Molecular Cloning and Characterization of STAMP1 , a Highly Prostate-specific Six Transmembrane Protein that Is Overexpressed in Prostate Cancer*

We have identified a novel gene, six transmembrane protein of prostate 1 ( STAMP1 ), which is largely specific to prostate for expression and is predicted to code for a 490-amino acid six transmembrane protein. Using a form of STAMP1 labeled with green fluorescent protein in quantitative time-lapse and immunofluorescence confocal microscopy, we show that STAMP1 is localized to the Golgi complex, predominantly to the trans -Golgi network, and to the plasma membrane. STAMP1 also localizes to vesicular tubular structures in the cytosol and colocalizes with the early endosome antigen 1 (EEA1), suggesting that it may be involved in the secre-tory/endocytic pathways. STAMP1 is highly expressed in the androgen-sensitive, androgen receptor-positive prostate cancer cell line LNCaP, but not in androgen receptor-negative prostate cancer cell lines PC-3 and DU145. Furthermore, STAMP1 expression is significantly lower in the androgen-dependent human prostate xenograft CWR22 compared with the relapsed de-rivative CWR22R, suggesting that its expression may be deregulated during prostate cancer progression. Cloning and Construction— 262-bp cDNA fragment originally obtained from a screen of a prostate-specific (16) and termed L74. RACE (5 (cid:1) -rapid amplification of cDNA ends) was per- formed (oligonucleotide sequences available upon request) using the Marathon-Ready cDNA that was prepared from normal prostate tissue (CLONTECH) and/or SMART-RACE LNCaP cDNA library (CLONTECH) that was generated according to the manufacturer’s recommendations. RACE products were cloned into pCRII-TOPO (Invitrogen), and positive clones were confirmed by Southern analysis and sequenced. In parallel, a (cid:2) gt10 cDNA library made from a pool of normal human prostates (CLONTECH) was screened by established procedures (19) to obtain additional clones. Ovelapping clones were used to deduce the full-length STAMP1 cDNA sequence. The full-length STAMP1 ORF was amplified by using primers cen-tered around the start and stop codons (sequences available upon re- quest) and fused in-frame to the C terminus of GFP using the vector pcDNA3.1-NT-GFP-TOPO (Invitrogen) to generate GFP-STAMP1. Protein Sequence Analysis— Primary sequence analysis for STAMP1 was performed in BLAST (www.ncbi.nlm.nih.gov/BLAST/). Secondary protein structure predictions were performed using the web tools SMART (smart.embl-heidelberg.de/), SOSUI (sosui.proteome.bio.tuat. and PSORT Analysis— by step gua- nidine thiocyanate Northern g of by random priming and had a specific activity of (cid:2) 3 (cid:3) 8 dpm/ (cid:3) g. A cDNA fragment of STAMP1 spanning 145–2202 bp was used as a probe. Bands visualized and

The prostate gland is a major secretory organ whose precise function is still not known (1). Through secretions into the male ejaculate, it is thought that the prostate protects the lower urinary tract from infection and increases fertility. Despite the unknown specific function, the prostate is the most common site of neoplastic transformation in men. Prostate cancer is the most commonly diagnosed cancer and the second leading cause of cancer mortality in men other than skin cancer (2). In the initial stages, prostate cancer is dependent on androgens for growth, which is the basis for androgen ablation therapy (3). However, in most cases, prostate cancer progresses to an androgen-independent phenotype for which there is no effective therapy available at present (for reviews, see Refs. 4 and 5).
