Increased Safety with Preserved Antitumoral Efficacy on Hepatocellular Carcinoma with Dual-Regulated Oncolytic Adenovirus

Hepatocellular carcinoma accounts for >90% of all primary liver cancers and has a dismal prognosis with a median life expectancy of 6 to 9 months. It ranks fifth in frequency among all malignancies worldwide and causes nearly 1 million deaths annually (1). In China, it ranks third in frequency among all malignancies due to high hepatitis B virus infection. As it is known, surgery such as hepatic resection and liver transplantation is the choice of treatment for liver cancers. Yet, no more than 20% of patients with hepatocellular carcinoma have opportunity to undergo surgery procedures. General treatment or combined treatments have drawn great attention due to the poor prognosis of hepatocellular carcinoma. Palliative treatments including chemoembolization, hepatic artery infusion, percutaneous interstitial ablation, cryosurgery, radiation therapy, and systemic chemotherapy have been beneficial complementarities for liver cancer treatment (2). A major limitation with many cancer treatments is the lack of tumor specificity. Therefore, development of effective alternative approaches with novel tumor-targeting mechanism is needed. Replication-selective virotherapy holds great promise for the treatment of cancer (3–5). Appealing features include tumor-selective targeting, viral self-spreading in cancer cells, and no cross-resistance to current treatments. Several types of oncolytic viruses have already been tested in clinical trials, including oncolytic adenovirus, herpes simplex virus, vaccinia virus, reovirus, and Newcastle disease virus.

Oncolytic adenoviruses have attracted considerable interest among these replicating viruses. There are three common strategies exploited to target the oncolytic adenoviruses selectively to tumor cells (6–8). The first strategy involves the deletion of adenovirus genes that are necessary for virus replication in normal cells but not in tumor cells (9, 10).The second strategy to achieve tumor specificity is the use of tumor- or tissue-specific promoters, such as α-fetoprotein, MUC1, PSA, kallikrein-1, and pS2, to drive adenoviral genes that are essential for replication, which is by transcriptional targeting (11–14). The third strategy is based on receptor-mediated targeting of tumor cells through genetic modification of the adenoviral capsid and can be termed transduction targeting (15).

To create an oncolytic adenovirus with strong antitumor efficacy targeting broad-spectrum cancer cells and with reduced side effect in normal cells, a novel adenovirus variant CNHK500, in which the tumor-selective expression of both E1a and E1b genes is accomplished by replacing the nonselective endogenous viral promoters with human telomerase reverse transcriptase promoter and hypoxia response promoter, respectively, was constructed. The human telomerase reverse transcriptase (hTERT) promoter is expected to be active in the majority (>90%) of human cancer cells with up-regulated telomerase expression. The hypoxia response promoter is expected to treat all solid tumors that develop hypoxia, regardless of their tissue of origin or genetic alterations.

The selective reactivation of telomerase in tumor cells offers an attractive therapeutic target for developing new broad-spectrum antitumor agents (16). Telomerases are essential elements at chromosome termini that preserve chromosomal integrity by preventing DNA degradation, end-to-end fusion, rearrangements, and chromosome loss. Each cell replication is associated with the loss of 30 to 150 bp of telomeric DNA that can be compensated by telomerase, an RNA-dependent DNA polymerase. Most human somatic cells exhibit no or very low levels of hTERT expression and telomerase activity, whereby the number of cell divisions is limited because of the reduction of telomeres to a critical length. In contrast to normal somatic cells, in highly proliferative cells, such as germ cells, hematopoietic stem cells, or transformed cancer cells, diverse molecular mechanisms are necessary to maintain telomere length. Although some tumors activated a yet unknown alternative mechanism of telomere extension, the majority (>90%) of human cancer cells acquire immortality by expression of the hTERT. It has been shown that hTERT expression is regulated at the transcriptional level, thereby providing a promising tool for tumor-specific gene expression (16, 17).

