining; M, moderate to weak staining; PCNA (P53'), PCNA staining in p53-positive areas; PCNA ( m a . ) , maximum PCNA staining; PCNA (av.), average PCNA staining.

TABLE I1 - ~ 5 3 , PCNA EXPRESSION AND CLINICO-PATHOLOGICAL VARIABLES IN COLORECTAL ADENOCARCINOMAS

Case Sex/ number Age

- PCNA (av.) PCNA (p53') PCNA (max.) pi3 S % M % S % M % S % M %

Sitel Duke's Size stage

1 2 3 4 5 6 7 8 9

10 11

MI64 R12.5cm Fl68 R/1 cm Fl85 Sl2.5 cm MI78 R13cm MI70 S17cm MI68 Dl7.5cm MI54 Rl5 cm F154 Al4.5 cm MI82 A12.3cm MI63 SlO.5 cm MI71 S10.5 cm

A A A A A A B B B B C

- +/ <0.1 + 110

++I85 ++I50 ++I30

++I50 +I10 ++I30

-

-

- 22 6 12 15 12 42 40 50 50 50

- 25 62 22 70 44 70

-

-

- - - - - - - -

- 18 15 45 62 45 62 17 36 17 36 27 30 27 30

-

15 5 7 13

28 40 20 42 22 41 - __ - - 7 12

21 36 10 25 17 18

S, strong staining; M, moderate staining; PCNA (P53'), PCNA staining in p53-positive sections; PCNA (max.), maximum PCNA staining; PCNA (av.), average PCNA staining.

followed by further washing and incubation with the avidin- biotin complex for a further 30 min (both from Dakopatts, High Wycombe, UK). Three more washes (10 min each) were performed, following which the peroxidase reaction was devel- oped using a freshly prepared, filtered solution of diaminoben- zidine (Sigma, St. Louis, MO), 5 mg in 20 ml PBS containing 0.03% hydrogen peroxide. The reaction was stopped after 3 min by washing in tap water. The slides were counterstained with Mayer's haemalum for 1 min, dehydrated in a graded

alcohol series, cleared in xylene and mounted in DPX (BDH). Control sections were incubated with either second or third layers alone, and no staining was observed.

The intensity of the p53 nuclear staining on neoplastic cells was scored arbitrarily as follows: very strong and uniform, +++; strong and uniform, ++; weak, f ; equivocal, t; and negative, -. The percentage of positive cells was evaluated on 2 occasions by 3 examiners (MP, GK, GWHS) by scoring at least 1,000 cells. In 10 cases there was a difference of one grade

OVER-EXPRESSION OF p53 IN COLORECTAL ADENOMAS 685

between the examiners, and consensus was reached following further review.

Proliferating-cell nuclear antigen (PCNA) was detected using the mouse MAb PC10, which was previously raised using genetically engineered rat PCNA (Waseem and Lane, 1990). PClO MAb was applied in a 1/10 dilution of the hybridoma supernatant in PBS with overnight incubation to sections previously incubated with normal rabbit serum. The ABC immunoperoxidase technique was used as before, except that biotin-conjugated rabbit anti-mouse immunoglobulin was em- ployed as second layer. The percentage of PC10-positive tumour cells was assessed by counting at least 1,000 nuclei, and divided into 3 groups: strong, moderate to weak, and negative. The average distribution of PC10-positive cells within each sub-group was also recorded in a minimum of 5 randomly selected high-power fields, using a 1 0 0 ~ objective lens im- mersed in oil, by 3 examiners (MP, GK, GWHS). In addition, the index of Pclo-positive cells was independently assessed in those areas of the tumours showing maximum p53 staining.

FIGURE 2 - Moderately dysplastic tubulo-villous adenoma show- ing strong and widespread p53 nuclear staining (ABC; bar = 150 w ) .

RESULTS

The results of the immunohistochemical staining in the adenomas and adenocarcinomas are summarized in Tables I and 11. The anti-p53 polyclonal antibody CM1 did not show any specific staining in 4 normal large-bowel mucosae and in 5 metaplastic polyps examined. Seven out of 11 adenocarcino- mas (65%) showed specific nuclear staining, with a percentage of positive cells ranging from 10% to 85% (Fig. 1). Five of these carcinomas had adenomatous areas at the margin, which stained positively for CM1 in similar proportions. These results are in agreement with previous reports (Van Den Berg et al., 1989; Rodrigues et al., 1990; Scott et al., 1991).

