[CANCER RESEARCH 59, 3392–3395, July 15, 1999]

Castration-induced Apoptosis of Androgen-dependent Shionogi Carcinoma IsAssociated with Increased Expression of Genes Encoding Insulin-likeGrowth Factor-binding Proteins1

Tara Nickerson, Hideaki Miyake, Martin E. Gleave, and Michael Pollak2

Lady Davis Research Institute of the Sir Mortimer B. Davis Jewish General Hospital and Departments of Oncology and Medicine, McGill University, Montreal, Quebec, H3T 1E2Canada [T. N., M. P.], and Division of Urology, Vancouver General Hospital, University of British Columbia, British Columbia, V5Z 3J5 Canada [H. M., M. E. G.]

ABSTRACT

Insulin-like growth factor (IGF) I has well-characterized mitogenic andantiapoptotic effects that are essential for maintenance of the normalprostate and may be important during regression of the normal prostateand/or prostate tumors induced by androgen-targeting therapies for pros-tate cancer. IGF-I activity is modulated by IGF-binding proteins(IGFBPs). Here we examine IGFBP expression during regression ofandrogen-dependent Shionogi carcinoma tumors after castration. In thismodel, we observe a 90% reduction in Shionogi tumors by 10 dayspostcastration. Northern blotting of RNA from tumors collected at varioustimes after castration indicates a rapid induction of IGFBP-5 concomitantwith apoptotic regression of tumors, as detected by Apoptag staining oftumor sections after castration. IGFBP-5 mRNA was not detectable intumors from control animals, but levels increased 120-fold in tumors 3days after castration. The mRNAs for IGFBP-3 and -4 were abundant inShionogi tumors from intact mice and decreased to;33% and ;20% ofcontrol, respectively. Castration had no significant effect on IGFBP-2expression. Treatment with calcium channel blockers inhibited castration-induced apoptosis and tumor regression and also significantly inhibitedup-regulation of IGFBP-5 after castration. These data provide strongevidence for a functional role of IGFBP-5 expression in mediating theapoptosis induced by androgen deprivation in androgen-dependent neo-plasia.

INTRODUCTION

Standard therapy for prostate cancer involves targeting androgensby blocking androgen production (by castration or the administrationof luteinizing hormone-releasing hormone agonists), or activity byadministration of antiandrogens. Androgen ablation therapy generallyleads to significant involution of prostate tumors, but the molecularmechanisms underlying this regression are incompletely described.

IGF-I3 has an important influence on the growth and survival ofboth normal and neoplastic prostate epithelial cells (1, 2). We haverecently demonstrated in a prospective study that a strong positivecorrelation exists between serum IGF-I levels and prostate cancer risk(3). The well-recognized mitogenic and antiapoptotic effects of IGF-I(reviewed in Refs. 4 and 5) are modulated by a family of high-affinityIGFBPs (IGFBPs 1–6; reviewed in Refs. 6 and 7).

IGFBPs have been shown to have apoptotic effects on both PC-3prostate cancer cells and MCF7 breast cancer cellsin vitro (8, 9). Wehave previously reported that castration-induced apoptosis in thenormal rat prostate gland is associated with an increased expression ofIGFBPs (10). The Shionogi carcinoma is a well-characterized model

of androgen-dependent neoplasia (11). We undertook these experi-ments to determine whether regression of androgen-dependentShionogi carcinoma after castration involves changes in local IGFphysiology.

