Filamin A Is Required for Optimal T Cell Integrin-Mediated Force Transmission, Flow Adhesion, and T Cell Trafficking

T cells are subjected to a range of different mechanical forces, as they spend part of their lifetime in the bloodstream subjected to shear stress and part of it inside lymphatic or peripheral tissues where they interact mechanically with other cells, such as APCs. Mechanotransduction encompasses the translation of a mechanical signal into a biochemical signal to change cell behavior, and it is becoming increasingly apparent that mechanotransduction is of fundamental importance in many biological systems, including immune cells (1).

Extrinsic mechanical signals can be sensed by cells through surface receptors, such as integrin adhesion receptors, and because they act as bidirectional signaling molecules, integrins also transduce mechanical signals from the actin cytoskeleton to the extracellular matrix (2). β₂ integrins are leukocyte-specific adhesion molecules expressed on the surface of T cells and other immune cells. They are essential for T cell trafficking out from the bloodstream and into tissues in homeostasis and inflammation because they mediate the firm adhesion to ICAMs expressed on endothelial cells under shear force conditions (blood flow). The β₂ integrin LFA-1 (αLβ₂) is also a component of the immunological synapse between a T cell and an APC, aiding in the stabilization of the contact, and it is required for optimal T cell activation in vitro and in vivo (3, 4).

To perform their functions in vivo, integrins need to be switched on and off at appropriate times (e.g., following TCR activation or chemokine receptor activation). In their active state, integrins are able to bind their ligands, such as ICAMs, on the surfaces of other cells. After binding ligands, β₂ integrins can also participate in signaling to change cell behavior (e.g., to mediate cell spreading, actin reorganization, or cell migration). Integrin activation and signaling is regulated by the binding of cytoplasmic proteins, such as talin, filamin, kindlin-3, and 14-3-3 proteins, to the β integrin cytoplasmic domain (5–11).

Filamin A (FlnA) is an important mechanosensitive linker that binds to integrins and couples them to the actin cytoskeleton. In addition to actin and integrins, FlnA also interacts with many other cytoplasmic partners and can therefore translate mechanical signals into biochemical signals, thereby acting as a mechanosensor in cells (12). Both extrinsic and cell-intrinsic forces (e.g., actomyosin contraction) acting on filamin can lead to changes in the interaction of FlnA with its various cytoplasmic partners, providing a mechanism by which FlnA can act as a cellular mechanosensor (12, 13). In contrast, FlnA has also been reported to function as an integrin inhibitor by competing with talin and/or by linking integrin α- and β-chains together (5, 8, 14). FlnA has also been reported to play a role in T cell activation through interactions with CD28 and PKC-θ (15). However, the in vivo roles of this integrin interactor in the immune system remain poorly understood.

In this study, we have investigated the role of FlnA in T cell functions in vivo. Using a T cell–specific FlnA knockout (KO) mouse model, we show that FlnA does not function as an integrin inhibitor in T cells. Instead, it is required for β₂ integrin–dependent traction forces in T cells as well as for formation of strong integrin–ligand bonds and therefore for optimal T cell adhesion to integrin ligands under shear flow conditions. Consequently, FlnA KO T cell homing into peripheral lymph nodes (LNs) and trafficking into sites of inflammation is reduced in vivo. In conclusion, our results show that FlnA is not an integrin inhibitor in T cells, but it instead plays fundamental roles in integrin-mediated force transmission and T cell trafficking in vivo.

FlnA-floxed mice (16) (The Jackson Laboratory, stock no. 010907) have a conditional KO allele of the filamin gene. When bred to mice expressing Cre recombinase in CD4⁺ T cells, the floxed exons are deleted in CD4 T cells expressing Cre in the offspring. CD4-Cre⁺ (Ctrl) mice were used as controls for FlnA KO–CD4-Cre mice. For in vivo T cell activation studies, FlnA KO mice were crossed with TCR–OT-II–transgenic mice (4). FlnA and CD4-Cre mice were ordered from The Jackson Laboratory. All experiments were in compliance with Social and Health Services of the State Provincial Office of Southern Finland. C57/Bl6 mice were obtained from Charles River.

