Abstract
We report the development of a novel replication-competent adenoviral vector, Ad-hOC-E1, containing a single bidirectional human osteocalcin (hOC) promoter to drive both the early viral E1A and E1B gene. This vector selectively replicated in OC-expressing but not non-OC-expressing cells, with viral replication enhanced at least 10-fold on vitamin D3 exposure. Both the artificial TATA-box and hOC promoter element in this bidirectional promoter construct were controlled by a common OC regulatory element which selectively activated OC expression in cells. The expression ofE1A and E1B gene by Ad-hOC-E1 can be markedly induced by vitamin D3. Unlike Ad-sPSA-E1, an adenoviral vector with viral replication controlled by a strong super prostate-specific antigen (sPSA) promoter which only replicates in PSA-expressing cells with androgen receptor (AR), Ad-hOC-E1 retarded the growth of both androgen-dependent and androgen-independent prostate cancer cells irrespective of their basal level of AR and PSA expression. A single i.v. administration of 2 × 109 plaque-forming units of Ad-hOC-E1 inhibited the growth of previously established s.c. DU145 tumors (an AR- and PSA-negative cell line). Viral replication is highly enhanced by i.p. administration of vitamin D3. Ultimately, enhancing Ad-hOC-E1 viral replication by vitamin D3 may be used clinically to treat localized and osseous metastatic prostate cancer in men.
INTRODUCTION
One of the biggest challenges facing gene therapy is the development of site-specific vectors for therapeutic genes. Two main approaches using Ad 3 -mediated cancer gene therapy have been attempted. The first is targeting tumor cells with genetically modified Ad through either direct modification of viral fiber protein (1, 2, 3, 4) or through conjugation with bispecific single-chain antibodies (5 , 6) that interact with specific cell surface receptors. The second approach is selective targeting with tumor- and tissue-specific promoters driving the expression of therapeutic genes or viral replication in a tumor cell-specific manner (7, 8, 9) . In replication-competent Ad vectors, the tissue or tumor-specific promoter-enhancer has exclusively been inserted proximal to the E1A promoter-enhancer region (10, 11, 12) with the rationale that expression of E1A and, therefore, the whole Ad transcription program will depend on these tissue- or tumor-specific promoters. However, leakage of foreign promoters in E1A control, yielding low levels of E1A, may result in loss of specificity (13) . To control the viral replication more stringently, separate promoter control of both the E1A and E1B genes significantly improves specificity (14 , 15) . However, with promoter interference and homologous recombination between closely juxtaposed promoters within the Ad vectors, modulation or loss of promoter activity and tissue specificity may result (16) . Therefore, we devised an alternate strategy to retain a high level of specificity for target cells by controlling expression of E1A and E1B genes with a single bidirectional promoter fused with a reversed artificial TATA-box upstream of the enhancer/promoter region.
In transgenic mice carrying the hOC promoter-chloramphenicol acetyltransferase fusion gene, OC expression was restricted to bone-associated tissues and the brain (17) . Our laboratory also demonstrated that OC expression is detectable in both primary and bone metastatic prostate tumor specimens (18) . We previously showed that mouse OC-mediated hsv-TK (OC-TK), plus the pro-drug ganciclovir (GCV) or acyclovir (ACV), efficiently blocks the growth of localized prostate tumors and their skeletal xenografts (19) . We recently showed that a single i.v. administrated dose of Ad-OC-E1a markedly inhibited previously established prostate tumor grown in the skeleton (18) . A Phase I OC dose escalation trial has demonstrated the safety of intratumoral delivery of Ad-OC-TK followed by an oral ACV analogue, valacyclovir (20) . To improve on mouse OC promoter, we developed a human version, hOC, which contains a VDRE. Its activity can be induced by vitamin D (21) . We now demonstrate that a hOC-E1 bidirectional E1A/E1B expression cassette (Ad-hOC-E1) can be effectively up-regulated by vitamin D3. Concomitant Ad-hOC-E1 and vitamin D3 treatment showed a highly specific and effective kill of AI and metastatic prostate cancer cells in vitro and markedly reduced the growth of both AR- and PSA-negative DU145 tumor xenografts in nude mice by a single systemic administration of Ad-hOC-E1.
MATERIALS AND METHODS
RT-PCR Analysis.
