Simultaneous Blockade of Multiple Immune System Inhibitory Checkpoints Enhances Antitumor Activity Mediated by Interleukin-15 in a Murine Metastatic Colon Carcinoma Model

Read the Commentary on this article by Carson, p. 5917

Interleukin-15 (IL-15), a member of the 4-α-helix bundle family of cytokines is required for the generation of natural killer (NK) cells and lasting CD8⁺ memory T-cell responses (1). IL-15 is transpresented by its receptor, IL-15Rα, expressed on the surface of antigen presenting cells to the IL-2Rβ and common gamma (γ_c) chains expressed on effector T, B and NK cells (2, 3). IL-15 facilitates the survival and differentiation of these cells, their activation, and the maintenance of memory T cells (4, 5). In preclinical studies, IL-15 has been shown to enhance both humoral and cell-mediated immune responses leading to the inhibition of tumor growth (6–9).

A number of inhibitory receptors have been demonstrated to dampen or terminate immune responses in the setting of chronic viral infections and in tumor-bearing animals. These include programmed death-1 (PD-1, CD274; ref. 10), and its ligands, PD-L1 and PD-L2 (11), and cytotoxic T-lymphocyte antigen 4 (CTLA-4). PD-1 is a member of the CD28/CTLA-4 family of regulatory T-cell receptors that is expressed on the surface of activated B and T cells and is involved in the induction of immune tolerance (12). PD-L1 is constitutively expressed at low levels on hematopoietic cells, including resting T, B, myeloid, and dendritic cells, as well as some nonhematopoietic cells in the lung, heart, and other organs (13). PD-L1 is upregulated during T-cell activation and it has been shown to interact with PD-1 and CD80. The interaction of PD-1 with its ligands results in an inhibitory signal in activated T cells that promotes apoptosis and anergy. Similarly, the interaction of PD-L1 and CD80 also delivers an inhibitory signal to activated T cells (14). Chronic stimulation of PD-1 may result in T-cell “exhaustion” and the attenuation of immune responses (15). Tumors may exploit these checkpoint controls to inhibit antitumor immune responses.

CTLA-4 (CD152), a member of the immunoglobulin superfamily is recognized as another important negative regulator of the immune response (16). CTLA-4 is expressed on the surface of T cells and exhibits immunosuppressive function, downregulating T-cell activation in response to engagement of the T-cell receptor. Blockade of the interaction of CTLA-4 with its ligands CD80 and CD86, using monoclonal antibodies, can reverse inhibition of activation and stimulate T-cell proliferation that leads to enhanced antitumor immunity in preclinical models and in patients (17, 18). The combined blockade of both the PD-1 and CTLA-4 pathways in conjunction with an antitumor vaccine was recently reported (19). The combination of anti-PD-1, anti-PD-L1, and anti-CTLA-4 antibodies along with a gene-modified tumor cell vaccine increased the numbers of tumor infiltrating T cells in a mouse B16 melanoma model. This strategy resulted in a greater tumor rejection compared with the vaccine without the blockade of the 2 critical checkpoints.

In the present study, we demonstrated that in addition to its immunostimulatory effect, IL-15 also increased expression of the negative T-cell regulator, PD-1, and enhanced secretion of the immunosuppressive cytokine IL-10. Blockade of multiple negative immune checkpoints by the simultaneous administration of monoclonal antibodies directed against PD-L1 and CTLA-4 reduced expression of PD-1, secretion of IL-10, and enhanced IL-15–mediated immune responses and provided greater therapeutic benefit in a metastatic murine CT26 colon carcinoma model.

Female 6- to 8-week-old BALB/c mice were obtained from NCI-Frederick and were maintained in the National Cancer Institute (NCI) animal holding facility. Animal use adhered to NIH Laboratory Animal Care Guidelines and was approved by the NCI Animal Care and Use Committee.

