Synthetic cannabinoid kidney damage speculation

[niflheim]

I spent some time yesterday doing some research into the reports of kidney damage associated with synthetic cannabinoids and trying to find a plausible mechanism that would explain the known (or really, given the dearth of actual information, likely) facts. My assumptions are:

1. Some synthetic cannabinoids, especially XLR-11, but possibly others, cause acute kidney injury associated with renal tubular necrosis.
2. The metabolism of synthetic cannabinoids is complicated, not fully understood and may involve pyrolysis products depending on the route of admission
3. Some synthetic cannabinoids that are structurally similar to XLR-11 do not appear connected to acute kidney injury, including the newer ADB series.
4. Toxicity associated with the use extremely potent cannabinoids (e.g. MMB-CHMINACA and variants) will be difficult to separate out from overdoses.

The structures of many synthetic cannabinoids are similar to non-steroidal anti-inflammatory drugs and there are known interactions between naturally occurring cannabinoids and NSAIDS (Decrease in efficacy and potency of nonsteroidal anti-inflammatory drugs by chronic delta(9)-tetrahydrocannabinol administration), indicating that cannabinoids (synthetics included) are likely to have effects on the cyclooxygenase system. NSAIDS work by inhibiting the COX-1 and COX-2 enzymes; AM-404, a metabolite of acetaminophen is an endocannabinoid reuptake inhibitor as well as inhibiting COX-1 and 2. In short, it's complicated.

It struck me that the kidney damage associated with XLR-11 seems quite similar to analgesic nephropathy. This is associated with NSAIDs, but occurs much less frequently since the withdrawal of phenacetin, which functions as a pro-drug for acetaminophen. While being pretty nasty by itself (chronic use is linked to renal/pelvic cancers and cardiovascular disease), phenacetin was heavily implicated in analgesic nephropathy when used in combination with other NSAIDs. However, phenacetin itself is quickly metabolised into acetamonophen and no direct mechanism for it's kidney-damaging effects (which are much worse than any associated with acetaminophen) has been found but it seems likely to be a minor metabolite.

This is extremely specullative, but one (possibly) plausible mechanism may be:

1. XLR-11 and other synthetic cannabinoids are likely to inhibit COX-1 and COX-2. This reduces prostaglandin production in the kidneys which reduces blood-flow in the same way that (most) NSAIDs do, which is generally not problematic.
2. A phenacetin-like metabolite (possibly the unidentified, kidney-damaging minor metabolite of phenacetin) is produced via P450 metabolism in the liver. In the liver itself, this undergoes glucuronidation and is harmlessly excreted
3. Another possible metabolite, 2,2,3,3-tetramethylcyclopropanecarboxylic acid functions as a potent analogue of valroic acid, which is known to inhibit glucuronidation in some complex ways.
4. With inhibitied glucuronidation of the phenacetin-like metabolite in the kidney, which is already succeptible to oxidative stress due to the COX-2 inhibition causing reduced blood flow, damage occurs.

If correct, this would imply that kidney damage is possible with other cannabinoids and renal cancers may be a risk with long term use. However, the production of potent glucuronidation-inhibiting metabolites turns the chronic risks into an acute danger. If similar dangers are confirmed for ADB series cannabinoids, then a similar mechanism could be involved. I can't find anything on the equivalent metabolite however. It would also suggest that co-administrating drugs (such as valproic acid itself) while using synthetic cannabinoids may have serious health implications.

I'm not a chemist and have no medical background and even if I was this is all based on looking up stuff on the internet rather than anything real, but hopefully I haven't made any truly stupid mistakes in the above (feel free to point them out, though!). This is also my first post here, so let me know if I've accidentally broken any rules or similar.

 

[niflheim]

Some further thoughts on syntetic cannabinoid safety.

The current generation of synthetic cannabinoids are very close to a number of pharmaceuticals, known as 'setron' drugs. These are 5-ht3 antagonists used mainly as anti-emetics following operations and during chemotherapy. A handful of them are so close that, although they lack cannabinoid activity (due to their alkane tail's being too short or non-existent), they can easily be described using the abbreviations that have evolved for the cannabinoids. Although I'm not aware of any research linking cannabinoids to ht3 antagonism, it seems likelu that some or all of the current set do have this property as vomiting and diarrhea are frequently reported as withdrawal symptoms and this is exactly what you'd expect with abrupt cessation of an ht3 antagonist (i.e. it's rebound nausea and increased intestinal motility). This means that it should be possible to speculate on other likely features drawing from the extensive research on these close pharmacological cousins. First, however, some details.

