Human Molecular Genetics, 2000, Vol. 9, No. 11 1563-1566
© 2000 Oxford University Press
Polymorphisms of the CYP2D6 gene increase susceptibility to ankylosing spondylitis
Wellcome Trust Centre for Human Genetics, Roosevelt Drive, Headington, Oxon OX3 7BN, UK, 1Department of Pharmacological Sciences, University of Newcastle-upon-Tyne, UK, 2Queen Elizabeth Hospital, Adelaide, Australia, 3Royal National Hospital for Rheumatic Diseases, Upper Borough Walls, Bath, UK, 4Division of Life Sciences, King’s College, London, UK and 5Department of Human Genetics, University of Pittsburgh, Pittsburgh, PA, USA
Received 1 February 2000; Revised and Accepted 25 April 2000.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
Ankylosing spondylitis (AS) is a common and highly familial rheumatic disorder. The sibling recurrence risk ratio for the disease is 63 and heritability assessed in twins >90%. Although MHC genes, including HLA-B27, contribute only 20–50% of the genetic risk for the disease, no non-MHC gene has yet been convincingly demonstrated to influence either susceptibility to the disease or its phenotypic expression. Previous linkage and association studies have suggested the presence of a susceptibility gene for AS close to, or within, the cytochrome P450 2D6 gene (CYP2D6, debrisoquine hydroxylase) located at chromosome 22q13.1. We performed a linkage study of chromosome 22 in 200 families with AS affected sibling-pairs. Association of alleles of the CYP2D6 gene was examined by both case–control and within-family means. For case–control studies, 617 unrelated individuals with AS (361 probands from sibling-pair and parent–case trio families and 256 unrelated non-familial sporadic cases) and 402 healthy ethnically matched controls were employed. For within-family association studies, 361 families including 161 parent–case trios and 200 affected sibling-pair families were employed. Homozygosity for poor metabolizer alleles was found to be associated with AS. Heterozygosity for the most frequent poor metabolizer allele (CYP2D6*4) was not associated with increased susceptibility to AS. Significant within-family association of CYP2D6*4 alleles and AS was demonstrated. Weak linkage was also demonstrated between CYP2D6 and AS. We postulate that altered metabolism of a natural toxin or antigen by the CYP2D6 gene may increase susceptibility to AS.
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
Ankylosing spondylitis (AS) is a common rheumatic disorder characterized primarily by inflammation in the spine and sacroiliac joints. The population prevalence of the disease in Caucasians is 1/1000 (1). AS is strongly linked with the MHC region (2,3), and associated with the genes HLA-B27 (B27) and HLA-B60 (4–7). About 95% of patients are B27 positive, and inheritance within families occurs almost exclusively in B27-positive individuals (8). However, twin and family studies suggest that the contribution of MHC-linked genes, including B27, is only 20–50% of the familial recurrence risk of the disease (2,3,9). Variance modelling in twins suggests that more than 90% of the population variance of AS is attributable to genetic factors, indicating that the majority of the familial recurrence risk for the disease is due to shared genes rather than the environment (9). A large non-MHC component of the genetic risk for the disease is suggested by the significant difference in concordance rates for MZ twins (63%) and B27-positive DZ twins (23%), and the large non-MHC component of the sibling recurrence risk ratio (
(non-MHC) = 14) (3). The pattern of reduction in the recurrence risk ratio in increasingly distant relatives of AS patients suggests that three to nine genes contribute multiplicatively to disease susceptibility (10). In a genome-wide scan of 105 affected sibling-pair families, we identified seven regions linked with AS with P-values of 
0.01 (3). One gene, cytochrome P450 2D6 (debrisoquine 4-hydroxylase, CYP2D6), which has previously been associated with AS (11), lies near a microsatellite CYP2D (12) which showed weak evidence of linkage in our genome screen (LOD = 0.6, P = 0.05). Approximately 5–10% of Caucasians have deficient function of this enzyme (poor metabolizer phenotype), which is inherited as an autosomal recessive trait. At least 15 allelic variants of CYP2D6 can cause poor metabolizer phenotype, but 75% of poor metabolizers are CYP2D6*4 homozygotes. In this study, we have investigated the role of this gene in susceptibility to AS by both linkage and association. |
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
The CYP2D6 allele frequencies in each patient collection and controls are given in Table 1. No significant difference in allele frequencies was observed between our controls and those previously reported by Beyeler et al. (Table 1) (11). Therefore, results are reported both separately and pooled together with this previously reported study. Neither case nor control genotype frequencies vary significantly from Hardy–Weinberg equilibrium.
