Carcinogenesis vol.23 no.6 pp.967–975, 2002

Expression profile of differentially-regulated genes duringprogression of androgen-independent growth in human prostatecancer cells

Dev Karan1, David L.Kelly2, Angie Rizzino1,2,Ming-Fong Lin1,2,3 and Surinder K.Batra1,2,4

1Department of Biochemistry and Molecular Biology, 2Eppley Institute forResearch in Cancer and Allied Diseases and 3Department of Surgery,Section of Urology, University of Nebraska Medical Center, Omaha, NE,USA4To whom correspondence should be addressed at: Department ofBiochemistry and Molecular Biology, The Eppley Institute for Research inCancer and Allied Diseases, University of Nebraska Medical Center, 984525Nebraska Medical Center, Omaha, NE 68198-4525, USAEmail: sbatra@unmc.edu

Because of the heterogeneous nature of prostate cancer,identifying the molecular mechanisms involved during thetransition from an androgen-sensitive to an androgen-independent phenotype is very complex. An LNCaP cellmodel that recapitulates prostate cancer progression, com-prising early passage androgen-sensitive (LNCaP-C33) andlate passage androgen-independent (LNCaP-C81) pheno-types, would help to provide a better understanding ofsuch molecular events. In this study, we examined the genesexpressed by LNCaP-C33 and LNCaP-C81 cells usingcDNA microarrays containing 1176 known genes. Thisanalysis demonstrated that 34 genes are up-regulated andeight genes are down-regulated in androgen-independentcells. Northern blot analysis confirmed the differencesidentified by microarrays on several candidate genes,including c-MYC, c-MYC purine-binding transcriptionfactor (PuF), macrophage migration inhibitory factor(MIF), macrophage inhibitory cytokine-1 (MIC-1), lactatedehydrogenase-A (LDH-A), guanine nucleotide-bindingprotein Gi, α-1 subunit (NBP), cyclin dependent kinase-2(CDK-2), prostate-specific membrane antigen (PSM), cyclinH (CCNH), 60S ribosomal protein L10 (RPL10), 60Sribosomal protein L32 (RPL32), and 40S ribosomal proteinS16 (RPS16). These differentially-regulated genes are cor-related with progression of human prostate cancer andmay be of therapeutic relevance as well as an aid inunderstanding the molecular genetic events involved in thedevelopment of this disease’s hormone-refractory behavior.

Introduction

Prostate cancer is the most commonly diagnosed malignancyand the second leading cause of cancer deaths in the malepopulation of the United States. Progression of prostate canceris characterized by a transition from an androgen-dependentto an androgen-independent phenotype, which determines theclinical outcome of the patient. The initial androgen-dependent

Abbreviations: CCNH, cyclin H; CDK-2, cyclin dependent kinase-2;LDH-A, lactate dehydrogenase-A; MIC-1, macrophage inhibitory cytokine-1;MIF, macrophage migration inhibitory factor; NBP, guanine nucleotide-bindingprotein Gi, alpha-1 subunit; PSA, prostate-specific antigen; PSM, prostate-specific membrane antigen; PuF, purine-binding transcription factor; RPL10,60S ribosomal protein L32, RPS16, 40S ribosomal protein S16.

© Oxford University Press 967

stage of prostate cancer is organ-confined and is curableby means of surgery or other modalities. After the diseasemetastasizes to other organs and acquires androgen-independ-ency, it frequently displays a fatal behavior. At the advancedstage of prostate cancer, most patients who are treated byandrogen deprivation therapy exhibit a dramatic regression ofthe androgen-dependent cancer cells (1). Initially, 70–80% ofthe patients respond to the androgen ablation therapy, but thetumor eventually progresses to an androgen-independent stage,thereby resulting in a poor prognosis (2,3). Unfortunately, noeffective alternative therapies are currently available for thehormone-refractory stage of this disease. Therefore, exploringthe molecular mechanisms underlying prostate cancer’s pro-gression to the androgen-independent stage is a key to introdu-cing new therapies. In this regard, using DNA microarraytechnology to define alterations in gene expression associatedwith prostate cancer may be an efficient way of uncoveringclues to the specific molecular derangements that contributeto its pathogenesis and thus of identifying potential targets fortherapeutic intervention.

DNA microarray technology allows simultaneous determina-tion of the expression level of thousands of genes. Thisapproach has dramatic ramifications for deciphering and classi-fying gene expression changes in human cancers in a way thatreflects the underlying molecular pathology or that anticipatestheir potential for progression or response to treatment. Overthe last 5 years, DNA array analyses have largely identifiedgenes that potentially play key roles in the malignant behaviorof melanomas (4), embryonic stem cells (5), solid tumors ofthe breast (6), cervix (7), ovaries (8,9), colorectal carcinomas(10), and glioblastomas (11,12). Several recent studies havealso suggested the classification of various cancers based ongene expression profiles established by DNA array analysesfor improving the efficacy of treatment (13–15).

Information concerning the gene expression pattern in pro-state cancer is limited due to the heterogeneous and multifocalbehavior of the disease. Therefore, only a few recent studieshave addressed this issue (16–20). Nevertheless, due to thecomplexity of prostate carcinomas, the results obtained fromarchival specimens would make it difficult to address theassociation between those gene expressions and hormone-refractory growth. To investigate the molecular mechanism ofandrogen-independent growth of prostate cancer cells, we havedeveloped a unique in vitro cell model (21) derived from theLNCaP cell line, which was originally established from humanprostate adenocarcinomas (22). This tumor cell model, whichincludes different LNCaP cells, closely resembles differentstages of tumor progression and the acquisition of hormone-refractiveness in advanced prostate cancer patients (21,23).We previously observed that genetic alterations increased withdecreased androgen-responsive behavior of the cells in thismodel. Furthermore, those genetic alterations identified in thecell model were seen in clinical specimens (24). Thus, this

D.Karan et al.

cell model recapitulates prostate cancer progression from theandrogen-sensitive to the androgen-independent stage.

