Discussion

SwissADME is a free web tool applied to evaluate the ADME (Absorption, Distribution, Metabolism,
and Excretion) behaviour, including physiochemical properties, lipophilicity, water solubility,

pharmacokinetics, drug likeness, and medicinal chemistry properties of the compounds. It has been
demonstrated that early estimation of ADME in the discovery phase reduces drastically the fraction
of pharmacokinetics-related failure in the clinical phases.
The first section presents 2D chemical structure and canonical SMILES. It shows on which chemical
form the predictions were calculated. Besides, the Bioavailability Radar is displayed for a rapid
appraisal of drug-likeness. Physicochemical properties such as lipophilicity, size, polarity, solubility,
flexibility, and saturation are taken into account and defined by descriptors. A range of optimal
values is depicted as a pink area which is inspired from commonly accepted bioavailability and drug-
likeness guidelines. For saturation, the ratio of sp hybridized carbons over the total carbon count of
the molecule (Fraction Csp3) should be at least 0.25. For size, the molecular weight should be
between 150 and 500 g/mol. For polarity, the TPSA should be between 20 and 130Å. For solubility,
log S (calculated with the ESOL model) should not exceed 6. For lipophilicity, XLOGP3 should be in
the range from −0.7 to +6.0. For flexibility, the molecule should not have more than 9 rotatable
bonds. To be estimated as drug-like, the red line of the compound under study must be fully
included in the pink area. Any deviation represents a suboptimal physicochemical property for oral
bioavailability.
Moreover, simple molecular and physicochemical descriptors like molecular weight (MW), molecular
refractivity (MR), count of specific atom types, and polar surface area (PSA) are compiled in this
section. The values are computed with OpenBabel. The PSA is calculated using the fragmental
technique called topological polar surface area (TPSA), considering sulphur and phosphorus as polar
atoms. This has proven a useful descriptor as it can quickly estimate some ADME properties,
especially with regards to biological barrier crossing such as absorption and brain access.
The partition coefficient between n-octanol and water (log Po/w) is the descriptor for lipophilicity. It is
advisable to use multiple predictors either to select the most accurate methods for a given chemical
series or to generate consensus estimation. The models behind the predictors should be as diverse
as possible to increase the prediction accuracy through consensus log Po/w. Thus, five freely available
predictive models are XLOGP3, an atomistic method including corrective factors and knowledge-
based library; WLOGP, an implementation of a purely atomistic method based on the fragmental
system of Wildman and Crippen; MLOGP, an archetype of topological method relying on a linear
relationship with 13 molecular descriptors; SILICOS-IT, an hybrid method relying on 27 fragments
and 7 topological descriptors; and iLOGP, an in-house physics-based method relying on free energies
of solvation in n-octanol and water calculated by the Generalized-Born and solvent accessible
surface area (GB/SA) model. Consensus log Po/w is the arithmetic mean of the values predicted by the
5 proposed methods.
Next, a soluble molecule greatly facilitates many drug development activities, primarily the ease of
handling and formulation. For discovery projects targeting oral administration, solubility is one main
property influencing absorption. In addition, a drug meant for parenteral usage should be highly
soluble in water to deliver a sufficient quantity of active ingredient in the small volume of such
pharmaceutical dosage. Therefore, ESOL and Ali are the topological methods used to predict water
solubility. Both differ from the seminal general solubility equation as they avoid the melting point
parameter; the latter being challenging to predict. They also demonstrate strong linear correlation
between predicted and experimental values (R 2 = 0.69 and 0.81, respectively). The third solubility
predictor was developed by SILICOS-IT. The linear correlation coefficient of this fragmental method
corrected by molecular weight is R 2 = 0.75. All three models predict log S values, which are translated
into solubility in mol/l and mg/ml. A qualitative estimation of the solubility class is given according to
the following log S scale: insoluble <−10 <poorly <−6 <moderately <−4 <soluble <−2 <very <0 <highly.
In the Pharmacokinetics section, specialized models are used to evaluate individual ADME

