Role of the EHD Family of Endocytic Recycling Regulators for TCR Recycling and T Cell Function

Understanding the mechanisms that underlie proper T cell function is of great interest in basic immunology and immunopathology. T cell activation requires recognition of an antigenic peptide bound to MHC proteins on APCs by the TCR and downstream signaling mediated by cytoplasmic regions of the TCR-CD3 complex. Effective generation of an immune response requires concurrent engagement of and signaling through costimulatory proteins (e.g., CD28) on T cells by their cognate ligands on the APC surface (1). Therefore, the mechanisms that ensure optimal levels of the TCR and accessory receptors on the T cell surface prior to Ag stimulation are fundamentally critical to generating an effective immune response. However, mechanisms that regulate the surface pool of TCR and its accessory receptors have been primarily investigated in the context of T cell activation. During T cell activation, TCR signaling elicited by the APC-presented Ag, together with costimulatory signals, leads to spatial reorganization of the TCR and accessory receptors, such as CD28 and lymphocyte function–associated 1 (LFA-1), to form an area of intimate contact with APCs, the immunological synapse (IS) (2, 3). The endocytic pathways supply the IS with newly synthesized and recycled receptors to replenish those that are targeted to the lysosome for degradation (4–6).

Cells use clathrin-dependent as well as clathrin-independent pathways to internalize surface receptors into early endosome/sorting endosome compartments, from where they traffic to late endosome/multivesicular bodies and lysosomes for degradation, or recycle back to the cell surface either through a Rab4⁺ fast recycling endosome (directly to the plasma membrane) or through a slow recycling route via the Rab11⁺ recycling endosome (7, 8). This sorting processes is known to be regulated by the Rab family of small GTPases and their interacting partners (9). TCRs such as TCR-CD3 (1), CD28 (10), and LFA-1 (11, 12) are known to be internalized and recycled to the IS during T cell activation. Internalized TCR-CD3 traverses the Rab5⁺ early endosome, followed by transport through Rab4⁺ and Rab11⁺ endosomes, indicating the involvement of both fast and slow recycling pathways (13, 14). Naive and resting T cells constitutively internalize and recycle their surface receptors, and the balance of these processes dictates their prestimulation surface levels and hence the levels of T cell activation (13). However, the mechanisms of constitutive basal traffic of TCR-CD3 are not fully understood.

Recently, a new family of endocytic regulators, the EPS15 homology domain–containing proteins (EHD1–4), has been identified. These proteins exhibit a conserved domain architecture with helical regions near the N terminus and the middle of the protein surrounding an ATP-binding G-domain that can hydrolyze ATP slowly, and a C-terminal EH domain that binds to proteins with asparagine-proline-phenylalanine–containing or related tripeptide motifs as a major mechanism of protein–protein interactions (15, 16). Structural studies of EHD2 reveal that the helical regions form a curved membrane lipid-binding interface, and that the ATPase domain folds similar to the GTPase domain of dynamin and mediates the formation of stable dimers (17). Other structural findings support a model that dimeric EHDs organize into oligomeric rings around curved membrane structures and function in membrane tubulation and vesiculation (17–21). Consistent with these biochemical observations, cell biological studies have shown that EHD proteins regulate recycling of several surface receptors that traffic through clathrin-dependent or clathrin-independent pathways, including transferrin receptor, MHC class I, β₁ integrin, 2-amino-5-hydroxy-5-methyl-4-isoxazolepropion acid receptors, insulin-like growth factor 1 receptor, insulin-responsive glucose transporter 4, and others (7, 22–29). However, the physiological relevance of the many in vitro–assigned roles of EHD proteins remains to be defined.

We used gene targeting in mice to reveal the in vivo functional roles of EHD proteins and cell surface receptors whose traffic is regulated by EHDs (30–36). For example, EHD1 knockout (KO) mice exhibit strain-dependent phenotypes varying degrees of embryonic lethality, male infertility, ocular developmental defects, and neural tube closure defects due to impairment of ciliogenesis and SHH signaling (37, 38).

To date, any roles of EHD family proteins in TCR traffic or T cell function are unknown. Given their importance in the regulation of a variety of other cell surface receptors and the consequences of deleting their genes, singly or in combination, on organ/cell function in vivo, we hypothesized that EHD proteins could play an important functional role in T cells by regulating receptor traffic. Consistent with such a hypothesis, EHD proteins have been shown to interact with Rab effectors, such as Rabenosyn-5, a dual Rab4 and Rab5 effector in the early endosome (39), and Rab11 effector, Rab11-FIP2, which regulates the exit of vesicles from the recycling endosome back to the plasma membrane (40). Moreover, EHD proteins (EHD1, EHD3, and EHD4) were shown to interact with SNARE complex proteins such as SNAP-25, SNAP29 (41, 42), Syndapin I and II (7, 43, 44), and other proteins in non–T cell systems. EHD proteins have been shown to localize in endosomes positive for Rab5, Rab4, Rab8, Rab11, or Rab35 (9, 25, 29, 45), compartments that are implicated in TCR traffic (46–48). Thus, we posited that EHD proteins could play an important functional role in T cells.

