Neutrophils Mediated Adhesion and Migration in − Integrin 2Β

The Journal of Immunology

Tetraspanin CD37 Regulates b2 Integrin–Mediated Adhesion and Migration in Neutrophils

Janet L. Wee,*,† Keith E. Schulze,‡ Eleanor L. Jones,* Louisa Yeung,*,† Qiang Cheng,† Candida F. Pereira,x Adam Costin,‡ Georg Ramm,‡ Annemiek B. van Spriel,{ Michael J. Hickey,†,1 and Mark D. Wright*,1

Deciphering the molecular basis of leukocyte recruitment is critical to the understanding of inﬂammation. In this study, we inves- tigated the contribution of the tetraspanin CD37 to this key process. CD37-deﬁcient mice showed impaired neutrophil recruitment in a peritonitis model. Intravital microscopic analysis indicated that the absence of CD37 impaired the capacity of leukocytes to follow a CXCL1 chemotactic gradient accurately in the interstitium. Moreover, analysis of CXCL1-induced leukocyte-endothelial cell interactions in postcapillary venules revealed that CXCL1-induced neutrophil adhesion and transmigration were reduced in the

absence of CD37, consistent with a reduced capacity to undergo b2 integrin–dependent adhesion. This result was supported by in vitro flow chamber experiments that demonstrated an impairment in adhesion of CD37-deficient neutrophils to the b2 integrin ligand, ICAM-1, despite the normal display of high-affinity b2 integrins. Superresolution microscopic assessment of localization of CD37 and CD18 in ICAM-1–adherent neutrophils demonstrated that these molecules do not significantly cocluster in the cell

membrane, arguing against the possibility that CD37 regulates b2 integrin function via a direct molecular interaction. Moreover, CD37 ablation did not affect b2 integrin clustering. In contrast, the absence of CD37 in neutrophils impaired actin polymerization, cell spreading and polarization, dysregulated Rac-1 activation, and accelerated b2 integrin internalization. Together, these data indicate that CD37 promotes neutrophil adhesion and recruitment via the promotion of cytoskeletal function downstream of integrin-mediated adhesion. The Journal of Immunology, 2015, 195: 5770–5779.

eukocyte recruitment from the bloodstream and migration thelial surface, interactions predominantly mediated by the throughout the body are of fundamental importance in the selectin family of adhesion molecules. Rolling leukocytes are then L function of the immune system. It is well established that able to undergo arrest and ﬁrmly adhere to the endothelial surface for circulating leukocytes to enter sites of inﬂammation, they must in response to chemoattractants presented on the endothelium. undergo a complex sequence of interactions with the endothelium Chemoattractant-generated signals cause leukocyte integrins to

by guest on October 1, 2021. Copyright 2015 Pageant Media Ltd. lining the microvasculature of the inflamed site (1). This process is increase their affinity for their endothelial ligands, thereby fa- initiated by the leukocyte tethering to and rolling on the endo- cilitating arrest of the rolling leukocyte. Adherent leukocytes then undergo integrin-dependent migration over the endothelial surface to encounter an optimal site for transmigration and exit *Department of Immunology, Monash University, Alfred Medical Research and the vasculature. Once in the interstitium, leukocytes migrate to † Education Precinct, Melbourne, Victoria 3004, Australia; Centre for Inflammatory the inflammatory focus via chemotaxis. Although the major Diseases, Monash University Department of Medicine, Monash Medical Centre, Clayton, Victoria 3168, Australia; ‡Monash Micro Imaging, Monash University, adhesion molecules responsible for these interactions have been Clayton, Victoria 3800, Australia; xBurnet Institute, Alfred Medical Research and { identified and studied in detail, a growing body of evidence in- Education Precinct, Melbourne, Victoria 3004, Australia; and Department of Tumor dicates that another family of cell membrane molecules, the Immunology, Radboud Institute for Molecular Life Sciences, Radboud University https://www.jimmunol.org Medical Center, 6525 GA Nijmegen, the Netherlands tetraspanins, has previously unrecognized functions in control- 1M.J.H. and M.D.W. contributed equally to this work. ling adhesion molecule function during leukocyte recruitment ORCIDs: 0000-0001-7261-3314 (K.E.S.); 0000-0003-1233-1168 (L.Y.); 0000- (2–5). 0003-3596-2288 (G.R.); 0000-0003-2354-357X (M.J.H.); 0000-0002-2177-5214 Tetraspanins are a large family of proteins present in all mul- (M.D.W.). ticellular eukaryotic organisms, and they are characterized by their Received for publication September 23, 2014. Accepted for publication October 14, prototypical four transmembrane domain structure. Tetraspanins 2015.

Downloaded from have been shown to contribute to diverse cellular functions, in- This work was supported by funding from the National Health and Medical Research cluding cell viability, proliferation, adhesion, and migration (6). Council (NHMRC) of Australia (Grant ID 1033198). M.J.H. is an NHMRC Senior Research Fellow (Grant ID 1042775). A.B.v.S. is supported by the Netherlands Or- Tetraspanins achieve these effects via organization of the cell ganization for Scientiﬁc Research (Innovational Research Incentives Scheme Vidi membrane into microdomains consisting of tetraspanins and a Grant 864.11.006) and the Dutch Cancer Society (KUN 2014-6845). diverse assortment of partner proteins. In immune cells, critical Address correspondence and reprint requests to Assoc. Prof. Mark D. Wright, Depart- cell surface receptors including CD4, CD8, CD28, CD19, pattern ment of Immunology, Monash University, Alfred Medical Research and Education Precinct, Commercial Road, Melbourne, VIC 3004, Australia. E-mail address: recognition receptors, and MHC class I and II have all been [email protected] reported to be components of these tetraspanin-enriched micro- The online version of this article contains supplemental material. domains (7). Moreover, several of the proteins associated with Abbreviations used in this article: AMREP, Alfred Medical Research and Education these domains, including integrins and other adhesion molecules, Precinct; MFI, mean ﬂuorescence intensity; STORM, stochastic optical reconstruc- signaling molecules such as protein kinase C, and proteins asso- tion microscopy; TR, transferrin receptor; WT, wild-type. ciated with cytoskeletal function, have important roles in leuko- Copyright Ó 2015 by The American Association of Immunologists, Inc. 0022-1767/15/$25.00 cyte recruitment and migration. As such, the role of tetraspanins in

www.jimmunol.org/cgi/doi/10.4049/jimmunol.1402414 The Journal of Immunology 5771

controlling the movement of immune cells is a growing area of chloride, 10 mg/kg, i.p.), and a heat pad was used to maintain body interest. In B cells, the tetraspanin CD81 has been shown to asso- temperature. The jugular vein was cannulated for administration of addi- ciate with the a integrin and to promote interaction with VCAM-1 tional anesthetic. The cremaster muscle was dissected free of surrounding 4 tissues and exteriorized onto an optically clear viewing pedestal. The (8). Similarly, in endothelial cells, tetraspanins CD9, CD81, and muscle was cut longitudinally with a cautery and held flat against the CD151 assemble in docking structures around adherent leukocytes, pedestal by attaching silk sutures to the edges of the tissue. The muscle where they associate with adhesion molecules ICAM-1 and VCAM-1 was then superfused with bicarbonate-buffered saline (pH 7.4, 37˚C) and and promote transendothelial migration (2, 3). However, although the placed under a coverslip. An intravital microscope (Axioplan 2 Imaging; Zeiss) with an 320 actions of tetraspanins have been examined in detail in lymphocytes, objective lens was used to examine the cremasteric microcirculation. A dendritic cells, and endothelial cells, little is known about their actions video camera (Sony SSC-DC50AP) was used to project the images onto in neutrophils. a monitor (Sony PVM-20N5E), and the images were recorded using a CD37 is a leukocyte-restricted tetraspanin with documented videocassette recorder. Unbranched postcapillary venules (25–40 mmin functions in many different leukocyte lineages. In B cells, the actions diameter, two per experiment) were selected for observation in each experiment. To minimize variability, the same section of venule was of CD37 contribute to development of humoral immunity (9–11). recorded at each time interval. Leukocyte rolling flux, rolling velocity, Moreover, CD37 is a target for novel immunotherapies to treat adhesion, and transmigration were quantitated offline using playback B cell malignancies (12–18). CD37 has also been shown to control analysis. Leukocyte rolling flux was defined as the number of leukocytes functions in T cells and dendritic cells (19–22). Notably, we removing at a velocity less than that of erythrocytes in the same vessel per minute. Leukocyte rolling velocity was determined by measuring the cently demonstrated that CD37 promotes migration of dendritic time taken for leukocytes to roll over a 100-mm length of the venule, and cells (23). CD37 is also expressed in granulocytes (24), but its role was determined for 20 cells per time point. Adherent leukocytes were in these cells has not been examined. In this study, we analyzed defined as cells that remained stationary for 30 s or longer. Leukocyte leukocyte recruitment in CD372/2 mice and found compelling transmigration was defined as the number of extravascular leukocytes per field of view adjacent to the selected venule. evidence that tetraspanins regulate b2 integrin function in neutrophils and thereby modulate their capacity to respond to inflamma- CXCL1-induced leukocyte recruitment tory stimuli in vivo. Intravital microscopic examination of wild-type 2/2 b Two models of CXCL1-induced leukocyte recruitment were examined (25). (WT) and CD37 neutrophils revealed that CD37 promotes 2 The first model was CXCL1-induced leukocyte adhesion and transmigration. integrin–mediated neutrophil adhesion and transmigration, as well After an initial baseline recording, the cremaster muscle was superfused with as extravascular chemotaxis. These responses were associated with 5.2 nM CXCL1 (R&D Systems, Minneapolis, MN) dissolved in bicarbonate a role for CD37 in modulating cytoskeletal-dependent parameters superfusion buffer, and leukocyte recruitment parameters were examined after including actin polymerization, neutrophil spreading and polariza- 30 and 60 min. The second model was CXCL1-induced directed extravas- b cular leukocyte migration. Directed leukocyte recruitment was induced using tion, and 2 integrin internalization. CXCL1 incorporated in an agarose bead directly applied to the exteriorized cremaster muscle, as described previously (25). The gel bead (∼1mm3), which also incorporated a small amount of charcoal to aid visualization, was Materials and Methods placed ∼350 mm from a postcapillary venule and held in place with a cov- Mice erslip. The cremaster muscle was superfused with bicarbonate buffer at a 2 2 minimum rate (,0.2 ml/min) to avoid disrupting the chemotactic gradient CD37 / mice were generated by homologous recombination (9) and 2 2 and viewed with transillumination microscopy. Experiments ran for ∼2h. backcrossed 10 times to a C57BL/6 background. WT and CD37 / mice

