Biomedical Engineering Reference
In-Depth Information
overlooked, since an increasing amount of attention has focused on this feature and
its correlation to drug-likeness.
Although complexity is desirable in screening collections, it is difficult to compute
compared to more straightforward physicochemical properties which have played an
important role in the design of small molecules for drug discovery. Lipinski's rule
of 5 focused on molecular weight, log P , hydrogen-bond donors and acceptors, and
their correlation to bioavailability [12]. Additional analyses have expanded these
filters to other standard physicochemical properties, but only recently have complex-
ity descriptors been calculated and used to correlate a small molecule's chances of
success through clinical trials. Work done by Lovering et al. from Wyeth looked
at both sp 3 content and stereogenic centers in this regard [13]. In this analysis, all
compounds from 1980 onward were retrieved from the GVK BIO database repre-
senting discovery through phases I, II, and III and approved drugs. Complexity was
measured by the fraction of sp 3 content (Fsp3), which is the number of sp 3 carbons
divided by the total number of all carbons. As Fsp3 increases, the compound contains
more three-dimensional character, which can correlate to increased complexity. The
analysis shows not only the general increase in the average sp 3 content for every
stage going from discovery to drugs, but that drugs have a 31% increase in the Fsp3
content compared to that of compounds in discovery. A similar analysis was done on
compounds that contain at least one stereogenic center, another descriptor for com-
plexity. The same enrichment is observed as the average is calculated for each class of
compounds, ranging from discovery to drugs. In this case, 46% of the compounds in
discovery contain at least one stereogenic center, while 61% of drugs contain at least
one stereogenic center, a 33% increase. Interestingly, the authors go on to correlate
increased sp 3 content with increased solubility and lower melting points, a descriptor
for predicting drug absorption.
In addition, work done by Luker et al. from AstraZeneca further correlates com-
plexity as measured by flatness and toxicity [14]. Their analysis focused on a set
of internal small-molecule candidates that were investigated in rat and dog in vivo
toxicology studies over the past 10 years. The authors were attempting to develop a
predictive model for preclinical toxicology based off structural descriptors. Although
many properties are described as being a contributing factor in a compound's toxicity,
flatness, as measured by the ratio of aromatic atoms to total heavy atoms [15], was
specifically cited as having a statistically relevant impact. Compounds that reached
humans had an average of 49% of the atoms aromatic, while compounds that were
stopped due to toxicity had an average aromatic atom count of 64%.
DOS is a synthetic strategy using short, modular pathways to access complex
scaffolds. One such strategy is the build/couple/pair (B/C/P) approach, which has
been utilized in several published library productions [16]. In the first step, highly
functionalized, stereochemically enriched building blocks are built (Figure 17.2).
Accessing all stereochemical permutations is highly beneficial for the generation
of stereostructure-activity relationships (SSAR) [17] in future HTS campaigns. The
building blocks are then coupled together with all possible stereochemical com-
binations to access all stereoisomers of a common intermediate. Pairing reactions
combine complementary functional groups to rigidify the scaffold and accentuate the
stereochemistry built into the scaffold.
Search WWH ::




Custom Search