Brain Structure and Function (2020) 225:2331–2345 https://doi.org/10.1007/s00429-020-02125-3 ORIGINAL ARTICLE Effect of BDNF Val66Met on hippocampal subfields volumes and compensatory interaction with APOE‑ε4 in middle‑age cognitively unimpaired individuals from the ALFA study Natalia Vilor‑Tejedor1,2,3,4 · Grégory Operto2,5,6 · Tavia E. Evans3 · Carles Falcon2,6,7 · Marta Crous‑Bou2,8,9 · Carolina Minguillón2,5,6 · Raffaele Cacciaglia2,5,6 · Marta Milà‑Alomà2,4,5,6 · Oriol Grau‑Rivera2,5,6,10 · Marc Suárez‑Calvet2,5,6,10 · Diego Garrido‑Martín1 · Sebastián Morán11 · Manel Esteller11,12,13,14 · Hieab H. Adams3,15,16 · José Luis Molinuevo2,4,5,6 · Roderic Guigó1,4 · Juan Domingo Gispert2,4,6,7 · for the ALFA Study Received: 23 April 2019 / Accepted: 30 July 2020 / Published online: 17 August 2020 © The Author(s) 2020 Abstract Background Current evidence supports the involvement of brain-derived neurotrophic factor (BDNF) Val66Met polymor- phism, and the ε4 allele of APOE gene in hippocampal-dependent functions. Previous studies on the association of Val66Met with whole hippocampal volume included patients of a variety of disorders. However, it remains to be elucidated whether there is an impact of BDNF Val66Met polymorphism on the volumes of the hippocampal subfield volumes (HSv) in cogni- tively unimpaired (CU) individuals, and the interactive effect with the APOE-ε4 status. Methods BDNF Val66Met and APOE genotypes were determined in a sample of 430 CU late/middle-aged participants from the ALFA study (ALzheimer and FAmilies). Participants underwent a brain 3D-T1-weighted MRI scan, and volumes of the HSv were determined using Freesurfer (v6.0). The effects of the BDNF Val66Met genotype on the HSv were assessed using general linear models corrected by age, gender, education, number of APOE-ε4 alleles and total intracranial volume. We also investigated whether the association between APOE-ε4 allele and HSv were modified by BDNF Val66Met genotypes. Results BDNF Val66Met carriers showed larger bilateral volumes of the subiculum subfield. In addition, HSv reductions associated with APOE-ε4 allele were significantly moderated by BDNF Val66Met status. BDNF Met carriers who were also APOE-ε4 homozygous showed patterns of higher HSv than BDNF Val carriers. Conclusion To our knowledge, the present study is the first to show that carrying the BDNF Val66Met polymorphisms partially compensates the decreased on HSv associated with APOE-ε4 in middle-age cognitively unimpaired individuals. Keywords APOE-ε4 · BDNF · Hippocampal subfields · Imaging genetics · Subiculum · Val66Met Introduction Brain-derived neurotrophic factor (BDNF) is a neurotro- The complete list of collaborators of the ALFA Study can be found phin involved in neurogenesis and synaptic plasticity in the in the Acknowledgements. central nervous system, especially in the hippocampus, and has been implicated in the pathophysiology of several neu- Electronic supplementary material The online version of this ropsychiatric disorders (Bathina and Das 2015; Autry and article (https​://doi.org/10.1007/s0042​9-020-02125​-3) contains Monteggia 2012; Numakawa et al. 2018). The single nucleo- supplementary material, which is available to authorized users. tide polymorphism (SNP) rs6265 (also known as Val66Met), * Natalia Vilor‑Tejedor causes a valine (Val) to methionine (Met) substitution at natalia.vilortejedor@crg.eu codon 66 of BDNF protein. Particularly, the study of Val- * Juan Domingo Gispert 66Met polymorphism within the BDNF gene is of special jdgispert@barcelonabeta.org interest because of its documented impact on hippocampal- Extended author information available on the last page of the article dependent functions (Notaras and van den Buuse 2018; Toh 13 Vol.:(0123456789) 2332 Brain Structure and Function (2020) 225:2331–2345 et al. 2018; Egan et al. 2003; Hariri et al. 2003). Hence, APOE-ε4 and BDNF have been suggested to influence their extensive research focuses on the discovery of associa- secondary effects on AD pathology (Álvarez et al. 2014), tions between BDNF Val66Met polymorphism and several and their influence on hippocampal volume (Li et al. 2016; hippocampal phenotypes. However, recent meta-analyses Shi et al. 2014; Liu et al. 2015a, b). In addition, a significant addressing hippocampal volumes for BDNF Val66Met have combined effect of APOE-ε4 and BDNF Val66Met poly- reported inconsistent statistically significant associations, as morphisms has been reported to moderate β-amyloid-related well as inconsistencies regarding the direction of the geno- cognitive decline in preclinical AD (Lim et al. 2015). Epi- type effects across individual studies (Harrisberger et al. sodic memory performance was also found to be impaired 2014, 2015). in MCI/AD individuals who were also carriers of both the Two recent large meta-analyses suggest that the analy- APOE-ε4 and BDNF Met polymorphisms (Gomar et al. sis of hippocampal subfield volumes may allow for more 2016), as well as in healthy individuals (Ward et al. 2014). accurate detection of genetic effects in genetic association Overall, evidence suggests biological interactions between analyses, compared with whole hippocampal volume (van APOE and BDNF for memory and other brain-related pro- der Meer et al. 2018; Hibar et al. 2017). Moreover, previ- cesses that may help to explain the increased AD risk in ous studies have shown that different pathological condi- APOE-ε4 carriers during the period that precedes the devel- tions affect subfields differently (West et al. 1994; Jin et al. opment of symptoms. 2004; Ezzati et al. 2014; Mueller et al. 2010; Hett et al. Therefore, the aim of the present study is to evaluate the 2018). In fact, the proven differential expression of BDNF impact of Val66Met polymorphism on hippocampal sub- and its receptors in different regions of the hippocampus fields in a large sample of in middle-age cognitively unim- (Kowiański et al. 2018; Vilar and Mira 2016; Franzmeier paired individuals CU participants and to assess whether an et al. 2019), reinforces distinct biological functions of BDNF interactive effect with the APOE-ε4 genotype exists. Val66Met polymorphism on the different subfields. How- ever, to our knowledge no previous studies have addressed the effects of the BDNF Val66Met polymorphism on hip- Materials and methods pocampal subfields in cognitively unimpaired (CU) indi- viduals. Most of the studies addressing the association of Study population and setting BDNF Val66Met polymorphism and hippocampal volumes (subfields and/or whole hippocampus) included patients of a Participants were drawn from the ALFA study (Alzhei- variety of neuropsychiatric disorders, such as major depres- mer and FAmilies) established at the Barcelonaβeta Brain sive disorder, schizophrenia and bipolar disorder (Zeni et al. Research Center (Molinuevo et al. 2016), which aims at 2016; Cao et al. 2016; Reinhart et al. 2015; Aas et al. 2014; identifying the neuroimaging and cognitive signatures in Frodl et al. 2014), showing also inconsistencies concerning preclinical AD. The ALFA study (Clinicaltrials.gov Identi- the impact of the BDNF Val66Met polymorphisms (Tsai fier: NCT01835717) entangles a cohort of 2,743 cognitively 2018). unimpaired participants, mostly adult children of patients The ε4 allele of apolipoprotein E (APOE) gene, the major with AD, and aged between 45 and 75 years. Cognitive sta- genetic risk factor for Alzheimer’s disease (AD) (Mueller tus was assessed at baseline as follows: Mini-Mental State and Weiner 2009), has also an impact on hippocampal sub- Examination (Folstein et al. 1975; Blesa et al. 2001) > 26, fields. APOE ε4-carriers have reduced volume of the subicu- Memory Impairment Screen (Buschke et al. 1999; Böhm lar/CA1 region