Alzheimer's, Memory Disorders and Diseases

Steven T. DeKosky, MD
Vice-President and Dean University of Virginia School of Medicine

---

Q:How does it feel to be back at the University of Florida? How was your experience here at UF Medical School? How would you rate the education you obtained here overall?

A:It’s always good to be here, I’ve been back fairly often since I left the second time (graduated in ’74 and finished my residency in ’78). I started with graduate school here for two years, then interned at Johns Hopkins (internal medicine) then back here for my neurology residency. My undergraduate medical education here was fantastic – the school was a bit different in those days. The classes were around 70 students and I think we knew the faculty a bit better than most medical students know faculty today. As virtually all the medical schools have gotten larger, the rotation of teaching physicians has gotten larger with smaller exposures to the students and of course what they learn and the options for going into different fields is broader and more complex. When I was a student, many of these areas of sub-specialization and sub-sub-specialization were just getting started. My undergraduate medical education prepared me extraordinarily well for my post-graduate studies and for my internship. The teaching, especially in the department of neurology (which was full of young hotshots like Ken Heilman and Mel Greer), was what brought me back here from Baltimore for residency and I had a terrific time. I’d have to say my four years of undergraduate medical school were probably the best years of my life. We had a fantastic time, learned a lot, played a lot of sports, had a lot of fun, and got straight enough to show up for graduation. All in all it was a good blend of academics and real life outside the hospital.

Q:Why did you choose to go into Neurology?

A:Neurology is by far the most interesting area there is. I came into medical school from graduate school in neuroscience and knew that I wanted to study the brain and its relationship to behavior, so that brought me to psychiatry, neurosurgery and neurology. I spent a lot of time debating among those three areas and had an especially difficult time with neurosurgery. At that time, Dr. Rhoton had just become the new chair of UF neurosurgery and I spent a lot of time with him trying to decide what I was going to do. I ultimately decided that neurology would give me the best chance to pursue the kinds of changes in behavior and relationship to brain function that I wanted to study.

Q:How did you become interested in Alzheimer’s disease?

A:Well, I had initially been a graduate student in comparative and physiological psychology on campus and was the first UF graduate student in neuroscience because I started the summer before the formal program opened in the fall. So I came to clinical training from wet lab experience with brain studies of rats, cats and other animals. I wanted to study the relationship between brain and behavior and when I went to do my postdoctoral fellowship in neurochemistry (after finishing neurology training), I talked to my advisor at the University of Virginia. He had just started to do some work on brain tissue of patients with Alzheimer’s, normal elderly, and young people using samples he’d brought with him from Harvard. These samples were not available in the majority of institutions around the country. Now with the expansion of human studies, there are brain banks centralized in the NIH as well as all of the Alzheimer’s centers and a variety of other institutions and for different kinds of diseases. Now, It’s a much more widespread operation that it was then. This was the perfect chance to start looking at disease and relationship of changes to actual tissue in the brain at a time when the only information you got about patients was in autopsy. So I started looking at cell counts and markers for synapses, connections between neurons and markers of the arbor or tree of individual neurons. These were all examined biochemically in microchemical assays and I moved from that to microchemical assays of neurotransmitters after the findings of abnormal neurotransmitters in Alzheimer’s disease began to emerge. As the field grew, it has proven to be a terrific place to both utilize the behavioral neurology training I had here as well as the neurochemical and molecular biology training I received later.

Q:What are the risk factors and research criteria for a diagnosis of Alzheimer’s?

