Evidence That Deletion at FCGR3B Is a Risk Factor for Systemic Sclerosis

ORIGINAL ARTICLE Evidence that deletion at FCGR3B is a risk factor for systemic sclerosis

C McKinney1, JCA Broen2, MC Vonk2, L Beretta3, R Hesselstrand4, N Hunzelmann5, G Riemekasten6, R Scorza3, CP Simeon7, V Fonollosa7, PE Carreira8, N Ortego-Centeno9, MA Gonzalez-Gay10, P Airo11, M Coenen12, J Martin13,15, TRDJ Radstake2,14,15 and TR Merriman1,15

There is increasing evidence that gene copy number (CN) variation influences clinical phenotype. The low-affinity Fc receptor 3B (FCGR3B) located in the FCGR gene cluster is a CN polymorphic gene involved in the recruitment of polymorphonuclear neutrophils to sites of inflammation and their activation. Given the genetic overlap between systemic lupus erythematosus and systemic sclerosis (SSc) and the strong evidence for FCGR3B CN in the pathology of SLE, we hypothesised that FCGR3B gene dosage influences susceptibility to SSc. We obtained FCGR3B deletion status in 777 European Caucasian cases and 1000 controls. There was an inverse relationship between FCGR3B CN and disease susceptibility. CN of p1 was a significant risk factor for SSc (OR ¼ 1.55 (1.13–2.14), P ¼ 0.007) relative to CNX2. Although requiring replication, these results suggest that impaired immune complex clearance arising from FCGR3B deficiency contributes to the pathology of SSc, and FCGR3B CN variation is a common risk factor for systemic autoimmunity.

Genes and Immunity (2012) 13, 458–460; doi:10.1038/gene.2012.15; published online 3 May 2012 Keywords: systemic sclerosis; FCGR3B; copy number; polymorphism; association

INTRODUCTION that deletion of FCGR3B is a risk factor in comparison with diploid Systemic sclerosis (SSc) is an uncommon autoimmune disease, and increased CN. in which abnormalities in the immune and vascular system lead to a variety of dermatological and systemic abnormalities. Although it differs from systemic autoimmune diseases such as RESULTS systemic lupus erythematosus (SLE) in several respects (the CN assignment, aided by neutrophil antigen typing on a subset of inflammatory infiltrates are milder, and neither immunoglobulin 56 samples selected because they were positioned near the CN deposits nor glomerulonephritis is a feature of SSc), a number of 1–2 boundary on histograms (Supplementary Figure 1a), showed loci associated with SSc appear to be common risk factors for that 11.3% of the SSc sample set carried a single-copy deletion of several autoimmune connective tissue and rheumatic disorders, FCGR3B compared with 7.6% of controls, with one SSc patient including PTPN22 rs2476601, STAT4 rs7574864 and IRF5 rs2004640.1 having a null FCGR3B genotype. (This patient also has Sjo¨gren’s Fc gamma receptor 3B (FcgR3B) is the most abundant syndrome with arthritis, which is possibly of clinical significance neutrophil binding site for polymeric IgG and immune complexes given a report of association of FCGR3B deletion with Sjo¨gren’s (IC), and is involved in both the recruitment of polymorphonuclear syndrome.14) The FCGR3B CN allele distribution among the neutrophils (PMN) to sites of inflammation and the clearance of controls (CNo2 ¼ 7.6%, CN 2 ¼ 82.6% and CN42 ¼ 9.8%) is circulating IC. Copy number (CN) variation in FCGR3B has been similar to that observed in other healthy Caucasian populations associated with a variety of autoimmune disorders, including as determined by multiplex ligand-dependent probe amplification 2–4 5 rheumatoid arthritis, ulcerative colitis and glomerular or paralog ratio testing (CNo2 ¼ 6.7–7.8%, CN 2 ¼ 81.1–82.2% and disease.6,7 In particular, CNo2 has been reported to be a risk CN42 ¼ 9.4–11.8%).8,12,13,15,16 7–13 factor for SLE in a number of studies. Given the evidence for The influence of possessing o2 copies of FCGR3B on SSc risk shared genetic susceptibility between SSc and SLE, we was tested under the hypothesis that any association of FCGR3B hypothesised that copy-number variation in FCGR3B has a role with disease would be similar to that evident in SLE, where CNo2 in the pathology of SSc. In order to evaluate this hypothesis, we is a risk factor.5–10 There was significant evidence that possessing tested for association between FCGR3B gene dosage and SSc in a fewer than two copies of FCGR3B is a risk factor for SSc (odds ratio Caucasian case–control sample set under the specific hypothesis (OR) ¼ 1.55 (1.13–2.14), P ¼ 0.007) (Table 1). This risk was not

