Fusogenic Effects of Murine Retroviruses and Cationic Enhancers of Transduction

Fusogenic effects of murine retroviruses and cationic enhancers of transduction

Guoping Xu, Fauzia Solaiman, Mary Ann Zink, and Clague P. Hodgson Creighton Cancer Center/Department of Biomedical Sciences, Creighton University School of Medicine, Omaha, Nebraska 68178.

The maturation of retrovirus particles involves proteolytic cleavage of the envelope glycoprotein transmembrane component, resulting in conversion of the virus particle to a fusogenic or infectious state. Susceptible murine cells exposed to virus-containing supernatants from ecotropic retroviral helper cells occasionally fused to neighboring cells, resulting in syncytia (giant cells with multiple nuclei). Polycationic molecules dramatically enhanced the effect, leading to widespread cell death. The degree of cell fusion was dependent upon the retroviral envelope subtype (ecotropic3amphotropic, gibbon ape leukemia virus was negative) as well as on the polycationic reagent used (G9 dendrimer3Lipofectamine3polybrene). Cell fusion effects were not mediated by the retroviral vector backbone, because virus-containing supernatants from helper cells (without vector) and vector producer cells had a similar effect. Human target cells were not fused by any type of murine retrovirus; in addition, amphotropic virus from human helper cells was not fusogenic toward murine cells. Thus, fusogenic effects were important during the propagation of vectors using murine helper cells but were not a signiﬁcant factor during the transduction of human cells. Cancer Gene Therapy (2000) 7, 53–58

Key words: Retrovirus; vector; gene therapy; syncytia; cell fusion.

etrovirus-derived vectors (RVs) are widely used for dendrimers, a high positive charge density appears to be Rgene delivery to human cells and tissues. A signifi- a significant factor related to its efficiency. cant advantage of RVs is their ability to integrate into Many retroviruses (including those that infect hu- chromosomal DNA, providing the basis for long-term mans) induce the formation of giant fused cells, or therapeutic gene expression without the expression of syncytia.4–7 In the case of human viruses, fusogenic viral genes. However, a significant problem associated cytocidal strains of HIV-1 have been implicated in the with RVs is their low or moderate efficiency of transduc- pathogenicity of AIDS.6 The fusion mechanism for 3 tion. Titers of retroviral vectors usually range from 10 ecotropic murine leukemia virus involves maturation of 6 1 to 10 transducing units/mL. Polycationic reagents the transmembrane (TM) component of the env glyco- (such as hexadimethrine bromide (polybrene) or prota- protein,8 and this mechanism takes place after virus buds mine sulfate) increase retroviral titers by approximately from the cell. The transition of immature TM to its one order of magnitude and are routinely used for gene fusogenic form involves proteolytic cleavage of 16 amino transfers in vitro. acid residues at the carboxyl terminus of TM, or glyco- Recently, more powerful transduction-enhancing re- protein 15, to a truncated and active glycoprotein 12. agents have been reported, including DOSPA:DOPE The fusion process of various retroviruses is reportedly liposomes (Lipofectamine, Life Technologies, Bethesda, 9 10 2 affected by changes in pH, protease activity, and Md) and polyamidoamine dendrimers (F. Solaiman and 11 12 3 amphotericin B, or by growth in murine XC cells. C.P.H., unpublished observations). These reagents in- The ability of cells to survive retroviral infections is crease the titers of murine retroviral vectors by 30- to apparently facilitated by the controlled maturation of 200-fold, apparently by neutralizing the negative charges wild-type TM after the virus leaves the cell. In addition associated with virus envelopes and target cell mem- to fusogenic mutations to TM, cell fusion can result from branes. Lipofectamine is postulated to interact with the massive infection of target cells by virus particles carry- retrovirus envelope (env) via the lipid moiety, permitting ing mature p12. One hypothesis is that the mature env the polycationic head group to more effectively neutral- glycoprotein could be incorporated into the cell mem- ize the surface charge of virus particles. In the case of brane during infection, converting the cell into a pseudovirus particle that is capable of fusing with its neighbors. Received November 10, 1998; March 4, 1999. Replication defective retroviral vectors are propa- Address correspondence and reprint requests to Dr. Clague P. Hodgson, gated by helper cells. Vectors are often transmitted to Nature Technology Corporation, 4701 Innovation Drive, Lincoln, NE complementary helper cells (containing a different env) 68521. E-mail address: [email protected] by cross-infection or by coculture of two complementa-

