Abstract

Integrin alpha M (ITGAM; CD11b) is a component of the macrophage-1 antigen complex, which mediates leukocyte adhesion, migration and phagocytosis as part of the immune system. We previously identified a missense polymorphism, rs1143679 (R77H), strongly associated with systemic lupus erythematosus (SLE). However, the molecular mechanisms of this variant are incompletely understood. A meta-analysis of published and novel data on 28 439 individuals with European, African, Hispanic and Asian ancestries reinforces genetic association between rs1143679 and SLE [Pmeta = 3.60 × 10−90, odds ratio (OR) = 1.76]. Since rs1143679 is in the most active region of chromatin regulation and transcription factor binding in ITGAM, we quantitated ITGAM RNA and surface protein levels in monocytes from patients with each rs1143679 genotype. We observed that transcript levels significantly decreased for the risk allele (‘A’) relative to the non-risk allele (‘G’), in a dose-dependent fashion: (‘AA’ < ‘AG’ < ‘GG’). CD11b protein levels in patients' monocytes were directly correlated with RNA levels. Strikingly, heterozygous individuals express much lower (average 10- to 15-fold reduction) amounts of the ‘A’ transcript than ‘G’ transcript. We found that the non-risk sequence surrounding rs1143679 exhibits transcriptional enhancer activity in vivo and binds to Ku70/80, NFKB1 and EBF1 in vitro, functions that are significantly reduced with the risk allele. Mutant CD11b protein shows significantly reduced binding to fibrinogen and vitronectin, relative to non-risk, both in purified protein and in cellular models. This two-pronged contribution (nucleic acid- and protein-level) of the rs1143679 risk allele to decreasing ITGAM activity provides insight into the molecular mechanisms of its potent association with SLE.

INTRODUCTION

Systemic lupus erythematosus (SLE or lupus) is a complex, multiorgan, autoimmune disease with significant morbidity and mortality. SLE has a strong genetic basis. To date, over 40 genetic associations (genes/loci) have been identified (P < 5 × 10−8) through genome-wide and candidate-gene association studies. However, little is known about possible molecular mechanisms through which associated variants contribute to disease. We identified a novel genetic variant, rs1143679, in exon-3 of  ITGAM (1), and confirmed that it is the only polymorphism that explains the observed ITGAM association with SLE (2–5). While this association is robust across most populations studied with European, African, Hispanic or Native American origin, the rs1143679 risk variant is absent or very rare in many East Asian populations (2,6); thus it deserves further exploration in Asian populations. ITGAM has been associated with SLE and systemic sclerosis (a skin-affecting autoimmune disease), but not with other autoimmune diseases (7,8). The rs1143679 risk allele ‘A’ has also been linked to specific SLE clinical subphenotypes, including renal disease, discoid rash and immunologic manifestations (9).

Missense mutation of rs1143679 changes amino acid arginine (R) to histidine (H) at position 77 (R77H) of the CD11b protein. This transmembrane glycoprotein is an integrin adhesion molecule mainly expressed in neutrophils, monocytes, macrophages and dendritic cells. Together with CD18 (integrin beta 2; ITGB2), CD11b forms the macrophage-1 antigen complex (Mac-1, or complement receptor 3, CR3). Mac-1 is involved in numerous trafficking and adherence functions in monocytes and neutrophils, including binding to stimulated endothelium, intravascular aggregation and signaling of complement-coated particles. The amino acid change is in the β-propeller domain of CD11b near the ‘I’ domain, potentially altering protein conformation and affecting key cell surface ligand interactions and other cellular functions (10–12).

We investigated the molecular mechanisms by which the rs1143679 risk allele alters gene/protein functions and contributes to SLE pathogenesis. We found significant differences between RNA and surface protein expression levels in monocytes from SLE patients with either the homozygous protective or homozygous risk genotype. Using allelic expression assays, we confirmed that reduced RNA expression is specific to the risk allele and is not attributed to aberrant splicing or degradation, but rather to the loss of specific and strong transcriptional enhancer activity. We also show that cells that stably express the CD11b risk allele bind Mac-1 ligands fibrinogen (FBN) and vitronectin (VTN) less efficiently than those expressing the wild-type allele. Recombinant protein studies confirm that this is because of reduced affinity of CD11b for the ligands and not to downstream signal transduction events. These results implicate a multifaceted risk allele-specific alteration of ITGAM function at both the RNA and protein expression levels, as well as interactions of the resulting expressed protein. The combination of these effects helps to explain the strong statistical association of this SNP with SLE.

RESULTS

Meta-analysis of published and novel data

Because the rs1143679 risk allele is absent or very rare in many Asian populations, our meta-analysis included new data from three Asian populations (Indian, Malayan and Chinese) alongside published data from European-derived, West African-admixed, Asian and Hispanic populations (19 countries, 27 independent data sets, N = 28 439) (1,2,6,13–17). Our results greatly reinforce ITGAM-SLE association [Pmeta = 3.60 × 10−90, odds ratio (OR) (95% CI) = 1.76 (1.67–1.86); Fig. 1]. Minor-allele frequencies in controls varied across populations (Table 1): from absent in Korean and Japanese to 0.51% in Chinese to 19% in Portuguese. However, tests for OR heterogeneity were not significant (Q-statistic, PHET = 0.43). Region-specific meta-analyses were also significant (European: P = 2.22 × 10−27, OR = 1.78; European-American: P = 1.82 × 10−35, OR = 1.79; African-American: P = 3.81 × 10−12, OR = 1.64; Hispanic: P = 5.88 × 10−14, OR = 1.83; East Asian: P = 1.38 × 10−7, OR = 2.60; Table 2). A Cochran–Mantel–Haenszel (CMH) test demonstrated that there was no population stratification within each ethnic/population subgroup (minimum P = 0.33). However, the CMH test across all populations was significant (P = 0.008). Therefore, we also performed a meta-analysis using a random effect model, with similar results to the fixed effect model [P = 3.05 × 10−83, OR = 1.76 (1.67–1.86)].

Table 1.

Populations from published reports of SLE-rs1143679 association

Ethnicity/region Country/population Case/control Case MAF Control
 
MAF P-value OR 95% CI References 
European Greece 191/186 0.26 0.19 1.30E−2 1.55 1.10–2.19 Gestal et al., 2009 
Portugal 94/95 0.26 0.19 1.58E−1 1.42 0.87–2.31 Gestal et al., 2009 
Spain 523/569 0.26 0.15 2.68E−9 1.91 1.54–2.37 Gestal et al., 2009 
Italy 292/316 0.25 0.17 7.85E−4 1.61 1.22–2.13 Gestal et al., 2009 
Netherlands 104/180 0.21 0.14 3.98E−2 1.59 1.02–2.48 Gestal et al., 2009 
Hungary 95/95 0.17 0.10 5.04E−2 1.82 0.99–3.35 Gestal et al., 2009 
Slovakia 94/93 0.12 0.11 7.71E−1 1.10 0.58–2.09 Gestal et al., 2009 
Czech-rep 101/99 0.13 0.11 4.92E−1 1.23 0.68–2.25 Gestal et al., 2009 
Germany 82/92 0.21 0.09 2.50E−3 2.57 1.37–4.80 Gestal et al., 2009 
Polish 154/276 0.19 0.12 7.90E−3 1.71 1.16–2.52 Warchol et al., 2011 
UK 445/528 0.16 0.09 6.20E−8 2.10 1.60–2.76 Han et al., 2009 
Finland 275/356 0.19 0.10 8.29E−6 2.13 1.54–2.94 Jarvinen et al., 2010 
European American EA-I 1914/1901 0.17 0.10 1.30E−17 1.78 1.56–2.04 Nath et al., 2008 
EA-II 736/1051 0.18 0.11 5.71E−9 1.75 1.45–2.12 Han et al., 2009 
EA-III 1183/1287 0.17 0.11 2.78E−12 1.72 1.46–2.02 Hughes et al., 2011 
East Asian Hong Kong 910/2360 0.011 0.005 2.10E−2 2.28 1.27–4.11 Yang et al., 2009 
Thai 278/383 0.06 0.02 4.10E−4 3.15 1.72–5.75 Yang et al., 2009 
Korean 661/781 Mono Mono – – – Han et al., 2009 
Japanese 176/365 Mono Mono – – – Han et al., 2009 
Chinesea 365/292 0.03 0.01 3.72E−2 2.57 1.02–6.49 present study 
Malaya 76/76 0.10 0.03 1.90E−2 3.27 1.16–9.23 present study 
South Asian Indiaa 239/248 0.19 0.09 8.16E−6 2.36 1.60–3.47 present study 
West African Derived AA-I 588/701 0.15 0.11 4.00E−4 1.52 1.20–1.92 Nath et al., 2008 
AA-II 1029/1758 0.16 0.10 2.94E−8 1.57 1.34–1.85 Kim et al., 2012 
Gullah 137/134 0.20 0.11 2.90E−3 2.07 1.27–3.36 Nath et al., 2008 
Latin American Hispanic-I 692/230 0.16 0.09 9.95E−5 1.99 1.40–2.82 Han et al., 2009 
Hispanic-II 961/336 0.16 0.11 1.20E−3 1.56 1.19–2.05 Kim et al., 2012 
Mexican 392/281 0.16 0.10 2.30E−3 2.01 1.44–2.81 Han et al., 2009 
Colombia 205/378 0.22 0.11 3.64E−7 1.57 1.08–2.29 Han et al., 2009 
Ethnicity/region Country/population Case/control Case MAF Control
 
MAF P-value OR 95% CI References 
European Greece 191/186 0.26 0.19 1.30E−2 1.55 1.10–2.19 Gestal et al., 2009 
Portugal 94/95 0.26 0.19 1.58E−1 1.42 0.87–2.31 Gestal et al., 2009 
Spain 523/569 0.26 0.15 2.68E−9 1.91 1.54–2.37 Gestal et al., 2009 
Italy 292/316 0.25 0.17 7.85E−4 1.61 1.22–2.13 Gestal et al., 2009 
Netherlands 104/180 0.21 0.14 3.98E−2 1.59 1.02–2.48 Gestal et al., 2009 
Hungary 95/95 0.17 0.10 5.04E−2 1.82 0.99–3.35 Gestal et al., 2009 
Slovakia 94/93 0.12 0.11 7.71E−1 1.10 0.58–2.09 Gestal et al., 2009 
Czech-rep 101/99 0.13 0.11 4.92E−1 1.23 0.68–2.25 Gestal et al., 2009 
Germany 82/92 0.21 0.09 2.50E−3 2.57 1.37–4.80 Gestal et al., 2009 
Polish 154/276 0.19 0.12 7.90E−3 1.71 1.16–2.52 Warchol et al., 2011 
UK 445/528 0.16 0.09 6.20E−8 2.10 1.60–2.76 Han et al., 2009 
Finland 275/356 0.19 0.10 8.29E−6 2.13 1.54–2.94 Jarvinen et al., 2010 
European American EA-I 1914/1901 0.17 0.10 1.30E−17 1.78 1.56–2.04 Nath et al., 2008 
EA-II 736/1051 0.18 0.11 5.71E−9 1.75 1.45–2.12 Han et al., 2009 
EA-III 1183/1287 0.17 0.11 2.78E−12 1.72 1.46–2.02 Hughes et al., 2011 
East Asian Hong Kong 910/2360 0.011 0.005 2.10E−2 2.28 1.27–4.11 Yang et al., 2009 
Thai 278/383 0.06 0.02 4.10E−4 3.15 1.72–5.75 Yang et al., 2009 
Korean 661/781 Mono Mono – – – Han et al., 2009 
Japanese 176/365 Mono Mono – – – Han et al., 2009 
Chinesea 365/292 0.03 0.01 3.72E−2 2.57 1.02–6.49 present study 
Malaya 76/76 0.10 0.03 1.90E−2 3.27 1.16–9.23 present study 
South Asian Indiaa 239/248 0.19 0.09 8.16E−6 2.36 1.60–3.47 present study 
West African Derived AA-I 588/701 0.15 0.11 4.00E−4 1.52 1.20–1.92 Nath et al., 2008 
AA-II 1029/1758 0.16 0.10 2.94E−8 1.57 1.34–1.85 Kim et al., 2012 
Gullah 137/134 0.20 0.11 2.90E−3 2.07 1.27–3.36 Nath et al., 2008 
Latin American Hispanic-I 692/230 0.16 0.09 9.95E−5 1.99 1.40–2.82 Han et al., 2009 
Hispanic-II 961/336 0.16 0.11 1.20E−3 1.56 1.19–2.05 Kim et al., 2012 
Mexican 392/281 0.16 0.10 2.30E−3 2.01 1.44–2.81 Han et al., 2009 
Colombia 205/378 0.22 0.11 3.64E−7 1.57 1.08–2.29 Han et al., 2009 

aUnpublished data.

