Annexin A2 Limits Neutrophil Transendothelial Migration by Organizing the Spatial Distribution of ICAM-1

This information is current as Niels Heemskerk, Mohammed Asimuddin, Chantal Oort, Jos of September 26, 2021. van Rijssel and Jaap D. van Buul J Immunol 2016; 196:2767-2778; Prepublished online 10 February 2016; doi: 10.4049/jimmunol.1501322

http://www.jimmunol.org/content/196/6/2767 Downloaded from

Supplementary http://www.jimmunol.org/content/suppl/2016/02/09/jimmunol.150132 Material 2.DCSupplemental http://www.jimmunol.org/ References This article cites 50 articles, 36 of which you can access for free at: http://www.jimmunol.org/content/196/6/2767.full#ref-list-1

Why The JI? Submit online.

• Rapid Reviews! 30 days* from submission to initial decision by guest on September 26, 2021 • No Triage! Every submission reviewed by practicing scientists

• Fast Publication! 4 weeks from acceptance to publication

*average

Subscription Information about subscribing to The Journal of Immunology is online at: http://jimmunol.org/subscription Permissions Submit copyright permission requests at: http://www.aai.org/About/Publications/JI/copyright.html Email Alerts Receive free email-alerts when new articles cite this article. Sign up at: http://jimmunol.org/alerts

The Journal of Immunology is published twice each month by The American Association of Immunologists, Inc., 1451 Rockville Pike, Suite 650, Rockville, MD 20852 Copyright © 2016 by The American Association of Immunologists, Inc. All rights reserved. Print ISSN: 0022-1767 Online ISSN: 1550-6606. The Journal of Immunology

Annexin A2 Limits Neutrophil Transendothelial Migration by Organizing the Spatial Distribution of ICAM-1

Niels Heemskerk, Mohammed Asimuddin, Chantal Oort, Jos van Rijssel, and Jaap D. van Buul

ICAM-1 is required for firm adhesion of leukocytes to the endothelium. However, how the spatial organization of endothelial ICAM- 1 regulates leukocyte adhesion is not well understood. In this study, we identified the calcium-effector protein annexin A2 as a novel binding partner for ICAM-1. ICAM-1 clustering promotes the ICAM-1–annexin A2 interaction and induces translocation of ICAM-1 into caveolin-1–rich membrane domains. Depletion of endothelial annexin A2 using RNA interference enhances ICAM-1 membrane mobility and prevents the translocation of ICAM-1 into caveolin-1–rich membrane domains. Surprisingly, this results in increased neutrophil adhesion and transendothelial migration under flow conditions and reduced crawling time, velocity, and lateral migration distance of neutrophils on the endothelium. In conclusion, our data show that annexin A2 limits neutrophil Downloaded from transendothelial migration by organizing the spatial distribution of ICAM-1. The Journal of Immunology, 2016, 196: 2767–2778.

ntracellular adhesion molecule-1 (ICAM-1) at the endothelial in neutrophil transmigration because it has been shown to be cell (EC) surface functions as an adhesive ligand for neu- rapidly coclustered into a ringlike structure around transmigrating I trophil b2-integrins, supporting the rolling, arrest, crawling, neutrophils (12, 13). Because the intracellular tail of ICAM-1 con- and transmigration of neutrophils in vivo (1–6). However, to what tains no identified signaling domains, activation of intracellular sig- http://www.jimmunol.org/ extent ICAM-1 is specifically involved and how ICAM-1 func- naling is thought to be mediated by receptor clustering (14, 15). tionally facilitates each distinct step during neutrophil transmi- Clustering of ICAM-1 using anti–ICAM-1 Abs is commonly gration has been controversial. Depletion of the intracellular used to study leukocyte induced ICAM-1 signaling. It has been domain of ICAM-1 abolishes T lymphocyte, as well as neutrophil reported that Ab-induced clustering of ICAM-1 initiates a rise in transmigration, but not adhesion (7–9). This work established the intracellular calcium, activates small Rho GTPases, Protein kinase importance of the intracellular domain of ICAM-1 as signaling C, Src family protein kinases, eNOS and MAPK (7, 15–18), and molecule in orchestrating leukocyte transmigration, but also induces the recruitment of several membrane-actin binding pro- showed the importance of the extracellular domain of ICAM-1 in teins to the intracellular tail of ICAM-1 (14, 19–22). Moreover, by guest on September 26, 2021 leukocyte adhesion. Studies that used LFA-1– (aLb2,CD11a/ Ab-induced cross-linking of ICAM-1 has been shown to translo- CD18) and MAC-1 (aMb2,CD11b/CD18)–deficient mice showed cate ICAM-1 to F-actin– and caveolin-1–rich areas at the EC that neutrophil arrest depends mainly on LFA-1–ICAM-1 in- periphery (23). This translocation has been suggested to link the teractions, whereas neutrophil crawling requires MAC-1–ICAM-1 sequential steps of leukocyte adhesion and transcellular transmi- interactions (2, 3). Moreover, the binding site of ICAM-1 gration (23). Studies that examined ICAM-1 surface distribution for MAC-1, but not LFA-1, is regulated by glycosylation, which showed that depletion or blockage of the intracellular tail of may provide a mechanistic explanation for sequential roles in ICAM-1, using cell-penetratin-ICAM-1 peptides competitive for ICAM-1–mediated arrest and crawling (10, 11). In addition to the RKIKK motif, induced a more homogeneous cell-surface LFA-1–ICAM-1 interactions in mediating neutrophil arrest, the distribution and loss of ICAM-1+ microvilli (24). ICAM-1 can LFA-1–ICAM-1 interaction has also been reported to be involved form at least three different topologies on the cell surface through dimerization of domain 1 that does not interfere with LFA-1 or MAC-1 binding (25). Moreover, structural basis of ICAM-1 di- Department of Molecular Cell Biology, Sanquin Research and Landsteiner Labora- merization on the cell surface provides the existence of preformed tory, Academic Medical Center, University of Amsterdam, 1066 CX Amsterdam, the ICAM-1 cis dimers on the cell membrane (26). Netherlands In line with this, Barreiro and colleagues (27, 28) showed that ORCIDs: 0000-0003-2340-537X (N.H.); 0000-0001-8077-1371 (J.v.R.); 0000-0003- 0054-7949 (J.D.v.B.). adhesive molecules such as ICAM-1 are physically organized Received for publication June 16, 2015. Accepted for publication January 5, 2016. and compartmentalized in preformed tetraspanin nanoplatforms. Those studies suggest the existence of multiple ICAM-1 com- This work was supported by Landsteiner Foundation for Blood Transfusion Research Fellowship 1028 (to J.D.v.B.). plexes with different protein content, spatial organization, or Address correspondence and reprint requests to Dr. Jaap D. van Buul, Department of subcellular localization. However, little is known about how the Molecular Cell Biology, Sanquin Research and Landsteiner Laboratory, Academic spatial organization of ICAM-1 affects its adhesive function and Medical Center, University of Amsterdam, Plesmanlaan 125, 1066 CX Amsterdam, what proteins regulate the distribution of ICAM-1 at the EC the Netherlands. E-mail address: [email protected] surface. The online version of this article contains supplemental material. In this study, we investigated the spatial organization of ICAM-1 Abbreviations used in this article: BG, background; cp, circular permutated; EC, endothelial cell; ECFP, enhanced cyan fluorescent protein; FN, fibronectin; FRET, at the EC surface and found annexin A2 as a novel binding partner fluorescence resonance energy transfer; HUVEC, human umbilical vein EC; for ICAM-1. Our data show that annexin A2 negatively regulates RhoGDI, Rho GDP-dissociation inhibitor; ROI, region of interest; shRNA, short the adhesion, crawling, and subsequent transmigration of neutro- hairpin RNA; TEM, transendothelial migration. phils across the endothelium by controlling the spatial organization Copyright Ó 2016 by The American Association of Immunologists, Inc. 0022-1767/16/$30.00 of ICAM-1. www.jimmunol.org/cgi/doi/10.4049/jimmunol.1501322 2768 ANNEXIN A2 CONTROLS ICAM-1 FUNCTION

