Abstract
Background/Aim: Oncolytic adenoviruses (OAds) have attracted much attention as novel anticancer therapeutics. The proper design of an expression cassette containing the E1A gene, which is indispensable for self-replication of the Ad genome, is crucial for efficient tumor cell-specific infection of an OAd. Various types of oncolytic adenoviruses (OAds) possessing different types of the E1A gene expression cassettes have been developed, but their oncolytic activities and safety profiles have not been systematically evaluated. Herein we examined the oncolytic activities and safety profiles of five types of OAds possessing different types of the E1A gene expression cassette in order to optimize the E1A gene expression cassette for development of an efficient and safe OAd. Materials and Methods: We prepared five types of OAds containing different types of E1 gene expression cassettes, and examined the oncolytic activities and safety profiles of the OAds. Results: Among the OAds examined, OAd-Δ24, which had a 24-bp deletion in the E1A gene, mediated the most efficient oncolytic activities against the human tumor cell lines, although OAd-Δ24 showed slightly higher cytotoxicity to normal human cells than the other OAds. Conclusion: These results provide important clues for the development of safe and efficient OAds.
Oncolytic virotherapies using oncolytic viruses, which replicate in a tumor cell-specific manner and mediate efficient tumor cell killing activities, have attracted much attention as a novel therapeutic approach for various types of tumors (1, 2). More than 10 species of oncolytic viruses, including adenovirus, herpes virus, and reovirus, have been developed. Pre-clinical and clinical trials using oncolytic viruses have been carried out worldwide and have shown promising results. Talimogene laherparepvec (T-VEC), a genetically modified herpes simplex virus type 1 (HSV-1), was approved by the U.S. Food and Drug Administration (FDA) in 2015 as a first-in-class oncolytic virus for the treatment of metastatic melanoma.
Among the various types of oncolytic viruses, the most promising are the oncolytic adenoviruses (OAds); their several advantages include efficient tumor cell killing activities and relatively large capacities for transgene insertion (3). In OAds, the E1A gene is crucial for regulation of the virus gene expression and self-replication of the virus genome. This gene must function or be expressed in a tumor cell-specific manner in order to permit tumor cell-specific infection, making the design of the E1A gene expression cassette crucial for efficient and tumor cell-specific infection with oncolytic adenoviruses (OAds) without cytotoxicity to normal cells. Various types of E1A gene expression cassettes have been developed, but they can be largely divided into two groups. In the most widely used type of E1A gene expression cassette, the E1A gene is driven by a tumor-specific promoter, including human telomerase reverse transcriptase (hTERT) promoter and survivin promoter (4, 5). In the other type of cassette, the E1A (a 24-bp deletion) or E1B gene (a 19-kDa or 55-kDa gene deletion mutant) is genetically modified for tumor cell-specific replication and tumor cell killing activities (6, 7). Cassettes using a combination of these two types of modification have also been tested (8, 9). However, the replication efficiencies, tumor cell killing activities, and safety profiles of OAds containing the different types of E1A gene expression cassette have not been systematically evaluated.
In this study, we prepared five types of OAds containing different types of E1 gene expression cassettes, and examined the E1A expression levels, virus replication levels, and cytotoxicities in various types of human tumor cell lines and two types of normal human cells. Among the five types of OAds tested, an OAd containing a 24-bp deletion in the E1A gene driven by the native E1A gene promoter (OAd-Δ24) showed the most efficient replication and cytotoxic effects on the tumor cells.
Materials and Methods
Cells. Panc-1 (a human pancreatic cancer cell line), HepG2 (a human hepatocellular carcinoma cell line: JRCB1648), HEK293 (a human transformed embryonic kidney cell line), HeLa (a human epithelial carcinoma cell line; JCRB9004), HCT116 (a human colorectal carcinoma cell line), SW620 (a human colorectal adenocarcinoma cell line), KK47 (a human bladder cancer cell line), and MRC-5 (normal human embryonic lung fibroblast: CCL-171) cells were cultured in Dulbecco’s modified Eagle’s medium supplemented with 10% fetal bovine serum (FBS), streptomycin (100 μg/ml), and penicillin (100 U/ml). MCF7 (a human breast cancer cell line) and H1299 (a human non-small cell lung carcinoma cell line) cells were cultured in RPMI1640 supplemented with 10% FBS, streptomycin (100 mg/ml), and penicillin (100 U/ml). Human umbilical vein endothelial cells (HUVEC) were cultured in the medium recommended by the supplier (Lonza, Basel, Switzerland).
