Skip Navigation


Rheumatology Advance Access originally published online on October 3, 2008
Rheumatology 2008 47(12):1761-1767; doi:10.1093/rheumatology/ken376
This Article
Right arrow Abstract Freely available
Right arrow FREE Full Text (PDF) Freely available
Right arrow Supplementary Data
Right arrow All Versions of this Article:
47/12/1761    most recent
ken376v1
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Add to My Personal Archive
Right arrow Download to citation manager
Right arrow Search for citing articles in:
ISI Web of Science (2)
Right arrowRequest Permissions
Right arrow Disclaimer
Google Scholar
Right arrow Articles by Harney, S. M. J.
Right arrow Articles by Brown, M. A.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Harney, S. M. J.
Right arrow Articles by Brown, M. A.
Related Collections
Right arrow Immunogenetics
Right arrow Rheumatoid Arthritis
Social Bookmarking
Add to CiteULike   Add to Connotea   Add to Del.icio.us  
What's this?

© The Author 2008. Published by Oxford University Press on behalf of the British Society for Rheumatology. All rights reserved. For Permissions, please email: journals.permissions@oxfordjournals.org

Fine mapping of the MHC Class III region demonstrates association of AIF1 and rheumatoid arthritis

S. M. J. Harney1, C. Vilariño-Güell1, I. E. Adamopoulos1, A.-M. Sims1,2, R. W. Lawrence3, L. R. Cardon3, J. L. Newton1, C. Meisel1, J. J. Pointon1, C. Darke4, N. Athanasou1, B. P. Wordsworth1 and M. A. Brown1,2

1Botnar Research Centre, Nuffield Orthopaedic Centre, University of Oxford, Oxford, UK, 2Diamantina Institute of Cancer, Immunology and Metabolic Medicine, University of Queensland, Brisbane, Australia, 3Wellcome Trust Centre for Human Genetics, University of Oxford, Oxford and 4Welsh Blood Service, Pontyclun, UK.

Correspondence to: M. A. Brown, Diamantina Institute of Cancer, Immunology and Metabolic Medicine, Princess Alexandra Hospital, Ipswich Road, Woolloongabba, Queensland 4102, Australia. E-mail: matt.brown{at}uq.edu.au


   Abstract
Top
 Abstract
Introduction
 Materials and methods
 Results
 Discussion
 Supplementary data
 Acknowledgements
 References
 
Objectives. The heritability of RA has been estimated to be ~55%, of which the MHC contributes about one-third. HLA-DRB1 alleles are strongly associated with RA, but it is likely that significant non-DRB1 MHC genetic susceptibility factors are involved. Previously, we identified two three-marker haplotypes in a 106-kb region in the MHC class III region immediately centromeric to TNF, which are strongly associated with RA on HLA-DRB1*0404 haplotypes. In the present study, we aimed to refine these associations further using a combination of genotyping and gene expression studies.

Methods. Thirty-nine nucleotide polymorphisms (SNPs) were genotyped in 95 DRB1*0404 carrying unrelated RA cases, 125 DRB1*0404-carrying healthy controls and 87 parent-case trio RA families in which the affected child carried HLA-DRB1*04. Quantitative RT–PCR was used to assess the expression of the positional candidate MHC class III genes APOM, BAT2, BAT3, BAT4, BAT5, AIF1, C6orf47, CSNK2β and LY6G5C, and the housekeeper genes, hypoxanthine-guanine phosphoribosyltransferase (HPRT) and β2-microglobulin (B2M) in 31 RA cases and 21 ethnically, age- and sex-matched healthy controls. Synovial membrane specimens from RA, PsA and OA cases were stained by an indirect immunoperoxidase technique using a mouse–anti-human AIF1 monoclonal antibody.

Results. Association was observed between RA and single markers or two marker haplotypes involving AIF1, BAT3 and CSNK. AIF1 was also significantly overexpressed in RA mononuclear cells (1.5- to 1.9-fold difference, P = 0.02 vs HPRT, P = 0.002 vs B2M). AIF1 protein was clearly expressed by synovial macrophages in all the inflammatory synovial samples in contrast to the non-inflammatory OA samples.

Conclusions. The results of the genotyping and expression studies presented here suggest a role for AIF1 in both the aetiology and pathogenesis of RA.

