This study was conducted by a 3 + 3 dose-escalation strategy at a single institution following IRB, IBC, RAC, and FDA approval. Participants were enrolled from July 2007 to April 2009. Eligible patients originally included histiologically documented persistent or recurrent epithelial ovarian or primary peritoneal adenocarcinoma and eventually was expanded to include fallopian tube and endometrial carcinoma. All patients were required to have previous treatment with conventional surgery and chemotherapy and have evidence of intra-abdominal disease. Patients were required to have adequate organ laboratory function defined as WBC > 3000 uL, granulocyte count > 1500 uL, platelets > 100,000, creatinine clearance > 80mg/dL, creatinine < 2.0, AST or ALT < 2.5× the upper limit of the normal range, bilirubin < 2.0, and PT/PTT/INR < 1.5× the upper limit of the normal range. Patients were required to have an ejection fraction > 55% on echocardiogram and an O2 saturation > 92%. Patients were required to be ≥ 19 years of age, have a GOG performance status of 0-2, have a life expectancy > 3 months, and signed an informed consent document. Patients with low malignant potential epithelial, stromal, or germ cell ovarian tumors were excluded. Patients with active heart disease, pulmonary disease, or coagulation disorders were excluded.
The Ad5-Δ24 mutant adenovirus containing the 24 nucleotide deletion from Ad5 bp 923 to 946 was originally provided by Dr. Juan Fueyo (MD Anderson Cancer Center, Houston, TX). An E1
fragment containing the 24bp deletion from this plasmid was cloned via homologous recombination into a ClaI digested plasmid pVK503 containing the RGD fiber as previously described (14
). Following PacI digestion the resulting genome was released from the plasmid backbone, transfected into A549 cells and rescued. RGD presence and Δ24 absence were verified via PCR.
Ad5-Δ24-RGD was manufactured with the support of the NCI RAID program at the Cell and Gene Therapy Center at Baylor College of Medicine and at the Biopharmaceutical Development Program/SAIC at NCI-Frederick. All viral doses were administered in 250 ml of 0.9% sodium chloride and kept refrigerated until administration.
General treatment plan and Ad5-Δ24-RGD dose cohorts
Pretreatment evaluation consisted of: history and physical, toxicity grading, performance status assignment, CBC, chemistry panel, liver profile, coagulation profile, CA-125, determination of ejection fraction by echocardiogram, O2 saturation, and CT of the abdomen and pelvis. Patients completing pretreatment evaluation and meeting all eligibility criteria were enrolled and had an intraperitoneal (IP) Quinton Curl, 22.4 inch, double cuffed, Tenchkhoff type catheters (Tyco Healthcare, Princeton, NJ) placed by interventional radiology at least one week prior to utilization.
Patients were then enrolled in successive escalating dose cohorts such that cohort 1 received 1×109 vp/d and each successive cohort dose increased by ½ log vp/d. The 7th and final cohort received 1×1012 vp/d. Assigned doses were instilled via the IP catheter daily for three consecutive days in an inpatient setting. On the 4th day, the patient was discharged. Dose escalation occurred 4 weeks after the final patient in the previous cohort was treated. No individual patient dose escalation was performed.
On days 0-3, 7, 14, and 28 patients were evaluated via history and physical, performance status assignment, toxicity grading, CBC, chemistry profile. Peritoneal aspirates for biologic ancillary studies and urine, saliva and serum specimens for viral shedding studies were obtained immediately preceding Ad5-Δ24-RGD administration on day 0, 3, 7, 14, and 28. Serum CA-125 and a CT of the abdomen and pelvis were repeated on day 28. All samples were processed to assure anonymity; individuals performing the biologic studies were blinded to patient identity.
Toxicity grading was performed utilizing NCI Common Toxicity Criteria v3.0. MTD was defined as the dose exceeded by the dose at which at least 2 patients experience dose-limiting toxicity (DLT). DLT was defined as any vector related grade 3 non-hematologic toxicity, not including nausea, vomiting, or fatigue. Dose limiting hematologic toxicities were defined as any admission for neutropenic fever, ANC < 500 for > 5 days, or platelet count < 20,000. Any patient experiencing vector related grade 3/4 toxicity prior to completion of scheduled treatment had subsequent days of treatment held until resolution of toxicity. If no resolution occurred within 72 hours or if a second dose interruption occurred, patients received no further study drug.