by guest on October 1, 2021. Copyright 2015 Pageant Media Ltd. During the first hour, leukocytes underwent adhesion and migrated into the were bred and housed under specific pathogen-free conditions in the Alfred interstitial tissue adjacent to the selected venule. In the second hour, interstitial Medical Research and Education Precinct (AMREP) or Monash Medical migration of leukocytes toward the gel bead was recorded using time-lapse Centre Animal facilities and were used between 6 and 12 wk of age. The microscopy (Ultraview; PerkinElmer) and Volocity image acquisition software animal ethics committees at AMREP or Monash Medical Centre approved (PerkinElmer). Leukocyte migration parameters (velocity, track length, dis- all animal experiments. placement and meandering index [displacement/track length]) were deter- Abs mined using Image J (National Institutes of Health). Flow chamber assay Anti-Ly6G conjugated to allophycocyanin and anti-CD11a (aL integrin, clone M17/4) conjugated to Alexa 488 were purchased from eBio- a b In vitro leukocyte adhesion was examined using a whole blood flow science (San Diego, CA). Nonblocking anti-LFA-1 ( L 2 integrin, chamber assay, as described previously (23, 26), with some modifications. https://www.jimmunol.org clone H155-78) conjugated to PE was obtained from BioLegend Blood was harvested from anesthetized mice by cardiac puncture, diluted (Atlanta, GA). Phalloidin-FITC was purchased from Sigma-Aldrich 2 a 1:10 in HBSS, and perfused at a physiological shear rate of 2 dynes/cm at (St. Louis, MO); unconjugated rat anti L (clone M17/4), rat IgG2a 37˚C through a parallel flow chamber apparatus (Glycotech, Rockville, isotype control and mouse Fc block (clone 2.4G2) were all purchased MD) assembled over 35-mm Petri dishes precoated with P-selectin (1 mg/ml), from BD Biosciences (San Jose, CA). Mouse anti-human CD11b (clone CXCL1 (5.2 nM), and either mouse ICAM-1-Fc or Fc (9 nM; all reagents MEM-174) was sourced from ImmunoTools, (Friesoythe, Germany). b were from R&D Systems). In some experiments, blood samples were in- Rabbit polyclonal Abs to human CD18 ( 2 integrin) and human cubated with blocking anti-b mAb 2E6 for 15 min at 37˚C prior to per- transferrin receptor were sourced from Abcam (Cambridge, U.K.). 2 Downloaded from b fusion. Leukocyte interactions were visualized using an inverted microscope Hamster anti-mouse CD18 ( 2 integrin, clone 2E6), mouse anti-human 3 a (Axiovert 200; Zeiss) and an 10 objective. After a 6-min equilibration CD37 (clone WR17), and rat anti-mouse CD11b ( M integrin, clone period, the chamber was flushed with HBSS and recordings were made along M1/70) were purified in-house. the length of the flow chamber using a camera (Sony SSC-DC50AP) and video recorder. The numbers of interacting (rolling or adherent) cells were Thioglycollate-induced peritonitis determined in 100 fields of view for each experiment, and rolling velocity Mice were inoculated i.p. with 500 ml of 3% thioglycollate (Sigma- was determined by measuring the time taken for cells (20 per experiment) to Aldrich); 24 h later, mice were euthanized and cells were harvested by roll a defined distance in the chamber. peritoneal lavage. Total cell counts were obtained, and neutrophil numbers were derived from the percentage of Ly6G+ cells determined by analysis of Neutrophil enrichment from bone marrow stained cells by flow cytometry (LSR-Fortessa Cell Analyzer, BD Bio- Neutrophils were enriched from bone marrow as described previously (27). sciences). Within these isolates, neutrophils were identified with flow cytometry as Ly6G+ cells. Intravital microscopy Flow cytometric analysis of neutrophil integrins CXCL1-induced leukocyte recruitment was examined in the mouse cremaster muscle by intravital microscopy, as described previously (25). Mice Expression of aL and aM on neutrophils (identified via Ly6G expression) were anesthetized (ketamine hydrochloride, 150 mg/kg; xylazine hydro- was determined using flow cytometry, by staining with M17/4 and M1/70 5772 CD37 AND b2 INTEGRIN–MEDIATED ADHESION

respectively. To assess b2 integrin ligand binding function, a soluble temperature. Cells were subsequently stained with anti-mouse Alexa405- ICAM-1 binding assay was used (28). Bone marrow–derived neutrophils Alexa647 and anti rabbit Cy3-Alexa647 activator-reporter dyes (1:100) for (1 3 106) were incubated with mICAM-1 Fc or Fc alone (100 mg/ml for 30 min at room temperature. both), in the presence of Mg2+ (10 mM) and EGTA (1 mM) for 30 min at 37˚C. Cells were then fixed with paraformaldehyde, incubated with Fc N-STORM imaging and analysis block (1:100) for 5 min, and then stained with FITC-conjugated anti- m human Fc (1:50) and allophycocyanin-conjugated anti-Ly6G (1:200) for Stained cells were imaged in an oxygen-scavenging buffer (562 g/ml m m 30 min. ICAM-1 binding was examined on Ly6G+ cells. glucose oxidase, 34 g/ml catalase, 100 mM cysteamine, 80 M Tris, 44 mM Tris-HCL, 9.4 mM NaCl pH 8) through a CFI Apochromat TIRF Integrin internalization assay 3100 oil objective (NA 1.49) using Immersion oil type B with a refractive index of 1.515 (Cargille, Cedar Grove, NJ). Two-dimensional STORM Integrin internalization was examined as described previously (29), with some images were acquired with an N-STORM microscope (Nikon). Each modifications. Bone marrow neutrophils were chilled on ice for 20 min prior sample was illuminated repetitively with each frame of an activation laser to labeling with anti–LFA-1-PE (clone H155-78, 1:400). The mean fluo- (405 or 561 nm) followed by three frames of the reporter laser (647 nm) rescence intensity (MFI) of total anti-CD11a-PE staining on neutrophils with an alternating sequence of the two activation lasers used for two-color (MFIT) was determined at this time point, identifying neutrophils on the imaging. Emission was recorded onto a 256 3 256-pixel region of an basis of anti-Ly6G (1:200) staining. Additional aliquots of these labeled cells electron multiplying charge coupled device camera (iXon DU-897; Andor were then resuspended in ice-cold RPMI 1640 containing 1% goat serum as Technology, Belfast, U.K.), giving a pixel size of 157.17 nm. The an activating stimulus (11), allowed to adhere onto ICAM-1-Fc or Fc coated N-STORM images were reconstructed from a series of 8000 images using coverslips (9 nM) at room temperature for 20 min, and then transferred to NIS-Elements Ar software (version Ar 4.11; Nikon). Background was 37˚C to initiate integrin internalization. After 15 min, the coverslips were determined by measuring fluorescence intensity in intracellular areas placed on ice to arrest integrin internalization, and nonadherent cells were where no blinking was visible. Minimum peak size was determined by washed off with ice-cold 0.5% BSA-RPMI 1640. Any integrin-Ab com- measuring the fluorescence intensity of the dimmest spots above back- plexes remaining on the cell exterior were then stripped of Ab via incubation ground while maximum peak size was set at 20,000. in 0.5% BSA-RPMI 1640 (pH 2.5) for 2 min. Control experiments revealed Localization coordinates from each image were exported from NIS- that this technique removed the vast majority of the surface-bound Ab. Elements as a text file and imported into the statistical analysis environ- Subsequently, adherent cells were detached from the coverslips with trypsin- ment R (31). These coordinates were used to generate Planar Point Pattern EDTA, fixed with 2% paraformaldehyde, and colabeled with anti-Ly6G prior objects to be analyzed further using the spatstat package (32). Coclustering to FACS analysis. The MFI of internalized anti–LFA-1-PE in cells (MFII) of CD37, b2 integrin, CD11b and the transferrin receptor (TR) were was established from cell populations gated for Ly6G expression. The in- assessed using univariate and bivariate Getis and Franklin’s local Ripley’s ternalization index was calculated from normalizing MFII values to MFIT.In K-function analysis (33, 34). Under this method, coclustering is quantified parallel experiments to examine the time course of LFA-1 internalization, by correlating the degree to which each localization clusters with locali- formyl peptide receptor agonist peptide, WKYMVm (5 nM; Phoenix Phar- zations of its own species (i.e., L[r]) and the degree to which they cluster maceuticals, Burlingame, CA), was used as an activating stimulus (30). Cells with localizations of the opposite protein species (i.e., L[r]cross), at a were removed from the assay after 5, 15, 30,and 60 min, and assessed as specified spatial scale r. Getis and Franklin’s local version of Ripley’s described above. K-function (33) was used to calculate L(r) and bivariate/cross-type L(r)cross Confocal microscopy values for each point as follows: vffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi u ÀÁ Bone marrow–derived neutrophils (in RPMI 1640/1% sheep serum) were u n dij tA+ exi;xj; r allowed to adhere to coverslips coated with ICAM-1 Fc (9 nM), in the i¼0 n 1ifdij r LðrÞ ¼ where dij ¼ presence of WKYMVm or sheep serum, for 20 min at room temperature, j p else 0 and then transferred to 37˚C for 15 min. After washing, adherent cells were

