**6.1 The metabolism hypothesis and cognitive aging**

The metabolism hypothesis suggests that if chronically elevated levels of resting brain activity over the lifetime drive Aß deposition with attendant cognitive dysfunction leading to AD, there should be amyloid deposition during healthy aging as well. Based on this mechanism, the PCC region should show major hypometabolism, atrophy, and amyloid deposition as seen in early AD [21, 86–88].

Yet, this phenomenon is not observed. The PCC in normal aging shows relative preservation (**Figure 3**). Among the regions showing the least decline in metabolism with aging is the PCC. Older healthy adults show minimal PCC atrophy rates over 12 months [89]. Older healthy adults, especially E4 non-carriers, do not show amyloid deposition in the PCC [90–92]. Young adult E4 carriers with positive family history of AD and at high risk of future AD already show PCC hypometabolism implying DMN hyperactivity related to AD must have occurred before then [93]. Of note, older healthy E4 non-carriers begin to show amyloid positivity at around 71 years of age, while the E4 carriers develop amyloid positivity about 20 years earlier. Interestingly, when separating the independent effects of aging vs. E4 load, amyloid deposition shows a more frontal involvement. Of note, the effects of aging and E4 load interact: the peak hazard ratio occurs ~60 years of age and declines thereafter; E4 is a risk factor for AD even for younger adults (<65 years) [94]. Furthermore, resting connectivity of the PCC/precuneus region to the ACC is reduced in older healthy adults who carry E4 even in the absence of detectable fibrillar amyloid or decreased CSF Aß42 suggesting both Aß-dependent and Aß-independent aging-related mechanisms [95].

The principal locus of declining metabolism in healthy elders does not map to the PCC but localizes instead to the ACC (**Figure 3**) [96–101]. ACC hypometabolism correlates also with aging-related decline in cognitive function [99]. Whereas

#### **Figure 3.**

*Redirecting Alzheimer Strategy - Tracing Memory Loss to Self Pathology*

decreased redox buffering) leading to accelerated tauopathy.

AD rises relentlessly with aging and oxidative stress.

refined the metabolism hypothesis of AD.

**5. Role of DMN and amyloid**

AG [73]. As summarized above, the PCC is very susceptible to amyloid deposition and is among the earliest dysfunctional regions in AD. These observations further

The default network shows overlap with brain regions high in AG which in turn show overlap with areas of amyloid deposition in AD [74]. Unlike oxidative phosphorylation used to generate energy, AG proceeds less efficiently energetically (2 ATP vs. 38 ATP per glucose molecule) but more suitably for reduction of biomolecules for anabolism [75]. Anabolism that appears to play a much greater role in early human development could also provide, albeit to a lesser extent, the substrates for plasticity related to learning and memory in adults [76, 77]. The metabolism hypothesis is important because it motivates the search for AD pathophysiology beyond amyloid deposition to some aspect of cerebral metabolism particularly AG. Normal aging is associated with the loss of AG in regions which sustain higher levels of AG in youth; these are the very regions showing susceptibility to amyloid deposition [7]. The metabolism hypothesis could help explain why the frequency of

Several lines of evidence support the metabolism hypothesis of AD through altered processing of Aß [78]. The processing pathways include both increased production and decreased clearance. Most AD-causing dominant mutations in APP, PSEN1, and PSEN2 increase Aß production [79]. One mutation that is protective for AD occurs near the APP BACE1 cleavage site impairing γ cleavage; it is associated in vitro with decreased amyloidogenic peptides [80]. Likewise, vibrissal stimulation of APP transgenic mice increases Aß in interstitial CSF and amyloid plaques while decreasing lactate, a proxy for neural activity [81]. In-vitro mouse slice preparations show based on microdialysis rapid increases in Aß correlated with synaptic activity [82]. In cognitively normal older adults, greater hippocampal activity during encoding at baseline correlates with longitudinal amyloid deposition and diminished cognitive performance [83]. APOE, the major risk locus for AD, plays a key role in Aß aggregation, fibrillogenesis, and maturation of neuritic plaques [84]. AD patients relative to controls have decreased clearance of CSF Aß with normal rate of

