http://orcid.org/0000-0002-2692-5365
ABSTRACT
Introduction: Immune thrombocytopenia (ITP) is an autoimmune blood disease of unknown etiology. The aim of our study was to investigate a possible role of FCGR2A and FCGR3A polymorphisms in the development of primary ITP.
Methods: We analyzed 125 adult patients with ITP and 120 healthy controls. Genotyping was performed by using PCR-RFLP methods.
Results: Our results showed significantly higher frequency of high-affinity FCGR3A-158V allele in patients with ITP compared with control subjects (47.2% versus 37.5%; p = 0.037). We did not find significant differences in the genotype distribution or allele frequencies for FCGR2A-131H/R between patients and controls, p = 0.652 and p = 0.478. In the groups of patients with unresponsive and responsive ITP we found significantly different genotype distribution and allele frequencies for FCGR3A, p = 0.036 and p = 0.008 respectively. There was no significant difference in genotype and allele frequencies for FCGR2A between these two groups of patients. Our results confirmed that the combination of high-affinity FCGR2A-131H and FCGR3A-158V allele was more common in patients with ITP than in controls (55% versus 40%; p = 0.024).
Conclusion: Our results suggest possible role of FCGR3A polymorphism in the etiology, development and clinical outcome of ITP, but larger prospective studies are needed to confirm these results.
Introduction
Immune thrombocytopenia, also known as idiopathic thrombocytopenic purpura (ITP), is an autoimmune condition in which defects in immune self-tolerance lead to humoral and cellular abnormal responses comprising auto-antibody production and cytotoxic effects [Citation1]. These immunological abnormalities are responsible for increased platelet destruction as well as decreased megakaryopoiesis and thrombopoiesis [Citation2–4], both leading to a thrombocytopenic state.
A large number of studies established the crucial role of auto-antibodies in ITP pathogenesis, demonstrating that main platelet antigenic targets are the fibrinogen receptor glycoprotein complex (GP) IIb-IIIa and the von Willebrand receptor GPIb/IX [Citation2,Citation5] while a lower proportion of auto-antibodies react with the collagen receptors GPIa/IIa and GPIV [Citation6]. Antibody-bound-platelet phagocytosis by the reticuloendothelial system is the primary pathogenic mechanism by which auto-antibodies induce thrombocytopenia, although lysis mediated by complement activation on antibody-bound platelets seems to have also a role in ITP too [Citation2,Citation7]. Apart from auto-antibodies, another mechanism involving direct T-cell mediated cytotoxicity was also shown to participate in platelet destruction [Citation8].
Previous studies have assessed the contribution of platelet apoptosis to ITP pathogenesis. Platelet apoptosis was first demonstrated in an animal model of ITP, in which injection of anti-GPIIb antibodies triggered features of platelet apoptosis, in murine platelets [Citation9]. Concerning human ITP, evidence of platelet apoptosis, including caspase 3, 8 and 9 activations, was shown in children with acute ITP, which was ameliorated by intravenous immunoglobulin infusion [Citation10].
The etiology of ITP remains unclear, but both genetic and environmental factors are thought to play role in the development of the disease. Several genes involved in immune system regulation like cytokine genes [Citation11–14], Fc gamma receptor genes [Citation12,Citation15], CTLA-4 gene [Citation16,Citation17] and HLA genes [Citation18] have been associated with susceptibility to ITP in several studies.