by guest on October 1, 2021. Copyright 2015 Pageant Media Ltd. fixed with 2% paraformaldehyde and labeled with allophycocyanin- where r is the spatial scale radius, A is the area of three analysis windows conjugated anti-Ly6G. Cells were then permeabilized with 0.5% Triton-X distributed across and contained within the cell (each 3 3 3 mm), d is the m ij prior to staining with phalloidin-FITC (0.5 g/ml). Cells were examined distance between two points i and j, n is the number of points, and e(xi, xj; r)is via confocal microscopy (Nikon A1R). The average fluorescence intensity the edge-correction weighting. Ripley’s isotropic edge correction weighting, + of phalloidin-FITC staining and cell surface area of Ly6G cells were as implemented in the spatstat package (32, 35), was used for all analyses in determined in captured images using Image J. Thirty to thirty-five cells per this study. In our analyses, a spatial scale of r = 50 nm was used, which experiment were examined. Cell polarization was assessed utilizing the corresponds approximately to the maxima observed for the overall Ripley’s K p 3 Circularity plugin in Image J. Circularity is defined as 4 ([Area] / analysis of CD37, b2 integrin, CD11b, and the TR (data not shown). [Perimeter]2), where values range from 0 (elongated polygon, maximally m 2 polarized) to 1 (perfect circle, no polarization). Cells with areas of 40 m or Electron microscopy analysis of the surface distribution of b2 less were excluded from the analysis. integrin https://www.jimmunol.org Labeling of samples for N-STORM imaging For whole-mount transmission electron microscopy, bone marrow–derived neutrophils from WT and CD372/2 mice were allowed to adhere for Anti-mouse IgG Alexa 405-Alexa 647 and anti-rabbit IgG Cy3-Alexa 647 20 min at room temperature on ICAM-1–coated 100-mesh nickel grids prel- activator-reporter-labeled secondary Abs were prepared by dissolving dyes ayered with Formvar and then transferred to 37˚C in the presence of the Alexa 405 and 647 (Invitrogen) and Cy3 (GE Healthcare, Melbourne, chemotactic peptide WKYMVm (Peprotech, Rocky Hill, NJ) and Ca2+ and Australia) in DMSO at 0.02 mg/ml, and then incubating these with anti- Mg2+ for 20 min for full activation. After washing with PBS, the cells mouse or anti-rabbit IgG (Jackson ImmunoResearch Laboratories; were fixed in 2% paraformaldehyde for 15 min, washed, blocked with Downloaded from 1.2 mg/ml in PBS at a molar ratio of activator dye Ab:reporter dye of I-buffer (PBS, 1% BSA, 0.25% gelatin, 20 mM glycine), and incubated 2.0–3.0:1:0.6–1.0 for 30 min at room temperature). Labeled secondary with anti-a integrin (clone M17/4) Ab or isotype control Ab in I-buffer Abs were purified by gel filtration using Nap-5 columns (GE Healthcare) L for 30 min at 4˚C. After washing, cells were fixed in 1% paraformaldehyde according to the manufacturer’s instructions. For stochastic optical re- containing 0.1% glutaraldehyde to prevent capping induced by secondary construction microscopy (STORM) analysis of CD37 and b integrin 2 Abs. After washing and blocking in I-buffer, samples were incubated with distribution, HL-60 promyelocytic leukemia cells, sourced from the anti-rat IgG (Cappel Laboratories) with 15 nm gold particles coupled to American Type Culture Collection (ATCC, Manassas, VA), were differ- 5 Protein A, and were fixed in 1% glutaraldehyde. Subsequently, samples entiated (dHL-60) by incubating 3 3 10 cells/ml in 10% FCS-RPMI 1640 were washed, dried, dehydrated, and analyzed in a Hitachi H-7500 trans- supplemented with 1.25% DMSO and 1 mM all-trans-retinoic acid 5 mission electron microscope operating at 80 kV. Digital images of electron (Sigma-Aldrich) for 4 d. dHL60 cells (2 3 10 ) were allowed to adhere to micrographs were collected on a Gatan Multiscan camera and were pro- ICAM-1 (9 nM)–coated Nunc Lab-Tek II chambered coverglass (Thermo cessed with custom-written software based on MATLAB (MathWorks). Fisher Scientific, Scoresby, Australia) in the presence of 100 nM fMLF The distribution pattern of a b integrin and interparticle distance was (Sigma-Aldrich) for 20 min at 37˚C, and then fixed in 4% paraformalde- L 2 analyzed as described previously (36). hyde for 20 min at room temperature. Cells were then permeabilized in 0.01% Triton-X (Sigma-Aldrich) and incubated with combinations of Assessment of Rho GTPase activation mouse anti-human CD37 Ab WR17 (4 mg/ml), rabbit anti-human b2 integrin Ab (2 mg/ml), mouse anti-human CD11b MEM-174 (2 mg/ml), Activation of RhoA, Rac-1, and Cdc42 were assessed in WT and CD372/2 and rabbit anti-human transferrin receptor (2 mg/ml) for 30 min at room neutrophils using G-LISA activation assay kits (Cytoskeleton, Denver, The Journal of Immunology 5773

CO). Neutrophils were either untreated or exposed to PMA (100 ng/ml, 30 s or 5 min), and the cells were immediately lysed and Rho activity was assessed in cell lysates as per the manufacturer’s instructions.

Statistical analysis All data are presented as mean 6 SEM. Comparisons between groups were determined using Student t tests. A p value ,0.05 was deemed signiﬁcant.

Results CD37 ablation impairs neutrophil recruitment and endothelial interactions We first addressed whether the absence of CD37 affects the ability of neutrophils to undergo recruitment to sites of inflammation. In the thioglycollate peritonitis model, CD372/2 mice showed a significant reduction in the number of recruited neutrophils (Fig. 1A), despite having normal numbers of circulating neutrophils (WT 9.1 6 1.8 3 105/ml versus CD372/2 7.7 6 1.0 3 105/ml; n = 11 per group). This provides evidence of a role for CD37 in neutrophil recruitment. To examine this contribution of CD37 further, we used intravital microscopy to examine neutrophil- endothelial interactions in the microvasculature of WT and CD372/2 mice. In the absence of an inflammatory stimulus, leukocyte rolling flux and velocity did not differ between WT and CD372/2 mice (Fig. 1B–E basal time point and data not shown; see also Supplemental Videos 1, 2), indicating that in contrast to the tetraspanin CD63 (4), CD37 does not regulate P-selectin–dependent leukocyte rolling interactions. We next examined the response during acute exposure to CXCL1, a chemokine previously shown to induce neutrophil adhesion and transmigration (25, 37, FIGURE 1. Neutrophil recruitment is impaired in CD372/2 mice. (A) 38). In WT mice, CXCL1 induced robust increases in leukocyte Comparison of thioglycollate-induced peritoneal neutrophil recruitment in adhesion and transmigration after 30 and 60 min, without mark- WT and CD372/2 mice. Neutrophils were enumerated by flow cytometry edly affecting leukocyte rolling (Fig. 1B–E, Supplemental Video 24 h after thioglycollate injection. Data points represent individual mice. 3). At these same time points, CD372/2 mice displayed signifi- Mean 6 SEM are also shown. (B–F) Leukocyte-endothelial cell interactions 2/2 cant reductions in adhesion and transmigration (Fig. 1D, 1E; in cremasteric postcapillary venules of WT and CD37 mice, prior to (“Basal”) and after 30 and 60 min of CXCL1 superfusion (n =5mice/ by guest on October 1, 2021. Copyright 2015 Pageant Media Ltd. Supplemental Video 4). In addition, leukocyte rolling velocity was B C significantly faster in CD372/2 mice after 30 and 60 min of ex- group). Data are shown for ( ) leukocyte rolling flux, ( ) leukocyte rolling velocity, (D) leukocyte adhesion, and (E) transmigration. (F) Transmigration posure to CXCL1 (Fig. 1C), possibly reflecting the reduction in efficiency (i.e., ratio of leukocyte adhesion to transmigration after 30 and leukocyte adhesion in these mice. Transmigration efficiency, 60 min CXCL1 superfusion). Data are shown as mean 6 SEM. See also which represents the ratio of leukocyte adhesion to transmigration, Supplemental Videos 1–4. *p , 0.05, **p , 0.01, ***p , 0.001 versus did not differ between the two strains of mice (Fig. 1F). This result basal conditions in same genotype; †p , 0.05 versus WT. indicated that the defect in leukocyte transmigration observed in CD372/2 mice stemmed primarily from an impairment in adhesion. These impaired responses were observed despite normal expres- fore, one possible explanation for the reduced capacity of CD372/2 https://www.jimmunol.org a a a sion of L-selectin, CXCR2, and the L, M,and 4 integrins on leukocytes to adhere to ICAM-1 is that CD37 regulates the ex- 2/2 CD37 neutrophils (Fig. 2). pression of high-affinity b2 integrin on the cell surface. To de- To confirm the role of CD37 in promoting leukocyte adhesion, termine whether CD37 affected integrin inside out signaling, we 2/2 we next compared the ability of WT and CD37 neutrophils to examined the display of high-affinity b2 integrins by measuring adhere to adhesion molecule-coated coverslips in an in vitro flow the binding of soluble ICAM-1–Fc to activated neutrophils. chamber assay (Fig. 3). Previous studies have demonstrated that Binding of ICAM-1–Fc to CD372/2 neutrophils did not differ

