Int. J. Cancer: 67, 785-790 (1996) 0 1996 Wiley-Liss, Inc. -0 Puolcatlon 01 tne Internal onal h o n Agamst Cancer Publcation ae ‘Union Internal onale Contre le Cancer CELL PROLIFERATION AND APOPTOSIS DURING PROSTATIC TUMOR XENOGRAFT INVOLUTION AND REGROWTH AFTER CASTRATION Franck BLADOU‘,’,Robert L. VESSELLA’, Kent R. BUHLER’,William J. ELLIS’,Lawrence D. TRUE’and Paul H. LANCE’ Departments of I Urology and ‘Pathology, University of Washington Medical School, Seattle, WA, USA. The biological mechanisms involved in androgen-dependent and -independent prostate cancer growth after castration were analyzed in the LuCaP 23. I human prostate cancer xenograft model. Athymic mice (n = 82) bearing LuCaP 23.1 xenograft were castrated and tumors were harvested at different time points from day 0 to day I I 2 post castration. In each group of mice, tumor growth rate (TGR), serum PSA concentration, percentage of tumor cells incorporating bromodeoxyuridine (BUdR index), percentage of apoptotic tumor cells assessed by morphological analysis (apoptotic index), and presence of apoptosis-related DNA “ladder” were analyzed. Castration induced a significant decrease in TGR and serum PSA from day I to day 7, and a progressive increase in the 2 parametersfrom day 14 to day I 12, heralding androgen-independenttumor relapse. Meanwhile the BUdR and apoptotic indexes varied as follows after castration: an increase was noted for both at day 3, a significant increase in apoptotic index with a decrease in BUdR index from day 5 to day 14, and a progressive decrease in apoptotic index while BUdR index remained at 50% of the pre-castration value from day 28 to day 112. DNA ladder was present sparsely in tumors grown in non-castrated hosts, universally present in tumors from day I to day 28 post castration, and frequent in tumors from day 56 to I 12. Castration-induced effects in LuCaP 23.1 tumors were characterized by an increase in number of apoptotic cells and a decrease in proliferative activity. The androgen-independenttumor relapse after castration was associated with a low apoptotic index with no increase in proliferative activity. o 1996 Wilqv-Liss,Inc. Advanced prostate cancer is at present not curable because androgen-independent (AI) tumor relapse occurs following medical or surgical castration, and eventually causes the death of the patient. The effects of androgen ablation and the mechanisms involved in A1 tumor growth are poorly understood. Indeed, prostate-cancer studies are limited because animal models are either animal tumors (i.e., Dunning and Noble rat models) or inadequate human prostate cancer models (since prostate cancer is a slow-growing tumor, most of the in vira and in vitro human models were established from explants of metastatic loci and do not mimic the clinical cancer). Androgen ablation induces genetically programmed cell death, or apoptosis, in the epithelial cells of the rat ventral prostate and in the androgen-dependent human prostate cancer xenograft PC82 within a few hours following castration (Kyprianou and Isaacs, 1988; Kyprianou et al., 1990). Within the first week post castration, apoptosis is responsible for the death of 80% of the epithelial cells in the rat ventral prostate (English et a/., 1989), and for a 9-fold increase in the percentage of apoptotic cells in the regressing PC82 tumor model (Kyprianou et al., 1990). However, 2 other androgen-sensitive prostate cancer models do not undergo apoptosis in response to castration. Gleave et al. (1992) in the LNCaP tumor model as well as Westin et al. (1993) and Brandstrom et al. (1994) in the Dunning R3327-PAP rat tumor model found no evidence of castration-induced apoptosis as assessed by the absence of DNA fragmentation, no change in the number of apoptotic cells, and no induction of the TRPM-2 apoptosis-related gene. Moreover, Westin et al. (1995) showed that the apoptotic index was increased in only a minority of prostate-cancer patients at day 7 following