Table 2.

Region and overall meta-analysis including published and unpublished data from Table 1

Region Case/control P-value OR 95% CI χ2-het het P-value 
European 2450/2884 2.22E−27 1.78 1.60–1.97 8.74 0.65 
European-American 3833/4239 1.82E−35 1.79 1.63–1.96 0.45 0.80 
West African-derived 1754/2593 3.81E−12 1.64 1.43–1.89 1.32 0.52 
Hispanic 2250/1225 5.88E−14 1.83 1.56–2.15 3.40 0.33 
East Asian 1629/3111 1.38E−7 2.60 1.76–3.83 0.73 0.70 
South Asian 294/298 1.10E−6 2.50 1.73–3.62 NA NA 
Total samplesa 12,155/14,300 3.60E−90 1.76 1.67–1.86 26.57 0.43 
Region Case/control P-value OR 95% CI χ2-het het P-value 
European 2450/2884 2.22E−27 1.78 1.60–1.97 8.74 0.65 
European-American 3833/4239 1.82E−35 1.79 1.63–1.96 0.45 0.80 
West African-derived 1754/2593 3.81E−12 1.64 1.43–1.89 1.32 0.52 
Hispanic 2250/1225 5.88E−14 1.83 1.56–2.15 3.40 0.33 
East Asian 1629/3111 1.38E−7 2.60 1.76–3.83 0.73 0.70 
South Asian 294/298 1.10E−6 2.50 1.73–3.62 NA NA 
Total samplesa 12,155/14,300 3.60E−90 1.76 1.67–1.86 26.57 0.43 

aDoes not include Korean (661/791) or Japanese (176/365) samples who were monomorphic for rs1143679.

Figure 1.

Meta-analysis of the rs1143679 using published data and new data from multiple ethnically diverse populations.

Figure 1.

Meta-analysis of the rs1143679 using published data and new data from multiple ethnically diverse populations.

The rs1143679 locus is an active chromatin region

Analysis of ENCODE Project data for ITGAM in GM12878 cells (a lymphoblastoid cell line closely related to monocytes) revealed that rs1143679 shows the greatest density of H3K27Ac ‘active histone’ marking (18,19) within the gene and flanking untranslated regions; significantly more than the promoters of the ITGAM or adjacent ITGAX genes (Supplementary Material, Fig. S1). Similarly, DNAse I hypersensitivity suggested open chromatin, and ENCODE data annotated (Hidden Markov Model) chr16: 31 276 300–31 277 700 (rs1143679 is at 31 276 811) as a strong enhancer in GM12878 cells (Supplementary Material, Fig. S1). ChIP-Seq experiments showed binding of a number of transcription factors linked to immune cell function, including EBF1 (early B-cell factor), which controls lymphocyte development; PAX5 and NFKB1, activator proteins specific to B-cells; JunD, a determinant of macrophage activation (20); TCF12 (21), BCL11a (22) and BATF (23), which coordinate B- and T-cell maturation; and SP1, which drives interleukin-21 receptor activation (24) (Supplementary Material, Fig. S2). In addition, the region containing rs1143679 (the base itself is part of a CpG motif) is highly onserved among primates, consistent with a functional regulatory role.

ITGAM RNA and surface CD11b expression are reduced in individuals with risk genotypes

We isolated genotype-specific [‘AA’ (n = 9), ‘GA’ (n = 24), ‘GG’ (n = 20)] monocytes from peripheral blood mononuclear cells (PBMCs) of 53 SLE patients (patient information in Supplementary Material, Table S1) and 3 controls (all female; genotype, GA) by FACS with fluorescently labeled anti-CD14 antibody (Supplementary Material, Fig. S3) and determined levels of RNA by Reverse Transcription-Quantitative Polymerase Chain Reaction (RT–qPCR). Surface-displayed CD11b protein was analyzed with anti-CD14 and anti-CD11b antibodies (Supplementary Material, Fig. S4). We found reduced levels of both transcript and surface-displayed protein [mean fluorescence intensity (MFI) of CD11b in CD14+/Cd11b+ cells; Fig. 2A and B] in individuals with ‘AA’ risk alleles compared with individuals with ‘GG’ protective alleles (RNA: 3.4-fold, P = 2.1 × 10−6; protein: 2.2-fold, P = 2.3 × 10−4). Further, these levels were dose-dependent on the number of ‘G’ alleles (‘AA’ versus ‘AG’: 2.5-fold lower RNA, P = 2.0 × 10−4; 1.8-fold lower protein, P = 7.8 × 10−3). These dose-dependent level changes are consistent with a decrease specific to risk allele-containing transcripts, either through transcriptional repression or through degradation. Similar genotype-dependent differences in surface protein expression were observed for the MFI of CD11b in total CD14+ cells (‘GG’ versus ‘AA’: 2.2-fold, P = 4.6 × 10−4) and the frequency of CD14+/CD11b+ cells among CD14+ cells (‘GG’ versus ‘AA’: 1.4-fold, P = 7.8 × 10−6) (Supplementary Material, Fig. S5).

Figure 2.

ITGAM RNA and CD11b protein expression in monocytes from ‘GG’, ‘GA’ and ‘AA’ patients. (A) RT–qPCR analysis suggests that ‘GG’ patients have significantly higher levels of RNA than ‘AA’ patients, with ‘GA’ patients having intermediate levels. (B) Surface CD11b protein expression was reduced in ‘AA’ patients compared with ‘GG’ patients, with ‘GA’ present at intermediate levels. (C) Risk allele ‘A’ carrying clones are reduced in monocytes from ‘GA’ individuals, suggesting that some individuals express little or no ‘A’ transcript. (D) Average number of clones carrying the ‘G’ and ‘A’ alleles.

Figure 2.

ITGAM RNA and CD11b protein expression in monocytes from ‘GG’, ‘GA’ and ‘AA’ patients. (A) RT–qPCR analysis suggests that ‘GG’ patients have significantly higher levels of RNA than ‘AA’ patients, with ‘GA’ patients having intermediate levels. (B) Surface CD11b protein expression was reduced in ‘AA’ patients compared with ‘GG’ patients, with ‘GA’ present at intermediate levels. (C) Risk allele ‘A’ carrying clones are reduced in monocytes from ‘GA’ individuals, suggesting that some individuals express little or no ‘A’ transcript. (D) Average number of clones carrying the ‘G’ and ‘A’ alleles.

Risk allele RNA is specifically reduced in heterozygous individuals

We explored the potential of differential regulation between ‘A’ and ‘G’ sequences by amplifying rs1143679 sequences from cDNA with two neighboring exonic primers, cloning and sequencing 434 ITGAM transcripts from 13 female heterozygous (‘GA’) individuals (11 SLE patients and 2 healthy controls). For a typical somatic gene, transcripts should be equally abundant for ‘A’ and ‘G’ alleles, barring differential regulation of the two sequences. Instead, we found an average ∼7-fold increase of ‘G’ transcripts over ‘A’ transcripts (385 ‘G’ versus 49 ‘A’; P = 1.3 × 10−11) (Fig. 2C and D). Eight of 13 individuals had very low levels of ‘A’ allele transcripts (G : A = 68 : 1) and for 6 of these individuals ‘A’ allele-carrying transcripts were completely absent (∼35 total transcripts sequenced from each individual). Sequence alignments of these individuals' genomic DNA (data not shown) at this region confirmed that there were no additional polymorphisms within either primer region, ruling out any ‘G’ allele-specific preferential amplification. Rather, overrepresentation of the ‘G’ allele-carrying RNA in heterozygous individuals' monocytes suggests that risk allele RNA is actively suppressed.

In a separate assay, sequencing-based allelic expression assays were performed for eight additional heterozygous individuals (7 female SLE cases and 1 healthy control). Although three of them (including the healthy control) showed comparable G : A levels, the other 5 had very few or no ‘A’ transcripts in the total cDNA pool (Supplementary Material, Fig. S6). The three individuals with higher A : G ratio (1 control, Patients 13 and 15) also showed higher surface expression of CD11b than the remaining heterozygous patients (individual expression data not shown, but included in Fig. 2A and B). Thus, using two independent methods, we analyzed 21 heterozygous individuals for allele-specific expression. Risk allele (‘A’) transcripts are lower than protective ‘G’ transcripts in 18/21 individuals; in 13/21 individuals, they are extremely low or completely absent.

Reduced transcript levels are not due to risk allele-specific splicing defects

rs1143679 lies 9 bases from the exon-3/intron-3 splice donor site and is predicted to be in an exonic splicing enhancer sequence (http://manticore.niehs.nih.gov/cgi-bin/snpinfo/splice.cgi?2_rs1143679). Thus the risk allele ‘A’ may impair normal RNA splicing. We used monocyte cDNA from 3 ‘AA’ patients for PCR with exonic primers located in exon-3 and exon-5 (Supplementary Material, Fig. S7) to identify any non-spliced pre-mRNA carrying intronic sequences. Stringent PCR conditions produced only fully spliced products (245 bp) (Fig. 3A). Non-stringent PCR conditions amplified several intermediate products (between 245 and 705 bp) (Fig. 3B), but cloning and sequencing these did not reveal any ITGAM sequences. Thus rs1143679 does not appear to affect proper maturation of the exon-3/intron-3 splice junction in producing mature ITGAM transcripts.

Figure 3.

Splicing and degradation analysis. (A) RT–PCR produces fully spliced 245 bp products. PCR for genomic DNA produces the expected 705 bp band. Three homozygous (‘AA’) patients' monocyte cDNA (i.e., P1, P2, P3). (B) cDNAs from the same ‘AA’ patients of (A) with 4 additional patients were used for RT–PCR under non-stringent PCR conditions. Intermediate products (b, c, d, e) between genomic (a) and fully spliced (f) products did not correspond to unspliced ITGAM transcripts. (C) Time course of RNA measurements on CD11b+ flow sorted cells. Levels of RNA were similar for all time courses for each sample. (D) Luciferase activity was driven by the rs1143679 locus in HeLa cells. (E) Luciferase activity was driven by the rs1143679 locus in MonoMac-6 cells. In both cells, the ‘G’ sequence showed more activity than the ‘A’ sequence. MCS, mGL.1 vector carrying multiple cloning site MCS and luciferase gene but no promoter; Tkmin, mGL.1 with Tk minimal promoter with luciferase gene G, mGL.1-TKmin with G (nonrisk allele); A, mGL.1-TKmin with A (risk allele).

Figure 3.

Splicing and degradation analysis. (A) RT–PCR produces fully spliced 245 bp products. PCR for genomic DNA produces the expected 705 bp band. Three homozygous (‘AA’) patients' monocyte cDNA (i.e., P1, P2, P3). (B) cDNAs from the same ‘AA’ patients of (A) with 4 additional patients were used for RT–PCR under non-stringent PCR conditions. Intermediate products (b, c, d, e) between genomic (a) and fully spliced (f) products did not correspond to unspliced ITGAM transcripts. (C) Time course of RNA measurements on CD11b+ flow sorted cells. Levels of RNA were similar for all time courses for each sample. (D) Luciferase activity was driven by the rs1143679 locus in HeLa cells. (E) Luciferase activity was driven by the rs1143679 locus in MonoMac-6 cells. In both cells, the ‘G’ sequence showed more activity than the ‘A’ sequence. MCS, mGL.1 vector carrying multiple cloning site MCS and luciferase gene but no promoter; Tkmin, mGL.1 with Tk minimal promoter with luciferase gene G, mGL.1-TKmin with G (nonrisk allele); A, mGL.1-TKmin with A (risk allele).

Reduced transcript level is not due to risk allele-specific RNA degradation

To determine whether risk allele-carrying RNA was targeted for degradation after transcription, we generated stable cells for both full-length protective and risk alleles of ITGAM under control of the ferritin promoter along with ITGB2 (CD18) coding sequence in a dual expression vector (see Materials and Methods, Supplementary Material, Fig. S8). We hypothesized that the construct would be transcribed at high levels in both cases. This allowed us to observe other potential sources of regulation such as risk allele-specific RNA degradation, which should manifest as a lower steady-state transcript level. We flow sorted (FACS) and cultured cells that stably expressed either allele of CD11b along with CD18. Because the CD11b RNA half-life is unknown we collected cells at both 2 and 24 h intervals, and RT–qPCR was performed on total RNA. RNA levels were similar at both collection times in R77 (non-risk) and H77 (risk) cells (Fig. 3C), showing a slight increase over time. This may be attributable to increased growth rate following FACS isolation.

