The effect of TOMM40 on spatial navigation in amnestic mild cognitive impairment

Neurobiology of Aging 36 (2015) 2024e2033

Contents lists available at ScienceDirect

Neurobiology of Aging journal homepage: www.elsevier.com/locate/neuaging

The effect of TOMM40 on spatial navigation in amnestic mild cognitive impairment Jan Laczó a, b, *, Ross Andel b, c, Martin Vyhnalek a, b, Vaclav Matoska d, Vojtech Kaplan d, Zuzana Nedelska a, b, Ondrej Lerch a, Ivana Gazova a, b, Scott D. Moffat e, Jakub Hort a, b a Memory Clinic, Department of Neurology, Charles University in Prague, Second Faculty of Medicine and Motol University Hospital, Prague, The Czech Republic b International Clinical Research Center, St. Anne’s University Hospital Brno, Brno, The Czech Republic c School of Aging Studies, University of South Florida, Tampa, FL, USA d Department of Clinical Biochemistry, Hematology and Immunology, Homolka Hospital, Prague, The Czech Republic e School of Psychology, Georgia Institute of Technology, Atlanta, GA, USA

a r t i c l e i n f o

a b s t r a c t

Article history: Received 11 November 2014 Received in revised form 11 February 2015 Accepted 7 March 2015 Available online 16 March 2015

The very long (VL) poly-T variant at rs10524523 (“523”) of the TOMM40 gene may hasten the onset of late-onset Alzheimer’s disease (LOAD) and induce more profound cognitive impairment compared with the short (S) poly-T variant. We examined the influence of TOMM40 “523” polymorphism on spatial navigation and its brain structural correlates. Participants were apolipoprotein E (APOE) ε3/ε3 homozygotes with amnestic mild cognitive impairment (aMCI). The homozygotes were chosen because APOE ε3/ε3 variant is considered “neutral” with respect to LOAD risk. The participants were stratified according to poly-T length polymorphisms at “523” into homozygous for S (S/S; n ¼ 16), homozygous for VL (VL/VL; n ¼ 15) TOMM40 poly-T variant, and heterozygous (S/VL; n ¼ 28) groups. Neuropsychological examination and testing in real-space human analog of the Morris Water Maze were administered. Both selfcentered (egocentric) and world-centered (allocentric) spatial navigation was assessed. Brain magnetic resonance imaging scans were analyzed using FreeSurfer software. The S/S group, although similar to S/VL and VL/VL groups in demographic and neuropsychological profiles, performed better on allocentric navigation (p  0.004) and allocentric delayed recall (p  0.014), but not on egocentric navigation. Both S/VL and VL/VL groups had thinner right entorhinal cortex (p  0.043) than the S/S group, whereas only the VL/VL group had thinner left entorhinal cortex (p ¼ 0.043) and left posterior cingulate cortex (p ¼ 0.024) than the S/S group. In conclusion, TOMM40 “523” VL variants are related to impairment in allocentric spatial navigation and reduced cortical thickness of specific brain regions among aMCI individuals with (LOAD neutral) APOE ε3/ε3 genotype. This may reflect a specific role of TOMM40 “523” in the pathogenesis of LOAD. Ó 2015 Elsevier Inc. All rights reserved.

Keywords: Alzheimer’s disease Apolipoprotein E Magnetic resonance imaging Morris Water Maze Neuropsychology Memory Hippocampus

1. Introduction Among early cognitive markers of Alzheimer’s disease (AD), increasing attention has been paid recently to spatial navigation impairment (Abbott and Callaway, 2014; Lithfous et al., 2013; Serino et al., 2014; Vl cek and Laczó, 2014). Difficulties with navigation in unfamiliar surroundings is a commonly reported complaint by patients with AD. Impairment of spatial navigation is often present early in the course of ADdin individuals with mild * Corresponding author at: Department of Neurology, Second Faculty of Medicine, Charles University in Prague and Motol University Hospital, V Úvalu 84, Praha 5 e Motol 150 06, The Czech Republic. Tel.: þ420 224 436 816; fax: þ420 224 436 875. E-mail address: [email protected] (J. Laczó). 0197-4580/$ e see front matter Ó 2015 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.neurobiolaging.2015.03.004

cognitive impairment (MCI) (Benke et al., 2014; Cushman et al., 2008; deIpolyi et al., 2007). This is especially so for individuals with amnestic MCI (aMCI) (Hort et al., 2007; Weniger et al., 2011), whose risk of conversion to AD is particularly high (Petersen, 2004), and its subgroups composed of individuals with hippocampal type of memory impairment (Laczó et al., 2009, 2012), hippocampal atrophy (Nedelska et al., 2012), and in apolipoprotein E (APOE) ε4 carriers (Berteau-Pavy et al., 2007; Laczó et al., 2010, 2011). Spatial navigation includes 2 basic navigation strategiesdthe self-centered (egocentric) and the world-centered (allocentric) (Maguire et al., 1998). Allocentric navigation, which uses distal orientation cues (landmarks) for navigation independent of the individual’s position, is closely tied with the function of the hippocampus (Maguire et al., 1998) and its connections with the

J. Laczó et al. / Neurobiology of Aging 36 (2015) 2024e2033

entorhinal cortex (Ekstrom et al., 2003). Egocentric navigation, which uses own position to set distances and directions to the final destination (is dependent on own position), is reflected in the function of the posterior-inferior parietal cortex (Maguire et al., 1998) and precuneus (Weniger et al., 2009), as well as the caudate nucleus when navigation is relatively simple (Hartley et al., 2003). Finally, the posterior cingulate cortex plays a crucial role for translation between egocentric and allocentric navigation strategies (Byrne et al., 2007). The APOE ε4 is the main known genetic risk factor for the most common form of AD, late-onset AD (LOAD) (Saunders et al., 1996), and it also increases the risk of conversion from MCI to dementia (Xu et al., 2013). Recently, the TOMM40 gene, which encodes the translocase of the outer mitochondrial membrane pore subunit (Humphries et al., 2005) and is adjacent to and in linkage disequilibrium with APOE, has been proposed as another genetic risk factor for LOAD (Roses et al., 2010). Roses et al. (2010) demonstrated that a variable length deoxythymidine homopolymer (poly-T) at rs10524523 (“523”) within intron 6 of the TOMM40 gene modulates risk and onset age of LOAD (Lutz et al., 2010; Roses et al., 2010). The length of the poly-T homopolymer was proposed to be categorized as short (14e20 repeats; i.e., S), long (21e29 repeats, i.e., L), or very long (>29 repeats, i.e., VL). Because of the linkage disequilibrium between APOE and TOMM40 genes, the APOE ε3 allele may be linked to either the S or VL variants of TOMM40 “523”, whereas APOE ε4 is almost exclusively linked to the L variant (Roses et al., 2010). In individuals with APOE ε3, including APOE ε3/ε3 carriers, who were considered to have neutral risk for LOAD, the VL variants were found to be associated with a higher risk for and earlier onset age of LOAD, whereas the S variants were associated with later age of onset (Roses et al., 2010). Researchers have suggested the importance of TOMM40 “523” in early LOAD pathogenesis and proposed that in asymptomatic individuals TOMM40 “523” may induce cognitive and brain changes similar to those found in AD (Caselli et al., 2012; Hayden et al., 2012; Johnson et al., 2011). The VL variants were associated with lower memory performance (Caselli et al., 2012; Hayden et al., 2012; Johnson et al., 2011), reduced gray matter volume in the posterior cingulate cortex, and greater age-related differences in gray matter volume of the right anterior medial temporal lobe (Johnson et al., 2011). We build on previous research demonstrating that spatial navigation impairment is present very early in the course of AD (deIpolyi et al., 2007; Hort et al., 2007; Weniger et al., 2011). It is also known that spatial navigation is highly susceptible to the influence of genetic risk factors, especially to the deleterious effect of APOE ε4 (Berteau-Pavy et al., 2007; Laczó et al., 2011), with the right hippocampal volume particularly affected (Laczó et al., 2014). Here we focused on individuals with aMCI with the APOE ε3/ε3 genotype, considered “neutral” with respect to LOAD risk, thus eliminating variance associated with protective and risk APOE alleles (Roses et al., 2010). We assessed the effects of a new promising genetic risk factor for LOAD, TOMM40 “523”, using data from performance on 2 basic spatial navigation strategies and the relevant structural brain correlates. The aims of this study were to evaluate the associations between the poly-T variants at rs10524523 of TOMM40 gene and (1) 2 basic spatial navigation strategies, allocentric and egocentric, in a realspace human analog of the Morris Water Maze (hMWM) and (2) volumes or cortical thicknesses of the brain regions known to be important for spatial navigation (i.e., hippocampus, caudate nucleus, entorhinal, posterior-inferior parietal, precuneal, and posterior cingulate cortices) among individuals with aMCI with the APOE ε3/ε3 genotype.

