- Email: [email protected]
Treatment of ovarian cancer with a novel dual targeted conditionally replicative adenovirus (CRAd)
Gynecologic Oncology 105 (2007) 113 – 121 www.elsevier.com/locate/ygyno
Treatment of ovarian cancer with a novel dual targeted conditionally replicative adenovirus (CRAd) Rodney P. Rocconi a,⁎, Zeng B. Zhu b , Mariam Stoff-Khalili b , Angel A. Rivera b , Baogen Lu b , Minghui Wang b , Ronald D. Alvarez a , David T. Curiel b , Sharmila K. Makhija a b
a Division of Gynecologic Oncology, University of Alabama at Birmingham, Birmingham, Alabama, USA Division of Human Gene Therapy, Departments of Medicine, Surgery, Pathology, and the Gene Therapy Center, University of Alabama at Birmingham, Birmingham, Alabama, USA
Received 15 June 2006 Available online 14 December 2006
Abstract Objectives. Current virotherapy strategies for ovarian cancer have been hampered by limitations in target cell infectivity and nonspecific tissue replication. In an effort to circumvent these limitations, we evaluated various CRAds modified to incorporate novel capsid targeting motifs (RGD and chimeric Ad5/3) with a novel tissue-specific promoter (CXCR4). Methods. Two novel CRAds (Ad5-CXCR4-F5/3 and Ad5-CXCR4-RGD) were constructed via homologous recombination and verified by PCR and DNA sequencing. The infectivity and viral replication rates of these two CRAds were analyzed via quantitative real-time PCR (QRTPCR) in cell line experiments using three ovarian cancer cell lines (SKOV3.ip1, Hey, and OV4) and compared to that achieved with a clinical grade CRAd (Δ24-RGD) to be evaluated in a Phase I trial. Cytocidal effects were determined by crystal violet staining in these same cell lines infected with different concentrations of viral particles per cell (0, 0.1, 1, 10, 100, and 500). Additionally, viral replication was evaluated by QRTPCR in primary ovarian cancer tissue slices from multiple patients with ovarian cancer as well as in primary human normal liver tissue slices in order to establish CRAd selectivity. All experiments incorporated appropriate controls and repeated in triplicate. Results. Compared to RGD-capsid CRAds (Δ24-RGD and CXCR4-RGD), the F5/3-capsid CRAd (CXCR4-F5/3) demonstrated significant improvements in infection rates (p = 0.025, 0.006, and 0.006) in all ovarian cancer cell lines tested (SKOV3.ip1, Hey, and OV4, respectively). In addition to improved transduction of virus into the cells, the TSP CXCR4-based CRAds demonstrated improved viral replication. Specifically, CXCR4-F5/3 further enhanced viral replication 89-fold (p = 0.009, 0.010, 0.003) in the same cancer cell lines. Furthermore, CXCR4-F5/3 showed a 4-log improvement in oncolytic potential over Δ24-RGD. In the ex vivo primary ovarian tissue slices, CXCR4-F5/3 showed a 58-fold improvement in viral replication (p = 0.005) compared to the clinical grade Δ24-RGD. Both CXCR4-F5/3 and CXCR4-RGD demonstrated significant reduction of viral replication in normal liver slices (p = 0.001). Conclusions. These data suggest that a dual targeted approach is feasible for the combined enhancement of infectivity and replication in ovarian cancer with a specificity that was attenuated in normal liver tissues. In fact, CXCR4-F5/3 outperformed our best CRAd agent to date nearly 60-fold in our most stringent ex vivo model of primary ovarian cancer tissue slices and suggests that this novel agent could be useful for the treatment of ovarian cancer. © 2006 Elsevier Inc. All rights reserved. Keywords: Ovarian cancer; Gene therapy; Virotherapy; Adenovirus; CXCR4; Tumor-specific promoter; Capsid modification
Introduction ⁎ Corresponding author. Department of Obstetrics and Gynecology, University of Alabama at Birmingham, 619 20th Street South, Old Hillman Building-Room 538, Birmingham, AL 35249-7333, USA. Fax: +1 205 975 6174. E-mail address: [email protected] (R.P. Rocconi). 0090-8258/$ - see front matter © 2006 Elsevier Inc. All rights reserved. doi:10.1016/j.ygyno.2006.10.057
Greater than 20,000 women are expected to be diagnosed with ovarian cancer during the year 2006 in the United States alone [1]. Although the majority of patients will present with advanced disease, most will respond to cytoreductive surgery and first-line platinum-based combination chemotherapy.
114
R.P. Rocconi et al. / Gynecologic Oncology 105 (2007) 113–121
Despite this response, effective curative therapy has yet to be determined and thus many patients with advanced disease will experience a recurrence and unfortunately succumb to progressive disease [2–4]. In this regard, novel therapeutics, such as virotherapy, are currently being sought for the treatment of advanced ovarian cancer. Specifically, virotherapy utilizing conditionally replicative adenoviruses (CRAd) represents one such advanced modality for the treatment of ovarian cancer. Despite the promise of CRAds, clinical outcomes with virotherapy have been disappointing to date. Two limitations restricting the clinical efficacy include imprecise tumor cell replicative specificity and inefficient tumor cell infectivity. The recognition of these limits has lead to strategies designed to address these barriers. On this basis, advanced generation CRAd agents have been proposed, which seek to achieve improved replicative specificity and/or enhanced infectivity. For improving tumor cell replication specificity, tumorspecific promoters (TSPs) selectively drive adenoviral E1 gene expression in tumor cells and thereby accomplish transcriptional targeting based upon the induced specificity of viral replication [5]. The optimal TSP restricts replication to tumor cells alone, thereby having the highest activity in tumor cells (“tumor on”) and lowest activity in normal cells (“normal off”). Of note, adenoviral vectors possesses a predilection for hepatocytes thereby making hepatotoxicity the dose-limiting toxicity for these vectors. In this regard, the ideal TSP-directed adenovirus would have the highest “tumor on/liver off” ratio and thus possess the optimal therapeutic index. One such promoter, CXCR4 has been recently investigated as a good candidate for cancer gene therapy. Several authors have demonstrated that CXCR4 gene expression has been undetectable in normal ovarian epithelial cells but markedly upregulated in ovarian cancer cell lines and primary ovarian cancers [6–8]. In our previous work, we have reported that the CXCR4 promoter had a “tumor on” and “liver off” phenotype in both in vitro and in vivo experiments in ovarian cancer [9]. In other studies, the therapeutic index of the CXCR4 promoter was evaluated with other promoters commonly evaluated in ovarian cancer (survivin, SLPI, and Cox-2) (Makhija, personal communication). In these studies, the CXCR4 promoter demonstrated a superior therapeutic index with greater activity in ovarian cancer cell lines and primary ovarian tumor, while demonstrating low levels of activity in human liver tissue. Thus, the CXCR4 promoter appears to be an excellent candidate for use in transcriptional targeting for ovarian cancer gene therapy. For enhancing tumor cell infectivity, the targeting adenoviral vector to tumor cells via alternative pathway is of importance. Native Ad5 tropism is mediated by two capsid proteins: the fiber and the penton base. These proteins bind to the primary high affinity cellular receptor, coxsackie–adenovirus receptor (CAR), and to the integrins αvβ3 and αvβ5, respectively. Tumor is resistant to adenovirus infection due to a relative paucity of the primary receptor CAR on tumor cell surface [10–12]. Based on these molecular interactions, a concerted effort has been made to modify Ad5 tropism, resulting in enhanced tumor cell transduction by retargeting
cellular entry through heterologous pathway or CAR-independent pathway. An example of this is the utilization of RGD motif in the fiber knob of the Ad. This capsid modification appears to facilitate Ad binding and entry into tumor cells via integrin receptors that are abundantly expressed on tumor cells [13,14]. Additional capsid modifications have been explored to obtain infectivity enhancement of Ads including AdF5/3, which substituted the Ad5 fiber with the Ad3 fiber [15,16], Ad5-pk7 in which a motif polylysine (pk7) was genetically inserted at C terminus of Ad fiber and bound to heparin sulfatecontaining receptor [17], and Ad5-CK in which canine Ad knob was used to replace Ad5 knob [18]. In this study, we utilized two capsid modifications (RGD and F5/3) that target to integrins and CD80, 86, and 46, respectively. These modifications enhance the viral infectivity via a CAR-independent pathway [19,20]. Our objective was to improve the viral-specific replication in tumor cells and enhance tumor cell infectivity via an optimal CRAd for the treatment of ovarian cancer. We constructed two novel CRAds (Ad5-CXCR4-RGD and Ad5-CXCR4-F5/3) in which the viral replication is under the control of the CXCR4 TSP and the viral infectivity enhanced via the capsid modifications (RGD and F5/3). Both tumor specificity and infectivity of these novel CRAd agents were evaluated and compared with Δ24RGD. Δ24RGD is a first-generation CRAd that utilizes a partial deletion (24 base pair) of the E1 region which allows a relative selectivity by depending upon the target cells' machinery to complete replication. This CRAd has been approved for use in an NIH-funded Phase I trial. Materials and methods Cells and tissues Human ovarian tumor cell lines, SKOV3.ip.1, OV4, and Hey were cultured in medium suggested by ATCC. 911 cells (a kind gift from Dr. Van Der Eb, Leiden University, The Netherlands) were maintained in Dulbecco's modified Eagle medium. Each medium was supplemented with 10% fetal calf serum, penicillin (100 IU/ml), and streptomycin (100 μg/ml). Cells were incubated at 37°C in a 5% CO2 environment under humidified conditions. Following IRB approval, human ovarian specimens were obtained for tissue slices from epithelial ovarian cancer remnants not needed for diagnostic purposes during primary cytoreductive surgery performed at the University of Alabama at Birmingham (UAB). Time from harvest to tissue slicing was kept at an absolute minimum (< 2 h). To generate tissue slices, a coring device (Alabama Research Development, Munford, AL) was used to create an 8-mm diameter core of tissue from the ovarian cancer. This core was cut in consecutive 0.25-mm-thick slices using the Krumdiek tissue slicer (Alabama Research Development, Munford, AL) with the reciprocating blade at 30 rpm. Sequential slices were then cultured in 12-well plates in RPMI medium supplemented with 10% bovine fetal serum, 100 U/ml of penicillin, and 100 μg/ml of streptomycin. Human liver samples were obtained (Department of Surgery, University of Alabama at Birmingham) from three seronegative donor livers prior to transplantation into recipients. All liver samples were flushed with University of Wisconsin (UW) solution (ViaSpan, Barr Laboratories, Inc., Pomona, NY) before harvesting and kept on ice in UW solution until slicing with the Krumdiek tissue slicer. Time from harvest to slicing was kept at an absolute minimum (<2 h). Liver slices were placed into 6-well plates (1 slice per well) containing 2 mL of complete culture media (William's Medium E with 1% antibiotics, 1% L-glutamine, and 10% bovine fetal serum).
