Abscopal Effects of Radiotherapy Are Enhanced by Combined Immunostimulatory mAbs and Are Dependent on CD8 T Cells and Crosspriming

Radiotherapy is a solid pillar of cancer treatment used to treat localized stages of a broad variety of malignant diseases and to alleviate local complications in advanced or metastatic cases as a palliative treatment. The mechanism of action of ionizing radiotherapy against cancer is thought to mainly rely on catastrophic damage of genomic DNA, leading to apoptotic tumor cell death. Many cellular genetic and epigenetic factors affect the sensitivity of each tumor to radiotherapy approaches. Recently, the tumor stroma component has been found to play a key role in the outcome of irradiated tumors (1). When radiotherapy is prescribed to a patient, it is assumed that the normal nonmalignant tissue will also be irradiated giving rise to multifarious biological effects including inflammation and scarring (1). Radiotherapy can be performed by applying an external beam of irradiation or by the temporal surgical insertion of radiation sources guided by catheters into the cancer tissue using techniques collectively known as brachytherapy.

Immunotherapy is emerging as another major pillar for the treatment of cancer treatment. mAbs acting on immune system receptors to derepress or agonistically augment antitumor immunity are being developed in the clinic (2). Antibodies against the inhibitory (checkpoint) receptor CTLA-4 were the first to be clinically developed with ipilimumab receiving FDA and European Medicines Agency (EMA) approval for metastatic melanoma (3). Among these checkpoint inhibitor monoclonal immunostimulatory antibodies, agents blocking the PD1/PD-L1 receptor/ligand pair have already attained FDA and EMA approval for metastatic melanoma (4), non–small cell lung cancer (5–7), and renal cell carcinoma (8) and other indications are under regulatory evaluation. This achievement was preceded by extensive and successful preclinical research in mouse models.

Agonist antibodies crosslinking CD137 (4-1BB) were also shown to enhance antitumor immunity in mice to the point of causing the rejection of transplanted tumors (9). Two antibodies against CD137 are undergoing phase II clinical trials with promising results (10, 11). Anti-PD1 and anti-CD137 mAbs act on T cells that express these receptors on their plasma membrane presumably as a consequence of an antigen-cognate activation process. Hence, the main mechanism of action is exerted on tumor-infiltrating lymphocytes that express such receptors on their surface, thus becoming amenable to pharmacologic modulation with the corresponding mAb. In preclinical mouse models, anti-CD137 and anti-PD1 mAbs exert powerful synergistic effects (12) that have given rise to two ongoing clinical trials testing such a combination (NCT02253992, NCT02179918).

The interphase between radiotherapy and immunotherapy is an exciting emerging topic. Radiotherapy causes biological effects known to both ignite (13, 14) and quench the cellular immune response (13, 14). The type of cell death induced by radiotherapy is considered immunogenic (15, 16), because it sets in motion multiple alarmins (15, 16) and proinflammatory mechanisms (17). Radiotherapy-induced cell death is a potential source of tumor antigens to be uptaken, processed, and presented by dendritic cells to CD8⁺ T lymphocytes, a process that is collectively known as crosspresentation (18) and termed crosspriming if it results in CTL activation. Crosspresentation to CD8⁺ T cells is mainly mediated by a specialized subset of dendritic cells, which are dependent for development on the Batf-3 transcription factor (19) and on sFLT-3L as a growth factor. We have published that this dendritic cell (DC) subset is critical for the therapeutic effects of anti-PD1 and anti-CD137 mAbs by means of crosspresentation of tumor antigens (20). This DC subset is also known to be involved in eliciting postradiotherapy CTL immune responses (21). However, other mechanisms such as irradiation-dependent TGFβ production and myeloid cell recruitment are considered immunosuppressive.

Immunostimulatory mAbs have already been combined with radiotherapy in preclinical models. Anti-CTLA4 mAb (22), anti-PD1 mAb (23, 24), and anti-CD137 mAb (25–27), show evidence for synergistic effects with external beam irradiation. Furthermore, triple combinations of radiotherapy with anti CTLA-4 plus anti PD1 exert efficacious synergistic effects against B16F10 melanoma tumors as seen against the directly irradiated tumor and a concomitant tumor, implanted outside the irradiation field (28), a phenomenon known as the abscopal effect of radiotherapy (29).