MATERIALS AND METHODS

Shionogi Tumor Growth. Animal studies were conducted in accordancewith local humane animal care standards. The Toronto subline of the trans-plantable SC-115 AD mouse mammary carcinoma (12) was used in all exper-iments. Shionogi tumor cells were maintained in DMEM (Life Technologies,Inc., Gaithersburg, MD) supplemented with 5% heat-inactivated FCS. Forinvivo study, approximately 53 106 cells of the Shionogi carcinoma wereinjected s.c. into adult male DD/S strain mice, which were bred within our ownfacilities (n 5 5 animals/group; two tumors/animal). When Shionogi tumorsreached 1–2 cm in diameter (usually 2–3 weeks after injection), castration wasperformed through an abdominal incision under methoxyflurane anesthesia.Details of the maintenance of mice, tumor stock, and operative procedureshave been described previously (13). Tumors were measured each day, andtumor volume was calculated asL 3 W 3 H 3 0.5236 (14). Animals weresacrificed at various times after castration, and tumors were excised and fixedin 10% neutral buffered formalin for paraffin embedding, and the remainderwas immediately frozen in liquid nitrogen for isolation of RNA.

Inhibition of Apoptosis by Calcium Channel Blockers. To determinewhether changes in IGFBPs after castration are regulated by androgen orassociated with apoptosis, calcium channel blockers [600mg of nifedipine and100 mg of amlodipine (Norvasc)] were given three times/day by oral admin-istration to mice bearing Shionogi tumors 1 day before castration to inhibitcastration-induced apoptosis. Tumor tissues were harvested 3, 8, and 10 daysafter castration.

Quantitation of Apoptosis after Castration. To characterize changes inapoptotic rates after castration, the number of apoptotic bodies/high powerfield were counted and compared in control tumors before castration, regress-ing tumors 3, 5, 10, and 15 days after castration. Tumors were harvested atvarious times postcastration, fixed in 10% neutral buffered formalin, andembedded in paraffin. Fixed sections (5mm) were cut from each specimen andstained with digoxigenin-dUTP antibodies (Apoptag; Oncor, Inc., Gaithers-burg, MD). The number of Apotag-positive cells/high power field in fiverandom fields was counted and averaged.

Northern Analysis. Total RNA was isolated from tissue using the RNAzolB method (Teltest). Total RNA (40mg) was fractionated on 1% agarose gelsand transferred onto Zeta-Probe membrane (Bio-Rad) in 50 mM NaOH. ThecDNAs for rat TRPM-2 (15); IGFBP-2, -3, -4, and -5 (16); IGF-I (AmericanType Culture Collection); and IGF-IR (American Type Culture Collection)were labeled with [a-32P]dCTP using the T7 Quick-Prime kit (Pharmacia).Prehybridization, hybridization, and washing were performed as describedpreviously (10). Blots were subjected to autoradiography with intensifyingscreen at280°C. Quantitative analysis of gene expression was accomplishedby densitometric scanning of autoradiograms, and results were corrected forminor loading differences by normalizing to 28S rRNA.

RESULTS

Apoptotic Regression of Shionogi Tumors after Castration.Androgen-dependent Shionogi tumors develop rapidly in intact mice,reaching a volume of;1000 mm3 by 4 weeks. In this model, weobserve a 90% reduction in tumor volume by 10 days postcastration

Received 3/1/99; accepted 5/17/99.The costs of publication of this article were defrayed in part by the payment of page

charges. This article must therefore be hereby markedadvertisementin accordance with18 U.S.C. Section 1734 solely to indicate this fact.

1 Supported by grants from the National Cancer Institute of Canada (to M. P.) andNational Cancer Institute of Canada Program Centre Grant 9002 (to M. E. G.).

2 To whom requests for reprints should be addressed, at Lady Davis Institute, SirMortimer B. Davis Jewish General Hospital, 3755 Cote Ste Catherine Road, Montreal,Quebec, H3T 1E2 Canada. Phone: (514) 340-8222, ext. 5527; Fax: (514) 340-7502;E-mail: md49@musica.mcgill.ca.

