Multiple Lupus-Associated ITGAM Variants Alter Mac-1 Functions on Neutrophils

Authors


Abstract

Objective

Multiple studies have demonstrated that single-nucleotide polymorphisms (SNPs) in the ITGAM locus (including the nonsynonymous SNPs rs1143679, rs1143678, and rs1143683) are associated with systemic lupus erythematosus (SLE). ITGAM encodes the protein CD11b, a subunit of the β2 integrin Mac-1. The purpose of this study was to determine the effects of ITGAM genetic variation on the biologic functions of neutrophil Mac-1.

Methods

Neutrophils from ITGAM-genotyped and -sequenced healthy donors were isolated for functional studies. The phagocytic capacity of neutrophil ITGAM variants was probed with complement-coated erythrocytes, serum-treated zymosan, heat-treated zymosan, and IgG-coated erythrocytes. The adhesion capacity of ITGAM variants, in adhering to either purified intercellular adhesion molecule 1 or tumor necrosis factor α–stimulated endothelial cells, was assessed in a flow chamber. Expression levels of total CD11b and activation of CD11b were assessed by flow cytometry.

Results

Mac-1–mediated neutrophil phagocytosis, determined in cultures with 2 different complement-coated particles, was significantly reduced in individuals with nonsynonymous variant alleles of ITGAM. This reduction in phagocytosis was related to variation at either rs1143679 (in the β-propeller region) or rs1143678/rs1143683 (highly linked SNPs in the cytoplasmic/calf-1 regions). Phagocytosis mediated by Fcγ receptors was also significantly reduced in donors with variant ITGAM alleles. Similarly, firm adhesion of neutrophils was significantly reduced in individuals with variant ITGAM alleles. These functional alterations were not attributable to differences in total receptor expression or activation.

Conclusion

The nonsynonymous ITGAM variants rs1143679 and rs1143678/rs113683 contribute to altered Mac-1 function on neutrophils. These results underscore the need to consider multiple nonsynonymous SNPs when assessing the functional consequences of ITGAM variation on immune cell processes and the risk of SLE.

Recent genome-wide association studies of human systemic lupus erythematous (SLE) have revealed strong associations between single-nucleotide polymorphisms (SNPs) in the ITGAM locus and susceptibility to SLE ([1, 2]). Following the initial reports of ITGAM SNP associations with SLE ([1, 2]), this observation has been replicated in many independent genetic studies across different ethnic groups ([3-5]). ITGAM encodes the α-subunit (known as CD11b) of the β2 integrin Mac-1 (also called CR3) ([6]). Notably, even before these results in prior genetic studies had implicated ITGAM as a major susceptibility locus in SLE, a study using an experimental mouse model demonstrated that lupus-prone MRL/MpJ-Faslpr mice rendered deficient in CD11b had an exaggerated autoimmune phenotype ([7]).

Mac-1 is broadly expressed on cells of the myeloid lineage and on a subset of lymphocytes ([8-11]). Mac-1 is a surface receptor involved in numerous cellular functions. On neutrophils, for example, Mac-1 is constitutively expressed, can be rapidly up-regulated upon cell activation, and is important for promoting firm adhesion to endothelial cells and subsequent transendothelial migration (via Mac-1 binding to ligands such as intercellular adhesion molecule 1 [ICAM-1] and ICAM-2, among others) ([12, 13]). Mac-1 also mediates neutrophil phagocytosis of both complement-opsonized and unopsonized particles ([14-17]). Furthermore, Mac-1 can modify the functions of other coexpressed receptors, such as Fc receptors and Toll-like receptors ([18-20]).

Based on the results of genetic studies, it has been suggested that the observed association of ITGAM with SLE in Caucasian and African American populations is attributable to the variation at the nonsynonymous SNP rs1143679 ([4]), which encodes an amino acid change from Arg to His at amino acid position 77 in the extracellular domain of CD11b. Since then, studies of the impact of ITGAM genetic variation on Mac-1–mediated biologic processes have almost exclusively focused on the influence of the rs1143679 SNP on the functions of Mac-1 in transduced cell lines and primary human monocytes ([21-23]). While these studies have variously demonstrated that rs1143679 can affect cell adhesion, phagocytosis, and cytokine production, it has also been observed that this SNP can occur in conjunction with other nonsynonymous ITGAM SNPs, which are in high linkage disequilibrium (LD) in this locus ([4]). The potential impact of these linked nonsynonymous SNPs on Mac-1–mediated functions has not been addressed in previous studies. Indeed, analyses in patients of different ethnicities have indicated a more complex pattern of association between ITGAM variation and SLE susceptibility ([5]). Therefore, multiple ITGAM SNPs, in addition to rs1143679, could be contributing to the genetic risk of SLE development.

