Document information
Molecular Static and Dynamic Analyses reveal Flaw in Murine Model used by US FDA to Detect Drug Carcinogenicity
Correspondence: (Login to view email address)
- Murdoch University, West Australia; Autoimmunity Research Foundation, California
PDF (697.7 KB)
- Document Type:
- Poster
- Date:
- Received 15 June 2007 18:05 UTC; Posted 15 June 2007
- Subjects:
- Biotechnology, Cancer, Immunology, Pharmacology, Bioinformatics
- Abstract:
The US FDA currently accepts carcinogenicity studies of pharmaceutical drugs based on murine models. In addition to 6 month studies with p53(+/-) and ras.H2 transgenic mice, lifetime studies (typically 2 years) in WT mice or rats are also considered as evidence that a drug lacks carcinogenic activity. This model is not always exhaustive. For example, during the acceptance testing of the ARB Olmesartan1, possible carcinogenicity observed in hamsters was not able to be duplicated in rats, or in transgenic mice. We have previously used the static molecular modeling of AutoDock to demonstrate that Olmesartan has agonostic activity in the PDB:1DB1 model of the human VDR Nuclear Receptor2, while it has antagonistic activity in the PDB:1RK3 model of the rat VDR. This agonism has now been confirmed with Molecular Dynamics, using GROMACS. The murine VDR indeed lost its ability to bind the DRIP-205 co-activator when Olmesartan was the ligand, while the human VDR was activated by Olmesartan similarly to its native ligand (1,25-dihydroxyvitamin-D). Since the VDR is believed to express 913 genes3, many of which are known to be associated with cancer pathogenesis, good homology between human VDR, and the animal model VDR, is exceedingly important. CONCLUSION: The murine environment is inadequate to accurately model drug carcinogenic activity in humans. A species should be chosen which has a VDR LBP homology closer to that of man. AutoDock and GROMACS molecular analyses are useful in resolving any remaining anomalies in the observed data.
References:
1. FDA CDER: NDA-21-286, Sankyo Pharma Inc Available from URL http://www.fda.gov/cder/foi/nda/2002/21-286_Benicar.htm
2. Marshall TG: VDR Nuclear Receptor Competence is the Key to Recovery from Chronic Inflammatory and Autoimmune Disease. Abstract presentation, DMM2006.
Available from URL http://AutoimmunityResearch.org/karolinska-handout.pdf
3. Wang TT, et al: Large-scale in silico and microarray-based identification of direct 1,25-dihydroxyvitamin-D3 target genes. Mol Endocrinol. 2005 Nov;19(11):2685-95.- Presented at:
- Days of Molecular Medicine, Cambridge, MA, 22 May 2007
Discussion
- Votes:
-
8 votes
- Comments:
-
6 comments
I would be interested to see the methods of this simulation. Energy minimization techniques and parameters(vacuum? water box? what type of electrostatics?) and MD simulation parameters. Are differences in H bonds over time statistically significant, maybe you should test this? Are the pictures you show PDB files or energy minimized files, averages of MD simulations? Have other methods shown similar results, computing Linear interaction energy? Your conclusion seems a bit drastic with no ex silico data.
I would be interested to see what actual parameters and techniques were used for the Energy minimization(if one was done) and MD simulation, gromacs mdp files &c. I see no statistical analysis of the H bond data just a comment that “it looks different”. Have you performed any other methods of in silico analysis such as computing linear interaction energy? Your conclusions seem pretty drastic with no ex silico data to back it up.
Josiah, If you look through the PDF file I think you will find the degree of detail and thoroughness you are seeking. Additionally, even better than mdp files, there is a link in that PDF to our VIDEO presentation of key portions of the MD simulations. If you are concerned that we have not competently performed the simulations, the receptor structures are cited, available from RCSB, and the ligand is available in the supplementary files attached to our 2006 ARBs paper ( doi:10.1186/1742-4682-3-1 ). You are welcome to check our conclusions, and I would ask that you do that before you publicly reject them. We have a close relationship with the FDA, who have had a more detailed copy of this work in their possession for some time now, and they have expressed none of the concerns you are voicing.
Beneficial discussion with well-supported detail.
Consider an expansion of the thesis beyond the pharmaceutical arena moving into the wider gross toxicology and the wheels may begin spinning with more energy yet.
Your study provides a novel line of reasoning on the age-old questions, “Are our experiments valid? Are men simply tall mice without tails? What conclusions may be trusted and where are the new species models?”
This age of instant genetic inquiry provides fertile ground for re-examination of our presuppositions. Are you considering a spectrum of animal models per homology ranked from most similar outcomes (man:specimen) to least similar outcomes? Where do you find the murine model to be most correlative? Will you continue to rely on transgenic mice when examing these sequences? Where are you certain of the digressions (man:murine)? Do you have suitable theoretic replacements?
I can see the labs filled to overflowing with 3 dozen species for an examination of one extended genetic sequence. Perhaps that kind of detail is necessary when investigating certain areas?
I never “publicly rejected” anything, I never said your simulations were or were not correct, I do not have enough information to say this. A video as you probably know is not nearly as helpful as an mdp file. How can I try and recreate a simulation when I have no idea even how long your simulation was run for, 10 ps or 10 ns?
I have no idea what type of minimization/optimization was performed before your production MD run or if any was even performed. The 2006 ARBs paper (doi:10.1186/1742-4682-3-1) mentions a minimization using pss and minimize. For MD simulations with GROMACS I thought the standard protocol was: remove bad contacts, position restrained dynamics, then production MD run (http://research.ozreef.org/GROMACS_MD_Flowchart.pdf). I do not see any mention of these things in the ARBs paper or this one on Nature Preceedings.
By my post I was suggesting if you would like someone as myself to understand and gain knowledge from your experiment these things would be useful for the next paper you post on Nature Preceedings. - (Login to share with a colleague)
Additional information
- License:
- This document is licensed to the public under the Creative Commons Attribution 2.5 License
- How to cite this document:
-
Marshall, Trevor. Molecular Static and Dynamic Analyses reveal Flaw in Murine Model used by US FDA to Detect Drug Carcinogenicity. Available from Nature Precedings <http://dx.doi.org/10.1038/npre.2007.52.1> (2007)
- Version info:
-
Other versions of this document in Nature Precedings
None.
Other versions of this document elsewhere on the web
None known.
Joyce Waterhouse on 16 June 2007 00:23 UTC
Excellent work. Important implications for cancer research and drug development.