R.P. Rocconi et al. / Gynecologic Oncology 105 (2007) 113–121 All cultures were maintained at 37°C in a humidified atmosphere of 95% air and 5% CO2. Tissue slices were incubated for up to 48 h on a plate rocker set at 60 rpm in order to agitate slices to ensure adequate oxygenation and viability.
Recombinant adenovirus The CRAd genomes were constructed via homologous recombination in Escherichia coli (Fig. 1). CRAd-CXCR4-RGD has the following characteristics: (1) The CRAd agent contains the human CXCR4 promoter nucleotide − 191/ +8826 to drive E1 expression. The CXCR4-controlled E1 expression cassette was placed in the original E1 region of the Ad gene. (2) A RGD-4C capsid modification was inserted into the Ad fiber knob region for enhancement of Ad infectivity.32 (3) The E3 gene was retained in the Ad genome for elevating the oncolytic effect of the CRAd agents.33 And (4) a poly-A signal was inserted between the inverted terminal repeat (ITR) and the CXCR4 promoter to stop the nonspecific transcriptional activity of the ITR and to retain the tumor specificity of the CXCR4 promoter. The CRAd-CXCR4-F5/3 vector has the following characteristics: (1) The CRAd agent contains the human CXCR4 promoter nucleotide −191/+88 to drive E1 expression [9]. The CXCR4-controlled E1 expression cassette was placed in the original E1 region of the Ad gene. (2) The capsid modification has modified fiber genes that contain an Ad5 shaft region and Ad3 knob region for enhancement of Ad infectivity [21]. (3) The E3 gene was retained in the Ad genome to preserve the oncolytic effect of the CRAd agent [22]. (4) A poly-A signal was inserted between the inverted terminal repeat (ITR) and the CXCR4
115
promoter to stop the nonspecific transcriptional activity of the ITR and to maintain the tumor specificity of the CXCR4 promoter. DNA fragments containing nucleotides −191/+ 88 were cut with BamHI and HindIII restriction endonucleases from the clone pBSKCAT/CXCR4 3B/4-1 [5′Δ3] [23] and subcloned into the plasmid pBSSK (Stratagene, La Jolla, CA) by use of the same restriction sites. A SV40 PA fragment was cut with XbaI/BamHI from a pGL3B vector (Invitrogen, Carlsbad, CA) and inserted into the pBSSK via the same restriction sites. pBSSK/PA/CXCR4, the generated clone, was used to create shuttle vectors. Next, DNA fragments containing both an SV40 PA and the CXCR4 promoter were cut with NotI/XhoI and subcloned into the E1 gene containing pScsE1 plasmid (a kind gift from Dr. Dirk Nettelbeck, Department of Dermatology, University Medical Center-Erlangen, Erlangen, Germany). Thus, the plasmid pScsE1/PA/CXCR4 was the result. The Ad vector, Pvk500F5/3 was a kind gift from Dr. Takayama and contains both the E3 gene and a capsid modified F5/3 in Ad5 backbone [24]. After cleavage with PmeI, the shuttle vector, pScsE1/PA/CXCR4, was recombined with Pvk500F5/3 to generate a CRAd genome with an F5/3 modified fiber. The resultant plasmids encoding the CXCR4 promoter were linearized with PacI and transfected into 911 cells using Lipofectamine (Qiagen, Valencia, CA). Resultant viruses that were generated were propagated in A549 cells (a lung cell line in which the CXCR4 gene is overexpressed) and then purified by double CsCl density gradient centrifugation, followed by dialysis against phosphate-buffered saline (PBS) with 10% glycerol. The viruses were titrated by plaque assay and the vp number was determined by spectrophotometrically based on the absorbance at a wavelength of 260 nm. Viruses were stored at − 80°C until use.
Fig. 1. Characterization of CRAd. (a) Construction of the CRAd agents. A 279 bp of the CXCR4 promoter was amplified from the clone pBSKCAT/CXCR4 3B/4-1 [5′Δ3] and constructed in a p ScE1 vector. A poly-A signal sequence was inserted between the ITR and CXCR4 promoter to terminate the transcription signal from the ITR. Constructed clones were recombined with pAdback5/3 to generate the CRAd-CXCR4.F5/3 (RGD-4C motif used to generate the CRAd-CXCR4.RGD). (b) The PCR products were run on an agarose gel (1%) that amplified the oligo pair b/c (CXCR4) and oligo pair a/c (PolyA signal + CXCR4) from the template DNAs of CRAdCXCR4.F5/3 and a negative control reAdGL3.
116
R.P. Rocconi et al. / Gynecologic Oncology 105 (2007) 113–121
In the analysis of CRAd agents, replication positive and negative controls were wild-type Ad5 (Adwt) and reAdGL3B-CXCR4, respectively.
Analysis of CRAd construct by PCR and sequencing The structures of the CRAds were verified by PCR and sequence data. The virus [5 × 108 viral particles (vp)] was processed with a blood DNA kit (Qiagen), and 1/50 of the DNA solution was analyzed by PCR for various regions with Tag polymerase (Qiagen). Thermal cycling conditions were as follows: initial denaturing, 5 min at 95°C; 30 cycles of 30 s each at 95°C, 30 s at 55°C, and 1 min at 72°C. The primers used for each region were as follows: (a) PA-F, 5′ GATGAGTTTGGACAAACCAC; (b) CXCR4-F, 5′TACCGACCACCCGCAAACA; (c) CXCR4-R, 5′AACCGCTGGTTCTCCAGATGC. The PCR products were analyzed by automatic sequence analysis performed by using the BigDye Terminator v3.1 Cycle Sequencing Ready Reaction kit and an Applied Biosystems 3100 Genetic Analyzer (Foster City, CA) at the Genomics Core of the Howell and Elizabeth Heflin Center for Human Genetics of the University of Alabama at Birmingham.
Detection of the E1 gene from the CRAd agent by using quantitative real-time PCR The CRAd-CXCR4 was diluted to 103, 104, and 105 viral particles (vp)/μl with distilled water and added to real-time (RT)-PCR reactions as described [26]. RT-PCR was performed, and the E1 gene copy number was determined by using an oligo pair (forward primer—5′AACCAGTTGCCGTGAGAGTTG, reverse primer—5′CTDGTTAAGCAAGTCCTCGATACT) and the probe (5′6FAM-CACAGCTGGCGACGCCCA-TAMRA-3′). The oligos were designed using Primer Express 1.0 (Perkin-Elmer, Foster City, CA) and synthesized by Applied Biosystems. A nonreplicative control, AdGL3BCXCR4, was used in parallel, in which the Ad5 E1 gene was deleted.
Analysis of infectivity of CRAd agents in tumor cell lines SKOV3.ip1, OV4, and Hey ovarian cancer cells at 105 were cultured as discussed previously. Cells were infected in triplicate with CXCR4.RGD, CXCR4. F5/3, Δ24.RGD, Adwt, or reAdGL3B.CXCR4 at 1, 10, 100, and 1000 vp/cell in 2% FBS infection medium. After a 3-h incubation at 37°C in a 5% CO2 environment, infection medium was aspirated and cells were washed three times with PBS so to remove uninternalized viruses. Cells were trypsinized for 15 min, and 1 ml of PBS was added. DNA was extracted from the specimens, via the DNeasy Tissue Kit (Qiagen, Valencia, CA). The Ad5 E1 gene was detected in DNA samples and quantitative RT-PCR was performed as previously described [25]. Ad E1 gene copy numbers were detected and normalized with human β-actin.
Analysis of replication of CRAd agents in tumor cell lines SKOV3.ip1, OV4, and Hey ovarian cancer cells at 105 were cultured as previously discussed with 100 vp/cell of CXCR4.RGD, CXCR4.F5/3, Δ24.RGD, Adwt, or reAdGL3B.CXCR4 in infection medium containing 2% FBS and incubated at 37°C in a 5% CO2 environment. After incubation for 3 h, infection medium was aspirated, and the cells were washed three times with PBS in order to remove uninternalized viruses. Fresh 10% FBS culture medium was then replaced and the cells were returned to incubation at 37°C in a 5% CO2 environment. Media from triplicate wells were collected 1, 3, and 9 days later. DNA was extracted from the specimens via the DNeasy Tissue Kit (Qiagen, Valencia, CA). The Ad5 E4 gene was detected in DNA samples by using an oligo pair (forward primer—5′ GGAGTGCGCCGAGACAAC; reverse primer—5′ACTAGGTCCGGCGTTCCAT; probe ORF6—TGGCATGACACTACGACCAACACGATCT), and quantitative RT-PCR was performed as described above. Ad E4 gene copy numbers were detected and normalized with human β-actin.
Replication of CRAds in human tissue slices Excess ovarian cancer tissues not needed for diagnostic purposes were obtained from three ovarian cancer patients and primary liver tissue was obtained as previously described. The Krumdieck tissue slicer was used to
generate ovarian cancer and liver tissue slices and was cultured at the conditions described previously. Each tissue slice was estimated to contain 1 × 106 cells/ slice based on a 10-cell thickness (∼ 250 μm) and an 8-mm-slice diameter. Slices were infected in triplicate with 500 vp/cell of CXCR4.RGD, CXCR4.F5/3, Δ24. RGD, Adwt, or controls in fresh infection medium. Slices were incubated at 37°C in a 5% CO2 environment on a rocking plate. After 24 h and 72 h of incubation times, respectively, total DNA was extracted from primary human ovarian tumor slices and liver slices via the DNeasy Tissue Kit (Qiagen, Valencia, CA). DNA samples were treated with DNase-free RNase to remove possible RNA contamination and stored at − 80°C until use. Ad E4 gene copy numbers were detected as previously described.
In vitro analysis of cytocidal effects The in vitro cytocidal effects of the CRAds were analyzed by determining the viability of cells with crystal violet staining after infection with respective CRAd. SKOV3.ip1, OV4, and Hey ovarian cancer cells were plated at 5 × 106 cells/well on a 24-well plates and infected with 0, 0.1, 1, 10, 100, and 500 vp/cell with CXCR4.RGD, CXCR4.F5/3, Δ24.RGD, or reAdGL3B. CXCR4. After a 3-h infection, the infection medium was replaced with the corresponding appropriate complete medium. After 10 days of cultivation, the cells were fixed with 10% buffered formalin for 10 min and stained with 1% crystal violet in 70% ethanol for 20 min, followed by washing with tap water and set to air dry.
