Research ArticleCancer

MEK inhibition enhances oncolytic virus immunotherapy through increased tumor cell killing and T cell activation

See allHide authors and affiliations

Science Translational Medicine  12 Dec 2018:
Vol. 10, Issue 471, eaau0417
DOI: 10.1126/scitranslmed.aau0417

Enabling immunotherapy with an oncolytic virus

Immunotherapy treatments have been pioneered in the setting of melanoma, and although subsets of patients are able to survive long term, some tumors are resistant. Bommareddy and colleagues inquired whether combining certain treatments may lead to even better therapeutic responses. They examined two approved therapies, an oncolytic herpesvirus and a MEK inhibitor. This combination led to increased survival in mouse melanoma models, which was further extended with the addition of PD-1 blockade. The immune response behind this enhanced antitumor activity was dissected. The results of these studies suggest that combining these treatments may be beneficial for patients with melanoma.

Abstract

Melanoma is an aggressive cutaneous malignancy, but advances over the past decade have resulted in multiple new therapeutic options, including molecularly targeted therapy, immunotherapy, and oncolytic virus therapy. Talimogene laherparepvec (T-VEC) is a herpes simplex type 1 oncolytic virus, and trametinib is a MEK inhibitor approved for treatment of melanoma. Therapeutic responses with T-VEC are often limited, and BRAF/MEK inhibition is complicated by drug resistance. We observed that the combination of T-VEC and trametinib resulted in enhanced melanoma cell death in vitro. Further, combination treatment resulted in delayed tumor growth and improved survival in mouse models. Tumor regression was dependent on activated CD8+ T cells and Batf3+ dendritic cells. We also observed antigen spreading and induction of an inflammatory gene signature, including increased expression of PD-L1. Triple therapy with the combination of T-VEC, MEK inhibition, and anti–PD-1 antibody further augmented responses. These data support clinical development of combination oncolytic viruses, MEK inhibitors, and checkpoint blockade in patients with melanoma.

INTRODUCTION

Melanoma is a metastatic tumor arising from melanocytes located in the stratum basale of the epidermis, mucosal membranes, and middle layer of the uvea. Metastatic melanoma has historically been associated with dismal prognoses; however, systemic therapies have transformed patient outcomes over the past decade, largely due to advances in molecular therapy targeting the RAS–RAF–MEK–ERK mitogen-activated protein kinase (MAPK) pathway in patients with tumors that harbor BRAF V600E/K mutations, and by immunotherapy, most notably with immune checkpoint blockade targeting cytotoxic T lymphocyte antigen 4 and programmed cell death 1 (PD-1) (1). Combination approaches within drug classes have shown improved therapeutic benefit, but treatment is associated with drug resistance in the case of MAPK inhibitors and increased toxicity with checkpoint blockade (2, 3). New combination strategies with agents that enhance therapeutic responses while limiting toxicity have become a high priority for drug development in melanoma.

Oncolytic viruses are a class of cancer drugs that use native or genetically modified viruses that selectively replicate in tumor cells (4). Oncolytic viruses mediate therapeutic activity through multiple mechanisms, including: direct immunogenic tumor cell killing; release of soluble tumor antigens, danger signals, and type 1 interferons (IFNs); and induction of host antitumor immunity (4). Talimogene laherparepvec (T-VEC) is an oncolytic herpes simplex virus, type 1 (HSV-1) encoding granulocyte-macrophage colony-stimulating factor (GM-CSF) and is approved for local treatment of advanced melanoma that has recurred after initial surgery (5). T-VEC has recently been combined with immune checkpoint inhibitors and was associated with improved response rates without an increase in immune-related adverse events (6, 7). The ability of oncolytic viruses to correct various aspects of tumor-mediated immune suppression and the favorable therapeutic window suggest that oncolytic viruses may be ideal candidates for combination approaches with other systemic agents as well (8).

Approximately 40 to 50% of cutaneous melanomas harbor mutations in BRAF, which serve as oncogenic drivers of the MAPK pathway promoting tumor progression. Small-molecule inhibitors of BRAF and MAPK kinase (MEK) in treatment-naïve patients with melanoma whose tumors harbor V600E or V600K BRAF mutations contribute to significant improvements in relapse-free and overall survival (9). Preclinical studies have further suggested improved therapeutic activity of the combination of MAPK inhibition and immune checkpoint blockade (10). Although these findings await further clinical validation, the potential for combining MAPK inhibition with immunotherapy is particularly appealing because MAPK inhibitors act directly on mutated tumor cells, resulting in release of soluble tumor-associated antigens, whereas immunotherapy acts on immune cells to promote innate and adaptive immune responses and/or prevent suppression of host antitumor immunity (11). Preclinical studies have suggested improvements in therapeutic antitumor activity between oncolytic viruses and MEK inhibition in a murine breast cancer model (12). The combination of MEK inhibition and oncolytic viruses has not been tested in melanoma and has not yet entered clinical trials. Thus, we hypothesized that MEK inhibition would improve oncolytic virus responses in melanoma and sought to test this with currently approved agents in melanoma.