In the present study, the LNCaP cell model was used toidentify differentially-regulated genes associated with acquisi-tion of the androgen-independent phenotype of the prostatetumor. A cDNA microarray gene filter containing 1176 knowngenes was employed to analyze genes expressed by earlypassage androgen-sensitive (C33) and late passage androgen-independent (C81) cells of the LNCaP cell model. We identifiedseveral genes that exhibit quantitative differences in theirexpression during progression of the androgen-independentphenotype. These genes appear to be helpful for understandingthe biology of this disease and may prove useful in developingimproved diagnosis and treatment of advanced stage prostatetumor.

Materials and methods

LNCaP cell model and cell culture

The LNCaP cell model used in the present study was developed by Lin et al.(21) and further characterized by Igawa et al. (23). Briefly, three stages ofLNCaP cells during regular maintenance were designated as C33 (LNCaPpassage number �33), C51 (LNCaP passage number between 45 and 70),and C81 (LNCaP passage number between 81 and 120). The cells in thesethree stages exhibit different androgen-responsiveness of growth stimulation;C33 cells are androgen-sensitive while C81 cells are androgen-independent.This cell model recapitulates the progression of human prostate tumor to theadvanced hormone-refractory stage (21,23,24). In the study reported here, weutilized C33 and C81 cells of this LNCaP cell model to analyze thedifferentially-expressed genes during tumor progression. Cells were maintainedin RPMI 1640 medium supplemented with 5% FBS, 1% glutamine, 1%penicillin–streptomycin (Gibco BRL, Grand Island, NY) and passaged weeklyby trypsinization.

RNA isolation

Total RNA was extracted from C33 and C81 stage LNCaP cells by usingguanidine isothiocyanate–cesium chloride ultracentrifugation method (25).RNA concentration was measured spectrophotometrically, and its integritywas analyzed by electrophoresis on a formaldehyde agarose gel.

cDNA microarray hybridization

The mRNA expression profiles of C33 and C81 cells were surveyed usingAtlas Human 1.2 cDNA arrays (Clontech Laboratories, Palo Alto, CA). Forparallel hybridizations, total RNA (5 µg) from each cell population wasreverse transcribed at 37°C for 90 min in a final reaction volume of 30 µlcontaining 5X first strand buffer, 0.1 M DTT, 20 mM dNTP, ReverseTranscriptase, oligo d(T) primers, and [33P]dCTP (10 µCi/ µl with a sp. act.of 3000 Ci/mmol). Probes were purified by chromatography with NICKTM

Column (Pharmacia Biotech, NJ), and then denatured for 3 min in a boilingwater-bath. Denatured radio-labeled probes were added to a 5 ml aliquotof MicroHyb hybridization solution (Research Genetics, Huntsville, AL)containing 5 µg of Cot-1 DNA and 5 µg of Poly d(A). This cDNA probesolution was applied to the pre-hybridized Atlas Human 1.2 cDNA array(2–3 h in MicroHyb solution) and hybridized overnight at 42°C. Afterhybridization, the filters were washed twice with 2X SSC and 1% SDSsolution at 55°C for 20 min followed by a third wash with 0.5X SSC and 1%SDS solution for 15 min. Following the final wash, the filters were exposedto phosphor imaging screens for varying lengths of time to ensure a propercomparison of weakly and strongly expressed genes.

Quantitative analysis of the expressed genes

The phosphor images of hybridized filters were scanned by PhosphoImager(Molecular Dynamics Inc., Sunnyvale, CA), and analyzed using AnalyticalImaging Station (AIS) software (version 3.0 r.1.3; Imaging Research,St Catharine’s, Ontario, Canada). The intensity of hybridization signals foreach spot was determined automatically and corrected for background.Calculated intensities correlate linearly with the concentration of mRNAspresent in the total RNA population from both C33 and C81 cells, since theamount of cDNA attached to the membrane was in excess (10 ng) and thebackground hybridization signals were sufficiently low. For a quantitativedifference in gene-expression between arrays (mRNA from C33 and C81),the intensity value of each known gene was normalized in two ways: to theintensity values of the designed housekeeping genes and to the sum of theintensity values of all of the genes. We found no significant difference in the

968

normalization coefficients generated by the two methods. Comparison of C33and C81 RNA population was performed in two separate parallel hybridizationexperiments. Genes that showed an average induction or reduction of 1.6-foldor higher in both hybridization experiments were considered to be differentiallyexpressed. Correlations and differences in gene expression between C33 andC81 cells were also compared by scatter plot analysis, in which each pointrepresents a particular gene. The location of the point is determined afternormalization by the intensity value of the gene in one cell populationrepresented by the x-axis, with its intensity value in the other cell populationrepresented by the y-axis. In each scatter plot, points that lie close to the45-degree ‘line of identity’ represent genes with similar levels of expressionin both LNCaP cells. The perpendicular distance of a point away from thediagonal line represents the degree to which a gene is differentially expressedbetween C33 and C81 cells of the LNCaP cell model.

Reverse transcription-polymerase chain reaction (RT-PCR)

Oligonucleotide primers were synthesized for genes of interest that showeddifferential expression in microarray analyses. Total RNA (1 µg) was reversetranscribed using the SUPERSCRIPTTMII Rnase H– Reverse TranscriptaseSystem. Samples were subjected to PCR amplification in a total reactionvolume of 50 µl containing 10X PCR buffer (GIBCO BRL®), 50 mM MgCl2(GIBCO BRL®), 10 mM dNTP, 5 pmol concentration of each specific primer,and 2.5 units of Taq DNA polymerase (GIBCO BRL®). The PCR reaction wascarried out in a programmable thermal controller (PTC-100, MJ Research,Inc., Watertown, MA). The reaction mixture was denatured at 94°C for 3 minfollowed by 30 cycles at 94°C for 45 s, annealing at 60°C for 45 s and 72°Cfor 1 min. The final elongation was extended for an additional 20 min. Forsome of the genes, the annealing temperature ranged from 56°-60°C. Theamplified PCR products were resolved electrophoretically on agarose gelstained with ethidium bromide to verify size of the amplified product. Afternucleotide sequence verification for some of the genes, the PCR productswere used to prepare the probes for northern analysis.