3 The abbreviations used are: IGF, insulin-like growth factor; IGFBP, IGF-bindingprotein; IGF-IR, IGF-I receptor.

3392

American Association for Cancer Research Copyright © 1999 on December 9, 2011cancerres.aacrjournals.orgDownloaded from

(Fig. 1). The extent of apoptosis in Shionogi tumors was determinedby Apoptag staining of tumor sections. We observed minimal apop-tosis in tumors from intact mice (Fig. 2,A andC). The number of cellsundergoing apoptosis increased 5-fold by 3 days after castration andpeaked around 5 days (Fig. 2,B andC). Northern blot analysis wasused to confirm the expression of TRPM-2/clusterin (Fig. 2D), anandrogen-repressed gene known to be associated with involution ofShionogi tumors (12).

Effects of Castration on IGFBP Expression in Shionogi Tu-mors. Castration resulted in a dramatic increase in IGFBP-5 ex-pression, decreased expression of IGFBP-3 and -4, and had noeffect on IGFBP-2 expression, as determined by Northern blotanalysis. IGFBP-2 mRNA is expressed in control tumors and doesnot undergo a significant change in response to castration (Fig.3A). IGFBP-3 mRNA is abundant in control tumors, and decreasesto ;33% of control 3 and 10 days after castration (Fig. 3B).IGFBP-4 mRNA is also strongly expressed in Shionogi tumorsfrom intact mice and decreases to;20% of control 3 days aftercastration (Fig. 3C). By 10 days postcastration, IGFBP-4 levelsincrease slightly, reaching;40% of control. IGFBP-5 mRNA isvery low in tumors from control mice but increases 120-fold in

Fig. 1. Shionogi tumor regression after castration. Shionogi cells were injected intomice as described in “Materials and Methods.” After 28 days, castration wasperformed on untreated mice (M) or mice treated with Ca21 channel blockersnifedipine and amlodipine (F). Tumor dimensions were measured every 4 days(n 5 10/group), and tumor volumes were calculated and are expressed as mean tumorvolume 6 SD.

Fig. 2. Castration-induced apoptosis in Shionogitumors. Apoptag methodology as described in“Materials and Methods” was used to detect apop-tosis in Shionogi tumorsin situ. Sections of paraf-fin-embedded Shionogi tumors harvested frommice before castration and 3, 5, 10, and 15 daysafter castration were stained with digoxigenin-dUTP antibodies (Apoptag) to identify apoptoticbodies. A, representative Shionogi tumor beforecastration.B, representative Shionogi tumor 3 dayspostcastration.C, quantitation of the apoptotic timecourse in Shionogi tumors postcastration. For eachtime point, 50 high power fields were examined,and the means are plotted. SE for each data pointwas #15% of the mean value.D, Northern blotanalysis was used to detect the 2.3-kb TRPM-2/clusterin transcript.

3393

APOPTOSIS AND IGFBP EXPRESSION IN SHIONOGI TUMORS

American Association for Cancer Research Copyright © 1999 on December 9, 2011cancerres.aacrjournals.orgDownloaded from

tumors 3 days after castration (Fig. 3D). By 10 days postcastration,IGFBP-5 mRNA levels had declined slightly, but remained 100-fold of control levels. In relation to the time course of tumorregression, this dramatic increase in IGFPB-5 expression occursearly on, before the peak of apoptotic cell death.

Effect of Calcium Channel Blockers on Castration-inducedTumor Regression and IGFBP Expression.Apoptosis in manycell types requires a rise in intracellular Ca12 levels. Use ofcalcium channel blockersin vivo prevents apoptosis in a number ofcell types (17–19). In particular, treatment with calcium channelblockers prevents apoptosis of androgen-dependent prostate cellsand delays prostate regression after castration (20 –22). We con-firmed that administration of calcium channel blockers to micebearing Shionogi tumors effectively prevents tumor regressionafter castration (Fig. 1). Importantly, the up-regulation of IGFBP-5in tumors by castration was significantly inhibited by calciumchannel blocker treatment (Fig. 3D), whereas the expression ofother IGFBPs after calcium channel blocker treatment was notsignificantly altered (Fig. 3,A–C).