CD4 T cells were isolated from the spleen and LNs by positive selection using magnetic beads to CD4 (MACS; Miltenyi Biotec, Germany). To culture effector T (T_eff) cells, CD4-isolated T cells were cultured for 48 h with 0.5 μg/ml anti-CD3 (clone 2C11; R&D Systems, Minneapolis, MN) and 20 ng/ml IL-2 (Novartis) in RPMI 1640 medium supplemented with 10% FBS, 2 mM l-glutamine, 100 U/ml penicillin–streptomycin, and 50 μM 2-ME. Cells were washed and maintained in IL-2 for a further 5 d. CD8 T cells were purified from the CD4 fraction from the spleen and LN cells and were isolated with CD8 MACS beads (Miltenyi Biotec). Neutrophils were purified by using a MACS mouse neutrophil purification kit (Miltenyi Biotec) and B cells by negative selection using CD43 (Ly 48) beads from Miltenyi Biotec. Dendritic cells (DCs) were cultured from bone marrow–derived cells and cultured for 10 d with GM-CSF (10 ng/ml; PeproTech) as described previously (17).

CD4 T_eff cells were fractionated as in (18). Cell pellets were lysed on ice for 15 min with a lysis buffer containing 150 mM NaCl, 10 mM EDTA, 50 mM Tris (pH 7.4), and 1% Triton X-100 (Sigma-Aldrich) in the presence of protease (Thermo Fisher Scientific) and phosphatase inhibitors (Pierce Protein Biology). The lysate was centrifuged, and the supernatant was further centrifuged at 159,900 × g for 2 h at 2°C. The pellet contained the cytoskeletal fraction and the supernatant the soluble fraction. The cytoskeletal fraction was solubilized with a buffer containing 350 mM NaCl, 10 mM EDTA, 50 mM Tris (pH 7.4), and 1% Triton X-100 (Sigma-Aldrich) in the presence of protease and phosphatase inhibitors, then treated with DNase (Promega). The cytoskeletal fraction was then centrifuged, and the supernatant contained the solubilized proteins. All fractions were balanced to have the same NaCl content.

T cells, B cells, neutrophils, and DCs were lysed in 1% Triton X-100, 150 mM NaCl, 10 mM EDTA, 50 mM Tris (pH 7.4), and Pierce Phosphatase and Protease Inhibitor Mini Tablets (Thermo Fisher Scientific). The primary Ab rabbit polyclonal FlnA was from Bethyl Laboratories. The primary Abs used for fractionation samples were against talin (8d4; Santa Cruz Biotechnology), CD18 (C71/16; Abcam), and β-actin (Cell Signaling Technology). After incubation in the HRP-conjugated secondary Ab, blots were developed by a standard chemiluminescence technique.

1 × 10⁶ CD4 T_eff cells on ICAM-1 (Bio-Techne) 6 μg/ml–coated coverslips were incubated for 30 min and fixed with 4% paraformaldehyde. F-actin was stained with FITC-phalloidin (Sigma-Aldrich). For phalloidin staining, all slides were imaged using a Leica SP5 II and LAS AF Lite Software (Leica Microsystems), with 20% 488-Argon laser power (10% visible laser power). Z-stacks were taken with the following parameters: spectral range 496–678 nm, QD405/488/561/635 mirror, smart gain 650 V, smart offset 0%, pinhole 111.49 μm, line average 3, zoom 1, objective 63×, z-distance 10.078 μm, z-step size 0.15 μm, and format 512 × 512. The area, Feret diameter (the longest distance between any two points along the cell), fluorescence intensity, and background were measured with ImageJ and used to calculate corrected total cell fluorescence (CTCF) values as previously described. Briefly, CTCF values were calculated for individual cells based on measurements of area, intensity, and background with ImageJ as previously described [CTCF = integrated density − (area of selected cell × mean fluorescence of background readings)] (19). Twenty-five to one hundred cells per condition were measured for each animal.