2025

We hypothesized that those with VL poly-T variant at rs10524523 of TOMM40 gene would perform worse on spatial navigation than those homozygous for S poly-T variant. We expected differences primarily in the allocentric navigation strategy, which is the first navigation strategy impaired in AD (Hort et al., 2007) and in normal aging (Rodgers et al., 2012) and which depends on the medial temporal lobe structures, an initial site of AD pathological processes (Braak and Braak, 1991) and TOMM40-induced mitochondrial dysfunction and apoptotic processes (Devi et al., 2006). Furthermore, we hypothesized that spatial navigation impairment among those with VL poly-T variant would be accompanied by atrophy of (1) the medial temporal lobe structures, which are affected early in AD and are also susceptible to TOMM40 “523” induced age-dependent gray matter volume reduction (Johnson et al., 2011), and (2) the posterior cingulate cortex, which is one of the earliest regions involved in AD pathogenesis (Rowe et al., 2007), susceptible to TOMM40 “523” induced gray matter volume reduction (Johnson et al., 2011) and a key structure for translation between egocentric and allocentric navigation strategies (Byrne et al., 2007). 2. Methods 2.1. Participants Fifty-nine right-handed APOE ε3 homozygotes with aMCI were recruited from the Czech Brain Aging Study cohort at the Memory Disorders Clinic at Second Faculty of Medicine, Charles University and Motol University Hospital in Prague, Czech Republic between February 2006 and February 2012. They underwent standard neurological, internal and laboratory evaluations, brain magnetic resonance imaging at 1.5 T, neuropsychological examination, and spatial navigation testing in a real-space hMWM within 2 months from the first visit. Participants were referred to the clinic by general practitioners, neurologists, psychiatrists, and geriatricians. Referral to the memory clinic was triggered by memory complaint from the patient or the informant. All participants met clinical criteria for aMCI (Petersen, 2004), including memory complaints reported by a patient or caregiver, evidence of memory dysfunction on neuropsychological testing, generally intact activities of daily living and absence of dementia. All participants had Clinical Dementia Rating global score no greater than 0.5, which commonly designates MCI (Morris, 1993). Memory impairment was established when the patient scored more than 1.5 standard deviations below the mean of age- and education-adjusted norms on any memory test (Laczó et al., 2011). The aMCI patients included both those with isolated memory impairment (single-domain aMCI) and those with memory impairment and additional impairment in any other non-memory domain (multiple-domain aMCI). Participants with depression (6 points on the 15-item Geriatric Depression Scale); (Yesavage, 1988), with a Hachinski Ischemic Scale score 5, and with a history of other primary neurological or psychiatric disorders were not included in this study. The aMCI patients were further stratified into 3 groups based on the TOMM40 poly-T length polymorphisms at rs10524523 using a standard procedure (Roses et al., 2010)dhomozygous for S (S/S; n ¼ 16) poly-T variant, homozygous for VL (VL/VL; n ¼ 15) poly-T variant, and heterozygous (S/VL; n ¼ 28) groups. Group-wise characteristics are listed in Table 1. The study was approved by an institutional ethical committee and the participants have signed written informed consent. 2.2. Neuropsychological battery The neuropsychological assessment comprised the following tests: (1) verbal memory measured with the Rey Auditory

2026

J. Laczó et al. / Neurobiology of Aging 36 (2015) 2024e2033

Table 1 Characteristics of study participants; means (SD) Variables

aMCI S/S (n ¼ 16)

aMCI S/VL (n ¼ 28)

aMCI VL/VL (n ¼ 15)

p-Values

Effect sizes

Women, n (%) Age in years Age in years (range) Education in years MMSE score GDS score RAVLT 1e5 score RAVLT 30 score ECRT free recall score ECRT total recall score TMT A score (in seconds) TMT B score (in seconds) COWAT score Digit span total numbers recalled Reversed digit span total numbers recalled ROCFT-C score ROCFT-IR score BNT error score Allocentric-egocentric navigation distance error value (cm) Egocentric navigation distance error value (cm) Allocentric navigation distance error value (cm) Delayed navigation distance error value (cm) Left hippocampal volume, eTIV adjusted (mm3) Right hippocampal volume, eTIV adjusted (mm3) Left caudate volume, eTIV adjusted (mm3) Right caudate volume, eTIV adjusted (mm3) Left entorhinal cortical thickness (mm) Right entorhinal cortical thickness (mm) Left posterior inferior parietal cortical thickness (mm) Right posterior inferior parietal cortical thickness (mm) Left precuneal cortical thickness (mm) Right precuneal cortical thickness (mm) Left posterior cingulate cortical thickness (mm) Right posterior cingulate cortical thickness (mm)

9 73.9 65e80 14.8 27.6 2.6 36.9 4.7 4.6 12.8 54.8 164.0 41.8 6.0 4.2 28.1 6.6 6.3 49.4 33.6 47.8 53.3 3654.3 3765.7 3313.0 3328.9 3.2 3.5 2.3 2.5 2.2 2.2 2.3 2.2

9 75.4 62e86 15.1 26.2 3.0 32.5 4.5 6.2 13.6 55.4 179.6 33.1 5.4 4.1 26.9 9.6 5.3 63.5 59.1 99.3 120.6 2907.3 3172.0 3451.7 3471.0 2.7 2.8 2.2 2.2 1.9 1.9 2.1 2.1

7 75.4 62e84 13.6 27.6 3.0 35.6 6.6 6.6 14.2 70.1 234.7 32.1 5.4 4.0 29.1 8.8 5.3 71.0 66.4 87.9 106.2 3059.7 3070.4 3143.3 3181.0 2.7 2.8 2.1 2.1 2.0 2.0 1.9 2.0

0.276 0.707 N/A 0.319 0.144 0.922 0.652 0.392 0.521 0.734 0.327 0.388 0.105 0.222 0.876 0.461 0.518 0.895 0.631 0.052 <0.001 0.001 0.214 0.074 0.218 0.328 0.045 0.022 0.138 0.118 0.407 0.115 0.013 0.171

0.21 0.01 N/A 0.04 0.07 0.01 0.03 0.05 0.04 0.02 0.08 0.05 0.08 0.07 0.01 0.04 0.04 0.02 0.08 0.12 0.37 0.32 0.08 0.16 0.10 0.07 0.20 0.24 0.12 0.13 0.06 0.13 0.24 0.11

(56) (6.0) (3.4) (2.4) (3.1) (9.9) (4.5) (4.0) (5.0) (18.8) (140.2) (5.9) (1.1) (1.1) (3.1) (5.0) (3.8) (29.4) (12.1) (23.0) (41.3) (784.1) (617.1) (280.8) (303.5) (0.3) (0.5) (0.2) (0.1) (0.1) (0.1) (0.2) (0.2)

(32) (6.2) (3.1) (2.7) (2.0) (12.5) (3.8) (3.9) (3.9) (20.0) (97.5) (12.3) (0.9) (1.0) (5.1) (6.6) (3.3) (32.2) (43.8) (29.7)*** (46.3)*** (611.0) (419.5) (528.0) (605.5) (0.6) (0.6)* (0.3) (0.2) (0.2) (0.2) (0.2) (0.3)

(47) (6.5) (2.6) (2.8) (2.1) (15.5) (4.3) (4.4) (3.0) (23.4) (150.5) (10.2) (0.9) (1.1) (3.8) (7.4) (3.8) (33.7) (39.7) (35.2)** (39.9)* (585.6) (639.1) (586.5) (649.8) (0.6)* (0.8)* (0.3) (0.3) (0.4) (0.3) (0.3)* (0.2)

Demographic, neuropsychological, spatial navigation, and MRI characteristics of the groups. Values are mean (SD) except for gender and age range. Effect sizes indicating the differences among all groups were calculated as Cramér’s V for c2 (gender) and partial eta2 for ANOVA, MANCOVA, and linear mixed effects regression models comparisons (all other variables). For p indicating the level of significance compared with aMCI S/S group are *p < 0.05; **p < 0.01; ***p < 0.001. Key: aMCI S/S, amnestic MCI TOMM40 homozygous for S; aMCI S/VL, amnestic MCI TOMM40 heterozygous group; aMCI VL/VL, amnestic MCI TOMM40 homozygous for VL; BNT, Boston Naming Test; COWAT, Controlled Oral Word Association Test; ECRT, Enhanced Cued Recall Test; eTIV, estimated total intracranial volume; GDS, Geriatric Depression Scale; MCI, mild cognitive impairment; MMSE, Mini-Mental State Examination; RAVLT, Rey Auditory Verbal Learning Test; RAVLT 1e5, trials 1 to 5 total; RAVLT 30, word recall after 30 min; ROCFT-C, Rey-Osterrieth Complex Figure Testdthe Copy condition; ROCFT-IR, Rey-Osterrieth Complex Figure Testdthe Immediate Recall condition; TMT A and B, Trail Making Tests A and B.

