Abstract
Despite improvements in chemotherapy and radical surgical debulking, peritoneal carcinomatosis (PC) remains among the most common causes of death from abdominal cancers. Immunotherapies have been effective for selected solid malignancies, but their potential in PC has been little explored. Here, we report that intraperitoneal injection of an infected cell vaccine (ICV), consisting of autologous tumor cells infected ex vivo with an oncolytic Maraba MG1 virus expressing IL12, promotes the migration of activated natural killer (NK) cells to the peritoneal cavity in response to the secretion of IFNγ-induced protein-10 (IP-10) from dendritic cells. The recruitment of cytotoxic, IFNγ-secreting NK cells was associated with reduced tumor burden and improved survival in a colon cancer model of PC. Even in mice with bulky PC (tumors > 8 mm), a complete radiologic response was demonstrated within 8 to14 weeks, associated with 100% long-term survival. The impact of MG1-IL12-ICV upon NK-cell recruitment and function observed in the murine system was recapitulated in human lymphocytes exposed to human tumor cell lines infected with MG1-IL12. These findings suggest that an MG1-IL12-ICV is a promising therapy that could provide benefit to the thousands of patients diagnosed with PC each year. Cancer Immunol Res; 5(3); 211–21. ©2017 AACR.
Introduction
Peritoneal carcinomatosis is one of the most common and problematic sites of metastases for abdominal malignancies, including gastrointestinal and ovarian cancers (1). Its presentation is associated with a significantly reduced quality of life and a very poor prognosis, with a median survival of 6 to 12 months. Chemotherapy is less effective in patients with peritoneal carcinomatosis, and it cannot be administered once a patient develops a complication, such as a bowel obstruction (2). Intraperitoneal chemotherapy with surgical debulking is associated with higher survival rates than systemic chemotherapy, but results are still disappointing, and treatment is rarely curative (3). However, several independent groups have reported that immune modulating agents can provide a significant therapeutic benefit in preclinical models of peritoneal carcinomatosis (PC; refs. 4–7). These promising findings suggest that strategies designed to boost the antitumor immune response within the peritoneal cavity should be pursued in the treatment of PC.
Natural killer (NK) cells play a critical role in the recognition and clearance of malignant cells (8). The therapeutic benefit achieved with boosting NK-cell activation and cytolytic activity through cytokine therapy has previously been reported for a number of cancers (9). IL12, arguably one of the best characterized proinflammatory cytokines, is a potent stimulator of NK cell–mediated tumor cell killing (10–12). In addition, the action of IL12 is not restricted to NK-cell activation. It also induces secretion of additional cytokines and chemokines through paracrine stimulation of conventional dendritic cells (cDC). These cDCs contribute to further activation of the innate and adaptive arms of the immune system, ultimately leading to a durable adaptive immune response to tumor-associated antigens (13, 14). However, the short half-life and toxicity following systemic administration precludes the widespread adoption of systemic IL12 therapy for the treatment of cancer. Furthermore, cytokine therapies such as IL12 are not as effective in enhancing immune-mediated tumor cell clearance as strategies that combine cytokine delivery with tumor antigens or other additional methods of immune system activation (9). Ultimately, cytokine therapy alone is associated with a less enduring immune response and a greater likelihood of tumor cell escape and disease recurrence (9).
Several groups reported nearly three decades ago that combining cytokine delivery with tumor cells yields an increased antitumor immune response capable of significantly delaying tumor growth and disease progression (15). The profound antitumor response elicited by an autologous “tumor cell vaccine” is thought to be a result of enhanced activation of the immune system in response to delivering tumor-specific antigens in a proinflammatory environment created by the cytokine. Existing data suggest that disease recurrence does not occur or is significantly delayed when patients successfully mount an immune response against the tumor, as evidenced by a delayed-type hypersensitivity response. Unfortunately, the majority of patients do not mount such a response, either because the cell vaccine is not immunogenic enough or because the host immune system is suppressed in response to the cancer. Our laboratory and others have endeavored to improve upon this vaccination paradigm by infecting autologous tumor cells with oncolytic viruses engineered to express immune modulating cytokines (16–18). We have previously provided evidence for this approach by demonstrating that an oncolytic rhabdovirus, VSVΔM51 expressing GM-CSF could significantly enhance immune activation and prevent tumor outgrowth (18). GM-CSF exerts its antitumor activity through enhancing DCs activation and antigen presentation. In addition, GM-CSF induces T-cell functions indirectly, in an interleukin-1–dependent manner (19). Unfortunately, this antitumor activity is limited, due to the counter-regulatory activities of GM-CSF overexpression that expand myeloid-derived suppressor cells, which in turn inhibit CTL and NK-cell functions (19). As such, we sought to determine whether the therapeutic potential and safety of the infected cell vaccine (ICV) platform could be improved by combining the superior NK cell–stimulating properties of IL12 with the tumor-targeting properties of a replicating oncolytic Maraba MG1 virus, a genetically modified virus with mutations in both the G and M proteins which expand its tropism for diverse cancer types and greatly attenuate its replication in healthy tissue when compared with VSVΔM51 (20).
Additionally, the immunostimulatory properties of replicating MG1 are in part due to the potent effects of this oncolytic virus upon NK-cell activation, as assessed by increased CD69 expression, IFNγ secretion, and granzyme B production (21).
Here, we report that intraperitoneal delivery of autologous cells infected with a replicating MG1 virus expressing IL12 (MG1-IL12-ICV) provided a significant therapeutic benefit to normally resistant murine models of established peritoneal disease. The IL12-producing vaccine was well tolerated by the experimental animals while inducing a robust recruitment of activated, cytotoxic NK cells to the peritoneal cavity. Magnetic resonance imaging revealed that the recruitment of activated NK cells was associated with a substantial reduction in the size of established peritoneal tumors while providing a durable cure, with no evidence of disease recurrence 14 weeks after treatment.
