Mac-1 Regulates IL-13 Activity in Macrophages by Directly Interacting with IL-13Rα1*

Background: Mac-1 strongly suppresses IL-13-induced JAK/STAT activation in macrophages, but the mechanism is unknown. Results: Our data demonstrate that Mac-1 interacts with the IL-13Rα1 subunit of IL-13R and thereby suppresses IL-13 signaling. Conclusion: Mac-1 regulates macrophage to foam cell transformation by binding to IL-13Rα1. Significance: This study identifies a novel interaction and provides a potential mechanism by which Mac-1 safeguards macrophages from foam cell differentiation. Mac-1 exhibits a unique inhibitory activity toward IL-13-induced JAK/STAT activation and thereby regulates macrophage to foam cell transformation. However, the underlying molecular mechanism is unknown. In this study, we report the identification of IL-13Rα1, a component of the IL-13 receptor (IL-13R), as a novel ligand of integrin Mac-1, using a co-evolution-based algorithm. Biochemical analyses demonstrated that recombinant IL-13Rα1 binds Mac-1 in a purified system and supports Mac-1-mediated cell adhesion. Co-immunoprecipitation experiments revealed that endogenous Mac-1 forms a complex with IL-13Rα1 in solution, and confocal fluorescence microscopy demonstrated that these two receptors co-localize with each other on the surface of macrophages. Moreover, we found that genetic inactivation of Mac-1 promotes IL-13-induced JAK/STAT activation in macrophages, resulting in enhanced polarization along the alternative activation pathway. Importantly, we observed that Mac-1−/− macrophages exhibit increased expression of foam cell differentiation markers including 15-lipoxygenase and lectin-type oxidized LDL receptor-1 both in vitro and in vivo. Indeed, we found that Mac-1−/−LDLR−/− mice develop significantly more foam cells than control LDLR−/− mice, using an in vivo model of foam cell formation. Together, our data establish for the first time a molecular mechanism by which Mac-1 regulates the signaling activity of IL-13 in macrophages. This newly identified IL-13Rα1/Mac-1-dependent pathway may offer novel targets for therapeutic intervention in the future.


Mac-1 exhibits a unique inhibitory activity toward IL-13-induced JAK/STAT activation and thereby regulates macrophage to foam cell transformation. However, the underlying molecular mechanism is unknown. In this study, we report the identification of IL-13R␣1, a component of the IL-13 receptor (IL-13R), as a novel ligand of integrin Mac-1, using a co-evolution-based algorithm. Biochemical analyses demonstrated that recombinant IL-13R␣1 binds Mac-1 in a purified system and supports
Mac-1-mediated cell adhesion. Co-immunoprecipitation experiments revealed that endogenous Mac-1 forms a complex with IL-13R␣1 in solution, and confocal fluorescence microscopy demonstrated that these two receptors co-localize with each other on the surface of macrophages. Moreover, we found that genetic inactivation of Mac-1 promotes IL-13-induced JAK/STAT activation in macrophages, resulting in enhanced polarization along the alternative activation pathway. Importantly, we observed that Mac-1 ؊/؊ macrophages exhibit increased expression of foam cell differentiation markers including 15-lipoxygenase and lectin-type oxidized LDL receptor-1 both in vitro and in vivo. Indeed, we found that Mac-1 ؊/؊ LDLR ؊/؊ mice develop significantly more foam cells than control LDLR ؊/؊ mice, using an in vivo model of foam cell formation. Together, our data establish for the first time a molecular mechanism by which Mac-1 regulates the signaling activity of IL-13 in macrophages. This newly identified IL-13R␣1/Mac-1-dependent pathway may offer novel targets for therapeutic intervention in the future.
High levels of LDL and its associated cholesterol in the circulation are well established risk factors for cardiovascular diseases (1)(2)(3). LDL accumulates within the subendothelial space in response to inflammation, where it is converted into oxidized LDL (oxLDL) 2 by reactive oxygen species, myeloperoxidase, and lipoxygenase such as 15-lipoxygenase , all of which are secreted by macrophages (3). Subsequently, oxLDL is internalized by these macrophages via specific cell surface receptors, including CD36 (4) and lectin-type oxidized LDL receptor-1 (LOX-1) (5), and transforms macrophages into foam cells. Accumulation of fat-laden foam cells in the vessel wall gives rise to the earliest vascular lesions, which gradually progress into fatty streaks, intermediate lesions, and ultimately atherosclerotic plaques (3,6). The critical role of foam cell formation in the pathogenesis of atherosclerosis is underscored by the observations that genetic inactivation of 15-LO or LOX-1 in mice protects against the development of severe plaques under hyperlipidemia (5,7).
