STING-activating nanoparticles normalize the vascular-immune interface to potentiate cancer immunotherapy

Abstract

The tumor-associated vasculature imposes major structural and biochemical barriers to the infiltration of effector T cells and effective tumor control. Correlations between stimulator of interferon genes (STING) pathway activation and spontaneous T cell infiltration in human cancers led us to evaluate the effect of STING-activating nanoparticles (STANs), which are a polymersome-based platform for the delivery of a cyclic dinucleotide STING agonist, on the tumor vasculature and attendant effects on T cell infiltration and antitumor function. In multiple mouse tumor models, intravenous administration of STANs promoted vascular normalization, evidenced by improved vascular integrity, reduced tumor hypoxia, and increased endothelial cell expression of T cell adhesion molecules. STAN-mediated vascular reprogramming enhanced the infiltration, proliferation, and function of antitumor T cells and potentiated the response to immune checkpoint inhibitors and adoptive T cell therapy. We present STANs as a multimodal platform that activates and normalizes the tumor microenvironment to enhance T cell infiltration and function and augments responses to immunotherapy.

Figure 1.. Systemic administration of STANs stimulates an anti-angiogenic inflammatory response in the tumor microenvironment to inhibit tumor growth.

(A) Schematic of RenCa tumor inoculation and treatment schedule. Mice with subcutaneous ~50 mm³ RenCa tumors were intravenously administered PBS (vehicle), empty nanoparticle (NP), free cGAMP, or STAN three times as indicated, and tumor volume was monitored. (B-D) Spider plots of individual tumor growth curves (B), tumor growth curves (C), and Kaplan-Meier survival curves (D) of RenCa tumor bearing mice treated with the indicated formulation (n=5 mice/group). Tumor growth curves are presented as mean ± SEM with P value determined by two-way ANOVA with post-hoc Tukey’s correction for multiple comparisons; ****P<0.0001 on day 21 for all groups compared to STAN. Kaplan-Meier survival curves of mice treated with indicated formulation using 1500 mm³ tumor volume as endpoint criteria with P value was determined by log-rank test; **P≤0.01 compared to PBS control. (E) Schematic of RenCa tumor inoculation, treatment schedule, and analysis time points. Mice with RenCa tumors of ~100 mm³ were administered PBS, empty nanoparticle (NP), free cGAMP, or STAN three times as indicated, and tumors were harvested at 24 h after the 3^rd injection for qRT-PCR and Western blot analysis. (F) Western blot analyses for p-IRF3, total IRF3, and STING in RenCa tumors 24 h following the final administration of STANs or controls (n=2 mice/group). Blots shown are representative of 3 independent experiments with similar results (see Fig. S3). (G-P) qRT-PCR analysis of tumor tissue 24 h following the final administration of STANs or indicated controls to evaluate changes in expression of type-I interferon, interferon-stimulated genes, and pro-inflammatory cytokines (G-J), markers of angiogenesis and vascular remodeling (K-N), and leukocyte adhesion molecules (O-P) (n=3–6 mice/group). Data shown as mean ± SEM. P value determined by one-way ANOVA with post-hoc Tukey’s correction for multiple comparisons; *P≤0.05, **P≤0.01, and ***P≤0.001 compared to PBS control.

Figure 2.. Systemic administration of STANs induces normalization of tumor vasculature.

