SEARCH

SEARCH BY CITATION

Abstract

  1. Top of page
  2. Abstract
  3. PATIENTS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. AUTHOR CONTRIBUTIONS
  7. ROLE OF THE STUDY SPONSOR
  8. Acknowledgements
  9. REFERENCES

Objective

The strongest susceptibility locus of psoriatic arthritis (PsA) is within the major histocompatibility complex (MHC) region (psoriasis susceptibility region 1, or PSORS1), and HLA–Cw*06:02 has been reported as the PSORS1 susceptibility allele. Non-HLA genes within the MHC region have also been implicated in PsA, but because of the strong linkage disequilibrium at chromosome 6p21, it is difficult to make a distinction between susceptibility alleles and linked markers. Recent studies have demonstrated that the association between PsA and the tumor necrosis factor (TNF) promoter polymorphism TNF*-857 is independent of PSORS1. The aim of this study was to replicate the independent association of TNF*-857 in patients with PsA.

Methods

A total of 909 patients with PsA and 1,315 healthy controls originating from the UK, Germany, and Italy were typed for TNF*-857 and for the estimated risk alleles of HLA–Cw*06:02.

Results

Overall, the results of genotyping in these 3 case–control cohorts replicated the finding that the frequency of carriers of TNF*-857 TT/CT who were negative for the PSORS1 risk allele was significantly higher among patients with PsA compared with control subjects (30% versus 21%; P = 9.17 × 10−5).

Conclusion

The results of this collaborative study indicate that TNF*-857T is a susceptibility allele for PsA independent of the PSORS1 allele.

Psoriatic arthritis (PsA) is an inflammatory arthritis associated with psoriasis that is known to affect as many as 30% of patients with typical skin manifestations of psoriasis vulgaris (PsV). In ∼70% of cases, skin manifestations of PsV precede the onset of arthritis by a mean of 10 years (1), whereas in 15% of cases, the arthritis precedes the onset of skin manifestations (2). PsA appears to have a variety of clinical features, in particular, the average age at disease onset and the degree of disease severity, that do not correlate with those of PsV (1).

PsA is a complex disease because of the interaction between several genetic and environmental factors. It has a strong familial clustering (3), with greater heritability than that reported for PsV. Because PsA is strongly related to PsV, the search for genetic factors for PsA has mainly focused on those factors previously found to have an association with PsV. The strongest and most replicated susceptibility region for PsV is within the major histocompatibility complex (MHC) region PSORS1 (for psoriasis susceptibility region 1) (4). This locus maps to chromosome 6p21.3, comprising many different class I antigens associated with disease expression. HLA–C has been repeatedly described as the PSORS1 gene, and HLA–Cw*06:02 as the susceptibility allele. Moreover, non-HLA genes within the MHC region have been implicated in PsV and PsA (5, 6), but the existence of long-range linkage disequilibrium (LD) at 6p21 does not allow a stringent distinction to be made between true susceptibility alleles and markers that are simply linked to the disease.

From a biologic point of view, the association between tumor necrosis factor α (TNFα) promoter polymorphisms and PsA (6) is of particular interest, because increased levels of various cytokines (TNFα, interleukin-1 [IL-1], IL-6, and IL-18), derived primarily from monocyte/macrophages, have been observed in the psoriatic skin, synovial fluid, and synovial membrane of patients with PsA (7). Although an association of polymorphisms in the TNFα promoter region has been reported in various studies and different populations (6, 8), results regarding the exact location of the susceptibility allele (TNF*-238, TNF*-308, or TNF*-857) have been conflicting. Recently, a German study suggested that strong LD with the PSORS1 risk allele caused an association of TNF*-238 with psoriasis, whereas an association of TNF*-857 (T allele) was reported in patients with PsA who were not carrying the PSORS1 risk allele (8).

In order to replicate the latter findings, the current study assessed allele and genotype frequencies of TNF*-857 in 3 independent cohorts, comprising independent case–control samples from Germany (374 PsA patients and 561 healthy controls), Italy (400 PsA patients and 400 healthy controls), and the UK (135 PsA patients and 354 healthy controls) (Table 1). All carriage rates were analyzed following stratification of the subjects according to the presence or absence of the PSORS1 susceptibility allele.