To assess degradation of ITGAM RNA in primary cells, we also determined RNA levels in PBMCs from 4 non-risk (‘GG’) homozygous, 2 heterozygous (‘GA’) and 2 risk (‘AA’) homozygous patients at two time points, 0 and 16 h. RNA levels did not significantly change in any patient, suggesting that the risk allele does not affect RNA stability (Supplementary Material, Fig. S9).

The rs1143679 locus contains a strong transcriptional enhancer damaged by the rs1143679 risk allele

As discussed above, the region surrounding rs1143679 shows numerous signs of active transcriptional modulation, including significant differences in transcript levels for patient monocytes between risk and protective alleles. As we did not observe splicing defects or degradation, we sought to directly measure transcriptional activity of the rs1143679 locus. We used PCR and cloned 110 bp homozygous double-stranded DNA (dsDNA) fragments surrounding the ‘G’ or ‘A’ rs1143679 allele into the TKmin-mGL.1 mammalian minimal expression vector, followed by a luciferase reporter assay. We transfected each construct into two cells types: HeLa and the immortalized monocyte line MonoMac-6. In HeLa cells, both allele-carrying fragments increased luciferase activity over the minimal promoter by a small but significant amount (Fig. 3D). In MonoMac-6 cells, this increase was several hundred-fold (Fig. 3E). This level of transcriptional enhancement from such a small DNA sequence is striking and confirms that this locus is a potent enhancer fragment, particularly in monocytes. In both cell lines, the risk allele transcribed at levels ∼20% less than the protective allele (P = 6.4 × 10−4 for HeLa; P = 3.9 × 10−4 for MonoMac-6) (Fig. 3D and E), indicating that the conserved rs1143679 ‘G’ base is needed for full transcriptional enhancer activity.

Protein-binding activity of the rs1143679 locus

We investigated the molecular mechanisms mediating the strong transcriptional enhancer activity of the rs1143679 locus (e.g. transcription factor binding). Homozygous ‘GG’ and ‘AA’ 110 bp dsDNA fragments (the same PCR products used in the luciferase assay) were used as bait in electrophoretic mobility shift assays (EMSAs) against nuclear protein extracts from MonoMac-6 cells. DNA fragments showed significant mobility shifts, indicating high occupancy of the rs1143679 locus by bound proteins (Fig. 4A, Supplementary Material, Fig. S10A). Moreover, the ‘G’ allele bound at least twice as much protein as the risk ‘A’ allele. Next, we used a competition assay (Fig. 4B) where ‘G’ allele-carrying DNA was labeled (‘hot’) in the presence of increasing quantities of non-labeled (‘cold’) DNA for ‘G’ or ‘A’. We observed that the intensity of the shifted band was nearly abolished in the presence of excess cold ‘G’. However, addition of excess cold ‘A’ did not compete as well as ‘G’, suggesting that protein binding is enhanced against the ‘G’ sequence. Similar results were observed for the converse experiment. Thus, we found that protein binding to the DNA varies quantitatively for alternative allele-carrying sequences, and protein binding with ‘bait’ DNA is sequence-specific. However, our competition assay does not rule out non-risk or risk allele-specific binding of different proteins in the cell extracts.

Figure 4.

In vitro protein binding and enhancer activity of rs1143679 locus. (A) dsDNA containing rs1143679 binds tightly to nuclear protein extract from MonoMac-6 cells. The ‘G’ sequence binds twice as much protein as the ‘A’ sequence. Nonspecific (NS) DNA did not produce shifted bands under these conditions. (B) Competition assay with labeled ‘G’ (hot) and unlabeled (cold) ‘G’ or ‘A’ carrying sequences. Similarly, labeled ‘A’ (hot) with unlabeled (cold) ‘A’ or ‘G’ carrying sequences. (C) Anti-NF-κB1, anti-EBF1 and anti-Ku70/80 antibodies super-shifted the EMSA band. (D) Oligo binding with recombinant Ku70/80 and NFKB1. Biotin-labeled nonspecific duplex PCR products (NS) do not bind any of these proteins. Duplex oligo DNA carrying ‘G’ binds with both of these proteins, with higher affinity for duplex sequences than single stranded oligo (Supplementary Material, Fig. S10). Protease digestion of proteins in the EMSA bound shifted band in excess duplex DNA abolished/reduced shifted bands.

Figure 4.

In vitro protein binding and enhancer activity of rs1143679 locus. (A) dsDNA containing rs1143679 binds tightly to nuclear protein extract from MonoMac-6 cells. The ‘G’ sequence binds twice as much protein as the ‘A’ sequence. Nonspecific (NS) DNA did not produce shifted bands under these conditions. (B) Competition assay with labeled ‘G’ (hot) and unlabeled (cold) ‘G’ or ‘A’ carrying sequences. Similarly, labeled ‘A’ (hot) with unlabeled (cold) ‘A’ or ‘G’ carrying sequences. (C) Anti-NF-κB1, anti-EBF1 and anti-Ku70/80 antibodies super-shifted the EMSA band. (D) Oligo binding with recombinant Ku70/80 and NFKB1. Biotin-labeled nonspecific duplex PCR products (NS) do not bind any of these proteins. Duplex oligo DNA carrying ‘G’ binds with both of these proteins, with higher affinity for duplex sequences than single stranded oligo (Supplementary Material, Fig. S10). Protease digestion of proteins in the EMSA bound shifted band in excess duplex DNA abolished/reduced shifted bands.

To identify these DNA-binding proteins, we sequenced DNA-bound protein(s) of the ‘GG’ EMSA band with mass spectrometry using MALDI-TOF from 2D gel electrophoresis after resolving the DNA-bound protein band (Supplementary Material, Fig. S10B and C). Database assignment of peptides from this band identified (Supplementary Material, Table S2) two heat shock proteins (HSP90AA1/AB1), DNA UV damage repair proteins [XRCC5 (Ku70)/XRCC6 (Ku80), a ‘lupus autoantigen’], actin (ACTB) and translin (TSN) as the major protein constituents of the band.

Mass spectrometric sequencing of the MonoMac-6-derived EMSA band produced few of the transcription factors discovered by the ENCODE project as labeling the rs1143679 locus (Supplementary Material, Fig. S2). Low expression in MonoMac-6 cells or loss during in vitro manipulation may have led to the failure to annotate these proteins by mass spectrometry. Thus we used direct production and testing of high-priority targets from ENCODE, including NF-κB1 and EBF1. These two proteins, as well as recombinant Ku70/80, were observed to super-shift the EMSA bait sequence (Fig. 4C), verifying that they bind to the rs1143679 locus as determined by ENCODE. However, quantitation of the super-shift band suggests that these three proteins are insufficient to explain the total protein content in the shifted band, suggesting that other proteins could be bound as well (e.g. other ENCODE-identified proteins). The molecular weights of commercial antibodies (IgG) are similar and are expected to run at this apparent molecular weight as well, and may contribute to the size of the band.

Smaller oligonucleotides containing the rs1143679 locus at the center also bound Ku70/80 and NF-κB1, and bound more tightly when double-stranded (Supplementary Material, Fig. S11), consistent with their primary binding to dsDNA. Protease digestion of the protein-bound shifted band from EMSA verified that shifting is caused by protein binding and not oligomerization of the oligonucleotides (Fig. 4D).

ITGAM stable cell lines show risk allele-specific reduced ligand-binding activity

The R77H mutation changes a highly conserved arginine to histidine in the CD11b β-propeller domain near the site that interacts with the I-domain of Mac-1 (25). A recent report indicates that there is decreased binding of monocytes from rs1143679 ‘AA’ patients (mutant, CD11b-77H) to iC3b, DC-SIGN, ICAM1 and fibrinogen (FBN) relative to rs1143679 ‘GG’ (wild-type, CD11b-R77) patients (26). Furthermore, both phagocytosis and Toll-like receptor 7/8 (TLR7/8)-induced cytokine release are also reduced for the mutant allele (26,27).

Here we focused on measuring the effects of ligand-binding affinity and specificity on the ability of biologically relevant CD11b/CD18 ligands to transduce cellular phenotypes. We assessed ligand-binding activity of K562 cell lines stably expressing wild-type Mac-1 or mutant Mac-1 to the natural ligands FBN and VTN. Mac-1 stable cell lines were stimulated with phorbol butyrate ester (PBT) and incubated with fluorescently labeled FBN, and then quantitated by flow cytometry. The percentage of FBN+ cells within the population of CD11b+/CD18+ positive cells was used as a measure of FBN recognition by the Mac-1 displaying cells. Cells containing the 77H mutant Mac-1 complex showed ∼3- to 5-fold weaker binding to FBN than those displaying wild-type Mac-1 (Fig. 5A; Supplementary Material, Fig. S12A). Similarly, 77H Mac-1 cells showed ∼3-fold weaker binding to VTN than wild type (Fig. 5B; Supplementary Material, Fig. S12B). Untransfected K562 cells bound FBN and VTN with <0.2% efficiency. The binding efficiency of both FBN and VTN were reduced (∼50%) in the presence of an anti-CD11b antibody (with ICRF44; data not shown), demonstrating specificity of the interaction. Thus, known ligand-binding activities of CD11b/CD18 measured on the surface of K562 cells were reduced by the risk allele, implicating altered protein function of CD11b due to the polymorphism.

Figure 5.

Mutant Mac-1 is polarized on the cell surface. (A) Cells stably expressing wild-type (‘G’ allele, R77) CD11b and CD18 bind fibrinogen (FBN+ cells) and (B) VTN (VTN+ cells) 3-4 times better than mutant (‘A’ allele, H77). (C) Recombinant mutant CD11b-H binds fibrinogen poorly compared with the wild-type (CD11b-R). Antibodies against the I-domain block the fibrinogen (Fb) binding abilities of the CD11b proteins. (D) Mac-1 distribution is polarized in ‘AA’ patients' monocytes. Merged pictures are shown for ‘GG’ (1) and ‘AA’ (3) patients' monocytes with distribution of CD11b (green). (2) and (4) are the same cells after incubation with ICAM1 (red). Nuclei are stained in blue. Overlap of the green and red channels produces yellow.

Figure 5.

Mutant Mac-1 is polarized on the cell surface. (A) Cells stably expressing wild-type (‘G’ allele, R77) CD11b and CD18 bind fibrinogen (FBN+ cells) and (B) VTN (VTN+ cells) 3-4 times better than mutant (‘A’ allele, H77). (C) Recombinant mutant CD11b-H binds fibrinogen poorly compared with the wild-type (CD11b-R). Antibodies against the I-domain block the fibrinogen (Fb) binding abilities of the CD11b proteins. (D) Mac-1 distribution is polarized in ‘AA’ patients' monocytes. Merged pictures are shown for ‘GG’ (1) and ‘AA’ (3) patients' monocytes with distribution of CD11b (green). (2) and (4) are the same cells after incubation with ICAM1 (red). Nuclei are stained in blue. Overlap of the green and red channels produces yellow.

In vitro ligand binding efficiency is directly affected by the risk-allele protein

The R77H mutation in the β-propeller domain appears to disrupt function of Mac-1 by modulating interactions with cellular ligands (27). However, I-domain conformational changes resulting in downstream signaling do not seem to be affected. We sought to confirm this by directly measuring CD11b interactions with ligands in vitro. We expressed CD11b-R (for the protective amino acid arginine) and CD11b-H (for the risk amino acid histidine) in Escherichia coli and purified the recombinant proteins. Binding to FBN (fragment D) was determined by a NABBA assay (see Materials and Methods) (11). CD11b-H had a ∼3.6-fold reduction in FBN binding (Fig. 5C) compared with wild-type CD11b-R. This is similar to cellular binding results and consistent with a model where the mutation primarily affects ligand binding rather than downstream signaling.

CD11b display is polarized in monocytes of homozygous risk genotype patients

While there is a relatively modest (3- to 5-fold) decrease in ligand-binding affinity of the mutant Mac-1 protein relative to wild-type, there is a marked cellular phenotype (e.g. abolished TLR7/8-induced cytokine release) (26) and significant association with SLE. Thus, we sought to determine whether other mechanisms operated to perturb Mac-1 signaling in the rs1143679 mutant. Recognition of cellular ligands by Mac-1 and subsequent signaling depends on proper display of cell surface complexes. In previous cellular binding experiments, Mac-1 components were overexpressed and were likely present at much higher levels than in endogenous phagocytes. Similarly, ligand binding was performed at steady-state with excess ligand. We reasoned that in endogenous cell types, lower expression levels of mutant Mac-1/R77H may give rise to more severe phenotypes, particularly in an in vivo setting where ligand recognition may also be limited by kinetic considerations.