Fc receptors are GPs and members of immunoglobulin superfamily of molecules. They are found on many different cells (neutrophils, macrophages, lymphocytes, platelets) and form a critical link between the humoral and cellular immune responses. Three different families of FcγR exist: FcγRI, FcγRII and FcγRIII and they are diverse in both their structure and function [Citation19,Citation20]. FcγRI has a strong affinity for monomeric IgG, while FcγRII and FcγRIII will only bind effectively to IgG in the form of immune complexes. FcγRII class is encoded by three genes (IIA, IIB and IIC) and FcγRIII is encoded by two genes (IIIA and IIIB). Genes for Fcγ Receptors are located on the long arm of chromosome 1. The gene for FcγRIIA (FCGR2A) is the only FcγR expressed on platelets. FCGR2A is polymorphic and has two codominantly expressed alleles, FCGR2a-H131 and FCGR2A-R131. This polymorphic variation of FCGR2A is due to a single base substitution at nucleotide position 494. Nucleotide adenine (A) is substituted for guanine (G) and this results in a change of amino acid 131 histidine (H131) to arginine (R131). This polymorphism influences the affinity of the receptor. The allele H131 is characterized by a high affinity for human IgG2, whereas the other allele 131 arginine (R131) has a low affinity for human IgG2 [Citation21]. The gene for FcγRIIIA (FCGR3A) is mainly expressed on mononuclear phagocytes and has two polymorphic variant alleles: 158 valine (V158) and phenylalanine (F158) due to a single base substitution of timidine to guanine (T to G) at nucleotide 559. This polymorphism is located in the proximal membrane domain (EC2) that influences ligand binding [Citation22,Citation23]. FCGR3A-158V allele variant has higher affinity for Fc fragment of IgG1 and IgG3 than 158F variant. These Fc gamma receptor polymorphisms may influence antibody-mediated phagocytosis and antigen presentation activity. These variations have been recently investigated in different autoimmune diseases like systemic lupus erythematosus (SLE) [Citation24,Citation25], multiple sclerosis [Citation26], Addison disease [Citation27], heparin-induced thrombocytopenia [Citation28,Citation29] and they have been shown to be involved in disease susceptibility.
In this study, we examined the possible role and involvement of FCGR2A and FCGR3A polymorphisms in the development of primary ITP.
Materials and Methods
We analyzed 125 unrelated adult patients with primary chronic ITP (35 men and 90 women) with median age of 47 (range 18–83 years old) and 120 healthy, age- and sex-matched controls (). Since ethnicity can influence gene polymorphisms we have analyzed only patients and controls from the same ethnicity. Both, patients and controls in this study were all Macedonian orthodox of Caucasian origin. The diagnosis of ITP was based on thrombocytopenia (platelet count < 100 × 109/l), normal or increased bone marrow megakaryocytes without morphological elements of dysplasia, and exclusion of other diseases that could be a cause for thrombocytopenia like underlying autoimmune disorders like SLE, increased platelet destruction due to enlarged spleen, hepatitis C, HIV [Citation30]. The median follow up of the patients was 44 months (12–384 months). Clinical information recorded for every patient included: sex, age at diagnosis, platelet count at diagnosis and last control, bleeding symptoms, treatment modality and response to treatment ().
Table 1. Patients’ and controls characteristics.
Table 2. Patients’ characteristics.
Thirty-three patients (26.4%) were asymptomatic, 82 (65.5%) had minor skin or mucosal bleeding and 14/125 (11.2%) had bleeding from gastrointestinal or genitourinary system. None of the patients had severe, life threatening bleeding symptoms. The median platelet count in patients with ITP at diagnosis was 13 × 109/l (range: 0–98 × 109/l) and 178 × 109/l (range: 2–687 × 109/l) at the last follow up visit. Numbers of patients with severe, moderate or mild thrombocytopenia at diagnosis and at the last control are shown in . Treatment modalities and responses are also summarized in . Corticosteroids were initial treatment for 95% of patients, 38 patients were splenectomized, 22 patients were treated with intravenous gamma globulins (IVIG), while five patients did not received any specific treatment due to mild, asymptomatic thrombocytopenia. Complete response according to the criteria by Rodeghiero et al. [Citation30] to initial treatment (prednisone ± splenectomy) was achieved in 38% of patients, partial response in 52% and no response in 12/120 (10%) of patients.
Refractory ITP was defined as a failure to achieve at least a response or as a loss of response after splenectomy [Citation3,Citation30]. Fourteen patients had refractory form of ITP, while 28 patients had ITP unresponsive to one or more agents. ‘Unresponsive ITP patients’ are unsplenectomized patients not responding to medical treatments. In total, 42 (34%) patients had unresponsive or refractory form of ITP according to new definition criteria by Rodighiero et al. [Citation30].
DNA was isolated from peripheral blood mononuclear cells with standard phenol–chloroform extraction. Genotyping was performed by using polymerase chain reaction (PCR) and restriction fragment length polymorphism (RFLP) method already published [Citation31,Citation32]. Written informed consent was obtained from all participants.
The distribution of genotypes and allele frequencies were compared between patients and controls using a chi-squared test. Differences were considered significant if p < 0.05.