Currently, there is limited information regarding the molecular details of normal prostate function as well as prostate cancer initiation and progression. Several independent approaches resulted in the identification of a few highly prostateenriched genes that may have unique roles in these processes. The first such gene discovered was prostate-specific antigen (PSA) 1 (for a review, see Ref. 6), which is currently used as a diagnostic tool and also as a marker for the progression of prostate cancer, albeit with significant limitations (7,8). More recently, several additional prostate-enriched genes were identified including prostate-specific membrane antigen (PSMA) (9), prostate carcinoma tumor antigen 1 (PCTA-1) (10), NKX3.1 (11,12), prostate stem cell antigen (PSCA) (13), DD3 (14), and PCGEM1 (15). Research on these genes and their products is likely to provide detailed information on normal and hyperplastic prostate biology as well as improving disease diagnosis, prognosis, and therapy. However, these goals have not yet been realized.
While searching for genes that are differentially expressed during early stages of prostate cancer (16), we have cloned a novel gene, named six transmembrane protein of prostate 1 (STAMP1). STAMP1 is highly specific to prostate for expression. GFP-tagged STAMP1 in immunocytochemistry and timelapse imaging studies indicate that STAMP1 may have a role in endocytic and/or secretory trafficking pathways. Furthermore, STAMP1 expression is increased in androgen-independent prostate cancer xenografts compared with their androgendependent counterparts as well as in prostate tumors compared with normal prostate. These data suggest that STAMP1 may have a key role in both normal prostate physiology and the progression of prostate cancer.

MATERIALS AND METHODS
Cell Culture-LNCaP, PC-3, and DU145 cells were routinely maintained and treated as described previously (16).
Xenograft Studies-Transplantation, growth, and harvesting of tumors from mice bearing the CWR22 and CWR22R xenografts were as previously described (17,18). * This work was supported by grants from the Norwegian Research Council, the Norwegian Cancer Society, and the University of Oslo (to F. S.). 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.
The nucleotide sequence(s) reported in this paper has been submitted to the GenBank TM /EBI Data Bank with accession number(s) AY008445.
¶ These authors contributed equally to this work. ‡ ‡ To whom correspondence should be addressed. Cloning and Plasmid Construction-A 262-bp cDNA fragment was originally obtained from a screen of a prostate-specific library (16) and termed L74. RACE (5Ј-rapid amplification of cDNA ends) was performed (oligonucleotide sequences available upon request) using the Marathon-Ready cDNA that was prepared from normal prostate tissue (CLONTECH) and/or SMART-RACE LNCaP cDNA library (CLONTECH) that was generated according to the manufacturer's recommendations. RACE products were cloned into pCRII-TOPO (Invitrogen), and positive clones were confirmed by Southern analysis and sequenced. In parallel, a gt10 cDNA library made from a pool of normal human prostates (CLONTECH) was screened by established procedures (19) to obtain additional clones. Ovelapping clones were used to deduce the full-length STAMP1 cDNA sequence.
The full-length STAMP1 ORF was amplified by using primers centered around the start and stop codons (sequences available upon request) and fused in-frame to the C terminus of GFP using the vector pcDNA3.1-NT-GFP-TOPO (Invitrogen) to generate GFP-STAMP1.
Northern Analysis-Total RNA was prepared by the single step guanidine thiocyanate procedure and used in Northern analysis (19). 15 g of total RNA were used per lane. Probes were generated by random priming and had a specific activity of Ͼ3 ϫ 10 8 dpm/g. A cDNA fragment of STAMP1 spanning 145-2202 bp was used as a probe. Bands were visualized and quantitated by phosphorimaging analysis (Amersham Biosciences).
Confocal Microscopy and Live Cell Imaging-COS-1 cells were transfected by electroporation using a BTX square-wave pulser at 150 V, 1 ms duration. Cells were grown either on coverslips placed in 6-well tissue culture plates for indirect immunofluorescence or on Lab-Tek Chambered Coverglass (Nalge Nunc International) for live-cell microscopy. Transiently transfected cells were observed 18 h after transfection by a Leica TCS-SP confocal microscope using a 488-nm argon laser line. All live-cell experiments were done at 37°C.
In Situ Hybridization-The STAMP1 riboprobe was made on a STAMP1 cDNA fragment corresponding to nucleotides 1339 -2074 of STAMP1 cDNA. In parallel and as a negative control, a sense probe was generated corresponding to nucleotides 171-830 of STAMP1 and used in in situ analysis that did not give any specific staining as shown. The procedure was essentially as described previously (51).

