From a425c020154dcd1dcab358ff0f078050f23f85d4 Mon Sep 17 00:00:00 2001 From: claude Date: Sat, 2 May 2026 15:30:20 +0800 Subject: [PATCH] Add METABOLISM AND ALZHEIMERS --- METABOLISM_AND_ALZHEIMERS.md | 1006 ++++++++++++++++++++++++++++++++++ 1 file changed, 1006 insertions(+) create mode 100644 METABOLISM_AND_ALZHEIMERS.md diff --git a/METABOLISM_AND_ALZHEIMERS.md b/METABOLISM_AND_ALZHEIMERS.md new file mode 100644 index 0000000..74ab116 --- /dev/null +++ b/METABOLISM_AND_ALZHEIMERS.md @@ -0,0 +1,1006 @@ +# Metabolism and Alzheimer's Disease: The Bioenergetic Theory +## Alzheimer's as a Disease of Impaired Brain Mitochondrial Function + +**Core thesis:** Alzheimer's disease is fundamentally a metabolic disease -- specifically, a disease of impaired brain energy metabolism. The amyloid plaques, tau tangles, and neuroinflammation that define AD are not the *cause* -- they are *consequences* of progressive brain mitochondrial dysfunction and glucose hypometabolism. The amyloid cascade hypothesis has dominated AD research for 30+ years and produced zero disease-modifying treatments, while metabolic evidence has been systematically ignored. The bioenergetic framework predicts that maintaining brain mitochondrial function is the most fundamental AD prevention strategy. + +This document is a companion to METABOLISM_AND_AGING.md and METABOLISM_AND_CANCER.md. The central argument is the same: **mitochondrial dysfunction is upstream, not downstream.** In aging, mitochondrial decline drives the hallmarks of aging. In cancer, mitochondrial damage drives the hallmarks of cancer. In Alzheimer's disease, brain mitochondrial decline drives amyloid deposition, tau hyperphosphorylation, neuroinflammation, synaptic loss, and cognitive failure. The three processes share a common root cause, and the prevention strategy for all three is identical -- maintain mitochondrial function. + +--- + +## Table of Contents + +1. [The Central Argument -- Alzheimer's Is a Metabolic Disease](#1-the-central-argument--alzheimers-is-a-metabolic-disease) +2. [The Amyloid Cascade Hypothesis -- Where It Went Wrong](#2-the-amyloid-cascade-hypothesis--where-it-went-wrong) +3. [Brain Energy Metabolism -- The Biochemistry](#3-brain-energy-metabolism--the-biochemistry) +4. [The APOE Connection -- Metabolism, Not Just Lipid Transport](#4-the-apoe-connection--metabolism-not-just-lipid-transport) +5. [Insulin Resistance in the Brain -- "Type 3 Diabetes"](#5-insulin-resistance-in-the-brain--type-3-diabetes) +6. [Tau -- A Metabolic Consequence, Not a Cause](#6-tau--a-metabolic-consequence-not-a-cause) +7. [Neuroinflammation -- The Metabolic-Inflammatory Nexus](#7-neuroinflammation--the-metabolic-inflammatory-nexus) +8. [Where the Amyloid, Metabolic, and Aging Frameworks Converge](#8-where-the-amyloid-metabolic-and-aging-frameworks-converge) +9. [Prevention and Intervention Through the Metabolic Lens](#9-prevention-and-intervention-through-the-metabolic-lens) +10. [The APOE e3/e4 Personal Prevention Architecture](#10-the-apoe-e3e4-personal-prevention-architecture) +11. [Key References and Intellectual Lineage](#11-key-references-and-intellectual-lineage) + +--- + +## 1. The Central Argument -- Alzheimer's Is a Metabolic Disease + +### 1.1 The Conventional View Is Backwards + +Mainstream neuroscience treats Alzheimer's disease (AD) as a proteinopathy -- amyloid-beta (Abeta) plaques accumulate in the brain, trigger tau hyperphosphorylation, provoke neuroinflammation, and ultimately destroy synapses and neurons. The interventions that follow from this view are protein-clearance focused: develop antibodies that bind and clear amyloid, inhibit the secretases that produce it, prevent tau aggregation. + +We propose the relationship is substantially reversed: + +``` +CONVENTIONAL (Amyloid Cascade): + APP mutations / aging --> Abeta accumulation --> Tau tangles --> Neurodegeneration + +PROPOSED (Metabolic): + Mitochondrial dysfunction --> Glucose hypometabolism --> Abeta + Tau + Inflammation + ^ | + └──────────────────────────────────────────────────────────────────┘ + (self-reinforcing loop) +``` + +Brain mitochondrial dysfunction creates the conditions for AD pathology to emerge: +- Impaired oxidative phosphorylation --> reduced ATP --> impaired proteostasis (Abeta clearance, tau degradation, synaptic maintenance all require ATP) +- Reduced Complex IV activity --> impaired cytochrome c oxidase --> neuronal energy crisis +- Shift from glucose oxidation toward compensatory pathways --> lactate accumulation, altered neurotransmitter synthesis +- Reduced mitochondrial membrane potential --> impaired calcium buffering --> excitotoxicity +- Increased mitochondrial ROS --> oxidative damage to lipids, proteins, DNA --> further mitochondrial damage +- Impaired mitochondrial transport along axons --> local energy failure at synapses --> synaptic dysfunction and loss +- Energy-depleted neurons upregulate Abeta production as a stress response (Abeta has antimicrobial and neuroprotective properties at low concentrations) +- Energy failure activates GSK3-beta --> tau hyperphosphorylation (the direct link to "tangles") + +The Abeta and tau then feed back to further impair mitochondria, creating the same self-reinforcing degenerative loop described for aging (METABOLISM_AND_AGING.md Section 1.1) and cancer (METABOLISM_AND_CANCER.md Section 1.1). + +### 1.2 The "Type 3 Diabetes" Concept + +In 2005, Suzanne de la Monte and Jack Wands at Brown University published a landmark paper in the *Journal of Alzheimer's Disease* demonstrating that post-mortem AD brains showed dramatically reduced expression of insulin, IGF-1, and their receptors -- specifically in the brain, independent of peripheral diabetes status. They coined the term **"type 3 diabetes"** to describe AD as a brain-specific form of insulin resistance with impaired insulin signalling, energy metabolism, and trophic support. + +This was not merely a clinical correlation. De la Monte showed that intracerebral administration of streptozotocin (a compound that destroys insulin-producing cells) in rats produced an AD-like phenotype: tau hyperphosphorylation, Abeta accumulation, neuroinflammation, synaptic loss, and cognitive impairment -- without any genetic manipulation. + +### 1.3 Why the Brain Fails First + +The brain is uniquely vulnerable to bioenergetic decline: + +| Parameter | Brain | Rest of Body | +|-----------|-------|--------------| +| **Mass** | ~2% of body weight (1.4 kg) | ~98% | +| **Energy consumption** | ~20% of total body oxygen, ~25% of glucose | ~75-80% | +| **Metabolic rate** | ~10x higher per gram than skeletal muscle | Variable | +| **Glycogen stores** | Minimal (~5 minutes of supply) | Substantial (liver, muscle) | +| **Antioxidant capacity** | Low (relatively low catalase, moderate SOD/GPx) | Higher in most tissues | +| **PUFA content** | Very high (~30-40% of grey matter fatty acids are DHA) | Lower | +| **Regenerative capacity** | Extremely limited (post-mitotic neurons) | Variable (most tissues regenerate) | +| **Iron accumulation** | Progressive with age (substantia nigra, hippocampus, basal ganglia) | Variable | + +The brain is running the highest metabolic rate of any organ, on the smallest energy reserves, with the least regenerative capacity, in a membrane environment (DHA-rich) that is maximally vulnerable to lipid peroxidation. It is the organ most likely to cross the **bioenergetic threshold** first -- the point below which mitochondrial ATP production can no longer sustain cellular function. This is why neurodegeneration is an early manifestation of the same metabolic decline that drives systemic aging (METABOLISM_AND_AGING.md Section 2). + +### 1.4 The 30-Year Failure of Amyloid-Targeting Therapeutics + +If Abeta accumulation is the root cause of AD, then drugs that reduce or clear Abeta should stop or reverse the disease. They have not. + +| Drug | Mechanism | Trial | Result | +|------|-----------|-------|--------| +| **Bapineuzumab** | Anti-Abeta monoclonal antibody | Phase III (2012) | No cognitive benefit; significant ARIA (amyloid-related imaging abnormalities) | +| **Solanezumab** | Anti-soluble Abeta antibody | Phase III x2 (2012, 2016); A4 prevention (2023) | No cognitive benefit in mild AD; no prevention benefit | +| **Aducanumab** | Anti-Abeta (aggregate-binding) | Phase III (EMERGE/ENGAGE 2019) | One trial "positive" (marginal, reanalysed after initial futility), one negative; FDA accelerated approval 2021 was highly controversial; CMS restricted coverage; Biogen withdrew it 2024 | +| **Lecanemab** | Anti-Abeta (protofibrils) | Phase III CLARITY AD (2023) | Statistically significant but **clinically marginal**: 27% slowing of decline on CDR-SB (0.45 points on an 18-point scale over 18 months); ~37% amyloid reduction; ARIA in 21.3%; 3 deaths potentially linked | +| **Donanemab** | Anti-Abeta (N-terminal pyroglutamate) | Phase III TRAILBLAZER-ALZ 2 (2023) | 35% slowing in early/intermediate amyloid; ARIA 24%; 3 deaths ARIA-related | +| **Semagacestat** | Gamma-secretase inhibitor | Phase III (2010) | WORSENED cognition; increased skin cancer risk; trial halted early | +| **Verubecestat** | BACE1 inhibitor | Phase III (2018) | WORSENED cognition in prodromal AD | +| **Tarenflurbil** | Gamma-secretase modulator | Phase III (2008) | No benefit | + +**~$40 billion** has been spent on amyloid-targeting drug development. The most optimistic results (lecanemab, donanemab) show effects so small they are below the threshold of clinical detectability by patients or caregivers, at the cost of dangerous brain edema (ARIA) and haemorrhage. Two drugs that successfully reduced amyloid production (semagacestat, verubecestat) actually **made patients worse** -- a result that is devastating for the amyloid hypothesis but perfectly consistent with the metabolic framework if Abeta has protective functions. + +The failure rate of amyloid-targeting therapeutics is not merely high -- it is essentially 100% for clinically meaningful disease modification. This is the single most important empirical fact in AD