Augmented IL-15Rα Expression by CD40 Activation Is Critical in Synergistic CD8 T Cell-Mediated Antitumor Activity of Anti-CD40 Antibody with IL-15 in TRAMP-C2 Tumors in Mice

Interleukin-2 has been approved by the U.S. Food and Drug Administration (FDA) for use in the treatment of patients with metastatic malignancy (1, 2). However, IL-2 is not optimal as an agent to inhibit tumor growth because it is associated with capillary leak syndrome and because it induces a series of checkpoints that inhibit immune responses that include its role in activation-induced cell death (AICD) and maintenance of IL-2–dependent regulatory T cells. In contrast, IL-15, with its ability to activate CD8 T cells and NK cells, its inhibition of AICD, and its role in the persistence of CD8 memory T cells, might be a better choice. Attempts to prevent tumor growth and to treat various mouse tumor models with IL-15 proved effective (3–8). In contrast to IL-2, IL-15 is a membrane-associated molecule that induces signaling at an immunological synapse between APCs and target NK or CD8 T cells (9). IL-15Rα on surfaces of monocytes and dendritic cells (DCs) presents IL-15 in trans to cells that express the other two IL-15 receptors, IL-2/IL-15Rβ (CD122) and common γ-chain (γc) alone (9). Mice with genetic deletions of IL-15 or its private receptor, IL-15Rα, are characterized by decreased numbers of NK, NKT, CD8⁺/CD44^high T, TCRγ⁺/δ⁺ T, and intestinal intraepithelial CD8α⁺/β⁻ T cells, suggesting that physiologically relevant IL-15 signals require IL-15Rα expression (10, 11). The FDA and National Cancer Institute (NCI) have approved a phase I dose-escalation trial of recombinant human IL-15 in patients with metastatic melanoma and renal cell cancer, and this trial has been initiated.

Although IL-15 administration may ultimately show efficacy in the treatment of metastatic malignancy, it may not be optimal when used in monotherapy. As noted for IL-15 long-term persistence and its optimal transpresentation to NK and CD8 T cells, it must be bound to IL-15Rα on the surface of DCs and monocytes (9–13). However, there is only modest expression of IL-15Rα on unactivated DCs; therefore, the level of IL-15Rα expression would be limiting in therapeutic trials that use IL-15 alone. In the current study, we examined the use of an agonistic anti-CD40 Ab to induce IL-15Rα expression that could be given to circumvent this problem.

CD40 is a member of the TNFR superfamily that plays a critical role in both cellular and humoral immune responses (14). CD40 ligation triggers a series of cellular functions, including activation of APCs (15, 16). Agonistic anti-CD40 Abs were shown to promote T cell-mediated immunity and effective treatment of neoplastic diseases in animal models (17, 18). It was reported that a stimulatory CD40 Ab indirectly activated NK cells by inducing IL-12 secretion by DCs, which resulted in significant antitumor effects (19). In another potential contribution to effective therapy of tumors, we demonstrated that activation of CD40 was associated with an increased expression of IL-15Rα by DCs and an enhanced antitumor efficacy in murine models (6). This induction of IL-15Rα expression by CD40 activation provided the scientific basis for this study in which we demonstrated therapeutic efficacy mediated by the combination regimen of murine IL-15 (mIL-15) and an agonistic anti-CD40 Ab in the TRAMP-C2 model of prostatic cancer. Furthermore, we used IL-15Rα^−/− mice and DCs from such mice to demonstrate that induction of IL-15Rα expression by an agonistic anti-CD40 Ab contributes to the synergistic antitumor effect observed with the combination therapy. The findings from this study suggest that the combination regimen is very promising for the treatment of patients with cancer.

mIL-15 was purchased from PeproTech (Rocky Hill, NJ). Rat anti-mouse CD40 (FGK4.5) and rat anti-mouse CD8 (clone 2.43) mAbs were obtained from Bio X Cell (West Lebanon, NH). Rat anti–asialo-GM1 was purchased from Wako Chemicals (Richmond, VA).

