and Prostatic Diseases (2010) 13, 369–375 & 2010 Macmillan Publishers Limited All rights reserved 1365-7852/10 www.nature.com/pcan ORIGINAL ARTICLE

Effect of androgen deprivation therapy on the expression of prostate cancer biomarkers MSMB and MSMB-binding CRISP3

A Dahlman1,2, A Edsjo¨ 2,5, C Hallde´n3, JL Persson4, SW Fine6, H Lilja3,7, W Gerald6 and A Bjartell1,2 1Department of Clinical Sciences, Division of Urological Cancers and Department of Laboratory Medicine, Lund University, University Hospital Malmo¨, Malmo¨, Sweden; 2Center of Molecular Pathology, Lund University, University Hospital Malmo¨, Malmo¨, Sweden; 3Division of Clinical Chemistry, Lund University, University Hospital Malmo¨, Malmo¨, Sweden; 4Division of Experimental Cancer Research, Lund University, University Hospital Malmo¨, Malmo¨, Sweden; 5University and Regional Laboratories Region Ska˚ne, Clinical Pathology, Malmo¨, Sweden; 6Department of Pathology, Memorial Sloan-Kettering Cancer Center, New York, NY, USA and 7Department of Clinical Laboratories, Memorial Sloan-Kettering Cancer Center, New York, NY, USA

We have investigated the effects of short-term neoadjuvant and long-term androgen deprivation therapies (ADTs) on b-microseminoprotein (MSMB) and cysteine-rich secretory protein-3 (CRISP3) expression in prostate cancer patients. We also studied if MSMB expression was related to genotype and epigenetic silencing. Using an Affymetrix cDNA microarray analysis, we investigated the expression of MSMB, CRISP3, androgen receptor (AR), KLK3 and Enhancer of Zeste Homologue-2 (EZH2) in tissue from prostate cancer patients receiving (n ¼ 17) or not receiving (n ¼ 23) ADT before radical prostatectomy. MSMB, CRISP3 and AR were studied in tissue from the same patients undergoing TURP before and during ADT (n ¼ 16). MSMB genotyping of these patients was performed by TaqMan PCR. MSMB and KLK3 expression levels decreased during ADT. Expression levels of AR and CRISP3 were not affected by short-term ADT but were high in castration-resistant prostate cancer (CRPC) and metastases. Levels of EZH2 were also high in metastases, where MSMB was low. Genotyping of the MSMB rs10993994 polymorphism showed that the TT genotype conveys poor MSMB expression. MSMB expression is influenced by androgens, but also by genotype and epigenetic silencing. AR and CRISP3 expression are not influenced by short-term ADT, and high levels were found in CRPC and metastases. Prostate Cancer and Prostatic Diseases (2010) 13, 369–375; doi:10.1038/pcan.2010.25; published online 3 August 2010

Keywords: microseminoprotein; PSP94; neoadjuvant; castration; tissue biomarker

Introduction tate-specific protein of 94 amino acids (PSP94). This protein is second only to PSA as the most predominant Prostate cancer is the most common form of cancer and a protein in seminal plasma.2,3 The function of MSMB is leading cause of cancer-related deaths in men in Western largely unknown, but in vitro and in vivo studies suggest countries.1 Patient prognosis will vary from a rapidly it may be involved in apoptosis, cell mobility, vascular- progressing disease with a high probability of death in a ization and tumor suppressor functions, all recently minority of patients, to a relatively indolent prostate reviewed by Whitaker et al.4 The suitability of MSMB as cancer that can be controlled for the remainder of the a biomarker for prostate cancer has been reported by patients life with little intervention in a majority of cases. several groups during the past two decades. It has been Despite identification of a number of promising new reported that MSMB mRNA and protein expression is biomarkers, there is still no absolute way of determining reduced in malignant prostatic epithelia compared with disease prognosis at the time of diagnosis. Inevitably, this benign epithelia.5 In prostate cancer, the MSMB may leads to overtreatment of a large number of men for the be subjected to transcriptional silencing by methylation, benefit of the few who need it. mediated by enhancer of zeste homologue-2 (EZH2), a Attention has been focused on the predictory abilities Polycomb group member which is often overexpressed in of b-microseminoprotein (MSMB); also known as pros- castration-resistant prostate cancer (CRPC).6,7 MSMB expression has been reported by several groups to be negatively associated with disease-free survival8 Correspondence: Dr A Bjartell, Department of Clinical Sciences, and outcome.9 MSMB expression has also been proposed Division of Urological Cancers and Department of Laboratory to be a significant prognostic indicator for prostate Medicine, Lund University, University Hospital Malmo¨, 205 02 cancer progression under endocrine therapy.10 Recently, Malmo¨, Sweden. E-mail: [email protected] the MSMB gene has gained further attention as one of Received 9 May 2010; accepted 22 May 2010; published online 3 the primary candidate prostate cancer susceptibility 11,12 August 2010 , and several causal risk alleles have been ADT affects MSMB and CRISP3 expression A Dahlman et al