Granisetron is a highly-specific 5-ht3 antagonist that is closely related to APINACA (AKB-48). Both are indazole-3-carboxadmides and differ only with the 'head' and 'tail'. Swap APINACA's amantidine moiety for a tropane and reduce the pentyl tail to a single methyl. Both tropane and amantidine are cyclic nitrogen-containing structures of a fairly similar size, so the difference isn't likely to be high. I'd be willing to bet that if you extended Granisetron's carbon-chain to match APINACA's, it would gain CB activity. You can call Granisetron Tr-MINACA. Ricasetron is similar, with an indole (rather than indazole) core placing it closer to APICA than APINACA. It also has a dimethyl on the indole.

Dolasetron is more similar to QUPIC (PB-22) being a carboxylate ester. Dolasetron has no carbon chain tain and instead of PB-22's quinolinyl head, Dolasetron features something quite similar to amantadine, but with the addition of a ketone. It's also attached via a carbon rather than the nitrogen common to the amantadine-containing cannabinoids. QUPIC steands for QUinolinyl Pentyl Indole Carboxylate, so we could call KADesIC (i.e. Keto-Amantadine Desmethyl Indole Caboxylate). Tropisetron is a cross between Ricasetron (as it has an indole without the extra dimethyl) and Dolasetron (as it's a carboxylate ester).

There's other Setron drugs which are a little further away from the current generation of cannabinoids. For example, some constraining the nitrogen of the carboxamide bridge in a third ring attached to the indole or indazole, which doesn't appear to be a feature of any current cannabinoids, although a recent JWH-018 analogue EG-018 does have a three-ring structure which suggests that this is something we'll probably see in the future. Granisetron and Dolastetron are both well researched, so at this stage, they're probably most useful to consider.

The 5-HT3 antagonists are reasonably side-effect free, however the small number of cannabinoid deaths do seem similar to some of the rare side-effects that can occur with both Dolasteron and Granisetron. Specifically both drugs are rarely associated with long QT syndrome [http://en.wikipedia.org/wiki/Long_QT_syndrome] (see 5-HT3 receptor antagonists and ECG effects [http://www.clevelandclinicmeded.com/medicalpubs/pharmacy/novdec2002/5ht.htm]). Long QT syndrome can lead to a potential fatal form of tachycardia called Torsades Des Pointes [http://en.wikipedia.org/wiki/Torsades_de_pointes]. The linked pharmacotherapy update does say that Torsades Des Pointes has not been found in relation to 5-HT3 antagnosists; however the update is from 2003 and that is no longer the case [http://www.healio.com/cardiology/ar...ial-of-5-ht3-antagonists-a-cause-for-concern], with changes to approved dosages/RoAs being made for Ondansetron and Dolasetron, and warnings being added to boxes. It's also worth noting that the likelihood of adverse events in the context of a carefully monitored and well-advised patient taking prescribed doses for time-limited periods is very different to the context of a recreational user without monitoring and possibly without any real control over dose. i.e. The latter activity is inherently a lot more risky and cardiac events may occur much more frequently under those conditions. There's also more potential for interactions due to the broader effect profile (i.e. the Setron drugs don't act on cannabinoid receptors and may also lack the likely NSAID effects of the cannabinoids).

There's also a possibility that some (or all) of the synthetic cannabinoids have effects at receptors beyond CB1/2 and 5-HT3. While most Setron drugs are fairly selective for 5-HT3, there is evience of some activity at other receptors for certain ones:

"For example, ondansetron has detectable binding to 5-HT1B, 5-HT1C, α1-adrenergic, and μ-opioid receptor sites. Although not proven, the binding of these agents to additional receptor subtypes other than their target receptor may underlie the inferior adverse-event profile seen with ondansetron compared with granisetron" [http://theoncologist.alphamedpress.org/content/9/6/673.full.

One 5-HT3 antagonist was also withdrawn following reports of serious (and in some cases fatal) gastrointestinal side-effects relating to slowed intestinal transit time [http://en.wikipedia.org/wiki/Alosetron#Adverse_effects].

Some people are genetically predisposed to long QT syndrome due to certain gene mutations; this may be completely unproblematic, unknown and undiagnosed up until the point that it suddenly becomes a problem - this may be a plausible explanation for the handful of reports of sudden death following cannabinoid use. It's also unpredictable (multiple short episodes of Torsades des Points can occur without incident before an episode escalates into potentially-fatal tachycardia.

