The transformation of a healthy cell into a malignant neoplasm involves numerous genetic mutations and aberrations in gene expression. As few of these changes are shared between individuals or types of cancer, the best source for eliciting broad-spectrum tumor immunity remains each patient's own tumor. Previously, we have shown that combining blockade of the T-cell–negative costimulatory molecule CTL-associated antigen 4 (CTLA-4) and vaccination with irradiated B16 tumor expressing granulocyte macrophage colony-stimulating factor (GM-CSF; Gvax) promotes rejection of established murine melanomas. Here we show that, like GM-CSF, the cytokine Flt3 ligand (Flt3L) expressed in B16 and coupled with CTLA-4 blockade promotes both prophylactic and therapeutic rejection of B16. When administered at the site of growing tumor, Gvax fails to prevent tumor outgrowth in any mice, whereas the B16-Flt3L vaccine (Fl3vax) induces the rejection of 75% of melanomas implanted 3 days before vaccination. Relative to Gvax, Fl3vax promotes greater infiltration of both the vaccine site and the tumor site by CD8+ T cells and “sentinel” and plasmacytoid dendritic cells. Gvax and Fl3vax did not synergize when used in combination in treating B16 melanoma even in the context of CD25+ regulatory T-cell depletion. Further, we show that a combination of Flt3L expression and CTLA-4 blockade can also promote the rejection of established TRAMP prostate adenocarcinomas, proving that the utility of this treatment extends beyond melanoma. Engineering Flt3L to be constitutively secreted and attaching an IgG2a tail yielded a B16 vaccine that, when combined with CTLA-4 blockade, prevented the outgrowth of significantly more 5-day implanted B16-BL6 tumors than did Gvax. [Cancer Res 2009;69(19):7747–55]

Irradiated tumor cells can be reintroduced as a vaccine; however, the induced antitumor responses are ineffective due to lack of immune costimulation (1). Granulocyte macrophage colony-stimulating factor (GM-CSF) has proved the most capable adjuvant for transforming irradiated tumor vaccines into mediators of tumor protection and sometimes rejection (2). Cellular tumor vaccines rely on cross-presentation of tumor antigens by professional antigen-presenting cells (APC; ref. 3). GM-CSF enhances this process both by inducing differentiation and maturation of APCs such as dendritic cells (DC) and by chemoattracting granulocytes, macrophages, and lymphocytes to the vaccine site (4).

We sought to determine whether other molecules that may enhance cross-presentation could substitute for GM-CSF in converting irradiated tumor into an effective vaccine. The cytokine Fms-like tyrosine kinase 3 ligand (Flt3L) supports the survival, proliferation, and differentiation of hematopoietic progenitors (5) and both induces and chemoattracts DCs (6, 7). Recombinant Flt3L promotes tumor regression in some tumor models (8), but alone cannot reject preimplanted poorly immunogenic tumors like B16 melanoma. The chemokine IP10/CXCL10 is known both to chemoattract monocytes, natural killer (NK) cells, and TH1 lymphocytes (9) and to have antiangiogenic properties (10). IP10 can function in immunotherapy and has been associated with graft rejection and GvHD in transplantation (1113). The chemokine MCP-3/CCL7 potently attracts lymphoid and myeloid cells (1417). MCP-3 can also promote antitumor responses and has been implicated in graft rejection and GvHD (1820). The adjuvant activity of GM-CSF for tumor vaccination has been best characterized in the B16 melanoma model; thus, we compared these candidate adjuvants in this setting.

Vaccination with irradiated B16 expressing GM-CSF can protect mice against subsequent tumor challenge and can slow the growth of, but not reject, preimplanted tumors (2). To reject established melanomas, B16-GMCSF (Gvax) vaccination must be combined with antibody blockade of the T-cell–negative costimulatory receptor CTL-associated antigen 4 (CTLA-4; ref. 21). Blockade of CTLA-4/B7 binding using antibodies exerts a powerful adjuvant effect on T cells and can induce the rejection of many transplantable tumors. In the case of less immunogenic tumors such as B16 melanoma or TRAMP prostate adenocarcinomas, CTLA-4 blockade alone may slow the growth of, but cannot reject, preimplanted tumors (2224). The combination of CTLA-4 blockade and Gvax vaccination, however, synergizes to induce rejection of established tumors in these models (21, 25, 26) and in some clinical trials (27, 28).

Using retroviral vectors, we created B16 cell lines expressing GM-CSF (Gvax), Flt3L (FL3vax), IP10, or MCP-3. We measured the efficacy of these vaccines in prophylaxis and treatment in combination with CTLA-4 blockade. We next compared the relative efficacy of each vaccine when administered at the tumor site versus on the opposite flank. These experiments revealed distinct treatment profiles for GM-CSF–, Flt3L-, and IP10–expressing vaccines and strongly validated Flt3L as a cellular tumor adjuvant. We found that Fl3vax and Gvax did not synergize when used for combination treatment of B16 melanomas or TRAMP adenocarcinomas. We thoroughly characterized the lymphocytic infiltrates of both the vaccine sites and tumors of mice receiving Gvax, Fl3vax, or thecombination in the context of CTLA-4 blockade. In addition, we sought to determine the utility of Flt3L as a cellular tumor adjuvant outside of the B16 system by comparing it to GM-CSF for the treatment ofestablished TRAMP tumors. Finally, we reengineered our FL3vax to optimize the bioactivity of Flt3L and compared the enhanced vaccine to Gvax in treating 5-day established B16 melanomas.

Mice. C57BL/6 mice (4–6 wk old males, The Jackson Laboratory) were cared for in accordance with the NIH and Association for Assessment and Accreditation of Laboratory Animal Care regulations. Experiments were all approved by the Memorial Sloan-Kettering Cancer Center Institutional Animal Care and Use Committee.