However, 15 out of 36 (41%) adenomas showed a definite positive nuclear staining (Fig. 2 ) , which in 4 cases was very focal and involved less than 0.1% of tumour cells (Figs. 3 and 5b). Interestingly, these cells were isolated rather than clus- tered within the adenomatous tissue, and showed no discern- able morphological differences from their neighbours. Spo- radic staining was seen especially in the villous areas. p53 immunoreactivity was seen in 83% of severely dysplastic adenomas (5/6), with approximately 50% positive nuclei in 3/5 adenomas (cases 32, 35, 36). The strongest staining was encountered in those glands with non-mucin-containing cells and well formed glandular structures (Figs. 2 and 4). In 2 cases, the strongest staining was seen in areas with the most dysplas- tic features. p53 immunoreactivity was seen only in 33% of

FIGURE - Sporadic p53 staining in a moderately dysplastic tubu~o-villous adenoma (ABc; bar = 150 ,,m),

FIGURE 4 - Severely dysplastic adenoma with p53-positive cells, but to a variable degree in some glands (ABC; bar = 150 pm).

adenomas showing mild to moderate dysplasia (10/30). In 3 cases (9,10,20) the p53 staining was consistently weak and was therefore scored as “equivocal” (k). No significant correlation was found between the P53 immunoreactivity and the tubular, tubulovillous or villous histology and/or size of the adenomas,

FIGURE 1 - Moderately differentiated adenocarcinoma, positive for p53 nuclear oncoprotein (ABC; bar = 150 pm).

686 PIGNATELLI ET AL.

which are also considered prognostic indicators of adenomas with malignant potential (Table I).

In order to correlate the proliferative compartment of the adenomas and adenocarcinomas with the p53 positive tumour cells, sections were stained with the PClO MAb, which recognizes the proliferating-cell nuclear antigen (PCNA) (Wa- seem and Lane, 1990). PClO immunostaining has been shown to identify the proliferative cellular compartment in a variety of conventionally fixed and processed tissues (Hall et al., 1990). A linear correlation between PClO immunostaining and other known parameters of cell proliferation has been demonstrated in normal tissues and some neoplasms (Celis and Celis, 1985; Hall et al., 1990).

In 1 metaplastic polyp, 2 adenomas (1 tubular with mild dysplasia, 1 tubulovillous with moderate dysplasia, both p53- negative) and in 2 adenocarcinomas (both p53-positive), no immunostaining with PClO was observed, attributed to over- fixation of the specimens (Hall et al., 1990). In 4 normal colonic biopsies and 4 metaplastic polyps, PCNA-positive cells were localized in the lower third of the crypts, which represents the normal proliferative compartment. Occasionally cells in the upper and middle crypts showed PCNA staining, probably secondary to the long half-life of the protein, as previously documented (Hall et al., 1990).

In the adenomas there was a higher index of strongly stained cells, and these occupied all of the crypt compartment and surface epithelium in some cases (Fig. 5a), as would be anticipated, although cells were more frequently strongly positive in the crypt bases in tubular adenomas. An index was obtained for the overall tumour (PCNA Av), so that the p53 areas could be re-assessed for the comparative index in that area (PCNA p53'). This was then compared with the areas of maximal PClO staining, both strong (S) and moderate to weak (M), which were independently combined. These figures showed that 11 of the 15 unequivocally positive adenomas (1, 12, 17, 18,19,24,25,26,31,32,36) showed closeiycorrespond- ing areas of strong and moderate staining for PCNA and for the highest p53 labelling (Fig. 5a,b) , and of the 5 severely dysplastic adenomas positive for p53 4 showed maximal PCNA positivity in those areas (31, 32, 35, 36). Interestingly, both of

the equivocally p53-positive adenomas which could be as- sessed (9, 10) showed an average PCNA index. The same relationship was seen for the carcinomas, but the numbers in this series are too small to be significant, and the wider range of positivity made assessment considerably more difficult.

DISCUSSION

We have demonstrated that over-expression of the p53 nuclear protein occurs in colorectal adenomas, as well as in adenocarcinomas of the large bowel as previously shown (Van Den Berg et al., 1989; Rodrigues et al., 1990; Scott et al., 1991). Our results also emphasise that the p53 expression in adeno- mas to a certain extent correlates with their degree of dysplasia but not with other known prognostic indicators of malignant potential, such as size and histological type (Muto et al., 1975). Since the high levels of expression of p53 appear to reflect the presence of point mutations in highly conserved regions of the gene (Rodrigues et al., 1990), our data suggest that the genetic alterations involving the p53 gene occur in pre-malignant lesions, earlier than previously described, and therefore may play a role in the dysplastic progression of adenomas with malignant potential (Muto et al., 1975). However, it would appear that p53 expression per se is insufficient to determine the fully malignant phenotype. In fact the sporadic staining observed in many of the tumour cells, rather than focal clusters of p53 positive cells within an otherwise negative adenoma, may indicate cells in which the initial mutation has occurred. In invasive adenocarcinoma, p53-positive tumour cells will eventually become the predominant cell type, by a process of clonal expansion.

