Association Alopecia Areata | Determination of aire risk haplotypes in families with alopecia areata

Determination of aire risk haplotypes in families with alopecia areata

Auteur : Dr. Rachid TAZI-AHNINI (Biomedical Genetics Division of Genomic Medicine University of Sheffield Medical School Royal Hallamshire Hospital SHEFFIELD ENGLAND).

Travail subventionné par l’AAA (subvention 2002) à hauteur de 12 000 €.

Subvention débloquée le 27 février 2003.

Compte rendu fourni le 30 octobre 2003.

COMPTE RENDU DÉTAILLE (texte intégral)

 Determination of AIRE Risk Haplotypes in Families

with Alopecia Areata


Patients with the autoimmune polyendocrinopathy-candidiasis-ectodermal dystrophy syndrome (APECED or APS1) have mutations in the autoimmune regulator (AIRE) gene resulting in autoimmune reactions towards endocrine and range of other organs. Ectodermal dystrophies affecting the dental enamel, nails and hair follicles are observed in APECED patients and alopecia areata is observed in approximately 40% of cases of APECED. The protein encoded by the AIRE gene is likely to be involved in the pathogenesis of the disorders associated with APECED including alopecia areata. We have performed systematic screening of the entire coding sequence of the AIRE gene for SNPs and identified 23 variants of which 6 gave an amino acid change in Aire protein. Subsequently, we demonstrated a disease association in alopecia areata with one AIRE variant of potential functional relevance, encoding a DNA binding domain of the Aire protein (Tazi-Ahnini et al. 2002).


In the present project, the SNPs described above were studied using both case-control and family-based methods. Association between each SNP and disease was analysed using the transmission disequilibrium test (TDT) as implemented in the software package GENEHUNTER. Linkage disequilibrium between SNPs was examined using the D’ measure.


Location of polymorphisms

We have located two definite and four possible functional polymorphisms in the coding region of the AIRE gene. Nine other potential functional polymorphisms required further investigation. However, these have been discounted as experimental artefacts and are therefore not listed in table 1 which also includes SNP’s identified by other groups.

Table 1: Positions of SNPs within the AIRE gene to date

Exon Intron Genomic Exon DNA Nucleic acid change Amino acid change Notes
1 No SNP detected
2 No SNPs detected
3 No SNPs detected
4 6333 A – G 2,5
4 6340 A – G 2,5
4 6999 A – G 2,7
5 7094 715 C – T Silent 2,4,5
5 7202 G – A 2,6
6 8385 808 C – T Silent 2,4,5
7 8723 961 G – C S278R 1,2,4,8
7 8744 982 C – T Silent 3,7
7 8859 G – A 2,5
7 8915 G – A 2,6
7 8924 T – C 2,4,5
7 9745 G – A 2,5
7 9760 G – A 3,6
8 9817 1029 T – C V301A 1,2
8 9863 1075 G – T Silent 2,6
8 9958 C – G 2,5
9 11107 G – A 3,4,6,8
9 11179 C – T 2,4,5
9 11182 C – T 2,5
9 11184 G – T 3,5
10 11794 1324 T – C Silent 2,4,5,8
11 12581 1498 C – T Silent 2,5
11 12587 1504 C – T Silent 3,6
12 No SNPs detected
13 No SNPs detected
14 16366 1705 T – C Silent 2,4,5,8
14 16642 1981 G – A Silent 2,6
14 16730 C – T 2,7
14 16731 T – G 2,7



1: Tested by us, confirmed to be a definite polymorphism.

2: Detected by us in the first analysis, when the alignments were initially done.

3: Found by us in the second analysis (July 2002), but not in the first.

4: Also reported in Scott et al. (1998).

5: Probable SNP (not an artefact).

6: Possible SNP (uncertain if this is an artefact).

7: Unlikely SNP (this is probably an artefact).

8: Also reported in Wang et al. (1998).

Analysis of polymorphisms in disease

G8723C: This has been assayed and confirmed to be a relevant functional polymorphism. The rare C allele is found at a higher frequency in patients with alopecia universalis. This polymorphism is located in the SAND domain, affecting DNA-binding and possibly nuclear compartmentalisation.

T9817C: This is a functional polymorphism, but assays have shown that the presence of either allele does not affect the chance of contracting alopecia areata. However, the mutation may affect nuclear organisation, as it is in a PHD domain.