The TRAMP-C2 cell line was derived from a prostate tumor of a male TRAMP mouse. MC38 is a metastatic murine colon cancer cell line syngeneic to C57BL/6 mice. IL-15Rα^−/− mice (Jackson Laboratory, Bar Harbor, ME) were backcrossed to the C57BL/6 strain for 10 additional generations. The tumor model was established by s.c. injection of 5 × 10⁵ TRAMP-C2 cells into the right flank of male C57BL/6 (NCI-Frederick) or IL-15Rα^−/− mice. All animal experiments were approved by the NCI Animal Care and Use Committee and were performed in accordance with NCI Animal Care and Use Committee guidelines.

The therapeutic studies were performed in both C57BL/6 wild-type and IL-15Rα^−/− mice bearing TRAMP-C2 tumors. Therapy started when the tumor volumes reached 40–110 mm³. Groups of TRAMP-C2 tumor-bearing mice received mIL-15 i.p., 2.5 μg/mouse every day, 5 d a week for 2 wk; the anti-CD40 Ab, 200 μg on day 0 and then 100 μg on days 3, 7, and 10; or a combination of mIL-15 with anti-CD40 Ab at the same doses and dosing schedules as those in the mIL-15 and the anti-CD40 Ab groups. An additional group of TRAMP-C2 tumor-bearing mice that received PBS injections served as a control. Throughout the experiments, survival of mice was recorded, and TRAMP-C2 tumor growth was monitored by measuring tumor size in two orthogonal dimensions. The tumor volume was calculated using the equation: 1/2(long dimension)(short dimension)².

TRAMP-C2 cells were seeded into a 24-well plate at the concentration of 5 × 10⁴/ml/well and cultured overnight. Then, luciferase-GFP fusion protein expressing lentivirus (pol2-ffluc-eGFP; Advanced Technology Program, NCI-Frederick) was added to the culture for 2 d. The GFP⁺ cells were collected by sorting. Male C57BL/6 wild-type mice were injected with luciferase-GFP–transduced TRAMP-C2 (TRAMP-C2/luc-GFP) cells s.c., and the therapeutic study was performed using the same therapeutic protocol as above. The bioluminescence images were obtained 1 d before and 16 and 40 d after initiation of therapy with an in vivo-imaging system, and the images were analyzed using Living Imaging software (both from Xenogen, Alameda, CA).

The TRAMP-C2 tumor-bearing wild-type mice received combination therapy of mIL-15 with the anti-CD40 Ab at the same doses and dosing schedule as above. NK and/or CD8⁺ T cells were depleted in vivo by i.p. injections of anti–asialo-GM1 (50 μl) or purified rat anti-mouse CD8 Ab (200 μg). One dose of anti–asialo-GM1 and anti-mouse CD8 Ab was administered 1 d before initiation of combination therapy; subsequent doses were administered three times weekly for 2 wk. Depletion of NK and CD8⁺ cells in PBMCs was evaluated by flow cytometry during the experiment. Survival of mice was recorded, and TRAMP-C2 tumor growth was monitored throughout the experiments.

TRAMP-C2 tumor-bearing wild-type mice were treated using the same therapeutic protocol as above. At day 12 after initiation of the treatment, all of the mice (four mice/group) were sacrificed, and the spleens were obtained. The proportion and absolute numbers of NK1.1⁺, CD8⁺, CD44^highCD8⁺, and TRAMP-C2 tumor-specific SPAS-1/SNC9-H₈ (20) tetramer⁺CD8⁺ cells in the spleens were analyzed by flow cytometry.

Surviving tumor-free mice from the group receiving the combination regimen of mIL-15 and anti-CD40 Ab were rechallenged with 5 × 10⁵ TRAMP-C2 tumor cells or with 1 × 10⁵ MC38 tumor cells, an irrelevant tumor cell line, at 40 d or 3.5 mo after initiation of therapy to assess the effect on tumor engraftment. One group of the surviving tumor-free mice received i.p. injections of purified rat anti-mouse CD8 Ab (200 μg/injection) three times weekly for 2 wk, and one dose of anti-CD8 Ab was administered 1 d before the rechallenge. Another group of mice received combination therapy initially but did not receive the TRAMP-C2 tumor cells before the rechallenge. The mice in the control groups did not receive either the initial TRAMP-C2 challenge or any therapeutic reagents prior to “rechallenge.”