The CT26 cell line, an N-nitro-N-methylurethane–induced BALB/c murine colon carcinoma cell line was purchased from American Type Culture Collection and maintained in RPMI 1640 (Invitrogen) supplemented with 10% FCS (fetal calf serum), l-glutamine, sodium pyruvate, streptomycin, and penicillin (Invitrogen).

Groups of 5 or 6 mice were injected with 2 × 10⁵ CT26 tumor cells by the tail vein on day 0. Treatment began 1 day later (day 1). Each mouse received 5 μg of mIL-15 (PeproTech) intraperitoneally daily, 5 times a week for 3 weeks. Along with mIL-15, some groups of animals also received anti-mouse-PD-L1 antibody (clone 9G2; Bio-express), anti-CTLA-4 antibody (clone UC10-4F10-11; Bio-express), or both anti-mouse-PD-L1 and anti-CTLA-4 antibodies or as a control, an irrelevant isotype IgG (Bio-express). The dose and schedule of these antibodies was 100 μg per injection administered twice a week for 2 weeks. Control mice received injections of phosphate-buffered saline (PBS; Biofluids). On day 21, the mice were euthanized by CO₂ inhalation, their lungs harvested, and the tumor nodules counted. In brief, the lungs were stained by intratracheal administration of India ink, removed, fixed in Fekete's solution, and metastatic nodules >1 mm counted under a dissection microscope. In a matched series of experiments, similar treatment groups were followed daily for survival. Mice were humanely sacrificed using prospective ACUC-approved criteria to generate survival curves.

Groups of mice (n = 10) were injected intravenously with 2 × 10⁵ CT26 cells on day 0 and then received mIL-15 as described above. Groups of animals received 200 μg of either rat anti-mouse-CD8 (clone 2.43; Bio-express) or an isotype control rat IgG (clone LTF-2; Bio-express) intraperitoneally beginning on days 0 and 1, and then 3 times a week for 3 weeks. Animal survival was followed daily.

Expression of MHC class I on CT26 cells was analyzed by staining with fluorescein isothiocyanate (FITC)-anti-mouse-H-2K^d. Isotype-matched IgG was utilized as a control. To compare the effects on the immunophenotype of CD8⁺ T cells of the mIL-15 and antibody combination treatments, surface expression of PD-1 on the CD8⁺ T cells as well as the CD8⁺CD44^high cell populations was evaluated using flow cytometry. Spleen cells were stained with allophycocyanin (APC)-conjugated anti-mouse-CD8 (53–6.7), phycoerythrin (PE)-Cy5.5–conjugated anti-mouse-CD44 (IM-7) and PE-conjugated anti-mouse-PD-1 (RPM-4) or with an isotype control, and incubated for 30 minutes on ice. Fc receptor binding was minimized by preincubation of the cells with rat anti-mouse-CD16/CD32. All fluorescent-labeled antibodies were obtained from BD Biosciences. Immunofluorescence analysis was performed on a FACSCalibur (BD Biosciences) and analyzed with FlowJo software (Tree Star).

Splenic CD8⁺ T cells of animals treated with mIL-15 alone or in combination with anti-PD-L1 and/or anti-CTLA-4 were isolated using a magnetic column (Miltenyi Biotec). In brief, each spleen was processed into a single cell suspension, incubated with anti-mouse-CD8α microbeads and sorted using magnetic column according to the manufacturer's instructions (Miltenyi Biotec). The CD8⁺ T-cell fractions were more than 94% pure as assessed by flow cytometry.

Cytokine secretion by splenic CD8⁺ T cells from tumor-bearing animals was assayed by incubating CD8⁺ T-cells (2 × 10⁵ per well) on anti-mouse CD3 (clone 2C11; BD Biosciences; 10 μg/mL)–coated plates with 1 μg/mL of soluble anti-mouse CD28 (clone 37.51; BD Biosciences) for 72 hours and then collecting the supernatants. Splenic CD8⁺ T cells cocultured with the same concentrations of isotype-matched antibodies were set up to control for background. IFNγ and IL-10 were measured by an ELISA (R&D Systems). The cytokine concentration in the supernatants was interpolated from the linear portion of the ELISA standard curve.

