Abstract
The herpes simplex virus type 1 thymidine kinase suicide gene (HSV1tk) together with ganciclovir (GCV) have been successfully used for in vivo treatment of various experimental tumors, and many clinical trials using this system have been launched. With the aim to improve this therapeutic system, we compared the potential efficacy of different herpes virus derived thymidine kinases (HSV1, varicella-zoster virus, equine herpes virus type-4 and Epstein–Barr virus) as suicide genes in association with the nucleoside analogs acyclovir, ganciclovir and bromovinyldeoxyur- idine. Using various murine and human cell lines expressing these viral tk, we show that HSV1- and EHV4tk are the more efficient suicide genes for the different nucleoside analogs tested. Moreover, EHV4tk expressing murine and human cells were three- to 12-fold more sensitive to GCV than HSV1tk expressing cells. This was correlated with the presence of five-fold higher amounts of the toxic triphosphated-GCV in EHV4- versus HSV1tk expressing cells. Altogether, these experiments underline the potential advantages of the EHV4tk as a suicide gene.
Main
The herpes simplex virus type 1 thymidine kinase gene (HSV1tk) is the most widely used suicide gene, both in experimental settings and clinical trials.1 Expression of the HSV1tk gene renders tumor cells sensitive to antiviral agents like acyclovir (ACV), ganciclovir (GCV) and bromovinyl deoxyuridine (BVDU). These nucleoside analogs are efficiently converted to their monophosphate form by HSV1tk, and are then converted to triphosphate compounds by host cellular kinases. Incorporation of these metabolites into elongating DNA blocks elongation leading to cell death.2,3,4 The HSV1tk/GCV system has proved efficient for inducing the regression of transplanted tumors in various animal models,5 as well as of carcinogen-induced tumors.6 Based on these favourable results, several clinical gene therapy trials are in progress aimed at assessing safety and efficacy of this treatment for treating malignancies, and preliminary results have now been reported.7,8
Despite the effectiveness of the HSV1tk/GCV system for killing tumor cells, it remains important to try to improve it, with the aim to maximize therapeutic efficacy and/or to facilitate treatment modalities. We thus looked for new suicide gene/prodrug combinations. Biochemical and in vivo studies have demonstrated different substrate specificity of herpes tk towards antiviral and cytostatic nucleoside analogs.9,10,11 We thus compared tk from HSV1, varicella-zoster virus (VZV), equine herpes virus type-4 (EHV4) and Epstein–Barr virus (EBV) for their capacity to sensitize tumor cells to ACV, GCV and BVDU.
We generated expression vectors with each of these tk genes fused in frame with the Sh Ble gene conferring resistance to zeocin.12 This should facilitate the selection of transduced cells and the detection of the chimeric proteins by immunodetection with anti-SH polyclonal antibodies. The dual functionality of such chimeric proteins expressed by these vectors has been previously demonstrated in bacteria lacking endogenous tk activity.12 From endogenous tk-deficient murine fibroblasts (tk(−) 3T3), we generated cells expressing the different viral tk based on zeocin selection. These cells could survive when grown in selective HAT medium and could incorporate tritiated thymidine (3HT) (data not shown) demonstrating the dual functionality of tk/SH fusion proteins. These results are in agreement with previous experiments showing that HSV1tk retain its functionality after amino- or carboxy-terminal fusion to various proteins.13,14,15,16
We next analyzed these selected bulk populations for their in vitro sensitivity to nucleoside analogs. We measured inhibition of cell proliferation in comparison with parental cells and determined 50% inhibitory concentrations (IC50) (Table 1). Incorporation of 3HT in cells in the absence of drug was in the same range for the four bulks (data not shown). ACV IC50 values were approximately 15-fold lower in HSV1-, EHV4- and VZVtk expressing cells compared with the parental NIH-3T3 cells, and only four-fold lower for EBVtk expressing cells. HSV1- and EHV4tk expressing cells were 350- and 700-fold more sensitive to GCV than NIH-3T3 cells, compared with only seven-fold for VZV- or EBVtk expressing cells. Unlike ACV and GCV, IC50 values for BVDU were in the same range for all herpes virus-tk expressing cell bulks, about 1300- to 3600-fold lower than for NIH-3T3 cells. Similar results were obtained by assessing directly the drug’s toxicity in a semi-quantitative assay evaluating the cell survival in the presence of increased drug concentrations (data not shown).17 These observations were extended using the murine pancreatic tumor PANC-O2 cell line. Again, it was found that IC50 values were three-fold lower for EHV4tk than for HSV1tk expressing cell bulk, 0.55 ± 0.07 and 1.7 ± 0.5 μM, respectively (means of two independent experiments).