Antibodies. Mouse α-CTLA-4 (9D9), hamster α-CTLA-4 (9H10), α-CD25 (PC61), α-CD4 (GK1.5), α-CD8 (2.43), and α-NK1.1 (PK136) were purified at the Memorial Sloan-Kettering Cancer Center antibody facility. Antibodies and gating for 9-color flow cytometry are outlined in Supplementary Table S2.

Cell lines. B16/BL6 cells were used for tumor challenge and for the creation of all vaccine lines and were maintained as described (21). TRAMP-C2 cells were used in all prostate cancer experiments and for creation of TRAMP-GMCSF and TRAMP-Flt3L and grown as described (29).

Retroviral vectors and virus production. Murine GM-CSF, Flt3L, IP10, and MCP-3 cDNAs were cloned into the pMG-Lyt2 retroviral vector. This vector resembles pGC-IRES except that a truncated form of murine CD8α is used for selection (30). Recombinant virus production and infection were done as described, except that vesicular stomatitis virus-G and 10A1 envelope proteins were used (31).

Tumor Challenge and Treatment Experiments

B16 tumor treatment. Mice were injected in the flank intradermally (i.d.) at day 0 with the indicated number of B16-BL6 cells and treated ondays 3, 6, and 9 (5, 8, and 11 for Figs. 3 and 6) with 1 × 106 irradiated (150 Gy) gene-modified B16 cells on the contralateral flank and 100-μg α-CTLA-4 (9D9) i.p. All vaccinations were on the opposite flank from tumor except in Fig. 3, with “local” groups receiving their vaccinations on the same flank as the tumor challenge. For FL3vax/Gvax combination vaccines, all were normalized to a total dose of 2 × 106 B16 cells per vaccine by addition of B16-Lyt2 cells. Where indicated, regulatory T-cell (Treg) depletion was achieved by a single 400-μg injection of α-CD25 on day −4.

TRAMP-C2 Tumor Treatment

Mice were injected in the flank i.d. at day 0 with 1 × 106 TRAMP-C2 prostatic adenocarcinoma cells and treated on days 2, 5, and 8 with 1 × 106 irradiated (120 Gy) gene-modified TRAMP-C2 cells on the contralateral flank and 100-μg α-CTLA-4 (9H10) i.p.

Vaccine Site Infiltration Analysis

Mice received two gene-modified B16 vaccines i.d., 4 d apart, in 30% collagen matrix (Matrigel, BD) coupled with 100 μg of α-CTLA-4 i.p. Gvax, Fl3vax, and Gvax/Fl3vax vaccines were formulated as described above. One day after the second vaccination, mice were euthanized, the vaccine site was excised, and infiltrating lymphocytes were obtained after disruption and Ficoll purification. Infiltrating cells were stained with fluorophore-conjugated monoclonal antibodies and analyzed by nine-color flow cytometry on a Cyan flow cytometer (Dako; Supplementary Table S2).

Tumor Infiltration Analysis

Mice were injected in the flank i.d. at day 0 with 10,000 B16/BL6 cells in 30% collagen matrix (Matrigel, BD). Mice were vaccinated as described for day −3 tumor treatment. Thirteen to 15 d after tumor challenge, mice were sacrificed, and tumor-infiltrating lymphocytes (TIL) were obtained from tumors and purified and stained as described earlier for intra-vaccine lymphocytes.

We compared the adjuvant efficacy of two chemokines, MCP-3 and IP10, and the cytokine Flt3L to the established adjuvant cytokine GM-CSF (Supplementary Table S1). Retroviral vectors were constructed with a marker gene directly linked to the adjuvant gene, ensuring that the levels of the two genes were proportional (Supplementary Fig. S1A). Cell lines were normalized by cell sorting to equivalent high levels of marker gene expression (Supplementary Fig. S1B and C) and then expression of the adjuvant gene was confirmed by intracellular cytokine staining (Supplementary Fig. S1D). This system allowed rapid generation of vaccine cell lines expressing levels of all adjuvant genes within or exceeding the known ED50 ranges by ELISA (Supplementary Table S1).

Flt3L and IP10 are effective adjuvants for B16 prophylactic vaccination. Mice were vaccinated twice with irradiated tumor cells before being challenged with B16-BL6 melanoma cells on the opposite flank. Compared with B16-YFP, B16-GMCSF (Gvax), B16-Flt3L (FL3vax), and B16-IP10 all showed statistically significant activity in preventing tumor growth (Fig. 1A). Although MCP-3 protected more mice than B16-YFP alone, its effect was not significant.

Figure 1.

Fl3vax, like Gvax, synergizes with CTLA-4 blockade to treat 3-d preimplanted B16 melanomas. A, percent tumor-free mice following vaccination on days −4 and −7 (n = 4, 5–15 mice per group). *, P < 0.05, compared with B16-YFP (Fisher's exact test). B, tumor growth is shown for monotherapy treatment of 5 mice per group on days 3, 6, and 9. C, Kaplan-Meier survival curves are shown for vaccine with CTLA-4 blockade on days 3, 6, and 9, where the end point was defined as tumor burden ≥1,000 mm3 (n = 3, 10 mice per group). Surviving mice were tumor-free. D, tumor growth is shown for a representative experiment of three.

Figure 1.

Fl3vax, like Gvax, synergizes with CTLA-4 blockade to treat 3-d preimplanted B16 melanomas. A, percent tumor-free mice following vaccination on days −4 and −7 (n = 4, 5–15 mice per group). *, P < 0.05, compared with B16-YFP (Fisher's exact test). B, tumor growth is shown for monotherapy treatment of 5 mice per group on days 3, 6, and 9. C, Kaplan-Meier survival curves are shown for vaccine with CTLA-4 blockade on days 3, 6, and 9, where the end point was defined as tumor burden ≥1,000 mm3 (n = 3, 10 mice per group). Surviving mice were tumor-free. D, tumor growth is shown for a representative experiment of three.

