Immunogenic cell death of human ovarian cancer cells induced by cytosolic poly(I:C) leads to myeloid cell maturation and activates NK cells



Owing to high rates of tumor relapse, ovarian cancer remains a fatal disease for which new therapeutic approaches are urgently needed. Accumulating evidence indicates that immune stimulation may delay or even prevent disease recurrence in ovarian cancer. In order to elicit proinflammatory signals that induce or amplify antitumor immune reactivity, we mimicked viral infection in ascites-derived ovarian cancer cells. By transfection or electroporation we targeted the synthetic double-stranded RNA poly(I:C) intracellularly in order to activate melanoma differentiation-associated gene-5 (MDA-5), a sensor of viral RNA in the cytosol of somatic cells. Cancer cells reacted with enhanced expression of HLA-class I, release of CXCL10, IL-6, and type I IFN as well as tumor cell apoptosis. Monocytes and monocyte-derived DCs (MoDCs) engulfed MDA-5-activated cancer cells, and subsequently upregulated HLA-class I/II and costimulatory molecules, and secreted CXCL10 and IFN-α. Further, this proinflammatory milieu promoted cytolytic activity and IFN-γ secretion of NK cells. Thus, our data suggest that the engagement of MDA-5 in a whole tumor cell vaccine is a promising approach for the immunotherapy of ovarian cancer.


Despite advances in treatment, epithelial ovarian cancer (EOC) continues to be the most lethal gynecologic cancer in developed countries with a relative 5-year survival rate of only 47%. Current standard management for primary disease combines radical cytoreductive surgery with platinum/taxane-based chemotherapy regimen and yields response rates of up to 80%. Therapeutic success, however, is limited due to the appearance of chemotherapy-resistant relapses. Thus, novel and effective treatment modalities are urgently needed to maintain a higher percentage of patients in complete response. The observation that tumor-infiltrating lymphocytes (TILs) are consistently correlated with improved survival 1–3 makes treatment approaches aimed at enhancing the spontaneous antitumor immune reactivity particularly plausible in EOC patients 4. In principle, EOC cells are sufficiently immunogenic to attract and activate myeloid and lymphoid immune cells. However, immune suppressive signals are often dominant in the tumor milieu, and may prevent effective clearance of tumor cells by the immune system.

In contrast, somatic cells that are virally infected are promptly eliminated by phagocytic and cytotoxic immune cells. Viruses are detected via their nucleic acids by innate immune receptors including members of the TLR and RIG-like receptor (RLR) families 5. TLRs that recognize viral nucleic acids are expressed in selective cell types, primarily immune cells 6. In contrast, RLRs are expressed ubiquitously and can be engaged in most cell types, including tumor cells 7–9. While not a typical consequence of TLR activation, engagement of RLRs can lead to apoptosis in many cell types, which has been found to provide an advantage for experimental immunotherapy of melanoma 10. Accordingly, we have previously shown that the intracellular delivery of RIG-I agonists led to apoptosis of human EOC cells 11.

Polyinosinic-polycytidylic acid (poly(I:C)) has long been used as a synthetic mimic of viral double-stranded RNA (dsRNA), and has been shown to activate innate immune pathways through both the cytosolic RLR melanoma differentiation-associated antigen-5 (MDA-5) 12, 13 and the endosomal receptor TLR3 14. Differential activation of MDA-5 over TLR3 can be achieved through the use of transfection reagents that promote efficient cytosolic delivery of poly(I:C) 7, 15. Targeted delivery of poly(I:C) to MDA-5 was found to promote autophagy, and IRF-3/NOXA-dependent apoptosis 7, 15, 16. Similar to what has been ascribed to the cytotoxicity of certain anthracycline chemotherapy agents 17, apoptosis induced via RLRs is considered as an immunogenic form of tumor cell death.

We show here that upon delivery to the cytosol, poly(I:C) is able to act directly on EOC cells, leading to the production of type I IFN, proinflammatory cytokines/chemokines as well as to cancer cell death. Following the uptake of poly(I:C)-induced apoptotic EOC cells, monocytes and monocyte-derived dendritic cells (MoDCs) become activated, upregulate HLA-class I/II and costimulatory molecules, secrete IFN-α and proinflammatory cytokines/chemokines. This milieu then promotes the activation and effector function of NK cells. Thus, engagement of MDA-5 in EOC cells exhibits anticancer effects via both apoptosis induction as well as NK-cell activation via an altered cytokine milieu.


