Molecular Profiling and Functional Analysis of Macrophage-Derived Tumor Extracellular Vesicles

Abstract

Extracellular vesicles (EVs), including exosomes, modulate multiple aspects of cancer biology. Tumor-associated macrophages (TAMs) secrete EVs, but their molecular features and functions are poorly characterized. Here, we report methodology for the enrichment, quantification, and proteomic and lipidomic analysis of EVs released from mouse TAMs (TAM-EVs). Compared to source TAMs, TAM-EVs present molecular profiles associated with a Th1/M1 polarization signature, enhanced inflammation and immune response, and a more favorable patient prognosis. Accordingly, enriched TAM-EV preparations promote T cell proliferation and activation ex vivo. TAM-EVs also contain bioactive lipids and biosynthetic enzymes, which may alter pro-inflammatory signaling in the cancer cells. Thus, whereas TAMs are largely immunosuppressive, their EVs may have the potential to stimulate, rather than limit, anti-tumor immunity.

Keywords: T cell response; exosome; extracellular vesicle; inflammation; lipid metabolism; lipidomics; proteomics; tumor microenvironment; tumor-associated macrophage.

Graphical abstract

Figure 1

EV Isolation from MC38 Tumors of IgG- and Anti-CSF1R-Treated Mice (A) Procedure to isolate EVs from IgG- and anti-CSF1R-treated tumors. (B) Flow cytometry of MC38-tumor-derived cells (day 14 post-tumor challenge; see Figure S1A). Data show percentage values (mean ± SEM; n = 5 mice/condition). Statistics by unpaired two-tailed Student’s t test. (C) Yield of EVs prior to sucrose fractionation, determined by BCA (mean ± SEM; n = 5 independent EV preparations). Statistics as in (B). (D) Representative TEM images of EVs obtained as in (C). One representative EV preparation is shown for IgG and anti-CSF1R-treated tumors. Scale bars, 200 nm. (E) EV concentration and size distribution by NTA (mean ± SEM; n = 3 acquisitions/sample). One representative EV preparation per condition is shown. (F) Correlation between EV protein content and EV concentration, determined by BCA and NTA, respectively (mean ± SD; n = 3 serial dilutions/sample). A simple linear regression function was used. One representative EV preparation per condition is shown. (G) WB analysis of cells and matched EVs from cultured MC38 cells or MC38 tumors. One representative cell or EV preparation per condition is shown. (H) EV protein content and EV concentration in each sucrose fraction, determined by BCA and NTA, respectively (mean of 2–3 technical replicates). One representative EV preparation per condition is shown. (I) EV concentration and size distribution by NTA (mean ± SEM; n = 3 acquisitions/sample). One representative EV preparation per condition is shown. (J) Yield of EVs recovered from the third top fraction of the sucrose gradient, determined by BCA (mean ± SEM; n = 5 independent EV preparations). Statistics as in (B). (K) Representative TEM images of EVs recovered from the third top sucrose fraction. One representative EV preparation per condition is shown. Scale bars, 200 nm. (L) WB analysis of EVs after sucrose gradient fractionation. Upper panel shows a representative experiment; equal sample volumes were loaded in each lane. Lower panels show relative band intensities of MRC1 and GAPDH (mean ± SEM; n = 3 independent EV preparations, one of which is shown in the WB above). For each protein, the relative signal intensity in each fraction is indicated as percentage of the total signal from all fractions. (M) Taqman analysis of selected microRNAs (normalized to miR-16-5p; fold-change versus anti-CSF1R-EVs) in EVs after sucrose gradient fractionation (mean ± SEM; n = 3 independent EV preparations). Statistics by two-way ANOVA with Sidak’s multiple comparison test. Statistical significance of the data: ^∗p < 0.05; ^∗∗p < 0.01; ^∗∗∗p < 0.001; ^∗∗∗∗p < 0.0001. See also Figures S1 and S2.

Figure 2

Enrichment and Interactome Analysis of EV Proteins (A) LC-MS/MS-based GO enrichment analysis, prior to and after sucrose gradient fractionation, of MC38 EV-associated proteins, whose fold-change (FC) abundance was > 1.7 in IgG-EVs versus anti-CSF1R-EVs. Data were generated from 2 independent EV preparations. (B and C) LC-MS/MS-based immune cell enrichment (B) and interactome (C) analysis of MC38 EV-associated proteins (IgG versus anti-CSF1R, FC > 1.7) obtained after sucrose gradient fractionation. Data were generated from 4 independent EV preparations. Three proteins (LRRC25, CYP4F3, and CD300LD4) were not detected in the Immuno-Navigator database.

