Cutting Edge: Vascular Endothelial Growth Factor-Mediated Signaling in Human CD45RO⁺ CD4⁺ T Cells Promotes Akt and ERK Activation and Costimulates IFN-γ Production

Vascular endothelial growth factor (VEGF), a well-established angiogenesis factor, has been found to have potent proinflammatory properties, including an ability to mediate leukocyte trafficking into sites of cell-mediated immunity (1–5). Some of its proinflammatory properties are dependent on direct interactions with its receptors expressed on monocytes (6), and some are thought to be related to the ability of VEGF to induce endothelial activation and chemokine production (3). Furthermore, a relatively underappreciated aspect of VEGF biology is that there is also evidence that VEGF has direct biological effects on T cells. For instance, VEGF has been observed to augment Ag-induced cytokine production, including both Th1 (7), Th2 (8), and Th17 (9) responses. However, the mechanism(s) and basis for these interactions are poorly understood.

The biological activities of VEGF are mediated by its receptors Flt-1 (VEGF receptor [VEGFR]1), KDR (VEGFR2, also called Flk1 in the mouse), and neuropilin-1 (NRP-1) (10, 11). All VEGFRs are expressed by endothelial cells, and select receptors are reported to be expressed by cells of the immune system (6, 7, 12–14). Flt-1 and NRP-1 are expressed on human monocytes and APCs (6, 15), and Flt-1 and KDR/Flk-1 have been identified on murine populations of T cells (13) and human leukemic T cell lines (14, 16). Furthermore, a recent report indicated that murine CD4⁺CD25⁺FoxP3⁺ T regulatory cells (but not effector cells) express NRP-1, which was found to function in Ag presentation (12). Moreover, other studies have suggested that NRP-1 is expressed on populations of human naive T cells where it functions in the initiation of T cell activation and in primary immune responses (17). Thus, there are several reports indicating that VEGFRs are expressed on different T cell subsets, suggesting the potential importance of VEGF–VEGFR interactions in immunity.

In this study, we observed that the VEGFRs, KDR and Flt-1, are expressed on the CD45RO⁺ subset of human CD4⁺ T cells. Furthermore, we demonstrate that VEGF stimulates KDR-induced signals within this subset of T cells, costimulates IFN-γ production, and mediates chemotaxis. Our results define a novel function for VEGF and KDR in the immune response and provide for the intriguing possibility that overexpressed VEGF at sites of chronic inflammation may facilitate the peripheral homing and local reactivation of memory populations of T cells.

Human rVEGF and IFN-γ–inducible protein of 10 kDa (IP)-10 were obtained from R&D Systems (Minneapolis, MN), and anti-CD3 and anti-CD28 were obtained from BD Pharmingen (San Diego, CA). For Western blotting, anti–phospho-ERK1/2 (polyclonal anti-Thr²⁰²/Tyr²⁰⁴) and anti–phospho-Akt (polyclonal anti-Ser⁴⁷³) were purchased from Cell Signaling Technology (Danvers, MA), Abs to total ERK, total Akt (clone C-20), NRP-1 (clone C-19), and Flt-1 (clone C-17) were purchased from Santa Cruz Biotechnology (Santa Cruz, CA), and anti-KDR (clone 55B11) was purchased from Cell Signaling Technology. For FACS analysis, PE-conjugated anti-KDR (clone 89106) and anti–Flt-1 (clone 49560) were purchased from R&D Systems, anti–NRP-1 (clone AD5-17F6) was obtained from Miltenyi Biotec (Auburn, CA), and isotype controls were purchased from BD Pharmingen. The pharmacological KDR inhibitor SU5416 and pertussis toxin were obtained from Sigma-Aldrich (St. Louis, MO), and the Akt signaling inhibitor LY294002 and the ERK signaling inhibitor PD 98059 were purchased from Calbiochem (San Diego, CA).

CD4⁺CD45⁺RO⁺ memory T cells were isolated from the blood of healthy volunteers using a negative isolation kit (Miltenyi Biotec, Auburn, CA) according to the manufacturer’s protocol. The purity of T cells was determined by FACS after each isolation. For occasional experiments, pooled populations of CD4⁺ T cells were purified using the Dynal positive isolation kit (Invitrogen, Carlsbad, CA), and RO⁺ and RA⁺ subsets were identified by FACS. Human T cells were cultured in RPMI 1640 (Lonza, Walkersville, MD) and supplemented with 10% FBS. Jurkat T cells were grown in modified RPMI 1640 medium from the American Type Culture Collection (Manassas, VA) supplemented with 10% FBS (HyClone, Logan, UT).