castration. The question of whether A1 growth of prostate cancer is due to adaptation of androgen-dependent cells or to clonal selection of A1 cells has been debated for years, and remains controversial (Isaacs and Coffey, 1981). A1 tumor relapse is also associated with abnormal regulation of autocrine and paracrine growth factors (Davies and Eaton, 1991; Ware, 1993). Finally, the equation: “cell number = cell proliferation - cell death” had led to the conclusion that uncontrolled tumor growth is caused by cell populations increasing their rate of proliferation, decreasing their rate of death, or both (Barr and Tomei, 1994). Recently. studies on oncogenesis have focused not only on the regulation of cell proliferation, but also on negative growth control such as apoptosis. In A1 prostatecancer relapse, the proliferation rate is low; thus, a logical conclusion is that apoptosis may be down-regulated in these tumors, although this hypothesis remains controversial (Westin et al., 1995). We have characterized a new prostate-cancer model, the LuCaP 23.x series of xenografts (Liu et al., 1996) that exhibits some of the salient properties of clinical prostate cancer, such as androgen sensitivity, PSA production and A1 tumor relapse following a period of regression after hormonal ablation. The aim of this study was to analyze the role of cell proliferation and apoptosis during tumor involution following androgen ablation and A1 prostate-cancer relapse in this promising animal model. MATERIAL AND METHODS LuCaP 23.1 human prostate cancerxenograft LuCaP 23.1 is a subline of the established human prostate cancer xenograft LuCaP 23.x. The donor was a 63-year-old white male diagnosed with a stage-D3 prostate adenocarcinoma. He had previously received androgen ablation therapy, radiotherapy and chemotherapy. Different metastatic foci were harvested within 2 hr post mortem for the generation of xenografts in athymic mice. Three sublines have been developed in our laboratory. Two of the xenograft sublines were established from lymph nodes (LuCaP 23.1 and LuCaP 23.8) and one from liver metastasis (LuCaP 23.12). The LuCaP 23.1 xenograft was chosen for this study because it produces high levels of PSA, shows androgen sensitivity and has the potential to undergo A1 growth post castration. Animals All animal uses and procedures were performed in compliance with the recommendations of the University of Washington Animal Care Committee. Tumor implantation and castration were performed under 130 mg ketamine/8.8 mg xylazineikg anesthesia. Male BALB/c nu/nu mice aged 6-8 weeks (Simonsen, Gilroy, CA) were implanted subcutaneously with 20-25 3To whom correspondence and reprint requests should be sent, at the Department of Surgical Oncology, Paoli-Calmettes Cancer Institute, 232 Bd Ste Marguerite, 13273 Marseille Cedex 9, France. Fax: 011 33 91 22 35 50. Received: December 19,1995 and in revised form April 25,1996. 786 BLADOU E T A L mm3 LuCaP 23.1 (passage 10). Tumor volume in individual animals was measured weekly with micro-calipers and calculated as: V (mm’) = (L x W x H) x 0.5236 (Janek et al., 1975). The percentage of tumor growth rate (TGR) was calculated by the formula: TGR = (tumor volume at end of the observation period)/(tumor volume at time of castration) (Landstrom et al., 1994). When the tumor volume reached 200 to 500 mm7, the mice were killed and LuCaP 23.1 tumors were harvested from 10 animals. Seventy-two mice were castrated transabdominally and tumors were harvested from randomized groups of 8 mice at day 1, 3, 5, 7, 14, 28, 56 and 112 post-castration. Following euthanasia, LuCaP 23.1 tumor volumes were measured and the xenografts were excised, weighed, and cut into 3 or 4 pieces. Tumor pieces were fixed in different histological media or flash-frozen in liquid nitrogen (see below). Measurement of serum PSA concentration Sequential serum PSA concentrations were measured weekly from samples obtained by capillary prick of the tail vein in all