Macrophages can polarize along two distinct pathways: stimulation with IFN-␥ and lipopolysaccharide generates the classically activated macrophages (also known as M1) that exhibit proinflammatory activities, whereas stimulation with IL-4 or IL-13 generates the alternatively activated macrophages (AAMs or M2) that possess anti-inflammatory properties (8 -10). Surprisingly, emerging evidence has implicated a pathological role of AAMs in the early stage of atherosclerosis development (9,(11)(12)(13). Indeed, AAMs are found to produce high levels of 15-LO and CD36 under the pathological setting of hyperlipidemia (2,9,10) and are thus more susceptible to foam cell transformation than classically activated macrophages (11,12). Recently, Yakubenko et al. (9,14) reported * This work was supported by National Institute of Health Grants NHLBI that integrin Mac-1 (CD11b/CD18 or ␣ M ␤ 2 ) functions to suppress IL-13-induced JAK/STAT activation in macrophages and reduces their uptake of oxLDL in vitro, suggesting that Mac-1 plays a regulatory role in foam cell development. However, the molecular mechanism by which Mac-1 inhibits IL-13-induced JAK/STAT activation in macrophages is unknown.
In this work, we identified IL-13R␣1, a subunit of the heterodimeric IL-13R, as a novel biological partner of Mac-1, using an unbiased co-evolution-based algorithm. Subsequently, we confirmed direct binding between these two receptors using purified full-length Mac-1 and recombinant IL-13R␣1. We also showed by co-immunoprecipitation that endogenous Mac-1 and IL-13R␣1 form a complex in solution and showed by confocal fluorescence microscopy that they reside in proximity on the cell surface. Biologically, we found that genetic inactivation of Mac-1 enhances IL-13-induced JAK/STAT activation in macrophages and promotes their polarization. Furthermore, we found that the absence of Mac-1 on macrophages increases their expression of foam cell differentiation markers. Using an LDLR Ϫ/Ϫ mouse model of foam cell formation, our data revealed that Mac-1 Ϫ/Ϫ mice develop significantly more foam cells in vivo. Together, our study established for the first time a molecular mechanism by which Mac-1 suppresses IL-13-induced macrophage activation and foam cell transformation. Thus, this newly identified IL-13R␣1/Mac-1 interaction could represent a potential target for therapeutical intervention that will not interfere with IL-13R-medaited immunity or Mac-1dependent host defense functions.

Experimental Procedures
Mice-WT, Mac-1 Ϫ/Ϫ , LDLR Ϫ/Ϫ , and Mac-1 Ϫ/Ϫ LDLR Ϫ/Ϫ mice were all in the C57BL/6J background and used at 8 -13 weeks of age (20 -22 g). WT and LDLR Ϫ/Ϫ mice were purchased from The Jackson Laboratory. Mac-1 Ϫ/Ϫ mice were kindly provided by Dr. Christie M. Ballantyne, Baylor College of Medicine (Houston, TX) and have been backcrossed to the C57BL/6J background for more than 10 generations. All mice were housed in a pathogen-free facility, and all procedures were performed in accordance with University of Maryland Institutional Animal Care and Use Committee approval.
Prediction of Protein-Protein Interactions-All protein sequences for the complete genomes of 184 species were downloaded from the National Center for Biotechnology Information (NCBI) Entrez database (16). The set of species contained 130 eukaryotes, 47 prokaryotes, and seven archaea. All sequences for the 184 species were updated on June 2011. Orthologous proteins were identified for each protein in each of the genomes by performing reciprocal BLAST searches (17) against all genomes. For each species, the protein with the smallest average E-value (average of the forward and reciprocal E-values), with E-value Ͻ1 ϫ 10 Ϫ40 and the largest sequence overlap with at least 80% alignment overlap, was selected as the ortholog of protein P for that species.
Calculating Distance Matrices-Multiple sequence alignments for each protein within the pairs of interacting and noninteracting proteins and their corresponding orthologs were obtained using MUSCLE version 3.52 (18). From these multiple sequence alignments, the distance matrices for each protein set were estimated using the ClustalW suite with default parameters (19).
Estimating Correlation Coefficients-For each protein pair from the interacting and non-interacting sets, the correspondence between their evolutionary histories was measured by computing the correlation between the distance matrices of each protein (the Mirrortree approach) (20). For two proteins A and B with n species in common (where n Ն9), let us denote A i,j as the distance between species i and j for protein A; and B i,j for protein B. The linear correlation coefficient (corr. coef. or CC AB ) between their distance vectors was calculated using the standard Pearson's corr. coef.