Results Attributes of the CXCR4-F5/3 CRAd The vectors shown in Fig. 1 were constructed, amplified in 911 cells, and upscaled in A549 cells. The yield was 2 × 1012 vp/ ml with a vp/plaque-forming unit ratio of 62. The vector structure was confirmed by PCR of the viral DNA (Fig. 1) and DNA sequence data from PCR products (not shown). Furthermore, CXCR4-F5/3 CRAd agent possessed the E1 gene. The E1 gene was detected by QRT-PCR, the tested dilutions of 103, 104, and 105 vp, in a dose response manner, but not in the nonreplicating viral vector of AdGL3B.CXCR4 in which the E1 gene was deleted. In our CRAd agent, the native E1 promoter was replaced by the CXCR4 promoter, which was evidenced by the propagation of the vector in the lung tumor cell line A549 in which the CXCR4 gene was overexpressed. The Ad type 5 backbone with the F5/3 modification allows the retargeting of the adenoviral vector towards a CAR-independent pathway for Ad infectivity enhancement [26]. Evidence of CRAd infectivity in tumor cell lines To further assay CRAd-CXCR4.F5/3, 3 ovarian cancer cell lines SKOV3.ip1, OV4, and Hey were plated in 24-well plates and cells were infected with 1, 10, 100, and 1000 vp/cell of CXCR4.RGD, CXCR4.F5/3, Δ24.RGD, or reAdGL3B.CXCR4. After a 3-h infection, the cells were washed three times with PBS to remove noninternalized viruses and provided with fresh substrate. The Ad5 E1 copy numbers should relate to the relative levels of viral CRAd vectors that entered the tumor cells and thus should correlate to the CRAd's ability to infect ovarian cancer cells. As expected, the number of viruses which infected the ovarian cell lines increased as the vp/cell increased (Fig. 2). The CRAd agent CXCR4.F5/3 showed superior results in all three
R.P. Rocconi et al. / Gynecologic Oncology 105 (2007) 113–121
117
with 100 vp/cell of CXCR4.RGD, CXCR4.F5/3, Δ24.RGD, Adwt, or reAdGL3B.CXCR4. After a 3-h infection, the cells were washed with PBS three times and provided fresh medium. The presence of E4 gene was determined by using QRT-PCR from DNA samples extracted after 1, 3, and 9 days postinfection. The Ad5 E4 copy numbers should relate to the levels of released virions from the tumor cells and therefore correlate to the replication rates of the CRAd agents. Data depicted in Fig. 3 demonstrate that CXCR4-F5/3 has superior replicative capabilities in the three ovarian cell lines tested. In Day 9 samples, this translated into 1.7- to 2.7-fold increase compared to CXCR4.RGD and 24.6- to 213-fold replication enhancement over Δ24.RGD. Compared to Δ24.RGD, CXCR4.F5/3
Fig. 2. Transductional activity in ovarian cancer cell lines. Levels depict the amount of viral CRAd vectors that infected the ovarian cancer cells in each of the 3 cell lines. In all cell lines tested, CXCR4-F5/3 demonstrated the highest level of infectivity at all MOIs. (a) SKOV3.ip.1: 1 vp/cell (p = 0.9); 10 vp/cell (p = 0.03); 100 vp/cell (p = 0.05); and 1000 vp/cell (p = 0.025). (b) Hey: 1 vp/cell (p = 0.08); 10 vp/cell (p = 0.03); 100 vp/cell (p =0.07); and 1000 vp/cell (p =0.006). (c) OV4: 1 vp/cell (p = 0.08); 10 vp/cell (p = 0.02); 100 vp/cell (p =0.03); and 1000 vp/cell (p =0.006).
ovarian cell lines tested at nearly all levels of MOI. Compared to RGD-capsid CRAds (Δ24.RGD and CXCR4.RGD), the F5/3capsid CRAd (CXCR4.F5/3) demonstrated significant improvements in infection rates at the highest MOI 1000 vp/cell (p = 0.025, 0.006, and 0.006) in SKOV3, Hey, and OV4 cell lines respectively. Specifically, the infection level of CXCR4.F5/ 3 (1000 vp/cell) resulted in a 2-fold, 6.7-fold, and nearly 3-fold improvement in infectivity compared to Δ24.RGD in SKOV3. ip1, OV4, and Hey cells, respectively. Compared to CXCR4. RDG, the infectivity levels of CXCR4-F5/3 showed a 5.9-fold, 4.2-fold, and 9.5-fold enhancement in the same cell lines. Evidence of CRAd replication in tumor cell lines To assess CRAd replication, SKOV3.ip1, OV4, and Hey cell lines were plated (105) in 24-well plates and cells were infected
Fig. 3. Transcriptional activity in ovarian cancer cell lines. Levels shown represent the amount of viral replication noted at 1, 3, and 9 days postinfection. (a) SKOV3.ip.1; CXCR4-F5/3 improved replication—Day 1 (p = 0.13); Day 3 (p = 0.008); and Day 9 (p = 0.009). (b) Hey; CXCR4-F5/3 improved replication—Day 1 (p = 0.003); Day 3 (p = 0.001); and Day 9 (p = 0.01). (c) OV4; CXCR4-RGD improved replication on Day 1 (p = 0.003) and Day 3 (p = 0.001); CXCR4-F5/3 improved replication on Day 9 (p = 0.01).
118
R.P. Rocconi et al. / Gynecologic Oncology 105 (2007) 113–121
with 500 vp/cell of CXCR4.RGD, CXCR4.F5/3, Δ24.RGD, Adwt, or reAdGL3B.CXCR4. Media were collected after 24 and 72 h, and DNAwas isolated as previously described for detection of E4 copy number via QRT-PCR. The ratios of DNA copy numbers between Day 1 and Day 3 for CXCR4.RGD were 272%, 254%, and 760% for patients 1, 2, and 3, respectively. The ratios for CXCR4.F5/3 were 1602%, 2220%, and 1854% while the ratios for Δ24.RGD were 180%, 85%, and 235% in the same 3 patients. As seen in the ovarian cell lines, CXCR4.F5/3 demonstrated superiority to the other CRAd's replication rates in all 3 patients evaluated. In these primary tumors, improvements in Day 3 replication were seen in comparison to both CXCR4. RGD (2.3- to 7.3-fold) and Δ24.RGD (14- to 84-fold). CXCR4. F5/3 showed a 58-fold improvement in viral replication (p = 0.005) compared to the clinical grade Δ24.RGD. Evidence of CRAd replication in human liver tissue slices Because Ad replication is species specific and because the major Ad toxicity in gene therapy trials is hepatotoxicity, we tested the CXCR4 promoter activity in human liver specimens in organ culture. Human liver slices were infected with CXCR4. RGD, CXCR4.F5/3, and Adwt. Replication was measured by detecting the presence of the E4 gene by using QRT-PCR from DNA samples extracted after 1 and 3 days post-infection. Compared to the positive control (Adwt), the TSP-directed CRAds demonstrated a significant reduction in replication in liver tissue on both Day 1 (p = 0.001) and Day 3 (p < 0.001) (Fig. 5), thus showing that the CXCR4 promoter has a “liver off” phenotype that should result in low toxicity to the human host liver. Evidence of CRAd cytolysis in tumor cell lines All CRAds were tested for its cell-killing potential via cytolysis in the three ovarian cell lines utilized. Oncolysis was evaluated after 10 days of incubation via crystal violet staining Fig. 4. Transcriptional activity in primary ovarian cancer tissue slices. Human ovarian cancer tissue slices were infected with 500 vp/cell of Ad vectors. After a two-day incubation period, DNA was isolated from each slice and the E4 levels determined by quantitative PCR after normalization. All patients showed a significant increase in Day 3 replication from Day1. CXCR4-F5/3 demonstrated superior replication rates in all 3 patients evaluated, with a 58-fold improvement compared to clinical grade Δ24.RGD.
enhanced the viral replication an average 89-fold in the three ovarian cancer lines evaluated (SKOV3.ip1, p = 0.009; Hey, p = 0.010; OV4, p = 0.003). Evidence of CRAd replication in ovarian tumor tissue slices In order to verify the replication of this promoter under nearclinical conditions, in two novel CRAd agents, we determined the replication of the CRAds in three primary ovarian cancer patients with histologically documented papillary serous adenocarcinomas (Fig. 4). The samples were handled and sliced as described in the Materials and methods section and infected
Fig. 5. Transcriptional activity in primary normal liver tissue slices. Human liver tissue slices were infected with 500 vp/cell of Ad vectors. DNA was isolated on Day 1 and Day 3 and E4 levels determined by QRT-PCR after normalization. Compared to the positive control (Adwt), the TSP-directed CRAds demonstrated a significant reduction in replication in liver tissue on both Day 1 (p = 0.001) and Day 3 (p < 0.001).
R.P. Rocconi et al. / Gynecologic Oncology 105 (2007) 113–121
(Fig. 6). While reAdGL3B.CXCR4 (replication-incompetent) possessed no cytotoxic effect as expected, the CXCR4-based CRAds demonstrated potent oncolytic potential in all 3 cell lines. In fact, CXCR4.RGD showed a 2- to 3-log improvement in oncolysis compared to Δ24.RGD, with some oncolysis at levels as low as 1 vp/cell. Likewise, CXCR4.F5/3 resulted in a 4- to 5-
Fig. 6. Cytotoxic efficiency of Ad vectors. 5 × 104 cells from each ovarian cancer cell line (SKOV3.ip1, Hey, and OV4) were plated on 24-well plates, and infected with Ad vectors at different MOIs (0, 0.1, 1, 10, 100, and 500). Cells were stained with crystal violet after a 10-day incubation. In all 3 cell lines evaluated, CXCR4-F5/3 demonstrated at least a 2-log and 4-log improvement in cell death over CXCR4-RGD and Δ24.RGD, respectively. (a) SKOV3.ip.1, (b) Hey, and (c) OV4.
