Prostate targeting: PSP94 gene promoter/enhancer region directed prostate tissue-specific expression in a transgenic mouse prostate cancer model

Gene Therapy (2002) 9, 1589–1599  2002 Nature Publishing Group All rights reserved 0969-7128/02 $25.00 www.nature.com/gt

RESEARCH ARTICLE

Prostate targeting: PSP94 gene promoter/enhancer region directed prostate tissue-specific expression in a transgenic mouse prostate cancer model MY Gabril1,3, T Onita1,5, PG Ji1, H Sakai5, FL Chan4, J Koropatnick2,3, JL Chin1, M Moussa3 and JW Xuan1,3 1

Department of Surgery, University of Western Ontario, Canada; 2Department of Microbiology and Immunology, University of Western Ontario, Canada; 3Department of Pathology, Western Ontario, Canada; 4Department of Anatomy, The Chinese University of Hong Kong; and 5Department of Urology, Nagasaki University School of Medicine, Nagasaki, Japan

To date, only a few prostate-specific vector genes have been tested for prostate targeting in gene therapy of prostate cancer (CaP). Current clinical trials of gene therapy of CaP utilize the only two available vector genes with a combination of a rat probasin promoter and a human PSA promoter sequence in an adenovirus vector to target CaP. There is an urgent need to establish additional vector gene systems to sustain and propagate the current research. Since PSP94 (prostate secretory protein of 94 amino acids) is one of the three most abundant proteins secreted from the human prostate and is generally considered to be prostate tissuespecific in both human and rodents, we performed a transgenic experiment to assess the promoter/enhancer region of PSP94 gene-directed prostate targeting. Firstly, a series of progressive deletion mutants of a 3.84 kb PSP94 gene promoter/enhancer region (including parts of the intron 1 sequence) linked with a reporter LacZ gene was constructed and assessed in vitro in cell culture. Next, transgenic mice were generated with two transgene constructs using the SV40 early region (Tag oncogene) as a selection marker.

PSP94 gene promoter/enhancer region-directed SV40 Tag expression specifically in the mouse was demonstrated in three breeding lines (A, B, C, n = 374) by immunohistochemistry staining of Tag expression. Specific targeting to the prostate in the PSP94 gene-directed transgenic CaP model was characterized histologically by correlation of SV40 Taginduced tumorigenesis (tumor grading) with puberty and age (10–32 weeks). Prostatic hyperplasia was observed as early as 10 weeks of age, with subsequent emergence of prostatic intraepithelial neoplasia (PIN) and eventually high grade carcinoma in the prostate. The PSP94 transgenic mouse CaP model was further characterized by its tumor progression and metastatic tendency at 20 weeks of age and also by its responsiveness and refractoriness to androgen manipulation. This study indicates that the PSP94 gene promoter/enhancer has the potential for prostate specific targeting and may ultimately be of use in gene therapy of CaP. Gene Therapy (2002) 9, 1589–1599. doi:10.1038/sj.gt.3301895

Keywords: PSP94/␤-microseminoprotein; prostate targeting; transgenic mouse; PSP-transgenic mice

Introduction The risk of prostate cancer (CaP), the most common cancer among adult males in North America, increases steeply with age by a projected 3–4% yearly and is the second leading cause of cancer death, for a review see Ref. 1. Currently surgery, radiotherapy and endocrine therapy either as mono-therapy or in combination, have not been totally effective in locally high-stage cancer, which ultimately leads to distant metastasis and death. To reduce morbidity and mortality of CaP, other novel treatment approaches to advanced or recurrent CaP are desperately needed.

Correspondence: JW Xuan, Urology Research Laboratory, London Health Sciences Center, 375 South Street, London, Ontario, Canada, N6A 4G5 The first two authors contributed equally to this work Received 6 March 2002; accepted 23 June 2002

The first critical component for gene therapy of CaP is a promoter/enhancer capable of directing prostate tissuespecific expression of therapeutic tumor suppressor genes in CaP. In this regard, several purported (nonprostate) tissue-specific promoters/enhancers have been tested for targeting the prostate in transgenic mouse experiments. These include the mouse mammary tumor virus (MMTV) long terminal repeat (LTR),2 mouse cryptdin-2 (CR-2) gene,3 human fatal G␥-globin,4 bovine keratin 5 promoter5 and gp91-phox6 (for review see Refs 7 and 8). However none of these promoter/enhancers have been found to be adequate for targeted specifically to the prostate (for reviews see Ref 7–9). Only a few prostate tissue-specific genes have been tested for targeting of heterologous genes in the transgenic mouse prostate. The rat prostate steroid-binding protein (PSBP or C3(1))10–12 gene was characterized in transgenic mice to target the prostate and mammary gland. Prostatic-specific antigen (PSA) was characterized for several prostate tissue-specific elements (TSE) by in

PSP94 gene promoter directed prostate targeting MY Gabril et al

1590

vitro tissue cell culture studies.13–15 A 657 bp promoter of the human PSA gene linked with T24-ras in transgenic mice showed tumor induction in the salivary gland and gastrointestinal tract.16 Two other PSA transgenes were constructed based on a 6 kb promoter/enhancer region17 and the whole human PSA gene (12 kb), and were reported to be essentially prostate-specific with LacZ as a reporter gene.17,18 The PSA promoter region has not yet been used to target prostatic function in transgenic mice.8 The transgenic adenocarcinoma mouse prostate (TRAMP)19–21 and the LPB-Tag22,23 models are all based on a prostate-specific gene, rat probasin (rPB), for targeting to the prostate and directing the SV40 Tag (SV40 T and t antigens) expression. The rPB gene as a targeting vector represents a major advance in the establishment of a mouse model for prostate cancer research. However, its targeting to non-prostate tissues has also been reported.24 Recently, an osteocalcin promoter was used in a novel strategy for cotargeting both tumor epithelial and bone stromal cells in gene therapy of androgen-independent CaP bone metastasis.25,26 Current gene therapy clinical trials of prostate cancer utilize a combination of the only two available targeting vector genes with both the rat probasin promoter sequence and human PSA promoter sequence13,20,27 without the benefit of prior testing in an appropriate transgenic animal model.8,28,29 Both vector genes of rat PB and human PSA have been challenged by the fact that neither a human counterpart of the rPB gene nor rodent counterpart of the human PSA gene has been clearly identified due to interspecies divergence. Thus, there is a need to establish additional vector gene systems to sustain and to propagate research on prostate targeting and gene therapy of CaP. For the purpose of gene therapy, the ideal candidate vector gene for targeting to the prostate should be a pluripotent targeting vector suitable for and effective in both animal models and human clinical studies, so that animal studies can be replicated in clinical trials. In this study, we used mouse PSP94 (prostate secretory protein of 94 amino acids), also known as ␤-MSP (microseminoprotein), a prostate-specific gene, as a targeting vector to the prostate. Along with PSA (also called ␤-MSP) and prostatic acid phosphatase (PAP), for a review see Ref. 30, PSP94 is one of the three most abundant (mg/ml) secretory proteins of the prostate gland. As with human PSA, the utility of PSP94 as a CaP biomedical and histological marker has been studied.31–33 Serumfree and bound forms of PSP94 were identified31 and a new differential test between free and bound forms of PSP94 showed an improved prognostic value.31,34 Studies in human, primate, pig, rat and mouse PSP94 have shown that PSP94, unlike rPB and PSA, is a conserved, but rapidly evolving protein, for a mini-review see Ref. 35. Since mouse PSP94 is expressed specifically in the prostate,33,35 we performed a transgenic mouse experiment using a 3.8 kb PSP94 gene promoter/enhancer region to explore its utility for prostate targeting.