Effects of Castration on IGF-IR and IGF-I Gene Expression.IGF-IR mRNA levels decreased by;50% 3 days after castration andthen partially recovered by 8 days (Fig. 4A). IGF-I gene expressionincreased slightly as a result of castration, reaching 1.3-fold of controlby 8 days (Fig. 4B). Expression of IGF-II was not detected in tumorsfrom either intact or castrated animals (data not shown).

DISCUSSION

We demonstrate that apoptotic regression of androgen-dependentShionogi tumors is associated with major changes in IGFBP geneexpression within the neoplasms. Whereas IGFBP-3 and -4 expressiondecrease slightly after castration, the magnitude of increase in expres-sion of IGFBP-5 is remarkable. Furthermore, gene expression ofIGFBP-2, -3, and -4 is strong in control tumors in contrast toIGFBP-5, which is barely detectable. The effects of castration on geneexpression of IGF-I and IGF-IR in Shionogi tumors were minimalcompared to the changes in expression of IGFBPs. In view of the largeincrease in IGFBP-5 gene expression in Shionogi tumors after andro-gen ablation, it is possible that the overall increase in IGF bindingcapacity in the tumor microenvironment results in lower IGF bioac-tivity. Alternatively, there is evidence that IGFBPs have direct effectson cell growth and survival that are independent of IGF-I (23, 24).These mechanisms are not mutually exclusive.

It has been recognized for some time that IGFs have a central rolein regulating proliferation and apoptosis of prostate epithelial cells (1,25). However, more attention has been given recently to the apopto-sis-inducing effects of IGFBPs. It has previously been hypothesizedthat IGFBP-5 serves to trigger apoptosis in the prostate after androgenablation (26). Our observation that castration is followed by dramaticincreases in IGFBP-5 expression in Shionogi tumors that precedethe time of maximum apoptosis provides strong support for thishypothesis.

It is well established that during the process of apoptotic execution,the intracellular free calcium concentration increases, and, conse-quently, endonucleases, which can degrade nucleosomal DNA, areactivated. On the basis of these findings, we treated mice bearingShionogi tumors with calcium channel blockers before castration toprevent castration-induced apoptosis. As expected, calcium channelblocker treatment inhibited regression of Shionogi tumors after cas-tration through the inhibition of apoptotic cell death precipitated byandrogen ablation. Furthermore, up-regulation of IGFBP-5 after cas-tration was significantly reduced by calcium channel blocker treat-ment. Although functional analysis of IGFBP-5 is required to deter-mine the specific role after castration, these finding suggest that

Fig. 3. IGFBP expression in Shionogi tumors after castration. Northern blot analysiswas used to detect mRNA levels of IGFBP-2, -3, -4, and -5 in Shionogi tumors at varioustimes after castration from untreated animals or from animals treated with calcium channelblockers as described in “Materials and Methods.” Northern blotting was performed threetimes independently, and the results from each experiment were similar. All of theNorthern blots shown, including a glyceraldehyde-3-phosphate dehydrogenase control,are from a representative filter that was stripped and reprobed. IGFBP gene expressionwas quantitated by densitometric scanning of the representative autoradiograms and isshown on theright. The variation between densitometric scanning values for each timepoint from the independent experiments was#18% of the mean value.

Fig. 4. Expression of IGF-IR and IGF-I in Shionogi tumors after castration. IGF-IRmRNA (A) and IGF-I mRNA (B) expression was detected in Shionogi tumors at varioustimes after castration by Northern blot analysis. Densitometric scanning of autoradiogramsis shown on theright. The densities of the two lanes for each group were averaged. ForIGF-I, the sum of density bands corresponding to IGF-I transcripts was calculated for eachlane.

3394

APOPTOSIS AND IGFBP EXPRESSION IN SHIONOGI TUMORS

American Association for Cancer Research Copyright © 1999 on December 9, 2011cancerres.aacrjournals.orgDownloaded from

IGFBP-5 is an apoptosis-related gene rather than an androgen-repressed gene.

