We have generated a transgenic rat with the SV40 T antigen under probasin promoter control, allowing prostate-specific gene expression. Males demonstrate atypical epithelial cell proliferation in the prostate from 4 weeks of age and develop prostate carcinomas at 100% incidence before they are 15 weeks old. Castration at 5 weeks of age was found to inhibit the prostate tumor formation completely, whereas testosterone propionate administration induced marked cell proliferation as well as microinvasion in prostate carcinomas. Castration at 20 weeks of age, after tumor development, even with testosterone propionate treatment, induced complete tumor involution within 5 weeks. To investigate the underling processes, sequential histological changes were monitored 1, 2, 3, 7, 14, and 21 days after castration. At days 1–3, many apoptotic bodies and inflammatory cells, including foam cells, were observed, and clear glandular structures were no longer evident in the tumors. Seven days after castration, most glands were involved, and nuclei of the cells did not show atypia. After 14 and 21 days, only atrophic glands were observed. During this process, expression of caspase 3, caspase 6, BAX, bcl-x, TRPM-2, and MMP7 genes was apparently increased. Comparison of the gene expression profile between a prostate carcinoma in a transgenic animal and a normal prostate of a wild-type rat by a cDNA array technique was also conducted. The results suggested that our model is suitable to investigate mechanisms of carcinogenesis, including androgen dependence, involution, and apoptosis.
Prostate cancer, the most common malignant disease in the Western world, has increased dramatically over the past decade. In the United States, it ranks second only to lung cancer as a cause of cancer related deaths (1) . Despite this serious situation, the basic biology, etiology, and risk factors for prostate cancer remain largely to be elucidated. Several mouse transgenic models have already been established (2) , which may contribute to analyze molecular, cellular, and physiological events in prostate carcinogenesis. However, rats have advantages for this purpose (2, 3, 4) , their larger size, e.g., allowing adequate materials to be obtained. Furthermore, several rat prostate models using chemical carcinogens have been established (5 , 6) , and data for hormone effects and modifying agents have accumulated (7, 8, 9, 10, 11, 12, 13, 14, 15, 16) . The problem is that these models are labor intensive and require long periods to induce tumors, which are usually microscopic and not suitable for molecular biological analysis. Therefore, we have established a rat transgenic prostate cancer model using the probasin gene promoter and the SV40 T antigen gene. The rat probasin gene encodes an androgen- and zinc-regulated protein specific to the dorsolateral epithelium of the prostate (17) . cis-acting androgen-response regions within the 5′ flanking region have been identified (18) , and recently, the ability of the prostate-specific rat probasin gene promoter to target heterologous genes specifically in the prostate of transgenic mice was demonstrated (19 , 20) . In rats also, the gene promoter would be expected to work in the same way because this was the species from which it was originally isolated (17) . The SV40 early-region tumor antigen has the ability to induce transformation in vivo (21) , the SV40 large tumor T antigen (Tag) acting as an oncoprotein through interactions with retinoblastoma (22) and p53 tumor-suppressor gene products (23 , 24) ; the small t antigen interacts with protein phosphatase 2A (25) . SV40 Tag in particular has been used successfully in transgenic mice to induce tumors in a variety of organs, including the prostate (20 , 26 , 27) . It also proved useful in the liver of transgenic rats (28) .
Probasin-SV40 T antigen transgenic rats were here generated by microinjection of recombinant DNA of the probasin gene promoter, fused to the SV 40 T antigen, into pronuclei of Sprague Dawley rat embryos. In this paper, we describe the resultant prostate tumors and their androgen dependence in the transgenic rats, as well as sequential changes in histology and gene expression during involution after castration. Gene expression profiles were also compared between a prostate tumor in a transgenic animal and a normal prostate of a wild-type rat by a cDNA array technique.
