Cutting Edge: Dendritic Cells Require Maturation via CD40 to Generate Protective Antitumor Immunity¹

The interaction between CD40 and its ligand, CD154 (glycoprotein 39), has been recognized as an important element in the development of cell-mediated immunity (CMI)³ (1). Studies have shown that CD40/CD154 interactions are critical for the priming and expansion of CD4⁺ Th cells and CD8⁺ cytotoxic T cells in response to protein Ag, adenovirus, and alloantigen (reviewed in 2 . It is not clear why this receptor/ligand pair is critical in triggering CMI; however, the involvement of CD40/CD154 interactions in the regulation of costimulatory molecule expression and cytokine production by APCs has been hypothesized (1). The observations that the triggering of CD40 induces an increase in the expression of both CD80 (B7-1) and CD86 (B7-2) on the surface of B cells (3), dendritic cells (DCs) (4), and macrophages (5) are in support of this hypothesis. Furthermore, both DCs and macrophages secrete IL-12 as well as a spectrum of other inflammatory cytokines and chemokines following CD40 ligation (4, 5, 6, 7, 8). IL-12 is of particular interest, in that this cytokine has been shown to be critical for the differentiation of Th cells to the Th1 (IFN-γ-producing) phenotype following cognate interaction with Ag-pulsed APCs (9).

Although the function of CD40 as a maturational signal for B cells to differentiate to competent APCs has already been established (10), the role of CD40 in DC maturation and function in vivo is unresolved. The question remains as to whether such “professional” APCs must be matured by a CD40 signal to generate functional cellular responses. In vitro studies have shown that the triggering of CD40 on DCs is superior at inducing IL-12 secretion and at enhancing the Ag presentation capabilities of DCs in comparison with other inflammatory mediators such as bacterial LPS or TNF-α (6). These data underscore the potential importance of CD40 as a central component for the maturation and APC function of DCs in vivo.

We have recently shown that the generation of protective tumor immunity following the administration of normally effective vaccine regimens is dependent upon CD40/CD154 interactions (11). Herein, we demonstrate that DCs fail to generate protective tumor immunity if they do not express CD40; we also provide evidence to suggest that CD40-mediated IL-12 production by DCs and/or macrophages is critical for the generation of effectors that are capable of Th1-type cytokine production following tumor vaccination.

We obtained 6- to 8-wk-old female BALB/c (H-2^d) and C57BL/6 (H-2^b) mice from the National Cancer Institute (Bethesda, MD). CD40- (H-2^d) and CD154-deficient mice (H-2^b) were bred at the Dartmouth-Hitchcock Medical Center (DHMC) and have been described previously (8, 12). All mice were maintained in a specific pathogen-free animal facility at DHMC, and all animal studies were preapproved by the DHMC Institutional Animal Care and Use Committee.

The MB49 bladder carcinoma cell line (13), the sarcomas MCA-105 and MCA-205, and the adenocarcinoma TS/A were maintained as described previously (11). The MB49 cell line was doubly infected with two murine stem cell virus retroviruses (14) that were separately carrying the cDNAs for the p35 and p40 subunits of murine IL-12 and were selected for stable expression using G418 and puromycin. It was determined that the MB/IL-12 cell line secreted 2700 pg/ml/2 × 10⁵ cells/36 h of bioactive IL-12 as assessed by ELISA, whereas the amount of IL-12 produced by both parental MB49 and MB/neo (infected with empty virus) was below the limit of detection (data not shown). The parental MB49, MB/IL-12, and MB/neo lines were kindly provided by Dr. John Leonard (Genetics Institute, Andover, MA).

The following Abs were used: M5-biotin (αMHCII; American Type Culture Collection (ATCC), Manassas, VA), N418-biotin/FITC (αCD11c; a gift of Ralph Steinman, Rockefeller University, New York, NY), GL1-biotin (αB7-2; ATCC), 1C10-biotin (αCD40; a gift of Maureen Howard, DNAX, Palo Alto, CA), rat Ig-biotin/FITC (Harlan, Indianapolis, IN), 6B2-biotin (αCD45R/B220; ATCC), αH-2D^d-biotin, and αB7-1-biotin (PharMingen, San Diego, CA). Abs that had been conjugated to biotin were detected using either streptavidin-phycoerythrin or streptavidin-FITC (Southern Biotechnologies, Birmingham, AL). Cells were analyzed by flow cytometry according to standard protocols with a FACScan (Becton Dickinson, Mountain View, CA).

