Skip to main page content
U.S. flag

An official website of the United States government

Dot gov

The .gov means it’s official.
Federal government websites often end in .gov or .mil. Before sharing sensitive information, make sure you’re on a federal government site.

Https

The site is secure.
The https:// ensures that you are connecting to the official website and that any information you provide is encrypted and transmitted securely.

Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
. 2023 Dec 11;42(24):e113590.
doi: 10.15252/embj.2023113590.

Extracellular vesicles and co-isolated endogenous retroviruses from murine cancer cells differentially affect dendritic cells

Federico Cocozza  1   2 , Lorena Martin-Jaular  1   3 , Lien Lippens  4 , Aurelie Di Cicco  5   6 , Yago A Arribas  1 , Nicolas Ansart  1 , Florent Dingli  7 , Michael Richard  7 , Louise Merle  1 , Mabel Jouve San Roman  8 , Patrick Poullet  9 , Damarys Loew  7 , Daniel Lévy  5   6 , An Hendrix  4 , George Kassiotis  10 , Alain Joliot  1 , Mercedes Tkach #  1 , Clotilde Théry #  1   3
Affiliations

Extracellular vesicles and co-isolated endogenous retroviruses from murine cancer cells differentially affect dendritic cells

Federico Cocozza et al. EMBO J. .

Abstract

Cells secrete extracellular vesicles (EVs) and non-vesicular extracellular (nano)particles (NVEPs or ENPs) that may play a role in intercellular communication. Tumor-derived EVs have been proposed to induce immune priming of antigen presenting cells or to be immuno-suppressive agents. We suspect that such disparate functions are due to variable compositions in EV subtypes and ENPs. We aimed to characterize the array of secreted EVs and ENPs of murine tumor cell lines. Unexpectedly, we identified virus-like particles (VLPs) from endogenous murine leukemia virus in preparations of EVs produced by many tumor cells. We established a protocol to separate small EVs from VLPs and ENPs. We compared their protein composition and analyzed their functional interaction with target dendritic cells. ENPs were poorly captured and did not affect dendritic cells. Small EVs specifically induced dendritic cell death. A mixed large/dense EV/VLP preparation was most efficient to induce dendritic cell maturation and antigen presentation. Our results call for systematic re-evaluation of the respective proportions and functions of non-viral EVs and VLPs produced by murine tumors and their contribution to tumor progression.

Keywords: antigen presenting cells; endogenous retrovirus; exosomes; extracellular vesicles; tumors.

PubMed Disclaimer

Conflict of interest statement

GK is co‐founder of EnaraBio and a member of its scientific advisory board. MT is currently an employee of Egle Therapeutics. Both companies are involved in search of anti‐cancer therapies. CT and MT are inventors of a patent on immuno‐therapeutic use of EVs. The other authors have no interests to disclose.

Figures

Figure 1
Figure 1. Characterization of the EVs and ENPs from EO771 cell line: presence of virus‐like particles (VLPs)

  1. Scheme of the protocol of separation of 10k, 200k and ENPs.

  2. Western blot showing total proteins (left) and various protein markers as indicated (right). Lysate from 2 × 105 EO771 cells, 10k, 200k and ENPs secreted by 20 × 106 EO771 cells were loaded on the gel. * indicates specific Hsp90 band of the expected size. ° indicates a shorter size band resulting from endogenous cleavage.

  3. Cryo‐EM of EO771 10k, 200k and ENPs. Arrows indicate typical capsid‐like structures inside EVs. One such structure highlighted by a white square is shown in a zoom (below, right).

  4. Quantification of size and presence (“enveloped capsids”) or absence (EVs) of a VLP structure in all particles of one representative sample of 200k analyzed by cryo‐EM (from the image at 6,500× magnification).

  5. Western blot of cell lysate (Lys) from 2 × 105 cells, 10k and 200k secreted by 20 × 106 of 10 tumor cell lines (EO771, TS/A, LLC1, KP, B16F10, MCA101, MB49, Raw264.7, 4T1 and MutuDC) and 2 non‐tumoral fibroblast cell lines (Pfa1 and Mus Dunni), showing hybridization with antibodies against env (top) and gag (middle) viral proteins and total proteins (bottom). Gag is observed with different sizes, especially the full‐length Pr65 (blue circle), and the mature cleaved p30 (blue cross) forms. Cell lysates of 5 additional cells (MC38, B3Z, EL4, EG7, splenocytes) are shown in Fig EV1C.