Close modal

The YFP marker created high variability in the B16-YFP control (Fig. 1A) and was difficult to normalize in expression between lines (Supplementary Fig. S1B); thus, for all subsequent experiments, a nonimmunogenic truncated Lyt2 marker was used instead. These Lyt2 vectors yielded similar prophylactic data to the YFP vectors (Supplementary Fig. S2) but were easier to normalize (Supplementary Fig. S1C) and lacked the potentially confounding immunogenicity of YFP. Compared with B16 alone, B16-Lyt2 grew at the same rate in untreated mice (data not shown).

B16-Flt3L is as effective as B16-GMCSF in treating 3-day implanted B16 melanomas when combined with antibody blockade of CTLA-4. To compare our adjuvant genes to GM-CSF, we tested their ability to synergize with CTLA-4 antibody blockade to treat 1 × 104 B16 melanoma cells implanted 3 days earlier. Individually, neither any of the cellular vaccines nor CTLA-4 blockade alone slows tumor growth (Fig. 1B) or is curative. When coupled with CTLA-4 blockade, treatment with Gvax or Fl3vax resulted in 63% and 60% tumor-free mice (Fig. 1C). By comparison, IP10 and MCP-3 showed much less capacity to prolong the survival of B16-bearing mice. Gvax and Fl3vax were also most effective in slowing tumor growth (Fig. 1D), whereas B16-MCP3 had no effect on tumor growth and was not studied further.

FL3vax and B16-IP10 increase in effectiveness when administered at the tumor site, whereas Gvax fails to prevent tumor outgrowth in any mice. As melanoma lesions can sometimes be directly accessed for treatment, we tested the efficacy of Gvax, Fl3vax, and B16-IP10 both locally (at the tumor site) and distally (on the opposite flank) in combination with CTLA-4 blockade.

Surprisingly, Gvax showed vastly diminished efficacy in treating mice that had been challenged with 2 × 104 B16 cells 3 days prior, when administered locally versus distally in terms of both tumor-free survival (Fig. 2A) and tumor size at first measurement (Fig. 2B). No tumor growth curve is provided because granulomas evoked bythe vaccine at the tumor site cannot be distinguished from tumor for 3 to 4 weeks. Unlike Gvax, Fl3vax was significantly more effectivewhen administered at the tumor site versus the opposite flank, resulting in 75% tumor-free mice. B16-IP10, which showed little effect distally, proved to be effective when given at the tumorsite.

Figure 2.

Fl3vax with CTLA-4 blockade prevents tumor outgrowth in nearly all mice when given at the tumor site where Gvax proves ineffective, whereas B16-IP10 is effective locally but not distally. A, percent tumor-free mice following vaccination on days 3, 6, and 9 (n = 2, 10 mice per group, except PBS vaccination was only distal to 5 mice per group). B, mean tumor size at the first measurement for one representative experiment of two independent experiments (10 mice per group). Growth curves could not be generated in these studies as no measurements could be made for 3 to 4 wk due to postvaccination granulomas at the tumor site.

Figure 2.

Fl3vax with CTLA-4 blockade prevents tumor outgrowth in nearly all mice when given at the tumor site where Gvax proves ineffective, whereas B16-IP10 is effective locally but not distally. A, percent tumor-free mice following vaccination on days 3, 6, and 9 (n = 2, 10 mice per group, except PBS vaccination was only distal to 5 mice per group). B, mean tumor size at the first measurement for one representative experiment of two independent experiments (10 mice per group). Growth curves could not be generated in these studies as no measurements could be made for 3 to 4 wk due to postvaccination granulomas at the tumor site.

Close modal

Gvax can induce myeloid suppressor cells under certain conditions, suggesting that the failure of Gvax proximal to tumor may be due to the interaction of the myeloid cells evoked by the vaccine and the tumor microenvironment (32, 33). Our Gvax line produces 360 ng/1 × 106 cells/24 hours of GM-CSF, which is within the optimal immunogenicity range and well below the levels that have been described as inherently suppressive (32, 34). These data indicated that there were likely to be significant differences in the cellular subsets evoked by each vaccine; therefore, we sought to determine if Gvax and Fl3vax would cooperate in eradicating B16 melanomas.

Combination treatment with Fl3vax and Gvax is no more effective than using either vaccine alone. To test the potential for Fl3vax and Gvax to synergize in rejecting B16 melanomas, we waited until 5 days following tumor challenge to begin treatment. Gvax and FL3vax showed similar efficacy; however, combination treatment with both vaccines prevented tumor outgrowth in equal or fewer mice compared with either alone (Fig. 3A). In addition, tumors in mice receiving the dual cytokine vaccine grew at similar or faster rates compared with tumors in mice receiving either vaccine alone (Fig. 3B).

Figure 3.

Fl3vax and Gvax fail to synergize when used in combination to treat B16 melanomas implanted 5 d prior, even in mice pre-depleted of Tregs. A, percent tumor-free mice following vaccination on days 3, 6, and 9 (n = 3, 10 mice per group). B, tumor growth is shown for a representative experiment of three. C and D, replicates of A and B except that mice were pre-depleted of CD25+ cells by PC61 injection on day −4 (n = 2, 10 mice per group).

Figure 3.

Fl3vax and Gvax fail to synergize when used in combination to treat B16 melanomas implanted 5 d prior, even in mice pre-depleted of Tregs. A, percent tumor-free mice following vaccination on days 3, 6, and 9 (n = 3, 10 mice per group). B, tumor growth is shown for a representative experiment of three. C and D, replicates of A and B except that mice were pre-depleted of CD25+ cells by PC61 injection on day −4 (n = 2, 10 mice per group).

