Purpose: We decided to construct a novel oncolytic adenovirus whose replication was driven by the CDC25B promoter for its use in preclinical models of pancreatic cancer.
Experimental Design: We placed the essential E1A gene under control of the CDC25B promoter. Based on preliminary data, we pseudotyped the adenovirus with a chimeric fiber of serotypes 5/3. We investigated the in vitro lytic effect and the in vivo therapeutic efficacy in combination with gemcitabine on human pancreatic tumor xenografts orthotopically growing in nude mice and in tumors growing in Syrian hamsters. We also assessed biochemical markers of hepatic toxicity and CA19.9 levels.
Results: AV25CDC exhibited a strong in vitro lytic effect on pancreatic cancer cells. In vivo administration of AV25CDC combined with gemcitabine in mice harboring subcutaneously growing SW1990 pancreatic tumors almost abrogated tumor growth. Nude mice harboring 15-day-old orthotopic tumors, treated intratumorally or systemically with AV25CDC combined with gemcitabine, exhibited 70% to 80% reduction in tumor size compared with control mice that lasted for at least 60 days. Chemovirotherapy treatment induced a return to normal levels of biochemical parameters of hepatic toxicity; these mice exhibited more than 90% reduction in CA19.9 serum levels compared with control. Chemovirotherapy efficacy was confirmed in mice harboring Mia PaCa-2 tumors and in Syrian hamster harboring HaP-T1 tumors. We observed that viral treatment disrupted tumor architecture and induced an increase in MMP-9 activity that might facilitate gemcitabine penetrability.
Conclusion: These data demonstrate that AV25CDC is an effective oncolytic agent candidate for pancreatic cancer chemovirotherapy combination. Clin Cancer Res; 21(7); 1665–74. ©2015 AACR.
Pancreatic cancer exhibits a high mortality rate, and currently no effective therapy is available. We have developed an oncolytic adenovirus named AV25CDC in which E1A expression is driven by the human CDC25B gene promoter. AV25CDC was used solely or combined with gemcitabine to treat nude mice harboring orthotopically established SW1990 human pancreatic cancer xenografts. Mice treated intratumorally or systemically with AV25CDC + gemcitabine exhibited up to 80% reduction in tumor size compared with control mice, and showed a return to normal levels of biochemical parameters of hepatic toxicity; these mice also exhibited more than 90% reduction in CA19.9 serum levels compared with control mice. Further studies confirmed these findings in additional preclinical models, such as Mia PaCa-2 xenografts in nude mice and HaP-T1 tumors in Syrian hamsters. These findings provide a proof of concept for the combined use of AV25CDC and gemcitabine as a powerful therapeutic modality in pancreatic cancer.
Worldwide, more than 200,000 people die annually of pancreatic cancer, making it the fourth leading cause of cancer-related death in the United States (1, 2). Despite the increased understanding of the molecular biology of the disease, the use of the purine analog gemcitabine remains the standard of care (3), with a median survival time that did not exceed 6.5 months (4). In fact, gemcitabine and erlotinib remained as the only two drugs approved for use in the advanced disease with modest benefit.
One rising area as a potential new therapeutic approach in cancer are oncolytic viruses (OV). These OVs are engineered or natural viruses with selective toxicity for malignant cells. Among OVs, conditionally replicative adenoviruses (CRAd) were developed based on the fact that the transcriptional activity of the E1A gene, that is essential for adenoviral replication is driven by tumor-specific promoters (5). Several CRAds have been evaluated in preclinical trials for their potential therapeutic effect on pancreatic cancer such as the cyclooxygenase 2 promoter-based CRAd (6), and especially the hTERT promoter-based CRAd (7). Several preclinical studies have demonstrated improved efficacy when hTERT- or COX-2-promoter–based oncolytic adenovirus (OAV), were combined with gemcitabine (6, 8); the replication-selective dl922-947 adenovirus, defective in pRb binding, improved mice survival when combined with 5-FU and gemcitabine (9). At the clinical level it was of note that intratumorally injected ONYX-15 OV combined with gemcitabine was well-tolerated in a phase I/II clinical trial of patients with pancreatic cancer (10). The HSV-1 OV, HF10, reached the clinic and was well tolerated in phase I trials (11) while a clinical trial with the HSV-1-based OV, BioVex GMCSF, is running (12). Thus, despite the severity of the disease and the certain success of OVs in preclinical trials, only very few of them reached the clinics.