animals. Approximately 40 pl of blood were collected. Serum was separated by centrifugation and stored at -20°C until assay. A 1 : l O dilution in Hanks’ BSS was made and 100 pl of diluted serum were used for PSA determination by the automated IMx PSA immunoassay system (Abbott, Chicago, IL). The percentage of serum PSA rate was also calculated by the formula: PSA rate = (PSA at end of the observation period)/(PSA at time of castration). Extraction and analysis of DNA LuCaP 23.1 tumors were flash-frozen and powdered under liquid nitrogen. DNA was extracted from SO mg powdered tumor using the A.S.A.P. Genomic DNA Isolation Kit (Boehringer Mannheim, Indianapolis, IN). Briefly, powdered tumors were incubated in lysis buffer with RNase DNase-free solution (10 mg/ml) at 37°C for 30 min. Proteinase K (20 mg/ml) and guanidine hydrochloride (5M solution) were added and samples were incubated at 55°C for 3 hr. DNA was eluted in a chromatography column, precipitated with isopropanol and rinsed in 70% ethanol. DNA was then resuspended in TE buffer and quantified by spectrophotometry at 260 nm, after which 10-pg aliquots of DNA were electrophoresed on a 1.5% agarose gel. Gels were stained with ethidium bromide ( 5 pg/lOO PI), and DNA was visualized by UV fluorescence. Morphological analysis and electron microscopy Tissue samples (30-40 mm3) were fixed in a mixture of 2.5% glutaraldehyde and 2% paraformaldehyde in cacodylate buffer at -4°C overnight. The tissue was post-fixed in osmium tetroxide, dehydrated in graded ethanol solutions and propylene oxide, infiltrated with a mixture of epoxy and propylene oxide, and finally embedded in cpoxy (Medcast, BALTEC Product, Middlebury, CT). Then, 1-pm sections were stained with Azure IIiMethylene blue. Morphological analysis to estimate the percentage of apoptotic cells per ~ 4 0 field 0 was performed with an Olympus BH-2 microscope in a 10 X 10 grid. An average of 3,000 cells in regions of tumor which were free of necrosis has counted per tumor and 3 tumors were counted per group (27 tumors). For electron microscopy, selected areas were thin-sectioned, stained with uranyl acetatellead citrate, and examined under a Philips (Eindhoven, The Netherlands) 410 electron microscope. Proliferative activig One hour before tumor harvest, mice were injected intraperitoneally with 80 mg/kg BUdR (Sigma, St. Louis, MO). BUdR incorporated into S-phase cell DNA was visualized in paraffinembedded tissue by incubation of 5-pm tissue slices with a monoclonal anti-BUdR antibody (Zymed, South San Francisco, CA). Incorporated BUdR was detected with an indirect streptavidin-biotin-peroxidase system. Quantificative estimates of proliferative activity were assessed under 1OOx magnification (Olympus BH-2 microscope) and were expressed as a percentage of positively staining cells per x 100 field on a 10 x 10 grid. An average of 3,000 cells was counted per tumor and 3 tumors were counted per group (27 tumors). Statistics Values are expressed as the mean and standard error of the mean. Statistical analysis was performed by Student’s t-test for unpaired comparison. RESULTS Tumor growth and PSA production In non-castrated hosts, serum PSA concentration and tumor volume increased proportionally with a correlation factor r2 = 0.56. Although castration induced a retardation of the TGR in most animals, tumor volume did not significantly decrease during the first few weeks post castration in all animals. As shown in Table I, TGR was significantly lower in the groups of animals killed between day 1 and day 28 following castration when compared to the group of non-castrated mice. After castration, serum PSA concentration decreased significantly during the same period (until 4 weeks post castration) in the majority of animals (mean serum PSA decrease = 54%, p < 0.001). Following a 2-week period, an increase in serum PSA rate and TGR was observed in almost all castrated animals, heralding A1 tumor growth (Table I ; Fig. 1). Castration-induced