Close modal

Both B16-Flt3L and B16-GMCSF vaccinations are known to elicit Tregs that dampen antitumor responses (35, 36). To determine if these two cytokines would synergize in the absence of Tregs, we repeated the above experiments in mice that had been pre-depleted of CD25+ cells. Once again, the combination of Gvax and Fl3vax was clearly no more effective than either vaccine alone (Fig. 3C), although the potency of all vaccines was enhanced by Treg pre-depletion. Tumors in mice vaccinated with both B16-Flt3L and B16-GMCSF, although small, grew at similar or faster rates compared with tumors in mice receiving either vaccine alone (Fig. 3D). To further understand the mechanisms of action of each these vaccines, we examined the cells infiltrating both the vaccine and tumor sites for each.

Fl3vax treatment relies on CD8+ T cells, NK1.1+ cells, and CD4+ cells and is opposed by the action of CD4+CD25+ Tregs. Published analysis for Gvax indicated that CD8+ T cells and NK1.1+ cells were the critical populations for rejecting B16 tumors (26). We found that in mice depleted of CD8+ cells or NK1.1+ cells, Fl3vax + α-CTLA-4 vaccination lost virtually all capacity to promote rejection of tumors implanted 3 days earlier (Supplementary Fig. S3A), and tumors grew at rates similar to those in untreated mice (Supplementary Fig. S3B). In addition, a much higher percentage of CD4-depleted mice developed tumors, and those tumors grew faster than the tumors in mice pre-depleted of Tregs (which are also depleted by α-CD4). Thus, we found that CD8+ T cells, CD4+ Tcells, and NK1.1+ cells all play important roles in rejecting tumors following Fl3vax vaccination with CTLA-4 blockade. We hoped that a more detailed study of the vaccine and TIL populations evoked by each vaccine would further clarify the adjuvant mechanics of each, and perhaps suggest reasons for their lack of synergy.

Relative to Gvax, Fl3vax vaccination sites contain a higher percentage of CD8+ T cells, NK cells, plasmacytoid DCs, and CD11b−/lo DCs. Mice were given two vaccinations with Gvax, FL3vax, or a combination of both with α-CTLA-4 antibody, and 24hours later, the lymphocytes infiltrating the vaccine site were isolated and typed (Supplementary Table S2). For both Gvax and Fl3vax, the vaccination site infiltrate was dominated by granulocytes and macrophages (Fig. 4A). Despite this similarity, the B16-Flt3L site contained a 2- to 3-fold higher fraction of B cells, NK cells, and CD11b−/lo DCs relative to the B16-GMCSF site. Most strikingly, plasmacytoid DCs (pDC) and CD8+ T cells comprised 5.5-fold and 8.5-fold higher fractions of the Fl3vax site compared with that of Gvax. Tregs, granulocytes, and monocytes/CD11bmed/hi DCs were more abundant in the B16-GMCSF site. The CD11b−/lo DC population favored by Flt3L consists primarily of “sentinel” DCs and pDCs, whereas the CD11c+CD11bmed/hi population elicited by GM-CSF is a mixture of monocytes and “inflammatory” DCs (37). The most striking difference between the Gvax/Fl3vax combination and either vaccine alone was the relative lack of NK cells and Tcells,especially CD8+ T cells.

Figure 4.

Fl3vax elicits strong infiltration of the vaccine and tumor sites by CD8+ T cells, sentinel DCs, and pDCs. A, percent of the total vaccine site infiltrating lymphocytes for each subset and the indicated vaccine (n = 2, 5 individually analyzed mice per group). The percent infiltration relative to Gvax is also shown. B, percent composition of total TIL for each subset and the indicated vaccine (n = 3, 5 individually analyzed mice per group). The percent infiltration relative to Gvax is also shown. C, changes in intratumoral CD8+ T cells, CD4+ Tregs, and the resulting intratumoral CD8+ T-cell/Treg ratios. D, percent of TIL composed of CD11b+GR1+ cells for each treatment group.

Figure 4.

Fl3vax elicits strong infiltration of the vaccine and tumor sites by CD8+ T cells, sentinel DCs, and pDCs. A, percent of the total vaccine site infiltrating lymphocytes for each subset and the indicated vaccine (n = 2, 5 individually analyzed mice per group). The percent infiltration relative to Gvax is also shown. B, percent composition of total TIL for each subset and the indicated vaccine (n = 3, 5 individually analyzed mice per group). The percent infiltration relative to Gvax is also shown. C, changes in intratumoral CD8+ T cells, CD4+ Tregs, and the resulting intratumoral CD8+ T-cell/Treg ratios. D, percent of TIL composed of CD11b+GR1+ cells for each treatment group.

Close modal

We next decided to investigate how the differences observed in the composition of the vaccine sites would affect the profile of the lymphocytes infiltrating the tumor.

Fl3vax induces higher percentages of CD8+ T cells, Tregs, sentinel DCs, and pDCs and fewer CD11b+GR1+ cells in tumors relative to those of mice receiving Gvax. To dissect the differential effects of Gvax, Fl3vax, and combination vaccination in treating B16, we undertook a more comprehensive analysis of the TIL population. Following tumor implantation and three vaccinations, TIL were isolated and characterized (Fig. 4B). We found that overall TIL frequency (i.e., the extent of infiltration) in mice receiving the Gvax, Fl3vax, and Gvax/Fl3vax vaccines with CTLA-4 blockade was similar and substantially higher than in untreated mice (Supplementary Fig. S4). Similar to the vaccine site infiltrate, TIL from Fl3vax-treated mice contained higher proportions of sentinel DCs, pDCs, NKDCs, and CD8+ T cells and lower levels of inflammatory DCs relative to TIL from mice receiving Gvax. Unlike the vaccine site, FL3vax-inoculated mice also had higher levels of Tregs in their tumors. Despite having large numbers of granulocytes, macrophages, and DCs, TIL from the combination vaccinated mice contained the lowest percentage of CD8+ T cells.

