Challenging Views of Alzheimer's Disease

Posted 2 August 2001

Cincinnati, Ohio,
July 27-29, 2001
Summarized by Keith Crutcher and Steve Robinson
Conference URL:
Related News: See BioMedNet news report (subscribers only)

This meeting was born of a disenchantment with the regular conference format that usually consists of numerous data-oriented talks with limited or no opportunity for critical discussion. Challenging Views explored a new and radically different format: a series of debates in which teams defended a viewpoint while exposing weaknesses in the opposing team's position. After rebuttals and counter-attacks, the audience had a generous 45 minutes to grill both teams.

This bold approach produced a depth of engagement and quality of discussion rarely before seen at conferences on Alzheimer's disease. The debates were passionate, riveting and sometimes hilarious as participants struggled to defend the indefensible. All teams conceded some ground, and none of the debates produced a clear winner. The real winners were the participants, as it became evident that opposing viewpoints could often be resolved by new experiments or novel interpretations. Those who attended the conference commented on how extraordinarily constructive the debates were, and how these interactions had sparked ideas and forged research collaborations.

By the close of the conference participants were united on one point: that Challenging Views must become a regular event. It was decided that Challenging Views of Alzheimer's disease will become a biannual event, with the next meeting to be held in Cincinnati in July, 2003.

Summaries of the debates are provided below:
Debate 1: Does amyloid or its precursor cause familial AD?
Debate 2: ApoE4: is it the absence of good or the presence of bad?
Debate 3: Is non-familial AD inherited?
Debate 4: Fire! Are oxidation and inflammation the culprits in AD?
Debate 5: Is AD a vascular or a metabolic disorder?
Debate 6: Were the tauists right all along?
Debate 7: Cyclin towards, or away from, dementia?

Debate 1: Does amyloid or its precursor cause familial AD?

Moderator: Steve Snyder

"Beta amyloid deposition is essential to AD neuropathology"
Steve Estus, Bob Vassar, Mark Kindy, and David Borchelt
"Dysfunction of the amyloid precursor protein is essential to AD neuropathology"
Rachael Neve, Craig Atwood, and Stephen Robinson

Moderator Steven Snyder tossed the coin (a quarter graciously donated by the NIH). The APP team won, and Steve Robinson was sent in to bat. He began with a blistering critique of the amyloid hypothesis, arguing that the pathological definition of AD presupposed a central involvement of amyloid plaques, and therefore the amyloid hypothesis was a tautology. He pointed out that there are massive discrepancies between the clinical and pathological diagnoses of AD, with 10-20% of patients clinically diagnosed with probable AD failing to meet the pathological criteria for AD. Conversely 20-50% of cognitively normal elderly patients meet the pathological criteria for AD. Finally he asserted that transgenic mouse models expressing familial AD mutations fail to develop significant neurodegeneration or other features of AD, depite an abundance of beta amyloid.

Craig Atwood argued that the toxicity ascribed to amyloid is due to transition metal ions that bind to the peptide. He noted surprising new evidence indicating that beta amyloid has anti-oxidant properties. Team captain Rachael Neve proposed that the emphasis on beta amyloid has been misplaced. She reviewed evidence showing that APP serves numerous roles during brain development and in regeneration, and concluded that any impairment in the function of this protein is bound to have deleterious effects. Mutations in familial AD directly involve the APP gene, or the presenilin-1 gene, and therefore affect either the structure of APP or its cleavage. Cleavage products such as C100 appear to be much more toxic than beta amyloid.

Opening for the amyloid team, Steve Estus countered that beta amyloid deposits are always present in brains affected by AD, and the distribution of plaques closely matches other indices of neurodegeneration. Furthermore, amyloid has been shown to be neurotoxic in vitro, and autosomal dominant AD involves mutations that appear to result in the increased production of A-beta. The team argued that cognitive impairments and synaptic loss are evident in some transgenic mouse models of AD, and that these impairments can be reversed by immunising the mice with A-beta. The presence of plaques in normal individuals can be explained by brain reserve. The team concluded that on the balance of evidence, Occam's razor demands that we accept a central role for A-beta in AD.

The APP team countered that William of Occam predated the scientific revolution, and that Descartes had warned against concluding that correlation proves causation. They pointed out that the reversibility of cognitive loss produced by A-beta immunisation proved that A-beta plaques do not cause permanent toxicity. They questioned whether the idea of 'brain reserve' could be formulated in a way that allowed it to be tested scientifically; if it couldn't then it was merely an 'escape clause'.