research, and it demands a fundamental reassessment of the underlying model. + +--- + +## 2. The Amyloid Cascade Hypothesis -- Where It Went Wrong + +### 2.1 The Founders and Their Framework + +The **amyloid cascade hypothesis (ACH)** was formally articulated by John Hardy and Gerald Higgins in 1992 (*Science*), building on earlier work by Karl Beyreuther and Colin Masters (who sequenced Abeta from plaques in 1984) and Dennis Selkoe (who demonstrated Abeta toxicity to cultured neurons). The hypothesis states: + +1. Abeta accumulation in the brain is the initial pathological event +2. Abeta triggers tau hyperphosphorylation and tangle formation +3. Tau pathology drives neuronal dysfunction and death +4. Everything else (inflammation, synaptic loss, cognitive decline) follows downstream + +Hardy's strongest evidence came from **familial (genetic) AD**, which accounts for <5% of all AD cases. Mutations in APP (amyloid precursor protein), PSEN1, and PSEN2 (presenilin 1 and 2, components of gamma-secretase) all increase Abeta42 production and cause aggressive early-onset AD (age 30-60). This genetic evidence is real and undeniable -- but its interpretation is contested. + +### 2.2 What the ACH Predicts and Where It Fails + +The ACH makes testable predictions: + +1. **Abeta accumulation should precede all other pathology** -- Partially true: amyloid PET positivity does precede clinical symptoms by 15-20 years (Jack et al. 2013, *Lancet Neurol*). But glucose hypometabolism on FDG-PET precedes amyloid deposition in many studies (Section 3.5). + +2. **More amyloid should mean worse disease** -- **False.** Up to 30-40% of cognitively normal elderly have substantial amyloid plaque burden at autopsy (Bennett et al. 2006, *Neurology*; Aizenstein et al. 2008, *Arch Neurol*). These individuals, termed "pathological agers" or "amyloid-positive cognitively normal," have plaques indistinguishable from AD patients but no dementia. Conversely, some clinically demented patients have minimal amyloid at autopsy. + +3. **Removing amyloid should stop or reverse the disease** -- **False.** See Section 1.4. Every trial that successfully cleared amyloid has failed to produce meaningful clinical benefit. + +4. **Reducing Abeta production should prevent the disease** -- **False, and worse.** BACE1 inhibitors (verubecestat, atabecestat, elenbecestat) and gamma-secretase inhibitors (semagacestat) that successfully reduced Abeta production worsened cognition -- suggesting Abeta has physiological functions that are being disrupted. + +5. **APP overexpression should reliably cause AD** -- **Complicated.** Down syndrome (trisomy 21) includes an extra copy of APP (located on chromosome 21) and is associated with near-universal AD pathology by age 40. This is the strongest genetic argument for the ACH. However, the extra chromosome 21 also duplicates ~200 other genes, including SOD1 (copper-zinc superoxide dismutase), which when overexpressed paradoxically increases oxidative stress through H2O2 overproduction. The Down syndrome-AD connection may be substantially metabolic, not purely amyloid-driven. + +### 2.3 The APP Processing Pathway + +Understanding why the metabolic framework reinterprets amyloid requires understanding how Abeta is produced: + +``` +AMYLOID PRECURSOR PROTEIN (APP) PROCESSING + + APP (770 amino acids, type I transmembrane) + | + ┌─────────────────────┴──────────────────────┐ + | | + NON-AMYLOIDOGENIC AMYLOIDOGENIC + (alpha-secretase) (beta-secretase) + | | + sAPP-alpha sAPP-beta + (neuroprotective, (less protective) + neurotrophic) | + | C99 fragment + C83 fragment | + | gamma-secretase + gamma-secretase (PSEN1/PSEN2) + | | + p3 peptide Abeta peptides + (non-toxic) Abeta40 (90%) -- less toxic + Abeta42 (10%) -- aggregation-prone + Abeta43 -- rare, very toxic + + KEY INSIGHT: The balance between alpha and beta processing + determines amyloid burden. Metabolic stress SHIFTS the balance + toward beta-secretase (BACE1) processing: + + - BACE1 expression is upregulated by oxidative stress, hypoxia, + NF-kappaB activation, and energy depletion + - Alpha-secretase (ADAM10) is activity-dependent and requires + healthy synaptic function + - Metabolically stressed neurons produce MORE Abeta, not because + of a genetic programme to create disease, but because BACE1 + is a stress-responsive enzyme +``` + +This is critical: **BACE1 is transcriptionally upregulated by the very conditions that define metabolic dysfunction** -- oxidative stress (Tamagno et al. 2002, *J Neurochem*), hypoxia/HIF-1alpha (Zhang et al. 2007, *Mol Cell Neurosci*), NF-kappaB activation (Chen et al. 2012, *Trends Mol Med*), and energy depletion (O'Connor et al. 2008, *Neuron*). The metabolic framework predicts exactly this: when neurons are energy-depleted and under oxidative stress, they shift APP processing toward the amyloidogenic pathway. Abeta is not an accidental toxic by-product -- it is a stress-induced response. + +### 2.4 Amyloid as a Protective Response -- The Antimicrobial Hypothesis + +In 2010, Soscia et al. (*PLoS ONE*) demonstrated that Abeta is a potent **antimicrobial peptide (AMP)** -- structurally and functionally homologous to the human cathelicidin LL-37. Abeta42 killed bacteria, fungi, and viruses in vitro at concentrations comparable to or exceeding LL-37. Robert Moir and Rudolph Tanzi at Harvard expanded this work: + +- Abeta forms oligomeric pores in microbial membranes (the same mechanism by which it damages neuronal membranes, reinterpreted as antimicrobial function) +- Abeta entraps and agglutinates microbes within its fibrillar aggregates -- the plaques themselves may be antimicrobial "nets" +- Transgenic mice that overexpress human Abeta survive brain infections (Salmonella, Candida, HSV-1) significantly better than wild-type (Kumar et al. 2016, *Sci Transl Med*) + +**The infection/inflammation trigger hypothesis:** Several pathogens have been detected in AD brains: HSV-1 (Itzhaki et al., multiple publications 1997-2020), Porphyromonas gingivalis (Dominy et al. 2019, *Sci Adv*), Chlamydia pneumoniae (Balin et al. 1998, *Med Microbiol Immunol*), and various fungi (Pisa et al. 2015, *Sci Rep*). The metabolic framework accommodates this: **chronic infection or inflammation creates metabolic stress --> metabolic stress upregulates BACE1 --> Abeta is produced as an antimicrobial defence --> if the underlying metabolic-inflammatory stimulus is not resolved, Abeta production becomes chronic --> chronic Abeta accumulation becomes pathological in its own right, further damaging mitochondria and perpetuating the cycle.** + +This interpretation explains why clearing amyloid without addressing the underlying metabolic-inflammatory stimulus fails to improve outcomes -- and may actually worsen them by removing an active antimicrobial defence. + +--- + +## 3. Brain Energy Metabolism -- The Biochemistry + +### 3.1 The Brain's Extraordinary Metabolic Demands + +The adult human brain: +- Weighs ~1.4 kg (~2% of body mass) +- Consumes ~20% of cardiac output (~750 mL blood/min) +- Uses ~20% of total body O2 consumption (~3.5 mL O2/100g/min for grey matter) +- Consumes ~120-130 g glucose/day (~25% of total body glucose utilisation) +- Produces ~20% of total body CO2 +- Generates and consumes ~6 kg of ATP per day (~4.2 mmol ATP/g/min) + +A single cortical neuron may have ~2,000-3,000 mitochondria to sustain its ~10,000 synapses. Synaptic transmission is extraordinarily energy-intensive: each action potential requires restoration of ion gradients via Na+/K+-ATPase (consuming ~50% of brain ATP), and each synaptic vesicle cycle (docking, fusion, endocytosis, refilling) consumes additional ATP. A single neuron may fire 10-100 times per second and maintain thousands of synapses simultaneously. + +**The consequence:** even a modest decline in mitochondrial ATP production has immediate functional effects on neural activity. A 10-15% decline in brain ATP -- undetectable in most tissues -- may be sufficient to impair synaptic plasticity, long-term potentiation (LTP), and memory consolidation. This is the bioenergetic threshold for cognitive function, and it is lower than for any other organ. + +### 3.2 Glucose Transport and Utilisation + +Brain glucose uptake depends on specific glucose transporters: + +``` +BRAIN GLUCOSE DELIVERY AND UTILISATION + + Blood + | + | GLUT1 (55 kDa isoform, endothelial cells of BBB) + | Km ~1-2 mM (high affinity, constitutive) + v + Brain interstitial fluid + | + | GLUT1 (45 kDa isoform, astrocyte end-feet) + | Surrounds >99% of brain capillary surface + v + Astrocytes + | (glycolysis --> lactate, via LDH5) + | (glycogen storage and breakdown) + v + MCT1/MCT4 (monocarboxylate transporters) + | (export lactate) + v + Extracellular space + | + | MCT2 (neuronal, high affinity, Km ~0.7 mM) + v + Neurons + | + | GLUT3 (neuronal glucose transporter) + | Km ~1.4 mM (highest affinity of any GLUT) + | Ensures neurons get glucose even at low concentrations + v + Glucose --> Glycolysis --> Pyruvate --> PDH --> Acetyl-CoA --> TCA --> ETC --> ATP + + KEY FINDINGS IN AD: + - GLUT1 expression reduced 25-30% in AD brain (Simpson et al. 1994) + - GLUT3 expression reduced in proportion to tau pathology (Liu et al. 2008) + - GLUT1 haploinsufficiency in mice accelerates AD-like pathology + (Winkler et al. 2015, Nat Neurosci) + - Glucose transporter decline PRECEDES amyloid deposition +``` + +**Winkler et al. (2015, *Nature Neuroscience*)** provided decisive evidence: crossing GLUT1-haploinsufficient mice with APP/PS1 transgenic mice dramatically accelerated Abeta pathology, tau phosphorylation, and neurodegeneration. Reducing brain glucose transport by ~50% was sufficient to convert a slowly progressing amyloid model into an aggressive neurodegenerative one. This demonstrates that **glucose hypometabolism drives amyloid pathology, not the reverse.