Cytosolic delivery of poly(I:C) induces apoptosis in EOC cells

By stimulating EOC cells that were derived from malignant ascites of EOC patients, we found that poly(I:C) was able to induce cell death in EOC cells exclusively when delivered intracellularly by transfection. Cytosolic stimulation of EOC cells resulted in an increase in the annexin V-positive population characteristic of early apoptosis (p<0.02; Fig. 1A and B). In contrast, untransfected poly(I:C) did not induce significant apoptosis even at a 10-fold higher dose (Fig. 1A and B). These observations were supported by the occurrence of a hypodiploid sub G0/G1-peak indicative of late stage apoptosis (p<0.03; Fig. 1C and D). Similar to these results obtained using patient-derived EOC cells, MDA-5-engagement following cytosolic delivery induced apoptosis also in the established EOC cell lines IGROV-1 and SKOV-3, as determined by annexin V binding (data not shown) and the appearance of a hypodiploid DNA content (p<0.02; Supporting Information Fig. S1A and B). Transfected polyadenylic acid (poly(A)) does not activate cytosolic nucleic acid receptors, and was used throughout this study as a negative control to reveal unspecific effects caused by the transfection. MDA-5-triggered apoptosis in EOC cells was mitochondria-dependent as reflected by the disruption of ΔΨm (8 h p<0.04, 11 h p<0.002, 24 h p<0.0003; Supporting Information Fig. S1C). In accordance with published results on different tumor cell types 7, 15, 16, these data indicate that poly(I:C)-mediated activation of MDA-5 rather than TLR3 was responsible for the EOC cell apoptosis observed. Of note, although tumor cells often show apoptosis resistance to chemotherapy, we found that transfected poly(I:C) was able to similarly induce cell death in all established lines and patient-derived EOC cells (OVCACE-1–OVCACE-4) studied. Figure legends indicate that the EOC-cell culture(s) used in the experiments yield results representative of all EOC cultures.

Figure 1.

Poly(I:C) induces apoptosis in ascites-derived EOC cells upon cytosolic delivery. (A) Patient-derived EOC cells were stimulated with poly(I:C) either transfected or untransfected. After 24 h cells were stained and the annexin V positive and Hoechst33258 negative phenotype was quantified by flow cytometry. Transfected poly(A) was used as a negative control. (B) The percentage of apoptotic EOC cells was quantified as in (A); cumulative data obtained independently with OVCACE-1, OVCACE-2; mean±SEM; *p<0.05, Student's t-test. (C) Stimulation of MDA-5 by transfected poly(I:C) increased late stage apoptosis in EOC cells. After 24 h cells were stained with propidium iodide and analyzed by flow cytometry to assess the hypodiploid DNA peak. (D) The percentage of apoptotic EOC cells was quantified as in (C); cumulative data obtained independently with OVCACE-2, OVCACE-4; mean±SEM; *p<0.05, Student's t-test. Data are representative of eight independent experiments.

Activation of MDA-5 enhances the immunogenicity of EOC cells

Downregulation of HLA-class I molecules is a common immune escape mechanism in EOC. We found that the cytosolic delivery of poly(I:C) instead promoted MHC-class I upregulation in tumor cells (p<0.02; Fig. 2A), which may increase EOC susceptibility to immune recognition and destruction by T cells. Furthermore, MDA-5 stimulation in EOC cells triggered the release of proinflammatory cytokines/chemokines that alter the tumor milieu in favor of immune activation. In a sensitive bioassay we found that EOC cells secreted type I IFN (p<0.0005; Fig. 2B) only when poly(I:C) was transfected. IFN subtype-specific ELISA revealed that stimulated EOC cells produced IFN-β (p<0.007) exclusively, while IFN-α could not be detected (Supporting Information Fig. S2). In addition, poly(I:C)-transfected EOC cells secreted CXCL10 (p<0.00007), IL-6 (p<0.02; Fig. 2C), as well as CCL5 and TNF-α (data not shown). Very high doses of untransfected poly(I:C) were able to induce CXCL10 and IL-6 release by EOC cells (Fig. 2C). However, untransfected poly(I:C) was not associated with enhanced HLA-class I expression or secretion of type I IFN (Fig. 2A and B), in line with its lack of proapoptotic activity in EOC cells.

Figure 2.

Stimulation of EOC cells via MDA-5 leads to upregulation of HLA-class I and the secretion of proinflammatory cytokines/chemokines. (A) Cell surface expression of HLA-class I antigens on OVCACE-3 cells was determined by flow cytometry 48 h after stimulation. Data are given as geometric MFI values (mean+SEM; *p<0.05, Student's t-test). (B) Release of type I IFN was quantified 18 h after OVCACE-3 stimulation using the HEK-Blue IFN-α/β reporter cell line (mean+SEM; *p<0.05, Student's t-test). (C) Secretion of CXCL10 and IL-6 was quantified in the supernatant by ELISA 18 h after EOC cell stimulation (OVCACE-1, OVCACE-4; mean+SEM; *p<0.05, Student's t-test). Data are representative of three independent experiments.

