VLA-4 activation at the leading edge of migrating human T cells. A. FITC-conjugated BIO1211 stained the leading edge of T cells during migration on ICAM-1 + CXCL12 (Supplemental video 5). ICAM-1 was used to trigger cell migration through LFA-1:ICAM-1 interaction. Under the condition, active VLA-4, which is not occupied by VCAM-1, can be detected by BIO1211 binding. B. PTX inhibited most of BIO1211 staining of migrating T cells on ICAM-1 + CXCL12, while 1mM MnCl₂ induced BIO1211 staining to T cells. Scale bars, 10 μm.

Dynamic VLA-4 activation at the lamellipodium during cell spreading. A. Time-lapse YFP and FRET images of dynamic TIRF/FRET were acquired from transient VLA-4 FRET sensor/GD25 cell every 10 sec for 10 min during spreading on immobilized VCAM-1 and PLL (Supplemental video 4). Images at 10 min from time-lapse movies of the protruding edge in the highlighted region of the left panels were magnified for comparison of FRET efficiency during cell spreading (middle panels). From the protruding edges highlighted with three dashed lines on FRET images of VCAM-1 and PLL in left panels, kymographs were generated in time-lapse FRET images (right panels). White and black dashed arrows in the right panels depict an angle (θ) of accumulation of active VLA-4 at the spreading edge on VCAM-1 and PLL. A representative cell image was shown from five independent experiments of each condition. D; spreading direction, T; time. The color bar represents FRET efficiency (highest [blue] to lowest [red]). Scale bars, 10 μm. B. The angles of active VLA-4 regions in FRET kymograph (A) were averaged from five individual cells on VCAM-1 and PLL, respectively.

CXCR4 and active Rap1 are localized to the leading edge of migrating human T cells. A. Anti-CXCR4 Ab staining showed that CXCR4 is localized mainly at the leading edge of migrating T cells on VCAM-1 + CXCL12. Scale bar, 20 μm. B. During T cells were migrating on ICAM-1 + CXCL12, BIO1211 was incubated for 5 min, and cells were fixed. Following CXCR4 Ab staining, BIO1211 and CXCR4 staining of cells were visualized. Scale bar, 10 μm. C. T cell expressing Rap1 FRET sensor was tracked during migration on VCAM-1 + CXCL12. Dynamic FRET images are shown every 2 min (Supplemental Video 6). YFP intensity is shown in gray and FRET efficiency is shown in rainbow colors (highest [red] to lowest [blue]). Scale bars, 10 μm. D. FRET efficiency (red) and YFP intensity (blue) of a T cell expressing Rap1 FRET sensor from an image in C (time = 8 min) were profiled along the dashed arrows. E. A kymograph was generated from the highlighted region of the left panel from the time-lapse images of migrating T cells (Supplemental Video 6). D; spreading direction, T; time. Scale bar, 10 μm.

Selective blocking of active VLA-4 is sufficient to inhibit T cell migration on VCAM-1. A. Human T cells were incubated with 4, 10, 200, and 400 nM, as well as 4 nM (+ 1 mM MnCl₂) BIO1211. Percentages of cells bound BIO1211 were measured by flow cytometry. MFI; mean fluorescence intensity. A data set from three independent experiments was shown. *, P < 0.0001 versus control. B. T cell migration trajectories on VCAM-1 + CXCL12 were tracked in the absence or presence of 4 nM and 400 nM BIO1211. T cell migration trajectories on VCAM-1 alone or PLL + CXCL12 were also tracked. 15 - 23 randomly selected cells were represented on the trajectory plots. C. Percentages of migrating cells, normalized to the control condition, were shown under different blocking conditions in B. The data were produced from at least 100 cells of each condition.

Dynamic FRET measurement under TIRF microscopy. A. YFP signal from transient VLA-4 FRET sensor/GD25 cell on VCAM-1 was visualized, and kymographs were generated from the highlighted area of the TIRF images (left two panels) and EPI fluorescence images (right two panels). Scale bars, 5 μm (TIRF image) and 10 μm (EPI image). D; spreading direction, T; time. B. A schematic representation of how the FRET signal is measured in TIRF microscopy. (1): Evanescent wave transmitting through the contact between cell surface and immobilized VCAM-1, (2): Glass surface coated with VCAM-1, (3): Objective lens. θ; the critical angle for a change in refractive index. C. α₄-mCFP/wt β₁, or α₄-mYFP/wt β₁ were transiently expressed, and CFP and YFP crosstalk with the FRET filter set (CFP_Ex/YFP_Em) were calculated for the sensitized emission FRET measurement. D. CFP and YFP intensities of transient VLA-4 FRET sensor/GD25 cells were measured every 10 sec for 30 min on randomly selected regions of interest (ROIs). FRET efficiency was also calculated based on CFP and YFP intensities. E. From individual ROIs selected on VLA-4 FRET sensor/GD25 cell, CFP and YFP images were acquired with a CFP/YFP dual filter for 3 min for FRET calculation by the sensitized emission method, and then CFP images were acquired before and after permanent photobleaching of YFP for FRET calculation by the acceptor photobleaching method. The dashed line depicts the theoretical correlation between these two FRET measurement methods.