Single-cell suspensions of LN, blood, thymus, and spleen were prepared. In some experiments, CD4 T cells were isolated, activated, and stained for flow cytometry. Single-cell suspensions from the ear tissue were prepared by cutting the skin sample into small pieces, passing it through a 70-μm cell strainer, washing it with PBS, and filtering it again through a 40-μm cell strainer (BD Biosciences). The following fluorescently labeled Abs were used: CD4 (RM4-5), CD3 (145-2C11), CD8a (53-6.7), B220 (RA3-6B2), CD44 (IM7), CD69 (H1.2F3), CD25 (PC61), CD18 (C71/16), CD62L (MEL-14), PSGL-1 (2PH1), CD29 (HMb1-1), CD49d (R1-2), β₇ (FIB27), and CD11a (2D7), all from eBioscience. Fc block (clone 2.4G2; BD Biosciences) was used in all staining. Intracellular staining for Foxp3 (FJK-16s) was performed according to the manufacturer’s instructions (eBioscience). For F-actin staining, nonadherent cultured T cells were fixed with 1% paraformaldehyde in PBS, washed, and stained with phalloidin-FITC (Sigma-Aldrich) in 0.1% saponin and 1% FBS/PBS. Data were acquired on an LSRFortessa (BD Biosciences) and analyzed using FlowJo software (Tree Star).

Static adhesion assays were performed as previously described (7). Briefly, the integrin ligand recombinant mouse ICAM-1/CD54 Fc chimera (0.5, 1, 3, and 6 μg/ml) was coated onto 96-well MaxiSorp plates (Nunc, Thermo Fisher Scientific) by overnight incubation at 4°C. T_eff cells at 1 × 10⁶ cells/ml were resuspended in an adhesion medium (RPMI 1640 supplemented with 0.1% BSA, 40 mM HEPES, and 2 mM MgCl₂) and added to the plate. Cells were allowed to adhere for 30 min at 37°C before gentle washing to remove unbound cells. Bound cells were lysed and detected with phosphatase substrate para-nitrophenyl phosphate (Sigma-Aldrich).

A shear flow adhesion assay was performed as in (20). ibidi VI 0.4 μ-slides were coated with 6 μg/ml ICAM-1 or with 6 μg/ml MAdCAM-1 (R&D Systems). ICAM-1 or MAdCAM-1 were coated overnight at 4°C, followed by SDF-1a (1 μg/ml) or CCL25 (4 μg/ml) coating at +37°C for 40 min. Cells flowed over coated VI 0.4 ibidi μ-slides at a 0.3–1 dynes/cm² (0.3 dynes/cm² for T_eff cells and 1 dyne/cm² for chemokine-activated naive T cells) continuous shear flow rate over a 5 min period. Cells were monitored by microscopy, and the number of adhered cells in the field of view was determined by manual counting. Rolling rates were analyzed from video recordings (Hamamatsu ORCA-Flash digital camera) with ImageJ. The average speed of 100 points per cell was measured, and a total of 30 cells were analyzed per condition.

T_eff cells were cultured for 1 h on elastic ICAM-1–coated silicone-based gel substrates (Young’s modulus = 2 kPa) (Matrigen). Substrates were surface coated with green fluorescent sulfate nanobeads (diameter 200 nm; Invitrogen, Carlsbad, CA). Single cells together with the underlying beads were imaged with the 3i Marianas imaging system (3i Intelligent Imaging Innovations) by using multipoint imaging. A 63×/1.2 W C-Apochromat Corr WD = 0.28 M27 objective was used, and the dish was placed in a heated sample chamber (37°C) controlled for CO₂. Following live cell imaging, the cells were detached from the gel substrates with 10× trypsin (Lonza), and a second set of nanobead images, serving as reference images, were obtained in a cell-free configuration. Spatial maps of cell-exerted nanobead displacements were achieved by comparing the reference bead images together with the experimental images. With the knowledge of the bead displacement fields, substrate stiffness (2 kPa), and a manual trace of the cell boundary, the cell-exerted traction fields were computed by using Fourier transform traction cytometry (21, 22). The root mean squared magnitude was computed from the traction field.