A validated KDR small interfering RNA (siRNA) (sense-r [CGC UGA CAU GUA CGG UCU A] dTdT antisense-r [UAG ACC GUA CAU GUC AGC G] dTdT) and control (AllStars negative control siRNA, number 1027280) were purchased from Qiagen (Valencia, CA) and were transfected (100 nM) into Jurkat T cells using Lipofectamine 2000 (Invitrogen). After 96 h, transfected cells were treated with VEGF (5–20 ng/ml) for 10 and 15 min and harvested. The efficiency of siRNA for knockdown was assessed by PCR and/or Western blot analysis in control cells.

Protein samples were lysed with ice-cold radioimmunoprecipitation assay buffer (Boston Bioproducts, Ashland, MA) were separated on a SDS-polyacrylamide gel and transferred to a polyvinylidene difluoride membrane (Millipore, Bedford, MA). Using standard methodology, proteins were detected in each membrane by chemiluminescence (Pierce, Rockford, IL).

Total RNA was prepared using the RNeasy isolation kit (Qiagen). cDNA synthesis and PCR were performed using the SuperScript one-step RT-PCR kit (Invitrogen) and gene-specific primers according to the manufacturer’s protocol. The oligonucleotide primers used were as follows: human KDR, forward, 5′-AAA GAC TAC GTT GGA GCA ATC CCT-3′, and reverse, 5′-CTG GAT TGT GTA CAC TCT GTC AAA-3′; human Flt-1, forward, 5′-ATG GCT CCC GAA TCT ATC TTT GAC-3′, and reverse, 5′-GCC CCG ACT CCT TAC TTT TAC TGG-3′; and human GAPDH, forward, 5′-ACC ACA GTC CAT GCC ATC AC-3′, and reverse, 5′-TCC ACC ACC CTG TTG CTG TA-3′. Quantitative real-time PCR was performed using the 7300 real-time PCR system and the Assays-on-Demand Gene Expression Product (TaqMan, MGC probes; Applied Biosystems, Foster City, CA). Gene-specific primers for the analysis of human KDR and GAPDH by real-time PCR were obtained from Applied Biosystems. Ct values for the evaluation of KDR expression were calculated.

Cell suspensions were stained with FITC- or PE-conjugated Abs, or isotype controls using standard techniques. Stained cells were subsequently analyzed using a FACSCalibur flow cytometer (BD Biosciences, Mountain View, CA), CellQuest, and FlowJo software.

Briefly, multiscreen-IP microtiter plates (Millipore, Bedford, MA) were coated with anti-human IFN-γ (BD Pharmingen) and anti-CD3 (0.1–1 μg/ml). CD45RO⁺ memory CD4⁺ T cells (5 × 10⁴ cells/well) were subsequently cultured in the plates alone, or with VEGF (10–20 ng/ml), or with VEGF and SU5416 (1 and 5 μM) for 24 h The cells were removed by washing and ELISPOT was performed using standard methodology (3). Negative controls were unstimulated cells, and positive controls were PHA (1 μg/ml)-treated cells.

Live-time chemotaxis assays were performed across 3-μm pore membrane Falcon FluoroBlok Transwell inserts (BD Biosciences, Franklin Lakes, NJ) in 24-well culture plates, similar to that described by others (18). Cell culture medium (180 μl) was added into the lower chamber in the absence or presence of VEGF (10–50 ng/ml). Addition of the T cell chemoattractant IP-10 (50 ng/ml) into the lower chamber was used as a positive control. Negatively isolated CD4⁺CD45RO T cells were labeled with 5 μM CFSE (Molecular Probes, Carlsbad, CA) and were added (1 × 10⁵ labeled cells/well in duplicate) into the upper chamber of each Transwell in a total volume of 180 μl of cell culture medium. Lymphocyte migration was monitored in the lower chamber of the Transwell by the assessment of increasing fluorescence every 15 min using an automated plate reader (Victor; Wallac, Turku, Finland).