The amyloid team responded by reminding us that mice are not humans, so we can't expect a perfect correlation with AD. At that point the audience had their chance and James Mortimer came to the defense of the brain reserve hypothesis, claiming that head/brain size is the key biological factor contributing to the reserve. After further discussion it was agreed that more research needs to be done to explain how brain size confers protection from AD at the cellular and molecular levels. David Morgan defended the transgenic mice by saying that their real value lay in helping us to understanding the processes of amyloid deposition and clearance. The consensus reached by the end of the debate was that amyloid is involved somehow in AD but the nature of that involvement is not yet understood. Whether it is the cause or a consequence of AD, will be answered definitively by the outcome of clinical trials using A-beta immunization.

Debate 2: ApoE4: is it the absence of good or the presence of bad?

Moderator: Wes Ashford

"The absence of apoE3 or apoE2 contributes to AD pathology"
Mary Jo LaDu, William Rebeck, and Mark Kindy
"The presence of apoE4 contributes to AD pathology"
Keith Crutcher, Britto Nathan, Jacob Raber, and Bruce Teter

That apoE is playing some role in AD is uncontested. What that role is, however, is still controversial. Supporters of the view that the absence of apoE3 is most relevant were pitted against those arguing that the presence of apoE4 is actually most important. The pro-apoE3 team clearly pleased the audience with their visual aids. Sporting customized angel-white caps emblazoned with "APOE3" they provided "APOE4" black caps to the opposing "dark side of the force" so the audience could clearly tell which was the winning side. Bill Rebeck also provided critical audiovisual support with his giant Reds foam finger that he used to emphasize critical points being made by team captain MaryJo LaDu.

The major arguments presented by the apoE4 team (which opened the debate after winning the coin toss) were that E4 exhibits a variety of "bad" effects in several systems ranging from cultured cells to transgenic animals. These include greater inhibition of neurite outgrowth, neuronal toxicity, more profound behavioral deficits, and greater risk of diseases other than AD. The apoE3 team was not willing to concede any of these points other than the fact that apoE does appear to exhibit an isoform-specific neurotoxicity under certain culture conditions. But where it counts most, which is in its interaction with amyloid, the white-capped team strongly upheld the view that E3 is a "good guy". This might take the form of increasing clearance of amyloid or otherwise neutralizing the negative effects of amyloid.

The Darth Vader side sought to bolster their position by citing studies indicating that polymorphisms in the apoE promoter region, which are thought to result in increased expression of apoE, appear to increase the risk of AD regardless of isoform. However, Bill Rebeck, who has been involved in a number of genetic studies, pointed out that there really wasn't sufficient evidence available to draw such a conclusion.

Wes Ashford did an admirable job of moderating the session and maintaining relative order. Part of the discussion period was devoted to questions regarding the metabolism of apoE and the extent to which specific intracellular pathways might be implicated in apoE function and re-cycling. Although neither team appeared to be persuaded to switch allegiance, there was agreement that a number of critical questions remain to be answered before the contribution of apoE to AD pathology can be identified.

Debate 3: Is non-familial AD inherited?

Moderator: Bob Petersen

"Non-familial AD is mainly due to genetic factors"
James Mortimer and Wes Ashcroft
"Non-familial AD is mainly due to environmental factors"
Arezoo Campbell, William Grant, Ruth Itzhaki, and John Savory

As for some of the other debates, this one brought together two teams who would agree that the opposition's view could be at least partly supported by available data. However, the contrast between these teams, in terms of whether genetic or environmental factors were most important in contributing risk to AD, was clearly evident. The environmental team cited a number of lines of evidence in support of their view that genes are not destiny. Thess include aluminum exposure, viral infections and diet. Although not refuting the role that genetics might play in conferring susceptibility, this team cited numerous instances where exposure to environmental factors enhanced AD risk.

The genetics team didn't concede any ground. Two lines of evidence were marshalled to support the primary role of genetic factors. The first is that the majority of the risk for AD appears to be established before the age of 20 (the "Nun study"). The second is that 50% of AD patients carry the E4 gene. Some of the resulting exchanges and discussion from the floor made it clear that separating the role of environment and genes is virtually impossible. One point made in rebutting the genetics argument was that even if the majority of risk is established early on, this does not support an exclusively genetic contribution. The dietary pattern, for example, is usually established during this period. One of the twin studies cited in support of the genetic hypothesis, for example, would not have controlled for similarities in diet.

