Identification and Validation of Novel Androgen-Regulated Genes in ...

13 downloads 117 Views 422KB Size Report
dihydrotestosterone-regulated genes. Representative clus- ters containing genes with similar expression patterns to prostate-specific antigen and other known ...
0013-7227/04/$15.00/0 Printed in U.S.A.

Endocrinology 145(8):3913–3924 Copyright © 2004 by The Endocrine Society doi: 10.1210/en.2004-0311

Identification and Validation of Novel AndrogenRegulated Genes in Prostate Cancer ANNE MARIE VELASCO, KIMBERLY A. GILLIS, YIZHENG LI, EUGENE L. BROWN, TAMMY M. SADLER, MARIA ACHILLEOS, LEE M. GREENBERGER, PHILIP FROST, WENLONG BAI, AND YIXIAN ZHANG Department of Genomics (A.M.V., K.A.G., Y.L., E.L.B.), Wyeth Research, Cambridge, Massachusetts 02140; Department of Oncology (T.M.S., M.A., L.M.G., P.F., Y.Z.), Wyeth Research, Pearl River, New York 10965; and Department of Pathology (W.B.), University of South Florida College of Medicine, Tampa, Florida 33612 Androgen-regulated genes (ARGs) are essential for the development of the prostate. Ironically, ARGs are also responsible for the pathogenesis of prostate cancer. We used oligonucleotide array technology to study the expression profiles of ARGs in LNCaP prostate cancer cells and identified 692 dihydrotestosterone-regulated genes. Representative clusters containing genes with similar expression patterns to prostate-specific antigen and other known ARGs are discussed. Based on functional information, we categorized several candidate targets for prostate cancer therapy and diagnosis. Although many of these candidate targets are known to play an important role in cancer development, several are novel genes to the field of prostate cancer. A cross-comparison study of our results with those that have been previously published from three other array experiments using a similar LNCaP model validated 13 of these candidate targets as an-

P

ROSTATE CANCER IS the second leading cause of male cancer deaths in the United States (1, 2). Normal prostate gland growth and prostatic carcinomas are regulated by androgen via its cognate receptor, the androgen receptor (AR) (3–5). Androgen ablation has been a promising approach for the treatment of advanced prostate cancer. However, the effectiveness of this therapy is hindered by the complexity of the disease and an inadequate understanding of the basic mechanisms underlying the initiation, promotion, and progression of prostate cancer. Frequently, the tumor reappears after therapy and becomes androgen independent (6, 7). The mechanism of androgen-independent cancer progression is poorly understood but may include the following events: mutation of AR, ligand-independent activation of AR, or altered expression of AR and its target genes (8). Studies have revealed that AR protein is present in most Abbreviations: AR, Androgen receptor; ARG, androgen-regulated gene; BPH, benign prostatic hyperplasia; CDK, cyclin-dependent kinase; DHT, dihydrotestosterone; FCS, fetal calf serum; FKBP51, FK506-binding immunophilin 51; GAPDH, glyceraldehyde-3-phosphate dehydrogenase; KLK2, Kallikrein-2; PacP, prostatic acid phosphatase; Pca, prostate cancer; PSA, prostate-specific antigen; PVDF, polyvinyl difluoride; Ref Seq, reference sequence; SAGE, serial analysis of gene expression; Seladin-1, selective Alzheimer’s disease indicator 1; SNRK, sucrose nonfermenting protein-1-related kinase; SOM, self-organizing map; TBST, TBS with 0.1%Tween 20. Endocrinology is published monthly by The Endocrine Society (http:// www.endo-society.org), the foremost professional society serving the endocrine community.

drogen-regulated. FKBP51 (FK506-binding immunophilin 51) was found in the same cluster as prostate-specific antigen and its protein expression was increased in LNCaP cells treated with either dihydrotestosterone or synthetic androgen R1881. Results from mining the Gene Logic BioExpress database showed that FKBP51 expression is significantly higher in the prostate cancer group than in the normal and normal adjacent group. Additionally, the androgen-independent prostate tumor xenograft, CWR22R, had higher FKBP51 protein levels than that of the androgen-dependent prostate tumor xenograft, CWR22. A tissue microarray study further revealed that FKBP51 protein expression was higher in prostate cancer specimens than in benign prostate tumor samples. These results suggest the potential value of FKBP51 as a novel diagnostic marker or target for prostate cancer therapy. (Endocrinology 145: 3913–3924, 2004)

androgen-independent tumors, raising the possibility that androgen-regulated genes (ARGs) may play a role in the acquisition of hormone-independent growth in prostate cancer cells (9, 10). Although very little is known about the downstream events in the AR pathway, the discovery of ARGs, such as prostate-specific antigen (PSA) and prostatespecific membrane antigen, has already shown potential in the early diagnosis and treatment of prostate cancer patients (11–14). The identification of additional ARGs will provide a better understanding of prostate cancer development. Until recently most studies have been limited to monitoring the expression of a few genes at a time. The development of microarray technology has allowed us to rapidly analyze gene expression patterns in tissues or cells on a genomic scale (15, 16). Microarrays have also been used to study ARGs and genes involved in the development of prostate cancer (17– 25). Because of their importance in normal prostate development and prostate cancer progression, we focused our initial study on ARGs in the well-established androgen-dependent prostate cancer cells, LNCaP (26). Similar studies using microarrays for identifying androgen-regulated genes in LNCaP have been previously published; however, subtle differences in experimental design can have a major impact on resulting data sets. In this study, we show how data sets from two serial analysis of gene expression (SAGE) and four separate microarray studies, including ours, have relatively little overlap among them and discuss possible explanations for these differences. By using oligonucleotide array technology in combination

3913

3914

Endocrinology, August 2004, 145(8):3913–3924

with two-way ANOVA and cluster analysis, we identified multiple genes that are differentially regulated by androgen in LNCaP cells, most of which have not yet been reported before as androgen-regulated genes. Of great interest is a novel kinase sucrose nonfermenting protein-1-related kinase (SNRK) (27), a novel transglutaminase, selective Alzheimer’s disease indicator 1 (Seladin-1) (27), and FKBP51 (FK506binding immunophilin 51), a member of the immunophilin family (28). In this study, androgen regulation of SNRK, Seladin-1, and FKBP51 in LNCaP cells was verified through quantitative RT-PCR. Additionally, we established up-regulation of FKBP51 protein by Western analysis. We performed a search and statistical analysis on FKBP51 mRNA expression among various human prostate tissue subtypes that were profiled by the Affymetrix GeneChip technology (Santa Clara, CA) and stored in the commercial Gene Logic BioExpress database system (Gaithersburg, MD). The results showed significant differential expression of FKBP51 among the four prostate tissue subtypes. Amler et al. (18) and Mousses et al. (19) recently showed that FKBP51 transcription was repressed in response to castration in a CWR22 prostate cancer xenograft model but enhanced in recurrent tumors. Through Western analysis of prostate xenograft tumor models, we confirmed that FKBP51 protein expression was increased as tumors evolved from the androgen-dependent stage to the androgen-independent stage. Using tissue microarrays, we further investigated the expression of FKBP51 in multiple human benign prostate hyperplasia (BPH) and diseased specimens, and found that FKBP51 was expressed significantly higher in prostate tumor samples relative to BPH samples. These studies suggest that FKBP51 may be used as a potential marker or therapeutic target for prostate cancer. Materials and Methods Cell cultures Human prostatic cancer cell lines LNCaP, DU-145, and PC-3 were obtained from American Type Culture Collection (Manassas, VA). LNCaP cancer cells were maintained in a humidified atmosphere of 5% CO2 and in RPMI 1640 medium supplemented with 10% fetal calf serum (FCS) (Life Technologies, Inc., Rockville, MD), 3 mm l-glutamine, 100 ␮g/ml streptomycin, and 100 U/ml penicillin. Other lines were maintained in DMEM containing 3 mm l-glutamine, 100 ␮g/ml streptomycin, 100 U/ml penicillin, 10% FCS in a humidified atmosphere of 5% CO2. To examine the effects of steroids, cells were cultured in RPMI 1640 medium containing 5% FCS treated with dextran-coated charcoal (Hyclone, Logan, UT) for 24 h before treatment. Cells were grown in the absence or presence of 10 nm dihydrotestosterone (DHT) for 0, 2, 4, 6, 12, 24, 48, and 72 h. They were collected and frozen at each time point. Two hundred microliters of medium were collected from each flask for the PSA assay.

