Androgen and Glucocorticoid Receptors in the Stroma and Epithelium of Prostatic Hyperplasia and Carcinoma1

Vol. 2. 889-895, Ma’ 1996 Clinical Cancer Research 889

James L. Mohler,2 Yeqing Chen, tatic carcinoma warrants further study of GR, particularly in the area of stromal-epithelial interaction. Katherine Hamil, Susan H. Hall, John A. Cidlowski, Elizabeth M. Wilson, INTRODUCTION Frank S. French, and Madhabananda Sar American men suffered 42,400 deaths and 244,000 new Departments of Surgery (Division of Urology) lJ. L. M., Y. C.]. Cell diagnoses of CaP3 in 1995 (1). In spite of the impact of CaP, Biology and Anatomy lM. S.], Physiology [J. A. C.]. Pediatrics (Laboratories for Reproductive Biology) [K. H., S. H. H.. E. M. W., little is known about its pathogenesis or predictors of response F. S. Fl, Biochemistry and Biophysics [E. M. WI. and University of to various therapies. Studies on prostate morphogenesis, growth, North Carolina-Lineberger Comprehensive Cancer Center U. L. M.. and function have focused upon androgenic regulation of epi- J. A. C., E. M. W., F. S. F.], University of North Carolina at Chapel thelial cells. However, recent studies indicate that at least some Hill. Chapel Hill, North Carolina 27599 responses of prostatic epithelial cells to hormonal stimuli may be the consequence of growth factors, morphogens, or inductors ABSTRACT produced in neighboring stromal cells (2-5). In view of the Differences in stromal and epithelial cell staining for potential importance of stromal-epithelial interactions during androgen and glucocorticoid receptors (ARs and GRs) were prostatic development, investigation of hormonal regulation by investigated in 20 patients with clinically localized prostatic the stromal compartment may lead to a better understanding of carcinoma treated by radical prostatectomy. Sections of prostatic disease and may provide insight into how abnormal benign prostatic hyperplasia and prostatic carcinoma from growth processes are modified by hormonal therapy. each patient were stained with antibodies to AR and GR Various programs of hormonal therapy have been used for using an avidin-biotin peroxidase technique. The specificity the treatment of patients with CaP since the pioneering work of of the GR immunoreactivity was established in benign pros- Huggins and Hodges (6). Approximately 80% of patients with tatic hyperplasia and prostatic carcinoma by immunohisto- metastatic CaP respond to androgen deprivation with an exten- chemistry using the GR antibody absorbed with synthetic sion of life expectancy from approximately 9 to 43 months. A peptide and Western blotting. Nuclear staining intensity and quantitative and/or functional difference of the AR may con- percentage of nuclei stained were summed to obtain AR and tribute to clinical response to androgen deprivation therapy (7, GR immunostaining scores. AR staining of prostatic carci- 8). However, Sadi et a!. (9) found no correlation between the noma epithelial [103 ± 58 (SD)] and stromal (126 ± 48) proportion of epithelial cells staining positive for AR and inter- nuclei was less than in benign prostatic hyperplasia (142 ± val to progression in 10 patients. Quantitative analysis in a 47 and 169 ± 56; paired Student’s t tests, P 0.02 and P = larger group of patients disclosed that heterogeneity of AR 0.01); however, no difference in staining intensity occurred staining among nuclei from individual tumors correlated with between stroma and epithelium in either tissue type. GR response (10). Similar to prostatic development, androgen-me- stained intensely in stromal cells from benign prostatic hy- diated growth of CaP may occur via paracrine factors produced perplasia (189 ± 50) and prostatic carcinoma (163 ± 60). by the stroma. The AR status of the epithelial compartment in However, prostatic carcinoma epithelial cells (34 ± 57) had early CaP may provide an incomplete description of disordered low levels of glucocorticoid receptor staining (P < 10), growth control. and benign prostatic hyperplasia epithelium (74 ± 51) was Approximately 20% of patients with androgen-independent intermediate. In most patients, GR could not be detected in CaP respond to aminoglutethimide or ketoconazole and corti- nuclei of prostatic carcinoma epithelial cells but was undi- costeroids (1 1-13). Responses have been attributed to the dim- minished in stromal cell nuclei. There was no relationship by ination of adrenal androgens (14). Antibodies specific for the multivariate regression analysis between AR or GR staining GR have been developed recently (15) but, to our knowledge, and age, serum prostate-specific antigen, Gleason grade, or they have not been applied to studies of prostatic tissue. pathological stage. In comparison with AR, the greater var- We reported previously ( I 6) methodology for sensitive and iability of GR staining in epithelium versus stroma of pros- specific detection of AR in frozen sections of fresh human prostate tissue using the polyclonal antibody AR-52 (17) and the avidin-biotin peroxidase method. Epithelial cell nuclei stained less intensively and more heterogeneously for AR in CaP than BPH. In the present study, we used antibodies specific for both Received 8/22/95: revised I 2/27/95: accepted 1/24/96. I Supported by Grant 5 UOl AG1 1343-02 from the NIH, National Institute on Aging. 2 To whom requests for reprints should be addressed, at Division of

Urology, Campus Box 7235, University of North Carolina at Chapel 3 The abbreviations used are: CaP, prostatic carcinoma; AR, androgen Hill, Chapel Hill, NC 27599-7235. Phone: (919) 966-2574; Fax: (919) receptor: GR, glucocorticoid receptor; BPH, benign prostatic hyperpla- 966-5722. sia.