CDC25B phosphatase has been found to be overexpressed in more than 70% of human pancreatic cancer samples, often associated with high-grade tumors and poor prognosis (13); interestingly, most pancreatic and gastric cancer overexpresses only this phosphatase isoform (13). Further studies demonstrated that CDC25B levels were augmented 2.2 and more than four times in metastatic pancreatic cancer compared with primary tumors and normal pancreatic tissue, respectively (14). Also, both malignant cells and cancer-associated fibroblasts exhibited strongly positive staining for CDC25B (14). We constructed a novel CRAd named AV25CDC, in which the adenoviral E1A gene was placed under the control of a 0.45-kb fragment of the CDC25B promoter. The combination of AV25CDC and gemcitabine exhibited the largest therapeutic effect on orthotopically implanted human xenografts tumors in nude mice and on tumors in Syrian hamsters; AV25CDC therapeutic effect was accompanied by a strong decrease in biochemical markers of hepatic toxicity and in the tumor biomarker CA19.9. Further evidence demonstrated that AV25CDC treatment induced the disruption of tumor architecture that might have helped gemcitabine to penetrate deeper into the tumor mass.
Materials and Methods
Pancreatic cancer cells (BxPC-3, MIA PaCa-2, Panc-1, and SW1990), colorectal cancer cells (HT 29 and LoVo), gastric cancer cells (MKN-45), and normal cells (CCD1140sk, HFL1, WI38) were purchased from the American Tissue Culture Collection (ATCC) between 2006 and 2008. All the cells were authenticated by ATCC. Upon arrival, cells were thawed, expanded once (P1) to obtain 106 cells, and stored in liquid nitrogen in five vials. When necessary, each one of the vials at P1 was thawed and expanded to obtain 10 vials (P2). Only P2 cells were used for the in vitro and in vivo experiments. The HaP-T1 hamster cell line was kindly provided by Dr. Ruben Hernandez Alcoceba (University of Navarra, Pamplona, Spain), and HaCaT cells were kindly provided by Fernando Larcher (Universidad Carlos III, Madrid, Spain). All the cell lines were grown in the recommended medium supplemented with 10% fetal bovine serum (Natocor), 2 mmol/L glutamine, 100 U/mL of penicillin, and 100 μg/mL of streptomycin and maintained in a 37°C atmosphere containing 5% CO2. All the cells were routinely tested for mycoplasma contamination by PCR, using the following primers: forward, 5′-ACACCATGGGAGYTGGTAAT-3′ and reverse, 5′-CTTCWTCGACTTYCAGACCCAAGGCAT-3′. Our slightly modified protocol (Tang and colleagues; ref. 15) is useful to detect contamination with the following strains: M. arginine, M. orale, M. hyorhinis, M. fermentans, M. hominis, M. salivarium, M. argininin, and M. laidlawii.
Assessment of CDC25B, CONEXIN 26, and E1A mRNA expression
Total RNA was extracted from each cell line to assess for CDC25B, CONEXIN 26, and E1A levels using quantitative real-time PCR. Detailed information is described in Supplementary Materials and Methods.
Assessment of viral replication
DNA was extracted from cells to assess for viral E4 levels by quantitative real-time PCR as readout of viral particles. Total E4 copies per sample were normalized with the amount of DNA present in each sample and reported as E4 copies/ng of DNA (16). Detailed information is described in Supplementary Materials and Methods.
Luciferase activity following cells transduction by the nonreplicative viruses was measured using a Genios luminometer (TECAN) and normalized by protein concentration in the cell lysate (Bio-Rad; ref. 16). Detailed information is described in Supplementary Materials and Methods.
In vitro assays combining AV25CDC with gemcitabine
Cells seeded in 96-well plates (2 × 103 cells per well) were infected with AV25CDC for 24 hours followed by the addition of fresh medium containing gemcitabine (2′-deoxy-2′,2′-difluorocytidine monohydrochloride, Sandoz S.A). Five days later, cell viability was established with MTS as described (16). All assays were carried out in six different replicates.
Cell viability assays
Cells were plated onto 96-well plates at a density of 2 × 103 cells per well and infected with the viruses at different multiplicity of infection (MOI). Six days after the amount of viable cells was determined by the MTS assay (CellTiter 96 Aqueous One Solution Cell Proliferation assay (Promega). The plates were incubated for 1 hour after which the absorbance of each well was read at a wavelength of 490 nm. All assays were performed in quadruplicate, and each assay was repeated at least twice (17).
