Retrovirus vectors designed for efficient transduction of ... - Nature

Gene Therapy (1999) 6, 1670–1678  1999 Stockton Press All rights reserved 0969-7128/99 $15.00 http://www.stockton-press.co.uk/gt

Retrovirus vectors designed for efficient transduction of cytotoxic or cytostatic genes M Ui1, M Takada1, T Arai1,2, K Matsumoto1, K Yamada3, T Nakahata3, T Nishiwaki4, Y Furukawa4, T Tokino4, Y Nakamura4 and H Iba1 1

Department of Gene Regulation 3Department of Clinical Oncology and 4Laboratory of Molecular Medicine, Institute of Medical Science, University of Tokyo, Tokyo; and 2Tsukuba Research Laboratories, Eisai Co, Ltd, Tsukuba, Japan

It is difficult to establish stable packaging cell lines producing retrovirus vectors for the expression of anti-oncogenes with cytotoxic or cytostatic potential, because these genes would also affect the growth of the packaging cell lines. To overcome this problem, we designed a transcriptional unit pBabeLPL for vector RNA production, in which the transcription of the exogenous genes is completely suppressed by the presence of a preceding insertion containing the puromycin resistance gene (puro) and a poly(A) addition signal. This insertion is flanked by a tandem pair of loxP, and is designed to be excised after the introduction of Cre recombinase, when transcription of the exogenous gene will be started from the 5⬘-LTR. The transcriptional unit car-

rying LacZ or p53 as the exogenous gene was introduced into a previously constructed prepackaging cell lines PtGS2, in which the expression of VSV-G is also designed to be initiated by the introduction of Cre recombinase, while the gag-pol gene is expressed continuously. After the introduction of Cre recombinase by an adenovirus vector, LacZor p53-expressing VSV-G-pseudotyped retrovirus vectors with the designed structure were produced at high virus titers. The p53 virus was shown to be able to transduce p53 into the entire population of several human cancer cell lines and to induce their growth arrest at the G1 phase, indicating that this vector-producing system will be advantageous for human gene therapy.

Keywords: retrovirus vector; Cre-loxP; p53; cell cycle arrest; VSV-G-pseudotypes

Introduction Retrovirus vectors based on murine leukemia virus (MLV) have been used as powerful gene delivery systems in basic research and human gene therapy.1–3 The introduced genes are stably integrated into the target cell genome and have the potential for long-term expression. Recently, retrovirus vectors pseudotyped with the G protein of vesicular stomatitis virus (VSV-G) have been constructed and shown to be advantageous in that they have a much broader host range, often yielding higher transduction efficiency, than conventional amphotropic retrovirus vectors and that virus stocks can be concentrated by ultracentrifugation to give a very high titer (approximately 1 × 109 IU/ml).4–7 Efficient production systems for these vectors have recently been developed by us8 and by other groups,9–11 making use of the CreloxP-mediated recombination system and inducible promoters, respectively. Retrovirus vectors that can efficiently transduce antioncogenes or tumor suppressor genes with cytotoxic or cytostatic potential could be useful for gene therapy of human cancer. Such vectors might be produced by transient transfection of the vector DNA into packaging cell lines, but these systems are not suitable for the reproducible preparation of certified vectors on a large scale. Correspondence: H Iba, Department of Gene Regulation, Institute of Medical Science, University of Tokyo, 4–6–1 Shirokanedai, Minato-ku, Tokyo 108–8639, Japan Received 12 April 1999; accepted 8 June 1999

Establishment of stable packaging cell lines for such retrovirus vectors is difficult, because the growth-suppressing activity of these genes would cause problems in isolation of stable packaging cellular clones and their propagation in large-scale culture. Furthermore, the propagation step may potentially result in the accumulation of cellular populations carrying loss-of-function mutants of the tumor suppressor genes because of their growth advantage. Such difficulties would lower the quality and safety of the vectors, as discussed previously for an amphotropic retrovirus vector encoding p53.12 To overcome these problems, we have designed a transcriptional unit for vector RNA production, pBabeLPL, in which the expression of the cytotoxic genes is completely silenced during cloning and propagation of the packaging cell lines. When this transcriptional unit was introduced into a prepackaging cell line PtG-S2, which we had previously developed,8 the expression of both the exogenous gene in the vector and the VSV-G gene was induced in an all-or-nothing manner upon the introduction of Cre recombinase in an adenovirus vector.13,14 We show here that pBabeLPL in combination with PtG-S2 functioned as designed and was able to produce high titer virus stocks of VSV-G-pseudotyped retrovirus vector encoding p53. The virus stock transmitted the p53 gene into the entire population of several human tumor cell lines in a single transduction and induced cell cycle arrest of these cultures.

