Preclinical Evaluation ofWholeCell Vaccines for Prophylaxis and

Document Sample
Preclinical Evaluation ofWholeCell Vaccines for Prophylaxis and Powered By Docstoc
					[CANCER RESEARCH 60, 1663–1670, March 15, 2000]

Preclinical Evaluation of “Whole” Cell Vaccines for Prophylaxis and Therapy Using a Disabled Infectious Single Cycle-Herpes Simplex Virus Vector to Transduce Cytokine Genes1
S. A. Ali, C. S. McLean, M. E. G. Boursnell, G. Martin, C. L. Holmes, S. Reeder, C. Entwisle, D. M. Blakeley, J. G. Shields, S. Todryk, R. Vile, R. A. Robins, and R. C. Rees2
Department of Life Sciences, Nottingham Trent University, Nottingham NG11 8NS, United Kingdom [S. A. A., S. R., R. C. R.]; Cantab Pharmaceuticals Research Limited, Cambridge CB4 4GN, United Kingdom [C. S. M., M. E. G. B., G. M., C. L. H., E. E., D. M. B., J. G. S.]; Imperial Cancer Research Fund Laboratory of Molecular Therapy, Imperial College of Science and Medicine, Hammersmith Hospital, London W12 0NN, United Kingdom [S. T., R. V.]; Division of Immunology, Queens Medical Centre, Nottingham, United Kingdom [R. A. R.]; and Molecular Medicine Program, Mayo Clinic, Rochester, Minnesota 55905 [R. V.]

ABSTRACT
The development of genetically modified “whole” tumor cell vaccines for cancer therapy relies on the efficient transduction and expression of genes by vectors. In the present study, we have used a disabled infectious single cycle-herpes simplex virus 2 (DISC-HSV-2) vector constructed to express cytokine or marker genes upon infection. DISC-HSV-2 is able to infect a wide range of tumor cells and efficiently express the -galactosidase reporter gene, granulocyte-macrophage colony-stimulating factor (GM-CSF), or IL-2 genes. Gene expression occurred rapidly after infection of tumor cells, and the level of production of the gene product ( -galactosidase, GM-CSF, or IL-2) was shown to be both time-and dose-dependent. Vaccination with irradiated DISC-mGM-CSF or DISChIL-2-infected murine tumor cells resulted in greatly enhanced immunity to tumor challenge with live parental tumor cells compared with control vaccines. When used therapeutically to treat existing tumors, vaccination with irradiated DISC-mGM-CSF-infected tumor cells significantly reduced the incidence and growth rates of tumors when administered locally adjacent to the tumor site, providing up to 90% protection. The prophylactic and therapeutic efficacy of DISC-mGM-CSF-infected cells was shown initially using a murine renal cell carcinoma model (RENCA), and the results were confirmed in two additional murine tumor models: the M3 melanoma and 302R sarcoma. Therapy with DISC-infected RENCA “whole” cell vaccines failed to reduce the incidence or growth of tumor in congenitally T-cell deficient (Nu /Nu ) mice or mice depleted of CD4 and/or CD8 T-lymphocytes, confirming that both T-helper and T-cytotoxic effector arms of the immune response are required to promote tumor rejection. These preclinical results suggest that this “novel” DISCHSV vector may prove to be efficacious in developing genetically modified whole-cell vaccines for clinical use.

INTRODUCTION Cancer vaccination strategies have focused on the use of autologous and allogeneic tumor cells genetically modified to express a range of different immunomodulatory genes which include cytokines, costimulatory molecules, and tumor antigens. Studies using animal models have shown that inoculation/immunization with tumor cells engineered to express IL3-2, IFN- , IL-4, tumor necrosis factor , GMCSF, IL-7, or IL-6 enhances antitumor immunity (1– 6). In most tumor models, this results not only in the rejection of the genetically modified tumor cells but also the induction of systemic immunity capable
Received 8/23/99; accepted 12/16/99. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. 1 Funded by Cantab Pharmaceuticals. 2 To whom requests for reprints should be addressed, at Department of Life Sciences, Nottingham Trent University, Clifton Lane, Nottingham NG11 8NS, United Kingdom. 3 The abbreviations used are: IL, interleukin; hIL, human IL; CSF, colony-stimulating factor; GM-CSF, granulocyte-macrophage CSF; mGM-CSF, murine GM-CSF; HSV, herpes simplex virus; DISC, disabled infectious single cycle; pfu, plaque-forming unit(s); HSP, heat shock protein; BrdUrd, bromodeoxyuridine; CMV, cytomegalovirus; -gal, -galactosidase; MOI, multiplicity of infection; TD50, tumor dose 50%; Mab, monoclonal antibody; gH, glycoprotein H.

of mediating the rejection of a subsequent challenge with parental, unmodified tumor cells. A major drawback and limitation in using autologous cellular vaccines to treat cancer patients is the need to establish in vitro tumor cell lines prepared from biopsy tumor tissue for the transduction of immuno-modulatory genes. Difficulties associated with establishing cell lines from human tumor biopsy material and the relative inefficiency of many of the transfection methodologies have led to renewed efforts to establish alternative strategies for the efficient delivery of genes into freshly prepared/isolated tumor cells. Vectors that efficiently deliver genes into tumor cells either in vivo or ex vivo are required, and several viral and nonviral vector systems have been investigated for their suitability in this regard (7). Viral vectors represent the most efficient means of transducing genes into tumor cells, and many replication-competent and replication-defective viruses have been used to deliver genes of interest to in vitro and in vivo targets. HSVs have been used recently for cancer therapy and gene transduction studies. Intratumoral injection of replication-competent attenuated mutants of HSV-1 were shown to be effective in killing malignant gliomas (8, 9), and Toda et al. (10) have reported recently that immunization with a defective HSV-1 vector encoding the IL-12 gene in combination with a HSV helper virus can induce local and systemic antitumor immunity to the CT26 murine colon carcinoma. Similarly, systemic therapy using a recombinant adenovirus encoding both subunits of IL-12 inhibited the formation of 3-day hepatic metastasis of murine tumors (11). We have reported previously the development of a genetically inactivated HSV-2 vector that is restricted to a single cycle of replication, DISC, for use as a vaccine against genital herpes infection (12–14). We have used this virus to deliver cytokine genes to tumor cells, and there are several reasons why this vector is potentially useful for cancer immunotherapy: (a) DISC-HSV-2 is unable to spread from cell to cell; replication of the virus is genetically restricted by deletion of the gH gene, which is essential for the production of infectious progeny; and (b) HSV-2 has a broad host cell range, making the DISC variant an appropriate vehicle for the delivery of genes to a variety of tumors. In addition, DISC-HSV infects nondividing as well as dividing cells and has been shown to rapidly and efficiently infect primary human leukemia and neuroblastoma cells (15), human carcinoma cells (16), and cultured murine tumor cells (17). In the present study, the DISC-HSV-2 vector has been used to deliver genes encoding murine GM-CSF, human IL-2, or the lacZ reporter genes to murine tumor cells and the efficacy of DISCHSV-2-infected “whole” tumor cell vaccines for prophylactic immunization prior to tumor challenge and for the therapy evaluated in three murine tumor models. The results show that DISC-HSV-2 is an efficient vector for cytokine gene delivery into tumor cells, and that the expression of mGM-CSF or hIL-2 enhances the immunogenicity of whole-cell vaccines. In this study, the therapeutic response was