We selected monocytes from homozygous ‘GG’ and ‘AA’ patients, the latter taken from patients with the highest levels of ITGAM transcript and CD11b protein from the previous assays. Using immunocytochemistry with an anti-CD11b antibody, we found that in wild-type (‘GG’) patients, Mac-1 was uniformly distributed on the monocyte membrane (Fig. 5D1,2) and bound its ligand ICAM1 (Fig. 5D2). In contrast, for risk-allele patients Mac-1 was completely polarized in the membrane (Fig. 5D3) and this polarization persisted even when bound with ICAM1 (Fig. 5D4). Because ligand binding on the surface of monocytes is primarily dependent on Mac-1 being exposed to the ligand, such polarization may significantly hamper recognition events involved in phagocytosis, such as the recruitment of the cytoskeleton (28), in a manner additive to direct ligand-binding affinity decreases (27).

Gene network analysis of ITGAM leading to SLE subphenotypes

Ingenuity pathway analysis (IPA) revealed an intricate interaction network of CD11b/CD18 (Supplementary Material, Fig. S13). The CD11b/CD18 (Mac-1) heterodimer interacts with a number of cell surface proteins throughout the immune system (10,29). Major functions of this complex include binding with deposited immune complexes (after antigen–antibody recognition) and debris in the kidneys' glomerular filtering devices as part of the phagocytic ‘sweeping’ activity of monocytes and neutrophils. Inhibition of such apoptotic clearance can induce cytokine release and subsequent inflammation (30). Based on IPA, Mac-1 impairment could lead to SLE phenotypes, including nephritis, autoantibody induction, cardiovascular complexities and immunological abnormalities. This supports our previous report, which found that lupus-related kidney disease was strongly associated with rs1143679 (9).

DISCUSSION

Our meta-analysis further confirms earlier reports (1–2,5) that rs1143679 explains the strong statistical association between ITGAM and SLE across diverse ethnic populations. Together, CD11b (encoded by ITGAM) and CD18 comprise the Mac-1 (CR3) receptor on neutrophils, monocytes, macrophages and dendritic cells. Interaction with extracellular ligands by Mac-1 is transduced into downstream signaling, leading to cell adhesion, phagocytosis, complement activation and inflammation (12,31–33). Thus it is unsurprising that it is linked to SLE (Supplementary Material, Fig. S13).

We explored how ITGAM influences SLE pathogenesis and found notable disruptions at both amino acid and nucleic acid levels by rs1143679. At the DNA level, rs1143679 is located in a highly conserved base near the exon-3 splice donor site within the most active chromatin region in ITGAM. This suggests that the mutation might disrupt splicing and/or transcription. rs1143679 is also inside the binding sites of many transcription factors, mainly leukocyte-specific. We observed significant dose-dependent reduction of ITGAM transcript and CD11b protein levels in monocytes from SLE patients carrying the risk allele ‘A’, consistent with allele-specific expression. Further, we found specific and significant repression of risk allele transcripts in heterozygous patients and most healthy controls, with many patients only expressing wild-type transcripts, albeit at reduced levels. The reduced levels of transcripts in healthy controls suggests that this reduction is likely not an indirect effect, e.g. of medicine use. Whether risk allele transcript levels correlate with SLE onset or subphenotypes is of prime interest for future research. Short DNA sequences surrounding rs1143679 significantly enhanced transcriptional activity in cell culture, particularly in MonoMac-6 monocyte cells, where a 110 bp fragment resulted in several hundred-fold greater reporter gene expression compared with the minimal promoter. The risk allele resulted in significantly lower expression levels compared with the wild-type allele in both cell lines tested.

Allele-specific expression revealed that some heterozygous individuals express little or no risk allele ‘A’ carrying RNA. In such cases, individuals would produce little or no mutant protein. Thus, the primary contribution of the rs1143679 risk allele to SLE may be because of reduced levels of functional protein. In individuals who express both risk RNA and protein, increased SLE risk may be precipitated by the deleterious effects on Mac-1 localization and ligand binding.

At the mechanistic level, we show that lupus autoantigen Ku70/80 binds directly to the rs1143679 ‘G’ allele, but significantly less to the ‘A’ allele. Ku70/80 is involved in numerous molecular processes such as DNA repair, recombination and transcription (34). Ku70/80 autoantibodies are a hallmark of SLE (35). Autoantibodies could impair function of the Ku70/80 protein, either by interfering with its enzymatic activity or by affecting binding to specific target sequences, including the rs1143679 risk allele. Both effects could contribute to SLE pathogenesis. Ku70/80 shows high affinity for another cis-acting locus in IFIH1, and mutations that disrupt Ku70/80 binding are associated with SLE in that context as well (36). Ku70/80 interacts with poly-A polymerase (37) and functions as a transcriptional regulator (38) and coactivator (39). It is the dsDNA-binding component of the DNA-dependent protein kinase (DNA-PK) assembly, which functions in DNA break repair and V(D)J antibody recombination (40). DNA-PK regulates numerous transcription factors as well as RNA polymerase II itself, giving Ku70/80 a prominent role in transcriptional regulation. Components identified in the EMSA complex, such as actin (ACTB) and heat shock proteins (HSP90AA1/AB1), also have important roles in transcription (34,41,42). In addition to a primary role as protein chaperones (43), HSP90 proteins regulate RNA polymerase activity (44–48). Moreover, HSP90 chaperoning of ‘self-DNA’ and CpG oligonucleotides to the static endosome is required for signaling through Toll-like receptor TLR9 for interferon (IFN-alpha) production from dendritic cells (41,49). Autoantibodies to HSP90 proteins are a distinct feature of SLE (50–53). Furthermore, HSP90 is a mediator for the metabolism of vitamin D, which is beneficial for SLE patients (54). Actin is critical for chromatin-dependent gene regulation, forming a core structure of RNApolII that directs chromatin modifiers and assembly of the preinitiation complex in addition to its other functions (42). Furthermore, actin and related cytoskeletal proteins play a key role in phagocytosis during formation of the cytoplasmic envelope surrounding foreign bodies. Impairment of these functions potentially contributes to autoimmunity development and SLE pathogenesis (55,56). We also identified translin (TSN), which is known to bind recombination hotspots and recombination complexes on DNA (57) and could bind the Ku70/80 complex.

Of the proteins identified by ENCODE (NF-κB, TCF12, CTCF, SP1, BATF, EBF1, BCL11a, PAX5) (Supplementary Material, Fig. S2), we confirmed a role for NF-κB and EBF1 using a super-shift assay. Oligonucleotide experiments recapitulated these results, and showed that the super-shift was indeed protein-dependent. Most of these transcription factors, including HSP90, Ku70/80 and translin, directly interact to produce a cellular network (Fig. 6A). NF-κB is a key player for immunological processes and regulates expression of numerous genes with autoimmune phenotypes (58–60). Disruption of the interactions of these proteins with the ITGAM locus correlates with, and may underlie, the gene expression phenotype we observed. The failure to validate some ENCODE-identified proteins by mass spectrometry may stem from low expression levels, sample handling or features of the MonoMac-6 cell line.

Figure 6.

Network interactions of proteins binding the rs1143679 locus and flow-chart for proposed CD11b function. (A) Network interactions of proteins binding the rs1143679 locus. ENCODE (light grey) identified proteins and EMSA (dark grey) identified proteins interact with one other and could potentially form a complex. Both NFKB1 and EBF1 (dark blue) were identified in EMSA and in ENCODE. Hexagonal box represents small molecules. (B) Model of ITGAM molecular function. The ITGAM risk allele affects function both at the level of transcript production and ligand binding activity, leading to impairment of multiple downstream functions. Heterozygous individuals (‘GA’) are affected by reduced levels of the wild-type protein and/or expression of the mutant protein.

Figure 6.

Network interactions of proteins binding the rs1143679 locus and flow-chart for proposed CD11b function. (A) Network interactions of proteins binding the rs1143679 locus. ENCODE (light grey) identified proteins and EMSA (dark grey) identified proteins interact with one other and could potentially form a complex. Both NFKB1 and EBF1 (dark blue) were identified in EMSA and in ENCODE. Hexagonal box represents small molecules. (B) Model of ITGAM molecular function. The ITGAM risk allele affects function both at the level of transcript production and ligand binding activity, leading to impairment of multiple downstream functions. Heterozygous individuals (‘GA’) are affected by reduced levels of the wild-type protein and/or expression of the mutant protein.

The rs1143679 SNP produces a non-synonymous mutation of the highly conserved Arg77 residue to His in the β-propeller domain (1–130 amino acids) of CD11b, adjacent to the I-domain (148–310 amino acids). Although the I-domain directly interacts with most extracellular ligands (25), the β-propeller domain has a role in conformational changes of the I-domain and may facilitate ligand binding (61). Recently, three groups reported that this risk allele affects binding of FBN, DC-SIGN, ICAM1 and iC3b ligands, and affects phagocytosis (26–27,62). Our binding studies, at the levels of cells and purified proteins, support these findings. Stable cells expressing mutant CD11b/CD18 bound inefficiently with FBN and VTN. Recombinant risk-allele CD11b protein showed similarly reduced FBN and VTN-binding activity, consistent with a role of the β-propeller domain in facilitating FBN binding (63,64). VTN competes with FBN binding, and could bind at the same sites as FBN on Mac-1 (65). The difference (2- to 3-fold) in binding between wild-type and mutant cells that we observed for FBN and VTN in the cellular model is similar to the observations of MacPherson et al. (27). Rhodes et al. (26) observed a 20% reduction in FBN binding with patients' macrophages ex vivo. It is expected that cellular binding studies should give clearer results than patients' cells ex vivo, as many confounding factors could mask the differences in studies with patient cells. However, differences in FBN binding are more distinct (3- to 4-fold) for in vitro FBN binding. Intriguingly, Mac-1 is also implicated as a primary receptor for oligonucleotide recognition and uptake, particularly of CpG nucleotides (66,67). Mac-1 and the CD87 complex are known to bind fibrin and VTN during leukocyte adhesion (65,68), and Mac-1 colocalizes with immuno-complexes in SLE nephritis (69). Inefficient binding of FBN to Mac-1 leads to increased FBN in plasma and results in increased polymerization of fibrin and systemic platelet activation (70). Indeed, this hypercoagulation due to excessive red blood cell aggregation is a distinct feature of SLE patients and is associated with SLE thrombolytic disease and nephritis (70–75). Epistatic interaction of FBN and platelet activating factor PAI-1 has also been observed in SLE nephritis patients (76). Altered Mac-1 function may disrupt immune responses to foreign nucleic acids (66) as well as proteins. This may impair complement activation and TLR9 signaling, leading to a feedback loop of immune system dysfunction.

We demonstrate a novel mechanism of Mac-1 disruption by the rs1143679 mutation, which produces highly polarized cell surface localization on patients' monocytes that contrasts with the uniform distribution of wild-type Mac-1. This polarized Mac-1 distribution on the monocyte cell surface leads to decreased recognition of target proteins, which is additive with decreased affinities for the ligands. The aberrant subcellular localization could also detrimentally affect cytoskeletal remodeling during phagocytosis. Whether such mislocalization results from differences in expression level stemming from the disruption of the enhancer, or whether the amino acid mutation interferes with proper trafficking, remains to be seen.

Our results clearly suggest that rs1143679 is most likely a causal SNP for lupus. However, the functional association of additional ITGAM SNPs with lupus, such as those identified in the calf and cytoplasmic regions (77), cannot be ruled out. These other SNPs may indeed also affect Mac-1 expression levels and/or activity. What we systematically show in this manuscript is the involvement and likely mechanism of a single SNP. A similar level of rigor would need to be applied to these other SNPs, after which point they may be found to plausibly contribute to SLE, or they may not. A separate study with large sample sizes would be required to make definitive conclusions in this regard.

These rs1143679-related protein-level changes and significant disruptions at the nucleic acid level may play prominent roles in its contribution to SLE predisposition and disease progression (Fig. 6B). Together, our results implicate the rs1143679 SNP risk allele in (i) disruption of a CpG sequence within a highly conserved locus associated with active chromatin regulation and transcription factor binding, (ii) weakened binding to Ku70/Ku80, NFKB1, EBF1, HSP90AA1/AB1 and other components of the RNApolII transcription complex, (iii) decreased activity of an exceptionally strong transcriptional enhancer in a reporter assay, (iv) risk allele-specific decreased ITGAM transcript levels in patients' monocytes leading to reduced surface protein expression, (v) reduced ligand-binding activity and (vi) mislocalization of the Mac-1 protein at the cell surface. Most likely a combination of these effects at different molecular levels (DNA and protein) results in the extremely robust contribution of this SNP to SLE susceptibility.

