Pharmacology REPOSTED: 3DQSAR of Opiate Receptor Ligands

[4DQSAR]

'Insights into subtype selectivity of opioid agonists by ligand-based and structure-based methods'

Sci-Hub: are you are robot?

J Mol Model (2011) 17:477–493
DOI 10.1007/s00894-010-0745-1

I feel I should apologise because I have been referring people to a foundational paper and only just noted that the link is no longer valid.

I'm sure most people have noted that the vast majority of high affinity MOR ligands have two aromatic rings and that from 'tip to tail' the two rings (inclusive) are roughtly 15 methylenes equivelent (10-14Å) apart. With BDPC it's 14 hence the para bromo, with etonitazene it's 13 hence the para ethoxy. R-4066 and even nitrometofoline also fulfil that 'rule' so it's more common that is commonly noted.

This concept is (currently) termed the 'twin site receptor model'.

I don't study opioids because I ever intend to harm others, I study them because after a century, we are still unable to find all of the key moieties and as @fastandbulbous noted 'design of opioids uses black magic'. It's a field where individuals who concentate on a key facet may uncover something of value to others.

 

[4DQSAR]

Sorry, yesterday I was too tired to provide reference to the 'twin site receptor model'

https://www.sciencedirect.com/science/article/pii/0024320586902171

I suppose the one example that at least some BLers may recognize is the subjective difference between codeine and dihydrocodeine. Although DHC is active in it's own right, most of it's activity of both is mediated by their O-demethylation into morphine and dihydromorphine.

I don't know anyone who hasn't noted the significant difference between the two and you will note that DHM has significantly higher affinity than morphine for the receptor subtype labelled μ₂.

Years ago I noted that thiofentanyl and furanylethylfentanyl (two analogues that on the face of it would be attractive to illicit producers) were conspicuous by their absence. In that case I posit the reason is that while both demonstrate more potent analgesic activity in animal models, that extra analgesia isn't mediated by increased μ₂ activity.

I use the naming used by the paper although I strongly suspect they have mixed up μ₁ and μ₂ as it's the latter that is responsible for euphoria and respiratory collapse. I suspect that is why fentanyl (an opioid with a TI of around 270) is still a massive killer. It's biased towards μ₁ activity so only becomes euphoric at doses close to it's toxic limit.

I've said it before but I believe that an opioid may be potent or an opioid may be euphoric, just not both. Wherever I look it seems that compounds estimated to have around 10 times the potency of morphine are the most euphoric and the reasons are:

1)Relatively low LogP so they permeate the blood-brain barrier quickly (i.e. produce a flash)
2)Relatively short mean receptor occupancy time and the ongoing euphoria of an opioid is medated by the binding-unbinding cycle
3)Relatively short overall duration of action so a user's body has essentially metabolized almost all of the previous dose

Yes, WSB is quite correct in noting 'all pleasure is relief' but I suggest some opioids are pleasurable for their ability to remove all psychological stressors producing mental detachment while others simply provide relief from the AWS. Hence my noting that in non-dependent users, fentanyl (for example) produces only a mild 'plastic euphoria'.

 

[vima]

this is extremely interesting; thanks for posting!

 

[4DQSAR]

Sci-Hub: are you are robot?

The phenylmorphans are not facile targets but I think it's interesting that neither of the two foundational books on opioids cover them since they are a relatively recently discovered class. But researchers did apply rational design to uncover some extremely potent examples.

Sci-Hub: are you are robot?

I admit, even I was amazed to discover that the patent application for the phenylmorphans was made as recently as 2019. Specifically:

https://patents.google.com/patent/US20210024532A1 - Biased potent opioid-like agonists as improved medications to treat chronic and acute pain and methods of using the same.

Note that example 52 (Ki 0.13±0.02 - full agonist) conforms to the 15 methylene equivelents.

Due to the sheer number and variety of experiments to identify the MOR affinity of fentanyl, the range is HUGE but most modern researchers apply standardized tests using recombinant human receptors yield an affinity of roughly 1.35 nM.

Just for reference, the MOR affinity of morphine using the same standardized testing yields a value of 14 nM i.e. affinity does not have a linear relationship to potency for manifold reasons but it still of value when guessing the rough range.

The patent does explicitly provide the efficacy of each example so while example 52 is listed as 100% i.e. full agonist, example (+)38 is a partial agonist, specifically 91%.

It appears that the term 'biased' may have a large overlap with the term 'partial agonist'. Sadly in the patent doesn't explicitly seperate affinity and activity for the μ₁ & μ₂ subtypes but it does pose the question that a ligand that is biased towards μ₁ affinity may present as a partial agonist.

Certainly that would explain why research ligands such as SR-17018 failed so spectacularly as an analgesic. It simply could be the case that when only beta-arrestin recruitment is studied, nobody is asking the deeper question i.e. are such ligands μ₁ selective?

