NEWS
The amyloid code
Is there a conspiracy afoot to keep one theory dominant in Alzheimer
disease research? Maybe not, but as Apoorva Mandavilli discovers, it may
still be high time for new ideas.
In the Church of the Holy Amyloid, the reigning
© 2006 Nature Publishing Group http://www.nature.com/naturemedicine
deity is a 42–amino acid protein, thought to be
the key to the mysteries that underlie Alzheimer
disease. Amyloid has many believers—and like
any good religion, a few staunch heretics who
question its supremacy.
As flippant as this might sound, this is how
some see the dominance of the amyloid-beta
protein in Alzheimer disease research.
For more than a decade, many scientists
have believed that the protein triggers a series
of steps—dubbed the amyloid cascade—that
leads to the memory loss and dementia in those
afflicted with Alzheimer disease.
Every aspect of the field reflects this theory’s
dominance: therapies being tested in the clinic,
animal models of the disease, most funded
grants and the most-cited papers, including
those selected for this special issue (see page
767), are all based on the premise that amyloid-
beta is the culprit in Alzheimer disease.
But some researchers in the community—
and a few outside it—say that those who believe
the amyloid cascade hypothesis have unwisely
pushed it, perhaps to the detriment of the
whole field.
“I think it’s become a little bit of a religion,”
says Mark Smith, professor of pathology at
Case Western Reserve University. “We have
very charismatic religious leaders leading the
way and there are many followers.”
Smith is one of the most vocal opponents
of the amyloid cascade hypothesis and coined
the phrases “Church of the Holy Amyloid,”
“amyloid code” and others to invoke what he
says is a virtual dictatorship of the hypothesis.
But even among those who are more
measured in their criticisms, the phrase “all of
your eggs in one basket” comes up with startling
frequency. These scientists say that despite many
unexplained aspects, there is a disproportionate
amount of attention paid to the amyloid
hypothesis that has prevented other ideas from
flourishing. With such a complex disease, they
warn, it is foolhardy, perhaps even dangerous,
to focus exclusively on one theory.
“I have mixed feelings about saying this
because I also feel that the evidence for the
amyloid hypothesis is overwhelming,” says
Tom Sudhof, director of the Center for Basic
Neuroscience at the University of Texas
Southwestern Medical Center. “The whole
NATURE MEDICINE VOLUME 12 | NUMBER 7 | JULY 2006
field is governed by an old boys’ network that
is not very positive. It needs new blood, new
movement, new ideas.”
Feeble alternatives
Alois Alzheimer first described plaques and
tangles that characterize the diseased brain
nearly 100 years ago. The dense tangles are
a feature in many different dementias, but
amyloid plaques in the brain are unique to
Alzheimer disease.
The strongest evidence for amyloid being
a cause—and not just a consequence—of the
disease comes from genetics. In familial or early-
onset forms of Alzheimer disease, mutations in
the amyloid precursor protein, and the enzymes
that clip it to form amyloid-beta, lead to a
massive overproduction of amyloid-beta and
a swift descent into disease.
It would be much easier to argue against this
theory if there were a worthy alternative to take
its place.
There are certainly several candidates. Smith
believes, for example, that amyloid-beta is
beneficial, and is produced as a compensatory
response to the disease. Oxidative stress,
inflammation, long-term response to injury
or infection and defects in normal brain
maintenance—such as the clearance of defective
proteins—are among other alternatives on offer.
Because Alzheimer disease is so prevalent,
the pathway must be one that can be easily
Normal neuron
Amyloid-beta plaques
Tau tangles
All tangled up: Nerve cells in an Alzheimer brain
are filled with dense fibrils inside and amyloid
plaques outside.
perturbed in an aging brain. But none of those
theories offers a detailed mechanism either.
“I don’t like things where people mutter vague
criticisms,” says John Hardy, laboratory chief of
neurogenetics at the US National Institute on
Aging. Hardy and Harvard University’s Dennis
Selkoe are widely credited with laying out the
amyloid cascade hypothesis. “I argue my case as
effectively as I can,” Hardy says. “This is a criticism
of other people not arguing theirs as well.”
The amyloid cascade hypothesis, despite its
monolithic ring, has undergone some revision.
Initially, scientists believed that the plaques
themselves were causing the disease. Based
on that idea, they proposed that removing the
plaques might retard or even reverse the memory
loss and cognitive decline seen in Alzheimer
disease. These days, the emphasis is more on
oligomers, or smaller aggregates of amyloid-
beta.
In March, University of Minnesota researcher
Karen Ashe and her colleagues isolated A-beta*,
a 56-kilodalton aggregate of amyloid-beta, from
an amyloid mouse model and showed that
injecting the aggregate into healthy rat brains
triggers memory problems (see page 760).
There are specific criticisms raised about that
report—that Ashe used an unusual memory
test, for example, or that the paper rehashes
results published in 1985 by Australian
researcher Colin Masters—but it has broad
support among amyloid proponents.
Unsure beginnings
Ashe’s finding, like most others in the field,
relies on a model that grossly overproduces
747
 NEWS

