Does<i>BDNF</i>Val66Met contribute to preclinical Alzheimer’s disease?

This scientific commentary refers to ‘ BDNF Val66Met moderates memory impairment, hippocampal function and tau in preclinical autosomal dominant Alzheimer’s disease’, by Lim et al. (doi: 10.1093/brain/aww200 ) . Brain-derived neurotrophic factor (BDNF) is a protein of 247 amino acids (NP_001137277) that is encoded on chromosome 11p13 in humans. During its biogenesis, the protein traverses through the endoplasmic reticulum and Golgi apparatus before it is stored in dense core vesicles. Following its activity-dependent release into the synaptic cleft, BDNF can bind to several receptors, including TrkB and p75. The activation of these and other cell surface receptors by an intricate balance of BDNF expression products, involving various BDNF splice isoforms and post-translational products, affects brain physiology and neural plasticity through a context-dependent promotion of long-term potentiation or depression. In addition, neurotrophic signalling elicited by BDNF modulates the architecture of the brain by influencing the stability of dendritic spines and promoting neurogenesis ( Martinowich et al. , 2007 ). In this issue of Brain , Lim and co-workers show that a common BDNF polymorphism modulates Alzheimer-related endophenotypes in individuals with mutations in the APP , PSEN1 or PSEN2 genes ( Lim et al. , 2016 ). Several coding variations have been reported in the human BDNF gene. The most studied BDNF variation, Val66Met (rs6265), maps to the pro-domain that is absent in mature BDNF ( Fig. 1 ) and results in a 30% reduction in BDNF secretion, most likely by partially impairing its packaging into dense core vesicles. The Met66 allele is common in human populations (e.g. ∼30% heterozygotes and ∼4% homozygotes in datasets of European origin). Following early reports proposing that the Val66Met polymorphism confers risk of sporadic Alzheimer’s disease ( Ventriglia et al. , 2002 ), the presence of this allele has been linked to a wide range of conditions from psychiatric and personality disorders, including depression, substance abuse disorders, eating disorders, schizophrenia and neuroticism, to lower mean intelligence or body mass index. More recently, the involvement of the Val66Met variation in the pathogenesis of neurodegeneration has again been a topic of intense study, following observations of associations with hippocampal atrophy, reduced hippocampal activity and impairments in episodic memory ( Lim et al. , 2013 ). A survey of this body of literature, which now comprises more than a hundred articles, reveals surprisingly inconsistent conclusions. The divergence of findings has been attributed to differences between studies in variables such as ethnicity, age, gender, and phenotypic assessment ( Hong et al. , 2011 ). Schematic depicting Alzheimer-related endophenotypes affected by familial Alzheimer’s disease-causing mutations and the BDNF Val66Met polymorphism. Green and blue outlined boxes depict previously described observations that emerged in separate studies of familial Alzheimer’s disease mutations and the Met66 allele, respectively. The figure depicts only preclinical Alzheimer’s disease endophenotypes. Background shading is used to highlight phenotypes emphasized by Lim et al. (2016) . The box shown on the right depicts key endophenotypes, which were either unaffected by BDNF Val66Met variation (amyloid deposition) or were exacerbated by it (CSF tau/p-tau levels, hippocampal volume, and episodic memory). NC = non-carriers. The current study by Lim and co-workers builds on previously published work by the same group, which documented an effect of the Val66Met polymorphism on the severity of endophenotypes in preclinical and prodromal sporadic Alzheimer’s disease ( Lim et al. , 2013 , 2014 , 2015 ). It extends these earlier data by showing that the presence of the Met66 allele correlates with a more severe preclinical presentation of Alzheimer-related endophenotypes in individuals carrying highly penetrant dominant mutations in the APP , PSEN1 or PSEN2 genes, who on average were up to 12 years younger than the mean age of onset for a given type of mutation observed in autosomal dominant Alzheimer’s disease ( Lim et al. , 2016 ). More specifically, the study investigated neuroimaging measures in 143 asymptomatic Alzheimer’s disease mutation carriers and 131 non-carriers, recruited via the Dominantly Inherited Alzheimer Network (DIAN) study that is based on ongoing work with 197 families with APP , PSEN1 or PSEN2 mutations ( http://clinicaltrials.gov/ct2/show/NCT01760005 ). Aside from its access to this unique preclinical dataset, the Lim et al. study stands out by virtue of its characterization of a broader spectrum of primary and secondary Alzheimer’s disease-related phenotypes and biomarkers. Cases and controls were assessed by Mini-Mental State Examinations, and were scored against the Clinical Dementia Rating and Geriatric Depression Scales. Participants were also subjected to a battery of neuropsychological tests, and their CSF analysed for levels of amyloid-β 42 , tau and phospho-tau (p-tau). Furthermore, their brains were imaged by PET for amyloid deposition and metabolic activity following bolus injections of radiolabelled Pittsburgh compound B and fluorodeoxyglucose probes, respectively. As a result of this formidable effort, this is the first report to establish that the Met66 allele is associated with increased tau and p-tau levels in the CSF in mutation