Chemistry Reference
In-Depth Information
for each company or organization. Furthermore, company compound collections may be
increasingly populated by compounds that are available to other companies; consequently,
any hits carry additional competitive or intellectual property risks. Such risks are lowered
for fragment-based hits because even common hits may be considerably evolved in novel
ways during the optimization phase.
A second consequence for fragment-based methods is that by restricting to hits with
high ligand efficiency, we improved the likelihood of evolving towards suitable drug
candidates with a target potency of
10 nM while keeping the molecular weight below
500 Da . The optimization of typical drug-like hits from HTS (typically molecular weight
450-500 and target potency 500-1000 nM) may require further increase in molecular
weight or efforts to truncate and redesign. Compound
21
has 26 heavy atoms (molecu-
lar weight 351 Da ) and a ligand efficiency of 0.37. Assuming that the ligand efficiency
can be maintained, the target potency (10 nM) should be obtainable for a compound with
molecular weight around 420 Da ; this would correspond to addition of as few as three
heavy atoms. Fragment-based lead generation methods were successful here because three
essential components were in place. First, NMR served as a robust and sensitive hit iden-
tification tool and methods were available to identify false positives quickly. Second,
we could quantitate binding affinities over a wide range using SPR spectroscopy and
thereby build and evaluate structure-activity relationships. Third, the design process was
accelerated because we could obtain highly detailed structural information through crys-
tallography on a routine basis, even for compounds with low affinity. Taken together,
these approaches allowed us to identify hit
1
and rapidly improve its affinity by 30 000-
fold to deliver lead
21
in an efficient manner, requiring only a small number of chemical
modifications.
∼
11.10 Acknowledgments
We gratefully acknowledge the contributions fromour colleagues and collaborators atAstex
Therapeutics, includingGianni Chessari,Miles Congreve, ChrisMurray and Sahil Patel.We
also thank Lise-Lotte Olsson for computational support and Tonny de Beer, Fredrik Edfeldt,
Rutger Folmer and Stefan Geschwinder for scientific input, support and illustrations.
Finally, we thank all our other collaborators whose names appear in the cited papers.
References
[1] Dominguez, D. I., De Strooper, B. Novel therapeutic strategies provide the real test for the
amyloid hypothesis of Alzheimer's disease.
Trends Pharmacol. Sci.
2002,
23
, 324-330.
[2] Luo, Y., Bolon, B., Damore, M.A., Fitzpatrick, D., Liu, H., Zhang, J., Yan, Q., Vassar, R., Citron,
M. BACE1 (-secretase) knockout mice do not acquire compensatory gene expression changes
or develop neural lesions over time.
Neurobiol. Dis.
2003,
14
, 81-88.
[3] Cai, H., Wang, Y., McCarthy, D., Wen, H., Borchelt, D. R., Price, D. L., Wong, P. C. BACE1
is the major -secretase for generation of A peptides by neurons.
Nat. Neurosci.
2001,
4
, 233-234.