Scott et al. (1991) have shown that, in colorectal adenocarci- noma, p53 over-expression does not correlate with established prognostic indicators such as tumour grade and Duke's stage and does not predict survival. These data further suggest that genetic alterations of p53 do not play a role in the late stages of growth and invasion characteristic of poorly differentiated tumours, but may be involved in the early steps of the adenoma-carcinoma sequence, as suggested by our study. It therefore appears that other genetic events may be required to

FIGURE 5 - (a) Proliferating-cell-nuclear-antigen (PCNA) and (b) corresponding p53 staining, showing some clusters of positive nuclei in a villous area of an adenoma which elsewhere showed few p53-positive cells and a lower PCNA index (ABC; bar = 150 bm).

OVER-EXPRESSION OF p53 IN COLORECTAL ADENOMAS 687

determine the invasion of colorectal tumours. For example, it is more likely that changes in genes encoding for molecules controlling cell-cell and cell-matrix interactions, which are 2 fundamental processes disrupted in malignancy (Fearon et al., 1990; Pignatelli and Bodmer, 1990), are more significant in the late stages of colorectal carcinogenesis.

Van Den Berg et al. (1989) reported that only 8% of tubular adenomas over-express the p53 protein, whereas we report here an overall p53 positivity of 41%. This discrepancy may be related either to the sensitivity of the polyclonal antibody used in our study (CM1) or to the pattern of very patchy positivity (less than 0.1% positive cells) found in 33% of the positive adenomas which could be appreciated only on paraffin- embedded material used in this study. This type of staining pattern has also been observed in breast cancer (BBrtek et al., 1990~).

The relationship between p53 over-expression and cell proliferation as measured by PCNA immunoreactivity is prob- ably not surprising, since wild-type p53 appears to inhibit cell-cycle progression into S-phase (Lane and Benchimol, 1990) and this is accompanied by selective down-regulation of PCNA mRNA and protein expression in vitro (Mercer et al., 1991). Nevertheless, this correlation suggests that mutated p53 gene may be the direct cause of the loss of the normal proliferative control seen in colorectal adenomas.

The p53 gene has been shown to be the most commonly mutated oncogene in a variety of human malignancies, such as small-cell lung carcinoma (Iggo et al., 1990), astrocytoma (James et al., 1989), hepatocellular (Bressacet at., 1991), breast (BBrtek et al., 1990a, b; Davidoff et al., 1991), ovarian (Marks et al., 1991) and colonic adenocarcinomas (Rodrigues et al., 1990), and in the majority of cases the mutations correlated with over-expression of the p53 nuclear protein. We have found that pre-malignant lesions, such as dysplastic adenomas

of the large bowel, over-express the p53 protein, suggesting that these changes may be responsible for the susceptibility to adenoma development and progression to carcinomas. Interest- ingly, Shirasawa et al. (1991) have detected p53 mutations also in 3 adenomatous polyps from patients with adenomatous polyposis coli (APC) (who have hundreds of adenomas with 100% susceptibility to the development of colorectal cancer). It has been postulated that this may be an interactive effect of the product of the FAP gene on chromosome 5q21 with p53 protein or other tumour-suppressor genes which are mutated in colorectal tumours (Fearon and Vogelstein, 1990).

Several reports have shown that high levels of p53 in tumour and tumour-cell lines appears to be due to a post-translational stabilization secondary to mutations of the p53 protein coding sequence (Finlay et al., 1988; Hinds et al., 1989). However, it cannot be excluded that changes in the cellular environment or complexing to other proteins specifically expressed by tumour cells but not by normal colonic epithelial cells may also contribute to stabilization of the protein, with its subsequent over-expression. Obviously, it will be of great interest to evaluate whether over-expression of the p53 protein in colonic adenomas is indeed due to the presence of a point mutation in the p53 gene and whether alterations involving the newly identified FAP gene (Groden et al., 1991; Kinzler et al., 1991), thought to be involved in the early stages of colorectal carcinogenesis, are found in the sporadic adenomas with malignant potential.