Cells were treated with 5 nm vitamin D3 analogue (Ro 25-9022; Roche, Nutley, NJ) or ethanol as the control group for 48 h. RNA was extracted using RNAzolB (Teltest, Friendswood, TX) and RT-PCR was performed according to the manufacturer’s protocol with Moloney Murine Leukemia Virus reverse transcriptase (Life Technologies, Inc., Rockville, MD). The primer sequences for hOC are 5′-ACACTCCTCGCCCTATTG-3′ (forward) and 5′-GATGTGGTCAGCCAACTC-3′ (reverse); for PSA 5′-GATGACTCCAGCCACGAC-3′ (forward) and 5′-CACAGACACCCCATCCTA-3′ (reverse); for AR 5′-ATGGAAGTGCAGTTAGGG-3′ (forward) and 5′-CAGGATGTCTTTAAGGTCAGC-3′ (reverse); for VDR are 5′-ATGGAAGTGCAGTTAGGG-3′ (forward) and 5′-TCAGGAGATCTCATTGCC-3′ (reverse); for GAPDH are 5′-ACCACAGTCCATGCCATCA-3′ (forward) and 5′-TCCACCACCCTGTTGCTGT-3′ (reverse).
Plasmid and Virus Construction.
A 3.9-kb hOC promoter was cloned from genomic DNA of DU145, using Genome Walker kits (Clontech, Palo Alto, CA). A short version (800-bp) of hOC promoter (22) was subsequently generated by PCR. A 600-bp super-PSA promoter (sPSA) was created by our laboratory as described previously (23) . Ad5 E1A and E1B cDNA were amplified from pXC548c (24) by PCR. An artificial 33-bp TATA-box containing fragment was obtained from the original pGL3/TATA (23) . These fragments were subcloned to generate the bidirectional hOC or sPSA promoter-driving E1A and E1B expression cassette (Fig. 2) ⇓ using standard cloning methods (25) . Ad-hOC-E1 and Ad-sPSA-E1 were generated in 293 cells by cotransfecting these cells with both the expression shuttle plasmid and a circular Ad genome plasmid (pJM17). After transfection, cells were cultured in agarose medium for up to 10–12 days to allow plaque formation. Individual plaques were picked up and screened by the PCR method. Viral DNA of recombinant Ad vectors obtained from the selected plaques was extracted and digested with HindIII. Ad vectors were amplified in 911 cells (26) to avoid recombinant Ad virus generation and purified according to the method of Graham and Prevec (27) . Wild-type Ad5 (Ad-w.t.), dl309 (28) , was a gift from Dr. Frank Graham (McMaster University, Ontario, Canada). A replication-defective Ad vector, Ad-CMV-pA, was constructed by our laboratory as described previously (29) . All of the Ad vectors were evaluated by particle count as determined by absorbance measurement of DNA and titered by plaque assay.
Northern Blot Analysis.
Cells were infected with 10 MOI of Ad-hOC-E1. After 2-h absorption, the virus was removed, and cells were incubated with fresh medium containing 5 nm vitamin D3 or ethanol for 24 h. RNA was extracted, and 5 μg of total RNA was fractionated on a 1% agarose, glyoxal-based denaturing gel, and transferred to a Hybond NX (Amersham Pharmacia Biotech) membrane. The blots were probed with 32P-labeled-full length E1A or E1B cDNA in Rapid-Hyb buffer (Amersham Pharmacia Biotech) according to the manufacturer’s protocol. Membranes were exposed to BioMax film (Kodak). Quantity one-4.1.1 Gel Doc gel documentation software (Bio-Rad) was used for quantification of intensity.
Western Blot Analysis.
Cells were infected with 10 MOI of Ad-hOC-E1 or Ad-CMV-pA. After 2 h of absorption, cells were washed with PBS twice and then cultured in fresh medium containing 5 nm vitamin D3 analogue or ethanol for 48 h. Cells were lysed in triple-detergent lysis buffer [50 mm Tris (pH 8.0), 150 mm NaCl, 0.02% NaN3, 0.1% SDS, 1% NP40, 0.5% sodium deoxycholate, 1 mm phenylmethylsulfonyl fluoride and 1× protease inhibitor cocktail (Roche, Nutley, NJ)]. Fifty μg of protein were used for immunoblotting using the NOVEX (Invitrogen, Carlsbad, CA) system. Membrane was probed with a 1:200 dilution of adenovirus-2 E1A antibody (13 S-5), followed by a 1:5000 dilution of horseradish peroxidase-conjugated secondary antibody against rabbit IgG. Both antibodies were purchased from Santa Cruz Biotechnology Inc. (Santa Cruz, CA). ECL plus system (Amersham Pharmacia Biotech) was used to detect the signal. Cell lysate from 293 cells was used as quantitative reference in each blot and the intensity of bands was measured by Quantity one-4.1.1 Gel Doc gel documentation software (Bio-Rad).
Southern Blot Analysis.
