- Email: [email protected]
Expression of a fusogenic membrane glycoprotein by an oncolytic herpes simplex virus potentiates the viral antitumor effect
ARTICLE
doi:10.1016/S1525-0016(03)00092-3
Expression of a Fusogenic Membrane Glycoprotein by an Oncolytic Herpes Simplex Virus Potentiates the Viral Antitumor Effect Xinping Fu,1,* Lihua Tao,1,* Aiwu Jin,1 Richard Vile,2 Malcolm K. Brenner,1 and Xiaoliu Zhang1,† 1
Center for Cell and Gene Therapy, Baylor College of Medicine, Houston, Texas 77030, USA 2 Molecular Medicine Program, Mayo Clinic, Rochester, Minnesota 55905, USA *Current address: The Fifth Affiliated Hospital of Zhongshan University, Zhuhai, China.
†
To whom correspondence and reprint requests should be addressed. Fax: (713) 798-1230. E-mail: [email protected].
Oncolytic viruses have shown considerable promise in the treatment of solid tumors, but their potency must be improved if their full clinical potential is to be realized. We inserted the gene encoding a truncated form of the gibbon ape leukemia virus envelope fusogenic membrane glycoprotein (GALV.fus) into an oncolytic herpes simplex virus, using an enforced ligation procedure. Subsequent in vitro and in vivo studies showed that expression of GALV.fus in the context of an oncolytic virus significantly enhances the antitumor effect of the virus. Furthermore, by controlling GALV.fus expression through a strict late viral promoter, whose activity depends on the initiation of viral DNA replication, we were able to express this glycoprotein in tumor cells but not in normal nondividing cells. It will be of interest to confirm whether functional expression of a strong fusogenic gene by an oncolytic herpes simplex virus enhances viral antitumor activity without increasing its toxicity. Key Words: oncolytic, herpes simplex virus, late viral promoter, fusogenic membrane glycoprotein, liver cancer
INTRODUCTION Replication-selective oncolytic viruses are potentially useful antitumor agents. Their genetic modifications enable them to replicate preferentially within tumor cells, while being restricted in their ability to replicate in normal cells. The principal antitumor mechanism used by these viruses is a direct cytopathic effect, produced as the viruses propagate and spread from initially infected tumor cells to surrounding tumor cells, achieving a progressively larger volume of distribution and enhanced tumor cell killing. Oncolytic herpes simplex viruses (HSVs) were initially designed and constructed for the treatment of brain tumors [1,2]. Subsequently, they were found to have activity against a variety of other human solid tumors, including breast cancer [3], prostate cancer [4], lung cancer [5], ovarian cancer [6], and colon and liver cancers [7,8]. The safety of oncolytic HSVs has been extensively tested in studies in mice [9] and in primates [10], which showed that these agents are safe for in vivo administration. Despite these encouraging preclinical studies, results
748
from early clinical trials suggested that current oncolytic viruses, although safe, may have only limited antitumor activity on their own [11–14]. One of the main reasons for this suboptimal oncolytic efficacy is that the viral gene deletions conferring tumor selectivity also reduce the replicative potential of the virus in tumors. For example, complete elimination of endogenous ␥34.5 gene function from HSV, one of the common approaches to construction of an oncolytic HSV, significantly reduces viral replication and compromises the ability of the virus to spread within targeted tumors [15]. Thus, strategies are needed to enhance the potency of oncolytic viruses. Recently, Bateman et al. [16] reported that envelope fusogenic membrane glycoproteins (FMGs) expressed by certain viruses are extremely cytotoxic to tumor cells. An example of one such FMG is a C-terminal truncated form of the gibbon ape leukemia virus envelope glycoprotein (GALV.fus) [16,17] that lacks the 16-amino-acid R-peptide of the wild-type protein, which normally serves to restrict fusion of the envelope until it is cleaved during viral maturation [18]. This alteration renders the protein con-
MOLECULAR THERAPY Vol. 7, No. 6, June 2003 Copyright © The American Society of Gene Therapy 1525-0016/03 $30.00
doi:10.1016/S1525-0016(03)00092-3
ARTICLE
stitutively highly fusogenic with human cells that express the Pit-1 receptor for GALV. Transduction of the GALV.fus gene into a range of human tumor cells efficiently kills the cells through a process of syncytial formation [19]. It also destroys neighboring cells through a “bystander” effect at least 10-fold greater than that observed with HSV-Tk or cytosine deaminase-mediated therapy [19,20]. The full therapeutic potential of GALV.fus will probably not be realized until means are devised for its efficient delivery and specific expression in tumor cells. We show here that these aims can be achieved by expressing the truncated glycoprotein in the context of an oncolytic HSV. In particular, expressing GALV.fus under control of a strictly late viral promoter, whose activity depends on the ability of the oncolytic virus to replicate in tumor cells, minimized the expression of this glycoprotein in normal tissues.
RESULTS Construction and Initial Characterization of the New Fusogenic Viruses We used an enforced ligation strategy to insert either GALV.fus or the enhanced green fluorescent protein (EGFP) gene into a bacterial artificial chromosome (BAC)based HSV construct that was deleted for both the ␥34.5 gene and the HSV packaging signal (fHSV-delta-pac). The genes encoding GALV.fus or EGFP were driven either by the CMV immediate early promoter or by a strict late viral promoter. We believe that directing GALV.fus expression with a strict late viral promoter, in the context of an oncolytic HSV, may selectively express the fusogenic glycoprotein in tumor tissue. This is because transcription of the HSV genome is a tightly regulated process in which early and late phases of gene expression are separated by viral DNA replication [21]. In particular, expression of strict late transcripts (e.g., those of UL38) depends rigorously on the initiation of viral DNA replication. The promoters of these strict late genes, such as UL38p, which drives UL38 expression, would be active only in tumor cells in which the oncolytic HSV can fully replicate; in nondividing or postmitotic normal cells, they would be silent since viral replication would be limited. We constructed a total of three viruses, including Baco-1 (containing EGFP driven by the CMV promoter), Synco-1 (containing GALV.fus driven by the CMV promoter), and Synco-2 (containing GALV.fus driven by UL38p). We analyzed the genomes of these newly generated viruses by enzyme digestion to confirm that their structure had not undergone significant rearrangement. DNA extracted from supernatant virion particles (Baco-1, Synco-1, and Synco-2) or from bacterial cultures (fHSVdelta-pac) was digested with BamHI, which cuts multiple times within the viral genome. The result of gel electrophoresis showed that the restriction band patterns among the viruses and the original fHSV-delta-pac construct are
MOLECULAR THERAPY Vol. 7, No. 6, June 2003 Copyright © The American Society of Gene Therapy
FIG. 1. Characterization of virion DNA by BamHI digestion. Virion DNA or fHSV-delta-pac DNA extracted from bacterial culture was digested with BamHI and run on a 0.8% gel together with a 1 kb DNA ladder as marker (M) at a constant voltage for 6 h.
