Article Figures & Data
Figures
- Figure S1. DC-EV are specifically taken up by monocytes.
PBMCs were incubated with GFP-labeled DC-EV for 3 h. Uptake of EV was determined by flow cytometry gating on lymphocytes and monocytes.
- Figure 1. DC-derived EV differentiate monocytes.
(A) Monocytes ingest EV. Monocytes were incubated with DC-derived and GFP-labeled EV for 3 h, washed, and subsequently analyzed by confocal microscopy using Z-stack imaging. DIC: differential interference contrast. (B) maDC-EV differentiate a DC-like morphology in monocytes. Monocytes were incubated with the dose of EV (30 μg for 106 cells) derived from imDC and maDC or stimulated with GM-CSF/IL-4 for 6 d. Subsequently, images were taken from representative cells. (C) DC-EV induce proliferation in monocytes. PBMCs were labeled with CFSE and treated either with a single dose of imDC or maDC-derived EV (50 μg) or cytokines or LPS as indicated. CFSE dilution in CD11b+ cells was determined at day 1 and day 10 by flow cytometry, and the number of cell divisions is indicated. The graph summarizes the results from four different donors; one representative result is shown on the left side of the panel. Results are presented as mean ± SEM; statistical significance was analyzed by one-way ANOVA: *P < 0.05, **P < 0.01, and ***P < 0.005. (D) maDC-EV–treated monocytes maintain a DC-like morphology upon exposure to maturation stimuli. Same experimental setup as in (B). Subsequently, cells were incubated for 24 h with a MC (IL-1β, IL-6, TNF-α, and PGE2) or LPS and images were taken from representative cells. (E) maDC-EV–treated monocytes that received a maturation stimulus induce T-cell proliferation. Monocytes incubated with imDC and maDC-derived EV (10 μg), or stimulated with GM-CSF/IL-4 (serving as positive control) for 6 d, either received a maturation stimulus (MC) or were left untreated. Subsequently, CFSE-labeled T cells were co-incubated at a defined ratio as indicated and proliferation of cells was determined by radiolabeled thymidine incorporation. Shown is one representative experiment of five performed with different donors (see Fig S3A). Scale bars represent 7.5 μm.
- Figure S2. Concentrations of CCF in DC-derived EV and DC culture supernatants.
(A) CCF concentrations in 10-μg EV preparations from different DC donors. maDC were generated as described in the Materials and Methods section from five different donors and analyzed for the indicated factors by multiplex technology (BioLegend). (B) Measurement of single factors in the EV pellet and the respective culture supernatants to demonstrate the relative amount secreted through EV and directly into the supernatant. Note: measurements of factors used for the generation of DC generated aberrant results in the culture supernatant and were not reported.
- Figure S3. Analysis of EV-treated monocytes by mixed lymphocyte reaction and FACS analysis.
(A) Same experimental procedure as described in Fig 1E using primary cells from five different donors. (B) Representative FACS plots/graphs from one donor for the FACS analysis described in Fig 2A.
- Figure 2. maDC-EV–treated monocytes develop DC-typical marker expression and factor secretion.
(A) DC-derived EV induce DC-typical marker proteins on monocytes. Peripheral monocytes (2 × 105) were incubated with EV (10 μg) derived from imDC and maDC, or stimulated with GM-CSF/IL-4 for 6 d. Subsequently, cells were analyzed by FACS for the indicated markers. Horizontal black bars represent mean values of all analyzed individual donors. Statistical significance was determined by one-way ANOVA: *P < 0.05, **P < 0.01, and ***P < 0.005. (B) maDC-EV–treated monocytes receiving a DC maturation stimulus secrete IL-12p70. Same experimental setup as in (A). The resulting cells were treated with R848 for 6 h or left untreated. Subsequently, the indicated cytokines were measured in the cell culture supernatant. (C) Monocytes receiving DC-EV and a DC maturation stimulus express surface markers typical for maDC. Same experimental setup as in (B), and subsequent analysis of indicated surface markers by FACS. In all plots of the figure, each symbol represents one individual donor. The experiments in (A–C) were repeated with different donors, indicated by individual data points.
