Induction of specific antitumor immunity in the mouse with the electrofusion product of tumor cells and dendritic cells

ARTICLE

doi:10.1016/S1525-0016(03)00044-3

Induction of Specific Antitumor Immunity in the Mouse with the Electrofusion Product of Tumor Cells and Dendritic Cells William M. Siders,1 Kristin L. Vergilis,1 Carrie Johnson,1 Jacqueline Shields,1 and Johanne M. Kaplan1,2,* 1

Genzyme Molecular Oncology, 31 New York Ave., Framingham, Massachusetts 01701, USA 2 Genzyme Corporation, 31 New York Ave., Framingham, Massachusetts 01701, USA

*To whom correspondence and reprint requests should be addressed. Fax: (508) 872-4091. E-mail: [email protected].

Dendritic cells (DCs) are potent antigen-presenting cells capable of inducing primary T-cell responses. Several immunotherapy treatment strategies involve manipulation of DCs, both in vivo and ex vivo, to promote the immunogenic presentation of tumor-associated antigens. In this study, an electrofusion protocol was developed to induce fusion between tumor cells and allogeneic bone marrow-derived DCs. Preimmunization with irradiated electrofusion product was found to provide partial to complete protection from tumor challenge in the murine Renca renal cell carcinoma model and the B16 and M3 melanoma models. Vaccinated survivors developed specific immunological memory and were able to reject a subsequent rechallenge with the same tumor cells but not a syngeneic unrelated tumor line. Antitumor protection in the B16 model was accompanied by the development of a polyclonal cytotoxic T-lymphocyte response against defined melanoma-associated antigens. The therapeutic potential of this type of approach was suggested by the ability of a Renca-DC electrofusion product to induce tumor rejection in a substantial percentage of hosts (60%) bearing pre-established tumor cells. These results indicate that treatment with electrofused tumor cells and allogeneic DCs is capable of inducing a potent antitumor response and could conceivably be applied to a wide range of cancer indications for which tumor-associated antigens have not been identified.

INTRODUCTION Even though tumor cells express tumor-associated antigens that can be recognized by cytotoxic T lymphocytes (CTLs), they typically fail to induce a productive immune response. Their poor immunogenicity can be attributed to a variety of influences, including lack of T-cell costimulation [1], loss of MHC class I expression [2,3], lack of MHC class II presentation to CD4⫹ T-helper cells [4], and production of immunosuppressive factors [5]. Consequently, several immunotherapeutic approaches are being developed to achieve immunogenic presentation of tumorassociated antigens. Dendritic cells (DCs) have been used extensively in this context because they are potent antigen-presenting cells that possess both MHC class I and class II molecules as well as co-stimulatory and adhesion molecules [6]. Results from numerous studies have indicated that DCs pulsed with defined tumor-associated peptides [7,8], tumor lysate [9,10], apoptotic or necrotic tumor cells [11,12], as well as DCs genetically modified to express tumor-associated antigens using viral vectors [13,14] or tumor RNA [15,16], are all capable of eliciting a

498

tumor-specific cytotoxic T-cell response. Recently, several groups have reported that hybrid cells generated by the fusion of tumor cells and DCs also possess immunogenicity and are able to induce a protective immune response against tumor cells [17–20]. The fusion hybrids combine the antigens from the tumor cells with the antigen-presenting and co-stimulatory properties of DCs thus allowing for effective presentation of the full complement of potential antigens within the tumor, both known and unknown. To date, such hybrids have been most commonly generated by the fusion of tumor cells and DCs with polyethylene glycol (PEG) [17–20], a method that requires culturing of the tumor cells and hybrids before vaccination. In this study, an electrofusion protocol was used to promote the fusion of tumor cells and DCs. This approach presents practical advantages in terms of clinical application in that no culturing of the hybrids is required and, although tumor cell lines were used in our studies, the same method can be applied to cells freshly isolated from a tumor sample without any prior culturing 21–23; unpublished results]. Kugler et al. have used such

MOLECULAR THERAPY Vol. 7, No. 4, April 2003 Copyright © The American Society of Gene Therapy 1525-0016/03 $30.00

doi:10.1016/S1525-0016(03)00044-3

ARTICLE

FIG. 1. Electrofusion process. (A) An electrofusion product was generated by mixing equal numbers of tumor cells, in this instance B16 tumor cells, with dendritic cells, in this instance from BALB/c mice, in a waxed electroporation cuvette. (B) The mixture was then subjected to an alignment pulse to promote cell to cell contact and then (C) a fusion pulse to cause cell membrane fusion. The entire process was repeated a second time to maximize fusion efficiency.

an approach to produce human DC-tumor cell hybrids for the treatment of renal cell carcinoma and have reported encouraging clinical results [23]. Although clinical trials have already been conducted, there is actually little or no information on the performance and nature of the response elicited by electrofused DC-tumor cells in animal tumor models. In this study, an electrofusion protocol was developed and optimized and the activity of the electrofusion products generated was tested in tumor models possessing different growth and immunogenic properties. The level of efficacy, specificity, and longevity of the immune response induced by vaccination were investigated.

