June 7, 2012
PKC Pharmaceuticals is working very hard to try and not only remedy the issues that have come out of the Bosutinib issue described in previous posts and in other media outlets, but to also really understand what went wrong, what is out there, and how to clear the air on this issue. I believe they should be commended for that.
In the Bosutinib Report #1, issued by PKC on February 20th, 2012 there is an exhaustive review on their analysis up to that date. They plan a report #2 in the near future.
Report #1 is an interesting read and contains several interesting observations about a historical precedent for this kind of issue in the biochemical reagent space, but most interesting for me was the posed question of “What is Bosutinib?”
You see, PKC is now reporting evidence that there is possibly not just 2 different versions of bosutinib being sold, but the potential for a 3rd variant!
The report goes on to explain exhaustive efforts on their part to collect melting points, TLC, HPLC data, and full spectroscopic data to identify the identities of the compounds claimed to be bosutinib.
At page 13, comes the big question. What is bosutinib? From the report:
For the question, “what is bosutinib?”, our opinion is that, in view of:
(i) the substantial structural ambiguities (for example, the 16 possible isomers for the simple anilinic raw material used to make bosutinib, many of whose structures cannot be rigourously established, distinguished or ruled out in the final molecule by NMR analysis
(ii) the general and specific difficulty of rigourously proving many complex chemical structures such as bosutinib by any methods other than x-ray crystallography and;
(iii) the fact that bosutinib has been or now is being administered to humans in clinical trials.
bosutinib should now be rigorously defined as that x-ray crystal structure that results from x-ray crystallography carried out on a sample taken directly from the actual “active pharmaceutical ingredient” supplies used to prepare the drug product administered to humans in the Pfizer-sponsored clinical trials- “from the same lot”
Now I don’t discount, at this point, the need for very riguorous analysis and structure confirmation of bosutinib by any means necessarry. And there is no question that the most reliable method for structure elucidation of this compound is by X-Ray.
But I also can’t help but state that while this will hopefully uncover this problem, it doesn’t address the overall problem, and it certainly does not prevent an instance like this from happening again. It doesn’t change the fact that false positives happen. Whether they are from degradation products, products from incorrect starting materials, incorrect compounds coming from an external supplier, etc.
X-Ray is a powerful tool but it is also not fast, nor routine, difficult to realize for some research groups, and still relatively expensive. In the industry, X-Ray really becomes the last resort for many of the above reasons. For the more routine cases, this would be like using a hand grenade to kill a housefly.
So the questions come back to the first statement about whether NMR analysis alone can be used to establish, distinguish, or rule out structural candidates. Of course this is a big job because to truly be 100% certain that you’ve got the correct structure you need to rule out all other possible alternatives. In the case of bosutinib, once you look at the molecular formula and generate all possible isomers, that is clearly a daunting task for a human.
Enter Computer-Assisted Structure Elucidation (CASE). ACD/Structure Elucidator is the most peer-reviewed CASE system available today with over 30 publications focused around the technology and performance. A full book was recently published by the RSC on CASE and much of it is devoted to ACD/Structure Elucidator as the co-authors are ACD/Labs employees past and present.
Specific to this blog posting, the methodology behind Structure Elucidator is such that it can generate ALL conceivable structures from spectral data and a priori knowledge. And it can do this pretty quickly depending on the NMR data made available. And of course the instrument vendors have made great strides in hardware development to make some of the more sophisticated 2D NMR experiments available in a reasonable amount of time for the special cases.
One potential article of interest would be Structureal Revisions of Natural Products by Computer-Assisted Structure Elucidation (CASE) systems. This publication highlights the application of CASE systems to a series of examples in which original structures published in the literature were later revised.
Another notable case study can be found here. it explains the structure elucidation of a molecule where the authors had to resort to residual dipolar couplings to solve a structure based on their belief that the structure could not be solved by classical spectroscopic means.
Which brings us back to Bosutinib. So far we have obtained NMR data of the bosutinib isomer from our collaborator Phil Keyes, Lexicon Pharmaceuticals. We already ran that through our ASV systems and presented the result. Next up was running the data through ACD/Structure Elucidator. We did our first run with the following experiments:
– 1D 1H NMR
– 1D 13C NMR
– 2D COSY
– 2D HSQC (DEPT-edited)
– 2D HMBC
Lo and behold, following structure generation the software generated 17 unique structures based on the data above. After ranking these structures based on average 13C chemical shift deviation between experimental and predicted values, the #1 best structure was indeed Bosutinib Isomer #1:
Note that the average 13C chemical shift deviation between experimental and predicted for the bosutinib isomer #1 was 1.980 ppm which in our experience is very low and suggests a strong level of consistency between the proposed structure and the data. Also worth noting that quite obviously, the actual proposed structure of bosutinib was not generated as any of the 17 possibilities based on the data provided.
That said, there were other interesting structures proposed and we do have some additional 2D NMR data to put into the software. This was a quick and dirty run to generate some quick results and to make some quick conclusions. More to follow. We are going to take a more rigourous look at all the data provided and generate some additional results. We also hope to soon get our hands on the authentic bosutinib material and run similar tests on that.
Special thanks to Philip Keyes for acquiring and sharing the NMR data, and to my colleague Joe DiMartino from promptly running the data in Structure Elucidator. I’d also like to acknolwedge the efforts being put forth by my colleagues Mikhail Elyashberg and Kirill Blinov for conducting some more rigourous tests with this data and I will share more results as they become available.