1st Master Biomedical Sciences
Pharmaceutical Medicine
Possible questions 2021
Questions of which 1 will be part of the examination
LESSON: Drug discovery and design
1ste Master Biomedical Sciences / Pharmaceutical Medicine / questions / version 2021 1
,1. Discuss the strengths and limitations of the two different approaches to drug discovery. Give
some examples – Bespreek de sterkten en zwakten van de twee verschillende wijzen voor de
ontdekking van geneesmiddelen. Geef ook enkele voorbeelden.
Phenotypic Drug Discovery
This approach is an empirical, holistic method that is based on observable characteristics (read-
outs) of an organism (cell, animal). Holistic means that you focus on an organism or on the cell
and not on a spec target.
Positive
(1) You don’t need any knowledge about the patho-mechanistic details, it is a target unbiased
approach. How your compound is leading to the effect is of no concern. (2) It has been claimed
that this increases the change on first in class medication (with new mech of action). (3) In vivo,
so indirectly ADME of a compound is being tested.
Negative
(1) There is no target involved, so there is no knowledge of the target, that makes it difficult to
optimize the structure. (2) This whole approach is low throughput, you can test a few compounds,
but definitely not 1000 of them.
Examples
In vitro antiproliferation assay (cells).
This is based on the fact that tumor cells would proliferate in an uncontrolled way, so investigators
isolate tumor cells from biopsies coming from tumors that are growing in pats, then those cells
proliferate in an uncontrolled way, plastic support are used to let them grow on (96 well plate).
Then you can expose those cells at this stage to compounds. To what extent inhibits this
compound the proliferation of these cells. Those cells are metabolically active and you can
expose those cells to a compound, a colourless one, and the cells will convert this compound into
a colourful compound, a stain, you can extract the stain, the more compound was used, the less
proliferation was obtained, so the less colour developed.
In vivo antitumoral assay (animal).
Animals are used, cost will go up dramatically. You start with an in vitro assay to kind of select a
tumor cell line which is sensitive toward the compound, then inject those tumor cells under the
skin of nude mice, they are immune deficient, under the skin you see the tumor cells who
proliferate and finally you will see a bump appearing under the skin. With a caliber you can follow
the growth of this tumor as a function of time, the growth of tumor is not as fast when a compound
is given every day. They also follow the bodyweight to see if it is a toxic compound.
Target-based Drug Discovery (TDD)
This is a rational, molecular method based on knowledge of the patho-mechanism, this
knowledge comes from fundamental biomedical research (often academic), universities over the
whole world publish a lot so data becomes public available and then from this data a target can
be defined.
Positive
(1) High throughput and monoclonal ab approach are possible. This method is rational because
you know the target, you have some background about it (intellectually attractive).
1ste Master Biomedical Sciences / Pharmaceutical Medicine / questions / version 2021 2
, Negative
Is the target validated and relevant in a larger context?
Examples
In vitro kinase assay (enzyme).
Biochemical assay, ELISA procedure, you have a peptide and add this to a 96 well plate, you let
it overnight adsorb to the plastic wall, then add your kinase enzyme + ATP, peptide will be
phosphorylated by the kinase, you do not see this. Later you add something for visualisation, add
ab that recognises the phosphorylated (P) peptide, linked to this ab there is an enzyme, HRP
horse radish peroxidase), this enzyme will convert a compound (TMB reagent) into a colourful
compound, the more P peptide, the more intense the colour. Add compound, inhibitor of kinase,
amount of colour will be lower, then you know that your compound of interest has a kinase
inhibiting activity.
In vitro kinase assay (reporter cell line) → see question 2.
In silico drug design (PC).
Design by computational method, it is possible to crystallise a prot in spec conditions, if those
crystals then are analysed by X ray diffraction methods you can get a detailed structural
information regarding the conformation of the prot. You can also co-crystallise an inhibitor
together with a prot and so you can design your molecule of interest.
2. Discuss the PathHunter cell-based assay – Bespreek de PathHunter cel-gebaseerde test.
= in vitro kinase assay (reporter cell line) = enzymes present in a cell
This system is called Eurofins: enzyme fragment complementation assay technology.
Tyrosine kinase R are located at the membrane, those are engineered prot and at the cytosolic
site you have a prot kinase (PK). In the cytosol there is also another prot present (EA = enzyme
acceptor), this is a fusion prot with SH2 (= spec domain of adaptor prots).
If those 2 prot are not linked but separated then there is no beta-gal act present.