Retrovirus vectors for cytotoxic genes M Ui et al

1671

Figure 1 (a) Schematic presentation of the system designed for the production of VSV-G-pseudotyped retrovirus vectors encoding cytotoxic or cytostatic genes. The transcriptional unit pBabeLPL-X contains both the puro gene and a poly(A) addition signal flanked with loxP sequences which are tandemly located. Before the introduction of Cre recombinase, the puro gene is transcribed from the vector 5⬘-LTR, while the X gene is completely silent because the RNA transcript terminates before its coding sequence. Arrows indicate the predicted transcript in each cell type. Cre recombinase excises the puro gene and poly(A) addition signal by site-specific recombination between the two loxP sequences, producing a proviral structure carrying the X gene with a loxP. This enzyme also excises the other insert carrying the neo gene, an mRNA-destabilizing signal and a poly(A) addition signal in the transcriptional unit, pCALNdLG and induces the transcription of the VSV-G gene, which initiates VSV-G-pseudotyped vector production. MoLTR, MoMLV long terminal repeat; ␺, packaging signal of retrovirus vector; puro, puromycin resistance gene; pA, polyadenylation signal; X gene, an arbitary gene; CAG, CAG promoter; neo, neomycin resistance gene; VSV-G, VSV-G-coding sequence; MoMLV gag-pol, MoMLV gag and pol genes; bsr, blasticidin resistance gene. (b, c) Predicted structural changes in pBabeLPL-lacZ (b) or pBabeLPL-p53 (c) before and after the introduction of Cre recombinase. The predicted lengths of the restriction fragments used for the chromosomal analysis are shown. K, KpnI sites; X, XbaI sites.


1672

Results Design of retrovirus vectors for the expression of cytotoxic or cytostatic genes To overcome the difficulties in generating packaging cell lines which produce retrovirus vectors expressing cytotoxic or cytostatic genes, we first designed a transcriptional unit pBabeLPL-X (Figure 1a) for vector RNA production. While it has a usual MLV-based proviral DNA structure, the exogenous gene X is preceded by an insertion carrying the puromycin resistance gene (puro) and a poly(A) addition signal flanked by a tandem pair of loxP sequences. The RNA transcript driven by the 5⬘-LTR would be expected to encode only the puro gene and to be terminated by the poly(A) addition signal present just before the X gene. Therefore, the X gene should not be transcribed from the 5⬘-LTR or from any other promoters, even when pBabeLPL-X is introduced into cells. In cells harboring pBabeLPL-X, the introduction of Cre recombinase will induce excision of the insertion (both the puro gene and the poly(A) addition signal) through loxP-specific recombination and the transcription of the X gene from the 5⬘-LTR will be initiated. As the virus-producing cell system, we used the prepackaging cell line PtG-S2, which we had previously developed.8 PtG-S2 constitutively expresses the gag and pol genes of MoMLV and also contains a Cre recombinase-inducible transcriptional unit, pCALNdLG, for the VSV-G gene (Figure 1a). Neither RCR nor adenovirus vector was detected in pseudotyped virus stocks of 1 × 107 IU produced from PtG-S2.8 After the introduction of an adenovirus vector encoding Cre recombinase, PtGS2 harboring pBabeLPL-X would be expected to begin transcription of both full-length vector RNA driven by the 5⬘-LTR and VSV-G mRNA driven by a CAG promoter.15 Thereafter, virus vector carrying the X gene will be packaged and produced from the packaging cell line just generated from the prepackaging cell line. To develop this vector-producing system, we first used nlslacZ, the ␤-galactosidase gene with a nuclear localization signal, as a control gene (hereafter abbreviated as LacZ) and basic conditions for virus production, as well as the characteristics of the viral structure, were examined in PtG-S2 harboring pBabeLPL-lacZ (Figure 1b). Using the conditions and procedures established for these cells, we further applied this system to a representative tumor suppressor gene p53, with cytostatic and cytotoxic potential, by generating PtG-S2 harboring pBabeLPL-p53 (Figure 1c). Isolation of prepackaging cell lines from which LacZ virus production can be induced pBabeLPL-lacZ (Figure 1b) was transfected into PtG-S2 by lipofection and stable transformants were selected with puromycin. When mixed populations of puromycinresistant clones were collected and grown, no LacZexpressing cells were detectable in the entire culture, indicating that the LacZ gene is initially silent, as designed. When a subculture of the mixed populations was infected with the adenovirus vector AxCANCre13,14 at an MOI of 10, cells expressing LacZ became detectable within 3 days. From a parallel subculture, culture fluids were recovered every day and virus production was determined by titrating the LacZ-expressing retrovirus vectors with rat fibroblast 3Y1 cells as the indicator. The

mixed populations of these transfectants of PtG-S2 produced 2 × 104 IU/ml of LacZ-expressing retrovirus vectors around 5 to 7 days after the introduction of Cre recombinase (Figure 2a). The induction kinetics of the vectors were slightly delayed compared with those of PtG-S2 harboring pMFGnlslacZ (PtG-S2 lacZ1) and the highest titer was about 1/50 of that of PtG-S2 lacZ1. The slight delay in the induction kinetics might be partially explained by the fact that two independent Cre-loxP recombination events are required for the virus production in PtG-S2 harboring pBabeLPL-lacZ. In order to improve the production of retrovirus