Downloaded from neutrophil adhesion induced via CXCR2 ligands is primarily de- from that of WT neutrophils under the conditions examined pendent on LFA-1 (39). Therefore, in these experiments, we (Fig. 3C), indicating that CD37 has no role in regulating the b assessed 2 integrin–dependent interactions via incorporation of conformational state of the b2 integrins. Together with the b the 2 integrin ligand ICAM-1 in the adhesive substrate, along in vivo findings, these data demonstrate that the absence of with P-selectin and CXCL1 to promote rolling and arrest re- CD37 results in a reduced capacity of leukocytes to undergo b2 spectively. Similar to the in vivo findings, in the in vitro flow integrin–dependent adhesion under flow conditions, but has no chamber assay neither the number of rolling leukocytes (Fig. 3A) effect on the conformational state of the b2 integrins. nor leukocyte rolling velocity (WT, 15.7 6 0.9 mm/s; CD372/2, 13.6 6 0.5 mm/s; n = 5 mice/group) differed between WT and Absence of CD37 impairs the ability of neutrophils to respond CD372/2 leukocytes. In contrast, adhesion of CD372/2 leuko- to chemotactic cues in the interstitium cytes was significantly impaired relative to WT leukocytes Effective neutrophil recruitment to sites of inflammation requires (Fig. 3B). In these experiments, adhesion could be inhibited using leukocytes to respond appropriately to chemotactic stimuli in the a function-blocking anti-b2 integrin mAb, confirming that it was interstitium. Therefore, we next assessed the effect of CD37 de- b2 integrin–dependent (Fig. 3B). Integrins can be displayed at the ficiency on interstitial leukocyte migration. This was achieved using cell surface in both low- and high-affinity conformations. There- an in vivo chemotaxis assay in which extravascular leukocytes 5774 CD37 AND b2 INTEGRIN–MEDIATED ADHESION

neutrophils, we analyzed their distribution using two-color superresolution microscopy (N-STORM). As an appropriate Ab against mouse CD37 is not available, for these experiments we used the human HL-60 cell line. Cells were imaged following adhesion to ICAM-1 and activation with fMLF. Coclustering of CD37 and the b2 integrin was assessed using univariate and bivariate Getis and Franklin’s local Ripley’s K-function analysis (34) (Fig. 5). As a positive control, coclustering of CD18 with CD11b was assessed, and coclustering of CD37 with the TR receptor was assessed as a negative control, as previous work has shown that these two molecules do not coassociate (40). As ex- pected, CD18 and CD11b achieved a high level of coclustering (Fig. 5A–C, Supplemental Fig. 1A). In contrast, the level of coclustering of CD37 and CD18 was significantly lower than that of CD18 with CD11b (Fig. 5C–E, Supplemental Fig. 1B) and not significantly different from that of the CD37/TR negative control (Fig. 5F–H, Supplemental Fig. 1C). Thus, using this approach, we were unable to detect significant interactions between CD37 and b2 integrins. To determine whether CD37 ablation affected b2 integrin clustering, WT and CD372/2 mouse neutrophils were adhered to ICAM-1, and LFA-1 was labeled with 15-nm gold particles. Samples were analyzed by whole-mount electron microscopy (Supplemental Fig. 2A) where cluster formation was determined by calculating interparticulate distances with a near-neighbor analysis, and cluster size was rated according to the number of beads per cluster (Supplemental Fig. 2B, 2C). Although we

FIGURE 2. Absence of CD37 does not alter neutrophil expression of

by guest on October 1, 2021. Copyright 2015 Pageant Media Ltd. homing molecules. Expression of (A) LFA-1 (aL), (B) Mac-1 (aM), (C) L-selectin (CD62L), (D) CXCR2, and (E) a4 integrin were compared on circulating neutrophils (Ly6G+ cells) from WT and CD372/2 mice using ﬂow cytometry. Histograms shown are representative of n = 3–4 independent experiments.

are visualized while responding to a chemotactic gradient, established via release of CXCL1 from an agarose bead immobilized on the cremaster muscle (25). In this assay, neither the velocity nor https://www.jimmunol.org the mean displacement of migrating leukocytes differed between WT and CD372/2 mice (Fig. 4A, B). In contrast, the track length of migrating leukocytes was significantly greater in CD372/2 versus WT mice (Fig. 4C). Similarly, the meandering index, in- dicative of the capacity of the leukocyte to follow the directional cues provided by the CXCL1 gradient effectively, was signifi- 2/2 2 2 FIGURE 3. CD37 leukocytes adhere less efficiently to ICAM-1

Downloaded from / cantly reduced in CD37 mice (Fig. 4D). The latter finding is under flow conditions, but display normal levels of high-affinity b2 further illustrated by comparison of the paths of migration of WT integrin. (A and B) An in vitro flow chamber adhesion assay was used to 2 2 and CD37 / leukocytes, with the WT leukocytes migrating more assess interactions of WT and CD372/2 leukocytes. Blood from WT or 2 2 directly toward the source of the chemotactic stimulus than the CD37 / mice was diluted in HBSS and perfused over a substrate con- CD372/2 leukocytes (Fig. 4E, 4F). These data indicate that CD37 sisting of P-selectin, CXCL1, and ICAM-1 Fc fusion protein. (A) Rolling of WT and CD372/2 leukocytes on complete substrate. (B) Adhesion of is not required for detection of a CXCL1-derived chemotactic 2/2 gradient by leukocytes, but is important in enabling the leukocyte WT and CD37 leukocytes on substrates containing either ICAM-1 Fc or Fc alone. In some experiments, inhibitory anti–b integrin Ab was in- to follow the gradient accurately. 2 cluded to confirm that adhesion was b2 integrin–dependent. For ICAM-1- b CD37 and b2 integrin do not cocluster significantly in human Fc/Fc–alone experiments, n = 5 mice per genotype. For anti– 2 integrin neutrophils experiments, n = 3. Data are shown as mean 6 SEM. (C) Assessment of high-affinity b integrin expression on WT and CD372/2 neutrophils. b 2 The previous data show that CD37 ablation impairs 2 integrin- Neutrophil binding of soluble ICAM-1 Fc, or Fc alone (both at 100 mg/ml), dependent processes of neutrophil adhesion to the vascular en- under static conditions, was determined using flow cytometry. Data are dothelium (Fig. 1) and to ICAM-1 (Fig. 3). To understand the shown as mean 6 SEM of n = 4 independent experiments. *p , 0.05, molecular relationship between CD37 and the b2 integrin in **p , 0.01 versus WT ICAM-1 Fc. The Journal of Immunology 5775

ternalization revealed that internalization progressively increased over 60 min in both WT and CD372/2 cells, but was greater in CD372/2 neutrophils at every time point examined (Fig. 6C). These ﬁndings were conﬁrmed with confocal microscopy (Fig. 6D). These data are consistent with a role for CD37 in stabiliz- ing expression of LFA-1 on the surface under ligand-binding conditions. Neutrophil CD37 promotes cytoskeletal rearrangement, cell spreading, and polarization and dysregulates Rac-1 activation

b2 integrin–mediated outside-in signaling is critical to stabiliza- tion of leukocyte adhesion via leukocyte spreading and cytoskeletal rearrangement, as well as subsequent polarization and migration (42). As such, alterations in these behaviors may represent a mechanism whereby the absence of CD37 modulates leukocyte adhesive function. Therefore, we next compared cytoskeletal function of WT and CD372/2 neutrophils by assessing actin polymerization and cell spreading on ICAM-1. FITC- phalloidin was used to label polymerized actin, and staining was quantified with flow cytometry and confocal microscopy (Fig. 7A, 7B). Both approaches revealed that actin polymerization was significantly impaired in CD372/2 neutrophils relative to WT cells. These changes were associated with a significant reduction in cell spreading of CD372/2 neutrophils (Fig. 7C). In- deed, although .40% of WT neutrophils achieved a cell area .120 mm2, only ∼20% of CD372/2 neutrophils reached this size (p , 0.05 versus WT; Fig. 7D). These findings were supported by a reduction in the capacity of adherent CD372/2 neutrophils to undergo polarization (Fig. 7E), as indicated by a circularity index

