Skip to main content

Biological News

Opioid Signaling: Think Again

Opioids are some of the most effective and most problematic drugs in the entire pharmacopeia. For severe, intractable pain we really have nothing to match them, despite decades of searching for alternatives. The history of research for new pain medications itself calls for pain medication, because it is a tapestry of expensive late-stage clinical failures. All sorts of receptors, enzymes, ion channels and other targets have been proposed as pain targets (on the basis of perfectly reasonable biological hypotheses) and many of these appear to work in when targeted in animal models. But when you go into human trials, they collapse. It’s been apparent for decades that there are severe problems in the translational medicine of new analgesics, but we still haven’t been able to fix the problem, and such programs continue to evaporate in the clinic.

Perhaps drugs targeting the opioid receptors can be fixed? Decoupling pain relief from addiction will be extremely difficult (likely impossible?) as long as you’re hitting the opioid receptors themselves, thus the long search for other pain targets. But for some years now it’s been thought that opioid drugs themselves could possibly be made less dangerous. There are all sorts of side effects to hitting one of the main subtypes, the mu receptor, with the neuromuscular problems high on the list. It’s that respiratory depression that generally kills people in the case of an overdose, and pain patients also experience severe constipation due to the effects on intestinal smooth muscle.

And that takes us into the wilds of G-protein coupled receptor (GPCR) signaling. For at least ten years now, it’s been a popular hypothesis that the mu-opioid receptor might be splitting its effects via two different signaling pathways – the “standard” G-protein second-messenger one and the beta-arrestin pathway. That’s thought to be involved in receptor desensitization, among many other things, and there are a number of examples of differential signaling (and differentiation of side effects) depending on arrestin binding. There’s some evidence that you might be able to get analgesia without development of opioid tolerance and without smooth muscle effects if you could find ligands that activated the G-protein pathway and not the beta-arrestin one, and naturally enough there’s been a lot of work devoted to that idea. Compounds trying to put this idea to the test have recently made it far into clinical trials, but results have not been encouraging.

Comes now the cold water. A new paper demonstrates that in mice that have had their key beta-arrestin-2 protein completely knocked out that morphine and other opioid ligands still produce respiratory depression (and constipation, for that matter). This ties in with another recent paper (from some of the same authors) that looked at a series of mutant proteins that biased the mu-opioid receptor towards G-protein signaling and away from arrestin pathways, and that one showed that the heightened sensitivity to analgesia part might well be real, but that respiratory depression was still showing up and might even be worse.

So why hasn’t this experiment been run before? Well, that’s the thing: it has. In fact, the 2005 paper that really called attention to the possible therapeutic split in mu-opioid signaling was a demonstration in beta-arrestin-2 knockout mice (!) This latest paper, in fact, is a consortium across research teams in Germany, the UK, and Australia to re-examine this whole hypothesis due to all the conflicting results. They used the same knockout rodent line as in the 2005 work, but their results flatly contradict the earlier study, and they have no explanation for why this should be (although there’s a possibility that the earlier paper’s knockout animal strains were not sufficiently characterized). The paper also mentions a recent conference presentation from yet another group that has failed to reproduce the results as well. So I would put my money down on the idea that there’s something off with the 2005 work, and with the whole idea of being able to manipulate opioid effects via G-protein versus arrestin signaling.

Back, once again, to the damned drawing board. And back, once again, to the realization that we don’t understand receptor signaling anywhere as much as we need to, or as we sometimes might think we do. . .

47 comments on “Opioid Signaling: Think Again”

  1. tally ho says:

    Unfortunately, this seems to correlate with Trevena’s trials and tribulations in the clinic (TRVN $0.76). After the Nobel Prize in 2012 “for studies of G-protein-coupled receptors”, you’d think a deeper understanding of GPCR and beta-arrestin pharmacology would clear the path for next generation GPCR therapeutics. alas…

    1. John says:

      Is the TRV compound an allosteric agonist?

      1. bhip says:

        Partial agonist

  2. zzt says:

    Opioids remind me of good old cough syrup with codeine.

    But what do I see here ?Cough Syrup for parkinsons!

    What do folks here think of Ambroxol for preventing parkinsons ?

  3. Vader says:

    Opioids with fewer side effects would certainly be welcome.

    But they might have the perverse effect of increasing opioid addiction.