We have previously reported that increases in IGFBP gene expres-sion are associated with apoptotic regression of the normal rat prostateinduced by either castration (10) or the antiandrogen bicalutamide.4

IGFBP expression has also been shown to increase during postlacta-tional involution of the rat mammary gland (27) and after treatment ofbreast cancer cells with antiestrogens (28). Thus, there is accumulat-ing evidence that apoptosis associated with steroid hormone depriva-tion involves up-regulation of IGFBP expression.

Our results represent the first report documenting up-regulation ofIGFBP-5 after castration in a model of androgen-dependent neoplasia.Study of the functional role of IGFBP-5 in castration-induced apo-ptosis is now justified.

ACKNOWLEDGMENTS

We thank S. Shimasaki for providing rat IGFBP-2, -3, -4, and -5 cDNAs andM. Tenniswood for providing the TRPM-2/clusterin cDNA.

REFERENCES

1. Cohen, P., Peehl, D. M., Lamson, G., and Rosenfeld, R. G. Insulin-like growth factors(IGFs), IGF receptors, and IGF-binding proteins in primary cultures of prostateepithelial cells. J. Clin. Endocrinol. Metab.,73: 401–407, 1991.

2. Peehl, D. M., Cohen, P., and Rosenfeld, R. G. Role of insulin-like growth factors inprostate biology. J. Androl.,17: 2–4, 1996.

3. Chan, J. M., Stampfer, M. K., Giovannucci, E., Gann, P. H., Ma, J., Wilkinson, P.,Hennekens, C. H., and Pollak, M. Plasma insulin-like growth factor-I and prostatecancer risk: a prospective study. Science (Washington DC),279: 563–566, 1998.

4. Baserga, R. The insulin-like growth factor I receptor: a key to tumor growth? CancerRes.,55: 249–252, 1995.

5. LeRoith, D., Parrizas, M., and Blakesley, V. A. The insulin-like growth factor Ireceptor. Endocrine,7: 103–105, 1997.

6. Clemmons, D. R. Insulin-like growth factor binding proteins and their role incontrolling IGF actions. Cytokine Growth Factor Rev.,8: 45–62, 1997.

7. Nunn, S. E., Gibson, T. B., Rajah, R., and Cohen, P. Regulation of prostate cellgrowth by the insulin-like growth factor binding proteins and their proteases. Endo-crine,7: 115–118, 1997.

8. Rajah, R., Valentinis, B., and Cohen, P. Insulin-like growth factor binding protein-3induces apoptosis and mediates the effects of transforming growth factor-b1 onprogrammed cell death through a p53- and IGF-independent mechanism. J. Biol.Chem.,272: 12181–12188, 1997.

9. Nickerson, T., Huynh, H., and Pollak, M. Insulin-like growth factor binding protein-3induces apoptosis in MCF7 breast cancer cells. Biochem. Biophys. Res. Commun.,237: 690–693, 1997.

10. Nickerson, T., Pollak, M., and Huynh, H. Castration-induced apoptosis in the ratventral prostate is associated with increased expression of genes encoding insulin-like growth factor binding proteins 2, 3, 4 and 5. Endocrinology,139: 807– 810,1998.

11. Minesita, T., and Yamaguchi, K. An androgen-dependent mouse mammary tumor.Cancer Res.,25: 1168–1175, 1965.

12. Rennie, P. S., Bruchovsky, N., Buttyan, R., Benson, M., and Cheng, H. Geneexpression during the early phases of regression of the androgen-dependent Shionogimouse mammary carcinoma. Cancer Res.,48: 6309–6312, 1988.

13. Bruchovsky, N., and Rennie, P. S. Classification of dependent and autonomousvariants of Shionogi mammary carcinoma based on heterogeneous patterns of andro-gen binding. Cell,13: 272–280, 1978.

