Skip to main content
NCBI home page
Immunology logoLink to Immunology
. 2003 Apr;108(4):531–538. doi: 10.1046/j.1365-2567.2003.01579.x

A CR1 polymorphism associated with constitutive erythrocyte CR1 levels affects binding to C4b but not C3b

Daniel J Birmingham *, Wei Chen *, Grace Liang *, Holly C Schmitt *, Katie Gavit *, Haikady N Nagaraja †
PMCID: PMC1782929  PMID: 12667215

Abstract

The erythrocyte type one complement receptor (E-CR1) mediates erythrocyte binding of complement-opsonized immune complexes (IC), and helps protect against random deposition of circulating IC. Two linked CR1 polymorphisms occur in binding domains, at I643T and Q981H. In Caucasians, the variant alleles (643T, 981H) are associated with low constitutive E-CR1 expression levels. This study was conducted to determine if these polymorphisms affect ligand binding, and if so, represent risk factors for the autoimmune IC disease, systemic lupus erythematosus (SLE). In an ELISA comparing relative ligand binding differences, E-CR1 from individuals homozygous for the variant residues (643TT/981HH) exhibited greater binding to C4b, but not C3b, than homozygous wild-type E-CR1. Analysis of single-binding domain CR1 constructs demonstrated that the 981H residue imparted this enhanced C4b binding. No differences were observed in the 981H allele frequency between Caucasian controls (0·170, n = 100) and SLE patients (0·130, n = 150, P = 0·133), or between African American controls (0·169, n = 71) and SLE patients (0·157, n = 67). In a subset of individuals assessed for CR1 size, excluding from this analysis those expressing at least one B allele revealed a trend for over-representation of the 981H allele in Caucasian controls (0·231 frequency, n = 26) versus SLE patients (0·139, n = 83, P = 0·089), but again no difference between African American controls (0·188, n = 24) and SLE patients (0·191, n = 34). These data suggest that the 981H residue compensates for low constitutive expression of E-CR1 in Caucasians by enhancing C4b binding. This may contribute protection against SLE.

Introduction

Primates have a unique pathway for ridding immune complexes (IC) from the circulation that involves the binding of IC by erythrocytes through a process known as immune adherence (reviewed in Birmingham & Hebert1). In humans, this binding is mediated by the type one complement receptor (CR1). Erythrocyte CR1 (E-CR1) binds the complement proteins C1q, C4b, C3b and C3bi, all of which are found on IC that activate the complement system. The binding of complement-activating IC by human erythrocytes allows the erythrocyte to shuttle IC through the circulation until passage through the liver or spleen, at which time the IC are transferred to the monocyte phagocytic system and removed from the circulation. The importance of this IC clearing system has been underscored by studies in humans and non-human primates, demonstrating that reduced immune adherence function results in tissue trapping of circulating IC in tissue such as the skin or kidney.2–6

The extracellular portion of the CR1 molecule is composed entirely of a series of repeating units, known as short consensus repeats (SCRs) or complement control protein repeats (CCPs).7 The SCRs are arranged in tandem groups of seven, known as long homologous repeats (LHRs). Variation of the LHR number constitutes a level of allelic polymorphism, with the most common A size allotype being composed of four LHRs. Of these four LHRs, the three most N-terminal define ligand binding domains8–10 with the first four SCRs in each LHR contributing to ligand binding (i.e. SCRs 1–4, SCRs 8–11, SCRs 15–18). The second most common B size allotype is composed of five LHRs, with the four most N-terminal LHRs defining ligand binding domains.11 The physiological relevance of an extra binding domain in the B size allotype is currently unclear.

Another type of allelic polymorphism occurs with constitutive levels of expression of E-CR1, which can vary by 10-fold among normal individuals. In Caucasians, this regulation is linked to a HindIII RFLP that is found in a CR1 intron.12 Initial analysis suggested that the low expression genotype might be responsible for the low E-CR1 levels that have been observed in SLE patients.13 Subsequent studies, however, indicated that the low E-CR1 levels in SLE reflect an acquired defect that is the result of SLE disease activity (reviewed in Walport and Lachmann14).

Recently, Wilson and colleagues identified a number of CR1 polymorphisms that, in Caucasians, are linked to constitutive E-CR1 expression levels.15,16 Two of these polymorphisms occur in regions of known ligand binding domains, at amino acid I643T (nucleotide T2078C) at the end of SCR 10 in LHR B, and at Q981H (nucleotide G3093T) in SCR 16 in LHR C. This led us to hypothesize that the CR1 encoded by the low expression allele could have ligand binding activity different from CR1 by the high expression allele. The purpose of this study was to test this hypothesis by comparing C3b and C4b binding activities of E-CR1 from individuals homozygous for the residues associated with high constitutive E-CR1 expression (643II/981QQ) or homozygous for the residues associated with low constitutive E-CR1 expression (643TT/981HH). Single-binding domain constructs were also used in transfection experiments to assess the effect of each polymorphism alone. Finally, we determined whether the distribution of these polymorphisms differed between controls and SLE patients.