in AD patients (Pievani et al. 2011), as well et al. 2005) > 6, Time-Orientation subtest of the Barcelona as in a pool of older adults that included healthy controls and Test II (Quinones-Ubeda 2009) > 68, semantic fluency patients with amnestic mild cognitive impairment (aMCI) (Ramier and Hecaen 1970; Peña-Casanova et al. 2009) and AD dementia, after controlling for the diagnostic group (animals) > 12 and Clinical Dementia Rating scale (Morris (Kerchner et al. 2014). In a recent report in CU partici- 1993) = 0. A subset of 430 participants from the ALFA study pants, we also showed that APOE-ε4 relates to significantly with available information on BDNF Val66Met polymor- reduced hippocampal tail in a gene dose-dependent manner phisms and APOE genotypes, as well as neuroimaging data (Cacciaglia et al. 2018a). (HSv) were included in this study (Fig. 1). The cognitive Moreover, recent evidence suggests that APOE geno- status of these participants was reviewed if cognitive test- types differentially affects the expression of BDNF through ing had not been conducted in the last 6 months. For this, the regulation of its maturation in human astrocytes and its mild cognitive impairment (MCI) was ruled out by clinical secretion (Sen, Nelson, and Alkon 2015). Astrocytes are judgment after interview and accounting for psychometric known to synthesise BDNF, and as brain APOE is primarily scores in the main variables of the Free and Cued Selec- produced by astrocytes, studying APOE and BDNF modula- tive Reminding Test [FCSRT] (Buschke et al. 2017). The tion becomes important. Specifically, interactions between study was conducted in accordance with the directives of 13 Brain Structure and Function (2020) 225:2331–2345 2333 heterozygosity 5 SD (N = 8 exclusions). Then, we checked ALFA Study sample N = 2,743 sex discordances (N = 4 exclusions). In total, we excluded 18 subjects (less than 2%). None of the individuals pre- No genetic data or data excluded after sented autosomal dominant mutations in APP, PSEN1, and applying quality control criteria (N = 1,805). PSEN2. The final genetic data set consisted of volunteers of European ethnic origin with available information regarding Genetic data BDNF Val66Met polymorphism and the APOE rs429358 N = 938 SNP: rs6265/Val66Met (BDNF) and rs7412 polymorphisms. Genotype and allele frequen- rs7412, rs429358 (APOE) cies of Val66Met, rs429358 and rs7412 polymorphisms were determined. Moreover, allele frequencies were inspected No neuroimaging data acquired/pre-processed for potential covariate-related differences. Departures from or data excluded after quality control criteria (N= 508). Hardy–Weinberg equilibrium were also examined (Ryck- man and Williams 2008). The APOE allelic variants were Neuroimaging data obtained from allelic combinations of the rs429358 and N = 430 rs7412 polymorphism (Radmanesh et al. 2014). According ROIS: Hippocampal subfield volumes (12 structures) and whole hippocampus to the genotypes of these polymorphisms, subjects were clas- sified depending on the number of ε4 alleles (non-carriers, one ε4 allele or two ε4 alleles). Fig. 1  Flow chart depicting the final sample size of the real appli- cation. Solid lines and boxes represent individuals remaining in the study. Dashed lines and boxes represent individuals excluded. Rea- Image acquisition and extraction of hippocampal son and number of individuals excluded is indicated in dashed boxes. subfield volumes SNP single nucleotide polymorphism, N size of the sample, ROIS brain regions of interest Scans were obtained with a 3 T scanner (Philips Ingenia CX). The MRI protocol was identical for all participants the Spanish Law 14/2007, of 3rd of July, on Biomedical and included high-resolution three-dimensional struc- Research (Ley 14/2007 de Investigación Biomédica). All tural images weighted in T1 with an isotropic voxel of participants accepted the study procedures by signing an 0.75 × 075 × 0.75 mm3. The acquisition parameters were informed consent form. A subset of 430 participants from TR/TE/TI = 9.9/4.6/900 ms, flip angle = 8° and a matrix the ALFA study with available information on BDNF Val- size of 320 × 320 × 240. Hippocampal subfields were seg- 66Met polymorphisms and APOE genotypes, as well as neu- mented using FreeSurfer version 6.0 (Iglesias et al. 2015). roimaging data (HSv) were included in this study (Fig. 1). We extracted raw volumes for 12 different HSv per hemi- sphere: the cornu ammonis region 1 (CA1), cornu ammonis Genotyping region 2/3 (CA3), cornu ammonis region 4 (CA4), dentate gyrus (DG), fimbria, hippocampal-amygdaloid transition DNA samples were obtained from whole blood samples by area (hata), tail, parasubiculum, presubiculum, subiculum, applying salting out protocol. DNA was eluted in 800 µl of fissure and molecular layer. The value of the subfields used H2O (milliQ) and quantified using Quant-iTT PicoGreen® as the outcomes of the study were calculated as the sum of dsDNA Assay Kit (Life Technologies). Integrity of DNA the regional value of each hemisphere ­(mm3). We visually was checked in a subset of samples by running a 1% aga- inspected the segmentation of the individuals included in rose gel. All the samples were within specification. Genome- the study (Fig. 2), and we removed outliers and/or abnormal wide genotyping was performed using the NeuroChip hippocampal subfields volume values. The whole hippocam- backbone(Blauwendraat et al. 2017), based on a genome- pal volume, as well as, total intracranial volume were also wide genotyping array (Infinium HumanCore-24 v1.0) con- calculated using Freesurfer (v. 6.0). taining 306,670 tagging variants and a custom content that has been updated and extended with 179,467 neurodegen- Statistical analysis erative disease-related variants at the Cancer Epigenetics and Biology Program (PEBC; IDIBELL). Previous step was Differences in demographic variables were tested using to normalize the quantity of DNA from each sample. The χ2 test and F test for gender, age, education, number of analysis was performed by the GenomeStudio (Illumina) APOE-ε4 carriers and total intracranial volume (TIV). The software using the genotyping module (standard analysis). additive, dominant, recessive, and codominant effects of the PLINK was used for genetic data quality control (Purcell BDNF Val66Met genotype on the hippocampal subfields et al. 2007). We applied the following sample quality con- volume were assessed using general linear models corrected trol thresholds: sample call rate > 97% (N = 6 exclusion) and by age, sex, years of education, number of APOE-ε4 alleles 13 2334 Brain Structure and Function (2020) 225:2331–2345 Fig. 2  T1 images of hippocampal segmentation and TIV. These covariates were selected based on previous assumes a common response to 1 or 2 copies of the Met associations reported using the ALFA study sample (Cac- allele. Finally, a recessive genetic model predicts a common ciaglia et al. 2018b). In brief, the genetic additive model response to 0 or 1 copies of the Met allele. predicts a linear increase of the phenotypic variable depend- The assumption of different genetic models was per- ing on the number of Met alleles, whereas the codominant formed to counteract a misspecification of the true underly- genetic model infers that the heterozygote mean differs from ing genetic model, which could have an adverse effect on both the homozygote means. The dominant genetic model the statistical power of an association, and on the effect size 13 Brain Structure and Function (2020) 225:2331–2345 2335 (Gaye and Davis 2017). The goodness-of-fit of each genetic subfield volumes analyzed in the study by BDNF genotype. model was evaluated based on the Akaike information cri- All morphometric subfield measures