Dysregulation of HOX gene expression has been implicated as a factor in malignancies for a number of years. However, no consensus has emerged regarding specific causative genes. Using a degenerate reverse transcription-PCR technique, we show up-regulation of genes from the HOXC cluster in malignant prostate cell lines and lymph node metastases. When relative expression levels of the four HOX clusters were examined, lymph node metastases and cell lines derived from lymph node metastases exhibited very similar patterns, patterns distinct from those in benign cells or malignant cell lines derived from other tumor sites. Specific reverse transcription-PCR for HOXC4 , HOXC5 , HOXC6 , and HOXC8 confirmed overexpression of these genes in malignant cell lines and lymph node metastases. Laser capture microdissection and examination of paired tumor/normal prostate epithelial cells also indicated overexpression of these HOXC genes in primary tumor cells. Our data indicate a possible link between expression of HOXC genes and malignancy in prostate cells. Overexpression of HOXC8 in LNCaP prostate cancer cells suppressed transactivation by androgen receptors. We speculate that HOXC overexpression may predispose tumor cells to androgen independence by necessitating adaptation to diminished androgen signaling.

Prostate cancer is the most common male cancer and the second leading cause of male cancer deaths in the United States (1). Autopsy data indicate nearly 30% of men 30–39 years old have small foci of carcinoma in their prostate (2), and it is believed nearly all men develop at least one small focus with age (3). Thus, although more men die with prostate cancer than of it, progression of the cancer in an individual usually results in a poor outcome. Metastasis to a site other than prostate has long been a problem poorly understood and usually leading to shortened life span. Although relatively little is known of the underlying molecular events that lead to progression of a cancer, clearly cellular functions such as proliferation, cell death, motility, signal transduction, and microenvironment interactions must play a role. The homeobox superfamily of genes and the HOX subfamily contain members that are transcription factors involved in controlling and coordinating complex functions during development via spatial and temporal expression patterns. There are growing lines of evidence that dysregulation of HOX gene expression plays important roles in cancer (see Refs. 4, 5, 6 for review).

Dysregulation of a variety of HOX genes has been implicated in several human cancers including leukemias (7, 8, 9, 10, 11); colorectal (12), breast (13, 14), and renal (15) carcinomas; melanomas (16); and squamous carcinomas of the skin (17). Because the genes implicated show little consensus, the dysregulation may be a tissue-specific perturbation of the existing HOX expression pattern rather than a single causative gene. Tissue-specific expression patterns have been reported in kidney and colon, by Northern blot analysis (12, 15). Primary tumors in both kidney and colon showed variations in specific HOX gene expression from the corresponding normal tissue, but overall expression patterns for individual tumors were not reported. Only primary kidney tumors were examined (15), but liver metastases from colon tumors reportedly displayed expression of specific HOX genes similar to that seen in either primary colon tumors or normal colon but not in normal liver (12). Translocations involving several HOXA , HOXC , and HOXD genes and the NUP 98 nucleoporin gene have been reported in hematopoietic malignancies (11, 18, 19, 20, 21, 22); however, translocations have not been reported for solid tumors.

In humans, HOX genes comprise the largest of several families of genes containing a sequence motif termed the homeobox. The homeobox was first identified as a motif shared among the Drosophila homeotic genes. HOX genes represent their human counterpart. HOX genes are highly conserved across a very broad range of animal organisms from Caenorhabditis elegans and Drosophila through humans. A uniform nomenclature for vertebrates was established in 1992 using naming conventions with uppercase letters for human genes (HOXA1 ) and lowercase letters for mouse genes [Hoxa1(23)]. In humans, the 39 members of the HOX family are organized into four clusters (A–D) on chromosomes 7, 17, 12, and 2, respectively, which arose through reduplication and divergence of the ancestral cluster (24). “Paralogy” is used to describe the relationship between genes from different clusters that demonstrate the greatest sequence similarity as well as the same linear arrangement along the chromosomes (e.g ., HOXA5 , HOXB5 , and HOXC5 ). Whereas 13 potential paralogy groups exist, no HOX cluster has members in all 13 groups, and each group therefore contains between 2 and 4 members. The sequence similarities, as well as overlapping expression in developing embryos (25), have suggested possible functional redundancy. Gene disruption studies in mice indicate some degree of functional cooperation during development (25, 26), and these studies are being extended into adult tissues (27). However, the relationships between paralogs appear complex, and the relative importance of individual HOX genes is currently unknown. Indeed, evidence is accumulating that the same HOX gene functions differently in different tissues (28, 29) and that distinct portions of the protein structure are important for this specificity (30). This may also prove to be the case between developing and adult tissues.

Very little has been published on HOX gene expression in human prostate. Detectable expression of HOXD10 and HOXA9 RNA was reported in normal prostate (31). Regulation of two unidentified homeobox-containing genes by extracellular matrix and androgens in prostatic carcinoma cells has been reported (32). One recent publication correlated HOXC8 expression with higher Gleason grades in prostate tumors (33). A divergent homeobox gene, NKX3.1 , is also expressed in normal human prostate and has been implicated as a tumor suppressor gene (34). This conclusion is bolstered by data from mice with targeted disruption of the Nkx3.1 gene, in which age-dependent hyperplastic and dysplastic lesions resembling prostate intraepithelial neoplasia are seen (35, 36, 37). Our laboratory has recently completed a large screen of expression patterns of the HOX family of genes in benign and malignant prostate cells in vitro , as well as in benign and malignant tissue specimens, to determine whether shifts in relative expression levels might provide evidence of a role for HOX genes in prostate cancer development or progression. Our findings indicate changes in HOX gene expression patterns with malignancy and, in particular, an up-regulation of HOXC cluster genes. In particular, lymph node metastases present a distinct pattern of HOX expression. Surprisingly, this pattern is maintained in established cell lines derived from lymph node metastases. Finally, we also show that overexpression of one of the HOXC genes, HOXC8 , suppresses androgen-dependent transcription in prostate cancer cells, suggesting that dysregulation of HOX gene expression has important consequences in prostate cancer.

Cell Lines and Culture.

The human prostate cancer cell lines were obtained from the following laboratories: LAPC-4 (38), Dr. C. Sawyers (University of California at Los Angeles, Los Angeles, CA); LNCaP (39, 40), Dr. J. Horoszewicz (Roswell Park Memorial Institute, Buffalo, NY); ALVA-31 (41), Dr. S. Loop (American Lake VAMC, Tacoma, WA); PPC-1 (42), Dr. A. Brothman (University of Utah, Salt Lake City, UT); 22Rv 1 (43), Dr. J. Jacobberger (Case Western Reserve University, Cleveland, OH); and PC-346C (44), Dr. W. van Weerden (Erasmus University, Rotterdam, the Netherlands). The cell lines PC-3 (45) and DU 145 (46) were obtained from the American Type Culture Collection (Manassas, VA). Stromal cells were from primary cultures established from BPH5 nodules obtained from radical prostatectomy specimens. Primary cultures of PrECs from three independent origins were obtained from BioWhittaker (San Diego, CA) and grown in PrEGM supplemented as per the manufacturer’s instructions. LAPC-4 cells were grown in Iscove’s modified Dulbecco’s medium (Invitrogen, Carlsbad, CA) with 7.5% FBS (Omega Scientific, Tarzana, CA) and 10−8m methyltrienolone (R1881; Perkin-Elmer Life Sciences, Boston, MA). PC-346C cells were grown in DMEM/Ham’s F-12 plus 10% FBS (Hyclone, Logan, UT). All other cell lines were grown in RPMI 1640 (Invitrogen) plus 10% FBS (Hyclone). The authenticity of each was verified by comparison with the original karyotype as well as by DNA profiling using the AMPFl STR Profiling Plus Kit [Applied Biosystems, Foster City, CA (47)]. While this work was in progress, ALVA-31 and PPC-1 were shown by our laboratory to be derivatives of PC-3 (48).

Tissue Samples.

Samples were snap frozen in liquid nitrogen immediately after procurement and stored there until harvest of RNA. Normal prostate tissue was obtained from organ donors of ages 1.5, 13, 15, 18, 20, 44, and 61 years. BPH specimens were obtained from patients undergoing radical prostatectomy. Lymph node metastases were obtained from two patients with primary prostate tumors. Highly epithelial areas, devoid of inflammation, were selected from all tissue used for whole tissue RNA extraction after examination of H&E-stained sections. Prostate tissue containing normal and malignant epithelium was obtained from two radical prostatectomy patients and one research organ donor.

LCM.

