Introduction
Human leukocyte antigen (HLA) class I molecule
HLA-B27 was the first genetic risk factor identified as associating with ankylosing spondylitis (AS) and remains the most important risk locus for this archetypal spondyloarthropathy[
1]. Twin and family studies estimate that
HLA-B27 accounts for 20 to 50% of the total genetic risk of AS[
2] and confers an odds ratio in European Caucasians >100 for AS[
1]. To date, 100 suballeles of
HLA-B27 have been described[
1]. Of these suballeles,
HLA-B*2701,
HLA-B*2702,
HLA-B*2704,
HLA-B*2705, and
HLA-B*2707 have been associated with AS[
3]. As yet there are few data on whether the other suballeles are associated with altered disease susceptibility.
Data from murine models suggest that HLA-B27 is directly involved in the pathogenesis of AS and it is recognised in human populations that the prevalence of the gene reflects the prevalence of AS[
1]. However, the mechanism by which this HLA protein contributes to disease remains a source of intense speculation. Hypotheses for the role of HLA-B27 in the pathogenesis of AS can be broadly divided into those related to aberrant processing of antigenic peptides and endoplasmic reticulum stress resulting from a tendency for HLA-B27 to misfold and form homo-dimers. The molecular mimicry/cross-tolerance theory relating to specific bacterial antigens is currently less favoured[
1].
The association of
HLAB-27 within the broader group of spondyloarthropathies (SpA) varies significantly, ranging from <50% in psoriatic, enteropathic, and inflammatory bowel disease-associated SpA, to 80% in reactive arthritis, to >95% in AS. The frequency of the
HLA-B27 allele also varies widely across populations[
4]. Both a north–south gradient and an east–west gradient have been observed for
HLA-B27 prevalence in the Northern Hemisphere. It is hypothesised that these gradients result from the negative selection pressure exerted by malaria[
5]. In regions where malaria is endemic, the prevalence of
HLAB-27 is low, and
vice versa[
5].
HLA-B27 is highly prevalent in Northern Eurasia and North America, with 10 to 16% of Norwegians, Swedes and Icelanders and 25 to 50% of Inuit, Yupik and Indigenous Northern Americans (for example, Haida and Bella Coola) carrying this allele[
4].
HLA-B27 prevalence decreases to 9.5% in the United Kingdom[
6], and further decreases to 2 to 6% in Mediterranean regions[
4]. In a similar manner,
HLA-B27 prevalence decreases from west to east. In Southeast Asia prevalence of
HLA-B27 can exceed 12%, but in mainland China the range is between 2 and 6%[
4].
HLA-B27 prevalence also varies significantly within the Pacific Islands. In Melanesia the prevalence is high, whereas
HLA-B27 is uncommon in Micronesia and absent in unmixed native populations of Southeast Polynesia[
4].
At present there is a paucity of prevalence data regarding the
HLA-B27 in the New Zealand population, including indigenous Maori. Three previous studies have included New Zealanders. In the first study, Gonzalez-Roces and colleagues conducted a worldwide survey of
HLA-B27 polymorphisms, and included 12 disease-free New Zealand Māori who tested positive for
HLA-B27[
3]
. The second study related to the prevalence of HLA-B27 in patients presenting to an acute eye service with a history of bilateral or recurrent anterior uveitis. In this study 124 consecutive patients undergoing uveitis screening were typed for HLA-B27. Of these patients, 44 were positive for HLA-B27 and 41% (
n = 18 out of 44) showed early radiologic evidence of AS[
7]. No ethnicity data were available for this patient dataset. In the third study, 116 patients with SpA, 23 patients with uveitis, and 47 patients of unknown disease status were identified by the New Zealand Blood Service as serologically positive for HLA-B27 and sequenced to determine the suballele prevalence[
8]. No healthy controls were included in either of the two latter studies, and none of the SpA patients in the study of Stewart were assessed clinically to determine whether they met classification criteria for AS[
8]. None of these three studies provided prevalence data for
HLA-B27 in New Zealand. Given that
HLA-B27 remains the most important genetic risk factor for the development of AS, and testing is frequently used to assist diagnosis, there is clinical relevance to establishing the prevalence of this
HLA allele, which could help with the planning of health resource allocation in New Zealand.
In countries with a high prevalence of AS, dedicated clinics are increasingly being established to assist with the assessment of patients an initiation of anti-tumour necrosis factor therapies. Furthermore, a recent
HLA-B27 prevalence study from the United States reported a significant decline in the prevalence of
HLA-B27 with age[
9]. The overall age-adjusted
HLA-B27 prevalence in the United States was estimated at 6.1% (95% confidence interval = 5.3 to 10.4), but when age groups were examined separately the prevalence in 40 to 49 year olds was 8.1% (95% confidence interval = 5.8 to 11.2) and dropped to 3.6% (95% confidence interval = 2.2 to 5.8) in 50 to 69 year olds[
9]. This observation raises the possibility that
HLA-B27 status has a detrimental effect on longevity.