119
log improvement compared to Δ24.RGD and a 2-log improvement compared to CXCR4.RGD. This cytolytic effect of CXCR4.F5/3 was seen at 0.1 vp/cell in all 3 ovarian cell lines. Discussion Effective virotherapy for the treatment of cancer is dependent on their capability to infect and replicate in target cells. Thereby, the oncolytic potential is directly relative to the ability of viral replication. Most adenoviruses utilized are serotype 5 (Ad5) which binds to CAR. Unfortunately, CAR expression varies widely on many tumor cells including ovarian cancer cells. This could potentially hinder viral infection, replication, and subsequent oncolysis [26]. Therefore, methods to overcome these limitations involve transductional and transcriptional targeting to improve the infectivity and selective replication of CRAd agents. Fiber capsid modifications, such as F5/3, enhance viral infectivity via utilizing CAR-independent pathways for transduction into the cell. Once infected, tumor-specific promoters (TSP) place the adenovirus under the control of heterologous control regions that limit the replication to specific tissues or tumors. In this study, we have demonstrated that capsid or fiber modification enhances infectivity and allows improved antitumor potential of replicative CRAds in ovarian cancer. Furthermore, by incorporating an ovarian cancer relevant TSP in these infectivity-enhanced capsid/fiber modified CRAds, we showed ovarian cancer cell-specific adenoviral replication and oncolysis. These principals were tested in comparison to a clinical grade CRAd in multiple ovarian cancer cell lines and in primary ovarian tissue slices, thus giving us results in stringent models in relation a CRAd set for human clinical trials. Of interest is the significant “liver off” profile demonstrated by the CXCR4 TSP, which should reduce the dose-limiting toxicity of hepatotoxicity of adenoviral trials to date. Additionally, CXCR4.F5/3 attains improved oncolysis by both enhanced infectivity and selective replication in vitro. Importantly, we confirmed this combined infectivity and replication in a stringent human model system utilizing primary human ovarian cancer tissue slices and normal liver slices. Therefore, our experiments achieved significant improvements in regards toward overcoming the limitations of poor infectivity and poor specificity previously seen in CRAd agents. Other limitations previously demonstrated and yet to be circumvented encompasses the human immune response, which limits the effectiveness of multiple treatments with adenoviral vectors. This investigation evaluated the efficacy of an infectivity enhanced capsid modified CRAd utilizing an ovarian cancer cell TSP, CXCR4. As previously mentioned, this TSP has been shown to exist preferentially in ovarian tumor cells while undetectable in normal ovarian tissue. Additionally, ovarian cancer proliferation, migration, and metastasis have been linked to the CXCR4 binding of its ligand, SDF-1, thereby making it a good candidate for a TSP. We demonstrated that the novel CRAd agents containing the CXCR4 TSP improved transcriptional targeting over a clinical grade CRAd, Δ24.RGD. Future investigations could be directed towards correlating the
120
R.P. Rocconi et al. / Gynecologic Oncology 105 (2007) 113–121
level of CXCR4 expression in tumor to the response to CXCR4 CRAds. The paucity of native adenoviral receptors (CAR) in ovarian tumor necessitates the evaluation of CAR-independent pathways to enhance infectivity. This enhanced infectivity was achieved by replacing the Ad5 fiber knob domain with the knob domain from human adenovirus serotype 3 (Ad5/3). Other capsid modification strategies have been investigated in order to circumvent the poor infectivity of adenoviruses in tumor cells by utilizing CAR-independent receptors. These studies included the incorporation of short heterologous peptides like RGD, or polylysine into the fiber knob domain [27], or knob switching strategies [29] such as xenotype knob switching [18,28,29]. Although both RGD and F5/3 serve as potential capsid modifications for the enhanced infectivity of adenoviruses, we confirmed the previously described superiority of F5/3 compared to RGD [16]. In conclusion, we present a CRAd addressing the obstacles seen in adenoviral cancer gene therapy, by infectivity enhancement and specific replication with subsequent cell killing in ovarian cancer. Specifically, the TSP CXCR4 emerges as an especially promising transcriptional targeting approach for ovarian cancer. Our study using CXCR4.F5/3 demonstrated improved infectivity, selective replication, and enhanced oncolytic potential over all other CRAds tested, including the clinical grade Δ24.RGD. Considering the significant improvements of a “tumor on/liver off” CRAd evaluated in stringent well-established ex vivo models (primary tissue slices), this CRAd potentially could overcome the limitations of poor infectivity and poor specificity previously seen in other novel CRAd agents in cancer therapy. Although, further evaluation is needed, the promoter CXCR4 and our construct CXCR4.F5/3 could represent a novel gene therapy agent for the treatment of ovarian cancer. Acknowledgment This work was performed with NIH grant #5R01CA09054703 (8/9/01–1/31/08). References [1] Jemal A, Murray T, Ward E, Samuels A, Tiwari RC, Ghafoor A, et al. Cancer statistics. CA Cancer J Clin 2005;55:10–30. [2] Alberts DS, Green S, Hannigan EV, O'Toole R, Stock-Novack D, Anderson P, et al. Improved therapeutic index of carboplatin plus cyclophosphamide versus cisplatin plus cyclophosphamide: final report by the Southwest Oncology Group of a phase III randomized trial in stages III and IV ovarian cancer. J Clin Oncol 1992;10:706–17. [3] McGuire WP, Hoskins WJ, Brady MF, Kucera PR, Partridge EE, Look KY, et al. Cyclophosphamide and cisplatin compared with paclitaxel and cisplatin in patients with stage III and stage IV ovarian cancer. N Engl J Med 1996;334:1–6. [4] Ozols RF, Bundy BN, Greer BE, Fowler JM, Clarke-Pearson D, Burger RA, et al. Phase III trial of carboplatin and paclitaxel compared with cisplatin and paclitaxel in patients with optimally resected stage III ovarian cancer: a Gynecologic Oncology Group study. J Clin Oncol 2003;21: 3194–200. [5] Haviv YS, Curiel DT. Engineering regulatory elements for conditionallyreplicative adeno-viruses. Curr Gene Ther 2003;3:357–85.
[6] Li F, Zhu HS, Han ZQ, Chen G, Gao QL, Jia P, et al. Effects of chemokine receptor and its ligand on migration of ovarian cancer cells. Ai Zheng 2005;24:23–7. [7] Scotton CJ, Wilson JL, Milliken D, Stamp G, Balkwill FR. Epithelial cancer cell migration: a role for chemokine receptors? Cancer Res 2001;61: 4961–5. [8] Scotton CJ, Wilson JL, Scott K, Stamp G, Wilbanks GD, Fricker S, et al. Multiple actions of the chemokine CXCL12 on epithelial tumor cells in human ovarian cancer. Cancer Res 2002;62:5930–8. [9] Zhu ZB, Makhija SK, Lu B, Wang M, Kaliberova L, Liu B, et al. Transcriptional targeting of adenoviral vector through the CXCR4 tumorspecific promoter. Gene Ther 2004;11:645–8. [10] Dmitriev I, Krasnykh V, Miller CR, Wang M, Kashentseva E, Mikheeva G, et al. An adenovirus vector with genetically modified fibers demonstrates expanded tropism via utilization of a coxsackievirus and adenovirus receptor-independent cell entry mechanism. J Virol 1998;72:9706–13. [11] Hemmi S, Geertsen R, Mezzacasa A, Peter I, Dummer R. The presence of human coxsackievirus and adenovirus receptor is associated with efficient adenovirus-mediated transgene expression in human melanoma cell cultures. Hum Gene Ther 1998;9:2363–73. [12] Miller CR, Buchsbaum DJ, Reynolds PN, Douglas JT, Gillespie GY, Mayo MS, et al. Differential susceptibility of primary and established human glioma cells to adenovirus infection: targeting via the epidermal growth factor receptor achieves fiber receptor-independent gene transfer. Cancer Res 1998;58:5738–48. [13] Bauerschmitz GJ, Lam JT, Kanerva A, Suzuki K, Nettelbeck DM, Dmitriev I, et al. Treatment of ovarian cancer with a tropism modified oncolytic adenovirus. Cancer Res 2002;62:1266–70. [14] Suzuki K, Fueyo J, Krasnykh V, Reynolds PN, Curiel DT, Alemany R. A conditionally replicative adenovirus with enhanced infectivity shows improved oncolytic potency. Clin Cancer Res 2001;7:120–6. [15] Kanerva A, Mikheeva GV, Krasnykh V, Coolidge CJ, Lam JT, Mahasreshti PJ, et al. Targeting adenovirus to the serotype 3 receptor increases gene transfer efficiency to ovarian cancer cells. Clin Cancer Res 2002;8: 275–80. [16] Kanerva A, Wang M, Bauerschmitz GJ, Lam JT, Desmond RA, Bhoola SM, et al. Gene transfer to ovarian cancer versus normal tissues with fibermodified adenoviruses. Mol Ther 2002;5:695–704. [17] Wu H, Seki T, Dmitriev I, Uil T, Kashentseva E, Han T, et al. Double modification of adenovirus fiber with RGD and polylysine motifs improves coxsackievirus–adenovirus receptor-independent gene transfer efficiency. Hum Gene Ther 2002;13:1647–53. [18] Glasgow JN, Kremer EJ, Hemminki A, Siegal GP, Douglas JT, Curiel DT. An adenovirus vector with a chimeric fiber derived from canine adenovirus type 2 displays novel tropism. Virology 2004;324:103–16. [19] Short JJ, Pereboev AV, Kawakami Y, Vasu C, Holterman MJ, Curiel DT. Adenovirus serotype 3 utilizes CD80 (B7.1) and CD86 (B7.2) as cellular attachment receptors. Virology 2004;322:349–59. [20] Sirena D, Lilienfeld B, Eisenhut M, Kalin S, Boucke K, Beerli RR, et al. The human membrane cofactor CD46 is a receptor for species B adenovirus serotype 3. J Virol 2004;78:4454–62. [21] Kawakami Y, Li H, Lam JT, Krasnykh V, Curiel DT, Blackwell JL. Substitution of the adenovirus serotype 5 knob with a serotype 3 knob enhances multiple steps in virus replication. Cancer Res 2003;63: 1262–9. [22] Yu DC, Chen Y, Seng M, Dilley J, Henderson DR. The addition of adenovirus type 5 region E3 enables calydon virus 787 to eliminate distant prostate tumor xenografts. Cancer Res 1999;59:4200–3. [23] Wegner SA, Ehrenberg PK, Chang G, Dayhoff DE, Sleeker AL, Michael NL. Genomic organization and functional characterization of the chemokine receptor CXCR4, a major entry co-receptor for human immunodeficiency virus type 1. J Biol Chem 1998;273:4754–60. [24] Takayama K, Reynolds PN, Short JJ, Kawakami Y, Adachi Y, Glasgow JN, et al. A mosaic adenovirus possessing serotype Ad5 and serotype Ad3 knobs exhibits expanded tropism. Virology 2003;309:282–93. [25] Zhu ZB, Makhija SK, Lu B, Wang M, Rivera AA, Kim-Park S, et al. Incorporating the survivin promoter in an infectivity enhanced CRAdanalysis of oncolysis and anti-tumor effects in vitro and in vivo. Int J Oncol 2005;27:237–46.
R.P. Rocconi et al. / Gynecologic Oncology 105 (2007) 113–121 [26] Bauerschmitz GJ, Barker SD, Hemminki A. Adenoviral gene therapy for cancer: from vectors to targeted and replication competent agents (review). Int J Oncol 2002;21:1161–74. [27] Wickham TJ, Roelvink PW, Brough DE, Kovesdi I. Adenovirus targeted to heparan-containing receptors increases its gene delivery efficiency to multiple cell types. Nat Biotechnol 1996;14:1570–3. [28] Krasnykh V, Dmitriev I, Navarro JG, Belousova N, Kashentseva E, Xiang
121
J, et al. Advanced generation adenoviral vectors possess augmented gene transfer efficiency based upon coxsackie adenovirus receptor-independent cellular entry capacity. Cancer Res 2000;60:6784–7. [29] Stoff-Khalili MA, Rivera AA, Glasgow JN, Le LP, Stoff A, Everts M, et al. A human adenoviral vector with a chimeric fiber from canine adenovirus type 1 results in novel expanded tropism for cancer gene therapy. Gene Ther 2005;12:1696–706.