were normally distrib- terion (AIC), for which lower numerical values indicate a uted (Kolmogorov Smirnov test, FDR > 0.05) and their vari- better fit of the model (Akaike 1998). ances were homogenous (Levene’s test, FDR > 0.05). Figure We also investigated whether the association between S1 shows the pattern of correlation (Pearson correlation sta- BDNF Val66Met and hippocampal subfield volumes was tistics) among all subfields included in the study. Hippocam- modified by the number of APOE-ε4 alleles, with a sec- pal subfield structures present high correlation among them ond model that included an interaction term between BDNF (r > 0.8) (i.e., structural covariance). Val66Met polymorphism and the number of ε4 alleles, co- varying for age, sex, years of education and total intracra- Effect of BDNF Val66Met polymorphism nial volume potential confounders. In this model, dominant on hippocampal subfields genetic effects were assumed for Val66Met polymorphism and additive genetic effects for APOE-ε4 alleles. General linear models revealed that Met carriers showed Moreover, in post-hoc analyses, we evaluated the effects statistically significant larger bilateral volumes of the sub- of BDNF Val66Met and APOE-ε4 status on cognitive iculum under the dominant model ( 𝛽dom = 2.53%, pFDRdom performance. = 3 × 10−3) (Table 4 and Fig. 3). For subiculum subfield, Statistical significance was set at False Discovery Rate an additive genetic model obtained the lowest AICs score (FDR) corrected p value < 0.05, and all statistical analyses (AIC = 4890.43), indicating that this model is the most par- and data visualization were carried out using R version simonious model for this subfield structure. Moreover, we 3.4.4. found statistically significant larger bilateral volumes of the subiculum under the additive genetic model (𝛽add = 2.39%, pFDRadd = 0.013) (Table S1). No significant results after Results FDR-correction were found under recessive and codominant genetic models. In addition, nominal significant results with- Demographic characteristics out FDR adjustment (p < 0.05) showed larger bilateral vol- umes of the molecular layer of the hippocampus (β: 1.55%, Descriptive data of the demographic and BDNF Val66Met p: 0.007), presubiculum (β: 1.74%, p: 0.041), and whole hip- polymorphism information are presented in Tables 1 and pocampal volume (β: 1.46%, p: 0.025) for Met carriers under 2. The mean age of the population was 57.1 ± 5.7 years the dominant genetic model. Results of all adjusted genetic old, with 61.4% women. The BDNF Val66Met genotype models for each HSv can be found in Table S1. groups did not significantly differ in the distribution of gender (χ2[2] = 0.55, p: 0.51), number of APOE-ε4 alleles Effect of the interaction between APOE‑ε4 and BDNF (χ2[2] = 6.13, p = 0.19), age (F[2,427] = 0.67, p: 0.76), years Val66Met on hippocampal subfields of education (F[2,247] = 0.107, p: 0.9), or total intracranial volume (TIV) (F[2,427] = 1.66, p: 0.19). The distribution of As expected, APOE ε4 allele was associated with lower BDNF Val66Met and APOE rs429358 and rs7412 polymor- bilateral volumes of the hippocampal subfields, even though phisms did not deviate from Hardy–Weinberg equilibrium on a trend-level (Table S2). Interestingly, when this associa- (χ2[1] = 0.42, p: 0.51). Table 3 summarizes the hippocampal tion was studied according to BDNF Val66Met genotypes, Table 1  Characteristics of the study according to rs6265 (Val/Met) status ValVal carriers ValMet carriers MetMet carriers Total (n = 430) Statistic p (n = 247) (n = 161) (n = 22) Age (m ± SD; years) 57.31 (5.63) 56.93 (5.84) 55.97 (5.82) 57.1 (5.72) F (2.427) = 0.67 0.76 Sex (female), n (%) 154 (62.35%) 98 (60.87%) 12 (54.55%) 264 (61.4%) Chi (2) = 0.549 0.512 Education (m ± SD; 13.87 (± 3.53) 14.03 (± 3.53) 13.95 (± 3.5) 13.93 (± 3.52) F (2.427) = 0.107 0.899 years) Number of APOE-ε4 0: 160 (64.78%); 1: 0: 89 (55.28%); 1: 0: 12 (54.55%); 1: 0: 261 (60.7%); 1: Chi (2) = 6.129 0.19 alleles, n (%) 75 (30.36%); 2: 12 59 (36.65%); 2: 13 7 (31.82%); 2: 3 141 (32.79%); 2: 28 (4.86%) (8.07%) (13.64%) (6.51%) TIV (m ± SD; ­cm3) 1442.91 (± 177.13) 1453.1 (± 163.66) 1511.62 (± 163.28) 1450.24 (± 171.79) F (2.427) = 1.656 0.192 Mean and SD are shown for continuous variables n sample size, m mean, SD standard deviation, TIV total intracranial volume P p value 13 2336 Brain Structure and Function (2020) 225:2331–2345 we observed a significant interaction with the presence of SNP single nucleotide polymorphisms, BP base position, A1 major allele, A2 minor allele, MAF minor allele frequency, MAF gp* MAF general population. Source: gnomAD genome aggrega- Array source** Blauwendraat et al., NeuroChip, an updated version of the NeuroX genotyping platform to rapidly screen for variants associated with neurological diseases. Neurobiology of Met alleles (p < 0.05) (Table 5). APOE-ε4 homozygotes ValVal: 247 / ValMet: 161 / MetMet: 22 0.51 Neurochip backbone (Infinium Human- carrying at least one Met allele presented nominally sig- nificant larger bilateral volumes of the CA4 (β: 7.25%, p: 0.016) and DG (β: 7.38%, p: 0.012) subfields, hippocampal tail (β: 8.33%, p: 0.05), and whole hippocampal volumes (β: 5.35%, p: 0.046) than the expected combined effect of the individual contribution of APOE-ε4 (reverse effect) and BDNF Val66Met (Fig. 4). Moreover, even though the results Core-24 v1.0) for the remainder hippocampal subfields were statistically not significant, changes on hippocampal subfields volumes HWE Array** follow the same general patterns (Figure S2). Post‑hoc analyses: Effect of BDNF Val66Met ε4-non carriers: 261; ε4 heterozygous: 0.137 and APOE‑ε4 on cognitive performance The post hoc analyses, although not significant, suggested better cognitive performance patterns for Met carriers in most FCSRT domains, and in depression scores of the Hos- 141; ε4 homozygous: 28 pital Anxiety and Depression Scale (HADS) (Table 6). In addition, APOE-ε4 status did not significantly influence Allele 1 Allele 2 MAF MAF gp* Genotype distribution Allele ε4 distribution the effects of BDNF Val66Met genotypes on cognitive performance. Discussion 27,658,369 (CRCh37) C [Val] T [Met] 0.238 0.19,437 To the best of our knowledge, this is the first study to show a 0.214 0.138 0.043 0.061 phenotypic effect of the BDNF Val66Met polymorphism in the hippocampal subfields of cognitively unimpaired (CU) individuals. Moreover, the present study is also the first in CU to find an effect modification by BDNF Val66Met poly- morphism of associations between APOE-ε4 status and hip- pocampal subfield volumes. C T Table 2  Characteristics of BDNF Val66Met and APOE polymorphisms We first found significantly larger bilateral subiculum vol- umes in CU middle-aged/late-middle-aged BDNF Val66Met carriers in a dose-dependent manner. The direction of the 45,412,079 (CRCh37) C 45,411,941 (CRCh37) T effects is consistent across different subfields and the entire hippocampal formation, as shown by the nominally signifi- cant difference between BDNF Val66Met carriers and non- tion database, HWE Hardy weinberg equilibrium carriers involving the molecular layer of the hippocampus, CHR position as well as the pre-subiculum and the whole hippocampus. Given that BDNF Val66Met polymorphism has been Aging. 