We were unable to confirm potential functional SNPs at T5723G, C5792G, G11709T and C12601T respectively. However, we did detect and confirm an SNP in the promoter region at position -103 not far from a CAT box (from –96 to –73) and within a growth factor independence-1 (GFI-1) binding site which completely overlaps a confirmed nuclear factor-Y (NF-Y) binding site /CCAAT box (from –96 to -73). GFI-1 is a transcriptional repressor and in adults it is only expressed in the thymus, spleen and testis. If GFI-1 is bound to the promoter region of AIRE, it could possibly impede the binding of NF-Y to the CCAAT box, and thus play a role in regulating AIRE expression. The three polymorphisms with potential functional relevance at positions –103, 8723 and 9817 respectively were selected to perform case-control and family studies.

Case-control study

We typed 188, 201 and 171 cases and 178, 360 and 244 controls respectively for the –103, 8723 and 9817 polymorphisms. Chi-square tests have been performed for each of these SNPs in the overall data set and in alopecia areata clinical subtypes. Results are shown in table2.

Table 2


AIRE-103 AIRE 8723 AIRE 9817
Controls 147 CC 303 CC 261   TT
43 CT 57 CG 4 TC
1 TT 3 GG 0 CC
Patients 158 CC 156 CC 204 TT
43 TC 52 CG 8 TC
3 TT 2 GG 1 CC
c2 0.13 7.04 3.24
p NS 0.008 0.06
Early AAO (≤30) 81 CC 80 CC 106 TT
24 CT 27 CG 5 TC
1 TT 2 GG 1 CC
c2 0.01 5.56 4.51
p NS 0.018 0.03
Late AAO (>30) 69 CC 67 CC 85 TT
15 TC 22 CG 3 TC
2 TT 0 GG 0 CC
c2 054 3.22 1.22
Patchy alopecia 85 CC 85 CC 106 TT
21 CT 25 CG 4 TC
1 TT 0 GG 1 CC
c2 0.62 2.20 3.01
Alopecia totalis 29 CC 29 CC 33 TT
4 CT 6 CG 2 TC
2 TT 0 GG 1 CC
c2 0.59 0.01 6.49
p NS NS 0.01
Alopecia universalis 28 CC 28 CC 43 TT
13 CT 14 CG 2 TC
0 TT 2 GG 0 CC
c2 1.36 10.17 1.74
p NS 0.001 NS


Table 2: allelic distribution of AIRE polymorphisms at position –103, 8723 and 9817 in cases and controls. Patients have been divided according to severity of disease (patchy alopecia areata, alopecia totalis and alopecia universalis) and age at onset of the disease (up to 30 and more than 30 years old). Results of Chi-square tests are also presented.

We have confirmed the association between the AIRE G8723C and alopecia areata as published. The association is stronger with alopecia universalis (2 = 10.17, p=0.001). This polymorphism is also significantly associated with early age at onset of the disease (p=0.018). There is also slight increase of individual heterozygotes for the AIRE T-103C in most severe forms of alopecia areata (totalis and universalis) but this is not statistically significant. In the overall dataset, association between the AIRE 9817 and alopecia areata is not statistically significant (p=0.06). However, the association became significant when we divided patients according to severity or onset of disease. There was a significant association with alopecia totalis and early age at onset (p= 0.01 & 0.03 respectively) with this polymorphism, but this became non-significant when p-values were corrected for multiple testing.

We then calculated risk haplotypes for AIRE polymorphisms –103, 8723 and 9817 using the Estimate Haplotype programme (Zhao et al. 2000); 198 cases and 168 control samples were typed for the three polymorphisms. We found no linkage disequilibrium (LD) between the three markers in this analysis, suggesting that they segregate independently. A comparison of the distribution of AIRE haplotypes in cases and controls revealed that there was no significant difference in the allelic distribution between them. This is not surprising because the –103 and 9817 polymorphisms are not in LD with the AIRE 8723 variant.

We also genotyped nuclear families for these three polymorphisms. There were ninety seven TDT and S-TDT units which have been typed for the AIRE –103, 8723 and 9817 polymorphisms. Taking genotype data for the AIRE –103 and AIRE 8723 polymorphisms, we used the S-TDT program (Spielman and Ewens, 1998) which calculates the standard TDT, the S-TDT and also a combined TDT. Due to the lack of heterozygosity amongst many of the parents, only a small number of families (n=39) could actually be used in the analysis. The AIRE 9817 polymorphism was not included due to a lack of heterozygosity in this set. Z scores for both TDT, S-TDT or a combined test were very low and therefore the p-values non significant. We also used TRANSMIT, version 2.5, a program for estimating genetic associations from probabilities of haplotype transmission to affected offspring when there may be uncertain marker haplotype assignment (Clayton and Jones, 1999). The number of families with transmissions to affected offspring was 52 and a global chi-squared test, on 2 degrees of freedom was 2.03 giving p-value >0.05.