In the present study, using a cohort of 1,815 healthy donors, we confirmed that multiple ITGAM nonsynonymous SNPs exist, including SNPs rs1143679, rs1143678 (amino acid change from Pro to Ser at position 1146), and rs1143683 (amino acid change from Ala to Val at position 858), and that these SNPs show strong LD. Furthermore, we provide the first experimental evidence that these multiple SLE-associated nonsynonymous ITGAM SNPs independently alter Mac-1–mediated neutrophil functions. The alterations in neutrophil Mac-1 functions associated with the variant ITGAM alleles, including decreased firm adhesion of neutrophils and reduced phagocytosis, could not be attributed to changes in the expression or activation of Mac-1, but could be linked to altered Fc receptor–mediated functions. The results of our study highlight the need for caution when interpreting the potential contribution to SLE of any single variant in ITGAM, as any functional differences observed between the common and variant forms of Mac-1 could be attributed to one or more highly linked variants.

MATERIALS AND METHODS

Reagents

RPMI medium, fetal bovine serum, phosphate buffered saline (PBS), tumor necrosis factor α (TNFα), and a SuperScript First-Strand Synthesis System reverse transcription–polymerase chain reaction (RT-PCR) kit were all from Invitrogen. Antibodies against CD11b (ICRF44 and CBRM1/5) and isotype-matched immunoglobulins were from eBioscience. Rabbit anti-sheep erythrocyte IgM antibodies were from Fitzgerald Industries International. Purified ICAM-1/Fc and P-selectin/Fc chimeras were from R&D Systems. Heavy Ficoll (Histopaque-1119), fMLP, protein A, gelatin, fibronectin, zymosan, human AB serum, rabbit anti-sheep erythrocyte IgG antibodies, and human C5–deficient serum were all from Sigma-Aldrich. Lymphocyte separation medium (light Ficoll) was from Mediatech. Human umbilical vein endothelial cells (HUVECs) and culture medium were from Lonza. Sheep erythrocytes were from Colorado Serum.

Human participants and genotyping

Healthy human donors with no chronic autoimmune diseases (n = 1,815) were recruited for our genotyping and functional assays. Healthy subjects were chosen in order to avoid the confounding effects of inflammation or medications on cellular functions. All donors recruited for these studies gave their informed consent to participate, and the study was approved by our Institutional Review Board.

Blood samples were obtained from all donors, and genomic DNA was isolated from the EDTA-treated whole blood samples using a Puregene DNA isolation kit (Qiagen). SNP rs1143679 was genotyped via TaqMan assay (Applied Biosystems), using an ABI 7900HT instrument. SNP rs1143678 was genotyped with the Pyrosequencing method, as previously described ([24]), using the sense primer 5′-biotin-AGC-TCG-GCT-TCT-TCA-AGC-GGC-A-3′, the antisense primer 5′-CAC-CGA-GAG-GCA-GCT-CTG-3′, and the pyrosequencing primer 5′-GGG-TTC-GGC-CCC-CGG-3′. In blood samples from a subset of 31 healthy donors whose neutrophils were used in the functional studies, the full-length ITGAM complementary DNA (cDNA) was directly sequenced using the Sanger sequencing strategy. The full-length ITGAM cDNA was then amplified with RT-PCR, using total RNA isolated from human leukocytes. A series of overlapping sequencing primers was used to sequence the whole open-reading frame of the ITGAM cDNA (details available from the corresponding author upon request).

Neutrophil isolation

Fresh anticoagulated blood was collected from participants by phlebotomy, and neutrophils were separated on a discontinuous gradient as previously described ([24]). The neutrophils were then washed and resuspended in complete RPMI medium. All functional assays were performed in duplicate, using neutrophils from different pairs of overlapping-genotype donors.

Phagocytosis assays

Sheep erythrocytes coated with rabbit anti-sheep RBC IgG antibody (referred to as EA), sheep erythrocytes coated with rabbit anti-sheep RBC IgM antibody plus complement (referred to as EAC), serum-treated zymosan (STZ), and heat-treated zymosan (HTZ) were prepared fresh each day, as described previously ([25]). Neutrophils from healthy donors were suspended in complete RPMI medium at a density of 5 × 106 cells per ml. Thereafter, the neutrophils (with or without priming for 10 minutes with 10−8M fMLP) were incubated with EA or EAC (each at a ratio of 1:20) for 30 minutes, or incubated with STZ or HTZ (each at a ratio of 1:10) for 15 minutes, at 37°C. Particles engulfed by neutrophils were counted by light microscopy with oil immersion. For each condition, at least 200 neutrophils were inspected, and the phagocytic index (number of ingested particles per 100 neutrophils) was calculated. The genotype of each neutrophil donor was unknown to the observers at the time of data acquisition.