“In vivo” studies
All the in vivo studies were approved by the Institutional Animal Care and Use Committee of Instituto Leloir (Protocol #30OP) that has an approved Animal Welfare Assurance as a foreign institution with the Office of Laboratory Animal Welfare, NIH number A5168-01. Whole-body images of each mouse were obtained by the Bioluminescence Assay using an in vivo bioluminescent system (IVIS50; Xenogen) and the Living Image 2.20.1 Software (18, 19).
“In vivo” studies on subcutaneous tumors
Five- to six-week-old female and male athymic N:NIH (S)-nu mice (obtained from the animal facility of the Faculty of Veterinary, University of La Plata, Argentina) were subcutaneously (s.c.) injected in one flank with 5 × 106 cells of SW1990. When the average tumor volume reached 100 mm3, mice received 1 × 109 viral particles per mouse of AV25CDC or PBS administered intratumorally on days 1, 4, and 7 after tumor cells' injection. For the combination of AV25CDC with gemcitabine, mice were injected with AV25CDC followed by intraperitoneal (i.p.) administration of gemcitabine (15 mg/kg for 5 days) starting 1 day after the last AV25CDC injection. In vivo experiments were performed following approval of the Institutional Animal Care and Use Committee (IACUC); all animals under study received food and water ad libitum.
Orthotopic xenograft model
Mice were anesthetized with i.p. injection of 80 mg/kg ketamine (Aveco Co., Inc.) and 10 mg/kg xylazine (Rugby Laboratories, Inc.). A small (1 cm) lateral subcostal laparotomy was performed. A total of 1 × 105 SW1990 cells, suspended in 50 μL PBS + Matrigel 20% v/v (Matrigel Basement Membrane Matrix, BD Biosciences), were injected beneath the capsule of the pancreas, and the abdominal wall and skin were closed. For intratumor treatment, mice received 1 × 109 viral particles per mouse of AV25CDC. For systemic treatment the mice received PBS (50 μL) or 1010 viral particles/mouse in PBS (50 μL) in tail vein. For the combination with gemcitabine, mice were injected i.p. with gemcitabine (15 mg/kg for 5 days) starting 1 day after the last AV25CDC injection. The last day of the experiment (40 or 60 days after first virus administration) animals were bled and serum was used to determine serum markers in a specialized laboratory. Mice were euthanized following institutional guidelines, tumors were removed, weighed, and fixed in 10% buffered formalin for immunohistochemical studies; part of the tumor was kept for E4 quantification. None of the animals showed any signs of toxicity or weight loss throughout the experiment.
Five-week-old male Syrian golden hamsters (weight, 70–80 g) were obtained from and housed at the animal facility of the National Commission of Atomic Energy, Argentina. The animals were anesthetized with i.p. injection of 80 mg/kg ketamine (Aveco Co., Inc.) and 10 mg/kg xylazine (Rugby Laboratories, Inc.). A small (1 cm) lateral subcostal laparotomy was performed. HaP-T1 cells (5 × 105), suspended in 100 μL PBS + Matrigel 20% v/v (Matrigel Basement Membrane Matrix, BD Biosciences), were injected beneath the capsule of the pancreas, and the abdominal wall and skin were closed. Animals were anesthetized and then administered a single intrajugular injection of 1 × 1010 viral particles per hamster of AV25CDC. For the combination of AV25CDC with gemcitabine, hamsters were injected with AV25CDC followed by i.p. administration of gemcitabine (15 mg/kg for every day for 3 days) starting 1 day after the AV25CDC injection. At the end of the experiments (40 days after adenoviral injection) hamsters were euthanized following institutional guidelines. Tumors were removed, weighed, cut, and fixed in 10% buffered formalin for immunohistochemical studies and the other half frozen for E4 quantification. Animals showed no signs of toxicity or weight loss throughout the experiment.
Statistical analyses are detailed in Supplementary Materials and Methods.
AV25CDC exhibited an in vitro cytocidal activity on pancreatic cancer cells
Because the sequence of the human version of the CDC25B promoter was not available, we designed oligonucleotides primers based on the Cdc25B murine promoter and cloned two different fragments of the human genomic sequence of 0.25 kb and 0.45 kb extensions (Supplementary Fig. S1A); the largest region of homology between the murine and the human version extends from −110 to +4 of the murine promoter that includes a TATA box, a NFY site and two SP1 sites (Supplementary Fig. S1A), which were described as functional in the murine promoter (20).