Table 1. Genotype distribution of the HLA–Cw*06:02/PSORS1 risk allele in patients with psoriatic arthritis and healthy control subjects (n = 2,224) of European ancestry from different populations
PopulationNo. of patients/ no. of controlsPSORS1 positivePSORS1 negative
Patients, %Controls, %Patients, %Controls, %
Germany374/56143.812.656.287.4
Italy400/40033.32566.775
UK135/35441.613.358.486.7
Total909/1,31538.916.561.183.5

PATIENTS AND METHODS

  1. Top of page
  2. Abstract
  3. PATIENTS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. AUTHOR CONTRIBUTIONS
  7. ROLE OF THE STUDY SPONSOR
  8. Acknowledgements
  9. REFERENCES

Study populations.

Germany.

All 374 patients with PsA in the case cohort from Germany were of German descent and fulfilled the 2006 Classification of Psoriatic Arthritis (CASPAR) Study Group criteria for PsA (9). This group of patients is a subset of a cohort of patients previously described in the study by Hüffmeier et al (10). The 561 German control probands had no PsV and no history or signs of inflammatory joint disease at the time of recruitment. The study was approved by the ethics committees of the medical faculties of the University of Erlangen–Nuremberg and University of Münster. Written informed consent was obtained from each patient and control proband before enrollment. The investigations were conducted according to the Declaration of Helsinki.

Italy.

The 400 Italian patients with PsA were recruited by clinicians of the Policlinico Tor Vergata in Rome. All of the patients were diagnosed as having PsA by rheumatologists and fulfilled the CASPAR Study Group criteria (9). All patients were negative for rheumatoid factor. The 400 Italian control subjects had no evidence of PsV or PsA at the time of recruitment. The study was conducted according to the Declaration of Helsinki, and informed consent was obtained from all subjects. The study protocol was approved by the ethics committee of the University of Rome Tor Vergata.

UK.

The 135 patients from the UK were recruited from the PsA clinic at the Royal National Hospital for Rheumatic Diseases in Bath. All patients were negative for rheumatoid factor and fulfilled the Moll and Wright criteria for PsA (3). The mean age of the patients was 54 years (range 21–85 years), and just under one-half of the patients (48%) were male. The control group comprised blood samples from 354 healthy Caucasian subjects living in the UK. Ethics approval for the study was given by the Bath Local Research Ethics Committee, and informed written consent was obtained from all participants.

Genotyping.

In the German cohort, for genotyping of HLA–Cw*06:02, a predesigned TaqMan assay (Applied Biosystems) that was previously described in the study by Hüffmeier et al (10) was used. The risk alleles were estimated by genotyping 3 variants in strong LD with the HLA–C risk allele. For genotyping of TNF*-857, another predesigned TaqMan assay (Applied Biosystems), as described in the study by Reich et al (8), was used. In the cohort from Italy, genotyping of HLA–Cw*06:02, as well as genotyping of TNF*-857, was performed using a predesigned TaqMan assay (Applied Biosystems) described previously in the study by De Bakker et al (11). In the UK cohort, for genotyping of HLA–Cw*06:02, sequence-specific primer–polymerase chain reaction (SSP-PCR) was used as previously described in a study by Bunce et al (12), and for genotyping of TNF*-857, SSP-PCR was used as described in a study by McHugh et al (13).

Statistical analysis.

Hardy-Weinberg equilibrium was evaluated in both cases and controls using Haploview software (available at http://www.broadinstitute.org/mpg/haploview). Odds ratios (ORs) and exact 95% confidence intervals (95% CIs) were calculated using an OR calculator (available at http://www.hutchon.net/ConfidOR.htm). Wald's chi-square test for differences between groups was performed using 2 × 2 contingency tables (available at http://www.physics.csbsju.edu/stats/contingency_NROW_NCOLUMN_form.html). To correct the OR values in analyses in which all populations were considered together, we performed the Cochran-Mantel-Haenszel test, with results calculated using R software (available at http://www.r-project.org/); the results from the different cohorts were combined in a meta-analysis. To exclude the presence of heterogeneity between the samples, a Q test was performed, and the associated I2 value, representing the percentage variability in effect estimates due to heterogeneity between the 3 populations, was calculated using online statistical software (available at http://StatTools.net).

