Letters to the Editor
1066
No association between dopaminergic polymorphisms and intelligence variability in attention-deficit/ hyperactivity disorder Molecular Psychiatry (2006) 11, 1066–1067. doi:10.1038/sj.mp.4001900
Attention-deficit/hyperactivity disorder (ADHD) is one of the most common and heritable psychiatric disorders.1 Although it is characterized by core symptoms (inattention, hyperactivity and impulsivity), there is considerable heterogeneity in clinical features among individuals with ADHD.2 The use of endophenotypes has been suggested as an alternative approach to decrease this clinical heterogeneity in genetic investigations addressing ADHD.3,4 Recently, Mill et al.5 demonstrated an association between the two most investigated polymorphisms in ADHD, the 40 bp variable number of tandem repeats (VNTR) at the DAT1 gene and the 48 bp VNTR at the DRD4 gene, with intelligence variability among children with ADHD in two independent birth cohorts from Britain and New Zealand. In that study, children with ADHD diagnosis were assessed based on mother and teacher reports using respectively DSM-IV or DSM-III criteria for the Britain and New Zealand cohorts. These investigators defined genetic risk by the presence of at least one DRD4 seven repeat allele and/or by the presence of homozygous 10/10 DAT1 genotype. Children were
scored 0 if they carried no genetic risk, 1 if they carried one genetic risk factor and 2 if they carried both risks. In the British cohort (n = 171), children with one genetic risk scored 2.6 intelligence quotient (IQ) points lower than children with no genetic risk, and children with both genetic risks scored 8.2 IQ points lower than children with no risk (P = 0.02). These results were replicated in the New Zealand Cohort (n = 55) (P = 0.03). In both samples, the association between number of genotypes at risk and IQ remained significant after controlling for number of symptoms of hyperactivity/impulsivity and inattention. As it is well-established that replication is an important condition to accept a given hypothesis in psychiatric genetics, we decided to examine the hypothesis proposed by Mill et al.5 in three independent samples of Brazilian subjects with ADHD. The first two samples were composed by 242 referred children with ADHD recruited at the Child and Adolescent Psychiatric Division of the Hospital de Clı´nicas de Porto Alegre (HCPA), and 220 adults with ADHD recruited at the adult clinic in the same hospital. An ADHD consensus diagnosis based on DSM-IV criteria was achieved as described in detail previously.6,7 The third group was a community sample including 100 children with ADHD inattentive type ascertained from 12 public schools. The extensive diagnostic process, also based in DSM-IV criteria, used in this sample have been fully described elsewhere.8 The estimated IQ score was obtained from the Vocabulary and Block Design subtests of the Wechsler Intelligence Scale – Third Edition (WISC III)9 for children and from the same sub-tests of the WAIS-R10 for adults, both administered by trained psychologists. The DRD4 region containing the 48 bp repeat and the 40 bp VNTR site at DAT1 gene was amplified by the polymerase chain reaction (PCR) as described by Roman et al.6 The IQ scores were compared among groups through analysis of variance in the unadjusted model and
Table 1 IQ scores for subjects with ADHD defined by the presence or absence of the seven-repeat allele of the DRD4 gene and by the presence or absence of the DAT1 10/10 genotypea Samples
Low-risk genotype
1 genotype risk
DRD4/DAT1 risk
P-value
Childrenb N = 242
92.8 (15.0) n = 71
93.0 (14.6) n = 129
91.2 (13.8) n = 42
0.769
Childrenc N = 100
94.1 (10.8) n = 29
93.9 (10.8) n = 49
93.7 (12.4) n = 22
0.992
Adults N = 220
99.0 (8.0) n = 58
99.2 n = 45
0.147
101.3 (8.9) n = 117
Abbreviations: ADHD, attention-deficit/hyperactivity disorder; IQ, intelligence quotient. a Data are presented as mean (s.d.). b Children ascertained at the hospital. c Children ascertained at public schools. Molecular Psychiatry
Letters to the Editor