Results
We did not find significant differences in the genotype distribution or allele frequencies for FCGR2A-131H/R between patients with ITP and controls, p = 0.652 and p = 0.478 (). Our results did not find significantly different distribution of the FCGR3A genotypes in patients with ITP comparing with healthy controls, p = 0.133 (). Allele frequencies for FCGR3A-158F/V were significantly different in patients with ITP (F allele 52.8%, V allele 47.2%) compared with controls (F allele 62.5%, V allele 37.5%), p = 0.037 with Yates correction ().
Table 3. Genotype distributions and allele frequencies for FCGR2A and FCGR3A in patients with ITP and control subjects.
We divided all ITP patients into two groups: those with refractory or unresponsive form of ITP and those with ITP responsive to initial treatment and compared these two polymorphisms in these two groups of patients. We found significantly different genotype distribution and allele frequencies for FCGR3A between patients with unresponsive and responsive ITP, p = 0.036 and p = 0.008 respectively (). There was no significant difference in genotype distribution and allele frequencies for FCGR2A between these two groups of patients ().
Table 4. Genotype distributions and allele frequencies for FCGR2A and FCGR3A in patients with responsive and unresponsive ITP patients.
We did not find significant differences in genotype distribution and allele frequencies for both gene polymorphisms between splenectomized and unsplenectomized ITP patients ().
Table 5. Genotype distributions and allele frequencies for FCGR2A and FCGR3A in splenectomized and unsplenectomized ITP patients.
Knowing that the combination of different polymorphic variants may have a synergistic effect, we compared patients that had the combination of at least one high-affinity allele variants (FCGR2A-131H or FCGR3A-158V) with control subjects (). The combination of FCGR2A-131H and FCGR3A-158V allele was more common in patients with ITP than in control subjects (55% versus 40%; p = 0.024). At the same time, the combination of low-affinity alleles FCGR2-131R and FCGR3A-158F was less common in ITP patients than in control subjects (50.4% versus 70%; p = 0.027).
Table 6. Combined FCGR2A and FCGR3A phenotypes.
Discussion
Over the last two decades, a number of groups have investigated the frequencies and clinical significance of variant alleles in FCGR2A and FCGR3A in several disease populations [Citation33]. FCGR2A polymorphism has been associated with different diseases like heparin-induced thrombocytopenia [Citation24,Citation25], SLE [Citation28,Citation29,Citation34], bacterial infection [Citation35] and response to rituximab-based treatment in lymphoma patients [Citation36,Citation37]. Our results did not confirm significant difference in genotype distribution or allele frequencies for FCGR2A polymorphisms in patients with ITP and control subjects. We also could not find such a difference between patients with responsive or unresponsive ITP. Such a difference was not found in the groups of splenectomized and unsplenectomized patients. Our results confirmed significantly higher allele frequencies for FCGR3A-158V in patients with ITP comparing to controls (V allele 47.2% versus 37.5%), p = 0.037. We also found significantly different genotype distribution and allele frequencies for FCGR3A between patients with unresponsive and responsive ITP, p = 0.036 and p = 0.008 respectively.
Our results are similar to the previously reported results by Foster et al. and Carcao et al. [Citation12,Citation32] considering the frequency of FCGR3A variant alleles. They found higher frequencies of high-affinity allele FCGR32-158V in patients with ITP compared to control subjects [Citation12,Citation32]. Contrary to our results, Carcao et al. [Citation32] also reported significant over-representation of the homozygous FCGR2A-131H/H genotype among ITP patients compared with the control subjects. Similar to our results, Fujimoto et al. [Citation38] did not find different allele or genotype frequencies for FCGR2A in ITP patients, but for FCGR3A polymorphism they found significantly lower frequency of FCGR3A-158F/F genotype in ITP patients (p < 0.005). Furthermore, they reported that FCGR3A-158V/V genotype was more common in patients with complete remission to immunosuppressive treatment (60% versus 10%). These results indicate that FCGR3A polymorphism, especially V allele predicts poor response to treatment in patients with ITP. Results from the study of Williams et al. [Citation39] are opposite from our results. They found higher frequency of FCGR2A-131R allele in patients with adult refractory ITP, while we did not find significant difference in the distribution of FCGR2A genotypes and allele frequencies in patients with different form of ITP.
