Publication of the International Union Against Cancer Publication de l’Union Internationale Contre le Cancer Int. J. Cancer: 70, 706–715 (1997) r 1997 Wiley-Liss, Inc. REGRESSION OF TUMORS IN MICE VACCINATED WITH PROFESSIONAL ANTIGEN-PRESENTING CELLS PULSED WITH TUMOR EXTRACTS Smita K. NAIR1, David SNYDER1, Barry T. ROUSE2 and Eli GILBOA1* 1Department of Surgery, Duke University Medical Center, Durham, NC 2Department of Microbiology, University of Tennessee, Knoxville, TN Vaccination with tumor extracts circumvents the need to identify specific tumor rejection antigens and extends the use of active immunotherapy to the vast majority of cancers, in which specific tumor antigens have not yet been identified. In this study we examined the efficacy of tumor vaccines comprised of unfractionated tumor material presented by professional antigen-presenting cells (APC): dendritic cells (DC) or macrophages (Mø). To enhance the relevance of these studies for human patients we used 2 poorly immunogenic murine tumor models and evaluated the effectiveness of the vaccination protocols in tumor-bearing animals. APC (in particular DC) pulsed with unfractionated extracts from these ‘‘poorly immunogenic’’ tumors were highly effective in eliciting tumorspecific cytotoxic T lymphocytes. A measurable CTL response could be detected after even a single immunization with tumor extract-pulsed DC. DC or Mø pulsed with tumor extract were also effective vaccines in tumor-bearing animals. In the murine bladder tumor (MBT-2) model a modest extension of survival and 40% cure rate was seen in the animal groups immunized with DC or Mø pulsed with MBT-2 tumor extract. DC or Mø pulsed with B16/F10.9 tumor extract were also remarkably effective in the B16 melanoma lung metastasis model, as shown by the observation that treatment with APC caused a significant reduction in lung metastases. Cumulatively, the CTL and immunotherapy data from the two murine tumor systems suggest that APC (in particular DC) pulsed with unfractionated cell extracts as a source of tumor antigen may be equally or more effective than genetically modified tumor vaccines. Int. J. Cancer 70:706–715, 1997. r 1997 Wiley-Liss, Inc. Animal studies support the notion that, with a few exceptions, tumor-specific CD81 CTL constitute an important effector arm of the antitumor immune response (Greenberg, 1991). Hence, antigens recognized by CD81 CTL cells are likely to function as tumor rejection antigens capable of eliciting protective immunity in vivo. The existence of specialized, or professional, antigen-presenting cells (APC) that are responsible for the presentation of Ag to naive CD81 T cells was based on observations that host MHC-restricted CTL can be primed in vivo to Ag that was introduced on MHC-disparate cells (Matzinger and Bevan, 1977). Additional evidence stems from transplantation studies suggesting that only a subset of allo-MHC donor cells, called passenger leukocytes, were responsible for the induction of an immune response and rejection of the mismatched tissue (Lafferty et al., 1983). Studies exploring the mechanism of action of interleukin-2 (IL-2) or granulocyte/ macrophage colony-stimulating factor (GM-CSF)-secreting tumor vaccines have also suggested that priming of an MHC class I-restricted antitumor response required the transfer of antigens from the tumor cell to a host-derived cell for presentation to CD81 CTL (Huang et al., 1994; Bannerji et al., 1994). As recently discussed by Bevan (1995), a number of possible pathways exist by which extracellular antigens can translocate into the cytosolic class I presentation pathway of host-derived professional APC. The main candidates for professional APC are macrophages (Mø) and the bone marrow-derived dendritic cells (DC). Several studies have documented the exceptional ability of DC to stimulate naive T cells, both in vitro and in vivo. DC pulsed with protein or peptide in the presence of lipid (Nair et al., 1993) or transfected with DNA (Rouse et al., 1994) are capable of eliciting primary CTL responses in vitro, and inoculation of mice with small numbers of allogeneic DC (McKinney and Streilein, 1989) or with peptide- pulsed DC (Takahashi et al., 1993) induces a potent CTL response in vivo. In general, Mø are less effective than DC at inducing T-cell responses in vitro or in vivo (Steinman, 1991). However, presentation of exogenous soluble antigens to CD81 T cells, a defining feature of a professional APC, is carried out by cells with Mø characteristics (Rock et al., 1993). A number of genes that encode tumor antigens recognized by CD81 T cells have been characterized (Boon et al., 1994). There are at least 3 advantages to using defined tumor antigens in cancer immunotherapy: 1) use of defined tumor antigen obviates the need for tumor tissue and therefore will benefit patients with low tumor burden; 2) the purity of the antigenic preparation is likely to enhance the effectiveness of the vaccines; and 3) the absence of irrelevant tumor material will minimize possible autoimmune reactions against ‘‘self antigen.’’ There are, however, 3 possible drawbacks in using defined CTL antigens: 1) it is unclear whether or which of the identified human tumor-specific antigens are the best choice to mount an effective anti-tumor immune response in vivo; This potential concern was underscored in a report by Anichini et al. (1996), who showed that the majority of CTL present in HLA-A2.1 melanoma patients were not directed to the tumor antigens, Melan-A/Mart-1, tyrosinase, gp100, or MAGE-3; 2) the use of vaccines consisting of a single antigen or a few tumor antigens carry the risk of generating escape mutants; and 3) human CTL antigens have been identified and isolated only from a small number of cancers. An alternative strategy, not encumbered by these limitations, is to use unfractionated tumor-derived antigens obtained from tumors, such as whole tumor cells, total peptide extracts or total protein extracts. The main advantages of using unfractionated tumor material as a source of tumor antigen are: 1) the identity of the effective tumor antigen(s) need not be known, a fact that expands significantly the type of cancer that can be treated; and 2) the (likely) presence of multiple tumor antigens reduces the risk of escape mutants. There are, however, three potential drawbacks in vaccinating cancer patients with unfractionated tumor-derived peptides or proteins, compared with use of purified antigens: 1) use of unfractionated tumor material as a source of tumor antigen will depend on the availability of substantial amounts of tumor tissue from the patient; 2) vaccination with unfractionated tumor-derived antigens could induce autoimmune responses directed against ‘‘self’’ antigens; and 3) immunization with unfractionated tumor material may be less effective due to the low concentration of effective tumor antigens in the mixture. Several studies have shown that immunization of mice with professional APC pulsed with unfractionated tumor proteins induced protective immunity in mice against a subsequent challenge with live tumor cells (Grabbe et al., 1991; Shimizu et al., 1989; Flamand et al., 1994; Cohen et al., 1994). Approximating the conditions prevailing in cancer patient more closely, genetically engineered whole cell tumor vaccines were capable of causing the regression of tumors devoid of intrinsic immunogenicity (Gilboa and Lyerly, 1994). *Correspondence to: Department of Surgery, Duke University Medical Center, Box 2601, Durham, NC 27710, USA. Fax: 919 681 7970. Received 26 August 1996; revised 21 November 1996 TUMOR EXTRACT-PULSED APC AND TUMOR IMMUNITY In this study we examined the efficacy of tumor vaccines comprised of unfractionated tumor material presented by professional APC, DC or Mø. Using 2 poorly immunogenic murine tumor models, we show that immunization with APC pulsed with unfractionated tumor cell extracts induce potent CTL responses in vivo and cause the regression of pre-existing tumors in tumorbearing animals. MATERIAL AND METHODS Mice Retired breeder female C57BL/6 mice (H-2b ), 7–8 weeks old and C3H/He mice (H-2k ) were obtained from the Jackson Laboratory (Bar Harbor, ME). In conducting the research described below, the investigators adhered to the ‘‘Guide for the Care and Use of Laboratory Animals’’ as proposed by the committee on care of Laboratory Animal Resources Commission on Life Sciences, National Research Council. The facilities are fully accredited by the American Association for Accreditation of Laboratory Animal Care. Cell lines The murine MBT-2 cell line, derived from a carcinogen-induced bladder tumor in C3H mice, was obtained from Dr. T. Ratliff (Washington University, St. Louis, MO). The MBT-2/IL-2 cell line is derived by transfecting MBT-2 cell with human IL-2 cDNA (Connor et al., 1993). The F10.9 clone of the B16 melanoma of C57BL/6 origin is a highly metastatic, poorly immunogenic and a low class I-expressing cell line. F10.9/K1 is a poorly metastatic and highly immunogenic cell line derived by transfecting F10.9 cells 707 H-2Kb cDNA (Porgador et al., 1995). Other with a class I molecule, cell lines used were EL4 (C57BL/6, H-2b, thymoma), E.G7-OVA (EL4 cells transfected with the cDNA of chicken ovalbumin (OVA) (Moore et al., 1988), A20 (H-2d, B cell lymphoma) and L929 (H-2k fibroblasts). Cells were maintained in DMEM supplemented with 10% FCS, 25 mM Hepes, 2 mM L-glutamine and 1 mM sodium pyruvate. E.G7-OVA cells and MBT-2/IL-2 cells were maintained in medium supplemented with 400 µg/ml G418 (GIBCO, Grand Island, NY), and F10.9/K1 cells were maintained in medium containing 800 µg/ml G418. APC and responder T cells Splenocytes obtained from naive C57BL/6 female retired breeders were treated with ammonium chloride Tris buffer for 3 min at 37°C to deplete red blood cells. Splenocytes (3 ml) at 2 3 107 cells/ml were layered over 2 ml metrizamide gradient column (Nycomed Pharma, Oslo, Norway; analytical grade, 14.5 g added to 100 ml PBS, pH 7.0) and centrifuged at 600g for 10 min. The DC-enriched low-density fraction from the interface was further enriched by adherence for 90 min to remove contaminating T and B cells. Adherent cells (mostly DC and a few contaminating Mø) were retrieved by gentle scraping and subjected to a second round of adherence at 37°C for 90 min to deplete the contaminating Mø. After the second round of adherence, non-adherent cells were pooled as splenic DC and FACS analysis showed approximately 80–85% DC (MAb 33D1), 1–2% Mø (mAb F4/80), 5–10% T cells and ,5% B cells (data not shown, Nair et al., 1993). The pellet was resuspended and enriched for Mø by 2 rounds of adherence at 37°C for 90 min each. More than 80% of the adherent FIGURE 1 – Induction of OVA-specific CTL responses in mice immunized with DC pulsed with OVA protein. Mature splenic DC were pulsed with OVA protein in the presence or absence of the cationic lipid, DOTAP, as described in Material and Methods. C57BL/6 mice were immunized once i.p., and after 7 days splenocytes were harvested and restimulated in vitro. CTL assay was done on day 5 with E.G7-OVA, EL4 and BALB/3T3 cells as targets. Control targets EL4 and BALB/3T3 showed insignificant lysis (data not shown). 