similar (Fig. 1), which is consistent with published data from Stavropoulos and Strathdee [22]. The small differences in the band patterns among them may reflect the circular form of fHSV-delta-pac and the different extra DNA sequences among the recombinant viruses. We evaluated the growth properties of the newly created viruses in both Vero cells and the Hep 3B human liver cancer cell line. We infected cells with the viruses at different multiplicities of infection (m.o.i.), harvested the viruses at different times, and titrated them by plaque assay. As shown in Fig. 2, Synco-1 and Synco-2 had virtually identical growth curves and maximal viral titers. Although the titers for the fusogenic viruses were slightly higher than those of Baco-1 at later stages of infection (i.e., 24 and 48 h after infection), these differences were not statistically significant. Phenotypic Characterization of the Fusogenic Viruses To characterize the phenotypes of the fusogenic oncolytic HSVs, we infected human cancer cells of different tissue origins with the viruses at a low m.o.i. At 36 h after infection, a clear syncytial phenotype was observed for all tumor cells infected by either Synco-1 or Synco-2 (Fig. 3a); the extent of tumor cell fusion did not differ appreciably between Synco-1 and Synco-2. By contrast, the same types of tumor cells infected with Baco-1 lacked this phenotype. To determine if GALV.fus-mediated syncytial formation after Synco-2 infection truly depends on viral DNA replication, we infected human tumor cells with either Synco-1 or Synco-2 in the presence or absence of acyclovir
749
ARTICLE
doi:10.1016/S1525-0016(03)00092-3
(ACV), which inhibits HSV DNA replication. Syncytial formation by Synco-1 in Hep 3B cells was largely unaffected by the presence of ACV (Fig. 3a), while cell membrane fusion by Synco-2 in the same cells was completely blocked by this drug. We also tested directly whether Synco-2 loses its ability to cause syncytial formation in normal nondividing human cells, by infecting primary human fibroblasts in either a quiescent or a cycling state. As shown in Fig. 3b, infection with either Synco-1 or Synco-2 caused syncytial formation in these normal human cells when they were in cycle; the syncytia formed by Synco-1 infection were noticeably larger than those produced by Synco-2 (3.6 ⫾ 0.4 vs 1.8 ⫾ 0.2 nuclei per syncytium). Synco-1-mediated cell fusion was only marginally affected in cells whose cycling was either slowed by serum starvation (2.3 ⫾ 0.2 nuclei per syncytium) or completely arrested with lovastatin (2.0 ⫾ 0.3 nuclei per syncytium). Synco-2-mediated cell fusion, on the other hand, was absent in cells whose cycling was decreased or arrested. The infected cells appeared essentially identical to the uninfected cells, except that after extended incubation, cytopathic effect resembling cell fusion did become evident in the serum-starved cells by day 4, while the cells treated with lovastatin began to die from effects of the drug by day 5 (data not shown). These results may imply that syncytial formation induced by the UL38 promoter-controlled GALV.fus expression in the context of an oncolytic HSV is cell-cycle-dependent. FIG. 3. Phenotypic characterization of Synco-1 and Synco-2 in tumor cells and in cycling or arrested embryonic fibroblasts. (a) Phenotypic characterization of viruses in tumor cells. Tumors of different tissue origin were infected with each of the viruses at 0.01 pfu/cell, with or without acyclovir (ACV) in the culture medium. Photos were taken at 48 h after initial virus infection. The photos of cells in the presence of ACV were taken 72 h after infection. (b) Phenotypic characterization of viruses in embryonic fibroblasts of cycling or arrested status. Fully cycling, serum-starved, or arrested (with lovastatin) human embryonic fibroblasts were infected with either Synco-1 or Synco-2 at 0.01 pfu/cell or left uninfected. The photos were taken 48 h after infection. Original magnification, ⫻200.
FIG. 2. Growth characterization of oncolytic HSVs. Vero cells (a) and Hep 3B cells (b) were seeded into 24-well plates in triplicate and infected with the viruses at 0.1 or 1 pfu/cell. The viruses were harvested at 0-, 12-, 24-, and 48-h intervals after infection and were titrated by plaque assay.
750
Comparison of Tumor Cell Killing Next, we determined if GALV.fus-mediated syncytial formation in the context of oncolytic HSV replication would increase tumor cell killing. We initially compared the viability of cells after they were infected with either Baco-1 or Synco-1 at a low m.o.i. (0.1 and 0.01 plaqueforming units (pfu)/cell), so that both the inherent cytotoxicity of the input virus and the ability of the virus to replicate and spread in these cells could be assessed. The cytotoxic effect of the virus infection on tumor cells was quantified by calculating the percentage of cells that survived at different time points after infection. In all tumor cell lines, Synco-1 killed tumor cells more effectively than Baco-1 at both time points and both viral doses (Figs. 4a and 4b, P ⬍ 0.01 for all comparisons). We also compared
MOLECULAR THERAPY Vol. 7, No. 6, June 2003 Copyright © The American Society of Gene Therapy
doi:10.1016/S1525-0016(03)00092-3
ARTICLE
the Hep 3B cell killing ability of Synco-1 and Synco-2 directly with that of Fu-10, a fusogenic oncolytic HSV selected from G207 through random mutagenesis [23]. The results showed that while all versions of the fusogenic oncolytic HSVs had a significantly greater ability to destroy tumor cells than did Baco-1, their individual oncolytic activities did not differ significantly against this particular tumor cell line (Fig. 4c).
FIG. 4. Increased tumor cell killing by fusogenic oncolytic HSVs. (a) Tumor cells of different tissue origin were infected with either Baco-1 or Synco-1 at 0.1 pfu/cell. (b) The same tumor cells were infected with either Baco-1 or Synco-1 at 0.01 pfu/cell. (c) Hep 3B cells were infected with the indicated viruses at either 0.1 or 0.01 pfu/cell. The percentage of surviving cells was calculated by dividing the number of live cells from the infected wells by the number of live cells in the well that was not infected. P ⬍ 0.01 for all comparisons between Synco-1/Synco-2 and Baco-1. No statistically significant difference between Synco-1 and Synco-2 was found.
MOLECULAR THERAPY Vol. 7, No. 6, June 2003 Copyright © The American Society of Gene Therapy
Enhanced Antitumor Effect in Xenografted Human Tumors To determine if the syncytial phenotype and enhanced ex vivo tumor cell killing shown by Synco-1 would translate into an enhanced antitumor effect in vivo, we injected the virus at 1 ⫻ 107 pfu directly into established xenografts of Hep 3B carcinomas with diameters of 5 to 8 mm. Tumor sizes were measured weekly for 4 weeks. Compared with the PBS controls, a single injection of either Baco-1 or Synco-1 had an immediate effect on tumor growth (Fig. 5a). Within 1 week of virus injection, the tumors in mice treated with these constructs were significantly smaller than tumors injected with PBS (P ⬍ 0.001). From week 2 to week 4, however, Synco-1 produced significantly greater antitumor effects than did Baco-1 (P ⬍ 0.001). Half of the animals (5 of 10) were tumor free by week 3 after Synco-1 administration, while the other half had significantly reduced tumor sizes. By contrast, only 1 mouse in the group injected with Baco-1 was tumor free. In the other 9 mice, tumors that had shrunk initially began to regrow by week 3 after virus injection. By week 4 —the end point of the experiment—the mice in the Synco-1injected group either were still tumor free (5/10) or had only small tumors, while the average tumor size in Baco1-injected mice had substantially increased from the pretreatment measurement (Fig. 5a). We then compared directly the antitumor potencies of Synco-1 and Synco-2 by injecting the viruses into xenografted human liver tumors in nude mice, as described earlier. The results showed that both constructs have significantly better antitumor effects than Baco-1, starting from week 2 after virus administration (P ⬍ 0.01). There was no significant difference in the tumor growth ratio between Synco-1- and Synco-2-treated mice (Fig. 5b). These results demonstrate that Synco-2 has antitumor potency equivalent to that of Synco-1, even though its GALV.fus expression is driven by a conditional viral promoter. To determine if syncytial formation is indeed part of the oncolytic process of Synco-1- and Synco-2-mediated antitumor therapy, we injected established Hep 3B tumors with oncolytic HSVs or PBS and excised the tumors 5 days later. Examination of stained tumor sections revealed that syncytia comprising a variable number of cell nuclei were frequently encountered across the tissue section of tumors injected with either Synco-1 or Synco-2
751
ARTICLE
doi:10.1016/S1525-0016(03)00092-3
FIG. 5. Enhanced oncolytic potency of fusogenic oncolytic HSVs in human liver cancer xenografts. (a) Enhanced antitumor activity of Synco-1. Treatment groups include Synco-1, Baco-1, or PBS. (b) Direct comparison of antitumor activities of Synco-1 and Synco-2. Treatment groups include Baco-1, Synco-1, and Synco-2. The tumor growth ratio was determined by dividing the tumor volume measured on the indicated week after virus injection by the tumor volume before treatment. The values represent means ⫾ standard deviation (n ⫽ 10 mice per group).
virus (Fig. 6). Such syncytia were not seen in tissue sections of tumors injected with either Baco-1 or PBS.
DISCUSSION Earlier attempts to clone the GALV.fus gene into an oncolytic adenovirus were not successful [20], for two possible reasons. First, the extensive and rapid syncytial formation after GALV.fus gene transfection may have interfered with the subsequent homologous recombination needed for the generation of recombinant virus. Sec-
FIG. 6. Virus-induced syncytial formation in tumor tissues. The syncytia in Synco-1- or Synco-2-infected tumor tissues are indicated by arrows (original magnification, ⫻400).