In parallel, we established the infective capacity of several adenoviral vectors carrying different engineered chimeric fibers. Using luciferase expression as a reporter gene, we demonstrated that the strongest activity was associated with the adenovirus carrying the chimeric fiber of serotypes 5/3 (Fig. 1A). Therefore, we constructed nonreplicative adenoviral vectors containing the chimeric 5/3 fiber expressing luciferase downstream of the 0.2 kb (AV25CDCs Luc 5/3) or the 0.45 kb (AV25CDC Luc 5/3) CDC25B promoter variants. We observed that the strongest luciferase activity was associated with the virus carrying the 0.45 kb CDC25B promoter variant (Fig. 1B). The transcriptional activity of the 0.45 kb promoter correlated with CDC25B mRNA levels, as the largest activity was observed in Panc-1 pancreatic cancer cells that exhibited the highest CDC25B mRNA levels (Supplementary Table S1).
Thus, we constructed a CRAd named AV25CDC in which E1A was cloned downstream of the 0.45 kb CDC25B promoter. E1A expression was confirmed by Western blot analysis after infection of SW1990 pancreatic cancer cells with AV25CDC (Supplementary Fig. S1B). AV25CDC was highly lytic on Panc-1, MIA Paca2, and BxPC3 cells even at low MOIs of 0.1–1, whereas SW1990 cells were more resistant to the lytic activity as AV25CDC was effective only at a starting MOI of 10 (Fig. 1C). Using E4 copy number as a readout, we observed up to 5-fold increase in AV25CDC viral particles 72 hours after infection confirming that the virus can indeed replicate in SW1990 cells (Fig. 1C, inset). Wild-type adenovirus 5/3 (Ad5/3WT) was more effective than AV25CDC in BxPC3 and SW1990, whereas both viruses exhibited a similar activity in Panc-1 and MIA PaCa-2 cells (Fig. 1C). AV25CDC also exhibited a strong cytopathic effect on HT29 and LoVo colon cancer cells and MKN45 gastric cancer cells (Supplementary Fig. S1C). On the other hand, AV25CDC showed a largely attenuated lytic effect on nonmalignant cells that express very low levels of CDC25B, compared with Ad5/3WT (Fig. 1D).
The combination of AV25CDC and gemcitabine exhibited the largest therapeutic effect on subcutaneously established human pancreatic tumor xenografts
Treatment with the nucleoside analog gemcitabine (GEM) is the mainstay chemotherapeutics for human pancreatic cancer (21). In initial in vitro studies, we observed that MIA Paca-2 and SW1990 cells were largely resistant to GEM with an IC50 of 1 μmol/L (Fig. 2A) that seemed to correlate with the lack of connexin-26 expression that was shown to facilitate cell to cell passage of GEM through gap junctions (Fig. 2A, inset). Further studies demonstrated that the combination of AV25CDC + GEM was more effective than each single agent alone, on the in vitro growth inhibition of SW1990 cells (Fig. 2B). Based on this evidence, we assessed initially whether a combined chemovirotherapy could be more effective than single agent treatment of s.c. established SW1990 tumors. To avoid any effect of GEM on the cell cycle that would hamper AV25CDC replication, the virus was administered 24 hours in advance of GEM in all the combination treatments. The combination of AV25CDC + GEM reduced tumor growth up to 90% (Fig. 2C) compared with control mice. Surprisingly, GEM alone had no therapeutic effect at all, whereas the virus alone had an intermediate effect (Fig. 2C). We confirmed that AV25CDC was able to replicate inside the tumor mass as we observed a 2-fold increase in E4 copy number at day 7 after a single intratumor administration of AV25CDC (Fig. 2C, inset). We next assessed the fractional tumor volume (FTV) to establish whether the combined chemovirotherapy effect was synergistic (22). Indeed, an FTV value of 5.7 indicated that the chemovirotherapy combination exerted a synergistic therapeutic effect (Supplementary Table S2). Because GEM alone had no effect at all, it was likely that viral administration 24 hours in advance facilitated GEM penetration inside the tumor mass.