Anecdotal evidence in the clinic suggests that in a patient treated with anti CTLA-4 mAb (ipilimumab) and subsequent palliative radiotherapy, there were objective responses outside the irradiation field, concurrent with increases in the titer of antibodies against the shared tumor antigen NY-ESO1 (30). In a phase II clinical trial testing the ipilimumab plus radiotherapy combination, there was a trend toward better overall survival in metastatic melanoma patients (28).

In this study, we use different mouse models to demonstrate that external beam radiotherapy synergizes with immunostimulatory anti-PD1 and anti-CD137 mAbs as single agents and when used in combination. The therapeutic effects were attributed to CD8 T cells by depletion experiments and involved profound changes in the tumor microenvironment that include and augment the expression of the receptors to be targeted by the immunomodulatory mAb.

Tumor cells lines, MC38, a colon adenocarcinoma cell line of C57BL/6 origin whose identity (Case 6592-2012) was provided to us by Dr. Karl E. Hellström (University of Washington, Seattle, WA). The 4T1 breast carcinoma cells of BALB/c origin were originally provided by Dr. Claude Lecrec, Institute Pasteur, Paris, France, and verifed in the master cell bank at Institute Pasteur (Paris, France). B16F10-OVA melanoma–derived cells that are transfected to express chicken ovalbumin (OVA) were verified by Idexx Radil in 2012 and kept in a master cell bank as vials thawn every 3–6 months and were cultured in RPMI1640 supplemented with 10% FBS, 2 mmol/L l-glutamine, 0.05 mmol/L 2-mercaptoethanol, HEPES, penicillin, and streptomycin at 37°C in a humidified atmosphere containing 5% CO₂. All these cells lines were certified as being free of contamination by Mycoplasma using the Mycoplasma detection kit (MycoAlert Mycoplasma Detection Kit from Lonza).

C57BL/6 female mice were injected subcutaneously with 5 × 10⁵ MC38 and 5 × 10⁵ B16OVA cells, respectively, in the right flank (primary tumor) and with 3 × 10⁵ MC38 and 3 × 10⁵ B16OVA cells in the left flank (secondary tumor). A similar scheme was used to subcutaneously engraft 4T1 cells in female BALB/c mice. Perpendicular tumor diameters were measured with a Vernier calipers every 2–3 days, and tumor volumes were calculated. On day 11, when both tumors were palpable, animals were randomly assigned to 8 groups receiving or not receiving radiotherapy (8 Gy × 3 fractions), to only one of the two tumors, in combination or not in combination with intraperitoneal immunostimulatory mAbs (anti-PD1, anti-CD137, or both). Anti-PD1, anti-CD137, the combination, or anti-rat IgG control antibody were administered intraperitoneally at the dose of 200 μg/mouse (10 mg/kg) or 100 μg/mouse (5 mg/kg) on days 13, 15, and 17. In some experiments, monoclonal immunostimulatory antibodies were administered on days 17, 19, 20. Tumor size was monitored every 2–3 days and mice were sacrificed when tumor size reached 4,000 mm³. Tumor radiotherapy procedures are detailed in Supplementary Materials section.

Tumor tissue was processed to obtain single-cell suspension for flow cytometry analysis (see Supplementary Methods). To estimate absolute numbers in cell suspension, perfect count microspheres were used as an internal standard according to the manufacturer's instructions (Cytognos).

Levels of human IFNγ in mouse plasma samples were measured by a commercial ELISA (Human IFNγ Elisa Set, BD OptEIA, BD Biosciences), following the manufacturer's instructions. All samples were measured in duplicate. The detection cut-off levels of the assay were 4.7 pg/mL for IFNγ. The coefficient of variation was <15%. For tumor antigen–specific CD8 T-cell assessment, a H-2K^b KSPWFTTL tetramer labeled with PE (manufactured by Biolegend) was used. For gating and costaining, the following mAbs were used: CD45.2 PerCP/Cy5.5 (clone 104 from Biolegend), CD4 BV421 (clone RM4-5 from Biolegend), CD8 BV510 (clone 53-6.7 from Biolegend), CD137 biotin (clone 17B5 from Biolegend), and PD-1 FITC (clone 29F.1A12 from Biolegend).