Among the four herpes viruses tk tested for their capacity to sensitize cells to nucleoside analogs toxicity, HSV1- and EHV4tk demonstrated the more interesting selectivity indexes. Because cell bulks represent a mixture of clones with different sensitivities, we next analyzed the drug cytotoxicity on individual clones expressing either of these two tk (Table 2). For ACV and GCV, the sensitivity of EHV4tk expressing cells was statistically more important than the one of HSV1tk expressing cells (P < 0.02 and 0.008, respectively). However, HSV1tk-expressing clones were three times more sensitive to BVDU than the EHV4tk expressing clones (Table 2). These differences in sensitivity to nucleoside analogs were not due to differences in tk expression levels. Indeed, expression of EHV4tk/ZEO and HSV1tk/ZEO proteins detected by Western blot using polyclonal antibodies against SH protein was similar in the four EHV4tk (+) and four HSV1tk (+) clones tested (number 1 to 4 on Table 2 for both fusion proteins) (data not shown).
Since nucleoside analog’s metabolism might be different in cells of different species we thus analyzed the efficiency of the system for killing human cells. Human tk(−) osteosarcoma 143B cell bulks expressing HSV1- or EHV4tk were thus generated. Herpes virus tk protein expression levels were similar for each cell bulk as assessed by dot-blot using polyclonal antibodies against SH protein (Figure 1b). We then, studies cell sensitivity to ACV, GCV and BVDU (Figure 1a). The toxicity of ACV was similar in HSV1- and EHV4tk-expressing cells. It is worth noting that EHV4tk expressing cells were approximately 12-fold more sensitive to GCV than HSV1tk expressing cells. On the contrary the toxicity of BVDU was greater in HSV1tk than in EHV4tk expressing cells (Figure 1a). These observations were extended to an additional tumor cell line, human colonic carcinoma HCT116 cells. EHV4- and HSV1tk expressing cells were 485- and 170-fold more sensitive to GCV than parental HCT116 cells (mean of IC50 values for GCV was 53.4 ± 15 μM). IC50 values for GCV were three-fold lower for EHV4tk than for HSV1tk expressing cells, 0.11 ± 0.02 and 0.32 ± 0.08 μM, respectively (means ± s.d. for three independent experiments).
Cytotoxic activity of ACV, GCV and BVDU against human tk(−) 143B cells transfected with HSV1- or EHV4tk. tk(−) 143B cells were grown in DMEM medium supplemented with 1% L-glutamine, 1% penicillin, streptomycin and neomycin mixture and 10% fetal calf serum (GibcoBRL). Cells were transfected using a calcium phosphate precipitation kit and cell bulks were selected with 50 μg/ml zeocin in grown medium. (a) Drug toxicity was assessed by measuring inhibition of cell proliferation as described in Table 1. Results were presented as mean of IC50 values ± s.d. The means, determined on three independent experiments, are indicated on the top of each histogram. (b) Dot blot analysis of tk expression in transfected and non-transfected tk(−) 143B cells. Cells were incubated in lysis buffer at the final cellular concentration of 20 × 106 cell/ml. 5 μl of successive dilutions of cellular extracts were dropped on nitrocellulose membrane. Zeo-fusion proteins were revealed as described in Cazaux et al.12 There was no detectable band corresponding to fusion protein in tk(−) 143B cell extracts.
In order to understand why GCV is more toxic in EHV4tk compared with HSV1tk expressing cells, we analyzed the GCV phosphorylation in tk(−) 143B and HSV1- or EHV4tk expressing cells. While there was no detectable phosphorylation of thymidine in tk(−) 143B cells (data not shown), thymidine was similarly phosphorylated in both TK expressing cells (Figure 2). For GCV, the proportions of the different phosphorylated forms were the same in both tk cells, while the total amount of intracellular GCV was five-fold higher for EHV4- than for HSV1tk expressing cells. Such a difference might be explained by the uptake of nucleoside and their analogs that has been reported to be correlated with tk expression. Indeed, the uptake of thymidine into tk(−) Escherichia coli is proportional to the amount of tk activity expressed from the heterologous HSV1tk gene.18 Furthermore, mammalian HSV1tk expressing cells showed higher GCV uptake and phosphorylation than control cells.19 Since tk expression was similar in both tk(+) cells, the greater amount of intracellular GCV in EHV4tk expressing cells could indicate a greater EHV4tk activity toward GCV compared with HSV1tk activity. Detailed kinetic studies of EHV4tk and HSV1tk enzymatic activities should confirm these observations.
