Carcinogenesis vol.21 no.8 pp.1623–1627, 2000
SHORT COMMUNICATION
Genomic instability-based transgenic models of prostate cancer
Christina Voelkel-Johnson, Dale J.Voeks2, Norman M.Greenberg3, Roberto Barrios4, Frideriki Maggouta, David T.Kurtz1, David A.Schwartz, Gina M.Keller, Thomas Papenbrock5, Gary A.Clawson6 and James S.Norris7 Department of Microbiology and Immunology and 1Department of Pharmacology, Medical University of South Carolina, 173 Ashley Avenue, BSB 201, PO Box 250504, Charleston, SC 29425, 2Oncology Research Center, Prince of Wales Hospital, Randwick, NSW 2031, Australia, 3Department of Cell Biology and 4Department of Pathology, Baylor College of Medicine, Houston, TX 77030, 5School of Biological Sciences, Division of Cell Sciences, University of Southampton, Southampton SO16 7PX, UK and 6Department of Pathology, Pennsylvania State University, Hershey, PA 17033, USA 7To
whom correspondence should be addressed Email:
[email protected]
To develop animal models that represent the broad spectrum of human prostate cancer, we created transgenic mice with targeted prostate-specific expression of two genes (EcoRI and c-fos) implicated in the induction of genomic instability. Expression of the transgenes was restricted to prostate epithelial cells by coupling them to the tissuespecific, hormonally regulated probasin promoter (PB). The effects of transgene expression were examined histologically in prostate sections at time points taken from 4 to 24 months of age. The progressive presence of regions of mild-to-severe hyperplasia, low- and high-grade prostatic intra-epithelial neoplasia, and well-differentiated adenocarcinoma was observed in both PBEcoRI lines but no significant pathology was detected in the PBfos line. Prostate tissue of PBEcoRI mice was examined for expression of p53, proliferating cell nuclear antigen (PCNA) and Ki67 at multiple time points. Although p53 does not appear to be mutated, levels of PCNA and Ki67 are elevated and correlate with the severity of the prostatic lesions. Overall, pre-neoplastic and neoplastic stages represented in the PBEcoRI model showed similarity to corresponding early stages of the human disease. This genomic instability-based model will be used to study the mechanisms involved in the early stages of prostate carcinogenesis and to investigate the nature of subsequent events necessary for the progression to advanced disease. Prostate cancer is the most commonly diagnosed malignancy and the second leading cause of cancer death in American men but its etiology is poorly understood. One in five American males will eventually develop the disease but the high rate of clinical incidence is exceeded by the prevalence of prostate carcinomas undetected until autopsy (1,2). Although several Abbreviations: PB, probasin promoter; PCNA, proliferating cell nuclear antigen; PIN, prostatic intra-epithelial neoplasia; TRAMP, transgenic adenocarcinoma mouse prostate. © Oxford University Press
genetic alterations have been documented (3), critical early molecular events involved in neoplastic initiation and progression from latent prostate cancer to the aggressive form of clinical disease remain poorly understood. Therefore, animal models that faithfully represent the broad spectrum of the human disease are required. A number of animal models for prostate cancer have been developed (4) but many lack a natural environment and characterization of early and late stages is not possible. The most successful mouse models to date utilize targeted expression of the SV40 early genes (T/t antigens) that function as oncoproteins, in part, by binding to and interfering with the p53 and/or Rb tumor suppressor proteins or by interfering with protein phosphatase 2A activity (5). In an attempt to create a model for early prostate cancer, we generated transgenic mice in which the restriction enzyme EcoRI or the proto-oncogene c-fos were expressed from a truncated rat probasin promoter. This promoter has been shown to express the SV40 early genes in an androgen-induced, developmentally regulated manner, which is restricted to prostate epithelial cells (6,7). Selection of the transgenes was based on their potential to induce genomic instability. Exposure to ionizing radiation increases the incidence of cancer presumably through the direct induction of cellular genomic instability by generation of a