Findings

The full aetiology of chronic widespread pain (CWP) is unknown but evidence suggests that genetic risk factors may contribute [1]. As a complex trait it is likely that multiple genes will be involved with each of them having a small effect on the phenotype. Sensitivity to painful stimuli has also been shown to be genetically influenced [2, 3]. The relationship between CWP and pain sensitivity is unclear. Clinic studies have shown increased pain sensitivity in fibromyalgia patients using measures such as thermal and mechanical stimuli and temporal summation [4–8]. However, it is unknown whether the relationship between CWP and pain sensitivity is causal or consequential. Genes which influence pain sensitivity may also be important in susceptibility to chronic pain syndromes such as CWP. Previous studies have reported associations between known and purported functional polymorphisms and pain sensitivity for two genes, GCH1 [9–11] and OPRM1[12, 13].

GCH1 encodes GTP cyclohydrolase which is a pathway synthesis enzyme for Tetrahydrobiopterin (BH4), a cofactor essential in neurotransmitter synthesis which is up-regulated in neuropathic and inflammatory pain [9]. In 2006, a haplotype comprised of 15 SNPs spanning GCH1 was reported to associate with a reduced hyperalgesic response to a mechanical pain stimulus and also with lower levels of lumbar pain following diskectomy [9], thus implicating GCH1 as a potentially important gene in both pain sensitivity and chronic pain. In 2007, Lotsch et al demonstrated that this haplotype could be captured with 100% specificity and sensitivity by genotyping only three single nucleotide polymorphisms (SNPs); rs10483639, rs3783641 and rs8007267, with the C, A and T alleles of these three SNPs respectively forming the "pain-protective" haplotype. Subsequent studies, however, have yielded conflicting results with some confirming the protective effect of the haplotype on pain sensitivity [10, 11] and others not [14]. Null associations between the haplotype and pain severity following dental surgery [14] and in subjects with pancreatitis [15] have also been reported. Tegeder et al (2006) also reported lower levels of GCH1 mRNA in subjects carrying the "pain-protective" haplotype [9]. A subsequent study also found that subjects homozygous for the CAT haplotype had reduced levels of GCH1 mRNA as well as reduced levels of BH4 compared to subjects not carrying the haplotype [10]. In addition, Zhang et al (2007) found that the 3' UTR SNP rs841, which is in LD with the "pain-protective" haplotype, also affects expression of the gene [16].

The second gene of interest, OPRM1, encodes the μ opioid receptor which binds both exogenous and endogenous opiates. The variant G allele of the non-synonymous SNP, A118G (rs1799971) which changes asparagine to aspartic acid, increases the ability of the receptor to bind β-endorphin and has been associated with increased pain thresholds [12]. More recently, Shabalina et al (2009) conducted a comprehensive analysis of the variation within OPRM1 and although they did not observe an association between A118G and pain sensitivity they found the minor (T) allele of rs563649 to be associated with increased pain sensitivity. They also found that rs563649 is located within the 5'UTR of an isoform of the μ opioid receptor in a putative internal ribosome entry site (IRES). In vitro, the T allele increased translational activity through increased ribosome binding and lowered mRNA levels suggesting the SNP may be functionally important [13].

The aim of this study was to determine if the GCH1 "pain-protective" haplotype and the OPRM1 SNPs, rs563649 and rs1799971 are associated with CWP in a UK population-based cohort.

Subjects, aged 25-65 were recruited from 3 primary care registers in the North-west of England into a prospective population-based cohort study (EPIFUND, Epidemiology of Functional Disorders). Pain data was collected at 3 time-points over a 4 year period via a postal survey. A detailed pain questionnaire and body manikins were used to ascertain CWP status using American College of Rheumatology Criteria (pain for ≥3 months in contra-lateral body quadrants above and below the waist and in the axial skeleton) at each time-point. DNA was obtained, using buccal swab sampling, from 1189 subjects who participated at all three time-points and had complete pain data. From this study population, a nested case-control study was conducted with cases being subjects who had CWP for ≥2 time-points and controls were subjects who were pain-free at all 3 time-points. There was a female preponderance (58%) in subjects included in the analysis and the mean ± standard deviation age of subjects was 50 ± 9.6 years old. The proportion of females did not significantly differ between cases and controls (p > 0.05) and cases were significantly older than controls (p < 0.01). Ethnicity was not determined in EPIFUND; however, subjects were recruited from a predominantly white Caucasian geographic area.

The three SNPs in GCH1; rs10483639, rs3783641 and rs8007267, which capture the pain protective haplotype [17] and two SNPs in OPRM1; rs563649 and rs1799971 (A118G), were genotyped in all cases (n = 197) and all controls (n = 197). Genotyping was carried out using Sequenom MassArray technology. Pair-wise LD was examined using Haploview [18]. All SNPs were in Hardy-Weinberg Equilibrium and their genotyping success rate was ≥97%.

The genomic position of the three GCH1 SNPs and the pair-wise LD between them in this UK population is shown in Figure 1. Haplotype frequencies in the cases (N = 164) and controls (N = 172) were estimated using the Expectation-Maximisation algorithm and compared using a χ2test in PLINK [19]. Allele frequencies were compared between cases and controls for the OPRM1 SNPs using a χ2 test. The effect of having one or two copies of the minor allele compared to zero copies (dominant model) was also tested using a χ2 test. Odds ratios (OR) and 95% confidence intervals (95%CI) were calculated using logistic regression. Analysis was conducted using STATA version 9.2.

Figure 1
figure 1

Linkage disequilibrium of the GCH1 haplotype. SNPs genotyped and their position in GCH1 is shown with pair-wise LD (coloured by D' and numbered with r2 values).

Full size image

Two common, GTC (76%) & CAT (17%); and two rare (≤3%), CTC and CAC; GCH1 haplotypes were identified in the population. There was no significant difference in the overall distribution of haplotypes between cases and controls (p = 0.488). The pain-protective (CAT) haplotype was less frequent in cases (15%) compared to controls (19%), although this was also non-significant (p = 0.129). Concurrently the most common haplotype, GTC, showed a non-significant (p = 0.190) increased frequency in cases (81%) compared to controls (77%) (Table 1). This result is in keeping with the findings of previous reports [14, 15] that did not replicate the findings of the initial study which observed association between the haplotype and both pain sensitivity and chronic pain [9].

Table 1 Results of haplotype analysis of GCH1 with CWP
Full size table

There was no evidence of association with CWP for either of the OPRM1 SNPs, although rs563649 showed a trend towards a protective effect of the T allele (Table 2). This is in contrast to the findings of Shabalina et al (2009) who found the T allele to be associated with increased pain scores [13].

Table 2 Results of analysis of OPRM1 SNPs with CWP
Full size table

The aim of this study was to determine if putative functional SNPs reported to influence pain sensitivity are genetic predictors of CWP susceptibility. There was no evidence of a significant association between the GCH1 "pain-protective" haplotype or the two SNPs in OPRM1 and CWP. It should be noted that sample size in the studies of pain sensitivity and chronic pain conducted on these two genes were relatively small with the maximum being N = 632. Equally in this study the sample size was modest therefore the decreased prevalence of the GCH1 CAT haplotype in subjects with persistent CWP compared to pain-free controls may not have reached significance in this study due to limited power. Further investigation of this relationship in larger independent cohorts is therefore warranted.