FULL PAPER
Association study of VDR gene with rheumatoid arthritis in
the French population
A Maalej
1, E Petit-Teixeira
2, L Michou
2,3, A Rebai
4, F Cornelis
2,3and H Ayadi
11Laboratoire de Ge´ne´tique Mole´culaire Humaine, Faculte´ de Me´decine de Sfax, Sfax, Tunisie;2GenHotel Laboratoire de Recherche
Europe´en pour la Polyarthrite Rhumatoı¨de, Universite´ d’Evry-Val d’Essonne, ECRAF-Universite´ Paris VII, Evry, France;3Unite´ de
Ge´ne´tique Clinique, Hoˆpital Lariboisie`re, Assistance Publique—Hoˆpitaux de Paris, Paris, France;4Unite´ de Bio-informatique, Centre de
Biotechnologie de Sfax, BP K, Sfax, Tunisie
Vitamin D is a potent regulator of calcium homeostasis and may have immunomodulatory effects. The influence of vitamin D on human autoimmune disease is controversial. The aim of this study was to investigate the role of vitamin D receptor gene (VDR) in rheumatoid arthritis (RA). Three polymorphisms for VDR gene FokI T4C (rs 10735810), BsmI A4G (rs 1544410) and TaqI C4T (rs 731236) were genotyped in 100 RA French nuclear families (set 1) and 100 additional French nuclear families for replication (set 2). The association analysis was performed using comparison of alleles frequencies (AFBAC), transmission disequilibrium test and genotype relative risk. Our results revealed a significant difference of F allele of FokI polymorphism between transmitted and nontransmitted frequencies (P¼ 0.01) in set 1. Furthermore, the F/F genotype was more frequent in RA patients compared to controls (P¼ 0.01) in set 1. The replication in set 2 showed similar patterns of transmission with a nonsignificant association. Association with FokI was found to be significant when the two sets were combined (P¼ 0.006). These data suggest that the F allele and F/F VDR genotype are associated with RA. The mechanisms by which distinct receptor variants might confer disease susceptibility remain to be elucidated.
Genes and Immunity (2005) 6, 707–711. doi:10.1038/sj.gene.6364260; published online 8 September 2005 Keywords: vitamin D receptor; rheumatoid arthritis; association study; candidate gene
Introduction
Rheumatoid arthritis (RA) is one of the most common human systemic autoimmune diseases. It is characterized by inflammation of synovial tissues and the formation of rheumatoid pannus, which is capable of eroding adjacent cartilage and bone and causing subsequent joint destruc-tion. Previous studies have indicated that the risk of developing the disease in siblings of affected individuals (lsib) is 2–17 times higher than in the general population,
suggesting the importance of genetic factors in RA.1The
only locus that has been conclusively associated is the HLA-DRB1 locus, which accounts for about one-third of the genetic component.2–4Many other potential candidate
genes of smaller effect have to be discovered.
Vitamin D plays an important role in the regulation of calcium and phosphorus metabolism.5 Recently, an
im-munoregulatory role of vitamin D has been postulated, based on the fact that the activation of human leukocytes causes the expression of the vitamin D receptor (VDR). Vitamin D is also involved in interleukin-2 (IL-2) inhibition and antibody production and in lymphocyte proliferation
suppression and the generation of cytotoxic lympho-cytes.6,7 In addition, monocytes constitutively express
VDR, as well as activated but not resting lymphocytes.8,9
Within the rheumatoid joint, the active form of vitamin D has been shown to be synthesized in RA synovium and is thought to be stimulated by interleukin-1 (IL-1) and/or IL-2.10The actions of 1,25 (OH)
2D3are mediated
via the nuclear VDR. Its gene, located on chromosome 12q12-14, harbors several polymorphisms and was found to be associated with several autoimmune diseases such as Addison’s disease, type I diabetes and systemic lupus erythematosus.11–14
Given the shared genetic susceptibility among auto-immune diseases and the association of VDR variants with other autoimmune diseases, we investigated the distribution of three VDR polymorphisms, FokI T4C (rs 10735810) located in exon 2, BsmI A4G (rs 1544410) located in intron 8 and TaqI C4T (rs 731236) located in exon 9, in 100 trios affected with one RA patient and both parents. A second set of 100 trios was used for replication.