number of DNA lesions (8). The restriction enzyme EcoRI is a prime candidate to use as a tool to model the genomic instability-inducing effects of DNA damaging agents, due to its high activity and nuclear localization (9,10). Induction of genomic instability may not always require direct damage to DNA and may also be caused by early response genes, such as c-fos, that are triggered by a variety of external stimuli including ionizing radiation (11). Overexpression of c-fos leads to chromosomal aberrations, gene mutations and recombination (12,13) and, therefore, seems to play a role in genomic instability. In this study, three lines of transgenic mice were generated by single-cell pronuclear microinjection. The transgenes were constructed by inserting the cDNA coding for EcoRI (Jasper Rine, University of California, Berkeley, CA) or c-fos between the (–426 to ⫹28) rat probasin gene promoter (Robert J.Matusik, Vanderbilt University, Nashville, TN) and an RNA cleavage and polyadenylation signal derived from SV40 to generate PBEcoRI and PBfos plasmids, respectively. Founder offspring were screened for the presence of the transgene by Southern blot and in subsequent generations by PCR. PBEcoRI lines 22L and 25N each were single copy and PBfos line 38L contained four copies of the transgene (data not shown). Mice were killed at age intervals from 4 to 24 months. Tissues collected at necropsy were halved and frozen in liquid nitrogen for protein purification, or immersed in 10% phosphatebuffered formalin for standard tissue processing. To demonstrate transgene expression, we analyzed prostates of transgenic 1623
C.Voelkel-Johnson et al.
animals by western blot and immunohistochemistry. For western blotting, protein was prepared in RIPA buffer containing leupeptin (1 µg/ml) and PMSF (100 µg/ml), separated on 8% Tris–glycine gels (Novex), transferred to nitrocellulose and probed with antibodies as indicated. Detection was performed with ECL (Amersham) or Supersignal (Pierce). Androgen-dependent PB-driven expression parallels sexual maturation with activity increasing 70-fold from 2 to 7 weeks of age (7). Protein harvested from prostate tissue of ~6-monthold PBEcoRI transgenic mice from the 22L and 25N lines was analyzed by western blot using an EcoRI-specific antibody. EcoRI was detected in both lines (Figure 1A). It was not possible to extend this analysis to tissue localization, as the EcoRI antibody is not functional in immunohistochemistry. Western blot analysis of prostate protein from the PBfos line at ~4 months of age demonstrated increased levels of c-fos expression compared with the non-transgenic control (Figure 1B). Elevated c-fos expression in the dorsolateral lobe of the PBfos line was also shown using immunohistochemical staining (Figure 1C and D). Human prostate cancer generally proceeds through low- and high-grade prostatic intra-epithelial neoplasia (PIN) to prostate adenocarcinoma (carcinoma in situ and invasive carcinoma); however, multiple grades of adenocarcinoma often coexist with PIN lesions in the same samples (14). Histological examination of dorsolateral prostate tissue from PBEcoRI and PBfos lines from 4 to 24 months of age showed the progressive presence of regions of mild-to-severe hyperplasia, low-grade PIN and, in the case of PBEcoRI animals, high-grade PIN and well-differentiated adenocarcinoma. Lesions were characterized according to a grading system, recently described for the transgenic adenocarcinoma mouse prostate (TRAMP) model of murine prostate cancer (15). In the PBfos line, nuclear atypia was observed beginning at 10 months of age and lowgrade PIN was detected at 16 months. However, occasional low-grade PIN with stratified epithelium was also observed in non-transgenic animals and probably reflects normal aging in the genetic background of FVB mice. Lesions in PBfos animals failed to progress to high-grade PIN or carcinoma. In PBEcoRI, low-grade PIN was detectable at 10 months and progressed to high-grade PIN and well-differentiated carcinoma, as characterized by invasion of epithelial cells through the basal membrane into the surrounding stroma, by 24 months (Figure 2). General features of the prostate epithelial cells and the prostatic acini in the transgenic mice as they progressed in the severity of lesions included loss of organization, nuclear pleiomorphism, hypercellularity, increased