Attention deficit hyperactivity disorder (ADHD) is a common childhood-onset neurodevelopmental disorder. Evidence from twin, adoption, and family studies provide support for a genetic contribution to the etiology of ADHD. Several candidate gene studies have identified an association between a 7-repeat variant in exon 3 of the dopamine 4 receptor gene (DRD4) and ADHD. However, in spite of the positive reports finding association of the exon 3 VNTR with ADHD, several other polymorphisms within DRD4 have been identified that conceivably could contribute to risk for ADHD. Recently, another common polymorphism of the DRD4 gene has been described involving a 120-bp repeat element upstream of the 5′ transcription initiation site. In this report, we describe results of analysis of the DRD4 120-bp repeat promoter polymorphism in a sample of 371 children with ADHD and their parents, using the transmission disequilibrium test (TDT). Results showed a significant preferential transmission of the 240-bp (long) allele with ADHD. Exploratory analyses of the Inattentive phenotypic subtype of ADHD strengthened the evidence for linkage. These data add further support for the role of DRD4 variants conferring increased risk for ADHD, and imply that additional studies of DRD4 and other related genes are needed.
A genetic contribution to the etiology of attention deficit hyperactivity disorder (ADHD) has received strong support from a variety of twin, adoption, and family studies.1, 2, 3 Both the syndrome of ADHD4 and the separate dimensions of inattention and hyperactivity appear to be under genetic regulation.5, 6 However, the mode of inheritance of ADHD appears to be complex and non-mendelian. Because of the robust effects on the dopamine system exerted by the psychostimulants,7 the most common form of treatment for the disorder, much interest has been focused on gene variants in dopamine-related genes using association study methods. Candidate gene studies have a role in the study of complex neurobehavioral disorders,8, 9 and initial studies of ADHD examining candidate genes from the dopamine system have yielded several positive findings.
One such candidate gene receiving a high degree of interest in relation to ADHD is the DRD4 gene. DRD4 is a member of the DRD2 receptor family and is highly expressed in frontal cortex, a region known to subserve executive functions, including aspects of working memory and attentional processes.10 Dopamine is a potent agonist at the DRD4 receptor, and DRD4 would be under increased agonist stimulation following the release of dopamine with inhibition of dopamine reuptake exerted by the commonly used stimulants.11 The DRD4 gene is located on chromosome 11p and has been identified to contain a number of polymorphic elements, including a 48-bp repeat motif in exon 3 which codes for the third cytoplasmic loop of the receptor.12 A limited amount of data suggests that receptors encoded by these variants containing differing numbers of repeat elements may vary in their agonist-binding affinity13 and receptor-coupled cyclic AMP response to agonist stimulation.14, 15 The gene variant containing seven repeats of the 48-bp motif (DRD4.7) has been assessed in a number of studies of novelty-seeking,16, 17, 18 Tourette syndrome,19 and other neuropsychiatric disorders,20, 21 with both positive and negative findings reported. In contrast, the majority of studies investigating the association of the DRD4.7 allele with ADHD have thus far found evidence that the gene variant contributes to susceptibility to ADHD,22, 23, 24, 25, 26 though one study which finds no association has been published.27 While the 48-bp repeat has functional consequences, it is not yet clear if this is the functional alteration in DRD4 which confers susceptibility to ADHD. Another possibility is that the DRD4.7 allele is in linkage disequilibrium with an unknown gene variant that conveys increased risk for ADHD. Yet another possibility would be that there are multiple DRD4 variants that independently contribute to ADHD risk, similar to that found in other complex disorders (eg, breast cancer and BRCA1).28 Therefore, these initial studies support the need to further investigate the possible role of other variations in the DRD4 gene with ADHD.