Splenic DCs were prepared as described previously (15). Flt3 ligand (Flt3L)-splenic DCs were generated in vivo as described previously (16) and then isolated from the spleen by the depletion of T cells using anti-Thy1.2 (HO-13-4; ATCC), anti-CD4 (RL172/4; ATCC), and guinea pig complement (Cedarlane, Ontario, Canada); B cells were subsequently removed by panning twice on petri dishes that had been coated with goat anti-mouse Ig (Zymed, San Francisco, CA). Normal and Flt3L-generated DCs were harvested, washed, and analyzed by flow cytometry for purity. Both BALB/c and CD40-deficient DCs expressed N418/CD11c as well as high levels of CD80, CD86, and MHC class II, whereas the expression of CD40 was restricted to the cells derived from BALB/c mice. Preparations were typically between 70 and 80% positive for N418/CD11c (data not shown).

Vaccines consisted of varying numbers of normal or Flt3L-generated splenic DCs that were admixed with irradiated TS/A tumor cells (3,500 or 10,000 rad) and were followed by a live TS/A challenge as indicated in the figure legends. The adjuvant-enhanced MCA-105 vaccine was administered as described previously (11). The mock- and IL-12-transduced MB49 tumor vaccines were administered as indicated in the figure legends. All tumor diameters were measured starting at 1 wk after live tumor injection using a caliper. Tumors were measured every 3 days, and mice whose tumors reached a diameter of >2 cm were sacrificed as per Institutional Animal Care and Use Committee guidelines. For protection studies, Fisher’s exact test was used to compare the proportions of surviving mice, and tumor growth was measured for ≥70 days in all experiments.

Supernatants from draining lymph node (DLN) cells were generated as indicated in the figure legends; the supernatants were subsequently assayed for cytokine content by ELISA using unconjugated rat anti-mouse IL-2, IL-4, and IFN-γ and hamster/rat anti-mouse IL-12 (PharMingen) as the capture Abs, followed by detection with biotinylated rat anti-mouse IL-2, IL-4, IFN-γ, and IL-12 (PharMingen) as per the manufacturer’s recommendations. The ELISAs were developed using streptavidin-horseradish peroxidase (Amersham, Arlington Heights, IL) and tetramethylbenzidine substrate (Sigma, St. Louis, MO).

The production of Th1-type inflammatory cytokines (IL-2 and IFN-γ) clearly aids in the effective generation of tumor-specific CD8⁺ CTLs (17, 18), and it is has been demonstrated that Th1-type responses afford optimal tumor protection (19). T cells are the therapeutic effectors that are generated by immunization with the sarcoma MCA-105 in conjunction with the bacterial adjuvant Corynebacterium parvum, and both CD4⁺ and CD8⁺ T cells are required for the generation of protective immunity following vaccination with MCA-105 tumor (17). We postulated that the requirement for CD40/CD154 interactions in the induction of systemic immunity following MCA-105 vaccination (11) is a reflection of the fact that a protective Th1-type response is not generated in the absence of this receptor/ligand pair. To test this hypothesis, mice that had been treated with anti-CD154 or control hamster Ig (HIg) were vaccinated with MCA-105 admixed with C. parvum, and DLN cells were assessed for cytokine production following in vitro restimulation as described in the figure legends. The production of both IFN-γ (Fig. 1,A) and IL-2 (Fig. 1 B) by DLN cells was completely blocked by the administration of anti-CD154, indicating that the priming of tumor-specific CD4⁺ and/or CD8⁺ T cells for Th1-type cytokine production is impaired in the absence of CD40/CD154 interactions. The amount of IL-4 production in these experiments was below detectable levels for all treatment groups (data not shown), suggesting that a blockade of this receptor/ligand pair did not result in an overt polarization from a Th1-type (IFN-γ-producing) to a Th2-type (IL-4-producing) response as has been described by others (8, 12, 20).

These data are consistent with other experimental systems in which both CD4⁺ and CD8⁺ effector T cell generation appears to be significantly impaired in the absence of CD40/CD154 interactions (2). The failure to generate cellular immunity in our tumor model system may be a reflection of the fact that host APCs that are deprived of CD40 maturation signals do not properly activate naive T cells because of a deficient expression of costimulatory molecules (3, 4, 5, 10), a deficient production of inflammatory cytokines and chemokines (4, 5), and/or the failure of APCs to accumulate long-lived MHC class II-peptide complexes capable of stimulating naive T cells over a significant period of time (21, 22).