An additional twist on this idea was provided from the floor by Sue Griffin, who wondered to what extent the behavioral patterns that contribute to increased risk (such as eating high-fat foods) might not, in themselves, ultimately be genetic risk factors. If so, one might hold the view that it is all in the genes. In the end, both teams seemed willing to admit that it is the interaction between genes and environment that determine risk. And the genetics team seemed to hedge its bets by noting that it can't hurt to watch one's diet and exercise.

Debate 4: Fire! Are oxidation and inflammation the culprits in AD?

Moderator: Hossein Ghanbari

"Oxidative stress and inflammation are essential to AD pathogenesis"
Allan Butterfield, Sue Griffin, Gerald Munch, and Giulio Maria Pasinetti
"Oxidative stress and inflammation are secondary phenomena to AD pathogenesis"
Craig Atwood, Stephen Robinson, and Mark A. Smith

There was agreement between both teams that inflammation and oxidation are pathogenic features of AD, and that the presence of amyloid plaques provokes both an inflammatory response and oxidative injury. The debate centered on the issues of whether amyloid deposition was the primary event or whether inflammation and oxidation in AD could also be caused by other factors.

Moderator Hossein Ghanbari's coin toss sent Allan Butterfield to begin the debate. His support for a central role of amyloid included in vitro evidence that amyloid provokes a neurotoxic response in microglia, and biochemical evidence that beta amyloid peptide is neurotoxic and can produce reactive oxygen species. In particular, he noted that amyloid peptide inactivates the astrocytic enzyme glutamine synthetase. Gerald Munch noted that the brains of APP transgenic mice display lipid peroxidation and elevations in inflammatory cytokines. He also suggested that the AD brain is like a slowly-baking fruit cake and that sugar moieties crosslink with amyloid peptide (and other proteins) to form AGE's that are the real source of toxicity and inflammation. Sue Griffin noted that increased oxidation and inflammation were also features of Down syndrome (which shares many similarities with AD), and presented a model to show that most of the neuropathological features of AD, including amyloid deposition, could be explained in terms of a cascade that is initiated by interleukin-1 expression.

Craig Atwood led the counter-attack and asserted that the toxicity of amyloid was clearly a consequence of transition metals bound to the peptide and not to an intrinsic property of the peptide. This point was supported by evidence that metal chelators completely block the toxicity of amyloid in biochemical assays and in cell culture. Amyloid might even be viewed as a natural anti-oxidant. A jet-lagged Steve Robinson speculated whether a disturbed melatonin rhythm might explain the cognitive features of AD. Then he presented evidence from his research team showing that amyloid peptide protects against metal-induced neurotoxicity in vivo. He also showed that the cellular expression of glutamine synthetase is reduced in AD brain, but in a pattern that does not correspond to the distribution of plaques, so it cannot be due to amyloid toxicity. Mark Smith noted that Sue Griffin had actually supported the opposing team's point of view and invited her to defect to the other side. He used the analogy of firefighters putting out a fire to suggest that while oxidation and inflammation may cause injury, that injury may be in a good cause. He also pointed out that inflammation and oxidation in APP transgenics precede plaque deposition.

Much of the question time concerned aspects of methological detail, and whether the conclusions reached by both teams were fully supported by the available evidence. The consensus at the end of the debate was that cause and effect in AD is almost impossible to unravel when we are limited to tissue from end stages of AD. Techniques are needed that will allow the assessment of amyloid deposition, inflammation and oxidation in the living brain.

Debate 5: Is AD a vascular or a metabolic disorder?

