David L. Duffy1, Sue C. Healey1, Georgia Chenevix-Trench1, Lon Cardon2, Jay Lichter2, Tim Harris2, Nick G. Martin1.
1 Genetic Epidemiology Laboratory, Queensland Institute of Medical Research, Brisbane, Australia
2 Sequana Therapeutics, La Jolla, California
Shirakawa et al  have recently reported a strong association between atopy and mutations in the sixth exon of the FcERI beta-subunit gene (FCER1B), notably I181L. This finding has been replicated in a population sample from Busselton, Western Australia [Hill et al 1995], where the frequency of Leu181 was 2.7% (28 carriers in 1020 individuals). We report here an Australia wide study of asthma and allergy in twins and their families who have been typed for this and a close (fifth intron of the FCER1B gene) short repeat polymorphism, FCER1Bca.
The Australian National Health and Medical Research Council Twin Registry (ATR) is a volunteer twin registry, whose members have been recruited through community organisations, schools and media appeals from 1979 to date. We surveyed a sample of 1961 twin pairs where one or both members previously reported a history of asthma and wheezing on questionnaires sent (1980, 1988, 1989) to a 15772 adults on the ATR. In 1992, 865 twins from this group (419 complete pairs) resident in seven cities around Australia, underwent a histamine inhalation challenge [Yan 1984] (a measure of bronchial responsiveness, increased in asthma), epicutaneous skin testing for eleven common aeroallergens (wheal diameter being measured after ten minutes as the mean of the maximum diameter and its perpendicular), and measurement of total serum Immunoglobulin E (IMx Total IgE assay, Abbott Laboratories, USA). Therefore the sample consists of 808 twins and their parents, as well as 131 unrelated controls recruited from our laboratory staff, spouses and blood donors.
Atopy was defined in terms similar to those previously published [Cookson et al 1989] as any skin test wheal 3 mm or greater in diameter than negative control or a serum IgE over 100 IU/ml. Serum IgE level was log transformed to approximate normality.
DNA was extracted from peripheral blood leucocytes [Miller et al 1988] collected from 610 of these twins, and from 198 of their parents. The 215 DZ twin pairs and their available parents were typed at the two short tandem repeat polymorphisms within the FCER1B as originally described by Sandford et al3 (FCER1B-ca, CCl11-319ca). The entire sample was typed for the Leu181 mutation via an allele-specific oligonucleotide polymerase chain reaction based assay modified from Ref. 1. Amplification was carried out using: 5FU, 5'-TGTATGTGTCACTTTAAAAGGACTGGTCAG-3'; and 3FU, 5'-TAACATATCAGTCCTATTATCCCAACCCTC-3' producing a 459 bp control product. The mutation (Leu181) was detected using 3M, 5'-AATGGTGAGAAACAGCATCATCATTACCAA-3', giving a 163 bp product; the wild type by 5WK, 5'- TTGTCATTTGTTGCTGTTCAATAGGAAGTT-3', giving a 353 bp product. A 20 ul reaction volume contained approximately 125 ng of the template DNA, 25 ng of the primer 3M, 125 ng of primers 3FU, 5FU and 5WK, 1.5 mM MgCl2, 10 mM Tris-HCl (pH 8.3), 50 mM KCl, 200 æM dNTPs, and 1 U Taq DNA polymerase. PCR conditions (96 well plate on a Hybaid PCR machine) consisted of 94øC for 3 minutes, 30 cycles of 45 s at 94øC, 45 s at 55øC, and 45 s at 72øC, followed by an elongation step of 7 minutes at 72øC. Visualisation was on a 3% agarose gel (Seakem ME) stained by ethidium bromide.
This was performed at Sequana Therapeutics. A region of the FcERI beta-subunit gene spanning part of the fifth intron and sixth exon and containing the codons for amino acids 181 and 183 was amplified by PCR and sequenced by automated, fluorescent cycle sequencing. Amplification was carried out using: Asm F5, 5'-GATGGTAGGGAGATGAAAACAGGAG-3'; and Asm B6, 5'-CAGCTCCACAGATTGTGAGTGACAC-3' producing a 177 bp product. A 20 ml reaction volume contained 100 ng of the template DNA, 4 pmol of each of the Asm B6 and Asm F5 primers, 10mM Tris-HCl, pH 8.3 (at room temperature), 2.5 mM MgCl2, 50 mM KCl, 200 mM dNTPs, and 0.5 U Taq DNA polymerase. PCR was performed in a PE 9600 with the following conditions: 95 C for 1 minute, 25 cycles of 15 seconds at 95 C, 30 seconds at 53 C, and 30 seconds at 72 C, followed by an elongation step of 2 minutes at 72 C. PCR products (5ml of each reaction) were run on an 1.5% agarose gel for quantification and to observe the quality of the amplification. Products were then ultrafiltered with Microcon 100's. Cycle sequencing reactions were performed according to the protocol in Applied Biosystem's "Taq Dye DeoxyTM Terminator Cycle Sequencing Kit" on a PE 9600 using both the Asm F5 and Asm B6 primers. Sequence analysis was completed with Applied Biosystem's Factura and Autoassembler programs.
Sib-pair linkage analysis was performed using the SIBPAL program included in the SAGE package for both discrete (atopy) and continuous (log sIgE level) phenotypes. Allelic association was examined using a generalised form of the transmission-disequilibrium test, implemented within SAS PROC CATMOD and a Fortran program Sib-pair (DLD).
This analysis is for a total of 219 DZ twin pairs typed at FCER1B-ca, of whom 208 pairs were typed at CCl11-319-ca. The range and frequency of alleles for FCER1B-ca in this sample was very close to that reported from the British population (Table 1).