Table 1 Raw data : Patient characteristics an d immunohistochemistry scores

Pathological AR intensity score GR inten sity score

BPH CaP BPH CaP PSA Patient AgeIRace’ (ng/ml) Clinical stageb Grade’9 Stage Epi Str Epi Str Epi Str Epi Str

1 47B 24.2 T3 7 OC 160 220 105 40 55 180 7 110 2 68W 0.7 T2a 5 OC 150 140 150 120 170 255 120 230 3 61W 10.6 T2b 7 CP, M 80 190 130 120 120 200 0 120 4 63W 6.4 T2a 6 CP 170 190 90 100 170 180 5 170 5 62W 8.8 T2a 6 OC 180 190 0 180 80 200 0 170 6 64W 25.4 T2b 7 SV 190 240 200 200 30 240 10 205 7 63W 5.9 T2a 8 OC 50 80 140 200 140 210 0 70 8 71W 10.5 T2a 7 OC 120 110 80 90 65 80 5 120 9 55W 12.1 T2b 9 CP, SV, M 190 130 100 100 95 190 20 190 10 66W 7.7 T2a 7 OC 140 160 150 80 60 255 0 190 1 1 53W 23.5 T2b 7 N 45 60 5 170 40 185 15 290 12 54W 18.0 T25 7 N 180 260 170 190 70 260 160 210 13 53W 12.3 T2a 7 CP, SV 190 220 90 90 0 130 0 120 14 69W 4.7 T2c 6 OC 140 190 55 150 0 140 0 150 15 75W 11.9 T2b 7 OC 75 110 160 120 75 115 60 180 16 75W 9.0 T2a 5 OC 150 170 150 140 60 140 0 70 17 58W 12.9 T2a 7 CP, M 170 200 140 80 140 255 180 235 18 64 B 27.6 T2 7 CP, M 120 160 0 155 40 200 5 170 19 75W 8.2 T2a 5 OC 150 110 80 50 20 160 0 60 20 59W 5.9 T, 7 OC 190 240 70 150 50 200 9 190 a B, black; W, white; PSA, prostate-specific antigen (Hybritech); OC, organ-confined; CP, capsular penetration; SV, seminal vesicle invasion; M, margin positive; N, nodes positive; AR, androgen receptor; GR, glucocorticoid receptor; Epi, epithelium; Str, stroma. b Schroder et al. (18).

C Gleason et al. (19).

AR and OR to compare intensity of staining of epithelium and frozen sections were air dried and fixed with PBS containing 4%

stroma in BPH and CaP. paraformaldehyde, 10% sucrose, and 0. 1 M sodium phosphate (pH 7.2). Sections were transferred into 2% hydrogen peroxide MATERIALS AND METHODS in PBS for 5 mm, washed, and permeablized in 0.2% Triton Patients. Tissue was obtained in the operating room from x-l00 for 10 mm. Nonspecific staining was blocked with 2% 20 patients [63 ± 8 (SD) years old] treated by radical prosta- normal goat serum. Sections were incubated with AR-52 (17) at tectomy for clinically localized CaP (Table 1). Serum prostate- an IgG protein concentration of 5 p.g/ml or GR-57 (15) at a specific antigen (Hybritech) was 12.3 ± 7.5 ng/ml. Pathological purified IgG protein concentration of 2 pg/ml in a moisture sum Gleason grade (19) was 6.7 ± 1.0 and CaP proved organ- chamber for 16-18 h at 4#{176}C.Sections were exposed at room confined in 1 1 patients (55%), specimen-confined in 3 patients temperature to biotinylated goat antirabbit IgG secondary anti- (15%), extended beyond surgical margins in 4 patients (20%), body (3.75 p.g/ml; I :400 dilution; Vector Labs Inc., Burl- and metastasized to obturator lymph nodes in 2 patients (10%). ingame, CA) for 60 mm, reacted with avidin-biotin peroxidase Tissue was obtained from the center of grossly recognizable (ABC; Vector Labs, Inc.), diluted 1:400 for 60 mm, and exposed nodules of CaP and from the transition zone opposite the car- for 10 mm with a solution containing 75 mg diaminobenzadine cinoma. Gross diagnoses were confirmed later histologically, (Aldrich Chemical Co., Milwaukee, WI) and 8.5 pA fresh 30% and no patient had undergone prior treatment for prostatic hydrogen peroxide in 100 ml 0.05 M Tris (pH 7.6). Negative disease, suffered endocrinological disease, or had received ste- controls included preimmune serum in place of AR and GR for roid therapy. each patient, AR and OR absorbed with synthetic peptide spe- Tissue Preparation. Tissue samples were placed into cific for each receptor protein in each run, and AR- and OR- iced 0.01 M PBS (pH 7.4), transported to the laboratory, and stained spleen for each run. Positive controls were rat liver processed immediately. Tissue was frozen in isopentane over present on each section as adhesive that stains uniformly and liquid nitrogen on brass mounts using rat liver cushions as an intensely for both AR and OR. To control for variability in adhesive. Tissue was stored in liquid nitrogen until sectioned at immunohistochemical preparation, all tissue sections including 8-jim thickness using a RMC cryostat at -31#{176}C.Ten adjacent controls were stained for AR in a single “run,” and all sections frozen sections from each tissue for each patient were prepared were stained for OR in a second “run.” Finally, all tissue for steroid receptor immunohistochemistry (8 sections) or alco- preparation and immunohistochemical staining was performed hol fixation and H&E staining for confirmation of histology by one highly trained technician (Y. C.). (two sandwiching sections). Steroid Receptor Intensity Scoring. All sections of Immunohistochemical Staining. The immunohisto- BPH and CaP from each patient were graded by one highly chemical technique has been described previously (16). Briefly, trained technician (Y.C.) for intensity of AR and OR staining in