changes in cell proliferation The average percentage of BUdR-stained epithelial cells, or BUdR index, in LuCaP 23.1 tumors grown in intact male mice was 8.6%. Within 3 days post castration, the BUdR index increased 1.25- to 1.5-fold, and then decreased progressively to a level of 5% 2 weeks after castration ( p < 0.01). This BUdR index remained low and stable up to 4 months post-castration, even in groups of relapsing tumors (Table I). Castration-induced changes in programmed cell death An early event in castration-induced apoptosis is the activation of a Ca++/Mg++-dependentendonuclease which causes a typical DNA fragmentation into multiples of a 180 nucleotide base-pair fragment, or “DNA ladder” (Kyprianou et al., 1988). Some low-molecular-weight nucleosomal fragments were detectable in a few LuCaP 23.1 grown in non-castrated hosts and in the TK-177 renal-cell carcinoma xenograft used as a tumor control. However, typical DNA ladder was universally present in tumors examined from day 1 to day 28 post castration and was frequent in the tumors harvested at days 56 and 112 post castration (Fig. 2). Programmed cell death was also noted in LuCaP 23.1 grown in intact mice, as indicated by the presence of apoptotic cells (less than 1% of the tumor cells). Apoptotic cells appeared as single rounded cells with condensed chromatin in semilunar nuclear caps surrounded by viable tumoral cells (Fig. 3). Apoptotic cells were often fragmented into apoptotic bodies presented as heterophagic vacuoles in neighboring tumor cells. When several apoptotic bodies were present at the same location, they were counted as pieces of the same cell. Castration induced a 2- to 3-fold increase in the percentage of LuCaP cells undergoing apoptosis 1 and 2 weeks post castration ( p < 0.05). The percentage of LuCaP apoptotic cells decreased progressively from day 14, to reach the precastration value 16 weeks post castration, as shown in Table I. The androgen-sensitive period after castration (1 week after castration) was characterized by a 50% decrease in prolifera- 787 CELL PROLIFERATION AND APOPTOSIS IN LUCAP 23.1 TABLE I - TGR, PSA RATE, PERCENTAGE OF APOPTOTIC, AND S-PHASE TUMOR CELLS, BEFORE AND AT DIFFERENT TIME POINTS AFTER CASTRATION ~~ Time Dostcastration TGR (%) Mean 2 SEM PSA Rate (%) Mean & SEM Apoptotic index (%) Mean & SEM Non-castrated Postcastration 1 day 3 days 5 days 7 days 14 days 28 days 56 days 112 davs 291.1 f 114.7 318.6 f 113.1 0.66 f 0.13 87 f 13.7** 115 2 44’ 116.5 ? 14.5* 98.8 f 18.8** 146.9 f 28.6* 150.2 f 57.8* 362.6 f 117.9 475 2 225.3 88.7 f 12.3** 71.8 f 16.4** 36.3 2 26.4** 36.5 2 22.5** 67.3 f 20.5** 120.2 f 86.2* 353.2 f 75 508.3 2 117.6 1.13 f 0.25 1.29 f 0.14* 1.69 f 0.41* 1.30 f 0.18 1.29 f 0.17 0.89 f 0.37 0.83 f 0.11 0.54 f 0.08 BUdR index (9%) Mean ? SEM 8.6 2 0.6 nd 10.7 f 1.1* 8.3 f 1.7 6.6 f 0.9* 5.0 f 0.6* 4.5 f 0.6** 5.1 f 1.1** 4.4 +. 0.9** * , p < 0.05; **, p < 0.01: Student’s t-test, unpaired comparisons of non-castrated group versus castrated groups. Castration induced a 2- to 3-fold increase in the apoptotic index within the first week, then the apoptotic index decreased steadily to reach the pre-castration value 4 months post-castration. The BUdR index (percentage of S-phase cells) increased during the first 3 days post-castration and decreased linearly during the next 3 weeks. The BUdR index remained at a low level (half of the pre-castration value) 4 months post-castration. ? 