Moderator: Tony Phelps

"AD is primarily due to vascular pathology"
Paula Grammas, Masahito Yamada, and Berislav Zlokovic
"AD is primarily a disorder of brain metabolism"
Jim Blass, Gary Gibson, and Siegfried Hoyer

Paula Grammas began the debate by noting that AD is accompanied by extensive deposits of vascular amyloid, and that many amyloid plaques in the brain are closely associated with capillaries. She showed evidence that intravascular amyloid deposits cause injury to the underlying capillary endothelial cells. She noted that abnormalities of blood-brain-barrier (BBB) permeability are a common feature of AD brains, and this might be explained by the actions of vascular amyloid deposits. In particular, the serum protein thrombin is present in AD brain, and this agent has been demonstrated to be highly neurotoxic. In addition, thrombin appears to enhance the toxicity associated with apoE4. Berislav continued this line by reviewing evidence from aged primates showing that amyloid peptide injected into the bloodstream crosses the BBB and binds to pre-existing amyloid deposits in the neuropil. He highlighted other abnormalities of BBB function, including the reduced export of glutamate from the AD brain. Masahito Yamada showed that perivascular amyloid deposits in AD brains are invariably associated with tau pathology.

Siegfried Hoyer noted that the earliest feature of AD is a reduced rate of oxidative metabolism. This reduction begins prior to the appearance of cognitive deficits and decreases further as the disease progresses. He showed that there were abnormalities in brain glucose transport, as well as a reduced contribution of insulin and insulin receptors to cerebral metabolism. Effectively, AD is a diabetes of the brain and the pathological features of the disease result from this condition. Gary Gibson pointed out that AD involves profound disorders of mitochondrial metabolism that cannot easily be attributed to beta amyloid or other features of the disease. John Blass commented that experimental induction of cerebrometabolic deficiency in humans reproduces many of the cognitive features of AD. Furthermore, cognitive impairment in AD patients can be ameliorated by elevating blood glucose levels or increasing cerebral blood flow.

In the rebuttals both teams made it clear that they thought the opposing team had a strong argument and that the two viewpoints were compatible. The audience was impressed by the wealth of evidence suggesting that cerebrometabolic abnormalities may be one of the initial features of AD. The theme running through most of the questions related to what might possibly cause these abnormalities, with broad agreement that this an area worthy of grant support.

Debate 6: Were the tauists right all along?

Moderator: Dave Morgan

"Neurofibrillary pathology is central to AD"
Skip Binder and Khalid Iqbal
"Synaptic dysfunction precedes cytoskeletal pathology"
David Borchelt, Bill Honer, Paul Coleman, and Hugo Geerts

Previous debates in the AD field have often pitted the ├čaptists (those who believe that amyloid is playing the most critical role) against the tauists (those who believe that neurofibrillary tangles, composed of tau, are more critical). In this round, the slant was a bit different. Although amyloid made an appearance, the debate was more focused on whether neuronal dysfunction in AD is initiated by disruption of the cytoskeleton or loss of synaptic function. There were able defenders on both sides.

The tauists provided summaries of various lines of evidence pointing to the importance of tau. Featured prominently were the findings of tau mutations in frontotemporal dementia. The neurofibrillary pathology in this disease is not accompanied by amyloid deposition. The argument was made that abnormal phosphorylation of tau results in loss of the normal function of this protein as well as a negative gain-of-function insofar as the phosphorylated tau tends to aggregate much more readily. But the team admitted that tau mutations do not apparently cause AD. However, if one examines the AD brain, one finds truncated forms of tau and modifications of tau that occur very early in the disease. The point being that these changes in tau would be expected to disrupt neuronal function long before neurofibrillary tangles are detected.

The synaptic team did not dispute that tau pathology is present, but they did call into question whether tangles were really much more than end stage indications of some other cellular dysfunction. For example, GSK3├č transgenic mice show brain shrinkage and functional impairment. But the shrinkage is apparently due to neuronal atrophy, not neuronal loss, and there is no evidence for tangles. In addition, there is evidence for declines in several synaptic markers in APP23 mice but also no evidence for tangles. Even in the human brain, the tangle-bearing neurons appear to survive for a very long time and the declines in expression of certain proteins, such as synaptophysin, appear to precede tangle formation. Therefore, it would seem that other alterations in synaptic dysfunction are separate from, and perhaps precede, tau pathology. This team characterized itself as taking more of a multifactorial approach rather than focusing on a single molecule.

Perhaps one of the most salient questions raised in the debate was at what point one could know whether tau alterations were "significant" in terms of functional effects on the neuron. There was agreement from both sides that it was not possible with existing methods to answer this question definitively. There was little evidence that either side was going to return to their lab and abandon their current focus. However, there also seemed to be greater clarification of exactly what questions remained to be answered and which ones were most important for resolving the issue of cytoskeletal versus synaptic pathology in AD.

