The Systemic Lupus Erythematosus–Associated Single Nucleotide Polymorphism rs1143678 in Integrin α_M Cytoplasmic Tail Generates a 14-3-3ζ Binding Site That Is Proinflammatory

Integrins are transmembrane heterodimers that are composed of α and β subunits, and they mediate cell–cell and cell–extracellular matrix interactions (1). The leukocyte β₂ (CD18) integrin family consists of α_Lβ₂ (CD11aCD18, LFA-1), α_Mβ₂ (CD11bCD18, Mac-1, CR3), α_Xβ₂ (CD11cCD18, p150,95), and α_Dβ₂ (CD11dCD18) (2). The importance of the β₂ integrins is underscored by the disease leukocyte adhesion deficiency type I in which ITGB2 (β₂ subunit) gene mutations lead to defective leukocyte adhesion and migration (3). Patients with leukocyte adhesion deficiency type I are therefore prone to recurrent microbial infections that often lead to life-threatening complications.

Integrin α_Mβ₂ is expressed in myeloid, NK, and γδ T cells (4, 5). Its ligands include iC3b, fibrinogen, ICAM-1, receptor for advanced glycation end products, myelin basic protein, LPS, and denatured proteins (2). In addition to leukocyte migration and phagocytosis, integrin α_Mβ₂ regulates neutrophil survival, monocyte/macrophage differentiation, and immune tolerance (6–11). The integrin α_L, α_M, α_X, and α_D cytoplasmic tails (CTs) are divergent except for a conserved GFFKR motif (2). Integrin α_M CT contains Tyr¹¹³⁷ and Ser¹¹⁴² that are potential phosphorylation sites. Although there is limited information on the function of Tyr¹¹³⁷, nonconserved mutation of Ser¹¹⁴² has been shown to reduce the binding of integrin α_Mβ₂ to ICAM-1 (12). Nuclear magnetic resonance study has shown that integrin α_M CT adopts a hairpin conformation that contains a membrane-proximal helix and a C-terminal loop, and Ser¹¹⁴² phosphorylation did not induce any significant change to the conformation (13). Hence, phospho-Ser¹¹⁴² most likely provides a docking site for yet to be identified cytosolic molecules that regulate integrin α_Mβ₂ function.

Systemic lupus erythematosus (SLE) is a chronic inflammatory disease with a complex etiology that involves genetic predispositions and environmental factors. Genome-wide association studies have identified several susceptibility variants in the ITGAM gene (α_M subunit) (14–16). The nonsynonymous single nucleotide polymorphism (SNP) variant rs1143679 substitutes Arg⁷⁷ for His in the integrin α_M β-propeller, leading to defective ligand-binding of integrin α_Mβ₂ (15–20). It was proposed that macrophages expressing integrin α_M(R77H) contribute to the onset of SLE because they are defective in clearing apoptotic bodies and they overproduce proinflammatory IL-6 (21). SNP rs1143679 together with SNPs rs1143678 and rs1143683 have been identified in some SLE patients: rs1143678 leads to P1146S substitution and rs1143683 leads to A858V substitution in the integrin α_M CT and calf-1 regions, respectively (17). Neutrophils from SLE patients with rs1143678 and rs1143683 have been shown to exhibit attenuated integrin α_Mβ₂ ligand-binding properties (22). To date, a detailed characterization of rs1143678, α_M(P1146S), is lacking. In this study, we show that P1146S substitution in integrin α_M CT promotes myeloid cell adhesion, spreading, migration, phagocytosis, and secretion of proinflammatory cytokines IL-6 and TNF-α. Mechanistically, P1146S substitution in the integrin α_M CT generates a noncanonical 14-3-3ζ binding site. 14-3-3ζ is a cytoplasmic molecule that binds to and positively regulates β₂ integrins (23, 24). Our data suggest that the aberrant association between 14-3-3ζ and integrin α_M(P1146S) CT leads to the proinflammatory phenotype of myeloid cells expressing integrin α_M(P1146S)β₂.

K562 and 293T cell lines were obtained from the American Type Culture Collection (Manassas, VA) and cultured in RPMI 1640 and DMEM media, respectively, supplemented with 10% (v/v) heat-inactivated FBS, 100 IU/ml penicillin, and 100 μg/ml streptomycin. Cell culture medium, FBS, and antibiotics were from HyClone (GE Healthcare Life Sciences, Logan, UT). The integrin α_l-deficient Jurkat cell line JB2.7 (provided by Prof. N. Hogg, London Research Institute, London, U.K. and Prof. S.K.A. Law, School of Biological Sciences, Nanyang Technological University, Singapore) was cultured in full RPMI 1640 medium. Platinum-E retroviral packaging and L929 cell lines (provided by Prof. K.E. Karjalainen, School of Biological Sciences, Nanyang Technological University, Singapore) were cultured in full DMEM and IMDM media, respectively.

Animal studies were approved by the Institutional Animal Care and Use Committee at Nanyang Technological University, Singapore. Bone marrow cells isolated from ITGAM^tm1Myd/g (integrin α_M^−/−) mice (The Jackson Laboratory, Bar Harbor, ME) were transduced by retroviral method with pMyc-IRES plasmid carrying the HOXB4 and Nup98 genes to generate long-term hematopoietic stem cells according to a previous study (25). In brief, Platinum-E cells were transfected with the plasmid using the calcium phosphate method. After 2 d, the culture supernatant containing the retrovirus was collected and used to transduce freshly isolated bone marrow cells. The cells were cultured and selected in IMDM medium containing 2% (v/v) FBS, 2 mM glutamine, 1.5 μg/ml puromycin supplemented with stem cell factor, and IL-6 conditioned media in a bacteriological petri dish. To generate immortalized bone marrow cells with rescued expression of integrin α_M (wild-type [WT] or P1147S), immortalized bone marrow cells (5 × 10⁶) were incubated in 5 ml of retrovirus-containing culture supernatant harvested from Platinum-E cells that were previously transfected with either empty pMyc-IRES-GFP plasmid or plasmid containing mouse integrin α_M(WT) or α_M(P1147S) cDNA. Polybrene (30 μg) was included in the culture mix. Cells were then harvested and sorted for GFP expression using FACS (FACSAria; BD Biosciences).