FIGURE 4. Directed migration of leukocytes in the interstitium is altered in the absence of CD37. (A–D) Directed extravascular leukocyte migration was induced in cremaster muscles of WT and CD372/2 mice via controlled local release of CXCL1, and was examined by intravital microscopy. Mi- gration was quantitatively assessed with analysis of the following parame- A B C D by guest on October 1, 2021. Copyright 2015 Pageant Media Ltd. ters: ( ) migration velocity, ( ) displacement, ( ) track length, and ( ) meandering index (displacement/track length). Data were derived from 7–8 mice per group, 8–10 cells per mouse. Data are shown as mean 6 SEM. (E and F) Representative migration paths of neutrophils in the CXCL1-induced in vivo chemotaxis assay in (E)WTand(F)CD372/2 mice. Cell migration tracks in XY plane are displayed as arising from a common point of origin. The position of the source of CXCL1 relative to the paths of migration is also depicted. *p , 0.05, **p , 0.01. https://www.jimmunol.org readily detected b2 integrin clusters on mouse neutrophils, the absence of CD37 did not affect their size and frequency. We conclude that CD37 does not cluster signiﬁcantly with CD18, and further that CD37 is not required for CD18 clustering.

CD37 restrains b2 integrin internalization

Downloaded from Integrin-mediated cellular adhesion is a dynamic process, partic- ularly in migrating cells, where endocytosis of integrins plays a critical role in detaching and removing integrins from the trailing

edge, and redistributing these to the leading edge (41). As such, the FIGURE 5. CD37 and the b2 integrin do not cocluster significantly in 2 2 defective adhesive and migratory behavior of CD37 / leukocytes human neutrophils. (A, D, and G) Reconstructed dual-color single mole- might be explained by dysregulation in integrin internalization in cule N-STORM images, and enlarged views (B, E, and H) showing the b A B b the absence of CD37 and consequent reduced stability of integrin distributions of CD11b and CD18 ( 2 integrin) ( and ), CD37 and 2 expression upon ligand engagement. To address this issue, we integrin (D and E) and CD37 and the transferrin receptor (TR) (G and H)in compared internalization of LFA-1 in WT and CD372/2 neutro- HL-60 cells following adhesion to ICAM-1 and activation with fMLF. White boxes indicate representative windows (3 3 3 mm) used for sub- phils quantitatively using flow cytometry and confocal micros- sequent second-order Ripley’s K analysis. Yellow denotes coclustering. copy. WT neutrophils underwent substantial LFA-1 internalization Scale bars represent 5 mm(A, D, and G) or 0.5 mm(B, E, and H). (C) within 15 min of adhesion to ICAM-1 (Fig. 6A, 6B, 6D). How- Proportion of CD18 localizations coclustered with CD37 or CD11b (n = 2/2 ever, LFA-1 internalization was significantly greater in CD37 7). (F) Proportion of CD37 localizations coclustered with CD18 (n =5)or neutrophils (Fig. 6A), as indicated by an increase in the inter- TR (n = 5). Data in (C) and (F) were derived from Quadrant 4 in nalization index (Fig. 6B). Analysis of the kinetics of LFA-1 in- Supplemental Fig. 1A–C. 5776 CD37 AND b2 INTEGRIN–MEDIATED ADHESION

FIGURE 6. LFA-1 internalization is dysregulated in CD372/2 neutrophils. Assessment of LFA-1 internalization in WT and CD372/2 neutrophils via ﬂow cytometry and confocal microscopy. (A) Flow cytometry plots illustrating the stages of the internalization assay. Initially, total cell surface LFA-1

(shown as Total) was labeled using anti–LFA-1 and quantitated as total MFI (MFIT). After incubation to allow for LFA-1 internalization, the residual surface anti–LFA-1 was stripped off the cell surface, and the remaining LFA-1 staining was quantitated (shown as Internalized) and deﬁned as in-

ternalized MFI (MFII). (B) Integrin internalization index was determined as the ratio of MFII to MFIT. Data are shown as mean 6 SEM of n =5 independent experiments. (C) Time course of LFA-1 internalization. LFA-1 internalization was assessed, as performed in (C), after 5, 15, 30, and 60 min in WT and CD372/2 neutrophils. Data are shown as mean 6 SEM of n = 6 independent experiments. (D) Confocal micrographs of neutrophils from experiments in (C) showing surface LFA-1 (left panels) and internalized LFA-1 (right panels)inWTandCD372/2 cells. Individual panels are shown by guest on October 1, 2021. Copyright 2015 Pageant Media Ltd. for anti-Ly6G (green indicating the neutrophil cell membrane) and anti–LFA-1 (aL in red), as well as the merged image. Scale bars, 10 mm. *p , 0.05 versus WT via two-way ANOVA.

of ,0.3 (Fig. 7F). We also compared activation of the Rho family idence that tetraspanins regulate b2 integrin function in leuko- 2/2 of GTPases in WT and CD37 neutrophils via quantitation of cytes. b2 integrins have been reported to be components of the GTP-bound forms of RhoA, Rac-1, and Cdc42 in activated tetraspanin-enriched microdomains (52, 53). However, the func- cells. After PMA stimulation, CD372/2 cells displayed signiﬁ- tional relevance of these observations is unclear. Currently the https://www.jimmunol.org cantly elevated active Rac-1 relative to that seen in WT cells most compelling evidence that tetraspanins regulate b2 integrins (Fig. 7G), whereas Cdc42 activation was comparable between comes from in vitro studies by Quast et al. (54), who showed in the genotypes and RhoA activation was undetectable under these migrating dendritic cells that the tetraspanin CD81 and aLb2 conditions (data not shown). Together, these data indicate that colocalized at the leading edge, and that silencing CD81 reduced CD37 plays an important role in the cytoskeletal rearrangements dendritic cell adhesion to ICAM-1. Whether tetraspanins regulate that occur after neutrophil binding to the b2 integrin ligand b2 integrins in neutrophils or in vivo immune responses has not

Downloaded from ICAM-1, and that Rac-1 activation is dysregulated in the absence been examined. of CD37. In this study, we examined the role of the tetraspanin CD37 in neutrophil function by analyzing the classical sequence of Discussion leukocyte-endothelial cell interactions in WT and CD372/2 mice, Tetraspanins regulate molecular function by organizing their revealing compelling evidence that CD37 regulates b2 integrin partner proteins into signal transducing microdomains. The best function in an in vivo inflammatory response. CD372/2 leuko- characterized of the tetraspanin molecular partners are integrins. cytes activated by the neutrophil-attracting chemokine CXCL1 Indeed, in nonimmune cells the evidence that tetraspanins can demonstrated impaired adhesion to the vascular endothelium regulate integrin-mediated functions is overwhelming, incorpo- (Fig. 1C–F), a response previously shown to be LFA-1–dependent rating roles in wound healing (43), angiogenesis (44), hemostasis (37). This resulted in diminished leukocyte extravasation. This (45), epidermal integrity (46), glomerular filtration (47, 48), and impairment in migration and adhesion was reflected in the reduced cancer cell metastasis (49–51). By comparison, the evidence that neutrophil inflammation induced by thioglycollate in the perito- tetraspanins regulate adhesion and migration of leukocytes is nitis model (Fig. 1A). In vitro assessment of CD37-deficient limited. Whereas CD81 and CD37 have been found to modulate neutrophils under flow conditions provided further evidence that the function of a4b1 in B cells (8, 11), there is little strong ev- CD37 modulates b2 integrin function, in that they showed reduced The Journal of Immunology 5777