Plasmids and viruses. OAd plasmids containing tumor-specific promoters were constructed by an improved in vitro ligation method (10, 11) using pAdHM3 (10) and shuttle plasmids encoding the E1A and E1B genes. Information about the shuttle plasmids encoding the E1A and E1B genes is available on request. An OAd plasmid for OAd-Δ24 was produced by homologous recombination between pTG3602 (12) and an Ad genome fragment (bp 121-1890) containing a 24-bp deletion (bp 923-946).
Recombinant OAds were prepared as follows: PacI-digested OAd plasmids were transfected into HEK293 cells using Lipofectamine 2000 (Thermo Fisher Scientific, San Jose, CA, USA), producing OAds. The OAds were then propagated in H1299 cells, purified by two rounds of cesium chloride gradient ultracentrifugation, dialyzed, and stored at –80°C. The determination of virus particle (VP) titers was accomplished according to Maizel et al. (13).
Cell viability assays. In the crystal violet staining assay, cells were seeded on a 24-well plate at a density of 2-5×104 cells/well. On the following day, cells were infected with OAds at the indicated VP/cell. After 5-10 days incubation, the medium was removed. Cells were then fixed with 4% paraformaldehyde phosphate buffer solution (FUJIFILM Wako Pure Chemical, Osaka, Japan) for 1 h at room temperature, then treated with 1 ml of 2% crystal violet in 100% methanol. The plates were washed, dried and observed.
In the WST-8 assay, cells were seeded on a 96-well plate at a density of 0.5-1×104 cells/well. On the following day, cells were infected with OAds at the indicated VP/cell. Cell viabilities were determined using a Cell Counting Kit-8 (Dojindo Laboratories, Kumamoto, Japan) on the indicated days.
Real-time PCR analysis. Cells were seeded at a density of 1×105 cells/well in a 12-well plate. On the following day, cells were infected with OAds at the indicated VP/cell. Total DNA, including OAd genomic DNA, was recovered from the cells using a DNeasy Blood & Tissue Kit (Qiagen, Valencia, CA, USA). After isolation, the Ad genome copy numbers were quantified using a StepOnePlus real-time PCR system (Applied Biosystems, Foster City, CA, USA) as previously described (14). For determination of the hTERT and survivin mRNA levels, total RNA was recovered from the tumor cells, followed by real-time RT-PCR analysis as previously described (15). For measurement of the E1A mRNA levels, total RNA was recovered from the tumor cells following an 8-h infection, followed by real-time RT-PCR analysis. The sequences of the primers used in this study are available on request.
Antibodies and western blotting. Human tumor cells were seeded at a density of 1×105 cells/well in a 12-well plate. On the following day, cells were infected with OAds and WT-Ad at 300 VP/cell. Total protein was recovered from the cells 8 h after infection. Western blotting analysis was performed as previously described (16). Briefly, whole-cell extracts were prepared and electrophoresed on 10% sodium dodecyl sulfate (SDS)-polyacrylamide gels under reducing conditions, followed by electrotransfer to polyvinylidene difluoride membranes (Millipore, Bedford, MA, USA). After blocking with 5% skim milk or 5% bovine serum albumin (BSA) prepared in TBS-T (Tween-20, 0.1%), the membranes were incubated with mouse anti-adenovirus 2/5 E1A antibody (M73, 1:500; Santa Cruz Biotechnology, Santa Cruz, CA, USA) and rabbit anti-human β-tubulin polyclonal antibody (1:10000; Abcam, Cambridge, UK), followed by incubation with horseradish peroxidase (HRP)-labeled anti-mouse and anti-rabbit IgG antibodies (1:5,000; Cell Signaling Technology, Danvers, MA, USA). The protein bands were visualized with a chemiluminescence kit (ECL Plus Western blotting detection system; Amersham Biosciences, Piscataway, NJ, USA).