Materials and Methods
Cell lines and mice
Murine CT26 colon carcinoma, murine B16F10 melanoma, human SW620 colorectal adenocarcinoma, human HCT15 colorectal adenocarcinoma, human A549 lung carcinoma, murine YAC-1 lymphoma, human K562 leukemic cell lines (all from ATCC) were propagated in DMEM (Hyclone) for the adherent cell lines, or RPMI Media (Hyclone) for nonadherent cell lines supplemented with 10% fetal calf serum (Cansera). RMA and RMA-S were obtained from Dr. A. Veillette (Institut de Recherches Clinique, Montreal, Quebec, Canada). Cell lines were obtained from collaborators as well as from ATCC between 2009 and 2013 and they were not authenticated; however, cells' morphology and growth characteristics are conformant, and they have been tested regularly for mycoplasma using PCR and/or Hoechst staining methods. MHC-I staining of all cell lines is conducted every 6 to 12 months. Cell lines were passaged every 3 days and maintained in culture no more than 2 weeks prior the experiments. Female Balb/C and C57BL/6 mice (6–8 weeks old) were purchased from Charles River Laboratories. Animals were housed in pathogen-free conditions, and all experiments were conducted with the approval of the University of Ottawa Animal Care and Veterinary Service.
MG1-IL12 construction and viability assays
Murine IL12 was PCR amplified from pORF-mIL12 [IL12elasti(p35::p40); InvivoGen] to add MluI (5′) and (3′) cloning sites to facilitate cloning into the rhabdovirus Maraba MG1 and the nonreplicating, glycoprotein deleted Maraba MG1, hereafter called G-less MRB (Supplementary Fig. S1; refs. 21, 22). The recombinant MG1-IL12 and G-less MRB-IL12 viruses were rescued as described elsewhere (20, 21). Viral cytotoxicity was assessed on the indicated cell lines, and cell viability was carried out as described previously (21).
Antibodies and FACS analysis
For splenic and lung lymphocyte population analyses, organs were harvested from mice and red blood cells lysed using ammonium chloride–potassium lysis (ACK) buffer. Monoclonal antibodies (mAb) to TCRβ (H57-597) and NK1.1 (PK136) were from eBiosciences. Spleen and lung NK-cell IFNγ and Granzyme B secretion were analyzed after a 1-hour GolgiPlug (BD Biosciences) incubation using anti-CD3 (17A2), anti-NK1.1 (PK136), anti-IFNγ (XMG1.2), and anti–Granzyme B (16G6), all from BD Biosciences. The mAbs used for human NK-cell migration and activation were anti-CD56 (HCD56) from Biolegend and anti-CD3 (UCHT1) and anti-CD69 (FN50) from eBiosciences. Fluorescence-activated cell sorting (FACS) acquisitions were conducted on a CyAn-ADP using Summit software (Beckman Coulter), and data were analyzed with Kaluza software (Beckman Coulter).
Ex vivo splenocyte cytotoxicity assay
The 51Cr-release assay was performed as previously described (23). Briefly, splenocytes were harvested from treated and control mice 2 days after treatment. ACK buffer–treated splenocytes were resuspended and mixed with 51Cr-labelled YAC-1 cells at specified effector-to-target (E:T) ratios.
In vivo tumor rejection assay
The in vivo rejection assay was performed as described previously (23). Briefly, RMA and RMA-S were labeled with 5 and 0.5 mu/L CFSE, respectively. Cells (1 × 106) containing a 1:1 mixture of each cell type were injected i.p. into C57BL/6 mice 24 hours after ICV treatment. Peritoneal cells were collected the next day (24 hours) by washing the peritoneum with 5 mL of PBS containing 2 mmol/L EDTA. Collected cells were analyzed by flow cytometry for the presence of CFSE labeling.
Viral infection of B16F10 cells and coculture with bone marrow–derived DCs (DC-ICV)
B16F10 cells were mock infected or infected with MG1 or MG1-IL12 (MOI = 0.1 pfu/cell) and harvested after 18 hours. Infected cells were cultured with bone marrow–derived DCs, described elsewhere (24) at a 3:1 ratio in DC medium (1% FBS; complete RPMI supplemented with 1× of 2-Mercapoethanol; cat #21985-023, Gibco, Life Technologies) in 96-well plates. Media were collected after 24 hours and stored at −80°C until further analysis.
Cytokine and chemokine analyses
Murine IFNγ was quantified from DC coculture supernatants by FlowCytomix (eBioscience) kits as per the manufacturer's instructions. For lung IL12 and IFNγ expression, lungs from C57BL/6 mice treated with irradiated B16F10, MG1-ICV, or MG1-IL12-ICV at 5 × 105 cells/100 μL/mouse intravenously (i.v.) were resected and homogenized in 1 mL PBS 24 hours after treatment. Murine MCP-1, SDF-1, and IP-10 chemokines were assayed from tissue culture supernatant using ELISArray kits (SABiosciences) as per the manufacturer's instructions.
Transwell chemotaxis assay
Tissue culture supernatants for assessment of chemokines or chemotaxis assay were generated in DC media. Chemotaxis of NK cells was assessed using a Transwell system as described previously (17). Briefly, 500 μL of conditioned media from DC cultures was added to the lower chamber of Transwell plates with 5-μm pores (Costar, Corning). DX5+-sorted NK cells (3 × 105) were added to the upper chamber, and plates were incubated for 3 hours at 37°C. Cells in the lower chambers were harvested, stained with trypan blue, and counted. For human NK-cell chemotaxis assay, conditioned media were generated in DC media through direct ICV-PBMCs coculture at 3:1 ratio for 18 hours. A migration percentage was calculated as (total number NK cells in bottom chamber/total number NK-cell input) × 100. The NK-cell index was calculated as follows: (NK-cell migration percentage/NK-cell migration percentage from media alone group).