Materials and methods

Subject recruitment

Two hundred and seventeen SLE patients (150 Caucasian, 67 African American) were recruited from the Division of Nephrology and Division of Rheumatology at The Ohio State University, and met at least four of 11 of the revised 1982 ACR criteria for the diagnosis of SLE. One hundred and seventy-one healthy control subjects (100 Caucasians, 71 African Americans) representing predominantly central Ohio residents were recruited from the OSU medical campus. Ten ml of EDTA blood were drawn from all subjects after signing consent forms, in accordance with an approved OSU Human Subjects IRB protocol.

Blood processing

Total leucocyte RNA was prepared from buffy coat cells isolated from 4 ml of EDTA blood, using RNAzol B according to the manufacturer's instructions (Tel-Test, Friendswood, TX). Leucocyte DNA was prepared from 1 ml of EDTA blood using Puregene according to the manufacturer's instructions (Gentra Systems, Minneapolis, MN).

To prepare soluble erythrocyte membranes containing CR1, EDTA blood was centrifuged through 50% isotonic Percoll (Sigma, St Louis MO), the pelleted erythrocytes were washed with 0·01 m Tris-HCl, pH 7·4 (TB), 0·15 m NaCl (TBS), and then lysed by freeze/thawing in the presence of TBS with 1% general protease inhibitor cocktail (Sigma). Erythrocyte membranes were solubilized in TBS containing 1% NP-40 with 1% protease inhibitors and stored at − 70°.

E-CR1 was partially purified from soluble membrane preparations by cationic exchange chromatography. Erythrocyte membranes were solubilized in TB, 0·02 m NaCl, 1% NP-40, and passed through 1 ml of HiTrap SP HP cationic exchanger (Amersham Biosciences, Piscataway, NJ). After washing the exchanger with TB, 0·02 m NaCl, 1% NP-40, bound CR1 was eluted with TB, 0·25 m NaCl, 1% NP-40 and dialysed against TB, 0·05 m NaCl, 1% NP-40.

Determination of E-CR1 size and expression levels

The CR1 size allotype was determined in a subset of study individuals by Western blot analysis of erythrocyte membranes under nonreducing conditions using the anti-CR1 monoclonal antibody, E11 (Accurate, Westbury, NY), as described previously.17 The average CR1 number per erythrocyte was also determined in a subset of study individuals using 125I-labelled E11 at saturation, as described previously.18

Determination of ligand-binding differences

ELISAs were used to determine relative concentrations of each E-CR1 or CR1 construct preparation, and to quantify differences in ligand binding between the E-CR1 allelic forms (643II/981QQ wild-type and 643TT/981HH variant). E-CR1 concentrations were determined by coating wells in duplicate with either anti-CR1 monoclonal antibody J3D3 (Beckman Coulter, Inc., Fullerton, CA) or 1B4 (gift from John P. Atkinson, Washington University, St Louis, MO), and detecting with either biotinylated E11 (Ancell, Bayport, MN) or a rabbit anti-CR1 antibody followed by biotinylated goat anti-rabbit IgG. The polyclonal anti-CR1 antibody was produced in our laboratory against baboon erythrocyte CR1-like17 and cross-reacts with human CR1. The use of different coating and detecting anti-CR1 antibodies assured both that equal concentrations of E-CR1 or CR1 constructs were compared within each assay, and that the affinities of the anti-CR1 antibodies used in the ELISAs did not differ between wild-type and variant CR1 (Fig. 1).

Figure 1.

Figure 1

Open in a new tab

Relative E-CR1 concentrations as determined using different capture and detecting anti-CR1 antibodies.

Ligand-binding differences between E-CR1 allelic forms were quantified using an ELISA in which wells were coated with C3b (Advanced Research Technologies, San Diego, CA), C4b (Advanced Research Technologies, containing a mix of C4A and C4B gene products), BSA (Sigma) or the anti-CR1 monoclonal antibody J3D3. For solubilized E-CR1, wells were coated in duplicate with C3b at 1·2–2·5 µg/ml or C4b at 2·5–5 µg/ml. For the single binding domain CR1 constructs, wells were coated in triplicate with C3b at 10 µg/ml or C4b at 50 µg/ml. For all assays, BSA was coated at 1 µg/ml and J3D3 was coated at 0·2 µg/ml. After blocking the wells with 3% BSA, 50 µl containing equal amounts of soluble E-CR1 or CR1 construct, diluted in TB, 0·05 m NaCl, 1% BSA, 1% NP-40, pH 7·4, were incubated in the wells at room temperature for 60 min on a constant orbital shaker table. Unbound receptor was removed, the wells were washed twice with TB, 0·05 m NaCl, 0·1% NP-40, pH 7·4, and bound receptor was detected by sequential incubations with 50 µl of 0·8 µg/ml biotinylated E11, 50 µl of strepavidin-HRP (1/1000, Zymed, S. San Francisco, CA) and 200 µl of 0·5 µg/ml OPD substrate (Sigma). Both E11 and strepavidin-HRP were diluted in TB, 0·05 m NaCl, 1% BSA, 1% NP-40. Substrate colour development was halted with 50 µl of 3 m HCl and read at OD490, using the BSA-coated wells as baseline. Each assay (ELISA plate) consisted of a comparison between two E-CR1 preparations (wild-type versus variant) or two CR1 constructs (643-I versus 643-T, 981-Q versus 981-H).