If those R are activated those 2 prots will come together and auto-P each other, these P sites will
be recognized by the SH2 domains and the EA + SH2 prot will locate close to the cellular
membrane, close to the R and it will come in contact with the PK prot.
So a functional beta galactosidase enzyme is formed. Lugal is added to see this happening. B-gal
is going to split off aglycon and D luciferin is formed. Then add luciferase, another enzyme + ATP,
this enzyme will convert luciferin into the oxidized form, and by doing so light is emitted. This is a
chemiluminescence assay and the light can be picked up.
More light = more luciferin = more B gal activity = more of the P tyrosines.
If you add a compound to the cell that would inhibit this P you would see less light.
1ste Master Biomedical Sciences / Pharmaceutical Medicine / questions / version 2021 3
,3. Discuss the different types of validity in drug discovery – Bespreek de verschillende typen van
validiteit in de ontdekking van geneesmiddelen.
Face validity: quality aspect which is referring to the degree of similarity between what the model
appears to show/measure (at the surface, symptoms) and what it claims to show/measure. Eg It
is not guaranteed that cells in 2D system show the same type of mechanistic background as in
case of the clinc. This can also be an obstacle for developing models using lower phylogenetically
animal species such as zebrafish and drosophila, but low face validity does not per se invalidate
a model!
Predictive validity: quality aspect referring to the ability to correctly identify the efficacy (and
safety) of a putative therapeutic. No false neg and pos and this is the most important criterion for
drug screening models. When chain of translatability is present, this means that what you find in a
model state can be translated to the disease state. 2D culture, 3D culture and 3D co-culture go
from low to high predictive validity. If you have results with compounds using the 3D culture, it is
much closer to the real situation and therefor the results you obtain are more relevant to the clinic,
the non-clinical proxi is better.
Construct validity: quality aspect referring to the degree of similarity between the (hypothesized)
mechanism(s) underlying the pathophysiology present in the model (below the surface) and the
human situation. This is the most important criterion for models. Possibly high construct validity
with low face validity. If there is a shared mechanistic basis available between the preclin disease
model (assay) and the human pathophysiology, a complete as possible overlap between what is
happening in those cells present in a model and what is happening in the cells present in the
clinic.
Predictive validity and construct validity are kind of the same → yes because if the construct
validity is high the predictive validity will also be high.
Example of the validities see slides 48 to 53 (Dravet Syndrome)!
4. Discuss the way in which and the importance to validate a target in drug discovery – Bespreek de
manier waarop en het belang van het valideren van een target in de ontdekking van
geneesmiddelen.
You need to make sure that when you hit a target you get the result that you expect.
Problem validation target; when an engine does not work anymore, you can replace the part that
is defect, BUT a cell is far more complex, we talk about 10 000-100 000 of interactions, so
basically what you are doing with target approach is there is a problem with this cell, tumor cell
and here is a possible target, if I inhibit that I will reverse the disease state of this cell, there is a
possibility that this strategy works but you should know that there are a lot of compensatory
mechanisms present in those cells, so if you hit one target, the cell might compensate this by
focusing on more other targets, so it is not because you inhibit one target that your strategy is
going to work!
1ste Master Biomedical Sciences / Pharmaceutical Medicine / questions / version 2021 4
, Loss-of-function vs inhibition;
So before you start searching for a selective inhibitor in a very early phase of drug discovery
work; this is going to take time, inhibitor needs to be non-toxic, selective, only interact with that
prot not with other prot, so before you do this there are a few simple tricks that you can use to
understand whether your target is indeed a good target. Nowadays with crispr-cas9 you can quite
easily get a very spec KO of a gene of interest, a gene (DNA) that ofc stands for the final
production of this prot. Wit small interfering RNA or antisense you can also interfere with mRNA
(KD). The end result is that in this cell a spec prot is no longer expressed, you can do this in a
couple of months in time, if it then is true that the prot is a good target for pharmacological
inhibition then you should indeed see reversal of pathological phenotype and then you can argue
that an inhibition will have a similar effect, this means that your target is a good target.
Assumption is that if you see a reversal of the pathological state of your cells by removing the
prot, that this is the same as the situation where the prot is present but the prots activity is
inhibited. This is easily done but finding a selective inhibitor is quite a bit of work and requires a
lot of money. So before you start to look for compounds, first validate your target, this is essential
LESSON: Target deconvolution
5. Discuss 2 target deconvolution strategies in drug discovery. – Bespreek 2 target identificatie
strategieën in de ontdekking van geneesmiddelen.
