Accumulation of p53 protein in chemically induced oval cells during early stages of rodent hepatocarcinogenesis. U.Wirnitzer, ff.Enzmann, M.Rosenbruch1 and.
Carcinogenesis vol.16 no.4 pp.697-701, 1995
Accumulation of p53 protein in chemically induced oval cells during early stages of rodent hepatocarcinogenesis
U.Wirnitzer, ff.Enzmann, M.Rosenbruch1 and E.M.Bomhard Institute of Industrial Toxicology and 1 Institute of Toxicological Pathology, Pharma Research Center, Bayer AG, 42096 Wuppertal, Germany
Up to now the possible involvement of p53 in rodent cancerogenesis has been based on results of the endpoint of chemically or virally induced carcinogenesis-tumors. To address the role of altered p53 expression in different stages of the multi-step process of rodent carcinogenesis in a systematic way we fed potent chemical carcinogens to male rats for 6, for 12 and for 6 weeks followed by a 6 week recovery period. Assessment of alterations of p53 expression was performed by immunoperoxidase staining with a polyclonal antiserum on frozen liver sections. Positive p53-immunostaining was localized to treatment-induced proliferating oval cells on liver sections of 21/21 2-acetylaminofluorene- (AAF) and 19/21 N-nitrosomorpholine (NNM)-treated rats irrespective of application scheme as weil as to foci of hepatocytes in 1/21 NNM-treated animals and in 3/21 AAF-treated animals after 6 weeks of treatment only. The induction of oval cell proliferation by AAF was more pronounced than by NNM, and for NNM appeared to be dependent on application scheme, with a similar lower abundance of oval cells after a 6 week treatment with and without recovery as compared to a 12 week treatment. These results are discussed with respect to the role of p53 in human and rodent carcinogenesis on the one band, and the disputed function of oval cells as facultative liver stem and tumour progenitor cells on the other. Introduction The wild-type tumour suppressor gene p53 is involved in cell cycle control, DNA repair and synthesis, differentiation and apoptosis, and its aberrant expression has been demonstrated in the majority of human malignancies ( 1--4 ). p53 is evolutionarily highly conserved on the DNA sequence and protein function level as the 41 amino acids in the conserved regions are 93% identical in mammals, whereas the overall homology approximates only 55% (5). Recent results suggest a role for altered p53 expression in rodent carcinogenesis (for review see 6), which in turn is usually the basis for human cancer risk assessment. Thus, studies on the alteration of p53 expression in chemically induced rodent tumors are hypothesized to have a direct impact on the understanding of the multi-step process of carcinogenesis in humans. Many of the known p53 alterations have been identified as genetic missense mutations resulting in a prolongation of the half-life of the corresponding gene product, probably due to conformational changes as demonstrated by distinct antigenicity. Due to a half-life of up to 20 min the wild-type p53 protein does not normally accumulate in amounts detectable *Abbreviations: AAF, 2-acetylaminofluorene; NNM, N-nitrosomorpholine; GST, glutathione S-transferase; FAH, foci of enzyme-altered hepatocytes.
© Oxford University Press
by routine immunohistochemistry. In contrast half-lives of several hours have been documented for mutant p53 proteins resulting in intracellular p53 protein concentrations high enough for immunodetection (7) as seen in many different human and rodent tumors (6). A number of monoclonal antibodies have been developed during recent years, though these have several disadvantages: (i) they have specificity that is defined, but confined to only one epitope; (ii) most of them are specific to human p53 protein; and (iii) with one exception (D07) they do not perform well on fixed and paraffinembedded tissue, which represents the routinely available material in animal experiments. For these reasons we decided to use a recently generated polyclonal antibody raised against recombinant human wild-type p53, CMl (8), with a known specificity against murine and rat p53 protein (C.C.Harris, NCI, personal communication). In order to investigate the potential role of p53 in rodent cancerogenesis in a systematic way during the early steps of tumour development we applied the potent hepatocancerogenic chemicals 2-acetylaminofluorene (AAF*) and N-nitrosomorpholine (NNM) in different regimens to rats; these compounds are known to induce a multitude of nodular changes and alterations considered to be preneoplastic. In this communication we present AAF- and NNM-induced oval cell proliferation and demonstrate elevated levels of p53 protein in these cells.