Flow chamber adhesion assay

Neutrophil adhesion to purified ICAM-1 or to activated HUVECs was assessed under flow conditions, as previously described ([26]). Briefly, a mixture of 25 mg/ml recombinant human ICAM-1/Fc chimera and 0.5 mg/ml recombinant human P-selectin/Fc chimera (to initiate cell capture and facilitate cell rolling and firm adhesion to ICAM-1) were used as the purified substrates. HUVECs were cultured to full confluence (in 25-mm Corning culture dishes) and primed with 20 ng/ml human TNFα for 6 hours prior to each experiment.

Neutrophils (5 × 105 cells/ml in complete RPMI medium) were injected into the flow chamber at a shear stress of 1.5 dyne/cm2 via a programmable syringe pump, and cell movement was viewed via microscopy and recorded with a CCD camera (30 images per second). All experiments were done at 37°C. The number of firmly adherent neutrophils was determined by 2 independent scorers (YZ and NBW), without knowledge of the donors' genotypes. Four-minute–long digital movies of each adhesion experiment were viewed, and cells that were captured and adhered in the field of view were determined (expressed as adherent cells/minute) (results available from the corresponding author upon request). A cell that moved a distance of <1 cell diameter in 5 seconds was considered to be firmly adhered. For each donor, a minimum of 3 separate experiments were performed.

Mac-1 expression analyses

Freshly obtained blood (2 ml) was centrifuged, and peripheral blood cells were washed and suspended in PBS. Aliquots of 50 μl blood were prepared and fMLP was added to each (final concentration of 5 × 10−9M). Thereafter, the cells were incubated at 37°C for 15 minutes. After washing, fluorochrome-labeled anti-human CD11b antibodies (ICRF44 and CBRM1/5) were added to measure the expression of total CD11b and activated CD11b ([27]), respectively. The cells were then incubated on ice for 30 minutes. Erythrocytes were lysed with BD lysis buffer, and flow cytometry was done on a FACScan instrument. The captured flow cytometry results were analyzed using the FlowJo software package (Tree Star). Neutrophils were gated based on forward and side scatter patterns.

Statistical analysis

All data are presented as the mean ± SEM. Because the phagocytosis probes and adhesion substrates were freshly prepared daily, all studies assessing differences in phagocytosis and adhesion, using cells from homozygous donors, were performed using a paired experimental design, and results were analyzed using Student's paired t-test (GraphPad Prism, version 5). For analysis of donors heterozygous for variation at the ITGAM locus, analysis of variance models were built to test whether the mean level of phagocytosis varied between genotypes. Least squares means were calculated and utilized for pairwise comparisons.

RESULTS

Characterization of nonsynonymous variants in the human ITGAM locus

Variants in the human ITGAM locus have been found to be strongly and reproducibly associated with SLE susceptibility, and one nonsynonymous SNP, rs1143679, has been purported to be causal for this genetic association ([4]). However, several other nonsynonymous ITGAM SNPs are in strong LD with rs1143679, and the impact of these additional SNPs is seldom considered. Indeed, based on currently available data and our own genotyping results, there are 3 other nonsynonymous SNPs, in addition to rs1143679, in the ITGAM locus whose minor allele frequency (MAF) is >2% (see Supplementary Table 1, available on the Arthritis & Rheumatism web site at http://onlinelibrary.wiley.com/doi/10.1002/art.38117/abstract). SNPs rs1143679, rs1143683, and rs1143678 have all been shown to have a strong association with SLE in Caucasian populations. Attributing disease causality solely to the ITGAM SNP rs1143679 may, therefore, be premature. To examine the possible biologic consequences of the presence of nonsynonymous variants in high LD with rs1143679, we characterized ITGAM SNPs and their impact on neutrophil Mac-1 function, using a large population of genotyped healthy donors.