RESULTS

Combination of MEK inhibition and oncolytic virus treatment augments oncolytic activity and viral replication in human and mouse melanoma cell lines

We sought to investigate whether combining T-VEC and MAPK inhibition can augment tumor cell killing in melanoma. T-VEC was able to replicate in and kill melanoma cell lines harboring BRAF V600E mutations and wild-type N-Ras (SK-MEL-28 and SK-MEL-5; fig. S1, A and B) and those cells with wild-type BRAF but an NRAS Q61R mutation (SK-MEL-2; fig. S1C). Infected cells exhibited dose-dependent cytotoxicity after viral infection at doses starting at 0.003 multiplicity of infection (MOI) (fig. S1, A to C). In addition, the BRAF-mutated murine D4M3A cell line (13) was susceptible to T-VEC at high doses (MOI ≥ 1; fig. S1D). Cytotoxicity was increased in all cell lines when they were pretreated with trametinib, a selective MEK inhibitor (MEKi; Fig. 1, A to D, left panels). Independent assays with vemurafenib, a selective BRAF inhibitor, enhanced T-VEC–mediated cytotoxicity in BRAF V600E–mutated SK-MEL-28 and SK-MEL-5 cell lines, but not in the BRAF wild-type SK-MEL-2 cell line (fig. S1, E to G). Increased viral replication was confirmed by plaque assay (Fig. 1, A to D, right panels) and Western blot showing increased amounts of HSV-1 glycoprotein D during combination treatment in the SK-MEL-28 cell line (Fig. 1E).

Fig. 1 MEK inhibition augments T-VEC–mediated cell lysis in vitro and increases viral replication.

Cell viability determined by MTS assay. (A to D) Left: Cells were treated with either T-VEC alone or trametinib (MEKi) or a combination of T-VEC and MEKi . Right: HSV-1 titers as measured by plaque assay from cells treated with either T-VEC alone (blue bar) or T-VEC and MEKi (purple bar). Only significant differences are indicated. PFU, plaque-forming units. (E) Western blot of cell lysate collected at 24 hours after mT-VEC (0.1 MOI) infection of SK-MEL-28, mock infected, MEKi (10 nM), or combination treatment. (F) Infection metric analysis by LumaCyte (left panel) of SK-MEL-28 cells (mock), treated with 10 nM trametinib (MEKi), 1 MOI of T-VEC, or trametinib and T-VEC. The right panel shows a time course for untreated cells (dotted black line), or those treated with 0.1 MOI of T-VEC (dotted blue line) or 1 MOI of T-VEC (solid blue line). (G) PCA of the infection metric. Each experiment was performed in triplicate and is conducted at least twice with similar results. Data are presented as means ± SEM, and statistical differences between groups were measured by using two-tailed Student’s t test. *P < 0.05, **P < 0.01, ***P < 0.001.

To confirm viral replication within infected cells, we used single-cell laser radiance-based quantitative technology (14) that allows detection of viral infection at a single-cell level (fig. S2A). As shown in Fig. 1F, the infection metric was increased at 18 hours for virally infected cells, with the highest value seen in cells treated with T-VEC and MEKi (Fig. 1F, left). A time course analysis on cells infected with T-VEC at low (0.01) or high (1.0) MOI or uninfected control cells showed the expected rapid increase in infection metric for cells infected with 1 MOI, whereas cells infected with 0.01 MOI demonstrated a delayed increase in infection metric at 36 hours when more virus had replicated (Fig. 1F, right). Principal components analysis (PCA) based on cell size (F1) and radiance (F2) was able to differentiate each of the treated cell populations (Fig. 1G).

Combination of T-VEC and MEK inhibition inhibits tumor growth in a melanoma xenograft model

Next, we sought to determine whether T-VEC and MEK inhibition had therapeutic activity in vivo. We used a murine xenograft model using the human SK-MEL-28 cell line (Fig. 2A). Delayed tumor growth was observed with MEK inhibition alone and T-VEC alone, but combination treatment was associated with a significant decrease in tumor growth and tumor regression compared to mock or monotherapy treatments (P < 0.001; Fig. 2B). Previously, MEK inhibition was shown to induce tumor cell apoptosis (15); therefore, we sought to determine how cells were killed in this model. We found that the combination of T-VEC and MEKi is associated with fewer proliferating cells, based on Ki67 immunostaining, compared to either treatment alone (Fig. 2C). HSV-1 in tumors was detected by immunostaining for the HSV-1 glycoprotein D, which was seen in the T-VEC-alone–treated tumors and significantly increased in tumors of mice treated with T-VEC and MEKi (Fig. 2D). We also observed decreased levels of phosphorylated extracellular signal–regulated kinase (pERK) in tumor cells treated with MEKi and in tumors treated with T-VEC alone; pERK levels were further decreased in tumors treated with the combination (Fig. 2E). Last, whereas T-VEC treatment alone resulted in significant increase in caspase 3 cleavage compared to mock treatment, combination therapy resulted in higher tumor cell apoptosis in vivo (Fig. 2F).

Fig. 2 MEK inhibition enhances T-VEC–induced inhibition of human melanoma xenograft growth in vivo and promotes tumor cell apoptosis.

(A) NSG (nonobese diabetic/severe combined immunodeficient interleukin-2 receptor gamma chain null) mice (n = 5 per group) were implanted subcutaneously with human melanoma SK-MEL-28 cells (8 × 106) on day 0 (d0), treated via intratumoral injection with sterile water or T-VEC (1 × 105 PFU) on days 35, 40, and 45, and MEKi (trametinib; 0.5 mg/kg) or vehicle [0.2% Tween 80 and 0.5% hydroxypropyl methyl cellulose (HPMC)] was given from days 35 to 43 via oral gavage. Red arrows indicate days when T-VEC was injected, and the blue bar on top indicates days of trametinib (MEKi) treatment. (B) Mean tumor area. (C) Representative images obtained from immunohistochemical staining of tumors for Ki67 at day 36, (D) HSV-1 gD, (E) pERK1/2, and (F) cleaved caspase 3. Right panels indicate quantification of positive cells. Scale bars are as indicated. Each experiment was repeated at least twice with similar results. Data are presented as means ± SEM, and statistical differences between groups were measured by using one-way analysis of variance (ANOVA). *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001. Only significant differences are indicated.