DC-ICV/splenocyte cocultures
DCs from DC-ICV cocultures were isolated by MACS CD11c+ selection (Miltenyi Biotec) and cocultured with naïve splenocytes at 1:5 ratio in DC medium for 24 hours, at 2 × 105 splenocytes/well in 96-well plate format. Cell-free supernatant was stored at −80°C for measurement of IFNγ. Intracellular IFNγ staining on splenocytes by intracellular FACS was also performed as described above.
Mouse models
Therapeutic treatment model.
CT26 and B16F10 peritoneal carcinomatosis in BALB/c and C57BL/6 mice, respectively, were treated with 1 × 104 ICV on day 3 after seeding 5 × 105 tumor cells intraperitoneally. For the CT26 bulky tumor model, 5 × 105 tumor cells were seeded within the peritoneum, and a treatment regimen of 6 doses of ICV was initiated following magnetic resonance (MR) scan confirmation of a tumor with a size of >3 mm. Animals were sedated with isoflurane gas, and MR scanning was performed with a 7 Tesla GE/Agilent MR 901 (GE Healthcare). For each mouse, three MR pulse sequences were used: one localizer and two fast spin echo (FSE) sequences in the coronal and axial planes. The parameters for the FSE sequences were as follows: echo train length 8, bandwidth = 16 kHz, echo time = 42 ms, repetition time = 1,500 ms, field of view = 35 mm, matrix 256 × 256, slice thickness = 1 mm. The total MR scan time per mouse was approximately 15 minutes. Follow-up MR scans were performed 1 week, 6 weeks, and 13 weeks after the start of the treatment using the same MR scan parameters.
Prophylactic treatment model.
C57BL/6 mice were vaccinated with a single dose of 1 × 103 irrB16, MG1-ICV, or MG1-IL12-ICV i.p. The following day, mice were challenged with 3 × 105 B16F10-LacZ cells i.v., sacrificed at 4 days after tumor cell injection followed by staining and quantification of lung metastases with X-gal (Bioshop) as described previously (25). The total number of lung surface metastases was determined on all lung lobes using a stereomicroscope (Leica Microsystems).
Statistical analysis
All statistical analyses were determined using GraphPad Prism 6.0 software. Statistical significance, where applicable, was determined by the Student t test or one-way ANOVA with a cutoff P = 0.05. Data are presented as mean ±SD. Survival analyses were performed according to the Kaplan–Meier method using the log-rank test, and the difference between groups was significant only when P < 0.05.
Results
Characterization of an MG1 oncolytic virus encoding murine IL12 (MG1-IL12)
A murine IL12 transgene (p70), which is composed of p35 and p40 subunits, was incorporated into the backbone of the oncolytic Maraba virus variant MG1 to create MG1-IL12 (Fig. 1A; Supplementary Fig. S1). This replication competent oncolytic virus could infect both murine and human tumor cell lines with an efficiency comparable to parental MG1, and expression of IL12 did not negatively affect viral replication or spreading (Fig. 1B and C and Supplementary Fig. S2). IL12 was detected in the culture media of B16F10 (22 pg/cell) and CT26 (180 pg/cell) cells infected with MG1-IL12 (Fig. 1D and E). Thus, MG1-IL12 could successfully infect murine tumor cells, the virus could replicate and IL12 was secreted, resulting in an MG1-IL12-ICV.
Characterization of a Maraba MG1 virus expressing murine IL12. A, A schematic of the MG1-IL12 genome. B and C, Cytotoxicity of MG1 and MG1-IL12 was assessed in B16F10 and CT26lacZ cells at the indicated MOIs 48 hours after infection using Alamar Blue. D and E, The amount of IL12 secreted from ICVs was examined 6 hours after infection by ELISA and reported as pg/mL of IL12 per cell. n = 3; error bars = SD (**, P < 0.005; ****, P ≤ 0.00005).
Characterization of a Maraba MG1 virus expressing murine IL12. A, A schematic of the MG1-IL12 genome. B and C, Cytotoxicity of MG1 and MG1-IL12 was assessed in B16F10 and CT26lacZ cells at the indicated MOIs 48 hours after infection using Alamar Blue. D and E, The amount of IL12 secreted from ICVs was examined 6 hours after infection by ELISA and reported as pg/mL of IL12 per cell. n = 3; error bars = SD (**, P < 0.005; ****, P ≤ 0.00005).
MG1-IL12-ICV enhanced NK cell–mediated tumor rejection
We previously demonstrated that infecting autologous tumor cells ex vivo with oncolytic viruses can elicit a robust immune response against established, nonpermissive, tumors in vivo (18). To determine whether MG1 and MG1-IL12 could similarly induce an immune response when used as an ICV, we i.v. injected 5 × 105 -irradiated B16F10 cells, either mock infected or infected with MG1 or MG1-IL12. Administration of ICVs i.v. is associated with a rapid and dose-dependent accumulation of injected cells that persist in the lung for up to 1 day in tumor-free animals (26). As expected, following ICV delivery we detected significantly more IL12 in lung homogenates from mice receiving MGI-IL12-ICV in comparison with animals receiving cells alone or MG1-ICV (Fig. 2A, t = 24 hours). To determine whether the increased concentrations of IL12 had any functional effect, we simultaneously measured the IL12-responsive cytokine IFNγ. In agreement with the increase in IL12, IFNγ concentrations were also elevated in the lungs of mice treated with MG1-IL12-ICV compared with mice receiving MG1-ICV or irradiated cells (Fig. 2B). Because IL12 targets both NK and T cells to promote IFNγ secretion (13), we next sought to determine which cell types were responding to treatment with our MG1-IL12-ICV. Vaccination with MG1-IL12-ICV was not found to affect the total number of T cells in the lung; however, the total number of NK cells present in the lung increased 3-fold, suggesting that MG1-IL12-ICV enhanced NK-cell recruitment (Fig. 2C and D). In addition, the total number of IFNγ and granzyme B positive NK cells increased approximately 7- and 4-fold, respectively, after injection of MG1-IL12-ICV, indicating an increase in NK-cell activation (Fig. 2E and F).