Cells were infected with 10 MOI of Ad-hOC-E1 for 2 h and then cultured in 5 nm vitamin D3- or ethanol-containing medium and harvested 4, 24, and 48 h p.i. Viral DNA was prepared as described previously (30) . Samples were digested with HindIII and fractionated on a 1.2% agarose gel and transferred to a Hybond NX (Amersham Pharmacia Biotech) membrane. The blots were probed with 32P-labeled HindIII-digested fragments of dl309 in Rapid-Hyb buffer (Amersham Pharmacia Biotech) based on the manufacturer’s protocol. Membranes were exposed to BioMax film (Kodak).
Indirect Immunofluorescence Analysis of CAR Antigen.
One × 106 cells of each cell line were washed extensively with cold PBS, incubated with 1 μg/ml mouse anti-CAR antibody (RmcB, provided by Dr. Jer-Tsong Hsieh, University of Texas Southwestern, Dallas, TX) or a 1:50 diluted isotope control antibody in 100 μl of FACS buffer (0.1% BSA and 1% sodium azide in PSA) at 4°C for 1 h. After incubation, cells were washed with cold FACS buffer twice. Cells were incubated with FITC-conjugated goat antimouse IgG (Jackson Immunoresearch Laboratories Inc.) in a 1:50 dilution at 4°C for 1 h and then analyzed by flow cytometry (FACSCalibur, Becton Dickinson, San Jose, CA). The intensity of fluorescence was determined for 10,000 cells.
In Vitro Cytotoxicity and Viral Replication Assay.
Cells were seeded onto 24-well plates and infected with Ad vectors in the range of 0–5 MOI. For viral replication assay, 2 MOI of Ad vectors were used. After 2-h adsorption, cells were cultured in 5 nm vitamin D3- or ethanol-containing medium. For in vitro cytotoxicity assay, crystal violet analysis was performed 1, 3, 5, and 7 days p.i. as described previously (29) . The relative cell number was assessed by absorbance at 590 nm after staining and was calculated as fold of uninfected cells. For in vitro viral replication assay, culture supernatant was harvested 3 days p.i. and viral titer was determined by plaque assay. Each experiment was carried out either in duplicate or triplicate.
Animal Studies.
Xenografts were established by s.c. injecting 2 × 106 DU145 cell suspension into the flanks of male nude mice (Charles River, Wilmington, MA). When tumors reached 100 mm3, mice were randomized (n = 8 in each group) and given 2 × 109 pfu of Ad-CMV-pA, Ad-hOC-E1, and PBS (mock injection) in a volume of 100 μl, via tail vein injection. One day after viral injection, mice were treated by i.p. administration of 100 μl vitamin D3 (4 ng/dose) or control vehicle (3.2% ethanol, 2.6% PEG-400, 2.2% Tween 80, and 92% PBS) every other day for 3 weeks. Vitamin D3-treated mice were fed a sterilized calcium deficient diet (ICN Research Diets). Tumor volume measurements were taken twice per week and calculated according to the formula: length × width2 × 0.5236 (31) . Data are expressed as fold of tumor volume increase, obtained by assessing tumor size relative to the initial size at the time of virus or vehicle injection. For viral distribution and in vivo replication analysis, tissue DNA was extracted using DNeasy Tissue Kit (Qiagen, Valencia, CA) and 200 ng of DNA were used for detecting the Ad5 DNA polymerase sequences by PCR amplification. The primers were 5′-TACGGCATCTCGATCCAC-3′ (forward) and 5′-TCGAGGACAGGCCTCTCA-3′ (reverse). For immunohistochemical staining, deparaffinized tumor and liver specimens were treated with 3% H2O2, blocked with SuperBlock (Scytex Laboratories, Logan, UT) and reacted with an Adenovirus-2 E1A antibody (13 S-5, Santa Cruz Biotechnology) in a 1:1000 dilution. The antibody staining signals were amplified by a biotinylated-peroxidase-conjugated streptavidin system (Bio-Genex Laboratories, San Ramen, CA). Background staining was obtained using control rabbit IgG. E1A stain was visualized after reacting with 3,3′-diaminobenzidine.
Statistical Analysis.
Differences between treatment groups were analyzed using Student’s t test and two-tailed distribution.
RESULTS
Vitamin D3 Up-Regulates hOC mRNA Expression in AI Prostate Cancer Cell Lines.