Surface staining to quantitate the number of NK1.1, CD8, CD44, and SPAS-1/SNC9-H₈ tetramer-expressing cells in PBMCs or splenocytes was performed using commercial FITC-, PE-, or allophycocyanin-conjugated Abs (eBioscience), with the exception of the PE-conjugated SPAS-1/SNC9-H₈ tetramer, which was provided by Dr. J.P. Allison. IL-15Rα expression on cell surfaces was detected using biotinylated anti–mIL-15Rα Ab or biotinylated normal goat IgG as an isotype control (R&D Systems), followed by PE-labeled streptavidin (eBioscience). CD40 expression on TRAMP-C2 cell surfaces was analyzed using the primary Ab FGK4.5 or normal rat IgG as an isotype control, followed by PE-labeled mouse anti-rat Ig (BD Biosciences, San Jose, CA). The samples were collected on a FACScan flow cytometer (BD Biosciences) and analyzed using FlowJo cytometry analysis software (TreeStar, Ashland, OR).

NK cells were purified from RAG1^−/− mice using negative-isolation MicroBeads (Miltenyi Biotec). Bone marrow-derived DCs (BMDCs) from wild-type or IL-15Rα^−/− mice were generated using 10 ng/ml IL-4 and 20 ng/ml GM-CSF (both from PeproTech). The BMDCs were pretreated with LPS (1 μg/ml; Sigma, St. Louis, MO), anti-CD40 Ab (10 μg/ml), or medium alone for ∼20 h. Then BMDCs were washed and cocultured with NK cells at a 2:1 ratio for an additional 20 h in the presence or absence of LPS (1 μg/ml), anti-CD40 Ab (10 μg/ml), mIL-15 (2.5 ng/ml), or anti-CD40 Ab plus mIL-15. ⁵¹Cr-labeled TRAMP-C2 cells were incubated with NK cells and BMDCs for 5 h at various E:T ratios. Radioactivity in the liquid phase was measured in a gamma counter (PerkinElmer). The percentage of specific lysis was determined using the following formula: percentage lysis = 100 × [(experiment cpm − spontaneous cpm)/(maximum cpm − spontaneous cpm)]. The maximum release value was determined from target cells treated with 1% (v/v) Triton X-100 (Sigma).

Comparison of cell numbers in spleens and luminescent signals of images among different treatment groups was analyzed using the Student t test. The statistical significance of differences in the survival of mice in different groups was determined using the log-rank test and the GraphPad Prism program (GraphPad, San Diego, CA).

We investigated the therapeutic efficacy of the combination regimen of mIL-15 with an agonistic anti-CD40 Ab in an established TRAMP-C2 tumor model. The rationale for this approach is that expression of IL-15Rα was induced by CD40 activation (6, 21). Therapy started when TRAMP-C2 tumors were well established (average tumor volume = 80 mm³). Treatment with mIL-15 alone at a dose of 2.5 μg/mouse, 5 d a week for 2 wk, provided a modest inhibition of tumor growth (Fig. 1A) and prolonged survival of the TRAMP-C2 tumor-bearing mice compared with mice in the PBS control group (Fig. 1B, p < 0.05). Treatment with anti-CD40 Ab at a dose of 200 μg/mouse on day 0, followed by 100 μg/mouse on days 3, 7, and 10, significantly inhibited tumor growth (Fig. 1A) and prolonged survival of TRAMP-C2 tumor-bearing mice compared with mice in either the PBS control group or the mIL-15–alone group (Fig. 1B, p < 0.001). Critically, combination therapy with mIL-15 and the anti-CD40 Ab provided greater therapeutic efficacy than did monotherapy with either mIL-15 or anti-CD40 Ab (Fig. 1A, 1B, p < 0.001). The therapeutic study was repeated, and comparable results were obtained. All mice in the PBS control and mIL-15–alone groups died of tumor progression by day 40, and 70–100% of mice in the anti-CD40 Ab-alone group died of tumor progression within 2 mo (Fig. 1B). In contrast, combination treatment resulted in a highly significant prolongation of survival, with 70–100% of mice in the combination group becoming and remaining tumor free (Fig. 1B). Flow cytometry analysis showed that TRAMP-C2 cells do not express CD40 on their cell surfaces (data not shown).

