Accelerated age-related cortical thinning in healthy carriers of apolipoprotein E ɛ4

Abstract

Effects of APOE genotype on age-related slopes of cortical thinning was estimated by measuring the thickness of the cerebral cortex on a point-by-point basis across the cortical mantle in 96 healthy non-demented volunteers aged 48–75 years. Fifty nine were APOE ɛ4− (no ɛ4 allele) and 37 were ɛ4+ (1 or 2 ɛ4 alleles). The genotype groups had similar age, sex and IQ. Two T₁-weighted MP-RAGE sequences were averaged for each participant to yield images with high signal-to-noise ratio, and quantified using semi-automated analysis tools. ɛ4 carriers had thicker cortex than non-carriers in several frontal and temporal areas in both hemispheres, but showed a steeper age-related decline in adjacent areas. Upon comparison of the ɛ4-specific age-related thinning with previously published patterns of thinning in normal aging and Alzheimer's disease (AD), we conclude that APOE ɛ4 may function to accelerate thinning in areas found to decline in aging (medial prefrontal and pericentral cortex), but also to initiate thinning in areas associated with AD and amyloid-β aggregation (occipitotemporal and basal temporal cortex).

Introduction

Gray and white matter volumes shrink with age (Bartzokis et al., 2003, Resnick et al., 2003) and the efficiency of many perceptual and cognitive functions decline in adult aging (e.g. Dobbs and Rule, 1989, Salthouse, 1996). Despite this general trend, there is considerable individual variation, and although many environmental factors modulate brain aging (Raz et al., 2005) genes also play a major role. Twin studies have shown that genetic factors contribute substantially to normal variation in brain volume (Toga and Thompson, 2005) and that heritability is generally high also in old age, but differs between structures (Pfefferbaum et al., 2001, Pfefferbaum et al., 2000, Sullivan et al., 2001). Candidate gene allelic association methods have been used to address the issue of specific genes involved in individual variation, and one of the most widely studied genes is the apolipoprotein E gene (APOE). The three APOE alleles, ɛ2, ɛ3 and ɛ4, produce three isoforms of a protein (ApoE) that delivers lipids to neurons in the service of synaptogenesis (Mauch et al., 2001). Inheritance of ɛ4 is associated with increased risk of AD in a gene dose dependent manner (Corder et al., 1993). In animals, the ɛ4 allele has been associated with impaired repair mechanisms following lesions (Teter et al., 2002). In humans, the ɛ4 is associated with reduced brain integrity following many kinds of neural insult (Fazekas et al., 2001, Sundstrom et al., 2004) and ɛ4 is also associated with reduction in cerebral glucose metabolism (Reiman et al., 1996) and reduced cognitive functioning in healthy middle-aged adults (Flory et al., 2000, Greenwood et al., 2000). It has been suggested that APOE may be primarily characterized as a gene involved in neuronal plasticity and repair, and only secondarily as an AD susceptibility gene (Greenwood and Parasuraman, 2003, Teter et al., 2002). Consistent with this, Reiman et al. (2004) reported an association between APOE ɛ4 and regional patterns of reduction in cerebral glucose metabolism in healthy volunteers as young as the 20s and 30s.

A growing literature shows that APOE ɛ4 modulates brain morphology in healthy elderly participants, including hippocampal volume (Cohen et al., 2001, Den Heijer et al., 2002, Lemaître et al., 2005, Moffat et al., 2000, Plassman et al., 1997) and amygdalar volume (Den Heijer et al., 2002). Two studies have also shown that APOE ɛ4 modulates the rate of hippocampal atrophy with increasing age (Cohen et al., 2001, Moffat et al., 2000), and behavioral results have also shown steeper age-related decline in cognitive functions for ɛ4 carriers (Bretsky et al., 2003, Wilson et al., 2002). Morphological studies have so far emphasized measures of memory-related subcortical structures, presumably because APOE ɛ4 has been thought of primarily as a susceptibility gene for AD, where memory disorder is a cardinal trait. However, functional MRI studies using memory tasks have also shown effects of APOE variation on cortical activation (Bondi et al., 2005, Bookheimer et al., 2000, Lind et al., 2006), and behavioral studies indicate APOE-associated modulation of several cognitive functions known to be dependent on cortical integrity (Espeseth et al., 2006, Greenwood et al., 2005a, Greenwood et al., 2000, Rosen et al., 2002).

Since APOE has been implied as the major susceptibility gene for late-onset AD (Corder et al., 1993), it is important to understand the effects of this gene on regional thickness of the cerebral cortex in older age cohorts. It is well known that widespread thinning of both primary as well as association cortex occurs from middle age and upwards (Salat et al., 2004). Salat et al. (2004) reported prominent age-related thinning in the prefrontal cortex, including areas close to the primary motor and premotor cortex, and also in the calcarine cortex near the striate cortex. Other regions such as the temporal lobes and areas in or near the anterior cingulate cortex seemed to be relatively spared. In fact, regions of the anterior cingulate and some medial orbitofrontal regions revealed significant age-related thickening. While prefrontal thinning and temporal cortex sparing is consistent with previous reports on age-related volume reduction, thinning of motor, premotor and calcarine cortex, and thickening of anterior cingulate cortex were novel findings (see Raz, 2005 for a review). Thus, strict conclusions regarding the influence of age on these cortical areas cannot be drawn at the present stage. However, since the present study focuses on cortical thickness and also employ the same methodology as Salat et al. (2004), we will compare our results with published reports on cortical thickness rather than cortical volume. There are also widespread morphometric effects associated with AD (Lerch et al., 2005). For example, Lerch et al. (2005) showed that AD patients had significantly thinner cortex in anterior and posterior cingulate cortex, the left dorsolateral prefrontal cortex and most of the temporal lobes. Comparison of the results from Salat et al. (2004) and Lerch et al. (2005) reveal that there is not complete overlap between regional patterns of cortical thinning in aging and AD. This fact suggests that it may be worthwhile to directly compare patterns of APOE ɛ4-related patterns of cortical thickness with age- or AD-related cortical thinning as reported in the literature. Conceivably, if APOE ɛ4 has an impact on cortical thickness, it could primarily act by speeding up thickness changes in areas reported to change in normal aging, or alternatively, could modulate cortical thickness in areas associated with AD-related cortical thinning.

As part of a larger study on genetic effects on several cognitive and brain parameters we obtained MRI scans from 96 persons aged 48–75 years and submitted these to morphometric analyses. We made use of a methodology that has been validated with histological (Rosas et al., 2002) and manual measurement (Kuperberg et al., 2003), and has proven capable of showing developmental changes in cortical thickness in aging samples (Salat et al., 2004) and different developmental trajectories over the life span dependent on cognitive abilities (Fjell et al., 2006). Thus, measures of cortical thickness are sensitive to changes in normal aging and may have predictive value on successful aging. Since this study is the first to test the effect of APOE genetic variations of cortical morphometry in an older age-cohort, it is important not to restrict the analyses to predefined areas. Thus, the effect of APOE on cerebral cortical thickness was tested on a point-by-point basis across the entire cortical mantle. The present sample's age range and the methodology were chosen to reveal possible age by genotype interactions on local variation in thickness of the cortical mantle in the life period where early morphological changes are beginning to emerge in the subset of individuals who are at risk for developing AD or other types of dementia. The main questions investigated are whether APOE isoforms differently affect cortical thickness and age-related changes in cortical thickness, and whether such changes are mostly similar to normal age-changes or changes that are found in pathological populations, such as participants suffering from AD.