Mitochondrial DNA sequence characteristics modulate the size of the genetic bottleneck

Human Molecular Genetics, 2016, Vol. 25, No. 5 1031–1041 doi: 10.1093/hmg/ddv626 Advance Access Publication Date: 5 January 2016 Association Studies Article A S S O C I AT I O N S T U D I E S A R T I C L E Mitochondrial DNA sequence characteristics modulate the size of the genetic bottleneck 1 Institute of Genetic Medicine, 2Wellcome Trust Centre for Mitochondrial Research and 3Institute of Neuroscience, Newcastle University, Newcastle upon Tyne, UK, 4Medical Research Council Mitochondrial Biology Unit, Cambridge, UK, 5Department of Clinical Neurosciences, School of Clinical Medicine, University of Cambridge, Cambridge, UK, 6Department of Clinical Genetics, Research Schools GROW/CARIM, Maastricht University Medical Center, Maastricht, Netherlands, 7Division of Molecular Neurogenetics, National Neurological Institute ‘C. Besta’, Milano, Italy, 8IRCCS Institute of Neurological Sciences of Bologna, Bellaria Hospital, Bologna, Italy, 9Unit of Neurology, Department of Biomedical and NeuroMotor Sciences (DIBINEM), University of Bologna, Bologna, Italy, 10Department of Neurology, Haukeland University Hospital, Bergen, Norway, 11Department of Clinical Medicine (K1), University of Bergen, Bergen, Norway, 12Vanderbilt Genetics Institute, Department of Molecular Physiology and Biophysics, Vanderbilt School of Medicine, Nashville, TN, USA and 13Department of Genetics, Faculty of Science, Kasetsart University, Bangkok, Thailand *To whom correspondence should be addressed at: Department of Clinical Neurosciences, University of Cambridge, Addenbrookes Biomedical Campus, Cambridge CB2 0QQ, UK. Tel: +44 1223217091; Fax: +44 122333694; Email: Abstract With a combined carrier frequency of 1:200, heteroplasmic mitochondrial DNA (mtDNA) mutations cause human disease in ∼1:5000 of the population. Rapid shifts in the level of heteroplasmy seen within a single generation contribute to the wide range in the severity of clinical phenotypes seen in families transmitting mtDNA disease, consistent with a genetic bottleneck during transmission. Although preliminary evidence from human pedigrees points towards a random drift process underlying the shifting heteroplasmy, some reports describe differences in segregation pattern between different mtDNA mutations. However, based on limited observations and with no direct comparisons, it is not clear whether these observations simply reflect pedigree ascertainment and publication bias. To address this issue, we studied 577 mother–child pairs transmitting the m.11778G>A, m.3460G>A, m.8344A>G, m.8993T>G/C and m.3243A>G mtDNA mutations. Our analysis controlled for inter-assay differences, inter-laboratory variation and ascertainment bias. We found no evidence of selection during transmission but show that different mtDNA mutations segregate at different rates in human pedigrees. m.8993T>G/C † I.J.W. and P.J.C. contributed equally to the study. Received: September 2, 2015. Revised: November 26, 2015. Accepted: December 22, 2015 © The Author 2016. Published by Oxford University Press. This is an Open Access article distributed under the terms of the Creative Commons Attribution Non-Commercial License (http://creativecommons.org/ licenses/by-nc/4.0/), which permits non-commercial re-use, distribution, and reproduction in any medium, provided the original work is properly cited. For commercial re-use, please contact 1031 Ian J. Wilson1,†, Phillipa J. Carling1,2,†, Charlotte L. Alston2,3, Vasileios I. Floros4,5, Angela Pyle1,2, Gavin Hudson1,2, Suzanne C.E.H. Sallevelt6, Costanza Lamperti7, Valerio Carelli8,9, Laurence A. Bindoff10,11, David C. Samuels12, Passorn Wonnapinij13, Massimo Zeviani4,7, Robert W. Taylor2,3, Hubert J.M. Smeets6, Rita Horvath1,2 and Patrick F Chinnery2,4,5,* 1032 | Human Molecular Genetics, 2016, Vol. 25, No. 5 segregated significantly faster than m.11778G>A, m.8344A>G and m.3243A>G, consistent with a tighter mtDNA genetic bottleneck in m.8993T>G/C pedigrees. Our observations support the existence of different genetic bottlenecks primarily determined by the underlying mtDNA mutation, explaining the different inheritance patterns observed in human pedigrees transmitting pathogenic mtDNA mutations. Introduction Results Determining the potential impact of ascertainment bias Given previous concerns about ascertainment bias when studying the inheritance of heteroplasmy in human pedigrees (9), first we performed a simulation experiment to determine the possible consequences of identifying pedigrees through a clinically affected child. We then determined whether the standard approach of omitting the affected proband minimizes any bias to an acceptable level. The simulations were based on an established model for the mtDNA genetic bottleneck using measurements of heteroplasmy made in human oocytes for neutral alleles (i.e. with no selection) (12,13). We studied the difference in heteroplasmy level between a mother and child (ΔM-O) in simulated pedigrees in silico, with three possible strengths of the mtDNA genetic bottleneck (where bottleneck parameter, b = 0.9, is a weak bottleneck; b = 0.7 is an intermediate strength bottleneck; and b = 0.2 is a strong Analysis of human pedigree data To generate the largest data set possible, we studied pedigrees derived from a meta-analysis of published data [n = 532 mother–child pairs transmitting five common heteroplasmic mtDNA mutations: m.11778G>A (n = 117), m.3460G>A (n = 74), m.8344A>G (n = 96), m.8993T>G/C (n = 117) and m.3423A>G (n = 128), citations shown in the Material and Methods]; and 45 new unpublished mother–child pairs [m.8344A>G (n = 9), m.8993T>G/C (n = 2) and m.3243A>G (n = 34)] measured in two centres. Given that there was no obvious difference between the two data sets (Fig. 3), we merged all of the data and minimized ascertainment bias by analysing separately data with and without the clinically affected probands. The final data set included 467 mother–child pairs for the uncorrected data. m.3243A>G was analysed before and after correcting for the known decrease in leucocyte heteroplasmy levels for this specific mutation using the published approach (14). For this correction, only individuals with a heteroplasmy level of <95% were included to avoid pairs where the mother or offspring values corrected to >100%. This reduced the sample size of m.3243A>G from 137 to 99 pairs. The entire data set was used to determine the likely size of the bottleneck parameter, b, using the model described previously (12,13), incorporating the laboratory assay and laboratory site as covariates. Bayesian statistical analyses were performed using JAGS (15). For each mutation, the average change in heteroplasmy was not significantly different from zero, consistent with no selection for or against the mtDNA mutations during transmission. The posterior differences in bottleneck strength, b, are shown in Figure 4. As predicted from the simulations, the bottleneck strength was overestimated (i.e. b (...truncated)