MATERIALS AND METHODS
Construction of the Transgene
The rat probasin gene promoter region (−426 to +33; Ref. 19 ) was obtained by PCR using primers GGAATTCAGCTTCCACAAGTGCATTTA and CATGCCATGGAGCTCTGTAGGTATCTGGAC with addition of restriction endonuclease recognition sites for EcoRI and NcoI, respectively. The SV40 Tag gene in pBluescriptII KS(−) (VG042, pBS-SVT) was obtained from Health Science Research Resource Bank (Osaka, Japan). The PCR products for the rat probasin gene promoter region and the SV40 Tag gene in pBluescriptII KS(−) were digested with EcoRI and NcoI and then ligated, resulting in a pBS-Probain-SVT plasmid. To prepare a linear transgene and remove plasmid sequences for microinjection, the plasmid was digested with EcoRI and BamHI, purified by agarose gel electrophoresis, and recovered using a QIAquick gel extraction kit (Qiagen, Tokyo, Japan).
Production and Screening of Transgenic Rats
Generation of transgenic rats was performed by DNX transgenic Science (Princeton, NJ). DNA isolation from rat tails, and the PCR-base screening assay were performed as described previously (3) . Sequences of the PCR primers were 5′-GTCAGCAGTAGCCTCATCAT-3′ and 5′-GGTTGATTGCTACTGCTTCG-3′. Heterologous transgenic males for the studies were routinely obtained by mating heterologous transgenic males and wild-type Sprague Dawley rats (Clea, Tokyo, Japan).
A total of 25 transgenic male rats were divided into five groups. In group 1, five rats were maintained until 25 weeks old, when they were killed. In group 2, five rats were castrated at 20 weeks of age and kept until 25 weeks old, when they were killed. In group 3, five rats were castrated at 5 weeks of age and were kept until 25 weeks. In groups 4 and 5, 1.5-cm long Silastic tubes containing 30 mg of testosterone propionate (Sigma Chemical Co., St. Louis, MO) were implanted into the subcutis of the dorsal region of five animals each at 5 weeks old of age; they were then killed when 25 weeks old. In group 5, castration and removal of the testosterone tubes were performed at 20 weeks of age.
In a total of 21 transgenic male rats, 1.5-cm long Silastic tubes containing 30 mg of testosterone propionate were implanted at 5 weeks of age. Castration and removal of the testosterone tubes were performed at 15 weeks of age, and rats were killed 0, 1, 2, 3, 7, 14, and 21 days thereafter.
Preparation and Analysis of Tissues
One-half of each prostate collected at necropsy was routinely fixed in 10% phosphate-buffered formalin for 48 h and then processed for embedding in paraffin. Five-μm-thick sections were cut and stained with H&E. Immunohistochemical analyses of SV40 Tag, and androgen receptor were performed using mouse anti-SV40 large T antigen monoclonal antibody (PharMingen, San Diego, CA), and rabbit polyclonal anti-androgen receptor antibody (Affinity Bioreagents, Golden, CO). Binding was visualized with a Vectastain Elite ABC kit (Vector Lab, Burlingame, CA) and light hematoxylin counterstaining was conducted to facilitate microscopic examination. Photographs were taken with a digital camera (DP11, Olympus, Tokyo, Japan) and printed out using a digital printer (Pictrography 3000, Fujifilm, Tokyo, Japan). The other half of each prostate was frozen in liquid nitrogen for extraction of RNA.