We found that the combination of high-affinity alleles of both FCGR (FCGR2A-131H and FCGR3A-158V) were more common in ITP patients than in control subjects. But this finding has its limitation because we could not know whether the increase in high-affinity alleles for both polymorphisms is caused by the more common appearance of V allele of FCGR3A and limited number of analyzed cases.
It is well known that Fc gama Receptor polymorphic variants have different affinities for different subclasses of immunoglobulins, resulting in variable clearance of immune complexes, antigen presentation or antibody-mediated phagocytosis, which may have role in some autoimmune diseases. An over-representation of the high-affinity binding alleles might result in increased clearance of IgG-sensitized platelets, which is the most important mechanism of thrombocytopenia in ITP patients. But different studies have shown different, sometimes opposite results for genotype or allele frequencies for FCGR2A and FCGR3A for the same autoimmune disease. These differences in the results from various studies could be explained with different allele frequencies in different populations, small number of analyzed patients and selection of different groups of patients.
Limitation of our study is the fact that not all FCGR polymorphisms were studied in our cohort of patients and controls. We did not analyze the FCGR2C-ORF allele, which was associated with ITP in the study reported by Breunis et al. [Citation15]. They reported significantly different prevalence of an open reading frame of FCGR2C, in healthy individuals compared with the ITP patients (18% versus 34.4% in ITP patients; p < 0.009). FCGR2C acts as an activating IgG receptor that exerts antibody-mediated cellular cytotoxicity by immune cells. Therefore, Breunis et al. [Citation15] propose that the activating FCGR2C-ORF genotype predisposes to ITP by altering the balance of activating and inhibitory FCGR on immune cells.
Another FCGR polymorphism that could be associated with ITP and is not analyzed in our study is FCGR2B-I232T. FCGR2B-I232T was identified by Kyogoku et al. [Citation40] as a single nucleotide polymorphism in FcγRIIB gene, coding for a nonsynonymous substitution within the transmembrane domain, Ile232Thr (I232T). Kyogoku et al. [Citation40] demonstrated that FCGR2B-232T/T genotype was significantly increased in Japanese patients with SLE.
In conclusion, our results suggest possible role of FCGR3A polymorphism in the etiology, development and clinical outcome of ITP, but further larger prospective studies of these and other FCGR polymorphisms are needed to confirm these results.
Disclosure statement
No potential conflict of interest was reported by the authors.
ORCID
Marica Pavkovic http://orcid.org/0000-0002-2692-5365
References
- McKenzie CG, Guo L, Freedman J, et al. Cellular immune dysfunction in immune thrombocytopenia (ITP). Br J Haematol. 2013;163(1):10–23. doi: 10.1111/bjh.12480
- McMillan R. Chronic idiopathic thrombocytopenic purpura. N Eng J Med. 1981;304(19):1135–1147. doi: 10.1056/NEJM198105073041904
- Cines DB, Blanchette VS. Immune thrombocytopenic purpura. N Eng J Med. 2002;346(13):995–1008. doi: 10.1056/NEJMra010501
- Lev PR, Grodzielski M, Goette NP, et al. Impaired proplatelet formation in immune thrombocytopenia: a novel mechanism contributing to decreased platelet count. Br J Haematol. 2014;165(6):854–864. doi: 10.1111/bjh.12832
- Woods VLJr, Oh EH, Mason D, et al. Autoantibodies against the platelet glycoprotein IIb/IIIa complex in patients with chronic ITP. Blood. 1984;63(2):368–375.
- He R, Reid DM, Jones CE, et al. Spectrum of Ig classes, specificities, and titers of serum antiglycoproteins in chronic idiopathic thrombocytopenic purpura. Blood. 1994;83(4):1024–1032.
- Najaoui A, Bakchoul T, Stoy J, et al. Autoantibody-mediated complement activation on platelets is a common finding in patients with immune thrombocytopenic purpura (ITP). Eur J Haematol. 2012;88(2):167–174. doi: 10.1111/j.1600-0609.2011.01718.x
- Zhao C, Li X, Zhang F, et al. Increased cytotoxic T-lymphocyte-mediated cytotoxicity predominant in patients with idiopathic thrombocytopenic purpura without platelet autoantibodies. Haematologica. 2008;93(9):1428–1430. doi: 10.3324/haematol.12889.
