DMSO to SAHA: Questions of toxicity

I alluded in the last post to Breslow and Friend's discovery of DMSO as a HDAC (histone deacetylase) targeting agent. The story of how such a simple molecule was made into a putative drug deserves a closer look. The Nature Biotechnology paper that was cited in the last post contains most of the details. Suffice it to say here that the following points were noted.

DMSO was postulated to bind to a metal ion. Since Zn-heteroatom interactions are well known, another known sulfur binding moiety (look here for Whistling's list of zinc binding organic motifs), hydroxamic acid, was modified after some SAR studies to give the end molecule. One of the interesting steps in the lead optimization was the simple observation that it was unlikely that there were two metal ions in the same active site which bound to two hydroxamic acid groupings. So instead, one hydroxamic acid grouping was converted into a phenyl ring to gain binding affinity from hydrophobic binding. As the picture above shows, the drug now nicely fits into the active site, with the expected metal-hydroxamic acid interaction.

A journal club meeting member realised that SAHA's structure was similar to that of TSA (trichostatin A shown above), which was known to modulate HDACs. Thus came the hypothesis, later validated, that SAHA binds to HDACs. The details however are not yet clear, as HDAC interaction leads to a cascade of different events, and there are a dozen HDACs which are crucial for transcription.

There is one paragraph in the paper however, which I don't completely agree with. The authors were testing a few analogs, and came up with a branched structure with two hydrophobic rings that was much more potent than SAHA. However, it was later found to be quite toxic. While the authors make a good point that increased potency (increased binding affinity) can translate into greater toxicity because of continuous target activation, they also say:

"SAHA has hit the happy medium. It is potent enough to be useful and tolerated in patients. If the dosing is intermittent, such as not to maintain a continuous ‘therapeutic’ level of SAHA, it can be released from the binding site periodically so as to allow the deacetylation activities in cells. This is a general consideration, which could well be true of many other medicinal compounds. Thus, it is probably a mistake for medicinal chemists to set out first to find the most potent compound they can achieve in a series and then to look at any question of toxicity, as is often done."

The first thought that came to my mind after reading this was; this can't be quite right. Aren't some of the best-selling drugs of all time both highly potent and not particularly toxic? Toxicity and potency cannot always coincide.

I think there are two issues here. First is the fact that 'toxicity is in the eye of the beholder'. Cancer drugs are some of the most toxic drugs ever produced, but they are tolerated because of their essential requirement. So toxicity is a relative phenomenon whose impact has to be judged based on the disease. Secondly, even for very potent compounds, toxicity will depend on the exact nature of the target modulated, and its role in life processes. There may be a target like HDAC which plays a central role in the crucial process of transcription. You wouldn't want to keep this target activated/deactivated for too long. On the other hand, there might be another target which could be activated or deactivated for relatively longer periods because of its less sensitive nature and role in metabolism. So while the opinion of Breslow et al. is not wrong, it misses the other big part of the picture- the protein- on whose nature toxicity will depend too.