Affinity chromatography-based methods
Ligand = compound
Target = large molecule, prot
Affinity = a ligand or compound of interest shows affinity for a target (Vanderwaals forces or
hydrophobic forces most of the time) and docks into a spec part of the target.The optimal value
for the dissociation constant Kd is between 10-7 to 10-15 M, Kd = [L][P]/[LP]. A complex is formed,
the more complexes formed the more affinity the ligand shows for the target or vice versa. This
situation is an equilibrium: L + P ↔ LP and the conc of the ligand can be calculated.
If you want to look for a target in a mixture of targets then the optimal situation is where the prot is
present in high concentration, then it is easier to find this target. But that high conc target prot will
obscure binding to lower-abundance proteins (if diff targets), because it might be that your
compound binds to a principle protein but also to other low abundance prot. More difficult to find
your target of interest if there are many targets because of low abundance. A chemical
modification with linker (binding to carrier) might inactivate your drug.
How does it work?
You have a column and you are looking for the target of a spec compound, which is covalently
bound to parts of that column. A cell homogenate is made and all those prots are added to the
column = loading step, afterwards wash the column = elude, prot that does not show any kind of
affinity for the compound will elude from the column. The ones that bind, are the target of interest.
Afterwards the pH can be changed so that the affinity is decreased of the target-compound so
that the target eludes from the column. All the fractions that you obtain are put on an SDS page.
To identify the prot of interest, you can cut out the band, digest the prot so that you end up with
peptides, inject them in liquid chromatography system with mass detection and by detecting and
identifying the masses of those peptides you would be able to identify the prot of interest.
1ste Master Biomedical Sciences / Pharmaceutical Medicine / questions / version 2021 5
,Also beads can be used to immobilize the ligand. In real life it is not so easy to know after the
SDS page which band if the target, because also some other prots will bind to your compound so
as a result you get an SDS page with multiple bands visible. A few tricks exist to identify the spec
target: (1) you can use a ligand analogue (inactive ligand) and (2) use an excess of ligand not
coupled to beads and then there is a competitive mechanism that occurs and it again gives
another pattern. Via these and a few other tricks you can decipher the situation and you might
select the prot of interest which is indeed interacting with your ligand.
Expression-cloning-based methods
This is also based on affinity, but it is a cell system that is being used. The advantage here is that
there is no low target prot abundance, the cells you are going to use will contain all possible prot
targets in approximately the same amount. Disadvantage is that if you for example use yeast cells
that there will be no or different post-translational modifications of your prot of interest (target) and
this might also lead to diff prot folding, so the target is not the same as present in mammalian
cells. But you can also use mammalian cells to solve this problem.
Three-hybrid system: this cellular based system is composed out of three different parts.
Part 1 - fusion protein (expressed in all cells)
Fusion = composed out of 2 different prots, but expressed as one. You have a DNA-binding
domain eg LexA operator and a ligand-binding domain eg DHFR (dihydrofolate reductase). MTX
(methotrexaat) eg binds DHFR with very high affinity. So via the DNA-binding domain the fusion
prot will find its way to the DNA. It is important that all cells you use (mammalian or yeast) will be
transfected with the gene that expresses this fusion prot so all cells will express this.
Part 2 – linker system (small molecule linked to MTX)
The purple square in the picture below is MTX, then you have the linker and this can be dextrane
or PEG. [MTX is typically used as an anti-tumoral drug and has as target DHFR, by interfering
with this reductase MTX will inhibit it, so in the clinic this results in an anti-tumoral effect.] MTX
makes sure that this yellow oval finds its way to the DNA. MTX on one side with the linker and
then the yellow oval = the molecule of interest for which you try to find a target. If you add a
certain amount of this system, so MTX with linker and molecule of interest, a part of those
molecules will be taken up by the cells and MTX will find DHFR and will drag along the yellow
oval so it comes very close to the DNA.