MATERIALS AND METHODS

Ethics statement

All individuals were de-identified prior to being genotyped. Written, informed consent was obtained from all study participants. This study was approved by the Institutional Review Boards of the OMRF and the ethical committees at the institutions where subjects were recruited.

Genotyping the rs1143679 SNP in new populations

We genotyped rs1143679 (680 cases, 616 controls) (Table 1, Fig. 1) from three previously unreported populations, Indian, Malayan and Chinese (Chinese living in Malaysia), using TaqMan assays (Applied Biosystems) designed with File Builder 2.0 software. Genotyping was performed in 384-well plates using Universal PCR Master Mix on an ABI 7900HT Sequence Detection System and SDS 2.0 software.

Meta-analysis

Meta-analyses used 28 439 unrelated individuals (12 992 cases, 15 447 controls) from published (1–2,6,13–17) and unpublished data for ITGAM SNP rs1143679 from 19 countries across Europe, North and Central America and Asia. Individuals included in more than one study were removed from analysis of subsequent studies for the purposes of this meta-analysis. Breslow–Day and Mantel–Haenszel tests assessed homogeneity of OR across populations. Since ORs were homogeneous, an Inverse Variance Fixed-Effects Model was appropriate for the meta-analysis, and was performed with R-Meta and CatMap (78).

Bioinformatics of the rs1143679 locus

The ITGAM locus (chromosome 16: 31 271 386 to 31 343 025; rs1143679 is at 31 276 811 bp) was annotated (∼20–30 kb upstream and downstream, 31 250 000 to 31 375 000) for chromatin regulation state using the UCSC Genome Browser (http://genome.ucsc.edu). The active chromatin mark H3K27Ac (Abcam antibody ab4729) track was taken from the ENCODE/Broad Institute data set (Bernstein Laboratory, Release 3) for GM12878, a lymphoblastoid cell line produced from the blood of a female donor of European ancestry. Data (‘Raw Signal’) was downloaded in BigWig format and converted to floats for plotting. To create the plot (Supplementary Material, Fig. S1), ‘start’ and ‘stop’ were averaged to give rise to a centroid genomic position, and raw signal was normalized to the peak of the H3K27Ac signal over the selected interval, which occurs at 31 277 076–31 277 100 (center 31 277 088); rs1143679 is at 31 276 811.

The DNase I hypersensitivity track was also downloaded from the ENCODE/University of Washington data set (‘Raw Signal’). Interval position was treated as with the H3K27Ac track. The signal was normalized to the peak of the DNase I replicate #1 track over the selected interval, which occurs at 31 276 844–31 276 864 (center 31 276 854). The GM12878 signal versus the chromosome was plotted; Replicate 1 is shown because both replicates are extremely similar. Transcription factor binding sites were downloaded from the ENCODE ChIP-Seq data track (Supplementary Material, Fig. S2), and were shown if they bound to rs1143679 or if they showed strong (>900/1000) labeling at some point in the 31 250 000–31 375 000 interval.

RNA and surface protein expression in protective and risk genotype individuals

Monocytes were collected from SLE patients (n = 53) and healthy controls (n = 3). Based on available ethnographic data, most of the individuals were from the Columbian Paisa community, which is historically considered highly homogeneous (79). They were between 10 and 58 years old. Most (nearly 90%) patients were female, which is consistent with the known male : female ratio for lupus (Supplementary Material, Table S1). All SLE patients were in remission when monocytes were collected; their genotype-specific clinical criteria are shown in Supplementary Material, Table S1. PBMCs were isolated from blood with a Ficoll-hypaque kit (Sigma-Aldrich, St.Louis, Mo, USA). These PBMCs were stained with fluorescently labeled (Alexa-488) antibodies against CD14 (monocyte specific, Abcam, USA) and FACS (Fluorescence-Activated Cell Sorting) sorted for CD14+ cells to purify monocytes (Supplementary Material, Fig. S3). For RNA quantitation, total RNA was isolated (Get Pure RNA purification kit, Qiagen, Valencia, CA, USA) from these monocytes and reverse transcribed to cDNA by random prime labeling (cDNA kit, Invitrogen now Life Technologies, Grand Island, NY, USA). ITGAM probe (00167304_m1; exon-8 to exon-9, Applied Biosystems now Life Technologies, Grand Island, NY, USA) was used for qPCR Taqman assays with preassayed glyceraldehyde-3-phosphate dehydrogenase (GAPD) as an internal control. For surface protein expression, PBMCs were stained with fluorescently labeled anti-CD14 (Alexa-488) and anti-CD11b (ICRF44, Alexa-700 conjugated, Abcam Cambridge, MA, USA) antibodies and analyzed for CD11b MFI (mean fluorescent intensity) of CD14+ or CD14+CD11b+ cells by flow cytometry with FLOWJO software (Supplementary Material, Fig. S4). We also estimated the frequency of CD14+CD11b+ cells within the CD14+ cells, and CD14+ cells from total cells. MFI of CD11b or frequency of these cells were compared between ITGAM genotypes, and statistical significance for each genotype was calculated using a t-test.

Quantitation of risk ‘A’ allele specific transcript in heterozygous individuals

cDNAs from rs1143679 heterozygous (‘GA’ genotype) individuals' monocytes were subjected to PCR amplification using neighboring exonic primers (forward primer-agagcgtggtccagcttcag, reverse primer- aatgtcactatcctcttgagg). RT–PCR products (347 bp) were cloned into a TA cloning vector, and over 30 clones from each individual were sequenced to identify ‘G’ or ‘A’ allele-carrying clones. A t-test was used to assess statistical significance.

For the sequencing-based allelic expression assay, we used a primer pair (forward primer-tgccaaccaaaggggcagcctctaccag; reverse primer-tgcacggtgggaccacagg) that amplifies 162 bp of cDNA and 622 bp of genomic DNA. cDNAs of seven SLE patients (GA) and one healthy control (GA) were amplified with these primers and directly sequenced with the forward primer. Genomic DNA from the healthy individual was also sequenced as a reference sample, as G : A ratio should be 50 : 50. The area under the intensity curve (as a triangle) for the ‘A’ or ‘G’ peak of the ABI electropherogram trace was calculated, and the ratio of the ‘A’ to ‘G’ was estimated for each sample.

RT–PCR assay for potential splicing defects

Monocyte cDNA from three homozygous risk genotype (‘AA’) SLE patients was used to amplify precursor RNA under stringent and non-stringent PCR conditions. Exonic primers (exon-3 forward primer (P1)-AGGAGATAGTGGCTGCCAACCAAAGG and exon-5 reverse primer (R1)-CTGCTGCCGTAGGTTGGATCC) were used for PCR. Fully spliced cDNA should produce a band of 245 bp and genomic DNA should produce a band of 705 bp. This 705 bp includes 2 introns; splicing of single introns is expected to yield bands of 393 bp (only intron-4 spliced) and 557 bp (only intron-3 spliced) (Supplementary Material, Fig. S7). The rs1143679 SNP is near the exon-3/intron-3 boundary, thus in case of non-splicing, inclusion of intron-3 in the pre-mRNA would be expected. From non-stringent PCR, all intermediate bands (between 245 and 705 bp) were excised from the gel, cloned into a TA cloning vector, and 30 clones were sequenced to look for unspliced introns in the pre-mRNA.

Generation of stable cell lines

Full-length cDNA of ITGAM (containing the rs1143679 ‘G’ allele) was obtained from Origene Technologies, Rockville, MD, USA (10) in pBR322. ITGAM cDNA was PCR amplified (forward primer-attctgacgcgtatgctctcagagtccttctg and reverse primer-ttgtaaatatcgatactggggttcggccc, amplicon size 3470 bp) and cloned into the MluI-ClaI multiple cloning site (MCS1) of the pVITRO2-neo-MCS dual expression vector (Invivogen, USA) for expression in mammalian cells. In the same vector, CD18 was cloned (full length 2306 bp cDNA obtained from Research Genetics and amplified with forward primer -aatacgaattcatgctgggcctgcgccccccacttc and reverse primer-agtttcgctagcactctcagcaaacttggggttc) into the EcoRI-NheI site in the second cloning site, MCSII. ITGAM and CD18 sequences were under control of a composite hFerH-hFerL human ferritin promoter in this dual expression vector. Both genes were cloned in the same vector to ensure formation of Mac-1 in the cells. These cDNA sequences were verified by sequencing from both directions with vector and internal primers. After cloning, the ‘G’ allele was replaced by the ‘A’ allele of rs1143679 in the ITGAM gene through site directed mutagenesis (XL site-directed mutagenesis kit, Stratagene now Agilent Technologies, Santa Clara, CA, USA), and sequences were verified to ensure that no unintended mutations were introduced during mutagenesis.

Both ‘A’ and ‘G’ allele-carrying ITGAM cDNAs (with CD18) were transfected to K562 cells, a myelogenous leukemia cell line in which neither of these integrins is endogenously expressed (80). Transfected cells were maintained with stringent Geneticin selection for at least 1 month and repeatedly flow sorted (Supplementary Material, Fig. S8A and B) with anti-CD11b (Alexa-700 conjugated) and/or anti-CD18 antibodies (FITC conjugated). CD11b and CD18 transcription were confirmed (Supplementary Material, Fig. S8C and D) by multiplexed RT–PCR analysis (for CD11b specific forward primer-agagcgtggtccagcttcag and reverse primer-aatgtcactatcctcttgagg, amplicon size 347 bp; for CD18, forward- tgccttcaggcacgtgctgaagctgacc, reverse-cgttgctcctcttgtacaagttgtcctcc; amplicon size 287 bp) with HPRT1 (hypoxanthine phosphoribosyltransferase 1) internal controls (forward—aagcttgctggtgaaaaggac, reverse—gtcaagggcatatcctacaac, 100 bp amplicon).

Degradation assay for ITGAM RNA

Cells stably expressing CD11b [R77 (non-risk) or H77 (risk)] and CD18 from three individual isolates were flow sorted and freshly cultured, and aliquots of cells were collected at 2-h and then 24-h intervals. Total RNA was isolated from cells collected at different time points and RT–qPCR was performed for these cDNAs with ITGAM and GAPD (as internal control) TaqMan probes (same probes used for expression analysis, Applied Biosystems now Life Technologies, Grand Island, NY, USA) in a time-dependent manner. Each cDNA was subjected to qPCR in triplicate and repeated twice. The average value of each data point was used to draw the curve.

Patient PBMCs were cultured in the standard culture medium containing RPM11640 and 10% fetal bovine serum. Cells were collected at start and after 16 h. Total RNA was isolated, and cDNAs (random-primed) were prepared. ITGAM RNA was quantitated using qPCR with the same ITGAM probe and internal control (GAPD) used for ITGAM expression in patients' monocytes.

Electrophoretic mobility shift assay

110-mer DNA sequences surrounding rs1143679, constituting exon-3 of ITGAM, were amplified from ‘GG’ (homozygous protective) and ‘AA’ (homozygous risk) patients using PCR. While the forward primer (F2) is common for both cases (5′-CCACAGGGTGGTGGTTGGAGC) the reverse primer differs for ‘GG’ genotype (R2a): 5′-AGGCGGATGGGCTCGCATGAGC and for ‘AA’ genotype (R2b): 5′-AGGTGGATGGGCTCGCATGAGC). As the variant lies near an exon/intron junction, primers were designed to include only exonic sequences so as to avoid any possible interaction of these sequences with the splicing machinery (in this in vitro setting we cannot explicitly rule out non-specific binding of spliceosome components, which typically recognize RNA, to the splice junction DNA template). PCR products were sequenced to confirm that no other mutations were incorporated during PCR. Nuclear extracts were prepared from MonoMac-6 cells (EpiQuik Nuclear Extraction Kit, Epigentek, Farmingdale, NY, USA). PCR-amplified DNA sequences carrying non-risk ‘G’ allele or risk ‘A’ allele, and a 140 bp nonspecific DNA (generated by a separate PCR from bisulfite modified genomic DNA: forward 5′-TGTTATTTAGATTGGAGTGTAGTGGTAC, reverse 5′-AATCCCAACACTTAAAAAAACTACG, thus non-existent in normal cellular conditions) were purified and incubated with increasing amounts of nuclear extracts (20 to 40 ng) for 25 min at 25°C in 10 µl reactions containing 2–3 ng of biotin-labeled DNA (3′-end labeling kit, Thermo Fisher Scientific, Waltham, MA, USA) and 1 μg of Poly dI : dC. For competition assay, ‘G’ or ‘A’ allele carrying PCR amplified non-labeled DNA (cold) was added in increasing quantities (6 and 12 ng) in the EMSA reaction. EMSA-bound proteins were fractionated in 4–15% Tris–HCl precast PAGE gels (Bio-Rad Laboratories, Herculis, CA, USA), run in TBE (under native conditions), transferred to BioDyn nylon membranes (Thermo Fisher Scientific, USA) and detected using Pierce nonradioactive chemiluminescence detection kit (Thermo Fisher Scientific, USA). Intensities (as a measure of DNA quantity) of the shifted and free DNA bands were calculated with TotalLab Quant (TotalLab Quant Limited, Newcastle upon Tyne, UK) for each lane. The ratio of DNA quantity of the shifted band to the free DNA band in each lane was calculated as a measure of binding efficiency.