Groups of six wild-type mice bearing TRAMP-C2/luc-GFP tumors were treated using the same therapeutic protocol as above. As shown in Fig. 1C, the average tumor volumes among groups were comparable at the beginning of therapy. Treatment with anti-CD40 Ab significantly inhibited tumor growth, with average total luminescent signals of 1.0 × 10⁹ photons/s on day 16, which were much lower than 1.5 × 10¹⁰ and 7.7 × 10⁹ photons/s in the PBS control and mIL-15 groups, respectively, on the same day (p < 0.01). However, only one of six mice in the anti-CD40 group was tumor free. In contrast, combination treatment led to regression of tumors, with all six mice becoming and remaining tumor free.

IL-15 and its private receptor, IL-15Rα, are both essential for the support of NK and CD8⁺ T cell homeostasis. Expression of IL-15Rα specifically on DCs is critical for the trans-presentation of IL-15 and activation of NK cells. A series of studies was performed to determine whether increased expression of IL-15Rα, mediated by administration of agonistic anti-CD40 Ab, played a role in the augmented IL-15–mediated therapeutic efficacy observed in the combination-treatment group. Previously, we reported that treatment with anti-CD40 Ab augmented expression of IL-15Rα on the CD11c⁺ population of splenocytes and increased serum concentrations of IL-15Rα (6). In the current study, we performed a therapeutic trial in the TRAMP-C2 tumor model using IL-15Rα^−/− mice. We used the same doses and dosing schedules for mIL-15 and anti-CD40 Ab as in the therapeutic studies in TRAMP-C2 tumor-bearing wild-type mice (Fig. 1). Treatment with mIL-15 showed very little therapeutic efficacy in IL-15Rα^−/− mice, as reflected by the very modest inhibition of tumor growth (Fig. 2A) and the prolongation of survival of the TRAMP-C2 tumor-bearing mice compared with those in the control group (Fig. 2B, p < 0.001). Treatment with anti-CD40 Ab inhibited tumor growth and prolonged the survival of TRAMP-C2 tumor-bearing IL-15Rα^−/− mice compared with mice in either the PBS control or mIL-15–alone group (Fig. 2, p < 0.001). However, treatment with anti-CD40 Ab at the same dose and dosing schedule showed much less therapeutic efficacy in IL-15Rα^−/− mice (Fig. 2) compared with that observed in wild-type mice (Fig. 1). Although the combination regimen in IL-15Rα^−/− mice provided greater therapeutic efficacy compared with monotherapy consisting of either mIL-15 or anti-CD40 Ab (Fig. 2, p < 0.001), only 10–20% of the mice in the combination group became tumor free, in contrast with 70–100% of the wild-type mice receiving combination therapy (Fig. 1). The therapeutic study in IL-15Rα^−/− TRAMP-C2 tumor-bearing mice was repeated, and comparable results were obtained. IL-15Rα^−/− mice had decreased numbers of NK and CD44^highCD8⁺ T cells (10, 11); therefore, the reduced therapeutic efficacy of combination therapy in IL-15Rα^−/− mice might be due to the lower numbers of cytolytic cells.

To determine whether the killing activity of cytolytic cells from IL-15Rα^−/− mice was lower than that of cells from wild-type mice, the lysis activity of NK cells toward TRAMP-C2 cells was examined ex vivo (Supplemental Fig. 1). Compared with NK cells isolated from TRAMP-C2 tumor-bearing wild-type mice that received PBS injections as a control, NK cells isolated from both TRAMP-C2 tumor-bearing wild-type and IL-15Rα^−/− mice that received the combination therapy for 5 d showed increased lysis activity toward TRAMP-C2 tumor cells (Supplemental Fig. 1). However, NK cells isolated from wild-type mice that received the combination regimen showed greater lysis activity against tumor cells than did NK cells from IL-15Rα^−/− mice (Supplemental Fig. 1). This suggested that both decreased numbers of NK and CD44^highCD8⁺ T cells and lower lysis activity per cytolytic cell contributed to the reduced antitumor activity obtained using the combination regimen in IL-15Rα^−/− mice bearing TRAMP-C2 tumor (Fig. 2) compared with that in wild-type mice (Fig. 1).

