Lecture 7 – APOE: Breaking Alzheimer’s – The Definitive Lecture Series

About one in four persons have a genetic predisposition for Alzheimer’s disease and dementia caused by having one or two copies of the ɛ4 allele of the gene APOE. Over half of the persons with Alzheimer’s have this gene allele. Population-based genotyping for this genetic risk is not recommended because there are no widely accepted protocols for treating APOE ɛ4 carriers that are different from what is recommended for all persons concerned about Alzheimer’s disease and dementia. However, due to easily accessible genotyping services, more and more people are discovering that they have an increased risk of developing Alzheimer’s disease and dementia, and many of these persons want to know the details, and they want these details presented to them in a manner that they can understand. If you are one of these persons, this lecture is for you.  

To access the FREE seminars with full presentations and videos please visit Dr. Goodenowe’s resource site here. This is the article for Lecture 7 – APOE: Breaking Alzheimer’s – The Definitive Lecture Series.

In 1992, it was discovered that persons diagnosed with senile dementia of the Alzheimer’s type, had a disproportionately high prevalence of a specific genotype of a gene (APOE) that codes for a specific cholesterol transport protein called apolipoprotein E (ApoE). The role of apolipoproteins in cholesterol transport was first discovered in the 1960s. In the late 1970s, it was discovered that the APOE gene contained two single nucleotide polymorphisms (SNPs) that alter the structure of the ApoE protein made by this gene. Everyone has two copies of each of their genes, one from each of their parents, except for the X and Y-chromosome in males, of which there is only one each. The individual gene copies are called alleles. In the APOE gene alleles, the two SNPs change the amino acid that gets put into the protein. Some people get the amino acid cysteine, and some get the amino acid arginine. These two SNPs result in three possible combinations, referred to as ɛ2, ɛ3, or ɛ4. An ɛ2 allele codes for two cysteines, an ɛ3 for one cysteine and one arginine, and an ɛ4 codes for two arginines. So, since everyone has two alleles, there are six allele combinations (ɛ2ɛ2, ɛ2ɛ3, ɛ2ɛ4, ɛ3ɛ3, ɛ3ɛ4, and ɛ4ɛ4). It is this combination that is your APOE genotype. The best way to think about your APOE genotype is like a different blood type. The most common genotype is ɛ3ɛ3 (about 60-65%), the next common is ɛ3ɛ4 (15–25%), then ɛ2ɛ3 (10–15%), and all the other combinations account for less than 5% of the population. The ɛ2ɛ2 and ɛ4ɛ4 are special cases with special disease risks.  

The discovery that these simple genotype differences were associated with the observed incidence of Alzheimer’s dementia ignited a firestorm of research activity the likes of which the world had never experienced before. Within a few short years, it was discovered that the ɛ4 genotype was associated with higher levels of brain amyloid in both demented and non-demented elderly persons and the incidence of dementia, with respect to the different APOE genotypes, behaved like a medicinal chemistry dose-response curve, with the rate of dementia increasing in each group of people based on their genotype from ɛ2ɛ3 (the group with the lowest incidence) to ɛ3ɛ3 to ɛ3ɛ4 to ɛ4ɛ4 (the group with the highest incidence). The most amazing thing about this discovery is the one thing that virtually nobody talks about. Everyone is so abuzz about how bad it is to have an ɛ4 allele that they miss the most important point of this discovery — that the epidemiology of the APOE gene dosage revealed that a low rate of dementia is possible. Persons with the ɛ2ɛ3 genotype have a very low rate of dementia. If we could reproduce the ɛ2ɛ3 phenotype in everyone, we would eliminate over 75% of all dementia cases; this is not a trivial point to be brushed over. This simple fact proves that Alzheimer’s disease and dementia are curable and preventable.  