    1. Derek Lowe says:

      That occurred to me as well – and at even higher dosages than people hit now, perhaps. . .

    2. Isidore says:

      Good point! I was given opioids some 19 years ago when I hurt my back really bad. The singularly unpleasant experience with the constipation that followed has made me refuse them ever since, even after a couple of major procedures when, to the dismay and some irritation of the doctors, I made do with ibuprofen and acetaminophen. Doesn’t mean I would have developed an addiction if I had taken opioids again (I don’t know what the statistics are but I am guessing that a relatively small number of those who have been prescribed opioids become addicted to them) but you certainly can’t get addicted to something that you do not take. And I suspect that I am not the only one who would think twice about taking opioids if severe constipation is part of the price one has to pay.

      1. Dr. Manhattan says:

        I’ll second that. Had a bit of minor surgery back in the 1980’s and was given an opioid for pain. Far more pain and problems from the constipation than from the surgery!

    3. loupgarous says:

      “Opioids with fewer side effects would certainly be welcome. But they might have the perverse effect of increasing opioid addiction.”

      True, unless the side effects that bring people to take opioids recreationally is also abolished. My cancer pain has, since an invasive tumor in my stomach wall was surgically debulked, been manageable with tramadol (with an abuse potential that places it in Schedule IV of the Controlled Substances Act, not Schedule II as my prior pain meds were classified). Looking back, my memory’s now more reliable than it was on the Schedule II drugs, and so is my judgment. My surgical oncologist prescribed tramadol after he cut away the stomach tumor. Afterward, I weaned myself off of the Schedule II meds (with no prompting to do so from my pain control specialists), with no noticeable withdrawal symptoms. The lack of withdrawal symptoms may have been connected partially to high doses of gabapentin, also for the original cancer pain.

      Tramadol’s a known SSNRI. My experience, like my tumors, may be rare, but it makes me wonder how much of the human experience of pain is connected to things like circulating serotonin and norepineprine levels that aren’t measured (or measurable) by our current animal models of pain.

      1. Knowledgeable says:

        Gabapentin is known among users as an efficient medication to manage withdrawal symptoms. In your case, with Tramadol, the withdrawal from the monoamine reuptake inhibitor might have been worse than the opioid withdrawal, certainly after the first few days.

    4. aairfccha says:

      Maybe it’s time to bury the bogeyman of addiction then? Addiction to opioids is certainly not an ideal condition, but its problems pale in comparison to the carnage of the paradigm of enforced abstention (aka War on Drugs).

      1. John says:

        So, are you saying the epidemic of opioid addiction is “fake news”?

        1. navarro says:

          i don’t think they are calling large number of opioid addicts fake news so much as they may be pointing out that the pendulum may have swung towards undertreatment of severe pain.

          currently i am recovering from a bout with sepsis over christmas. one part of the infection colonized my left knee with which i was already suffering moderate to severe arthritis. i’m currently in out-patient physical therapy and had been seeing slow but steady improvements since my release from the hospital. about 10 days ago i began suffering a setback which has included a serious uptick in pain. all of the following is based on the 0-10 pain scale. where two weeks ago i had been experiencing an unmedicated pain level of 4 and a medicated (10 mg hydrocodone/325 mg acetaminophen) pain level of 2. currently i am experiencing an unmedicated pain level of 7-7.5 (almost at the level at which the pain is the only thing i am aware of) and a medicated pain level of 6. because of the laws of the state of texas my pain management doctor cannot allow me more than 4 tablets each day which means i spend a total of about 6 hours a day in a great deal of pain waiting to be able to take another dose.

          it might be possible for my doctor to prescribe a limited quantity of a more potent opioid to use for “breakthrough” pain if he is willing to go through the trouble of all the paperwork and potential for auditing. you might say that i am biased under the circumstances but i feel that my circumstances represent a case of undertreatment, looked at objectively.

          1. John says:

            I’m very sorry to hear of your case.
            Perhaps this is more a “political” than a scientific problem, as if your doctor was willing to take the risk you mention, it seems you’d have a palliative to your condition.
            You can also argue: why should a doctor put themselves at risk to treat a patient?

          2. aairfccha says:

            @John: Q. E. Duh.

