Solubility-related Toxicity and DDI

The impact of low aqueous solubility is well documented, from unreliable assay results to limited oral bioavailability. What I had not realised is that it didn't stop there...

A recent Genentech JMedChem article on the discovery of their NLRP3 inhibitor GDC-2394 highlighted a major potential issue with insoluble compounds, or more precisely with pH-dependent low soluble compounds. The first lead compound the authors described was compound 17, which progressed all the way to preclinical tox. Unfortunately renal toxicity was detected in cynomolgus monkey (increased kidney weight, elevated blood urea nitrogen levels and presence of "ghost crystals"). The Genentech team was able to determine that the cause of this toxicity was the precipitation of the parent compound in the kidney fluid. It turns out that the pH of this fluid can go down to 6.4 and, as shown in the table below, the solubility of the 17 drops down with lower pH (Solubility at higher pHs increases due to the acidic sulfonyl urea group(pKa~5)).

The Genentech scientists were able to develop a zwitterionic chemotype with an incredibly improved solubility profile while maintaining a reasonable PK profile (and increasing the potency). Gratifyingly, the toxicity signal disappeared. Of note, the Vss was increased in all species, which is one of the real advantage of zwitterions. Of note, also, is the drop in permeability, which is consistant with Paul Gleeson's simple ADMET rule of thumbs for acids vs zwitterions with clogP<3 but not general for zwitterions vs acids.

Another solubility-related renal toxicity example was shared by Janssen scientists in the c-Met inhibitor literature. Following similar toxicity observations, the authors were able to identify JNJ-38877605's metabolites as the "renal crystals", which precluded any further development of this compound.

The metabolites (M3 and M5) depicted on the right, come from Aldehyde Oxidase mediated oxidation of the quinoline ring. In this case, the pH-dependency of the solubility from pH 6-7.4 is less clear, but, not surprisingly, these compounds display extremely low aqueous solubility (<5μM).

Scientists from the Shanghai Institute of Materials Medica were able to elegantly tackle the liability by using a deuterium atom to slow down the AO-metabolism. This strategy (even better than atom-economical, neutron-economical!) was also used by Vertex for their DNA-PK inhibitor VS-984 (highlighted in this very well written and exhaustive review on Deuterium).

Scientists from Hutchison MediPharma were very conscious of the issues around the quinoline ring when they designed Volitinib. Ironically, they were able to identify another AO-mediated metabolite with low solubility on the central core this time...Thankfully, this metabolite was only formed in minuscule amount and no renal toxicity was reported.

Another liability of low solubility compounds I had not appreciated, is a potential for DDI. Drug-drug-interactions are very present in the medchemist's mind and the main reason we screen for CYP inhibition, CYP time-dependent inhibition and CYP induction. It turns out pH-dependent soluble compounds can be the victims of DDI with Proton-pump inhibitors (PPI), H2 antagonists and Antacids, all of which will lead to an increase in the intragastric pH. For low solubility basic drugs, this can lead to precipitation and eventually less drug being absorbed.

This turns out to be quite problematic and needs to be derisked using clinical trials as exemplified here: interaction of PPI lansoprazole and quizartinib.

Thankfully, there are clear strategies to mitigate this risk like formulation optimisation (Razaxaban), staggered administrations (the impact on pH is only for a couple of hours) or drinking well-known soft drinks to bring the gastric pH back down.

To give you an idea of how often this happens, I encourage you to take a look at this review on DDI with kinase inhibitors (not surprisingly this happens quite a lot for poly aromatic basic compounds like kinase inhibitors). The table on the right highlights the impact of these DDI on the AUC in clinic.

As mentioned above, with clear workarounds, this DDI is probably not a major issue in the end (not like the renal toxicity mentioned at the beginning). However, this will mean additional studies for derisking.

Adding ionisable groups is the number one strategy recommended by most medchemists to tackle solubility issues. This strategy comes well-documented downsides (permeability and potential for reactive C-glucuronides for acidic groups/ efflux and hERG for basic ones). Renal toxicity through precipitation/potential DDI with intragastric pH-modifying drugs should be added to the list of things to key an eye on.

Back in my AZ days on the Reims site, the head of chemistry, Laurent Hennequin, used to drill into us to focus on neutral compounds and tackle their solubility by optimising their melting point (I was this close to tattooing the Yalkowski equation on my forearm :-) )...Maybe he was on to something...Roule ma poule!