Research ArticleLYMPHOCYTE MIGRATION

RhoB controls the Rab11-mediated recycling and surface reappearance of LFA-1 in migrating T lymphocytes

See allHide authors and affiliations

Sci. Signal.  12 Dec 2017:
Vol. 10, Issue 509, eaai8629
DOI: 10.1126/scisignal.aai8629

Recycling integrins

T cells must rapidly migrate along endothelial cell surfaces and transmigrate to reach sites of infection and inflammation in tissues. The cell surface integrin LFA-1 binds to its ligand ICAM-1 on endothelial cells in a dynamic manner so that T cells rapidly adhere to and detach from the endothelial surface to migrate quickly. Using biochemical and imaging analyses, Samuelsson et al. identified a process by which LFA-1, internalized after its interaction with ICAM-1, is recycled back to the T cell surface to maintain migration. LFA-1 bound to the GTPase RhoB in the cytosol, which mediated its recruitment to Rab11-positive rapidly recycling endosomes. Loss of RhoB function resulted in the cytosolic accumulation of LFA-1 at the rear of the cell, reduced cell surface abundance of LFA-1, and impaired T cell migration.

Abstract

The regulation of cell adhesion and motility is complex and requires the intracellular trafficking of integrins to and from sites of cell adhesion, especially in fast-moving cells such as leukocytes. The Rab family of guanosine triphosphatases (GTPases) is essential for vesicle transport, and vesicles mediate intracellular integrin trafficking. We showed that RhoB regulates the vesicular transport of the integrin LFA-1 along the microtubule network in migrating T lymphocytes. Impairment in RhoB function resulted in the accumulation of both LFA-1 and the recycling endosomal marker Rab11 at the rear of migrating T lymphocytes and decreased the association between these molecules. T lymphocytes lacking functional RhoB exhibited impaired recycling and subsequently decreased surface amounts of LFA-1, leading to reduced T cell adhesion and migration mediated by the cell adhesion molecule ICAM-1 (intercellular adhesion molecule–1). We propose that vesicle-associated RhoB is a regulator of the Rab11-mediated recycling of LFA-1 to the cell surface, an event that is necessary for T lymphocyte motility.

INTRODUCTION

Leukocytes need to quickly transmigrate from the blood vessels into tissues upon inflammation or infection. An essential mechanism regulating this process is the subcellular trafficking of adhesion molecules, mainly integrins (1). LFA-1 (also known as CD11a/CD18) is composed of the αL and β2 integrin subunits. LFA-1 is the major integrin used by T lymphocytes to adhere and migrate. For LFA-1 to function effectively, it must switch between active and inactive conformations. In the absence of stimulation, LFA-1 has a bent, inactive, non–ligand-binding conformation on circulating leukocytes. When activated through receptors such as chemokine receptors or T cell receptors, the inside-out signaling cascade initiates a conformational change that makes the ligand-binding domain of LFA-1 accessible by generating an extended and intermediate-affinity conformation of LFA-1 (2). To strengthen adhesion and initiate cell spreading and motility, LFA-1 outside-in signaling must occur. The interaction of intermediate-affinity LFA-1 with its ligand intercellular adhesion molecule–1 (ICAM-1), together with exposure to shear force, promotes the high-affinity conformation of LFA-1, leading to its full activation (2). LFA-1 needs to quickly alter its affinity to function effectively. For example, locking LFA-1 in either the intermediate-affinity or the high-affinity conformation causes T lymphocytes to adhere without detachment, giving a lack-of-motility phenotype in vitro (3). LFA-1–mediated adhesions are thus considered to be highly dynamic structures, being constantly reused in migrating cells. Nevertheless, in contrast to integrins with β1 or β3 subunits, little is known about the recycling of integrins with β2 subunits (4). However, LFA-1–dependent T lymphocyte migration decreases when intracellular vesicle transport is blocked (5).

Rab proteins are members of the Ras superfamily of guanosine triphosphatases (GTPases), and they function as molecular switches to control diverse stages of vesicular traffic (6). Integrins with various subunit combinations traffic through Rab-containing compartments in the cell (4). The exact pathways are complex and likely depend not only on multiple factors, including integrin activation state and integrin family, but also on cell type and extracellular stimuli (7). LFA-1 is associated with the early endosomal marker Rab5 in migrating T lymphocytes, as well as with both Rab5 and the recycling endosomal marker Rab11 in migrating neutrophils (810). In addition, Nishikimi et al. (11) reported that active Rab13 is associated with the kinase Mst1 and facilitates the delivery of LFA-1 to the leading edge of migrating lymphocytes. RAPL (regulator for cell adhesion and polarization enriched in lymphoid tissues) is an Mst1- and Rap1-binding effector protein (12). The Rap1-RAPL-Mst complex increases LFA-1 affinity and avidity, thus strengthening the LFA-1–ICAM-1 interaction (1315). The RAPL-Mst1 complex is also dynamically translocated together with LFA-1 to the leading edge of lymphocytes, suggesting that this complex has a role in the regulation of LFA-1 intracellular transport (14).

The Rho subfamily of the Ras GTPases controls cytoskeletal reorganization to indirectly regulate endocytosis and vesicle traffic (16). In particular, RhoB is involved in the transition of endosomal vesicles from the peripheral actin cytoskeleton to the microtubular network in epithelial cells and regulates endosomal traffic back to the plasma membrane (17, 18). In addition, RhoB is important for the endosomal trafficking of various receptor tyrosine kinases, such as epidermal growth factor receptor and platelet-derived growth factor receptor (1922). The role of RhoB in intracellular integrin trafficking has not been thoroughly investigated. RhoB-deficient murine macrophages have impaired integrin-dependent attachment. These cells have normal β1 integrin abundance at the cell surface but have reduced amounts of β2 and β3 integrins at the cell surface (23). A human RhoB-deficient prostate cancer cell line has a decreased amount of β1 integrins at the cell surface, which correlates with an increase in cell motility (24). Whether RhoB is involved in intracellular transport of integrins is currently unknown.

Here, we investigated the function of RhoB in the intracellular transport of LFA-1 in human T lymphocytes. In T lymphocytes migrating on ICAM-1–coated surfaces, we found that LFA-1 colocalized with RhoB in vesicular structures, some of which were distributed along the α-tubulin network. Our data revealed a role for RhoB in the Rab11-dependent recycling of LFA-1. RhoB knockdown resulted in the accumulation of both Rab11 and LFA-1 in the uropod of migrating T lymphocytes. Moreover, T lymphocytes expressing dysfunctional RhoB not only had impaired LFA-1–mediated migration but also impaired LFA-1 recycling to the cell surface and reduced association between LFA-1 and Rab11.

RESULTS

RhoB interacts with LFA-1 and regulates LFA-1–dependent migration

To investigate the role of RhoB in LFA-1–mediated T lymphocyte migration, we used RhoB-specific small interfering RNA (siRNA) to reduce its abundance in T lymphocytes (Fig. 1A). Knocking down RhoB in T lymphocytes substantially reduced the mean cell migration distance and speed compared to those of control cells (Fig. 1, B and C), and a similar reduction in migration was also seen in the T lymphocyte cell line HSB-2 using two different RhoB-specific siRNAs (fig. S1A). In addition, the uropods in most RhoB-knockdown (RhoB-KD) T lymphocytes were attached to the immobilized ICAM-1–coated substrate rather than projecting upwards as is commonly seen for migrating T lymphocytes (fig. S1, B and C). Together, these data suggest that reducing RhoB function impairs T lymphocyte migration.

Fig. 1 RhoB interacts with LFA-1, and a reduction in RhoB abundance impairs LFA-1–dependent migration in T lymphocytes.