Extraction of Total RNA, Quantitative RT-PCR, and cDNA Array Analysis
Total RNA extraction and DNase treatment were performed according to the manufacturer’s instructions using the Atlas Pure Total RNA Labeling system (Clontech, Palo Alto, CA). One microgram of the RNA was converted to cDNA with avian myoblastosis virus reverse transcriptase (Takara, Otus, Japan) in 20 μl of reaction mixture. Aliquots of 2 μl of cDNA samples were subjected to quantitative PCR in 20-μl reactions using FastStart DNA Master SYBR Green I and a Light Cycler apparatus (Roche Diagnostics, Mannheim, Germany). Primers used for caspase 3 were 5′-GCCGACTTCCTGTATGCTTA-3′ and 5′-CACGGGATCTGTTTCTTTGC-3′; for caspase 6, 5′-AGACCTTGACTGGCTTGTTC-3′ and 5′-GCTGAGAGACCTTCCTGTTC-3′; for bcl-x, 5′-TACCAGGAGAACCACTACTA-3′ and 5′-AGCTGATCTGAGGAAAAACC-3′; for BAX, 5′-GGAGCTGCAGAGGATGATTG-3′ and 5′-TGAGCGAGGCGGTGAGGACT-3′; for MMP7, 5′-TGGGTCTGGGTCACTCTTCT-3′ and 5′-CACAGCTTGTTCCTCTTTCC-3′; for TRPM-2, 5′-CATGGAATGAGACAGAAGCA-3′ and 5′-GAACCCAGAGGAAGGAGAGG-3′; for cyclophilin, 5′-TGCTGGACCAAACACAAATG-3′ and 5′-GAAGGGGAATGAGGAAAATA-3′. Initial denaturation at 95°C for 10 min was followed by 40 cycles with denaturation at 95°C for 15 s, annealing at 55°C (except at 50°C for MMP7 and at 45°C for TRPM-2) for 5 s, and elongation at 72°C for 30 s. The fluorescence intensity of the double-strand specific SYBR Green I, reflecting the amount of formed PCR-product, was monitored at the end of each elongation step. Cyclophilin mRNA levels were used to normalize the sample cDNA content. Three samples per each point (carcinomas from transgenic rats 0, 1, 2, 3, and 7 days after castration, and normal prostates of wild-type littermates) were analyzed.
For probe synthesis of cDNA arrays, poly(A)+ RNA enrichment and 32P labeling were performed according to the manufacturer’s instructions (Atlas Pure Total RNA Labeling System; Clontech). Atlas Rat 1.2 Array (Clontech) was used for comparing gene expression profiles between normal prostate of a wild-type littermate and a prostate carcinoma in a probasin-SV40 Tag transgenic rat at 15 weeks of age. Signals were detected and analyzed by image analyzer (FLA-3000G; Fujifilm, Tokyo) with Array Gauge software (Fujifilm).
Four male founder rats were obtained. Two of them fertilized females, which gave birth to viable pups. The transgenic rats of both lines developed prostate lesions from an early age, just after weaning at 4 weeks. One line made more pups; therefore, the transgenic rats from this line were used for the experiments. In the preliminary observation period, prostate lesions in the transgenic rats showed marked epithelial proliferation with formation of irregular glands and luminal bridging to give a cribriform pattern. Their nuclei demonstrated enlargement and severe atypia. These lesions are compatible with adenocarcinomas in human cases and were, therefore, diagnosed as such. Glands with less proliferation were also observed. These exhibited crowding of stratified epithelial cells with irregular spacing and occasional luminal bridging. Although nuclear atypia were severe, basic glandular structures were maintained, similar to normal prostates, and the lesions were diagnosed as PIN, 3 comparable with the human case. Prostate adenocarcinomas in the ventral, dorsolateral, and anterior lobes were observed at 100% incidence before 15 weeks of age, but no tumors were noted in the seminal vesicles. In addition to the prostate carcinomas, the transgenic rats developed taste bud neuroblastomas (details published in Ref. 29 ).
