CCL2/CCR2-Dependent Recruitment of Functional Antigen-Presenting Cells into Tumors upon Chemotherapy

Accumulating evidence indicates that the antineoplastic effects of anthracyclines largely depend on the induction of a T-cell–mediated tumor-specific immune response. In patients with breast cancer subjected to one cycle of anthracycline-based chemotherapy, a high ratio of CD8⁺ cytotoxic versus FOXP3⁺ regulatory T cells within the tumor predicts the advent of a complete pathological response after 6 cycles of neoadjuvant treatment (1, 2). In mice, the therapeutic efficacy of anthracyclines is abolished when the extravasation of CD11b⁺ myeloid cells is blocked (3), when several distinct T-cell subsets (i.e., α/β or γ/δ T cells, CD8⁺ cells) are eliminated (4, 5) as well as when essential cytokines such as interferon (IFN)-γ, interleukin (IL)-1β, and IL-17 are neutralized (5–7). Thus, the therapeutic efficacy of anthracyclines requires specific cytokines to orchestrate a complex immune response that involves antigen-presenting cells (APC) exhibiting a dendritic cell–like (i.e., CD11b⁺CD11c⁺Ly6C^highLy6G⁻MHCII⁺) phenotype and several T-cell subsets (3, 6). The actual source of such intratumoral APCs has not yet been elucidated.

One of the most critical processes for the initiation of an immune response against cancer cells succumbing to anthracyclines is the secretion of ATP (8). Extracellular ATP indeed operates not only as a potent chemotactic signal (by binding to purinergic P2Y2 receptors; ref. 9), but also as a proinflammatory agent, stimulating the activation of the NLRP3 inflammasome (by binding to purinergic P2RX7 receptors) and hence the release of IL-1β (10). In line with this notion, ATP is essential for the accumulation of monocytic precursors within the tumor bed as well as for their local differentiation into dendritic cell–like APCs (3). Thus, shortly (24–72 hours) after the administration of immunogenic chemotherapy (11–13), ATP levels increase in the tumor interstitium as a consequence of the apoptotic demise of cancer cells. Of note, autophagy is required for such an immunogenic cell death–associated release of ATP to be optimal (8). Myeloid and lymphoid cells indeed are not recruited by anthracycline-treated autophagy-deficient tumors, correlating with the absence of anticancer immune responses and hence with therapeutic failure (8). Along similar lines, cancers can be rendered virtually insensitive to the antineoplastic effects of anthracyclines by the overexpression of CD39, a cell surface ATP-degrading enzyme (8). Besides ATP, other chemotactic factors including multiple chemokines are known to participate into the recruitment of innate and cognate immune effectors to inflammation sites (14). However, the precise contribution of chemokines to anticancer immune responses elicited by immunogenic chemotherapeutics has not yet been established.

Driven by these premises and incognita, we investigated whether the APCs that accumulate within neoplastic lesions in response to anthracycline-based chemotherapy derive from local precursors or are recruited from external sources and to which extent chemokines regulate their biogenesis. We found that dendritic cell–like APCs are freshly recruited into the tumor bed in response to immunogenic chemotherapy and that the host-derived chemokine Ccl2 plays a major, ATP-independent role in this process. In line with this notion, Ccl2 and its major receptor, Ccr2 (also known as CD192), turned out to be fundamental for the antineoplastic effects of anthracycline-based chemotherapy in mice.

Mouse fibrosarcoma MCA205 cells (H2^b), their ovalbumin (OVA)-expressing derivatives (a generous gift from Dr. A.D. Weinberg, Portland Providence Medical Centre, Portland, OR), and AT3 mammary carcinoma cells (H2^b) were cultured in GlutaMAX-I–containing RPMI 1640 medium supplemented with 10% fetal bovine serum (FBS), 1 mmol/L sodium pyruvate, 10 mmol/L HEPES buffer, 1 mmol/L nonessential amino acids, 100 units/mL penicillin G sodium, and 100 μg/mL streptomycin sulfate. Unless otherwise indicated, authenticated wild-type cells were obtained from American Type Culture Collection, immediately amplified to constitute liquid nitrogen stocks and (upon thawing) never passaged for more than 1 month before use in experimental determinations. Cell culture products were from Gibco-Life Technologies.