Figure 2 Time-course of VSV-G-pseudotyped retrovirus production after transduction with AxCANCre at an MOI of 10. (a) Induction kinetics of virus carrying LacZ from cellular clones of PtG-S2 harboring pBabeLPLlacZ; A48 (closed circles), B19 (open triangles) or from the parental mixed population (closed squares). (b) Induction kinetics of virus carrying p53 from cellular clones of PtG-S2 harboring pBabeLPL-p53; C1 (open triangles), C58 (closed circles), C72 (closed squares) and D18 (open squares). None of these cultures produced retrovirus vectors in the absence of Cre recombinase. Results from A48 ((a); open circles) and C58 ((b); open circles) that were kept without the introduction of AxCANCre are shown. Cells were kept at 37°C until 2 days after the transduction and were shifted to 32°C thereafter.


vectors, we isolated 182 clones of stable transformants of pBabeLPL-lacZ after selection with puromycin. None of the clones expressed LacZ or produced LacZ-expressing retrovirus vectors before the introduction of Cre recombinase, as observed above. A subculture of each clone was infected with AxCANCre at an MOI of 10 and the induction of LacZ expression in the culture was examined 6 days later. Twenty clones showing high level induction as monitored by X-gal staining were selected, and shown to produce LacZ with titers more than 2 × 104 IU/ml on day 6. Among them, two clones (A48 and B19) produced high titer retrovirus vectors (A48; 4 × 105 IU/ml) with similar kinetics of virus induction to mixed populations (Figure 2a). After high levels of vector production (6 or 7 days after adenovirus infection), the virus titer declined gradually possibly due to the cytotoxic effects of VSV-G protein. Experiments using different MOIs of AxCANCre indicated that an MOI of 10 reproducibly gave the highest titers, and this condition was employed thereafter. Since the structural analysis of PtG-S2 harboring pBabeLPL-lacZ and of 3Y1 transduced with the stocks of the LacZ virus indicated that DNA recombination and virus production proceeded precisely as designed (see below), we next applied this system for p53 virus production.

Isolation of prepackaging cell lines from which p53 virus production can be induced To produce p53-expressing retrovirus vectors, we constructed pBabeLPL-p53 (Figure 1c), transfected it into PtG-S2, and isolated 168 clones of stable transformants by puromycin selection. Parallel cultures of these clones were infected with AxCANCre at an MOI of 10, and culture fluids were collected 6 and 8 days after the introduction of Cre recombinase. The titers of the virus stocks were determined by transduction on to 3Y1 and counting of p53-producing cellular clones after immunocytochemical staining. Twelve cellular clones that produced virus stocks with higher titers than 1 × 104 IU/ml were selected; none produced p53-expressing retrovirus vectors before the introduction of Cre recombinase. We further selected four clones (C1, C58, C72 and D18) on the basis of reproducibility in virus induction and high virus-producing activity (C58; 4 × 104 IU/ml), though the induction kinetics of the vectors were variable among them (Figure 2b). These virus stocks could easily be concentrated by ultracentrifugation to 1 × 108 IU/ml.

Figure 3 Proviral DNA analysis by Southern blotting. (a) Chromosomal DNA was prepared from A48 before (−) or 5 days after (+) the introduction of Cre recombinase. As a control, chromosomal DNA of PtG-S2 lacZ1 (lacZ1), known to harbor four copies of proviral DNA per diploid, was also prepared. DNA samples were digested with KpnI, separated on 1.0% agarose gel, analyzed by Southern blotting using the LacZ probe shown in Figure 1b (10 ␮g per lane) and visualized by autoradiography. (b) Chromosomal DNA was prepared from C1, C58, C72, D18 and the parental PtG-S2 before (−) and 4 days after (+) the introduction of Cre recombinase, doubly digested with XbaI and KpnI and analyzed by Southern blotting using the p53 probe shown in Figure 1c, top panel (10 ␮g per lane). The same sets of chromosomal DNA were designed with KpnI and analyzed by Southern blotting using the gag probe shown in Figure 1c, bottom panel. (c) Autoradiogram of Southern blotting of KpnI digests of chromosomal DNA isolated from 3Y1(−) or virus-transduced 3Y1(+) (10 ␮g per lane). The genomic DNA was hybridized to the LacZ probe (left panel) or the p53 probe (right panel).