Combination of AV25CDC and gemcitabine was highly efficient on orthotopically xenografted human tumors
To further define the therapeutic efficacy of AV25CDC, we administered 5 × 105 Matrigel-embedded SW1990 cells expressing luciferase directly into the mice pancreas. Fifteen days later, mice were split in four groups that received either PBS; GEM alone administered i.p. three times; a single intratumoral (i.t.) injection of 1 × 109 viral particles of AV25CDC; or AV25CDC followed 24 hours later by GEM (AV25CDC + GEM). Mice were sacrificed 25 days later and an autopsy was performed that included removal of the entire tumor area. Microscopic analysis of control tumors showed an infiltrating mucinous adenocarcinoma of the pancreas and areas of remnant pancreatic acinar cells (data not shown). Macroscopic analyses evidenced a clear reduction in tumor size especially between mice treated with AV25CDC + GEM compared with control and GEM-treated mice (Fig. 2D, left). The combination of AV25CDC + GEM induced the largest reduction of 75% in tumor weight, compared with control mice, although no statistically significant differences in tumor size were observed between the AV25CDC-treated group compared with the combination of the OAV + GEM (Fig. 2D, right); even not when mice were injected i.t. twice with the OAV (Supplementary Fig. S1D). The presence of adenoviral E4 gene copies inside the tumor mass at the end of the experiment confirmed that AV25CDC replicated in vivo (Supplementary Fig. S1D, inset).
We also assessed biochemical parameters associated with organ toxicity. Control mice exhibited increased serum levels of amylase, GOT-AST and GPT-ALT indicating that both the pancreas and the liver were compromised by tumor growth (Table 1). GEM treatment induced a reversion to normal levels of GPT-ALT but not of GOT-AST and amylase, indicating that the pancreas and the liver remained highly compromised (Table 1). However, treatment with AV25CDC or AV25CDC + GEM induced a reversion to normal levels in all the biochemical parameters tested (Table 1). Interestingly, control mice exhibited huge serum levels of the tumor biomarker CA 19.9 released by SW1990 cells. At the end of the experiment, the combined treatment with AV25CDC + GEM induced almost 95% reduction in CA 19.9 levels compared with the control, confirming that the largest therapeutic benefit was observed following the combined chemovirotherapy treatment (Table 1).
To further confirm AV25CDC efficacy mice harboring orthotopic SW1990 tumors were split in six different groups that received PBS, PBS + GEM, OAV (i.t.), OAV (i.t.) + GEM, OAV (i.v.), and OAV (i.v.) + GEM. Mice were followed for additional 60 days when control mice were considered as not survivors. Confirming the previous studies, the combination of AV25CDC + GEM exhibited the largest therapeutic effect regardless of whether the OAV was administered i.t. or systemically, although GEM combined with i.t. administered OAV was slightly superior (Fig. 3A). Interestingly, the chemovirotherapy treatment was statistically significantly more efficacious than the OAV or GEM treatments as single agents (Fig. 3A). Moreover, the chemovirotherapy combination also showed the complete reversion to normal levels of all the serum markers of organ toxicity (data not shown). Mice follow up with bioluminescence imaging demonstrated that the chemovirotherapy treatment that included i.t. administered OAV combined with GEM seemed to completely arrest tumor growth (Fig. 3B and C). Similar studies performed with mice harboring orthotopic Mia PaCa-2 tumors confirmed that the combination of AV25CDC + GEM was more efficacious than each single agent alone (Fig. 4A); the combination of GEM with i.t. administered OAV was slightly superior to the combination of GEM with systemically injected OAV although not statistically significant difference was observed (Fig. 4A). On the other the combination of the OAV administered i.t. combined with GEM exhibited statistically significant higher efficacy than the OAV alone (Fig. 4A).