          3. Wallace Grommet says:

            The variation in individual pain response to traumatic injury is mystifying. I crashed my motorcycle at 60 mph and suffered compound fractures of both my femur and tibia, was unconscious for five days, in traction, with a head swollen up like a black and blue pumpkin. After a lengthy operation to insert enough pins, plates, rods, and screws, I went home to recover and heal. At no time did I take any opioid medications. Somehow, perhaps because I was only 25, who knows? Not done yet! Three months later, a massive staph infection is taking my calf to dinner, I became delirious, feverish, and emergency surgery was performed on the infected calf, with almost half a liter of flesh debrided in a desperate attempt to stop the infection. I awakened hours after the surgery, and felt immeasurably better. The entire next month was spent in the hospital for wound care. Incredibly, despite having to close a pint size hole in my calf, I never needed any pain medication.

        2. aairfccha says:


        3. James Millar says:

          It is impossible to know what portion of the problems belongs to the drug, and what belongs to the enforcement – although the evidence is pretty clear that the “drug war” definitely makes things worse.

          What is almost certain is that the current boom in fentanyl and analogs is mostly due to enforcement. From the recreational user or abuser’s point of view, fentanyl really has nothing to recommend it versus heroin or a semisynthetic. The advantage is to the supplier. Laws skew things so that the cost (economic, risk) is almost entirely the transport. That’s fentanyl’s value. Orders of magnitude less product crossing borders.

          How many fentanyl deaths would still have occurred if the drug was heroin etc? Can’t know. but compouinds that strong are not easy to dose safely in a clandestine environment.

          1. loupgarous says:

            The problem gets worse. Carfentanil is beginning to show up in emergency rooms in the US, UK and Europe, very often in users dying or dead of overdoses, along with the tide of fentanyl. In 2016, Canadian customs intercepted a box labeled “printer accessories” filled with Laser Jet toner cartridges containing a kilogram of carfentanil, estimated to be enough to kill every resident of Canada.

            While the State of the Union message referred to agreements with China to reduce the flow of those highly potent opioid agonists, the Chinese president’s public attitude toward the trade in fentanyls to other countries lays the blame on our abuser population. Their comments refer to China’s own relative lack of opiate abusers, which may be as fictional as official Chinese denials of a coronavirus outbreak until recent weeks.

            A few years ago China was helpfully shipping THC analogues to the US which weren’t covered by the Controlled Substance Act then. We can probably rely on chemists somewhere in the world, if not China, to discover, mass-produce and sell our addicts new drugs with the same effects as the fentanyls currently sold on our streets (e.g., furanyl fentanyl and other fentanyls long imported and sold here as “China white”).

            We can’t win the War on Drugs because the goalposts keep moving in that game. When we succeed in making old-school heroin prohbitively expensive and troublesome to import and market here, there’s a whole alternate pharmacopeia of opioids, many of them available on prescription for very legitimate uses.

            State legislatures are beginning to strictly regulate drugs like gabapentin with very wide therapeutic indices, despite the fact that they are often useful in the treatment of intractable pain without opiates. Northern Ireland already has a problem with pregabalin abuse, which suggests to me that we’re in an arms race between the perverse ingenuity of substance abusers and the mentality of legislators and regulatory officals who have created a large, unwieldy compendium of banned and highly regulated drugs to no real avail.

            In the 1800s Johann Friedrich von Schiller wrote “Against stupidity, the gods themselves contend in vain”. He could say the same thing now just as accurately. It’s stupid to create a ceremonial chemistry (as Thomas Szasz called it) forbidding the ingestion of any of a growing list of chemicals and create an illicit demand for them. This has made billionaires of drug smugglers while we spend billions more on repressive measures against those citizens who violate our drug taboos. I don’t practice or endorse drug abuse, but I don’t see the logic in punishing it.

          2. aairfccha says:

            “It is impossible to know what portion of the problems belongs to the drug, and what belongs to the enforcement”

            On the contrary, it’s quite possible if you don’t freeze in fear of the addiction bogeyman (or are a useful idiot of the cartels). In general, the success stories of drug policies outside police states (opioid substitution/maintenance, needle exchange, drug checking, over-the-counter naloxone) are a violation of the abstentionist dogma.

            Second, the discussed studies above have demonstrated in several different contexts that the implementation of HAT [Heroin-assisted treatment] is feasible, effective, and safe as a therapeutic intervention. This should not be seen as a conclusion that could be taken for granted because many observers expected disastrous consequences from the provision of medical heroin prescription.