To confirm melanoma cell apoptosis, we treated SK-MEL-28 cells in vitro and found an increase in annexin V staining in cells treated with the combination compared to monotherapy or mock treatment (fig. S3, A and B), and this effect was partially blocked by a pan-caspase inhibitor (Z-VAD) (fig. S3C). Further, there was increased cleaved poly(ADP-ribose) polymerase in tumor cells treated with both T-VEC and trametinib (fig. S3D). Collectively, these data demonstrate that the combination of T-VEC and MEK inhibition can delay melanoma xenograft growth in vivo and that treatment is associated with decreased tumor cell viability and increased apoptosis.

Combination of T-VEC and MEK inhibition enhances therapeutic effectiveness and improves survival in the immune-competent D4M3A murine melanoma model

To determine the effects of the combination of T-VEC and trametinib (MEKi) in an immune-competent D4M3A BRAF V600E melanoma model, we used a modified T-VEC encoding murine GM-CSF (mT-VEC), as described in Materials and Methods. D4M3A cells are susceptible to T-VEC infection and killing (fig. S4A) and exhibit up-regulation of pERK, characteristic of BRAF V600E–mutated cells (fig. S4B). In D4M3A tumor–bearing mice, mT-VEC alone exhibited no significant delays in tumor growth (Fig. 3, B and C), whereas MEKi (MEKi) alone showed significant delays in tumor growth (Fig. 3, B and C). Combination treatment, however, was associated with significant tumor growth inhibition and improved survival with complete tumor eradication in four of nine (44%) mice (Fig. 3C). Mice with complete tumor regression remained tumor free (Fig. 3C) and were rechallenged with twice the number of D4M3A cells implanted in the opposite flank (left). In this experiment, 70% (five of seven) of mice completely rejected rechallenged tumor (Fig. 3D). We further observed a significant increase in tumor-infiltrating CD8+ T cells in mice treated with combination therapy (Fig. 3E). CD8+ T cells exhibited increased levels of interferon-γ (IFN-γ), granzyme B, and Ki67 (Fig. 3E), indicative of an activated cytotoxic phenotype (Fig. 3E). The increased number of CD8+ T cells was further confirmed by immunohistochemistry (Fig. 3, F and G). Both mT-VEC and MEKi demonstrated an increase in CD8+ T cells after treatment, which were further increased by combination therapy (Fig. 3, F and G). There was no significant change in the total number of CD3+CD4+ T cells (Fig. 3H), but there was a decrease in CD4+FoxP3+ regulatory T cells (Tregs) in mice treated with mT-VEC alone or in combination with MEKi (Fig. 3H). This resulted in a significant increase in the CD8+/Treg ratio in mice treated with mT-VEC alone and in combination with MEKi (Fig. 3H).

Fig. 3 Combination of T-VEC and MEK inhibition enhances tumor regression in an immune-competent murine melanoma model, promotes recruitment of CD8+ T cells, and establishes long-term memory.

(A) Treatment schema: Red arrows indicate days of mT-VEC treatment, and the blue bar on top indicates trametinib (MEKi) treatment. (B) Mean tumor area of mice from treated groups at day 45. (C) Survival of mice. (D) Rechallenge of mice cured in (C). (E) Flow cytometry analysis of tumors at day 24. Bar graphs (n = 6) indicating the percentage of positive CD8 T cells, CD8+IFN-γ, CD8+granzyme B, and CD8+Ki67 T cells, respectively. (F) Immunohistochemical staining of CD8+ T cells in the tumor. Scale bar is as indicated. (G) Quantification of CD8+ cells. (H) Bar graph indicating CD4+ and CD4+FoxP3 (Tregs) and ratio of CD8+ T cells to Tregs. Each experiment was conducted at least twice with similar results. Data are presented as means ± SEM, and statistical differences between groups were measured by using one-way ANOVA. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001. Only significant differences are indicated.

T-VEC and MEKi combination therapy is CD8+ T cell-dependent

To determine which immune cells are involved in the antitumor activity, we repeated the in vivo tumor experiments using depletion antibodies against CD4+ and CD8+ T cells and liposomal clodronate to deplete macrophages. All cell depletions were confirmed by fluorescence-activated cell sorting (FACS) analysis of splenocytes (fig. S5, A and B). Mice bearing D4M3A tumors were treated as described in the survival experiments in Fig. 3, and depletion antibodies were injected as shown in Fig. 4A and described in Materials and Methods. Neither macrophage depletion nor CD4+ T cell depletion significantly affected antitumor activity, but CD8+ T cell depletion completely abrogated the antitumor activity and survival benefit (Fig. 4, B and C). FACS analysis confirmed the loss of CD4+ and CD8+ T cells in tumors collected from mice treated with immune cell–depleting antibodies (Fig. 4, D and E). A compensatory increase in CD4+ T cells in the tumor microenvironment of mice depleted of CD8+ cells (Fig. 4D) and in CD8+ T cells in tumors of mice depleted of CD4+ cells (Fig. 4E) was seen.

Fig. 4 Depletion of CD8+ T cells abrogates the effects of T-VEC and MEKi combination therapy.

(A) C57BL/6J mice (n = 5 per group) were implanted with D4M3A murine melanoma cells, and mice were treated as described in Materials and Methods. Red arrows indicate days of mT-VEC treatment, the blue bar on top indicates days of trametinib (MEKi) treatment, and black arrows indicate days where depletion antibodies against CD4, CD8, and clodronate were injected. (B) Mean tumor area of mice treated from different groups as indicated. (C) Survival of mice. (D and E) Flow cytometric analysis of tumor-infiltrating T cells on day 24. Bar graphs show the percentage of (D) CD45+CD3+CD4+ and (E) CD45+CD3+CD8+ cells. Each experiment was repeated at least twice with similar results. Data are presented as means ± SEM, and statistical differences between groups were measured by using one-way ANOVA. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001. Only significant differences are indicated.