SNP rs1143679 encodes a c.328G>A, p.Arg77His change in exon 3 of ITGAM, which encodes part of the extracellular β-propeller domain of CD11b (Figure 1). It has been hypothesized that this change is sufficient to alter the overall structure of the CD11b protein and, thereby, alter its activity ([3]). Investigations have revealed the presence of conserved amino acids between CD11a and CD11b that have been mapped to the calf-1 region of CD11a ([28]). SNP rs1143683 (c.2671C>T) encodes a conservative amino acid change (p.Ala858Val) in or near the extracellular calf-1 region of CD11b, while rs1143678 (c.3436C>T) encodes a nonconservative amino acid change (p.Pro1146Ser) in the cytoplasmic tail of CD11b (Figure 1).

Figure 1.

Schematic diagram of the CD11b protein domains and the location of 3 nonsynonymous single-nucleotide polymorphisms (SNPs). Variant alleles of SNPs rs1143678, rs1143683, and rs1143679 are underlined. The codes of each donor with the indicated genotypes are shown. CY = cytoplasmic domain; EC = extracellular domain.

Existing HapMap data, the results of prior studies ([5]), and the ITGAM cDNA sequencing results for all donors used in our functional studies demonstrated that the SNPs rs1143683 and rs1143678 were in perfect LD. In addition, it was found that strong LD exists between SNPs rs1143678 and rs1143679 (D′ = 0.97, r = 0.82; n = 1,815) (see ref.[4] and Supplementary Table 2, available on the Arthritis & Rheumatism web site at http://onlinelibrary.wiley.com/doi/10.1002/art.38117/abstract). Notably, in the present study, 86% of the donors carrying the rs1143679 variant also carried the nonsynonymous rs1143678 variant (Supplementary Table 2, available on the Arthritis & Rheumatism web site at http://onlinelibrary.wiley.com/doi/10.1002/art.38117/abstract). Strong LD between rs1143678 and rs1143679 has also been found in patients with SLE (see refs.[3] and[4] and Edberg JC: unpublished results). Remarkably, of the 21 donors who were homozygous for the variant (A) allele of SNP rs1143679, all carried at least 1 variant (T) allele of SNP rs1143678 (Supplementary Table 2, available on the Arthritis & Rheumatism web site at http://onlinelibrary.wiley.com/doi/10.1002/art.38117/abstract). Variation at rs1143680 was not observed in any of the donors used in our functional studies. These results underscore the notion that any functional effects attributed to SNP rs1143679 must take into account the potential role of SNP rs1143678 (and rs1143683, which was in perfect LD with rs1143678 in our donors).

Association of ITGAM variant alleles with differing capacity of neutrophils to phagocytose opsonized particles

To reveal the possible functional consequences of SLE-associated ITGAM variants, we examined Mac-1–mediated phagocytosis by neutrophils harvested from genotyped donors. Full-length ITGAM cDNA from all 31 donors used in the functional studies was prepared and sequenced, to confirm the allelic status at the SNP trio rs1143678/rs1143683/rs1143679 (Figure 1). In these 31 individuals, we did not detect any other nonsynonymous ITGAM genetic variation.

In cultures of fMLP-primed neutrophils from donors homozygous for the common alleles at rs1143678/rs1143683/rs1143679 (hereafter referred to as the common CC/CC/GG genotype) and those from donors homozygous for the respective minor alleles (hereafter referred to as the fully variant TT/TT/AA genotype) (see Figure 1), we observed a consistent and statistically significant decrease in phagocytosis of EAC and STZ (complement-opsonized particles that are known to engage the I-domain of Mac-1 [14–17]) by neutrophils from donors with the variant TT/TT/AA genotype of ITGAM (Figures 2A and B). In contrast, in neutrophils incubated with HTZ, a particle that can bind to Mac-1 via a lectin domain different from the complement-binding site ([17]) and via other cell surface receptors such as dectin-1 ([29]), phagocytosis was not significantly different between the fully common and fully variant genotypes (Figure 2C).

Figure 2.

Nonsynonymous ITGAM variants alter complement-mediated neutrophil phagocytosis. Quantitative phagocytosis by neutrophils from genotyped (rs1143678/rs1143683/rs1143679) healthy donors was assessed. Neutrophils were pretreated with 10−8M fMLP for 10 minutes, and then incubated with sheep erythrocyte antigen coated with complement (EAC) (A and D), serum-treated zymosan (STZ) (B and E), or heat-treated zymosan (HTZ) (C and F). All experiments were performed in a paired manner. The TT/TT/AA genotype was compared with the CC/CC/GG genotype (A–C) using 3 pairs of donors (donors 1–3 and donors 20–22). The TT/TT/GG genotype was compared with the CC/CC/GG genotype (D–F) using 5 pairs of donors (donors 6–10 and donors 24–28). Results are the mean ± SEM phagocytic index (P. I.) (the number of internalized probes per 100 neutrophils). ∗ = P < 0.05; ∗∗ = P < 0.01 by Student's paired t-test.

