Drug Hunter Letters #2: On Carboxylic Acids and Acylglucuronides
Recently, Dr. Alf Claesson (CEO of Awametox AB, inventor of SpotRM, and previously Principal Scientist and Section Head at AstraZeneca/Astra AB, Sweden) reached out about our resource on the bioactivation of structural alerts. Dr. Claesson shared the helpful notes that:
sulfonylureas have limited evidence to be a concern
hydantoins may not have sufficient evidence to be a concern
Here Dr. Claesson elaborates on carboxylic acids in particular.
On carboxylic acids and acylglucuronides in medicinal chemistry
from Alf Claesson
Awametox AB, Lilldalsvägen 17 A, 14461 Rönninge, Sweden. Email: email@example.com
On every team there are differing views on which functional groups are allowed and which should be avoided. One functional group that has been troublesome and bewildering for decades is the carboxylic acid (CA), particularly with respect to its formation of acylglucuronides (AG) and their role in idiosyncratic drug induced liver injury (IDILI). For example, Hammond et al. state about AGs: “these commonplace metabolites have been linked persistently, and at times almost generically, but always somewhat uncertainly, with the varied adverse reactions of carboxylate drugs.”
I believe that the controversies around these abundant metabolites are settled (but I’m not quite sure!).
To summarize our recent Mechanism in Context report (within the application SpotRM) on the topic:
Traditionally, AG formation was the main concern with carboxylic acids. Recent FDA guidelines on MIST from 2016 even mention AGs as “toxic metabolites” despite a lack of solid evidence for in vivo toxicity attributable to an acyl glucuronide (see extensive discussion in a review on RM alerts by Kalgutkar 2019).
This 2015 paper from AstraZeneca scientists highlighted the fact that the formation of CoA thioesters (40-70x more reactive than AGs) is a more likely culprit for toxicities associated with carboxylic acid
One of the most puzzling examples had always been ibufenac (withdrawn from Europe in 1968) vs. ibuprofen (one of the safest NSAIDs), whose AGs only differ by 2-3-fold in reactivity, neither of which are very reactive. The CoA thioesters of each, however, readily form covalent adducts, with ibufenac-CoA being an order of magnitude more reactive
The exaggerated focus on AG formation from CAs should come to an end but likely continues because of FDA guidance documents. Here is a key excerpt from reactive metabolite expert Jack Uetrecht’s 2020 views:
“The simple correlation that has been suggested between the reactivity of acyl glucuronides and the risk of IDRs is not at all clear. There is simply not convincing clinical evidence that the covalent binding associated with acyl glucuronide formation is responsible for serious IDRs. It is possible that acyl glucuronides can lead to IDRs but testing their reactivity during drug development would lead to a large number of false positive and false negative results. Alternatively, acyl glucuronides may rarely or never cause IDRs. Yet, as stated earlier, many pharmaceutical companies routinely study the reactivity of acyl glucuronides of drug candidates that are carboxylic acids. In part, this is likely due to FDA guidance documents that imply that acyl glucuronides are a significant source of risk. It is possible to generate a large amount of data with high throughput in vitro assays; however, such data may do little to improve drug safety.”
Summary Guidance for Carboxylic Acids as Structural Alerts
A review of a large number of specialist reports in the last 10-15 years suggests that carboxylic acids can be treated as distinct groups (see also perspectives in CRT2018 by Claesson & Minidis, in JMC2019 by Kalgutkar, and in JMMC20 by Uetrecht).
Arylacetic acids (i.e., alpha-unsubstituted) remain a definite structural alert along both AG and CoA thioester formation hypotheses.
α-Alkyl arylacetic acids, although likely to be less hazardous than arylacetic acids, are not totally harmless. Ibuprofen, for example, gives rise to extensive covalent protein binding via the acyl CoA route (Darnell et al. 2015). General awareness and guidance by in vitro evaluations are recommended.
Aryloxyacetic acids, which are shown to form highly reactive AGs, have few drug examples, and largely represent unprecedented territory (the hepatotoxic tienilic acid is an aryloxyacetic acid that also contains a thiophene, and bendazac (a 3-indazolyloxyacetic acid) was withdrawn in 1993 due to hepatotoxicity (Spain).
Other alkyl (normal or branched) carboxylic acids deserve a vigilant view based on the cited valproic acid and fasiglifam/ MK8666 cases, and possibly on amineptine.
Arylcarboxylic and benzoic acids seem to have little or no empirical concerns.