Our study had three aims. The first aim was to determine the prevalence of HLA-B27 in three large control datasets (two Caucasian, one New Zealand Māori) and one AS dataset using single-specific primer polymerase chain reaction (PCR-SSP) and the HLA-B27 tagging single nucleotide polymorphisms (SNPs) rs4349859 and rs116488202. The second aim was to determine HLA-B27 suballeles in AS patients of New Zealand Māori ancestry. The third aim was to investigate whether the decline in HLA-B27 prevalence with increasing age, observed in US data, is also observed in the New Zealand population.
Results
A total of 117 Caucasian controls, 111 New Zealand Māori controls and 176 AS patients were typed for
HLA-B27 using PCR-SSP. The prevalence of
HLA-B27 in New Zealand Caucasians, Māori, and AS patients was noted as 7.7%, 6.5% and 93.2%, respectively. These study participants, as well as 1,103 additional Caucasian controls, were then genotyped for the
HLA-B27 tagging SNPs rs4349859 and rs116488202 (Table
2). The concordance between
HLA-B27 and tagging SNP genotypes was 100% in New Zealand Caucasian controls and 98.7 to 99.3% in New Zealand Caucasians with AS, but declined to 76.9 to 85.7% in AS patients and controls of New Zealand Māori ancestry (Table
2). The Caucasian controls comprised two datasets (A1 and A2). The combined prevalence of
HLA-B27 in these datasets, inferred by rs4349859 genotype, was 9.2% (11/1,220). Significant deviation from Hardy–Weinberg equilibrium was observed for rs4349859 in the AS dataset (
P = 2.09 × 10
–11), whereas no deviations from Hardy–Weinberg equilibrium were observed in any of the control datasets for this SNP (
P >0.05).
Table 2
Concordance between
HLA-B27
and tagging SNP rs4349859 and rs116488202 genotypes in New Zealand controls and ankylosing spondylitis patients
Controls | Caucasian | 9/117 (7.7) | 9/117 (7.7) | 9/117 (7.7) | 100.0 | 100.0 |
| New Zealand Māori | 7/107 (6.5) | 6/107 (5.6) | 6/107 (5.6) | 85.7 | 85.7 |
AS | Overall | 164/176 (93.2) | 159/176 (90.3) | 159/176 (90.3) | 96.9 | 97.6 |
| Caucasian | 151/162 (93.2) | 149/162 (91.9)c | 150/162 (92.6)d | 98.7 | 99.3 |
| New Zealand Māori | 13/14 (92.9) | 10/14 (71.4) | 10/14 (71.4) | 76.9 | 76.9 |
A total of 14 patients in the AS dataset were identified as New Zealand Māori by self-report (Table
1). Of these 14 patients one was
HLA-B27-negative, and 13 were
HLA-B27-positive. All 14 patients went forward to DNA sequence-based typing to determine their
HLA-B27 suballele status. Ten were positive for the European suballele
HLAB*2705, and three for the Asian suballele
HLAB*2704.
The two Caucasian control datasets, A1 and A2, were used to determine whether
HLA-B27 prevalence declined with increasing age. Participants in the A1 dataset had a mean age of 44.9 years whereas participants in the A2 dataset had a mean age of 69.9 years (Table
3).
HLA-B27 prevalence for each dataset was inferred from the rs4349859 genotype. In the A1 dataset, 9.1% (44/485) were heterozygous or homozygous for the minor (A) allele compared with 9.2% (68/735) in the A2 dataset (
P = 0.92) (Table
3).
Table 3
Genotype and minor allele frequency of the
HLA-B27
tagging SNP rs4349859 in New Zealand Caucasian controls stratified according to age
A1 (n = 485) | 44.9 | 441 (0.909) | 42 (0.087) | 2 (0.004) | 46 (0.047) | 44 (0.091) | 0.92 |
A2 (n = 735) | 69.9 | 667 (0.907) | 65 (0.088) | 3 (0.004) | 71 (0.048) | 68 (0.092) | |
Discussion
To our knowledge this is the first study to investigate
HLA-B27 in a large number of Caucasian (
n = 1,220) and Māori (
n = 111) controls to determine prevalence of this AS-associated allele in the New Zealand population. The prevalence of
HLA-B27 in New Zealand Caucasian controls (9.2%) was similar to the prevalence of 9.5% previously reported in a dataset of 5,926 UK controls[
6]. Similarly, the overall
HLA-B27 prevalence of 93.2% in our predominately Caucasian AS patient dataset was consistent with the prevalence previously reported in overseas Caucasian AS datasets[
6]. There is a paucity of existing data regarding the prevalence of AS in Māori.