where Ā and B are the means of A i,j and B i,j respectively. Correcting for Speciation-As shown by Pazos et al. (21) and Sato et al. (22), prediction of protein interactions using co-evolutionary analysis is greatly improved by excluding the background co-evolutionary signal from phylogenetic relationships (signal from speciation). The signal from speciation for species i and j, represented by matrix S i,j , was computed by averaging the evolutionary distance matrices of all proteins in species i with their orthologs in species j. To estimate the corr. coef. with this correction, the distance matrices A i,j and B i,j in Eq. 1 were replaced by AЈ i,j and BЈ i, Preparation of Peritoneal Macrophages-Mice were injected intraperitoneally with 4% thioglycollate. Four days later, peritoneal macrophages were collected by lavages with PBS. After lysis of red blood cells with ammonium chloride (8.3 g/liter in 10 mM Tris-HCl, pH 7.4), peritoneal cells were plated on 10-cm tissue culture Petri dishes in 10% FBS in DMEM and incubated in a humidified incubator with 5% CO 2 at 37°C for 2-4 h. Adherent macrophages were collected, and their purity was verified by flow cytometry based on positive staining for F4/80 and M1/70 and by morphological examination of Hema 3 (Fisher Scientific)-stained smears prepared with a Cytospin (Shandon).
Alternative Activation of Macrophages and qRT-PCR-Peritoneal macrophages were treated with or without 4 nM IL-13 for various times, ranging from 5 min to 3 days. STAT6 activation was assessed by Western blot using phospho-STAT6-specific antibody. Total STAT6 protein and ␤-actin were used as loading controls. Macrophage polarization along the alternative activation pathway was assessed by quantitative real-time RT-PCR (qRT-PCR) based on expression of specific markers, including Arginase-1, FIZZ1, YM1, and YM2. Briefly, total RNA was extracted from peritoneal macrophages using the Absolutely RNA miniprep kit (Stratagene) according to the manufacturer's instructions. One microgram of total RNA was used for cDNA synthesis using SuperScript II and random hexamers (Invitrogen). qRT-PCR was performed on an ABI Prism 7500 HT sequence detection system (Applied Biosystems), using specific primer pairs ( Table 1). The PCR reaction was done in a 25-l solution containing 12.5 l of SYBR Green PCR master mix, 10 ng of cDNA, and 400 nM of each primer, with the following settings: activation of the AmpliTaq Gold polymerase at 95°C for 10 min and 40 cycles of 95°C for 15 s and 60°C for 1 min. The melting curve was analyzed using the ABI software, and quantification of gene expression was done based on the 2 Ϫ⌬⌬Ct method (RQ Manager 1.4). Each experiment was run in duplicates. All data were normalized to ␤-actin expression in the same cDNA set.
Ligand Binding-Binding of recombinant IL-13R␣1 to purified full-length Mac-1 was carried out in 96-well microtiter plates. Detergent-solubilized Mac-1 was purified from Mac-1expressing HEK293 cells (23) by affinity chromatography using mAb LM2/1-coupled agarose. The purity of Mac-1 was verified by SDS-PAGE and by its ability to bind conformation-dependent mAbs (data not shown). Fifty microliters of 10 g/ml purified Mac-1 in 100 mM KCl, 3 mM MgCl 2 , 10 mM PIPES, pH 7.0, were used to coat the wells of 96-well microplates at 4°C overnight. The plate was washed and blocked with 1% BSA. Recombinant mouse IL-13R␣1, which contains a histidine tag at its C terminus (Leinco Technologies Inc.), was added at different concentrations (0 -2 M) in the above buffer plus 0.1% BSA. After incubation at 37°C for 2 h, the plate was washed, and bound IL-13R␣1 was detected using HisProbe-HRP (Thermo Scientific) with its substrate 3,3Ј,5,5Ј-tetramethylbenzidine, measuring absorbance at 450 nm. The binding curve was fitted to the Michaelis-Menten equation based on a single binding site model using SigmaPlot (Systat Software, Inc., San Jose, CA).
Cell Adhesion Assays-Cell adhesion was conducted based on our published methods (23). Briefly, 24-well polystyrene microtiter plates were coated with increasing concentrations of recombinant mouse IL-13R␣1, BSA, or 10 g/ml ICAM-1-Fc (24). The plates were blocked with 400 l of 0.05% polyvinylpyrrolidone in Dulbecco's PBS. A total of 2 ϫ 10 6 Mac-1expressing HEK293 cells in Hanks' balanced salt solution containing 1 mM Ca 2ϩ and 1 mM Mg 2ϩ in the presence or absence of 50 nM neutrophil inhibitor factor (NIF; a Mac-1-specific antagonist (25)) or LFA-1-expressing HEK293 cells in Hanks' balanced salt solution containing 2 mM EGTA and 2 mM Mg 2ϩ with or without 10 g/ml mAb 38 (a LFA-1 function-blocking mAb) were added to each well. After incubation at 37°C for 30 min, non-adherent cells were removed by three washes with Dulbecco's PBS and adherent cells were stained with Crystal Violet. Cell adhesion was quantified based on absorption at 595 nm.