Different from the intrinsic high telomerase activity in tumor cells, hypoxia is an integral component of the tumor microenvironment that develops in most solid tumors regardless of their origin, location, or genetic alterations. Hypoxia, a unique feature of human solid tumors, with median oxygen tensions of 1.3% to 3.9% compared with 3.1% to 8.7% in normal tissues, has been considered to be a major factor in the resistance of cancers to radiotherapy and chemotherapy and can also accelerate malignant progression and metastasis (18). Hypoxic conditions initiate a cascade of physiologic responses and lead to the induction of genes involved in glycolysis, erythropoiesis, and angiogenesis. The hypoxia-inducible factor I (HIF-1) transcriptional factor, composed of two subunits, HIF-1α and HIF-1β, mediates transcriptional response to hypoxia by binding to a highly conserved cis-activating hypoxia response element (HRE) present in target genes. In addition to intratumoral hypoxia, mutations in several tumor suppressors, including von Hippel-Lindeu, phosphatase and tensin homologue, and p53, and constitutive activity of AKT have been implicated in promoting HIF-1α accumulation. Several studies have already shown that using HIF/HRE system can specifically target therapeutic gene expression to solid tumors (19, 20).

In our previous study, we have constructed a mono-regulated oncolytic adenovirus, CNHK300, which showed strong antitumor efficacy in many telomerase-positive cancers in vitro and in vivo by using hTERT promoter to control the adenoviral E1a gene (21). Considering that CNHK300 might replicate in normal cells with high telomerase activity, like germ cells and hematopoietic stem cells, we further produced dual-regulated oncolytic adenovirus CNHK500 by replacing the endogenous E1b promoter of CNHK300 with hypoxia response promoter. In this communication, we have evaluated tumor selectivity, cytotoxicity, and antitumor efficacy of CNHK500 in a variety of preclinical models.

Cells and cell cultures. The following cells were purchased from the American Type Culture Collection (Manassas, VA): Hep3B, HepGII (human hepatocellular carcinoma cell lines), A549 (human lung cancer cell line), PANC-1 (human pancreas cancer cell line), HT29 (human colon cancer cell line), HeLa (human cervical adenocarcinoma cell line), MBD231 (human breast cancer cell line), BJ (normal human fibroblast cell line), MRC-5 (embryonic fibroblast cell line), and WRL-68 (normal human fetal liver cell line). Human embryonic kidney (HEK) 293 cell line was obtained from Microbix Biosystem, Inc. (Toronto, Ontario, Canada). Human gastric cancer cell line SGC-7901 and human hepatocellular carcinoma cell line SMMC-7721 were obtained from Institute of Biochemistry and Cell Biology, Chinese Academy of Sciences (Shanghai, China). Human primary hepatocytes were isolated and cultured as described by Liddle et al. (22). All the cells were cultured in the media suggested by the providers, supplemented with 10% heat-inactivated fetal bovine serum (Life Technologies, Inc., Rockville, MD), 4 mmol/L l-glutamine, and 100 units/mL penicillin and 100 μg/mL streptomycin under a 5% CO₂ atmosphere at 37°C. Hypoxic conditions (1% O₂) were achieved with a sealed hypoxia chamber (Billups-Rothenberg, DelMar, CA).

Synthesis of hypoxia response promoter and luciferase assays. Dual luciferase Reporter Assay System Kit and luciferase reporter plasmids (pGL3-basic, pGL3-control, and pRL-TK) were purchased from Promega Corp. (Madison, WI). All restriction endonucleases were purchased from New England Biolabs, Inc. (Beverly, MA). The plasmid pGL3-linker containing an SpeI site (underlined) was generated by introducing linker 1 (annealing primer i and primer ii) into pGL3-basic with BglII/BamHI digestion (Table 1). The hypoxia response promoter (HREp), containing five tandem inverted HREs derived from vascular endothelial growth factor and a mini cytomegalovirus promoter, was synthesized and attached with an SpeI site at each end. This synthesized HREp fragment was first inserted into pUC19 plasmid by BamHI and XbaI digestion, and then released by SpeI digestion and introduced into plasmid pGL3-linker in forward direction to generate pGL3-HREp for promoter activity assays following the methods previously described (21). For hypoxic exposure, cells were exposed to 1% O₂ for 16 hours at 37°C before harvest.