Northern blotting

Total RNA (10 µg) was fractionated by gel electrophoresis on 1.0% agarosegel containing 0.66 M formaldehyde and was transferred to nitrocellulosefilter via capillary blotting. cDNA probes were labeled with [32P]dCTP usinga random primed labeling kit (Pharmacia Biotech, Piscataway, NJ). Pre-hybridization and hybridization of the filter were carried out in a solution of50% formamide, 5X SSPE, 5X Denhardt’s reagent, 200 µg/ml of shearedsalmon sperm DNA, and a minimum of 106 c.p.m./ml of probe at 42°C for18 h. After hybridization, filters were washed twice with 2X SSC and 0.1%SDS solution at room temperature for 15 min followed by two washes witha solution of 0.2X SSC and 0.1% SDS at 55°C. These filters were exposedto phosphor imaging screens, and scanned with PhosphoImager (MolecularDynamics, Sunnyvale, CA). Northern blot analysis for the selected genes wasperformed in duplicates.

Results

Filter hybridization and cDNA array quantificationTo address the question of differentially expressed genes duringthe progression of human prostatic adenocarcinomas, we useda cDNA microarray gene filter (Atlas Human 1.2 cDNA arrays)representing 1176 identified genes. Total RNA from earlypassage androgen-sensitive (C33), and late passage androgen-independent (C81) LNCaP cells was reverse transcribed, andthe generated probes were hybridized to microarray gene filters(Figure 1). On these filters, each spot represents a singleknown gene separated from the housekeeping genes locatedat the bottom portion of the filter. The intensity of each spotfor the known genes was calculated after normalization of thesignals with the housekeeping genes as well as with thebackground intensity. The distribution profile of the differenti-ally expressed genes is represented by a log transformed scatterplot analysis (Figure 2). The majority of the genes are locatedclose to the ‘line of identity’ (line with a slope of 1) representingno difference in expression, and only a few genes appearedbelow and above the line of identity, indicating that thesegenes exhibited �2 fold higher expression in either C33 orC81 passage cells. A visual examination of the expressionpatterns between the filters revealed that some of the genes

Differentially-regulated genes in prostate cancer

Fig. 1. The 33P-labeled probes from (A) androgen-sensitive C33 and (B) androgen-independent C81 cells of the LNCaP cell model hybridized to microarraygene filters (Human Cancer 1.2: ClonTech). Each cDNA spot corresponds to a single known gene. The six housekeeping genes are located at the bottom. Atthe upper left and right corners of the gene filter, there is one genomic DNA spot directing the orientation of the filter. The arrow marks indicatedifferentially-regulated genes confirmed by northern blot analysis (→: up-regulated; →: down-regulated in C81 stage cells).

showed false positive results. These were subsequentlyexcluded from the study.

Differential gene expression profile

A detailed analysis of differentially expressed genes identifiedin the present study, explaining their localization on the genefilter and the GenBank accession number, is presented inTable I. Out of the 1176 known genes from the Atlas Array1.2gene filter, only 42 genes (3.6%) showed differences (1.6-foldor higher) in their level of expression between C33 and C81cells. The mRNA expression level for 34 genes (2.9%)

969

exhibited up-regulation, while 8 genes (0.7%) displayed down-regulation in C81 cells as compared to C33 cells.

Confirmation of the gene expression profile by northern blotanalysis

To confirm the quantitative differences in the gene expressionprofile associated with the progression of the androgen-inde-pendent phenotype developed from the androgen-sensitivestage of the LNCaP cell model, northern blot analysis wasperformed on several selected target genes. These included:c-MYC, c-MYC purine-binding transcription factor (PuF),

D.Karan et al.

Fig. 2. A scatter plot of log transformed data viewing the distributionpattern of differentially regulated genes in androgen-sensitive (C33) andandrogen-independent (C81) cells. The solid line indicates the ‘line ofidentity,’ while the dashed lines indicate �2-fold limits.

macrophage migration inhibitory factor (MIF), macrophageinhibitory cytokine-1 (MIC-1), lactate dehydrogenase-A (LDH-A), guanine nucleotide-binding protein Gi, α-1 subunit (NBP),cyclin dependent kinase-2 (CDK-2), prostate specific mem-brane antigen (PSM), cyclin H (CCNH), 60S ribosomal proteinL10 (RPL10), 60S ribosomal protein L32 (RPL32), and 40Sribosomal protein S16 (RPS16). As shown in Figure 3, datafrom northern blot analyses of these genes paralleled closelythe results obtained by DNA microarrays, indicating thesegenes indeed have differential expressions in androgen-sensit-ive and androgen-independent prostate cancer cells.

Discussion

Research in prostate cancer stands to benefit considerably fromDNA microarrays, as evidenced by recent studies (16–20,26).Although, certain differentially-regulated genes have beenidentified, it remains a challenge to prove whether these genescontribute to neoplastic processes or androgen-independency.The molecular events underlying the progression of prostaticadenocarcinomas remain largely unsolved because of theheterogeneity of these tumors. Molecular and biochemicalcharacterization of LNCaP cell model indicates that the cellsat various passages mimic tumor progression as seen in clinicalpatients, indicated by the expression of biomarkers includingthe elevated secretion of prostate specific antigen (PSA) inandrogen depleted conditions (21,23). Other cell models ofandrogen-independent cells, which were derived under andro-gen depletion conditions, may represent only a subpopulationof cancer patients, since the expression of biomarkers (suchas PSA) in those cells may not be consistent with the clinicalphenomenon (27,28). Thus, our LNCaP cell model, to a certainextent, eliminates the possibility of heterogeneity in clinicalsamples, which is largely responsible for the complexity ofthis malignancy. Combining the LNCaP cell model withmicroarray studies has provided an efficient means of identify-ing genes that are differentially-regulated during tumor pro-gression.

It is clear that in the early stage of prostate tumor progressionmost of the cells are androgen-sensitive and gradually acquirethe phenotypes of androgen-independency. This may be dueto the accumulation of genetic and epigenetic changes, leading

970

to the acquisition of androgen-refractiveness. Alternatively, thedevelopment and/or progression of the androgen-independentphenotype may arise from a heterogeneous mass of cellpopulations, i.e. a mixture of androgen-sensitive and androgen-independent cells, in the beginning of cancer. Both thesenotions are well supported by our recent study with progressionof LNCaP cell model (24). Therefore, it is essential to definethe sets of genes that facilitate the phenomenon of prostatetumor development and progression.