1663

PRECLINICAL EVALUATION OF WHOLE-CELL VACCINES

shown to be depend on the functionality of CD4 phocytes.

and CD8

lym-

MATERIALS AND METHODS
Tumors. RENCA-3 is a BALB/c renal carcinoma cell line of spontaneous origin and was generously provided by Dr. Robert Wiltrout (National Cancer Institute, Bethesda, MD). The immunogenicity of RENCA has been determined as being low to moderate. RENCA-3 cells were maintained by serial in vitro passage in RPMI 1640 supplemented with 10% FCS, sodium pyruvate, and NEAA. The M3 cell line is a DBA/2 melanoma cell line that was obtained from American Type Culture Collection. M3 cells were grown and maintained in Hams F-12 media supplemented with 15% FCS. The 302R is C57Bl/6 mouse sarcoma cell line, derived through repeated in vivo passage in mice, and was kindly supplied by Dr. B. Fox (Portland, OR). 302R cells were grown and maintained in DMEM media supplemented with 10% FCS. Animals. Female DBA/2, C57Bl/6, BALB/c, and BALB/c Nu /Nu mice were purchased from Harlan (UK) Ltd. and were maintained in accordance with the Home Office Codes of Practice for housing and care of animals. Infection of Tumor cells with DISC-HSV lacZ Virus. Tumor cells were either cultured on glass slides precoated with fibronectin to increase the attachment of the cells or in 24-well plates at 1 105 cells/well and infected with DISC-HSV lacZ virus at a MOI of 1.25–10 pfu/cell. At various times postinfection, the cells were either fixed in acetone and stained for the presence of HSV-2 antigen using a polyclonal anti-HSV-2 antibody (Dako) or fixed in glutaraldehyde and stained for -gal by 5-bromo-4-chloro-3-indolyl- -D-galactopyranoside staining (Promega). Cytokine Assays. Expression of cytokines after infection of tumor cells with DISC-mGM-CSF and DISC-hIL-2 was determined by ELISA (R&D Systems, United Kingdom). Tumor cells were cultured in 24-well plates at a concentration of 1 105 cells/well overnight. The medium was removed, and cells were infected with 1.25–10 pfu/cell of each virus in a total volume of 100 –200 l for 1 h at 37°C. The medium was removed and replaced by 1 ml of serum-free medium, and the plates were reincubated at 37°C for various times, up to 48 h. Supernatants were collected and stored at 20°C and assayed for mGM-CSF or hIL-2. Apoptosis and HSP Expression. To determine whether DISC-HSV infection of tumor cells induced cell death by apoptosis, 5 105 cells were cultured in T25 flasks and infected with DISC lacZ virus at 10 pfu/cell for 24 h. Floating and adherent (trypsinized) cells were pooled together for analysis. An ABO-BrdUrd kit from PharMingen (San Diego, CA) was used according to the manufacturer’s instructions. Briefly, the cells were prefixed in 1% paraformaldehyde and then stored in 70% ethanol for 24 h. The cells were then washed and incubated with terminal deoxynucleotidyl transferase and bromo-dUTP, followed by FITC-anti- BrdUrd and propidium iodide. DNA breaks are indicative of apoptosis. Prophylactic Immunization and Therapy with DISC-HSV-infected Cells. Tumor cells were infected with 5–10 pfu/cell with either DISC-mGMCSF, DISC-hIL-2, or DISC-lacZ viruses for 1 h. The virus inoculum was removed, and cells were washed two times in medium. Fresh serum-free medium was added, and the cells were then irradiated (15,000 rads) using a Gammacell cesium-137 source; uninfected tumor cells were prepared in similar manner and used as control. Cells (1 106) cells infected with DISCmGM-CSF or DISC-hIL-2 viruses (before and after irradiation) were cultured in 24-well plates to assess cytokine production (as detailed above). To assess Fig. 1. Infectability of DISC-lacZ and reporter gene expression in RENCA cells. A, the effect of prophylactic vaccination using RENCA cells, animals were 4 immunized s.c. two times on the right flank at 2-week intervals with irradiated, 2 10 RENCA cells were cultured in wells on fibronectin-coated glass microscope slides (MIC3412; Scientific Laboratory Supplies Ltd., Nottingham, United Kingdom) in noninfected RENCA cells or irradiated, DISC-infected RENCA cells in a 100 l of culture medium at 37°C/5% CO for 3 h. The attached cells were then infected 2 volume of 200 l (see individual experiments for details). Unless otherwise with 0.5, 1, 5, and 10 pfu/cell of DISC-lacZ virus for 3 h. The cells were then fixed and stated, animals were challenged s.c 7 days after the second inoculation, with stained for both -gal and HSV-2 antigen expression and examined by light and fluorescent microscopy to determine the percentage of stained cells. Bars, SD. B, 1 105 cells 5 104 (10 times TD50) parental RENCA cells on the opposite flank. were cultured in 24-well plates overnight. Culture medium was removed, and the virus To assess the efficacy of DISC-HSV infected whole-cell vaccines in ther- inoculum was added for 1 h (experiments were performed in triplicate). Residual virus 4 apy, mice received injections in the right flank with 5 10 tumor cells and was removed, and fresh serum-free medium was added, and the cells were reincubated for were vaccinated with 1 106 DISC-HSV infected-irradiated RENCA cells at 24 h. Thereafter, the cells were fixed by adding 500 l of 1% glutaraldehyde solution and the same site or contralaterally on day 0 or 3. Mice then received two stained for -gal expression. At least 100 cells were evaluated, and the percentage of stained cells was determined. Bars, SD. C, experiments were performed as outlined in B. additional immunizations at 3-day intervals. Similar protocols using the M3 RENCA cells were stained for -gal expression at time intervals of 0.5 to 24 h after melanoma and 302R sarcoma were used to confirm the findings obtained with infection with DISC-lacZ virus at a MOI of 10 pfu. the RENCA model. 1664