A defective priming of T cells to produce Th1-type cytokines as the result of a CD40/CD154 blockade (Fig. 1) may suggest that APCs require a CD40 signal to effectively trigger T cell activation and differentiation in response to tumor Ag. As CD40-deficient mice exhibit impaired antitumor responses (11), we next assessed whether the injection of splenic CD40^+/+ DCs with tumor Ag would reconstitute protective tumor immunity in these animals. As indicated in Figure 2 A, CD40-deficient mice that had been immunized with tumor cells alone succumbed to a live tumor challenge with kinetics that were similar to those observed with unimmunized BALB/c mice. On the other hand, 50% of CD40-deficient mice that had been immunized with tumor cells and CD40-bearing DCs were protected from tumor challenge. This experiment is representative of three experiments that were performed. In total, 10 of 15 (67%) CD40-deficient mice receiving the DC plus tumor cell vaccine were protected from tumor challenge compared with only 2 of 18 (11%) mice that had been immunized with TS/A tumor cells alone (p = 0.001).

To directly test whether isolated splenic DCs require activation via CD40 to restore protective immunity, CD40-deficient mice were immunized with irradiated TS/A cells with or without splenic DCs that had been isolated from CD40^+/+ or CD40^−/− donors. Mice were then challenged with live parental TS/A and followed for tumor growth and survival as a measure of the induction of protective tumor immunity. Whereas CD40-deficient mice receiving irradiated tumor cells alone were only minimally protected (one of six mice, Fig. 2 B, group 3), those mice receiving CD40^+/+ DCs with tumor cells were completely protected from live tumor challenge (five of five mice, group 4, p = 0.01). Of note, CD40-deficient mice that were injected with the same dose of TS/A and the same number of CD40^−/− DCs were poorly protected (one of six mice, group 5); this finding was identical with that seen for the CD40-deficient mice receiving irradiated tumor cells alone (group 3). In light of the fact that only 50% of CD40^+/+ BALB/c mice that had been immunized with irradiated TS/A cells alone typically exhibited long-term protection in our study (two of six mice (33%) in this case, group 2), the ability of the CD40^+/+ DCs admixed with irradiated TS/A cells to induce complete protection in CD40-deficient mice (group 4) illustrates the efficiency with which these professional APCs are able to prime naive T cells. Earlier studies involving the use of DC vaccines have suggested that the activation of T cells against FCS components within the media such as BSA leads to a significant degree of nonspecific lytic activity following the administration of these APCs in vivo (23). Our observation that CD40^−/− DCs do not generate protective tumor immunity even in the presence of serum suggests that CD40 is critical for the APC function of DCs, regardless of whether there is some contribution of FCS-specific T cells to the clearance of live tumors in our model system.

Similar reconstitution experiments were performed in CD40-deficient mice using CD40^+/+ or CD40^−/− DCs that were derived from mice treated with Flt3L; Flt3L is a recently identified hematopoietic stem cell growth factor that induces a 17-fold increase in splenic N418⁺ MHC class II⁺ DCs, which are functionally equivalent to conventional splenic DCs (16). BALB/c and CD40-deficient mice were vaccinated with irradiated TS/A cells with or without CD40^+/+ or CD40^−/− Flt3L DCs and then challenged with live parental TS/A cells as indicated in the figure legends. As shown in Figure 2 C, CD40-deficient mice that were vaccinated with TS/A alone exhibited only 13% long-term survival (one of eight mice, group 3), whereas CD40^−/− mice that were immunized with tumor cells and CD40^+/+ Flt3L DCs were almost completely protected (seven of eight mice, group 4, p = 0.01). However, vaccination with equal numbers of tumor cells and CD40^−/− Flt3L DCs afforded the CD40-deficient mice no protection (zero of seven mice, group 5). This experiment was repeated with analogous results. These data indicate that Flt3L-generated DCs, like conventional splenic DCs, require activation via CD40 for their maturation into efficient APCs that are capable of generating protective tumor immunity.

