|
Return to Past Workshop
Reports main page
Workshop Reports
Hereditary Disease Foundation
In conjunction with:
The ALS Association and
The Jennifer Jones Simon Foundation
Protein-Protein Interaction in
Neurodegenerative Disease
June 19 and June 20, 1996
Los Angeles, California
Prepared by Sandra Ackerman
Protein-Protein Interactions in Neurodegenerative Disease
June 19 - 20, 1996
Los Angeles, California
Participants
Robert V. Abendroth
ALS Association
Helga Ahrens
University of Wisconsin
Melanie Bennett
University of Pennsylvania
Azad Bonni
Harvard University
David Borchelt
Johns Hopkins University
Robert Brown
Massachusetts General Hospital
Steven Clarke
University of California Los Angeles
Don Cleveland
University of California San Diego
Gregory Cole
Sepulveda Medical Center
David Eisenberg
University of California Los Angeles
Stanley Fields
University of WashingtonH. Robert Horvitz
Massachusetts Institute of Technology
Karen K. Hsiao
University of Minnesota
Ivan Lieberburg
Athena Neurosciences, Inc.
John E. Maggio
Harvard Medical School
Diane Merry
University of Pennsylvania
Donald L. Price
Johns Hopkins University
Ethan Signer
Pharmaceutical Peptides, Inc.
Lawrence Steinman
Stanford University
Allan Tobin
University of California Los Angeles
Joan S. Valentine
University of California Los Angeles
Nancy S. Wexler
Columbia University
Protein-Protein Interactions in Neurodegenerative Disease
June 19-20, 1996
Los Angeles, California
This workshop was convened under the sponsorship of the Hereditary
Disease Foundation, the ALS Association, and the Jennifer Jones Simon
Foundation. Dr. Allan Tobin served as moderator throughout the two-day
discussion.
On June 19, Dr. Nancy Wexler opened the discussion by explaining the
rationale for bringing together researchers from many different areas of
expertise. Such meetings, because they offer fresh perspectives, can be
energizing. Moreover, findings from one disease are very likely to be
important for others as well; as an example, work on the gene for
Huntington s disease led directly into work on ALS.
Dr. Tobin s introductory remarks highlighted the theme of cell death. Dr.
Robert Horvitz, in an overview, then said it is clear that two families of
molecules are involved in cell death. One family, including the bcl-2 and
ced-9 genes, serves a protective function; the other, including the ced-3 and
ICE genes, has a destructive function, coding for proteases that cleave with
great specificity after aspartate residues. The two families of genes may, in
theory, interact in any of several ways: the protective genes can work by
preventing killers; the killers can work by preventing protection; or some
more complicated interaction may be going on. Inhibitors of the proteases--
the cowpox virus protein crmA, the baculovirus protein p35, and assorted
peptides--may be useful for possible therapies, as well as for investigation.
Dr. Donald Price suggested that apoptosis, as a rather rapid cellular
process, may actually represent a late stage in a long cascade of
abnormalities. Dr. Ethan Signer observed that diseases such as ALS and
Huntington s are characterized not by cell death but by cellular dysfunction.
Dr. Karen Hsaio pointed out that in a host of diseases (schizophrenia,
manic-depression, dystonia, Tourette s syndrome, and various movement
disorders) the brain is clearly not functioning perfectly, and yet there is no
cell death. Dr. Price maintains, however, that even in these cases there may
be cell death, albeit at a more subtle level.
Dr. Joan Valentine raised the question whether apoptosis is just one
specialized form of programmed cell death, pointing out that certain primitive
life forms exhibit other ways of programmed cell death. Dr. Gregory Cole
likened a preoccupation with apoptosis, or cell suicide, to the situation of a
person who has many serious problems: the person may be saved from
suicide but still has the problems. Similarly, whether or not cell death in
itself is a problem, cell damage clearly is.
Dr. Horvitz discussed recent work identifying cells that are "partial killers,"
either causing some damage but not cell death, or killing the cell but taking a
very long time. He said that "apoptosis" and "programmed cell death" are
not synonymous and that the way to clarify the relationship among various
cell deaths will be to look more closely at the mechanisms involved. Dr.
Diane Merry pointed out that in a disease such as Huntington s, the
degenerative process is a very slow one; this raises the question, in the
case of a hypothetical 20-year-old who begins to develop the disease, of
when to intervene.
