Bacterial protoplast-derived nanovesicles carrying CRISPR-Cas9 tools re-educate tumor-associated macrophages for enhanced cancer immunotherapy

Abstract

The CRISPR-Cas9 system offers substantial potential for cancer therapy by enabling precise manipulation of key genes involved in tumorigenesis and immune response. Despite its promise, the system faces critical challenges, including the preservation of cell viability post-editing and ensuring safe in vivo delivery. To address these issues, this study develops an in vivo CRISPR-Cas9 system targeting tumor-associated macrophages (TAMs). We employ bacterial protoplast-derived nanovesicles (NVs) modified with pH-responsive PEG-conjugated phospholipid derivatives and galactosamine-conjugated phospholipid derivatives tailored for TAM targeting. Utilizing plasmid-transformed E. coli protoplasts as production platforms, we successfully load NVs with two key components: a Cas9-sgRNA ribonucleoprotein targeting Pik3cg, a pivotal molecular switch of macrophage polarization, and bacterial CpG-rich DNA fragments, acting as potent TLR9 ligands. This NV-based, self-assembly approach shows promise for scalable clinical production. Our strategy remodels the tumor microenvironment by stabilizing an M1-like phenotype in TAMs, thus inhibiting tumor growth in female mice. This in vivo CRISPR-Cas9 technology opens avenues for cancer immunotherapy, overcoming challenges related to cell viability and safe, precise in vivo delivery.

Fig. 1. Schematic design of E. coli protoplast-derived nanovesicles (sgPik3cg-DHP/DGA-NVs) for TAM-selective genome editing to enhance anti-tumor efficacy.

In this illustration, we outline the creation of sgPik3cg-DHP/DGA-NVs for targeted genome editing in tumor-associated macrophages (TAMs) to enhance anti-tumor effects. The process begins with the construction of E. coli expressing the Cas9-sgPik3cg complex. Subsequently, the bacterial outer membrane, which possesses high endotoxicity, is removed. This results in the formation of sgPik3cg-DHP/DGA-NVs, which encapsulate a substantial amount of Cas9-sgPik3cg ribonucleoproteins (RNPs) and CpG-rich genomic DNA. These nanovesicles (NVs) are produced through a series of extrusion steps and are further modified with a pH-responsive phospholipid derivative (DHP) and a phospholipid derivative targeted specifically to TAM (DGA). Upon intravenous injection, sgPik3cg-DHP/DGA-NVs accumulate in tumor tissues due to their prolonged circulation capability and the enhanced permeability and retention (EPR) effect. Within the acidic microenvironment of the tumor, PEG₂₀₀₀ separates from DHP, triggering the recognition and internalization of DGA-functionalized NVs by TAMs via macrophage galactose-type lectin (MGL) receptor-mediated endocytosis. This process enables TAM-specific genome editing of Pik3cg and activation of toll-like receptor 9 (TLR9) in vivo, resulting in the reprogramming of M2-like TAMs into an anti-tumor M1-like phenotype and facilitating tumor immunotherapy.

Fig. 2. The preparation and characterization of sgPik3cg-DHP/DGA-NVs.

a A total of 1 × 10⁹ colony forming unit (CFU) E.coli derived protoplasts were added with indicated amounts of RhB-DHP and Cy5.5-DGA for physical extrusion. The decoration efficiency and quantitative analysis of RhB-DHP and Cy5.5-DGA for nanovesicles were determined through flow cytometry analysis. n = 3 biologically independent samples. b Representative TEM images illustrate sgPik3cg-NVs, both with and without DHP/DGA decoration. Scale bar, 200 nm. c The diameters of sgPik3cg-NVs and sgPik3cg-DHP/DGA-NVs were analyzed using Nanosight. For (b, c), experiments were independently conducted three times with similar results. d The ratio of RhB-positive NVs was quantified through flow cytometry analysis after treating 1 × 10¹⁰ CFSE-labeled NVs with RhB-DHP and DGA in 100 μL of pH 6.5 PBS at indicated time points (0 h and 24 h). e Quantitative analysis of sgPik3cg-RhB-DHP/Cy5.5-DGA-NVs treated with different pH values of PBS at indicated time points. n = 3 biologically independent samples for (d, e). Statistical analysis was performed using two-way ANOVA with Bonferroni’s multiple comparison test. f Direct injection of 50 μL PBS containing 5 × 10⁹ CFSE-labeled sgPik3cg-RhB-DHP/DGA-NVs into 4T1 tumor tissues and tumor-adjacent tissues. Corresponding samples were harvested from tumor-bearing mice at 24 h post-injection for fluorescence photography. The representative images presented are from a sample size of n = 3 mice. Scale bar = 50 μm. g Measurement of IL-6 and TNF-α levels in 100 μL serum of 4T1 tumor-bearing mice at 2 h and 24 h after in vein injections of 100 μL PBS containing different types of NVs (dose: 1 × 10¹⁰ per mouse). n = 5 mice per group. Statistical analysis was performed using one-way ANOVA with Dunnett’s multiple comparison test. Data are represented as mean ± SD. *P < 0.05, **P < 0.01 and ***P < 0.001, ns, no significant change. The exact P-value and source data are provided as a Source Data file.

