androgen receptor gene amplification: a possible ......reaction; mab, maximal androgen blockade. 314...

[CANCER RESEARCH57, 314-319, January 15. 19971

ABSTRACT

Progression of prostate cancer during endocrine therapy is a majorclinical problem, the molecular mechanisms of which remain poorly understood. Amplification of the androgen receptor (AR) gene was recentlydescribed in recurrent prostate carcinomas from patients who had failedandrogen deprivation therapy. To evaluate the hypothesis that amplificalion of the AR gene Is a cause for the failure of androgen deprivationtherapy in prostate cancer, we studied whether AR amplification leads togene overexpression, whether the amplified AR gene Is structurally Intact,and whether tumors with AR amplification have distinct biological andclinical characteristics. Tumor specimens were collected from 54 prostatecancer patients at the time of a local recurrence following therapy failure.In 26 cases, paired primary tumor specimens from the same patients prior

to therapy were also available. Fifteen (28%) of the recurrent therapyresistant tumors, but none of the untreated primary tumors, contained ARgene amplification as determined by fluorescence in situ hybridization.According to single-stranded conformation polymorphism analysis, theAR gene was wild type in all but one of the 13 AR amplified casesstudied.In one tumor, a presumed mutation in the hormone-binding domain atcodon 674 leadIng to a Gly â€”@Ala substitution was found, but functionalstudies indicated that this mutation did not change the transactivatlonalproperties of the receptor. AR amplification was associated with a substantlally increased level of mRNA expression of the gene by in siftshybridization. Clinicopathological correlations indicated that AR ampilfication was most likely to occur In tumors that had Initially respondedwell to endocrine therapy and whose response duration was more than 12months. Tumors that recurred earlier or those that showed no Initialtherapy response did not contain AR amplification. The median survivaltime after recurrence was two times longer for patients with AR ampliflcation in comparison to those with no amplification (P = 0.03, WIllcoxon-Breslow test). In conclusion, failure of conventional androgen deprivatlon therapy in prostate cancer may be caused by a clonal expansionof tumor cells that are able to continue androgen-dependent growthdespite of the low concentrations of serum androgens. Amplification andthe Increased expression ofa wild-typeAR gene may play a key role In thisprocess.

INTRODUCTION

Prostate cancer is the most common malignancy in men in theUnited States and Western Europe. The proportion of operable earlystage (stages A and B) prostate cancers has increased rapidly in recentyears as a result of the widespread use of serum prostate-specificantigen measurements to screen for cancer in asymptomatic men (1).However, many prostate cancers are still diagnosed at stage C or D,

Received 8/12196;accepted I 1/18/96.The costs of publication of this article were defrayed in part by the payment of page

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

I Supported by the Finnish Science Academy, Finnish Cancer Society, The Dutch

CancerSociety,TampereUniversityHospitalResearchFund,andthe ReinoLabtikariandSigridJuseliusFoundations.In addition,P.K. hasreceivedsupportfromthe PirkanmaaCancer Society, Ida Montin Foundation,Finnish Culwral Foundation,Leiras ScienceFoundation,and FinnishSociety of Clinical Chemistry.

2 To whom requests for reprints should be addressed, at Laboratory of Cancer Genetics, National Center for Human Genome Research. NIH, 49 Convent Drive MSC 4470,Room 4A24, Bethesda, MD 20892-4470. Phone: (301) 435-2896; Fax: (301) 402-7957.

when cure by radical surgery is not possible. Patients with advanceddisease are usually treated by endocrine therapy, either by classicalandrogen deprivation (orchiectomy or luteinizing hormone-releasinghormone agonists) or by maximal androgen blockade (castrationcombined with antiandrogens). About 70â€”80% of the patients respond favorably and disease palliation is achieved for several monthsor years. Eventually, however, the disease progresses despite thetherapy (2).

Molecular mechanisms of recurrence and endocrine therapy failurein prostate cancer have remained unclear. Many investigators havestudied whether mutations in the AR3 gene would explain therapyresistance. Mutations leading to activation of AR by other steroids,and paradoxically also by antiandrogens, have been found, but theirprevalence in primary and locally recurrent prostate cancer is low and

clinical significance unknown (3â€”8).However, Taplin et a!. (9) recenfly reported AR mutations in 5 of 10 distant metastases from

patients who had failed endocrine therapy and suggested that ARmutations would play a major role in the metastatic progression ofprostate cancer.

Another molecular aberration involving the AR gene may alsounderlie endocrine therapy failure. In our recent study (10), amplification of the AR gene was found in about 30% of locally recurrent

hormone-refractory prostate cancers. In vitro studies of cancer celllines have revealed amplifications of specific genes associated withthe development of drug resistance to cancer chemotherapeutic agentsand metabolic inhibitors (11, 12). This prompted us to suggest that ARgene amplification in recurrent prostate cancer would represent ananalogous in vivo mechanism responsible for failure of androgendeprivation therapy. AR gene amplification, by leading to increased

expression of the AR gene, could allow the cancer cells to continueandrogen-dependent growth, although the serum androgen level isdrastically reduced after castration (10, 13).

