Patients and controls

This research was approved by the local Research Ethics Committee at the Sultan Qaboos University (SQU) and the consent was obtained from patients for blood sample collection. The samples were obtained from 30 patients (age = 15 years old, 27 females and 3 males) seen in the outpatient rheumatology clinic in Sultan Qaboos University Hospital (SQUH). Patients having active disease and those in remission were included in this study. Among these, 16 individuals were sero-positive for rheumatoid factor (RF+) and 14 individuals were sero-negative for rheumatoid factor (RF-). Fourteen healthy volunteers from the same ethnic background were enrolled in the study as controls. The diagnosis was based according to the American College of Rheumatology criteria for RA (Arnett et al. (1988)).

Laboratory and radiographic studies

Patient’s history, physical results and laboratory analysis recorded in the Hospital Information System (HIS) were retrieved. Patients with additional inflammatory connective tissue diseases, HIV infection, as well as patients undergoing chemo or radiotherapies were excluded from the study. Patient age, age at diagnosis, previous treatments, history of systemic disease, family history and morning stiffness were recorded. Swollen and tender joints were examined and number of such joints was recorded. DAS28 scores were calculated using C-reactive protein (CRP) values, number of swollen and tender joints. Anti-CCP and RF values were also determined for each patient.

PCR amplification and sequencing

Blood samples (2-5 ml) were collected from patients and genomic DNA was isolated using QIAamp Blood Maxi kit DNA extraction according to the manufacturer’s instructions (QIAGEN, USA). The full coding sequence of the HLA-DRB1 gene was amplified using primers specific for each of the 6 exons. Amplification included a 10-min denaturation set up at 94°C, followed by 35 cycles each consisting of 50s denaturation at 94°C, 50s of annealing at temperatures ranging from 51 to 62°C depending on each exon's (melting temperature), and 50s extension at 72°C. After the last cycle, the samples were incubated for 10 min at 72°C for final extension. The PCR product from each exon was visualized using 1.5% agarose gel electrophoresis.

PCR products of HLA-DRB1 gene were sequenced using the ABI BigDye® Terminator v3.1 Cycle Sequencing Kit. The conditions of the sequencing reaction included 25 cycles at 95°C (70s), 55°C (5s), 60°C (60s), 4°C (holding temperature The products were cleaned with 95% and 70% ethanol, solubilized in deionized formamide and analyzed using the ABI Prism Genetic Analyzer.

Test sequences were aligned with the reference sequence obtained from the Ensembl database, and analyzed using the Sequencher 5.0 Demo software. In addition, protein 3-dimensional structures were visualized with the 3D program Ras Win Molecular Graphics windows version 2.7.5.2. The PolyPhen-2 and PROVEAN (PROtein Variation Effect Analyzer v1.1) online tools were utilized to determine the functional alterations caused by the identified variations in the RA patients. The insertions, deletions and premature stop codon mutations were further analyzed to predict the resulting loss of information using the Indelz online tool (http://www.moseslab.csb.utoronto.ca/amin/indelz.html).

Clinical data results

Thirty blood samples from Omani patients affected with RA and fourteen samples from healthy subjects were examined for HLA-DRB1 gene mutations. Females constituted 90% of the patient cohort and the mean age was 40 ± 13.4 with the age range between 19 and 69 years. 63.3% of patients presented with positive anti-cyclic citrullinated peptide and 56.7% were positive for rheumatoid factor. According to the calculated DAS score, 30% of patients presented with high active disease, 47% presented with moderate active disease and 23% were in remission.

Sequence analysis of HLA-DRB1 gene

The exon sequences of the HLA-DRB1 gene from the study cohort were examined using the publicly available Sequencher Demo version 5 software. Numerous mutations were identified in exons 1 and 2 (Tables 1 and 2). Analysis of exons 3, 4 and 5 revealed no aberration when compared to sequence references. However, in exon 6, although patients 1-4, 9 and 11-30 displayed a specific product for sequencing, five patients numbered 5, 6, 7, 8 and 10 failed to display any product consistently after three consecutive assays suggesting that exon 6 may be deleted in people affected with this disease.

As summarized in Figure 1, among the total aberrations identified in the HLA-DRB1 gene in this study, 20% were polymorphisms, 26.6% of which were silent. The remaining 68% were mutations, 25.49% of them being silent mutations; thus showing no change in the amino acid sequence. Additionally, 6 deletions and 3 insertions were identified.

Figure 1: Summary of geneticaberration identified in the HLADRB1 gene in an Omani rheumatoid arthritis cohort.