Mice were housed in specific pathogen-free conditions at either the Institut Gustave Roussy (IGR, Villejuif, France), the Peter MacCallum Cancer Center (East Melbourne, Australia), or the QIMR Berghofer Medical Research Institute (Herston, Australia). All animal experiments were performed in accordance with the Federation of European Laboratory Animal Science Association guidelines and were approved either by the IGR Ethics Committee (CEEA IRCIV/IGR No. 26, registered with the French Ministry of Research) or by the Animal Experimentation Ethics Committee of Peter MacCallum Cancer Center and QIMR Berghofer Medical Research Institute. Female C57BL/6 (H2^b) mice (aged between 6 and 8 weeks) were purchased from Janvier, OVA^257–264-specific T-cell receptor (TCR) transgenic OT1 mice (H2^b) and congenic mice harboring an alternative CD45/Ly-5 alloantigen (Ly-5.1) were obtained from Charles River, whereas Ccl2^−/− and Ccr2^−/− mice were bred at the Peter MacCallum Cancer Center or QIMR Berghofer Medical Research Institute. Mice were maintained under controlled light cycle (12 hours lights ON, 12 hours lights OFF), allowed food and water ad libitum, and were invariably used for experiments between the age of 7 and 16 weeks.

Mice were subjected to skin decontamination with 10% povidone–iodine (commercial Betadine solution) and anesthesia was maintained by continuous mask inhalation of isoflurane flow during surgery. A narrow incision (3 mm) was performed on the thigh, tumor-draining lymph nodes (DLN) were removed using a dissecting forceps, and the incision was eventually closed with sterile skin closure clips. Sham-operated mice were treated with a similar procedure (but were not subjected to lymphadenectomy), and the duration of surgery was standardized between the 2 groups. After surgery, mice were placed on a heater plate set at 37°C until recovery.

For establishing syngeneic solid tumors, wild-type, Ccl2^−/−, or Ccr2^−/− CB57BL/6 mice were inoculated with 5 × 10⁵ MCA205, OVA-expressing MCA205, or AT3 cells subcutaneously. Thereafter, tumor surface (longest diameter multiplied by its perpendicular diameter) was routinely monitored by means of a common caliper, and when it reached 25 to 45 mm², mice received either 2.9 mg/kg doxorubicin intratumorally (i.t.; in 50 μL PBS) or an equivalent volume of solvent.

CD45.2 CB57BL/6 mice were anesthetized under continuous isoflurane flow and a small incision was made on the skin near the thigh. MCA205 fibrosarcomas (surface = ∼35–50 mm²) freshly harvested from CD45.1 mice were then implanted as a whole subcutaneously, with the help of grooved probes and forceps. Eventually, the incision was closed with sterile skin closure clips.

Freshly recovered tumors or spleens were cut into small pieces (using surgical scissors) in digestion buffer (serum-free GlutaMAX-I–containing RPMI 1640 medium supplemented with 0.4 Wünsch units/mL Liberase TL, from Roche, and 200 U/mL DNase I, from Calbiochem), and then transferred to 24-well plates for enzymatic dissociation (30 minutes, 37 °C). Upon washing with PBS, single-cell suspensions were obtained by filtering samples through a 70-μm cell strainer. Lymph nodes were gently squeezed with dissecting forceps in digestion buffer to obtain single-cell suspensions. Bone marrow cells were recovered by repeatedly flushing the femoral or tibial bone marrow cavities with PBS using 26G needles. The peripheral blood was harvested from the inner canthus of the eye with heparin-coated Pasteur pipettes. Red blood cells were removed with the RBC Lysis Buffer (BioLegend), following the manufacturer's instructions.