Construction of recombinant virus CNHK500. A dual-regulated shuttle, plasmid pSG500, was generated by the following steps, in which the hTERT promoter regulates the E1a gene and HREp regulates the E1b gene. The first step was to replace the endogenous E1a promoter with hTERT promoter in the pXC1 plasmid, which encodes Ad5 sequences from 22 to 5,790 bp containing the E1 gene (Microbix Biosystem). To remove the endogenous E1a promoter, overlapping PCR was done. A 912-bp fragment was generated by primer iii and primer iv and an 836-bp fragment was generated by primer v and primer vi. These two fragments were used as templates for overlapping PCR with primer iii and primer vi. A 1,716-bp fragment, which lacked endogenous E1a promoter but contained new restriction endonucleases sites AgeI-BstBI-NotI-SpeI-XhoI-SwaI upstream of E1a gene, was generated and cloned into the pXC1 to produce the plasmid pQW1. The 313-bp hTERT promoter was released from CNHK300 and cloned into pQW1 to produce the plasmid pQW1-hTERTp by NotI and SwaI digestion. Thus, the E1a gene in pQW1-hTERTp was under the control of hTERT promoter. Similarly, the E1B endogenous promoter in pQW1-hTERTp was further removed by overlapping PCR. A 326-bp PCR fragment was amplified by primer vii and primer viii and a 363-bp fragment was amplified by primer ix and primer x using pXC1 as templates. A 670-bp fragment was produced by overlapping PCR with primer vii and primer x, which lacked the endogenous E1b promoter but contained an SpeI site upstream of the E1b gene. The 670-bp fragment replaced the 710-bp fragment in pQW-hTERTp by KpnI and XbaI digestion to produce the plasmid pQW2. The final step was to insert the fragment of HREp from pUC19-HREp by SpeI digestion into pQW2 in forward orientation to produce the dual-regulated plasmid pSG500. A 1,538-bp fragment of the enhanced green fluorescent protein (EGFP) expression cassette, released from the plasmid pXC7C-GFP that we had previously constructed, was inserted into pSG500 upstream of the hypoxia response promoter by AgeI and NotI digestion to produce pSG500-EGFP. The plasmids pSG500 and pSG500-EGFP were cotransfected with pBHGE3 (Microbix Biosystem) into HEK293 cells by Lipofectamine 2000 (Life Technologies) to rescue recombinant virus CNHK500 and CNHK500-EGFP as previously described (21).

Western blot analysis of E1A and E1B expression. HepGII, Hep3B, SMMC-7721, human primary hepatocytes, and BJ cells were seeded in six-well plates at a density of 5 × 10⁵ per well and infected with CNHK500 or wild-type adenovirus 5 (wtAd5) at indicated multiplicities of infection (MOI) after 24 hours of incubation. Two days after viral infection, cells were harvested and lysed with M-PER Mammmalian Protein Extraction Reagent (Pierce, Rockford, IL). Equal amount of 20 μg of protein was separated on 10% SDS-polyacrylamide gel, electroblotted onto PROTRAN nitrocellulose transfer membrane (Schleicher & Schuell, Inc., Dassel, Germany), and blocked with 5% fat-free milk in TBS [10 mmol/L Tris (pH 7.5), 0.9% NaCl] containing 0.1% Tween 20 (TBST) at room temperature for 1 hour. The membrane was incubated with either rabbit polyclonal antibody against Ad2 E1A protein (Santa Cruz Biotechnology, Santa Cruz, CA) or rat anti–Ad5 E1B 55-kDa monoclonal antibody overnight at 4°C and washed thrice in TBST. After incubation for 1 hour with horseradish peroxidase–conjugated secondary antibodies and extensive washing with TBST, immunocomplexes on the membrane were detected with LumiGLO reagent and visualized with Kodak BiomaxMR film. To detect the expression of E1A and E1B under hypoxia, cells infected with adenoviruses were exposed to hypoxia for 16 hours before harvest.

In vitro viral replication assay. Contact-inhibited normal cells and log-phase tumor cells (plated at 60-70% confluence) were prepared in six-well plates and infected with recombinant adenoviruses at an MOI of 5.0. Virus inocula were removed after 2 hours. The cells were then washed twice with PBS and incubated at 37°C for 0, 12, 24, 48, or 96 hours. Cell lysates were prepared with three cycles of freezing and thawing. Serial dilutions of the lysates were titered on HEK293 cells with the tissue culture 50% infective dose (TCID₅₀) method, normalized with that at the beginning of infection and reported as multiples. WtAd5, mono-regulated adenovirus CNHK300, and ONYX-015 (a gift from A.J. Berk, University of California-Los Angeles, Los Angeles, CA) were used as controls.

To observe the replication of CNHK500-EGFP, liver cancer cell lines (Hep3B and SMMC-7721) and normal cells (human primary hepatocytes and BJ cells) were plated in six-well plates at a density of 4 × 10⁵ per well, cultured for 24 hours, then infected with the virus at an MOI of 0.01. On days 3, 7, and 10 postinfection, cells were observed under a fluorescent microscope and photographs were taken.