For visualization, EMSA was performed with a nonradioactive EMSA kit (Invitrogen, now Life Technologies, Grand Island, NY, USA) with SYBR Green, which detects DNA to 1 ng sensitivity. Reaction mixtures were loaded onto a 4–20% Tris–HCl neutral PAGE gel (Bio-Rad laboratories) (Supplementary Material, Fig. S10A). DNA bands were visualized with SYBR Green staining. Intensities (as a measure of DNA quantity) of the shifted and free DNA bands were calculated with TotalLab Quant (TotalLab Quant Limited, Newcastle upon Tyne, UK) for each lane. The ratio of DNA quantity of the shifted band to the free DNA band in each lane was calculated as a measure of binding efficiency.

Super-shift assay

EMSAs for super-shift assay were performed as described above except: (i) with ‘G’ allele-carrying sequences and (ii) in the presence of EMSA compatible antibodies (0.8 µg in 10 µl reaction) against NFKB1 (raised against p50 subunit, Santa Cruz Biotechnology, USA), EBF1 (Santa Cruz Biotechnology, Dallas, TX, USA) or Ku70/80 (GeneTex, Irvine, CA, USA). Similarly, as described for EMSA assay, precast 4–15% neutral Tris–HCl PAGE gels were used for fractionating these reactions, followed by transfer to nylon membranes and detection using chemiluminescence as described above.

Identification of proteins from EMSA complex by 2D electrophoresis

EMSA reaction was performed in large scale and a neutral PAGE gel (4–20%, Bio-Rad Laboratories) was run in the first dimension. The gel was stained with SYBR Green to locate the DNA-bound protein band (Supplementary Material, Fig. S10B). For the second dimension, the gel was rotated 90° and run for 6 more hours. It was again stained with SYBR Green to locate the position of the DNA-bound protein band (Supplementary Material, Fig. S10C) then stained with Coomassie Blue. The DNA-bound protein bands were excised from the gel and sequenced in a mass spectrometry core facility (www.lmbcr.ouhsc.edu). Data obtained from these analyses were submitted to the MASCOT (Matrix Science) server for protein identification against the SwissProt protein database (2011).

Oligo binding assay with recombinant proteins

Biotin labeled oligonucleotides (50 pg) (sequence, 5′-TGCGAGCCCATCCGCCTGCAGGGTGAGTCACTGCCCCGC) were incubated with 1 ng of purified protein (Ku70/80, Biorbyt, UK; and NFKB1, Abcam, USA) at 37°C for 30 min. Duplex oligo DNA was prepared by heating the oligo with the complementary oligo (5′-GCGGGGCAGTGACTCACCCTGCAGGCGGATGGGCTCGCA) at 96°C and then cooling slowly to room temperature. A non-specific DNA (a PCR amplification that contained 124 and 530 bp, generated by PCR of bisulfite treated genomic DNA) was used as a control (this non-specific control was distinct from that used for EMSA). Each excess duplex oligo-carrying EMSA reaction mixture (for Ku70/80 or NFKB1) was divided into two aliquots, and each was incubated with excess protease (proteinase K, 2 µg in 10 µl reaction) (Sigma, USA) to digest the proteins in the shifted band, in order to confirm that the heavier band was not due to higher-order structures of the duplex DNA.

Luciferase assay

Luciferase assay was performed using the same DNA sequences from EMSA in order to test whether these sequences act as transcription regulatory elements and to correlate bound proteins with gene expression activity. These sequences were amplified by primer F2, with R2a or R2b (described for EMSA) from ‘GG’ and ‘AA’ individuals and cloned into a minimal promoter enhancer/silencer assay vector (TKmin-mGL.1, Xactagen, Shoreline, WA, USA; under minimal TKmin promoter). Luciferase activity in HeLa and MonoMac-6 cells was measured at 48 and 6 h, respectively, after transfection (luciferase activity in MonoMac-6 cells peaks at 6 h) (81). Luciferase activity was measured in 32 independent transfections for HeLa cells, and 9 independent transfections for MonoMac-6 cells. For HeLa cells, for each transfection, cells were grown in one well of a 24-well plate, trypsinized and cells from each well were divided into five tubes for electroporation of five constructs (MCS-mGL.1, the luciferase reporter gene with no promoter, having only the multiple cloning site; Tkmin-mGL.1; CMV-mGL.1; 679G-TKmin-mGL.1; and 679A-Tkmin-mGL.1) and each well was divided for two assay reads. For MonoMac-6 cells, the suspension culture was not trypsinized, but harvested and divided into five tubes for transfection, and subsequently divided for three assay reads. As expected, the potent viral CMV promoter showed strong luciferase activities in each transfection set (data not shown). Each measurement was normalized to the value of non-transfected cells. Mean values for 64 measurements of HeLa cells, 27 of MonoMac-6 cells, and equivalent numbers for the controls (MCS-mGL.1, TKmin-mGL.1,679G-Tkmin-mGL.1 and 679A-Tkmin-mGL.1) were used to generate graphs from this large number of replicates. Statistical significance was calculated using a t-test.

Cellular ligand binding assay

Cells stably expressing either allele of CD11b and CD18 were activated with PBT (200 ng/ml). Activated cells were incubated for 10–15 min at 37°C with anti-CD11b-Alexa700 and anti-CD18-FITC antibodies, and FBN conjugated with Alexa594, 200 µg/ml (Invitrogen). Among CD11b+/CD18+ cells, FBN+ cells were analyzed in flow cytometry, and data were analyzed with t-tests.

For the VTN binding assay, cells were incubated with anti-CD11b-Alexa700, anti-CD18-FITC, VTN (Antigenix America, USA), 200 µg/ml for 15 min at 37°C followed by incubation with a mouse anti-VTN antibody. VTN binding was assessed with donkey anti-mouse antibody conjugated with Alexa594. Alexa594 positive (VTN+) cells were analyzed among CD11b+/CD18+ cells.

Cloning and purification of recombinant CD11b proteins in E. coli, and in vitro ligand binding assays

We cloned a 930 bp (1–310 amino acids) cDNA carrying the β-propeller domain (1–130 amino acids) and the von Willebrand factor type A (VWA, I) domain (148–310 amino acids) of CD11b into the NdeI /HindIII sites of the pET28 expression vector, which adds an N-terminal poly-histidine tag for purification and a maltose binding protein (MBP) fusion to improve solubility. Proteins were produced for risk and protective rs1143679 alleles, and are designated as CD11b-R (for arginine, protective amino acid) and CD11b-H (for histidine, risk amino acid). We expressed proteins in E. coli (BL21 codon plus, Invitrogen, now Life Technologies, Grand Island, NY, USA) and purified them using Ni2+-NTA agarose (Sigma-Aldrich, St. Louis, MO, USA) beads followed by FPLC.

For the FBN binding ‘NABBA’ assay (Ni2+–NTA agarose beads binding assay), we immobilized these CD11b-R and CD11b-H His-tagged proteins on Ni2+-NTA-agarose beads, then incubated with Fibrinogen fragment D (Fb) for 1 h with gentle shaking at room temperature. Beads were washed extensively with wash buffer to remove any unbound FBN. Finally, CD11b-bound FBN was eluted with 100 mm imidazole and fractionated in a denaturing SDS–PAGE gel, releasing CD11b and FBN. Antibodies to CD11b I-domain block the ligand binding ability of CD11b protein, and were used to compete off CD11b. In wild-type and mutant lanes, FBN and CD11b bands were quantitated by densitometric scanning with TotalLab Quant Limited, Newcastle upon Tyne, (UK), and the ratio of the intensity of the FBN band to the CD11b band was used to determine binding efficiency.

Immunocytochemistry

Purified monocytes from ‘GG’ and ‘AA’ genotype-carrying patients were plated on plastic cover slips and fixed with 50% methanol/50% acetone. After fixation, cells were incubated with mouse monoclonal anti-CD11b antibody (ICRF44, Abcam, USA) followed by incubation with conjugated secondary antibody (goat anti-mouse-FITC) and DAPI for nuclei. Some cells were also stained with an anti-ICAM1 (Abcam, rabbit) antibody and secondary antibody conjugated with rhodamine (donkey anti-rabbit). Cells were initially visualized on an epifluorescence microscope and then photographed on a confocal microscope.

Ingenuity pathway analysis

Ingenuity pathway was constructed using IPA (Ingenuity Pathway Analysis; www.ingenuity.com), as we performed earlier (82). IPA is knowledge-based software suitable for identifying direct or indirect interactions between two or more molecules. IPA provides two types of connections: solid lines for direct interactions and dotted lines for indirect interactions (through a connecting molecule).

SUPPLEMENTARY MATERIAL

Supplementary Material is available at HMG online.

FUNDING

This work was supported by grants from the US National Institutes of Health (AI103399, AR060366, AI094377).

ACKNOWLEDGEMENTS

We thank the patients and their families for their cooperation. We are grateful to John Knight, PhD, for critical reading and helpful comments on this manuscript. Jonathan Marvin assisted in the formatting of Supplementary Material, Figures S1 and S2.

Conflict of Interest statement. None declared.