Part 3 – fusion prot (expressed in individual cells)
Another fusion prot is being expressed in the cells. This prot also exists out of two moieties, on
the one hand you have a transcriptional activation domain that will express some gene
(expression of lacZ, beta-galactosidase). The other part is a prot that comes from a
complementary library (cDNA library = potential target). The idea is that each cell which is present
here will express a spec prot from the library that is linked to that act domain. The ideal situation
is that you just have one target prot expressed. If this light green prot is expressed and is that
target for the yellow oval, they will interact (see bottom situation). In that way also the activation
domain will get very close to the DNA and there you have a reporter gene present. Molecule of
interest gets very close to the DNA by means of MTX, the target also finds the molecule of
interest and in that way the act domain gets close to the reporter gene. That gene will express
beta-galactosidase. If you then expose the cells to X-gal (colourless), beta-gal will split off
galactose and the other part 5-bromo-4-chloro-3-hydroxyindole is released, this molecule will
undergo spontaneous dimerization and oxidation and in that way form a blue stain.
1ste Master Biomedical Sciences / Pharmaceutical Medicine / questions / version 2021 6
,Afterwards the blue clones can be isolated and then you can see with what that clone was
transfected, with what kind of target prot and so you know the target of the compound of interest.
EXTRA from that ppt: from slide 13 onwards
Hit to lead
HIT = interaction with or activity on target.
LEAD = drug-like characteristics = acceptable ADME.
Assume that we started with TDD, a robot which does 1000 and more assays per week with a
chemical library to find some compounds that show activity, if we find a hit we know that a
compound was found that interacts with the target. Is this a new drug that will arrive on the
market? Maybe or maybe not, because it does not automatically imply that the compound can
been given orally and that it will distribute over the body and find the target in the body. It can be
that it is eg not absorbed, so it has act in the artificial environment but maybe no act in the in vivo
clinical environment.
Absorption: you have to make sure that the compound when it is given orally will be absorbed
from the intestinal tract into the bloodstream.
1ste Master Biomedical Sciences / Pharmaceutical Medicine / questions / version 2021 7
,Distribution: distributed over the tissues and the organs.
Metabolism: exposed to the liver, there you have all kinds of enzymes that will metabolise your
compound, most of the time that inactivates your compound and it has not reached the target yet.
Excretion: it can be that the excretion is extremely fast, that the half life of your compound is only
10 minutes, in that time the compound of interest is not going to be able to distribute over the
body. Taking then a lot of pills is not ideal, you need a half life which matches the expectations
(once a day eg).
Discovery testing in TDD.
(1) Funnel figure (see above); the top of the funnel is the screening and then you come to a hit
triage.
They look at:
Potency, IC50 value = inhibitory conc, the more potent a compound is, less likely it is going to
interact with other prot. If more potent then it has a higher affinity for spec prot.
Physicochemical characteristics (drug-like? Acceptable ADME after oral administration), the rule
of 5 will be applied (Ro5).
RULE OF FIVE → Lipinski
Without doing any kind of experimental work, you just look at the structure and then you can
predict whether your compound has an acceptable solubility, absorption and permeation.
< 5 hydrogen bond donors (O-H, N-H)
< 10 hydrogen bond acceptors (O, N)
< 500 Mr, molecular weight
log P < 5
4 rules, but they are all a plural of 5. If a compound or a chemical class does not obey these
rules, they can already be excluded. You did not spend any money here, you just looked at the
structure.
Intellectual property IP (freedom to operate), it is possible that you found compounds which are
already used by other companies.
Practical aspects (synthesis, costs, scalability?), when you are screening for activity you use a
library with 1000 and 1000 of compounds, but in that library the compounds are only present in
amounts of 2 to 5 mg (lab scale), but if your compound needs to be synthesised in grams preclin
or kg clin application, then you have to upscale the chemical synthesis, it is possible that there
are major problems with upscaling.
(2) Hit to lead (100s of compounds)
Synthesis of focused libraries: based on first triage you will find chemical classes that are of
possible interest and then people will start synthesising focused libraries with all kind of
analogues of a first hit. There will be an exploration of the SAR for target binding (structure-
activity relationship) and of the IC50 values of the compounds to make the libraries better and
better, the idea is to synthesise 100s of compounds. Use Tier I ADME experiments to see if it is
drug-like and safety tests on selectivity are also being done in vitro.
1ste Master Biomedical Sciences / Pharmaceutical Medicine / questions / version 2021 8
,(3) Lead optimization (dozens of compounds)
After previous phase you get a good idea about the act of the molecules, what molecules would
be more toxic, you know the SAR etc. Synthesis of focused libraries again of dozens of
compounds. Do Tier II ADME experiments and in vivo efficacy assays.
In vitro, Tier 1 (1-2 mg x4)
You do simple, relevant tests and not to expensive to perform.
