Association analysis of the monoamine oxidase A and B genes with attention deficit hyperactivity disorder (ADHD) in an Irish sample: Preferential transmission of the MAO-A 941G allele to affected children
Abstract
Pharmacological and genetic studies suggest the importance of the dopaminergic, serotonergic, and noradrenergic systems in the pathogenesis of attention deficit hyperactivity disorder (ADHD). Monoamine oxidases A and B (MAO-A and MAO-B) degrade biogenic amines such as dopamine and serotonin and thereby control the levels of these neurotransmitters in the central nervous system. We examined four polymorphisms in the MAO-A gene (30 bp promoter VNTR, CA microsatellite in intron 2, 941G/T SNP in exon 8, and A/G SNP in intron 12) as well as two markers in the MAO-B gene (CA microsatellite in intron 2 and T/C SNP in intron 13) for association with ADHD in an Irish sample of 179 nuclear families. TDT analysis of the examined MAO-A markers revealed a significant association of the more active MAO-A 941G allele with the disorder (χ2 = 5.1, P = 0.03, OR = 1.7). In addition, haplotype analysis revealed a significantly increased transmission of a haplotype consisting of the shorter allele of the promoter VNTR (allele 1), the 6-repeat allele of the CA microsatellite and the G-allele of the 941G/T SNP (famhap global statistic 34.54, P = 0.01) to ADHD cases. No significant distortion in the number of transmitted alleles was observed between the two examined MAO-B polymorphisms and ADHD. These findings suggest the importance of the 941G/T MAO-A polymorphism in the development of ADHD at least in the Irish population. © 2005 Wiley-Liss, Inc.
INTRODUCTION
Attention deficit hyperactivity disorder (ADHD) is one of the most common behavioral disorders affecting 2%–6% of school-aged children worldwide [Tannock, 1998]. It is characterized by inattention, hyperactive, and impulsive behavior leading to difficulty organizing tasks, excessive motor activity and risk taking behavior. Although the aetiology of ADHD is not fully understood, a strong genetic component in the pathogenesis of the disease with an estimated heritability of 70%–90% has been reported [Levy et al., 1997].
Several lines of evidence have implicated the dopaminergic, noradrenergic, and serotonergic systems in the development of ADHD. Monoamine oxidase A (MAO-A) is an enzyme that degrades biogenic amines such as dopamine, noradrenaline, adrenaline, and serotonin by oxidative deamination and thereby plays a key role in the modification of signal transduction in these neurotransmitter systems [Shih and Thompson, 1999]. In addition, MAO-A inhibitors such as tranylcypromine have been shown to be effective in the pharmacological treatment of ADHD [Zametkin et al., 1985]. Furthermore, MAO-A has also been reported to be involved in the pathogenesis of the intermediate phenotypes “impulsivity” and “aggression”. Brunner et al. [1993] described a rare point mutation in the MAO-A gene associated with loss of function, which resulted in a highly impulsive and aggressive behavioural phenotype. Moreover, MAO-A knockout mice have been observed to exhibit a significantly increased aggressive behavior accompanied by elevated levels of serotonin, noradrenaline, and dopamine [Cases et al., 1995]. Finally, maltreated children with a genotype conferring high levels of MAO-A expression were reported to be less likely to develop antisocial problems [Caspi et al., 2002].
In children with ADHD, significantly lower levels of platelet MAO-A activity associated with increased impulsivity and inattention were reported [Shekim et al., 1986]. These findings indicate that MAO-A is a good candidate gene for ADHD and that DNA variations in this gene may play a role in the predisposition to the disorder. The gene maps to chromosome Xp11.4-p11.3. A functionally relevant 30 bp VNTR has been identified in the promoter region of the MAO-A gene [Sabol et al., 1998]. The longer alleles (3a, 4, and 5) have been shown to affect the transcription of the gene three to four times more efficiently than the shorter allele 3 [Deckert et al., 1999]. Significant association between the longer alleles of the VNTR and ADHD has been observed [Manor et al., 2002], but could not be replicated in an independent study [Lawson et al., 2003]. The short allele has been reported to be associated with impulsivity and aggression [Manuck et al., 2000], which was confirmed in a subgroup of patients with ADHD and broadly defined concurrent conduct disorder [Lawson et al., 2003]. Another functional variant, a silent 941G/T polymorphism in exon 8 has been reported to be associated with low (941T) and high (941G) MAO-A activity, respectively [Hotamisligil and Breakefield, 1991]. However, a recent study did not find any association of this polymorphism with either ADHD in general or a subgroup of patients with ADHD and broadly defined concurrent conduct disorder [Lawson et al., 2003]. Finally, a CA-repeat microsatellite located in intron 2 [Black et al., 1991] was reported to be associated with ADHD in a Chinese sample [Jiang et al., 2001]. This has been supported by similar findings describing a trend towards preferential transmission of the 122 bp allele of the CA(n) microsatellite, which, however, is not the same allele that Jiang et al. reported to be linked with the disease [Payton et al., 2001].