Recently, an additional polymorphism of the 5′ untranslated region of the DRD4 gene has been reported as common in the population.29 The polymorphism consists of a 120-bp tandem duplication 1.2 kb upstream from the initiation codon. The duplication may give rise to a PstI RFLP previously reported.30 Given that the region contains consensus sequences for several transcription factors, the variant may be hypothesized to be functionally relevant to protein expression.29 Because of our interest in examining the role of DRD4 gene variants as risk alleles for ADHD, we tested the 120-bp repeat polymorphism as a hypothesized risk gene for ADHD in a large sample of triads (ADHD children and parents) drawn from multiplex and simplex families, and demonstrate significant linkage by transmission disequilibrium testing (TDT). In addition, using data from our prior studies of the DRD4.7 allele and ADHD,24 we tested the transmission of haplotypes derived from the 120-bp repeat variant and the exon 3 7-repeat variant in this expanded sample of ADHD trios.
Materials and methods
The study was approved by the UCLA Human Subject Protection Committee. All children and parents provided written informed consent for their participation. As a portion of ongoing molecular-genetic studies of ADHD, 197 families including 371 affecteds and their parents underwent phenotype assessments to assess lifetime psychopathology and to identify families with both single and multiple children affected with ADHD. Families were accepted into the multiplex ADHD study if at least two children were affected with ADHD (any subtype) based on DSM-IV criteria, and at least one member met definite criteria with the other member definite or probable ADHD. Probable ADHD was defined as falling one symptom short of diagnostic criteria (including number of behavioral symptoms, age of onset, duration, or presence in two settings), but the child still had demonstrated impairment (cf sample description24).
All probands and first-degree relatives were directly interviewed and assessed using the K-SADS-PL31 for children and adolescents, and the SADS-LA-IV32 for adults. Both parent report of child symptoms and child report (for subjects 8 years or older) were obtained. The structured interview data were supplemented by teacher and parent rating scales (SNAP-IV, CBCL, TRF). All interviews were performed by experienced, research trained clinical psychologists familiar with ADHD who had completed reliability training on the instruments. Inter-rater reliability (kappa) exceeded 0.98 for all ADHD diagnoses, 0.95 for conduct disorder, and 0.84 for 10 other major psychiatric diagnoses (ie major depression, dysthymic disorder, etc) with greater than 5% frequency in the sample.
A consensus diagnostic panel led by senior child psychiatrists (JJM, JTM) generated ‘best estimate’ summary DSM-IV and DSM-III-R diagnoses from structured interview and questionnaire data reviewed blind to family status. In addition, full scale IQ estimates were determined for all probands using either the WISC-III or the WAIS-R, with all children required to have an IQ estimate of 70 or greater. All children evaluated were screened by history to rule out the presence of any potentially confounding comorbid medical or genetic disorders. The mean age of the affected children was 10.6 ± 3.5 years. The mean IQ was 105 ± 15. The families were predominantly Caucasian (81%) and represented a range of SES classes according to the Hollingshead ranks of: I (21%), II (38%), III (27%), IV (12%), and V (2%).
Blood samples were collected from family members and genomic DNA was extracted from whole blood with the Qiagen Kit using the manufacturer's protocol (Qiagen, Valencia, CA, USA). The DRD4 120-bp tandem repeat polymorphism was assayed according to methods described elsewhere29 with minor modifications. PCR amplication of 100 ng of genomic DNA was performed in 25-μl reactions containing 0.2 mM dNTPs, 0.2 μM of forward and reverse primers, Tris 10 mM (8.4 pH), 50 mM KCl, 2.5 mM MgCl2, and 0.5 U of Taq DNA polymerase (PE Applied Biosystems, Foster City, CA, USA). The primer sequences used were D4upstrFor2 (5′-GTT GTC TGT CTT TTC TCA TTG TTT CCA TTG -3′) and D4upstrRev3 (5′-GAA GGA GCA GGC ACC GTG AGC -3′). Using the MJ Research Tetrad thermal cycler, DNA was denatured at 95°C for 15 min, followed by 10 cycles of 95°C (30 s), 66°C (30 s), and 72°C (90 s), and 25 cycles of 95°C (30 s), 55°C (30 s), and 72°C (90 s), followed by a final extension at 72°C for 10 min. Ten microliters of PCR products were electrophoresed in 1.5% agarose gels in 1 × TBE buffer for 1 h at 120 V. The gels were stained with ethidium bromide and alleles were determined by comparison with known molecular weight standards. The PCR reaction yields distinct bands at 429 bp (short allele) and 549 bp (long allele). All gels were scored by two readers and discrepancies clarified.
