Characterization of KLK4 Expression and Detection of KLK4-Specific

Oncogene (2002) 21, 7114 – 7120 ª 2002 Nature Publishing Group All rights reserved 0950 – 9232/02 $25.00 www.nature.com/onc SHORT REPORTS Characterization of KLK4 expression and detection of KLK4-speciﬁc antibody in prostate cancer patient sera

Craig H Day*,1, Gary R Fanger1, Marc W Retter1, Bonnie L Hylander2, Remedios B Penetrante2, Raymond L Houghton1, Xinqun Zhang1, Patricia D McNeill1, Aristides Maltez Filho3, Marcos Nolasco3, Roberto Badaro3, Martin A Cheever1, Steven G Reed1,4,5, Davin C Dillon1 and Yoshihiro Watanabe1

1Corixa Corporation, 1124 Columbia Street, Suite 200, Seattle, Washington, WA 98104, USA; 2Department of Immunology, Roswell Park Cancer Institute, Elm and Carlton Street, Buﬀalo, New York, NY 14263, USA; 3Hospital Aristides Maltez, Bahia, Brazil; 4Department of Pathobiology, University of Washington, Box 357238, Seattle, Washington, WA 98195, USA; 5Infectious Disease Research Institute, 1124 Columbia Street, Suite 600, Seattle, Washington, WA 98104, USA