500 500 400 400 300 300 200 200 100 LOO h s a -s v 0 v a2 c v) CO 0 control day3 day7 day28 day112 Time post Castration FIGURE 1 - Evolution of average tumor growth rate (TGR (0-0) and PSA rate (B-B) in non-castrated hosts (controls1 and at different time points after castration. The decrease in PSA rate was higher than the decrease in TGR during the 4 weeks post castration. tive activity and a 2- to 3-fold increase in apoptotic cells. A1 tumor relapse was characterized by a progressive decrease in the apoptotic index with no increase in the proliferative activity (Fig. 4). Electron niicroscopic analysis Electron microscopic analysis showed the presence of clusters of round to oval osmiophilic “pseudo-granules” within the cytoplasm of tumor cells. These inclusions ranged in diameter from 0.5 to 5 km, their electron density varied from cell to cell, and they did not have a polar distribution within the tumor cells. The electron-dense inclusions were interpreted as lipid droplets and/or secretory granules. They were markedly increased in the tumor cells of castrated hosts, and particularly in the cells surrounding necrotic areas (Fig. 5). DISCUSSION In this study, androgen ablation induced both an early increase in apoptosis and a decrease in cell proliferation in the FIGURE 2 - Electrophoretic analysis of DNA isolated from LuCaP 23.1 prostate cancer xenografts from intact and castrated hosts, and from TK 177 renal cell carcinoma xenograft. Lane 1, 100-bp DNA ladder as a molecular weight marker; Lanes 2 and 3, control LuCaP 23.1 tumor DNA from non-castrated mice; Lanes 4 , 5 , 6 , 7 and 8, DNA isolated from LuCaP 23.1 tumors I, 7, 14,56 and 112 days post-castration; lane 9, DNA isolated from TK 177 xenograft. There is no evidence of DNA fragmentation in the first LuCaP tumor in a non-castrated host but some low-molecularweight nucleosomal fragments are visible in lanes 2 and 9 due to spontaneous apoptosis occurring in non-treated growing tumors. A typical DNA ladder is visible in LuCaP 23.1 tumors following castration (lanes 4-8). LuCaP 23.1 human prostate-cancer xenograft. The progression to androgen independence was concordant with a progressive decrease in apoptosis without an increase in cell proliferation. LuCaP 23.1 is a unique and promising model for in vivo prostate cancer studies. This model exhibits many of the defining characteristics of clinical prostate cancer, for example, heterogeneous growth kinetics, PSA production, castration-induced apoptosis, and A1 tumor relapse. Tumor growth and PSA production are heterogeneous within the same subline LuCaP 23.1 xenograft grown in non-castrated athymic male mice. The castration-induced response is variable as well. Castration induced a moderate regression in tumor volume 788 BLADOU E T A . FIGURE3 - Apoptotic cells and apoptotic bodies (arrows) shown in a 1-km section of plastic-embedded LuCaP 23.1 tumor on day 7 post-castration. Scale bar: 5 Krn. and a significant decline in serum PSA concentration during the first few weeks. We speculate that there is a population of androgen-dependent cells in such tumors, although the initial tumor was from a stage-D3 prostate cancer patient treated with androgen ablation. However, some tumors exhibited a relatively androgen-insensitive status. This heterogeneity in tumor growth and response to castration in LuCaP 23.1 is similar to that seen in patients treated with androgen ablation for advanced prostate cancer (Westin et al., 1995). Apoptosis is the most important contributory factor to the continuous cell loss from most growing tumors (Green et al., 1994; Wyllie, 1992). This process occurs in small as well as large tumors and may be caused by numerous factors such as hypoxia. inadequate levels of growth factors or hormones, or any sublethal injury (Cotter e f al., 1990). In LuCaP 23.1 xenografts in intact hosts, apoptosis occurs spontaneously, as evidenced by the presence of some degree of DNA fragmentation and a 0.5 to l%level of apoptotic cells. There are varying opinions regarding the best method of measuring apoptosis. Certain authors (Green et al., 1994; Collins et al., 1992) report that apoptosis can occur in the absence of DNA ladder, and thus DNA fragmentation should not be considered as the definitive feature of apoptosis. According to Green et al. (1994), cell morphology analysis is the most important parameter defining apoptosis. For these reasons, we employed a combination of 2 parameters (DNA analysis and cell morphol- ogy) to identify and quantify