An engineered oncolytic virus expressing PD-L1 inhibitors activates tumor neoantigen-specific T cell responses

Abstract

Oncolytic viruses offer an in situ vaccination approach to activate tumor-specific T cell responses. However, the upregulation of PD-L1 expression on tumor cells and immune cells leads to tumor resistance to oncolytic immunotherapy. In this study, we generate an engineered oncolytic virus that coexpresses a PD-L1 inhibitor and GM-CSF. We find that the oncolytic virus is able to secrete the PD-L1 inhibitor that systemically binds and inhibits PD-L1 on tumor cells and immune cells. Importantly, the intratumoral injection with the oncolytic virus overcomes PD-L1-mediated immunosuppression during both the priming and effector phases, provokes systemic T cell responses against dominant and subdominant neoantigen epitopes derived from mutations, and leads to an effective rejection of both virus-injected and distant tumors. In summary, this engineered oncolytic virus is able to activate tumor neoantigen-specific T cell responses, providing a potent, individual tumor-specific oncolytic immunotherapy for cancer patients, especially those resistant to PD-1/PD-L1 blockade therapy.

Fig. 1. Generation and characterization of an oncolytic vaccinia virus coexpressing a mouse PD-L1 inhibitor and GM-CSF.

a A schematic diagram of recombinant vaccinia virus (VV) shuttle vectors that express GM-CSF or/and iPDL1 (soluble PD-1-Fc). vTK, VV thymidine kinase gene; R and L, right and left flank sequences; RFP, red fluorescent protein. b Expression and secretion of iPDL1 from infected MC38 tumor cells infected with the indicated VVs. Anti-IgG Fc (Licor 926-32210; upper) or anti-PD-1 (Biolegend 114101; lower) was used for western blot with reducing or non-reducing loading buffer. The experiment was repeated twice. c, d Serum iPDL1 and GM-CSF levels in different VV-treated MC38-bearing mice at 2 days post-virus injection c. Kinetics of iPDL1 levels in injected tumors or sera of the VV-iPDL1/GM-treated mice d. n = 3 independent samples. Data are presented as the means ± SD. The experiment was repeated twice. e Purified iPDL1 binds to PD-L1⁺ tumor cell. Upper panel: flow cytometric analysis of PD-L1 expression on shPD-L1/MC38 tumor cells that were transduced with PD-L1-shRNA and wild-type MC38 cells. Lower panel: shPD-L1/MC38 cells and wild-type MC38 cells were incubated with 50 μg/mL of purified iPDL1, an irrelevant MAGE3-IgG Fc fusion protein, or IgG control, followed by staining with an anti-IgG Fc for flow cytometry. f Inhibition of PD-1/PD-L1 binding by purified iPDL1 protein using ELISA. An anti-PD-L1 antibody was used as a positive control; n = 3 independent samples. g iPDL1-mediated ADCC. ADCC Reporter Bioassays were performed in triplicate wells, and the concentrations of iPDL1 protein and control IgG Fc used for this assay are indicated; n = 3 independent samples. Data presented as the means ± SD. The experiment was repeated twice. Significant differences are indicated as ***P < 0.001, or ****P < 0.0001 using two-tailed student’s t-test. h CD11c⁺ DC frequency in monocyte cultures in the presence of culture media of MC38 cells infected with VV-RFP, VV-GM, VV-iPDL1/GM, or GM-CSF as a positive control, and IL-4. i Viral replication in vitro; n = 3 independent samples. j Replication and biodistribution of VV after intratumor injections. Data presented as the means ± SD. The experiment was repeated twice.

Fig. 2. PD-L1 inhibitors secreted from VV-iPDL1/GM-infected cells bind to PD-L1 on tumor cells and immune cells.

a MC38 tumor cells were infected with VV-RFP, VV-iPDL1/GM at an MOI = 0.5, or PBS for 24 h. The percentage of IgG Fc⁺ population representing iPDL1 (soluble PD-1-IgG Fc)-bound VV-infected (RFP⁺) or uninfected (RFP⁻) PD-L1-expressing tumor cells was measured by flow cytometry. b MC38 tumor cells that were stimulated with IFN-γ (20 ng/mL) for 48 h were infected with the indicated VVs. PD-L1 expression of infected (RFP⁺) or uninfected (RFP⁻) cells was determined by flow cytometry. c–g MC38 cells were subcutaneously inoculated into the left (1 × 10⁶) and right (5 × 10⁵) flanks of C57BL/6 mice. When left flank tumor sizes reached ~100 mm³ (counted as day 0), the tumors of the left flank were intratumorally injected with 50 μL of PBS, VV-RFP, VV-GM, or VV-iPDL1/GM (5 × 10⁷ pfu per tumor), or 200 μg of anti-PD-L1 antibody (clone 10F.9G2) intravenously on days 0 and 3. Two days post-second VV treatment, VV-treated (primary tumor) and untreated, distant tumors were collected, weighed and digested with collagenase type I and DNase. Tumor cell suspensions were blocked with anti-CD16/32 antibody and then stained with antibodies against CD45, CD3, CD8, CD4, CD11c, CD11b, Gr-1, FoxP3, PD-L1, and IgG Fc to assess PD-L1 expression or IgG Fc⁺ frequency on infected (RFP⁺) or uninfected (RFP⁻) tumor cells from the treated primary tumors c or untreated distant tumors (CD45⁻ cells) d, and PD-L1 expression on infiltrating immune cells from treated e or untreated distant tumors f, IgG Fc⁺ frequency on infiltrating immune cells g. Infiltrating immune cells include cytotoxic T cells (CD45⁺CD3⁺CD8⁺), DCs (CD45⁺CD11c⁺), myeloid-derived suppressor cell MDSCs (CD45⁺CD11c⁻CD11b⁺Gr-1⁺), and Treg (CD45⁺CD3⁺CD4⁺FoxP3⁺); n = 5 mice. Significant differences are indicated as **P < 0.01, ***P < 0.001, or ****P < 0.0001 determined by two-tailed Student’s t-test.