- Figure 3. DC-EV induce GM-CSF signaling and convey a cornucopia of effector molecules.
(A) DC-EV induce Stat5 phosphorylation. Peripheral monocytes (2 × 105) were incubated with EV (10 μg) derived from imDC and maDC and 293T cells (control EV) or stimulated with GM-CSF/IL-4 (each for 15 min) or left untreated. Subsequently, cells were fixed and analyzed for Stat5 phosphorylation by intracellular FACS. FACS blots depict one representative experiment. Three healthy donors were analyzed to calculate the mean and SEM. (B) Anti-GM-CSF blocks DC-EV–induced Stat5 phosphorylation. Same experimental setup as in (A); however, one cell aliquot of each culture was left untreated or was supplemented with anti-GM-CSF. Triplicate cultures were performed for each donor (three donors) to calculate the mean and SEM. (C) DC-EV–derived GM-CSF is derived from the producer DC. Lysates of purified DC-EV and control EV (from 293T cells) was blotted for endogenous (endg.) GM-CSF using lysates of maDC (maDC lys.) and recombinant (recomb.) GM-CSF as control. (D, E) DC-derived EV contain multiple CCF. (D) EV were collected from monocytes and monocyte-derived imDC and maDC (50 μg) and subsequently analyzed for the indicated factors using commercially available protein arrays (RayBiotech). The pixel intensity of each dot was determined by ImageJ, and the value was adjusted in relation to the internal positive control, which was set to 1. Shown is one representative analysis performed with four different donors (see also Fig S4C). (E) Same experimental setup as in (D); however, the EV contents were analyzed using bead-based quantitative immunoassays (BioLegend). imDC-EV were analyzed from six and eight different donors for cytokine and chemokine content, respectively (gray columns). maDC-EV from six different donors were analyzed (black columns). Heat maps depict the common logarithm (log(10)) of the cytokine and chemokine concentrations of each individual sample. Bar graphs indicate mean values ± SEM. Statistical significance was analyzed by the t test: *P < 0.05, **P < 0.01, and ***P < 0.005.
Source data are available for this figure.
Source Data for Figure 3[LSA-2018-00093_SdataF1.pdf]
- Figure S4. DC-EV–induced monocyte differentiation depends on GM-CSF signaling, and CCF analysis in EV by a protein array.
(A) Monocyte differentiation through maDC-EV is inhibited by anti-GM-CSF. Peripheral monocytes were incubated with 10-μg imDC-EV or maDC-EV or were stimulated with GM-CSF/IL-4 for 5 d, and either supplemented with 10-μg anti-GM-CSF Ab or matched anti-IgG. Subsequently, cells were stained and analyzed for CD11b. Percentage of living FSC-high/CD11b+ cells was determined by flow cytometry. (B) FACS analysis of maDC-derived EV coupled to latex beads. maDC-derived EV were coupled to latex beads as described in the Materials and Methods section. Subsequently, the beads were analyzed for surface markers by FACS as indicated. (C) DC-derived EV contain multiple CCF. EV were collected from monocytes and monocyte-derived imDC and maDC (50 μg) and subsequently analyzed for the indicated factors using commercially available protein arrays (RayBiotech). This is the same analysis as described in Fig 3D using a second donor.
Source data are available for this figure.
Source Data for Figure S4[LSA-2018-00093_SdataF1.pdf]
- Figure S5. CCF–containing DC-EV float at densities characteristic for exosomes.
DC-EV contain CCF at floating densities characteristic for exosomes. EV from culture supernatants of maDC were obtained first by ultracentrifugation and subsequently fractionated by an iodixanol (OptiPrep) density gradient. Densities (g/ml) of gradient fractions (Nos. 1–11) were determined by a refractometer. Cytokine concentrations in each fraction were determined by an antibody bead array. For additional evidence, electron micrographs were taken from fractions 1 and 8. Scale bars represent 250 nm.