Results
Construction of OAds containing different types of E1A gene expression cassette. In order to directly compare the E1A gene expression cassettes for the development of safe and efficient OAds, we produced five OAds containing different types of the E1A gene expression cassette (Figure 1). OAd-tAIB contained the E1A and E1B genes linked with an internal ribosome entry site (IRES) under the control of the hTERT promoter. OAd-tANB possessed the hTERT promoter-driven E1A gene expression cassette. The E1B gene was regulated by the native E1B promoter in OAd-tANB. Three amino acid substitution (L122V, C124S, E126D) was introduced into the E1A gene of OAd-tANB according to a previous study (17), producing OAd-tmANB. This substitution was demonstrated to disrupt not only retinoblastoma protein (Rb) binding but also cyclic guanosine monophosphate-adenosine monophosphate synthase (cGAS)-stimulator of interferon genes (STING) pathway inhibition (18). OAd-Δ24 was produced by a 24-bp deletion (bp 922-947) responsible for the Rb binding in the E1A gene according to the previous study (7). OAd-smANB was produced by replacing the hTERT promoter of OAd-tmANB with a survivin promoter for transcription of the E1A gene. All OAds were normally propagated to high titers in the packaging cells.
Tumor cell lysis activities of OAds. Next, in order to examine the tumor cell lysis activities of OAds, eight types of human tumor cell lines were infected with OAds. The expression levels of the hTERT and survivin differed among the tumor cell lines (Figure 2). A crystal violet staining assay demonstrated that OAd-Δ24 mediated the highest levels of tumor cell killing activities among the tumor cell lines examined (Figure 3A). Almost all tumor cells were lysed at 3 VP/cell following a 5-day incubation with OAd-Δ24 in HepG2, H1299, and SW620 cells. In addition, tumor cell viabilities following OAd-Δ24 infection were significantly lower than those following WT-Ad infection. The other OAds mediated similar levels of tumor cell lysis activities, although the viabilities of HepG2 and SW620 cells were slightly higher in the OAd-smANB-treated group, compared to the other OAd-treated groups. Tumor cell viabilities following OAd infection were also determined by WST-8 assay (Figure 3B). The cell viabilities progressively decreased over the incubation periods for all types of OAds. In keeping with the results of the crystal violet staining assay, OAd-Δ24 showed the most efficient tumor cell lysis in H1299, HepG2, and Panc-1 cells. These results indicated that OAd-Δ24 mediated the highest tumor cell lysis activities among the OAds examined.
OAd genome levels in the tumor cells following infection. In order to examine the replication levels of OAds in the tumor cells, the OAd genome copy numbers in the tumor cells were determined by real-time PCR analysis following infection. The Ad genome copy numbers 3 h after infection were similar for all OAds and WT-Ad in the tumor cells (Figure 4), suggesting that the OAds and WT-Ad exhibited similar levels of cellular binding and uptake in the tumor cells. Approximately 2-5-fold higher levels of Ad genome copy numbers were found for OAd-Δ24, compared with the other OAds and WT-Ad, 72 h after infection. These data indicated that the OAd-Δ24 genome replicated more efficiently in the tumor cells than the other OAds or the WT-Ad.
E1 gene expression levels in the tumor cells following OAd infection. In order to examine whether OAd-Δ24 mediated the highest levels of E1A gene expression, leading to the highest levels of OAd genome copy numbers and tumor cell lysis activities in the tumor cells, we examined the E1A gene expression levels following an 8-h infection with OAds and WT-Ad in the tumor cells. Real-time RT-PCR analysis demonstrated that WT-Ad, OAd-Δ24, and OAd-smANB exhibited higher levels of the E1A mRNA expression in H1299 and Panc-1 cells than the other OAds (Figure 5A). There were no apparent differences in the E1A mRNA levels between WT-Ad, OAd-Δ24, and OAd-smANB. We also examined the E1A protein levels 8 h after OAd infection by western blotting analysis. OAd-Δ24 mediated the higher levels of E1A protein expression than WT-Ad and the other OAds (Figure 5B). The other OAds produced comparable or slightly lower levels of the E1A protein, compared to the WT-Ad. These data indicated that OAd-Δ24 mediated the highest levels of E1A protein production, leading to efficient oncolysis of tumor cells.