PRECLINICAL EVALUATION OF WHOLE-CELL VACCINES

Construction of the DISC-HSV Viruses. Construction of the basic vector DISC-HSV-2 (DISC-HSV) by plasmid recombination was described previously (17). A similar process was used to construct dH2B (DISC-mGM-CSF), which required a two-stage recombination strategy. For the first stage, sodium iodide-purified, wild-type DNA and plasmid DNA (pIMMB56) were transfected into gH expressing complementing CR1 cells. The plasmid pIMMB56 contains the lacZ gene under the control of the SV40 promoter; the expression cassette is flanked by HSV sequences to enable recombination into viral genome and is similar in construction to pIMMB47 (17). The resulting virus is designated dH2D. For the second stage, sodium iodide-purified dH2B viral DNA and plasmid DNA pIMR3 were transfected into CR1 cells as described above. Plasmid pIMR3 was constructed by ligation of the mGM-CSF gene from plasmid pJL3.2 (received as a gift; Ref. 18) into the shuttle vector pIMMB46. Plasmid pIMMB46 had been adapted previously to contain the CMV promoter and bovine growth hormone poly(A) addition signal from the plasmid PPRC/CMV (R&D Systems). The resulting virus was passaged three times on BHK gH /TK cells in the presence of methotrexate to select for TK virus. The final virus was designate dH2B. dH2J (DISC hIL-2) was constructed by ligation. This process is simpler than traditional recombinant techniques and yields a high frequency of recombinant viruses. A linker sequence containing a unique PacI restriction site was engineered into the basic vector during construction to allow subsequent manipulation. dH2J was constructed by a ligation process similar to basic plasmid construction. The IL-2 gene was excised from the plasmid BBG30 (R&D systems); the plasmid was engineered to construct the native signal sequence and insert a Kosak consensus sequence upstream of the gene. The modified IL-2 gene was ligated into the “Pac ligation” vector (pIMJ2), which contains two PacI sites, before insertion into the virus. The vector pIMJ2 was digested with PacI to release a fragment containing the modified IL-2 gene downstream of the CMV promoter. The expression cassette was ligated with PacI-digested HSV in a similar process to basic plasmid construction. The ligated DNA was transfected into CR2 cells, and the resulting virus was designated DH2J. Depletion of CD4 and CD8 Cells. The effect of the in vivo depletion of CD4 or CD8 T-cells or both CD4 and CD8 T-cells on the therapeutic efficacy of the DISC vaccine was investigated. Groups of 10 mice were given three i.p. injections of 1 mg of anti-CD4 (YTS191.1.2), anti-CD8 (YTS 169.4.2.1), control isotype antibody (YTH24), or a combination of anti-CD4 and anti-CD8 antibodies over a period of 1 week (19). Ten days after the last injection, three representative animals from each group were tail bled to determine the efficiency of depletion by flow cytometric analysis of the blood cells using anti-CD4 and anti-CD8 antibodies (Serotec Ltd. Oxford, United Kingdom). Mice received injections of RENCA cells (5 104 cells/mouse) 7 days after the last antibody injection. Three vaccinations, with irradiated DISC-infected RENCA cells, were given 3 days apart at an adjacent body site, commencing on day 3 after the injection of tumor cells.

Fig. 2. Time course for the production of mGM-CSF and hIL-2 by cells infected with DISC-mGM-CSF and DISC-hIL-2 viruses. RENCA cells were cultured in 24-well plates overnight and infected with 10 pfu/cell of DISC-mGM-CSF virus (A) and DISC-hIL-2 virus (B) for 1 h, washed twice in serum-free medium, and incubated for up to 72 h. Supernatant from individual wells was collected at each indicated time points and stored at 20°C for cytokine analysis. Bars, SD.

RESULTS Reporter and Cytokine Gene Expression in RENCA Cells. DISC-HSV-2 viruses have been constructed to express the genes for LacZ, mGM-CSF, and hIL-2, respectively, after infection. In vitro studies were performed to confirm the ability of these viruses to infect murine tumor cells and to express the gene of interest. Expression of the LacZ gene was observed in RENCA cells coexpressing HSV viral glycoproteins, as shown by dual staining for -gal and HSV protein expression. RENCA cells infected with DISC-LacZ (0.5–10 pfu) demonstrated an increase in immunostaining for the virus using antiHSV fluorescent-labeled antibody only in cells expressing the -gal gene (Fig. 1A). The infectability of RENCA cells using a MOI ranging from 0.3 to 10 pfu/cell was determined. A MOI of 1.25 pfu/cell resulted in 50% of the cells staining for -gal 24 h after infection, whereas a MOI of 2.5 pfu/cell or greater resulted in -gal protein expression in virtually all of the cells (Fig. 1B) 3 h after infection (Fig. 1C). Similar results were obtained for 302R and M3 cells (results not shown). RENCA cells infected with DISC-mGM-CSF released up to 400 pg

of m-GM-CSF/ml/105 cells in 24 h (Fig. 2A), which was the maximum amount detected in this time course. The mGM-CSF release from RENCA cells increased proportionally with virus MOI (0.3 to 10 pfu/cell). Similar results were obtained for RENCA cells infected with DISC-hIL-2 virus, with the maximum release occurring 48 h after infection (Fig. 2B). For RENCA cells infected with DISC-hIL-2 virus at a MOI of 10 pfu/cell, 548 pg/ml/105 cells of hIL-2 was released into the supernatant at 24 h. In comparison, the mouse sarcoma 302R cells released 440 pg/ml/105 cells 24 h after infection with 10pfu/cell of DISC-hIL2, and the melanoma cell line M3 released 600 pg/ml/105 cells at 24 h. Cytokine production and release into culture medium by DISC-HSV-infected cells peaked at 24 and 48 h for mGM-CSF and IL-2, respectively, and thereafter and by day 5 declined to negligible levels; this was demonstrated using transwells when culture medium was replaced with fresh medium at 24-h intervals (results not shown). The level of cytokine production by DISC-HSV-2-infected tumor cells is high in comparison to conventional gene transfection is transitional but of limited duration (72 h) because of the cytopathic effect of the virus. Because it was necessary to use irradiated tumor cells for in vivo vaccination studies, the effect of the irradiation on cytokine release was determined. The release of mGM-CSF varied, depending on the cell line. For RENCA cells, the release of mGM-CSF remained at the same level after irradiation (Table 1), whereas 302R cells showed an