Further confirmation of the therapeutic efficacy of the OAV was attempted with HaP-T1 pancreatic cancer cells on Syrian hamsters. First, we confirmed that HaP-T1 cells indeed expressed high levels of CDC25B (Supplementary Fig. S2A insert). AV25CDC was able to lyse HaP-T1 cells in vitro at MOIs similar to that of Ad5WT and Ad5/3WT (Supplementary Fig. S2A, left); moreover, AV25CDC lytic effect was enhanced in the presence of GEM (Supplementary Fig. S2A right). For the in vivo studies Syrian hamsters harboring 15-day-old orthotopic Hap-T1 tumors were split in four groups that received either AV25CDC Luc 5/3; GEM alone; a single systemic administration of 1 × 1010 viral particles of AV25CDC; or AV25CDC + GEM. The nonreplicative adenovirus was used as a control as it had no effect on the growth of Hap-T1 tumors (data not shown). Hamsters were followed for 40 days when control animals were considered not survivors. The results confirmed that the combined chemovirotherapy exhibited the largest statistically significant therapeutic effect compared with the control or the OAV or GEM treatments as single agents (Fig. 4B). E4 gene copies assessment in the tumor mass confirmed that AV25CDC indeed replicated in vivo (Fig. 4C). In comparison, we observed only neglectable levels of viral particles in lung, kidney, spleen, and liver that are permissive for adenoviral replication (Fig. 4C). Moreover, no histologic abnormality was found in the normal organs of all treated Syrian hamsters (Supplementary Fig. S2B). In addition, AV25CDC was unable to replicate in explants obtained from four different normal Syrian hamster organs compared with Ad5/3WT, which replicated in all of them (Supplementary Fig. S2C).
AV25CDC administration disrupted tumor architecture
The whole in vitro and in vivo data clearly demonstrated a cooperative effect between AV25CDC and GEM suggesting that the prior administration of the OAV might facilitate GEM activity. On the other hand, we were unable to see any stimulatory effect of GEM on the transcriptional activity of the CDC25B promoter (Supplementary Fig. S3A) nor on the cell-surface expression of the adenoviral serotype 3 receptors Desmoglein-2 and CD46 (Supplementary Fig. S3B).
Similar to human tumors, pancreatic tumor xenografts growing s.c. or orthotopically in nude mice contained tumor cells nests surrounded by a dense extracellular matrix of packed collagen fibers with intermingled fibroblasts. These fibers were more prominent in s.c. tumors compared with orthotopic tumors (compare Fig. 5Aa and 5Ab). Histologic analysis of the tumor mass 24 hours after administration of AV25CDC showed a severe disruption of the tumor architecture with large necrotic areas devoid of collagen fibers both in the s.c. (Fig. 5AC) as in the orthotopic tumors treated i.t. or systemically (Fig. 5AD and 5Ae, respectively). Deep analysis evidenced tumor architecture disruption in large areas accounting for 10% to 40% of the entire tumor mass (data not shown). Interestingly, systemic or i.t. administration of the Ad5/3WT virus also disrupted tumor architecture (Supplementary Fig. S3C). Immunohistochemical analysis at 24 hours showed no difference in the amount of infiltrating neutrophils and macrophages (the infiltrate was devoid of NK cells), that located mostly at the tumor periphery (data not shown). The disruption of tumor architecture was increasingly evident at 7 days after OAV administration (Fig. 5Bb and 5Bd) compared with control tumors (Fig. 5Ba and 5Bc). Interestingly, tumor architecture disruption was associated with highly increased levels of activated MMP-9 (Fig. 5C) inside the E1A-expressing tumor mass (Fig. 5D).
We have developed a novel and potent replicative oncolytic adenovirus that was able to inhibit the in vivo growth of locally advanced orthotopically growing human pancreatic tumor xenografts in nude mice and pancreatic tumors in Syrian hamsters.
It was of note that GEM had almost no effect on s.c. tumors and only a partial one on orthotopic tumors, but its combination with the OAV exhibited the largest therapeutic effect that in most cases was also statistically significant different compared with the effect exerted by each single agent. Different studies have reported that the innate resistance of pancreatic cancer to systemic therapies is related to the poor delivery of chemotherapies due to a dense stromal matrix (23–25). Gene expression analysis of GEM resistance in pancreatic cancer demonstrated enrichment in stroma-related genes (26). Targeting the stroma also enhanced macrophage-mediated anti-stroma depletion that increased GEM efficacy (27). The present in vivo studies demonstrated a remarkable disruption in tumor architecture following viral treatment that was coincidental with the increased activation mainly of MMP9. The capacity to disrupt the tumor matrix was also observed upon administration of a wild-type virus pseudotyped with a chimeric fiber 5/3 suggesting that this is an intrinsic characteristic of adenovirus activity. This change in tumor architecture would certainly be permissive for GEM penetration deep inside the tumor mass and provide an explanation for the superior therapeutic efficacy of the chemovirotherapy combination. Because no increase in the inflammatory infiltrate was observed at 24 hours following AV25CDC administration, it is likely that the enhanced MMP9 activity and the disruption of tumor architecture occurs as a direct effect of the OAV on malignant cells. The fact that AV25CDC was able to lyse in vitro WI38 and HFL1 fetal fibroblasts that resemble cancer-associated fibroblasts (CAF) indicates that AV25CDC can be also active on CAF in a clinical setting. In addition, we cannot rule out that E1A can also sensitize pancreatic cancer cells to GEM (28).