A:The diagnostic criteria are fairly straightforward, especially for the typical case. You have to have a slow and insidious onset of a cognitive deficit associated with short term memory and at least two cognitive domains affected. The four major domains we look for are memory, language function, executive or frontal lobe functions and visio-spatial functions. Any one of those produces a specific disorder (e.g. an isolated problem with language disorders is an aphasia of some type). In neurology, the NINCDS ADRDA criteria (which are now over 20 years old) only require that you have those cognitive changes and that you can document them with a neuropsychological assessment and that you don’t have any other disease which might have caused the cognitive problem, e.g. hepatic/renal failure. In psychiatry there is a higher bar - these behavioral problems must interfere with social or occupational function. So you have to arguably be a little more severe to meet the DSM-IV criteria. There are some good arguments that if you have a memory problem severe enough to bring you to the attention of a physician that it’s likely interfering with your social and occupational activities. We don’t necessarily know that, but it’s one of the issues that must be squared away as we have moved to earlier detection of Alzheimer’s disease or other incipient dementias. We see people now who quite certainly have Alzheimer’s disease but actually do not meet the formal criteria for it yet. We have just published a paper suggesting some new diagnostic criteria through use of a biomarker and recognition of early cognitive changes (Dubois B, Feldman HH, Jacova C, DeKosky ST, et al. Research criteria for the diagnosis of Alzheimer’s disease: revising the NINCDS-ADRDA criteria. Lancet Neurol. 2007 6(8):734-746) – this is probably what we are going to focus on for the next 5-10 years – nailing down as specific a diagnosis as early as you can in anticipation of medicines that will interfere with the pathology and the decline as early as we can. There doesn’t appear to be a major difference in your susceptibility to develop the disease as a function of gender. On the other hand, “early” doesn’t mean in your 40’s or 50’s as opposed to in your 70’s or 80’s, it means early in the course of the development of the pathology that leads to the behavioral manifestations. So, I would rather find someone to give a medication to who is having early changes in memory loss if I can prove with a biomarker that this is incipient Alzheimer’s disease, than wait until someone meets full criteria which may be 4, 5, 6, even 10 years later wherein likely irreversible damage has already been done. What we would like to do is get to where most of our preventive medicine is now anyway – ask who is at risk long before they are symptomatic or even have pathology. I think you can compare this now to where cardiology was 25 years ago. If you want to battle heart disease in a major way, you aggressively pursue hypercholesterolemia, hypertension, other markers of being obese, not exercising, etc. all of those things that we know in midlife make a difference 30 years later. It’s likely that Alzheimer’s disease is going to end up being a disorder that has a similar way to fight it off. There are still people who discuss this, but in terms of trying to prove whether or not you could make a difference in the huge epidemic that’s going to occur over the next 2 or 3 decades – this is probably the most powerful argument we could make.

Q:You have conducted a number of studies on traumatic brain injury, and we know that caspase and calpain inhibitors are major players in brain injury. Can you elaborate?

A:Brain trauma is very complicated and we’ve actually done some of the first studies – staining the brains of football players who become demented in midlife – they have cognitive changes resembling Alzheimer’s or have neurofibrillary tangles from multiple traumas. It’s widespread, but why it appears to cause late life dementia, we’re just not sure. It may be that it interferes with the repair or removal characteristics, e.g. APOEe4. Look at boxers – cognitive impairment later in life is correlated directly with the number of rounds they box and it doesn’t matter whether they won or lost (although I guess if you lose early and consistently you don’t box as many rounds). The exposure to repeated head trauma plus being positive for APOEe4 makes you all the more likely to develop cognitive impairment. Many of the boxers in the U.S. are African-American and African-Americans have higher incidence and prevalence of APOEe4 (24% vs. 12-15% of Caucasians), thus have a higher risk of cognitive impairment. We haven’t addressed that in this country and  may not do so in football as it is such a popular pastime, but most neurologists think boxing should be banned because the objective is to hurt our organ!  I ask urologists all the time – how would they feel if a sport’s objective was just to beat up their organ? Overall it will take chemistry to figure out what the best therapies are, e.g. calpain-related metabolism. It will be difficult, but we need to figure out a way to undo axonal shearing. In the meantime, prevention is the watchword – keep your helmets on.

Q:What biomarkers are used in traumatic brain injury?

A:A good biomarker in Alzheimer’s, when you can’t always get direct access to the brain, would be something that changes in a relatively noninvasive way to tell you about the presence of a disease and its severity. There are a host of biomarkers in brain trauma based upon what is being disrupted and where therapies should be targeted. The trouble with brain injury is the initial injury – the instant loss of cells causes a massive cytotoxic release of glutamate, the likely irreversible tearing of axons. The use of biomarkers or the examination of the pathologic processes is key to figuring out the various types of therapies people will need. Not many people believe one kind of intervention is helpful - the best course of action would likely be combination therapies; however, in the drug creation process you must do one drug at a time, so this creates more challenges for us in this field.

Q:What occurs in the brain of someone who ages normally versus someone with dementia? In other words, what is causing memory loss?