To investigate possible involvement of NK or CD8⁺ T cells as effectors in the combination regimen-mediated antitumor activity, we treated TRAMP-C2 tumor-bearing mice with anti–asialo-GM1 or purified rat anti-mouse CD8 Ab to eliminate NK or CD8⁺ T cells, together with the combination regimen. One dose of the anti–asialo-GM1 and anti-CD8 Ab was administered 1 d before the initiation of combination therapy, and subsequent doses were administered three times weekly for 2 wk. FACS analysis showed that >90% of NK cells (Supplemental Fig. 2A) and CD8⁺ cells (Fig. 3A) from PBMCs were depleted by administration of anti–asialo-GM1 and anti-CD8 Ab, respectively. Compared with the group receiving the combination therapy alone, depletion of CD8⁺ cells significantly, but not completely, eliminated antitumor efficacy (Fig. 3B, p < 0.001). Administration of anti–asialo-GM1 decreased antitumor efficacy (Fig. 3B, p < 0.001). Simultaneous depletion of CD8⁺ and NK cells abrogated antitumor efficacy mediated by the combination therapy (Supplemental Fig. 2B). It should be noted that administration of anti–asialo-GM1 depleted NK cells (Supplemental Fig. 2A), as well as a subset of CD8⁺ cells, especially CD44^highCD8⁺ and TRAMP-C2 tumor Ag-specific SPAS-1/SNC9-H₈ tetramer⁺CD8⁺ cells (Fig. 3A). TRAMP-C2 cells isolated from in vivo tumors expressed high levels of MHC class I (MHCI) (Supplemental Fig. 2C). TRAMP-C2 cells cultured in vitro with IFN-γ increased the expression of MHCI on their cell surfaces (Supplemental Fig. 2C), which was associated with increased cytolytic activity of CD8⁺ T cells toward TRAMP-C2 cells ex vivo (Supplemental Fig. 2D). These studies support the view that CD8⁺ T cells played a major role in combination regimen-mediated antitumor efficacy, whereas NK cells might also contribute to the therapeutic efficacy observed.

To further elucidate the effectors underlying the synergistic efficacy of mIL-15 plus anti-CD40 Ab, we defined the impact of combination therapy on numbers of splenic tumor Ag-specific CD8⁺ T cells using the TRAMP-C2 tumor-specific SPAS-1/SNC9-H₈ tetramer. TRAMP-C2 tumor-bearing mice were treated with the same therapeutic protocol used in the therapeutic studies shown in Fig. 1. At day 12 after initiation of therapy, all mice (four mice/group) were sacrificed. The percentages of CD8⁺, CD44^highCD8⁺, and TRAMP-C2 tumor Ag-specific SPAS-1/SNC9-H₈ tetramer⁺CD8⁺ cells in spleens were analyzed by flow cytometry, and absolute cell numbers were calculated. Treatment with mIL-15, anti-CD40 Ab, or their combination increased total cell numbers and absolute numbers of CD8⁺ and CD44^highCD8⁺ cells in the spleens compared with those in the control group (Fig. 4A–C, p < 0.05). Most critically, the combination regimen dramatically increased the absolute number of TRAMP-C2 tumor Ag-specific SPAS-1/SNC9-H₈ tetramer⁺CD8⁺ cells in the spleens compared with those of all other groups (Fig. 4D, p < 0.05).