The ability to identify prostate tumor or prostate tissue radiation therapy. Either technique can fail to eliminate specific genes that are expressed at high levels and use their all tumor cells, resulting in recurrence of disease. This protein products as targets could greatly aid in the possibility is increased if diagnosis of prostate cancer is diagnosis and treatment of prostate cancer. Using a made late, with extracapsular involvement. polymerase chain reaction (PCR)-based subtraction A number of human prostate-specific antigens have technique, we have recovered the recently described previously been identified, including prostate-specific KLK4 (prostase) gene from human prostate cDNA. In antigen (PSA), prostatic acid phosphatase and pros- this study, KLK4 gene expression in human prostate tate-specific membrane antigen. PSA is currently a tumors was further characterized using cDNA quantitative well-established clinical marker as an indicator for PCR and immunohistochemistry, demonstrating that the prostate cancer (Daher and Beaini, 1998), with the gene is specifically expressed at both the mRNA and advantages that it is secreted and enters the circulatory protein levels in normal human prostate tissue, and in both system, and that virtually all primary prostate tumors primary and metastatic prostate tumor samples. Quanti- maintain PSA expression. Increasing levels of PSA in tative mRNA analysis also demonstrated low level peripheral blood has been shown to be an effective expression including adrenal gland, salivary gland and clinical marker useful in detecting the presence of thyroid. Finally, it was demonstrated that prostate cancer prostate cancer, either in the initial diagnosis or in the patient sera contain antibodies that bind specifically to detection of disease progression following procedures recombinant KLK4 protein. This antibody has been used to such as radical prostatectomy (Abi-Aad et al., 1992). detect KLK4-specific peptides in epitope mapping experi- However, two problematic issues remain with utilizing ments. The relatively specific expression profile and PSA as a marker for the diagnosis of prostate cancer elevated level of KLK4 mRNA and protein in both tumor and the detection of residual disease. First, there is a and normal prostate tissues, in addition to detectable high level of false positive tests due to PSA expression KLK4-specific antibody in cancer patient sera, supports in benign prostatic hyperplasia that result in unneces- additional efforts to determine if KLK4 can play a role in sary biopsies. Second, PSA can fail as a marker for the diagnosis of prostate cancer, the monitoring of residual residual disease since not all metastases maintain disease, or act as a target for immunotherapy. expression of PSA (Daher and Beaini, 1998; Sissons Oncogene (2002) 21, 7114 – 7120. doi:10.1038/sj.onc. et al., 1992), and the presence of the PSA gene by 1205786 reverse transcription (RT) – PCR from blood does not unequivocally indicate specific prostatic metastases Keywords: antibody; KLK4; prostase; prostate cancer; (Thiounn et al., 1997). Due to these limitations, protein expression additional markers for prostate cancer are needed. Recently, an additional prostate-specific androgen- regulated serine protease was identified, termed Prostate cancer is the most common cancer among men, prostase (Nelson et al., 1999), KLK-L1 (kallikrein- and is the second leading cause of male cancer deaths in like1) (Yousef et al., 1999) and also KLK4 (Stephenson the US (Grenlee et al., 2000). The current treatment et al., 1999). The initial characterization of this gene options include prostatectomy or prostate ablation by (Nelson et al., 1999) demonstrated that prostase mRNA expression is restricted to normal prostate tissue and the prostate tumor cell line LNCaP. The *Correspondence: C Day; E-mail: [email protected] prostase cDNA was shown to encode a putative 254 Received 22 March 2002; revised 12 June 2002; accepted 18 June amino acid polypeptide, and was radiation-hybrid- 2002 panel mapped to chromosome 19q13, a region contain- Prostase characterization and expression in prostate CH Day et al 7115 ing several other serine proteases including three expression was detected in normal lung or colon tissue known human kallikreins (hK1, hK2 and hK3 samples. [PSA]). A genomic analysis of this region more Quantitative PCR was then performed to determine precisely mapped and localized the prostase gene and mRNA levels of KLK4 in prostate tumors, normal the related positions of other known genes in the area prostates and other normal human tissues. As shown in including PSA, KLK1, KLK2, zyme and neuropsin Figure 1b, increased expression of KLK4 mRNA is (Yousef et al., 1999). An additional study mapped the evident in all 22 prostate tumors and all three normal kallikrein-related prostase gene to the KLK locus, and prostate samples analysed. Elevated expression of based upon sequence similarities and like-character- KLK4 is evident in one metastatic prostate tumor istics amongst the three human kallikrein genes, sample (met-lymph node), and low but detectable proposed prostase be designated the fourth human expression is apparent in a second metastatic prostate kallikrein (KLK4) (Stephenson et al., 1999). Fluores- tumor sample (met-ascites). Elevated expression of cence in situ hybridization analysis confirmed the KLK4 is also evident in all three of the benign localization of KLK4, or PRSS17, to the long arm of prostatic hyperplasia (BPH)-positive samples examined human chromosome 19, and more specifically as compared to non-prostate samples. In addition, low narrowed the gene’s location to 19q13.3?q13.4 but detectable levels of expression were observed in the (DuPont et al., 1999). normal adrenal, salivary and thyroid gland samples, In this study we quantitatively characterize KLK4 while the remaining 29 normal tissues tested yielded expression, extending the mRNA expression profile of exceedingly low or no detectable quantities of amplified KLK4 by examining primary and metastatic human KLK4 mRNA. prostate tumors as well as mRNA expression char- To characterize KLK4 expression at the protein acteristics of a large panel of human cell lines. In level, both rabbit polyclonal antisera and monoclonal addition, anti-KLK4 antibodies were generated and antibodies were produced. Recombinant KLK4 was used to demonstrate that KLK4 is a secreted protein. generated by expression of a six histidine-tagged, C- Using immunohistochemical analysis, the prostate- terminal 159 amino acid fragment in E. coli and the specific expression profile was confirmed at the protein product was purified using nickel-nitrilotriacetic acid level. Lastly, we demonstrate that sera from patients (NTA) column chromatography for subsequent use as with prostate cancer contain antibody specific to an immunogen in rabbits. Polyclonal antisera was recombinant KLK4 protein. Collectively, these findings tested for KLK4 specificity and then affinity purified suggest further evaluation of KLK4 as a diagnostic for use in Western blot and IHC analysis. marker and as a potential therapeutic target. To determine if the KLK4 gene encoded a secreted To date, characterization of the expression of this protein, it was cloned into the pcDNA3.1 mammalian new serine protease family member has focused on expression vector and used to transfect HEK293 cells. mRNA levels in human normal tissues and prostate Analysis of transiently transfected HEK293 cell lysates tumor cell lines (Nelson et al., 1999; Yousef et al., revealed the expression of an *22 kD protein that is 1999), hormonal regulation of a human endometrial recognized by the anti-KLK4 affinity purified rabbit cancer cell line (Myers and Clements, 2001), serous sera (data not shown). Analysis of supernatants ovarian carcinomas (Dong et al., 2001) and on recovered from the same transfection revealed the expression levels as an indicator of ovarian cancer expression of an *25 kD protein due to glycosylation. patient prognoses (Obiezu et al., 2001). In order to Furthermore, N-terminal amino-acid sequence analysis extend the characterization of the tissue-specific revealed the cleavage of the putative signal sequence expression of KLK4, cDNA quantitative PCR was (residues 1 – 26) (Nelson et al., 1999), retaining the used to analyse the expression of KLK4 in a large presence of the four amino acid propeptide (SCSQ) panel of samples, including normal prostate, primary indicating incomplete cleavage of the protein into its and metastatic prostate tumors, human cell lines and a predicted active form (data not shown). large panel of normal tissues. In order to assess KLK4 expression by immunohis- Quantitative PCR was performed to determine tochemistry methods, a monoclonal antibody (mAb) mRNA levels of KLK4 in a large panel of human cell specific for KLK4 was generated and used to stain lines (Figure 1a). KLK4 was found to be expressed in tissue sections. Figure 2 shows representative examples LnCaP prostate tumor cell lines, but not in DU145 or of KLK4 immunoreactivity in tissue sections derived PC-3 prostate tumor cell lines, a finding consistent with from a prostate tumor and normal prostate. In both previously published data (Nelson et al., 1999). KLK4 normal prostate and prostate tumor samples, KLK4 was also found to be expressed in pooled normal expression was predominantly restricted to the epithe- prostate tissues as expected, but had no detectable lial cells lining the ductal regions (Figure 2, upper expression in a large number of non-prostate tumor panels). Overall, KLK4 expression was typically lines, including breast, colon, ovarian, and lung tumor granular, cytoplasmic and dispersed throughout the cell lines. In addition, extremely low or no detectable cell. Four different prostate cancer tumor samples and expression was observed in non-prostate tumor four normal prostate samples were analysed, all of samples, including pooled breast, colon, ovarian, and which were positive for anti-KLK4 immunoreactivity. lung tumor tissue. Finally, extremely low expression Breast cancer and normal kidney tissues were also was detected in normal breast tissues whereas no stained with the anti-KLK4 mAb and showed no anti-