708 NAIR ET AL. population was identified as Mø by FACS analysis with 5% lymphocytes and ,5% DC. Pulsing of APC APC were washed twice in Opti-MEM medium (GIBCO). Cells were resuspended in Opti-MEM medium at 5–10 3 106 cells/ml and added to 50 ml polypropylene tubes (Falcon, Oxnard, CA). The cationic lipid DOTAP (Boehringer Mannheim, Indianapolis, IN) was used to deliver protein or tumor extracts into cells. Tumor extracts were obtained by sonicating tumor cells in Opti-MEM (107 cells/500 µl) using a Special Ultrasonic Cleaner (Laboratory Supplies, Hicksville, NY). Cell sonicates were used without any further manipulation as tumor extracts. Tumor cell sonicates or tumor extracts (500 µl) and DOTAP (125 µg in 500 µl Opti-MEM medium) were mixed in 12 3 75 mm polystyrene tubes at room temperature (RT) for 20 min. The complex was added to the APC (107 cells) in a total volume of 2–5 ml and incubated at 37°C in a water bath with occasional agitation for 2 hr. The cells were washed, irradiated at 3,000 rads and resuspended in PBS (2 3 106 DC pulsed with 2 3 106 tumor cell extract and 25 µg DOTAP in 500 µl PBS/mouse) for intraperitoneal immunizations. Purified OVA protein was used at a concentration of 50 µg/mouse. Induction of antigen-specific CTL in vivo Cells were pulsed with tumor extracts or OVA protein as described above. Naive, syngeneic mice were immunized intraperitoneally with 2 3 106 APC (irradiated at 3,000 rads)/mouse in 500 µl PBS. Irradiated tumor cells were used at a concentration of 5 3 106 cells per mouse. Splenocytes were harvested after 7–10 days and depleted of red blood cells with ammonium chloride Tris buffer. Splenocytes (1.5 3 107 ) were cultured with 7.5–10 3 105 irradiated stimulator cells (E.G7-OVA cells irradiated at 20,000 rads, MBT-2/IL-2 cells irradiated at 7,500 rads and F10.9/K1 cells irradiated at 7,500 rads) in 5 ml of IMDM with 10% FCS, 1 mM sodium pyruvate, 100 IU/ml penicillin, 100 mg/ml streptomycin and 5 3 1025 M b-mercaptoethanol/well in a 6-well tissue culture plate. Cells were cultured for 5 days at 37°C and 5% CO2. Effectors were harvested on day 5 on Histopaque 1083 gradient prior to use in a CTL assay. Alternatively, human recombinant IL-2 (5 units/ml, Schiapparelli Biosystems, Columbia, MD) was used to restimulate effector cells instead of irradiated tumor cells. Cytotoxicity assay Target cells (5–10 3 106 ) were labeled with europium diethylenetriamine penta-acetate for 20 min at 4°C. After several washes, 104 europium-labeled targets and serial dilutions of effector cells at an effector/targer ratio of 80:1 to 5:1 were incubated in 200 µl of RPMI 1640 with 10% heat-inactivated FCS in 96-well V-bottomed plates. The plates were centrifuged at 500g for 3 min and incubated at 37°C and 5% CO2 for 4 hr; 50 µl of the supernatant were harvested, and europium release was measured by timeresolved fluorescence (Delta fluorometer, Wallac, Gaithersburg, MD) (Saito et al., 1994). Spontaneous release was less than 25%. Standard errors (SE) of the means of triplicate cultures were less than 5%. FIGURE 2 – Priming of OVA-specific CTL in mice immunized with DC pulsed with cellular extracts from OVA expressing tumor cells. Mature splenic DC were pulsed with OVA protein, EG7, EL4 or A20 extracts in the presence of the cationic lipid, DOTAP, EG7 extract/DOTAP alone or DC 1 EG7 extract without DOTAP as described in Material and Methods. C57BL/6 mice were immunized once i.p., and after 10 days splenocytes were harvested and restimulated in vitro. CTL assay was done on day 5 with E.G7-OVA, EL4, RMA-S cell pulsed with OVA peptide and A20 cells as targets. Control targets A20 showed insignificant lysis (data not shown). TUMOR EXTRACT-PULSED APC AND TUMOR IMMUNITY Immunotherapy MBT-2 murine model.Orthotopic implantation of MBT-2 cells into the bladder of C3H mice was performed as described previously (Connor et al., 1993). Briefly, under magnification, a 0.8 cm incision was made transversely in the abdomen just above the pubis. The anterior abdominal wall muscles were incised, and the bladder was delivered into the surgical field. Using a 1 ml tuberculin syringe, MBT-2 cells (2 3 104 ) in 50 µl PBS were injected into the bladder wall. The incision was closed in one layer using a 5.0 prolene suture. The procedure was well tolerated, and postoperative mortality was less than 5%. Mice were vaccinated on days 4, 8 and 12 following implantation of MBT-2 cells. Mice were evaluated on a daily basis and were sacrificed when moribund. Each treatment group consisted of 5–10 mice. F10.9-B16 melanoma model.Mice were injected