The ability of naive BALB/c splenic CD8⁺ T cells to respond to mIL-15, anti-PD-L1, and anti-CTLA-4 ex vivo was also examined. Naive BALB/c splenic CD8⁺ T cells were incubated with mIL-15 at a concentration of 20 ng/mL alone or in combination with anti-mouse-PD-L1 (10 ng/mL) and/or 10 ng/mL of anti-mouse-CTLA-4 or 10ng/mL of isotype IgG. Meanwhile, CD8⁺ T cells cocultured with different concentrations of mIL-15 were included. These cells were incubated on anti-mouse-CD3–coated plates (10 or 4 μg/mL as indicated) with 1 μg/mL of soluble anti-mouse-CD28 for 72 hours and the concentration of IL-10 secretion was determined by ELISA, as described above. Coculture with media without mIL-15 was used as a control.

Single cell suspensions of spleen cells were prepared from mice sacrificed on day 21 after CT26 tumor challenge and cultured with irradiated (100 Gy) CT26 tumor cells at a ratio of 50:1. Recombinant hIL-2 (Hoffmann-LaRoche Inc) was added to a concentration of 10 to 15 U/mL. After 4 days, 1 × 10⁶ effector cells were cultured with CT26 tumor cells at a ratio of 20:1 for 6 hours. Brefeldin A (10 μg/mL; Sigma) was added to the cultures for the last 5 hours to prevent secretion of the intracellular protein. Cells from each group were first incubated with APC-conjugated anti-mouse-CD8 Ab (clone 53–6.7; BD Biosciences) on ice for 30 minutes. The cells were fixed and permeabilized using a cytotofix/cytoperm kit as instructed by the manufacturer (BD Bioscience). Briefly, cells were resuspended in a fixing buffer for 20 minutes at room temperature and then washed with permeabilization buffer. Cells were then stained with anti-mouse-IFNγ-PE (clone XMG1.2; BD Bioscience) in permeabilization buffer at 4°C for 30 minutes. Cells were washed in permeabilization buffer, resuspended in FACS buffer, and analyzed on a FACScalibur flow cytometer (Becton Dickinson).

A cytotoxicity assay was performed as follows: effector cells from each of the treatment groups were cultured with 10⁴ CT26 target cells per well in triplicate at varying effector/tumor (E/T) ratios and incubated at 37°C for 4 hours. Cytotoxic activity was measured by LDH (lactate dehydrogenase) release using the CytoTox 96 nonradioactive cytotoxicity assay (Promega). The percentage of cytotoxicity was calculated as 100 × [(experimental release) − (effector spontaneous release) − (target spontaneous release)]/[(target maximum release) − (target spontaneous release)].

Kaplan–Meier nonparametric regression analysis was performed to assess the survival time of tumor-bearing animals with significance determined by the log-rank test using JMP statistical software (SAS Institute). The comparison of cytokine secretion among groups was analyzed by using an unpaired Student's t test. A value of P < 0.05 was considered significant.

To examine the potential for CT26 tumors to serve as targets of CD8⁺ CTL, we examined MHC class I expression on these cells. CT26 cells in vitro (Fig. 1A) or ex vivo from lung nodules of tumor-bearing animals (Fig. 1B) demonstrated high levels of MHC class I expression with mean fluorescence intensities (MFI) of 391.2 and 407.4, respectively suggesting that these tumors may be targeted by CD8⁺ cytolytic T cells.

To see if IL-15 could induce CD8⁺ CTL-mediated protection against tumor, mIL-15 was administered beginning 1 day after intravenous injection of 2 × 10⁵ CT26 tumor cells and the mice followed for survival (Fig. 1C). Animals treated with mIL-15 demonstrated longer survivals compared with control mice receiving PBS control (P < 0.05). The median survival of mIL-15 treated mice was 44 days whereas those treated with PBS survived a median of 19 days. To determine whether CD8⁺ T cells played a role in this effect, we used anti-CD8 antibody to deplete CD8⁺ T cells. The survival advantage afforded by mIL-15 was largely abrogated by the depletion of CD8⁺ T cells indicating that they play a major role in protection from tumor in this model (Fig. 1C). Animals administered IL-15 along with an IgG control exhibited no significant difference in survival when compared with IL-15 alone (data not shown).