Examining cases and controls only from the current study, significant association was observed between homozygosity for CYP2D6*4 and AS [genotype relative risk (GRR) = 2.1, 95% confidence interval 1.3–3.4, P = 0.002)]. In contrast, the risk of disease among heterozygotes was not increased, the overall GRR being 1.1 (95% confidence interval 0.8–1.4, P = 0.5). Pooling all patients and controls, homozygosity for the CYP2D6*4 allele was strongly associated with AS, with a GRR of 2.1 (95% confidence interval 1.5–3.1, P = 0.0005). Again, heterozygosity for allele 4 was not associated with AS (GRR = 1.0, 95% confidence interval 0.9–1.2, P = 0.7).
CYP2D6*4 homozygosity and CYP2D6*4/CYP2D6*5 heterozygosity both result in the poor metabolizer phenotype. Considering both these poor metabolizer genotypes, the association with AS is highly significant both in the current study population (GRR = 2.2, 95% confidence interval 1.4–3.6, P = 0.0007) and with the pooled results (GRR = 2.1, 95% confidence interval 1.5–3.1, P = 0.0003). Significant transmission disequilibrium of the CYP2D6 polymorphism was observed using Transmit [
2 = 7.3 (1 d.f.), empirical P = 0.014].
Weak linkage of the CYP2D6 polymorphism and AS was observed with a LOD score of 0.9 [P = 0.02, mean identical-by-descent (IBD) sharing 0.6]. Weak linkage was also observed for the microsatellite marker CYP2D, with a LOD score of 0.6 (P = 0.05). For the CYP2D6 polymorphism the proportion of families sharing one allele was considerably reduced from the expected (z1 expected 0.5, observed 0.4), implying the presence of dominance variance. Multipoint LOD scores at the CYP2D6 locus were slightly lower than single-point scores (LOD = 0.5). No other marker showed significant evidence of linkage (mean IBD sharing < 0.55, P > 0.05).
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
This report confirms the previously reported finding of significant association between poor metabolizer genotypes and AS (11). Within-family association was also demonstrated for the CYP2D6*4 polymorphism, demonstrating that the case–control findings indicate true association rather than the presence of population stratification. Weak evidence of linkage was also noted both for the CYP2D intragenic microsatellite and the CYP2D6 gene.
CYP2D6 poor metabolizer phenotype can be due to at least 15 different genetic variants of the CYP2D6 gene (13). Homozygosity for CYP2D6*4 is genetically responsible for ~75% of poor metabolizers (14,15). The lack of association between heterozygosity for CYP2D6*4 and AS suggests either that we have observed true association with a recessively acting polymorphism, or that CYP2D6*4 is in strong linkage disequilibrium with another recessively acting allele. IBD sharing analysis at the CYP2D6 locus revealed a marked reduction in the sharing of one allele IBD (z1), supporting the presence of dominance variance at this locus. Where a gene acts codominantly, the proportion z1 is 0.5 and deviation below this occurs with dominant or recessive acting genes. In the current study the reduction of z1 from the expected value of 0.5 to the observed value 0.4 is consistent with a recessive genetic effect.
The use of the GRR statistic depends on the assumption that the control genotypes are in Hardy–Weinberg equilibrium. In this circumstance the statistic is more powerful than the standard chi-squared statistic (16). In this study the control genotypes do not deviate significantly from those expected under Hardy–Weinberg equilibrium, and therefore the use of the GRR is justified.
While case–control studies are susceptible to error due to population stratification, the Transmit transmission disequilibrium statistic has been shown to be highly robust to violation of the assumption of population stratification (17). The finding of significant association using this statistic is therefore highly significant in that it indicates that the case–control results are unlikely to be due to this source of error. The CYP2D6 allele frequencies we observed are very similar to those reported in other studies of British Caucasians both from London (18) and the northeast of England (11), also supporting the conclusion that true association is the cause of the findings reported here.