Materials and methods Experimental animals The study was performed in male SPF-bred Wistar rats of the strain Hsd/ Win:WU kept individually (9) under optimized hygienic conditions. Groups of seven animals were treated with a pelleted fixed-formula standard diet (Altromin GmbH, Lage) containing 200 p.p.m. AAF or 100 p.p.m. NNM for 6 weeks, for 12 weeks and for 6 weeks followed by a 6 week recovery period respectively. Feed and water were available ad libitum. Control animals were kept for 6 or 12 weeks under the same conditions and on the above-mentioned standard diet. Tissue preparation Animals were killed by exsanguination in deep ether anaesthesia and dissected. For immunohistochemistry liver samples were shock-frozen, submersed in 2-methylbutane at about -120°C and stored at -80°C. Cryostat sections of 6 µm were prepared at about - l 7°C (Slec MTE) and transferred to 'super frost plus' slides (Menzel), fixed for 10 min in acetone at room temperature, air-dried, wrapped in aluminium foil and stored at -20°C (for up to 2 months) or -80°C (for longer than 2 months). Immunohistochemistry After inactivation of tissue peroxidase activities (0.5% H202) and blocking of unspecific antibody cross-reactivity (10% normal goat serum, Dako NS, Glostrup, Denmark) sections were incubated for p53 protein detection overnight at 4°C with a 1:1000 dilution of a polyclonal rabbit serum, CM! (8, Medac, Hamburg, Germany) raised against full-length wild-type human p53 protein produced in a bacterial expression system. Primary antibody binding was detected by a horseradish peroxidase-conjugated goat anti-rabbit secondary antibody used according to the manufacturer's specifications (Dako). Diaminobenzidine (final concentration 0.5 mg/ml) with nicke! sulfate (final concentration 0.2 mg/ml) was used as chromogen, resulting in black granular precipitates. The counterstain was 'Kemechtrot' (5% Al2S04, 0.1 % 'Kemechtrot', Bayer AG, Leverkusen, Germany, aqueous solution). For positive control, A431-a human epidermoid carcinoma cell line (ATCC, Rockville, MD, USA), grown on chamber slides according to the specifications of the supplier-which
697
U.Wirnitzcr et al.
'fable L p5J-�pecific immunostuining after AAF and NNM trearment Cell type AAF 6
AAf 6-6
AAF 12
Nl\'M 6
Hepatocytc Oval cells
3/7
+
717 7 717 4-
7/7
717
+
NNM 6+6
717 -
1/7
+
.m +
+
717
317 +/-
+
NNM 12
717 717 -
717 -
6 ( 12) = 6 (12) wceks treatment. 6+6 = 6 weeks rrearment folJowed by 6 weeks of recovery. + = p53-specitic immunostaining, - = no signals afrer spccific p53-immunostaining. x/y: y = number of animals in a treatmenr group; x = numbcr of animals with indicated result.
Fig. 3. Foci of p53-positive bepatocytes. NNM trcatmcnt for 6 weeks. Objective magnification 20 X. Fig. 1. NNM treatmeni for 12 weeks. Objective rnagnification l 6X (Arrows indicate oval cells).
Fig. 2. AAF treatment for 6 wccks. Objective rnagnification 16X. (Arrows indicatc oval cells ).
Fig. 4. p53-positivc (arrow) and p53-ncgativc (arrowhcad) oval cells. NNM treatrnent for 6 weeks. Objective magnification 20X.
overexpresses consritutively p53 was stained undcr idcntical conditions in each experiment. Results were reproduced at least once.