Source data are available online for this figure.
Figure EV1
Figure EV1. Characterization and viral content evaluation of a panel of murine cells and their EVs

  1. NTA analysis of 10k, 200k and ENPs of EO771 breast tumor cells: Particle quantification (left) and size distribution (right), n = 4 biological replicates (i.e. independent EV preparations). Error bars = SEM.

  2. Proteomic identification of viral proteins in 200k pellets of EO771, 4T1 and MutuDC. Blue: Proteins identified by specific proteotypic peptides. Green: Proteins from our curated endogenous retrovirus database.

  3. Western blot of cell lysate from 2 × 105 cells of 14 tumor cell lines (EO771, TS/A, 4T1, LLC1, KP, B16F10, MCA101, MB49, MC38, Raw264.7, MutuDC, B3Z, EL4 and EG7), 1 non‐tumoral fibroblast cell line (Pfa1) and primary cell (splenocytes), hybridized with antibodies against env and gag viral proteins and actin, as indicated.

  4. Cryo‐EM of 200k from MutuDC, showing presence of enveloped capsids (arrows).

  5. Infectivity capacity of the 200k from EO771 and MutuDC cells, shown by production of pseudotyped XG7 GFP‐encoding virus from Mus Dunni‐XG7 cells, 14 days after exposure to the EO771 or MutuDC 200k. Scheme of the protocol (left, created by Biorender.com). Pseudotyped virus in the supernatant of Mus Dunni‐XG7 was evidenced by detection of GFP expression in parental Mus Dunni exposed to this supernatant (red), as compared to cells exposed to the supernatant of Mus Dunni‐XG7 not exposed to the 200k (blue).

Source data are available online for this figure.
Figure EV2
Figure EV2. Asymmetric flow field‐flow fractionation (AF4)‐based characterization and separation of EVs by size

  1. Detector‐flow and cross‐flow settings used for the AF4‐based separation of EVs from a EO771 200k pellet.

  2. UV (280 nm), Light Scattering, and particle size measured in the fractions recovered overtime (one experiment). UV signal in the absence of light scatter signal and of detected particles ≥35 nm (P2) corresponds to non‐vesicular proteins or particles.

  3. Differential weight fraction calculated from the UV (280 nm) absorption, as a function of the diameter. Two peaks of mass are observed, suggesting the presence of 2 populations of particles of around 35–80 nm and 80–180 nm, named P3 and P4.

  4. cryo‐EM of the total input and P1, P2, P3, P4 and P5 fractions of B, showing the presence of VLPs in both P4 and P5, together with EVs.

Source data are available online for this figure.
Figure 2
Figure 2. Separation of sEVs and VLPs present in the 200k pellet

  1. Density of fractions of 2 ml (1–8) of the velocity gradient.

  2. Percentage of number of particles in each fraction over the total measured for all the gradient. Each line corresponds to an independent experiment. n = 3.

  3. Protein stain‐free image of a representative gel of fractions 1–8 from the gradient. * ° = prominent bands in fractions 5–7 of the same size as Pr65 and p30 gag bands in 2D.

  4. Representative western blot stained with antibodies against MLV, EV and ENP markers. * ° in Hsp90 = position of full‐length and short bands as observed in the 10k and 200k of Fig 1B.

  5. Cryo‐EM of fractions 1–3 (sEVs) and fractions 5–7 (VLPs) from EO771.

  6. Summary of the combined protocols used to recover 10k, 200k, sEVs, VLPs, ENPs and Mix.

Source data are available online for this figure.
Figure EV3
Figure EV3. Refined separation protocol combining differential ultracentrifugation and velocity gradient applied to different mixed pellets of EVs and characterization of the resulting EV or ENP subtypes

  1. Quantification of particles in fractions recovered from an iodixanol velocity gradient top‐loaded with a EO771 10k or 200k pellet from the same CCM. A single peak of particles recovered in fractions 5–6 is obtained from the 10k.

  2. Western blot analysis of the velocity gradient fractions. Stain‐free total protein (top) and hybridization with antibodies against gag and env MLV proteins or CD9 and CD63 tetraspanins (bottom).