Purified control OT-II⁺ and FlnA KO/OT-II⁺ CD4⁺ T cells were labeled with CFSE (Life Technologies) according to the manufacturer’s instructions, and 2 × 10⁶ CD4⁺ T cells were injected i.v. in a volume of 200 μl into the recipient wild type (WT) mice. After 24 h, WT mice were immunized i.v. into the tail vein with 100 μg OVA peptide (AnaSpec) in 10 μg LPS (Sigma-Aldrich) in a total volume of 200 μl. Spleens were collected, and the splenic cells were stained for the T cell activation marker CD44 and analyzed by flow cytometry.

CD4 T cells were isolated with CD4 beads (MACS; Miltenyi Biotec) and stimulated with 2.5 μg/ml soluble anti-CD3 plus IL-2 for 24 h. IL-2 and IL-10 were measured from supernatants by ELISA. Abs for ELISA were purified anti-mouse IL-2 (clone JES6-1A12; eBioscience) and biotinylated anti-mouse IL-2 (clone JES6-5H4; eBioscience). IL-10 was detected with the DuoSet ELISA kit (R&D Systems).

Control and FlnA KO CD4 T cells were labeled with CFSE and CellTrace Violet (Life Technologies) according to the manufacturer’s instructions. Reversal of fluorescent dyes yielded the same result. Ctrl and FlnA KO CD4⁺ T were mixed at a 1:1 ratio, and 5–6 × 10⁶ cells were injected i.v. into recipient WT mice. At 1.5 or 18 h later, donor cells in the LNs, spleen, and blood were identified by flow cytometry.

ibidi μ-slide eight-well plates were coated with 3 μg/ml ICAM-1 in PBS overnight, after which wells were washed with an adhesion medium before 70,000 cells in 400 μl adhesion media were added to each well. Cells were imaged in doublets, with controls and KO cells on the same plate. Cells were imaged in brightfield with a 20× magnification using the 3i Marianas imaging system (3i Intelligent Imaging Innovations) with multipoint imaging. Cells were imaged for 1 h in 1-min intervals. Cells were tracked with the manual tracking plugin from ImageJ.

Mice were sensitized on their shaved backs with 1% oxazolone (OXA) on day 0 and challenged with 0.3% OXA on both ears on day 5, followed by sample collection and analysis 24 h later on day 6 as in (23). Ear thickness was measured with a micrometer (Mitutoyo, Kanagawa, Japan) at 24 h after challenge, after which the ear lobes were collected for RNA isolation and flow cytometry.

Ear samples were homogenized with T10 Ultra-Turrax (IKA) and RNA was extracted with the NucleoSpin RNA kit (Macherey-Nagel). cDNA was synthesized from 0.5 μg of total RNA with a High Capacity cDNA Reverse Transcription Kit (Applied Biosystems, Life Technologies). TaqMan gene expression assays were performed as in (7) with TaqMan Fast Advanced Master Mix and real-time quantitative PCR with a Bio-Rad CFX96 Real-Time PCR Detection System.

The Student t test or Mann–Whitney U test was used for statistical analysis. In all cases, p values are defined as *p < 0.05, **p < 0.01, and ***p < 0.001.