Using the TDTEX program from the SAGE package,33 the asymptotic McNemar test was used to test for linkage disequilibrium of the 120-bp repeat gene in the 197 families. Multiple affected siblings per family were included in the TDT as independent affected-parent trios.34 Genotype relative risk of the 120-bp repeat gene was calculated according to the method of Risch and Merikangas.35 IBD sharing among the subset of affected sib pairs was determined using SIBPAL. Genotype and phenotype data were entered and double-checked using double-entry verification procedures. Linkage disequilibrium of the 120-bp and 48-bp VNTR loci was tested using the Associate program in LINKAGE.36 Genotype inconsistencies were identified using LODLINK. Univariate analyses were conducted using SAS Version 6.12.37
The sample consisted of 371 children with ADHD from 197 families, including 46 families of a single child with ADHD (23%), 132 families with two children with ADHD (67%), 15 families with three affected children with ADHD (8%), and four families with four affected children with ADHD (2%). Genotyping data were available for 308 parents (156 mothers and 152 fathers). Characteristics of the sample are shown in Table 1. In general, the demographics and clinical features of the sample were consistent with other ADHD samples in reported family-genetic and molecular genetic studies.
Table 2 displays the results of genotype analyses for the sample. The frequency of the allele in the overall sample is similar to that reported from analysis of a small number of subjects of mixed European origin, with the long allele (240 bp) appearing more common (81%) than the short allele (19%). The long allele is presumed to represent the derived allele, as it has not been found in samples from chimpanzees, gorillas, or orangutans.29 As expected, the 120-bp repeat gene was in linkage disequilibrium (allelic association) with the exon 3 VNTR from the same sample (χ2 = 8.32, P < 0.004). Results from the transmission disequilibrium test (TDT) in families selecting only one affected child per family revealed a significant association of the 240 allele and ADHD (P = 0.042). All subsequent linkage analyses were conducted using all affected children. Results from the TDT showed preferential transmission of the 240-bp (long) allele in the trios (P = 0.020; 138 vs 102 transmissions) for probable or definite ADHD by DSM-IV criteria. The results of the TDT test are also marginally significant (P = 0.053) when applied to the sample diagnosed by the earlier DSM-III-R criteria for probable or definite ADHD. Due to evidence suggesting possible genetic independence of ADHD subtypes,4, 5, 6 we also conducted exploratory TDT analyses restricting the sample to those trios by ADHD subtype (Inattentive vs Combined/Hyperactive-Impulsive). As seen in Table 2, the most robust evidence for linkage with the 240-bp allele was observed when the TDT analysis was restricted to those trios formed by probands meeting DSM-IV criteria for ADHD, Inattentive subtype (P = 0.005; 63 vs 35 transmissions). Although the heritability of ADHD does not appear to be influenced by comorbidity with major depression,2 we also performed secondary analyses after excluding all trios containing an ADHD proband with any diagnosis of a mood disorder (major depression, bipolar disorder, dysthmic disorder). The results of these analyses, in spite of reducing the sample, continued to show significant linkage of the 240-bp allele with both ADHD, all subtypes as well as ADHD, Inattentive subtype only (χ2 = 5.03, P < 0.025 and χ2 = 5.39, P < 0.02, respectively). The genotype relative risk (GRR) for the 240-bp allele was 1.4, based on the number of transmitted and non-transmitted proportions, when examined for the entire sample for probable or definite DSM-IV ADHD. Calculated for the Inattentive subtype only, the GRR rose to 1.8. Our estimated GRR for the 240-bp allele is comparable to the risk calculated for the DRD4.7 allele of 1.5 reported by our group24 and others.