Fig. 3. Enhanced antitumor activities against primary tumors.

a–c C57BL/6 mice were subcutaneously inoculated with 5 × 10⁵ luciferase-expressing B16-F10 (B16-F10-Luc) cells. When tumor sizes reached ~100 mm³ (counted as day 0), the mice were intratumorally injected with 50 μL of VV-RFP, VV-GM, or VV-iPDL1/GM (5 × 10⁷ pfu per tumor) or PBS at days 0, 3, and 7. Bioluminescence monitoring a, b and caliper measurement c of B16-F10-Luc cells were performed on the indicated days. Data are presented as the means ± SD (n = 5 mice). Significant differences are indicated as *P < 0.05 determined by two-tailed Student’s t-test. d, e Py230 d or MC38 e tumor volume was monitored by caliper measurement using the same treatment schedule as in a–c. Data are presented as the means ± SD (n = 5 mice).

Fig. 4. Enhanced antitumor activities against untreated, distant tumors.

a–d Inhibition of rechallenged tumor growth. B16-F10 melanoma cells were implanted intradermally to the left flank of C57B/6 mice. When tumor sizes reached ~100 mm³ (counted as day 0), the mice were intratumorally injected with the indicated VVs on days 0, 3, and 7. Treated mice were s.c. rechallenged with B16-F10-Luc cells 30 days after the last VV injection (counted as day 0 for rechallenge data). Bioluminescence monitoring a, b and caliper measurement of B16-F10-Luc cells c were performed. Data are presented as means ± SD (n = 5 mice). d Survival curve of B16-F10 rechallenged mice. e–h The volumes of rechallenged Py230 e or MC38 g tumors were monitored using a similar treatment schedule as in a, except that 5 × 10⁵ of Py230 or MC38 tumor cells were rechallenged. Data are presented as means ± SD (n = 5 mice). *P < 0.05, ***P < 0.001 determined by two-way ANOVA. Survival curve of Py230 f and MC38 h rechallenge mice. *P < 0.05, *** P < 0.001 by two-tailed Log rank test. i CD8 T cell depletion. Surviving mice treated with VV-iPDL1/GM for the original left flank tumor implantation were rechallenged with 5 × 10⁵ MC38 cells at right side without or with weekly i.v. injections of anti-CD8 antibody for two times. Data are presented as means ± SD (n = 5 mice). ****P < 0.01 by two-tailed repeated-measures two-way ANOVA. j–l Inhibition of untreated, established tumor growth. B16-F10 melanoma cells were implanted to the left and right flanks of C57B/6 mice. The mice were intratumorally injected to the left flank tumors with indicated VVs without or with i.v. injections of anti-PD-L1 antibody on days 0, 3, and 7. j Individual curves are depicted for each tumor. Numbers indicate complete tumor regression out of total tumors in each group. k Distribution of tumor volumes determined on day 30 after virus injection; n = 10 mice. Bars represent mean values ± SD. *P < 0.05 by two-tailed Mann–Whitney U test. l Cumulative survival curves. Data are from two independent experiments. *P < 0.05; **P < 0.01; NS, not significant by two-tailed Log rank test.

Fig. 5. Enhanced tumor infiltration and activation of immune cells.

A similar treatment schedule as in Fig. 2c was used, except that 5 days after the second VV injection, VV-treated MC38 tumors were harvested, weighed, and digested for preparation of single-cell suspensions followed by antibody staining against CD45, CD8, CD4, CD11c, CD11b, Gr-1, and FoxP3. a–c Representative plots of the percentages of infiltrating CD45⁺ immune cells, DCs, MDSCs, CD4⁺ T cells, CD8⁺ T cells, and Tregs in treated tumors a. Absolute numbers of the above immune cells and CD8⁺ T cell/Treg ratio values in treated tumors b and distant, untreated tumors c. n = 5 mice. Data presented as the means ± SD. *P < 0.05, **P < 0.01 by two-tailed Student’s t-test. d, e Expression of IFN-γ, TNF-α, and CD 107a of tumor-infiltrating CD8⁺ T cells in response to restimulation with MC38 tumor lysate-pulsed DCs in the presence of Golgi-plug for 8 h were measured by intracellular staining d. e Quantitative presentation of d. n = 5 mice. Data presented as the means ± SD. *P < 0.05 by two-tailed Student’s t-test.