To investigate the role of FlnA in T cells in vivo, we generated T cell–specific FlnA KO mice. Mating of conditional Flna^fl/fl mice with CD4-Cre transgenic mice generated mice lacking FlnA in CD4 (and CD8) T cells. Western blotting confirmed the deletion of FlnA in cultured CD4 T_eff cells and in CD8 T cells (Fig. 1A). Deletion of FlnA did not lead to upregulation of other filamin isoforms in T cells (e.g., FlnB or FlnC) (Fig. 1B). Furthermore, expression of FlnA was normal in B cells, neutrophils, and cultured DCs (Supplemental Fig. 1A), showing that the deletion is T cell specific. Deletion of FlnA in CD4 T cells did not affect the proportion of CD4 or CD8⁺ T cells in the LNs, spleen, or thymus (Fig. 1C, Supplemental Fig. 1B). Also, the proportions of B220⁺ B cells in the spleen and the LNs were similar in control and FlnA KO/CD4-Cre mice (Fig. 1C). Therefore, filamin does not appear to be essential for T cell development in vivo.

FlnA has an important role in coupling the actin cytoskeleton to cell surface receptors, such as integrins, thereby influencing membrane–cytoskeleton interactions, and FlnA has previously been reported to function as an integrin inhibitor in cells (5, 8). We therefore investigated β₂ integrin expression and cytoskeletal coupling and function in FlnA KO T cells. The surface expression of β₂ integrins (CD18 and CD11a) in CD4 T cells was normal in the absence of FlnA (Supplemental Fig. 2A). In addition, the subcellular fractionation experiments of T_eff cells revealed that β₂ integrin distribution between soluble and cytoskeletal fractions was not significantly affected by FlnA deficiency (Fig. 2A). In addition, static adhesion assays with CD4⁺ T cells revealed that FlnA KO T cell adhesion to the β₂ integrin ligand ICAM-1 was similar to control cells (Fig. 2B), indicating that FlnA does not function as a β₂ integrin inhibitor in primary T cells.

As FlnA cross-links actin in cells, we asked whether FlnA deficiency affects cell morphology or the actin cytoskeleton in T cells. We cultured CD4⁺ T_eff cells from FlnA KO mice (as these have a more pronounced actin cytoskeleton than naive CD4⁺ lymphocytes) and investigated the F-actin cytoskeleton in these cells plated on the β₂ integrin ligand ICAM-1 by phalloidin staining. We found that the cell spreading of CD4⁺ T_eff cells was similar in FlnA KO and in control cells plated on ICAM-1 (Fig. 2C, 2E). Also, cell morphology assessed by Feret diameter analysis was similar in control and FlnA KO T cells (Fig. 2C). However, FlnA KO CD4 T cells displayed significantly increased F-actin content when compared with control cells (Fig. 2E). This result is in line with previous research showing increased F-actin levels in platelets lacking FlnA, which have one or more abnormal F-actin foci inside the cell when FlnA is absent (24), and in a human melanoma cell line lacking FlnA (25). The abnormal actin cytoskeleton in some FlnA KO T cells on ICAM-1 may be related to the important role of FlnA as a cross-linker of actin filaments into high-angle orthogonal networks linked to membrane receptors, such as integrins. This process may be disturbed in FlnA KO cells, resulting in denser, more parallel actin filaments in these cells when they are spreading on integrin ligands, as has been previously reported in human melanoma cells lacking FlnA (25). In contrast, F-actin content in T cells in suspension was normal (Fig. 2D).

In summary, our results show that FlnA-deficient T cells display relatively normal integrin-mediated static adhesion and spreading, and filamin therefore does not function as an integrin inhibitor in primary T cells.

Adhesion sites mediate actin-dependent cellular traction forces on exterior substrates, and FlnA has been reported to stabilize adhesion sites under applied force (26, 27). To investigate whether the integrin/FlnA signaling node in T cells affects transmission of force through the integrin to an extracellular substrate, traction force microscopy experiments with WT and FlnA KO T cells on hydrogels coated with the β₂ integrin ligand ICAM-1 were performed. In this method, the integrin-mediated cellular traction forces are measured by assessing integrin-mediated fluorescent bead displacement in the ligand-coated hydrogels. Importantly, we found that filamin-deficient CD4 T_eff cells generated significantly less β₂ integrin–dependent traction forces than control cells (Fig. 3A).