- Leytin V, Mykhaylov S, Starkey AF, et al. Intravenous immunoglobulin inhibits anti-IIb-induced platelet apoptosis in a murine model of immune thrombocytopenia. Br J Haematol. 2006;133(1):78–82.
- Winkler J, Kroiss S, Rand M, et al. Platelet apoptosis in paediatric immune thrombocytopenia is ameliorated by intravenous immunoglobulin. Br J Haematol. 2012;156(4):508–515. doi: 10.1111/j.1365-2141.2011.08973.x
- Satoh T, Pandey JP, Okazaki Y, et al. Single nucleotide polymorphisms of the inflammatory cytokine genes in adults with chronic immune thrombocytopenic purpura. Br J Haematol. 2004;124(6):796–801. doi: 10.1111/j.1365-2141.2004.04843.x
- Foster CB, Zhu S, Erichsen HC, et al. Polymorphisms in inflammatory cytokines and Fcγ receptors in childhood chronic immune thrombocytopenic purpura: a pilot study. Br J Haematol. 2001;113(3):596–599. doi: 10.1046/j.1365-2141.2001.02807.x
- Rocha AM, DeSouza C, Rocha GA, et al. Research paper: IL1RN VNTR and IL2−330 polymorphic genes are independently associated with chronic immune thrombocytopenia. Br J Haematolol. 2010;150(6):679–684. doi: 10.1111/j.1365-2141.2010.08318.x
- Zhao H, Du W, Gu D, et al. DNMT3B 579g>T promoter polymorphism and the risk for idiopathic thrombocytopenic purpura in a Chinese population. Acta Haematol. 2009;122(1):31–35. doi: 10.1159/000235616
- Breunis WB, van Mirre E, Bruin M, et al. Copy number variation of the activating FCGR2C gene predisposes to idiopathic thrombocytopenic purpura. Blood. 2007;111(3):1029–1038. doi: 10.1182/blood-2007-03-079913
- Pavkovic M, Georgievski B, Cevreska L, et al. CTLA-4 Exon 1 polymorphism in patients with autoimmune blood disorders. Am J Hematol. 2003;72(2):147–149. doi: 10.1002/ajh.10278
- Li H, Ge J, Zhao H, et al. Association of cytotoxic T-lymphocyte antigen 4 gene polymorphisms with idiopathic thrombocytopenic purpura in a Chinese population. Platelets. 2011;22(1):37–42. doi: 10.3109/09537104.2010.521601
- Kuwana M, Kaburaki J, Pandey JP, et al. HLA class II alleles in Japanese patients with immune thrombocytopenic purpura. Associations with anti-platelet glycoprotein autoantibodies and responses to splenectomy. Tissue Antigens. 2000;56(4):337–343. doi: 10.1034/j.1399-0039.2000.560405.x
- Van der Pol W, van de Winkel JGJ. Igg receptor polymorphisms: risk factors for disease. Immunogenetics. 1998;48(3):222–232. doi: 10.1007/s002510050426
- Van de Winkel JGJ, Capel PJA. Human IgG Fc receptor heterogeneity: molecular aspects and clinical implications. Immunol Today. 1993;14(5):215–221. doi: 10.1016/0167-5699(93)90166-I
- Salmon JE, Edberg JC, Brogle NL, et al. Allelic polymorphisms of human Fc gamma receptor IIA and Fc gamma receptor IIIB. Independent mechanisms for differences in human phagocyte function. J Clin Invest. 1992;89(4):1274–1281. doi: 10.1172/JCI115712
- De Haas M, Koene HR, Kleijer M, et al. A triallelic Fcγ receptor IIIA polymorphism influences the binding of human IgG by NK cell Fcγ receptor IIIA. J Immunol. 1996;156(8):2948–2955.
- Tamm A, Schmidt RE. The binding epitopes of human CD 16 (Fcγ RIII) monoclonal antibodies: implication for ligand binding. J Immunol. 1996;157(4):1576–1581.