We found that EHD1, EHD3, and EHD4 are expressed in lymphoid tissues and in purified CD4⁺ T cells. Accordingly, we generated myelin oligodendrocyte glycoprotein (MOG)–specific TCR transgene (2D2-TCR Tg)–bearing mice in which floxed EHD1, EHD3, and EHD4 were deleted using a CD4-Cre transgene. In vitro, CD4⁺ T cells of EHD1^fl/fl; EHD3^fl/fl; EHD4^fl/fl; Cd4-cre; 2D2Tcr mice (EHD1/3/4 knockout) exhibit reduced Ag-driven cell proliferation and IL-2 secretion. In vivo, these mice exhibit a reduced severity of experimental autoimmune encephalomyelitis (EAE) elicited by immunization with the MOG peptide. We show that surface levels of TCR, CD28, and LFA-1 are reduced and TCR recycling is impaired in CD4⁺ T cells of mice with a CD4-Cre–mediated EHD1/3/4 knockout. Therefore, our studies reveal a novel role of the EHD family of endocytic recycling proteins in the surface display of TCR in unstimulated T cells that contributes to their positive functional role in Ag-induced T cell activation and immunopathology.

BSA (catalog no. A7906-100G), paraformaldehyde (catalog no. 158127-500G), Triton X-100 (catalog no. 93418), EGTA (catalog no. E8145-50G), sodium orthovanadate (catalog no. S6508-50G), sodium deoxycholate (catalog no. D6750-100G), 4-hydroxtamoxifen (catalog no. T176-10MG), and brefeldin A (catalog no. B7651) were from Sigma-Aldrich (St. Louis, MO). Propidium iodide staining solution (catalog no. 00-6990-42) was from eBioscience. [³H]thymidine (catalog no. 2407001, 2.0 Ci/mmol) was from MP Biomedicals. Bafilomycin A1 (catalog no. BML-CM110-0100) was from Enzo Life Sciences. MOG peptide (aa 35–55) (catalog no. BP001328-PRO-371) was from Syd Labs. OVA peptide (aa 323–336) (sequence: ISQAVHAAHAEINE) was synthesized from Tufts University Core Facility (Medford, MA). CFA (catalog no. 231131) and heat-inactivated Mycobacterium tuberculosis were from Difco Laboratories. NaF (catalog no. S299-500), NaCl (catalog no. S271-10), and Tris (catalog no. BP152-5) were from Fisher Chemicals. Pertussis toxin was from Enzo Life Sciences. ECL development reagent (catalog no. 32106), BSA for bicinchoninic acid assay (catalog no. 23209), and PMSF (catalog no. 36978) were from Thermo Scientific. CFSE (catalog no. C34554), CellTrace Violet stain (catalog no. C34557), RPMI 1640 (catalog no. SH30027.02), penicillin/streptomycin (catalog no. 15140-122), and FBS (catalog no. 10427-028; lot no. 1662765A120-01) were from Life Technologies. Live/Dead fixable blue dead cell stain kit was from Molecular Probes. IL-2 ELISA kit (catalog no. M2000) was from R&D Systems.

Anti-CD28 (catalog no. 122007) and anti–LFA-1 (catalog no. 141011) were from BioLegend. Anti-CD3ε (referred to as anti-CD3 in this study) (catalog nos. 553064 and 553057), anti-CD25 (catalog no. 558642), anti-CD4 (catalog nos. 553051 and 553047), anti-CD8a (catalog no. 560469), anti-CD28 (catalog no. 553294), and anti-CD44 (catalog no. 553135) were from BD Biosciences. Anti-B220 (catalog no. 25-0452-81) and anti-CD3ε (referred to as anti-CD3 in this study) (catalog no. 56-0032-80) were from eBioscience. Anti-HSC70 (catalog no. sc-7298) and anti-LAMP1 1D4B (catalog no. sc-19992) were from Santa Cruz Biotechnology. Anti-EHD1 (catalog no. ab109311) was from Abcam. Anti–β-actin (catalog no. A5316) was from Sigma-Aldrich. Allophycocyanin-conjugated anti–annexin V (catalog no. 17-8007-72) was from eBioscience. Polyclonal rabbit Abs recognizing EHD1 plus EHD4, EHD2, or EHD3 have been described previously (45). Secondary fluorochrome-conjugated Abs were from Life Technologies.

Whole-body knockout mice (Ehd1-null) derived from Ehd1^flox/flox mice have been described previously (49). Ehd1-null mice were maintained on a mixed 129, B6 background. EHD1^fl/fl; EHD3^fl/fl; EHD4^fl/fl mice in a predominantly C57BL/6 background were crossed with B6.Cg-Tg (Cd4-cre) 1Cwi/BfluJ mice to generate Cd4-cre; EHD1^fl/fl; EHD3^fl/fl; EHD4^fl/fl mice. These mice were further crossed with C57BL/6-Tg (Tcra2D2, Tcrb2D2)1Kuch/J, 2D2 Tcr, or MOG Tcr to generate Cd4-cre; 2D2Tcr; EHD1^fl/fl; EHD3^fl/fl; EHD4^fl/fl mice. EHD1^fl/fl; EHD3^fl/fl; EHD4^fl/fl mice in a predominantly C57BL/6 background were also crossed with tamoxifen-inducible CreERT2-expressing mice from The Jackson Laboratory [Gt (ROSA)26Sor^{tm1(cre/ERT2) Tyj}; strain 008463] to generate Cre^ERT2; EHD1^fl/fl; EHD3^fl/fl; EHD4^fl/fl mice. These mice were further crossed with C57BL/6-Tg (Tcra2D2, Tcrb2D2)1Kuch/J, 2D2 Tcr, or MOG Tcr) to generate Cre^ERT2; 2D2Tcr; EHD1^fl/fl; EHD3^fl/fl; EHD4^fl/fl mice. Genotypes were confirmed by subjecting tail clip DNA to PCR analysis using the KAPA mouse genotyping kit (Kapa Biosystems). Mice were treated humanely according to the National Institutes of Health and University of Nebraska Medical Center guidelines. Animal studies were preapproved by the Institutional Animal Care and Use Committee (no. 14-067).