Materials and Methods Pull-out assay and immunoprecipitation DNA and RNA constructs A total of 1.2 mg/ml dynabeads goat-x-ms IgG (Dynal; Invitrogen) per YCam3.6 in a mammalian expression vector, based on a clontech-style condition was washed once with 1 ml buffer 1 containing PBS + 2 mM pEGFP-C1 backbone and driving expression from a CMV promotor, was EDTA and 0.1% BSA (Millipore) using a magnetic holder. Dynabeads were provided by Joachim Goedhart (Swammerdam Institute for Life Sciences, coated with 1.6 mg anti–ICAM-1 CD54 (BBIG-I1/IIC81, cat. no. BBA9; Amsterdam, the Netherlands). In brief, the single-chain Ycam3.6 biosensor R&D Systems) Abs per condition and incubated head-over-head at 4˚C for 45 min. The beads were washed twice using buffer 1 and resuspended in consists of an enhanced cyan fluorescent protein (ECFP), calmodulin 2+ E104Q, the calmodulin-binding peptide M13p, and a circular permu- PBS containing 0.5 MgCl2 and 1 mM CaCl2. Overnight TNF-a–treated tated (cp)-Venus. Binding of calcium makes calmodulin wrap around (10 ng/ml) HUVECs (2–5 million cells) were pretreated with 1 mM ion- the M13 domain, increasing the fluorescence resonance energy transfer omycin (I-24222; Invitrogen) for 10 min. A total of 1.2 mg/ml dynabeads per condition was used to cluster ICAM-1 for 10–30 min. Cells were (FRET) between the flanking FPs (29). Short hairpin RNA (shRNA) in 2+ pLKO.1 targeting annexin A2 (A9) (TRCN 56145), annexin A2 (G5) washed once on ice using PBS . Next, cells were lysed in 1 ml cold pH7.4 (TRCN 56144) were purchased from Sigma-Aldrich mission library. radioimmunoprecipitation assay buffer containing 50 mM Tris, 100 mM ICAM-1–GFP was a kind gift from Dr. F. Sanchez-Madrid (University NaCl, 10 mM MgCl2, 1% Tergitol-type NP-40, 10% glycerol, 0.1% SDS, of Madrid, Madrid, Spain). pE AnnexinA2-GFP (N3) was purchased 1% DOC (Sigma-Aldrich), DNAse inhibitor, and protease phosphatase from Addgene. inhibitor mixture for 5 min. Cells were scraped together and lysates were transferred to a new tube. Then, ICAM-1–coated dynabeads were added to Abs nonclustered control cells. Fifty microliters whole cell lysate was taken from all conditions. Beads and cell lysates were subsequently incubated Polyclonal rabbit Abs against ICAM-1 for Western blot were purchased head-over-head for 1–2 h at 4˚C. Next, cells were washed twice with from Santa Cruz Biotechnology (cat. no. SC-7891). Polyclonal rabbit Ab radioimmunoprecipitation assay buffer and three times with Tergitol-type 3 against annexin A2 (H50)(SC-9061) and the Alexa Fluor 405 monoclonal NP-40 lysis buffer. Beads were resuspended in 30 ml2 SDS sample Downloaded from mouse Ab against ICAM-1 CD54 (15.2)(Sc-107 AF405) were purchased buffer and assessed by Western blotting. from Santa Cruz (Bio-Connect). Monoclonal ICAM-1 CD54 (BBIG-I1/ IIC81) Abs used for ICAM-1 clustering were purchased from R&D Sys- Western blotting tems (cat. no. BBA9). Polyclonal goat anti-mouse Fcg IgG was purchased Cells were washed twice with PBS and lysed with 95˚C SDS sample buffer from Jackson (cat. no. 115-005-071). mAb against annexin A2 was a kind containing 4% 2-ME. Samples were boiled at 95˚C for 4 min to denature gift from Dr. Gerke (Institute of Medical Biochemistry ZMBE Center for proteins. Proteins were separated on 4–12% NuPAGE Novex Bis-Tris gels Molecular Biology of Munich, Germany). Rabbit monoclonal annexin A2