Concordance between the tagging SNPs and
HLA-B27 genotypes was 98.7% for rs4349859 and 99.3% for rs116488202 in the Caucasian AS patients but declined to 76.9% for both SNPs in AS patients of New Zealand Māori ancestry (Table
2). A previous study found that SNP rs116488202 tagged
HLA-B27 more accurately than rs4349859 in both Caucasians and Asians[
17]. In our study, whilst SNP rs116488202 tagged HLA-B27 more accurately for Caucasians this was not the case for New Zealand Māori. Subsequent high-resolution
HLA typing found the three Māori patients with discordant genotypes were all positive for the Asian AS-associated
HLAB*2704 suballele, which neither rs4349859 nor rs116488202 genotyping identified[
16]. To assess the ability of rs116488202 to tag
HLA-B27 suballeles more fully, we sent DNA from three Asian patients (one Filipino, one Han Chinese, one Thai) with AS for high-resolution HLA typing. One patient was found to carry
HLAB*2705 and two patients tested positive for
HLAB*2704. Subsequent rs116488202 genotyping was only able to identify the patient who carried the suballele
HLAB*2705 as HLA-B27-positive. These findings suggest that whilst rs116488202 is a more accurate tagging SNP for HLA-B27 in Caucasians, it does not tag HLA-B27 any more strongly than rs43439859 in Asians or New Zealand Māori. However, it is important to note that our study only had 14 New Zealand Māori and three Asian participants, and therefore does not have the power to allow us to draw any firm conclusions on the relative performance of these two tagging SNPs in individuals of Māori or Asian ancestry.
Direct
HLA typing by PCR-SSP typically costs around NZD$55 per sample, requires 2 μg DNA, and has a 14-day turnaround. In comparison, genotyping by TaqMan® costs NZD$5 per sample, requires 8 to 16 ng DNA, and can be completed within 4 to 8 hours. Given the high concordance between
HLA-B27 and SNP genotypes observed in New Zealand Caucasians (Table
2), genotyping for rs4349859 or rs116488202 could serve as a rapid and inexpensive way of reliably establishing
HLA-B27 status in predominately Caucasian AS datasets.
The distribution of AS-associated
HLA-B27 suballeles differs significantly across populations.
HLAB*2705 is found in all racial groups and thus is widely considered to be the ancestral suballele from which all other variants of
HLA-B27 have arisen.
HLAB*2705 is the most common suballele in Caucasians, but a second, Caucasian-specific suballele (
HLAB*2702) is also seen, albeit at a much lower frequency.
HLAB*2704 is the most common AS-associated suballele in East Asians but has not been reported in Caucasians[
3]. Of the 11 Māori AS patients who underwent high-resolution typing, 10 carried
HLAB*2705 and three carried
HLAB*2704. In the only previous study of
HLAB-27 prevalence to include New Zealand Māori[
3], seven of 12 (58%) individuals carried the
HLAB*2705 allele and five individuals (42%) carried the
HLAB*2704 allele. The presence of the Asian-specific suballele
HLAB*2704 in New Zealand Māori supports a recent investigation into genetic structure of Pacific Islanders, suggesting Polynesians are closely related to Taiwanese Aboriginal populations[
19], a finding strongly supported by anthropological evidence and a common Austronesian language structure. A study of Taiwanese Aboriginal populations demonstrated a differential prevalence of
HLA-B27 amongst different tribes, with a high prevalence of
HLA-B27 (9.2%) in Atayal Aborigines, 2.1% in the Paiwan and none in the Rukai.
HLAB*2704 was the only suballele identified in the aboriginal population[
20]. In the current study, AS patients of New Zealand Māori ancestry have a low prevalence of this
HLAB*2704, indicating that genetic susceptibility to AS in this population is primarily due to significant recent admixture with Caucasians.
The final aim of our study was to investigate the potential link between
HLA-B27 status and longevity. A recent study of
HLA-B27 in 2,320 US adults reported a significant decline in the prevalence of the
HLA-B27 allele with age, which remained significant after adjustment for sex and race[
9]. Although the decreasing linear trend in
HLA-B27 prevalence with age was not entirely consistent in this dataset[
9], the result suggests that
HLA-B27 may negatively impact on longevity, a potentially devastating observation, given that
HLA-B27 is one of the most widely employed clinical genetic tests. Previous investigations offer a possible biological rationale for this observation, although clinical studies following patients with AS have actually shown a reduced standardised mortality ratio in patients with AS[
21].
HLA-B27 has been shown to increase susceptibility to intracellular bacterial pathogens that could in turn increase the risk of atherosclerosis and valvular heart disease, and thus reduce life expectancy. In the current study, however, we found no evidence of a similar decline in
HLA-B27 prevalence in two New Zealand Caucasian datasets of significantly different mean age.
Competing interests
The authors declare that they have no competing interests.
Authors’ contributions
RLR and SMS contributed equally to this study. RLR participated in the design of the study, data analysis, and drafting the manuscript. MCW performed the tagging SNP genotyping of patients and controls, and participated in data analysis and the drafting of the manuscript. GTJ and AMvR provided the A2 control cohort. TRM provided the A1 control cohort. AH, DW, LKS, DC, and JH provided AS patients and associated phenotype data. SMS participated in the design of the study, data analysis, drafting of the manuscript, and coordinated patient recruitment from different centres within New Zealand. All authors read and approved the final manuscript.