Here, issues informing discussions about the amyloid hypothesis of AD and relationships to cerebral metabolism are outlined. These raise questions about the strong version of the metabolism hypothesis and its implications, or at least suggest a need for revision. Significant circumstantial evidence centers on several

In a group of mostly cognitively intact elders, those globally without amyloid did not have tau accumulation in areas prone toward tau deposits (precuneus, amygdala, entorhinal, inferior temporal, inferior and superior parietal, fusiform, and lateral occipital cortices) and did not have decreased CMRO or AG [69]. They showed a positive correlation between AG and CMRglu; no correlations surfaced between CMRO, CMRglu, or tau deposition. In contrast, those who were amyloid positive globally showed an inverse relationship between tau and AG but not between tau and CMRO or CMRglu. These data suggest the loss of AG in tau-prone regions with tau accumulation leads to decreased plasticity and decreased neuroprotection (i.e.,

**20**

Aß production [85].

**6. Gaps in the metabolism hypothesis**

*Decline of brain activity with aging in healthy volunteers. Voxel-wise Pearson correlation (r) map of glucose uptake vs. age in stereotactically normalized brain (3D-SSP; S. Minoshima, University of Utah). A, lateral; B, medial; C, dorsal (left)/ventral (right). Color scale shows peak r = −0.8. Note the greatest decline in glucose uptake with age localizes to the ACC (BA 32/9; region 1). A midline circuit (regions 1-3) includes dorsomedial thalamus and basal forebrain/subgenual cingulate; the metabolism in this circuit correlates with declining executive function (verbal fluency). Other regions with lower correlations are not associated with cognitive performance. Reprinted from Neuroimage, Vol 35(3), Pardo JV et al., Where the brain grows old: Decline in anterior cingulate and medial prefrontal function with normal aging. Copyright (2007), with permission from Elsevier.*

AG localizes to ACC, PCC, and parietal regions, loss of regions high in AG during youth appears to occur in all three regions without selectivity for any one of these regions [76]. So, unless AD per se involves hyperactivity of the DMN beyond that in normal elders for which there is no evidence, the hypothesis does not address the inconsistency between mechanisms of cognitive aging versus AD to account for the observed dissociation between ACC and PCC findings.

Several observations related to cognitive aging need reconciliation with the metabolism hypothesis. Healthy elders free from amyloid deposition show a remarkable disconnection between the anterior and posterior default networks (i.e., ACC and PCC) [102]. Yet, nothing about the metabolism hypothesis explains why DMN regions with high AG, where amyloid will be deposited as AD develops, should disconnect—both regions should show aging-related hypometabolism as a result of white matter damage. Similarly, cognitively normal elders not at high risk for AD (no E4) show increases and decreases within anterior DMN connectivity, while showing only decreases in posterior DMN FC [103]. Cognitively intact elders with minimal amyloid deposition without E4 have greater connectivity of the ACC to the precuneus than those with E4 [95]. The aging-related anterior vs. posterior dissociations in connectivity within DMN networks remain theoretically difficult to predict based solely on the metabolism hypothesis.

The ACC also has high glucose metabolism, flow, oxygen consumption, and AG; yet, the ACC does not show amyloid deposition akin to the PCC with healthy aging (those >60 years without amyloid positive scans [23]). Thus, chronic neural activity along with AG during the lifespan per se is not sufficient to lay down amyloid. Just as PCC hypometabolism marks focal atrophy early in AD, the region of ACC hypometabolism with aging should likewise display cortical thinning. However, several large studies do not support the prediction [104, 105], although not all findings are convergent [106]. Additionally, recent studies of tau deposition in preclinical AD show early deposition in the PCC but not in the ACC [107]. Those elders with cognitive function akin to much younger subjects (i.e., "SuperAgers") show thickening of the ACC and increased spindle cells suggestive of plasticity with aging; or alternatively, "SuperAgers" may be endowed with ACC thickening before aging [108, 109]. Also, age can confound years of education; the latter is associated with increased ACC thickness and metabolism [110, 111]. However, studies of the effects of aerobic fitness exercise on cognition and cortex show ACC thickening in older adults in support of the potential for plasticity in this region [112].

