T cell receptor internalization from the immunological synapse is mediated by TC21 and RhoG GTPase-dependent phagocytosis

Abstract

The immunological synapse (IS) serves a dual role for sustained T cell receptor (TCR) signaling and for TCR downregulation. TC21 (Rras2) is a RRas subfamily GTPase that constitutively associates with the TCR and is implicated in tonic TCR signaling by activating phosphatidylinositol 3-kinase. In this study, we demonstrate that TC21 both cotranslocates with the TCR to the IS and is necessary for TCR internalization from the IS through a mechanism dependent on RhoG, a small GTPase previously associated with phagocytosis. Indeed, we found that the TCR triggers T cells to phagocytose 1-6 μm beads through a TC21- and RhoG-dependent pathway. We further show that TC21 and RhoG are necessary for the TCR-promoted uptake of major histocompatibility complex (MHC) from antigen-presenting cells. Therefore, TC21 and RhoG dependence underlie the existence of a common phagocytic mechanism that drives TCR internalization from the IS together with its peptide-MHC ligand.

Figure 1

The TCR is internalized from the IS by a TC21-dependent process. (A) TC21 is co-internalized with the TCR from the IS. Selected frames of images taken by time-lapse videomicroscopy of Jurkat cells transfected with CD3ζ-yellow fluorescent protein (YFP) and TC21-cyan fluorescent protein (CFP), and stimulated with SEE-loaded Raji APCs. The positions of two sequentially contacted APCs are indicated by circles in the merged images. For clarity, TC21-CFP emission is shown in red and CD3ζ-YFP in green. (B) Dominant negative and constitutively active forms of TC21 block TCR internalization from the IS. Jurkat cells were transfected with CD3ζ-YFP and the indicated TC21 mutants (S28N: dominant negative; G23V: constitutively active). Arrows indicate the accumulation of the TC21 mutants and CD3ζ at the IS, while arrowheads indicate the existence of an intracellular pool of CD3ζ-YFP before stimulation (upper panel), and the emergence of a small pool of endocytic vesicles that contain CD3ζ but not TC21 (lower panel). The bright field (BF) image in the lower panel shows the formation of extensive membrane protrusions by the T cell at the T cell:APC contact area (red dots). (C) Quantification of TC21 mutant effect on CD3ζ internalization from the IS. A number of 15-20 single T:B cell conjugates were scored according to the presence of CD3ζ in vesicles immediately below the IS. Each value represents the mean and standard deviation of three datasets.**P < 0.005 (two-tailed Mann-Whitney test).

Figure 2

The TC21-dependent mechanism of TCR internalization is RhoG-dependent but independent of clathrin. (A) TCR and TC21 double positive endocytic vesicles are negative for markers of clathrin-dependent endocytosis. Jurkat cells transfected with CD3ζ-Cherry and TC21-GFP were stimulated with SEE-loaded Raji for 30 min and were either incubated with Alexa633-labeled transferrin during the last 5 minutes (upper panels) before fixation or stained with anti-clathrin heavy chain (CHC; lower panels). Arrows indicate the position of the IS and arrowheads internal vesicles double positive for TC21 and CD3ζ. The images are representative of 50 cells examined per condition. (B) The cSMAC is devoid of coated pits. Primary naïve mouse T cells were plated on patterned arrays containing a central anti-CD3 spot surrounded by ICAM-1. Cells were unroofed and prepared for freeze-etch electron microscopy. A typical pattern with actin rings surrounding a central bare spot was seen (micrograph a, scale bar-1 μm). More vigorous sonication removed the actin ring revealing large numbers of clathrin cages in the pSMAC (yellow arrows, micrograph c, scale bar-100 nm) but not in the cSMAC (micrograph b, scale bar-100 nm). Quantification of the number of clathrin cages in the pSMAC vs. cSMAC areas was performed from a series of 17 photographs corresponding to the pSMAC and 3 photographs to the cSMAC and expressed in d as number of cages per square micrometer. ***P < 0.0005 (two-tailed Mann-Whitney test). Scale bars (C) TCR is internalized from the IS in RhoG-positive vesicles. Jurkat cells were transfected with both wild type RhoG-GFP and wild type CD3ζ-Cherry, and they were stimulated with SEE-loaded Raji for the times indicated. The Raji APCs are encircled in the merged images. The arrow indicates the accumulation of TCR at the IS, while the arrowhead indicate the presence of IS-derived vesicles containing TCR that are positive for RhoG. ICA of RhoG and CD3ζ across the section indicated with a white line at times 0 and 2 min in the merged images are shown at the bottom to illustrate how RhoG concentrates in CD3ζ-positive vesicles. (D) Both the dominant negative and constitutively active RhoG mutants block TCR internalization from the IS. Jurkat cells transfected with CD3ζ-Cherry and either dominant negative (T17N) or constitutively active (Q61L) mutants of RhoG-GFP were stimulated for the times indicated. The arrows indicate the presence of TC21 at the IS. (E) Both dominant negative and constitutively active RhoG mutants block TC21 internalization from the IS. Jurkat cells transfected with wild type TC21-DsRed, and either the T17N or the Q61L mutants of RhoG-GFP, were stimulated for the times indicated. The presence of TC21 at the IS is indicated by arrows, while the presence of RhoG-positive, TC21-negative internal vesicles is indicated by arrowheads. (F) Quantification of the effect of RhoG mutants on CD3ζ and TC21 internalization from the IS. 15–20 single T:B cell conjugates were scored according to the presence of either CD3ζ or RhoG in vesicles immediately below the IS. Each value represents the mean and standard deviation of three datasets. *P < 0.05, and **P < 0.005 (two-tailed Mann-Whitney test).