(A) Western blots (representative) from T lymphocytes transfected with scrambled control or RhoB-specific siRNA were incubated with antibodies against RhoB and actin (knockdown efficiency, 52.6 ± 8.9%; means ± SEM; n = 7 experiments). (B) T lymphocytes transfected with the indicated siRNAs were incubated on immobilized ICAM-1 before random migration was observed by time-lapse microscopy. Individual cells were tracked and plotted with a common origin. A representative pattern of a single cell from 25 cells in each of three independent experiments. (C) Quantification of mean cell migration speed from experiments similar to those shown in (B). Data are means ± SEM. (D) T lymphocytes were applied to an Ibidi μ-Slide containing a confluent human umbilical cord endothelial cell (HUVEC) layer with a shear stress of 1 dyne/cm2. The activity of interacting T lymphocytes was recorded at a rate of 60 frames/min during 3 min. T lymphocytes were manually scored according to their rate of interaction with HUVEC. Adherent: T lymphocytes adhered for more than 3 s. Transendothelial migration (TEM): T lymphocytes transmigrated beneath the HUVEC layer. Means ± SEM; n = 3 experiments. Statistical significance between siCTRL and siRhoB for both Adherent and TEM is indicated. (E) Guanosine triphosphate (GTP)–Rho isolated from T lymphocytes migrating on immobilized intercellular adhesion molecule–1 (ICAM-1) with or without CXCL12 was analyzed by Western blotting for RhoB. Mean horseradish peroxidase (HRP) intensity was normalized to unstimulated control (Unstim) of cells grown in suspension without CXCL12. Means ± SEM; n = 4 experiments. Statistical significance between Unstim and ICAM-1 and ICAM-1 + CXCL12 is indicated. See also fig. S1F for representative Western blotting analysis. (F) RhoB or rabbit immunoglobulin G (IgG) control immunoprecipitates (IP) from T lymphocytes grown in suspension were analyzed by Western blotting for the presence of RhoB and LFA-1. Representative blot; n = 2 experiments. The total cell lysate (TCL) and immunoprecipitated samples were from different parts of the same gel; intervening lanes have been spliced out. (G) GTP-Rho immunoprecipitates from T lymphocytes migrating on immobilized ICAM-1 (or unstimulated control cells grown in suspension without ICAM-1) were analyzed by Western blotting for the presence of LFA-1. Representative blot; n = 3 experiments. (H) Mean HRP intensity from experiments similar to those shown in (G), for Unstim and ICAM-1, respectively. Data are normalized to Unstim and shown as means ± SEM; n = 3 experiments. (I) Representative confocal images of a T lymphocyte attached to immobilized ICAM-1 and stained with antibodies against LFA-1 (green) or RhoB (red). Magnified areas (1, middle; 2, bottom) are indicated with dashed rectangles in the top left image. White arrows indicate fluorescence overlap. Images are deconvoluted and Gaussian-fitted. Scale bars, 5 μm (low magnification) and 2 μm (high magnification); n = 5 experiments (with 15 to 20 cells per experiment). (J) Mander’s colocalization coefficient (MCC) quantification of LFA-1 fluorescence colocalization with RhoB (left) and vice versa (right) in T lymphocytes. Mean overlap from three regions averaged per cell (total cell) and overlap from single regions grouped by location within the cell (uropod and lamellipodium). Pooled data from two independent experiments (n = 36 cells) are shown. Means ± SEM; n = 2 experiments. (K) High magnification of a representative STED image of a T lymphocyte attached to immobilized ICAM-1 and stained with antibodies against RhoB (red) or LFA-1 (green). White arrows indicate fluorescence overlap. Scale bar, 1 μm. *P <0.05, **P < 0.01, ***P < 0.001, unpaired (two-tailed) t test.

To study the effect of knocking down RhoB in T lymphocytes in a more physiologically relevant setting, we used a confluent monolayer of human umbilical vein endothelial cells (HUVECs) in Ibidi μ-Slide chambers to mimic a blood vessel environment. RhoB-KD T lymphocytes suspended in migration buffer were applied with a shear stress of 1 dyne/cm2 to a μ-Slide with a confluent HUVEC layer. The interaction of T lymphocytes with HUVECs was manually scored as the number of binding events and transmigration events, reflecting their adherent capacity and their ability to transmigrate through the HUVEC layer, respectively. Compared to control T lymphocytes, RhoB-KD T lymphocytes exhibited substantially reduced adhesion to and transmigration through the HUVEC layer (Fig. 1D). In addition, RhoB-KD T lymphocytes showed substantially reduced chemotaxis on ICAM-1 toward the chemokines CXCL12 and CCL19 in Transwell assay compared to that of control cells (fig. S1D). As a control, placing CXCL12 and CCL19 in both the upper and lower chambers abrogated directed T lymphocyte movement (fig. S1E). Overall, inhibiting RhoB impaired T lymphocyte migration in this more hemodynamically relevant setting.

We next analyzed the activity of RhoB in T lymphocytes during LFA-1–mediated migration. We isolated the active, guanosine triphosphate (GTP)–bound form of Rho from T lymphocytes migrating on immobilized ICAM-1 in the presence or absence of CXCL12 and performed Western blotting to detect RhoB. We observed a substantial increase in the amount of GTP-bound RhoB in T lymphocytes that migrated on ICAM-1 in the presence or absence of chemokine compared to that in T lymphocytes kept in suspension (Fig. 1E and fig. S1F). However, the amount of GTP-bound RhoB was similar in the presence and absence of CXCL12 for cells that migrated on ICAM-1 (Fig. 1E and fig. S1F). In summary, LFA-1–mediated migration on ICAM-1 increased the amount of the active GTP-bound form of RhoB in T lymphocytes, and this response was independent of chemokine receptor stimulation.

Next, we used coimmunoprecipitation to investigate whether RhoB and LFA-1 were present in a protein complex. We detected LFA-1 in RhoB immunoprecipitates in T lymphocytes grown in suspension (Fig. 1F). The extent of the interaction between active (GTP-bound) Rho and LFA-1 increased as a function of LFA-1–ICAM-1 engagement (Fig. 1, G and H). In addition, we detected LFA-1 in immunoprecipitates from HSB-2 cells expressing green fluorescent protein (GFP)–tagged wild-type RhoB (WT-RhoB) (fig. S1G). LFA-1 did not coimmunoprecipitate with GFP expressed in control cells. In addition, in the presence of an antibody against GFP, no nonspecific binding was observed in the lysates of cells that did not express GFP, which further verifies the specificity of an antibody recognizing GFP (fig. S1H). These findings suggest that RhoB is part of a complex with LFA-1 and that the interaction between both proteins increases with LFA-1–ICAM-1 engagement.

Intracellular LFA-1 exists in vesicular structures partly distributed together with RhoB in migrating T lymphocytes

To further define the RhoB–LFA-1 complex, we investigated RhoB and LFA-1 distribution in T lymphocytes migrating on immobilized ICAM-1. Confocal analysis showed that RhoB was localized in small intracellular vesicles, starting from the uropod and along the longitudinal axis of the cell toward the center. When costaining RhoB together with LFA-1, we found that LFA-1 was associated in similar intracellular structures (Fig. 1I). Fluorescence colocalization analysis using Mander’s colocalization coefficient (MCC) showed that about one-third of LFA-1 associated with RhoB, and similar results were obtained for the overlap between RhoB and LFA-1 (Fig. 1J). Fluorescence quantification also showed that LFA-1 and RhoB associated in similar quantities in the uropod and in the lamellipodium. We also found RhoB and LFA-1 in similar intracellular structures when analyzed in super-resolution images obtained with stimulated emission depletion (STED) microscopy (Fig. 1K).

Thus, in migrating T lymphocytes, we found a functional and interactive association between RhoB and LFA-1. To define where this interaction functions within the cell, we looked at the distribution of RhoB and LFA-1 in relation to the cytoskeleton. The general role of the Rho GTPase subfamily in the regulation of endocytic traffic is mediated through modulation of the actin cytoskeleton (25); however, in T lymphocytes migrating on immobilized ICAM-1, our confocal image analysis of cells stained for RhoB, LFA-1, and actin showed that most RhoB did not colocalize with the actin cytoskeleton and that some LFA-1 did colocalize with actin (Fig. 2A and fig. S2, A and B). Because intracellular LFA-1–positive vesicles can also be found along the tubulin network (Fig. 2B), we investigated the association between RhoB, LFA-1, and tubulin using both confocal and STED microscopy. RhoB-positive intracellular vesicles were distributed along the α-tubulin network in T lymphocytes migrating on immobilized ICAM-1 (Fig. 2, B and C, and fig. S2, A and B). The same observation was also true for LFA-1 using confocal microscopy (Fig. 2B and fig. S2, A and B). These observations suggested that some RhoB and LFA-1 interaction occurred along the tubulin network.

Fig. 2 LFA-1 is present in cytosolic clusters similar to those containing RhoB and tubulin in migrating T lymphocytes, and reducing RhoB abundance impairs the localization of LFA-1.