Monoamine oxidase B (MAO-B) catalyzes the oxidative deamination of neurotransmitters such as dopamine, phenylethylamine, and benzylamine. Additionally, pharmacological studies provide preliminary evidence for a beneficial effect of monoamine oxidase B inhibitors such as selegiline or deprenyl in the treatment of ADHD [Feigin et al., 1996; Akhondzadeh et al., 2003]. Administration of deprenyl has also been shown to significantly reduce impulsiveness in an animal model of ADHD [Boix et al., 1998]. Thus, the MAO-B gene was regarded as a good candidate gene for ADHD. The gene coding for MAO-B maps to chromosome Xp11.23 in close proximity to the MAO-A gene. The MAO-A and MAO-B genes are functionally related to each other by exhibiting an identical exon-intron organization as well as a high sequence similarity [Grimsby et al., 1991]. A CA-repeat polymorphism in intron 2 of the MAO-B gene (rs3838196) has been analyzed in a Chinese sample with ADHD, but was not found to be associated with the disorder [Jiang et al., 2001]. Additionally, a single nucleotide T/C polymorphism (rs1799836) in intron 13 has been reported as a possible risk factor for Parkinson disease [Costa et al., 1997], but to date has not yet been examined in relation to ADHD.
In the present study, we attempted to assess the importance of the above-mentioned three variants at the MAO-A locus in a sample of 179 Irish ADHD nuclear families. Furthermore, we also examined an additional MAO-A marker in an effort to conduct linkage disequilibrium and haplotype analysis of the studied markers. We also examined two markers at the MAO-B locus for possible association/linkage with ADHD.
MATERIALS AND METHODS
Subjects
The Irish clinical ADHD sample comprising 179 nuclear families was recruited from child psychiatric clinics and schools in West County Dublin and from the Hyperactive and Attention Deficit Children's Support Group of Ireland. Further details on the sample used in the present study can be found in Kirley et al. [2004].
Genotyping
DNA was extracted from whole blood using standard phenol chloroform procedure or from buccal cells as described in Daly et al. [1999]. Fragments containing the respective polymorphisms were amplified by PCR on a PTC-225 Peltier Thermocycler (MJ Research, Dublin, Ireland) and subsequently genotyped using the conditions given in Table I. The two microsatellite markers (MAO-A CA(n) and MAO-B CA(n) (rs3838196)) were genotyped using the semi automated florescent method on an ABI 3100 DNA sequencer (ABI Applied Biosystems, Warrington, UK).
| Primer (5′–3′) | Annealing temperature (°C) | Restriction enzyme | Reference | |
|---|---|---|---|---|
| MAO-A | ||||
| 30 bp VNTR | F: CCCAGGCTGCTCCAGAAAC | 56 | — |
Deckert et al. [1999] |
| R: GGACCTGGGCAGTTGTGC | ||||
| CA(n) repeat | F: AGAGACTAGACAAGTTGCAC | 56 | — |
Black et al. [1991] |
| R: CACTATCTTGTTAGCTCACT | ||||
| 94IG/T | F: GACCTTGACTGCCAAGAT | 54 | Fnu4HI |
Hotamisligil and Breakefield [1991] |
| R: CTTCTCCTTCCAGAAGGCC | ||||
| A/G (rs979605) | F: GCTGCTACACGGCCTACTTC | 64 | TspRI | — |
| R: AGAAATGGGGATTTTGACAAC | ||||
| MAO-B | ||||
| T/C (rs1799836) | F(T): CACTGGCAAATAGCAAAAGT | 64 | — |
Costa et al. [1997] |
| F(C): CACTGGCAAATAGCAAAAGC | ||||
| R: GGATTTACTTTGCAGGCACC | ||||
| CA(n) repeat (rs3838196) | F: GTGTTGGTGTGAAGGAAGCA | 64 | — | — |
| R: GATTGAGTAAGAGGGAAATGG |
Statistical Analysis
For single marker analysis, we used the transmission disequilibrium test (TDT) [Spielman et al., 1993]. The McNemar χ2 test was used to assess the significance level. For the analysis of multi-allelic markers, we applied the ETDT (TDTPHASE v2.40) [Dudbridge, 2003]. Since MAO-A and MAO-B are x-linked markers, genotype information from fathers was only used in the analysis of female patients with ADHD. Linkage disequilibrium (LD) between the markers expressed as D′ was assessed by means of the GOLD program (www.well.ox.ac.uk/asthma/gold). Only parents' (in female probands) and mothers' (in male probands) genotypes were used to assess LD. Hardy–Weinberg equilibrium was examined using the online site (http://kursus.kvl.dk/shares/vetgen/_Popgen/genetik/applets/kitest.htm). Haplotype analysis was carried out using the program FAMHAP (version 14) [Becker and Knapp, 2004]. In this program, an expectation maximization (EM) algorithm is implemented to estimate transmitted and nontransmitted haplotype frequencies. The test statistic is computed from the table of transmitted and nontransmitted haplotypes. FAMHAP is unique in taking the smallest P-value found among the combinations as test statistic. The transmission/nontransmission status is then permuted to obtain the distribution of the test statistic. The empirical P-value is the fraction of permutation replicates resulting in a test statistic greater than or equal to the test statistics of the real data. The program was run on genotypes from 170 parent child trios, where genotype information for all four MAO-A SNPs was obtainable. For families with a male proband, transmitted and nontransmitted haplotypes of the mother only were used. Statistical analyses were not corrected for multiple testing.