For cell surface immunostaining, primary antibodies targeting the following proteins were used (at 2 μg/mL unless indicated, staining at 4°C for 20 minutes): CD45.1 (A20), CD45.2 (104), TCR Vβ 5.1/5.2 (MR9-4), and CD69 (H1.2F3), all from eBioscience; CD11b (M1/70), CD11c (N418), and Ly6C (AL-21), all from BD Pharmingen; B220 (RA3-6B2), I-A/I-E (M5/114.15.2; used at 0.2 μg/mL), Ly6G (1A8), and CD103 (2E7), all from BioLegend; Ly-6B.2 (7/4), from AbD Serotec; and CCR2 (MC-21), a generous gift from Dr. Matthias Mack (University of Regensburg, Regensburg, Germany). Cytofluorometric assessments were performed on a BD LSR II Flow Cytometer (BD Biosciences) and data were analyzed by means of the FlowJo software package (Tree Star, Inc.). Tumor cell suspensions were sorted with a MoFlo XDP (Beckman-Coulter) or a FACSAria (BD Biosciences) cell sorter.

For testing antigen-specific T-cell proliferation in situ, splenic OT1 cells were stained with 5 μmol/L CFSE (Molecular Probes-Life Technologies) for 10 minutes at 37°C, followed by quenching of the extra dye with an equal volume of 20% FBS in PBS. These cells were then inoculated intravenously into OVA-expressing MCA205 fibrosarcoma-bearing mice (7 × 10⁶ cells/mouse), 4 days after tumor cell inoculation (day-5). Chemotherapy was administered on day 0 and mice were subjected to lymphadenectomy or sham surgery on day 1. OT1 cell proliferation in tumors and DLNs was assessed by CFSE dilution on day 3.

Alternatively, CD45.2⁺ OT1 cells were inoculated intravenously into CD45.1 congenic mice (1 × 10⁶ cells/mouse) 1 day before the subcutaneous inoculation of OVA-expressing MCA205 cells. Doxorubicin-based chemotherapy or PBS was administered i.t. 7 days later. Tumors, DLNs, contralateral lymph nodes, and spleens were harvested 60 hours after chemotherapy to analyze CD69 expression on OT1 cells.

Total RNA was isolated from established MCA205 fibrosarcomas 2 and 7 days after the i.t. administration of doxorubicin or PBS. Biotin-labeled cRNA was synthesized using the Illumina TotalPrep RNA Amplification Kit (Applied Biosystems). cRNA was hybridized to Mouse WG-6 V.2 Expression BeadChips (Illumina) and scanned with a BeadStation 500× (Illumina). Raw data collection from Illumina BeadChips was performed using the BeadStudio software v. 3.1.1.0 (Illumina). Data were further analyzed by on an R-based informatics platform by means of packages from the Bioconductor project (15). Upon quantile normalization, genes were defined as differentially expressed when associated with absolute changes in expression levels (absolute fold change) >1.7. This transcriptomic dataset has been deposited in the Gene Expression Omnibus database (accession number GSE46275; ref. 16).

Total RNA extraction and genomic DNA removal were performed by means of the RNeasy Micro Kit (Qiagen), following the manufacturer's instructions. Total RNA (0.5–2 μg per sample) was then reverse transcribed into cDNA with the SuperScript III Reverse Transcriptase (Life Technologies), in the presence of random primers, the RNasin Plus RNase Inhibitor (Promega) and the Deoxynucleoside Triphosphate Set, PCR grade (Roche). The expression of genes coding for inflammatory cytokines, chemokines, M1 or M2 macrophage-related factors, and protein involved in MHC class I or class II antigen presentation was analyzed with TaqMan Gene Expression Assays using the Universal Master Mix II (with UNG) on a StepOnePlus Real-Time PCR System (everything from Applied Biosystems). Quantitative RT-PCR data were invariably normalized to the expression levels of the housekeeping gene peptidylprolyl isomerase A (Ppia).