RESULTS
Isolation and Characterization of the STAMP1 Gene and mRNA-While searching for prostate-specific genes that are regulated in the early stages of prostate cancer (16), we cloned a 262-bp novel cDNA fragment, termed L74, which did not show any sequence similarity to the sequences in GenBank TM . By screening a normal prostate cDNA library and by 5Ј-and 3Ј-RACE analysis, we obtained the full-length cDNA for L74. Because computer-aided secondary structure prediction of the deduced amino acid sequence of L74 suggested the presence of a six transmembrane domain in its C-terminal half, we named L74 STAMP1.
When the full-length STAMP1 cDNA was used in BLAST analysis we found that it completely matched a BAC clone (GenBank TM accession no. AC002064) except for a 313-bp repetitive unit in the 3Ј-UTR region (data not shown), thereby identifying it as the STAMP1 gene and localizing it to Chr7q21. The repetitive region is likely to be a cloning or sequencing artifact of the BAC clone. Computational exon/intron junction analysis and alignment of the full-length cDNA sequence with the BAC clone revealed that the STAMP1 gene is composed of six exons and five introns (data not shown). The STAMP1 gene spans around 26 kb, which is in part due to the extremely large size of intron 2 (12713 bp). There are three different predicted promoters within 4 kb upstream of the STAMP1 initiation codon (data not shown), none of which has any significant TATA or CAAT box consensus sequences, indicating that STAMP1 is transcribed from a TATA-less promoter.
The STAMP1 cDNA has a predicted 5Ј-UTR of ϳ1 kb (deduced by RACE analysis, data not shown) and an unusually long 3Ј-UTR of ϳ4 kb that comprises ϳ77% of the total cDNA sequence (data not shown). The ORF starts within the third exon and is predicted to encode a 490 amino acid protein (Fig.  1A). A search for known protein motifs identified six predicted transmembrane domains in the C-terminal half of STAMP1 starting at Phe-209 (Fig. 1B).
STAMP1 Belongs to a New Subfamily of Six Transmembrane Domain Proteins-BLAST analysis in GenBank TM with the predicted STAMP1 amino acid sequence identified two recently discovered proteins that showed significant similarity to STAMP1 over the entire ORF: the rat protein pHyde (21), which, when overexpressed, can cause apoptosis in prostate cancer cells and the tumor necrosis factor ␣-induced adiposerelated protein (TIARP) (22), a mouse protein which may have a role in adipocyte differentiation. TIARP and pHyde are 43 and 51% identical, respectively, to STAMP1 at the amino acid level. In addition, there was significant similarity in the Cterminal half of STAMP1 to STEAP, a recently discovered human cell membrane protein enriched in prostate for expression (23). An alignment of these sequences is shown in Fig. 1B. These data suggest that STAMP1, TIARP, and pHyde are structurally, and possibly also functionally, related proteins, and they therefore form a small six transmembrane protein subfamily that does not include STEAP.
STAMP1 Is Highly Enriched to Prostate for Expression-We next determined the expression profile of STAMP1 in various human tissues by Northern analysis in which a multiple tissue Northern blot was hybridized to the STAMP1 probe. As shown in Fig. 2A, STAMP1 hybridized to a major mRNA species of 6.5 kb and three minor mRNA species of 2.2, 4.0, and 4.5 kb in the normal prostate tissue. There was 15-20-fold lower mRNA expression of the 6.5-kb band in the heart, brain, kidney, pancreas, and ovary but not in other tissues. In contrast, the three lower molecular weight species, which are likely to be encoded by alternatively spliced forms of STAMP1, 2 were only detectable in the prostate. Hybridization with a glyceraldehyde 3-phosphate dehydrogenase (G3PDH) cDNA probe resulted in approximately similar signals in all lanes except for the heart and skeletal muscle where G3PDH is known to be more abundant compared with other tissues. These data show that STAMP1 is highly specific to prostate and that it may have isoforms that are restricted to prostate for expression.