370 identified in the region upstream of the coding bone (two samples); lung (one sample) and lymph node sequence.13–15 Serum levels of MSMB, its binding protein (three samples)) and metastatic CRPC (three samples, PSP94-binding protein (PSPBP) and the ratio free/bound sites not specified) were also included. Details of patient MSMB, have all been suggested as independent prog- characteristics have previously been described.25 nostic factors in prostate cancer.16–19 analysis were performed as described We have previously reported that the cysteine-rich previously.25 In brief, total RNA was extracted from secretory family (CRISP) family member CRISP3 also frozen tissue by Trizol (Invitrogen, Carlsbad, CA, USA), forms a high-affinity complex with MSMB in human purified and evaluated for integrity. Complementary seminal plasma.20 CRISP3 has been found to be one of DNA (cDNA) was synthesized from total RNA, and gene the most upregulated genes in prostate cancer compared expression analysis was performed using Affymetrix to benign tissue,21,22 and was proposed to be a useful U95 human gene arrays with 63 175 probes for genes and biomarker for prostate cancer.23,24 Recently we reported expressed sequence tags as described by the manufac- that in a tissue microarray (TMA) with samples from 945 turer (Affymetrix Inc., Santa Clara, CA, USA). Samples prostate cancer patients undergoing radical prostatect- were analyzed using Affymetrix Microarray Suite ver- omy (RP), high CRISP3 and low MSMB expression were sion 4.0 as previously reported.25 associated with poor outcome.9 As the impact of androgen availability on the expres- sion of certain tumor biomarkers is currently unclear, we Tissue specimens and clinical data from patients wanted to investigate how the candidate markers MSMB undergoing TURP and CRISP3 are regulated by androgens during andro- Formalin-fixed paraffin-embedded human tissue sam- gen deprivation therapy (ADT) and in progressive ples were obtained from 16 prostate cancer patients prostate cancer. First, we investigated the effect of short (mean age 70 years, range 52–78) undergoing repeated term neoadjuvant ADT on CRISP3 and MSMB gene therapeutic TURP at the University Hospital Malmo¨, expression in primary prostate cancer, as well as in a Sweden. Immunohistochemistry was performed as pre- limited set of prostate cancer metastases. CRISP3 and viously described,9 using antibodies and dilutions as MSMB protein levels were then evaluated in serial tissue described in Table 1. Descriptive characteristics of the samples from a small but unique cohort of men patients are given in Table 2. Samples were collected with prostate cancer undergoing TURP before and between 1971 and 2001. From each patient, samples were during long-term ADT. Finally, as MSMB expression collected at diagnosis, before ADT and from a later TURP may be differentially regulated in primary prostate due to local tumor progression. Nine patients were cancers and in progressive disease, we also studied subjected to medical castration with GnRH analogue, whether MSMB expression levels were related to a single one patient was treated with flutamide monotherapy and nucleotide polymorphism (SNP) in the MSMB promotor five patients underwent bilateral orchiectomy. Mean region and associated with epigenetic silencing of MSMB follow-up time was 73 months (range 12–167 months) by EZH2. and mean time of hormonal treatment was 37 months (range 3–118 months). Patients were considered to have progressive disease and CRPC at the second TURP date as levels of serum PSA or acid phosphatase were rising Materials and methods despite ongoing ADT. The study was approved by the local ethics committee at Lund University, Sweden. Sample preparation and data analysis of Affymetrix U95 human gene array Microarray tissue samples were obtained as previously Genotype analysis described.25 In brief, tissue samples were collected from Paraffin-embedded tissues from 13 of the 16 patients patients undergoing therapeutic or diagnostic proce- undergoing TURP were available for genotype analysis. dures at Memorial Sloan-Kettering Cancer Center, New Fresh sections were microscopically studied, and tissue York, NY. All samples were snap-frozen, histologically was manually punched out of the paraffin block with analyzed, and tissues containing approximately 60–80% a 2 mm biopsy needle. DNA was isolated using prostate cancer were manually dissected from the frozen the QIAamp DNA FFPE tissue kit (Qiagen, Hilden, block. Samples included 23 primary prostate cancer Germany) according to the manufacturer’s protocol. The tissues from patients not receiving adjuvant therapy rs10993994 polymorphism was determined using Taq- before RP, 17 primary prostate cancer tissues from Man primers and probes (Applied Biosystems, Foster patients collected at RP after three months of neoadju- City, CA, USA), using 1 ng per ml template DNA in a vant ADT (monthly injections of 3.6 mg of goserelin and reaction volume of 25 ml. The analysis was performed on 250 mg flutamide three times daily). A limited set of a 7900HT Fast Real-Time PCR System (Applied Biosys- 9 prostate cancer metastases (the secondary sites were in tems) according to the manufacturer’s instructions.