As noted earlier, in our cohort of 1,815 healthy donors, none were homozygous for the variant allele at rs1143679 (i.e., none had the CC/CC/AA genotype) (results available from the corresponding author upon request). Assessment of the influence of the extracellular SNP rs1143679 in isolation was, therefore, not possible. However, compared to neutrophils from donors with the common CC/CC/GG genotype, neutrophils from donors with the common genotype at rs1143679 but variant genotype at rs1143678/rs1143683 (i.e., a partially variant TT/TT/GG genotype) (see Figure 1) still showed a statistically significant defect in neutrophil phagocytosis of EAC and STZ (Figures 2D and E). Moreover, phagocytosis of HTZ was not affected (Figure 2F). These results showed that ITGAM variation in the cytoplasmic/calf-1 domain is sufficient to alter the biologic functions of neutrophils, even when there is no SNP variation in the β-propeller domain.

Importantly, whereas the phagocytosis of EAC required priming of neutrophils with fMLP, STZ and HTZ were internalized by unprimed neutrophils, albeit less efficiently. This allowed us to investigate the influence of ITGAM variation on phagocytosis using unprimed neutrophils. In these experiments, we again found that neutrophil phagocytosis of STZ, but not HTZ, was significantly reduced in cultures of neutrophils from donors with the fully variant TT/TT/AA genotype and in those from donors with the partially variant TT/TT/GG genotype (results available from the corresponding author upon request). In summary, these results demonstrate that ITGAM variation affects phagocytosis of complement-opsonized particles, such as EAC and STZ, but not the nonopsonized particle HTZ. Furthermore, these data are the first to show that, in addition to rs1143679, other SLE-associated nonsynonymous SNPs in ITGAM can alter the biologic functions of Mac-1.

Lack of effect of ITGAM variation on expression and activation of Mac-1.

Similar to other β2 integrins, the conformation of Mac-1 is normally very dynamic, with different conformations associated with different states of activation and affinity ([11, 30]). When neutrophils are at rest, most of the cell surface–expressed Mac-1 is in a constrained conformation with the major ligand-binding domain, the I-domain (binding site for ICAM-1 and iC3b, among others), buried within the protein's complex 3-dimensional structure. When neutrophils are activated with fMLP or phorbol myristate acetate, however, Mac-1 undergoes a conformational change that exposes the I-domain, and thus increases the receptor's affinity for its ligands ([31, 32]). Accordingly, it is possible that the expression of Mac-1 and/or the activation of its I-domain may be altered by ITGAM variation. If so, this could explain the observed changes in neutrophil phagocytosis.

However, our results showed that the expression of total CD11b (assessed using the monoclonal antibody ICFR44) was not different in neutrophils from donors with variant ITGAM alleles compared to those from donors with common ITGAM alleles (Figures 3A–C). Moreover, we also found that the activation of CD11b (assessed with the I-domain–specific monoclonal antibody CBRM1/5) did not differ by genotype (Figures 3D–F). Based on these findings, we conclude that the observed differences in neutrophil phagocytosis associated with ITGAM variation (Figure 2) could not be attributed to altered expression of Mac-1 or to changes in its activation state.

Figure 3.

Expression and activation of CD11b on neutrophils are not affected by ITGAM variation. Total CD11b expression and I-domain activation were measured on neutrophils from genotyped (rs1143678/rs1143683/rs1143679) healthy donors, using flow cytometry. A and D, Representative histograms of total CD11b expression (A) and CD11b I-domain expression (D) are shown. Red line represents neutrophils incubated with isotype control antibody, blue line represents neutrophils incubated with CD11b antibody without stimulus, and green line represents neutrophils incubated with CD11b antibody after 15 minutes of stimulation with 5 × 10−9M fMLP. B, C, E, and F, Anti-CD11b monoclonal antibodies ICRF44 and CBRM1/5 were used to measure total CD11b expression (B and C) and CD11b I-domain expression (E and F), respectively, in fMLP-stimulated and unstimulated cells. All experiments were performed in a paired manner, with 4 different pairs of donors (donors 1–4 and donors 20–23 in B and E; donors 6–9 and donors 24–27 in C and F). Results are the mean ± SEM values of mean fluorescence intensity (MFI) in donors with the TT/TT/AA genotype (B and E) or the TT/TT/GG genotype (C and F) compared to donors with the CC/CC/GG genotype. PE = phycoerythrin.