Characterization of STAMP1 Expression in Prostate Cancer Cell Lines and Xenografts-Because androgen is a major hormonal stimulus for the normal prostate gland and for early stage prostate cancer (3), we assessed the possible androgen regulation of STAMP1 by Northern analysis in the androgen responsive prostate cancer cell line LNCaP (24). Cells were either left untreated or treated with the synthetic androgen R1881 for 24 h and harvested, and total RNA was isolated and used in Northern analysis with STAMP1 cDNA as probe. As shown in Fig. 2B, STAMP1 displayed similar expression levels in untreated and R1881-treated LNCaP cells. In contrast, the mRNA accumulation of the androgen-regulated gene NKX3.1 dramatically increased upon androgen stimulation in a time-dependent manner, as expected (11,25), reaching ϳ10-fold higher levels by 24 h. Time course analysis of androgen treatment did not result in significant differences in STAMP1 expression. 2 These data suggested that STAMP1 expression is not significantly regulated by androgens in LNCaP cells.
We also assessed the possible regulation of STAMP1 expression in an in vivo setting using the recently developed androgendependent xenograft model CWR22, which is derived from a primary human prostate tumor (17). Because it is androgendependent for growth, the CWR22 tumors in nude mice display marked regression upon castration and may in fact regress completely (17). CWR22 xenografts were grown in nude mice in the presence of a sustained-release testosterone pellet (17). After the tumors had grown, mice were castrated, the testosterone pellets were removed, and the regressing tumors were collected at 1, 2, or 4 weeks postcastration. Total RNA was prepared from these tumor samples and used in Northern analysis. As shown in Fig. 2B, similar to that observed in LNCaP cells, STAMP1 mRNA accumulation in the CWR22 tumors showed no significant change upon castration and was not significantly affected by the presence of androgens. In contrast, the mRNA accumulation of the androgen-regulated gene NKX3.1 was dramatically decreased upon castration, dropping to ϳ4% of precastrate levels by 4 weeks postcastration. These results are consistent with the findings in LNCaP cells and suggest that STAMP1 expression is not significantly regulated by androgens in prostate cancer cells. Interestingly, STAMP1 expression was substantially lower in the CWR22 tumors compared with LNCaP cells.
We also analyzed the expression profile of STAMP1 in the androgen receptor-negative and therefore androgen-independent prostate cancer cell lines PC3 and DU145, as well as in  (TM 1-6). B, BLAST analysis identified TIARP (GenBank TM NP473439), pHyde (Gen-Bank TM AAK00361.1), and STEAP (GenBank TM AF186249) as closely related recently identified proteins. The multiple sequence alignment obtained by Clustal and GenDoc programs is shown. Completely conserved residues are shaded in blue. Residues that are conserved in two or three of the sequences are shaded light and dark gray, respectively. four independent, relapsed derivatives of CWR22 tumors, named CWR22R (18), which are representative of advanced prostate cancer. Interestingly, there was no STAMP1 expression in the androgen-independent prostate cancer cell lines PC-3 or DU145, similar to that observed for NKX3.1 (Fig. 2B). In contrast, there was significant STAMP1 expression in tumors from all four independent CWR22R xenograft lines tested, ranging between ϳ30 and 60% of that observed in LNCaP cells. A similar overexpression pattern was also observed for NKX3.1 (Fig. 2B), consistent with previous findings (25). These data suggest that STAMP1 expression may be deregulated once prostate cancer progresses from an androgendependent to an androgen-independent state.