** + +### 3.3 The Astrocyte-Neuron Lactate Shuttle + +Pierre Magistretti and Luc Pellerin proposed the **astrocyte-neuron lactate shuttle (ANLS)** hypothesis in 1994 (*PNAS*), fundamentally changing our understanding of brain fuel metabolism: + +1. Glutamate released at synapses is taken up by astrocyte end-feet via EAAT1/EAAT2 transporters +2. Glutamate uptake is coupled to Na+ influx --> activates Na+/K+-ATPase --> increases astrocyte ATP demand +3. Astrocytes respond by increasing glycolysis (not OxPhos) --> produce lactate +4. Lactate is exported via MCT1/MCT4 and imported by neurons via MCT2 +5. Neurons convert lactate to pyruvate (via LDH1) and oxidise it through the TCA cycle and ETC + +**The metabolic coupling is elegant:** neuronal activity at synapses directly signals astrocytes to increase fuel delivery. But it depends on intact neuronal mitochondria to oxidise the delivered lactate. When neuronal mitochondria are dysfunctional, this coupling breaks down -- neurons cannot use the lactate being delivered, creating a local energy crisis despite adequate fuel supply. This is analogous to the systemic insulin resistance-metabolic dysfunction loop described in METABOLISM_AND_AGING.md Section 4. + +### 3.4 Ketone Bodies -- The Alternative Brain Fuel + +Ketone bodies (beta-hydroxybutyrate/BHB, acetoacetate/AcAc) provide an alternative brain fuel that bypasses the glucose transport impairment seen in early AD: + +| Property | Glucose | Ketone Bodies | +|----------|---------|---------------| +| **BBB transport** | GLUT1 (reduced in AD) | MCT1 (preserved in AD) | +| **Neuronal uptake** | GLUT3 (reduced in AD) | MCT2 (preserved in AD) | +| **Entry into TCA cycle** | Glycolysis --> PDH --> acetyl-CoA | Directly to acetyl-CoA (bypasses glycolysis and PDH) | +| **ATP yield per carbon** | ~30 ATP per glucose (6C) | ~23 ATP per BHB (4C) -- more efficient per carbon | +| **Complex I dependence** | Full | Full (NADH from BHB metabolism enters at Complex I) | +| **Complex II contribution** | Some (via FADH2 from succinate) | Greater relative contribution via succinyl-CoA | +| **ROS generation** | Baseline | **Reduced** (Maalouf et al. 2007, *Neurobiol Dis*) -- ketone oxidation produces less ROS per ATP | +| **Availability in AD** | Impaired (GLUT1/GLUT3 decline) | **Preserved** (Castellano et al. 2015, *Neurobiol Aging*) | + +Stephen Cunnane and colleagues at the University of Sherbrooke have demonstrated using **carbon-11 acetoacetate PET** that brain ketone uptake is preserved in early AD and MCI -- even when glucose uptake is markedly reduced (Castellano et al. 2015; Croteau et al. 2018). This means the brain's ketone-utilising machinery remains intact while its glucose-utilising machinery is failing. Providing ketone bodies (via ketogenic diet, MCT oil, or exogenous ketone supplements) can partially compensate for the glucose hypometabolic deficit. + +### 3.5 FDG-PET -- The Decisive Temporal Evidence + +**Fluorodeoxyglucose positron emission tomography (FDG-PET)** measures regional brain glucose metabolism. The FDG-PET findings in AD are among the most replicated in all of clinical neuroscience: + +**The pattern:** Bilateral reductions in glucose metabolism in the **posterior cingulate cortex (PCC)**, **precuneus**, **temporal-parietal junction**, and **lateral temporal cortex**. This pattern is distinct from other dementias and is diagnostically useful (85-95% sensitivity and specificity for AD). + +**The critical temporal finding:** Glucose hypometabolism on FDG-PET is detectable **decades before** clinical symptoms and, in many studies, **before amyloid deposition:** + +- **Reiman et al. (2004, *PNAS*):** Young adult APOE e4 carriers (age 20-39, cognitively normal, no amyloid on any available measure) already showed reduced glucose metabolism in the same posterior cingulate/precuneus/temporal regions that decline in clinical AD. This was detected **40+ years** before the typical age of AD onset. +- **Mosconi et al. (2008, *Ann Neurol*):** FDG-PET hypometabolism at baseline predicted conversion from MCI to AD with ~85% accuracy -- outperforming amyloid PET, CSF Abeta, and clinical measures. +- **Gordon et al. (2018, *Nat Med*):** In autosomal dominant (familial) AD, FDG-PET changes were detectable ~18 years before expected symptom onset, concurrent with or slightly preceding amyloid changes. + +**The implication is profound:** if glucose hypometabolism precedes amyloid deposition, then metabolic dysfunction cannot be a downstream consequence of amyloid. It must be either upstream (causing amyloid, as the metabolic framework proposes) or parallel (an independent co-pathology). Either way, it demolishes the strict linear causality of the ACH. + +### 3.6 Brain Mitochondrial Changes in AD + +Post-mortem and biopsy studies reveal systematic mitochondrial dysfunction in AD brains: + +| Component | Change in AD | Key Evidence | +|-----------|-------------|--------------| +| **Complex IV (CcO)** | **30-50% reduction** in activity in temporal cortex, hippocampus | Maurer et al. 2000, *Neurobiol Aging*; Valla et al. 2010 | +| **Complex I** | Reduced activity | Kim et al. 2001, *Neurobiol Aging* | +| **PDH (pyruvate dehydrogenase)** | **Reduced activity** -- the gatekeeper enzyme | Gibson et al. 1998, *Ann Neurol* -- "PDH complex deficit is the most consistent enzyme change in AD" | +| **Alpha-ketoglutarate dehydrogenase** | **~50% reduction** -- TCA cycle enzyme, thiamine-dependent | Gibson et al. 2000, *Arch Neurol* | +| **Cardiolipin** | Oxidised and depleted | Monteiro-Cardoso et al. 2015 | +| **Mitochondrial membrane potential** | Reduced | Increased mPTP susceptibility | +| **Mitochondrial dynamics** | Excessive fission (DRP1 increase), reduced fusion | Wang et al. 2009, *J Neurosci* | +| **Mitochondrial transport** | Impaired axonal transport of mitochondria | Calkins et al. 2011 | +| **mtDNA** | Increased deletions and point mutations | Lin et al. 2002, *Mol Brain Res* | +| **Cytochrome c oxidase (histochemistry)** | Reduced staining in vulnerable regions | Gonzalez-Lima lab, multiple studies | + +**John Blass and Gary Gibson** at Weill Cornell were among the first to systematically document these changes in the 1990s. Their "mitochondrial cascade" model of AD predated and was largely overshadowed by the amyloid cascade but has proven more predictive. Gibson's finding that alpha-ketoglutarate dehydrogenase (alpha-KGDH, a thiamine-dependent TCA cycle enzyme) is reduced by ~50% in AD brain is particularly notable -- it directly implicates TCA cycle dysfunction and explains why thiamine (vitamin B1) deficiency mimics many features of AD (Wernicke-Korsakoff syndrome; see SUPPLEMENTS.md Section 1.2 on benfotiamine). + +--- + +## 4. The APOE Connection -- Metabolism, Not Just Lipid Transport + +### 4.1 APOE Biology + +**Apolipoprotein E (ApoE)** is a 299-amino acid lipid transport protein with three common human isoforms (E2, E3, E4) differing at positions 112 and 158: + +| Isoform | Position 112 | Position 158 | Frequency | AD Risk (vs E3/E3) | +|---------|-------------|-------------|-----------|---------------------| +| **APOE e2** | Cys | Cys | ~7% | 0.6x (protective) | +| **APOE e3** | Cys | Arg | ~78% | 1x (reference) | +| **APOE e4** | Arg | Arg | ~15% | ~3x (one copy), ~12x (two copies) | + +APOE is the dominant lipoprotein in the brain (the brain does not use peripheral lipoproteins -- it makes its own, primarily in astrocytes). Brain ApoE mediates cholesterol and phospholipid transport between astrocytes and neurons, synaptic repair, membrane maintenance, and Abeta clearance. The Cys112Arg substitution in ApoE4 alters the protein's structure: the arginine at position 112 forms a salt bridge with Glu109, changing the domain interaction between the N-terminal receptor-binding domain and the C-terminal lipid-binding domain. This single amino acid change has consequences across multiple systems. + +### 4.2 APOE e4 and Brain Glucose Metabolism + +The most important APOE e4 finding from the metabolic perspective is the FDG-PET data: + +**Reiman et al. (2004, *PNAS*; 2005, *Neurobiol Aging*)** demonstrated that cognitively normal APOE e4 carriers aged 20-39 already had significantly reduced cerebral metabolic rate of glucose (CMRgl) in posterior cingulate, precuneus, and lateral temporal cortex. These individuals were decades away from possible AD onset and had no detectable amyloid pathology. + +**The magnitude:** 5-10% reduction in CMRgl in young e4 carriers, progressing to 10-20% reduction by middle age, and 25-40% in those who develop MCI/AD. The pattern is the same in young carriers and AD patients -- it is the same metabolic signature, separated by decades of progression. + +**Mosconi et al. (2009, *Biol Psychiatry*)** showed that e4 carriers had significantly reduced mitochondrial cytochrome oxidase activity in platelets (a peripheral proxy for brain mitochondrial function) -- suggesting the metabolic deficit is not limited to the brain but reflects a systemic bioenergetic vulnerability. + +### 4.3 ApoE4 Mitochondrial Toxicity + +Robert Bhatt and Robert Mahley (Gladstone Institutes, UCSF) demonstrated a direct mechanism by which ApoE4 damages mitochondria: + +1. ApoE is synthesised in neurons under stress conditions (normally, astrocytes are the primary ApoE source) +2. Neuronal ApoE4 is proteolytically cleaved by a chymotrypsin-like serine protease to generate **C-terminal truncated fragments** (ApoE4 1-272) +3. These fragments escape the secretory pathway and translocate to mitochondria +4. Within mitochondria, truncated ApoE4 fragments bind to and impair **Complex IV subunits** (COX IV, COXVA) and **Complex V (ATP synthase)** +5. This binding reduces electron transport chain efficiency, increases ROS production, and can trigger mitochondrial permeability transition + +**Critically, ApoE3 is cleaved less efficiently and its fragments have lower mitochondrial toxicity.