Genetically engineered nanovesicles mobilize synergistic antitumor immunity by ADAR1 silence and PDL1 blockade

Abstract

Growing evidence has proved that RNA editing enzyme ADAR1, responsible for detecting endogenous RNA species, was significantly associated with poor response or resistance to immune checkpoint blockade (ICB) therapy. Here, a genetically engineered nanovesicle (siAdar1-LNP@mPD1) was developed as an RNA interference nano-tool to overcome tumor resistance to ICB therapies. Small interfering RNA against ADAR1 (siAdar1) was packaged into a lipid nanoparticle (LNP), which was further coated with plasma membrane extracted from the genetically engineered cells overexpressing PD1. siAdar1-LNP@mPD1 could block the PD1/PDL1 immune inhibitory axis by presenting the PD1 protein on the coating membranes. Furthermore, siAdar1 could be effectively delivered into cancer cells by the designed nanovesicle to silence ADAR1 expression, resulting in an increased type I/II interferon (IFN-β/γ) production and making the cancer cells more sensitive to secreted effector cytokines such as IFN-γ with significant cell growth arrest. These integrated functions confer siAdar1-LNP@mPD1 with robust and comprehensive antitumor immunity, as evidenced by significant tumor growth regression, abscopal tumor prevention, and effective suppression of lung metastasis, through a global remodeling of the tumor immune microenvironment. Overall, we provided a promising translatable strategy to simultaneously silence ADAR1 and block PDL1 immune checkpoint to boost robust antitumor immunity.

Keywords: ADAR1 silence; PDL1 blocking; cancer immunotherapy; genetically engineered cell membrane; lipid nanoparticle.

Graphical abstract

Figure 1

Preparation and characterization of siAdar1-LNP@mPD1 nanovesicles (A) Schematic illustration of siAdar1-LNP@mPD1 nanovesicle preparation and the principle of antitumor immune response. (B–D) CLSM images (B), flow cytometry analysis (C), and western blot (WB) analysis (D) of stably expressing PD1 on engineered CHO cell membrane. Scale bar, 5 μm. (E) The interaction between mPD1 NVs and anti-PD1 antibody detected by co-immunoprecipitation (coIP). For immunoprecipitation, anti-PD1 antibody was used to pull down mPD1 NVs, while IgG was used as the control. (F) CLSM images of mPD1 NVs binding to PDL1-EGFP that is intentionally overexpressed on 4T1 cells, the co-localization of mPD1 NVs and PDL1 proteins are represented by the yellow colored dots as indicated by white arrows. Scale bar, 20 μm). mPD1 NVs (40 μg/mL) were incubated with 4T1 cells at 4°C for 1 h. (G) CLSM images of siAdar1-LNP with Cy5-labeled siAdar1 (red fluorescence) and LNP loaded with DiO (green fluorescence). Scale bars, 20 μm. (H) Gel fluorescence imaging of free siAdar1 and siAdar1-LNP. (I and J) Transmission electron microscopy images (I) and dynamic light scattering (DLS) results (J) of siAdar1-LNP and siAdar1-LNP@mPD1. Scale bar, 100 nm. (K and L) SDS-PAGE electrophoresis patterns (K) and WB analysis (L) of siAdar1-LNP (I), mPD1 NVs (II), and siAdar1-LNP@mPD1 (III).

Figure 2

In vitro immune checkpoint inhibition and ADAR1 silence of siAdar1-LNP@mPD1 nanovesicle (A) Relative viability of 4T1 cells and NIH3T3cells after incubation with siAdar1-LNP, as determined by CCK-8 assays. Error bar: mean ± SD (n = 5). (B) CLSM images of 4T1 cells treated with 5-FAM-labeled siAdar1-LNP for 4, 24, and 48 h, or free siAdar1 (for 48 h). Scale bar, 20 μm. (C–E) The qRT-PCR analysis of the relative expression of ADAR1 mRNA (C), WB analysis (D), and the relative expression (E) of ADAR1 protein from 4T1 cells after different treatments (G1 to G5). G1, blank control without any treatment; G2, siNC-LNP; G3, free siAdar1; G4, siAdar1-LNP; G5, positive control with siAdar1 transfected by Lipofectamine 3000 (mean ± SD; ∗∗∗∗p < 0.0001; n = 3). (F) CLSM images of the nanoparticle internalization behaviors in 4T1 cells when treated with siAdar1-LNP@m, siAdar1-LNP@mPD1, and siAdar1-LNP@mPD1 + anti-PDL1. The siAdar1 was labeled with 5-FAM for fluorescence imaging. Scale bar, 20 μm. (G) Schematic diagram of IFN-γ response (left) and the relative viability of 4T1 cells treated by INF-γ with different concentrations (0.1–0.5 μg/mL) (right). 4T1 cells were pretreated with PBS or siAdar1-LNP@mPD1 before INF-γ incubation (mean ± SD; ∗∗∗∗p < 0.0001; n = 5).