In 2019, I published detailed epidemiological research data describing the relationship between the APOE genotypes and blood plasmalogens in 1205 persons aged 58 to 104 (average age was 84). The overall finding of the research was that high blood plasmalogens reduced the risk of dementia in ɛ4 carriers to that of an ɛ3 carrier, the risk of dementia in an ɛ3 carrier to that of an ɛ2 carrier, and the risk of dementia in an ɛ2 carrier to virtually zero. This was the first scientific evidence that the increased risk of Alzheimer’s associated with the APOE ɛ4 gene allele can be neutralized. This is the first crack in what has been a 30-year siege on the seemingly impenetrable APOE genetic wall. DHA-plasmalogens can change the biochemistry relating to amyloid deposition and the risk of dementia of an ɛ4 carrier to that of an ɛ3 carrier. This is an observable fact, but the question is how? More importantly, how far can we extrapolate this biochemical mechanism? Can we reduce the risk of Alzheimer’s in everyone to that of an ɛ2 carrier or even less, regardless of their underlying genotype? Time will tell.  

A genetic risk factor is not magic. It must act through a biochemical mechanism. The mechanism is somehow related to cholesterol regulation and maintenance in the brain. Plasmalogens and the ApoE protein are like two partners in a three-legged race. APOE ɛ4 carriers have a reduced ability to efflux and regulate membrane cholesterol levels in the brain due to the physical limitations of the ApoE ɛ4 protein structure — like running on a sprained ankle. Persons with high brain DHA-plasmalogens have an increased ability to transport and regulate membrane cholesterol due to the improved function of the other systems involved in cholesterol efflux. Because the APOE genotype affects a lot of people (one in four) and is responsible for up to three out of four Alzheimer’s cases, let’s build up our knowledge from the facts and try to build the best hypothesis we can that fits the data.  

ApoE is an apolipoprotein. Apolipoproteins are lipid transport proteins. The bulk (60%) of the human body is water. About 90% of the plasma in your circulatory system is water, and about 70% of the cytoplasm inside each of your cells is water. However, your body needs to be able to transport lipid molecules throughout the body in extracellular fluids (blood plasma and brain cerebral spinal fluid), in intracellular fluids between compartments within the cells of your body (called the cytoplasm), and in the fluid that exists between cells (called the interstitial fluid). Lipoproteins are what your body uses to do this. They are very water-soluble — so soluble that lipids, which are insoluble in water, can catch a ride on them to go wherever the lipoprotein is going. The apolipoproteins act like a scaffold upon which various lipids can bind, and they also act like a lasso that holds lipids in a tight group. The different apolipoproteins have different abilities to bind different lipids. Different apolipoproteins also have different cellular and compartmental docking stations, and that is how the body manages the distribution of lipids from one cell type to another. The body uses the various apolipoproteins to build the larger and final lipoproteins that are used in the circulation. The main purpose of lipoproteins is the transport and regulation of cholesterol. The transport of other lipids like triacylglycerols and phospholipids is important too, but their regulation is secondary to cholesterol. This is abundantly clear if you view the human body from a system design perspective.  

Maintaining cholesterol homeostasis is a “system critical” process. So, before we get into the weeds regarding the special weirdness of all things APOE, it is important to understand why we have such sophisticated cholesterol transport and regulation processes in the first place. This story starts with what makes cholesterol unique relative to the other lipids in the body. All the phospholipids and triacylglycerols have interchangeable units of head groups and fatty acid tails. Their structures resemble something like sticky noodles hanging from a fork. Cholesterol is more like a sticky pancake with a handle like a frying pan. A phospholipid bilayer that has only noodles in the center can barely stick together. However, if you start putting some sticky pancakes in there, the noodles stick to the pancake and cannot move around as much. The more pancakes you put in, the less the noodles can move around.  

The other big difference between cholesterol and other lipids is that the chemical structure of cholesterol is virtually indestructible, which makes it very hard to make. It takes more than 35 biochemical steps to make cholesterol, and once it is made, the pancake structure might as well be titanium — the body cannot even digest it. All we can do is oxidize the handle and edges and then excrete it in the bile and feces. So, why make something so indestructible? So you can recycle it and reuse it repeatedly in all compartments of the body — unlike the other lipids in the body, which can be digested into food energy.  

By varying the concentration of cholesterol here and there, the body can create any type of membrane structure. The different proteins embedded in the membrane all have different cholesterol likes and dislikes that influence what part of the membrane they prefer and how much of one protein versus another is in a given membrane region. Cholesterol is also the precursor for all the steroid hormones in the body. So, cholesterol is pretty important. Because of this importance, it is critical that cells can make very fine adjustments to their cholesterol “settings” when needed. To do this, a cell must be able to intake or make more cholesterol when it needs more, and it equally needs to remove cholesterol when it needs less. Since it is energetically demanding and time-consuming to make cholesterol, there needs to be a quick access supply. Since cholesterol cannot be digested, it needs to be exported on demand when there is too much. This is the basis of the general mechanism of cholesterol regulation and why the import and the export side are mostly independent of each other. It is the cell that is in control of its needs.  