β₂ Integrins are essential for firm T cell adhesion to ligands on endothelial cells under conditions of shear flow. We hypothesized that the reduced integrin-mediated traction forces generated by FlnA-deficient T cells may influence the ability of the cell to resist shear forces. Therefore, adhesion of freshly isolated FlnA-deficient and control CD4 T cells was studied under shear flow conditions. FlnA-deficient CD4 T cells adhered slightly less to ICAM-1 in the presence of the chemokine SDF-1 (CXCL12) under low shear forces, although statistically significant reduction was only seen after 2 min of starting the flow (Fig. 3B). However, FlnA-deficient T cells displayed significantly faster integrin-mediated rolling on ICAM-1⁺SDF-1 than control cells, indicating that the bonds between integrin and ICAM-1 are weaker in FlnA KO than in control T cells. Because FlnA is known to interact also with β₇ integrins, which mediate homing into mucosal sites, adhesion to the β₇ integrin ligand MAdCAM-1 in the presence of the chemokine CCL25 under shear flow conditions was explored. We found that FlnA-deficient T cells also rolled faster on the β₇ integrin ligand MAdCAM-1 (Fig. 3B).

We further investigated the adhesion of cultured T_eff cells to ICAM-1 under flow conditions, which does not require chemokine activation (20), and found that FlnA is indeed required for optimal T_eff cell firm adhesion to the β₂ integrin ligand ICAM-1 under shear flow conditions (Fig. 3C).Our results therefore show that FlnA is required for formation of strong integrin–ligand bonds and therefore for firm adhesion or cell arrest under the conditions of shear flow.

FlnA binds to the LFA-1 integrin (5, 8), and we show above that FlnA is necessary for the formation of strong LFA-1–ICAM-1 bonds under force and for integrin force transmission. Functional LFA-1 is important for T cell activation in vivo (4). Therefore, to investigate the role of the integrin regulator FlnA in T cell activation in vivo, FlnA KO mice were crossed with TCR-transgenic (OT-II) mice. We purified and labeled CD4 T cells from either OT-II⁺–Ctrl or OT-II⁺FlnA KO mice with CFSE and adoptively transferred them into WT mice. Recipient mice were immunized with OVA peptide, which is recognized by the OT-II TCR transgene specific for OVA peptide (OVA 323–339). We found that the percentage of labeled cells in the spleen recovered 5 d postimmunization was increased, rather than decreased, in the FlnA KO group compared with the control group (Fig. 4A, Supplemental Fig. 3A). This reflects increased proliferation of FlnA KO T cells, as demonstrated by reduced expression of CFSE in FlnA KO T cells. In addition, the expression of the T cell activation marker CD44 was also increased in the Flna KO group (Fig. 4A). In vitro–stimulated T cells showed no difference in expression of CD69, CD25, or CD44 (Fig. 4B, Supplemental Fig. 3B) or IL-2 production (Fig. 4C). Therefore, FlnA is not required for T cell activation; instead, it appears to restrict T cell activation in vivo. In addition, the amount of Foxp3⁺ regulatory T cells (Treg) was reduced in the LNs of nontreated FlnA KO mice (Fig. 4C), and IL-10 production in nontreated FlnA KO T cells was reduced (Fig. 4C). These results demonstrate that FlnA deficiency does not lead to reduced T cell activation in vivo, but rather, T cell activation is increased in FlnA KO mice. In addition, Treg numbers and functionality (e.g., IL-10 production) are slightly reduced in these mice (Fig. 4C).