These experiments also showed that CD37 ablation impairs outside-in dependent parameters, including actin polymerization and cell spreading and polarization (Fig. 7), but does not affect the display of high-affinity b2 integrins on the neutrophil surface (Fig. 3C). These findings are in line with previous reports that have shown that tetraspanin-mediated effects on integrin function occur at the level of outside-in signaling, as evidenced by the effects on cell spreading (45, 55) and adhesion strengthening under flow (8, 56), observed for b1andb3 integrins. In addition, we observed a role for CD37 in restraining the endocytosis of b2 integrins following adhesion to ICAM-1 (Fig. 6), a function that is also dependent on the cytoskeleton (57). Together with the effects on b2 integrin–dependent adhesion, these observations indicate that the leukocyte-restricted tetraspanin CD37 has a key role in regulating the function of the leukocyte-restricted b2 integrin family. However, the underlying mechanism is likely to be different to the well-characterizedregulationofb1 integrins by tetraspanins, where several robust and direct molecular partner interactions have been demonstrated (58). In contrast, in the current study, high-stringency N-STORM analyses of CD37 and b2 integrin distribution in human neutrophils adherent to ICAM-1 failed to detect coclustering that was significantly different from negative controls (Fig. 5). Moreover, the absence of CD37 from mouse neutrophils did not affect the size and frequency of CD18 clusters (Supplemental Fig. 2). Although we cannot rule out a functionally relevant transient or weak interaction between CD37 and b2 integrin, we argue that the most likely model is that CD37 facilitates communication between b2 integrins and the cytoskeleton. This model is consistent with several reports that tetraspanins modulate the activity of the cytoskeleton-regulating Rho GTPases (44, 51, 59–62), and interact with and regulate the activity of Ezrin-radixin-moesin proteins that can control cytoskeletal rearrangement through the Rho pathway (63, 64). Thus, the impairment in the ability 2/2

by guest on October 1, 2021. Copyright 2015 Pageant Media Ltd. of CD37 neutrophils to bind stably to the vascular endothelium can be explained first by a defect in cytoskeletal rearrangement (Fig. 7), which leads to an impairment of neutrophil spreading on the endothelium. Second, an elevated rate of microtubule-dependent b2 integrin endocytosis in the absence of CD37 (Fig. 6) would result in weakening of the ad- FIGURE 7. Absence of CD37 impairs neutrophil actin polymerization, cell hesive interaction of the neutrophil with the endothelium. spreading, and polarization. (A) Flow cytometric histogram showing phal- Impairments in spreading, polarity, and integrin recycling are also loidin staining as an index of actin polymerization, in WT and CD372/2 likely to contribute to the lack of directionality we observed in 2 2 https://www.jimmunol.org neutrophils after adhesion to ICAM-1. Data are representative of n =3in- CD37 / neutrophil migration in the interstitium (Fig. 4). In this dependent experiments. (B) MFI of phalloidin-FITC staining in WT and study, we note that extracellular neutrophil migration is dependent 2/2 CD37 neutrophils adherent to ICAM-1, as determined from confocal on integrins other than the b family, with studies indicating that C D 2/2 2 microscopy images. ( and ) Surface area of adherent WT and CD37 a b is of particular importance to this function (65, 66). This raises neutrophils, shown for individual cells (C), and as a frequency distribution of 2 1 the possibility that in addition to its effect on the function of b surface areas, categorized as 0–60, 60–120, and .120 mm2 (D). Results were 2 pooled from n = 3 experiments. Data in (B)and(C) show results for individual integrins, CD37 also modulates the actions of other integrins that Downloaded from cells and mean 6 SEM. (E and F) Quantitative analysis of polarization of contribute to neutrophil extravascular migration. This would be adherent WT and CD372/2 neutrophils. Images of cells (E) were assessed by consistent with our previous observations on the actions of CD37 in quantitation of cell circularity (F)whereavalue,0.3 equates to a highly regulating a4b1 integrin function in B cells (11). polarized cell [indicated by arrows in (E)], 0.30–0.49 equates to a moderately In conclusion, these data demonstrate that the tetraspanin CD37 polarized cell, and $0.5 equates to a minimally polarized cell [indicated by is a key player in neutrophil recruitment and, therefore, acute arrowheads in (E)]. Data in (F) were derived from analysis of n =3miceper inflammation. These observations raise the possibility that immu- ∼ G 2/2 group, and 30 cells per mouse. ( ) Rac-1 activation in WT and CD37 notherapeutic reagents that target CD37 to treat B cell malignancies neutrophils. Neutrophils were either untreated or activated with PMA, and cell might have unforeseen consequences in impairing neutrophil lysates prepared for Rac-1 activity assessment after 30 s. Data represent an recruitment, and potentially the capacity to mount an effective analysis of six independent assays and are shown as mean 6 SEM. *p , 0.05, ***p , 0.001, †p , 0.05 versus WT using contingency analysis. acute inflammatory response. Acknowledgments b adhesion to the 2 integrin ligand, ICAM-1 (Fig. 3A, 3B). To- We thank the staff members of the Monash Animal Facility at both gether, these data provide clear evidence that CD37 has an im- Clayton and AMREP campuses, the staff members of Monash Micro portant role in supporting b2 integrin function. Imaging Facility for assistance with confocal imaging experiments, Alie 5778 CD37 AND b2 INTEGRIN–MEDIATED ADHESION

