Simultaneous Targeting of Tumor Cells and Tumor-Associated Macrophages To Reprogram Glioblastoma Using Trypsinized Extracellular Vesicles Carrying Tumor Suppressive MicroRNA

Abstract

Glioblastoma (GBM) remains difficult to treat due to poor drug delivery across the blood-brain barrier and an immunosuppressive tumor microenvironment (TME). Tumor-suppressive microRNAs (miRNAs) offer a promising strategy to reprogram both tumor cells and the TME, but inefficient delivery systems limit their clinical application. We previously reported that tumor-suppressive miR-138 regresses tumor growth in preclinical GBM models. Here, we demonstrate that trypsin digestion of extracellular vesicles (EVs) enhances labeling efficiency with folate (FA), enhancing selective targeting of folate receptor (FR)-positive GBM cells and enabling simultaneous targeting of tumor-associated macrophages (TAMs). FA-labeled trypsinized EVs (tEVs) loaded with miR-138 inhibit tumor growth, depolarize TAMs, and enhance antitumor immunity. This study represents the first preclinical attempt to modulate tumor cells and innate immunity via miRNA-loaded tEVs, offering a novel and more effective therapeutic approach to GBM treatment.

Keywords: Extracellular vesicle (EV); Glioblastoma (GBM); Trypsin; Tumor microenvironment (TME); Tumor-associated macrophages (TAMs); microRNA-138 (miR-138).

Figure 1.. Surface trypsinization of extracellular vesicles (EVs) for enhanced ligand labeling.

(A) Engineering scheme of the EV surface by trypsin digestion to generate trypsinized EV (tEV) from untrypsinized EV (uEV). (B) Size determination of uEV and tEV by dynamic light scattering (DLS). (C) Representative transmission electron microscopy (TEM) images of uEV and tEV at magnification of 50 000. Scale bar represents 50 nm. (D) Comparison of total protein contents of uEV or tEV relative to those in input EV (n = 3). (E) Schematic design of surface labeling of tEV with 27-mer DNA linkers (D27) harboring Alexa-647 fluorophore (AF647) on its 5′-end and cholesterol-TEG on 3′-end (AF647-D27-Chol) yielding AF647-D27-tEV or AF647-D27-uEV as a negative control. Unbound free AF647-D27-Chol remained in the supernatant after the centrifugation of precipitated EVs. (F) Fluorescence imaging and relative fluorescence intensity of the EV pellets and supernatant compared to the initial input of AF647-D27-Chol linkers for the comparison of surface labeling efficiency with AF647-D27-Chol ligands between uEV and tEV as described in (E) (n = 3). (G) Schematic design of two-step engineering of the tEV for cargo loading and ligand labeling. (H) Fluorescence imaging and relative fluorescence intensity of the EV pellets and supernatants during two-step engineering compared to the initial input of D27-AF647 cargo for the comparison of cargo loading and surface labeling efficiencies between uEV and tEV as described in (G) (n = 3). All error bars indicate standard deviation (s.d.). ***p < 0.001, ****p < 0.0001.

Figure 2.. Patient-derived human primary GBM cell targeting by FA-D27-tEV in vitro.

(A) Schematic design of a folate (FA)-conjugated DNA linker for tEV surface labeling. The 5′ and 3′ ends of random 27-base scrambled DNA linker sequences (D27) were labeled with folate (FA) and cholesterol-TEG linker, respectively, to generate FA-D27-Chol. (B) Schematic design of FA-free negative control linker (D27-chol) labeled with only the cholesterol-TEG linker. (C) Schematic preparation design of fluorescently labeled FA-D27-tEVs for cell targeting. Fluorescent cargo (D27-AF647) was encapsulated into tEV before labeling with FA-D27-Chol (FA-D27-tEV-AF647) or FA-free negative control D27-Chol (D27-tEV-AF647) on their surface. (D) Flow cytometry to measure differential targeting of human GBM28 cells by FA-D27-tEV-AF647 in vitro compared to that of D27-tEV-AF647. No EV-treated cells were used as gating control (n = 3). (E) Representative fluorescence confocal microscopy images from human GBM cells (GBM6, GBM12, GBM28, and GBM59) treated for 16 h with FA-D27-tEV-AF647 or D27-tEV-AF647. Pseudocolor (red) was used to visualize tEV-AF647, while the nucleus and actin filaments were counterstained with DAPI (blue) or Phalloidin-Alexa488 (green). Scale bar indicates 50 μm. (F) 3D-scanned fluorescence confocal microscope images from human GBM28 visualizing the distribution of FA-D27-tEV-AF647 throughout the cytoplasmic area while D27-tEV-AF647 was found on the cell surface (arrows). Scale bar indicates 50 μm.

Figure 3.. Co-targeting of tumor cells and tumor-associated macrophages in orthotopic GBM tumor by FA-D27-tEV in vivo.