Combination treatment with T-VEC and MEKi augments melanoma antigen–specific T cell responses

We sought to further characterize the antigen specificity of the CD8+ T cell responses in mice treated with mT-VEC/MEKi combination therapy. Initially, flow cytometry using major histocompatibility complex class I (MHC-I) dextramers for two defined melanoma antigens, gp100 and TRP2, and one viral antigen, HSV-1 gB, was used to determine antigen specificity of tumor-infiltrating CD8+ T cells during T-VEC treatment alone in a time course study (fig. S6). We saw an initial increase of HSV-1 gB–specific CD8+ T cells at day 19, which plateaued by day 24 (fig. S6). gp100- and TRP2-specific CD8+ T cells emerged between days 19 and 24 (fig. S6). mT-VEC treatment induced HSV-1 gB–specific CD8+ T cells (Fig. 5A), and the combination of mT-VEC and MEKi resulted in a significant increase in the relative frequency of tumor-infiltrating HSV-1 gB–specific CD8+ T cells (Fig. 5A). We also observed an increase in gp100- and TRP2-specific CD8+ T cells during combination treatment (Fig. 5, B and C). Although the increase in melanoma-specific CD8+ T cells was especially high within the tumor-infiltrating lymphocyte population, we did not detect high frequencies of HSV-1 gB–specific CD8+ T cells in the spleen of treated animals but did observed a minimal, but significant, increase in both gp100- and TRP2-specific CD8+ T cells in the spleen (fig. S7). These data suggest that T-VEC and MEKi can induce antigen spreading.

Fig. 5 Combination of T-VEC and MEK inhibition induces viral-specific CD8+ T cells and increases melanoma antigen–specific CD8+ T cell responses.

C57BL/6J mice implanted subcutaneously in the right flank with 3 × 105 D4M3A cells and treated with mT-VEC (1 × 106 PFU) or sterile water intratumorally for three doses on days 15, 19, and 22 and/or trametinib (0.5 mg/kg) or vehicle (0.2% Tween 80 and 0.5% HPMC) orally once daily on days 15 to 19. Tumors were harvested on day 24, and cells were dissociated and analyzed by flow cytometry. Percentages of live CD45+ cells, CD3+ cells, and CD3+CD8+ subsets from the mock, T-VEC monotherapy, MEKi monotherapy, and T-VEC + MEKi combination groups were analyzed and compared. Tumor-infiltrating CD8+ T cells were analyzed with (A) HSV-1–specific H-2Kb–HSV-1 gB dextramer, (B) melanoma antigen–specific H-2Db–gp100 dextramer, and (C) H-2Kb–TRP2 dextramer. Quantitative analysis is shown in the bar graphs on the right. These experiments were conducted at least twice with similar results. Data are presented as means ± SEM, and statistical differences between groups were measured by using one-way ANOVA. *P < 0.05, **P < 0.01, ****P < 0.0001. Only significant differences are indicated.

Combination treatment with T-VEC and MEKi is dependent on Batf3+ dendritic cells

To determine the role of CD8+CD103+ dendritic cells (DCs) in mediating antitumor immunity (16, 17), we implanted D4M3A tumors into Batf3−/− mice. The lack of CD8+CD103+ DCs in Batf3−/− mice did not alter the ability to establish D4M3A tumors (fig. S8, A and B). Tumors from Batf3−/− mice had significantly reduced frequency of CD45+MHCII+CD11c+CD103+ cells, as well as CD45+MHCII+CD11c+CD8+ cells after combination therapy (fig. S8C). To determine the effects of Batf3+ DCs on combination therapy, D4M3A tumors were implanted in C57BL/6J and Batf3−/− mice and treated as described in Fig. 3. Although combination treatment resulted in delayed tumor growth in C57BL/6J mice, as previously seen (Fig. 3C), this effect was significantly diminished in Batf3−/− mice (Fig. 6, A and B).

Fig. 6 Batf3+ DCs play a role in the antitumor activity and antigen spreading associated with combination treatment with T-VEC and MEK inhibition.

C57BL/6J mice (B6, n = 7) and Batf3−/− mice (n = 7) were implanted with D4M3A murine melanoma cells and either mock treated or treated with T-VEC and trametinib as described in Materials and Methods. (A) Survival of mice. (B) Mean tumor area. (C to F) Flow cytometry analysis of tumors obtained from B6 and Batf3−/− mice on day 24. (C) Bar graph indicating the percentage and number of tumor-infiltrating total CD8+ T cells and the frequency of CD8+IFN-γ+ and CD8+granzyme B+ T cells, respectively. (D) CD8+Ki67+ T cells. (E) CD4+FoxP3+ Tregs. (F) Percentage of HSV 1-gB+, murine gp100+, and TRP2+CD8+ T cells, respectively. These experiments were repeated at least twice with similar results. Data are presented as means ± SEM, and the statistical differences between groups were measured by two-tailed Student’s t test. **P < 0.01, ***P < 0.001, ****P < 0.0001.

Treated Batf3−/− mice demonstrated a significant decrease in the percentage and number of CD8+ T cells compared to C57BL/6J (B6) (Fig. 6C). There was also a significant decrease in the percentage of CD8+ T cells expressing IFN-γ and granzyme B (Fig. 6D) and proliferating (Ki67+) CD8+ T cells (Fig. 6D) and increased Tregs (Fig. 6E) seen after combination treatment. We also observed a significant decrease in HSV-1 gB–specific CD8+ T cells and gp100- and TRP2-specific CD8+ T cells in Batf3−/− mice compared to wild-type mice treated with combination therapy (Fig. 6F).