Results In vitro characterization of the promoter/enhancer region of mouse PSP94 in tissue cultures of prostate cancer cell lines In order to determine if the 3.84 kb PSP94 promoter/ enhancer region is capable of directing prostate tissueGene Therapy

specific transgene expression, a series of progressive deletions of this region fused with a reporter gene (LacZ) were constructed (see Figure 1) and tested in transient transfectants in a human prostate cancer cell line LNCaP. As shown in Figure 1a, starting from 3.84 kb, down to a 2.5 kb (BglII) deletion construct, the PSP94 promoter/enhancer region maintained the ability to direct LacZ expression in LNCaP transient transfectants. This transient expression required co-transfection of the androgen receptor cDNA at a two times higher dose than transgene constructs, as compared with results without co-transfection. The internal positive control was pSVgal; the negative control groups were: no DNA, mouse androgen receptor cDNA only, and PstI and HindIII deletion constructs. Although the positive transfectants (X-gal staining) in the two constructs were easily repeatable, the frequency was very low, at about 1/50 to 1/100 of the frequency of positive controls (pSVgal, positive staining rate at ~10 %), ie 5 × 10⫺4–10⫺4. In other murine, nonprostate cancer cell lines (NIH 3T3 fibroblasts and C127 epithelial mammary tumour-derived cells), PSP94 did not drive reporter expression. The intron 1 sequence was tested in transfectant construct 167 (Figure 1b) to assess if the long (~6 kb) intron closest to the PSP94 promoter region35could enhance the transfection frequency, and results of in vitro test in LNCaP cells were positive, but only with slightly higher frequencies. Northern blotting experiments with multiple mouse tissue total RNA

Figure 1 (a) Construction and testing of progressive deletions of a 3.842 kb promoter/enhancer region of mouse PSP94 gene linked with a reporter gene LacZ. Positive (+) and negative (⫺) results are indicated on right side. Deletion mutants were indicated by -bp numbers upstream of the first transcriptional initiation site (+1). KpnI site is in the first exon at +16bp. (b) Testing of both promoter/enhancer region and part of intron 1 sequences with a reporter gene LacZ (shown plasmid structure of 167). The wild-type intron 1 structure was compared with part of intron 1 subcloned in this construction. The linker sequence of 3⬘-end of intron 1 and LacZ (␤-galactosidase) coding region is indicated on the bottom, which contains the restriction sites: HindIII (exon 2 of PSP94)-ClaIHincII/BsaA1. BsaA1 site of LacZ DNA sequence located at seven codons upstream of the third possible translational initiation Met codon. The LacZ DNA sequence is listed in italic and the reading frame is also indicated by triplets (last line). B, BamHI; Bg, BglII; H, HindIII; K, KpnI; P, PstI; R, EcoRI.

PSP94 gene promoter directed prostate targeting MY Gabril et al

samples were performed with the whole intron 1 probe35 to confirm that intron 1 did not code other structural genes (data not shown). Establishment and characterization of three PSP94gene directed transgenic mouse CaP breeding lines targeted with the expression of SV40 Tag oncogene Since results from in vitro tissue cell culture analyses indicated the possibility of a 3.84 kb PSP94 promoter/enhancer region-driven prostate targeting in vivo, transgenic mouse experiments were performed. SV40 Tag was used as a selection marker gene for prostate targeting testing, since any non-prostate targeting events in transgenic mice could be easily detected by gross pathology at necropsy. One transgene (183, Figure 2a) was constructed using a 3.84 kb mouse PSP94 gene promoter/enhancer region to direct targeting of expression of a full length of SV40 Tag (T and t antigens) gene to the prostate. For enhancing transgenic expression by intron sequences,36 a second transgene (186, Figure 2b) was constructed linking the promoter/enhancer-exon1part of intron 1 and exon 2 followed by the SV40 Tag gene with the same reading frame. After breeding with the wild type of (C57BL/6 × CBA) F1 hybrids, three transgenic founder breeding lines (F0 of A, B and C) were characterized and successfully established out of 18 transgenic mice: A from transgene 183 (183-2), B and C from 186 (named F0186-3, F0186-9, respectively). The copy numbers of the three transgenic mouse founder breeding lines (F0A, B and C) and F1, F2, F3 progenies were characterized by Southern blotting experiments

Figure 2 Construction of prostate targeting transgene 183 (a) and 186 (b) for transgenic mice study. PSP94 promoter/enhancer region was linked with SV40 Tag (both T and t) gene. Whole transgene of 183 is in a 6.4 kb XbaI fragment and two inside PstI (P) sites were indicated. Locations of primer pairs for identification of transgenic mice are indicated by arrows. Part of the intron 1 sequences were subcloned in the same way as in Figure 1b. The linker sequence of 3’-end of intron 1 and SV40 Tag coding region is indicated on the bottom of b, which contains the restriction sites HindIII (exon 2 of PSP94) ClaI-HincII/StuI. StuI site of SV40Tag is 27 bp upstream of the first Met codon of Tag (in italics) and followed by another two Met codons (reading frame in triplets). Intron and exon sequences were indicated by white and black bars separately. The whole transgene is in a 8.2 kb XbaI fragment and the three PstI (P) sites were indicated.