1Department of Biochemistry, University of Otago, Dunedin, New Zealand; 2Department of Rheumatology, Radboud University Nijmegen Medical Center, Nijmegen, The Netherlands; 3Referral Center for Systemic Autoimmune Diseases, University of Milan, Milan, Italy; 4Department of Rheumatology, Lund University, Lund, Sweden; 5Department of Dermatology, University of Cologne, Cologne, Germany; 6Department of Rheumatology and Clinical Immunology, Charite´ University Hospital and German Rheumatism Research Centre, Leibniz, Germany; 7Internal Medicine Department, Hospital Valld0Hebron, Barcelona, Spain; 8Rheumatology Division, Hospital 12 de Octubre, Madrid, Spain; 9Internal Medicine Department, Hospital Universitario San Cecilio, Granada, Spain; 10Rheumatology Division, Hospital Universitario Marques de Valdecilla, IFIMAV, Santander, Spain; 11Servizio di Reumatologia ed Immunologia Clinica, Spedali Civili, Brescia, Italia; 12Department of Human Genetics, Radboud University Nijmegen Medical Center, Nijmegen, The Netherlands; 13Instituto de Parasitologı´a y Biomedicina Lo´pez-Neyra, CSIC, Granada, Spain and 14Department of Rheumatology and Clinical Immunology, Utrecht Medical Center, Utrecht, The Netherlands. Correspondence: Dr TR Merriman, Department of Biochemistry, University of Otago, 710 Cumberland Street, Dunedin, Otago 9001, New Zealand. E-mail: [email protected] 15These authors contributed equally to this work. Received 26 October 2011; revised 3 April 2012; accepted 4 April 2012; published online 3 May 2012 SSc and deletion at FCGR3B C McKinney et al 459 Table 1. Testing for association of integer values of FCGR3B CN with SSc

Method CN Case (frequency) Control (frequency) OR (95% CI)a P

CopyCaller 450% conﬁdence o2 88 (0.117) 78 (0.078) 1.58 (1.15–2.18) 0.005 X2 661 (0.883) 927 (0.922) Standard curve o2 88 (0.113) 76 (0.076) 1.55 (1.13–2.14) 0.007 X2 689 (0.887) 924 (0.924)

Abbreviations: CN, copy number; OR, odds ratio; SSc, systemic sclerosis. aInclusion of country of origin as a covariate in linear regression gave OR ¼ 1.91 (1.36–2.69), P ¼ 2 Â 10 À 4 for CopyCaller (450% conﬁdence) and OR ¼ 1.89 (1.34–2.67), P ¼ 3 Â 10 À 4 for the standard curve method.