Statistical differences between survival curves were analyzed with the Mantel–Cox, log-rank test, nonlinear regression and differences between other groups were analyzed with the Mann–Whitney U test using GraphPad Prism (GraphPad Software Inc.)

Mice bearing bilateral tumors derived from subcutaneous engraftment of MC38 colorectal carcinoma cells were used as a model to monitor the abscopal effects of radiotherapy in combination with immunostimulatory mAbs. Eight Gy fractionated doses of external beam radiotherapy were selectively applied only to one of the tumor lesions, while a contralateral tumor was set outside the irradiation field (see representative dosimetry in Supplementary Fig. S1A). Contralateral concomitant tumors were inoculated the same day with 10-fold fewer tumor cells. Radiotherapy given every other day was followed on alternate days by three doses of anti-CD137 or/and anti-PD1 mAbs. The mAbs were given as single agents or in combination as detailed in Supplementary Fig. S1B. Supplementary Table S1A individually shows the statistical comparisons of the evolution of irradiated and nonirradiated tumor lesions. Results collectively indicate that both anti-PD1 and anti-CD137 mAb contributed to control contralateral tumor growth when in conjunction with unilateral radiotherapy. Strikingly, the mice receiving radiotherapy and the combination of the two immunostimulatory mAbs were the group that achieved faster and almost constant complete responses (Supplementary Fig. S1C), translated in 100% long-term overall survival (Supplementary Fig. S1D). Of note, cured mice were immune 3 months later to MC38 tumor cell rechallenge, while able to engraft B16OVA melanoma cells as an antigenically unrelated control (Supplementary Fig. S1E).

Of note, combined treatment was well tolerated by the mice in terms of safety. Given the fact that CD137 mAb can cause liver inflammation (31), we assessed ALT serum levels and checked liver pathology specimens that ruled out increased toxicity due to the addition of local radiotherapy to the immunostimulatory antibody combination (unpublished observations).

Similar experiments were carried out with bilateral B16OVA melanoma (Fig. 1A), known to be of difficult treatment by immunotherapy (32, 33). In this case, mice bearing tumors for 11 days showed a radiotherapy-dependent control of contralateral tumors, when distant radiotherapy was combined with either anti-PD1 or anti-CD137 mAb. When both antibodies were combined together, all the tumors regressed bilaterally, even though combined immunotherapy without irradiation also induced the regression of most tumors (Fig. 1B; Supplementary Table S1A) achieving long-term survival (Fig. 1C). When mice cured by the radiotherapy plus mAb combination were rechallenged 3 months later, 3 of 5 mice were protected from B16OVA, while growing MC38 as a contralateral antigenically unrelated control tumor (unpublished observations). Remarkably, measurements of the concentration of IFNγ in sera from mice undergoing triple combined treatment (radiotherapy plus anti-PD1 plus anti-CD137) showed much higher levels than any other treatment regimen on day +18 (Fig. 1D). This fact strongly indicated an ongoing cellular immune response of far greater intensity.

4T1 breast cancer is an exceedingly difficult tumor model to treat with immunotherapy (34) that causes spontaneous lung metastases. In this setting, we again performed experiments with bilateral tumors, irradiating only one of the lesions (Fig. 2A). Tumor growth analyses indicated better local and distant tumor control when radiotherapy was combined with immunotherapy but without achieving complete responses (Fig. 2B; Supplementary Table S1A), although this treatment did lead to longer survival (unpublished observations). Furthermore, spontaneous metastases to the lung were followed by CT scans and by surgical inspection upon sacrifice with the quantification of the number and size of metastases (Fig. 2C and Supplementary Fig. S2). As can be seen in Fig. 2C, overall numbers of spontaneous lung metastases were reduced in the radiotherapy plus combined mAb immunotherapy group. In Supplementary Fig. S2, individual CT scan sections and representative excised lungs are shown.