Oncogene Prostase characterization and expression in prostate CH Day et al 7116

Figure 1 Quantitative PCR analysis was performed on the Perkin-Elmer/Applied Biosystems (Foster City, CA, USA) GeneAmp 5700 Sequence Detection System. Matching primers were designed for KLK4 according to the Primer Express software program (Perkin-Elmer/Applied Biosystems). Primer sequences used were as follows: (5’ primer: 5’-CTTGCTCGC-TAACGACCTCAT-3’ and 3’ primer: 5’-GCCAGAAACGAGGCAAGAGT-3’). Primer concentrations were 300 nM each. SYBR Green reaction mixes were generated according to manufacturer’s protocol (Perkin-Elmer/Applied Biosystems). Each reaction contained 2 ml of diluted template. cDNA from RT reaction prepared as above was used as template diluted 1 : 10 for each gene of interest and 1 : 100 for human b-actin quantitation. The human b-actin primer sequences were (5’ primer: 5’-ACTGGAACGGTGAAGGTGACA-3’ and 3’ primer: 5’-CGGCCACATTGTGAACTTTG-3’) and were used in a similar manner to quantitate the presence of b-actin in the samples. A portion of KLK4 was used as the template control. It was isolated by using prostate-speciﬁc PCR-select cDNA (Clontech, Palo Alto, CA, USA). Suppression subtractive hybridization was performed using a pool of normal human prostate cDNA as the tester and a pool of 10 other normal human tissues as a driver (Diatchenko et al., 1996, 1999). (a) KLK4 mRNA expression analysis in tumor cell lines, primary tumor and normal tissue samples. Cell line sources used for the analysis were as follows: prostate, [Du145, LNCaP, PC-3 (ATCC, Manassas, VA, USA)]; breast [MCF7, SKBR3, MDA-MB 435, MDA-MB 415, MDA-MB 453, BT-474, T-47D (ATCC)]; lung [A549, 391-06 (ATCC); TL-1, LT140-68 (kindly contributed by E Repasky, Roswell Park Cancer Institute, Buffalo, NY, USA); 390T, 84T (kindly contributed by J Siegfried, University of Pittsburgh Cancer Institute, Pittsburgh, PA, USA)]; [(ad) is adenocarcinoma, (sq) is squamous, (lg) is large cell], ovarian [SKOV3, OV-1063 (ATCC); OT391-73 (endometrial, generated internally)] and colon [SW480, Colo 320 (ATCC)]. In addition, both pooled tumor (5 – 10 samples) and pooled normal tissue (5 – 10 samples) mRNAs were included from these tumor types. RNA was isolated from tissue samples by homogenization in Trizol reagent (Life Technologies, Rockville, MD, USA) and extracted as per manufacturer’s protocol. cDNA was prepared from 26 mg of total RNA by incubating with 100 ng/ml oligo dT (Roche, Indianapolis, IN, USA) at 708C for 10 min followed by incubation at 428C for 1 h with 0.5 mM dNTP, 1000 units RNasin, 0.02 mM dithiothreitol and 2500 units of