Table 1 Antibody characteristics Antibody/clone Species/type Dilution Retrieval Manufacturer/reference

MSMB Rabbit, polyclonal 1:40 000 MW 3 CRISP3 Rabbit, polyclonal, affinity purified 3.5 mgmlÀ1 MW 20 AR, A441 Mouse, monoclonal 1:300 MW Neomarker (Thermo Fisher Scientific Inc., Fremont, CA, USA)

Abbreviation: AR, androgen receptor; CRISP3, cysteine-rich secretory protein-3; MSMB, b-microseminoprotein; MW, microwave-mediated retrieval.

Prostate Cancer and Prostatic Diseases ADT affects MSMB and CRISP3 expression A Dahlman et al 371 Table 2 Patient description and tumor sample characteristics Case Age at Gleason TNM stage Local Biochemical Time between TURPs diagnosis score progression progression (months) At time of diagnosis Before ADT During ADT During ADT During ADT

1 74 4 T1NxMx T3NxMx Yes mSerum PSA 42 2 73 6 T1NxMx T1NxM0 Yes NA 94 3 78 7 T2NxM0 T3NxM0 Yes mSerum PSA 51 4 66 7 T1N1M0 T3N1M1 Yes mSerum PSA 167 5 75 9 T1NxM0 T3NxM1 Yes mSerum PAP 54 6 67 6 T1NxM0 T3NxM0 Yes NA 71 7 74 6 T1NxM0 T3NxM0 Yes mSerum PAP 119 8 74 8 T1NxM0 T3NxM1 Yes mSerum PAP 108 9 52 7 T1NxM0 T4NxM0 Yes mSerum PAP 136 10 76 8 T1NxM0 T1NxMx Yes mSerum PSA 35 11 74 7 T1NxM0 T3NxM1 Yes mSerum PSA 12 12 62 8 T2NxM0 T2NxM1 Yes mSerum PAP 80 13 78 6 T2NxM0 T3NxM0 Yes mSerum PSA 41 14 52 7 T1NxM0 T3NxM0 Yesa mSerum PAP 48 15 77 7 T2NxM0 T3NxM0 Yes mSerum PSA 58 16 73 7 T2NxMx T2NxMx Yes mSerum PSA 84

Abbreviations: ADT, androgen deprivation therapy; NA, data not available. aTreated with estradiol injections during the period between TURP.