In vitro cell viability assay. 3-(4,5-Dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) assay was done to determine cell viability at various viral MOIs. Hep3B, SMMC-7721, HepGII, human primary hepatocytes, WRL-68, and BJ cells were plated in 96-well plates (BD Falcon, Bedford, MA) at a density of 1 × 10⁴ per well. Twenty-four hours later, the cells were infected with CNHK500 at serial MOIs from 0.001 to 1,000. After 7 days of incubation, cell viability was measured by the MTT assay with the nonradioactive Cell Proliferation Kit (Roche Molecular Biochemicals, Indianapolis, IN) according to the kit protocol. The spectrophotometric absorbance of the samples was measured with Microplated Reader Model 550 (Bio-Rad Laboratories, Tokyo, Japan) at 570 nm with a reference wavelength of 655 nm. The percentage of cell survival was calculated using the following formula: % cell survival = (absorbance value of infected cells / absorbance value of uninfected control cells) × 100%. Eight replicative samples were measured at each MOI and each experiment was done at least thrice. The IC₅₀ for each cell line was calculated and compared.

In vivo tumor killing assay. Three protocols were followed to evaluate the efficiency of CNHK500 in killing hepatocellular carcinoma xenografts in vivo. All procedures were approved by the Committee on the Use and Care on Animals and done in accordance with the institution guidelines.

In the first protocol, SMMC-7721 tumor xenografts were established by s.c. inoculation of 5 × 10⁶ cells into the right flanks of 6- to 8-week-old BALB/c nude mice (Institute of Animal Center, Chinese Academy of Sciences, Shanghai, China). When tumors reached 7 to 9 mm in diameter, 40 mice were randomly assigned to three treatment groups (CNHK300, CNHK500, and ONYX-015) and one control group (treated with virus preservation buffer). Preestablished tumors were then injected with 100 μL of the control buffer or 2 × 10⁸ plaque-forming units (pfu) of viruses in the same solution. The injections were repeated five times every other day, with a total dosage of 1 × 10⁹ pfu. Tumor growth was monitored by periodic measurements with calipers and tumor volume was calculated using the following formula: (maximal length) × (perpendicular width)² / 2. All animals were sacrificed 6 weeks posttreatment and tumors were dissected and weighed. The tumor reduction rate was calculated using the following formula: [1 − (V_{treatment group end point} − V_{treatment group initial point}) / (V_{control group end point} − V_{control group initial point)}] × 100%.

In the second protocol, Hep3B tumor xenografts were established and treated in a similar way as mentioned above to evaluate the antitumor efficacy of CNHK500 on Hep3B tumor-bearing mice by i.t. administration.

In the third protocol, viruses were administrated by tail vein injection. Twenty-eight mice bearing Hep3B tumor xenografts were randomly assigned into four groups and treated five times every other day: CNHK500 high-dosage group, injected with 2 × 10⁹ pfu viruses; CNHK500 intermediate-dosage group, injected with 1 × 10⁹ pfu viruses; CNHK500 low-dosage group, injected with 2 × 10⁸ pfu viruses; and control group, injected with virus preservation buffer. Periodic tumor measurements were taken and the behavior of mice was observed until the experiments were terminated because of heavy tumor burden.

Histopathologic and immunohistochemical study. The transplanted tumors, dissected from nude mice, were fixed in 10% formalin, paraffin embedded, and cut into 4-μm-thick sections. All sections were baked, deparaffinized, and heated in citrate buffer (10 mmol/L citric acid, pH 6.0) in a microwave oven. After inactivation by exposure to 1.5% H₂O₂/methanol for 10 minutes for blocking the endogenous peroxidase, the sections were incubated with blocking serum (goat serum) at room temperature for 30 minutes. Immunohistochemistry was carried out with anti–adenoviral hexon protein antibody (Biodesign International, Saco, ME) and the UltraSensitive Streptavidin-Peroxidase Kit (Maxim Pharmaceuticals, San Diego, CA).

Statistical analysis. Absorbance value in MTT assay and mean tumor volume were presented as mean ± SD and compared at a given time point by unpaired, two-tailed t test. Survival was analyzed by the Kaplan-Meier method. Results were compared for statistical significance by applying the generalized Wilcoxon test. Data were considered statistically significant when P < 0.05.