Oncogene Prostase characterization and expression in prostate CH Day et al 7117 KLK4 immunoreactivity (Figure 2, lower panels). In existence of Abs specific to KLK4 protein. The pooled addition to these tissues, various other malignant and sera (1000-fold dilution) of either patient or normal non-malignant tissues were examined with the intent of donor sera were used in the Western blot analysis. The better defining the specificity of KLK4 expression. No pooled patient sera recognized a KLK4-specific band anti-KLK4 immunoreactivity could be observed in cancerous lung tissue, or in normal lung, colon and liver tissue samples (data not shown). A summary of Table 1 Summary of KLK4 immunohistochemistry analysis of the relative levels of anti-KLK4 immunoreactivity is various tissue sections. KLK4 expression levels range from a relative shown in Table 1. score of 0, indicating no expression, to a score of 3+, which indicates maximum expression levels Furthermore, we have addressed whether or not prostate cancer patients have had KLK4-specific B-cell IHC analysis of KLK4 expression Tissue analysed Anti-KLK4 Immunoreactivity (0 – 3+) responses induced. Although normal prostate epithelial cells express KLK4 at the protein level, it is known Prostate cancer 3+ that tumor regions are frequently associated with Normal prostate/BPH 3+ inflammation and invasion of inflammatory cells. Normal liver 0 Normal colon 0 Given this information, we asked if prostate tumor Normal kidney 0 patient sera contain antibodies specific to KLK4. Normal lung 0 Initially, eight prostate cancer patient sera and eight Lung cancer 0 normal donor sera were pooled and examined for the Breast cancer 0

Figure 2 Immunohistochemistry analysis of KLK4 expression in prostate tumor and normal tissues. Representative examples of KLK4 expression in prostate cancer (upper left), normal prostate (upper right), breast cancer (lower left) and kidney (lower right). Rabbits (R and R Rabbitry, Stanwood, WA, USA) were immunized with 400 mg of purified, E. coli-generated KLK4 in 1 ml of incomplete Freund’s adjuvant (IFA; Bethesda Research Laboratories) together with 100 mg of muramyl dipeptide (Calbiochem) de- livered s.c. Rabbits were boosted 4 weeks later with 200 mg of recombinant KLK4 protein emulsified in 1 ml IFA, followed by i.v. every 4 weeks with 100 mg of soluble recombinant KLK4 protein. KLK4-specific antisera was purified by affinity chromatography against the E. coli-derived recombinant KLK4 antigen. Anti-KLK4 specific monoclonal antibodies were generated from the same immunized rabbits using the selected lymphocyte activation method (Babcook et al., 1996). For immunohistochemistry, paraffin-em- bedded formalin fixed tissue was sliced into eight micron sections. Steam heat induced epitope retrieval (SHIER) in 0.1 M sodium citrate buffer (pH 6.0) was used for optimal staining conditions. Sections were incubated with 10% serum/PBS for 5 min. Primary antibody was added to each section for 25 min at indicated concentrations followed by a 25 min incubation with either an anti-rabbit or anti-mouse biotinylated antibody. Endogenous peroxidase activity was blocked by three 1.5 min incubations with hydrogen peroxidase. The avidin biotin complex/horse radish peroxidase (ABC/HRP) system was used along with DAB chromogen to visua- lize antigen expression. Slides were counterstained with hematoxylin