intrafootpad with 2 3 105 F10.9 cells. The postsurgical protocol was used as described previously with a few modifications (Porgador et al., 1995). Mice were amputated when the local tumor in the footpad was 5.5–7.5 mm in diameter. Postamputation mortality was less than 5%. Two days postamputation mice were immunized intraperitoneally followed by weekly vaccinations twice, for a total of 3 vaccinations. Mice were sacrificed based on the metastatic death in the non-immunized or control groups (28–32 days postamputation). Metastatic loads were assayed by weighing the lungs and by counting the number of metastatic nodules. Statistical methods For the MBT-2 model, overall significance of the study was calculated using the Kaplan-Meier method, and the significance of 709 the differences between the survival rates was calculated using the log-rank test. In the B16 melanoma model, the different experimental groups within the study were compared using the KruskalWallis test. Comparisons of significance for differences in lung weights between specific pairs of groups were then compared by the Mann-Whitney U-test. A probability of less than 0.05 ( p , 0.05) was used for statistical significance. RESULTS Dendritic cells pulsed with OVA protein in the presence of the cationic lipid DOTAP induce OVA-specific CTL responses in vivo To test whether DC pulsed with protein are capable of inducing CTL in vivo, mice were immunized once with DC pulsed with OVA protein in the presence or absence of DOTAP. The OVA protein encodes an H-2Kb epitope (aa 257–264, SIINFEKL) that is recognized by CD81 CTL in C57BL/6 mice. As controls, DC were incubated with DOTAP with no antigen, or with antigen in the absence of DOTAP. Splenocytes were restimulated in vitro with E.G7-OVA cells, which are EL4 (H-2b ) thymoma cells transfected with and expressing the chicken OVA protein (Moore et al., 1988). Cytotoxic activity of the splenocytes was tested on E.G7-OVA, EL4 and BALB/3T3 cells as targets. As shown in Figure 1, only DC pulsed with OVA protein in the presence of DOTAP were able to induce a strong and specific CTL response in the treated mice. Presumably, the function of DOTAP is to facilitate cytoplasmic incorporation of the exogenous antigen for MHC class I presentation to CD81 T cells. Control targets, EL4 and BALB/3T3 cells (data not shown) were not lysed by the CTL. Mice immunized FIGURE 3 – Induction of MBT-2 tumor-specific CTL by immunization with tumor extract pulsed APC. Splenic DC and Mø were pulsed with MBT-2 tumor cell sonicates in the presence or absence of DOTAP as described in Material and Methods. C3H mice were immunized i.p. 3 times at weekly intervals with 2 3 106 APC or with 5 3 106 tumor cells. Splenocytes were harvested and restimulated with recombinant IL-2, and a CTL assay was done after 5 days. EL4 cells were used as targets for MHC restriction and showed insignificant lysis. This experiment was repeated 3 times with similar results. 710 NAIR ET AL. intraperitoneally with OVA 1 DOTAP without DC were much less effective at in vivo CTL induction as shown in Figure 1. Comparison between intraperitoneal and intravenous routes of immunization did not show any significant differences in the levels of cytotoxicity generated (data not shown). Figure 1 demonstrates that it is possible to induce an OVAspecific CTL in vivo following a single immunization with DC pulsed with OVA protein and DOTAP. The aim of our study was to determine if we could generate CTL responses using crude tumor extracts instead of using purified protein antigens as a practical alternative for active immunotherapy of cancer patients when tumor antigens have not been identified. To determine if we could generate OVA-specific and tumor-specific CTL using unfractionated tumor-derived antigen, we pulsed DC with tumor cell sonicates from EL4 thymoma cells and E.G7-OVA in the presence of DOTAP. Splenocytes were harvested from immunized mice to determine CTL activity (refer to Material and Methods). E.G7OVA, EL4, A20 and RMA-S pulsed with OVA peptide were used as targets. As shown in Figure 2, DC pulsed with OVA protein/DOTAP were more effective at generating OVA-specific CTL responses compared with DC 1 E.G7-OVA extract/DOTAP (70% lysis of RMA-S cells pulsed with OVA peptide compared with 30% lysis), whereas both groups were comparable in their ability to lyse E.G7-OVA targets. Similarly, DC pulsed with EL4 or E.G7-OVA cell sonicates plus DOTAP were equally effective at generating EL4-specific CTL responses in vivo. This is also evident in Figure 2 as shown by the lysis of E.G7-OVA targets by effectors generated from mice immunized with DC 1 EL4 extract/DOTAP. Thus a single immunization with DC pulsed with unfractionated E.G7OVA cell sonicates primed OVA-specific CTL that were capable of lysing E.G7-OVA (55% specific lysis), RMA-S pulsed with OVA peptide (30% specific lysis) and EL4 cells (17% specific lysis). Mice immunized with DC pulsed with E.G7-OVA extract in the absence of DOTAP or with E.G7-OVA extract/DOTAP complexes in the absence of DC generated insignificant E.G7-OVA-specific CTL responses in vivo. Control targets A20 showed insignificant lysis (data not shown). This experiment demonstrates that immunization with DC pulsed with extracts from