Lymphoid tissues or isolated cells were lysed in ice-cold Triton X-100 lysis buffer (0.5% Triton X-100, 50 mM Tris [pH 7.5], 150 mM NaCl, 1 mM PMSF, 10 mM NaF, 1 mM VO₄) or with RIPA lysis buffer (same as Triton X-100 lysis buffer with an increase of Triton X-100 to 1% and an addition of 5 mM EDTA, 1 mM EGTA, 1% SDS, and 0.5% sodium deoxycholate). Lysates were vortexed, centrifuged at 13,000 rpm for 30 min at 4°C either immediately or after overnight rocking in the cold room, and supernatants were collected. Protein lysates were quantified using the bicinchoninic acid assay. Forty-microgram aliquots of lysate protein per sample were resolved by SDS-PAGE and transferred to polyvinylidene difluoride membranes (from Immobilon-P, catalog no. IPVH00010). In certain experiments, lysates from equal numbers of cells were resolved by SDS-PAGE. The membranes were blocked in TBS/5% BSA, incubated with the appropriate primary Abs diluted in TBS–0.1% Tween 20 for 1 h, washed in TBS–0.1% Tween 20 (three times for 5 min each) followed by a 45-min incubation with HRP-conjugated secondary Ab in the same buffer. The membrane was then washed in TBS–0.1% Tween 20 (three times for 5 min each) and ECL-based detection was performed.

To isolate primary CD4⁺ T cells, a negative selection protocol was performed as described (50) using magnetic beads (biotin binder kit, catalog no. 11533D; Invitrogen) and biotinylated Abs (BioLegend), and purity was established to be 91–95% based on flow cytometry.

T cells were incubated on ice in the dark for 15–30 min (depending on the experiment) with appropriate conjugated Abs at the manufacturer’s recommended dilution in FACS buffer (0.1% BSA in PBS). Cells were pelleted, washed twice, and suspended in 400 μl of cold FACS buffer. In other cases, cells were fixed with 4% cold paraformaldehyde for 15 min at room temperature after staining, then washed and suspended in 400 μl of cold FACS buffer. Cells were protected from light until analyses using either an LSR II Green or LSR II cytometer (BD Biosciences). FACS data were analyzed using DIVA (BD Biosciences FACSDiva software), FlowJo data analysis software (FlowJo, Ashland, OR) and ModFit LT software (Verity Software House, Topsham, ME).

Spleen cells (5 × 10⁶ cells/ml) from CD4-Cre; 2D2Tcr; EHD1^fl/fl; EHD3^fl/fl; EHD4^fl/fl mice and control (2D2Tcr; EHD1^fl/fl; EHD3^fl/fl; EHD4^fl/fl) mice were stained with CFSE or CellTrace Violet according to the manufacturer’s instructions. Cells were treated with 50 μg/ml MOG_35–55 peptide for 72 h. On the indicated day, cells were stained with FITC-CD4 before analysis. Dilution of CFSE or CellTrace Violet fluorescence as an indicator of cell division was assessed via FACS analysis. Data were analyzed using FlowJo and ModFit LT software to delineate the percentage of cells that had undergone increasing number of divisions and to determine the proliferation index. In some instances, the cells were not stained with the proliferation dye, but were stained with anti–annexin V and propidium iodide staining solution after 72 h of stimulation and analyzed by FACS for cell death analysis.

Spleen cells (5 × 10⁵ cells per well) from CD4-Cre; 2D2Tcr; EHD1^fl/fl; EHD3^fl/fl; EHD4^fl/fl or control (2D2Tcr; EHD1^fl/fl; EHD3^fl/fl; EHD4^fl/fl) mice were seeded in 96-well U-bottom plates in 100 μl of medium in the presence of varying concentrations of MOG_35–55 peptide for 72 h. Cells were pulsed with 1 μCi of [³H]thymidine per well for the last 6 h of incubation, harvested onto filter disks, and the radioactivity (counts per minute) was counted using a scintillation counter (Packard). For T cell expansion, spleen cells or lymph node (LN) cells from CD4-Cre; 2D2Tcr; EHD1^fl/fl; EHD3^fl/fl; EHD4^fl/fl mice and control (2D2Tcr; EHD1^fl/fl; EHD3^fl/fl; EHD4^fl/fl) mice were stimulated with 10–20 μg/ml MOG _35–55 peptide and T cells were expanded in the presence of 30 U/ml IL-2 for 7–8 more days (fresh IL-2–containing media every 2 d after 3 d of stimulation).

Spleen cells or LN cells from EHD1^fl/fl; EHD3^fl/fl; EHD4^fl/fl; Cre^ERT2; 2D2Tcr mice were prestimulated with either the MOG peptide or with anti-CD3 and anti-CD28 for 3–4 d in the presence (+) or absence (−) of 4-hydroxytamoxifen (4-OHT) (200–300 nM) for deletion of EHD1, EHD3, and EHD4. Cells were washed in PBS, stained with CFSE, and restimulated for proliferation as previously described.

Isolated spleen cells (5 × 10⁵ cells per well) were cultured in 96-well U-bottom plates in the presence of 10 μg/ml of MOG_35–55 peptide at 37°C for indicated time points. IL-2 secretion was measured in culture supernatants using an ELISA kit according to the manufacturer’s instructions.