ACKNOWLEDGEMENTS

Support for this research was provided in part by the Imperial Cancer Research Fund and by a grant to MP and GWHS (Grant RC/188) from the Hammersmith and Queen Charlotte’s Special Health Authority.

REFERENCES

BAKER, S.J., MARKOWITZ, S., FEARON, E.R., WILLSON, J.K.V. and VOGELSTEIN, B., Suppression of human colorectal carcinoma cell growth by wild-type p53. Science, 249,912-915 (1990). BARTEK, J., BARTKOVA, J., VOJTESEK, B., STASKOVA, Z., REJTHAR, A., K O V A ~ ~ K , J. and LANE, D.P., Patterns of expression of the p53 tumour suppressor in human breast tissues and tumours in situ and in vitro. Int. J. Cancer, 46,839-844 (1990~). BARTEK, J., IGGO, R., GANNON, J. and LANE, D.P., Genetic and immunochemical analysis of mutant p53 in human breast-cancer cell lines. Oncogene, 5,893-899 (19906). BARTKOVA, J., BARTEK, J., L u ~ s , J., VOJTESEK, B., STASKOVA, Z., REJTHAR, A,, K O V A ~ ~ K , J., MIDGLEY, C. and LANE, D.P., p53 protein alterations in human testicular cancer including pre-invasive intratubu- lar germ-cell neoplasia. Int. J. Cancer, 49,196-202 (1991). BRESSAC, B., KEW, M., WANDS, J. and OZTURK, M., Selective G to T mutations of p53 gene in hepatocellular carcinoma from southern Africa. Nature (Lond.), 350,429-431 (1991). CELIS, J.E. and CELIS, A,, Cell-cycle-dependent variations in the distribution of the nuclear protein cyclin proliferating-cell nuclear antigen in cultured cells: subdivision of S phase. Proc. nut. Acad. Sci.

DAVIDOFF, A.M., HUMPHREY, P.A., ICLEHART, J.D. and MARKS, J.R., Genetic basis for p53 over-expression in human breast cancer. Proc. nut. Acad. Sci. (Wash.), 88,5006-5010 (1991). FEARON, E.R., CHO, K.R., NIGRO, J.M., KERN, S.E., SIMONS, J.W., RUPPERT, J.M., HAMILTON, S.R., PREISINGER, A.C., THOMAS, G., KINZLER, K.W. and VOGELSTEIN, B., Identification of a chromosome 18q gene that is altered in colorectal cancers. Science, 247, 49-56 (1990). FEARON, E.R. and VOGELSTEIN, B., A genetic model for colorectal tumorigenesis. Cell, 61,759-767 (1990). FINLAY, C.A., HINDS, P.W., TAN, T.-H., ELIYAHU, D., OREN, M. and LEVINE, A.J., Activating mutations for transformation by p53 produce a gene product that forms on hsp 70-p53 complex with an altered half-life. Mot. cell. Biol., 8,531-539 (1988).

(Wash.), 82,3262-3266 (1985).