KEY WORDS: Rheumatoid arthritis, Immunogenetics, Major histocompatibility complex


   Introduction
Top
 Abstract
 Introduction
Materials and methods
 Results
 Discussion
 Supplementary data
 Acknowledgements
 References
 
RA is a common complex disease affecting ~1% of most populations worldwide. Twin studies have estimated the heritability of RA to be ~55%, and that about one-third is contributed by the MHC region [1]. The HLA-DRB1 locus in the class II region of the MHC is consistently associated with RA susceptibility in diverse ethnic groups. Susceptibility to RA is conferred by specific HLA-DRB1 alleles that encode a conserved linear sequence of amino acids in the DRβ1 chain, at positions 67–74, known as the shared epitope (SE) [2]. However, the SE does not explain all of the MHC associations in RA. RA is associated with non-SE HLA-DRB1 alleles in some populations [3, 4], different SE alleles have different strengths of disease association [5] and some MHC haplotypes carrying the same DRB1 alleles are differentially associated with RA [6, 7]. Recently, Amos and colleagues [8] have demonstrated significant linkage of RA to the MHC in families where the affected relative pairs do not carry SE alleles [logarithm of the odds ratio (LOD) = 3.9]. Consequently, it appears that other genes within the MHC) are involved in disease susceptibility. Although many studies have investigated the possibility of non-HLA-DRB1 MHC susceptibility genes, most of these studies have taken inadequate account of the extensive linkage disequilibrium in the MHC, resulting in failure to distinguish true associations from those due to linkage disequilibrium with HLA-DRB1 alleles [9]. Very few studies have been performed that control for the DRB1 associations of RA at allelic level, and can thus accurately study association as opposed to linkage disequilibrium with HLA-DRB1 [6, 10, 11].

We have previously identified a single nucleotide polymorphism (SNP) haplotype, which includes the lymphotoxin alpha (LTA) and TNF loci in the MHC class III region of chromosome 6 that shows differential association with RA on HLA-DRB1*0404 and DRB1*0401 haplotypes, suggesting the presence of additional non-HLA-DRB1 RA susceptibility genes on these haplotypes [7]. These findings are supported by Jawaheer et al. [6], who demonstrated association of RA with different MHC haplotypes carrying either HLA-DRB1*0404 or DRB1*0401. Using microsatellite markers we refined the regions identified in our own studies, identifying two three-marker haplotypes in a region of 122 kb in the MHC class III region immediately centromeric to TNF which are strongly associated with RA on HLA-DRB1*0404, but not on DRB1*0401, haplotypes [10]. The most strongly associated haplotype, d6s2671/3-MN6s1997/5-d6s273/5, spans 106 kb and encompasses 12 genes. The genes in this interval include allograft inflammatory factor 1 (AIF1), HLA-B associated transcripts (BAT2, BAT3, BAT4, BAT5), Apolipoprotein M (APOM), Ly6 Superfamily members LY6G5B, LY6G5C, LY6G6C and LY6G6D, Casein Kinase 2subunit beta (CSNK2B) and C6orf47.

In the present study, we aimed to refine these associations further using a combination of genotyping and gene expression studies. Over 600 SNPs were identified in the candidate region by a search of electronic databases including dbSNP. Of these, 39 SNPs, either exonic or lying close to exon–intron boundaries, and with published heterozygosity data, were genotyped in HLA-DRB1*0404 unrelated cases and controls and in parent-case trio families with children carrying HLA-DRB1*0404. Nine of these positional candidate genes showed nominal association with RA (P < 0.05) and we investigated the expression of these using real-time quantitative PCR in peripheral blood mononuclear cells (PBMCs) from RA patients and matched healthy controls. We also compared the expression of AIF1 in the synovium from inflammatory and non-inflammatory disorders.


   Materials and methods
Top
 Abstract
 Introduction
 Materials and methods
Results
 Discussion
 Supplementary data
 Acknowledgements
 References
 
Study population
Cases and controls in this study were British Caucasians. All subjects gave informed written consent and approval had been granted from the local ethics committee. For the genotyping studies, 95 unrelated HLA-DRB1*04-positive RA cases, 125 DRB1*04-positive unrelated controls and 87 parent-case trios were available. All cases satisfied the 1987 ACR criteria for RA [12], and all unrelated cases and controls carried DRB1*0404, as the specific aim of this project was to map the HLA-class III associations we had previously observed on the background of HLA-DRB1*0404-bearing haplotypes. In total, there were 306 DRB1*0401, 290 DRB1*0404 and 366 non-DRB1*0401/0404 haplotypes available for study. Of the cases, 103 DRB1*0401, 128 DRB1*0404 and 127 non-DRB1*0401/0404 haplotypes were available for study.

For the RTq–PCR study, we tested 31 Caucasian RA patients fulfilling the 1987 ACR criteria for RA [12]. These patients were taking standard DMARDs and were compared with 21 ethnically, age- and sex-matched healthy controls.