Superscript II (Life Technologies). Relative expression levels of KLK4 for the given samples are shown in copies of KLK4/ng b- actin. (b) KLK4 mRNA expression levels in human prostate tumor tissues, prostate normal tissues and an extensive panel of normal tissues. Relative expression levels of KLK4 for the given samples are shown in copies of KLK4/ng b-actin

Oncogene Prostase characterization and expression in prostate CH Day et al 7118 at the predicted molecular weight (*30 kDa) of an Table 2 KLK4-specific Ab-positives in prostate cancer patients and NS1-KLK4 fusion protein consisting of the mature normal donors. Twenty sera from prostate cancer patients and 13 normal male donors have been assayed for the presence of KLK4 protein fused to the 8 kD NS1 protein, while antibodies specific to NS1-KLK4-Histidine-tag protein made in E. the pooled sera of normal donors did not reveal any coli. Wells of a 96-flat-well Nunc plate were coated with 200 ng of specific signal (Figure 3a). The rabbit polyclonal protein in 0.1 M carbonate buffer (pH 9.5), then blocked with 10% KLK4-specific antisera was used as a positive control, (w/v) non-fat dry milk (NFDM)/PBS. Patient and normal donor sera was diluted 300-fold with 10% goat sera/PBS, added to the reacting strongly with the NS1-KLK4 protein, but not wells and incubated at 48C overnight. In parallel, bovine serum NY-ESO-1 (Chen et al., 1997), a control protein albumin (BSA) was used as the negative control. The ELISA was (Figure 3b). In Figure 3c, Western blot analysis of developed with HRP-conjugated anti-human IgG and monitored at individual patient’s sera were shown, and patient sera 450 nm absorption. The signal values considered positive (+) are #1, #2 and #3 but not #4 contain Abs specific to the 53-times higher than the mean value obtained from the wells coated with BSA NS1-KLK4 fusion protein. These data have also been confirmed using baculoviral-KLK4 protein, which Prostate cancer KLK4-specific Normal KLK4-specific contained no fusion partner or tags (data not shown). patients antibody donors antibody It is therefore likely that the NS1-KLK4-specific Abs #1 + C1 7 observed in these prostate patients are specific to #2 + C2 7 KLK4 and not to NS1. #3 + C3 7 #4 7 C4 7 Next, the sera of 20 prostate cancer patients and 13 #5 7 C5 7 normal male donors were assessed by ELISA for the #6 7 C6 7 existence of antibodies specific to the NS1-KLK4 protein. #7 + C7 7 As a negative control, BSA was assayed in parallel. As #8 7 C8 7 #9 7 C9 7 shown in Table 2, the ELISA data demonstrated that #10 7 C10 7 seven of 20 prostate cancer patients have antibodies #11 7 C11 7 specific to the NS1-KLK4 antigen. Sera #1 through #4 #12 7 C12 7 correspond with those patients examined by Western blot #13 + C13 7 (Figure 3c), and the ELISA results correlate with those #14 + #15 + results. In marked contrast, none of the 13 normal donors #16 7 had NS1-KLK4-specific antibodies. #17 7 Lastly, in order to further confirm the specificity of #18 7 these Abs in patients, epitope-analysis using KLK-4 #19 7 #20 7 peptides was instigated using patient sera. As a Total 7/20 Total 0/13 negative control, normal donors’ sera were also set in

