Human Endothelial Cells Modulate CD4+ T Cell Populations and Enhance Regulatory T Cell Suppressive Capacity

Abstract

Endothelial cells (ECs) line the luminal surface of blood vessels and have an active role in the recruitment of leukocytes, including immune cell activation. Regulatory T cells (Tregs) are immune suppressor cells that maintain peripheral tolerance and must interact with the endothelium as they traffic into tissue. We hypothesized that human ECs could modulate Tregs and their suppressor function. Cocultures of CD4⁺ T cells with human umbilical vein ECs (HUVECs) or dermal microvascular ECs (HDMECs) were conducted and analyzed for activation and proliferation after 72 and 120 h using flow cytometry. In monocyte-depleted cultures, human ECs were found to support CD4⁺ T cell proliferation in the presence of external mitogens phytohemagglutinin or anti-CD3/28 antibodies (aCD3/28). Activation was shown by CD25 expression in these cells that also transiently expressed the Treg transcription factor FOXP3. HUVECs supported the specific concurrent proliferation of both effector T cells and Tregs when cocultured with aCD3/28. Purified Tregs were also functionally activated by prior coculture with EC to suppress effector T (Teff) cell proliferation. Both direct coculture and indirect coculture of EC and Treg showed activation of the Treg suppressive phenotype. However, whereas HUVEC showed enhancement of suppression by both mechanisms, HDMEC only supported Treg suppressive activity via the contact-independent mechanism. In the contact-independent cultures, the soluble mediators IL-6, GM-CSF, or G-CSF released from ECs following interferon-γ activation were not responsible for the enhanced Treg suppressor function. Following direct coculture, Treg expression of inhibitory receptors PD-1 and OX40 was elevated while activated EC expressed the counter ligands programmed death ligand (PD-L)1 and PD-L2. Therefore, human ECs have a role in supporting T cell proliferation and increasing Treg suppressor function. This ability of EC to enhance Treg function could offer novel targets to boost Treg activity during inflammatory disorders.

Keywords: CD4+ T cells; cell–cell interaction; endothelial cells; immune regulation; regulatory T cells.

Figure 1

Proliferation of CD4⁺ T cells in endothelial cell (EC) cocultures and expression of FOXP3 of proliferated CD4⁺ T cells. Cells were harvested from cocultures of CD4⁺ T cells with HUVECs at 72 or 120 h and analyzed by flow cytometry, where T cells were discriminated based on CD4 expression. (A) CD4⁺ T cells with the highest CFSE fluorescence were gated as unproliferated whereas those with diluted/low CFSE fluorescence were gated as the proliferated population. The CD25 and FOXP3 expression of proliferated cells (blue line) were analyzed with the aid of the isotype control (red line). (B) Proliferation of CD4⁺ T cells cultured with unstimulated (uEC) or IFN-γ (iEC) or tumor necrosis factor (TNF)-α-stimulated (tEC) HUVECs were analyzed by CFSE dilution at 72 h, in which coculture had no external stimuli or were stimulated with phytohemagglutinin (PHA) or aCD3/28 antibodies. One-way analysis of variance (ANOVA) with Tukey test was conducted; ****P < 0.0001, *P < 0.05 cf. control; n = 10. (C) Comparison of CD4⁺ T cell proliferation in cocultures with uEC, iEC, or tEC at 72 and 120 h, in which cocultures had no external stimuli or were stimulated with PHA or aCD3/28 antibodies. Two-way ANOVA analysis with Tukey test was conducted; ****P < 0.0001; **P < 0.01 for 120 h cf. 72 h data; ^####P < 0.0001, ^##P < 0.01 cf. control for 120 h data; n = 10/6. (D) FOXP3 expression of the proliferated CD4⁺ T cell populations at 72 and 120 h from the different HUVEC-CD4⁺ T cell cocultures were compared, where data are expressed as percentage of CD4⁺FOXP3⁺ cells from the total number of CD4⁺ T cell analyzed, with mean ± SD shown. Two-way ANOVA with Tukey test was conducted, where ****P < 0.0001, ***P < 0.001, *P < 0.05 for data at 120 h cf. data at 72 h; n = 4.

Figure 2

Proliferation of CD4⁺ T cells in HDMEC cocultures and their expression of FOXP3. (A) Phase contrast micrographs of HDMEC-CD4⁺ T cell cocultures using unstimulated or interferon (IFN)-γ-stimulated HDMECs in the absence or presence of stimulatory aCD3/28 antibodies were taken at 72 h; micrographs were taken at 100× original magnification; scale bar of 50 µm included. (B) CFSE-labeled CD4⁺ T cells were cocultured with HDMECs used either unstimulated (uEC) or stimulated with 10 U/mL IFN-γ (iEC) for 24 h; proliferation of CD4⁺ T cells was analyzed by CFSE dilution at 72 h. Proliferation of CD4⁺ T cells in HDMEC-CD4⁺ T cells cocultures stimulated with aCD3/28 antibodies were also measured. One-way analysis of variance (ANOVA) with Dunnett’s test was conducted; *P < 0.05 cf. control, n = 5. (C) FOXP3 expression of proliferated CD4⁺ T cells from the same HDMEC-CD4⁺ T cell cocultures was also analyzed at 72 h, where data are expressed as percentage of CD4⁺FOXP3⁺ cells from the total CD4⁺ population with means shown. One-way ANOVA with Dunnett’s test was used; **P < 0.01 cf. control; n = 5.