Macrophages were derived from these immortalized bone marrow cells by culturing in DMEM medium containing 10% (v/v) FBS, 2 mM glutamine, and 15% (v/v) conditioned medium of L929 cells that contained M-CSF-1. The latter was prepared by collecting the DMEM culture supernatant from L929 cells after 7 d in culture. Bone marrow–derived macrophage (BMDM) differentiation medium was changed once after 4 d in culture.

Mouse primary neutrophils were isolated by negative selection using the EasySep mouse neutrophil enrichment kit (Stemcell Technologies, Vancouver, BC, Canada). The isolated neutrophils were resuspended in appropriate medium and were used immediately in subsequent experiments.

Integrin α_M^−/− mice (6–8 wk of age) were subjected to two 5.5 Gy doses of gamma irradiation (each for 2 min and 4 h apart). After 24 h, transplantation was performed. Immortalized bone marrow cells that were reconstituted with integrin α_M (WT, P1147S, or empty vector [EV]) were washed twice in PBS and resuspended in PBS containing 2% (v/v) FBS to a final concentration of 2.5 × 10⁶ cells/100 μl. Two hundred microliters of cells was injected into each mouse via the retro-orbital venous route. Mice were fed with sterile water containing 1.17 g/ml neomycin for 2 wk.

Full-length human integrin α_M and β₂ cDNAs in expression plasmid pcDNA 3.0 were previously described (26). Human integrin α_M P1146S, β₂ Q746*, and β₂ T758G were generated by site-directed mutagenesis using the QuikChange II site-directed mutagenesis kit (Agilent Technologies, Santa Clara, CA) and relevant primers. Full-length mouse integrin α_M (WT) cDNA was purchased from Source BioScience LifeSciences (Nottingham, U.K.). To subclone the full-length mouse integrin α_M (WT) cDNA into the pMyc-IRES-GFP plasmid (a gift from Prof. K.E. Karjalainen, School of Biological Sciences, Nanyang Technological University, Singapore), a silent point mutation was introduced in the α_M cDNA to destroy the XhoI restriction site. The mouse integrin α_M cDNA was PCR amplified using relevant primers containing NotI and XhoI restriction sites and subcloned into pMyc-IRES-GFP digested with the same enzymes. Site-directed mutagenesis was performed using relevant primers to generate mouse integrin α_M P1147S in pMyc-IRES-GFP. An N-terminal hemagglutinin (HA)-tagged 14-3-3ζ (M1-S230) in pcDNA 3.1/Zeo(−) was generated by PCR subcloning using the template plasmid pGEX-4T1–14-3-3ζ obtained from Addgene (Cambridge, MA). The region Met¹–Ser²³⁰ was selected based on previous studies on 14-3-3ζ in complex with CT peptides of integrins α₄ (Protein Data Bank: 4HKC) and β₂ (Protein Data Bank: 2V7D) (27, 28). For fluorescence resonance energy transfer (FRET) analyses, full-length human integrin α_M (WT) and α_M (P1146S) in pEYFP-N1 (Clontech Laboratories, Mountain View, CA) (29) were generated by PCR subcloning of the respective cDNAs into the plasmids using the KpnI and AgeI restriction sites. The construct therefore carries the integrin α_M cDNA sequence with a C-terminal EYFP sequence. 14-3-3ζ (M1-S230) in pECFP-C1 was generated by PCR subcloning using the KpnI and BamHI restriction sites. The construct therefore contains an N-terminal ECFP sequence followed by the 14-3-3ζ sequence. Escherichia coli expression plasmid pET14b containing 14-3-3ζ (M1-S230) was generated by Shanghai ShineGene Molecular Biotech (Shanghai, China).

TRIzol reagent (Invitrogen) was used to extract total RNA from cells. RNA was air dried and resuspended in RNAse-free water. Quantitative RT-PCR (qRT-PCR) was performed using the Ambion Cells-to-CT kit (Life Technologies, Carlsbad, CA) on a CFX96 real-time PCR detection system (Bio-Rad Laboratories, Hercules, CA). The cycling parameters were 95°C for 10 min, 95°C for 30 s, 58°C for 1 min, and repeated from step 2 for 39 times, followed by melt-curve analysis from 58°C to 95°C with a 0.5°C increment every 10 s. The specificity of the primers was verified by analyzing the melt-curve data and by performing agarose gel electrophoresis of amplified products from standard RT-PCR. Mouse IL-6 primers were: forward, 5′-GCCAGAGTCCTTCAGAGAGA-3′, reverse, 5′-TCTTGGTCCTTAGCCACTCC-3′. Mouse TNF-α primers were: forward, 5′-CTCCCAGGTTCTCTTCAAGG-3′, reverse, 5′-TGGAAGACTCCTCCCAGGTA-3′. Mouse GAPDH primers were: forward, 5′-GGTGAAGGTCGGTGTGAACG-3′, reverse, 5′-CTCGCTCCTGGAAGATGGTG-3′.

Cells (2 × 10⁶) were transfected with relevant plasmids (8 μg each) by electroporation on a Microporator MP-100 according to the manufacturer’s instructions (Invitrogen) (30). Electroporated cells were resuspended in 2 ml of culture medium containing 20% (v/v) heat-inactivated FBS and maintained in culture for 24 h before being used in subsequent experiments.