using a full length SV40 Tag gene (~2.7 kb) as a probe. As shown in Figure 3, transgenic mouse line A had approximately 100 copies of transgene integrated into the mouse chromosome, and B had a higher copy number (approximately 150 copies per genome), while C founder mouse (186-9) and some offspring (generations F1, F2, F3) possessed five times more copies (approximately 500 copies/genome). Transmittance of this higher copy transgene in total breeding offspring occurred at a ratio (high copy/total) of: 5/25 (20%) in F1, and 6/108 (5.5%) in F2, and 0/41 in F3 respectively, indicating a transmittance mechanism of non-sex chromosomal linkage, non-Mendelian segregation. Targeting of transgene expression in three transgenic breeding lines (210 male and 164 female) to the prostate was demonstrated. First, autopsy studies of mice in breeding lines A (transgene 183) and B (186-2)(shown in Figure 4a and b) were performed revealing that SV40 Tag expression induced morphological and histological changes detectable only in the ventral and dorsolateral prostate lobes, while non-prostate targeting including the seminal vesicle and coagulation gland, was not found. This pattern of expression was demonstrated in two mouse populations: A line 116 mice (64 male, 52 female), for a period of 16 weeks up to 52 weeks of age (Figure 4a); B line 84 mice (47 male, 37 female), for a period of 12 weeks up to 52 weeks of age (Figure 4b). Founder mouse F0 (186-9) of line C (Figure 4c) with extraordinary high copies of transgene was the only mouse founder which showed non-prostate targeting tumor formation in the mesentery at autopsy (at 23 weeks), while only PIN was found in the prostate tissue, as demonstrated by Western blotting and immunohistochemistry experiments (data not shown). Total non-prostate targeting proportion in line C mice was found in 4.3% (9/210) (Figure 4c and d). Twenty percent (1 male, 4 female of 25 F1) of F1 progenies in line C showed non-prostate targeting

1591

Figure 3 Southern blotting analysis of three transgenic breeding mice lines: A (183) and B (186-3), C (186-9). For semi-quantitative determination and comparison of the copy number of transgenes, different amounts of SV40 Tag fragment DNA (from 0–500 relative copies as indicated, one relative copy of the 2.7 kb SVTag gene is ~2 pg DNA with ~6.6 × 105 copies) comparable to copy numbers of mouse chromosomal DNA (4 ␮g/per lane digested with PstI representing ~6.6 × 105 copies of mouse genome DNA) were loaded (left). The high intensity must be from the repeated copy of transgenes. The exposure time of Southern blots was adjusted to differentiate signals higher than 10 copies, and the integrated chromosomal flanking fragments (only one copy) are not visible in this blot. Gene Therapy

PSP94 gene promoter directed prostate targeting MY Gabril et al

1592

was also found (Figure 4d), since the rate of high copy transgene mice also correspondingly decreased from F0 to F3, and the majority of non-prostate targeting mice are high copy mice (6/11). The ratio of total numbers (male and female) of non-prostate targeting/high copy mice in F1, F2 and F3 was 5/3, 4/3 and 0/0, respectively. While in breeding, the proportion of homozygous (mating among C mice) over heterozygous (mating between C mice and wild type) in the resultant progenies was unchanged from the F2 (4/6 families) to F3 offspring (2/3 families).

Figure 4 Graphs showing prostate targeting of three transgenic mouse lines (a, b, c) was indicated as % (y-xis) for non-prostate targeting (NPT) (for details see Materials and methods) of total transgenic mice tested in A line: 116, B (84) and C (174). Both male (M) and female (F) transgenic mice were assessed. (d) Elimination of non-prostate targeting (shown by % of total number) of line C mice in the process of breeding in three generations (F1, F2, F3). HC (high copy of transgene 186 as shown in Figure 3) was shown by black bars. n, number of non-prostate targeting (blank bars)/total number of mice (M + F) tested.

with tumor detected in skin, liver, intestine and bone, etc, while in F2 progenies non-prostate targeting decreased significantly to 3.7% (3 male, 1 female of 108 F2) and in F3, no non-prostate targeting was found in three breeding families of 41 offspring. Correlation of non-prostate targeting with transmittance of high copy of transgene Gene Therapy

Histopathological characterization of prostate tumor development in the PSP94 promoter/enhancer region directed transgenic mice Targeting of transgene expression in the prostate directed by the PSP94 promoter/enhancer region in 210 transgenic male mice from the three transgenic breeding lines (A, B and C) was confirmed by histopathological studies (Figure 5). Figure 6 demonstrates the correlation of age (puberty of the prostate organ in transgenic mice) and sequential histopathological changes. Hyperplasia was detected by 10 weeks of age in three lines, a few A line mice showed an unexpected appearance of hyperplasia at 16 to 32 weeks of age. By 12 weeks of age, all B and C line mice (in A line it started by 16 weeks of age) showed changes of PIN (both low and high grades of PIN). All three lines showed correlation between PIN and age: low grade PIN was found in 44% (11/25) of mice in all three lines at age 12–19 weeks, then gradually decreased to being unidentifiable at or after 28 weeks of age (Figure 6). High grade PIN started in line C as early as 12 weeks (4/6), and was detected in 48% (12/25) of mice in all three lines from 12–19 weeks. By 24 weeks, some high grade PIN foci in lines B and C mice showed microinvasion of the stroma (Figure 5e). Well-differentiated CaP started by 24 weeks of age in all three lines, mostly in the ventral and dorsolateral prostate lobes (not in the coagulation gland and seminal vesicles) and was detected in 13% (2/15), 40% (2/5), and 25% (2/8) of lines A, B and C, respectively. By 32 weeks or later, well-differentiated CaP was detected in 40%, 100% and 50% of lines A, B and C, respectively. Moderately differentiated CaP was mainly detected in line A mice by 16–19 weeks of age. Poorly differentiated CaP started as early as 16 weeks of age in line A mice (10/40). Eight of 10 poorly differentiated CaP showed large visible (palpable) tumors, and the remainder (2/10) were non-visible tumor. By 28 weeks, poorly differentiated CaP was detected in line C mice (1/4). Correlation of SV40 Tag expression and tumor progression in targeted prostate tissues in three PSPdirected transgenic breeding lines Consistent with results of histopathological studies, the expression of SV40 Tag as detected by immunohistochemistry in three established transgenic mouse breeding lines revealed a positive correlation with tumor progression. In the initiation of hyperplastic epithelium, there was no positive nuclear staining for Tag oncoprotein under the detection limit of immunohistochemistry. The development of low grade PIN correlated with the SV40 Tag expression in the prostate epithelium, and Tag expression was localized by immunohistochemistry as shown in Figure 5b, d, h, j, and l, being mostly in the