Co-immunoprecipitation-Co-immunoprecipitation experiments were conducted based on our published methods (26). WT peritoneal macrophages were stimulated with IL-13, washed, and then lysed. The cell lysates were immunoprecipitated using a rabbit anti-IL-13R␣1 antibody, and the immunoprecipitates were separated on 10% SDS-PAGE, transferred to PVDF membrane, and probed with anti-CD18 antibody ARC22 (15). Equal loading was verified by the presence of equal amounts of CD18 and ␤-actin in the cell lysates.
Confocal Laser Scanning Fluorescence Microscopy-Co-localization between IL-13R␣1 and Mac-1 was determined by confocal fluorescence microscopy based on our published method (26). Peritoneal macrophages were seeded on multicoverslips overnight. The cells were washed with PBS and fixed with 4% paraformaldehyde in PBS for 30 min at room temperature. After blocking with 5% bovine serum albumin in PBS at room temperature for 30 min, these cells were incubated with 20 g/ml rabbit anti-IL-13R␣1 antibody and 20 g/ml rat anti-Mac-1 mAb (M1/70) in 1% bovine serum albumin in PBS at room temperature for 60 min. After washing, these different coverslips were incubated with Alexa Fluor 488-conjugated goat anti-rabbit IgG and Alexa Fluor 568-conjugated goat antirat IgG (Molecular Probes). Non-immune rabbit and rat IgGs were used as specificity controls. The stained macrophages were analyzed using a Bio-Rad Radiance 2000 confocal laser scanning fluorescence microscope system equipped with a Nikon Eclipse E800 upright light microscope. The images

Mac-1 Binds IL-13R␣1 and Suppresses IL-13 Signaling
were collected using 100ϫ oil objectives with a slice thickness of 2.6 m. (27) with minor modifications. Briefly, Mac-1 Ϫ/Ϫ LDLR Ϫ/Ϫ mice and their LDLR Ϫ/Ϫ littermates were fed a high fat "Western" diet (21% fat and 0.2% cholesterol) for 5 weeks. Peritoneal macrophages were prepared from these mice by intraperitoneal injection of 4% thioglycollate, followed by peritoneal lavage 4 days later. Foam cells within the collected peritoneal cells were identified by staining the fat droplets within the cytosol with Oil Red O following centrifuging in the Cytospin. To quantify the amount of cholesterol accumulated within these foam cells, peritoneal cells were pelleted at 1000 rpm for 7 min, extracted in 500 l of isopropanol by sonication (2 ϫ 10 s), and then centrifuged at 14,000 rpm for 10 min. The pellet was used to determine the amount of total protein by Bio-Rad protein assay. The supernatant was assayed for cholesterol using Amplex Red kit (Invitrogen) according to the manufacturer's protocol. Free cholesterol was determined by omitting the cholesterol esterase. Total cholesterol was determined in the presence of the cholesterol esterase. Cholesterol esters were calculated from the difference in total and free cholesterol. Data are plotted as mg of cholesterol per mg of protein.

Foam Cell Formation in Vivo-The formation of foam cells in vivo was investigated based on published methods
Statistical Analysis-Statistical analyses were performed using two-tailed Student's t test. p values less than 0.05 were considered significant.

Co-evolution between Mac-1 and IL-13R␣1-Recently,
Yakubenko et al. (9,14) reported that Mac-1 suppresses IL-13induced JAK/STAT activation in macrophages and reduces their uptake of oxLDL in vitro; however, the molecular mechanism underlying this unique regulatory function of Mac-1 is not yet established. To identify potential biological partners that confer the inhibitory activity of Mac-1 toward IL-13-induced macrophage activation, we used an unbiased informatics-based method, termed Mirrortree (28 -30), to predict protein-protein interactions. The premise of Mirrortree is that biological partners, such as parasite/host, must change in sync during evolution to maintain their interdependence (28). Accordingly, we downloaded the entire protein sequences encoded in 184 different species from the NCBI Entrez database (16) and identified the orthologous protein representatives for each species. Multiple sequence alignments for each protein within the pairs of interacting and non-interacting proteins and their corresponding orthologs were then calculated using MUSCLE version 3.52 (18). From these multiple sequence alignments, the distance matrices for each protein set were estimated using the ClustalW suite with default parameters (19). Linear correlation coefficients for each protein pair from the interacting and noninteracting sets was calculated using the standard Pearson's correlation coefficient. Finally, background co-evolutionary signals, e.g. due to speciation, were excluded from phylogenetic relationships, based on published methods (21,22).