.. . 1’ A I_ 4 ‘

‘44 U %1i ‘ ee. ‘4 c. ‘,. $ 111: : &‘: , C lt *

s. , ., U,

C ‘U $:‘ ,.

B ..# D 4; a $1 +#{149}* 1% L 4. :: ;

_1 C I

, * . ‘U, Fig. 1 Frozen sections of BPH were prepared for steroid receptor immunohistochemistry (reduced from X800) and stained with H&E (A), AR-52 against the AR (B), GR-57 against the GR (C), or GR-57 competed with synthetic peptide antigen comprising 22 amino acids of the GR (D).

epithelial cells and stromal fibroblasts (0, absent; 1, weak; 2, sized as described previously and used as the antigen to raise the moderate; or 3, strong), and the percentage of nuclei (0-100%) polyclonal OR-57 antibody in rabbits (15). Tissue sections were stained at each intensity level was determined. Total intensity prepared and stained as described above except that synthetic score (0-300) was the summation of the products of each peptide was added to OR-57 at a concentration ranging from 2 intensity score and its corresponding percentage (16). One hun- to 10 p.g/ml (1-5 times the concentration of OR-57); (b) frozen dred cells for each tissue type and steroid receptor stain were specimens of histologically confirmed BPH and CaP were pul- selected by 20 random movements of the microscope stage and verized and solubilized in 8 M urea, incubated 1 h at 0-4#{176}C,and grading of the five nuclei (AR) or cells (OR) closest to a point centrifuged to pellet insoluble residue. Seventy-five p.g of each superimposed upon the center of the field of view. Images were tissue extract were subjected to denaturing electrophoresis in collected and stored upon videotape to document accuracy of 8.0% polyacrylamide gels and transferred to Immobilon P (Mil- grading. Random sampling of 100 cells assured accurate de- lipore Corp., Bedford, MA; Refs. 2, 20). OR was detected using scription of average as well as variation in staining intensity for the ECL chemoluminescence (Amersham Corp., Arlington AR (16) and for OR (data not shown). Student’s t tests were Heights, IL) Western detection protocol, except that the primary used to compare the 20 paired specimens of BPH and CaP antibody was diluted at 1 :3000 (0.3 mg/ml) and the secondary obtained from the same radical prostatectomy specimens and antibody was diluted 1:10000. The primary antibodies used stromal-epithelial differences within patients. Multivariate re- were preimmune serum, OR-57, and GR-57 preincubated with a gression analyses were performed to test for dependence of 1: 1 concentration of the synthetic peptide described above (15). staining scores upon patients’ ages, serum prostate-specific antigens, sum Gleason grades, and clinical and pathological stages. Specificity of GR Antibody in Prostate Tissue. Since RESULTS OR-57 has not been used previously in prostate tissue, its Frozen sections of BPH and CaP treated with control specificity for OR was investigated in BPH and CaP by two preimmune serum stained negligibly. In BPH (Fig. 1), AR methods: (a) OR-57 was competed with the NH2-terminal do- staining was confined to the nucleus, whereas OR staining main peptide (amino acids 346-367) of the human OR synthe- occurred to some extent in cytoplasm but was more intense