VLA-4 is activated at the lamellipodia during cell migration on VCAM-1. A. Human T cells were allowed to adhere to VCAM-1 in the presence or absence of CXCL12, or PLL in the presence of CXCL12. Migrating T cells were tracked over 10 min at 37°C, and videos were generated using DIC images acquired every 10 sec (Supplemental video 1). The left corner of each image is the magnified image of a randomly selected region. Scale bars, 100 μm. B. Human T cells were incubated on VCAM-1 + CXCL12-coated cover slips for 30 min at 37°C. After fixation, dual immunofluorescence labeling with M106 and B44 antibodies was performed and samples were visualized using TIRF microscopy to detect total and activated VCAM-1 at the contact between cells and immobilized VCAM-1. From a fixed cell during migration, M106 and B44 labeling intensities were measured following the dashed arrow and profiled from the tail to the head for intensity comparison on the lower panel. Scale bar, 5 μm. Ratio images were generated by subtracting the background and dividing B44 intensity by M106 intensity. The color bar represents fluorescence intensity ratio (B44/M106). C. From three independent experiments of B, 40 cells were randomly selected and ratio images were generated as above. Cells were carefully analyzed, and scored for the presence of B44 staining enriched at the anterior region, the posterior region, or middle, based on the ratio images (B). Each bar represents the percentage of cells whose B44 staining was dominant over M106 staining (B44/M106 ratio > 1.0). D. Stable GD25 cells expressing wt VLA-4 were allowed to settle on immobilized VCAM-1 for 10 min at 37°C. Spreading cells were tracked over 10 min and DIC images were acquired every 10 sec (Supplemental video 2). Time-lapse images of the highlighted region of a cell spreading (left panel) show lamellipodial protrusion during cell spreading on VCAM-1 (right panels). The arrow shows spreading direction. Scale bar, 20 μm. E. Stable wt VLA-4/GD25 cells were grown on VCAM-1 + CXCL12 overnight. Dual immunofluorescence labeling with M106 and B44 antibodies was performed as described in B. M106 and B44 labeling intensities were measured following the dashed arrow (left panels) and profiled from an edge to an opposite edge for intensity comparison (right panels). Scale bar, 10 μm.

Development of VLA-4 FRET sensor. A. Schematic representation of VLA-4 activation. In the basal, bent conformation, cytoplasmic domains of α₄ and β₁ subunits of VLA-4 are in close proximity. In this case, energy transfer occurs between CFP (α₄-mCFP) and YFP (β₁-mYFP). In the active, extended conformation, there is low or no FRET because of the distal location of the cytoplasmic domains. Ex: extracellular domain, TM: transmembrane, Cy: cytoplasmic domain. B. Inter-subunit FRET measurements in GD25 cells transiently expressing VLA-4 FRET construct pairs containing different linker lengths. *, P < 0.0001 versus 22_5 (n = 10 cells with acceptor intensity of 100 to 400). C. GD25 cells were transiently transfected with α₄-mCFP, α₄-mYFP and wt β₁. Interheterodimer FRET between neighboring α₄ and β₁ integrins for individual cells (diamonds) was fit to the saturable one-site binding model E% = E%maxF/F(F + K), where FRET efficiency (E%) is a hyperbolic function of the YFP acceptor intensity (F), and K is analogous to a dissociation constant. The nonlinear least squares regression fits of FRET efficiency between α₄-mCFP/wt β₁ and α₄-mYFP/wt β₁ (red curve) yielded K = 301, showing little association. D. The cell surface expression levels of VLA-4 FRET sensor (transient) and wt VLA-4 (stable) on GD25 were determined by flow cytometry. Parental GD25 cells were used as control. E. Whole cell lysates of non-transfected control, α₄-mCFP-expressing, β₁-mYFP-expressing and C/YFP-expressing 293T cells were subjected to SDS-PAGE and immuoblotting with polyclonal anti-GFP Ab that has cross-reactivity to CFP and YFP. The band at around 75 kDa is a non-specific band detected by anti-GFP Ab. We carried out the transfection of 293T cells, since we were unable to detect protein bands from the immunoblotting with GD25 cells that were used as a host cell for transfection for other functional studies. F. FRET efficiency of VLA-4 FRET sensor transiently expressed in GD25 cells was measured after stimulation with 1mM MnCl₂, 10 μg/ml β₁ activating Ab (TS2/16), and both of them. *, P < 0.0001 versus control without stimulation (n = 10 cells with acceptor intensity of 100 to 400). G. Human T cells were activated with 1mM MnCl₂ in the absence and presence of VCAM-1 or ICAM-1. The cells then were stained with B44 and KIM127 antibodies, and then measured by flow cytometry to detect activation of VLA-4 and LFA-1, respectively. Control was parental T cells without stimulation. MFI; mean fluorescence intensity. H. DIC, CFP and YFP images were shown every 10 min during spreading of GD25 cell transiently expressing VLA-4 FRET sensor on immobilized VCAM-1, (Supplemental video 3). Scale bar, 20 μm. I. VLA-4 FRET sensor/GD25 were treated with 1 mM MnCl₂ and incubated with soluble human VCAM-1_Ig fusion protein. Cells were fixed and then labeled with Alexa Fluor 647-goat anti-human IgG. DIC, VCAM-1 (Alexa Fluor 647), CFP and YFP images were acquired. The arrows represent that a cell expressing only α₄-mCFP without β₁-mYFP does not bind VCAM-1. Scale bars, 10 μm. J. From three independent experiments in I and with stable wt VLA-4/GD25 cells, more than 100 cells were randomly selected and carefully analyzed to count the percentages of cells that bound VCAM-1.