** The Arg112 in ApoE4 renders the protein more susceptible to proteolytic cleavage and the resulting fragments more prone to mitochondrial translocation. + +``` +ApoE4 MITOCHONDRIAL TOXICITY PATHWAY + + Neuronal stress / injury / aging + | + v + Increased neuronal ApoE expression (normally astrocytic) + | + v + ApoE4 protein (Arg112, Arg158) + | + Proteolytic cleavage (serine protease) + | + v + Truncated ApoE4 (aa 1-272) + | + Escapes ER/Golgi --> translocates to mitochondria + | + v + ┌─────────────────────────────────────────┐ + │ MITOCHONDRIAL DAMAGE │ + │ - Binds Complex IV subunits (COX IV) │ + │ - Impairs ETC electron flow │ + │ - Reduces ATP production │ + │ - Increases ROS generation │ + │ - Sensitises mPTP opening │ + │ - Fragments mitochondrial network │ + └─────────────────────────────────────────┘ + | + v + Neuronal energy crisis + | + v + ┌────────┴────────┐ + │ │ + BACE1 upregulation GSK3-beta activation + (Abeta production) (Tau phosphorylation) +``` + +Mahley (2012, *J Lipid Res* review) synthesised decades of this work and concluded: "the fundamental problem with apoE4 is its detrimental effects on mitochondrial function and oxidative stress." The increased AD risk conferred by APOE e4 is, in this framework, **primarily a metabolic vulnerability** -- a structurally determined propensity for mitochondrial damage -- rather than simply impaired Abeta clearance (though it is that too). + +### 4.4 APOE e4 and Blood-Brain Barrier Integrity + +Bell et al. (2012, *Nature*) demonstrated that ApoE4 accelerates blood-brain barrier (BBB) breakdown via pericyte degeneration. Pericytes express LRP1 (low-density lipoprotein receptor-related protein 1), and ApoE4 activates a CypA-NF-kappaB-MMP9 pathway in pericytes that degrades the basement membrane. This is relevant because: + +- BBB breakdown allows peripheral inflammatory mediators, albumin, fibrinogen, and red blood cells to enter the brain parenchyma +- Extravasated fibrinogen directly activates microglia toward a pro-inflammatory state (Ryu et al. 2018, *Nat Neurosci*) +- BBB disruption impairs clearance of Abeta and metabolic waste products +- BBB dysfunction is detectable in living APOE e4 carriers via dynamic contrast-enhanced MRI (Montagne et al. 2020, *Nature*) + +The metabolic interpretation: BBB integrity requires energy-dependent tight junction maintenance. ApoE4 impairs pericyte mitochondrial function --> pericyte degeneration --> BBB breakdown --> neuroinflammation + impaired waste clearance --> further metabolic stress. + +### 4.5 The Christchurch Mutation -- A Natural Experiment + +In 2019, Arboleda-Velasquez et al. (*Nat Med*) reported a Colombian woman from a presenilin-1 E280A kindred (the world's largest familial AD pedigree, in which carriers develop AD in their 40s) who carried two copies of the **APOE3 Christchurch mutation (R136S)** and remained cognitively intact until her 70s -- three decades beyond the expected age of onset. Her brain showed massive amyloid burden (the highest ever recorded in the kindred) but minimal tau pathology and neurodegeneration. + +This case is devastating for a strict amyloid cascade interpretation: she had extreme amyloid but was protected. The Christchurch mutation reduces ApoE binding to heparan sulfate proteoglycans (HSPGs), which are involved in both Abeta uptake/aggregation and tau propagation. But critically, the Christchurch mutation also preserves lipid transport function -- ApoE3ch maintains metabolic support of neurons while altering its interaction with pathological aggregates. The metabolic framework interprets this as: **maintaining neuronal metabolic support (which ApoE3ch does) is more protective than removing amyloid (which she had in abundance).** + +--- + +## 5. Insulin Resistance in the Brain -- "Type 3 Diabetes" + +### 5.1 Brain Insulin Signalling + +Insulin in the brain is not primarily a metabolic hormone (neurons are largely insulin-independent for basal glucose uptake via GLUT3). Instead, brain insulin functions as a **neurotrophic and neuromodulatory signal:** + +- Enhances synaptic plasticity and long-term potentiation (LTP) -- the cellular basis of memory +- Promotes dendritic arborisation and synaptogenesis +- Modulates neurotransmitter release (dopamine, acetylcholine, norepinephrine) +- Inhibits apoptosis via PI3K/Akt signalling +- Regulates tau phosphorylation via GSK3-beta inhibition + +The brain insulin signalling cascade: + +``` +BRAIN INSULIN SIGNALLING --> TAU PHOSPHORYLATION + + Insulin + | + v + Insulin Receptor (IR, tyrosine kinase) + | + v + IRS-1 (insulin receptor substrate 1) + | + v + PI3K (phosphoinositide 3-kinase) + | + v + Akt (protein kinase B) -- ACTIVE + | + v + GSK3-beta -- PHOSPHORYLATED (Ser9) --> INACTIVE + | + | (GSK3-beta is the MAJOR tau kinase) + | + v + TAU PHOSPHORYLATION -- NORMAL (low level) + Microtubule stability -- MAINTAINED + Axonal transport -- FUNCTIONAL + + WHEN INSULIN SIGNALLING FAILS: + + No insulin signal + | + v + IRS-1 phosphorylated on SERINE residues (inhibitory) + (by JNK, IKKbeta, PKC -- all activated by inflammation) + | + v + PI3K -- INACTIVE + | + v + Akt -- INACTIVE + | + v + GSK3-beta -- DEPHOSPHORYLATED --> ACTIVE + | + v + TAU HYPERPHOSPHORYLATION (at >40 sites) + Microtubule DESTABILISATION + Axonal transport FAILURE + Neurofibrillary tangle formation +``` + +**The critical insight:** GSK3-beta is simultaneously the major tau kinase and a downstream target of insulin signalling. Brain insulin resistance directly activates the kinase responsible for tau pathology. This is not a correlation -- it is a direct mechanistic link between metabolic dysfunction and the second defining feature of AD. + +### 5.2 Evidence for Brain Insulin Resistance in AD + +De la Monte's 2005 findings have been extensively replicated and extended: + +- **Steen et al. (2005, *J Alzheimers Dis*):** AD brains showed progressive reductions in insulin receptor expression, IRS-1 expression, and PI3K activity from Braak stage 0 through VI, correlating with disease severity. These changes were independent of peripheral diabetes status. +- **Talbot et al. (2012, *J Clin Invest*):** Brain insulin resistance (measured by ex vivo insulin stimulation of hippocampal tissue) was present in AD brains and correlated with tau pathology and cognitive impairment -- including in non-diabetic individuals. +- **Bomfim et al. (2012, *J Clin Invest*):** Abeta oligomers directly induced neuronal insulin resistance by activating JNK and TNF-alpha signalling --> IRS-1 serine phosphorylation --> PI3K/Akt inhibition --> GSK3-beta activation. This demonstrates a bidirectional relationship: metabolic dysfunction produces Abeta, and Abeta worsens metabolic dysfunction. + +### 5.3 The Diabetes-AD Epidemiological Connection + +The epidemiological evidence is consistent: + +- **Type 2 diabetes increases AD risk ~1.5-2.0x** (Ott et al. 1999, *Neurology*; Luchsinger et al. 2004, *Am J Epidemiol*) +- **Insulin resistance (even without diabetes) increases AD risk** (Craft et al. 1998, *Arch Neurol*) +- **Hyperinsulinaemia in midlife predicts AD decades later** (Ronnemaa et al. 2008, *Neurology*) +- **Metformin use in diabetics is associated with reduced AD risk** (Ng et al. 2014, *Diabetes Care*) -- though metformin's mechanism may be Complex I-related, not purely insulin-sensitising +- **The APOE e4 + diabetes combination is synergistic** -- e4 carriers with diabetes have ~5x AD risk, far exceeding the additive prediction (~3x + ~1.7x) + +### 5.4 Intranasal Insulin -- Targeting Brain Insulin Resistance Directly + +Suzanne Craft at Wake Forest School of Medicine pioneered intranasal insulin delivery as an AD intervention: + +- **Reger et al. (2008, *Neurology*):** Acute intranasal insulin (20 or 40 IU) improved verbal memory in AD patients and MCI subjects. APOE e4 carriers showed a **differential response** -- they benefited from 20 IU but not 40 IU, while non-carriers benefited from 40 IU. +- **Craft et al. (2012, *Arch Neurol*):** 4-month intranasal insulin (20 or 40 IU/day) trial in MCI/mild AD (n=104). 20 IU/day group showed preserved cognition and functional ability; FDG-PET showed preserved glucose metabolism in treated group. +- **The SNIFF trial (2020, *NEJM*):** Phase II/III (n=240, 12 months). **Primary endpoints not met** overall. However, significant device malfunction was documented (the delivery device was changed mid-trial, and the new device had inferior brain deposition). A subsequent analysis using only the original device showed positive signals. + +The intranasal insulin story is instructive: the metabolic approach (restoring brain insulin signalling) showed early promise, but execution challenges in clinical trials obscured potentially real effects. This parallels the broader difficulty of translating metabolic interventions into the rigid structure of pharmaceutical trials designed for single-target drugs. + +### 5.5 TCF7L2 TT and Brain Insulin -- The Genotype Connection + +The TCF7L2 rs7903146 TT genotype confers ~1.7x type 2 diabetes risk through impaired beta-cell function and incretin (GLP-1) signalling. While TCF7L2 is primarily studied in peripheral glucose metabolism, Wnt/beta-catenin signalling (which TCF7L2 mediates) is critical in the brain: + +- TCF7L2 is expressed in hippocampal neurons where it regulates synaptic gene expression +- Wnt signalling inhibits GSK3-beta (the same tau kinase regulated by insulin) +- TCF7L2 risk variants may impair this Wnt-mediated GSK3-beta inhibition in brain +- The convergence of impaired insulin signalling (TCF7L2 TT-driven beta-cell failure) and impaired Wnt signalling (TCF7L2 TT-driven GSK3-beta disinhibition) creates a dual-pathway vulnerability for tau hyperphosphorylation + +**Combined with APOE e4, TCF7L2 TT creates a metabolic risk architecture where both systemic insulin resistance (driving peripheral metabolic syndrome) and brain-specific signalling impairment (driving GSK3-beta activation) converge on the same neurodegenerative pathway.