Transfected JB2.7 and K562 cells were analyzed by flow cytometry as previously described (31). Cells were incubated in PBS containing 20 μg/ml primary Ab (mAb LPM19c) for 30 min on ice. Cells were washed in PBS and incubated in PBS containing FITC-conjugated sheep anti-mouse F(ab′)₂ secondary Ab (1:400 dilution) (Sigma-Aldrich) for 30 min on ice. Stained cells were washed in PBS and analyzed on a FACSCalibur, FACS LSR II, or FACS LSRFortessa X-20 flow cytometer (BD Biosciences). Data were analyzed and presented using FlowJo software (Tree Star, Ashland, OR). Neutrophils and BMDMs were analyzed by flow cytometry using the following Abs: allophycocyanin/Cy7 anti-mouse/human CD11b, allophycocyanin/Cy7 rat IgG2b κ isotype control (BioLegend, San Diego, CA), PE anti-mouse Ly6G/Ly6C (Gr-1), PE anti-mouse F4/80, and PE anti-rat IgG2b κ isotype control (eBioscience, San Diego, CA).

An uncoated μ-slide I^0.4 Luer flow chamber (ibidi, Munich, Germany) was coated with 7.5 μg/ml iC3b (Complement Technology, Tyler, TX) in PBS at 4°C overnight (32). The solution was removed and replaced with 0.2% (w/v) polyvinylpyrrolidone (10,000) in PBS at room temperature (RT) for 1 h to block nonspecific binding sites. Bone marrow–isolated neutrophils (5 × 10⁵) were incubated in 1 ml of HBSS (1 mM CaCl₂, 1 mM MgCl₂, 5% [v/v] heat-inactivated FBS, 10 mM HEPES [pH 7.4]) without or with 200 nM phorbol dibutyrate (PdBu) for 5 min at 37°C. Cells were then infused into the chamber at a constant shear stress of 0.4 dyne/cm² with an automated syringe pump (Harvard Apparatus, Holliston, MA). The experiment was performed at 37°C in a custom-built microscope-stage incubator equipped with a ×10 objective lens and a CCD camera. The number of adherent cells in four different fields (dimension 1 mm versus 1 mm) was recorded. The average number of adherent cells in four fields was calculated. Data were analyzed and presented based on the following: number of adherent cells with PdBu stimulation − number of adherent cells without stimulation.

Similar experiments were performed using transfected JB2.7 and K562 cells except that 6 × 10⁵ cells were used and the applied shear stresses were 0.6 and 0.5 dyne/cm², respectively. Data for JB2.7 cells were analyzed and presented based on the following: number of adherent cells with PdBu − number of adherent cells without PdBu. In the experiment using transfected K562 cells, the β₂ integrin activating Ab KIM185 (10 μg/ml) was included (33).

In the experiment involving HA–14-3-3ζ overexpression in K562 cells, cells were transfected with β₂ T758G and α_M(WT)enhanced yellow fluorescent protein (eYFP) or α_M(PS)eYFP and HA–14-3-3ζ or empty plasmid by electroporation. Cells were pretreated without or with 200 nM PdBu for 5 min at 37°C followed by shear flow analyses on iC3b-coated flow chamber as described above. LPM19c (10 μg/ml) was included to demonstrate ligand-binding specificity.

BMDMs (1 × 10⁵) were resuspended in 200 μl of culture medium and dispensed into one well of the PolySorp microtiter plate (Nalge Nunc, Rochester, NY) that was precoated overnight with 80 μg/ml fibrinogen (Sigma-Aldrich) in sodium bicarbonate buffer (pH 9). Cells were incubated under culture conditions for 15 and 30 min. Four wells were used for each time point. Untreated cells were used for the 0 min time point. At each time point, cells were harvested and lysed in lysis buffer containing protease inhibitors (Roche, Basel, Switzerland) and phosphatase inhibitors (Nacalai Tesque, Kyoto, Japan). Total cell lysate was used in subsequent immunoblotting assays. To examine the activation of MAPKs in BMDMs on iC3b, the same procedure was used except that cells were dispensed into wells of standard cell culture microtiter plates that were preincubated with 1:1 (v/v) mixture of mouse serum and HBSS buffer for 2 h at 37°C. This method is based on serum iC3b deposition on cell culture plastic as previously described (34, 35). For experimental conditions that required integrin activation, 100 ng/ml LPS (Sigma-Aldrich) or 1 mM MnCl₂ were used. In the experiments involving Src family kinase (SFK) inhibition, cells were pretreated with 20 μM PP2 or the inactive analog PP3 for 2 h under culture conditions. Syk inhibition in cells was performed by pretreating cells with 20 μM piceatannol or DMSO carrier for 1 h under culture conditions. Inhibitors previously mentioned were purchased from Abcam (Cambridge, U.K.).