RESULTS
The distribution of genotypes (from parents with no history of ADHD) for all examined markers did not significantly differ from those expected according to the Hardy–Weinberg equilibrium.
MAO-A
TDT analysis conducted on all examined markers of the MAO-A gene (Table II) showed significantly higher transmission of the more active 941G allele (allele 2) to ADHD cases as compared to the lower active 941T allele (allele 1) (transmitted: 45, not transmitted: 26; χ2 = 5.1, P = 0.03, OR = 1.7). Additionally, stratifying the sample for parental history of ADHD, an increased transmission of the more active 941G allele from parents with a positive history for ADHD was observed (transmitted: 26, not transmitted: 13; χ2 = 4.3, P = 0.06). None of the remaining examined markers showed any significant distortion in the transmission to ADHD cases. LD measured as D′ among the MAO-A markers was significant between all examined markers (Table III) with D′ ranging between 0.63 and 0.85. Haplotype analysis using a two-marker window revealed a significantly increased transmission of a haplotype containing the shorter 3 allele of the 30 bp VNTR (allele 1) and allele 6 of the CA-repeat (transmitted: 24, not transmitted: 11; famhap global statistic: 24.86, P = 0.05). This association was further enhanced when a three-marker window (including the 941G/T) was analyzed. A haplotype comprising allele 1 of the 30 bp VNTR, allele 6 of the CA-repeat and the 941G allele (allele 2) was preferentially transmitted to ADHD cases (transmitted: 14, not transmitted: 5; famhap global statistic: 34.54, P = 0.01).
| MAO-A | MAO-B | ||||||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Marker allelesa | T | NT | Marker allelesb | T | NT | Marker alleles | T | NT | Marker alleles | T | NT | Marker alleles | T | NT | Marker allelesc | T | NT |
| 30 bp VNTR | CA(n) | 941G/T | A/G(rs979605) | T/C (rs1799836) | CA(n) (rs3838196) | ||||||||||||
| 1(3) | 39 | 48 | 1 | 7 | 9 | 1 (T) | 26 | 45 | 1 (A) | 35 | 31 | 1 (T) | 47 | 47 | 1 | 0 | 0 |
| 2(3a,4,5) | 48 | 39 | 2 | 2 | 0 | 2 (G) | 45 | 26 | 2 (G) | 31 | 35 | 2 (C) | 47 | 47 | 2 | 5 | 10 |
| 3 | 27 | 20 | 3 | 37 | 27 | ||||||||||||
| 4 | 2 | 9 | 4 | 23 | 23 | ||||||||||||
| 5 | 4 | 4 | 5 | 25 | 34 | ||||||||||||
| 6 | 15 | 18 | 6 | 17 | 17 | ||||||||||||
| 7 | 42 | 34 | 7 | 8 | 5 | ||||||||||||
| 8 | 6 | 11 | 8 | 1 | 0 | ||||||||||||
| 9 | 0 | 0 | |||||||||||||||
| 10 | 1 | 1 | |||||||||||||||
| 11 | 0 | 0 | |||||||||||||||
| 12 | 0 | 0 | |||||||||||||||
| χ2 = 0.9, df = 1, P = 0.39 | LRS = 10.8, df = 7, P = 0.15 | χ2 = 5.1, df = 1, P = 0.03* | χ2 = 0.2, df = 1, P = 0.71 | χ2 = 0, df = 1, P = 1.0 | LRS = 5.8, df = 7, P = 0.56 | ||||||||||||
- a As in previous studies [Manor et al., 2002; Lawson et al., 2003], due to their functional roles [Deckert et al., 1999] the alleles of the MAO-A 30 bp VNTR were grouped into two classes (3 allele vs. all longer alleles (3a, 4, and 5) and named allele 1 and 2, respectively. Alleles 3 and 4, however, in concordance with other studies constituted over 95% of all observed alleles.