RT-PCR was performed to compare the expression profiles of OC, PSA, and AR mRNA among several human cancer cell lines. Fig. 1 ⇓ shows the basal level of OC mRNA expression, based on RT-PCR, can be detected clearly in human osteogenic sarcoma MG-63, and AI metastatic prostate cancer cell lines, PC3, PC3M, and DU145. A very low level of OC transcripts was detected in LNCaP, an AD line, and its derivative AI and metastatic subline C4-2. PSA mRNA can only be detected in AR-expressing prostate cancer cells, such as LNCaP and C4-2, but not in the AR-expressing osteoblastic cell line MG-63 or other AR-negative but OC-expressing prostate cancer cell lines. In addition, OC mRNA expression can be up-regulated markedly by a vitamin D3 analogue in all metastatic prostate cancer cell lines including C4-2 cells, which only express a trace basal level of hOC. In renal cancer RCC 52 cells, the basal level of OC mRNA was undetectable. However, vitamin D3 can also activate OC promoter by an unknown mechanism (Fig. 1) ⇓ . Cell lines in which OC can be induced by vitamin D3, all express VDR, which suggests that the VDR complex with VDRE in the proximal region of hOC promoter must be responsible for up-regulating OC transactivating activity in prostate and bone cancer cell lines. We observed that vitamin D3 also slightly enhanced PSA mRNA expression in LNCaP and C4-2 cells by an unknown mechanism (31 , 32) . These results show that the hOC promoter has a broader spectrum of activity than the PSA promoter in AD and AI prostate cancers, and could be highly inducible by vitamin D3. The results suggest that therapeutic genes driven by the hOC promoter can target primary and metastatic prostate cancers.
Basal and vitamin D3-induced OC, PSA, and AR VDR mRNA expression in human prostate cancer cell lines. RT-PCR was performed using total RNA prepared from a series of human prostate cancer cell lines (LNCaP, C4-2, PC3, PC3M, and DU145), a human osteosarcoma cell line (MG63), and a human renal cell carcinoma cell line (RCC52) cultured in the presence or absence of 5 nm vitamin D3 for 48 h. MG63 was used as a positive control, and RCC52 without vitamin D3 treatment was used as a negative control of OC mRNA expression. GAPDH expression by RT-PCR was used as an internal standard of RNA loading in each sample.
E1A and E1B genes Are Expressed in OC-expressing Cells by Ad-hOC-E1.
To control both E1A and E1B genes with a single promoter, we generated bidirectional hOC and strong sPSA promoters by inverting an artificial TATA box lined in the opposite direction to hOC or sPSA enhancer/promoter. The E1A cDNA was cloned downstream of an artificial TATA box promoter, and E1B cDNA was cloned downstream of the hOC or sPSA enhancer/promoter region. Replication-competent Ad-hOC-E1 and Ad-sPSA-E1 vectors were constructed by inserting these bidirectional E1A/E1B expression cassettes at the deleted E1 region of the replication-defective Ad5 virus. In parallel, a replication defective Ad vector, Ad-CMV-pA, was also constructed by inserting the polyadenylated[poly(A)] signal-linked CMV promoter fragment at the same region as E1A/E1B expression cassette (Fig. 2) ⇓ . The bidirectional hOC promoter, shown to be functional for driving both E1A and E1B gene expression, can be induced by vitamin D3 in both directions. OC-expressing prostate cancer cell lines, C4-2, PC3, and DU145, and a non-OC-expressing cell line, RCC52, were infected with Ad-hOC-E1 and the transcription of E1A and E1B mRNA was assessed by Northern blot analysis. The basal level of two major E1B transcripts, 12S and 22S mRNAs transcribed from hOC-enhancer/promoter, was detected only as trace in C4-2, PC3, and DU145, and was nondetectable in RCC52 cells. After vitamin D3 induction, however, these transcripts were enhanced 10- to 50-fold above the basal level of gene expression in these cell lines (Fig. 3) ⇓ . E1A transcript can be clearly seen in all of the OC-expressing cell lines, but is barely visible in the OC-nonexpressing RCC52 cell line without vitamin D3 induction. On vitamin D induction, the steady-state level of E1A mRNA was dramatically enhanced (Fig. 3) ⇓ . These results suggest that both the artificial TATA-box and the hOC promoter element can be controlled by a common OC regulatory element in a bidirectional manner. OC transcription in cells can be markedly induced by vitamin D3.
Organization of the vectors used in this study. A replication defective Ad vector, Ad-CMV-pA, carrying an expression cassette driven by the human CMV-IE promoter and terminated by the SV40 pA was inserted in the E1 region of first generation E1-deleted adenovirus. In the replication competent vectors Ad-hOC-E1 and Ad-sPSA-E1, Ad5 E1A and E1B expression are driven by bidirectional hOC and sPSA enhancer/promoter (E/P), respectively.