Exposure to MDA-5-triggered EOC cells induces DC maturation

We next examined whether MDA-5-triggered EOC cells would activate DCs whose presence was shown to be correlated with a favorable prognosis in EOC 18. Indeed, MDA-5-mediated apoptotic tumor cells induced MoDCs to increase the expression of HLA-class I/II and costimulatory molecules (Supporting Information Fig. S3A). Differences in these immune parameters were statistically significant (HLA-class II p<0.05, CD86 p<0.04, CD83 p<0.03; Supporting Information Fig. S3B). Also, MoDCs secreted IFN-α (p<0.000008), CXCL10 (p<0.0002) and IL-12p70 (p<0.01; Supporting Information Fig. S3C) upon coculture with MDA-5-induced apoptotic EOC cells.

Activation of monocytes upon exposure to MDA-5-triggered EOC cells

Our observation of DC maturation following the exposure to MDA-5-triggered EOC cells prompted us to also characterize the activation of monocytes, which constitute the dominant myeloid cell population in the ascites of EOC patients 19, making them an attractive target for therapeutic interventions. In response to MDA-5-stimulated EOC cells, monocytes enhanced the expression of molecules related to activation and costimulation, namely HLA-class I/II, CD69 and CD86 (Fig. 3A). Accordingly, monocytes exposed to MDA-5-triggered EOC cells showed a dose-dependent release of IFN-α (poly(A)-treated tumor cells p<0.005, directly treated monocytes p<0.006, directly treated tumor cells p<0.005; Fig. 3B) and CXCL10 (poly(A)-treated tumor cells p<0.007, directly treated monocytes p<0.002, directly treated tumor cells p<0.002; Fig. 3C). Of note, the low levels of cytokine/chemokine secretion observed after direct stimulation of monocytes were not due to monocyte apoptosis (Supporting Information Fig. S4A), in contrast to the dose-dependent apoptosis of EOC cells (p<0.002; Supporting Information Fig. S4B).

Figure 3.

MDA-5-triggered EOC cells activate monocytes and elicit a proinflammatory state. (A) Monocytes were exposed to MDA-5-stimulated EOC cells (OVCACE-1 and OVCACE-4). After 48 h cell surface expression of HLA-class I/II, CD69 and CD86 was determined by flow cytometry. Data are given as geometric MFI values. (B) Secretion of IFN-α by monocytes in response to exposure to MDA-5-triggered OVCACE-1 cells as determined after 18 h by ELISA (mean+SEM; *p<0.05, Student's t-test). Control experiments included single culture of EOC cells and monocytes. (C) The release of CXCL10 was quantified as in (B); mean+SEM; *p<0.05, Student's t-test. Data are representative of three independent experiments.

MDA-5-activation is of benefit over chemotherapeutic treatment in terms of monocyte activation

Some of the established chemotherapy compounds have been described to induce an immunogenic form of tumor cell death 17. We therefore compared MDA-5-triggered immune-enhancing effects with chemotherapeutic-induced alterations. All agents were used at doses at which they were similarly effective in inducing EOC-cell death (poly(I:C) p<0.0001, paclitaxel p<0.002, doxorubicin p<0.04; Fig. 4A). Given that apoptotic cells externalized phosphatidylserine (Fig. 1A and B), which serves as an “eat-me” signal, we analyzed monocyte phagocytosis, and found that apoptotic EOC cells were readily ingested irrespective of cell-death inducer (Fig. 4B). Our data indicate an active uptake rather than a tight association of monocytes with apoptotic cells, since doublet events were excluded, the acquired fluorescence of monocytes remained always lower than that of EOC cells alone, and the process was considerably inhibited at 4°C, a temperature that blocks phagocytosis but not surface adhesion (Fig. 4B and data not shown). We further characterized the activation status of monocytes upon exposure to apoptotic EOC cells. Poly(I:C)-mediated but not chemotherapeutic-treated EOC cells induced monocytes to release IFN-α (poly(A)-treated tumor cells p<0.02, directly treated monocytes p<0.004, directly treated tumor cells p<0.02; Fig. 4C). Moreover, the release of immunosuppressive IL-10, which is induced in monocytes upon exposure to EOC cells, was significantly diminished when EOC cells were pretreated with poly(I:C) (p<0.02; Fig. 4D). Both paclitaxel and doxorubicin treatment reduced IL-10 secretion by monocytes but with a less pronounced effect.