Proviral DNA analysis confirmed virus production as designed To test whether the loxP-specific recombination occurred in pBabeLPL as designed, proviral DNA in PtG-S2 chromosome was analyzed by Southern blotting using a LacZ

1673


1674

probe before or 5 days after the introduction of AxCANCre (Figure 3a). A single 5.9-kb band was detected in the KpnI digests before the introduction of Cre recombinase, but the density of the band was drastically reduced and a new 4.8-kb band appeared instead after the introduction. This change in the size of the LacZ fragment is consistent with the expected structural change accompanying Cre recombinase-mediated recombination between the two loxP sequences in pBabeLPLlacZ, as illustrated in Figure 1b. By comparing the density of chromosomal DNA with that of control chromosomal DNA from PtG-S2 lacZ18 harboring four copies of MFGnlslacZ per diploid, A48 was shown to contain a single copy of pBabeLPL-lacZ in the diploid cell before the introduction of Cre recombinase. After the introduction, the copy number increased to two to three copies per diploid, as shown in Figure 3a. This result would indicate that the LacZ virus has propagated in the packaging cell lines by 5 days after the introduction of Cre recombinase, because VSV-G-pseudotyped virus was previously shown to be fully infectious to VSV-G-expressing cells such as PtG-S2 into which Cre recombinase had been introduced.8 We next analyzed chromosomal DNA of C1, C58, C72 and D18 by Southern blotting to examine the structural changes in pBabeLPL-p53. Chromosomal DNA was isolated before or 4 days after the introduction of AxCANCre. The samples were doubly digested with KpnI and XbaI and analyzed with a p53 probe, or digested with KpnI and analyzed with a gag probe (Figure 1c). The size changes observed by the p53 probe (2.5 kb to 3.4 kb) or by the gag probe (4.6 kb to 3.5 kb) are consistent with the idea that Cre recombinase had excised the insertion between the two loxP sequences by site-specific recombination. A comparison of the densities of the bands indicates that the copy numbers were drastically increased by the introduction of Cre recombinase in all the clones (Figure 3b). In C1, we additionally detected an unexpected 3.3-kb band with either the p53 or gag probe before the introduction of Cre recombinase (Figure 3b, top and bottom panels), and found that it disappeared after the introduction. This result suggests that the C1 chromosome contains one copy of pBabeLPL-p53 which has an altered structure from that in the original plasmid and we eliminated this clone from further analysis. By chromosomal analysis of clone A48, or C58, we also showed that in pCALNdLG the loxP-specific recombination was almost completed by 4 days after the introduction of Cre recombinase (data not shown), as has also been shown for the parental strain PtG-S2. These chromosomal changes are consistent with the kinetics of LacZ virus or p53 virus production; the virus production was highest from 4 days after the introduction of Cre recombinase. To examine the integrated structures of transduced retrovirus vectors, LacZ virus or p53 virus was transduced into 3Y1 at an MOI of 10 or 3, respectively. Chromosomal DNA was isolated before or 3 days after the transduction, digested with KpnI, and analyzed by Southern blotting using LacZ or p53 probe. A single 4.8kb band was detected in the LacZ virus-transduced cells and a 3.5-kb band was detected in the p53 virus-transduced cells (Figure 3c). These bands were not detected before the virus transduction. Therefore only retrovirus

vectors with the designed structure were integrated into the target cells.

Transduction of the p53 virus into several cell lines and its biological effects To test the biological activity of the p53-carrying virus prepared here, a pair of growing cultures of either 3Y1 (rat fibroblast) or U2-OS (originated from human osteosarcoma) was prepared and transduced with the p53 virus at an MOI of 10. In both cell lines, the entire populations of the transduced culture were shown to express p53 protein at high levels within 65 h as judged by immunocytochemical staining using anti-p53 (human) antiserum (Figure 4 right panels) and no cells were unstained. The cell numbers of transduced cultures were much lower than those of cultures left untransduced, indicating that cell growth was strongly inhibited in the transduced cultures. It was also clear that all the cells transduced with the p53 virus exhibited a much flatter morphology than the untransduced cells (Figure 4, left panels). Endogenous p53 expression in 3Y1 was not detectable because the antiserum is non-crossreactive to rat p53, while low level expression of endogenous p53 was detected in untransduced U2-OS. When the amounts of human p53 protein in these cells were analyzed by Western blotting using the same antip53 (human) antiserum, strong induction of p53 protein was confirmed. The analysis of parallel cultures transduced with p53 virus at an MOI of 1 indicated that the p53 expression levels increased in a dose-dependent manner and further that the p53 expression level induced by UV irradiation was roughly equivalent to that induced by retrovirus transduction at an MOI of 1 (Figure 5). To assess the effects of p53 virus on cell growth more precisely, these two cell lines as well as PtG-S2, the parental prepackaging cell line originated from the human fibrosarcoma cell line, HT1080, were transduced with this virus at an MOI of 10 and analyzed by flow cytometry after propidium iodide staining. All three cell lines transduced with p53 virus showed a marked decrease in the

Figure 4 Transduction of p53 vector into the entire cellular population. p53 vector was transduced into 3Y1 or U2-OS at an MOI of 10. Cells were fixed before and 3 days after the transduction and observed under a Normarski’s differential interference microscope after immunocytochemical staining using monoclonal anti-p53 (human) IgG. The bar indicates 100 ␮m.