As integrin functionality was affected in FlnA KO T cells, we next asked the question whether FlnA plays a role in T cell homing into lymphoid organs in vivo. Therefore, control and FlnA KO T cells were purified and labeled with fluorochromes, and the labeled cells were injected i.v. at a 1:1 ratio into the tail veins of WT mice. Lymphoid organs were collected either 1.5 or 18 h later, and labeled cells were detected by flow cytometry. Interestingly, FlnA KO T cell trafficking was indeed significantly reduced into axillary LNs at both time points (Fig. 5). In addition, FlnA KO T cell homing was also decreased into cervical and inguinal LNs at the shorter time point, although this was not statistically significant after 18 h. FlnA KO T cell homing was decreased to mesenteric LNs only after 18 h. The different results for the short-term and long-term experiments may reflect effects other than homing in the long-term experiment (e.g., T cell egress, migration of T cells between different LNs, or T cell apoptosis). In contrast, similar numbers of control and FlnA-deficient T cells were detected in the blood and spleen, which is not surprising, as trafficking of T cells to the spleen is integrin independent. FlnA deficiency did not affect the expression of other T cell homing receptors important for T cell trafficking, such as α₄ integrin, β₁ integrin, β₇ integrin, CD62L, or CD44 (Supplemental Fig. 2). In contrast to in vivo migration and shear flow adhesion experiments, in vitro two-dimensional migration on ICAM-1 was similar for FlnA KO T cells and control cells (Supplemental Fig. 1C). These data confirm that FlnA is required for normal integrin function during in vivo homing to LNs, most likely by affecting the formation of strong integrin–ligand bonds under conditions of shear stress.

T cells play fundamental roles in the pathogenesis of allergic skin inflammation, and β₂ integrins are important in the trafficking of T cells into the inflamed skin (23, 28). The mouse model of contact hypersensitivity can be used to study T cell migration into inflammatory sites in vivo. We therefore used a contact hypersensitivity model to investigate whether FlnA plays a role in recruitment of T cells to the inflamed skin. OXA-sensitized Ctrl and FlnA KO/CD4-Cre mice had similar ear swelling (Fig. 6A), and there was no difference in the expression levels of inflammatory cytokines in the inflamed skin (Fig. 6C, Supplemental Fig. 4). Importantly, however, we found decreased numbers of CD4 and CD8 T cells in the inflamed skin of FlnA KO/CD4-Cre mice (Fig. 6B). In summary, our results show that FlnA is important for integrin-mediated adhesion under shear flow, and it is important not only for CD4⁺ T cell homing into LNs but also for T cell trafficking into sites of inflammation.

T cell trafficking between the blood, LNs, and peripheral tissues is essential for their functions in the immune system. Integrins are essential for T cell trafficking, as they mediate firm adhesion under shear flow conditions (blood flow) between the T cell and the endothelium. In addition, integrins are important for the formation of the immunological synapse between the T cell and the APC, and therefore, integrins are required for T cell activation. The actin-binding protein FlnA provides an important link between the integrin and the actin cytoskeleton. FlnA is a large (280 kDa) actin cross-linking protein with important roles in cellular mechanosensing. It has a key role in regulating the structural stability, adhesion, and motility of cells by cross-linking actin. FlnA links to transmembrane proteins, including integrins, as well as many other proteins in cells (12). FlnA binding to the integrin β-chain cytoplasmic domain has been previously shown to inhibit integrin activation by competing with talin, a well-known integrin activator (5, 8), and/or by linking α and β integrin chains together (14). However, the role of FlnA in integrin regulation in primary immune cells has remained incompletely understood. Our results clearly demonstrate that FlnA does not function as an integrin inhibitor in primary T cells, and it is not required for T cell spreading on integrin ligands or for T cell adhesion or migration under static conditions. However, it has an important role in integrin-mediated mechanotransduction in T cells, and it is needed for optimal T cell trafficking in vivo.