Figure 4.

The proinflammatory state of monocytes induced by MDA-5-stimulated EOC cells surpasses activation achieved with immunogenic chemotherapy. (A) OVCACE-4 cells were treated with paclitaxel, doxorubicin or transfected with poly(I:C) at the doses indicated. After 24 h apoptotic cells were detected by flow cytometry using propidium iodide staining after permeabilization and appeared as the hypodiploid DNA peak (mean+SEM; *p<0.05, Student's t-test). (B) Monocytes were exposed to apoptotic PKH26-labeled OVCACE-3 cells for 4 h, stained and analyzed by flow cytometry. Control experiments with untreated EOC cells were conducted at 4° and 37°C. Double-positive cells indicate the uptake of apoptotic EOC cells by monocytes. (C) Monocytes were exposed to apoptotic OVCACE-4 cells. Control experiments included single culture of EOC cells and monocytes. After 18 h IFN-α secretion was quantified in the supernatant by ELISA (mean+SEM; *p<0.05, Student's t-test). (D) Secretion of IL-10 was quantified in the supernatant as in (C); OVCACE-4; mean+SEM; *p<0.05, Student's t-test. Data are representative of three independent experiments.

Immunogenic apoptosis and monocyte activation are independent of the mode of MDA-5 ligand delivery

As an alternative method for cytosolic delivery of poly(I:C), we used electroporation of EOC cells, and found that electroporation of the MDA-5 ligand also induced EOC-cell death (p<0.006; Supporting Information Fig. S5A) and type I IFN secretion (p<0.04; Supporting Information Fig. S5B). Moreover, poly(I:C)-electroporated EOC cells showed enhanced tumor immunogenicity as they triggered monocytes to secrete CXCL10 (poly(A)-treated tumor cells p<0.002, directly treated monocytes p<0.0003, directly treated tumor cells p<0.007) and to reduce the release of IL-10 (p<0.0002; Supporting Information Fig. S5C). Thus, the observed effects of DCs and monocytes upon exposure to MDA-5-stimulated EOC cells were not due to residues of poly(I:C) in complex with a transfection reagent.

MDA-5 stimulation in EOC cells induces autologous antitumor immune reactivity

Based on our findings of immune activation through MDA-5 engagement in EOC cells, we aimed to determine whether a preexisting tumor-specific immune response might in this way be restimulated. Using cells isolated from malignant ascitic fluid, which contains not only exfoliated tumor but also immune cells, we found that even in a tumor-primed system, activation of MDA-5 in EOC cells significantly induced lymphocytes to secrete IFN-γ (p<0.05; Fig. 5A). Although effector cells in malignant ascites are dominated by CD4+ and CD8+ T cells, a substantial amount of NK cells is always present 11. Thus, IFN-γ release may be due to both NK- and T-cell activation.

Figure 5.

Monocytes activated by MDA-5-induced apoptotic EOC cells elicit a tumor-specific NK-cell response. (A) Secretion of IFN-γ after in vitro restimulation of autologous ascitic immunocytes with MDA-5-triggered OVCACE-4 cells measured by ELISPOT after a 48 h coculture (mean+SEM; *p<0.05, Student's t-test). Controls included untreated EOC and ascites cells. (B) PBMCs were incubated for 18 h with the (co)culture supernatants indicated. NK-cell degranulation was then triggered by the addition of OVCACE-4 cells for 4 h. Surface accumulation of CD107a was analyzed by flow cytometry. Data are given as geometric MFI values (cumulative data obtained with two different donors; mean+SEM; *p<0.05, Student's t-test). (C) PBMCs were incubated for 18 h with the (co)culture supernatants indicated. DiO-labeled OVCACE-3 cells were then added in different E:T ratios, and their specific lysis was determined by flow cytometry after 4 h. (D) Expression of CD69 on purified NK cells following an 18 h incubation with the indicated (co)culture supernatants (OVCACE-3) as determined by flow cytometry after 18 h. Data are given as geometric MFI values (mean+SEM; *p<0.05, Student's t-test). (E) Purified NK cells were plated in the (co)culture supernatants (OVCACE-4) indicated. After 18 h secretion of IFN-γ was quantified in the supernatant by ELISA (mean+SEM; *p<0.05, Student's t-test). Data are representative of three independent experiments.

