Introduction

Alopecia areata is characterized by patchy hair loss with perifollicular and intrafollicular T-cell infiltration (Kalish et al., 1992; McDonagh & Messenger, 1996). The association of alopecia areata with autoimmunity, including thyroid disorders, pernicious anaemia and vitiligo, is well established (Muller & Winkelmann, 1963; Cunliffe et al., 1968) and alopecia areata itself is conventionally regarded as an autoimmune disease.

There are no reliable population-based data on the prevalence of alopecia areata in the UK but the lifetime risk of disease is thought to be about 1.7% in the USA with similar incidence in males and females (Safavi et al., 1995). Up to 40% of individuals with alopecia areata have a positive family history of the disease (Anderson, 1950; van der Steen et al., 1992) and twin studies (Cole & Herzlinger, 1984; Scerri & Pace, 1992) have confirmed that there is a genetic component to the disorder. Several investigators have shown an association between alopecia areata and particular HLA class I alleles including A28, B12, B13, B18, B27 and Cw3 (Kianto et al., 1977; Hacham-Zadeh et al., 1981; Orecchia et al., 1987; Price & Colombe, 1996); but HLA Class II associations including DR4, DR11, DPw4, DQw3, DQw7 and DQw8 are considered more important (Frentz et al., 1986; Odum et al., 1990; Duvic et al., 1991). A high relative risk of disease for HLA DR5 (RR=3.14, P < 0.01) was found in patients with severe disease and early age at onset (Price & Colombe, 1996) and the importance of HLA genes has been confirmed in the only family study reported to date (de Andrade et al., 1999). However, the HLA contribution alone cannot explain the entire genetic basis of alopecia areata, and we have found evidence to suggest the presence of another alopecia areata locus on chromosome 21 (Tazi-Ahnini et al., 2000). Interaction between HLA and other loci is probably required to produce the disease phenotype, as shown for systemic lupus erythematosus (Tjernström et al., 1999).

Interleukin-1 is a highly pro-inflammatory cytokine that promotes recruitment of T-lymphocytes, neutrophils and macrophages to inflamed tissues (Dinarello, 1996). IL-1α, IL-1β and TNF-α are also known to have an inhibitory effect on hair growth in vitro (Harmon & Nevins, 1993; Philpott et al., 1996). Inhibitory doses of these three cytokines have similar effects on the morphology of cultured explant hair follicles resulting in a dystrophic anagen pattern, characterized by condensation of the dermal papilla with disruption and abnormal keratinization of the pericortical cells of the hair matrix. These features are similar to the follicular pathology of alopecia areata (MacDonald-Hull et al., 1991) and the genes for these cytokines are therefore candidate genes in alopecia areata. The IL-1 system consists of at least two agonist molecules, IL-1α and IL-1β, and a structurally related receptor antagonist molecule, IL-1ra. IL-1ra binds to type 1 IL-1 receptors but does not transduce a signal, and therefore acts as a potent anti-inflammatory molecule (Dinarello, 1996).

The IL1A, IL1B and IL1RN (IL-1 receptor antagonist) genes are clustered within a 430-kb interval on chromosome 2q13 (Nicklin et al., 1994). IL1RN variants are associated with the severity of several inflammatory/autoimmune diseases, including ulcerative colitis (Mansfield et al., 1994), lichen sclerosus (Clay et al., 1994), psoriasis (Tarlow et al., 1997), myasthenia gravis (Huang et al., 1998), multiple sclerosis (Schrijver et al., 1999) and rheumatoid disease (Cox et al., 1999). We previously reported an association between the rare allele of the IL1RN VNTR and alopecia universalis, the severest form of alopecia areata (Tarlow et al., 1994). This has been confirmed in a recent study in which we showed a strong association with severity of alopecia areata using further markers within the IL1RN gene as well as the new IL1RN analogue, IL1L1 (Tazi-Ahnini et al., in press). In the present study, we have tested alopecia areata for association with three different markers within the IL-1 cluster. Genotypes for the marker IL1B+3954 are known to influence production of IL-1β (Pociot et al., 1992). We chose the IL1A+4845 polymorphism because it has 100% linkage disequilibrium with IL1A-889 (Cox et al., 1998), which has also been shown to influence IL-1β production (Hulkkonen et al., 2000). There is strong linkage disequilibrium between IL1A+4845 and IL1B+3954. Weak linkage disequilibrium exists between IL1B−511 and these two markers (Cox et al., 1998).

Materials and methods

Patients and clinical assessment

One-hundred and sixty-five patients with alopecia areata were recruited from dermatology clinics in Sheffield, UK (aged 46.8 ± 15.4 years, female:male ratio 1.98). DNA from healthy controls was obtained from 1150 consecutive sample donations to the Trent Blood Transfusion Service, Sheffield; aged 43.9 ± 11.9 years, female:male ratio 1.03. Controls were ethnically matched to the disease population (Caucasian, northern European).

The alopecia areata patients entered into this project were managed by three consultant dermatologists (MJC, AMcD and AGM) and had been followed up for 1–6 years. The clinical diagnosis of alopecia areata was based on the presence of initially patchy alopecia with exclamation mark hairs and exclusion of other causes of alopecia (Messenger & Simpson, 1997). Detailed clinical information was obtained from each patient, including age at onset, family history of alopecia areata, type of disease (patchy alopecia, alopecia totalis and alopecia universalis), concomitant autoimmune diseases and atopy (Table 1). The clinical information was updated at follow-up visits.