In the case of bilateral MC38 tumors, experiments were also performed starting treatment as late as day +14 (Fig. 3A) after tumor cell engraftment to ascertain the limits of the strategy and demonstrate radiotherapy synergy with the anti-CD137 plus anti-PD1 combination regimen. In this case, the treatments were not curative in any case (Fig. 3B and C), but the delay in tumor progression induced by the triple combination (radiotherapy plus anti-PD1 plus anti-CD137) was readily seen in comparison when monitoring the contralateral tumor. No noticeable effects were exerted by each of the mAb when used separately in this regimen, or when the immunotherapy combination was employed without radiotherapy (Fig. 3B and C; Supplementary Table S1B).

Using treatment conditions comparable with those in Fig. 1 and Supplementary Fig. S1, on bilateral MC38-derived tumors, we repeatedly depleted CD8_β⁺ T cells, CD4⁺ T cells, and NK 1.1⁺ lymphocytes with specific mAbs (Fig. 4A). As can be seen in Fig. 4B and C, in mice treated with the combinatorial regimen of radiotherapy plus anti-CD137 and anti-PD1 mAbs, we observed that CD8 T cells were absolutely required for the contralateral antitumor effects. In contrast, CD4 T-cell depletion resulted in a more pronounced therapeutic effect with all animals achieving complete bilateral regressions and showing a striking effect on overall survival. NK1.1 depletion had little effect on the outcome of the contralateral tumors (Fig. 4B and C; Supplementary Table S1C). These results show the involvement of cytolytic T lymphocytes in the beneficial effect of the combinatorial regimen. CD4 depletion also eliminates regulatory T cells likely explaining the better outcome upon depletion with anti-CD4 mAb. Induction of CTLs against tumor antigens is mainly mediated by BATF-3–dependent DC (20). Accordingly, we performed similar experiments in BATF-3–deficient mice, which showed no abscopal effects and had weaker local tumor control by radiotherapy (Fig. 5A and B). In line with this, BATF-3^−/− mice showed no increases in serum IFNγ following combined treatment (Fig. 5C)

As both CTLs (35) and crosspriming (36) are known to be dependent on type I IFNs, we next studied treatment in the same experimental setting (Fig. 5A) of mice devoid or not of type I IFN receptor (IFNAR^−/−). As can be observed in Fig. 5D, the abscopal effects exerted by combined radioimmunotherapy were completely abrogated in IFNAR-deficient mice. Importantly, the local effects on the directly irradiated tumors were also decreased to same extent, suggesting an important role of the IFNα/β system on the therapeutic effects exerted by radiotherapy.

Observations of therapeutic effects on tumor lesions outside the irradiation fields prompted us carry out experiments to investigate changes in the immune contexture of the tumor microenvironment due to irradiation. Our 8 Gy fractionated doses on alternate days (scheme in Supplementary Fig. S3A) were applied to mice bearing bilateral MC38-derived tumors. Absolute numbers and density of T lymphocyte subsets were quantitated at the end of the regimen on day +17. We observed that CD4 T-cell and CD8 T-cell numbers were clearly reduced both at the tumor site receiving radiation and, importantly, at the tumor lesion outside the irradiation field. FOXP3⁺ CD4⁺ Treg cells were also reduced at the contralateral site and less clearly so at the irradiated site (Supplementary Fig. S3B and S3C). Myeloid-derived suppressor cells (MDSC) were also evaluated as CD11b (Ly6C or Ly6G) positive cells in the tumors. Our results indicate a trend toward a decrease in G-MDSC in the irradiated tumor, whereas in the nonirradiated sites M-MDSCs were decreased to some extent (Supplementary Fig. S3D and S3E). However, our repeated experiments did not reach statistical significance. The results on radiotherapy-dependent reduction of tumor-infiltrating T lymphocytes were confirmed in mice bearing MC38 tumors in which abscopal effects were noted (Supplementary Fig. S4A–S4C). However, in this case, combined radiotherapy plus combined immunotherapy gave rise to dramatic increases of CD4 and CD8 T cells infiltrating the irradiated and distant tumors (Supplementary Fig. S4C).