RNA extraction and preparation Total RNA was isolated from LNCaP cells using the RNeasy midikit (Qiagen, Santa Clarita, CA) following the manufacturer’s recommendations. For polyA (⫹) selection, the PolyATract kit (Promega, Madison, WI) was used according to the manufacturer’s procedures. One microgram of poly A(⫹) RNA was used as template for synthesis of doublestranded cDNA using the cDNA synthesis kit (GibcoBRL, Gaithersburg, MD), with an oligo dT primer incorporating a T7 RNA polymerase promoter [10 min at 70 C for priming, 65 min at 37 C for first-strand synthesis with Superscript II reverse transcription, followed by 150 min at 15.8 C for second-strand synthesis with Escherichia coli ligase, E. coli

Velasco et al. • Androgen-Regulated Genes in LNCaP Prostate Cancer Cells

polymerase, and RNase H]. The double-stranded cDNA was purified by solid-phase reversible immobilization (29) using Perseptives paramagnetic beads. Approximately 50 ng of double-stranded cDNA was used as template for in vitro transcription to make labeled cRNA (16 h at 37 C, Epicenter T7 RNA polymerase, Enzo Laboratories (Farmingdale, NY) bio-11-CTP, bio-11-uridine 5-triphosphate). The cRNA was purified by solid-phase reversible immobilization using paramagnetic beads (Bangs Laboratories, Fishers, IN), and total molar concentration was determined from the absorbance at 260. Before hybridization, 10 ␮g of labeled cRNA was fragmented randomly to an average length of approximately 50 bases by heating at 94 C in 40 mm Tris-acetate (pH 8.1), 100 mm potassium acetate, and 30 mm magnesium acetate for 35 min.

Chip hybridization and data reduction Hybridization cocktail was made using 10 ␮g of fragmented cRNA, 2 ⫻ 2-(N-morpholine) ethane sulfonic acid buffer with BSA, herring sperm DNA, control prokaryotic transcripts for internal control, and biotinylated control oligo 948 (for chip quality control). Diethylpyrocarbonate-water was added to bring the volume to 200 ␮l. Two sets of cocktail from each cRNA sample were prepared for this experiment. Before hybridization, the hybridization cocktails were heated to 99 C for 10 min and then 37 C for an additional 10 min before loading into Hu6800FL arrays (Affymetrix GeneChips). The Hu6800FL array is comprised of 6800 known full-length genes, about 250,000 25-mer oligonucleotide probes with 20 probe pairs per gene. Array hybridization proceeded overnight at 45 C with 50 rpm. After hybridization, hybridization cocktails were recovered from the arrays and replaced with 6⫻ SSPET [0.9 m NaCl, 60 mm NaH2PO4, 6 mm EDTA, 0.01% Triton X-100 (pH 7.6)]. The arrays were washed and stained using the manufacturer’s recommendations and procedures. Nonstringent wash buffer–20⫻ SSPE [3 m NaCl, 0.2 m NaH2PO4, 20 mm EDTA (pH 7.0)]; 1.0 ml of 10% Tween 20, and water–at 25 C and stringent wash buffer (20⫻ SSPE, 5 m NaCl, 10% Tween 20, and water) at 50 C were used for the wash steps. The arrays were then stained with streptavidin-conjugated phycoerythrin (Molecular Probes, Eugene, OR), followed by biotinylated antistreptavidin and a second round of streptavidin-conjugated phycoerythrin for signal amplification at 25 C. Each stain step was done for 10 min. All arrays were then scanned using the Genearray scanner (Hewlett Packard, Portland, OR), and the resulting fluorescence emissions were collected and quantified using Affymetrix GeneChip software. Within the software, the signal intensities for all the probes on each array were calculated from the scanned image, and the appropriate probe array algorithm was applied to generate a qualitative call (absent, marginal, or present) and a quantitative measurement (average difference) of expression level for each gene. Average difference values were converted to transcript abundance estimates, in units of parts per million, by reference to a standard curve of 11 spiked in vitro transcripts as described elsewhere (30).

Data filtering and statistics Initial data were reduced by filtering for all genes called present by GeneChip. A two-way ANOVA was then performed on the replicate data for each of these genes in the statistical computing package S-plus. The potential effects of two experimental factors (treatment and time) and the interaction of both factors on the expression level were evaluated in the ANOVA model, and the P values (probability values that the observations made are due to chance) for the main effects (treatment, time) and the interaction were obtained. Only those genes that were statistically significant, defined by having a P ⱕ 0.05, due to the treatment effect alone and/or the interaction of two factors were considered for interpretation in the present context and were subsequently classified by clustering analysis. First, the average was taken for baseline and experimental replicate mRNA frequencies of the 692 genes that passed this P value criterion. Average frequencies obtained for each gene were then standardized across all samples to have a mean of 0 and sd of 1. A modified version of the original self-organizing map (SOM) algorithm developed by Kohonen et al. (31), created using the MATLAB toolbox (Mathworks, Natick, MA), was then applied to the standardized expression values to generate a 6 ⫻ 6 matrix of 36 clusters (32). Several public databases such as SOURCE (33) and Swiss-Prot (34) were used for gene annotation.

Velasco et al. • Androgen-Regulated Genes in LNCaP Prostate Cancer Cells

Quantitative Taqman RT-PCR One set of total RNA samples that were used for the GeneChip experiments were analyzed using a Taqman EZ RT-PCR kit (PE Applied Biosystems, Foster City, CA). Total RNA samples were diluted to a concentration of 50 ng/␮l, and a total of 50 ng was used for each reaction. Primers and fluorescence probes for PSA, FKBP51, Seladin-1, and SNRK were designed using the Primer Express software (Foster City, CA) and were chosen based on the manufacturer’s recommendations for primer selection. The primers used were of 100 ␮m concentration and were as follows: 1) PSA-F, CGTGGCCAACCCCTGA (forward primer), PSA-R CTTGGCCTGGTCATTTCCAA (reverse primer), PSA-P CACCCCTATCAACCCCCTATTGTAGTAAACTTGGA (probe); 2) FKBP51-F, CTGTGACAAGGCCCTTGGA (forward primer), FKBP51-R, CTGGGCTTCACCCCTCCTA (reverse primer), FKBP51-P, ACAAGCCTTTCTCATTGGCACTGTCCA (probe); 3) Seladin-1-F, CAAGATCCTTCCTTCAACCCC (forward primer), Seladin-1-R, TGGCACCTGGAATGACAAGA (reverse primer), Seladin-1-P, AGCTCCCATCTCATTTCCAGAAAGGCTCAT (probe); 4) SNRK-F, GTCATGTGTCTGAGGTGACGGA (forward primer), SNRK-R, TGAAGAAACAGTGACCACAGCAAT (reverse primer), SNRK-P, TGGTCCTGTAATTCAGAGAGTGGGCACATCACC (probe). Samples were prepared using a reagent mix of manufacturer-supplied RT-PCR components [(5 ⫻ Taqman EZ buffer, manganese acetate (25 mm), dATP (10 mm), dCTP (10 mm), dGTP (10 mm), and dexoyuridine 5-triphosphate (20 mm), rTth DNA polymerase (2.5 U/␮l), AmpErase UNG (1 U/␮l), primers (final concentration 1 ␮m), and RNA (50 ng)], following manufacturer’s recommendations. In addition, glyceraldehyde-3-phosphate dehydrogenase (GAPDH) control samples for standard curve generation and subsequent quantitation of sample RNA was prepared. Primers and probe for GAPDH were included in the kit (GAPDH forward and reverse primers 10 ␮m, GAPDH probe 5 ␮m). Dilutions were made for GAPDH that ranged from 5 ⫻ 101 copies to 5 ⫻ 106 copies. The assay was performed on a Perkin-Elmer/ Applied Biosystems 7700 Prism, and the PCR cycling parameters were chosen based on the manufacturer’s recommendations. RNA samples were quantified and normalized to GAPDH.

Western blot analysis LNCaP cells were harvested in MPER reagent (Pierce, Rockford, IL) containing 400 mm NaCl. CWR22 and CWR22R tumor samples were lysed by homogenization in buffer containing 20 mm Tris (pH 7.5) 250 mm NaCl, 3 mm EDTA, 3 mm EGTA, 100 ␮m Na3VO4, 1 mm dithiothreitol, 0.5% Nonidet P-40). Protein from all samples was quantified by the Bradford method (35). Tissue lysates containing 100 ␮g protein or cell lysates containing 30 ␮g protein were electrophoresed on a 12% SDSPAGE gel and transferred to a polyvinyl difluoride (PVDF) membrane using a liquid transfer apparatus (Bio-Rad Laboratories, Hercules, CA). The PVDF membrane was incubated in TBST (TBS with 0.1%Tween 20) with 3% milk for 15 min before the addition of the first antibody, ␣-FKBP51 (a gift kindly given by Dr. David Smith, Department of Biochemistry and Molecular Biology, Mayo Clinic, Scottsdale, AZ). After overnight incubation, the PVDF membrane was washed three times with TBST and incubated with secondary antibodies coupled with horseradish peroxidase (Transduction Laboratories, Lexington, KY) for 1 h. The PVDF membrane was then washed three times with TBST, and protein was detected using an enhanced chemiluminescence detection system (Pierce).

Endocrinology, August 2004, 145(8):3913–3924 3915

expression among subgroups, we used the Tukey method of multiple comparisons to calculate simultaneous 95% confidence intervals for all pairwise differences among the four prostate tissue subtypes.