Cytotoxic effects of OAds on normal human cells. Next, in order to evaluate their safety profiles, the OAds were added to normal human cells. Neither the OAds nor the WT-Ad mediated apparent cytotoxicities in MRC-5 cells, which are human fetal lung fibroblasts, at the virus doses examined (Figure 6). This was partly due to the negligible levels of coxsackievirus-adenovirus (CAR) receptor expression on MRC-5 cells (19). On the other hand, significant cytotoxicities were found in HUVEC following infection with OAds and WT-Ad at high virus doses. WT-Ad exhibited detectable levels of cytotoxic effects on HUVEC at more than 30 VP/cell. Among the OAds tested, OAd-Δ24 showed the highest levels of cytotoxicities in HUVEC. Significant reduction in the viabilities of HUVEC were found at 30 VP/cell of OAd-Δ24, on the other hand, the other types of OAds did not appear to mediate cytotoxicities in HUVEC at 30 VP/cell. These data indicated that the cytotoxicity of OAd-Δ24 to normal human cells was slightly higher than those of the other OAds.
Discussion
In this study, we optimized the E1A gene expression cassette by comparing the tumor cell lysis activities and safety profiles of five OAds containing different types of the E1A gene expression cassette or the WT-Ad for development of efficient and safe OAds. Various types of OAds containing the different E1A expression cassettes, which play a crucial role in self-replication of OAds, have been developed so far; however, tumor cell killing activities and safety profiles in normal human cells of various types of OAds containing the different E1A expression cassettes have not been systematically examined. This study demonstrated that among the OAds examined, OAd-Δ24 exhibited the most efficient tumor cell lysis activities in the human tumor cell lines. OAd-Δ24 mediated higher levels of oncolytic activities in the tumor cells than the WT-Ad (Figure 3). Similar results were observed in a previous study (18). A 24-bp deletion responsible for the Rb binding in the E1A gene prevented the association between the E1A protein and Rb, leading to the attenuation of virus replication in normal cells; however, Heise et al. proposed that this mutation in the E1A gene enhanced the favorable potency of the E1A protein for Ad replication via not only ablation of the Rb binding but also abrogation of the E1A autoregulation and alteration in the E1A phosphorylation (18). The 24-bp deletion in the E1A gene enhanced the Ad replication via several mechanisms, leading to more efficient oncolysis than the WT-Ad.
Significant cytotoxicity levels were not found in MRC-5 cells following treatment with OAd-Δ24 or WT-Ad, but the cytotoxicity levels of OAd-Δ24 in HUVEC were similar to those of the WT-Ad and higher than those of the other OAds. Previous studies demonstrated that OAd-Δ24 exhibited significantly lower cytotoxicities and replication in primary normal human cells than the WT-Ad (7, 20, 21). These contradictory results might be due to the proliferation status of the primary normal human cells. The HUVEC used in this study were proliferating. Jiang et al. demonstrated that an OAd with a 24-bp deletion in the E1A gene showed higher cytotoxicities in proliferating normal human astrocytes than in quiescent normal human astrocytes (22). Most normal human cells are not actively proliferating in vivo, suggesting that OAd-Δ24 would be less cytotoxic to normal human cells in vivo. Indeed, no severe safety concerns were reported in cancer patients receiving an OAd with a 24-bp deletion in the E1A gene in clinical studies (23, 24).
Higher levels of E1A protein production were found for OAd-Δ24, compared to WT-Ad and the other OAds (Figure 5B). Efficient E1A protein expression contributed to higher levels of tumor cell killing activity of OAd-Δ24. Jiang et al. also reported that the E1A protein level is a crucial determinant of oncolytic activity of OAds (25). Although OAd-Δ24 mediated the highest E1A protein production among the OAds and WT-Ad, OAd-Δ24 produced the E1A mRNA at levels similar to WT-Ad and OAd-smANB. These data suggest that translation of the E1A mRNA and/or stability of the E1A protein would be improved by a 24-bp deletion (bp 923-946) in the E1A gene. The proteasome pathway is involved in degradation of the E1A protein (26), but, there is no lysine codon in the 24-bp deletion region, indicating that the 24-bp deletion region does not contain ubiquitination residue(s). It remains to be clarified why the E1A protein levels were increased by a 24-bp deletion in the E1A gene.