For additional measurements of E-CR1 binding to C4b, ELISAs were performed as above except that wells were coated with purified human IgG, and C4b was deposited on the IgG-coated wells by incubating the wells with 0·02 mg/ml C1 (Sigma) at 37° for 15 min followed by 0·05 mg/ml C4 (Advanced Research Technologies) at 4° for 15 min. Both C1 and C4 were diluted in TBS, 1% BSA, 2 mm MgCl2, 0·3 mm CaCl2. The wells were washed before and after C4 incubation, and the ELISAs were developed to completion as described above. No CR1 binding was observed if either C1 or C4 was omitted. Each assay compared binding among E-CR1 from one soluble erythrocyte membrane preparation (a wild-type) and two preparations (one wild-type, one variant) that were partially purified by cationic exchange chromatography.

CR1 constructs design

To determine which polymorphism affected C4b binding, single binding domain CR1 constructs were designed that expressed either the C3b/C4b binding domain containing the I643T polymorphism at the end of SCR 10 or the C3b/C4b binding domain containing the Q981H polymorphism in SCR 16 (Fig. 2). These constructs were created from a naturally occurring chimpanzee sequence encoding CR1 SCRs 1–6 followed SCR 28, 29, 30 and the transmembrane and cytoplasmic region.19 The cDNA for the chimpanzee sequence, termed CR1a, was cloned previously into the CMV-driven expression vector, pBK-CMV (Stratagene, La Jolla, CA).20 Initially, chimpanzee cDNA sequence for SCRs 1–4 and 28–30 were replaced in the expression vector by human CR1 sequence SCRs 1–4 and SCRs 28–30 using polymerase chain reaction (PCR)-amplified human CR1 cDNA and restriction enzyme sites common to both the human and chimpanzee CR1 sequence. Human SCRs 1–4 were then replaced by human sequence for SCR 8–11 (for the I643T polymorphism) or SCRs 15–18 (for the Q981H polymorphism) that was amplified by PCR from leucocyte cDNA from an individual genotyped as 643IT/981QH. Isolated clones were fully sequenced to identify plasmids with SCR 8–11 containing either 643I or 643T, and plasmids with SCR 15–18 containing either 981Q or 981H, and to insure 100% homology with the expected sequence.

Figure 2.

Figure 2

Open in a new tab

Single binding domain CR1 constructs, containing SCRs 8–11 for the analysis of the I643T polymorphism, or SCRs 15–18 for the analysis of the Q981H polymorphism.

Transfection of CR1 constructs

Plasmids were transfected into COS-7 cells (ATCC, Manassas, VA) using Lipofectamine (Life Technologies) and COS-7 membranes containing the receptor constructs were solubilized as described previously.17 Western blot analysis using E11 confirmed the expression of the single-binding receptor constructs (data not shown).

CR1 Q981H genotype determination

For determining Q981H polymorphism frequencies, encoded by the G3093T polymorphism, genomic DNA was amplified using a forward primer specific for the intron upstream from the 5′ exon of SCR 16 (16iF4, GCTACATGCACGTTGAGACCTTAC) and a reverse primer specific for the intron downstream from the 5′ exon for SCR 16 (16iR6, AGCAAGCATACAGATTTTCCCC). Amplification yielded a 366-bp fragment containing a BstNI cleavage site at nucleotide 54 in all samples, and at residue 145 in sequence containing the wild-type 3093G residue. Thus, amplification and digestion of genomic DNA containing the wild-type 3093G residue yielded fragments of 54 bp, 91 bp and 221 bp (Fig. 3). Amplification and digestion of genomic DNA containing the 3093T variant yielded fragments of 54 and 312 bp.

Figure 3.

Figure 3

Open in a new tab

Ethidium bromide-stained gel following PCR amplification and BstNI digestion for determining G3093T genotype.

Statistical analysis

ELISA data were analysed for significance using two-way analysis of variance models with multiple observations per cell, and the E-CR1 preparations and the assays were used as the two factors. Residuals were examined for the validity of the model assumptions. Significance of the factor effects were checked using appropriate F-tests. Single specific pairwise comparisons were carried out using two-tailed t-tests (Figs 4 and 5), and multiple comparisons among all preparations were carried out using Tukey's honest significant difference test, with a 0·05 level of significance (Fig. 6). Mean effects of the factors were estimated using least squares estimates and the associated standard errors were estimated using the relevant analysis of variance model. Statistical analyses were carried out using SAS JMP Version 4 software (SAS Institute, Cary, NC).

Figure 4.