(A and B) Representative confocal images of T lymphocytes attached to immobilized ICAM-1 stained with antibodies against LFA-1 (green), RhoB (red), and actin (blue) or tubulin (blue). Magnified areas (1, top right; 2, bottom right) are indicated with dashed rectangles in large image. White arrows indicate fluorescence overlap between LFA-1 and RhoB or between LFA-1 and actin in (A) and between LFA-1, RhoB, and tubulin in (B). Images are deconvoluted and Gaussian-fitted. Scale bars, 5 μm (low magnification) and 2 μm (high magnification). Images are representative of five experiments with 15 to 20 cells analyzed per experiment. See also fig. S2 for triple-color overlaps and quantification. (C) High-magnification image from a representative STED image of a T lymphocyte attached to ICAM-1 and stained with antibodies against RhoB (red) and tubulin (green). White arrows indicate fluorescence overlap; n = 3 experiments. Scale bar, 1 μm. (D) Representative confocal images of T lymphocytes transfected with scrambled control or RhoB-specific siRNA attached to immobilized ICAM-1 and stained with antibody against LFA-1. Marked cell area (3 × 3 μm) taken for analysis: uropod (white squares) and lamella (yellow squares). Top images are grayscale and bottom images are rainbow scale of graded fluorescence intensity. z1, low z-position; z2, middle z-position; and z3, high z-position; n = 3 experiments [15 to 20 cells per small interfering RNA (siRNA) per experiment]. Scale bar, 5 μm. (E and F) Mean fluorescence intensity (MFI) of LFA-1 from cells in experiments similar to those in (D) in the uropod (E) and in the lamellipodium (F). Pooled data of designated cell area (3 × 3 μm) from two independent experiments (n = 40 cells) are shown. Normalized to cells with control siRNA. Means ± SEM; n = 2 experiments. (G) Representative confocal image of HSB-2 cells expressing green fluorescent protein (GFP)–tagged wild-type (WT)–RhoB (green) and stained for internalized (Int.) LFA-1 (red). Magnified area (bottom) is indicated with dashed rectangles in the top left image; white arrows indicate fluorescence overlap. Scale bars, 5 μm (low magnification) and 2 μm (high magnification); n = 3 experiments (with 10 cells per experiment). *P < 0.05, ****P < 0.001, unpaired (two-tailed) t test.

Reducing the abundance of RhoB in T lymphocytes increases the amount of LFA-1 in the rear of a migrating cell

When examining the effect of knocking down RhoB on LFA-1 location in migrating T lymphocytes, we observed an accumulation of LFA-1 in the uropod (Fig. 2, D and E). To exclude the possibility of an overall increase in LFA-1 abundance in RhoB-KD T lymphocytes, we also analyzed LFA-1 in the lamellipodium. Analysis of the same z-position as in the uropod analysis, mean fluorescence intensity data showed a reduction in LFA-1 fluorescence in the lamellipodium of RhoB-KD T lymphocytes (Fig. 2F). This finding suggests that RhoB regulates LFA-1 localization within the T lymphocyte.

To further investigate the functional role of RhoB on LFA-1 distribution, we studied whether internalized LFA-1 localized with RhoB-containing vesicles. We stained live HSB-2 cells expressing GFP-tagged WT-RhoB with the non–function-blocking LFA-1–specific antibody Ts2.4. After the stained cells migrated on immobilized ICAM-1, which would promote the internalization of the Ts2.4–LFA-1 complex, we fixed the cells and further stained them for membrane-bound and internalized complexes with Alexa Flour 568–conjugated antibody recognizing mouse immunoglobulin G (IgG). We found that the internalized Ts2.4–LFA-1–containing vesicles colocalized with GFP-tagged WT-RhoB (Fig. 2G), suggesting a role for RhoB in the regulation of intracellular LFA-1 trafficking.

LFA-1 locates within EEA-1–positive early endosomal structures and within Rab11-positive recycling endosomal structures in migrating T lymphocytes

Intracellular sorting fate depends on the coordinated action of multiple endocytic regulatory molecules, including the Rab proteins, which regulate distinct endocytic pathway events. The endosomal marker of the early endosome is endosomal early antigen–1 (EEA-1; an effector of Rab5), and the recycling endosomal markers for fast and slow recycling pathways are Rab4 and Rab11, respectively. We analyzed and quantified the association of LFA-1 and RhoB in migrating T lymphocytes with these markers of the endosomal compartment using confocal analysis (Fig. 3, A and B) with MCC quantification (Fig. 3C). Both LFA-1 (Fig. 3, A to C) and RhoB (Fig. 3, B and C) associated with EEA-1 and Rab11. LFA-1 showed a very weak extent of association with Rab4 (Fig. 3, A to C). The association between RhoB and Rab11, as well as between Rab11 and LFA-1, was also supported by coimmunoprecipitation experiments (fig. S3, A and B). A lower association was seen between RhoB and Rab4 in migrating T lymphocytes compared to Rab11 (Fig. 3, C and D). No overlap was detected between GFP-tagged Rab7, a late endosomal marker, and RhoB in migrating HSB-2 cells (fig. S3, C and D). These observations suggest a main role for Rab11 and EEA-1 in the interplay between LFA-1 and RhoB in migrating T lymphocytes.

Fig. 3 LFA-1 locates within EEA-1–positive early endosomal structures and within Rab11-positive recycling endosomal structures in migrating T lymphocytes.

(A) Representative confocal image of T lymphocytes attached to immobilized ICAM-1 and stained with antibodies against LFA-1 and EEA-1, Rab4, or Rab11. Magnified area (right) is indicated with dashed rectangles in the leftmost image. White arrows indicate fluorescence overlap. Scale bars, 5 μm (low magnification) and 2 μm (high magnification). Images are representative of three experiments, with 15 to 20 cells analyzed for each antibody per experiment. (B) T lymphocytes were stained with antibodies against RhoB and EEA-1 (top), Rab4 (middle), or Rab11 (bottom) as described in (A). (C) MCC quantification of fluorescence colocalization from experiments similar to those shown in (A). Mean overlap from three regions averaged per cell was pooled from two independent experiments (n = 25 to 36 cells). (D) Quantification of fluorescence colocalization in (B) as described in (C). Data from 25 to 36 cells are shown. In (C) and (D), data are presented as means ± SEM; n = 2 experiments. Statistical significance between EEA-1 versus Rab4 and Rab11 is indicated as asterisks above bars and between Rab4 and Rab11 is indicated by asterisks above black lines. **P < 0.01, ***P < 0.001, unpaired (two-tailed) t test.

Dysfunctional RhoB impairs the LFA-1–mediated migration of HSB-2 cells

We next expressed GFP, GFP-tagged WT-RhoB, the GFP-tagged constitutively active Q63L mutant RhoB (CA-RhoB), or the GFP-tagged dominant-negative T19N mutant RhoB (DN-RhoB) in HSB-2 cells. Because of the low transfection rate, we sorted the cells based on GFP intensity (fig. S4A) and then analyzed random cell migration on immobilized ICAM-1 using time-lapse microscopy. The average migration speeds of HSB-2 cells expressing either CA-RhoB or WT-RhoB were similar, but the migration speed of HSB-2 cells expressing DN-RhoB was substantially reduced (Fig. 4, A and B). The average attachment (quantified as the percent of adherent cells) to immobilized ICAM-1 of HSB-2 cells expressing DN-RhoB was substantially reduced compared with HSB-2 cells expressing either WT-RhoB or CA-RhoB (Fig. 4C). Flow cytometry analysis revealed that HSB-2 cells expressing DN-RhoB displayed a reduced abundance of LFA-1 at the cell surface compared with T lymphocytes expressing GFP, WT-RhoB, or CA-RhoB (Fig. 4D). In summary, HSB-2 cells expressing DN-RhoB exhibited impaired attachment to ICAM-1, which likely results from the lower amount of LFA-1 at the cell surface, and accordingly showed impaired LFA-1–mediated migration.

Fig. 4 Dysfunctional RhoB impairs the LFA-1–mediated migration of HSB-2 cells.

(A) HSB-2 cells expressing GFP, GFP-tagged WT-RhoB, GFP-tagged CA-RhoB, or GFP-tagged DN-RhoB were sorted (24 hours after transfection) based on GFP expression and incubated on immobilized ICAM-1 before random migration was observed by time-lapse microscopy. Individual cells were tracked and plotted with a common origin. A representative pattern tracked by a single cell from 45 cells from three experiments is shown. (B) Quantification of mean speed of the cells shown in (A). Means ± SEM; n = 3. Statistical significance between GFP-WT-RhoB and GFP-DN-RhoB (black line) is indicated. (C) Sorted HSB-2 cells expressing GFP, GFP-WT-RhoB, GFP-CA-RhoB, or GFP-DN-RhoB were stained with CellTracker Blue CMAC and incubated on immobilized ICAM-1. Excitation was measured, and nonadherent cells were removed by gentle washing before excitation was measured again. Adhesion index: Percentage of adherent cells = (average fluorescence intensity read in the washed wells)/(average fluorescence intensity read in the unwashed cells) × 100%. Means ± SEM; n = 4 experiments. Statistical significance between GFP-WT-RhoB and GFP-DN-RhoB (black line) is indicated. (D) Sorted HSB-2 cells expressing GFP, GFP-WT-RhoB, GFP-CA-RhoB, or GFP-DN-RhoB were analyzed by flow cytometry based on LFA-1 surface expression. Data are pooled and normalized to GFP. Means ± SEM; n = 3 experiments. Statistical significance between GFP-WT-RhoB and GFP-DN-RhoB (black line) is indicated. **P < 0.01, ***P < 0.001, unpaired (two-tailed) t test.