Fig. 3. The components analysis of sgPik3cg-DHP/DGA-NVs.

a, b Size distribution and abundance of DNA and RNA from E. coli, protoplast, OMVs and sgPik3cg-DHP/DGA-NVs were determined using LabChip analyzer and Agilent 2100 bioanalyzer. c, d The presence of CpG-rich genomic DNA sequences and sgRNA targeting Pik3cg was confirmed through PCR amplification and agarose gel electrophoresis. e Subcellular localization of the top 200 abundant proteins, ranked according to their iBAQ protein intensity, was identified in E. coli, protoplasts, OMVs and sgPik3cg-DHP/DGA-NVs through proteomic analysis, with 100 μg of protein used for each sample. A pie chart illustrates the variety of protein types categorized by their subcellular localization in each sample. f Analysis of the Clusters of Orthologous Groups (COG) was conducted for the top 200 abundant proteins identified in sgPik3cg-DHP/DGA-NVs. g The presence of Cas9 protein in E. coli and sgPik3cg-DHP/DGA-NVs was determined using western blotting. All experiments for (a–g) were independently repeated three times, yielding consistent results. Source data are provided as a Source Data file.

Fig. 4. In vitro uptake of NVs by M2-like macrophages via MGL-mediated endocytosis.

a–c 5 × 10⁵ M2-like bone marrow-derived macrophages (M2-BMDMs) were seeded in 24-well plates and treated with 6 × 10⁸ CFSE-sgPik3cg-DGA-NVs for 3 h, followed by flow cytometry analysis and fluorescent microscopy. In some cases, macrophages were pre-incubated with GlcNAc or GalNAc (100 mmol/L) for 1 h before NVs treatment. Red, F4/80; green, CFSE labeled NVs; blue, DAPI nuclear staining. Scaled bar = 50 μm. n = 3 biologically independent samples. d A total of 1 × 10⁵ macrophages with above mentioned treatments was harvested, lysed and the fluorescence intensity was quantified using a microplate reader (Em: 488 nm; Ex: 530 nm). n = 3 biologically independent samples. e–h 5 × 10⁵ M2-BMDMs were treated with 6 × 10⁸ CFSE-sgPik3cg-DGA-NVs or CFSE-sgPik3cg-DHP/DGA-NVs (pre-treated with pH 6.5 or 7.4 PBS for 24 h) for 3 h and then subjected to flow cytometry analysis, microphotography and fluorescence intensity quantification. Red, F4/80; green, CFSE labeled NVs; blue, DAPI nuclear staining. Scaled bar = 50 μm. n = 3 biologically independent samples. i M2-BMDMs subjected to the aforementioned NVs treatment were fixed, stained with fluorescent-labeled F4/80 and Cas9 antibodies and imaged at indicated time points. Red, F4/80; green, Cas9; blue, DAPI nuclear staining. Scaled bar = 10 μm. j, k The levels of Cas9 and sgRNA targeting Pik3cg in 5 × 10⁵ M2-BMDMs treated with 6 × 10⁸ sgPik3cg-DHP/DGA-NVs (pH 6.5 PBS pre-treatment) for 3 h after which the medium was replaced with fresh medium for another 9 h. The cells were examined by western blotting and agarose gel electrophoresis. l The NVs-mediated delivery of CpG-rich genomic DNA fragments into 5 × 10⁵ M2-BMDMs incubated with 6 × 10⁸ NVs for 3 h was detected by PCR amplification and agarose gel electrophoresis. The experiments for (i–l) were independently repeated three times with similar results. Data are presented as the means ± SD. Statistical analyses were performed using one-way ANOVA with Dunnett’s multiple comparison test. *P < 0.05, **P < 0.01 and ***P < 0.001, ns, no significant change. The exact P-value and source data are provided as a Source Data file.