Freshly harvested tumors were processed for the immunofluorescence microscopy–assisted quantification of caspase-3 activation and infiltration by CD11c⁺CD86⁺ cells as previously reported (3). Thereafter, 5-μm-thick cryosections were stained with the following antibodies or antibody conjugates: CD11c-AlexaFluor 488 (N418, from eBioscience), CD86-AlexaFluor 647 (GL-1, from BioLegend), and anti-cleaved caspase-3 (Asp175, from Cell Signaling Technology), followed by AlexaFluor 568 goat anti-rabbit immunoglobulin G (H+L; from Molecular Probes-Life Technologies). Nuclear counterstaining was achieved by incubating slides in 10 μmol/L Hoechst 33342 (Molecular Probes-Life Technologies) for 5 minutes before mounting. Finally, 20 to 40 view-fields were acquired for each slide by means of a confocal SPE microscope (Leica Microsystems GmbH), and 8 to 10 images were randomly picked for statistical analysis.

Unless otherwise indicated, results are expressed as means ± SEM or means ± SD, as appropriate, of n = 3 parallel assessments. All experiments were repeated at least twice, yielding similar results. Normal distributions were compared by unpaired, 2-tailed Student t tests; tumor growth curves with Mann–Whitney U tests. Statistical analyses were performed with Prism 5 (GraphPad) or Excel 2007 (Microsoft). P-values < 0.05 were considered as statistically significant.

Shortly (38 hours) after the i.t. administration of doxorubicin, a prototypical anthracycline, MCA205 fibrosarcomas established in immunocompetent C57BL/6 mice contain an elevated percentage of CD11b⁺Ly6C^highLy6G⁻ and CD11c⁺MHCII⁺ cells, as compared with vehicle-injected lesions (Fig. 1A and Supplementary Fig. S1; ref. 3). Besides becoming more abundant with time, this subset of myeloid cells overexpresses genes involved in MHC class I– and II–restricted antigen presentation (Fig. 1B and C), as well as genes coding for multiple cytokines, including IL-1β, TNF-α, IL-10, and IL-12 (Fig. 1D). However, these cells do not manifest an obvious polarization of gene expression toward a profile associated with M1 or M2 macrophages (Supplementary Fig. S2). At this time point, no increase in CD11b⁺Ly6C^highLy6G⁻, CD11b⁺, CD11b⁺Ly6C^highLy6G⁻, and CD11b⁺Ly6G⁺ cells was observed in the DLNs of mice bearing doxorubicin-treated MCA205 fibrosarcomas. Conversely, CD11b⁺Ly6C^highLy6G⁻ and CD11b⁺Ly6G⁺ neutrophil-like cells were found to accumulate in the bone marrow and spleen of these animals, respectively (Supplementary Fig. S1). The biological significance of this latter observation remains obscure.

In mice carrying OVA-expressing MCA205 fibrosarcomas, doxorubicin-based chemotherapy increased the i.t. proliferation of adoptively transferred CFSE-labeled OT1 cells expressing a transgenic TCR specific for the MHC class I H2K^b-restricted OVA-derived peptide SIINFEKL (Fig. 1E and Supplementary Fig. S3). Sixty hours after the administration of doxorubicin, the percentage of activated (CD69⁺) OVA-specific OT1 cells was increased in the tumor bed, but not in the spleen, DLNs, and contralateral lymph node (Supplementary Fig. S3). Of note, the chemotherapy-induced proliferation of i.t. OT1 cells was maintained upon the surgical ablation of DLNs (Fig. 1E). This is in line with previous findings from our group demonstrating that anticancer immune responses elicited by doxorubicin-based chemotherapy are insensitive to the administration of a lymphotoxin β decoy that impedes the formation of DLNs (3). Moreover, the proliferation of adoptively transferred CFSE-labeled OT1 cells in DLNs was not affected by chemotherapy (Supplementary Fig. S4). Taken together, these findings indicate that the presentation of tumor antigens elicited by immunogenic chemotherapy occurs within the tumor bed.