Basal and vitamin D3-induced E1A and E1B mRNA transcription by Ad-hOC-E1. AI human prostate cancer cell lines C4-2, PC3, and DU145, and human renal cell carcinoma cell line RCC52, infected with 10 pfu/cell Ad-hOC-E1 or Ad-CMV-pA, were cultured in the presence or absence of 5 nm vitamin D3. RNA was extracted at 48 h p.i. E1A and E1B mRNA were detected by Northern blot and probed with 32P-labeled E1A and E1B cDNA probes, respectively. 28S RNase was used as an internal control of RNA loading of each sample.
To test whether E1A transcript driven by artificial promoter element can be successfully translated into E1A protein, we performed an immunoblotting analysis of E1A protein from Ad-hOC-E1- or Ad-CMV-pA-infected C4-2, PC3, DU145, and RCC52. Cell lysate prepared from 293 cells expressing endogenous E1 gene was used as a positive control. Fig. 4 ⇓ shows that with the exception of DU145 cells, the induction of E1A mRNA by vitamin D3 (Fig. 3) ⇓ is consistent with both basal and vitamin D3-induced E1A protein accumulation in cells. A trace amount of E1A protein accumulated in Ad-hOC-E1-infected RCC52 cells, and this level was greatly enhanced by vitamin D3. E1A was undetectable in cells infected with Ad-CMV-pA, either with or without vitamin D3.
Basal and vitamin D3-induced E1A protein expression by Ad-hOC-E1. AI human prostate cancer cell lines C4-2, PC3, and DU145, and human renal cell carcinoma cell line RCC52, infected with 10 pfu/cell of Ad-hOC-E1 or Ad-CMV-pA, were cultured in the presence or absence of 5 nm vitamin D3. Proteins were prepared 48 h p.i. E1A protein was detected by Western blot and probed with a polyclonal antibody to Ad2 E1A protein. Cell lysate from 293 cells was used as a positive control of E1 protein expression. KDa, molecular weight (Mr) in thousands.
Ad-hOC-E1 Selectively Replicates in OC-expressing Cells.
To examine Ad-hOC-E1 viral DNA replication in C4-2, PC3, DU145, and RCC52 cells, we performed Southern blot to detect the viral DNA accumulated in the cells. With vitamin D3 induction, Ad-hOC-E1 DNA was detected in C4-2 and DU145 cells 24 h earlier than without vitamin D3 (Fig. 5) ⇓ . Similarly, vitamin D3 induced a 3- to 10-fold increase over the basal level of Ad-hOC-E1 DNA replication. The induction of viral replication in C4-2, PC3 and DU145 correlated with hOC mRNA expression (Fig. 1) ⇓ . In RCC52 cells, Ad-hOC DNA was strongly detected at 48 h p.i. with vitamin D3 induction but was barely detectable without induction. This result is consistent with the expression of hOC mRNA (Fig. 1) ⇓ . Because RCC52 cells expressed a high level of CAR on the cell surface (Fig. 6) ⇓ , obviously the failed Ad-hOC-E1 replication in non-hOC-expressing cells is attributable to the stringent specificity of hOC promoter and is not related to the efficiency of viral entry.
Basal and vitamin D3-induced Ad-hOC-E1 replication. A, cells infected with 10 pfu/cell Ad-hOC-E1 were cultured in the presence or absence of 5 nm vitamin D3. Viral DNA was isolated at the indicated hours p.i. DNA from an equal number of cells was digested withHindIII and subjected to Southern blot analysis. The HindIII-digested DNA fragments of Ad-w.t. were labeled with [32P[CTP and used as a probe to detect all viral HindIII fragments. B, a serial diluted Ad-hOC-E1 DNA was digested with HindII and used as a positive control of Southern blot.
Immunoflurorescence FACS analysis of CAR expression. The indicated cells were incubated with either a monoclonal antibody to CAR (open area) or an isotope-identical control antibody (shaded area) and subjected to flow cytometric analysis.
To further assess whether the amplified viral DNA can be packed to form the infectious particle, we harvested culture medium and performed a plaque assay. The differential titer of Ad vectors in various human cell lines is shown in Table 1 ⇓ and demonstrates that Ad-hOC-E1 grew well in OC-expressing prostate cancer cell lines, such as C4-2, PC3, and DU145, and that vitamin D3 can induce a 5- to 25-fold increase in viral replication. This induced viral titer is equal to Ad-w.t. in PC3. However, Ad-OC-E1 cannot grow in non-OC-expressing RCC52 cells. The titer of Ad-hOC-E1 in these cells is as low as that of Ad-CMV-pA, a replication-defective Ad vector.
Titer of Ad vectors in human cell line (pfu/ml)
Cells were plates and infected as described in “Materials and Methods.” Viral titer of culture supernatants was determined by plaque assay. Data represent the mean of three experiments. Titers were normalized to 1 × 108 pfu/ml in 293 cells.