Poor metabolizer phenotype has been associated with a wide range of diseases including Parkinson’s disease (19), cancers (20), scleroderma (21) and systemic lupus erythematosus (22), although replication of the latter two findings has not been reported. Whilst CYP2D6 is known to be involved in metabolism of a wide range of drugs, no endogenous CYP2D6 substrate has yet been identified, and therefore its biological function remains uncertain. It has been postulated that poor metabolism of xenobiotics (i.e. drugs, metals, industrial and naturally occurring chemicals) or plant toxins by the defective CYP2D6 variants may explain these associations (22). Xenobiotics are known to have numerous pro-inflammatory effects on T-cells (23), which may be critical cells in the pathophysiology of B27-related arthritis (24). These effects include uncovering of cryptic peptides by hapten formation producing autoimmune T-cell reactions, and direct recognition of xenobiotics or their metabolites by T-cells (23). Well-known examples of autoimmune diseases due to xenobiotics include Stevens–Johnson syndrome and dihydralazine-induced systemic lupus erythematosus, in which impaired drug metabolism is a significant disease risk factor. Whilst these diseases are clearly quite different to AS, they illustrate that impaired xenobiotic metabolism can induce autoimmune diseases in humans.
This study strongly suggests that dysfunction of the CYP2D6 gene increases the risk of AS, although only contributing a small proportion of the overall risk of the disease. Further studies of enzymes involved in xenobiotic metabolism are warranted in AS, and in other T-cell mediated autoimmune diseases.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
Patients
We studied linkage of the CYP2D6 gene, and association of the main poor metabolizer genotype in 617 unrelated AS patients and 402 healthy controls, and in 361 families with AS consisting of 161 parent–case trios and 200 affected sibling-pair families, containing 276 affected sibling-pairs. The 105 affected sibling-pair families used in our initial whole genome screen are included in the affected sibling-pair families used in this study. In the parent–case trios all parents were available for genotyping, whereas of the affected sibling-pair families, 71 had both parents available, 63 one parent and 66 neither parent. The probands from these families were also used in the case–control association studies. AS was defined by the definite category of the modified New York criteria for AS consisting of radiographic and clinical criteria (25). Four hundred and two control healthy blood donors were recruited through the Oxford Regional Transfusion Service. An additional panel of 662 UK historic controls and 54 AS patients previously reported were used in some of the analyses (11). All patients and controls were British Caucasian subjects. Ethical approval for this study was obtained from the Central Oxford Research Ethics Committee.
Laboratory methods
We used a modified version of the typing method reported by Beyeler et al. (11), by which all samples were genotyped for the CYP2D6*4 allele. This method involves a restriction digest using the enzyme BstNI, which cuts at position 1934 in wild-type but not mutant alleles. Because this method fails to distinguish between reactions where the restriction digest has failed and homozygous mutant samples, we altered the position of the 5'-PCR primer to also include a non-polymorphic BstNI restriction site at position 1772. The primers used were D6L2 (5'-GGT GTT CCT CGC GCG CTA TG-3') and P5 (5'-CTC GGT CTC TCG CTC CGC AC-3') which produce a 421 bp product. The inclusion of this second restriction site produces a constant 77 bp fragment, so that failure of digestion can clearly be distinguished from homozygote mutant samples. The primer D6L2 has a deliberate mismatch of the 10th base to remove a further BstNI restriction site in the normal sequence. Thus CYP2D6*4 homozygotes produce two fragments of lengths 77 and 344 bp, whereas non-CYP2D6*4 homozygotes produce three fragments of lengths 77, 161 and 183 bp. Amplifications were carried out in 10 µl reactions consisting of 50 ng DNA, 25 ng each primer, 200 µM each dNTP, 2.0 mM MgCl2, 1 µl 10x NH4 buffer (Bioline, London, UK) and 0.2 U Taq polymerase (Bioline). Amplifications were carried out in 96-well plates (Costar, Corning, NY) using Omnigene thermal cyclers (Hybaid, Teddington, UK). The standard cycling conditions used were denaturation at 94°C 1 min, annealing 60°C 1 min and extension at 72°C for 1 min. Digestions were carried out using 10 µl of PCR product, 2 µl of NE Buffer 2 (New England Biolabs, Beverly, MA), 0.2 µl bovine serum albumin, 10 U of BstNI, and water to a final volume of 20 µl, heated overnight at 60°C. Digestion products were electrophoresed on 3% agarose gels and the PCR product was visualised by ethidium bromide staining and fluorescence under UV light. Titration experiments indicated that complete digestion could be obtained with <1 U of BstNI and a 2 h digestion period, indicating that our protocol ensured complete PCR product digestion. This method distinguishes CYP2D6*4 from all other CYP2D6 alleles except CYP2D6*5. CYP2D6*5 is a 13 kb deletion of the CYP2D6 gene, with an allele frequency of 3.9% in British Caucasians (11). CYP2D6*5 heterozygotes genotype as homozygotes for either CYP2D6*4 or wild-type by the BstNI restriction digest typing method, as the chromosomal strand carrying the CYP2D6*5 allele will not amplify. Homozygotes for CYP2D6*5 fail to amplify altogether, but are rare (0.1%) (11) and were therefore not specifically genotyped in this study. We typed all apparent CYP2D6*4 homozygotes for CYP2D6*5 using a variation of a previously reported method (26). The same PCR primers were used, but rather than ‘long-range’ DNA polymerase, we employed standard DNA polymerase (Bioline) and Q-solution (Qiagen, Crawley, UK), which gave more reliable amplification of this GC-rich region.