Log P (partitiecoëfficiënt), log D (distributiecoeff): determines solubility, absorption, membrane
penetration, plasma protein binding, distribution, CNS penetration, excretion route and interaction
with targets. In the very early phase you want to select compounds with log P less than 5 (see
Ro5), but now you want to determine is experimentally because the calculated log P is not always
very accurate.
How? You will use two solvents, one is water and the other one is octanol (liphophilic). Bring
those two solvents together (it will not mix, bilayer system) and add your compound, shake, let it
go back to 2 layers, then determine the conc of your compound in octanol phase and in the water
phase [octanol]/[water]. This gives you a good indication of the lipophilicity of the compound.
Water-solubility: determines bio-availability. If you take a pill and your compound does not
dissolve at all in the liquid of the stomach or GI tract, it will pass the tract completely unabsorbed.
Microsomal stability: stability or conversion by liver microsomes (ER) with metabolizing enzymes.
If a compound is very rapidly converted in inactive metabolites, compound of interest is not going
to be active. In vitro predict this; from rat liver you obtain the microsomes by centrifugation
methods, ER of the liver cells (represents a high amount of metabolizing enzymes), incubate in
phosphate buffer (compound + enzymes) and after 10 to 30 min you examine to what extent your
compound has been metabolised, if completely metabolised so low stability your compound might
be an interesting hit but is not a good lead.
Plasma stability: stability in presence of plasma enzymes (esterases).
In vitro, Tier 2 (1-2 mg x3, 5-7 mg)
Plasma protein binding: determines free fraction (active!). Albumin is a plasma prot with whom a
lot of compounds will bind. If compound of interest binds with a very high affinity to this albumin
then the free fraction (what moves into tissues and finds the target) is too low and the compound
of interest is never going to reach the cells of interest, then you have to change your compound,
modify it chemically so that the plasma prot binding is lower.
Hepatotoxicity test: with primary hepatocytes (and possibly other cells). You want to know
whether your compound is toxic to cells as a whole.
CYP450 inhibition profile: as in case of microsome-stability testing, but in presence of compounds
that are selectively metabolized by CYP1A2, 2B6, 2C9, 2D6, 3A4 (drug-drug interaction). Those
are enzymes which are for a big part responsible for liver metabolism. Change that the compound
that you bring on the market will compete with another compound for metabolization by CYP is
very high. So we want already avoid at this stage that the compound is going to interact with
typical medication which is on the market that the pat will possibly take as well.
1ste Master Biomedical Sciences / Pharmaceutical Medicine / questions / version 2021 9
, Permeability through PAMAP (parallel artificial membrane permeability assay) or cellular
monolayer: determines intestinal permeability and oral bio-availability. With the artificial
membrane; bring your compound of interest in the blue zone, take samples from the green zone
and assess the permeability over the membrane (see figure left), this shows that your compound
penetrates well or does not penetrate at all. Or use a caco cel (carcinoma colon cell line); human
cell line that differentiates in the typical epithelial layer as you see it in the intestinal tract, basically
the same system, you have a compartment which is an apical compartment and then the
basolateral compartment, also used to understand the absorption capability of your compound of
interest (see figure right).
In vivo, Tier 1 (5-50 mg)
RACE = rapid assessment of compound exposure, we are going to give a single dose of
compound of interest po or ip to animals (5-50 mg/kg, n = 2-4 animals). The molecules are
being compared to gain insight to what extent you have this permeability/bio-av in an in vivo
situation and compare po with ip dosing, so here we do a comparison between molecules
and/or formulations.
In vivo, Tier 2 (10-100 mg)
If some compounds are selected we go for a more extensive PK analysis = the way
molecules of pharmaceutical interest move in the body and are finally excreted. This is
typically done like this; you give 1 mg/kg iv (= 100% dose) vs 2 mg/kg oral and if you see
that the final conc (after x time points you take blood samples and determine the conc) of
your compound of interest given orally is 20% as compared to iv, you can conclude that the
oral avail is rather low, you want to have 60-80%. Meanwhile also the half life is being
determined, time you need to reach half of the original conc, if low, then your compound
does not have drug like characteristics (n = 6-12 animals (rat), analysis; 5, 15, 30 min, 1, 2,
4, 6, 8, 24 h). Using liquid chromatography with mass detection you can also find what the
metabolites are (earlier in vitro, now in animals, to see what kind of metabolites would be
formed in humans).
1ste Master Biomedical Sciences / Pharmaceutical Medicine / questions / version 2021 10