a b ilarly, and stromal cells stained more intensely than epithelial cells. For example, patient 17 (Fig. 3, C and D) showed OR ( r-l--r- 2 -ir- 3’-, C., r-l--ir- 2 ---3 staining intensity scores for epithelium and stroma of 140 and 255, respectively, for BPH and I 80 and 235, respectively, for i4 U p’s -217 CaP. In 16 patients, OR staining of malignant epithelial cells was absent or barely visible, whereas staining remained intense -130 in the remaining cell types. For example, patient 3 (Fig. 3, E and -+ . ,1,_ C $ ‘*; F) showed no OR staining in neoplastic epithelial cells. but staining intensity scores ranged from I 20 to 200 in stroma of , neoplastic areas and stromal and epithelial cells of benign areas. Among all 20 patients, OR stained intensely in stromal cells of BPH (189 ± 50) and CaP (163 ± 60). However, CaP epithelial ‘ ..‘ . a .. =- cells (34 ± 57) had very low levels ofGR staining (P < l0), ‘ . and BPH epithelium (74 ± 5 1) was intermediate. #{149}‘: ‘. - -31.1 Data on immunohistochemical staining of AR and OR receptors for each type of tissue subtype are presented in sim- - - -- --- -18.1 BCB CBC, ,BCBCBC, plified form in Fig. 4. The greatest change in steroid receptor GR-57 GR-57 + Peptide expression, the decrease in OR staining observed in epithelial cells in CaP, would be even more significant within the sub- Fig. 2 Specificity of GR-57 for the GR was demonstrated in prostate group of 16 patients in whom OR could not be detected at all or tissue by Western blotting. Proteins were extracted from samples of BPH (B) and CaP (C) from radical prostatectomy specimens from each was barely visible, although the GR staining intensity was of three patients (numbered 1, 2, and 3), and GR was a preparation of undiminished in CaP stromal cells or benign tissue. human recombinant GR expressed in baculovirus-infected Sf9 cells. Multivariate regression analyses showed no relationship GR-57-stained control GR (arrow) as well as a Mr 97,000 protein in five between AR or OR staining and patients’ ages, serum prostate- of six samples (a). Staining was abolished (b) by preincubation of specific antigens, sum Gleason grades, and clinical and patho- GR-57 with a I : 1 concentration of the 22-amino acid synthetic peptide used to generate GR-57. Right, molecular weight markers. logical stages. For example, sum Gleason grade did not correlate with epithelial (r = -0.29, P 0.22) or stromal (r -0.10,

P = 0.69) AR staining. Similarly, differences between AR and OR staining in stroma versus epithelium in BPH and CaP or within the nucleus. Specificity of OR-57 for OR in prostate epithelium in BPH versus CaP were not related to clinical tissue was demonstrated immunohistochemically when staining parameters. For example, sum Gleason grade did not correlate was progressively decreased upon the addition of 2, 4, and 6 with differences in AR staining between benign and neoplastic

xg/ml of synthetic peptide. OR-57 staining was abolished by the epithelium (r = -0.20, P = 0.39) or between neoplastic stroma

addition of a 4-fold (Fig. 1D) and 5-fold concentration of and epithelium (r = 0.05; P = 0.83). Among all relationships synthetic peptide. Western blotting oftissue extracts of BPH and examined, no clinical factor was responsible for greater than 8% CaP revealed no detectable immunoreactivity when preimmune of the variation observed in steroid receptor staining. serum was substituted for OR-57. OR-57 detected protein of Mr 97,000 that corresponds to the molecular weight of human OR (Fig. 2a). This protein was not found when OR-57 was prein- DISCUSSION cubated with synthetic peptide (Fig. 2b). The specificity of OR Steroid receptors may have a role in the development, immunostaining as demonstrated by peptide adsorption of the progression, and treatment failure of human prostatic carcino- antibody for Western blotting is in agreement with the results of mas. Point mutations have been described in the AR steroid tissue immunostaining. Patient characteristics, grade, stage of binding domain in the human LNCaP cell line (21, 22) and but CaP, and AR and OR staining scores for epithelial and stromal occur rarely in clinically localized (23) and androgen-indepen- cells from BPH and CaP are detailed in Table I. dent metastatic CaP (24). Previous investigations of AR content A representative example of AR immunohistochemistry in benign and malignant prostatic tissue have demonstrated (Fig. 3, A and B) demonstrates the relationships between benign immunohistochemical staining of AR predominantly in the and malignant tissues. Malignant epithelial (staining intensity glandular epithelial cells (8, 16. 25-27). The recent demonstra- score, 80) and stromal (staining intensity score, 50) cell nuclei tion of the modulation of the epithelial compartment by the stain similarly. However, benign epithelial (staining intensity stroma (4) led us to investigate whether the AR content of the score, 150) and stromal (staining intensity score, 1 10) cells stain stroma might govern the biological behavior of the epithelium more intensely for AR protein. In the group of twenty patients, since epithelial AR content appears not to. We found no quan- the mean AR intensity scores in CaP epithelial (103 ± 58) and titative differences in AR staining between stromal and epithe- stromal (126 ± 48) nuclei were less than those in BPH (142 ± hal nuclei in either BPH or CaP. However, the abundance of AR 47 and 169 ± 56; P < 0.02 and 0.01, respectively). No signif- in the stromal and epithelial compartments suggests that, in icant difference occurred between stroma and epithelium in addition to direct actions on epithelial cells, hormonal-depen- either tissue type. dent carcinogenesis and progression may be regulated through The group of 20 patients exhibited two types of OR stain- stromal-epithelial interactions, as occurs during embryological ing. In four patients, malignant and benign tissues stained sim- development of the prostate (5).

a” I .. \ 4

‘ 44 U #{149} . ‘-#{149} .:#{231}.:..,. , B #{149} #:

p .. as .w, , . ‘p . ? , I,, ij’ #{149}#{149} :_#{149} p

... I a q,’4p& .‘ S ‘ - I, #{149} .‘.