nuclear-to-cytoplasm ratio, increased number of apoptotic bodies, appearance of cribriform structures, loss of luminal secretions, hyperchromatic nuclei, presence of mitotic figures and stromal thickening. Each of these features became more pronounced with increasing dysplasia and cancer progression. Beyond determination of transgene expression and histological evaluation of the models, the relative levels of several markers that have been shown to correlate with progression and prognosis in human prostate cancer, including p53 (16), proliferating cell nuclear antigen (PCNA) (17,18) and Ki67 were assessed. We were unable to detect p53 by immunohistochemical staining in the 24-month-old animals, although prostate tissue from TRAMP mice in which p53 is stabilized by SV40 early genes (T/t antigens) resulted in a positive signal (data not shown). PCNA is a 36 kDa molecule that functions as a required cofactor for DNA polymerase in both the S 1624
Fig. 1. Transgene expression in prostate tissue. (A) EcoRI western blot. PC-3 cells were transiently transfected with the pMENs plasmid and served as a positive control for expression of EcoRI (pMENs). Protein of prostate tissue harvested from a 20-week-old non-transgenic male (wt), a 25-weekold PBEcoRI transgenic male from the 25N line (25N) and a 22-week-old PBEcoRI transgenic male from the 22L line (22L). An aliquot of 25 µg protein was used for western blotting with an EcoRI antibody (1:100). (B) c-fos western blot. Protein of prostate tissue was harvested from 20 week old non-transgenic (wt) and transgenic males (PBfos). An aliquot of 50 µg protein was used for western blotting with a fos antibody (1:40 dilution, Ab-2 from Calbiochem, LaJolla, CA). The endogenous c-fos protein in wild-type animals was detected following prolonged exposure of the blots. (C and D) Immunohistochemical detection of c-fos expression was performed after permeabilization with 0.1% pepsin in 0.2 N HCl, using a 2.5 µg/ml dilution of the fos antibody Ab-2 and the ABC detection system (Vector Laboratories, Burlingame, CA). Dorsolateral prostate tissue from 23-week-old PBfos transgenic (C) and non-transgenic males (D) were probed with fos antibody. Brown/black-stained nuclei were indicative of fos expression. Sections are shown at 40⫻ and were weakly counterstained with H&E.
phase of the cell cycle and in synthesis associated with repair of damaged DNA (17). Since expression of the EcoRI transgene is expected to induce DNA damage, we also included Ki67 as a marker that is not DNA repair sensitive. To measure PCNA and Ki67 expression, protein was isolated from prostates of transgenic lines and age-matched controls at 4, 10, 16 and 24
Transgenic models of prostate cancer
Fig. 2. Histological analysis of dorsolateral prostate tissue at 24 months. Routine sections (4 µm) from the paraffin-embedded tissue were stained with H&E or Masson trichrome and examined by light microscopy for the presence of prostate cancer. H&E-stained sections from a non-transgenic mouse at 20⫻ (A) and 40⫻ (B) show that glands are composed of columnar epithelial cells with round to oval basolateral nuclei. H&E-stained sections (20⫻) from a PBfos mouse show regions of low-grade PIN (C) and poorly organized cells (arrows pointing right) with associated stromal thickening (arrows pointing left) (D). H&E-stained sections (40⫻) from a PBEcoRI mouse of the 25N line reveals epithelial proliferation and invasion (E) as well as numerous apoptotic bodies (F). H&E-stained section from a PBEcoRI mouse of the 22L line (20⫻) shows generally well organized acini with cribriform structures characteristic of highgrade PIN (G). Masson trichrome-stained section of a PBEcoRI mouse from the 22L line (40⫻) shows invasion (see arrow) through the basal membrane (H).
months. PCNA and Ki67 were either absent or only faintly detectable upon overexposure in the blots of the 4- and 10month-old animals (data not shown). At 16 and 24 months of age, a marked increase in both PCNA and Ki67 was observed (Figure 3). To support the western blot analysis, we also performed immunohistochemistry using the same Ki67 monoclonal antibody. Serial sections of dorsolateral prostate from a 24-month-old PBEcoRI (25N) animal were stained with Ki67 or hematoxylin and eosin (H&E). As shown in Figure 4, Ki67 expression correlates with the more proliferative regions of the tissue.