Activation of NK cells upon exposure to MDA-5-triggered EOC cells cocultured with monocytes

Type I IFN plays an integral role in NK-cell activation 20. Based on our data of MDA-5-induced type I IFN release by EOC cells and the subsequent initiation of IFN-α secretion by APCs, we next analyzed NK-cell cytolytic activity within peripheral blood mononuclear cells (PBMCs) that were exposed to the supernatant derived from EOC cell-monocyte coculture. By measuring degranulation-dependent surface exposure of CD107a 21 we found that NK-cell cytotoxicity was significantly upregulated following incubation with supernatants from MDA-5-stimulated EOC cells both individually cultured (p<0.03), and cocultured with monocytes (p<0.02; Fig. 5B). Yet, NK-cell lysis of EOC cells was elevated only when PBMCs had been incubated with supernatant from cocultures of MDA-5-stimulated EOC cells with monocytes (Fig. 5C). Of note, (co)culture supernatant was able to promote NK-cell degranulation and cytotoxicity only for NK cells within PBMCs, indicating a requirement for cell contact with other cell types. In contrast, exposure of isolated NK cells to the supernatants of monocytes cocultured with MDA-5-stimulated EOC cells was sufficient for NK-cell upregulation of CD69 (p<0.0007; Fig. 5D), as well as IFN-γ production (p<0.0003; Fig. 5E).

Whole cell vaccination with MDA-5-stimulated EOC cells leads to in vivo activation of DCs and NK cells

Our finding that MDA-5 stimulation enhanced the immunogenicity of EOC cells may have relevance for the development of a tumor vaccine based on patient-derived EOC cells. However, before a clinical trial can be conceived, a strong foundation of preclinical data needs to be obtained in a suitable animal model in which the immune reactivity of the whole organism can be evaluated. We therefore transposed our approach by transfecting C57BL/6 syngeneic murine ID8 EOC cells 22 with either poly(A) or poly(I:C) 4 h before intraperitoneal injection into C57BL/6 mice. Six hours after injection, IFN-α was detectable only in the serum samples of mice that had received MDA-5-stimulated EOC cells (p<0.03; Fig. 6A). After 24 h, spleen DCs in mice that had been injected with MDA-5-stimulated EOC cells showed an upregulation of CD69 (p<0.004) and CD86 (p<0.007; Fig. 6B). Similarly, CD69 was upregulated on NK cells in spleen (p<0.004), LNs (p<0.003), and the peritoneal cavity (p<0.000006; Fig. 6C), again exclusively in mice treated with MDA-5-stimulated EOC cells. In the same mice NK-cell proliferation (p<0.005; Fig. 6D), measured via the short-term incorporation of Bromodeoxyuridine (BrdU), and IFN-γ production by NK cells (p<0.02; Fig. 6E) were significantly elevated. Overall, these observations are congruent with the results we obtained in vitro using primary human EOC cells.

Figure 6.

Intraperitoneal administration of MDA-5-stimulated ID8 cells leads to the activation of NK cells. Mice were injected i.p. with poly(A)- or poly(I:C)-transfected ID8 cells, or were left untreated. (A) Serum IFN-α 6 h after injection was determined by ELISA (mean+SEM; *p<0.05, Student's t-test). (B) Splenocytes from mice were collected 24 h after injection. Cell surface expression of CD69 and CD86 on CD11c+ DCs was determined by flow cytometry. Data are given as geometric MFI values (mean+SEM; *p<0.05, Student's t-test). (C) Cell surface expression of CD69 on CD3 NK1.1+ NK cells was determined 24 h after injection by flow cytometry. Data are given as geometric MFI values (mean+SEM; *p<0.05, Student's t-test). (D) In vivo NK-cell proliferation 24 h after injection was determined by short-term BrdU incorporation. Data are given as the percentage of BrdU+ cells of all NK cells identified (mean+SEM; *p<0.05, Student's t-test). (E) For intracellular staining of IFN-γ in NK cells, spleen cells were isolated 24 h after injection and restimulated for 4 h with PMA/ionomycin in the presence of Brefeldin A. Data are given as the percentage of IFN-γ+ cells of all NK cells identified (mean+SEM; *p<0.05, Student's t-test). Data are representative of two independent experiments.


Targeted intracellular delivery of poly(I:C) was effective in mediating apoptosis in all EOC specimens analyzed suggesting an intact MDA-5 signaling pathway even in poorly differentiated tumors. This result is consistent with previous data obtained in melanoma or hepatocellular carcinoma cells 7, 15, 23. In contrast to studies with different tumor entities (melanoma, breast and prostate cancer) that evaluated untransfected poly(I:C) as a TLR3 agonist 24–26, incubation of EOC cells with even high doses of untransfected poly(I:C) did not induce apoptosis. Expression of TLR3 in tumor cells is variable and a proapoptotic effect of TLR3 engagement was identified only in patients with strong TLR3 expression 27. Since TLR3 was expressed only at marginal levels in all EOC cells we tested (data not shown) our results suggest that MDA-5 rather than TLR3 is the major mediator of tumor-cell apoptosis in EOC cells.