Figure 3

Enhancement of Treg suppressor function after coculture with activated human umbilical vein ECs (HUVECs). (A) Treg suppression assays were set up where CFSE-labeled Teffs were either stimulated with 5 µg/mL plate-bound aCD3 antibody (in the presence of gamma-irradiated autologous PBMCs) or with CD4⁺CD25^hiCD127^low Tregs at a 1:1 ratio. Teff proliferation was analyzed by CFSE dilution and the suppressive capacity of Tregs was measured by comparing the reduction of Teff proliferation in the presence of Tregs. Tregs from 16 different donors were found to suppress Teff proliferation in varying capacity. Two-tailed paired T test was conducted; ****P < 0.0001; n = 16. (B) CD4⁺CD25^hiCD127^low cells (Tregs) were cocultured with unstimulated (uEC), interferon (IFN)-γ- (iEC), tumor necrosis factor (TNF)-α- (tEC), or both IFN-γ and TNF-α (itEC) HUVECs for 24 h, before they were recovered and plated in a Treg suppression assay to measure suppressive capacity. Raw data is expressed as % proliferated Teff, with each experiment represented by a different color. Data are also expressed where Teff proliferation is normalized to proliferation in 1:1 cultures as 100%, showing mean ± SEM. One-way analysis of variance (ANOVA) with Dunnett’s test was used; *P < 0.05, n = 8. (C) The contact-dependent nature of the modulation of Treg function by HUVEC was investigated where Tregs were plated in direct contact with uEC or iEC, or suspended above unstimulated (uEC TW) or IFN-γ-stimulated (iEC TW) HUVECs in Transwells for 24 h, before plating in a Treg suppression assay. Raw data are expressed as % proliferated Teffs, with each experiment represented by a different color. Data are also expressed where Teff proliferation is normalized to proliferation in 1:1 cultures as 100%, showing mean ± SEM. One-way ANOVA with Dunnett’s test was used; **P < 0.01, *P < 0.05, n = 8.

Figure 4

Increased Treg suppressive capacity through contact-independent interaction with activated HDMECs. (A) The effect of HDMEC interaction on Treg function was investigated whereby CD4⁺CD25^hiCD127^low were plated in direct contact with unstimulated (uEC) or IFN-γ-stimulated (iEC) HDMECs, or suspended above unstimulated (uEC TW) or IFN-γ-stimulated (iEC TW) HDMECs in Transwells for 24 h, then recovered and plated in Treg suppression assays. Data are expressed as % proliferated Teffs, with each experiment represented by a different color. Data are also expressed where Teff proliferation is normalized to proliferation in 1:1 cultures as 100%, showing mean ± SEM. One-way analysis of variance (ANOVA) with Dunnett’s test was used; *P < 0.05; n = 8. (B) HDMECs were left unstimulated or stimulated with 10 U/mL IFN-γ, 1 ng/mL TNF-α, or a combination of both cytokines for 24 h. Cells were washed and media replaced with 1 mL complete RPMI. Conditioned media was collected after 24 h and secreted cytokines detected using bead-based multiplex assays or ELISAs. Production of IL-6, G-CSF, and GM-CSF by HDMECs are shown with means drawn. One-way ANOVA with Dunnett’s test was conducted; ****P < 0.0001, ***P < 0.001, **P < 0.01, and *P < 0.05 cf. control; n = 6. (C) Tregs were cocultured in the upper chambers of Transwells with IFN-γ-stimulated (iEC TW) HDMECs in the lower chambers for 24 h, where 5 µg/mL of anti-IL-6, anti-G-CSF, or anti-GM-CSF blocking antibody was added in the lower chambers. Tregs were recovered and plated in a Treg suppression assay where Teff proliferation was measured by CFSE dilution. Data are expressed as % proliferated Teffs, with each experiment represented by a different color. Data are also expressed where Teff proliferation is normalized to proliferation in 1:1 cultures as 100%, showing mean ± SEM. One-way ANOVA with Dunnett’s test was used; *P < 0.05; n = 8.

Figure 5

Involvement of PD-1/programmed death ligand (PD-L)1/2 in the contact-dependent interactions between HUVEC and Tregs. Human umbilical vein ECs (HUVECs) were stimulated with 10 U/mL IFN-γ, 1 ng/mL tumor necrosis factor (TNF)-α, or combination of both cytokines for 24 h. Cells were then stained and analyzed for (A) PD-L1 or (B) PD-L2 expression using flow cytometry. Data are expressed as mean fluorescence index (MFI) and percentage of positive cells, with means shown. One-way analysis of variance (ANOVA) with Dunnett’s test was conducted; ****P < 0.0001, ***P < 0.001, *P < 0.05; n = 3. Modulation of Treg phenotype after HUVEC interactions were also investigated whereby Tregs were plated in direct contact with unstimulated (uEC Treg) or IFN-γ-stimulated (iEC Treg) HUVECs, or in the upper chambers of Transwells with unstimulated (uEC TW Treg) or IFN-γ-stimulated (iEC TW Treg) HUVECs. Tregs were then recovered and plated in a Treg suppression assay and analyzed for (C) PD-1 and (D) OX40 expression after 72 h using flow cytometry. One-way ANOVA with Dunnett’s test was conducted; *P < 0.05; n = 3. Plots of a representative experiment are shown in Figures (E,F).