The main aberrations found in exon 1 are summarized in Table 1 where the status of the HLA-DRB1 in rheumatoid arthritis patients is compared to healthy controls. Out of 30 cases and 14 controls analyzed we confirmed 2 polymorphisms, which were previously reported, and eleven mutations, 3 of which have been previously reported. Six of the identified mutations were silent mutations. Amongst the non-synonymous mutations, we identified two novel mutations, L4P and C11R in the RA patient cohort and were absent in the controls. The PolyPhen-2 and PROVEAN (protein variation effect analyzer v1.1) online tools were utilized to determine the functional alterations caused by the identified variations in the RA patients. The PolyPhen-2 variation score ranges from 0 (Benign) to 1.0 (Damaging) whereas the PROVEAN tool has a set threshold score of -2.5, below which variations are classed as deleterious. Prediction of the functional effect of the variation using the PolyPhen-2 (v2.2.2r398) online tool, classed these two mutations as “benign” (score=0.082 and 0.001, respectively). The S29A and T13A variations, which have been previously reported, were present at a high frequency in the patient cohort compared to the other identified polymorphisms/mutations. However, these reported variations have not been associated with RA or any other disease and were also classes as “benign” by the PolyPhen-2 prediction tool. The identified benign variants may be in strong linkage disequilibrium with other causal alleles.

Table 1: A summary of aberrations found in exon 1 of the HLA-DRB1 gene. Nucleotide numbering starts from the first codon (ATG) encoding for Methionine.

Several mutations were identified in exon 2 of HLA-DRB1 in rheumatoid arthritis patients compared to healthy controls and are listed in Table 2. Thirteen polymorphisms were recorded, of which 11 have been previously reported, in addition to three insertions and six deletions. Moreover, 40 mutations were identified in total, 10 of which have not been previously reported, 7 of which were synonymous and one caused a stop codon (W38-, G124A). Only 3 mutations displayed overlapping results in both the PolyPhen-2 and PROVEAN tools as functionally deleterious variations: V67C, E51K and N48T (Table 3), all 3 of which occur in the conserved MHC II beta domain.

Table 2: A summary of aberrations found in exon 2 of the HLA-DRB1 gene. Nucleotide numbering starts from the first (ATG) codon.

Table 3: Mutations identified by both the PolyPhen2 and PROVEAN online tools as functionally deleterious. Mutations identified by both the PolyPhen2 and PROVEAN online tools as functionally deleterious.

The insertions, deletions and premature stop codon mutations were further analyzed to predict the resulting loss of information by utilizing the Indelz online tool (http://www.moseslab.csb.utoronto.ca/amin/indelz.html), which provides a Loss score D ranging from 0 to 1, where 0 signifies no change to the protein and 1 signifies complete loss of amino acid information for proper protein function [26]. Interestingly, all the identified insertions and deletions led to a significant loss of information (>75%) due to protein truncation. The most significant effect was displayed by the deletions at nucleotide position 109, 111, 115, insertion at position 119 and a premature stop codon at amino acid position 38, all of which span the first beta strand domain of the HLA-DRB1 protein.

Amongst the variations identified in exon 2, amino acid substitution at position 41 occurred at the highest frequency in the studied RA cohort, whereby the native threonine residue was more frequent than the lysine residue in the RA cases (43.33% compared to none in the control samples). However, this variation was classed as benign/neutral by both the PolyPhen-2 (score=0.003) and PROVEAN tools (score=3.691) indicating the possibility of strong linkage disequilibrium between this variation and another causal mutation. Three-dimensional modeling of the HLA-DRB1 protein with amino acid variation at position 41 indicated a conformational change between the native (A) and K41 (B) form. The affected beta strand domain harboring the mutation is indicated by an arrow. The function of the affected domain juxtaposed to the folding domain on the DRB1 is not yet known.

The heritability of rheumatoid arthritis has been predicted to be approximately 50%, laying emphasis on the strong role of genetics in RA disease susceptibility [27]. Moreover, only 30% of the disease risk i.e. half of the heritability, could be explained by common causal variants identified by GWAS studies, thus pointing out to the role of rare causal variants in RA patients that need to be identified by sequencing. Although GWAS studies have identified several disease risk loci [28] in genes such as peptidyl arginine deiminase type 4 (PADI4), protein tyrosine phosphatase nonreceptor 22 (PTPN22), tumour necrosis factor alpha- induced protein 3 (TNFAIP3), signal transducer and activator of transcription 4 (STAT4) and chemokine receptor 6 (CCR6) in numerous RA cohorts, the HLA-DRB1 gene has been the major focus of RA studies since the identification of variant alleles that confer risk or protection against the disease in RA patients. The purpose of this study was to establish the pattern of known and novel mutations of HLA-DRB1 in Omani patients affected with rheumatoid arthritis and the association of these genetic variants with serological and clinical data in the cohort.