Intracellular Localization of STAMP1-To gain insight into the cellular localization of STAMP1, we labeled it with GFP to generate GFP-STAMP1. Such use of GFP fusion proteins has recently become a standard method to assess intracellular localization and dynamics of proteins (26,27). COS-1 cells were transiently transfected with GFP-STAMP1, fixed, and processed for confocal microscopy. The series of 11 confocal sec-tions along the z-axis were collected through a single cell at nominal 100-nm intervals. Three of the consecutive sections and projection of all 11 sections are shown in Fig. 3A. In all 11 z-plane sections, GFP-STAMP1 showed bright juxtanuclear distribution patterns, characteristic of the Golgi complex. Additionally, GFP-STAMP1 was dispersed in dots with variable size throughout the cytoplasm and at the cell periphery (Fig. 3,  z-7, projection). Some of these bright fluorescent spots were tubular (z-6, arrow) or vesicular (z-5, arrow) in morphology.
To determine more directly whether GFP-STAMP1 was localized to the Golgi complex, we compared its intracellular distribution with those of two well characterized Golgi markers, the medial Golgi enzyme mannosidase II (ManII) (28) and the coat protein ␤-COP (29). GFP-STAMP1 was transfected into COS-1 cells, which were then fixed and labeled with the appropriate primary and secondary antibodies, and then optical sections were analyzed by laser scanning confocal microscopy. As shown in Fig. 3B, the distribution of GFP-STAMP1 extended throughout the Golgi complex, as evidenced by significant colocalizations with both ManII and ␤-COP. However, some areas of non-overlap between the GFP-STAMP1 and both Golgi markers were observed, suggesting that STAMP1, at least in part, is differentially localized within the Golgi complex compared with these two markers. Virtually identical results were obtained by visualization of endogenous STAMP1 in LN-CaP cells using a STAMP1-specific antiserum (data not shown).
Because GFP-STAMP1 appeared to be associated with vesicular tubular structures (VTS) (Figs. 3A and 4), we assessed whether it may be more specifically localized to the TGN, an important site for the sorting of proteins destined to the plasma membrane, secretory vesicles, or lysosomes (30 -32). To that end, we used an antibody against TGN46, a TGN resident protein that shuttles between the TGN and the plasma membrane, (33,34) in immunofluorescence microscopy experiments as above. As shown in Fig. 3B, GFP-STAMP1 extensively (ϳ80%) co-localized with TGN46, greater than that observed with ManII and ␤-COP, suggesting that in the Golgi complex STAMP1 is primarily localized to the TGN. STAMP1 Shuttles Between the Golgi and the Plasma Membrane and Colocalizes to the Endosomes-To gain insight into the possible function of STAMP1, the kinetic properties of GFP-STAMP1 distribution and trafficking were studied using confocal time-lapse imaging in living cells. COS-1 cells were transfected with GFP-STAMP1. 18 h after transfection, images were captured from live cells every 20 s at 37°C by confocal laser scanning microscopy (see QuickTime movie sequence at www.biologi.uio.no/mcb/fs/project2.html).
As shown in Fig. 4, there was anterograde-retrograde trafficking of GFP-STAMP1 to and from the Golgi complex predominantly in the form of VTS. Some of the VTS followed straight or curvilinear paths, some moved in a stop-and-go fashion, and some showed saltatory movements (see Quicktime movie). The mobile structures indicated at the top panel (white arrows) extended away from and then retracted back to the Golgi. The VTS in the middle panel and the first image in the lower panel (red arrows) detached from the Golgi complex, paused, and moved toward the cell periphery until it disappeared at the cell edge, suggesting that STAMP1 is associated with the secretory pathway. The VTS (yellow arrow) in the lower panel moved from the cell periphery toward the Golgi body, suggesting that STAMP1 is also associated with the endocytic pathway.