Dr. Price said that in ALS, two factors contribute to paralysis: destruction
of the lower motor neurons, which leads to weakness and atrophy, and
destruction of the upper motor neurons, which leads to spasticity and
increased reflexes. By contrast, according to Dr. Robert Brown, polio affects
the lower motor neurons almost exclusively. He observed that ALS has the
striking characteristic of appearing first focally and then disseminating;
Huntington s and Parkinson s disease may also begin focally and then
spread either by proximity or by connectivity. Dr. Ivan Lieberburg pointed
out, however, that the cells involved in these diseases each use different
neurotransmitters.
Dr. Don Cleveland then posed the question, What is it about these cells
that makes them vulnerable? One possible answer is that they are "long"
cells (i.e., with long axons), but in fact they are not the body s longest.
Another possibility is that they are simply among the bigger cells, and
somehow their greater size puts them at risk. Currently, mutations in any of
three genes are implicated in the degeneration of motor neurons: an
expanded repeat in the androgen receptor; defects of superoxide dismutase
(SOD); and overproduction of neurofilament.
Dr. Merry addressed the question of why motor neurons die, using as an
example Kennedy s disease, or spinobulbar muscular atrophy. This disease
is caused by an expanded polyglutamine repeat (CAG) in the androgen
receptor: perhaps 40-60 repeats, as compared with about 13-30 in a normal
individual. This is thought to be a protein-protein interaction rather than a
protein-RNA interaction because, according to Max Perutz among others,
the expanded repeats can make other protein products. The evidence so far
is only circumstantial, but it is accumulating. Dr. Merry thinks the expanded-
repeat androgen receptor may have something to do with the lower level of
androgen receptor protein. This is mainly a toxic gain of function, but there
may also be some decrease of normal function. Interestingly, even though
the androgen receptor is expressed in a wide variety of neurons, the only
ones affected by the disease are the motor neurons.
In answer to a question from Dr. Tobin, Dr. Merry said the protein is
somewhat more abundant in motor neurons than in other neurons; it is
particularly abundant in the sexually dimorphic nucleus that affects the
perineal muscles. Discussion turned to the question of whether the number
of repeats in excess of normal might in fact be quite small; according to Dr.
Cleveland, new evidence from Dr. Mandel at Cold Spring Harbor points
toward a threshold effect. Dr. Steven Clarke observed that a number of
aging processes also show a threshold effect.
Dr. Cleveland spoke under the heading of "How to Kill a Motor Neuron."
The most abundant protein in large motor neurons is neurofilament;
incorporation of this protein enlarges the diameter of the axon from less than
1 micron to about 10 microns. The growth begins at the time of myelination,
at birth, but continues into old age, so there is a large (up to one
hundredfold) increase of axonal volume over the course of life.
When Dr. Cleveland s lab made mice with an overabundance of the NF
protein, the mice developed motor neuron disease and then died--before the
neurons themselves had died. Clearly, then, this is not a good mimic of
ALS. Subsequent mice expressing a neurofilament mutation do show
selective death of motor neurons, accompanied by aberrant aggregates of
neurofilaments very similar to those found in several studies of human ALS.
However, in mouse models of ALS, upper motor neurons are spared; but this
is perhaps not surprising, since mice do not have any upper motor neurons
larger than 1 micron in diameter. It is possible that the neurofilamentous
excess puts neurons at risk by disrupting axonal transport. Dr. Cleveland
argues that these proteins are long-lived (requiring one to two years to
transmit from the cell body to the end of the longest nerves) and that
aberrant accumulations can build slowly, ultimately strangling the neuron by
interfering with transport. Apparently, ample amounts of wild-type NF
proteins are, at the least, much less deleterious than modified or mutated
proteins, making neurofilaments attractive candidates for SOD mutant
mediated damage.
The "stuff" that occludes transport may depend on the mutant. Some
SOD mutants, at least in mice, yield unusual membrane-bounded structures,
first in cell bodies and later in axons and dendrites. In others (as found by
the Hopkins/UCSD consortia), membrane-bounded "vacuoles," apparently
derived from degenerating mitochondria, appear first in axons and dendrites,
and with cell bodies largely spared. In another, newer set (from San Diego
and Baltimore), there are accumulations in glial cells, but no vacuoles.
Dr. Price added that one mutant, when expressed at lower levels, is an
even better mouse model for human ALS, as it develops no vacuoles but
has abnormal accumulations of SOD and ubiquitin, features known in human
disease.
Dr. Tobin, calling for common themes, asked, Is ALS a clogging disease?
If so, where do you intervene? He asked for perspectives from people who
work on mouse models for Alzheimer s disease.
Dr. Cleveland said that the various mouse models that show progressive
disease from expressing familial ALS-linked SOD mutants display lower
levels of endogenous proteins or activity, a finding inconsistent with disease
arising from reduction in overall SOD activity.