FlnA has been proposed to have an important role in CD28-mediated costimulation of human T cells in vitro (15, 29, 30). However, the role of filamin in regulating T cell responses in vivo has remained poorly understood. In this article, we show that in vivo, FlnA is not required for T cell activation but instead negatively regulates T cell activation in the spleen, although exactly how filamin regulates T cell activation is still unclear. Interestingly, FlnA deficiency resulted in decreased levels of Foxp3 expression in inflamed skin and in decreased numbers of Treg in LNs of nontreated mice. Therefore, the absence of FlnA leads to abnormalities in Foxp3⁺Treg numbers or functionality. Interestingly, β₂ integrins and the integrin activator talin-1 have also been previously shown to be important for Treg development in vivo (31, 32). Furthermore, it was previously shown that T cell–specific deletion of talin leads to spontaneous lymphocyte activation (33). We did not see spontaneous activation in FlnA–CD4-Cre mice, which may be explained by the higher percentage of Foxp3⁺ T cells in the absence of filamin (5% decrease as compared with control) than talin (10% decrease as compared with control). However, it is clear that the role of filamin in T cell activation and Treg function needs further investigation.

To enter LNs or inflamed tissues, T cells need to be able to arrest on endothelial cells surrounding the blood vessels under conditions of shear flow. Integrins mediate this firm adhesion and transmit mechanical forces by integrating the cell exterior and the cytoskeleton inside the cell. In this article, we show that rather than being an integrin inhibitor, FlnA is required for the formation of strong integrin–ligand bonds and for optimal T cell adhesion to integrin ligands under conditions of shear flow. LFA-1–ICAM-1 bonds are so-called catch bonds, which become stabilized under force, such as shear flow in blood vessels (34, 35). External shear force stabilizes LFA-1 in the extended conformation, leading to slow rolling or arrest on the endothelium in blood vessels. In the absence of the link between integrin, FlnA, and actin, this stabilization of the integrin-extended conformation may not occur. As a result of weakened integrin–ligand bonds, FlnA-deficient T cells display increased rolling rates of T cells on integrin ligands ICAM-1 and MadCAM, and T_eff cells display reduced adhesion to ICAM-1 under shear flow conditions. In addition, we show that in the absence of FlnA, traction forces generated by T cells on integrin–ligand-coated substrates are significantly weaker than in the presence of FlnA, indicating that FlnA is essential for integrin force transmission in these cells. Therefore, FlnA is required for optimal T cell adhesion to β₂ and β₇ integrin ligands under shear flow conditions. Consequently, we show that in the absence of FlnA, T cell trafficking into lymphoid organs is significantly reduced. Our results therefore clearly demonstrate the physiological importance of FlnA in integrin-mediated mechanotransduction in T cells.

Talin and kindlin-3 are integrin interactors that have previously been shown to be of fundamental importance in integrin-mediated leukocyte adhesion under shear flow (7, 31, 36). In the absence of these proteins, integrin-mediated shear flow adhesion is completely abolished (36). In addition, both the absence of talin and the mutation of the kindlin-3 binding site in β₂ integrin lead to reduced amounts of T cells in peripheral LNs (7, 31) and reduced homing of T cells in vivo. In this study, we also show that FlnA plays an important role in T cell shear flow adhesion and lymphocyte trafficking. However, the absence of FlnA results in increased integrin-mediated rolling rather than a complete abolishment of adhesion, showing that these integrin-interacting proteins regulate different phases of integrin-mediated shear flow adhesion. Also, although homing of FlnA-deficient lymphocytes is clearly reduced in short-term homing assays, the numbers of T cells in peripheral LNs are normal in these mice, indicating that, over time, FlnA-deficient T cells can accumulate in LNs. In addition to its role in lymphocyte homing, we also show that FlnA, like β₂ integrins (28) and kindlin-3 (23, 37), is required for T cell trafficking into sites of inflammation (e.g., to the inflamed skin in a contact hypersensitivity model).

Taken together, our results show that FlnA regulates force transmission through integrins in T cells and T cell shear flow adhesion. Because of this role in integrin force transmission, FlnA is required for T cell homing and trafficking into sites of inflammation. However, it is not required for T cell activation in vivo, a process that occurs in lymphoid organs in the absence of shear flow. Our data therefore provide novel insights into how FlnA regulates the function of T cells in vitro and in vivo.

The authors have no financial conflicts of interest.