Figure 3 Western blot observation of KLK4-specific antibodies in prostate cancer patient sera but not in normal donor sera. Eight prostate cancer patient sera and eight normal male donor sera have been assayed for the existence of antibodies (Abs) specific to NS1-KLK4 fusion protein and NY-ESO-1-Histidine tag protein purified form E. coli. Each protein (200 ng/ lane) was electrophoresed, then blotted to nitrocellulose membrane. After blocking membranes with 10% NFDM/PBS, pooled sera or individual sera were incubated with the membranes. After washing with PBS/Tween and incubating with HRP-conjugated goat anti-human IgG, the membranes were developed by chemo-luminescence substrate. MM indicates molecular weight in kD

Oncogene Prostase characterization and expression in prostate CH Day et al 7119 parallel. Patient #1 and #7 have Abs specific to NS1- intestine. Additionally, we have demonstrated that this KLK4 (Figure 3c, Table 2), and both sera also expression is retained both in primary prostate tumors recognized Bac-KLK4 by Western blot analysis (data of various stage and grade and is also maintained in not shown). These sera and two sera from normal two metastatic prostate samples. donors were assessed for specific Abs by KLK4- The demonstration of KLK4 mRNA expression in peptide-based ELISA, wherein 96-well plates were all of the primary prostate tumors and metastatic coated with 0.5 mg of individual peptide covering the samples tested indicates a potential role for use of this KLK4 coding sequence (Nelson et al., 1999). As shown transcript for diagnosis and monitoring for residual in Figure 4a, both sera recognized one specific peptide disease by RT – PCR amplification from patient blood. (aa229-aa244). In contrast, the normal donor sera did Potential for this type of diagnosis has been demon- not have specific binding to this KLK4 peptide (Figure strated for the mammaglobin gene in metastatic breast 4b). cancer (Houghton et al., 2001; Watson et al., 1999). The recent discovery and reporting of KLK4 by four We have also examined various human tissue other groups (DuPont et al., 1999; Nelson et al., 1999; preparations with a KLK4-specific monoclonal anti- Stephenson et al., 1999; Yousef et al., 1999) has body by immunohistochemistry. This analysis resulted in some question of the specificity of mRNA indicates that the KLK4 protein expression deter- expression of this gene. Herein, we clarify the level of mined thus far has the same restricted expression mRNA expression in prostate tumors, normal prostate, pattern as that found for KLK4 mRNA. Both a and non-prostate normal tissues. Using a quantitative bioinformatic analysis protein prediction program PCR method and normalizing results using human (PSORT) (Nakai and Kanehisa, 1992) and the actin mRNA expression, we have determined that relatively high level of homology to secreted serine mRNA expression is significantly elevated in both proteases, argues that KLK4 is a secreted protein. The normal prostate and prostate tumors, compared to all demonstration of KLK4 secretion from transfected other normal tissues tested. However, low levels of HEK293 cells is consistent with this prediction. KLK4 mRNA expression were detected in three Although this study confirms the potential for normal human glandular tissues – adrenal, salivary KLK4 to be secreted, this has yet to be confirmed and thyroid. In contrast to data reported by Nelson et in either normal or malignant prostate tissues. Our al. (1999) we do not detect elevated expression in small attempts to detect KLK4 protein in LNCaP cells were

Figure 4 Epitope-analysis of KLK4-speciﬁc Abs in prostate cancer patients and normal donors’ sera. Prostate cancer patient sera #1 and #7 were diluted 300-fold with 5% goat sera and 5% NFDM-containing PBS. They were incubated in 96-well plate (Nunc, Denmark) wherein individual KLK4 peptides were coated with bicarbonate buffer (9.5) and blocked with 10% NFDM/PBS. After washing plates, HRP-conjugated anti-human IgG Abs (Jackson ImmunoResearch, West Grove, PA, USA) was used for developing ELISA as described in the legend to Table 2