PSP94 gene promoter directed prostate targeting MY Gabril et al

1593

Figure 5 Histological analysis of the prostate samples in PSP-transgenic mice. Sections were stained with hematoxylin/eosin (H&E) and immunohistochemistry (with hematoxylin counterstaining) for a SV40 Tag antibody. (a) Low grade PIN (LGPIN) in transgenic C line (12 weeks), H&E, ×40. (b) Immunohistochemistry of a serial slide of (a) section, localizing SV40 Tag expression (dark brown areas) in nuclei. ×40. (c) High grade PIN (HGPIN) in transgenic C line (13 weeks), H&E ×40. (d) Immunohistochemistry of a serial slide of (c). ×40. (e) Well-differentiated CaP (WDCaP) arising from a high grade PIN with microinvasion in transgenic C line (24 weeks), H&E ×25. (f) Well-differentiated CaP (WDCaP) shows invasive glandular differentiation, H&E ×40. (g) Well-differentiated CaP shows cribriform pattern in transgenic C line (28 weeks), H&E ×10. (h) Immunohistochemistry of a serial slide of (g) in well-differentiated CaP, ×25. (i) Moderately differentiated CaP (MDCaP) from transgenic A line at age (20 weeks), arrows pointing apoptotic bodies. H&E ×10. (j) High magnification of a section of (i) by immunohistochemistry, ×25. (k) Poorly differentiated CaP (PDCaP) from transgenic A line at age 16 weeks, H&E ×25. (l) Immunohistochemistry of a serial slide of (k) ×25. Solid and open arrows indicate apoptotic bodies and invasive structures, respectively.

nuclei although there was cell to cell variation in any given single gland in the different histological grades. The number of the positive epithelial cells as well as the intensity of staining (ie degrees of immunoreactivity) increased gradually from low grade PIN to high grade PIN (compare Figure 5b and d) in three lines. Line C mice showed stronger immunoreactivity in terms of numbers of positive stained cells and also the intensity of staining than other two lines (A, B). Well and poorly differentiated CaP mice showed stronger expression of the Tag than those of PINs (Figure 5h, j, l). To confirm the prostate-specific origin and to study gene expression in different stages of transgenic mice, RT-PCR analysis of PSP94 and probasin gene expression was performed in dorsolateral prostate samples from relatively homogenous tissues samples (n = 3) with different grades of tumor. As shown in Figure 7, both PSP94 and rPB gene expression was eliminated in well-differen-

tiated and poorly differentiated CaP. PSP94 gene expression decreased earlier than rPB in tumor development. Because the SV40 Tag is known to stabilize the accumulated p53 protein in the nucleus, immunohistochemistry analysis of p53 was conducted in prostate samples from different grades of PSP-transgenic mice (n = 4). Normal prostate as control showed no p53 signal (data not shown). Mice with low grade PIN showed weak signals in the nucleus (not shown) and with high grade PIN the signals were appearing only in nuclei with typical high grade PIN changes (Figure 9a). In poorly differentiated CaP samples, p53 signals were found to be strong and homogenous in isolated areas (Figure 9b). Metastatic prostate cancer in the PSP94 gene-driven transgenic CaP model In order to decide if the PSP-transgenic CaP model has metastatic potential, three out of eight mice (16 weeks Gene Therapy

PSP94 gene promoter directed prostate targeting MY Gabril et al

1594

nodes were performed on serial slides separately (Figure 9d). As shown in Figure 9c, the metastatic lesions in lymph nodes were composed of sheets of prostate malignant cells affecting most of the lymph node areas, and were histologically similar to the prostate poorly differentiated CaP malignant cells in line A mice. The replacement of the lymph node cortex with undifferentiated epithelial cells was evident. Using the immunohistochemistry test (Figure 9d), the metastatic cells were found to be reactive to Tag. In lines B and C, high grade PIN with local invasion (ie well-differentiated CaP, n = 4, see Figure 5e) was detected at 24–31 weeks of age, but no distant metastasis was found.

Figure 6 Correlation of age (weeks) and sequential tumor development in three PSP-transgenic breeding lines (a, b, c). n (x-xis), age groups (weeks). Number of mice in each group is indicated (n); % (y-xis), percentage of mice tested positive for the six histological grades in each group.

Responsiveness to the androgen deprivation of the prostate in PSP94 gene driven transgenic CaP model Since androgen responsiveness is a physiologic characteristic of the prostate, transgenic mice targeted by PSP94 gene with SV40Tag expression were tested and characterized by responsiveness to castration. Two groups (A and C) of mice at 20–26 weeks of age were selected: group A was from a faster tumor growing family of A line and was tested for tumor downsizing after castration, while group C was from an established prostate targeting family of line C with 100% immunohistochemistry reactivity at 12 weeks of age and was tested for changes of immunohistochemistry signal after castration. Two weeks after castration, the prostate autopsy samples (n = 3) from each group were analyzed. Figure 9e showed involution and atrophic changes in the entire prostate including areas of PIN, nuclear atypia and hyperplastic acini. Immunohistochemistry analysis on a serial slide of Figure 9e showed weak immunoreactivity for SV40 Tag protein (Figure 9f) compared with non-castrated control mice (data not shown), indicating responsiveness to androgen deprivation of the PSP-transgenic mouse CaP model. One month after castration, 1/3 of castrated mice of line A show high grade PIN with weak immunoreactivity, while 2/3 showed visible, fair-sized carcinoma (H&E, see Figure 9g) with positive immunoreactivity (Figure 9h). In line C mice (4/4), glands showed atrophic changes 1 month after castration (Figure 9i) with marked increase in the stroma ratio, and immunohistochemistry negative immunoreactivity to the Tag oncoprotein (Figure 9j).

Discussion Figure 7 RT-PCR analysis of the dorsolateral prostate tissue samples from transgenic mice line A shown by 1.5% agarose gel electrophoresis. Homogeneous tissue samples of different tumor grades at necropsy were selected: Hyp, hyperplasia; LGPIN, low grade PIN; WDCaP, well-differentiated adenocarcinoma; PDCaP, poorly differentiated CaP.

of age) in line A mice with visible prostate tumors were selected for gross pathological examination and all three showed enlarged renal lymph nodes. Western blotting experiments were performed using lysates of biopsy samples from different tissues and tested for SV40 Tag expression. As shown in Figure 8, metastatic Tag expression was seen in lymph nodes and the upper pole of the kidney. To demonstrate that metastatic cells in the lymph nodes were of prostatic origin, histological and immunohistochemistry analyses on samples of lymph Gene Therapy

For cancer gene therapy, the tactic of using a vector gene as a pharmacological agent for targeting inherited and acquired diseases was adopted only 5–10 years ago. In order to extrapolate from pre-clinical animal studies to clinical trials, the ideal vector gene selected for gene therapy requires: (1) a prostate tissue specific element in a relatively strong promoter/enhancer region; and (2) a relatively conserved evolving gene structure and function. Only two prostate tissue specific genes (rPB and PSA) have been initially considered as targeting vectors.9 However, neither can be utilized for both animal CaP model (transgene targeting) and clinical trial studies due to the functional and structural divergency in mammalian species. For example, although rat probasin amino sequence reveals that it possibly belongs to a lipocalin superfamily, no human counterpart has been clearly identified.8 Similarly, human PSA (or hk3) is one member