2017 vol: 57 pp: 247.e9-247.e13 related to impaired hippocampal long-term potentiation which underlies learning and memory (Spriggs et al. BDNF Val66Met (rs6265) 11 19 19 2018), our results may underline compensatory mecha- nisms in the Met-carriers to achieve normative episodic recall, which is highly specialized in the subiculum (Eldridge et al. 2005; Suthana et al. 2015). However, APOE rs429358 rs7412 although most studies showed that large hippocampal vol- SNP umes lead to better memory performance and may protect from dementia (Pohlack et al. 2014; Whitwell 2010; Erten- Gene Lyons et al. 2009), the impact of hippocampal volume on 13 Table 3  Characteristics of hippocampal subfield volumes Hippocampal subfield ValVal carriers (n = 247) ValMet carriers (n = 161) MetMet carriers (n = 22) Total (n = 430) Mean (SD) Min Median Max Mean (SD) Min Median Max Mean (SD) Min Median Max Mean (SD) Min Median Max Brain Structure and Function (2020) 225:2331–2345 CA1, ­mm3 1193 (123) 920 1176 1595 1214 (119) 950 1207 1518 1255 (161) 924 1250 1630 1204 (124) 920 1192 1630 CA3, ­mm3 361 (42) 241 357 511 364 (40) 273 361 475 385 (62) 299 375 525 364 (43) 241 358 525 CA4, ­mm3 454 (44) 299 448 594 459 (38) 374 458 564 480 (59) 377 464 619 457 (43) 299 455 619 GC-ML-DG, ­mm3 535 (52) 351 530 706 541 (45) 432 535 656 566 (67) 450 553 724 539 (51) 351 533 724 Subiculum, ­mm3 802 (91) 594 804 1113 822 (86) 639 816 1067 866 (104) 688 859 1061 812 (91) 594 809 1113 Presubiculum, ­mm3 601 (69) 412 599 797 614 (60) 482 616 786 630 (75) 528 616 781 608 (66) 412 604 797 Parasubiculum, ­mm3 125 (18) 83 123 194 127 (17) 92 125 175 131 (21) 94 127 194 126 (18) 83 125 194 Hippocampal fissure, m­ m3 339 (45) 214 338 514 345 (40) 255 346 448 360 (50) 293 348 457 343 (44) 214 341 514 Hippocampal tail, m­ m3 1068 (126) 767 1058 1388 1086 (128) 808 1087 1488 1080 (133) 787 1069 1328 1076 (127) 767 1066 1488 3 Fimbria, ­mm 174 (30) 108 173 308 175 (30) 105 175 267 182 (29) 132 181 241 175 (30) 105 174 308 Hata, ­mm3 117 (14) 72 116 179 118 (13) 83 117 156 120 (16) 85 119 147 118 (14) 72 117 179 Molecular layer, ­mm3 1059 (99) 767 1052 1311 1077 (93) 889 1070 1328 1121 (122) 910 1110 1404 1069 (99) 767 1061 1404 whole hippocampus, ­mm3 6489 (597) 4623 6435 8235 6598 (552) 5407 6567 8014 6817 (747) 5351 6793 8582 6547 (594) 4623 6492 8582 Means, standard deviations (SD), Median, and ranges values are shown Segmentation of hippocampal subfields performed with FreeSurfer version 6.0 image analysis suite CA1 cornu ammonis region 1, CA3 cornu ammonis region 23, CA4 cornu ammonis region 4, GC-ML-DG granule cells in the molecular layer of the dentate gyrus, hata hippocampal-amygda- loid transition region, HP hippocampus, CI95 confidence interval, FDR95 false discovery rate corrected p value < 0.05 13 2337 2338 Brain Structure and Function (2020) 225:2331–2345 Table 4  Main effects of Val66Met genotype on hippocampal subfields (­ mm3) Hippocampal subfield Best genetic model Effect ­(mm3) CI 95% Effect (%) p value FDR95% AIC CA1 Dominant 18.575 (− 0.67, 37.82) 1.54% 0.059 0.531 5188.564 CA3 Recessive 14.291 (− 1.43, 30.01) 3.93% 0.076 0.684 4321.049 CA4 Recessive 14.04 (− 0.69, 28.77) 3.07% 0.062 0.671 4264.9 GC-ML-DG Recessive 16.756 (− 0.19, 33.71) 3.11% 0.053 0.636 4385.858 Subiculum Additive 19.435 (8.06, 30.81) 2.39% 0.001 0.013* 4890.434 Presubiculum Dominant 10.556 (0.44, 20.67) 1.74% 0.041 0.41 4635.326 Parasubiculum Dominant 1.845 (− 1.06, 4.75) 1.46% 0.214 1 3562.288 Hippocampal fissure Dominant 5.226 (− 2.05, 12.5) 1.52% 0.16 1 4351.502 Hippoccampal tail Dominant 12.37 (− 9.43, 34.17) 1.15% 0.267 1 5295.52 Fimbria Recessive 1.601 (− 10.09, 13.29) 0.91% 0.789 1 4065.996 Hata Dominant 0.333 (− 1.95, 2.61) 0.28% 0.775 1 3354.183 Molecular layer Additive 16.578 (4.59, 28.57) 1.55% 0.007 0.084 4937.902 Whole hippocampus Dominant 95.904 (12.06, 179.75) 1.46% 0.025 0.275 6454.188 All models were adjusted by sex, years of education, number of APOE-ε4 allele and total intracranial volume CA1 cornu ammonis region 1, CA3 cornu ammonis region 23, CA4 cornu ammonis region 4, GC-ML-DG granule cells in the molecular layer of the dentate gyrus, hata hippocampal-amygdaloid transition region, HP hippocampus, CI95 confidence interval, FDR95 false discovery rate cor- rected p value < 0.05. AIC akaike information criterion cognitive performance in middle-aged CU individual’s deleterious effect of the APOE-ε4 genotype in the age range remains controversial. For instance, smaller hippocampal of the studied sample. volumes have been related to better episodic memory, due Another possible explanation to this finding could raise to efficient synaptic pruning (Van Petten 2004). Thus, our from an interaction of the BDNF Val66Met polymorphism results could suggest a moderating role of BDNF in the with pathological markers of AD, as the APOE-ε4 allele neurobiology of hippocampal subfields, which may stress has also been strongly linked to a dose-dependent increase the importance to consider the hippocampal formation at in the prevalence of abnormally elevated cerebral amy- the subfield level to disentangle potential opposite effects loid deposition in CU individuals (Reiman et al. 2009). leading to the aforementioned conflicting results. In addi- By the mean age of our APOE-ε4 homozygote group tion, we cannot rule out that the BDNF Val66Met poly- (56.62 ± 5.71y), about half of them are expected to display morphism may differentially influence the morphology of abnormally high amyloid levels (Jansen et al. 2015). How- other brain areas. This calls for additional whole-brain ever, previous longitudinal reports on amyloid-positive voxel-wise studies addressing distinct genetic models of CU individuals have described that the BDNF Val66Met penetrance of BDNF Val66Met. allele was associated with a steeper decline in cognitive Second, we also observed that Met-carriers compensate function and hippocampal atrophy (Yen Ying Lim et al. for the deleterious impact of the number of APOE-ε4 alleles 2013). Moreover, this deleterious effect is more severe in on hippocampal subfield volumes. As expected, we observed APOE-ε4 carriers (Lim et al. 2015). These studies, how- that APOE-ε4 homozygotes showed a tendency towards dis- ever, were performed in significantly older cohorts (aver- playing reduced volumes of the subiculum and hippocampal age age of 70y) than that in our work. Another recent study tails, in accordance with previous reports (Kerchner et al. performed in subjects with an age range similar to ours 2014; Cacciaglia et al. 2018a; Pievani et al. 2011). These (55y) confirmed this longitudinal pattern of decline in individuals are at increased higher risk (× 15) to develop cognition, particularly in amyloid-positive CU individu- AD as compared to APOE-ε4 non-carriers. The lower hip- als (Boots et al. 2017). However, in this work, Boots et al. pocampal volumes in APOE-ε4 carriers are often inter- also reported that, at baseline, Met carriers showed a sig- preted as brain marker that confers vulnerability towards nificantly better cognitive performance (verbal learning developing the clinical picture of AD. Strikingly, we found and memory, speed and flexibility, and working memory), that APOE-ε4 homozygotes who were also Met-carriers even if amyloid positive. Similar patterns of effect were countered the effect of the APOE genotype and presented observed in our sample. Specifically, we found that Met HSv within the ranges expected for APOE-ε4 non-carriers, carriers suggested patterns of better cognitive performance particularly in the CA4, GC-ML-DG and the hippocampal on the Free and Cued Selective Reminding Test (FCSRT) tails. It could be argued that Met-carriers can counter the domains, and on the Depression score of the Hospital 13 Brain Structure and Function (2020) 225:2331–2345 2339 Fig. 3  Box plot of change in hippocampal subfield volumes between *Significant difference between groups at a nominal level (p < 0.05). BDNF Val66Met (rs6265) genotypes under additive, dominant, reces- ***Significant difference between groups after multiple comparison sive and codominant genetic models. Middle line in box represents correction (FDR < 0.05). CA1 cornu ammonis region 1, CA3 cornu the median; lower box