Taking Cannabinoids alongside medications or other drugs that are known to cause Long QT syndrome will also increase the likelihood of experiencing episodes of Torsades des Pointes with a concomitant increase in the risk of sudden death. A useful list of drugs that can cause long QT syndrome is available here [http://www.sads.org.uk/drugs_to_avoid.htm]. Some common ones (well, common for people likely to be reading this forum anyway) are:

Stimulants - amphetamines (Adderall, Vyvanse, etc.), phenidates (Ritalin, Concerta, ethylphenidate, etc.), atomexetime (Strattera), phentermine, ephedrine and the various substitues commonly used in cougha, cold and flu medications and so on.m
Most SSRIs (Prozac, Celexa, Lexapro, Paxil, Efexor, etc.) and many antipsychotics (Seroquel, Risperdal)
Various antibiotics/anti-fungals, including the -mycin, -azole and -floxacin groups.
And lots more besides.

Long QT syndrome can also be aggravated by consumption of drugs, foods or supplements that slow down the metabolism of the causative agent. For both the 5-HT3 antagonists and (as far as the research goes) cannabinoids, the relevant enzyme is P450 [http://en.wikipedia.org/wiki/Cytochrome_P450]. Lots of drugs can affect this system, and again, a sensible harm-reduction strategy is for users to disclose their use to anyone presribing medication.

Nicotine increases P450 activity so smoking cessation can also slow P450 metabolism down. Users should be aware that if they quit smoking (or vaping nicotine, etc.), the cannabinoid dose they were taking previously will now be effectively higher.

There are also non-chemical causes of long QT syndrome. Electrolyte imbalances such as hypomagnesemia (low magnesium) and hypokalemia (low potassium) are a cause and are often the result of dehydration (e.g. hot weather, illnesses that cause diarrhoea, low intake of water or poor nutition) and various other less common illnesses.

Assuming that the synthetic cannabinoids are also 5-HT3 antagonists also indicates that serotonin syndrome is a possibility, especially when used alongside other serotonergic drugs. The Ondansetron prescribing guidelines are very clear:

"The development of serotonin syndrome has been reported with 5-HT3 receptor antagonists alone. Most reports have been associated with concomitant use of serotonergic drugs (e.g., selective serotonin reuptake inhibitors (SSRIs), serotonin and norepinephrine reuptake inhibitors (SNRIs), monoamine oxidase inhibitors, mirtazapine, fentanyl, lithium, tramadol, and intravenous methylene blue). Some of the reported cases were fatal. Serotonin syndrome occurring with overdose of Ondansetron alone has also been reported. The majority of reports of serotonin syndrome related to 5-HT3 receptor antagonist use occurred in a post-anesthesia care unit or an infusion center." [http://www.drugs.com/pro/ondansetron.html]

With both the potential for long QT syndrome and serotonin syndrome, a sensible harm-reduction strategy may be to disclose use to medical professionals involved in prescribing drugs. Medical professionals may not be aware of the possibility that synthetic cannabinoids may be have effect profiles significantly different to cannabis, so discussion of these possibilities and the similarities in structure between these drugs and commonly used 5-HT3 antagonists may be needed.

Another minor risk that users should be aware of is the possibility that ht3 antagonism may prevent normal vomiting. This is probably a comparatively minor concern, but could lead to discomfort during illnesses where relief is normally gained by vomiting (e.g. food poisoning) or masking diseases and delaying diagnosis of diseases where vomiting is a major or only symptom.

It would also seem reasonable that one of the Setron drugs could be useful for treatment of cannabinoid withdrawal syndrome as part, by alleviating the rebound symptoms without providing any cannabinoid agonist activity. I'm not aware of any current use in this context (and I'm not a doctor, so this is not medical advice as I'm completely unqualified to provide that; and even if I were qualified it wouldn't be ethical to provide it like this. At best, actual medical professionals may possibly find it a useful starting point for their own research). Presumably this would need to be carefully controlled as part of a treatment plan and the dosage lowered over time until the 5-HT3 neurons had down-regulated sufficiently). Who knows what dosage or timescale would be involved, though?