Treatment of ovarian cancer with a novel dual targeted conditionally replicative adenovirus (CRAd) Rodney P. Rocconi a,⁎, Zeng B. Zhu b , Mariam Stoff-Khalili b , Angel A. Rivera b , Baogen Lu b , Minghui Wang b , Ronald D. Alvarez a , David T. Curiel b , Sharmila K. Makhija a b
a Division of Gynecologic Oncology, University of Alabama at Birmingham, Birmingham, Alabama, USA Division of Human Gene Therapy, Departments of Medicine, Surgery, Pathology, and the Gene Therapy Center, University of Alabama at Birmingham, Birmingham, Alabama, USA
Received 15 June 2006 Available online 14 December 2006
Abstract Objectives. Current virotherapy strategies for ovarian cancer have been hampered by limitations in target cell infectivity and nonspecific tissue replication. In an effort to circumvent these limitations, we evaluated various CRAds modified to incorporate novel capsid targeting motifs (RGD and chimeric Ad5/3) with a novel tissue-specific promoter (CXCR4). Methods. Two novel CRAds (Ad5-CXCR4-F5/3 and Ad5-CXCR4-RGD) were constructed via homologous recombination and verified by PCR and DNA sequencing. The infectivity and viral replication rates of these two CRAds were analyzed via quantitative real-time PCR (QRTPCR) in cell line experiments using three ovarian cancer cell lines (SKOV3.ip1, Hey, and OV4) and compared to that achieved with a clinical grade CRAd (Δ24-RGD) to be evaluated in a Phase I trial. Cytocidal effects were determined by crystal violet staining in these same cell lines infected with different concentrations of viral particles per cell (0, 0.1, 1, 10, 100, and 500). Additionally, viral replication was evaluated by QRTPCR in primary ovarian cancer tissue slices from multiple patients with ovarian cancer as well as in primary human normal liver tissue slices in order to establish CRAd selectivity. All experiments incorporated appropriate controls and repeated in triplicate. Results. Compared to RGD-capsid CRAds (Δ24-RGD and CXCR4-RGD), the F5/3-capsid CRAd (CXCR4-F5/3) demonstrated significant improvements in infection rates (p = 0.025, 0.006, and 0.006) in all ovarian cancer cell lines tested (SKOV3.ip1, Hey, and OV4, respectively). In addition to improved transduction of virus into the cells, the TSP CXCR4-based CRAds demonstrated improved viral replication. Specifically, CXCR4-F5/3 further enhanced viral replication 89-fold (p = 0.009, 0.010, 0.003) in the same cancer cell lines. Furthermore, CXCR4-F5/3 showed a 4-log improvement in oncolytic potential over Δ24-RGD. In the ex vivo primary ovarian tissue slices, CXCR4-F5/3 showed a 58-fold improvement in viral replication (p = 0.005) compared to the clinical grade Δ24-RGD. Both CXCR4-F5/3 and CXCR4-RGD demonstrated significant reduction of viral replication in normal liver slices (p = 0.001). Conclusions. These data suggest that a dual targeted approach is feasible for the combined enhancement of infectivity and replication in ovarian cancer with a specificity that was attenuated in normal liver tissues. In fact, CXCR4-F5/3 outperformed our best CRAd agent to date nearly 60-fold in our most stringent ex vivo model of primary ovarian cancer tissue slices and suggests that this novel agent could be useful for the treatment of ovarian cancer. © 2006 Elsevier Inc. All rights reserved. Keywords: Ovarian cancer; Gene therapy; Virotherapy; Adenovirus; CXCR4; Tumor-specific promoter; Capsid modification
Introduction ⁎ Corresponding author. Department of Obstetrics and Gynecology, University of Alabama at Birmingham, 619 20th Street South, Old Hillman Building-Room 538, Birmingham, AL 35249-7333, USA. Fax: +1 205 975 6174. E-mail address: [email protected] (R.P. Rocconi). 0090-8258/$ - see front matter © 2006 Elsevier Inc. All rights reserved. doi:10.1016/j.ygyno.2006.10.057
Greater than 20,000 women are expected to be diagnosed with ovarian cancer during the year 2006 in the United States alone [1]. Although the majority of patients will present with advanced disease, most will respond to cytoreductive surgery and first-line platinum-based combination chemotherapy.
114
R.P. Rocconi et al. / Gynecologic Oncology 105 (2007) 113–121
Despite this response, effective curative therapy has yet to be determined and thus many patients with advanced disease will experience a recurrence and unfortunately succumb to progressive disease [2–4]. In this regard, novel therapeutics, such as virotherapy, are currently being sought for the treatment of advanced ovarian cancer. Specifically, virotherapy utilizing conditionally replicative adenoviruses (CRAd) represents one such advanced modality for the treatment of ovarian cancer. Despite the promise of CRAds, clinical outcomes with virotherapy have been disappointing to date. Two limitations restricting the clinical efficacy include imprecise tumor cell replicative specificity and inefficient tumor cell infectivity. The recognition of these limits has lead to strategies designed to address these barriers. On this basis, advanced generation CRAd agents have been proposed, which seek to achieve improved replicative specificity and/or enhanced infectivity. For improving tumor cell replication specificity, tumorspecific promoters (TSPs) selectively drive adenoviral E1 gene expression in tumor cells and thereby accomplish transcriptional targeting based upon the induced specificity of viral replication [5]. The optimal TSP restricts replication to tumor cells alone, thereby having the highest activity in tumor cells (“tumor on”) and lowest activity in normal cells (“normal off”). Of note, adenoviral vectors possesses a predilection for hepatocytes thereby making hepatotoxicity the dose-limiting toxicity for these vectors. In this regard, the ideal TSP-directed adenovirus would have the highest “tumor on/liver off” ratio and thus possess the optimal therapeutic index. One such promoter, CXCR4 has been recently investigated as a good candidate for cancer gene therapy. Several authors have demonstrated that CXCR4 gene expression has been undetectable in normal ovarian epithelial cells but markedly upregulated in ovarian cancer cell lines and primary ovarian cancers [6–8]. In our previous work, we have reported that the CXCR4 promoter had a “tumor on” and “liver off” phenotype in both in vitro and in vivo experiments in ovarian cancer [9]. In other studies, the therapeutic index of the CXCR4 promoter was evaluated with other promoters commonly evaluated in ovarian cancer (survivin, SLPI, and Cox-2) (Makhija, personal communication). In these studies, the CXCR4 promoter demonstrated a superior therapeutic index with greater activity in ovarian cancer cell lines and primary ovarian tumor, while demonstrating low levels of activity in human liver tissue. Thus, the CXCR4 promoter appears to be an excellent candidate for use in transcriptional targeting for ovarian cancer gene therapy. For enhancing tumor cell infectivity, the targeting adenoviral vector to tumor cells via alternative pathway is of importance. Native Ad5 tropism is mediated by two capsid proteins: the fiber and the penton base. These proteins bind to the primary high affinity cellular receptor, coxsackie–adenovirus receptor (CAR), and to the integrins αvβ3 and αvβ5, respectively. Tumor is resistant to adenovirus infection due to a relative paucity of the primary receptor CAR on tumor cell surface [10–12]. Based on these molecular interactions, a concerted effort has been made to modify Ad5 tropism, resulting in enhanced tumor cell transduction by retargeting
cellular entry through heterologous pathway or CAR-independent pathway. An example of this is the utilization of RGD motif in the fiber knob of the Ad. This capsid modification appears to facilitate Ad binding and entry into tumor cells via integrin receptors that are abundantly expressed on tumor cells [13,14]. Additional capsid modifications have been explored to obtain infectivity enhancement of Ads including AdF5/3, which substituted the Ad5 fiber with the Ad3 fiber [15,16], Ad5-pk7 in which a motif polylysine (pk7) was genetically inserted at C terminus of Ad fiber and bound to heparin sulfatecontaining receptor [17], and Ad5-CK in which canine Ad knob was used to replace Ad5 knob [18]. In this study, we utilized two capsid modifications (RGD and F5/3) that target to integrins and CD80, 86, and 46, respectively. These modifications enhance the viral infectivity via a CAR-independent pathway [19,20]. Our objective was to improve the viral-specific replication in tumor cells and enhance tumor cell infectivity via an optimal CRAd for the treatment of ovarian cancer. We constructed two novel CRAds (Ad5-CXCR4-RGD and Ad5-CXCR4-F5/3) in which the viral replication is under the control of the CXCR4 TSP and the viral infectivity enhanced via the capsid modifications (RGD and F5/3). Both tumor specificity and infectivity of these novel CRAd agents were evaluated and compared with Δ24RGD. Δ24RGD is a first-generation CRAd that utilizes a partial deletion (24 base pair) of the E1 region which allows a relative selectivity by depending upon the target cells' machinery to complete replication. This CRAd has been approved for use in an NIH-funded Phase I trial. Materials and methods Cells and tissues Human ovarian tumor cell lines, SKOV3.ip.1, OV4, and Hey were cultured in medium suggested by ATCC. 911 cells (a kind gift from Dr. Van Der Eb, Leiden University, The Netherlands) were maintained in Dulbecco's modified Eagle medium. Each medium was supplemented with 10% fetal calf serum, penicillin (100 IU/ml), and streptomycin (100 μg/ml). Cells were incubated at 37°C in a 5% CO2 environment under humidified conditions. Following IRB approval, human ovarian specimens were obtained for tissue slices from epithelial ovarian cancer remnants not needed for diagnostic purposes during primary cytoreductive surgery performed at the University of Alabama at Birmingham (UAB). Time from harvest to tissue slicing was kept at an absolute minimum (< 2 h). To generate tissue slices, a coring device (Alabama Research Development, Munford, AL) was used to create an 8-mm diameter core of tissue from the ovarian cancer. This core was cut in consecutive 0.25-mm-thick slices using the Krumdiek tissue slicer (Alabama Research Development, Munford, AL) with the reciprocating blade at 30 rpm. Sequential slices were then cultured in 12-well plates in RPMI medium supplemented with 10% bovine fetal serum, 100 U/ml of penicillin, and 100 μg/ml of streptomycin. Human liver samples were obtained (Department of Surgery, University of Alabama at Birmingham) from three seronegative donor livers prior to transplantation into recipients. All liver samples were flushed with University of Wisconsin (UW) solution (ViaSpan, Barr Laboratories, Inc., Pomona, NY) before harvesting and kept on ice in UW solution until slicing with the Krumdiek tissue slicer. Time from harvest to slicing was kept at an absolute minimum (<2 h). Liver slices were placed into 6-well plates (1 slice per well) containing 2 mL of complete culture media (William's Medium E with 1% antibiotics, 1% L-glutamine, and 10% bovine fetal serum).