- Figure S6. EV derived from mature BMDC are efficiently taken up by murine monocytes and have a rich content of CCF.
(A) Murine PBMCs were incubated with stained (PKH26) immature (im) or mature (ma) BMDC-EV for 2 h. Percent of PKH26+ cells was determined by flow cytometry gating on monocytes (Ly6C-high, CD11b+, and Ly6G−), B cells (B220+, CD11b−, Ly6C−, and Ly6G−), and granulocytes (SSC-high, Ly6G+, CD11b+, Ly6C-int., and B220−). (B) EV derived from matured BMDC contain a rich CCF. EV derived from mature BMDC, which had been stimulated with Poly IC/R848 (EV-BMDC + Poly I:C) or LPS (BMDC-EV + LPS) or unstimulated (BMDC-EV), were analyzed by an antibody-based bead array (BioLegend). Values were normalized to number of EV-producing cells. Five different donors were analyzed. The resulting numbers were used to calculate mean values ± SEM. (C) Lysates from EV as described in (B), but derived from cells of a different donor animal, were analyzed by a protein array as described in Fig 3. The pixel intensity of each dot was determined by ImageJ, and the value was adjusted in relation to the internal positive control, which was set to 1. Shown is one representative analysis performed with four different donors.
- Figure S7. EV from maBMDC induce a DC-like morphology, whereas EV from imBMDC induce a macrophage-like phenotype in monocytes.
BM-derived monocytes were incubated with medium (no stim.), EV derived from imBMDC (imBMDC-EV) or maBMDC (maBMDC-EV), or GM-CSF for 5 d. Subsequently, images were taken from representative cells. The mean fluorescence intensity (MFI) of surface marker expression was determined by flow cytometry.
- Figure 4. Skin-injected BMDC-EV attract immune cells.
(A) Cartoon depicting the injection site of PKH26-labeled BMDC-EV, and images from excised skin patches and draining lymph nodes used for MELC analysis and marker quantification. For control, PKH26-containing medium and PBS were injected. (B) BMDC-derived EV attract immune cells in the skin. Tissue sections from skin patches described in (A) were subjected to a MELC analysis. Images represent an overlay of four markers (CD45, cytokeratin-14, PI, and EV). Cytokeratin-14 and CD45 were stained by antibodies, whereas EV (red stain) were visualized through PKH26. Tissue sections from two animals (6 h and 24 h) are presented. Using the StrataQuest software, the relative presence (in percentage) of common immune cell markers was quantified in EV-containing tissue areas. The relative presence of cells (percentage of cells with marker in EV areas) is depicted by a bar diagram as indicated. Note: individual images for these markers are presented in Fig 5. Quantifications of MELC analyses from four different injection sites were used to determine the SEM. Scale bars represent 100 μm. (C) Monocytes and neutrophils in imDC-EV and maDC-EV areas. Monocytes (CD11b+/Ly6C+/Ly6G−) and neutrophils (CD45+/Ly6C+/Ly6G+) were quantified in the EV areas using the StrataQuest software as explained in Table 1 and Fig S8A. epi, epidermis; de, dermis; sc, subcutaneous.
- Figure S8. Marker identification and quantification by StrataQuest, and confocal analysis of EV-containing tissue sections.
