Biased Agonism: The End Of The Opioid Crisis Or A Pathway To More Pain?
Whether it’s a killer headache after a night out or a particularly nasty cold, everyone has experienced the need for pain relief at some point in their life. Whilst the likes of ibuprofen and paracetamol usually suffice, in the arena of severe, postoperative, or chronic pain, no class of drugs can match the opioids for analgesic efficacy.
Whilst opioids are among humanity’s most essential medications, it is difficult to reconcile their importance in the clinic with the havoc that is wreaked through their illegal consumption on American streets. Overdose has now eclipsed car accidents to become the number one cause of accidental death in the United States, which is in no small part due to Purdue Pharma and the Sackler family’s notorious advertising campaign to drive OxyContin into the everyday pain niche.
There is a risk, however, of throwing the proverbial baby out with the bathwater. Relief from pain is a fundamental therapeutic right, and an over-reluctance to prescribe opioids, even when highly recommendable, is problematic too. This is the case in Italy and a handful of other European countries, where an acknowledgement that NSAIDs can be used for certain acute conditions has led to a slightly oversimplified one-size-fits-all consensus against using opioids to treat any acute pain conditions.
Well, according to some research groups, it can. In 2020, the FDA approved a novel µ-opioid receptor-targeting drug called Oliceridine. The developers of the drug, Trevena, argue that it is the first ever pain medication to successfully leverage the phenomenon of biased agonism for therapeutic gain. Biased agonists are drugs that bind to a target receptor and selectively activate specific signalling pathways downstream of that receptor preferentially over others, by changing the shape of the receptor to favour the binding of one of the next proteins in the pathway over another. While other potential drugs at various stages of development claim to have achieved the same thing, in the past three years, the claims of these industry-backed groups have found themselves on increasingly unstable ground. The current data suggests that we may have some way to go before we can truly achieve the level of targeted drug design afforded by biased agonism.
The Promise of Biased Agonism
Our current model of opioid signalling at the µ-opioid receptor (broadly based on gene knockout experiments in mice) describes a system where both the analgesic and respiratory depressive effects of opioids are downstream of the µ-opioid receptor. However, since they are transduced through two different pathways: G proteins and β-arrestin 2, respectively, a drug that, when bound at the µ opioid receptor, biases the system towards G protein signalling, would, in theory, offer the pain relief of the opioids without the potentially lethal respiratory depression.
This is exactly what Trevena claim to have achieved with oliceridine. Indeed, initial testing illustrated less respiratory depression than morphine at equivalent doses, while phase II human studies indicated a statistically significant difference in the prevalence of nausea, vomiting, and respiratory effects between oliceridine and morphine. The FDA even granted the drug breakthrough status, and a handful of other biased lead compounds at earlier stages of development have shown even more dramatic and promising side bias levels and side effect profiles. So why aren’t all these drugs being fast-tracked into trials, the clinic, and even harm reduction programmes on the street so we can start phasing out dangerous clinical opioids and sparing patients the nausea, constipation, and threat of respiratory failure?
A Biased or Partial Perspective?
The problem is that accurately quantifying bias in a lab setting is incredibly difficult. Biased agonist sceptics have detailed how simple, low-efficacy (weaker) opioids could give the impression of bias in the wrong test. For example, the G protein signalling pathway is complex and massively amplified at each stage of the pathway, so that if a downstream step is used as the proxy for G protein pathway activation, even low-efficacy drugs might saturate downstream signalling and achieve a large response in tests for G protein activation. Conversely, β-arrestin 2 is simpler and more one-to-one in its downstream effects, meaning that a low-efficacy agonist might elicit a response too small for an insensitive test to measure. This would give the appearance of a profound bias between the pathways that doesn’t accurately reflect how the drug would act in vivo.
If these tests for bias are indeed misleading, the drugs may be merely low-efficacy agonists, which are neither a new technology nor are they able to offer the same possible upside as biased agonists. In vivo, partial agonism often manifests with slightly improved side effect profiles (e.g., buprenorphine and methadone); however, they retain some side effects and do not generally generate the same degree of pain relief. A reexamination of the pharmacological profiles of the three most promising G protein-biased agonists (oliceridine, PZM21, and SR-17018) with mitigation for test sensitivity and other previously confounding effects did indeed find all three to in fact be low-efficacy drugs with no statistically significant bias (though this was of course contested by the developers of these drugs).
With the debate surrounding pharmacological analyses, perhaps the most serious blow for biased agonist proponents and developers is the recent interrogation of β-arrestin 2’s role in µ-opioid receptor signalling. The value of bias in this system relies on separating the G protein and β-arrestin 2 pathways downstream of the receptor; however, if both analgesic and adverse effects result from the same pathway, this no longer becomes desirable. Three independent laboratories repeated the original genetic experiments that implicated β-arrestin 2 in opioid side effects, and all three demonstrated dose-dependent constipation and respiratory depression in mice with the β-arrestin 2 gene knocked out.
The presence of opioid-induced side effects supposedly downstream of β-arrestin 2 in mice that don’t actually have any β-arrestin 2 is certainly a serious blow for those supporting biased agonism. That said, it’s not quite a knock out. Deleting a ubiquitous and highly integrated signalling protein such as β-arrestin 2 could have all sorts of bizarre effects. Ironically then, it was the elegant knock-in genetics of Kliewer et al. that delivered the final blow for the biased agonist theory at the µ-opioid receptor. Kliewer and co. created a mice line with a mutated µ-opioid receptor incapable of recruiting β-arrestin 2 but functional with respect to G protein binding. All of the knock-in mice displayed profound respiratory depression and constipation, with the correlation between the potency values for morphine-mediated analgesia and respiratory depression being highly significant. This work strongly suggests that β-arrestin signalling is not actually responsible for canonical opioid side effects but rather that both the analgesic and adverse effects result from the G protein pathway.
Looking forward
Where does this leave the so-called biased agonists currently in development and the theory as a whole? Well, regardless of the mechanism (low-efficacy rather than biased agonism), opioids with improved side effect profiles may well have clinical value. However, given the current evidence, it seems that true biased agonism is not a strategy that is likely to yield safer opioid pain medication. As is so often the case in the field of drug design, we simply do not have a complete enough picture of the underlying signalling networks to achieve the level of targeting that biased agonism could afford. There isn’t much value in trying to find drugs that are biased towards a certain pathway when we don’t truly understand how that pathway relates to those around it or to the physiological effects of a drug. Biased agonism is still a phenomenon of immense potential and may well soon open the doors to highly specific novel drugs at new and previously untouchable receptor targets, but in the case of the µ-opioid receptor, this looks to have been a false dawn.