Discussion

Figure 5 p53 expression in cells before and 3 days after the transduction of p53 virus at an MOI of 1 or 10. The parallel culture of logarithmically growing U2-OS was UV-irradiated using a 254-nm germicidal lamp at the dose of 20 J/m2 14 h before preparation. Cell lysates were prepared under denaturing conditions, separated by SDS-PAGE (15 ␮g per lane) and detected by Western blotting using murine monoclonal anti-p53 (human) IgG.

Table 1 Effects of p53 virus or control MFGnlslacZ on the cell cycle distribution Cell line

3Y1 U2-OS PtG-S2

Virus

LacZ p53 LacZ p53 LacZ p53

Cell cycle distribution (%) G0-G1

S

G2-M

27.73 50.26 36.82 59.64 31.01 59.08

45.67 30.06 43.47 19.70 24.66 9.63

26.60 19.68 19.71 20.66 44.33 31.29

proportion of cells in S phase and a significant increase in the proportion of cells in G1 phase (Figure 6, Table 1), when compared with control cultures transduced with LacZ virus at an MOI of 10, as well as untransduced cultures (data not shown). These results indicate the expression of p53 results in the G1 arrest of these cell lines at the G1-S transition.

We have presented here a unique system for the production of retrovirus vectors carrying cytotoxic or cytostatic genes. pBabeLPL can be employed as an inducible transcriptional unit for the vector RNA encoding an exogenous gene in an all-or-nothing manner using the Cre-loxP recombination system. We can select prepackaging cell lines without any production of vector RNA bearing the exogenous gene during the selection and subsequent passaging procedures. Even before the introduction of Cre recombinase, a part of the vector RNA with the puromycin resistance gene is transcribed from the 5⬘-LTR of pBabeLPL and terminated by the SV 40 poly(A) addition signal. In the culture medium of PtGS2 derivatives such as A48, B19, C1, C58, C72 and D18, however, we have never detected any vector particles transmitting puro resistancy to 3Y1 (our unpublished result) either before or after the introduction of Cre recombinase. While the Cre-loxP recombination system was used here for the induction of the cytotoxic exogenous genes, it has also been used in the context of retrovirus-mediated gene transfer for the generation of selfdeleting vectors,16,17 for the removal of proviral marker genes,18–20 and for the exchange of the exogenous genes in a packaging cell line.21 In our system, induction of both the X gene in pBabeLPL-X and the VSV-G gene in pCALNdLG is mediated through Cre-loxP recombination. While site-specific recombination by Cre recombinase has been reported to be quite rare if the two loxP sequences are distantly located in the same chromosome or located in different chromosome,22 it is theoretically possible for the site-specific recombination to occur between the loxP sequences in pBabeLPL-X and pCALNdLG. However, we did not detect any of the bands that would be expected from such recombinations (Figure 3b), even in an autoradiogram of the same gel exposed for 10 times longer (our unpublished results). Such recombinations would generate a structure, 5⬘-LTR-loxP-VSV-G-poly(A) addition signal, which could be further converted to produce packageable and transmissible transcripts after a single non-homolo-

Figure 6 Cell cycle analysis of 3Y1, U2-OS and PtG-S2 transduced with either p53 virus or control MFGnlslacZ at an MOI of 10. Cells were harvested 60 h after the transduction with p53 virus or control MFGnlslacZ, stained with propidium iodide and analyzed by flow cytometry.

