Extracellular Vesicles Generated by Gram-Positive Bacteria Protect Human Tissues Ex Vivo From HIV-1 Infection

Abstract

Vaginal microbiota dominated by lactobacilli protects women from sexually transmitted infection, in particular HIV-1. This protection is, in part, mediated by Lactobacillus-released extracellular vesicles (EVs). Here, we investigated whether EVs derived from other Gram-positive bacteria also present in healthy vaginas, in particular Staphylococcus aureus, Gardnerella vaginalis, Enterococcus faecium, and Enterococcus faecalis, can affect vaginal HIV-1 infection. We found that EVs released by these bacteria protect human cervico-vaginal tissues ex vivo and isolated cells from HIV-1 infection by inhibiting HIV-1-cell receptor interactions. This inhibition was associated with a diminished exposure of viral Env by steric hindrance of gp120 or gp120 modification evidenced by the failure of EV-treated virions to bind to nanoparticle-coupled anti-Env antibodies. Furthermore, we found that protein components associated with EV's outer surface are critical for EV-mediated protection from HIV-1 infection since treatment of bacteria-released EVs with proteinase K abolished their anti-HIV-1 effect. We identified numerous EV-associated proteins that may be involved in this protection. The identification of EVs with specific proteins that suppress HIV-1 may lead to the development of novel strategies for the prevention of HIV-1 transmission.

Keywords: Enteroccoccus faecalis; Enteroccoccus faecium; Gardnerella vaginalis; HIV-1; Staphylocccus aureus; extracellular vesicles (EVs); gram positive bacteria; vaginal microbiota.

Figure 1

Bacterial EV-size and concentration. Nanoparticle tracking analysis of EVs derived from S. aureus, E. faecium, E. faecalis, and G. vaginalis. (A) EV-sizes expressed as mean ± SEM of particle diameter (nm). (B) Mean ± SEM of the EV concentration (particles/mL). Presented are the results of three independent measurements. **p < 0.01.

Figure 2

Antiviral effect of bacterial-derived supernatants, EVs, and EV-free supernatants. A mixture of HIV-1_LAI.04 and bacterial-derived supernatants (0.5%), EVs (1×10¹⁰ EVs/mL), or EV-free supernatants (0.5%) incubated for 1 h was added to MT-4 cell cultures. Cells were washed and cultured for 3 days in the presence of bacterial supernatants, EVs, or EV-free supernatants. In control experiments, EV-free, particles isolated from fresh MRS medium, and MRS medium were tested as well. Replication of HIV-1 was evaluated from measurements of the capsid protein p24_gag, in the cell culture medium; data are presented as percentages of HIV-1 replication compared with untreated controls. (A) The effects of bacterial supernatants (0.5%) derived from S. aureus, E. faecium, E. faecalis, G. vaginalis, and MRS medium (0.5%) on HIV-1 replication in MT-4 cells. (B) The effects of 1×10¹⁰ EVs/mL bacterial EVs derived from S. aureus, E. faecium, E. faecalis, G. vaginalis, and MRS medium-derived particles on HIV-1 replication in MT-4 cells. (C) The effects of bacterial EV-free supernatants (0.5%, after pulling down EVs by ultracentrifugation) derived from S. aureus, E. faecium, E. faecalis, and G. vaginalis on HIV-1 replication in MT-4 cells. Presented are means ± SEM from at least three independent measurements. Asterisks indicate statistical significance by one-way ANOVA multiple comparison with Dunnett’s correction (*p < 0.05, ***p < 0.001, ****p < 0.0001).

Figure 3

Concentration effect of bacterial EVs on HIV-1 replication. MT-4 cells infected with HIV-1_Lai.04 were cultured in presence of bacterial EVs [S. aureus (A), E. faecium (B), E. faecalis (C), and G. vaginalis (D)] at different EV concentrations, 1×10⁷, 1×10⁸, 1×10⁹, and 1×10¹⁰ EVs/mL. After 3 days of culture, HIV-1 replication was evaluated from measurements of the concentration of p24_gag in the cell culture medium. The data are presented as percentages of HIV-1 replication compared with untreated controls. Presented are means ± SEM from six independent measurements. Asterisks indicate statistical significance by one-way ANOVA multiple comparison with Dunnett’s correction (*p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001).