752
ond, expression of GALV.fus in the context of an adenovirus might have interfered with the infectious process. Unlike adenovirus, which is nonenveloped, HSV is an enveloped virus whose infection cycle naturally involves membrane fusion. The successful insertion of GALV.fus into an oncolytic HSV genome and subsequent demonstration of their syncytial phenotype indicate that HSV can withstand the membrane-fusion effect from genes such as GALV.fus and continue to grow. The virus construction strategy of direct in vitro ligation followed by transfection into Vero cells, which does not require homologous recombination, may have also contributed to the successful construction of the fusogenic viruses. In our early studies we showed that a syncytial mutant (Fu-10) selected from the well-characterized oncolytic HSV G207 through random mutagenesis has a significantly greater antitumor capacity than does the parental virus on xenografted lung metastatic breast cancer [23]. Here we demonstrate that expression of GALV.fus by a conditionally replicating HSV also leads to strong syncytial formation in tumor cells of different tissue origins, a property translating into a significantly increased tumor cell killing in vitro. Our in vivo data on xenografted human liver tumor demonstrate that fusogenic oncolytic HSVs constructed by this strategy have significantly elevated antitumor potency. Although direct comparison between Fu-10 and Synco-2 showed that they had similar antitumor potencies on established solid tumors of different tissue origins, including prostate cancer, liver cancer, and glioblastoma (X. Fu and X. Zhang, unpublished data), we believe that fusogenic oncolytic HSVs generated by incorporation of FMGs from viruses other than HSV may potentially be more advantageous than Fu-10. This is because syncytial formation of FMGs relies on the initial
MOLECULAR THERAPY Vol. 7, No. 6, June 2003 Copyright © The American Society of Gene Therapy
doi:10.1016/S1525-0016(03)00092-3
binding of the fusogenic glycoproteins with their receptors on the target cells. Unlike xenografted tumors established from cultured tumor cells that are relatively homogeneous, cancer cells in naturally occurring tumors may be heterogeneous in their sensitivity to HSV infection and/or syncytial formation due to variation of receptors on the cell surface. Insertion of a FMG from another virus that uses a receptor totally different from that of the oncolytic virus (Pit-1 vs heparan sulfate in the case of GALV FMG in oncolytic HSV) should significantly reduce the frequency of therapy-resistant tumor cells. This is because tumor cells that become resistant to oncolytic virus infection/replication may still be sensitive to the FMG-mediated syncytial formation and vice versa. Although expression of GALV.fus can add to the antitumor effects of oncolytic HSV, uncontrolled expression of the gene, even in this context, raises safety concerns. To minimize this potential problem, we used a strict late viral promoter to direct the expression of the GALV.fus gene in the virus. Our data show that this strategy provides strong tissue-selective expression of the gene. The UL38 promoter appears to have several advantages over currently available tissue-specific promoters. First, it is probably substantially more active than most tissue-specific promoters. Direct comparison of the UL38 promoter with the CMV immediate early promoter (one of the strongest mammalian promoters) showed that the former has only slightly lower activity than the latter (X. Fu and X. Zhang, unpublished data), whereas the activity of most tissue-specific promoters is 10 to 100 times lower than that of the CMV promoter. Second, activation of the UL38 promoter upon the initiation of lytic HSV infection in the tumor cells may synchronize the actions of these two antitumor mechanisms. Finally, the demonstration that GALV.fus-mediated fusion is essentially ablated when the cell cycle of an human embryonic fibroblast is arrested suggests that Synco-2 may be no more toxic to normal cells than the first generation of oncolytic HSVs, whose safety record has been established in both preclinical studies and phase I clinical trails. Due to the species specificity of GALV.fus, we could not test the safety of Synco-2 in our rodent models; however, confirmation of its safety in nonhuman primates (which are sensitive to GALV.fusmediated cell membrane fusion) will permit clinical exploration of this new generation of oncolytic HSVs.
MATERIALS
AND
METHODS
Cell lines. African green monkey kidney (Vero) cells, human embryonic fibroblasts (HF 333.We), human tumor cell lines U-87 MG (glioblastoma) and Hep 3B (hepatocellular carcinoma), and the human prostate cancer line DU 145 were obtained from American Tissue Culture Collection (Rockville, MD). All cells were cultured with DMEM containing 10% fetal bovine serum (FBS). Generation of fusogenic oncolytic HSVs. All oncolytic HSVs were derived from fHSV-delta-pac, a BAC-based construct that contains a mutated HSV genome, in which the diploid gene encoding ␥34.5 and both copies of the
MOLECULAR THERAPY Vol. 7, No. 6, June 2003 Copyright © The American Society of Gene Therapy
ARTICLE
HSV packaging signal (pac) have been deleted [24]. Infectious HSV cannot be generated from this construct unless an intact HSV pac is provided in cis. To insert foreign genes and to generate infectious viruses from fHSVdelta-pac, we used an enforced ligation strategy. Initially, three gene cassettes (GALV.fus driven by the CMV promoter, GALV.fus driven by the UL38 promoter, and the green fluorescent protein gene driven by the CMV promoter) were cloned into pSZ-pac so that each was linked with an HSV sequence containing HSV pac. The EGFP gene cassette was cut from pEGFP-N1 (Clontech, Palo Alto, CA) with AseI and AflIII. The GALV.fus gene was cut from pCR3.1-GALV with NheI and NotI and linked with the CMV promoter in pcDNA3. The promoter and enhancer region (DAS) of UL38, a well-defined 145-bp DNA sequence covering the 84324 – 84469 nucleotide region of the HSV genome [25], was amplified from HSV-1 DNA with the following pair of primers: forward 5⬘-GTGGGTTGCGGACTTTCTGC-3⬘ and reverse 5⬘-ACACTCACGCAAGGCGGAAC-3⬘. The GALV.fus or EGFP gene cassette that was linked with the HSV pac was then cut from the plasmids and directly ligated into the unique PacI site of fHSV-deltapac. The DNA ligation mixture was transfected into Vero cells using Lipofectamine (Gibco BRL, Grand Island, NY) and incubated for 3–5 days to permit generation of infectious virus. The viruses were subsequently plaque purified and were designated Baco-1 (containing the DNA fragment of EGFP and pac), Synco-1 (containing the DNA fragment of CMV-GALV. fus and pac), and Synco-2 (containing UL38p-GALV.fus and pac). Virus stocks were prepared by infecting Vero cells with the viruses at 0.01 pfu per cell, harvested after 2 days, and stored at ⫺80°C. Viral structure and growth characterization. Virion DNA was prepared from cell culture supernatants as described [26]. Briefly, 48 h after viral infection (at 0.01 pfu/cell), virion particles were pelleted by centrifuging the supernatants at 27,000g for 3 h. Viral pellets were resuspended in TE (10 mM Tris, 1 mM EDTA, pH 8.0) and incubated with SDS and proteinase K for 4 h at final concentrations of 0.1% and 50 g/ml, respectively. Virion DNA was then gently phenol-extracted three times before the DNA was precipitated with ethanol. For viral growth characterization, Vero cells seeded in triplicate into 48-well plates were infected with the viruses at 0.1 or 1 pfu/cell for 1 h. Cells were washed once with PBS to remove unadsorbed and uninternalized viruses before fresh medium was added. Cells were harvested at 0, 12, 24, and 48 h after infection. Viruses were released by repeated freezing and thawing and sonication. Virus titers were determined on Vero cells by a plaque assay. Phenotypic characterization of fusogenic oncolytic HSVs. Cells were seeded into six-well plates and infected the following day with each virus at a dose ranging from 0.1 to 0.0001 pfu/cell. Cells were cultured in a maintenance medium (containing 1% FBS) and were left for up to 2 days to allow the fusion pattern and plaques to develop. To block HSV DNA replication, we added ACV to the culture medium at a final concentration of 100 M. To slow the division or arrest of embryonic fibroblasts, we either starved cells for FBS for 48 h or starved them for FBS during incubation with 20 M lovastatin for 24 h, before the cells were infected with virus. Lovastatin is a chemical that induces cell-cycle arrest but does not interfere with HSV replication [27]. To quantify the size of syncytia induced by virus infection, the infected cell monolayer was stained with a solution containing 0.1% crystal violet (Sigma, St. Louis, MO) and 20% ethanol for 30 min at room temperature. The cells were then thoroughly rinsed with water. On average, 10 syncytia were randomly selected from each sample and the nuclei in each syncytium were enumerated. The result was given as mean ⫾ deviation. In vitro cell killing assay. Each of the three human tumor cell lines was seeded into 48-well plates and infected with each virus at 0.1 and 0.01 pfu/cell or left uninfected. Cells were harvested 24 and 48 h later by trypsinization. The number of surviving cells was counted on a hemacytometer following trypan blue staining. The percentage of surviving cells was calculated by dividing the number of viable cells from the infected well by the number of cells from the well that was left uninfected. The experiments were done in triplicate and the mean numbers were used for the final calculation.