RESULTS Electrofusion Process An electrofusion method was used to induce the formation of DC-tumor cell hybrids. The process is illustrated in Figure 1. Bone marrow-derived DCs were mixed with tumor cells at a 1:1 ratio and added to an electroporation cuvette. The mixture was submitted to an alignment pulse to promote cell-cell contact, followed by a fusion pulse to induce cell membrane fusion. Delivery of the alignment and fusion pulses was then repeated a second time to maximize fusion. To establish optimal fusion conditions, various combinations of alignment and fusion pulses with varying intensities and durations were tested. The fusion efficiency resulting from each test condition was assessed by fluorescence-acti-

MOLECULAR THERAPY Vol. 7, No. 4, April 2003 Copyright © The American Society of Gene Therapy

vated cell sorting (FACS) analysis. The fusion products generated were stained for expression of both a tumor cell marker (for example, gp100 in the case of melanoma) and a DC marker not expressed on tumor cells (CD11b). The fusion efficiency was defined as the percentage of double-positive cells expressing both markers. Using the ECM 830 electroporator and the Gene Pulser II electroporator, we determined that optimal fusion conditions consist of two rounds of an alignment pulse of 50 V for 5 s followed by a fusion pulse of 250 V. This process typically yielded a fusion product containing 5–20% double-positive tumor-DC hybrids. This level of fusion is in agreement with results from investigators using similar electrofusion protocols [22– 24]. FACS analysis of a typical fusion product obtained with B16 tumor cells and DCs is shown in Figure 2. Staining of a mixture of B16 tumor cells and DCs with a fluorescein isothiocyanate (FITC)-labeled antibody against CD11b allowed for the identification of the DC population (Fig. 2A), whereas staining with an unlabeled primary antibody against the gp100 melanomaassociated antigen and then by a phycoerythrin (PE)conjugated secondary antibody served to identify the tumor cell population (Fig. 2B). Co-staining against both markers allows for the detection of double-positive hybrids generated subsequent to electrofusion (Fig. 2C). As shown in Figure 2D, simple mixing of the two cell populations did not result in a substantial number of double-positive cells, indicating a lack of fusion or transfer of markers.

499

ARTICLE

doi:10.1016/S1525-0016(03)00044-3

FIG. 2. FACS analysis of electrofusion products. A typical profile obtained with the fusion of B16 tumor cells and BALB/c DCs is shown in this figure. (A) DCs were stained with a FITC-labeled antibody against CD11b, whereas (B) tumor cells were stained with a PE label using an antibody against gp100. (C) Fusion efficiency was defined as the percentage of double-positive cells (18%). (D) Simple mixing of the DCs and tumor cells did not produce a substantial number of double-positive cells (1.5%).

Immunization Against Tumor Challenge The ability of irradiated, electrofused tumor-DC products to induce an antitumor response and provide protection against tumor challenge was evaluated in three different tumor models: (1) the murine B16 melanoma tumor model, characterized by an aggressive growth rate (palpable tumors by day 7–10) and low levels of MHC class I expression by the tumor cells; (2) the M3 melanoma tumor model, which displays slower growth kinetics (palpable tumors around day 20) and constitutive MHC class I expression by the tumor cells; and (3) the Renca renal cell carcinoma model, which is moderately aggressive (palpable tumors around day 14) and also possesses constitutive MHC class I expression. In these studies, mice were immunized intradermally with 5 ⫻ 105 to 2 ⫻ 106 electrofused cells on days 0 and 14. Dosing was based on the number of cells that underwent the electrofusion process (for example, a dose of 5 ⫻ 105 corresponds to the electrofusion product of 2.5 ⫻ 105 DCs and 2.5 ⫻ 105 tumor cells). The intradermal route of immunization was chosen because it has been reported that mature DCs injected by this route can traffic to the draining lymph nodes [25], and we have observed that labeling of electrofused cells with 5-(-6)-carboxyfluorescein diacetate (CFSE) allows for the detection of CFSE⫹ cells in the local draining lymph nodes as long as 24 hours after injection (results not shown).

500

One week after the second immunization (day 21), the mice were challenged with a lethal s.c. dose of tumor cells and were then monitored for tumor growth and survival over time. Representative results obtained in each tumor model are shown in Figure 3. The best level of antitumor protection was observed with the slowest growing, most immunogenic tumor type, that is, the M3 tumor model, in which 80 –100% of mice immunized with 2 ⫻ 105 to 5 ⫻ 105 electrofused cells were typically able to reject tumor challenge (Fig. 3A). In the more aggressive, poorly immunogenic B16 tumor model, preimmunization with fusion product was less effective and levels of antitumor protection ranged from 20 to 60% in mice immunized with 5 ⫻ 105 to 2 ⫻ 106 electrofused cells (Fig. 3B). In the moderately aggressive Renca tumor model, 50 –75% of mice immunized with 5 ⫻ 105 electrofused cells were typically able to reject tumor challenge and remain tumor-free (Fig. 3C). In contrast to results obtained with the fusion product, immunization with tumor cells that underwent the electrofusion process alone, DCs that underwent electrofusion alone, or a mixture of these two populations, failed to provide the same level of antitumor protection, indicating that optimal antitumor efficacy did in fact require the presence of both the tumor and the DC component during the electrofusion process and could not be solely attributed to immunization with inactivated tumor cells