behaviours of the molecule under investigation. One model is a multiple linear regression which
predicts the skin permeability coefficient (Kp). Kp is linearly correlated with molecular size and
lipophilicity (R2 = 0.67). The more negative the log Kp (with Kp in cm/s), the less skin permeant is the
molecule. The BOILED-Egg model, which is an intuitive graphical classification model, consists of the
predictions for passive human gastrointestinal absorption (HIA) and blood-brain barrier (BBB)
permeation. Other binary classification models are intended to focus on the propensity for a given
small molecule to be substrate or inhibitor of proteins governing important pharmacokinetic
behaviours. The knowledge about compounds being substrate or non-substrate of the permeability
glycoprotein (P-gp, suggested the most important member among ATP-binding cassette transporters
or ABC-transporters) is vital to appraise active efflux through biological membranes, for instance,
from the gastrointestinal wall to the lumen or from the brain. The role of P-gp is to protect the
central nervous system from xenobiotics. Nonetheless, P-gp is overexpressed in some tumour cells
and leads to multidrug-resistant cancers. Likewise, it is necessary to understand the interaction of
molecules with cytochromes P450 (CYP). This superfamily of isoenzymes is responsible for drug
elimination through metabolic biotransformation. It has been reported that CYP and P-gp can
process small molecules synergistically to improve protection of tissues and organisms.
Approximately 50 to 90% of therapeutic molecules are substrate of five major isoforms (CYP1A2,
CYP2C19, CYP2C9, CYP2D6, CYP3A4). Inhibition of these isoenzymes is one major factor of
pharmacokinetics-related drug-drug interactions resulting in toxic or unwanted adverse effects due
to the lower clearance and accumulation of the drug or its metabolites. Numerous inhibitors of the
CYP isoforms have been identified. Some are affecting different CYP isoforms, while other
compounds show selectivity for specific isoenzymes. Therefore, it is essential in drug discovery to
predict the propensity with which the molecule will cause significant drug interactions through
inhibition of CYPs, and to determine which isoforms are affected.
Furthermore, drug-likeness assesses qualitatively the chance for a molecule to become an oral drug
with respect to bioavailability. Drug-likeness was established from structural or physicochemical
inspections of development compounds advanced enough to be considered oral drug-candidates. It
is routinely employed to perform filtering of chemical libraries to exclude molecules with properties
most probably incompatible with an acceptable pharmacokinetics profile. The following rule-based
filters often originate from analyses aiming to improve the quality of the proprietary chemical
collections. The Lipinski (Pfizer), Ghose (Amgen), Veber (GSK), Egan (Pharmacia), and Muegge (Bayer)
filters are implemented. Multiple estimations allow consensus views or selection of methods best
fitting the end-user’s specific needs in terms of chemical space or project-related demands. Any
violation of any rule described here appears explicitly in the output panel. The Abbot Bioavailability
Score seeks to predict the probability of a compound to have at least 10% oral bioavailability in rat
or measurable Caco-2 permeability. This semi-quantitative rule-based score relying on total charge,
TPSA, and violation to the Lipinski filter defines four classes of compounds with probabilities of 11%,
17%, 56% or 85%. It focuses on the fast screening of chemical libraries to select the best molecules
to be purchased, synthetized or promoted at a further stage of a medicinal chemistry project.
Finally, two complementary pattern recognition methods allow for identification of potentially
problematic fragments. PAINS (Pan Assay Interference Compounds) are molecules containing
substructures showing potent response in assays irrespective of the protein target. Such fragments,
yielding false positive biological output, have been identified by Baell in analysing six orthogonal
assays and breaking down the molecules active on 2 or more assays into 481 recurrent fragments,
considered as potentially leading to promiscuous compounds. SwissADME returns warnings if such
moieties are found in the molecule under evaluation. Apart from that, Structural Alert which consists
in a list of 105 fragments is identified by Brenk to be putatively toxic, chemically reactive,
metabolically unstable or to bear properties responsible for poor pharmacokinetics. In SwissADME, it
is possible to have a chemical description of the problematic fragments found in a given molecule.
This is implemented for both PAINS and Brenk filters. By applying these and other physicochemical
filters to design screening libraries, Brenk observed that most of the remaining compounds satisfy
criteria for “leadlikeness”. This concept is similar to drug-likeness, yet focusing on physicochemical
boundaries defining a good lead, for example, a molecular entity suitable for optimization. By
definition, leads are subjected to chemical modifications that will most likely increase size and
lipophilicity. As a consequence, leads are required to be smaller and less hydrophobic than drug-like
molecules. Another factor to be considered in the selection of the most promising virtual molecules
that will be synthetized and submitted to biological assays or other experiments is Synthetic
accessibility (SA). Then, the SA score is normalized to range from 1 (very easy) to 10 (very difficult to
synthetize).
Results obtained:
Molecule 1
- Consensus Log Po/w = 4.10; and water solubility is moderately soluble
- GI absorption is high; and is a BBB permeant