1675


1676

gous recombination. We have carefully checked whether such particles that can transmit the VSV-G gene were present in our virus stocks, but we were not able to detect such particles at all in 5 × 106 IU of the vectors. Therefore, such recombination was negligible, at least at 5–6 days after the introduction of Cre recombinase. The proviral DNA analysis further confirmed that only DNA forms with the designed structure were transmittable to the transductants. Using this retrovirus production system, we have succeeded in generating stable packaging cell lines which can produce high-titer p53-expressing VSV-G-pseudotyped retrovirus vector. When this vector was transduced into 3Y1, U2-OS and PtG-S2 at an MOI of 10, exogenous p53 expression was induced in all the cells at quite high levels and essentially none of the cells in the entire culture escaped transduction. Therefore this system should permit us to examine the effect of introducing cytotoxic genes into an entire cellular population without any prior selection. Since the transduction of the p53-carrying virus into the parental prepackaging cell line, PtG-S2, induced cell cycle arrest (Figure 6, Table 1), a stable cell line which constitutively expresses p53 should not be established. Due to its widespread alterations in malignant cells and its central position in both proliferation arrest and apoptosis induction, p53 has been the most investigated gene for cancer gene therapy.23–28 In most experiments, gene transfer efficiencies have been below 100%,29 so the antitumoral effect of p53 relies strongly on ‘by-stander killing’ effects.24,27,30,31 In contrast by using the VSV-Gpseudotyped retrovirus vectors developed here, it should be possible to transduce exogenous genes into the entire population of monolayer cultures originated from human cell lines (Figure 4a), even if the exogenous genes are cytostatic or cytotoxic, like p53.

Materials and methods Plasmid construction The 1.0-kb SaII–ClaI fragment (carrying an internal promoter and the puro gene) of pBabe puro32 (a kind gift from Dr H Land, Imperial Cancer Research Fund) was deleted and replaced with the linker oligonucleotides (5⬘ TCGACGCAGATCTCACGTGATTTAAATAT 3⬘ and 5⬘ CGATATTTAAATCACGTGAGATCTGCG 3⬘) to generate pBabe. The 1.1-kb HindIII–BamHI fragment carrying the puro gene and SV 40 poly(A) addition signal was excised from pPUR (Gibco-BRL, Rockville, MD, USA) and ligated into the 2.5-kb HindIII–BamHI fragment of pBS246 (Gibco-BRL) that is flanked with loxP sequences at both ends. The 1.2-kb loxP-puro-SV 40 poly(A)-loxP fragment was excised from the generated plasmid by digesting with both EcoRI and ScaI, blunt-ended with Klenow fragment, and inserted into the unique SnaBI site of pBabe to generate pBabeLPL. The BamHI fragment encoding the entire nlslacZ or p53 was excised from pMFGnlslacZ33 or pCMV-NEO-BAM-p5323 and inserted into the unique BgIII site of pBabeLPL to generate pBabeLPL-lacZ and pBabeLPL-p53, respectively. Cell lines The prepackaging cell line PtG-S28 (a derivative of human fibrosarcoma cell line HT1080, which carries the gag and pol genes from Moloney MLV (MoMLV) and a

VSV-G gene (Indiana serotype) that is silent before the introduction of Cre recombinase), 3Y1 (rat fibroblast) and U2-OS (human osteosarcoma) were maintained in Dulbecco’s modified Eagle’s medium (DMEM) (high glucose) supplemented with 10% fetal calf serum and kept at 37°C. PtG-S2 and its derivatives were grown in the presence of 4 ␮g of blasticidin S (Funakoshi, Tokyo, Japan) and 1 mg of G418 (Gibco-BRL) per milliliter. Drug selection for the transfected PtG-S2 was performed with 2 ␮g of puromycin (Sigma, St Louis, MO, USA) per ml.

DNA transfection and cloning of the transfectants PtG-S2 was seeded at 2 × 106 cells/100-mm diameter dish, kept for 1 day and transfected with either pBabeLPL-lacZ or pBabeLPL-p53 (4 ␮g) by the use of Lipofectamine Plus Reagent (Gibco-BRL). Two days after the transfection, the cultures were split at several ratios, and puromycin (final concentration, 2 ␮g/ml) was added to the medium 3 or 4 days after the transfection to select stable transfectants. Puromycin-resistant colonies were picked up with cloning cylinders and propagated. Adenovirus vector infection for pseudotyped retrovirus production PtG-S2 and its derivatives were kept at 37°C, seeded at 7.5 × 105 cells per 100-mm diameter dish, kept for 1 day and infected with AxCANCre at a multiplicity of infection (MOI) of 10. G418 and puromycin were removed from the culture medium just before the AxCANCre infection. From two days after the adenovirus infection, the cultures were kept at 32°C to obtain a slightly more stable production of VSV-G-pseudotypes. The medium was changed every day and at each medium change, retrovirus vector stocks were recovered. Southern blot analysis Total chromosomal DNA was prepared using standard techniques from PtG-S2, 3Y1 and their derivatives. The digested DNA was separated on 1.0% agarose gel and transferred to a nylon membrane (Hybond N+; Amersham, Buckinghamshire, UK) by the capillary transfer method. The 3.1-kb LacZ probe was isolated from pMFGnlslacZ by BamHI digestion. The 1.8-kb p53 probe was isolated from pCMV-NEO-BAM-p53 by BamHI digestion. The 0.7-kb gag probe was isolated from pBabe by SpeI and BamHI digestion. The probes were labeled with 32P-␣-CTP by using the RTS RadPrime DNA Labeling System (Gibco-BRL), hybridized and detected by autoradiography. Protein analysis For Western blot analysis, cellular lysates were prepared under denaturing conditions and separated by sodium dodecyl sulfate-10% polyacrylamide gel electrophoresis. The gels were transferred on to polyvinylidene difluoride membrane (Immobilon; Millipore, Bedford, MA, USA) with a semi-dry electroblotter. Filters were immunoblotted first with monoclonal anti-p53 (human) IgG (Pab 1801; Santa Cruz, Santa Cruz, CA, USA) that is non-crossreactive to either mouse or rat p53 and then with antimouse IgG horseradish peroxidase-linked whole antibody (Amersham). Protein bands were detected with an ECL kit (Amersham). For immunocytochemical staining, cells were fixed with PBS containing 2% paraformaldehyde and 0.1% Triton X-100 at 4°C for 15 min and