Given that our initial interest in DRD4 was spurred in part by the finding of suggested linkage of the DRD4.7 allele with ADHD in our earlier report from our first 220 ADHD probands and their parents,24 with the expanded sample herein, we re-analyzed the relationship between DRD4.7 and ADHD in our full sample of 371 trios. Due to predominance of data from our own sample and others suggesting that it is the 7-repeat allele which confers risk for ADHD, we grouped all of the other alleles into a single category (DRD4 7- and not 7-repeat). While the 7-repeat continued to show greater transmission than not 7-repeats (111 to 96 transmissions), the difference did not reach a 0.05 significance level in the larger sample (P < 0.297). We also constructed haplotypes composed of genotype data of the 120-bp repeat and the exon 3 VNTR. Resulting haplotypes (see Table 3) were analyzed by TDT, with evidence found for preferential non-transmission of the 120-bp short allele/not 7 allele haplotype (P < 0.04; 68 transmitted vs 94 not transmitted).
As another test of the role of 120-bp repeat allele in ADHD, we tested the subsample of affected sibpair families for linkage of the DRD4 120-bp repeat polymorphism and ADHD by comparing the observed rate of IBD sharing with that expected by the null hypothesis (π = 0.5) for each informative meiosis. None of the linkage analyses was significant, even after stratifying the sample by ADHD subtype.
In this large sample of ADHD families, we report the first evidence of a possible role for another polymorphism of the DRD4 gene conferring increased susceptibility for ADHD, in this case a 120-bp repeat polymorphism in the 5′ untranslated region. Our data suggest that the 240-bp (long) allele is preferentially transmitted from parents to their ADHD offspring. Although a functional role for the polymorphism is not known at present, the observation that the duplication contains sequences that bind to several transcription factors raises the question of differences in degree of transcription rates between the variants.7 It is also of note that another recent report has identified a C to T polymorphism in the 5′ UTR at −521 which was observed to reduce DRD4 mRNA expression by 40%.38 Therefore, a possible mechanism for the role of the 240-bp repeat polymorphism in risk for ADHD may be via a reduction in transcription or expression of the DRD4 receptor. This hypothesis is testable with appropriate postmortem samples.
While the estimate of genotype relative risk associated with the 240-bp repeat variant is minor for the entire sample for all subtypes of ADHD (1.4), the finding is comparable to that reported earlier by our group for the DRD4.7 allele (1.5),24 and for DRD4.7 as reported by other groups.23, 26 The finding of a minor contribution to risk from the 240-bp polymorphism is consistent with the prevailing notion of ADHD as a polygenic disorder of multifactorial etiology, in which multiple risk genes interact to produce disease liability. Based on these data, it is also possible that DRD4.7 plays a minor role and that the 240-bp allele is the variant contributing the majority of risk at the DRD4 locus for ADHD. Consistent with that, our re-analysis of the role of the exon 3 VNTR did not confirm our earlier finding in the expanded sample. Similarly, analysis of haplotypes composed of the 120-bp polymorphisms and the exon 3 VNTR provided no support for an additional contribution of risk from the DRD4.7 allele. Our findings should encourage other groups to re-visit prior findings in relation to the exon 3 VNTR by also genotyping the 5′ UTR 120-bp variant in their ADHD samples.