OAd-tmANB and OAd-smANB, which contained the hTERT and survivin promoters, respectively, for the E1A expression, mediated similar levels of tumor cell killing activities and virus genome replication in the tumor cells, although the E1A mRNA and protein levels in OAd-smANB-infected tumor cells were higher than those in OAd-tmANB-treated cells at 8 h after infection. It remains unclear why OAd-tmANB showed levels of oncolytic activities and virus genome replication similar to those of OAd-smANB despite OAd-smANB having higher E1A expression than OAd-tmANB; however, Glasspool et al. reported that the hTERT promoter was activated by the E1A protein (27). The E1A gene expression gradually increased at the late phase of infection via the hTERT promoter-E1A protein positive-feedback loop, resulting in efficient oncolytic activities of OAds containing the hTERT promoter.
The oncolytic activities of OAds possessing the hTERT promoter-driven E1A expression cassette did not correlate with the hTERT mRNA levels. For example, SW620 cells expressed lower levels of hTERT mRNA than HCT116 and Panc-1 cells, but OAds possessing the hTERT promoter-driven E1A expression cassette mediated higher levels of oncolytic activities in SW620 cells than HCT116 and Panc-1 cells. A similar phenomenon was found in OAd-smANB. Although Panc-1 cells showed the highest levels of survivin mRNA among the tumor cells, Panc-1 cells were less susceptible to OAd-smANB, compared with H1299 and HepG2 cells. The expression levels of tumor cell-specific genes whose promoters are used for transcription of the E1A gene in the tumor cells are not a sole determinant of the oncolytic activities of OAds.
The tumor cell killing activities of OAd-tANB and OAd-tmANB were similar in the human tumor cells tested (Figure 3), suggesting that the three amino acid substitution in the STING binding domain of the E1A protein did not enhance the in vitro oncolytic activities of OAds in human tumor cells. Lau et al. reported that the E1A protein inhibited the cGAS-STING pathway, which is crucial for DNA virus-induced innate immunity, and that the three amino acid substitution in the STING binding domain in the E1A protein resulted in loss of the blockade of the cGAS-STING pathway (18). The STING binding domain in the E1A protein was also crucial for Rb binding. At the beginning of the study, these findings let us to hypothesize that OAd genome activated innate immunity via the cGAS-STING pathway by the substitution of three amino acids in the STING binding domain of the E1A protein, leading to higher levels of oncolytic activities in tumor cells by inducing innate immunity-mediated cytotoxicity. However, three amino acid substitution in the E1A gene did not enhance oncolytic activities of an OAd in the tumor cells in this study, probably because type I IFN expression was not detectably induced in the tumor cells following OAd infection (data not shown), indicating that the OAds did not significantly stimulate the cGAS-STING pathway in the tumor cells, at least under the experimental conditions in this study.
Although OAd-Δ24 mediated the most efficient oncolytic activities in the cultured human tumor cell lines in this study, it remains to be evaluated which oncolytic Ads exhibit the most efficient antitumor effects in vivo. Tumor cell conditions are largely different under in vitro culture condition and in vivo tumor tissues. For example, activities of the survivin and hTERT promoters are higher in the hypoxia condition, compared to those in normoxia condition (28, 29). Inflammatory cytokines, which are produced by OAd-mediated innate immunity, also activated the survivin and hTERT promoters (30, 31). Further examination is necessary for optimization of the E1A expression cassette in OAds.
In summary, we systematically evaluated the oncolytic activities and safety profiles of five OAds. The results indicated that OAd-Δ24 exhibited the most efficient tumor cell killing activities. These data provide important information for the development of superior OAds and OAd-mediated oncotherapy.
Acknowledgements
The Authors thank Toshiyoshi Fujiwara (Okayama University, Okayama, Japan) and Toshiro Shirakawa (Kobe University, Hyogo, Japan) for kindly providing experimental materials. This study was supported by grants-in-aid for Scientific Research (A) from the Ministry of Education, Culture, Sports, Science, and Technology (MEXT) of Japan.
Footnotes
↵* These Authors contributed equally to this study.
Authors’ Contributions
FS designed the experiments, analyzed the data, and wrote the article. FN and KT designed and performed the experiments, and analyzed the data. HM designed the experiments and reviewed the article.
Conflicts of Interest
The Authors declare no competing interests in relation to this study.
- Received December 31, 2020.
- Revision received January 19, 2021.
- Accepted January 21, 2021.
- Copyright © 2021 International Institute of Anticancer Research (Dr. George J. Delinasios), All rights reserved.