We first examined linkage for atopy. This found no evidence for linkage (Table 2). The results of the regressive test was: overall mean estimated ibd score for FCER1B-ca was pi=0.5006, t=0.54 (df=217, P=0.70); CCl11-319-ca pi=0.5277, t=0.47 (df=206, P=0.68), with similar results as a multiple regression. The analyses for continuous traits also gave negative results. For log sIgE: FCER1B-ca mean ibd score pi=0.5024, t=- 1.1272 (df=193, P=0.13); CCl11-319-ca, mean ibd score pi=0.5338, t=0.0605 (df=184, P=0.52). For inverse cube-root transformed histamine dose-response slope, FCER1B-ca pi=0.5033, t=0.4397 (df=201, P=0.67); CCl11-319-ca pi=0.5307, t=-1.0074 (df=191, P=0.16).
We then tested for an allelic association for the two markers. The simple analysis we have performed was to examine genotype distributions in non-skin-atopic versus skin-atopic twins, selecting one twin from each pair - randomly where concordant, but a nonatopic twin in preference to an atopic twin otherwise. For this analysis there were 211 subjects. One test was for homogeneity of genotype distributions. For FCER1B-ca, the contingency Pearson X2(df=17)=12.47 (P=0.77), and LR X2=15.8 (P=0.54); for CCl11-319-ca Pearson X2(df=24)=36.8 (P=0.05), LR X2=44.8 (P=0.01). However, an examination of CCl11-319-ca for differences in gene frequencies (under Hardy-Weinberg constraints), excluding alleles 181, 185 and 203 gave a LR X2(df=7)=10.8 (P=0.15), and Pearson X2=8.1 (P=0.32). The findings then seem to hinge on the lack of Hardy-Weinberg equilibrium seen for this marker. The TDT did not find any evidence for association (Table 3), and the negative affected sib-pair linkage results for atopy also exclude strong allelic association [Hodge 1993]. Negative results also were obtained for ANOVA of log sIgE and inverse cube-root transformed DRS either under an additive allelic model or genotypic models. We conclude there is no evidence for allelic association.
|Moffat [pers comm]||.02||.00||.42||.01||.28||.23||.01||.00||.03||.00|
|Marker||Number of twins affected||Number of DZ pairs||Mean estimated IBD score||SD||SE||T-value||P-value|
|Parental Genotype||Proportion where first allele transmitted||Contribution to global TDT|
|First allele||Second allele|
|Histamine Challenge DRS > 2.56 l/umol|
|D pter SPT wheal > 3.5 mm|
|Total sIgE > 100 IU/ml|
These results were consistent with the finding that no Leu181 mutations were present in the total 939 subjects tested (MZ and DZ twins, available parents of DZ twins and unrelated controls). Each run of the AS-PCR assay for Leu181 was accompanied by a successful positive control run, and the wild type allele (Ile181) was detected in all cases.
As these DNAs have amplified at many other STR loci successfully, the only artefactual explanation for this finding considered was the presence of interfering factors introduced during processing that might specifically affect the Leu181 primer's binding. However, dilutions of up to 1:99 of the known positive control in pooled DNA from the twins did not interfere at all with detection of the (control) mutant band present, making any contaminant hypothesis untenable. In addition, we did not detect the Leu181 or Leu183 mutations using either the original ARMS assay, or direct sequencing, in the 19 subjects with the highest summed skin prick test wheal diameters (geometric mean sIgE of 321.5 IU/ml) or their available parents (N=20; 12 mothers, 8 fathers). Four of this group had a maternal but no paternal history of symptomatic atopy. These individuals would have a high probability of carrying a mutant allele according to the earlier work.
There have been numerous negative linkage studies of atopy and the 11q13 region since the original report in 1989. This failure to replicate has variously been ascribed to (1) the presence of maternal inheritance of atopy (possibly mediated by imprinting); (2) use of different disease definitions; and/or (3) genetic heterogeneity. The finding of a mutation in the FCER1B receptor gene associated with atopy when transmitted from the mother confirmed the first mechanism listed. The report that the mutation frequency was much lower than that needed to explain a 30% population prevalence of atopy supported the third possibility, given that disequilibrium between the Leu181 mutation and the disease allele is large. The recent publication of linkage of IgE level to the IL-4 region, is further evidence that the atopic phenotype is under polygenic control.
Our finding of a high heritability of atopy in the twins tested despite the absence of any Leu181 mutations also supports the hypothesis of genetic heterogeneity. It seems most likely from these results that the Leu181 mutation is far less common than 2.7% in most of the Australian population, including atopic asthmatics. We can only speculate why the mutations found in the Busselton sample were as frequent as they were - the most likely suggestion that migration from Oxfordshire and neighbouring regions is higher to West Australia than it is to the other Australian states. We feel confident that we have excluded experimental artefact as a reason for the strong discordance between our results and those of Hill et al and Shirakawa et al.
We wish to thank Teresa Barrington who assisted with the field phase of the study. The study was funded by the Asthma Foundation of Queensland and the Australian National Health and Medical Research Council. D.L.D. was a Public Health Research and Development Committee Postgraduate Scholar. We also thank all the twins who took part in the study. The SAGE package is supported by U.S. Public Health Service Resource grant 1 P41 RR03655 from the division of Research Resources. Finally, we are most grateful to Drs W.O.C.M. Cookson, M. Hill, and coworkers from the Asthma Genetics Group at John Radcliffe Hospital, Oxford, for making positive control DNA samples available.