RESULTS

  1. Top of page
  2. Abstract
  3. PATIENTS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. AUTHOR CONTRIBUTIONS
  7. ROLE OF THE STUDY SPONSOR
  8. Acknowledgements
  9. REFERENCES

Distribution of the HLA–Cw*06:02/PSORS1 risk allele was assessed in a total cohort of 2,224 individuals of European ancestry (Table 1). Moreover, the same cohort was typed for TNF*-857. Overall, the frequency of allele TNF*-857T was significantly higher in individuals with PsA (27%) than in control subjects (20%) (P = 3.36 × 10−6) (results not shown).

To verify the existence of an association independent of the PSORS1 risk allele, we calculated the genotype frequencies of TNF*-857 (C or T allele) in samples either positive or negative for the PSORS1 risk allele (Table 2). In the Italian PsA cohort, a nominally/weakly significant association with allele TNF*-857T was observed in carriers as well as noncarriers of the PSORS1 risk allele (OR 1.83, 95% CI 1.01–3.31 [P = 0.044] and OR 1.48, 95% CI 1.04–2.10 [P = 0.028]). In the German PsA cohort, a weak protective effect of carriage of the TNF*-857T allele was observed in samples positive for the PSORS1 risk allele (OR 0.54, 95% CI 0.30–0.99; P = 0.046). In the UK cohort, the differences in genotype frequencies between PsA patients and controls did not reach significance (Table 2). Overall, the frequency of heterozygous or homozygous carriers of TNF*-857T (TT/CT) in individuals negative for the PSORS1/HLA–C risk allele was significantly higher (P = 9.17 × 10−5) in patients with PsA (30%) when compared with healthy control subjects (21%) (OR 1.35, 95% CI 1.06–1.72) (Table 2). It should be noted that similar results were obtained when the full German cohort, previously described in the study by Hüffmeier et al (10), was included in the analysis (OR 1.56, 95% CI 1.26–1.92; P = 1.49 × 10−7).

Table 2. Frequencies of TNF*-857 genotypes in carriers and noncarriers of the HLA–Cw*06:02/PSORS1 risk allele among patients with psoriatic arthritis compared with healthy control subjects in the different populations*
 Genotype frequency (no. of subjects)OR (95% CI)P
  • *

    Odds ratios (ORs) with 95% confidence intervals (95% CIs) indicate the likelihood of being a carrier or noncarrier of the PSORS1 risk allele among patients with the TNF-857T allele in comparison with healthy controls. NS = not significant.