A:It’s important to keep in mind that dementia is a syndrome with probably 75 or so different causes, but the vast majority of cases do turn out to be Alzheimer’s disease or Alzheimer’s along with some other pathology; most commonly they are cerebrovascular diseases. So Alzheimer’s disease is a disorder that probably starts with early biochemical changes – we’re not exactly sure what precipitates it. In people who have a mutation that forces them to get it (these are the early onset people who get it between the ages of 30 and 50), the disease is represented in 50% of every generation – depending upon whether they inherit the defective gene or not. All of those people have gene defects that cause an alteration in the amount of a particular protein in the brain: amyloid precursor protein (APP) and the metabolism of the APP is altered so you have elevation of a particular snippet of it: Aβ1-40 or Aβ1-42. This piles up in the brain to form the plaques that Alzheimer saw in his first patient. We also think that the initiation of the rest of the pathology, (e.g. cell death and especially formation of neurofibrillary tangles) stems from this altered amyloid metabolism. In people who only carry a risk gene - a variant of a gene that increases the likelihood but does not guarantee that you get Alzheimer’s is the e4 variant apolipoprotein (APOE) – there is poor removal of beta amyloid protein in the brain. If you replace an animal’s endogenous APOE with the human e4 variant, they don’t repair their brain and peripheral nerves as well as they could. So there’s something about APOE whose normal function is to ferry cholesterol around – which may be why it doesn’t move the cholesterol effectively in the periphery, which is associated with decreased repair of nerves. There may be something about the e4 variant that interferes with removal of beta amyloid from the brain, and that may be why it builds up in even normal people as they age, precipitating Alzheimer’s. There’s a lot more to the story, but that is the major preferred theory. This is in part because in animal models we have found ways to interfere with the metabolism of Aβ and these strategies are underway in clinical trials in humans.

Q:How do amyloid-beta and tau relate to Alzheimer’s and traumatic brain injury?

A:We’re not sure how amyloid and tau relate to one another. The current thinking is that it is amyloid itself – before it forms plaques (which may be some sort of end stage or deposition of elevated concentrations of beta peptides) – the damage may really come from single peptides or trimers or oligomers and these somehow alter metabolism associated with oxidative stress with neuronal dysfunction or death precipitating abnormalities in tau metabolism. Tau is a normal protein that can be thought of as a railroad tie on a railroad track – the metal part of the rail is what lets the axon of the neuron carry information or proteins back and forth from the synapse. If you disturb the ties i.e. tau, you can’t transport, which may be the reason why synapses die off in this disease even before neurons are lost. This happens with tau – it gets hyperphosphorylated, it gets decorated with too many pieces of the metabolic process that it normally controls very carefully and it irreversibly crosslinks with itself and forms an insoluble protein kind of like a fried egg. So, we’ll never remove tau in neurofibrillary tangle form from humans, only attempt to cease further formation, which is what anti-tau medications are directed toward. In brain trauma, APP goes flying up immediately - within minutes to hours of brain injury and gets broken down into soluble APP and also increases the amount of beta amyloid in the brain. This is detectable – you can actually see plaques in the brain of a 30-year-old two hours after brain trauma because so much beta amyloid has been created. Unfortunately we don’t know the fate of those cells as that example is of a sample of cells that were removed to save the patient by decreasing pressure and removing a hematoma. However, we know that in people who have suffered severe brain traumas (i.e. loss of consciousness for 30 minutes or more) in their lifetime are much more likely to develop Alzheimer’s disease later. The hypothesis is that they were probably going to get it anyhow, but if you look at the rates of acquisition of Alzheimer’s in epidemiological studies, in people who have had head trauma vs. people who just get Alzheimer’s anyway, you see the curve shift back several years. So they probably had this load of beta amyloid which may have done damage or initiated some of the pathological processes, the trauma may have had an affect on the brain that left it less able to fight off Alzheimer’s – all of those things would lead to an earlier manifestation of disease in someone destined to get it. If you carry the e4 allele and are severely head-injured, you are also more likely to develop Alzheimer’s. If e4 is associated with “worse” removal of Aβ and you have a massive increase of Aβ after trauma, you have stressed the brain’s removal system and that may carry on for a long time.

Q:Please tell us about Pittsburgh Compound B (PIB).

A:We didn’t name it that, but we’re forever grateful to our colleagues at Stockholm who did. We sent them 2 compounds which they termed Pittsburgh Compound A and Pittsburgh Compound B. Pittsburgh Compound A wasn’t as good of a PET ligand as PIB but it is the result of 8 years of incredibly painstaking work. My colleagues Chet Mathis (head of the PET center) and Bill Clunk (geriatric psychiatrist with a PhD in pharmacology) and I searched for a ligand that would get in, bind to amyloid plaque, and get out. We actually started with studying stains – what exactly is it about them that allows them to stain amyloid and nothing else? We started with thioflavins and other compounds that stained beta-pleated sheets. Pittsburgh Compound B turned out to be a very effective way of seeing amyloid. If we used 6-cyano-PIB in our histology, it stains plaque and nothing else. If you build up levels far beyond what you’d see in humans or in PET scan you can see some other amyloid stain but in the levels possible in the brain, you see nothing but plaque (no amyloid in white matter or the cerebellum). We’re about to publish a paper on the first case who died – a large correlative study on a woman that had a PIB scan 10 months before her death with dissections of the high vs. low PIB retention areas of her brain and these reveal a strong correlation with amyloid. We believe this represents a safe and painless way to determine if someone has amyloid in their brain or not in normal individuals, in those with cognitive changes, and those with varying stages of Alzheimer’s. Although many people think PIB is used for diagnosing, it’s a powerful biomarker originally developed to test anti-amyloid compounds currently in development. In current studies we are looking to see if PIB will show loss of amyloid in the brain following administration of medications.