In this era of functional genomics, microarray technologyhas provided a great opportunity to analyze the expressionprofile of thousands of transcripts simultaneously. In thiscontext, to elucidate the transcriptional differences associatedwith the acquisition of the androgen-independent phenotypeof human prostatic adenocarcinomas, we performed the studiesusing in vitro LNCaP cell model and DNA microarray techno-logy. The microarray gene filters used in the present studycontain 1176 known genes, categorized into six groups: (A)oncogenes, tumor suppressors, cell-cycle regulators; (B) trans-porters, signal transduction; (C) GDP/GTP exchangers andGTPase stimulators/inhibitors, apoptosis; (D) transcription fac-tors, cell signaling and extracellular communication; (E) cell-surface antigens, cell adhesion, receptors; and (F) stressresponse, cell–cell communication. Further details are availablefrom the ClonTech website, www.clontech.com. It is note-worthy that a large number of genes exhibited up-regulationin C81 cells. Although some of the genes listed in Table Ihave been confirmed by northern blot analysis, not all of theseidentified genes are likely to play an important role in tumorprogression leading to hormone-refractiveness. Some of thesegenes are cell cycle regulated, and may be the outcome of thedisease. We selected 12 genes, which may potentially play animportant role or serve as a useful biomarker for prostatepathogenesis, to confirm their differential expressions bynorthern blotting. Among these, c-MYC, PuF, LDH-A, MIF,MIC-1, NBP, CDK-2, RPL32, RPL10, RPL16 genes exhibitedup-regulation, while PSM and CCNH revealed down-regulationin androgen-independent C81 passage cells.

MYC genes include c-MYC, N-MYC, and L-MYC encodinga family of helix–loop–helix transcription factors containingDNA binding phosphoproteins, which regulate cell growthproliferation and differentiation (29,30). PuF stimulatesc-MYC transcription by binding to the promoter of c-MYCgene (31). In this study, the data clearly show that c-MYCand its transcription factor PuF are up-regulated in androgen-independent C81 cells. The results thus correlated well withprevious reports of overexpression and amplification ofc-MYC gene in prostate cancer (32–34).

Lactate dehydrogenase (LDH) is a tetrameric enzyme com-prising two subunits (LDH-A, and LDH-B). LDH is involvedin metabolic activities, such as glycolytic pathways. Alterationsof the glycolytic pathway, including an enhanced level ofLDH-A, are the distinguishing features of cancer cells(35–37). A significantly higher expression of LDH-A and itstyrophosphorylation have been observed in prostatic carcin-omas compared with benign prostatic hyperplasia (38,39). Inovarian cancer, tyrosylphosphorylation of the LDH-A subunitwas also observed in both normal and malignant tissues, andthe degree of phosphorylation is correlated with the stage ofthis cancer (40). Interestingly, a higher level of LDH-A isdirectly correlated with the c-MYC gene up-regulation (37).

It is well established that macrophages are important inorchestrating the chronic inflammatory response that is respons-

Differentially-regulated genes in prostate cancer

Table I. Differentially expressed genes in androgen-sensitive (LNCaP-C33) and androgen-independent (LNCaP-C81) cells of LNCaP cell model identified bymicroarray analysis

Name of the gene Gene location Accession number Fold difference

Leukemia associated gene1 A 01i Y15227 �2.9MYC proto-oncogene (c-Myc) A03c V00568 �2.5a

C-Myc binding protein MM-1 A03g D89667 �1.9Cyclin-dependent kinase-2 A03j M68520 �3.4Transforming growth factor β-3 (TGF-β3) A046 J03241 �1.8Cell division protein kinase-4 (CDK4) A04j M14505 �1.9Cyclic B1, G2/mitotic-specific A05i M25753 �5.3Prohibitin A06b S85655 �1.9N-Myc proto-oncogene A06c M13228 �3.1Cell cycle progression restoration 2 protein (CPR2) A06m AF011792 �1.9C-MYC transcription factor PuF (PuF) A09b L16785 �2.7a

CDC28 protein kinase 2 A11k X54942 �2.0Putative serine/threonine-protein kinase p78 A13b M80359 �2.0Proto-oncogene RhoA, multidrug resistance protein A13e L25080 �2.0STE-20 related kinase DCHT B04e AF017635 �2.7TSC-22 (TGF-B1 induced transcript 4) C11m U35048 �2.3Guanine nucleotide-binding protein G(1) (NBP) D01a M17219 �8.4a

High-mobility group protein 2 D07b X62534 �2.2Gluthione-S-transferase homolog D09m U90313 �2.0Semaphorin (CD100) D10e U60800 �2.3Macrophage migration inhibitory factor (MIF) E07h M25639 �2.3Macrophage inhibitory cytokine-1 (MIC-1) E09e AF019770 �2.6a

Collagen type VI alpha-2 E12n M34570 �1.960S ribosomal protein L10 (RPL10) F01k M73791 �2.7a

Regulator of G-protein signaling 3 F04k U27655 �2.0Lactate dehydrogenase-A (LDH-A) F05d X02152 �3.2a

60S ribosomal protein L32 (RPL32) F05k X03342 �2.8a

G-rich sequence factor-1 F08i U07231 �2.4Splicing factor, arginine/serine-rich 7 F10i L22253 �2.1Ribosomal protein S21 F11k L04483 �2.1Ribonuclease 6 F12i U85625 �2.2Adenylosuccinate lyase F14c X65867 �2.140S ribosomal protein S16 (RPS16) F14j M60854 �1.6a

Adducin 3 F141 U37122 �2.6Prostate-specific membrane antigen (PSM) A09g M99487 �2.1a

Cyclin H (CCNH) A11i U11791 �3.5a

Serine/threonine protein kinase C01d Y10256 �4.2NET transcription factor C01j Z36715 �2.1Tumor necrosis factor receptor-2 C03a M32315 �2.3TGF-B superfamily receptor type 1 ALK-1 D07i L17075 �2.0Teratocarcinoma-derived growth factor E05c M96956 �2.3Interleukin-10 E14f M57627 �2.4

�ve number indicates up-regulation and the –ve number indicates the down-regulation of the genes in androgen-independent, highly aggressive, late stagecells (LNCaP-C81) of the LNCaP cell model.aExpression level of these genes was confirmed by northern blot analysis.

ible for several human diseases. Results of this study showedan elevated expression level of two macrophage regulatoryfactors, MIC-1 and MIF, in androgen-independent C81 cells.MIC-1, also known as prostate derived factor (PDF), is amember of the transforming growth factor-β superfamily(41,42) and is highly regulated by androgens, as orchitectomyresults in a remarkable decrease in prostate size and a similardecrease in MIC-1 expression level compared to those ofcontrol animals (43). This molecule is unique in that it doesnot require a signal peptide for correct folding or secretion ofthe mature peptide (44). Recent studies have clearly shown anincreased level of expression of MIC-1 in archival prostatetumor specimens (20). Similarly, MIF is also associated withthe metastatic stage of prostate cancer (45,46), and a significantpositive correlation was observed with higher grades ofGleason’s score (47). Interestingly, it is proposed that MIFfunctions as a mediator of several malignancies (48,49).