Chemotaxis assays were also performed across type 1 collagen coated 3-μm pore polycarbonate filters using the standard Boyden Chamber assay, according to the manufacturer’s instructions (Neuro Probe, Gaithersburg, MD). Negatively isolated CD4⁺CD45RO T cells (1 × 10⁵ cells/well) were added into the upper wells, in duplicate, in a total volume of 50 μl of cell culture medium. VEGF and IP-10 were added into the lower wells (also in 50 μl of medium), and the chamber was incubated for 3 h at 37°C in 5% CO₂. Cells were removed from the upper surface of the filters, and they were fixed in methanol and stained with Wright–Giemsa stain (Richard-Allan Scientific,Waltham, MA). The number of transmigrated stained T cells on the undersurface of the filter was counted by light microscopy in a mean of six fields (at ×400 magnification)/condition.

We initially evaluated the expression of VEGFRs on human CD4⁺ T cell subsets and a Jurkat T cell line. By FACS analysis, we found that NRP-1, KDR, and Flt-1 were expressed at low levels on our Jurkat T cells (Fig. 1A), but in multiple experiments, we failed to find notable expression of any VEGFR on unactivated pooled populations of human CD4⁺ T cells or on naive CD45RA⁺CD4⁺ human T cell subsets (data not shown). In contrast, although NRP-1 was at low negligible levels, we found that the expression of KDR and Flt-1 was notable at the protein level by FACS and by Western blot analysis on purified CD45RO⁺ populations of CD4⁺ T cells (Fig. 1B–D). Both KDR and Flt-1 were also expressed at the mRNA level in these T cells (Fig. 1E). However, in multiple experiments, we found that there were variations in baseline levels of expression of KDR on CD4⁺CD45RO⁺ T cells, and in occasional donors, KDR was expressed at high levels comparable to that observed in endothelial cells (data not shown). We used quantitative real-time PCR to evaluate whether induced T cell expression of KDR was related to the activation status of the CD4⁺ T cells. As illustrated in Fig. 1F, we found a 10-to 20-fold increase in KDR mRNA expression following stimulation of pooled populations of CD4⁺ T cells with anti-CD3. Thus, VEGFRs are expressed on recently activated and memory subsets of human T cells.

VEGF is well established to activate protective, proliferative, and migratory signaling in endothelial cells via the MAPK and the PI3K-Akt pathways (11, 19). We initially cultured our Jurkat T cell line with increasing concentrations of VEGF (0–20 ng/ml), and we subsequently evaluated the activation of these classical VEGF-inducible signaling pathways by Western blot analysis. We found that the treatment of Jurkats with VEGF (for time periods from 2–15 min) resulted in the phosphorylation of ERK and Akt (Fig. 2 and data not shown).

Because KDR is well established to function in VEGF-induced activation of ERK and Akt in endothelial cells (19), we next transfected our Jurkat T cells with siRNA to KDR (or control siRNA) prior to treatment with VEGF. We found that VEGF increased the expression of pERK (Fig. 2A) and pAkt (Fig. 2B) in control siRNA-transfected cells, and the response was significantly reduced following transfection with KDR siRNA (p < 0.05, n = 3 experiments; Fig. 2A, 2B). We also cultured purified populations of CD4⁺CD45RO⁺ T cells with VEGF in the absence or presence of SU5416, a pharmacological KDR signaling inhibitor. As illustrated in Fig. 2C–E, we found that the treatment of these cells with VEGF also resulted in activation/phosphorylation of ERK and Akt, and furthermore, we found that the response was reduced in the presence of SU5416 (Fig. 2D, 2E).

TCR-mediated signaling also activates the MAPK and PI3K-Akt pathways. To determine whether there is cross talk between VEGF- and TCR-induced responses, we next cultured our Jurkat T cells with anti-CD3 in the absence or presence of VEGF, and we subsequently evaluated the expression of pERK by Western blot analysis. As expected, we found that TCR-mediated signals markedly induced the expression of pERK (Fig. 2F). Moreover, we observed that the treatment of Jurkats with both VEGF and anti-CD3 in combination resulted in an additive effect on pERK expression. Collectively, these findings define a potential role for VEGF in the costimulation and/or the reactivation of human memory subsets of CD4⁺ T cells.

To next determine whether VEGF elicits functional reactivation responses, purified populations of CD4⁺CD45RO⁺ T cells were stimulated with increasing concentrations of VEGF alone or in combination with anti-CD3. As illustrated in Fig. 3A–C, although we found differences in baseline IFN-γ responses to anti-CD3 in different donors, VEGF consistently enhanced anti-CD3–stimulated production of IFN-γ (p < 0.001, n > 10 experiments). VEGF also increased IL-2 production, but its effect was not as marked as that observed for IFN-γ, and the addition of VEGF to anti-CD3–treated cultures failed to costimulate IL-4 production in memory T cell subsets (data not shown).