Verbal Learning Test trials 1e5, the Rey Auditory Verbal Learning Test 30-minute Delayed Recall trial, and a 16-item picture version of the Enhanced Cued Recall Test (free and total recall scores); (2) non-verbal memory measured with the Rey-Osterrieth Complex Figure Testdthe Immediate Recall condition; (3) visuospatial function measured with the ReyOsterrieth Complex Figure Testdthe Copy condition; (4) executive function measured with the Trail Making Test B and Controlled Oral Word Association Test; (5) attention and working memory measured with the Forward and Backward Digit Span and Trail Making Test A; and (6) language measured with the Boston Naming Test (30-item version). The MiniMental State Examination (MMSE) was administered to measure global cognitive function. Group-wise neuropsychological characteristics are listed in Table 1.

2.4. TOMM40 rs10524523 analysis TOMM40 rs10524523 polymorphism was analyzed by a novel high-resolution melting screening method (Fig. 1). This method uses the fact of bimodal distribution of TOMM40 alleles in APOE ε3/ε3 patients (Johnson et al., 2011). Polymerase chain reaction primers: FOR: 50 -ACT GGC ATG AGC CAT TGC ATC TGG C-30 ; REV: 50 -CGG GCA ACA TGG TGA GAC CCC ATC TC-30 ; polymerase chain reaction conditions: 25-mL reaction; BioTaq 0.2 mL (5 U/mL); 1-mL primers (10 mM/L); 2.5-mL buffer (10); 0.5-mL dNTPs (10 mM); 2-mL MgCl2 (50 mM). Thermocycling conditions: hot start 95  C/8 minutes; (95  C/20 seconds þ 74  C/20 seconds)/40 cycles. Subsequently, 2 mL of LCGreen were added for high-resolution melting analysis (70e90  C). 2.5. Magnetic resonance imaging acquisition and analysis

2.3. APOE genotyping To determine the APOE genotype, DNA was isolated from blood samples (ethylenediaminetetraacetic acid; Qiagen extraction) and genotyping was performed according to Idaho-tech protocol (LunaProbes Genotyping Apolipoprotein [ApoE] Multiplexed Assay) for high-resolution melting analysis (Hixson and Vernier, 1990; Laczó et al., 2011).

Brain scans were performed on a 1.5 T scanner (Siemens AG, Erlangen, Germany) using T1-weighted 3-dimensional highresolution magnetization-prepared rapid acquisition with gradient echo sequence with the following parameters: TR/TE/ TI ¼ 2000/3.08/1100 ms, flip angle 15 , 192 continuous partitions, slice thickness 1.0 mm, and in-plane resolution 1 mm. Scans were visually inspected by a neuroradiologist to ensure appropriate data

J. Laczó et al. / Neurobiology of Aging 36 (2015) 2024e2033

Fig. 1. High-resolution melting analysis of TOMM40 rs10524523 polymorphism. Separation of distinct genotypes by high-resolution melting analysisdVL/VL genotypes are represented by red curves; S/VL (heterozygous) genotypes are represented by gray curves; S/S genotypes are represented by blue curves. Abbreviations: S, short; VL, very long. (For interpretation of the references to color in this Figure, the reader is referred to the web version of this article.)

quality and to exclude patients with a major brain pathology that could interfere with cognitive functioning such as cortical infarctions, neoplasm, subdural hematoma, or hydrocephalus. We used fully automated FreeSurfer algorithm (version 5.3.0; http:// surfer.nmr.mgh.harvard.edu), described in detail elsewhere (Fischl et al., 2004), to measure cortical thicknesses and hippocampal and caudate volumes. The CPU-intensive FreeSurfer analyses were performed on computer cluster using neuGRID portal (http:// neuGRID.eu). To limit the number of multiple comparisons, only regions known to be involved in spatial navigation (Ekstrom et al., 2003; Maguire et al., 1998; Ohnishi et al., 2006) and associated with risk for AD (Mosconi et al., 2007; Reiman et al., 1996) were included in the analyses. These included hippocampal and caudate volumes and thickness measurements of entorhinal, inferior parietal, precuneal, and posterior cingulate cortices. Measurements were performed and reported separately for the left and right side. The outputs from FreeSurfer were visually checked for any potential errors. Furthermore, the distributions of hippocampal and caudate volumes and all thickness measures were assessed and no substantial outliers were identified. Hippocampal volumes were corrected for the differences in head size by regressing the estimated total intracranial volume (eTIV) using this formula: Hippocampal volumei (adjusted) ¼ hippocampal volumei (baseline)  B(eTIVieeTIV mean), where eTIVi ¼ the ith subject’s eTIV, eTIV mean ¼ overall average eTIV, and B ¼ the slope of the hippocampal volume regression on eTIV (Jack et al., 1989). Caudate volumes were also eTIV adjusted using this formula. Cortical thicknesses were not eTIV adjusted. Morphometric characteristics of the participants are listed in Table 1.

2.6. Spatial navigation testing with the Hidden Goal Taskdthe human analog of the MWM test The procedure was described in detail previously (Hort et al., 2007). Briefly, spatial navigation testing was performed in the Laboratory of Spatial Cognition, a joint workplace of the Department of Neurology, Second Faculty of Medicine, Charles University in Prague, Czech Republic and Institute of Physiology, Academy of Sciences of the Czech Republic, Prague, Czech Republic. For spatial navigation testing, we used the real-space version of the Hidden Goal Task, a hMWM test, that is designed to test 2 basic distinct types of spatial navigation strategiesdthe allocentric and the egocentric. The hMWM test mimics the conditions in the original

2027

MWM task, where animals are trained to locate the position of the hidden platform: (1) in relation from the constant releasing position at the pool periphery (egocentric subtask), (2) in relation to the distal orientation cues in the room while releasing from the different positions at the pool periphery (allocentric subtask), and (3) in the probe trial when hidden platform is removed and distal orientation cues are used for navigation (allocentric delayed subtask). To familiarize participants with the testing procedure, we initially presented trials from the allocentric-egocentric subtask where both spatial navigation strategies could be used. The real-space version of the Hidden Goal Task was performed in a real-space navigation setting called the Blue Velvet Arena, a fully enclosed cylindrical arena 2.8 m in diameter and 2.9 m high surrounded by a dark blue velvet curtain and described in detail elsewhere (Fig. 2A) (Hort et al., 2007; Laczó et al., 2011). The participant was asked to locate an invisible goal in 4 different subtasks using the start position or 2 distal orientation cues, respectively (Fig. 2B). The position of the goal on the arena floor was constant across all trials (Fig. 2C). To begin the task, participants were asked to enter the arena. They were given a long-standing pole with a light-emitting diode and shown the position of the goal on the arena floor. Before each trial, they were asked to stand at the side of the arena, go directly from their start position to the goal, and to place the pole directly on the presumed goal position. The goal was briefly shown after each trial to facilitate learning (although the goal was not shown any time during the delayed subtask). The allocentric-egocentric subtask was performed first. It involved locating the goal using its spatial relationship with both the start position and the 2 distal orientation cues on the arena wall. This was considered a training subtask designed to familiarize participants with the testing procedure. This subtask was followed with an egocentric subtask, which involved using only the start position to locate the goal with no distal orientation cues displayed. It is analogous to the condition in the original MWM task, where animals are released from a constant position at the pool periphery to search for a hidden platform. The egocentric subtask was followed with an allocentric subtask, which involved using only 2 distal orientation cues at the perimeter of the arena for navigation to the goal as the start position was unrelated to the goal position. It is analogous to the condition in the original MWM task, where animals are released from “different positions” at the pool periphery to search for the hidden platform and thus only distal orientation cues are used for navigation. Finally, the delayed subtask, which was a repeat of the allocentric subtask, was administered 30 minutes after the initial allocentric subtask was completed. It is analogous to the probe trial in the original MWM task, where the hidden platform is removed and only distal orientation cues are used for navigation. As in the probe trial in the original MWM task, no feedback through showing the hidden goal was provided. Each subtask involved 8 trials performed in direct sequence with the exception of the delayed subtask that involved only 2 consecutive trials. The positions of the goal were stable across all trials relative to (1) the positions of the start location and both orientation cues in the allocentric-egocentric subtask, (2) positions of the start location in the egocentric subtask, and (3) positions of both orientation cues in the allocentric and delayed subtask. Performance was recorded automatically by the computer as the distance error between the standing pole’s final position and the actual goal location (in centimeters). There was no time limit to find the goal, mainly to reduce bias by differences in sensory and physical functioning. Examiners were blinded to the results of the other examinations and they supervise the correct performance of the task without interference beyond standard instructions.