MG1-IL12-ICV enhances NK-cell cytokine secretion and cytotoxicity. A, IL12 and B, IFNγ concentrations were measured in lung homogenates of tumor naïve mice 18 hours after i.v. delivery of 5 × 105 ICVs. C–F, Total number of CD3+ T cells, NK1.1+ cells, NK1.1+IFNγ+ cells, and NK1.1+GranzymeB+ cells present in the lungs were quantified 18 hours after i.v. treatment with 5 × 105 ICVs in tumor naïve mice. G, The ex vivo chromium release cytotoxicity assay of splenocytes (effector, E) isolated from tumor naïve C57BL/6 18 hours following treatments with 1 × 106 of either ICVs toward target YAC1 (T) cells at the indicated (E:T) ratios. H, B16F10 lung micrometastases observed in the lungs 4 days after i.v. vaccination with either ICVs. All data are presented as mean of n = 3 mice per group. Error bars = SD (*, P < 0.05; **, P < 0.005; ****, P ≤ 0.00005).
MG1-IL12-ICV enhances NK-cell cytokine secretion and cytotoxicity. A, IL12 and B, IFNγ concentrations were measured in lung homogenates of tumor naïve mice 18 hours after i.v. delivery of 5 × 105 ICVs. C–F, Total number of CD3+ T cells, NK1.1+ cells, NK1.1+IFNγ+ cells, and NK1.1+GranzymeB+ cells present in the lungs were quantified 18 hours after i.v. treatment with 5 × 105 ICVs in tumor naïve mice. G, The ex vivo chromium release cytotoxicity assay of splenocytes (effector, E) isolated from tumor naïve C57BL/6 18 hours following treatments with 1 × 106 of either ICVs toward target YAC1 (T) cells at the indicated (E:T) ratios. H, B16F10 lung micrometastases observed in the lungs 4 days after i.v. vaccination with either ICVs. All data are presented as mean of n = 3 mice per group. Error bars = SD (*, P < 0.05; **, P < 0.005; ****, P ≤ 0.00005).
We measured the cytotoxic activity of NK cells against YAC-1 target cells ex vivo and found that splenocytes isolated from MG1-IL12-ICV–treated mice exhibited a significantly more YAC-1 killing (Fig. 2G). These data supported a role for MG1-IL12-ICV in promoting NK-cell recruitment to the site of delivery and a concomitant systemic activation of splenic NK cells (Supplementary Fig. S3). To determine if this effect translated into improved tumor clearance, we treated mice with B16F10 lung metastases with either irradiated cells, MG1-IL12-ICV, or MG1-ICV alone by i.v. delivery (Fig. 2H). Systemic delivery of MG1-IL12-ICV was sufficient to significantly attenuate the number of detectable lung metastasis in comparison to treatment with MG1-ICV or irradiated cells. These results suggested that MG1-IL12-ICV could stimulate NK-cell recruitment and effector function to significantly improve the antitumor efficacy of the ICVs. Unfortunately, when doses of higher than 1 × 106 infected cells were administered i.v., mice began to experience immediate respiratory distress, likely secondary to pulmonary embolism (data not shown). Given the potential risks of intravenous delivery of tumor cell vaccine in the clinical setting, including the risk of pulmonary embolism, stroke, and systemic cytokine storm, we hypothesized that intraperitoneal delivery (i.p.) of the ICV, in the setting of peritoneal carcinomatosis, would take advantage of NK-cell recruitment to the site of ICV delivery, while maintaining safety.
Enhanced NK-cell activation and improved mouse survival by MG1-IL12-ICV
Our initial findings suggest that the improved antitumor response elicited by MG1-IL12-ICV, in comparison with MG1-ICV, are in part due to potent the chemotactic properties of IL12, which contributed to the enhanced recruitment of cytotoxic NK cells to the site of delivery (Fig. 2). Therefore, we next sought to assess whether vaccinating mice i.p. with MG1-IL12-ICV could improve clearance of tumors within the peritoneal cavity and promote improved survival. More mice vaccinated with MG1-IL12-ICV survived longer than when treated with MG1-ICV (Fig. 3A), a difference that was statistically significant (P = 0.02). In addition, we observed an increased proportion of NK cells (19% vs. 49%, P = 0.0073) in the peritoneum 24 hours after i.p. vaccination as compared with MG1-ICV (Fig. 3B), similar to our previous observations with i.v. vaccination (Fig. 2D). The infiltrating NK cells also displayed a significant upregulation of the activation marker CD69, indicating that the NK cells accumulating in the peritoneum in MG1-IL12 vaccinated mice were more highly activated (Fig. 3C). To complement these findings, we performed an in vivo NK-cell cytotoxicity assay by challenging vaccinated mice with the NK-sensitive RMA-S and parental RMA tumor cell lines to investigate whether the activated NK cells that migrated into the peritoneal cavity were tumorcidal. Following vaccination with MG1-IL12-ICV, tumor cell clearance was significantly improved, demonstrating that ICV-mediated recruitment and activation of NK cells can effectively promote tumor cell clearance from the peritoneum (Fig. 3D). In support of this conclusion, the protective effect of vaccinating mice bearing B16F10 peritoneal tumors with MG1-IL12-ICV was completely abrogated upon depletion of NK cells or CD8+ T cells, further suggesting that the therapeutic benefit of this treatment strategy was dependent upon both NK-cell and CD8+ T-cell recruitment (Fig. 3E). In contrast, no significant impact upon survival was observed following depletion of CD4+ T cells. In agreement with these findings, the survival benefit of the MG1-IL12-ICV was lost upon depletion of both CD4+ and CD8+ T cells (Fig. 3E), further highlighting the importance of CD8+ T cells in mediating the response to this treatment.