Cells were lysed in lysis buffer (1% [v/v] Igepal, 150 mM NaCl, and 50 mM Tris [pH 7.5]) containing protease inhibitors (Roche) and phosphatase inhibitors (Nacalai Tesque). Total protein concentration was determined using the bicinchoninic acid assay kit (Thermo Scientific). Proteins were resolved by SDS-PAGE under reducing conditions and electrotransferred onto polyvinylidene difluoride (PVDF) membrane (Bio-Rad Laboratories). For total protein immunoblotting, PVDF membrane was incubated in TBST blocking buffer containing 5% (w/v) nonfat milk for 30 min at RT. The membrane was incubated in TBST buffer containing relevant primary Abs at 4°C overnight. The following primary Abs were used: rabbit anti-MAPK1/2 (Erk1/2) (Millipore, Billerica, MA), rabbit anti-p38 MAPK (Cell Signaling Technology, Danvers, MA), rabbit anti-HA (Delta Biolabs, Gilroy, CA), and rabbit anti–14-3-3ζ (Abcam). To detect the phosphorylated proteins, PVDF membrane was incubated in TBST blocking buffer containing 5% (w/v) BSA for 30 min at RT. The following Abs were used: mouse anti-p44/42 MAPK (Erk1/2) (Thr²⁰²/Tyr²⁰⁴) and rabbit anti–phospho-p38 MAPK (Thr¹⁸⁰/Tyr¹⁸²) (both from Cell Signaling Technology). The membrane was washed three times in TBST buffer followed by incubation in TBST buffer containing appropriate HRP-conjugated secondary Abs for 1 h at RT. Protein bands were detected by enhanced chemiluminescence using autoradiography or on a ChemiDoc Touch Gel and Western blot imaging System (Bio-Rad Laboratories). To determine the level of phosphorylated Erk1/2 or p38, the membrane was immunoblotted for the phosphorylated protein and the same membrane was stripped and reimmunoblotted for total protein.

Acceptor-photobleaching FRET analysis was performed as previously described but with modifications (29). K562 cells were transfected by electroporation with plasmids β₂ Q746* and α_M(WT)eYFP or α_M(PS)eYFP and enhanced cyan fluorescent protein (eCFP)–14-3-3ζ or pECFP-C1 EV (10 μg each). Transfected cells were treated without or with 200 nM PdBu for 5 min at 37°C and centrifuged onto poly-l-lysine slides by cytospinning. FRET was performed on a Zeiss LSM 510 confocal microscope (Carl Zeiss, Thormwood, NY) using the following laser lines and emission filters: eCFP (excitation wavelength at 458 nm and band-pass filter 470–500 nm); eYFP (excitation wavelength at 514 nm and low-pass filter 539 nm). eYFP was photobleached between the fifth and sixth time points by scanning the entire cell 20 times using the 514 argon laser line that was set at the maximum intensity. The eCFP signal of the cell membrane was acquired before and after the photobleach. FRET efficiency was calculated using the equation: FRET efficiency = [(I₆ – I₅)/I₆] × 100, where I is the eCFP intensity at the nth time point. Control samples that were not subjected to eYFP photobleaching (unbleached) were also analyzed.

Macrophage adhesion and spreading on iC3b and fibrinogen were determined by the electric cell–substrate impedance sensing (ECIS) method (36). The wells of a 16-well E-plate (ACEA Biosciences, San Diego, CA) were incubated with 4 mg/ml dithiobis(succinimidyl propionate) (Pierce; Thermo Fisher Scientific, Rockford, IL) in DMSO (Thermo Scientific) for 30 min at RT. The wells were washed twice in molecular biology–grade water and incubated with 80 μg/ml mouse fibrinogen (Haematologic Technologies, Essex Junction, VT) in PBS for 3 h at RT. Thereafter, the wells were washed twice with PBS. Each well was filled with 100 μl of culture medium and the background signal was measured on a real-time cell electronic system (ACEA Biosciences). Macrophages (1 × 10⁵) were seeded into each well, without or with 100 ng/ml LPS. The cell index, which corresponds to cell adhesion/spreading, was determined during a period of 100 min. Each datum point represents the SD of technical triplicates. Data points were plotted as a function of time at 10-min intervals. For macrophage adhesion/spreading on iC3b, each dithiobis(succinimidyl propionate)–treated well was coated with 7.5 μg/ml iC3b in PBS for 3 h at RT.

Transwell migration assays were performed using the Corning Costar Transwell permeable supports (Thermo Fisher Scientific, Waltham, MA). The Transwell insert (8 μM) was coated with 80 μg/ml fibrinogen in PBS overnight at 4°C. The insert was washed with PBS and placed into a well (24-well format) filled with 500 μl of BMDM culture medium containing 100 ng/ml mouse SDF-1α (PeproTech, Rocky Hill, NJ). Cells were incubated in serum- and M-CSF-1–free culture medium for 6 h under culture conditions. Cells (8 × 10⁴) were resuspended in 300 μl of serum- and M-CSF-1–free culture medium without or with 100 ng/ml LPS and were placed in the upper chamber followed by incubation under culture conditions for 16 h. A cotton swab was used to remove cells that did not migrate through the upper chamber insert. Cells underneath the upper chamber insert were fixed in 4% (w/v) paraformaldehyde in PBS for 10 min at RT. Cells were permeabilized with cytoskeletal buffer containing 0.25% (v/v) Triton X-100 for 2 min at RT followed by staining in crystal violet solution (0.5% w/v) for 20 min at RT. Stained cells were extensively washed in PBS. The dye was extracted using 0.25% (v/v) Triton X-100 in double distilled H₂O and absorbance at 595 nm was measured on a Benchmark Plus microplate spectrophotometer (Bio-Rad Laboratories).

FluoSpheres carboxylate-modified microspheres (1.0 μm, crimson fluorescent, excitation/emission of 625/645) (Invitrogen) were incubated in HBSS buffer containing 20% (v/v) mouse serum at 4°C overnight. Beads were washed and resuspended in HBSS before use. BMDMs (5 × 10⁵) were preincubated in 500 μl of culture medium for 30 min at 4°C or 37°C. Cells were mixed with the beads at a cell/bead ratio of 1:100 in the presence or absence of LPS (100 ng/ml) for 30 min at 4°C or 37°C. Nonphagocytosed beads were removed by washing cells in 500 μl of PBS containing 4 mM EDTA and 1% (w/v) BSA. Cells were subjected to flow cytometry analysis. Expression index (EI) was calculated based on percentage gated positive × geometric mean. Phagocytic index was calculated and plotted based on EI at 37°C − EI at 4°C.

