Extracellular vesicles in human semen modulate antigen-presenting cell function and decrease downstream antiviral T cell responses

Abstract

Human semen contains trillions of extracellular vesicles (SEV) similar in size to sexually transmitted viruses and loaded with potentially bioactive miRNAs, proteins and lipids. SEV were shown to inhibit HIV and Zika virus infectivity, but whether SEV are able also to affect subsequent immune responses is unknown. We found that SEV efficiently bound to and entered antigen-presenting cells (APC) and thus we set out to further dissect the impact of SEV on APC function and the impact on downstream T cell responses. In an APC-T cell co-culture system, SEV exposure to APC alone markedly reduced antigen-specific cytokine production, degranulation and cytotoxicity by antigen-specific memory CD8+ T cells. In contrast, inhibition of CD4+ T cell responses required both APC and T cell exposure to SEV. Surprisingly, SEV did not alter MHC or co-stimulatory receptor expression on APCs, but caused APCs to upregulate indoleamine 2,3 deoxygenase, an enzyme known to indirectly inhibit T cells. Thus, SEV reduce the ability of APCs to activate T cells. We propose here that these immune-inhibitory properties of SEV may be intended to prevent immune responses against semen-derived antigens, but can be hi-jacked by genitally acquired viral infections to compromise adaptive cellular immunity.

Fig 1. Size profile of SEV and binding of SEV to leukocytes.

(A) Pooled preparations of SEV were analyzed using nanoparticle tracking analysis. A representative size profile is presented along with the published sizes of common sexually transmitted viruses to demonstrate that SEV and viral STIs substantially overlap in size. Sexually transmitted human herpesviruses include EBV, CMV, HSV-1, HSV-2, and KSHV. (B) Fluorescently labeled SEV were incubated with PBMCs at 10⁵ per cell. At 2, 4, 6 or 8, and 10 hours post addition samples were moved to 4° to arrest SEV binding and entry. The next day, samples were stained for phenotypic markers and analyzed by flow cytometry. Data from two different PBMC donors are shown as the percent of cell type positive for DiI. (C) Histograms of DiI staining on PBMCs gated on the indicated populations. (D) Confocal microscopy of blood dendritic cells (DC) or Langerhans cells (LC) from vaginal mucosal tissue exposed to DiI-labeled SEV for 4 hrs. DC or LC were stained for expression of MHC class II (HLA-DQ) and are shown in green, DiI SEV in red, and cell nuclei in blue (DAPI). Nuclei of unstained cells conjugated to LCs are T cells. (E) Annexin and cytochalasin D inhibit SEV binding and entry. Annexin-pretreated SEV (10 μg/mL) were incubated with PBMCs; untreated DiI-labeled SEV were added to cytochalasin D-pretreated (20 μM) or untreated PBMCs as above. Data are presented as percent of cells positive for DiI-SEV relative to the condition without annexin or cytochalasin D treatment, averaged across two independent PBMC donors.

Fig 2. SEV inhibit EBV and CMV-specific memory immune responses.

(A) PBMC from 10 individuals were exposed to CMV peptides, CMV lysate, EBV peptides or EBV lysate and 10⁵ SEV per cell (or left unexposed). Production of cytokines was assessed by intracellular cytokine staining for IFNγ, IL-2 and TNFα, and the sum of the percent of cells responding with any cytokine is reported, separately for CD4+ and CD8+ T cells. Each color indicates a different blood donor. Each donor was tested in 2–3 independent experiments and responses are averaged. Significance by Wilcoxon matched-pairs signed rank test. (B) For each donor, the average percent reduction in the fraction of cytokine producing cells from antigen alone-exposed cells is plotted, separately for CD4+ and CD8+ T cells. Colors indicate different donors, as in (A). Significance by unpaired t test. (C) PBMC were exposed to CMV peptides at t = 0 and 10⁵ SEV per cell were added at the indicated times. “Never” indicates no addition of SEV. Data were analyzed for cytokine production as in (A).

Fig 3. SEV inhibit T cell cytokine production upstream of protein kinase C activation.

(A) PBMC were stimulated with staphylococcal enterotoxin B (SEB) at 1 μg/mL or phorbol myristate acetate (PMA) at 50 ng/ml plus ionomycin at 1 μg/mL for 6 hrs, in the presence or absence of SEV. Production of cytokines was assessed by intracellular cytokine staining for IFNγ, IL-2 and TNFα, and the sum of the percent of cells responding with any cytokine is reported. Each color indicates a different blood donor. The reduction in the fraction of cytokine-producing cells in SEV-exposed cells is significant by Wilcoxon matched-pairs signed rank test (CD4+ SEB p = 0.0078; CD4+ PMA p = 0.0039; CD8+ SEB p = 0.0078; CD8+ PMA p = 0.0039). (B) The percent reduction in SEV-exposed cells from stimulated cells alone is plotted. Differences in the percent reduction for SEB compared to PMA/ionomycin are significant by unpaired t test (CD4+ p = 0.02; CD8+ p = 0.006). (C) DC differentiated from blood precursors were exposed to CMV peptides and a maturation stimulus. SEV were added to DC or autologous purified T cells or both at the time of antigen loading and incubated for 20 hrs. After washing, T cells were mixed with antigen-loaded DC at a ratio of 1 DC to 30 T cells. Percent reduction in the sum of cytokine-producing cells compared to CMV-loaded DC mixed with SEV-untreated T cells is plotted. Significance by one-way ANOVA with repeated measures (p = 0.0003). (D) DC were exposed to CMV peptides and 50% or 20% of the DC were simultaneously treated with 10⁵ SEV/cell. DC were washed, SEV-treated and -untreated DC were combined, and the DC mixtures were added to autologous T cells at a DC to T cell ratio of 1:30. Production of cytokines was assessed as in (C) and percent reduction from antigen-loaded, SEV-unexposed DC + T cells is plotted.