Cancer Gene Therapy, Vol 7, No 1, 2000: pp 53–58 53 54 XU ET AL: FUSOGENIC EFFECTS OF RETROVIRAL HELPER CELLS

tion cell lines. The observation of massive cell fusion duced with supernatant from PA317/pBAG cells (bearing a during cross-infection of vector producer cells (VPCs) BAG18 retrovirus with a ␤-galactosidase marker gene) and led us to investigate cause and effect. were selected with G418. The selected HT1080/BAG cells showed nearly 100% blue cells when stained with 5-bromo-4- In this report, we describe extensive cell fusion asso- ␤ ciated with the use of ecotropic (and, to a lesser extent, chloro-3-indolyl -D-galactoside and were used as marker- bearing cells. Supernatants from VPCs were filtered and amphotropic) retrovirus particles. The cytotoxic effects applied to marker-bearing cells using Lipofectamine as de- are enhanced by the use of polycationic reagents. The scribed above. Transduced marker cells were maintained for 2 results have relevance to the fields of gene therapy, virus weeks and were split twice to permit amplification of any RCR detection, and retroviral pathogenicity. that might be present. The supernatants were then filtered and added to fresh HT1080 target cells. The 5-bromo-4-chloro-3- indolyl ␤-D-galactoside staining procedure was used to detect MATERIALS AND METHODS infected cells. No RCRs were detected. Cell lines and cell culture Determination of syncytia formation 13 14 15 Cell lines NIH 3T3 and PA317, HT1080, PG13, and 19 16 Syncytia were scored using a syncytial index, by estimating SK-LU-1 were obtained from the American Type Culture 17 the percentage of nuclei present in cells with more than two Collection (Manassas, Va). The GPϩE86 cell line was kindly nuclei. The scores reported in Tables 1–6 are as follows: Ϫ, provided by Arthur Bank (Columbia University, New York, Ͻ5%; ϩ, 5–24%; ϩϩ, 25–74%; ϩϩϩ, 75–100%. The results NY). Normal human epidermal keratinocytes (NHEK) were represent a consensus of two observations. obtained from Clonetics (San Diego, Calif). The HTam1 helper cell line was developed in our laboratory, and will be pH experiment described elsewhere (F. Solaiman, M. A. Zink, and C.P.H., unpublished observations). NHEK cells were grown in kera- Complete media with 10% FBS were buffered with 20 mM tinocyte growth medium provided by the supplier (Clonetics). N-2-hydroxyethylpiperazine-NЈ-2-ethanesulfonic acid (Life Other cells were grown in Dulbecco’s modified Eagle’s me- Technologies) and adjusted to pH values of 7.25 (normal dium (DMEM) (Life Technologies, Gaithersburg, Md) supple- medium pH), 6.50, and 6.00, respectively. Donor and target mented with 10% (vol/vol) fetal bovine sera (FBS) (Life cells were incubated overnight with media of different pH Technologies). PA317 helper cells were preselected for 3 days values, before transfer of donor media. in HAT medium (DMEM containing 20 ␮M hypoxanthine, 30 ␮M thymidine, and 1 ␮M aminopterin) and grown in HT Sera experiment medium (HAT lacking aminopterin). GPϩE86 cells were Two brands (Life Technologies and