Internalization of the cell surface pool of TCR-CD3 was assessed as previously described (13, 51) with modifications. Briefly, freshly isolated LN cells at 5 × 10⁶ cells/ml in 96-well U-bottom plates (in triplicates) were incubated in the presence of 10 μg/ml brefeldin A at 37°C for the indicated times to allow the surface TCR to internalize in the absence of newly synthesized TCR transport to the cell surface. At each time point, cells were transferred to ice and subsequently stained with PE-conjugated anti-CD3 followed by FACS analysis to quantify the surface TCR levels remaining at each time point. The extent of the TCR remaining at the cell surface was calculated by expressing the mean fluorescence intensity (MFI) of PE–anti-CD3 staining at each time point relative to time 0, which was set as 100%.

The recycling of pre-existing intracellular pools of TCR-CD3 was carried out by adapting a previously described protocol (52). Briefly, the accessibility of the cell surface CD3 on freshly isolated LN cells (at 5 × 10⁶ cells/ml in 96-well U-bottom plates in triplicates) was blocked by staining with a predetermined saturating concentration of unconjugated anti-CD3 at 4°C for 30 min. The cells were extensively washed and incubated at 37°C. At the indicated times, the cells were transferred to 4°C and stained with a PE-conjugated version of the same anti-CD3 Ab that was used to block the staining of the initial cell surface cohort of CD3. The MFI of PE–anti-CD3 cell surface FACS staining was used as a measure of the recycling of intracellular TCR-CD3. Recycling of the cell surface TCR-CD3 pool following internalization was assayed by a modified protocol (53, 54). Freshly isolated LN cells were first stained with PE-conjugated anti-CD3 for 30 min on ice. After extensive washes, the cells were incubated at 37°C for 30 min (in complete RPMI) to allow the Ab-labeled TCR-CD3 to undergo internalization. The cells were then subjected to an acid wash to strip any remaining Ab bound to CD3 on T cell surface. The cells were washed in cold FACS buffer and resuspended in complete RPMI followed by incubation at 37°C for the indicated times to allow recycling of the labeled internalized pool of TCR-CD3 back to the cell surface. The cells were spun down, acid washed, and resuspended in FACS buffer for FACS analysis to determine the remaining intracellular fraction of the internalized TCR-CD3. The percentage of TCR-CD3 recycling was calculated from the equation [(T_{0 −} T_x)/T₀] × 100, where T₀ is the MFI of cells incubated for 0 min before the second acid wash, and T_x is the MFI of cells subjected to a second acid wash after incubation for the indicated times.

LN T cells were expanded as described under T cell expansion. A total of 5 × 10⁶ expanded T cells in complete RPMI medium were plated onto poly-l-lysine–coated coverslips placed in 24-well plates and allowed to attach at 37°C for 10 min. Cells on coverslips were fixed in 4% ice-cold paraformaldehyde for 20 min and then incubated with 0.1 M glycine (in PBS) for 3 min at room temperature. After washing, the cells were incubated with the blocking plus permeabilization buffer (2% BSA plus 0.1% Triton X-100 in PBS) for 30 min at room temperature. Cells were stained with anti–Lamp-1 and anti-CD3 primary Abs overnight a 4°C. After three washes, the cells were stained with secondary anti-rat (for Lamp-1) and anti-Armenian hamster Abs for 45 min at room temperature. After washes, the cells were mounted with Vectashield HardSet mounting medium with DAPI (catalog no. H-1500; Vector Laboratories). Images were acquired at room temperature using a Zeiss ELYRA S.1 or a superresolution structured illumination microscopy. The sCMOS camera mounted on side port was used [objective lenses, Plan-Apochromat ×63/1.40 oil differential interference contrast; for resolution: lateral resolution (XY), 120 nm; axial resolution (Z), 300 nm (typical experimental full width a half maximum values with objective lens Plan-Apochromat ×63/1.40 oil differential interference contrast, subresolution beads of 40 nm diameter and excitation at 488 nm)]. Merged fluorescence images were generated and analyzed using ZEN 2012 software from Carl Zeiss.

Six to eight-week-old female CD4-Cre; 2D2Tcr; EHD1^fl/fl; EHD3^fl/fl; EHD4^fl/fl or control (2D2Tcr; EHD1^fl/fl; EHD3^fl/fl; EHD4^fl/fl) mice were s.c. immunized in the flank with 200 μg of MOG_35–55 peptide emulsified (1:1) in CFA containing 0.5 mg of heat-killed Mycobacterium tuberculosis H37RA in a total of 200 μl of emulsion per mouse, together with an i.p. injection of pertussis toxin (200 ng per mouse) in 200 μl of PBS followed by a second dose of pertussis toxin 2 d later. Mice were monitored daily and scored for clinical signs of EAE on a scale of 0–5 as follows: 0, no disease; 1, decrease tail tone; 2, hindlimb weakness or partial paralysis; 3, complete hindlimb paralysis; 4, forelimb and hindlimb paralysis; and 5, moribund state or dead (55). Composite scores of each cohort were used to determine statistical significance.

An unpaired Student t test was used to calculate the p values. Data are presented as mean ± SEM, and a p value <0.05 served as the threshold for statistical significance.