Evaluation of clinical efficacy
RECIST criteria version 1.0 was utilized to define patients' best radiographic response to treatment. Measurable disease was defined as at least one lesion that could be accurately measured in one dimension and greater than 1 cm. Up to 5 lesions per organ or 10 lesions total were identified as target lesions. Complete response (CR) required disappearance of all target lesions and normalization of CA-125. Partial response (PR) was defined as > 30% decrease in the sum of target lesions' recorded dimensions. Progressive disease (PD) was defined as > 20% increase in sum of target lesions recorded dimensions. Stable disease was a condition that did not qualify for PR or PD.
Evaluation of viral infection, CRAd replication, generation of wild type virus in peritoneal fluid cells
Genomic DNA of tumor cells in peritoneal fluid was isolated using a QIAamp DNA Mini Kit (QIAGEN, Valencia, CA) according to the manufacturer's protocols. Real-Time quantitatitve PCR (RTqPCR) was utilized to evaluate gene transfer and CRAd replication. RGD copies in genomic DNA were determined by amplification of the RGD gene with forward primer CACACTAAACGGTACACAGGAAACA, reverse primer ATGCAGATGGGCAGAAACAGT and probe: 6-FAM-AGACACAACTTGTGACTGCCGCGG-BHQ-1. Resultant RGD copies were normalized to a human cellular house-keeping gene (human β-Actin) which was amplified with forward primer CCAGCAGATGTGGATCAGCA, reverse primer CTAGAAGCATTT-GCGGTGGAC and probe 6-HEX-AGGAGTATGACGAG-TCCGGCCCCTC-BHQ-1.
To determine whether potential contaminating wild type adenovirus were replicating, the wild type E1 (WT-E1) gene was amplified from cells in the ascites fluid with forward primer TGCCAAACCTTGTACCGGA, reverse primer CGTCGTCACTGGGGTGGAAA and probe 6-FAM-ATCGATCTTACCTGCCACGAGGCTGG–BHQ. Resulted WT-E1 copies were normalized to house-keeping genes to allow for comparison between patients and at different time points and varying amount of cellular material.
All primers and probes were designed by the Primer Express 1.5 software and synthesized by Sigma-Aldrich (St. Louis, MO). FastStart TaqMan Probe Master (Roche Applied Science, Indianapolis, IN) was used for duplexing the PCR reaction on a LightCycler™ 480 (Roche Applied Science, Indianapolis, IN). Thermal cycling conditions began with 8 minutes at 95 °C followed by 45 cycles of 10 seconds at 95 °C and 40 seconds at 60 °C. Data were analyzed with LightCycler 480 1.5.0 SP1 software. WT and RGD specific primers were confirmed to only amplify the virus being tested (data not shown).
After being transported to the laboratory, DMSO (Dimethyl Sulfoxide,(Sigma Aldrich, St. Louis, MO) was added to ascites specimens containing tumor cells to reach 5% and immediately frozen at −80°C until assays and analysis commenced. Upon thawing, cells were suspended in cell culture media (Sigma Aldrich, St. Louis, MO). Cell suspensions of ascites were used to prepare at least two cytospin slides per sample using the ThermoShandon 3 Cytospin (Dreieich, Germany) at 2000rpm for 10 minutes at room temperature. After the cytospin, slides were removed and fixed in 70% ethanol overnight (Sigma Aldrich, St. Louis, MO). Slides were incubated with rabbit polyclonal anti-hexon antibody (Abcam, Cambridge, MA) at a dilution of 1:2000 and CC-49 anti-Tag72 antibody (provided by M. B. Khazaeli, PhD, University of Alabama at Birmingham) at a dilution of 5ug/ul for one hour. CC-49 anti-Tag72 antibody is an antibody to a tumor associated glycoprotein known to be expressed by ovarian cancer cells (15
). Anti-hexon is an anti-adenovirus antibody. Negative controls were performed by omitting primary antibodies. Ascites cells from patients untreated with adenovirus were used to confirm that in the absence of adenovirus, staining with anti-hexon is negative. Ascites cells from patients not in this trial were treated ex vivo with adenovirus and stained with anti-hexon antibody to confirm that anti-hexon will detect present adenovirus. (Supplemental data, ). Secondary antibodies Alexafluor 488 (Invitrogen) and Alexafluor 594 (Invitrogen) were incubated at 1:100 for one hour in phosphate buffered saline. Slides were then mounted with 0.25% Propyl Gallate in a 9:1 (v/v) glycerol:PBS solution to prevent photobleaching (Sigma Aldrich, St. Louis, MO). Fluorescence microscopy was performed with an inverted IX-70 microscope (Olympus, Melville, NY) equipped with a Magnifire digital CCD camera (Optronics, Goleta, CA) or a DP71 digital camera (Olympus). All images were at 40× magnification. The images of fluorescent signals for tumor cells and adenovirus were merged using Adobe Photoshop CS (San Jose, CA).