MOLECULAR THERAPY Vol. 7, No. 4, April 2003 Copyright © The American Society of Gene Therapy

doi:10.1016/S1525-0016(03)00044-3

ARTICLE

FIG. 3. Induction of antitumor protection by immunization with tumor cell-allogeneic DC fusion vaccine. The antitumor response induced by irradiated tumor-DC electrofusion product was evaluated in three different tumor models. Representative results for each model are shown in this figure. (A) In the M3 melanoma model, DBA/2 mice were immunized with M3 cells electrofused to C57BL/6 DCs (dose of 5 ⫻ 105), or M3 cells that underwent the electrofusion process alone (2.5 ⫻ 105), on days 0 and 14 and then challenged s.c. with 1.5 ⫻ 104 M3 cells on day 21. Results are shown as the percentage of tumor-free mice on day 42 after tumor injection. (B) In the B16 melanoma model, C57BL/6 mice were immunized with B16 cells electrofused with BALB/c DCs (dose of 2 ⫻ 106), or B16 cells that underwent electrofusion alone (1 ⫻ 106), on days 0 and 14, and then challenged s.c. with 1.5 ⫻ 104 B16 cells on day 21. The percentage of tumor-free mice on day 42 after tumor injection is shown. (C) In the renal cell carcinoma Renca model, BALB/c mice were immunized with DCs that underwent the electrofusion process alone (2.5 ⫻ 105), or with Renca cells that underwent the electrofusion process alone (2.5 ⫻ 105), a mixture of the two (5 ⫻ 105), or Renca cells electrofused with C57BL/6 DCs (dose of 5 ⫻ 105), on days 0 and 14, and then challenged s.c. with 2 ⫻ 104 Renca cells on day 21. Results are shown as the percentage of tumor-free mice on day 60 after tumor injection. Naive unvaccinated mice were used as a negative control in each study.

or their by-products (Fig. 3). Moreover, in a fractionation study, immunization of mice with the supernatant fraction from a centrifuged Renca-DC electrofusion product (lysate, cell debris) provided little or no protection against challenge with Renca tumor cells (30-day median survival, 25% long-term survivors), and the antitumor activity appeared to reside primarily in the pelleted cellular component (50-day median survival, 50% long-term survivors vs. 65.5-day median survival, 50% long-term survivors with unfractionated vaccine). Comparison of Syngeneic Versus Allogeneic Dendritic Cells as Fusion Partner In the studies just shown, allogeneic DCs were selected as a fusion partner (for example, H-2d BALB/c DCs fused to H-2b B16 tumor cells) with the aim to stimulate highfrequency alloreactive CD4⫹ T cells and provide a potent source of help for the development of tumor-specific CTLs. The use of a tumor-allogeneic DC fusion product would be expected to have the advantage of simultaneously providing antigen presentation, co-stimulation, and allogeneic help in the same local microenvironment. To confirm the validity of this approach, we conducted a side-by-side comparison of the antitumor response elicited by fusion products generated with syngeneic versus

MOLECULAR THERAPY Vol. 7, No. 4, April 2003 Copyright © The American Society of Gene Therapy

allogeneic DCs. In the M3 tumor model, immunization with M3 tumor cells electrofused to syngeneic (H-2d DBA/2) or allogeneic (H-2b C57BL/6) DCs resulted in complete protection from tumor challenge in both instances, suggesting that both types of DCs can function as effective fusion partners (Fig. 4A). In the more stringent B16 model, immunization with a fusion product generated with tumor cells and allogeneic DCs (H-2d BALB/c) appeared to be superior to the product generated with syngeneic DCs as assessed by the greater percentage of mice capable of rejecting tumor challenge (Fig. 4B). Longevity and Specificity of the Antitumor Response As described earlier, immunization with DC-tumor fusion product was found to elicit protective activity against tumor challenge (Figs. 3 and 4). To determine whether the immunization resulted in long-term immunological memory, mice injected with M3-allogeneic DC fusion product that survived an M3 tumor challenge were rechallenged with a second lethal dose of tumor cells 56 days later. A parallel group of age-matched naive mice injected with M3 cells gradually succumbed to tumor growth. In contrast, all of the fusion product-immunized mice were able to reject this secondary tumor challenge, indicating that they had developed a long-term memory response

501

ARTICLE

doi:10.1016/S1525-0016(03)00044-3

were pooled and divided into four parallel sets of cultures that were stimulated with syngeneic SVB6KHA fibroblasts infected with adenovirus (Ad) vector encoding one of four melanoma-associated antigens—namely, gp100, MART-1, TRP-1, or TRP-2-to expand specifically any CTLs reactive against each of these antigens. As shown in Figure 6, effector cells recovered from these cultures showed low but detectable reactivity against the gp100, MART-1, and TRP-2 antigens, suggesting that a polyclonal CTL response against several antigens expressed by the tumor had developed as a result of immunization. As expected, lysis of control target cells infected with Ad vector lacking a transgene (Ad2/empty vector) was minimal. In addition, spleen cells from naive mice cultured under the same sets of conditions failed to show any substantial lytic activity against any of the targets (not shown).