amyloid, based on mutations seen in the
familial disease.
Familial Alzheimer disease accounts for only
about five percent of the more than 4 million
individuals with Alzheimer disease in the US
alone. What sets off the disease in the vast
majority of cases?
“I think the idea of what gets it all started is
kind of murky,” says Ashe. “Nobody has fantastic
theories about how sporadic Alzheimer disease
© 2006 Nature Publishing Group http://www.nature.com/naturemedicine
its own pancreatic cells. Type 2, or adult-onset
diabetes, is much more common, occurs much
later in life and is characterized by insulin
resistance.
Both diseases eventually result in abnormal
processing of blood sugar and can have the same
dire consequences, but the initial trigger and
progression of the disease are sharply different.
It’s possible that familial and sporadic
Alzheimer disease are similarly distinct,
4 million — Number of people
in the US with Alzheimer disease.
14 million — Projected number
for 2050 if no preventive
treatments become available.
$100 billion — Annual cost

begins.” suggests Kaj Blennow, professor of clinical

Amyloid proponents are betting that even
if the initiating event is different, the disease
mechanisms are bound to be similar in both
the familial and sporadic forms of the disease.
But what if they’re not?
That kind of model is not entirely without
precedent. For example, the name ‘diabetes’ is
used to refer to two very distinct conditions:
type 1, or juvenile diabetes, as its name suggests,
strikes children and young adults, and is an
autoimmune disorder in which the body attacks
neurochemistry at the Sahlgrenska Academy of
Göteborg University.
If that’s the case, scientists may have
misguidedly spent too much time studying
models that merely replicate the familiar, and
rarer, form of the disease (see “Of mice and men,”
page 750), Blennow says, and neglected research
on mechanisms that might explain the sporadic
form.
“My hope is of course that it’s correct and that
we’ll have treatments soon,” Blennow says. “But if
this proves not to be the case, then this hypothesis
has been too dominating and has hindered other
hypotheses from coming up.”
Open war
There are controversies in every scientific arena,
but the Alzheimer field is a rare instance in which
scientists openly accuse each other of bias.
If you don’t endorse the amyloid hypothesis
and actively work on some aspect related to
it, some say, it’s difficult to win funds, publish
papers or present at conferences. To have any
credibility, people have to prove that their idea
has some merit. But without funds and some
element of receptiveness, that’s hard to do, notes
Larry Goldstein, professor of molecular medicine
at the University of California in San Diego (see
page 750).
of caring for people with the
disease in the US.
“There’s a certain catch-22, I think,” Goldstein
says. “I’m not trying to insult these folks, but it’s
quite a mafia if you really look at the field.”
To be fair, Goldstein and others also say that
it’s difficult to know whether some scientists’
ideas are being ignored because they’re on the
periphery, because they’re bad—or simply as a
matter of course. Journal editors and funding
agencies deny any bias on their part and say they
simply don’t see many papers or grants with new
ideas. What’s more, they add, no amount of bias
could keep a genuinely good idea down for very
long.
“Of course the truth is that scientists are
inevitably paranoid,” says Hardy. “You get your
grant turned down and you think it’s those
amyloid bastards who did you in. But my grants
have been turned down with monotonous
regularity too.”
Some favoritism is also true in any field. People
who organize conferences tend to invite others
with similar ideas. Scientists who review grants
are naturally more interested in ideas that mesh
with their own.
“The enemy is us,” Goldstein says. “In a peer-
review system, the only way to change what