Figure 4

Open in a new tab

Results of ELISA for measuring binding E-CR1 from solubilized erythrocyte membranes from two individuals who were homozygous wild-type (WT; 643II/981QQ) or homozygous variant (Var; 643TT/981HH). Soluble erythrocyte membranes containing CR1 were incubated in parallel in wells coated with J3D3, C3b or C4b, and bound CR1 were detected using biotinylated E11, followed by strepavidin-HRP. A single WT and Var sample were assayed together in each ELISA. For assays comparing WT1 versus Var1 (n = 4, d.f. = 11), C3b was coated at 1·2 µg/ml and C4b was coated at 2·5 µg/ml. For assays comparing WT2 versus Var2 (n = 3, d.f. = 8), C3b was coated at 2·5 µg/ml and C4b was coated at 5 µg/ml. Mean and standard error are shown.

Figure 5.

Figure 5

Open in a new tab

Results of ELISA comparing the binding of partially purified wild-type (WT1) or variant (Var1) E-CR1 or solubilized erythrocyte membranes containing a second wild-type E-CR1 (WT2), to C4b deposited on IgG-coated wells (n = 4, df = (2,18)). Mean and standard error are shown.

Figure 6.

Figure 6

Open in a new tab

Results of ELISA comparing the binding of 643-I and 643-T constructs (n = 2, df = 9), or the 981-Q and 981-H constructs (n = 3, df = 14) to C3b and C4b. Mean and standard error are shown.

Differences in average E-CR1 levels (Fig. 7) were determined by t-tests between 981QQ and 981QH genotypes, or between controls and SLE patients within a CR1 genotype. The 981HH genotype was not analysed statistically due to low sample numbers.

Figure 7.

Figure 7

Open in a new tab

Average E-CR1 levels in a subset of race-matched controls and SLE patients. (a)Average E-CR1 levels measured in normal control Caucasians and African Americans grouped by Q981H genotype. For Caucasians, n = 25 QQ, 10 QH, 2 HH. For African Americans, n = 23 QQ, 11 QH. Comparison between E-CR1 levels of controls (same as (a)) and SLE patients, according to Q981H genotype. For Caucasian SLE patients, n = 74 QQ, 19 QH, 1 HH. For African American SLE patients, n = 18 QQ, 9 QH. Mean and standard error are shown. **P < 0·005 for Caucasian QQ versus QH (a) or controls versus SLE patients (b).

Genotype and allele frequency differences for the Q981H polymorphism were tested statistically by a 2 × 2 Fisher's exact test.

Results

Effect of I643T/Q981H on ligand binding

To identify initially normal individuals whose CR1 contained residues 643II/981QQ (also termed wild-type) or 643TT/981HH (also termed variant), primer pairs were used to amplify two CR1 cDNA segments, one from SCR 7–13 and one from SCR 13–19. After purifying and partially sequencing the amplified cDNA, a number of controls were identified as 643II/981QQ, and two controls were identified as 643TT/981HH. Both individuals homozygous for the variant CR1 (Var1 and Var2) were also homozygous for the common A size CR1 allotype. Two wild-type CR1 controls (WT1 and WT2) were chosen who were also homozygous for the A size allele, and the amplified CR1 cDNA segments from these two wild-types and from the two variants were sequenced between SCR 8–11 and from SCR 15–18. No other polymorphisms were identified. Blood samples were taken from these four individuals, and CR1 was solubilized from erythrocyte membranes and assessed for binding to C3b and C4b.

An ELISA was used to assess differences in E-CR1 binding to C3b and C4b. One wild-type and one variant sample were analysed in parallel in each assay, and equal E-CR1 concentrations were confirmed in each assay by including wells coated with an anti-CR1 capture antibody (J3D3). As can be seen in Fig. 4, there were no differences between wild-type and variant E-CR1 in the level of binding to C3b. In contrast, in both sets of parallel comparisons, the variant E-CR1 exhibited greater binding to C4b than did wild-type E-CR1.

To further assess the difference between wild-type and variant CR1, an ELISA was used in which C4b was physiologically deposited on IgG through the action of C1. In this experiment, wild-type and variant E-CR1 preparations partially purified by cation exchange chromatography were assessed along with a solubilized erythrocyte membrane preparation containing a different wild-type E-CR1. The results of these assays (Fig. 5) show no difference between the two different wild-type CR1 preparations and again demonstrate that the variant E-CR1 exhibits greater binding to C4b than does wild-type E-CR1.

Ligand binding by CR1 constructs

To determine which of the two polymorphic residues imparts the higher C4b binding activity, CR1 constructs were made that contained single binding domains, defined either by SCRs 8–11 (for the I643T polymorphism) or SCRs 15–18 (for the Q981H polymorphism). Following transient expression in COS-7 cells, the single binding domain CR1 constructs were solubilized from the COS-7 membranes and the binding activities were compared. As shown in Fig. 6, there was no difference in binding to C3b or C4b between the constructs 643-I and 643-T. In contrast, while no difference was noted in binding C3b between the 981-Q and 981-H constructs, the 981-H construct exhibited greater binding to C4b. Thus, the higher binding to C4b displayed by the full-length CR1 variant was due to the histidine present at residue 981 in SCR 16.