Results Although caution is a prerequisite when interpreting small sample numbers, we found that MSMB expression MSMB but not CRISP3 transcript levels decrease in is substantially decreased in metastatic tumors compared response to short-term neoadjuvant ADT with primary prostate cancer, whereas CRISP3 is Changes in expression of MSMB and CRISP3 genes due expressed at high levels in metastases (Figure 1). Inter- to androgen deprivation were analyzed using data from estingly, CRISP3 expression pattern appears to be some- Affymetrix U95 gene expression microarrays. The clin- what similar to that of AR in that both transcipts are ical and pathological features related to the patients expressed at the same level and neither AR nor CRISP3 is and tissue samples included in the study (patient age, affected by ADT (Figure 1). preoperative serum PSA levels, Gleason score, patholo- gical TNM stage and recurrence) have been described 25 Protein expression of MSMB and CRISP3 during ADT and earlier. MSMB expression was found to be high in primary prostate cancer from patients not receiving prostate cancer progression adjuvant ADT, and expression was significantly de- To evaulate the link between long-term ADT, progressing creased on ADT (Figure 1; Mann–Whitney Po0.001). disease and the expression levels of MSMB and CRISP3 No significant reduction in gene expression levels , we used a small but unique set of serially was observed in either CRISP3 or androgen receptor collected tissue samples from 16 patients undergoing (AR) transcript levels in patients receiving neoadjuvant TURP before and during ADT. At the time of the second ADT. However both CRISP3 and AR were highly TURP, all patients were in biochemical progress and expressed in metastatic prostate cancer lesions considered to be in a CRPC stage (Table 2). Tissue (Figure 1). There was considerable interpatient varia- sections were immunhistochemically stained for MSMB, bility in expression of both MSMB and CRISP3. Interest- CRISP3 and AR, and protein expression was evaluated ingly, the MSMB expression level and decrease during by a pathologist for both intensity and percentage of ADT were equal to those of KLK3, a well-known target of positive tumor cells. Representative staining is depicted androgen signaling (Figure 1). As expected, KLK3 gene in Figure 2. expression levels decreased in the group receiving A majority of the patients had very low MSMB levels neoadjuvant ADT compared with the nontreated already at the time of the first TURP, and decreased group (Mann–Whitney Po0.001). However, as in the expression was therefore not readily detected (Figure 3). case of MSMB, considerable interpatient variability was In the few patients with 450% tumors cells positive for detected. MSMB at the time of the first TURP, there was a dramatic decrease in expression at the time for the second TURP. No patient had more than 45% tumor cells positive for MSMB at a CRPC stage. Tumor tissue from these CRPC Prostate cancer metastases express low transcript levels of patients during long-term ADT showed higher levels of MSMB, and high levels of CRISP3 CRISP3 and AR compared with tissue from primary A majority of prostate cancer patients develop CRPC tumors from the same patient at the time of diagnosis within a few years following ADT. Such tumors often (Figure 3). show highly upregulated levels of AR though the Increased expression levels of CRISP3 and AR were signaling may be dysfunctional.26,27 To investigate found in terms of an increased fraction of positive cells. MSMB and CRISP3 transcriptional levels in metastases, MSMB and CRISP3 staining intensity did not change we studied gene expression in nine remote tumors, out with progressing disease, but AR staining was stronger of which three were from patients with CRPC. at the time for the second TURP. The decreased MSMB

Prostate Cancer and Prostatic Diseases ADT affects MSMB and CRISP3 expression A Dahlman et al 372

Figure 1 Transcription profile changes before and after androgen deprivation therapy (ADT) in primary prostate cancer tumors and in metastases. Patients with primary prostate cancer undergoing routine surgery either received (n ¼ 17) or did not receive (n ¼ 23) neoadjuvant ADT before surgery. Transcription is represented as graphs and box plots for b-microseminoprotein (MSMB), cysteine-rich secretory protein-3 (CRISP3), androgen receptor (AR) (reproduced from Am J Pathol 2004, 164:217–227 with permission from the American Society for Investigative Pathology), KLK3 and Enhancer of Zeste Homologue-2 (EZH2). The cDNA array included two probe sets for AR, and the corresponding boxplot is based on the average.