To our knowledge, these data are the first demonstration of the absolute requirement for CD40 expression on DCs for their function as mature APCs in vivo. Consistent with the hypothesis that CD40 plays a critical role in DC function, CD40-deficient mice are incapable of generating protective cell-mediated tumor immunity (11). The immune deficiency in these mice appears to be the result of a lesion in APC function, in that the coadministration of CD40-bearing DCs with irradiated tumor cells restores the ability of these mice to clear a live parental tumor challenge (Fig. 2, A–C). The failure of CD40^−/− DCs to reconstitute tumor protection in this manner (Fig. 2, B and C) suggests a crucial role for this receptor in the maturation of both conventional and Flt3L-generated splenic DCs into efficient APCs. Interestingly, a number of laboratories have demonstrated that Ag-pulsed DCs fail to generate protective immune responses in vivo in the absence of CD4⁺ T cells (24, 25). Our data suggest that the requirement for Th cells in these other studies is a reflection of the fact that DCs require the CD40 maturation signals induced during cognate interaction with CD154-bearing CD4⁺ T cells to become efficient APCs.

The failure of CD40^−/− DCs to restore protective immunity in CD40-deficient mice (Fig. 2, B and C) may be due to insufficient costimulatory capacity and/or to their inability to produce inflammatory cytokines, both of which are important for the generation of CMI. The cytokine IL-12 has been shown to play a pivotal role in the differentiation of naive CD4⁺ T cells toward the Th1 phenotype (9), and recent studies suggest that IL-12 secretion by DCs following cognate interaction with Th cells is largely CD40/CD154-dependent (7). As it is generally accepted that inflammatory Th1-type responses are most beneficial in the context of tumor protection (18, 19), it is reasonable to postulate that CD40-mediated IL-12 production may be critical for the generation of tumor protection, and that a source of exogenous IL-12 may allow for the induction of systemic tumor immunity in the absence of CD40/CD154 interactions. To test this hypothesis, both C57BL/6 mice that had been treated with either anti-CD154 or control HIg and CD154-deficient mice were vaccinated with the syngeneic bladder carcinoma line MB49; this line was either mock-transduced or transduced with the genes encoding the p35 and p40 chains of IL-12, leading to the secretion of bioactive IL-12 from the tumor (see Materials and Methods). Mice were vaccinated, treated with Ab, and challenged with live parental MB49 cells as described in Table I. Vaccination with the mock-transduced MB49 line (MB/neo) almost completely protected all of the C57BL/6 mice that had been treated with control HIg (96% protection), indicating that the MB49 line itself was highly immunogenic. However, the ability of this line to induce systemic tumor immunity is dependent upon CD40/CD154 interactions, in that the immunization of CD154-deficient or anti-CD154-treated mice with the same MB/neo vaccine dose failed to generate a protective immune response in these groups (only 3 of 18 mice protected (17%) and 0 of 10 mice protected (0%), respectively). However, the secretion of IL-12 by the tumor vaccine appears to partially bypass the requirement for CD40/CD154 interactions, in that CD154-deficient and anti-CD154-treated mice that had been vaccinated with the IL-12-transduced cell line (MB/IL-12) were able to generate significant protective responses vs those animals receiving the MB/neo vaccine (58% vs 0% for anti-CD154-treated C57BL/6 mice, p = 0.001; 67% vs 17% for CD154-deficient mice, p = 0.01).

These data suggest that the failure of anti-CD154-treated mice to generate protective tumor immunity (11) is a result of their failure to induce a strong Th1-type T cell response following tumor vaccination and is largely due to the lack of IL-12 production by DCs and/or macrophages following CD40 ligation (6, 7, 8). However, because IL-12 has been shown to have profound effects on NK cell activation (26), the role of this cell type in the induction of tumor immunity by the MB/IL-12 vaccine remains a formal possibility. The restoration of tumor protection by IL-12 is particularly impressive in light of our earlier observations that tumor vaccines that are engineered to secrete granulocyte-macrophage CSF or are admixed with bacterial adjuvant fail to generate systemic immunity in the presence of anti-CD154 (11). The data generated with the IL-12-transduced tumor vaccine also suggest that host APCs that are deprived of CD40 maturation signals retain some capacity to activate naive T cells but are inefficient at inducing the effectors that are capable of producing inflammatory Th1-type cytokines. Our observations are consistent with studies characterizing the effect of a CD40/CD154 blockade on the onset of inflammatory bowel disease (20) and on the generation of protective cellular responses following Leishmania infection (8, 12); in these studies, exogenous IL-12 was shown to restore Ag-specific Th1 responses in the absence of the CD40/CD154 receptor/ligand pair (8, 20).

Taken together, the data presented in this study suggest that the inability to generate protective tumor immunity in the absence of CD40/CD154 interactions is due to the lack of a strong Th1-type response following vaccination and results from the failure of DCs and/or macrophages to produce IL-12 following CD40 ligation.