Production of cytokines in cells was detected using the GolgiPlug kit (BD Biosciences). BMDMs were incubated in serum- and M-CSF-1–free medium (1 ml) containing 1 μg of GolgiPlug (BD Biosciences) for 2 h under culture conditions. Cells (5 × 10⁵) were resuspended in 200 μl of full medium containing 0.2 μg of GolgiPlug without or with 100 ng/ml LPS and plated onto iC3b-coated culture dish (35) for 6 h under culture conditions. Cells were harvested and stained with either allophycocyanin-conjugated anti-mouse TNF-α or anti-mouse IL-6 Ab (eBioscience) according to the manufacturer’s instructions followed by flow cytometry analyses. Fold expression was calculated based on EI and the expression level in cells bearing integrin α_M(WT)β₂ in the absence of LPS was normalized to 1.0.

In brief, E. coli BL21(DE3) cells were transformed with pET14b–6xHis–14-3-3ζ (Met¹–Ser²³⁰). Transformed cells were grown at 37°C in Luria–Bertani medium until it reached an OD₆₀₀ reading of 0.6–0.7. Isopropyl β-d-thiogalactoside was added to the culture to an end concentration of 0.5 mM followed by incubation for 10–12 h at 18°C. Cells were pelleted by centrifugation and lysed by sonication in 50 mM sodium phosphate buffer (pH 8) containing 300 mM NaCl followed by affinity purification on a Ni-NTA column (GE Healthcare, Amersham, U.K.). Recombinant protein was further purified by size exclusion chromatography in 20 mM sodium phosphate buffer (pH 6) containing 50 mM NaCl using a Hiload Superdex 75 16/26 GL preparative column (flow rate of 0.3 ml/min) on a ÄKTAFPLC UPC-900 system (GE Healthcare).

N-terminal biotin-labeled peptides of full-length integrin α_M (WT) and α_M (pS1142) tails were synthesized by ChinaPeptides (Shanghai, China). N-terminal biotin-labeled peptides of full-length α_M (PS) and α_M (P1146pS) tails were synthesized by GL Biochem (Shanghai, China). Pull-down assays were performed by mixing 60 μl of streptavidin agarose beads (Sigma-Aldrich) with or without (for uncoated) 20 μg of peptide and 3 μg of purified His⁶–14-3-3ζ (M1-S230) in 25 mM Tris (pH 7.4) containing 150 mM NaCl, 0.1% (w/v) BSA, 0.02% (v/v) Tween 20, and 0.5 mM TCEP. The mixture was rotated for 30 min at RT followed by washing in the same buffer. Proteins were eluted by boiling the beads in SDS sample loading buffer containing 40 mM DTT followed by SDS-PAGE and immunoblotting to detect 14-3-3ζ.

E. coli BL21(DE3) cells were transformed with GST–rhotekin binding domain (RBD) pGEX-2T plasmid (Addgene) or pGEX-KG empty plasmid. Cells in 500 ml of Luria–Bertani medium with an OD₆₀₀ reading of 0.6 were induced to express the GST-RBD or GST by adding isopropyl β-d-thiogalactoside to a final concentration of 0.2 mM followed by incubation for 6 h at 30°C. Cells were pelleted by centrifugation and lysed by sonication in lysis buffer (100 mM NaCl, 0.5 mM PMSF, 0.5 mM TCEP, and 50 mM Tris-HCl [pH 7.4]) containing protease inhibitors. GST-RBD or GST was affinity purified on a glutathione-Sepharose column. An active RhoA pull-down assay was performed according to the protocol of the RhoA G-LISA activation assay kit (Cytoskeleton, Denver, CO) but with modifications. Glutathione-Sepharose bead suspension (15 μl) was washed once in 50 mM Tris-HCl (pH 7.5) buffer containing 2% (v/v) Igepal, 500 mM NaCl, and 10 mM MgCl₂. Beads and 45 μg of either GST-RBD or GST were incubated in the same buffer for 1 h at 4°C with rotation. At the end of the incubation, beads were washed in the same buffer. BMDMs were serum and M-CSF-1 starved for 6 h. Thereafter, 1 × 10⁵ cells were resuspended in 200 μl of full culture medium containing 100 ng/ml LPS and seeded into one well of the PolySorp microtiter plate that was precoated with 80 μg/ml mouse fibrinogen. For each time point, cells from 12 wells were used. Cell lysate containing 60 μg of total proteins as determined by bicinchoninic acid assay was used for each pull-down assay.

Statistical analyses were performed using the Prism software (GraphPad Software, La Jolla, CA). A Student t test (two-tailed and unequal variance) was used unless indicated otherwise. A p value <0.05 was considered significant.

The nonsynonymous SNP (rs1143678) leads to P1146S substitution in human integrin α_M CT (Fig. 1A). Human and mouse integrin α_M CTs have high sequence identity. The residue corresponding to Pro¹¹⁴⁶ of human integrin α_M CT is Pro¹¹⁴⁷ in mouse integrin α_M CT. P1146S (human) and P1147S (mouse) substitutions are henceforth referred to as PS. To examine the effects of PS substitution on integrin α_Mβ₂ function, the human T cell line JB2.7 that does not express the α subunits of the β₂ integrins was transfected to express either human α_M(WT) or α_M(PS) (Fig. 1B). Under shear flow and in the presence of PdBu, which induces integrin activation, the number of integrin α_M(PS)β₂ cells that adhered to iC3b was significantly higher than that of integrin α_M(WT)β₂ cells. To determine whether PS substitution promotes integrin α_Mβ₂ inside-out or outside-in signaling, similar experiments were performed using transfected human chronic myelogenous leukemia K562 cells without treatment or treated with the β₂ integrin–activating mAb KIM185 that bypasses inside-out signaling (33) (Fig. 1C). In the presence of mAb KIM185, the number of integrin α_M(PS)β₂ cells that adhered to iC3b was markedly higher than that of integrin α_M(WT)β₂ cells. These data suggest that PS substitution promotes integrin α_Mβ₂ outside-in signaling.