LCM was performed using a PixCell II System (Arcturus Engineering, Inc., Mountain View, CA) on 8-μm-thick frozen sections of normal prostate from one prepubertal (1.5 year), two pubertal (13 and 15 years), and two adult specimens (18 and 20 years) to separately capture epithelial and stromal cells for RNA isolation. Eight micrometer frozen sections were also used to capture the paired tumor/normal epithelium samples.

The cryostat blade was wiped with RNase Zap (Ambion, Austin, TX) before each specimen was cut, and solutions were made with RNase-free water. Tissue sections were fixed in 70% ethanol for 30 s, washed in water for 10 s, and then stained in Gill’s hematoxylin (Electron Microscopy Sciences, Fort Washington, PA) for 1–2 min. After another water wash, they were put in 1% ammonia for 30–60 s and then in 70% and 95% ethanol washes for 10 s each. After staining in eosin for 1–2 min, they were dehydrated in 95% and 100% ethanol washes for 10 s each, washed in xylene for 5–10 min, and then air dried for 60 min before capture of cells.

RNA Isolation.

Total RNA was isolated by the guanidinium isothiocyanate/acid phenol method of Chomczynski and Sacchi (49), with minor modifications.

Cultured cells were grown to 70–80% confluence before harvest. Briefly, cells were scraped from the flasks, washed in PBS, and then lysed in guanidinium isothiocyanate buffer. After extraction with acid-phenol, the aqueous layer was precipitated overnight with an equal volume of isopropanol. The precipitate was dissolved in guanidinium buffer and reprecipitated overnight with isopropanol. The pellet was washed twice with 75% ethanol and then dried briefly before resuspending in diethylpyrocarbonate-treated H2O.

Frozen tissue was weighed and manually pulverized before addition of the appropriate volume of guanidinium buffer. Samples were homogenized using a Polytron 2000 (Brinkmann Instruments Inc., Westbury, NY). RNA isolation was as performed for cell cultures with the addition of a second phenol extraction step. Analysis of RNA was done by 260/280 absorbance ratio and electrophoresis on a nondenaturing 1% agarose gel.

Laser-captured specimens were extracted with additional slight modifications. Two hundred μl of guanidinium extraction buffer containing 1.6 μl of β-mercaptoethanol per CapSure cap (Arcturus Engineering, Inc.) were used with 2-min tissue lysis times. Centrifugation times were 30 min each. One μl of glycogen (10 μg/μl in water) was added to each sample before precipitation in isopropanol at −20°C overnight followed by one 75% ethanol wash. RNA was redissolved in diethylpyrocarbonate-treated water and used for reverse transcription in volumes used for 1 μg of RNA.

RT and PCR.

RT was preceded by treatment with DNase I (Invitrogen) as per the manufacturer’s instructions. For each microgram of RNA, the volume was adjusted to 20 μl containing 2.5 μm random hexamers, dNTPs (1 mm each), 1 unit of RNasin (Promega, Madison, WI), and 1× RT buffer (Promega). A mock RT sample of 10 μl was removed, and 50 units of Moloney murine leukemia virus reverse transcriptase (Promega) were added to the actual RT tube. Samples were incubated at 22°C for 10 min, 42°C for 45 min, 95°C for 5 min, and 4°C for 5 min. The equivalent of 250 ng of starting RNA from the RT or mock sample were used for each degenerate PCR reaction. Degenerate PCR primers corresponded to the conserved peptide motifs ELEKEF and KIWFQN. Sequences of all primers used are listed in Table 1. Degenerate PCR reaction volumes were 25 μl each, containing the following: template or mock; 2 μm of each primer; 200 μm each dNTP; 4.5 mm MgCl2; and 0.625 unit of Taq polymerase (Promega) in 1× reaction buffer. Cycle number was 32 for the degenerate primers. PCR products were electrophoresed on gels of 2% NuSieve 3:1 agarose (BioWhittaker Molecular Applications, Rockland, ME), and the appropriate size band was cut out and extracted with QIAquick Gel Extraction Kit (Qiagen, Valencia, CA). All PCR assays included negative controls using mock RT templates as well as a water template sample. Only samples showing no PCR amplicons in the corresponding mock tubes were used for subsequent cloning and sequencing.

Amplification volumes for nondegenerate RT-PCR reactions were 25 μl containing template (equivalent to 100 ng of starting RNA), 200 μm each dNTP, 200 nm each primer, 0.625 unit of Taq polymerase (Promega), and 1.5–2.5 mm MgCl2 (optimized for each primer pair) in 1× reaction buffer (see Table 1). Specific HOXC reactions with tissue templates ran for 35 cycles, whereas those from cell line templates ran for 32 cycles. Reactions for β-actin were cycled 30 times, and reactions for PSA were cycled 32 times.

Cloning and Sequencing.

Purified PCR products were cloned using a TOPO TA Cloning Kit (Invitrogen) as per the manufacturer’s instructions. After overnight incubation of plates, colonies were selected for each sample and checked for presence of the correct-sized insert by direct PCR with the degenerate primers. Bacterial cultures were grown overnight, and plasmids were isolated using QIAwell 8 Ultra Kits (Qiagen). Sequencing was done by the University of Colorado Cancer Center DNA Sequencing Core using dye-terminator chemistry on ABI 373A and ABI 377 automated sequencing machines (Applied Biosystems). Resulting sequences were compared with the nonredundant nucleotide database and the SwissProt protein database of GenBank using the BLAST search algorithm (50).

RNA Amplification.

RNA from LCM-captured malignant and normal epithelial cells was amplified through two rounds using the RiboAmp RNA Amplification Kit (Arcturus Engineering, Inc.) as per the manufacturer’s instructions. Optional maximum times for the in vitro transcription incubations were used in both rounds to increase yield. Amplified RNA was subjected to an additional DNase I treatment before RT for RT-PCR.

FISH.

FISH analysis was performed by the Cytogenetics Core Laboratory of the University of Colorado Cancer Center. Cells were blocked in metaphase with Colcemid (0.05 μg/ml) for 1–2 h before hypotonic swelling in a 4:1 mixture of 0.075 m KCl and 1% sodium citrate. Fixation was done using a 3:1 mixture of methanol and glacial acetic acid. Two PAC clones that map to 12q12–14 were used as probes; one contains the complete HOXC cluster, and the other contains the vitamin D receptor gene, which is located proximal to the HOXC cluster. The probes were labeled using the Vysis nick translation kit (Vysis, Downers Grove, IL). The HOXC PAC was labeled with dUTP SpectrumGreen, and the vitamin D receptor PAC was labeled with dUTP-SpectrumRed. The prehybridization process consisted of cleaning the slides with 70% acetic acid and then rinsing them, followed by an ethanol dehydration series. Next the slides were digested in 0.008% pepsin/0.01 m HCl at 37°C for 10 min and fixed in 1% formaldehyde at room temperature for 10 min. Two hundred ng of each probe DNA, dissolved in 50% formamide hybridization mix, were added to the selected hybridization area, which was covered with a 177-mm2 glass coverslip. Probe and chromosomal DNA were codenatured for 8 min at 80°C and then hybridized at 37°C for 48–72 h. Approximately 13 metaphase cells were analyzed per cell line. Additionally, red and green signals were scored in 100 interphase nuclei. Analysis and imaging were performed using an Olympus BX-60 epifluorescence microscope coupled with SmartCapture software (Vysis).

Transfection/Luciferase Assays.

Sixteen h before transfection, 3 × 105 LNCaP cells were seeded into each well of a 6-well dish in medium containing charcoal-stripped serum. Plasmids pAHluc (51), pCMV6c (52), and pCMVHOXC8 were transfected into LNCaP cells using Lipofectamine 2000 (Invitrogen) according to the manufacturer’s instructions. The total amount of DNA per well was 3 μg, with pCMV6c being used as empty vector to normalize DNA amounts per transfection. The day after transfection, the cells were treated with 10 nm R1881 for 24 h, and luciferase assays were performed as described previously (53). Protein concentration of the extracts was determined by Bradford assay (Bio-Rad Laboratories, Hercules, CA) and used to normalize the luciferase activity of the corresponding extract. The plasmid pCMVHOXC8 was constructed by RT-PCR amplification of HOXC8 cDNA and cloning the amplicon into the expression vector pCMV6c. The insert was sequenced to affirm that no enzymatic errors had been introduced.

Western Analysis.