Figure 3

Proteomic Analysis of IgG-EVs and Anti-CSF1R-EVs and Comparison with BMM-EVs and MC38-EVs (A–G) LC-MS/MS and classification of TAM-EV proteins (criteria: IgG versus anti-CSF1R FC > 1.7) according to their associated pathways. Data show normalized quantitative values (mean ± SEM; n = 4 independent EV preparations). Statistics by two-way ANOVA with Holm-Sidak’s multiple comparison test. (H) WB analysis of cells and EVs from tumors, MC38 cells, and M2 macrophages (BMM). One representative EV preparation per condition is shown. (I) Schematic of EV isolation (left) and Venn diagram (right) comparing proteins detected by LC-MS/MS in MC38-EVs (3 independent EV preparations) and M1 and M2 BMM-EVs (2 independent EV preparations). (J) Classical macrophage markers in M1 and M2 BMM-EVs. Data show mean quantitative values normalized to ALIX (2 independent EV preparations). (K and L) Venn diagrams comparing proteins enriched in IgG-EVs versus anti-CSF1R-EVs (FC > 1.7; K) or anti-CSF1R-EVs versus IgG-EVs (FC > 1.7; L) with those detected in MC38-EVs and M1 and M2 BMM-EVs. (M) MC38-tumor-derived EV proteins of likely neutrophil origin (anti-CSF1R/IgG FC > 1.7). Data show normalized quantitative values (mean ± SEM; n = 4 independent EV preparations). Statistics as in (A).

Figure 4

Quantification of TAM-EVs and Validation of the TAM-EV Protein Signature by IP of CD11b⁺ Tumor-Derived EVs (A) Procedure to isolate EVs from MC38 tumors grown in LysM.Cre/ROSA^mT/mG mice. (B) WB analysis of EVs isolated from cultured MC38 cells or MC38 tumors grown in LysM.Cre/ROSA^mT/mG mice, after IP with anti-CD9 or control-coated magnetic beads. One representative EV preparation per condition is shown. (C and D) Flow cytometry analysis of tumor-EVs isolated from wild-type (WT) or LysM.Cre/ROSA^mT/mG mice. Data show representative dot plots of GFP and tdTomato (C) and quantitative values (mean ± SD; n = 3 independent experiments; D). Statistics by two-way ANOVA, using Sidak’s multiple comparison test. (E) LC-MS/MS analysis of EVs showing quantitative values (mean ± SEM; n = 4 or 3 independent preparations) for CD11b, ALIX, and CD81, normalized to CD9. Statistics by two-way ANOVA, using Tukey’s multiple comparison test. (F) IP procedure for capturing CD11b⁺ EVs. (G) Flow cytometry analysis of PKH67-labeled EVs bound to anti-CD9, anti-CD11b or isotype-control beads. Data show percentage values (mean ± SEM; n = 4 and 5 preparations for CD11b and CD9 IP, respectively). (H) WB analysis of tumor-derived EVs after IP. One representative EV preparation per condition is shown. (I) Venn diagrams comparing proteins enriched in IgG-EVs versus anti-CSF1R-EVs (FC > 1.7; see Table S2) with those detected in CD11b⁺ IgG-EVs after IP (FC versus isotype control beads > 2; see Table S4). IP was performed on the same IgG-EV preparation shown in (H). (J) Radial table reporting the 62 proteins of the MC38 TAM-EV signature and their association with biological pathways. The graphical representation limits connection of each protein to two pathways (see also Table S5). (K) The data in (I) are shown after removing potential protein contaminants identified according to the CRAPome database. See also Figures S3, S4, S5, and S6.

Figure 5

Immunomodulatory Profile and Functions of TAM-EVs (A and B) Gene set enrichment analysis (GSEA) plots showing the correlation of TAM (TAM-Cell) and TAM-EV protein signatures with genes expressed in FACS-sorted M1-like or M2-like TAMs (A) or prognostic genes across human cancers recorded in the PRECOG database (B). (C) Flow cytometry analysis of CD8⁺ T cells obtained after priming OT-I splenocytes. (D) Schematic of the experiment shown in (E–G). (E and F) Flow cytometry analysis of CD8⁺ OT-I proliferation assessed by CellTrace dilution. (E) Representative flow profiles. (F) Quantification of the data (mean ± SEM; n = 4 cell cultures/condition). Statistics by two-way ANOVA, using Sidak’s multiple comparison test. (G) ELISA-based quantification of IFNγ in medium conditioned by OT-I CD8⁺ T cells. Data are shown as mean ± SEM (n = 4 cell cultures/condition). Statistics as in (F). (H) Flow cytometry analysis of CD8⁺ T cells purified from the spleen of C57BL/6 mice. (I) Schematic of the experiment shown in (J and K). (J) Flow cytometry analysis of CD8⁺ T cell proliferation assessed by CellTrace dilution (mean ± SEM; n = 4 cell cultures/condition). Statistics by one-way ANOVA, using Tukey’s multiple comparison test. (K) ELISA-based quantification of IFNγ in medium conditioned by CD8⁺ T cells (mean ± SEM; n = 4 cell cultures/condition). Statistics as in (J). (L) Flow cytometry analysis of CD4⁺ T cells purified from the spleen of BALB/c mice. (M) Schematic of the experiment shown in (N). (N) Flow cytometry analysis of CD4⁺ T cell proliferation assessed by CellTrace dilution (mean ± SEM; n = 4 cell cultures/condition). Statistics as in (J). (O) Schematic of the experiment shown in (P). (P) Flow cytometry analysis of activation markers (mean fluorescence intensity, MFI) in CD11b⁺CD11c⁺ BMDCs (mean ± SEM; n = 4 cell cultures/condition). Statistics by two-way ANOVA, using Dunnett’s multiple comparison test (each treatment condition versus control DMSO).

