Insulin and cholesterol play important roles in basic metabolic processes in peripheral tissues. Both insulin and cholesterol can also act as signalling molecules in the central nervous system that participate in neuronal function, memory and neurodegenerative diseases. A high-cholesterol diet improves spatial memory in experimental animals. β-Amyloid, the toxic peptide in neurons of AD (Alzheimer's disease) patients, binds cholesterol and catalyses its oxidation to 7β-hydroxycholesterol, a highly toxic oxysterol that is a potent inhibitor of α-PKC (α-protein kinase C), an enzyme critical in memory consolidation and synaptic plasticity and implicated in AD. Oxidized cholesterol also can act as a second messenger for insulin. Oxidized low-density lipoprotein inhibits insulin-dependent phosphorylation of the signalling kinases ERK (extracellular-signal-regulated kinase) and PKB/Akt. In sporadic AD patients, insulin levels are decreased, suggesting links between AD and diabetes. Insulin signalling is also important in synaptic plasticity. Insulin receptors are up-regulated and undergo translocation after spatial learning. Insulin modulates the activity of excitatory and inhibitory receptors including the glutamate and γ-aminobutyric acid receptors and activates two biochemical pathways: the shc-ras-mitogen-activated protein kinase pathway and the PI3K (phosphoinositide 3-kinase)/PKC pathway, both of which are involved in memory processing. These findings point to a convergence at the biochemical level between pathways involved in AD and those important for normal memory.
- Alzheimer's disease (AD)
- protein kinase C (PKC)
Insulin and memory
Both insulin and cholesterol have long been considered part of the biochemical ‘furniture’ of the cell. However, recent evidence has shown that both insulin and cholesterol also perform specific roles in the brain during learning. There are also many potential interactions between pathways involving insulin and cholesterol in the brain and many commonalities in the roles of insulin and cholesterol in neurodegenerative diseases such as AD (Alzheimer's disease).
Insulin is a small protein consisting of a 21-amino acid α chain linked to a 30-amino acid β chain by two disulphide bridges. In the periphery, it is synthesized primarily in pancreatic β-cells. Glucose entering the β-cells through the GLUT1 and GLUT2 transporters increases the rate of metabolism and increases the ATP/ADP ratio, which leads to inhibition of ATP-sensitive K+ (KATP) channels, causing the β-cells to depolarize, which leads to Ca2+ influx and insulin secretion . Acetylcholine and stimulatory hormones, such as glucagon-like peptide I, potentiate insulin release by further increasing intracellular Ca2+ and activating PKC (protein kinase C) and PKA . Insulin increases glucose uptake in muscle and adipose tissue by translocating glucose transporters from an intracellular pool to the cell surface. Insulin also crosses the blood–brain barrier via specific receptors by receptor-mediated transcytosis.
Besides regulating glucose metabolism, insulin stimulates lipogenesis, diminishes lipolysis, increases amino acid transport into cells, modulates transcription and stimulates growth, DNA synthesis and cell replication. The finding that insulin also has specific effects on the brain has gained credence in the past 20 years. However, the question of whether insulin acts solely to regulate glucose uptake, or plays a more direct role in memory, is still unresolved. Distinguishing between insulin's effects on glucose levels and insulin's possible role as a neurohormone is difficult, because changes in peripheral insulin produce a variety of effects unrelated to memory. For example, animals with streptozotocin-induced diabetes have impaired memory  and insulin-enhancing therapeutic agents enhance memory . However, it has been difficult to determine precisely what is happening biochemically in these experiments. Potential use of insulin in neurodegenerative disorders is also hampered by insulin's well-known ability to induce profound and life-threatening hypoglycaemia.
Insulin signalling may also be important in synaptic plasticity. Insulin receptors are up-regulated and undergo translocation after spatial learning . Insulin modulates the activity of excitatory and inhibitory receptors, including the glutamate and GABA (γ-aminobutyric acid) receptors and activates two biochemical pathways: the shc-ras-mitogen activated protein kinase pathway and the PI3K (phosphoinositide 3-kinase)/PKC pathway, both of which are involved in memory processing. Activation of insulin receptors also activates muscarinic acetylcholine receptors by a PKC-dependent pathway, leading to potentiation of GABAA receptor currents . PKC may also contribute to insulin resistance and activation of atypical PKC could explain the hyperlipidaemia associated with the insulin-resistant state .
