Vicinal diols: two alcohols for the "cost" of one

Adding a -CH2OH group to a molecule comes with significant benefits (it will reduce your logD and most likely increase your solubility, potentially reduce your CLint (if not introducing a new metabolism route) but the lower logD and the additional Hydrogen Bond Donor can (will) really hurt your permeability, your efflux....

But what about adding an additional alcohol to an existing -CH2OH group?... There is a good chance that the alcohols in a diol moiety will make an Intra Molecular Hydrogen Bond essentially tying up the additional HBD/HBA and potentially minimising its effect on permeability.

The fact that diols are showing up in drugs/clinical candidates certainly would suggest that their impact on properties must not be detrimental to bioavailability.

What better reason to start digging in the literature and to try to document the impact of diols vs single alcohols!

The first example comes from the Genentech/Argenta collaboration on their JAK2 inhibitor programme (JMedChem2012). When exploring the solvent facing vector off of the pyrazole, the team profiled ethanol and diol substituents. Both modifications maintain most of the potency, while increasing solubility. Keep in mind this is kinetic solubility, it is what you would expect from a logD perspective. CLint is slightly lower on the diols which is also encouraging, but it probably does not drop as much as it should given the likely lower logD. Unfortunately, the JAK selectivity of these compounds was eroded and they were not profiled any further.

As part of their effort to identify pan-BET inhibitors, the following example from GSK scientists (JMedChem2022) gives us additional insights. Measured logD shows a strong reduction with the second hydroxy group- this is slightly different from the other examples in this post: we are only adding a hydroxy whereas we are adding CH2OH in the other examples. The potency is maintained and , although it takes a hit, the permeability is relatively maintained given that the logD dropped by 1.5 units. Further optimisation led to the identification of diol 41, an orally-bioavailable candidate quality compound. The authors comment on the instrumental role of the IMHB of the diol with respect to permeability.

The third example comes again from the Genentech/Argenta collaboration, this time from their MEK allosteric inhibitors programme (BMCL2014). The diol analog displays similar potency, presumably lower logD which translates into lower hERG inhibition and lower Clint in both rat and human. The beneficial impact is maintained in vivo with lower CL (check out this post if you think this was a given). Surprisingly the thermodynamic solubility is worse. Although this is from amorphous material, it is still a slight flag that the second hydroxy might not always bring the desired solubility boost, most likely due to tighter crystal packing/higher melting point in this case. More importantly, the permeability/efflux profile is still maintained for the most part, suggesting again that if an alcohol-bearing molecule is permeable, its diol analog will also probably be permeable.

The fourth example comes from Novartis scientists in their quest for brain-sparing GIRK1/4 inhibitors (BMCL2023). This is a nice example of the diol displaying equivalent permeability to the alcohol - and also somewhat surprisingly to the methoxy analog. The benefits on kinetic solubility are a bit higher than the alcohol one (with the caveats mentioned above).

The final example also comes from Novartis scientists. The article covers the efforts leading to the identification of the novel brain-penetrant proteasome inhibitors 7 (JMedChem2022). Unfortunately, we do not have exact matched pair, but along the way, the authors did profile a diol, which had quite an interesting profile, with similar Microsomal Clint as the candidate (although at presumably lower logD), and comparable in vitro permeability/efflux. However, for this particular series, kpuu in mouse did not correlate with P-gp mediated efflux, with Compound 3 having a significantly lower kp,uu. These result could hint to the diol being recognised by another transporter.

In conclusion, the title of this post was a bit click-baity and clearly there is a price to pay for the additional alcohol...however it is sometimes fairly low and definitely worth exploring and profiling!

Just to emphasise the context of this post, if a primary alcohol is tolerated on your ligand (potency and ADME wise) then I would suggest trying the diol out... not directly trying a diol out of the blue....

If you need a little more convincing, I should also add that diols have been used as unionized carboxylic acid mimetics as highlighted in this recent review (BMC2024)!

#drugdesign #medchem