diminished after accounting for country of origin in the as a partial validation of cluster analysis, as presence of hetero- association analysis (Table 1). zygosity in individuals assigned as CN ¼ 1 would indicate We tested for association between FCGR3B CNo2 and limited misidentification. cutaneous or diffuse cutaneous SSc, and the presence or absence Another potential source of error is the presence of SNPs within of anti-topoisomerase I and anti-centromere antibodies. There was the amplification primers; the forward primer was originally no significant association of CNo2 with any of these stratifications designed so that the two bases at the 30 end of the forward (Supplementary Table 1). primer differ between FCGR3B and the closely related FCGR3A.2 The 1000 Genomes project has recently identified two SNPs at these 30 positions (rs115802971 and rs115130696) in individuals of DISCUSSION Yoruban descent in which the minor allele corresponds to the Our results provide evidence that deletion of FCGR3B is associated FCGR3A sequence. Although an individual carrying these minor with SSc (OR ¼ 1.55, P ¼ 0.007). This is the first report in SSc and is alleles would be falsely determined to have a deletion, the allele consistent with the association reported for SLE, where published frequencies in Caucasian populations are very low, with the minor ORs range from 1.47 to 1.65.7–13 Although the P value is modest allele of these SNPs being observed in only one of the first 380 (0.007), this likely represents genuine association given the prior Caucasians sequenced in the 1000 Genomes project. evidence supporting a role for FCGR3B deletion in SSc. Membrane- Given the complexity of the FCGR locus and the fact that bound FcgR3B is involved in the adherence of neutrophils to deletion of FCGR3B can extend to FCGR2C and/or FCGR2B,11,15 it is endothelial cells at sites of infection as well as in IC uptake. The also possible that the apparent association with FCGR3B is in fact soluble form of the receptor (sFcgR3B) released from activated caused by variations in one of these other genes. Although PMN also binds IC, preventing these complexes being taken up by FCGR2C has been associated with idiopathic thrombocytopenic other, activating Fc receptors. FcgR3B expression, PMN adhesion, purpura,18 CN variation is unlikely to have a major biological IC formation and sFcgR3B levels are all correlated with FCGR3B effect; in most individuals the gene contains a premature stop CN.12 FCGR3B deletion is therefore likely to decrease recruitment codon. It is the presence of a SNP (found in B18% of the of PMN to sites of inflammation as well as reduce IC sequestration population) within this stop codon that creates an open reading and clearance. It has been hypothesised that the primary effect of frame and results in the expression of this inhibitory receptor that deletion in the FCGR3B gene is a reduced uptake of IgG appears to be the aetiological variant. Despite the presence of complexes, leading to tissue deposition of IC.5,6,12,15 Our results, linkage disequilibrium between FCGR3B CN and a functional however, suggest that this may not be the only explanation for polymorphism within FCGR2B that abolishes its inhibitory capacity, the observed associations. Although SLE and SSc are characterised the effects of FCGR3B deletion and homozygosity for the by the presence of autoantibodies, the absence of IgG deposits in inactivated FcgR2B allele are independently associated with SSc is one of the major differences between SLE and SSc.17 This SLE.11 This suggests that FCGR3B CN is also likely to contribute suggests that decreased FcgR3B expression leading to the to disease risk independently of an accompanying deletion in reduced clearance of circulating IC containing autoantibodies FCGR2B. may contribute to systemic inflammation, with tissue-specific In conclusion, the results presented here suggest that effects if IC deposition occurs. Although the role of autoantibodies FCGR3B CN influences susceptibility to SSc. Future work should in the pathogenesis of SSc is a subject of debate, it is known that focus on development of more efficient approaches for measuring toll-like receptor-mediated production of type 1 interferons by CN at FCGR3B, replication of this association and investigation of plasmacytoid dendritic cells is a characteristic feature of SLE and a association in clinically-defined subgroups. Given the role of this similar interferon signature has been reported in SSc patients.17 receptor in the uptake of IC and the fact that impaired clearance This observation is consistent with a model whereby FCGR3B of autoantibodies is strongly implicated in the pathogenesis of deficiency decreases IC uptake by PMN, leading to higher levels of SLE,19 it is possible that a similar deficiency contributes to the circulating IC and enhanced autoantibody stimulation of development of SSc. This association also adds to the list of plasmacytoid dendritic cells. diseases in which FCGR3B CN has been implicated, suggesting that Although it is possible that the association is an artefact FCGR3B may be a common autoimmune-related locus. of experimental methodology, we are confident that this is not the case. The quantitative PCR assay method we have used has 2 been validated previously. The use of an internal reference in MATERIALS AND METHODS the PCR removes variations due to amplification efficiency and Clinical material experimental conditions that can occur when test and reference genes are measured in separate wells. Cases and controls were All the patients fulfilled the 1980 American College of Rheumato- logy (ACR) classification criteria for SSc.20 The local ethical committee also assayed concurrently, and a-posteriori assignment of CN from each centre approved the study. Both patients and controls were rather than arbitrary rounding of non-integer values was done included in the study after obtaining written informed consent. All patients to minimise batch effects and systematic errors. Furthermore, included in this study were classified as having limited cutaneous (LcSSc) all plates where the cutoff between CN ¼ 1 and CN ¼ 2 was or diffuse cutaneous SSc (DcSSc) using the criteria postulated by LeRoy unclear were excluded from the analysis. NA typing was also used et al.21 Patients with scleroderma changes limited to the skin distal to the