Figure 3

Therapeutic effect on the subcutaneous 4T1 tumor model (A) Schematic representation of the therapeutic timeline of siAdar1-LNP@mPD1 on the subcutaneous 4T1 tumor model. (B) In vivo fluorescence imaging of the 4T1 tumor-bearing mouse at different time points after intravenous (i.v.) injection of siAdar1-LNP@mPD1, siAdar1-LNP@m, or free siAdar1. (C) Ex vivo fluorescence imaging of excised major organs and tumors from the mice after 24 h i.v. injection of different formulation. (D) Analysis of the mean fluorescence intensity (MFI) of major organs and tumors by ImageJ software. (E) Average and individual tumor growth kinetics of mice after i.v. injection with PBS, siAdar1-LNP@m, siNC-LNP@mPD1, or siAdar1-LNP@mPD1 (mean ± SD; n = 10; ∗∗p < 0.01, ∗∗∗∗p < 0.0001). (F and G) Photographs (F) and weights (G) of tumors excised from all treated mice on day 20 (mean ± SD; n = 10; ∗∗p < 0.01, ∗∗∗∗p < 0.0001). (H) H&E, Ki67, and TUNEL staining of tumor slices after different treatments. Scale bars, 100 μm. (I–K) The qRT-PCR analysis of the relative expression of ADAR1 mRNA (I), WB analysis (J), and the relative expression (K) of ADAR1 protein from excised 4T1 tumor cells after different treatments (mean ± SD; ∗∗∗∗p < 0.0001; n = 3).

Figure 4

Immune response in the tumor site (A) Schematic diagram of IFN production after treatment. (B) Cytokine levels (IFN-β, IFN-γ, TNF-α, IL-12, IL-10, and TGF-β1) in excised tumors after different treated mice by ELISA analysis. G1, PBS; G2, siAdar1-LNP@m; G3, siNC-LNP@mPD1; G4, siAdar1-LNP@mPD1 (mean ± SD; ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001; n = 3). (C–F) Typical FCM plots and quantification results of CD3⁺CD8⁺ T cells (C), CD3⁺CD4⁺ T cells (D), Tregs (E), and MDSCs (F) in tumors of four different treatment groups (mean ± SD; ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001; n = 3). (G) Multiple immunofluorescence images displayed the infiltrated CD4⁺ T cells (red) and CD8⁺ T cells (green) within tumors tissue. Scale bar, 100 μm. An enlarged image of the upper panel is shown in the lower panel. Scale bar, 25 μm. (H) Volcano plot of disregulated genes in tumor tissues with siAdar1-LNP@mPD1 treatment versus PBS treatment. (I) Heatmap showing disregulated genes in mice tumors after siAdar1-LNP@mPD1 and PBS treatments (fold change ≥ ± 10, false discovery rate < 0.05; n = 3). (J) KEGG enrichment analysis of the disregulated genes between siAdar1-LNP@mPD1 and PBS treatment.

Figure 5

Therapeutic effect on a lung metastasis 4T1 tumor model (A) Schematic representation of the timeline of a lung metastasis inhibition model of 4T1 tumor-bearing mice after different treatments. (B) Representative photographs (upper two panels) and H&E staining (lower panel) of isolated lungs on the 40th day post-treatment. Scale bar, 5 mm. (C) Summary data analysis of lung metastasis nodules in (B) (mean ± SD; ∗∗p < 0.01, ∗∗∗∗p < 0.0001; n = 5). (D–F) Typical FCM plots of CD44⁺ cells (gated on CD4⁺ or CD8⁺ T cells) (D), effector memory T cells (T_EM) (gated on CD4⁺CD44⁺ or CD8⁺CD44⁺ T cells) (E), and relevant quantification results (F) in the spleens after various treatments (mean ± SD; ∗∗∗p < 0.001, ∗∗∗∗p < 0.0001; n = 3).

Figure 6

Therapeutic effect on a bilateral 4T1 tumor model (A) Schematic representation of the therapeutic timeline of siAdar1-LNP@mPD1 on a bilateral 4T1 tumor model. The primary tumors were treated by intratumoral injection of different formulations, and the abscopal tumors were not treated. (B and C) Growth curves and the average volume of primary tumors (B) and abscopal tumors (C) in different groups were recorded every other day, including: G1, PBS; G2, siAdar1-LNP@m; G3, siNC-LNP@mPD1; G4, siAdar1-LNP@mPD1 (mean ± SD; n = 10). (D) Quantification results of average tumor volume on the day 20 post-treatment (mean ± SD; ∗∗p < 0.01, ∗∗∗∗p < 0.0001; n = 3). (E) Survival curves of the mice after different treatments as indicated (∗∗∗∗p < 0.0001; n = 10). (F) Schematic diagram of tumor microenvironment reshaped by siAdar1-LNP@mPD1. (G) Multiple immunofluorescence images displayed the infiltrated CD4⁺ T cells (red) and CD8⁺ T cells (green) within abscopal tumors. Scale bar, 100 μm. Enlarged images of the upper panel are shown in the lower panel. Scale bar, 25 μm. (H) Quantification results of different tumor-infiltrating immune cells in abscopal tumors after different treatments as indicated (mean ± SD; ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001, ∗∗∗∗p < 0.0001; n = 3). (I) Cytokine levels in excised tumors after different treatments examined by ELISA analysis (mean ± SD; ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001; n = 3). (J) Typical FCM plots of CD3⁺CD8⁺ T cells and quantification results of CD3⁺CD8⁺ and CD3⁺CD4⁺ T cells in the blood after various treatments. CD8⁺ and CD4⁺ values (right panel) are calculated by multiplying the proportion of positive cells, respectively, indicated in the left panel and in Figure S37B (gated on P1 region) by a normalization factor. The normalization factor is the percentage of P1 region in ALL EVENTS (FSC/SSC) as shown in Figure S37A (mean ± SD; ∗p < 0.05, ∗∗p < 0.01; n = 3).