Total cellular extracts were made by resuspending the cell pellets in radioimmunoprecipitation assay buffer [150 mm NaCl, 1% NP40, 0.5% sodium deoxycholate, 0.1% SDS, 50 mm Tris (pH 7.5)] containing protease inhibitors (Protease Inhibitor Cocktail; Sigma-Aldrich, St. Louis, MO). After 10 min on ice, the extracts were cleared by microcentrifugation for 10 min, the supernatants were transferred to a new tube, and the protein concentration of each extract was determined by Bradford assay. Equivalent amounts of extract were separated on a 4–12% NuPAGE gel (Invitrogen) and transferred to polyvinylidene difluoride membrane. HOXC8 was detected with monoclonal antibody C953-7E (Covance Research Products, Berkeley, CA). Western blots were developed using enhanced chemiluminescence Western Lightning (Perkin-Elmer Life Sciences). Blots were stripped and reprobed for β-actin using monoclonal antibody AC-15 (Sigma).

Statistical Analysis.

Statistical analyses were performed using the Intercooled Stata 5.0 software package (Stata Corp., College Station, TX). χ2 and Fisher ’s exact test results were considered to be statistically significant when P ≤ 0.05.

Relative Expression of HOX Clusters in Benign and Malignant Prostate.

We used degenerate RT-PCR combined with cloning and sequencing to simultaneously detect expression of multiple HOX genes. This sampling method permits examination of relative expression levels within each specimen. A summary of the raw data collected is presented in Table 2. Divergence in the conserved sequences in the homeodomain prevented detection of eight of the HOX genes (HOXA11 , HOXA13 , HOXB13 , HOXC11 , HOXC12 , HOXC13 , HOXD11 , and HOXD13 ). Three HOX genes (HOXB1 , HOXB2 , and HOXD3 ) were not detected by the degenerate primers in any of these samples, despite retention of the conserved sequences and theoretical match with the primers. A total of 548 HOX inserts from benign sources and 342 HOX inserts from malignant sources were identified from cell lines, whole tissue, and LCM-captured epithelium and stroma.

We first examined overall expression patterns of the four HOX clusters (A–D) in the broad categories of benign and malignant epithelium (Fig. 1). Benign stroma, from laser capture-microdissected specimens, is shown for comparison because whole tissue specimens, by definition, contained some stromal components. However, the highly epithelial areas selected for extraction were approximately 2:1 epithelial to stromal cells. We noted that less total RNA was isolated from stromal cells and that overall HOX expression seemed lower from stromal samples (data not shown). We thus felt the overall contribution of stromal expression in whole tissue samples to be minimal. Additionally, when we examined epithelium and stroma separately, we found the overall expression patterns to be remarkably similar. The malignant epithelial pattern shows definite changes in the relative expression of the HOXA , HOXB , HOXC , and HOXD clusters when compared with the benign expression pattern (P < 0.001). Benign epithelial cells showed approximately 70% of the HOX genes expressed to be from the A cluster, with virtually no expression of C cluster genes. However, malignant cells expressed A and C cluster genes at approximately equal rates (roughly 40% of total). Malignant epithelium also showed a decrease in expression of the D cluster compared with benign epithelium (P < 0.001), and particularly HOXD10 (P < 0.05).

As can be seen in Fig. 2, the change in relative expression level of A and C cluster genes is not simply the result of substitution of HOXC paralogs for HOXA paralogs. The increased expression in the C cluster is concentrated in C4, C5, and C6, although C8, C9, and C10 are also expressed, whereas the main decrease in the A cluster appears to be in A9 and A10.

HOX Expression Patterns in Prostate Cell Lines.

Because the differences between benign and malignant HOX expression patterns were quite clear, in view of the heterogeneous nature of prostate cancer, we decided to examine subsets of the malignant cell lines. Based on originating tumor location, the malignant prostate cell lines could be divided into three groups. LNCaP and LAPC-4 are derived from lymph node metastases. 22Rv 1 and PC-346C are derived from primary prostate tumors xenografted into mice and then established as cell lines. PC-3 and DU 145 are both metastases from other sites, specifically bone and brain, respectively (because ALVA-31 and PPC-1 are clonal derivatives of PC-3 and not unique cell lines, we have excluded them from this analysis). For statistical analyses, we chose to group the cell lines into the three pairs mentioned above to compare expression patterns in lymph node metastasis-derived cell lines, primary tumor-derived cell lines, and cell lines from other metastatic sites. Fig. 3 shows expression of the four HOX clusters in these categories and in lymph node metastases. Interestingly, breaking out the malignant subgroups revealed that the patterns of relative expression of the HOX clusters demonstrated by lymph node metastases and prostate cell lines derived from lymph node metastases were very similar. In contrast, the relative expression of the HOXB and HOXC clusters in lymph node-derived cell lines differed significantly (P < 0.001) from that observed in cell lines derived from primary tumors (22Rv 1 and PC-346C) or other metastatic sites (PC-3 and DU 145).

Examination of expression patterns within each HOX cluster also revealed differences characteristic of the origin of the cell lines (Fig. 4). We focused our attention on the HOXC cluster, which was overexpressed by each of the malignant cell line subgroupings. When the expression pattern of individual HOXC genes was compared in lymph node metastases and cell lines derived from lymph node metastases, the patterns appeared very similar, particularly with respect to HOXC4 , HOXC5 , and HOXC6 (Fig. 5). In contrast, the pattern of genes expressed in primary tumor-derived cell lines or other metastases-derived cell lines exhibited low expression of HOXC4 , HOXC5 , and HOXC6 but stronger expression of HOXC9 or HOXC8 , respectively. The differences in expression of HOXC6 were highly significant between the lymph node metastases-derived cell lines and either primary tumor-derived or other metastases-derived cell lines (P < 0.001). Due to sample sizes, the smaller differences detected in HOXC8 and HOXC9 expression between all three subgroupings do not reach statistical significance. However, the similarities between HOX expression patterns in lymph node metastases and cell lines derived from lymph node metastases and the differences between these samples and primary tumor-derived or other metastases-derived cell lines suggest that these patterns relate to the origination site of the cell line.

Expression of Specific HOXC Genes.

Of 548 inserts examined by the degenerate technique from all benign sources (whole tissue, cultured cells, and LCM samples of both epithelium and stroma) only 4 were HOXC genes. Two of these were HOXC10 from the prepubertal stromal sample, and two were HOXC6 from the whole BPH sample. In contrast, 134 HOXC inserts were identified out of 342 inserts examined from malignant sources (malignant cell lines and lymph node metastases). The degenerate data provide information on the relative expression of HOX clusters and individual genes within a particular sample, based on sampling. Thus we can compare relative expression within a sample, but we do not have information regarding absolute HOX expression level either overall or for specific genes.

To confirm the implications of the results seen by degenerate RT-PCR, we performed RT-PCR for specific HOXC genes on several different benign and malignant human tissue samples (Fig. 6 A ). Little or no expression of HOXC4 , HOXC5 , or HOXC6 was detected in normal tissue, and very low HOXC8 expression was seen in two of the four normal samples. BPH samples showed very low level expression of HOXC4 , HOXC5 , and HOXC6 , whereas both lymph node metastasis samples expressed all four of the HOXC genes examined.

Specific RT-PCR on cell lines confirmed that HOXC products were overexpressed in malignant lines compared with cultures of normal epithelial and stromal cells (Fig. 6 B ). Except in the derivative cell lines ALVA-31 and PPC-1, HOXC4 , HOXC5 , HOXC6 , and HOXC8 were broadly expressed in the malignant prostate cell lines examined. The benign stromal cells did not express detectable levels of these HOXC genes, whereas PrECs show some HOXC4 , HOXC6 , and HOXC8 .

Some samples, which did not show HOXC expression by degenerate RT-PCR, did show faint products with specific primers. This was not totally unexpected, due to the increased sensitivity of RT-PCR using specific primers. Whereas the degenerate primers should amplify all matching HOX gene products, those with low representation might not be detected by our sampling unless much larger numbers of inserts were examined for each specimen. Pooling data from all of the malignant cell lines and the malignant lymph node tissue was designed to minimize this effect by increasing the total number of inserts examined for the malignant category and allowing comparison of benign and malignant expression.

We also confirmed expression of HOXC5 and HOXC8 on the cell lines and whole tissue samples using intron-spanning primers that gave products of 333 and 388 bp, respectively. Relative expression levels were similar to those shown in Fig. 6, A and B.