            The science is there, we just have to treat the war on drugs as the crime against humanity it is.

  4. Barry says:

    Splitting out the addiction and respiratory depression remain hard. But splitting out the constipation may be more tractable. Portoghese showed back in the 80s that by quaternizing the amine, you can build opiates that don’t get out of the gut. They agonize opioid receptors there w/o an systemic exposure/effects. It remains to be shown that one could build an opiate antagonist that similarly acts only in the gut. But it doesn’t look hard.

    1. AQR says:

      Three such compounds are currently on the market, naloxegol (Movantik), methylnaltrexone (Relistor) and naldemedine (Symproic).

    2. Skeptical says:

      So you’re saying we should dose patients with an opioid agonist for pain relief and a gut-specific opioid antagonist to prevent constipation? Interesting idea.

      1. loupgarous says:

        It’s happening. My pain doctor’s got ads in his exam rooms for a gut-specific mu-opioid antagonist. The ad copy specifically targets patients receiving opioid therapy who are experiencing intractable constipation. And I’ve seen direct-to-patient TV ads for the same medication.

  5. MoBio says:

    One wonders here if the problem lies in the background of the mice. In the original Bohn paper (JPET 314 (3) 1195-1201; 2005) the background strain is not specified. In the most recent paper they state:

    “β-Arrestin2 knockout mice were backcrossed over seven generations to wild-type control JAX™C57BL/6J (RRID:IMSR_JAX:000664) mice from Charles River Laboratories (DE), which were also used for breeding of the knockout strain and as controls in all experiments.”

    It is well known that genetic background matters quite a bit for opioid effects (see Anesthesiology. 2006 May;104(5):1054-62. for instance) and perhaps the answer here is a more trivial and perhaps interesting one.

    1. Jeff says:

      Yeah, they sure don’t make mice the way they used to. 😉

  6. Christophe Verlinde says:

    Reminds me of the following:

    Nature volume 537, pages185–190(2016). “Structure-based discovery of opioid analgesics with reduced side effects … Structure-based optimization yields PZM21—a potent Gi activator with exceptional selectivity for μOR and minimal β-arrestin-2 recruitment. Unlike morphine, PZM21 is more efficacious for the affective component of analgesia versus the reflexive component and is devoid of both respiratory depression and morphine-like reinforcing activity in mice at equi-analgesic doses.”

    Br J Pharmacol. 2018 Jul; 175(13): 2653–2661. “The novel μ‐opioid receptor agonist PZM21 depresses respiration and induces tolerance to antinociception … Contrary to a previous report, PZM21 depresses respiration in a manner similar to morphine, the classical opioid receptor agonist.”

  7. luysii says:

    It’s time for some clinical experience with many people receiving opiates, which is not to denigrate in any way the individual experiences described here.

    Back in the day, disc surgery required general anesthesia, dissection of the back muscles down to the spine, sometimes chipping away at the bones of the spine to remove a bone spur (osteophyte) and/or removal of the offending herniated intervertebral disc. This meant a hospital stay (unlike my ophthalmologist who had a microdiscectomy as an outpatient a few years ago). This was the era of the discovery of the protein receptor for morphine and other opiates, and we were all hopeful that this would lead to the development of a nonAddicting opiate (narcotic). Spoiler alert — it hasn’t happened and likely won’t.

    Often, I was the neurologist who diagnosed the disc and told the surgeon where it was likely to be found (this was in the preCT and later the preMRI era). I’d developed a relationship with most of those I’d referred for surgery (since it was never recommended, without a trial of rest — unless there were compelling reasons not to — trouble controlling bowels and bladder, progressive weakness etc. etc.). I was their doc while they tried to heal on their own.

    So post-operatively I’d always stop by to see how the surgery had worked for them. All were on a narcotic (usually Demerol back then) as even if the source of their preoperative pain had been relieved, just getting to the problem had to cause significant pain (see above).

    If the original pain was much improved (as it usually was), I’d ask them how they liked the way the demerol made them feel. There were two types of responses.

    #1 I hate feeling like this. I don’t care about anything. I’m just floating, and feel rather dopey. I’m used to being in control.

    #2 I love it ! ! ! ! I don’t have a care in the world. All my troubles are a million miles away as I just float along.