OVA-expressing EL4 tumors is capable of eliciting CTL responses against OVA antigen as well as undefined EL4 tumor antigens. In vivo priming of tumor-specific CTL responses using DC or Mø pulsed with unfractionated tumor material To test whether unfractionated proteins from tumors can elicit CTL responses in vivo, we used tumor cell extracts from MBT-2 (H-2k ) murine bladder tumor cells. The MBT-2 tumor cell line is a poorly immunogenic tumor, as indicated by the fact that repeated immunizations with irradiated MBT-2 cells fail to elicit a measurable CTL response (Connor et al., 1993; Saito et al., 1994; Fig. 3). However, as we have previously shown, C3H (H-2k ) mice immunized 3 consecutive times at weekly intervals with genetically modified MBT-2 cells expressing the human IL-2 gene (MBT-2/ IL-2) elicit a strong tumor-specific CTL response (Connor et al., 1993; Saito et al., 1994; Fig. 3). As shown in Figure 3, under the same experimental conditions, DC or Mø pulsed with MBT-2 FIGURE 4 – A single immunization with DC pulsed with MBT-2 or with L929 tumor cell extracts elicits antigen-specific CTL responses in vivo. C3H mice were immunized with DC or Mø pulsed with MBT-2 or with L929 tumor extracts as described in Material and Methods. Ten days later splenocytes were harvested and cultured in the presence of IL-2. CTL assay was done on day 5 with MBT-2, L929 and EL4 cells as targets. EL4 targets showed insignificant lysis (data not shown). TUMOR EXTRACT-PULSED APC AND TUMOR IMMUNITY extracts in the presence of the lipid DOTAP were also capable of eliciting a strong CTL response, although DC were reproducibly more effective than Mø. DC or Mø pulsed with tumor extracts in the absence of DOTAP also elicited CTL, albeit, a low level. EL4 tumor cells used as control targets (H-2b ) showed no specific lysis. Also shown is the fact that immunization with MBT-2 tumor extract 1 DOTAP alone was consistently ineffective at inducing CTL responses in vivo. As shown in Figures 1 and 2, a single immunization with DC pulsed with a defined CTL antigen, the chicken OVA protein, was sufficient to induce a strong CTL response in vivo, while 3 immunizations with MBT-2 tumor extract-pulsed DC or Mø were used to elicit the tumor-specific cytotoxic responses shown in Figure 3. We therefore tested whether a single immunization with unfractionated MBT-2 tumor extracts loaded onto APC would suffice to generate CTL in vivo. As shown in Figure 4, a single immunization with tumor extract-pulsed DC elicited a CTL response in vivo, whereas Mø pulsed with tumor extract were not effective stimulators of CTL induction. Irradiated MBT-2/IL-2 cells were also incapable of CTL priming in vivo following a single immunization. Similarly, a single immunization with DC, but not Mø, pulsed with L929 (H-2k ) tumor extracts elicited an L929specific CTL response. These observations show that in this experimental system DC pulsed with unfractionated tumor extracts were superior to either Mø pulsed with tumor extracts or to IL-2-secreting MBT-2 tumor cells in eliciting CTL in vivo. The antigen specificity of the CTL response is illustrated by the fact that DC pulsed with MBT-2 extracts generated CTL responses capable 711 of lysing only MBT-2 targets, and DC pulsed with L929 extracts generated only CTL capable of lysing L929 cells. EL4 tumor cells used as control targets (H-2b ) showed no specific lysis (data not shown). Treatment of MBT-2 tumor-bearing animals with APC pulsed with tumor extracts The mouse MBT-2 tumor is an excellent model for human bladder cancer. Using experimental conditions that simulate as closely as possible the conditions prevailing in the cancer patients, we have previously shown that IL-2- or GM-CSF-secreting irradiated MBT-2 cell preparations were capable of curing animals of orthotopically established tumors and also engendered protective immunological memory in the cured animals (Connor et al., 1993; Saito et al., 1994). In the experiment shown in Figure 5, tumors were established orthotopically by implanting MBT-2 cells in the bladder wall. Four days post-tumor implantation, mice were vaccinated 3 times with APC pulsed with tumor extracts or with irradiated tumor cells (MBT-2 and MBT-2/IL-2). All mice in the group treated with irradiated MBT-2 cells or with DC pulsed with L929 tumor extract died within 4 weeks. Mice treated with irradiated IL-2-secreting MBT-2 cells had a modest survival advantage, similar though less pronounced, than that observed in previous studies (Connor et al., 1993; Saito et al., 1994). The most pronounced therapeutic benefit was seen in mice treated with DC or with Mø pulsed with MBT-2 extract. Two of 5 mice (40%) in each treatment group survived for over 60 days, after which mice were challenged again with MBT-2 cells. Mice FIGURE 5 – Treatment of tumor-bearing mice with APC pulsed with MBT-2 tumor extracts. Tumors were established in C3H mice by orthotopic implantation of 2 3 104 MBT-2 cells into the bladder of the animal. Vaccinations consisting of 5 3 106 tumor cells or 2 3 106 APC pulsed with tumor extracts were given as described in Material and Methods. Mice were vaccinated a total of 3 times and were evaluated on a daily basis and sacrificed when moribund. Each treatment group consisted of 5 mice. Data