To further examine HOXC gene expression in malignancy, we performed LCM on frozen sections from biopsy punch specimens of three prostates to collect corresponding benign and malignant epithelium. Due to limited tumor availability, RNA from both tumor and normal cells was extracted and first amplified to provide sufficient template to allow examination of HOXC4 , HOXC5 , HOXC6 , HOXC8 , and β-actin in all specimens. This technique has been used for microarray assays and maintains differential mRNA expression (54), although not necessarily identical to unamplified levels (55). Fig. 6 C shows a representative gel showing the specific RT-PCR products. HOXC5 was expressed in all six samples. Two of the tumors showed increased expression over their normal component, and one showed decreased expression. The relatively strong expression in normal cells was somewhat unexpected, based on the degenerate data as well as on the absence of expression seen with specific primers on normal primary-culture cells and normal whole tissue templates. As a control, we then isolated RNA from 8-μm sections of normal tissue from two of the samples used for the tumor/normal pairs, and we performed RT-PCR using the HOXC5 primers without initial RNA amplification. A faint product could be detected in the sample with the strongest β-actin signal. Thus, it would appear the HOXC5 expression seen after two rounds of amplification may be enhanced but does not appear to be solely an artifact of amplification. One hypothesis is that the HOXC5 expression in these normal samples is related to the presence of tumor within the same prostate because the previously examined normal samples were taken from patients without tumors. A larger sample size designed specifically to address this question will be needed before drawing further conclusions. Neither HOXC6 nor HOXC8 was expressed in any of the normal specimens, and their absence despite RNA amplification appears to confirm lack of expression. One tumor sample expressed both HOXC5 and HOXC6 strongly, whereas the other two tumors expressed HOXC5 and HOXC8 . No expression of HOXC4 was seen from any of these tumor or normal cells (data not shown).

To test whether RNA expression levels of HOXC genes were reflected in protein expression, Western analysis was performed on extracts of the same prostate cell lines analyzed for HOX RNA using an antibody to a unique peptide in HOXC8. The cell lines showing the highest levels of expression of HOXC8 RNA (PC-3, DU 145, PC-346C, 22Rv 1, and LAPC-4) also showed the highest levels of HOXC8 protein expression (Fig. 6 D ). Lower to undetectable levels of HOXC8 were seen in LNCaP, ALVA-31, and PPC-1, consistent with the RT-PCR data.

Reciprocal Regulation of HOXC8 and Androgen Signaling.

Homeodomain proteins have been reported to inhibit the related transcriptional coactivators CBP and p300 by inhibiting their intrinsic histone acetyltransferase activity (56). CBP and p300 are two of the best-studied coactivators of the steroid receptor family of hormone-dependent transactivators (57, 58). Because CBP is an androgen receptor coactivator (59, 60), we reasoned that overexpression of HOXC genes might abrogate androgen-dependent gene activation. HOXC8 cDNA was cloned into an expression vector, and the vector was transfected into LNCAP cells along with an androgen-responsive luciferase reporter (Fig. 7 A ). Increasing expression of HOXC8 resulted in a progressive decrease of reporter gene induction by the synthetic androgen R1881. At the highest levels of HOXC8 expression, the hormone induction was completely abolished. Thus, overexpression of HOXC8 results in a major alteration in a critical signaling pathway in prostate cancer cells. Additional studies will investigate the influence of other members of the HOX family on androgen signaling.

Conversely, we examined the consequences of androgen signaling on HOXC8 gene expression. LNCaP cells in medium containing charcoal-stripped serum were treated with R1881 or hydroxyflutamide. After isolation of RNA, RT-PCR was used to assess levels of HOXC8 , HOXC5 , and PSA . Expression of PSA was induced as expected by R1881 and also by hydroxyflutamide (Fig. 7,B ). Although hydroxyflutamide is an androgen antagonist, it displays agonist activity with the mutated androgen receptor found in LNCaP cells (61). In contrast to PSA , both ligands inhibited expression of HOXC8 (Fig. 7 B ). HOXC5 expression showed no significant response to either ligand (data not shown). Thus, HOXC8 and androgens exhibit reciprocal inhibitory actions.

Analysis of Chromosomal Abnormalities in Prostate Cell Lines.

We performed FISH analysis on the prostate carcinoma cell lines to see whether there was a correlation between HOXC gene expression and chromosomal abnormalities such as amplification and/or translocation. No such correlation was found. Three lines (LNCaP, PC-346C, and 22Rv 1) displayed only normal copies of chromosome 12 equivalent to the modal ploidy for that line (tetraploid, diploid and mixed tetraploid/diploid, respectively). Two triploid lines (LAPC-4 and DU 145) both showed two normal copies of chromosome 12 plus one differing abnormal chromosome in each cell line exhibiting rearrangement outside the 12q12–14 region. The final three triploid cell lines (PC-3, ALVA-31, and PPC-1) showed complex rearranged karyotypes, as we have published previously (48). These lines exhibited from two to five different derivatives of chromosome 12. Interestingly, two of these cell lines (PPC-1 and ALVA-31) show very low HOXC gene expression, even though they are derived from the parental PC-3 line (47). Whether a connection exists between the rearranged chromosomes and the apparent decrease in HOXC expression seen in the clonal derivatives of PC-3 needs further study. The existence of small deletions or mutations within individual HOXC genes as potential contributing factors cannot be ruled out.

Comparison of expression patterns of benign and malignant prostate cells indicates overexpression of HOXC cluster genes in malignancy. RT-PCR using degenerate primers that detect 31 of the 39 HOX genes revealed that expression from the HOXC cluster was low compared with the other HOX clusters in benign cells. RT-PCR for several specific HOXC genes confirmed very low expression in benign primary cultures, benign whole tissue, and laser-captured normal tissue. Malignant cell lines and lymph node metastases demonstrate substantially higher expression of these genes. Because the HOXC genes we examined do not appear to be uniformly up-regulated but show varied expression levels by gene in each malignant sample, it appears that independent control mechanisms are operating rather than a single switch for the entire HOXC cluster.

A recent study also found little or no expression of HOXC8 in normal tissue by in situ hybridization (33). However, increased HOXC8 expression in tumors correlated with higher Gleason score (decreasing differentiation). These investigators also detected HOXC8 in PC-3 and DU 145 cells by RT-PCR but found no expression in LNCaP. We also find little or no expression of HOXC8 in LNCAP and more robust expression in PC-3 and DU 145. Moreover, we have detected expression of HOXC8 protein in both of the latter cell lines.

Detectable expression of HOXA9 and HOXD10 has been reported in normal prostate tissue by Northern blot (31). We found HOXA9 to be broadly expressed in almost all prostate samples, both benign and malignant, although individual samples showed some variation. We also found HOXD10 to be expressed rather widely in normal prostate cells, but not in malignant ones. Adult mouse prostate has been shown to express Hoxb13(62) and Hoxd13(63). Due to sequence divergence in the region of the degenerate primers, we do not have expression information for these genes in our human samples. Exploration of expression of the paralogous groups 11–13 will require a more individualized approach.

Mammalian downstream targets of HOX genes have proven difficult to identify. The mouse homologue of the Drosophila tumor suppressor l(2)gl , called mgl-1 , was identified as a target of Hoxc8, and a possible inverse relationship was postulated (64). Increasing evidence links various HOX genes to cell adhesion molecules. Hoxc6 and Hoxb9 have been reported to increase NCAM promoter activity, whereas Hoxb8 repressed this activity in NIH 3T3 fibroblasts (65, 66). HOXD9 was found to increase expression of the L-CAM enhancer in the same cell type (67). Human prostate tissue has been shown to contain NCAM-like molecules (68). Other investigators (69) have reported that HOXD3 down-regulates E-cadherin expression, up-regulates integrins α3 and β3, and produces de novo expression of N-cadherin and integrin α4 in HOXD3 -transfected A549 lung cancer cells. Mobility and invasion assays indicated increased activity in the transfected cells compared with the parental line or control transfectants. Transfected cells also produced a greater number of metastatic foci compared with the parental cells or control transfectants when injected into nude mice. Decreased E-cadherin expression in human prostate cancers has been correlated with the degree of differentiation of the tumor and the presence of metastases (70). Taken together, these studies indicate the possibility that dysregulation of HOX genes in the malignant prostate could play a role in metastasis by facilitating migration away from the original tissue through modulation of homeostatic cell adhesion molecules. Whether expression of specific cell adhesion molecules could also play a role in selecting metastatic sites remains to be demonstrated.

In view of the vastly different cellular environments between in vivo and in vitro conditions, particularly the lack of stromal cell components in vitro , the similarity in relative HOX cluster expression between lymph node metastases and lymph node metastasis-derived cell lines is quite remarkable. Several investigators (71, 72) have suggested the importance of stroma in prostatic tissue homeostasis as well as cancer progression. An interesting question is whether prostatic stromal signaling could be responsible for modulating epithelial HOX expression. The similarity in relative expression of the four HOX clusters, as well as genes within the C cluster, between lymph node metastases and lymph node metastases-derived cell lines, two extraprostatic environments both lacking prostatic stromal components, suggests this not to be the case. Alternatively, stromal influences may have permanently entrained a constitutive pattern of HOX gene expression. The observed similarities between cell lines and tumors provide supplemental evidence that some in vivo characteristics of cells can be maintained in culture, further validating the use of cell lines as specific investigative tools, especially for initial studies.