Intergen, Inc., Purchase, selected in HXM medium (DMEM containing 10% FBS, 15 NY, three lots from each) of sera were tested for fusion effects. ␮M/mL hypoxanthine, 250 ␮M/mL xanthine, and 25 ␮M/mL Ten percent (vol/vol) of sera was added to DMEM. mycophenolic acid) and subsequently maintained in 0.5ϫ HXM medium. RESULTS Transduction procedures Cell fusion observed during cross-infection Target cells were plated on the day before transduction at a density of 1 ϫ 105 cells/well in a 6-well plate. Helper cells were In preliminary experiments (Fig 1), ecotropic virus par- grown to ϳ90% confluence in 10-cm dishes and were fed 7 mL ticles (from GPϩE86 cells) were incubated with G9 of fresh media on the day before collection. The supernatants dendrimers prior to infection of PA317 amphotropic ␮ were collected the next morning by 0.45- m filtration. Next, helper cells. The infected cells formed syncytia within 16 the supernatants were used for transduction or were frozen at Ϫ hours, resulting in widespread cell death. In later stages, 80°C for later use. All experiments in this report were done giant cells contained multiple pools of ϳ10–30 nuclei with frozen supernatants except for those specifically men- tioned. In the standard transduction procedure, 8 ␮Lof each, scattered throughout a continuous cytoplasm. This ␮ occurred regardless of whether retroviral vectors were Lipofectamine (Life Technologies), 9 L of dendrimer G9 (3 ϩ mg/mL, Life Technologies) or 8 ␮g of polybrene (Sigma, St. present in the GP E86 cells. The supernatant from the Louis, Mo) was incubated in 100 ␮L of serum-free DMEM for recipient PA317 cells, however, had no apparent effect 20 minutes at room temperature. A total of 0.05–1.0 mL of when passed to fresh PA317 cells, and there was no various supernatants were added to the polycation preparation evidence of RCRs as determined by a sensitive marker (above) and incubated for 20 minutes. Media were removed rescue assay in HT1080 cells. from target cells, and reagent-supernatant mixtures were added to the cells. Fresh medium with 10% FBS was then Polycations enhanced retroviral fusion effects added for a total volume per well of 1 mL. Transduced cells This cell fusion process is reportedly enhanced by the were incubated at 37°C in a 5% CO2 ambient environment 11 overnight and examined microscopically the next morning polyene antibiotic amphotericin B. However, at a (ϳ16 hours) or at the times indicated. Photographs were taken concentration of 0.25 ␮g/mL (recommended for tissue with a Nikon camera mounted on an Olympus inverted phase culture), amphotericin B had no apparent effect on the contrast microscope. fusion process when ecotropic virus was used (data not shown). Therefore, we asked whether the polycations Marker rescue assay were primarily involved in the fusion process (Table 1). A marker rescue procedure was used to detect replication- Although addition of dendrimer alone (in the absence of competent retrovirus (RCR). HT1080 cells were first trans- virus) caused no apparent fusion, addition of 1 mL of