One of the proteins that co-evolve with Mac-1 is IL-13R␣1 (corr. coef. ϭ 0.76 ϫ 10 Ϫ40 ). Fig. 1 shows the concerted changes in protein sequence of Mac-1 and IL-13R␣1 during a million years of evolution, suggesting that these two proteins may interact with each other to maintain important biological functions. IL-13R␣1 associates with IL-4R␣ and forms the heterodimeric receptor IL-13R␣1/IL-4R␣ (IL-13R) that is responsible for IL-13-induced intracellular signaling. Accordingly, we hypothesized that Mac-1 suppresses IL-13-induced macrophage activation and its subsequent polarization by binding to IL-13R␣1 on the macrophage surface.
Direct Interaction between purified Mac-1 and Recombinant IL-13R␣1-To test our hypothesis, we first examined whether Mac-1 interacts directly with IL-13R␣1 in a purified system. We isolated full-length Mac-1 from Mac-1-expressing HEK293 cells by affinity chromatography. Purified Mac-1 was then incubated with recombinant soluble IL-13R␣1 in a 96-well microtiter plate, with or without the addition of a neutralizing anti-IL-13R␣1 antibody. We found that IL-13R␣1 bound immobilized Mac-1 in a dose-dependent manner with a K d of ϳ5 M ( Fig.  2A), which could be inhibited by an anti-IL-13R␣1 antibody, thus verifying the specificity of the binding assay. Next, we evaluated whether IL-13R␣1 can support Mac-1-mediated cell adhesion. We coated 24-well microtiter plates with increasing concentrations of IL-13R␣1 or a control protein BSA. After blocking the plates, Mac-1-or LFA-1-expressing HEK293 cells were added in the presence or absence of a Mac-1-specific antagonist NIF (25) or LFA-1-blocking mAb 38 and incubated at 37°C for 30 min. The plates were washed to remove nonadherent cells and the adherent cells were quantified. Fig. 2B shows that IL-13R␣1 supported adhesion of Mac-1-expressing HEK293 cells in a dose-dependent manner with an EC 50 of ϳ0.5 M. In comparison, little adhesion of Mac-1/HEK293 cells was observed to BSA-coated wells. Moreover, the addition of a Mac-1-specific antagonist NIF blocked more than 65% Mac-1/ HEK293 cell adhesion to IL-13R␣1, thus verifying specificity. Furthermore, we found that LFA-1/HEK293 cells adhered strongly to ICAM-1, which could be inhibited by mAb 38; however, they failed to adhere to IL-13R␣1 (Fig. 2C). As expected, Mac-1/HEK293 cells adhered to both ICAM-1 and IL-13R␣1, both of which could be blocked by NIF. These results demonstrate that Mac-1 interacts specifically with IL-13R␣1, thus  AUGUST 28, 2015 • VOLUME 290 • NUMBER 35 supporting our hypothesis that Mac-1 modulates the function of IL-13 by interacting with its receptor IL-13R.

Mac-1 Binds IL-13R␣1 and Suppresses IL-13 Signaling
Endogenous Mac-1 and IL-13R␣1 Form a Complex in Solution-To test whether Mac-1 and IL-13R␣1 interact with each other when expressed on macrophages, we conducted coimmunoprecipitation experiments using primary macrophages harvested from thioglycollate-treated mice. We collected peritoneal macrophages and prepared cell lysates. IL-13R␣1 was then immunoprecipitated with its specific antibody. After separation of the immunoprecipitates on SDS-PAGE, the presence of Mac-1 in the immunoprecipitates was determined by Western blot using an anti-CD18 antibody ARC22 (15). We found that Mac-1 was pulled down by an anti-IL-13R␣1 antibody but not by a control IgG (Fig. 2D), demonstrating that Mac-1 and IL-13R␣1 formed a complex in solution.
Mac-1 Is Associated with IL-13R␣1 on the Surface of Macrophages-To determine whether Mac-1 and IL-13R␣1 associate with one another on the cell surface, peritoneal macrophages were stained with a rabbit anti-IL-13R␣1 antibody and a rat anti-Mac-1 antibody and then incubated with their corresponding secondary antibody conjugated with Alexa Fluor 488 or Alexa Fluor 568, respectively. The stained macrophages were analyzed by confocal fluorescence microscopy. We found that IL-13R␣1 co-localized with Mac-1 (Fig. 2E, yellow color in the merged image), suggesting that these two receptors reside proximally on the cell surface.
Genetic Inactivation of Mac-1 Enhances IL-13-induced JAK/ STAT Activation-IL-13R␣1 is a subunit of IL-13R. Stimulation of this heterodimeric IL-13R receptor by IL-13 leads to activation of the JAK/STAT pathway (8,31). To investigate whether Mac-1 association with IL-13R␣1 plays a functional role in IL-13R signaling, we stimulated WT and Mac-1 Ϫ/Ϫ macrophages with IL-13. Activation of the IL-13R signaling pathway was measured by Western blot using an antibody that specifically recognizes the phosphorylated form of STAT6 (at Tyr-641). The results showed that treatment of macrophages with IL-13 resulted in a time-dependent increase in phosphorylated STAT6, plateauing at 30 min (Fig. 3A). Genetic inactivation of Mac-1 significantly enhanced STAT6 phosphorylation as compared with their WT counterparts (Fig. 3A). These data demonstrated that Mac-1 functions to suppress IL-13R signaling in IL-13-stimulated macrophages.