Combination of T-VEC and MEK inhibition induces an inflammatory gene signature and increases PD-L1 expression in the tumor microenvironment

Previous studies have identified an inflammatory gene signature in patients responding to PD-1 checkpoint blockade (18). Because T-VEC is associated with type 1 IFN release and CD8+ T cell recruitment to the tumor microenvironment, we evaluated the established D4M3A tumors on day 24 from mice treated with mT-VEC, MEKi, or both for inflammatory gene expression profile. Treatment with mT-VEC was associated with an increased inflammatory signature compared to both mock- and MEKi-treated tumors, and this profile was highest in tumors treated with the combination of mT-VEC and MEKi (Fig. 7A). We also restricted the gene expression profile to five genes related to T cell activation (interferon-γ, CD8α, tumor necrosis factor-α, granzyme B, and perforin 1) and found a correlation between T cell activation gene expression and therapeutic effects (Fig. 7B), as well as categories of genes related to immune cells and antiviral immunity (fig. S9). Whereas MEKi inhibited expression of most antiviral genes, the combination of mT-VEC and MEKi resulted in increased gene expression compared to mT-VEC alone, except for interleukin-34 (IL34) and NKG2D ligand [UL16 binding protein 1 (Ulbp1)].

Fig. 7 T-VEC and MEK inhibition reprogram immune silent tumors into immune inflamed tumors and induces expression of PD-1 and PD-L1.

C57BL/6J mice were implanted subcutaneously in the right flank with 3 × 105 D4M3A cells and treated with mT-VEC (1 × 106 PFU) intratumorally for three doses on days 15, 19, and 22 and/or trametinib (0.5 mg/kg) orally once daily on days 15 to 19. Tumors were harvested on day 24, total RNA was isolated, and gene expression analysis was performed using the NanoString PanCancer Immune panel as described in Materials and Methods. (A) An inflammatory 16-gene expression profile was generated from mice (n = 3) treated (as described in Fig. 3D) with mock control (black), trametinib alone (MEKi; blue), mT-VEC alone (red), or a combination of mT-VEC and MEKi (purple). (B) A selected five-gene expression signature represented by genes highly associated with CD8+ T cell activation. (C) Gene expression of PD-1 (right panel) and PD-L1 (left panel). (D) Bar graphs show the mean fluorescence intensity (MFI) of CD45+PD-1+ (left panel) and CD45-PD-L1+ (right panel). Each experiment was performed at least twice with similar results. Data are presented as means ± SEM, and the statistical differences between groups were measured by two-tailed Student’s t test. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001.

Although PD-1 and programmed cell death ligand 1 (PD-L1) expression was significantly increased in the inflammatory gene panel in mT-VEC and MEKi–treated animals (Fig. 7C), this was confirmed by flow cytometry analysis of CD45+ cells harvested from the tumor microenvironment at day 24 as described in Fig. 3D. An increase in both PD-1 and PD-L1 was observed in tumors treated with mT-VEC alone, but they were highest in tumors treated with the combination (Fig. 7D).

Triple treatment with T-VEC, MEK inhibition, and PD-1 blockade further enhances therapeutic activity

Although combination therapy using mT-VEC and MEK inhibition reduced tumor burden and enhanced survival of treated mice (Fig. 3, B and C), tumors were completely eradicated in only 30 to 40% of mice. On the basis of the flow cytometry analysis and gene expression profiling showing an increase in PD-1 and PD-L1 expression in the tumor microenvironment (Fig. 7D), we reasoned that therapeutic activity might be further expanded by addition of PD-1 blockade to the combination regimen. To test this, we treated D4M3A tumor–bearing mice with mT-VEC, MEKi, or both as previously described in Fig. 3C, with or without αPD-1 antibody. There was limited impact of αPD-1 when given with mT-VEC alone or MEK inhibition alone (Fig. 8, B and C), as compared to monotherapy (Fig. 3, B and C). However, the combination of αPD-1 with both mT-VEC and MEKi resulted in complete durable responses in almost all mice (six of seven), compared to two of seven mice with mT-VEC and MEKi (Fig. 8C). All mice who cleared primary tumors with mT-VEC/MEKi/αPD-1 therapy rejected subsequent tumor rechallenge (Fig. 8D). Flow cytometry analysis performed on tumors showed a significant decrease in CD45+PD-1+ and CD8+PD-1+ cells in mice treated with triple therapy compared to mT-VEC and MEKi (Fig. 8E and fig. S10). No significant changes were observed in Tregs or the CD8+/Treg ratio (Fig. 8F). Triple combination elicited an increase in the percentage of total CD8+ T cells (Fig. 8G, left panel), as well as granzyme B and Ki67 expression (Fig. 8G). There was no overt toxicity observed in the mice as evidenced by changes in body weight, feeding habits, stool frequency, or coat appearance, including the absence of vitiligo.

Fig. 8 Triple-combination treatment with T-VEC, MEK inhibition, and PD-1 blockade improves therapeutic treatment of melanoma and colon cancer models.

(A) Treatment schema: Red arrows indicate T-VEC, the blue bar on top indicates trametinib, and brown arrows indicate αPD-1. (B) Mean tumor area. (C) Survival of mice. (D) Rechallenge of mice cured from (B). (E to G) Flow cytometry of tumors at day 24. Bar graph indicating percentage of positive (E) CD45+PD-1+ cells (left panel) and CD8+PD-1+ cells (right panel), (F) CD4+FoxP3+ (left panel) and ratio of effector T cells (Teff) to Tregs (right panel). (G) CD8+ T cells, granzyme B+CD8+ T cells, and Ki67+CD8+ T cells, respectively. (H) Evaluation of triple combination in the CT26 murine colon carcinoma model. Mice were treated as described in Materials and Methods. Each experiment was conducted at least twice with similar results. Data are presented as means ± SEM, and statistical differences between groups were measured by two-tailed Student’s t test. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001.