1665

PRECLINICAL EVALUATION OF WHOLE-CELL VACCINES

Table 1 Effect of irradiation on cytokine production by RENCA cells infected with DISC mGM-CSF virus and DISC-IL-2 virus Experimenta Irradiationb Irradiation

A. mGM-CSF present in supernatant (pg/ml) 1 2 3 4 Average 244c 256 264 290 263.5 19.4 304 248 240 214 251.5 37.8

SD

B. hIL-2 present in supernatant (pg/ml) 1 2 3 Average
a

SD

1022 1036 1170 1076 81.7d

548 480 540 522.6 37.1

RENCA cells were infected with 10 pfu/cell of DISC-HSV-mGM-CSF virus or DISC-HSV-IL-2 virus for 1 h, and the cells were washed and exposed to irradiation. Cells (1 106) in 1 ml of serum-free medium were cultured in 24-well plates, and supernatants were collected after 24 h. b Irradiation (15,000 rads) of RENCA cells was performed 1 h after virus infection. c Cytokine release was measured by ELISA; assays were performed in duplicate (values are mean SD). d Significant at P 0.01.

increase in the amount of mGM-CSF release of 65% and M3 cells a decrease of 40% (results not shown). A significant (P 0.01) increase in hIL-2 release occurred after irradiation of DISCIL-2-infected RENCA cells (Table 1; mean of three experiments). Treatment of RENCA cells with the protein synthesis inhibitor cyclo-

heximide resulted in a 95% inhibition of the reporter (LacZ) and cytokine gene expression (results not shown), confirming that virus replication is essential for gene expression. Apoptosis and Heat Shock Protein Expression in RENCA Cells Infected with DISC-HSV Virus. Cell death by apoptosis was investigated in RENCA cells infected with the DISC-lacZ virus. RENCA cells infected with DISC-lacZ at a MOI of 10 pfu/cell for 24 h induced apoptotic cell death in 7.8% of all cells compared with 2.1% of noninfected control cells (Fig. 3). Analysis of the DNA profile of cells showed that infected cells contained more DNA than uninfected cells, indicative of a block in the S or G2-M phases of the cell cycle, together with an increase in sub-G1 cells, which is suggestive of an increase in necrotic cell death (Fig. 3). Furthermore, infecting RENCA cells with DISC-lacZ did not induce the expression of the HSP protein HSP70, as shown by quantitative mRNA expression (results not shown), inferring that virus infection does not induce the early onset of stress response proteins that has been reported previously to influence the immunogenicity of tumor cells (20). Prophylactic Immunization Using Irradiated Tumor Cells Infected with DISC-HSV. Having demonstrated that RENCA cells release mGM-CSF and hIL-2 after infection with DISC-mGM-CSF and DISC-hIL-2 viruses, respectively, experiments were performed to compare the immunogenicity of DISC-infected whole-cell vaccines with standard vaccines prepared from irradiated, uninfected RENCA cells. Cultured RENCA cells were infected with DISC-HSV-2-containing either the lacZ, mGM-CSF, or hIL-2 gene for 1 h, irradiated, and injected

Fig. 3. Necrosis versus apoptosis in RENCA cells infected with DISC-lacZ virus. RENCA cells (5 105) were seeded into T25 tissue culture flasks and infected with 10 pfu/cell of DISC-lacZ for 24 h. Floating and adherent (trypsinized) cells were pooled and analyzed using an APO-BrdUrd kit to determine the apoptosisindicative DNA breaks. The DNA profile (a and b) and DNA breaks (c and d) for noninfected (a and c) and infected (b and d) tumor cells are shown.

1666

PRECLINICAL EVALUATION OF WHOLE-CELL VACCINES

Fig. 4. Prophylactic immunization of DISCmGM-CSF-infected RENCA cells. RENCA cells were infected with 5 pfu/cell of either DISC-mGMCSF virus or DISC-lacZ ( -gal) virus for 1 h and irradiated (15,000 rads). Groups of up to 10 mice 105 irradiated were injected with 1 104 or 1 virus-infected or irradiated noninfected RENCA cells, and cells were injected s.c. on the right flank on two occasions 2 weeks apart; control mice were untreated. One week after the second immunization, the mice were challenged s.c with 5 104 (10 TD50) parental RENCA cells on the opposite flank, and tumor growth and incidence were monitored on a regular basis. Animals immunized with 104 cells and challenged with 5 104 RENCA cells (A), immunized with 105 cells and challenged with 5 104 RENCA cells (B), immunized with 104 cells and challenged with 105 RENCA cells (C), or immunized with 104 cells and challenged with 2 105 RENCA cells (D) are shown.

into groups of up to 10 mice. Two immunizations were performed at 2-week intervals, and the animals were challenged s.c. with 5 104 live RENCA cells (10 TD50) 7 days after the second immunization. Results representative of several experiments are shown in Fig. 4. The degree of protective immunity was increased after immunization with 1 104 DISC-infected cells expressing mGM-CSF compared with nonimmunized control mice or mice receiving irradiated RENCA cells or irradiated RENCA cells infected with DISC-lacZ (Fig. 4A). In addition, mice immunized with 1 105 DISC-mGM-CSF-infected RENCA cells showed an increased level of protection to tumor challenge compared with mice immunized with irradiated (noninfected) RENCA cells and control (nonimmunized) mice (Fig. 4B). It was observed that immunization with 1 105 irradiated RENCA cells infected with DISC-lacZ could also increase resistance to tumor challenge compared with nonimmunized control mice, although the degree of protective immunity was consistently lower than that observed for mice immunized with DISC-mGMCSF-infected cells. The level of protective immunity was decreased after immunization with whole-cell vaccines (1 104 cells/inoculum) and challenge with 20 TD50 or 40 TD50 RENCA cells (Fig. 4, C and D, respectively). Vaccination with DISC-hIL-2-infected RENCA cells gave results that were similar to those obtained using DISC-mGM-CSF-infected RENCA cells (results not shown). These experiments were repeated in a second tumor model. Prophylactic immunization with irradiated DISC- mGM-CSF- or DISC-hIL-2-infected 302R cells enhanced immunity to tumor challenge and confirmed the results obtained using the RENCA model (results not shown). Tumor Therapy Using DISC-HSV-infected Tumor Cells. The ability of DISC-infected tumor cells to influence the growth of established tumors was investigated in three tumor models: RENCA, 302R, and M3. Groups of 10 mice were injected s.c. on the right flank with