analysis was performed hy courning the foci of a complete cross-section of
Demonstration of enzyme-altered foci
Glutathione S-transferase (placeatal form. GST1t) was detected by a 2 h lncubation at room ternperature of liver freeze scctions wlth a 1: l 00 dilution of u polyclonal rabbit serum (Oncor, Gairhersburg, PA. USA). Localization of firsr antibody was achieved as described Ior p53 antigen. Glycerol-Svphospherase dehydrogenase (G3PDH) and glucose-ö-phosphate dehydrogenase (G6PDH) activity was demonstrated enzyrne histochemically in freeze seciions transferred to a semipermeable mcmbrane as described ( l 0.11 ). Foci with increased glucose-ö-phosphaie, glycerol-3-phosphate dehydrogenase and GST activity were defined as described (l2. IO.l3). Morphornetric
the [cft liver lobe. Areas of the sections were mcasurcd intcractivcly wiih a videoplan system (Konten, Munich) (10,12). Histology
For histclogical investigations, liver lobes were fixed in buffered 4%, formaldehyde for 24 h and ernbedded in paraffin. Sections (4 µm) were prepared and stained with H&E.
Results There were ciear diffcrcnccs between thc histological changes induced by AAF and NNM. Thus, 6 weeks and-slightly more
698
•
Accumulation of p53 protein in oval cells
Fig. 5. p5J-_posuive oval cells (arrowheads), AAF rrearment for 12 weeks. Objective magnification 5X. Fig, 7. Untreatcd animal. sorne p53-immunostained cholangiocytes (arrow).
Objecrive magnification 5X.
Tuble ß. Morphometric analysis of enzyme-altered liver foci (foci/cm2) Ko 6
Ko 12
AAF6
AAF6+6
AAF 12
SD
0 0
0.23 0.51
3.04 3.07
2.38 3.10
7J.02 30.62
SD
0.19 0.53
0.86 0.82
0.14 0.36
2.81 2.19
51.13 85.29
0 0 0.19
0 1.09
1.35 1.24 4.53
5.52 2.68 10.71
3.42 2.42 125.57
Ko 6
Ko 12
NNM6
NNM 6+6 NNM 12
()
0
0.23 0.51
10.53 21.20
18.90 14.22
62.93 22.33
x SD
0.19 0.53
0.86 0.82
0.57 0.77
12.37 6.16
3.89 3.48
x
()
0 0 1.09
22.09 21.70 33.97
51.49 21.l4 82.76
L26.91 59.00 193.73
G6PDH lt
G3PDH x
GST-P lt
SD I:
G6PDH
x
so
G3PDH Fig. 6. p53-positive oval cclls. NNM treatrnent for 6 weeks rccovery, Objective magnification 20X.
+
6 weeks
pronounced-12 weeks of NNM treatment yielded altered hepatocytes (partly enlarged with prominent nuclei. clear cell like) and nests of oval cells in periportal fields that often formed ring-like structures, resembling biliary duct cells, and only rarely radiated into the lobule (Figure 1). After a 6 week exposure to NNM and a subsequent recovery period. altered hepatocytes and oval cells were less abundant as compared to 6 and 12 weeks of treatrnenr. After AAF administration highly abundant oval cells radiating in rows from periportal fields, only rarely arranged around srnall lumina (as compared to NNM-trealed animals), were seen (Figure 2) in approximately equal abundance in all AAF-treated groups. Additiona1Jy. in the 12 week AAF group, morphoJogically altered hepatocytes were apparent. In untreated animals no oval cells were scen. p5J-specific immunostaining was detected (see Table I) after AAF treatment in all animals irrespective of treatment schedule and after NNM exposure in all animals treated for 6 weeks. in 6n animals treated for 12 wceks and in 6/7 animals treated for 6 weeks followcd by a 6 week recovery. After 6 weeks of trearment this specific p53 detection was localized to a few foci of hepatocytes (Figure 3) in 117 NNM-trealed animals and in 3/7 AAF-treated animals rcspectively, as weil as ro all
GST-P
so I:
0 0.19
()
oval cells except for 3/7 NNM-treated animals that had p5J positive as weil as p53-negative oval cells (Figure 4). In all other treated animals (6 weeks treatment + 6 weeks rccovery, 12 weeks trcatrnent) no p53-positive hepatocytes were seen, but a streng p5J-positive immunoreaction was detected in abundant oval cells (Figures 5 and 6). Immuoostaining was mainly nuclear, but a faint reaction in tbe cytoplasm cannot be excluded. Cornpetition with increasing arnounts of recombinant p53 antigen during primary antibody incubation of liver sections confirmed the specificity of the reaction (data not shown).