Thymidine and GCV phosphorylation in transfected and non-transfected tk(−) 143B cells. 3 × 106 cells were incubated for 7 h with 1 μM of 8-3H-GCV (3 Ci/mmol, Isotopchim, France) or methyl-3H-thymidine (5 Ci/mmol, Amersham, UK) in a final volume of 5 ml of medium. After trypsin treatment and washes with PBS, the nucleotides were extracted with 1 ml of ice cold methanol 60% over the night at −20°C. The samples were analyzed by Fast Protein Liquid Chromatography (FPLC, Pharmacia-LKB Instruments, Uppsala, Sweden) on an anion-exchange column. Radioactivity was counted at the column exit with a radiamatic Flo-one Beta A-500 apparatus (Packard, Meriden, CT, USA), and peaks were analyzed with Flo-one/Data software.25 The percentages of thy-MP, thy-DP and thy-TP were respectively 29.3, 27.5 and 39.2% in HSV1-tk expressing cells and 36.8, 27.5 and 28.8% in EHV4-tk expressing cells. The percentages of GCV-MP, GCV-DP and GCV-TP were respectively 15.5, 36.5 and 45% in HSV1-tk expressing cells and 23.9, 29.3 and 45% in EHV4-tk.
In conclusion, significant differences in the activity of the different herpes tk were observed. EBV- and VZVtk did not demonstrate any advantages over HSV1tk whatever the nucleoside analog used. Two recent studies investigated VZVtk as a suicide gene in human breast cancer and osteosarcoma cells.20,21 In these articles IC50 values for BVDU ranged from 0.06 to 0.6 μM, quite similar to our values, but no comparison between VZV- and HSV1tk was shown. It is worth nothing that EHV4tk expressing cells were always more sensitive to GCV than HSV1tk expressing ones whether as bulks that provide an estimate of the overall enzyme activity in a polyclonal population, or as clones. This enhanced toxicity was not due to a difference in tk expression levels as assessed by Western blot. This is also further indicated by the observation of an opposite effect with BVDU that is more toxic for HSV1tk expressing cells, demonstrating a qualitative rather than a quantitative difference between the activity of these enzymes. Depending on the cell lines, EHV4tk was three- to 12-fold more potent than HSV1tk. Such a difference appears significant when contemplating the clinical use of suicide genes. Indeed the treatment of experimental tumors in mice or rats demonstrate that the GCV dosage is often critical for obtaining efficacy, notably for brain tumors. When administered at 10 mg/kg twice a day, the usual dosage in humans, the GCV plasmatic concentration range between 0.44 to 2.2 μg/ml, and the cerebral spinal fluid (CSF) concentrations are approximately three-fold lower.22,23 It is therefore possible that CSF GCV concentrations are sub-optimal for the treatment of brain tumors transduced with HSV1tk. In any case, a better efficiency of the enzyme should lead to a better treatment efficiency and/or a decrease of the GCV dose used. This latter point is also significant since GCV toxicity in human appears cumulative depending on the overall amounts administered.24 Altogether, our results underline the potential advantages of the EHV4tk as a suicide gene.
References
Anonymous . Human gene marker/therapy clinical protocols (complete updated listings) Hum Gene Ther 1998 9: 2805–2852
Balzarini J, De Clercq E . Inhibitory effects of (E)-5-(2-bromovinyl)-2′-deoxyuridine (BVDU) and related compounds on herpes simplex virus (HSV)-infected cells and HSV thymidine kinase gene-transformed cells Meth Find Exp Clin Pharmacol 1989 11: 379–389
Furman PA et al. Inhibition by acyclovir of cell growth and DNA synthesis of cells biochemically transformed with herpesvirus genetic information Virology 1980 102: 420–430
St Clair MH, Lambre CU, Furman PA . Inhibition by ganciclovir of cell growth and DNA synthesis of cells biochemically transformed with herpes virus genetic information Antimicrob Agents Chemother 1987 31: 844–849
Freeman SM et al. In situ use of suicide genes for cancer therapy Semin Oncol 1996 23: 31–45
Wei MX et al. Suicide gene therapy of chemically induced mammary tumor in rat: efficacy and distant bystander effect Cancer Res 1998 58: 3529–3532
Ram Z et al. Therapy of malignant brain tumors by intratumoral implantation of retroviral vector-producing cells Nature Med 1997 3: 1354–1361