reacted with monoclonal anti-p53 (human) IgG (DO-1; Santa Cruz), which is non-crossreactive to either mouse or rat p53. p53-producing cells were visualized using biotinylated anti-mouse IgG and a Vectastain ABC kit (Vector, Burlingame, CA, USA).

Retrovirus transduction and titration For the titration of VSV-G-pseudotyped retrovirus vectors, the indicator cell line 3Y1 was plated at 7.5 × 102– 1.5 × 103 cells per well in 96-well plates and kept for 1 day before transduction. For transduction, cells were incubated with serial dilutions of virus supernatants in the presence of 8 ␮g/ml polybrene (Sigma) and kept for a further 3 days. For X-gal staining transduced 3Y1 was fixed with 1.25% glutaraldehyde and stained with 5 mm K4[Fe(CN)6], 5 mm K3[Fe(CN)6], 2 mm MgCl2 and 1 mg/ml X-gal (5-bromo-4-chloro-3-indolyl-␤-d-galactopyranoside) (Wako, Osaka, Japan) for more than 4 h. The numbers of cellular clones with blue-stained nuclei were counted. Expression of p53 (human) protein was evaluated by immunocytochemical staining using monoclonal anti-p53 (human) IgG (DO-1; Santa Cruz) and visualized as described above. The numbers of cellular clones with brown-stained nuclei were counted. Cell cycle analysis Cells were trypsinized, fixed with 90% ethanol and resuspended in PBS solution containing 2 mg/ml of Rnase A (Sigma) and 50 ␮g/ml of propidium iodide (Sigma). The stained cells were analyzed in a fluorescence-activated cell sorter (FACScan; Becton Dickinson, Franklin Lakes, NJ, USA). The percentages of cells in various cell cycle phases were determined by using the CellQuest and ModFit program.

Acknowledgements We thank Yuriko Yoshikawa for assistance in preparing the manuscript. This work was supported in part by grants and endowments from Eisai Co, Ltd, and by a Grant-in-Aid for Scientific Research on Priority Areas from the Ministry of Education, Science and Culture, Japan.

References 1 Miller AD. Human gene therapy comes of age. Nature 1992; 357: 455–460. 2 Mulligan RC. The basic science of gene therapy. Science 1993; 260: 926–932. 3 Vile R, Russell SJ. Gene transfer technologies for the gene therapy of cancer. Gene Therapy 1994; 1: 88–98. 4 Burns JC et al. Vesicular stomatitis virus G glycoprotein pseudotyped retroviral vectors: concentration to very high titer and efficient gene transfer into mammalian and nonmammalian cells. Proc Natl Acad Sci USA 1993; 90: 8033–8037. 5 Yee JK et al. A general method for the generation of high-titer, pantropic retroviral vectors: highly efficient infection of primary hepatocytes. Proc Natl Acad Sci USA 1994; 91: 9564–9568. 6 Matsubara T et al. Pantropic retroviral vectors integrate and express in cells of the malaria mosquito, Anopheles gambiae. Proc Natl Acad Sci USA 1996; 93: 6181–6185. 7 Sharma S, Cantwell M, Kipps TJ, Friedmann T. Efficient infection of a human T cell line and of human primary peripheral blood leukocytes with a pseudotyped retrovirus vector. Proc Natl Acad Sci USA 1996; 93: 11842–11847.