As increased PD-1 expression on CD8⁺ T cells has been shown to attenuate immune responses (9–11), we examined the effects of mIL-15 on the expression of PD-1 on CD8⁺ T cells from tumor-bearing animals. Mice treated with mIL-15 showed increased PD-1 expression on CD8⁺ T cells (Fig. 2A and C; compared with PBS group, P < 0.05) as well as on CD8⁺CD44^high cells in the spleen (Fig. 2B and D; compared with PBS group, P < 0.05). When anti-PD-L1 was administered, PD-1 expression decreased compared with that seen with mIL-15 treatment alone. This was seen in both the CD8⁺ T cells and the CD8⁺CD44^high T-cell population (Fig. 2C and D, P < 0.05). Anti-CTLA-4 when used alone did not reduce PD-1 expression on CD8⁺ T cells (data not shown); however, when anti-CTLA-4 and anti-PD-L1 were combined, a further decrease in the expression of PD-1, most notably in the CD8⁺CD44^high population, was noted (Fig. 2D, P < 0.05).

On day 14 after tumor inoculation, CD8⁺ T cells obtained from the spleens of mice in the different treatment groups were stimulated with anti-mouse-CD3 and anti-mouse-CD28, and IFNγ and IL-10 secretion were measured. CD8⁺ T cells isolated from animals treated with mIL-15 demonstrated an increased IFNγ secretion compared with the PBS control group (P < 0.05; Fig. 3A). The addition of anti-PD-L1, or the combination of anti-PD-L1 and anti-CTLA-4 to treatment with mIL-15 resulted in further increases in IFNγ secretion.

Splenic CD8⁺ T cells isolated from animals treated with mIL-15 also secreted higher levels of IL-10 compared with those isolated from PBS-treated animals (P < 0.05; Fig. 3B). Coadministration of antibodies to PD-L1 and CTLA-4 in the mIL-15 treated mice inhibited the secretion of IL-10 below detectable levels (<16 pg/mL). Anti-CTLA-4 alone had no significant effect on either IFNγ or IL-10 secretion compared with PBS-treated control animals (data not shown).

To clarify the relationship between IL-15 and IL-10 secretion, splenic CD8⁺ T cells were isolated from naive mice and stimulated with anti-mouse-CD3 and anti-mouse-CD28 as described. Different concentrations of mIL-15 were provided to the cultures, supernatants were collected, and IL-10 was quantitated (Fig. 3C). There was a dose-dependent relationship between IL-10 secretion and the concentration of mIL-15. Moreover, if both anti-PD-L1 and anti-CTLA-4 antibodies were added to the culture system along with mIL-15, they significantly reduced IL-10 secretion by CD8⁺ T cells compared with mIL-15 alone (P < 0.05; Fig. 3D) No statistical difference was observed between mIL-15 (20 ng/mL) alone treated cells and mIL-15 (20 ng/mL) combined with isotype IgGs (10 ng/mL) treated cells (data not shown).

Tumor-specific IFNγ secretion was detected by intracellular staining of splenic CD8⁺ T cells. There was no obvious IFNγ secretion by PBS-treated CD8⁺ T cells, but mIL-15 treatment as well as mIL-15 combined with anti-PD-L1 and anti-CTLA-4–treated tumor-bearing animals showed a significantly higher number of IFNγ secreting CD8⁺ T cells in the spleen (Fig. 4A) compared with the PBS control group (P < 0.01). The ability of CD8⁺ T cells isolated from tumor-bearing mice to induce CT-26–specific cell lysis was examined (Fig. 4B). CD8⁺ T cells isolated from PBS-treated animals did not demonstrate any lytic activity against CT26 cells, whereas CD8⁺ T cells isolated from mice treated with mIL-15 alone or in combination with anti-PD-L1 and/or anti-CTLA-4 showed increased CD8⁺ T-cell lytic activity against CT26 cells in an effector:tumor cell ratio–dependent manner.