Figure 6

Molecular and Functional Lipidomic Profile of TAM-EVs (A) Schematic illustrating AA metabolism. (B) LC-MS/MS proteome analysis of EVs showing enzymes involved in eicosanoid synthesis. Data show quantitative values (mean ± SEM; n = 4 and 3 independent preparations of MC38-tumor-derived EVs and MC38-EVs, respectively). Statistics by two-way ANOVA, using Tukey’s multiple comparison test. (C) WB analysis of cells and EVs from cultured MC38 cells or MC38 tumors. One representative EV preparation per condition is shown. (D) AA concentration (mean ± SEM; n = 4 independent EV preparations) determined by LC-MS/MS-based absolute quantification using calibration curves of internal standards. Statistics by unpaired two-tailed Student’s t test. (E) TXB₂ concentration (mean ± SEM; n = 4 independent EV preparations) determined by LC-MS/MS-based relative quantification of ion counts. Statistics as in (D). (F) Schematic illustrating putative thromboxane synthesis in TAM-EVs. (G) Quantification of eicosanoids, mostly PGs (left) and their precursor AA (right), in MC38 tumors (mean ± SEM; n = 6 mice/condition) by LC-MS/MS lipidomics. Statistics by two-way ANOVA, using Sidak’s multiple comparison test (left) or unpaired two-tailed Student’s t test (right). (H) CD8⁺ T cell proliferation in response to PGE₂ or PGF_2α; the left panel shows the experimental design. The right panel shows flow cytometry analysis of CD8⁺ T cell proliferation assessed by CellTrace dilution (mean ± SD; n = 3 cell cultures/condition). Statistics by one-way ANOVA, using Dunnett’s multiple comparison test (each PG concentration versus control DMSO). (I) Confocal analysis of MC38 cells incubated with PKH67-labeled MC38 IgG-EVs (green). Nuclei are stained with 4′,6-diamidino-2-phenylindole (DAPI; blue). The merged panel is shown on the left. One representative experiment is shown. Scale bar, 10 μm. (J) WB analysis of MC38 cells incubated with EVs isolated from LysM.Cre/ROSA^mT/mG mice. One representative EV preparation per condition is shown. (K) Schematic of the experiment shown in (L–N). (L and M) Quantification of PUFAs (L) and eicosanoids (M) in medium conditioned by MC38 cells (mean ± SEM; n = 4 cell cultures/condition) by LC-MS/MS. Statistics as in (B). (N) ELISA-based quantification of PGE₂ and TXB₂ (mean ± SEM; n = 4 cell cultures/condition) in medium conditioned by MC38 cells. Statistics by one-way ANOVA, using Tukey’s multiple comparison test. See also Figure S7.

Figure 7

Molecular and Functional Analysis of E0771 TAM-EVs (A) Schedule of subcutaneous E0771 cancer cell inoculation in C57BL/6 mice and drug administration. (B) Flow cytometry analysis of immune infiltrates in E0771 tumors. Data show percentage values (mean ± SD; n = 3 mice/condition). Statistics by unpaired two-tailed Student’s t test (left and middle) or two-way ANOVA, using Sidak’s multiple comparison test (right). (C) Yield of EVs recovered from E0771 tumors prior to and after sucrose gradient fractionation, determined by BCA (mean ± SD; n = 3 EV preparations/condition). Statistics by unpaired two-tailed Student’s t test. (D) EV concentration and size distribution by NTA in the 6 sucrose fractions (mean ± SEM; n = 3 EV preparations/condition). (E) EV protein content and EV concentration in each sucrose fraction determined by BCA and NTA, respectively (mean ± SD; n = 3 EV preparations/condition). (F) Representative TEM images of EVs. One representative EV preparation per condition is shown. Scale bars, 200 nm. (G) WB analysis of cultured E0771 cells and matched EVs. One representative EV preparation is shown. (H and I) WB analysis of E0771 tumor-EVs (H). Relative signal quantification of MRC1, COX1, and TBXAS1 is shown in (I) as mean band intensity normalized to CD9 (n = 3 EV preparations/condition). Statistics as in (C). (J) Venn diagrams comparing proteins enriched in IgG-EVs versus anti-CSF1R-EVs from E0771 tumors (FC > 1.7; see Table S7) with those of the MC38 TAM-EV signature (see Table S5). (K) LC-MS/MS analysis of TBXAS1 in E0771 tumor-EVs (mean ± SD; n = 3 EV preparations/condition). Statistics as in (C). (L) ELISA-based quantification of TXB₂ in medium conditioned by E0771 cells (mean ± SD; n = 3 EV preparations/condition). Statistics by one-way ANOVA, using Tukey’s multiple comparison test.