Alteration of Fc receptor function by ITGAM variation

Previous studies assessing cells from patients with leukocyte adhesion deficiency (in whom CD18 is not expressed, leading to lack of expression of β2 integrin) or transduced cells expressing various Fc receptors have demonstrated that full Fc receptor–mediated phagocytosis requires the presence of Mac-1 ([33, 34]). Interestingly, SLE patients have also been reported to have reduced Fc receptor functions ([35]). Therefore, we investigated whether ITGAM variation might affect Fc receptor–mediated phagocytosis of IgG antibody–opsonized erythrocytes (EA).

Surprisingly, we found that neutrophils from donors with the fully variant TT/TT/AA genotype and those from donors with the partially variant TT/TT/GG genotype displayed significantly reduced phagocytosis of EA when compared to neutrophils from controls with the common CC/CC/GG genotype, under both fMLP-primed and unprimed conditions (Figures 4A–D). Thus, these results are the first to show that SNP variations in ITGAM will render neutrophils less able to phagocytose IgG-coated particles via Fc receptors.

Figure 4.

Nonsynonymous variants in ITGAM alter Fcγ receptor–mediated neutrophil phagocytosis. Quantitative phagocytosis by neutrophils from genotyped (rs1143678/rs1143683/rs1143679) healthy donors was assessed using sheep erythrocyte antigen coated with rabbit anti-sheep erythrocyte IgG antibodies (EA). Neutrophils without priming (A and C) or after 10 minutes of priming with 10−8M fMLP (B and D) were used. All experiments were performed in a paired manner. The TT/TT/AA genotype was compared with the CC/CC/GG genotype (A and B) using 3 pairs of donors (donors 1–3 and donors 20–22). The TT/TT/GG genotype was compared with the CC/CC/GG genotype (C and D) using 5 pairs of donors (donors 6–10 and donors 24–28). Results are the mean ± SEM phagocytic index (P. I.) (the number of internalized probes per 100 neutrophils). ∗ = P < 0.05; ∗∗ = P < 0.01 by Student's paired t-test

Alteration of neutrophil firm adhesion by ITGAM variation

To investigate whether ITGAM variation alters additional Mac-1–dependent functions that could increase the susceptibility to SLE, we compared firm adhesion of neutrophils isolated from donors with different ITGAM genotypes. In the first series of experiments, we used an in vitro flow chamber–based assay to analyze neutrophil firm adhesion to the purified Mac-1 ligand ICAM-1. P-selectin was also included on the surface to initiate cell capture and facilitate cell rolling and subsequent firm adhesion to ICAM-1 ([26]). We found that neutrophil firm adherence was significantly reduced in cell cultures derived from donors with the fully variant TT/TT/AA genotype and those with the partially variant TT/TT/GG genotype compared to donors with the common genotype (Figures 5A and B).

Figure 5.

Nonsynonymous ITGAM variants alter neutrophil firm adhesion under flow conditions. Firm adhesion under flow conditions was assessed using neutrophils from genotyped (rs1143678/rs1143683/rs1143679) healthy donors after 10 minutes of pretreatment with 10−8M fMLP, followed by incubation with intercellular adhesion molecule 1 (ICAM-1)/P-selectin (A and B) or human umbilical vein endothelial cells (HUVECs) (C and D) coupled in a flow chamber. Firm adhesion was defined as movement of the neutrophil a distance of <1 cell diameter for 5 seconds. The number of adherent cells/minute was calculated from a 4-minute video. All experiments were performed in a paired manner. For ICAM-1 adhesion, 3 different pairs of donors (donors 2, 3, 5, 20, 30, and 31 in A and donors 6–8, 21, 24, and 29 in B) were used. For HUVEC adhesion, 4 different pairs of donors (donors 1–4, 21, 23, 26, and 31 in C and donors 6–9, 23, 25, 26, and 28 in D) were used. Results are the mean ± SEM of adherent cells/minute from a minimum of 3 independent videos recorded for each donor. ∗ = P < 0.05; ∗∗ = P < 0.01 by Student's paired t-test.

We next analyzed neutrophil adhesion to TNFα-stimulated HUVECs. In this series of experiments, we observed the same genotype-associated reduction in neutrophil adhesion (Figures 5C and D). These results demonstrated that ITGAM variations have functional consequences in neutrophils, including the impairment of Mac-1–mediated firm adhesion to both purified and natural ligands under conditions of shear stress.