van der Schaff for TEM analysis, Mariam Sofi for technical assistance, and 23. Gartlan, K. H., J. L. Wee, M. C. Demaria, R. Nastovska, T. M. Chang, Maria Demaria for technical assistance and reading the manuscript. E. L. Jones, V. Apostolopoulos, G. A. Pietersz, M. J. Hickey, A. B. van Spriel, and M. D. Wright. 2013. Tetraspanin CD37 contributes to the initiation of cellular immunity by promoting dendritic cell migration. Eur. J. Immunol. 43: 1208–1219. Disclosures 24. Schwartz-Albiez, R., B. Do¨rken, W. Hofmann, and G. Moldenhauer. 1988. The The authors have no financial conflicts of interest. B cell-associated CD37 antigen (gp40-52). Structure and subcellular expression of an extensively glycosylated glycoprotein. J. Immunol. 140: 905–914. 25. Hickey, M. J., M. Forster, D. Mitchell, J. Kaur, C. De Caigny, and P. Kubes. References 2000. L-selectin facilitates emigration and extravascular locomotion of leukocytes during acute inflammatory responses in vivo. J. Immunol. 165: 7164–7170. 1. Ley, K., C. Laudanna, M. I. Cybulsky, and S. Nourshargh. 2007. Getting to the 26. Cheng, Q., A. Hoi, M. J. Hickey, and E. F. Morand. 2012. Lymphocytes from site of inflammation: the leukocyte adhesion cascade updated. Nat. Rev. Immu- systemic lupus erythematosus patients display increased spreading on VCAM-1, nol. 7: 678–689. an effect associated with active renal involvement. Lupus 21: 632–641. 2. Barreiro, O., M. Yań˜ez-Mo´, M. Sala-Valde´s, M. D. Gutie´rrez-Lo´pez, S. Ovalle, 27. Khan, A. I., S. M. Kerfoot, B. Heit, L. Liu, G. Andonegui, B. Ruffell, P. Johnson, A. Higginbottom, P. N. Monk, C. Cabanãs, and F. Sańchez-Madrid. 2005. En- and P. Kubes. 2004. Role of CD44 and hyaluronan in neutrophil recruitment. J. dothelial tetraspanin microdomains regulate leukocyte firm adhesion during Immunol. 173: 7594–7601. extravasation. Blood 105: 2852–2861. 28. Chan, J. R., S. J. Hyduk, and M. I. Cybulsky. 2003. Detecting rapid and transient 3. Barreiro, O., M. Zamai, M. Yań˜ez-Mo´, E. Tejera, P. Lo´pez-Romero, P. N. Monk, upregulation of leukocyte integrin affinity induced by chemokines and chemo- E. Gratton, V. R. Caiolfa, and F. Sańchez-Madrid. 2008. Endothelial adhesion attractants. J. Immunol. Methods 273: 43–52. receptors are recruited to adherent leukocytes by inclusion in preformed tetra- 29. Cera, M. R., M. Fabbri, C. Molendini, M. Corada, F. Orsenigo, M. Rehberg, spanin nanoplatforms. J. Cell Biol. 183: 527–542. C. A. Reichel, F. Krombach, R. Pardi, and E. Dejana. 2009. JAM-A promotes 4. Doyle, E. L., V. Ridger, F. Ferraro, M. Turmaine, P. Saftig, and D. F. Cutler. neutrophil chemotaxis by controlling integrin internalization and recycling. J. 2011. CD63 is an essential cofactor to leukocyte recruitment by endothelial Cell Sci. 122: 268–277. Blood P-selectin. 118: 4265–4273. 30. Le, Y., W. Gong, B. Li, N. M. Dunlop, W. Shen, S. B. Su, R. D. Ye, and 5. Termini, C. M., M. L. Cotter, K. D. Marjon, T. Buranda, K. A. Lidke, and J. M. Wang. 1999. Utilization of two seven-transmembrane, G protein-coupled J. M. Gillette. 2014. The membrane scaffold CD82 regulates cell adhesion by al- receptors, formyl peptide receptor-like 1 and formyl peptide receptor, by the tering a4 integrin stability and molecular density. Mol. Biol. Cell 25: 1560–1573. synthetic hexapeptide WKYMVm for human phagocyte activation. J. Immunol. 6. Yań˜ez-Mo´, M., O. Barreiro, M. Gordon-Alonso, M. Sala-Valde´s, and F. Sańchez- 163: 6777–6784. Madrid. 2009. Tetraspanin-enriched microdomains: a functional unit in cell 31. Ihaka, R., and R. Gentleman. 1996. R: A Language for Data Analysis and plasma membranes. Trends Cell Biol. 19: 434–446. Graphics. J. Comput. Graph. Stat. 5: 299–314. 7. Jones, E. L., M. C. Demaria, and M. D. Wright. 2011. Tetraspanins in cellular 32. Baddeley, A., and R. Turner. 2005. Spatstat: an R package for analyzing spatial immunity. Biochem. Soc. Trans. 39: 506–511. point patterns. J. Stat. Softw. 12: 1–42. 8. Feigelson, S. W., V. Grabovsky, R. Shamri, S. Levy, and R. Alon. 2003. The 33. Getis, A., and J. Franklin. 1987. 2nd-order neighborhood analysis of mapped CD81 tetraspanin facilitates instantaneous leukocyte VLA-4 adhesion strength- point patterns. Ecology 68: 473–477. ening to vascular cell adhesion molecule 1 (VCAM-1) under shear flow. J. Biol. 34. Rossy, J., E. Cohen, K. Gaus, and D. M. Owen. 2014. Method for co-cluster Chem. 278: 51203–51212. analysis in multichannel single-molecule localisation data. Histochem. Cell Biol. 9. Knobeloch, K. P., M. D. Wright, A. F. Ochsenbein, O. Liesenfeld, J. Lo¨hler, 141: 605–612. R. M. Zinkernagel, I. Horak, and Z. Orinska. 2000. Targeted inactivation of the 35. Ripley, B. D. 1991. Statistical Inference for Spatial Processes. Cambridge tetraspanin CD37 impairs T-cell-dependent B-cell response under suboptimal University Press, Cambridge. costimulatory conditions. Mol. Cell. Biol. 20: 5363–5369. 36. Cambi, A., B. Joosten, M. Koopman, F. de Lange, I. Beeren, R. Torensma, 10. van Spriel, A. B., M. Sofi, K. H. Gartlan, A. van der Schaaf, I. Verschueren, ´ R. Torensma, R. A. Raymakers, B. E. Loveland, M. G. Netea, G. J. Adema, et al. J. A. Fransen, M. Garcia-Parajo, F. N. van Leeuwen, and C. G. Figdor. 2006. 2009. The tetraspanin protein CD37 regulates IgA responses and anti-fungal Organization of the integrin LFA-1 in nanoclusters regulates its activity. Mol. immunity. PLoS Pathog. 5: e1000338. Biol. Cell 17: 4270–4281. 11. van Spriel, A. B., S. de Keijzer, A. van der Schaaf, K. H. Gartlan, M. Sofi, 37. Lee, J., G. Cacalano, T. Camerato, K. Toy, M. W. Moore, and W. I. Wood. 1995. A. Light, P. C. Linssen, J. B. Boezeman, M. Zuidscherwoude, I. Reinieren- Chemokine binding and activities mediated by the mouse IL-8 receptor. J. a b Immunol. 155: 2158–2164. by guest on October 1, 2021. Copyright 2015 Pageant Media Ltd. Beeren, et al. 2012. The tetraspanin CD37 orchestrates the (4) (1) integrin-Akt signaling axis and supports long-lived plasma cell survival. Sci. Signal. 5: ra82. 38. Santos, L. L., H. Fan, P. Hall, D. Ngo, C. R. Mackay, G. Fingerle-Rowson, DOI: 10.1126/scisignal.2003113. R. Bucala, M. J. Hickey, and E. F. Morand. 2011. Macrophage migration in- 12. Beckwith,K.A.,F.W.Frissora,M.R.Stefanovski,W.H.Towns,C.Cheney,X.Mo, hibitory factor regulates neutrophil chemotactic responses in inflammatory ar- J. Deckert, C. M. Croce, J. M. Flynn, L. A. Andritsos, et al. 2014. The CD37- thritis in mice. Arthritis Rheum. 63: 960–970. targeted antibody-drug conjugate IMGN529 is highly active against human CLL and 39. Phillipson, M., B. Heit, P. Colarusso, L. Liu, C. M. Ballantyne, and P. Kubes. in a novel CD37 transgenic murine leukemia model. Leukemia 28: 1501–1510. 2006. Intraluminal crawling of neutrophils to emigration sites: a molecularly 13. Byrd, J. C., J. M. Pagel, F. T. Awan, A. Forero, I. W. Flinn, D. P. Deauna- distinct process from adhesion in the recruitment cascade. J. Exp. Med. 203: Limayo, S. E. Spurgeon, L. A. Andritsos, A. K. Gopal, J. P. Leonard, et al. 2014. 2569–2575. A phase 1 study evaluating the safety and tolerability of otlertuzumab, an anti- 40. Escola, J. M., M. J. Kleijmeer, W. Stoorvogel, J. M. Griffith, O. Yoshie, and CD37 mono-specific ADAPTIR therapeutic protein in chronic lymphocytic H. J. Geuze. 1998. Selective enrichment of tetraspan proteins on the internal leukemia. Blood 123: 1302–1308. vesicles of multivesicular endosomes and on exosomes secreted by human B-lymphocytes. J. Biol. Chem. 273: 20121–20127.

https://www.jimmunol.org 14. Deckert, J., P. U. Park, S. Chicklas, Y. Yi, M. Li, K. C. Lai, M. F. Mayo, C. N. Carrigan, H. K. Erickson, J. Pinkas, et al. 2013. A novel anti-CD37 41. Jacquemet, G., M. J. Humphries, and P. T. Caswell. 2013. Role of adhesion antibody-drug conjugate with multiple anti-tumor mechanisms for the treat- receptor trafficking in 3D cell migration. Curr. Opin. Cell Biol. 25: 627–632. ment of B-cell malignancies. Blood 122: 3500–3510. 42. Abram, C. L., and C. A. Lowell. 2009. The ins and outs of leukocyte integrin 15. Flinn, I. W. 2011. CD37: the comeback kid. Blood 118: 4007–4008. signaling. Annu. Rev. Immunol. 27: 339–362. 16. Heider, K. H., K. Kiefer, T. Zenz, M. Volden, S. Stilgenbauer, E. Ostermann, 43. Cowin, A. J., D. Adams, S. M. Geary, M. D. Wright, J. C. Jones, and A. Baum, H. Lamche, Z. Kupc€ u,€ A. Jacobi, et al. 2011. A novel Fc-engineered L. K. Ashman. 2006. Wound healing is defective in mice lacking tetraspanin monoclonal antibody to CD37 with enhanced ADCC and high proapoptotic CD151. J. Invest. Dermatol. 126: 680–689. activity for treatment of B-cell malignancies. Blood 118: 4159–4168. 44. Takeda, Y., A. R. Kazarov, C. E. Butterfield, B. D. Hopkins, L. E. Benjamin, Downloaded from 17. Palomba, M. L., and A. Younes. 2013. In the spotlight: a novel CD37 antibody- A. Kaipainen, and M. E. Hemler. 2007. Deletion of tetraspanin Cd151 results in drug conjugate. Blood 122: 3397–3398. decreased pathologic angiogenesis in vivo and in vitro. Blood 109: 1524–1532. 18. Zhao, X., R. Lapalombella, T. Joshi, C. Cheney, A. Gowda, M. S. Hayden- 45. Lau, L. M., J. L. Wee, M. D. Wright, G. W. Moseley, P. M. Hogarth, Ledbetter, P. R. Baum, T. S. Lin, D. Jarjoura, A. Lehman, et al. 2007. Targeting L. K. Ashman, and D. E. Jackson. 2004. The tetraspanin superfamily member CD37-positive lymphoid malignancies with a novel engineered small modular CD151 regulates outside-in integrin alphaIIbbeta3 signaling and platelet func- immunopharmaceutical. Blood 110: 2569–2577. tion. Blood 104: 2368–2375. 19. Gartlan, K. H., G. T. Belz, J. M. Tarrant, G. Minigo, M. Katsara, K. C. Sheng, 46. Karamatic Crew, V., N. Burton, A. Kagan, C. A. Green, C. Levene, F. Flinter, M. Sofi, A. B. van Spriel, V. Apostolopoulos, M. Plebanski, et al. 2010. A R. L. Brady, G. Daniels, and D. J. Anstee. 2004. CD151, the first member of the complementary role for the tetraspanins CD37 and Tssc6 in cellular immunity. J. tetraspanin (TM4) superfamily detected on erythrocytes, is essential for the Immunol. 185: 3158–3166. correct assembly of human basement membranes in kidney and skin. Blood 104: 20. van Spriel, A. B., K. L. Puls, M. Sofi, D. Pouniotis, H. Hochrein, Z. Orinska, 2217–2223. K. P. Knobeloch, M. Plebanski, and M. D. Wright. 2004. A regulatory role for 47. Baleato, R. M., P. L. Guthrie, M. C. Gubler, L. K. Ashman, and S. Roselli. 2008. CD37 in T cell proliferation. J. Immunol. 172: 2953–2961. Deletion of CD151 results in a strain-dependent glomerular disease due to severe 21. Meyer-Wentrup, F., C. G. Figdor, M. Ansems, P. Brossart, M. D. Wright, alterations of the glomerular basement membrane. Am. J. Pathol. 173: 927–937. G. J. Adema, and A. B. van Spriel. 2007. Dectin-1 interaction with tetraspanin 48. Sachs, N., M. Kreft, M. A. van den Bergh Weerman, A. J. Beynon, T. A. Peters, CD37 inhibits IL-6 production. J. Immunol. 178: 154–162. J. J. Weening, and A. Sonnenberg. 2006. Kidney failure in mice lacking the 22. Sheng, K. C., A. B. van Spriel, K. H. Gartlan, M. Sofi, V. Apostolopoulos, tetraspanin CD151. J. Cell Biol. 175: 33–39. L. Ashman, and M. D. Wright. 2009. Tetraspanins CD37 and CD151 differen- 49. Copeland, B. T., M. J. Bowman, and L. K. Ashman. 2013. Genetic ablation of tially regulate Ag presentation and T-cell co-stimulation by DC. Eur. J. Immunol. the tetraspanin CD151 reduces spontaneous metastatic spread of prostate cancer 39: 50–55. in the TRAMP model. Mol. Cancer Res. 11: 95–105. The Journal of Immunology 5779