The complexity of expression patterns of HOX genes in prostate tissue means that much work remains to clarify the role played by these genes in prostate cancer. We have attempted to examine HOX expression simultaneously as an overall pattern rather than simply focusing on single genes. The lack of expression of HOXC genes in normal prostate and up-regulation in cultured malignant cell lines, lymph node metastases, and primary prostate tumors suggests a role in prostate cancer. No correlation between chromosomal abnormalities and HOXC expression was found. Similarity of expression patterns between lymph node metastases and cell lines derived from lymph node metastases suggests possible links between HOX expression and metastatic site. An intriguing speculation is whether the HOX expression patterns within a malignant cell could be instrumental in determining where that cell might lodge and metastasize. Whereas this first broad study of HOX gene expression in human prostate has exposed some intriguing clues, much more work remains to elucidate the role of HOX genes in prostatic malignancy.

One functional consequence of overexpression of HOXC genes was suggested by work reporting that homeodomain proteins inhibit the histone acetyltransferase activity of the transcriptional coactivator, CBP (56). In androgen-responsive LNCaP cells, we have demonstrated that increasing expression of HOXC8 progressively inhibited transactivation by the androgen receptor. It may seem paradoxical to suggest that a gene whose expression is associated with malignancy would inhibit a signaling pathway that is growth-promoting. However, if overexpression of HOXC occurs at a relatively early stage of prostate tumorigenesis, then the tumor must adapt to the diminished androgen signaling that accompanies expression of HOXC genes, thereby predisposing the tumor to survive in the face of a subsequent withdrawal of androgens. Thus, we postulate that the tumor would already be at least partially androgen resistant at the onset of ablation therapy, allowing some tumor cells to escape therapy and eventually progress. This speculation suggests an important role for HOXC in the pathogenesis of androgen-resistant prostate cancer.

Fig. 1.

Relative expression of the four HOX clusters (A–D) in benign and malignant cell types. Benign epithelium (ben.epi. ) includes PrECs, normal prostate tissue, and LCM epithelium. Benign stroma (ben.str. ) includes LCM stroma. Malignant epithelium (malig.epi. ) includes eight malignant cell lines and lymph node metastases. Error bars represent 95% confidence intervals for sample sizes. Differences in relative expression of the four HOX clusters between benign and malignant epithelium are significant (P < 0.001).

Fig. 1.

Relative expression of the four HOX clusters (A–D) in benign and malignant cell types. Benign epithelium (ben.epi. ) includes PrECs, normal prostate tissue, and LCM epithelium. Benign stroma (ben.str. ) includes LCM stroma. Malignant epithelium (malig.epi. ) includes eight malignant cell lines and lymph node metastases. Error bars represent 95% confidence intervals for sample sizes. Differences in relative expression of the four HOX clusters between benign and malignant epithelium are significant (P < 0.001).

Close modal
Fig. 2.

Relative gene expression within the four HOX clusters in benign and malignant cells. Benign epithelium includes PrECs, normal prostate tissue, and LCM epithelium. Malignant epithelium includes eight malignant cell lines and lymph node metastases. Percentages are based on total HOX genes identified within each sample category.

Fig. 2.

Relative gene expression within the four HOX clusters in benign and malignant cells. Benign epithelium includes PrECs, normal prostate tissue, and LCM epithelium. Malignant epithelium includes eight malignant cell lines and lymph node metastases. Percentages are based on total HOX genes identified within each sample category.

Close modal
Fig. 3.

Relative expression levels of the four HOX clusters in malignant cells. l.n.mets, lymph node metastases. Percentages are based on total for each category.

Fig. 3.

Relative expression levels of the four HOX clusters in malignant cells. l.n.mets, lymph node metastases. Percentages are based on total for each category.

Close modal
Fig. 4.

Relative expression of genes within the four HOX clusters by malignant cell line groups. Differences in expression of HOXC6 between lymph node-derived cell lines and primary tumor-derived or other metastases-derived cell lines are statistically significant (P < 0.001).

Fig. 4.

Relative expression of genes within the four HOX clusters by malignant cell line groups. Differences in expression of HOXC6 between lymph node-derived cell lines and primary tumor-derived or other metastases-derived cell lines are statistically significant (P < 0.001).

Close modal
Fig. 5.

Comparison of relative expression of HOXC genes in malignant cells. benign , PrECs, normal donor prostate tissue, and LCM epithelium and stroma; l.n.mets , lymph node metastases; other categories are malignant prostate cell lines.

Fig. 5.

Comparison of relative expression of HOXC genes in malignant cells. benign , PrECs, normal donor prostate tissue, and LCM epithelium and stroma; l.n.mets , lymph node metastases; other categories are malignant prostate cell lines.

Close modal
Fig. 6.

M, molecular weight marker lane. A, RT-PCR analysis of specific HOXC genes from tissue samples. Normal 1–4, donor prostate tissues; BPH 1 and 2, epithelial BPH nodules; l.n.met 1 and 2, lymph node metastases. B, RT-PCR analysis of specific HOXC genes from cell line samples. STR, primary stromal culture; other names refer to specific malignant prostate cell lines. C, RT-PCR analysis of specific HOXC genes from tumor/normal epithelial cells selected by LCM. T, tumor; N, normal. Numbers indicate paired material from the same patient. D, Western analysis for HOXC8 protein expression in the same cell lines used in B .

Fig. 6.

M, molecular weight marker lane. A, RT-PCR analysis of specific HOXC genes from tissue samples. Normal 1–4, donor prostate tissues; BPH 1 and 2, epithelial BPH nodules; l.n.met 1 and 2, lymph node metastases. B, RT-PCR analysis of specific HOXC genes from cell line samples. STR, primary stromal culture; other names refer to specific malignant prostate cell lines. C, RT-PCR analysis of specific HOXC genes from tumor/normal epithelial cells selected by LCM. T, tumor; N, normal. Numbers indicate paired material from the same patient. D, Western analysis for HOXC8 protein expression in the same cell lines used in B .

Close modal
Fig. 7.

Mutual inhibition by HOXC8 and androgen receptor in prostate cancer cells. A, LNCaP cells were transfected with increasing levels of an expression vector for HOXC8 along with an androgen-responsive mouse mammary tumor virus-luciferase reporter, pAHluc. After treatment with the synthetic androgen R1881 for 24 h, cells were harvested for assay of luciferase activity. A representative of three identical experiments is shown. Error bars represent the SE of triplicate wells. B, LNCAP cells were treated with 10 nm R1881 or 10 μm hydroxyflutamide (OH-FLU ) for 24 h and then harvested for RT-PCR to analyze gene expression of PSA and HOXC8 . M, molecular weight marker lane.

Fig. 7.

Mutual inhibition by HOXC8 and androgen receptor in prostate cancer cells. A, LNCaP cells were transfected with increasing levels of an expression vector for HOXC8 along with an androgen-responsive mouse mammary tumor virus-luciferase reporter, pAHluc. After treatment with the synthetic androgen R1881 for 24 h, cells were harvested for assay of luciferase activity. A representative of three identical experiments is shown. Error bars represent the SE of triplicate wells. B, LNCAP cells were treated with 10 nm R1881 or 10 μm hydroxyflutamide (OH-FLU ) for 24 h and then harvested for RT-PCR to analyze gene expression of PSA and HOXC8 . M, molecular weight marker lane.

Close modal

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

1

This work was supported in part by NIH Grant CA84269 and United States Army Medical Research Grant DAMD17-00-1-0047 as well as NIH Grant CA46934 to the University of Colorado Cancer Center. This work is dedicated to Gary J. Miller, valued colleague, friend and mentor, beloved husband and father. His contributions toward the diagnosis and eradication of prostate cancer are severely missed.

5

The abbreviations used are: BPH, benign prostatic hyperplasia; FBS, fetal bovine serum; FISH, fluorescence in situ hybridization; LCM, laser capture microdissection; PrEC, benign prostate epithelial cell; RT, reverse transcription; RT-PCR, reverse transcription-PCR; PSA, prostate-specific antigen; dNTP, deoxynucleotide triphosphate; PAC, P1-derived artificial chromosome; CBP, cAMP-responsive element binding protein (CREB)-binding protein.