Results
Linkage disequilibrium and Hardy–Weinberg equilibrium tests in set 1
The observed FokI, BsmI and TaqI genotype frequencies were in accordance with the Hardy–Weinberg
equili-Received 10 June 2005; revised 25 July 2005; accepted 2 August 2005; published online 8 September 2005
Correspondence: Dr H Ayadi, Laboratoire de Ge´ne´tique Mole´culaire Humaine, Faculte´ de Me´decine, Avenue Majida Boulila, 3029 Sfax, Tunisie. E-mail: hammadi.ayadi@fmsf.rnu.tn
brium in controls (data not shown). Analysis of linkage disequilibrium between the FokI and both BsmI and TaqI markers did not show any significant result (P ¼ 0.87 and 0.62, respectively). However, a complete linkage dis-equilibrium between BsmI and TaqI polymorphisms was found.
Test for association of VDR markers in set 1 RA trios AFBAC analysis revealed a significant difference of F and f alleles between transmitted and nontransmitted frequencies (P ¼ 0.01) (Table 1a). The transmission of the F allele from heterozygous parents (61 of 97 transmis-sions) (62.88%) showed an excess over that would be expected by Mendel’s law (50%) (w2¼ 6.44, P ¼ 0.01)
(Table 1a). Global GRR analysis showed a significant difference between RA patients and controls (w2¼ 6.21,
P ¼ 0.04). The F/F genotype was more frequent in RA patients compared to controls (45 RA cases vs 30 controls, P ¼ 0.01) (Table 1b).
To establish whether FokI association with RA was dependent on the presence of HLA-DRB1 SE allele, the transmission of F allele was analyzed in the subgroup of families where patient carried at least one HLA-DRB1 SE allele. Our results showed that F allele was preferentially transmitted (50 transmitted vs 29 nontransmitted, P ¼ 0.02). Similar results were found when patients were stratified according to the presence of erosion (53 transmitted vs 31 nontransmitted, P ¼ 0.02) as well as the presence of an autoimmune disease including RA in first-degree relative and/or in index (24 transmitted vs 12 nontransmitted, P ¼ 0.05).
In order to test haplotype transmission of VDR FokI, BsmI and TaqI alleles, haplotypes were determined. Five different haplotypes at the VDR gene locus were inferred with a frequency higher than 5%. None of these haplotypes showed a significant disequilibrium of transmission to patients (P40.05) (Table 2).
Results of test for association ofFokI marker in set 2 The association found in the first set of RA trios was tested in a second set of 100 families. AFBAC analysis did not reveal any significant difference between transmitted and nontransmitted frequencies (P ¼ 0.21) (Table 3a). The transmission of the F allele from heterozygous parents (49 of 86 times (57%) also trends toward overtransmission, but does not reach statistical significance (w2¼ 1.86, P ¼ 0.2). At the genotypic level,
the F/F genotype showed more evidence for association with RA (P ¼ 0.14) than at the allelic level (Table 3b).
Stratifying patients according to the presence of at least one HLA-DRB1 SE allele and erosion did not reveal any evidence of association of allele F with RA by TDT with 36 transmitted vs 30 nontransmitted (P ¼ 0.55) and 35 transmitted vs 32 nontransmitted (P ¼ 0.72) for HLA-DRB1 and erosion, respectively. Similar results were found when patients were stratified according to the presence of an autoimmune disease including RA in first-degree relative and/or in index (15 transmitted vs 12 nontransmitted, P ¼ 0.56).