Untreated control animals showcd slight p53 immunestaining in somc of the much less abundant cholangiocytes (Figure 7). The results of morphometric measurements of foci of enzyme altered hepatocytes (FAHs) are summarized in Table II. indicating the means (x) and standard deviation (SD) for the individual enzymes investigated as weil as tbe sum of the number of all FAHs per treatment. As can be deduced from these data the total number of FAHs (irrespective of pbenotype)
699
• r
U.Wirnitzer et al.
per treatment regimen increased for AAF and NNM with time of exposure but also during recovery as compared to identical ingestion time without recovery. Considering the three parameters separately, a difference in phenotype becomes apparent with respect to the two carcinogens applied and for each treatment regimen. Thus, for example, after treatment for 12 weeks with AAF, most FAHs on liver sections showed G6PDH or G3PDH activity with only a few FAHs positive for GST-P, while after the same time of exposure to NNM most FAH were positive for GST-P, many for G6PDH and only a few for G3PDH activity.
Discussion Induction of oval cell proliferation by high doses of a variety of hepatocarcinogens is a known phenomenon (14,15) and dose dependence of oval cell induction after NNM application has been extensively studied (16). The role of this cell type as a 'facultative stem cell' and during hepatocarcinogenesis is controversial. Results from autoradiographic, morphologic and biochemical analysis in rat liver after carcinogen treatment support a precursor-product relationship between oval cells and hepatocellular carcinoma (17-19) while others demonstrated this for cholangiocarcinoma (20-22). Considering the association of aberrant p53 expression in the majority of human malignancies and its normal function for the integrity of the genome (1-4), our demonstration of the elevated intracellular levels of p53 protein in oval cells in the early stages of rodent hepatocancerogenesis further supports the hypothesis that at least some forms of liver cancer arise from this cell type. The presence of p53-positive as well as p53-negative oval cells in 3/7 NNM-treated animals after 6 weeks of treatment, no p53-negative oval cells but an apparently more intense staining in the other treatment groups seems to suggest a gradual increase in intracellular p53 protein concentration in the course of cancerogenesis. p53-positive hepatocytes were only seen in the 6 week treatment groups, although morphologically altered hepatocytes became more abundant with prolonged NNM exposure or appeared after 12 weeks of AAF exposure. A multitude of phenotypic markers have been investigated for correlation to hepatocancerogenesis (23) out of which we chose G6PDH, G3PDH and GST1t. The total number of FAHs on liver sections increased for both substances with time of exposure, while after the recovery period FAHs were less frequent than after a continuous 12 week treatment, but more frequent than after a 6 week exposure without recovery, possibly indicating the progression of hepatocarcinogenesis. Tue FAH phenotype was distinct for the substances tested and at each endpoint of exposure. No correlation between the induction of oval cell proliferation and the occurrence of FAH could be derived. Together with the results of positive p53 immunodetection in oval cells, we interpret the distinct pattem of focal enzyme alterations seen in hepatocytes as a reversible phenotypic modulation (23) due to acute toxic effects of the administered carcinogens and as discussed in the context of different carcinogenic regimens by others (17). Thus, in some models for hepatocarcinogenesis with severe cytotoxic effects, most nodules of enlarged hepatocytes induced by exposure to carcinogenic chemicals reshape and disappear, while only a few develop into persistent nodules that may provide an environment for malignant transformation of a rare cell type of particular characteristics. lt may be this cell type, for which mutant p53 protein within FAHs could be demonstrated 700
immunohistochemically (24). Nevertheless, the definite identity of this cell type is still elusive. Concluding from the results obtained in our model we postulate that the p53-positive oval cells are progenitors of liver tumours known to be inducible by the applied carcinogens. The immunohistochemical detection of p53 protein on frozen liver sections strongly suggests intracellular accumulation due to a missense mutation (25) before the start of or in an early step of oval cell proliferation, but definite evidence can only be provided by polymerase chain reaction amplification and subsequent sequencing of DNA from single p53-positive oval cells. Although excess intracellular p53 protein, irrespective of the mechanism, has been demonstrated in humans as a distinct marker of preinvasive lesions and invasive tumours (26), the possible role of p53 in apoptosis and repair after DNA damage inducing a transient p53-positive immunohistochemistry should be kept in mind. Thus, immunohistochemically detectable amounts of p53 after UV exposure in the stem cell compartment of normal skin (27) and after y-irradiation in the stem cell compartment of different segments of the gastrointestinal tract correlating with differential sensitivity to p53-dependent apoptosis have recently been demonstrated (28). However, the immunohistochemically p53-positive phenotype after both UV- and y-irradiation was only transient (in the range of hours). Thus, the demonstration of increased p53 protein levels in an early stage of chemically induced rodent hepatocarcinogenesis furthers the hypothesis of a role of altered p53 expression in tumour development in rodents. Studies on other substances and tissues are under way in order to establish altered p53 expression as an early marker for rodent cancerogenesis, which due to the high evolutionary conservation of p53 on the DNA sequence and protein function level will have a direct impact on human cancer risk assessment.