    Love it or hate it, both groups are describing the same feeling. Neuropharmacology can help to tell us why opiates produce this feeling, but it can’t tell us why some like it (about 5%) and the majority (95%) do not. This clearly is the province of psychology and psychiatry. It’s the Cartesian dualism between flesh (opiate receptor) and spirit (whether you like what it does). It also shows the limitation of purely physical reductionism of the way we react to physical events.

    For more see —

  8. Scott says:

    The news media leaves out the side of the persons with a disease or syndrome that involves high levels of pain for an extended period of time. I have had RSD Syndrome five times. The second time was the worst – 6 years. Every 6 months they had to increase the milligrams of the pain med I was taking. My average pain level was 7. I did hit a 10 a number of times. I never got addicted to them. I in fact have had ten surgeries in my lifetime. For people like myself, they are needed. Of course, anyone that is terminal, they definitely should not be restricted.

  9. zuzusaurus says:

    The problem is the over interpretation of the translatability of the original work. The original Science (1999) and Nature (2000) papers are beautiful and elegant. Yes, Morphine clearly has differential effects on OR signaling WRT B-arr2 KO vs Litter mates.

    But that is where the story should have ended for 3 reasons.

    1) The antinociceptive B-Arr2 KO effect does not translate to other opioids. See Neuropharmacology. 2011 Jan; 60(1): 58–65. Oxycodone, methadone and fentanyl do not show the same effects as morphine.
    2) When looking at Signal Bias writ large, morphine is vastly the outlier wrt non-addictive endogeneous opioids. See Molecular Pharmacology, 2015, August 2015; 88, 335-346. Or Figure 2 of Trends in Pharmacological Sciences, November 2015, 36(11) 705-706.
    3) Did you really expect that B-Arr2 KO wouldn’t cause havoc among all the other GCPRs that use it as a regulator? Likely some compensatory mechanism.

    When talking signal bias… it is probe dependent. The Opioid B-arr2 research is a story about Morphine… nothing more, nothing less. Endogeneous ligands are not addictive… why? No idea. But it probably has something to do with Ligand Bias and / or Spatio-temporal nature of the signaling.

    1. gpcrMedChemist says:

      Excellent comment. Thank you for your points and perspective.

      1. Dr CNS says:

        Agreed. Very good point.
        So, considering the many known endogenous peptidic ligands for the mu opioid receptor, how do you establish a preclinical translational strategy to start understanding the in vivo effects of a drug candidate?

  10. Emjeff says:

    Error of recency. Why are all the old data being thrown out on the basis of a single study?

    1. Jolt says:

      Maybe because the recent data looks stronger:
      1) Multiple backcrossed WT control
      2) Multisite replication

      1. MrXYZ says:

        Were all sites using the same source of animals?

        As MoBio mentions above, the genetic background of the mice matters.

        1. Jolt says:

          Apparently two labs used the same source while the third had a different one.

  11. metaphysician says:

    So, two questions come to my mind, both recurring when the topic of opioids and pain control come up:

    1. I have read elsewhere that there is some evidence for addiction being connected to preexisting psychological issues. Essentially, an opioid used to treat “physical” pain is unlikely to addict, but opioids don’t distinguish between physical and psychological “pain”, and if they treat the latter? Addiction. Is this true ( or at least something which has supporting evidence )?

    If so, does this suggest that patients with pain issues need to be screened for psychiatric illness? And what do you do about comorbidity? Traumatic injury, severe illness, or chronic pain can all cause depression and other psychological problems, after all.

    2. Consider this hypothesis: “Pain and addiction are inextricably linked on a mechanistic level. Anything which acts upon pain levels in the central nervous system will have the potential to induce addiction, because the trigger for addiction is the reduction in central pain signaling”. This seems at least a plausible hypothesis, based on my admittedly layman’s knowledge ( addiction being a malfunction in motivational behavior, and pain/pleasure being the feedback loop for motivation ).

    First, is there any evidence out there that would tend to *disprove* this hypothesis? And second, if this hypothesis is true: what the hell do we *do*, ethically? Sure, there might be a lot of diseases and ailments where treatment “downstream” of the pain is the best option, but there definitely are also ailments where this might be impossible ( like anything which exists only within the central nervous system ).