Treatment with IL-15 significantly reduced the number of tumor nodules in the lungs of mice injected with CT26 cells compared with animals receiving PBS, P < 0.05 (Fig. 5A). Coadministration of anti-PD-L1 and mIL-15 also led to a reduction in the number of tumor nodules compared with PBS treatment, P < 0.05 (Fig. 5A). The combination of mIL-15, anti-PD-L1, and anti-CTLA-4 led to a further decrease in the number of pulmonary metastases relative to that observed in the mIL-15–treated group (P < 0.05). Representative lungs are shown in Figure 5B.

We next examined if mIL-15 in combination with anti-PD-L1 and anti-CTLA-4 could improve the survival of mice with metastatic CT26 tumors (Fig. 6). Mice treated with PBS survived a median of 21 days (range: 19–23 days). Animals treated with anti-CTLA-4 did not show a survival advantage over animals treated with PBS (P = 0.075). In contrast, animals treated with anti-PD-L1 alone did exhibit a survival advantage over those treated with PBS (P < 0.05); however, when compared with the mIL-15 treated group, their survival was significantly less (P < 0.05). Mice receiving mIL-15 demonstrated a significant prolongation of survival when compared with the PBS-treated group (P < 0.01). Combining mIL-15 with an irrelevant IgG provided no survival advantage compared with mIL-15 alone (data not shown). Combining mIL-15 and anti-CTLA-4, we found that there was no survival advantage compared with treatment of mIL-15 alone (median survival: 46 vs. 45 days; P = 0.399). However, when mIL-15 was combined with anti-PD-L1, there was a significant survival benefit with this addition over mIL-15 alone, (median survival: 59 vs. 45 days; P < 0.01). The addition of anti-CTLA-4 to mIL-15 and anti-PD-L1 treatment gave the greatest survival benefit compared with mIL-15 treatment alone or treatment with mIL-15 in combination with anti-PD-L1 alone, (median survival: 74 vs. 59 days; P < 0.05).

Many immunotherapeutic strategies directed against cancer focus on enhancing effector cell function in an attempt to increase tumor killing. Despite attempts to augment antitumor immunity by a variety of approaches, the enthusiasm for many of these approaches has been diminished by the fact that the generation of measurable increases in antitumor effector responses have not often correlated with clinical responses. It is now recognized that at least part of this failure is attributable to the presence of numerous naturally occurring negative regulatory pathways that act to limit immunological responses.

IL-15 is a cytokine that is important in the development and homeostasis of several lymphocyte populations including NK, NK/T, and CD8⁺ T cells (20–25). In addition, IL-15 can also stimulate proliferation, enhance cytotoxicity, and upregulate production of IFNγ by NK- and T cells (20, 21). In the current study, we showed that IL-15 upregulated IFNγ secretion by CD8⁺ T cells and prolonged the survival of tumor-bearing animals. However, IL-15 as monotherapy also suffers from the potential limitations imposed by upregulation of negative regulatory checkpoints. In this study, we showed that treatment with IL-15 led to increased expression of PD-1 on CD8⁺ T cells, most notably on CD8⁺CD44^high memory phenotype T cells. Further, we demonstrated that IL-15 also induced the production of IL-10, a negative regulatory cytokine (26), suggesting that IL-15 may work as a “double-edged sword.” In an attempt to address this issue, we studied the effects of IL-15 combined with the removal of a number of these negative immune regulatory checkpoints.