753
ARTICLE
Animal experiments. Six-week-old male Hsd athymic (nu/nu) mice were purchased from Harlan (Indianapolis, IN). Hep 3B cells were cultured under standard conditions and were harvested in log phase with 0.05% trypsin–EDTA. The cells were washed twice with serum-free medium before they were resuspended in PBS at a concentration of 5 ⫻ 107 cells/ml. A total of 5 ⫻ 106 cells (in a l00-l suspension) were subcutaneously injected into the right flank of each mouse. When the tumors reached approximately 5– 8 mm in diameter, they were injected with virus (1 ⫻ 107 pfu in a 100-l volume); control tumors received an injection of PBS only. Tumor size was measured weekly for 4 weeks, and tumor volume was calculated by the formula tumor volume (mm3) ⫽ 0.5 ⫻ [length (mm)] ⫻ [width (mm)]2. For pathological examination of the tumor tissues, mice were sacrificed 5 days after virus or PBS injection. The tumors were removed, fixed, and stained with hematoxylin and eosin.
doi:10.1016/S1525-0016(03)00092-3
9.
10.
11.
12.
13. 14.
Statistical evaluation. Values are presented as means ⫾ standard deviation of the mean; statistical comparisons were made with Student’s t test (statistical significance was defined as P ⬍ 0.05).
15.
ACKNOWLEDGMENTS
16.
We thank Yoshinaga Saeki (Massachusetts General Hospital) for the generous gift of fHSV-delta-pac. This work was supported in part by NIH Grant P50CA058204 (X.Z.) and R01 CA85931 (R.V.).
17.
RECEIVED FOR PUBLICATION DECEMBER 16, 2002; ACCEPTED MARCH 4, 2003.
18.
REFERENCES
19.
1. Martuza, R. L., Malick, A., Markert, J. M., Ruffner, K. L., and Coen, D. M. (1991). Experimental therapy of human glioma by means of a genetically engineered virus mutant. Science 252: 854 – 856. 2. Mineta, T., Rabkin, S. D., Yazaki, T., Hunter, W. D., and Martuza, R. L. (1995). Attenuated multi-mutated herpes simplex virus-1 for the treatment of malignant gliomas. Nat. Med. 1: 938 –943. 3. Toda, M., Rabkin, S. D., and Martuza, R. L. (1998). Treatment of human breast cancer in a brain metastatic model by G207, a replication-competent multimutated herpes simplex virus 1. Hum. Gene Ther. 9: 2177–2185. 4. Walker, J. R., et al. (1999). Local and systemic therapy of human prostate adenocarcinoma with the conditionally replicating herpes simplex virus vector G207. Hum. Gene Ther. 10: 2237–2243. 5. Toyoizumi, T., et al. (1999). Combined therapy with chemotherapeutic agents and herpes simplex virus type 1 ICP34.5 mutant (HSV-1716) in human non-small cell lung cancer. Hum. Gene Ther. 10: 3013–3029. 6. Coukos, G., et al. (1999). Use of carrier cells to deliver a replication-selective herpes simplex virus-1 mutant for the intraperitoneal therapy of epithelial ovarian cancer. Clin. Cancer Res. 5: 1523–1537. 7. Carroll, N. M., Chiocca, E. A., Takahashi, K., and Tanabe, K. K. (1996). Enhancement of gene therapy specificity for diffuse colon carcinoma liver metastases with recombinant herpes simplex virus. Ann. Surg. 224: 323–329, discussion 329 –330. 8. Pawlik, T. M., et al. (2000). Oncolysis of diffuse hepatocellular carcinoma by intravas-
754
20. 21. 22. 23. 24.
25.
26.
27.
cular administration of a replication-competent, genetically engineered herpesvirus. Cancer Res. 60: 2790 –2795. Sundaresan, P., Hunter, W. D., Martuza, R. L., and Rabkin, S. D. (2000). Attenuated, replication-competent herpes simplex virus type 1 mutant G207: safety evaluation in mice. J. Virol. 74: 3832–3841. Todo, T., et al. (2000). Viral shedding and biodistribution of G207, a multimutated, conditionally replicating herpes simplex virus type 1, after intracerebral inoculation in aotus. Mol. Ther. 2: 588 –595. Markert, J. M., et al. (2000). Conditionally replicating herpes simplex virus mutant, G207 for the treatment of malignant glioma: results of a phase I trial. Gene Ther. 7: 867– 874. Rampling, R., et al. (2000). Toxicity evaluation of replication-competent herpes simplex virus (ICP 34.5 null mutant 1716) in patients with recurrent malignant glioma. Gene Ther. 7: 859 – 866. Kirn, D. (2001). Oncolytic virotherapy for cancer with the adenovirus dl1520 (Onyx015): results of phase I and II trials. Expert Opin. Biol. Ther. 1: 525–538. Nemunaitis, J., et al. (2001). Phase II trial of intratumoral administration of ONYX-015, a replication-selective adenovirus, in patients with refractory head and neck cancer. J. Clin. Oncol. 19: 289 –298. Kramm, C. M., et al. (1997). Therapeutic efficiency and safety of a second-generation replication-conditional HSV1 vector for brain tumor gene therapy. Hum. Gene Ther. 8: 2057–2068. Bateman, A., et al. (2000). Fusogenic membrane glycoproteins as a novel class of genes for the local and immune-mediated control of tumor growth. Cancer Res. 60: 1492– 1497. Fielding, A. K., et al. (2000). A hyperfusogenic gibbon ape leukemia envelope glycoprotein: targeting of a cytotoxic gene by ligand display. Hum. Gene Ther. 11: 817– 826. Januszeski, M. M., Cannon, P. M., Chen, D., Rozenberg, Y., and Anderson, W. F. (1997). Functional analysis of the cytoplasmic tail of Moloney murine leukemia virus envelope protein. J. Virol. 71: 3613–3619. Higuchi, H., et al. (2000). Viral fusogenic membrane glycoprotein expression causes syncytia formation with bioenergetic cell death: implications for gene therapy. Cancer Res. 60: 6396 – 6402. Diaz, R. M., et al. (2000). A lentiviral vector expressing a fusogenic glycoprotein for cancer gene therapy. Gene Ther. 7: 1656 –1663. Weir, J. P. (2001). Regulation of herpes simplex virus gene expression. Gene 271: 117–130. Stavropoulos, T. A., and Strathdee, C. A. (1998). An enhanced packaging system for helper-dependent herpes simplex virus vectors. J. Virol. 72: 7137–7143. Fu, X., and Zhang, X. (2002). Potent systemic antitumor activity from an oncolytic herpes simplex virus of syncytial phenotype. Cancer Res. 62: 2306 –2312. Saeki, Y., et al. (1998). Herpes simplex virus type 1 DNA amplified as bacterial artificial chromosome in Escherichia coli: rescue of replication-competent virus progeny and packaging of amplicon vectors. Hum. Gene Ther. 9: 2787–2794. Guzowski, J. F., and Wagner, E. K. (1993). Mutational analysis of the herpes simplex virus type 1 strict late UL38 promoter/leader reveals two regions critical in transcriptional regulation. J. Virol. 67: 5098 –5108. Zhang, X., Efstathiou, S., and Simmons, A. (1994). Identification of novel herpes simplex virus replicative intermediates by field inversion gel electrophoresis: implications for viral DNA amplification strategies. Virology 202: 530 –539. Schang, L. M., Phillips, J., and Schaffer, P. A. (1998). Requirement for cellular cyclindependent kinases in herpes simplex virus replication and transcription. J. Virol. 72: 5626 –5637.