A clear dissociation can arise also between amyloid deposition and FDG metabolism in the ACC in AD that is difficult to explain with the metabolism hypothesis. Patients who initially had mild AD, as confirmed with metabolic and amyloid biomarkers, were followed for 2 years along with a matched, amyloid negative control group. Despite extensive amyloid deposition in the ACC of the AD patients, no hypometabolism colocalized in the ACC during follow-up [22]. The agingrelated ACC hypometabolism noted previously likely led to a floor effect across groups. Likewise, aging-related ACC hypometabolism would tend to spare amyloid deposition in patients appearing inconsistent with the metabolism hypothesis given extensive amyloid involvement of the ACC in AD.

#### **6.2 The metabolism hypothesis and autosomal dominant AD**

A corollary of the metabolism hypothesis suggests those with familial AD would show similar patterns of hypometabolism and amyloid deposition to late-onset, sporadic AD. However, those with mutations (APP, PSEN1, PSEN2) show greatest amyloid deposition in the basal ganglia, a site that only becomes involved late in typical sporadic AD [113–115]. However, the specific pattern of

**23**

transgenes.

*Fact, Fiction, or Evolution: Mechanism Hypothesis of Alzheimer's Disease*

**6.3 Role of APOE genotype in the metabolism hypothesis**

**6.4 Role of tau in the metabolism hypothesis**

and provides useful staging of disease progression [17].

amyloid deposition may depend to some extent on the specific mutation. For example, the PS1 mutation, E280A, shows amyloid deposition more like lateonset sporadic AD than many other mutations with early onset. This variability may not surprise given the complexity of the underlying biology of different

If regional brain activity drives amyloid deposition in the pattern seen in AD, then the AD metabolic pattern should arise in those at highest risk for the future development of AD—asymptomatic *APOE\*E4* homozygotes, who have a 12-fold increased risk of LOAD. So far, there are seven such individuals in the Alzheimer's Disease Neuroimaging Initiative's (ADNI) database. Their pattern of amyloid deposition highlights bilateral lenticular nuclei and the ACC/medial prefrontal involvement with the PCC notably unaffected (see **Figure 3**). In fact, the deposition of amyloid in the homozygotes is reminiscent of that seen in Down's syndrome and most mutations found in autosomal dominantly inherited forms of AD arising in APP, PS1, or PS2 [113–115, 117]. This pattern of amyloid deposition in E4 homozygotes is consistent with findings reported previously in an independent group of eight homozygotes [91].

Another difficulty with the metabolism hypothesis of AD is the notable absence of tau involvement in this theory. The role of tau, its modifications, and its etiologic role in neurodegeneration in AD has been reviewed previously [118]. Although the metabolism hypothesis of AD focuses on amyloid deposition, there is increasing evidence that tau plays at least as great if not greater etiopathological role. Of interest in this context, recent studies show tau deposition during preclinical AD in the PCC; the metabolism hypothesis cannot explain this dissociation between PCC and ACC [107]. AD cases with neurofibrillary changes (neuritic plaques, neurofibrillary tangles, neurofibrillary threads, tau tangles) typically show extensive amyloid deposition. However, not all cases with extensive amyloid deposition show neurofibrillary changes [17]. Amyloid deposits and neuritic plaques vary widely across individuals both temporally and regionally; so, they do not provide useful biomarkers for staging of AD [17]. In contrast, the distribution of tau is consistent across individuals

Hyper-phosphorylated intraneuronal tau ("pretangle") has been reported even in young adults in the absence of amyloid particularly in subcortical nuclei such as the locus coeruleus [119]. The significance of these findings in the context of AD remains uncertain as the pretangle material may be transient, related incidentally to other processes (e.g., traumatic brain injury), or the earliest manifestations of AD. Furthermore, studies of transgenic mice with APOE isoform knock-in and APOE knock-out show that even in the absence of amyloid, E4 is particularly neurotoxic in mice with mutant tau transgenes, and this toxicity is in part mediated by neuroinflammation via the innate immune system produced by microglia and type A1 astrocytes [120].