In an effort to create representative animal models of early human prostate cancer, we generated transgenic mice based on genomic instability as the underlying cause of neoplastic transformation. The number of mutations that accumulate during tumor progression is too large to be explained by the rate of spontaneous mutations; thus, early events in carcinogenesis may be the acquisition of a mutator phenotype resulting from genomic instability (19). The TRAMP model, which is most representative of human prostate cancer, utilizes, in part, the SV40 large T antigen to interfere with p53 function. In general, mutation and loss of p53 function is considered a 1625
C.Voelkel-Johnson et al.
Fig. 3. Expression of markers of proliferation by western blot analysis. An aliquot of 50 µg of prostatic protein from 16 month (A) or 24 month (B) non-transgenic (wt) and PBEcoRI mice was probed with antibodies specific for PCNA (1:1000, clone PC10 from Novocastra), Ki67 (1:100, Novocastra) or β-actin (1:2000, Sigma).
Fig. 4. Immunohistochemical analysis of Ki67 expression. (A) Dorsolateral prostate of a 24-month-old PBEcoRI (25N) male was stained with an antibody specific for Ki67 (1:100) after permeabilization with 0.1% Triton X-100. No counterstain was used. Dark nuclei indicate positive staining for Ki67. (B) H&E stain of an adjacent serial section. All sections shown are at 20⫻.
late-stage event, making its role in the initiation of genomic instability questionable (20). Here we show that expression of the EcoRI transgene leads to the advancement of prostatic lesions in the dorsolateral lobe. Progression in the PBEcoRI model was similar to that observed in the early stages of human prostate cancer development with areas of both lowand high-grade PIN mixed with regions of well-differentiated adenocarcinoma in the same sections of prostate tissue from the more advanced 24-month-old mice. The multifocality and heterogeneity of prostatic lesions obtained in the PBEcoRI model are important to further our understanding of the disease. Increased expression of proliferative markers PCNA and Ki67 paralleled the onset of pathology. However, we failed to detect p53 by immunohistochemistry. Since the frequency of p53 mutations in human prostate cancer increases with advanced disease (21), failure to detect p53 mutations is in agreement with the lack of late-stage tumors in the PBEcoRI model of prostate cancer. Onset of prostate pathology and expression of markers of proliferation was preceded by expression of the transgenes, suggesting a contributive role in the development of lesions. Similar to the preferential effects of radiation on expressed DNA, EcoRI could selectively create DNA lesions at transcriptionally active sites of the genome which are more available for attack due to structural considerations. The most likely 1626
mechanism by which EcoRI initiates neoplastic transformation is by loss of genetic information. This is supported by the observation that the restriction enzymes PstI, PvuII and XbaI, all of which transform mouse embryo cells, cause small deletions (1–36 bp), deletion of DNA fragments between two restriction sites and deletions of large regions spanning the restriction sites (22). This would, however, affect only one allele of any given gene and may explain why even early stage lesions are not observed until the animals are ⬎1 year old. Alterations in genes controlling cell proliferation and differentiation such as c-fos may indirectly lead to genomic instability by encouraging the cell cycle progression that is needed to create permanent mutations (23). Alternatively, neoplastic transformation by c-fos may depend on accumulation of a putative AP-1-induced gene product that has to reach a critical level (24,25). Although c-fos was expressed in the prostate of our PBfos line, the animals did not display any significant pathology. This is in contrast to an earlier study in which transgenic mice expressing c-fos from an MHC class I promoter developed bone tumors (26). However, a subsequent study by the same group revealed that induction of bone tumors by c-fos depended on the presence of a retroviral LTR that had been used to replace the 3⬘-untranslated region of c-fos. Removal of this retroviral LTR still allowed for high levels of c-fos expression, but did not result in tumor formation (27). These results indicate that although c-fos has genome destabilizing effects in cultured cells (12,13), overexpression of c-fos does not