Fig. 3 Patient 19 was representative of AR immunohistochemistry. BPH reduced from X800 showed AR intensity scores of 150 for epithelium and 1 10 for stroma (A); CaP reduced from X400 showed AR intensity scores of 80 for epithelium and 50 from stroma (B). Patient 17 was representative of four cases of GR immunohistochemistry, where BPH and CaP epithelial staining was similar but less intense than stromal staining. BPH reduced from X400 showed GR intensity scores of 140 for epithelium and 255 for stroma (C); CaP reduced from X400 showed GR intensity scores of 180 for epithelium and 235 for stroma (D). Patient 3 was representative of 16 cases where GR staining of epithelial cells was absent or nearly absent but CaP stroma and BPH epithelium and stroma stained similarly. BPH reduced from X800 showed GR intensity scores of 120 for epithelium and 200 from stroma (E); CaP reduced from X800 showed GR intensity scores of 0 for epithelium and 120 for stroma (F).

Our earlier findings (16), as well as those of Ruizeweld de numbers of cases studied or by the absence of a relationship Winter et a!. (27) and van der Kwast et al. (28), demonstrated between tumor grade and response to endocrine therapy (30). that CaP exhibits decreased intensity of nuclear staining that is OR belongs to the family of ligand-activated, transcriptional heterogeneous from gland to gland and even from cell to cell regulators that convey endocrine information to target cells. Many when compared to BPH. We found decreased intensity and of the physiological effects of steroid hormones are mediated similar heterogeneity of AR staining within the stromal com- through selective interaction of steroid receptor complexes with partment that was comparable to that observed within the epi- enhancer-like elements in the promoter and intron regions of ste- thelial compartment in CaP. The intensity scores for AR staining roid-responsive genes (31, 32). GR, like AR, contains a unique in this study were diminished compared to our prior report. amino terminal variable region that includes a transactivation do- Neoplastic nuclei stained with an intensity score of 103 ± 58 main that is important for the regulation of gene expression (33, versus 166 ± 69 in the previous work, and benign nuclei scored 34). The central DNA-binding domain contains zinc fingers crucial 142 ± 47 versus 246 ± 41 previously. However, the same for the specific interaction of receptor with DNA sequences that relationship between CaP and BPH was maintained; CaP nuclei contain glucocorticoid regulatory elements (35, 36). Finally, the stained with 67% of the intensity score and BPH nuclei with carboxyl terminus of the OR contains a ligand binding domain 72% of the intensity score. These differences emphasize the involved in the process of receptor activation. need for internal controls when comparing immunohistochem- A specific role for the OR receptor in CaP remains un- ical studies from different laboratories and even from different known. Once CaP becomes androgen independent, responses to investigators within the same laboratory. Interobserver variation secondary treatment that may be mediated through OR are could be minimized by quantitative analysis of AR staining unusual but do occur. Medical adrenalectomy using aminoglu- intensity as performed by Sadi and Barrack (10) using cytopho- tethimide or ketoconazole produced subjective improvement in tometry of immunoperoxidase-reacted sections. 33, 48, and 60% of 58, 40 and 20 patients with androgen- In the present study of 20 patients, neither AR content nor independent CaP (37-39). A more critical test of the effect of heterogeneity correlated with Gleason grade. We reported pre- these two drugs was conducted using the objective response viously in a smaller group of 10 patients significant inverse criteria devised by the U.S. National Prostate Cancer Project. correlation (r = -0.77, P = 0.009) between Gleason grade and Two groups of investigators studied prospectively previously intensity of AR staining (16). Masai et al. (29) also reported an castrated men with progressive metastatic CaP treated with inverse correlation between Gleason grade and percentage of aminoglutethimide. Worgul et a!. (1 1) found one complete “strongly stained” cancer cells. The different relationship be- response, four partial responses, and six men with stable disease tween grade and AR staining may be explained by the small among 25 patients; Drago et al. (12) found one complete re-