Because we find that VEGF-KDR interactions mediate cell signaling responses in our T cells (Fig. 2), we also wished to test whether they costimulate IFN-γ production. CD4⁺CD45RO⁺ T cells were cultured with anti-CD3 and VEGF in combination, in the absence or presence of SU5416 (Fig. 3C). We observed that SU5416 inhibited VEGF-inducible IFN-γ production to baseline levels, indicating that VEGF–KDR interactions play a major role in the costimulation of cytokine production.

To evaluate whether VEGF is functional for T cell chemotaxis, CFSE-labeled CD4⁺CD45RO⁺ T cells were placed in the upper chamber of FluoroBlok Transwells, and the response to VEGF was assessed live-time every 15 min, for up to 90 min using an automated assay. As illustrated in Fig. 4A, we found that the migratory response to VEGF was significantly higher than controls at all time points examined. Using the Boyden chamber chemotaxis assay, we also found a marked migratory response to VEGF (Fig. 4B, p < 0.05) suggesting that VEGF mediates motility response(s) in this T cell subset. To test whether VEGF-inducible activation of the Akt and/or the ERK pathways function in the chemotaxis response, we next cultured T cells with LY294002 (to inhibit Akt signaling) or PD98059 (to inhibit ERK signaling) prior to, and during, the chemotaxis assay. We used pertussis toxin, a well-established G protein-coupled receptor signaling inhibitor as a control. As illustrated in Fig. 4B, we found that VEGF elicited a migratory response in the cells treated with pertussis toxin. However, the cells treated with either LY294002 or PD98059 failed to migrate in response to VEGF. As expected, there was a marked chemotactic response to IP-10, which was significantly reduced in the cells treated with pertussis toxin (Fig. 4C). Collectively, these observations indicate that VEGF functions to elicit migratory responses in human CD4⁺CD45RO⁺ subsets of T cells, and furthermore, that this effect involves VEGF-inducible activation of the Akt and ERK signaling pathways.

VEGFRs are expressed on select subpopulations of T cells, indicating potential role(s) for VEGF in the regulation of immune responses. In this study, we demonstrate that KDR and Flt-1 are expressed at both the mRNA and protein levels on CD45RO⁺ memory subsets of human CD4⁺ T cells. However, similar to others (12), we failed to find detectable levels of the VEGFR NRP-1 on effector/memory CD45RO⁺ T cells. NRP-1 is reported to be expressed on murine CD4⁺CD25⁺FoxP3⁺ T regulatory cells and on human subpopulations of naive T cells (17). These observations suggest that NRP-1 may be functional in naive and T regulatory responses. Our observations in this report indicate that KDR and Flt-1 but not NRP-1 are functional in VEGF-mediated activation of CD45RO⁺ T cells.

We observed that VEGF mediates the activation of both the ERK and Akt signaling pathways in human T cells. Interestingly, it has been demonstrated that there is an intricate cross talk between these pathways to regulate the outcome of the signaling response (20). Akt activation may negatively regulate ERK signaling, which may serve, for instance, to maintain a physiological balance in terms of proliferation and/or survival. Thus, it is possible that the effect of VEGF to mediate ERK activity may be limited due to simultaneous Akt activation.

Our findings, as well as those by others (13) indicate that VEGF has direct chemoattractant effects on T cells. Furthermore, we find that VEGF-inducible activation of ERK and Akt signaling within T cells mediate this migratory response. Because VEGF is expressed at high levels in many chronic inflammatory disease processes (1, 2), the implications of our observations are that targeting of T cell VEGF or KDR, or its signaling partners, may be therapeutic to inhibit the trafficking and/or the peripheral localization of human memory T cells. On the other hand, augmenting KDR-induced signaling in T cells could have implications, for instance, in the directed migration of T cell subsets into tumors.

Collectively, our studies defined in this report indicate that VEGF has direct effects on human memory T cells and may qualitatively and quantitatively regulate migratory and reactivation responses. Thus, a novel implication of our studies is that VEGF, a classical endothelial cell mitogen and growth factor, has potent effects on the inflammatory response.

We thank Debabrata Mukhopadhyay, Ph.D., for ongoing constructive comments and Olivier Dormond, M.D., Ph.D., for helpful discussions about the techniques used in this report.

Disclosures The authors have no financial conflicts of interest.