Amongst the genetic aberrations identified in the Oman RA patients, a majority segregated in exon 2 of the HLA-DRB1 gene. Three novel non-synonymous mutations were predicted to have a damaging effect on protein function: V67C, E51K and N48T, which occur in the MHC class II beta domain. The V67 amino acid occurs in the beta-chain region of the HLA-DRB1 protein and the substitution of valine, a non-polar amino acid, with cysteine, a polar amino acid, may potentially affect the antigen-presentation properties of the HLA-DRB1 protein. On the other hand, E51 and N48 both flank the first beta-turn region of HLA-DRB1. Mutations at these amino acid residues resulting in charge alterations could affect the proper folding of the protein. These mutations were identified in only one patient each highlighting their rare occurrence. Although these mutations do not occur in the shared epitope conserved domain (spanning amino acid residues 70-74) that has been associated with RA and the levels of citrullinated protein antibody (ACPA) [29], they may indirectly affect protein function by altering the 3-dimensional structure of HLA-DRB1. Moreover, several insertions and deletions in exon 2 of HLA-DRB1 were identified in the RA patients that resulted in significant loss of genetic information and protein truncation. The inability to amplify exon 6 in 5 RA patients also highlighted the possibility of the presence of deletions in this region of the gene.

Analysis of a large reference population identified the risk associated with amino acid positions 11, 13, 71 and 74 of HLA-DRB1 in seropositive rheumatoid arthritis patients [20]. Although a rare mutation (Cysteine to Arginine) in the amino acid 11 was identified in the Omani RA cohort, none of the previously highlighted high-risk alleles coding for Val-11 or Leu-11 are reported here. A non-synonymous mutation at position 13 (Threonine to Alanine) was also reported in 50% of the Omani RA patients, the most common variation in the studied cohort. This variant was not also associated with RA in the Raychaudhuri et al study and hence, may indicate a novel rare mutation or strong linkage with alternative causal alleles at position 11.

In exon 1, S29A is a well-established polymorphism associated with RA in a Korean population. This polymorphism occurs in 21.4% of healthy subjects and 76.66% of patients and may affect the folding of the protein structure due to the change from a polar (serine) to a nonpolar amino acid (alanine).

A 67-year old RA female suffering from small tender swelling joints and positive RF with an abnormally high level of mutations in the HLA-DRB1 gene, when compared to other patients, displayed 2 polymorphisms, 19 mutations, 3 deletions and one insertion. Among these aberrations, two mutations (W38-, R101-) resulted in the incorporation of premature stop codons in the protein. Thus, these aberrations in combination may have contributed to the formation of an abnormal protein or its complete absence, in line with the critically high RF value displayed by this patient implying the highly aggressive nature of the disease.

The large number of polymorphisms and mutations identified in this study albeit the small size of the population, further confirms the highly polymorphic nature of the HLA-DRB1 gene, the presence of broad linkage disequilibrium and the contribution of multiple genetic aberrations in determining disease aggressiveness.

Both ACCP and RF levels have shown strong association with rheumatoid arthritis severity in several studies and has hence, been included in the RA classification criteria [29,30]. Autoantibodies specific to cyclic citrullinated peptides appear at early stages of RA onset and their specificity and production have been shown to be regulated by HLA SE alleles [15]. It also has a better predictive value compared to RF, which is non-specific and absent in early stages of RA [31]. It has been suggested that the higher affinity of HLA SE alleles to bind to citrullinated peptides may play a key role in activating the T cell response and B-cell mediated production of autoantibodies [32]. However, the lack of binding of certain citrullinated peptides (e.g. Epstein Barr virus-derived) or the ability of some HLA-DRB1^* alleles to equally bind both the native and citrullinated forms of the peptide indicates the presence of alternative mechanisms than antigen-presentation by which ACCP is involved in RA pathogenesis. In this study no association was identified between the number of HLA-DRB1 genetic aberrations and ACCP or RF status. This observation is in line with results from a study indicating that the various HLA-DRB1 genotypes affect the expression level of the protein irrespective of disease activity or severity [33]. Hence, this further strengthens the claim that serum markers such as ACCP and RF and the HLA-DRB1 gene aberration status independently contribute to the pathogenesis of the RA. Thus, the laboratory analysis criteria used to diagnose RA does not reflect the patients’ genomic status and may affect appropriate disease diagnosis. It is of primary interest to dissect the contribution of each of these parameters and their effect on RA prognosis with molecular biology studies to ensure systematic patient and disease management.

Although this study is the first investigation of the association of HLA-DRB1 genetic aberrations, in rheumatoid arthritis within the Omani population, there are some caveats that need to be addressed in the future. Larger cohort with matched controls and detailed description of diagnostic parameters are necessary to determine the significance of these mutations and their association with the clinical and pathological data. The gender skew in the sample population reflects the higher disease prevalence of RA in females with a worldwide ratio of 3:1 [2], although this ratio was higher in the Oman cohort studied despite the relatively lower prevalence of RA in Oman, further stressing on the need to collect larger sample populations to prevent biased observations. The possibility that the rare variations identified in this study display ethnic specificities needs to be tested by sequencing studies in different ethnic populations.

In conclusion, this study identified several novel polymorphisms and mutations in the HLA-DRB1 gene in an Omani rheumatoid arthritis population irrespective of disease severity. The functional consequences of mutations predicted to have a deleterious effect on the protein need to be elucidated in detail, in addition to the confirmation of these genetic aberrations in larger populations.