Here, we sought to evaluate this hypothesis by identifying biological and clinical characteristics of prostate carcinomas that contain ARgene amplification as well as by studying whether amplification leadsto gene overexpression. Furthermore, because of the proposed role ofAR mutations in endocrine therapy failure, we evaluated whether theamplified AR gene is structurally intact.

MATERIALS AND METHODS

Patients and Tumor Specimens. The material consisted of patients withadvanced prostate cancer who were treated at the Tampere University Hospitalin Finland (n 51) or at the Dijkzigt Hospital in Rotterdam, the Netherlands(n 3) during 1972â€”1995.Conventional androgen deprivation therapy consisting of either orchiectomy (37 cases), luteinizing hormone-releasing hor

mone analogue (6), estrogen (6), or orchiectomy and estrogen (5) was theprimary treatment for these patients. All patients experienced a local tumorrecurrence,asevidencedby symptomsof urethralobstructionandan increaseof serum prostatic acid phosphatase (before 1991) or serum prostate-specific

3 The abbreviations used are: AR, androgen receptor; FISH, fluorescence in situ

hybridizaton; SSCP, single-strand conformation polymorphism; PCR, polymerase chainreaction; MAB, maximal androgen blockade.

314

Androgen Receptor Gene Amplification: A Possible Molecular Mechanism forAndrogen Deprivation Therapy Failure in Prostate Cancer'

Pasi Koivisto, Juha Kononen, Christian Palmberg, Teuvo Tammela, Eija Hyytinen, Jorma Isola, Jan Trapman,Kitty Cleutjens, Arjan Noordzij, Tapio Visakorpi, and Offi-P. Kallioniemi2

Laboratory of Cancer Genetics, Institute of Medical Technology. University of Tampere (P. K., J. K., E. H., J. I., T V., 0-P. K.J and Division of Urology (C. P.. T. TI. TampereUniversity Hospital (P. K., J. K., E. H., J. L, T. V., 0-P. K.), 33521 Tampere, Finland; Departments of Pathology If. T., K. C.] and Urology (A. N.J. Erasmus University,Rotterdam, the Netherlands; and Laboratory of Cancer Genetics, National Centerfor Human Genome Research, NIH, Bethesda, Maryland 20892-4470 [7'. V., 0-P. K.J

Research. on October 7, 2020. © 1997 American Association for Cancercancerres.aacrjournals.org Downloaded from

http://cancerres.aacrjournals.org/

ANDROGEN RECEPTOR GENE AMPLIFICATION

antigen(PSA) levels (after 1991). ThM stage andhistologicalgradedistribution of the tumors at the time of diagnosis are summarized in Table 1. Theresponse of the patients to primary endocrine therapy was classified as complete response, partial response, or stable disease according to the clinicalcriteria of Murphy et a!. (14) or the change in serum PSA levels. Pairedprimary tumor specimens taken before administration of any therapy wereavailable from 26 patients. Formalin-fixed, paraffin-embedded tumor specimens were used in all molecular analyses. These specimens had been collectedeither by Tru-Cut needle biopsy (17 primary, 12 recurrent tumors) or by

transurethralresection(9 primary,42 recurrenttumors).Analysis of Gene Amplification by Interphase FISH. The most repre

sentative paraffin-embedded blocks containing more than 70â€”80%malignantcells were selected for interphase FISH analysis. Many blocks were trimmed

by removing surrounding hyperplastic tissue. FISH analysis of AR gene amplification was done from disaggregated nuclei obtained by a proteinase K

treatment from 100-pm sections of the tumors (10). The nuclei were depositedon Vectabond-treated (Vector Laboratories, Burlingame, CA) slides and pre

treated by heating in a 50% glycerol/0. I X SSC (0. 15 M NaCI and 0.015 M

sodium citrate) solution for 3 mm at 90Â°C(15). Thirty ng of a biotin-labeledP1probefor theAR geneand5 ng of digoxigenin-labeledreferenceprobeforthe alphoid repeat of the X chromosome (DXZI) were mixed with 10 @gofplacental DNA and pipetted on denatured and proteinase K-treated cells onslides.Hybridizationwascarriedout in amoistchamberfor 24h at37Â°C.Afterhybridization, the slides were washed and the biotin-labeled AR probe wasdetectedwith avidin-FITC and the digoxigenin-labeledDXZ1 probe withanti-digoxigenin-rhodamine.The slides were counterstainedwith 4',6-diamidino-2-phenylindolein an antifadesolution.

Either a Nikon SA (Nikon Corporation, Tokyo, Japan) or an Olympus BX6O(Olympus Corporation, Tokyo, Japan) epifluorescence microscope was usedfor scoring signal copy numbers of both probes from a minimum of 70 nucleiperhybridization.Bothmicroscopeswereequippedwith multibandpassfiltersas well as both X60 and X 100 objectives. AR amplification was considered to

be present when the average number ofAR signals was more than five per cellor when the copy number of the AR gene was at least 2-fold higher than thatof the DXZ1. In practice,amplificationwaseasilyrecognizablebasedon thefinding ofindividual tumor cells with very high numbers ofAR signals, usuallyarranged in tight clusters (10). Aneuploidy or polysomy for chromosome Xwere defined as the presence of an equally elevated copy number for both ARand the DXZ1 reference probe and could thereby be clearly distinguished fromgene amplification.