Drug Manufacturer Mechanism of action Launched US sales (2005) Side effects

Donepezil Eisai Inc. Cholinesterase inhibitor; prevents the 1997 $1.1 billion Nausea, diarrhea, insomnia,
(Aricept) breakdown of acetylcholine in the brain vomiting, muscle cramps,
fatigue, anorexia
Memantine Forest NMDA receptor antagonist; blocks toxic effects 2003 $498 million Dizziness, headache, confusion,
(Namenda) Pharmaceuticals associated with excess glutamate and regulates constipation
glutamate activation
Rivastigmine Novartis Cholinesterase inhibitor; prevents the 2000 $226 million Nausea, vomiting, loss of
(Exelon) Pharmaceuticals breakdown of acetylcholine and butyrylcholine appetite, indigestion, weakness/
Corporation in the brain lack of energy, dizziness, diarrhea,
headache, stomach pain
Galantamine Ortho-McNeil Cholinesterase inhibitor; prevents the 2001 $223 million Nausea, vomiting, diarrhea,
(Razadyne) Neurologics Inc. breakdown of acetylcholine and stimulates anorexia, weight loss
nicotinic receptors to release more acetylcholine
Galantamine Ortho-McNeil Cholinesterase inhibitor; prevents the 2005 $24 million Nausea, vomiting, diarrhea,
(Razadyne ER) Neurologics Inc. breakdown of acetylcholine and stimulates anorexia, weight loss
nicotinic receptors to release more acetylcholine
Source: IMS Health

Weak weapons: The drugs available for Alzheimer disease have shown only modest benefits.

748 VOLUME 12 | NUMBER 7 | JULY 2006 NATURE MEDICINE

 NEWS

Drug Manufacturer Mechanism of action Stage of Side effects

development
New drugs 3APS Neurochem, Inc. inhibits amyloid-beta aggregates, binds and reduces Phase 3 Nausea, vomiting
(Alzhemed) soluble amyloid-beta
MPC-7869 Myriad NSAID derivative; inhibits amyloid-beta aggregates Phase 3 None disclosed
(Flurizan) Pharmaceuticals and reduces their levels of amyloid-beta with little or
no anti-inflammatory effect
AAB-001 Elan monoclonal antibody binds to and clears amyloid- Phase 2 None disclosed
Pharmaceuticals beta, is designed to directly deliver antibodies to
amyloid-beta
© 2006 Nature Publishing Group http://www.nature.com/naturemedicine

Neramexane Forest NMDA receptor antagonist: blocks the effects of Phase 3 None disclosed
Laboratories excessive glutamate at the receptor
Drugs Simvastatin Merck Statin; reduces cholesterol-carrying protein that Phase 3 None disclosed for the trial, but
for other (Zocor) promotes amyloid-beta aggregation Zocor has been known to cause
conditions nausea, diarrhea, abdominal
pain and muscle cramps
VP4896 Voyager Hormone drug leuprolide acetate; decreases amount Phase 3 None disclosed
Pharmaceutical of luteinizing hormone in body, might prevent brain
cell death
Valproate Manufacturer not Anticonvulsant drug; neuroprotective properties may Phase 3 None disclosed
disclosed delay clinical progression of Alzheimer disease
Dietary Gingko Antioxidants neutralize free radicals and may reduce Phase 3 Headache, upset stomach,
supplements biloba or prevent the damage they cause in brain cells allergic reactions
Vitamin E Antioxidants neutralize free radicals and may reduce Phase 3 None disclosed
Selenium or prevent the damage they cause in brain cells

Target practice: Most candidates being tested for Alzheimer disease are based on the amyloid hypothesis.