Active RhoB regulates LFA-1 transport toward membrane compartments in HSB-2 cells

To investigate whether a defect in trafficking contributed to the reduced abundance of LFA-1 at the cell surface, defective cell migration, and impaired cell adhesion to ICAM-1, we examined the location of GFP-tagged WT-RhoB, CA-RhoB, and DN-RhoB in relation to LFA-1 in HSB-2 cells. Confocal analysis showed that WT-RhoB was localized in small intracellular vesicular structures similar to the localization of endogenous RhoB (fig. S4B). The analysis also demonstrated the specificity of the polyclonal antibody against RhoB in use. CA-RhoB was also partly distributed in intracellular vesicular structures; however, more CA-RhoB accumulated toward the rear of the cell (fig. S4B). LFA-1 was found in similar intracellular structures compared to those positive for WT-RhoB or CA-RhoB (fig. S4B). DN-RhoB was mostly found toward the uropod and formed even higher accumulation of vesicles than those formed by CA-RhoB. In the DN-RhoB–expressing cells, LFA-1 also accumulated in intracellular structures toward the rear of the cell (fig. S4B). These observations may explain why HSB-2 cells expressing DN-RhoB had less LFA-1 at the cell surface and consequently exhibited reduced adhesion to ICAM-1.

To gain quantitative insight into the effects of WT-RhoB, CA-RhoB, and DN-RhoB on the localization of LFA-1 and Rab11, we obtained cytoplasmic and membrane protein fractions of T lymphocytes and HSB-2 cells expressing these GFP-tagged RhoB proteins and examined the extracts by Western blotting. We found that endogenous RhoB, as well as all of the GFP-tagged RhoB proteins, was present in membrane compartments, but was relatively enriched in the cytoplasm (fig. S4, C and D). In contrast to RhoB, endogenous Rab11 in T lymphocytes and in HSB-2 cells expressing RhoB fusion proteins was more evenly distributed between the cytoplasmic and membrane compartments (fig. S4, C and D). Endogenous LFA-1 in T lymphocytes had an even distribution between the two compartments; however, there was a trend toward a more membrane-bound localization in HSB-2 cells expressing WT-RhoB, CA-RhoB, or DN-RhoB (fig. S4, C and D). In summary, both our confocal and biochemical analysis of HSB-2 cells expressing CA-RhoB or DN-RhoB indicated that LFA-1 localization depended on the proper balance of RhoB activity. Balanced activity was needed to guide LFA-1 between the cytoplasmic and membrane compartments. When RhoB activity was compromised, for example, in cells expressing DN-RhoB, LFA-1 accumulated in intracellular structures, resulting in lower LFA-1 surface membrane amounts and consequently impairment of LFA-1–mediated migration of HSB-2 cells.

Dysfunctional RhoB impairs LFA-1 internalization and recycling in HSB-2 cells

To investigate the mechanism by which HSB-2 cells expressing DN-RhoB had lower LFA-1 surface amounts and consequently were less adherent to ICAM-1, we assessed LFA-1 internalization and recycling using a previously established flow cytometry–based integrin internalization assay (26). We used primaquine, which reduces LFA-1–mediated migration by preventing the intracellular transport and primary recycling of LFA-1 (5), to validate that the assay quantified LFA-1 trafficking. Primaquine impaired the recycling of LFA-1 in a dose-dependent manner in T lymphocytes (fig. S5A). In HSB-2 cells expressing GFP, GFP-tagged CA-RhoB, or GFP-tagged DN-RhoB, we found that, at 60 min, the percentage of LFA-1 at the cell surface was reduced compared to the initial amount (Fig. 5A). Over the following 60 min, the percentage of LFA-1 at the surface of HSB-2 cells expressing either GFP or GFP-tagged CA-RhoB returned to their initial amounts. In contrast, LFA-1 recycling in HSB-2 cells expressing GFP-tagged DN-RhoB was defective (Fig. 5, A and B). As a negative control, GFP-expressing HSB-2 cells were exposed to mouse antibody against IgG, which internalized over time without recycling back to the surface (fig. S5B).

Fig. 5 Dysfunctional RhoB impairs LFA-1 internalization and recycling in HSB-2 cells.

(A) Flow cytometric analysis of the cell surface expression of LFA-1 over time in sorted HSB-2 cells expressing GFP, GFP-tagged CA-RhoB, or GFP-tagged DN-RhoB. Pooled data normalized to 0 min (start point). Means ± SEM; n = 5 experiments. Statistical significance calculated by ordinary one-way analysis of variance (ANOVA) multiple comparisons test (*P < 0.05) between GFP and GFP-DN-RhoB is indicated. (B) LFA-1 surface expression from (A) was quantified at 120 min after staining. Data are normalized to GFP-expressing cells and are means ± SEM. Statistical significance between GFP and GFPDN-RhoB is indicated. (C and D) Sorted HSB-2 cells expressing GFP, GFP-tagged WT-RhoB, CA-RhoB, or DN-RhoB were attached to ICAM-1 and stained with antibodies against LFA-1 and Rab11 (C) or Rab4 (D). MCC quantification of confocal images of cells stained for LFA-1 and Rab11 (C) or Rab4 (D). Pooled data from three regions averaged per cell from 22 to 35 cells in two independent experiments are shown. Data are means ± SEM. Statistical significance of differences between means of GFP-WT-RhoB and GFP-DN-RhoB (black line) is indicated in (C). (B and C) *P < 0.05, ***P < 0.001, unpaired (two-tailed) t test.

We examined internalization and recycling by biotinylating the cell membrane proteins of HSB-2 cells (transfected with scrambled control siRNA or RhoB-specific siRNA) and allowed the cells to migrate on immobilized ICAM-1. We removed biotin from the cell surface and assessed the internalized biotinylated LFA-1 by immunoprecipitating LFA-1 and using streptavidin–horseradish peroxidase (HRP) to detect biotin bound to the immunoprecipitated LFA-1 on Western blots (fig. S5, C and D). We detected recycling of LFA-1 to the cell surface by reincubating the T lymphocytes on immobilized ICAM-1 after the removal of biotin from the cell surface. We then repeated the removal of biotin from the cell surface, immunoprecipitated LFA-1, and analyzed the samples by Western blotting with streptavidin-HRP. After the first migration on immobilized ICAM-1, most of the internalized biotinylated LFA-1 in RhoB-KD–expressing HSB-2 cells was still detectable, suggesting that, compared to the situation in cells transfected with control siRNA, less LFA-1 left the intracellular compartment of RhoB-KD cells during the reexposure time period on ICAM-1 and consequently did not reappear at the plasma membrane in these cells (fig. S5E). This finding is suggestive of a role for RhoB in the recycling of LFA-1 to the cell surface in HSB-2 cells.

Dysfunctional RhoB impairs the Rab11-mediated intracellular trafficking of LFA-1 in HSB-2 cells

To determine whether RhoB regulated the intracellular trafficking of LFA-1 through an effect on the endosomal trafficking step involving Rab11, we analyzed the association of LFA-1 with Rab11 in HSB-2 cells expressing GFP, GFP-tagged WT-RhoB, GFP-tagged CA-RhoB, or GFP-tagged DN-RhoB. In agreement with our data using freshly isolated T lymphocytes (Fig. 3C), about one-third of LFA-1 associated with Rab11-positive early endosomes in HSB-2 cells expressing GFP or GFP-tagged WT-RhoB (Fig. 5C). A large reduction in the association between LFA-1 and Rab11 was seen in HSB-2 cells expressing DN-RhoB compared to the association in cells expressing WT-RhoB or CA-RhoB. In parallel, we also analyzed the association of Rab11 with WT-RhoB, CA-RhoB, or DN-RhoB in HSB-2 cells. Fluorescence colocalization analysis with MCC quantification showed that Rab11 associated with each fusion protein (fig. S5F) and the extent of the association was similar to that detected between endogenous RhoB and Rab11 in T lymphocytes (median MCC ~ 0.33 ± 0.09) (Fig. 3C). Because we saw a large reduction in the association between DN-RhoB and Rab11, we hypothesized that there was a shift to the Rab4-mediated recycling pathway of LFA-1. Therefore, we investigated the association between LFA-1– and Rab4-positive vesicles and found a very low association between LFA-1 and Rab4 in HSB-2 cells expressing the RhoB fusion proteins (Fig. 5D).