carriers. Moreover, among the mutation carriers group, only individuals with the Met66 allele, but not Val66 homozygotes, were observed to exhibit additional impairments in hippocampal metabolism and episodic memory. Consistent with previous reports, the presence of the APOE -ε4 allele, but not the Met66 allele, was associated with an increased burden of amyloid-β deposits in the mutation carriers cohort. The data endorse a cascade of events that sees APOE influence mutation-dependent amyloid-β pathology independent of BDNF genotype, which plays out downstream by exacerbating increases in CSF tau/p-tau levels, characteristic of Alzheimer’s disease, and by impairing hippocampal glucose metabolism and memory functions. This study raises several questions, some of them easily addressed and others requiring deeper investigation. For instance, currently no conclusions can be drawn on possible relationships between Met66-dependent phenotypes and subtypes of familial Alzheimer’s disease mutations (e.g. those targeting the PSEN1 or PSEN2 enzymes versus the APP substrate), because a breakdown of this information was not provided. Similarly, BDNF genotypes were available only for the mutation carriers cohort (i.e. 95 Val66 homozygotes and 48 Met66 carriers) but not for the non-carriers cohort, leaving some uncertainty regarding the degree to which phenotypes in Met66 carriers manifest in dependence or independence of an underlying pathology caused by the familial Alzheimer’s disease mutations (see also below). Moreover, because the family relationships amongst study participants are not detailed, it is also difficult to assess whether inclusion of closely related individuals from large pedigrees might have introduced inadvertent biases. For instance, where P -values were determined to be borderline significant, the question arises as to whether sample selection choices (ethnicity, pedigree structure, and types of mutations) might have masked or exaggerated differences. The study did not distinguish between the effects of Val/Met and Met/Met genotypes raising the prospect that a future study based on either a larger sample size or a cohort selectively enriched for Met66 homozygotes might strengthen conclusions by allowing stratification of data based on allele dosage. A future study may also need to clarify discrepancies in the conclusions drawn by Lim et al. in the current study and a previous report by a subset of the authors ( Kim et al. , 2015 ), which suggested that Val66Met genotypes and plasma BDNF levels are not associated with hippocampal volume or memory impairments in middle-aged and older adult cohorts (with or without early-stage dementia). In addition, because the Val66Met polymorphism exists in linkage disequilibrium with several other BDNF polymorphisms ( Kazantseva et al. , 2015 ) that may influence its interactions and downstream phenotypes, future studies will need to address the extent to which these other polymorphisms influence the phenotypes presented in the Lim et al. study. It is also possible that next generation sequencing may reveal rare functional BDNF variations that contribute to Alzheimer’s disease. Notably, the Ensemble database currently has a record of 58 missense BDNF substitutions, including variations predicted to be deleterious (e.g. rs780128716 and rs758638310). Finally, it remains to be determined whether the variations in well-confirmed Alzheimer’s disease loci identified by large genome-wide association studies could modify the phenotype of patients with mutations in Mendelian Alzheimer’s disease genes, in addition to BDNF variation(s). The model proposed by Lim et al. , whereby consequences of the Val66Met polymorphism feed into and exacerbate the cascade of amyloid-β-dependent Alzheimer’s disease pathogenesis events, is attractive but not yet conclusive. Thus, the current data do not exclude the possibility that Met66-mediated impairments manifest in parallel and are merely additive to amyloid-β-dependent damage because they affect a subset of overlapping phenotypes. Even an effect of the Val66Met variation on the levels of tau and p-tau in the CSF could be additive to the Alzheimer-related tau pathology and not be dependent on it. To begin to resolve this question, it would be informative to compare CSF tau and p-tau levels in mutation non-carriers. Answers to this question may also emerge from longitudinal data on whether the future age of onset of Alzheimer’s disease symptoms in the mutation carriers cohort reflects the individuals’ BDNF Val66Met genotypes. Most importantly, the precise mechanism by which the Val66Met polymorphism causes differences in the pathophysiology of neurodegenerative disorders remains to be identified. To this end, a mouse model is available in which the endogenous Bdnf gene was point-mutated using a knock-in strategy to reproduce the human Val66Met polymorphism ( Chen et al. , 2006 ). Consistent with reports in humans, cultured neurons from homozygous Met66 mice were observed to exhibit a 30% decrease in activity-dependent BDNF release. Moreover, the knock-in mice exhibited a decrease in the complexity of dendritic arbours on neurons within the dentate gyrus, which may or may not be causative for reductions in hippocampal volume. It will be interesting to cross these mice with Alzheimer’s disease or frontotemporal dementia (tau) mouse models in order to dissect the time-course and nature of the molecular events that underlie Met66-related impairments.

Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.