Likewise, our results identified higher expression of NBP(a member of G protein signal transduction family) in androgen-independent cells. An increased level of G protein α subunits

971

has been linked to various human diseases such as cervicalcancer (7,50), hepatocellular carcinomas (51), heart failure(52), and goiter (53), and stimulates uncontrolled thyroidcell proliferation (54). The up-regulation of cyclin-dependentkinase-2 (CDK-2) in androgen-independent cells reveals itsinvolvement in prostate tumor progression, as the cell cycle isregulated by sequential events of cyclin-dependent kinases(55). In human lung carcinomas, a higher expression ofCDK-2 is associated with uncontrolled tumor cell proliferation(56). Simultaneously, three ribosomal proteins (RPL10, RPL32,and RPS16) also showed up-regulation in C81 cells. It issuggested that ribosomal proteins mainly participate in proteinsynthesis, but additionally they are likely to be involved inthe neoplastic transformation of the cells (57,58). Overexpres-sion of several ribosomal proteins including RPS16 has beenreported in colon, breast, liver, and pancreatic tumors (59–62).Until this study, analysis of ribosomal protein genes in prostatecancer was very limited.

Although many genes were up-regulated, we also observeddown-regulation of several genes in androgen-independent

D.Karan et al.

Fig. 3. Northern blot analysis confirming the expression profile for some of the genes obtained from the microarray gene filter. (A) Up-regulated genes; (B)down-regulated genes in androgen-independent cells of the LNCaP cell model.

972

Differentially-regulated genes in prostate cancer

cells (Table I). Northern blot analysis further confirmed down-regulation of the selected genes that included the prostatespecific membrane antigen (PSM) and cyclin H (CCNH).CCNH is a regulatory component of the cyclin-dependentkinase (CDK)-activating kinase (CAK) that is part of transcrip-tional factor IIH multiprotein complex. The CAK complex isrequired for RNA polymerase II transcription, nucleotideexcision repair, and phosphorylation of p53 (63,64). Althoughthe functional role of CCNH has not been defined in prostatecancer, its down-regulation can affect regulation of p53 withDNA repair and the basal transcriptional machinery in tumorcells (64). PSM is a membrane-bound glycoprotein highlyrestricted to prostate (65,66). An alternatively spliced variant,PSM’, is also found in normal prostate. The variant PSM’cDNA is shorter (2387 nucleotides) than PSM (2653 nucleot-ides) and lacks signal and transmembrane sequences. Prostatetumors and tumor cell lines express predominantly PSMcompared to PSM’ (67). A higher expression of PSM is seenin hormone-refractory and metastatic conditions and, therefore,has been suggested as a diagnostic marker (68–70). On theother hand, a study on 236 normal individuals and cancerpatients by Beckett et al. (71) showed an increased level ofPSM in healthy male and female population. Its expressionlevels increased with age, and the highest level was observedin breast cancer patients. In addition, PSM is not completelyrestricted to prostatic tissue but also expressed in brain, heart,liver, kidney, lung, and spleen (72). Hence, the use of PSMas a specific biomarker for prostate cancer is not clear. Furtherinvestigations are needed in the functional/diagnostic roles ofPSM as well as CCNH in prostate cancer progression.

In conclusion, we have identified 34 genes that appear tobe up-regulated and eight genes that are down-regulated asLNCaP cells progress in vitro from an androgen-sensitive toandrogen-independent phenotype. An association of MIC-1,MIF, and c-MYC genes with prostate tumor progressionhas already been reported in clinical samples (20,34,45–47).Additional studies will be needed to determine whether thegenes identified in this study are key factors in the alteredphenotype of the LNCaP cells, and whether these genes canserve as useful biomarkers in androgen-independent growthof prostate cancer cells.

Acknowledgements

This work was supported, in part, by grants from the NationalCancer Institute (CA 88184), the Nebraska Health Department/Eppley Cancer Center LB 595 and the Program of NebraskaResearch Initiative. We thank Mr Erik Moore and Ms Fen-FenLin for their technical assistance, and Ms Kristi L.W.Berger,communications specialist and editor, for editing the manu-script.

References

1.Koivisto,P., Kolmer,M., Visakorpi,T. and Kallioniemi,O.P. (1998) Androgenreceptor gene and hormonal therapy failure of prostate cancer. Am. J.Pathol., 152, 1–9.

2.Blackard,C.E., Byar,D.P. and Jordan,W.P.J. (1973) Orchiectomy foradvanced prostatic carcinoma. A reevaluation. Urology, 1, 553–560.

3.Gittes,R.F. (1991) Carcinoma of the prostate [see comments]. N. Engl. J.Med., 324, 236–245.

4.Huang,F., Adelman,J., Jiang,H., Goldstein,N.I. and Fisher,P.B. (1999)Identification and temporal expression pattern of genes modulated duringirreversible growth arrest and terminal differentiation in human melanomacells. Oncogene, 18, 3546–3552.

973

5.Kelly,D.L. and Rizzino,A. (2000) DNA microarray analyses of genesregulated during the differentiation of embryonic stem cells. Mol. Reprod.Dev., 56, 113–123.

6.Hoch,R.V., Thompson,D.A., Baker,R.J. and Weigel,R.J. (1999) GATA-3 isexpressed in association with estrogen receptor in breast cancer. Int. J.Cancer, 84, 122–128.