Tissue microarray construction and analysis as performed by Clinomics Biosciences (Pittsfield, MA) After fixation in 10% neutral buffered formalin, tissues were selected, trimmed, and placed in a processing cassette. The cassette was then placed in a processing basket on a Shandon Hypercenter (Atlantic, IA) tissue processor in which the tissues were exposed to a series of buffers over a 16-h processing cycle (10% neutral buffered formalin, 70, 95, 100% ethanol, xylene, and melted paraffin embedding media). All steps were carried out under vacuum at 40 C except for the paraffin step, which was at 58 C. After processing, the tissues were removed from the cassettes and embedded in paraffin blocks. The resulting blocks were sectioned at 5 ␮m and mounted on glass slides. The slides were heated at 58 C for 30 min before staining. Antibody ␣-FKBP51 was titered to a 1:150 dilution using antibody diluent (Dako, Carpinteria, CA). Staining of test specimen was performed employing HIER (heat-induced epitope retrieval) in citrate buffer (pH 6.0) with no pretreatment. Tissues were then stained using the Ventana (Tucson, AZ) ES automated immunohistochemistry stainer. Grading of the immunohistochemical staining was based on the intensity of the cytoplasmic staining of the epithelial components of both the tumor and the normal tissues. The strength of the staining was scored using a 1⫹ to 4⫹ scale, 1⫹ indicating faint staining and 4⫹ indicating strongest staining. A score of 0 indicated no staining.

Data query and comparison Data from published articles (24, 36, 37) were downloaded and the data sets of all differentially expressed genes determined. Reference sequence (Ref Seq) accession numbers were used for gene identification. If Ref Seq accession numbers were not assigned to the genes by the original authors, the accession numbers were used to pull sequences for subsequent blasting against the human Ref Seq database in GenBank. If no Ref Req accession number was found, comparisons were done by gene description.

Results GeneChip hybridization and analysis

We used Affymetrix GeneChip technology to monitor the expression of about 6000 full-length human genes in response to a natural androgen DHT in LNCaP cells. Figure 1 illustrates the general scheme used for sample preparation, hybridization, and analysis. Total RNA was prepared in duplicate from LNCaP cells treated with or without DHT for 0, 2, 4, 6, 12, 24, 48, and 72 h. cRNAs were prepared and hybridized also in duplicate to Affymetrix chips. Only those genes that were called present in either the baseline or the experiment in at least one time point and in either replicate passed our initial data reduction filter. Of about 6000 genes represented on the chip, 4491 passed this initial filter (75%).

Data mining of Gene Logic BioExpress database

Statistical analysis of replicates

We performed a database search and statistical analysis on FKBP51 mRNA expression among various human prostate tissue subtypes that were profiled by the Affymetrix GeneChip technology and stored in the commercial Gene Logic BioExpress database system (http://www. genelogic.com/solutions/bioexpress/). Based on sample annotation and quality assessment, we identified a total of 117 prostate tissue samples of four subtypes for the current study. These include normal prostate tissues from nondiseased individuals (trauma death, n ⫽ 6), normal adjacent (n ⫽ 34), BPH (n ⫽ 20), and prostate cancer (n ⫽ 57). The expression values (average difference) were normalized by the Affymetrix global scaling method. The data were then log transformed (base 2), and a one-way ANOVA was performed. To identify differential

Using the two-way ANOVA model, we determined the statistical significance of 4491 gene expression changes. Based on a 0.05 significance level cut-off, our results show that 934 of 4491 gene expression changes were significant. One hundred ninety-five genes were significant due to androgen treatment alone independent of time, whereas 497 genes were significant due to an interaction between androgen treatment and time (i.e. varied treatment effect over time). Because we were interested only in identifying androgen-regulated genes, the remaining 242 genes that were sig-

3916

Endocrinology, August 2004, 145(8):3913–3924

nificantly modulated due to the time effect alone were not considered. Rapid classification of expression profiles using SOMs

For rapid classification and to understand the potential function of candidate genes, expression profiles of the 692 genes found to be regulated by androgen (alone or differentially modulated over time) in ANOVA were clustered using an adaptation of the SOM algorithm. Shown in Fig. 2 are selected genes from two clusters based on patterns of induced (Fig. 2A) or repressed expression (Fig. 2B).

Velasco et al. • Androgen-Regulated Genes in LNCaP Prostate Cancer Cells

Genes that are induced in response to androgen

Table 1 shows several genes found in Cluster (1,1) (Fig. 2A), including several prominent androgen-regulated genes such as PSA and Kallikrein-2 (KLK2), Seladin-1, and a 54-kDa immunophilin, FKBP51. Genes were annotated for description, function, and categories using several public databases such as SOURCE and Swiss-Prot. The fact that several members of the serine protease family, including PSA and KLK2, are found in this cluster suggests that genes that resemble each other in expression patterns are likely to participate in similar physiological programs. This method also allows insight into possible novel functions of newly identified androgen-regulated genes such as FKBP51 and Seladin-1. A complete list of genes in this cluster is available on The Endocrine Society’s Journals online web site, http://endo. endojournals.org/. Genes that are repressed in response to androgen

FIG. 1. RNA sample preparation, Affymetrix GeneChip hybridizations, and analysis. Total RNA was extracted from cell lysates. Amplified Poly (A⫹) selected RNA was hybridized in duplicate to the Hu6800 array, which monitors approximately 6000 full-length genes. After GeneChip analysis, the resulting data for quality control genes were assessed for chip quality and sample quality. After passing QC (quality control), the data were normalized using our own proprietary analysis subsystem, and average differences for all genes were converted into mRNA frequency estimates (in molecules per million) based on the standard spike-in control transcripts. The data reduction was carried out by querying only those genes called present, in either baseline or experiment, in at least one time point in either replicate. A total of 692 known genes were identified as being significantly differentially regulated by androgen treatment at a 95% confidence level. The genes were subsequently clustered and annotated based on functional classifications and information from various public databases. IVT, In vitro transcription.

FIG. 2. Two examples of SOM cluster analysis of 692 ARGs. Baseline values are shown in red, and experimental values (DHT treatment) are shown in blue. A, Cluster (1,1) illustrates a few genes that shared a similar profile of induced expression upon androgen treatment. B, Cluster (6,6) includes a few genes that had a pattern of repressed expression upon androgen treatment.

In Table 2, we show a selection of genes that fall within Cluster (6,6) (Fig. 2B), which share a pattern of repressed expression relative to baseline. One of the genes is prostatic acid phosphatase (PAcP), which like PSA, is another prostate-specific protein. PAcP has previously been shown to be suppressed by DHT in LNCaP (38). TRPM2, or Clusterin, also found in cluster (6,6), is another gene previously hypothesized to be androgen regulated. Mattmueller and Hinton (39) found that the caudal fluid of orchiectomized rats contain increased amounts of Clusterin, compared with control animals. Several kinases such as PCTAIRE-1 and cyclindependent kinase (CDK)-8, which in general are known to play key roles in cell cycle regulation and cell differentiation and have often been targeted for cancer therapy, were also found in this cluster. In addition, a novel serine/threonine kinase, SNRK, was also found. A complete list of genes in this cluster is available on The Endocrine Society’s Journals online web site, http://endo.endojournals.org/. Identification of ARGs and candidate targets

Based on the ranking of significant P values and cluster classification, candidate target genes were identified and categorized by area of therapeutic potential. In Table 3, we show a subset of these genes and their respective target category. Shown in red are genes that are proposed as

Velasco et al. • Androgen-Regulated Genes in LNCaP Prostate Cancer Cells

Endocrinology, August 2004, 145(8):3913–3924 3917

TABLE 1. A sample of genes within cluster (1, 1) Name

Accession no.

DRG1 (RTP)

D87953

Description

Categories

KLK2

S39329

FKBP51

U42031

SGTPBP

U57094

54-kDa progesterone receptorassociated immunophillin Small GTP-binding protein

PSA

X07730

Prostate-specific antigen

Seladin-1

D13643

Human Diminuto/Dwarf1 homolog

Novel marker of androgen-induced differentiation Kallikrein-2 precursor

Differentiation-related gene 1 protein; suppressor Hydrolases; serine protease; kininogenase; glycoprotein; multigene family; zymogen; signal Isomerase; rotamase; repeat; nuclear protein GTP-binding; lipoprotein; prenylation; polymorphism Hydrolase; serine protease; glycoprotein; antigen; zymogen; signal; 3D-structure Oxidoreductase; cell elongation; steroid synthesis

The genes listed in this table are representative of many up-regulated genes in response to DHT treatment. A complete list of genes in this cluster is available as supplemental data on The Endocrine Society’s Journals Online web site, http://endo.endojournals.org/. TABLE 2. A sample of genes within cluster (6, 6) Name

Accession no.