through analysis of covariance controlling for age, and sex. No significant association was seen both in the unadjusted model (Table 1) (children with ADHD from the clinical sample: P = 0.769; ADHD-I community sample: P = 0.992; sample of adults with ADHD: P = 0.147), and in the model adjusted for potential confounders (P = 0.648, P = 0.907 and P = 0.847, respectively from children with ADHD from the clinical sample, ADHD-I community sample and adults; data available upon request). Therefore, our results did not replicate previous findings from Mill et al.5 suggesting that DAT1 and DRD4 VNTR polymorphisms did not influence IQ variation in three independent samples of Brazilian subjects with ADHD. Although Mill et al.5 replicated their findings in two independent cohorts, their two ADHD samples were smaller than those investigated in Brazil. Thus, our negative results do not seem to be owing to lack of statistical power to replicate their findings. More important, these negative results suggest that between-study heterogeneity in the association between ADHD and DRD4/DAT1 might not be explained by variability in intellectual ability across subjects with ADHD. The role of these polymorphisms in the neurobiology of ADHD needs to be further investigated. JP Genro1, T Roman2, CP Zeni3, EH Grevet3, M Schmitz3, PB de Abreu3, CHD Bau1, LA Rohde3 and MH Hutz1 1 Department of Genetics, Federal University of Rio Grande do Sul, Porto Alegre, RS, Brazil; 2 Department of Morphological Sciences, Federal School of Medical Sciences of Porto Alegre, Porto Alegre, RS, Brazil and 3 Department of Psychiatry, Federal University of Rio Grande do Sul, Porto Alegre, RS, Brazil E-mail:
[email protected]
References 1 Faraone SV, Sergeant J, Gillberg C, Biederman J. World Psychiatry 2003; 2: 104–113. 2 Nigg JT, Willcutt EG, Doyle AE, Sonuga-Barke EJ. Biol Psychiatry 2005; 57: 1224–1230. 3 Castellanos FX, Tannock R. Nat Rev Neurosci 2002; 3: 617–628. 4 Sergeant J. Rev Bras Psiquiatr 2005; 27: 262–263. 5 Mill J, Caspi A, Williams BS, Craig I, Taylor A, Polo-Tomas M et al. Arch Gen Psychiatry 2006; 63: 462–469. 6 Roman T, Schmitz M, Polanczyk G, Eizirik M, Rohde LA, Hutz MH. Am J Med Genet 2001; 105: 471–478. 7 Grevet EH, Bau CH, Salgado CA, Fischer AG, Kalil K, Victor MM et al. Eur Arch Psychiatry Clin Neurosci 2006; 256: 311–319 (E-pub ahead of print). 8 Schmitz M, Denardin D, Silva TL, Pianca T, Roman T, Hutz MH et al. Biol Psychiatry 2006 (E-pub ahead of print), doi:10.1016/j.biopsych.2006.02.035. 9 Wechsler D. Examiner’s Manual: Wechsler Adult Intelligence Scale-Revised. Psychological Corporation: Cleveland, 1981. 10 Wechsler D. Examiner’s Manual: Wechsler Intelligence Scale for Children-Third Edition. Psychological Corporation: New York, 1991.
1067
Expression of ribosomal subunit genes increased coordinately with postmortem interval in human brain Molecular Psychiatry (2006) 11, 1067–1069. doi:10.1038/sj.mp.4001901
It has been well established that postmortem interval (PMI), the time between death and freezing of brain tissue, did not greatly affect the gene expressions.1–5 In fact, messenger RNA (mRNA) in human brain is relatively stable after death and its integrity was maintained over the long period of time.1–5 From the data analysis of two independent large sets of DNA microarray experiments, we unexpectedly found that expression of ribosomal subunit genes was consistently increased with PMI, implicating the existence of coordinated gene expression changes with PMI in human brain. Using DNA microarray data of postmortem human brains (set 1: n = 102),6 we systematically assessed the effect of demographic variables including PMI on gene expressions (See Supplementary information for detail). Expression of a total of 1754 out of 22 000 probe sets showed significant correlations with PMI (Pearson’s correlation, P < 0.05), consistent with the previous notion that PMI is a minor confounding factor compared with other factors (Supplementary Figure 1). In contrast to sample pH, which is closely associated with terminal condition of death and a major confounding factor for gene expression studies,1–5,7,8 PMI did not correlate with indicators of RNA integrity or sample quality (Figure 1a). Together with the fact that PMI did not correlate with sample pH (Supplementary Figure 2), we concluded that PMI did not affect quality of DNA microarray data. GeneOntology analysis of PMI-correlated probe sets revealed significant over-representation of ribosomerelated categories (Figure 1b). Detailed examination of annotations for all the PMI-correlated probe sets revealed that they included 77 probe sets, which corresponded to 54 distinctive ribosomal subunit genes (Supplementary Table 1). They included ribosomal protein components of both small and large subunits, and, strikingly, all of their expressions showed positive correlations with PMI (Figure 1c, for example). Positive correlation between expression of ribosomal subunit genes and PMI was also clearly found in the other independent data set of DNA microarray experiments (set 2: n = 50)9 (Supplementary Table 2). In this set, among 1106 PMI-correlated probe sets out of 12 000 probe sets, 59 probe sets corresponding to 56 distinctive ribosomal subunit genes could be identified (Supplementary Table 3). In both sample sets, positive correlation was also observed when we normalized Molecular Psychiatry