PSP94 gene promoter directed prostate targeting MY Gabril et al

1595

Figure 8 Western blotting analysis of prostate targeting and metastasis. A transgenic mouse (A line) at 16 weeks was subjected to thoroughly anatomical detection and SV40 Tag protein expression analysis. Blots were first tested for both first and second antibodies, and (a) shows the signals produced by the first antibody only. Next, the chemoluminescent signals were stripped and probed with a second antibody only (b). The second body against mouse IgG recognizes mouse IgG (both heavy and light chains), indicating that approximately comparable amounts of tissue lysate was tested in each lane.

of the serine proteinase and kallikrein superfamily, where three similar members (hk1, hk2, hk3) were aligned in the same chromosome locus, and each member has a different tissue-specific expression pattern. In the mouse, although approximately 30 kallikrein genes have been identified, no prostate specific kallikrein has been reported (for a mini-review, see Refs 16 and 18). The human PSA enhancer elements have not been designed and tested in transgenic mice to target transgenes that disrupt prostate function.8 The current clinical trial using a combination of rat probasin sequence and human PSA promoter/enhancer regions employs the human LNCaP cell line xenograft tumors in nu/nu mice.13,27,37 The rationale for using PSP94 as a clinically competent targeting vector for gene therapy of CaP is based on: (1) Both human PSP94 and PSA are considered to be restricted in the prostate secretory fluids (for a review see Ref. 30). Although both have shown a similar insignificant, non-prostate tissue-specific distribution (for example, breast, trachea, gastric juice, saliva) at significantly lower and minimal levels (mg/ml viz ng-pg/ml),31,33 it possibly represents a ‘leakiness threshold’, similar to all abundant secretory proteins. (2) The exceptionally high content of both PSA and PSP94 in the prostate fluid and semen (g/l) indicates the possibility of using the strong promoter/enhancer region to direct therapeutic gene expression while targeting to the prostate cancer tissue. In this study, the low efficiency expression of reporter gene (LacZ) directed by a mouse PSP94 promoter/enhancer region in a human prostate cancer cell culture (in vitro), may be explained by lack of cellular background requested by mouse gene to human prostate cell. In a transgenic study (in vivo), we have demonstrated that a 3.84 kb promoter/enhancer region of the PSP94 gene is capable of directing heterogeneous gene expression in a strictly prostate tissue-specific mode. A breeding population of close to 1000 for three generations with observation time for 38 weeks (some up to 1 year) was utilized. PSP-transgenic mice (A, B and C lines, n = 374) showed targeting mostly to the ventral prostate and dorsolateral prostate lobes (no coagulation gland and seminal vesicles), with cancer development in both of these two lobes. In our model, the neoplastic changes induced by SV40 Tag expression in the prostate in PSP94transgenic mice temporally occurred (10 weeks) in a prostate specific mode, ie it correlates with male puberty. All mice in lines A, B and C have hyperplastic changes by 10 weeks of age. 100% of mice in B and C lines develop

PIN transformation (low or high grade PIN), detectable and shown in association with immunohistochemistry analysis of SV40 Tag expression by 12 weeks. Most line A mice developed immunohistochemistry detectable PIN changes in the prostate after 20 weeks (Figure 6) and showed accelerated tumor growth of up to 2 × 3 cm size as early as 19 weeks (10/40). The PSP-transgenic CaP model has been further characterized as an experimental metastasis CaP model. Mice which showed distant metastatic deposits were characterized by: (1) gross pathology detection of large palpable poorly differentiated CaP (n = 8) co-incidently with visibly enlarged perinepheric lymph nodes; (2) immunohistochemistry and Western blotting analyses of SV40 Tag expression; (3) correlation of SV40 Tag with p53 expression in normal, low grade and high grade PIN and various grades of carcinoma (by immunohistochemistry, Figure 9); (4) loss of gene expression of rPB and PS94 in metastatic (poorly differentiated CaP) tissues (Figure 7). The possibility that these metastatic deposits are non-prostate targeting can be excluded by the location and extent of the primary tumor. For instance, PIN changes have no metastatic potential. Surgical castration provided the androgen deprivation for the PSP-transgenic mice and demonstrated the prostatic tissue specific responsiveness and in some cases, refractoriness to androgen deprivation. Our prostatic targeting experiment was characterized and demonstrated by SV40 Tag expression, where Tag as an oncogene served as a reporter gene. Compared with other reporter genes (CAT or LacZ), detection of SV40 Tag oncoprotein using a Tag-specific antibody is very sensitive and specific with low background in all mouse tissues. More importantly, SV40 Tag is easily detectable visually at necropsy in the majority of the mouse population in the various stages of tumor development and progression. It is impractical to examine every section of all tissues of every mouse and in a large number of mice by LacZ and CAT assay, which usually produce a high background in immunohistochemistry tests in mouse prostate tissue samples. We have attributed the non-prostate targeting in the founder and the next two generations of C-breeding line to the extraordinarily high copy, which we assume will result in a very complicated array of transgenes (186). There were different reports of copy numbers in transgenic mice, ranging from a few38 to hundreds.10,17 In this study we utilized a Southern blotting method to semiquantitatively determine and compare the copy number of three PSP-transgenic breeding lines. Assuming that the Gene Therapy

PSP94 gene promoter directed prostate targeting MY Gabril et al

1596

Figure 9 Histological analysis of high grade and metastatic CaP in PSP94 gene driven PSP-transgenic CaP model. (a) Immunohistochemistry of p53 expression in HGPIN (line A mice) ×10. (b) Immunohistochemistry of p53 expression in PDCaP. (line A mice) ×10. (c) A lymph node section from A line mouse of 16 weeks of age. Arrow points to residual normal lymph node tissue. H&E ×10. (d) Immunohistochemistry of SV40 Tag oncoproteins in a serial slide of (c) ×10. (e) A section of the prostate from a castrated mouse of line C, 2 weeks after castration. H&E ×25. (f) Immunohistochemistry analysis of a serial slide of (e) shows weak positive immunoreactivity to Tag oncoprotein. Hematoxylin ×30. (g) A section of the prostate from castrated mouse from line A 5 weeks after castration. Arrows, atrophic glands. H&E ×10. (h) Higher magnification of a section shown in (g), demonstrating immunohistochemistry analysis of Tag expression. ×25. (i) A section of the prostate from castrated mouse (line of C) 5 weeks after castration showing atrophy. H&E x4. (j) Immunohistochemistry analysis of serial slide of (i) shows negative immunoreactivity. ×10.