8 Arai T et al. A new system for stringent, high-titer vesicular stomatitis virus G protein-pseudotyped retrovirus vector induction by introduction of Cre recombinase into stable prepackaging cell lines. J Virol 1998; 72: 1115–1121. 9 Yang Y et al. Inducible, high-level production of infectious murine leukemia retroviral vector particles pseudotyped with vesicular stomatitis virus G envelope protein. Hum Gene Ther 1995; 6: 1203–1213. 10 Chen ST et al. Generation of packaging cell lines for pseudotyped retroviral vectors of the G protein of vesicular stomatitis virus by using a modified tetracycline inducible system. Proc Natl Acad Sci USA 1996; 93: 10057–10062. 11 Ory DS, Neugeboren BA, Mulligan RC. A stable human-derived packaging cell line for production of high titer retrovirus/vesicular stomatitis virus G pseudotypes. Proc Natl Acad Sci USA 1996; 93: 11400–11406. 12 Estreicher A, Iggo R, Roth JA. Retrovirus-mediated p53 gene therapy. Nature Med 1996; 2: 1163. 13 Kanegae Y et al. Efficient gene activation in mammalian cells by using recombinant adenovirus expressing site-specific Cre recombinase. Nucleic Acids Res 1995; 23: 3816–3821. 14 Kanegae Y et al. Efficient gene activation system on mammalian cell chromosomes using recombinant adenovirus producing Cre recombinase. Gene 1996; 181: 207–212. 15 Niwa H, Yamamura K, Miyazaki J. Efficient selection for highexpression transfectants with a novel eukaryotic vector. Gene 1991; 108: 193–200. 16 Choulika A, Guyot V, Nicolas JF. Transfer of single gene-containing long terminal repeats into the genome of mammalian cells by a retroviral vector carrying the cre gene and the loxP site. J Virol 1996; 70: 1792–1798. 17 Russ AP, Friedel C, Grez M, von Melchner H. Self-deleting retrovirus vectors for gene therapy. J Virol 1996; 70: 4927–4932. 18 Bergemann J et al. Excision of specific DNA sequences from integrated retroviral vectors via site-specific recombination. Nucleic Acids Res 1995; 23: 4451–4456. 19 Fernex C, Dubreuil P, Mannoni P, Bagnis C. Cre/loxP-mediated excision of a neomycin resistance expression unit from an integrated retroviral vector increases long terminal repeat-driven transcription in human hematopoietic cells. J Virol 1997; 71: 7533–7540. 20 Wildner O et al. Generation of a conditionally neor-containing retroviral producer cell line: effects of neor on retroviral titer and transgene expression. Gene Therapy 1998; 5: 684–691. 21 Vanin EF et al. Development of high-titer retroviral producer cell lines by using Cre-mediated recombination. J Virol 1997; 71: 7820–7826. 22 van Deursen J, Fornerdo M, van Rees B, Grosveld G. Cremediated site- specific translocation between nonhomologous mouse chromosomes. Proc Natl Acad Sci USA 1995; 92: 7376– 7380. 23 Baker SJ et al. Suppression of human colorectal carcinoma cell growth by wild-type p53. Science 1990; 249: 912–915. 24 Cai DW et al. Stable expression of the wild-type p53 gene in human lung cancer cells after retrovirus-mediated gene transfer. Hum Gene Ther 1993; 4: 617–624. 25 Fujiwara T et al. A retroviral wild-type p53 expression vector penetrates human lung cancer spheroids and inhibits growth by inducing apoptosis. Cancer Res 1993; 53: 4129–4133. 26 Fujiwara T et al. Therapeutic effect of a retroviral wild-type p53 expression vector in an orthotopic lung cancer model. J Natl Cancer Inst 1994; 86: 1458–1462. 27 Xu M et al. Parenteral gene therapy with p53 inhibits human breast tumors in vivo through a bystander mechanism without evidence of toxicity. Hum Gene Ther 1997; 8: 177–185. 28 Favrot M, Coll JL, Louis N, Negoescu A. Cell death and cancer: replacement of apoptotic genes and inactivation of death suppressor genes in therapy. Gene Therapy 1998; 5: 728–739. 29 Polyak K et al. Genetic determinants of p53-induced apoptosis and growth arrest. Genes Dev 1996; 10: 1945–1952.

1677


1678

30 Freeman SM et al. The ‘bystander effect’: tumor regression when a fraction of the tumor mass is genetically modified. Cancer Res 1993; 53: 5274–5283. 31 Mesnil M et al. Bystander killing of cancer cells by herpes simplex virus thymidine kinase gene is mediated by connexins. Proc Natl Acad Sci USA 1996; 93: 1831–1835. 32 Morgenstern JP, Land H. Advanced mammalian gene transfer:

high titre retroviral vectors with multiple drug selection markers and a complementary helper-free packaging cell line. Nucleic Acids Res 1990; 18: 3587–3596. 33 Dranoff G et al. Vaccination with irradiated tumor cells engineered to secrete murine granulocyte–macrophage colony-stimulating factor stimulates potent, specific, and long-lasting antitumor immunity. Proc Natl Acad Sci USA 1993; 90: 3539–3543.