Independent inhibition of Mac-1 function on neutrophils by SNP rs1143679.

Our analysis of donors who were homozygous for ITGAM variant alleles demonstrated that the rs1143678/rs1143683 variants can independently alter Mac-1 functions in human primary neutrophils. To directly address the ability of the rs1143679 ITGAM variation to alter Mac-1 function, we studied neutrophils from donors heterozygous at rs1143679 but homozygous for the common alleles at rs1143678/rs1143683 (i.e., the partially variant CC/CC/GA genotype [see Figure 1]). Neutrophils from these donors displayed significantly reduced phagocytosis of STZ to neutrophils from donors with the common ITGAM genotype (CC/CC/GG) (Figure 6A).

Figure 6.

The ITGAM single-nucleotide polymorphism rs1143679 variant independently alters neutrophil phagocytosis. Quantitative phagocytosis by neutrophils from genotyped (rs1143678/rs1143683/rs1143679) healthy donors was assessed using serum-treated zymosan (STZ) (A and B) or heat-treated zymosan (HTZ) (C and D). Neutrophils without priming (fMLP−) or after 10 minutes of priming with 10−8M fMLP (fMLP+) were used. All experiments were performed in a paired manner. For comparisons of donors with the CC/CC/GA genotype and those with the CC/CC/GG genotype (A and C) or donors with the CT/CT/GA genotype and those with the CT/CT/GG genotype (B and D), 3 pairs of donors (donors 11–20, 23, and 26) were assessed. Variations are shown in boldface. Results are the mean ± SEM phagocytic index (P. I.) (number of internalized probes per 100 neutrophils). ∗ = P < 0.005 by analysis of variance.

Further supporting the notion of a functional role of variation at rs1143679, we also observed a significant reduction in neutrophil phagocytosis of STZ in donors heterozygous for all 3 SNPs (the partially variant CT/CT/GA genotype [see Figure 1]), when compared to donors heterozygous at rs1143678/rs1143683 alone (the partially variant CT/CT/GG genotype [see Figure 1]) (Figure 6B). In contrast, none of the ITGAM variant genotypes affected phagocytosis of HTZ (Figures 6C and D), which is consistent with the results reported above (Figures 2C and F). These results demonstrated that SNP variation at rs1143679 has an independent impact on Mac-1 function in primary neutrophils.

DISCUSSION

Our study, like others before it ([21-23]), demonstrates that ITGAM variation alters leukocyte phagocytosis of complement-opsonized particles. In addition, we have extended the observation that ITGAM variation has impact on other biologically important neutrophil functions, i.e., firm adhesion under shear stress and Fc receptor–mediated phagocytosis. Importantly, we have shown that these functional alterations can be the consequence of multiple SLE-associated nonsynonymous SNPs, not just solely at rs1143679. Indeed, it is likely that these multiple ITGAM SNPs co-segregate as a haplotype and act in concert to affect Mac-1–mediated functions, and thereby modify the risk of SLE development.

Among the SLE-associated nonsynonymous SNPs studied herein, SNPs rs1143679 and rs1143678/rs1143683 independently alter Mac-1–mediated neutrophil functions. Our genotyping results suggest that rs1143679 is in strong LD with rs1143678/rs1143683, and that 86% of carriers of the rs1143679 variant allele also carry an rs1143678 variant allele. This strong LD has been confirmed in another large cohort, which comprised >10,000 SLE cases and controls (Edberg JC: unpublished results). We acknowledge that ITGAM genetic association studies have shown different strengths of association between ITGAM SNPs and SLE in different ethnic groups, perhaps because the LD patterns and MAF values may differ across ethnicities. Nevertheless, our results should serve as a reminder that a full understanding of the reasons for the genetic association between the ITGAM locus and SLE will require studies of multiple ITGAM SNPs, not single ones.

Similar to the findings in human monocytes and transfected cell lines ([21-23]), our results in human neutrophils demonstrated that neither total Mac-1 receptor expression nor activation of its major ligand-binding site I-domain is affected by ITGAM variation. Thus, Mac-1 expression or I-domain exposure is unlikely to be the mechanism by which neutrophil phagocytosis and adhesion is reduced by ITGAM variation. The caveat is that the monoclonal antibody CBRM1/5 may not detect subtle changes in the I-domain, which might nevertheless occur as a consequence of ITGAM variation. This is especially true for the rs1143679 SNP, which is located in the extracellular domain of CD11b and proximal to the I-domain. If this SNP does alter I-domain structure, it is likely to also affect ligand-binding affinity ([3]).