Table 1a Transmission disequilibrium test and frequency analysis (AFBAC) for the FokI, BsmI and TaqI polymorphisms of VDR gene in RA set 1
Alleles TDT Frequencies
Trans Nontrans P-value Trans Nontrans P-value
F 61 36 0.665 0.540 f 36 61 0.01 0.335 0.460 0.01 B 50 52 0.376 0.387 b 52 50 0.84 0.624 0.613 0.83 T 50 46 0.626 0.605 t 46 50 0.68 0.374 0.395 0.67
Table 1b Genotype relative risk for the FokI, BsmI and TaqI markers in RA set 1
Genotypes RA Controls GRR P-value
FF 45 30 Global 0.04 Ff 43 48 FF vs Ff +ff 0.01 ff 12 22 ff vs Ff+FF 0.07 BB 19 13 Global 0.15 Bb 35 48 BB vs Bb+bb 0.27 bb 42 35 bb vs Bb+BB 0.36 TT 42 33 Global 0.12 Tt 35 49 TT vs Tt+tt 0.25 tt 18 13 tt vs Tt+TT 0.44
Table 2 Transmission disequilibrium test and frequency analysis (AFBAC) for the FokI/BsmI/TaqI haplotypes in RA set 1
Alleles TDT Frequencies
Trans. Nontrans P-value Trans Nontrans P-value
FBT 5 13 0.06 0.043 0.078
FBt 28 29 0.9 0.190 0.198
FbT 39 38 0.91 0.302 0.276
fBt 17 16 0.86 0.112 0.086
fbT 33 27 0.44 0.241 0.233 0.98
Table 3a Transmission disequilibrium test and frequency analysis for the FokI marker in RA set 2
Alleles TDT Frequencies
Trans Nontrans P-value Trans Nontrans P-value
F 49 37 0.680 0.620
f 37 49 0.20 0.320 0.380 0.21
Power computation
In these calculations, we used the frequency of the F allele in the general population (60%) as being that of the risk allele. Four GRR values were considered (g ¼ 1.1, 1.4, 1.7 and 2.0) and power computations were made under multiplicative gene effect model as suggested by previous studies.15,16 Note that the estimate of GRR
based on transmissions in sets 1 and 2 combined is about 1.5 (b ¼ 109, c ¼ 71 giving g ¼ 109/71 ¼ 1.53).
A power of 50% is reached for a g value close to 1.7. For g ¼ 1.5, we found a power of about 38% in set 1 and 31% in set 2. This means that if the actual g value is 1.5, the probability of finding a positive association when it actually exists is 0.38 and 0.31 in sets 1 and 2, respectively. The probability of finding a positive result in both sets is thus 0.38 0.31 ¼ 0.12 (as both sets are independent). This means that out of 100 samples (of 100 triads each), the replication would be found only in 12. To detect a gene having this GRR with 80% power (at 5% significance level) a number of informative families between 80 and 100 is needed (depending on the risk allele frequency).
Results of test for association ofFokI marker in sets 1 and 2
When we combined set 1 and 2, AFBAC analysis revealed a significant difference of F and f alleles between transmitted and nontransmitted frequencies (P ¼ 0.007). The transmission of the F allele from heterozygous parents (110 of 183 times (60.1%)) showed an excess over that expected by Mendel0s law (50%, P ¼ 0.006). Global GRR analysis showed a significant difference between RA patients and controls (P ¼ 0.02) and the F/F genotype was more frequent in RA patients compared to controls (93 RA cases vs 67 controls (P ¼ 0.005).
Discussion
We studied the association of three VDR polymorphisms (FokI, BsmI and TaqI) in RA and found that allele F and genotype FF appeared to be associated with suscept-ibility to RA in set 1. Following this descriptive analysis on set 1, replication study on a second set of 100 trios showed the same tendency. Stratification for HLA-DRB1 SE alleles, erosion or autoimmune disease including RA in first-degree relative and/or in index did not show stronger association. The results obtained in the ‘auto-immune’ subgroup may outline a specific association of VDR with RA and not with the general process of autoimmune diseases.
The lack of replication evidence in set 2 could be explained by the smaller number of informative families with at least one heterozygous parent (64 vs 74 in set 1). In fact, power calculation showed that the probability of detecting the gene when it exists was weak (below 1/3) for a gene with minor effect (GRRo1.5) under a multi-plicative mode. At least 80 triads are needed to get an 80% power for a positive replication. This association was emphasized when both sets are combined.