** + +--- + +## 6. Tau -- A Metabolic Consequence, Not a Cause + +### 6.1 Normal Tau Function + +**Tau (microtubule-associated protein tau, MAPT gene, chromosome 17)** is a critically important structural protein in neurons. Normal tau function: + +- Binds and stabilises axonal microtubules -- the "railway tracks" along which mitochondria, vesicles, neurotransmitters, and organelles are transported +- Regulates microtubule spacing and bundling +- Modulates axonal transport rates (kinesin/dynein motor proteins) +- Interacts with the actin cytoskeleton at growth cones +- May have signalling functions in dendrites (recently recognised) + +Tau is particularly abundant in long-projecting neurons -- precisely the neurons that have the highest metabolic demands and longest axons (hippocampal, entorhinal, cholinergic basal forebrain, locus coeruleus). These are the neurons that degenerate first in AD. + +### 6.2 How Energy Failure Drives Tau Pathology + +The metabolic framework provides a clear mechanistic link from energy failure to tau hyperphosphorylation: + +1. **Insulin/IGF-1 pathway failure --> GSK3-beta activation** (Section 5.1) +2. **CDK5 activation:** Energy depletion and calcium dysregulation activate calpain, which cleaves p35 to p25. p25/CDK5 is a potent tau kinase with prolonged activity (deregulated, unlike normal p35/CDK5) +3. **PP2A inactivation:** Protein phosphatase 2A (PP2A) is the major tau phosphatase -- it normally removes phosphate groups from tau, maintaining the balance. PP2A activity is reduced in AD brain by ~20-30% (Gong et al. 2000, *J Biol Chem*). PP2A is sensitive to oxidative stress and metabolic state -- its regulatory subunit is methylated by SAM-dependent methyltransferases, linking PP2A activity to the methylation cycle (MTHFR C677T het relevance -- see SUPPLEMENTS.md Section 1.2) +4. **AMPK activation:** Chronic energy depletion activates AMPK, which can phosphorylate tau at multiple sites (Thornton et al. 2011, *J Biol Chem*). AMPK is the cell's energy sensor -- its activation in tau hyperphosphorylation directly implicates energy failure. + +``` +ENERGY FAILURE --> TAU HYPERPHOSPHORYLATION (multiple converging pathways) + + Mitochondrial dysfunction + | + ┌────────────────┼────────────────┐ + | | | + ATP depletion ROS increase Ca2+ dysregulation + | | | + v v v + AMPK activation PP2A oxidation Calpain activation + | (inactivation) | + | | p35 --> p25 + | | | + v v v + Tau kinase Tau phosphatase CDK5/p25 + activation FAILURE activation + | | | + └────────────────┼─────────────────┘ + | + Insulin resistance --> GSK3-beta activation + | + v + TAU HYPERPHOSPHORYLATION (~40 sites) + | + ┌───────┴───────┐ + | | + Microtubule Tau aggregation + destabilisation (paired helical + | filaments, NFTs) + v + Axonal transport failure + | + ┌──────┴──────┐ + | | + Mitochondria Synaptic vesicles + cannot reach cannot reach + synapses terminals + | | + v v + Local energy Neurotransmitter + failure at depletion + synapse | + | v + v Cognitive decline + FURTHER mitochondrial + dysfunction (vicious cycle) +``` + +### 6.3 The Vicious Cycle of Axonal Transport Failure + +This vicious cycle is perhaps the most important mechanistic insight in the metabolic model of AD: + +- Neurons depend on **anterograde axonal transport** (kinesin motors, ATP-dependent) to deliver mitochondria from the cell body to synapses, which can be up to 1 metre away in motor neurons and 5-10 cm in cortical neurons +- Tau hyperphosphorylation destabilises microtubules --> disrupts the kinesin tracks +- Mitochondria cannot reach synapses --> local energy failure at the synapse +- Local energy failure --> local tau hyperphosphorylation (via the mechanisms above) +- More tau pathology --> more transport failure --> more distal energy failure +- The pathology spreads from synapse to cell body in a retrograde wave of energy failure + +This explains why **synaptic loss is the strongest correlate of cognitive decline in AD** -- stronger than amyloid plaques, stronger than total NFT burden (Terry et al. 1991, *Ann Neurol*). Synapses are the most metabolically demanding compartment of the neuron, the most distal from the cell body (where mitochondrial biogenesis occurs), and therefore the first to fail when axonal transport is compromised. + +### 6.4 Braak Staging and Metabolic Vulnerability + +Heiko and Eva Braak (1991, *Acta Neuropathol*) defined six stages of tau pathology progression (Braak stages I-VI). The progression follows a stereotyped anatomical path: transentorhinal cortex --> hippocampus --> temporal and parietal association cortex --> primary sensory and motor cortex. This staging correlates far better with cognitive decline than amyloid distribution. + +The metabolic framework notes that Braak staging correlates with **metabolic vulnerability:** the transentorhinal cortex and hippocampus have among the highest metabolic rates, the greatest synaptic density, the most Complex IV activity (and thus the most to lose), and the most reliance on long-range axonal projections. The "prion-like" spreading of tau pathology along neural circuits may not require a literal prion mechanism -- it may simply reflect the spread of metabolic failure along connected neural networks, with the most metabolically demanding circuits failing first. + +--- + +## 7. Neuroinflammation -- The Metabolic-Inflammatory Nexus + +### 7.1 Microglia as Metabolic Sensors + +**Microglia** (the brain's resident immune cells, derived from yolk sac macrophages) are increasingly recognised as metabolic sensors, not merely immune responders: + +- Resting (homeostatic) microglia maintain synaptic health through complement-mediated synaptic pruning, neurotrophic factor secretion, and debris clearance -- all ATP-intensive processes +- Microglial surveillance depends on continuous process extension and retraction -- a highly energy-dependent cytoskeletal activity +- Microglia express TREM2 (triggering receptor expressed on myeloid cells 2), which regulates microglial metabolism, phagocytosis, and survival. TREM2 loss-of-function variants (R47H, R62H) are the second-strongest genetic risk factors for AD after APOE e4. TREM2 promotes aerobic glycolysis and mTOR signalling in activated microglia -- its loss impairs the metabolic reprogramming required for effective phagocytic clearance. + +### 7.2 The Metabolic Shift in Activated Microglia + +When microglia encounter danger signals (Abeta, damaged neurons, DAMPs, PAMPs), they undergo **metabolic reprogramming** strikingly similar to the Warburg effect in cancer cells (see METABOLISM_AND_CANCER.md Section 3): + +| State | Metabolism | Function | Consequence | +|-------|-----------|----------|-------------| +| **Homeostatic (M0/surveillant)** | Oxidative phosphorylation | Synaptic pruning, trophic support, debris clearance | Neuroprotective | +| **Anti-inflammatory (M2-like)** | OxPhos + fatty acid oxidation | Phagocytosis, tissue repair, anti-inflammatory cytokines (IL-10, TGF-beta) | Reparative | +| **Pro-inflammatory (M1-like)** | **Aerobic glycolysis (Warburg shift)** | ROS/RNS production, pro-inflammatory cytokines (TNF-alpha, IL-1beta, IL-6), NO via iNOS | Neurotoxic if sustained | + +The M1/M2 dichotomy is a simplification (microglia exist on a spectrum), but the metabolic shift is real and well-documented: pro-inflammatory microglia increase glucose uptake (upregulate GLUT1), increase glycolysis, and suppress OxPhos -- exactly the Warburg shift. This is driven by: + +- HIF-1alpha stabilisation --> glycolytic gene expression +- NF-kappaB activation --> iNOS expression --> NO inhibits Complex IV +- mTOR activation --> biosynthetic programmes for cytokine production +- Succinate accumulation --> HIF-1alpha stabilisation (same oncometabolite mechanism as in cancer) + +**The vicious cycle:** Metabolically reprogrammed (glycolytic, pro-inflammatory) microglia consume glucose that neurons need, produce inflammatory mediators that further damage neuronal mitochondria, and release ROS that exacerbate oxidative stress. The microglia themselves are now competing with neurons for the limited glucose supply -- and winning, because glycolysis-dependent immune cells upregulate GLUT1 more aggressively than neurons can upregulate GLUT3. + +### 7.3 TNF-alpha, NF-kappaB, and the -308 AA Genotype + +TNF-alpha is a central mediator of AD neuroinflammation. The TNF-alpha -308 G/A genotype (AA homozygous) results in constitutively elevated TNF-alpha production. In the context of AD: + +- TNF-alpha directly impairs neuronal mitochondrial function (via TNF receptor signalling --> ceramide generation --> Complex III inhibition) +- TNF-alpha activates NF-kappaB in microglia --> more TNF-alpha, IL-1beta, IL-6 (positive feedback loop) +- TNF-alpha promotes brain insulin resistance via JNK and IKKbeta phosphorylation of IRS-1 (the same peripheral insulin resistance mechanism -- see METABOLISM_AND_AGING.md Section 4) +- TNF-alpha impairs Abeta clearance across the BBB +- TNF-alpha activates BACE1 expression via NF-kappaB binding to the BACE1 promoter + +``` +TNF-ALPHA -308 AA POSITIVE FEEDBACK IN AD BRAIN + + Constitutively elevated TNF-alpha (genetic baseline) + | + v + NF-kappaB activation (in microglia, astrocytes, neurons) + | + ┌────┼────────────────────────┐ + | | | + v v v + More BACE1 upregulation Brain insulin + TNF-alpha | resistance + (+feedback)| | + v v + Abeta production GSK3-beta activation + | | + v v + Microglial Tau hyperphosphorylation + activation | + | v + v Axonal transport failure + MORE TNF-alpha, | + IL-1beta, IL-6 v + | Neuronal energy crisis + v | + FURTHER mitochondrial | + damage <------------------┘ +``` + +For the TNF-alpha -308 AA genotype, this NF-kappaB-driven positive feedback loop starts from a **higher baseline** -- the genetic predisposition to elevated TNF-alpha means the inflammatory component of AD may be amplified. The multi-level NF-kappaB suppression strategy described in SUPPLEMENTS.md (curcumin, boron, zinc, PQQ, pranayama/vagal activation) takes on additional importance for AD prevention when the TNF-alpha AA genotype coexists with APOE e4. + +### 7.4 Why NSAIDs Failed in Trials Despite Epidemiological Promise + +Multiple observational studies showed that long-term NSAID use was associated with 30-60% reduced AD risk (McGeer et al. 1996, *Neurology*; in 't Veld et al. 2001, *NEJM*). However, randomised trials of NSAIDs for AD prevention (ADAPT trial -- naproxen and celecoxib; multiple others) uniformly failed. The ADAPT trial was halted early due to cardiovascular safety concerns without demonstrating cognitive benefit. + +The metabolic framework explains the discrepancy: **NSAIDs address inflammation but not the underlying metabolic dysfunction.