(Invitrogen), transferred to Immobilon-PVDF transfer membranes (Milli- http://www.jimmunol.org/ (D11G2) was purchased from Cell Signaling (BIOKE´ ) (8235). Monoclonal pore, Billerica, MA), and subsequently blocked with 2.5% (w/v) BSA mouse Ab against Actin (AC-40) was purchased from Sigma (cat. no. in TBST for 60 min. The immunoblots were analyzed using primary A3853). The Alexa Fluor 405 goat anti-rabbit IgG (cat. no. A31556), Abs incubated overnight at 4˚C and secondary Abs linked to HRP (GE Alexa Fluor 647 chicken anti-goat IgG (cat. no. A21469), Alexa Fluor 488 Healthcare); after each step immunoblots were washed six times with chicken anti-rabbit IgG (cat. no. A21441), and Texas Red 568 phalloidin TBST. Signals were visualized by ECL and light-sensitive films (GE (cat. no. T7471) were purchased from Invitrogen. Alexa Fluor 405 phal- Healthcare). loidin was purchased from Promokine (cat. no. PK-PF405-7-01). Poly- clonal rabbit Ab against caveolin-1 was purchased from BD Biosciences FRET measurements: microscopy, image acquisition, and (cat. no. 610059). Mouse monoclonal actin Abs were purchased from analyses Sigma-Aldrich (cat. no. A3853). Mouse monoclonal human ezrin (18) was purchased from Transduction Laboratories (BD) (cat. no. 610602). Mouse We use a Zeiss Observer Z1 microscope with 403 NA 1.3 oil immersion by guest on September 26, 2021 monoclonal Rho GDP-dissociation inhibitor (RhoGDI) Abs were pur- objective, a HXP 120 V excitation light source, a Chroma 510 DCSP di- chased from Transduction Laboratories (cat. no. 610255), mouse mono- chroic splitter, and two Hamamatsu ORCA-R2 digital CCD cameras for clonal Abs against VCAM-1 were purchased from R&D Systems (cat. no. simultaneous monitoring of ECFP and Venus emissions. Image acquisition 2090), and mouse monoclonal filamin A Abs were purchased from Serotec was performed using Zeiss Zen 2011 microscope software. The lowest (cat. no. MCA464S). Secondary HRP-conjugated goat anti-mouse, swine achievable HXP excitation power, through a FRET filter cube [Exciter ET anti-rabbit Abs were purchased from Dako (Heverlee, Belgium). All Abs 436/203, and 455 DCLP dichroic mirror (Chroma), the emission filter is were used according to the manufacturer’s protocol. removed], was used to excite the ECFP donor. The emission is directed to the left-side port by a 100% mirror, to an attached dual camera adaptor Cell cultures, treatment, and transfections (Zeiss) controlling a 510 DCSP dichroic mirror. Emission wavelengths between 455 and 510 nm are directed to an emission filter (ET 480/40; Pooled human umbilical vein ECs (HUVECs), purchased from Lonza Chroma) and then captured by the left Hamamatsu ORCA-R2 camera (P938, cat. no. C2519A), were cultured on fibronectin-coated dishes in resulting in ECFP image acquisition. The emission wavelength $510 nm EGM-2 medium and supplemented with SingleQuots (Lonza, Verviers, are directed to a six-position LEP filter wheel (Ludl Electronic Products) Belgium). HUVECs were cultured until passage 9. Human embryonic placed in front of the second right Hamamatsu ORCA-R2 camera. Position kidney-293T were maintained in DMEM (Invitrogen, Breda, the Nether- 1 in the emission filter wheel is equipped with an ET 540/40m used for lands) containing 10% (v/v) heat-inactivated FCS (Invitrogen), 300 mg/ml 3 Venus image acquisition. Position 2 is left empty to allow mCherry image L-glutamine, 100 U/ml penicillin and streptomycin, and 1 sodium py- acquisition (EX BP 572/25, BS FT 590, EM BP 629/62; Zeiss). The LEP ruvate (Invitrogen). Cells were cultured at 37˚C and 5% CO2. HUVECs filter wheel is controlled by the MAC 6000 controller system (Ludl a were pretreated with 10 ng/ml rTNF- (PeproTech, Rocky Hill, NJ) 24 h Electronic Products). Exposure time of the differential interference con- before each leukocyte transendothelial migration (TEM) experiment, trast image was set to 76 ms, and exposure time of simultaneous ECFP and m pretreated with 1 M ionomycin (I-24222; Invitrogen) for periods as in- Venus acquisition was set to 6800 ms; images were subsequently recorded dicated. Cells were transfected with the expression vectors according to the every 5 s by two 1344 (H) 3 1024 (V) Hamamatsu ORCA-R2 digital CCD manufacturer’s protocol with Trans IT-LT1 (Myrus, Madison, WI). Len- cameras (2 3 2 binning) for periods as indicated. Offline ratio analysis tiviral constructs were packaged into lentivirus in human embryonic kid- between ECFP and Venus images was processed using the MBF ImageJ ney-293T cells by means of third generation lentiviral packaging plasmids collection (Tony Collins). The raw ECFP and Venus image stacks were (30). Lentivirus containing supernatant was harvested on days 2 and 3 after background (BG) corrected using the plug-in “region of interest (ROI), BG transfection. Lentivirus was concentrated by Lenti-X concentrator (Clon- subtraction from ROI.” Then, the ECFP and Venus stacks were aligned tech; cat. no. 631232). Transduced target cells were used for assays after using the registration plug-in “Registration, MultiStackReg.” A smooth 72 h. ICAM-1–GFP was delivered into HUVECs by adenoviral trans- filter was applied to both image stacks to improve image quality by re- duction 48 h before imaging. ducing the noise. Next, both image stacks were converted to a 32-bit image Pull-down assay format, required for subsequent masking. A threshold was applied exclu- sively to the Venus image stack, converting the BG pixels to “not a Synthetic, biotinylated peptides encoding the intracellular domains of number.” It allows elimination of artifacts in ratio image stemming from human ICAM-1 and VCAM-1 were used in pull-down assays as previously the BG noise. Finally, the Venus/ECFP ratio was calculated using the plug- described (15). The following sequences were used: ICAM-1: NH2- in “Ratio Plus,” and a custom lookup table was applied to generate a color- RQRKIKKYRLQQAQKGTPMKPNTQATPP-COOH; VCAM-1: IIY- coded image illustrating the high red and low blue activities. Note that FARKANMKGSYSLVEAQKSKV-COOH. some of the plug-ins, namely MultiStackReg, and Ratio Plus are not in- The Journal of Immunology 2769 cluded in the basic MBF ImageJ collection and should be downloaded a–stimulated (10 ng/ml) HUVEC monolayer for 30 min. ICAM-1 was from the plug-in page in the ImageJ Web site (http://rsb.info.nih.gov/ij/ subsequently cross-linked using secondary goat anti-mouse IgG Abs for plugins/index.html). Normalized intensity graphs, the intensity of an ROI another 30 min. Next, procedures were all carried out on ice. Two 10-cm2 and the BG of the raw ECFP and Venus image stacks, were measured using confluent HUVECs were washed and scraped into base buffer (20 mM the plug-in ROI, Multi Measure in ImageJ. The raw ECFP and Venus Tris-HCL, pH 7.8, 250 mM sucrose) to which had been added 1 mM CaCl2 intensities were BG subtracted using the equation ECFP = (ECFP raw 2 and 1 mM MgCl2. Cells were pelleted by centrifugation for 2 min at 250 3 g BG); subsequently, Venus only was corrected for bleed-through using the and resuspended in 1 ml base buffer containing 1 mM CaCl2 and 1 mM equation VenusC = (Venus) 2 (0.5748*ECFP), and normalized using the MgCl2 and protease inhibitors (Invitrogen). The cells were then lysed by equation ECFPNorm = ECFP/average (ECFP 3–13), and finally the Venus/ passage through a 22 g 3 3-inch needle 203. Lysates were centrifuged at ECFP ratio was calculated using the equation Venus/ECFP = VenusNorm/ 1000 3 g for 10 min. The resulting postnuclear supernatant was collected ECFPNorm using Excel. and transferred to a separate tube. The pellet was again lysed by the addition of 1 ml base buffer, followed by sheering 203 through a needle and Immunofluorescence staining syringe. After centrifugation at 1000 3 g for 10 min, the second post- HUVECs were cultured on fibronectin (FN)-coated glass bottom (14–30 nuclear supernatant was combined with the first. An equal volume (2 ml) mm) until confluency. After treatment, cells were washed with cold PBS, of base buffer containing 50% OptiPrep (Sigma-Aldrich, St. Louis, MO) containing 1 mM CaCl and 0.5 mM MgCl , and fixed in 4% (v/v) was added to the combined postnuclear supernatants and placed in the 2 2 bottom of a 12-ml centrifuge tube. An 8-ml gradient of 0–18% OptiPrep in formaldehyde for 10 min. After fixation, cells were permeabilized in base buffer was poured on top of the lysate, which was now 25% in PBS supplemented with 0.5% (v/v) Triton X-100 for 10 min followed by a 3 g blocking step in PBS supplemented with 2.5% (w/v) BSA. Cells were OptiPrep. Gradients were centrifuged for 90 min at 52,000 using an incubated with primary and secondary Abs, and after each step washed SW-41 rotor in a Beckman ultracentrifuge. After centrifugation, gradients with PBS and mounted on microscope glasses using Mowiol. Image were fractionated into 0.8-ml fractions, and the distribution of various acquisition was performed on a confocal laser scanning microscope proteins was assessed by Western blotting. This method was developed by Macdonald and Pike (31). Downloaded from (LSM510/Meta; Carl Zeiss Micro-Imaging) using a voxel size of 0.06 3 0.06 3 0.48 mm and a 633 NA 1.4 oil immersion objective. Line profile plots were generated in ImageJ using the plot profile plug-in. ICAM-1 mobility assays Detergent-free membrane fractionation Live-cell imaging was conducted with a ZEISS LSM510 confocal microscope (Carl Zeiss MicroImaging) using ZEN 2007 (Carl Zeiss Micro- HUVECs were transduced with shRNA control or shRNA targeting annexin Imaging) image acquisition software with appropriate filter settings. To A2 and left for 72 h. Anti–ICAM-1 Abs were allowed to bind to 4-h TNF- assess the mobility of ICAM-1 on the plasma membrane, we bleached a small fraction of ICAM-1–GFP+ plasma membrane and measured the http://www.jimmunol.org/ fluorescent recovery of ICAM-1–GFP into this region for 200 s. Data were further analyzed in ImageJ. Intensities were normalized with the following equation: normalized intensity = (intensity 2 intensityAfterbleach)/ (intensityPrebleach 2 intensityAfterbleach)*100. To examine ICAM-1 mobility, we incubated TNF-a–treated ECs with 10-mm polystyrene beads (Polysciences) that were coated with an mAb against ICAM-1 CD54 (BBIG-I1/IIC81, cat. no. BBA9; R&D Systems) for 30 min. Percentage ICAM-1–GFP+ rings were quantified as the number of ICAM-1–GFP+ rings divided by the number of beads per ECs. by guest on September 26, 2021 Neutrophil isolation Neutrophils were isolated from whole blood derived from healthy donors. Whole blood was diluted (1:1) with 5% (v/v) trisodium citrate in PBS. Diluted whole blood was pipetted carefully on 12.5 ml Percoll (room temperature) 1.076 g/ml. Tubes were centrifuged (Rotanta 96R) at 2000 rpm, slow start, low brake for 20 min. Ring fraction containing lymphocytes and monocytes was discarded. After erythrocyte lysis in an ice-cold isotonic lysis buffer [155 mM NH4CL, 10 mM KHCO3, 0.1 mM EDTA, pH 7.4 in Milli-Q (Millipore)], neutrophils were centrifuged at 1500 rpm for 5 min at 4˚C, incubated once with lysis buffer for 5 min on ice, centrifuged again at 1500 rpm for 5 min at 4˚C, washed once with PBS, centrifuged again at 1500 rpm for 5 min at 4˚C, resuspended in HEPES medium [20 mM HEPES, 132 mM NaCl, 6 mM KCL, 1 mM CaCL2, 1 mM MgSO4, 1.2 mM K2HPO4, 5 mM glucose (all from Sigma- Aldrich), and 0.4% (w/v) human serum albumin (Sanquin Reagents), pH 7.4] and kept at room temperature for not .4 h until use. Neutrophil counts were determined by cell counter (Casey). Neutrophil TEM under physiological flow HUVECs were cultured to 70% confluence in FN-coated six-well plate and transduced with lentivirus containing shRNA targeting Annexin A2 and shRNA control for 24 h. HUVECs were cultured in a FN-coated ibidi m-slide VI0.4 (ibidi, Munich, Germany) the day before the experiment was FIGURE 1. The intracellular tail of ICAM-1, but not VCAM-1, interacts executed and stimulated overnight with TNF-a (10 ng/ml). Freshly iso- with annexin A2. (A) Immunoblot analysis of protein extracts prepared lated neutrophils were resuspended at 1 3 106 cells/ml in HEPES medium from TNF-a–stimulated ECs that were incubated with peptides encoding and were incubated for 30 min at 37˚C. Cultured HUVECs in ibidi flow the intracellular tail (PD: C-term) of ICAM-1 or VCAM-1. Extracts were chambers were connected to a perfusion system and exposed to 0.5 ml/min 2 probed with Abs directed against annexin A2 and filamin A. Streptavidin HEPES shear flow for 10 min (0.8 dyne/cm ). Neutrophils were subse- beads were used as a negative control. (B) Immunoprecipitation of ICAM- quently injected into the perfusion system, and real-time leukocyte–en- 1 using magnetic beads coated with anti–ICAM-1 Abs. Anti–VCAM-1 or dothelial interactions were recorded for 20 min by a Zeiss Observer Z1 microscope using a 403 NA 1.3 oil immersion objective. All live imaging anti-IgG isotype control Ab-coated beads were used as controls. Beads was performed at 37˚C in the presence of 5% CO2. Transmigrated neu- were incubated 30 min at the cell surface of TNF-a–stimulated ECs before trophils were distinguished from those adhering to the apical surface of the lysis. Extracts were probed with Abs directed against ICAM-1, VCAM-1, endothelium by their transition from bright to phase-dark morphology. filamin A, and annexin A2. Data are from three experiments (A and B). Percentage adherent or transmigrated neutrophils were manually quantified WCL, whole cell lysate. using the ImageJ plug-in Cell Counter (type 1, adherent cells, type 2, 2770 ANNEXIN A2 CONTROLS ICAM-1 FUNCTION Downloaded from http://www.jimmunol.org/ by guest on September 26, 2021