Surviving tumor-free mice from the combination-therapy group in studies shown in Fig. 1 were rechallenged at day 40 (Fig. 5A) or 3.5 mo (Supplemental Fig. 3) after initiation of therapy with 5 × 10⁵ TRAMP-C2 cells s.c. Another group of these surviving tumor-free mice was rechallenged with 1 × 10⁵ MC38 cells, an irrelevant tumor cell line (Fig. 5B). Mice that initially received the TRAMP-C2 tumor and that became and remained tumor free after receiving the combination regimen demonstrated resistance to TRAMP-C2 tumor development compared with mice in the control group that had not received either TRAMP-C2 tumor cells or any previous treatment (Fig. 5A, Supplemental Fig. 3A, p < 0.001). However, such tumor-free mice did not show any resistance to MC38 tumor development (Fig. 5B). FACS analysis showed that the percentages of CD44^highCD8⁺ and TRAMP-C2 tumor Ag-specific SPAS-1/SNC9-H₈ tetramer⁺CD8⁺ cells in PBMCs of mice surviving the initial tumor challenge were higher than those of mice in the control groups (Fig. 5C, Supplemental Fig. 3B). The protection from TRAMP-C2 tumor development mediated by the combination therapy was nearly abrogated by depletion of CD8⁺ cells (Fig. 5A, p < 0.01). To further verify that the protection from tumor development is tumor specific, we treated normal C57BL/6 mice with the combination regimen and then challenged the mice with TRAMP-C2 cells at day 40. The percentages of CD8⁺ and CD44^highCD8⁺ cells in the PBMCs of these mice increased, whereas the percentage of TRAMP-C2 tumor Ag-specific SPAS-1/SNC9-H₈ tetramer⁺CD8⁺ cells did not (Supplemental Fig. 4A); these mice did not show resistance to TRAMP-C2 tumor development on challenge (Supplemental Fig. 4B). These findings support the view that protection from tumor development on rechallenge correlated with development of TRAMP-C2 tumor Ag-specific SPAS-1/SNC9-H₈ tetramer⁺CD8⁺ T cells. Surviving tumor-free mice from the combination-therapy group shown in Fig. 1C were rechallenged at day 40 after initiation of therapy with 5 × 10⁵ TRAMP-C2/luc-GFP cells s.c. on the left flank, and photographs were taken at day 23 after rechallenge. These surviving tumor-free mice demonstrated resistance to TRAMP-C2/luc-GFP tumor development compared with mice in the control group that had not received either TRAMP-C2/luc-GFP tumor cells or any previous treatment (Fig. 5D).

We reported that administration of the anti-CD40 Ab increased the expression of IL-15Rα on the CD11c⁺ population of splenocytes (6). In the current study, we further examined IL-15Rα expression on other cell types by flow cytometry analysis. We demonstrated that treatment with anti-CD40 Ab increased the expression of IL-15Rα on CD11c⁺ cells, as well as on other cell types, such as B cells, the CD11b⁺ population, and CD8⁺ cells in spleen, compared with PBS-treated mice (Fig. 6A).

It was shown that murine DCs require IL-15Rα to prime NK cells (22). To further investigate the mechanism of synergetic antitumor effect of the combination regimen, we used an approach similar to that used by Koka et al. (22) to examine the abilities of BMDCs derived from wild-type or IL-15Rα^−/− mice to support NK cytolytic activity. BMDCs were pretreated with LPS, anti-CD40 Ab, or medium alone for 20 h and then cocultured with freshly isolated NK cells for an additional 20 h in the presence of LPS, anti-CD40 Ab, mIL-15, or mIL-15 plus anti-CD40 Ab. Subsequently, these NK cells were assayed for their ability to kill ⁵¹Cr-labeled TRAMP-C2 cells. NK cells cocultured with wild-type BMDCs in the presence of LPS or anti-CD40 Ab showed stronger cytolytic activity compared with those in medium alone (Fig. 6B). In contrast, lytic activity of NK cells primed by IL-15Rα^−/− BMDCs that were stimulated with LPS or anti-CD40 Ab did not increase (Fig. 6B). When mIL-15 was added to the coculture with anti-CD40 Ab, wild-type BMDC-primed NK cells demonstrated stronger cytolytic activity to the target cells than did the wild-type BMDC-primed NK cells in the presence of either mIL-15 or anti-CD40 Ab alone (Fig. 6C). In contrast, IL-15Rα^−/− BMDC-primed NK cells in the presence of both mIL-15 and anti-CD40 Ab showed cytolytic activity that was similar to that of IL-15Rα^−/− BMDC-primed NK cells in the presence of mIL-15 alone (Fig. 6D). These findings suggest that anti-CD40 Ab-induced expression of IL-15Rα by BMDCs is critical for the synergistic effect of the combination regimen on NK cell activation.