Previous studies claimed that at least five times as many families are needed to replicate an original linkage finding if criteria for significance are the same at the two stages.17 However, with multiple susceptibility genes
contributing to a trait, some genes will by chance be most prevalent and are detected in a given data set, whereas they would be very difficult to find in a new data set because they are unlikely to be again the most frequent in this data set.18
Neither BsmI nor TaqI alleles showed an association with RA in set 1. No association was found with any haplotype of three markers; this result could be explained by the lack of association of both BsmI and TaqI polymorphisms and the lack of linkage disequili-brium between the associated FokI polymorphism and the two others. In fact, Nejentsev et al19found that FokI
SNP is not in linkage disequilibrium with any other SNP of the VDR gene. Moreover, these authors showed that BsmI and TaqI SNP localized to a single block of linkage disequilibrium and did not capture any information on a large part of the VDR gene, especially on the 50region.
Linkage studies in our whole-genome scan of Cauca-sian RA families have revealed no evidence for linkage with the 12q14 region, where VDR gene is located.20 It
should, however, be noted that failure to detect linkage of a region does not exclude the possibility of a disease risk gene in that region.21
Our results are in contrast with a case/control study in German population, which showed no evidence of RA association with VDR.22 This could be due to
hetero-geneity between populations or lack of power in the German study caused by the small samples used (62 cases and 40 controls). Moreover, we used a case family-based control design to avoid the heterogeneity caused by mismatch between populations in case/controls studies. The additional advantages of such studies are described in Huizinga et al.23
The F associated allele in a VDR protein has three amino acids less than the f variant.24 FokI alleles differ
functionally25,26due to altered VDR affinity and
transac-tivation of VDR elements containing promoter construct in HeLa and COS-7 cells.27 An in vitro study
demon-strated an increased transcription rate (1.7-fold) of the VDR gene in cells with the F/F genotype.28 An
over-expression of VDR gene may affect the over-expression of genes containing such VDR response element. This could dysregulate the Th1/Th2 balance and therefore could cause the development of autoimmune process of RA. Our present findings suggest that polymorphisms in the VDR gene could have some effects on RA etiopathogen-esis. These effects are of interest given recent studies29
demonstrating the existence of different VDR-mediated immunoregulatory properties of vitamin D. The clinical course of RA is characterized by bone and joint destruction and it is therefore conceivable that poly-morphic genes whose products have a direct effect on Table 3b Genotype relative risk for the FokI marker in RA set 2
Genotypes RA Controls GRR P-value
FF 48 37 Global 0.28 Ff 40 50 FF vs Ff +ff 0.14 ff 12 13 ff vs Ff+FF 0.63 709
calcium/vitamin D metabolism may play a role in its pathogenesis. On the other hand, the discovery of VDR in monocytes and activated, but not resting, lympho-cytes8,9 suggested a role in immunoregulation and
possibly that joint inflammation could be influenced by VDR polymorphisms. Alternatively, the VDR gene may not be responsible for the primary disease association but may be in linkage disequilibrium with a nearby novel disease-related locus.
In summary, our findings suggest that VDR or a closely linked gene is associated with RA in the French population. However, the mechanism by which such polymorphisms are associated with RA is unknown and our findings need to be confirmed by a functional study.
Patients and methods
Patients
Set 1 RA sample. All subjects provided informed consent, and the ethics committee of Hoˆpital Biceˆtre (Kremlin-Biceˆtre, Assistance Publique—Hoˆpitaux de Paris, France), approved the study. RA families were recruited through a national media campaign followed by selection of individuals who fulfilled the 1987 American College of Rheumatology criteria for RA30
according to the physicians in charge of the patients. All clinical data were reviewed by a rheumatologist uni-versity fellow. A total of 100 French Caucasian trio families (with four French Caucasian grand parents) were investigated. Among the 100 RA patients, 87 were women and 13 were men; their mean age at disease onset was 32 years. In all, 72 were rheumatoid factor (RF) positive, 78 carried at least one HLA-DRB1 susceptibility alleles ‘shared epitope’ (SE) (DRB1*0101, 0102, 0401, 0404, 0405, 0408, 1001),2 90 presented with erosive
disease and 31 presented nodules.