apoptosis in the LuCaP 23.1 model. Castration induced a 3-step response in the LuCaP 23.1 model. During the first 3 days post-castration, an increase was noted in BUdR index and apoptotic index. Two different hypotheses have been suggested to explain the increase in the BUdR index during the first few days post castration in the rat ventral prostate model (Colombel et al., 1992; Berges et al., 1993). The first hypothesis is that BUdR is incorporated into the DNA of GO epithelial cells as they enter an abortive cell cycle which ends in apoptosis (Colombel et al., 1992). The second hypothesis is that BUdR is incorporated during the cellular repair of damaged DNA, called “futile DNA repair,” which proceeds independently of the programmed cell death (Berges et al., 1993). However, the increase in the BUdR index noted on day 3 in this study was only approximately 20%. In contrast, within 1 day following castration there was more than a doubling of the apoptotic index and by day 5 there was a nearly 3-fold increase in this apoptotic index. These data suggest that androgen-dependent cells undergo programmed cell death rapidly following androgen ablation. During the second week after castration, the tumor response was characterized by a significant decrease in serum PSA concentration, a stable decrease in BUdR index, and a progressive decrease in the apoptotic index. As previously described in clinical studies and the LNCaP model (Leo et al., CELL PROLIFERATION AND APOPTOSIS IN LUCAP 23.1 300 250 B M 200 C 2 u s 150 100 50 - ._- . . _ _ _- -. ..-_ . .I 0 1 0 - I I r e I I I I I I I I I - I n P z % I -n - , Time post Castmtion (days) F~GURE 4 - Relative percentage of BUdR index (0-0), apoptotic index (-), TGR (-) and PSA (----)rate plotted against time post-castration in the LuCaP 23.1 xenograft model. The decrease in the BUdR index, TGR and PSA rate was concomitant with an increase in the apoptotic index during the first week following castration. The androgen-independent tumor relapse heralded by an increase in TGR and PSA rate was characterized by a progressive decrease in apoptosis while proliferative activity remained at a low level. FIGURE5 - Electron-micrograph of LuCaP 23.1 tumoral cells showing round to oval, osmiophilic, intracytoplasmic, electrondense inclusions ranging in diameter from 0.5 to 5 km in the cells surrounding necrotic areas. Scale bar: 3 km. 1991; Gleave et al., 1992), we observed a loss of correlation between tumor volume and serum PSA post-castration in the LuCaP 23.1 model: the PSA rate decrease was more important than the T G R decrease during the 4 weeks following castration (see Fig. 1). Androgen ablation results in the reduction of PSA production on a per-cell basis and the enrichment of undifferentiated, non-secreting cells. In our model, tumor volume and morphology did not vary markedly after castration. 789 Thus, the major effect of androgen withdrawal appears to be a tumor-cell repression revealed by a significant decrease in TGR, PSA production and proliferative activity. This observation is concordant with the experimental (Gleave et al., 1992; Westin et al., 1993; Brandstrom et al., 1994) and clinical findings (Murphy et al., 1991; Westin et al., 1995) reported in the literature. The mechanism of androgen withdrawal must be considered as suppression rather than as effective ablation in prostate cancer cells, and androgen withdrawal may induce a cell-cycle repression in prostate cancer cells for a limited period of time (Isaacs, 1994). Electron micrographic analysis showed an increase in cytoplasmic electron-dense inclusions in tumor cells after castration, and particularly around necrotic areas. Further studies are ongoing in our laboratory to determine whether these electron-dense inclusions are secretory granules containing PSA protein. Post-castration tumor relapse, the third step in the tumor response, occurred in almost all the LuCaP 23.1 xenografts. Tumor relapse was heralded by an increase in