Receptors for IL-2 and IL-15 are heterotrimeric and both contain the γc (23) and IL-2/IL-15Rβ (24–27). In addition, the high-affinity forms of IL-2R and IL-15R contain a unique cytokine-specific α subunit (28). In light of the common receptor components, the two cytokines share several functions (13, 25, 26, 29, 30). In addition, there are distinct differences between the actions of IL-2 and IL-15 (30–33). In contrast to IL-2, IL-15 has no marked effect on regulatory T cells and is an antiapoptotic factor that inhibits IL-2–induced AICD in select systems (32). A critical factor in the functional differences between IL-2 and IL-15 involves distinct modes of action of these two cytokines. IL-2 is a secreted molecule that binds to preformed high-affinity heterotrimeric receptors (34, 35). In contrast, IL-15 is predominantly membrane bound and induces signaling in the context of an immunological synapse (9). Furthermore, IL-15Rα on activated DCs presents IL-15 in trans to cells that express IL-2/IL-15Rβ and γc but not IL-15Rα (9).

IL-2 has been approved by the FDA for the treatment of patients with metastatic malignancy (1, 2). However, IL-15 might be superior to IL-2 for cancer therapy (13, 25). IL-15 administration has shown therapeutic efficacy in animal models (4, 6, 36–38); however, it is not optimal as a single agent. IL-15Rα expression on monocytes and DCs is critical for transpresentation of IL-15 (9–11, 13). However, there is only a low level of expression of IL-15Rα on nonactivated DCs; therefore, IL-15Rα would be very limiting in therapeutic trials involving IL-15 alone.

In the current study, combination therapy with both mIL-15 and anti-CD40 Ab showed enhanced efficiency compared with monotherapy with either mIL-15 or anti-CD40 Ab (Fig. 1). With the combination treatment there was a highly significant prolongation of survival of TRAMP-C2 tumor-bearing mice, with 70–100% of the mice with established tumors becoming tumor free. Furthermore, surviving tumor-free mice in the combination group demonstrated resistance to TRAMP-C2 tumor development when they were rechallenged (Fig. 5, Supplemental Fig. 3). A number of studies was performed to define the mechanisms underlying this synergy between IL-15 and anti-CD40 Ab. In contrast to our previous study (6), in which we demonstrated that NK cells appear to be the primary effector cells mediating the killing of CT26 and MC38 tumors after combination treatment, in the TRAMP-C2 model, CD8⁺ T cells were shown to play a major role in the combination regimen-mediated antitumor efficacy (Fig. 3). Depletion of CD8⁺ T cells significantly reduced the antitumor efficacy mediated by the combination therapy (Fig. 3). Furthermore, the combination regimen dramatically increased the absolute number of TRAMP-C2 tumor-specific SPAS-1/SNC9-H₈ tetramer⁺CD8⁺ T cells compared with that of all other groups (Fig. 4D), and the protection from tumor development on rechallenge with TRAMP-C2 tumor cells was associated with an increased number of TRAMP-C2 tumor-specific tetramer⁺CD8⁺ T cells (Fig. 5). Therefore, the combination therapy of IL-15 plus anti-CD40 Ab has the potential to enhance the antitumor efficacy of both MHCI- and non-MHCI–restricted cytotoxic cells, which may broaden its application in human cancer therapy.

Nevertheless, there was also evidence to suggest that NK cells might also contribute to the therapeutic efficacy in the TRAMP-C2 model. It is known that, in addition to supporting NK cell survival, IL-15 supports NK cell activation. In the current study, the administration of anti–asialo-GM1 decreased antitumor efficacy mediated by the combination regimen (Fig. 3). Although anti–asialo-GM1 largely eliminated the NK cell population (Supplemental Fig. 2A), it also reduced the proportion of CD8⁺ T cells, especially CD44^highCD8⁺ and TRAMP-C2 tumor-specific tetramer⁺CD8⁺ T cells (Fig. 3A). Previously, it was demonstrated that non-NK cell expression of IL-15Rα is essential to support the generation, proliferation, and survival of NK cells (39). We demonstrated that administration of anti-CD40 Ab increased the expression of IL-15Rα on DCs (6), as well as other cell types, such as B cells, the CD11b⁺ population, and CD8⁺ cells. Furthermore, we showed that IL-15Rα on activated human monocytes and DCs presents IL-15 in trans to NK cells not expressing IL-15Rα (9). Thus, expression of IL-15Rα on DCs, as well as other cell types, appears to be critical for the activation of NK cells and for cytotoxicity against tumor cells.