To probe whether GFP-STAMP1 was associated with the endocytic pathway, we compared the intracellular distribution of GFP-STAMP1 with that of the early endosome protein EEA1 (35). GFP-STAMP1 was transfected into COS-1 cells that were then fixed, immunostained with EEA1 antibodies, and ob- Similary, PC3 and DU145 cells were cultured but were left untreated. Total RNA was isolated and used in Northern analysis with STAMP1 as the probe. The same membrane was also probed for the androgen-dependent gene PSA. Relative induction of mRNA accumulation is indicated at the bottom of the lanes determined by phosphorimaging analysis (Amersham Biosciences). The CWR22 xenograft was grown in nude mice, and tumor samples were collected either before (t ϭ 0) or 1, 2, or 4 weeks after castration. Total RNA was isolated and then used in Northern analysis with the same probes. Note that CWR22 week 2 lane was underloaded (compare 18 S RNA). In parallel, four independent lines of the androgen-independent human prostate cancer xenograft CWR22R were grown in nude mice, tumors were collected, and total RNA was isolated and used in Northern analysis with STAMP1 or the androgen target gene NKX3.1 cDNAs as probes, as indicated. served by laser scanning confocal microscopy. As shown in Fig.  5, EEA1 had similar intracellular distribution in both transfected and untransfected cells. Furthermore, GFP-STAMP1 significantly colocalized with EEA1 both in the cell periphery and also in the perinuclear area (Fig. 5, indicated by arrows), suggesting that STAMP1 is associated with early endosomes and the endocytic pathway.
Analysis of STAMP1 Expression in Normal Versus Cancerous Prostate-Because STAMP1 is highly enriched in the prostate compared with other tissues, we studied its expression in normal prostate and compared it with prostate cancer. To assess the location of STAMP1 expression in prostate tissue and the possible differences in normal compared with neoplastic cells, we performed in situ hybridization analysis on six embedded tissue blocks from prostatectomy specimens (data not shown) or a tissue microarray that contained 46 tissue samples from 13 prostatectomy cases (example shown in Fig. 6). STAMP1 was expressed exclusively in the epithelial cells of the prostate, both in normal and tumor glands (Fig. 6A). However, the level of STAMP1 expression was higher in the tumor glands, ϳ2.5-fold overall, compared with normal epithelium (Fig. 6B). These data indicate that STAMP1 may directly contribute to, or be a marker for, the progression of prostate cancer.

DISCUSSION
Proteins that contain six transmembrane domains have key roles in a variety of important physiological processes. For example, they function as ion channels (36), as signal transducers of painful stimuli (37), as water channels or aquaporins (38), as lipid phosphate phosphatases or LPPs (39), and as ATP-binding cassette (ABC) transporters and multidrug resistance (MDR) proteins (40). Given these important functions, it is not surprising that deregulation/mutation of some of these proteins is implicated in the pathogenesis of major human diseases such as Alzheimer's (41) and Tangier disease (42). Because STAMP1 is highly prostate-specific for expression, it is tempting to speculate that deregulation or mutation of STAMP1 may have a role in the pathogenesis of prostate cancer.
STAMP1 shows significant similarity to three recently identified genes. Although there is high sequence similarity with both TIARP and pHyde over the whole ORF, the similarity with STEAP is restricted to the predicted transmembrane domain in the C terminus. In fact, STEAP is a smaller protein and does not have an N-terminal region compared with STAMP1 and TIARP. Both pHyde and STEAP have previously been implicated in the prostate; pHyde for its ability to induce growth arrest and apoptosis of prostate cancer cell lines (21) and STEAP for its significantly prostate-enriched expression pattern (23). There is no information on the expression of TIARP in the prostate to date, but our preliminary analysis with the human TIARP cDNA suggests that it has a ubiquitous expression profile, similar to that for human pHyde. 3 Based on the secretory/endocytic pathway localization of GFP-STAMP1, TIARP, pHyde, and STEAP may also be involved in intracellular trafficking pathways. However, there are some differences in their subcellular distribution patterns compared with STAMP1, suggesting that they have distinct functions in the cell. Further work is in progress to assess whether the sequence similarities of these proteins also extend to an overlap in cellular localization and function.