    1. MrRogers says:

      Consider ibuprofen and acetaminophen. Both are effective (if weak) analgesics. Neither has any addictive potential.

      1. loupgarous says:

        A contrast and comparison between the opiates, ibuprofen and acetaminphen suggests that ibuprofen reduces pain by inhibiting COX and inflammation outside the CNS. It’s possible acetaminophen acts either by reducing COX in the brain (COX must be oxidized to work) and/or by a major metabolite inhibiting reuptake of a major endocannabinoid, anandamide and activating the vanilloid receptor.

        Drugs which act on the endocannabinoid system (FAAH inhibitors that block degradation of anandamide) have gone into clinical trials for various indications, including pain relief. One of them, BIA 10-2474, put five patients in the hospital during a healthy volunteer trial in France, killing one of them and causing permanent brain damage to another. Other FAAH inhibitor drugs in human trials before and after then haven’t reported serious adverse effects.

        It’s possible that the future of pain relief involves highly selective inhibition of FAAH to increase higher levels in the brain of anandamide with consequent reduction of pain and anxiety without addiction, respiratory depression or other adverse opioid effects. But the cautionary tale of BIA 10-2474 seems to have slowed that research down.

        1. exGlaxoid says:

          The BIA drug trial was a diaster, and done badly, obviously. The most apparent cause of the side effects was off-target activity, which the BIA drug had a large number non-selective activities.

          Sadly, the BIA trial set back FAAH and MAGL (a related target) drug research and trials by years or maybe forever, as the other compounds tested were much cleaned in action, more potent (so less off target potential), and some appeared to have promise. But they have almost all been dropped. I was working inthat area when it happened, and that trial killed all work for us.

          There is clear evidence (the women with no pain articles and more) that they have promise, but once an area is tainted by any risk, the likelyhood of future research becomes much lower. I hate to see that, as I think many less-than-ideal drugs could have a second or third generation that is much better and effective, but we might never know. Examples include the glitizones for diabetes/ALZ, COX2 inhibitors, and many others. Another examples was the long time it took thalidomide analogs to be used for cancers due to the initial horrors it caused in birth defects. So we need to be careful both in putting drugs into the clinic, but also more careful in pulling them and never allowing them again, even for cases where they have a clear benefit.

  12. scripps guy says:

    The problem with the failed Trevena clinical compound is that it isn’t biased enough to give a respiratory benefit. Some with most of the “biased ligands” people have touted. That is spoiling the whole field, plus the fact that pharma is scared to death of the opioid mechanism and of CNS drugs in general.

    Some of the highly biased Scripps Bohn/Bannister compounds clearly show full analgesia with NO statistically significant respiratory depression on overdose, and have no effects at all in mu opioid knockout mice. There is something there. The animal data is compelling

    1. XXopolis says:

      The Trevena compound is far from failed. Scripps guy, your bias is showing – you should have disclosed the numerous biased ligands you have at your institute.

      I would just point out a couple of clinical realities for those of you tied to the bench, dreaming of doing clinical research. It turns out that measuring respiratory depression in human patients is not easy. No one can agree on what the “best measure is, and if you ask the FDA, they will shrug their shoulders. So, it appears that Trevena tried to do the only thing really open to them – use an endpoint that clinicians use. Although it was not clinically significant, it went the way you would predict. I suspect that if Trevena used a more quantitative endpoint, the results would look much better. But, what that endpoint actually is, no one knows…

  13. Peripheral kappa guy says:

    Peripheral Kappa opioid receptors are known to exist on neural extremities. Cara Therapeutics is ardently pursuing this approach (I have no connection with them). Kappa opiate agonists which do not cross the blood/brain barrier. We had a project team working on this at the pharma company that employed me way back in the mid 80’s. The commercial guys insisted on oral bioavailability. The k-selective compounds were zwitterions and had very poor oral bioavailability. The project was terminated.

  14. Jolt says:

    Outside of the opioid field, what is the strongest evidence that the biased agonism approach may ever work?

    1. Derek Lowe says:

      There are a couple of good papers that have just appeared about the angiotensin II receptor that address this – I’ll be blogging about them next week. . .

    2. Patrick says:

      It certainly works in the case of psychedelics (5-HT[2] agonists) – that’s how they produce their effects, while the usual 5-HT (serotonin) doesn’t.

Comments are closed.