We next wanted to verify whether RhoB-dependent LFA-1 trafficking might also be clathrin-dependent. Confocal imaging and fluorescence colocalization analysis by MCC also showed that most LFA-1 did not associate with clathrin-associated structures (fig. S5, H and I), which perform clathrin-mediated endocytosis (8). We observed no difference in the amount of clathrin that associated with LFA-1 between HSB-2 cells transfected with control siRNA or RhoB-specific siRNA (fig. S5, H and I). Both sets of cells had mean MCC values of <0.2. We also evaluated clathrin-mediated endocytosis by monitoring the abundance of the transferrin receptor, which is a cargo for internalization through clathrin-mediated endocytosis, at the cell surface over time in HSB-2 cells transfected with control siRNA, siRNA targeting RhoB, or siRNA targeting Rab11. When comparing transferrin receptor surface abundance over time by flow cytometry in RhoB-KD cells and control cells, we observed similar reduction and a subsequent recovery back to initial levels. In contrast, the abundance of transferrin at the surface was stable over 150 min in Rab11-KD cells (fig. S5J). In summary, these data suggest that most LFA-1 endocytosis in HSB-2 cells did not involve clathrin and that clathrin-dependent endocytosis of the transferrin receptor was independent of RhoB but depended on Rab11.

Reducing the abundance of RhoB in T lymphocytes impairs the localization of Rab11 during migration

Because RhoB-KD T lymphocytes showed an accumulation of LFA-1 in the uropod, we next investigated the localization of Rab5 (early endosomal marker), Rab7 (late endosomal marker), and Rab11 (recycling endosomal marker) in RhoB-KD T lymphocytes. Confocal microscopy imaging and subsequent fluorescence colocalization analysis by MCC revealed that RhoB-KD T lymphocytes accumulated Rab11 in the rear of the cell (Fig. 6, A and B). We could not see any effect of the loss of RhoB on the localization of Rab5 or Rab7. Similar to our data from HSB-2 cells expressing DN-RhoB (Fig. 5C), MCC analysis indicated that the extent of the association between LFA-1 and Rab11 in RhoB-KD cells was substantially reduced compared to that in control cells (Fig. 6C). In the absence of functional RhoB, T lymphocytes showed not only an accumulation of both LFA-1 and Rab11 in the uropod but also a reduced association between the two proteins.

Fig. 6 Reduced RhoB abundance impairs the localization of Rab11 in migrating T lymphocytes.

(A) Representative confocal images of T lymphocytes transfected with scrambled control or RhoB-specific siRNA, attached to immobilized ICAM-1, and stained with antibodies against Rab11 (left), Rab5 (middle), or Rab7 (right). Dotted line represents cell shape. Left image is grayscale; right image is a rainbow scale of graded fluorescence intensity. Marked uropod cell area (3 × 3 μm) taken for analysis (yellow squares). Scale bar, 5 μm. Images are representative of two experiments with 15 cells per siRNA analyzed per experiment. (B) Mean fluorescence intensity of Rab11 (left), Rab5 (middle), and Rab7 (right) in the uropod. Pooled data of designated cell areas (3 × 3 μm) from 30 cells are shown. Normalized to cells with control siRNA. Means ± SEM; n = 2 experiments. (C) MCC quantification of LFA-1 fluorescence colocalization with Rab11 (left) and vice versa (right) in T lymphocytes transfected with the indicated siRNAs. Pooled data from three regions averaged per cell from 22 cells in two independent experiments. Means ± SEM; n = 2 experiments. **P = 0.01, ***P < 0.001, unpaired (two-tailed) t test.

Reducing the abundance of Rab11 in T lymphocytes impairs LFA-1 recycling and consequently LFA-1–dependent migration

To further define the functional role of Rab11 in LFA-1–mediated T lymphocyte migration, we used Rab11-specific siRNA to reduce its abundance in T lymphocytes (Fig. 7A). Rab11-KD T lymphocytes exhibited reduced migratory distance and a substantially reduced mean cell migration speed compared to those of control cells (Fig. 7, B and C). Similar to RhoB-KD T lymphocytes, most of the Rab11-KD T lymphocytes had their uropods projecting downwards compared to control cells (fig. S6, A and B).

Fig. 7 Reduced Rab11 abundance impairs LFA-1 recycling and consequently LFA-1–dependent migration in T lymphocytes.

(A) Representative Western blots of T lymphocytes transfected with scrambled control or Rab11-specific siRNA were incubated with antibodies against total Rab11 and actin (mean knockdown efficiency, 88.7 ± 3.9%; means ± SEM; n = 3 experiments). (B) T lymphocytes transfected with the indicated siRNAs were incubated on immobilized ICAM-1 before random migration was observed by time-lapse microscopy. Individual cells were tracked and plotted with a common origin. A representative experiment of the pattern tracked by a single cell from 40 cells tracked is shown; n = 2 experiments. (C) Quantification of mean cell migration speed from the cells in experiments described in (B). Means ± SEM. (D) HSB-2 cells transfected with the indicated siRNAs and with biotinylated cell surface proteins were incubated on immobilized ICAM-1. The remaining biotin on surface-located cell membrane receptors was removed with reducing glutathione, and the internalized biotinylated LFA-1 was analyzed by Western blotting analysis of LFA-1 immunoprecipitates with streptavidin-HRP (first and second lanes). Reexposure of biotinylated LFA-1 to the surface was analyzed by reincubating T lymphocytes on immobilized ICAM-1 after the removal of biotinylated membrane receptors. Reexposed biotinylated membrane receptors were then removed with reducing glutathione, and the remaining intracellular biotinylated LFA-1 was subsequently analyzed by Western blotting analysis of LFA-1 immunoprecipitates with streptavidin-HRP (third and fourth lanes). (E) Mean band intensity from experiments similar to that represented in (D) (first and second lanes). Data are pooled and normalized to control. Means ± SEM; n = 3 experiments. (F) Mean band intensity from experiments similar to that represented in (D) (third and fourth lanes). Data are pooled and normalized to HRP intensity after internalization for siCTRL and siRhoB, as shown in (E). Means ± SEM; n = 3 experiments. (G) Representative Western blots from the analysis of isolated cytoplasmic (Cyt) or membrane (Mem) fractions from HSB-2 cells transfected with the indicated siRNAs. Blots were incubated with antibodies against total RhoB and LFA-1. (H) Quantification of the ratio between the mean band intensities of the cytoplasmic and membrane fractions from experiments similar to that shown in (G). Means ± SEM; n = 3 experiments. (I) Representative Western blots from isolated cytoplasmic or membrane fractions of HSB-2 cells transfected with scrambled control or RhoB siRNA and analyzed for total Rab11 and LFA-1. (J) Quantification of the mean band intensity ratio between the cytoplasmic fraction and the membrane fraction in (I). Left, Rab11; right, LFA-1. Pooled data from three independent experiments are shown. Means ± SEM. (H and J) *P < 0.05, **P < 0.01, ***P < 0.001, unpaired (two-tailed) t test.

We also investigated the effect of knocking down Rab11 on LFA-1 internalization and recycling with the surface biotinylation assay. There was no difference between Rab11-KD HSB-2 cells and control cells in terms of the amount of internalized biotinylated LFA-1 detected by Western blotting analysis of LFA-1 immunoprecipitates with streptavidin-HRP (Fig. 7, D and E). However, the reexposure of biotinylated LFA-1 at the cell surface was markedly impaired in Rab11-KD HSB-2 cells compared to that in control cells (Fig. 7, D and F). A similar dysfunction in LFA-1 recycling was also seen in T lymphocytes by the flow cytometry–based study approach. Here, Rab11-KD T lymphocytes and RhoB-KD T lymphocytes exhibited defective recycling of LFA-1 to the cell surface over time (fig. S6C). These findings support the hypothesis that both RhoB and Rab11 play functional roles in LFA-1 recycling in HSB-2 cells. We measured the distribution of LFA-1 between the cytoplasmic and membrane fractions of HSB-2 cells transfected with scrambled control siRNA, RhoB-specific siRNA, or Rab11-specific siRNA. LFA-1 accumulated in the cytoplasm in the Rab11-KD HSB-2 cells compared to the control cells (Fig. 7, G and H). In contrast, LFA-1 distribution remained similar in RhoB-KD HSB-2 cells (fig. S6, D and E). RhoB tended to accumulate less in the cytoplasm in Rab11-KD HSB-2 cells (Fig. 7, G and H). Rab11 shifted to the membrane fraction in RhoB-KD HSB-2 cells (fig. S6, D and E). RhoB and Rab11 appeared to affect the distribution of both LFA-1 and each other within the HSB-2 cells.

DISCUSSION

An essential mechanism regulating the process of cell adhesion and motility in T lymphocytes is the turnover of the β2 integrin LFA-1. Despite this, the mechanisms targeting LFA-1 to specific endocytic compartments have not been extensively investigated. Here, we set out to study the functional role of RhoB in the intracellular transport of LFA-1 in human T lymphocytes. Our data suggest a functional and interactive link between RhoB and LFA-1 in migrating T lymphocytes. T lymphocytes expressing dysfunctional RhoB show not only impaired LFA-1–mediated migration but also impaired LFA-1 recycling and a reduced association between LFA-1 and Rab11. Overall, this study contributes to our understanding of the signaling pathways that mediate the recycling of LFA-1 in T lymphocytes.