** In observational studies, NSAID users may have been captured at a stage when metabolic decline was still early and amenable to inflammatory modulation. By the time of randomised trials (typically enrolling older, higher-risk individuals), the metabolic damage was too advanced for anti-inflammatory agents alone to reverse. Moreover, some NSAIDs (particularly at higher doses) inhibit mitochondrial Complex I -- potentially worsening the underlying bioenergetic deficit while suppressing the inflammatory response. + +--- + +## 8. Where the Amyloid, Metabolic, and Aging Frameworks Converge + +### 8.1 A Unified Model + +The metabolic model does not deny that Abeta plaques and tau tangles exist or that they contribute to neuronal damage. It reframes their role from primary cause to amplifying consequence: + +``` +THE UNIFIED MODEL OF ALZHEIMER'S DISEASE + + AGING (the #1 risk factor) + | + v + Progressive mitochondrial decline + (see METABOLISM_AND_AGING.md) + | + +--- APOE e4 ACCELERATES (mitochondrial toxicity, impaired lipid transport) + +--- TNF-alpha AA AMPLIFIES (higher baseline inflammation) + +--- TCF7L2 TT COMPOUNDS (insulin resistance, GSK3-beta disinhibition) + +--- Seed oils/PUFA (membrane peroxidation, cardiolipin damage) + +--- Sedentary lifestyle (reduced mitochondrial biogenesis) + +--- Sleep deprivation (impaired glymphatic clearance) + +--- Chronic stress (cortisol --> hippocampal mitochondrial damage) + | + v + BRAIN GLUCOSE HYPOMETABOLISM + (FDG-PET detectable 20-40 years before symptoms) + | + ┌────┼──────────────────────┐ + | | | + v v v + Abeta Tau hyper- Neuro- + production phosphorylation inflammation + (BACE1 up) (GSK3-beta) (microglial Warburg shift) + | | | + └────┬─────┘ | + | | + v v + Synaptic dysfunction Neuronal damage + (energy failure + (ROS, NO, cytokines) + transport failure) | + | | + └──────────┬──────────┘ + | + v + COGNITIVE DECLINE + (memory, then executive function, + then global impairment) +``` + +### 8.2 How This Model Explains the Major Observations + +| Observation | Amyloid Cascade Explanation | Metabolic Explanation | +|-------------|---------------------------|----------------------| +| **Aging is the #1 risk factor** | Amyloid accumulates with time | Mitochondrial function declines with age -- AD is accelerated aging of the brain | +| **APOE e4 is the strongest genetic risk** | Impaired Abeta clearance | Impaired mitochondrial function, brain glucose metabolism, BBB integrity, and lipid transport -- a **metabolic variant** | +| **Exercise is the most consistent protective factor** | Unclear mechanism in ACH | Enhances mitochondrial biogenesis, PGC-1alpha, brain BDNF, cerebral blood flow -- **metabolic enhancement** | +| **Diabetes increases AD risk 1.5-2x** | Hyperglycaemia somehow promotes amyloid | Insulin resistance --> brain energy failure + GSK3-beta activation -- **metabolic disease causes metabolic neurodegeneration** | +| **FDG-PET hypometabolism precedes amyloid** | Awkward; requires ad hoc explanation | Predicted -- metabolic decline is upstream | +| **30-40% of elderly have plaques without dementia** | "Cognitive reserve" hand-waving | These individuals maintain mitochondrial function -- they tolerate amyloid because their neurons have adequate energy | +| **Amyloid-clearing drugs don't work** | "Wrong target within amyloid" (oligomers vs plaques) | Removing a downstream marker doesn't address the metabolic cause | +| **BACE1 inhibitors worsen cognition** | Unexplained; devastating for ACH | Abeta has protective/antimicrobial functions; removing it worsens neuronal vulnerability | +| **Christchurch mutation protects despite massive amyloid** | Requires revision of ACH to include tau separately | Maintains metabolic/lipid support of neurons while blocking tau propagation | + +### 8.3 The Same Strategy That Prevents Aging and Cancer Prevents AD + +This is the unifying insight across the three companion documents: + +| Prevention Strategy | Aging | Cancer | Alzheimer's | +|--------------------|-------|--------|-------------| +| **Maintain mitochondrial function** | Core thesis | Core thesis | Core thesis | +| **Regular aerobic exercise** | Most robust anti-aging intervention | 20-40% cancer risk reduction | Most consistent AD-protective factor | +| **Avoid seed oils/excess PUFA** | Membrane protection, reduced lipid peroxidation | Reduced ferroptosis vulnerability, less inflammation | Brain membrane protection (DHA preservation, cardiolipin integrity) | +| **Maintain insulin sensitivity** | Metabolic flexibility, appropriate nutrient sensing | Reduced insulin/IGF-1 mitogenic signalling | Brain insulin signalling maintenance, GSK3-beta inhibition | +| **Manage inflammation** | Reduced inflammaging | Anti-tumour immune function | Reduced microglial Warburg shift | +| **Optimise thyroid function** | Metabolic rate maintenance | Metabolic rate maintenance | Brain metabolic rate maintenance | +| **Protect sleep** | Cellular repair, autophagy | Immune surveillance, DNA repair | Glymphatic clearance, synaptic maintenance | +| **Supplement metabolic cofactors** | ETC support | ETC support | Brain ETC support | + +--- + +## 9. Prevention and Intervention Through the Metabolic Lens + +### 9.1 Exercise -- The Strongest Evidence + +Exercise is the single most evidence-supported AD prevention strategy: + +- **Larson et al. (2006, *Ann Intern Med*):** Regular exercise (3+x/week) reduced dementia risk by 32% in 1,740 elderly followed for 6.2 years +- **Hamer & Chida (2009, *Psychol Med* meta-analysis):** Physical activity reduced dementia risk by 28% and AD specifically by 45% +- **Ahlskog et al. (2011, *Mayo Clin Proc* review):** Exercise preserved hippocampal volume, improved cerebral blood flow, enhanced BDNF, and reduced FDG-PET hypometabolism +- **Etnier et al. (2007):** APOE e4 carriers who exercised showed **no cognitive decline** compared to marked decline in sedentary carriers -- exercise may disproportionately benefit e4 carriers (consistent with the metabolic framework: e4 carriers have the greatest metabolic vulnerability and therefore the most to gain from metabolic enhancement) + +The metabolic mechanisms of exercise in AD prevention: +- Increases PGC-1alpha --> mitochondrial biogenesis in brain +- Increases BDNF (brain-derived neurotrophic factor) --> neuroplasticity, synaptic maintenance +- Increases cerebral blood flow --> improved oxygen and glucose delivery +- Increases insulin sensitivity --> brain insulin signalling restoration +- Increases autophagy --> clearance of damaged mitochondria, protein aggregates +- Increases Complex IV activity in brain (exercise is to brain mitochondria what PBM is to skin mitochondria) + +### 9.2 Ketogenic Diet and Medium-Chain Triglycerides + +If glucose hypometabolism is the proximate metabolic deficit in AD, providing an alternative fuel that bypasses impaired glucose transport is a rational strategy: + +- **Henderson et al. (2009, *Nutr Metab*):** AC-1202 (caprylic acid MCT formulation, later marketed as Axona) improved cognitive scores in AD patients who were APOE e4 **negative**. E4 carriers did not respond -- possibly because ApoE4 impairs ketone body metabolism or transport as well, though this remains debated. +- **Taylor et al. (2018, *J Alzheimers Dis*):** 3-month ketogenic diet in AD patients (n=15) showed improved ADAS-cog scores during ketosis, returning to baseline when ketosis was discontinued. +- **Fortier et al. (2021, *Alzheimers Dement*):** MCT supplementation (30 g/day for 6 months) in MCI patients increased brain ketone uptake (confirmed by 11C-acetoacetate PET), improved episodic memory, language, executive function, and processing speed. + +**The framework's caution:** While ketogenic interventions are mechanistically sound for AD, the bioenergetic framework favours glucose oxidation as the optimal fuel for healthy mitochondria (see METABOLISM_AND_AGING.md Section 3). Ketogenic diets are a therapeutic intervention for brains with impaired glucose metabolism -- they are not the optimal preventive diet. Prevention should focus on **maintaining the brain's ability to oxidise glucose efficiently** through mitochondrial protection, insulin sensitivity, and adequate glucose transporter expression. Ketones are the backup generator when the main power grid fails. + +### 9.3 The Supplement Stack -- AD-Relevant Components + +The existing supplement framework (SUPPLEMENTS.md) contains multiple compounds with AD-relevant mechanisms. The table below maps each to its specific AD-related action: + +| Supplement | AD-Relevant Mechanism | Evidence Level | Section | +|-----------|----------------------|----------------|---------| +| **CoQ10 / Ubiquinol** | ETC electron carrier, Complex III support, reduces mitochondrial ROS; age-related brain decline 30-35% (Kalen 1989) | Strong mechanistic, moderate clinical | 1.3 | +| **Methylene blue** | Alternative electron carrier bypassing Complex I/III; 10x brain concentration; 25-30% Complex IV enhancement; tau aggregation inhibitor; APOE e4-specific bioenergetic rescue | Strong mechanistic, Phase III tau data (monotherapy signal), strong animal neuroprotection | 3.19 | +| **B vitamins (B1/B2/B6/B9/B12)** | VITACOG trial: 30% brain atrophy reduction (7-fold in AD-vulnerable regions with omega-3); homocysteine lowering; NADH (B3), FAD (B2), TPP (B1) are literal ETC cofactors; alpha-KGDH (B1-dependent) reduced 50% in AD | Strong clinical (VITACOG), strong mechanistic | 1.2 | +| **Creatine** | Brain PCr buffer; compensates for cerebral hypometabolism; Mi-CK stabilisation; spares SAM for methylation | Strong mechanistic, moderate clinical (cognitive enhancement in vegetarians/stress) | 1.6 | +| **Magnesium (Mg-threonate)** | Mg-threonate (MgT) crosses BBB; Slutsky et al. 2010 -- enhanced hippocampal synaptic density, LTP, learning/memory in aged rats; NMDA receptor modulation; mPTP inhibition | Moderate-strong (MgT-specific preclinical; general Mg clinical) | 1.1 | +| **Lion's mane** | NGF and BDNF stimulation; cholinergic neuron support; Li et al. 2020 pilot -- improved MMSE in mild AD with erinacine A | Moderate preclinical, emerging clinical | 3.7 | +| **Curcumin** | NF-kappaB inhibition (BACE1 downregulation); synergistic Abeta clearance with vitamin D3 (Masoumi 2009); anti-inflammatory | Moderate clinical (bioavailable forms), strong mechanistic | 3.10 | +| **Alpha-GPC** | Acetylcholine precursor; De Jesus Moreno 2003 n=261 AD -- cognitive improvement; cholinergic substrate for APOE e4-vulnerable basal forebrain neurons | Moderate clinical | 3.16 | +| **Nicotine (low-dose, transdermal/lozenge)** | Alpha7 nAChR agonism --> cholinergic anti-inflammatory pathway; NF-kappaB suppression in microglia; Newhouse et al. 2012 pilot -- improved attention/memory in MCI; APOE e4 neuroprotection | Moderate clinical (MCI pilot), strong mechanistic (alpha7 nAChR) | 3.12 | +| **Vitamin D3** | VDR-mediated neuroprotection; Abeta clearance enhancement; Littlejohns 2014 -- 2.25x dementia risk at <10 ng/mL; neurotrophin support (NGF, GDNF) | Strong epidemiological, moderate interventional | 1.7 | +| **Vitamin K2 (MK-4)** | Brain sulfatide support (depleted early in AD); Gas6/TAM receptor neuroprotection | Emerging mechanistic | 1.8 | +| **Selenium** | Selenoprotein P brain delivery; GPx4 prevents ferroptosis (hippocampal neuron death); TrxR2 mitochondrial protection | Strong mechanistic, moderate clinical (Cardoso 2010 correlation) | 1.4 | + +### 9.4 Photobiomodulation -- Transcranial PBM for APOE e4 + +Transcranial photobiomodulation (tPBM) delivers 600-1100 nm light through the skull to enhance Complex IV (cytochrome c oxidase) activity in cortical neurons (see THERAPIES.md Section 1.1). The AD relevance: + +- Complex IV is the photoacceptor for PBM (CuA centre absorbs NIR) +- Complex IV activity is specifically reduced in AD brain +- tPBM increases cytochrome c oxidase activity, ATP production, and NO release +- Gonzalez-Lima (the same group that characterised MB's CcO enhancement) has shown tPBM improves cognitive function in healthy adults (Barrett & Gonzalez-Lima 2013, *Neurosci*) +- The MB + PBM synergy: MB delivers electrons TO Complex IV, while PBM enhances Complex IV's ACTIVITY -- the two interventions converge on the same enzyme from complementary directions + +``` +METHYLENE BLUE + PBM SYNERGY AT COMPLEX IV + + Methylene blue Photobiomodulation + (oral/sublingual) (transcranial, 660-810 nm) + | | + Crosses BBB, 10x brain concentration Penetrates skull ~1-3 cm + | | + v v + MBH donates electrons Photon absorbed by CuA + to cytochrome c copper centre in Complex IV + | | + v v + Increased electron SUPPLY Increased enzyme ACTIVITY + to Complex IV (NO dissociation, conformational + | change, enhanced turnover) + | | + └──────────────────┬───────────────────────┘ + | + v + Enhanced Complex IV output + More O2 consumed, more ATP, more H2O + Improved neuronal bioenergetics + | + v + Addresses the SPECIFIC deficit + in APOE e4 carriers: + reduced brain Complex IV activity +``` + +### 9.5 The Three-Pronged Cholinergic Strategy + +The **cholinergic hypothesis** of AD (Bartus et al. 1982; Whitehouse et al. 1982, *Science* -- 75-90% loss of cholinergic neurons in the nucleus basalis of Meynert) identified acetylcholine deficiency as a proximate cause of cognitive symptoms. Current AD drugs (donepezil, rivastigmine, galantamine) are all cholinesterase inhibitors -- they prevent ACh breakdown to prolong the signal from remaining cholinergic neurons. They are symptomatic band-aids that do not address why cholinergic neurons are dying. + +The metabolic-cholinergic framework proposes a three-pronged strategy: + +1. **Substrate supply -- Alpha-GPC:** Provides the choline precursor for ACh synthesis (see SUPPLEMENTS.md Section 3.16). Cholinergic neurons cannot make ACh without choline, and age-related PEMT decline reduces endogenous PC --> choline supply. + +2. **Trophic support -- Lion's mane:** Stimulates NGF synthesis (the survival factor for cholinergic neurons). NGF-secreting cell transplants improved cognition in AD (Tuszynski et al. 2005, *Nat Med* Phase I). Lion's mane provides a non-invasive, sustained NGF stimulus. + +3. **Receptor activation -- Nicotine (low-dose):** Directly activates alpha7 nicotinic acetylcholine receptors (nAChRs) on neurons and microglia. The alpha7 nAChR on microglia triggers the cholinergic anti-inflammatory pathway (Tracey 2002; see THERAPIES.md Section 2.1 and SUPPLEMENTS.md Section 3.12), suppressing NF-kappaB and reducing neuroinflammation. On neurons, alpha7 nAChR activation enhances synaptic plasticity and LTP. + +This three-pronged approach (substrate + trophic factor + receptor activation) addresses the cholinergic deficit at three levels simultaneously, rather than the single-level approach of cholinesterase inhibitors. + +### 9.6 Sleep and Glymphatic Clearance + +The **glymphatic system** (Iliff et al. 2012, *Sci Transl Med*; Nedergaard lab, University of Rochester) is a brain-wide waste clearance system that operates primarily during sleep: + +- CSF flows along periarterial spaces into the brain parenchyma +- AQP4 (aquaporin-4) water channels on astrocyte end-feet facilitate convective flow +- Interstitial solutes -- including Abeta and tau -- are flushed along perivascular channels to cervical lymphatic drainage +- Glymphatic flow increases ~60% during sleep compared to waking (Xie et al. 2013, *Science*) +- Sleep deprivation increases brain Abeta accumulation (Shokri-Kojori et al. 2018, *PNAS* -- one night of sleep deprivation increased Abeta in hippocampus and thalamus on PET) + +APOE e4 carriers may have impaired glymphatic function: ApoE4 disrupts AQP4 polarisation at astrocyte end-feet (Achariyar et al. 2016), and Abeta clearance via the glymphatic system depends on ApoE-mediated transport. The CLOCK CC genotype may provide some advantage as the evening chronotype, if leveraged for adequate sleep duration. + +### 9.7 The FINGER Trial and Multi-Domain Interventions + +The **Finnish Geriatric Intervention Study to Prevent Cognitive Impairment and Disability (FINGER)** is the most important clinical trial supporting the metabolic/multi-domain approach to AD prevention: + +- **Ngandu et al. (2015, *Lancet*):** n=1,260 at-risk elderly, 2-year multi-domain intervention (diet, exercise, cognitive training, vascular risk monitoring) vs general health advice +- **Result:** 25% relative improvement in overall cognitive score, 83% in executive function, 150% in processing speed +- **Significance:** A multi-domain intervention addressing metabolic, cardiovascular, and lifestyle factors significantly improved cognitive outcomes -- without any drug targeting amyloid or tau + +The FINGER model validates the metabolic approach: address multiple metabolic and lifestyle risk factors simultaneously, rather than targeting a single molecular pathway. The ongoing **World-Wide FINGERS** initiative is replicating this across 40+ countries. + +--- + +## 10. The APOE e3/e4 Personal Prevention Architecture + +### 10.1 The Risk Profile + +For APOE e3/e4 heterozygotes: +- Lifetime AD risk: ~20-25% (vs ~10-12% for e3/e3) +- Median age of onset (if AD develops): ~73 years (vs ~78 for e3/e3) +- FDG-PET changes detectable from ~age 20-30 +- BBB permeability changes potentially beginning in 40s +- Synaptic/cognitive compensation may mask pathology for 1-2 decades + +**The e3/e4 position is critical: the risk is real (~3x) but NOT deterministic.** The majority of e3/e4 carriers (75-80%) do NOT develop AD. The genetic loading creates vulnerability; whether that vulnerability manifests depends on metabolic, lifestyle, and environmental factors -- precisely the factors the bioenergetic framework addresses. + +### 10.2 Tiered Intervention Strategy for Age 36 + +| Tier | Intervention | Rationale | When to Start | +|------|-------------|-----------|---------------| +| **1 -- Non-Negotiable** | Aerobic exercise 150+ min/week | Strongest evidence for AD prevention; mitochondrial biogenesis; APOE e4 carriers benefit disproportionately | NOW | +| **1 -- Non-Negotiable** | Sleep 7-8+ hours, dark room, consistent schedule | Glymphatic clearance; synaptic maintenance | NOW | +| **1 -- Non-Negotiable** | Metabolic cofactor stack (CoQ10, B vitamins, Mg, Se, D3, K2) | Brain ETC support; homocysteine control; target Hcy <8 | NOW | +| **1 -- Non-Negotiable** | Insulin sensitivity maintenance (avoid seed oils, maintain body composition, exercise) | Prevent brain insulin resistance; GSK3-beta control | NOW | +| **2 -- Strongly Recommended** | Methylene blue (low-dose, 0.5-1 mg/kg) | Alternative electron carrier; Complex IV enhancement; tau inhibition; 10x brain concentration | Consider at 36-40 | +| **2 -- Strongly Recommended** | Transcranial PBM (810 nm, 2-3x/week) | Complex IV enhancement; synergistic with MB; targets the specific deficit in e4 carriers | NOW | +| **2 -- Strongly Recommended** | Three-pronged cholinergic strategy (alpha-GPC + lion's mane + low-dose nicotine) | Substrate + trophic + receptor activation; addresses cholinergic vulnerability | NOW | +| **2 -- Strongly Recommended** | NF-kappaB suppression stack (curcumin + boron + zinc + pranayama) | TNF-alpha AA amplification control; microglial activation modulation | NOW | +| **2 -- Strongly Recommended** | Creatine (3-5 g/day) | Brain PCr buffer; compensates cerebral hypometabolism | NOW | +| **3 -- Monitor and Consider** | MCT oil / C8 caprylic acid (10-30 mL/day) | Alternative brain fuel; bypasses GLUT1 decline; evidence mixed in e4 carriers | If biomarkers worsen | +| **3 -- Monitor and Consider** | Pregnenolone (topical or low-dose oral) | Neurosteroid; microtubule stabilisation; reduced in AD brain | If biomarkers worsen | +| **3 -- Monitor and Consider** | Intranasal insulin (if available via clinical trial) | Direct brain insulin restoration; Craft et al. early positive signals | If available | + +### 10.3 Biomarker Monitoring Protocol + +| Biomarker | What