Here is how cholesterol regulation works: each of the cells of the body can make cholesterol, but it is mostly made in the liver in the periphery and in astrocytes in the brain. There are three supply chains for cholesterol in the liver and two in astrocytes: endogenously synthesized, recycled, and dietary absorption (liver only). The brain and the periphery have the same conceptual cholesterol regulatory system, but the mechanisms are very different. The liver takes these three sources of cholesterol and either converts them into a cholesterol ester to package it on a lipoprotein particle that enters the blood to eventually supply cholesterol to all the cells of the body, or it converts cholesterol into bile acids and sends it to the intestines where it is used to solubilize fatty foods and gets recycled or excreted. This cholesterol ester is essentially the pancake with a sticky noodle attached to it so that it more closely resembles a fatty acid lipid, like a triacylglycerol. All these various fatty acid-based lipids are packaged on lipoproteins for distribution. If a cell is hungry and needs cholesterol, it will grab onto an LDL string and pull it into the cell. The hand that grabs the LDL string is called the LDL receptor. Once the cell pulls it into the cell, it puts the LDL particle in a special compartment called the lysosome, where it gets digested. The fatty acid cholesterol tail is removed, and the free cholesterol is sent to the membrane and other intracellular compartments for the cell to use as needed. 80-85% of all the body’s cholesterol is stored in the membranes as free cholesterol, not a cholesterol ester like that found in LDL cholesterol. The body only uses esterified cholesterol to transport it or temporarily store it. Esterified cholesterol is essentially inactivated cholesterol. If a cell has a prolonged demand for cholesterol, the cell will upregulate the number of LDL receptors so that it can import more. Accordingly, circulating LDL levels are determined by the combination of the liver’s ability to make cholesterol and LDL particles and the metabolic demand of the cells of the body for cholesterol. Blood LDL levels do not force themselves on the cells of the body, just like the gas in your car’s gas tank does not force itself into your car’s engine. Your cells, just like your car engine, take what they need — when and as they need it.  

All the cholesterol inside the cell and in the membrane is free cholesterol. The outer plasma membrane contains most of the cholesterol, and this is the primary reservoir that the cell uses to regulate its need for cholesterol. If the membrane is full of free cholesterol, this excess free cholesterol stimulates cholesterol efflux, inhibits internal cholesterol synthesis, and inhibits cholesterol uptake. The efficiency and capacity of your cells to export cholesterol are roughly correlated to your blood HDL (high-density lipoprotein) levels. For the most part, the LDL system and the HDL system are in different universes. The main apolipoprotein used to make HDL (ApoA) is different from the apolipoprotein used to make LDL (ApoB). HDL does not get absorbed into cells. Whereas blood LDL levels equilibrate to your cellular import needs, your blood HDL levels equilibrate to your cellular cholesterol export needs. Healthy cells make cholesterol. Healthy levels of cholesterol recirculation mean that cells can rapidly adjust cholesterol levels so that membranes are neither starved of nor fat with cholesterol. A cell signals to the circulating HDL particles that it has cholesterol to be collected by up-regulating HDL receptors. HDL is like a trucking company. If the demand for cholesterol shipments is consistently high, the company buys more trucks, and you notice more trucks on the freeway. Just like LDL cannot force a cell to accept cholesterol, HDL cannot force a cell to release cholesterol. Approximately 90% of the plasma lipoproteins in the blood are either LDL or HDL, and you typically have two to three times more LDL than HDL. The healthiest levels of total cholesterol, as measured by all-cause mortality, are in the 220–240 mg/dL range, and for HDL, in the 60–90 mg/dL range. 

