Letters to the Editor 707 SADS-LB) and diagnoses met DSM-IV criteria. Asso- References ciation analyses were performed using the TDT,4 ETDT5 and Transmit6 programmes. 1 Tosato S, Dazzan P, Collier D. Association between the neuregulin 1 and schizophrenia: a systematic review. Schizophr Bull From the single marker association analysis, there 2005; 31: 613–617. was no evidence of association of NRG1 with BPAD in 2 Maier W, Hofgen B, Zobel A, Rietschel M. Genetic models our sample (Table 1). D8S1810 did exhibit a signi- of schizophrenia and bipolar disorder: overlapping inheritance ficant over transmission of allele 5 to bipolar probands or discrete genotypes? Eur Arch Psychiatry Clin Neurosci 2005; 255: 159–166. (uncorrected P = 0.04) but the global P-value for this 3 Green EK, Raybould R, Macgregor S, Gordon-Smith K, Heron J, marker was not significant (Table 1: P =0.212). Inter- Hyde S et al. Operation of the schizophrenia susceptibility estingly, this marker was selected based on its presence gene, neuregulin 1, across traditional diagnostic boundaries to within an Irish schizophrenia-associated haplotype, increase risk for bipolar disorder. Arch Gen Psychiatry 2005; 62: HapB .7 A range of two to three marker haplotypes, 642–648. IRE 4 Spielman RS, McGinnis RE, Ewens WJ. Transmission test for including HAPICE and HapBIRE, also failed to exhibit linkage disequilibrium: the insulin gene region and insulin- any evidence of association with BPAD (Table 1). A dependent diabetes mellitus. Am J Hum Genet 1993; 52: 506–516. haplotype comprised of alleles from the three micro- 5 Sham PC, Curtis D. An extended transmission/disequilibrium satellite markers (M3-M4-M5) did exhibit a marginal test for multi-allele marker loci. Ann Hum Genet 1995; 59(Pt3): 323–336. association (Table 1, haplotype 1, 3, 5, uncorrected 6 Clayton D. A generalization of the transmission/disequilibrium P = 0.04) which was not significant at the global test for uncertain-haplotype transmission. Am J Hum Genet 1999; haplotype level (Table 1: P = 0.31). We hypothesized 65: 1170–1177. that if NRG1 is principally a gene for schizophrenia, a 7 Corvin AP, Morris DW, McGhee K, Schwaiger S, Scully P, Quinn J stratified analysis restricted to trios with probands that et al. Confirmation and refinement of an ’at-risk’ haplotype for schizophrenia suggests the EST cluster, Hs.97362, as a potential have experienced at least one psychotic episode may susceptibility gene at the Neuregulin-1 locus. Mol Psychiatry 2004; reveal an association with BPAD, similar to the 9: 208–213. findings of Green et al.3 However, restriction of the 8 Thiselton DL, Webb BT, Neale BM, Ribble RC, O’Neill FA, association analysis to the subset of 59 trios with Walsh D et al. No evidence for linkage or association of neuregulin-1 (NRG1) with disease in the Irish study of high- psychosis did not reveal any evidence for association density schizophrenia families (ISHDSF). Mol Psychiatry 2004; 9: of NRG1 with BPAD (data not shown). 777–783. To conclude, we have failed to replicate the findings 9 Morris DMK, Schwaiger S, Nangle J, Murphy K, Donohoe G, of Green et al.3 in a family-based collection which is not Clarke S et al. Abstracts for the XIIth World Congress of subject to the population substructure effects of case– Psychiatric Genetics. Dublin, Ireland, 9–13 October 2004. Am J Med Genet B Neuropsychiatr Genet 2004; 130: 143 (poster control association studies. These findings suggest that 14.62). NRG1 is not a susceptibility gene for BPAD in the Irish 10 Gardner M, Gonzalez-Neira A, Lao O, Calafell F, Bertranpetit J, population. Interestingly, despite a plethora of positive Comas D. Extreme population differences across Neuregulin 1 associations of NRG1 with schizophrenia in diverse gene, with implications for association studies. Mol Psychiatry 2006; 11: 66–75. populations, the two Irish schizophrenia investigations have also been negative.8,9 Although a two-marker