Figure 3

TCR-triggered phagocytosis. (A) T cells phagocytose anti-CD3-coated latex beads. Jurkat and primary mouse naïve T cells were incubated with 1–6 μm diameter fluorescent beads (shown in green) for 30 min and stained with fluorescent phalloidin to visualize the cortical F-actin (in red). Optical sections along the xy, yz and xz axis are shown to demonstrate that some beads are completely internalized. A secondary Alexa 647-labeled anti-mouse Ig (shown in blue) was added to determine if the 3 μm and 6 μm antibody-coated beads were accessible or not. This distinguished fully internalized (arrowheads) from external beads, and even detected phagocytotic cups (arrows) in which the bead is partly inaccessible to the secondary antibody. (B) T cells form a tight seal in the phagocytotic cup. Naïve mouse cells stimulated for 30 min with anti-CD3 coated 6 μm beads were fixed, permeabilized and stained with phalloidin-TRITC (shown in green), To-Pro to visualize the nucleus (n, orange), and with an Alexa 488-labeled secondary anti-hamster Ig antibody (shown in blue) to distinguish internal from non-phagocytosed beads. Arrows indicate a tight apposition of the actin cytoskeleton to the bead at the growing edges of the phagocytotic cup. Arrowheads indicate the part of the bead facing the cell that is not accessible to the anti-Ig antibody and that therefore, is red but not blue. (C) Phagocytosed beads are associated with both the TCR and polymerized actin. Human primary T cell lymphoblasts were incubated with anti-CD3-coated 1 μm beads for 60 min and stained with with phalloidin and an anti-CD3ζ antibody to visualize the endogenous TCR. Arrowheads indicate the presence of polymerized actin around the phagocytosed beads. (D) Quantitative assay of bead phagocytosis. Jurkat cells were incubated at 37°C with red fluorescent 1 μm beads coated with anti-CD3 for the times indicated and subsequently, with a secondary Alexa 488 anti-Ig antibody at 0°C. (E) TCR-triggered phagocytosis is dependent on the actin cytoskeleton. Jurkat cells were incubated for 30 min with either 1, 3 or 6 μm anti-CD3-coated beads in the presence or absence of 1 μg/ml cytochalasin D (CytD). The percentage of phagocytosis was calculated according to the in/out ratio as shown in panel D. Each value represents the mean and standard deviation of three datasets. (F) Particle size and PI3K-dependence. Jurkat T cells pre-incubated with the indicated inhibitors were incubated with 1–6 μm anti-CD3-coated beads for 30 min and phagocytosis was assessed as in panel D. Each value represents the mean and standard deviation of three datasets. (G) Partially and totally phagocytosed beads are tightly surrounded by membranes. Jurkat cells were incubated with 1 μm anti-CD3-coated beads for 5 (a), 15 (b), 30 (c) or 60 min (d), and they were prepared for transmission electron microscopy. Arrows indicate the presence of a phagocytic cup in micrograph a, and the presence of membranes around phagocytosed beads in b. Arrowheads indicate the presence of membrane vesicles in what appears a multivesicular body in c and of particles associated to egressing beads in d. Scale bars: 500 nm (a), 200 nm (b), 1 μm (c and d)