Table 1

Primer sequences and PCR cycling conditions

GeneSequencePCR conditionsProduct size (bp)
Degenerate HOX Forward: 5′-gctctagarytngaraargartt-3′ 94°C-1 min, 40°C-1 min, 72°C-1 min 125 
 Reverse: 5′-ggaattcrttytgraaccadatytt-3′   
HOXC4 Forward: 5′-caccaccccgagaaatcaca-3′ 94°C-30 s, 55°C-30 s, 72°C-30 s 187 
 Reverse: 5′-aattggggtttcaccgtgctaac-3′   
HOXC5 Forward: 5′-cagcctgctagctagctcaact-3′ 94°C-30 s, 56°C-30 s, 72°C-30 s 159 
 Reverse: 5′-caggaaggacccagagtcaata-3′   
HOXC6 Forward: 5′-ctgagcagggcaggact-3′ 94°C-30 s, 58°C-30 s, 72°C-1 min 186 
 Reverse: 5′-gcgttaggtagcgattgaag-3′   
HOXC8 Forward: 5′-cagtggacggcaaacttac-3′ 94°C-30 s, 53°C-30 s, 72°C-30 s 170 
 Reverse: 5′-cctttttccacttcatccttc-3′   
β-Actin Forward: 5′-atctggcaccacaccttctacaatgagctgcg-3′ 94°C-45 s, 60°C-45 s, 72°C-2 min 838 
 Reverse: 5′-cgtcatactcctgcttgctgatccacatctgc-3′   
Alternate β-actin Forward: 5′-acacaggggaggtgatagcatt-3′ 94°C-30 s, 52°C-30 s, 72°C-30 s 150 
 Reverse: 5′-atctcaagttgggggacaaaaa-3′   
PSA Forward: 5′-gcctctcgtggcagggcagt-3′ 94°C-1 min, 60°C-1 min, 72°C-1 min 440 
 Reverse: 5′-gggtgaacttgcgcacacac-3′   
GeneSequencePCR conditionsProduct size (bp)
Degenerate HOX Forward: 5′-gctctagarytngaraargartt-3′ 94°C-1 min, 40°C-1 min, 72°C-1 min 125 
 Reverse: 5′-ggaattcrttytgraaccadatytt-3′   
HOXC4 Forward: 5′-caccaccccgagaaatcaca-3′ 94°C-30 s, 55°C-30 s, 72°C-30 s 187 
 Reverse: 5′-aattggggtttcaccgtgctaac-3′   
HOXC5 Forward: 5′-cagcctgctagctagctcaact-3′ 94°C-30 s, 56°C-30 s, 72°C-30 s 159 
 Reverse: 5′-caggaaggacccagagtcaata-3′   
HOXC6 Forward: 5′-ctgagcagggcaggact-3′ 94°C-30 s, 58°C-30 s, 72°C-1 min 186 
 Reverse: 5′-gcgttaggtagcgattgaag-3′   
HOXC8 Forward: 5′-cagtggacggcaaacttac-3′ 94°C-30 s, 53°C-30 s, 72°C-30 s 170 
 Reverse: 5′-cctttttccacttcatccttc-3′   
β-Actin Forward: 5′-atctggcaccacaccttctacaatgagctgcg-3′ 94°C-45 s, 60°C-45 s, 72°C-2 min 838 
 Reverse: 5′-cgtcatactcctgcttgctgatccacatctgc-3′   
Alternate β-actin Forward: 5′-acacaggggaggtgatagcatt-3′ 94°C-30 s, 52°C-30 s, 72°C-30 s 150 
 Reverse: 5′-atctcaagttgggggacaaaaa-3′   
PSA Forward: 5′-gcctctcgtggcagggcagt-3′ 94°C-1 min, 60°C-1 min, 72°C-1 min 440 
 Reverse: 5′-gggtgaacttgcgcacacac-3′   
Table 2

Number of various HOX genes identified from different sample types

HOX geneBenignMalignant
Cell linesLCM tissueaWhole tissueCell linesbWhole tissue
n = 3c Epithelialdn = 1 Stromalen = 5 Epithelialn = 5 Stromaln = 5 Normal donorn = 8 Epithelialn = 2 l.n. metsf
  A1      
  A2       
  A3 11 13 
  A4  
  A5    
  A6 
  A7 24 14 
  A9 36 52 37 27 61 
  A10 38 32 28 35 
        
  B1        
  B2        
  B3 14 10 
  B4       
  B5       
  B6    17 
  B7    11 
  B8       
  B9   14 
        
  C4      18 
  C5      12 
  C6     51 13 
  C8      10  
  C9      10  
  C10     
        
  D1    10 
  D3        
  D4    
  D8   
  D9   
  D10 18 18 23    
  D12      
        
Total HOX 102 46 141 146 113 293 49 
HOX geneBenignMalignant
Cell linesLCM tissueaWhole tissueCell linesbWhole tissue
n = 3c Epithelialdn = 1 Stromalen = 5 Epithelialn = 5 Stromaln = 5 Normal donorn = 8 Epithelialn = 2 l.n. metsf
  A1      
  A2       
  A3 11 13 
  A4  
  A5    
  A6 
  A7 24 14 
  A9 36 52 37 27 61 
  A10 38 32 28 35 
        
  B1        
  B2        
  B3 14 10 
  B4       
  B5       
  B6    17 
  B7    11 
  B8       
  B9   14 
        
  C4      18 
  C5      12 
  C6     51 13 
  C8      10  
  C9      10  
  C10     
        
  D1    10 
  D3        
  D4    
  D8   
  D9   
  D10 18 18 23    
  D12      
        
Total HOX 102 46 141 146 113 293 49 
a

Paired epithelial and stromal cells from prepubertal, pubertal, and adult prostates.

b

Includes LNCaP, LAPC-4, DU 145, PC-3, ALVA-31, PPC-1, 22Rv 1, and PC-346C.

c

n, number of cell lines or patients.

d

PrEC.

e

Primary stromal culture.

f

l.n. mets, lymph node metastases.

We thank Drs. Johanna Nardi-Korver and Francisco G. La Rosa for assistance with LCM and Dr. Scott Stewart for assistance with statistical analysis.

