Combination of KRAS ASO and RIG-I agonist in extracellular vesicles transforms the tumor microenvironment towards effective treatment of KRAS-dependent cancers

Abstract

Rationale: Mutations in the KRAS gene drive many cancers, yet targeting KRAS mutants remains a challenge. Here, we address this hurdle by utilizing a nucleic acid-based therapeutic strategy delivered via extracellular vesicles (EVs) to simultaneously inhibit KRAS mutants and activate the RIG-I pathway, aiming to enhance anti-tumor immunity. Methods: Antisense oligonucleotides against KRAS mutants (KRAS ASOs) and RIG-I agonist immunomodulatory RNA (immRNA) were loaded into EVs and administered to KRAS-mutant cancer models. The therapeutic effects were assessed in colorectal and non-small cell lung cancer (NSCLC) tumor models, as well as patient-derived pancreatic cancer organoids. Immune responses were evaluated by analyzing tumor microenvironment's changes, dendritic cell activation, and T cell memory formation. The treatment efficacy was evaluated based on the tumor development and overall survival. Results: The KRAS-ASO and immRNA combination treatment induced immunogenic tumor cell death and upregulated interferons in KRAS-dependent cancers. In a colorectal tumor model, the therapy shifted the tumor microenvironment to an immunogenic state, activated dendritic cells in sentinel lymph nodes, and promoted memory T cell formation. In an aggressive NSCLC model, the treatment resulted in a strong anti-tumor activity and extended survival without any adverse effects. Validation in patient-derived pancreatic cancer organoids confirmed the clinical translation potential of this approach. Conclusions: EV-mediated delivery of ASOs and immRNA effectively inhibits KRAS mutants and activates RIG-I, leading to a robust anti-tumor immune response. This strategy holds promise for effectively treating KRAS-driven cancers and improving clinical outcomes.

Keywords: KRAS mutation; RIG-I agonist; antisense oligonucleotides; cancer therapy; extracellular vesicles.

Figure 1

RBCEVs can be efficiently loaded with small nucleic acids for delivery to KRAS-mutated tumor cells. (A) Schematic for the RBCEV purification process from red blood cells (RBCs). (B) Western blot analysis of proteins in RBCs and purified RBCEVs. (C-D) Dynamic light scattering analysis of (C) size distribution and (D) zeta potential of unloaded EVs, ASO-loaded EVs (ASO-EVs), and immRNA-loaded EVs (immR-EVs). (E) Loading efficiency of NC ASO in RBCEVs using REG1 transfection reagent determined by gel electrophoresis (n=3). (F) Representative transmission electron microscopy images of RBCEVs and NC-ASO-loaded RBCEVs. Scale bar: 100 nm.

Figure 2

Combination treatment of ASOs targeting KRAS mutations with RIG-I agonist immunomodulatory RNA in RBCEVs synergistically activates the RIG-I signaling pathway and induces apoptosis in KRAS-mutated cancer cells. (A) Heatmap visualization of DDX58 and IFNB expression in A427 lung cancer cells (KRAS G12D), AsPC-1 pancreatic cancer cells (KRAS G12D), PDO67 patient-derived pancreatic cancer organoids (KRAS G12D), CT26 colorectal cancer cells (Kras G12D), and H441 lung cancer cells (KRAS G12V) after treatments with 25 µg/mL RBCEVs or RBCEVs loaded with either NC ASO (NC-ASO-EVs, 50 µg/mL EVs), KRAS ASO (KRAS-ASO-EVs, 25 µg/mL EVs), immRNA (immR-EVs, µg/mL EVs), combined NC ASO and immRNA (NC-ASO-EVs + immR-EVs, 50 µg/mL EVs), or combined KRAS ASO and immRNA (KRAS-ASO-EVs + immR-EVs, 50 µg/mL EVs) for 48 h. Z-Score was computed using the formula: (Gene expression value - mean expression across all samples) / Standard Deviation. The Gplots package in R software was used to plot a heatmap based on the computed Z-score. (B) Flow cytometric analysis of Annexin V/SYTOX Blue staining in A427, AsPC-1, and H441 cells 24 h post-treatment with different RBCEV formulations. (C-E) Proportion of late apoptosis (Annexin⁺, SYTOX Blue⁺) in (C) A427, (D) AsPC-1, and (E) H441 cells which were treated as described in (B) (n = 3). (F) Assessment of apoptosis in PDO67 patient-derived pancreatic cancer organoids after treatments with different RBCEV formulations for 72 h by staining the treated organoids with acridine orange (AO) and propidium iodide (PI). The graphs present mean ± SEM. *P < 0.05, **P < 0.01, and ***P < 0.001, determined using One-Way ANOVA test.

Figure 3

Combination of KRAS ASO and immRNA delivered by RBCEVs promotes immune cell infiltration into KRAS-addicted cancer spheroids. (A) Schematic for the evaluation of peripheral blood mononuclear cell (PBMC) infiltration into 3D cultures of A427 (KRAS G12D) and H441 (KRAS G12V) cells pre-treated with KRAS-ASO-EVs and/or immR-EVs, assessed by flow cytometry. (B-C) Flow cytometric analysis showing the percentage of CD45⁺ immune cells infiltrated into spheroids formed by (B) A427 and (C) H441 cells 48 h after the co-culture. The graphs present mean ± SEM. **P < 0.01, and ***P < 0.001, determined using One-Way ANOVA test.