MG1-IL12-ICV–dependent NK-cell activation enhances tumor cell clearance and improves survival. A, Mice were treated once with 1 × 104 cells of the indicated vaccines on day 3. n = 5 mice per group, P < 0.02, log-rank test. B and C, The effect of either ICVs upon the recruitment and activation status (CD69+) of NK cells (NK1.1+CD3−) to the peritoneal cavity, respectively, was examined by flow cytometry 24 hours after i.p. vaccination with 1 × 104 ICV. D, Mice vaccinated i.p. with 1 × 104 irrB16, MG1-ICV, or MG1-IL12-ICV and were subsequently challenged with a mixture of 1 × 106 RMA and RMA-S cells in a 1:1 ratio. The percentage of RMA-S cells rejected was calculated as described in Materials and Methods. E, Kaplan–Meier survival analysis of C57BL/6 mice-bearing i.p. B16F10 tumors seeded at 5 × 105. On day 1 following tumor seeding, NK cells, CD8+ T cells, and CD4+ T cells were depleted using antibodies to NK1.1, CD8, and CD4, respectively, and continued every other day for a total of 7 doses. n = 6 to16 mice per group.
MG1-IL12-ICV–dependent NK-cell activation enhances tumor cell clearance and improves survival. A, Mice were treated once with 1 × 104 cells of the indicated vaccines on day 3. n = 5 mice per group, P < 0.02, log-rank test. B and C, The effect of either ICVs upon the recruitment and activation status (CD69+) of NK cells (NK1.1+CD3−) to the peritoneal cavity, respectively, was examined by flow cytometry 24 hours after i.p. vaccination with 1 × 104 ICV. D, Mice vaccinated i.p. with 1 × 104 irrB16, MG1-ICV, or MG1-IL12-ICV and were subsequently challenged with a mixture of 1 × 106 RMA and RMA-S cells in a 1:1 ratio. The percentage of RMA-S cells rejected was calculated as described in Materials and Methods. E, Kaplan–Meier survival analysis of C57BL/6 mice-bearing i.p. B16F10 tumors seeded at 5 × 105. On day 1 following tumor seeding, NK cells, CD8+ T cells, and CD4+ T cells were depleted using antibodies to NK1.1, CD8, and CD4, respectively, and continued every other day for a total of 7 doses. n = 6 to16 mice per group.
IP-10 from DCs enhanced NK-cell activation and migration after MG1-IL12-ICV
Our data provide evidence for the ability of MG1-IL12-ICV to promote NK-cell activation, migration, and function. However, it was unclear whether DCs, a key mediator of NK-cell function in vivo, were involved in this process. To understand the interaction between NK cells and DCs in the presence of MG1-IL12, we quantified IFNγ production from splenocytes cultured in the presence of bone marrow–derived DCs that were either untreated or cultured with mock-, MG1-, or MG1-IL12–infected B16F10 cells. Splenocytes cultured with DCs that had been previously exposed to MG1-IL12-ICV secreted significantly more NK cell–specific IFNγ, suggesting that DCs promote NK-cell cytokine secretion (Fig. 4A and B). We next investigated the ability of DCs to promote NK-cell migration using an in vitro Transwell chemotaxis assay (Fig. 4C). The migration of NK cells across a 5-μm membrane was significantly increased by either MG1-ICV or MG1-IL12-ICV compared with the irrB16 control, with MG1-IL12-ICV inducing a higher percentage of NK cells to migrate. NK-cell migration was further increased by media conditioned in the presence of DCs, suggesting that DCs provide the stimuli for increased NK-cell activation and migration. Next, we sought to identify which chemokines commonly secreted by DCs were mediating the observed effects. Although we could not detect any effect on MCP-1 (monocytic chemotactic protein-1) and SDF-1 (stromal cell-derived factor-1) secretion, MG1-IL12-ICV induced a significant increase in IP-10 (IFN-inducible protein-10) in both in vitro and in vivo settings (Fig. 4D and E). The neutralization of IP-10 in conditioned media derived from DCs cultured with MG1-IL12-ICV significantly inhibited the migratory capacity of NK cells in vitro confirming its central role (Fig. 4F).
MG1-IL12-ICV enhances NK-cell migratory behavior and activation in vitro by stimulating secretion of IP-10 from dendritic cells. A, IFNγ secretion and B, population of IFNγ+ present in naïve splenocytes (2 × 105) was measured by ELISA and flow cytometry, respectively, following coculture (5:1) with isolated CD11c+ cells previously stimulated with either ICVs (t = 18 hours). C, Migration of isolated, naïve NK cells across a Transwell membrane (pore size = 5 μm) into cell free media conditioned as indicated was measured following a 3-hour incubation as described in Materials and Methods. D, The release of MCP-1, SDF-1, and IP-10 into the culture media following coculture of DCs (CD11c+) with the indicated treatments was determined by ELISA. E, In vivo expression levels of MCP-1, SDF-1, and IP-10 from the peritoneal exudates (18 hours following ICV treatments) were determined by ELISA. F, Quantification of the effects of IP-10 neutralization on NK-cell chemotaxis following the addition of IP-10–neutralizing antibodies (100 ng/mL) in conditioned media with DCs-(MG1-IL12-ICV) cocultures. n = 3; error bars = SD; *, P < 0.05; **, P < 0.005; ***, P ≤ 0.0005 and ****, P ≤ 0.00005).