REFERENCES

1
Nath
S.K.
Han
S.
Kim-Howard
X.
Kelly
J.A.
Viswanathan
P.
Gilkeson
G.S.
Chen
W.
Zhu
C.
McEver
R.P.
Kimberly
R.P.
, et al.  . 
A nonsynonymous functional variant in integrin-alpha(M) (encoded by ITGAM) is associated with systemic lupus erythematosus
Nat. Genet.
 , 
2008
, vol. 
40
 (pg. 
152
-
154
)
2
Han
S.
Kim-Howard
X.
Deshmukh
H.
Kamatani
Y.
Viswanathan
P.
Guthridge
J.M.
Thomas
K.
Kaufman
K.M.
Ojwang
J.
Rojas-Villarraga
A.
, et al.  . 
Evaluation of imputation-based association in and around the integrin-alpha-M (ITGAM) gene and replication of robust association between a non-synonymous functional variant within ITGAM and systemic lupus erythematosus (SLE)
Hum. Mol. Genet.
 , 
2009
, vol. 
18
 (pg. 
1171
-
1180
)
3
Harley
J.B.
Alarcon-Riquelme
M.E.
Criswell
L.A.
Jacob
C.O.
Kimberly
R.P.
Moser
K.L.
Tsao
B.P.
Vyse
T.J.
Langefeld
C.D.
, et al.  . 
International Consortium for Systemic Lupus Erythematosus Genetics (SLEGEN)
Genome-wide association scan in women with systemic lupus erythematosus identifies susceptibility variants in ITGAM, PXK, KIAA1542 and other loci
Nat. Genet.
 , 
2008
, vol. 
40
 (pg. 
204
-
210
)
4
Hom
G.
Graham
R.R.
Modrek
B.
Taylor
K.E.
Ortmann
W.
Garnier
S.
Lee
A.T.
Chung
S.A.
Ferreira
R.C.
Pant
P.V.
, et al.  . 
Association of systemic lupus erythematosus with C8orf13-BLK and ITGAM-ITGAX
N. Engl. J. Med.
 , 
2008
, vol. 
358
 (pg. 
900
-
909
)
5
Molineros
J.E.
Kim-Howard
X.
Deshmukh
H.
Jacob
C.O.
Harley
J.B.
Nath
S.K.
Admixture in Hispanic Americans: its impact on ITGAM association and implications for admixture mapping in SLE
Genes Immun.
 , 
2009
, vol. 
10
 (pg. 
539
-
545
)
6
Yang
W.
Zhao
M.
Hirankarn
N.
Lau
C.S.
Mok
C.C.
Chan
T.M.
Wong
R.W.
Lee
K.W.
Mok
M.Y.
Wong
S.N.
, et al.  . 
ITGAM is associated with disease susceptibility and renal nephritis of systemic lupus erythematosus in Hong Kong Chinese and Thai
Hum. Mol. Genet.
 , 
2009
, vol. 
18
 (pg. 
2063
-
2070
)
7
Anaya
J.M.
Kim-Howard
X.
Prahalad
S.
Chernavsky
A.
Canas
C.
Rojas-Villarraga
A.
Bohnsack
J.
Jonsson
R.
Bolstad
A.I.
Brun
J.G.
, et al.  . 
Evaluation of genetic association between an ITGAM non-synonymous SNP (rs1143679) and multiple autoimmune diseases
Autoimmun. Rev.
 , 
2012
, vol. 
11
 (pg. 
276
-
280
)
8
Carmona
F.D.
Simeon
C.P.
Beretta
L.
Carreira
P.
Vonk
M.C.
Rios-Fernandez
R.
Espinosa
G.
Navarrete
N.
Vicente-Rabaneda
E.
Rodriguez-Rodriguez
L.
, et al.  . 
Association of a non-synonymous functional variant of the ITGAM gene with systemic sclerosis
Ann. Rheum. Dis.
 , 
2011
, vol. 
70
 (pg. 
2050
-
2052
)
9
Kim-Howard
X.
Maiti
A.K.
Anaya
J.M.
Bruner
G.R.
Brown
E.
Merrill
J.T.
Edberg
J.C.
Petri
M.A.
Reveille
J.D.
Ramsey-Goldman
R.
, et al.  . 
ITGAM coding variant (rs1143679) influences the risk of renal disease, discoid rash and immunological manifestations in patients with systemic lupus erythematosus with European ancestry
Ann. Rheum. Dis.
 , 
2010
, vol. 
69
 (pg. 
1329
-
1332
)
10
Corbi
A.L.
Kishimoto
T.K.
Miller
L.J.
Springer
T.A.
The human leukocyte adhesion glycoprotein Mac-1 (complement receptor type 3, CD11b) alpha subunit. Cloning, primary structure, and relation to the integrins, von Willebrand factor and factor B
J. Biol. Chem.
 , 
1988
, vol. 
263
 (pg. 
12403
-
12411
)
11
Loike
J.D.
Sodeik
B.
Cao
L.
Leucona
S.
Weitz
J.I.
Detmers
P.A.
Wright
S.D.
Silverstein
S.C.
CD11c/CD18 on neutrophils recognizes a domain at the N terminus of the A alpha chain of fibrinogen
Proc. Natl. Acad. Sci. USA.
 , 
1991
, vol. 
88
 (pg. 
1044
-
1048
)
12
Diamond
M.S.
Staunton
D.E.
Marlin
S.D.
Springer
T.A.
Binding of the integrin Mac-1 (CD11b/CD18) to the third immunoglobulin-like domain of ICAM-1 (CD54) and its regulation by glycosylation
Cell
 , 
1991
, vol. 
65
 (pg. 
961
-
971
)
13
Suarez-Gestal
M.
Calaza
M.
Dieguez-Gonzalez
R.
Perez-Pampin
E.
Pablos
J.L.
Navarro
F.
Narvaez
J.
Marenco
J.L.
Herrero-Beaumont
G.
Fernandez-Gutierrez
B.
, et al.  . 
Rheumatoid arthritis does not share most of the newly identified systemic lupus erythematosus genetic factors
Arthritis Rheum.
 , 
2009
, vol. 
60
 (pg. 
2558
-
2564
)
14
Sanchez
E.
Comeau
M.E.
Freedman
B.I.
Kelly
J.A.
Kaufman
K.M.
Langefeld
C.D.
Brown
E.E.
Alarcon
G.S.
Kimberly
R.P.
Edberg
J.C.
, et al.  . 
Identification of novel genetic susceptibility loci in African American lupus patients in a candidate gene association study
Arthritis Rheum.
 , 
2011
, vol. 
63
 (pg. 
3493
-
3501
)
15
Warchol
T.
Lianeri
M.
Lacki
J.K.
Olesinska
M.
Jagodzinski
P.P.
ITGAM Arg77His is associated with disease susceptibility, arthritis, and renal symptoms in systemic lupus erythematosus patients from a sample of the Polish population
DNA Cell Biol.
 , 
2011
, vol. 
30
 (pg. 
33
-
38
)
16
Jarvinen
T.M.
Hellquist
A.
Koskenmies
S.
Einarsdottir
E.
Panelius
J.
Hasan
T.
Julkunen
H.
Padyukov
L.
Kvarnstrom
M.
Wahren-Herlenius
M.
, et al.  . 
Polymorphisms of the ITGAM gene confer higher risk of discoid cutaneous than of systemic lupus erythematosus
PLoS One
 , 
2010
, vol. 
5
 pg. 
e14212
 
17
Kim
K.
Brown
E.E.
Choi
C.B.
Alarcon-Riquelme
M.E.
Kelly
J.A.
Glenn
S.B.
Ojwang
J.O.
Adler
A.
Lee
H.S.
Boackle
S.A.
, et al.  . 
Variation in the ICAM1-ICAM4-ICAM5 locus is associated with systemic lupus erythematosus susceptibility in multiple ancestries
Ann Rheum Dis.
 , 
2012
, vol. 
71
 (pg. 
1809
-
1814
)
18
Wang
Z.
Zang
C.
Rosenfeld
J.A.
Schones
D.E.
Barski
A.
Cuddapah
S.
Cui
K.
Roh
T.Y.
Peng
W.
Zhang
M.Q.
, et al.  . 
Combinatorial patterns of histone acetylations and methylations in the human genome
Nat. Genet.
 , 
2008
, vol. 
40
 (pg. 
897
-
903
)
19
Tie
F.
Banerjee
R.
Stratton
C.A.
Prasad-Sinha
J.
Stepanik
V.
Zlobin
A.
Diaz
M.O.
Scacheri
P.C.
Harte
P.J.
CBP-mediated acetylation of histone H3 lysine 27 antagonizes Drosophila Polycomb silencing
Development
 , 
2009
, vol. 
136
 (pg. 
3131
-
3141
)
20
Behmoaras
J.
Bhangal
G.
Smith
J.
McDonald
K.
Mutch
B.
Lai
P.C.
Domin
J.
Game
L.
Salama
A.
Foxwell
B.M.
, et al.  . 
JunD is a determinant of macrophage activation and is associated with glomerulonephritis susceptibility
Nat. Genet.
 , 
2008
, vol. 
40
 (pg. 
553
-
559
)
21
Bain
G.
Murre
C.
The role of E-proteins in B- and T-lymphocyte development
Semin Immunol.
 , 
1998
, vol. 
10
 (pg. 
143
-
153
)
22
Liu
P.
Keller
J.R.
Ortiz
M.
Tessarollo
L.
Rachel
R.A.
Nakamura
T.
Jenkins
N.A.
Copeland
N.G.
Bcl11a is essential for normal lymphoid development
Nat. Immunol.
 , 
2003
, vol. 
4
 (pg. 
525
-
532
)
23
Betz
B.C.
Jordan-Williams
K.L.
Wang
C.
Kang
S.G.
Liao
J.
Logan
M.R.
Kim
C.H.
Taparowsky
E.J.
Batf coordinates multiple aspects of B and T cell function required for normal antibody responses
J. Exp. Med.
 , 
2010
, vol. 
207
 (pg. 
933
-
942
)
24
Wu
Z.
Kim
H.P.
Xue
H.H.
Liu
H.
Zhao
K.
Leonard
W.J.
Interleukin-21 receptor gene induction in human T cells is mediated by T-cell receptor-induced Sp1 activity
Mol. Cell Biol.
 , 
2005
, vol. 
25
 (pg. 
9741
-
9752
)
25
Diamond
M.S.
Garcia-Aguilar
J.
Bickford
J.K.
Corbi
A.L.
Springer
T.A.
The I domain is a major recognition site on the leukocyte integrin Mac-1 (CD11b/CD18) for four distinct adhesion ligands
J. Cell Biol.
 , 
1993
, vol. 
120
 (pg. 
1031
-
1043
)
26
Rhodes
B.
Furnrohr
B.G.
Roberts
A.L.
Tzircotis
G.
Schett
G.
Spector
T.D.
Vyse
T.J.
The rs1143679 (R77H) lupus associated variant of ITGAM (CD11b) impairs complement receptor 3 mediated functions in human monocytes
Ann. Rheum. Dis.
 , 
2012
, vol. 
71
 (pg. 
2028
-
2034
)
27
MacPherson
M.
Lek
H.S.
Prescott
A.
Fagerholm
S.C.
A systemic lupus erythematosus-associated R77H substitution in the CD11b chain of the Mac-1 integrin compromises leukocyte adhesion and phagocytosis
J. Biol. Chem.
 , 
2011
, vol. 
286
 (pg. 
17303
-
17310
)
28
May
R.C.
Machesky
L.M.
Phagocytosis and the actin cytoskeleton
J. Cell Sci.
 , 
2001
, vol. 
114
 (pg. 
1061
-
1077
)
29
Mazzone
A.
Ricevuti
G.
Leukocyte CD11/CD18 integrins: biological and clinical relevance
Haematologica.
 , 
1995
, vol. 
80
 (pg. 
161
-
175
)
30
Magil
A.B.
Frohlich
J.J.
Monocytes and macrophages in focal glomerulosclerosis in Zucker rats
Nephron.
 , 
1991
, vol. 
59
 (pg. 
131
-
138
)
31
Hickstein
D.D.
Ozols
J.
Williams
S.A.
Baenziger
J.U.
Locksley
R.M.
Roth
G.J.
Isolation and characterization of the receptor on human neutrophils that mediates cellular adherence
J. Biol. Chem.
 , 
1987
, vol. 
262
 (pg. 
5576
-
5580
)
32
Graham
I.L.
Gresham
H.D.
Brown
E.J.
An immobile subset of plasma membrane CD11b/CD18 (Mac-1) is involved in phagocytosis of targets recognized by multiple receptors
J. Immunol.
 , 
1989
, vol. 
142
 (pg. 
2352
-
2358
)
33
Lo
S.K.
Van Seventer
G.A.
Levin
S.M.
Wright
S.D.
Two leukocyte receptors (CD11a/CD18 and CD11b/CD18) mediate transient adhesion to endothelium by binding to different ligands
J. Immunol.
 , 
1989
, vol. 
143
 (pg. 
3325
-
3329
)
34
Ohno
M.
Kunimoto
M.
Nishizuka
M.
Osada
S.
Imagawa
M.
Ku proteins function as corepressors to regulate farnesoid X receptor-mediated gene expression
Biochem. Biophys. Res. Commun.
 , 
2009
, vol. 
390
 (pg. 
738
-
742
)
35
Elkon
K.
Autoantibodies in systemic lupus erythematosus
Curr Opin Rheumatol.
 , 
1995
, vol. 
7
 (pg. 
384
-
388
)
36
Molineros
J.E.
Maiti
A.K.
Sun
C.
Looger
L.L.
Han
S.
Kim-Howard
X.
Glenn
S.
Adler
A.
Kelly
J.A.
Niewold
T.B.
, et al.  . 
Admixture mapping in lupus identifies multiple functional variants within IFIH1 associated with apoptosis, inflammation, and autoantibody production
PLoS Genet.
 , 
2013
, vol. 
9
 pg. 
e1003222
 
37
Idogawa
M.
Masutani
M.
Shitashige
M.
Honda
K.
Tokino
T.
Shinomura
Y.
Imai
K.
Hirohashi
S.
Yamada
T.
Ku70 and poly(ADP-ribose) polymerase-1 competitively regulate beta-catenin and T-cell factor-4-mediated gene transactivation: possible linkage of DNA damage recognition and Wnt signaling
Cancer Res.
 , 
2007
, vol. 
67
 (pg. 
911
-
918
)
38
Grote
J.
Konig
S.
Ackermann
D.
Sopalla
C.
Benedyk
M.
Los
M.
Kerkhoff
C.
Identification of poly(ADP-ribose)polymerase-1 and Ku70/Ku80 as transcriptional regulators of S100A9 gene expression
BMC Mol. Biol.
 , 
2006
, vol. 
7
 pg. 
48
 