protein expression on long-term ADT is corroborating previously immunohistochemically examined for MSMB the studies at the transcriptional level, and is also expression. TaqMan genotyping identified the TT geno- emphasizing that low MSMB expression is a feature of type previously associated with lower MSMB expression aggressive disease, whereas CRISP3 and AR show the in two patients. Both patients were showing low levels of opposite with high expression levels in CRPC. MSMB. The remaining 11 patients all had CC or CT genotypes, and taken as a group, a higher level of MSMB expression (Figures 3 and 4). Low MSMB protein levels are associated with the TT genotype of SNP rs10993994 High transcriptional levels of the EZH2 gene are associated The MSMB promoter contains an SNP reported to significantly affect MSMB gene expression. To investi- with low MSMB levels in progressive disease and gate the impact of this SNP on MSMB expression in our metastases experimental setup, we performed TaqMan genotyping MSMB expression before and during ADT varies con- of the rs10993994 locus on 13 of the 16 patients siderably between patients, and cannot be explained by

Prostate Cancer and Prostatic Diseases ADT affects MSMB and CRISP3 expression A Dahlman et al 373

Figure 2 Immunohistochemical staining showing expression pattern of b-microseminoprotein (MSMB), cysteine-rich secretory protein-3 (CRISP3) and androgen receptor (AR) in TURP specimens derived from the same patient undergoing diagnostic TURP before androgen deprivation therapy (ADT, upper panel), and later during ADT when TURP was performed due to local tumor progression at a castration- resistant prostate cancer stage (lower panel). Images were generated at an original magnification of  20.

Figure 4 Correlation of the rs10993994 genotype and the corre- sponding percentage of b-microseminoprotein (MSMB)-positive tumor cells in tissue collected from patients undergoing TURP before initiation of androgen deprivation therapy. Figure 3 Bar charts representing the fraction of cancer cells positively immunostained for b-microseminoprotein (MSMB), cysteine-rich secretory protein-3 (CRISP3), and androgen receptor (AR) in TURP samples from patients undergoing TURP before receiving any treatment (gray bars), and later at tumor progression during androgen deprivation therapy (ADT, black bars). The Discussion rs10993994 single nucleotide polymorphism genotype for each patient is inserted below the upper chart, illustrating MSMB protein In this study, we wanted to examine impact of short- and expression. *Tumor material missing. long-term ADT on prostate cancer outcome predictors MSMB and the MSMB-binding protein CRISP3. We show that short-term ADT significantly reduces both transcript either androgen availability or genotype. Previous and protein levels of MSMB, whereas CRISP3 levels do studies have shown that methylation, mediated by the not change. Conversely, CRISP3 is upregulated in Polycomb group member EZH2, is yet another way by parallell with AR and KLK3 in progressive disease and which MSMB expression may be regulated. Therefore, during long-term ADT. we analyzed the transcriptional levels of EZH2 in the In concordance with studies conducted by others, the Affymetrix U95 gene expression microarray described results of our study support that MSMB expression above. Although this is a limited sample set, we found decrease in patients receiving neoadjuvant ADT indicat- that the EZH2 gene was highly expressed in metastatic ing androgen dependent expression.28 Importantly, and lesions from prostate cancer, where MSMB expression in contrast to KLK3, MSMB expression is continuously levels were very low. EZH2 gene expression does not low in progressive and metastatic disease (Figure 1). appear to be affected by neoadjuvant ADT (Figure 1). Rising level of PSA is considered a hallmark for disease