bounds the first quartile; upper box bounds ammonis region 3, CA4 cornu ammonis region 4, GC-ML-DG gran- the 3rd quartile. Whiskers represent the 95% confidence interval of ule cells in the molecular layer of the dentate gyrus, hata hippocam- the mean. Open circles are outliers from 95% confidence interval. pal-amygdaloid transition region, HP hippocampus Anxiety and Depression Scale (HADS). Although these a subset of our participants will undergo a lumbar puncture differences were not statistically significant, one poten- to assess cerebrospinal fluid levels of core AD biomarkers tial explanation to reconcile our findings with the existing (Aβ42, total Tau, and phosphorylated Tau). literature would be that the Met genotype might provide A strong feature of our study that may sustain our ability a limited beneficial effect during middle age. When no to detect a significant effect of Val66Met genotype (and its longer capable of compensating for the deleterious down- interaction with APOE) on hippocampal subfields is that stream effects of amyloid accumulation, then Met carri- the studied cohort presents a higher prevalence of BDNF ers would experience faster hippocampal atrophy and a Met and APOE-ε4 homozygotes compared with previous steeper decline in cognition. Neither shall it be excluded studies. While most of the studies reported allele frequen- that APOE-ε4 and Met carriers could be the most vulner- cies between 0% (studies without MetMet carriers) to 6% able to an inflammatory response at the beginning of the (Harrisberger et al. 2015), the minor allele frequency in our amyloid-pathology. Nevertheless, the current unavailabil- study achieves 23%, which is even higher to the population ity of core AD biomarkers and the cross-sectional nature frequency (15–19%, Source: Genome Aggregation Data- of this study constitute a limitation for the full inter- base https​://gnoma​d.broad​insti​tute.org/varia​nt/11-27679​ pretation of the interaction between the BDNF Met and 916-C-T). Similarly, the high number of APOE-ε4 carriers APOE-ε4 genotypes. Nevertheless, this will be mitigated in the ALFA participants compared to the general popula- in the longitudinal follow-up of the cohort here studied, as tion (19% vs. 14%, respectively; p < 0.001) has allowed us to 13 2340 Brain Structure and Function (2020) 225:2331–2345 Table 5  Interaction Effects Hippocampal subfield Dominant genetic model Between number of APOE-ε4 alleles (under additive Effect ­(mm3) CI 95% Effect (%) p value FDR95% model) and BDNF Val66Met polymorphism (under dominant CA1 − 10.788 (− 52.36, 30.79) − 0.90% 0.611 1 genetic model) on hippocampal 55.535 (− 23.33, 134.4) 4.61% 0.168 1 subfields CA3 10.889 (− 4.31, 26.09) 2.99% 0.161 1 27.425 (− 1.4, 56.25) 7.53% 0.063 0.567 CA4 7.797 (− 6.39, 21.98) 1.71% 0.282 1 33.155 (6.25, 60.06) 7.25% 0.016 0.192 GC-ML-DG 7.94 (− 8.39, 24.27) 1.47% 0.341 1 39.772 (8.79, 70.75) 7.38% 0.012 0.156 Subiculum − 13.434 (− 42.9, 16.03) − 1.65% 0.372 1 25.529 (− 30.36, 81.42) 3.14% 0.371 1 Presubiculum 0.013 (− 21.92, 21.94) ~ 0% 0.999 1 13.726 (− 27.87, 55.33) 2.26% 0.518 1 Parasubiculum 2.587 (− 3.7, 8.88) 2.05% 0.421 1 2.78 (− 9.15, 14.71) 2.21% 0.648 1 Hippocampal fissure − 1.375 (− 17.1, 14.35) − 0.40% 0.864 1 7.756 (− 22.08, 37.59) 2.26% 0.611 1 Hippoccampal tail 21.561 (− 25.48, 68.6) 2.00% 0.369 1 89.676 (0.45, 178.9) 8.33% 0.05 0.506 Fimbria − 7.8 (− 19.07, 3.47) − 4.46% 0.176 1 8.845 (− 12.54, 30.23) 5.05% 0.418 1 Hata − 0.304 (− 5.25, 4.64) − 0.26% 0.904 1 4.091 (− 5.29, 13.47) 3.47% 0.393 1 Molecular layer − 2.31 (− 33.37, 28.75) − 0.22% 0.884 1 49.911 (− 9.01, 108.83) 4.67% 0.098 0.784 Whole hippocampus 16.151 (− 164.67, 196.97) 0.25% 0.861 1 350.445 (7.43, 693.46) 5.35% 0.046 0.506 All models were adjusted by sex, years of education, age and total intracranial volume CA1 cornu ammonis region 1, CA3 cornu ammonis region 23 CA4 cornu ammonis region 4, GC-ML-DG granule cells in the molecular layer of the dentate gyrus, hata hippocampal-amygdaloid transition region, HP hippocampus, CI95 confidence interval, FDR95 FDR corrected p value < 0.05 disentangle specific effects in ε4 homozygotes, while most subfield volumes and compared them to whole hippocampal studies pool them with heterozygotes in a single APOE-ε4 volumetry, even though the independent genetic variation carrier group. In our study, the high prevalence of these less specific to hippocampal subfields (Elman et al. 2019). Thus, frequent genotypes has allowed us to achieve a relatively our analyses based on hippocampal subfields increased the higher inferential power, allowing for testing additive, reces- sensitivity of the results, which make our study more robust sive and codominant genetic effects, as well as, gene–gene and consistent than previous ones. interactions. Finally, in contrast to previous studies in which the diag- Another substantial strength is the use of a high-reso- nostic value of hippocampal subfield volumes related to lution T1 scan as compared to previous studies, combined Val66Met polymorphism was only assessed by comparing with the use of the most recent version (v.6.0) of the hip- patients with psychiatric disorders, our study includes CU pocampal subfield segmentation toolbox in Freesurfer, middle-aged/late-middle-age participants. This is also a rel- which overcomes significant shortcomings of previous ver- evant strength because the misrepresentation of the general sions (Iglesias et al. 2015; Wisse et al. 2014; Mueller et al. population could constitute a bias in the assessment of the 2018). Thus, subfield volumes available for this analysis are diagnostic utility of hippocampal subfield volumes, due to of substantially better quality, which, combined with a con- potential etiologies associated with neurodegenerative pro- siderably higher sample size, has allowed us to achieve a cesses (de Flores et al. 2015). significantly superior statistical power than previous reports. Altogether, our findings suggest that the BDNF Met allele Moreover, only a few studies have assessed hippocampal might confer a time-limited resilience, which protects the 13 Brain Structure and Function (2020) 225:2331–2345 2341 Fig. 4  Differences according BDNF Val66Met genotypes in associa- gyrus (GC-ML-DG) subfield volume, c hippocampal tail subfield vol- tions between APOE-ε4 and a Cornu ammonis region 4 (CA4) sub- ume, and d whole hippocampal volume field volume, b Granule cells in the molecular layer of the dentate 13 2342 Brain Structure and Function (2020) 225:2331–2345 Table 6  Main and interaction effects of BDNF-Val66Met genotype Plan (PERIS) 2016–2020 grant# SLT002/16/00201). CM was sup- and APOE-ε4 status (dominant model) on cognitive performance ported by the Spanish Ministry of Economy and Competitiveness (HADS, FCSRT tests) (grant n° IEDI-2016-00690). MSC received funding from the Euro- pean Union’s Horizon 2020 Research and Innovation Program under Cognition (Test) Effect CI 95% p value FDR-95% the Marie Sklodowska-Curie action grant agreement No 752310. H.H.H.A. was supported by ZonMW grant numbers 916.19.15 and HADS-anxiety 916.19.151. J.D.G. holds a ‘Ramón y Cajal’ fellowship (RYC-2013- Val66Met − 0.040 (− 0.29, 0.21) 0.756 13054). This publication is part of the ALFA (ALzheimer and FAmi- APOE-ε4 − 0.188 (− 0.59, 0.22) 0.364 lies) study. The authors would like to express their most sincere grati- Val66Met x APOE-ε4 0.243 1 tude to the ALFA project participants, without whom this research would have not been possible. With recognition and heartfelt gratitude HADS-depression 0.318 to Mrs. Blanca Brillas for her outstanding and continued support to Val66Met 0.036 (− 0.43, 0.5) 0.879 the Pasqual Maragall Foundation to make possible a Future without APOE-ε4 0.32 (− 0.36, 0.99) 0.354 