Genotyping studies
SNP selection and genotyping
Cases, controls and families were genotyped for 39 SNPs in this region. Eleven SNPs (rs2736182, rs9688644, rs1062968, rs6921213, rs1065503, rs3130049, rs2242655, rs2295664, rs2254083, rs10298 and rs2242654) proved to be monomorphic and were excluded from further analysis. The average spacing of the remaining SNPs was 3.9 kb. Details of the polymorphic SNPs, including their locations are summarized in Table 1. All SNPs were genotyped by KBiosciences (Hoddesdon, Herts, UK) using a mixture of competitive allele PCR (KASPar) and TaqMan® genotyping assays (Applied Biosystems, CA, USA). Details of SNP genotyping protocols are available on request. HLA-DRB1*04 subtyping was performed by PCR/amplification refractory mutation system (ARMS) as previously described [7].


View this table:
[in this window]
[in a new window]

  		TABLE 1. Details of 28 polymorphic SNPs typed in this study

 
RTq–PCR study
RNA extraction
Peripheral blood was collected into pre-cooled EDTA tubes and all samples were processed within 60 min. Mononuclear cells were isolated from 15 ml of peripheral whole blood by Ficoll density gradient centrifugation. RNA extraction was performed using the Qiagen RNAeasy kit according to the manufacturer's instructions (Qiagen, Sussex, UK). The RNA samples were DNase treated. The quality and concentration of the RNA samples were assessed using an Agilent 2100 Bioanalyser (Agilent Technologies, CA, USA).

cDNA synthesis
The same reverse transcriptase and protocol was used for all cDNA synthesis reactions. Each reaction contained 2 µg of total RNA, and 2 µl of oligodT primer (100 ng/µl) in a 20-µl reaction volume, which was denatured at 75°C for 10 min then cooled on ice for 2 min. Samples were then treated with DNase by the addition of 8 µl of 5x first strand buffer, 4 µl of 0.1 M dithiothreitol (DTT), 2 µl of dNTPs (10 mM each), 1 µl of DNase (100 U/µl), 0.5 µl of RNase inhibitor (40 U/µl) and 2.5 µl of water. The mixture was then incubated 37°C for 30 min. The DNase was inactivated at 75°C for 5 min, following which the mixture was immediately placed on ice for 2 min. One microlitre each of Moloney Murine Leukemia Virus (M-MLV) (200 U/µl) and RNase inhibitor were added and the sample incubated at 42°C for 90 min then heat inactivated at 94°C for 5 min. The final cDNA concentration was diluted 1:3 with sterile RNAse free water, and stored at –20°C.

Primer and probes for RTq–PCR
In our real-time study, mRNA expression of the nine MHC class III genes that showed evidence of genetic association with RA in any analysis: APOM, BAT2, BAT3, BAT4, BAT5, AIF1, C6orf47, CSNK2B and LY6G5C, and of two housekeeper genes, hypoxanthine-guanine phosphoribosyltransferase (HPRT) and β2-microglobulin (B2M), was measured. The primers and probes for HPRT were designed using Primer Express 1.5 (Applied Biosystems, sequences available on request). The primers and probes for the 11 genes studied were from ‘Assay on Demand’ kits (Applied Biosystems).

RTq–PCR protocol
RTq–PCR was performed using a Rotor-gene 3000 (Corbett Research, Sydney, Australia), and a primer/probe-based system. RTq–PCRs contained 5 µl of universal qPCR mastermix (Eurogentec, Hampshire, UK), 0.5 µl of ‘Assay on Demand’ (Applied Biosystems), 3.5 µl of RNAse free water and 1 µl of cDNA template. Two housekeeping genes, HPRT and B2M, were studied. After an initial single cycle of 50°C for 2 min and 95°C for 10 min, the PCR programme was 45 cycles of 95°C for 15 s and 60°C for 60 s. All RTq–PCRs included a standard curve (SC) using pooled cDNA diluted over a representative range of concentrations, and a common reference calibrator to check for intra-run variability. All standards were run in triplicate; all samples were run in duplicate.

The effect of SNPs on expression of the genes listed above was investigated using publicly available data from lymphoblastoid cell lines from 206 families of British descent (n = 830 individuals). In these samples, expression has been assessed using Affymetrix gene-expression chips (U133 Plus 2.0 chips, including 54 675 distinct probe targets), and tested for association with genomewide SNP genotype data from Illumina HumHap300 and Sentrix-1 chips (covering in total 408 273 SNPs) [13]. Heritability of expression levels and the contribution of individual SNPs to that heritability was also assessed within the studied families.