FIGURE 2. Rolling neutrophils transiently increase intracellular calcium levels in ECs driving annexin A2 to associate with the plasma membrane and ICAM-1. (A) Schematic illustration of the Ycam3.6 calcium sensor design containing ECFP (blue), calmodulin E104Q (gray), the calmodulin-binding peptide M13p (white), and a cp-Venus (yellow). (B) Time-lapse Venus/CFP ratio images of Ycam3.6 calcium biosensor simultaneously recorded with an epifluorescent microscope showing spatiotemporal calcium measurements during the onset of flow (0.9 dyne/cm2) starting after 30 s, and upon neutrophil rolling, adhesion, and transmigration (arrow). Calibration bar shows the increase in intracellular calcium concentrations (red) relative to basal calcium concentration (blue). Scale bar, 150 mm. (C) Quantification of temporal intracellular calcium concentration in multiple ECs during the onset of flow, and neutrophil rolling, adhesion, and transmigration. Data represent mean and SEM of three experiments including at least 12 cells per experiment. (D) Confocal imaging of cytosolic annexin A2–GFP translocation to the plasma membrane after ionomycin stimulation. Scale bar, 20 mm. (E) Immunoblot analysis of protein extracts prepared from TNF-a–stimulated ECs that were incubated with anti–ICAM-1–coated beads for (Figure legend continues) The Journal of Immunology 2771 transmigrated cells). Neutrophil rolling velocity, crawling time, crawling calcium makes calmodulin wrap around the M13 domain, increas- velocity, and crawling distance were quantified using the ImageJ plug-in ing the FRET between the flanking FPs (29). ECs transfected with manual tracking. Ycam3.6 biosensor were grown to confluence in FN-coated perfusion Statistics chambers, stimulated with TNF-a overnight, and subsequently ex- 2 Filamin and annexin A2 binding to anti–ICAM-1–coated beads was tested posed to shear stress (0.8 dyne/cm ). Intracellular changes in calcium using a one-way ANOVA assuming no matching or pairing, comparing the levels were measured by donor/acceptor ratio imaging and showed mean of each column with the mean of every other column, that were that the initial onset of flow rapidly triggered the transient influx of corrected for multiple comparisons by Tukey multiple comparisons test, calcium, as was also shown by others (33–35) (Supplemental Fig. 1). with a single pooled variance. The Student t test performed statistical comparison between experimental groups. A two-tailed p value ,0.05 was The levels of calcium influx normalized within 2 min. To exclude considered significant. Unless otherwise stated, a representative experi- any flow-induced calcium signals, we perfused the neutrophils after ment out of at least three independent experiments is shown. at least 20 min of initial flow onset. The results showed that initial neutrophil contact during the rolling stage rapidly induced a transient Results increase in intracellular calcium concentration in ECs, preceding the The intracellular tail of ICAM-1 interacts with annexin A2 firm adhesion step (Fig. 2B, Supplemental Video 1). Interestingly, To identify new proteins involved in the regulation of ICAM-1– firm adhesion of neutrophils to the endothelium was observed di- mediated leukocyte TEM, we performed a pull-down assay using rectly after the temporary burst of released intracellular calcium (Fig. biotinylated peptides encoding the intracellular tail of ICAM-1. 2B, Supplemental Video 1). We measured a 5-fold increase in sensor Peptide-bound proteins were subsequently separated by SDS- ratio almost instantly when the neutrophil touched the endothelium, PAGE, visualized by silver staining, and specific protein bands in line with previous findings by Ziegelstein and colleagues (36). Downloaded from were analyzed by mass spectrometry. A protein band interacting Surprisingly, however, only 20–30% of the sensor-expressing ECs with the ICAM-1 peptide that migrated at a molecular mass of 36 showed an increase in calcium signal upon rolling neutrophils. This kDa was identified by mass spectrometry as the calcium effector may reflect the heterogenic expression of the adhesion molecules like protein annexin A2. Biochemical pull-down assays from cell ly- selectin and ICAM-1 between mutual ECs. Quantification of the sates using peptides were repeated, analyzed by Western blot, and biosensor activity showed the significance of the transient and rapid confirmed the binding of annexin A2 to the intracellular tail of intracellular calcium changes preceding neutrophil adhesion (Fig. http://www.jimmunol.org/ ICAM-1, and not to the intracellular domain of VCAM-1 or 2C). To test whether the supernatant of the neutrophils may increase streptavidin beads, to corroborate the interaction of annexin A2 intracellular calcium levels, we treated ECs with supernatant of with the intracellular domain of ICAM-1 (Fig. 1A). Filamin A, PMA-stimulated neutrophils and did not observe any calcium influx previously reported to interact with the intracellular domain of (Supplemental Fig. 2). This excludes that secreted molecules from ICAM-1 (14), was used as a positive control (Fig. 1A). To in- the neutrophils induced the calcium influx in the endothelium. vestigate the endogenous interaction between annexin A2 and In addition, increased levels of intracellular calcium drives the ICAM-1, we immunoprecipitated endogenous ICAM-1 using majority of annexin A2 to the plasma membrane, as is indicated by magnetic beads coated with anti–ICAM-1 Abs that induced the the rapid translocation of cytosolic annexin A2-GFP to the plasma by guest on September 26, 2021 clustering of ICAM-1 (14). Anti–VCAM-1 or anti-IgG isotype membrane after ionomycin treatment (Fig. 2D, Supplemental Video control Ab-coated beads were used as controls. Using this method, 2). Also, endogenous stimuli like histamine induced annexin A2 we effectively precipitated endogenous ICAM-1 and VCAM-1 translocation to the plasma membrane (Supplemental Fig. 3). To from TNF-a–treated ECs (Fig. 1B). Analysis of the binding of study whether the interaction between endogenous ICAM-1 and annexin A2 and filamin A showed that these proteins interacted annexin A2 is regulated by high calcium, we immunoprecipitated with ICAM-1 using anti–ICAM-1 Ab–coated beads (Fig. 1B). In ICAM-1 using magnetic beads coated with anti–ICAM-1 Abs on contrast with annexin A2 that was only coprecipitated with anti– intact ECs treated with ionomycin (prelysis). The coprecipitation of ICAM-1 Ab–coated beads, a low amount of filamin A did also annexin A2 with anti–ICAM-1 Ab–coated beads was enhanced coprecipitate with anti–VCAM-1 Ab–coated beads (Fig. 1B). upon high intracellular calcium conditions (Fig. 2E). Quantification Annexin A2 and filamin A were not coprecipitated with anti-IgG showed that pretreating ECs with ionomycin significantly promoted isotype control Ab-coated beads (Fig. 1B). Thus, our findings the coprecipitation of annexin A2 with anti–ICAM-1 Ab–coated show that annexin A2 binds to the intracellular tail of ICAM-1, but beads, whereas filamin A levels were not significantly altered (Fig. not to the intracellular tail of VCAM-1. 2E). Note that addition of anti–ICAM-1 Ab–coated beads to an EC lysate (postlysis) did not result in the coprecipitation of annexin A2 Rolling neutrophils transiently increase intracellular calcium or filamin A, indicating that the interaction with endogenous annexin levels in EC driving annexin A2 to associate with the plasma A2 required intact membranes to promote the interaction between membrane and ICAM-1 annexin A2 and ICAM-1 upon ICAM-1 clustering. Together, these Annexin A2 is a calcium effector protein comprising five calcium- findings showed that rolling neutrophils transiently increase intra- binding domains that facilitate the binding of annexin A2 to the cellular calcium levels in ECs that may transiently drive annexin A2 plasma membrane upon release of intracellular calcium (32). To to associate with the plasma membrane and ICAM-1. investigate calcium signaling during leukocyte diapedesis under shear stress conditions, we used the FRET-based genetically ICAM-1 is localized in F-actin– and annexin A2–rich microvilli encoded calcium indicator Ycam3.6 (29) (Fig. 2A). The single- To study the spatial organization of ICAM-1 at the EC surface chain Ycam3.6 biosensor consists of an ECFP, calmodulin E104Q, under non–ligand-bound conditions, we costained TNF-a–stimu- the calmodulin-binding peptide M13p, and a cp-Venus. Binding of lated ECs for endogenous ICAM-1, F-actin, and annexin A2.