(A) Schematic of RenCa tumor inoculation, treatment schedule, and analysis time points. Mice with ~100 mm³ RenCa tumors were administered PBS (vehicle), empty nanoparticle (NP), free cGAMP, or STAN three times as indicated. Two days after the 3^rd dose, pimonidazole hydrochloride was injected intravenously 60 min before euthanasia and tumor harvesting; dextran or lectin was intravenously injected 30 min prior to euthanasia. (B) Representative fluorescent micrographs and (C-G) quantification of RenCa tumor tissue following (C) Hypoxyprobe (green) staining; (D) CD31⁺ blood vessel density (green) and (E) αSMA pericyte (yellow) coverage; (F) lectin perfusion of tumor vessels (green); and (G) dextran leakage area in tumor tissue (green) for PBS, NP, cGAMP, and STAN treated mice. (H) Schematic of RenCa tumor inoculation, treatment schedule, and analysis time points. Starting on day 13, mice were administered PBS (vehicle) or STAN three times as indicated. Depleting or blocking antibodies (αIFNAR1, αTNFα, or αCD8 antibody) were administered on days 12, 15, 18 (indicated by green arrows) by intraperitoneal (IP) injection. Tumors were harvested for staining two days after the 3^rd STAN dose. (I) Tumor growth curves of RenCa tumor bearing mice treated with PBS, STANs, or STANs with depleting/blocking antibodies (n=6–8 mice/group). Tumor growth curves are presented as mean ± SEM with P value determined by two-way ANOVA with post-hoc Tukey’s correction for multiple comparisons; *P≤0.05 and **P≤0.01 on day 21 compared to STAN. (J) Representative fluorescent micrographs and (K-L) quantification of RenCa tumor tissue following indicated treatment and stained for (K) CD31⁺ blood vessel density (red) and (L) NG2⁺ pericyte coverage (green). (B,J) For all images, scale bars are 50 μm. (C-G, K-L) *P≤0.05, **P≤0.01, ***P≤0.001, and ****P<0.0001 indicate a statistically significant difference compared to PBS using a 1-way ANOVA with post-hoc Tukey’s correction for multiple comparisons (C-G, n=3 mice/group; K-L, 6–8 mice/group).

Figure 3.. STANs potentiate STING signaling in human renal cell carcinoma.

(A) Schematic of ex vivo culture model of human renal cell carcinoma tissue. (B) qRT-PCR analysis of ex vivo cultured human renal cell carcinoma tissue following the final administration of STAN or controls to evaluate changes in gene expression. *P≤0.05, **P≤0.01, ***P≤0.001, and ****P<0.0001 indicate a statistically significant difference compared to STAN using a 1-way ANOVA with post-hoc Tukey’s correction for multiple comparisons (n=2–3 sections of tumor per patient). (C) Volcano plots of changes in gene expression (log₂) between human RCC specimens treated with PBS and STANs. (D) Heat map of normalized gene expression represented by z-scores for five human RCC specimens treated with PBS or STAN. (E) Scatter plots for individual patients of the fold change in selected genes associated with vascular remodeling in response to treatment with STAN.

Figure 4.. Vascular normalization induced by systemic STAN administration enhances antibody and nanoparticle accumulation and penetration in tumors.

(A) Schematic of RenCa tumor inoculation, treatment schedule, and time points for analysis of tumor antibody distribution. Mice with ~100 mm³ RenCa tumors were administered PBS or STANs 3 times as indicated. Two days after the 3^rd dose, AlexaFluor680-IgG (10 mg/kg) antibody was administered intravenously (IV), and mice were imaged using IVIS (B) 6 h or (C) 24 h after antibody injection. (B-C) Left: Representative fluorescence IVIS images showing the accumulation of AlexaFluor680-IgG antibody in RenCa tumors. Right: Quantification of fluorescent antibody accumulation in tumors via measurement of fluorescent radiant efficiency at the tumor site. (D) Schematic of RenCa tumor inoculation, treatment schedule, and time points for analysis of Cy5-labeled polymer nanoparticle (Cy5-NP) distribution. Mice with ~100 mm³ RenCa tumors were administered PBS or STAN 3 times as indicated. One day following the 3^rd STAN administration, Cy5-NP were then injected IV, and tumors were harvested 24 h later for analysis by flow cytometry and immunofluorescence microscopy (IF). (E) Flow cytometry analysis of cellular uptake of Cy5-NP 24 h post-injection. Data are plotted as the percentage of each indicated cell population that are Cy5⁺ within the tumor (EC: endothelial cells; DC: dendritic cells; Mφ: macrophages). (F) Representative images and (G) quantification of Cy5-NP colocalization with CD31 (red), F4/80 (red), and CD11c (red). DAPI staining is represented in blue. Image magnification: 200x; scale bar: 100 μm. (H) Schematic of RenCa tumor inoculation, treatment schedule, and time points for analysis of gold nanoparticles accumulation in tumors. Mice with ~100 mm³ RenCa tumors were administered PBS or STAN 3 times as indicated. PEGylated 20 nm gold nanoparticles were then injected IV, and tumors were harvested for analysis of gold content via ICP-MS. (I) ICP-MS quantification of gold accumulation within tumors. All results are the mean ± SEM (B-C, n=8–12 mice/group; E, n=10–13 mice/group; G, n=5 mice/group; I, n=8 mice/group). **P≤0.01 and ****P<0.0001 indicate a statistically significant difference between STAN and PBS as determined by Student’s t-test.