serum PSA from a nadir within 2 weeks post-castration, in much the same way as PSA rises in advanced prostate cancer patients who have undergone androgen ablation. In this study, tumors growing in an A1 manner had a decreased proliferative activity compared to the tumors before castration. This finding is concordant with the fact that the overall proliferation rate of recurrent prostate tumors is low, as shown by Landstrom et a1. (1994) in rat Dunning R3327-PAP tumors and that anti-proliferative chemotherapeutic agents are not effective in prostate cancer (Murphy, 1988). The significance of a decreased proliferative activity in an A1 relapsing tumor is not clear, and requires further investigation. Interestingly, tumor relapse after castration seemed to be related to a decrease in apoptosis rather than an increase in proliferative activity, as allready reported in a study on the Dunning R3327-PAP rat prostatic tumor model (Landstrom et al., 1994). A new concept in the mechanism of prostate cancer relapse (as well as chemotherapeutic resistance, for example) focuses on the role of new factors, ie., anti-apoptosis oncoproteins such as mutant p53 and bcl-2 proteins (Westin et al., 1995). Indeed bcl-2 is overexpressed in A1 prostate cancer (McDonnell et al.. 1992; Colombel et al., 1993) and in a recent report, we showed that hormoneindependent LuCaP 23.1 tumors expressed bcl-2 as well (Liuet al., 1996). Castration-induced apoptosis could be blocked (by bcl-2 or mutant p53) and induce A1 prostate-cancer relapse. If this hypothesis is confirmed, activation of programmed cell death may be a new therapeutic strategy for advanced prostate cancer, either by induction of apoptosis-activating genes or by inhibition of apoptosis-inhibiting genes, in conjunction with androgen withdrawal. Further studies are ongoing on bcl-2 and mutant p53 expressions in androgen-dependent and -independent LuCaP xenografts, as well as on induction of apoptosis in A1 tumors. In conclusion, the LuCaP 23.1 prostate cancer xenograft appears to be a promising in vivo model for the study of PSA production, effects of androgen ablation and mechanisms involved in A1 tumor relapse. Castration-induced effects observed in the LuCaP 23.1 model may be explained by both the decrease in proliferative activity and the increase in apoptotic cells. Tumor relapse after castration seemed to be related to a progressive decrease in apoptosis rather than an increase in proliferative activity. ACKNOWLEDGEMENTS This work was supported in part by a Veterans Administration RAGS grant, the Lucas Foundation and the O’Brien Center grant. F.B. was supported by a grant from the French Ligue Nationale contre le Cancer. 790 BLADOU E T A L REFERENCES BARR,P.J. and TOMEI,L.D., Apoptosis and its role in human disease. dependent GR mouse mammary tumors. Cancer Res., 35, 3698-3704 BiolTechnology, 12,487-493 (1994). (1975). BERGES,R.R., FURUYA, Y., REMINGTON, L., ENGLISH, H.F., JACKS,T. KYPRIANOU, N.. ENGLISH,H.F. and ISAACS,J.T., Activation of a endonuclease as an early event in castrationand ISAACS,J.T.. Cell proliferation, DNA repair, and p53 function are Ca2+-Mg2+-dependent not required for programmed death of prostatic glandular cells induced prostatic cell death. Prostate, 13,103-1 17 (1988). induced by androgen ablation. Proc. nut. Accid. Sci. (Wash.), 90, KYPRIANOU, N., ENGLISH, H.F. and ISAAC'S, J.T., Programmed cell 8910-8914 (1993). death during regression of PC-82 human prostate cancer following BRANDSTROM, A,, WESTIN, P., BERGH, A,, CAJANDER, S. and DAMBER, androgen ablation. Cancer Re$., 50,3748-3753 (1990). J.E., Castration induces apoptosis in the ventral prostate but not in an KYPRIANOU, N. and ISAACS, J.T., Activation of programmed cell death andro en sensitive rostatic adenocarcinoma in the rat. Cancer Rex, in the rat ventral prostate after castration. Endocrinobgy, 122,552-562 54,3564-;601 (19945. (1988). COLLINS, R.J., HARMON, B.V., GOBE,G.C. and KERR,J.F.R., Inter- LANDSTROM, M., DAMBER,J.E. and