BTW the phenylmorphan class is a subset of the phenanthracine (morphine) class of opioid. In this case the A-ring, C-Ring and D-ring. That chiral hydroxy overlays the 14-position of the phenanthracines.

 

[4DQSAR]

Whoops - linked to the paywalled paper of the potent example. Now edited so everyone can read it.

 

[4DQSAR]

OK - I was not aware of the fact since my education was in the late 80s-early 90s BUT my noting that ALL highly potent MOR ligands are roughly 15 methylenes 'long' has not one but two formal names:

The aromatic-aromatic vector distance AKA bi-aromatic pharmacore envelope.

So we can now firmly state that any candidate in which the AAVD is >8Å will demostrate lower or no activity. That number is from the electronic centre of the A-aromatic to the electronic centre of the B-aromatic.

That's why etonitazene has a para ethoxy moiety - to move the centre of the aromatic. Or why BDPC has a para bromo moiety.

Wth flexible ligands such as fentanyl or phenoperidine, it's entirely possible that the distance isn't perfect but can be 'tuned'. For example, a para '-F added so some fentanyl derivatives allows them to bind in a lower energy state.

Even tianeptine follows that rule but the unusual bit is rather than a second aromatic, it has a long alkyl chain terminating in a carboxylic acid. Usually such a long chain of rotatable bonds would render such a ligand inactive. But the carboxylic acid acts as a sort of 'velcro' i.e. as the ligand changes conformation as small amounts of energy are imputted out outputted, as it passes through the active conformation, that carboxylic acid forms a VERY strong hydrogen-bond.

It's nice because I felt I was close to finding a set of rules then someone discovered tianeptine and I had no answer as to why such a 'floppy' ligand would be active.

 

[AnIrregularApocalypse]

Is μ₁ what we now call "g-protein", and μ₂ what we now call "b-arrestin"?

 

[4DQSAR]

Is μ₁ what we now call "g-protein", and μ₂ what we now call "b-arrestin"?

Role of β-Arrestin 2 in the antinociceptive and side effect profile of morphine and the novel mu opioid receptor agonists, kurkinorin and kurkinol - Neuropsychopharmacology

The development of safer mu opioid receptor (MOR) agonists with reduced side effects is a key focus of pain research. Some studies have suggested that MOR agonists with reduced β-arrestin 2 (βArr2) signaling (i.e. G-protein biased agonists) may have greater therapeutic windows. However, there...

www.nature.com

It's pretty 'bleeding edge' as it appears BOTH mediate analgesia hence highly-biased ligands not being particularly active in that department and while the makers of oliceridine (Olinvyk™) claiming a HUGE TI, the protocol THEY supply for it's use does suggest a narrow theraputic window and the list of side-effects is the same as 'classic' opioids - tolerence being the sole exception.

We know from SR-17018 that once the slow-onset is bypassed, it can be abused just as much as a 'classic' opioid.

So this whole concept of biased ligands is now termed a theory which does suggest that the predicted activity doesn't match actual human data.

This is always the thing with studying opioids. You will find axioms used for decades being discarded.

A classic is Nathen B. Eddy at Hoffman la Roche taking the simple morphinan scaffold and systematically just trying every N-substitution possible. But the analgesic activity in an animal model tells us almost nothing about the subjective effects in man. In the case of Ro4-1539 (and most of the orphines), the analgesic activity is mostly mediated by other classes of opiate receptor. But makers of such products don't seem to recognize that issue.

In man I think the best example I can point to is that while hundreds if not thousands of fentanyl homologues have turned up on the street, the N-2-phenylethyl moiety is never substituted for another aromatic. The reason is the same - Sure, the N-2-(2-thienyl)ethyl and moreso the N-2-(2-furanyl)ethyl and indeed some other heterocycles all produced FAR more potent analgesics in animal models... it is modtly mediated by other classes of receptor.

KOR ligands are dysphoric, DOP ligands are pro-convulsant and NOP ligands seem to produce all of the side-effects of 'classic opioids' EXCEPT for the euphoria. So ironically NOP ligands might be the next class because nobody would ever take them for fun.

But tomorrow, who knows? Another axiom may change and off we go again.

I think that is why studying opioid ligands is interesting - there isn't a clean QSAR because some moieties are incompatible e.g. adding a 14-OH to the morphinan scaffold increases actvity BUT then swapping the N-methyl for an N-2-phenylethyl only increases potency slightly (x1.5 compared to x8 if the 14-OH is absent).

Everyone who has tried to draw a QSAR of a single class has since had their work disproven hence any sort of PiHKAL for opioids would be incredibly dangerous and incredibly inaccurate unless someone is REALLY prepared to syntheize hundreds of examples just to have a few from each class... I doubt the sales would cover the costs of the research.