Germany   
 HLA–Cw*06:02/PSORS1 positive   
  Cases   
   TNF-857 CC0.76 (124)  
   TNF-857 CT/TT0.24 (39)0.54 (0.30–0.99)0.046
  Controls   
   TNF-857 CC0.63 (45)  
   TNF-857 CT/TT0.37 (26)  
 HLA–Cw*06:02/PSORS1 negative   
  Cases   
   TNF-857 CC0.76 (158)  
   TNF-857 CT/TT0.24 (49)1.27 (0.86–1.88)NS
  Controls   
   TNF-857 CC0.80 (393)  
   TNF-857 CT/TT0.20 (96)  
Italy   
 HLA–Cw*06:02/PSORS1 positive   
  Cases   
   TNF-857 CC0.65 (88)  
   TNF-857 CT/TT0.35 (46)1.83 (1.01–3.31)0.044
  Controls   
   TNF-857 CC0.77 (77)  
   TNF-857 CT/TT0.23 (22)  
 HLA–Cw*06:02/PSORS1 negative   
  Cases   
   TNF-857 CC0.62 (166)  
   TNF-857 CT/TT0.38 (100)1.48 (1.04–2.10)0.028
  Controls   
   TNF-857 CC0.71 (214)  
   TNF-857 CT/TT0.29 (87)  
UK   
 HLA–Cw*06:02/PSORS1 positive   
  Cases   
   TNF-857 CC0.90 (47)  
   TNF-857 CT/TT0.10 (5)0.78 (0.21–2.56)NS
  Controls   
   TNF-857 CC0.87 (41)  
   TNF-857 CT/TT0.13 (6)  
 HLA–Cw*06:02/PSORS1 negative   
  Cases   
   TNF-857 CC0.83 (61)  
   TNF-857 CT/TT0.17 (12)1.14 (0.57–2.29)NS
  Controls   
   TNF-857 CC0.85 (262)  
   TNF-857 CT/TT0.15 (45)  
All   
 HLA–Cw*06:02/PSORS1 positive   
  Cases   
   TNF-857 CC0.74 (259)  
   TNF-857 CT/TT0.26 (90)0.99 (0.67–1.47)NS
  Controls   
   TNF-857 CC0.75 (163)  
   TNF-857 CT/TT0.25 (54)  
 HLA–Cw*06:02/PSORS1 negative   
  Cases   
   TNF-857 CC0.70 (385)  
   TNF-857 CT/TT0.30 (161)1.35 (1.06–1.72)9.17 × 10−5
  Controls   
   TNF-857 CC0.79 (869)  
   TNF-857 CT/TT0.21 (228)  

In order to directly test the significance of a multiplicative interaction term for these 2 loci, a binary logistic regression analysis, in which HLA–Cw*06:02 and TNF*-857 were considered the independent variables and each participant's status (case or control) was considered the dependent variable, was performed in the 3 populations singly, as well as in a merged cohort. In the Italian cohort, the contributions of the TNF*-857 variant and the HLA–Cw*06:02 variant to disease susceptibility were both statistically significant (in case–control comparisons, Wald's χ2 = 7.56 [P = 0.006] and Wald's χ2 = 8.36 [P = 0.004], respectively). In the German and UK cohorts, only HLA–Cw*06:02 showed a significant contribution to disease susceptibility (Wald's χ2 = 105.16 and Wald's χ2 = 41.17, respectively; P < 0.001 for both populations). Finally, in the combined cohort, both TNF*-857 and HLA–Cw*06:02 were significantly associated with PsA (Wald's χ2 = 13.22 and Wald's χ2 = 135.38, respectively; P < 0.001 for both).

As expected, the haplotype analysis confirmed that TNF*-857T was a susceptibility factor independent of the PSORS1/HLA–C risk allele (Table 3). The association of PsA with TNF*-857T was significant in individuals who were not carrying the PSORS1/HLA–C risk allele (OR 1.27, 95% CI 1.05–1.54; P = 0.014). To test the specificity of these variants for the arthritis phenotype, a further typing of PsV samples from Italy (n = 374) and Germany (n = 729) revealed evidence of association with disease susceptibility in the German subset only (P = 0.003 and P = 0.007 in carriers and noncarriers of the HLA–C susceptibility allele, respectively). Further analyses will be necessary to better evaluate the role of TNF*-857 in the pathogenesis of PsA.

Table 3. Frequencies of haplotypes of the HLA–Cw*06:02/PSORS1 risk allele and TNF*-857 in 909 patients with psoriatic arthritis compared with 1,315 healthy control subjects*
HLA–Cw*06:02/PSORS1 status and TNF*-857 alleleHaplotype frequency (no. of subjects)OR (95% CI)P
CasesControls
  • *

    Odds ratios (ORs) with 95% confidence intervals (95% CIs) indicate the likelihood of being a carrier of each haplotype among cases in comparison with healthy controls.