NRG1 haplotype (HapBIRE) was initially reported to be associated with schizophrenia in an Irish case–control sample,7 the association was no longer evident after Serotonin transporter further investigation within an extended sample.9 We cannot rule out the possibility that our collection had polymorphism and insufficient power to replicate the association of NRG1 with BPAD. Indeed, there are moderate differences in LDL-cholesterol the allele frequencies for SNPs surrounding the core haplotype in European populations, which would impact the power of replication studies, particularly Molecular Psychiatry (2006) 11, 707–709. for rarer alleles/haplotypes.10 However, geographical doi:10.1038/sj.mp.4001837 variationisunlikelytobeanissueinthiscaseasthe estimated frequency of the core haplotype in our samples (7.7%) was very similar to that reported by The short allele of the serotonin transporter poly- Green et al.3 (7.8% in controls). To elucidate whether morphism (5HTTLPR) has been linked to psychiatric NRG1 is truly a susceptibility gene for BPAD, addi- disorders, including depression or anxiety. Our tional studies in family-based collections are required. analyses concerning the impact of 5HTTLPR on late- onset depression in a large community-based age- F Cassidy1, S Roche1, E Claffey2 and P McKeon2,3 cohort unexpectedly detected that 5HTTLPR strongly 1Smurfit Institute of Genetics, Trinity College, influences low-density lipoprotein (LDL)-cholesterol Dublin, Ireland; 2Depression Research Unit, levels. St Patrick’s Hospital, Dublin, Ireland and The serotonin transporter (SERT) regulates the 3Department of Psychiatry, Trinity College, serotonergic system via modulation of extracellular Dublin, Ireland. fluid serotonin concentrations. Transcriptional acti- E-mail: [email protected] vity of human SERT is modulated by a repetitive

Molecular Psychiatry Letters to the Editor 708 Table 1 Demographic data in different 5HTTLPR genotypes; unless otherwise indicated, data are means7s.d.

SS SL LL

n (LDL-cholesterol measured) 92 (91) 270 (267) 204 (203) Female/male 53/39 160/110 120/84 LDL-cholesterol (mg/dl)a 137737 145739 154743 HDL-cholesterol (mg/dl) 59715 59716 58714 Triglycerides (mg/dl) 125761 136777 138767 Systolic blood pressure (mm Hg) 147720 147720 147721 Diastolic blood pressure (mm Hg) 84710 82711 81710 Body mass index 277427742774 HbA1c (five of total haemoglobin) 5.871.1 6.073.5 5.870.9 History of smoking, yes/no 44/48 112/158 90/114 Myocardial infarction or stroke, yes/no 12/80 36/234 37/167 On lipid-lowering drugs yes/no 16/76 61/209 42/162

Abbreviations: HDL, high-density lipoprotein; LDL, low-density lipoprotein. aANOVA, df = 558, F = 6.44, P = 0.002; level of significance P = 0.006 (Bonferroni corrected).