FIG. 4. Activity of electrofusion products generated with syngeneic versus allogeneic DCs. A comparison of electrofusion products generated with allogeneic versus syngeneic DCs was conducted in the M3 and B16 melanoma models. (A) In the M3 model, DBA/2 mice were immunized with the irradiated electrofusion product of M3 tumor cells and syngeneic DBA/2 DCs or allogeneic C57BL/6 DCs (dose of 5 ⫻ 105) on days 0 and 14 and were then challenged s.c. with 1.5 ⫻ 104 M3 cells on day 21. (B) In the B16 model, C57BL/6 mice were immunized with the irradiated electrofusion product of B16 tumor cells and syngeneic C57BL/6 DCs or allogeneic BALB/c DCs (dose of 2 ⫻ 106) on days 0 and 14 and were then challenged s.c. with 1.5 ⫻ 104 B16 cells on day 21. Naive unvaccinated mice were used as a negative control in each study. Results are shown as the percentage of surviving mice over time after tumor challenge.

Therapeutic Activity of Tumor-DC Fusion Product To determine whether tumor-DC fusion vaccines may be useful in a clinically relevant setting, the therapeutic activity of Renca-allogeneic DC fusion product was tested in mice bearing pre-established Renca tumor cells. Mice were injected s.c. with 1.5 ⫻ 104 Renca cells on day 0 and were then treated with fusion vaccine at a dose of 5 ⫻ 105 or 2 ⫻ 106 electrofused cells on days 3, 7, and 14. The level of tumor rejection obtained was reduced compared with that observed in a preimmunization setting (Figs. 3 and 7). Little or no antitumor protection was observed in mice treated with 5 ⫻ 105 electrofused cells, a dose sufficient to achieve substantial tumor protection in pretreatment

(Fig. 5). The specificity of the immune response was also examined. Survivors from the second M3 tumor challenge were rechallenged with a lethal dose of the unrelated syngeneic P815 mastocytoma cell line. As shown in Figure 5, the mice rapidly developed P815 tumors and had to be killed because of excessive tumor burden. Taken together, these observations indicate that mice immunized with tumor-DC fusion product develop a specific long-term immune response. Induction of a Specific Cytotoxic T-Lymphocyte Response by Tumor-DC Fusion Product The presence of specific immunological memory in mice immunized with fusion product suggested that specific effector cells such as CTLs likely developed as a result of immunization. Because B16 tumor cells have been shown to express several known melanoma-associated tumor antigens [26,27], a CTL assay was done to determine whether immunization with B16-allogeneic DC fusion product resulted in the development of reactivity against specific melanoma tumor antigen(s). Spleen cells collected from mice immunized with the fusion product

502

FIG. 5. Longevity and specificity of the immune response elicited by electrofusion vaccine. DBA/2 mice were immunized with the irradiated electrofusion product of M3 tumor cells and allogeneic C57BL/6 DCs (dose of 5 ⫻ 105) on days 0 and 14 and were then challenged s.c. with 1.5 ⫻ 104 M3 cells on day 21. Vaccinated mice that survived the challenge (seven out of eight) received a second s.c. injection of M3 cells, 56 days after the first challenge. All immunized mice survived and were then challenged s.c. with 2 ⫻ 105 cells from the unrelated syngeneic P815 mastocytoma cell line, 128 days after the first M3 challenge. Results are shown as the percentage of surviving mice over time after tumor challenge. Naive unvaccinated mice served as a negative control.

MOLECULAR THERAPY Vol. 7, No. 4, April 2003 Copyright © The American Society of Gene Therapy

doi:10.1016/S1525-0016(03)00044-3

FIG. 6. CTL activity against individual melanoma antigens. C57BL/6 mice (n ⫽ 10) were immunized with the irradiated electrofusion product of B16 tumor cells and allogeneic BALB/c DCs (dose of 2 ⫻ 106) on days 0 and 14. Spleen cells collected on day 21 were pooled and divided into four parallel cultures that were stimulated with SVB6KHA fibroblasts infected with Ad vector encoding one of four known melanoma-associated antigens: gp100, MART-1, TRP-1, and TRP-2. After 6 days of culture, the cells recovered were tested for cytolytic activity against 51Cr-labeled fibroblasts infected with Ad vector lacking a transgene (Ad2/EV; background lysis) or Ad vector encoding the same antigen that was used for the in vitro expansion. Results are shown as the percentage of lysis for each of the targets over a range of effector/target ratios. Spleen cells from naive mice stimulated in the same manner failed to display any substantial cytolytic activity against any of the targets (% lysis ⱕ 8).

studies (Fig. 3C). However, substantial antitumor protection (60% long-term survivors) was obtained with a higher dose of the vaccine (2 ⫻ 106 electrofused cells). These results suggest that active treatment of tumor growth with tumor-DC electrofusion vaccines can provide a therapeutic benefit.