Replication-competent Ad Vectors Induce AI Prostate Cancer Cell Death.
To test whether replication-competent Ad vectors can grow and lyse prostate cancer cells, an in vitro cytotoxicity assay comparing Ad-hOC-E1 and Ad-sPSA-E1 was performed. As controls, Ad-w.t. (positive control) and Ad-CMV-pA (negative control) either effectively lysed or were completely ineffective in all of the tested cell lines (data not shown). Fig. 7A ⇓ shows that in response to Ad-hOC-E1, marked cell lysis was observed in C4-2 cells (AR- and PSA-positive) at a dose level of 1 MOI (P < 0.05 versus mock-infected group) at 7 days posttreatment and day 5 by vitamin D3 induction. The 10-fold-enhanced cell kill of Ad-hOC-E1 by vitamin D3 also occurred in PC3 and DU145 cells (Fig. 7A) ⇓ , whereas there was no vitamin D-induced kinetic change of cytotoxicity observed in Ad-sPSA-E1, Ad-CMV-pA, and Ad-w.t.-treated groups in any of the tested cell lines (data not shown). With vitamin D3 treatment (Fig. 7B) ⇓ , in C4-2 cells, Ad-sPSA-E1 showed higher cell killing activity than did Ad-hOC-E1. These data correlate with the endogenous PSA and OC promoter activity shown in Fig. 1 ⇓ . In contrast, however, when Ad-hOC-E1 or Ad-sPSA-E1 was added to PC-3 cells (AR- and PSA-negative), a differential inhibition of cell growth by Ad-hOC-E1 but not by Ad-sPSA-E1 was observed. Because of the lower level of CAR associated with PC-3 cells (Fig. 6) ⇓ , a higher dose of Ad vector (e.g., 1–5 MOI) was necessary to lyse these cells in vitro. This is supported by DU145 experimental data (AR- and PSA-negative cells), which have a higher level of CAR (Fig. 6) ⇓ and were inhibited by Ad-w.t. and Ad-hOC-E1 at a dose of 0.1–1.0 MOI 5–7 days after viral exposure. Ad-sPSA-E1 and Ad-CMV-pA were ineffective against the growth of DU145 cells in vitro. As expected, the growth of RCC52 cells (completely deficient in OC expression) was sensitive only to Ad-w.t. and completely insensitive to growth inhibition by Ad-OC-E1, Ad-sPSA-E1, or Ad-CMV-pA (Fig. 7C) ⇓ .
Vitamin D3-induced cytotoxicity by replication-competent Ad vectors. OC-positive C4-2, PC-3, and DU145 cells were infected with (A) Ad-hOC-E1, (B) Ad-hOC-E1, or Ad-sPSA-E1 at the indicated dose, and OC-negative RCC52 cells were infected with (C) indicated Ad vectors. Cells were subsequently cultured in the presence (A, B) or absence (A, C) of 5 nm vitamin D3, and the cytotoxicity assay was performed using crystal violet staining at the indicated day after infection. The relative cell number was assessed by absorbance at 590 nm after staining; versus mock-infected group: ∗, P < 0.05; ‡, P < 0.005.
Ad-hOC-E1 Combined with Vitamin D3 Is Highly Effective against the Growth of DU145 Tumors in Vivo.
To determine the therapeutic efficacy of Ad-hOC-E1 in AI prostate cancer in vivo, we evaluated the therapeutic effect of Ad-hOC-E1 in a DU145 xenograft model in nude mice. DU145 xenograft was shown to be a very aggressive tumor that grew to 40-fold of its initial volume at 5 weeks (Fig. 8A) ⇓ . A single tail vein injection of replication-defective Ad-CMV-pA barely inhibited tumor growth, but the identical protocol of Ad-hOC-E1 administration suppressed tumor growth significantly (P < 0.05). Similarly, vitamin D3 administration alone also inhibited DU145 tumor growth (P < 0.05) in vivo. The growth of DU145 tumors was markedly repressed when animals were treated with Ad-hOC-E1 plus vitamin D3 (P < 0.005). In controls, Ad-CMV-pA plus vitamin D3 did not further enhance tumor volume reduction when compared with vitamin D treatment alone. These results demonstrated that Ad-hOC-E1 and vitamin D3 combination therapy achieved additive antitumor efficacy (P < 0.05 versus Ad-hOC-E1 alone or vitamin D3 alone).