Fig 4. CD8+ T cells activated by SEV-exposed APCs are impaired in degranulation and killing capacity.

(A) PBMC were exposed to SEB and SEV (or left unexposed) for 6 hrs and washed to remove free SEV. Degranulation of CD8+ T cells was assessed by surface staining of cells with anti-CD107a antibody and flow cytometry (gated on CD3+ T cells); percent of CD4+ and CD8+ T cells positive for CD107a is shown. Significance by Wilcoxon matched-pairs signed rank test (p = 0.008). (B) Comparison of percent reduction in the production of cytokines compared to degranulation in SEV-exposed T cells (relative to SEV-unexposed cells seeing only SEB). (C) CMV-loaded DC, T cells, or both, were separately exposed to SEV for 20 hrs before co-culturing the cells for 6 hrs, as in Fig 5C. Cell surface expression of CD107a was analyzed as in (A) and the percent reduction in CD107a+ CD8+ T cells from the no SEV condition is shown. (D) Gating scheme for killing assay. CMV-specific T cells were expanded for 6 days on CMV-peptide-loaded or unloaded DC. CD8+ T cells were then isolated by negative selection and mixed at 10:1, 5:1, and 2.5:1 ratios with autologous target DC. Target DC consisted of an equal 1:1 mix of CellTrace Violet (CTV)-stained DC loaded with CMV peptides and carboxyfluorescein succinimidyl ester (CFSE)-stained DC loaded with irrelevant HIV peptides. Cells were gated on live high side scatter target cells. Specific killing was calculated as (1 –(CTV/CFSE live cells for experimental condition / CTV/CFSE live cells for target alone control)) x 100 (E) Percent specific killing for each PBMC donor tested, calculated as in D. Non-specific CD8+ T cells were cultured with DC not exposed to any antigen but otherwise generated in the same way as CMV-specific CD8+ T cells. For “Targets”, SEV at 10⁵ per cell were added to labeled DC targets at the time of antigen-loading and DC targets were washed before mixing with CMV-specific CD8+ T cells. For “T cells & targets”, additional SEV were added at the time of mixing CMV-specific CD8+ T cells with labeled DC targets. E:T signifies the ratio between CD8+ T cell effector cells and DC target cells. (F) P values from comparison between SEV-treated and -untreated conditions in the killing assay in panel E, by one-way ANOVA with repeated measures. (G) The percent reduction by SEV in specific killing is plotted relative to CMV-specific CD8+ T effector cells and DC target cells not exposed to SEV. There was no specific killing at the 2.5 ratio for donor A, hence no percent reduction is plotted.

Fig 5. PBMC donors vary in susceptibility to SEV-mediated impairment of CD8+ T cell responses.

(A) The percent reduction from control (no SEV) is plotted by PBMC donor for all tested assays. Black horizontal lines indicate mean percent reduction. Differences between donors is significant by one-way ANOVA (p = 0.024). (B) Correlation between reduced cytokine production and reduced degranulation as measured by CD107a expression on SEV-exposed CD8+ T cells. Colors as in (A). Gray areas indicate the 95% confidence interval for the correlation.

Fig 6. MHC and co-stimulatory marker, and IDO expression, in SEV-treated dendritic cells (DC).

(A) Monocyte-derived DC were left alone or treated with a maturation cocktail (monocyte-conditioned medium, MCM), with or without 10⁵ SEV per cell. Cells were stained for surface expression of the indicated markers and analyzed by flow cytometry. One representative donor of n = 3 tested is presented. (B) DC treated overnight with 10⁶ SEV per cell were analyzed for IL-10 and TGF-β mRNA expression by qRT-PCR. Fold expression relative to mock treated DC is presented. Gray color indicates blood donors not used in previous experiments. The other colors match donors in previous experiments. (C) DC treated overnight or for 6 hours with 10⁴, 10⁵ or 10⁶ SEV per cell were analyzed for IDO mRNA expression by qRT-PCR. Fold change of IDO expression in SEV-treated cells compared to mock-treated cells was calculated by the delta-delta Ct method. (D) Gating scheme for calculating the percent of cells expressing high levels of IDO protein in one representative donor. Cells treated overnight as in C were analyzed for intracellular IDO protein expression by flow cytometry. Cells were gated on singlets (not shown), live high side scatter cells, then IDO expression in HLA-DQ⁺ cells. Percent of cells plotted in (E) are based on the IDO gates shown here. (E) Plot of the percent of cells from 8 different donors expressing high IDO after treatment with 10⁴, 10⁵, or 10⁶ SEV per cell overnight. High IDO is defined as in D.