Moreover, analyses with a MHC tetramer that detects specific CD8 T cells recognizing the gp70 immunodominant antigen in MC38 tumor cells showed a clear tendency to higher numbers of such tumor-reactive CD8 T lymphocytes in the tumor microenvironment observed in mice undergoing combined treatment (Supplementary Fig. S4D and S4E).

It was also important to assess functionality in terms of the ability of tumor-infiltrating T lymphocytes to produce IFNγ. To this end, we treated mice as shown in Supplementary Fig. S1 (Fig. 6A) and mice were sacrificed on day +16 to monitor tumor-infiltrating T cells. In this setting, experimental groups undergoing combined treatment showed smaller tumor lesions on both sides (Fig. 6B). Intratumoral CD4 and CD8-gated T cells were assessed for the intensity of intracellular IFNγ staining with further stimulation ex vivo with PMA and ionomycin. As seen in Fig. 6C and D, lymphocytes from mice undergoing combined radiotherapy plus immunotherapy attained more intense IFNγ production both in lymphocytes from the irradiated and nonirradiated lesions. Of note, these differences were also observed if the lymphocytes were not stimulated with PMA plus ION (unpublished observations).

One possible explanation of the synergy between radiotherapy and immunostimulatory mAbs was that radiotherapy resulted in a more intense expression of CD137 and/or PD1 on tumor-infiltrating T lymphocytes. Using multicolor immunofluorescence and flow cytometry, an increase of CD137 expression levels was observed among TILs 48 hours after a 20-Gy single dose (Supplementary Fig. S5A). Such an effect was more conspicuous on CD8⁺ TILs, while PD1 expression was preserved at a similar bright level as in the case of the nonirradiated tumor. PD-L1 was expressed on TILs with slight increases related to radiotherapy. On TILs in the contralateral nonirradiated lesion, CD137 levels also increased on CD8 T cells but not on CD4 T cells. Increases in PD1 were also noted but only on CD8 T cells in the contralateral side (Supplementary Fig. S5A and S5B). This was not observed when only a single dose of radiotherapy was given, as after three fractionated 8 Gy doses, an increase in the expression of PD-1 and CD137 was also documented (Supplementary Fig. S5C and S5D).

To study these effects on human tumor samples, we irradiated freshly surgically explanted adenocarcinomas (two gastric carcinomas, five colon cancers, and one chondrosarcoma). Fragments of the excised tumor received 20 Gy, while the other fragments were left without irradiation (mock irradiated). Tumors fragments were subsequently maintained in culture medium and 48 hours later, samples were formalin-fixed and paraffin-embedded for immunohistochemical analysis. As seen in Supplementary Fig. S6A and S6B, there was a clear increase in the percentage of TILs with CD137 and PD1-detectable surface expression, while the total number of T cells remained without noticeable changes. Supplementary Figure S6B shows microphotographs of representative IHC fields of one of the cases. To study these increased expression of CD137 and PD-1 at the single-cell level in a more quantitative fashion, flow cytometry analyses were performed in two cases of colon cancer as freshly surgically excised tumors. Figure 7A shows clear increases in the immunofluorescence intensity for PD-1 and CD137 on viable CD8 or CD4 T cells that was contingent upon irradiation. In Fig. 7B, dot plots of these cases show that CD8 and CD4 T cells frequently coexpressed CD137 and PD-1 after irradiation. Furthermore, multiplex tissue immunofluorescence on a surgical specimen of gastric cancer showed augmented expression of CD137 and PD-L1 upon irradiation, and representative images are shown in Fig. 7C.

All in all, radiotherapy-induced increases in expression of the mAb-targeted receptors PD1 and CD137 are thus likely to account, at least in part, for the combinatorial synergistic effect of radiotherapy and infusion of immunomodulatory mAb targeted to such receptors.

Cancer therapeutics are likely to benefit from the combination of radiotherapy and immunotherapy strategies (13). Our study also strongly indicates that radiotherapy can become a treatment with systemic beneficial antitumor effects (abscopal effects), in addition to the well-known local effects of irradiation. In our hands, radiotherapy modifies the immune microenvironment of distant nonirradiated tumors, but only the addition of immunostimulatory mAbs was able to elicit a meaningful therapeutic effect against nonirradiated tumors and consolidate or enhance the response against the directly irradiated malignant lesions.