Four patients have another autoimmune disease, and in 29 families at least one first-degree relative had an auto-immune disease including RA, thyroid auto-im-mune diseases, myasthenia gravis, systemic lupus erythematosus, vitiligo and pernicious anemia.
Set 2 RA sample. A total of 100 French Caucasian trio families (with four French Caucasian grand parents) were investigated. Among the 100 RA patients, 90 were women and 10 were men; their mean age at disease onset was 31.2 years. In all, 72 were RF positive, 80 carried at least one HLA-DRB1 SE, 78 had erosive disease and 20 presented nodules. In all, 10 patients have another autoimmune disease, and in 23 families at least one first-degree relative had an autoimmune disease includ-ing RA, thyroid autoimmune diseases, celiac disease, vitiligo, insulin-dependant diabetes mellitus and pri-mary Sjo¨gren’s syndrome.
Genotype analysis
DNA was isolated from whole blood according to standard protocols. Genotypes for three restriction polymorphic sites (FokI, BsmI and TaqI) were identified by the polymerase chain reaction followed by restriction fragment length polymorphism (PCR/RFLP) method. The FokI, BsmI and TaqI polymorphisms were studied using primers described by Pani et al.12PCR
amplifica-tions were performed on each sample in a 25 ml reaction
volume consisting of 10 PCR buffer (Perkin Elmer, Boston, MA, USA), 0.5 mMof each primer, 0.1 mMof each dNTP, 1.25 U of Taq Gold DNA polymerase (Perkin Elmer), 3 mMMgCl2and 50 ng of genomic DNA, diluted
to the final volume with H2O. PCR amplification was
carried out using Techne thermocycler. Each PCR was performed using a hot start procedure, and amplification was carried out using 35 cycles of denaturation at 941C for 40 s, with annealing temperature for 30 s followed by an elongation step at 721C for 1 min. One final cycle of the extension was performed at 721C for 2 min.
Regarding FokI polymorphism (annealing tempera-ture: 601C), the 265-bp amplified fragment was digested with FokI, generating two fragments of 196 and 69 bp when the FokI restriction site is present. For examining the BsmI polymorphism (annealing temperature: 661C), the resulting 825-bp fragment was digested with BsmI, generating two fragments of 175 and 650 bp when the BsmI restriction site is present. The TaqI polymorphism (annealing temperature: 601C) was examined in a 740-bp fragment of intron 8/exon 9. Digestion with TaqI resulted in three fragments of 290, 245 and 205 bp in the presence of the polymorphic site and two fragments of 245 and 495 bp in its absence, due to an additional monomorphic TaqI site. In samples heterozygous for the respective restriction sites, both digested and undigested DNA fragments were visible. Genotypes were assessed blindly by two independent investigators (AM and EPT) and were designated by a lower case letter for the presence of a restriction site and by a capital letter for its absence. In order to test the quality of genotyping, we performed a control quality test by randomly genotyping 10% of the sample. The results were the same as in the first genotyping.
Statistical analysis
Prior to association tests, we looked at Hardy–Weinberg equilibrium in virtual controls (constituted by parental alleles nontransmitted to RA patients).
Association between each polymorphism and RA was examined by three different methods: affected family-based controls (AFBAC) was used to compare trans-mitted and nontranstrans-mitted allele frequencies across all families.31 Transmission disequilibrium test (TDT) was
used to detect preferential transmission of the RFLP alleles to the affected subjects32 and genotype relative
risk (GRR) (a single genotype vs the others), was used to compare the genotype distribution in controls and patients.33
VDR haplotypes were inferred using the algorithm implemented in Genehunter (‘haplo’ option). AFBAC, TDT and GRR were performed on haplotypes as described above and linkage disequilibrium was tested for each pair of markers. The haplotype frequencies were estimated by the Arlequin version 2 program. Signifi-cance of P-values was assessed using a Bonferroni correction at 5% (a P-value less than 0.05/3 ¼ 0.017 is considered significant).
Power calculation
In order to evaluate the information of the two sets, we have carried out power calculation using TDT-Power calculator.34 This program allows computation of power
of the TDT given a number of informative families, the frequency of the risk allele in the population and the 710
genotypic relative risks (defined by the parameter g of Risch and Merikangas).15 Calculation of the number of
families needed to obtain an 80% power is also carried out.