In addition to the marker CYP2D, 10 chromosome 22 fluorescently tagged microsatellites spread across chromosome 22 (D22S1158, D22S1173, D22S1176, D22S274, D22S277, D22S280, D22S283, D22S315, D22S420, D22S423—average marker spacing 4.6 cM) were PCR amplified and products separated on ABI 373 semiautomated sequencers by standard means (3). Mendelian inheritance was verified both manually and using the program Pedcheck (27).
Statistical methods
Non-parametric two-point linkage analysis was performed using the package, Analyze (28). IBD sharing assessment and multipoint analysis was performed using the non-parametric program MapMaker/Sibs, conservatively weighting the numbers of affected sibling-pairs per family by 2/number of affecteds (29). GRRs were calculated according to the method of Lathrop (16). In this method, genotype frequencies and their variances are calculated from the observed allele frequencies assuming Hardy–Weinberg equilibrium, resulting in a power gain because the statistic variances are smaller compared with those from standard contingency table methods. The method is valid to use in our case, as the control genotype frequencies are in Hardy–Weinberg equilibrium. Unlike the GRR definition of Risch and Merikangas (30), this method makes no assumptions about the relative GRR of homozygotes and heterozygotes.
Within-family association was studied using the program, Transmit (17), a form of transmission–disequilibrium test, which permits use of families where parental genotypes may be missing. P-values were obtained by bootstrap simulation using 10 000 replicates. When using bootstrapping, the statistic is robust to inclusion of multiple affected offspring in each family, even in the presence of linkage.
|
|
FOOTNOTES |
|---|
+ To whom correspondence should be addressed. Tel: +44 1865 287647; Fax: +44 1865 287501; Email: mbrown@well.ox.ac.uk
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
1 van der Linden, S., Valkenburg, H. and Cats, A. (1983) The risk of developing ankylosing spondylitis in HLA-B27 positive individuals: a family and population study. Br. J. Rheumatol., 22, 18–19.
2 Rubin, L.A., Amos, C.I., Wade, J.A., Martin, J.R., Bale, S.J., Little, A.H., Gladman, D.D., Bonney, G.E, Rubenstein, J.D. and Siminovitch, K.A. (1994) Investigating the genetic basis for ankylosing spondylitis. Linkage studies with the major histocompatibility complex region. Arthritis Rheum., 37, 1212–1220.[Web of Science][Medline]
3 Brown, M.A., Pile, K.D., Kennedy, L.G., Campbell, D., Andrew, L., March, R., Shatford, J.L., Weeks, D.E., Calin, A. and Wordsworth, B.P. (1998) A genome-wide screen for susceptibility loci in ankylosing spondylitis. Arthritis Rheum., 41, 588–595.[Web of Science][Medline]
4 Brewerton, D.A., Hart, F.D., Nicholls, A., Caffrey, M., James, D.C. and Sturrock, R.D. (1973) Ankylosing spondylitis and HL-A 27. Lancet, 1, 904–907. [Web of Science][Medline]
5 Schlosstein, L., Terasaki, P.I., Bluestone, R. and Pearson, C. (1973) High association of an HL-A antigen W27, with ankylosing spondylitis. N. Engl. J. Med., 288, 704–705.
6 Robinson, W.P., van d. L.S., Khan, M.A., Rentsch, H.U., Cats, A., Russell, A. and Thomson, G. (1989) HLA-Bw60 increases susceptibility to ankylosing spondylitis in HLA-B27+ patients. Arthritis Rheum., 32, 1135–1141.[Web of Science][Medline]
7 Brown, M.A., Pile, K.D., Kennedy, L.G., Calin, A., Darke, C., Bell, J., Wordsworth, B.P. and Cornelis, F. (1996) HLA class I associations of ankylosing spondylitis in the white population in the United Kingdom. Ann. Rheum. Dis., 55, 268–270.