7.Shim,C., Zhang,W., Rhee,C.H. and Lee,J.H. (1998) Profiling ofdifferentially expressed genes in human primary cervical cancer bycomplementary DNA expression array. Clin. Cancer Res., 4, 3045–3050.

8.Wang,K., Gan,L., Jeffery,E., et al. (1999) Monitoring gene expressionprofile changes in ovarian carcinomas using cDNA microarray. Gene, 229,101–108.

9.Welsh,J.B., Zarrinkar,P.P., Sapinoso,L.M., Kern,S.G., Behling,C.A.,Monk,B.J., Lockhart,D.J., Burger,R.A. and Hampton,G.M. (2001) Analysisof gene expression profiles in normal and neoplastic ovarian tissue samplesidentifies candidate molecular markers of epithelial ovarian cancer. Proc.Natl Acad. Sci. USA, 98, 1176–1181.

10.Kitahara,O., Furukawa,Y., Tanaka,T., Kihara,C., Ono,K., Yanagawa,R.,Nita,M.E., Takagi,T., Nakamura,Y. and Tsunoda,T. (2001) Alterations ofgene expression during colorectal carcinogenesis revealed by cDNAmicroarrays after laser-capture microdissection of tumor tissues and normalepithelia. Cancer Res., 61, 3544–3549.

11.Sehgal,A., Boynton,A.L., Young,R.F., Vermeulen,S.S., Yonemura,K.S.,Kohler,E.P., Aldape,H.C., Simrell,C.R. and Murphy,G.P. (1998)Application of the differential hybridization of Atlas Human expressionarrays technique in the identification of differentially expressed genes inhuman glioblastoma multiforme tumor tissue. J. Surg. Oncol., 67, 234–241.

12.Fuller,G.N., Rhee,C.H., Hess,K.R., Caskey,L.S., Wang,R., Bruner,J.M.,Yung,W.K. and Zhang,W. (1999) Reactivation of insulin-like growth factorbinding protein 2 expression in glioblastoma multiforme: a revelation byparallel gene expression profiling. Cancer Res., 59, 4228–4232.

13.Su,A.I., Welsh,J.B., Sapinoso,L.M., et al. (2001) Molecular classificationof human carcinomas by use of gene expression signatures. Cancer Res.,61, 7388–7393.

14.Golub,T.R., Slonim,D.K., Tamayo,P., et al. (1999) Molecular classificationof cancer: class discovery and class prediction by gene expressionmonitoring. Science, 286, 531–537.

15. Alizadeh,A.A., Eisen,M.B., Davis,R.E., et al. (2000) Distinct types ofdiffuse large B-cell lymphoma identified by gene expression profiling.Nature, 403, 503–511.

16. Bull,J.H., Ellison,G., Patel,A., Muir,G., Walker,M., Underwood,M.,Khan,F. and Paskins,L. (2001) Identification of potential diagnostic markersof prostate cancer and prostatic intraepithelial neoplasia using cDNAmicroarray. Br. J. Cancer, 84, 1512–1519.

17.Chetcuti,A., Margan,S., Mann,S., Russell,P., Handelsman,D., Rogers,J. andDong,Q. (2001) Identification of differentially expressed genes in organ-confined prostate cancer by gene expression array. Prostate, 47, 132–140.

18.Dhanasekaran,S.M., Barrette,T.R., Ghosh,D., Shah,R., Varambally,S.,Kurachi,K., Pienta,K.J., Rubin,M.A. and Chinnaiyan,A.M. (2001)Delineation of prognostic biomarkers in prostate cancer. Nature, 412,822–826.

19.Luo,J., Duggan,D.J., Chen,Y., Sauvageot,J., Ewing,C.M., Bittner,M.L.,Trent,J.M. and Isaacs,W.B. (2001) Human prostate cancer and benignprostatic hyperplasia: molecular dissection by gene expression profiling.Cancer Res., 61, 4683–4688.

20.Welsh,J.B., Sapinoso,L.M., Su,A.I., Kern,S.G., Wang-Rodriguez,J.,Moskaluk,C.A., Frierson,H.F., Jr. and Hampton,G.M. (2001) Analysis ofgene expression identifies candidate markers and pharmacological targetsin prostate cancer. Cancer Res., 61, 5974–5978.

21.Lin,M.F., Meng,T.C., Rao,P.S., Chang,C., Schonthal,A.H. and Lin,F.F.(1998) Expression of human prostatic acid phosphatase correlates withandrogen-stimulated cell proliferation in prostate cancer cell lines. J. Biol.Chem., 273, 5939–5947.

22.Horoszewicz,J.S., Leong,S.S., Kawinski,E., Karr,J.P., Rosenthal,H.,Chu,T.M., Mirand,E.A. and Murphy,G.P. (1983) LNCaP model of humanprostatic carcinoma. Cancer Res., 43, 1809–1818.

23. Igawa,T., Lin,F.F., Lee,M.S., Karan,D., Batra,S.K. and Lin,M.F. (2002)Establishment and characterization of androgen-independent humanprostate cancer LNCaP cell model. Prostate (in press).

24.Karan,D., Schmied,B.M., Dave,B.J., Wittel,U.A., Lin,M.F. and Batra,S.K.(2001) Decreased androgen-responsive growth of human prostate canceris associated with increased genetic alterations. Clin. Cancer Res., 7,3472–3480.

25.Batra,S.K., Kern,H.F., Worlock,A.J., Metzgar,R.S. and Hollingsworth,M.A.(1991) Transfection of the human Muc 1 mucin gene into a poorlydifferentiated human pancreatic tumor cell line, Panc1: integration,expression and ultrastructural changes. J. Cell Sci., 100, 841–849.

D.Karan et al.

26.Xu,J., Stolk,J.A., Zhang,X., Silva,S.J., Houghton,R.L., Matsumura,M.,Vedvick,T.S., Leslie,K.B., Badaro,R. and Reed,S.G. (2000) Identificationof differentially expressed genes in human prostate cancer using subtractionand microarray. Cancer Res., 60, 1677–1682.

27.Lu,S., Tsai,S.Y. and Tsai,M.J. (1999) Molecular mechanisms of androgen-independent growth of human prostate cancer LNCaP-AI cells.Endocrinology, 140, 5054–5059.