Last, to test the efficacy of triple combination in a different model, we tested the triple combination in BALB/c mice bearing established CT26 murine colon cancer tumors. In the CT26 model, both mT-VEC/MEKi and mT-VEC/αPD-1 antibody dual combinations elicited significant antitumor activity, with regression observed in 5 of 10 mice (Fig. 8H and fig. S11). In addition, the triple combination, using mT-VEC/MEKi/αPD-1, caused regression of all tumors, producing complete responses not observed with double therapy treatment (Fig. 8H and fig. S11). There were no signs of toxicity or weight loss in any of these animals.

DISCUSSION

Here, we demonstrate that the combination of T-VEC and MEK inhibition increases melanoma tumor cell killing through increased viral replication and apoptosis in vitro and enhances melanoma-specific adaptive immune responses in vivo. Previous reports have described interactions between MAPK pathway inhibition and other oncolytic viruses (12, 19). MEK inhibition was found to increase oncolytic adenovirus replication and tumor cell killing, possibly through up-regulation of coxsackievirus and adenovirus receptor (20). In glioma cells, MEKi PD98059 inhibited autophagy and increased cell killing without increasing oncolytic adenovirus replication (21). Oncolytic reovirus increased tumor cell killing because of increased endoplasmic reticulum stress-induced apoptosis and not because of increased virus replication (19). The effects of MEK inhibition are more complex with oncolytic HSV and depend on the particular virus strain. An internal repeat-deleted oncolytic HSV-1 inhibited p-MAPK activation in vitro and synergized with PD98059 in killing triple-negative breast cancer cell lines (12). In contrast, the tumor cell cytotoxicity of ICP34.5-deleted oncolytic HSV-1 (R3616) in vitro correlated with constitutive MEK activation, due to suppression of protein kinase R so that MEK inhibition reduced virus replication by about 15-fold (22). The R3616 virus was also found to be more effective in vivo against tumors with high MEK activity (23). Compared to R3616, T-VEC has an additional ICP47 deletion, resulting in early Us11 expression and PKR suppression, which likely explains the favorable interaction of T-VEC with MEKi. Thus, different oncolytic viruses interact with MEK inhibition in different ways.

In our model, we used trametinib, which is a more selective MEK 1/2 inhibitor and has been previously shown to have fewer side effects compared to other MEKis (24). Thus, in the context of combination therapy, trametinib might provide a particularly improved therapeutic window. Another factor known to influence the replicative ability of viruses is the status of the antiviral machinery, which is composed of intracellular factors that detect viral nucleic acids and molecular elements that promote viral clearance, such as type 1 IFN, and viral DNA sensors such as cGAS-STING (cyclic GMP-AMP synthase–stimulator of interferon genes). The expression of the antiviral machinery is typically defective in tumor cells, which allows preferential replication for many oncolytic viruses (25). In our model, we also confirmed that MEK inhibition was associated with decreased antiviral response expression in vivo, including STING (Tmem173) and IFN response factors 3 and 7. MEK inhibition inhibits expression of these factors, thus establishing favorable intracellular conditions for viral replication.

Although MEK inhibition in tumor cells may promote apoptosis and drive immunogenic cell death, MEK inhibition may also block T cell activation (26). Thus, it may be unexpected that we observed strong antiviral and antitumor CD8+ T cell responses in our model (Figs. 3 and 5). One explanation may have been related to a recent finding that MEK inhibition selectively blocks activation of naïve T cells but not antigen-experienced effector T cells (10). Others have shown that MEK inhibition selectively suppresses alloreactive T cells in a model of graft-versus-host disease (GVHD), demonstrating that trametinib blocks GVHD-inducing CD8+ T cells but spares graft-versus-tumor–specific CD8+ T cells in vivo (27). Thus, established tumors may contain antigen-experienced T cells and, in this setting, MEK inhibition would be expected to promote T cell activation, consistent with the NanoString data from tumor-bearing mice treated with trametinib and mT-VEC. The increased expression of STING and TLRs may also promote recruitment of antitumor CD8+ T cells because induction of these innate immune sensors has been associated with restoration of effective antitumor immunity (28).

Another important observation in our study was that BRAF inhibition only enhanced T-VEC oncolysis in BRAF-mutant tumor cells. In contrast, MEK inhibition improved T-VEC replication and oncolytic activity in both BRAF-mutant and BRAF wild-type cell lines. This suggests that MEK inhibition may be better than BRAF inhibitors in combination with HSV-based oncolytic viral therapy and might be active regardless of BRAF mutation status. This will, however, require further clinical validation.

In our studies, therapeutic effectiveness was seen in both human xenograft and immune-competent melanoma models. We found that the combination of T-VEC and MEK inhibition is also associated with increased accumulation of activated CD8+ T cells, characterized by production of IFN-γ and granzyme B, within the tumor microenvironment as well as an increase in CD8+/Treg ratio. We also confirmed the importance of CD8+ T cells through selective immune cell depletion studies. This is consistent with previous reports in patients with melanoma treated with single-agent T-VEC in which injected tumors have been found to have higher numbers of MART-1–specific effector CD8+ T cells and decreased numbers of CD4+Foxp3+ Tregs (29). Because HSV-1 can promote IFN production, we also found an increase in PD-1 and PD-L1 expression within the tumor microenvironment (30), and this is likely related to the counter-regulatory induction of immune checkpoints in the setting of excessive IFN-γ (31). This comports with recent clinical data showing increased PD-L1 expression in tumors from patients with melanoma treated with T-VEC and pembrolizumab (6) and provides a biologic rationale for the addition of PD-1 blockade to T-VEC and MEKi treatment, where we showed more than 80% complete tumor eradication and increased survival without overt signs of toxicity. This triple-drug regimen is particularly appealing because all three agents are currently approved by the U.S. Food and Drug Administration for the treatment of melanoma, and it is consistent with recent reports suggesting that MAPK pathway inhibition and PD-1/PD-L1 blockade are associated with improved therapeutic responses in preclinical models (10, 32).