Tregs suppress cytotoxic T cells, and in the case of B16, elevated ratios of CD8+ T cells to FoxP3+ Tregs within the tumor correlate with successful treatment (35). Both Gvax and Fl3vax increase the ratio of CD8+ T cells to Tregs within the tumor, especially in conjunction with α-CTLA-4 antibody (Fig. 4C). Of the mice receiving CTLA-4 blockade, this ratio is lowest in the combination-vaccine group but not to an extent, which explains its reduced immunogenicity. Although Tregs do make up a larger percentage of the TIL elicited by Fl3vax relative to Gvax, the even higher level of CD8+ T cells also elicited by Fl3vax yielded the highest CD8/Treg ratios in TIL.

As CD11b+,GR1+ candidate myeloid suppressor cells can be classified as either macrophages or granulocytes in our typing system depending on the other markers they express, we decided to analyze their relative enrichment in the tumors of vaccinated mice as a separate population. We find that vaccination with Gvax, with or without CTLA-4 blockade, results in significantly higher percentages of CD11b+,GR1+ cells in TIL compared with mice receiving FL3vax or compared with untreated mice (Fig. 4D). This enhanced infiltration by myeloid suppressor cells may explain the failure of Gvax to cure tumors when given at the tumor site where much higher concentrations of GM-CSF would be present around the tumor.

Having shown the efficacy of Fl3vax for treating preimplanted B16 melanomas in conjunction with CTLA-4 blockade, we sought to determine whether it would also have utility for the treatment of other poorly immunogenic tumors.

Flt3L expression synergizes with CTLA-4 blockade in rendering irradiated autologous TRAMP-C2 cells capable of preventing the outgrowth of 2-day implanted TRAMP-C2 prostate adenocarcinomas. We used retroviral vectors to create GM-CSF– and Flt3L-expressing vaccines derived from the TRAMP-C2 prostate adenocarcinoma. The majority of TRAMP-C2 tumors cannot be cured by either Gvax or CTLA-4 blockade alone, but are susceptible to combination therapy. We found that both TRAMP-GMCSF and TRAMP-Flt3L could protect 100% of mice from outgrowth of 1-day preimplanted low-dose (5 × 105 cells) TRAMP-C2 tumors when combined with CTLA-4 blockade (data not shown). For this reason, we sought to block the outgrowth of a higher TRAMP-C2 challenge (1 × 106 cells) with vaccination on days 2, 5, and 8.

Treatment with TRAMP-GMCSF, TRAMP-Flt3L, and a combination of both resulted in similar percentages of tumor-free mice when combined with CTLA-4 blockade (Fig. 5A). The overall tumor growth rate was also not significantly different between mice receiving either vaccine (Fig. 5B). As with B16, combination therapy with TRAMP-GMCSF and TRAMP-Flt3L failed to show any efficacy beyond that of either treatment alone. These data show that the adjuvant utility of Flt3L extends beyond the B16 melanoma system.

Figure 5.

Flt3L converts irradiated TRAMP-C2 prostatic adenocarcinoma cells into an effective vaccine, which synergizes with CTLA-4 blockade to protect against outgrowth of preimplanted TRAMP tumors. A, percent tumor-free mice following vaccination on days 2, 5, and 8 (n = 2, 10 mice per group). B, tumor growth from a representative experiment of two.

Figure 5.

Flt3L converts irradiated TRAMP-C2 prostatic adenocarcinoma cells into an effective vaccine, which synergizes with CTLA-4 blockade to protect against outgrowth of preimplanted TRAMP tumors. A, percent tumor-free mice following vaccination on days 2, 5, and 8 (n = 2, 10 mice per group). B, tumor growth from a representative experiment of two.

Close modal

An improved FL3vax synergizes with CTLA-4 blockade and protects more mice from outgrowth of B16 melanoma implanted 5 days before treatment than does Gvax. Whereas GM-CSF has been shown to plateau in adjuvant efficacy at a relatively modest level (34), we wondered if the efficacy of Flt3L might increase with higher levels of expression. A preliminary study suggested that Gvax decreased in efficacy with increasing dose, whereas FL3vax maintained or slightly increased its effect at higher doses (Supplementary Fig. S5). This system suffered from many limitations, however, including veterinary complications due to the large vaccine dose (5 × 106 per injection), as well as the potential to saturate the amount of serum protease available to release Flt3L from the cell membrane in its bioactive form. We decided instead to create two new vaccines that constitutively secrete Flt3L. One of these is truncated at the site where membrane cleavage occurs in the wild-type form (secFL3vax), and the other is truncated at the same site and joined to a mouse IgG2a constant region for increased stability and to enable dimerization (sFL3vaxIg). Each of these lines was normalized to the same Lyt2 expression level as FL3vax and found to produce between 900 and 1,000 ng/1 × 106 cells/24 hours of Flt3L in vitro.

To compare these new vaccines to the original FL3vax and Gvax, we tested their ability to synergize with CTLA-4 blockade to block outgrowth of B16 melanoma cells injected 5 days before treatment. In this setting, sFL3vaxIg cured a significantly higher percentage of mice than did Gvax or any of the other Flt3L-based vaccines (Fig. 6A). The tumors of mice receiving the sFL3vaxIg vaccine with CTLA-4 blockade also grew at a significantly slower rate than those of any other treatment group in each of three individual experiments (Fig. 6B). These experiments showed that this enhanced Flt3L-expressing B16 vaccine was superior to Gvax for use in treating B16 melanoma in conjunction with CTLA-4 blockade.