Preoperative CA 19.9 levels below certain threshold were associated with increased overall and recurrence-free survival (29). The evidence that the chemovirotherapy treatment reduced CA 19.9 levels by 95% was a clear confirmation of the effective reduction in the tumor mass. Amylase levels have been historically associated in humans with pancreatic duct obstruction and pancreatic cancer (30). The fact that the combined chemovirotherapy and at a lesser extent AV25CDC treatment but not GEM alone, reduced amylase levels, clearly indicates that the treatment was able to reconstitute pancreatic function. GEM treatment reduced ALT but not AST levels. Previous studies in humans have shown that an increase in the ratio of aspartate to alanine aminotransferase (AST/ALT) is an early marker of hepatic fibrosis (31). Thus, it is likely that GEM treatment alone was unable to avoid tumor dissemination to the liver.
Previous studies have shown therapeutic efficacy of CRAds on human s.c. growing xenografts (32, 33). Only in few cases oncolytic activity was also assessed on s.c. tumors in combination with GEM or 5-FU (34–36). Only a single study reported the use of a CRAd based on the Cox-2 promoter to systemically treat orthotopically injected human pancreatic cancer xenografted but without combination with chemotherapy (37). Although the COX-2 promoter–based CRAd appeared very efficient in vivo, this gene is highly expressed in inflammatory situations raising the possibility of unwanted collateral effects on normal organs. It is of note that most pancreatic and gastric cancer tissues expressed only the CDC25B isoform (13) narrowing OAV specificity. Mostly important, all the studies performed in mice and Syrian hamsters demonstrated the lack of toxicity of AV25CDC even when administered in combination with GEM. The chemovirotherapy combination was efficacious regardless of whether the OAV was injected i.t or systemically, although i.t. injection, a usual procedure in pancreatic cancer treatment (10, 38), was more effective. The fact that most pancreatic cancer are diagnosed at an advanced stage makes it unrealistic an intratumor administration unless a secondary systemic immune response can be elicited. In this regard, few oncolytic viral vectors were modified to express immunostimulatory genes such as GM-CSF (12). For systemic administration, it will be necessary to circumvent liver uptake and preexisting immunity elicited by neutralizing antibodies, adenovirus-specific T cells and NK response. One possibility relies on varying the administration routes (39); in addition, the use of carrier cells might act as shields to protect the virus from the immune attack (40). AV25CDC seems to have the capacity to eliminate also colon cancer and gastric cancer cells with clear biosafety parameters. Thus, it can be assumed that AV25CDC might have a broader therapeutic use in different types of cancers from gastrointestinal origin.
Disclosure of Potential Conflicts of Interest
No potential conflicts of interest were disclosed.
Conception and design: D.T. Curiel, E.G. Cafferata, O.L. Podhajcer
Development of methodology: C. Rotondaro, G. Acosta Haab, D.T. Curiel, O.L. Podhajcer
Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): H.L. Weber, S. Werbajh, E. Salvatierra, L. Sganga, E.G. Cafferata
Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): H.L. Weber, E. Salvatierra, E.G. Cafferata, O.L. Podhajcer
Writing, review, and/or revision of the manuscript: H.L. Weber, E.G. Cafferata, O.L. Podhajcer
Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): M. Gidekel
Study supervision: O.L. Podhajcer
Other (awardee of grants): O.L. Podhajcer
This work was supported by the PIA CTE-06, World Bank CONICYT Project, Chile; the National Agency for Promotion of Science and Technology and Amigos de la Fundacion Leloir para la Investigacion en Cancer (AFULIC), Argentina; and the NIH Pancreatic Cancer SPORE grant 2P50CA101955, USA. H.L. Weber was the recipient of the Ph.D. fellowship 21070513 from CONICYT and fellowship 75100015 from the Applied Cellular and Molecular Biology PhD Program, Universidad de La Frontera.
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.