Integrins Modulate T Cell Receptor Signaling by Constraining Actin Flow at the Immunological Synapse

Abstract

Full T cell activation requires coordination of signals from multiple receptor-ligand pairs that interact in parallel at a specialized cell-cell contact site termed the immunological synapse (IS). Signaling at the IS is intimately associated with actin dynamics; T cell receptor (TCR) engagement induces centripetal flow of the T cell actin network, which in turn enhances the function of ligand-bound integrins by promoting conformational change. Here, we have investigated the effects of integrin engagement on actin flow, and on associated signaling events downstream of the TCR. We show that integrin engagement significantly decelerates centripetal flow of the actin network. In primary CD4⁺ T cells, engagement of either LFA-1 or VLA-4 by their respective ligands ICAM-1 and VCAM-1 slows actin flow. Slowing is greatest when T cells interact with low mobility integrin ligands, supporting a predominately drag-based mechanism. Using integrin ligands presented on patterned surfaces, we demonstrate that the effects of localized integrin engagement are distributed across the actin network, and that focal adhesion proteins, such as talin, vinculin, and paxillin, are recruited to sites of integrin engagement. Further analysis shows that talin and vinculin are interdependent upon one another for recruitment, and that ongoing actin flow is required. Suppression of vinculin or talin partially relieves integrin-dependent slowing of actin flow, indicating that these proteins serve as molecular clutches that couple engaged integrins to the dynamic actin network. Finally, we found that integrin-dependent slowing of actin flow is associated with reduction in tyrosine phosphorylation downstream of the TCR, and that this modulation of TCR signaling depends on expression of talin and vinculin. More generally, we found that integrin-dependent effects on actin retrograde flow were strongly correlated with effects on TCR signaling. Taken together, these studies support a model in which ligand-bound integrins engage the actin cytoskeletal network via talin and vinculin, and tune TCR signaling events by modulating actin dynamics at the IS.

Keywords: T cell; actin; costimulation; immunological synapse; integrin; signaling; talin; vinculin.

Figure 1

Integrin engagement slows actin flow. Jurkat T lymphoma cells and primary human CD4⁺ T cell blasts were allowed to interact with coverslips coated with anti-CD3 (OKT3) plus ICAM-1, VCAM-1, or both, and imaged for 4 min. (A) Still images of responding Jurkat cells stimulated on coverslips coated with anti-CD3 alone (left) or in the presence of VCAM-1 (right), together with corresponding kymographs of F-actin dynamics generated along the yellow lines. The white lines in the kymographs show example slopes used to calculate actin flow rates. (B) Kymographic analysis of F-actin features in Jurkat cells, showing the distribution of F-actin velocity across the immunological synapse (IS). The marked area displays the peripheral lamellipodial region (LP). (C,D) Surface expression of integrin LFA-1, as detected by CD11a staining and VLA-4, as detected by CD29 staining on Jurkat T cells (pink) and primary human CD4⁺ T cell blasts (blue). (E,F) Actin flow rates in the LP region for Jurkat cells (E) and primary human CD4⁺ T cell blasts (F). Error bars show mean ± SD. *p < 0.05, ***p < 0.001. Scale bar = 10 µm.

Figure 2

Integrin–ligand mobility influences actin flow. Jurkat T cells (A) and primary human CD4⁺ T cell blasts (B) were allowed to interact with mobile bilayers functionalized with anti-CD3 (OKT3) alone, together with ICAM-1, VCAM-1, or both. Actin flow rates in the LP region were analyzed as described for Figure 1. Solid dots show flow rates from cells responding to mobile ligand (M). Open dots show data for cells responding to immobile ligands (I). Data for immobile surfaces are from Figure 1, plotted here to facilitate comparison. Error bars show mean ± SD. **p < 0.01, ***p < 0.001.

Figure 3

The effects of integrin engagement are distributed across the actin network. Jurkat T cells were allowed to interact with mixed-mobility surfaces bearing patterns of immobilized VCAM-1, surrounded by planar bilayers coated with anti-CD3 and imaged for 4 min. Pattern (A) has 1 µm diameter spots with center-to-center distance of 3 µm. Patterns (B,C) have 2 µm diameter spots with center-to-center distances of 4 and 5 µm, respectively. (D) Actin flow rates were assessed within the lamellipodial region, choosing areas that overlie the immobile spots (on pattern, yellow dashed lines) or bypass them completely (off pattern, blue dashed lines). The red dashed line marks the actin flow on bilayers coated with anti-CD3. (E) Jurkat T cells were allowed to interact with surfaces patterned with bars of immobilized VCAM-1 (1 µm wide, 5 µm center-to-center), surrounded by planar bilayers coated with anti-CD3 and imaged for 4 min. (F) Actin flow rates were assessed within the lamellipodial region, choosing areas overlying the immobile bars (on pattern, yellow dashed lines), bypass them completely (off pattern, blue dashed lines) or where the free edge of the lamellipodium lies between bars (outside pattern, white dashed line). The red dashed line marks actin flow for cells spread on bilayers coated with anti-CD3. (G) Measurements of actin velocity were made parallel to the bars in the space between two bars. Mean ± SD is calculated for measurements made in 20–40 cells for each condition, *p < 0.05, ***p < 0.001. Scale bars = 10 µm.

