Q&A with Dr. Joshua Kritzer - a novel yeast-based drug discovery strategy

Dr. Joshua Kritzer, a postdoctoral fellow with Dr. Susan Lindquist at the Whitehead Institute, discusses his new publication in Nature Chemical Biology with Dr. Michael Rogan of MJFF. Dr. Kritzer describes a novel yeast-based drug discovery strategy, and utilizes that strategy to identify cyclic peptides that reduce alpha-synuclein toxicity in a Parkinson's disease model.

MR: Can you summarize your findings and what next steps you are taking to move this discovery further?

JK: We have successfully demonstrated a new drug discovery strategy that is rapid, inexpensive, and broadly applicable to models of human disease.  This strategy uses baker's yeast to generate millions of compounds, called cyclic peptides, and to test them simultaneously to discover which prevent cell death.  The yeast become, in effect, an army of millions of medicinal chemists, each synthesizing and testing a different cyclic peptide.  We applied this technology to a yeast-based Parkinson's disease model previously developed in the Lindquist lab, testing five million different cyclic peptides in a single week and pulling out two that rescued the cells in a disease-specific manner.  We also demonstrated that they can be expressed in dopaminergic neurons of live C. elegans worms, and that they are effective in a worm-based Parkinson's model.  Our next steps are to identify the cellular targets of the cyclic peptides, in order to understand how they work.  We are doing this by forming hypotheses based on the cyclic peptide's structure, which is similar to those of natural proteins involved in diverse processes including oxidation and metal binding.  We are also currently developing second-generation cyclic peptides by testing millions of cyclic peptides based on the two hits described in the paper, in effect repeating the original strategy to generate even more potent hits.

MR: I gather that the kind of cyclic proteins you have identified are not 'drug-like' - that is, they are not, themselves, candidates for drug development.  What are the steps that would need to be taken to move the results of this screening technology to a drug intervention suitable for clinical testing?

JK: These cyclic peptides are not traditionally "drug-like", in that they do not have the properties most associated with existing drugs.  But they could definitely be used as drug leads.  Peptides have been used as drug leads in the past and more and more biologics such as peptides are being tested in clinical trials.  These cyclic peptides are macrocycles, which puts them in the same chemical space as potent natural products such as cyclosporine and rapamycin.  These natural products are often cited as the exception to the rules of thumb that govern what chemists consider "drug-like," so there is reason to believe that cyclic peptides could indeed be developed into viable drugs.  Just as with any other drug leads, they would need an intensive medicinal chemistry effort to prove their worth as drug candidates.

What is exciting about these cyclic peptides is that we know that macrocyclic natural products can be potent inhibitors of protein-protein interations and other therapeutic targets that small molecule drugs can't even begin to attack. This technology gives us a unique opportunity to look for new drug leads in this extremely underexplored chemical space.  Thus, for diseases such as Parkinson's where there are few known cellular targets, using cyclic peptides will allow us to look for inhibitors of a broader range of targets than those that are targeted by traditional compound libraries.

MR: You have not yet identified the targets of your effective peptides - Do you have any clues as to their structure or function, or what PD-related pathological pathways may be involved?

JK: Yes -- the genetic encoding of the cyclic peptides allows us to test variants very rapidly, helping us identify a core motif within the two cyclic peptides responsible for their function.  This motif, composed of two cysteine amino acids, is involved in many cellular processes.

Most relevant to Parkinson's disease, a double cysteine motif is used by cellular proteins that act to balance oxidants and reductants in the cell, and similar motifs are also used to bind metal ions.

Parkinson's is characterized by rampant oxidative damage in affected neurons, and imbalances of certain metals, most notably manganese, have been linked to certain kinds of Parkinsonism.  Thus, we are using these pathways as starting points to elucidate how these cyclic peptides work at the molecular level.

Joshua A Kritzer, Shusei Hamamichi, J Michael McCaffery, Sandro Santagata, Todd A Naumann, Kim A Caldwell, Guy A Caldwell & Susan Lindquist Rapid selection of cyclic peptides that reduce alpha-synuclein toxicity in yeast and animal models. Nat Chem Biol. published online 13 July 2009; doi:10.1038/nchembio.193