Our previous Western blotting studies of EHD protein expression in the mouse thymus and spleen (49, 56, 57), which we extended here to LN (Fig. 1A), showed that multiple EHD family members are expressed, albeit unequally, in lymphoid organs. This suggested that multiple EHD family members are likely to be expressed in individual immune cell types, consistent with our recent findings that bone marrow–derived macrophages express both EHD1 and EHD4 (36). With the focus of this study on functional analyses in the context of CD4⁺ T cells, our additional analyses showed that EHD1, EHD3, and EHD4 are expressed in primary murine CD4⁺ T cells (Fig. 1B) whereas EHD2 protein was undetectable to barely detectable (Fig. 1C). Analyses of mRNA expression databases (ImmGen and BioGPS) further supported this expression pattern (Supplemental Fig. 1). Notably, Western blotting of purified LN T cells at various time points of anti-CD3/CD28 stimulation showed that levels of EHD proteins vary as a function of CD4⁺ T cell activation, with a time-dependent increase in EHD1 and EHD4 levels and a reduction in EHD3 levels (Fig. 1B). These findings suggested that EHD proteins may play important roles in T cell function.

Our previous knockout studies have revealed relatively unique as well as redundant functional roles of EHD proteins, with the knockout of EHD1 alone producing by far the most obvious developmental and adult organ functional aberrations, depending on the strain background (27, 37, 38, 49). Therefore, we assessed whether EHD1 deficiency had any impact on T cell activation. Notably, the initial analysis of EHD1^−/− splenic T cells did not reveal any detectable abnormality in anti-CD3/CD28–induced cell proliferation (Fig. 2A). These findings supported a redundant functional role of EHD proteins, at least in the context of T cell proliferation. Consistent with this, a slight increase in EHD4 levels was noted in LN from EHD1^−/− mice (Fig. 2A). Initial analysis of EHD3^−/− CD4⁺ T cells showed no difference in anti-CD3/CD28–induced proliferation compared with CD4⁺ T cells from control mice (Fig. 2B). We observed that deletion of EHD3 did not cause an increase in the level of either EHD1 or EHD4 in LN cells (Fig. 2B). However, EHD4^−/− CD4⁺ T cells showed reduced proliferation compared with control CD4⁺ T cells, although the decrease was not statistically significant (Fig. 2C). We also noted that the deletion of EHD4 did not cause an increase in the level of either EHD1 or EHD3 in LN cells. Given the lack of impact or a mild impact of single EHD gene knockouts on T cell proliferation and the expression of multiple family members, we reasoned that EHD1, EHD3, and EHD4 might function redundantly in T cells, and we therefore used a combinatorial knockout strategy to further explore their role in T cells.

Given the likelihood of redundant EHD function, we generated mice with a CD4-Cre–dependent, T cell–directed, conditional deletion of floxed Ehd1, Ehd3, and Ehd4 genes. To examine the impact of EHD1/3/4 deletions in T cells in the context of Ag-specific responses, we also incorporated the MOG peptide-specific 2D2-Tcr transgene (55) into our conditional knockout model (referred to as CD4-Cre⁺ mice). Littermates without a CD4-Cre transgene were used as controls (referred to as CD4-Cre⁻ mice). The presence of floxed EHD1/3/4 gene alleles targeted for CD4-Cre–dependent deletion was detected by tail PCR using specific primers for each floxed allele (Supplemental Fig. 2A). Western blot analysis of CD4⁺ T cells isolated from the 2D2-Tcr transgenic mice showed that the presence of the transgenic TCR by itself did not have any significant impact on the levels of EHD proteins (Supplemental Fig. 2B). CD4-Cre⁺ mice were developmentally indistinguishable from CD4-Cre⁻ mice, and no gross phenotypic changes were seen in adult mice. Western blot analysis confirmed the CD4-dependent deletion of EHD1/3/4 (Fig. 3A).

Next, to examine whether EHD1/3/4 deletion impacted T cell development, we isolated single cells from lymphoid organs of 4-wk-old mice and analyzed these for changes in T cell numbers using cell counting combined with FACS. We found that total numbers of thymocytes, splenocytes, and LN cells were comparable in mice from both groups (Fig. 3B). Similarly, comparable cell numbers were observed when spleens of older (6–8 wk old) mice of the two genotypes were analyzed (Fig. 3C). We carried out FACS analyses on the live/B220⁻ T cell population and assessed the proportion of cells at various developmental stages based on staining with Abs against CD4, CD8, CD25, and CD44 (58–60). Single-positive thymocytes in 2D2-Tcr Tg mice are known to be skewed toward the CD4⁺ T cell compartment; this is also seen in spleen and LN (55). CD4-Cre–mediated deletion of EHD1/3/4 did not affect the relative percentages or total cell numbers within various thymic developmental stages when analyzed in 4-wk-old mice (Fig. 3D, 3E).

In contrast, CD4-Cre–mediated EHD1/3/4 deletion led to a decrease in the percentage of T cells in the periphery. This phenotype was seen in both the LN and spleen (Fig. 3F, 3G), although only the splenocyte reduction was statistically significant. However, the absolute number of CD4⁺ T cells in these organs did not differ significantly between the CD4-Cre–positive and control mice. Additionally, an apparent increase in peripheral B cell numbers and their percentages were seen (Fig. 3F, 3G); the basis of this phenotype is unknown but is likely to reflect altered T cell/B cell interactions due to T cell deficiency of EHD1/3/4. We also saw a decrease in the percentage of CD8⁺ T cells in spleens of CD4-Cre–positive mice, but this did not reflect an increase in CD8⁺ T cell numbers in these organs. From these results, we conclude that expression of EHD proteins is dispensable during thymic T cell development, but may play a role in peripheral T cell function. To exclude the possibility that the lack of impact on T cell development and changes in peripheral lymphocyte numbers might reflect the presence of an autoreactive TCR transgene, we also analyzed the non–TCR-transgenic mice for the effect of CD4-Cre–mediated EHD1/3/4 deficiency during thymic development and in spleen cells. We observed a lack of impact of T cell development similar to that seen in the TCR-transgenic mice (Fig. 4A–C), further supporting the conclusion that EHD proteins are dispensable for thymic T cell development beyond the CD4/8⁺ stage. The analysis of spleen cells showed a slight increase in the percentage of B220⁺ cells (B cells) and a slight decrease in the percentages of CD4⁺ and CD8⁺ T cells, but these differences were not statistically significant (Fig. 4D).