2028

J. Laczó et al. / Neurobiology of Aging 36 (2015) 2024e2033

Fig. 2. The Hidden Goal Task. (A) Scheme of the real-space version from an overhead view with the largest circle representing the arena, the small circle in the arena representing the goal position, the mid-size circle on the edge of the arena representing the start position and the 2 lines on the edge of the arena representing the cues. The line representing tracking by a subject between the start and the goal positions is also depicted. (B) The scheme of the first 3 individual subtasks: allocentric-egocentric, egocentric, and allocentric. The delayed subtask (not shown here) is the same as the allocentric subtask. (C) The scheme of the real-space version from the overhead view rotated 90 clockwise from the previous trial shown in Fig. 2A.

2.7. Statistical analyses An analysis of variance (ANOVA) with post hoc Tukey’s honest significant differences test evaluated the mean differences between groups in age, years of education, MMSE scores, and results of neuropsychological tests. The c2 test evaluated the differences in gender proportions. The Pearson’s correlation coefficients were calculated to explore the bivariate relationships. All other quantitative data were found to be adequate for parametric analysis. First, to properly account for the repeated measures structure of the spatial navigation data, we used linear mixed effects regression models (Singer and Willett, 2003). This method of analysis yields the same output as repeated measures ANOVA, but it is more versatile with respect to the specification of variance-covariance matrix, can handle repeated measures data properly, and allows for the specification of random effects. The distance between the participant’s choice of the goal position and the correct goal position (distance error) in the spatial navigation task (in centimeters) was the outcome. All distance error values were converted into z scores (mean ¼ 0, standard deviation ¼ 1), which allowed us to present the main results in standard deviation units. The TOMM40 status (group membership) was the independent variable. Our main interest was in the main effect for group (S/S vs. S/VL vs. VL/VL). We also report the main effect for trial (trials 1e8 for the egocentric or allocentric subtasks) and the group-by-trial interaction, which reflect learning and differential learning by group, respectively. The intercept, trial, and a person identifier were specified as random effects. Based on model fit, the final models used the autoregressive covariance structure. Holm-Bonferroni correction for multiple comparisons was used in the linear mixed effects regression models (Holm, 1979). Second, to show whether TOMM40 status had a different effect on allocentric and egocentric navigational strategies we used a linear mixed effects regression model where the distance error in allocentric and egocentric subtasks was the outcome and the TOMM40 status (S/S vs. S/VL vs. VL/VL) and the type of subtask

(egocentric ¼ 0, allocentric ¼ 1) were the independent variables. This approach allowed a direct comparison between the magnitude of the association with the allocentric and egocentric subtasks. The intercept and a person identifier were specified as random effects. Finally, to assess the differences between groups in volumes and cortical thicknesses of the brain structures relevant for spatial navigation, we estimated multivariate analysis of covariance (MANCOVA) controlling for age and sex. Education was not included because, unlike age and sex, it did not correlate with any volumetric and thickness measures. Initially, we performed multivariate tests. When the Wilks’ Lambda indicating overall differences between groups in morphometric characteristics of the brain structures was significant, we subsequently performed univariate tests with post hoc Tukey’s honest significant difference test, where the TOMM40 status was the independent variable and right and left hippocampal and caudate volumes and cortical thicknesses of right and left entorhinal, posterior-inferior parietal, precuneal, and posterior cingulate cortices converted into z scores were separately entered as the outcome. Statistical significance was set at 2-tailed alpha of 0.05. Effect sizes are reported using Cramér’s V for the c2 test (Cramér, 1999) and partial eta2 for ANOVA, MANCOVA, and linear mixed effects regression analyses (Tabachnick and Fidell, 2007). Partial eta2 of 0.2 corresponds to Cohen’s d of 1.0. With our sample size, Cramér’s V of about 0.47 corresponds to Cohen’s d of 1.0. All analyses were conducted by using IBM SPSS for Windows version 20.0. Observed power, outputted automatically by SPSS, is reported with the 2-tailed alpha of 0.05. 3. Results 3.1. Basic characteristics The S/S, S/VL, and VL/VL groups did not differ in age, gender, education, MMSE, and Geriatric Depression Scale scores or in scores

J. Laczó et al. / Neurobiology of Aging 36 (2015) 2024e2033

of any neuropsychological test. The results are presented in detail in Table 1. The allele distribution in our sample (N ¼ 59) exhibited a Hardy-Weinberg equilibration (c2 ¼ 0.15). In correlational analyses (Supplementary Table 1), age did not correlate with spatial navigation subtasks but it correlated negatively with hippocampal volumes and posterior-inferior parietal cortical thicknesses (r  0.39; p < 0.05). Gender did not correlate with spatial navigation subtasks, but correlated negatively with posterior-inferior parietal, precuneal, and posterior cingulate cortical thicknesses, indicating thinner cortex among women in these regions (r  0.36; p < 0.05). Education did not correlate with spatial navigation subtasks or volumetric and thickness measures of brain structures. Extent of errors on the egocentric navigation subtask correlated negatively with posterior-inferior parietal and posterior cingulate cortical thicknesses (r  0.40; p < 0.05). Extent of errors on the allocentric navigation subtask correlated negatively with hippocampal volumes and entorhinal cortical thicknesses (r  0.37; p < 0.05). 3.2. TOMM40 “523” polymorphism and spatial navigation in aMCI Using the linear mixed models, we found significant main effects for group in the allocentric (F[2,55] ¼ 8.36; p < 0.001; partial eta2 ¼ 0.37; observed power ¼ 0.98) and delayed (F[2,55] ¼ 9.10; p ¼ 0.001; partial eta2 ¼ 0.32; observed power ¼ 0.95) subtasks, but not in the allocentric-egocentric (F[2,55] ¼ 0.46; p ¼ 0.631; partial eta2 ¼ 0.08; observed power ¼ 0.31) and egocentric (F[2,55] ¼ 3.01; p ¼ 0.052; partial eta2 ¼ 0.12; observed power ¼ 0.47) subtasks. The S/VL and VL/VL groups exhibited poorer overall navigation accuracy than the S/S group in the allocentric (p  0.004) and delayed (p  0.014) subtasks. There were no differences between the S/VL and VL/VL groups in the allocentric (p ¼ 0.642) and delayed (p ¼ 0.768) subtasks (Fig. 3). Application of the Holm-Bonferroni correction for multiple comparisons did not affect the results. The between-group comparisons of spatial navigation performance in each subtask are listed in Table 2. The main effects for trial were not significant in the allocentricegocentric (F[1,59] ¼ 0.03; p ¼ 0.874; partial eta2 ¼ 0.01; observed power ¼ 0.10), egocentric (F[1,59] ¼ 0.27; p ¼ 0.602; partial eta2 ¼ 0.02; observed power ¼ 0.25), and allocentric (F[1,59] ¼ 0.28; p ¼ 0.598; partial eta2 ¼ 0.02; observed power ¼ 0.27) subtasks, indicating no significant learning effect across consecutive trials in the sample overall. Finally, there were no significant group-by-trial interactions, suggesting no differences in learning between the groups in the allocentric-egocentric (F[2,59] ¼ 0.44; p ¼ 0.643; partial eta2 ¼ 0.02; observed power ¼ 0.22), egocentric (F[2,59] ¼ 1.32; p ¼ 0.271; partial eta2 ¼ 0.05; observed power ¼ 0.43), and allocentric (F[2,59] ¼ 0.04; p ¼ 0.996; partial eta2 ¼ 0.01; observed power ¼ 0.09) subtasks (Fig. 4). 3.3. The effect of TOMM40 “523” polymorphism on egocentric versus allocentric navigation strategies In addition, we used the linear mixed models to test statistically whether allocentric navigation is related to TOMM40 “523” polymorphism significantly more than egocentric navigation. We found significant main effects for group indicating overall differences in spatial navigation performance among the groups (F[2,55] ¼ 12.14; p < 0.001; partial eta2 ¼ 0.15; observed power ¼ 1.00). Specifically, the S/VL and VL/VL groups exhibited poorer overall navigation accuracy than the S/S group (p ¼ 0.004 and p ¼ 0.016, respectively). There were no differences between the S/VL and VL/VL groups (p ¼ 0.996). Furthermore, there was a significant main effect for the type of the subtask (F[1,56] ¼ 5.98; p ¼ 0.015; partial eta2 ¼ 0.02; observed power ¼ 0.66), indicating poorer allocentric compared