14. Gleave, M. E., Hsieh, J. T., Wu, H-C., and Chung, L. W. K. Serum PSA levels in micebearing human prostate LNCaP tumor are determined by tumor volume and endocrineand growth factors. Cancer Res.,52: 1598–1605, 1992.

15. Wong, P., Pineault, J., Lakins, J., Taillefer, D., Leger, J., Wang, C., and Tenniswood,M. Genomic organization and expression of the rat TRPM-2 (clusterin) gene, a geneimplicated in apoptosis. J. Biol. Chem.,268: 5021–5031, 1993.

16. Shimasaki, S., and Ling, N. Identification and molecular characterization of insulin-like growth factor binding proteins (IGFBP-1,-2,-3,-4,-5 and -6). Prog. Growth FactorRes.,3: 243–266, 1991.

17. Schrier, R. W., Arnold, P. E., Van, P. V., and Burke, T. J. Cellular calcium inischemic acute renal failure: role of calcium entry blockers. Kidney Int.,32: 313–321,1987.

18. Cheung, J. Y., Bonventre, Mallis, C. D., and Leaf, A. Calcium and ischemic injury.N. Eng. J. Med.,31: 1670–1676, 1984.

19. White, B. C., Winegar, C. D., Wilson, R. F., and Krause, G. S. Early amelioration ofneurologic deficit by lidoflazine after fifteen minutes of cardiopulmonary arrest indogs. Ann. Emerg. Med.,12: 8–9, 1983.

20. Kyprianou, N., English, H. F., and Isaacs, J. T. Activation of a Ca21-Mg21-dependent endonuclease as an early event in castration-induced prostatic cell death.Prostate,13: 103–117, 1988.

21. Connor, J., Sawczuk, I. S., Benson, M. C., Tomashefsky, P., O’Toole, K. M., Olsson,C. A., and Buttyan, R. Calcium channel antagonists delay regression of androgen-dependent tissues and suppress gene activity associated with cell death. Prostate,13:119–130, 1988.

22. Martikainen, P., and Isaacs, J. Role of calcium in the programmed death of ratprostatic glandular cells. Prostate,17: 175–187, 1990.

23. Oh, Y., Muller, H. L., Lamson, G., and Rosenfeld, R. G. Insulin-like growth factor(IGF)-independent action of IGF-binding protein-3 in Hs578T human breast cancercells. J. Biol. Chem.,268: 14964–14971, 1993.

24. Oh, Y. IGFBPs and neoplastic models. Endocrine,7: 111–113, 1997.25. Resnicoff, M., Abraham, D., Yutanawiboonchai, W., Rotman, H. L., Kajstura, J.,

Rubin, R., Zoltick, P., and Baserga, R. The insulin-like growth factor I receptorprotects tumor cells from apoptosisin vivo. Cancer Res.,55: 2463–2469, 1995.

26. Guenette, R. S., and Tenniswood, M. The role of growth factors in the suppression ofactive cell death in the prostate: an hypothesis. Biochem. Cell Biol.,72: 553–559,1994.

27. Tonner, E., Barber, M. C., Travers, M. T., Logan, A., and Flint, D. J. Hormonalcontrol of insulin-like growth factor binding protein-5 production in the involutingmammary gland of the rat. Endocrinology,138: 5101–5107, 1997.

28. Huynh, H. T., Yang, X. F., and Pollak, M. Estradiol and antiestrogens regulate agrowth inhibitory insulin-like growth factor binding protein 3 autocrine loop inhuman breast cancer cells. J. Biol. Chem.,271: 1016–1021, 1996.

4 T. Nickerson and M. Pollak. Bicalutamide (Casodex)-induced prostate regressioninvolves increased expression of genes encoding IGFBPs, submitted for publication.

3395

APOPTOSIS AND IGFBP EXPRESSION IN SHIONOGI TUMORS

American Association for Cancer Research Copyright © 1999 on December 9, 2011cancerres.aacrjournals.orgDownloaded from