- Karassa FB, Trikalinos TA, Ioannidis JP. FcgammaRIIa-SLE Meta-Analysis Investigators. Role of the Fcγ receptor IIa polymorphism in susceptibility to systemic lupus erythematosus and lupus nephritis: a meta-analysis. Arthritis Rheum. 2002;46(6):1563–1571. doi: 10.1002/art.10306
- Manager K, Repp R, Jansen M, et al. Fcgamma receptor IIa, IIIa and IIIb polymorphisms in German patients with systemic lupus erythematosus: association with clinical symptoms. Ann Rheum Dis. 2002;61(9):786–792. doi: 10.1136/ard.61.9.786
- Vedeler CA, Myhr KM, Nyland H. Fc receptors for immunoglobulin G – a role in the pathogenesis of guillain–barré syndrome and multiple sclerosis. J Neuroimmunol. 2001;118(2):187–193. doi: 10.1016/S0165-5728(01)00344-7
- Wolff AS, Myhr KM, Vedeler CA, et al. Fcγ receptor polymorphisms are not associated with autoimmune Addison's disease. Scand J Immunol. 2007;65(6):555–558. doi: 10.1111/j.1365-3083.2007.01932.x
- Brandt JT, Isenhart C, Osborne JM, et al. On the role of platelet FcγRIIA phenotype in heparin induced thrombocytopenia. Thromb Haemost. 1995;74(6):1564–1572.
- Arepally G, McKenzie SE, Jiang XM, et al. FcγRIIA H/R 131 polymorphism, subclass-specific IgG anti-heparin/platelet factor 4 antibodies and clinical course in patients with heparin induced thrombocytopenia and thrombosis. Blood. 1997;89(2):370–375.
- Rodeghiero F, Stasi R, Gernsheimer T, et al. Standardization of terminology, definitions and outcome criteria in immune thrombocytopenic purpura of adults and children: report from an international working group. Blood. 2009;113(11):2386–2393. doi: 10.1182/blood-2008-07-162503
- Foster CB, Lehrnbecher T, Mol F, et al. Host defense molecule polymorphisms influence the risk for immune-mediated complications in chronic granulomatous disease. J Clin Invest. 1998;102(12):2146–2155. doi: 10.1172/JCI5084
- Carcao MD, Blanchette VS, Wakefield CD, et al. Fcγ receptor IIa and IIIa polymorphisms in childhood immune thrombocytopenic purpura. Br J Haematol. 2003;120(1):135–141. doi: 10.1046/j.1365-2141.2003.04033.x
- Lehrnbecher T, Foster CB, Zhu S, et al. Variant genotypes of the low affinity Fcγ receptor polymorphisms in control and disease populations. Blood. 2000;94(7):4220–4232.
- Grecco M, Santos VC, Pereira KM, et al. Fc gamma receptor IIIa polymorphism is not associated with susceptibility to systemic lupus erythematosus in Brazilian patients. Rev Brazil Reumatol. 2016;56(6):515–520. doi: 10.1016/j.rbr.2016.06.004
- Wu J, Edberg JC, Redecha PB, et al. A novel polymorphism of FcgammaRIIIa (CD16) alters receptor function and predisposes to autoimmune disease. J Clin Invest. 1997;100(5):1059–1070. doi: 10.1172/JCI119616
- Liu D, Tian Y, Sun D, et al. The FCGR3A polymorphism predicts the response to rituximab-based therapy in patients with non-hodgkin lymphoma: a meta-analysis. Ann Hematol. 2016;95(9):1483–1490. doi: 10.1007/s00277-016-2723-x
- Burkhardt B, Yavuz D, Zimmermann M, et al. Impact of Fc gamma-receptor polymorphisms on the response to rituximab treatment in children and adolescents with mature B cell lymphoma/leukemia. Ann Hematol. 2016;95(9):1503–1512. doi: 10.1007/s00277-016-2731-x
- Fujimoto TT, Inoue M, Shimomura T, et al. Involvement of Fcγ receptor polymorphism in the therapeutic response of idiopathic thrombocytopenic purpura. Br J Haematol. 2001;115(1):125–130. doi: 10.1046/j.1365-2141.2001.03109.x
- Williams Y, Lynch S, McCann S, et al. Correlation of platelet FcγRIIA polymorphism in refractory idiopathic (immune) thrombocytopenic purpura. Br J Haematol. 1998;101(4):779–782. doi: 10.1046/j.1365-2141.1998.00802.x
- Kyogoku C, Dijstelbloem HM, Tsuchiya N, et al. Fcγ receptor gene polymorphisms in Japanese patients with systemic lupus erythematosus: contribution of FCGR2B to genetic susceptibility. Arthritis Rheum. 2002;46(5):1242–1254. doi: 10.1002/art.10257