MOLECULAR THERAPY Vol. 7, No. 6, June 2003 Copyright © The American Society of Gene Therapy
doi:10.1016/S1525-0016(03)00092-3
Expression of a Fusogenic Membrane Glycoprotein by an Oncolytic Herpes Simplex Virus Potentiates the Viral Antitumor Effect Xinping Fu,1,* Lihua Tao,1,* Aiwu Jin,1 Richard Vile,2 Malcolm K. Brenner,1 and Xiaoliu Zhang1,† 1
Center for Cell and Gene Therapy, Baylor College of Medicine, Houston, Texas 77030, USA 2 Molecular Medicine Program, Mayo Clinic, Rochester, Minnesota 55905, USA *Current address: The Fifth Affiliated Hospital of Zhongshan University, Zhuhai, China.
†
To whom correspondence and reprint requests should be addressed. Fax: (713) 798-1230. E-mail: [email protected].
Oncolytic viruses have shown considerable promise in the treatment of solid tumors, but their potency must be improved if their full clinical potential is to be realized. We inserted the gene encoding a truncated form of the gibbon ape leukemia virus envelope fusogenic membrane glycoprotein (GALV.fus) into an oncolytic herpes simplex virus, using an enforced ligation procedure. Subsequent in vitro and in vivo studies showed that expression of GALV.fus in the context of an oncolytic virus significantly enhances the antitumor effect of the virus. Furthermore, by controlling GALV.fus expression through a strict late viral promoter, whose activity depends on the initiation of viral DNA replication, we were able to express this glycoprotein in tumor cells but not in normal nondividing cells. It will be of interest to confirm whether functional expression of a strong fusogenic gene by an oncolytic herpes simplex virus enhances viral antitumor activity without increasing its toxicity. Key Words: oncolytic, herpes simplex virus, late viral promoter, fusogenic membrane glycoprotein, liver cancer
INTRODUCTION Replication-selective oncolytic viruses are potentially useful antitumor agents. Their genetic modifications enable them to replicate preferentially within tumor cells, while being restricted in their ability to replicate in normal cells. The principal antitumor mechanism used by these viruses is a direct cytopathic effect, produced as the viruses propagate and spread from initially infected tumor cells to surrounding tumor cells, achieving a progressively larger volume of distribution and enhanced tumor cell killing. Oncolytic herpes simplex viruses (HSVs) were initially designed and constructed for the treatment of brain tumors [1,2]. Subsequently, they were found to have activity against a variety of other human solid tumors, including breast cancer [3], prostate cancer [4], lung cancer [5], ovarian cancer [6], and colon and liver cancers [7,8]. The safety of oncolytic HSVs has been extensively tested in studies in mice [9] and in primates [10], which showed that these agents are safe for in vivo administration. Despite these encouraging preclinical studies, results
748
from early clinical trials suggested that current oncolytic viruses, although safe, may have only limited antitumor activity on their own [11–14]. One of the main reasons for this suboptimal oncolytic efficacy is that the viral gene deletions conferring tumor selectivity also reduce the replicative potential of the virus in tumors. For example, complete elimination of endogenous ␥34.5 gene function from HSV, one of the common approaches to construction of an oncolytic HSV, significantly reduces viral replication and compromises the ability of the virus to spread within targeted tumors [15]. Thus, strategies are needed to enhance the potency of oncolytic viruses. Recently, Bateman et al. [16] reported that envelope fusogenic membrane glycoproteins (FMGs) expressed by certain viruses are extremely cytotoxic to tumor cells. An example of one such FMG is a C-terminal truncated form of the gibbon ape leukemia virus envelope glycoprotein (GALV.fus) [16,17] that lacks the 16-amino-acid R-peptide of the wild-type protein, which normally serves to restrict fusion of the envelope until it is cleaved during viral maturation [18]. This alteration renders the protein con-
MOLECULAR THERAPY Vol. 7, No. 6, June 2003 Copyright © The American Society of Gene Therapy 1525-0016/03 $30.00
doi:10.1016/S1525-0016(03)00092-3
ARTICLE
stitutively highly fusogenic with human cells that express the Pit-1 receptor for GALV. Transduction of the GALV.fus gene into a range of human tumor cells efficiently kills the cells through a process of syncytial formation [19]. It also destroys neighboring cells through a “bystander” effect at least 10-fold greater than that observed with HSV-Tk or cytosine deaminase-mediated therapy [19,20]. The full therapeutic potential of GALV.fus will probably not be realized until means are devised for its efficient delivery and specific expression in tumor cells. We show here that these aims can be achieved by expressing the truncated glycoprotein in the context of an oncolytic HSV. In particular, expressing GALV.fus under control of a strictly late viral promoter, whose activity depends on the ability of the oncolytic virus to replicate in tumor cells, minimized the expression of this glycoprotein in normal tissues.
RESULTS Construction and Initial Characterization of the New Fusogenic Viruses We used an enforced ligation strategy to insert either GALV.fus or the enhanced green fluorescent protein (EGFP) gene into a bacterial artificial chromosome (BAC)based HSV construct that was deleted for both the ␥34.5 gene and the HSV packaging signal (fHSV-delta-pac). The genes encoding GALV.fus or EGFP were driven either by the CMV immediate early promoter or by a strict late viral promoter. We believe that directing GALV.fus expression with a strict late viral promoter, in the context of an oncolytic HSV, may selectively express the fusogenic glycoprotein in tumor tissue. This is because transcription of the HSV genome is a tightly regulated process in which early and late phases of gene expression are separated by viral DNA replication [21]. In particular, expression of strict late transcripts (e.g., those of UL38) depends rigorously on the initiation of viral DNA replication. The promoters of these strict late genes, such as UL38p, which drives UL38 expression, would be active only in tumor cells in which the oncolytic HSV can fully replicate; in nondividing or postmitotic normal cells, they would be silent since viral replication would be limited. We constructed a total of three viruses, including Baco-1 (containing EGFP driven by the CMV promoter), Synco-1 (containing GALV.fus driven by the CMV promoter), and Synco-2 (containing GALV.fus driven by UL38p). We analyzed the genomes of these newly generated viruses by enzyme digestion to confirm that their structure had not undergone significant rearrangement. DNA extracted from supernatant virion particles (Baco-1, Synco-1, and Synco-2) or from bacterial cultures (fHSVdelta-pac) was digested with BamHI, which cuts multiple times within the viral genome. The result of gel electrophoresis showed that the restriction band patterns among the viruses and the original fHSV-delta-pac construct are
MOLECULAR THERAPY Vol. 7, No. 6, June 2003 Copyright © The American Society of Gene Therapy
FIG. 1. Characterization of virion DNA by BamHI digestion. Virion DNA or fHSV-delta-pac DNA extracted from bacterial culture was digested with BamHI and run on a 0.8% gel together with a 1 kb DNA ladder as marker (M) at a constant voltage for 6 h.