Figure 4

Integrin engagement modulates signaling downstream of the T cell receptor. (A) Jurkat T cells were stimulated on coverslips coated with 10 µg/ml OKT3 alone or together with VCAM-1 at varying concentrations (0–4 µg/ml). After 5 min, cells were fixed and stained for phospho-tyrosine (green) and phospho-ZAP70 (Y319, red). (B) Quantitative analysis of phospho-tyrosine labeling intensity. (C) Quantitative analysis of phospho-ZAP70 labeling intensity. (D–F) Jurkat T cells were stimulated on coverslips coated with 0.4, 2, or 10 µg/ml OKT3, alone or together with 2 µg/ml VCAM-1. After 5 min, cells were fixed and stained for phospho-tyrosine (D), phospho-ZAP70 Y319 (E) or phospho-PLCγ1 Y783 (F). Quantitative analysis of labeling intensity is shown below each image panel. (G) Primary human T cells were stimulated for 5 min on coverslips coated with 10 µg/ml OKT3 alone or together with ICAM-1 or VCAM-1 (each at 4 µg/ml). Cells were fixed and labeled for phospho-tyrosine (green) and phospho-ZAP70 Y319 (red). (H,I) Quantitative analysis of labeling intensity from (G). Mean ± SD for 50–60 cells per condition is shown. (J) Jurkat T cells loaded with Fura-2 were stimulated on coverslips coated with 1 or 3 µg/ml OKT3, alone or together with 2 µg/ml VCAM-1, and Ca²⁺ responses were monitored by ratiometric imaging. Lines represent averages of 14–23 cells per condition, after alignment of the traces based on the earliest detectable signal over baseline. Lines are artificially extended (before time 0) to better show the starting baseline intensities. Data from one representative experiment (of three) is shown. *p < 0.05; **p < 0.01.***p < 0.001. Scale bars = 10 µm.

Figure 5

Focal adhesion proteins are recruited to sites of integrin engagement. (A) Primary human T cells were stimulated on coverslips patterned with ICAM-1 surrounded with OKT3 for 15 min, and labeled with m24 to detect the active, extended-open conformation of LFA-1, and Kim127 to detect the extended form of LFA-1. (B,C) Jurkat T cells were stimulated on VCAM-1 patterns surrounded with anti-CD3 for 15 min. Cells were then fixed and labeled with phalloidin to detect F-actin, and with antibodies to talin, vinculin, and paxillin. (D,E) Primary human T cells were allowed to interact with VCAM-1 (D) or ICAM-1 (E) patterns surrounded with OKT3 for 17 min. Cells were then fixed and labeled with phalloidin to detect F-actin, and with antibodies to talin and vinculin. Far right panels in (B–E) show cropped regions in which the four channels have been merged. (E) Scale bars = 10 µm.

Figure 6

Focal adhesion-like complexes can be detected in T cells spreading on uniform fields of integrin ligand. Primary human T cells were stimulated on coverslips coated with OKT3 and VCAM-1 (A) or ICAM-1 (B,C) for 17 min, then fixed and stained for talin and paxillin. The inset in (B) shows focal adhesion-like structures in a different cell that was stimulated on OKT3 + ICAM-1 for 25 min. Scale bar = 10 µm; inset in B, 8 µm².

Figure 7

Vinculin and talin couple integrins to actin. (A) Jurkat T cells were stably transfected with shRNA vectors targeting talin (shT), vinculin (shV), and α-actinin 4 (shA4). Untransfected cells (UT) and cells expressing the empty shRNA vector (EV) were used as controls. Expression of each protein was tested by Western blotting with the indicated antibodies; the antibody against α-actinin recognizes both the 1 and 4 isoforms. Panels are regions from one blot, adjusted individually to enhance brightness and contrast, using only linear tools. The complete blot is available in Figure S2 in Supplementary Material. (B) Cells generated as in (A) were stimulated on coverslips coated with OKT3 and VCAM-1, and actin flow rates in the LP region were quantified. Mean ± SD were calculated from measurements made in 10–40 cells for each condition, ***p < 0.001 compared with EV control. (C–E) Untransfected (C), talin-suppressed (D), and vinculin-suppressed (E) Jurkat T cells were stimulated on coverslips patterned with VCAM-1 surrounded with OKT3 for 15 min, fixed, and labeled as indicated. (F) Untransfected Jurkat T cells were treated with Y-27632 and Jasplakinolide to arrest actin flow, and stimulated on coverslips patterned with VCAM-1 surrounded with OKT3. Scale bar = 10 µm.

Figure 8

Clutch molecules modulate T cell receptor (TCR) signaling. Jurkat T cells, untransduced (UT), or suppressed for talin (shT) or vinculin (shV) were stimulated on coverslips coated with OKT3 (10 µg/ml) in the presence and absence of VCAM-1 (2 µg/ml). After 5 min, cells were fixed and stained for phospho-tyrosine (A) or phospho-ZAP70 (Y319). (B) Quantitative analysis of labeling intensity is shown below each image panel. Mean ± SD of measurements from at least 50 cells is shown. *p < 0.05; ***p < 0.001. Scale bars = 10 µm. (C) Normalized phospho-tyrosine labeling intensity from (A) was plotted against actin retrograde flow rates from Figure 7B. Linear regression analysis showed vinculin-suppressed cells responding to OKT3 alone to be an outlier. Excluding that data point resulted in a strong correlation R² = 0.90. Dotted line shows the 95% confidence interval. Data represent means ± SEM from three independent experiments.