BERGH,A., Prostatic tumor nucleosomal DNA cleavage should not be the sole criterion for regrowth after initially successful castration therapy may be related to identifying apoptosis. Int. J. Radiat. Biol.. 61,451-453 (1992). a decreased apoptotic cell death rate. Cancer Res.. 54, 4281-4284 COLOMBEL, M.. OLSSON, C.A., NG, P.Y. and BUTTYAN, R., Hormone- (1994). regulated apoptosis results from re-entry of differentiated prostate LEO,M.E., BILHARTZ, D.L., BERGSTRALH, E.J. and OESTERLING, J.E., cells onto a defective cell cycle. Cancer Res., 52,4313-4319 (1992). Prostate specific antigen in hormonally treated stage D2 prostate COLOMBEL, M., SYMMANS, F., GIL, S., O'TOOLE,K.M., CHOPIN, D., cancer: Is it always an accurate indicator of disease status? J. Urol., 145,802-806 (1991). BENSON, M., OLSSON, C.A., KORSMEYER, S. and BUTTYAN, R., Detection of the apoptosis-suppressing oncoprotein bcl-2 in hormone- Lru, A.Y., COREY,E., BLADOU, F., LANGE,P.H. and VESSELLA, R.L., refractory human prostate cancers. Amer. J. Pathol., 143, 390-400 Prostatic cell lineage markers: emergence of bc12+ cells of human (1993). prostate cancer xenograft LuCaP 23 following castration. Int. J. COTTER,T.G., LENNON, S.V., GLYNN, J.G. and MARTIN,S.J., Cell Cancer, 65,85-89 (1996). death via apoptosis and its relationship to growth. Development and MCDONNELL, T.J., TRONCOSO, P., BRISBAY. S.M.. LOGOTHETIS, C., differentiation of both tumour and normal cells. Anticancer Res., 10, CHUNG, L.W.K., HSIESH,J.T., Tu, S.M. and CAMPBELL, M.L., Expres1153-1160 (1990). sion of the protooncogene bcl-2 in the prostate and its association with emergence of androgen-independent prostate cancer. Cancer Res., 52, DAVIES,P. and EATON,C.L., Regulation of prostate growth. J. 6940-6944 (1992). Endocrinol., 131,s-17 (1991). MURPHY, G.P., Is chemotherapy effective in the treatment of prostatic ENGLISH, H.F., KYPRIANOU, N. and ISAACS, J.T., Relationship between cancer? Clin. Oncol., 2,641-646 (1988). DNA fragmentation and apoptosis in the programmed cell death in the rat prostate following castration. Prostate, 15,233-150 (1989). MURPHY. W.M., SOLOWAY, M.S. and BARROWS.G.H., Pathologic GLEAVE,E.M., HSIEH,J.T., Wu, H.C.. VON ESCHENBACH, A.C. and changes associated with androgen deprivation therapy for prostate cancer. Cancer, 68,821-828 (1991). CHLING, L.W.K.. Serum prostate specific antigen levels in mice bearing human prostate LNCaP tumors are determined by tumor volume and WARE,J.L., Growth factors and their receptors as determinants in the endocrine and growth factors. Cancer Rex, 52, 1598-1605 (1992). proliferation and metastasis of human prostate cancer. Cancer Mefast. GREEN,D.R., BISSONNETTE, R.P. and COTTER,T.G., Apoptosis and Rev., 12,287-301 (1993). Cancer. Iniportanr advances in oncology, pp. 37-52, J.B. Lippincott Co., WESTIN, P., BERGH,A. and DAMBER, J.E., Castration rapidly results in Philadelphia (1994). a major reduction in epithelial cell numbers in the rat prostate but not in the highly differentiated Dunning R-3327PAP prostatic adenocarciISAACS.J.T. and COFFEY,D.S., Adaptation versus selection as the mechanism responsible for the relapse of prostatic cancer to androgen noma. Prosrate, 22,65-74 (1993). therapy as studied in the Dunning R-3327-H adenocarcinoma. Cancer WESTIN,P., STATTIN, P., DAMBER, J.E. and BERGH,A,, Castration Res., 41,5070-5075 (1981). therapy rapidly induces apoptosis in a minority and decreases cell proliferation in a majority of human prostatic tumors. Amer. J. Pathol., ISAACS, J.T., Advances and controversies in the study of programmed cell deathiapoptosis in the development of and therapy for cancer. 146,1368-1375 (1995). Curr. Opin. Oncol., 6,82-89 (1994). WYLLIE, A.H., Apoptosis and the regulation of cell numbers in normal JANEK,P., BRIAND,P. and HARTMAN, N.R., The effect of estrone- and neoplastic tissues: an overview. Cancer Mefast. Rev., 11, 95-103 progesterone treatment on cell proliferation kinetics of hormone- (1992).