Analysis of Gene Expression by mRNA in Situ Hybridization. Seven.sm sections were cut from the paraffin blocks representing both primary and

recurrenttumorsfrom six patientswithARamplification.Sectionswereplacedon SuperFrost (Menzel-Glhser,Germany) slides, deparaffinized, and subjectedto mRNA in situ hybridizationas described(16). An oligonucleotideprobeagainst the ligand-binding domain (nucleotides 810â€”854)of human AR (GenBank no. M34233; Ref. 17) was used. The probe was labeled with [33PJdATP(DuPont-NEN Research Products, Boston, MA) using terminal deoxynucleotidyltransferase (Amersham International, Buckinghamshire, United Kingdom). Hybridization was carried out in a humid chamber at 42Â°Cfor 18 h usingl0@cpm/mlof labeledprobe.Thesectionswerethenwashed,dehydrated,andexposed for 21 days to Hyperfilm (3-max X-ray film (Amersham International).As a control, a 100-foldexcess of unlabeled oligomer probe was added to thehybridization mixture. This abolished all hybridization signals. The hybridization properties of the paraffin sections and the quality of the target RNA were

ascertained by parallel hybridizations with a probe for the enzyme glycerol

dehyde-3-phosphate dehydrogenase (nucleotides 1022â€”1066,GenBank no.M33l97). After a 48-h exposure,this resultedin a virtually uniform hybridization of all tumor-containing regions of the sections.

Images of mRNA in situ hybridizations were processed with NIH image

software. To better visualize the mRNA expression levels, the original 256

gray level scalewasconvertedto 32 pseudocolors.The samemagnification,brightness, contrast values, and color conversion tables were used to processthe images from different slides. Amersham â€˜4Clow-activity autoradiographplasticstandards(AmershamInternational)were includedto verify thatexposure was within the linear range of the film and to provide a basis forcategorizing the expression levels. Expression levels were semiquantitatively

scored by two observers (P. K. and J. K.) to have either as no hybridizationsignal above tissue background (â€”), or low ( +), moderate (+ +), or high

(+ + + ) AR mRNA expression. The evaluation was done by inspecting autoradiographic films exposed for 21 days. The scoring was done without knowl

edge of the AR copy number.Structural Analysis of the Human AR Gene. Twelve cases with

AR-amplified tumors were screened for the presence of mutations in the ARgene using standard SSCP analysis. To this purpose, genomic fragmentscontaining exons 2â€”8of the AR gene were amplified using primers as follows:

exon 2, 5'-CAlTI'ATGCCTGCAGGTT and 5'-ATGGCTCTAlTI'CT

GAGATG; exon3, 5'-GmGGTGCCATACTCTGTCCAC and5'-CTGATGGCCACGUGCCTATGAA; exon 4, 5'-AAGTCTCTC'fltCTfC and 5'-TGCAAAGGAGUGGGCTGGUG; exon5, 5'-GACTCAGACTTAGC-TCAACC and 5'-ATCACCACCAACCAGGTCTG; exon 6, 5'-CAATCAGAGACAUCCTCTGG and 5'-AGTGGTCCTCTCTGAATCTC; exon 7, 5'-TGCTCCTrCGTGGGCATGCT and 5'-TGGCTCTATCAGGCTGTl@-CTC;and exon 8, 5'-ACCTCCflGTCACCCTGT and 5'- AAGGCACTGCAGAG

GAGTA (7, 18).StandardPCRconditionswere:25 cyclesof 1 mm at 94Â°C,2 mm at 55Â°Cor 60Â°C,and 2 mm at 72Â°Cin the presenceof 1 @.sCiof[32PIdATPin a 15-,.slreaction mix (Perkin-Elmer Thermal Cycler). One p1 ofthe reaction product was added to 9 p.1of a solution containing 98% formamide, 10 mM EDTA (pH 8), and xylene cyanol and bromphenol blue as dyemarkers. After denaturation (5 mm at 100Â°C)the solution was chilled on iceand I .5 pA was loaded onto a 6% nondenaturating polyacrylamide gel in 0.5 X

TBE and 5 or 10%glycerol. Electrophoresiswas done overnightat 7 W atroom temperature. Subsequently, the gel was dried and exposed to X-ray filmfor 24 to 72 h at â€”80Â°Cusing intensifying screens. The sensitivity of the SSCP

conditionsused was estimatedto be approximately90%, as judged from theanalysis of known point mutations in the AR in patients with androgen

insensitivity syndrome. Nontumor DNA served as control. Fragments showing

an aberrant PCR-SSCP pattern on gel were cloned into the vector pCRII(Invitrogen,Palo Alto, CA) and sequencedusing the standarddideoxy chaintermination method.

Transfections. To asses the transactivating function of mutated AR, themutation was introduced into AR eDNA in the AR expression vector pSVAR@)

(19) utilizing standard PCR technology with mutated primers. PCR fragmentsused for construction of the expression vector containing the mutated ARcDNA were completely sequenced. Functionality of the AR was analyzed inHep3B cells, which were cotransfected with wild-type or mutant pSVAR@)andthe reporter vector GRE2tkLuc with the calcium phosphate precipitationmethod. Following transfection, cells were incubated for 24 h in the absence orin the presence of the synthetic androgen Rl 88 1 (l0@ M) or the anti-androgen

OH-flutamide(l0@ M).Cell lysateswereprepared,andluciferaseactivity wasmeasured as described (20). All experiments were done in 4-fold, and corrections were made for protein content of the samples.