CR1 Q981H frequency determination

To test for an association of between the Q981H polymorphism (nucleotide G3093T) and SLE, a Caucasian cohort of SLE patients (n = 150) and controls (n = 100) were assessed for the Q981H polymorphism frequency. For comparison, a group of African American SLE patients (n = 67) and controls (n = 71) were also genotyped. The results of this analysis in all of the controls and SLE patients, stratified by race, are shown in Table 1. All genotype frequencies were in excellent agreement with the Hardy–Weinberg prediction. No differences were observed in genotype or allele frequency between controls and SLE patients, regardless of the race.

Table 1.

Distribution of Q981H amino acid polymorphism (nucleotide G3093T) in SLE patients and race-matched controls. P-values determined by Fisher's exact test for significant differences in allele frequency between controls and SLE patients within the same race

Genotype Allele frequency
QQ QH HH Q H P
Caucasian
″Control 69 28 3 0·830 0·170
″SLE 113 35 2 0·870 0·130 0·133
African American
″Control 48 22 1 0·831 0·169
″SLE 47 19 1 0·843 0·157 0·455
Open in a new tab

The analysis shown in Table 1 potentially includes individuals expressing all four of the CR1 size alleles (A, B, C or D). The B and D alleles express the most C4b binding domains (4 and 5, respectively), a trait that could minimize any physiological effect of differential C4b binding imparted by the Q981H polymorphism. To gauge this effect, the Q981H frequency analysis was repeated in a subset of individuals from each cohort in whom CR1 size genotypes were determined at the time of recruitment. Specifically, individuals expressing at least one B allele were excluded (no D alleles were found), resulting in a subset with a 96% A allele frequency and 4% C allele frequency. Analysis of the Q981H frequencies in this subset, shown in Table 2, demonstrated a trend towards an increase of the 981H allele in Caucasian controls (0·231) compared to Caucasian SLE patients (0·139, P = 0·089). The 981H allele frequency was again no different between African American controls (0·188) and SLE patients (0·191).

Table 2.

Distribution of Q981H amino acid polymorphism in SLE patients and race-matched controls expressing the A or C size allele. The CR1 size determination was done in a subset of individuals in whom E-CR1 genotyping was performed at recruitment. P-values were determined by Fisher's exact test for significant differences in allele frequency between controls and SLE patients within the same race

Genotype Allele frequency
QQ QH HH Q H P
Caucasian
″Control 16 8 2 0·769 0·231
″SLE 61 21 1 0·861 0·139 0·089
African American
″Control 16 7 1 0·813 0·188
″SLE 21 13 0 0·809 0·191 0·579
Open in a new tab

E-CR1 levels

Average E-CR1 levels were determined for some of the study individuals upon entry into the study. Comparing E-CR1 levels within the normal control populations confirmed the association of the 981H allele with low constitutive E-CR1 expression in the Caucasian population but not in the African American population, as expected (Fig. 7a). Average E-CR1 levels were higher in the controls compared to SLE patients in both the Caucasians (493 ± 41, n = 35 versus 408 ± 24, n = 93: P = 0·022) and African Americans (570 ± 27, n = 34 versus 334 ± 26, n = 27: P < 0·0001), confirming previous observations. Interestingly, when the E-CR1 levels were compared between controls and SLE patients within each Q981H genotype (Fig. 7b), differences were observed in Caucasians only in the 981QQ genotype. In contrast, in African Americans, SLE patients exhibited lower E-CR1 levels compared to controls in both the 981QQ and 981QH genotype groups.

Discussion

The range of average E-CR1 expression in the normal population is roughly 10-fold, with absolute values dependent on the antibody used to enumerate receptor number. Recent reports have identified a number of polymorphic residues that are linked to the low constitutive expression CR1 allele in Caucasians,15,16 including two occurring in ligand binding domains. Thus, the CR1 protein that occurs with low constitutive erythrocyte expression in Caucasians is different in primary sequence from the CR1 protein that occurs with high expression. This study was performed to determine if ligand binding differences also exist between these two CR1 allelic forms. The results from this study demonstrate that the low expression allele encodes a CR1 with greater binding activity for C4b than the high expression allele, and that this difference in binding is due to an histidine (versus a glutamine) present at amino acid residue 981 in SCR 16.

The observation that the change of glutamine to histidine at residue 981 in SCR 16 affects ligand binding is not surprising, as this is a significant change from a neutral residue to a charged residue in SCR 16 that is part of the binding domain for both C3b and C4b within LHR C. Studies by Krych et al. have identified numerous residues within a virtually identical domain in LHR B that affect both C3b and C4b binding activity, and residues that affect C3b or C4b preferentially.9,21,22 A common theme from this work is that increasing the positive charge of the ligand binding domain increases binding affinity for C3b and C4b. The present study is in agreement with this theme, as the change from a Q to an H increases the positive charge at this site. The fact that the change to H increases the binding only for C4b indicates that residue 981 appears to be part of the binding domain in LHR C that is specific for C4b.

