Satori Makes Good (Steroids)

Whether it's because of the evergreen interest in Carl Djerassi, the discovery of a (new-to-me!) blog, or Percy Julian's recent Google Doodle, everything in the blogosphere seems to be coming up steroids.

Take a peek at these back-to-back OPRD articles, both from chemists at the former Satori Pharmaceuticals.

As a "cub blogger" for CENtral Science, I still remember gaping at how radically different their root-derived phytosterols were from the rest of the anti-Alzheimer's medicines. Now, almost a year to the day after Satori shut its doors, we have some insight from the team responsible, just before they scattered to the winds.*

So, that means 6.99 tons of plant waste to dispose? [rubs eyes]
OPRD, 2014 ASAP

To access their preclinical candidate, Satori scientists first needed a reliable supply of the glycosylated intermediate. Enter seven metric tons of dried black cohosh, a traditional medicine used as a pain reliever by Native American healers. A third-party vendor crushed the root and sent it to another firm, where they extracted it with ethanol, passing about 300 kg of "solids" back to Satori. Suspension in brine / DCM partitioned the desired compounds into the organic phase, which was treated with triethylamine and catalytic zirconium tetrachloride. This ejects the pesky E-ring alcohol, and the resulting compound performs a net oxidation to yield a diastereomeric mixture of ketones (above).


(Warning: I guess you haven't lived 'til you've purified 70 kg of crude, brown extract with DCM over silica...yuck!)

So, once Satori had in hand ~11 kg of compounds 1 + 2 (see right), they needed to advance the structures to their desired candidate (below) which you'll notice has a few little changes from the plant-derived drug. Gone are the acetyl group. Reduced is the ketone. Ripped apart is the sugar, making way for a morpholine.

The scientists' first-gen synthesis, an 8-step sequence, wasn't up to snuff for kilogram work. The trouble? Too many chemically similar hydroxyls, prompting some protecting group manipulation to target one or two selectively. Process work on the ethyl ether step - utilizing diethylsulfate / tert-butoxide in place of an earlier NaH / EtI mix - allowed a telescoped 5 step route, cutting out 2 silica gel columns and upgrading the final process purity to >95%.

Interestingly, the authors, ever circumspect, reflected on the limitations of their final process:

"While this reported route was sufficient to provide the kg-scale quantities of target compound for preclinical studies, we acknowledge that it has limitations that would make it impractical at the 100 kg scale."

Silica gel chromatography strikes again! Still, their candidate came through on 1 kilo, at about 30% overall yield after HCl salt formation. Not too shabby.

Update (5/30/14): Want to see how Satori chose these molecules? One of the authors (Hubbs) writes in to recommend their 2012 J. Med. Chem. optimization paper.

*According to the author lists, everyone on the team ended up in a different place: Enanta, AstraZeneca, Celgene, ETH, Genzyme, Resilientx, Sanofi.