& 2012 Macmillan Publishers Limited Genes and Immunity (2012) 458 – 460 SSc and deletion at FCGR3B C McKinney et al 460 elbows and/or knees, regardless of facial involvement, fulfil the definition REFERENCES for LcSSc. Those SSc patients with more proximal scleroderma skin 1 Allanore Y, Dieude P, Boileau C. Genetic background of systemic sclerosis: auto- changes were classified as having DcSSc. The presence of pulmonary immune genes take centre stage. Rheumatology 2010; 49: 203–210. fibrosis was investigated by a high-resolution computed tomography scan. 2 McKinney C, Fanciulli M, Merriman ME, Phipps-Green A, Alizadeh BZ, Koeleman BP Restrictive syndrome and diffusion capacity of the lungs was defined as a et al. Association of variation in Fcgamma receptor 3B gene copy number forced vital capacity o75% of the predicted value, and the diffusion with rheumatoid arthritis in Caucasian samples. Ann Rheum Dis 2010; 69: capacity of the lung for carbon monoxide was defined as o75% of that 1711–1716. predicted. Controls were derived from the same country and region as the 3 Graf SW, Lester S, Nossent H, Hill CL, Proudman S, Lee A et al. Low copy number patients originated from and included 723 previously genotyped samples of the FCGR3B gene and rheumatoid arthritis: a case-control study and meta- 2 from the Netherlands that exhibited good clustering. Clinical details of the analysis. Arthritis Res Ther 2012; 14: R28. SSc patients are given in Supplementary Table 2. Collection of blood 4 Robinson JI, Carr IM, Cooper DL, Rashid LH, Martin SG, Emery P et al. Confirmation samples and clinical information from case and control subjects was of association of FCGR3B but not FCGR3A copy number with susceptibility undertaken with informed consent and relevant ethical review board to autoantibody positive rheumatoid arthritis. Hum Mutat 2012; 33: approval from each contributing centre in accordance with the tenets of 741–749. the Declaration of Helsinki. 5 Asano K, Matsushita T, Umeno J, Hosono N, Takahashi A, Kawaguchi T et al. A genome-wide association study identifies three new susceptibility loci for Determination of FCGR3B copy number ulcerative colitis in the Japanese population. Nat Genet 2009; 41: 1325–1329. FCGR3B CN was measured using Taqman Real-Time PCR as described 6 Aitman TJ, Dong R, Vyse TJ, Norsworthy PJ, Johnson MD, Smith J et al. Copy previously, with detection probe 50-CCTCAGCTTTCCC-30.2 To minimise DNA number polymorphism in Fcgr3 predisposes to glomerulonephritis in rats and degradation (which can increase the risk of false-positive calls22), lysophilised humans. Nature 2006; 439: 851–855. samples were first assayed within 24 h of resuspension and stored at À 20 1C 7 Fanciulli M, Norsworthy PJ, Petretto E, Dong R, Harper L, Kamesh L et al. FCGR3B between experiments. All samples were assayed on two separate occasions copy number variation is associated with susceptibility to systemic, but not in duplicate on each occasion and the raw CN values averaged, with an organ-specific, autoimmunity. Nat Genet 2007; 39: 721–723. overall average difference of 0.21 copies between duplicates. A maximum of 8 Mamtani M, Anaya JM, He W, Ahuja SK. Association of copy number variation 183 samples were assayed simultaneously in 384-well plates. Genotyping was in the FCGR3B gene with risk of autoimmune diseases. Genes Immun 2010; 11: successful for 90.1% (1284/1425) of cases and 89.4% (1218/1362) of controls. 155–160. Two approaches were used to convert continuous CN estimates to integers. 9 Molokhia M, Fanciulli M, Petretto E, Patrick AL, McKeigue P, Roberts AL et al. FCGR3B copy number variation is associated with systemic lupus erythematosus In the first approach, DDCT analysis was carried out using Applied Biosystems CopyCaller V2.0 software with the assumption that the most common CN risk in Afro-Caribbeans. Rheumatology 2011; 50: 1206–1210. estimate was 2, and by discarding samples where bin assignment was done 10 Morris DL, Roberts AL, Witherden AS, Tarzi R, Barros P, Whittaker JC et al. Evidence with o50% confidence. The second approach was based on the standard for both copy number and allelic (NA1/NA2) risk at the FCGR3B locus in systemic curve method described previously,2 whereby the distribution of calculated lupus erythematosus. Eur J Hum Genet 2010; 18: 1027–1031. CN for each experiment was examined by displaying them histographically, 11 Niederer HA, Willcocks LC, Rayner TF, Yang W, Lau YL, Williams TN et al. Copy with an average of 80 