Cancer Gene Therapy, Vol 7, No 1, 2000 XU ET AL: FUSOGENIC EFFECTS OF RETROVIRAL HELPER CELLS 55

helper cells. The effect was greater when dendrimers were used. Amphotropic (PA317) retrovirus by itself had no noticeable fusogenic effect on any cell line. However, when dendrimers were added to amphotropic virus, 5–75% fusion was observed in susceptible murine (but not human) cells. In addition to the data shown in Table 2, PA317 virus plus G9 also tested negative on human SK-LU-1 and NHEK cells (data not shown). Gibbon ape leukemia virus (GALV) pseudotype virus (PG13 cells) had no detectable fusogenic effect, with or without dendrimers (Table 2). Non-helper (control) supernatants from NIH 3T3 cells or HT1080 cells were also nonfusogenic, with or without dendrimers. Finally, the virus produced by a human amphotropic helper cell line, HTam1 (F. Solaiman, M. A. Zink, and C.P.H., unpublished observations) (made from HT1080 cells) was not fusogenic in any cell line tested. Effects of medium pH on the fusion process The syncytia induced by murine ecotropic retrovirus are pH-independent within a wide range of tested pH values (4.5–7.5).19 We asked whether polycation-enhanced retroviral fusion was pH-sensitive. N-2-hydroxyethylpiperazine-NЈ-2-ethanesulfonic acid buffer (20 mM) was used to stabilize medium pH at 7.25, 6.5, and 6.0. For uniformity, the buffered medium was used on both GPϩE86 donor cells and PA317 recipient cells (Table 3). pH alterations within the range of 6.0–7.25 had no effect on the fusogenic process. Effects of sera from different manufacturers on the fusion process Figure 1. Fusion of PA317 cells by GPϩE86 cell supernatant and G9 dendrimers. A total of 1 mL of GPϩE86 cell supernatant containing We also investigated the possible role of the sera in the 9 ␮g of G9 dendrimer was incubated for 16 hours with a confluent induction of syncytia. We compared three lots each of plate of PA317 cells. two different brands of sera (Life Technologies and Intergen, Inc.), obtaining very similar results in each case. Results indicated about the same efficiency in the viral supernatant by itself was sufficient to induce fusion induction of fusogenic effects by ecotropic supernatant in 5–24% of cells. This effect was enhanced by the on PA317 cells in the presence of G9 or Lipofectamine addition of polybrene (25–74% fusion) and especially by Ͼ (data not shown). The results indicated that different Lipofectamine and G9 dendrimer ( 75% fusion). sera had no significant differential effects on cell fusion. Fusogenic effect of amphotropic retroviral particles Dose responses The above results led to the investigation of fusion The dosage effects of ecotropic supernatant and G9 effects by comparing different combinations of viral dendrimer were investigated to determine the relative stocks and target cells, with and without the addition of contribution of each (Table 4). At a constant concentra- dendrimers (Table 2). Ecotropic retrovirus (by itself) tion of dendrimers (9 ␮g/mL), GPϩE86 cell supernatant had effects on murine NIH 3T3 cells and 3T3-derived had a virus dose-dependent effect on PA317 cells. Den- drimers also had a dose-dependent effect when viral Table 1. Enhancement of Fusogenic Effects of Ectropic supernatant was held constant and dendrimers were Retrovirus by Cationic Reagents on PA317 Cells varied. The data indicated that both ecotropic retrovirus and dendrimers played important contributory roles in ϩ GP E86 the fusion process. supernatant (mL/well) Cationic reagents Fusion

1.0 Dendrimer G9 (9 ␮g/well) ϩϩϩ Fusogenic effects of fresh versus frozen ecotropic 1.0 Lipofectamine (8 ␮l/well) ϩϩϩ supernatants on PA317 cells in the presence 1.0 Polybrene (8 ␮g/well) ϩϩ of G9 dendrimer Ϫ ϩ 1.0 ( ) Interestingly, it appeared that the freeze-thawed eco- 0 Dendrimer G9 (9 ␮g/well) Ϫ tropic supernatant had a stronger fusogenic effect than

Cancer Gene Therapy, Vol 7, No 1, 2000 56 XU ET AL: FUSOGENIC EFFECTS OF RETROVIRAL HELPER CELLS

Table 2. Fusogenic Effects of Retroviral Supernatants from Different Cell Lines in the Presence of Dendrimer Molecules

Recipient cells Supernatant Treatment (1 mL/well) (9 ␮g/well) GP ϩ E86 PA317 PG13 NIH 3T3 HT1080 HTam1