R.P. Rocconi et al. / Gynecologic Oncology 105 (2007) 113–121 All cultures were maintained at 37°C in a humidified atmosphere of 95% air and 5% CO2. Tissue slices were incubated for up to 48 h on a plate rocker set at 60 rpm in order to agitate slices to ensure adequate oxygenation and viability.
Recombinant adenovirus The CRAd genomes were constructed via homologous recombination in Escherichia coli (Fig. 1). CRAd-CXCR4-RGD has the following characteristics: (1) The CRAd agent contains the human CXCR4 promoter nucleotide − 191/ +8826 to drive E1 expression. The CXCR4-controlled E1 expression cassette was placed in the original E1 region of the Ad gene. (2) A RGD-4C capsid modification was inserted into the Ad fiber knob region for enhancement of Ad infectivity.32 (3) The E3 gene was retained in the Ad genome for elevating the oncolytic effect of the CRAd agents.33 And (4) a poly-A signal was inserted between the inverted terminal repeat (ITR) and the CXCR4 promoter to stop the nonspecific transcriptional activity of the ITR and to retain the tumor specificity of the CXCR4 promoter. The CRAd-CXCR4-F5/3 vector has the following characteristics: (1) The CRAd agent contains the human CXCR4 promoter nucleotide −191/+88 to drive E1 expression [9]. The CXCR4-controlled E1 expression cassette was placed in the original E1 region of the Ad gene. (2) The capsid modification has modified fiber genes that contain an Ad5 shaft region and Ad3 knob region for enhancement of Ad infectivity [21]. (3) The E3 gene was retained in the Ad genome to preserve the oncolytic effect of the CRAd agent [22]. (4) A poly-A signal was inserted between the inverted terminal repeat (ITR) and the CXCR4
115
promoter to stop the nonspecific transcriptional activity of the ITR and to maintain the tumor specificity of the CXCR4 promoter. DNA fragments containing nucleotides −191/+ 88 were cut with BamHI and HindIII restriction endonucleases from the clone pBSKCAT/CXCR4 3B/4-1 [5′Δ3] [23] and subcloned into the plasmid pBSSK (Stratagene, La Jolla, CA) by use of the same restriction sites. A SV40 PA fragment was cut with XbaI/BamHI from a pGL3B vector (Invitrogen, Carlsbad, CA) and inserted into the pBSSK via the same restriction sites. pBSSK/PA/CXCR4, the generated clone, was used to create shuttle vectors. Next, DNA fragments containing both an SV40 PA and the CXCR4 promoter were cut with NotI/XhoI and subcloned into the E1 gene containing pScsE1 plasmid (a kind gift from Dr. Dirk Nettelbeck, Department of Dermatology, University Medical Center-Erlangen, Erlangen, Germany). Thus, the plasmid pScsE1/PA/CXCR4 was the result. The Ad vector, Pvk500F5/3 was a kind gift from Dr. Takayama and contains both the E3 gene and a capsid modified F5/3 in Ad5 backbone [24]. After cleavage with PmeI, the shuttle vector, pScsE1/PA/CXCR4, was recombined with Pvk500F5/3 to generate a CRAd genome with an F5/3 modified fiber. The resultant plasmids encoding the CXCR4 promoter were linearized with PacI and transfected into 911 cells using Lipofectamine (Qiagen, Valencia, CA). Resultant viruses that were generated were propagated in A549 cells (a lung cell line in which the CXCR4 gene is overexpressed) and then purified by double CsCl density gradient centrifugation, followed by dialysis against phosphate-buffered saline (PBS) with 10% glycerol. The viruses were titrated by plaque assay and the vp number was determined by spectrophotometrically based on the absorbance at a wavelength of 260 nm. Viruses were stored at − 80°C until use.
Fig. 1. Characterization of CRAd. (a) Construction of the CRAd agents. A 279 bp of the CXCR4 promoter was amplified from the clone pBSKCAT/CXCR4 3B/4-1 [5′Δ3] and constructed in a p ScE1 vector. A poly-A signal sequence was inserted between the ITR and CXCR4 promoter to terminate the transcription signal from the ITR. Constructed clones were recombined with pAdback5/3 to generate the CRAd-CXCR4.F5/3 (RGD-4C motif used to generate the CRAd-CXCR4.RGD). (b) The PCR products were run on an agarose gel (1%) that amplified the oligo pair b/c (CXCR4) and oligo pair a/c (PolyA signal + CXCR4) from the template DNAs of CRAdCXCR4.F5/3 and a negative control reAdGL3.
116
R.P. Rocconi et al. / Gynecologic Oncology 105 (2007) 113–121
In the analysis of CRAd agents, replication positive and negative controls were wild-type Ad5 (Adwt) and reAdGL3B-CXCR4, respectively.
Analysis of CRAd construct by PCR and sequencing The structures of the CRAds were verified by PCR and sequence data. The virus [5 × 108 viral particles (vp)] was processed with a blood DNA kit (Qiagen), and 1/50 of the DNA solution was analyzed by PCR for various regions with Tag polymerase (Qiagen). Thermal cycling conditions were as follows: initial denaturing, 5 min at 95°C; 30 cycles of 30 s each at 95°C, 30 s at 55°C, and 1 min at 72°C. The primers used for each region were as follows: (a) PA-F, 5′ GATGAGTTTGGACAAACCAC; (b) CXCR4-F, 5′TACCGACCACCCGCAAACA; (c) CXCR4-R, 5′AACCGCTGGTTCTCCAGATGC. The PCR products were analyzed by automatic sequence analysis performed by using the BigDye Terminator v3.1 Cycle Sequencing Ready Reaction kit and an Applied Biosystems 3100 Genetic Analyzer (Foster City, CA) at the Genomics Core of the Howell and Elizabeth Heflin Center for Human Genetics of the University of Alabama at Birmingham.
Detection of the E1 gene from the CRAd agent by using quantitative real-time PCR The CRAd-CXCR4 was diluted to 103, 104, and 105 viral particles (vp)/μl with distilled water and added to real-time (RT)-PCR reactions as described [26]. RT-PCR was performed, and the E1 gene copy number was determined by using an oligo pair (forward primer—5′AACCAGTTGCCGTGAGAGTTG, reverse primer—5′CTDGTTAAGCAAGTCCTCGATACT) and the probe (5′6FAM-CACAGCTGGCGACGCCCA-TAMRA-3′). The oligos were designed using Primer Express 1.0 (Perkin-Elmer, Foster City, CA) and synthesized by Applied Biosystems. A nonreplicative control, AdGL3BCXCR4, was used in parallel, in which the Ad5 E1 gene was deleted.
Analysis of infectivity of CRAd agents in tumor cell lines SKOV3.ip1, OV4, and Hey ovarian cancer cells at 105 were cultured as discussed previously. Cells were infected in triplicate with CXCR4.RGD, CXCR4. F5/3, Δ24.RGD, Adwt, or reAdGL3B.CXCR4 at 1, 10, 100, and 1000 vp/cell in 2% FBS infection medium. After a 3-h incubation at 37°C in a 5% CO2 environment, infection medium was aspirated and cells were washed three times with PBS so to remove uninternalized viruses. Cells were trypsinized for 15 min, and 1 ml of PBS was added. DNA was extracted from the specimens, via the DNeasy Tissue Kit (Qiagen, Valencia, CA). The Ad5 E1 gene was detected in DNA samples and quantitative RT-PCR was performed as previously described [25]. Ad E1 gene copy numbers were detected and normalized with human β-actin.
Analysis of replication of CRAd agents in tumor cell lines SKOV3.ip1, OV4, and Hey ovarian cancer cells at 105 were cultured as previously discussed with 100 vp/cell of CXCR4.RGD, CXCR4.F5/3, Δ24.RGD, Adwt, or reAdGL3B.CXCR4 in infection medium containing 2% FBS and incubated at 37°C in a 5% CO2 environment. After incubation for 3 h, infection medium was aspirated, and the cells were washed three times with PBS in order to remove uninternalized viruses. Fresh 10% FBS culture medium was then replaced and the cells were returned to incubation at 37°C in a 5% CO2 environment. Media from triplicate wells were collected 1, 3, and 9 days later. DNA was extracted from the specimens via the DNeasy Tissue Kit (Qiagen, Valencia, CA). The Ad5 E4 gene was detected in DNA samples by using an oligo pair (forward primer—5′ GGAGTGCGCCGAGACAAC; reverse primer—5′ACTAGGTCCGGCGTTCCAT; probe ORF6—TGGCATGACACTACGACCAACACGATCT), and quantitative RT-PCR was performed as described above. Ad E4 gene copy numbers were detected and normalized with human β-actin.
Replication of CRAds in human tissue slices Excess ovarian cancer tissues not needed for diagnostic purposes were obtained from three ovarian cancer patients and primary liver tissue was obtained as previously described. The Krumdieck tissue slicer was used to
generate ovarian cancer and liver tissue slices and was cultured at the conditions described previously. Each tissue slice was estimated to contain 1 × 106 cells/ slice based on a 10-cell thickness (∼ 250 μm) and an 8-mm-slice diameter. Slices were infected in triplicate with 500 vp/cell of CXCR4.RGD, CXCR4.F5/3, Δ24. RGD, Adwt, or controls in fresh infection medium. Slices were incubated at 37°C in a 5% CO2 environment on a rocking plate. After 24 h and 72 h of incubation times, respectively, total DNA was extracted from primary human ovarian tumor slices and liver slices via the DNeasy Tissue Kit (Qiagen, Valencia, CA). DNA samples were treated with DNase-free RNase to remove possible RNA contamination and stored at − 80°C until use. Ad E4 gene copy numbers were detected as previously described.
In vitro analysis of cytocidal effects The in vitro cytocidal effects of the CRAds were analyzed by determining the viability of cells with crystal violet staining after infection with respective CRAd. SKOV3.ip1, OV4, and Hey ovarian cancer cells were plated at 5 × 106 cells/well on a 24-well plates and infected with 0, 0.1, 1, 10, 100, and 500 vp/cell with CXCR4.RGD, CXCR4.F5/3, Δ24.RGD, or reAdGL3B. CXCR4. After a 3-h infection, the infection medium was replaced with the corresponding appropriate complete medium. After 10 days of cultivation, the cells were fixed with 10% buffered formalin for 10 min and stained with 1% crystal violet in 70% ethanol for 20 min, followed by washing with tap water and set to air dry.