RhoB has previously been implicated in cell adhesion (23); however, the exact mechanism involved and its function are unclear. RhoB-deficient murine macrophages show increased β1-dependent cell motility and adhesion on fibronectin but reduced β2-dependent adhesion on ICAM-1 (23). In agreement with this, we found that HSB-2 cells expressing dysfunctional DN-RhoB had impaired adhesion to ICAM-1. Here, we showed that there was a substantial decrease in LFA-1–mediated migration in HSB-2 cells that expressed DN-RhoB or in human T lymphocytes that had reduced amounts of RhoB. Moreover, using a monolayer of endothelial cells in a flow chamber and applying shear stress, we found that the knockdown of RhoB in T lymphocytes impaired their adhesion to endothelial cells and reduced transmigration. We also showed that active WT-RhoB was present in a complex together with LFA-1 and that this interaction increased in response to exposure of the cells to ICAM-1. Together, these findings suggest that RhoB plays a substantial role in integrin-dependent migration. However, we and others have observed that the underlying mechanism seems to vary between different subsets of integrins, various substrates, and different cell types (23).

The functional role of RhoB and its location remain controversial, and different studies suggest that the cellular distribution of RhoB is cell type–specific. In fibroblasts and in Madin-Darby canine kidney (MDCK) epithelial cells, RhoB is found in late endosomes and pre-lysosomal multivesicular bodies (27, 28). Other studies showed that RhoB is localized to the plasma membrane in epithelial MDCK cells (29) and in early endosomes in epithelial HeLa cells (17). In T lymphocytes, we found RhoB at the plasma membrane, in cytosolic vesicles, and in the cytosol. We demonstrated that RhoB, in contrast to LFA-1 and other Rho GTPase family members, such as RhoA, Rac1, and Cdc42 (3033), did not interact with the actin cytoskeleton. Instead, most of the vesicular RhoB was distributed along the microtubule network of migrating T lymphocytes. Microtubules participate in several aspects of cell migration, and one of their major functions is to facilitate polarized vesicular trafficking (34). RhoB is involved in the transition of endosomal vesicles from the peripheral actin cytoskeleton to the microtubular network in epithelial cells. Fernandez-Borja et al. (17) showed that RhoB transiently inhibited internalized cargo from binding to microtubules for further endosomal transport. On the basis of these results, RhoB could potentially function as a transport mediator that enables LFA-1 to transition between different endosomal transporting vesicles in migrating T lymphocytes. In our study of migrating T lymphocytes, we showed that WT-RhoB was dispersed in small clusters on the tubulin network, whereas DN-RhoB accumulated in larger clusters that interfered with the movement of LFA-1 to the correct recycling compartment. This consequently resulted in the altered LFA-1 pattern of expression, leading to an increase in the amount of LFA-1 in the rear of the cell, reduced cell surface LFA-1 amounts, and consequently less adhesion to ICAM-1. Together, these data suggest that RhoB has an important role in intracellular vesicular trafficking (17, 23, 34), although we and others have observed that the localization of RhoB seems to vary between different cell types (17, 2729).

One of the possible mechanisms underlying the involvement of RhoB in adhesion and migratory cell behavior could be the control of cell surface integrin density, because we and Wheeler and Ridley (23) found a decrease in the cell surface abundance of β2 integrin in cells expressing dysfunctional RhoB. This decrease could be explained by either increased degradation or impaired intracellular transport affecting either the internalization or the reexpression of β2 integrin on the plasma membrane. We found that the late endosome– or lysosome-associated Rab7 did not associate with RhoB in migrating T lymphocytes, which suggests that RhoB lacks a role in lysosomal LFA-1 degradation. In contrast, RhoB associated with both EEA-1, the effector of the early endosome–associated Rab5, and with the long recycling endosome–associated Rab11. LFA-1 associates with Rab5 in migrating T lymphocytes and with both Rab5 and Rab11 in migrating neutrophils (8, 10). Consistent with this, our work on migrating T lymphocytes also revealed the enrichment of LFA-1 in Rab5/EEA-1– and Rab11-associated endosomal compartments. We also found RhoB within these LFA-1–containing Rab-associated endosomal compartments specifically associating with internalized LFA-1, which further supports a role for RhoB in LFA-1 recycling. In summary, our results suggest that the observed decrease in β2 integrin abundance on the surface of T cells may be a consequence of the dysfunctional recycling of β2 integrin rather than its increased degradation.

One of the driving forces in cell migration is the mechanism of integrin recycling. Several studies provide evidence for a relationship between cell migration and the intracellular transport of β1 and β3 integrins through Rab-mediated endocytosis pathways (3545). The mechanism underlying the intracellular transport of β2 integrins in migrating leukocytes is less established. The Rap2 GTPase and Gαq/11 proteins promote LFA-1 recycling in migrating lymphocytes, and knockdown of these proteins inhibits cell migration (3, 10). Nishikimi et al. (11) revealed an additional mechanism whereby Mst1 mediates the Rab13-dependent transport of β2 integrin in migrating pro-B BAF cells, in contrast to the β1 integrins used by cells that form focal adhesions, such as fibroblasts (46). Our work suggests that RhoB may be involved in the regulation of intracellular LFA-1 transport in migrating T lymphocytes. Our results indicate that LFA-1 internalization is independent of RhoB in migrating T lymphocytes. In the absence of RhoB, EEA-1 was localized normally and LFA-1 internalization was not affected. LFA-1 was assumed not to use the clathrin-dependent pathway to become internalized in a previous study (8). We confirmed this finding in T lymphocytes and showed that clathrin-dependent transferrin receptor endocytosis is independent of RhoB expression. In contrast, we found that RhoB was critical for the trafficking of LFA-1 back to the cell surface because LFA-1 transport to plasma membrane was reduced in T lymphocytes expressing dysfunctional DN-RhoB.

RhoB-depleted T lymphocytes accumulated both LFA-1 and Rab11 in the uropod. DN-RhoB expression also led to a reduced association between LFA-1 and Rab11, whereas the RhoB-Rab11 association was unaffected. Potentially, intracellular accumulation of LFA-1 in the uropod in the absence of functional RhoB could be the result of disrupted Rab11-mediated endosomal transport. A study showed that in the presence of dysfunctional Rab11, internalized LFA-1 accumulates in the cytoplasm of nonmigrating Chinese hamster ovary cells (8). We suggest that this is also the case in the T lymphocyte cell line, because our experiments revealed increased LFA-1 localization in the cytoplasm in Rab11-KD HSB-2 cells. To support this claim further, we showed that in the absence of functional RhoB, T lymphocytes not only exhibited the accumulation of both LFA-1 and Rab11 in the uropod, as previously discussed, but also displayed a reduced association between both proteins. This could be explained by increased Rab11 localization in the membrane, and together, these findings further suggest that RhoB is controlling Rab11-mediated transport. We also found that neither LFA-1 nor RhoB was associated with short recycling endosome–associated Rab4 in migrating T lymphocytes. T lymphocytes expressing DN-RhoB displayed no increased LFA-1–Rab4 association. In addition, we demonstrated that LFA-1 recycling and consequently LFA-1–dependent migration were dysfunctional in Rab11-KD HSB-2 cells and T lymphocytes, respectively. Thus, Rab11-dependent recycling seems to be a major route for the reentry of LFA-1 into the plasma membrane.

In conclusion, our results suggest an essential role for RhoB in the regulation of β2 integrins. If the presence of functional RhoB is compromised in T lymphocytes, LFA-1 recycling and subsequently LFA-1–mediated cell migration are impaired. We hypothesize that this is a consequence of defective RhoB-mediated, Rab11-directed recycling of LFA-1. Here, in the absence of functional RhoB, LFA-1 is incapable of being guided to Rab11-containing recycling compartments and subsequently disperses from small clusters into larger clusters. This consequently results in alteration of not only LFA-1 but also Rab11 intracellular structures. The lack of appropriate Rab11-mediated recycling further results in reduced amounts of surface LFA-1, less attachment of LFA-1 to ICAM-1, and finally impaired T lymphocyte motility. Many questions still remain unanswered regarding the regulation of LFA-1 recycling. Future studies should elucidate which affinity forms of LFA-1 are being recycled and whether clustered forms of LFA-1 are also internalized and recycled back to the plasma membrane.