The role that charge plays in the effect that the 981H residue has on C4b binding suggests that the charge characteristic of C4b is also relevant to the 981H effect. Two different genes exist for C4, one encoding an acidic form (C4A) and one encoding a basic form (C4B), with multiple allotypes of each.23 The thioester bond of C4A preferentially reacts with amino groups and appears to be the predominant form that binds immune complexes24,25 although whether this predominance occurs under physiological conditions is not clear.26 Perhaps more relevant to the handling of circulating IC is the observation that C4b formed from the C4A gene product (C4Ab) binds to CR1 with higher affinity than C4Bb.27,28 The charge effect of the 981H residue would probably enhance the binding of CR1 to the negatively charged C4Ab, while having a minimal or opposite effect on C4Bb. The specific allotype of C4Ab and C4Bb might also influence the binding effect of the Q981H polymorphism.

The association of the 981H residue with low constitutive expression of E-CR1 in Caucasians, and its effect on C4b binding, suggest that in Caucasians the 981H residue may compensate for low E-CR1 levels in the interaction between erythrocytes and IC in the circulation. Consequently, we hypothesized that the 981H allele might protect against IC-mediated pathologies. Our data did not support this hypothesis when the SLE patients were compared to race-matched controls (Table 1). However, this comparison did not take into account the CR1 size alleles. The two most common alleles, the A and B alleles (together accounting for >95% of the size allele frequency) differ in that the A allele encodes one less binding domain. The C allele, occurring in less than 5% of our study population, encodes two less binding domains. As additional domains could theoretically override any physiological effect of the Q981H polymorphism, we re-analysed the frequency data in a subset of individuals in whom CR1 size was determined and excluded those who expressed at least one B allele, leaving just individuals with A or C size allele expression. This analysis revealed a trend for over-representation of the 981H allele in Caucasian controls (0·231 frequency) versus Caucasian SLE patients (0·139, P = 0·089), suggesting that the higher affinity 981H allotype is associated with protection against SLE in Caucasians expressing CR1 with limited numbers of ligand binding domains. In contrast to the Caucasian cohorts, the 981H frequencies between African American controls (0·188) and African American SLE patients (0·191) were strikingly similar.

Stratification of the Q981H frequency data by CR1 size reduced substantially the number of individuals analysed in this study, and these data need to be expanded to confirm the association between the 981H allele combined with the A or C size allele and protection against SLE in Caucasians. This association, if valid, has a number of interesting implications. First, the requirement for a specific combination of the CR1 Q981H polymorphism and the CR1 size polymorphism to be a risk factor for SLE would reflect the complex nature of SLE, where multiple genetic (and environmental) factors contribute to disease susceptibility, and susceptibility requires the occurrence of specific combinations of these factors.

Secondly, the fact that this association is restricted to the Caucasian population would suggest that the role that the Q981H polymorphism plays as a risk factor is related to constitutive E-CR1 expression. An acquired loss of measurable E-CR1 occurs in SLE patients29 and our data reflect this, as low-E-CR1 levels were demonstrated in SLE patients in both Caucasians and African Americans. However, when control and SLE patient E-CR1 levels were compared in those with the 981QH genotype (and probably the 981HH genotype), no differences were found between Caucasian controls and SLE patients, suggesting that E-CR1 were not loss from Caucasian SLE patients expressing at least one 981H allele. Why no difference is observed is currently unclear. There may be a critical expression level below which no measurable E-CR1 loss occurs, and the constitutive E-CR1 levels in Caucasians expressing at least one H allele could be below this critical level. Alternatively, the H allotype could provide a stabilizing influence against overall E-CR1 loss, an explanation that would require an interaction between the H and Q allotype. Regardless, the data suggest that if the 981H allele provides protective compensation for acquired E-CR1 loss, it would be associated with protection against SLE in African Americans. This clearly is not the case.

Perhaps the most intriguing implication of an association of the Q981H polymorphism with SLE susceptibility is that this susceptibility would be related specifically to CR1/C4b interactions. This is somewhat surprising, as our initial hypothesis concerning the Q981H polymorphism was based on the role of E-CR1 in clearing IC. In this pathway, C3b is thought to be the major physiological ligand for the interaction between E-CR1 and complement opsonized IC, because more C3b is probably deposited than C4b on any given activation surface, and the affinity of CR1 for C3b is higher than for C4b.30,31 The apparent physiological effect of the Q981H polymorphism would suggest that CR1/C4b interactions are uniquely important in protecting against SLE, and not replaceable by CR1/C3b interactions. This unique connection with C4b would parallel the observations that C4 deficiencies, but not C3 deficiencies, are associated predominantly with SLE,32,33 suggesting that defective IC–C4b interaction with E-CR1 is one mechanism behind this association.

Acknowledgments

The authors wish to thank Dr C. Yung Yu for helpful discussions. This work was supported by NIH grants AI41729 (DJB), DK55546 (DJB) and MO1 RR-00034 (GCRC).