30 min prelysis (meaning clustered ICAM-1) or postlysis (meaning nonclustered ICAM-1 conditions). ECs were stimulated with ionomycin for 2 min to induce high intracellular calcium levels. Extracts were probed with Abs directed against filamin A, ICAM-1, and annexin A2. Quantification of annexin A2 and filamin A binding, presented as quantum level (sum of grayness of each pixel in blot) minus BG (QL-Bg), normalized to that of ICAM-1. *p , 0.05 AnxA2 versus AnxA2 + ionomycin (ANOVA). Data represent mean and SEM of seven experiments (B and C), five experiments (D), or three experiments (E). 2772 ANNEXIN A2 CONTROLS ICAM-1 FUNCTION

Overnight TNF-a–stimulated ECs showed heterogeneous distri- showed enrichment of both proteins in microvilli (Fig. 3C, 3D). In bution of ICAM-1 on the endothelial surface (Fig. 3A). Moreover, contrast with ICAM-1, annexin A2 is also enriched in membrane ICAM-1 is particularly enriched in F-actin–rich extensions called domains other than F-actin–rich microvilli (Fig. 3C, 3D). Thus, microvilli that protrude out of the EC apical surface (Fig. 3B) unengaged ICAM-1 is distributed at microvilli where it colo- (24, 37). Immunoﬂuorescent analysis of annexin A2 and ICAM-1 calized with F-actin and annexin A2. Downloaded from http://www.jimmunol.org/ by guest on September 26, 2021

FIGURE 3. ICAM-1 at the EC surface is localized in microvilli rich in F-actin and annexin A2. (A) ECs show heterogeneous ICAM-1 surface expression after TNF-a stimulation. Immunofluorescence analyses of ICAM-1 and F-actin in TNF-a–stimulated ECs. Open and filled arrows indicate low and high endothelial surface expression of ICAM-1, respectively. Scale bar, 30 mm. (B) Immunofluorescence analyses of ICAM-1 and F-actin in TNF-a–stimulated ECs. Scale bar, 10 mm. In the zoom panel, filled arrows indicate ICAM-1 localization at the plasma membrane in microvilli. Scale bar, 5 mm. (C) Immunofluorescence analyses of ICAM-1, F-actin, and annexin A2 in TNF-a–stimulated ECs. Annexin A2 is localized in nanometer- scale clusters at the plasma membrane. Scale bar, 10 mm. (D)LineprofilesofROImarkedin(C). Original magnification 36. Line profile 1 indicates colocalization among ICAM-1, F-actin, and annexin A2 in microvilli. Line profile 2 indicates annexin A2 membrane domains outside microvilli that do not colocalize with ICAM-1. Line profile 3 indicates membrane domains that are rich in ICAM-1, annexin A2, and F-actin and domains that contain only annexin A2 or F-actin. The Journal of Immunology 2773 Downloaded from http://www.jimmunol.org/ by guest on September 26, 2021

FIGURE 4. Annexin A2 regulates the spatial distribution of ICAM-1 at the EC surface and is required for ICAM-1 translocation to caveolin-1–rich membrane domains upon ICAM-1 clustering. (A) Schematic representation of ICAM-1 cross-linking prior membrane fractionation. ICAM-1 at the EC surface is clustered by incubation with the BBIG-I1 mAb for 30 min followed by incubation with a secondary goat anti-mouse Ab for 30 min. (B) Im- munoblot analysis of protein extracts that were separated on density by OptiPrep gradient centrifugation to visualize the spatial distribution of proteins at the plasma membrane in nonclustered and clustered ICAM-1 conditions. Extracts used for membrane fractionation were prepared 72 h after transduction with control (B and C) or annexin A2 shRNA (D and E). Extracts were probed with Abs directed against ICAM-1, caveolin-1, annexin A2, RhoGDI, and Ezrin. Lanes 1–5 indicate the cytosolic protein fraction. Lanes 6–10 indicate high-density membrane fractions rich in caveolin-1. Lanes 11–14 indicate low- density membrane fractions rich in ezrin and Annexin A2. (F) Quantification of the spatial distribution of ICAM-1 at the plasma membrane in nonclustered and clustered ICAM-1 conditions. RhoGDI indicates the cytosolic fraction. (G) Quantification of spatial distribution of ICAM-1 in control and annexin A2– depleted ECs after ICAM-1 clustering. (H) Quantification of the spatial distribution of caveolin-1 in control and annexin A2–depleted ECs after ICAM-1 clustering. (I) Quantification of the spatial distribution of ICAM-1 and caveolin-1 at the plasma membrane in nonclustered and clustered ICAM-1 conditions. RhoGDI indicates the cytosolic fraction. 2774 ANNEXIN A2 CONTROLS ICAM-1 FUNCTION Downloaded from http://www.jimmunol.org/ by guest on September 26, 2021

FIGURE 5. Annexin A2 regulates the lateral mobility of ICAM-1 at the EC surface. (A) Time-lapse confocal images showing fluorescent recovery after photobleaching of ICAM-1–GFP indicative for the lateral mobility of ICAM-1 at the plasma membrane and in microvilli. White box indicates the ICAM-1–GFP+ membrane part that has been bleached. Scale bar, 10 mm. (B) Quantification of ICAM-1–GFP recovery kinetics after photobleaching showing the mobile fraction of ICAM-1–GFP at the EC surface in control (green line) and annexin A2–depleted (red line) ECs. Bar graph shows quantification of the mobile fraction of ICAM- 1 after 150 s. (C) Endothelial annexin A2 depletion by shRNA. Immunoblot analysis of protein extracts prepared from HUVECs expressing ICAM-1–GFP. Extracts were prepared 72 h after transduction with control or annexin A2 shRNA. Extracts were probed with Abs directed against ICAM-1, annexin A2, and actin. Quantification of actin, annexin A2 ICAM-1, and ICAM-1–GFP protein levels after 72 h after transduction with control or annexin A2 shRNA, presented as quantum level (sum of grayness of each pixel in blot) minus BG (QL-Bg). (D) Confocal imaging of ICAM-1–GFP dynamics during incubation of anti–ICAM-1– coated beads at TNF-a–stimulated control and annexin A2–deficient primary ECs. (E) Kinetics of ICAM-1–GFP recruitment to anti–ICAM-1–coated beads in TNF-a–stimulated control and annexin A2–deficient primary ECs. (F) Quantification of ICAM-1–GFP+ rings percentage after 30-min incubation with anti–ICAM- 1–coated beads in TNF-a–stimulated control and annexin A2–deficient primary ECs. ****p , 0.0001, control versus AnxA2 (Student t test) (B), ****p , 0.0001, control versus AnxA2 (ANOVA) (C), **p , 0.01 control versus AnxA2 (Student t test) (F). Data represent mean and SEM of 20 (B) or 7 experiments (E and F).

Annexin A2 regulates the spatial organization of ICAM-1 at the brane domains (fractions 6–10) (Fig. 4B). Ab-mediated cross- plasma membrane linking of ICAM-1 drives ICAM-1 to the EC poles, regions that To examine the spatial organization of ICAM-1 at the plasma are rich in caveolin-1 (23). In line with these ﬁndings, we found membrane during ICAM-1–mediated neutrophil TEM, we used a that Ab-mediated clustering of ICAM-1 increased cosedimenta- detergent-free membrane fractionation (31). Neutrophil-mediated tion of ICAM-1 with caveolin-1 to fractions 8–10, but also in- ICAM-1 clustering was mimicked using Ab cross-linking, duced the dissociation of ezrin with the plasma membrane according to previous studies (38–40) (Fig. 4A). Without (Fig. 4C). Note that the membrane distribution of annexin A2 was Ab-mediated clustering, ICAM-1 cosedimented with ezrin and not affected by ICAM-1 clustering (Fig. 4B, 4C). To investigate annexin A2 in fractions 10–12 outside of caveolin-1–rich mem- whether annexin A2 regulated the spatial organization of ICAM-1 The Journal of Immunology 2775

FIGURE 6. Annexin A2 depletion in ECs enhanced the adhesive capacity of ICAM-1 to bind neutrophil b2-integrins. (A) Schematic representation of the experimental setup used to simultaneously measure two conditions with same donor neutrophils. (B) Quantification of neutrophil rolling velocities over TNF-a– stimulated control and annexin A2–deficient primary ECs. (C) Epifluorescent live-cell imaging of Downloaded from neutrophil TEM through TNF-a– stimulated control and annexin A2–deficient primary ECs under physiological flow conditions (0.9 dyne/ cm2). (D) Quantification of adherent neutrophils through TNF-a–treated ECs under physiological flow con- http://www.jimmunol.org/ ditions after 72-h transduction with control shRNA (open bar), annexin A2 shRNA A9 (black bar), or annexin A2 shRNA G5 (gray bar). *p , 0.05, **p , 0.01 control versus AnxA2 (ANOVA) (D). Data are representative of five independent experiments with .5 rolling events

per group (B) or ﬁve experiments by guest on September 26, 2021 with .150 events per group (D) [error bars in (B) and (D), SEM].