MATERIALS AND METHODS

Antibodies, complementary DNA, and reagents

Primary antibodies used in this study were as follows: m38 (pan-LFA-1; a gift from N. Hogg, The Francis Crick Institute, UK), TS2.4 (47, 48), and antibodies against EEA-1 (610457), Rab4 (610888), Rab11 (610657), and GFP (11814460001; all from BD Transduction Laboratories); RhoB (sc180; Santa Cruz Biotechnology); Rab5 (3547S; Cell Signaling Technology); Rab7 (ab137029; Abcam); tubulin (T9026) and actin (A5441; both from Sigma-Aldrich); clathrin (ab2731; Abcam); and IgG (ab2025; Santa Cruz Biotechnology). Secondary antibodies used in this study were as follows: goat antibody against rabbit IgG-HRP (31460; Thermo Fisher Scientific), sheep anti-mouse IgG-HRP (NA9310V; GE Healthcare), STAR 580 and STAR Red (both Abberior), and Alexa Fluor–conjugated antibodies. Alexa Fluor conjugation kits were from Molecular Probes (Thermo Fisher Scientific). Streptavidin-HRP conjugate (RPN1231) was from GE Healthcare. Phalloidin–Alexa Fluor 647 (A22287) and CellTracker Blue CMAC Dye (C2110) were from Molecular Probes (Thermo Fisher Scientific). ICAM-1–Fc was produced as described previously (49). The complementary DNA (cDNA) encoding hemagglutinin-RhoB was provided by Fernandez-Borja et al. (17). The cDNA encoding RhoB was subcloned into the pEGFP-C1 vector (Clontech) using Eco RI/Bam HI restriction sites. The GFP-tagged RhoB Q63L and T19N mutants were generated by point mutagenesis using the Stratagene kit protocol. The following primers were used for mutagenesis: GFP-RhoB (Q63L), 5′-GGACACGGCGGGCCTGGAGGACTACGACC-3′ (forward) and 5′-GGTCGTAGTCCTCCAGGCCCGCCGTGTCC-3′ (reverse); GFP-RhoB (T19N), 5′-GCGCGTGTGGCAAGAACTGCCTGCTGATCGTG-3′ (forward) and 5′-CACGATCAGCAGGCAGTTCTTGCCACACGCGC-3′ (reverse).

Cell isolation and culture

Peripheral blood mononuclear cells (PBMCs) were isolated from single-donor leukocyte concentrates obtained from the Skåne University Hospital. PBMCs were isolated with Lymphoprep (Axis-Shield) and cultured in RPMI media (Gibco) supplemented with 10% fetal calf serum (FCS) (Gibco) and 1% penicillin-streptomycin-glutamine (PAA). Cells were stimulated with phytohemagglutinin (1 μg/ml; Thermo Fisher Scientific) for the first 72 hours and then with interleukin-2 (20 ng/ml; aldesleukin, Novartis). The human T lymphocyte cell line HSB-2 (American Type Culture Collection, CCL-120.1) was maintained in RPMI, 10% FCS, and 1% penicillin-streptomycin-glutamine.

Cell transfection

HSB-2 cells were washed in Opti-MEM (Gibco) and electroporated using a Gene Pulser with Capacitance Extender (Bio-Rad) set at 975 μF and 300 mV. Plasmid (5 μg) or 400 nmol RhoB siRNA (Qiagen or GE Dharmacon) or siCONTROL Non-targeting (Qiagen) was used to transfect 2 × 107 HSB-2 cells. GFP-positive cells were sorted before use with a FACSAria cell sorter (BD Biosciences). Transfection of 1 × 107 primary human T lymphocytes was performed with the Amaxa Nucleofector and Human T Cell Nucleofector Kit using program T-023 (Lonza). The efficiency of knockdown by individual siRNAs in T lymphocytes and HSB-2 cells was evaluated by Western blotting.

Western blotting and immunoprecipitation

Cells were lysed in radioimmunoprecipitation assay cell lysis buffer containing phosphatase and protease inhibitors for 20 min and centrifuged before being resolved by polyacrylamide gel electrophoresis (PAGE) with the Mini-PROTEAN Tetra Cell gel electrophoresis system (Bio-Rad) and hand-casted 8 to 15% polyacrylamide gels. Proteins were transferred to Amersham Hybond LFP 0.2 polyvinylidene difluoride or 0.45 nitrocellulose (GE Healthcare) membranes and blocked with 5% milk in phosphate-buffered saline (PBS)/0.1% Tween 20 or tris-buffered saline/0.1% Tween 20. After incubation with the appropriate primary and secondary antibodies, immunoreactive bands on the blots were visualized by incubation with Clarity Western ECL Substrate (Bio-Rad) followed by exposure to Amersham Hyperfilm ECL (GE Healthcare). For immunoprecipitations, 1 to 3 μg of control or specific antibody were added to cell lysates overnight at 4°C, which was followed by the addition of Dynabeads Protein A and/or G (Thermo Fisher Scientific) and further incubation at 4°C for 1 hour. Beads were washed five times in lysis buffer before elution with 2× SDS-PAGE reducing sample buffer. Lysates and immunoprecipitates were used immediately or stored at −80°C until analysis.

Subcellular location of proteins

The Subcellular Protein Fractionation Kit (78840, Thermo Fisher Scientific) was used according to the manufacturer’s instructions. Isolated cytoplasmic and membrane fractions were subjected to Western blotting as described earlier, and blots were incubated with antibodies against RhoB, Rab11, and LFA-1.

RhoB activation assay

The amount of activated RhoB in cells was measured with the RhoB Activation Assay Kit (STA-403-B, Cell Biolabs) according to the manufacturer’s instructions. Cells suspended in migration buffer were applied to an ICAM-1–coated coverslip and stimulated with CXCL12 (10 μg/μl) for 30 min at 37°C. Unstimulated control cells were incubated in migration buffer alone. Protein lysates were prepared and quantified with the Bio-Rad Protein Assay Dye Reagent Concentrate and standard procedure based on the Bradford dye–binding method. Equal amounts of protein were incubated for 1 hour with Rhotekin RBD Agarose beads at 4°C before being eluted with 2× SDS-PAGE reducing sample buffer and analyzed by Western blotting as described earlier. Membranes were incubated with antibodies against LFA-1 and RhoB.

LFA-1 internalization and reexposure assay

The protocol used was adapted from Fabbri et al. (50) and Stanley et al. (5). Glass coverslips (32 mm) were coated with ICAM-1–Fc (5 μg/ml) in PBS at 4°C overnight and then were blocked with 2.5% bovine serum albumin (BSA). To biotinylate membrane proteins, washed T lymphocytes were resuspended in EZ-link sulfo-NHS-SS-biotin (0.5 mg/ml; 21331, Thermo Fisher Scientific) and incubated on ice for 1 hour. After washing, T lymphocytes were suspended in migration buffer [Hanks’ balanced salt solution (Gibco) with 10 mM Hepes (Gibco)] and applied to an ICAM-1–coated coverslip. Cells were incubated for 45 min at 37°C to allow adhesion and internalization of receptors. To remove membrane-bound biotin, freshly made cold reduced glutathione buffer (46 mM glutathione, 75 mM NaCl, 1 mM EDTA, 1% BSA, and 75 mM NaOH) was added and the cells were incubated on ice for 30 min. To investigate the reexposure of biotinylated LFA-1 on the cell surface, membrane-bound biotin with glutathione buffer was removed and T lymphocytes suspended in migration buffer were placed onto new ICAM-1–coated coverslips and incubated at 37°C for 45 min followed by incubation on ice for 30 min with glutathione buffer to remove reexposed membrane-bound biotin. To detect intracellular biotinylated LFA-1, T lymphocytes were lysed with a standard cell lysis buffer containing 0.2% NP-40 and Complete Mini EDTA-free Protease Inhibitor Cocktail (Roche) for 20 min. After centrifugation, the m38 antibody was added to the supernatant and incubated at 4°C overnight, followed by the addition of Dynabeads Protein A and G (Thermo Fisher Scientific) and a further incubation at 4°C for 1 hour. Beads were washed five times in lysis buffer before being eluted with nonreducing gel sample buffer and analyzed by Western blotting as described earlier. To detect biotinylated LFA-1, blots were incubated with enhanced chemiluminescence streptavidin-HRP conjugate, washed, and visualized. To check for equivalent protein loading between samples, a nonbound immunoprecipitation fraction was also probed with antibody against actin. HRP signal densities were analyzed with ImageJ software.

Random cell migration

Either uncoated μ-Slide IV (Ibidi) slides or MatTek dishes (MatTek Corporation) were coated with ICAM-1–Fc (3 μg/ml) in PBS at 4°C overnight and then blocked with 2% BSA. T lymphocytes (2 × 106/ml) were washed in migration buffer and allowed to adhere to the ICAM-1–Fc–coated bottom of the slide or dish. Random migration was studied with a Zeiss Axiovert 200M inverted microscope using a 20× Plan-Neofluar 0.5 numerical aperture (NA) M27 objective with a Hamamatsu Orca ER camera. Images were captured at a rate of 4 frames/min for 20 min. Individual cells were tracked with Volocity software (Perkin Elmer).