Statistical Analyses. BMDP Statistical Software Package (21) was used inall statistical analyses. Clinicopathological correlations were analyzed using

the Pearson's@ test (BMDP4F program). The statistical significance ofsurvival differences between patient groups was determined with Mantel-Cox

and Breslow tests (BMDP1L).

RESULTS

Prevalence of AR Amplification. According to interphase FISH

analysis, 15 of the 54 (28%) recurrent prostate cancers showed spe

cific amplification of the AR gene using FISH. The mean AR copynumber in these tumors ranged from 2.7 to 28/cell, with individual

cells showing much higher copy numbers,ranging from 5 to morethan 60 AR copies/cell in the various tumors. Aneuploidy or polysomyfor chromosome X, seen in 10 tumors (19%), was distinguished fromAR gene amplification based on the finding of an equally high copynumber for both the AR and the DXZI probe. The remaining 29recurrent tumors (53%) had only a single copy ofAR and DXZ1. ARamplification was detected both in tumors from Finnish and Dutchpatients.

AR gene amplification was not detected in any of the 26 primarytumors studied. Between 1000 and 1500 nuclei/specimen werescreened from 10 primary tumors available from those patients whosetumors had AR gene amplification at the time of recurrence. No

315



Table I Clinicopathological characteristics ofprostate cancer patients at the timeofdiagnosisThe

patients were classified into those who developed recurrent tumors withARamplificationand those with no AR amplification at the time of diseaserecurrence.AR

amplification No ARamplification(n15)(n39)Mean

age (range), yr 71 (51â€”77) 71(55â€”79)TstageT1@,

1(7%)5(13%)T3@413(86%)33(85%)T@I(7%) 1(2%)M

stageM07(47%) 17(44%)M13(20%)12(31%)M@5(33%)10(25%)Histopathological

grade oftheprimarytumorI+ II 7(47%) 26(67%)III

7(47%)12(31%)Unknown1(6%) 1 (2%)


evidence of cells with unambiguous gene amplification was foundeven in this extensive analysis. Occasional doublet signals that wereobserved are likely to reflect the presence of the S-phase and G2-Mcells.

Clinicopathological Characteristics and Predictive Value of ARAmplification. AR gene amplification was not significantly associated with the age of the patient, tumor size, presence of distantmetastases, nor with the tumor grade evaluated at the time of diagnosis (Table 1). In contrast, the appearance of AR amplification was

associated with both the degree and duration of response to primaryendocrine therapy (Fig. 1). Recurrent tumors from patients who hadnot initially responded to androgen deprivation did not show AR geneamplification (Fig. IA). In contrast, two of the three patients withcomplete response and one third of those with a partial response toprimary endocrine therapy had AR-amplified recurrent tumors. ARgene amplifications were not found in any of the 11 recurrent specimens from patients who progressed less than I year after initiation ofandrogen deprivation therapy (Fig. 1B). In contrast, AR amplificationwas found in 30â€”38%of the cases whose response duration was morethan 1 year.

Patients with AR gene amplification had a more favorable overallsurvival than those with no amplification (Fig. 2A). Since the AR geneis only amplified at the time of recurrence, we also evaluated themedian postrecurrence survival of the patients which was almost twotimes longer in the AR-amplified cases as compared with the nonamplified ones (Fig. 2B). The survival differences were most prominentduring the first 48 months of follow-up.

AR Amplification Leads to Increased mRNA Expression. According to mRNA in situ hybridization, the AR gene was highlyexpressed when the gene was amplified (Fig. 3). Based on a semiquantitative analysis of six cases, a wide range of expression levelswas seen in both primary and recurrent tumors. However, geneexpression was systematically higher in the recurrent AR-amplifiedtumors than in the corresponding primary unamplified tumors(Table 2).

Amplification and Mutations of the AR Gene With one exception, the coding region of the AR gene was not structurally altered inany of the 13 tumors with AR amplification studied using PCR-SSCPand sequencing. In the only structural aberration detected, codon 674(GGT) in exon 4 was changed into GCT, leading to Gly â€”@Alasubstitution. Although normal control DNA from the same patient wasnot available, an Ala at position 674 has never been found in any ofmany hundreds of DNA samples analyzed, arguing against a polymorphism. The mutation was introduced into AR cDNA and tested forfunctionality. Comparison of wild-type and mutated AR-transactivat

70 67%

60

50

40

30

20

10 0%

SD(n5)

Response to hormonal therapy

38%

50

40

30

20

10

0%

1-3 >3(n21) (n=20)

A

04-.