at the plasma membrane, we depleted annexin A2 using shRNA. of ICAM-1–GFP in the bleached region was measured using Annexin A2 depletion shifted nonclustered ICAM-1 to less dense confocal laser scanning microscopy and showed that the mobile fractions 11–13 (Fig. 4D). Strikingly, cross-linking ICAM-1 under fraction of ICAM-1 was around 40% in control conditions these conditions prevented the cosedimentation of ICAM-1 with (Fig. 5B). However, in annexin A2–deficient ECs, the mobile caveolin-1, whereas annexin A2 depletion did not alter the dis- fraction of ICAM-1 was significantly increased to .55% (Fig. 5B). tribution of caveolin-1 or ezrin dissociation from the plasma Subsequently, annexin A2, F-actin, endogenous ICAM-1, and membrane upon ICAM-1 clustering (Fig. 4E). These findings ICAM-1–GFP levels were analyzed by Western blot. Importantly, showed that annexin A2 regulates the spatial ICAM-1 distribution annexin A2–deficient ECs did not affect endogenous ICAM-1, at the EC surface and that annexin A2 is required for the trans- as well as ICAM-1–GFP protein levels, in TNF-a–treated ECs location of ICAM-1 to caveolin-1–rich membrane domains upon (Fig. 5C). Moreover, using anti–ICAM-1 Ab–coated beads, we ICAM-1 clustering. found that ICAM-1–GFP was more rapidly and more frequently recruited to beads when annexin A2 was depleted than under Annexin A2 negatively regulates ICAM-1 mobility at the EC control conditions (Fig. 5D, 5E, Supplemental Video 3). Quantifi- surface cation showed that the kinetics of ICAM-1–GFP recruitment to To investigate whether annexin A2 affects ICAM-1 dynamics at anti–ICAM-1 Ab–coated beads was significantly faster in annexin the plasma membrane, we examined ICAM-1 mobility using A2–deficient ECs than in control ECs (Fig. 5F). Also, the number of fluorescent recovery after photobleaching. ECs were transiently ICAM-1–GFP+ rings was increased in annexin A2–deficient ECs transfected with ICAM-1–GFP, treated with TNF-a, and subse- compared with the control (Fig. 5F). Thus, annexin A2 negatively quently a region of interest was bleached (Fig. 5A). The recovery regulates ICAM-1 mobility at the EC surface. 2776 ANNEXIN A2 CONTROLS ICAM-1 FUNCTION

Downregulation of endothelial annexin A2 increases neutrophil rolling velocity (Fig. 6B). However, the adhesion of neu- neutrophil adhesion to TNF-a–stimulated ECs trophils to annexin A2–deficient ECs was significantly increased ICAM-1 at the EC surface has been described to support the compared with control ECs (Fig. 6C, Supplemental Video 4). Quan- rolling, arrest, crawling, and transmigration of neutrophils (1–6). tification of the number of adherent neutrophils showed a 2- to To investigate how altered ICAM-1 membrane distribution and dy- 3-fold increase in neutrophil adhesion when annexin A2 was silenced namics in annexin A2–deficient ECs affect neutrophil TEM under using two distinct short hairpins targeting annexin A2 (Fig. 6D). physiological flow conditions, we simultaneously measured neu- Surprisingly, neutrophil crawling distance on the EC surface, trophil TEM through control and annexin A2–deficient ECs in a as well as its velocity, was significantly reduced in annexin A2– parallel flow setup (Fig. 6A). This setup allowed us to use primary deficient primary ECs (Fig. 7A–C). Despite slower apical move- human neutrophils from the same donor. Neutrophils were si- ment, neutrophils spend less time crawling over the EC surface multaneously perfused over TNF-a–stimulated ECs under physi- before transmigration (Fig. 7D). In line with this, we observed that ological shear stress conditions (0.8 dyne/cm2). Neutrophil rolling most neutrophils crawled in a direct line to the EC junctions where velocity, adhesion, crawling, and transmigration were recorded they breeched the endothelial monolayer (data not shown). Con- by a wide-field microscope and subsequently quantified using sequently, the absolute number of neutrophils that finished the ImageJ software. Annexin A2 depletion in ECs did not affect final step, that is, crossing the endothelial barrier, was significantly Downloaded from http://www.jimmunol.org/

FIGURE 7. Neutrophil crawling distance and velocity were significantly reduced in annexin A2–deficient ECs, but neutrophil diapedesis was not affected. (A) Graphical representation of neutrophil crawling tracks over TNF-a–stimulated control and annexin A2–deficient primary ECs. (B) Quantification of by guest on September 26, 2021 neutrophil crawling distance, crawling velocity (C), and crawling time (D) over TNF-a–stimulated control and annexin A2–deficient primary ECs. (E) Quantification of transmigrated neutrophils through TNF- a–treated ECs under physiological flow conditions after 72-h transduction with control shRNA (open bar), annexin A2 shRNA A9 (black bar), or annexin A2 shRNA G5 (gray bar). (F) Quantification of percentage of adherent cells that completed transmigration after 20 min. ****p , 0.0001 control versus AnxA2 (Stu- dent t test) (B), *p , 0.05 control versus AnxA2 (Student t test) (C and D), *p , 0.05 control versus AnxA2 (ANOVA) (E). Data are representative of five independent experiments with .5 crawling events per group (A–D) or five experiments with .150 events per group (E and F) [error bars in (A)–(F), SEM]. The Journal of Immunology 2777 increased when annexin A2 was depleted (Fig. 7E). When cal- tion (47–50). In contrast with multiple transient calcium spikes at culating the number of transmigrated neutrophils relative to the each stage of diapedesis, we found that EC intracellular calcium number of adhering neutrophils, there was no difference in per- was released upon initial neutrophil–EC contact before neutrophil centage of neutrophils that crossed the endothelium in the absence adhesion under flow conditions. As an alternative model for calcium or presence of annexin A2 (Fig. 7F). Altogether, our results showed being involved in junctional opening, we show that the calcium- that annexin A2 regulates the adhesion, crawling, and subsequent effector annexin A2 is recruited to the plasma membrane upon re- transmigration of neutrophils across the endothelium by controlling lease of intracellular calcium. Increased annexin A2 at the plasma ICAM-1 membrane distribution and dynamics that, in turn, me- membrane may modulate ICAM-1 function and thereby affect diates the efficiency of neutrophil TEM. neutrophil TEM efficiency. In conclusion, our work discovers that annexin A2 regulates Discussion ICAM-1 function through its spatial distribution on the endothelial ICAM-1 at the EC surface functions as an adhesive ligand for surface, which limits neutrophil TEM efficiency. These findings neutrophil b2-integrins required for effective neutrophil TEM. may provide new targets for drug therapy to treat inflammatory- However, little is known about how cell-surface distribution of based diseases. ICAM-1 is regulated and how this spatial organization controls the adhesive function of ICAM-1. It has been hypothesized that Acknowledgments multiple ICAM-1 complexes with different protein content, spatial Annexin A2-GFP was a kind gift from Dr. V. Gerke (University of organization, and subcellular localization regulate leukocyte be- Muenster, Germany). ICAM-1–GFP was a kind gift from Dr. O. Barreiro havior at different stages of leukocyte TEM (21, 23, 28). In and Dr. F. Sanchez-Madrid (Madrid, Spain). We thank Dr. Joachim Downloaded from agreement with this hypothesis, we found that altered ICAM-1 Goedhart (University of Amsterdam). We thank Prof. Dr. P. L. Hordijk membrane distribution significantly alters neutrophil TEM effi- for critically reading the manuscript. Mark Hoogenboezem and Floris van ciency. We found that clustered ICAM-1 translocates from ezrin- Alphen are greatly acknowledged for technical assistance. We thank rich membrane domains to caveolin-1–rich membrane domains, Bart-Jan de Kreuk, Erik Mul, Ilse Timmerman, Jeffrey Kroon, Mar which were impaired in annexin A2–depleted ECs. This altered Fernandez-Borja, and Simon Tol for helpful discussions.