To explain the abscopal effects of radiotherapy, many mechanisms have been invoked. Radiotherapy causes vascular inflammation (13) and activation of antigen-presenting DCs (37). Recently, the key contribution of sensing tumor-released DNA by the cytoplasmic pattern recognition receptor STING was found to be crucially important. This mechanism critically causes a local release of type I IFN involving DCs (36). Irradiation is also reported to kill tumor cells showing the hallmarks of immunogenic cell death, as defined by Kroemer and colleagues (38). In this study, we found that abscopal effects are contingent on a DC subset specialized in antigen crosspriming to induce CTLs (20). It is tempting to speculate that such antigen-presenting cell subset is the main mediator of productive tumor antigen presentation to CD8 T cells.

CTL responses and crosspriming are known to be dependent on type I IFN in mice (36, 38). This cytokine system has evolved to raise the alarm upon acute viral infection and is involved in setting in action an optimal immune response for viral clearance. Our findings demonstrate that IFNα/β is critically involved in the abscopal effects of radiotherapy. DNA released from dying tumor cells is probably involved in eliciting IFNα/β via STING (39) and, in turn, IFNα/β may act both on crosspriming DCs (36) and of CD8 T cells (35) to favor, as a necessary factor, the CTL immune response. Strategies aiming at local enhancement of IFNα/β could render radiotherapy-induced tumor cell death more immunogenic as recently shown for chemotherapy (40).

However, immunogenic cell death as induced by radiotherapy only exceptionally offers systemic control of the spread of disease. This could be due to the relative weakness of the immunizing effects or because of concomitantly elicited immunosuppressing factors and mechanisms such as those mediated by TGFβ (41).

Strategies to enhance the antitumor immune effects of radiotherapy have been explored in preclinical models, and results from pioneering clinical research have also been reported (28). For instance, mouse tumor lesions were treated with the TLR7 agonist imiquimod cream (42, 43), or injected with TLR9 CpG agonist nucleotides showing evidence for stronger immunity with the ability to partially tackle distant disease (44, 45). In the clinic, strategies based on combinations of radiotherapy with imiquimod (42) or subcutaneous GM-CSF (46) have been reported with promising proof-of-concept results.

Regarding the optimal combinations of radiotherapy and immunotherapy, several parameters are to be optimized including dose, fractionation, and interval between doses. We chose three fractions of 8 Gy based on published evidence (47) suggesting that this regimen attains better results from the immunologic point of view at least when combining radiotherapy with anti-CTLA-4 mAb (22). However, this issue remains open to debate.

mAbs that tamper with immunoinhibitory receptors (checkpoints; ref. 2) or agonist antibodies to lymphocyte costimulatory receptors (48) have taken the centerstage of oncology drug development. A plethora of clinical trials are exploring their efficacy against multiple malignant diseases, first when used as single agents and then in combinations (49). To date, very little clinical experience exists with combining radiotherapy with immunostimulatory mAbs. Only results from two clinical trials combining radiotherapy and ipilimumab are available for metastatic melanoma (28) and prostate cancer (50). These showed limited efficacy that might be patient subset–specific. As for abscopal effects, evidence is even more scanty although there is reported anecdotal evidence (30).

Systemic effects of anti-PD1 mAb have not yet been reported to potentiate abscopal effects of radiotherapy in patients. In immunogenic mouse models, anti-CD137 (25) and anti-PD1 mAb (21) have been reported to enhance the antitumor effects of radiotherapy. In our case, we report that these antibodies, and especially their combination, can unleash a very potent therapeutic effect against the contralateral tumors (abscopal effects), when both irradiated and nonirradiated tumor lesions were very well established for longer than one week.

In our experiments, potentiation of abscopal effects resulted in long-term survival, comparable with recently reported effects with the combination of anti-PD1 and anti-CTLA-4 mAb in B16F10-bearing mice (28). In this case, the combined mechanism resulted from a reinvigoration of antitumor CTLs that were not repressed by the PD1/PD-L1 axis if combined treatment was given. Our selective depletion experiments point in the same direction.