Quantification of Ad5-Δ24-RGD in ascites/peritoneal lavage samples (A) and immunohistochemical evidence of colocalization of Ad5-Δ24-RGD and ovarian cancer cells (B)
Evaluation of viral shedding
Viral DNA from urine specimens was isolated with a QIAamp Viral RNA Mini Kit (QIAGEN, Valencia, CA) after being concentrated with Millipore* Amicon* Ultra-4 and Amicon Ultra-15 Centrifugal Filter Units (Billerica, MA). Viral DNA from saliva specimens was isolated using a QIAampMinElute Virus Spin Kit (QIAGEN, Valencia, CA) following the manufacturer's instructions. Viral DNA from sera specimens was isolated with the DNeasy Tissue Kit's blood protocol (QIAGEN, Valencia, CA). One microliter from resultant samples were used as a template for RTqPCR since many of these samples lacked genomic DNA for normalization of results. We amplified a fragment of RGD with forward primer CACACTAAACGGTACACAGGAAACA, reverse primer ATGCAGATGGGCAGAAACAGT and probe: 6-FAM-AGACACAACTTGTGACTGCCGCGG-BHQ-1 (Sigma Aldrich, St. Louis, MO). RTqPCR conditions were similar to those previously described to evaluate viral infection, CRAd replication, and generation of wild type virus.
Evaluation of an anti-adenovirus neutralizing antibody (Nabs) response
For evaluation of induced anti-adenovirus Nabs response in serum and ascites specimens after treatment, a nonreplicative luciferase expressing virus, Ad5-RGD-Luc1, was neutralized by either serum or ascites prior to infection of SKOV3.ip1 cells. SKOV3.ip1 cells are a cell line derived from the implantation of SKOV3 cells (ATCC, Manassas, VA) in nude mice. These cells were last tested and authenticated for epithelial staining via pooled AE1/AE3 antibody in December, 2009. These antibodies stain normal and neoplastic cells of epithelial origin.
Following neutralization in respective samples, Ad5-RGD-Luc1 transduction efficacy was determined by a luciferase assay. Triplicates of SKOV3.ip1 cells were plated into 96-well plates (10,000/well) and allowed to grow overnight before infection. A 1:2 dilution of serum or ascites of each day point specimen was prepared in Opti-MEM (Media Preparation Shared Facility, UAB), in a normalized volume. Non-replicative Ad5-RGD-Luc1 at 100 PFU/cell was mixed with each dilution for 30 minutes at room temperature before adding to appropriate wells. This infection was allowed to proceed for 48 hours. A luciferase assay was carried out using a luciferase assay system (Promega, Madison, WI) on an Orion microplate luminometer (Berthold, Pforzheim, Germany) reading Culturplate-96 (Research Parkway, Meriden, CT) according to manufacturer's protocols.
Demographic and baseline characteristics of the treated patients are summarized descriptively. The incidence of adverse events and laboratory tests are also briefly summarized respectively. For the analysis of biological effects, a repeated measures analysis of variance was used to compare baseline values to the other study day values. The raw data were transformed into logarithms to the base 10 (log10) to meet the normality assumption before statistical testing.