MG1-IL12-ICV enhances NK-cell migratory behavior and activation in vitro by stimulating secretion of IP-10 from dendritic cells. A, IFNγ secretion and B, population of IFNγ+ present in naïve splenocytes (2 × 105) was measured by ELISA and flow cytometry, respectively, following coculture (5:1) with isolated CD11c+ cells previously stimulated with either ICVs (t = 18 hours). C, Migration of isolated, naïve NK cells across a Transwell membrane (pore size = 5 μm) into cell free media conditioned as indicated was measured following a 3-hour incubation as described in Materials and Methods. D, The release of MCP-1, SDF-1, and IP-10 into the culture media following coculture of DCs (CD11c+) with the indicated treatments was determined by ELISA. E, In vivo expression levels of MCP-1, SDF-1, and IP-10 from the peritoneal exudates (18 hours following ICV treatments) were determined by ELISA. F, Quantification of the effects of IP-10 neutralization on NK-cell chemotaxis following the addition of IP-10–neutralizing antibodies (100 ng/mL) in conditioned media with DCs-(MG1-IL12-ICV) cocultures. n = 3; error bars = SD; *, P < 0.05; **, P < 0.005; ***, P ≤ 0.0005 and ****, P ≤ 0.00005).
MG1-IL12-ICV was effective in treating established peritoneal disease in mice
Our findings suggest that MG1-IL12-ICV can significantly slow the outgrowth of B16F10 tumors within the peritoneal compartment by stimulating the recruitment of activated NK cells. Because peritoneal carcinomatosis is a common presentation for late-stage gastrointestinal and gynecological malignancies, we sought to determine whether MG1-IL12-ICV could provide therapeutic benefit in a clinically relevant model of colon cancer (CT26) with peritoneal disease at time of treatment. To accomplish this, BALB/c mice were seeded with CT26 tumor cells (Fig. 5A). Three days later, mice were treated with a single dose of irradiated cells alone, virus alone, or the ICVs. Mice treated with irradiated CT26 cells, MG1, MG1-IL12, and MG1-ICV all had significantly lower median survival times and increased peritoneal tumor burden in comparison with mice receiving MG1-IL12-ICV (>90%, 26 of 28 mice survived more than 200 days; Fig. 5A and B). The cured mice developed a long-lasting immunity such that when the surviving mice were rechallenged with 5 × 105 CT26 cells on the flank, 148 days after treatment, they rejected the tumors (5/5 mice). However, this antitumor memory immune response was specific to CT26 tumors: all mice developed tumors (5/5 mice) when challenged with syngeneic 4T1 tumor cells on the opposite flank. The route of vaccination was critical. MG1-IL12-ICV had superior efficacy when given intraperitoneally, compared with intravenous or subcutaneous injections (Supplementary Fig. S4).
MG1-IL12-ICV improves survival and clearance of CT26 tumors. A, BALB/c mice received a single dose of either virus alone (MG1 or MG1-IL12, MOI = 1 × 105 pfu/mL), an ICV (MG1-ICV or MG1-IL12-ICV, 1 × 104 infected cells), irradiated cells alone (irrCT26lacZ) or left untreated. 3 days following i.p. tumor seeding (5 × 105), n = 10–12 mice per group, P < 0.0005, log-rank test. B, Weight of peritoneal organs including any associated tumor burden was measured 16 days after tumor seeding. C, Kaplan–Meier survival analysis for BALB/c mice-bearing tumors greater than 3 mm that received two doses of either ICVs each week over a 3-week period (total number of intended doses per animal = 6). n = 7–14 mice per group, P < 0.05, log-rank test. D, Coronal T2-weighted FSE MR images (TR/TE 1,500/42 ms, 256 × 256 matrix, 35 mm FOV, 1-mm slice thickness) for one representative mouse bearing class I tumor (>8 mm). (1) Day 10, bulky tumor mass with intermediate T2 signal evident prior to treatment. (2) Day 17, an increase in tumor size is evident, (3 and 4) are two different views of the same tumor on week 6 demonstrating a marked tumor size reduction 6 weeks after treatment initiation, and (5 and 6) at 13 weeks after treatment initiation with no residual tumor signal at site of the previous large peritoneal tumor deposit.
MG1-IL12-ICV improves survival and clearance of CT26 tumors. A, BALB/c mice received a single dose of either virus alone (MG1 or MG1-IL12, MOI = 1 × 105 pfu/mL), an ICV (MG1-ICV or MG1-IL12-ICV, 1 × 104 infected cells), irradiated cells alone (irrCT26lacZ) or left untreated. 3 days following i.p. tumor seeding (5 × 105), n = 10–12 mice per group, P < 0.0005, log-rank test. B, Weight of peritoneal organs including any associated tumor burden was measured 16 days after tumor seeding. C, Kaplan–Meier survival analysis for BALB/c mice-bearing tumors greater than 3 mm that received two doses of either ICVs each week over a 3-week period (total number of intended doses per animal = 6). n = 7–14 mice per group, P < 0.05, log-rank test. D, Coronal T2-weighted FSE MR images (TR/TE 1,500/42 ms, 256 × 256 matrix, 35 mm FOV, 1-mm slice thickness) for one representative mouse bearing class I tumor (>8 mm). (1) Day 10, bulky tumor mass with intermediate T2 signal evident prior to treatment. (2) Day 17, an increase in tumor size is evident, (3 and 4) are two different views of the same tumor on week 6 demonstrating a marked tumor size reduction 6 weeks after treatment initiation, and (5 and 6) at 13 weeks after treatment initiation with no residual tumor signal at site of the previous large peritoneal tumor deposit.