Abbreviations

CCP

complement control protein repeat

CR1

type one complement receptor

E-CR1

erythrocyte CR1

IC

immune complexes

LHR

long homologous repeat

SCR

short consensus repeat

SLE

systemic lupus erythematosus

References

  • 1.Birmingham DJ, Hebert LA. CR1 and CR1-like: the primate immune adherence receptors. Immunol Rev. 2001;180:100–11. doi: 10.1034/j.1600-065x.2001.1800109.x. [DOI] [PubMed] [Google Scholar]
  • 2.Waxman FJ, Hebert LH, Smead WL, Van Aman ME, Kraut EH, Birmingham DJ, Taguim JM. Complement depletion accelerates the clearance of immune complexes from the circulation of primates. J Clin Invest. 1984;74:1329–40. doi: 10.1172/JCI111543. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 3.Waxman FJ, Hebert LA, Cosio FG, Smead WL, VanAman ME, Taguiam JM, Birmingham DJ. Differential binding of immunoglobulin A and immunoglobulin G1 immune complexes to primate erythrocytes in vivo. J Clin Invest. 1986;77:82–9. doi: 10.1172/JCI112306. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 4.Schifferli JA, Ng YC, Paccaud JP, Walport MJ. The role of hypocomplementaemia and low erythrocyte complement receptor type 1 numbers in determining abnormal immune complex clearance in humans. Clin Exp Immunol. 1989;75:329–35. [PMC free article] [PubMed] [Google Scholar]
  • 5.Madi N, Steiger G, Estreicher J, Schifferli JA. Immune adherence and clearance of hepatitis B surface Ag/Ab complexes is abnormal in patients with systemic lupus erythematosus (SLE) Clin Exp Immunol. 1991;85:373–8. doi: 10.1111/j.1365-2249.1991.tb05734.x. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 6.Davies KA, Peters AM, Beynon HLC, Walport MJ. Immune complex processing in patients with systemic lupus erythematosus in in vivo imaging and clearance studies. J Clin Invest. 1992. pp. 2075–83. [DOI] [PMC free article] [PubMed]
  • 7.Klickstein LB, Wong WW, Smith JA, Weis JH, Wilson JG, Fearon DT. Human C3b/C4b receptor (CR1). Demonstration of long homologous repeating domains that are composed of the short consensus repeats characteristics of C3/C4 binding proteins. J Exp Med. 1987;165:1095–112. doi: 10.1084/jem.165.4.1095. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 8.Klickstein LB, Bartow TJ, Miletic V, Rabson LD, Smith JA, Fearon DT. Identification of distinct C3b and C4b recognition sites in the human C3b/C4b receptor (CR1, CD35) by deletion mutagenesis. J Exp Med. 1988;168:1699–717. doi: 10.1084/jem.168.5.1699. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 9.Krych M, Clemenza L, Howdeshell D, Hauhart R, Hourcade D, Atkinson JP. Analysis of the functional domains of complement receptor type 1 (C3b/C4b receptor; CD35) by substitution mutagenesis. J Biol Chem. 1994;269:13273–8. [PubMed] [Google Scholar]
  • 10.Kalli KR, Hsu PH, Bartow TJ, Ahearn JM, Matsumoto AK, Klickstein LB, Fearon DT. Mapping of the C3b-binding site of CR1 and construction of a (CR1)2-F(ab′)2 chimeric complement inhibitor. J Exp Med. 1991;174:1451–60. doi: 10.1084/jem.174.6.1451. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 11.Wong WW, Cahill JM, Rosen MD, et al. Structure of the human CR1 gene. Molecular basis of the structural and quantitative polymorphisms and identification of a new CR1-like allele. J Exp Med. 1989;169:847–63. doi: 10.1084/jem.169.3.847. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 12.Wilson JG, Murphy EE, Wong WW, Klickstein LB, Weis JH, Fearon DT. Identification of a restriction fragment length polymorphism by a CR1 cDNA that correlates with the number of CR1 on erythrocytes. J Exp Med. 1986;164:50–9. doi: 10.1084/jem.164.1.50. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 13.Wilson JG, Wong WW, Murphy E, Ed Schur PH, Fearon DT. Deficiency of the C3b/C4b receptor (CR1) of erythrocytes in systemic lupus erythematosus: analysis of the stability of the defect and of a restriction fragment length polymorphism of the CR1 gene. J Immunol. 1987;138:2706–10. [PubMed] [Google Scholar]
  • 14.Walport MJ, Lachmann PJ. Erythrocyte complement receptor type 1, immune complexes, and the rheumatic diseases. Arth Rheum. 1988;31:153–8. doi: 10.1002/art.1780310201. [DOI] [PubMed] [Google Scholar]
  • 15.Herrera AH, Xiang L, Martin SG, Lewis J, Wilson JG. Analysis of complement receptor type 1 (CR1) expression on erythrocytes and of CR1 allelic markers in Caucasian and African American populations. Clin Immunol Immunopathol. 1998;87:176–83. doi: 10.1006/clin.1998.4529. [DOI] [PubMed] [Google Scholar]