transgene is arrayed as a tandem repeating fragment39 integrated on to the mouse chromosome (as shown in Figure 3), high copies of transgene 186-9 transgenic DNA may be oriented by the same XbaI site in either the headto-tail or tail-to-head direction, which may account for divergent targeting patterns as shown in some of the C line mice. Although lines B and C have the same transgene structure and a similar immunohistochemistry positive rate in terms of breeding record, line C is the only Gene Therapy

one which showed non-prostate targeting, albeit at a low incidence rate of 4.3%. Non-prostate targeting is correlated with a high copy in three generations of C line, with the majority of non-prostate targeting mice in the first two breeding generations and mostly high copy mice (Figure 4d). The non-Mendelian segregation of PSP-transgenic C-line of the founder mouse (186-9 F0) may indicate that the extraordinarily higher copy number of transgene is possibly unstable in the transmittance of transgenic

PSP94 gene promoter directed prostate targeting MY Gabril et al

material in the chromosome. We have demonstrated by segregation record that non-prostate targeting in C line has been established in several non-prostate targeting breeding families (Figure 4d). We previously reported40 the prostate tissue-specific role of intron sequences and mRNA splicing. In this study, part of the intron 1 sequence was introduced to enhance the prostate specific targeting in transgenic CaP model. By comparing transgene structures of A with B and C, we believe that the additional intron sequence in the transgene may enhance prostate targeting and transgene expression (SV40 Tag-induced tumorigenesis) in the transgenic mice, since lines B and C showed earlier PIN (and SV40 Tag expression) than A (Figure 6). However, since line A is the only breeding line showing large sized, metastatic CaP at 15–25 weeks, the intron 1 sequence in lines B and C may play an attenuation role for promoter/enhancer driven SV40 oncogene expression leading to tumor progression and invasion (possibly also due to only parts of the intron sequence used in this study). The proliferative response as a consequence of Tag oncoprotein expression in a transgenic mouse is a prerequisite, but not sufficient for neoplastic transformation.19 There are other stochastic events which contribute to determining the incidence and growth rate of prostate cancer in each individual mouse.10,19 For example, in a small proportion of A line mice, neoplastic changes did not correlate with age (shown in Figure 6a). Additional factors need to be taken into account in further modification of our PSP-transgenic CaP model. We previously reported that rat PSP94 is more specific than probasin to the dorsolateral prostate,33 however in this study we observed mouse PSP94 promoter directed ventral prostate targeting. To clarify this question, we need to establish a knock-out (in) mouse CaP model with SV40 Tag insertion immediately adjacent to the PSP94 promoter/enhancer region. This research is currently underway in our laboratory.

Materials and methods Construction and in vitro testing of progressive deletion of the mouse PSP94 promoter/enhancer region linked with LacZ gene A 3842 bp promoter/enhancer region of the mouse PSP94 gene (GenBank accession No. AF087140) was screened from a mouse 129Sv genomic library (Stratagene, La Jolla, CA, USA) and completely sequenced. All sequencing reactions were on double strands and all data were based on overlapping fragments in a DNA sequencer (ABI Model 377, Perkin Elmer, Foster City, CA, USA). Introns of the mouse PSP94 gene have been reported previously.35 A series of progressive deletions of the mouse 3.842 kb promoter/enhancer region was constructed (Figure 1a) with a LacZ reporter gene (plasmid pSVgal, Promega, Toronto, Canada) using restriction enzymes: XbaI, BglII, BamHI, PstI and HindIII upstream of the first exon of mouse PSP94 cDNA. Figure 1b shows another plasmid construction (167) for testing additional intron sequences in cell lines, which contains a 3.842 kb promoter/enhancer region-exon 1-5’ intron 1 (KpnI–PstI, 1.5 kb)-3’ intron 1 (EcoRI–HindIII) and exon 2 ligated by blunt ends of HindII/BsaAI with the LacZ gene coding

region in the same reading frame (Figure 1b). Part of intron 1 sequence was selected only for the convenience of the cloning process. Transfection to LNCaP was performed in six-well plates using a Lipofectin reagent (Gibco/BRL, Burlington, Ontario, Canada) and was detected by X-gal staining. Approximately 10 ␮g plasmid DNA was used for 3 × 105 LNCaP cells and co-transfected by a mouse androgen receptor cDNA plasmid (CMV5mAR, cytomegalovirus promoter-driven, obtained from Dr D Robbins, Michigan University) at a 1:2 ratio.

1597

Construction of transgenes directed by 3.84 kb promoter/enhancer region of mouse PSP94 gene and intron 1sequences linked with SV40 Tag Two transgenes (183 and 186) were constructed using both SV40 large T and small t antigens from pBSV-1 (by courtesy of Dr Janet Butel, Baylor College of Medicine). Transgene 183 contains exon 1 (at KpnI site +16) to ⫺3842 bp and was constructed by insertion at EcoRV/Sal1 sites in a PSP94 subclone with a StuI/XhoI restriction fragment (2.7 kb) of SV40 Tag gene from pBSV-1 (Figure 2a). A 6.5 kb XbaI fragment of 183 was used for transgenic injection. The second transgene (186, shown in Figure 2b) was similar to 167, containing a 3.842 kb promoter/enhancer region-exon 1-5’ intron 1 (KpnI–PstI, 1.5 kb)-3’ intron 1 (EcoRI–HindIII) and a StuI/SalI fragment of SV40 Tag subclone at HincII/XhoI sites. The transgenic DNA fragment of 186 (~8.2 kb) was digested by XbaI digestion. Both transgenes, especially the junctional area between PSP94-SV40 Tag, have been confirmed by DNA sequencing analysis. Creation of transgenic mice, including micro-injection into fertilized eggs of mouse (C57BL/6 × BA) F1 hybrids, was performed in a transgenic core facility in London Regional Cancer Centre, Ontario, Canada. Identification and beeding of tansgenic mce by PCR and Southern botting Transgenic mice were first identified by a quick PCR test of toe chromosomal DNA using a nonionic detergent buffer with proteinase K digestion. Primer pairs were indicated in Figure 2a: Pr36: 5’ GGC AAC AGC GTG TCA AAG 3’; mPr70: 5’ GCC TTA GTC TCT GAT TGC TC3’; PrSVtag 5’ CAA GAC CTA GAA GGT CCA TTA GC 3’. After weaning, tail DNA was purified using an EASY kit (Qiagen, CA, USA) and PCR was repeated. Southern blots were prepared using ~5 ␮g mouse tail chromosomal DNA for each lane digested by PstI restriction enzyme and in a 0.5% agarose gel. A 2.7 kb DNA fragment of the SV40 Tag gene labeled by 32P dCTP was used as a probe. Breeding of transgenic mice was carried out in a strain of (C57BL/6 × CBA) F1 hybrids, first by mating with the wild type, then gradually by mating among transgenic mice. Mouse anatomy, animal handling, preparation and analysis of tissue samples Anatomy of the prostate complexes and the male accessory gland, the ventral prostate, the dorsolateral prostate lobes, coagulation gland and seminal vesicles, had been reported previously.33 Surgical castration under anaesthesia was performed via the scrotal route. All animal experiments were conducted according to standard protocols approved by the University Animal Care Committee. All organs and tissue samples (as indicated in the Gene Therapy