Next, we measured the effects of treatment in established bulky tumors. Between days 10 and 17 after implantation, tumors were visualized by MRI and mice bearing significant tumor masses (class 1 > 8 mm and class 2 > 3 mm) were randomly allocated into a treatment group. Mice were treated with 6 doses of irradiated cells, MG1-ICV, or MG1-IL12-ICV administered over a 3-week period (Fig. 5C). Despite the lethal tumor burden, evident by the loss of all animals treated with irradiated cells by day 15, MG1-IL12-ICV provided complete protection (21/21 survived > 100 days, study ongoing). Follow-up MRI scans, which confirmed the presence of large tumor masses at the early stages of treatment, were dramatically reduced at later time points (Fig. 5D). This efficacy is dependent on viral replication, in that a nonreplicative MG1-IL12 virus (G-less MRB-IL12-ICV) was not efficacious in reducing tumors, despite the fact that in vitro IL12 expression from G-less MRB-IL12-ICV was significantly higher than MG1-IL12-ICV (Fig. 5C and Supplementary Fig. S5). Furthermore, the injection of exogenous recombinant IL12 protein (20 ng) along with MG1-ICV was not superior to MG1-ICV (Fig. 5C). Collectively, these results demonstrated that MG1-IL12-ICV is an effective approach for promoting the clearance of large, established tumors within the peritoneum in a murine model of peritoneal carcinomatosis.
MG1-IL12-ICV enhanced human NK-cell cytotoxicity and migratory capacity
Given the fact that murine p40 and p35 subunits of IL12 share 70% and 60% homology with their human counterparts, respectively, they can functionally activate human NK and T cells (27). We next sought to confirm that the vaccine could elicit a similar effect on human NK cells ex vivo. To accomplish this, irradiated SW620 colon cancer cells were infected with MG1 or MG1-IL12 and cultured with peripheral blood mononuclear cells (PBMC) isolated from a healthy donor as part of a perioperative blood collection protocol (approved by the Ottawa Health Science Network Research Ethics Board #2011884). In agreement with our previous findings, MG1-IL12-ICV resulted in a significant increase in the expression of CD69, an established marker of NK-cell activation, in the NK-cell (CD56+CD3−) subset of PBMCs (Fig. 6A). In addition, IP-10 chemotactic protein secretion was also significantly increased in the supernatant of PBMCs coculture with MG1-IL12-ICV. This supernatant enhanced the migration of NK cells in the ex vivo Transwell system, suggesting that the MG1-IL12-ICV vaccine elicits similar responses from NK cells of human and murine origin (Fig. 6B and C). Finally, stimulating PBMCs with MG1-IL12-ICV resulted in an increased cytotoxic activity toward K562 target tumor cells, suggesting that the activation and enhanced migratory capacity of human NK cells cultured in the presence of MG1-IL12-ICV is associated with increased ability to eradicate tumor cells (Fig. 6D and E and Supplementary Fig. S6). Together, these results provide support for our hypothesis that autologous infected tumor cell vaccines may provide a much-needed therapeutic benefit in the treatment of patients with PC.
MG1-IL12-ICV enhances human NK cells migration and activation in vitro. A, Percentage of CD56+ CD3− CD69+ NK cells in PBMCs cultured with irrSW620, MG1-ICV, or MG1-IL12-ICV was quantified by flow cytometry (t = 18 hours). B, The stimulatory effects of cell free conditioned media prepared from the indicated groups upon human NK-cell migration was assessed at 3 hours as described in Materials and Methods by a flow cytometric method. C, IP-10 secretion was measured by ELISA following culture of PBMCs media conditioned by the indicated groups. D and E, Ex vivo cytotoxicity of human NK cells from two individual cancer patients were isolated and cultured with either ICVs (t = 18 hours). Cytotoxic potential of NK cells against K562 tumor cells was quantified by 51Cr release as described in Materials and Methods. Data represent means of triplicate wells; error bars = SD. *, P < 0.05 and **, P < 0.005.
MG1-IL12-ICV enhances human NK cells migration and activation in vitro. A, Percentage of CD56+ CD3− CD69+ NK cells in PBMCs cultured with irrSW620, MG1-ICV, or MG1-IL12-ICV was quantified by flow cytometry (t = 18 hours). B, The stimulatory effects of cell free conditioned media prepared from the indicated groups upon human NK-cell migration was assessed at 3 hours as described in Materials and Methods by a flow cytometric method. C, IP-10 secretion was measured by ELISA following culture of PBMCs media conditioned by the indicated groups. D and E, Ex vivo cytotoxicity of human NK cells from two individual cancer patients were isolated and cultured with either ICVs (t = 18 hours). Cytotoxic potential of NK cells against K562 tumor cells was quantified by 51Cr release as described in Materials and Methods. Data represent means of triplicate wells; error bars = SD. *, P < 0.05 and **, P < 0.005.
Discussion
ICVs are a powerful personalized cancer therapy platform that exploits the efficacy of oncolytic viruses, autologous cell vaccines, and immunostimulatory cytokines, while overcoming key barriers. It is becoming increasingly clear that oncolytic viruses can direct an innate and adaptive immune response against the cancer at both the site of administration and at distant sites (28). In order for this to occur, the oncolytic virus must infect and express viral genes within a cancer cell (29). Ensuring viral delivery and robust infection of cancer cells can pose a significant barrier to efficacy (30). Indeed, the recently approved GM-CSF expressing herpes virus (talmogene laherparepvec) is directly injected into an accessible tumor, to avoid the problems of viral delivery, effectively turning the oncolytic virus into an in situ ICV (31).
Vaccination of patients with their own cancer cells (autologous cell vaccine) has been tried in the past with variable success (32, 33). Most have used mixing the whole cell vaccine with nonspecific adjuvants, such as BCG, but difficulties in overcoming immune suppression within the tumor microenvironment have limited results (34). Nonetheless, clinical trials have consistently shown that survival is significantly better in those patients that are able to mount an immune response to the whole cell vaccine, suggesting that when an immune response is generated, prognosis is improved (35). Cytokines, such as IL12, have also been used to direct an antitumor immune response but the short half-life of these cytokines, when administered as proteins, and the dose-limiting toxicities encountered following systemic administration have diminished their potential effectiveness (36). That said, the strong immunologic rationale for IL12-based vaccines continues to drive the development of novel experimental approaches in numerous laboratories worldwide. Indeed, more than 20 active anticancer clinical trials are listed on clinicaltrials.gov covering just IL12.