are representative of 3 experiments performed with similar results. Differences in survival for the groups MBT-2/IL-2, Mø 1 MBT-2 extract and DC 1 MBT-2 extract are significant compared with the control MBT-2 group, based on log-rank analysis ( p 5 0.0017 for MBT-2/IL-2 and p 5 0.0013 for Mø 1 MBT-2 extract and DC 1 MBT-2 extract). p 5 0.395 for the control DC 1 L929 extract group compared with the MBT-2 group. The overall significance of this study is p 5 0.0001 based on the Kaplan-Meier test. 712 NAIR ET AL. immunized with DC or Mø pulsed with tumor extracts in the presence of DOTAP survived the rechallenge and were sacrificed on day 90. In a separate experiment, all survivors were tested for CTL activity and demonstrated MHC-restricted, MBT-2-specific cytotoxic responses (data not shown). This experiment therefore suggests that vaccines based on DC or Mø pulsed with unfractionated tumor extracts are equally or more potent than IL-2-secreting MBT-2 cells as vaccines. Induction of tumor immunity in the B16/F10.9 melanoma metastasis model The potency of APC pulsed with unfractionated tumor extracts was further evaluated in the B16/F10.9 (H-2b ) melanoma metastasis model. The B16/F10.9 melanoma tumor is poorly immunogenic, expresses low levels of MHC class I molecules and is highly metastatic in both experimental and spontaneous metastasis assay systems (Porgador et al., 1995). Porgador et al. (1995) have shown that when vaccinations are carried out after the removal of the primary tumor implant, only irradiated tumor cells transduced with both the IL-2 and the H-2Kb genes were capable of significantly impacting the metastatic spread of B16/F10.9 tumor cells in the lung. Thus, the B16/F10.9 melanoma model and the experimental design used by Porgador et al. (1995) constitute a stringent and highly informative experimental system to assess the efficacy of adjuvant treatments for metastatic cancer. We first tested whether a single immunization of C57BL/6 (H-2b ) mice with DC or Mø pulsed with extracts from B16/F10.9 cells elicits a CTL response against B16/F10.9 cells, or against B16/F10.9 cells expressing H-2Kb (F10.9/K1 cells). As shown in Figure 6, only DC pulsed with B16/F10.9 tumor extracts were able to elicit a strong tumor-specific CTL response against either target, while tumor extract-pulsed Mø elicited a barely detectable CTL response. No CTL could be detected in mice immunized with irradiated B16/F10.9 cells, or with DC or Mø pulsed with EL4 tumor extracts. In another experiment EL4 cells and BALB/3T3 cells were also used as control targets. No lysis of BALB/3T3 cells and EL4 was observed when effectors were generated using DC pulsed with F10.9 cells (data not shown). As shown in Figure 6, F10.9 tumor extracts 1 DOTAP in the absence of DC or Mø were not capable of priming F10.9-specific CTL responses in vivo, even after 3 consecutive immunizations. To test whether immunization with DC or Mø was capable of causing the regression of pre-existing lung metastases, primary tumors were induced by implantation of B16/F10.9 tumor cells in the footpad. When the footpad reached 5.5–7.5 mm in diameter, the tumors were surgically removed, and 2 days later mice were immunized with irradiated B16/F10.9 cells, or with APC pulsed with tumor extract (Fig. 7). The mice received a total of 3 vaccinations given at weekly intervals. The average lung weight of a normal mouse is 0.18–0.22 g. Mice immunized with irradiated B16/F10.9 cells were overwhelmed with metastases. The mean lung weight of mice in this treatment group was 0.84 g, about three-quarters of the weight was contributed by the metastases, which were too many to count. A similar metastatic load was seen in animals treated with PBS (data not shown), which confirms numerous previous observations that treatment with irradiated B16/F10.9 tumor cells alone has no therapeutic benefit in this tumor model (Porgador et al., 1995). As also previously shown, immunization with H-2Kb expressing B16/F10.9 cells (F10.9/K1) had a modest therapeutic benefit, as indicated by a statistically FIGURE 6 – Induction of tumor-specific CTL following a single immunization with DC pulsed with B16/F10.9 tumor extracts. DC or Mø were pulsed with extracts derived from B16/F10.9 tumor cells. C57BL/6 mice were immunized once i.p.; splenocytes were then harvested 10 days later and restimulated with irradiated K1 cells (B16/F10.9 cells expressing H-2Kb ). CTL assay was done 5 days later with F10.9 cells and F10.9/K1 cells as targets. DC and Mø pulsed with EL4 cell extract were used as controls for antigen specificity. TUMOR EXTRACT-PULSED APC AND TUMOR IMMUNITY 713 FIGURE 7 – Regression of lung metastases in mice treated with APC pulsed with F10.9 extracts. B16/F10.9 tumors established in the footpads of C57BL/6 mice were amputated when they reached 5.5–7.5 mm in diameter. Two days after the amputation, and weekly thereafter, mice were vaccinated i.p. for a total of 3 vaccinations. Mice were sacrificed 28–32 days post-amputation (as determined by the metastatic death monitored for the control F10.9 group), and lung weights and metastatic loads were determined. Columns represent mean lung weight, and dots represent individual lung weights (6–7 mice/group). The results are representative of 3 different experiments. For more details see Material and Methods. Relative to the control F10.9 immunized group, p values were 0.0379, 0.62, 0.0012 and 0.0023 for F10.9/K1, DC 1 EL4 extract, DC 1 F10.9 extract and Mø 1 F10.9 extract immunized mice, respectively. The overall significance of the study as determined by the Kruskal-Wallis test is p 5 0.0002. NAIR ET AL. 714 significant decrease in the average lung weight of the animals in this treatment group. One of 7 animals in this group had no visible metastasis, 4 of 7 animals had 20–35 nodules and 1 mouse did not respond to treatment (. 