ANDROGEN RECEPTOR The relative decrease in OR in CaP epithelium is intriguing and could result from loss of a differentiated cell function. However, we found no relationship between GR content or Epithellum Stroma heterogeneity and sum Gleason grade. In clinically localized BPH 142 ±47 NS 169±56 CaP, cell maintenance and growth may depend upon androgens and glucocorticoids. Once the patient has been androgen de- prived, OR may retain a role in transcriptional regulation. Glu- 0.021 NS cocorticoid could influence androgen-independent CaP through CaP 103±57 C, 126±48 OR interactions with hormone response elements in androgen- regulated genes (46). Glucocorticoids may enhance the actions of androgens on certain genes through positive transcriptional GLUCOCORTICOID RECEPTOR effects mediated by hormone response elements. Simple response elements, 1 5-bp partial palindromes, mediate transacti- Epithellum Stroma vation by both AR and OR. Alternatively, glucocorticoids might 10 inhibit androgen-specific transactivation mediated by complex BPH 74±51 C, 189±50 response elements such as the intron element of the Mr 20,000 protein gene (47) that is transcriptionally unresponsive to glu- 0.02 i cocorticoid, although it binds OR. The differential loss of the 10 OR in the epithelial and stromal compartments of CaP is an CaP 34±57 163±60 interesting phenomenon that should be investigated further. Fig. 4 Summary ofdata available from Table 1. Immunohistochemical AR and GR staining intensity scores for epithelium and stroma from REFERENCES BPH and CaP were compared using paired Student’s t tests. Arrows, statistical comparisons were made and are labeled with a P value if 1. Wingo, P. A., Tong, T., and Bolden, S. Cancer Statistics, 1995. CA statistically significant or NS if not statistically significant. Cancer J. Clin., 45: 8-30, 1995. 2. Truong, L. D., Kadmon, D., McCune, B. K., Flanders, K. C., Scardino, P. T., and Thompson, T. C. Association of transforming growth factor-31 with prostate cancer: an immunohistochemical study. sponse, six partial responses, and 10 men with stable disease Hum. Pathol., 24: 4-9, 1993. among 43 patients. The complete response patients remained in 3. Kan, M., Wang, F., Xu, J., Crabb, J. W., Hou, J., and McKeehan, W. L. An essential heparmn-binding domain in the fibroblast growth remission after 275 and 290 weeks of therapy. Finally, using the factor kinase. Science (Washington DC), 259: 1918-1921, 1993. same objective response criteria, ketoconazole produced com- 4. Yan, G., Fukabon, Y., McBride, G., Nikolaropolous, S., and Mc- plete and partial responses in 1 and 14% of patients, respec- Keehan, W. L. Exon switching and activation of stromal and embryonic tively, when administered to patients with disease progression fibroblast growth factor (FGF)-FGF receptor genes in prostate epithelial despite adequate androgen deprivation (13). cells accompany stromal independence and malignancy. Mol. Cell Biol., 13:4513-4522, 1993. These responses to ketoconazole and aminoglutethimide 5. Cunha, G. R., Alarid, E. T., Turner, T., Donjacour, A. A., Boutin, may he due to these drugs or may result from glucocorticoid E. L., and Foster, B. A. Normal and abnormal development of the male replacement therapy required to prevent the Addisonian state urogenital tract: role of androgens, mesenchymal-epithelial interactions, these drugs produce. However, Trachtenberg (40) reported sim- and growth factors. J. Androl., 13: 465-475, 1992. ilar response rates with ketoconazole administered with and 6. Huggins, C., and Hodges, C. V. Studies on prostatic cancer. I. The without glucocorticoids. Furthermore, low-dose hydrocortisone effect of castration, of estrogen and of androgen injection on serum monotherapy prior to entry upon a suramin protocol at the phosphatases in metastatic carcinoma of the prostate. Cancer Res., 1: 293-297, 1941. National Cancer Institute produced no responses among 18 7. Gonor, S. E., Lakey, W. H., and McBlain, W. A. Relationship heavily pretreated patients with extensive androgen-independent between concentrations of extractable and matrix-bound nuclear andro- CaP (41). These responses could also result from further reduc- gen receptor and clinical response to endocrine therapy for prostatic tion in serum androgen levels. The administration of aminoglu- adenocarcinoma. J. Urol., 131: 1196-1201, 1984. tethimide plus hydrocortisone was shown to decrease testoster- 8. Fentie, D. D., Lakey, W. H., and McBlain, W. A. Applicability of one and dihydrotestosterone levels by 56 and 40%, respectively, nuclear androgen receptor quantification to human prostatic adenocarcinoma. J. Urol., 135: 167-173, 1986. compared to castrate levels (1 1). However, reduction of testos- 9. Sadi, M. V., Walsh, P. C., and Barrack, E. R. Immunohistochemical terone concentration from 28 to 13 ng/l00 ml and dihydrotest- study of androgen receptors in metastatic prostate cancer: comparison of osterone levels from 12 to 5 ng/ml may be biologically insig- receptor content and response to hormonal therapy. Cancer (Phila.), 67: nificant (42). In addition, Ponder et a!. (38) found no difference 3057-3064, 1991. in measured serum androgen levels between responders and 10. Sadi, M. V., and Barrack, E. R. Image analysis ofandrogen receptor nonresponders to aminoglutethimide and cortisone acetate ther- immunostaining in metastatic prostate cancer. Cancer (Phila.), 71: apy. The effects of ketoconazole and aminoglutethimide may 2574-2580, 1993. also result from inhibition of estrogen production by inhibition I 1. Worgul, T. J., Santen, R. J., Samojlik, E., Veldhuis, J. D., Lipton, A., Harvey, H. A., Drago, J. R., and Rohner, T. J. Clinical and bio- of aromatases (43) or cytochrome p450-dependent cyclooxyge- chemical effect of aminoglutethimide in the treatment of advanced nase, which is involved in arachidonic acid metabolism and prostatic carcinoma. J. Urol., 129: 51-55, 1983. prostaglandin synthesis (44). Finally, ketoconazole may have a 12. Drago, J. R., Santen, R. J., Lipton, A., Worgul, T. J., Harvey, H. A., direct cytotoxic effect upon CaP cells (45). Boucher, A., Manni, A., and Rohner, 1. J. Clinical effect of aminoglu-