Figure 6.

A secreted, immunoglobulin-tailed FL3vax synergizes with CTLA-4 blockade and more effectively treats mice of 5-d preimplanted B16 melanomas than does Gvax. A, percent tumor-free mice following vaccination on days 5, 8, and 11 (n = 3, 10 mice per group). B, tumor growth from a representative experiment of three.

Figure 6.

A secreted, immunoglobulin-tailed FL3vax synergizes with CTLA-4 blockade and more effectively treats mice of 5-d preimplanted B16 melanomas than does Gvax. A, percent tumor-free mice following vaccination on days 5, 8, and 11 (n = 3, 10 mice per group). B, tumor growth from a representative experiment of three.

Close modal

We pursued a rational approach to determine if molecules designed to enhance infiltration of a vaccine site by APCs could substitute for GM-CSF in rendering autologous tumor immunogenic. In our hands, Flt3L and IP10 function effectively as adjuvants for B16 melanoma vaccination, as they protected mice from subsequent tumor challenge when used prophylactically (Fig. 1A), and were capable of effectively treating preimplanted B16 tumors when combined with CTLA-4 blockade (Fig. 1C). Most impressively, Fl3vax, like Gvax, proved capable of synergizing with α-CTLA-4 treatment to completely prevent the outgrowth of a majority of 3-day established B16 melanomas (Fig. 1C and D).

As we investigated the application and mechanics of FL3vax further, we found that in some settings, its efficacy exceeded that of Gvax. One of the most striking findings we report is the inability of Gvax + α-CTLA-4 therapy to prevent tumor growth in any mice bearing 3-day preimplanted B16 melanomas when given at the tumor site versus on the opposite flank (Fig. 2). In contrast, Fl3vax showed similar efficacy to Gvax when given on the opposite flank, but resulted in 75% of mice being tumor-free when given locally. As further evidence for the lack of local immunogenicity of Gvax, we and others have found little or no reduced growth of B16-GMCSF cells in vivo relative to untransduced B16 cells (38). Most interestingly, the Fl3vax-elicited infiltrate seems to be resistant to co-optation by the tumor microenvironment and likely benefits from reduced trafficking requirements and sustained antigen availability. Relevantly, we later found that even distal Gvax inoculation increases the percentage of CD11b+GR1+ candidate myeloid suppressor cells in B16 tumors relative to FL3vax-treated or untreated mice (Fig. 4D). For malignancies such as melanoma in which tumor sites are accessible to direct injection, these data provide a compelling rationale for clinical evaluation of Fl3vax in conjunction with CTLA-4 blockade.

The differences in local versus distal efficacy between Fl3vax and Gvax suggested that there were distinct mechanisms underlying their immunogenicity, and we hoped that they might synergize when used in combination. In both the B16 and TRAMP tumor models, however, Gvax and FL3vax failed to cooperate when used in combination even in the context of Treg depletion (Figs. 3 and 5). To better understand this observation as well as the underlying differences between Fl3vax and Gvax, we isolated and characterized the lymphocytes infiltrating both the vaccine site and tumor in response to each vaccine.

Compared with vaccines expressing GM-CSF, we find a higher percentage of CD11b−/lo DCs and much higher proportion of pDCs and CD8+ T cells infiltrating the Flt3L site (Fig. 4A). CD4+FoxP3+ Tregs are known to suppress the activity both of CD8+ T cells and DCs (39, 40), and we have reported that higher CD8/Treg ratios within tumors are indicative of effective vaccination (35). Within the Fl3vax vaccination site there are few Tregs, suggesting a favorable environment for immune priming. Compared with either vaccine alone, the combination-vaccine site contains very low levels of T and NK cells, suggesting a possible qualitative defect in the APCs generated by FL3vax + Gvax.

The TIL evoked by FL3vax + α-CTLA-4 contain the highest levels of CD8+ T cells, pDCs, and CD11b−/lo DCs we observed (Fig. 4B and C) and the lowest levels of CD11b+GR1+ cells (Fig. 4D). The tumors of FL3vax recipient mice are substantially infiltrated by Tregs; however, the increase in CD8+ T cells exceeds that of Tregs and preserves a highly advantageous CD8+ T-cell/Treg ratio. The synergy we observed between Treg depletion and Fl3vax combination therapy may be due to the removal of suppression from these large numbers of CD8+ T cells and DCs (Supplementary Fig.S3).

A prior publication reported that B16-GMCSF was more effective preventively compared with B16-Flt3L, and that GM-CSF elicited, based on higher B7-1 expression, a qualitatively and quantitatively superior DC infiltration of the vaccine (38). We found that the levels of Flt3L produced by their vaccine are substantially lower than ours by both ELISA (290 versus 950 ng/1 × 106 cells/24 hours) and flow cytometry (Supplementary Fig. S6), which we believe may explain the disparity between our results. Also the differences in our B16 systems (BL6 versus F10) or in the timing of our prophylactic vaccinations could be relevant. Besides our own observations, a wealth of data now exists, which suggests not only that Flt3L potently generates highly stimulatory DCs (7, 8, 41) but also that those DCs, despite lower expression of B7-1, may be superior for T-cell priming compared with those generated by GM-CSF (42, 43).

We observed that the efficacy of Flt3L seemed to increase with higher doses (Supplementary Fig. S5), whereas it is known that the adjuvant activity of GM-CSF decreases at higher expression levels (32). To increase the bioactivity of our Flt3L transgene, we fused the extracellular domain to a mouse immunoglobulin constant region, creating an enhanced vaccine termed sFL3vaxIg. Compared with Gvax, sFL3vaxIg was able to prevent outgrowth of more 5-day preimplanted B16-BL6 tumors in conjunction with CTLA-4 blockade (Fig. 6). Taken together, these observations suggest that Fl3vax may be a broadly useful vaccine for the treatment of human malignancies especially in conjunction with CTLA-4 blockade.