A recent study identified the significant role of tau in the context of amyloid deposition [121]. High resolution fMRI of cerebral blood volume (CBV), coupled to regional metabolism, mapped the earliest changes in preclinical AD to lateral entorhinal (LEC), transentorhinal, and perirhinal cortices, as predicted from neuropathological studies [17]. The former region's CBV correlated significantly with a test of delayed retention. Three lines of mice were generated with differential expression in entorhinal cortex of pathological human APP, tau, or both

*DOI: http://dx.doi.org/10.5772/intechopen.83824*

mutations in PS1 [116].

*Redirecting Alzheimer Strategy - Tracing Memory Loss to Self Pathology*

observed dissociation between ACC and PCC findings.

predict based solely on the metabolism hypothesis.

extensive amyloid involvement of the ACC in AD.

**6.2 The metabolism hypothesis and autosomal dominant AD**

AG localizes to ACC, PCC, and parietal regions, loss of regions high in AG during youth appears to occur in all three regions without selectivity for any one of these regions [76]. So, unless AD per se involves hyperactivity of the DMN beyond that in normal elders for which there is no evidence, the hypothesis does not address the inconsistency between mechanisms of cognitive aging versus AD to account for the

Several observations related to cognitive aging need reconciliation with the metabolism hypothesis. Healthy elders free from amyloid deposition show a remarkable disconnection between the anterior and posterior default networks (i.e., ACC and PCC) [102]. Yet, nothing about the metabolism hypothesis explains why DMN regions with high AG, where amyloid will be deposited as AD develops, should disconnect—both regions should show aging-related hypometabolism as a result of white matter damage. Similarly, cognitively normal elders not at high risk for AD (no E4) show increases and decreases within anterior DMN connectivity, while showing only decreases in posterior DMN FC [103]. Cognitively intact elders with minimal amyloid deposition without E4 have greater connectivity of the ACC to the precuneus than those with E4 [95]. The aging-related anterior vs. posterior dissociations in connectivity within DMN networks remain theoretically difficult to

The ACC also has high glucose metabolism, flow, oxygen consumption, and AG; yet, the ACC does not show amyloid deposition akin to the PCC with healthy aging (those >60 years without amyloid positive scans [23]). Thus, chronic neural activity along with AG during the lifespan per se is not sufficient to lay down amyloid. Just as PCC hypometabolism marks focal atrophy early in AD, the region of ACC hypometabolism with aging should likewise display cortical thinning. However, several large studies do not support the prediction [104, 105], although not all findings are convergent [106]. Additionally, recent studies of tau deposition in preclinical AD show early deposition in the PCC but not in the ACC [107]. Those elders with cognitive function akin to much younger subjects (i.e., "SuperAgers") show thickening of the ACC and increased spindle cells suggestive of plasticity with aging; or alternatively, "SuperAgers" may be endowed with ACC thickening before aging [108, 109]. Also, age can confound years of education; the latter is associated with increased ACC thickness and metabolism [110, 111]. However, studies of the effects of aerobic fitness exercise on cognition and cortex show ACC thickening in

older adults in support of the potential for plasticity in this region [112].

A clear dissociation can arise also between amyloid deposition and FDG metabolism in the ACC in AD that is difficult to explain with the metabolism hypothesis. Patients who initially had mild AD, as confirmed with metabolic and amyloid biomarkers, were followed for 2 years along with a matched, amyloid negative control group. Despite extensive amyloid deposition in the ACC of the AD patients, no hypometabolism colocalized in the ACC during follow-up [22]. The agingrelated ACC hypometabolism noted previously likely led to a floor effect across groups. Likewise, aging-related ACC hypometabolism would tend to spare amyloid deposition in patients appearing inconsistent with the metabolism hypothesis given

A corollary of the metabolism hypothesis suggests those with familial AD would show similar patterns of hypometabolism and amyloid deposition to late-onset, sporadic AD. However, those with mutations (APP, PSEN1, PSEN2) show greatest amyloid deposition in the basal ganglia, a site that only becomes involved late in typical sporadic AD [113–115]. However, the specific pattern of

**22**

amyloid deposition may depend to some extent on the specific mutation. For example, the PS1 mutation, E280A, shows amyloid deposition more like lateonset sporadic AD than many other mutations with early onset. This variability may not surprise given the complexity of the underlying biology of different mutations in PS1 [116].
