SfN 2013—New Tools for Rational Drug Design
See Allison Curley's snapshots from the conference.
January 28, 2014. On November 12, 2013, the penultimate morning of the Neuroscience 2013 conference in San Diego, Bryan Roth of the University of North Carolina, Chapel Hill, gave a special lecture titled, “How Synthetic and Chemical Biology Will Transform Neuroscience.” In actuality, it would be a talk on how team science can transform neuroscience, prefaced Roth, emphasizing that a great number of people have been involved in the work he would discuss.
Roth played a video of a young man experiencing florid hallucinations shortly after smoking the psychotomimetic drug salvinorin A, the main ingredient in the salvia plant. A member of the mint family, salvia has been used for centuries in the spiritual practices of the Mazatec people in Mexico, but more recently, salvinorin A has become a recreational drug that is marketed and sold as a legal alternative to marijuana. Throughout his talk, Roth used salvinorin A to illustrate the approach his group has taken to improve drug discovery. The salvinorin A story “is emblematic of a large field of drug discovery that’s currently ongoing … to create drugs that engage specific signaling pathways,” he said.
Scanning the receptor-ome
The plasma membrane is home to over 1,000 different flavors of receptors. In order to identify the specific molecular targets of salvinorin A, Roth and his team screened the compound against all known drug targets (collectively termed the receptor-ome). Using robotic screening and automated analysis of competition radioligand binding assays, this approach can screen hundreds of compounds against the known drug targets at once, an approach Roth noted was the opposite of the single-compound approach typically used by drug companies.
Using this high-throughput approach, Roth and colleagues found that salvinorin A selectively activates κ-opioid G protein-coupled receptors (GPCRs; see Roth et al., 2002). GPCRs represent the largest target class and comprise 4 percent of human DNA, making them “very good molecular targets,” said Roth. Functional studies revealed that salvinorin A is a very potent and selective κ-opioid receptor (KOR) agonist, with very rapid pharmacokinetics. It is quickly converted to the inactive salvinorin B, which explains why users only hallucinate for a short period of time. Surprisingly, salvinorin A showed no action on serotonin 5HT2A receptors, the predominant target of other hallucinogens.
In addition to characterizing compounds of unknown pharmacology, this approach has broader implications for drug discovery. The salvinorin A example suggests that opioid receptor antagonists could be useful for treating psychotic disorders such as schizophrenia and implicates these receptors in the modulation of perception.
Designing drugs and receptors
Next, Roth and his team turned to how salvinorin A engages with its target, KOR. Although it took over five years and many collaborations, he said, the crystal structure of all four opioid receptors was finally reported in 2012 (Wu et al., 2012), opening up the possibility for structure-based drug design.
KOR agonists are promising analgesics, but side effects such as hallucinations limit their clinical potential. Activation of a GPCR stimulates two different downstream pathways: the G protein second messenger system and signaling through β-arrestin. Some drug-dependent behaviors are mediated by G protein signaling, while others stem from β-arrestin’s effect, Roth said. He and colleagues have developed ligands that show a preference for one pathway over the other, and the hope is that this biased signaling can be harnessed to develop drugs that have only the desired effect(s) (see SRF related news report).
Roth’s graduate student Kate White explored the functional selectivity of KOR ligands in a poster she presented the following day. Using β-arrestin 2 knockout mice (that exclusively rely on G protein signaling), White has identified behaviors that seem to rely on one pathway versus the other. She has also identified biased KOR ligands in vitro, and in-vivo experiments with these tools are underway.
Drug design is complicated by the fact that ligands often act at more than one receptor type, many times leading to toxicity and side effects, said Roth. One way to get at this multi-target problem is the Similarity ensemble approach (SEA), which can be used to predict the off-target effects of drugs. Developed by Roth in collaboration with Brian Shoichet at the University of California, San Francisco, SEA is a computer-screening tool that relates targets by ligand similarity irrespective of protein sequence or structure (Keiser et al., 2007).
However, multiple targets aren’t always problematic, Roth continued, and may actually be helpful in many cases. With so many risk alleles linked to complex diseases such as schizophrenia (see SRF related conference report), treatment will require either combinations of medications or drugs that act at multiple targets. Roth developed an automated approach to create the latter that uses machine learning algorithms and large databases of pharmacological information to create novel drug-like compounds in silico (Besnard et al., 2012). By permitting the de novo, rational design of compounds, this approach removes a major roadblock to central nervous system drug discovery, he added.
To discover how salvinorin A’s profound effects on human consciousness are achieved, Roth harnessed the “awesome power of yeast genetics” to create a new technology: Designer Receptors Exclusively Activated by Designer Drugs (DREADDs). Like a chemical version of optogenetics, DREADDs are receptors that have been engineered to respond to the otherwise inert ligand clozapine-N-oxide. By adding this small molecule to animals’ food or water, researchers can control GPCR signaling and, therefore, neural activity. Complementary to optogenetics, this chemogenetic approach allows for longer lasting and broader effects, and provides researchers with another powerful tool with which to dissect neural signaling. DREADDs have spread like wildfire through the neuroscience research community for a wide variety of applications such as stimulating receptors in the cortex and hippocampus, and silencing neurons (including the parvalbumin-expressing ones of particular interest to the schizophrenia research community (see SRF related news report; SRF news report).
Together, predicted Roth, novel structural biology technologies and the non-invasive modulation of neuronal networks will, within the next several years, allow for any drug with a beneficial or harmful effect to be screened against all known druggable targets. This in turn could pave the way for safer and more effective drugs.—Allison A. Curley.