5 104 (10 TD50) viable RENCA cells prior to vaccination with irradiated DISC-infected or irradiated noninfected RENCA cells. Three immunizations (each containing 1 106 irradiated tumor cells) were given s.c. on the same or contralateral flank on days 3, 6, and 9, and the tumor incidence and tumor size were recorded during a 9-week observation period. By day 3 after inoculation of live tumor cells, defined tumor foci were detected by H&E histological analysis (results not shown), and for most of the experiment performed, this was the start date for initiation of therapy. The results (Fig. 5A) demonstrate a slight but insignificant delay in the onset of tumors in mice receiving the irradiated (noninfected) RENCA cell vaccine compared with control mice; this difference was not apparent when the average tumor sizes of the groups were compared (results not shown). Mice receiving DISC-mGM-CSFinfected RENCA cells showed a significant delay in the onset of tumors, and a high proportion of mice remained tumor free up to 9 weeks after challenge. In addition, vaccination with DISC-hIL2-infected cells significantly inhibited tumor growth; 60% of mice remained free of tumor throughout the observation period (Fig. 5A). These results were reproducible and occurred when immunization was initiated on day 0 or day 3 after tumor cell inoculation. Up to 80% of mice immunized with either DISC-mGM-CSF or DISC-hIL-2 vaccines remained free of tumor throughout the study period (data from several experiments, not shown). Similar results were obtained with the M3 melanoma (Fig. 5B) model. Vaccination, beginning on day 0 after live tumor cell implantation, with DISC-mGM-CSF or DISC– hIL-2-infected M3 cells inhibited tumor growth in 60 and 20% of vaccinated animals, respectively (Fig. 5C), and caused a delay in tumor growth in the remaining mice (results not shown). Inhibition of established M3 tumor growth also occurred in 40% of mice treated with DISC-lacZ M3-infected cells. One important feature of this immunotherapy model was the development of local immunity in vaccinated mice. Vaccination at a site

1667

PRECLINICAL EVALUATION OF WHOLE-CELL VACCINES

Fig. 5. Treatment of established RENCA tumors with DISC-infected cells. A and B, animals were implanted with either 5 104 RENCA (A) or 1 104 M3 (B) tumor cells on the right flank (RF). Three vaccinations (1 106 cells/inoculum; arrows) were given into the RF, 3 days apart, starting on day 3 and day 0 for RENCA and M3, respectively, with either irradiated noninfected (tumor cells) or irradiated tumor cells infected with either DISC-lacZ, DISC-mGM-CSF, or DISC-IL-2 virus. Tumor growth was monitored regularly throughout the observation period. C, animals were implanted with 5 104 RENCA cells on the right flank (RF). Three vaccination (1 106 cells/inoculum; arrows) were given 3 days apart, starting on day 3 with irradiated RENCA cells infected with DISC-mGM-CSF injected either on the right flank (DISC-mGM-CSF RHS) or on the left flank (DISC-mGM-CSF LHS). Tumor growth was monitored throughout the observation period.

adjacent to the tumor implantation site with the DISC-mGM-CSF RENCA cell vaccine was effective in delaying the onset and growth of tumors; however, immunization on the contralateral flank was less effective (Fig. 5C). These results demonstrate that an increased therapeutic benefit can be derived by local administration of tumor cells infected with DISC-HSV-2 engineered to express either mGM-CSF or hIL-2 and were confirmed using the 203R tumor model (results not given). Vaccine Therapy in T-Cell-deficient Mice. To determine that T-lymphocytes were required for effective immunotherapy using DISC whole tumor-cell vaccines, BALB/c nude mice (Nu /Nu ) received injections s.c. with 10 TD50 of RENCA cells on the right flank and vaccinated (starting on day 0) three times (3 days apart) on the same flank with irradiated noninfected RENCA cells or irradiated DISC-mGM-CSFinfected RENCA cells. The tumor incidence and growth rate were similar in control and vaccinated Nu /Nu mice, indicating that T lymphocytes play a pivotal role in promoting tumor rejection (results not given). To establish the involvement of CD4 and CD8 T lymphocytes in immunotherapy, mice were depleted of the respective T-cell populations by the administration of Mabs raised to either CD4 or CD8 antigens. Seven days after antibody treatment, mice were injected with 5 104 RENCA cells and vaccine therapy (irradiated RENCA cells infected with DISC-mGMCSF) given on days 3, 6, and 9. The results demonstrate that CD4, CD8, and CD4/CD8 “knock out” mice failed to respond to whole-cell vaccine therapy, whereas mice inoculated with isotype control serum or untreated mice were successfully treated by vaccination with DISCmGM-CSF-infected RENCA cells (Fig. 6). The abrogation of therapeutic

efficacy was proportional to the reduction in the subpopulations of T lymphocytes; the administration of CD8 Mab caused a 50% depletion of circulating CD8 T cells and abrogated the effect of immunotherapy in 60% of mice. Administration of CD4 Mab reduced circulating CD4 T cells by 95% and completely abrogated the effect of vaccine therapy. Collectively, these results demonstrate an absolute requirement for both CD4 and CD8 T lymphocytes for effective immunotherapy using DISC-mGM-CSF-infected whole-cell vaccination. DISCUSSION Murine tumor models have been used to evaluate whole tumor cell vaccines, genetically modified by gene transfection to produce cytokines, for their ability to promote anticancer immunity. These vaccines have been shown to elicit systemic immunity against tumor challenge and in some instances induce the regression of small tumors when given therapeutically (1– 6). Thus, a number of viral and nonviral vector systems have been investigated as vehicles for gene delivery into tumor cells (21). In the present study, a DISC-HSV-2 was used as a vector for gene transfection of tumor cells in preclinical studies to assess its potential for human application. We previously constructed gH-deleted HSV-2 to be used as a vaccine for the prevention of HSV-induced disease. This virus, which we term DISC, can only complete one replication cycle in normal cells and was shown to stimulate broad humoral and cell-mediated antiviral immune responses (12). DISC-HSV offers advantages as a vector system for gene transfer; they are safe because they are unable to spread from cell to cell