Animal studies of combination therapy with Ad-hOC-E1 and vitamin D3. A, antitumor efficacy of Ad-hOC-E1, vitamin D3, or combination, on DU145 tumor xenografts in nude mice; versus PBS control group: ∗, P < 0.05; ‡, P < 0.005. n = 8 at day 1 and n = 4 at day 38 in all of the groups. B, viral distribution of Ad-hOC-E1 in the indicated tissues 1 week after a single i.v. administration of Ad-hOC-E1A alone. C, viral spread in s.c. DU145 tumor and liver for the indicated treated mice at 1, 3, and 5 weeks p.i. Ad5 DNA polymerase (Ad DNA Pol.) sequences (5197–5792bp in Ad5) in tissue DNA were detected using PCR amplification (B, C). One pg of purified Ad-hOC-E1 DNA was used as a positive control of PCR. D, pathologic analysis of viral-induced cytopathic effects (H&E staining) and E1A expression (immunohistochemical staining) within tumor and liver tissue in the Ad-hOC-E1 plus vitamin D3-treated mice at the indicated time frame.
To assess viral distribution after a single i.v. Ad-hOC-E1 administration, we detected by PCR analysis the Ad viral DNA sequences in the prostate, liver, lung, brain, and tumor tissues. Liver and lung were the major organs trafficking viruses. Only a few viruses were found at the s.c. tumor site at week 1 (Fig. 8B) ⇓ . Viral DNA accumulated significantly thereafter and markedly increased in weeks 3 and 5. Vitamin D3 administration enhanced viral replication/accumulation consistently in tumor tissues, but not liver, during the entire course of the treatment period (week 1 to 5, see Fig. 8C ⇓ ). Toxicology studies with Ad vectors were hampered because human adenoviruses replicate only in human cells. Immunohistochemistry data are consistent with the characteristics of Ad type 5 virus in which the E1A viral protein was expressed only in human tumor tissue but not in mouse liver (Fig. 8D ⇓ , E1A), although a steady accumulation of Ad-DNA was observed in mouse liver over 5 weeks (Fig. 8C) ⇓ . Tumor xenografts maintained in mice treated with Ad-hOC-E1 plus vitamin D3 together underwent a strong necrotic reaction within the tumor region without affecting the normal hepatocellular architecture (Fig. 8D ⇓ , H&E). These results provide preclinical evidence of the specificity, efficacy, and safety of Ad-hOC-E1 and vitamin D3 for prostate cancer gene therapy.
DISCUSSION
The progression of prostate cancer to the AI bone metastatic state is lethal. Patients with hormone-refractory bone-metastatic prostate cancer survive 9 months or less (34) . There is no effective therapy broadly targeting tumor cells with variable levels of AR and PSA. Our objective is an inducible gene therapy approach using replication-competent Ad vectors targeting the growth of human prostate cancer cells whether they express AR and PSA or not. Ad-hOC-E1 has advantages over Ad-sPSA-E1: (a) unlike As-sPSA-E1, Ad-hOC-E1 has a broad spectrum of cytotoxicity against the growth of human prostate cancer tumor cells in vitro and tumor xenografts in vivo irrespective of their basal AR and PSA status; and (b) the replication of Ad-hOC-E1 but not Ad-sPSA-E1 can be induced by vitamin D, with enhanced cytotoxicity against tumor cells in vitro and tumor xenografts in vivo.
One approach to creating conditionally replication-competent Ad vectors is to use tissue-specific promoters such as PSA to regulate the early viral gene, E1. CN706 has been engineered to include other tissue-specific promoters in tandem, such as human glandular kallikrein (hK2) or probasin promoter to selectively control transcription and translation of early viral E1A and E1B genes (14 , 15) with efficient oncolytic action in cells that only express PSA. We have designed a novel Ad-vector, Ad-hOC-E1 with a single bidirectional hOC promoter to drive the expression of both E1A and E1B genes. Because of the VDRE in the hOC promoter, viral replication can be promoted 10-fold or more with vitamin D exposure. This new version of Ad-hOC-E1 was highly efficient in destroying human prostate tumor cells irrespective of their basal AR and PSA status. We have detailed the ability of hOC promoter to drive the expression of both E1A and E1B in a bidirectional manner and the ability of this virus to inhibit the growth of human prostate cancer cell lines in vitro and tumor xenografts in vivo. There are potential advantages of using a single bidirectional promoter to drive the adenoviral replication or adenovirus-directed toxic gene expression. We have observed that two copies of the same promoter that drives both E1A and E1B genes resulted in the deletion of the E1A sequence during viral replication (data not shown). Juxtaposing promoters with a homologous region could result in homologous recombination and deletion of transgenes important for viral replication. Conversely, juxtaposing promoters with heterologous sequences could prevent homologous recombination and result in promoter competition and the squelching of transcription factors. A single promoter to drive both E1A and E1B genes could avoid the homologous recombination, promoter competition, and squelching of transcription factors during gene transcription. In the present study, we demonstrated that a single hOC promoter, driving bidirectional E1A and E1B genes, replicated efficiently in AI prostate cancer cell lines (C4-2, PC-3, and DU145). The selectivity of this bidirectional hOC promoter and its inducibility by vitamin D was demonstrated in RCC52 and prostate cancer cell lines. The inducibility of viral replication by vitamin D3 enhances cytotoxicity and efficient viral replication during the early onset of virus accumulation at the site of tumor implantation. Vitamin D-induced viral replication can also be observed in cells that have undetectable basal OC promoter activity (such as RCC52 cell lines).