U

a.EtS

I.00

$

I

B

C00U

a.E0

00

$C0UI-0a-

<1(n11)

Timeto localrecurrence(years)Fig. I . A, frequency of AR gene amplification in recurrent prostate cancer according to

the initial response of the corresponding primary tumor to endocrine therapy. Tumors thatresponded well to androgen deprivation are most likely to recur with AR amplification,whereas tumors that showed no objective response never contained AR amplification atthe time of recurrence. CR, complete response; PR, partial response; SD, stable disease.B, frequency of AR gene amplification in recurrent prostate cancers according to the timefrom diagnosis to local recurrence. None of the tumors that recurred less than 1 year afterdiagnosis contained AR gene amplification.

ing activity in Hep3B cells did not reveal any difference (data notshown).

DISCUSSION

The present results emphasize the close association between ARgene amplification and failure of androgen deprivation therapy inprostate cancer as originally reported by us in a smaller series (10).Several new findings that emerged from the present study substantiatethe hypothesis that AR amplification is directly contributing to thedevelopment therapy failure by allowing cells to resume hormonedependent growth in the low concentrations of androgens. First,amplification occurs exclusively in the recurrent tumors from patientswhose disease has progressed during the therapy. Not even individualamplified cells were found in the untreated primary tumors, many ofwhich came from patients with late-stage metastatic prostate carcinoma. Thus, AR amplification is not involved in the genesis of

316



although it does not exclude the fact that there may be other importantgenes in the same amplicon. Results from the mRNA in situ hybridization are compatible with those from previous immunohistochemical studies (7, 22) in that the AR gene is virtually always highlyexpressed in both primary and recurrent prostate cancer. However,

immunohistochemical studies are less quantitative than the mRNA insitu hybridization used here, and detection of the increased level ofexpression in this subgroup of recurrent tumors has previously notbeen reported.

Fourth, the amplified AR is usually not mutated and has therebyretained specificity for androgens. Although 1 of the 15 AR-amplifiedcases showed a mutation in the steroid-binding domain, this had noinfluence on the transactivational properties of the receptor, suggesting that this mutation did not lead to impaired steroid specificity. Thefinding of predominantly wild-type AR gene in the amplified tumorssupports the concept that recurrent AR-amplified tumors may still beandrogen dependent. Our results on AR mutations are compatible withprevious studies of both primary and locally recurrent prostate cancersthat have shown a very low frequency ofAR mutations (3â€”5,7â€”8).Arecent study indicated that micrometastases of prostate cancer may

much more often contain AR gene mutations (9).Fifth, prostate cancer is known to be an extraordinarily hor

mone-dependent tumor type. Since the AR protein functions as akey mediator of androgen-dependent growth, amplification andconsequent increased expression of such a receptor gene is alogical mechanism by which the growth of tumor cells could bemaintained when this critical ligand for the receptor is present at avery low concentration. There are more than 20 analogous in vitroexamples of specific gene amplifications that have been foundwhen cultured tumor cells have been deprived of an essential agentfor growth or treated by chemotherapeutic drugs (12). For examplc, DHFR gene amplification is selected for during methotrexatetreatment in response to deprivation of folic acid in vitro ( I 1) andthe recent report by Goker et a!. (23) suggests that a similarmechanism may be operational in therapy-resistant leukemic cellsin vivo. AR amplification was never seen in the tens of thousandsof cells screened from many different untreated tumors. Furthermore, close to I00% of the cells in the recurrent tumors usually hadAR amplification. These findings, taken together with the fact thata long period of therapy (12 months) was required before ARamplification emerged, all resemble the situation in the previous invitro studies where a strong selection force has been applied on thecultured cells to induce a stepwise selection for gene amplificationover many cell generations (24, 25).

Patients whose tumors contained AR amplification at the time ofrecurrence had a somewhat more favorable prognosis than thosewhose tumors were not amplified. It is unexpected that DNA amplification, usually associated with genetically highly unstable (26â€”28)and aggressive tumors (29â€”30),is showing an association with favorable prognosis. One can speculate that the AR-amplified tumor cells,although being genetically unstable, are still subject to a rigorouslyregulated hormone-dependent growth and therefore show a lowerlevel of malignant potential. Based on this material, however, it isdifficult to separate the effects of a favorable response to androgendeprivation and survival from one another and, thus, the prognosticfindings should be confirmed in a prospective setting.

The discovery of AR gene amplification and the emerging understanding of the molecular mechanisms of therapy failure may in thefuture translate into new therapy options for patients with prostatecancer. The results illustrate the fact that recurrent hormone-refractory

tumors may not always be androgen independent as often thought(3 1). These cells may in fact be highly dependent on the residual lowlevels of circulating androgens that originate from the adrenals and are

317

1.0

0.8

0.6

0.4

0.2

0.00


A

0

0

S

EC.)

B

C0

4-'U

[email protected]

EC)

pâ€”0.09(Mantel-Cox)p=0.03 (Broslow)

I@ I I I I I

24 48 72 96 120 144 168

Overallsurvival (months)

1.0

0.8 _________________

0.6

0.4

:@ 24 48 7@ :@

Survival after recurrence (months)

Fig. 2. Survival of the prostate cancer patients according to the AR gene amplification.A, overall survival (from diagnosis to death). The median survival for AR-amplified casesis 135 months in comparison to 70 months in the nonamplified cases. B, survival afterrecurrence.The medianpostrecurrencesurvival for AR-amplifiedcasesis 55 monthsincomparison to 24 months in the nonamplified cases.

prostate cancer, nor does it occur during the clinical progression of

cancer in patients who have not received androgen deprivationtherapy.