1
Jemal A., Murray T., Samuels A., Ghafoor A., Ward E., Thun M. J. Cancer statistics, 2003.
CA Cancer J. Clin.
,
53
:
5
-26,  
2003
.
2
Sakr W. A., Haas G. P., Cassin B. F., Pontes J. E., Crissman J. D. The frequency of carcinoma and intraepithelial neoplasia of the prostate in young male patients.
J. Urol.
,
150
:
379
-385,  
1993
.
3
Guileyardo J. M., Johnson W. D., Welsh R. A., Akazaki K., Correa P. Prevalence of latent prostate carcinoma in two U.S. populations.
J. Natl. Cancer Inst. (Bethesda)
,
65
:
311
-316,  
1980
.
4
Cillo C., Cantile M., Faiella A., Boncinelli E. Homeobox genes in normal and malignant cells.
J. Cell. Physiol.
,
188
:
161
-169,  
2001
.
5
Abate-Shen C. Deregulated Homeobox gene expression in cancer: cause or consequence?.
Nat. Rev. Cancer
,
2
:
777
-785,  
2002
.
6
Owens B. M., Hawley R. G. HOX and non-HOX homeobox genes in leukemic hematopoiesis.
Stem Cells
,
20
:
364
-379,  
2002
.
7
Lawrence H. J., Stage K. M., Mathews C. H. W., Detmer K., Scibienski R., MacKenzie M., Migliaccio E., Boncinelli E., Largman C. Expression of HOXC homeobox genes in lymphoid cells.
Cell Growth Differ.
,
4
:
665
-669,  
1993
.
8
Celetti A., Barba P., Cillo C., Rotoli B., Boncinelli E., Magli M. C. Characteristic patterns of HOX gene expression in different types of human leukemia.
Int. J. Cancer
,
53
:
237
-244,  
1993
.
9
Lawrence H. J., Sauvageau G., Ahmadi N., Lopez A. R., LeBeau M. M., Link M., Humphries K., Largman C. Stage-and lineage-specific expression of the HOXA10 homeobox gene in normal and leukemic hematopoietic cells.
Exp. Hematol.
,
23
:
1160
-1166,  
1995
.
10
Perkins A., Kongsuwan K., Visvader J., Adams J. M., Cory S. Homeobox gene expression plus autocrine growth factor production elicits myeloid leukemia.
Proc. Natl. Acad. Sci. USA
,
87
:
8398
-8402,  
1990
.
11
Borrow J., Shearman A. M., Stanton V. P., Jr., Becher R., Collins T., Williams H. J., Dube I., Katz F., Kwong Y. L., Morris C., Ohyashiki K., Toyama K., Rowley J., Housman D. E. The t(7;11)(p15;p15) translocation in acute myeloid leukemia fuses the genes for nucleoporin NUP98 and class I homeoprotein HOXA9.
Nat. Genet.
,
12
:
159
-167,  
1996
.
12
DeVita G., Barba P., Odartchenki N., Givel J-C., Freschi G., Bucciarelli G., Magli M., Boncinelli E., Cillo C. Expression of homeobox-containing genes in primary and metastatic colorectal cancer.
Eur. J. Cancer
,
29A
:
887
-893,  
1993
.
13
Chariot A., Castronovo V. Detection of HOXA1 expression in human breast cancer.
Biochem. Biophys. Res. Commun.
,
222
:
292
-297,  
1996
.
14
Chariot A., Castronovo V., Le P., Gillet C., Sobel M. E., Gielen J. Cloning and expression of a new HOXC6 transcript encoding a repressing protein.
Biochem. J.
,
319
:
91
-97,  
1996
.
15
Cillo C., Barba P., Freschi G., Bucciarelli G., Magli M. C., Boncinelli E. HOX gene expression in normal and neoplastic human kidney.
Int. J. Cancer
,
51
:
892
-897,  
1992
.
16
Caré A., Silvani A., Meccia E., Mattia G., Stoppacciaro A., Parmiani G., Peschle C., Colombo M. P. HOXB7 constitutively activates basic fibroblast growth factor in melanomas.
Mol. Cell. Biol.
,
16
:
4842
-4851,  
1996
.
17
Rieger E., Bijl J. J., van Oostveen J. W., Soyer H. P., Oudejans C. B., Jiwa N. M., Walboomers J. M., Meijer C. J. Expression of the homeobox gene HOXC4 in keratinocytes of normal skin and epithelial skin tumors is correlated with differentiation.
J. Investig. Dermatol.
,
103
:
341
-346,  
1994
.
18
Nakamura T., Largaespada D. A., Lee M. P., Johnson L. A., Ohyashiki K., Toyama K., Chen S. J., Willman C. L., Chen I. M., Feinberg A. P., Jenkins N. A., Copeland N. G., Shaughnessy J. D., Jr. Fusion of the nucleoporin gene NUP 98 to HOXA9 by the chromosome translocation t(7;11)(p15;p15) in human myeloid leukemia.
Nat. Genet.
,
12
:
154
-158,  
1996
.
19
Fujino T., Suzuki A., Ito Y., Ohyashiki K., Hatano Y., Miura I., Nakamura T. Single-translocation and double-chimeric transcripts: detection of NUP98-HOXA9 in myeloid leukemias with HOXA11 or HOXA13 breaks of the chromosomal translocation t(7;11)(p15;p15).
Blood
,
99
:
1428
-1433,  
2002
.
20
Arai Y., Kyo T., Miwa H., Arai K., Kamada N., Kita K., Ohki M. Heterogenous fusion transcripts involving the NUP 98 gene and HOXD13 gene activation in a case of acute myeloid leukemia with the t(2;11)(q31;p15) translocation.
Leukemia (Baltimore)
,
14
:
1621
-1629,  
2000
.
21
Taketani T., Taki T., Shibuya N., Kikuchi A., Hanada R., Hayashi Y. Novel NUP98-HOXC11 fusion gene resulted from a chromosomal break within exon 1 of HOXC11 in acute myeloid leukemia with t(11;12)(p15;q13).
Cancer Res.
,
62
:
4571
-4574,  
2002
.
22
Panagopoulos I., Isaksson M., Billström R., Strömbeck B., Mitelman F., Johansson B. Fusion of the NUP 98 gene and the homeobox gene HOXC13 in acute myeloid leukemia with t(11;12)(p15;q13).
Genes Chromosomes Cancer
,
36
:
107
-112,  
2003
.
23
Scott M. P. Vertebrate homeobox gene nomenclature.
Cell
,
71
:
551
-553,  
1992
.
24
Ruddle F. H., Bartels J. L., Bentley K. L., Kappen C., Murtha M. T., Pettaway C. A. Evolution of HOX genes.
Annu. Rev. Genet.
,
28
:
423
-442,  
1994
.
25
Maconochie M., Nonchev S., Morrison A., Krumlauf R. Paralogous HOX genes: function and regulation.
Annu. Rev. Genet.
,
30
:
529
-556,  
1996
.
26
Greer J. M., Puetz J., Thomas K. R., Capecchi M. R. Maintenance of functional equivalence during paralogous Hox gene evolution.
Nature (Lond.)
,
403
:
661
-665,  
2000
.
27
Chen F., Capecchi M. R. Paralogous mouse Hox genes, Hoxa69 , Hoxb9 , and Hoxd9 , function together to control development of the mammary gland in response to pregnancy.
Proc. Natl. Acad. Sci. USA
,
96
:
541
-546,  
1999
.
28
Zhao Y., Potter S. S. Functional comparison of the Hoxa 4, Hoxa 10 , and Hoxa 11 homeoboxes.
Dev. Biol.
,
244
:
21
-36,  
2002
.
29
van den Akker E., Fromental-Ramain C., de Graaff W., Le Mouellic H., Brûlet P., Chambon P., Deschamps J. Axial skeletal patterning in mice lacking all paralogous group 8 Hox genes.
Development (Camb.)
,
128
:
1911
-1921,  
2001
.
30
Chauvet S., Merabet S., Bilder D., Scott M. P., Pradel J., Graba Y. Distinct Hox protein sequences determine specificity in different tissues.
Proc. Natl. Acad. Sci. USA
,
97
:
4064
-4069,  
2000
.
31
Redline R. W., Hudock P., MacFee M., Patterson P. Expression of AbdB-type homeobox genes in human tumors.
Lab. Investig.
,
71
:
663
-670,  
1994
.
32
Robbins S. E., Shu W-P., Kirschenbaum A., Levine A. C., Miniati D. N., Liu B. C-S. Bone extracellular matrix induces homeobox proteins independent of androgens: possible mechanism for androgen-independent growth in human prostate cancer cells.
Prostate
,
29
:
362
-370,  
1996
.
33
Waltregny D., Alami Y., Clausse N., de Leval J., Castronovo V. Overexpression of the homeobox gene HOXC8 in human prostate cancer correlates with loss of tumor differentiation.
Prostate
,
50
:
162
-169,  
2002
.
34
Dong J-T. Chromosomal deletions and tumor suppressor genes in prostate cancer.
Cancer Metastasis Rev.
,
20
:
173
-193,  
2001
.
35
Bhatia-Gaur R., Donjacour A., Sciavolino P., Kim M., Desai N., Young P., Norton C., Gridley T., Cardiff R., Cunha G., Abate-Shen C., Shen M. Roles for NKX3.1 in prostate development and cancer.
Genes Dev.
,
13
:
966
-977,  
1999
.
36
Kim M. J., Bhatia-Gaur R., Banach-Petrosky W. A., Desai N., Wang Y., Hayward S. W., Cunha G., Cardiff R., Shen M. M., Abate-Shen C. Nkx3.1 mutant mice recapitulate early stages of prostate carcinogenesis.
Cancer Res.
,
62
:
2999
-3004,  
2002
.
37
Abdulkadir S. A., Magee J. A., Peters T. J., Kaleem Z., Naughton C. K., Humphrey P. A., Milbrand J. Conditional loss of Nkx3.1 in adult mice induces prostatic intraepithelial neoplasia.
