What has computer-aided molecular design ever done for drug discovery? Flashcards Preview

Topics in Drug Discovery > What has computer-aided molecular design ever done for drug discovery? > Flashcards

Flashcards in What has computer-aided molecular design ever done for drug discovery? Deck (14)
Loading flashcards...
1
Q

Examples of drugs developed from ligand-based drug design

A

Norfloxacin
Losartan
Zolmitriptan

2
Q

Norfloxacin

A

Discovered in Japan at Kyorin but licensed to Merck
First fluoroquinolone antibiotic on the market (1984)
Used for the treatment of UTIs
QSAR model for 6-, 7- or 8-monosubstituted compounds showed there was a correlation between antibacterial activity and parameters describing the steric effects of substituents at both the 6- and 8-positions
No such relationship found for substituents at the 7-position, but it was found that a piperazinyl substituent in the 7-position was promising
QSAR model predicted 6-fluoro-7-(1-piperazinyl) derivative would be 10x more active than the parent without the fluoro substituent - was actually found to be 16x more active
Parabolic relationship (draw) between log(1/MIC) and Es showed 6-fluoro-7-(1-piperazinyl) had highest activity

3
Q

Es

A

Steric parameter
Obtained by comparing the rates of acid-catalysed hydrolysis of substituted esters against a standard ester (usually X = Me)
Es = logkx - logk0 where kx = rate of hydrolysis of ester bearing substituent X and k0 = rate of hydrolysis of reference ester

4
Q

Positive Es value

A

kx > k0

Smaller X groups (e.g. F, H) have faster rates of hydrolysis than a Me group so give positive values of Es

5
Q

Negative Es value

A

kx < k0

Substituents larger than Me have slower rates of hydrolysis so give negative Es values

6
Q

Disadvantage of Es values

A

They are a measure of an intramolecular steric effect, whereas drugs interact with targets in an intermolecular fashion

7
Q

Losartan

A

Discovered at Dupont (but co-developed by Merck)
First non-peptide oral angiotensin II receptor antagonist on the market (1995)
Used for the treatment of hypertension
Computational molecular graphics were used to display an overlay of the geometries of an early lead compound (weakly potent) and the known angiotensin II structure in solution to identify similarities
This work suggested that the para position of the benzyl ring in the lead compound (draw) would be a favourable position for adding another substituent in order to increase the degree of similarity between the 2 structures in a simple space-filling sense
This led to the discovery of the biphenyl tetrazole motif
- this is characteristic of the class of active compounds known as ‘-sartans’

8
Q

Examples of drugs developed from structure-based drug design

A

Dorzolamide
Zanamivir
Amprenavir

9
Q

Dorzolamide

A

Discovered at Merck (1994)
Used for the treatment of high intraocular pressure (e.g. glaucoma)
Carbonic anhydrase II inhibitor
S-enantiomer of racemic prototype drug found to be 100x more active than R-enantiomer
X-ray crystal structures of carbonic anhydrase II with each enantiomer showed conformational differences between the two
The X-ray structure of the S-enantiomer revealed that the isobutylamine side chain was in the normally less-favoured pseudo axial conformation - but computational modelling indicated that this conformation would actually be favoured if a Me group was introduced at the 6-position, because this would decrease the energy penalty experienced by the compound upon binding to the enzyme
The increase in lipophilicity as a result of the extra Me group could be compensated for by replaced the isobutylene group with a smaller Et group
Finally, the trans-(S,S) diastereogmer was found to be the best inhibitor
(draw)

10
Q

Zanamivir

A

Discovered by GSK and other collaborators in 1999
Used for the treatment of influenza
Neuraminidase inhibitor
An X-ray crystal structure of DANA (natural inhibitor) bound to neuraminidase was analysed by computational modelling using the GRID program in order to identify favourable substitutions that could be made synthetically
It was predicted that the 4-OH groups could be replaced by a positively charged group so as to interact with a negative glutamate residue located at the bottom of an empty pocket in the enzyme active site
This 4-guanidino compound (= Relenza) is highly potent, but has a drawback of low oral bioavailability
A later derivative (Roche’s oseltamivir, Tamiflu) overcomes this problem

11
Q

Amprenavir

A

Discovered by Vertex and co-developed by GSK and Kissei (1999)
Used for the treatment of HIV/AIDs
HIV-1 protease inhibitor
Computational experiments indicated that the amide NH in the P1’ group of the substrate was not required for activity - formed the weakest H bond to the enzyme out of all those made by an early inhibitor
Comp experiments also indicated that an N,N-dialkylsulfonamide is good for binding to the ‘flap’ water molecule and acts as a scaffold for the P1’ and P2’ groups
Co-crystallisation of the enzyme with the compounds designed and synthesised in light of this knowledge showed that the bound conformations were substantially similar to the modelling predictions

12
Q

Examples of drugs developed from structure-based drug design (virtual screening)

A

4SD-101

Naluzotan

13
Q

4SD-101

A

While in clinical trials for the treatment of rheumatoid arthritis, a dihydroorotate dehydrogenase inhibitor was discovered by 4SC using their own software for molecular docking
Structures from a library of commercially available compounds were docked into the active site of the published enzyme X-ray crystal structure in order to obtain a lead compound
Virtual screening of 1000s of substituted structures, followed by the synthesis and testing of only 55 compounds, allowed QSAR analysis to identify even better compounds - activity was found to improve with increased negatively-charged and hydrophobic vdW surface area, and worsen with increased MW and volume
(draw)

14
Q

Naluzotan

A

Discovered by EPIX
Previously in clinical trials for the treatment of anxiety and depression
Selective 5-HT(1A) receptor partial agonist
A 3D model of the 5-HT(1A) receptor was generated from its amino acid sequence using computational protein structure prediction
Virtual screening of a library of 40,000 compounds was performed by docking each one into the model
Only 31 real compounds were actually synthesised in the entire discovery campaign - only took 6 months for the whole optimisation process with a team of only 3 medicinal chemists and 3 computational chemists!