  3. Western blot analysis of the velocity gradient fractions of 200k obtained from Pfa1. Stain‐free total protein stain (top) and hybridization with antibodies against gag and env MLV proteins or CD9, CD63 and Alix EV proteins (bottom).

  4. Particle number (left) and protein amount (right) in each subtype of particle (10k, sEVs, VLPs, ENPs and Mix) secreted by 106 EO771 cells. Kruskal‐Wallis was perfomed with Dunn's multiple comparaisons test.

  5. Median size (left) and zeta potential (right) of each subtype of particle. Mixed‐effects with Geisser–Greenhouse correction and Tukey's multiple comparison statistical analysis was performed. *P < 0.05; **P < 0.01 (D–E). Error bars = SEM (D‐E). n = 13 biological replicates (i.e. independent EV preparations) (D–E).

  6. Percentage of eMLV+ Mus Dunni cells analyzed by FACS after 24 h of exposure to different amounts of proteins of 10k, sEVs, VLPs, ENPs and Mix coming from EO771 cells, and staining with anti‐env 83A25 antibody.

Source data are available online for this figure.
Figure 3
Figure 3. Proteomic characterization of subtypes of particles

  1. Venn diagram of all (mouse + viral) proteins identified with at least three peptides among five biological replicates in at least one sample type among the 10k, sEVs, VLPs and ENPs. All proteins are listed in Dataset EV2, VennDiagram table.

  2. 3D view of the first three components of the principal component analysis (PCA) of all samples based on their protein abundance (LFQ). Samples are colored by sub‐types: 10k = red; sEVs = orange, VLPs = yellow, ENPs = green, Mix = blue.

  3. GO term‐enrichment analysis for cellular components in the top‐100 most specific proteins (from Dataset EV2, SSPA tab): the only 3 significant GO‐terms (for VLPs) or the 15 most significant GO‐terms (for 10k, sEVs and ENPs) are shown. GO‐terms are ordered by adjusted P‐value. Statistical analysis was performed using the default parameters of Enrichr package.

  4. Heatmap of protein abundance of the 15 most specific proteins (or the only 10 specific proteins for VLPs) of each group (from SSPA analysis, Dataset EV2). Clustering is shown to confirm that specific proteins are not identified due to outlier replicates. NA = not detected/absent.

  5. Heatmap of protein abundance of the 46 quantified viral proteins in the different fractions. All gag‐pol sequences are primarily present in the 10k and VLPs, whereas envelope proteins (env) are more equally distributed between the fractions. NA = not detected/absent.

Figure EV4
Figure EV4. Label‐free quantification analysis of the proteomic results

  1. Heatmap of the protein abundance of all proteins identified in one of the 5 groups. NA = not detected/absent.

  2. GO‐term enrichment analysis of the top‐100 most abundant proteins (according to LFQ) in the 4 groups. GO‐terms with P < 0.0001 are represented. GO‐terms are ordered by adjusted P‐value. Statistical analysis was performed using the default parameters of Enrichr package.

Figure 4
Figure 4. Phenotypic changes induced by the different subtypes of particles on MutuDC cells

  1. Scheme of the in vitro experiment and the parameters assessed.

  2. Viability of MutuDC after 16 h of exposure to 10 or 20 μg/ml of the different particles coming from EO771 or Pfa1, measured by flow cytometry with eFluor 780 fixable viability dye (Appendix Fig S1A), normalized to the control non‐treated condition. n = 3.

  3. Viability of MutuDC after 16 h of exposure to 0.5, 1, 2 and 4 × 1010 particles/ml of the different particles coming from EO771 cells, graphed in function of their corresponding concentration of protein. Graph in function of the particle concentration is shown in Fig EV5A. n = 4.

  4. Viability of MutuDC cells after 16 h of exposure to different ratios (1/7, 1/3, 1/1, 3/1, 7/1 and 1/0) of the mixture between sEVs and VLPs reaching 4 × 1010 particles/ml. n = 4.

  5. Viability of MutuDC after 16 h of exposure to 0.5, 1, 2 and 4 × 1010 particles/ml of 10k and 200k coming from a panel of tumor cell lines, 3 carrying MLV gag and env in their EVs (EO771, B16F10 and MCA101), 1 carrying only env (LLC1) and 1 carrying none (KP). n = 2–4.