destabilize the genome sufficiently in vivo to result in tumor formation. Our PBEcoRI model of prostate cancer may reflect what is known as latent prostate cancer. Although latent prostate cancers harbor a number of chromosomal abnormalities (28), critical events that lead to a more malignant phenotype may not yet have occurred. Similarly, given the length of time before manifestation of the neoplastic phenotype in our transgenic animals, the cancers appear to be slow growing, suggesting the need for additional mutations for further transformation. Therefore, genomic instability in the PBEcoRI model is associated with early events in prostate cancer progression, predisposing to focal disease, but is not sufficient to cause the development of late-stage disease. Nevertheless, the lesions already represented in this model may harbor valuable information on the genetic changes associated with the earlier stages of disease. In the future, the PBEcoRI model may be used as a resource to examine both the earliest events in the development and progression of prostate cancer and the critical missing events that contribute to the onset of advanced disease. Acknowledgements We thank Deanne King and Qiao Fang for technical support, Boo Schmidt for histology assistance, and Dr Margaret Kelly for critical reading of the manuscript. This work was supported by NIH grants CA69596 (to J.S.N.), CA75741 (to C.V.-J.), NIH Prostate Cancer SPORE grant CA58204 (to N.M.G.), National Cancer Institute Specialized Program of Research Excellence CA58204 (to N.M.G.) and Mouse Models of Human Cancer Consortium CA84296 (to N.M.G.).
References 1. Parker,S.L., Tong,T., Bolden,S. and Wingo,P.A. (1997) Cancer Statistics, 1997. CA Cancer J. Clin., 47, 5–27. 2. Breslow,N., Chan,C.W., Dhom,G., Drury,R.A., Franks,L.M., Gellei,B., Lee,Y.S., Lundberg,S., Sparke,B., Sternby,N.H. and Tulinius,H. (1977) Latent carcinoma of prostate of autopsy in seven areas. Int. J. Cancer, 20, 680–688.
Transgenic models of prostate cancer 3. Isaacs,W.B., Bova,G.S., Morton,R.A., Bussemakers,M.J.G., Brooks,J.D. and Ewing,C.M. (1994) Genetic alterations in prostate cancer. Cold Spring Harb. Symp. Quant. Biol., 59, 653–659. 4. Green,J.E., Greenberg,N.M., Ashendel,C.L., Barrett,C.J., Boone,C., Getzenberg,R.H., Henkin,J., Matusik,R., Janus,T.J. and Scher,H.I. (1998) Transgenic and reconstitution models of prostate cancer. The Prostate, 36, 59–63. 5. Ludlow,J.W. (1993) Interactions between SV40 large-tumor antigen and the growth suppressor proteins pRB and p53. FASEB J., 7, 866–871. 6. Greenberg,N.M., DeMayo,F., Finegold,M.J., Medina,D., Tilley,W.D., Aspinall,J.O., Cunha,G.R., Donjacour,A.A., Matusik,R.J. and Rosen,J.M. (1995) Prostate cancer in a transgenic mouse. Proc. Natl Acad. Sci. USA, 92, 3439–3443. 7. Greenberg,N.M., DeMayo,F.J., Sheppard,P.C., Barrios,R., Lebovitz,R., Finegold,M., Angelopoulou,R., Dodd,J.G., Duckworth,M.L., Rosen,J.M. and Matusik,R.J. (1994) The rat probasin gene promoter directs hormonally and developmentally regulation of a heterologous gene specifically to the prostate in transgenic mice. Mol. Endocrinol., 8, 230–239. 8. Kronenberg,A. (1994) Radiation-induced genomic instability. Int. J. Rad. Biol., 66, 603–609. 9. Barnes,G. and Rine,J. (1985) Regulated expression of endonuclease EcoRI in Saccharomyces cerevisiae: nuclear entry and biological consequences. Proc. Natl Acad. Sci. USA, 82, 1354–1358. 10. Morgan,W.F., Fero,M.L., Land,M.C. and Winegar,R.A. (1988) Inducible expression and cytogenetic effects of the EcoRI restriction endonuclease in Chinese Hamster Ovary cells. Mol. Cell. Biol., 8, 4204–4211. 11. Woloschak,G.E. and Chang-Liu,C.-M. (1990) Differential modulation of specific gene expression following high- and low-LET radiations. Radiat. Res., 124, 183–187. 12. Van den Berg,S., Kaina,B., Rahmsdorf,H.J., Ponta,H. and Herrlich,P. (1991) Involvement of fos in spontaneous and ultraviolet light induced genetic change. Mol. Carcinog., 4, 460–466. 13. Van den Berg,S., Rahmsdorf,H.J., Herrlich,P. and Kaina,B. (1993) Overexpression of c-fos increases recombination frequency in human osteosarcoma cells. Carcinogenesis, 14, 925–928. 14. Bostwick,D.G. (1995) High grade prostatic intraepithelial neoplasia: the most likely precursor of prostate cancer. Cancer Suppl., 75, 1823–1836. 15. Gingrich,J.R., Barrios,R.J., Fooster,B.A. and Greenberg,N.M. (1999) Pathologic progression of autochthonous prostate cancer in the TRAMP model. Prostate Cancer Prostat. Dis., 2, 70–77.