To our knowledge, STAMP1 is the first six transmembrane protein described to date that is localized to the Golgi and TGN. Most Golgi-resident proteins studied so far are transmembrane proteins that share a common structure, the majority of which have a single type II transmembrane domain (43,44). Expression of STAMP1 as a six transmembrane protein in the Golgi suggests that it may have unique functions compared with the other Golgi-resident proteins identified so far. Photobleaching experiments provided evidence that GFP-STAMP1 is rapidly exchanged between different parts of the Golgi with kinetics consistent for it being a transmembrane protein (data not shown), supporting the notion that it indeed functions as a transmembrane protein in the Golgi. In addition, in the presence of Brefeldin A, a known inhibitor of the formation of TGN (45,46), GFP-STAMP1 localization to TGN is inhibited which is reversible (data not shown). Further work is needed for the biochemical and functional characterization of STAMP1 in the Golgi and TGN.
The Golgi complex has a central role in the secretory pathway (30,31,47,48). It is involved in the processing and sorting of proteins and lipids to their final destinations, i.e. to the cell surface, secretory granules, endosomes, or recycling back to the endoplasmic reticulum. Because prostate is a major secretory organ (49), trafficking through the Golgi is expected to be important and tightly regulated in prostate cells. However, in contrast to some other specialized organs, such as the pancreas (50), there is no specific knowledge about functioning of the Golgi in the prostate. For example, it is possible that the secreted molecules of the prostate, such as PSA, are modified or sorted in a regulated fashion in the Golgi complex. Resident enzymes of the Golgi in the prostate, such as STAMP1, may therefore have an important role in this type of regulation. Alternatively, STAMP1 may be a receptor for an endogenous or exogenous ligand that may be delivered to different intracellu-FIG. 6. In situ analysis of STAMP1 expression in normal versus neoplastic prostate tissues. In situ hybridization of paraffin-embedded human prostate cancer specimens (data not shown) or tissue arrays derived from them were performed using probes derived from either sense or antisense strands of STAMP1. Hematoxylin and eosin staining is shown to the left and in situ images are to the right. The other panels are photographed using dark-field microscopy after a 14-day exposure. Positive hybridization is seen as bright spots. A, in situ analysis on prostate cancer tissue array. S, sense probe; AS, antisense probe. B, comparison of STAMP1 expression in normal versus tumor cells in the prostate by in situ analysis. Sections on a prostate array were scored by a pathologist and assigned a relative value of 0, 1, 2, or 3 according to increasing levels of STAMP1 expression in 18 normal and 28 tumor sections. Each score is presented: normal, open circles; tumor, filled circles. Student's t test analysis indicates that the means of the normal and tumor samples are significantly different (p Ͻ 0.001). lar compartments through the secretory and endocytic pathways to be involved in cellular processes. Further studies are required to assess these possibilities.
An interesting property of STAMP1 expression profile is that it is not expressed in the AR-negative prostate cancer cell lines PC-3 and DU145 but is expressed at high levels in the ARpositive cell line LNCaP. It is also present in the CWR22 and CWR22R xenografts, which are AR-positive. Thus, expression of STAMP1 is correlated with the presence of a functional AR in the cell. Whether AR is directly involved in STAMP1 expression will require the characterization of STAMP1 flanking sequences and further functional studies. Another related area to pursue is the expression of STAMP1 and its variants in the different cell populations that are found in the prostate. Laser capture microdissection coupled to quantitative RT-PCR is in progress to look for these possible differences.
It has been known for over 60 years that androgens play a key role both in the development and maintenance of the normal prostate and the initiation and progression of prostate cancer (3). This is the basis for androgen withdrawal therapy, which unfortunately fails in most cases after a few months or years. At this point, there is no effective therapy and prognosis for survival is extremely poor (3)(4)(5). Because STAMP1 expression may be increased during the progression to this advanced state, it may prove to be a useful tool in diagnostic and therapeutic applications for prostate cancer, in addition to helping define the molecular dynamics of the prostate cancer cell.