- Not a P-gp substrate
- Inhibits all the CYP450 enzymes except CYP1A2 and CYP2D6
- Log Kp= -5.38 cm/s
- 0 violation
- Bioavailability Score= 0.55
- 0 alert for PAINS; 1 alert for Brenk due to michael_acceptor_1; 2 violations for leadlikeness
due to MW>350, XLOGP3>3.5
- Synthetic accessibility= 3.97
Molecule 2
- Consensus Log Po/w = 5.18; and water solubility is poorly soluble

- Inhibits all the CYP450 enzymes except CYP1A2, CYP2D6, and CYP3A4
- 1 violation for Ghose due to WLOGP>5.6; 1 violation for Muegge due to XLOGP3>5
Molecule 3
- Consensus Log Po/w = 4.29; water solubility is moderately soluble based on ESOL, and poorly
soluble based on ALI and SILICOS-IT

- 0 violation
Molecule 4
- GI absorption is high; and is not a BBB permeant

- 0 violation
Molecule 5
- Consensus Log Po/w = 4.98; and water solubility is poorly soluble

- Inhibits all the CYP450 enzymes except CYP1A2, CYP2D6, and CYP3A4
- 1 violation for Lipinski due to MW>500; 2 violations for Ghose due to MW>480, WLOGP>5.6;
1 violation for Muegge due to XLOGP3>5
Molecule 6

- Inhibits all the CYP450 enzymes except CYP2C19 and CYP2D6
- 0 violation
- 0 alert for PAINS; 1 alert for Brenk due to michael_acceptor_1; 1 violation for leadlikeness
due to MW>350
Molecule 7

- Inhibits all the CYP450 enzymes except CYP2C19 and CYP2D6
- 0 violation
Molecule 8
- Consensus Log Po/w = 2.77; and water solubility is moderately soluble based on ESOL and
SILICOS-IT, and poorly soluble based on ALI
- GI absorption is low; and is not a BBB permeant

- Is a P-gp substrate
- Inhibits all the CYP450 enzymes except CYP1A2, CYP2C19, and CYP2D6
- 1 violation for Veber due to TPSA>140; 1 violation for Egan due to TPSA>131.6; 1 violation
for Muegge due to TPSA>150
- 0 alert for PAINS; 2 alerts for Brenk due to michael_acceptor_1, thiol_2; 1 violation for
leadlikeness due to MW>350
Molecule 9
- Consensus Log Po/w = 2.19; and water solubility is soluble based on ESOL, and moderately

- Inhibits all the CYP450 enzymes except CYP2C19, CYP2D6, and CYP3A4
- 1 violation for Veber due to TPSA>140; 1 violation for Egan due to TPSA>131.6
- 0 alert for PAINS; 1 alert for Brenk due to michael_acceptor_1; 1 violation for leadlikeness
due to MW>350
Molecule 10
- Consensus Log Po/w = 1.98; and water solubility is soluble based on ESOL, and moderately

- Inhibits all the CYP450 enzymes except CYP2C19, CYP2D6, and CYP3A4
- 1 violation for Veber due to TPSA>140; 1 violation for Egan due to TPSA>131.6; 1 violation
for Muegge due to TPSA>150
- 0 alert for PAINS; 2 alerts for Brenk due to aniline, michael_acceptor_1; 1 violation for
leadlikeness due to MW>350
Most of the molecules have poor to moderate water solubility, except for molecule 9 and molecule
10 having moderate to soluble water solubility. Molecules 1, 2, and 3 are BBB permeants while the
other molecules are not BBB permeants. Molecules 8, 9, and 10 are P-gp substrates while the others
are not P-gp substrates. They inhibit most of the CYP450 enzymes which will possibly lead to drug-
drug interactions, toxicity or unwanted adverse effects. In general, the molecule should have no
alert or violation, however, all molecules designed have alerts or violations. Regarding to synthetic
accessibility, they are considered moderately easy to be synthesized. Therefore, I suggest that all 10
molecules are not suitable for future drug discovery research.
SwissTargetPrediction is an online tool to predict the targets of bioactive small molecules in human
and other vertebrates. This is useful to understand the molecular mechanisms underlying a given
phenotype or bioactivity, to rationalize possible side-effects or to predict off-targets of known
molecules. The predictions are performed by searching for similar molecules, in 2D and 3D, within a
larger collection of 376 342 compounds known to be experimentally active on an extended set of
3068 macromolecular targets. The probable protein targets are predicted by the dual score ligand-
based reverse screening of the query molecule towards the collection of known actives. A
Combined-Score higher than 0.5 predicts that the molecules are likely to share a common protein
target. The probability values are calculated from the Combined-Scores of the most similar
compounds to the query molecule known to be active on a given protein.
Table 1 shows the probability of molecule 1 to 10 toward the protein targets.
Molecule Target Probability