It has been suggested that, because of its importance for glucose utilization, insulin's role in memory could be most important in those aspects of memory relevant to finding food sources. Neuron-specific disruption of the insulin receptor in mice produces an increase in food intake, leading to obesity . Intraventricular administration of insulin increased retention of a passive avoidance training in rats  (Figure 1). Insulin receptor mRNA is found in the olfactory bulb, choroid plexus, pyriform cortex, amygdaloid nucleus, hippocampus and cerebellar cortex . With the exception of the hippocampus and the cerebellum, the mRNA levels generally parallel the protein levels . Although most of the insulin found in brain is imported from the periphery, some researchers [5,11] have provided tantalizing data suggesting that small amounts may also be synthesized in neurons.
Interactions between insulin and cholesterol
One of insulin's main effects in the periphery is to stimulate the activity of HMGR (3-hydroxy-3-methylglutaryl-CoA reductase), which catalyses the conversion of 3-hydroxy-3-methylglutaryl-CoA into mevalonate, in the rate-limiting step in cholesterol biosynthesis. Insulin increases HMGR levels by at least 10-fold in rat hepatoma cells . While it was originally believed that insulin acted primarily by reducing the phosphorylation state of HMGR (which increases its activity), recent results have shown that insulin also acts through phosphorylation of cAMP-response element protein . Paradoxically, cholesterol-lowering drugs (statins) that are HMGR inhibitors also act to stabilize HMGR and promote increased transcription and translation. Cholesterol is a feedback inhibitor of HMGR and also reduces expression of the enzyme. Another link between cholesterol and insulin is that Type II diabetes is associated with high synthesis and low absorption of cholesterol. Insulin-resistant patients have increased cholesterol synthesis . Under hyperglycaemic conditions, insulin also activates ACAT (acyl-CoA:cholesterol acyltransferase), the principal enzyme involved in cholesteryl ester synthesis .
Insulin also down-regulates the gene transcription of cholesterol 7α-hydroxylase and sterol 27-hydroxylase , two enzymes that degrade cholesterol to produce oxysterols in the pathway to bile acids. Both oxysterols and insulin are required for expression of transcription factors known as SREBPs (sterol-regulatory-element-binding proteins). SREBP-lc is an important mediator of insulin's ability to express glucokinase and genes involved in cholesterol and triacylglycerol synthesis . Binding of oxysterols to the nuclear LXR (liver X receptor) promotes binding of LXR to the SREBP-lc promoter. Oxidized cholesterol also can act as a second messenger for insulin. Copper-oxidized LDL (low-density lipoprotein) inhibits insulin-dependent phosphorylation of the signalling kinases ERK (extracellular-signal-regulated kinase) and PKB/Akt , suggesting that oxysterols may be involved in insulin resistance.
Cholesterol and memory
A high cholesterol diet increases memory retention in rabbits  and young rats (F. DuFour, V. Micale, T. Khan and D. Alkon, unpublished work). The importance of cholesterol for brain function is attested by the fact that brain itself is >2% cholesterol by weight. However, the mechanism by which cholesterol affects memory is unknown. Cholesterol biosynthesis is inhibited by circulating cholesterol, and cholesterol import into the brain is tightly regulated. Circulating cholesterol is bound up in lipoproteins (LDL and high-density lipoprotein), which enter cells by transcytosis at LDL receptors . However, almost all of the cholesterol in the brain is synthesized in the brain and little or none of the peripheral cholesterol crosses the blood–brain barrier. Thus it has been unclear to what extent dietary cholesterol can produce increases in brain cholesterol. Many studies have reported effects of a high-cholesterol diet on levels of β-amyloid and other molecules in the brain without measuring whether levels of brain cholesterol are changed by the treatment. Others have found only minor increases. For example, Kirsch et al.  reported that a high-cholesterol diet enhanced the amyloid β-peptide burden in the brains of transgenic mice with little effect on brain cholesterol levels. However, Sparks  found increased cholesterol levels in brains of rabbits fed a high-cholesterol diet.
