QDIS Blog

Recently two new chemical routes to Tamiflu have been reported in the chemical literature:

A Short Enantioselective Pathway for the Synthesis of the Anti-Influenza Neuramidase Inhibitor Oseltamivir from 1,3-Butadiene and Acrylic Acid
J. Am. Chem. Soc., ASAP Article 10.1021/ja0616433 S0002-7863(06)01643-X
Web Release Date: April 25, 2006

De Novo Synthesis of Tamiflu via a Catalytic Asymmetric Ring-Opening of meso-Aziridines with TMSN₃
J. Am. Chem. Soc., ASAP Article 10.1021/ja061696k S0002-7863(06)01696-9
Web Release Date: April 25, 2006

Both of these reports use cheaper more readily available starting materials than shikimic acid that Roche currently uses.

The first article is from E.J. Corey’s lab and describes a synthesis starting from butadiene and an acrylic flouro-ester. It involves 11 steps and from a development and scale-up perspective the synthesis has several reagents that would need to be removed. This includes TMSOTf and SnBr₄ (although it is catalytic) as well as CCl₄ solvent. These are all problems that development chemists deal with on a daily basis and should be fairly easily overcome. The overall yield is 28% according to my calculations (they claim 30%) with only one low yielding step (61% but not optimized) but these were all on a very small scale (total of 80 mg of the final intermediate was made). As those who spend time scaling up chemical processes these yields rarely translate directly on scale and usually require optimization. However, in general, the reactions used are fairly common and should scale readily. It is also interesting to note that the paper specifically states that this route is in the public domain and presumably they do not intend to patent this process.

The second paper is from the Masakatsu Shibasaki group in Japan and has more serious concerns from a scale-up perspective since the key step involves TMSN₃. Not only is this a hazardous and expensive reagent but it is also used in large excess (3 molar eqv). It also involves a large catalytic amount of a Y(OiPr)₄ (10mol%) and a chiral additive (20mol%). Also the sequence is rather lengthy at 15 steps and overall yield of 0.01% according to the reported yields. Several steps are low to moderate yields and the last two steps have yields of 53% and 50%, the worst possible place to have low yielding steps. It also involves a Dess-Martin oxidation; these are never scaled up due to the hazards and so another method would have to be demonstrated. Ni(COD), SeO₂, TMSCN are other reagents that would have to be eliminated. I don’t hold much optimism for this route as a great amount of development work would be needed before even kilo-lab quantities could be produced.

In the development process there is always the need to weigh how quickly the material is needed with how much material needs to be produced. That is, it doesn’t make sense to spend a large amount of time perfecting a synthesis only to find the drug fails in phase 1 or early phase 2 trials. However, in this case, the drug is already approved and so perfecting the route should be a high priority. However, post-apporval chances are time consuming and costly since that route would need to be validated and shown to produce equal or better material than is currently produced. Also the different route will result in different impurities that must be dealt with both in terms of analytical chemistry as well as toxicology.

So, although it is interesting to see these developments, it won’t have any immediate impact on the Tamiflu supply issues.

Technorati Tags: Tamiflu

Email Ed Vawter

Bookmark it with: | del.icio.us | Digg it | Furl | Simpy | Spurl | My Yahoo! |