OK, so what does all this have to do with the APOE genotype and Alzheimer’s disease and dementia? ApoE is a relatively minor apolipoprotein in the periphery (<5%), and it is found in all lipoprotein density classes, but mostly in VLDL and HDL particles. It interacts with the LDL receptor and is absorbed into the cell and is digested in the lysosome. What makes ApoE special is that ApoE is the only apolipoprotein synthetized in the brain (>99%). It makes up about 75% of the cerebral spinal fluid lipoprotein with approximately 25% ApoA-I coming from the blood (ApoA-I is the predominant apolipoprotein in your blood HDL). So, in the brain, there are two types of HDL particles — ApoE and ApoA-I. The brain ApoE HDL is essentially equal to blood LDL, and the brain ApoA-I HDL is equivalent to blood HDL.  

The concept of cholesterol regulation is the same in the brain as it is in the periphery. The only difference is that the brain has a different organizational structure and different cholesterol needs. The brain only makes up about 2% of the total body mass, but it has about 23% of all the cholesterol in the body. Maintaining cholesterol homeostasis is so important for the brain that it makes and regulates its own cholesterol with no reliance on cholesterol from the liver or blood. The periphery has a central cholesterol control center — the liver. This control center is physically remote from the cells that it supplies. Accordingly, in the periphery, there is a long-distance interstate delivery and return system, which is provided by the circulatory system. However, in the brain, the site of cholesterol synthesis (the astrocytes) is locally controlled and physically proximate to the cells that are being supplied. Just like you would not use a big semi-tractor trailer to make local deliveries, the brain does not need large LDL particles. So, it only uses small HDL and even immature HDL particles. Likewise, the cerebral spinal fluid and the interstitial fluid in the brain are not a system of fast-flowing one-way arteries and veins. It is more like a bunch of narrow two-way local streets. The cerebral vascular system still provides bulk delivery of nutrients and bulk removal of waste, but this is not done directly. The blood-brain-barrier (BBB) is how nutrients and waste pass between the brain and the blood. The BBB is also a protective barrier. Its purpose is to only allow into the brain materials that the brain needs and nothing more.  

Since most of the metabolic activity and regulation in the brain is local, and there is no appreciable circulatory system; metabolic recycling is critical for maintaining efficiency. The most efficient system of recycling is cellular symbiosis. This is a system where the waste of one cell is used as fuel or as a precursor for biochemical activities in another cell. This dramatically reduces both energy supply and waste removal needs. Besides passive metabolic symbiosis, there is also active signaling between symbiotic cells where support cells can turn up or down certain metabolic processes based on signals from the active cell.  

The more specialized a cell becomes, the more dependent it becomes on its support system. In the brain, neurons are very specialized. They cannot survive without the biochemical support of astrocytes for all the biochemical activities in their cell bodies or without the oligodendrocytes that support and protect the specialized bioelectrical transmissions carried out by their axons. The purpose of the brain support structure is so that neurons can focus all their energy on efficiently executing their duties. Neurons have most of the machinery to do most things if they must. If the support structure begins to fail, neurons must divert their energy to other basic survival tasks, which reduces their function.  

The symbiotic relationship between astrocytes and neurons in the working brain makes it very difficult for scientists to study how neurons work and how astrocytes work in the laboratory. Most scientific studies look at neurons and astrocytes in isolation from each other. For the most part, these studies are valid and provide valuable information regarding what biochemical processes occur in one type of cell versus another. However, the limitation is that it is very difficult, if not impossible, to understand how these systems behave in a normal functioning brain. Therefore, different hypotheses are proposed. Often, more than one hypothesis can be developed that fits a single set of data. Under these circumstances, scientists often have philosophical arguments that can become quite heated while we await additional data that will disprove one or more of the tentative hypotheses.  

For all practical purposes, ApoE is the only apolipoprotein produced in the brain. This means that the brain uses the same apolipoprotein (ApoE) for both LDL-like and HDL-like functions, and it does this in both cholesterol-producing and cholesterol-receiving cells. These different cells are adjacent to each other, and these cells are biochemically integrated with one another in a symbiotic manner. This makes it almost impossible for scientists to design experiments to understand these relationships.  