Figure 4

TCR-triggered phagocytosis requires TC21 and RhoG activities. (A) Phagocytosed beads are associated to the TCR, TC21 and RhoG. Jurkat cells were transfected with TC21-GFP or RhoG-GFP and they were incubated for 60 min with fluorescent 1 μm anti-CD3 coated beads. Untransfected Jurkat cells were incubated with beads in the same conditions and the endogenous TCR was visualized with a CD3ζ antibody. (B) Phagocytosed beads are associated with endogenous proteins. Untransfected Jurkat cells were incubated with 3 μm anti-CD3 coated green beads for 30 min and stained with either anti-TC21 or anti-RhoG. (C) TCR-triggered phagocytosis is TC21- and RhoG-dependent. Jurkat cells were transfected with the constructs indicated, incubated with fluorescent anti-CD3-coated beads and phagocytosis was analyzed as in Fig. 3D. WT, cells transfected with wild type TC21-GFP; vector, cells transfected with the empty vector. (D) Dose dependency of the effect of mutant TC21 and RhoG. The efficiency of phagocytosis is shown as a function of the amount of GFP fusion protein. (E) TCR-triggered phagocytosis is impaired in T cells genetically deficient in TC21 or RhoG. Lymph node T cells of the indicated genotypes that were stimulated for 60 min with anti-CD3-coated 1μm red fluorescent latex beads were examined by confocal microscopy after staining with phalloidin (green). In parallel, phagocytosis was quantified by flow cytometry (as in Fig. 3D) at the times indicated (bottom). Each value represents the mean and standard deviation of three datasets. *P < 0.05, and **P < 0.005 (two-tailed Mann-Whitney test).

Figure 5

The TCR is internalized from the IS by a TC21- and RhoG-dependent mechanism. (A) The TCR is retained at the IS in the absence of TC21 or RhoG. Lymph node T cells from AND TCR transgenic mice with the indicated genotypes were transduced with a CD3ζ-GFP lentivirus and stimulated with MCC peptide antigen-loaded DCEK cells as APCs. Formation of the IS and TCR internalization from the IS was followed by time-lapse videomicroscopy according to the distribution of CD3ζ-GFP. Arrows indicate the accumulation of CD3ζ-GFP at the IS. The WT T cell makes two ISs, first with APC b and later with APC a. The emergence of internal vesicles derived from both ISs is indicated with yellow arrowheads. (B) Quantification of TCR internalization from the IS. For quantification, AND T cells were stained on ice with the anti-TCRβ antibody H57-597 (H57) before they were stimulated for 30 min with MCC-loaded DCEK. The formation of an IS was quantified according to the percentage of T:APC cell conjugates with CD3ζ concentrated at the synapse. The internalization of the TCR from the IS was quantified according to the percentage of T cells with internal vesicles positive for both the anti-CD3ζ and H57 antibodies. Arrows indicate the position of the IS and arrowheads the presence of endocytic vesicles containing TCR that have been internalized from the IS. A total of 20–30 conjugates per genotype were studied in triplicate (right panel). *P < 0.05 (two-tailed Mann-Whitney test); n.s, not significant. (C&D) TC21 and RhoG deficiencies have a weak effect on TCR downregulation. AND T cells of the indicated genotype were stimulated for 3 hr with DCEK cells loaded with the indicated concentrations of MCC peptide (C) or for the times indicated with DCEK loaded with 100 μM MCC (D). TCR downregulation was calculated according to the mean fluorescence intensity of the AND TCR with an anti-Vβ3 antibody. Each value represents the mean and standard deviation of three datasets. *P < 0.05, and **P < 0.005 (two-tailed Mann-Whitney test).