PSP94 gene promoter directed prostate targeting MY Gabril et al

1598

text) were subjected to gross pathological inspection, and any suspicious or abnormal looking tissue was sampled for immediate histology slide processing. By performing these procedures, non-prostate targeting (NPT) was designated as neoplastic changes (the maximum not beyond the PIN) undetectable in the prostate, but detectable in non-prostate tissues. Histopathological characterization and definitions of various degrees of CaP in transgenic mice Normal ventral and dorsolateral were composed of acinic glands lined by columnar secretory epithelial cells with round oval nuclei and pale cytoplasm. No abnormality was seen in the prostates of the normal control (n = 5, 6– 10 months of age). To study tumor development in the PSP-transgenic model, the mice were killed at different ages between 8 to 38 weeks. According to the diagnostic criteria established for human CaP and the basic conceptual similarity, prostatic neoplastic changes in the three breeding lines of the transgenic mice (A, B and C) were classified into the following six histological grading categories: (1) hyperplasia (Hyp) showing circumscribed epithelial proliferation with round, small and uniform nuclei; (2) low-grade PIN (LGPIN, Figure 5a, b) showing increased focal cellularity with mild enlarged nuclei of variable size and shape, increased chromatin density and scant cytoplasm (Figure 5a); (3) high grade PIN (HGPIN, Figure 5c, d) characterized by marked nuclear atypia, more hyperchromatic and chromatin clumping than the low grade PIN, increased mitotic and apoptotic rates and increased cellularity with stratification pattern; (4) welldifferentiated adenocarcinoma (WDCaP, Figure 5e, f, g, h) featuring several patterns: microinvasion identified as microacinar glands at the base of high grade PIN glands compressing the surrounding stroma (Figure 5e); invasive glandular differentiation pattern (Figure 5f) and multiple histological patterns including papillary, tufting, cribriform pattern (Figure 5g); (5) moderately differentiated adenocarcinoma (MDCaP, Figure 5i, j) showing some acini completely filled with malignant cells and more frequent appearance of apoptotic bodies, while maintaining the glandular structure (Figure 5i); and (6) poorly differentiated carcinoma (PDCaP, Figure 5k, l) containing sheets of malignant cells with no glandular features. Total cellular RNA isolation and RT-PCR Total RNA was purified by a Trizol solution (Gibco-BRL) and RT-PCR was performed as previously reported.40,41 The primer pairs were used: mPB-PR1: 5’ AAG ATA AAT GAA GGC TCA CCA TTG 3’; mPB-PR2: 5’ CAT ATT GAT GTT TCA GGT TCC AGG 3’; mPSP-1: 5’CCT GTA AGG AGT CCT GCT TTG TC 3’; mPSP-2 5’ATG CTG GCT CTG CCT TCT GAG T 3’. For semi-quantitative RT-PCR, samples (10 ␮l) were taken every five cycles from up to 20 cycles (for GAPDH: glyceraldehyde-3phosphate dehydrogenase, as controls), and after the initial 30 cycles (for rPB and PSP94), as reported previously.41 Immunohistochemistry and Western blotting analysis Monoclonal antibodies against SV40 Tag oncogene (Calbiochem, La Jolla, CA, USA) and p53 (AB-1, Oncogene, Cambridge, MA, USA) were used for immunohistochemistry by an ABC kit (StreptABC complex kit, DAKO,

Gene Therapy

Mississauga, Ontario, Canada) at 1:100 and 1:20 dilution separately. All immunochemistry slides were counterstained by hematoxylin. For Western blotting experiment, an ECL kit (enhanced chemiluminescent method, from Amersham, Oakville, Ontario, Canada) was used and a horseradish peroxidase conjugated goat anti-mouse IgG (Dimension Laboratories, Mississauga, Ontario, Canada) was used as a secondary antibody.

Acknowledgements The authors would greatly appreciate Dr Robert J Matusik and Dr Norman Greenberg for their help and support and for their critical review of the manuscript. This work was supported by grants from the Medical Research Council of Canada (MT-15390), The Kidney Foundation of Canada, The Prostate Cancer Research Foundation of Canada, an RGC Earmarked Research Grant (CUHK 4131/00M to FLC), a grant-in-aid (No.11671562) from the Ministry of Education, Science, Sports and Culture of Japan, and also from Procyon Biopharma Inc. (Montreal, Quebec, Canada).

References 1 Steiner MS, Gingrich JR. Gene therapy for prostate cancer: where are we now? J Urol 2000; 164: 1121–1136. 2 Kitsberg DI, Leder P. Keratinocyte growth factor induces mammary and prostatic hyperplasia and mammary adenocarcinoma in transgenic mice. Oncogene 1996; 13: 2507–2515. 3 Garabedian EM, Humphrey PA, Gordon JI. A transgenic mouse model of metastatic prostate cancer originating from neroendocrine cells. Proc Natl Acad Sci USA 1998; 62: 227–237. 4 Perez-Stable C et al. Prostate cancer progression, metastasis, and gene expression in transgenic mice. Cancer Res 1997; 57: 900–906. 5 DiGiovanni J et al. Deregulated expression of insulin-like growth factor 1 in prostate epithelium leads to neoplasia in transgenic mice. Proc Natl Acad Sci USA 2000; 97: 3455–3460. 6 Blackburn RV, Galoforo SS, Corry PM, Lee YJ. Adenoviralmediated transfer of a heat-inducible double suicide gene into prostate carcinoma cells. Cancer Res 1998; 58: 1358–1362. 7 Sharma P, Schreiber-Agus N. Mouse models of prostate cancer. Oncogene 1999; 18: 5349–5355. 8 Matusik RJ et al. Transgenic mouse models of prostate cancer. In: Matuk M, Brown CW, Kumar TR (eds). Comtemporary Endocrinology: Transgenics in Endocrinology. Humana Press: Totowa, 2001, pp 401–425. 9 Green JE et al. Workgroup 3: transgenic and reconstitution models of prostate cancer. Prostate 1998; 36: 59–63. 10 Maroulakou IG, Anver M, Garrett L, Green JE. Prostate and mammary adenocarcinoma in transgenic mice carrying a rat C3(1) simian virus 40 large tumor antigen fusion gene. Proc Natl Acad Sci USA 1994; 91: 11236–11240. 11 Shibata MA et al. Progression of prostatic intraepithelial neoplasia to invasive carcinoma in C3(1)/SV40 large T antigen transgenic mice: histopathological and molecular biological alterations. Cancer Res 1996; 56: 4894–4903. 12 Shibata MA et al. Haploid loss of bax leads to accelerated mammary tumor development in C3(1)/SV40-TAg transgenic mice: reduction in protective apoptotic response at the preneoplastic stage. EMBO J 1999; 18: 2692–2701. 13 Rodriguez R et al. Prostate attenuated replication competent adenovirus (ARCA) CN706: a selective cytotoxic for prostate— specific antigen-positive prostate cancer cells. Cancer Res 1997; 57: 2559–2563. 14 Pang S et al. Identification of a positive regulatory element responsible for tissue-specific expression of prostate-specific antigen. Cancer Res 1997; 57: 495–499. 15 Schuur ER et al. Prostate-specific antigen expression is regulated by an upstream enhancer. J Biol Chem 1996; 271: 7043–7051.