tethimide, medical adrenalectomy, in treatment of 43 patients with 28. van der Kwast, T. H., Schalken, J., Ruizeweld de Winter, J. A., van advanced prostatic carcinoma. Cancer (Phila.), 53: 1447-1450, 1984. Vroonhoven, C. C. J., Mulder, E., Boersma, W., and Trapman, J. 13. Jubelirer, S. J., and Hogan, T. High dose ketoconazole for the Androgen receptors in endocrine-therapy-resistant human prostate can- treatment of hormone refractory metastatic prostate carcinoma: 16 cases cer. Int. J. Cancer, 48: 189-193, 1991. and review of the literature. J. Urol., 142: 89-91, 1989. 29. Masai, M., Sumiya, H., Akimoto, S., Yatani, R., Chang, C. S., Liao, 14. Labrie, F., Dupont, A., Giguere, M., Borsanyi, J-P., Lacourciere, Y., S. S., and Shimazaki, J. Immunohistochemical study of androgen re- Mongette, G., Emond, J., and Bergeron, N. Benefits of combination ceptor in benign hyperplastic and cancerous human prostates. Prostate, therapy with flutamide in patients relapsing after castration. Br. J. Urol., 17: 293-300, 1990. 61: 341-346, 1988. 30. Reiner, W. G., Scott, W. W., Eggleston, J. C., and Walsh, P. C. IS. Cidlowski, J. A., Bellingham, D. L., Powell-Oliver, F. E., Lubahn, Long-term survival after hormonal therapy for stage D prostatic cancer. D. B., and Sar, M. Novel antipeptide antibodies to the human glucocor- J. Urol., 122: 183-184, 1979.