Fig. 5. The genome editing efficiency of Pik3cg and phenotypic analysis in M2-like macrophages after sgPik3cg-DHP/DGA-NVs treatment.

a, b 1.5 × 10⁶ M2-BMDMs in 6-well plates were treated with 1.8 × 10⁹ sgPik3cg-DHP/DGA-NVs (pH 6.5 PBS pre-treatment) for 6 h, followed by replacement with fresh medium. Macrophages were harvested 48 h post-incubation with NVs for T7E1 analysis to assess indel formation and western blotting to determine PI3Kγ levels. The experiments were repeated three times independently with similar results. c 5 × 10⁵ 293 T cells transfected by TLR9 overexpression plasmid, NF-κB reporter plasmid and β-gal reference plasmid (each plasmid: 2 μg) for 24 h were further treated with 1.8 × 10⁹ different types of NVs or ODN1826 (TLR9 agonist, 10 μmol/L) for 6 h, followed by replacement with fresh medium for another 18 h to assess luciferase activity. In some cases, ODN2088 (TLR9 inhibitor, 10 μmol/L) and the corresponding scramble control were added 30 min before NVs treatment and co-incubated for another 24 h. n = 3 biologically independent samples. d 1.5 × 10⁶ M2-BMDMs were treated with 1.8 × 10⁹ different types of NVs for 6 h, followed by replacement with fresh medium for another 42 h. ODN1826 (10 μmol/L) or IPI549 (PI3Kγ inhibitor, 1 μmol/L) were added for 48 h, and then cells were harvested for western blotting to examine the downstream proteins of PI3Kγ and TLR9 (except p-IRAK4). p-IRAK4/IRAK4 was examined 1 h after NVs or reagents treatment. In some cases, 10 μmol/L ODN2088 was added 30 min before NVs treatment and co-incubated for 48 h. The experiments were repeated three times independently with similar results. e–h M2-BMDMs with the above-mentioned NVs and IPI549 treatments were used to examine PIP2 transition ratio, macrophage phenotype markers by RT-qPCR and flow cytometry, and cytokine levels by ELISA. n = 3 biologically independent samples. Data are represented as mean ± SD. Statistical analyses were performed using one-way ANOVA with Dunnett’s multiple comparison test. *P < 0.05, **P < 0.01 and ***P < 0.001, ns, no significant change. The exact P-value and source data are provided as a Source Data file.

Fig. 6. In situ genetic reprogramming and repolarization of TAM in 4T1 tumor-bearing mice via sgPik3cg-DHP/DGA-NVs.

a Fluorescence images of the 4T1 tumor-bearing mice and quantitative analysis of tumor site fluorescence (red circle) at indicated time points after injection of different types of Cy5-NVs (dose: 1 × 10¹⁰ NVs per mouse). n = 3 mice per group. b Organ tissues from mice treated with the above-mentioned NVs at 72 h were photographed, and fluorescence intensity was analyzed. n = 3 mice per group. c, d Cy5 fluorescence intensity in plasma and tumor of mice with above-mentioned NVs treatment was detected at indicated time points (Em:644 nm; Ex: 665 nm). n = 3 mice per time point. e Immunofluorescence staining of tumor sections harvested from mice at 72 h after Cy5-sgPik3cg-DHP/DGA-NVs injection. Red, Cy5-NVs; green, F4/80; blue, DAPI nuclear staining. Scaled bar = 50 μm for 400× and 10 μm for amplification. The representative images presented are from a sample size of n = 3 mice. f Fluorescence co-localization analysis of Cy5-NVs and TAMs using image J software. Fluorescence intensity profile representing the value indicated by the red rectangle. g, h Flow cytometry was used to analyze the ratio of Cy5⁺ TAMs from mice at 72 h after different types of NVs injection (dose: 1 × 10¹⁰ NVs per mouse). n = 3 mice per group. i, j The ratio of Cy5⁺ TAMs at different time points after treatment with 1 × 10¹⁰ sgPik3cg-DHP/DGA-NVs measured by flow cytometry. n = 3 mice per time point. k, l T7E1 analysis for indel formation in TAMs from mice at different time points after treatment with sgPik3cg-DHP/DGA-NVs. A representative image is presented. n = 3 mice per time point. Data are presented as the means ± SD. Statistical analyses were performed using one-way ANOVA with Dunnett’s multiple comparison test. *P < 0.05, **P < 0.01 and ***P < 0.001, ns, no significant change. The exact P-value and source data are provided as a Source Data file.