Description

Categories

TRMP2 SNRK CDK8

M63379 D43636 X85753

Clusterin SNF-1-related kinase Cyclin-dependent kinase 8

SBP PAcP

U29091 M24902

Selenium-binding protein Prostatic acid phosphatase

PCTAIRE1

X66363

Serine/threonine protein kinase

Glycoprotein, apoptosis Novel serine/threonine kinase Serine/threonine kinase; ATP binding; cell cycle; cell division; transcription regulation; transferase Selenium Hydrolase; glycoprotein; signal; riboflavin metabolism; protein tyrosine phosphatase ATP binding; transferase

The genes listed in this table are representative of many up-regulated genes in response to DHT treatment. A complete list of genes in this cluster is available as supplemental data on The Endocrine Society’s Journals Online web site, http://endo.endojournals.org/.

ARGs. The confirmation of 10 of them (in red) as ARGs in our analysis has validated our approach. We also identified many genes that either were never previously found to be androgen regulated in LNCaP or are just beginning to be investigated as potential prostate cancer targets (a complete list of 692 ARGs is available on The Endocrine Society’s Journals online web site, http://endo.endojournals. org/). Based on both significant P values and an induction pattern similar to PSA, we selected two up-regulated genes, FKBP51 and Seladin-1, as potential candidate targets (both genes had P ⬍ 0.01). In response to androgen treatment, PSA expression increased 3-fold relative to control at 12 h and maintained its high expression through 72 h at which point it was induced approximately 4-fold (Fig. 3A). Similarly, FKBP51 expression in the control samples maintained a consistent pattern of relatively low expression throughout the time course. However, with androgen treatment, FKBP51 was rapidly induced 2-fold at 6 h and peaked at 24 h, at which point it was overexpressed approximately 4-fold relative to baseline (Fig. 3B). Seladin-1 also shares this pattern of induced expression in response to androgen. Its induction began at 12 h, and at 24 h it was overexpressed 2-fold relative to control (Fig. 3C). Like PSA, Seladin-1 maintained its overexpression in response to androgen through 72 h, at which point it was still induced 3-fold relative to baseline. We were also interested in identifying genes that are down-regulated in response to androgen in LNCaP. Based on protein and nucleotide homology analysis, we found SNRK to be a novel serine/threonine kinase. Its expression, although consistently repressed relative to control throughout the time course, was most significantly altered

at 24 h at which point it was down 5-fold relative to baseline (Fig. 3D). Quantitative RT-PCR analysis of RNA samples

Using Quantitative RT-PCR, we were able to confirm the gene expression changes from the GeneChip analysis for four genes: PSA, FKBP51, Seladin-1, and SNRK. Expression of PSA, FKBP51, and Seladin-1 increased, whereas SNRK expression decreased in response to DHT treatment (Fig. 4). Production of FKBP51 was regulated by DHT and a synthetic androgen R1881

To investigate whether the observed effects of DHT on FKBP51 mRNA were accompanied by changes in its protein level, we performed Western analysis on the samples obtained from the time-course experiment. DHT up-regulated FKBP51 expression in a time-dependent manner (Fig. 5A). Similarly, a synthetic androgen, R1881, could also up-regulate FKBP51 expression (Fig. 5A). Interestingly, the protein level increased after 24 h, which was 12 h later than the transcript, suggesting that protein synthesis was required for the induction. To validate that protein synthesis is involved in DHT stimulation of FKBP51, we repeated the 48-h time point with or without protein synthesis inhibitor cycloheximide (Fig. 5B). In the presence of cycloheximide, DHT failed to increase the protein level of FKBP51 (Fig. 5B). Data mining and statistical analysis of FKBP51 mRNA levels in prostate tissue subtypes

To validate and supplement in vitro observations on FKBP51, we sought in vivo evidence of differential expression

3918

Endocrinology, August 2004, 145(8):3913–3924

Velasco et al. • Androgen-Regulated Genes in LNCaP Prostate Cancer Cells

TABLE 3. Candidate target genes Target genes

Angiogenesis ID1

Affymetrix probe set

Unigene ID

GenBank description

HG3342HT3519_s_at X69111_at

Hs.75424

U61276_s_at M57730_at

Hs.91143 Hs.1624

Inhibitor of DNA binding 1; dominant negative helix-loophelix protein Inhibitor of DNA binding 3, dominant negative helix-loophelix protein Jagged 1 (Alagille syndrome) Ephrin-A1

X85753_at

Hs.25283

Cyclin-dependent kinase 8

U42031_at

Hs.7557

FK506-binding protein 5

X65873_at X01703_at

Hs.149436 Hs.272897

Kinesin family member 5B Tubulin, ␣, brain-specific

X07730_at S39329_at M24486_s_at

Hs.171995 Hs.181350 Hs.76768

Kallikrein 3, (prostate specific antigen) Kallikrein 2, prostatic Procollagen-proline, 2-oxoglutarate 4-dioxygenase (proline 4-hydroxylase), ␣-prolypeptide 1

M76180_at

Hs.150403

M68840_at U82256_at U08854_s_at

Hs.183109 Hs.172851 Hs.150207

Dopa decarboxylase (aromatic L-amino acid decarboxylase) Monoamine oxidase A Arginase, type II UDP glycosyltransferase 2 family, polypeptide B15

D13643_at D43636_at

Hs.75616 Hs.79025

KIAA0018 KIAA0096 protein

D87953_at

Hs.75789

N-myc downstream regulated

S76978_s_at

Hs.1915

Folate hydrolase (prostate-specific membrane antigen) 1

U65093_at

Hs.82071

Z19002_at

Hs.37096

CBP/p300-interacting transactivator, with Glu/Asp-rich carboxy-terminal domain, 2 Zinc finger protein 145 (Kruppel-like, expressed in promyelocytic leukemia)

Gene regulation SMARCD3

U66619_at

Hs.71622

SWI/SNF related, matrix associated, actin dependent regulator of chromatin, subfamily d, member 3

Promoter and survival TRPM-2 (CLU)

M63379_at

Hs.75106

L00058_at U44103_at

Hs.79070 Hs.28726

Clusterin (complement lysis inhibitor, SP-40, 40, sulfated glycoprotein 2, testosterone-repressed prostate message 2, apolipoprotein j) V-myc avian myelocytomatosis viral oncogene homolog RAB9, member RAS oncogene family

ID3 JAG1 EFNA1 Cell cycle CDK8 Cell signaling EKBP51 Cell structure Kinesin ␣-Tubulin Invasion and metastasis PSA(KLK3) KLK2 P4HA1 Metabolism DDC MAOA Arginase UGT2B15 Novel enzyme Seladin-1 KIAA0096 Novel suppressor NDRG1(RTP) Surface antigen PSMA(FOH1) Transcription CITED2 ZNF145

c-Myc RAB9

Hs.76884

From the 692 genes found to be significant by ANOVA and cluster classification, candidate gargets were identified and categorized by area of therapeutic potential. Genes shown in bold are confirmed ARGs. A complete list of the 692 genes is available as supplemental data on The Endocrine Society’s Journals Online web site, http://endo.endojournals.org/.

by querying the Gene Logic BioExpress database, which stores mRNA expression data in Affymetrix GeneChip format for a large number of human samples. The ANOVA performed on the queried samples (n ⫽ 117) indicated significant differential expression of FKBP51 among the four prostate tissue subtypes (P ⫽ 0.001). Specifically, the prostate cancer (Pca) group mean expression was significantly higher than normal and normal adjacent group means as indicated by the boxplots (Fig. 6A) and the corresponding 95% confidence intervals located to the right of the vertical reference line at zero (Fig. 6B). Increased FKBP51 in androgen-independent CWR22 vs. androgen-dependent CW22

To determine whether FKBP51 protein is expressed at a higher level in hormone refractory prostate cancers than in

androgen sensitive ones, we analyzed the expression of FKBP51 protein in CWR22 human prostate cancer xenograft and various CWR22R tumors that have become hormone independent by repeated propagation in mice (40). As shown in Fig. 7, FKBP51 protein was expressed at a higher level in all six CWR22Rs relative to the parental CWR22 tumor. Immunohistochemistry staining of prostatic adenocarcinoma with anti-FKBP51

We sought to determine the level of FKBP51 protein in normal and cancerous human prostate tissue using tissue microarrays (Clinomics Biosciences). Because there were only a few normal prostate samples available to obtain a significant result, we analyzed immunohistochemistry-staining data from a chip containing 92 BPH and 95 prostate

Velasco et al. • Androgen-Regulated Genes in LNCaP Prostate Cancer Cells

Endocrinology, August 2004, 145(8):3913–3924 3919

FIG. 3. Expression profiles of PSA (A), FKBP51 (B), Seladin-1 (C), and SNRK (D) in response to androgen treatment. mRNA frequencies (in molecules per million) are plotted on the y-axis for each gene. Baseline values (white) and values for the gene after androgen treatment (gray) are shown for each time point. Range bars reflect actual frequency differences between the two replicates at each time point. FKBP51 and Seladin-1, an immunophilin and novel transglutaminase, respectively, show similar induction patterns to PSA and are being further investigated. SNRK expression is repressed in response to androgen treatment.