Deficiency of Mac-1 Promotes Macrophage Polarization along the Alternative Activation Pathway-IL-13 signaling promotes macrophage polarization along the alternative pathway, generating AAMs (8). To determine whether the increased JAK/STAT activation in Mac-1 Ϫ/Ϫ macrophages leads to enhanced alternative activation, we quantified the levels of several AAM markers, including Arginase-1, FIZZ1, YM-1, and YM2 (8), by qRT-PCR. The results showed that Mac-1 Ϫ/Ϫ macrophages had higher transcription levels of these four AAM markers, both in the absence and in the presence of IL-13 stimulation, as compared with WT cells (Fig. 3B). We also observed that Mac-1 Ϫ/Ϫ macrophages expressed significantly higher levels of IL-6, a proinflammatory cytokine (Fig. 3B).

Mac-1 Suppresses Macrophage Expression of Foam Cell Differentiation Markers in Vitro and in Vivo-To investigate
whether Mac-1 deficiency predisposes macrophages to foam (contains a C-terminal His tag) were added in 96-well microtiter plates, which were precoated with purified Mac-1, in the absence (F) or presence (E) of an IL-13R␣1-neutralizing antibody. Bound IL-13R␣1 was detected using a His-Probe-HRP conjugate and its substrate 3,3Ј,5,5Ј-tetramethylbenzidine, measuring absorbance at 450 nm. Data shown are means Ϯ S.D. of duplicate experiments. B, cell adhesion. Mac-1-expressing HEK293 cells were added to 24-well microtiter plates, which were precoated with increasing concentrations of IL-13R␣1 (F) or BSA (OE), in the absence or presence of a Mac-1-specific antagonist NIF (E). After incubation at 37°C for 30 min, non-adherent cells were removed and adherent cells were stained with Crystal violet. Cell adhesion was quantified based on absorption at 595 nm. Data shown are means Ϯ S.D. of duplicate experiments. C, Mac-1-or LFA-1-expressing HEK293 cells were added to 24-well microtiter plates, which were precoated with IL-13R␣1, ICAM-1, or BSA, in the absence or presence of NIF or mAb 38. Cell adhesion was carried out as above. Mock-transfected HEK293 cells were used as a control. Data shown are means Ϯ S.D. of duplicate experiments. D, co-immunoprecipitation. Total lysates of peritoneal macrophages were incubated with an anti-IL-13R␣1 antibody or a control IgG and then immunoprecipitated (IP) with protein A-agarose. Immunoprecipitates were separated on 10% SDS-PAGE and subjected to immunoblot (IB) with an anti-CD18 antibody ARC22. Total cell lysates were probed for CD18 and ␤-actin to verify equal protein loading. E, co-localization. Peritoneal macrophages were stained with a rabbit anti-IL-13R␣1 antibody and a rat anti-Mac-1 mAb (M1/70), followed by Alexa Fluor 488-anti-rabbit IgG and Alexa Fluor 568-anti-rat IgG. Specificity was verified using non-immune rabbit and rat IgGs (data not shown). Representative images shown were taken with 100ϫ objective oil lens with a slice thickness of 2.6 m. Co-localization between IL-13R␣1 (in green) and Mac-1 (in red) was shown in yellow color in the merged image.
cell differentiation, we evaluated the expression of several differentiation markers for foam cells. The results showed that Mac-1 Ϫ/Ϫ macrophages exhibited increased gene transcription of 15-LO and LOX-1, both of which are critical to foam cell formation, with or without IL-13 stimulation (Fig. 4A). Western blot analysis showed that the protein levels of 15-LO and LOX-1 in macrophages increased as well upon IL-13 stimulation, but they exhibited distinct patterns: the level of 15-LO was up-regulated early but peaked on Day 1, whereas the LOX-1 level continued to rise on Day 3 (Fig. 4B). Importantly, we found that Mac-1 Ϫ/Ϫ macrophages exhibited higher protein levels for both 15-LO and LOX-1 antigen on Day 1 and Day 3 (Fig. 4B), which agreed well with their increased gene transcription. No significant effect of Mac-1 deficiency was observed on the expression of CD36, SR-A1, or SR-BI (data not shown).