Theoretically, the APCs that accumulate within anthracycline-treated neoplastic lesions might either originate from preexisting precursors that differentiate locally or be recruited de novo from external sources. To discriminate between these possibilities, we established MCA205 fibrosarcomas in CD45.1 mice, treated them with vehicle or doxorubicin, and (1 day later) transplanted them into isogenic CD45.2 C57BL/6 recipients. This allowed us to quantify the intratumoral abundance of sessile (CD45.1⁺) versus freshly recruited (CD45.2⁺) leukocytes, more specifically CD11c⁺MHCII⁺ cells. In this setting, we observed that most tumor-sessile leukocytes are replaced by their recipient-derived counterparts within the first 4 days after transplantation, irrespective of the administration of vehicle or doxorubicin (Fig. 2A–C). Globally, the ratio of recipient-derived (CD45.2⁺) versus donor-derived (CD45.1⁺) cells was not altered by chemotherapy (Fig. 2E). However, in doxorubicin-treated tumors, the ratio of CD45.2⁺ versus CD45.1⁺ CD11c⁺MHCII⁺ dendritic cell–like cells was higher than in PBS-receiving lesions (Fig. 2E), supporting the notion that anthracyclines trigger the recruitment of dendritic cell precursors rather than the mere differentiation of preexisting myeloid cells in situ. Accordingly, CD11b⁺Ly6C^highLy6G⁻ cells accumulating within neoplastic lesions in response to anthracyclines phenotypically resembled a cell subset present in the bone marrow (both were also CD11c⁺F4/80^int7/4⁺CD103⁺; Supplementary Fig. S1). Moreover, the frequency of several distinct circulating CD11b⁺ cell subsets fluctuated in response to anthracyclines (Supplementary Fig. S5), in concert with the possibility that at least a fraction of CD11b⁺Ly6C^highLy6G⁻ cells found in the tumor bed upon immunogenic chemotherapy is blood-borne.

ATP is known to operate as a chemotactic and danger signal in the context of immunogenic chemotherapy (3, 8). In line with this notion, MCA205 fibrosarcoma cells that had been engineered to overexpress the ecto-ATPase CD39 on their surface generated lesions in immunocompetent mice that were unable to recruit myeloid, especially CD11b⁺Ly6C^highLy6G⁻ cells in response to doxorubicin (Fig. 3A). To identify additional factors that are involved in the chemotherapy-induced recruitment of antigen-presenting leukocytes into the tumor bed, we performed microarray-based transcriptomic analyses. In this setting, several chemokine-encoding genes were found to be upregulated (at the mRNA level) 2 days after the i.t. administration of doxorubicin (Fig. 3B). In particular, anthracycline-based chemotherapy turned out to promote the expression of Ccl2 and Ccl7 (two chemokines involved in myeloid cell chemotaxis; refs. 17 and 18) by intratumoral CD11b⁺Ly6C^highLy6G⁻ cells. This phenomenon was not influenced by CD39 and hence must be independent from the immunogenic cell death–associated secretion of ATP (Fig. 3C). The levels of Cx3cl1 in CD11b⁺Ly6C^highLy6G⁻ cells were not influenced by anthracycline-based chemotherapy, yet were a priori more elevated in the context of CD39 overexpression (Supplementary Fig. S6). Conversely, extracellular ATP seemed to have a positive impact on the expression levels of Cxcl1, Cxcl9, Cxcl10, Cxcl11, and Ccl5 (Fig. S6). Of note, both CD45⁺ leukocytes and CD45⁻ tumor cells responded to doxorubicin by producing Ccl2 and Ccl7, yet only the former did so in a statistically significant fashion (Fig. 3D). Moreover, tumors growing in Ccl2^−/− mice failed to recruit CD11b⁺Ly6C^highLy6G⁻ cells in response to doxorubicin (Fig. 3E). Interestingly, several CD11b⁺ cell subsets including CD11b⁺Ly6C^highLy6G⁻ leukocytes were found to express the Ccl2 receptor Ccr2 (which also binds Ccl7), irrespective of their previous exposure to doxorubicin. In this context, the highest levels of Ccr2 were expressed by CD11b⁺ cells isolated from neoplastic lesions (as opposed to the circulation, spleen, bone marrow and DLNs; Fig. 3F). Conversely, we failed to detect significant amounts of Ccr2 on the surface of malignant cells, be them recovered from PBS- or doxorubicin-treated tumors (Supplementary Fig. S7). Taken together, these observations suggest that Ccl2 produced by host leukocytes plays a prominent role in recruitment of CD11b⁺Ly6C^highLy6G⁻ cells into the tumor bed upon anthracycline-based chemotherapy.