In experiment 1, effects of administration of testosterone propionate and castration on prostate carcinogenesis were investigated in probasin-SV40 Tag transgenic rats. Macroscopically, prostates of the transgenic rats in the nontreated group (group 1) showed slight enlargement and irregular surfaces but no apparent nodule or mass formations (Fig. 1a) ⇓ . Castration at 20 weeks of age (group 2) caused severe atrophy in both the prostate and the seminal vesicles (Fig. 1b) ⇓ . Testosterone propionate treatment, in contrast, caused enlargement, especially in the anterior lobes and seminal vesicles. Surfaces of these prostates were irregular; and in anterior lobes, hemorrhage was frequently observed (Fig. 1c) ⇓ . Castration after testosterone propionate treatment induced prostate and seminal vesicle atrophy (Fig. 1d) ⇓ . Data for weights of prostates and seminal vesicles with urinary bladders are summarized in Table 1 ⇓ . Treatment with testosterone propionate (group 4) caused increase, whereas castration (groups 2, 3, and 5) resulted in dramatic decrease.
In the nontreatment group (group 1), the transgenic rats developed PINs and adenocarcinomas in the ventral, dorsolateral, and anterior prostates at very high incidences (Fig. 2a) ⇓ . Almost all of the glands consisted of atypical cells, diagnosed as PINs or adenocarcinomas. This may be a reason that the prostate of the transgenic rats looks enlarged, but without obvious formation of masses or nodules. Castration at 20 weeks of age (group 2) caused involution completely after 5 weeks, when no neoplastic lesions were found, and only atrophic glands without atypia and fibrosis were observed. Castration at 5 weeks of age (group 3) suppressed tumor development completely. In this group, atrophic glands without atypia surrounded by slight fibrosis were found (Fig. 2b) ⇓ . Testosterone propionate treatment enhanced atypical cell proliferation with formation of solid adenocarcinomas featuring cribriform patterns (Fig. 2c) ⇓ . Microinvasion was often observed in this group, but, no metastases were noted. Even after enhancement of cell proliferation by testosterone, castration caused complete involution, leaving only atrophic glands without atypia and severe fibrosis and no neoplastic lesions retained in group 5 at week 25 (Fig. 2d ⇓ , see Table 2 ⇓ ).
In experiment 2, involution of the adenocarcinomas with enhanced cell proliferation attributable to testosterone propionate was induced by castration. One to 3 days after castration, numerous apoptotic bodies were observed (Fig. 3, c and d) ⇓ , and inflammatory cells including neutrophils, lymphocytes, and foamy cells had infiltrated the lesions (Fig. 3, e and f) ⇓ . After 7 days, atrophic glands and fibrosis were frequently observed, and nuclear atypia of the glands were minimal (Fig. 3, g and h) ⇓ . Therefore, these lesions could no longer be diagnosed as carcinomas. Degree of atrophy of the glands and fibrosis were increased 14 and 21 days after castration (Fig. 3, i and j) ⇓ . SV40 Tag expression was detected immunohistochemically in almost all cells of the prostate adenocarcinomas until 2 days after castration (Fig. 4a) ⇓ , but at day 3, only a small proportion showed positive reactions (Fig. 4b) ⇓ . After 7 days, almost no expression of the transgene was detected (Fig. 4, c and d) ⇓ . Expression of androgen receptors was dramatically diminished from just 1 day after castration (Fig. 4, e and f) ⇓ , and after day 2, no positive signals were detected immunohistochemically (Fig. 4, g and h ⇓ ; see Table 3 ⇓ ).
cDNA array analysis revealed differences in the gene expression profile between an adenocarcinoma in a transgenic rat and normal ventral prostate tissue in a wild-type animal. Genes with expression increased or decreased >2-fold are listed in Table 4 ⇓ . Among these, increased expression of TRPM-2 and MMP7 in carcinomas was confirmed by quantitative RT-PCR (Fig. 5) ⇓ .
Using the prostate materials obtained in experiment 2, expression of caspase 3, caspase 6, bcl-x, and BAX, as apoptosis-related genes, was investigated by quantitative RT-PCR, in addition to TRPM-2 and MMP7. Expression of caspase 3 and 6 in carcinomas before castration was higher than that in normal prostate tissue and increased to a peak 2 days after castration. Expression of bcl-x and BAX was also increased, but the peaks were at day 7. TRPM-2 and MMP7 expression also increased with peaks at day 3 (Fig. 5) ⇓ . Changes in TRPM-2 were furthermore confirmed by western blotting (data not shown).