Acknowledgements
We thank the RA members for their participation, Docteur Sandra Lasbleiz (Unite´ de Ge´ne´tique Clinique, Hoˆpital Lariboisie`re) for reviewing the clinical data and all the other members of GenHotel team. This work was funded by the Association Franc¸aise des Polyarthri-tiques, Association de Recherche pour la Polyarthrite, Association Polyarctique, Association Rhumatisme et Travail, Socie´te´ Franc¸aise de Rhumatologie, Genopole, Universite´ d’Evry-Val d’Essonne, Shering-Plough, Pfizer, Amgen, Conseil Re´gional Ile de France, Conseil Ge´ne´ral de l’ Essone, Ministe`re de la Recherche et de l’Enseigne-ment Supe´rieur, Fondation pour la Recherche Me´dicale (France), Ministe`re de l’Enseignement Supe´rieur and Ministe`re la Recherche Scientifique et de la Technologie (Tunisie). We are grateful to Pr. Nadir Farid for his critical reading of the manuscript.
References
1 Seldin MF, Amos CI, Ward R, Gregersen PK. The genetics revolution and the assault on rheumatoid arthritis. Arthritis Rheum 1999; 42: 1071–1079.
2 Gregersen PK, Silver J, Winchester RJ. The shared epitope hypothesis. An approach to understanding the molecular genetics of susceptibility to rheumatoid arthritis. Arthritis Rheum 1987; 30: 1205–1213.
3 Nepom GT. Major histocompatibility complex-directed susceptibility to rheumatoid arthritis. Adv Immunol 1998; 68: 315–332.
4 Weyand CM, Goronzy JJ. Association of MHC and rheuma-toid arthritis. HLA polymorphisms in phenotypic variants of rheumatoid arthritis. Arthritis Res 2000; 2: 212–216.
5 DeLuca HF, Schnoes HM. Metabolism and mechanism of action of vitamin D. Annu Rev Biochem 1976; 45: 631–666. 6 Manolagas SC, Provvedini DM, Tsoukas CD. Interactions of
1,25-dihydroxyvitamin D3 and the immune system. Mol Cell Endocrinol 1985; 43: 113–122.
7 Rigby WF. The immunobiology of vitamin D. Immunol Today 1988; 9: 54–58.
8 Bhalla AK, Amento EP, Clemens TL, Holick MF, Krane SM. Specific high affinity receptors for 1,25-dihydroxivitamin D3 in human peripheral blood mononuclear cells: presence in monocytes and induction in T lymphocytes following activa-tion. J Clin Endocrinol Metab 1983; 57: 1308–1310.
9 Provvedini DM, Tsoukas CD, Deftos LJ, Manolagas SC. 1,25-Dihydroxyvitamin D3 receptors in human leukocytes. Science 1983; 221: 1181–1183.
10 Inaba M, Yukioka K, Furumitsu Y et al. Positive correlation between levels of IL-1 or IL-2 and 1,25 (OH)2D/25-OH-D ratio in synovial fluid of patients with rheumatoid arthritis. Life Sci 1997; 61: 977–998.
11 McDernott MF, Ramachandran A, Ogunkolade BW et al. Allelic variation in the vitamin D receptor influences susceptibility to IDDM in Indian Asians. Diabetologia 1997; 40: 971–975. 12 Pani MA, Knapp M, Donner H et al. Vitamin D receptor allele
combinations influence genetic susceptibility to type 1 diabetes in Germans. Diabetes 2000; 49: 504–507.
13 Huang CM, Wu MC, Wu JY, Tsai FJ. Association of vitamin D receptor gene BsmI polymorphisms in Chinese patients with systemic lupus erythematosus. Lupus 2002; 11: 31–34.
14 Pani MA, Seissler J, Usadel KH, Badenhoop K. 2002 Vitamin D receptor genotype is associated with Addison’s disease. Eur J Endocrinol 2002; 147: 635–640.
15 Risch N, Merikangas K. The future of genetics studies of complex human diseases. Science 1996; 273: 1516–1517. 16 Schaid DJ. Likelihoods and TDT for the case-parents design.