Dr. Clarke presented the issue as a war between biology and chemistry:
You make a perfect protein, and then biology takes over. He observed that
probably half the reactions are oxidative. Problems arise from residues that
kink the polypeptide chain, residues that move just 0.4-0.5 mm from their
proper place.
As to ubiquitin, the significance of this protein is that it signals the cells for
degradation. It is present normally at such low levels that once it can be
detected, this is already a bad sign. Ubiquitin is found in Alzheimer s
disease and Down syndrome, for example, and its presence is thought to
come much earlier than DNA fragmentation.
Dr. Valentine suggested another possible reason that motor neurons may
be vulnerable to these particular diseases. The SOD protein is synthesized
and then takes about 2 years in transport; all the while it is aging and is
subject to environmental "hits."
Dr. David Eisenberg discussed proteins that bind to themselves by means
of "domain-swapping." The result is a protein that is very tightly bound, hard
to get apart.
Dr. Stanley Fields talked about the possible applications of methods from
yeast studies to the diseases of interest. He asked hypothetically, If x-dose
means you will develop a disease in 40-50 years, does 0.5x-dose delay the
development of the disease until age 80 or 90? One can think of yeast
assays as a way to disrupt protein-protein interactions, he suggested, but it
is not clear which interactions are to be disrupted. To take Huntington s
disease as an example, the question is whether the function of huntingtin
contributes to the development of the disease? The function of this protein
is not yet known; a rescue of the huntingtin-knockout mouse has not yet
been developed.
On the question of how to interfere with protein-protein interactions, Dr.
Signer observed that small organic molecules are not typically effective at
this, but he thinks that peptides are effective. Dr. Horvitz noted that it has
not yet been convincingly established that protein-protein interactions are
the crux of the problem.
In the case of Alzheimer's disease, Dr. Lieberburg remarked that there
are other targets for intervention that are more tractable.
Dr. John Maggio said that one thing done so far has been to consider the
effect of concentration on a protein s ability to interact with itself: reactions
that occur at high concentrations of a peptide may not take place at a lower
concentration. Conformation, the way the protein folds, should also be
considered, and a peptide may fold very differently in vitro from the way it
folds within the amyloid plaque in vivo. Dr. Hsaio remarked that disrupting
prion proteins actually can increase their interactivity by a factor of 10,
because it opens up more surface for interaction.
Dr. Signer thought it best to try to develop an assay first, without worrying
about a drug at this point. His pharmaceutical company has decided to
focus on amyloid and has developed an in-vitro assay for an in-vivo disease.
Dr. David Borchelt added that in contrast to methods from tumor studies, in
which the cells are transformed and then remain alive for long studies, in
studies of neurodegenerative disease the cells do not get transformed but
die. One does not have the luxury of time.
Discussion then turned to the possible role of oxidative stress in
Alzheimer s disease. Dr. Lieberburg said that most studies so far have had
small populations and have been quite disparate, making comparison
difficult. In progress now is a Phase II trial of prednisone and vitamin E. Dr.
Robert Brown said that in the case of ALS, there are not good data
demonstrating that oxidative stress is important. Giving ALS mice vitamin E
can delay the onset of disease but does not affect the course of disease
once it starts. Dr. Brown thinks these two are different processes, and that
dissemination of the disease is still another. Also, according to several
researchers present, the ability of vitamin E to penetrate the brain is either
fairly poor or fairly slow.
The group then had the opportunity to talk with a woman with
Huntington s disease who had been a professional dancer for many years
before developing symptoms of the disease. The woman s husband and
primary caretaker accompanied her and mainly spoke for her, since her
speech was impaired. The couple described the enormity and complexity of
the physical and emotional toll of the disease. One burden may be resolved,
only to face a new one.
On June 20, the workshop began with a visit from a 67-year-old man who
first observed possible symptoms of ALS in 1991 and has taken an active
role in his own diagnosis and in decisions about treatment. He has
participated in both the CNTF and the BDNF II drug trials and keeps abreast
of new developments in ALS research by means of the Internet.
His visit focused discussion on ALS, particularly the role of SOD
mutations. Dr. Horvitz observed that although the enzyme is involved with
free radicals, it remains conceivable that this disease may have nothing at
all to do with free radicals. The mutant enzyme may simply trigger an entry
into the disease pathway. Dr. Price has found that in mice expressing
mutant SOD, more wild-type SOD present causes the disease to develop
faster. Dr. Brown said he thinks SOD could account for the spread of the
degeneration as well as for the initiation.