Immunohistochemistry
Fresh specimens of SM were obtained from patients undergoing hip or knee arthroplasty for OA (n = 6), RA (n = 4) or PsA (n = 2). SM was fixed in formalin and embedded in paraffin wax; 5 µm sections were stained by an indirect immunoperoxidase technique using mouse–anti-human AIF1 mAb (ab15690, Abcam, Cambridge, UK) at a 1:100 dilution.

Serial sections were similarly stained with the mAbs, F7.2.38, L26 and KP1 (Dakopats, UK) directed against CD3 (T cells), CD20 (B cells) and the CD68 (macrophages), respectively. Antigens were detected by incubation with labelled polymer and diaminobenzidine (Envision + kit; Dakopats). The sections were counterstained with haematoxylin. Sections treated with an isotype control IgG2b antibody, (ab9403-1, Abcam) in primary antibody diluent alone were negative controls.

Statistical analysis
Genotyping studies
Genotypes were checked for Hardy–Weinberg equilibrium by chi-squared test (1df) compared with expected values. Mendelian inheritance of markers was verified in single-case families.

Coverage of genetic variation by genotyped SNPs in genes targeted was assessed using HapMap data from Caucasian families (HapMap data release 23a, www.hapmap.org), using Haploview 4.1.

Association analysis was performed using the programme ‘Whap’ (http://pngu.mgh.harvard.edu/purcell//whap/). Whap uses the expectation-maximisation (EM) algorithm to determine the posterior probability of possible haplotypes, then uses all possible haplotypes in the analysis, weighted by their likelihood. To allow both case/control and parent-case trio data to be combined in the analysis, the parent-case trio contribution was restricted to the within-family component of variation. The population prevalence of RA was set at 1%. Single-marker analysis was performed, either controlling for the effect of linkage disequilibrium with HLA-DRB1 or not, as specified in the results, using logistic regression implemented within Whap. Where single-marker findings achieved P < 0.1, two-marker haplotypes with neighbouring SNPs were also studied. Only haplotypes making up >5% of the total available for study were analysed. Results are presented as uncorrected P-values. P-values corrected for the number of markers and haplotypes studied were calculated by permutation testing within Whap. To correct for multiple testing, the number of independent markers genotyped was assessed using spectral decomposition as implemented by the programme SNPSpD [14], which takes into account linkage disequilibrium between marker pairs. Note that because the analysis involves within-family association, conventional measures of effect sizes such as odds ratios are not available.

RTq–PCR analysis
RTq–PCR measures the increase in fluorescent signal ({delta}Rn) generated by the accumulation of product. The threshold cycle (CT) is the cycle at which a statistically significant increase in {delta}Rn is first detected and is related to the amount of target cDNA. All measurements were performed in duplicate. Analysis was done using the two SC method. Patient and control findings were compared by analysis of variance.


   Results
Top
 Abstract
 Introduction
 Materials and methods
 Results
Discussion
 Supplementary data
 Acknowledgements
 References
 
Genotyping studies
All 28 SNP analysed in these studies were in Hardy–Weinberg equilibrium. Spectral decomposition analysis indicates that an uncorrected P-value of 0.0035 (corrected for 14.4 independent observations) was equivalent to a corrected P-value of 0.05.

Of the 28 SNPs genotyped, 26 had been typed by the International HapMap project (rs813115 in BAT3, and rs5872 in CSNK2B had not been). These 26 SNPs captured 78/96 SNPs (81%) in the 12 positional candidate genes with r2 > 0.8 (mean r2 = 0.989).

Single- and two-marker haplotype results are summarized in Table 2. Genotype counts from unrelated cases and controls (excluding parent-case trios) are provided in Supplementary Table 1 (see supplementary data available at Rheumatology Online). Three regions demonstrate nominally significant association with RA (P < 0.05 uncorrected), involving SNPs in AIF1, BAT3 and CSNK2B. At AIF1, association with the two-marker combination AIF1.1-AIF1.2 showed global association with a P-value of 0.0067 (Pcorrected = 0.1), and of the three-marker combination AIF1.1-AIF1.2-BAT2.1 (P = 0.003, Pcorrected = 0.043). The associations at BAT3 and CSNK2B were not significant (Pcorrected > 0.3) and were therefore not studied further.