Second, among the patients who had been treated, AR amplificationwas most likely to occur in recurrent tumors from patients who hadinitially responded well to androgen deprivation and whose endocrinetherapy response had lasted for more than 12 months. These aretypical features of the most androgen-sensitive tumors whose proliferation is critically dependent on androgens and, therefore, any increase of AR copy number is likely to provide a survival or proliferative advantage to the cells by allowing them to resume androgendependent growth (12). In fact, none of the recurrent tumors frompatients who showed no initial response to androgen deprivation orwhose disease recurred less than 12 months after therapy containedAR amplification. This suggests that AR amplification represents amechanism for acquired therapy resistance, whereas primary resistance to androgen deprivation is apparently caused by mechanismsunrelated to AR copy number.

Third, we observed an increased AR mRNA expression level in therecurrent AR-amplified tumors as compared with those seen in thecorresponding nonamplified primary tumors from the same patients.Increased expression is a typical feature of amplification target genes,

SIAR amplification:â€”-â€”Yos(n14)â€” No (n=38)

[email protected](Mantel-Cox)pâ€”0.03(Breslow)



Table 2 Cotnparis)n ofAR copy numbers by interphase FISH analysis and AR expression levels by mRNA in situ hvbridizatiotumors were amplifledfor AR at the time of recurrencen

in six paired specimens from patientswhosePatientPrimary

tumorRecurrenttumorCopy

no. by FISHâ€•Expression by mRNA ISHb Copy no. by FISHâ€•Expression by mRNAISHâ€•I

234561.0

1.01.11.11.4

N.S.'+

11.5+ 11.4

+++ 10.5

+ 4.1+ 20.2

N.S. 5.7+++

+++++++

++++++

ANDROGEN RECEVI'OR GENE AMPLIFICATION

BA

LOW D@ C

@z@-@ :P*, :@.,4

@@ ,@ 4*

@ .â€¢@i@-@

-@@ 1 mm

i-;; HIGH __

Fig. 3. Androgen receptor gene copy number analysis using FISH and mRNA in situ hybridization of a tissue sections from an untreated primary tumor and a locally recurrent tumorafter failure of androgen deprivation therapy. The specimens came from the same patient. A. interphase FISH analysis of nuclei from the primary tumor with probes for AR (green)and DXZI (red). Only one copy of both AR and DXZI is seen. B, interphase FISH analysis of nuclei from the recurrent tumor demonstrating high level AR amplification. C, mRNAin situ hybridization analysis of an entire tissue section from the untreated primary tumor shows moderate expression level of the AR gene; D, mRNA in situ hybridization of the locallyrecurrent, therapy-resistant tumor shows highly elevated AR expression. The intensity of mRNA in situ hybridization is expressed in pseudocolors.

converted to bioactive dihydrotestosterone in the prostatic tissue by5a-reductase (32). The critical question for additional studies iswhether more complete removal of these androgens could improvesurvival in the subgroup of patients with AR-amplified tumors. Itshould be emphasized that all of the patients evaluated here receivedconventional androgen deprivation and were mostly treated in the1980s. Studies are now required to assess whether MAB allows for theeradication of tumors that have recurred after androgen deprivation aswell as whether AR amplification also arises if the patients havereceived MAB as the first-line hormonal therapy. We have thus fardocumented one patient whose tumor recurred with AR amplification

and was subsequently treated with MAB. An excellent initial treatment response was obtained, but the relapse was short-lived (33).Such short-term responses have previously been documented in about20â€”35%of unselectedpatientswhohavereceivedMABtherapyaftercastration has failed (34, 35). AR amplification could turn out to be auseful marker to identify patients who are most likely to benefit fromMAB as a second-line therapy. Finally, the results from the retrospective studies reported here warrant the incorporation of AR copynumber analysis in prospective studies of hormonal therapy in prostate cancer. We have now gained extensive experience on such analyses and strongly recommend the use of the highly sensitive inter

(,AveragenumberofARcopiespercell(correctedforthenumberofcopiesofreferenceprobe).,ExpressionbymRNAinsituhybridization(ISH)weresemiquantitativelyscoredtohaveeitherlow(+),moderate(++),orhigh(+++)hybridizationintensity.@ N.S., no sample available for study.

318




specific expression of pituitary adenylate cyclase activating polypeptide (PACAP)mRNA in the rat seminiferous tubules. Endocrinology, /35: 2291â€”2294,1994.

17. Govindan, M. V. Specific region in hormone binding domain is essential for hormonebinding and trans-activation by human androgen receptor. Mol. Endocrinol., 4:417â€”427,1990.

18. Ris-Stalpers, C., Trifiro, M. A., Kuiper, G. G. J. M., Jenster, 0., Romalo, G., Sri I.,van Rooij,H. C. J., Kaufman,M., Rosenfield,R. L., Liao,S., Schweikert,H-U.,Trapman, J., Pinsky, L.. and Brinkmann, A. 0. Substitution of aspartatic acid 686 byhistidineor asparginein thehumanandrogenreceptorleadsto functionally inactiveandrogen receptors with altered hormone binding characteristics. Mol. Endocrinol., 5:1562â€”1569,1991.