50. Fei, Y., J. Wang, W. Liu, H. Zuo, J. Qin, D. Wang, H. Zeng, and Z. Liu. 2012. 59. Delaguillaumie, A., C. Lagaudrie`re-Gesbert, M. R. Popoff, and H. Conjeaud. CD151 promotes cancer cell metastasis via integrins a3b1anda6b1 in vitro. 2002. Rho GTPases link cytoskeletal rearrangements and activation processes Mol. Med. Rep. 6: 1226–1230. induced via the tetraspanin CD82 in T lymphocytes. J. Cell Sci. 115: 433–443. 51. Takeda, Y., Q. Li, A. R. Kazarov, M. Epardaud, K. Elpek, S. J. Turley, and 60. Liu, W. M., F. Zhang, S. Moshiach, B. Zhou, C. Huang, K. Srinivasan, M. E. Hemler. 2011. Diminished metastasis in tetraspanin CD151-knockout S. Khurana, Y. Zheng, J. M. Lahti, and X. A. Zhang. 2012. Tetraspanin CD82 mice. Blood 118: 464–472. inhibits protrusion and retraction in cell movement by attenuating the plasma 52.Mantegazza,A.R.,M.M.Barrio,S.Moutel,L.Bover,M.Weck, membrane-dependent actin organization. PLoS One 7: e51797. P. Brossart, J. L. Teillaud, and J. Mordoh. 2004. CD63 tetraspanin slows 61. Takahashi, M., T. Sugiura, M. Abe, K. Ishii, and K. Shirasuna. 2007. Regulation down cell migration and translocates to the endosomal-lysosomal-MIICs of c-Met signaling by the tetraspanin KAI-1/CD82 affects cancer cell migration. route after extracellular stimuli in human immature dendritic cells. Blood Int. J. Cancer 121: 1919–1929. 62. Tejera, E., V. Rocha-Perugini, S. Lo´pez-Martıń, D. Pe´rez-Hernańdez, 104: 1183–1190. A. I. Bachir, A. R. Horwitz, J. Va´zquez, F. Sańchez-Madrid, and M. Yań˜ez-Mo. 53. Skubitz, K. M., K. D. Campbell, and A. P. Skubitz. 2000. CD63 associates with 2013. CD81 regulates cell migration through its association with Rac GTPase. CD11/CD18 in large detergent-resistant complexes after translocation to the cell Mol. Biol. Cell 24: 261–273. surface in human neutrophils. FEBS Lett. 469: 52–56. 63. Coffey, G. P., R. Rajapaksa, R. Liu, O. Sharpe, C. C. Kuo, S. W. Krauss, Y. Sagi, 54. Quast, T., F. Eppler, V. Semmling, C. Schild, Y. Homsi, S. Levy, T. Lang, R. E. Davis, L. M. Staudt, J. P. Sharman, et al. 2009. Engagement of CD81 C. Kurts, and W. Kolanus. 2011. CD81 is essential for the formation of mem- induces ezrin tyrosine phosphorylation and its cellular redistribution with fila- brane protrusions and regulates Rac1-activation in adhesion-dependent immune mentous actin. J. Cell Sci. 122: 3137–3144. cell migration. Blood 118: 1818–1827. 64. Sala-Valde´s, M., A. Ursa, S. Charrin, E. Rubinstein, M. E. Hemler, F. Sańchez- 55. Zhang, X. A., A. R. Kazarov, X. Yang, A. L. Bontrager, C. S. Stipp, and Madrid, and M. Yań ˜ez-Mo´. 2006. EWI-2 and EWI-F link the tetraspanin web to M. E. Hemler. 2002. Function of the tetraspanin CD151-alpha6beta1 integrin the actin cytoskeleton through their direct association with ezrin-radixin-moesin complex during cellular morphogenesis. Mol. Biol. Cell 13: 1–11. proteins. J. Biol. Chem. 281: 19665–19675. 56. Lammerding, J., A. R. Kazarov, H. Huang, R. T. Lee, and M. E. Hemler. 2003. 65. Werr, J., J. Johansson, E. E. Eriksson, P. Hedqvist, E. Ruoslahti, and L. Lindbom. Tetraspanin CD151 regulates alpha6beta1 integrin adhesion strengthening. Proc. 2000. Integrin alpha(2)beta(1) (VLA-2) is a principal receptor used by neutro- Natl. Acad. Sci. USA 100: 7616–7621. phils for locomotion in extravascular tissue. Blood 95: 1804–1809. 57. Pellinen, T., and J. Ivaska. 2006. Integrin traffic. J. Cell Sci. 119: 3723–3731. 66. Werr, J., X. Xie, P. Hedqvist, E. Ruoslahti, and L. Lindbom. 1998. beta1 58. Hemler, M. E. 2005. Tetraspanin functions and associated microdomains. Nat. integrins are critically involved in neutrophil locomotion in extravascular tissue Rev. Mol. Cell Biol. 6: 801–811. In vivo. J. Exp. Med. 187: 2091–2096. by guest on October 1, 2021. Copyright 2015 Pageant Media Ltd. https://www.jimmunol.org Downloaded from Wee et al. Supplementary information

Supplementary Information – Wee JL et al

Tetraspanin CD37 regulates 2 integrin-mediated adhesion and migration in neutrophils

Supplemental Video Descriptions

Video S1: Basal leukocyte rolling in postcapillary venule of wild-type mouse. Cremaster muscle preparation in wild-type mouse prior to inflammatory stimulation, illustrating constitutive leukocyte interactions in cremasteric postcapillary venules. Duration – 4 seconds.

Video S2: Basal leukocyte rolling in postcapillary venule of a CD37-/- mouse. Cremaster muscle preparation in a CD37-/- mouse prior to inflammatory stimulation, illustrating constitutive leukocyte interactions in cremasteric postcapillary venules. Duration – 7 seconds.

Video S3: CXCL1-induced leukocyte-endothelial cell interactions in wild-type mouse. Cremaster muscle preparation in wild-type mouse after 60 min exposure to CXCL1. Numerous adherent leukocytes are visible in the postcapillary venule, and transmigrated leukocytes (circles) are also visible adjacent to the blood vessel. Duration – 7 seconds.

Video S4: CXCL1-induced leukocyte-endothelial cell interactions in CD37-/- mouse. Cremaster muscle preparation in a CD37-/- mouse after 60 min exposure to CXCL1. Leukocyte adhesion is minimal and few leukocytes are present in the area adjacent to the venule. Duration – 7 seconds.

1 Wee et al. Supplementary information

Supplementary Figure 1. Co-cluster analysis of CD11b, 2 integrin, CD37 and transferrin receptor. Combined univariate and bivariate local Ripley’s K analysis

for CD11b and 2 integrin (A), CD37 and β2 integrin (B), and CD37 and the transferrin receptor (TR) (C) at a spatial scale of 50 nm. L(50) represents clustering within the same species and L(50)cross represents cross clustering with the second species. Cluster thresholds of 100 for L(50) and L(50)cross are indicated by horizontal and vertical black lines that divide the plot into 4 quadrants (Q1-Q4). These 4 quadrants therefore represent different types of clustering: Q1 represents free molecules showing no significant clustering; Q2 represents molecules in clusters with other molecules of the same species (i.e. self clusters); Q3 represents free molecules that by chance reside in clusters of molecules of the other species; and Q4 represents molecules that reside in genuine co-clusters of both species. Colours indicate the number of molecules at each spot. Trendlines showing the correlation between L(50) and L(50)cross were generated by fitting a linear model to the data using least- squares. Data are compiled from n=5 experiments in A, B and n=7 in C.

2 Wee et al. Supplementary information

Supplementary Figure 2: CD37 ablation does not affect LFA-1 clusters in mouse neutrophils. WT and CD37-/- neutrophils were adhered to ICAM-1 and the presence

of 2 integrin clusters was detected by anti-LFA-1 conjugated with 15 nm gold particles and assessment via transmission electron microscopy. A: Representative images of WT and CD37-/- neutrophils, showing i) standard view of immunogold staining, and ii) image analysis-based bead detection performed for clustering analysis. B: Histogram showing distribution of nearest-neighbor distances for LFA-1 in wild-type and CD37-/- neutrophils. C: Histogram of number of beads per LFA-1 cluster, as a read-out of cluster size, on wild-type and CD37-/- neutrophils. Data were generated from 10 micrographs per sample, and pooled from two experiments. Data are shown as mean ± SEM.