View this table:
[in this window]
[in a new window]

  		TABLE 2. Single-marker and two-marker haplotype results, uncontrolled and controlled for the effect of linkage disequilibrium between SNPs and HLA-DRB1, and the association of DRB1 with RA

 
Linkage disequilibrium (LD) in the region was assessed and r2-values are shown in Fig. 1. A high degree of LD was observed between markers, which did not decay with distance. There was no significant correlation between LD and intermarker distance (r = –0.06, P = 0.34). Nine blocks of LD were identified, including one block containing markers AIF1.1, AIF1.2 and BAT2.1, in which significant association with disease was observed.


Figure 1
View larger version (60K):
[in this window]
[in a new window]
[Download PowerPoint slide]
  		FIG. 1. Pairwise linkage disequilibrium (r2) across study region. Haplotype blocks define by ‘solid spine of LD’ option in Haploview 4.1. The significantly associated markers AIF1.1, AIF1.2 and BAT2.1 form one haplotype block. The block shades represent strength of linkage disequilibrium ranging from r2 = 1 in black to r2 = 0 in white.

 
RTq–PCR study
Reproducibility and accuracy
Two housekeeping genes were run for each positional candidate gene to improve accuracy of results. Both of the housekeeping genes, HPRT and B2M, were stably expressed. In addition, the common reference sample did not vary between reactions. Lastly, the SC produced nearly identical results in reactions performed on separate occasions. These quality control measures demonstrate that these RTq–PCR findings are highly reproducible.

RTq–PCR findings
A summary of the RTq–PCR results is shown in Table 3. AIF1 was significantly over expressed in RA case mononuclear cells (1.5- to 1.86-fold difference, P = 0.02 vs HPRT, P = 0.002 vs B2M). C6orf47 was significantly underexpressed in RA mononuclear cells (4-fold, P = 0.03 vs HPRT, P = 0.02 vs B2M). We were unable to detect expression of LY6G5C; it may not be expressed in mononuclear cells, or, its expression may be too low to be detected by this method. There was no significant difference in mRNA expression in RA case mononuclear cells compared with controls for the following six genes: APOM, BAT2, BAT3, BAT4, BAT5 and CSNK2B. However, BAT3 showed a non-significant, 2-fold decrease in expression in cases compared with controls.


View this table:
[in this window]
[in a new window]

  		TABLE 3. Findings of RTq–PCR study in RA cases and unrelated gender, age- and ethnicity-matched controls

 
For the case data, disease activity measures were recorded. Expression levels did not significantly correlate with CRP, ESR or anti-cyclic citrullinated peptide (CCP) antibody levels. Notably, all of the RA cases had moderate disease activity (mean ESR = 47 mm/h, mean CRP = 51 mg/l, mean anti-CCP antibody level = 59 U/ml). All patients were receiving disease modifying treatment at the time of study (16 were on anti-TNF biologics, six on MTX and seven on corticosteroids).

Data from lymphoblastoid cell lines indicate that AIF1 expression is highly heritable (heritability = 53%), and strongly influenced by variants close to or within AIF1. The SNP rs2269475 genotyped in our study was strongly associated with AIF1 expression in lymphoblastoid cell lines (P = 4.4 x 10–6, explaining 12% of the heritability of AIF1 expression). The SNP most strongly associated with AIF1 expression, rs2857694 (P = 2.3 x 10–10, explaining 21% of heritability of AIF1 expression), is in strong LD with marker rs2259571 which was associated with RA in the current study (r2 = 0.84, D' = 1). Apart from markers in and around AIF1, no other marker genomewide achieved experiment-wise association with AIF1 expression in this data set. No SNP was associated with expression levels of the other positional candidate genes studied here at experiment-wise significance levels.

Immunohistochemistry
AIF1-expressing cells were found in the subintima of all RA and PsA synovial specimens (Fig. 2a). In serial sections, it was evident that AIF1+ cells were CD68+ subintimal macrophages and not CD3 or CD20 lymphocytes. AIF1 expression was also noted in scattered synovial lining cells in the thickened intima of some RA and PsA cases. The OA synovial specimens which did not contain significant inflammatory infiltrate showed little or no staining for AIF1 (Fig. 2b). In contrast, increased AIF1 staining of subintimal macrophages was seen in the one case of ‘inflammatory OA’, in which the synovium contained numerous scattered chronic inflammatory cells including lymphocytes, plasma cells and macrophages.


Figure 2
View larger version (79K):
[in this window]
[in a new window]
[Download PowerPoint slide]
  		FIG. 2. Representative histological sections of (a) RA synovium showing AIF1 expression in intimal (red arrow) and subintimal macrophages (black arrows). In contrast, (b) OA synovium shows no staining for AIF1 (immunoperoxidase, original magnification 300x).