Figure 6

TC21 and RhoG mediate trogocytosis and the uptake of MHC class II by T cells. (A) TC21 and RhoG mediate TCR-induced trogocytosis of APC membrane fragments. AND lymph node T cells of the indicated genotypes were transduced with CD3ζ-GFP and stimulated with MCC-loaded DCEK cells previously labeled with the red fluorescent dye PKH26. Conjugate formation was recorded by time-lapse videomicroscopy. The arrows indicate the formation of an IS and the arrowhead the presence of an endocytotic vesicle containing fragments of the red-labeled APC membrane. The inset shows a magnification of the T cell:APC contact area of the wild type T cell 8.3 min after stimulation (scale bar-1μm). Quantification of PKH26 uptake by the T cells (right) was performed by flow cytometry after gating of CD4⁺ T cells. Each value represents the mean and standard deviation of three datasets. *P < 0.05 (two-tailed Mann-Whitney test). (B) Trogoctyosis of APC membrane fragments is PI3K-dependent. Wild type AND lymph node T cells were stimulated with MCC-loaded (+) or unloaded (-) DCEK cells previously labeled with the red fluorescent dye PKH26 in the presence of the indicated concentrations of either wortmanin or LY294002 PI3K inhibitors. APC membrane acquisition by the T cell was quantified by flow cytometry according to the acquisition of PKH26 label by the T cells. CytD indicates cells stimulated in the presence of 1 μg/ml cytochalasin D. (C) T cells take up MHC class II from APCs from the IS. DCEK cells were transiently transfected with the I-Eα^k-GFP construct and incubated with unlabeled purified naïve wild type AND T cells and recorded by time-lapse videomicroscopy. Arrowheads indicate the presence of vesicles in the T cell containing I-Eα^k-GFP that have been acquired from the IS. (D) T cells take up MHC class II from APCs by a TC21 and RhoG-dependent mechanism. DCEK cells were transiently transfected with the I-Eα^k-GFP construct and incubated with AND T cells of the indicated genotype. The formation of IS was visualized after staining with anti-CD3ζ. Yellow arrowheads show IS-derived vesicles that contain both the TCR (CD3ζ) and I-Eα^k-GFP acquired from the APC. Arrows indicate the accumulation of CD3ζ at the IS of Rras2^-/- and Rhog^-/- mice. In the RhoG-deficient T cell, I-Eα^k-GFP also accumulates at the IS but it is not trogocytosed by the T cell. Quantification of I-Eα^k-GFP uptake by the T cells (right) was performed by flow cytometry after gating on CD4⁺ T cells. Each value represents the mean and standard deviation of three datasets. *P < 0.05 (two-tailed Mann-Whitney test).

Figure 7

RhoG is activated by the TCR via a TC21 and PI3K-dependent pathway. (A) TCR triggering with immobilized anti-CD3 antibody activates RhoG. Jurkat cells were stimulated for 2 min with anti-CD3-coated 1 m beads (40:1 bead/cell ratio). Pull-down (Pd) with GST-ELMO was performed on cell lysates that were then analyzed in western blots (WB) probed with anti-RhoG to estimate RhoG activity. WB probed with anti-phosphoERK and total ERK antibodies were performed as controls of activation and loading, respectively. (B) Antigen stimulated activation of RhoG is TC21-dependent. AND T cells of the different genotypes were incubated for the times indicated with DCEK APCs preloaded with 10 μM MCC. Pull-down with GST-ELMO was used to isolate active RhoG. WB probed with anti-RhoG from whole cell lysates were used as a loading control. Quantification was carried out by densitometry and dividing the intensity of the bands in the GST-ELMO pull-down by those of the whole cell lysates. Comparison of the resulting values was performed by dividing them by this of the WT cells at time 0, that was considered as 1. (C&D) TCR-triggered phosphorylation of Akt is TC21-dependent. AND T cells of the indicated genotypes were incubated with DCEK APCs preloaded with 10 μM MCC (C) or stimulated with anti-CD3 coated 1 μm beads (D) for the times indicated. WB were probed with anti-phosphoAkt antibody of whole cell lysates as an indication of PI3K activity. (E) RhoG activation but not TC21 activation is PI3K dependent. Wild type T cell lymphoblasts were stimulated with anti-CD3 coated 1 μm beads for 5 min in the presence or absence of 50 μM LY294002 PI3K inhibitor. Pull-down with GST-ELMO was used to isolate active RhoG and pull-down with GST-Raf1(RBD) to isolate active TC21. Whole cell lysates were used as loading controls.