16. Shurbaji,M.S., Kalbfleisch,J.H. and Thurmond,T.S. (1995) Immunohistochemical detection of p53 protein as a prognostic indicator in prostate cancer. Hum. Pathol., 26, 106–109. 17. Ikido,H.A. (1996) Expression of proliferating cell nuclear antigen in nodenegative human prostate cancer. Anticancer Res., 16, 2607–2612. 18. Harper,M.E., Glynne-Jones,E., Goddard,L., Wilson,D.W., Matenhelia,I.G., Peeling,W.B. and Griffiths,K. (1992) Relationship of proliferating cell nuclear antigen (PCNA) in prostatic carcinomas to various clinical parameters. The Prostate, 20, 243–253. 19. Loeb,L.A. (1991) Mutator phenotype may be required for multistage carcinogenesis. Cancer Res., 51, 3075–3079. 20. Fishel,R. (1996) Genomic instability, mutators and the development of cancer: is there a role for p53? J. Natl Cancer Inst., 88, 1608–1609. 21. Heidenberg,H.B., Sesterhenn,I.A., Gaddipati,J.P., Weghorst,C.M., Buzard,G.S., Moul,J.W. and Srivastava,S. (1995) Alteration of the tumor suppressor gene p53 in a high fraction of hormone refractory prostate cancer. J. Urol., 154, 414–421. 22. Winegar,R.A., Lutze,L.H., Rufer,J.T. and Morgan,W.F. (1992) Spectrum of mutations produced by specific types of restriction enzyme-induced double-strand breaks. Mutagenesis, 7, 439–445. 23. Cohen,S.M. and Ellwein,L.B. (1991) Genetic errors, cell proliferation and carcinogenesis. Cancer Res., 51, 6493–6505. 24. Miao,G.G. and Curran,T. (1994) Cell transformation by c-fos requires an extended period of expression and is independent of the cell cycle. Mol. Cell. Biol., 14, 4295–4310. 25. Wang,Z.-Q., Liang,J., Schellander,K., Wagner,E.F. and Grigoriadis,A.E. (1995) c-fos induced osteosarcoma formation in transgenic mice: cooperativity with c-jun and the role of endogenous c-fos. Cancer Res., 55, 6244–6251. 26. Ruether,U., Komitowski,D., Schubert,F.R. and Wagner,E.F. (1986) c-fos expression induces bone tumors in transgenic mice. Oncogene, 4, 861–865. 27. Ruther,U. (1998) Induction of bone tumors in transgenic mice by C-FOS depends on the presence of a retroviral long terminal repeat. Cancer Genet. Cytogenet., 105, 123–127. 28. Erbersdobler,A., Bardenhagen,P. and Henke,R.-P. (1999) Numerical chromosomal anomalies in latent adenocarcinomas of the prostate. The Prostate, 38, 92–99. Received January 14, 2000; revised April 17, 2000; accepted May 1, 2000
1627