J.P. Allison: patent license, Medarex; consultant/advisory board, Medarex, Bristol Myers Squibb, and Cell Genesys. M.A. Curran disclosed no potential conflicts of interest.

Note: Supplementary data for this article are available at Cancer Research Online (http://cancerres.aacrjournals.org/).

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.

J.P. Allison is an investigator of the Howard Hughes Medical Institute and holds the David H. Koch Chair in Immunologic Studies at the Memorial Sloan-Kettering Cancer Center.

1
Pardoll D. New strategies for active immunotherapy with genetically engineered tumor cells.
Curr Opin Immunol
1992
;
4
:
619
–23.
2
Dranoff G. GM-CSF-secreting melanoma vaccines.
Oncogene
2003
;
22
:
3188
–92.
3
Huang AY, Bruce AT, Pardoll DM, Levitsky HI. In vivo cross-priming of MHC class I-restricted antigens requires the TAP transporter.
Immunity
1996
;
4
:
349
–55.
4
Shi Y, Liu CH, Roberts AI, et al. Granulocyte-macrophage colony-stimulating factor (GM-CSF) and T-cell responses: what we do and don't know.
Cell Res
2006
;
16
:
126
–33.
5
Shurin MR, Esche C, Lotze MT. FLT3: receptor and ligand. Biology and potential clinical application.
Cytokine Growth Factor Rev
1998
;
9
:
37
–48.
6
Maraskovsky E, Brasel K, Teepe M, et al. Dramatic increase in the numbers of functionally mature dendritic cells in Flt3 ligand-treated mice: multiple dendritic cell subpopulations identified.
J Exp Med
1996
;
184
:
1953
–62.
7
Onai N, Obata-Onai A, Schmid MA, Manz MG. Flt3 in regulation of type I interferon-producing cell and dendritic cell development.
Ann N Y Acad Sci
2007
;
1106
:
253
–61.
8
Dong J, McPherson CM, Stambrook PJ. Flt-3 ligand: a potent dendritic cell stimulator and novel antitumor agent.
Cancer Biol Ther
2002
;
1
:
486
–9.
9
Luster AD, Leder P. IP10, a -C-X-C- chemokine, elicits a potent thymus-dependent antitumor response in vivo.
J Exp Med
1993
;
178
:
1057
–65.
10
Luster AD, Greenberg SM, Leder P. The IP10 chemokine binds to a specific cell surface heparan sulfate site shared with platelet factor 4 and inhibits endothelial cell proliferation.
J Exp Med
1995
;
182
:
219
–31.
11
Hensbergen PJ, Wijnands PG, Schreurs MW, Scheper RJ, Willemze R, Tensen CP. The CXCR3 targeting chemokine CXCL11 has potent antitumor activity in vivo involving attraction of CD8+ Tlymphocytes but not inhibition of angiogenesis.
JImmunother (1997)
2005
;
28
:
343
–51.
12
Piper KP, Horlock C, Curnow SJ, et al. CXCL10-CXCR3 interactions play an important role in the pathogenesis of acute graft-versus-host disease in the skin following allogeneic stem cell transplantation.
Blood
2007
;
110
:
3827
–32.
13
Tominaga M, Iwashita Y, Ohta M, et al. Antitumor effects of the MIG and IP10 genes transferred with poly [d,l-2,4-diaminobutyric acid] on murine neuroblastoma.
Cancer Gene Ther
2007
;
14
:
696
–705.
14
Allavena P, Bianchi G, Zhou D, et al. Induction of natural killer cell migration by monocyte chemotactic protein-1, -2 and -3.
Eur J Immunol
1994
;
24
:
3233
–6.
15
Loetscher P, Seitz M, Clark-Lewis I, Baggiolini M, Moser B. Monocyte chemotactic proteins MCP-1, MCP-2, and MCP-3 are major attractants for human CD4+ and CD8+ T lymphocytes.
FASEB J
1994
;
8
:
1055
–60.
16
Sozzani S, Sallusto F, Luini W, et al. Migration of dendritic cells in response to formyl peptides, C5a, and adistinct set of chemokines.
J Immunol
1995
;
155
:
3292
–5.
17
Van Damme J, Proost P, Lenaerts JP, Opdenakker G. Structural and functional identification of two human, tumor-derived monocyte chemotactic proteins (MCP-2 and MCP-3) belonging to the chemokine family.
J Exp Med
1992
;
176
:
59
–65.
18
Fioretti F, Fradelizi D, Stoppacciaro A, et al. Reduced tumorigenicity and augmented leukocyte infiltration after monocyte chemotactic protein-3 (MCP-3) gene transfer: perivascular accumulation of dendritic cells in peritumoral tissue and neutrophil recruitment within the tumor.
J Immunol
1998
;
161
:
342
–6.
19
New JY, Li B, Koh WP, et al. T cell infiltration and chemokine expression: relevance to the disease localization in murine graft-versus-host disease.
Bone Marrow Transplant
2002
;
29
:
979
–86.
20
Wetzel K, Struyf S, Van Damme J, et al. MCP-3 (CCL7) delivered by parvovirus MVMp reduces tumorigenicity of mouse melanoma cells through activation of Tlymphocytes and NK cells.
Int J Cancer
2007
;
120
:
1364
–71.
21
van Elsas A, Hurwitz AA, Allison JP. Combination immunotherapy of B16 melanoma using anti-cytotoxic T lymphocyte-associated antigen 4 (CTLA-4) and granulocyte/macrophage colony-stimulating factor (GM-CSF)-producing vaccines induces rejection of subcutaneous and metastatic tumors accompanied by autoimmune depigmentation.
