Abstract
One particular strategy to render anticancer therapies efficient consists of converting the patient’s own tumor cells into therapeutic vaccines, via the induction of immunogenic cell death (ICD). One of the hallmarks of ICD dwells in the active release of ATP by cells committed to undergo, but not yet having succumbed to, apoptosis. We observed that the knockdown of essential autophagy-related genes (ATG3, ATG5, ATG7 and BECN1) abolishes the pre-apoptotic secretion of ATP by several human and murine cancer cell lines undergoing ICD. Accordingly, autophagy-competent, but not autophagy-deficient, tumor cells treated with ICD inducers in vitro could induce a tumor-specific immune response in vivo. Cancer cell lines stably depleted of ATG5 or ATG7 normally generate tumors in vivo, and such autophagy-deficient neoplasms, upon systemic treatment with ICD inducers, exhibit the same levels of apoptosis (as monitored by nuclear shrinkage and caspase-3 activation) and necrosis (as determined by following the kinetics of HMGB1 release) as their autophagy-proficient counterparts. However, autophagy-incompetent cancers fail to release ATP, to recruit immune effectors into the tumor bed and to respond to chemotherapy in conditions in which autophagy-competent tumors do so. The intratumoral administration of ecto-ATPase inhibitors increases extracellular ATP concentrations, re-establishes the therapy-induced recruitment of dendritic cells and T cells into the tumor bed, and restores the chemotherapeutic response of autophagy-deficient cancers. Altogether, these results suggest that autophagy-incompetent tumor cells escape from chemotherapy-induced (and perhaps natural?) immunosurveillance because they are unable to release ATP.
Anticancer chemotherapeutics are commonly conceived as antibiotics that selectively eliminate malignant cells. However, beyond their cancer cell-autonomous cytotoxic effects, some anticancer agents including anthracyclins and oxaliplatin can trigger a particular type of cellular demise, which we termed “immunogenic cell death,” mobilizing the host’s immune system against residual tumor cells. Accumulating evidence confirms that chemotherapy-induced anticancer immune responses have a major impact on the efficacy of anthracyclin and oxaliplatin-based therapies, both in experimental mouse models and in patients.
ICD differs from conventional apoptosis in that it is coupled to three unique molecular determinants, namely (1) the early, pre-apoptotic exposure of calreticulin on the outer leaflet of the plasma membrane [where it serves as an engulfment signal for antigen presenting cells, in particular dendritic cells (DCs)]; (2) the late, post-apoptotic release of the high mobility group box 1 (HMGB1) protein (which binds to toll-like receptors and stimulates the presentation of tumor cell antigens by DCs); and (3) the active secretion of ATP during apoptosis. By binding to purinergic receptors, extracellular ATP stimulates the recruitment of immune effectors in the proximity of dying cells and favors the activation of the NLRP3 inflammasome, in particular in DCs, allowing for the caspase-1-mediated proteolytic maturation of interleukin-1β, a cytokine that is required for tumor lysis ().
Figure 1. Involvement of autophagy in cancer therapy-relevant immunogenic cell death (ICD). In response to a selected panel of chemotherapeutics (including anthracyclins and oxaliplatin), cancer cells manifest two premortem stress responses [endoplasmic reticulum, (ER) stress and autophagy]. ER stress is required for the pre-apoptotic exposure of calreticulin (CRT) on the cell surface, whereas autophagy is necessary for the active secretion of ATP. Please note that autophagy alone cannot trigger ATP release, which also relies on the apoptosis-related activation of caspases. In this setting, secondary necrosis accounts for the release of HMGB1. CRT, ATP and HMGB1 act on specific receptors on the surface of antigen-presenting cells, mainly dendritic cells (DCs), to stimulate immune-relevant processes: engulfment, which depends on the interaction between CRT and a hitherto unidentified receptor; recruitment of immune effectors into the tumor bed and activation of the NLRP3 inflammasome, allowing for the release of mature interleukin-1β (IL-1β), both of which rely on the ATP-mediated activation of purinergic receptors; and tumor antigen presentation, which is stimulated by the HMGB1-mediated activation of Toll-like receptor 4 (TLR4). These processes ultimately result in the priming or re-activation of tumor antigen-specific T cells that can control the residual disease.