(A) Schematic design of in vivo intratumoral targeting experiment in mice brain bearing an orthotopic syngeneic GBM tumor. (B) Representative ex vivo fluorescence IVIS images visualizing intracranial GBM tumor targeting and relative fluorescence intensity by FA-D27-tEV-AF647 compared to negative controls, such as FA-D27-uEV-AF647 or D27-tEV-AF647 or PBS (n = 3). (C) Representative ex vivo image for the biodistribution profile of intratumorally injected FA-D27-tEV-AF647 or negative controls, such as FA-D27-uEV-AF647 or D27-tEV-AF647 or PBS. (D) Representative confocal fluorescence microscopy images from the harvested mouse brain tumor tissues implanted with 005-GFP syngeneic GBM tumors and treated with FA-D27-tEV-AF647 or D27-tEV-AF647 intratumorally. The 005-GFP brain tumor cells (green) were counterstained with DAPI for nucleus (blue) against FA-D27-tEV-AF647 or D27-tEV-AF647 (red). Scale bar 50 μm. (E) Representative contour plots from flow cytometry analysis of the tumor cells dissociated from the harvested 005-GFP GBM tumor tissues. The percentiles of AF647+ GFP+ double positive cells were plotted for FA-D27-tEV-AF647 (n = 3) in comparison to D27-tEV-AF647 (n = 3) using PBS-treated tumor cells (n = 3) as gate controls against AF647+ cells. (F) Representative t-SNE maps from analytical multicolor flow cytometry analysis on the CD45+ immune cell populations that were dissociated from the 005-GFP tumor to dissect the profile of cell type specificity of FA-tEV-D27-AF647. Pseudo colors were used to indicate each population of immune cells in the t-SNE map, such as CD45+/F4/80+ macrophages in red. Representative t-SNE maps from each tested group showed an association of FA-tEV-D27-AF647 with macrophage populations compared to those of D27-tEV-AF647 or PBS (blue-red scale). (G) Representative contour plots from flow cytometry analysis on the CD45+ F4/80+ macrophages in immune cells enriched with GBM tumor tissues. The percentile of F4/80+ AF647+ double positive macrophages was plotted for FA-D27-tEV-AF647 (n = 3) in comparison to D27-tEV-AF647 (n = 3) or PBS (n = 3). All error bars indicate standard deviations (s.d.). *p < 0.05, **p < 0.01, ***p < 0.001, and n.s. indicates not significant.

Figure 4.. Targeted delivery of miR-138 into GBM tumor cells by FA-D27-tEV.

(A) Step-by-step schematic illustration of the tEV engineering to generate FA-tEV-miR-138. (B) In vitro delivery efficiency of miR-138 into human GBM cells by FA-D27-tEV-miR-138 as measured by TaqMan miRNA expression assay to detect the expression level of miR-138 using U6 snRNA as normalizing control (n = 4). (C) MTT-based cell viability assay on human GBM cells after incubation with FA-D27-tEV-miR-138 or FA-D27-tEV-miR-Ctrl for 96 h (n = 9). (D) Western blotting assay on the known targets of miR-138, such as CD44, Survivin, or 4E-BP1, in GBM28 cells treated with FA-D27-tEV-miR-138 or FA-D27-tEV-miR-Ctrl for 72 h. GAPDH was used as a loading control. (E) Representative fluorescence confocal microscopy images of GBM28 cells treated with FA-D27-tEV-miR-138 or FA-D27-tEV-miR-Ctrl for 72 h. The known targets of miR-138, such as CD44 (green) or Survivin (red), were counterstained with DAPI (blue). Scale bar: 100 μm. (F) In vivo antitumor efficacy of FA-D27-tEV-miR-138 against human orthotopic GBM tumor in nude mice. Each group of mice (n = 10) was intratumorally injected with FA-D27-tEV-miR-138 or FA-D27-tEV-miR-Ctrl for a total of five times with 3 days interval. The percentile of mice survival rates was calculated using Kaplan−Meier curve fitting. *p < 0.05, **p < 0.01, and ***p < 0.001.

Figure 5.. Enhanced anti-GBM activity of FA-D27-tEV-miR-138 through the reprogramming of TAMs.

(A) Expression level of miR-138 in intratumoral F4/80+ macrophages isolated from intracranial 005 syngeneic GBM tumor implanted into immunocompetent C57BL/6 mice brain. The expression level of miR-138 in M1 (CD86+) or M2 (CD206+)-like macrophages was measured by TaqMan miRNA expression assay using U6 snoRNA as normalizing control (n = 16). (B) Relative miR-138 expression levels in M1 or M2-like F4/80+ macrophages polarized from human PBMC (n = 6). (C) Representative contour plots from flow cytometry analysis revealing the repolarization of M2-like macrophages derived from PBMC into CD68+CD86+ M1 phenotypes after the transfection with miR-138 or miR-Ctrl (n = 5). (D) Relative cell viability of human GBM cells cocultured with M2-polarized PBMC-derived F4/80+ macrophages overexpressing miR-138 or miR-Ctrl as measured by CellTiter-Glo luminescent cell viability assay and normalized to that of GBM/M2 coculture control (n = 4). (E) Schematic survival test design in intracranial 005 syngeneic GBM tumor-bearing C57BL/6 mice after intratumoral injection of FA-D27-tEV-miR-138 or FA-D27-tEV-miR-Ctrl five times (20 mice/group). Kaplan−Meier survival curve reveals anti-GBM tumor efficacy of FA-D27-tEV-miR-138 in the immunocompetent mice GBM models. (F) Representative confocal fluorescence microscopy images on cryosectioned mouse GBM tumor tissues after the fifth treatment. GFP+ 005 tumor cells (green) were counterstained against DAPI (blue) and CD206 (red). Scale bar indicates 100 μm. (G) Detection of M2-like TAM population changes by flow cytometry on the immune cell population dissociated from 005-GFP GBM tumor tissues. Double positive cell populations for CD206 and PD-L1 in total F4/80+ macrophages were identified as intratumoral M2-like TAMs (n = 3). (H) Intratumoral level of inflammatory cytokine, IL1β, in the 005-GFP tumor tissues measured by ELISA (n = 4). (I) Detection of CD8+ T cell population changes by flow cytometry on the immune cell population dissociated from 005-GFP GBM tumor tissues. Representative contour plots compare the CD3+ CD8+ double positive T cell population between FA-D27-tEV-miR-138 and FA-D27-tEV-miR-Ctrl. The relative population of CD3+ CD8+ T cells among the CD45+ immune cell population was shown in percentile (n = 4).