1 Ornithine decarboxylase 0.14853177
Nitric oxide synthase, inducible (by homology) 0.140333098
MAP kinase-activated protein kinase 5 0.123936621
2 Protein kinase C delta 0.115736675
Nitric oxide synthase, inducible 0.115736675
4 Nitric oxide synthase, inducible (by homology) 0
Ornithine decarboxylase 0
Protein kinase C delta 0
5 Ornithine decarboxylase 0
Protein-tyrosine phosphatase 1B 0
MAP kinase-activated protein kinase 5 0
Protein kinase C delta 0.095663487
7 Nitric oxide synthase, inducible 0.106165761
Ornithine decarboxylase 0.106165761
8 Nitric oxide synthase, inducible (by homology) 0.106542926
Ornithine decarboxylase 0.106542926
Angiotensin-converting enzyme (by homology) 0.106542926
Neprilysin 0.106542926
Endothelin-converting enzyme 1 0.106542926
Table 1: The probability of molecule 1 to 10 towards the protein targets.
Importantly, the probability values precisely illustrate the probability for a bioactive molecule to
have a given protein as target, but not the probability of being bioactive. From Table 1, 0.14853177,
0.140333098, and 0.123936621 are the top three probability values for molecule 1. Molecule 3 has
probabilities of 0.123936621 and 0.115736675 while molecule 6 has probabilities of 0.104440651
and 0.095663487. Besides, molecule 2 and molecule 7 have probabilities of 0.115736675 and
0.106165761 respectively. Molecules 8, 9, and 10 have same probability values of 0.106542926. On
the other hand, molecule 4 and molecule 5 shows no binding affinity to any target.
The probability values and targets vary due to the differences in the structural side chains or
substituents. Majority of the molecules (molecules 1, 3, 6, 7, 8, 9, 10) have binding affinity towards
the Ornithine decarboxylase, followed by MAP kinase-activated protein kinase 5, MAP kinase-
activated protein kinase 2, and then Nitric oxide synthase, inducible. Predictions obtained from
molecules active on homologous proteins are tagged ‘by homology’.
In summary, all 10 molecules are having low binding affinities, around 9.5% to 14.8% towards the
protein targets, revealing a very low selectivity effect. Therefore, I suppose that all 10 molecules are
not suitable for future drug discovery research.

Discussion

Uploaded by

Copyright:

Available Formats

Discussion

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Discussion

Uploaded by

Copyright:

Available Formats

SwissADME is a free web tool applied to evaluate the ADME (Absorption, Distribution, Metabolism,

and Excretion) behaviour, including physiochemical properties, lipophilicity, water solubility,

In the Pharmacokinetics section, specialized models are used to evaluate individual ADME

- Consensus Log Po/w = 4.10; and water solubility is moderately soluble

- GI absorption is high; and is a BBB permeant

- Consensus Log Po/w = 5.18; and water solubility is poorly soluble

- GI absorption is high; and is a BBB permeant

- GI absorption is high; and is a BBB permeant

- GI absorption is high; and is not a BBB permeant

- GI absorption is high; and is not a BBB permeant

- GI absorption is high; and is not a BBB permeant

- GI absorption is high; and is not a BBB permeant

- GI absorption is low; and is not a BBB permeant

- GI absorption is low; and is not a BBB permeant

- GI absorption is low; and is not a BBB permeant

Table 1 shows the probability of molecule 1 to 10 toward the protein targets.

Molecule Target Probability

You might also like