8 Brown, M.A. and Wordsworth, B.P. (1997) Predisposing factors to spondyloarthropathies. Curr. Opin. Rheumatol., 9, 308–314.[Medline]
9 Brown, M.A., Kennedy, L.G., MacGregor, A.J., Darke, C., Duncan, E., Shatford, J.L., Taylor, A., Calin, A. and Wordsworth, P. (1997) Susceptibility to ankylosing spondylitis in twins: the role of genes, HLA, and the environment. Arthritis Rheum., 40, 1823–1828.[Web of Science][Medline]
10 Brown, M.A., Laval, S.H., Brophy, S. and Calin, A. (2000) Recurrence risk studies in ankylosing spondylitis favour oligogenic epistatic disease model. Ann. Rheum. Dis., in press.
11 Beyeler, C., Armstrong, M., Bird, H.A., Idle, J.R. and Daly, A.K. (1996) Relationship between genotype for the cytochrome P450 CYP2D6 and susceptibility to ankylosing spondylitis and rheumatoid arthritis. Ann. Rheum. Dis., 55, 66–68.
12 Polymeropoulos, M., Rath, D., Xiao, H. and Merril, C. (1992) Dinucleotide repeat polymorphism at the human debrisoquine 4-hydroxylase (CYP2D) locus. Nucleic Acids Res., 19, 1961.
13 Daly, A.K., Brockmoller, J., Broly, F., Eichelbaum, M., Evans, W.E., Gonzalez, F.J., Huang, J.D., Idle, J.R., Ingelman, Sundberg, M. and Ishizaki, T. (1996) Nomenclature for human CYP2D6 alleles. Pharmacogenetics, 6, 193–201.[Web of Science][Medline]
14 Daly, A.K., Armstrong, M., Monkman, S.C., Idle, M.E. and Idle, J.R. (1991) Genetic and metabolic criteria for the assignment of debrisoquine 4-hydroxylation (cytochrome P4502D6) phenotypes. Pharmacogenetics, 1, 33–41.[Medline]
15 Sachse, C., Brockmoller, J., Bauer, S. and Roots, I. (1997) Cytochrome P450 2D6 variants in a caucasian population: allele frequencies and phenotypic consequences. Am. J. Hum. Genet., 60, 284–295.[Web of Science][Medline]
16 Lathrop, G.M. (1983) Estimating genotype relative risks. Tissue Antigens, 22,160–166.[Web of Science][Medline]
17Clayton, D. (1999) A generalization of the transmission/disequilibrium test for uncertain-haplotype transmission. Am. J. Hum. Genet., 65, 1170–1177.[Web of Science][Medline]
18Bandmann, O., Vaughan, J., Holmans, P., Marsden, C.D. and Wood, N.W. (1997) Association of slow acetylator phenotype for N-acetyltransferase 2 with familial Parkinson’s disease. Lancet, 350, 1136–1139.[Web of Science][Medline]
19 McCann, S., Pond, S., James, K. and Le Couteur, D. (1997) The association between polymorphisms in the cytochrome P-450 2D6 gene and Parkinson’s disease: a case-control study and meta-analysis. J. Neurol. Sci., 153, 50–53.[Web of Science][Medline]
20 Wolf, C.R., Dale Smith, C.A., Gough, A.C., Moss, J.E., Vallis, K.A., Howard, G., Carey, F.J., Mills, K., McNee, W., Carmichael, J. and Spurr, N.K. (1992) Relationship between the debrisoquine hydroxylase polymorphism and cancer susceptibility. Carcinogenesis, 13, 1035–1038.