 

   Discussion
Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
Supplementary data
 Acknowledgements
 References
 
In this study, we have further refined the 106-kb interval in the MHC class III region that we have previously shown to be involved in RA independently of HLA-DRB1 [10]. Using high density SNP typing and a case–control association study design, we identified one region (AIF1-BAT2) with significant association with RA. We were not able to determine whether AIF1.1 and AIF1.2, or BAT2.1, were associated with RA independently of each other.

Although many associations were apparent in the analyses involving HLA-DRB1*04 matched cases and controls most did not persist after fully controlling for linkage disequilibrium with HLA-DRB1 (Table 2). This is consistent with our previous demonstration of marked differences in MHC haplotypes even between closely related HLA-DRB1*04 subtypes [7, 10]. This highlights the necessity of controlling for linkage disequilibrium with HLA-DRB1 at the allelic level in such studies; simple matching for possession of HLA-DR4 or shared-epitope carriage is insufficient to distinguish true association from linkage disequilibrium effects.

No other published study to date has studied this region in this depth, controlling completely for linkage disequilibrium with HLA-DRB1. In a study of 54 MHC microsatellite markers in 469 multicase RA families, with appropriate control for the effects of HLA-DRB1, Jawaheer and colleagues [6] reported association with one microsatellite haplotype on the ancestral A1-B8-DRB1*03 haplotype, a non-SE encoding haplotype. The region of association covered an ~500 kb segment of the central MHC, which did not include HLA-DRB1. There was also evidence of an additional susceptibility element in the class I region of the MHC against the background of DRB1*0404. These findings are consistent with our results showing different disease-modifying elements on the background of DRB1*0401 and DRB1*0404 [7, 10]. Whether the genetic effects observed in the current study are also seen in association with non-DRB1*0404 MHC haplotypes will require further studies matching for each specific MHC-DRB1 haplotype studied. Kilding and colleagues [11] studied 13 SNPs across the MHC class III region controlling for HLA-DRB1 in 164 British Caucasian families, and reported association with SNPs in LST1, BAT1 and PG8, but none of the associations was significant after correction for the number of markers studied. Association of RA with alleles of the D6S273 microsatellite, lying between HSP70 and BAT2, has previously been reported [15]. Zanelli and colleagues [16] also reported association of the region immediately telomeric to TNF in RA, consistent with our own findings. However, neither of these studies fully controlled for linkage disequilibrium with HLA-DRB1 at allelic levels.

To assist in clarifying which specific genes were involved in RA, we investigated the expression of the positional candidate genes by RTq–PCR. Significant differential expression of two genes (AIF1 and C6orf47) in the original 106-kb region of interest was demonstrated by PBMCs from RA patients. C6orf47 is a novel protein, which interacts with fibroblast growth factor receptor 3 (FGFR3), an important mediator of cartilage development in the growth plate; FGFR3 mutations are involved in achondroplasia and thanatophoric dysplasia and it seems an unlikely candidate for RA [17]. Further, it was not associated with RA in the current genetic association study. Conversely, AIF1, which encodes a 17 kDa IFN-{gamma} inducible Ca++-binding EF hand protein, is a good candidate gene for involvement in RA, a Th1 cytokine-mediated disease. Historical data from lymphoblastoid cell lines indicate that AIF1 expression is strongly influenced by cis-acting genetic variants, in contrast to each of the other genes in the region studied. This would support the hypothesis that the association of AIF1 with RA operates through effects on AIF1 transcription. Our own findings, and the range of conditions in which AIF1 has been implicated, suggest that the gene has a role in a diverse range of inflammatory conditions rather than being specific for RA. This is further supported by the finding that AIF1 overexpression is found not only in synovium from RA patients, but also in PsA, and not in non-inflammatory OA controls.

Our findings support recent observations made by Kimura and colleagues [18], who not only demonstrated AIF1 expression in RA (but not OA) synovium but also that this was strongly correlated to IL-6 levels. They also noted that exposure to AIF1 stimulated synovial proliferation and IL-6 production. Overexpression of AIF1 has also been identified in a microarray study of RA synovium; eight other MHC genes were differentially expressed in RA synovium but none of these lie within 300 kb of AIF1 [19]. This makes it unlikely that other genes in linkage disequilibrium with AIF1 but not studied in our RTq–PCR study are also differentially expressed in RA synovium. Our findings extend these observations by demonstrating that AIF1 polymorphisms are directly associated with RA, and thus suggest that AIF1 is not only involved in the pathogenesis of RA but also in its genetic aetiology.