similar (Fig. 1), which is consistent with published data from Stavropoulos and Strathdee [22]. The small differences in the band patterns among them may reflect the circular form of fHSV-delta-pac and the different extra DNA sequences among the recombinant viruses. We evaluated the growth properties of the newly created viruses in both Vero cells and the Hep 3B human liver cancer cell line. We infected cells with the viruses at different multiplicities of infection (m.o.i.), harvested the viruses at different times, and titrated them by plaque assay. As shown in Fig. 2, Synco-1 and Synco-2 had virtually identical growth curves and maximal viral titers. Although the titers for the fusogenic viruses were slightly higher than those of Baco-1 at later stages of infection (i.e., 24 and 48 h after infection), these differences were not statistically significant. Phenotypic Characterization of the Fusogenic Viruses To characterize the phenotypes of the fusogenic oncolytic HSVs, we infected human cancer cells of different tissue origins with the viruses at a low m.o.i. At 36 h after infection, a clear syncytial phenotype was observed for all tumor cells infected by either Synco-1 or Synco-2 (Fig. 3a); the extent of tumor cell fusion did not differ appreciably between Synco-1 and Synco-2. By contrast, the same types of tumor cells infected with Baco-1 lacked this phenotype. To determine if GALV.fus-mediated syncytial formation after Synco-2 infection truly depends on viral DNA replication, we infected human tumor cells with either Synco-1 or Synco-2 in the presence or absence of acyclovir
749
ARTICLE
doi:10.1016/S1525-0016(03)00092-3
(ACV), which inhibits HSV DNA replication. Syncytial formation by Synco-1 in Hep 3B cells was largely unaffected by the presence of ACV (Fig. 3a), while cell membrane fusion by Synco-2 in the same cells was completely blocked by this drug. We also tested directly whether Synco-2 loses its ability to cause syncytial formation in normal nondividing human cells, by infecting primary human fibroblasts in either a quiescent or a cycling state. As shown in Fig. 3b, infection with either Synco-1 or Synco-2 caused syncytial formation in these normal human cells when they were in cycle; the syncytia formed by Synco-1 infection were noticeably larger than those produced by Synco-2 (3.6 ⫾ 0.4 vs 1.8 ⫾ 0.2 nuclei per syncytium). Synco-1-mediated cell fusion was only marginally affected in cells whose cycling was either slowed by serum starvation (2.3 ⫾ 0.2 nuclei per syncytium) or completely arrested with lovastatin (2.0 ⫾ 0.3 nuclei per syncytium). Synco-2-mediated cell fusion, on the other hand, was absent in cells whose cycling was decreased or arrested. The infected cells appeared essentially identical to the uninfected cells, except that after extended incubation, cytopathic effect resembling cell fusion did become evident in the serum-starved cells by day 4, while the cells treated with lovastatin began to die from effects of the drug by day 5 (data not shown). These results may imply that syncytial formation induced by the UL38 promoter-controlled GALV.fus expression in the context of an oncolytic HSV is cell-cycle-dependent. FIG. 3. Phenotypic characterization of Synco-1 and Synco-2 in tumor cells and in cycling or arrested embryonic fibroblasts. (a) Phenotypic characterization of viruses in tumor cells. Tumors of different tissue origin were infected with each of the viruses at 0.01 pfu/cell, with or without acyclovir (ACV) in the culture medium. Photos were taken at 48 h after initial virus infection. The photos of cells in the presence of ACV were taken 72 h after infection. (b) Phenotypic characterization of viruses in embryonic fibroblasts of cycling or arrested status. Fully cycling, serum-starved, or arrested (with lovastatin) human embryonic fibroblasts were infected with either Synco-1 or Synco-2 at 0.01 pfu/cell or left uninfected. The photos were taken 48 h after infection. Original magnification, ⫻200.
FIG. 2. Growth characterization of oncolytic HSVs. Vero cells (a) and Hep 3B cells (b) were seeded into 24-well plates in triplicate and infected with the viruses at 0.1 or 1 pfu/cell. The viruses were harvested at 0-, 12-, 24-, and 48-h intervals after infection and were titrated by plaque assay.
750
Comparison of Tumor Cell Killing Next, we determined if GALV.fus-mediated syncytial formation in the context of oncolytic HSV replication would increase tumor cell killing. We initially compared the viability of cells after they were infected with either Baco-1 or Synco-1 at a low m.o.i. (0.1 and 0.01 plaqueforming units (pfu)/cell), so that both the inherent cytotoxicity of the input virus and the ability of the virus to replicate and spread in these cells could be assessed. The cytotoxic effect of the virus infection on tumor cells was quantified by calculating the percentage of cells that survived at different time points after infection. In all tumor cell lines, Synco-1 killed tumor cells more effectively than Baco-1 at both time points and both viral doses (Figs. 4a and 4b, P ⬍ 0.01 for all comparisons). We also compared
MOLECULAR THERAPY Vol. 7, No. 6, June 2003 Copyright © The American Society of Gene Therapy
doi:10.1016/S1525-0016(03)00092-3
ARTICLE
the Hep 3B cell killing ability of Synco-1 and Synco-2 directly with that of Fu-10, a fusogenic oncolytic HSV selected from G207 through random mutagenesis [23]. The results showed that while all versions of the fusogenic oncolytic HSVs had a significantly greater ability to destroy tumor cells than did Baco-1, their individual oncolytic activities did not differ significantly against this particular tumor cell line (Fig. 4c).
FIG. 4. Increased tumor cell killing by fusogenic oncolytic HSVs. (a) Tumor cells of different tissue origin were infected with either Baco-1 or Synco-1 at 0.1 pfu/cell. (b) The same tumor cells were infected with either Baco-1 or Synco-1 at 0.01 pfu/cell. (c) Hep 3B cells were infected with the indicated viruses at either 0.1 or 0.01 pfu/cell. The percentage of surviving cells was calculated by dividing the number of live cells from the infected wells by the number of live cells in the well that was not infected. P ⬍ 0.01 for all comparisons between Synco-1/Synco-2 and Baco-1. No statistically significant difference between Synco-1 and Synco-2 was found.
MOLECULAR THERAPY Vol. 7, No. 6, June 2003 Copyright © The American Society of Gene Therapy
Enhanced Antitumor Effect in Xenografted Human Tumors To determine if the syncytial phenotype and enhanced ex vivo tumor cell killing shown by Synco-1 would translate into an enhanced antitumor effect in vivo, we injected the virus at 1 ⫻ 107 pfu directly into established xenografts of Hep 3B carcinomas with diameters of 5 to 8 mm. Tumor sizes were measured weekly for 4 weeks. Compared with the PBS controls, a single injection of either Baco-1 or Synco-1 had an immediate effect on tumor growth (Fig. 5a). Within 1 week of virus injection, the tumors in mice treated with these constructs were significantly smaller than tumors injected with PBS (P ⬍ 0.001). From week 2 to week 4, however, Synco-1 produced significantly greater antitumor effects than did Baco-1 (P ⬍ 0.001). Half of the animals (5 of 10) were tumor free by week 3 after Synco-1 administration, while the other half had significantly reduced tumor sizes. By contrast, only 1 mouse in the group injected with Baco-1 was tumor free. In the other 9 mice, tumors that had shrunk initially began to regrow by week 3 after virus injection. By week 4 —the end point of the experiment—the mice in the Synco-1injected group either were still tumor free (5/10) or had only small tumors, while the average tumor size in Baco1-injected mice had substantially increased from the pretreatment measurement (Fig. 5a). We then compared directly the antitumor potencies of Synco-1 and Synco-2 by injecting the viruses into xenografted human liver tumors in nude mice, as described earlier. The results showed that both constructs have significantly better antitumor effects than Baco-1, starting from week 2 after virus administration (P ⬍ 0.01). There was no significant difference in the tumor growth ratio between Synco-1- and Synco-2-treated mice (Fig. 5b). These results demonstrate that Synco-2 has antitumor potency equivalent to that of Synco-1, even though its GALV.fus expression is driven by a conditional viral promoter. To determine if syncytial formation is indeed part of the oncolytic process of Synco-1- and Synco-2-mediated antitumor therapy, we injected established Hep 3B tumors with oncolytic HSVs or PBS and excised the tumors 5 days later. Examination of stained tumor sections revealed that syncytia comprising a variable number of cell nuclei were frequently encountered across the tissue section of tumors injected with either Synco-1 or Synco-2
751
ARTICLE
doi:10.1016/S1525-0016(03)00092-3
FIG. 5. Enhanced oncolytic potency of fusogenic oncolytic HSVs in human liver cancer xenografts. (a) Enhanced antitumor activity of Synco-1. Treatment groups include Synco-1, Baco-1, or PBS. (b) Direct comparison of antitumor activities of Synco-1 and Synco-2. Treatment groups include Baco-1, Synco-1, and Synco-2. The tumor growth ratio was determined by dividing the tumor volume measured on the indicated week after virus injection by the tumor volume before treatment. The values represent means ⫾ standard deviation (n ⫽ 10 mice per group).
virus (Fig. 6). Such syncytia were not seen in tissue sections of tumors injected with either Baco-1 or PBS.
DISCUSSION Earlier attempts to clone the GALV.fus gene into an oncolytic adenovirus were not successful [20], for two possible reasons. First, the extensive and rapid syncytial formation after GALV.fus gene transfection may have interfered with the subsequent homologous recombination needed for the generation of recombinant virus. Sec-
FIG. 6. Virus-induced syncytial formation in tumor tissues. The syncytia in Synco-1- or Synco-2-infected tumor tissues are indicated by arrows (original magnification, ⫻400).