ARTICLE

against tumor cells. When a side-by-side comparison of the performance of fusion products generated with allogeneic versus syngeneic DCs was carried out, it appeared that syngeneic DCs were in fact quite potent and provided equal levels of tumor protection when used as a fusion partner in the M3 tumor model but that allogeneic DCs may provide an advantage against more aggressive tumor types such as the B16 melanoma (Fig. 4). The option of using allogeneic DCs as a fusion partner also presents a practical advantage because, in a clinical setting, allogeneic DCs can be generated conveniently from stored peripheral mononuclear cells from normal healthy volunteers from the general population. Interestingly, the tumor protective activity just described was obtained with electrofusion products containing only 5–20% actual tumor-DC hybrids as determined by FACS staining. The remainder of the product consists mostly of unfused cells, tumor-tumor hybrids, DC-DC hybrids, as well as debris and lysate from cells that die during the process. A fusion efficiency of 5–20% is typical of that reported by other investigators who used electrofusion methods and were able to induce antitumor immunity [21–23]. Most notably, Kugler et al. [23] obtained promising clinical results in renal cell carcinoma patients treated with allogeneic DC-autologous tumor cell fusion vaccines that contained only 10 –15% hybrids. It remains to be determined to what extent the hybrids themselves are responsible for the induction of antitumor immunity and to what extent other components of the product may contribute. Results from ourselves and others indicate that inactivated tumor cells or their by-products alone are not sufficient to induce optimal antitumor immunity and, as reported here, a mixture of DC and tumor cells that underwent electrofusion separately was also found to lack potency [18 –20,28,29] (Fig. 3). Mixtures of DCs and tumor cells were also found to be generally ineffective in the induction of antitumor immunity when compared with cells fused chemically with PEG [17,18]. These ob-

DISCUSSION The use of tumor-DC fusion hybrids for the induction of antitumor immunity is a promising new approach for the immunotherapy of cancer. Our results indicate that immunization with the fusion product generated by the electrofusion of allogeneic DCs and tumor cells is capable of inducing an immune response leading to tumor rejection in both pretreatment and therapeutic settings. The level of efficacy achieved varied in different tumor models and, not surprisingly, vaccination appeared to be most successful against less aggressive, MHC class I-bearing tumor cells such as the M3 melanoma tumor line. Allogeneic DCs were used as the fusion partner in most studies in light of the hypothesis that the foreign MHC class II molecules expressed by these cells would serve to recruit abundant, highly reactive allogeneic CD4⫹ T cells as a source of help for the development of a CTL response

MOLECULAR THERAPY Vol. 7, No. 4, April 2003 Copyright © The American Society of Gene Therapy

FIG. 7. Active treatment of mice bearing Renca tumor cells with electrofusion vaccine. BALB/c mice were injected s.c. with 1.5 ⫻ 104 Renca cells on day 0 and were then treated with the irradiated electrofusion product of Renca tumor cells and allogeneic C57BL/6 DCs at a dose of 5 ⫻ 105 or 2 ⫻ 106 on days 3, 7, and 14. Naive unvaccinated mice were used as a negative control. Results are shown as the percentage of surviving mice over time after tumor challenge.

503

ARTICLE

doi:10.1016/S1525-0016(03)00044-3

servations suggest that actual DC-tumor hybrids generated by fusion are required for activity of the product but do not rule out the possibility that other components may also be involved. Holmes et al. [30] used human tumor cells fused to autologous DCs with PEG to conduct a side-by-side comparison of the ability of unfractionated fusion product (10% hybrids) and FACS-purified hybrids (⬎95% purity) to induce tumor-specific CTLs in vitro. The purified hybrids stimulated the highest level of CTL activity (⬃70% specific lysis), but substantial cytolytic activity was also elicited by the unfractionated fusion product (⬃50% specific lysis), which contained a disproportionately much lower percentage of hybrids. These results may indicate that the hybrids are particularly potent or that other components of the preparation may add to their activity. One important advantage of immunization with an electrofusion product is the potential to induce an immune response against all possible tumor antigens, whether known or unknown, in any given haplotype. Such a possibility is supported by the observed development of a polyclonal CTL response against the gp100, MART-1, and TRP-2 antigens in mice immunized with a B16-allogeneic DC fusion product (Fig. 6). In addition, the use of a fusion product not only circumvents the need for the identification of tumor-associated antigens, but also potentially offers protection against the outgrowth of tumor escape variants that downregulate or lose expression of a given antigen under immunological pressure. In the face of a polyclonal response, it is less likely that tumor cells could undergo a coordinate loss of expression of several distinct antigens and evade the immune response entirely. The potential activation of natural killer cells and their relative contribution to the tumor rejection process is also a possibility that remains to be explored more fully. In summary, our results indicate that vaccination with electrofused tumor cells and allogeneic DCs is technically feasible and is capable of inducing a potent antitumor response. This type of approach could conceivably be applied to a wide range of cancer indications for which tumor-associated antigens have not been identified. Electrofusion conditions have been established for use with human cells (results not shown), and we have initiated an electrofusion clinical trial for the treatment of renal cell carcinoma.