FIG. 4. Quantitative RT-PCR analysis of samples. Using quantitative RT-PCR, we were able to confirm the gene expression changes from the GeneChip analysis for four targets: PSA (A), FKBP51 (B), Seladin-1 (C), and SNRK (D). Copy number is plotted on the y-axis for each gene. Baseline values (white) and values for the gene after androgen treatment (gray) are shown for each time point.

3920

Endocrinology, August 2004, 145(8):3913–3924

FIG. 5. Androgen induction of FKBP51. LNCaP cells were plated in a 6-well plate at 1 ⫻ 106 cells/well in charcoal-stripped serum containing medium. A, Cells were treated with 10 nM DHT and harvested at 0, 2, 4, 12, 24, 48, and 72 h post treatment. Protein levels of FKBP51 were detected by Western analysis. B, Cells were treated with or without 10 nM DHT in the presence or absence of cycloheximide (CHX) as marked for 48 h before Western analysis.

Velasco et al. • Androgen-Regulated Genes in LNCaP Prostate Cancer Cells

FIG. 7. Differential FKBP51 expression in androgen-sensitive and hormone-refractory prostate cancers. Tissue lysates of CWR22 and CWR22R tumors containing 100 ␮g protein were separated on a 10% SDS-PAGE and the protein levels of FKBP51 (top panel) or ␤-actin (bottom panel) were detected on immunoblots as described in Materials and Methods. The density of the bands on immunoblots was scanned. The signals were quantified using Scion Image ␤ 4.02 software. The numbers represent adjusted ratio of FKBP51 over ␤-actin (ratio of LNCaP without treatment and CWR22 are set to 1).

FIG. 8. Immunohistochemical staining of a tissue microarray containing samples of BPH and prostatic adenocarcinomas with antiFKBP51 antibody. Mean staining scores from multiple BPH (n ⫽ 92) and prostate adenocarcinoma samples (n ⫽ 95) as marked; bars, ⫾SE. *, P ⫽ 3 ⫻ 10⫺7, Student’s t test.

Discussion FIG. 6. Analysis of differential expression of FKBP51 mRNA in four prostate tissue subtypes (total n ⫽ 117) from the Gene Logic BioExpress database. A, Box plot displays of FKBP51 mRNA expression in each group. The y-axis represents global-scaled average difference. The middle horizontal line within each box indicates the median value of each group. The cross symbol within each box indicates the mean value of each group. Panel B shows fold-change (FC) estimates for all pairwise comparisons by the Tukey method and their simultaneous 95% confidence intervals (dashed lines with brackets at both ends). Both linear and log scale of FC are indicated (top and bottom, respectively) in the graph. A comparison of two groups is statistically significant if the confidence interval of the mean difference exclude zero on log scale (visually, the confidence interval does not intersect the vertical reference line at zero), in which cases, a set of four asterisks are marked next to the sample pairs.

adenocarcinomal samples. Figure 8 shows that FKBP51 staining in Pca was significantly higher (50%) than that in BPH, indicating that increased FKBP51 may play a role in the progression of prostate cancers to hormone refractory status.

In this study, we used oligonucleotide array technology in conjunction with statistical filtering methods to identify 692 ARGs in LNCaP prostate cancer cells, many of which have not yet been previously reported as being androgen regulated or relevant to prostate cancer. Using quantitative PCR, we validated the expression of three candidate targets (SNRK, Seladin-1, FKBP51), a control gene PSA, and further demonstrated that FKBP51 may play a role in prostate cancer development. We focused on ARGs because they may be responsible not only for normal growth of prostate cancer but also for hormone-independent growth if deregulation of their expression occurs. Insight into these newly identified candidate ARGs may lead to a better understanding of the molecular mechanisms leading to the proliferation, differentiation, and function of the normal and diseased human prostate. LNCaP prostate cancer cells were used for our experi-

Velasco et al. • Androgen-Regulated Genes in LNCaP Prostate Cancer Cells

ments because they maintain responsiveness to androgen (26). For example, their ability to proliferate (41, 42), express differentiated secretory function (43– 46), and control processes such as lipid synthesis and accumulation (47) all remains androgen responsive. Others have used the LNCaP model and performed similar comprehensive array studies on ARGs (24, 36, 37). However, when we compared the results from these studies, we found that the number of identified ARGs varied considerably from study to study and that the number of overlapping genes between any two studies was relative low (Table 4; see also the supplemental data published on The Endocrine Society’s Journals Online web site, http://endo.endojournals.org/ for details). Moreover, only 13 genes were identified as androgen regulated by all four studies (Table 5). The differences among these studies are alarming and may be caused by multiple factors. First, most studies (24, 36, 37) used a synthetic androgen, R1881, whereas we used the natural in vivo ligand, DHT. Whereas R1881 can achieve pharmacological activation of AR and appear to produce similar results as DHT as measured by array experiments (36), determining a suitable and relevant concentration of R1881 is not trivial and can impact binding of the ligand, global gene expression changes, and ultimately its physiological effects (48). Second, two studies (24, 36) made use of cDNA spotted arrays rather than GeneChips. GeneChip technology has considerable advantages TABLE 4. Cross-comparisons of ARGs between any two studies

Velasco Nelson DePrimo Segawa Xu Waghray

Velasco

Nelson

DePrimo

Segawa

Xua

Waghraya

692 25 48 113 12 4

25 105 49 30 8 3

48 49 567 53 6 5

113 30 53 579 7 5

12 8 6 7 55b 1

4 3 5 5 1 121c

Data sets from each study were downloaded and compared with each other. Each number represents the overlapped ARGs in any two studies compared. Bold number indicates the number of genes from individual study. The detailed results are available as supplemental data on The Endocrine Society’s Journals Online web site, http://endo. endojournals.org/. a Used SAGE technology. b Data set from Table 1 of Xu et al. (51). c Data set from Table 3 of Waghray et al. (52).

Endocrinology, August 2004, 145(8):3913–3924 3921

over spotted cDNA microarrays because it allows for quantifying absolute values of gene expression, which is necessary for performing powerful statistical analyses across chips and experiments (49). Another advantage is that the probability of cross-hybridization is largely minimized, leading to greater reproducibility and accuracy of the data (50). Another major difference between our study and others’ is that we prepared duplicate sets of cells treated or not treated with androgen in parallel for 2, 4, 6, 12, 24, 48, and 72 h. To include nontreated controls for each time point may be critical because we noticed that the expression levels of several genes in untreated cells changed over time (Fig. 2, A and B). Studies using only the 0 h as a general control in a time-course study may result in false-positive or -negative hits by artificially inflating or deflating fold-change values. Furthermore, because we were concerned with reproducibility such as chipto-chip variation, we also performed the GeneChip hybridizations of the two sets of biological replicates in duplicate. This time-matched, replicated, and balanced-design approach substantially increases the statistical power and sensitivity of data analyses and allows separation of interesting biological phenomena from the confounding experimental variations. The unique gene sets that were included on the arrays and filters used to qualify ARGs greatly contributed to the differences. The results from our study most closely overlapped with Segawa et al. (37), largely due to the fact that in both experiments, Affymetrix GeneChips were used and the gene probe sets that were on the arrays had significant overlap. Even so, only 113 of several hundred differential expressed genes overlapped. Despite all of these differences, however, we believe our study complements previously published microarray studies on ARGs by both validating recently identified ARGs and building on existing gene expression data with respect to androgen action on prostate cancer cells. We also compared all four array studies with two previously published SAGE studies (51, 52). The number of overlapping genes between the microarray studies and the SAGE studies was relatively small (Table 4). Interestingly, by including the two SAGE experiments in the cross-study comparison, only two genes, FKBP51 and Seladin-1, were also found to be identified by one of the SAGE studies (51). There

TABLE 5. Thirteen ARGs that are found among all four studies Ref. Seq. ID

Accession no.

Description

NM_014762a NM_014606 NM_005941 NM_006096 NM_000055 NM_004578 NM_001328 NM_004117a NM_003629 NM_002026 NM_001648 NM_002310 NM_001260

D13643 D25215 D83646 D87953 M16474 M28211 U37408 U42031 U90907 X02761 X07730 X61615 X85753

Homo sapiens mRNA for KIAA0018 protein, partial cds Human mRNA for KIAA0032 gene, complete cds H. sapiens mRNA for metalloproteinase, complete cds Human mRNA for RTP, complete cds Human fetal butyrylcholinesterase mRNA, complete cds H. sapiens GTP-binding protein (RAB4) mRNA, complete cds H. sapiens phosphoprotein CtBP mRNA, complete cds Human 54-kDa progesterone receptor-associated immunophilin FKBP54 mRNA, partial cds Phosphoinositide-3-kinase, regulatory subunit Fibronectin, Alt. Splice 1 Human mRNA for prostate specific antigen H. sapiens mRNA for leukemia inhibitory factor (LIF) receptor H. sapiens mRNA for CDK8 protein kinase

Data set from four different studies were downloaded and compared with each other. Thirteen genes that are present in all four studies are shown. a Gene also identified by Xu et al. (51).