Our present studies were directed toward defining whether anti-CD40–induced IL-15Rα expression is an important feature of antitumor efficacy of the combination therapy of IL-15 plus agonistic anti-CD40 Ab. Although the combination treatment was quite effective in the therapy of TRAMP-C2 tumor-bearing wild-type mice (Fig. 1), treatment with the combination regimen showed less therapeutic efficacy on tumor growth in IL-15Rα^−/− mice (Fig. 2). Furthermore, NK cells isolated from wild-type mice that received the combination regimen manifested greater lysis activity toward TRAMP-C2 tumor cells than did NK cells from IL-15Rα^−/− mice, suggesting that both decreased numbers of NK cells and lower lysis activity per living cell contributed to reduced antitumor activity. Although these studies in IL-15Rα^−/− mice support the view that anti-CD40 Ab-induced IL-15Rα expression contributes to efficacy of the combination regimen, the results are difficult to interpret because IL-15Rα^−/− mice are deficient in NK and CD44^highCD8⁺ T cells. In an attempt to circumvent this problem, we used an approach similar to that initiated by Koka et al. (22). These scientists found that mIL-15Rα^−/− DCs failed to support NK cell elaboration of IFN-γ and cytolytic activity. In parallel with their studies, we analyzed the ability of BMDCs derived from wild-type or IL-15Rα^−/− mice to prime NK cell cytolytic activity. We used anti-CD40 Ab to preactivate BMDCs and then cocultured these cells with freshly isolated NK cells in the presence of mIL-15, anti-CD40 Ab, or the combination of mIL-15 plus anti-CD40 Ab. Subsequently, these NK cells were assayed for their ability to kill TRAMP-C2 cells. NK cells cocultured with wild-type BMDCs in the presence of LPS or anti-CD40 Ab showed increased cytolytic activity. In contrast, the lytic activity of NK cells primed by IL-15Rα^−/− BMDCs stimulated with LPS or anti-CD40 Ab did not increase over control levels. Koka et al. (22) demonstrated that wild-type and IL-15Rα^−/− BMDCs expressed comparable levels of the activation/maturation markers CD40 and CD86 both before and after activation and that elaboration of IL-12 by DCs after stimulation was comparable between wild-type and IL-15Rα^−/−. However, in studies of Koka et al. (22) and ourselves IL-15Rα^−/− BMDCs were unable to fully activate NK cells to produce IFN-γ secretion or prime them to become cytolytic effectors. Taken together, these studies support the view that CD40 activation induces IL-15Rα expression on the surface of BMDCs, and this cytokine receptor presents IL-15 in trans to NK cells, leading to NK cell activation and increased cytolytic activity. This augmented IL-15Rα expression mediated by anti-CD40 Ab appears to be a critical element for an optimal response in IL-15–associated sustained remissions observed in TRAMP-C2 tumor-bearing mice receiving combination therapy.

In summary, given the multiple mechanisms of action of agonistic anti-CD40 Abs, we cannot conclude that the sole mode of action in the current study was the observed augmentation of IL-15Rα expression on DCs, as well as other cell types. Nevertheless, in addition to its other antitumor actions, the administration of agonistic anti-CD40 Ab and its associated increased expression of IL-15Rα facilitate the transactivation action of IL-15 on target effector CD8⁺ and NK cells. These findings in the current study provide the scientific basis for human clinical trials of patients with cancer to determine whether a similar synergistic antitumor activity is observed after administration of agonistic Abs to CD40 in combination with IL-15.

We thank Dr. Richard Bamford (Transponics, Jacobus, PA) for helpful discussions.

The authors have no financial conflicts of interest.