2029

Fig. 3. Average spatial navigation performance in each subtask. Mean distance errors (SD) from the goal are depicted. Asterisks (*) indicate significant differences from the S/S group. Abbreviations: aMCI, amnestic mild cognitive impairment; S, short; VL, very long.

with egocentric navigation in the sample overall. Finally, there was significant group-by-subtask interaction (F[2,59] ¼ 4.81; p ¼ 0.008; partial eta2 ¼ 0.03; observed power ¼ 0.77), indicating that the S/VL and VL/VL groups had poorer allocentric than egocentric navigation compared with the S/S group (Table 3). 3.4. TOMM40 “523” polymorphism and brain structures relevant for spatial navigation Using the MANCOVA adjusted for age and gender, the multivariate tests showed significant main group effects for the differences in volumes and cortical thicknesses of the brain structures relevant for spatial navigation (Wilks’ Lambda ¼ 2.44; p ¼ 0.049; partial eta2 ¼ 0.89; observed power ¼ 0.83). In the subsequent univariate tests for the each brain structure, the significant main group effect was found for the left and right entorhinal cortical thicknesses (F[2,53] ¼ 3.43; p ¼ 0.045; partial eta2 ¼ 0.20; observed power ¼ 0.53 and F[2,53] ¼ 4.33; p ¼ 0.022; partial eta2 ¼ 0.24; observed power ¼ 0.65, respectively) and for the left posterior cingulate cortex thickness (F[2,53] ¼ 4.99; p ¼ 0.013; partial eta2 ¼ 0.24; observed power ¼ 0.66). The main group effect for the Table 2 Group-wise comparisons of adjusted mean error distances from the goal in spatial navigation subtasks p-Value 95% Confidence interval for (I) Groupcode (J) Groupcode Mean difference difference (IJ) Lower bound Upper bound Allocentric-egocentric subtask aMCI S/S aMCI S/VL aMCI VL/VL aMCI S/VL aMCI VL/VL Egocentric subtask aMCI S/S aMCI S/VL aMCI VL/VL aMCI S/VL aMCI VL/VL Allocentric subtask aMCI S/S aMCI S/VL aMCI VL/VL aMCI S/VL aMCI VL/VL Delayed subtask aMCI S/S aMCI S/VL aMCI VL/VL aMCI S/VL aMCI VL/VL

0.29 0.44 0.15

0.458 0.82 0.178 1.00 0.874 0.69

0.24 0.13 0.39

0.51 0.63 0.13

0.179 1.17 0.107 1.36 0.953 0.79

0.16 0.10 0.54

0.97 0.76 0.21

<0.001 1.47 0.004 1.30 0.642 0.28

0.48 0.21 0.71

1.16 0.92 0.24

<0.001 1.85 0.014 1.68 0.768 0.45

0.47 0.15 0.93

Linear mixed models. Mean differences are in standard deviation units. Key: aMCI, amnestic mild cognitive impairment; S/S, TOMM40 homozygous for S; S/VL, TOMM40 heterozygous group; VL/VL, TOMM40 homozygous for VL.

2030

J. Laczó et al. / Neurobiology of Aging 36 (2015) 2024e2033

Fig. 4. Spatial navigation performance across trials. Mean distance errors (SD) from the goal are depicted. Abbreviations: aMCI, amnestic mild cognitive impairment; S, short; VL, very long.

right hippocampal volume (F[2,53] ¼ 2.83; p ¼ 0.074; partial eta2 ¼ 0.16; observed power ¼ 0.42) and other brain structures relevant for spatial navigation (F[2,53]  2.32; p  0.115; partial eta2  0.13; observed power  0.36) was not significant. The post hoc tests showed that the VL/VL group had thinner left entorhinal cortex than the S/S group (p ¼ 0.043), whereas there were no differences between the S/S and S/VL (p ¼ 0.473) or between the S/VL and VL/VL groups (p ¼ 0.267). Furthermore, the VL/VL and S/VL groups had thinner right entorhinal cortex than the S/S group (p ¼ 0.043 and p ¼ 0.031, respectively), whereas there were no differences between the S/VL and VL/VL groups (p ¼ 0.997). Finally, the VL/VL group had thinner left posterior cingulate cortex than the S/S and S/VL groups (p ¼ 0.024 and p ¼ 0.034, respectively), whereas there were no differences between the S/S and S/VL groups (p ¼ 0.859). The between-group comparisons of hippocampal and caudate volumes or thicknesses of cortical brain structures are listed in Table 4. 4. Discussion We examined the influence of TOMM40 “523” polymorphism on spatial navigation in a real-space hMWM and its brain structural correlates using a sample of patients with aMCI with an APOE ε3/ε3 genotype. We found that those who carried at least 1 TOMM40

Table 3 Group-by-subtask interaction of adjusted mean error distances from the goal in spatial navigation subtasks Groupcode

Subtask

Mean

aMCI S/S

Egocentric Allocentric Egocentric Allocentric Egocentric Allocentric

0.27 0.30 0.24 0.67 0.36 0.46

aMCI S/VL aMCI VL/VL

95% Confidence interval for difference Lower bound

Upper bound

0.63 0.66 0.05 0.38 0.01 0.10

0.09 0.06 0.52 0.96 0.72 0.81

Linear mixed models. Mean differences are in standard deviation units. Key: aMCI, amnestic mild cognitive impairment; S/S, TOMM40 homozygous for S; S/VL, TOMM40 heterozygous group; VL/VL, TOMM40 homozygous for VL.

“523” VL poly-T variant performed worse on allocentric spatial navigation tasks than those homozygous for S poly-T variant. The results are particularly intriguing because APOE ε3/ε3 is considered neutral with respect to LOAD risk (Lutz et al., 2010; Roses et al., 2010), thus tapping into a potentially important new genetic pool of patients with aMCI at risk of LOAD. Substantially greater influence of TOMM40 “523” polymorphism on spatial navigation performance was observed in subtasks designed to examine allocentric navigation compared with subtask testing egocentric navigation. The allocentric strategy involves finding the position of a hidden goal using a configuration of landmarks as opposed to own position (egocentric navigation). Allocentric navigation reflects the ability to form, use, and retain a mental representation of the environment, and is one of the first functions impaired in normal aging (Gazova et al., 2013; Rodgers et al., 2012), aMCI (Laczó et al., 2009, 2010, 2012), and AD (Hort et al., 2007; Weniger et al., 2011). Differences in spatial navigation as a function of TOMM40 “523” poly-T length were not found in the egocentric navigation subtask, which involves finding the position of a hidden goal using own position (Maguire et al., 1998) and which has been shown to be impaired later in the course of AD than allocentric navigation (Hort et al., 2007). The key structures for allocentric navigation are the medial temporal lobe structures (Feigenbaum and Morris, 2004). Among the medial temporal lobe structures, allocentric navigation primarily reflects the function of the hippocampus (Maguire et al., 1998) and the entorhinal cortex (Ekstrom et al., 2003), the regions affected early in the course of AD (Braak and Braak, 1991). In the present study, there was a negative moderate correlation between the right and left hippocampal volumes and thicknesses of the right and left entorhinal cortices with the extent of errors in the allocentric subtask, which underscores the importance of the medial temporal lobe structures for allocentric navigation. In addition, the participants showed TOMM40 “523” related differences in the right and left entorhinal cortices, where the individuals with aMCI who carried at least 1 TOMM40 “523” VL poly-T variant had thinner right entorhinal cortex than those homozygous for S poly-T variant. Furthermore, an overall thinner left entorhinal cortex was found in the individuals with aMCI