100 nodules). A more marked response was seen in the animal groups treated with APC pulsed with tumor extract. The mean lung weight of mice treated with DC pulsed with B16/F10.9 tumor extract was 0.30 g, only 33% above the normal lung weight. Four mice in this group (n 5 7) had no visible nodules, 2 mice had less than 5 nodules and 1 mouse had 15 nodules. Mø were almost as effective as DC, the average lung weight being 0.38 g, 80% above the normal lung average. In this treatment group (n 5 7) 3 mice had no visible metastasis, 3 mice had between 5 and 10 nodules and 1 mouse had 22 nodules. The antigen specificity of the therapeutic benefit seen in this experiment is indicated by the fact that no decrease in metastatic load was seen in mice treated with DC pulsed with EL4 tumor extract. DISCUSSION In this study we have shown that immunization with unfractionated tumor extracts presented by professional APC, in particular DC, elicits potent anti-tumor immunity in mice. To enhance the relevance of these results to human patients, we used two poorly immunogenic murine tumor models and evaluated the effectiveness of the vaccination protocols in tumor-bearing animals. Neither B16/F10.9 nor MBT-2 tumor cells alone were capable of eliciting CTL, a fact that reflects the poor intrinsic immunogenicity of the tumors used here. On the other hand, APC pulsed with unfractionated extracts from these ‘‘non-immunogenic’’ tumors were highly effective in eliciting tumor-specific CTL. A measurable CTL response could be detected after even a single immunization with B16/F10.9- or MBT-2 extract-pulsed DC (Figs. 4, 6). The observation that 3 consecutive immunizations with IL-2-secreting MBT-2 cells were required to induce a measurable CTL response suggests that MBT-2 tumor extract-pulsed DC are more potent inducers of CTL than the genetically engineered, IL-2-secreting tumor cells. DC or Mø pulsed with tumor extract were also effective vaccines in tumor-bearing animals. A modest extension of survival and 40% cure rate were seen in the animal groups immunized with DC or Mø pulsed with MBT-2 tumor extract, while immunization with IL-2-secreting MBT-2 cells led to extended survival but no cures. This observation, together with the CTL data presented in Figures 3 and 4, shows that DC pulsed with MBT-2 tumor proteins are more potent inducers of tumor immunity than IL-2-secreting MBT-2 tumor cells (Connor et al., 1993; Saito et al., 1994). DC or Mø pulsed with tumor extract were also remarkably effective in a metastasis tumor model. The B16/F10.9 melanoma tumor system used in this study is an excellent model for minimal residual metastatic disease; it measures the efficacy of adjuvant therapy in reducing the growth of pre-existing lung metastases in animals in which the primary tumor is surgically removed. The only treatment that has shown significant therapeutic benefit in this disease model was vaccination with doubly transduced tumor cells, B16/F10.9 cells transduced with both the IL-2 and H-2Kb genes (Porgador et al., 1995). As shown above (Fig. 7), treatment of the tumor-bearing mice with DC or Mø pulsed with tumor extract caused a significant reduction in lung metastasis. The magnitude of the effect, especially in the group treated with tumor extract-pulsed DC, was similar to the effect reported with the doubly transduced vaccine (Porgador et al., 1995). Cumulatively, the CTL and immunotherapy data from the 2 murine tumor systems suggest that vaccines based on APC, in particular DC, pulsed with unfractionated cell extracts as a source of tumor antigen may be equally or more effective than genetically modified tumor vaccines. However, additional studies will be required to assess more definitively the comparative value of APC-based versus genetically modified tumor cell-based vaccines, or the use of DC vs. Mø as APC pulsed with tumor antigen. We are currently characterizing the active components in the tumor extracts and studying the cellular and immunological mechanisms responsible for the induction of tumor immunity mediated by APC. Vaccination with tumor extracts circumvents the need for identifying specific tumor rejection antigens and hence extends the use of active immunotherapy to the vast majority of cancers, in which specific tumor antigens have not been identified. Zitvogel et al. (1996) have shown that vaccination of mice with bone marrow-derived DC pulsed with unfractionated acid-eluted tumor peptides was capable of reducing the growth of subcutaneously established, weakly immunogenic tumors. One potential drawback, however, of immunizing with unfractionated tumor material, compared with the use of defined tumor antigens, is