  • 16.Xiang L, Rundles JR, Hamilton DR, Wilson JG. Quantitative alleles of CR1: coding sequence analysis and comparison of haplotypes in two ethnic groups. J Immunol. 1999;163:4939–45. [PubMed] [Google Scholar]
  • 17.Birmingham DJ, Logar CM, Shen X-P, Chen W. The baboon erythrocyte complement receptor is a glycophosphatidylinositol-linked protein encoded by a homologue of the human CR1-like genetic element. J Immunol. 1996;157:2586–92. [PubMed] [Google Scholar]
  • 18.Birmingham DJ, Cosio FG. Characterization of the baboon erythrocyte C3b-binding protein. J Immunol. 1989;142:3140–4. [PubMed] [Google Scholar]
  • 19.Birmingham D, Shen X-P, Hourcade D, Nickells M, Atkinson J. Primary sequence of an alternatively spliced form of CR1: a candidate for the 75,000 Mr complement receptor expressed on chimpanzee erythrocytes. J Immunol. 1994;153:691–700. [PubMed] [Google Scholar]
  • 20.Chen W, Cm L, Shen X, Birmingham D. The chimpanzee and cynomolgus monkey erythrocyte immune adherence receptors are encoded by CR1-like genes. Immunogenet. 2000;52:46–52. doi: 10.1007/s002510000250. [DOI] [PubMed] [Google Scholar]
  • 21.Krych M, Hourcade D, Atkinson JP. Sites within the complement C3b/C4b receptor important for the specificity of ligand binding. Proc Natl Acad Sci USA. 1991;88:4353–7. doi: 10.1073/pnas.88.10.4353. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 22.Krych M, Hauhart R, Atkinson JP. Structure-function analysis of the active sites of complement receptor type 1. J Biol Chem. 1998;273:8623–9. doi: 10.1074/jbc.273.15.8623. [DOI] [PubMed] [Google Scholar]
  • 23.Blanchong C, Zhou B, Rupert K, et al. Deficiencies of human complement component C4A and C4B and heterozygosity in length variants of RP-C4-CYP21-TNX (RCCX) modules in Caucasians. The load of RCCX genetic diversity on major histocompatibility complex-associated disease. J Exp Med. 2000;191:2183–96. doi: 10.1084/jem.191.12.2183. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 24.Law SKA, Dodds AW, Porter RR. A comparison of the properties of two classes, C4A and C4B, of the human complement component C4. EMBO J. 1984;3:1819–23. doi: 10.1002/j.1460-2075.1984.tb02052.x. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 25.Kishore N, Shah D, Skanes VM, Levine RP. The fluid-phase binding of human C4 and its genetic variants, C4A3 and C4B1, to immunoglobulins. Mol Immunol. 1988;25:811–9. doi: 10.1016/0161-5890(88)90117-4. [DOI] [PubMed] [Google Scholar]
  • 26.Reilly BD. Analysis of human C4A and C4B binding to an immune complex in serum. Clin Exp Immunol. 1999;117:12–8. doi: 10.1046/j.1365-2249.1999.00940.x. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 27.Gatenby PA, Barbosa JE, Lachmann PJ. Differences between C4A and C4B in the handling of immune complexes: the enhancement of CR1 binding is more important than the inhibition of immunoprecipitation. Clin Exp Immunol. 1990;79:158–63. doi: 10.1111/j.1365-2249.1990.tb05172.x. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 28.Reilly BD, Mold C. Quantitative analysis of C4Ab and C4Bb binding to the C3b/C4b receptor (CR1, CD35) Clin Exp Immunol. 1997;110:310–6. doi: 10.1111/j.1365-2249.1997.tb08333.x. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 29.Ross GD, Yount WJ, Walport MJ, et al. Disease-associated loss of erythrocyte complement receptors (CR1, C3b receptors) in patients with systemic lupus erythematosus and other diseases involving autoantibodies and/or complement activation. J Immunol. 1985;135:2005–14. [PubMed] [Google Scholar]
  • 30.Cooper N. Immune adherence by the fourth component of complement. Science. 1969;165:396–8. doi: 10.1126/science.165.3891.396. [DOI] [PubMed] [Google Scholar]
  • 31.Ross GD, Newman SL, Lambris JD, Devery PJ, Cain JA, Lachmann PJ. Generation of three different fragments of bound C3 with purified factor I or serum. II. Location of binding sites in the C3 fragments for factors B and H, complement receptors, and bovine conglutinin. J Exp Med. 1983;158:334–52. doi: 10.1084/jem.158.2.334. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 32.Morgan B, Walport M. Complement deficiency and disease. Immunol Today. 1991;12:301–6. doi: 10.1016/0167-5699(91)90003-C. [DOI] [PubMed] [Google Scholar]
  • 33.Sullivan KE. Complement deficiency and autoimmunity. Curr Opin Pediatr. 1998;10:600–6. doi: 10.1097/00008480-199810060-00011. [DOI] [PubMed] [Google Scholar]

Articles from Immunology are provided here courtesy of British Society for Immunology

RESOURCES