element of varying length, in the 50 flanking region LPR in women (43 women with SS: 147 (37); 119 located B1.4 kb from the transcription start site, women with SL: 153 (33); 97 women with LL: 170 termed the serotonin transporter gene-linked poly- (41); ANOVA, df = 256, F = 7.66, P = 0.001), but non- morphic region (5HTTLPR). 5HTTLPR alleles are significantly in men (32 men with SS: 137 (35); 87 most commonly composed of either 14 (‘short’ men with SL: 143 (44); 64 men with LL: 149 (36)). allele S) or 16 (‘long’ allele L) repeated elements. Therapy with lipid-lowering drugs, high-density The 5HTTLPR genotype distribution in a Caucasian lipoprotein-cholesterol levels, triglyceride levels, population is 32% LL, 49% LS and 19% SS.1 The blood pressure, HbA1c, history of smoking and body serotonin transporter may contribute to several dis- mass index were not associated with 5HTTLPR. A orders and most reports comprise neuropsychiatric history of myocardial infarction or stroke in indivi- phenotypes such as anxiety disorders, depression and duals not taking lipid-lowering drugs was found in bipolar disorder linked to the S allele.2 There are only 16% of individuals with LL genotype, but with a few examples of apparent SERT-related dysfunction significantly lower frequency (8.8%) in individuals in the periphery. Nevertheless, serotonin plays an carrying the S allele (Fisher’s exact test: P = 0.029). important role outside the central nervous system, Data from this epidemiological study replicated the particularly in sustaining gastrointestinal function.3 finding of an association between artherosclerosis Two studies found significantly greater risk for events and the LL genotype of 5HTTLPR. Our myocardial infarction in individuals with the LL analyses further detected that 5HTTLPR influences genotype.4,5 The latter findings remain unexplained. LDL-cholesterol levels, especially in women. The The Vienna Transdanube Aging (VITA) study L-allele was associated with higher LDL-cholesterol, investigates an age cohort (mean age 75.7 years, and the S-allele with lower LDL-cholesterol. In case s.d. = 0.45 years) of all inhabitants of a defined of replication, this finding will induce a great deal geographical area in Vienna, Austria.6 Six hundred of research on the serotonergic influence on and ninety-seven out of 1505 inhabitants of this LDL-cholesterol. age participated; blood could be investigated in 606 participants and 5HTTLPR could be examined in 566 P Fischer1, E Gruenblatt2, P Pietschmann3 individuals (Table 1). and K-H Tragl4 The 5HTTLPR genotype distribution in the VITA 1Department of General Psychiatry, Medical population is 36% LL, 48% SL and 16% SS, which University of Vienna and Ludwig Boltzmann Institute does not deviate from Hardy–Weinberg equilibrium of Aging Research, Vienna, Austria; 2Institute of (w2 = 0.0281, P = 0.986). Mean fasting LDL-cholesterol Clinical Neurochemistry and National Parkinson levels (mg/dl) were higher in individuals with the Foundation, Centre of Excellence, Laboratory, Clinic L-allele than in individuals with the S-allele (Table 1). for Psychiatry and Psychotherapy, Bayerische, Julius- After exclusion of 119 probands who took lipid- Maximilians-University of Wu¨rzburg, Wurzburg, lowering drugs, the LDL-cholesterol values were 161 Germany; 3Center of Physiology and Pathophysiology, (741) for 161 individuals with LL, 149 (738) for 206 Medical University Vienna and Ludwig Boltzmann individuals with SL and 143 (736) for 75 individuals Institute of Aging Research, Vienna, Austria and with the SS genotype (analysis of variance (ANOVA), 4Ludwig Boltzmann Institute of Aging Research, df = 439, F = 7.47, P = 0.001). Low-density lipoprotein- Vienna, Austria cholesterol levels depended significantly on 5HTT E-mail: [email protected]

Molecular Psychiatry Letters to the Editor 709 References from reports that autistic symptoms are common in patients with chromosomal abnormalities of this 1 Lesch KP, Bengel D, Heils A, Sabol SZ, Greenberg BD, Petri S et al. Association of anxiety-related traits with a polymorphism in the region, especially if there is duplications of the serotonergic transporter gene regulatory region. Science 1996; 274: maternally derived 15, as outlined in 1527–1531. Bolton et al.2 There have also been a number of 2 Murphy DL, Lerner A, Rudnick G, Lesch KP. Serotonin transporter: suggestive linkage findings.3–7 gene, genetic disorders, and pharmacogenetics. Mol Intervent 2004; Our group had marginally suggestive results from 4: 109–123. 3 Blakely RD. Physiological genomics of antidepressant targets: an association analysis of ASD with five microsatel- 8 keeping the periphery in mind. J Neurosci 2001; 21: 8319–8323. lite markers spread across the PWACR. Furthermore, 4 Fumeron F, Betoulle D, Nicaud V, Evans A, Kee F, Ruidavets JB association studies have found evidence implicating et al. Serotonin transporter gene polymorphism and myocardial candidate in the region. GABRB3, composed of infarction. Circulation 2002; 105: 2943–2945. 5 Coto E, Reguero JR, Alvarez V, Morales B, Batalla A, Gonzales P nine exons spanning 250 kb, is thought probably not et al. 5-Hydroxy-tryptainine 5-HT2A receptor and 5-hydroxytrypt- to be imprinted. Several reports have documented amine transporter polymorphisms in acute myocardial infarction. findings of allelic association at microsatellite mar- Clin Sci (London) 2003; 104: 241–245. kers near/in the gene GABRB3; 155CA-29,10 and 6 Fischer P, Jungwirth S, Krampla W, Weissgram S, Kirchmeyr W, GABRB311 and more recently a dense set of single Schreiber W et al. Vienna-Transdanube Aging (VITA): Study Design, Recruitment, and Level of Participation. J Neural Transm 2002; nucleotide polymorphism (SNP) markers on GABRB3 62(Suppl): 105–116. found nominal significance for association with five SNPs across the gene.12 GABRG3 has been reported to be associated with ASD.13 A recent detailed examina- tion of GABA receptor genes in autism found An association analysis of evidence for some of these GABA receptor gene subtypes to be involved in aetiology, perhaps invol- candidate genes on ving complex gene–gene interactions between the subtypes.14 However, none of the GABA genes in the q11–13 and 15q11–13 region were found to be associated with autism in that study. autism spectrum disorder Evidence to support a role for the GABA neuro- transmitter system in autism susceptibility include;