(A) StrataQuest analysis algorithm. First, nuclei are identified and quantified (gated) through propidium iodide (PI) assessment (gated nuclei: 86.53%; first two images and gate 1). Around the nuclei, the software calculates and demarcates a small area that represents the cell cytoplasm and boundary (green net-like structure in image 2). Subsequently, a marker image is superimposed, demarcating the area of interest (here area of EV deposition in a lymph node, images 3 and 4, gate 2), which is set to 100% in order to calculate cell sub-populations in this area. Subsequently, an additional marker is superimposed and gated on the previous gate (43.22% of cells in the EV area, gate 3). This calculation can be performed for different individual markers (e.g., Fig 4), or continued for assessing multiple marker co-expression in one cell (e.g., Table 1). Scale bars represent 100 μm. (B) Analysis of cellular EV uptake in tissue. A tissue section as described in Fig 4B (maDC-EV 6 h after injection) was analyzed by confocal microscopy in order to demonstrate that EV were ingested by cells and not merely deposited by injection. To indicate the cytoplasm, the cells were stained for CD11b. EV are indicated by red color and nuclei by DAPI. Scale bars represent 25 μm.
- Figure S9. Representative examples of EV injection sites.
Shown are examples of im- and maBMDC-EV injection sites after 24 h in two different animals (four skin injections per animal). The EV areas are demarcated with white interrupted lines. For each time point and each EV type (immature/mature), one animal with four injection sites was analyzed. Scale bars represent 100 μm.
- Figure 5. Immune cells attracted by maBMDC-EV express myeloid activation markers (see also Fig S7).
(A) Individual images of a MELC analysis, assessing immune cell markers in BMDC-EV–containing skin tissue areas. The same skin tissue sections shown in Fig 4B from imBMDC-EV–, maBMD-EV–, and medium-injected areas, obtained after 6 h and 24 h, were analyzed for the indicated markers (green) and EV colocalization (yellow; see also Fig S7). Note: for better orientation, the images on the left were duplicated from Fig 4B. (B) Co-expression of myeloid activation and differentiation markers with monocytes (CD11b+/Ly6C+/Ly6G−). Using the StrataQuest software, monocytes were identified (see Fig 4C) and analyzed for co-expression of the indicated markers in the EV-containing tissue area, expressed in percentage of total monocytes found in the EV area. The analysis was performed for each time point in tissue sections from four different injection sites. The obtained numbers were used to determine the SEM. Scale bars represent 100 μm.
- Figure 6. BMDC-EV induce phenotypically different myeloid cell sub-populations.
(A) Protein expression profiles of individual immune cells in imBMDC-EV- and maBMDC-EV–containing tissue areas 6 h and 24 h after injection. The upper panels (also analyzed in Figs 4 and 5) show the whole tissue areas, whereas the lower panels depict the MELC protein profile of individual cells from these areas. Individual cells were chosen randomly and assigned an ID number (cell ID). Cells were selected in areas where EV were concentrated (EVca; yellow demarcated area, upper panels) or less concentrated (EVpa; blue demarcated areas). The latter was determined by the immunoreactivity score (IRS) (Remmele & Stegner, 1987). Colored boxes were inserted for explanations in the main text. Scale bar represents 50 μm (upper panels) and 7.5 μm (lower panels). (B) Expression levels of indicated markers for each numbered cell (cell ID in [A]). For better understanding, protein expression was divided into three levels (no, low, and high expression), which were color-coded. (C, D) Relative presence of cells with triple marker combinations in imBMDC-EV- and maBMDC-EV–containing tissue areas shown in (A). Triple combinations of the most abundant markers (CD45, Ly6C, Ly6G, and F4-80) and CD169 (serving as internal control) were assessed by StrataQuest software as explained in Fig S8 and Table 1 and displayed in bar diagrams.
- Figure 7. BMDC-EV attract immune cells in lymph nodes similarly as in the skin.