MATERIALS

AND

METHODS

Mouse and cell lines. From Taconic Laboratories (Germantown, NY) were purchased 6- to 8-week-old C57BL/6, DBA/2, and BALB/c mice. The B16F10 melanoma cell line syngeneic to C57BL/6 mice was obtained from the National Cancer Institute (NCI; Bethesda, MD). The M3 cell line (Cloudman S91 melanoma) derived from melanocytes of the DBA/2 mouse, and the P815 mastocytoma derived from the same strain, were purchased from the American Type Culture Collection (ATCC; Manassas, VA). The renal cell carcinoma Renca cell line derived from the BALB/c mouse strain was

504

provided by Robert Wiltrout (NCI, Frederick, MD). All cell lines were confirmed to be mycoplasma-free by routine testing. Preparation of electrofusion vaccine. DCs were derived from bone marrow by negative selection of precursors and subsequent culture in medium containing granulocyte-macrophage colony-stimulating factor (GM-CSF) as described [14], and were matured for 2 days by the addition of 50 ng/ml recombinant mouse tumor necrosis factor-␣ (TNF-␣) to the culture medium (R&D Systems, Minneapolis, MN). For generation of an electrofusion product, DCs and tumor cells were each resuspended at a density of 107 cells/ml in a solution of 0.3 M glucose in water, pH 7.0. Equal volumes of the DC and tumor cell suspensions were mixed and an 800-␮l aliquot containing 4 ⫻ 106 DCs and 4 ⫻ 106 tumor cells was added to an electroporation cuvette. The cuvettes were precoated on one side with paraffin wax (Surgipath Medical Industries Inc., Richmond, IL; ⬃50 ␮l per cuvette) to create an inhomogeneous dielectrophoretic field as described by Kugler et al. [23]. The optimal conditions for maximum electrofusion efficiency without substantial cell death (⬃70% viability by Trypan Blue staining) were found to consist of two consecutive rounds of an alignment pulse of 50 V for 5 s administered with an Electro Square Porator ECM 830 electroporator from BTX (San Diego, CA) followed by a fusion pulse of 250 V administered by a Gene Pulser II electroporator (BioRad, Hercules, CA). The electrofusion product was then irradiated with 200 Gy with an RS 2000 Biological Irradiator X-ray irradiator (Rad Source Technologies Inc., Boca Raton, FL) to ensure inactivation of the tumor cells and DCs. Samples from several cuvettes were pooled (before irradiation) when required to achieve the number of cells needed for a study. The irradiated pooled sample was then diluted in 0.3 M glucose to the appropriate concentration for immunization. The dosing was based on the input number of cells that underwent the electrofusion process (for example, a dose of 5 ⫻ 105 corresponds to the electrofusion product of 2.5 ⫻ 105 DCs and 2.5 ⫻ 105 tumor cells). Assessment of electrofusion efficiency. The electrofusion efficiency was evaluated by FACS analysis and was defined as the percentage of cells that stained positive for both a tumor cell marker and a DC surface marker. CD11b was selected as a DC marker that was not shared by any of the tumor lines and was detected by direct staining with a FITC-labeled rat anti-mouse CD11b antibody (BD PharMingen, San Diego, CA). The tumorassociated gp100 antigen was selected as a marker for the B16 and M3 melanoma cell lines, whereas H-2d was chosen as a marker for the Renca cell line because this cell line lacks known tumor-associated antigens but could be distinguished from the H-2b allogeneic DC partner on the basis of MHC staining. After permeabilization, indirect staining for gp100 was accomplished using an unlabeled mouse anti-human gp100 primary antibody (DAKO Corporation, Carpinteria, CA) and subsequent PE-conjugated rat anti-mouse kappa-light chain secondary antibody (BD PharMingen). Direct staining for H-2d was done with a PE-labeled mouse antimouse H-2d ␣3 domain (BD PharMingen). Cells were analyzed for staining on a FACS Calibur system (Becton Dickinson, San Diego, CA). Fusion efficiencies typically ranged from 5 to 20% using the optimized electrofusion conditions just described. Immunization with electrofusion vaccine and tumor challenge. Groups of 8 –10 mice were immunized with doses of electrofused cells ranging from 2.5 ⫻ 105 to 2 ⫻ 106 as specified in the text. Fresh irradiated electrofusion product was diluted to the appropriate concentration in 0.3 M glucose and delivered intradermally with a 27-gauge needle in a total volume of 200 ␮l divided between two sites on the abdomen (100 ␮l per site). Injection into the skin resulted in the formation of a raised “bubble” indicative of an intradermal location, although some of the material may also have been absorbed into the subcutaneous layer. In the pretreatment model, the mice were immunized twice, once on day 0 and again on day 14. The electrofusion product was prepared fresh each time. A lethal subcutaneous tumor challenge was carried out 1 week after the last immunization (day 21). For tumor challenge, vaccinated C57BL/6 mice were injected with 1.5 ⫻ 104 B16-F10 cells, DBA/2 mice received 1.5 ⫻ 104 M3 cells, and BALB/c mice were challenged with 2.0 ⫻ 104 Renca cells. In the therapeutic model, tumor cells were injected on day 0 and vaccination