Results Attributes of the CXCR4-F5/3 CRAd The vectors shown in Fig. 1 were constructed, amplified in 911 cells, and upscaled in A549 cells. The yield was 2 × 1012 vp/ ml with a vp/plaque-forming unit ratio of 62. The vector structure was confirmed by PCR of the viral DNA (Fig. 1) and DNA sequence data from PCR products (not shown). Furthermore, CXCR4-F5/3 CRAd agent possessed the E1 gene. The E1 gene was detected by QRT-PCR, the tested dilutions of 103, 104, and 105 vp, in a dose response manner, but not in the nonreplicating viral vector of AdGL3B.CXCR4 in which the E1 gene was deleted. In our CRAd agent, the native E1 promoter was replaced by the CXCR4 promoter, which was evidenced by the propagation of the vector in the lung tumor cell line A549 in which the CXCR4 gene was overexpressed. The Ad type 5 backbone with the F5/3 modification allows the retargeting of the adenoviral vector towards a CAR-independent pathway for Ad infectivity enhancement [26]. Evidence of CRAd infectivity in tumor cell lines To further assay CRAd-CXCR4.F5/3, 3 ovarian cancer cell lines SKOV3.ip1, OV4, and Hey were plated in 24-well plates and cells were infected with 1, 10, 100, and 1000 vp/cell of CXCR4.RGD, CXCR4.F5/3, Δ24.RGD, or reAdGL3B.CXCR4. After a 3-h infection, the cells were washed three times with PBS to remove noninternalized viruses and provided with fresh substrate. The Ad5 E1 copy numbers should relate to the relative levels of viral CRAd vectors that entered the tumor cells and thus should correlate to the CRAd's ability to infect ovarian cancer cells. As expected, the number of viruses which infected the ovarian cell lines increased as the vp/cell increased (Fig. 2). The CRAd agent CXCR4.F5/3 showed superior results in all three
R.P. Rocconi et al. / Gynecologic Oncology 105 (2007) 113–121
117
with 100 vp/cell of CXCR4.RGD, CXCR4.F5/3, Δ24.RGD, Adwt, or reAdGL3B.CXCR4. After a 3-h infection, the cells were washed with PBS three times and provided fresh medium. The presence of E4 gene was determined by using QRT-PCR from DNA samples extracted after 1, 3, and 9 days postinfection. The Ad5 E4 copy numbers should relate to the levels of released virions from the tumor cells and therefore correlate to the replication rates of the CRAd agents. Data depicted in Fig. 3 demonstrate that CXCR4-F5/3 has superior replicative capabilities in the three ovarian cell lines tested. In Day 9 samples, this translated into 1.7- to 2.7-fold increase compared to CXCR4.RGD and 24.6- to 213-fold replication enhancement over Δ24.RGD. Compared to Δ24.RGD, CXCR4.F5/3
Fig. 2. Transductional activity in ovarian cancer cell lines. Levels depict the amount of viral CRAd vectors that infected the ovarian cancer cells in each of the 3 cell lines. In all cell lines tested, CXCR4-F5/3 demonstrated the highest level of infectivity at all MOIs. (a) SKOV3.ip.1: 1 vp/cell (p = 0.9); 10 vp/cell (p = 0.03); 100 vp/cell (p = 0.05); and 1000 vp/cell (p = 0.025). (b) Hey: 1 vp/cell (p = 0.08); 10 vp/cell (p = 0.03); 100 vp/cell (p =0.07); and 1000 vp/cell (p =0.006). (c) OV4: 1 vp/cell (p = 0.08); 10 vp/cell (p = 0.02); 100 vp/cell (p =0.03); and 1000 vp/cell (p =0.006).
ovarian cell lines tested at nearly all levels of MOI. Compared to RGD-capsid CRAds (Δ24.RGD and CXCR4.RGD), the F5/3capsid CRAd (CXCR4.F5/3) demonstrated significant improvements in infection rates at the highest MOI 1000 vp/cell (p = 0.025, 0.006, and 0.006) in SKOV3, Hey, and OV4 cell lines respectively. Specifically, the infection level of CXCR4.F5/ 3 (1000 vp/cell) resulted in a 2-fold, 6.7-fold, and nearly 3-fold improvement in infectivity compared to Δ24.RGD in SKOV3. ip1, OV4, and Hey cells, respectively. Compared to CXCR4. RDG, the infectivity levels of CXCR4-F5/3 showed a 5.9-fold, 4.2-fold, and 9.5-fold enhancement in the same cell lines. Evidence of CRAd replication in tumor cell lines To assess CRAd replication, SKOV3.ip1, OV4, and Hey cell lines were plated (105) in 24-well plates and cells were infected
Fig. 3. Transcriptional activity in ovarian cancer cell lines. Levels shown represent the amount of viral replication noted at 1, 3, and 9 days postinfection. (a) SKOV3.ip.1; CXCR4-F5/3 improved replication—Day 1 (p = 0.13); Day 3 (p = 0.008); and Day 9 (p = 0.009). (b) Hey; CXCR4-F5/3 improved replication—Day 1 (p = 0.003); Day 3 (p = 0.001); and Day 9 (p = 0.01). (c) OV4; CXCR4-RGD improved replication on Day 1 (p = 0.003) and Day 3 (p = 0.001); CXCR4-F5/3 improved replication on Day 9 (p = 0.01).
118
R.P. Rocconi et al. / Gynecologic Oncology 105 (2007) 113–121
with 500 vp/cell of CXCR4.RGD, CXCR4.F5/3, Δ24.RGD, Adwt, or reAdGL3B.CXCR4. Media were collected after 24 and 72 h, and DNAwas isolated as previously described for detection of E4 copy number via QRT-PCR. The ratios of DNA copy numbers between Day 1 and Day 3 for CXCR4.RGD were 272%, 254%, and 760% for patients 1, 2, and 3, respectively. The ratios for CXCR4.F5/3 were 1602%, 2220%, and 1854% while the ratios for Δ24.RGD were 180%, 85%, and 235% in the same 3 patients. As seen in the ovarian cell lines, CXCR4.F5/3 demonstrated superiority to the other CRAd's replication rates in all 3 patients evaluated. In these primary tumors, improvements in Day 3 replication were seen in comparison to both CXCR4. RGD (2.3- to 7.3-fold) and Δ24.RGD (14- to 84-fold). CXCR4. F5/3 showed a 58-fold improvement in viral replication (p = 0.005) compared to the clinical grade Δ24.RGD. Evidence of CRAd replication in human liver tissue slices Because Ad replication is species specific and because the major Ad toxicity in gene therapy trials is hepatotoxicity, we tested the CXCR4 promoter activity in human liver specimens in organ culture. Human liver slices were infected with CXCR4. RGD, CXCR4.F5/3, and Adwt. Replication was measured by detecting the presence of the E4 gene by using QRT-PCR from DNA samples extracted after 1 and 3 days post-infection. Compared to the positive control (Adwt), the TSP-directed CRAds demonstrated a significant reduction in replication in liver tissue on both Day 1 (p = 0.001) and Day 3 (p < 0.001) (Fig. 5), thus showing that the CXCR4 promoter has a “liver off” phenotype that should result in low toxicity to the human host liver. Evidence of CRAd cytolysis in tumor cell lines All CRAds were tested for its cell-killing potential via cytolysis in the three ovarian cell lines utilized. Oncolysis was evaluated after 10 days of incubation via crystal violet staining Fig. 4. Transcriptional activity in primary ovarian cancer tissue slices. Human ovarian cancer tissue slices were infected with 500 vp/cell of Ad vectors. After a two-day incubation period, DNA was isolated from each slice and the E4 levels determined by quantitative PCR after normalization. All patients showed a significant increase in Day 3 replication from Day1. CXCR4-F5/3 demonstrated superior replication rates in all 3 patients evaluated, with a 58-fold improvement compared to clinical grade Δ24.RGD.
enhanced the viral replication an average 89-fold in the three ovarian cancer lines evaluated (SKOV3.ip1, p = 0.009; Hey, p = 0.010; OV4, p = 0.003). Evidence of CRAd replication in ovarian tumor tissue slices In order to verify the replication of this promoter under nearclinical conditions, in two novel CRAd agents, we determined the replication of the CRAds in three primary ovarian cancer patients with histologically documented papillary serous adenocarcinomas (Fig. 4). The samples were handled and sliced as described in the Materials and methods section and infected
Fig. 5. Transcriptional activity in primary normal liver tissue slices. Human liver tissue slices were infected with 500 vp/cell of Ad vectors. DNA was isolated on Day 1 and Day 3 and E4 levels determined by QRT-PCR after normalization. Compared to the positive control (Adwt), the TSP-directed CRAds demonstrated a significant reduction in replication in liver tissue on both Day 1 (p = 0.001) and Day 3 (p < 0.001).
R.P. Rocconi et al. / Gynecologic Oncology 105 (2007) 113–121
(Fig. 6). While reAdGL3B.CXCR4 (replication-incompetent) possessed no cytotoxic effect as expected, the CXCR4-based CRAds demonstrated potent oncolytic potential in all 3 cell lines. In fact, CXCR4.RGD showed a 2- to 3-log improvement in oncolysis compared to Δ24.RGD, with some oncolysis at levels as low as 1 vp/cell. Likewise, CXCR4.F5/3 resulted in a 4- to 5-
Fig. 6. Cytotoxic efficiency of Ad vectors. 5 × 104 cells from each ovarian cancer cell line (SKOV3.ip1, Hey, and OV4) were plated on 24-well plates, and infected with Ad vectors at different MOIs (0, 0.1, 1, 10, 100, and 500). Cells were stained with crystal violet after a 10-day incubation. In all 3 cell lines evaluated, CXCR4-F5/3 demonstrated at least a 2-log and 4-log improvement in cell death over CXCR4-RGD and Δ24.RGD, respectively. (a) SKOV3.ip.1, (b) Hey, and (c) OV4.