people do is for reviewers of grants and papers
and conference organizers to broaden their
perspective and shift emphasis.”
But with conference presentations and
publications devoted primarily to the amyloid
hypothesis, even those entering the field are
exposed mainly to this view of the disease.
This result is particularly insidious, says Keith
Crutcher, a neuroscientist at the University of
Cincinnati. “Newer people coming into the field
get the impression that the case has already been
solved,” Crutcher says. “It has the subtle effect
of dissuading other people from entering the
field.”
In the past few years, the US National Institutes
of Health has begun initiatives to encourage
young researchers and more peripheral or risky
projects. For example, the R03 grant offers young
scientists up to $50,000 for two years to help
gather data for larger grants, and the R21 grant
offers $275,000 for two years.
Steve Snyder, program director for the etiology
of Alzheimer disease at the US National Institute
of Aging, says he has also on occasion made a
special effort to pick up grants that refute the
amyloid hypothesis.
Of the $75 million Snyder allocates in
grants, fully half goes to research related to
the amyloid hypothesis and the other half to
work on the remaining topics such as learning
and memory, tau, genetics, glial cells and
apolipoprotein E.
Although none of the researchers are really
biased, “there are people on the panels whose
own research interests mesh quite nicely with Lone target
amyloid-beta,” Snyder says. “When those people If amyloid gets so much attention from
are making decisions, that outlook can have a the scientists, it’s because there’s almost
bearing on what happens next. I don’t know how overwhelming evidence that it’s correct
to get around that.” fundamentally, says Harvard neurologist Selkoe.
An eye on...
If there is a holy grail of genetic mysteries, Alzheimer disease
certainly qualifies—with Richard Mayeux hot on its trail. Mayeux has
spent the past 18 years trying to tease out the complex genetic causes
of late-onset Alzheimer disease.
Mayeux, co-director of Columbia University’s Taub Institute for
Research on Alzheimer’s disease and the Aging Brain, appears
to be on the brink of nailing a culprit, nearly 13 years after the
discovery of APOE, the only gene known in its mutated form to
be associated with the common, late-onset form of the disease.
Mayeux has painstakingly built a database of clinical histories
and cell lines from about 500 Dominican families prone to the
late-onset form of the disease. He won’t reveal more about the
candidate he and his collaborators are pursuing because the
findings are unpublished.
The mere mention of a new Alzheimer gene is sure to
stoke the ongoing debate about genetic testing. Like many
scientists, Mayeux supports genetic tests for individuals
from families that carry mutations for the early-onset
disease, especially for those considering having children.
But for the late-onset disease, which comprises more than
98% of cases, Mayeux is vocally opposed to testing.
“If there were a treatment or a cure, then by all means it
would be important to have genetic testing,” says Mayeux.
“Until that happens, I really think it’s inappropriate.”
Meredith Wadman, Washington, D.C.

NATURE MEDICINE VOLUME 12 | NUMBER 7 | JULY 2006 749

 NEWS

“Amyloid-beta is necessary and it’s very early—

An eye on... and that’s what makes it a beautiful target.”
Indeed, most therapies in trials are aimed at
Larry Goldstein, famous for his plainspoken
reducing aggregates of the protein. Even if only
attitude, is one of the few big names in
one of them works, it would lend much-needed
the Alzheimer field who bucks the amyloid
clinical credibility to the amyloid camp.
hypothesis.
That resolution came tantalizingly close
Goldstein says the long-favored hypothesis,
which proposes that the amyloid-beta protein
once before. In late 2001, Elan Corporation
is the initiating factor in the disease, doesn’t and Wyeth-Ayerst Laboratories set out to test
an amyloid vaccine, AN-1792, in about 300
© 2006 Nature Publishing Group http://www.nature.com/naturemedicine