Figure 4

Cell viability of cells treated with bacterial supernatants or EVs. MT-4 cells were treated or not treated for 3 days with bacterial supernatants (0.5%) or EVs (1×10¹⁰ EVs/mL) derived from S. aureus, E. faecium, E. faecalis, and G. vaginalis. MRS medium or particles derived from MRS were used as controls. The numbers of viable and dead cells were counted according to orange acridine/propidium-iodide-based assay. Results are expressed as percentages of cell viability in EV-free or EV-treated cells. Presented are means ± SEM from five independent measurements. Cell viability in presence of bacterial supernatants (A) or EVs (B) derived from S. aureus, E. aecium, E. faecalis, and G. vaginalis is shown. Results are expressed as percentages of viable cells treated or not treated with bacterial supernatant or EVs. Presented are means ± SEM from at least three independent measurements.

Figure 5

Anti-HIV-1 effect of bacterial EVs in human cervico-vaginal tissues ex vivo. Cervico-vaginal tissue blocks were infected with EV-pretreated HIV-1_BaL and cultured for 12 days, with replacement every 3 days of tissue culture medium containing or not containing 1×10¹⁰ EVs/mL derived from S. aureus, E. faecium, E. faecalis, and G. vaginalis. Replication of HIV-1 was evaluated from measurements of the capsid protein p24_gag in tissue culture medium and is represented as a percentage of HIV-1 replication in untreated control. Presented are means ± SEM from tissues of at least four donors. Asterisks indicate statistical significance by one-way ANOVA multiple comparison with Dunnett’s correction (**p < 0.01, ****p < 0.0001).

Figure 6

Effect of cell pre-exposure with bacterial EVs on HIV-1 infection. MT-4 cells were preincubated with bacterial EVs for 24 h, washed off, infected with HIV-1_LAI.04, and incubated for 3 days. HIV-1 replication was evaluated from measurements of the capsid protein p24_gag in cell culture medium and is represented as a percentage of HIV-1 replication in untreated control. Presented are means ± SEM from at least three independent measurements.

Figure 7

HIV-1 capture. HIV-1_LAI.04 was pre-treated with 1×10¹⁰ EVs derived from S. aureus, E. faecium, E. faecalis, or G. vaginalis, or with particles derived from MRS medium, or with PBS (control) for 1 h. Next, HIV-1_LAI.04 virions were captured with PG9 (A), VRC01 (B), or 4B3 (C) antibodies coupled to magnetic nanoparticles (MNPs). PG9 antibody recognizes HIV-1 trimeric gp120 proteins, VRC01 recognizes the CD4 binding site on the viral gp120, and 4B3 antibodies recognize the viral gp41. Data are presented as percentage of p24_gag concentration compared with the control. Presented are means ± SEM from three independent measurements. Asterisks indicate statistical significance by one-way ANOVA multiple comparison with Dunnett’s correction (*p < 0.05, ***p < 0.001, ****p < 0.0001).

Figure 8

Bacterial EV-associated protein role during HIV-1 infection. The proteins associated to bacterial EVs were digested with PK followed by an ulterior ultracentrifugation to obtain bacterial EVs free of surface proteins. HIV-1_LAI.04 was pre-treated with 1×10¹⁰ PK-treated EVs derived from S. aureus, E. faecium, E. faecalis, or G. vaginalis, or with PK (PBS treated with PK and purified by ultracentrifugation), or with PBS (control) for 1 h. Next, HIV-1_LAI.04 virions were captured with PG9 (A) or VRC01 (B) antibodies. Also, the antiviral effect of PK-treated bacterial EVs in MT-4 cells infected with HIV_LAI.04 was evaluated (C). The amount of virus captured and the amount of virus present on cell culture medium were determined from measurement of the levels of viral p24_gag. Data are presented as percentages of p24_gag concentration compared with the control (PBS). Presented are means ± SEM from three independent measurements. Statistical analysis was performed with one-way ANOVA multiple comparison.

Figure 9

Cellular localization and molecular function GOs of S. aureus, E. faecium, E. faecalis and G.vaginalis-derived EV proteome. (A) Relative abundance of the proteins predicted to have a cytoplasmic, extracytoplasmic, or intracellular localization within each strain-derived EV proteome. Detailed results of TMHMM and SignalP analysis used to predict TM and signal peptide motifs in the EV-related proteins are reported in Tables S1–S4. (B) Relative abundance of the proteins belonging to each molecular function’s GO within each strain-derived EV proteome; ‘other’ includes GO terms with a relative abundance < 1%. Molecular functions were extracted from strain-derived EV proteomes, analyzed in topGO, and categorized in CateGOrizer (Hu et al., 2008) against the GOSlim2 database. Raw data are reported in Tables S1–S11; sa and gv indicate molecular functions that were identified only in the S. aureus and G. vaginalis-derived EVs proteome, respectively.

Figure 10

Molecular function GOs significantly enriched in the EV derived proteins that are shared among S. aureus, E. faecium, E. faecalis and G.vaginalis. Raw data used for the plot are reported in Tables S1–S11.