PD-1 is a surface molecule that is overexpressed on apoptotic cells (27). It moved into the immunologists' sights as a molecule that is also expressed on “exhausted” T cells from chronic virus-infected patients and in patients with advanced cancer that contributed to poor T-cell proliferative capacity and incompetent effector function (28–30). One study demonstrated that the numbers of T cells expressing PD-1 were increased in patients with high-risk renal cell carcinoma (31). Patients with increased numbers of PD-1⁺ T cells were at greater risk of cancer-specific death compared with patients whose T cells exhibited a low expression of PD-1. In addition, one of the ligands of PD-1, PD-L1, has also become a focus of interest due to its broad expression and negative regulatory function. Blockade of the PD-1/PD-L1–negative checkpoint has been demonstrated to restore effector CD8⁺ T-cell responses in chronic infections such as HIV, hepatitis C, and lymphocytic choriomeningitis virus (32–34). In addition, monoclonal antibodies to PD-1 have produced tumor responses in clinical trials in patients with advanced cancer (35). In this study, we utilized an anti-PD-L1 antibody to block the PD-1/PD-L1 pathway in a murine metastatic colon cancer model. We found that administration of anti-PD-L1 antibody was able to reduce PD-1 expression on CD8⁺ T cells following treatment with IL-15 in vivo. Mice treated with the combination of IL-15 with anti-PD-L1 also had CD8⁺ T cells that secreted lesser amounts of the immune inhibitory cytokine, IL-10. Moreover, this combination also enhanced IFNγ secretion and prolonged the survival of tumor-bearing animals.

To examine if we could further increase the therapeutic benefit provided by eliminating negative checkpoints, we also inhibited CTLA-4. Antibody-mediated CTLA-4 blockade has been shown to improve the outcome of tumor-bearing animals and patients with advanced malignant melanoma (18, 36). Blockade of CTLA-4 has also been shown to inhibit the secretion of IL-10 (37). The removal of CTLA-4 function, in our system, when used alone, compared with PBS, or in combination with IL-15, compared with IL-15 alone, did not lead to a statistically significant survival advantage. However, when used in combination with PD-L1 inhibition and IL-15, it showed a significantly prolonged survival when compared with any of these agents used individually or in single combination with IL-15. Both PD-1 and CTLA-4 ligation inhibit Akt activation and block T-cell responses by targeting distinct signaling molecules (38). PD-1 engagement inhibits an upstream proximal step blocking phosphatidylinositol-3-kinase (PI3K) activation. In contrast, signaling by CTLA-4 preserves PI3K activity; instead, it functions through its binding to phosphatase PP2A leading to the inhibition of Akt phosphorylation and ultimately the inhibition of T-cell activation (38). By blocking both PD-L1 and CTLA-4, we are able to target 2 pathways leading to Akt activation, the result being a likely synergistic response allowing effective T-cell activation and ultimately an increased therapeutic effect. A recent study by Curran et al. found that the combination of CTLA-4 and PD-1/PD-L1 blockade may also have effects on the tumor microenvironment (19). They showed that this combination resulted in increased tumor-infiltrating T cells with a reduction in the regulatory T- cell populations within the tumor. Whether immune stimulation through IL-15 in combination CTLA-4 and PD-1/PD-L1 blockade leads to similar changes in tumor-infiltrating cells is yet to be determined. Although Curran and colleagues also used a whole cell antitumor vaccine, their results in terms of increased animal survival were similar to ours highlighting the benefit of multiple immune checkpoints removal.

In the present study, we found that to achieve an effective immune response to a tumor not only is the activation of the immune system by IL-15 required but also the simultaneous inhibition of 2 independent negative immune regulatory checkpoints (CTLA-4 and PD-l/PD-Ll) that act to attenuate the immune response. Although IL-15 effectively activated the immune system against tumor, it also activated negative regulators such as IL-10 and PD-1, potentially limiting its therapeutic efficacy. By using a strategy that combines IL-15 with the simultaneous blockade of both CTLA-4 and PD-1 pathways, we reduced the expression of IL-10 and PD-1 allowing a stronger immune response and an increased animal survival.

The authors disclosed no potential conflicts of interest.

This work was supported by the Intramural Research Program of the NIH, National Cancer Institute, Center for Cancer Research.