28.Thalmann,G.N., Sikes,R.A., Wu,T.T., Degeorges,A., Chang,S.M., Ozen,M.,Pathak,S. and Chung,L.W. (2000) LNCaP progression model of humanprostate cancer: androgen-independence and osseous metastasis. Prostate,44, 91–103.

29.Ryan,K.M. and Birnie,G.D. (1996) Myc oncogenes: the enigmatic family.Biochem. J., 314, 713–721.

30.Schmidt,E.V. (1999) The role of c-myc in cellular growth control.Oncogene, 18, 2988–2996.

31.Agou,F., Raveh,S., Mesnildrey,S. and Veron,M. (1999) Single strand DNAspecificity analysis of human nucleoside diphosphate kinase B. J. Biol.Chem., 274, 19630–19638.

32.Buttyan,R., Sawczuk,I.S., Benson,M.C., Siegal,J.D. and Olsson,C.A. (1987)Enhanced expression of the c-myc protooncogene in high-grade humanprostate cancers. Prostate, 11, 327–337.

33.Fleming,W.H., Hamel,A., MacDonald,R., Ramsey,E., Pettigrew,N.M.,Johnston,B., Dodd,J.G. and Matusik,R.J. (1986) Expression of the c-myc protooncogene in human prostatic carcinoma and benign prostatichyperplasia. Cancer Res., 46, 1535–1538.

34. Jenkins,R.B., Qian,J., Lieber,M.M. and Bostwick,D.G. (1997) Detectionof c-myc oncogene amplification and chromosomal anomalies in metastaticprostatic carcinoma by fluorescence in situ hybridization. Cancer Res., 57,524–531.

35.Matusiewicz,S.P., Williamson,I.J., Sime,P.J., Brown,P.H., Wenham,P.R.,Crompton,G.K. and Greening,A.P. (1993) Plasma lactate dehydrogenase:a marker of disease activity in cryptogenic fibrosing alveolitis and extrinsicallergic alveolitis? Eur. Respir. J., 6, 1282–1286.

36.Miyajima,H., Takahashi,Y. and Kaneko,E. (1995) Characterization of theglycolysis in lactate dehydrogenase-A deficiency. Muscle Nerve, 18,874–878.

37.Shim,H., Dolde,C., Lewis,B.C., Wu,C.S., Dang,G., Jungmann,R.A., Dalla-Favera,R. and Dang,C.V. (1997) c-Myc transactivation of LDH-A:implications for tumor metabolism and growth. Proc. Natl Acad. Sci. USA,94, 6658–6663.

38.Li,S.S., Pan,Y.E., Sharief,F.S., Evans,M.J., Lin,M.F., Clinton,G.M. andHolbrook,J.J. (1988) Cancer-associated lactate dehydrogenase is atyrosylphosphorylated form of human LDH-A, skeletal muscle isoenzyme.Cancer Invest., 6, 93–101.

39.Sharief,F.S., Mohler,J.L., Sharief,Y. and Li,S.S. (1994) Expression ofhuman prostatic acid phosphatase and prostate specific antigen genes inneoplastic and benign tissues. Biochem. Mol. Biol. Int., 33, 567–574.

40.Chow,S.N., Lin,J.K., Li,S.S. and Chien,C.H. (1997) Identification of LDH-ras p21 protein complex and expression of these genes in human ovariancancer. Gynecol. Oncol., 64, 114–120.

41.Bootcov,M.R., Bauskin,A.R., Valenzuela,S.M., et al. (1997) MIC-1, anovel macrophage inhibitory cytokine, is a divergent member of the TGF-beta superfamily. Proc. Natl Acad. Sci. USA, 94, 11514–11519.

42.Fairlie,W.D., Moore,A.G., Bauskin,A.R., Russell,P.K., Zhang,H.P. andBreit,S.N. (1999) MIC-1 is a novel TGF-beta superfamily cytokineassociated with macrophage activation. J. Leukoc. Biol., 65, 2–5.

43.Paralkar,V.M., Vail,A.L., Grasser,W.A., et al. (1998) Cloning andcharacterization of a novel member of the transforming growth factor-beta/bone morphogenetic protein family. J. Biol. Chem., 273, 13760–13767.

44.Fairlie,W.D., Zhang,H.P., Wu,W.M., Pankhurst,S.L., Bauskin,A.R.,Russell,P.K., Brown,P.K. and Breit,S.N. (2001) The propeptide of thetransforming growth factor-beta superfamily member, macrophageinhibitory cytokine-1 (MIC-1), is a multifunctional domain that canfacilitate protein folding and secretion. J. Biol. Chem., 276, 16911–16918.

45.Arcuri,F., del Vecchio,M.T., de Santi,M.M., Lalinga,A.V., Pallini,V.,Bini,L., Bartolommei,S., Parigi,S. and Cintorino,M. (1999) Macrophagemigration inhibitory factor in the human prostate: identification andimmunocytochemical localization. Prostate, 39, 159–165.

46.Meyer-Siegler,K. and Hudson,P.B. (1996) Enhanced expression ofmacrophage migration inhibitory factor in prostatic adenocarcinomametastases. Urology, 48, 448–452.

47.del Vecchio,M.T., Tripodi,S.A., Arcuri,F., Pergola,L., Hako,L., Vatti,R. andCintorino,M. (2000) Macrophage migration inhibitory factor in prostaticadenocarcinoma: correlation with tumor grading and combinationendocrine treatment-related changes. Prostate, 45, 51–57.

974

48.Bernhagen,J., Calandra,T. and Bucala,R. (1998) Regulation of the immuneresponse by macrophage migration inhibitory factor: biological andstructural features. J. Mol. Med., 76, 151–161.

49.Donnelly,S.C. and Bucala,R. (1997) Macrophage migration inhibitoryfactor: a regulator of glucocorticoid activity with a critical role ininflammatory disease. Mol. Med. Today, 3, 502–507.

50.Achary,M.P., Jaggernauth,W., Gross,E., Alfieri,A., Klinger,H.P. andVikram,B. (2000) Cell lines from the same cervical carcinoma but withdifferent radiosensitivities exhibit different cDNA microarray patterns ofgene expression. Cytogenet. Cell Genet., 91, 39–43.