The aforementioned findings indicate that Ccl2 is required for the doxorubicin-stimulated intratumoral accumulation of CD11b⁺Ly6C^highLy6G⁻ dendritic cell–like cells, including the APCs that are essential for the optimal efficacy of anthracycline-based chemotherapy (3). MCA205 fibrosarcomas developing in Ccl2^−/− and Ccr2^−/− mice exhibited a similar apoptotic response to doxorubicin than tumors growing in wild-type hosts (Fig. 4A and B). However, the absence of Ccl2 or Ccr2 in the host dampened the accumulation of CD11c⁺CD86⁺ dendritic cells in tumor areas enriched of dead cells (Fig. 4A and C). In line with this notion, Ccl2^−/− and Ccr2^−/− mice bearing MCA205 fibrosarcomas were less sensitive to doxorubicin-based chemotherapy than their wild-type counterparts (Fig. 4D). Similarly, AT3 breast carcinomas only responded to the i.t. administration of doxorubicin in mice expressing both Ccl2 and Ccr2. Conversely, the growth of AT3 cancers in Ccl2^−/− or Ccr2^−/− hosts was poorly affected by anthracycline-based chemotherapy (Fig. 4E). Altogether, these data indicate that the CCL2/CCR2 signaling axis is essential for optimal therapeutic responses to anthracyclines, at least in mice.

Taken together, the results presented here indicate that anthracycline-based chemotherapy promotes the intratumoral accumulation of (presumably blood-borne) myeloid cells, including cells with a dendritic cell–like phenotype that mediate antigen presentation. Immunogenic chemotherapy seems to cause major shifts in the composition of myeloid cell subsets that affect the tumor microenvironment but not lymphoid organs. In addition, available evidence suggests that neoplastic lesions contain all the cellular and humoral elements that are required to mount a local anticancer immune response even in the absence of DLNs, hence constituting a sort of minimal lymphoid organ. Independently from the accumulation of ATP in the tumor microenvironment, anthracyclines stimulated distinct types of tumor-associated cells, but mainly CD45⁺ myeloid cells, to produce a series of chemokines including the 2 Ccr2 ligands Ccl2 and Ccl7. Tumors established in Ccl2^−/− hosts failed to recruit dendritic cell–like APCs and exhibited a suboptimal therapeutic response to upon anthracycline-based chemotherapy. Along similar lines, tumor growing in Ccr2^−/− did not respond to anthracyclines as efficiently as tumors implanted in Ccl2- and Ccr2-proficient mice. These findings support the importance of host-derived Ccl2 acting on host Ccr2 for the anticancer immune response elicited by immunogenic cell death.

The CCL2/CCR2 system has been previously implicated in the accumulation of tumor-associated macrophages and myeloid-derived suppressor cells within neoplastic lesions, hence supporting tumor growth via proinflammatory and immunosuppressive mechanisms (19). In addition, an autocrine CCL2-dependent signaling pathway has been suggested to promote the survival and mobility of cancer cells (20), whereas tumor cell–derived CCL2 has been shown to stimulate metastasis (21) and angiogenesis (22). In line with these observations, CCL2- or CCR2-targeting agents have been shown to improve the antineoplastic activity of distinct immunotherapeutic regimens (23, 24) as well as the efficacy of gemcitabine- and cisplatin-based chemotherapy (25, 26). Based on these observations, several clinical trials have recently been launched to evaluate the safety and therapeutic potential of CCL2- or CCR2-blocking antibodies, either as standalone therapeutic interventions or combined with conventional chemotherapy, in patients with cancer (source http://clinicaltrials.gov).