Oncogene Prostase characterization and expression in prostate CH Day et al 7120 unsuccessful. It is also not known if KLK4 enters the Cancer Initiative Clinical Trials Grant from the Cancer circulation in individuals with a normally functioning Research Institute. Some antibodies provided in support prostate, benign prostatic hyperplasia, or prostate of this work were contributed by Dr Teresa Cabezon of cancer. While the secretion of the KLK4 protein has GlaxoSmithKline, Rixensart, Belgium. We thank Dr not been unambiguously established, the evidence Thomas S Vedvick, Dr Jiangchun Xu, Sandra Koseoglu and Joe Parsons of Corixa Corporation for expert protein presented here does indicate that the KLK4 protein characterization, cDNA library construction, mRNA is recognized by the immune system of prostate cancer analysis and DNA sequencing respectively. We thank Dr patients, with the generation of KLK4-speciﬁc anti- Samuel Li for baculoviral recombinant protein synthesis, body. and Dr Darrick Carter and Charisa Cornellison for bac- protein puriﬁcation, also of Corixa Corporation. Some tissue samples were obtained from the Cooperative Acknowledgements Human Tissue Network, which is funded by the National This work is supported in part by grant CA76822 from Cancer Institute, and from the National Disease Research the National Institutes of Health and by a Prostate Interchange.

References

Abi-Aad AS, Macfarlane MT, Stein A and deKernion JB. Myers SA and Clements JA. (2001). J. Clin. Endocrin. (1992). J. Urol., 147, 952 – 955. Metab., 86, 2323 – 2326. Babcook JS, Leslie KB, Olsen OA, Salmon RA and Schrader Nakai K and Kanehisa M. (1992). Genomics, 14, 897 – 911. JW. (1996). Proc. Natl. Acad. Sci. USA, 93, 7843 – 7848. Nelson PS, Gan L, Ferguson C, Moss P, Gelinas R, Hood L Chen YT, Scanlan MJ, Sahin U, Tureci O, Gure AO, Tsang and Wang K. (1999). Proc. Natl. Acad. Sci. USA, 96, S, Williamson B, Stockert E, Pfreundschuh M and Old LJ. 3114 – 3119. (1997). Proc.Natl.Acad.Sci.USA,94, 1914 – 1918. Obiezu CV, Scorilas A, Katsaros D, Massobrio M, Yousef Daher R and Beaini M. (1998). Clin. Chem. Lab. Med., 36, GM, Fracchioli S, Rigault de la Longrais R, Arisio R and 671 – 681. Diamandis EP. (2001). Clin. Cancer Res., 7, 2380 – 2386. Diatchenko L, Lau YF, Campbell AP, Chenchik A, Sissons GR, Clements R, Peeling WB and Penney MD. Moqadam F, Huang B, Lukyanov S, Lukyanov K, (1992). Br. J. Radiol., 65, 861 – 864. Gurskaya N, Sverdlov ED and Siebert PD. (1996). Proc. Stephenson SA, Verity K, Ashworth LK and Clements JA. Natl. Acad. Sci. USA, 93, 6025 – 6030. (1999). J. Biol. Chem., 274, 23210 – 23214. Diatchenko L, Lukyanov S, Lau YF and Siebert PD. (1999). Thiounn N, Saporta F, Flam TA, Pages F, Zerbib M, Methods Enzymol., 303, 349 – 380. Vieillefond A, Martin E, Debre B and Chevillard S. (1997). Dong Y, Kaushal A, Loan B, Chu S, Fuller PJ, Nicklin J, Urology, 50, 245 – 250. Samaratunga H and Clements JA. (2001). Clin. Cancer Watson MA, Dintzis S, Darrow CM, Voss LE, DiPersio J, Res., 7, 2363 – 2371. Jensen R and Fleming TP. (1999). Cancer Res., 59, 3028 – DuPont BR, Hu CC, Reveles X and Simmer JP. (1999). 3031. Cytogenet. Cell. Genet., 86, 212 – 213. YousefGM,ObiezuCV,LuoLY,BlackMHandDiamandis Grenlee RT, Murray T, Bolden S and Wingo PA. (2000). CA EP. (1999). Cancer Res., 59, 4252 – 4256. Cancer J. Clin., 50, 7–33. Houghton RL, Dillon DC, Molesh DA, Zehentner BK, Xu J, Jiang J, Schmidt C, Frudakis A, Repasky E, Filho AM, Nolasco M, Badaro R, Zhang X, Roche PC, Persing DH and Reed SG. (2001). Mol. Diagn., 6, 79 – 91.

Oncogene