3922

Endocrinology, August 2004, 145(8):3913–3924

were no genes that overlapped all six studies. Again, these large differences in data sets may be caused by the factors that we have already previously discussed and the fact that the complete data sets for the SAGE experiments were not readily available. Furthermore, the few number of overlapping genes is not necessarily surprising, given the fact that SAGE itself is an entirely different approach, which is beyond the scope of this discussion. Using quantitative PCR, we have begun to validate the GeneChip expression results from one set of LNCaP samples for a few candidate ARGs (Fig. 5). In addition, because we have identified many genes that have been previously reported as ARGs, namely PSA, prostate-specific membrane antigen, PacP, KLK2 (53), selenium binding protein (54), ␣-tubulin (55, 56), UGT2B15 (57, 58), DRG (differentiationrelated gene) 1 (59), TRPM (testosterone-repressed prostate message)-2 (60 – 62), and cyclin-dependent kinase-8 (24), it is conceivable that other genes identified in our study are also ARGs. In particular, those genes that had similar expression profiles and were found to cluster with known ARGs are of great interest. Several of these genes, such as Id3 and FKBP51, have recently been reported by investigators as being relevant to prostate cancer using different prostate tumor models (18, 19). We further investigated FKBP51 because of its near identical expression pattern to that of PSA. Both the transcription and protein levels of FKBP51 are regulated by DHT as well as R1881 and were actually higher in androgen-independent tumor cells, compared with androgen-dependent cells. To the best of our knowledge, our study is the first to examine FKBP51 protein level using tissue microarrays. Our results showed that protein expression of FKBP51 was significantly higher in Pca than in BPH samples. Unfortunately, our sample size for healthy control samples was too small (n ⫽ 4) for a statistically significant comparison (data not shown). Interestingly, our data mining and statistical analysis on human prostate tissues using the GeneLogic database provided further evidence of higher expression of FKBP51 mRNA in Pca (n ⫽ 57) relative to normal (n ⫽ 6), normal adjacent (n ⫽ 34), and possibly BPH samples (n ⫽ 20) (Fig. 6). Whereas the difference between Pca and BPH FKBP51 mRNA is not as significant as what we expected based on the tissue microarray results, it is not necessarily surprising, given the fact that differences in sample size can largely impact the statistical significance of any differences in gene expression data. Furthermore, in some cases, mRNA levels may not be predictive of protein levels. Our observation of higher FKBP51 protein levels in tumor samples relative to BPH samples is somewhat consistent with what was reported by Dhanasekaran et al. (61, 62a), who demonstrated through cDNA microarrays that FKBP51 is down-regulated in BPH, normal adjacent prostate samples, and normal prostate pool, compared with Pca samples. However, in their studies, FKBP51 mRNA levels were higher in BPH samples (n ⫽ 4), compared with normal adjacent samples (n ⫽ 8), but lower, compared with normal prostate pool (61). These inconsistencies in FKBP51 expression from normal to the diseased state may be attributed to sample size used in each study. It is critical to have a large sample size in each disease group to minimize the intrinsic tissue heterogeneity and

Velasco et al. • Androgen-Regulated Genes in LNCaP Prostate Cancer Cells

biological variation such as disease stage, age, ethnic background, and/or technical variation such as sample collection and preservation. We hope future studies using larger number of normal adjacent prostate and normal prostate control samples will provide a better understanding of differing mRNA and protein levels of FKBP51 relative to diseased samples. FKBP51 belongs to a family of intracellular receptors for immune suppressant drugs, such as FK506 and rapamycin. FKBPs are known as helper proteins of the enzyme class of peptidyl prolyl cis/trans isomerases and was initially identified as a gene regulated by glucocorticoids in murine thymocytes (63). Recently it was shown that peptidyl prolyl cis/trans isomerases associate with heat shock protein 90 and steroid receptors to form mature active heterocomplexes (64). Two recent studies showed that inhibition of cell growth by silymarin (65) and geldanamycin (66) correlated with the reduced protein level and transcripts of FKBP51, respectively. Interestingly, FKBP38, a member of the FKBP family, was reported to block apoptosis when overexpressed, whereas its functional inhibition promoted apoptosis (67). Although the role of FKBP51 in prostate cancer pathogenesis remains unclear, our results, data obtained from the CWR22R model by others (18, 19), and a study showing inhibition of FKBP51 expression by two Cox-2 inhibitors (68) suggest that FKBP51 can be used as a diagnostic marker and is potentially important for the hormone-independent prostate cancers. In examining other genes regulated by androgen, we further investigated two novel ARGs, SNRK and Seladin-1 [While we were preparing our manuscript, three studies (24, 36, 37) reported Seladin-1 as an ARG]. We found that Seladin-1 exhibited an expression pattern very similar to PSA and was significantly up-regulated in LNCaP cancer cells upon treatment of androgen. Seladin-1, a homolog of diminuto/ dwarf1 gene, is conserved among plants and Caenorhabditis elegans (69). In Arabidopsis, diminuto plays a role in cell elongation (70). It has also been shown that diminuto/ dwarf1 is involved in the conversion of isofucosterol to sitosterol and 24-methylenecholesterol to campesterol during steroid synthesis (71). The function of Seladin-1 in mammalian cells has yet to be understood. Recently, Greeve et al. (72) cloned Seladin-1 from human brain and found that it confers resistance to Alzheimer’s disease-associated neurodegeneration and protects cells from oxidative stress-induced cell death. Interestingly, a study from Sarkar et al. (73) suggests an important role of Seladin-1 in adrenocortical tumorigenesis by facilitating steroid synthesis and cell growth. It is worth mentioning that ␤-sitosterol, perhaps via a feedback mechanism to reduce Seladin-1 level, is effective in treating BPH in men, suggesting that Seladin-1 may play a part in reducing the growth of prostate cells (74). SNRK, on the other hand, is down-regulated by androgen in LNCaP prostate cancer cells. Based on protein homology analysis, SNRK is a putative serine/threonine kinase. Kinases, as a class, play important roles in cell cycle and proliferation, and have been recently increasingly targeted as candidate drug targets for a variety of diseases, including a variety of cancers. Interestingly, our subsequent tissue analysis has revealed that SNRK levels increased with tumor

Velasco et al. • Androgen-Regulated Genes in LNCaP Prostate Cancer Cells

grade (data not shown). It is tempting to think that deregulation of SNRK is directly involved in cellular proliferation of recurrent tumors and could potentially serve as a target for developing an anticancer drug. Among the 692 ARGs, up- and down-regulated genes were approximately evenly distributed. Functional classification shows that these genes belong to a broad set of categories (Table 3), suggesting that androgen regulation is very complex. Interestingly, several repressors including RTP (reducing agents and tunicamycin-responsive protein) were up-regulated by androgen, raising the possibility that their expression in tumors may decrease during androgen ablation therapy, providing a growth advantage to tumor cells. Whether this may help to explain why androgen ablation therapy often fails remains to be understood. Nevertheless, specific inhibition of ARGs, which are involved in promoting tumor growth, may offer some advantages. Newly defined ARGs may also facilitate the identification of novel targets for therapy of hormone-independent prostate cancer. Based on the results of our study, several genes including FKBP51, Seladin-1, and a novel kinase, SNRK, hold promise as potential markers or targets in the diagnosis and treatment of prostate cancer and warrant further investigation of their functional roles in Pca development. Acknowledgments The authors thank Dr. David Smith for the FKBP51 expression vector and ␣-FKBP51 antibody, Dr. Thomas P. Pretlow for the CWR22 and CWR22R tumor samples, Dr. Andrew Hill for his assistance with the clustering software, and Ms. Linda Sauter for her technical assistance with the quantitative RT-PCR experiments. Received March 10, 2004. Accepted April 26, 2004. Address all correspondence and requests for reprints to: Yixian Zhang, Department of Oncology Research, 401 North Middletown Road, New York, New York 10965. E-mail: [email protected]. A.M.V. and K.A.G. contributed equally to the work. Present address for A.M.V.: Ambit Biosciences, 9875 Towne Centre Drive, Suite 100, San Diego, California 92121. Present address for K.A.G.: U.S. Genomics, 6H Gill Street, Woburn, Massachusetts 01801.