Q:What other biomarkers are useful in Alzheimer’s detection?

A:Well, it depends on what you’re looking for. The most obvious marker is beta amyloid which also exists in the blood. Levels of 1-42 - which is probably the more pathological form of beta amyloid - are very low (you can measure it but it’s difficult to do). Levels of 1-40 are higher but none look useful in individual patients. It probably elevates at some point before someone develops manifest disease, but the variability is so high (and also with individuals varying one from another) it would be hard to see how that would ever be a test that would say you have an elevated level, therefore we should treat you in a preventive fashion. There are a couple of studies looking at blood proteomics to see if there is a pattern of protein changes that may be associated with incipient onset of Alzheimer’s. Typically you look at normal elderly, those with mild cognitive impairment and those with Alzheimer’s to see what their patterns look like and compare with the younger population with indication of risk.
Biopsy would be great but unfortunately amyloid doesn’t build up anywhere else in the body. The most sensitive marker we have now is loss of volume or shrinkage of the hippocampus – the human “RAM chip” needed for short term memory (also shrinkage of the cortex). As a neurologist it offends me that the marker should be shrinkage – seems like that’s locking the door after the horse is gone. I’d rather like to find something before observable structural changes begin to occur but for now this is what we have. I predict that in 5 years an MRI will be able to give an immediate volume determination of the hippocampus – it will say that if this patient has cognitive problems there is an 80% chance they have Alzheimer’s disease. If they do not, there is.a 60% chance that they will. Right now there is a multi-center study of neuroimaging and peripheral biomarkers occurring over 3 years of people with Alzheimer’s (normal’s vs. MCI’s) which will give us the ‘norms’ after correcting for head size, e.g. is your hippocampus the right size for your age and educational level? If the values are smaller than expected, then you are likely at risk. This type of biomarker might be useful in people who are asymptomatic.

Q:What is Ginkgo biloba? Does it really delay onset of dementia in “normal” elderly adults?

A:Darwin called Ginkgo biloba a living fossil - it’s millions of years old and essentially is still in its same form as eons ago. The Chinese have used it for at least two millennia for a variety of illnesses. It’s important to note that a special extraction process is used to extract the powerful antioxidants (from the leaves only) as the Ginkgo fruit does not have a pleasant odor and Ginkolic acid upsets the stomach and can actually kill neurons and basically any cell it is exposed to in culture. We did a study a few years ago and did not find a difference in disease progression among 250 Alzheimer’s patients given Ginkgo vs. 250 Alzheimer’s patients given placebo. Antioxidants given at the time of full-blown disease might not be effective. In fact, no known antioxidant has been shown to be effective in slowing progression when you already have the disease, including Vitamin E. We’ll see about red wine…The question is, if you give Ginkgo early, would it stop the disease?

Q:It has been rumored for years that an Alzheimer’s vaccine is on the horizon. What compounds make this possible? How soon do you foresee human trials beginning? Any other new developments in the pathology of Alzheimer’s? What are the most promising new therapies? Are there any current clinical trials?

A:We all call it a vaccine, but it’s really not – it is not part of a live biological like a virus, it is an immunization of protein. The first human trial has already been done. After success with mice, beta amyloid was given with an adjuvant in humans. It does work – it does remove beta amyloid plaque from the brain, but in about 6% of people it also produces meningoencephalitis, thus the trial was stopped and so we do not know how successful the vaccine was overall - no one was followed to ascertain who still developed Alzheimer’s, etc. The alternative to urging the body to make antibodies is to just immunize with the antibody and there are a few trials in place right now to test this. Immunotherapy, e.g. IgG infusion, although not a mainstream therapy would be another alternative. There are a few new drugs – enzyme inhibitors, anti-RAGE receptors.

Q:Which animal model do you believe is the best representative to study neurodegenerative disease like Alzheimer’s?

A:It depends. For Alzheimer’s there will probably never be a perfect model – mice do not develop neurofibrillary tangles, only amyloid. The best model we could have is a primate, but it remains to be seen whether we’d do that. Mice have tau but it’s only 50-60% homologous to human tau (probably the reason they don’t develop tangles). We’ve been working with primate stem cells to insert mutations but the cells are very unstable. Would we deliberately make an animal that would become demented? That brings up other issues about use of primates and animals as models. However, cell cultures may turn out to be powerful predictive models of whether medicines would work in humans.