GRODEN, J., THILIVERIS, A., SAMOWITZ, W., CARLSON, M., GELBERT, L., ALBERTSEN, H., JOSLYN, G., STEVENS, J., SPIRIO, L., ROBERTSON, M., SARGEANT, L., KRAPCHO, K., WOLFF, E., BURT, R., HUGHES, J.P., WARRINGTON, J., MCPHERSON, J., WASMUTH, J., LE PASLIER, D., ABDERRAHIM, H., COHEN, D., LEPPERT, M. and WHITE, R., Identification and characterization of the familial adenomatouspolyp- osis coli gene. Cell, 66,589-600 (1991). HALL, P.A., LEVISON, D.A., WOODS, A.L., Yu, C.C.-W., KELLOCK, D.B., WATKINS, J.A., BARNES, D.M., GILLETT, C.E., CAMPLEJOHN, R., DOVER, R., WASEEM, N.H. and LANE, D.P., Proliferating-cell-nuclear- antigen (PCNA) immunolocalization in paraffin sections: an index of cell proliferation with evidence of deregulated expression in some neoplasms. J. PathoZ., 162,285-294 (1990). HINDS, P., FINLAY, C. and LEVINE, A.J., Mutation is required to activate the p53 gene for co-operation with the rus oncogene and transformation. J. Virol., 63,739-746 (1989). IGGO, R., GATTER, K., BARTEK, J., LANE, D. and HARRIS, A.L., Increased expression of mutant forms of p53 oncogene in primary lung cancer. Lancet, I, 675479 (1990). JAMES, C.D., CARLBOM, E., NORDENSKJOLD, A., COLLINS, V.P. and CAVANEE, W.K., Mitotic recombination of chromosome 17 in astrocy- toma. Proc. nut. Acad. Sci. (Wash.), 86,2858-2862 (1989). JASS, J.R., Do all colorectal carcinomas arise in pre-existing adeno- mas? WorZdJ. Surg, 13,45-51(1989). . . KINZLER, K.W., NILBERT, M.C., Su, L.-K., VOGELSTEIN, B., BRYAN, T.M., LEVY, D.B., SMITH, K.J., PREISINGER, A.C., HEDGE, P., MCKECH- NIE, D., FINNIEAR, R., MARKHAM, A., GROFFEN, J., BOGUSKI, M.S., AL,TSCHUL, S.F., HORII, A,, ANDO, H., MIYOSHI, Y., MIKI, Y., NISH- ISHO, I. and NAKAMURA, Y., Identification of FAP locus genes from chromosome 5q21. Science, 253,661-665 (1991). LANE, D.P. and BENCHIMOL, S . , p53: oncogene or anti-oncogene? Genes Develop., 4, 1-8 (1990). MARKS, J.R., DAVIDOFF, A.M., KERNS, B.J., HUMPHRIES, P.A., PENCE, J.C., DODGE, R.K., CLARKE-PEARSON, D., IGLEHART, J.D., BAST, R.C. and BERCHUCK, A,, Over-expression and mutation of p53 in epithelial ovarian cancer. Cancer Rex, 51,2979-2984 (1991).

688 PIGNATELLI ET AL.

MATHEWS, M.B., BERNSTEIN, R.M., FRANZA, B.R. and CARRELS, J.I., Identity of the proliferating-cell nuclear antigen and cyclin. Nature (Land.), 303,374-376 (1984). MERCER, W.E., SHIELDS, M.T., LIN, D., APPELLA, E. and ULLRICH, S.J., Growth suppression induced by wild-type p53 protein is accompa- nied by selective down-regulation of proliferating-cell nuclear antigen expression. Proc. nut. Acad. Sci. (Wash.), 88,1958-1962 (1991). MUTO, T., BUSSEY, H.J.R. and MORSON, B.C., The evolution of cancer of the colon and rectum. Cancer, 36,2251-2270 (1975). PARADA, L.F., LAND, H., WEINBERG, R.A., WOLF, D. and ROTTER, W., Co-operation between gene-encoding p53 turnour antigen and ras in cellular transformation. Nature (Lond.), 312,649-651 (1984). PIGNATELLI, M. and BODMER, W.F., Integrin cell adhesion molecules and colorectal cancer. J. Pathol., 162,95-97 (1990). RODRIGUES, N.R., ROWAN, A,, SMITH, M.E.F., KERR, I.B., BODMER, W.F., GANNON, J.V. and LANE, D.P., p53 mutations in colorectal cancer. Proc. Nut. Acad. Sci. (Wash.), 87,7555-7559 (1990).

Scorn, N., SAGAR, P., STEWART, J., BLAIR, G.E., DIXON, M.F. and QUIRE, P., p53 in colorectal cancer: clinicopathological correlation and prognostic significance. Brit. J. Cancer, 63,317-319 (1991). SHIRASAWA, S., URABE, K., YANAGAWA, Y., TOSHITANI, K., IWAMA, T. and SASAZUKI, T., p53 gene mutations in colorectal turnours from patients with familialpolyposis coli. Cancer Rex, 51,2874-2878 (1991). VAN DEN BERG, F.M., TIGGES, A.J., SCHIPPER, M.E.I., DEN HARTOG- JAGER, F.C.A., KROES, W.G.M. and WALBOOMERS, J.M.N., Expres- sion of the nuclear oncogene p53 in colon turnours. J. Pathol., 157, 193-199 (1989). VOGELSTEIN, B., FEARON, E.R., HAMILTON, S.R., KERN, S.E., PREIS- INGER, A.C., LEPPERT, M., NAKAMURA, Y., WHITE, R., SMITS, A.M. and BOS, J.L., Genetic alterations during colorectal tumour develop- ment. New Engl. J. Med., 319,525-532 (1988). WASEEM, N.H. and LANE, D.P., Monoclonal antibody analysis of the proliferating-cell nuclear antigen (PCNA). Structural conservation and the detection of the nucleolar f0rrn.J. CellSci., 96,121-129 (1990).