The Batf3+ (CD8+CD103+) DC population was initially identified as critical for priming antiviral CD8+ T cell responses (16). These DCs are also critical for antitumor immunity and recruitment of lymphocytes through chemokines, such as CXCL9 (33). We demonstrated that Batf3+ DCs are also critical for the recruitment of CD8+ T cells in our melanoma model after treatment with T-VEC and trametinib and observed an increase in CXCL9 expression. These data support a role for Batf3+ DCs and local CXCL9 expression to generate and recruit effector CD8+ T cells into the tumor microenvironment. We found evidence for both viral (HSV-1 gB–specific) and melanoma (gp100- and TRP2-specific) CD8+ T cell responses, consistent with initial viral responses followed by cross presentation of tumor-associated antigens (34). Antigen spreading has previously been reported as a predictive biomarker of therapeutic responses for other forms of immunotherapy, including tumor vaccines and immune checkpoint inhibitors (35, 36). We further observed that combination therapy promoted an IFN-γ–regulated gene signature profile that has been associated with therapeutic responses to PD-1 blockade in patients with melanoma (18). On the basis of the increased expression of PD-1 and PD-L1, we added anti–PD-1 therapy to the mT-VEC + MEKi combination, which resulted in a survival benefit. We also observed a similar therapeutic effect for triple combination in the genetically distinct BALB/c murine CT26 model, suggesting that this approach may be broadly applicable for solid tumors beyond melanoma. Last, our data may also have implications for other microbial pathogens being evaluated as cancer therapeutics. In a patient-derived orthotopic xenograft model, treatment with Salmonella typhimurium combined with vemurafenib or trametinib resulted in improved therapeutic responses in a BRAF V600E–mutated melanoma (37).

Our study does have certain limitations, such as the use of murine models for oncolytic virus studies, which may be influenced by decreased viral tropism in murine tumor cells compared to human tumor cells. In addition, xenograft models are not adequate for representing an intact host immune system, and implanted tumors may not adequately reflect the biology of spontaneously arising cancers. We did observe infection of the D4M3A cell line, which allowed therapeutic studies to be conducted, and these were supplemented with human xenograft tumor experiments, which are susceptible to T-VEC infection and treatment with MEKi.

In summary, we evaluated the combination of MAPK inhibition and T-VEC in murine and human melanoma cell lines and found a synergistic effect between T-VEC and MEK inhibition regardless of BRAF mutation status. We also confirmed that therapeutic responses could be further improved by addition of anti–PD-1 therapy. In our studies, we did not observe overt signs of toxicity in mice supporting an improved therapeutic window although clinical confirmation is needed. Collectively, these data provide preclinical rationale for triple-combination treatment of T-VEC, MEK inhibition, and PD-1 blockade in patients with melanoma.

MATERIALS AND METHODS

Study design

In this hypothesis-driven study, we tested the therapeutic potential of two clinically approved agents in melanoma, an oncolytic virus (T-VEC) and the selective MEKi trametinib. Combination therapy was evaluated in human and murine melanoma cell lines in vitro, in human-derived xenograft tumor models in vivo, and in a transplantable murine model using the D4M3A cell line that is sensitive to HSV-1 infection. The effect of the two-drug regimen on viral replication was assessed by plaque assay. In addition, immune studies were performed to determine the impact of the combination on various immune cell populations, antigen spreading, and induction of gene signature profiles. The initial experiments found that the combination treatment induced expression of PD-1 and PD-L1, and further in vivo studies were conducted to assess triple-combination treatment with oncolytic virus, MEKi, and PD-1 blockade. In all experiments, animals were assigned to various experimental groups at random, but investigators were not blinded. For survival studies, sample sizes of 7 to 10 mice per group were used. Mice were euthanized when tumors reached 400 mm2. All outliers were included in the data analysis. Primary data showing tumor growth are reported in table S6.

Cell lines

Human melanoma cells SK-MEL-28, SK-MEL-2, and SK-MEL-5 [American Type Culture Collection (ATCC)] and mouse cell line CT26 (ATCC) were cultured in monolayers using RPMI supplemented with 10% heat-inactivated bovine serum (Thermo Fisher Scientific), 10 mM l-glutamine (Corning), and 0.5% penicillin G–streptomycin sulfate (Corning). Cells were detached using 0.25% trypsin EDTA (Corning) for passaging. The murine melanoma cell line D4M3A was generated from Tyr::CreER;BrafCA;Ptenlox/lox mice (13) and provided by D. Mullins (Dartmouth University, Hanover, NH). D4M3A cells were cultured as previously described (13). All cells were low passage and confirmed to be mycoplasma free (LookOut mycoplasma kit; Sigma-Aldrich).

Viruses

T-VEC is a modified JS1 strain of HSV-1 encoding human GM-CSF and has been previously reported (38). T-VEC is commercially available and was purchased from the Rutgers Cancer Institute Pharmacy. For immune-competent murine studies, a modified virus (mT-VEC) in which the human GM-CSF gene was replaced by murine GM-CSF was used and provided by P. Beltran (Amgen Inc.). All human cell lines and xenograft experiments were performed using T-VEC, and all murine cell line and syngeneic experiments were performed using mT-VEC.