Shear stress assay

The chambers of an IbiTreat μ-Slide IV (Ibidi) were seeded with HUVECs (5 × 106/ml, CellWorks). The μ-Slide was incubated for 24 hours to enable a confluent HUVEC layer to form. The HUVECs were pretreated overnight with tumor necrosis factor–α (25 μg/ml; PeproTech) and then with CXCL12 (1 μg/ml; PeproTech) for 30 min before the assay was performed, as previously described by Evans et al. (51). HSB-2 suspended in migration buffer was applied to the μ-Slide with a shear stress of 1 dyne/cm2 using an automated syringe pump (Aladdin-1000, World Precision Instruments). The activity of interacting T lymphocytes was recorded with a Zeiss Axiovert 200M inverted microscope with a Hamamatsu Orca ER camera using a 10× Plan-Neofluar 0.3 NA objective lens. Time lapses from three separate areas of each flow chamber were recorded at a rate of 60 frames/min for a period of 3 min. Cells were manually scored according to their rate of interaction with the HUVECs. T lymphocytes that adhered for more than 3 s were scored as adherent, and those that had transmigrated below the HUVEC layer were scored as having transendothelially migrated.

Cell attachment assays

The protocol used was adapted from that of Svensson et al. (49). Flat-bottom Immulon-1 96-well plates (Thermo Fisher Scientific) were coated overnight at 4°C with ICAM-1–Fc (3 μg/ml) in PBS and blocked with 2.5% BSA. Sorted cells were incubated with CellTracker Blue CMAC dye for 30 min at 37°C, washed, and resuspended in migration buffer and plated at a density of 5 × 105 cells per well in triplicate wells. The cells were incubated for 30 min in migration buffer supplemented with 5 μM Mn2+ and then measured with a Varioskan plate reader (Thermo Fisher Scientific) with excitation at 353 nm and emission at 466 nm. Nonadherent cells were removed by gentle washing. The adhesion index was calculated according to the following formula: Percentage of adherent cells = (average fluorescence intensity read in the washed wells)/(average fluorescence intensity read in the unwashed cells) × 100%.

Chemotaxis assay

T lymphocytes (5 × 106 cells/ml) were allowed to migrate through 5-μm pore size Transwell insert wells (Corning) coated with ICAM-1–Fc, as previously described (10). The lower wells contained either 600 μl of medium (RPMI 1640 and 0.1% BSA) alone (unstimulated) or medium supplemented with 10 nM CXCL10 or 10 nM CXCL12 (both from PeproTech). After 90 min of incubation at 37°C and 5% CO2, the inserts were discarded, and the numbers of migrating T lymphocytes were counted by flow cytometry after recovery with ice-cold 5 mM EDTA/PBS. All samples were tested in triplicate.

Immunohistochemistry and imaging

Coverslips were coated with ICAM-1–Fc (3 μg/ml) in PBS at 4°C overnight and blocked with 2% BSA. T lymphocytes were washed and allowed to adhere for 20 min before pH-shifted fixation [3% paraformaldehyde (PFA) in 80 mM K-Pipes (pH 6.8) supplemented with 2 mM Mg2+ and 5 mM EGTA, followed by treatment with 3% PFA in 100 mM borax (pH 11)] (10) and permeabilized with 0.1% Triton X-100 in PBS. Autofluorescence was quenched with NaBH4 (1 mg/ml; Sigma-Aldrich). Coverslips were incubated with primary antibody overnight at 4°C in PBS with 0.1% BSA, washed, and incubated with secondary antibody for 45 min at room temperature. Coverslips were mounted with Mowiol (for confocal microscopy) or Abberior Antifade Mount (Abberior, for STED microscopy). Confocal microscopy was performed with a Zeiss LSM 700 Axio Imager M2 system using a 63× Plan-Apo λ/1.40 NA M27 oil objective (Carl Zeiss). Images were collected, processed, and analyzed using Slidebook 6 software (3i Intelligent Imaging Innovations Inc.). STED microscopy was performed on a Quad Scan Super Resolution Microscope system (Abberior Instruments) using a 60× Plan-Apo λ/1.40 NA oil objective (Nikon). Images were collected using Imspector software (Abberior).

Image quantification

Fluorescence intensity quantification was performed with ImageJ software. A mean fluorescence value per designated cell area (2.5 × 2.5 μm) was determined. Quantification of colocalization between different Rab proteins and LFA-1 or RhoB as well as between LFA-1 and RhoB was analyzed according to Dunn et al. (52). Confocal images were processed with an automatic method of local background subtraction, median filtering, and finally by small-value subtraction using Slidebook6 software. The three regions of interest (2 × 2 μm) were analyzed using MCC (Slidebook6, 3i). Results from three regions were averaged per cell and statistically analyzed in Prism 6 (GraphPad Software).

Flow cytometry

Cells were stained in cold fluorescence-activated cell sorting buffer (0.5% BSA and 0.01% sodium azide) supplemented with antibodies (Ts2.4, antibody against IgG or clathrin, depending on the readout), washed, fixed with 3% PFA, and analyzed with an LSR II flow cytometer (BD Biosciences). Data analysis was performed with FlowJo software (Tree Star Incorporated). Sorting was performed with a FACSAria II flow cytometer (BD Biosciences).

Internalization assay

Analysis of LFA-1 internalization was performed as described previously (26) with modifications. Cell surface LFA-1 was labeled with TS2.4 (10 μg/ml) in 60 μl (6 × 105 cells) and incubated for 30 min on ice. After washing with ice-cold migration buffer supplemented with 2 mM Mg2+, cells were resuspended in warm migration buffer and incubated at 37°C for 0 to 150 min. After washing with ice-cold migration buffer, the cells were incubated with Alexa Fluor 647–conjugated donkey antibody against mouse IgG (H+L) for 30 min on ice. After washing and resuspending in ice-cold migration buffer, cells were immediately fixed in 3% PFA and then subjected to flow cytometry analysis with an LSR II flow cytometer (BD Biosciences). For image analysis, cell surface LFA-1 was labeled with TS2.4 (10 μg/ml) in 60 μl (6 × 105 cells) and incubated for 30 min on ice. After washing with ice-cold migration buffer supplemented with 2 mM Mg2+, cells were resuspended in warm migration buffer and allowed to migrate on ICAM-1–Fc coated and blocked with 2% BSA coverslips. Cells were allowed to adhere for 60 min before fixation with warm 4% PFA in migration buffer. Cells were further stained with secondary antibodies as described in the “Immunohistochemistry and imaging” section.

Statistical analysis

Statistical analysis was performed with Prism 6 software (GraphPad Software). The statistical tests used and significant differences are indicated in the legends and figures. *P < 0.05; **P < 0.01; ***P < 0.001.

SUPPLEMENTARY MATERIALS

www.sciencesignaling.org/cgi/content/full/10/509/eaai8629/DC1

Fig. S1. RhoB interacts with LFA-1, and a reduction in RhoB abundance impairs LFA-1–dependent migration in T lymphocytes.

Fig. S2. LFA-1 is present in cytosolic clusters similar to those containing RhoB and tubulin in migrating T lymphocytes.

Fig. S3. Both LFA-1 and RhoB interact with Rab11 in T lymphocytes.

Fig. S4. Active RhoB directs LFA-1 between cytoplasmic and membrane compartments in HSB-2 cells.

Fig. S5. Decreased abundance of functional RhoB impairs LFA-1 internalization and recycling.

Fig. S6. RhoB activity regulates Rab11 localization, and both proteins control LFA-1 recycling in HSB-2 cells.

REFERENCES AND NOTES

Acknowledgments: We thank A. Lellouch for critical reading of the manuscript and N. Hogg for fruitful discussions, critical reading of the manuscript, and antibodies. We also thank M. Fernandez-Borja for providing RhoB cDNA. Funding: This work was supported by Swedish Research Council awards K2010-80P-21592-01-4 and K2010-80X-215917-01-4, Foundation Olle Engquist Byggmästare, IngaBritt & Arne Lundberg Research Foundation, Royal Swedish Academy of Science, Royal Physiographic Society of Lund, Åke Wiberg, Jeanssons Foundation, Kocks Foundation, Per-Eric and Ulla Schybergs Foundation, Gyllenstiernska Krapperup Foundation, Gustav V 80 Jubilee Fund, Österlund Foundation, and Nanna Svartz and Crafoord awards (to L.S.); Anna-Greta Crafoord postdoctoral fellowship and Royal Physiographic Society of Lund (K.P.); and Royal Physiographic Society of Lund (M.S.). Author contributions: M.S. carried out study concept and design, performed experiments and statistical analysis, analyzed and interpreted data, and wrote and revised the manuscript. K.P. and J.L. carried out study concept and design, performed experiments, analyzed and interpreted data, and revised the manuscript. J.P.B., E.S., and H.U.-H. performed experiments. L.S. carried out study concept and design, analyzed and interpreted data, wrote and revised the manuscript, supervised the study, and obtained funding. Competing interests: The authors declare that they have no competing interests.
View Abstract

Navigate This Article