GPE ϩ 86 G9 ϩϩ ϩϩϩ ϩϩϩ ϩϩϩ Ϫ Ϫ Control ϩϩϩϩϪϪ PA317 G9 ϩ Ϫ ϩϩ ϩϩ Ϫ Ϫ Control ϪϪϪϪϪϪ PG13 G9 ϪϪϪϪϪϪ Control ϪϪϪϪϪϪ 3T3 G9 ϪϪϪϪϪϪ Control ϪϪϪϪϪϪ HT1080 G9 ϪϪϪϪϪϪ Control ϪϪϪϪϪϪ HT1080/am G9 ϪϪϪϪϪϪ Control ϪϪϪϪϪϪ fresh supernatant from the same cells. The results of one DISCUSSION such experiment are shown in Table 5. The effect was most apparent at low doses of supernatant. At doses of Retroviral helper cells are often used in cell complemen- 50–300 ␮L/well, the fresh supernatant induced syncytia tation procedures, either to move vectors between dif- involving Ͻ5–10% nuclei. The frozen supernatant ferent helper cells or to increase the vector copy number caused 10–75% syncytia. Even without G9 dendrimer, (ping-pong).20 In addition, murine cells or mouse animal 500 ␮L of the frozen supernatant could induce 15% models are often used in gene therapy research. The syncytia. results are relevant to these areas, but may also have importance with regard to understanding retroviral Course of the fusion process pathogenicity or increasing the sensitivity of retroviral We examined the morphological changes taking place detection assays. during the fusion process (Fig 2). Typically, the fusion Our studies indicated that the retrovirus particles process was complete in Ͻ9 hours (Table 6). In extreme produced in murine helper cells sometimes had the cases, such as that shown in Figure 1, the cellular ability to induce fusion in murine ﬁbroblast cells. In contents of a tissue culture plate appeared fused into a particular, ecotropic virus was more fusogenic than giant cell, with aggregating pools of ϳ10–30 nuclei each. amphotropic virus. GALV env did not appear fusogenic to human ﬁbroblasts. Polycationic molecules dramati- Table 3. Fusion of PA317 Cells at Different pH Values cally enhanced the fusion process mediated by murine retroviruses, and the fusion-enhancing effects of polyca- pH of medium for PA317 cells before transduction tions were proportional to the relative titer-enhancing Supernatant pH capabilities of the individual polycations (G9 of GP ϩ E86 cells 7.25 6.50 6.00 dendrimer3Lipofectamine3polybrene). This result suggested that the polycations acted by increasing con- 7.25 ϩϩϩ ϩϩϩ ϩϩϩ 6.50 ϩϩϩ ϩϩϩ ϩϩϩ tact between the virus and recipient cells. However, 6.00 ϩϩϩ ϩϩϩ ϩϩϩ susceptible human cell types were not fused by polyca- No supernatant ϪϪϪtion-treated amphotropic retrovirus, suggesting that the recipient cells played an important role in determining susceptibility to fusion effects. Furthermore, the ampho- Table 4. Syncytia Formation Dosage Effects of Ecotropic tropic virus produced in human cells was not fusogenic Retrovirus and Dendrimers on Recipient PA317 Cells

Treatment Table 5. Syncytia Induced in Recipient PA317 Cells by Fresh and Frozen Ecotropic, G9 Dendrimer-Treated Supernatant (% ϩ ␮ GP E86 supernatant (mL) G9 dendrimer ( g) Fusion Nuclei in Syncytia) ϩϩϩ 1.00 9 Ecotropic supernatant (mL/well) Fresh Frozen 0.50 9 ϩϩ 0.20 9 ϩϩ 1.00 ϩϩϩ ϩϩϩ 0.05 9 ϩ 0.50 ϩϩϩ ϩϩϩ 0.00 9 Ϫ 0.30 ϩϩ ϩϩϩ 0.10 ϩϩϩ 1.00 9 ϩϩϩ 0.05 Ϫϩ 1.00 6 ϩϩ 0 ϪϪ 1.00 3 ϩ 0.50 (no G9) Ϫϩϩ 1.00 0 Ϫ 10% FBS DMEM ϪϪ