Debate 7: Cyclin towards, or away from, dementia?
Moderator: John Blass

"Re-expression of cell cycle proteins induces neuronal cell death during AD"
Karl Herrup and Thomas Arendt
"Re-expression of cell cycle proteins is a response to neuronal injury in AD"
Mark Smith

The other debates in this conference can be thought of as dealing with relatively mature hypotheses. And one might be forgiven for thinking that the source of ideas for what might be happening in AD has been exhausted because of the extensive exposure given to one or two "leading" hypotheses. But it is clear that the source has not dried up (a la Manon des Sources?). And the final debate provided focus to a relatively new concept involving the role of cell cycle proteins in this disease.

Again, there was considerable agreement, at least insofar as both sides presented data showing that cell cycle proteins are expressed in neurons and glia in several areas of AD brains where neuropathology is abundant. This includes immunohistochemical data showing expression of PCNA, cyclin B, cyclin D and cdk4. There is also expression of endogenous cell cycle inhibitors such as p16. And there is clear evidence for neuronal polyploidy. So far so good. Everyone seems happy. Cell cycle proteins are upregulated in AD brain.

But at this point the debate lived up to its billing. The two sides were divided in how the results should be interpreted. Team one defended the view that neurons in AD are being driven back into the cell cycle, a state they should have left once they became fully differentiated. The consequence of this, they contended, is that neurons attempt to divide again and die in the process. The argument derives from what has been established in other systems where it has been shown that other cell types can be induced to die by being forced into the cell cycle. Neurons are trying to revert to their youth, perhaps in response to injury, but die in the process.

The opposition said, "Hold on". There are a number of cell cycle proteins expressed in neurons in the AD brain. True enough. But there is no evidence that these cells are being driven back into the cell cycle. In fact, many markers from different stages of the cell cycle appear to be expressed in the same neurons, which should not be the case if the neurons are truly "cycling". So maybe the expression of cell cycle proteins is actually an attempt at survival since many of the proteins are localized to the cytoplasm rather than the nucleus (where they should be if they are involved in division). As one of the debaters (Smith) put it, "The opposition is finding 5 wheels and a pedal here, a handlebar there and then trying to build a bicycle and go riding off into the sunset".

The ensuing discussion ranged from potential contributions of GRH to the phenomenon of cell cycle expression to the uniqueness of neuronal polyploidy, evidence of the wellspring of ideas that emerged during the debate. Although no "Eureka!" moment occurred, the attendees left with a healthier respect for the idea that cell cycle expression may be highly relevant to AD.

Wrap-up: Promising new directions for drugs and diagnostics?

The final session did not serve as a debate but, rather, as an opportunity for reflection. All seven moderators provided their own summaries of the debates they presided over as well as thoughtful comments on where the future might be taking us in regards to diagnostics and therapeutics. The importance of this perspective was provided by Chris Kircher, a neurologist who graciously filled in for Stephen Post, who unfortunately was not able to attend at the last minute. Dr. Kircher reviewed some patient histories, including a dramatic slide session of the decline of the first patient he cared for with AD. This sobering reminder helped focus the discussion on how best to move forward in developing new treatments.

John Blass was effusive in his praise of the meeting format, noting that he "hadn't enjoyed an Alzheimer's conference this much since the 1980's." He also noted that the session (cell cycle) he moderated was a "model session" in that everyone agreed on the data but that the interpretation is still open. Wes Ashford made a compelling case for the efficacy of available therapies in postponing admission into chronic care facilities, leading to tremendous economic benefit. Tony Phelps pointed out that significant improvements in imaging are being made. Dave Morgan said he found it interesting to be in the minority (as an avid supporter of the amyloid hypothesis) and noted that one limitation of the debate format was that it didn't allow for the conclusion that the disease encompasses all of the hypotheses being considered. This sentiment was echoed by Petersen who wondered if it was possible that in some sense "everyone is right"? All of the moderators agreed that the meeting had provided an invaluable opportunity to explore and refine ideas about the causes of Alzheimer's disease.

This meeting was co-sponsored by the Alzheimer Research Forum



Make a Comment

To make a comment you must login or register.

Comments on this content

No Available Comments


Other Citations

  1. Debate 1

External Citations

  2. See BioMedNet news report (subscribers only)

Further Reading

No Available Further Reading