In contrast, the SNP rs1143683 is not as likely to alter the structure of Mac-1, since this SNP introduces a conservative amino acid change (Ala-to-Val) in the extracellular calf-1 region of CD11b. Furthermore, although cation-binding sites in the thigh/calf region of integrins are important for its proper activation ([28]), SNP rs1143683 is not within those sites. Thus, the association between SNP rs1143683 and SLE likely is a reflection of its perfect LD with SNP rs1143678.

Importantly, SNP rs1143678 resides in the cytoplasmic tail of CD11b, which is known to be involved in Mac-1 signaling ([36]), and is likely to be important in integrin–cytoskeleton interactions and, thus, could play a role in a variety of integrin functions ([37-39]). The Pro-to-Ser amino acid change introduced by SNP rs1143678 could potentially affect the structure of the cytoplasmic tail, and thus impact its signaling or interaction with cytoskeleton proteins. Indeed, our results demonstrate that there is a functional consequence of variation at SNP rs1143678, and that this variation alone is sufficient to significantly inhibit neutrophil adhesion, Mac-1–mediated phagocytosis, and Fc receptor–mediated phagocytosis. This SNP has been overlooked or discounted in previous ITGAM functional studies. The fact that both rs1143679 and rs1143678/rs1143683 appear to independently alter Mac-1 functions suggests that multiple mechanisms of functional alterations are involved. All of these hypotheses need further verification.

In this study, we demonstrated that the function of neutrophils from healthy donors carrying SLE-associated ITGAM risk alleles is impaired. This finding is consistent with the findings from a previous study using CD11b-deficient mice, the results of which showed loss of Mac-1–increased autoimmune disease severity in the MRL/MpJ-Faslpr mouse strain ([7]). Our data also suggest that the reduction in Fc receptor function that has been observed in SLE patients ([35]) could be partly attributed to ITGAM variants, thus extending the knowledge regarding the contributions of these genetic variants to both complement and Fc receptor functions. Although our current study does not address the mechanism by which ITGAM variants alter Fc receptor function, prior work has demonstrated physical and functional interactions between Fcγ receptor IIIb and Mac-1 ([25, 40]). Moreover, our study does not address the physiologic ligands that interact with Mac-1 to manifest diminished receptor function related to the SLE-associated ITGAM variants. Despite these caveats, our results demonstrate that a deficiency in CD11b function is associated with development of an autoimmune phenotype.

The ITGAM variation–associated reduction in Mac-1–mediated phagocytosis (Figure 2 and results available from the corresponding author upon request) ([21, 22]) and Fc receptor–mediated phagocytosis (Figure 3) ([35]) could contribute to altered immune complex clearance and deposition. Indeed, impaired clearance of immune complexes has been observed in patients with SLE ([35, 41-43]), and it is tempting to speculate that ITGAM variation may provide one mechanism for this deficiency. It is also possible that reduced adhesion resulting from ITGAM variation could affect normal leukocyte trafficking in vivo, and thus may contribute to the pathogenesis of SLE. Other Mac-1 functions not probed in our study could also be altered by ITGAM variants and contribute to the pathogenesis of SLE ([20]).

To our knowledge, this study is the first to focus on the biologic impact of multiple disease-associated genetic variants in the ITGAM locus on neutrophils, a cell type that plays an important role in SLE ([44]). Although prior studies have investigated the impact of ITGAM variation on Mac-1–mediated adhesion, this study is the first to examine this question in terms of the impact on the more physiologically common biologic function of neutrophil adhesion, under conditions of shear stress, to both purified ligand (ICAM-1) and endothelial cells. Future functional studies of the ITGAM locus are needed to discern the potential role of other nonsynonymous variants yet to be identified at this locus.

AUTHOR CONTRIBUTIONS

All authors were involved in drafting the article or revising it critically for important intellectual content, and all authors approved the final version to be published. Dr. Edberg had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Study conception and design. Zhou, Wu, Kucik, Szalai, Bullard, Edberg.

Acquisition of data. Zhou, Kucik, White, Szalai.

Analysis and interpretation of data. Zhou, Kucik, Redden, Szalai, Bullard, Edberg.

Acknowledgments

We thank Dr. Robert P. Kimberly for helpful discussions and support. We also thank Dr. Carl Langefeld for advice on the statistical analysis, Mark McCrory, Deborrah McDuffie, and Ellen Sowell for technical assistance, and Dr. Susan Bellis for helpful discussions.

Ancillary