Numerous elegant experiments have resulted in several solid observations. There are four types of ApoE receptors on the outer surface of brain cells. The first is the LDL receptor. This serves the same purpose as in the rest of the body — to internalize and feed the cell cholesterol. The other three receptors (ABCA1, ABCG1, and ABCG4) are all HDL receptors. It is the selective use of these three receptors that allows the brain to use a single apolipoprotein (ApoE) for different purposes. We know that astrocytes make HDL-ApoE to act and behave like peripheral LDL. But astrocytes do not have the same enzymes that the liver uses to make LDL (the liver creates the esterified cholesterol LDL core inside the cell using an enzyme called ACAT2). Instead, astrocytes use the equivalent of peripheral HDL mechanisms. This means that it is a two-step process. The first step is to make the ApoE protein and the cholesterol metabolite and export both out of the cell as a nascent or immature discoidal HDL-cholesterol particle. In this form, it is pure LDL-like cholesterol food. Neurons can directly absorb these particles as cholesterol food. The formation of a mature HDL particle requires the formation of an esterified cholesterol core. For this to happen, the nascent HDL particle needs to bind to the outside of the astrocyte (using the ABCA1 docking site) to stimulate the efflux of new cholesterol, and then this newly exported cholesterol (not bound to ApoE) reacts with an extracellular enzyme called LCAT (Lecithin Cholesterol Acyl Transferase). This is how astrocytes re-create the liver LDL synthetizing process using an HDL system. The production of ApoE, the release of ApoE-cholesterol particles by astrocytes, and the uptake of ApoE by neurons do not appear to be affected by the APOE genotype.  

It is the ability of a neuron or astrocyte to efflux cholesterol that is affected by the APOE genotype. Several robust scientific publications describe how the interaction between the ApoE protein and the three ABC docking sites are affected by the different APOE gene alleles. ApoE-HDL, made using the ɛ2 allele, has very strong cholesterol efflux capacity, whereas ɛ3 has moderate efflux capacity, and ɛ4 has very weak efflux capacity. The difference in efflux capacity between the three proteins is related to the ability of the protein to form disulfide bridge dimers using the cysteine amino acid (cysteine has a free S-H bond that can react with another cysteine to form the S-S di-sulfide bridge). The two cysteine amino acids in ɛ2 and one cysteine amino acid in ɛ3 allow for the formation of this dimer. The ɛ4 protein cannot form dimers because it does not contain cysteine. This effect appears to occur at all three ABC docking sites. The ABCG1 and ABCG4 docking sites are used in the brain the way that normal HDL function works in the rest of the body. These sites are for the traditional fully formed HDL particle to dock, and they signal the inside of the cell to efflux free cholesterol out of the cell, which is then captured by the docked HDL particle. The astrocyte uses ABCG1, and neurons use ABCG4.  

The APOE genotype is a predictor of amyloid accumulation and is only useful if we do not know brain amyloid levels. We now know that the most proximate structural effect of the different APOE gene alleles is the formation of structurally different proteins. The structural differences relate to the ability of the different ApoE proteins to form disulfide bridges based upon their respective number of available cysteines resulting in a gene dosage effect that is 4x for ɛ2ɛ2, 3x for ɛ2ɛ3, 2x for ɛ3ɛ3, 1x for ɛ3ɛ4, and 0x for ɛ4ɛ4. Cholesterol efflux efficiency and capacity are affected by the ability of ApoE proteins to form these disulfide bridges. The interaction that is affected by the impaired ability to form these disulfide bridges is the interaction between the ApoE protein with ABC docking sites. It is the ability of the ApoE-ABC complex to stimulate cholesterol efflux that is affected by the different APOE genotypes. I mentioned above that any cellular modifications to the lipid raft region of the cell modifies Aβ1-42 production, and this is specifically true of the ABCG1 and ABCG4 docking sites. Increasing ABCG1 and ABCG4 activity decreases Aβ1-42 formation and decreasing ABCG1 and ABCG4 increases Aβ1-42 formation.  

Based on this data, the accumulation of amyloid that is observed to correlate with the APOE genotype alleles is caused by the different effects the alleles have on cholesterol efflux. If the ability of brain ApoE lipoproteins to stimulate cholesterol efflux is one of the underlying mechanisms that causes amyloid accumulation in the brain, then if you take the cerebral spinal fluid from persons with amyloid accumulation and test it for cholesterol efflux capacity in a separate laboratory setting, then this cerebral spinal fluid should show reduced cholesterol efflux capacity — and that is exactly what happens.  