  6. Expression of maturation markers on viable MutuDC cells after 16 h of exposure to 10 or 20 μg/ml of the different particles coming from EO771 and Pfa1, measured by flow cytometry (Appendix Fig S1A, GeoMean). LPS and CpG treatments were used as positive controls of maturation. n = 5.

  7. Expression of maturation markers on viable MutuDC cells after 16 h of exposure to 0.5, 1, 2 and 4 × 1010 particles/ml of the 10k and 200k coming from the same cells lines as in E. n = 3–4.

  8. Quantification of cytokines secreted by DCs in the supernatant of MutuDC exposed to the subtype of particles (top) and in the particles themselves (bottom), presented as heatmap (log10 scale). “Amount 1 μg/2 μg” refers to the amount of EV/ENP used to treat MutuDC (in 100 μl final volume). n = 3. Statistical analyses were performed using mixed‐effects model with Dunnett's multiple comparison to the PBS, the mean of both concentrations was used for the comparisons, with different concentrations as repeated measures (B, F–H). n = number of biological replicates, i.e. independent EV production and subtype separation from the respective producing cells (B–H). Error bars = SEM (B, D–G). *P < 0.05, **P < 0.01, ***P < 0.001 and ****P < 0.0001 (B, F–H).

  9. Heatmap of protein abundance of proteins qualifying as DAMPs and PAMPs as quantified by LFQ in our proteomic analysis of the different types of particles (Dataset EV2). NA = not detected/absent.

Source data are available online for this figure.
Figure EV5
Figure EV5. Analysis of dendritic cells' phenotype upon exposure to EVs/ENPs, and effects on primary murine and human DCs
  1. A

    Viability of MutuDC after 16 h of exposure to 0.5, 1, 2 and 4 × 1010 particles/ml of the different particles coming from EO771 cells. n = 4 biological replicates (i.e. independent EV preparations).

  2. B

    Viability of spleen DCs from C57BL/6 mice after 16 h of exposure to 20 μg/ml of the different particles coming from EO771tumor cells. Gating strategy follows green arrows in Appendix Fig S1B, i.e. gating DCs or DC subtypes before quantifying DAPIdim cells. Results are shown for total DCs (CD11c+), cDC1 (XCR1+), cDC2 (CD172a+), pDC (B220+) as indicated above each graph. n = 3 biological replicates (i.e. independent EV preparations).

  3. C

    Viability of human MoDCs exposed to 4 × 1010 particles/ml of EO771 EV/ENPs. Gating strategy is shown in Appendix Fig S1D. n = 3 independent EV preparations and 2 MoDCs donors. Ordinary one‐way ANOVA test was performed with Dunnett's multiple comparaison test.

  4. D, E

    Additional maturation markers (CD86, PD‐L1 and MHC‐II) analyzed in parallel in the experiments of Fig 4F and G respectively. n = 5 biological replicates (i.e., independent EV preparations). Mixed‐effects model with Dunnett's multiple comparison to the PBS was performed, the mean of both concentrations was used for the comparisons, with different concentrations as repeated measures. *P < 0.05, **P < 0.01, ***P < 0.001 and ****P < 0.0001 (C–E). Error bars = SEM (B, D‐E).

Source data are available online for this figure.
Figure 5
Figure 5. Protein transfer capacity of the different subtypes of particles

  1. Scheme of the construct containing the myristoylation and palmitoylation sequences fused to mCherry, which was used to create the EO771 myr/palm‐mCherry stable cells (top) and scheme of expected intracellular and intravesicular distribution of myr/palm‐mCherry (bottom, created by Biorender.com).

  2. Confocal microscopy of EO771‐myr/palm‐mCherry cells showing DAPI in blue and mCherry in red, CD9 in green and CD63 in magenta (overlay and close‐ups).

  3. Western blot loaded with EVs/ENPs from 20 × 106 cells hybridized with anti‐mCherry, showing its presence in the different subtypes of particles (same gels as in Figs 1B and 2C and D).

  4. Immunostaining with anti‐mCherry (5 nm gold particles) and transmission electron microscopy (TEM) analysis of saponin permeabilized 10k, 200k and ENPs from EO771 m/p‐mCherry.