Here, we investigated the impact of the CCL2/CCR2 system in a rather peculiar context. First, the growth of the tumor models that we chose was unaffected by the absence or presence of Ccl2 or Ccr2 in the host (Fig. 4), implying that none of the aforementioned protumorigenic functions of CCL2/CCR2 operate in this scenario. Second, we focused our attention on the therapeutic immune response elicited by the immunogenic cell death inducer doxorubicin. In this experimental setting, the negative impact of CCL2/CCR2 ablation on the therapeutic activity of doxorubicin correlated with a reduced intratumoral recruitment of dendritic cell–like APCs. It should be noted that the CCL2/CCR2 system mediates the accumulation of myeloid-derived dendritic cells during allergic (27) and autoimmune inflammation (28), as well as at sites of parasitic (29), fungal (30), and mycobacterial infection (31). Thus, the inhibition of the CCL2/CCR2 signaling axis may blunt a wide range of inflammatory and immune reactions, including those required for the control of infectious pathogens and (pre)malignant cells.

Based on these considerations, the use of CCL2- or CCR2-blocking agents in patients with cancer must be carefully considered on a personalized basis. This may indeed constitute a therapeutic option to subvert CCL2-dependent metastatic spread, angiogenesis, and immunosuppression. However, blocking the CCL2/CCR2 system should be avoided in conditions in which the recruitment of dendritic cell–like APCs into the tumor bed constitutes a limiting step for the elicitation of therapeutic anticancer immune responses.

No potential conflicts of interest were disclosed.

Conception and design: Y. Ma, L. Zitvogel, M.J. Smyth, G. Kroemer

Development of methodology: Y. Ma, H. Yang, L. Aymeric, O. Kepp

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): Y. Ma, S.R. Mattarollo, S. Adjemian, H. Yang, L. Aymeric, D. Hannani, J.P.P. Catani, M.W.L. Teng, Y. Wang, A. Sistigu, J.L. Schultze, M.J. Smyth

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): Y. Ma, D. Hannani, J.L. Schultze, G. Stoll, L. Zitvogel, M.J. Smyth

Writing, review, and/or revision of the manuscript: Y. Ma, S.R. Mattarollo, Y. Wang, J.L. Schultze, L. Galluzzi, L. Zitvogel, M.J. Smyth

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): Y. Ma, S. Adjemian, J.P.P. Catani, H. Duret

Study supervision: L. Zitvogel, M.J. Smyth, G. Kroemer

We thank our colleagues from the animal facility of the IGR. We thank Dr. A.D. Weinberg for OVA-expressing MCA205 cells.

G. Kroemer and L. Zitvogel were supported by the Ligue Nationale contre le Cancer (Equipes labellisées), AXA Chair for Longevity Research, Agence Nationale pour la Recherche (ANR AUTOPH, ANR Emergence), Assocation pour la Recherche sur le Cancer, European Research Council, Fondation pour la Recherche Médicale (FRM), Institut National du Cancer (INCa), Fondation de France, Cancéropôle Ile-de-France, Fondation Bettencourt-Schueller, the LabEx Immuno-Oncology, and the Paris Alliance for Cancer Research Institutes. Y. Ma and H. Yang were supported by China Scholarship Council (CSC). S. Adjemian was supported by the Ligue Nationale contre le Cancer. J.P.P. Catani was supported by Coordenação de Aperfeicoamento Pessoal de Nível Superior (CAPES/Brazil). S.R. Mattarollo, H. Duret, and M.J. Smyth were supported by a National Health and Medical Research Council (NH&MRC) Australia Fellowship and Program Grant and a grant from the Victorian Cancer Agency. M.W.L. Teng was supported by a NH&MRC CDF1 award. J.L. Schultze is a member of the Excellence Cluster ImmunoSensation and supported by the German Research Foundation (Sonderforschungsbereich 704, SFB 832, INST 217/575-1).

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.