Incorporation of an MOG Ag-specific T cell transgene in our EHD1/3/4 deletion models allowed analyses of CD4⁺ T cell response to a specific Ag, both in vitro and in vivo, with the latter in the context of the development of EAE. We asked whether the mature T cells in these mice exhibit any functional deficits. Initial experiments using spleen cells from the parental 2D2-TCR transgenic and control mice with MOG or OVA peptides established the optimal concentrations of MOG peptide to induce T cell proliferation (data not shown). Analysis of CD4⁺ T cells from spleen using the CellTrace Violet stain and [³H]thymidine incorporation showed a significant decrease in proliferation of T cells from CD4-Cre⁺ mice compared with control CD4-Cre⁻ mice (Fig. 5A, 5B). Visualization of progressive cell divisions using ModFit LT software showed that whereas CD4-Cre⁺ and CD4-Cre⁻ CD4⁺ T cells progress through a comparable number of divisions, EHD1/3/4-null T cells exhibit a significant underrepresentation at late divisions and an increased proportion of cells at earlier divisions (Fig. 5A). The EHD1/3/4-null T cells also exhibited a significantly lower overall proliferation index (Fig. 5A), further supporting the conclusion that these cells proliferate slower than the control T cells. To exclude the possibility that the apparent decrease in the number of EHD1/3/4-null T cells at late stages of proliferation might be due to their apoptosis, we stimulated the CD4⁺ T cells as previously described and stained these with propidium iodide and anti–annexin V followed by FACS analysis. We observed no increase in the percentage of apoptotic cells in Ag-stimulated CD4⁺ T cells isolated from CD4-Cre⁺ compared with CD4-Cre⁻ (control) mice (Fig. 5C). Thus, the mature peripheral CD4⁺ T cells of CD4-Cre⁻ mice exhibit a marked defect in the specific Ag-induced cell proliferation in vitro.

To assess whether the impact of EHD1/3/4 protein deficiency on CD4⁺ T cell proliferation reflects their abnormal development or an actual importance in the events associated with T cell activation, we used T cells from 2D2-TCR Tg–bearing mice carrying a tamoxifen-inducible ERT-Cre that allows in vitro deletion of EHD1/3/4. EHD1/3/4 deletions in vitro led to a significant decrease in Ag-induced proliferation compared with their control T cells (Fig. 5D–F). These results support a T cell–intrinsic positive role of EHD proteins in sustaining Ag-driven T cell responses in mature peripheral T cells.

Among the cytokines that initiate Ag-activated T cells into proliferation is the activation-induced secretion of IL-2 and the induced expression of the corresponding receptor on the T cell surface (61, 62). Therefore, we assessed the impact of EHD1/3/4 deletions on IL-2 secretion. We observed a significantly lower level of Ag-induced IL-2 secretion by EHD1/3/4-null CD4⁺ T cells (Fig. 5G) with a difference detectable at the earliest time point (12 h) analyzed, making it unlikely that the reduction is due to a smaller pool of T cells caused by reduced proliferation. The significant reduction in IL-2 secretion likely represents one mechanism for the defective T cell proliferation upon deletion of EHD1/3/4.

The impact of EHD1/3/4 deficiencies was not restricted to in vitro T cell responses alone. Using immunization with the MOG_35–55, we could test the impact of EHD1/3/4 deficiencies on the development of EAE in vivo. We immunized cohorts of control (CD4-Cre⁻) versus experimental (CD4-Cre⁺) mice with MOG_35–55 and assessed onset and severity of EAE over time. Interestingly, although no significant difference in the disease onset was seen, the EHD1/3/4-null group showed a decrease in the disease severity compared with control littermates (Fig. 5H). These results further support our conclusion that the expression of EHD proteins is important for full CD4⁺ T cell responses to Ag. Collectively, these in vitro and in vivo analyses established, to our knowledge for the first time, that EHD proteins play an important positive role in Ag-driven T cell activation and immune response.

At a cellular level, EHD proteins have been characterized as regulators of the endocytic traffic of different cell surface receptors as mentioned in the 1Introduction. Endocytic traffic of these receptors is also key to their targeting to the IS (1, 10–13, 63–66). To assess whether EHD proteins may indeed regulate the traffic of TCR-CD3 or the accessory receptors, we first examined the cell surface levels of TCR-CD3, CD28, CD4, CD25, CTLA-4, and LFA-1 by FACS analysis of unstimulated T cells (Fig. 6A). As assessed by comparing the MFI values, T cells from CD4-Cre⁺ mice expressed significantly lower levels of TCR-CD3 on the cell surface compared with T cells from control mice (Fig. 6A, 6C). These analyses also showed that in vitro–activated CD4⁺ T cells from CD4-Cre⁺ mice express lower levels, although not statistically significant, of TCR-CD3, CD28, CD4, CD25, CTLA-4, and LFA-1 compared with those from control mice (Fig. 6B, 6C).