Figure 4

Combination of Kras G12D ASO and immRNA in RBCEVs induces potent anti-cancer activity in a genetically engineered mouse model of NSCLC with Kras G12D mutation. (A) Schematic for the tumor induction in the lung of Kras^LSL-G12D/+;p53^fl/fl (KP) mice and intratracheal administration schedule. 8 weeks (56 days) post-infection with Cre recombinase-expressing lentivirus (Lenti-Cre), the infected KP mice were intratracheally administered with (1) PBS or RBCEVs loaded with either (2) NC ASO (NC-ASO-EVs, 20 mg/kg RBCEVs), (3) Kras G12D ASO (Kras-ASO-EVs, 10 mg/kg RBCEVs), (4) immRNA (immR-EVs, 20 mg/kg RBCEVs), or (5) combined Kras G12D ASO and immRNA (Kras-ASO-EVs + immR-EVs, 20 mg/kg RBCEVs) every three days for three weeks (n = 6-7 mice/group). (B) Representative images of Haematoxylin and eosin (H&E) stained lung tissues from KP mice at the end of the study. Cell types were detected and analyzed using QuPath software. Red: tumor cells, blue: immune cells, green: other cell types. Scale bar: 1 mm. (C) Percentages of tumor cells in the lung of treated mice as described in (B) detected by QuPath software. (D) Representative immunohistochemistry (IHC) images of NK1.1 and CD8 protein expression (brown) in lung tumors of treated mice at the end of the study. (E-F) Densities of (E) NK cells and (F) CD8⁺ T cells within tumor areas quantified by IHC analysis as described in (D). (G) Change in body weight of mice receiving treatments as described in (A). (H-L) Serum concentrations of (H) Creatin, (I) blood urea nitrogen (BUN), (J) Albumin, (K) Aspartate-Aminotransferase (AST), and (L) Alanine Transaminase (ALT) in the treated mice as described in (A) at the end of the study. All bar graphs represent mean ± SEM. Ns - not significant, *P < 0.05, **P < 0.01, and ***P < 0.001 determined by One-Way ANOVA test.

Figure 5

Combination treatment of Kras ASO and immRNA in RBCEVs induces potent anti-cancer effect against aggressive cold tumors. (A) Schematic for generating aggressive orthotopic lung tumor with mCherry-Luc-KP tumor cells in C57BL/6 mice and an intratracheal administration schedule. Mice with strong tumor signal at the lung were administrated with PBS or RBCEVs loaded with either NC ASO (NC-ASO-EVs, 20 mg/kg RBCEVs), Kras G12D ASO (Kras-ASO-EVs, 10 mg/kg RBCEVs), immRNA (immR-EVs, 10 mg/kg RBCEVs), or combined Kras G12D ASO and immRNA (Kras-ASO-EVs + immR-EVs, 20 mg/kg RBCEVs) every three days. (B) Representative bioluminescent images of orthotopic KP lung tumors expressing luciferase in treated xenograft mice over time. (C) Quantification of tumor progression in mice using the average bioluminescent signals following treatments over time (n = 4-5 mice/group). (D) Kaplan-Meier survival curves of treated mice as described in (A) (n = 6-8 mice/group). (E-G) Representative flow cytometry plot (left) and quantification (right) of (E) granulocytic and monocytic myeloid-derived suppressor cells (gMDSCs and mMDSCs, respectively) in CD45⁺CD11⁺ cells, (F) M1- and M2-like tumor-associated macrophages, and (G) populations of CD11c⁺MHC-II⁺ dendritic cells within the lung of treated mice at the end of study as described in (A) and (B). The graphs present the mean ± SEM. Ns - not significant, *P <0.05, **P < 0.01, ***P < 0.001 determined by two-way ANOVA (C), Gehan-Breslow-Wilcoxon test (D), or One-Way ANOVA test t-test (E-G).

Figure 6

Combination of KRAS ASO and immRNA shapes the immunogenic and tumoricidal tumor microenvironment of CT26 tumor. (A) Intratumoral (i.t.) administration scheme for BALB/c mice with CT26 tumor. Mice with 50 mm³ subcutaneous tumors were intratumorally administered with PBS or RBCEVs loaded with either NC ASO (NC-ASO-EVs, 10 mg/kg RBCEVs), Kras G12D ASO (Kras-ASO-EVs, 5 mg/kg RBCEVs), immRNA (immR-EVs, 5 mg/kg RBCEVs), or combined Kras G12D ASO and immRNA (Kras-ASO-EVs + immR-EVs, 10 mg/kg RBCEVs) every three days, n = 5 mice/group. (B) Tumor volume of the CT26 tumor-bearing mice over time after the treatment. (D) Weight of tumor masses collected from treated mice at the end of the study. (E) Heatmap visualization of gene expression in tumors collected from treated mice at the end of the study using Gplots package in R software. (F-I) Representative flow cytometry histogram (left) and quantification (right) of (F) CD8⁺ T cells and (G) NK cells in total live tumor single cells, (H) M1- and M2- like tumor-associated macrophages, and (I) granulocytic myeloid-derived suppressor cells in intratumoral CD45⁺CD11⁺ cells. The graphs present the mean ± SEM. *P <0.05, **P < 0.01, ***P < 0.001 determined by two-way ANOVA (B) or One-Way ANOVA test (D, F-I).

Figure 7

Combination of Kras ASO and immRNA induces immunogenic cell death effect in a CT26 tumor model. (A) Representative flow cytometry plot and (B) quantification of CD86 expression in dendritic cells at the tumor-draining lymph node of treated mice at the end of the study. (C) Representative flow cytometry plot and quantification of (D) effector memory CD4⁺ T cells (CD3⁺CD4⁺CD44⁺CD62L^-) and effector memory CD8⁺ T cells (CD3⁺CD8⁺CD44⁺CD62L^-) in the spleen of treated mice at the end of the study. The graphs present the mean ± SEM. *P <0.05, **P < 0.01, ***P < 0.001 determined by One-Way ANOVA test.