Fig. 6. Enhanced T cell responses against dominant and subdominant tumor neoantigen epitopes.

a Enhanced T cell responses against a pool of neoantigen peptides. MC38 tumor-bearing mice were intratumorally injected with various VVs at days 0, 3, and 7. One group of C57BL/6 mice were i.v. injected with 200 μg of anti-PD-L1 antibody. Ten days later, splenocytes were cultured in the presence of a mixture of 11 neoepitope peptides (10 μg/mL/each). After 80 h of incubation, supernatants were collected for IFN-γ ELISA (right). [³H] thymidine incorporation was measured (left). The graph shows the results from three mice of each group. Data presented as the means ± SD. *P < 0.05 by two-tailed Student’s t-test. b Enhanced T cell responses against individual neoantigens. The splenocytes from VV-treated mice were cocultured with each of the 11 neoepitope peptides (100 μg/ml) as above described above. [³H] thymidine incorporation (left) and ELISA IFN-γ concentrations (right) are shown; n = 3 mice. One bar or one dot represents one mouse. Data presented as the means ± SD. *P < 0.05 by two-tailed Student’s t-test. c Enhanced T cell responses against the neoantigenic peptide 11. The splenocytes isolated from VV-treated mice were cocultured with various concentrations of the neoepitope peptide 11 as above described above. [³H] thymidine incorporation was used to analyze T cell proliferation; n = 3 mice. Data presented as the means ± SD. *P < 0.05, **P < 0.01, ***P < 0.001 by two-tailed Student’s t-test. d, e Enhanced tumor infiltration of neopeptide 4-specific T cells. Tumor cell suspensions from various VV-treated mice using the same treatment schedule as Fig. 5a were stained with the neopeptide 4 (Pep4, ASMTNMEL)-loaded, H-2D^b-labeled pentamers, anti-CD45, and anti-CD8. Data are representative of five independent experiments. d Dot plots of flow cytometry; e quantification of peptide 4-pentamer⁺ CD8⁺ T cells. Data presented as the means ± SD. *P < 0.05 by two-tailed Student’s t-test. f Enhanced generation of neopeptide-specific memory T cells. Forty days after the virus injection, splenocytes were restimulated with neopeptide 4-loaded DCs in the presence of Golgi-plug followed by surface staining with anti-CD8 and intracellular staining with anti-107a, anti-IFN-γ, anti-IL-2, and anti-TNF-α.

Fig. 7. Enhanced neoantigen presentation and cytolytic activity of neoantigen-specific CTLs.

a Enhanced stimulatory potency of tumor-infiltrating DCs. Tumor-infiltrating DCs from VV-treated mice were loaded with neopeptide 4, 9, or 11, and cocultured with the neoantigens-primed T cells from mice immunized with the 11 neopeptide mixture to assess IFN-γ production; n = 3 mice. Data presented as the means ± SD. *P < 0.05, ***P < 0.001 by two-tailed Student’s t-test. b Enhanced maturation of tumor-infiltrating DCs. Using a similar treatment schedule as described in Fig. 5a, cell suspensions prepared from VV-treated tumors were analyzed by flow cytometry. c Enhanced tumor infiltration of CD103⁺ DCs. Using the same treatment schedule as in Fig. 5a, tumor cell suspensions from VV-treated mice were analyzed by FACS; n = 5 mice. Data presented as the means ± SD. **P < 0.01 by two-tailed Student’s t-test. d Intracellular staining of IL-12 and CXCL9 of CD103⁺ DCs from VV-treated tumors. e qRT-PCR analysis of CXCL10 mRNA levels in CD103⁺ DCs isolated from VV-treated tumors; n = 5 mice. Data presented as the means ± SD. **P < 0.01 by two-tailed Student’s t-test. f Neoantigens-primed T cells proliferated more efficiently in VV-iPDL1/GM-treated mice. The neoantigens-primed T cells were labeled with 5 μM CFSE and i.v. injected into VV-treated mice. Three days later, T cell proliferation was assessed by FACS. g Enhanced stimulatory effect of VV-iPDL1/GM-infected tumor cells. MC38 tumor cells infected with VVs at MOI = 1 were cocultured with the neoantigens-primed T cells. IFN-γ production (left) and T cell proliferation (right) were measured. Data presented as the means ± SD. *P < 0.05 by two-tailed Student’s t-test. h Serum of VV-iPDL1/GM-treated mice enhanced the cytolytic activity of neoantigens-primed T cells. MC38-Luc cells were cocultured with the neoantigen-specific T cells in the presence of the sera from treated MC38-bearing mice. Cytolytic activity was calculated using luciferase emission value. Data are presented as means ± SD. **P < 0.01 by two-tailed Student’s t-test. i PD-1⁺ CD8⁺ T cells isolated from VV-treated MC38 tumors were cocultured with MC38 cells in the presence of purified iPDL1 or IgG. IFN-γ⁺ frequencies of PD-1⁺ T cells were shown from one of two independent experiments.