References 1. Coffey DS 1993 Prostate cancer. An overview of an increasing dilemma. Cancer 71:880 – 886 2. Carter HB, Coffey DS 1990 The prostate: an increasing medical problem. Prostate 16:39 – 48 3. Mooradian AD, Morley JE, Korenman SG 1987 Biological actions of androgens. Endocr Rev 8:1–28 4. Culig Z, Hobisch A, Hittmair A, Peterziel H, Cato Andrew CB, Bartsch G, Klocker H 1998 Expression, structure, and function of androgen receptor in advanced prostatic carcinoma. Prostate 35:63–70 5. Koivisto P, Kolmer M, Visakorpi T, Kallioniemi Olli P 1998 Androgen receptor gene and hormonal therapy failure of prostate cancer. Am J Pathol 152:1–9 6. Gittes RF 1991 Carcinoma of the prostate. N Engl J Med 324:236 –245 7. Schroder F 1998 Endocrine treatment of prostate cancer. In: Walsh PC, Retik AB, Vaughan ED, Wein AJ, eds. Campbell’s urology. Philadelphia: WB Saunders; 2627–2644 8. Jenster G 1999 The role of the androgen receptor in the development and progression of prostate cancer. Semin Oncol 26:407– 421 9. Bubendorf L, Kononen J, Koivisto P, Schraml P, Moch H, Gasser TC, Willi N, Mihatsch MJ, Sauter G, Kallioniemi OP 1999 Survey of gene amplifications during prostate cancer progression by high throughput fluorescence in situ hybridization on tissue microarrays. Cancer Res 59:803– 806 10. Linja MJ, Savinainen KJ, Saramaki OR, Tammela TLJ, Vessella RL, Visakorpi T 2001 Amplification and overexpression of androgen receptor gene in hormone-refractory prostate cancer. Cancer Res 61:3550 –3555

Endocrinology, August 2004, 145(8):3913–3924 3923

11. Oesterling JE 1993 PSA leads the way for detecting and following prostate cancer. Contemp Urol 5:60 – 81 12. Oesterling JE 1992 Prostate-specific antigen. Improving its ability to diagnose early prostate cancer. JAMA 267:2236 –2238 13. Peace D, Qin G, Kelley L, Medin JA 2000 Immunogene therapy for prostate cancer by engineered expression of prostate specific membrane antigen (PSMA). Cancer Gene Ther 7:S29 14. Chang Sam S, Reuter Victor E, Heston WDW, Bander Neil H, Grauer Lana S, Gaudin Paul B 1999 Five different anti-prostate-specific membrane antigen (PSMA) antibodies confirm PSMA expression in tumor-associated neovasculature. Cancer Res 59:3192–3198 15. Lockhart DJ, Dong H, Byrne MC, Follettie MT, Gallo MV, Chee MS, Mittmann M, Wang C, Kobayashi M, Horton H, Brown EL 1996 Expression monitoring by hybridization to high-density oligonucleotide arrays. Nat Biotechnol 14:1675–1680 16. Wodicka L, Dong H, Mittmann M, Ho MH, Lockhart DJ 1997 Genome-wide expression monitoring in Saccharomyces cerevisiae. Nat Biotechnol 15:1359 –1367 17. Vaarala Markku H, Porvari K, Kyllonen A, Vihko P 2000 Differentially expressed genes in two LNCaP prostate cancer cell lines reflecting changes during prostate cancer progression. Lab Invest 80:1259 –1268 18. Amler LC, Agus DB, LeDuc C, Sapinoso ML, Fox WD, Kern S, Lee D, Wang V, Leysens M, Higgins B, Martin J, Gerald W, Dracopoli N, Cordon-Cardo C, Scher HI, Hampton GM 2000 Dysregulated expression of androgen-responsive and nonresponsive genes in the androgen-independent prostate cancer xenograft model CWR22–R1. Cancer Res 60:6134 – 6141 19. Mousses S, Wagner U, Chen Y, Kim JW, Bubendorf L, Bittner M, Pretlow T, Elkahloun Abdel G, Trepel Jane B, Kallioniemi Olli P 2001 Failure of hormone therapy in prostate cancer involves systematic restoration of androgen responsive genes and activation of rapamycin sensitive signaling. Oncogene 20:6718 – 6723 20. Elek J, Park KH, Narayanan R 2000 Microarray-based expression profiling in prostate tumors. In Vivo 14:173–182 21. Howell SB 1999 DNA microarrays for analysis of gene expression in prostate cancer. Mol Urol 3:295–299 22. Welsh JB, Sapinoso LM, Su AI, Kern SG, Wang-Rodriguez J, Moskaluk CA, Frierson HF, Hampton GM 2001 Analysis of gene expression identifies candidate markers and pharmacological targets in prostate cancer. Cancer Res 61:5974 –5978 23. Chaib H, Cockrell EK, Rubin MA, Macoska JA 2001 Profiling and verification of gene expression patterns in normal and malignant human prostate tissues by cDNA microarray analysis. Neoplasia 3:43–52 24. Nelson PS, Clegg N, Arnold H, Ferguson C, Bonham M, White J, Hood L, Lin BY 2002 The program of androgen-responsive genes in neoplastic prostate epithelium. Proc Natl Acad Sci USA 99:11890 –11895 25. Karan D, Kelly DL, Rizzino A, Lin MF, Batra SK 2002 Expression profile of differentially regulated genes during progression of androgen-independent growth in human prostate cancer cells. Carcinogenesis 23:967–975 26. Horoszewicz JS, Leong SS, Kawinski E, Karr JP, Rosenthal H, Chu TM, Mirand EA, Murphy GP 1983 LNCaP model of human prostatic carcinoma. Cancer Res 43:1809 –1818 27. Nomura N, Miyajima N, Sazuka T, Tanaka A, Kawarabayasi Y, Sato S, Nagase T, Seki N, Ishikawa K, Tabata S 1994 Prediction of the coding sequences of unidentified human genes. I. The coding sequences of 40 new genes (KIAA0001-KIAA0040) deduced by analysis of randomly sampled cDNA clones from human immature myeloid cell line KG-1 (supplement). DNA Res 1:47–56 28. Nair Satish C, Rimerman Ronald A, Toran Eric J, Chen S, Prapapanich V, Butts Rinalda N, Smith David F 1997 Molecular cloning of human FKBP51 and comparisons of immunophilin interactions with Hsp90 and progesterone receptor. Mol Cell Biol 17:594 – 603 29. De Angelis Margaret M, Wang David G, Hawkins Trevor L 1995 Solid-phase reversible immobilization for the isolation of PCR products. Nucleic Acids Res 23:4742– 4743 30. Hill AA, Brown EL, Whitley MZ, Tucker-Kellogg G, Hunter CP, Slonim DK: Springer 2001 Evaluation of normalization procedures for oligonucleotide array data based on spiked cRNA controls. Genome Biol 2:RESEARCH0055 31. Kohonen T 1997 Self organizing maps. 2nd extended ed. New York: Springer 32. Tamayo P, Slonim D, Mesirov J, Zhu Q, Kitareewan S, Dmitrovsky E, Lander ES, Golub TR 1999 Interpreting patterns of gene expression with self-organizing maps: methods and application to hematopoietic differentiation. Proc Natl Acad Sci USA 96:2907–2912 33. Diehn M, Sherlock G, Binkley G, Jin H, Matese JC, Hernandez-Boussard T, Rees CA, Cherry JM, Botstein D, Brown PO, Alizadeh AA 2003 SOURCE: a unified genomic resource of functional annotations, ontologies, and gene expression data. Nucleic Acids Res 31:219 –223 34. Appel RD, Bairoch A, Hochstrasser DF 1994 A new generation of information retrieval tools for biologists: the example of the ExPASy WWW server. Trends Biochem Sci 19:258 –260 35. Bradford MM 1976 A rapid and sensitive method for the quantitation of microgram quantities of protein utilizing the principle of protein-dye binding. Anal Biochem 72:248 –254 36. DePrimo SE, Diehn M, Nelson JB, Reiter RE, Matese J, Fero M, Tibshirani

3924

37.

38. 39. 40.

41. 42. 43.

44.

45.

46. 47. 48.

49. 50. 51.

52.

53.

54. 55.