Transcriptional activity of Hypoxia response promoter under normoxic and hypoxic conditions. The 240-bp hypoxia response promoter (HREp), containing five tandem inverted copies of HRE derived from the human vascular endothelial growth factor promoter and a mini cytomegalovirus promoter, was synthesized and cloned into the plasmid pGL3-linker to produce pGL3-HREp. Its transcriptional activity was measured. Under normoxic conditions, the pGL3-HREp showed 66% to 67% activity of the positive plasmid pGL3-control (with SV40 promoter/enhancer) in Hep3B and HepGII cancer cells and 6.7% to 11.3% activity of the control in normal cells (Fig. 1A). When cells were exposed to hypoxia, the luciferase activity of pGL3-HREp increased 6.7- to 10.8-fold either in cancer cells or in normal cells, indicating the sensitivity and functionality of the hypoxia response promoter. The luciferase activity of the pGL3-control showed no obvious response to hypoxia treatment (Fig. 1B).

Selective expression of adenoviral early genes of CNHK500 in vitro. We developed a dual-regulated adenoviral variant CNHK500 in which the E1a gene was controlled by the hTERT promoter and the E1b gene by the hypoxia response promoter, as depicted in Fig. 2A. The E1A and E1B proteins were checked in cells infected with CNHK500 or wtAd5 as a positive control. In Fig. 2B, the E1A protein of CNHK500 was detected in telomerase-positive cancer cells Hep3B, SMMC-7721, HepGII but not in normal human primary hepatocytes and fibroblast cells, either under normoxia or hypoxia, showing restricted control of E1a gene by hTERT promoter. When cells were infected with CNHK500 at an MOI of 1, the E1B protein was hardly detected in all cells under normoxia but was clearly detected in cancer cells under hypoxia (Fig. 2C and D). When MOI increased, E1B protein was detected both in cancer cells (MOI, 10) and normal cells (MOI, 100) under normoxia, but the protein level was lower than that under hypoxia (Fig. 2D). For wtAd5, E1A and E1B proteins were detected both in cancer cells and in normal cells (Fig. 2B and C).

Selective replication and oncolysis of CNHK500 in telomerase-positive cancer cells. According to the mechanism of oncolytic virus, the most important character of oncolytic virus is its ability of selectively replicating in tumor cells. As shown in Fig. 3A, CNHK500 replicated efficiently in all tested cancer cells but weakly in normal cells. The improvement of replication specificity of CNHK500 was validated in a series of cancer cells and normal cells (Fig. 3A and B). At 48 hours after viral infection, the replication of CNHK500 was severely impaired in human primary hepatocytes, MRC-5, and BJ cells when compared with CNHK300, ONYX-015, or wtAd5 (Fig. 3B). For example, the replication of CNHK500 was 302- to 777-fold less than that of wtAd5 in MRC-5 or BJ cells and 9- to 19.5-fold less than that of CNHK300. In all tested cancer cells, CNHK500 replicated as efficiently as mono-regulated CNHK300 and better than ONYX-015. The replication of CNHK500 was slightly increased in tested cells under hypoxia compared with that under normoxia (Fig. 3C). The selective replication and efficient report gene expression of CNHK500-EGFP, which contains the green fluorescent protein expression cassette, was also observed in liver cancer cell lines and human primary hepatocytes, as shown in Fig. 3D.

In addition, MTT assays were done to quantify the selective oncolytic effect of CNHK500. Within 7 days, CNHK500 killed 50% of liver cancer cells (IC₅₀) at MOIs between 0.01 and 2 whereas 50% killing of BJ or WRL-68 cells requires an MOI of between 398.56 and 1,360.14 (Fig. 3E).

Treatment of human liver cancer xenografts with CNHK500. The selective killing of hepatocellular cancer cells with CNHK500 was observed in vitro. So we evaluated the antitumor efficacy of CNHK500 on two types of hepatocellular carcinoma xenografts in vivo by different routes of viral administration and diverse treatment dosages.

In the SMMC-7721 xenografts models, CNHK500 i.t. treatment significantly inhibited tumor growth compared with the control (Fig. 4A; P < 0.001) from day 21 posttreatment. Compared with the control group, on days 21, 28, and 35 after initial virus administration, tumor mean volumes were reduced by CNHK500 to 83.45%, 70.88%, and 59.19%, and by CNHK300 to 74.18%, 64.37%, and 48.97%, respectively. Both CNHK500 and CNHK300 treatment groups showed stronger antitumor activity than the ONYX-015 treatment group from day 21 posttreatment (P < 0.005).