21 May, D., Black, C., Olsen, N., Csuka, M., Tanner, S., Bellino, L., Porter, J., Wilkinson, G. and Branch, R. (1990) Scleroderma is associated with differences in individual routes of drug metabolism: a study with dapsone, debrisoquin, and mephenytoin. Clin. Pharmacol. Ther., 48, 286–295.[Web of Science][Medline]
22 Baer, A., McAllister, C., Wilkinson, G., Woosley, R. and Pincus, T. (1986) Altered distribution of debrisoquine oxidation phenotypes in patients with systemic lupus erythematosus. Arthritis Rheum., 29, 843–850.[Web of Science][Medline]
23 Griem, P., Wulferink, M., Sachs, B., Gonzalez, J. and Gleichmann, E. (1998) Allergic and autoimmune reactions to xenobiotics: how do they arise. Immunol. Today, 19, 133–141.[Web of Science][Medline]
24 Sieper, J. and Braun, J. (1995) Pathogenesis of spondyloarthropathies – persistent bacterial antigen, autoimmunity, or both? Arthritis Rheum., 38, 1547–1554.[Web of Science][Medline]
25 van der Linden, S., Valkenburg, H.A. and Cats, A. (1984) Evaluation of diagnostic criteria for ankylosing spondylitis. A proposal for modification of the New York criteria. Arthritis Rheum., 27, 361–368.[Web of Science][Medline]
26 Steen, V.M., Andreassen, O.A., Daly, A.K., Tefre, T., Borresen, A.L., Idle, J.R. and Gulbrandsen, A.K. (1995) Detection of the poor metabolizer-associated CYP2D6(D) gene deletion allele by long-PCR technology. Pharmacogenetics, 5, 215–223.[Web of Science][Medline]
27 O’Connell, J.R. and Weeks, D.E. (1998) PedCheck: a program for identification of genotype incompatibilities in linkage analysis. Am. J. Hum. Genet., 63, 259–266.[Web of Science][Medline]
28 Satsangi, J., Parkes, M., Louis, E., Hashimoto, L., Kato, N., Welsh, K., Terwilliger, J.D., Lathrop, G.M., Bell, J.I. and Jewell, D.P. (1996) Two stage genome-wide search in inflammatory bowel disease provides evidence for susceptibility loci on chromosomes 3, 7 and 12. Nature Genet., 14, 199–202.[Web of Science][Medline]
29 Kruglyak, L. and Lander, E.S. (1995) Complete multipoint sib-pair analysis of qualitative and quantitative traits. Am. J. Hum. Genet., 57, 439–454.[Web of Science][Medline]
30Risch, N. and Merikangas, K. (1996) The future of genetic studies of complex human diseases. Science, 273, 1516–1517.
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  M. A. Brown Breakthroughs in genetic studies of ankylosing spondylitis Rheumatology, February 1, 2008; 47(2): 132 - 137. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
K. W. Carter, A. Pluzhnikov, A. E. Timms, C. Miceli-Richard, C. Bourgain, B. P. Wordsworth, H. Jean-Pierre, N. J. Cox, L. J. Palmer, M. Breban, et al. Combined analysis of three whole genome linkage scans for Ankylosing Spondylitis Rheumatology, May 1, 2007; 46(5): 763 - 771. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. Damaraju, D. Murray, J. Dufour, D. Carandang, S. Myrehaug, G. Fallone, C. Field, R. Greiner, J. Hanson, C. E. Cass, et al. Association of DNA Repair and Steroid Metabolism Gene Polymorphisms with Clinical Late Toxicity in Patients Treated with Conformal Radiotherapy for Prostate Cancer Clin. Cancer Res., April 15, 2006; 12(8): 2545 - 2554. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. S. Seo, S.-S. Lee, S. I. Kim, W. H. Ryu, K. H. Sa, S. U. Kim, S. W. Han, E. J. Nam, J. Y. Park, W. K. Lee, et al. Influence of VEGF gene polymorphisms on the severity of ankylosing spondylitis Rheumatology, October 1, 2005; 44(10): 1299 - 1302. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 C. Miceli-Richard, H. Zouali, R. Said-Nahal, S. Lesage, F. Merlin, C. de Toma, H. Blanche, M. Sahbatou, M. Dougados, G. Thomas, et al. Significant linkage to spondyloarthropathy on 9q31-34 Hum. Mol. Genet., August 1, 2004; 13(15): 1641 - 1648. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. C. C. Wei, W. C. Tsai, H. S. Lin, C. Y. Tsai, and C. T. Chou HLA-B60 and B61 are strongly associated with ankylosing spondylitis in HLA-B27-negative Taiwan Chinese patients Rheumatology, July 1, 2004; 43(7): 839 - 842. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. R. Whitney, M. Diehn, S. J. Popper, A. A. Alizadeh, J. C. Boldrick, D. A. Relman, and P. O. Brown Individuality and variation in gene expression patterns in human blood PNAS, February 18, 2003; 100(4): 1896 - 1901. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||