adequately explain many observations—such as

the poor correlation between the deposition of individuals. In animal models, the vaccine
amyloid plaques and cognitive problems. “It’s had effectively cleared plaques and improved
molecularly not explicit,” he says. “What is it memory. But in January 2002, the companies
that amyloid does to neurons? Does it activate had to halt a phase 2 human trial after 18
a receptor, poke a hole, activate kinases? There participants developed meningoencephalitis, a
are a lot of different suggestions, but they’re not potentially fatal inflammation of the brain.
precise.” Evidence from the trial showed that, as in the
Goldstein, a professor of cellular and mouse models, the vaccine successfully cleared
molecular medicine and a Howard Hughes amyloid plaques from human brains. Whether
investigator at the University of California, San it improved memory is more controversial,
Diego, favors an alternate model that pins the because the trial ended abruptly and there is
start of the downward spiral in Alzheimer disease to transport defects in the axons of nerve limited information from follow-up studies of
cells. The resulting ‘traffic jam’ of motor proteins, organelles and vesicles causes the axons to the participants. But at the very least, says Dale
swell, he says, and eventually leads to Alzheimer disease. Schenk, chief scientific officer of Elan, “what that
In a paper published last year, Goldstein and colleagues identified these defects in means is that at least one therapeutic approach
postmortem brains from individuals who showed early symptoms of the disease, but no extrapolates to patients, which is pretty darn
amyloid plaques (Science 307, 1282–1288; 2005). He also identified the defects in mouse great.”
models—more than a year before the mice developed classical disease pathology. The company has since discontinued work
Goldstein says one of the big problems in the field is that there are no good animal models on the vaccine, but is testing bapineuzumab, a
of the disease. He is instead turning to human embryonic stem cells that he has coaxed to
monoclonal antibody to amyloid-beta, to treat
differentiating into human neurons. “The problem is that humans are not just big mice,” he
the disease. Preliminary reports of the antibody’s
says. “We don’t know that the physiology and biochemistry of the human neuron is identical
effects have been positive, and results from
to that in animals.”
a phase 2 trial are expected by the end of this
Meredith Wadman, Washington D.C.
year.

For instance, only mice that carry two

Of mice and men
Every animal model for Alzheimer disease
is based on the same premise—that
overproduction of the amyloid-beta protein
somehow triggers a cascade that eventually
damages neurons and causes dementia.
Some of the mice have mutations seen
in familial or early-onset Alzheimer disease
and form the characteristic plaques seen
in diseased brains. Others carry mutations
in the protein tau and recreate the tangles,
another hallmark of the disease.
“In that capacity, the animal models have
been extremely successful and that’s no
trivial thing,” notes Karen Duff, associate
professor of neuroscience at New York
University.
or more mutations reproduce all the
physical features of an Alzheimer-addled
Blennow notes that in the animal models, at
least 47 molecules—a bizarre list ranging
from cholesterol drugs and painkillers to
brain. And most never exhibit the extent of
blueberries and curry spice—have been
neurodegeneration seen in people afflicted
shown to reduce the number of amyloid
with the disease.
plaques.
More important, the mice have never shed
“Hopefully, some of this is right, but I
much light on the mechanisms involved
think there’s a risk that there must be some
in the disease, such as the relationship
difference between the mice and patients
between amyloid-beta and tau, the two key
that makes it easier to improve things
molecules.
in mice,” he says. “As I see it, this calls
Because the
for caution when transferring data from
mice are based on
transgenic mice to man.”
mutations in familial
Apoorva Mandavilli, New York
Alzheimer disease,
it may be that they
only model the rare
form of the disease and
not the more common
sporadic form, suggests
Courtesy: Martin Ramsden

Research on the disease, for both

mechanisms and treatments, has been based Kaj Blennow, professor of
almost exclusively on these mice. But even clinical neurochemistry at
their staunchest supporters acknowledge the Sahlgrenska Academy of
that there are glaring flaws. Göteborg University.