More significant inhibition of tumor growth was observed in Hep3B tumor xenograft model by CNHK500 i.t. treatment (Fig. 4B). The mean volumes of Hep3B xenografts in CNHK500 treatment group increased a little during the whole observation course, whereas those in the control group increased dramatically (P < 0.001). CNHK500 showed stronger antitumoral effect on Hep3B xenografts than on SMMC-7721 xenografts, which is consistent with the viral replication and oncolysis in vitro.

When different dosages of CNHK500 were administrated in mice bearing Hep3B tumor xenografts by tail vein injection, all treatment groups showed significant antitumor efficacy compared with the control group (P < 0.01; Fig. 4C). Obvious dose-effect relation was observed among all treatment groups. The antitumor efficacy was enhanced with the increase in CNHK500 dosage. On day 91 posttreatment, seven mice in CNHK500 high-dosage treatment group and six of seven mice in CNHK500 intermediate- or low-dosage treatment group survived, whereas all mice in the control group died. All CNHK500 treatment groups manifested a significant survival benefit compared with the control group (P < 0.0001). The median survival times in the control group, CNHK500 low-dosage treatment group, and intermediate-dosage treatment group were 71.04, 107, and 127.5 days, respectively (Fig. 4D). Five of seven mice still survived on day 130 posttreatment before the whole experiment was terminated.

All tumor samples were examined histologically by H&E staining and immunohistochemical staining for hexon protein. In the control group, the cancer cells grew luxuriantly with small foci of necrosis. In the treatment groups, many wide areas of necrosis were observed and hexon protein was detected in the cytoplasm of tumor cells. These results indicated that CNHK500 selectively replicated in, and efficiently lysed, tumor cells (Fig. 5).

Specificity and efficiency of viral replication in target tumor cells are the two important variables used to evaluate the potential of any novel oncolytic vectors (7, 23–28). In the current study, we have successfully constructed an oncolytic adenovirus with broad-spectrum antitumor activity, CNHK500, in which hTERT promoter drives the E1a gene and hypoxia response promoter controls the E1b gene. Although both of these promoters have been used as single controlling elements in other oncolytic adenoviruses or in two dual promoter-controlled viruses (21, 24, 27, 29–36), this is the first report to use them simultaneously to build a new dual-regulated oncolytic adenovirus.

Infection of cells with CNHK500 showed that E1A protein was selectively expressed in telomerase-positive cancer cells but not in normal cells under normoxia or hypoxia. E1B protein expression was detected in tumor cells under hypoxia but not under normoxia when MOI was <5. Only trace E1B protein was detected in normal cells when MOI reached 100. Under the same MOI enough for infecting cancer cells, neither E1A nor E1B protein expression was detected in normal cells under normoxia or hypoxia, which means the replication of CNHK500 would be greatly attenuated in normal fibroblast cells and human primary hepatocytes if used as viral therapy for liver cancers. Although some researchers reported restricted replication of adenovirus under hypoxia (37, 38), we did not get the similar results in our study. CNHK500 replicated efficiently in tested hepatocellular carcinoma cells under either normoxia or hypoxia. The possible reason is that the five tandem inverted HREs derived from vascular endothelial growth factor in hypoxia response promoter acted as enhancer and were effectively activated under hypoxia. Thus, the essential viral gene expression of CNHK500 was increased instead of decreased, which compromised hypoxia-induced reductions in E1A levels that were mediated at the posttranscriptional level. The viral replication of CNHK500 was moderately increased when the infected cells, including tumor and normal cells, were exposed to hypoxia, implying the potency of synthesized hypoxia response promoter. Likewise, Post et al. (35) also observed efficient viral replication of a hypoxia/HIF-dependent oncolytic adenovirus, HYPR-Ad, under hypoxic microenviroment, and Cho et al. reported potent oncolytic effect of Ad.Δ55.HRE under hypoxia by using HRE-expression system to control E1a gene (36).