In conclusion, it is plausible that synthetic cannabinoids and commonly prescribed 5-HT3 antagonists share some of their effect profiles with an overlap in potential dangers and interactions. In addition to the possibilities of NSAID-like interactions mentioned previously, long-term and heavy use of synthetic cannabinoids may cause significant 5-HT3 antagonism resulting in a withdrawal syndrome involving rebound symptoms including vomiting and diarrhoea. There is a known risk of long QT syndrome associated with 5-HT3 antagonists, which can lead to tachycardia and sudden death, and this may also be a risk with synthetic cannabinoids which may be mitigated (though not eliminated) by avoiding combinations with other drugs (prescribed or otherwise) that also have this effect, and avoiding use when electrolyte imbalances are likely. Another danger is serotonin syndrome and similar avoidance tactics are likely to be prudent. There are also likely to be further complications due to the broader effects profile of the synthetic cannabinoids, and at least one 5-HT3 antagonist has been associated with severe, occasionally fatal gastrointestinal disease. Users of synthetic cannabinoids should not assume that the effects or safety profile of the drug can be purely attributed to the difference between partial and full cannabinoid receptor agonism.

 

[Pomzazed]

Interesting speculation indeed, im still speechless atm and has yet nothing to add,

Btw, isnt -setronlike compounds are actually metabolites of the indole/indazole CBs?
I think "tail" dealkylation is one of their metabolic fate.

If this were true, lone pair on -INACA esters on the headgroup could also contribute (mis)binding to 5HT targets - imitating the nitrogen lone pair on -setron drugs?
I also have a similar speculation that these inaca produces stronger convulsant effect upon test subject withdrawal phase.

 

[serotonin2A]

I think the distance between the carbonyl group and the amine in APINICA is to small to fit the established pharmacophore for 5-HT3 antagonists. The optimal distance is supposedly greater than 5 angstroms. These questions could easily be resolved by locating binding data for the cannabinoids, which has probably been published.

 

[niflheim]

Dealkylation of the amine chain has been found in metabolic studies of JWH-018 and 5F-AKB-48 (and probably others).

serotonin2A - since the most potent 5-ht3 antagonists have doses as low as 10ug, and cannabinoid doses of 1mg or upwards aren't uncommon (especially with tolerant users), even a decrease in potency of a couple of magnitudes could still produce significant ht3 antagonism. I've looked for ht3 binding data for synthetic cannabinoids and it doesn't seem to be out there, but if it does exist I'd be happy to be pointed in the right direction.

 

[niflheim]

Quick correction - that's 10ug/kg i.v. (i.e. about 0.6mg for an average human). However, the granisetron transdermal patches release 3.2mg/day and tolerant users can go through a gram of their cannabinoid of choice in a week or less, so the point probably holds.

 

[serotonin2A]

Quick correction - that's 10ug/kg i.v. (i.e. about 0.6mg for an average human). However, the granisetron transdermal patches release 3.2mg/day and tolerant users can go through a gram of their cannabinoid of choice in a week or less, so the point probably holds.

This type of dose comparison is really meaningless. Yes, granisetron is relatively potent, but there is no way to extrapolate from granisetron to members of an entirely different class of compounds with unknown activity at 5-HT3. You are assuming that there are no differences in the pharmacokinetics of the compounds.

You could make a more reasonable prediction by comparing the relative affinities of the cannabinoids for CB1 and 5-HT3, but that info isn't available here. So while it may be true that people who use the cannabinoids take very high doses due to tolerance, there is no way to predict based on the activity of granisetron whether the doses that are being used are actually high enough to block 5-HT3.

 

[niflheim]

serotonin2A - I agree - this is speculation. Proof would require 5-ht3 binding studies on APINACA / QUPIC and metabolites, particularly those without the alkyl chain which look like reasonable (though, as you point out, not perfect) matches to the setrol SAR. I've described the circumstantial evidence which suggests to me that 5-ht3 antagonism occurs with at least some of these chemicals, but I'm not claiming I have proof. Indeed, the research that would be required to prove or disprove this doesn't currently exist so far as I can tell.

You were arguing (or appeared to be arguing - please clarify if I'm misstating this) that the differences between APINACA and the setrol pharmacophore mean that 5-ht3 antagonism can be ruled out; by highlighting the potency of some of the setrol drugs in comparison to cannabinoid dosages I was suggesting that this, by itself, does not rule out the possibility of 5-ht3 antagonism at a harmful level. In the absence of proof, it's up to individual users to consider whether the possibility of the interactions and other dangers I describe is sufficient to avoid using these cannabinoids or to adopt some or all of the harm-reduction strategies derived from the research on -setron drugs.