1668

PRECLINICAL EVALUATION OF WHOLE-CELL VACCINES

Fig. 6. Immunotherapy using a DISC-mGM-CSF-infected RENCA vaccine in mice depleted of CD4 or CD8 T lymphocytes. Treatment of mice with anti-CD4 monoclonal antibody, given 3 times i.p. over a period of 1 week, reduced the circulating CD4 lymphocytes by 95% (CD4 depl, CD4 depleted); anti-CD8 antibody treatment reduced the circulating CD8 lymphocyte population by 50% (CD8 depl, CD8 depleted); as determined by antibody staining and flow cytometry. Control mice were similarly inoculated with CD8 depleted). medium or with an isotype control antibody (isotype). A group of mice was also treated with both anti-CD4 and anti-CD8 antibody (CD4 CD8 depl, CD4 One week after depletion treatment, all mice were inoculated with RENCA cells (5 104 cells/animal) on the right flank and vaccinated (1 106cells/inoculum) on the right flank. Three vaccinations were given 3 days apart starting on day 3. mGM-CSF, mice inoculated with medium and vaccinated with DISC-mGM-CSF-infected RENCA cells; CD4 depl, CD4-depleted mice receiving vaccination; CD8 depl, CD8-depleted mice receiving vaccination; CD4 CD8 depl, mice depleted of CD4 and CD8 T lymphocytes receiving vaccination; isotype, mice inoculated with isotype control antibody and receiving vaccination; medium, mice inoculated with medium and not vaccinated. Tumor growth was monitored throughout the observation period.

within the patient, and they have a broad host cell range, making them suitable for the delivery of genes to a variety of tumors. Initial studies have shown that the DISC-HSV-2 will infect a wide range of murine and human tumor cells, including primary human leukemia and neuroblastoma cells (15, 16, 22). Here we show that DISC-HSV is able to infect murine carcinoma, sarcoma, and melanoma cells. We have used three murine tumors, RENCA, 302R, and M3, as models to assess the ability of DISC-HSV-2 to deliver cytokine genes into tumor cells and undertaken preclinical evaluation of whole-cell vaccines expressing the cytokines mGM-CSF or hIL-2 to assess their ability to promote protective and therapeutic immunity. The RENCA tumor model was used extensively in this study, and the results were confirmed using 302R and M3 tumors. The relationship between the expression of the reporter -gal gene and viral proteins after in vitro infection with the DISC-HSV-lacZ virus was established by dual staining for the expression HSV glycoprotein and the -gal protein. Infection and reporter gene expression were time and dose dependent. A correlation between -gal expression and the MOI was shown and confirmed that the virus was incapable of lateral spread. In a study by Lowstein et al. (23), recombinant HSV type 1 mutant tsk vectors containing -gal were shown to infect neurocortical cells; -gal expression was directly related with the MOI of the virus. Infecting RENCA cells with DISC-HSV-2 encoding mGM-CSF and recombinant hIL-2 genes resulted in the release of cytokines into the culture supernatant in a time-dependent manner. After infection with DISC-HSV- -gal, an increase in necrotic cell death versus apoptosis occurred. RENCA cells undergoing death by necrosis may provide addition activation of the immune system in vivo by promoting tumor antigen processing and presentation by professional antigen-presenting cells, leading to an increase in T-cell activation (24, 25). In situ killing of tumor cells using suicide gene transfer to induce cell death through a nonapoptotic pathway is associated with enhanced immunogenicity and may in some cases require the induction of HSP expression (20), although in the present study RENCA cells infected with DISC-HSV failed to show elevated expression of HSP. On the basis of these in vitro results and because of the potential of these cytokines to activate effector T cells (26), DISC-mGM-CSF and DISC-hIL-2 vectors were chosen for in vivo studies. In animal models, mGM-CSF expression by tumor cells results in potent systemic

antitumor immunity, which can potentiate the rejection of weakly immunogenic murine tumors (27). IL-2 is also a potent mediator of antitumor immunity and can promote CTL activation and T-cell differentiation, enhance the activation status of natural killer cells, and induce lymphokine-activated killer cell activity (28, 29). Interestingly, hIL-2 production and release by DISC-hIL-2-infected cells were significantly increased after irradiation, an effect observed previously by Simova et al. (30), where administration of irradiated IL-2-secreting plasmacytoma cells was shown to be more effective than nonirradiated cells in promoting tumor immunity. Here, we demonstrate that immunization with irradiated RENCA cells infected with DISC encoding either mGM-CSF or hIL-2 cytokine genes protects mice against challenge with parental tumor cells in a dose-related manner. DISC-mGM-CSF vaccine therapy prevented tumor growth in a high percentage of mice. The response to therapy was T-lymphocyte dependent and required the participation of both CD4 and CD8 T lymphocytes. One important consideration in this therapy model is the relative contribution of HSV infection versus cytokine production. Partial protection was observed after immunization with tumor cells infected with the DISC -gal virus (used as a control for cells expressing cytokine), indicating that protection against tumor challenge may, in part, be a consequence of viral infection of the tumor cells; immunization with the -gal protein alone does not illicit a measurable antitumor immune response (10). We suggest that DISC-HSV infection can act as an additional stimulus to enhance the immunogenicity of the tumor cells. HSV infection of mice has been shown to lead to the up-regulation of IL-12 expression (31) and to have potent antitumor effects in animal models (32), most probably by promoting a Th1 response to tumor antigen(s). Preexisting immunity to HSV infection did not seem to affect the efficacy of this vaccine because no significant difference was shown when animals were preimmunized with HSV prior to tumor implantation and subsequent therapy (data not shown). Irradiated RENCA cell vaccination also induced a degree of protection against rechallenge with the parental tumor line. These data are consistent with previous observations demonstrating that RENCA cells are weakly to moderately immunogenic (33), and where irradiation itself may affect the immunogenicity of tumor cells through the up-regulation of H-2Kd class I MHC antigens (34). For the reasons outlined, there is a precedent for using DISC-HSV to deliver immune response genes to tumor cells, in the present study by ex vivo