PSP94 gene promoter directed prostate targeting MY Gabril et al

16 Schaffner DL et al. Transgenic mice carrying a PSArasT24 hybrid gene develop salivary gland and gastrointestinal tract neoplasms. Lab Invest 1995; 72: 283–290. 17 Cleutjens KB et al. A 6-kb promoter fragment mimics in transgenic mice the prostate-specific and androgen-regulated expression of the endogenous prostate-specific antigen gene in humans. Mol Endocrinol 1997; 11: 1256–1265. 18 Wei C et al. Tissue-specific expression of the human prostatespecific antigen gene in transgenic mice: implications for tolerance and immunotherapy. Proc Natl Acad Sci USA 1997; 94: 6369–6374. 19 Gingrich JR et al. Metastatic prostate cancer in a transgenic mouse. Cancer Res 1996; 56: 4096–4102. 20 Greenberg NM, Demayo F, Finegold MJ et al. Prostate cancer in a transgenic mouse. Proc Natl Acad Sci USA 1995; 92: 3439–3443. 21 Gingrich JR et al. Androgen-independent prostate cancer progression in the TRAMP model. Cancer Res 1997; 57: 4687–4691. 22 Kasper S et al. Development, progression and androgen-dependence of prostate tumors in probasin-large T antigen transgenic mice: a model for prostate cancer. Lab Invest 1998; 78: 319–333. 23 Zhang J, Thomas TZ, Kasper S, Matusik RJ. A small composite probasin promoter confers high levels of prostate-specific gene expression through regulation by androgens and glucocorticoids in vitro and in vivo. Endocrinology 2000; 141: 4698–4710. 24 Wu X et al. Generation of a prostate epithelial cell-specific Cre transgenic mouse model for tissue-specific gene ablation. Mech Dev 2001; 101: 61–69. 25 Hsieh CL, Yang L, Miao L et al. A novel targeting modality to enhance adenoviral replication by vitamin D(3) in androgenindependent human prostate cancer cells and tumors. Cancer Res 2002; 62: 3084–3092. 26 Matsubara S et al. A conditional replication-competent adenoviral vector, Ad-OC-E1a, to cotarget prostate cancer and bone stroma in an experimental model of androgen-independent prostate cancer bone metastasis. Cancer Res 2001; 61: 6012–6019. 27 Yu DC et al. Antitumor synergy of CV787, a prostate cancerspecific adenovirus, and paclitaxel and docetaxel. Cancer Res 2001; 61: 517–525. 28 Latham JPF, Searle PF, Mautner V, James ND. Prostate-specific antigen promoter/enhancer driven gene therapy for prostate cancer: construction and testing of a tissue specific adenovirus vector. Cancer Res 2000; 60: 334–341.

29 Yu DC, Sakamoto GT, Henderson DR. Identification of the transcriptional regulatory sequences of human Kalilkrein 2 and their use in the construction of Calydon virus 764, an attenuated replication competent adenovirus for prostate cancer therapy. Cancer Res 1999; 59: 1498–1504. 30 Hara M, Kimura H. Two prostate-specific antigens, -seminoprotein and ␤-microseminoprotein. J Lab Clin Med 1989; 113: 541–548. 31 Wu D et al. Serum bound forms of PSP94 (prostate secretory protein of 94 amino acids) in prostate cancer patients. J Cell Biochem 1999; 76: 71–83. 32 Imasato Y et al. PSP94 expression after androgen deprivation therapy: a comparative study with prostate specific antigen in benign prostate and prostate cancer. J Urol 2000; 164: 1819–1824. 33 Imasato Y et al. Rodent PSP94 gene expression is more specific to the dorsolateral prostate and less sensitive to androgen ablation than probasin. Endocrinol 2001; 142: 2138–2146. 34 Bauman GS et al. PSP94: evaluation of prognostic utility in patients treated with radiotherapy for nonmetastatic prostate cancer. Prostate J 2000; 2: 94–101. 35 Xuan JW et al. cDNA, genomic cloning and gene expression analysis of mouse PSP94 (prostate secretory protein of 94 amino acids). DNA Cell Biol 1999; 18: 11–26. 36 Palmiter RD et al. Heterologous introns can enhance expression of transgenes in mice. Proc Natl Acad Sci USA 1991; 88: 478–482. 37 Yu DC et al. The addition of adenovirus type 5 region E3 enables Calydon virus 787 to eliminate distant prostate tumor xenografts. Cancer Res 1999; 59: 4200–4203. 38 Yan Y et al. Large fragment of the probasin promoter targets high levels of transgene expression to the prostate of transgenic mice. Prostate 1997; 32: 129–139. 39 Hogan B, Beddington R, Costantini F, Lacy E. Manipulating the Mouse Embryo: A Laboratory Manual. In: Cold Spring Harbor Laboratory: New York 1994. 40 Xuan JW et al. Alternative splicing of PSP94 (prostate secretory protein of 94 amino acids) mRNA in prostate tissue. Oncogene 1995; 11: 1041–1047. 41 Xuan JW et al. Molecular cloning and gene expression analysis of PSP94 (prostate secretory protein of 94 amino acids) in primates. DNA Cell Biol 1997; 16: 627–638.

1599

Gene Therapy