Prostate Cancer and Prostatic Diseases ADT affects MSMB and CRISP3 expression A Dahlman et al 374 progression. This indicates that despite similarities in 3 Abrahamsson PA, Andersson C, Bjork T, Fernlund P, Lilja H, androgen effect on KLK3 and MSMB expression in the Murne A et al. Radioimmunoassay of beta-microseminoprotein, normal prostate and primary prostate cancer, it is a prostatic-secreted protein present in sera of both men and obvious that they are differentially regulated in progres- women. Clin Chem 1989; 35: 1497–1503. sive disease. This could be due to silencing of MSMB 4 Whitaker HC, Warren AY, Eeles R, Kote-Jarai Z, Neal DE. The expression by EZH2-mediated methylation of the MSMB potential value of microseminoprotein-beta as a prostate cancer Prostate 70 promotor.6 We find high levels of EZH2 in metastatic biomarker and therapeutic target. 2009; : 333–340. 5 Tsurusaki T, Koji T, Sakai H, Kanetake H, Nakane PK, Saito Y. lesions, which is consistent with our findings of low Cellular expression of beta-microseminoprotein (beta-MSP) MSMB expression in those tumors (Figure 1). ADT did mRNA and its protein in untreated prostate cancer. Prostate not appear to affect EZH2 levels, thus methylation is 1998; 35: 109–116. unlikely linked to the rapid MSMB downregulation 6 Varambally S, Dhanasekaran SM, Zhou M, Barrette TR, observed in hormonally treated primay prostate cancer. Kumar-Sinha C, Sanda MG et al. The polycomb group protein Yet another way by which MSMB can be regulated is EZH2 is involved in progression of prostate cancer. Nature 2002; by an SNP in the promotor region. In our study, the TT 419: 624–629. allele was only observed in 2 of 13 patients (Figures 3 7 Beke L, Nuytten M, Van Eynde A, Beullens M, Bollen M. The and 4). Both patients had very low levels of MSMB, but gene encoding the prostatic tumor suppressor PSP94 is a target not lower than many patients carrying the CC or CT for repression by the Polycomb group protein EZH2. Oncogene 2007; 26: 4590–4595. alleles. Despite much recent interest in the rs10993994 8 Girvan AR, Chang P, van Huizen I, Moussa M, Xuan JW, Stitt L SNP genotype, our findings indicate that it does not fully et al. Increased intratumoral expression of prostate secretory account for the low MSMB levels generally observed in protein of 94 amino acids predicts for worse disease recurrence prostate cancer patients with progressive disease. and progression after radical prostatectomy in patients with Our study data showed that neoadjuvant therapy does prostate cancer. Urology 2005; 65: 719–723. not appear to affect expression of CRISP3. However, the 9 Bjartell AS, Al-Ahmadie H, Serio AM, Eastham JA, Eggener SE, high CRISP3 expression in metastatic and recurrent Fine SW et al. Association of cysteine-rich secretory protein 3 and tumors may be indicative of a role for CRISP3 in the beta-microseminoprotein with outcome after radical prostatect- progression of prostate cancer. omy. Clin Cancer Res 2007; 13: 4130–4138. In this study, we find that MSMB expression may be 10 Sakai H, Tsurusaki T, Kanda S, Koji T, Xuan JW, Saito Y. regulated by several mechanisms, including androgen Prognostic significance of beta-microseminoprotein mRNA expression in prostate cancer. Prostate 1999; 38: 278–284. deprivation. In all, MSMB being suggested to function as 11 Thomas G, Jacobs KB, Yeager M, Kraft P, Wacholder S, Orr N a candidate tumor suppressor, it is most thought- et al. Multiple loci identified in a genome-wide association study provoking to note that according to this study, MSMB of prostate cancer. Nat Genet 2008; 40: 310–315. is downregulated by hormonal treatment and subse- 12 Eeles RA, Kote-Jarai Z, Giles GG, Olama AA, Guy M, quently it may be silenced by different mechanisms. The Jugurnauth SK et al. Multiple newly identified loci associated results of our study favor the view of MSMB being a with prostate cancer susceptibility. Nat Genet 2008; 40: 316–321. suitable marker for prostate cancer disease progression 13 Chang BL, Cramer SD, Wiklund F, Isaacs SD, Stevens VL, Sun J in CRPC. et al. Fine mapping association study and functional analysis implicate a SNP in MSMB at 10q11 as a causal variant for prostate cancer risk. Hum Mol Genet 2009; 18: 1368–1375. Conflict of interest 14 Lou H, Yeager M, Li H, Bosquet JG, Hayes RB, Orr N et al. Fine mapping and functional analysis of a common variant in MSMB The authors declare no conflict of interests. on 10q11.2 associated with prostate cancer suscept- ibility. Proc Natl Acad Sci USA 2009; 106: 7933–7938. 15 Kote-Jarai Z, Leongamornlert D, Tymrakiewicz M, Field H, Guy M, Al Olama AA et al. Mutation analysis of the MSMB gene in Acknowledgements familial prostate cancer. Br J Cancer 2009; 102: 414–418. 16 Wu D, Guo Y, Chambers AF, Izawa JI, Chin JL, Xuan JW. We thank Jeff Holzbeierlein and the American Society for Serum bound forms of PSP94 (prostate secretory protein of 94 Investigative Pathology for permission to reprint the amino acids) in prostate cancer patients. J Cell Biochem 1999; 76: microarray data on AR expression; Elise Nilsson for 71–83. excellent technical skills in the field of immunohisto- 17 Reeves JR, Xuan JW, Arfanis K, Morin C, Garde SV, Ruiz MT et al. chemistry; Lene Udby for providing the CRISP3 anti- Identification, purification and characterization of a novel human blood protein with binding affinity for prostate secretory body and Per Fernlund for the MSMB antibody. protein of 94 amino acids. Biochem J 2005; 385 (Part 1): 105–114. Supporting grants: European Union 6th Framework 18 Reeves JR, Dulude H, Panchal C, Daigneault L, Ramnani DM. (P-Mark) (Grant number LSHC-CT-2004-503011), Swedish Prognostic value of prostate secretory protein of 94 amino acids Cancer Society, Swedish Research Council (Medicine), and its binding protein after radical prostatectomy. Clin Cancer Cancer and Research Foundation at University Hospital Res 2006; 12 (20 Part 1): 6018–6022. Malmo¨, Gunnar Nilsson Cancer Foundation and the 19 Nam RK, Reeves JR, Toi A, Dulude H, Trachtenberg J, Emami M Crafoord Foundation. et al. A novel serum marker, total prostate secretory protein of 94 amino acids, improves prostate cancer detection and helps identify high grade cancers at diagnosis. JUrol2006; 175: 1291–1297. References 20 Udby L, Lundwall A, Johnsen AH, Fernlund P, Valtonen-Andre C, Blom AM et al. beta-Microseminoprotein binds CRISP-3 in 1 Jemal A, Siegel R, Ward E, Hao Y, Xu J, Murray T et al. Cancer human seminal plasma. Biochem Biophys Res Commun 2005; 333: statistics, 2008. CA Cancer J Clin 2008; 58: 71–96. 555–561. 2 Lilja H, Abrahamsson P-A. Three predominant proteins secreted 21 Asmann YW, Kosari F, Wang K, Cheville JC, Vasmatzis G. by the human prostate gland. Prostate 1988; 12: 29–38. Identification of differentially expressed genes in normal and