It Measures | Frequency | Action Threshold | +|-----------|-----------------|-----------|-----------------| +| **HbA1c** | Glycaemic control | Every 6 months | >5.5% -- intensify insulin sensitisation | +| **Fasting insulin** | Insulin resistance | Every 6 months | >8 uIU/mL -- investigate and address | +| **Homocysteine** | Methylation, B vitamin status | Every 6-12 months | >10 umol/L -- increase B6/B9/B12; target <8 | +| **hs-CRP** | Systemic inflammation | Every 6 months | >1.0 mg/L -- intensify NF-kappaB suppression | +| **Lipid panel (ApoB focus)** | Cardiovascular-metabolic | Every 6 months | ApoB >90 mg/dL -- dietary/lifestyle intervention | +| **Omega-3 index** | Membrane composition | Annually | <8% -- increase wild fish intake | +| **Vitamin D (25-OH)** | Vitamin D status | Every 6-12 months | <50 ng/mL -- adjust dosing; target 50-70 | +| **p-tau217** | Tau pathology (blood test, emerging) | Consider from age 45+ | Any positivity -- escalate all interventions | +| **Neurofilament light (NfL)** | Neurodegeneration (blood test) | Consider from age 45+ | Rising trend -- investigate, escalate | +| **FDG-PET** | Brain glucose metabolism | Baseline at 40-45, then every 5 years if normal | PCC/precuneus hypometabolism -- full protocol escalation | +| **Amyloid PET** | Amyloid deposition | Only if FDG-PET or fluid biomarkers are abnormal | Positive -- does NOT change the metabolic strategy, but adds urgency | + +### 10.4 The Framework's Prediction + +**If brain mitochondrial function is maintained through the interventions described in this document, the APOE e4-associated risk can be substantially mitigated.** This prediction is based on: + +1. APOE e4's primary mechanism is metabolic (mitochondrial toxicity, glucose hypometabolism), not directly amyloidogenic +2. Exercise (the strongest metabolic enhancer) disproportionately benefits e4 carriers +3. The 75-80% of e3/e4 carriers who never develop AD presumably maintain adequate brain bioenergetics +4. FDG-PET hypometabolism is a *progression marker* -- if it can be prevented or slowed, clinical AD should not manifest +5. The Christchurch case demonstrates that even massive amyloid can be tolerated if neuronal metabolic support is maintained + +This prediction is **not proven** -- no randomised trial has demonstrated that a comprehensive metabolic programme prevents AD in e4 carriers. But the convergent mechanistic, epidemiological, and early interventional evidence makes it the strongest framework-aligned prevention strategy available. + +--- + +## 11. Key References and Intellectual Lineage + +### The Metabolic Hypothesis Lineage + +- **Otto Warburg (1920s-1950s):** First to observe that cells with impaired respiration shift to fermentation. While his work focused on cancer, the principle applies to any cell with mitochondrial dysfunction -- including neurons. +- **John Blass and Gary Gibson (1990s-2000s, Weill Cornell):** Systematically documented mitochondrial enzyme deficiencies in AD brain. Gibson's finding of ~50% alpha-KGDH reduction in AD was among the first direct evidence of TCA cycle failure. Proposed the "mitochondrial cascade" model of AD. +- **Russell Swerdlow (University of Kansas, 2004 onwards):** Formalised the **"mitochondrial cascade hypothesis"** of late-onset AD, proposing that inherited mitochondrial function (influenced by both nuclear and mitochondrial genetics) determines the age at which bioenergetic decline crosses the threshold for AD pathology. Swerdlow explicitly proposed that mitochondrial dysfunction is **upstream** of amyloid and tau, not downstream. +- **Suzanne de la Monte and Jack Wands (Brown University, 2005 onwards):** Coined "type 3 diabetes" and demonstrated brain-specific insulin resistance in AD independent of peripheral diabetes. Showed streptozotocin-induced brain insulin resistance reproduces AD pathology in rats. +- **Robert Mahley and Yadong Huang (Gladstone Institutes, UCSF, 2000s-2010s):** Demonstrated ApoE4 mitochondrial toxicity -- truncated ApoE4 fragments enter mitochondria and impair Complex IV. Reframed APOE e4 from a lipid transport variant to a **metabolic vulnerability variant**. +- **Eric Reiman (Banner Alzheimer's Institute, 2004 onwards):** FDG-PET studies in young APOE e4 carriers demonstrating glucose hypometabolism decades before possible symptom onset. Arguably the most important single finding for the metabolic hypothesis. +- **Stephen Cunnane (University of Sherbrooke, 2010s-present):** Dual-tracer PET studies (FDG + 11C-acetoacetate) demonstrating preserved ketone uptake in AD brains with impaired glucose uptake. Provided the metabolic rationale for ketogenic interventions. +- **Dale Bredesen (Buck Institute, 2014 onwards):** ReCODE protocol -- the first clinically applied multi-domain metabolic approach to AD treatment. Published case series showing cognitive improvement in early AD/MCI patients with a personalised protocol addressing 36 metabolic and lifestyle parameters. Criticised for lack of controlled trials, but the framework is consistent with the metabolic hypothesis and the FINGER trial results. +- **Francisco Gonzalez-Lima (University of Texas at Austin, 1990s-present):** Systematic characterisation of methylene blue and photobiomodulation as brain cytochrome oxidase enhancers. Demonstrated that cognitive enhancement by MB is mechanistically linked to Complex IV activity -- not monoamine or glutamate modulation. +- **Suzanne Craft (Wake Forest, 2000s-present):** Intranasal insulin trials demonstrating that restoring brain insulin signalling improves cognition in MCI/AD. The only therapeutic approach that directly targets the "type 3 diabetes" mechanism. +- **Thomas Seyfried (Boston College, 2012 onwards):** While primarily focused on cancer (see METABOLISM_AND_CANCER.md), Seyfried's metabolic disease framework applies to neurodegeneration: cells that cannot maintain OxPhos undergo metabolic stress, and the downstream consequences depend on cell type -- cancer in proliferative tissues, degeneration in post-mitotic neurons. + +### Evidence Summary Table + +| Claim | Evidence Level | Key Evidence | +|-------|---------------|--------------| +| FDG-PET hypometabolism precedes amyloid deposition | **Strong** | Reiman 2004, 2005 (young e4 carriers); Mosconi 2008; Gordon 2018 | +| Amyloid-clearing drugs fail to improve cognition | **Strong (multiple Phase III)** | Bapineuzumab, solanezumab, aducanumab, lecanemab (marginal), donanemab (marginal) | +| 30-40% of elderly have plaques without dementia | **Strong (autopsy series)** | Bennett 2006; Aizenstein 2008; multiple neuropathology cohorts | +| APOE e4 impairs brain glucose metabolism in young carriers | **Strong** | Reiman 2004, 2005; Mosconi 2009 | +| Truncated ApoE4 enters mitochondria and impairs Complex IV | **Moderate-strong (mechanistic)** | Mahley 2012; Bhatt 2013; Chang 2005 | +| Brain insulin resistance occurs in AD independent of diabetes | **Strong** | De la Monte 2005; Steen 2005; Talbot 2012 | +| GSK3-beta links insulin resistance to tau phosphorylation | **Well-established** | Multiple labs; Hooper 2008 review | +| Exercise reduces AD risk ~30-45% | **Strong (meta-analyses)** | Hamer & Chida 2009; Larson 2006; Ahlskog 2011 | +| B vitamins reduce brain atrophy 30% (7-fold in AD regions) | **Strong (RCT)** | Smith 2010; Douaud 2013 (VITACOG) | +| Brain ketone uptake preserved when glucose uptake impaired | **Strong** | Castellano 2015; Croteau 2018 (dual-tracer PET) | +| Methylene blue enhances brain Complex IV activity 25-30% | **Strong (preclinical)** | Callaway 2004; Rojas 2012; Gonzalez-Lima lab | +| FINGER multi-domain intervention improves cognition | **Strong (RCT)** | Ngandu 2015 (n=1260, 2-year) | +| T2D increases AD risk 1.5-2x | **Strong (epidemiological)** | Ott 1999; Luchsinger 2004; multiple meta-analyses | +| Microglia undergo Warburg shift in AD | **Moderate-strong (mechanistic)** | Ulland 2017; Baik 2019; Holland 2018 | +| TNF-alpha drives BACE1 upregulation via NF-kappaB | **Strong (mechanistic)** | Chen 2012; Cho 2009; Bourne 2007 | +| Glymphatic clearance impaired in APOE e4 | **Emerging** | Achariyar 2016; Peng 2016; mechanistic plausibility | + +### Cross-References to Other Framework Documents + +| Document | Relevant Sections | +|----------|------------------| +| **METABOLISM_AND_AGING.md** | Section 1 (metabolic decline as driver); Section 2 (ETC biochemistry); Section 3 (fuel selection); Section 5 (membrane pacemaker); Section 10 (hallmarks of aging) | +| **METABOLISM_AND_CANCER.md** | Section 1 (mitochondrial dysfunction upstream); Section 3 (ETC defects); Section 5 (PUFA/seed oil connection); Section 11 (convergence with aging framework) | +| **Genotype-specific analysis** | APOE e3/e4; TCF7L2 TT; TNF-alpha -308 AA; COMT Val/Met; BDNF Val/Met; DIO2 Thr92Ala het | +| **SUPPLEMENTS.md** | Section 1.2 (B vitamins/VITACOG); Section 1.3 (CoQ10); Section 1.6 (creatine); Section 3.7 (lion's mane); Section 3.10 (curcumin); Section 3.12 (nicotine/alpha7 nAChR); Section 3.16 (choline/alpha-GPC); Section 3.19 (methylene blue) | +| **THERAPIES.md** | Section 1.1 (PBM/transcranial); Section 2.1 (pranayama/cholinergic anti-inflammatory pathway) | +| **PLAN.md** | Pillar VII (mitochondrial rejuvenation); Pillar IV (proteostasis); Pillar XI (chronic inflammation) | + +--- + +*This document presents a framework-aligned interpretation of Alzheimer's disease research. The metabolic hypothesis is supported by substantial evidence but remains a minority position in mainstream neuroscience. The amyloid cascade hypothesis, despite its clinical failures, retains dominant institutional support. The interventions described here are based on biological plausibility and available evidence but have not been validated in a controlled trial specifically designed to prevent AD in APOE e4 carriers through metabolic enhancement. Decisions about personal prevention strategies should be informed by this evidence but also by consultation with qualified healthcare providers.* + +--- + +**Last updated:** 2026-03-15 \ No newline at end of file