Ad-hOC-E1 plus vitamin D3 may be useful for the treatment of human renal cancers as long as these cancer cell lines contain functional VDR. Because of the presence of VDR in liver cells, potential hepatotoxicity is a concern, although other transcription factors may be required for OC promoter activation. Recently, Yeung et al. found that three groups of transcription factors, Runx2, JunD/Fra-2, and Sp1, were responsible for the high hOC promoter activity in AI prostate cancer cells by binding to the OSE2, AP-1/VDRE, and OSE1 elements, respectively (35) . The specific requirement of general transcription factors and specific interaction among these transcription factors and their binding to the OC promoter could result in differential gene transcription (36, 37, 38) . The involvement of vitamin D in triggering the formation of the VDR-retinoid X receptor complexes could trigger a threshold transcriptional factor that preferentially activates the hOC promoter, raising the possibility that Ad-hOC-E1, combined with vitamin D3, may be a useful regimen for the treatment of not only prostate cancer but also other cancers, including renal cancer.
Vitamin D has been found to affect the growth of prostate cancer in preclinical experiments (39) , and there are indications that it may be useful for both prevention and treatment of prostate cancer (40, 41, 42) . However, vitamin D-mediated antiproliferation of AI prostate cancer cells is still controversial (43) . In this study, VDR was prevalently expressed in PC3 cells, followed by DU145 and C4-2 cells (Fig. 1) ⇓ . On vitamin D exposure, growth inhibition was most pronounced in C4-2 but was not observed in PC3 and DU145 cells (data not shown). In contrast to the cells grown in culture, the growth of DU145 tumor in nude mice was markedly inhibited by vitamin D3 alone without additional Ad vector (Fig. 7A) ⇓ . The precise antiproliferation mechanism(s) of vitamin D3 against DU145 tumor growth as xenograft but not in cell culture is unclear. Possibly, vitamin D is a potent antiangiogenic agent (44) that inhibits neovascularization during tumor development.
The major side effect of high-dose vitamin D administration is hypercalcemia, which could jeopardize its clinical utility. Although vitamin D3 analogue, Ro 25-9022, used in the present study, has not been tested at the maximum tolerated dose, when mice treated with 4 ng of Ro 25-9022 twice a week for 3 weeks were compared with vehicle-treated mice, they showed only a mild side effect, a 10% body weight reduction for a period of 3 weeks, and they recovered after the vitamin D3 treatment was stopped (43) . The next goal of our study is to establish the optimal vitamin D3 treatment protocol to minimize its side effects. A dual modality strategy combining Ad-E1 and vitamin D3 may be translated rapidly into the clinic for the treatment of men with hormone-refractory metastatic prostate cancer.
Acknowledgments
We are grateful to Drs. Mary Hitt and Robert Sikes for advice and discussion and to Gary Mawyer for excellent editorial assistance.
Footnotes
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
↵1 Supported in part by NASA (NCC8-171), NIH (CA85555), and CaP CURE Foundation.
↵2 To whom requests for reprints should be addressed, at Molecular Urology and Therapeutics Program, Department of Urology, Emory University School of Medicine, Atlanta, GA 30322, Phone: (404) 778-4845; Fax: (404) 778-3965; E-mail: chsieh2{at}emory.edu
↵3 The abbreviations used are: Ad, adenoviral vector; hOC, human osteocalcin; PSA, prostate-specific antigen; sPSA, super PSA; VDRE, vitamin D-responsive element; AR, androgen receptor; VDR, vitamin D receptor; AD, androgen dependent; AI, androgen independent; MOI, multiplicity/multiplicities of infection; FACS, fluorescence-activated cell sorting/sorter; p.i., postinfection; pfu, plaque-forming unit(s); RT-PCR, reverse transcription-PCR; CMV, cytomegalovirus; GAPDH, glyceraldehyde 3-phosphate dehydrogenase; CAR, coxsackie and adenovirus receptor.
- Received December 20, 2001.
- Accepted April 15, 2002.
- ©2002 American Association for Cancer Research.