Cancer Gene Therapy, Vol 7, No 1, 2000 XU ET AL: FUSOGENIC EFFECTS OF RETROVIRAL HELPER CELLS 57

together, all of these results supported the hypothesis that retroviral env proteins contribute to the cell mem- branes of target cells, turning them into virus-like particles with fusion activity. For example, polycations were necessary for maximal effect, and their effect was proportional to their relative ability to enhance retroviral infections. This result supported other data indicating that the amount of virus infecting the cells was directly proportional to the magnitude of the fusogenic effect. Second, fusion was observed regardless of whether a helper cell line or NIH 3T3 cells were used as targets for infection. This finding suggested that env proteins (re- cycled from the incoming virus) were responsible for the fusion activity. However, we did not rule out the possibility that murine cells already had some env proteins expressed on their surfaces that could contribute to the overall effect. Third, fusion was decreased in target helper cells exposed to the homologous retrovirus (GPϩE86 or PA317), indicating that viral resistance to superinfection prevented some (ecotropic) or all (amphotropic) of the fusion activity. Surprisingly, although ecotropic helper cells by themselves did not appear to fuse under the influence of their own supernatants, they did fuse in the presence of incoming ecotropic virus if the virus was first treated with G9 dendrimers. This suggested that the polycation was able to force a small amount of the highly fusogenic virus back into donor cells. An important consideration is whether or not the viruses used for human gene therapy are fusogenic in Figure 2. Time course of fusion of PA317 cells by GPϩE86 supernatant and G9 dendrimers. A: PA317 cells before addition of human cells. Fortunately, neither amphotropic viruses supernatant. B: PA317 cells at 30 minutes after the addition of 1 mL or GALVs were fusogenic in the human cells tested; in of GPϩE86 supernatant containing 9 ␮g/mL of G9 dendrimers. C: addition, amphotropic virus particles made in human PA317 cells at 60 minutes after the addition of 1 mL of GPϩE86 helper cells were not fusogenic toward murine cells. supernatant containing 9 ␮g/mL of G9 dendrimers. D: PA317 cells These results indicated that an important role is played at 120 minutes after the addition of 1 mL of GPϩE86 supernatant by host cells in the mediation of cell fusion. The lack of containing 9 ␮g/mL of G9 dendrimers. fusion activity associated with the use of human (donor or recipient) cells is viewed as a positive sign for human in murine or human cells. Taken together, these results gene therapy, although we did not rule out the possibility suggested a role for both the VPC and recipient cell that some human cells could be susceptible to fusion by types. murine retroviruses or VPCs. Most of the fusion activity Our results were in agreement with previous work was associated with ecotropic retrovirus, which is not showing that fusion is associated with the mature form of directly used in human gene therapy. TM8 and with labeling studies demonstrating persistence In conclusion, cell fusion effects are associated with of this integral membrane protein p12 (TM) in target the use of ecotropic3amphotropic murine retroviruses cells for Ն18 hours after retrovirus infection.21 Taken in conjunction with murine target cells. The effects are exacerbated by the use of polycationic enhancers of Table 6. Time Course for the Fusion of PA317 Cells by transduction. Human target cells have not been affected Ecotropic Supernatant and Dendrimer G9 thus far.

Time (minutes) Process Fused nuclei

0 Normal Ϫ Ϫ 10 Start ACKNOWLEDGMENTS 20 ϩ 30 ϩϩ 40 ϩϩ We thank Alan Rein (National Cancer Institute) for helpful 50 ϩϩϩ advice and Kris Hodgson for art work. This work was sup- 60 ϩϩϩ ported in part by grants from The National Institute of General 90 ϩϩϩ* Medical Sciences (2R55 GM41314-09, 1R43 GM58361), The State of Nebraska Smoking and Tobacco-Related Diseases *100% fused. Program, and Transgene, SA (Strasbourg, France).