The effects of PS substitution on integrin α_Mβ₂ function were further verified using mouse myeloid cells. Adhesion of integrin α_M^−/− neutrophils to iC3b in the presence of PdBu was significantly less than that of WT neutrophils (Fig. 1D). We next isolated bone marrow cells from integrin α_M^−/− mice and performed retrovirus-mediated transduction of these cells with the Nup-98-HOXB4 expression plasmid. This procedure has been reported to enable self-renewal of bone marrow hematopoietic stem cells in vitro but they retained their capacity to differentiate in vitro and in vivo (25). These cells were then transduced with pMyc-IRES-GFP EV, pMyc-mouse α_M(WT)-IRES-GFP or pMyc-mouse α_M(PS)-IRES-GFP. GFP⁺ cells were isolated by FACS and then transplanted into irradiated integrin α_M^−/− mice. Bone marrow neutrophils were isolated from these mice 7 wk later. Integrin α_M(WT) and α_M(PS) expression levels were analyzed by flow cytometry (Fig. 1E). In line with previous data, the number of neutrophils expressing integrin α_M(PS)β₂ that adhered to iC3b under shear flow was significantly higher than that of integrin α_M(WT)β₂ and EV neutrophils (Fig. 1E).

Considering the low percentage (∼21%) of integrin α_M⁺ neutrophils that could be isolated from the transplanted mice, we decided to perform the rest of the study using F4/80⁺ macrophages that were obtained from in vitro differentiation of integrin α_M(WT) and α_M(PS) bone marrow cells (37, 38) (Fig. 2A). Integrin α_M(PS)β₂ BMDMs showed relatively higher levels of basal and LPS-activated adhesion and spreading on ligands than those of integrin α_M(WT)β₂ and EV cells as determined by the ECIS method (Fig. 2B). LPS was included because it activates macrophages via TLR4 signaling (39, 40). These data are in line with the increased adhesive properties of mouse neutrophils and human cell lines expressing integrin α_M(PS)β₂. Furthermore, the migration of LPS-stimulated integrin α_M(PS)β₂ BMDMs on fibrinogen was significantly faster than integrin α_M(WT)β₂ and EV cells in Transwell assays (Fig. 2C).

Integrin α_Mβ₂–mediated activation of Erk1/2 has been reported (41). Therefore, we asked whether the kinetics of Erk1/2 activation induced by integrin α_M(PS)β₂ and α_M(WT)β₂ are different. BMDMs expressing these integrins were seeded into empty culture wells in the presence of LPS- or ligand-coated culture wells in the absence of LPS with different incubation times. Cells were collected and lysed followed by immunoblotting to determine the levels of activated Erk1/2 and p38 MAPK (42) (Fig. 3A). There was no significant difference in the activation kinetics of Erk1/2 or p38 MAPK between the two groups of cells under all three conditions. We next examined cells that were seeded into ligand-coated culture wells in the presence of LPS or Mn²⁺, which activates integrin α_Mβ₂ (32, 43). Under these conditions, faster kinetics of Erk1/2 activation was detected in integrin α_M(PS)β₂ BMDMs as compared with those in integrin α_M(WT)β₂ BMDMs (Fig. 3B). There was, however, no significant difference in the kinetics of p38 activation, suggesting signaling specificity that involves integrin α_Mβ₂ and Erk1/2. Taken together, these data suggest that the overt activation of Erk1/2 mediated by integrin α_M(PS)β₂ requires both integrin activation and ligand engagement. SFKs and Syk are downstream effectors of integrin α_Mβ₂ signaling (2). The Syk inhibitor piceatannol and the pan-SFK inhibitor PP2 but not its inactive analog PP3 effectively reduced the level of Erk1/2 activation in both integrin α_M(PS)β₂ and α_M(WT)β₂ BMDMs (Supplemental Fig. 1).

qRT-PCR was performed to examine the mRNA expression levels of IL-6 and TNF-α in the BMDMs. In the presence of ligand and LPS, the expression levels of IL-6 and TNF-α transcripts were higher in integrin α_M(PS)β₂ BMDMs than in integrin α_M(WT)β₂ BMDMs (Fig. 4A). These findings were verified by performing intracellular staining of IL-6 and TNF-α using the GolgiPlug kit and then analyzed by flow cytometry (Fig. 4B). We next performed a bead-based phagocytosis assay to examine the phagocytic capacity of these cells (Fig. 4C). The level of serum-opsonized bead engulfment in integrin α_M(PS)β₂ BMDMs was markedly higher than that of integrin α_M(WT)β₂ BMDMs. RhoA is involved in integrin α_Mβ₂–mediated phagocytosis of complement-opsonized particles (44). To determine whether these cells have different levels of active RhoA, we performed a RhoA pull-down assay (Fig. 4D). The level of active RhoA was significantly higher in integrin α_M(PS)β₂ BMDMs than in integrin α_M(WT)β₂ BMDMs at the 15 min time point. Taken together, these data suggest that PS substitution in integrin α_M CT causes BMDMs to become proinflammatory and to exhibit enhanced phagocytosis capacity.

