Professor of biological sciences
University of Illinois at Chicago
Name: Susan Liebman
Position: Professor of biological sciences, University of Illinois at Chicago, 1977-present
Background: PhD and postdoc, genetics, University of Rochester; MS, biophysics, Harvard University; BS, biology, Massachusetts Institute of Technology
Long a workhorse of genetics, yeast cells are increasingly being explored as a model system for cell-based drug-discovery assays. In one of the latest examples of this trend, University of Illinois at Chicago researchers Susan Liebman and Sviatoslav Bagriantsev have developed a yeast-based Alzheimer’s disease model that can be used to examine the aggregation of even tiny amounts of β-amyloid, a process that has recently been shown to trigger the disease.
Liebman and Bagriantsev believe that the reporter system could be the basis of a high-throughput cell-based assay to screen small molecules for their ability to inhibit aggregation of β-amyloid, also known as A-beta. The assay is described in the Sept. 26 issue of BMC Biology, and the paper was the most viewed article online in October, according to the journal.
This week, Liebman took a few moments to discuss the work with CBA News.
This work is based on recent work that suggests that very small aggregates of B-amyloid may trigger Alzheimer’s disease, as opposed to large deposits or plaques. Can you elaborate on this?
This is not our work. But now the correlation between the presence of tangles and fibrils with toxicity is no longer one-hundred percent. More often, the correlation seems to be with the presence of oligomers. The size of the oligomer that might cause this toxicity isn’t clear, but it has been reported that even the low-end oligomers can form toxicity. It’s not sure that the low-end oligomers turn into larger oligomers and then turn into fibers, but that’s what people believe. The idea is, then, if you inhibit the appearance of fibers or stop fibers, that might be bad, because then you’ll have more of these oligomers around, which might really be the toxic entity. The goal is to inhibit these before the toxic species is formed. Whether it’s the fiber or oligomer, inhibiting it at the time of A-beta aggregation would catch both of those.
What are the relative sizes of these oligomers and these A-beta aggregates?
The oligomers we’re looking at are dimers, trimers, and tetramers. These are small oligomers – two, three, four, or five molecules. The oligomer that is kind of universal is somewhat larger, in the teens. We don’t actually see that in yeast – that large of an oligomer does not form. We see oligomers that are similar to the low-end ones that have been seen in cell culture.
Is most current Alzheimer’s drug discovery focused on these larger aggregates? How are Alzheimer’s drugs typically screened?
A lot of the academic papers have been focused on looking at drugs that would prevent fiber formation, and the simplest assay is to look for the appearance of the fibers. There are assays, for example, in which thioflavin fluorescence when amyloid fibers are formed. But that might already be too far down the road, when you already have the fiber. There isn’t an easy way to screen for the oligomer. In the very few studies where people have tried to look for compounds that might affect oligomers, they have looked for things that might prevent fiber formation. Then, they screen those again using a tedious gel-based screen to see if there is an oligomer. There are certainly other theories about the toxicity of Alzheimer’s disease. One of the main theories is that it is either these A-beta oligos or fibers that are causing the toxicity. But there are other molecules involved, like the Tau molecules. This is not a universal theory, and it may not be the universal cause.
Why did you choose yeast as the model for your cell-based assay system?
Yeast is a very simple organism; it’s non-toxic, it’s very easy to work with, and we have a lot of experience with it. We also had worked with other aggregating molecules in yeast, and we though the assay we used for those might work for A-beta.
For other disease models?
We worked with yeast prions, and they aggregate. It’s not a direct correlation. The diseases associated with prions are mad cow and Creutzfeldt-Jakob disease, and they are associated with the PRP protein in humans. Yeast doesn’t have a PRP protein, but it does turn out to have infectious proteins that are infectious to yeast. It’s not clear if there is disease, or if they are just changing a phenotype. One of these that we worked with is the translational release factor – so in protein synthesis, there is a stop codon that tells you to stop making proteins, and there is a highly conserved factor that helps stop the proteins.
It turns out that attached to that protein, in yeast, is a domain that causes it to form a prion. A prion is an altered state of a protein where its function is probably eliminated or very much reduced, because it forms an amyloid fiber. Once it forms that fiber, all the daughter cells form that. Indeed if you make that fiber in a test tube and inject it back into the yeast, it changes the yeast into having fibers that are inherited. We call that a self-propagating heritable protein. We work with this particular prion and others in yeast, and we can score in the yeast the appearance of the fiber or non-appearance of the fiber with an assay for whether that protein is functional or not.
My graduate student Sviatoslav Bagriantsev was instrumental in developing that concept into this assay. We took off the domain that makes the prion and replaced it with the A-beta to see if we can use the same assay, and it worked. Surprisingly, they didn’t form these large fibers, which we thought would happen. But they formed these oligomers, and that’s enough to inactivate the function of the protein.
But this is a simple growth assay instead of a direct assessment of the protein inactivation, right?
This is an assay that we’ve developed and that worked well. Once we know using this assay that the protein is inactivated, we can then use different kinds of reporters. We have a growth/no-growth scheme with this assay, and that seems like it would work quite well in a high-throughput screen.
How would you know in a screen that the yeast cells are dying or not proliferating due to some off-target or toxic effects of a test compound, as opposed to the inactivation of this protein?
We’re going to do a screen where only when the yeast has the compound, that’s when it will grow – otherwise it doesn’t grow. It’s going to promote growth. There are also other controls. When you have the aggregate – the dimer and trimer – that causes a certain phenotype in the yeast. That phenotype is misreading a stop codon. There are other ways to cause this misreading, such as a mutation in the release factor gene. We’ll use that as a control. If a drug also prevents the misreading of stop codons for those mutations, then that’s not a drug we want. We want drugs that only prevent the misreading of stop codons that’s caused by the making of the dimers and trimers that are inactivating the release factor.
You said that you have an interest in developing this into a high-throughput screening method for inhibitors of A-beta oligomer formation. What needs to be done to develop that?
We have a colleague who has a compound library that we can play with. It is a set of compounds that are known to cross the blood-brain barrier, so it’s sort of a head start if there did turn out to be an effective drug. And he also has some robots, so we can start off with a smaller screen – maybe 2,000 drugs – to try out our different assays and make sure we don’t get too many hits. And there are some other drugs that have been shown to be functional in other assays, so we can test them in this assay. And then through either an NIH grant or a company, we would try to ramp this up. We have an R-21 grant to do this first part.
Through a company that would [license the technology] from the University of Illinois at Chicago?
Yes, and we’ve applied for a patent through the university.