Alzheimer´s. Collaborators of the ALFA study are: Eider M. Arenaza- Val66Met x APOE-ε4 0.318 1 Urquijo, Annabella Beteta, Anna Brugulat-Serrat, Alba Cañas, Carme Deulofeu, Ruth Dominguez, Maria Emilio, Karine Fauria, Sherezade Delayed recall (DR) Fuentes, Laura Hernandez, Gema Huesa, Jordi Huguet, Paula Marne, Val66Met 0.143 (− 0.29, 0.57) 0,517 Tania Menchón, Albina Polo, Sandra Pradas, Aleix Sala-Vila, Gonzalo APOE-ε4 0.067 (− 0.61, 0.74) 0.847 Sánchez-Benavides, Anna Soteras, Gemma Salvadó, Mahnaz Shekari, Val66Met x APOE-ε4 0.85 1 Marc Vilanova. Total recall (TR) Val66Met 0.111 (− 1.22, 1.44) 0.87 Compliance with ethical standards APOE-ε4 1.398 (− 0.67, 3.47) 0.188 Conflict of interest JLM has served/serves as a consultant or at adviso- Val66Met x APOE-ε4 0.45 1 ry boards for the following for-profit companies, or has given lectures Delayed free recall (DFR) in symposia sponsored by the following for-profit companies: Roche Val66Met − 0.513 (− 1.30, 0.27) 0.204 Diagnostics, Genentech, Novartis, Lundbeck, Oryzon, Biogen, Lilly, APOE-ε4 0.528 (− 0.69, 1.75) 0.4 Janssen, Green Valley, MSD, Eisai, Alector, BioCross, GE Healthcare, ProMIS Neurosciences, NovoNordisk, Zambón, Cytox and Nutricia. Val66Met x APOE-ε4 0.722 1 The rest of the authors have no conflict of interest to declare. Free recall (FR) Val66Met 2.099 (− 2.4, 6.59) 0.387 Open Access This article is licensed under a Creative Commons Attri- APOE-ε4 − 0.853 (− 6.367, 4.662) 0.77 bution 4.0 International License, which permits use, sharing, adapta- Val66Met x APOE-ε4 0.385 1 tion, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, Retention index provide a link to the Creative Commons licence, and indicate if changes Val66Met − 0.029 (− 0.091, 0.033) 0.391 were made. The images or other third party material in this article are APOE-ε4 − 0.008 (− 0.085, 0.068) 0.837 included in the article’s Creative Commons licence, unless indicated Val66Met x APOE-ε4 0.948 1 otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not All models were adjusted by sex, years of education, and total intrac- permitted by statutory regulation or exceeds the permitted use, you will ranial volume need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creat​iveco​mmons​.org/licen​ses/by/4.0/. CI95 confidence interval, FDR95 FDR corrected p value < 0.05, HADS hospital anxiety and depression scale (HADS), HADS-Anxiety anxiety score of the HADS, HADS-Depression depression score of the HADS, FCSRT free and cued selective reminding test References hippocampi from the downstream deleterious effects of age- Aas M, Haukvik UK, Djurovic S, Tesli M, Athanasiu L, Bjella T, Hans- ing and/or amyloid accumulation, thus mediating the risk son L et al (2014) Interplay between childhood trauma and bdnf effect of APOE-ε4. Hence, these results prompt us to further val66met variants on blood bdnf mrna levels and on hippocampus subfields volumes in schizophrenia spectrum and bipolar disor- explore hippocampal atrophy rates and cognitive trajectories ders. J Psychiatr Res 59:14–21. https​://doi.org/10.1016/j.jpsyc​ of BDNF Met carriers compared to Val homozygotes, also as hires​.2014.08.011 a function of their APOE genotype and long-term accumula- Akaike H (1998) Information theory and an extension of the tion of amyloid beta. maximum likelihood principle. Springer NY. https ​ : //doi. org/10.1007/978-1-4612-1694-0_15 Álvarez A, Aleixandre M, Linares C, Masliah E, Moessler H (2014) Acknowledgements NV-T is funded by a post-doctoral grant, Juan Apathy and APOE4 are associated with reduced BDNF levels in de la Cierva Programme (FJC2018-038085-I), Ministerio de Ciencia, Alzheimer’s disease. J Alzheimer’s Dis 42(4):1347–1355. https​ Innovación y Universidades – Spanish State Research Agency. The ://doi.org/10.3233/JAD-14084​9 research leading to these results has received funding from “la Caixa” Foundation (LCF/PR/GN17/10300004) and the Health Department of the Catalan Government (Health Research and Innovation Strategic 13 Brain Structure and Function (2020) 225:2331–2345 2343 Autry AE, Monteggia LM (2012) Brain-derived neurotrophic factor Elman JA, Panizzon MS, Gillespie NA, Hagler DJ, Fennema-Notes- and neuropsychiatric disorders. Pharmacol Rev 64(2):238–258. tine C, Eyler LT, McEvoy LK et al (2019) Genetic architecture https​://doi.org/10.1124/pr.111.00510​8 of hippocampal subfields on standard resolution MRI: how the Bathina S, Das UN (2015) Brain-derived neurotrophic factor and its parts relate to the whole. Hum Brain Mapp 40(5):1528–1540. clinical implications. Arch Med Sci AMS 11(6):1164–1178. https​ https​://doi.org/10.1002/hbm.24464​ ://doi.org/10.5114/aoms.2015.56342​ Erten-Lyons D, Woltjer RL, Dodge H, Nixon R, Vorobik R, Calvert Blauwendraat C, Faghri F, Pihlstrom L, Geiger JT, Elbaz A, Lesage JF, Leahy M, Montine T, Kaye J (2009) Factors associated with S, Corvol J-C et al (2017) NeuroChip, an updated version of the resistance to dementia despite high Alzheimer disease pathol- NeuroX genotyping platform to rapidly screen for variants associ- ogy. Neurology 72(4):354–360. https​: //doi.org/10.1212/01. ated with neurological diseases. Neurobiol Aging 57:247.e9–247. wnl.00003​41273​.18141​.64 e13. https:​ //doi.org/10.1016/J.NEUROB ​ IOLAG ​ ING.2017.05.009 Ezzati A, Zimmerman ME, Katz MJ, Sundermann EE, Smith JL, Blesa R, Pujol M, Aguilar M, Santacruz P, Bertran-Serra I, Hernán- Lipton ML, Lipton RB (2014) Hippocampal subfields differ- dez G, Sol JM et al (2001) Clinical validity of the ‘mini-mental entially correlate with chronic pain in older adults. Brain Res state’ for spanish speaking communities. Neuropsychologia 1573:54–62. https​://doi.org/10.1016/j.brain​res.2014.05.025 39(11):1150–1157. https:​ //doi.org/10.1016/S0028-​ 3932(01)00055​ Folstein MF, Folstein SE, McHugh PR (1975) ‘Mini-mental state’. -0 A practical method for grading the cognitive state of patients Böhm P, Peña-Casanova J, Gramunt N, Manero RM, Terrón C, for the clinician. J Psychiatr Res 12(3):189–198. https​://doi. Quiñones-Ubeda S (2005) Versión española del Memory Impair- org/10.1016/0022-3956(75)90026​-6 ment Screen (MIS): datos normativos y de validez discriminativa Franzmeier N, Ren J, Damm A, Monté-Rubio G, Boada M, Ruiz [Spanish version of the Memory Impairment Screen (MIS): nor- A, Ramirez A et al (2019) The BDNFVal66Met SNP modu- mative data and discriminant validity]. Neurologia 20(8):402-411 lates the association between beta-amyloid and hippocampal Boots EA, Schultz SA, Clark LR, Racine AM, Darst BF, Koscik disconnection in Alzheimer’s disease. Mol Psychiatry. https​:// RL, Carlsson CM et al (2017) BDNF Val66Met predicts cogni- doi.org/10.1038/s4138​0-019-0404-6 tive decline in the wisconsin registry for Alzheimer’s preven- Frodl T, Skokauskas N, Frey E-M, Morris D, Gill M, Carballedo tion. Neurology 88(22):2098–2106. https​: //doi.org/10.1212/ A (2014) BDNFVal66Met genotype interacts with childhood WNL.00000​00000​00398​0 adversity and influences the formation of hippocampal subfields. Buschke H, Kuslansky G, Katz M, Stewart WF, Sliwinski MJ, Eck- Hum Brain Mapp 35(12):5776–5783. https​://doi.org/10.1002/ holdt HM, Lipton RB (1999) Screening for dementia with the hbm.22584​ memory impairment screen. Neurology 52(2):231–238. https​:// Gaye A, Davis SK (2017) Genetic model misspecification in genetic doi.org/10.1212/wnl.52.2.231 association studies. BMC Res Notes 10(1):569. https​: //doi. Buschke H, Mowrey WB, Ramratan WS et al (2017) Memory bind- org/10.1186/s1310​4-017-2911-3 ing test distinguishes amnestic mild cognitive impairment and Gomar JJ, Conejero-Goldberg C, Huey ED, Davies P, Goldberg TE dementia from cognitively normal elderly [published correction (2016) Lack of neural compensatory mechanisms of bdnf val- appears in Arch Clin Neuropsychol. 