Although only moderately powered, this is one of the largest studies to date of non-HLA-DRB1 MHC genes in RA [9], and one of few that can differentiate true from spurious association due to the known differential associations of HLA-DRB1 SE alleles with RA and the effects of linkage disequilibrium in the MHC. Nonetheless, studies involving larger sample sizes and more dense genotyping will still be required to replicate these findings and provide a comprehensive analysis of the region. The level of statistical association we observed is also amongst the strongest reported for non-DRB1 MHC associations in RA, particularly having corrected as we have for the number of SNPs genotyped. As is the case with all genetics studies, replication of these findings is required. The combination of our genetic, gene-expression and immunohistochemistry data, along with the observations of Kimura and colleagues strengthens the likelihood that these are true positive findings.

AIF1 originally came to attention because it is highly expressed in response to balloon-induced vascular injury and allograft rejection [20]. Its expression is induced by a variety of pro-inflammatory cytokines, including IFN-{gamma}, IL-1β and TNF-{alpha}. Transient transfection of macrophage cell lines with AIF1 constructs reduce macrophage apoptosis, promote iNOS production and increases macrophage migration, whereas down-regulation of AIF1 using siRNA has the opposite effect [21]. AIF1 overexpressing transgenic mice are partially resistant to 2,4,6-trinitrobenzene sulfonic acid (TNBS)-induced colitis, an IL-1-dependent model of IBD, raising the possibility that AIF1 is involved in the control of Th-1 responses [22]. The effect of loss-of-function of AIF1 has not yet been reported in in vivo models. The expression of AIF1 in different forms of arthritis, its correlation with disease severity, and the effect of genetic variation in AIF1 on its expression and function await further studies.

In summary, we have shown a significant association between RA and an MHC haplotype containing AIF1-BAT2, independent of HLA-DRB1. Using RTq–PCR, we have demonstrated increased expression of AIF1, but not BAT2, in PMBCs of RA patients compared with healthy controls. Finally, we have shown expression of AIF1 in synovial macrophages in the lining layer of all the inflamed SMs (including RA, PsA and inflammatory OA) compared with non-inflamed OA controls. These findings combined with the recent data from Kimura and colleagues [18], strongly suggest that AIF1 has a significant role in the aetiopathogenesis of RA.

Formula


   Supplementary data
Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 Supplementary data
Acknowledgements
 References
 
Supplementary data are available at Rheumatology Online.


   Acknowledgements
Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 Supplementary data
 Acknowledgements
References
 
The researchers would like to thank the individuals who provided samples for this study, and are grateful for the financial support of the Arthritis Research Campaign (UK), and Arthritis Australia.

Disclosure statement: I.E.A. is an employee of Schering-Plough Corp., Inc and receives salary/stock from them. L.R.C., following completion of the work detailed in the paper, accepted an appointment at GlaxoSmithKline. All other authors have declared no conflicts of interest.


   References
Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 Supplementary data
 Acknowledgements
 References
 