119
log improvement compared to Δ24.RGD and a 2-log improvement compared to CXCR4.RGD. This cytolytic effect of CXCR4.F5/3 was seen at 0.1 vp/cell in all 3 ovarian cell lines. Discussion Effective virotherapy for the treatment of cancer is dependent on their capability to infect and replicate in target cells. Thereby, the oncolytic potential is directly relative to the ability of viral replication. Most adenoviruses utilized are serotype 5 (Ad5) which binds to CAR. Unfortunately, CAR expression varies widely on many tumor cells including ovarian cancer cells. This could potentially hinder viral infection, replication, and subsequent oncolysis [26]. Therefore, methods to overcome these limitations involve transductional and transcriptional targeting to improve the infectivity and selective replication of CRAd agents. Fiber capsid modifications, such as F5/3, enhance viral infectivity via utilizing CAR-independent pathways for transduction into the cell. Once infected, tumor-specific promoters (TSP) place the adenovirus under the control of heterologous control regions that limit the replication to specific tissues or tumors. In this study, we have demonstrated that capsid or fiber modification enhances infectivity and allows improved antitumor potential of replicative CRAds in ovarian cancer. Furthermore, by incorporating an ovarian cancer relevant TSP in these infectivity-enhanced capsid/fiber modified CRAds, we showed ovarian cancer cell-specific adenoviral replication and oncolysis. These principals were tested in comparison to a clinical grade CRAd in multiple ovarian cancer cell lines and in primary ovarian tissue slices, thus giving us results in stringent models in relation a CRAd set for human clinical trials. Of interest is the significant “liver off” profile demonstrated by the CXCR4 TSP, which should reduce the dose-limiting toxicity of hepatotoxicity of adenoviral trials to date. Additionally, CXCR4.F5/3 attains improved oncolysis by both enhanced infectivity and selective replication in vitro. Importantly, we confirmed this combined infectivity and replication in a stringent human model system utilizing primary human ovarian cancer tissue slices and normal liver slices. Therefore, our experiments achieved significant improvements in regards toward overcoming the limitations of poor infectivity and poor specificity previously seen in CRAd agents. Other limitations previously demonstrated and yet to be circumvented encompasses the human immune response, which limits the effectiveness of multiple treatments with adenoviral vectors. This investigation evaluated the efficacy of an infectivity enhanced capsid modified CRAd utilizing an ovarian cancer cell TSP, CXCR4. As previously mentioned, this TSP has been shown to exist preferentially in ovarian tumor cells while undetectable in normal ovarian tissue. Additionally, ovarian cancer proliferation, migration, and metastasis have been linked to the CXCR4 binding of its ligand, SDF-1, thereby making it a good candidate for a TSP. We demonstrated that the novel CRAd agents containing the CXCR4 TSP improved transcriptional targeting over a clinical grade CRAd, Δ24.RGD. Future investigations could be directed towards correlating the
120
R.P. Rocconi et al. / Gynecologic Oncology 105 (2007) 113–121
level of CXCR4 expression in tumor to the response to CXCR4 CRAds. The paucity of native adenoviral receptors (CAR) in ovarian tumor necessitates the evaluation of CAR-independent pathways to enhance infectivity. This enhanced infectivity was achieved by replacing the Ad5 fiber knob domain with the knob domain from human adenovirus serotype 3 (Ad5/3). Other capsid modification strategies have been investigated in order to circumvent the poor infectivity of adenoviruses in tumor cells by utilizing CAR-independent receptors. These studies included the incorporation of short heterologous peptides like RGD, or polylysine into the fiber knob domain [27], or knob switching strategies [29] such as xenotype knob switching [18,28,29]. Although both RGD and F5/3 serve as potential capsid modifications for the enhanced infectivity of adenoviruses, we confirmed the previously described superiority of F5/3 compared to RGD [16]. In conclusion, we present a CRAd addressing the obstacles seen in adenoviral cancer gene therapy, by infectivity enhancement and specific replication with subsequent cell killing in ovarian cancer. Specifically, the TSP CXCR4 emerges as an especially promising transcriptional targeting approach for ovarian cancer. Our study using CXCR4.F5/3 demonstrated improved infectivity, selective replication, and enhanced oncolytic potential over all other CRAds tested, including the clinical grade Δ24.RGD. Considering the significant improvements of a “tumor on/liver off” CRAd evaluated in stringent well-established ex vivo models (primary tissue slices), this CRAd potentially could overcome the limitations of poor infectivity and poor specificity previously seen in other novel CRAd agents in cancer therapy. Although, further evaluation is needed, the promoter CXCR4 and our construct CXCR4.F5/3 could represent a novel gene therapy agent for the treatment of ovarian cancer. Acknowledgment This work was performed with NIH grant #5R01CA09054703 (8/9/01–1/31/08). References [1] Jemal A, Murray T, Ward E, Samuels A, Tiwari RC, Ghafoor A, et al. Cancer statistics. CA Cancer J Clin 2005;55:10–30. [2] Alberts DS, Green S, Hannigan EV, O'Toole R, Stock-Novack D, Anderson P, et al. Improved therapeutic index of carboplatin plus cyclophosphamide versus cisplatin plus cyclophosphamide: final report by the Southwest Oncology Group of a phase III randomized trial in stages III and IV ovarian cancer. J Clin Oncol 1992;10:706–17. [3] McGuire WP, Hoskins WJ, Brady MF, Kucera PR, Partridge EE, Look KY, et al. Cyclophosphamide and cisplatin compared with paclitaxel and cisplatin in patients with stage III and stage IV ovarian cancer. N Engl J Med 1996;334:1–6. [4] Ozols RF, Bundy BN, Greer BE, Fowler JM, Clarke-Pearson D, Burger RA, et al. Phase III trial of carboplatin and paclitaxel compared with cisplatin and paclitaxel in patients with optimally resected stage III ovarian cancer: a Gynecologic Oncology Group study. J Clin Oncol 2003;21: 3194–200. [5] Haviv YS, Curiel DT. Engineering regulatory elements for conditionallyreplicative adeno-viruses. Curr Gene Ther 2003;3:357–85.
[6] Li F, Zhu HS, Han ZQ, Chen G, Gao QL, Jia P, et al. Effects of chemokine receptor and its ligand on migration of ovarian cancer cells. Ai Zheng 2005;24:23–7. [7] Scotton CJ, Wilson JL, Milliken D, Stamp G, Balkwill FR. Epithelial cancer cell migration: a role for chemokine receptors? Cancer Res 2001;61: 4961–5. [8] Scotton CJ, Wilson JL, Scott K, Stamp G, Wilbanks GD, Fricker S, et al. Multiple actions of the chemokine CXCL12 on epithelial tumor cells in human ovarian cancer. Cancer Res 2002;62:5930–8. [9] Zhu ZB, Makhija SK, Lu B, Wang M, Kaliberova L, Liu B, et al. Transcriptional targeting of adenoviral vector through the CXCR4 tumorspecific promoter. Gene Ther 2004;11:645–8. [10] Dmitriev I, Krasnykh V, Miller CR, Wang M, Kashentseva E, Mikheeva G, et al. An adenovirus vector with genetically modified fibers demonstrates expanded tropism via utilization of a coxsackievirus and adenovirus receptor-independent cell entry mechanism. J Virol 1998;72:9706–13. [11] Hemmi S, Geertsen R, Mezzacasa A, Peter I, Dummer R. The presence of human coxsackievirus and adenovirus receptor is associated with efficient adenovirus-mediated transgene expression in human melanoma cell cultures. Hum Gene Ther 1998;9:2363–73. [12] Miller CR, Buchsbaum DJ, Reynolds PN, Douglas JT, Gillespie GY, Mayo MS, et al. Differential susceptibility of primary and established human glioma cells to adenovirus infection: targeting via the epidermal growth factor receptor achieves fiber receptor-independent gene transfer. Cancer Res 1998;58:5738–48. [13] Bauerschmitz GJ, Lam JT, Kanerva A, Suzuki K, Nettelbeck DM, Dmitriev I, et al. Treatment of ovarian cancer with a tropism modified oncolytic adenovirus. Cancer Res 2002;62:1266–70. [14] Suzuki K, Fueyo J, Krasnykh V, Reynolds PN, Curiel DT, Alemany R. A conditionally replicative adenovirus with enhanced infectivity shows improved oncolytic potency. Clin Cancer Res 2001;7:120–6. [15] Kanerva A, Mikheeva GV, Krasnykh V, Coolidge CJ, Lam JT, Mahasreshti PJ, et al. Targeting adenovirus to the serotype 3 receptor increases gene transfer efficiency to ovarian cancer cells. Clin Cancer Res 2002;8: 275–80. [16] Kanerva A, Wang M, Bauerschmitz GJ, Lam JT, Desmond RA, Bhoola SM, et al. Gene transfer to ovarian cancer versus normal tissues with fibermodified adenoviruses. Mol Ther 2002;5:695–704. [17] Wu H, Seki T, Dmitriev I, Uil T, Kashentseva E, Han T, et al. Double modification of adenovirus fiber with RGD and polylysine motifs improves coxsackievirus–adenovirus receptor-independent gene transfer efficiency. Hum Gene Ther 2002;13:1647–53. [18] Glasgow JN, Kremer EJ, Hemminki A, Siegal GP, Douglas JT, Curiel DT. An adenovirus vector with a chimeric fiber derived from canine adenovirus type 2 displays novel tropism. Virology 2004;324:103–16. [19] Short JJ, Pereboev AV, Kawakami Y, Vasu C, Holterman MJ, Curiel DT. Adenovirus serotype 3 utilizes CD80 (B7.1) and CD86 (B7.2) as cellular attachment receptors. Virology 2004;322:349–59. [20] Sirena D, Lilienfeld B, Eisenhut M, Kalin S, Boucke K, Beerli RR, et al. The human membrane cofactor CD46 is a receptor for species B adenovirus serotype 3. J Virol 2004;78:4454–62. [21] Kawakami Y, Li H, Lam JT, Krasnykh V, Curiel DT, Blackwell JL. Substitution of the adenovirus serotype 5 knob with a serotype 3 knob enhances multiple steps in virus replication. Cancer Res 2003;63: 1262–9. [22] Yu DC, Chen Y, Seng M, Dilley J, Henderson DR. The addition of adenovirus type 5 region E3 enables calydon virus 787 to eliminate distant prostate tumor xenografts. Cancer Res 1999;59:4200–3. [23] Wegner SA, Ehrenberg PK, Chang G, Dayhoff DE, Sleeker AL, Michael NL. Genomic organization and functional characterization of the chemokine receptor CXCR4, a major entry co-receptor for human immunodeficiency virus type 1. J Biol Chem 1998;273:4754–60. [24] Takayama K, Reynolds PN, Short JJ, Kawakami Y, Adachi Y, Glasgow JN, et al. A mosaic adenovirus possessing serotype Ad5 and serotype Ad3 knobs exhibits expanded tropism. Virology 2003;309:282–93. [25] Zhu ZB, Makhija SK, Lu B, Wang M, Rivera AA, Kim-Park S, et al. Incorporating the survivin promoter in an infectivity enhanced CRAdanalysis of oncolysis and anti-tumor effects in vitro and in vivo. Int J Oncol 2005;27:237–46.
R.P. Rocconi et al. / Gynecologic Oncology 105 (2007) 113–121 [26] Bauerschmitz GJ, Barker SD, Hemminki A. Adenoviral gene therapy for cancer: from vectors to targeted and replication competent agents (review). Int J Oncol 2002;21:1161–74. [27] Wickham TJ, Roelvink PW, Brough DE, Kovesdi I. Adenovirus targeted to heparan-containing receptors increases its gene delivery efficiency to multiple cell types. Nat Biotechnol 1996;14:1570–3. [28] Krasnykh V, Dmitriev I, Navarro JG, Belousova N, Kashentseva E, Xiang
121
J, et al. Advanced generation adenoviral vectors possess augmented gene transfer efficiency based upon coxsackie adenovirus receptor-independent cellular entry capacity. Cancer Res 2000;60:6784–7. [29] Stoff-Khalili MA, Rivera AA, Glasgow JN, Le LP, Stoff A, Everts M, et al. A human adenoviral vector with a chimeric fiber from canine adenovirus type 1 results in novel expanded tropism for cancer gene therapy. Gene Ther 2005;12:1696–706.