Because the TCR-CD3 complex is the primary determinant of Ag-specific T cell activation and subsequent responses, we focused on the TCR for further analyses. We examined the kinetics of TCR-CD3 internalization and recycling in unstimulated CD4⁺ T cells. Analysis of TCR-CD3 internalization using FACS showed that there was not a significant difference in internalization of TCR in EHD1/3/4-null T cells compared with control T cells (Fig. 7A). Next, we analyzed recycling to the cell surface of the pre-existing intracellular pool of TCR-CD3. For this purpose, anti-CD3 Ab recognition epitope on the pre-existing surface TCR-CD3 was blocked with an unconjugated Ab, and the appearance of intracellular TCR-CD3 on the cell surface at various time points was monitored by staining with a conjugated anti-CD3. These analyses revealed that EHD1/3/4-null CD4⁺ T cells exhibit a significantly reduced recycling of pre-existing intracellular TCR-CD3 to the cell surface at all time points analyzed compared with control T cells (Fig. 7B). Next, we assessed the recycling of the pre-existing cell surface pool of TCR-CD3 by prelabeling it with a PE-conjugated anti-CD3 and allowing it to internalize. After acid wash (to strip any remaining Ab bound to TCR-CD3 on the T cell surface), the internalized PE-labeled TCR-CD3 was allowed to reappear at the cell surface and restripped to remove the labeled Ab. Even under these assay conditions, we observed a significant decrease in the recycling of TCR-CD3 in EHD1/3/4-null CD4⁺ T cells compared with control cells (Fig. 7C). Overall, these results show a key requirement of EHD1, EHD3, and EHD4 in basal recycling of TCR-CD3 from intracellular endocytic recycling compartments back to the cell surface.

Given the impairment of the endocytic recycling (Fig. 7B, 7C) together with the reduced steady-state surface TCR-CD3 levels in freshly isolated EHD1/3/4-null T cells (Fig. 6), we considered the possibility that the pool of TCR-CD3 that is retained in intracellular compartments may be aberrantly targeted for lysosomal degradation. To address this, EHD1/3/4-null T cells (from CD4-Cre–mediated and 4OHT-induced ERT-Cre deletion) were examined by Western blot for total levels of CD3ε. In both cases, we found that deletion of EHD proteins led to a significant decrease in the total levels of CD3ε (Fig. 8). As in unstimulated T cells, we also observed a reduction in the total level of CD3ε in activated CD4⁺ T cells (Fig. 8C). Given our findings with the CSF-1 receptor in macrophages (36), we hypothesized that this reduction was likely due to lysosomal targeting and degradation. To test this hypothesis, we treated the EHD1/3/4-null T cells with or without Bafilomycin-A1, an inhibitor of lysosomal degradation, for 4 h and analyzed the impact on total CD3ε levels using Western blots and confocal microscopy (Fig. 9). Western blots of cells treated with Bafilomycin-A1 for 4 h showed a recovery of the CD3ε protein level in EHD1/3/4-null T cells with levels comparable to those in control cells (in treated cells) (Fig. 9A, 9B). We performed confocal imaging of control and EHD1/3/4-null T cells incubated with or without Bafilomycin-A1. Whereas relatively low CD3ε staining was colocalized with the lysosomal marker LAMP-1 in untreated cells (Fig. 9D), some colocalization was seen in Bafilomycin A1-treated control T cells (Fig. 9C, upper). Notably, significantly more lysosome-localized CD3ε was seen in EHD1/3/4-null cells treated with bafilomycin-A1 (Fig. 9C, lower, 9E). These results support the conclusion that expression of EHD1/3/4 is required for efficient transport of TCR-CD3 from endosomal compartments back to the cell surface and to prevent its aberrant traffic into lysosomes.

Cellular studies have identified members of the EHD protein family as key new regulatory elements of endocytic trafficking of several cell surface receptors, but their physiological functions remain relatively unexplored and their roles in the immune system are unknown. In this study, we investigate the role of EHD proteins in the context of Ag-specific CD4⁺ T cells in vitro and in vivo. Our genetic studies, ablating genes encoding three members of the EHD protein family (EHD1, EHD3, and EHD4) by CD4⁺ T cells in the context of MOG transgenic TCR, demonstrate, to our knowledge for the first time, an important, positive role of EHD proteins in Ag-specific T cell activation in vitro and an autoimmune response in vivo. Mechanistically, we show that EHD proteins are required for basal recycling of TCR-CD3 from intracellular endocytic pools to the cell surface, thereby determining the subsequent level of T cell activation. Combined with our recent study demonstrating a key positive role of EHD1 in the transport of newly synthesized CSF-1R from the Golgi to the cell surface of bone marrow–derived macrophages (36), and previous cell biological findings that EHD1 facilitates MHC class I recycling from the endocytic recycling compartment to the cell surface in a cell line system (25), our results strongly support the likelihood that EHD proteins play critical and potentially diverse roles in the immune system.

Analysis of thymic T cell development did not reveal any significant alterations in total cellularity, T cell subsets, or relative proportions of T cells at various developmental stages in mice with a CD4-Cre–mediated deletion of EHD1/3/4, indicating that EHD proteins are largely dispensable for thymic development beyond the CD4/8 double-positive stage. Whether this might reflect any role of the remaining EHD family member, EHD2, remains a possibility that will require future studies. It is also possible that EHD proteins are important at earlier stages of thymic development, and future use of other Cre elements active at these stages will be needed to address this possibility.