Cholesterol metabolites such as neurosteroids or oxysterols do not suffer from the same limitation as cholesterol. Oxysterols are known to pass through the blood–brain barrier more readily than cholesterol . Neurosteroids can be synthesized in the brain from peripherally derived progesterone . Cholesterol metabolites synthesized in the periphery would find it much easier to reach the brain than cholesterol itself. Thus one possible explanation for the effect of cholesterol on memory is that elevations in cholesterol could increase the production of neurosteroids such as dehydroepiandrosterone, δ5-androstene-3β,17β-diol, pregnenolone and 7α-OH-dehydroepiandrosterone. GABAA, NMDA (N-methyl-D-aspartate) and cholinergic and sigma opioid systems are all potential targets of neurosteroids. Neurosteroids bind with nanomolar affinity to inhibitory GABAA receptors and with sub-micromolar affinity to excitatory glutamatergic NMDA, α-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid, 5-hydroxytryptamine 3, glycine, sigma type 1 and nicotinic acetylcholine receptors. Allopregnanolone and pregnanolone are the most potent known modulators of GABAA receptors; allopregnanolone is 20 times more potent than benzodiazepines at potentiating GABA-ergic neurotransmission.
Some data also suggest that there may be lower levels of neurosteroids in different brain regions in Alzheimer's patients. Neurosteroid levels were negatively correlated in one report with levels of phosphorylated tau protein and β-amyloid, two biochemical markers of AD.
Another possibility is that the high-cholesterol diet may affect the ratio of esterified to non-esterified cholesterol in the brain. Cholesterol esterification by ACAT has been implicated in the production of β-amyloid, the toxic peptide associated with AD. Cholesteryl esters are directly correlated with β-amyloid production  and pharmacological inhibitors of ACAT can decrease β-amyloid synthesis. While most of the cholesterol in the brain is free, much of this free cholesterol is found in myelin membranes . Elsewhere, most cholesterol is not free; in blood, for example, 68% of the cholesterol is esterified, resulting in a free cholesterol concentration of 1.6 mM. Dufour et al.  found that 66% of the total cholesterol in Alzheimer's fibroblasts is non-esterified, compared with only 23% in controls. This could represent an attempt by the cell to decrease the production of β-amyloid.
The improvement of memory by dietary cholesterol is also puzzling in view of cholesterol's effects on membrane fluidity. Quails fed a high-cholesterol diet develop severe atherosclerosis and hypercholesterolaemia, resulting in decreased membrane fluidity in erythrocyte membranes . Similar effects have been observed in rats and in human patients. Decreases in membrane fluidity have also been observed in patients with AD. Decreases in membrane fluidity can also be caused by inhibition of phospholipase A2. Decreasing membrane fluidity impairs memory, while increasing fluidity improves memory.
Cholesterol in AD
Several lines of evidence have implicated a role for cholesterol in AD:
(a) elevated serum cholesterol is associated with increased risk for AD ;
(b) AD patients have elevated levels of total serum cholesterol and LDL-associated cholesterol ;
(c) the ε4 mutant in the gene for ApoE (apolipoprotein E), an important cholesterol transport protein associated with LDL, is a risk factor for AD;
(d) cholesterol and cholesteryl esters can also directly regulate the generation of the toxic β-amyloid peptide in affected brain regions  and can inhibit β-amyloid clearance;
(e) aberrations in cholesterol transport produced by redistribution of presenilin increase cellular levels of β-amyloid ;
(f) cholesterol accumulates in the dense cores of amyloid plaques ;
(g) patients with advanced atherosclerosis have an increased risk of AD ;
(h) both APP (amyloid precursor protein) and β-amyloid bind cholesterol ;
(j) β-amyloid reacts catalytically with cholesterol, producing 7β-hydroxycholesterol, an oxysterol that is extremely toxic to hippocampal neurons .