In addition, we found an immune-enhancing effect of transfected poly(I:C) on EOC cells through the upregulation of HLA-class I molecules and the secretion of proinflammatory cytokines and chemokines. A sufficient expression level of HLA-class I is critical for tumor cell lysis by CD8+ T cells as evidenced by the improved prognosis that correlate with HLA-class I-positive tumors 28. We further observed that MDA-5-triggered EOC cells induced the maturation of DCs and monocytes. Upon exposure to MDA-5-stimulated EOC cells, APCs shifted their secretion profile toward a proinflammatory pattern. In particular, type I IFN was detected, which is considered pivotal for efficient tumor rejection by the host immune system 29, and by itself has been shown to induce clinical responses in EOC patients 30. MDA-5-triggered EOC cells also reversed immune suppression by lowering EOC cell-induced IL-10 release from APCs. This shift in the cytokine milieu may help the recruitment of further tumor control immune mechanisms, such as activation and increased cytotoxicity of NK cells. Poly(I:C) is known to induce activation of NK cells, an effect that has recently been shown to be primarily mediated by MDA-5, while TLR3 played only a minor role 31. We found that APCs, which had engulfed MDA-5-triggered apoptotic EOC cells, released soluble factors that stimulated NK cells to upregulate the activation marker CD69, which is considered to be a tumor-specific triggering receptor on NK cells 32, to secrete IFN-γ, and to augment antitumor cytotoxicity. IFN-γ was shown to be an independent prognostic factor for overall survival in EOC patients thus playing an important role in antitumor immunity 33. Accordingly, a beneficial effect of IFN-γ administration on survival times has been observed 34. Support for the importance of NK-cell-mediated cytotoxicity in tumor control comes from a study that reported the association of enhanced NK-cell-dependent lysis and clinical responses by using a virus-modified EOC-cell extract 35.

In EOC, initial remission can be achieved in the majority of cases by standard therapy. However, chemotherapeutic agents often fail to destroy all EOC cells at clinically tolerated doses. Especially in cases with complete response, the application of non-cross-resistant agents as a maintenance therapy to prevent recurrence appears useful. Based on our results, we propose an immunotherapeutic strategy that uses poly(I:C)-triggered apoptotic EOC cells as a whole cell tumor vaccine 36 to be initiated once remission has been achieved by standard therapy, i.e. at a time when residual tumor burden is minimal, and immune suppression established by the primary tumor is reverted. Since EOC typically remains limited to the peritoneum cavity it offers a unique milieu in that the peritoneal cavity is an immune-rich environment accessible to the key innate and adaptive immune contributors to an effective antitumor response. In this respect, the local administration of immunotherapeutic agents appears attractive.

Overall, we here provide proof of concept that the engagement of MDA-5 in EOC cells exhibits anticancer effects via both cytodestructive activity and mobilizing antitumor immune responses. Our study of tumor-specific immune reactivity triggered by patient-derived, MDA-5-stimulated EOC cells is necessarily limited to the spectrum of immune parameters that can be assessed in vitro. In particular, the effects of MDA-5-stimulated EOC cells on the induction and restimulation of tumor-specific T cells cannot be readily assessed with the amount of autologous immune cells obtained from malignant ascites. We therefore analyzed the immune stimulatory capacity of MDA-5-stimulated EOC cells in mice with a whole cell vaccination approach. The results obtained confirmed our in vitro observations. Using this mouse model of EOC we currently evaluate whether MDA-5-triggered cancer cells can also be functionalized as a tumor vaccine to induce a potent anticancer cytolytic CD8+ T-cell response by synergistically engaging both antiviral and antitumor effector mechanisms.