Positive for PSORS1    
 Carrier of allele C0.36 (654)0.15 (401)3.20 (2.77–3.69)1.62 × 10−31
 Carrier of allele T0.02 (44)0.01 (33)1.98 (1.25–3.12)3.19 × 10−3
Negative for PSORS1    
 Carrier of allele C0.49 (864)0.73 (1,924)0.34 (0.30–0.39)3.93 × 10−40
 Carrier of allele T0.13 (228)0.11 (270)1.27 (1.05–1.54)0.014

DISCUSSION

  1. Top of page
  2. Abstract
  3. PATIENTS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. AUTHOR CONTRIBUTIONS
  7. ROLE OF THE STUDY SPONSOR
  8. Acknowledgements
  9. REFERENCES

Multiple genes within the MHC region have been found to be associated with PsA, but the existence of high long-range LD makes it difficult to ascertain the relevance of these genes in disease susceptibility independent of HLA–C. In this collaborative work, we replicated the association of TNF*-857T as a susceptibility allele for PsA independent of the main PSORS1 risk allele. The effect of this promoter variant is weak, and no significant association with PsA has been found in the limited number of samples. However, in the single replication cohorts studied herein, similar trends were observed among all samples negative for the PSORS1 risk allele, with ORs ranging from 1.14 to 1.48 (Table 2). In this large genetics study, we typed 909 patients with PsA and 1,315 healthy controls, and the results robustly demonstrated that TNF*-857T represents a risk allele for PsA independent of the PSORS1 main locus.

Although the functional role of TNF*-857T remains to be determined, previous data have shown that allele T increases the transcription of TNFα (14). Moreover, it is well known that TNFα plays a pivotal role in both the activation and the extravasation of T cells in the highly vascularized synovium, as well as in the promotion of bone erosions during subchondral osteoclastogenesis. Thus, genes encoding for TNFα as well as for the other cytokines associated with PsA (IL-11, IL-15, and IL-23 receptor) might represent candidate pharmacogenetic markers.

In this respect, it is important to note that, in the treatment of rheumatoid arthritis, the presence of TNF*-857T was associated with a good therapeutic response to etanercept, a dimeric, soluble form of the TNF receptor (15). From this perspective, the existence of an international multicenter consortium, such as the Psoriatic Arthritis Genetics European Consortium, should facilitate further prospective pharmacogenomic studies and create the needed link between clinicians and geneticists to properly assess the impact and value of genetic testing of patients with PsA in clinical practice.

AUTHOR CONTRIBUTIONS

  1. Top of page
  2. Abstract
  3. PATIENTS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. AUTHOR CONTRIBUTIONS
  7. ROLE OF THE STUDY SPONSOR
  8. Acknowledgements
  9. REFERENCES

All authors were involved in drafting the article or revising it critically for important intellectual content, and all authors approved the final version to be published. Dr. Novelli had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Study conception and design. Giardina, Hüffmeier, Ravindran, McHugh, Korendowych, Burkhardt, Novelli, Reis.

Acquisition of data. Giardina, Hüffmeier, Ravindran, Lepre, McHugh, Korendowych, Novelli, Reis.

Analysis and interpretation of data. Giardina, Hüffmeier, Behrens, Lepre, Korendowych, Burkhardt, Novelli, Reis.

ROLE OF THE STUDY SPONSOR

  1. Top of page
  2. Abstract
  3. PATIENTS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. AUTHOR CONTRIBUTIONS
  7. ROLE OF THE STUDY SPONSOR
  8. Acknowledgements
  9. REFERENCES

Wyeth Pharmaceuticals had no role in the study design or in the collection, analysis, or interpretation of the data, the writing of the manuscript, or the decision to submit the manuscript for publication. Publication of this article was not contingent upon approval by Wyeth.

Acknowledgements

  1. Top of page
  2. Abstract
  3. PATIENTS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. AUTHOR CONTRIBUTIONS
  7. ROLE OF THE STUDY SPONSOR
  8. Acknowledgements
  9. REFERENCES

We are grateful to all patients and control probands for participation in this study. We thank Petra Badorf for excellent technical assistance. We thank Ken Welsh and Tariq Ahmad for help in supplying the UK control data.

REFERENCES

  1. Top of page
  2. Abstract
  3. PATIENTS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. AUTHOR CONTRIBUTIONS
  7. ROLE OF THE STUDY SPONSOR
  8. Acknowledgements
  9. REFERENCES