Fig. 7. Anti-tumor immunotherapy was achieved through systemic injection of TAM-targeted sgPik3cg-DHP/DGA-NVs in 4T1 tumor-bearing mice.

a Schematic diagram of 4T1 tumor-bearing mice with in-vein injections of NVs (1 × 10¹⁰ NVs every two days) or intragastric administration of IPI549 (15 mg/kg every day). b, c Tumor images and tumor weights in mice treated with different NVs or IPI549 on day 16 post-tumor cell inoculation. d Tumor volume changes in various treatment groups of 4T1 tumor-bearing mice. e Representative H&E staining images of 4T1 tumor sections. Scale bar, 100 μm. f Phosphoinositide 3-kinase activity in TAMs, characterized by the conversion ratio of PIP2 to PIP3, was determined by ELISA on day 16 after tumor model establishment. n = 5 mice per group for (b–f). g Levels of PI3Kγ and TLR9 pathways related molecules (p-C/EBPβ, p-AKT, p-IRAK4 and p-p65) in TAMs isolated from mice receiving the mentioned treatments, detected by western blotting on day 8 and 16 post-tumor cell inoculation. The experiments were repeated three times independently with similar results. h, i Effects of sgPik3cg-DHP/DGA-NVs on TAM phenotype were determined by flow cytometry analysis (CD86 and CD206) and qRT-PCR assay (iNOS and Arg1). n = 5 mice for each time point. j Cytokine levels in TAMs from 4T1 tumor-bearing mice with different treatments were measured on day 16 after tumor model establishment. n = 5 mice per group. Data are presented as the means ± SD. Two-way ANOVA with Dunnett’s multiple comparison test was used in (d). Other statistical analyses were performed using one-way ANOVA with Dunnett’s multiple comparison test except the ELISA assay of IL-6 in (j) (Kruskal-Wallis test with Dunn’s multiple comparisons test). *P < 0.05, **P < 0.01 and ***P < 0.001, ns, no significant change. The exact P-value and source data are provided as a Source Data file.

Fig. 8. sgPik3cg-DHP/DGA-NVs treatment reshaped the tumor immune microenvironment and induced potent antitumor immunity against breast cancer.

a Volcano plot of the differentially expressed genes (DEGs) based on RNA-seq analysis of 4T1 tumor tissue from the sgPik3cg-DHP/DGA-NVs group (1 × 10¹⁰ NVs every two days) compared to the control group on day 16 post-model establishment. n = 3 mice per group. b Enriched gene ontology (GO) terms associated with DEGs related to immune activation. n = 3 mice per group. c GSEA enrichment analysis showing the enrichment of genes upregulated in positive regulation of innate immune response and T cell activation. n = 3 mice per group. d Cytokine levels in 4T1 tumors of mice after treated with either NVs (1 × 10¹⁰ NVs every two days) or IPI549 (15 mg/kg every day) were determined by ELISA on day 16 following the establishment of the tumor model. n = 5 mice per group. e The influence of sgPik3cg-DHP/DGA-NVs on intratumoral T cell activation and proliferation (percentage of IFN-γ⁺, ki67⁺ and Granzyme B⁺ cells in CD4⁺ T cells and CD8⁺ T cells) of mice with above mentioned treatments was evaluated using flow cytometry analysis at 8 days and 16 days post-tumor cell inoculation. n = 5 mice for each time point. Data are presented as the means ± SD. Statistical analyses were performed using one-way ANOVA with Dunnett’s multiple comparison test except the ELISA assay of IL-6 in d (Kruskal-Wallis test with Dunn’s multiple comparisons test). *P < 0.05, **P < 0.01 and ***P < 0.001. ns, no significant change. The exact P-value and source data are provided as a Source Data file.