Materials and methods

Cell culture

Ascites were collected at the time of surgical treatment from therapy-naïve patients with EOC (serous papillary). Cell cultures were established modifying a previously published technique 37. Briefly, cells isolated from ascites were cultured in MCDB 105/M199 (1:1 v/v) supplemented with 10% FCS, 100 units/mL penicillin, and 100 μg/mL streptomycin. Upon the formation of confluent cell monolayers cancer cells were enriched using anti-epithelial cell adhesion molecule (Epcam) microbeads (Miltenyi, Bergisch Gladbach, Germany). We newly derived four ovarian cancer cell cultures (OVCACE-1–4) that were first characterized in our previous study 11. The ethics committee of the University of Bonn approved the study protocol and all patients gave written consent. We also used the human EOC cell lines IGROV-1 38 and SKOV-3 39 which were cultured in RPMI-1640 supplemented with 10% FCS, 4 mM L-glutamine, 200 units/mL penicillin, 200 μg/mL streptomycin, and 1 mM sodium bicarbonate. Cells were grown at 37°C in a humidified 5% CO2 atmosphere under strictly endotoxin-free conditions. The identity of each cell line was verified by flow cytometry at regular intervals.


Monocytes were enriched from buffy coat-derived peripheral blood mononuclear cells (PBMCs) of healthy blood donors using the monocyte isolation kit II (Miltenyi). A purity of more than 98% was achieved, as determined by flow cytometry. Monocytes were maintained in RPMI-1640 supplemented with 2% human AB serum, 2 mM L-glutamine, 100 units/mL penicillin, and 100 μg/mL streptomycin. For MoDCs, monocytes were purified from PBMCs using CD14 microbeads (Miltenyi) and cultured for six days in CellGro DC (CellGenix, Freiburg, Germany) adding granulocyte-macrophage colony-stimulating factor (rhGM-CSF, 800 units/mL) and interleukin 4 (rhIL-4, 1000 units/mL; ImmunoTools, Friesoythe, Germany) on the first and fourth days.


The synthetic dsRNA poly(I:C) (InvivoGen, Toulouse, France) was used as MDA-5 agonist 13. Poly(I:C) and poly(A) (Sigma-Aldrich, Munich, Germany) were transfected using FuGENE HD (Roche, Mannheim, Germany), TransIT-LT1 (Mirus Bio LCC, Madison, WI, USA) or Lipofectamine 2000 (Invitrogen, Karlsruhe, Germany) according to the manufacturer's recommendations. Cells were electroporated with one pulse at 300 V for 6 ms (Gene Pulser II, Biorad, Munich, Germany). Opti-MEM (Invitrogen) was used for mock-transfection; poly(A) as a non-stimulatory control. Untransfected poly(I:C) was used as a ligand for TLR3 14. Furthermore, tumor cells were exposed to antineoplastic regimens. Drug concentrations were chosen to be of clinical relevance by corresponding to the achievable plasma peak concentration after administration of a standard dose (14 μg/mL paclitaxel corresponds to 175 mg/m2 i.v.; 0.5 μg/mL doxorubicin corresponds to 60 mg/m2 i.v.) 40.

Cell viability and apoptosis assays

Viable cells were quantified using the CellTiter-Blue assay according to the manufacturer's instructions (Promega, Mannheim, Germany). Apoptosis was assessed by staining with annexin V FLUOS (Roche) in conjunction with Hoechst 33258 41. Fragmentation of genomic DNA was evaluated by permeabilizing cells in absolute ethanol, and staining with propidium iodide in the presence of DNase-free RNase A 42. Changes in mitochondrial membrane potential (ΔΨm) were assessed using the MitoProbeTM DiOC2(3) assay kit according to the manufacturer's instructions (Invitrogen).

Coculture and phagocytosis assay of EOC cells with monocytes or MoDCs

Apoptosis was induced in EOC cells by poly(I:C)-electroporation or transfecting poly(I:C) for 4 h. Cells were then washed extensively with fresh culture medium before coculture with monocytes in a 1:2 ratio or with MoDCs in a 1:1 ratio. Fluorescence-labeled antibody against Epcam was used to exclude tumor cells during flow cytometric analysis. To evaluate phagocytosis, tumor cells were labeled with the membrane dye PKH26 (Sigma-Aldrich) following the manufacturer's instructions; 18 h after induction of apoptosis EOC cells were cocultured for 4 h with monocytes, which were then stained for CD14 and analyzed by flow cytometry.

NK-cell assays

NK cells were isolated from buffy coat-derived PBMCs using the NK-cell Isolation Kit (Miltenyi). A purity of more than 98% was achieved as determined by flow cytometry. A total of 4×105 PBMCs or 3×104 NK cells were treated with the supernatant of EOC-monocyte coculture for 18 h; for the lysis assay 50 U/mL IL-2 (ImmunoTools) were added. Degranulation of NK cells was quantified by flow cytometric detection of membranous CD107a as described previously 21. Briefly, PBMCs were incubated for 4 h with EOC cells in a 10:1 ratio in the presence of CD107a antibody and monensin (5 μg/mL; Sigma-Aldrich) added during the final 3 h. Samples were stained with anti-CD3 and anti-CD56 antibodies to identify NK cells. EOC lytic activity was determined using a flow cytometric assay as described 43. Briefly, EOC cells were labeled with 0.03 μM DiO (Invitrogen) and incubated at 37°C for 10 min. PBMCs were added for 4 h and cells were then stained with Hoechst 33258. Lysis was determined flow cytometrically as Dio+Hoechst+ cells/DiO+Hoechst+ cells+DiO+Hoechst cells×100, followed by subtraction of spontaneous lysis.