J. Laczó et al. / Neurobiology of Aging 36 (2015) 2024e2033 Table 4 Group-wise comparisons for adjusted hippocampal and caudate volumes and cortical thickness measures p-Value 95% Confidence interval for (I) Groupcode (J) Groupcode Mean difference difference (IeJ) Lower bound Upper bound Adjusted left hippocampal volumea aMCI S/S aMCI S/VL 0.68 0.231 aMCI VL/VL 0.56 0.493 aMCI S/VL aMCI VL/VL 0.13 0.977 Adjusted right hippocampal volumea aMCI S/S aMCI S/VL 0.74 0.168 aMCI VL/VL 0.96 0.085 aMCI S/VL aMCI VL/VL 0.22 0.891 a Adjusted left caudate volume aMCI S/S aMCI S/VL 0.10 0.994 aMCI VL/VL 0.55 0.546 aMCI S/VL aMCI VL/VL 0.64 0.244 Adjusted right caudate volumea aMCI S/S aMCI S/VL 0.05 0.999 aMCI VL/VL 0.52 0.631 aMCI S/VL aMCI VL/VL 0.57 0.387 Adjusted left entorhinal thickness aMCI S/S aMCI S/VL 0.57 0.473 aMCI VL/VL 1.22 0.043 aMCI S/VL aMCI VL/VL 0.66 0.267 Adjusted right entorhinal thickness aMCI S/S aMCI S/VL 1.22 0.031 aMCI VL/VL 1.29 0.043 aMCI S/VL aMCI VL/VL 0.07 0.997 Adjusted left inferior parietal thickness aMCI S/S aMCI S/VL 0.14 0.964 aMCI VL/VL 0.67 0.215 aMCI S/VL aMCI VL/VL 0.53 0.24 Adjusted right inferior parietal thickness aMCI S/S aMCI S/VL 0.65 0.221 aMCI VL/VL 0.81 0.146 aMCI S/VL aMCI VL/VL 0.16 0.947 Adjusted left precuneus thickness aMCI S/S aMCI S/VL 0.54 0.500 aMCI VL/VL 0.53 0.593 aMCI S/VL aMCI VL/VL 0.01 0.999 Adjusted right precuneus thickness aMCI S/S aMCI S/VL 0.85 0.119 aMCI VL/VL 0.71 0.323 aMCI S/VL aMCI VL/VL 0.14 0.972 Adjusted left posterior cingulate thickness aMCI S/S aMCI S/VL 0.24 0.859 aMCI VL/VL 1.08 0.024 aMCI S/VL aMCI VL/VL 0.83 0.034 Adjusted right posterior cingulate thickness aMCI S/S aMCI S/VL 0.29 0.805 aMCI VL/VL 0.74 0.197 aMCI S/VL aMCI VL/VL 0.45 0.433

0.28 0.52 1.01

1.65 1.63 0.75

0.21 0.10 0.64

1.69 2.02 1.09

1.11 0.59 0.28

0.92 1.68 1.57

1.12 0.68 0.41

1.02 1.71 1.54

0.51 0.03 0.32

1.64 2.42 1.63

0.09 0.03 0.96

2.36 2.56 1.1

0.69 0.26 0.23

0.97 1.60 1.29

0.26 0.2 0.66

1.55 1.82 0.98

0.51 0.64 0.96

1.58 1.70 0.95

0.16 0.42 1.06

1.86 1.83 0.77

0.62 0.12 0.05

1.11 2.04 1.62

0.61 0.26 0.36

1.18 1.74 1.26

Multivariate analysis of covariance adjusted for age and gender. Mean differences are in standard deviation units. Key: aMCI, amnestic mild cognitive impairment; S/S, TOMM40 homozygous for S; S/VL, heterozygous group; VL/VL, TOMM40 homozygous for VL. a Also adjusted for estimated total intracranial volume.

homozygous for VL variant compared with those homozygous for S poly-T variant. This may reflect structural changes of the medial temporal lobe influenced by TOMM40 “523” and corresponds with the recent findings of accelerated atrophy of the right anterior medial temporal lobe with age as a function of TOMM40 “523” VL variant (Johnson et al., 2011). One possibility to explain these findings is that the individuals with TOMM40 “523” VL poly-T variant may experience TOMM40-induced mitochondrial dysfunction and apoptotic processes caused by a high burden of amyloid precursor protein in mitochondrial import channels, which occur primarily in the medial temporal lobe (Devi et al., 2006). This possibility needs further exploration. We did not detect TOMM40 “523” related differences in the right and left hippocampal volumes. The effect sizes for the differences of

2031

the right hippocampal volume were moderate to strong (partial eta2 ¼ 0.16) in our sample, but the observed power was low at 0.42, which probably indicates that our sample size was too small. Another possibility is that the TOMM40 “523” poly-T length may induce mitochondrial dysfunction in the hippocampus, leading to cognitive changes including allocentric navigation impairment, but not to the extent that could be observed in this study as significantly reduced hippocampal volume (Swerdlow and Khan, 2004). In our sample, the participants showed TOMM40 “523” related differences in the left posterior cingulate cortex, where individuals with aMCI homozygous for VL variant had a thinner left posterior cingulate cortex compared with those homozygous for S variant. These results are in agreement with recent findings of left ventral posterior cingulate volume reduction in individuals with VL variants (Johnson et al., 2011) and very recent findings of TOMM40 “523” related differences in white matter integrity of the left cingulate gyrus between individuals with VL and S variants (Lyall et al., 2014). These findings may reflect LOAD-related changes in individuals with aMCI induced by TOMM40 “523” as the posterior cingulate cortex is one of the earliest regions involved in AD pathogenesis (Rowe et al., 2007). Egocentric navigation is processed in the posterior-inferior parietal cortex (Maguire et al., 1998) and precuneus (Weniger et al., 2009) and its simple form, which was not tested in our hMWM test, is processed in the caudate nucleus (Hartley et al., 2003). In the present study, there was a negative moderate correlation between thicknesses of the right and left posterior-inferior parietal cortices with the extent of errors in the egocentric subtask, which highlights the importance of these structures for egocentric navigation in our sample. Furthermore, TOMM40 “523” related differences were found neither in the egocentric navigation nor in thickness of its supporting brain regions, which is in agreement with findings that egocentric navigation is impaired later in the course of AD than allocentric navigation (Hort et al., 2007) and that structural and functional changes in parietal cortical areas occurred later in the course of AD than changes of the medial temporal lobe structures, especially of the entorhinal cortex (de Leon et al., 2001; Zhang et al., 2013). Finally, a negative moderate correlation between thicknesses of the right and left posterior cingulate cortices with extent of errors in the egocentric subtask may reflect the importance of these structures for translation between egocentric and allocentric navigation strategies in our sample. One of the strengths of this study is the fact that it is the first study to date to focus on the influence of TOMM40 “523” polymorphism on spatial navigation in patients with aMCI. Spatial navigation is a frequently neglected cognitive function that is affected very early in the course of AD (Hort et al., 2007). In addition, we used a real-space hMWM, a well-established method mimicking navigation in the real world, to examine spatial navigation. Finally, we included measurement of brain structures relevant for spatial navigation and affected very early in the course of AD (Killiany et al., 2000; Kogure et al., 2000), which may be influenced by TOMM40 “523” polymorphism. This study also has limitations. The number of subjects in S/S and VL/VL groups was relatively small. This may limit power to detect statistically significant associations, especially differences in morphometric characteristics of the brain structures as indicated by computed observed power. Since the age range of the participants was 62e86 years, we were unable to investigate the early deleterious effects of the VL allele, which may occur before the age of 60 years (Caselli et al., 2012). Finally, this was a cross-sectional study, which does not allow for tracking of aMCI patients to clinical endpoints such as conversion to dementia because of AD, but longitudinal follow-up is ongoing.