Figure 5.. STANs activate vascular endothelial cells to enhance T cell transmigration.

(A) Schematic of RenCa tumor inoculation, treatment schedule, and time points for flow cytometric analysis of tumor. (B) Flow cytometric analysis of the frequency of CD54⁺ (ICAM-1) and CD106⁺ (VCAM-1) CD34⁺CD31⁺ ECs in subcutaneous RenCa tumors following treatment with STANs or PBS. (C) Flow cytometric analysis of the number (cells/mg tumor) of CD8⁺ T cells and CD4⁺ T cells in RenCa tumors 24 h following treatment with STANs or PBS. Flow cytometric analysis of the percentage of CD69⁺ CD8⁺ (D) and CD4⁺ (E) T cells in RenCa tumors and the spleen following treatment with STAN or PBS. (F) Western blot analyses for p-IRF3, total IRF3, and STING in bEnd.3 and MS1 murine endothelial cells. STAN treated lanes are normalized to PBS per cell type. (G-H) qRT-PCR analyses for comparison of Icam1 (G) and Vcam1 (H) expression in bEnd.3 and MS1 murine endothelial cells in response to STAN. (I) Quantification of activated CD8⁺ T cell transendothelial migration in response to treatment of MS1 ECs with STANs and effect of combined antibody blockade of ICAM1 and VCAM1 on T cell migration. All data are shown as mean ± SEM. (B-E, n=4–10 mice/group; G-H, n=3 biological replicates). *P≤0.05, ***P≤0.001, and ****P<0.0001 indicate a statistically significant difference determined by Student’s t test. (I, n=3 biological replicates). ****P<0.0001 indicates a statistically significant difference compared to the STAN only treated group using a 1-way ANOVA with post-hoc Tukey’s correction for multiple comparisons.

Figure 6.. STAN-mediated vascularization remodeling increases the infiltration of endogenous and adoptively transferred, antigen-specific activated CD8+ T cells.

(A) Schematic of EO771-OVA tumor inoculation, treatment and T cell transfer schedule, and time point for flow cytometric analysis of T cell infiltration. (B) Flow cytometric analysis of the number (cells per mg tumor) of endogenous (CD45.2⁺) and transferred OT-I (CD45.1⁺) CD8⁺ T cell into tumors following administration of PBS (vehicle) or STAN. (C) Schematic of EO771-OVA tumor inoculation, treatment and T cell transfer schedule, and time point for flow cytometric analysis of T cell phenotype. (D) Flow cytometric analysis of the frequency of adoptively transferred Thy1.1⁺ OT-I positive CD8⁺ T cell in tumors and spleens following treatment with STANs of PBS. (E-G) Flow cytometric analysis of the frequency of CD69⁺ (E), Ki67⁺ (F), and IFNγ⁺TNFα⁺ (G) endogenous (Thy1.1⁻) CD8⁺ T cells in tumors and spleens following STAN or PBS treatment. (H-J) Flow cytometric analysis of the frequency of CD69⁺ (H), Ki67⁺ (I), and IFNγ⁺TNFα⁺ (J) adoptively transferred OT-I (Thy1.1⁺) CD8⁺ T cells in tumors and spleens following STAN or PBS treatment. (K-L) Flow cytometric analysis of the frequency of Ki67⁺ (K) and IFNγ⁺TNFα⁺ (L) CD4⁺ T cells in tumors and spleens following STAN or PBS treatment. All data shown as mean ± SEM. (B, n=6–11 mice/group; D-L, n=8–10 mice/group). *P≤0.05, **P≤0.01, ***P≤0.001, ****P<0.0001 determined Student’s t-test.