Drugs, chemicals, and antibodies

Drugs and antibodies and their respective suppliers are listed in table S1. Chemical agents are listed in table S2.

Cytotoxicity and viral plaque assays, Western blotting, LumaCyte analysis, immune histochemistry, and flow cytometry

All commercial kits used in these experiments are listed in table S3 and described in more detail in Supplementary Methods.

Gene expression studies

Gene expression analysis was performed using the NanoString PanCancer Immune panel as described in Supplementary Methods.

Murine tumor treatment

The murine treatment and survival studies, as well as the immune cell depletion experiments, are described in detail in Supplementary Methods. In tumor treatment studies, tumor growth was measured in two dimensions, recording the greatest length and width using digital calipers. Tumor area was calculated by multiplying length with width. Tumor sizes were plotted as average size for each group. For survival experiments, mice were monitored for tumor growth and euthanized when tumors reached 400 mm2. Kaplan-Meier curves were used to document survival. Mice were weighed twice a week, and no weight loss was observed during the treatment. Each experiment was performed at least two times, and all animal experiments were approved by the Rutgers Institutional Animal Care and Use Committee.

Statistical analyses

All statistical analyses were performed using GraphPad Prism software version 7.0a. Survival data were analyzed by Kaplan-Meier survival curves, and comparisons were performed by log-rank test. Cell viability data, flow cytometric data, and immunohistochemistry counts were compared using an unpaired Student’s t test (two tailed) or one-way ANOVA when multiple comparisons were made. P values of less than 0.05 were considered significant.

SUPPLEMENTARY MATERIALS

www.sciencetranslationalmedicine.org/cgi/content/full/10/471/eaau0417/DC1

Methods

Fig. S1. BRAF inhibitors enhance T-VEC cell killing in BRAF-mutant melanoma cell lines.

Fig. S2. LumaCyte laser flow cytometry analysis.

Fig. S3. T-VEC and MEKi–induced apoptosis.

Fig. S4. Characterization of murine D4M3A cells.

Fig. S5. Validation of immune cell depletion.

Fig. S6. Time course analysis of tumor-infiltrating CD8+ T cells during mT-VEC treatment.

Fig. S7. Analysis of CD8+ T cells from spleen during mT-VEC + MEKi treatment.

Fig. S8. Characterization of D4M3A tumor cells in Batf3 knockout mice.

Fig. S9. NanoString gene expression heat maps for all genes profiled and by gene function.

Fig. S10. MFI expression of PD-1 expression and frequency of PD-1+ cells.

Fig. S11. Individual tumor growth curves of BALB/c mice bearing CT26 tumors.

Table S1. Antibodies.

Table S2. Chemicals.

Table S3. Commercial assays.

Table S4. Experimental cell lines.

Table S5. Experimental models.

Table S6. Tumor area from mouse studies.

Data file S1. Gene expression raw data.

References and Notes

Acknowledgments: We wish to thank C. Hebert, S. Hart, and Z. Evans for help with LumaCyte experiments. We also thank P. Beltran, J. Gansert, and C. Keegan at Amgen Inc. for reagents and research funding. We thank D. Mullins at Dartmouth University for providing the D4M3A cell line. We thank D. Saha (Rabkin laboratory) for helpful discussion and useful comments with experimental methods. We also thank C. Peters (Rabkin laboratory) for helpful discussions. We thank S. Jhawar, S. Li, R. Pepe, J. Newman, and members of the Zloza laboratory for experimental help and useful discussions. We thank NanoString Technologies for help with gene expression analysis. We also thank R. Coffin for useful comments and suggestions. We thank S. Laddha from Systems Biology Group at Rutgers Cancer Institute of New Jersey for helpful discussions with NanoString analysis; W. Rodriguez, Rutgers Cancer Institute Central Laboratory Core Equipment Services for facilities and equipment; and D. Medina, Rutgers Cancer Institute Immune Monitoring Core for running NanoString. We also thank Biospecimen Repository Service for help with immunohistochemistry and Imaging Core for help with quantifying immunohistochemistry images. Funding: H.L.K. received partial funding to support this research through a grant from Amgen Inc. S.D.R. was supported by an NIH grant (R01CA160762). Author contributions: Conceptualization: P.K.B. and H.L.K. Methodology: P.K.B., A.Z., S.D.R., and H.L.K. Investigation: P.K.B. Technical assistance: S.A. Formal analysis: P.K.B., S.A., A.Z., S.D.R., and H.L.K. Writing: P.K.B. and H.L.K. (original draft) as well as P.K.B., A.Z., S.D.R., and H.L.K. (review and editing). Visualization: P.K.B. and H.L.K. Funding acquisition: H.L.K. Resources: A.Z., S.D.R., and H.L.K. Supervision: A.Z. and H.L.K. Competing interests: H.L.K. is an employee of Replimune Inc. H.L.K. and P.K.B. are inventors on the patent application submitted by Rutgers University that covers the use of T-VEC, MEKis, and PD-1 blockade (U.S. Provisional Patent Application no. 62/666,390, filed 3 May 2018). S.D.R. is an inventor on patents relating to oncolytic HSV owned by Georgetown University and Massachusetts General Hospital that have been licensed to Amgen Inc., for which he receives royalties, and is also a paid consultant for Replimune Inc. All other authors declare that they have no competing interests. Data and materials availability: All data associated with this study are present in the paper or the Supplementary Materials.
View Abstract

Stay Connected to Science Translational Medicine


Editor's Blog

Navigate This Article