Elucidation of the pathogenetic basis of prostate carcinogenesis can be facilitated greatly by laboratory models of the disease. We have developed a rat transgenic model producing well-differentiated prostate adenocarcinomas in all of the animals and in a short period using SV40 Tag under control of the probasin gene promoter. A transgenic mouse model of prostate carcinomas TRAMP, using the same gene construct was earlier established by Greenberg et al. (20) , featuring high-grade PIN and/or well-differentiated adenocarcinomas of the prostate by 10–12 weeks of age, with spontaneous development of invasive primary tumors that frequently metastasize to the lymph nodes and lungs in 30–36 week old animals (20) . Castration at 12 weeks of age did not inhibit progression to poorly differentiated and metastatic prostate carcinomas in TRAMP mice (30) . Earlier castration at 4 weeks of age reduced prostate cancer development, but some of the mice suffered from androgen-independent prostate carcinomas (31) . In our transgenic rats having the same transgene construct, well-differentiated adenocarcinomas similarly developed in the prostate. However, castration at 20 weeks of age caused complete regression or involution of these carcinomas and earlier castration at 5 weeks of age completely inhibited prostate carcinoma development, in clear contrast to the TRAMP case. Rapid decrease of expression of SV40 Tag and androgen-receptor in the prostate carcinomas after castration is in line with complete androgen-dependence. Because of this complete regression of the lesions and the macroscopic appearance of the prostates with no apparent masses or nodules, the lesion might be suspected of being hyperplastic in nature. However, histological characteristics such as structural and nuclear atypia and microinvasion are in line with a malignant neoplastic phenotype. Furthermore, the prostate lesions proved tumorigenic in nude mice (data not shown) without any pretreatment such as testosterone injection or Matrigel (32 , 33) , giving rise to adenocarcinomas similar to the original prostate lesions. Therefore, the data suggest that the prostate lesions in the transgenic rats are true malignancies (adenocarcinomas).
Small cell carcinomas were also occasionally observed in the prostates of transgenic rats, reminiscent of the poorly differentiated carcinomas noted in the TRAMP mouse prostate (20 , 30) . However, such small cell carcinomas in the transgenic rats demonstrated neuroendocrine tumor characteristics such as synaptophysin (34) and protein gene product 9.5 expression (35) , which were lacking in the TRAMP mice. Therefore, the origins of well-differentiated adenocarcinomas and small cell carcinomas may be very different, and development of small cell carcinomas is probably not a result of progression from well-differentiated adenocarcinomas. The transgenic rats also develop metastasizing taste bud neuroblastomas with the same histology, and these show androgen-independent growth (29) . Immunohistochemical phenotypes of most tumors were almost the same in the prostates and taste-buds (SV40 Tag, positive; protein gene product 9.5, positive; synaptophysin, positive; androgen receptors, negative). Whether the small cell carcinomas in the prostate are primary tumors or metastases from the taste bud tumors is, therefore, unclear. However, some showed weak positivity for androgen-receptor staining, which may indicate a prostate cell origin. In human cases, a small cell carcinoma with neuroendocrine character is an aggressive subtype of the prostate carcinomas (36, 37, 38) , and transgenic mouse models have been reported (39 , 40) . The present transgenic rats could also be useful in this respect.
The present prostate tumors are androgen-dependent and androgen-receptor positive, whereas the taste-bud tumors are androgen-independent and androgen-receptor negative. Both are positive for SV40 Tag. Therefore, in the prostate, the probasin gene promoter may function in a specific androgen-dependent fashion, whereas in the taste buds, other unknown factors may act in an androgen-independent way. In taste buds, probasin gene itself is not expressed.