 

[serotonin2A]

On the one hand, I'm not arguing that the pharmacophore model makes what you are proposing impossible, because there may be multiple modes of binding. But I don't think it is accurate to say that the SAR matches are "reasonable". In addition to the issue of the carbonyl-amine distance, it is entirely possible that the 5-HT3 receptor cannot tolerate the steric bulk of the alkyl chains present in the cannabinoid agonists. The comparison you are making is entirely based on the superficial appearance of the molecules, but the SAR data that I've seen indicates that those superficial similarities have nothing to do with how molecules actually bind to the receptor. But if you know of other SAR data that is consistent with your proposal than I would be happy to take a look!

Your speculation would be reasonable if there was some binding data showing that the cannabinoid agonists have some degree of affinity for 5-HT3. But until that evidence exists, I just don't think that it makes any sense for people to worry about this possibility. There are plenty of reasons, based on actual evidence, why people should probably avoid using these drugs chronically. I don't think it is helpful to start warning people about issues like this when there is absolutely no evidence to support your speculation.

I also wanted to respond to one thing you said. Even if in vitro studies show that certain cannabinoids have affinity for 5-HT3, that still wouldn't be "proof" that your speculation is correct. It would just be evidence that supports your speculation. In vitro binding data is not sufficient to confirm that the interactions you are proposing are actually occurring in animals or humans.

 

[niflheim]

OK, thanks for explaining, I'm not interested in arguing about what constitutes evidence, so I'll leave that side of the argument alone at this point. I don't think we're likely to come to a consensus on that anyway.

However, I do agree that in vitro studies only provide supporting evidence not proof, and probably should have phrased that better. I've also been doing some further reading around the 5-ht3 SAR and looking at some possible binding distances for various cannabinoids. You're right - APINACA is actually a failry poor match (the lack of an amine in the adamantyl group is what really kills it, I think). However, the AB and MMB variants look better. I'm not sure the conformation of these chemicals or their metabolites has actually been resolved yet, so this doesn't really do much.

Something I did find was these molecules. They're oxoquinolone analogues so not directly comparable to the indole/indazoles we've been talking about, but hopefully interesting anyway. The 5-ht3 antagonist (ki 0.48nM - not far off granisetron) has a butyl chain, while the cannabinoid has a very similar structure a pentyl chain and is very selective for CB2 (ki 29.4nM). No data on whether the 5-ht3 antagonist also has cannabinoid activity or vice versa, though.

 

[serotonin2A]

Something I did find was these molecules. They're oxoquinolone analogues so not directly comparable to the indole/indazoles we've been talking about, but hopefully interesting anyway. The 5-ht3 antagonist (ki 0.48nM - not far off granisetron) has a butyl chain, while the cannabinoid has a very similar structure a pentyl chain and is very selective for CB2 (ki 29.4nM). No data on whether the 5-ht3 antagonist also has cannabinoid activity or vice versa, though.

The ketone-amide distance is still very short in those cannabinoids. I don't think you mentioned them, but compounds such as AB-PINACA seem like better bets for cannabinoids that might bind to 5-HT3. I have no idea though if they actually do bind though...

But I still don't think the possibility of 5-HT3 effects is something that should be worrying people about using those compounds. Morphine is a 5-HT3 antagonist, but I haven't seen any evidence that morphine produces problems or should be avoided because it acts at 5-HT3. That action could potentially contribute to the morphine withdrawal syndrome, but it doesn't seem to make morphine more hazardous than other opioids.

 

[niflheim]

So, 2nd pass metabolism of the synthetic cannabinoids is not very well understood but it is known that the various first-pass metabolic products are extensively conjugated with glucuronide and cysteine. The putative cysteine conjugate of PB-22, identified as M14 (Wohlfarth et al, 2014),, is an example of how metabolism can change the location and properties of the basic centre, i.e. potentially placing it in a more optimal location with respect to the 5-ht3 antagonist SAR. It's still electronically different to the azacyclic ring structures seen in the -setron series, but if a histidine conjugate was formed, this would be much closer. Whether this happens is another question.

 

[serotonin2A]

What you are saying about the biotransformation of PB-22 is certainly true, but it is quite a stretch to link this to 5-HT3.