Endocrinology, August 2004, 145(8):3913–3924

R, Brown PO, Brooks JD 2002 Transcriptional programs activated by exposure of human prostate cancer cells to androgen. Genome Biol 3:RESEARCH0032 Segawa T, Nau ME, Xu LL, Chilukuri RN, Makarem M, Zhang W, Petrovics G, Sesterhenn IA, McLeod DG, Moul JW, Vahey M, Srivastava S 2002 Androgen-induced expression of endoplasmic reticulum (ER) stress response genes in prostate cancer cells. Oncogene 21:8749 – 8758 Lin M-F, Lee M-S, Garcia-Arenas R, Lin F-F 2000 Differential responsiveness of prostatic acid phosphatase and prostate-specific antigen mRNA to androgen in prostate cancer cells. Cell Biol Int 24:681– 689 Mattmueller DR, Hinton BT 1992 Clusterin Sgp-2 in epididymal luminal fluid and its association with epididymal spermatozoa in androgen-deprived rats. Mol Reprod Dev 32:73– 80 Nagabhushan M, Miller CM, Pretlow TP, Giaconia JM, Edgehouse NL, Schwartz S, Kung H-J, de Vere White RW, Gumerlock PH, Resnick MI, Amini SB, Pretlow TG 1996 CWR22: the first human prostate cancer xenograft with strongly androgen-dependent and relapsed strains both in vivo and in soft agar. Cancer Res 56:3042–3046 Kokontis J, Takakura K, Hay N, Liao S 1994 Increased androgen receptor activity and altered c-myc expression in prostate cancer cells after long-term androgen deprivation. Cancer Res 54:1566 –1573 Schuurmans AL, Bolt J, Veldscholte J, Mulder E 1991 Regulation of growth of LNCaP human prostate tumor cells by growth factors and steroid hormones. J Steroid Biochem Mol Biol 40:193–197 Swinnen JV, Esquenet M, Heyns W, Rombauts W, Verhoeven G 1994 Androgen regulation of the messenger RNA encoding diazepam-binding inhibitor/acyl-CoA-binding protein in the human prostatic adenocarcinoma cell line LNCaP. Mol Cell Endocrinol 104:153–162 Cleutjens KB, van Eekelen CC, van der Korput HA, Brinkmann AO, Trapman J 1996 Two androgen response regions cooperate in steroid hormone regulated activity of the prostate-specific antigen promoter. J Biol Chem 271: 6379 – 6388 Henttu P, Liao SS, Vihko P 1992 Androgens up-regulate the human prostatespecific antigen messenger ribonucleic acid (mRNA), but down-regulate the prostatic acid phosphatase mRNA in the LNCaP cell line. Endocrinology 130:766 –772 Murtha P, Tindall DJ, Young CY 1993 Androgen induction of a human prostate-specific kallikrein, hKLK2: characterization of an androgen response element in the 5⬘ promoter region of the gene. Biochemistry 32:6459 – 6464 Swinnen JV, Van Veldhoven PP, Esquenet M, Heyns W, Verhoeven G 1996 Androgens markedly stimulate the accumulation of neutral lipids in the human prostatic adenocarcinoma cell line LNCaP. Endocrinology 137:4468 – 4474 Castoria G, Lombardi M, Barone MV, Bilancio A, Di Domenico M, Bottero D, Vitale F, Migliaccio A, Auricchio F 2003 Androgen-stimulated DNA synthesis and cytoskeletal changes in fibroblasts by a nontranscriptional receptor action. J Cell Biol 161:547–556 Schulze A, Downward J 2001 Navigating gene expression using microarrays—a technology review. Nat Cell Biol 3:E190 –E195 Harrington CA, Rosenow C, Retief J 2000 Monitoring gene expression using DNA microarrays. Curr Opin Microbiol 3:285–291 Xu LL, Su YP, Labiche R, Segawa T, Shanmugam N, McLeod DG, Moul JW, Srivastava S 2001 Quantitative expression profile of androgen-regulated genes in prostate cancer cells and identification of prostate-specific genes. Int J Cancer 92:322–328 Waghray A, Feroze F, Schober MS, Yao FY, Wood C, Puravs E, Krause M, Hanash S, Chen YQ 2001 Identification of androgen-regulated genes in the prostate cancer cell line LNCaP by serial analysis of gene expression and proteomic analysis. Proteomics 1:1327–1338 Sun Z, Pan J, Balk SP 1997 Androgen receptor-associated protein complex binds upstream of the androgen-responsive elements in the promoters of human prostate-specific antigen and kallikrein 2 genes. Nucleic Acids Res 25:3318 –3325 Yang M, Sytkowski AJ 1998 Differential expression and androgen regulation of the human selenium-binding protein gene hSP56 in prostate cancer cells. Cancer Res 58:3150 –3153 Gregory CW, Hamil KG, Kim D, Hall SH, Pretlow TG, Mohler JL, French

Velasco et al. • Androgen-Regulated Genes in LNCaP Prostate Cancer Cells

FS 1998 Androgen receptor expression in androgen-independent prostate cancer is associated with increased expression of androgen-regulated genes. Cancer Res 58:5718 –5724 56. Buttyan R, Zakeri Z, Lockshin R, Wolgemuth D 1988 Cascade induction of c-fos, c-myc, and heat shock 70K transcripts during regression of the rat ventral prostate gland. Mol Endocrinol 2:650 – 657 57. Guillemette C, Levesque E, Beaulieu M, Turgeon D, Hum DW, Belanger A 1997 Differential regulation of two uridine diphospho-glucuronosyltransferases, UGT2B15 and UGT2B17, in human prostate LNCaP cells. Endocrinology 138:2998 –3005 58. Chang GT, Blok LJ, Steenbeek M, Veldscholte J, van Weerden WM, van Steenbrugge GJ, Brinkmann AO 1997 Differentially expressed genes in androgen-dependent and -independent prostate carcinomas. Cancer Res 57:4075– 4081 59. Ulrix W, Swinnen JV, Heyns W, Verhoeven G 1999 The differentiation-related gene 1, Drg1, is markedly up-regulated by androgens in LNCaP prostatic adenocarcinoma cells. FEBS Lett 455:23–26 60. Miyake H, Nelson C, Rennie PS, Gleave ME 2000 Testosterone-repressed prostate message-2 is an antiapoptotic gene involved in progression to androgen independence in prostate cancer. Cancer Res 60:170 –176 61. Leger JG, Montpetit ML, Tenniswood MP 1987 Characterization and cloning of androgen-repressed mRNAs from rat ventral prostate. Biochem Biophys Res Commun 147:196 –203 62. Rennie PS, Bruchovsky N, Akakura K, Goldenberg SL, Otal N, Akakura S, Wong P, Tenniswood M 1994 Effect of tumour progression on the androgenic regulation of the androgen receptor, TRPM-2 and YPT1 genes in the Shionogi carcinoma. J Steroid Biochem Mol Biol 50:31– 40 62a.Dhanasekaran SM, Barrette TR, Ghosh D, Shah R, Varambally S, Kurachi K, Pienta KJ, Rubin MA, Chinnaiyan AM 2001 Delineation of prognostic biomarkers in prostate cancer. Nature 412:822– 826 63. Baughman G, Harrigan MT, Campbell NF, Nurrish SJ, Bourgeois S 1991 Genes newly identified as regulated by glucocorticoids in murine thymocytes. Mol Endocrinol 5:637– 644 64. Schiene-Fischer C, Yu C 2001 Receptor accessory folding helper enzymes: the functional role of peptidyl prolyl cis/trans isomerases. FEBS Lett 495:1– 6 65. Zhu W, Zhang JS, Young CYF 2001 Silymarin inhibits function of the androgen receptor by reducing nuclear localization of the receptor in the human prostate cancer cell line LNCaP. Carcinogenesis 22:1399 –1403 66. Vanaja DK, Mitchell SH, Toft DO, Young CYF 2002 Effect of geldanamycin on androgen receptor function and stability. Cell Stress Chaperones 7:55– 64 67. Shirane M, Nakayama KI 2003 Inherent calcineurin inhibitor FKBP38 targets Bcl-2 to mitochondria and inhibits apoptosis. Nat Cell Biol 5:28 –37 68. Pan Y, Zhang JS, Gazi MH, Young CY 2003 The cyclooxygenase 2-specific nonsteroidal anti-inflammatory drugs celecoxib and nimesulide inhibit androgen receptor activity via induction of c-Jun in prostate cancer cells. Cancer Epidemiol Biomarkers Prev 12:769 –774 69. Koch CA, Chrousos GP 2001 Is the diminuto/dwarf1 gene involved in physiologic adrenocortical size regulation and tumor formation? J Clin Endocrinol Metab 86:5127–5129 (Editorial) 70. Takahashi T, Gasch A, Nishizawa N, Chua NH 1995 The DIMINUTO gene of Arabidopsis is involved in regulating cell elongation. Genes Dev 9:97–107 71. Klahre U, Noguchi T, Fujioka S, Takatsuto S, Yokota T, Nomura T, Yoshida S, Chua NH 1998 The Arabidopsis DIMINUTO/DWARF1 gene encodes a protein involved in steroid synthesis. Plant Cell 10:1677–1690 72. Greeve I, Hermans-Borgmeyer I, Brellinger C, Kasper D, Gomez-Isla T, Behl C, Levkau B, Nitsch RM 2000 The human DIMINUTO/DWARF1 homolog seladin-1 confers resistance to Alzheimer’s disease-associated neurodegeneration and oxidative stress. J Neurosci 20:7345–7352 73. Sarkar D, Imai T, Kambe F, Shibata A, Ohmori S, Siddiq A, Hayasaka S, Funahashi H, Seo H 2001 The human homolog of Diminuto/Dwarf1 gene (hDiminuto): a novel ACTH-responsive gene overexpressed in benign cortisolproducing adrenocortical adenomas. J Clin Endocrinol Metab 86:5130 –5137 74. Berges RR, Kassen A, Senge T 2000 Treatment of symptomatic benign prostatic hyperplasia with ␤-sitosterol: an 18-month follow-up. BJU Int 85:842– 846

Endocrinology is published monthly by The Endocrine Society (http://www.endo-society.org), the foremost professional society serving the endocrine community.