51.Schmidt,C.M., McKillop,I.H., Cahill,P.A. and Sitzmann,J.V. (1997)Alterations in guanine nucleotide regulatory protein expression and activityin human hepatocellular carcinoma. Hepatology, 26, 1189–1194.

52.Bohm,M. (1998) Catecholamine refractoriness and their mechanisms incardiocirculatory shock and chronic heart failure. Thorac. Cardiovasc.Surg., 46 (Suppl. 2), 270–275.

53.Derwahl,M., Hamacher,C., Russo,D., Broecker,M., Manole,D., Schatz,H.,Kopp,P. and Filetti,S. (1996) Constitutive activation of the Gs alphaprotein-adenylate cyclase pathway may not be sufficient to generate toxicthyroid adenomas. J. Clin. Endocrinol. Metab., 81, 1898–1904.

54.Selzer,E., Wilfing,A., Schiferer,A., Hermann,M., Grubeck-Loebenstein,B.and Freissmuth,M. (1993) Stimulation of human thyroid growth viathe inhibitory guanine nucleotide binding (G) protein Gi: constitutiveexpression of the G-protein alpha subunit Gi alpha-1 in autonomousadenoma. Proc. Natl Acad. Sci. USA, 90, 1609–1613.

55.Fotedar,R. and Fotedar,A. (1995) Cell cycle control of DNA replication.Prog. Cell Cycle Res., 1, 73–89.

56.Dobashi,Y., Shoji,M., Jiang,S.X., Kobayashi,M., Kawakubo,Y. andKameya,T. (1998) Active cyclin A-CDK2 complex, a possible criticalfactor for cell proliferation in human primary lung carcinomas. Am. J.Pathol., 153, 963–972.

57.Chiao,P.J., Shin,D.M., Sacks,P.G., Hong,W.K. and Tainsky,M.A. (1992)Elevated expression of the ribosomal protein S2 gene in human tumors.Mol. Carcinog., 5, 219–231.

58.Wong,J.M., Mafune,K., Yow,H., Rivers,E.N., Ravikumar,T.S., Steele,G.D.,Jr. and Chen,L.B. (1993) Ubiquitin-ribosomal protein S27a geneoverexpressed in human colorectal carcinoma is an early growth responsegene. Cancer Res., 53, 1916–1920.

59.Barnard,G.F., Staniunas,R.J., Bao,S., Mafune,K., Steele,G.D., Jr.,Gollan,J.L. and Chen,L.B. (1992) Increased expression of human ribosomalphosphoprotein P0 messenger RNA in hepatocellular carcinoma and coloncarcinoma. Cancer Res., 52, 3067–3072.

60.Batra,S.K., Metzgar,R.S. and Hollingsworth,M.A. (1991) Molecularcloning and sequence analysis of the human ribosomal protein S16. J. Biol.Chem., 266, 6830–6833.

61.Chester,K.A., Robson,L., Begent,R.H., et al. (1989) Identification of ahuman ribosomal protein mRNA with increased expression in colorectaltumours. Biochim. Biophys. Acta, 1009, 297–300.

62.Henry,J.L., Coggin,D.L. and King,C.R. (1993) High-level expression ofthe ribosomal protein L19 in human breast tumors that overexpress erbB-2. Cancer Res., 53, 1403–1408.

63.Schultz,P., Fribourg,S., Poterszman,A., Mallouh,V., Moras,D. and Egly,J.M.(2000) Molecular structure of human TFIIH. Cell, 102, 599–607.

64.Ko,L.J., Shieh,S.Y., Chen,X., Jayaraman,L., Tamai,K., Taya,Y., Prives,C.and Pan,Z.Q. (1997) p53 is phosphorylated by CDK7-cyclin H in ap36MAT1-dependent manner. Mol. Cell Biol., 17, 7220–7229.

65.Murphy,G.P., Elgamal,A.A., Su,S.L., Bostwick,D.G. and Holmes,E.H.(1998) Current evaluation of the tissue localization and diagnostic utilityof prostate specific membrane antigen. Cancer, 83, 2259–2269.

66.Sweat,S.D., Pacelli,A., Murphy,G.P. and Bostwick,D.G. (1998) Prostate-specific membrane antigen expression is greatest in prostateadenocarcinoma and lymph node metastases. Urology, 52, 637–640.

67.Su,S.L., Huang,I.P., Fair,W.R., Powell,C.T. and Heston,W.D. (1995)Alternatively spliced variants of prostate-specific membrane antigen RNA:ratio of expression as a potential measurement of progression. CancerRes., 55, 1441–1443.

68.Bostwick,D.G., Pacelli,A., Blute,M., Roche,P. and Murphy,G.P. (1998)Prostate specific membrane antigen expression in prostatic intraepithelialneoplasia and adenocarcinoma: a study of 184 cases. Cancer, 82, 2256–2261.

69.Kawakami,M. and Nakayama,J. (1997) Enhanced expression of prostate-specific membrane antigen gene in prostate cancer as revealed by in situhybridization. Cancer Res., 57, 2321–2324.

70.Rochon,Y.P., Horoszewicz,J.S., Boynton,A.L., Holmes,E.H., Barren,R.J.,III, Erickson,S.J., Kenny,G.M. and Murphy,G.P. (1994) Western blot assayfor prostate-specific membrane antigen in serum of prostate cancer patients.Prostate, 25, 219–223.

Differentially-regulated genes in prostate cancer

71.Beckett,M.L., Cazares,L.H., Vlahou,A., Schellhammer,P.F. andWright,G.L., Jr. (1999) Prostate-specific membrane antigen levels in serafrom healthy men and patients with benign prostate hyperplasia or prostatecancer. Clin. Cancer Res., 5, 4034–4040.

72.Renneberg,H., Friedetzky,A., Konrad,L., Kurek,R., Weingartner,K.,Wennemuth,G., Tunn,U.W. and Aumuller,G. (1999) Prostate specific

975

membrane antigen (PSM) is expressed in various human tissues:implication for the use of PSM reverse transcription polymerase chainreaction to detect hematogenous prostate cancer spread. Urol. Res., 27,23–27.

Received December 21, 2001; revised February 28, 2002;accepted March 11, 2002