References 1. Hollstein,M., Sidransky,D., Vogelstein,B. and Harris,C.C.(1991) p53 mutations in human cancer. Science, 253, 49-53. 2.Levine,A.J., Momand,J. and Finlay,C.A. (1991) The p53 tumor suppressor gene. Nature, 351, 453-456. 3.Harris,C.C. and Hollstein,M.C. (1993) Clinical implications of the p53 tumor suppressor gene. New Engl. J. Med., 18, 1318-1327. 4.0ren,M. (1992) p53: the ultimate tumor supressor gene? FASEB J., 6, 3169-3176. 5. Soussi,T., Caron de Fromental,C. and May,P. (1990) Structural aspects of the p53 protein in relation to gene evolution. Oncogene, 5, 945-952. 6.Caamano,J., Ruggeri,B.A. and Klein-Szanto,A.J.P. (1992) A catalog of p53 alterations in selected human and laboratory animal neoplasms. In Klein-Szanto,A.J.P. (ed.), Comparative Molecular Carcinogensis. WileyLiss, New York, pp. 331-355. 7.Shay,J.W., Werbin,H., Funk,W.D. and Wright,W.E. (1992) Cellular and molecular advances in elucidating p53 function. Mutat. Res., 277, 163-171. 8.Midgley,C.A., Fisher,C.J., Bartek,J., Vojtesek,B., Lane,D.P. and Bames,D.M. (1992) Expression of human p53 in bacteria: application to the analysis of p53 expression in human tumours. J. Cell. Sei., 101, 183-189. 9.Spiegel,A. and Gönnert,R. (1961) Neue Käfige für Mäuse und Ratten. Z. Versuchstierkunde, 1, 38-46. 10.Enzmann,H., Ohlhauser,D., Dettler,T., Benner,U., Hacker,H.J. and Bannasch,P. (1989) Unusual histochemical pattem in preneoplastic hepatic foci characterized by hyperactivity of several enzymes. Virchow's Arch., B57, 99-108. 11.Meijer,A.E.F.H. and de Vries,G.P. (1974) Semipermeable membranes for improving the histochemical demonstration of enzyme activities in tissue sections. IV. Glucose-phosphate-dehydrogenase and 6-phosphogluconate dehydrogenase (decarboxylating). Histochemistry, 40, 349-359. 12. Hacker,H.J., Moore,M.A., Mayer,D. and Bannasch,P. (1982) Correlative histochemistry of some enzymes of carbohydrate metabolism in preneoplastic and neoplastic lesions in the rat liver. Carcinogenesis, 3, 1265-1272.