We next examined whether PS substitution in integrin α_M CT could lead to aberrant interaction with cytosolic protein, thereby modulating the functions of integrin α_Mβ₂. The 14-3-3 family of proteins, which forms homo- or heterodimers, binds phospho-Ser/Thr–containing sequences, and they regulate diverse cellular processes (45). 14-3-3ζ interacts with the phospho-Thr triplet of the integrin β₂ CT and it positively regulates β₂ integrin–ligand binding (23, 24). This interaction could induce integrin clustering based on structure analysis of a 14-3-3ζ homodimer that was in complex with two integrin β₂ CT peptides (28). We hypothesized that PS substitution in integrin α_M CT could generate an aberrant phospho-Ser site that is recognized by 14-3-3ζ. To this end, we expressed and purified recombinant 14-3-3ζ(M1-S230) as previously described (27, 28). A pull-down assay was performed using the recombinant 14-3-3ζ, streptavidin-coated agarose beads, and the following biotin-tagged full-length integrin α_M CT peptides: WT, pS1142, P1146S, and P1146pS (Fig. 5A). The peptide pS1142 was included in the assay because phosphorylation of Ser¹¹⁴² in integrin α_M(WT) CT has been shown to positively regulate the ligand-binding function of integrin α_Mβ₂ (12) (Fig. 1A). 14-3-3ζ associated with P1146S and P1146pS but not with WT and pS1142 peptides. These data suggest that PS substitution in integrin α_M CT generates a 14-3-3ζ binding site, and the interaction apparently does not require the phosphorylation of Ser¹¹⁴⁶.

To verify the interaction between integrin α_M(PS) CT and 14-3-3ζ, we performed acceptor-photobleaching FRET on transfected cells. K562 cells were transfected with different combinations of the following expression plasmids: β₂ Q746*, α_M(WT)eYFP, α_M(PS)eYFP, eCFP–14-3-3ζ, and eCFP EV (Fig. 5B). β₂ Q746* is a truncated integrin β₂ mutant in which a stop codon was introduced after Gln⁷⁴⁶, thereby removing the 14-3-3ζ binding site from the β₂ CT. Hence, any FRET signal detected is a result of the interaction between 14-3-3ζ and α_M but not β₂ integrin CT. Without PdBu treatment, the basal FRET level of cells expressing α_M(WT), β₂ Q746*, and 14-3-3ζ was not significantly different from that of cells expressing α_M(PS), β₂ Q746*, and 14-3-3ζ (Fig. 5B). Basal FRET levels in these two groups of cells were higher than those of control groups (cells expressing the eCFP EV), which could be attributed to the method’s high sensitivity. In the presence of PdBu, the FRET level of cells expressing α_M(PS), β₂ Q746*, and 14-3-3ζ was significantly higher than that of cells expressing α_M(WT), β₂ Q746*, and 14-3-3ζ (Fig. 5C), suggesting that the interaction between integrin α_M(PS) CT and 14-3-3ζ in cells requires integrin activation. These data are in line with our earlier observations that cellular stimulation triggers overt responses from cells expressing integrin α_M(PS)β₂.

To verify the dependency of the interaction between integrin α_M(PS) CT and 14-3-3ζ on cellular stimulation, we performed additional functional assays. In order not to disrupt interactions between other cytosolic proteins and integrin β₂ CT, which are required for the function of integrin α_Mβ₂, we generated an expression plasmid containing full-length integrin β₂ with a point mutation T758G. We showed by FRET assay that this mutant interacted minimally with 14-3-3ζ (Fig. 5D). We next transfected K562 cells with HA–14-3-3ζ and integrin α_M(WT)β_2(T758G) or α_M(PS)β_2(T758G) and examined the binding of these cells to iC3b under shear flow (Fig. 5E). In the absence of PdBu stimulation, there was no significant difference between the two groups of cells despite them overexpressing HA–14-3-3ζ. When treated with PdBu, a significantly higher number of cells expressing integrin α_M(PS)β_2(T758G) and HA–14-3-3ζ adhered to iC3b compared with cells expressing integrin α_M(WT)β_2(T758G) and HA–14-3-3ζ. Ligand-binding specificity was demonstrated by including the function-blocking Ab LPM19c.

The ITGAM SNP rs1143678 leads to P1146S substitution in the integrin α_M CT. In this study, we showed that PS substitution promotes integrin α_Mβ₂-mediated cell adhesion, spreading, migration, and phagocytosis. The effects of P1146S substitution on integrin α_Mβ₂–mediated functions were most pronounced when cells were stimulated with LPS or PdBu and in the presence of ligands, suggesting that PS substitution primarily modulates integrin α_Mβ₂ outside-in signaling. This is also evident from the shear flow iC3b-binding data of transfected K562 cells treated with β₂ integrin–activating mAb KIM185, which bypassed inside-out integrin α_Mβ₂ activation.

The clustering of integrin α_Mβ₂ activates Erk1/2 through the Src/Syk pathway (10). We showed that integrin α_M(PS)β₂ also signals through this pathway. However, BMDMs expressing integrin α_M(PS)β₂ that were treated with LPS together with iC3b or fibrinogen exhibited faster kinetics of Erk1/2 activation and produced higher levels of IL-6 and TNF-α when compared with integrin α_M(WT)β₂ BMDMs. These findings corroborate well with the overproduction of IL-6 and TNF-α detected in SLE patients (46, 47). Apart from genetic factors, environmental factors can contribute to the onset of SLE (48). TLRs are involved in the etiology of SLE (49). TLR-4 recognizes Gram-negative bacteria LPS (50). Overt signaling from TLR-4 has been shown to break immune tolerance, leading to lupus-like renal disease in mice (51). When lupus-prone C57BL/6^lpr/lpr mice were crossed with TLR-4 ^−/− mice, their progenies produced lower levels of autoantibodies and they were less susceptible to renal injuries, suggesting that TLR-4 is involved in the onset of SLE (52). It is therefore physiologically relevant to treat BMDMs with LPS to examine the effects of PS substitution on integrin α_Mβ₂ function. In addition to LPS, heat shock protein 60 and high mobility group box 1 protein released from damaged cells interact with TLR-4 (53, 54). Hence microbial infections and tissue injuries may exacerbate the inflammatory responses in individuals carrying the ITGAM SNP rs1143678 allele.