 

[BallsStapledToLeg]

Interesting theory with the 5HT3 involvement, however in murine models overactivation of CB1 receptors alone has been associated with nephropathy and CB1 blockade has been associated with improved kidney function in an obese disease model. Turns out CB1 agonism actually may be pro-inflammatory and pro-apoptosis via immune and oxidative stress related pathways.

http://www.ncbi.nlm.nih.gov/pubmed/20590569
http://www.ncbi.nlm.nih.gov/pubmed/17882151

My bet is that its more likely due to full agonism of the CB1 receptor, which is unlike the partial agonism and complex pharmacology of the traditional phytocannabinoids. However, you may be on to another mechanism at play

 

[niflheim]

Actually it's not. Take a look at these three bioassays on PubChem: 6314, 6369, 6041. These cover thiazole 5-ht3 inhibitors. The indole-based analogues are of note:

2-[[4-(1H-indol-3-yl)-1,3-thiazol-2-yl]methyl]guanidine - ki 3.3nM
2-(diaminomethylideneamino)ethyl 4-(1H-indol-3-yl)-1,3-thiazole-2-carboxylate ki 7.1nM
4-(1H-indol-3-yl)-2-[(5-methyl-1H-imidazol-4-yl)methyl]-1,3-thiazole - ki 10.4nM

The quinoline thiazole is also interesting:

2-(1H-imidazol-5-ylmethyl)-4-quinolin-8-yl-1,3-thiazole - ki 0.31nM

When evaluated with the same protocol (6136), granisetron came out with a ki of 2.1nM, while ondansetron the next most potent commercially produced ht3 antagoinist has a ki of 7.6, so these thiazole analogues are around the same potency and potentially much higher. The routes from the synthetic cannabinoids and their metabolites to thiazole ht3 antagonists could be formed by conjugation with amino acids (as with the cysteine conjugate from PB-22), or with acetyl-CoA or succinyl-CoA. The NSAIDs form reactive conjugates in this way, which is one explanation for their nephrotoxicity, and the structures of the cannabinoids are quite similar. Cyclization may also occur as with indomethacin as a result of P450 metabolism or spontanously as the result of enzymatic production of unstable products (i.e. the thiaminyl metabolite of PB-22 may spontanously cyclize to form the required thiazole).

This also ties into the XLR-11 kidney toxicity theory from my first post. If the glucuronide conjugates are less reactive and glucuronidation is surpressed, then more cysteine (and possibly other amino acids) conjugates should be produced with more oxidative damage as a result. It's also possible that other 5-ht ligands may be produced from these routes, especially 5-ht4 and 5-ht7. Alpha adrenergic agonists and anticholinergic activity also seems a fair bet.

 

[niflheim]

Balls - It's also possible that this mechanism is at play with natural cannabinoids, although the more highly cyclic structures would make the formation of these kinds of toxic metabolites less likely. Metabolic processes would have to both open and reclose rings - cyclization is fairly uncommon (but crucially for this hypothesis, is known to occur with some NSAIDs which have similar structures to the synthetics and are known to cause kidney damage).

 

[BallsStapledToLeg]

In humans the CB1 but not the CB2 receptor have been found in adult and fetal kidneys (link), this is interesting as in animal models non-selective cannabinioids can be somewhat protective against nephropathy due to CB2 activation . My thinking is that with the more potent cannabinoids we are seeing greater damage in humans rather than mouse/rat models due to the lack of CB2 receptor activation opposing CB1 receptor activation. I'm not discounting toxic metabolites, but in theory all CB1 full agonists could cause this

A case for overactivation of cannabinoid receptors including GPC55 and TRPV1 in human proximal tubule hypertrophy can be made in common disease states as well, suggesting that its a common mechanism for several forms of nephropathy (link)

 

[niflheim]

Balls. There's lots of interrelated systems involved here and I don't think simple conclusions can be drawn from the available data. Endocannabinoids are produced by metabolic pathways involving COX, which is also essential for production of prostaglandins that regulate kidney function. The endocannabinoids vary in selectivity for CB1/CB2 which (as you note) have differing effects on COX production. Cannabinoids are transported between and into cells by HSP70, which is a stress response protein involved in repair of oxidative damage. Oxidative damage is also a risk of fatty acid metabolism (of which endocannabinoid production is a part), which is a two-step process with reactive intermediates. NSAIDs also cause oxidative damage and paracetemol (at least) has metabolites which act as endocannabinoid reuptake inhibitors. Dysregulation of fatty acid metabolism in diabetes also results in oxidative damage from increased production of reactive metabolites. And there's more when you look at the relationships with 5-ht, alpha-adrenergic and acetylcholine receptors/production. As well as up/down regulation of all of this.