ICAM-1 membrane distribution enhanced the lateral mobility and http://www.jimmunol.org/ adhesive capacity of ICAM-1 to bind neutrophils. These findings Disclosures suggest that the translocation of ICAM-1 into caveolae has a The authors have no financial conflicts of interest. limiting effect on ICAM-1–mediated leukocyte adhesion. A previous study showed that ICAM-1 in caveolae is rapidly transcytosed References and transported to the basolateral surface of ECs (23). This event 1.Bullard,D.C.,L.Qin,I.Lorenzo,W.M.Quinlin,N.A.Doyle,R.Bosse, together with the expression level of ICAM-1 on the cell surface has D. Vestweber, C. M. Doerschuk, and A. L. Beaudet. 1995. P-selectin/ICAM-1 been suggested to regulate transcellular migration, but not para- double mutant mice: acute emigration of neutrophils into the peritoneum is com- pletely absent but is normal into pulmonary alveoli. J. Clin. Invest. 95: 1782–1788. cellular migration of T cells and neutrophils (23, 41, 42). 2. Gorina, R., R. Lyck, D. Vestweber, and B. Engelhardt. 2014. b2 integrin- Depletion of the intracellular domain of ICAM-1 impairs leu- mediated crawling on endothelial ICAM-1 and ICAM-2 is a prerequisite for by guest on September 26, 2021 kocyte transmigration, but not adhesion (8, 9, 43). Partly in line transcellular neutrophil diapedesis across the inflamed blood-brain barrier. J. Immunol. 192: 324–337. with these studies, we found that annexin A2–mediated alteration 3. Phillipson, M., B. Heit, P. Colarusso, L. Liu, C. M. Ballantyne, and P. Kubes. 2006. of ICAM-1 distribution affects neutrophil adhesion, but not the Intraluminal crawling of neutrophils to emigration sites: a molecularly distinct percentage of transmigrating neutrophils. This suggests that the process from adhesion in the recruitment cascade. J. Exp. Med. 203: 2569–2575. 4. Sligh, J. E., Jr., C. M. Ballantyne, S. S. Rich, H. K. Hawkins, C. W. Smith, spatial organization of ICAM-1 at the endothelial surface is of A. Bradley, and A. L. Beaudet. 1993. Inflammatory and immune responses are primary importance for the adhesion, and less for the transmi- impaired in mice deficient in intercellular adhesion molecule 1. Proc. Natl. Acad. gration or diapedesis of leukocytes. Studies using LFA-1– (a b2, Sci. USA 90: 8529–8533. L 5. Kadono, T., G. M. Venturi, D. A. Steeber, and T. F. Tedder. 2002. Leukocyte CD11a/CD18) and MAC-1 (aMb2,CD11b/CD18)–deficient mice rolling velocities and migration are optimized by cooperative L-selectin and show that neutrophil firm arrest depends mainly on LFA-1–ICAM- intercellular adhesion molecule-1 functions. J. Immunol. 169: 4542–4550. 6. Bourdillon, M. C., R. N. Poston, C. Covacho, E. Chignier, G. Bricca, and 1 interactions, whereas neutrophil crawling requires MAC-1– J. L. McGregor. 2000. ICAM-1 deficiency reduces atherosclerotic lesions in ICAM-1 interactions (2, 3). Interestingly, depletion of annexin A2 double-knockout mice (ApoE(-/-)/ICAM-1(-/-)) fed a fat or a chow diet. Arte- increased neutrophil adhesion efficiency, but reduced the time, rioscler. Thromb. Vasc. Biol. 20: 2630–2635. 7. Etienne-Manneville, S., J. B. Manneville, P. Adamson, B. Wilbourn, velocity, and distance of crawling neutrophils. Possibly, the J. Greenwood, and P. O. Couraud. 2000. ICAM-1-coupled cytoskeletal rear- increased ability of ICAM-1 to move through the membrane rangements and transendothelial lymphocyte migration involve intracellular may result in an increased efficiency to interact with neutrophil calcium signaling in brain endothelial cell lines. J. Immunol. 165: 3375–3383. 8. Greenwood, J., C. L. Amos, C. E. Walters, P. O. Couraud, R. Lyck, integrins. As a consequence, neutrophils adhere more rapidly to B. Engelhardt, and P. Adamson. 2003. Intracellular domain of brain endothelial the endothelium. Why neutrophil crawling depends on the spatial intercellular adhesion molecule-1 is essential for T lymphocyte-mediated signaling and migration. J. Immunol. 171: 2099–2108. distribution of ICAM-1 needs further investigation. Interestingly, 9. Lyck, R., Y. Reiss, N. Gerwin, J. Greenwood, P. Adamson, and B. Engelhardt. overexpression of annexin A2 did not result in decreased leuko- 2003. T-cell interaction with ICAM-1/ICAM-2 double-deficient brain endothe- cyte adhesion or TEM (data not shown). For annexin A2 to lium in vitro: the cytoplasmic tail of endothelial ICAM-1 is necessary for transendothelial migration of T cells. Blood 102: 3675–3683. properly localize at the plasma membrane, it requires binding to 10. Diamond, M. S., D. E. Staunton, S. D. Marlin, and T. A. Springer. 1991. Binding the small molecule S100A10. However, if the expression of of the integrin Mac-1 (CD11b/CD18) to the third immunoglobulin-like domain S100A2 is limiting, this may prevent the translocation of annexin of ICAM-1 (CD54) and its regulation by glycosylation. Cell 65: 961–971. 11. Otto, V. I., E. Damoc, L. N. Cueni, T. Schurpf,€ R. Frei, S. Ali, N. Callewaert, A2 to the membrane and consequently its functionality in regu- A. Moise, J. A. Leary, G. Folkers, and M. Przybylski. 2006. N-glycan structures lating ICAM-1. and N-glycosylation sites of mouse soluble intercellular adhesion molecule-1 revealed by MALDI-TOF and FTICR mass spectrometry. Glycobiology 16: Calcium signaling in EC has been implicated in all steps of 1033–1044. leukocyte TEM (17, 36, 44–46). In fact, the major role for calcium 12. Carman, C. V., C. D. Jun, A. Salas, and T. A. Springer. 2003. Endothelial cells signaling is believed to be involved in myosin light-chain kinase proactively form microvilli-like membrane projections upon intercellular adhesion molecule 1 engagement of leukocyte LFA-1. J. Immunol. 171: 6135–6144. activation and subsequent actomyosin contraction to induce tran- 13. Shaw, S. K., S. Ma, M. B. Kim, R. M. Rao, C. U. Hartman, R. M. Froio, L. Yang, sient openings in the EC junctions allowing leukocyte transmigra- T. Jones, Y. Liu, A. Nusrat, et al. 2004. Coordinated redistribution of leukocyte 2778 ANNEXIN A2 CONTROLS ICAM-1 FUNCTION