We also compared the replication capacity, safety, and cytotoxicity of dual-regulated oncolytic adenovirus CNHK500 with mono-regulated CNHK300 and ONYX-015, the effectiveness and safety of which have been extensively investigated in a number of cancer patients. CNHK500 showed more efficient replication and oncolysis in all tested cancer cells and greater attenuation in tested normal cells compared with ONYX-015 even under normoxia. Moreover, CNHK500 showed excellent in vitro oncolysis comparable to CNHK300, but was ∼10 to 100 times attenuated in normal cells, confirming better cancer cell specificity by simultaneous regulation of E1a and E1b genes by hTERT and HRE promoters. The results also support the use of multiple transcriptional regulatory elements to control multiple, essential viral genes that can lead to greater specificity for the target tissue without further attenuation of viral replication, which was also validated in other researchers' studies (7, 27, 29). Furthermore, the possible replication of CNHK300 in normal cells with telomerase activity, like germ cells and hematopoietic stem cells, can be reduced by restricted regulation of E1b gene by hypoxia response promoter. In MTT assay, CNHK500 killed half of liver cancer cells at an MOI of 0.01 to 2.03 and killed half of normal cells at an MOI of 398.56 to 1,360.14. The data suggested that 196- to 128,301-fold more viral particles were needed to achieve similar cytolysis in normal cells compared with hepatocellular carcinoma cells, again indicating that the dual-regulated oncolytic adenovirus CNHK500 would be a safer and more efficient strategy for liver cancer viral therapy.

Finally, the dual-regulated oncolytic adenovirus CNHK500 had shown significantly high antitumor activity compared with ONYX-015 in tested SMMC-7721 and Hep3B xenografts. CNHK500 showed stronger antitumor efficacy on Hep3B xenografts by i.t. treatment compared with SMMC-7721 xenografts, which was consistent with the effective replication and oncolysis of CNHK500 in Hep3B cells in vitro. We also observed the antitumor ability of CNHK500 on Hep3B xenografts with different dosages of CNHK500 by tail vein injection. Compared with the control group, all CNHK500 treatment groups with different dosages showed significant tumor growth repression and prolonged survival. Obvious dose-effect relation was observed among three CNHK500 treatment groups, in which the high-dosage treatment group showed the strongest antitumor potency with five of seven tumor-bearing mice surviving 4 months after treatment. Although they were able to block the growth of human tumor xenografts in nude mice, they were unable to completely eradicate them in most of the animals. There may be physical barriers that limit the dissemination of the virus in the tumor mass. This problem remains to be resolved to enhance the oncolytic potential of these agents. Although CNHK500 showed greatly attenuated replication in human primary hepatocytes in vitro, it is still hard to assess the cytotoxicity of CNHK500 in vivo due to nonreplication of human adenovirus in murine cells and lack of appropriate animal models (21). Wang et al. (39) reported a novel assay to assess primary human cancer infectibility of oncolytic adenovirus using ex vivo culture of primary tumor tissues. It would be more persuasive to evaluate the specific replication and oncolysis of CNHK500 on ex vivo culture of primary hepatocellular carcinoma tissues and normal liver tissues.

In our present study, we have proved that the simultaneous regulation of E1a and E1b transcription by appropriate combination of promoters can increase the tumor specificity of oncolytic adenovirus with preserved antitumor activity. These observations strongly suggest that the use of replication-competent oncolytic adenovirus under the control of two different intrinsic promoters may hold strong promise for treating broad-spectrum tumors. The appropriate selection and disposal of exogenous promoter is important for replication-competent adenovirus. It must be very interesting to exchange the positions of the two exploited promoters in CNHK500 to produce another novel dual-regulated oncolytic adenovirus, in which hypoxia response promoter drove the E1a gene and hTERT promoter regulated the E1b gene. Stronger antitumor activity was expected by combining i.t. injection of CNHK500 with i.v. administration or by continuous treatments. The efficacy of CRAds can be further increased when armed with therapeutic genes, which is also called gene-viral therapy (40–42). The containing genes will be easily introduced to tumor cells and expressed efficiently along with the viral replication and will accumulate to therapeutic concentration in the tumor region. Many genes can be chosen as therapeutics including tumor suppressor genes, encoding cytokines or prodrug activating enzymes, and antiangiogenic genes. Our preliminary research on constructing a highly specific and wide-ranged oncolytic adenovirus paves the way for our future studies. We have built a series of such recombinant adenoviruses as CNHK500-mE (containing mouse endostatin), CNHK500-hγ (containing human IFN-γ), and CNHK500-p53 (containing p53 gene). The efficacy assessment of these adenoviruses is under way and we do hope it will provide a promising strategy for cancer treatment.

We thank Linfang Li, Yanzhen Qian, and Lihua Jiang for their assistance with viral recombination and cell culture in our laboratory.