samples assayed in the same batch. Each histogram number, linkage disequilibrium and disease association in the FCGR locus. was drawn in a manner blinded to the origin of the samples and the cutoff Hum Mol Genet 2010; 19: 3282–3294. between CN ¼ 1andCN41 independently determined by two people based 12 Willcocks LC, Lyons PA, Clatworthy MR, Robinson JI, Yang W, Newland SA et al. on the clustering of values and the location of the first-identified NA Copy number of FCGR3B, which is associated with systemic lupus erythematosus, heterozygote (Supplementary Figure 1a). Samples exhibiting discrepancy correlates with protein expression and immune complex uptake. J Exp Med 2008; with confident (450%) CopyCaller assignments were excluded. Ten plates 205: 1573–1582. (containing 507 cases and 218 controls) were excluded from the analysis 13 McKinney C, Merriman TR. Meta-analysis confirms a role for deletion in FCGR3B in because of unclear clustering between CN ¼ 1andCN¼ 2 (Supplementary autoimmune phenotypes. Hum Mol Genet 2012; 21: 2370–2376. Figure 1b). 14 Nossent JC, Rischmueller M, Becker-Merok A, Lester S. Low copy number of Fcgamma receptor 3B gene is a disease susceptibility and severity factor in 0 NA typing primary Sjogren s syndrome. Arthritis Rheum 2011; 63: S389. 15 Breunis WB, van Mirre E, Geissler J, Laddach N, Wolbink G, van der Schoot E et al. When sufficient DNA was available for standard PCR to be performed, the Copy number variation at the FCGR locus includes FCGR3A, FCGR2C and FCGR3B functional NA1 and NA2 alleles of FCGR3B were genotyped over samples but not FCGR2A and FCGR2B. Hum Mutat 2009; 30: E640–E650. spanning the boundary between CN ¼ 1 and CN ¼ 2 using conventional 16 Marques RB, Thabet MM, White SJ, Houwing-Duistermaat JJ, Bakker AM, Hendriks 23 PCR and gel visualisation using an established assay. GJ et al. Genetic variation of the Fc gamma receptor 3B gene and association with rheumatoid arthritis. PLoS One 2010; 5: e13173. Data analysis 17 Katsumoto TR, Whitfield ML, Connolly MK. The pathogenesis of systemic sclerosis. Subjects were grouped as those with FCGR3B deletion and were compared Annu Rev Pathol 2011; 6: 509–537. with the reference group of those having two or more copies of FCGR3B. 18 Breunis WB, van Mirre E, Bruin M, Geissler J, de Boer M, Peters M et al. Copy Chi-square analysis was used to compare the distribution of FCGR3B CN number variation of the activating FCGR2C gene predisposes to idiopathic between patients and controls after CN assignment. Logistic regression thrombocytopenic purpura. Blood 2008; 111: 1029–1038. analysis was used to measure the SSc-associated risks of low FCGR3B CN as 19 Davies KA, Peters AM, Beynon HL, Walport MJ. Immune complex processing the OR relative to CNX2 using the STATA 8.0 statistical package. in patients with systemic lupus erythematosus. In vivo imaging and clearance studies. J Clin Invest 1992; 90: 2075–8203. 20 Subcommittee for scleroderma criteria of the American Rheumatism Association CONFLICT OF INTEREST Diagnostic and Therapeutic Criteria Committee. Preliminary criteria for the classification of systemic sclerosis (scleroderma). Arthritis Rheum 1980; 23: 581–590. The authors declare no conflict of interest. 21 LeRoy EC, Black C, Fleischmajer R, Jablonska S, Krieg T, Medsger Jr TA et al. Scleroderma (systemic sclerosis): classification, subsets and pathogenesis. J Rheumatol 1988; 15: 202–205. ACKNOWLEDGEMENTS 22 Cukier HN, Pericak-Vance MA, Gilbert JR, Hedges DJ. Sample degradation leads to The Health Research Council of NZ and NZ Lottery Health are thanked for false-positive copy number variation calls in multiplex real-time polymerase chain financial support. JM was funded by grant SAF2009-11110 from the Spanish reaction assays. Anal Biochem 2009; 386: 288–290. Ministry of Science, CTS-4977 from Junta de Andalucıá, Spain, by Redes Tema´ticas 23 Morgan AW, Barrett JH, Griffiths B, Subramanian D, Robinson JI, Keyte VH et al. de Investigacioń Cooperativa Sanitaria Programme, RD08/0075 (RIER) from Instituto Analysis of Fcgamma receptor haplotypes in rheumatoid arthritis: FCGR3A de Salud Carlos III (ISCIII), Spain, and by Fondo Europeo de Desarrollo Regional remains a major susceptibility gene at this locus, with an additional contribution (FEDER). from FCGR3B. Arthritis Res Ther 2006; 8:R5.

Supplementary Information accompanies the paper on Genes and Immunity website (http://www.nature.com/gene)

Genes and Immunity (2012) 458 – 460 & 2012 Macmillan Publishers Limited