MOLECULAR THERAPY Vol. 7, No. 4, April 2003 Copyright © The American Society of Gene Therapy

ARTICLE

doi:10.1016/S1525-0016(03)00044-3

with electrofusion product was done on days 3, 7, and 14. Tumor size was measured twice a week with electronic digital calipers, and mice were killed when tumor size reached ⱖ 175 mm2. CTL assay. CTL activity was evaluated 1 week after the second immunization of C57BL/6 mice with B16-allogeneic DC electrofusion vaccine. Naive mice were used as a negative control. Pooled spleen cells from naive or immunized mice were stimulated in four parallel sets of cultures with mitomycin C-inactivated syngeneic SVB6KHA fibroblasts (gift from Linda Gooding, Emory University, Atlanta, GA) infected with Ad vector encoding one of four melanoma-associated antigens: gp100, MART-1, TRP-1, or TRP-2 [14,31]. Cells were cultured in 24-well plates containing 5 ⫻ 106 spleen cells and 6 ⫻ 104 stimulator fibroblasts in a 2-ml volume of culture medium without the addition of exogenous interleukin-2. Cytolytic activity was assayed after 6 days of culture. Target cells consisted of SVB6KHA fibroblasts infected for 48 hours with an Ad vector expressing one of the four melanoma-associated antigens or with an Ad vector lacking a transgene (Ad/empty vector) as a control. Target cells were treated with 100 U/ml recombinant mouse ␥-interferon (R&D Systems) for 24 hours to enhance MHC class I presentation, labeled with 51Cr (New England Nuclear, Boston, MA) overnight (25 ␮Ci/1 ⫻ 105 cells) and plated in roundbottom 96-well plates at 5 ⫻ 103 cells/well. Effector cells were added at various effector/target cell ratios in triplicate in a total volume of 200 ␮l. After a 5-hour incubation, 25 ␮l of cell-free supernatant was collected from each well and counted in a MicroBeta Trilux Scintillation Counter (Wallac, Gaithersburg, MD). The amount of 51Cr spontaneously released was obtained by incubating target cells in medium alone. Spontaneous release from target cells was typically ⬍20%. The total amount of 51Cr incorporated was determined by adding 1% Triton X-100 in distilled water, and the percentage lysis was calculated as follows: [(sample c.p.m.-spontaneous c.p.m.)/(total c.p.m.-spontaneous c.p.m.)] ⫻ 100.

8. 9. 10. 11.

12.

13.

14.

15.

16. 17. 18.

19. 20. 21.

ACKNOWLEDGMENTS We thank Kimberly Stencel for technical help and the Genzyme Animal Care and Technical Service group for animal husbandry. We also thank Michael Vasconcelles for useful discussions.

22. 23. 24.

RECEIVED FOR PUBLICATION DECEMBER 18; ACCEPTED JANUARY 30, 2003.

REFERENCES 1. Townsend, S. E., and Allison, J. P. (1993). Tumor rejection after direct costimulation of CD8⫹ T cells by B7-transfected melanoma cells. Science 259: 368 –370. 2. Maeurer, M. J., et al. (1996). Tumour escape from immune recognition: lethal recurrent melanoma on a patient associated with downregulation of the peptide transporter protein TAP-1 and the loss of the immunodominant MART-1/Melan-A antigen. J. Clin. Invest. 98: 1633–1641. 3. Scliger, B., Maeurer, M. J., and Ferrone, S. (1997). TAP off-tumour on. Immunol. Today 18: 292–299. 4. Ossendorp, F., Mengede, E., Camps, M., Filius, R., and Melief, C. J. (1998). Specific T helper cell requirement for optimal induction of cytotoxic T lymphocytes against major histocompatibility complex class II negative tumors. J. Exp. Med. 187: 693–702. 5. Wojtowicz-Praga, S. (1997). Reversal of tumor-induced immunosuppression: a new approach to cancer therapy. J. Immunother. 20: 165–177. 6. Steinman, R. (1991). The dendritic cell system and its role in immunogenicity. Annu. Rev. Immunol. 9: 271–296. 7. Mayordomo, J. I., et al. (1995). Bone marrow-derived dendritic cells pulsed with

MOLECULAR THERAPY Vol. 7, No. 4, April 2003 Copyright © The American Society of Gene Therapy

25.

26.