39
Mayeur
G.L.
Kung
W.J.
Martinez
A.
Izumiya
C.
Chen
D.J.
Kung
H.J.
Ku is a novel transcriptional recycling coactivator of the androgen receptor in prostate cancer cells
J. Biol. Chem.
 , 
2005
, vol. 
280
 (pg. 
10827
-
10833
)
40
Collis
S.J.
DeWeese
T.L.
Jeggo
P.A.
Parker
A.R.
The life and death of DNA-PK
Oncogene.
 , 
2005
, vol. 
24
 (pg. 
949
-
961
)
41
Okuya
K.
Tamura
Y.
Saito
K.
Kutomi
G.
Torigoe
T.
Hirata
K.
Sato
N.
Spatiotemporal regulation of heat shock protein 90-chaperoned self-DNA and CpG-oligodeoxynucleotide for type I IFN induction via targeting to static early endosome
J. Immunol.
 , 
2010
, vol. 
184
 (pg. 
7092
-
7099
)
42
Zheng
B.
Han
M.
Bernier
M.
Wen
J.K.
Nuclear actin and actin-binding proteins in the regulation of transcription and gene expression
FEBS J.
 , 
2009
, vol. 
276
 (pg. 
2669
-
2685
)
43
Csermely
P.
Schnaider
T.
Soti
C.
Prohaszka
Z.
Nardai
G.
The 90-kDa molecular chaperone family: structure, function, and clinical applications. A comprehensive review
Pharmacol Ther.
 , 
1998
, vol. 
79
 (pg. 
129
-
168
)
44
Sakurai
H.
Ota
A.
Regulation of chaperone gene expression by heat shock transcription factor in Saccharomyces cerevisiae: importance in normal cell growth, stress resistance, and longevity
FEBS Lett.
 , 
2011
, vol. 
585
 (pg. 
2744
-
2748
)
45
Powell
E.
Wang
Y.
Shapiro
D.J.
Xu
W.
Differential requirements of Hsp90 and DNA for the formation of estrogen receptor homodimers and heterodimers
J. Biol. Chem.
 , 
2010
, vol. 
285
 (pg. 
16125
-
16134
)
46
DeZwaan
D.C.
Freeman
B.C.
HSP90 manages the ends
Trends Biochem. Sci.
 , 
2010
, vol. 
35
 (pg. 
384
-
391
)
47
Kurdrid
P.
Phuengcharoen
P.
Cheevadhanarak
S.
Tanticharoen
M.
Hongsthong
A.
Identification of a heat shock-responsive cis-acting DNA sequence and its transcriptional regulator: Their roles in the expression of the Spirulina-desD gene in response to heat stress
J. Biosci. Bioeng.
 , 
2010
, vol. 
109
 (pg. 
205
-
210
)
48
Davies
T.H.
Ning
Y.M.
Sanchez
E.R.
A new first step in activation of steroid receptors: hormone-induced switching of FKBP51 and FKBP52 immunophilins
J. Biol. Chem.
 , 
2002
, vol. 
277
 (pg. 
4597
-
4600
)
49
Mandrekar
P.
Catalano
D.
Jeliazkova
V.
Kodys
K.
Alcohol exposure regulates heat shock transcription factor binding and heat shock proteins 70 and 90 in monocytes and macrophages: implication for TNF-alpha regulation
J. Leukoc. Biol.
 , 
2008
, vol. 
84
 (pg. 
1335
-
1345
)
50
Sherer
Y.
Gorstein
A.
Fritzler
M.J.
Shoenfeld
Y.
Autoantibody explosion in systemic lupus erythematosus: more than 100 different antibodies found in SLE patients
Semin Arthritis Rheum.
 , 
2004
, vol. 
34
 (pg. 
501
-
537
)
51
Stephanou
A.
Latchman
D.S.
Isenberg
D.A.
The regulation of heat shock proteins and their role in systemic lupus erythematosus
Semin Arthritis Rheum.
 , 
1998
, vol. 
28
 (pg. 
155
-
162
)
52
Latchman
D.S.
Isenberg
D.A.
The role of hsp90 in SLE
Autoimmunity.
 , 
1994
, vol. 
19
 (pg. 
211
-
218
)
53
Twomey
B.M.
Dhillon
V.B.
McCallum
S.
Isenberg
D.A.
Latchman
D.S.
Elevated levels of the 90 kD heat shock protein in patients with systemic lupus erythematosus are dependent upon enhanced transcription of the hsp90 beta gene
J Autoimmun.
 , 
1993
, vol. 
6
 (pg. 
495
-
506
)
54
Angelo
G.
Lamon-Fava
S.
Sonna
L.A.
Lindauer
M.L.
Wood
R.J.
Heat shock protein 90beta: a novel mediator of vitamin D action
Biochem Biophys Res Commun.
 , 
2008
, vol. 
367
 (pg. 
578
-
583
)
55
Oikonomou
K.G.
Zachou
K.
Dalekos
G.N.
Alpha-actinin: a multidisciplinary protein with important role in B-cell driven autoimmunity
Autoimmun Rev.
 , 
2011
, vol. 
10
 (pg. 
389
-
396
)
56
Antoni
A.
Graham
L.H.
Rauch
J.
Levine
J.S.
Altered cell-cell and cell-matrix interactions in the development of systemic autoimmunity
Autoimmunity.
 , 
2009
, vol. 
42
 (pg. 
278
-
281
)
57
Lluis
M.
Hoe
W.
Schleit
J.
Robertus
J.
Analysis of nucleic acid binding by a recombinant translin-trax complex
Biochem Biophys Res Commun.
 , 
2010
, vol. 
396
 (pg. 
709
-
713
)
58
Peng
S.L.
Transcription factors in the pathogenesis of autoimmunity
Clin. Immunol.
 , 
2004
, vol. 
110
 (pg. 
112
-
123
)
59
Gregersen
P.K.
Olsson
L.M.
Recent advances in the genetics of autoimmune disease
Annu. Rev. Immunol.
 , 
2009
, vol. 
27
 (pg. 
363
-
391
)
60
Kuryłowicz
A.
Nauman
J.
The role of nuclear factor-kappaB in the development of autoimmune diseases: a link between genes and environment
Acta Biochim. Pol.
 , 
2008
, vol. 
55
 (pg. 
629
-
647
)
61
Oxvig
C.
Lu
C.
Springer
T.A.
Conformational changes in tertiary structure near the ligand binding site of an integrin I domain
Proc Natl Acad Sci U S A.
 , 
1999
, vol. 
96
 (pg. 
2215
-
2220
)
62
Rosetti
F.
Tsuboi
N.
Chen
K.
Nishi
H.
Ernandez
T.
Sethi
S.
Croce
K.
Stavrakis
G.
Alcocer-Varela
J.
Gómez-Martin
D.
, et al.  . 
Human lupus serum induces neutrophil-mediated organ damage in mice that is enabled by Mac-1 deficiency
J. Immunol.
 , 
2012
, vol. 
189
 (pg. 
3714
-
3723
)
63
Kamata
T.
Tieu
K.K.
Irie
A.
Springer
T.A.
Takada
Y.
Amino acid residues in the alpha IIb subunit that are critical for ligand binding to integrin alpha IIbbeta 3 are clustered in the beta-propeller model
J. Biol. Chem.
 , 
2001
, vol. 
276
 (pg. 
44275
-
44283
)
64
Shimaoka
M.
Takagi
J.
Springer
T.A.
Conformational regulation of integrin structure and function
Annu. Rev. Biophys. Biomol. Struct.
 , 
2002
, vol. 
31
 (pg. 
485
-
516
)
65
Kanse
S.M.
Matz
R.L.
Preissner
K.T.
Peter
K.
Promotion of leukocyte adhesion by a novel interaction between vitronectin and the beta2 integrin Mac-1 (alphaMbeta2, CD11b/CD18)
Arterioscler Thromb. Vasc. Biol.
 , 
2004
, vol. 
24
 (pg. 
2251
-
2256
)
66
Benimetskaya
L.
Loike
J.D.
Khaled
Z.
Loike
G.
Silverstein
S.C.
Cao
L.
el Khoury
J.
Cai
T.Q.
Stein
C.A.
Mac-1 (CD11b/CD18) is an oligodeoxynucleotide-binding protein
Nat. Med.
 , 
1997
, vol. 
3
 (pg. 
414
-
420
)
67
Mangsbo
S.M.
Sanchez
J.
Anger
K.
Lambris
J.D.
Ekdahl
K.N.
Loskog
A.S.
Nilsson
B.
Totterman
T.H.
Complement activation by CpG in a human whole blood loop system: mechanisms and immunomodulatory effects
J. Immunol.
 , 
2009
, vol. 
183
 (pg. 
6724
-
6732
)
68
Simon
D.I.
Rao
N.K.
Xu
H.
Wei
Y.
Majdic
O.
Ronne
E.
Kobzik
L.
Chapman
H.A.
Mac-1 (CD11b/CD18) and the urokinase receptor (CD87) form a functional unit on monocytic cells
Blood
 , 
1996
, vol. 
88
 (pg. 
3185
-
3194
)
69
Dahlbäck
K.
Löfberg
H.
Dahlbäck
B.
Vitronectin colocalizes with Ig deposits and C9 neoantigen in discoid lupus erythematosus and dermatitis herpetiformis, but not in bullous pemphigoid
Br. J. Dermatol.
 , 
1989
, vol. 
120
 (pg. 
725
-
733
)
70
Spengler
M.I.
Svetaz
M.J.
Leroux
M.B.
Bertoluzzo
S.M.
Carrara
P.
Van Isseldyk
F.
Petrelli
D.
Parente
F.M.
Bosch
P.
Erythrocyte aggregation in patients with systemic lupus erythematosus
Clin. Hemorheol. Microcirc.
 , 
2011
, vol. 
47
 (pg. 
279
-
285
)
71
Hsieh
C.
Chang
A.
Brandt
D.
Guttikonda
R.
Utset
T.O.
Clark
M.R.
Predicting outcomes of lupus nephritis with tubulointerstitial inflammation and scarring
Arthritis Care Res (Hoboken)
 , 
2011
, vol. 
63
 (pg. 
865
-
874
)
72
Sailer
T.
Vormittag
R.
Pabinger
I.
Vukovich
T.
Lehr
S.
Quehenberger
P.
Panzer
S.
Lechner
K.
Zoghlami-Rintelen
C.
Inflammation in patients with lupus anticoagulant and implications for thrombosis
J. Rheumatol.
 , 
2005
, vol. 
32
 (pg. 
462
-
468
)
73
Ekdahl
K.N.
Bengtsson
A.A.
Andersson
J.
Elgue
G.
Rönnblom
L.
Sturfelt
G.
Nilsson
B.
Thrombotic disease in systemic lupus erythematosus is associated with a maintained systemic platelet activation
Br. J. Haematol.
 , 
2004
, vol. 
125
 (pg. 
74
-
78
)
74
de Pablo
P.
Ramírez
A.
Cortina
E.
de la Peña
A.
Zamora
J.
Izaguirre
R.
Amigo
M.C.
Increased fibrin polymerization rate in patients with primary antiphospholipid syndrome and systemic lupus erythematosus
Clin. Appl. Thromb. Hemost.
 , 
2003
, vol. 
9
 (pg. 
221
-
225
)
75
Kahn
M.J.
Hypercoagulability as a cause of stroke in adults
South Med. J.
 , 
2003
, vol. 
96
 (pg. 
350
-
353
)
76
Gong
R.
Liu
Z.
Li
L.
Epistatic effect of plasminogen activator inhibitor 1 and beta-fibrinogen genes on risk of glomerular microthrombosis in lupus nephritis: interaction with environmental/clinical factors
Arthritis Rheum.
 , 
2007
, vol. 
56
 (pg. 
1608
-
1617
)
77
Zhou
Y.
Wu
J.
Kucik
D.F.
White
N.B.
Redden
D.T.
Szalai
A.J.
Bullard
D.C.
Edberg
J.C.
Multiple lupus associated ITGAM variants alter mac-1 function on neutrophils
Arthritis Rheum.
 , 
2013
, vol. 
65
 (pg. 
2907
-
2916
)
78
Nicodemus
K.K.
Catmap: case-control and TDT meta-analysis package
BMC Bioinformatics
 , 
2008
, vol. 
9
 pg. 
130
 
79
Bravo
M.L.
Valenzuela
C.Y.
Arcos-Burgos
O.M.
Polymorphisms and phyletic relationships of the Paisa community from Antioquia (Colombia)
Gene Geogr.
 , 
1996
, vol. 
10
 (pg. 
11
-
17
)
80
Hickstein
D.D.
Grunvald
E.
Shumaker
G.
Baker
D.M.
Back
A.L.
Embree
L.J.
Yee
E.
Gollahon
K.A.
Transfected leukocyte integrin CD11b/CD18 (Mac-1) mediates phorbol ester-activated, homotypic cell:cell adherence in the K562 cell line
Blood
 , 
1993
, vol. 
82
 (pg. 
2537
-
2545
)
81
Klan
N.
Steinhilber
D.
Transient transfection of the human myeloid cell line Mono Mac 6 using electroporation
Biotechniques.
 , 
2003
, vol. 
34
 (pg. 
142
-
147
)
82
Maiti
A.K.
Nath
S.K.
Gene network analysis of small molecules with autoimmune disease associated genes predicts a novel strategy for drug efficacy
Autoimmun Rev.
 , 
2013
, vol. 
12
 (pg. 
510
-
522
)

Supplementary data