Genet Epidemiol 1999; 16: 250–260.
17 Suarez BK, Hampe CL. Linkage and association. Am J Hum Genet 1994; 54: 554–559.
18 Lernmark A, Ott J. Sometimes it’s hot, sometimes it’s not. Nat Genet 1998; 19: 213–214.
19 Nejentsev S, Godfrey L, Snook H et al. Comparartive high-resolution analysis of linkage disequilibrium and tag single nucleotide polymorphism between populations in the vitamin D receptor gene. Hum Mol Genet 2004; 13: 1633–1639. 20 Osorio Y, Fortea J, Bukulmez H et al. Dense genome-wide
linkage analysis of rheumatoid arthritis, including covariates. Arthritis Rheum 2004; 50: 2757–2765.
21 Van Heel DA, Fisher SA, Kirby A, Daly MJ, Rioux JD, Lewis CM, Genome Scan Meta-analysis group of the IBD interna-tional Genetics Consortium. Inflammatory bowel disease susceptibility loci defined by genome scan meta-analysis of 1952 affected relative pairs. Hum Mol Genet 2004; 13: 763–770. 22 Garcia-Lozano JR, Gonzalez-Escribano MF, Valenzuela A, Garcia A, Nunez-Roldan A. Association of vitamin D receptor genotypes with early onset rheumatoid arthritis. Eur J Immunogenet 2001; 28: 89–93.
23 Huizingua TW, Pisetsky DS, Kimberly RP. Associations, populations, and the truth: recommendations for genetic association studies in arthritis and rheumatism. Arthritis Rheum 2004; 50: 2066–2071.
24 Gross C, Eccleshall TR, Malloy PJ, Villa ML, Marcus R, Feldman D. The presence of a polymorphism at the translation initiation site of the vitamin D receptor gene is associated with low bone mineral density in postmenopausal Mexican-American women. J Bone Miner Res 1996; 11: 1850–1855. 25 Colin EM, Weel AE, Uitterlinden AG et al. Consequences of
vitamin D receptor gene polymorphisms for growth inhibition of cultured human peripheral blood mononuclear cells by 1,25-dihydroxyvitamin D3. Clin Endocrinol 2000; 52: 211–216. 26 Jurutka PW, Remus LS, Whitfield GK et al. The polymorphic N
terminus in human vitamin D receptor isoforms inflences transcriptional activity by modulating interaction with tran-scription factor IIB. Mol Endocrinol 2000; 14: 401–420. 27 Miyamoto K, Takentani Y, Tohna E, Nishisho T. A novel
polymorphism in the vitamin D receptor and bone mineral density: a study of vitamin D receptor expression and function in COS-7 cells. J Bone Miner Rese 1996; 11: 116S.
28 Arai H, Miyamoto K, Taketani Y et al. A vitamin D receptor gene polymorphism in the translation initiation codon: effect on protein activity and relation to bone mineral density in Japenese women. J Bone Miner Res 1997; 12: 915–992. 29 Merlino LA, Curtis J, Mikuls TR, Cerhan JR, Criswell LA, Saag
KG. Vitamin D intake is inversely associated with rheumatoid arthritis. Arthritis Rheum 2004; 50: 72–77.
30 Arnett FC, Edworthy SM, Bloch DA et al. The American Rheumatism Association 1987 revised criteria for the classi-fication of rheumatoid arthritis. Arthritis Rheum 1988; 31: 315–324.
31 Thomson G. Mapping disease genes: family-based association studies. Am J Hum Genet 1995; 57: 487–498.
32 Spielman RS, McGinnis RE, Ewens WJ. Transmission test for linkage disequilibrium: the insulin gene region and insulin-dependent diabetes mellitus (IDDM). Am J Hum Genet 1993; 52: 506–516.
33 Lathrop GM. Estimating genotype relative risks. Tissue Antigens 1983; 22: 160–162.
34 Chen WM, Deng HW. A general approach for computing the statistical power of transmission disequilibrum test for complex disease genes. Genet Epidemiol 2001; 21: 53–67.
