Buck Institute for Age Research
Who: Robert Hughes
Position: Assistant professor, Buck Institute for Age Research
Background: Director of therapeutic biology, Prolexys Pharmaceuticals, 2002 to 2005; acting assistant professor, medical genetics division, department of medicine, University of Washington, 2002 to 2002; PhD, Yale University, department of biology, 1992.
Robert Hughes was the lead investigator on research looking into protein-protein interaction in Huntington’s disease, a condition estimated to affect as many as 200,000 Americans.
Using technologies developed by Prolexys Pharmaceuticals, Hughes and his colleagues identified 234 proteins that could be associated with the disease.
They then validated 17 loss-of-function suppressors that could prove to be potential targets for therapeutics. Their work is published in the April edition of PLoS Genetics.
Below is an edited version of a conversation ProteoMonitor had this week with Hughes.
Can you describe the work you did and how you did it?
So Huntington’s disease is caused in the protein huntingtin. And we wanted to do a comprehensive survey of the protein interactions involving huntingtin. At Prolexys in Salt Lake City, they built a high throughput protein interaction discovery system, using two parallel technologies: one high throughput yeast 2-hybrid, and the other was high throughput pulldowns with mass spectrometry analysis.
We took huntingtin and we put it through these massive amounts of genetic screening in the yeast 2-hybrid, and a very significant number of pull-downs in the mass spec [analysis]. Then, after getting all this information and filtering the data to very high levels of stringency, we came up with 234 protein interactors. Now, roughly half of those were discovered with Y2H and half with mass spec.
And what we wanted to do then was to validate these interactions as having some potential role in modifying Huntington’s disease. The way that we chose to do that was with our colleague Juan Botas [at Baylor College of Medicine]. He had a fly model of HD where he expressed mutant huntingtin in the eye of the fly. This was set up to do genetic screens to find mutations that could enhance or suppress the huntingtin-mediated neurodegeneration in the fly.
We took our 234 genes that we discovered in human cDNA libraries or in mouse tissue extracts and we converted those into Drosophila and then obtained a large mutation in a large number of Drosophila genes that were orthologous to the genes that we picked in our protein interactions.
We got mutations in about 60 different genes with multiple alleles in many cases, and we took these mutant flies and we crossed them to the huntingtin fly, and basically found that almost 50 percent of the fly orthologs when mutated should modify the huntingtin’s mediated neurodegeneration in the fly.
That essentially validated this large protein interaction dataset. I think one of the things that makes this an interesting study is the fact that [through] the validation of the study using a high-content genetic model, we’re able to show that these protein interactions really do have a potential to being involved in Huntington’s disease because they can modify the huntingtin neurodegeneration in the fly.
We selected a group of about 60 just to test, and about half of those were modifiers, so if we scaled that up to the larger dataset it suggests that up to 100 of these things could be candidate modifiers for the disease.
What needs to be done now? You say they could be candidate modifiers.
What we’re doing now is testing them. Using siRNAs, we’re testing them in mammalian cell-based models of HD toxicity. The idea is we discovered all these things in mammalian systems, i.e. mammalian cDNA libraries or mammalian proteomic extracts. We then validated them in the fly, so now we’re going to go back to mammalian cells. I think once we validate them in mammalian cells, then we’re going to get our hands on the really valuable targets.
Was this the first comprehensive survey of proteins for Huntington’s disease?
No, people have done sort of protein-interaction screenings, and there was one other large scale screening that was published a while ago. But this is, by far, the biggest that’s ever been published.
And these initial 234 proteins that were identified, they were novel proteins?
Most, about 90 percent, were novel.
What did we know about the role of proteins in Huntington’s before your study?
We knew that a number of protein interactions were involved in Huntington’s disease and the possibility that huntingtin, when mutated, actually interferes with other proteins and affects their functions, and this may have to do with the cellular pathology of the disease.
Is there any way to characterize these 234 proteins, what their functions are, or roles are?
Most of them are annotated, so we published the full list in the paper in PLoS.
How long did you work on this?
I’ve been working on this for four years, and there are a lot of authors on this study. It was a very long and very time-consuming and labor-intensive study.
What were some of the obstacles and difficulties you encountered during this time?
I’d say the biggest obstacles were purifying the huntingtin proteins in sufficient quantities to do a large number of pulldowns.
How did you do that?
Just by scaling E. coli production of small huntingtin fragments. And then optimizing the methods for getting reproducible pulldowns for the mass spec analysis. That took a lot of work.
The only reason we were able to do it was because Prolexys has such advanced high-throughput proteomic technology. The primary data in this study could never have been done solely in an academic group, just because of the scale. I mean it really was large scale proteomic discovery.
Does this move us significantly closer to finding a therapy, if not a cure, for Huntington’s?
Well, I don’t know if I’d want to make such a bold claim. Let me put it this way. I think that drug discovery for a disease requires a number of steps, and one is target discovery, and then assay development, and then chemistry and pre-clinical stuff.
So target discovery is really the first step. And I think that the significance of this study is that it significantly expands the number of candidate targets that can be looked at with higher content validation technologies.
I think its significance is in expanding the portfolio of HD targets that can be worked on.
The real thing to do next is to confirm the Drosophila results in mice. We’ve got about two targets now in mice that we’re already studying … now to see if they modify HD.