Since some of the cannabinoids in use in these experiments may themselves cause oxidative damage via metabolites (which could differ depending on whether the experiment is in vivo in in vitro due to liver/kidney specific metabolic mechanisms and whether glucuronidation is inhibited for any reason) so getting a clear picture of any of the mechanisms involved isn't easy.

 

[serotonin2A]

There is one thing that is confusing me. The cannabinoids that are known to be nephrotoxic (XLR-11 and AM-2201) do not have an ester linkage and therefore wouldn't form the amino acid conjugates. On the other hand, there is no evidence that the cannabinoids that are capable of forming the conjugates are nephrotoxic. Isn't that the opposite of what would be expected if your explanation is correct?

Is there a consensus at this point that full cannabinoid agonists are inherently nephrotoxic? If that is the case, wouldn't there be a lot more cases from a wider range of substances? I think it is telling that (as far as I am aware) there are only two nephrotoxic cannabinoids and they both have a similar structure. Although CB1 activation itself could potentially cause toxicity, isn't it just as plausible that XLR-11 and AM-2201 might share a common metabolite or some other property that is causing the toxicity? Also, is it possible that there was an impurity in the batches that was actually the cause of the problem? I'm guessing that they are synthesized by the same route.

 

[niflheim]

I think the 5-ht3 antagonism and nephrotoxicity have related but separate underlying causes. I'll recap my arguments here as it's developed a bit since the initial posts (mostly as a result of your counter-arguments, which has been very useful).

So, NSAIDs are associated with Analgesic Nephrotoxicity. This was much more common when phenacetin was in use and especially in phenacetin/NSAID combinations. Long-term use of NSAIDs is still associated with increased risk of renal cancers and impaired kidney function. So there's two issues - acute renal injury and chronic renal injury. While there's no consensus on the cause of these, and there may in fact be multiple factors, the explanation that makes most sense to me is that chronic use causes cumulative oxidative damage due to reactive glucuronidated metabolites, while acute injury is caused by much more reactive cysteine conjugates, in particular cysteine-S-conjugates which can be cyclized via cysteine lyases resulting in the production of reactive thioketones. Depletion (possibly local to the kidney) or inhibition of glucuronidation would resulting in much higher cysteine conjugate formation, and results in preferential formation of cysteine conjugates and oxidative damage to renal tubules.

XLR-11 has an unusual moiety which resembles a potent analogue of valproic acid, a drug which is known to inhibit glucuronidation. This seems likely to increase production of cysteine conjugates and cause damage via the same mechanisms as the phenacetin/NSAID combinations. Combinations of other cannabinoids with glucuronidation inhibitors may also produce this effect. The situation is a little more complicated since glucuronidation is mediated by a number of enzymes and an inhibitor of one enzyme may not inhibit others.

The 5-ht3 antagonism hypothesis shares the same initial steps. Having done some further research, I can now describe the entire mechanism for the production of a potent ht3 antagonist from PB-22.

1. PB-22 is metabolised. M14, the S-cysteine conjugate is produced. It may autocyclize or this may happen in the next step.
2. The conjugate acts as an anlogue of lipoic acid and enters the pyruvate dehydrogenase enzyme responsible for production of co-enzyme A. This adds the required imidazole. Due to the differences between thiazole and the dithiane of lipoic acid I'm not sure exactly what structure this produces. This is a known mechanism for biotransformation (Xenobiotic Incorporation into Pyruvate Dehydrogenase complex Can Occur Via the Exogenous Lipoylation Pathway). Also, having run M14 through a docking target predictor, pyruvate dehydrogenase was one of the top hits. This presumably does not affect the metabolism of most other xenobiotics as they are not fatty acids. If anyone knows of any research involving thiazole analogues of lipoic acid, that would be useful. I'm not sure it exists, though.
3. The resulting molecule may either immediately act as a 5-ht3 antagonist or may be involved in further metabolic activity and biotransformation,possibly causing various metabolic effects along the way.

Edit: There are probably some other routes including adenylate cyclase, which is involved in the production of cAMP. 5F terminated chains may or may not work for these. Other tail variants may have different paths (e.g. fluorophenylmethyl analogues could possibly have a parallel pathway via NAD+ synthase). Other variants, who knows?