ticoid receptor: recognition of multiple receptor forms in vitro and 3 1 . Yamamoto, K. R. Steroid receptor regulated transcription of spe- distinct localization of cytoplasmic and nuclear receptors. Mo!. Endo- cific genes and gene networks. Annu. Rev. Genet., 19: 209-252, 1985. crinol., 4: 1427-1437, 1990. 32. Burnstein, K. L., and Cidlowski, J. A. Regulation of gene expres- 16. Miyamoto, K. K., McSherry, S. A., Dent, G. A., Sar, M., Wilson, sion by glucocorticoids. Annu. Rev. Physiol., 51: 683-699, 1989. E. M., French, F. S., Sharief, Y., and Mohler, J. L. Immunohistochem- 33. Giguere, V., Hollenberg, S. M., Rosenfeld, M. G., and Evans, R. M. istry of the androgen receptor in human benign and malignant prostate Functional domains of the glucocorticoid receptor. Cell, 46: 645-652, 1015-1019, 1993. tissue. J. Urol., 149: 1986. 17. Tan, J., Joseph, D. R., Quarmby, V. E., Lubahn, D. B., Sar, M., 34. Hollenberg, S. M., and Evans, R. M. Multiple and cooperative French, F. S., and Wilson, E. M. The rat androgen receptor: primary trans-activation domains of the human glucocorticoid receptor. Cell, 55: structure, autoregulation of its messenger ribonucleic acid, and immu- 899-906, 1988. nocytochemical localization of the receptor protein. Mol. Endocrinol., 2: 35. Umesono, K., and Evans, R. M. Determinants of target gene specificity 1276-1285, 1988. for steroid/thyroid hormone receptors. Cell, 57: 1139-1146, 1989. 18. Schroder, F. H., Hermanek, P., Denis, L., Fair, W. R., Gospoda- 36. Danielson, M., Hinck, L., and Ringold, G. M. Two amino acids rowicz, M. K., and Pavone-Macaluso, M. The TNM classification of within the knuckle of the first zinc finger specify DNA response element prostate cancer. Prostate, 4 (Suppl.): 129-138, 1992. activation by the glucocorticoid receptor. Cell, 57: 1 131-1 138, 1989. 19. Gleason, D. F., Mellinger, G. T., and the Veterans Administration 37. Murray, R., and Pitt, P. Treatment of advanced prostatic cancer Cooperative Urological Research Group. Prediction of prognosis for resistant to conventional therapy with aminoglutethimide. Eur. J. Cancer prostatic adenocarcinoma by combined histological grading and clinical Clin. Oncol., 21: 453-458, 1985. staging. J. Urol., 111: 58-64, 1974. 38. Ponder, B. A. J., Shearer, R. J., Pocock, R. D., Miller, J., Easton, D., 20. Towbin, H., Staehelin, T., and Gordon, J. Electrophoretic transfer of Chilvers, C. E. D., Dowsett, M., and Jeffcoate, S. L. Response to proteins from polyacrylamide gels to nitrocellulose sheets: procedure and aminoglutethimide and cortisone acetate in advanced prostate cancer. some applications. Proc. Natl. Acad. Sci. USA, 76: 4350-4354, 1979. Br. J. Cancer, 50: 757-763, 1984. 21. Harris, S. E., Harris, M. A., Rong, Z., Hall, J., Judge, S., French, 39. Robinson, M. R. G., Shearer, R. J., and Fergusson, J. D. Adrenal F. S., Joseph, D. R., Lubahn, D. B., Simental, J. A., and Wilson, E. M. suppression in the treatment of carcinoma of the prostate. Br. J. Urol., Androgen regulation of HBGF-1 (a FGF) mRNA and characterization 46: 555-559, 1974. of the androgen receptor mRNA in the human prostate carcinoma cell line- LNCaP/A-Dep. In: J. P. Karr, D. S. Coffey, R. G. Smith, and D. J. 40. Trachtenberg, J. Ketoconazole therapy in advanced prostatic cancer. Tindall (eds.), Molecular and Cellular Biology of Prostate Cancer, pp. J. Urol., 137: 959, 1987. 315-330. New York: Plenum Publishing Co., 1991. 41. Dawson, N. A., Cooper, M. R., Figg, W. D., Headlee, D. J., 22. Veldscholte, J., Ris-Stalpers, C., Kuiper, G. G., Jenster, G., Ber- Thibault, A., Bergan, R. C., Steinberg, S. M., Sausville, E. A., Myers, revoets, C., Claassen, E., van Rooij, H. C., Trapman, J., Bnnkmann, C. E., and Sartor, 0. Antitumor activity of suramin in hormone-refrac- A. 0., and Mulder, E. A mutation in the ligand binding domain of the tory prostate cancer controlling for hydrocortisone treatment and flut- androgen receptor of human LNCaP cells affects steroid binding char- amide withdrawal as potentially confounding variables. Cancer (Phila.), acteristics and response to anti-androgens. Biochem. Biophys. Res. 76: 453-462, 1995. Commun., 173: 534-540, 1990. 42. Trachtenberg, J. Optimal testosterone concentration for the treat- 23. Newmark, J. R., Hardy, D. 0., Tonb, D. C., Carter, B. S., Epstein, ment of prostatic cancer. J. Urol., 133: 888-890, 1985. J. I., Isaacs, W. B., Brown, T. R., and Barrack, E. R. Androgen receptor 43. Harris, A. L., Dowsett, M., Smith, I. E., and Jeffcoate, S. L. gene mutations in human prostate cancer. Proc. Natl. Acad. Sci. USA, Endocrine effects of low dose aminoglutethimide alone in advanced 89: 6319-6323, 1992. postmenopausal breast cancer. Br. J. Cancer, 47: 621-627, 1983. 24. Culig, Z., Hobisch, A., Cronauer, M. V., Cato, A. C. B., Hittmair, 44. Harris, A. L., Mitchell, M. D., Smith, I. E., and Powles, T. J. A., Radmayr, C., Eberle, J., Bartsch, G., and Klocker, H. Mutant Suppression of plasma 6-keto-prostaglandin F1 181 and 13,14-dihydro- androgen receptor detected in an advanced-stage prostatic carcinoma is keto-prostaglandin F2 I 8 1 by aminoglutethimide in advanced breast can- activated by adrenal androgens and progesterone. Mol. Endocrinol., 7: cer. Br. J. Cancer, 48: 595-598, 1983. 1541-1550, 1993. 45. Eichenberger, T., Trachtenberg, J., Toor, P., and Keating, A. Ke- 25. Lubahn, D. B., Joseph, D. R., Sar, M., Tan, J., Higgs, H. N., Larson, toconazole: a possible direct cytotoxic effect on prostate carcinoma R. E., French, F. S., and Wilson, E. M. The human androgen receptor: cells. J. Urol., 141: 190-191, 1989. complementary deoxyribonucleic acid cloning, sequence analysis and 46. Marschke, K. B., Ho, K-C., Power, S. G. A., Tan, J-A., Kupfer, gene expression in prostate. Mol. Endocrinol., 2: 1265-1275, 1988. S. R., Wilson, E. M., and French, F. S. Response element diversity in 26. Chang, C., Chodak, G., Sarac, E., Takeda, H., and Liao, S. Prostate androgen regulated genes. In: Dufau, M. L., Isidori, A., and Fabbri, A. androgen receptor: immunohistochemical localization and mRNA char- (eds.), Cell and Molecular Biology of the Testis, pp. 69-79. New York: acterization. J. Steroid Biochem., 34: 31 1-313, 1989. Raven Press, 1994. 27. Ruizeweld de Winter, J. A., Trapman, J., Brinkmann, A. 0., 47. Ho, K-C., Marschke, K. B., Tan, J., Power, S. G., Wilson, E. M., Boersma, W. J. A., Mulder, E., Schroeder, F. H., Claassen, E., and van and French, F. S. A complex response element in intron-1 of the der Kwast, T. H. Androgen receptor heterogeneity in human prostatic androgen regulated 20-kDa protein gene displays cell type-dependent carcinomas visualized by immunohistochemistry. J. Pathol., 161: 329- androgen receptor specificity. J. Biol. Chem., 268: 27226-27235, 332, 1990. 1993.

Downloaded from clincancerres.aacrjournals.org on September 23, 2021. © 1996 American Association for Cancer Research. Androgen and glucocorticoid receptors in the stroma and epithelium of prostatic hyperplasia and carcinoma.

J L Mohler, Y Chen, K Hamil, et al.

Clin Cancer Res 1996;2:889-895.

Updated version Access the most recent version of this article at: http://clincancerres.aacrjournals.org/content/2/5/889

E-mail alerts Sign up to receive free email-alerts related to this article or journal.

Reprints and To order reprints of this article or to subscribe to the journal, contact the AACR Publications Subscriptions Department at [email protected].

Permissions To request permission to re-use all or part of this article, use this link http://clincancerres.aacrjournals.org/content/2/5/889. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.