In contrast to the relatively unperturbed thymic development, CD4-Cre–mediated EHD1/3/4 deletion led to a decrease in the percentage of T cells in the periphery. This phenotype was seen in secondary lymphoid tissues, including the LN and the spleen, but was only significant in the spleen. These results suggested that EHD proteins may play a more important role in mature peripheral T cells. The precise basis for this phenotype remains unclear, but lower functional responses to TCR engagement, as seen in ex vivo analyses (discussed below), could play a role because TCR signals are important in peripheral T cell maintenance (67–69). Additionally, an apparent increase in the peripheral B cell numbers and their percentages were seen in the EHD1/3/4-null mice in the context of 2D2 transgene; the basis of this phenotype is unknown but is likely to reflect altered T cell/B cell interactions due to T cell deficiency of EHD1/3/4 genes. There was also a decrease in the percentage of CD8⁺ T cells in the periphery, but this was only significant in the spleens of transgenic mice. Further studies are needed to precisely assess the basis of these effects.

The in vitro analyses of T cells from the CD4-Cre deletion model demonstrated significantly defective Ag-elicited CD4⁺ T cell proliferation and IL-2 cytokine secretion responses upon EHD1/3/4 deletion. Similar results were observed using CD4⁺ T cells derived from an alternate model where in vitro deletion of floxed EHD1, EHD3, and EHD4 genes was induced with 4-OHT to bypass any in vivo developmental abnormalities. These results support a T cell–intrinsic positive role of EHD proteins in sustaining Ag-driven T cell responses in mature peripheral T cells, a role independent of any potential involvement of EHD proteins in events associated with thymic T cell development.

Coordination of multiple EAE-inducing T cell functions, including their early activation, expansion, migration to the brain across the blood–brain barrier, and effector functions to mediate demyelination in the CNS are required for full disease induction (70, 71). Therefore, we analyzed the impact of EHD1/3/4 deficiencies on the onset and progression of EAE. Because in vitro studies showed the T cell proliferation and cytokine (IL-2) secretion to be impaired upon EHD1/3/4 deficiencies, we expected to see a delayed onset and reduced severity of EAE. Although we saw no difference in the onset of disease, EHD1/3/4-null mice showed a decrease in the disease severity compared with control littermates. Therefore, further in vitro and in vivo studies using models developed in the present study should help to comprehensively reveal other aspects of T cell function that also rely on EHD protein expression.

Given the reduced TCR-CD3 levels in the CD4⁺ T cells of our EHD1/3/4-null mouse model prior to deliberate T cell activation, our focus in the present study has been on the role of EHD proteins in basal (prior to antigenic stimulation) TCR-CD3 traffic, which has not been mechanistically dissected in much detail. Therefore, further studies to identify partner proteins through which EHDs regulate TCR traffic will be of great interest. Notably, endocytic trafficking pathways play key roles during T cell activation by orchestrating the polarized transport of TCR-CD3 and signaling molecules such as Lck to maintain an active immunologic synapse, and Rab4⁻ and Rab11⁺ endosomes are implicated in these processes (72).

Cell biological studies have linked EHD proteins to the regulation of traffic in Rab11⁺ recycling endosomes as well as compartments regulated by other Rabs. Thus, further studies to examine the potential importance of EHD proteins in regulating endocytic traffic into and out of the IS during T cell activation will be of great interest. In this regard, we and others have recently identified a key role of EHD proteins in primary cilia biogenesis (37, 42), and recent studies have shown that the maintenance of IS in cilia-less T cells involves ciliogenesis-related proteins such as IFT20, which localize in the IS and are required for recycling of TCR-CD3 (46, 47, 73). Although IFT20 KO under CD4-Cre led to only minor defects in T cell development and collagen-induced arthritis (an experimental model of rheumatoid arthritis), earlier deletion using Lck-Cre led to defective thymic development and T cell function in vivo (74); the latter are partially similar to the phenotype of our EHD1/3/4 KO mice. Interestingly, the requirement of EHD1 and EHD3 were found to be important for the recruitment of transition zone proteins and IFT20 during ciliary vesicle formation (42). Alhough our study has been done mainly on freshly isolated and rested cells, it seems plausible that EHD proteins may also regulate the traffic of TCR-CD3 in the IS to promote T cell activation and subsequent responses.

Notably, the CD4-Cre–mediated Wiskott–Aldrich syndrome protein and SCAR homolog (WASH) deletion led to T cell defects strongly resembling the phenotype of EHD1/3/4 deficiency with minimal thymic developmental defects, reduced proliferation and cytokine production, a decrease in the surface TCR-CD3 levels, an increase in lysosomal degradation of TCR-CD3, and reduced severity of EAE (75). WASH is an Arp2/3 activator that localizes on distinct endosomal subdomains and functions in endosomal trafficking (76). Given these phenotypic similarities of the two KO models and the interaction of the WASH protein with retromer complex and the involvement of EHD proteins in retromer-dependent transport (77, 78), further exploration of a link between EHD proteins and the WASH complex in regulating TCR-CD3 traffic will be of great interest.

Overall, the present studies using genetic models and functional analyses in vivo and in vitro reveal an important albeit redundant positive role of the newly described EHD family of endocytic recycling regulatory proteins in T cell function. Mechanistically, we show that EHD proteins are required to facilitate basal recycling of TCR-CD3 from intracellular endocytic compartments back to the cell surface to ensure the surface expression of optimal TCR-CD3 levels for subsequent Ag-driven T cell activation. Future studies using the unique models described in the present study should add to our understanding of the importance of endocytic traffic in the control of TCR-CD3 and other receptors that regulate T cell functions as well as the roles of these proteins in other cells in the innate and adaptive arms of the immune system.

We thank the members of the Band Laboratories for helpful discussions and Dr. Vijay Kuchroo (Brigham and Women’s Hospital, Harvard Medical School) for advice on the use of 2D2 mice.

The authors have no financial conflicts of interest.