19. Brinkmam, A. 0., Faber, P. W., van Rooij, H. C. J., Kuiper, G. G. J. M., Ris, C.,Klaassen, J. A. G. M., van der Korput, J. A. G. M., Voorhorst@M. M., van Laar, J. H.,Mulder, E.. and Trapman, J. The human androgen receptor domain structure,genomicorganizationandregulationofexpression.J. SteroidBiochem.,34: 307â€”310,1989.

20. Cleutjens, K. B. J. M., van Eekelen, C. C. E. M., van der Korput, H. A. 0. M.,Brinkman, A. 0., and Trapman, J. Two androgen response regions cooperate insteroid hormone regulated activity of the prostate specific antigen promoter. J. Biol.Chem.,271:6379â€”6388,1996.

21.Dixon,W.J.BMDPStatisticalSoftware.Berkeley:Universityof CalifomiaPress,1981.

22. Van der Kwast, I. H., Schalken, J. Ruizeeld de Winter, J. A., van Vroonhoven,C. C. J., Mulder, E., Boersma, W.. and Trapman, J. Androgen receptors in endocrinetherapy-resistant human prostate cancer. tnt. J. Cancer, 48: 189â€”193,1991.

23. Gdker, E., Waltham, M., Kheradpour, A., Trippett, I., Mazumdar, M., Elisseyeff, Y.,Schnieders, B., Steinherz, P., Tan, C., Berman, E., and Bertino, J. R. Amplification ofthe dihydrofolate reductase gene is a mechanism of acquired resistance to methotrexate in patients with acute lymphoplastic leukemia and is correlated with p53 genemutations. Blood, 86: 677â€”684,1995.

24. Hamlin, J. L., Leu, I-H., Vaughn, J. P., Ma, C., and Dijkwell, P. A. Amplification ofDNA sequences in mammalian cells. Prog. Nucleic Acid Res., 4!: 203â€”239,1991.

25. WahI, 0. M. The importance of circular DNA in mammalian gene amplification.CancerRes.,49: 1333â€”1340,1989.

26. Wright, J. A., Smith, H. S., Watt, F. M., Hancock, M. C., Hudson, D. L., and Stark,G. R. DNA amplificationis rarein normalhumancells.Proc.NatI. Acad.Sci.USA,87: 1791â€”1795,1990.

27. Tlsty, I. D., White, A., and Sanchez, J. Suppression of gene amplification in humancell hybrids. Science (Washington DC), 255: 1425â€”1427,1992.

28. Stark, 0. R. Regulation and mechanisms of mammalian gene amplification. Adv.Cancer. Res., 61: 87â€”113, 1993.

29. Slamon, D. J., Godolphin, W., Jones, L. A., Holt, J. A., Wong, S. G., Keith, D. E..,Levin, W. J., Stuart, S. G., Udove, J., Ullrich, A.. and Press, M. F. Studies ofHER-2/neu proto-oncogene in human breast and ovarian cancer. Science (WashingtonDC),244:707â€”712,1989.

30. Brison, 0. Gene amplification and tumor progression. Biochim. Biophys. Acts, I 155:25â€”41,1993.

31. Logothetis, C. J., Hoosein, N. M., and Hsieh, J. I. The clinical and biological study ofandrogen independentprostatecancer (A! Pca). Semin. Oncol., 21: 620â€”629,1994.

32. Labrie, F., Belanger, A., Dupont, A., Luu-The, V., Simard, J., and Labrie, C. Sciencebehind total androgen blockade: from gene to combination therapy. Clin. Invest.Med., 16: 475â€”492,1993.

33. Palmberg, C., Koivisto, P., Hyytinen, E., Isola, J., Visakorpi, I., Kallioniemi, 0-P.,and Tammela, I. Androgen receptor gene amplification in a recurrent prostate cancerafter monotherapy with nonsteroidal pure antiandrogen â€œCasodexâ€•and favorableresponse to maximal androgen blockade. Eur. Urol., in press, 1996.

34. Crawford, E. D., Eisenberger, M. A., McLeod, D. G., Spaulding, J. I., Benson, R.,Dorr, A., Blumenstein, B. A., Davis, M. A., and Goodman, P. J. A controlled trial ofleuprolide with and without flutamide in prostatic carcinoma. N. Engl. J. Med., 321:419â€”424,1989.

35. Prostate Cancer Trialists' Collaborative Group. Maximum androgen blockade inadvanced prostate cancer: an overview of 22 randomised trials with 3283 deaths in5710 patients. Lancet, 346: 265â€”269,1995.

319

phase FISH technique for this purpose. Southern and slot blot studiesare complicated by the inadequate specificity of most AR cDNAprobes and the poor histological representativeness of many specimens obtained by biopsy or transuretheral resection.

ACKNOWLEDGMENTS

We thankLeena Pankko,ArjaAlkula,Anne Luuri,Hem Sleddens,Connyvan Eckelen, and Marcel van der Weiden for technical assistance.