It is possible to identify histologically different cells that are or are not
affected, and those affected are not always contiguous. According to Dr.
Price, with SOD the principal target appears to be mitochondria, and the
process appears similar to excitotoxicity. Dr. Valentine mentioned two
models for a gain of function; her model focuses on hydrogen peroxide and
the binding of copper. Dr. Horvitz called attention to the importance of Dr.
Price s earlier observation that increasing the wild-type SOD accelerates the
onset of disease. This could be because of increased stability, because of
increased normal function, or because of a completely unrelated or
unidentified function.
Dr. Clarke presented a protein-chemical perspective: polyglutamine has
hydrogen bonding not just in the main chain but also in the side chain. This
gives a very sticky protein, sometimes to the point of complete insolubility.
In amyloid, there may somehow be a platform for oxygen reactions to occur;
glutamines, also, may have such a platform. Dr. Clarke said the
spontaneous chemistry of aging includes cross-linking, oxidative chemistry,
and enhanced spontaneous degradation.
Dr. Tobin asked what kind of experiments would be needed to establish
that a disease arises from an accumulation of these spontaneous changes
rather than from some kind of threshold effect. According to Dr. Clarke this
would be quite difficult to establish, with the multitude of reactions that are
continually occurring among 80,000 genes and a similar number of
polypeptides. Ongoing processes of genetic repair would also complicate
the picture. Dr. Hsaio raised the perspective of biological time as distinct
from chemical time, which may include half-lives on the order of days or
even minutes.
After some general discussion, Dr. Tobin singled out the issue of what
constitutes an adequate demonstration of causality. He asked Drs. Hsaio
and Lieberburg each to talk about their mouse models of Alzheimer s
disease. Dr. Hsaio said that the host strain was turning out to be far more
important than had been thought. Her transgenic mouse model, starting with
the FVB/N strain, was made to overexpress the 695-amino acid isoform of
human Alzheimer beta-amyloid precursor protein (APP) [Science 274:99-
102, 1996]. At 150 to 500 days of age, about 20% of these mice develop a
disorder of the brain: agitation, "neophobia," or tendency not to explore new
surroundings, and gliosis. In high-resolution PET scans, they show
decreased utilization of glucose in the temporal and parietal cortex; most
heavily affected is the entorhinal area, but the somatosensory and occipital
areas are spared. Dr. Hsaio considers this functionally similar to
Alzheimer s disease because similar areas of the brain are affected in the
two diseases; equally important, similar areas are spared. These mice do
not, however, develop neurofibrillary plaques or tangles, so this model does
not serve for the pathology.
By contrast, Dr. Lieberburg s transgenic mouse model focuses on the
pathology rather than the behavior, and was developed primarily as a way of
testing drugs. This mouse, which overexpresses human APP with a codon
717 mutation [Nature 373:523-527, 1995], shows the same regions spared
as those in Dr. Hsaio s model. In other regions, there is significant
pathology in heterozygotes at 7-9 months, and in homozygotes at 3-4
months. The pathology involves gliosis and loss of synapses but not tangles,
and is heritable over at least 10 generations. Dr. Lieberburg said that if his
group had found some substance that reduced amyloid burden by about
30% or more, they would probably want to proceed to clinical trials
regardless of whether it changed rodent behavior, since he does not think
the latter extrapolates well to human behavior.
As the discussion began to wind up, Dr. Signer called for people to
generate testable hypotheses, such as testing the idea of a possible
polyglutamine zipper (from the work of Perutz). Dr. Fields recommended
using yeast hybrids to turn down the production of SOD mutants. Dr. Clarke
suggested testing ideas about the accumulation of "hits" to aging proteins,
tests on polyglutamine or on nearby proline-rich regions. In work on mutant
SOD, Dr. Eisenberg called for cell-biological assays in addition to transgene
assays.
Dr. Tobin then asked each participant to give their best ideas and
recommendations for action.
Dr. Cleveland, agreeing with Dr. Price, wants to learn more about the
mouse model developed by Dr. Bates for Huntington s disease. He also
likes the idea of experiments exploiting the combinatorial assembly of
peptides to search for initiators of enzymes such as SOD or to inhibit
protein-protein interactions.
Dr. Merry emphasizes the importance of developing animal models,
whether for diseases (including behavioral effects) or for the underlying
pathogenesis. She would like to use combinatorial approaches to study
interactions and ways to disrupt these interactions, and looks for better
communication among protein chemists, geneticists, and molecular
biologists.
