7.2 Applications of Pharmacophores in Predicting
Pharmacokinetic Properties
Poor pharmacokinetic properties contribute majorly to failures of many drugs
during development and clinical trials. Hence, these properties (also known as
ADMET) must be profiled during the early drug discovery process so as to avoid
failure at the later stages. Pharmacophore modelling approaches can be of great use
for prediction of the ADMET properties. If one can identify the possible interactions made by a group of drug molecules having a well-defined ADMET profile
with enzymes involved in drug metabolism, the common interacting features can be
captured as pharmacophore models and equivalent features of the query molecules
can be matched with the models. The cytochrome P450 (CYP) constitute the major
group of enzymes involved in drug metabolism out of which isoenzymes 3A4, 2E1,
2D6, 2C19, 2C9 and 1A2 carry out 90% of the metabolism. Many recent studies
report successful implementations [97, 98] of structure-based pharmacophore
models trained from the known drugs CYP enzyme interactions to predict the
suitability of query molecules to bind to a certain CYP. Also models to assess the
probability of chemical alteration of the molecules by a CYP enzyme [99, 100]
have been successfully developed and implemented. Inhibitors of the drug clearance enzymes such as the uridine 5′-diphospho-glucuronosyltransferases and
transporters like P-glycoprotein/organic cation transporter have also been utilized to
build pharmacophore models [101]. Pharmacophore models may also be employed
to predict the possibilities of off-target binging of compounds accounting for the
side effects, thereby helping design more target-specific compounds [102].
7.2.1 A Case Study with Hexadecahydro-1H-Cyclopenta[a]
Phenanthrene Framework (HHCPF)
One of the recent studies from our group [20] reports implementation of ligandbased pharmacophore model features in combination with the QSAR techniques to
establish a relationship between the number and type of pharmacophoric feature at a
particular position of the core scaffold of a group of drugs with their drug-like
properties and target binding affinities. A set of 110 FDA approved drugs containing the Hexadecahydro-1H-Cyclopenta[a]Phenanthrene Framework (HHCPF)
(Fig. 5) was considered for the study to understand their structural and functional
diversities and target specificities. Analyses of the target information collected from
DrugBank, UniProt and PDB show the selectivity of the scaffolds for different
targets and vice versa. The substituents present at 17 different positions of the
scaffolds were classified as six pharmacophoric features, viz. H-bond donors,
H-bond acceptors, aromatic rings, hydrophobic, charged and halogen groups.
ADMET (human intestinal absorption, biodegradability, P-glycoprotein binding,
carcinogenicity, Caco2 cell permeability, Ames test positivity, blood brain barrier
permeability, hERG, CYP450 binding, Rat LD50, etc.)/physicochemical properties
Pharmacophore Modelling and Screening: Concepts, Recent …
45
Pharmacokinetic Properties
Poor pharmacokinetic properties contribute majorly to failures of many drugs
during development and clinical trials. Hence, these properties (also known as
ADMET) must be profiled during the early drug discovery process so as to avoid
failure at the later stages. Pharmacophore modelling approaches can be of great use
for prediction of the ADMET properties. If one can identify the possible interactions made by a group of drug molecules having a well-defined ADMET profile
with enzymes involved in drug metabolism, the common interacting features can be
captured as pharmacophore models and equivalent features of the query molecules
can be matched with the models. The cytochrome P450 (CYP) constitute the major
group of enzymes involved in drug metabolism out of which isoenzymes 3A4, 2E1,
2D6, 2C19, 2C9 and 1A2 carry out 90% of the metabolism. Many recent studies
report successful implementations [97, 98] of structure-based pharmacophore
models trained from the known drugs CYP enzyme interactions to predict the
suitability of query molecules to bind to a certain CYP. Also models to assess the
probability of chemical alteration of the molecules by a CYP enzyme [99, 100]
have been successfully developed and implemented. Inhibitors of the drug clearance enzymes such as the uridine 5′-diphospho-glucuronosyltransferases and
transporters like P-glycoprotein/organic cation transporter have also been utilized to
build pharmacophore models [101]. Pharmacophore models may also be employed
to predict the possibilities of off-target binging of compounds accounting for the
side effects, thereby helping design more target-specific compounds [102].
7.2.1 A Case Study with Hexadecahydro-1H-Cyclopenta[a]
Phenanthrene Framework (HHCPF)
One of the recent studies from our group [20] reports implementation of ligandbased pharmacophore model features in combination with the QSAR techniques to
establish a relationship between the number and type of pharmacophoric feature at a
particular position of the core scaffold of a group of drugs with their drug-like
properties and target binding affinities. A set of 110 FDA approved drugs containing the Hexadecahydro-1H-Cyclopenta[a]Phenanthrene Framework (HHCPF)
(Fig. 5) was considered for the study to understand their structural and functional
diversities and target specificities. Analyses of the target information collected from
DrugBank, UniProt and PDB show the selectivity of the scaffolds for different
targets and vice versa. The substituents present at 17 different positions of the
scaffolds were classified as six pharmacophoric features, viz. H-bond donors,
H-bond acceptors, aromatic rings, hydrophobic, charged and halogen groups.
ADMET (human intestinal absorption, biodegradability, P-glycoprotein binding,
carcinogenicity, Caco2 cell permeability, Ames test positivity, blood brain barrier
permeability, hERG, CYP450 binding, Rat LD50, etc.)/physicochemical properties
Pharmacophore Modelling and Screening: Concepts, Recent …
45