1669

PRECLINICAL EVALUATION OF WHOLE-CELL VACCINES

infection of the tumor cells, but additionally by direct in vivo injection into the tumor using a murine colon carcinoma model, where 40% of tumors regressed completely (16); we have shown that both approaches are efficacious for therapy in animal models. GM-CSF has been used to potentiate antitumor immunity by promoting the maturation and function of professional antigen-presenting cells (35) and by recruiting additional antigen-specific and nonspecific effector cells. One noticeable feature of the immune response after inoculation of whole-cell vaccines expressing GM-CSF is the prevalence of a delayedtype hypersensitivity response at the site of vaccination and at the site of tumor rejection. Infiltration of tumors by eosinophils in response to GM-CSF has been reported in preclinical and clinical studies (27, 36), and a similar response is also observed after immunization with IL-4 gene-transduced vaccines (37, 38). There is also evidence that patients treated with an autologous GM-CSF gene-transduced vaccine can undergo an objective clinical response (36), although it is unclear to what extent eosinophils and effector cells other than CD8 and CD4 lymphocytes actually contribute to tumor rejection. Eosinophil infiltration of small established RENCA tumors occurs within 24 h after vaccine therapy with irradiated DISC-mGM-CSF RENCA cells4 and may represent a response associated with the production of Th2 cytokines IL-4 and IL-5 (36). In conclusion, the results obtained in this study allow us to propose a clinical approach to cancer immunotherapy based on the use of a novel DISC-HSV vector for the efficient delivery of cytokine genes to tumor cells. ACKNOWLEDGMENTS
We are grateful to Glenda Kill for typing the manuscript.

13.

14.

15.

16.

17.

18.

19.

20.

21. 22.

23.

24. 25.

REFERENCES
1. Fearon, E. R., Pardoll, D. M., Itaya, T., Golumbek, P., Levitsky, H. I., Simons, J. W., Karasuyama, H., Vogelstein, B., and Frost, P. Interleukin-2 production by tumour cells bypasses the T helper function in the generation of an anti-Tumour response. Cell, 60: 397– 403, 1990. 2. Gansbacher, B., Zier, K., Daniels, B., Cronin, K., Bannerji, R., and Gilboa, E. Interleukin-2 gene transfer into tumour cells abrogates tumourigenicity and induces protective immunity. J. Exp. Med., 172: 1217–1224, 1990. 3. Watanabe, Y., Kuribayashi, K., Miyatake, S., Nishihara, K., Nakayama, E. I., Taniyama, T., and Sakata, T. A. Exogenous expression of mouse interferon cDNA in mouse neuroblastoma C1300 cells results in reduced tumorigenicity by augmented anti-tumor immunity. Proc. Natl. Acad. Sci. USA, 86: 9456 –9460, 1989. 4. Li, W., Diamantstein, T., and Blankestein, T. Lack of tumorigenicity of interleukin-4 autocrine growing cells seems related to the antitumor function of interleukin-4. Mol. Immunol., 27: 1331–1337, 1990. 5. Asher, A. L, Mule J. J., Kasid, A., Restifo, N. P., Salo, J. C., Reihert, C. M., Jaffee, G., Fendly, B., Kriegler, M., and Rosenberg, S. A. Murine tumor cells transduced with genes for tumor necrosis factor. J. Immunol., 146: 3227–3234, 1991. 6. Porgador, A., Tzehoval, E., Katz, A., Vadai, E., Revel, M., Feldman, M., and Eisenbach, L. Interleukin-6 gene transfection into Lewis lung carcinoma tumor cells suppresses the malignant phenotype and confers immunotherapeutic competence against parental metastatic cells. Cancer Res., 52: 3679 –3686, 1992. 7. Rakhmilevich, A. L., Turner, J., Ford, M. J., McCabe, D., Sun, W. H., Sondel, P. M., Grota, K., and Yang, N. S. Gene gun-mediated skin transfection with interleukin 12 gene results in regression of established primary and metastatic murine tumors. Proc. Natl. Acad Sci. USA, 93: 6291– 6296, 1996. 8. Mineta, T., Rabkin, S. D., Yazaki, T., Hunter, W. D., and Martuza, R. L. Attenutated multi-mutated herpes simplex virus-1 for the treatment of malignant gliomas. Nat. Med., 1: 938 –943, 1995. 9. Yazaki, T., Manz, H. J., Dubkin, S. D., and Martuza, R. L. Treatment of human malignant meningiomas by G207, a replication-competent multimutated herpes simplex virus. Cancer Res., 55: 4752– 4756, 1986. 10. Toda, M., Martuza, R. L., Kojima, H., and Rabkin, S. D. In situ cancer vaccination: an IL-12 defective vector/replication-competent herpes simplex virus combination induces local and systemic antitumour activity. J. Immunol., 160: 4457– 4464, 1998. 11. Siders, W. M., Wright, P. W., Hixon, J. A., Alvord, W. G., Back, T. C., Wiltrout, R. H., and Fenton, R. G. T cell- and NK cell-independent inhibition of hepatic metastases by systemic administration of an IL-12-expressing recombinant adenovirus. J. Immunol., 160: 5465–5474, 1998. 12. Boursnell, M. E. G., Entwisle, C., Blakeley, D., Roberts, C., Duncane, I. A., Chisholm, S. E., Martin, G. M., Jennings, R., Challanain, D. N., Sobek, I., Inglis,
4

26.

27.

28. 29.

30.

31.

32.

33.

34.

35. 36.

37.