752
ond, expression of GALV.fus in the context of an adenovirus might have interfered with the infectious process. Unlike adenovirus, which is nonenveloped, HSV is an enveloped virus whose infection cycle naturally involves membrane fusion. The successful insertion of GALV.fus into an oncolytic HSV genome and subsequent demonstration of their syncytial phenotype indicate that HSV can withstand the membrane-fusion effect from genes such as GALV.fus and continue to grow. The virus construction strategy of direct in vitro ligation followed by transfection into Vero cells, which does not require homologous recombination, may have also contributed to the successful construction of the fusogenic viruses. In our early studies we showed that a syncytial mutant (Fu-10) selected from the well-characterized oncolytic HSV G207 through random mutagenesis has a significantly greater antitumor capacity than does the parental virus on xenografted lung metastatic breast cancer [23]. Here we demonstrate that expression of GALV.fus by a conditionally replicating HSV also leads to strong syncytial formation in tumor cells of different tissue origins, a property translating into a significantly increased tumor cell killing in vitro. Our in vivo data on xenografted human liver tumor demonstrate that fusogenic oncolytic HSVs constructed by this strategy have significantly elevated antitumor potency. Although direct comparison between Fu-10 and Synco-2 showed that they had similar antitumor potencies on established solid tumors of different tissue origins, including prostate cancer, liver cancer, and glioblastoma (X. Fu and X. Zhang, unpublished data), we believe that fusogenic oncolytic HSVs generated by incorporation of FMGs from viruses other than HSV may potentially be more advantageous than Fu-10. This is because syncytial formation of FMGs relies on the initial
MOLECULAR THERAPY Vol. 7, No. 6, June 2003 Copyright © The American Society of Gene Therapy
doi:10.1016/S1525-0016(03)00092-3
binding of the fusogenic glycoproteins with their receptors on the target cells. Unlike xenografted tumors established from cultured tumor cells that are relatively homogeneous, cancer cells in naturally occurring tumors may be heterogeneous in their sensitivity to HSV infection and/or syncytial formation due to variation of receptors on the cell surface. Insertion of a FMG from another virus that uses a receptor totally different from that of the oncolytic virus (Pit-1 vs heparan sulfate in the case of GALV FMG in oncolytic HSV) should significantly reduce the frequency of therapy-resistant tumor cells. This is because tumor cells that become resistant to oncolytic virus infection/replication may still be sensitive to the FMG-mediated syncytial formation and vice versa. Although expression of GALV.fus can add to the antitumor effects of oncolytic HSV, uncontrolled expression of the gene, even in this context, raises safety concerns. To minimize this potential problem, we used a strict late viral promoter to direct the expression of the GALV.fus gene in the virus. Our data show that this strategy provides strong tissue-selective expression of the gene. The UL38 promoter appears to have several advantages over currently available tissue-specific promoters. First, it is probably substantially more active than most tissue-specific promoters. Direct comparison of the UL38 promoter with the CMV immediate early promoter (one of the strongest mammalian promoters) showed that the former has only slightly lower activity than the latter (X. Fu and X. Zhang, unpublished data), whereas the activity of most tissue-specific promoters is 10 to 100 times lower than that of the CMV promoter. Second, activation of the UL38 promoter upon the initiation of lytic HSV infection in the tumor cells may synchronize the actions of these two antitumor mechanisms. Finally, the demonstration that GALV.fus-mediated fusion is essentially ablated when the cell cycle of an human embryonic fibroblast is arrested suggests that Synco-2 may be no more toxic to normal cells than the first generation of oncolytic HSVs, whose safety record has been established in both preclinical studies and phase I clinical trails. Due to the species specificity of GALV.fus, we could not test the safety of Synco-2 in our rodent models; however, confirmation of its safety in nonhuman primates (which are sensitive to GALV.fusmediated cell membrane fusion) will permit clinical exploration of this new generation of oncolytic HSVs.
MATERIALS
AND
METHODS
Cell lines. African green monkey kidney (Vero) cells, human embryonic fibroblasts (HF 333.We), human tumor cell lines U-87 MG (glioblastoma) and Hep 3B (hepatocellular carcinoma), and the human prostate cancer line DU 145 were obtained from American Tissue Culture Collection (Rockville, MD). All cells were cultured with DMEM containing 10% fetal bovine serum (FBS). Generation of fusogenic oncolytic HSVs. All oncolytic HSVs were derived from fHSV-delta-pac, a BAC-based construct that contains a mutated HSV genome, in which the diploid gene encoding ␥34.5 and both copies of the
MOLECULAR THERAPY Vol. 7, No. 6, June 2003 Copyright © The American Society of Gene Therapy
ARTICLE
HSV packaging signal (pac) have been deleted [24]. Infectious HSV cannot be generated from this construct unless an intact HSV pac is provided in cis. To insert foreign genes and to generate infectious viruses from fHSVdelta-pac, we used an enforced ligation strategy. Initially, three gene cassettes (GALV.fus driven by the CMV promoter, GALV.fus driven by the UL38 promoter, and the green fluorescent protein gene driven by the CMV promoter) were cloned into pSZ-pac so that each was linked with an HSV sequence containing HSV pac. The EGFP gene cassette was cut from pEGFP-N1 (Clontech, Palo Alto, CA) with AseI and AflIII. The GALV.fus gene was cut from pCR3.1-GALV with NheI and NotI and linked with the CMV promoter in pcDNA3. The promoter and enhancer region (DAS) of UL38, a well-defined 145-bp DNA sequence covering the 84324 – 84469 nucleotide region of the HSV genome [25], was amplified from HSV-1 DNA with the following pair of primers: forward 5⬘-GTGGGTTGCGGACTTTCTGC-3⬘ and reverse 5⬘-ACACTCACGCAAGGCGGAAC-3⬘. The GALV.fus or EGFP gene cassette that was linked with the HSV pac was then cut from the plasmids and directly ligated into the unique PacI site of fHSV-deltapac. The DNA ligation mixture was transfected into Vero cells using Lipofectamine (Gibco BRL, Grand Island, NY) and incubated for 3–5 days to permit generation of infectious virus. The viruses were subsequently plaque purified and were designated Baco-1 (containing the DNA fragment of EGFP and pac), Synco-1 (containing the DNA fragment of CMV-GALV. fus and pac), and Synco-2 (containing UL38p-GALV.fus and pac). Virus stocks were prepared by infecting Vero cells with the viruses at 0.01 pfu per cell, harvested after 2 days, and stored at ⫺80°C. Viral structure and growth characterization. Virion DNA was prepared from cell culture supernatants as described [26]. Briefly, 48 h after viral infection (at 0.01 pfu/cell), virion particles were pelleted by centrifuging the supernatants at 27,000g for 3 h. Viral pellets were resuspended in TE (10 mM Tris, 1 mM EDTA, pH 8.0) and incubated with SDS and proteinase K for 4 h at final concentrations of 0.1% and 50 g/ml, respectively. Virion DNA was then gently phenol-extracted three times before the DNA was precipitated with ethanol. For viral growth characterization, Vero cells seeded in triplicate into 48-well plates were infected with the viruses at 0.1 or 1 pfu/cell for 1 h. Cells were washed once with PBS to remove unadsorbed and uninternalized viruses before fresh medium was added. Cells were harvested at 0, 12, 24, and 48 h after infection. Viruses were released by repeated freezing and thawing and sonication. Virus titers were determined on Vero cells by a plaque assay. Phenotypic characterization of fusogenic oncolytic HSVs. Cells were seeded into six-well plates and infected the following day with each virus at a dose ranging from 0.1 to 0.0001 pfu/cell. Cells were cultured in a maintenance medium (containing 1% FBS) and were left for up to 2 days to allow the fusion pattern and plaques to develop. To block HSV DNA replication, we added ACV to the culture medium at a final concentration of 100 M. To slow the division or arrest of embryonic fibroblasts, we either starved cells for FBS for 48 h or starved them for FBS during incubation with 20 M lovastatin for 24 h, before the cells were infected with virus. Lovastatin is a chemical that induces cell-cycle arrest but does not interfere with HSV replication [27]. To quantify the size of syncytia induced by virus infection, the infected cell monolayer was stained with a solution containing 0.1% crystal violet (Sigma, St. Louis, MO) and 20% ethanol for 30 min at room temperature. The cells were then thoroughly rinsed with water. On average, 10 syncytia were randomly selected from each sample and the nuclei in each syncytium were enumerated. The result was given as mean ⫾ deviation. In vitro cell killing assay. Each of the three human tumor cell lines was seeded into 48-well plates and infected with each virus at 0.1 and 0.01 pfu/cell or left uninfected. Cells were harvested 24 and 48 h later by trypsinization. The number of surviving cells was counted on a hemacytometer following trypan blue staining. The percentage of surviving cells was calculated by dividing the number of viable cells from the infected well by the number of cells from the well that was left uninfected. The experiments were done in triplicate and the mean numbers were used for the final calculation.