Figure 7.. Systemic administration of STAN enhances responses to immune checkpoint blockade.

(A) Schematic of RenCa tumor inoculation and treatment schedule with STAN (3 doses) and ICB (5 doses of αPD-L1 antibody). (B-D) Spider plots of individual tumor growth curves (B), tumor growth curves (C), and Kaplan-Meier survival curves (D) of RenCa tumor bearing mice treated as indicated. (E) Schematic of RenCa tumor inoculation and treatment schedule with STAN (3 doses) and ICB (5 doses of αB7-H3 antibody). (F-H) Spider plots of individual tumor growth curves (F), tumor growth curves (G), and Kaplan-Meier survival curves (H) of RenCa tumor bearing mice treated as indicated. (C,G) Tumor growth curves are presented as mean ± SEM (n=7–10 mice/group). ****P<0.0001 determined by 2-way ANOVA with post-hoc Tukey’s correction for multiple comparisons; with comparisons shown between STAN and STAN + ICB as well as PBS against both STAN/STAN + ICB, are on (C) day 20 and (G) day 22. (D,H) Kaplan-Meier survival curves of mice treated with indicated formulation using 1500 mm³ tumor volume as endpoint criteria and P value was determined by log-rank test. **P≤0.01, ***P≤0.001, and ****P<0.0001 compared to PBS.

Figure 8.. Systemic administration of STANs increases response to adoptive T cell therapy in murine model of orthotopic breast cancer.

(A) Schematic of EO771-OVA tumor inoculation and treatment schedule with STANs and activated OT-I T cells. (B-D) Spider plots of individual tumor growth curves with number of complete responders (CR) indicated (B), tumor growth curves (C), and Kaplan-Meier survival curves (D) of EO771-OVA tumor bearing mice. (E-H) At day 135, mice that exhibited complete responses to the combination of STANs and OT-I T cell transfer treatment were re-challenged with EO771-OVA cells (CRs from STAN + 0.5×10⁶ OT-I treatment) or parental EO771 cells (CRs from STAN + 3×10⁶ OT-I treatment). (F) Tumor growth was monitored and compared with treatment-naïve mice inoculated with EO771-OVA or EO771 cells. For spider plots of individual tumor growth curves, number of CRs are indicated. Kaplan-Meier survival curves for EO771-OVA challenge (G) and EO771 challenge (H). (I) Schematic of larger EO771-OVA tumor inoculation (~200 m³) and treatment schedule. (J-L) Spider plots (J), tumor growth curves (K), and Kaplan-Meier survival curves (L). Tumor growth curves are presented as mean ± SEM (C, n=8–10 mice/group; E, n=5–9 mice/group; K, n=4–7 mice/group) with P value determined by (C,K) 2-way ANOVA with post-hoc Tukey’s correction for multiple comparisons or (E) by Student’s t-test. (C) ****P<0.0001 on day 49 shown comparing STAN to both STAN/OT-I T cell groups, (E) ****P<0.0001 at day 30 for EO771-OVA and day 23 for EO771, (K) ***P≤0.001 on day 40 shown comparing STAN to both PBS/OT-I T cells and ****P<0.0001 on day 49 shown comparing STAN to STAN + OT-I T cells. Kaplan-Meier survival curves of mice treated as indicated using 1500 mm³ tumor volume as endpoint criteria and P value was determined by log-rank test; **P≤0.01, ***P≤0.001, and ****P<0.0001 compared to PBS.