(A) Individual images of a MELC analysis, assessing EV and immune cell markers in lymph nodes. Tissue sections from draining lymph nodes after skin injections of imBMD-EV and maBMD-EV (red label) obtained after 24 h were analyzed by MELC for the indicated markers and for colocalization with EV. Numbers in the left-most panels depict the localization of cells analyzed in (D). (B) The relative presence of cells in EV-containing areas (in percentage) was assessed by StrataQuest software as in Figs 4 and 5. The analysis was performed in tissue sections from four different lymph nodes. The respective numbers served to calculate the SEM. (C) Monocytes in imDC-EV and maDC-EV areas. Monocytes (CD11b+/Ly6C+/Ly6G−) were quantified in the EV areas using the StrataQuest software as explained in Table 1 and Fig S8A. (D) Identified monocytes were analyzed for co-expression of additional markers, expressed in percentage of total monocytes found in the EV area. The analysis was performed in tissue sections from four different injection sites. The obtained numbers were used to determine the SEM. Scale bars represent 100 μm. (E) Expression levels of indicated markers for each cell numbered in (A). For better understanding, protein expression was divided into three levels (no, low, and high expression), which were color-coded as in Fig 6B. Scale bars represent 100 μm.
- Figure S10. maBMDC-EV–attracted immune cells express myeloid activation markers (II).
Individual images taken from a MELC analysis (same analysis as described in Fig 5) assessing immune cell markers in BMDC-EV–injected skin tissue areas. The same skin tissue sections shown in Fig 4B were analyzed by MELC for the indicated markers and EV colocalization (yellow). Note: for better orientation, the images on the left were duplicated from Fig 4B. Scale bars represent 100 μm.
- Figure S11. Confocal analysis of tissue sections in EV-injected areas.
(A, B) Tissue sections described in Fig 4B (maDC-EV 24 h after injection) were analyzed for the indicated markers by confocal microscopy as described in the Materials and Methods section. Scale bars represent 10 μm.
- Figure 8. Speculative models of monocyte-derived imDC generation through EV-induced signaling in endosomal compartments.
(A) Generation of monocyte-derived DC through DC-derived EV (“wind mill” model). (1) An imDC receives a maturation stimulus and differentiates into a maDC. (2) The resulting maDC secretes CCF-containing EV, which (3) attract and differentiate monocytes into imDC. (4) In case the maturation stimulus (danger signal) is still present, this DC precursor develops into a mature inflammatory DC. (5) It is assumed (not shown in this manuscript) that the mature inflammatory DC produces more EV and hence perpetuates this cycle. (B) GM-CSF signaling through EV. We depict two principle mechanisms of EV-mediated GM-CSF signaling, a mechanism that suggests sub-membrane signaling (steps 1–5) and outer membrane–mediated signaling (6–8). EV-mediated sub-membrane signaling: (1) GM-CSF precursor proteins are produced in the ER/Golgi and packaged into endosomal compartments, and, upon an activation stimulus (not shown), fuse with compartments containing effector proteases (activator). (2) This leads to the maturation and sequestration of GM-CSF into EV and subsequently (3) secretion of these vesicles through the membrane as shown in Muratori et al (2009). Both steps have been analyzed for the TNF precursor protein (Ostalecki et al, 2016). In the model here, we assume that GM-CSF is also incorporated into the membrane of EV as demonstrated by FACS in Fig S4B. (4) These vesicles are preferentially ingested by monocytes and (5) fuse with endosomal compartments that contain GM-CSF receptors. This leads to GM-CSF–signaling events from endosomal compartments (sub-membrane). EV-mediated outer membrane signaling: (6) EV that have GM-CSF predominantly on the EV membrane are accumulating in multivesicular bodies (MVBs). (7) These MVBs fuse with the outer membrane of the cell and release the EV. (8) In extracellular space, these EV attach to the next GM-CSF receptor, e.g., of a neighboring cells, and induce GM-CSF signaling from the outer membrane.
Tables
Using StrataQuest software, cells were identified and quantified in EV-containing areas through propidium iodide (PI) assessment (41.47% of all cells were found in EV area, first two images). Subsequently, MELC images for monocyte markers (CD11b+/Ly6C+/Ly6G−) were superimposed and colocalizing signals were assessed (EV and markers: 28%). A similar approach was taken to assess neutrophils (CD45+/Ly6C+/Ly6G+) and monocytes in EVca and EVpa areas described in Fig 6A. Scale bars represent 100 μm.