2032

J. Laczó et al. / Neurobiology of Aging 36 (2015) 2024e2033

5. Conclusion Among APOE ε3/ε3 homozygotes with aMCI, those carrying at least 1 TOMM40 “523” poly-T VL variant showed greater allocentric, but not egocentric, spatial navigation impairment compared with those homozygous for S variants. These findings indicate that TOMM40 “523” may play a specific and important role in spatial navigation in the aMCI population. Our findings also indicate that, in individuals with aMCI, TOMM40 “523” poly-T VL variants could be related to thinner entorhinal and posterior cingulate cortices. These structures are very important for spatial, especially allocentric, navigation and are affected very early in the course of AD (Killiany et al., 2000; Kogure et al., 2000). Spatial navigation testing may have a potential to identify individuals with incipient AD among heterogeneous population of individuals with aMCI and may thus help to stratify subjects in clinical trials with anti-Alzheimer drugs. The focus of future studies should be to evaluate the effect of the TOMM40 “523” polymorphism on longitudinal spatial navigation changes in aMCI individuals and spatial navigation performance in the broader MCI population such as those with non-aMCI and also in asymptomatic individuals between middle and older age groups. Our findings underline a specific role of TOMM40 “523” genotype in the pathogenesis of LOAD. Disclosure statement Dr Laczó has consulted for Pfizer and between June 2012 and May 2014 held shares of Polyhymnia-TS. Dr Hort has consulted for Pfizer, Janssen, Merck, Axon, Sotio, Novartis, Elan, Zentiva, Ipsen and between June 2012 and May 2014 held shares of Polyhymnia-TS. Dr Andel, Dr Vyhnalek, Dr Matoska, Mr. Kaplan, Dr Nedelska, Mr. Lerch, Dr Gazova, and Dr Moffat report no disclosures. Acknowledgements We would like to thank Dr K. Vl cek for technical support, Dr J. Cerman, Dr I. Mokrisová, Dr E. Hyn cicová, and Dr H. Marková for help with data collection, Mr. P. Kala and Ms. B. Zemlickova for preparing the MR data and Prof. C. H. Rhodes for helpful comments. This study was supported by Grant Agency of Charles University in Prague Grants No. 624012, 546113, and 1108214; European Regional Development FunddProject FNUSA-ICRC (No. CZ.1.05/1.1.00/02.0123) and by project ICRC-ERA-HumanBridge (No. 316345); European Social Fund and the State Budget of the Czech Republic; European Social Fund within the project Young Talent Incubator II (Reg. No. CZ.1.07/ 2.3.00/20.0117); Ministry of Health, Czech Republicdconceptual development of research organization, University Hospital Motol, Prague, Czech Republic 00064203; Institutional Support of Laboratory Research Grant No. 2/2012 (699002); research projects AV0Z50110509 and RVO:67985823. NeuGRID4you portal for cloud computing that we used to perform FreeSurfer analyses, has received funding from the European Commission’s Seventh Framework Programme (FP7/2007e2013) under grant agreement No. 283562. Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.neurobiolaging. 2015.03.004. References Abbott, A., Callaway, E., 2014. Prize for place cells. Nature 514, 153. Benke, T., Karner, E., Petermichl, S., Prantner, V., Kemmler, G., 2014. Neuropsychological deficits associated with route learning in Alzheimer disease, MCI, and normal aging. Alzheimer Dis. Assoc. Disord. 28, 162e167.

Berteau-Pavy, F., Park, B., Raber, J., 2007. Effects of sex and APOE ε4 on object recognition and spatial navigation in the elderly. Neuroscience 147, 6e17. Braak, H., Braak, E., 1991. Neuropathological staging of Alzheimer related changes. Acta Neuropathol. 82, 239e259. Byrne, P., Becker, S., Burgess, N., 2007. Remembering the past and imagining the future: a neural model of spatial memory and imagery. Psychol. Rev.114, 340e375. Caselli, R.J., Dueck, A.C., Huentelman, M.J., Lutz, M.W., Saunders, A.M., Reiman, E.M., Roses, A.D., 2012. Longitudinal modeling of cognitive aging and the TOMM40 effect. Alzheimers Dement. 8, 490e495. Cramér, H., 1999. Mathematical methods of statistics (PMS-9), Vol. 9. Princeton University Press, Princeton, NJ. Cushman, L.A., Stein, K., Duffy, C.J., 2008. Detecting navigational deficits in cognitive aging and Alzheimer disease using virtual reality. Neurology 71, 888e895. deIpolyi, A.R., Rankin, K.P., Mucke, L., Miller, B.L., Gorno-Tempini, M.L., 2007. Spatial cognition and the human navigation network in AD and MCI. Neurology 69, 986e997. de Leon, M.J., Convit, A., Wolf, O.T., Tarshish, C.Y., DeSanti, S., Rusinek, H., Tsui, W., Kandil, E., Scherer, A.J., Roche, A., Imossi, A., Thorn, E., Bobinski, M., Caraos, C., Lesbre, P., Schlyer, D., Poirier, J., Reisberg, B., Fowler, J., 2001. Prediction of cognitive decline in normal elderly subjects with 2-[18F]fluoro-2-deoxy-D-glucose/positronemission tomography (FDG/PET). Proc. Natl. Acad. Sci. U. S. A. 98, 10966e10971. Devi, L., Prabhu, B.M., Galati, D.F., Avadhani, N.G., Anandatheerthavarada, H.K., 2006. Accumulation of amyloid precursor protein in the mitochondrial import channels of human Alzheimer’s disease brain is associated with mitochondrial dysfunction. J. Neurosci. 26, 9057e9068. Ekstrom, A.D., Kahana, M.J., Caplan, J.B., Fields, T.A., Isham, E.A., Newman, E.L., Fried, I., 2003. Cellular networks underlying human spatial navigation. Nature 425, 184e188. Feigenbaum, J.D., Morris, R.G., 2004. Allocentric versus egocentric spatial memory after unilateral temporal lobectomy in humans. Neuropsychology 18, 462e472. Fischl, B., van der Kouwe, A., Destrieux, C., Halgren, E., Segonne, F., Salat, D.H., Busa, E., Seidman, L.J., Goldstein, J., Kennedy, D., Caviness, V., Makris, N., Rosen, B., Dale, A.M., 2004. Automatically parcellating the human cerebral cortex. Cereb. Cortex 14, 11e22. Gazova, I., Laczó, J., Rubinova, E., Mokrisova, I., Hyncicova, E., Andel, R., Vyhnalek, M., Sheardova, K., Coulson, E.J., Hort, J., 2013. Spatial navigation in young versus older adults. Front. Aging Neurosci. 5, 94. Hartley, T., Maguire, E.A., Spiers, H.J., Burgess, N., 2003. The well-worn route and the path less travelled: distinct neural bases of route following and way-finding in humans. Neuron 37, 877e888. Hayden, K.M., McEvoy, J.M., Linnertz, C., Attix, D., Kuchibhatla, M., Saunders, A.M., Lutz, M.W., Welsh-Bohmer, K.A., Roses, A.D., Chiba-Falek, O., 2012. A homopolymer polymorphism in the TOMM40 gene contributes to cognitive performance in aging. Alzheimers Dement. 8, 381e388. Hixson, J.E., Vernier, D.T., 1990. Restriction isotyping of human apolipoprotein E by gene amplification and cleavage with Hha I. J. Lipid Res. 31, 545e548. Holm, S., 1979. A simple sequentially rejective multiple test procedure. Scand. J. Stat. 6, 65e70. Hort, J., Laczó, J., Vyhnalek, M., Bojar, M., Bures, J., Vlcek, K., 2007. Spatial navigation deficit in amnestic mild cognitive impairment. Proc. Natl. Acad. Sci. U. S. A.104, 4042e4047. Humphries, A.D., Streimann, I.C., Stojanovski, D., Johnston, A.J., Yano, M., Hoogenraad, N.J., Ryan, M.T., 2005. Dissection of the mitochondrial import and assembly pathway for human Tom40. J. Biol. Chem. 280, 11535e11543. Jack Jr., C.R., Twomey, C.K., Zinsmeister, A.R., Sharbrough, F.W., Petersen, R.C., Cascino, G.D., 1989. Anterior temporal lobes and hippocampal formations: normative volumetric measurements from MR images in young adults. Radiology 172, 549e554. Johnson, S.C., La Rue, A., Hermann, B.P., Xu, G., Koscik, R.L., Jonaitis, E.M., Bendlin, B.B., Hogan, K.J., Roses, A.D., Saunders, A.M., Lutz, M.W., Asthana, S., Green, R.C., Sager, M.A., 2011. The effect of TOMM40 poly-T length on gray matter volume and cognition in middle-aged persons with APOE ε3/ε3 genotype. Alzheimers Dement. 7, 456e465. Killiany, R.J., Gomez-Isla, T., Moss, M., Kikinis, R., Sandor, T., Jolesz, F., Tanzi, R., Jones, K., Hyman, B.T., Albert, M.S., 2000. Use of structural magnetic resonance imaging to predict who will get Alzheimer’s disease. Ann. Neurol. 47, 430e439. Kogure, D., Matsuda, H., Ohnishi, T., Asada, T., Uno, M., Kunihiro, T., Nakano, S., Takasaki, M., 2000. Longitudinal evaluation of early Alzheimer’s disease using brain perfusion SPECT. J. Nucl. Med. 41, 1155e1162. Laczó, J., Vlcek, K., Vyhnalek, M., Vajnerova, O., Ort, M., Holmerova, I., Tolar, M., Andel, R., Bojar, M., Hort, J., 2009. Spatial navigation testing discriminates two types of amnestic mild cognitive impairment. Behav. Brain Res. 202, 252e259. Laczó, J., Andel, R., Vyhnalek, M., Vlcek, K., Magerova, H., Varjassyova, A., Tolar, M., Hort, J., 2010. Human analogue of the Morris water maze for testing subjects at risk of Alzheimer’s disease. Neurodegener. Dis. 7, 148e152. Laczó, J., Andel, R., Vlcek, K., Macoska, V., Vyhnalek, M., Tolar, M., Bojar, M., Hort, J., 2011. Spatial navigation and APOE in amnestic mild cognitive impairment. Neurodegener. Dis. 8, 169e177. Laczó, J., Andel, R., Vyhnalek, M., Vlcek, K., Magerova, H., Varjassyova, A., Nedelska, Z., Gazova, I., Bojar, M., Sheardova, K., Hort, J., 2012. From Morris Water Maze to computer tests in the prediction of Alzheimer’s disease. Neurodegener. Dis. 10, 153e157. Laczó, J., Andel, R., Vyhnalek, M., Vlcek, K., Nedelska, Z., Matoska, V., Gazova, I., Mokrisova, I., Sheardova, K., Hort, J., 2014. APOE and spatial navigation in amnestic MCI: results from a computer-based test. Neuropsychology 28, 676e684.