Considerable evidence has accumulated supporting the relative independent heritability of behavioral dimensions of inattention vs hyperactivity-impulsivity5, 6 and the different categorical ADHD subtypes (Inattentive, Combined, Hyperactive-Impulsive), although some studies have been negative.39 Similarly, the genetic heterogeneity of ADHD has been suggested by additional family-genetic studies of ADHD comorbid with conduct disorder3, 39 or major depressive disorder.40 With the availability of our large sample of affected ADHD probands and their parents, we tested the association of the DRD4 120-bp repeat polymorphism in the ADHD subgroup of those with Inattentive subtype only. Restricting the analysis to those 142 trios containing the diagnosis of ADHD, Inattentive subtype, suggested stronger linkage with this subtype and the 240-bp allele (P = 0.005). Conversely, the association of the 240-bp allele was not significant when the TDT was restricted to those trios containing only probands with ADHD, Hyperactive-Impulsive or Combined subtypes, representing approximately 59% of the sample. In general, the results of these subgroup analyses suggest that if the DRD4 240-bp repeat allele is exerting a role in the etiology of ADHD, it may be more relevant to the inattentive dimension of the ADHD phenotype.
Interestingly, one prior report based on a smaller sample with a case-control design noted significant association of the DRD4.7 allele with the ADHD-Inattentive subtype25 and not with other ADHD subtypes. Given that the bulk of the ADHD probands in the Combined subtype group, comprising 50% of our ADHD sample, also by definition display inattentive behaviors, the lack of significant association with the 240-bp allele in the Hyperactive-Impulsive and Combined subtypes cases may be interpreted to show the influence of other risk alleles for ADHD.
Taken together, these findings support a tentative hypothesis that variants of DRD4 may more likely be associated with the inattentive behavioral dimension of ADHD, and hence, possible deficits of attentional functioning. Additional studies examining the role of DRD4 variants in relation to direct measures of attentional processing may be informative. Conversely, other risk alleles, such as variants at the DAT1 locus41 may be more relevant to risk for hyperactive-impulsive and combined subtype symptomatology. Given the modest estimates for GRR conferred by these allelic variants, large samples and other behavioral/cognitive testing paradigms are necessary to definitely test these hypotheses.
We found no evidence for increased IBD sharing at the DRD4 locus, which may reflect the lower power of IBD-based methods to detect effects of minor risk alleles. Alternatively, the absence of evidence for IBD sharing may be a reflection of the observed lack of significant association within this subsample of affected sibpairs for ADHD specific subtypes,39 in contrast to other studies finding evidence for genetic independence of subtypes.4, 5, 6 Given the high frequency of the 240-bp allele in the population, the effect of the 120-bp allele on ADHD risk may be hypothesized to exert a ‘protective’ influence on individuals, until an actual functional consequence of the variant is revealed.
The etiology of ADHD involves the complex interaction of multiple genetic factors with effects of gender, and environmental influences. These findings underscore the likely genetic heterogeneity of ADHD (and its subgroups). Given the data reported herein, exploration of functional consequences of the DRD4 120-bp repeat variant appear warranted. With the identification of a functional polymorphism also located in the 5′ UTR of the DRD4 gene at −521, additional efforts should be made to determine whether other polymorphisms exist in this region, and what relationship they have with risk for ADHD. Expanded research incorporating the phenotypic heterogeneity of ADHD also appears needed. Such research will require even larger samples of probands with ADHD than tested in this report. The interaction of the 120-bp polymorphism with other gene variants at the DRD4 locus and with related neurotransmitter genes is unknown, but merits investigation. Lastly, the 120-bp polymorphism could be tested as a predictor of treatment response in ADHD, and perhaps in other disorders.
This paper is dedicated to the memory of Dennis P Cantwell, MD. The authors also extend their appreciation to the families for their participation. The helpful contributions of T Kim, J NewDelman, E Gordon, E Carr, E Cantwell, and A Woodward are gratefully appreciated. Some of the results of the paper were obtained with the use of SAGE, which is supported by a US PHS Service Resource Grant (P41 RR03655) from the National Center for Research Resources. The project was supported by NIH grant MH 58277 (SLS) and the UCLA Tourette Syndrome Research Fund (JTM).
About this article
Molecular Neurobiology (2012)