Dr. Hsaio says the animals used should be studied in and of themselves,
because it is unrealistic to expect their diseases to mirror those of humans:
in particular, the idea of Alzheimer s as a uniquely human disease is
incorrect.
Dr. Cole proposes to look at ubiquitin as a marker of damage in cells and
to study slices of adult striatum from HD mice for earlier signs of
neurodegeneration.
Dr. Borchelt plans to make an expression plasmid library for peptides that
might "plug" the hole and prevent SOD from acting.
Dr. Brown would like to look for CNS-permeable small molecules that can
find their way into channels and block copper toxicity. As a clinician, he
sees his clinical practice in ALS as the gold standard for testing hypotheses
about the disease.
Dr. Azad Bonni would work with the yeast-2 hybrid, using libraries of
genes from the regions involved in Huntington s disease. He would look at
"all-or-none" differences in polyglutamine and at the problem of instability of
the repeats. He is also interested in the possibility of post-translational
modification of signaling compounds, especially tyrosine.
Dr. Melanie Bennett s plan is to try replacing glutamine zippers with
asparagine zippers. She will also look at polyglutamine repeats in other
species; for instance, why does the pufferfish have only 1 repeat?
Dr. Maggio calls for some "fishing" for other proteins involved and for
markers for pathology. He would also use combinatorial methods to find out
more about protein-protein interactions and what interferes with them.
Dr. Signer, struck by the high caloric consumption that is a feature of
Huntington s disease, suggests looking at glyceraldehyde-3-phosphate
dehydrogenase (GAPDH). Also, taking into consideration the cost of
transgenic mice, he suggests working lower on the phylogenetic scale, and
thinking in terms of an assay rather than a model.
Dr. Fields feels it is still very early to tell about what the proteins are doing
or what interactions are involved. He thinks one viable route toward therapy
might be finding compounds to bind to a protein, then asking if they do
anything in transgenic models.
In Dr. Price s view, the workshop boiled down to markers and targets,
neither of which is very clear yet. He wants to know whether the mice in Dr.
Bates s model really show anything in the striatum. He reminds everyone
that there has been extraordinary progress in the last few years: we now
have models and approaches for the prion diseases, ALS, AD, and HD.
Dr. Clarke wonders whether the high caloric consumption of HD means
that with more energy burned and more oxygen required, there is more
oxidative damage. He sees an exciting possibility that spontaneous damage
may be part of the disease process, and says that with the prospects good
for developing a mouse that lacks a repair mechanism, he would like to mix it
with transgenic mouse models for pathology and see what happens.
Dr. Lawrence emphasizes that direct studies on human tissues should still
be encouraged.
Mr. Robert Abendroth is struck by the focus that has shifted from "do what
is doable to reaching out, fishing in waters that haven t been fished in yet;
he encourages everyone to continue this.
Dr. Valentine now thinks it worthwhile to look for something that binds to
SOD; she is looking at yeast, in which the only SOD is mutant. She
suggests studying the possible role of zinc, because it binds competitively
with copper and may be useful in dealing with possible copper damage. She
also raises the question, Is there a possible role for redox metals in these
diseases?
Dr. Helga Ahrens would like to create a peptide library and combine it
with a technique for getting the peptide into the cell. She would like to work
with someone with a mouse model, to check bioactivity.
Dr. Lieberburg thinks the hypermetabolic state of HD patients indicates
there is clearly something wrong; it may be an important clue.
Dr. Eisenberg would look at crystal structure and at threading, as possibly
showing the different structure of domains in the HD gene. He needs pure
material (and wonders if the Hereditary Disease Foundation can get
involved).
Dr. Horvitz wonders whether the protein is needed for the pathology. He
would ask this of the mouse model and find out mechanistically what the
expanded repeat is doing. He also asks, Does the pathology seen in the
SOD/ALS mutant mouse depend upon the presence of the neurofilament
gene? This issue should be easy to resolve using the methods of mouse
genetics.
Dr. Wexler observes that since we now have an antibody, a gene, and a
protein for HD, research in this area has in fact come further and faster than
anyone could have anticipated. A major issue will be how to get the protein
purified.
Dr. Signer asks whether anyone can put nuclei for the disease into cells
with normal mitochondria.
Dr. Wexler calls on everyone to think about what might be therapeutic
without worrying too much about how it can pass through the blood-brain
barrier, since others are working on that challenge. An advantage of being
able to identify disease genes carriers is to start therapy very early, possibly
even in utero.
Finally, Dr. Tobin thanked the participants and said that while
methodological and conceptual problems remain, he hopes the workshop
will have been a magnet for further efforts. |
|