J. Laczó et al. / Neurobiology of Aging 36 (2015) 2024e2033 Lithfous, S., Dufour, A., Després, O., 2013. Spatial navigation in normal aging and the prodromal stage of Alzheimer’s disease: insights from imaging and behavioral studies. Ageing Res. Rev. 12, 201e213. Lutz, M.W., Crenshaw, D.G., Saunders, A.M., Roses, A.D., 2010. Genetic variation at a single locus and age of onset for Alzheimer’s disease. Alzheimers Dement. 6, 125e131. Lyall, D.M., Harris, S.E., Bastin, M.E., Muñoz Maniega, S., Murray, C., Lutz, M.W., Saunders, A.M., Roses, A.D., Valdés Hernández Mdel, C., Royle, N.A., Starr, J.M., Porteous, D.J., Wardlaw, J.M., Deary, I.J., 2014. Alzheimer’s disease susceptibility genes APOE and TOMM40, and brain white matter integrity in the Lothian Birth Cohort 1936. Neurobiol. Aging 35, 1513.e25e1513.e33. Maguire, E.A., Burgess, N., Donnett, J.G., Frackowiak, R.S., Frith, C.D., O’Keefe, J., 1998. Knowing where and getting there: a human navigation network. Science 280, 921e924. Morris, J.C., 1993. The clinical dementia rating (CDR): current version and scoring rules. Neurology 43, 2412e2414. Mosconi, L., Brys, M., Switalski, R., Mistur, R., Glodzik, L., Pirraglia, E., Tsui, W., De Santi, S., de Leon, M.J., 2007. Maternal family history of Alzheimer’s disease predisposes to reduced brain glucose metabolism. Proc. Natl. Acad. Sci. U. S. A. 104, 19067e19072. Nedelska, Z., Andel, R., Laczó, J., Vlcek, K., Horinek, D., Lisy, J., Sheardova, K., Bures, J., Hort, J., 2012. Spatial navigation impairment is proportional to right hippocampal volume. Proc. Natl. Acad. Sci. U. S. A. 109, 2590e2594. Ohnishi, T., Matsuda, H., Hirakata, M., Ugawa, Y., 2006. Navigation ability dependent neural activation in the human brain: an fMRI study. Neurosci. Res. 55, 361e369. Petersen, R.C., 2004. Mild cognitive impairment as a diagnostic entity. J. Intern. Med. 256, 183e194. Reiman, E.M., Caselli, R.J., Yun, L.S., Chen, K., Bandy, D., Minoshima, S., Thibodeau, S.N., Osborne, D., 1996. Preclinical evidence of Alzheimer’s disease in persons homozygous for the epsilon 4 allele for apolipoprotein E. N. Engl. J. Med. 334, 752e758. Rodgers, M.K., Sindone 3rd, J.A., Moffat, S.D., 2012. Effects of age on navigation strategy. Neurobiol. Aging 33, 202.e15e202.e22. Roses, A.D., Lutz, M.W., Amrine-Madsen, H., Saunders, A.M., Crenshaw, D.G., Sundseth, S.S., Huentelman, M.J., Welsh-Bohmer, K.A., Reiman, E.M., 2010.

2033

A TOMM40 variable-length polymorphism predicts the age of late-onset Alzheimer’s disease. Pharmacogenomics J. 10, 375e384. Rowe, C.C., Ng, S., Ackermann, U., Gong, S.J., Pike, K., Savage, G., Cowie, T.F., Dickinson, K.L., Maruff, P., Darby, D., Smith, C., Woodward, M., Merory, J., Tochon-Danguy, H., O’Keefe, G., Klunk, W.E., Mathis, C.A., Price, J.C., Masters, C.L., Villemagne, V.L., 2007. Imaging beta-amyloid burden in aging and dementia. Neurology 68, 1718e1725. Saunders, A.M., Hulette, O., Welsh-Bohmer, K.A., Schmechel, D.E., Crain, B., Burke, J.R., Alberts, M.J., Strittmatter, W.J., Breitner, J.C., Rosenberg, C., 1996. Specificity, sensitivity, and predictive value of apolipoprotein-E genotyping for sporadic Alzheimer’s disease. Lancet 348, 90e93. Serino, S., Cipresso, P., Morganti, F., Riva, G., 2014. The role of egocentric and allocentric abilities in Alzheimer’s disease: a systematic review. Ageing Res. Rev. 16, 32e44. Singer, J.D., Willett, J.B., 2003. Applied Longitudinal Data Analysis: Modeling Change and Event Occurrence. Oxford University Press, New York, NY. Swerdlow, R.H., Khan, S.M., 2004. A “mitochondrial cascade hypothesis” for sporadic Alzheimer’s disease. Med. Hypotheses 63, 8e20. Tabachnick, B.G., Fidell, L.S., 2007. Using Multivariate Statistics. Allyn & Bacon, Boston, MA. Vl cek, K., Laczó, J., 2014. Neural correlates of spatial navigation changes in mild cognitive impairment and Alzheimer’s disease. Front. Behav. Neurosci. 8, 89. Weniger, G., Ruhleder, M., Wolf, S., Lange, C., Irle, E., 2009. Egocentric memory impaired and allocentric memory intact as assessed by virtual reality in subjects with unilateral parietal cortex lesions. Neuropsychologia 47, 59e69. Weniger, G., Ruhleder, M., Lange, C., Wolf, S., Irle, E., 2011. Egocentric and allocentric memory as assessed by virtual reality in individuals with amnestic mild cognitive impairment. Neuropsychologia 49, 518e527. Xu, W.L., Caracciolo, B., Wang, H.X., Santoni, G., Winblad, B., Fratiglioni, L., 2013. Accelerated progression from mild cognitive impairment to dementia among APOE ε4ε4 carriers. J. Alzheimers Dis. 33, 507e515. Yesavage, J.A., 1988. Geriatric depression scale. Psychopharmacol. Bull. 24, 709e711. Zhang, Y., Schuff, N., Camacho, M., Chao, L.L., Fletcher, T.P., Yaffe, K., Woolley, S.C., Madison, C., Rosen, H.J., Miller, B.L., Weiner, M.W., 2013. MRI markers for mild cognitive impairment: comparisons between white matter integrity and gray matter volume measurements. PLoS One 8, e66367.