2017 Dec 1;32(8):1037- 66met met carriers and APOE E4 carriers in healthy aging, 1038]. Arch Clin Neuropsychol 32(1):29-39. https ​ : //doi. mild cognitive impairment, and Alzheimer’s disease. Neuro- org/10.1093/arcli​n/acw08​3 biol Aging 39:165–173. https​://doi.org/10.1016/j.neuro​biola​ Cacciaglia R, Molinuevo JL, Falcón C, Brugulat-Serrat A, Sánchez- ging.2015.12.004 Benavides G, Gramunt N, Esteller M et al (2018a) Effects of Hariri AR, Goldberg TE, Mattay VS et al (2003) Brain-derived neu- APOE -Ε4 allele load on brain morphology in a cohort of rotrophic factor val66met polymorphism affects human memory- middle-aged healthy individuals with enriched genetic risk for related hippocampal activity and predicts memory performance. Alzheimer’s disease. Alzheimer’s Dement 14(7):902–912. https​ J Neurosci 23(17):6690–6694. https​://doi.org/10.1523/JNEUR​ ://doi.org/10.1016/j.jalz.2018.01.016 OSCI.23-17-06690​.2003 Cacciaglia R, Molinuevo JL, Sánchez-Benavides G, Falcón C, Gra- Harrisberger F, Spalek K, Smieskova R, Schmidt A, Coynel D, Milnik munt N, Brugulat-Serrat A, Grau O, Gispert J, ALFA Study A, Fastenrath M et al (2014) The association of the BDNF Val- (2018b) Episodic memory and executive functions in cogni- 66Met polymorphism and the hippocampal volumes in healthy tively healthy individuals display distinct neuroanatomical cor- humans: a joint meta-analysis of published and new data. Neu- relates which are differentially modulated by aging. Hum Brain rosci Biobehav Rev 42:267–278. https​://doi.org/10.1016/j.neubi​ Mapp 39(11):4565–4579. https​://doi.org/10.1002/hbm.24306​ orev.2014.03.011 Cao Bo, Bauer IE, Sharma AN, Mwangi B, Frazier T, Lavagnino L, Harrisberger F, Smieskova R, Schmidt A, Lenz C, Walter A, Wittfeld Zunta-Soares GB et al (2016) Reduced hippocampus volume K, Grabe HJ, Lang UE, Fusar-Poli P, Borgwardt S (2015) BDNF and memory performance in bipolar disorder patients carrying Val66Met polymorphism and hippocampal volume in neuropsy- the BDNF Val66met met allele. J Affect Disord 198(July):198– chiatric disorders: a systematic review and meta-analysis. Neu- 205. https​://doi.org/10.1016/j.jad.2016.03.044 rosci Biobehav Rev 55:107–118. https​://doi.org/10.1016/j.neubi​ de Flores R, La Joie R, Chételat G (2015) Structural imaging of hip- orev.2015.04.017 pocampal subfields in healthy aging and Alzheimer’s disease. Hett K, Ta VT, Catheline G, Tourdias T, Manjon J, Coupe P (2018) Neuroscience 309:29–50. https​://doi.org/10.1016/j.neuro​scien​ Multimodal hippocampal subfield grading for Alzheimer’s disease ce.2015.08.033 classification. BioRxiv. https​://doi.org/10.1101/29312​6(293126) Egan MF, Kojima M, Callicott JH et al (2003) The BDNF val- Hibar DP, Adams HHH, Jahanshad N, Chauhan G, Stein JL, Hofer E, 66met polymorphism affects activity-dependent secretion of Renteria ME et al (2017) Novel genetic loci associated with hip- BDNF and human memory and hippocampal function. Cell pocampal volume. Nat Commun 8:13624. https:​ //doi.org/10.1038/ 112(2):257–269. https​://doi.org/10.1016/s0092​-8674(03)00035​ ncomm​s1362​4 -7 Iglesias JE, Augustinack JC, Nguyen K, Player CM, Player A, Wright Eldridge LL, Engel SA, Zeineh MM, Bookheimer SY, Knowlton BJ M, Roy N et al (2015) A computational atlas of the hippocampal (2005) A dissociation of encoding and retrieval processes in the formation using Ex Vivo, ultra-high resolution MRI: application human hippocampus. J Neurosci 25(13):3280–3286. https​://doi. to adaptive segmentation of in vivo MRI. NeuroImage 115:117– org/10.1523/JNEUR​OSCI.3420-04.2005 137. https​://doi.org/10.1016/j.neuro​image​.2015.04.042 13 2344 Brain Structure and Function (2020) 225:2331–2345 Jansen WJ, Ossenkoppele R, Knol DL, Tijms BM, Scheltens P, Verhey Numakawa T, Odaka H, Adachi N (2018) Actions of brain-derived FRJ, Visser PJ et al (2015) Prevalence of cerebral amyloid pathol- neurotrophin factor in the neurogenesis and neuronal function, ogy in persons without dementia. JAMA 313(19):1924. https​:// and its involvement in the pathophysiology of brain diseases. Int doi.org/10.1001/jama.2015.4668 J Mol Sci 19(11):3650. https​://doi.org/10.3390/ijms1​91136​50 Jin K, Peel AL, Mao XO, Xie L, Cottrell BA, Henshall DC, Greenberg Peña-Casanova J, Quiñones-Ubeda S, Gramunt-Fombuena N, Quin- DA (2004) Increased hippocampal neurogenesis in Alzheimer’s tana-Aparicio M, Aguilar M, Badenes D, Cerulla N et al (2009) disease. Proc Natl Acad Sci USA 101(1):343–347. https​://doi. Spanish multicenter normative studies (NEURONORMA Pro- org/10.1073/pnas.26347​94100​ ject): norms for verbal fluency tests. Arch Clin Neuropsychol Kerchner GA, Berdnik D, Shen JC, Bernstein JD, Fenesy MC, Deutsch Off J Nat Acad Neuropsychol 24(4):395–411. https​: //doi. GK, Wyss-Coray T, Rutt BK (2014) APOE Ε4 worsens hippocam- org/10.1093/arcli​n/acp04​2 pal CA1 apical neuropil atrophy and episodic memory. Neurology Pievani M, Galluzzi S, Thompson PM, Rasser PE, Bonetti M, 82(8):691–697. https:​ //doi.org/10.1212/WNL.000000​ 00000​ 00154​ Frisoni GB (2011) APOE4 is associated with greater atrophy Kowiański P, Lietzau G, Czuba E, Waśkow M, Steliga A, Moryś J of the hippocampal formation in Alzheimer’s disease. Neuro- (2018) BDNF: a key factor with multipotent impact on brain sign- Image 55(3):909–919. https​: //doi.org/10.1016/j.neuro​i mage​ aling and synaptic plasticity. Cell Mol Neurobiol 38(3):579–593. .2010.12.081 https​://doi.org/10.1007/s1057​1-017-0510-4 Pohlack ST, Meyer P, Cacciaglia R, Liebscher C, Ridder S, Flor H Lim YY, Villemagne VL, Laws SM, Ames D, Pietrzak RH, Ellis KA, (2014) Bigger is better! Hippocampal volume and declarative Harrington KD et al (2013) BDNF Val66Met, Aβ amyloid, and memory performance in healthy young men. Brain Struc Func cognitive decline in preclinical Alzheimer’s disease. Neurobiol 219(1):255–267. https​://doi.org/10.1007/s0042​9-012-0497-z Aging 34(11):2457–2464. https​://doi.org/10.1016/j.neuro​biola​ Purcell S, Neale B, Todd-Brown K, Thomas L, Ferreira MAR, Bender ging.2013.05.006 D, Maller J et al (2007) PLINK: a tool set for whole-genome asso- Lim YY, Villemagne VL, Laws SM, Pietrzak RH, Snyder PJ, Ames ciation and population-based linkage analyses. Am J Hum Genet D, Ellis KA et al (2015) APOE and BDNF polymorphisms mod- 81(3):559–575. https​://doi.org/10.1086/51979​5 erate amyloid β-related cognitive decline in preclinical Alzhei- Quinones-Ubeda S (2009) Desenvolupament, normalització i validació mer’s disease. Mol Psychiatry 20(11):1322–1328. https​://doi. de la versió estandard de la segona versió del Test Barcelona. org/10.1038/mp.2014.123 Ramon Llull University, Barcelona Liu Y-H, Jiao S-S, Wang Y-R, Xian-Le Bu, Yao X-Q, Xiang Y, Wang Radmanesh F, Devan WJ, Anderson CD, Rosand J, Falcone GJ, Alz- Q-H et al (2015b) Associations between ApoEε4 carrier status heimer’s Disease Neuroimaging Initiative (ADNI) (2014) Accu- and serum BDNF levels—new insights into the molecular mech- racy of imputation to infer unobserved APOE epsilon alleles in anism of ApoEε4 actions in Alzheimer’s disease. Mol Neurobiol genome-wide genotyping data. Eur J Hum Genet 22(10):1239– 51(3):1271–1277. https​://doi.org/10.1007/s1203​5-014-8804-8 1242. https​://doi.org/10.1038/ejhg.2013.308 Liu Y, Yu J-T, Wang H-F, Han P-R, Tan C-C, Wang C, Meng X-F, Ramier AM, Hecaen H (1970) Role respectif des atteintes frontales Risacher SL, Saykin AJ, Tan L (2015a) APOE genotype and et de la lateralisation lesionnelle dans les deficits de la “fluence neuroimaging markers of Alzheimer’s disease: systematic verbale”. Rev Neurol 123:17–22 review and meta-analysis. J Neurol Neurosurg Psychiatry Reiman EM, Chen K, Liu X, Bandy D, Meixiang Yu, Lee W, Ayu- 86(2):127–134. https​://doi.org/10.1136/jnnp-2014-30771​9 tyanont N et al (2009) fibrillar amyloid-beta burden in cogni- Li