MG1 has a 100-fold higher maximum tolerated dose as compared with wild-type Maraba, and when compared with other oncolytic viruses, it has an extremely wide tropism for infection of cancer cells, a very rapid life cycle with the release of more infectious progeny and therefore rapid and robust transgene expression (20, 37). Additionally, the lack of preexisting neutralizing antibodies in human populations, a major hurdle that is associated with many other oncolytic viruses, warrants further development for clinical applications (20, 38–41). Our multimodal approach that utilizes the engineered oncolytic MG1-IL12 virus in the form of ICV has a combined effect on tumors by (i) infecting and directly lysing the tumor tissues as an initial tumor debulking process, (ii) recruiting the innate (NK cell) immune system to the tumor microenvironment by overexpressing IL12, (iii) activating and maturing DCs through expression of viral proteins and IFNγ secreted by activated NK cells, and (iv) the activated DC, in turn secrete IP-10, which further recruits NK cells to the vaccination site (Supplementary Fig. S7). Direct intratumoral delivery (i.p. injection in the case of peritoneal carcinomatosis) of MG1-IL12-ICV has superior efficacy, and this may be partly explained by the chemotactic properties of IL12. Utilizing IL12 in oncolytic viral therapy improves the therapeutic index in murine models (14), but we now clearly define the role of IP-10, secreted from DCs, in NK-cell chemotaxis and activation. This recruitment synergizes with the IL12-mediated direct effects on cytotoxicity and activation of NK cells, leading to a robust targeting and killing of tumor cells in the peritoneal carcinomatosis model. However, for the antitumor NK response to be curative, peripheral NK cells need to migrate to the site of tumor. Although not investigated in our model, it is possible that i.p. delivery allows for antigen presentation in the multitude of intra-abdominal and retroperitoneal lymph nodes, which has previously been documented to be essential for efficacy (42).
Although NK cells appear paramount for the efficacy of the MG1-IL12-ICV, clearly a specific adaptive immune response is mounted during curative treatment, as illustrated by the rejection of implanted CT26 tumors, but not of syngeneic implanted 4T1 tumors. Moreover, the pattern of tumor response visualized by MRI, with pseudoprogression seen at 1 week and ongoing tumor responses beyond 6 weeks, is suggestive of an adaptive T-cell response (43, 44). In addition, the abrogated efficacy of MG1-IL12-ICV in B16F10 tumor model that was observed following CD8+ T cells depletion (Fig. 3E) further suggests the involvement of CD8+ T cells to the observed efficacy following MG1-IL12-ICV treatment. In short, oncolytic viruses have demonstrated a potent ability to generate an antitumor T cell–mediated immune response but, in general, it is thought to require viral infection of cancer cells (45, 46). The ICV platform optimized tumor cell infection by incubating tumor cells ex vivo at a high multiplicity of infection. This may explain, in part, the improved efficacy of the MG1-IL12 ICV over viral administration alone.
In this era of emerging immunotherapies, including checkpoint inhibitors, combining treatments with an ICV hold significant promise (29). However, as with the addition of IL12 to MG1, it is important to consider how immune activation might hinder viral replication and oncolysis and transgene expression (47). Indeed in various contexts, immunotherapies such as anti-CTLA4 can be both complementary and inhibitory towards oncolytic viral therapy (48, 49). In the present study, we have shown that the addition of IL12 to the oncolytic rhabdovirus MG1 improves efficacy by recruiting and activating NK cells to the site of tumor burden. Further studies exploring relative contributions of viral oncolysis and innate and adaptive immunity will allow us to harness the immunotherapeutic potential of the ICV platform.
Disclosure of Potential Conflicts of Interest
D.F. Stojdl has ownership interest (including patents) in Turnstone Biologics and is a consultant/advisory board member for the same. J.C. Bell is in the Board of Directors of, has ownership interest (including patents) in, and is a consultant/advisory board member for Turnstone Biologics. No potential conflicts of interest were disclosed by the other authors.
Authors' Contributions
Conception and design: A.A. Alkayyal, L.-H. Tai, J.C. Bell, R.C. Auer
Development of methodology: A.A. Alkayyal, J. Zhang, C. Lefebvre, A.A. Ananth, A.B. Mahmoud, G.O. Cron, B. Macdonald, R.C. Auer
Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): A.A. Alkayyal, L.-H. Tai, M.A. Kennedy, S. Sahi, G.O. Cron, B. Macdonald, D.F. Stojdl, R.C. Auer
Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): A.A. Alkayyal, L.-H. Tai, M.A. Kennedy, G.O. Cron, B. Macdonald, R.C. Auer
Writing, review, and/or revision of the manuscript: A.A. Alkayyal, L.-H. Tai, M.A. Kennedy, A.A. Ananth, G.O. Cron, B. Macdonald, R.C. Auer
Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): A.A. Alkayyal, C. Tanese de Souza, J. Zhang, S. Sahi, A.B. Mahmoud, A.P. Makrigiannis, R.C. Auer
Study supervision: A.A. Alkayyal, R.C. Auer
Other (pathology-slides exam and interpretation): E.C. Marginean
Grant Support
This work was supported by grants from the Terry Fox Research Institute, Canadian Cancer Society Research Institute, Canadian Institute of Health Research New Investigator Award (to R.C. Auer), University of Tabuk scholarship (to A.A. Alkayyal), Canadian Institutes of Health Research (to L.-H. Tai), and Ontario Ministry of Research and Development Early Researcher Award (to A.A. Ananth and J. Zhang).
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.