We have confirmed the association of the AIRE 8723 polymorphism and alopecia universalis in larger case-control groups than those reported in our previous publication. There was a significant association between the AIRE 9817 and alopecia totalis and early age at onset (p= 0.01 & 0.03 respectively) which become non-significant when p-values were corrected for multiple testing. There was no apparent LD between the AIRE –103, 8723 and 9817 polymorphisms and these seem to segregate independently. Family analysis did not show any association with these polymorphisms but we cannot definitely say that the results were negative because we did not have a sufficient number of informative families with heterozygous parents. Weak heterozygosity for AIRE –103 and 9817 polymorphisms means that much larger family collections would be required to perform a conclusive analysis of these polymorphisms and we aim to undertake this using the US alopecia areata registry collection in the future. In the meantime, we are extending our study by typing other polymorphisms within the AIRE gene including synonymous and intronic substitutions and deriving estimate haplotypes in patient and control groups. This strategy will allow us to increase the number of hetorozygote parents and therefore have more power to detect the difference, if any, between transmitted and non-transmitted haplotypes.

Future Work

Confirmation of the 4 remaining potential functional polymorphisms has to be completed, and assays performed for those that are present.

Screening of the promoter region will be completed. Functional studies will be performed on any relevant polymorphism found in the promoter region.

When the assays have been completed for all relevant polymorphisms, the results can be used to produce a disease haplotype for the AIRE gene. Transmission disequilibrium/tests would also be performed at this stage.

References :

1) Clayton D, Jones H. Transmission/disequilibrium tests for extended marker haplotypes. Am J Hum Genet 1999 ; 65 : 1161-1169.

2) Scott HS, Heino M, Peterson P, Mittaz L, Lalioti MD, Betterle C, Cohen A, Seri M, Lerone M, Romeo G, Collin P, Salo M, Metcalfe R, Weetman A, Papasavvas MP, Rossier C, Nagamine K, Kudoh J, Shimizu N, Krohn KJ, Antonarakis SE. Common mutations in autoimmune polyendocrinopathy-candidiasis-ectodermal dystrophy patients of different origins. Mol Endocrinol. 1998 ; 12 : 1112-9.

3) Spielman RS & Ewens WJ. A sibship test for linkage in the presence of association : the sib transmission/disequilibrium test. 1998 ; 62 : 450-458.

4) Wang CY, Davoodi-Semiromi A, Huang W, Connor E, Shi JD, She JX. Characterization of mutations in patients with autoimmune polyglandular syndrome type 1 (APS1). Hum Genet. 1998 ; 103 : 681-5.

5) Zhao JH, Cutis D, Sham PC. Model-free analysis and permutation test for allelic associations. Human Heredity. 2000 ; 50 : 133-139.

Publications 2002-03

1) R Tazi-Ahnini, MJ Cork, DJ Gawkrodger, MP Birch, D Wengraf, AJG McDonagh, AG Messenger. Role of the Autoimmune Regulator (AIRE) Gene in Alopecia Areata: Association of Potentially Functional AIRE Polymorphism with Alopecia Universalis (2002). Tissue Antigens, 60 : 489-495.

2) Rachid Tazi-Ahnini, Janine M Timms, Angela Cox, & Anthony G Wilson. Identification of novel single nucleotide polymorphisms within the NOTCH4 gene and determination of their association with MHC alleles (2003). Eur J Immonogenet 30, 101-105.

3) Tazi-Ahnini R, Cork MJ, Wengraf D, Wilson AG, Gawkrodger DJ, Birch MP, Messenger AG, McDonagh AJ. Notch4, a non-HLA gene in the MHC is strongly associated with the most severe form of alopecia areata. 2003 Hum Genet. 112(4) : 400-403.

4) McDonagh AJ & R Tazi-Ahnini. Alopecia areata-epidemiology and genetics. Clinical and Experimental Dermatology. 2002. 27(5):405-9 (review).

5) McDonagh AJG, Tazi-Ahnini R, Messenger AG. Alopecia areata: an update on etiology and pathology. Skin, Hair & Nails Structure & Function. Bo Forslind & Magnus Lindberg, Eds, Marcel Dekker, New York 2004 ISBN 0 8247 4313 X.

Rachid Tazi-Ahnini & Andrew. J. McDonagh

Division of Genomic Medicine

University of Sheffield-Medical School

Beech Hill Road


S10 2RX

Tel/Fax: (+44) (0) 114 271 2933

Tel: (+44) (0) 114 271 3785