  5. Uptake by MutuDC, quantified by flow cytometry (expressed as % of mCherry+ cells), of 10 or 30 ng of mCherry from each subtype of particle (left, n = 5), or a dose–response of recombinant mCherry (right, n = 2).

  6. Uptake by spleen DCs of 30 ng of mCherry from each subtype of particle. Results are presented for total CD11c+ DCs, and subtypes of DCs: cDC1, cDC2, and pDCs (defined as in Appendix Fig S1B). One way ANOVA with Tukey's multiple comparison test was performed, n = 3. *P < 0.05 and **P < 0.01. n = number of biological replicates (i.e. independent EV preparations) (E and F). Error bars = SEM (E and F).

Source data are available online for this figure.
Figure 6
Figure 6. Cross‐presentation of OVA antigen carried by the subtypes of particles
  1. A

    Scheme of the construct containing the myristoylation and palmitoylation sequences fused to OVA, myc tag, P2A cleavage site and puromycin resistance gene for selection, which was used to create the EO771 myr/palm‐OVA stable cells.

  2. B

    Analysis of EVs/ENPs from 20 × 106 EO771‐m/p‐OVA by western Blot. Four known amounts of recombinant OVA were loaded on the same gel to allow quantification of OVA. Blots were revealed with a polyclonal antibody against OVA. One representative blot out of 6. Arrowheads indicate the positions of myr/palm‐OVA (upper band) and recombinant OVA (lower band).

  3. C

    Quantification of OVA in the different EVs/ENPs, was done on the western blot images as compared to the OVA dose–response curve. Graphs show ng OVA/106 cells (left) or ng OVA/particles (right). Friedman test with Dunn's multiple comparison was performed. n = 6.

  4. D

    Scheme of the protocol of cross‐presentation of EV/ENP‐OVA cargo by MutuDC to OT1 T cells.

  5. E

    Dot plots of flow cytometry analysis of OTI cells (CD69 and CD25 expression), 18 h after exposure to MutuDC that had been fed with OVA present in the various types of EVs/ENPs (one representative replicate).

  6. F, G

    Quantification of CD69− (F) and double CD69/CD25 (G)‐ expressing OT1 T cells after exposure to MutuDC that had been fed with OVA present in the various types of EVs/ENPs. Mixed‐effects model was performed with Dunnett's multiple comparison to the PBS, the mean of both concentrations was used for the comparisons, with different concentrations as repeated measures, n = 6.

  7. H

    Cross‐presentation dose–response using recombinant OVA.

  8. I

    Cross‐presentation efficacy was calculated by dividing the amount of rOVA needed to elicit a certain degree of CD69 expression (fitting the dose–response curve shown in Fig 6H) to the actual amount of OVA present in each EV/ENPs. n = 6 (only using the 20 ng/ml condition). Ordinary one‐way ANOVA test was performed with Tukey's multiple comparaison. *P < 0.05, **P < 0.01, and ***P < 0.001 (C, F, G, I). Error bars = SEM (C, F, G, I). n = number of biological replicates (i.e. independent EV preparations).

Source data are available online for this figure.
See this image and copyright information in PMC

References

    1. Attig J, Young GR, Stoye JP, Kassiotis G (2017) Physiological and pathological transcriptional activation of endogenous retroelements assessed by RNA‐sequencing of B lymphocytes. Front Microbiol 8: 2489 - PMC - PubMed
    1. Balaj L, Lessard R, Dai L, Cho YJ, Pomeroy SL, Breakefield XO, Skog J (2011) Tumour microvesicles contain retrotransposon elements and amplified oncogene sequences. Nat Commun 2: 180 - PMC - PubMed
    1. Balint S, Muller S, Fischer R, Kessler BM, Harkiolaki M, Valitutti S, Dustin ML (2020) Supramolecular attack particles are autonomous killing entities released from cytotoxic T cells. Science 368: 897–901 - PMC - PubMed
    1. Banchereau J, Steinman RM (1998) Dendritic cells and the control of immunity. Nature 392: 245–252 - PubMed
    1. Beck R, Verrax J, Gonze T, Zappone M, Pedrosa RC, Taper H, Feron O, Calderon PB (2009) Hsp90 cleavage by an oxidative stress leads to its client proteins degradation and cancer cell death. Biochem Pharmacol 77: 375–383 - PubMed

Publication types