  1. MacGregor AJ, Snieder H, Rigby AS, et al. Characterizing the quantitative genetic contribution to rheumatoid arthritis using data from twins. Arthritis Rheum (2000) 43:30–7.[CrossRef][Web of Science][Medline]
  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) 30:1205–13.[Web of Science][Medline]
  3. Jacobsson LT, Pillemer SR. What can we learn about rheumatic diseases by studying Pima Indians? J Rheumatol (1994) 21:1179–82.[Web of Science][Medline]
  4. Harvey J, Lotze M, Arnett FC, et al. Rheumatoid arthritis in a Chippewa band. II. Field study with clinical serologic and HLA-D correlations. J Rheumatol (1983) 10:28–32.[Web of Science][Medline]
  5. Hall FC, Weeks DE, Camilleri JP, et al. Influence of the HLA-DRB1 locus on susceptibility and severity in rheumatoid arthritis. Q J Med (1996) 89:821–9.
  6. Jawaheer D, Li W, Graham RR, et al. Dissecting the genetic complexity of the association between human leukocyte antigens and rheumatoid arthritis. Am J Hum Genet (2002) 71:585–94.[CrossRef][Web of Science][Medline]
  7. Newton J, Brown MA, Milicic A, et al. The effect of HLA-DR on susceptibility to rheumatoid arthritis is influenced by the associated lymphotoxin alpha-tumor necrosis factor haplotype. Arthritis Rheum (2003) 48:90–6.[CrossRef][Web of Science][Medline]
  8. Amos CI, Chen WV, Lee A, et al. High-density SNP analysis of 642 Caucasian families with rheumatoid arthritis identifies two new linkage regions on 11p12 and 2q33. Genes Immun (2006) 7:277–86.[CrossRef][Web of Science][Medline]
  9. Newton JL, Harney SM, Wordsworth BP, Brown MA. A review of the MHC genetics of rheumatoid arthritis. Genes Immun (2004) 5:151–7.[CrossRef][Web of Science][Medline]
  10. Newton JL, Harney SM, Timms AE, et al. Dissection of class III major histocompatibility complex haplotypes associated with rheumatoid arthritis. Arthritis Rheum (2004) 50:2122–9.[CrossRef][Web of Science][Medline]
  11. Kilding R, Iles MM, Timms JM, Worthington J, Wilson AG. Additional genetic susceptibility for rheumatoid arthritis telomeric of the DRB1 locus. Arthritis Rheum (2004) 50:763–9.[CrossRef][Web of Science][Medline]
  12. Arnett FC, Edworthy SM, Bloch DA, et al. The American rheumatism association 1987 revised criteria for the classification of rheumatoid arthritis. Arthritis Rheum (1988) 31:315–24.[Web of Science][Medline]
  13. Dixon AL, Liang L, Moffatt MF, et al. A genome-wide association study of global gene expression. Nat Genet (2007) 39:1202–7.[CrossRef][Web of Science][Medline]
  14. Nyholt DR. A simple correction for multiple testing for single-nucleotide polymorphisms in linkage disequilibrium with each other. Am J Hum Genet (2004) 74:765–9.[CrossRef][Web of Science][Medline]
  15. Singal DP, Li J, Ye M, Lei K. D6S273 microsatellite polymorphism and susceptibility to rheumatoid arthritis. Tissue Antigens (1998) 52:353–8.[Web of Science][Medline]
  16. Zanelli E, Jones G, Pascual M, et al. The telomeric part of the HLA region predisposes to rheumatoid arthritis independently of the class II loci. Hum Immunol (2001) 62:75–84.[CrossRef][Web of Science][Medline]
  17. Shiang R, Thompson LM, Zhu YZ, et al. Mutations in the transmembrane domain of FGFR3 cause the most common genetic form of dwarfism, achondroplasia. Cell (1994) 78:335–42.[CrossRef][Web of Science][Medline]
  18. Kimura M, Kawahito Y, Obayashi H, et al. A critical role for allograft inflammatory factor-1 in the pathogenesis of rheumatoid arthritis. J Immunol (2007) 178:3316–22.[Abstract/Free Full Text]
  19. van der Pouw Kraan TC, van Gaalen FA, Huizinga TW, Pieterman E, Breedveld FC, Verweij CL. Discovery of distinctive gene expression profiles in rheumatoid synovium using cDNA microarray technology: evidence for the existence of multiple pathways of tissue destruction and repair. Genes Immun (2003) 4:187–96.[CrossRef][Web of Science][Medline]
  20. Utans U, Arceci RJ, Yamashita Y, Russell ME. Cloning and characterization of allograft inflammatory factor-1: a novel macrophage factor identified in rat cardiac allografts with chronic rejection. J Clin Invest (1995) 95:2954–62.[Web of Science][Medline]
  21. Yang Z, Ho D, Lau C, et al. Allograft inflammatory factor-1 (AIF-1) is crucial for the survival and pro-inflammatory activity of macrophages. Int Immunol (2005) 17:1391–7.[Abstract/Free Full Text]
  22. Morohashi T, Iwabuchi K, Watano K, et al. Allograft inflammatory factor-1 regulates trinitrobenzene sulphonic acid-induced colitis. Immunology (2003) 110:112–9.[CrossRef][Web of Science][Medline]
Submitted 22 January 2008; revised version accepted 19 August 2008.
Add to CiteULike CiteULike   Add to Connotea Connotea   Add to Del.icio.us Del.icio.us    What's this?



This Article
Right arrow Abstract Freely available
Right arrow FREE Full Text (PDF) Freely available
Right arrow Supplementary Data
Right arrow All Versions of this Article:
47/12/1761    most recent
ken376v1
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Add to My Personal Archive
Right arrow Download to citation manager
Right arrow Search for citing articles in:
ISI Web of Science (2)
Right arrowRequest Permissions
Right arrow Disclaimer
Google Scholar
Right arrow Articles by Harney, S. M. J.
Right arrow Articles by Brown, M. A.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Harney, S. M. J.
Right arrow Articles by Brown, M. A.
Related Collections
Right arrow Immunogenetics
Right arrow Rheumatoid Arthritis
Social Bookmarking
Add to CiteULike   Add to Connotea   Add to Del.icio.us  
What's this?