REFERENCES

1. Potosky, A. L., Miller, B. A., Albertsen, P. C., and Kramer, B. S. The role ofincreasing detection in the rising incidence of prostate cancer. J. Am. Med. Assoc.,273: 548â€”552,1995.

2. Gittes, R. F. Carcinoma of the prostate. N. Engl. J. Med., 324: 236â€”245, 1991.3. Newmark, J. R., Hardy, D. 0., Tonb, D. C., Carter, B. S., Epstein, J. I., Isaacs, W. B.,

Brown, I. R., and Barrack, E. R. Androgen receptor gene mutations in humanprostate cancer. Proc. Ned. Aced. Sci. USA, 89: 6319â€”6323,1992.

4. Suzuki, H., Sato, N., Watabe, Y., Masai, M., Seino, S., and Shimazaki, J. Androgenrecept@genemutationsinhumanprostatecancer.3.SteroidBiochem.Mol.Biol.,46:759â€”765,1993.

5. Culig, Z., Hobisch, A., Cronauer, M.. V., Cato, A. C., Hittmair, A., Radmayr, C.,Eberle, J., Bartsch, G., and Klocker, H. Mutant androgen receptor detected in anadvan@d-stageprostaticcarcinomais activatedby adrenalandrogensandprogesterone. Mol. Endocrinol., 7: 1541â€”1550,1993.

6. Gaddi@nti,J. P., McLeod,D. 0., Heidenberg,H. B., Sestethenn,I. A., Finger,M. J., MoW,J, w., and Srivastava, S. Frequent detection of codon 877 mutation in the androgenreceptca@genein advancedprostatecancers.CancerRes.,54: 2861â€”2864,1994.

7. Ruizeveld de Winter. J. A., Janssen, P. J., Sleddens, H. M., Verleun-Mooijman, M. C..Trapman, I., Brinkmann, A. 0., Santerse, A. B.. Schroder, F. H., and van der Kwast,T. H. Androgen receptor status in localized and locally progressive hormone refractory human prostate cancer. Am. J. Pathol., 144: 735â€”746,1994.

8. Schoenberg,M. P.,Hakimi,J.M.,Wang,S.,Bova,G.S.,Epstein,J. I., Fiscbeck,K. H., lasses, W. B., Walsh, P. C., and Barrack, E. R. Microsatellite mutation(CAG,,, _. 18) in the androgen receptor gene in human prostate cancer. Biochem.Biophys.Rca.Commun.,198:74â€”80,1994.

9. Taplin, M. E., Bubley, G. J., Shuster, I. D., Frantz, M. E., Spooner, A. E., Ogata,G. K., Keer, H. N., and Balk, S. P. Mutation of the androgen-receptorgene inmetastatic androgen-independent prostate cancer. N. EngI. J. Med., 332: 1393â€”1398,1995.

10. Visakorpi, 1., Hyytinen, E., Koivisto, P., Tanner, M., Keinanen, R., Palmberg, C.,Palotie, A., Tammela, I., Isola, J., and Kallioniemi, 0-P. In vivoamplification of theandrogen receptor gene and progression of human prostate cancer. Nat. Genet., 9:401â€”406, 1995.

11. Alt, F. W., Kellems, R. E.. Bertino, J. R., and Schimke, R. I. Selective multiplicationofdihydrofolate reductase genes in methotrexate-resistant variants of cultured murinecells.I. Biol.Chem.,253: 1357â€”1370,1978.

12. Kellens, R. E. Gene Amplification in Mammalian Cells. A comprehensive guide.New York: Marcel Dekker Inc., 1993.

13. Kallioniemi, 0-P., and Visakorpi, I. Genetic basis and clonal evolution of humanprostate cancer. Adv. Cancer Res., 68: 225-255, 1996.

14. Murphy, G. P., Slack, N. H.. and Minelman, A. Use ofestramustine phosphate by theNational Prostatic Cancer Project and by Roswell Park Memorial Institute. Urology,23 (Suppl. 6): 54â€”63,1984.

15. Hyytinen, E., Visakorpi, I., Kallioniemi, A., Kallioniemi, 0-P., and Isola, J. Improved technique for analysis of formalin-fixed, paraffin-embedded tumors by fluorescence in situ hybridization. Cytometry, 16: 93â€”99,1994.

16. Kononen,J., Paavola,M., PenttilS,T-L, Parvinen,M., andPelto-Huikko,M. Stage



1997;57:314-319. Cancer Res Pasi Koivisto, Juha Kononen, Christian Palmberg, et al. Prostate CancerMechanism for Androgen Deprivation Therapy Failure in Androgen Receptor Gene Amplification: A Possible Molecular

Updated version

http://cancerres.aacrjournals.org/content/57/2/314

Access the most recent version of this article at:

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

Subscriptions

Reprints and

[email protected] at

To order reprints of this article or to subscribe to the journal, contact the AACR Publications

Permissions

Rightslink site. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC)

.http://cancerres.aacrjournals.org/content/57/2/314To request permission to re-use all or part of this article, use this link



http://cancerres.aacrjournals.org/cgi/alerts

mailto:[email protected]



androgen receptor gene amplification: a possible ......reaction; mab, maximal androgen blockade. 314...

Documents