Prostate Cancer and Prostatic Diseases ADT affects MSMB and CRISP3 expression A Dahlman et al 375 malignant prostate by electronic profiling of expressed sequence 25 Holzbeierlein J, Lal P, LaTulippe E, Smith A, Satagopan J, Zhang L tags. Cancer Res 2002; 62: 3308–3314. et al. Gene expression analysis of human prostate carcinoma during 22 Udby L, Bjartell A, Malm J, Egesten A, Lundwall A, Cowland JB hormonal therapy identifies androgen-responsive genes and et al. Characterization and localization of cysteine-rich secretory mechanisms of therapy resistance. Am J Pathol 2004; 164: 217–227. protein 3 (CRISP-3) in the human male reproductive tract. 26 Chen CD, Welsbie DS, Tran C, Baek SH, Chen R, Vessella R et al. J Androl 2005; 26: 333–342. Molecular determinants of resistance to antiandrogen therapy. 23 Kosari F, Asmann YW, Cheville JC, Vasmatzis G. Cysteine-rich Nat Med 2004; 10: 33–39. secretory protein-3: a potential biomarker for prostate cancer. 27 Tran C, Ouk S, Clegg NJ, Chen Y, Watson PA, Arora V et al. Cancer Epidemiol Biomarkers Prev 2002; 11: 1419–1426. Development of a second-generation antiandrogen for treatment 24 Bjartell A, Johansson R, Bjork T, Gadaleanu V, Lundwall A, of advanced prostate cancer. Science 2009; 324: 787–790. Lilja H et al. Immunohistochemical detection of cysteine-rich 28 Imasato Y, Xuan JW, Sakai H, Izawa JI, Saito Y, Chin JL et al. secretory protein 3 in tissue and in serum from men with cancer PSP94 expression after androgen deprivation therapy: a com- or benign enlargement of the prostate gland. Prostate 2006; 66: parative study with prostate specific antigen in benign prostate 591–603. and prostate cancer. J Urol 2000; 164: 1819–1824.

Prostate Cancer and Prostatic Diseases