![Figure 1. Involvement of autophagy in cancer therapy-relevant immunogenic cell death (ICD). In response to a selected panel of chemotherapeutics (including anthracyclins and oxaliplatin), cancer cells manifest two premortem stress responses [endoplasmic reticulum, (ER) stress and autophagy]. ER stress is required for the pre-apoptotic exposure of calreticulin (CRT) on the cell surface, whereas autophagy is necessary for the active secretion of ATP. Please note that autophagy alone cannot trigger ATP release, which also relies on the apoptosis-related activation of caspases. In this setting, secondary necrosis accounts for the release of HMGB1. CRT, ATP and HMGB1 act on specific receptors on the surface of antigen-presenting cells, mainly dendritic cells (DCs), to stimulate immune-relevant processes: engulfment, which depends on the interaction between CRT and a hitherto unidentified receptor; recruitment of immune effectors into the tumor bed and activation of the NLRP3 inflammasome, allowing for the release of mature interleukin-1β (IL-1β), both of which rely on the ATP-mediated activation of purinergic receptors; and tumor antigen presentation, which is stimulated by the HMGB1-mediated activation of Toll-like receptor 4 (TLR4). These processes ultimately result in the priming or re-activation of tumor antigen-specific T cells that can control the residual disease.](/cms/asset/90e5797a-e669-4f55-831e-73349cf6b48b/kaup_a_10919009_f0001.gif)
Autophagy is frequently disabled in cancer, in particular during early oncogenesis, due to the genetic or functional inactivation of multiple oncosuppressor proteins that also exert pro-autophagic roles (including BECN1, DAPK1, PTEN, TSC1, TSC2, LKB1/STK11 and several pro-apoptotic “BH3-only” proteins from the BCL-2 family) or to the hyperactivation of oncogene products that also operate as autophagy inhibitors [including phosphatidylinositol 3-kinase (PtdIns3K), AKT1, and anti-apoptotic members of the BCL-2 family]. Based on these considerations, we evaluated whether disabled autophagy might interfere with ICD. We found indeed that the suppression of autophagy (by the transient or permanent depletion of essential gene products) strongly reduces the capacity of cancer cells and mouse embryonic fibroblasts to release ATP in response to ICD inducers. In contrast, autophagy has no impact on the surface exposure of calreticulin or on the release of HMGB1 (). Of note, autophagy induction by pharmacological stimuli such as rapamycin is not sufficient to trigger ATP release. In line with this observation, we found that the autophagy-dependent secretion of ATP only occurs in the context of imminent cell death, when caspases are activated, shortly before phosphatidylserine is exposed on the cell surface and the plasma membrane eventually ruptures. The exact mechanisms through which autophagy contributes to ATP release remain elusive, although it may be speculated that the premortem activation of autophagy might preserve a bioenergetic status in which, in spite of massive cellular damage and futile (energy consuming) attempts of repair, high intracellular ATP levels are maintained.
Having established that autophagy is essential for the release of ATP by dying cells, we determined the impact of autophagy inhibition on ICD using two distinct experimental systems. First, we exposed autophagy-proficient or -deficient colon carcinoma or fibrosarcoma cells to ICD inducers in vitro, in standardized conditions (optimized so as to induce 70 ± 10% of cell death, as assessed by AnnexinV/propidium iodide co-staining), washed them (to remove ICD inducers) and injected them subcutaneously into immunocompetent mice. The capacity of these cells to elicit an anticancer immune response was tested by re-challenging mice with live tumor cells of the same type, which were injected one week later into the contralateral flank. All manipulations aimed at reducing the autophagic flux (such as the knockdown of ATG3, ATG5, ATG7, ATG10 and BECN1, as well as the addition of chloroquine or bafilomycin A1) attenuate the capacity of dying tumor cells to successfully vaccinate mice against a re-challenge with live cells of the same type. Importantly, the co-injection of pharmacological inhibitors of extracellular ATPases (also called ecto-ATPases or apyrases) such as ARL67156 and NGXT191, which increase extracellular ATP concentrations, is able to restore the immunogenicity of autophagy-deficient dying cancer cells. These data corroborate the notion that the absence of a single hallmark of ICD (in this case ATP release) is sufficient to abolish the immunogenicity of cell death in functional assays.