J Exp Med
1999
;
190
:
355
–66.
22
Leach DR, Krummel MF, Allison JP. Enhancement of antitumor immunity by CTLA-4 blockade.
Science
1996
;
271
:
1734
–6.
23
Kwon ED, Hurwitz AA, Foster BA, et al. Manipulation of T cell costimulatory and inhibitory signals for immunotherapy of prostate cancer.
Proc Natl Acad Sci U S A
1997
;
94
:
8099
–103.
24
Yang YF, Zou JP, Mu J, et al. Enhanced induction of antitumor T-cell responses by cytotoxic T lymphocyte-associated molecule-4 blockade: the effect is manifested only at the restricted tumor-bearing stages.
Cancer Res
1997
;
57
:
4036
–41.
25
Hurwitz AA, Foster BA, Kwon ED, et al. Combination immunotherapy of primary prostate cancer in a transgenic mouse model using CTLA-4 blockade.
Cancer Res
2000
;
60
:
2444
–8.
26
van Elsas A, Sutmuller RP, Hurwitz AA, et al. Elucidating the autoimmune and antitumor effector mechanisms of a treatment based on cytotoxic Tlymphocyte antigen-4 blockade in combination with a B16 melanoma vaccine: comparison of prophylaxis and therapy.
J Exp Med
2001
;
194
:
481
–9.
27
Peggs KS, Quezada SA, Korman AJ, Allison JP. Principles and use of anti-CTLA4 antibody in human cancer immunotherapy.
Curr Opin Immunol
2006
;
18
:
206
–13.
28
Phan GQ, Yang JC, Sherry RM, et al. Cancer regression and autoimmunity induced by cytotoxic Tlymphocyte-associated antigen 4 blockade in patients with metastatic melanoma.
Proc Natl Acad Sci U S A
2003
;
100
:
8372
–7.
29
Foster BA, Gingrich JR, Kwon ED, Madias C, Greenberg NM. Characterization of prostatic epithelial cell lines derived from transgenic adenocarcinoma of the mouse prostate (TRAMP) model.
Cancer Res
1997
;
57
:
3325
–30.
30
Costa GL, Benson JM, Seroogy CM, Achacoso P, Fathman CG, Nolan GP. Targeting rare populations of murine antigen-specific T lymphocytes by retroviral transduction for potential application in gene therapy for autoimmune disease.
J Immunol
2000
;
164
:
3581
–90.
31
Kinsella TM, Nolan GP. Episomal vectors rapidly and stably produce high-titer recombinant retrovirus.
Hum Gene Ther
1996
;
7
:
1405
–13.
32
Serafini P, Carbley R, Noonan KA, Tan G, Bronte V, Borrello I. High-dose granulocyte-macrophage colony-stimulating factor-producing vaccines impair the immune response through the recruitment of myeloid suppressor cells.
Cancer Res
2004
;
64
:
6337
–43.
33
Jinushi M, Nakazaki Y, Dougan M, Carrasco DR, Mihm M, Dranoff G. MFG-E8-mediated uptake of apoptotic cells by APCs links the pro- and antiinflammatory activities of GM-CSF.
J Clin Invest
2007
;
117
:
1902
–13.
34
Borrello I, Pardoll D. GM-CSF-based cellular vaccines: a review of the clinical experience.
Cytokine Growth Factor Rev
2002
;
13
:
185
–93.
35
Quezada SA, Peggs KS, Curran MA, Allison JP. CTLA4 blockade and GM-CSF combination immunotherapy alters the intratumor balance of effector and regulatory T cells.
J Clin Invest
2006
;
116
:
1935
–45.
36
Berhanu A, Huang J, Alber SM, Watkins SC, Storkus WJ. Combinational FLt3 ligand and granulocyte macrophage colony-stimulating factor treatment promotes enhanced tumor infiltration by dendritic cells and antitumor CD8+ T-cell cross-priming but is ineffective as a therapy.
Cancer Res
2006
;
66
:
4895
–903.
37
Xu Y, Zhan Y, Lew AM, Naik SH, Kershaw MH. Differential development of murine dendritic cells by GM-CSF versus Flt3 ligand has implications for inflammation and trafficking.
J Immunol
2007
;
179
:
7577
–84.
38
Mach N, Gillessen S, Wilson SB, Sheehan C, Mihm M, Dranoff G. Differences in dendritic cells stimulated in vivo by tumors engineered to secrete granulocyte-macrophage colony-stimulating factor or Flt3-ligand.
Cancer Res
2000
;
60
:
3239
–46.
39
Oderup C, Cederbom L, Makowska A, Cilio CM, Ivars F. Cytotoxic T lymphocyte antigen-4-dependent down-modulation of costimulatory molecules on dendritic cells in CD4+ CD25+ regulatory T-cell-mediated suppression.
Immunology
2006
;
118
:
240
–9.
40
von Boehmer H. Mechanisms of suppression by suppressor T cells.
Nat Immunol
2005
;
6
:
338
–44.
41
Gilliet M, Boonstra A, Paturel C, et al. The development of murine plasmacytoid dendritic cell precursors is differentially regulated by FLT3-ligand and granulocyte/macrophage colony-stimulating factor.
J Exp Med
2002
;
195
:
953
–8.
42
Taieb A, Breitinger JJ, Unadkat JV, et al. Intrinsic ability of GM + IL-4 but not Flt3L-induced rat dendritic cells to promote allogeneic T cell hyporesponsiveness.
Clin Immunol
2007
;
123
:
176
–89.
43
Daro E, Butz E, Smith J, Teepe M, Maliszewski CR, McKenna HJ. Comparison of the functional properties of murine dendritic cells generated in vivo with Flt3 ligand, GM-CSF and Flt3 ligand plus GM-SCF.
Cytokine
2002
;
17
:
119
–30.

Supplementary data