IFN-γ-secreting cells were detected using the enzyme-linked immunospot (ELISPOT) assay. Apoptotic OVCACE cells were cocultured with unseparated autologous ascites cells in a 1:45 ratio in plates previously coated with mouse anti-human IFN-γ monoclonal antibody (clone 1-D1K; Mabtech AB, Nacka Strand, Sweden). After 48 h, plates were developed using biotinylated mouse anti-human IFN-γ monoclonal antibody (clone 7-B6-1), streptavidin alkaline phosphatase, and substrate reagent (Mabtech AB). Spots were counted with an automated ImmunoSpot® S5 UV analyzer using the ImmunoSpot software 5.0.9 (Cellular Technology, Shaker Heights, OH, USA).

Measurement of cytokines and chemokines

Secretion of IFN-α (Bender MedSystems, Vienna, Austria), IFN-β (Invitrogen), IFN-γ, IL-6, IL-10, IL-12p70 and CXCL10 (IP-10) (BD Biosciences, Heidelberg, Germany) was quantified after 18 h in cell-free supernatants by ELISA according to the manufacturer's instructions. Bioactive type I IFN was quantified using the HEK-Blue IFN-α/β reporter cell line (InvivoGen) according to the manufacturer's instructions.

Flow cytometry

Antibodies used are provided in Supplementary Table S1. Data were obtained on an LSR II flow cytometer (BD Biosciences) evaluating at least 10 000 events per sample by gating on Hoechst 33258-negative cells after excluding doublets. Analysis was performed by FlowJo software (TreeStar, Olten, Switzerland).

Animal experiments

C57BL/6J mice were obtained from Janvier (Le Genest Saint-Isle, France). Animal experiments were performed at the Central Animal Facility of the University of Bonn in adherence to the standards of the German law for the care and use of laboratory animals. Mouse ovarian surface epithelial cells (MOSEC clone ID8) were cultured as described 22. Totally, 4×106 ID8 cells were transfected for 3 h with 15 μg poly(A) or poly(I:C) using TransIT-LT1. Cells were then washed and injected i.p. into 12-wk-old C57BL/6J mice; untreated mice served as controls. Blood samples were taken after 6 h. Mice were injected after 21 h with 1 mg BrdU, and after 24 h peritoneal lavage was performed, and single-cell suspensions were prepared from spleens and lymph nodes. Serum IFN-α was quantified by ELISA according to the manufacturer's instructions (PBL Biomedical Laboratories, Piscataway, NJ, USA). Flow cytometric analysis was carried out using the antibodies provided in Supplementary Table S2. IFN-γ was stained intracellularly following treatment with PMA (50 ng/mL, Enzo Life Sciences, Lörrach Germany), ionomycin (250 ng/mL, Sigma-Aldrich) and Brefeldin A (1 μg/mL, Sigma-Aldrich) for 4 h at 37°C, using the Cytofix/Cytoperm kit (BD Biosciences). The in vivo proliferation of NK cells was determined using the FITC BrdU Flow kit (BD Biosciences) according to the manufacturer's instructions.

Statistical analysis

Values are given as mean±standard error of the mean (SEM). The F-test was used to verify the assumption of equal variances; the unpaired one-tailed Student's t-test was used to determine statistical significance of differences. The results with a p-value<0.05 were considered to be signicant.


The authors thank Bettina Langhans for help in establishing ELISPOT assays, Juan José Hernando and Martin Schlee for helpful discussions and providing IGROV-1 SKOV-3, and HEK-Blue IFN-α/β reporter cell lines, Katherine Roby for kindly providing ID8 cells, and Jasper van den Boorn for critical reading of the manuscript. This work was supported by Deutsche Forschungsgemeinschaft grants BA3544/1-1 and SFB704 (W. Barchet), BONFOR grants of the University of Bonn (K. Kübler and W. Barchet) and a Maria von Linden grant of the University of Bonn (K. Kübler).

Conflict of interest: The authors declare no financial or commercial conflict of interest.