B, Shi J, Gutman BA, Baxter LC, Thompson PM, Caselli RJ, tively normal people at 3 levels of genetic risk for Alzheimer’s Wang Y, Alzheimer’s Disease Neuroimaging Alzheimer’s Dis- disease. Proc Natl Acad Sci USA 106(16):6820–6825. https:​ //doi. ease Neuroimaging Initiative (2016) Influence of APOE geno- org/10.1073/pnas.09003​45106​ type on hippocampal atrophy over time - an N=1925 surface- Reinhart V, Bove SE, Volfson D, Lewis DA, Kleiman RJ, Lanz TA based ADNI study. PLoS ONE 11(4):e0152901. https​: //doi. (2015) Evaluation of TrkB and BDNF transcripts in prefrontal org/10.1371/journ​al.pone.01529​01 cortex, hippocampus, and striatum from subjects with schizophre- Molinuevo JL, Gramunt N, Gispert JD, Fauria K, Esteller M, Min- nia, bipolar disorder, and major depressive disorder. Neurobiol Dis guillon C, Sánchez-Benavides G et al (2016) The ALFA project: 77:220–227. https​://doi.org/10.1016/j.nbd.2015.03.011 a research platform to identify early pathophysiological features Ryckman K, Williams SM (2008) Calculation and use of the Hardy- of Alzheimer’s disease. Alzheimer’s Dement Translat Res Clin Weinberg model in association studies. Curr Protoc Hum Genet. Inter 2(2):82–92. https​://doi.org/10.1016/j.trci.2016.02.003 https​://doi.org/10.1002/04711​42905​.hg011​8s57 Morris JC (1993) The clinical dementia rating (Cdr): current ver- Sen A, Nelson TJ, Alkon DL (2015) ApoE4 and A oligomers sion and scoring rules. Neurology 43(11):2412–2414. https​:// reduce bdnf expression via HDAC nuclear translocation. J doi.org/10.1212/wnl.43.11.2412-a Neurosci 35(19):7538–7551. https​://doi.org/10.1523/JNEUR​ Mueller SG, Weiner MW (2009) Selective effect of Age, Apo E4, and OSCI.0260-15.2015 Alzheimer’s disease on hippocampal subfields. Hippocampus Shi J, Leporé N, Gutman BA, Thompson PM, Baxter LC, Caselli RJ, 19(6):558–564. https​://doi.org/10.1002/hipo.20614​ Wang Y, Alzheimer’s Disease Neuroimaging Initiative (2014) Mueller SG, Schuff N, Yaffe K, Madison C, Miller B, Weiner MW Genetic influence of apolipoprotein E4 genotype on hippocampal (2010) Hippocampal atrophy patterns in mild cognitive impair- morphometry: an N = 725 surface-based Alzheimer’s disease neu- ment and Alzheimer’s disease. Hum Brain Mapp 31(9):1339– roimaging initiative study. Hum Brain Mapp 35(8):3903–3918. 1347. https​://doi.org/10.1002/hbm.20934​ https​://doi.org/10.1002/hbm.22447​ Mueller SG, Yushkevich PA, Das S, Wang L, Van Leemput K, Spriggs MJ, Thompson CS, Moreau D, McNair NA, Wu CC, Lamb Iglesias JE, Alpert K et al (2018) Systematic comparison of YN, Mckay NS et al (2018) Human sensory long-term potentia- different techniques to measure hippocampal subfield vol- tion (LTP) predicts visual memory performance and is modulated umes in ADNI2. NeuroImag Clin 17:1006–1018. https​://doi. by the brain-derived neurotrophic factor (BDNF) Val66Met poly- org/10.1016/j.nicl.2017.12.036 morphism. BioRxiv. https​://doi.org/10.1101/28431​5 Notaras M, van den Buuse M (2018) Brain-derived neurotrophic Suthana NA, Donix M, Wozny DR, Bazih A, Jones M, Heidemann RM, factor (bdnf): novel insights into regulation and genetic vari- Trampel R et al (2015) High-Resolution 7T FMRI of human hip- ation. Neurosci. https​: //doi.org/10.1177/10738​5 8418​8 1014​2 pocampal subfields during associative learning. J Cognit Neurosci (107385841881014) 27(6):1194–1206. https​://doi.org/10.1162/jocn_a_00772​ 13 Brain Structure and Function (2020) 225:2331–2345 2345 Toh YL, Ng T, Tan M, Tan A, Chan A (2018) Impact of brain-derived West MJ, Coleman PD, Flood DG, Troncoso JC (1994) Differences neurotrophic factor genetic polymorphism on cognition: a sys- in the pattern of hippocampal neuronal loss in normal ageing tematic review. Brain Behav 8(7):e01009. https:​ //doi.org/10.1002/ and Alzheimer’s disease. Lancet 344(8925):769–772. https​://doi. brb3.1009 org/10.1016/S0140​-6736(94)92338​-8 Tsai S-J (2018) Critical Issues in BDNF Val66Met genetic studies Whitwell JL (2010) The protective role of brain size in Alzheimer’s of neuropsychiatric disorders. Front Mol Neurosci. https​://doi. disease. Expert Rev Neurother 10(12):1799–1801. https​://doi. org/10.3389/fnmol​.2018.00156​ org/10.1586/ern.10.168 van der Meer D, Rokicki J, Kaufmann T, Córdova-Palomera A, Mober- Wisse LEM, Biessels GJ, Geerlings MI (2014) A critical appraisal get T, Alnæs D, Bettella F et al (2018) Brain scans from 21,297 of the hippocampal subfield segmentation package in freesurfer. individuals reveal the genetic architecture of hippocampal sub- Front Aging Neurosci 6:261. https ​ : //doi.org/10.3389/fnagi​ field volumes. Mol Psychiatry. https​://doi.org/10.1038/s4138​ .2014.00261​ 0-018-0262-7 Zeni CP, Mwangi B, Cao Bo, Hasan KM, Walss-Bass C, Zunta-Soares Van Petten C (2004) Relationship between hippocampal volume and G, Soares JC (2016) Interaction between BDNF rs6265 met allele memory ability in healthy individuals across the lifespan: review and low family cohesion is associated with smaller left hippocam- and meta-analysis. Neuropsychologia 42(10):1394–1413. https​:// pal volume in pediatric bipolar disorder. J Affect Disord 189:94– doi.org/10.1016/j.neuro​psych​ologi​a.2004.04.006 97. https​://doi.org/10.1016/j.jad.2015.09.031 Vilar M, Mira H (2016) Regulation of neurogenesis by neurotrophins during adulthood: expected and unexpected roles. Front Neurosci Publisher’s Note Springer Nature remains neutral with regard to 10:26. https​://doi.org/10.3389/fnins​.2016.00026​ jurisdictional claims in published maps and institutional affiliations. Ward DD, Summers MJ, Saunders NL, Janssen P, Stuart KE, Vickers JC (2014) APOE and BDNF Val66Met polymorphisms combine to influence episodic memory function in older adults. Behav Brain Res 271:309–315. https:​ //doi.org/10.1016/j.bbr.2014.06.022 Affiliations Natalia Vilor‑Tejedor1,2,3,4 · Grégory Operto2,5,6 · Tavia E. Evans3 · Carles Falcon2,6,7 · Marta Crous‑Bou2,8,9 · Carolina Minguillón2,5,6 · Raffaele Cacciaglia2,5,6 · Marta Milà‑Alomà2,4,5,6 · Oriol Grau‑Rivera2,5,6,10 · Marc Suárez‑Calvet2,5,6,10 · Diego Garrido‑Martín1 · Sebastián Morán11 · Manel Esteller11,12,13,14 · Hieab H. Adams3,15,16 · José Luis Molinuevo2,4,5,6 · Roderic Guigó1,4 · Juan Domingo Gispert2,4,6,7 · for the ALFA Stu dy 1 9 Centre for Genomic Regulation (CRG), The Barcelona Cancer Epidemiology Research Program, Catalan Institute Institute for Science and Technology, C. Doctor Aiguader of Oncology (ICO), Hospitalet de Llobregat, Barcelona, 88, Edif. PRBB, 08003 Barcelona, Spain Spain 2 10 Barcelonaβeta Brain Research Center (BBRC), Pasqual Servei de Neurologia, Hospital del Mar, Barcelona, Spain Maragall Foundation, Barcelona, Spain 11 Cancer Epigenetics and Biology Program (PEBC), Bellvitge 3 Erasmus MC University Medical Center Rotterdam, Biomedical Biomedical Research Institute (IDIBELL), Department of Clinical Genetics, Rotterdam, Barcelona, Spain The Netherlands 12 Josep Carreras Leukaemia Research Institute (IJC), 4 Universitat Pompeu Fabra (UPF), Barcelona, Spain Badalona, Barcelona, Spain 5 13 CIBER Fragilidad y Envejecimiento Saludable (CIBERFES), Institució Catalana de Recerca i Estudis Avançats (ICREA), Madrid, Spain Barcelona, Spain 6 14 IMIM (Hospital del Mar Medical Research Institute), Physiological Sciences Department, School of Medicine Barcelona, Spain and Health Sciences, University of Barcelona (UB), 7 Barcelona, Spain Centro de Investigación Biomédica en Red de Bioingeniería, 15 Biomateriales y Nanomedicina (CIBER-BBN), Madrid, Erasmus MC University Medical Center Rotterdam, Spain Department of Epidemiology, Rotterdam, The Netherlands 8 16 Department of Epidemiology, Harvard T.H. Chan School Erasmus MC University Medical Center Rotterdam, of Public Health, Boston, MA, USA Department of Radiology, Rotterdam, The Netherlands 13