Table 1 Clinical characteristics of the patient group
Full size table

DNA analysis

Genomic DNA was extracted from whole blood according to standard protocols and stored at 100 ng/μL. Three single-nucleotide polymorphisms (SNPs) in the IL-1 cluster IL1A (+4845), IL1B (+3954), IL1B (–511), were analysed. 25μL PCR reactions comprised 8% glycerol, 200 μM each dATP, dGTP and dCTP, 400 μM dUTP, 1.25 U AmplitaqGold (Perkin-Elmer, USA), 1.25 U Uracil-N-Glycosylase (Perkin-Elmer, USA), 5 mM MgCl2 and 500–900 nM each primer. Allelic discrimination at these loci was performed using a 5′ nuclease assay (TaqMan allelic discrimination test). This is based on 5′ nuclease activity of Taq polymerase and the detection by fluorescence-resonance energy transfer (FRET) of the cleavage of two probes designed to hybridize to either allele during PCR. Double fluorescent probes were provided by ABI-PE (Forster City, CA; Warrington, UK). Probe and primer sequences, and cycling conditions are detailed in Table 2. Probes were labelled with carboxyfluorescein (FAM) and carboxy -4,7,2′,7′-tetrachlorofluorescein (TET) fluorescent dyes at the 5′ end, and the quencher carboxytetramethylrhodamine (TAMRA) at the 3′ terminus. Concentrations of FAM and TET probes ranged between 20 and 50 nM and 50–350 nM, respectively, depending on the probes used. Plates were scanned in an LS50-B or a PE7200 fluorimeter (ABI/Perkin-Elmer).

Table 2 Probes and PCR conditions used for genotyping IL1A (+4845), IL1B (+3954), IL1B (−511)
Full size table

In keeping with standard nomenclature, we denoted the common allele ‘1’ and the rarer allele ‘2’ for each polymorphic site. Homozygosity is indicated by 1,1 or 2,2 and heterozygosity by 1,2.

Statistical analysis

Genotype distributions in cases and controls were compared using χ2-tests on 2 × 3 contingency tables. Odds ratios were calculated by combining 1,2 and 2,2 genotypes against 1,1 in 2 × 2 tables in the overall dataset and in disease subgroups according to severity.

Results

Having checked that there was no significant deviation from the Hardy–Weinberg equilibrium for any of the markers examined (IL1A+4845, IL1B+3954 and IL1B−511) in patients or controls, we examined genotypic distributions of the three markers in alopecia areata patients compared to healthy controls (Table 3). There was no significant difference in the allelic distribution of the IL1B+3954 and IL1B−511 polymorphisms between patients and controls. However, there was suggestive but nonsignificant evidence of association with IL1A+4845 (Table 3). We then divided the patients and controls into two groups by genotype (1,1 vs. 1,2/2,2). There was borderline association between the IL1A+4845 polymorphism and disease in the overall dataset [OR 1.39 (1.00, 1.93)]. Subgroup analysis by disease severity suggested that this association was due mainly to the contribution of cases of mild disease (patchy AA) [OR 1.48 (0.96, 2.29)], an apparently weaker association being noted in severe disease (alopecia totalis/universalis) [OR 1.12 (0.69, 1.82)].

Table 3 Allelic distribution of the three polymorphisms in the IL-1 cluster: IL1A (+4845), IL1B (+3954), IL1B (−511) in patients and controls. P-values are shown uncorrected for the number of tests performed
Full size table

Discussion

In the present investigation, we have shown that there is a possible association between alopecia areata and a polymorphism of IL1A (+4845). The effect of this polymorphism may be more pronounced in the milder (patchy) forms of the disease. This is the opposite pattern to our previous findings with IL1RN in which we showed that homozygosity for the rare allele was associated with a greatly increased risk of severe disease (Tarlow et al., 1994). Our current observations add to the evidence suggesting genetic heterogeneity in alopecia areata and the need for larger studies with analysis of well-defined clinical subgroups of disease is emphasized.

We found no association with either of the IL1B+3954 polymorphisms studied in the overall dataset or in disease severity subgroups. This contrasts with the work of (Galbraith et al., 1999) who reported a weak association with the same marker. However, the power of the Galbraith study was low, around 41%, whereas our study had approximately 90% power to detect such an effect. It is also possible that the contrasting results may reflect different clinical characteristics in the two patient populations. On the other hand, the main finding in that study was that IL1B co-operates with immunoglobulin κ light chain genotypes to increase susceptibility to alopecia areata. Other investigators have reported associations with polymorphisms of IL1B in a range of inflammatory disorders including inflammatory bowel disease (Stokkers et al., 1998; Nemetz et al., 1999) and insulin-dependent diabetes mellitus (Pociot et al., 1992; Loughrey et al., 1998).

Recently, in our study of the MX1 gene on chromosome 21q22.3 in alopecia areata (Tazi-Ahnini et al., 2000) we showed a significant association of this gene with patchy disease. The role of IL-1 cluster polymorphisms as well as MX1 variants and HLA genotypes in the pathogenesis of alopecia areata requires further examination in large numbers of patients and especially in family studies.