Human trials are underway to determine whether statins slow progression or lower AD risk. However, it is already clear that the observed clinical benefit with statin therapy is greater than that expected through the reduction of cholesterol levels alone, suggesting that the benefits of statin may be mediated by other effects, such as improvement of endothelial function, reduction of the vascular inflammation, or antioxidation.
Despite this evidence, it would be premature to ascribe the effects of high cholesterol to a side effect of vascular disease. Not only is the correlation between atherosclerosis and AD not perfect, but other diseases associated with age and diet, such as diabetes, are also correlated with AD . Moreover, feeding cholesterol to rabbits produces a number of pathological signs similar to those observed in early AD, including amyloid-like plaques, and interferes with the ability of rabbits to perform difficult memory tasks. These pathological effects require not just cholesterol, but copper , suggesting that AD pathology may start at a point where cholesterol and copper biochemistry converge.
There is mounting evidence that cholesterol not only promotes AD indirectly by exacerbating cardiovascular disease, but also acts directly by interacting with APP. Cholesterol binds to APP and β-amyloid near the α-secretase cleavage site, and β-amyloid 1–42 competitively inhibits cholesterol binding to ApoE and LDL . β-Amyloid peptides contain a hydrophobic region or ‘patch’ (LVFFA) near the copper-binding site, containing two phenylalanine residues and three other hydrophobic residues. This hydrophobic region may participate in the binding of cholesterol. Increased solvation of this region has been suggested as a possible factor in the pathogenic differences between wild-type and the E22Q (Glu22→Gln) (‘Dutch’) mutation . Point substitution of Phe19 with threonine also affects the folding and plaque competence of β-amyloid .
Far from being a chemically inert lipid, cholesterol can mediate oxidative stress by reacting with oxygen in the presence of transition metals such as copper. In the presence of β-amyloid, the chemical reaction between cholesterol, copper and oxygen is greatly accelerated and produces 7β-hydroxycholesterol, an oxygenated form of cholesterol that alters the balance from production of the relatively benign sAPP to the toxic β-amyloid .
Insulin in neurodegenerative diseases
Both insulin and insulin-like growth factor type I promote neurite outgrowth, synapse formation and neuronal survival. The incidence of insulin resistance, a symptom of Type II diabetes, has been shown to be correlated with AD. The Rotterdam study showed that diabetes mellitus is associated with a higher prevalence of AD . Many Alzheimer's patients have abnormal insulin levels in the cerebrospinal fluid, suggesting that insulin processing may be abnormal. It was also found that insulin regulates the formation of phosphorylated tau, a major component of neurofibrillary tangles . Intracerebroventricular injection of streptozotocin or depletion of neuronal insulin receptors produces effects similar to AD . Taken together, these results point to a potentially important role for insulin in AD. It has even been suggested  that AD is caused in part by neuronal insulin resistance, making it a form of ‘brain diabetes’.
Cholesterol and insulin, two molecules involved in the fundamental aspects of energy metabolism in the cells, also play important roles in the central nervous system. Their interrelationships and their involvement in memory and in neurodegenerative diseases are only beginning to be explored.
Proteins in Disease: A Focus Topic at BioScience2005, held at SECC Glasgow, U.K., 17–21 July 2005. Edited by B. Austen (St George's Hospital Medical School, London, U.K.), C. Connolly (Dundee, U.K.), B. Irvine (Belfast, U.K.), M. Sugden (Queen Mary, London, U.K.) and V. Zammit (Hannah Research Institute, Ayr, U.K.).
Abbreviations: ACAT, acyl-CoA:cholesterol acyltransferase; AD, Alzheimer's disease; ApoE, apolipoprotein E; APP, amyloid precursor protein; PKC, protein kinase C; GABA, γ-aminobutyric acid; HMGR, 3-hydroxy-3-methylglutaryl-CoA reductase; IRS-1, insulin receptor substrate; LDL, low-density lipoprotein; LXR, liver X receptor; NMDA, N-methyl-D-aspartate; SREBP, sterol-regulatory-element-binding protein
- © 2005 The Biochemical Society