Taking advantage of the rat prostate size, RNA was extracted from one-half of a transgenic prostate carcinoma and a wild-type littermate prostate in the present experiment to allow gene expression profiles to be investigated by cDNA array techniques. Although the number of genes on the array membrane (<1200) was limited, genes of interest for prostate carcinogenesis such as TRPM-2 and MMP7 were picked up as being up-regulated. TRPM-2, also known as clusterin or sulfated glycoprotein-2, was originally isolated from ram rete testes fluid (41) . It has been considered as a marker for cell death because of strong expression in various normal and malignant tissues undergoing apoptosis (42, 43, 44) . However, recently, conflicting data on the association between enhanced TRPM-2 expression and apoptotic activity were reported (45) . Introduction of TRPM-2 cDNA into LNCaP prostate cancer cells increases resistance to apoptosis induced by tumor necrosis factor α treatment (46) and promotes progression to androgen independence (47) . Elevated expression of TRPM-2 in human prostate cancer also correlates with increase in the Gleason score (48) and has been also reported in carcinogen-induced prostate cancers in rats (49) . In our transgenic rat model, TRPM-2 expression was found to be increased in carcinomas as compared with normal prostate, with further elevation during involution. Tissue undergoing apoptosis contains both apoptotic and remaining nonapoptotic cells. Therefore, these results may not conflict with the proposal that TRPM-2 is an antiapoptotic gene (47) , although TRPM-2 may have both apoptotic and antiapoptotic activities because of variation in glycosylation patterns (50) .
Degradation of the ECM is an essential step in tumor invasion and metastasis. The MMPs each have different substrate specificity within ECM and are important for this purpose; MMP7, for example, demonstrated increased expression in the prostate carcinomas of our transgenic rats and also in various human carcinomas, including prostate carcinomas (51, 52, 53, 54, 55) . MMP7 overexpression is also observed in involution of normal prostate (56) and prostate carcinomas in this study. During this process, tissue structure is dramatically changed, and degradation of ECM may be essential. MMP7 may, thus, have important roles for involution processes in addition to tumor invasion.
In our transgenic rat model of prostate carcinogenesis, the developed carcinomas are completely androgen dependent and castration causes complete regression in contrast to the compatible mice model. This has advantages for investigation of involution processes including apoptosis at the tissue level. Histologically, inflammatory cell infiltration was found to be involved in the present case. Increased expression of caspases 3 and 6, and BAX may reflect apoptotic activity and increased bcl-x may be caused by other cells escaping from apoptosis (57) .
In summary, we have developed a transgenic rat model featuring prostate carcinomas that are completely androgen dependent. This model may be useful to investigate mechanisms of development of androgen-dependent prostate carcinomas and processes of involution, and to evaluate strategies for prevention and treatment, including gene therapy. It should be remembered in this context that most human prostate cancers are androgen responsive before androgen ablation treatments.
We thank Dr. Malcolm A. Moore for his kind linguistic advice during preparation of the manuscript.
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
↵1 Supported by research grants from CREST (Core Research for Evolutional Science and Technology) of the Japan Science and Technology Corp. (JST), the Ministry of Health and Welfare of Japan, the Ministry of Education, Science, Sports and Culture of Japan, and the Society for Promotion of Toxicologic Pathology, Nagoya.
↵2 To whom requests for reprints should be addressed, at the First Department of Pathology, Nagoya City University Medical School, Kawasumi 1, Mizuho-cho, Mizuho-ku, Nagoya 467-8601, Japan.
↵3 The abbreviations used are: PIN, prostatic intraepithelial neoplasia; RT-PCR, reverse transcription PCR; MMP, matrix metalloproteinase; TRPM-2, testosterone-repressed prostate message 2; Tag, T antigen; ECM, extracellular matrix.
- Received November 28, 2000.
- Accepted April 17, 2001.
- ©2001 American Association for Cancer Research.