To test whether Mac-1 deficiency results in an enhanced foam cell development in vivo, we employed an LDLR Ϫ/Ϫ mouse model of foam cell formation (27). We crossed Mac-1 Ϫ/Ϫ mice with LDLR Ϫ/Ϫ mice and then fed Mac-1 Ϫ/Ϫ LDLR Ϫ/Ϫ mice and their littermate LDLR Ϫ/Ϫ controls a high fat diet for 5 weeks. After intraperitoneal injection of thioglycollate, peritoneal macrophages were harvested on Day 4 by lavage. Expression of 15-LO and LOX-1 in the peritoneal macrophages was determined by qRT-PCR. We found that macrophages derived from Mac-1 Ϫ/Ϫ LDLR Ϫ/Ϫ mice exhibited significantly higher transcription levels of 15-LO, LOX-1, and IL-6 than those of their WT counterparts (Fig. 4C).
Mac-1-deficient Mice Exhibit Increased Foam Cell Formation in Vivo-To directly investigate whether Mac-1-deficient mice develop more foam cells in vivo, we subjected the peritoneal cells obtained from the above high fat diet-fed LDLR Ϫ/Ϫ and Mac-1 Ϫ/Ϫ LDLR Ϫ/Ϫ mice to Oil Red O staining. The results showed that peritoneal macrophages from Mac-1 Ϫ/Ϫ LDLR Ϫ/Ϫ mice contained more fat droplets than those from LDLR Ϫ/Ϫ littermates ( Fig. 5A; red color), indicating that Mac-1 deficiency enhanced foam cell transformation. To quantify the amount of foam cells generated in vivo, we determined the total cholesterol content of the peritoneal macrophages harvested from the above mice. The results showed that the cells from Mac-1 Ϫ/Ϫ LDLR Ϫ/Ϫ mice contained significantly higher levels of total cholesterol and cholesterol ester than those from LDLR Ϫ/Ϫ littermates (Fig. 5B). No difference in free cholesterol contents was observed between the macrophages from Mac-1 Ϫ/Ϫ LDLR Ϫ/Ϫ mice and their littermate controls (Fig. 5B).

Discussion
A major step in the development of vascular diseases, such as atherosclerosis, is the generation of foam cells within the subendothelial space, which marks the formation of the earliest vascular lesions (1,6). Recent studies demonstrated that Mac-1, a major leukocyte integrin expressed on macrophages and macrophage-derived foam cells, exhibits potent inhibitory activities toward JAK/STAT activation in IL-13-stimulated macrophages, which reduces their ability to uptake oxLDL in vitro (9,14). However, the molecular mechanism by which Mac-1 inhibits IL-13 signaling in macrophages is still unknown. Here, we report for the first time the identification of IL-13R␣1 as a biological partner based on a combination of bioinformatics and biochemical approaches. Our results demonstrate that Mac-1 directly interacts with IL-13R␣1 and thereby suppresses IL-13-induced JAK/STAT activation. Hence, this study establishes a novel mechanism by which Mac-1 regulates IL-13 signaling in macrophages, thus suppressing their transformation into foam cells.
Using a Mirrortree-based algorithm, we first observed that Mac-1 co-evolves with IL-13R␣1 during evolution (Fig. 1), suggesting that these two receptors may dependent on each other to function. Next, we confirmed that Mac-1 binds directly to IL-13R␣1 in a purified system using recombinant proteins ( Fig.  2A). The ability of endogenous Mac-1 and IL-13R␣1 expressed on primary macrophages to interact with each other in solution was verified by co-immunoprecipitation experiments (Fig. 2D), and their association on the cell surface was confirmed by confocal fluorescence microscopy (Fig. 2E). Biochemical assays demonstrated that the formation of the Mac-1/IL-13R complex on macrophages dampens IL-13-induced JAK/STAT activation (Fig. 3A) and thus suppresses macrophage polarization along the alternative activation pathway (Fig. 3B). Consequently, our data showed that Mac-1 deficiency predisposes macrophages to foam cell differentiation both in vitro and in vivo (Fig. 4). Most importantly, we demonstrated that genetic inactivation of Mac-1 promotes the generation of foam cells in vivo using an LDLR Ϫ/Ϫ mouse model of foam cell development (Fig. 5). Together, this study establishes a major role of the Mac-1/IL-13R␣1 interaction in safeguarding macrophages from aberrant foam cell development.