753
ARTICLE
Animal experiments. Six-week-old male Hsd athymic (nu/nu) mice were purchased from Harlan (Indianapolis, IN). Hep 3B cells were cultured under standard conditions and were harvested in log phase with 0.05% trypsin–EDTA. The cells were washed twice with serum-free medium before they were resuspended in PBS at a concentration of 5 ⫻ 107 cells/ml. A total of 5 ⫻ 106 cells (in a l00-l suspension) were subcutaneously injected into the right flank of each mouse. When the tumors reached approximately 5– 8 mm in diameter, they were injected with virus (1 ⫻ 107 pfu in a 100-l volume); control tumors received an injection of PBS only. Tumor size was measured weekly for 4 weeks, and tumor volume was calculated by the formula tumor volume (mm3) ⫽ 0.5 ⫻ [length (mm)] ⫻ [width (mm)]2. For pathological examination of the tumor tissues, mice were sacrificed 5 days after virus or PBS injection. The tumors were removed, fixed, and stained with hematoxylin and eosin.
doi:10.1016/S1525-0016(03)00092-3
9.
10.
11.
12.
13. 14.
Statistical evaluation. Values are presented as means ⫾ standard deviation of the mean; statistical comparisons were made with Student’s t test (statistical significance was defined as P ⬍ 0.05).
15.
ACKNOWLEDGMENTS
16.
We thank Yoshinaga Saeki (Massachusetts General Hospital) for the generous gift of fHSV-delta-pac. This work was supported in part by NIH Grant P50CA058204 (X.Z.) and R01 CA85931 (R.V.).
17.
RECEIVED FOR PUBLICATION DECEMBER 16, 2002; ACCEPTED MARCH 4, 2003.
18.
REFERENCES
19.
1. Martuza, R. L., Malick, A., Markert, J. M., Ruffner, K. L., and Coen, D. M. (1991). Experimental therapy of human glioma by means of a genetically engineered virus mutant. Science 252: 854 – 856. 2. Mineta, T., Rabkin, S. D., Yazaki, T., Hunter, W. D., and Martuza, R. L. (1995). Attenuated multi-mutated herpes simplex virus-1 for the treatment of malignant gliomas. Nat. Med. 1: 938 –943. 3. Toda, M., Rabkin, S. D., and Martuza, R. L. (1998). Treatment of human breast cancer in a brain metastatic model by G207, a replication-competent multimutated herpes simplex virus 1. Hum. Gene Ther. 9: 2177–2185. 4. Walker, J. R., et al. (1999). Local and systemic therapy of human prostate adenocarcinoma with the conditionally replicating herpes simplex virus vector G207. Hum. Gene Ther. 10: 2237–2243. 5. Toyoizumi, T., et al. (1999). Combined therapy with chemotherapeutic agents and herpes simplex virus type 1 ICP34.5 mutant (HSV-1716) in human non-small cell lung cancer. Hum. Gene Ther. 10: 3013–3029. 6. Coukos, G., et al. (1999). Use of carrier cells to deliver a replication-selective herpes simplex virus-1 mutant for the intraperitoneal therapy of epithelial ovarian cancer. Clin. Cancer Res. 5: 1523–1537. 7. Carroll, N. M., Chiocca, E. A., Takahashi, K., and Tanabe, K. K. (1996). Enhancement of gene therapy specificity for diffuse colon carcinoma liver metastases with recombinant herpes simplex virus. Ann. Surg. 224: 323–329, discussion 329 –330. 8. Pawlik, T. M., et al. (2000). Oncolysis of diffuse hepatocellular carcinoma by intravas-
754
20. 21. 22. 23. 24.
25.
26.
27.
cular administration of a replication-competent, genetically engineered herpesvirus. Cancer Res. 60: 2790 –2795. Sundaresan, P., Hunter, W. D., Martuza, R. L., and Rabkin, S. D. (2000). Attenuated, replication-competent herpes simplex virus type 1 mutant G207: safety evaluation in mice. J. Virol. 74: 3832–3841. Todo, T., et al. (2000). Viral shedding and biodistribution of G207, a multimutated, conditionally replicating herpes simplex virus type 1, after intracerebral inoculation in aotus. Mol. Ther. 2: 588 –595. Markert, J. M., et al. (2000). Conditionally replicating herpes simplex virus mutant, G207 for the treatment of malignant glioma: results of a phase I trial. Gene Ther. 7: 867– 874. Rampling, R., et al. (2000). Toxicity evaluation of replication-competent herpes simplex virus (ICP 34.5 null mutant 1716) in patients with recurrent malignant glioma. Gene Ther. 7: 859 – 866. Kirn, D. (2001). Oncolytic virotherapy for cancer with the adenovirus dl1520 (Onyx015): results of phase I and II trials. Expert Opin. Biol. Ther. 1: 525–538. Nemunaitis, J., et al. (2001). Phase II trial of intratumoral administration of ONYX-015, a replication-selective adenovirus, in patients with refractory head and neck cancer. J. Clin. Oncol. 19: 289 –298. Kramm, C. M., et al. (1997). Therapeutic efficiency and safety of a second-generation replication-conditional HSV1 vector for brain tumor gene therapy. Hum. Gene Ther. 8: 2057–2068. Bateman, A., et al. (2000). Fusogenic membrane glycoproteins as a novel class of genes for the local and immune-mediated control of tumor growth. Cancer Res. 60: 1492– 1497. Fielding, A. K., et al. (2000). A hyperfusogenic gibbon ape leukemia envelope glycoprotein: targeting of a cytotoxic gene by ligand display. Hum. Gene Ther. 11: 817– 826. Januszeski, M. M., Cannon, P. M., Chen, D., Rozenberg, Y., and Anderson, W. F. (1997). Functional analysis of the cytoplasmic tail of Moloney murine leukemia virus envelope protein. J. Virol. 71: 3613–3619. Higuchi, H., et al. (2000). Viral fusogenic membrane glycoprotein expression causes syncytia formation with bioenergetic cell death: implications for gene therapy. Cancer Res. 60: 6396 – 6402. Diaz, R. M., et al. (2000). A lentiviral vector expressing a fusogenic glycoprotein for cancer gene therapy. Gene Ther. 7: 1656 –1663. Weir, J. P. (2001). Regulation of herpes simplex virus gene expression. Gene 271: 117–130. Stavropoulos, T. A., and Strathdee, C. A. (1998). An enhanced packaging system for helper-dependent herpes simplex virus vectors. J. Virol. 72: 7137–7143. Fu, X., and Zhang, X. (2002). Potent systemic antitumor activity from an oncolytic herpes simplex virus of syncytial phenotype. Cancer Res. 62: 2306 –2312. Saeki, Y., et al. (1998). Herpes simplex virus type 1 DNA amplified as bacterial artificial chromosome in Escherichia coli: rescue of replication-competent virus progeny and packaging of amplicon vectors. Hum. Gene Ther. 9: 2787–2794. Guzowski, J. F., and Wagner, E. K. (1993). Mutational analysis of the herpes simplex virus type 1 strict late UL38 promoter/leader reveals two regions critical in transcriptional regulation. J. Virol. 67: 5098 –5108. Zhang, X., Efstathiou, S., and Simmons, A. (1994). Identification of novel herpes simplex virus replicative intermediates by field inversion gel electrophoresis: implications for viral DNA amplification strategies. Virology 202: 530 –539. Schang, L. M., Phillips, J., and Schaffer, P. A. (1998). Requirement for cellular cyclindependent kinases in herpes simplex virus replication and transcription. J. Virol. 72: 5626 –5637.
MOLECULAR THERAPY Vol. 7, No. 6, June 2003 Copyright © The American Society of Gene Therapy