Mechanistically, PS substitution in integrin α_M CT generates an aberrant 14-3-3ζ binding site. 14-3-3ζ positively regulates β₂ integrin–mediated functions by interacting with a phospho-Thr triplet motif in the β₂ CT (23, 24). Although 14-3-3 proteins bind phospho-Ser– or phospho-Thr–containing motifs in their binding partners, non–phospho-Ser/Thr ligands have been identified (55–58). Our data suggest that the interaction between 14-3-3ζ and integrin α_M(PS) CT is not dependent on the phosphorylation of Ser¹¹⁴⁶. Structural study has shown that the nonphosphorylated exo-S peptide interacts with a flexible binding groove in 14-3-3β (59). Whether integrin α_M(PS) CT interacts with the binding groove or another site in 14-3-3ζ remains to be investigated.

Hetero- or homodimeric 14-3-3 proteins have been identified (60, 61). Structural study demonstrated that a 14-3-3ζ homodimer can associate with two integrin β₂ CT peptides, suggesting a role of 14-3-3ζ in promoting the formation of integrin cluster (28). We therefore proposed a model in which the PS substitution in integrin α_M CT could lead to overt outside-in signaling and inflammatory response (Fig. 6). In the absence of cellular stimulation by LPS or phorbol ester, the binding of 14-3-3ζ to the PS site is not favorable because the CTs of integrin α_M(PS)β₂ are in a clasped conformation, and the site may be occluded by negative regulators that directly interact with the integrin CTs. This is supported by FRET data that showed a significant increase in the interaction between 14-3-3ζ and integrin α_M(PS) CT compared with integrin α_M(WT) CT only under conditions that promote integrin activation.

When cells are stimulated by LPS or phorbol ester, this leads to integrin activation through the dissociation of the negative regulator from the integrin α_M(PS)β₂ CTs and the separation of the integrin α_M(PS) CT from the β₂ CT, thereby permitting 14-3-3ζ to bind to the PS site. A 14-3-3ζ homodimer can simultaneously bind to two integrin α_M(PS) CTs and induce integrin clustering, which in turn triggers downstream cellular responses. Because 14-3-3 proteins interact with a large number of cytosolic signaling molecules, the aberrant interaction between 14-3-3ζ and integrin α_M(PS) CT could activate signaling pathways other than the SFK–Syk–Erk1/2 pathway.

An interesting point of discussion refers to ITGAM SNP rs1143679 that leads to R77H substitution in the integrin α_M β-propeller. It has been shown to disrupt the formation of catch bonds between integrin α_M and its ligand, which impairs the ligand-binding function of integrin α_Mβ₂ (18–20). However, there are different observations reported in terms of the phagocytic capacity and cytokine production of cells bearing rs1143679 (62). Integrin α_Mβ₂ is involved in immune tolerance (63). It was therefore proposed that rs1143679 contributes to the onset of SLE because it reduces macrophage clearance of apoptotic bodies and it leads to deregulated production of inflammatory cytokines (21). Our findings that ITGAM SNP rs1143678 did not abrogate the ligand-binding function of integrin α_Mβ₂ and that it is proinflammatory may appear contrary to the effects of ITGAM SNP rs1143679. However, studies have shown that integrin α_Mβ₂ can serve both pro- and anti-inflammatory functions that are dependent on cell type and modulated by its cross-talk with membrane and cytosolic molecules, including CD14, FcγRIIA, FcγRIIIB, uPAR, and TLRs (62). Zhou et al. (22) reported that neutrophils from donors having all three allelic variants ITGAM SNP rs1143678 (P1146S), ITGAM SNP rs1143679 (R77H), and ITGAM SNP rs1143683 (A858V substitution in the calf-1 domain of integrin α_M) were defective in complement-opsonized phagocytosis and adhesion to ICAM-1. It is likely that R77H is dominant over P1146S because R77H directly abrogates integrin α_M–ligand interaction. This would have blunted the effects of P1146S that are dependent on integrin activation and ligand engagement. In the same study (22), phagocytosis and adhesion defects were also observed in neutrophils from donors having both ITGAM SNP rs1143678 (P1146S) and ITGAM SNP rs1143683 (A858V). A detailed functional characterization of A858V substitution is needed to explain the phenotype presented.

In conclusion, we showed that the SLE-associated nonsynonymous ITGAM SNP rs1143678 promotes integrin α_Mβ₂ proinflammatory functions. Mechanistically, P1146S substitution in the integrin α_M CT generates a noncanonical 14-3-3ζ binding site that leads to aberrant integrin α_Mβ₂ outside-in signaling. ITGAM SNP rs1143678 could therefore exacerbate inflammatory responses and tissue injuries in SLE patients that are exposed to not only self-antigens but also a milieu of environmental factors, including microbial pathogens.

We are very grateful to K.E. Karjalainen and S.C. Loh (School of Biological Sciences, Nanyang Technological University, Singapore) for providing advice and assistance for the generation and transplantation of the immortalized bone marrow cells. We thank Z.H. Xue (School of Biological Sciences, Nanyang Technological University, Singapore) for advice on the Syk-Erk signaling study. We thank L.G. Ng and laboratory members (Singapore Immunology Network [SIgN0], A*STAR, Singapore) for helpful advice.

The authors have no financial conflicts of interest.