Klatzmann D et al. A phase I/II study of herpes simplex virus type 1 thymidine kinase ‘suicide’ gene therapy for recurrent glioblastoma. Study Group on Gene Therapy for Glioblastoma Hum Gene Ther 1998 9: 2595–2604
Machida H . Comparison of susceptibilities of varicella-zoster virus and herpes simplex viruses to nucleoside analogs Antimicrob Agents Chemother 1986 29: 524–526
Shigeta S et al. Comparative efficacy of antiherpes drugs against various strains of varicella-zoster virus J Infect Dis 1983 147: 576–584
Tung PP, Summers WC . Substrate specificity of Epstein–Barr virus thymidine kinase Antimicrob Agents Chemother 1994 38: 2175–2179
Cazaux C et al. Phosphorylation and cytotoxicity of therapeutic nucleoside analogues: a comparison of alpha and gamma herpesvirus thymidine kinase suicide genes Cancer Gene Ther 1998 5: 83–91
Beck C et al. The thymidine kinase/ganciclovir-mediated ‘suicide’ effect is variable in different tumor cells Hum Gene Ther 1995 6: 1525–1530
Loimas S, Wahlfors J, Janne J . Herpes simplex virus thymidine kinase-green fluorescent protein fusion gene: new tool for gene transfer studies and gene therapy Biotechniques 1998 24: 614–618
Marini FR et al. In vivo marking of spontaneous or vaccine-induced fibrosarcomas in the domestic house cat, using an adenoviral vector containing a bifunctional fusion protein, GAL-TEK Hum Gene Ther 1995 6: 1215–1223
Schwartz F et al. A dominant positive and negative selectable gene for use in mammalian cells Proc Natl Acad Sci USA 1991 88: 10416–10420
Brady HJ, Miles CG, Pennington DJ, Dzierzak EA . Specific ablation of human immunodeficiency virus Tat-expressing cells by conditionally toxic retroviruses Proc Natl Acad Sci USA 1994 91: 365–369
Dube DK, Horwitz MS, Loeb LA . The association of thymidine kinase activity and thymidine transport in Escherichia coli Gene 1991 99: 25–29
Haberkorn U et al. Ganciclovir uptake in human mammary carcinoma cells expressing herpes simplex virus thymidine kinase Nucl Med Biol 1998 25: 367–373
Degreve B et al. Varicella-zoster virus thymidine kinase gene and antiherpetic pyrimidine nucleoside analogues in a combined gene/chemotherapy treatment for cancer Gene Therapy 1997 4: 1107–1114
Grignet-Debrus C, Calberg-Bacq CM . Potential of Varicella zoster virus thymidine kinase as a suicide gene in breast cancer cells Gene Therapy 1997 4: 560–569
Oldfield EH et al. Gene therapy for the treatment of brain tumors using intra-tumoral transduction with the thymidine kinase gene and intravenous ganciclovir Hum Gene Ther 1993 4: 39–69
Faulds D, Heel RC . Ganciclovir. A review of its antiviral activity, pharmacokinetic properties and therapeutic efficacy in cytomegalovirus infections Drugs 1990 39: 597–638
Buhles WC . Clinical safety of intravenous ganciclovir In:Spector SA (ed) . Ganciclovir Therapy for Cytomegelovirus Infection Marcel Dekker: New York 1991 pp 31–69
Guettari N, Loubiere L, Brisson E, Klatzmann D . Use of herpes simplex virus thymidine kinase to improve the antiviral activity of zidovudine Virology 1997 235: 398–405
Acknowledgements
This work was supported in part by the Agence Nationale de Recherche sur le SIDA, the Université Pierre et Marie Curie, the Centre National de la Recherche Scientifique and the Assistance Publique – Hôpitaux de Paris. Laurence Loubière was a fellow of the Association pour la Recherche sur le Cancer.
Author information
Affiliations
Rights and permissions
About this article
Cite this article
Loubière, L., Tiraby, M., Cazaux, C. et al. The equine herpes virus 4 thymidine kinase is a better suicide gene than the human herpes virus 1 thymidine kinase. Gene Ther 6, 1638–1642 (1999). https://doi.org/10.1038/sj.gt.3300993
Received:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1038/sj.gt.3300993
Keywords
- nucleoside analogs
- cancer
- fusion proteins
Further reading
-
Replicative retroviral vectors for cancer gene therapy
Cancer Gene Therapy (2003)
-
Selective killing of cancer cells based on translational control of a suicide gene
Cancer Gene Therapy (2002)
-
Herpes simplex virus thymidine kinase/ganciclovir–induced cell death is enhanced by co-expression of caspase-3 in ovarian carcinoma cells
Cancer Gene Therapy (2001)