(a) reduced binding of radiolabeled GABAA receptor Molecular Psychiatry (2006) 11, 709–713. ligands in autopsy brain specimens from individuals doi:10.1038/sj.mp.4001839 with autism,15 (b) elevated levels of circulating GABA and its essential precursor glutamate have been Autism spectrum disorder (ASD) is conceptualized as observed in children with autism16 and (c) during a spectrum of related diagnostic categories, compris- development it is known to have neurotrophic ing subtypes: autism, atypical autism, Asperger’s syn- effects.17 Furthermore, there is evidence to suggest drome and other pervasive developmental disorders that disruption of GABAergic interneuron develop- (ICD-10). Available evidence suggests that several ment may represent a common point of convergence genes contribute to the underlying genetic risk for the underlying the etiologies of developmental disorders development of ASD. However, etiologic and pheno- including autism.18 Also, we reported in vivo, using typic heterogeneity impede the discovery of ASD- magnetic resonance imaging (MRS), that people with susceptibility genes. We tested 148 simplex UK ASD have increased levels of the GABA precursors families, characterized by both ADI-R and ADOS to glutamate/glutamine in the amygdala-hippocampal have an autism spectrum disorder (ASD), for associa- complex (a brain region crucial to social communica- tion with candidate genes, GABRB3, GABRG3, tion (Page et al., in press 19). UBE3A and ATP10C, on chromosome 15q11–13. A UBE3A encodes a member of a family of function- dense map of single nucleotide polymorphisms (SNPs) ally related important in ubiquitin transfer that had not been previously parsed in autism, spaced and is maternally expressed in brain with cell and on average by 13 kb, were genotyped across each region specificity.20 Markers in UBE3A have also been gene. We found evidence for an association with reported to be associated with autism6 and a mixed GABRB3 and ASD. A haplotype analysis of epigenetic/genetic model for oligogenic inheritance of five adjacent SNPs in high linkage disequilibrium autism with a limited role for UBE3A has been (LD) on GABRB3 was significant with a Bonferroni- proposed.21 ATP10C is a probable calcium-transport- corrected P-value of 0.00013 for haplotypes of ing ATPase with wide expression including the brain. frequency > 1% (99.9% of all possible haplotypes Meguro et al.22 reported that the ATP10C gene is covered were represented by seven haplotypes). maternally expressed, that it maps within the most Further systematic examination of GABRB3 in relation common interval of deletion responsible for Angel- to ASD is required. man syndrome (15q11–q13), and that ATP10C expres- Chromosome 15q11–q13 (the Prader-Willi/Angel- sion is virtually absent from Angelman syndrome man Critical Region – PWACR) has been identified as patients with imprinting mutations, as well as a strong candidate region.1 Evidence for this comes from patients with maternal deletions of 15q11–q13.

Molecular Psychiatry