LFA-1 and endothelial cell ICAM-1 accompany neutrophil transmigration. J. 32. Gokhale, N. A., A. Abraham, M. A. Digman, E. Gratton, and W. Cho. 2005. Exp. Med. 200: 1571–1580. Phosphoinositide specificity of and mechanism of lipid domain formation by 14. Kanters, E., J. van Rijssel, P. J. Hensbergen, D. Hondius, F. P. Mul, annexin A2-p11 heterotetramer. J. Biol. Chem. 280: 42831–42840. A. M. Deelder, A. Sonnenberg, J. D. van Buul, and P. L. Hordijk. 2008. Filamin 33. Geiger, R. V., B. C. Berk, R. W. Alexander, and R. M. Nerem. 1992. Flow- B mediates ICAM-1-driven leukocyte transendothelial migration. J. Biol. Chem. induced calcium transients in single endothelial cells: spatial and temporal 283: 31830–31839. analysis. Am. J. Physiol. 262: C1411–C1417. 15. van Buul, J. D., M. J. Allingham, T. Samson, J. Meller, E. Boulter, R. Garcıá- 34. Shen, J., F. W. Luscinskas, A. Connolly, C. F. Dewey, Jr., and M. A. Gimbrone, Mata, and K. Burridge. 2007. RhoG regulates endothelial apical cup assembly Jr. 1992. Fluid shear stress modulates cytosolic free calcium in vascular endo- downstream from ICAM1 engagement and is involved in leukocyte trans- thelial cells. Am. J. Physiol. 262: C384–C390. endothelial migration. J. Cell Biol. 178: 1279–1293. 35. Dull, R. O., and P. F. Davies. 1991. Flow modulation of agonist (ATP)-response 16. Martinelli, R., M. Gegg, R. Longbottom, P. Adamson, P. Turowski, and (Ca2+) coupling in vascular endothelial cells. Am. J. Physiol. 261: H149–H154. J. Greenwood. 2009. ICAM-1-mediated endothelial nitric oxide synthase acti- 36. Ziegelstein, R. C., S. Corda, R. Pili, A. Passaniti, D. Lefer, J. L. Zweier, vation via calcium and AMP-activated protein kinase is required for A. Fraticelli, and M. C. Capogrossi. 1994. Initial contact and subsequent ad- transendothelial lymphocyte migration. Mol. Biol. Cell 20: 995–1005. hesion of human neutrophils or monocytes to human aortic endothelial cells 17. Pfau, S., D. Leitenberg, H. Rinder, B. R. Smith, R. Pardi, and J. R. Bender. 1995. releases an endothelial intracellular calcium store. Circulation 90: 1899–1907. Lymphocyte adhesion-dependent calcium signaling in human endothelial cells. 37. van Buul, J. D., J. van Rijssel, F. P. van Alphen, M. Hoogenboezem, S. Tol, J. Cell Biol. 128: 969–978. K. A. Hoeben, J. van Marle, E. P. Mul, and P. L. Hordijk. 2010. Inside-out 18. van Rijssel, J., J. Kroon, M. Hoogenboezem, F. P. van Alphen, R. J. de Jong, regulation of ICAM-1 dynamics in TNF-alpha-activated endothelium. PLoS E. Kostadinova, D. Geerts, P. L. Hordijk, and J. D. van Buul. 2012. The Rho- One GEF Trio controls leukocyte transendothelial migration by promoting docking 5: e11336. 38. Thompson, P. W., A. M. Randi, and A. J. Ridley. 2002. Intercellular adhesion structure formation. Mol. Biol. Cell. 23: 2831–2844. 19. Carpeń, O., P. Pallai, D. E. Staunton, and T. A. Springer. 1992. Association of molecule (ICAM)-1, but not ICAM-2, activates RhoA and stimulates c-fos and intercellular adhesion molecule-1 (ICAM-1) with actin-containing cytoskeleton rhoA transcription in endothelial cells. J. Immunol. 169: 1007–1013. and alpha-actinin. J. Cell Biol. 118: 1223–1234. 39. Tilghman, R. W., and R. L. Hoover. 2002. E-selectin and ICAM-1 are incor- 20. Heiska, L., K. Alfthan, M. Gro¨nholm, P. Vilja, A. Vaheri, and O. Carpeń. 1998. porated into detergent-insoluble membrane domains following clustering in Association of ezrin with intercellular adhesion molecule-1 and -2 (ICAM-1 and endothelial cells. FEBS Lett. 525: 83–87. Downloaded from ICAM-2). Regulation by phosphatidylinositol 4, 5-bisphosphate. J. Biol. Chem. 40. van Buul, J. D., C. Voermans, V. van den Berg, E. C. Anthony, F. P. Mul, S. van 273: 21893–21900. Wetering, C. E. van der Schoot, and P. L. Hordijk. 2002. Migration of human 21. Schaefer, A., J. Te Riet, K. Ritz, M. Hoogenboezem, E. C. Anthony, F. P. J. Mul, hematopoietic progenitor cells across bone marrow endothelium is regulated by C. J. de Vries, M. J. Daemen, C. G. Figdor, J. D. van Buul, and P. L. Hordijk. vascular endothelial cadherin. J. Immunol. 168: 588–596. 2014. Actin-binding proteins differentially regulate endothelial cell stiffness, 41. Abadier, M., N. Haghayegh Jahromi, L. Cardoso Alves, R. Boscacci, ICAM-1 function and neutrophil transmigration. J. Cell Sci. 127: 4470–4482. D. Vestweber, S. Barnum, U. Deutsch, B. Engelhardt, and R. Lyck. 2015. Cell 22. Yang, L., J. R. Kowalski, P. Yacono, M. Bajmoczi, S. K. Shaw, R. M. Froio, surface levels of endothelial ICAM-1 influence the transcellular or paracellular

D. E. Golan, S. M. Thomas, and F. W. Luscinskas. 2006. Endothelial cell cor- T-cell diapedesis across the blood-brain barrier. Eur. J. Immunol. 45: 1043–1058. http://www.jimmunol.org/ tactin coordinates intercellular adhesion molecule-1 clustering and actin cyto- 42. Yang, L., R. M. Froio, T. E. Sciuto, A. M. Dvorak, R. Alon, and skeleton remodeling during polymorphonuclear leukocyte adhesion and F. W. Luscinskas. 2005. ICAM-1 regulates neutrophil adhesion and transcellular transmigration. J. Immunol. 177: 6440–6449. migration of TNF-alpha-activated vascular endothelium under flow. Blood 106: 23. Millań, J., L. Hewlett, M. Glyn, D. Toomre, P. Clark, and A. J. Ridley. 2006. 584–592. Lymphocyte transcellular migration occurs through recruitment of endothelial 43. Sans, E., E. Delachanal, and A. Duperray. 2001. Analysis of the roles of ICAM-1 ICAM-1 to caveola- and F-actin-rich domains. Nat. Cell Biol. 8: 113–123. in neutrophil transmigration using a reconstituted mammalian cell expression 24. Oh, H. M., S. Lee, B. R. Na, H. Wee, S. H. Kim, S. C. Choi, K. M. Lee, and model: implication of ICAM-1 cytoplasmic domain and Rho-dependent sig- C. D. Jun. 2007. RKIKK motif in the intracellular domain is critical for spatial naling pathway. J. Immunol. 166: 544–551. and dynamic organization of ICAM-1: functional implication for the leukocyte 44. Huang, A. J., J. E. Manning, T. M. Bandak, M. C. Ratau, K. R. Hanser, and adhesion and transmigration. Mol. Biol. Cell 18: 2322–2335. S. C. Silverstein. 1993. Endothelial cell cytosolic free calcium regulates neu- 25. Jun, C. D., C. V. Carman, S. D. Redick, M. Shimaoka, H. P. Erickson, and trophil migration across monolayers of endothelial cells. J. Cell Biol. 120: 1371– T. A. Springer. 2001. Ultrastructure and function of dimeric, soluble intercellular 1380. by guest on September 26, 2021 adhesion molecule-1 (ICAM-1). J. Biol. Chem. 276: 29019–29027. 45. Kielbassa-Schnepp, K., A. Strey, A. Janning, L. Missiaen, B. Nilius, and 26. Yang, Y., C. D. Jun, J. H. Liu, R. Zhang, A. Joachimiak, T. A. Springer, and V. Gerke. 2001. Endothelial intracellular Ca2+ release following monocyte ad- J. H. Wang. 2004. Structural basis for dimerization of ICAM-1 on the cell hesion is required for the transendothelial migration of monocytes. Cell Calcium surface. Mol. Cell 14: 269–276. 30: 29–40. 27. Barreiro, O., M. Yań˜ez-Mo´, M. Sala-Valde´s, M. D. Gutie´rrez-Lo´pez, S. Ovalle, 46. Su, W. H., H. I. Chen, J. P. Huang, and C. J. Jen. 2000. Endothelial [Ca(2+)](i) A. Higginbottom, P. N. Monk, C. Cabanãs, and F. Sańchez-Madrid. 2005. En- Blood dothelial tetraspanin microdomains regulate leukocyte firm adhesion during signaling during transmigration of polymorphonuclear leukocytes. 96: extravasation. Blood 105: 2852–2861. 3816–3822. 28. Barreiro, O., M. Zamai, M. Yań˜ez-Mo´, E. Tejera, P. Lo´pez-Romero, P. N. Monk, 47. Hixenbaugh, E. A., Z. M. Goeckeler, N. N. Papaiya, R. B. Wysolmerski, E. Gratton, V. R. Caiolfa, and F. Sańchez-Madrid. 2008. Endothelial adhesion S. C. Silverstein, and A. J. Huang. 1997. Stimulated neutrophils induce myosin receptors are recruited to adherent leukocytes by inclusion in preformed tetra- light chain phosphorylation and isometric tension in endothelial cells. Am. J. spanin nanoplatforms. J. Cell Biol. 183: 527–542. Physiol. 273: H981–H988. 29. Nagai, T., S. Yamada, T. Tominaga, M. Ichikawa, and A. Miyawaki. 2004. 48. Etienne, S., P. Adamson, J. Greenwood, A. D. Strosberg, S. Cazaubon, and Expanded dynamic range of fluorescent indicators for Ca(2+) by circularly P. O. Couraud. 1998. ICAM-1 signaling pathways associated with Rho activation permuted yellow fluorescent proteins. Proc. Natl. Acad. Sci. USA 101: 10554– in microvascular brain endothelial cells. J. Immunol. 161: 5755–5761. 10559. 49. Saito, H., Y. Minamiya, S. Saito, and J. Ogawa. 2002. Endothelial Rho and Rho 30. Dull, T., R. Zufferey, M. Kelly, R. J. Mandel, M. Nguyen, D. Trono, and kinase regulate neutrophil migration via endothelial myosin light chain phos- L. Naldini. 1998. A third-generation lentivirus vector with a conditional pack- phorylation. J. Leukoc. Biol. 72: 829–836. aging system. J. Virol. 72: 8463–8471. 50. Wo´jciak-Stothard, B., L. Williams, and A. J. Ridley. 1999. Monocyte adhesion 31. Macdonald, J. L., and L. J. Pike. 2005. A simplified method for the preparation and spreading on human endothelial cells is dependent on Rho-regulated re- of detergent-free lipid rafts. J. Lipid Res. 46: 1061–1067. ceptor clustering. J. Cell Biol. 145: 1293–1307.