Opioids remain a mainstay of pain therapy, but their addictive qualities and sometimes life-threatening side effects, such as respiratory depression, have fueled the search for better drugs. Now, a collaborative research effort has identified an opioid-like molecule that not only eases pain in mice, but does so without the usual adverse consequences of opioids.
By simulating nearly four trillion interactions between known compounds and the µ-opioid receptor, the researchers narrowed their search to about two dozen molecules. From these, they converged on PZM21, a µ-opioid receptor-specific agonist that biases the receptor toward G protein signaling, thought to produce analgesia, and away from β-arrestin-2 signaling, which is associated with opioid side effects. Much like morphine, a single dose of PZM21 reduced pain behaviors both in the hotplate test and formalin injection nociception assay. But unlike the classic opioid, PZM21 left breathing unaffected and did not appear to be rewarding.
Michael Bruchas, Washington University, St. Louis, US, who was not involved in the current study, said the discovery is timely. “If there’s any example of a rationally designed drug that people have been waiting for, it’s something for the µ-opioid receptor that doesn’t have these other side effects."
“You have these different spaces coming together,” he added, with the study exploiting the recently revealed structure of the µ-opioid receptor to find biased ligands, at a time when opioid abuse is widespread.
The study was published online August 17 in Nature and led by senior authors Brian Kobilka, Stanford University School of Medicine, US; Peter Gmeiner, Friedrich-Alexander-Universität Erlangen-Nürnberg, Germany; Bryan Roth, University of North Carolina, Chapel Hill Medical School, US; and Brian Shoichet, University of California, San Francisco, US.
From millions to one
In 1999, Laura Bohn and colleagues reported that mice genetically lacking β-arrestin-2 were quite unlike their wild-type counterparts (Bohn et al., 1999). These mice showed greater analgesia in response to morphine, and were later found to be less liable to have respiratory depression and constipation (Raehal et al., 2005). Together, these findings and others suggested that, by biasing signaling of the µ-opioid receptor away from the β-arrestin-2 pathway, researchers could perhaps tip the balance of opioid effects toward analgesia without side effects.
This idea took on new life once the structures of the opioid receptors, including the µ-opioid receptor, were captured (see related PRF news stories here and here). “After the structures were solved, we realized that it would be possible to use them to discover new and potentially useful drug-like molecules,” said co-senior author Roth. Roth’s lab helped discover the structures of the nociceptin and κ-opioid receptors.
In the new work, lead authors Aashish Manglik, Henry Lin, Dipendra Aryal, and colleagues thus began by computationally docking three million commercially available “virtual molecules” against the µ-opioid receptor’s binding pocket, testing 1.3 million configurations for each. Based on the molecules’ novelty and interaction with important amino acid residues, the researchers whittled down their list to just 23 compounds. Ultimately, they arrived at one potent drug that activated Gi/o (a G protein subunit) far better than another µ-opioid receptor agonist, DAMGO, in human embryonic kidney cells, and weakly recruited β-arrestin-2 in human bone osteosarcoma cells. The researchers had therefore found a biased agonist.
They next optimized the drug to improve its affinity for the µ-opioid receptor, and its potency and efficacy. One isomer had greater affinity for the receptor and, compared to three other isomers, was the most potent and efficacious at activating Gi/0. The investigators named this version PZM21. The authors went on to find that PZM21 was selective for the µ-opioid receptor, as it failed to activate the κ-opioid receptor or the nociceptin receptor, and only weakly activated the δ-opioid receptor. Additionally, PZM21 had no appreciable activity at hundreds of other G protein-coupled receptors, and was a much weaker agonist at several monoamine transporters.
Fewer side effects
How did PZM21 compare to other drugs? The new compound activated Gi/o to a similar degree as morphine and TRV130, a biased ligand being developed by the pharmaceutical company Trevena and currently in Phase 3 clinical trials for acute pain (Soergel et al., 2014). And like TRV130, but not morphine, PZM21 did not recruit β-arrestin-2. Given these results, “the potential was that, at least in theory, PZM21 would be effective for suppressing pain sensations but might not have the side effects associated with morphine,” said Roth.
As predicted, subcutaneous injection of the drug into mice produced profound analgesia compared to saline, both in the hotplate test and after formalin was injected into the hindpaw (a model of inflammatory pain). In fact, PZM21’s effect in the hotplate test plateaued earlier than morphine and lasted longer than the effect of both morphine and TRV130. But in stark contrast to all known opioids, the researchers found no change in the tail-flick test with PZM21.
Why was there a difference in results between the hotplate and tail-flick tests? The authors thought that it might reflect where in the central nervous system PZM21 acted. Circuits in both the spinal cord and brain control the many behaviors seen in the hotplate test, but spinal reflexes account for responses in the tail-flick test. When the researchers categorized behaviors in the hotplate test as “reflexive” (e.g., retracting the paw) versus “affective” (e.g., jumping away), they found that PZM21 only changed affective responses, implying that it acted in the brain.
Importantly, PZM21 lacked several side effects common to opioids. While morphine depressed respiration, PZM21, as well as TRV130, looked comparable to saline in this regard. Reward circuits also appeared untouched, since PZM21 did not increase locomotion or result in conditioned place preference. These findings argue that PZM21 may be safer than morphine and other opioids, and may not be addictive. However, the authors reported that the drug caused constipation, another typical side effect of opioids, though less so than morphine.
Interestingly, biasing µ-opioid receptor signaling away from β-arrestin may not necessarily be entirely beneficial. In a recent study (Chen et al., 2016), mechanical allodynia in β-arrestin-2 knockout mice was extended by two days after DAMGO administration, relative to wild-type controls (opioid-induced hyperalgesia is a paradoxical effect of µ-opioid agonists). Furthermore, this work also found that the knockouts displayed prolonged pain in inflammatory and neuropathic pain models, through effects on NMDA receptor function. It remains to be tested whether a compound such as PZM21 that biases signaling away from β-arrestin may have similar adverse consequences.
Looking toward the future
As for the current work, Bruchas thinks there are two important extensions of the discovery of PZM21. “There’s the therapeutic one—having an opioid that doesn’t have as much addiction liability or doesn’t create respiratory depression is great,” he said. The second is scientific. “Now we have a biased ligand that we can use to understand how these properties and pathways work,” perhaps leading to even better drugs.
Another noteworthy aspect of the study, Bruchas said, is the “extensive collaborative network that made it possible, with multiple layers of expertise from different labs.”
He suggests, however, that to better assess PZM21’s addictive potential, future research should test if animals will self-administer the drug. “In a self-administration study, where the animals have control of how much they take, what does their behavior look like?”
There’s also the question of whether PZM21 will suppress pain as strongly after the first dose, which Roth intends to answer in future studies. “The prediction would be that animals would show less tolerance to PZM21, because β-arrestin is one of the main factors involved in tolerance, but we need to do the experiments.”
Matthew Soleiman is a science writer currently residing in Nashville, Tennessee. Follow him on Twitter @MatthewSoleiman
Image credit: maxxyustas/123RF Stock Photo