Mac-1 is a leukocyte integrin of the CD18 subfamily and expressed abundantly on macrophages and macrophage-de-rived foam cells. A unique feature of Mac-1 is its ability to recognize a wide range of diverse protein and non-protein ligands. These unique interactions may underlie the ability of Mac-1 to facilitate both pro-inflammatory and anti-inflammatory responses in a context-dependent manner (32)(33)(34)(35)(36)(37)(38)(39)(40). For example, Mac-1 binding of C3bi has been shown to suppress inflammation and induce immune tolerance (41,42); on the other hand, Mac-1-mediated adhesion to fibrin exacerbates inflammatory diseases (36,39). Given the suppressive activity of Mac-1 toward IL-13 signaling in macrophages (9,14), we sought in this work to identify candidate partners of Mac-1 that mediate this unique regulatory activities. To achieve our objective, we developed a computational method to predict protein- protein interactions, based in part on Mirrortree, using the entire genetic and protein sequence database. This method relies on the conservation of gene order (43,44), domain/gene fusion (45), and co-evolution (46 -48). The assumption behind co-evolution-based Mirrortree methods is that for a given interacting protein pair, sequence changes in one protein are mirrored in the other protein to maintain their interaction and thus biological function. Using this unbiased method, we found that IL-13R␣1, but not IL-4R␣, co-evolves with Mac-1, suggesting that Mac-1 may regulate IL-13 signaling activity in macrophages by interacting with the IL-13R␣1 subunit of the IL-13R. In support of this prediction, we showed that Mac-1 binds IL-13R␣1 directly with a relative low affinity (K d of ϳ5 M; Fig.  2A), which could be further enhanced by multivalent receptorreceptor interactions on the cell surface. Indeed, we found that immobilized IL-13R␣1 supported Mac-1-mediated cell adhesion with a relatively higher affinity (EC 50 of 0.5 M; Fig. 2B). Together, this study identifies IL-13R␣1 as a new ligand of Mac-1 that contributes to its ability to suppress foam cell development, especially in the setting of hyperlipidemia. Finally, as genetic ablation of the Mac-1/IL-13R␣1 interaction leads to enhanced IL-13 signaling (Fig. 3), our data support a working model that Mac-1 association with IL-13R␣1 interferes with the function of the IL-4R␣/IL-13R␣1 heterodimer by sterically hindering IL-13 binding. Experiments to test this hypothesis are currently underway.
Mac-1 and IL-13R␣1 are both composed of multiple independent domains (49,50). In particular, the ␣ M subunit of Mac-1 contains an I-domain, a ␤-propeller, a Thigh domain, two Calf domains, a transmembrane domain, and a cytoplas-mic tail (49). Major ligand binding activities have been localized to the I-domain, although other domains may also contribute to the binding of specific ligands. The IL-13R␣1 subunit contains three domains: an N-terminal Ig-like domain (D1) and two Fibronectin-III homologous domains (D2 and D3). All three domains participate directly in IL-13 binding (50). To determine putative domains of Mac-1 and IL-13R␣1 that may interact with each other, we conducted bioinformatics analysis on individual domain pairs that co-evolve. We delineated individual domains within Mac-1 and IL-13R␣1 by introducing "breaks" at the domain boundary during multiple sequence alignments and then performed Mirrortree analysis as discussed above. Several domain pairs were found to coevolve significantly. A major candidate interacting pair involves the fifth blade of the ␣ M ␤-propeller (residues 453-511) with one or more domains of IL-13R␣1, including the D3 domain (resides 233-346; corr. coef. ϭ 0.79 ϫ 10 Ϫ40 ), the D2 domain (residues 129 -219; corr. coef. ϭ 0.78 ϫ 10 Ϫ40 ), and/or the D1 domain (residues 27-117; corr. coef. ϭ 0.75 ϫ 10 Ϫ40 ). Other potential interactions include the ␣ M I-domain (residues 149 -324) with the IL-13R␣1 D2 domain (corr. coef. ϭ 0.67 ϫ 10 Ϫ40 ) and the ␣ M leg region (residues 615-1033) with the IL-13R␣1 D3 domain (corr. coef. ϭ 0.55 ϫ 10 Ϫ40 ). The validity of these predicted interactions needs to be assessed experimentally in the future.
In summary, we have identified IL-13R␣1 as a novel Mac-1 partner based on an unbiased co-evolution-based informatics approach, which was subsequently confirmed by biochemical assays. We demonstrated that Mac-1 interacts with IL-13R␣1 and associates with IL-13R on the surface of macrophages. Genetic inactivation of Mac-1 increases IL-13-induced JAK/ STAT activation and predisposes macrophages to foam cell differentiation. Altogether, our study identified a unique mechanism by which Mac-1 protects macrophages against foam cell transformation in disease settings such as hyperlipidemia. This newly identified IL-13R␣1/Mac-1-dependent pathway may offer novel targets for therapeutic intervention in the future.
Author Contributions-C. C. designed, performed, and analyzed the experiments. J. Z. performed animal studies. E. K. D. and M. G. K. conducted bioinformatics analysis. M. M. and D. K. S. measured cholesterol contents of peritoneal macrophages. L. Z. conceived and coordinated the study and wrote the paper. All authors reviewed the results and approved the final version of the manuscript.