Another interesting thing about Alzheimer’s is that the level of Aβ1-42 in the cerebral spinal fluid is lower in persons with Alzheimer’s versus healthy controls. Recently, a definitive study on the relationship between APOE genotype, brain amyloid deposition, and cerebral spinal fluid amyloid levels was published. In this study, it was confirmed that the level of Aβ1-42 in cerebral spinal fluid was dependent upon APOE genotype with ɛ4ɛ4 levels less than ɛ3ɛ4 which was less than non-ɛ4 carriers. More importantly, when the researchers corrected for brain amyloid levels, they observed that the cerebral spinal fluid Aβ1-42 levels were correlated with brain amyloid levels and that this correlation had nothing to do with the person’s APOE genotype.  

The APOE genotype is a genetic predisposition for elevated membrane cholesterol due to different cholesterol efflux capacities. If this genetic predisposition is active, we see elevated brain amyloid. Brain amyloid levels are a biomarker of the net cholesterol efflux capacity of brain neurons.  

Just for completeness, there is one more mechanism relating to cholesterol efflux that is affected by the type of ApoE protein that you make. When cells make the ApoE protein, which all brain cells can, including your neurons, it is made specifically to be secreted to the outside of the cell and to then become an HDL particle. When it is secreted, it takes cholesterol from the membrane with it. This is another cholesterol efflux mechanism, and it has been extensively studied in macrophages like those found in atherosclerotic plaques. It turns out that this process is affected by how strongly the ApoE protein binds to the membrane LDL receptor. Strong binding reduces cholesterol efflux. ApoE containing ɛ2 is a weak binder, ɛ3 is a moderate binder, and ɛ4 is a strong binder. This mechanism has not been extensively studied in the brain, but it could also contribute to the cholesterol efflux differences between the APOE genotypes. 

Many scientists see this data as frustrating, but I see it as remarkable and enlightening. The “in-your-face” APOE genotype effect on amyloid deposition and dementia has forced and incentivized the scientific community to perform a vast array of experiments. As you can see from the above analysis, everything leads back to defects in membrane cholesterol homeostasis, which, based upon the core function of the ApoE protein, was the most logical point of failure all along.  

What this data also reveals is that your APOE genotype is not the only factor influencing cholesterol efflux and amyloid deposition. Both the deposition of amyloid and the risk of dementia associated with the different APOE genotypes are modified or dependent upon brain and blood plasmalogen levels. Why is this so? Just like genetic risk factors are not magic, biochemical risk factors are not magic either.  

The ApoE-ABC docking system interaction described above occurs on the outside of the membrane. This interaction only represents half of the process. The inside of the membrane part of the process involves an enzyme called ACAT1 (Acyl-CoA:cholesterol acyltransferase). All the mechanisms involving ACAT1 are not known. However, we know that when ABC docking sites are activated, it is the ACAT1 cholesterol pool that is rapidly depleted. This is the first cholesterol pool to leave the cell. We also know that ACAT1 appears to pull cholesterol out of the membrane into the cytosol in preparation for its efflux out of the cell. Reduced ACAT1 activity results in cholesterol accumulating inside the cell. Increased ACAT1 counteracts genetic cholesterol accumulation disease mechanisms, such as Neimann-Pick Type C disease. Increasing ACAT1 activity reduces Aβ1-42 formation and reducing ACAT1 activity increases Aβ1-42 formation. Membrane DHA-plasmalogen levels alter the activity of ACAT1 such that low membrane DHA-plasmalogen levels result in reduced ACAT1 activity and reduced cholesterol efflux and elevated membrane DHA-plasmalogen levels result in increased ACAT1 activity and increased cholesterol efflux. Elevated membrane DHA-plasmalogen levels also have a separate and independent effect of increasing α-secretase activity, and decreased Aβ1-42 formation. This the biochemical mechanism through which high DHA plasmalogen levels neutralize the increased risk of Alzheimer’s in persons with an APOE ɛ4 allele. 

In Lecture 7 – The Biochemistry of APOE, Dr. Goodenowe explains the relevant research and literature relating to the biochemistry of the APOE genotype. The lectures integrate Dr. Goodenowe’s own research and over 50 years of research from leading researchers from around the world.