Mol. Cell. Biol.
,
22
:
1495
-1503,  
2002
.
38
Klein K. A., Reiter R. E., Redula J., Moradi H., Zhu X. L., Brothman A. R., Lamb D. J., Marcelli M., Belldegrun A., Witte O. N., Sawyers C. L. Progression of metastatic human prostate cancer to androgen independence in immunodeficient SCID mice.
Nat. Med.
,
3
:
402
-408,  
1997
.
39
Horoszewicz J. S., Leong S. S., Chu T. M., Wajsman Z. L., Friedman M., Papsidero L., Kim U., Chai L. S., Kakati S., Arya S. K., Sandberg A. A. The LNCaP cell line: a new model for studies on human prostatic carcinoma.
Prog. Clin. Biol. Res.
,
37
:
115
-132,  
1980
.
40
Horoszewicz J. S., Leong S. S., Kawinski E., Karr J. P., Rosenthal H., Chu T. M., Mirand E. A., Murphy G. P. LNCaP model of human prostatic carcinoma.
Cancer Res.
,
43
:
1809
-1818,  
1983
.
41
Loop S. M., Rozanski T. A., Ostenson R. C. Human primary prostate tumor cell line, ALVA-31: a new model for studying the hormonal regulation of prostate tumor cell growth.
Prostate
,
22
:
93
-108,  
1993
.
42
Brothman A. R., Lesho L. J., Somers K. D., Wright G. L., Merchant D. J. Phenotypic and cytogenetic characterization of a cell line derived from primary prostatic carcinoma.
Int. J. Cancer
,
44
:
898
-903,  
1989
.
43
Sramkoski R. M., Pretlow T. G., II, Giaconia J. M., Pretlow T. P., Schwartz S., Sy M. S., Marengo S. R., Rhim J. S., Zhang D., Jacobberger J. W. A new human prostate carcinoma cell line, 22Rv 1.
In Vitro Cell. Dev. Biol. Anim.
,
35
:
403
-409,  
1999
.
44
Romijn J. C., Erkens-Schulze S., Schröder F. H. Perspectives of the use of tissue culture methods as an alternative to human prostate cancer xenografts in nude mice Arnold W. Köpf-Maier P. Micheel B. eds. .
Immunodeficient Animals: Models for Cancer Res
,
51
:
209
-213, Karger Basel  
1996
.
45
Kaighn M. E., Narayan K. S., Ohnuki Y., Lechner J. F., Jones L. W. Establishment and characterization of a human prostatic carcinoma cell line (PC-3).
Investig. Urol.
,
17
:
16
-23,  
1979
.
46
Stone K. R., Mickey D. D., Wunderli H., Mickey G. H., Paulson D. F. Isolation of a human prostate carcinoma cell line (DU 145).
Int. J. Cancer
,
21
:
274
-281,  
1978
.
47
van Bokhoven A., Varella-Garcia M., Korch C., Hessels D., Miller G. J. Widely used prostate carcinoma cell lines share common origins.
Prostate
,
47
:
36
-51,  
2001
.
48
Varella-Garcia M., Boomer T., Miller G. J. Karyotypic similarity identified by multiplex-FISH relates four prostate adenocarcinoma cell lines: PC-3, PPC-1, ALVA-31, and ALVA-41.
Genes Chromosomes Cancer
,
31
:
303
-315,  
2001
.
49
Chomczynski P., Sacchi N. Single-step method of RNA isolation by guanidinium thiocyanate-phenol-chloroform extraction.
Anal. Biochem.
,
162
:
156
-159,  
1987
.
50
Altschul S. F., Madden T. L., Schäffer A. A., Zhang J., Zhang Z., Miller W., Lipman D. J. Gapped BLAST and PSI-BLAST: a new generation of protein database search programs.
Nucleic Acids Res.
,
25
:
3389
-3402,  
1997
.
51
Nordeen S. K., Moyer M. L., Bona B. J. Modulation of glucocorticoid-regulated transcription by purines: novel characteristics and implications for tissue specificity of steroid responses.
Endocrinology
,
136
:
1120
-1127,  
1995
.
52
Chapman B. S., Thayer R. M., Vincent K. A., Haigwood N. L. Effect of intron A from human cytogmegalovirus (Towne) immediate-early gene on heterologous expression in mammalian cells.
Nucleic Acids Res.
,
19
:
3979
-3986,  
1991
.
53
Nordeen S. K., Bona B. J., Moyer M. L. Latent agonist activity of the steroid antagonist, RU486, is unmasked in cells treated with activators of protein kinase A.
Mol. Endocrinol.
,
7
:
731
-742,  
1993
.
54
Phillips J., Eberwine J. H. Antisense RNA amplification: a linear amplification method for analyzing the mRNA population from single living cells.
Methods
,
10
:
283
-288,  
1996
.
55
Luzzi V., Hotschlag V., Watson M. A. Expression profiling of ductal carcinoma in situ by laser capture microdissection and high-density oligonucleotide arrays.
Am. J. Pathol.
,
158
:
2005
-2010,  
2001
.
56
Shen W., Krishnan K., Lawrence H. J., Largman C. The HOX homedomain proteins block CBP histone acetyltransferase activity.
Mol. Cell. Biol.
,
21
:
7509
-7522,  
2001
.
57
McKenna N. J., Xu J., Nawaz Z., Tsai M. J., O’Malley B. W. Nuclear receptor coactivators: multiple enzymes, multiple complexes, multiple functions.
J. Steroid Biochem. Mol. Biol.
,
69
:
3
-12,  
1999
.
58
Dilworth F. J., Chambon P. Nuclear receptors coordinate the activities of chromatin remodeling complexes and coactivators to facilitate initiation of transcription.
Oncogene
,
20
:
3047
-3054,  
2001
.
59
Aarnisalo P., Palvimo J. J., Janne O. A. CREB-binding protein in androgen receptor-mediated signaling.
Proc. Natl. Acad. Sci. USA
,
95
:
2122
-2127,  
1998
.
60
Fronsdal K., Engedal N., Slagsvold T., Saatcioglu F. CREB binding protein is a coactivator for the androgen receptor and mediates cross-talk with AP-1.
J. Biol. Chem.
,
273
:
31853
-31859,  
1998
.
61
Berrevoets C. A., Veldscholte J., Mulder E. Effects of antiandrogens on transformation and transcription activation of wild-type and mutated (LNCaP) androgen receptors.
J. Steroid Biochem. Mol. Biol.
,
46
:
731
-736,  
1993
.
62
Sreenath T., Orosz A., Fujita K., Bieberich C. J. Androgen-independent expression of hoxb-13 in the mouse prostate.
Prostate
,
41
:
203
-207,  
1999
.
63
Oefelein M., Chin-Chance C., Bushman W. Expression of the homeotic gene Hox-d13 in the developing and adult mouse prostate.
J. Urol.
,
155
:
342
-346,  
1996
.
64
Tomotsune D., Shoji H., Wakamatsu Y., Kondoh H., Takahashi N. A mouse homologue of the Drosophila tumour-suppressor gene l(2)gl controlled by Hox-c8 in vivo.
Nature (Lond.)
,
365
:
69
-72,  
1993
.
65
Jones F. S., Prediger E. A., Bittner D. A., De Robertis E. D., Edelman G. M. Cell adhesion molecules as targets for Hox genes: neural cell adhesion molecule promoter activity is modulated by cotransfection with Hox-2.5 and -2.4..
Proc. Natl. Acad. Sci. USA
,
89
:
2086
-2090,  
1992
.
66
Jones F. S., Holst B. D., Minowa O., De Robertis E. D., Edelman G. M. Binding and transcriptional activation of the promoter for the neural cell adhesion molecule by HoxC6 (Hox-3.3 ).
Proc. Natl. Acad. Sci. USA
,
90
:
6557
-6561,  
1993
.
67
Goomer R. S., Holst B. D., Wood I. C., Jones F. S., Edelman G. M. Regulation in vitro of an L-CAM enhancer by homeobox genes HoxD9 and HNF-1.
Proc. Natl. Acad. Sci. USA
,
91
:
7985
-7989,  
1994
.
68
Lipford G. B., Wright G. L., Jr. Comparative study of monoclonal antibodies TURP-27 and HNK-1: their relationship to neural cell adhesion molecules and prostate tumor-associated antigens.
Cancer Res.
,
51
:
2296
-2301,  
1991
.
69
Hamada J-I., Omatsu T., Okada F., Furuuchi K., Okubo Y., Takahashi Y., Tada M., Miyazaki Y. J., Taniguchi Y., Shirato H., Miyasaka K., Moriuchi T. Overexpression of homeobox gene HOXD3 induces coordinate expression of metastasis-related genes in human lung cancer cells.
Int. J. Cancer
,
93
:
516
-525,  
2001
.
70
Umbas R., Isaacs W. B., Bringuier P. P., Schaafsma H. E., Karthaus H. F. M., Oosterhof G. O. N., Debruyne F. M. J., Schalken J. A. Decreased E-cadherin expression is associated with poor prognosis in patients with prostate cancer.
Cancer Res.
,
54
:
3929
-3933,  
1994
.
71
Hayward S. W., Cunha G. R., Dahiya R. Normal development and carcinogenesis of the prostate.
Ann. N. Y. Acad. Sci.
,
784
:
50
-62,  
1996
.
72
Tuxhorn J. A., Ayala G. E., Rowley D. R. Reactive stroma in prostate cancer progression.
J. Urol.
,
166
:
2472
-2483,  
2001
.