Accumulation of p53 protein in oval cells 13. Sato.K., Kitahara,A.,Satoh,K., Ishikawa,T., Taternatsu, M. and Ito,N. (1984) The placental form of glutathione-S-transferase as a new marker for preneoplasia in rat chemical hepatocarcinogenesis. Jap. J. Cancer Res., 75, 199-202. 14.Sirica,A.E. and Cihla,H.P. (1984) Isolation and partial characterizations of oval and hyperplastic bile ductular cell-enriched populations from the livers of carcinogen and noncarcinogen-treated rats. Cancer Res., 44, 3454-3466. 15. Farber,E. (1955) Similarities in the sequence of early histological changes induced in the liver of the rat by ethionine, 2-acetylaminofluorene, and 3 '-methyl-4-dimethylaminoazobenzene. Cancer Res., 3, 142-148 16. Bannasch,P. ( 1980) Dose dependence of early cellular changes during liver cancerogenesis. Arch. Toxicol. Supp/., 3, 111-128. 17. Sell,S. and Dunsford,H.A. ( 1989) Evidence for the stem cell origin of hepatocellular carcinoma and cholangiocarcinoma. Am. J. Pathol., 134, 1347-1363. 18. Yukio,I. ( 1967) Significance of the so-called oval cell proliferation during azo-dye hepatocarcinogenesis. Gann, 58, 355-366. 19. Faris,R.A., Monfils,B.A., Dunsford,H.A. and Hixson,D.C. (1991) Antigenie relationship between oval cells and a subpopulation of hepatic foci, nodules and carcinomas induced by the 'resistant hepatocyte' model system. Cancer Res., 51, 1308-1317. 20.Cameron,R.G., Blanck,K. and Armstrong,D. (1990) Sex differences in response to four promotion regimens in spite of common first cellular steps in the hepatocellular cancer process initiated by diethylnitrosamine. Cancer lett., 50, 109-113. 21. Steinberg,P., Hacker,H.J ., Dienes,H.P., Oesch,F. and Bannasch,P. (1991) Enzyme histochemical and immunocytochemical characterization of oval and parenchymal cells proliferating in livers of rats fed a choline-deficient/ DL-ethionine-supplemented diet. Carcinogenesis, 12, 225-231. 22.Gerlyng,P., Grotmol,T., Stokke,T., Erikstein,B. and Seglen,P. (1994) Flow cytometric investigations of a possible precursor+ product relationship between oval cells and parenchymal cells in the rat liver. Carcinogenesis, 15, 53-59. 23. Moore,M.A., Weber,E. and Bannasch,P. ( 1988) Modulating influence of dehydroepiandosterone administration on the morphology and enzyme phenotype of dimethylaminoazobenzene-induced hepatocellular foci and nodules. Virchows Arch. B, 55, 337-343. 24.Smith.M.L., Yeleswarapu,L., Locker,J. and Lombardi,B. (1991) Expression of p53 mutant protein(s) in diethylnitrosamine-induced enzyme-altered hepatocytes in male Fischer-344 rats. Carcinogenesis, 12, 1137-1141. 25.Gannon,J.V., Greaves,R., Iggo,R. and Lane,D.P. (1990) Activating mutations in p53 produce a common conformational effect. EMBO J., 9, 1595-1602. 26. Bennett,W.P., Colby,T.V., Travis,W.D., Borkowski,A., Jones,R.T., Lane,D.P., Metcalf,R.A., Samet,J.M., Takeshima,Y., Gu,J.R., Vähäkangas, K.H., Soini,Y., Pääkkö,P., Welsh,J.A., Trump,B.F. and Harris,C.C. (1993) p53 protein accumulates frequently in early bronchial neoplasia. Cancer Res., 53, 4817-4822. 27.Hall,P.A., McKee,P.H., P.Menage,H.du, Dover.R. and Lane,D.P. (1993) High levels ofp53 protein in UV-irradiated normal human skin. Oncogene, 8, 203-207. 28. Merritt,A., Potten,C.S., Kemp,C.J., Hickmann,J.A., Balmain,A., Lane,D.P. and Hall,P.A. (1994) The role of p53 in spontaneous and radiation-induced apoptosis in the gastrointestinal tract of normal and p53-deficient mice. Cancer Res., 54, 614-617. Received on September 15, 1994; revised on January 3, 1995; accepted on January 4, 1995
701