27. 28.

29. 30.

31.

synthetic tumour peptides elicit protective and therapeutic antitumour immunity. Nat. Med. 1: 1297–1302. Mayordomo, J. I., et al. (1997). Bone marrow-derived dendritic cells serve as potent adjuvants for peptide-based anti-tumor vaccines. Stem Cells 15: 94 –103. Nestle, F. O., et al. (1998). Vaccination of melanoma patients with peptide-or tumor lysate-pulsed dendritic cells. Nat. Med. 4: 328 –332. Lambert, L. A., et al. (2001). Intranodal immunization with tumor lysate-pulsed dendritic cells enhances protective antitumor immunity. Cancer Res. 61: 641– 646. Paczesny, S., Beranger, S., Salzmann, J. -L., Klatzmann, D., and Colombo, B.M. (2001). Protection of mice against leukemia after vaccination with bone marrow-derived dendritic cells loaded with apoptotic leukemia cells. Cancer Res. 61: 2386 –2389. Kotera, Y., Shimizu, K., and Mule, J. J. (2001). Comparative analysis of necrotic and apoptotic tumor cells as a source of antigen(s) in dendritic cell-based immunization. Cancer Res. 61: 8105– 8109. Brossart, P., Goldrath, A. W., Butz, E. A., Martin, S., and Bevan, M. J. (1997). Virusmediated delivery of antigenic epitopes into dendritic cells as a means to induce CTL. J. Immunol. 158: 3270 –3276. Kaplan, J. M., et al. (1999). Induction of antitumor immunity with dendritic cells transduced with adenovirus vector encoding endogenous tumor-associated antigens. J. Immunol. 163: 699 –707. Boczkowski, D., Nair, S. K., Snyder, D., and Gilboa, E. (1996). Dendritic cells pulsed with RNA are potent antigen-presenting cells in vitro and in vivo. J. Exp. Med. 184: 465– 472. Koido, S., et al. (2000). Induction of anti-tumor immunity by vaccination of dendritic cells transfected with MUC1 RNA. J. Immunol. 165: 5713–5719. Gong, J., Chen, D., Kashiwaba, M., and Kufe, D. (1997). Induction of antitumor activity by immunization with fusions of dendritic and carcinoma cells. Nat. Med. 3: 558 –561. Wang, J., Saffold, S., Cao, X., Krauss, J., and Chen, W. (1998). Eliciting T-cell immunity against poorly immunogenic tumors by immunization with dendritic cell-tumor fusion vaccines. J. Immunol. 161: 5516 –5524. Lespagnard, L., et al. (1998). Dendritic cells fused with mastocytoma cells elicit therapeutic antitumor immunity. Int. J. Cancer 76: 250 –258. Akasaki, Y., et al. (2001). Antitumor effect of immunizations with fusions of dendritic and glioma cells in a mouse brain tumor model. J. Immunother. 24: 106 –133. Kugler, A., et al. (1998). Autologous and allogeneic hybrid cell vaccine in patients with metastatic renal cell carcinoma. Br. J. Urol. 82: 487– 493. Trefzer, U., et al. (2000). Hybrid cell vaccination for cancer immune therapy: first clinical trial with metastatic melanoma. Int. J. Cancer 85: 618 – 626. Kugler, A., et al. (2000). Regression of human metastatic renal cell carcinoma after vaccination with tumor cell-dendritic cell hybrids. Nat. Med. 6: 332–336. Scott-Taylor, T. H., et al. (2000). Human tumour and dendritic cell hybrids generated by electrofusion: potential for cancer vaccines. Biochim. Biophys. Acta 1500: 265–279. Barratt-Boyes, S. M., et al. (2000). Maturation and trafficking of monocyte-derived dendritic cells in monkeys: implications for dendritic cell-based vaccines. J. Immunol. 164: 2487–2495. Zhai, Y., et al. (1997). Cloning and characterization of the genes encoding the murine homologues of the human melanoma antigens MART1 and gp100. J. Immunother. 20: 15–25. Bloom, M. B., et al. (1997). Identification of tyrosinase-related protein 2 as a tumor rejection antigen for the B16 melanoma. J. Exp. Med. 185: 453– 459. Stuhler, G., and Walden, P. (1994). Recruitment of helper T cells for induction of tumour rejection by cytolytic T lymphocytes. Cancer Immunol. Immunother. 39: 342– 345. Souberbielle, B. E., et al. (1998). Comparison of four strategies for tumour vaccination in the B16-F10 melanoma model. Gene Ther. 5: 1447–1454. Holmes, L. M., Li, J., Sticca, R. P., Wagner, T. E., and Wei, Y. (2001). A rapid, novel strategy to induce tumor cell-specific cytotoxic T-lymphocyte responses using instant dendritomas. J. Immunother. 24: 122–129. Zhai, Y., et al. (1996). Antigen-specific tumor vaccines. Development and characterization of recombinant adenoviruses encoding MART1 or gp100 for cancer therapy. J. Immunol. 156: 700 –710.

505