38. Unpublished results.

S. C., and McLean, C. S. A genetically inactivated herpes simplex virus type 2 (HSV-2) vaccine provides effective protection against primary and recurrent HSV-2 disease. J. Infect. Dis., 175: 16 –25, 1997. Forrester, A., Farrell, H., Wilkinson, G., Kaye, J., Davispoynter, N., and Minoson, T. Construction and properties of a mutant of herpes-simplex virus type-1 with glycoprotein-h coding sequences deleted. J. Virol., 66: 341–348, 1992. McLean, C. S., Erturk, M., Jennings, R., Challanain, D. N., Minson, A. C., Duncane, I., Boursnell, M. E. G., and Inglis, S. C. Protective vaccination against primary and recurrent disease cause by herpes simplex virus type-2 using a genetically disabled herpes simplex type 1 virus. J. Infect. Dis., 170: 1100 –1109, 1994. Dillo, D., Rill, D., Entwisle, C., Bournsnell, M. E., Zhong, W., Holden, W., Holladay, M., Inglis, S., and Brenner, M. A novel herpes vector for the high-efficiency transduction of normal and malignant hemapoietic cells. Blood, 89: 119 –127, 1997. Todryk, S., McLean, C., Ali, S., Entwistle, C., Boursnell, M., Rees, R., and Vile, R. Disabled infectious single cycle herpes simplex virus vector for immunotherapy of colorectal cancer. Hum. Gene Ther., 10: 2757–2768, 1999. Boursnell, M. E. G., Rutherford, E., Hickling, J. K., Rollinson, E. A., Munro, A. J., Rolley, N., McLean, C. S., Borysiewicz, L. K., Vousden, K., and Inglis, S. C. Construction and characterisation of a recombinant vaccinia was expressing human papillomavirus proteins for immunotherapy of cervical cancer. Vaccine, 16: 1485– 1494, 1996. Gough, N. M., Metcalf, D., Gough, J., Grail, D., and Dunn, A. R. Structure and expression of the messenger RNA for murine granulocyte-macrophage colony stimulating factor. EMBO J., 4: 645– 653, 1985. Benjamin, R. J., Cobbold, S. P., Clark, M. R., and Waldmann, H. tolerance to rat monoclonal antibodies–implications for serotherapy. J. Exp. Med., 163: 1539 –1552, 1986. Melcher, A., Todryk, S., Hardwick, N., Ford, M., Jacobson, M., and Vile, R. G. Tumour immunogenicity is determined by the mechanism of cell death via induction of heat shock protein expression. Nature Med., 4: 581–587, 1998. Zhang, J., and Russell, S. J. Vectors for cancer gene therapy. Cancer Metastasis Rev., 15: 385– 401, 1996. Boursnell, M. E. G., Entwisle, C., Ali, S. A., Shields, J. G., Inglis, S. C., and Rees, R. C. Disabled infectious single cell (DISC) herpes simplex virus as a vector for immunotherapy of cancer. Adv. Exp. Med. Biol., 451: 379 –384, 1998. Lowstein, P. R., Fournel, S., Bain, D., Tomasec, P., Clissold, P., and Castro, M. G. Herpes-simplex virus-1 (hsv-1) helper coinfection affects the distribution of an amplicon encoded protein in glia. Neur. Rep., 5: 1625–1630, 1994. Matzinger, P. Tolerance, danger and the extended family. Annu. Rev. Immunol., 12: 991–1043, 1994. Kovacsovics-Bankowski, M., and Rock, K. L. A phagosome-to-cystol pathway for exogenous antigens presented on MHC class I molecules. Science (Washington DC), 267: 243, 1995. Haung, A. Y., Golumbek, P., and Ahmadzadeh, M. Role of bone marrow-derived cells in presenting MHC class I-restricted tumour antigens. Science (Washington DC), 264: 961–965, 1994. Dranoff, G., Jaffee, E., Lazenby, A., Golumbek, P., Levitsky, H., Brose, K., Jackson, V., Hamada, H., Pardoll, D., and Mulligan, R. Vaccination with irradiated tumor cells engineered to secrete murine granulocyte-macrophage colony-stimulating factor stimulates potent, specific, and long-lasting anti-tumor immunity. Proc. Natl. Acad. Sci. USA, 90: 3539 –3545, 1993. Rees, R. C. Cytokines as biological response modifiers. J. Clin. Pathol., 45: 93–98, 1992. Pawelec, G., Rees, R. C., Kiessling, R., Madrigal, A., Dodi, A., Baxevanis, C., Gambacorti-Passerini, C., Masucci, G., and Zeuthen, J. Cells, and Cytokines in Immunotherapy and Gene Therapy of Cancer. Crit. Rev. Oncog., 10: 83–127, 1999. Simova, J., Bubenik, J., Jandlova, T., and Indrova, M. Irradiated IL-2 gene-modified plasmacytoma vaccines are more efficient than live vaccines. Int. J. Oncol., 12: 1195–1198, 1998. Kanangat, S., Thomas, J., Gangappa, S., Babu, J. S., and Rouse, B. T. Herpes simplex virus type 1-mediated up-regulation of IL-12 (p40) mRNA expression: implications in immunopathogenesis acid protection. J. Immunol., 156: 1110, 1996. Mu, J., Zou, J. P., Yamamoto, N., Tsutsui, T., Tai, X. G., Kobayashi, M., Herrmann, S., Fujiwara, H., and Hamaoka, T. Administration of recombinant interleukin 12 prevents outgrowth of tumour cells metastasizing spontaneously to lung and lymph nodes. Cancer Res., 55: 4404, 1995. Sayers, T. J., Wiltrout, T. A., McCormick, K., Husted, C., and Wiltrout, R. H. Antitumor effects of -interferon and -interferon on a murine renal-cancer (Renca) in vitro and in vivo. Cancer Res., 50: 5414 –5420, 1990. Younes, E., Haas, G. P., Dezso, B., Ali, E., Maughan, R. L., Montecillo, E., Pontes, J. E., and Hillman, G. G. Radiation-induced effects on murine kidney tumour cells–role in the interaction of local irradiation and immunotherapy. J. Urol., 153: 2029 –2033, 1995. Pardoll, M. D. Paracrine cytokine adjuvants in cancer immunotherapy. Annu. Rev. Immunol., 13: 399 – 415, 1995. Simons, J. W., Jaffee, E. M., Weber, C. E., Levitsky, H. I., Nelson, W. G., Carducci, M. A., Laze, A. J., Cohen, L. K., Finn, C. C., Clift, S. M., Hauda, K. M., Beck, L. A., Leiferman, K. M., Owens, A. H., Piantadosi, S., Dranoff, G., Mulligan, R. C., Pardoll, D. M., and Marshall, F. F. Bioactivity of autologous irradiated renal cell carcinoma vaccines generated by ex vivo granulocyte-macrophage colony-stimulating factor gene transfer. Cancer Res., 57: 1537–1546, 1997. Golumbek, P. T., Lazenby, A. J., Levitsky, H. I., Jaffee, L. M., Karasuyama, H., Baker, M., and Pardoll, D. M. Treatment of established renal cancer by tumour cells engineered to secrete IL-4. Science (Washington DC), 254: 713–716, 1991. Tepper, R. I., Pattengale, P. K., and Leder, P. Murine interleukin-4 displays potent anti-tumor activity in vivo. Cell, 57: 503–512, 1989.

1670