Next, we determined the impact of autophagy inhibition on ICD-based chemotherapy in vivo. To this aim, we created tumor cell lines that are stably depleted for either of two essential autophagy gene products (ATG5 or ATG7). Autophagy-deficient fibrosarcoma or colon cancer cells proliferate normally in vivo and develop tumors that are histologically indistinguishable from their isogenic, autophagy-competent counterparts. In response to systemic chemotherapy with anthracyclins or oxaliplatin, both autophagy-competent and -incompetent tumors exhibit a similar apoptotic and necrotic response. Thus, ICD-based chemotherapy enhances the frequency of cells exhibiting nuclear pyknosis, active caspase-3 or HMGB1-negative nuclei to a similar extent in autophagy-proficient and -deficient neoplasms. However, in response to ICD inducers, ATG5 or ATG7-depleted tumors do not exhibit signs of autophagy such as the aggregation of LC3 in cytoplasmic dots, and they fail to release ATP into the extracellular milieu. These latter measurements were performed on tumors engineered to stably express luciferase on the cell surface, taking advantage of in vivo imaging techniques to quantify the ATP-dependent luciferase activity before and after chemotherapy.
Autophagy-competent tumors are strongly infiltrated by DCs and activated T lymphocytes, two and nine days after the systemic administration of ICD inducers, respectively. On the contrary, autophagy-deficient neoplasms fail to recruit such innate and cognate immune effectors into the tumor bed, correlating with a limited release of ATP in response to ICD-based chemotherapy. In this context, autophagy-incompetent tumors continue to proliferate in spite of the chemotherapeutic regimen, while their autophagy-competent counterparts exhibit a reduced growth rate after therapy. This therapeutic success is entirely dependent on the immune system, as the administration of anthracyclins or oxaliplatin fail to affect the proliferation of tumors growing on athymic mice (which lack T cells). To augment extracellular ATP levels, we intratumorally injected ARL67156 or NGXT191, a manipulation that restores the recruitment of DCs and T cells into autophagy-deficient cancers responding to systemic chemotherapy with anthracyclins. In addition, the intratumoral administration of ecto-ATPase inhibitors re-establishes an entirely immunodependent reduction in the growth rate of autophagy-deficient tumors that otherwise would not have responded to chemotherapy with ICD inducers.
Altogether, these results suggest that only autophagy-competent tumors can respond to chemotherapy by reinstating a period of immunosurveillance, underscoring the importance of autophagy for therapy-relevant ICD. It remains an open question whether similar mechanisms may be involved in the evasion of autophagy-deficient, (pre-)malignant cells from natural immunosurveillance during early oncogenesis.
Acknowledgments
G.K. is supported by the Ligue Nationale contre le Cancer (Equipes labelisée), Agence Nationale pour la Recherche (ANR), Fondation Axa (Chair For Longevity Research), European Commission (Apo-Sys, ArtForce, ChemoRes), Fondation pour la Recherche Médicale (FRM), Institut National du Cancer (INCa), Cancéropôle Ile-de-France and Fondation Bettencourt-Schueller. I.M. and S.A. are supported by La Ligue Nationale contre le Cancer, M.M. by FRM, Y.M. by the China Scholarship Council, A.Q.S. by the Higher Education Commission of Pakistan and L.G. by Apo-Sys.