750 VOLUME 12 | NUMBER 7 | JULY 2006 NATURE MEDICINE


~5% — Average lifetime risk
of developing Alzheimer disease
by age 65.
20–40% — Average lifetime
risk of developing Alzheimer
disease by age 85.
© 2006 Nature Publishing Group http://www.nature.com/naturemedicine
If drug companies have targeted amyloid-
beta, it has less to do with an overwhelming
belief in the hypothesis and more because it is
the most practical strategy, says Schenk. “Let’s be
very frank, as long as something is a hypothesis,
it’s still in testing,” he says. “It’s really in the eye
of the beholder.”
Schenk says the series of molecular steps that
lead to amyloid-beta, such as the enzymes that
cleave the precursor protein, yield clear targets for
drug development. At the same time, developing
therapies has been far from easy because the main
target is in the brain, one of the key enzymes has
a difficult structure and another is likely to have
wide-ranging effects in the body.
Companies are increasingly also looking
for therapies based on tau—which binds
and supports the microtubules that enable
molecular transport down neurons—such as
microtubule-stabilizing agents.
Tau tales
Not long ago, there was an acrimonious debate
between two groups of researchers who called
themselves BAPtists—or those who supported
the amyloid-beta protein—and tauists, who
believed that tau was the central molecule in
Alzheimer disease.
“I was vocal about tau because I thought the
field was putting too much emphasis on one
42–amino acid protein,” says John Trojanowski,
who along with his wife and fellow Alzheimer
disease researcher Virginia Lee directs the
Center for Neurodegenerative Research at the
University of Pennsylvania (see Profile, page
752). “We don’t know enough about the disease
to put all of our eggs in one basket.”
But the two groups have since settled their
differences, merging their ideas into a central
‘Unitarian’ premise. According to this theory,
amyloid-beta is involved at an early stage of the
disease and triggers a series of steps, at some
point involving the protein tau, that result in
the plaques, tangles and memory loss.
One of the most credible criticisms of the
amyloid hypothesis is that it has not yet been
able to explain exactly how this happens. Critics
raise various other points: for example, the most
widely used mouse model, Tg2576, overexpresses
the amyloid precursor protein—one of the
NATURE MEDICINE VOLUME 12 | NUMBER 7 | JULY 2006
NEWS
mutations seen in familial disease—but doesn’t
form tangles and more closely resembles age-
associated memory impairment than Alzheimer
disease. And some tau models, which express little
amyloid, develop the severe memory problems
associated with Alzheimer disease.
As a result of surprises like this, many sworn
amyloid supporters have begun to work on tau.
“I don’t think amyloid-beta is sufficient to cause
Alzheimer disease, that’s why I’m interested in
tau,” says Ashe, who developed the Tg2576 and
several dozen other mouse models.
Ashe and several others are trying to uncover
the molecular links between amyloid-beta and
tau that might explain how Alzheimer disease
develops. The many trials based on removing
amyloid aggregates might confirm—or refute—
the amyloid hypothesis long before then.
But in the meantime, a small group of
researchers continue to plead for greater
attention to alternative theories. With a disease
such as this, they say, the more ideas about how
it all began, the better.
“You don’t need a new religious leader,” says
Smith. “You need to tell all the followers to
disband and get thinking.”
Apoorva Mandavilli is senior news editor of
Nature Medicine.
An eye on...
The thing Mony de Leon finds most
disheartening about his job is watching
healthy study volunteers develop Alzheimer
disease.
A psychiatrist at New York University,
de Leon is developing imaging tools and
biomarkers to diagnose the debilitating
disease before symptoms develop. In studies
that can last 20 years, he often has to watch
once-healthy individuals develop diseased
brains.
In those afflicted with Alzheimer disease,
abnormal forms of proteins called tau and
amyloid-beta accumulate in the brain, and
the hippocampus—the region of the brain responsible for memory—shrinks. But here’s the
rub: there are no clear estimates for how much of these proteins exist in a normal brain. The
key to creating diagnostic tools, de Leon says, is to observe healthy people as young as age 50
and assess how these disease markers change over time. The results would help scientists set
standards for comparison and start treatment early, when it is likely to be most effective.
Using data from his studies, de Leon says he can predict with 90% accuracy which
individuals with mild cognitive impairment, a precursor to Alzheimer disease, will go on to
develop the disease (Neurobiol. Aging 27, 394–401; 2006). He hopes to collect enough
information to be able to set standards for disease markers even in healthy, young people.
When de Leon began his career in the 1970s, few scientists recognized the difference
between normal aging and cognitive disease. “Developing early diagnostic tools seemed so
remote and so daunting a task,” he says, “that I considered it the last thing I’d do before
I died.”
Emily Waltz, New York
751