Cancer Gene Therapy, Vol 7, No 1, 2000 58 XU ET AL: FUSOGENIC EFFECTS OF RETROVIRAL HELPER CELLS

REFERENCES 10. Anderson KB, Skov H. Retrovirus-induced cell fusion is enhanced by protease treatment. J Gen Virol. 1989;70: 1. Toyoshima K, Vogt P. Enhancement and inhibition of 1921–1927. avian sarcoma viruses by polycations and polyanions. Vi- 11. Pinter A, Chen TE, Lowy A, et al. Ecotropic murine rology. 1969;38:414–426. leukemia virus-induced fusion of murine cells. J Virol. 2. Hodgson CP, Solaiman F. Virosomes: cationic liposomes 1986;57:1048–1054. enhance retroviral transduction. Nat Biotechnol. 1996;14: 12. Zarling DA, Keshet I. Fusion activity of virions of murine 339–342. leukemia virus. Virology. 1979;95:185–195. 3. Hodgson CP, Solaiman F, Zink MA, et al. Semi-synthetic 13. Miller AD, Buttimore C. Redesign of retrovirus packaging systems for gene delivery. In: Kabanov AV, Felgner PL, cell lines to avoid recombination leading to helper virus Seymour LW, eds. Self-Assembling Complexes for Gene production. Mol Cell Biol. 1986;6:2895–2902. Delivery: From Laboratory to Clinical Trial. London: John 14. Rasheed S, Nelson-Rees WA, Toth EM, et al. Character- ization of a newly derived human sarcoma cell line Wiley and Sons, Ltd.; 1998:373–385. (HT1080). Cancer. 1974;33:1027–1033. 4. Klement V, Powe WP, Hartley JW, et al. Mixed culture 15. Miller AD, Garcia JV, von Suhr N, et al. Construction and cytopathogenicity: a new test for growth of murine leuke- properties of retrovirus packaging cells based on gibbon Proc Natl Acad Sci USA. mia virus in tissue culture. ape leukemia virus. J Virol. 1991;65:2220–2224. 1969;63:753–758. 16. Perucho M, Goldfarb M, Shimizu K, et al. Human tumor- 5. Nagy K, Clapham P, Cheingsong-Popov R, et al. Human T derived cell lines contain common and different transform- cell leukemia virus type I: induction of syncytia and ing genes. Cell. 1981;27:467–476. inhibition by patients sera. Int J Cancer. 1983;32:321–328. 17. Markowitz D, Goff S, Bank A. A safe packaging line for 6. Lifson JD, Feinberg MB, Reyes GR, et al. Induction of gene transfer: separating viral genes on two different CD4-dependent cell fusion by the HTLV-III/LAV enve- plasmids. J Virol. 1988;862:1120–1124. lope glycoprotein. Nature. 1986;323:725–728. 18. Price J, Turner D, Cepko C. Lineage analysis in the 7. Ragheb JA, Anderson WF. pH-independent murine leu- vertebrate nervous system by retrovirus-mediated gene kemia virus ecotropic envelope-mediated cell fusion: im- transfer. Proc Natl Acad Sci USA. 1987;84:156–160. plications for the role of the R peptide and p12E TM in 19. McClure MO, Sommerfelt MA, Marsh M, et al. The pH viral entry. J Virol. 1994;68:3220–3231. independence of mammalian retrovirus infection. J Gen 8. Rein A, Nirro J, Haynes JG, et al. Function of the Virol. 1990;71:767–773. cytoplasmic domain of a retroviral transmembrane pro- 20. Bestwick RK, Kozak SL, Kabat D. Overcoming interfer- tein: p15E-p2E cleavage activates the membrane fusion ence to retroviral superinfection results in ampliﬁed ex- capability of the murine leukemia virus env protein. J Virol. pression and transmission of cloned genes. Proc Natl Acad 1994;68:1773–1781. Sci USA. 1988;85:5404–5408. 9. Redmond S, Peters G, Dickson C. Mouse mammary tumor 21. Andersen KB. The fate of the surface protein gp70 during virus can mediate fusion at reduced pH. Virology. 1984; entry of retrovirus into mouse ﬁbroblasts. Virology. 1985; 133:393–402. 142:112–120.

Cancer Gene Therapy, Vol 7, No 1, 2000