- b Eight alleles were detected, which were name according to fragment length from 1 (longest) through 8 (shortest).
- c Twelve alleles were detected, which were named according to fragment length from 1 (longest) through 12 (shortest). Only eight of these alleles, however, were found to be transmitted from heterozygous mothers.
- * Significant P-value at a significance level of 0.05.
| Marker 1 | Marker 2 | χ2 | P-value | D′ |
|---|---|---|---|---|
| 30 bp VNTR | CA(n) repeat | 145.2 | 0.0000 | 0.746 |
| 30 bp VNTR | 941G/T | 70.7 | 0.0000 | 0.630 |
| 30 bp VNTR | A/G (rs979605) | 88.9 | 0.0000 | 0.709 |
| CA(n) repeat | 941G/T | 105.5 | 0.0000 | 0.690 |
| CA(n) repeat | A/G (rs979605) | 127.3 | 0.0000 | 0.758 |
| 941G/T | A/G (rs979605) | 153.1 | 0.0000 | 0.853 |
MAO-B
No difference in the transmission of T or C alleles of the MAO-B rs1799836 marker to affected children was observed (Table II). ETDT analysis performed on the CA(n) (rs3838196) marker showed a slight, but insignificant increase in the transmission of allele 3 (Table II). D′ analysis showed no evidence for LD between the two markers (D′ = 0.27, P = 0.2).
DISCUSSION
In the present study, we conducted TDT analysis on four markers in the MAO-A gene, three of which have been previously analyzed for association with ADHD in several different studies [Jiang et al., 2001; Payton et al., 2001; Manor et al., 2002; Lawson et al., 2003]. The examined markers are spaced about 20 kb along the MAO-A gene. Significant linkage/association between ADHD and the more active MAO-A 941G allele was observed, as it was the case for a haplotype containing the 941G allele.
Evidence arising from functional imaging, genetic and biochemical studies point towards a hypodopaminergic state in the pathogenesis of ADHD [Solanto, 2002]. The increased frequency of the more active MAO-A 941G allele in ADHD can be explained on the basis that elevated MAO-A activity would result in a higher turnover and thereby decreased levels of dopamine, which possibly contributes to the dopamine deficit proposed for ADHD. Besides degrading dopamine, MAO-A is additionally involved in the metabolic degradation of serotonin (5-HT), which has also been implicated in the pathogenesis of ADHD. Peripheral measures of blood serotonin as well as central 5-HT function have been reported to be reduced in children with ADHD [Kruesi et al., 1990; Spivak et al., 1999]. Also, there is some evidence that serotonin agonists such as fluoxetine might be effective in the pharmacological treatment of ADHD [Gainetdinov et al., 1999]. A higher activity variant of MAO-A generating a higher serotonin turnover would be likely to exacerbate serotonin deficiency in ADHD and thereby increase the susceptibility to the disorder. Furthermore, although the MAO-A 941G/T polymorphism is functional, it is possible that the observed association results from another functional variant within the coding region or the boundaries of the MAO-A gene. However, the present study failed to confirm previous studies reporting positive linkage/association results of ADHD with the longer, higher activity alleles of the MAO-A 30 bp VNTR [Manor et al., 2002] and the MAO-A CA(n) microsatellite [Jiang et al., 2001; Payton et al., 2001]. An explanation for these diverging findings could be population-specific genetic heterogeneity as reflected by a Chinese study reporting strong evidence for association of ADHD with the DXS7 marker closely linked to the MAO-A gene [Jiang et al., 2000], which, however, could not be confirmed in an Irish study [Lowe et al., 2001]. Additionally, applying a TDT design in the investigation of x-chromosomal markers such as MAO-A and MAO-B restricts the use of informative transmissions to heterozygous mothers only. This decreases the statistical power and thereby increases the risk of false negative results. In this case, as suggested by Manor et al. [2002] a case-control-based approach might be of additional value.
With regard to the MAO-B gene, the present study confirms previous reports of no linkage/association of the MAO-B CA(n) microsatellite (rs3838196) in a Chinese sample of patients with ADHD [Jiang et al., 2001]. Additionally, we did not observe association of MAO-B rs1799836 with the disorder suggesting no major influence of the examined MAO-B markers on the pathogenesis of ADHD at least in the Irish population.
In conclusion, the findings of the present study indicate that the higher activity variant of the monoamine oxidase A (941G) might be a risk factor in the development of ADHD. This is in keeping with the hypodopaminergic and/or hyposerotonergic hypothesis of the pathogenesis of the disorder. In the future, it will be of interest to investigate the role of this marker by using intermediate- or endophenotypes such as electrophysiological markers, functional neuroimaging or performance on neuropsychological tasks.
Acknowledgements
We would also like to thank the families that participated in the study.