the increased risk of inducing autoimmune responses, with pathological consequences. While no evidence of autoimmunity was seen in the current or previous animal vaccination studies employing unfractionated tumor material as a source of tumor antigens, the development of increasingly potent vaccines may very well lead to some autoimmune manifestations. REFERENCES ANICHINI, A., MORTARINI, R., MACCALLI, C., SQUARCINA, P., FLEISCHHAUER, K., MASCHERONI, E. and PARMIANI, G., Cytotoxic T cells directed to tumor antigens not expressed on normal melanocytes dominate HLA-A2.1restricted immune repertoire to melanoma. J. Immunol., 156, 208–217 (1996). BANNERJI, R., ARROYO, C.D., CORDON-CARDO, C. and GILBOA, E., The role of IL-2 secreted from genetically modified tumor cells in the establishment of antitumor immunity. J. Immunol., 152, 2324–2332 (1994). BEVAN, M.J., Antigen presentation to cytotoxic T lymphocytes in vivo. J. exp. Med., 182, 639–641 (1995). BOON, T., CEROTTINI, J. and VAN DEN EYNDE, B., Tumor antigens recognized by T lymphocytes. Ann. Rev. Immunol., 12, 337–365 (1994). COHEN, P.J., COHEN, P.A., ROSENBERG, S.A., KATZ, S.I. and MULÉ, J.J., Murine epidermal Langerhans cells and splenic dendritic cells present tumor-associated antigens to primed T cells. Europ. J. Immunol., 24, 315–319 (1994). CONNOR, J., BANNERJI, R., SAITO, S., HESTON, W., FAIR, W. and GILBOA, E., Regression of bladder tumors in mice treated with IL-2 gene-modified tumor cells. J. exp. Med., 177, 1127–1133 (1993). FLAMAND, V., SORNASSE, T., THIELEMANS, K., DEMANET, C., BAKKUS, M., BAZIN, H., TIELEMANS, F., LEO, O., URBAIN, J. and MOSER, M., Murine dendritic cells pulsed in vitro with tumor antigen induce tumor resistance in vivo. Europ. J. Immunol., 24, 605–610 (1994). GILBOA, E. and LYERLY, H.K., Specific active immunotherapy of cancer using genetically modified tumor vaccines. In: Biologic Therapy of Cancer Updates, Vol. 4, pp. 1–16, J.B. Lippincott, Philadelphia (1994). GRABBE, S., BRUVERS, S., GALLO, R.L., KNISELY, T.L., NAZARENO, R. and GRANSTEIN, R.D., Tumor antigen presentation by murine epidermal cells. J. Immunol., 146, 3656–3661 (1991). GREENBERG, P.D., Adoptive T cell therapy of tumors: mechanisms operative in the recognition and elimination of tumor cells. Adv. Immunol., 49, 281–355 (1991). HUANG, A.Y.C., GOLUMBEK, P., AHMADZADEH, M., JAFFEE, E., PARDOLL, D. and LEVITSKY, H., Role of bone marrow-derived cells in presenting MHC class I-restricted tumor antigens. Science, 264, 961–965 (1994). LAFFERTY, K.J., PROWSE, S.J., SIMEONOVIC, C.J. and WARREN, H.S., Immunobiology of tissue transplantation: a return to the passenger leukocyte concept. Ann. Rev. Immunol., 1, 143–190 (1983). MATZINGER, P. and BEVAN, M., Induction of H-2 restricted cytotoxic T cells: in vivo induction has the appearance of being unrestricted. Cell. Immunol., 33, 92–100 (1977). MCKINNEY, E.C. and STREILEIN, J.W., On the extraordinary capacity of TUMOR EXTRACT-PULSED APC AND TUMOR IMMUNITY allogeneic epidermal Langerhans cells to prime cytotoxic T cells in vivo. J. Immunol., 143, 1560–1564 (1989). MOORE, M.W., CARBONE, F.R. and BEVAN, M.J., Introduction of soluble protein into the class I pathway of antigen processing and presentation. Cell, 54, 777–785 (1988). NAIR, S., BABU, J.S., DUNHAM, R.G., KANDA, P., BURKE, R.L. and ROUSE, B.T., Induction of primary, antiviral cytotoxic and proliferative responses with antigens administered via dendritic cells. J. Virol., 67, 4062–4069 (1993). PORGADOR, A., TZEHOVAL, E., VADAI, E., FELDMAN, M. and EISENBACH, L., Combined vaccination with major histocompatibility class I and interleukin 2 gene-transduced melanoma cells synergizes the cure of postsurgical established lung metastases. Cancer Res., 55, 4941–4949 (1995). ROCK, K.L., ROTHSTEIN, L., GAMBLE, S. and FLEISCHACKER, C., Characterization of antigen-presenting cells that present exogenous antigens in association with class I MHC molecules. J. Immunol., 150, 438–446 (1993). ROUSE, R.J.D., NAIR, S.K., LYDY, S.L., BOWEN, J.C. and ROUSE, B.T., Induction in vitro of primary cytotoxic T-lymphocyte responses with DNA encoding herpes simplex virus proteins. J. Virol., 68, 5685–5689 (1994). 715 SAITO, S., BANNERJI, R., GANSBACHER, B., ROSENTHAL, F.M., ROMANENKO, P., HESTON, W., FAIR, W. and GILBOA, E., Immunotherapy of bladder cancer with cytokine gene-modified tumor vaccines. Cancer Res., 54, 3516–3520 (1994). SHIMIZU, J., SUDA, T., YOSHIOKA, T., KOSUGI, A., FUJIWARA, H. and HAMAOKA, T., Induction of tumor-specific in vivo protective immunity by immunization with tumor antigen-pulsed antigen-presenting cells. J. Immunol., 142, 1053–1059 (1989). STEINMAN, R.M., The dendritic cell system and its role in immunogenicity. Ann. Rev. Immunol., 9, 271–296 (1991). TAKAHASHI, H., NAKAGAWA, Y., YOKOMURO, K. and BERZOFSKY, J.A., Induction of CD81 cytotoxic T lymphocytes by immunization with syngeneic irradiated HIV-1 envelope derived peptide-pulsed dendritic cells. Int. Immunol., 5, 849–857 (1993). ZITVOGEL, L., MAYORDOMO, J.I., TJANDRAWAN, T., DELEO, A.B., CLARKE, M.R., LOTZE, M.T. and STORKUS, W.J., Therapy of murine tumors with tumor peptide-pulsed dendritic cells: dependence on T cells, B7 costimulation, and T helper cell 1-associated cytokines. J. exp. Med., 183, 87–97 (1996).