Our results on abscopal effects contrast with the fact that radiotherapy by itself reduced in our hands the content of T cells in the tumors, although it also slightly reduced the number of MDSCs both in the irradiated tumor and in the contralateral site. However, radiotherapy enhanced IFNγ production on a per cell basis and the level of CD137 and PD1 expression on T cells, in this way making them more amenable to pharmacologic therapeutic costimulation. This was also observed in human tumor fragments irradiated ex vivo. Interestingly, our depletion experiments reveal that only the function of CD8⁺ T cells is an absolute requirement. Moreover, at least a subset of CD4⁺ cells seems to be operating in detriment of efficacy. These are likely to be Treg cells. Importantly, at the single-cell level there are tumor-infiltrating lymphocytes that coexpress the CD137 and PD-1 receptors, arguing in favor of a double hit by the immunostimulatory mAbs on single T-cell basis.

In our hands, radiotherapy plus immunostimulatory mAbs dramatically enhance the T-cell infiltrate after 8 days of combined treatment with evidence for more CD8 T cells recognizing the gp70 tumor antigen in the MC38 tumor model. These tetramer-positive cells are CD137⁺ and PD-1⁺ in keeping with previous reports showing that CD137⁺ cells in human melanomas tend to be specific for tumor neoantigens (51).

Overall, our data strongly support initiation of clinical trials testing anti-CD137 mAb in combination with PD1/PD-L1 blockade together with concomitant irradiation of some of the tumor metastatic sites, in search of a way powerful to make the most of these novel immunotherapies.

M. Jure-Kunkel has ownership interest (including patents) in Bristol-Myers Squibb. I. Melero reports receiving a commercial research grant from Pfizer and is a consultant/advisory board member for Bristol-Myers Squibb, Boehringer Ingelheim, Roche-Genentech, Incyte, and Astra Zeneca. No potential conflicts of interest were disclosed by the other authors.

Conception and design: M.E. Rodriguez-Ruiz, I. Rodriguez, M.A. Aznar, I. Melero

Development of methodology: M.E. Rodriguez-Ruiz, I. Rodriguez, B. Barbes, S. Labiano, A. Azpilikueta, E. Bolaños, A. Rouzaut

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): M.E. Rodriguez-Ruiz, I. Rodriguez, J.L. Solorzano, J.L. Perez-Gracia, A. Rouzaut, K.A. Schalper

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): M.E. Rodriguez-Ruiz, I. Rodriguez, J.L. Perez-Gracia, A. Azpilikueta, A.R. Sanchez-Paulete, A. Rouzaut, K.A. Schalper, I. Melero

Writing, review, and/or revision of the manuscript: M.E. Rodriguez-Ruiz, I. Rodriguez, B. Barbes, J.L. Perez-Gracia, S. Labiano, M.F. Sanmamed, A. Azpilikueta, E. Bolaños, A.R. Sanchez-Paulete, M. Jure-Kunkel, I. Melero

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): M.E. Rodriguez-Ruiz, S. Garasa, B. Barbes, J.L. Solorzano, M.F. Sanmamed

Study supervision: M.E. Rodriguez-Ruiz, I. Melero

We acknowledge generous help by Drs. Martinez-Monge, Aristu, and Gil-Bazo from the Department of Oncology at CUN. We are also grateful for the advice from Drs. Lozano, Echeveste, and Idoate from the pathology department at CUN. We are grateful for Sciencific discussion with Drs. Mariano Ponz, David Sancho, Nicola Tinari, and Antonio Rullán. Excellent dosimetry by Arantza Zubiria and dedicated animal care by Eneko Elizalde are also acknowledged.

This work was financially supported by grants from MICINN (SAF2011-22831 and SAF2014-52361-R). I. Melero was also funded by the Departamento de Salud del Gobierno de Navarra, Redes temáticas de investigación cooperativa RETICC, European Commission VII Framework and Horizon 2020 programs (AICR and PROCROP), SUDOE-IMMUNONET, Fundación de la Asociación Española Contra el Cáncer (AECC), Fundación BBVA and Fundación Caja Navarra. M.E. Rodriguez-Ruiz receives a Rio Hortega contract from ISCIII. S. Labiano is recipient of predoctoral scholarship from MICINN.

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.