Experience the power of conversion leads optimization for Animal science with airSlate SignNow

How to create outlook signature

in this video we're going to look at the phase of drug discovery called lead discovery and optimization so lead finding and optimization come immediately after target selection you've chosen your target and now you have to find compounds that will affect the target in the way you want let's think about one of the examples that we looked at in another video we're looking for a drug that will block angiotensin receptors how do we do this we need to find a compound that's going to act on our target and there aren't many that are going to do that if you think about it so that means we need to be able to screen tens of thousands of compounds so however we do it we're going to be able to handle this number of compounds so that means we need to have a readout something which tells us whether our drug is actually doing anything to the target we need a model that is some sort of assay how am i actually going to find out whether the drug acts on a target and i need compounds i need lots of compounds to act and to test them i'll just take these off so what kind of assay can i do there are things like enzymatic assays where you measure the output of an enzyme you have a membrane response such as voltage changes you might look at how a cell responds or you might look at a behavioral phenotype so all of these are simply ways of trying to find out does my drug actually do anything the model needs to be simple enough to do the assay but as predictive as possible of the disease so this is essential point of your assay the model you use has got to be a good model of the disease let's say for example you're working on alzheimer's disease it's no good working on animals which take many many years to get old you want a simple model one which you can assay very quickly so here's an example of a scalable screen typically what you would do is you would use this is called a 96 well plate in industry you'll probably be talking about much bigger plates than this but you would maybe add cells to each one of this plate you would then add one compound to each of these wells so there's 96 compounds but of course you need to control as well and then you add a reagent and that reagent tells changes color if your proteins being produced so let's say for example you're looking at angiotensin receptors angiotensin respond like angiotensin might evoke a calcium response so you'd be looking for some sort of indicator of a rise of calcium and what you're looking for is of all the many chemicals that you try you don't see a response but here we have this well is a hit so we're very happy we've found a drug which actually evokes a response you want a hit to stand out just like it does in this example if you're screening 10 000 compounds and you're running against a p-value of 0.05 when you're doing your stats test that means you're going to have 200 false leads so when you screen drugs you need to have a very robust statistical method and we often use a thing called the z-score and the z-score is usually defined as the difference of the means so mu 1 minus mu 2 that is the difference between blue and pink here divided by the standard deviation so this this is what we call the z score here we've got a big z-score because as you can see the the standard deviation is quite small and the difference between blue and pink is very big so that's the kind of result that we want what compounds shall we start with well you could maybe go to a chemical supplier such as sigma and order everything they got that's probably not a good idea what we usually do is start off with standard chemical libraries drug companies often have their own libraries in-house developed ing to their own particular needs but there are some standard libraries as well so for example we have a thing called the ninz library which is the national institute of neurological disorders and stroke another one called medicines for malaria venture we also have commercial libraries there are libraries of natural compounds and there are custom made ones when we say a library what we mean is a collection of hundreds of thousands of chemicals and we're going to try them all against our target so that's how we find our lead now just going back when we got to this situation here where we got our hit what we don't know is is this the best chemical to use so you would look up and see which chemical affected it and you might find it has a certain structure so now just making this up now some sort of chemical structure and what you don't know is okay this works but maybe it's quite a weak effect and maybe without a particular part of the chemical they would actually do better so then what you would do is once you've got your hit you then create a new library of related compounds that are similar to it but differ in slightly different ways and that's called lead optimization it's finding the structure which not only works but actually works best so that's how you develop leads against a target but there's another aspect to this that's a bit surprising so what we've been saying is that we choose your target you create your assay and then you scream compounds this is target based screening but you know there is actually another way of doing this which isn't which it completely bypasses using a target altogether and that's called phenotype based screening and it does it completely the other way around you start off with a model of the disease so for example they might be mice with alzheimer's i can't spell it but there you go got the idea alzheimer's disease you then take the animal model and you screen them with thousands and thousands of compounds and you figure out which compound actually cure the disease once you've done that you know which compound it is you then use other methods to find out which target that compound was acting at now this wasn't a very good example because screening millions and millions of mice is quite difficult especially with a behavioral screen like this but it is possible and it is being done or you could use invertebrates such as nematode worms such as c elegans why would we bother doing it this way well the reason is target-based screening is actually has a very low yield for new first-in-class drugs in other words it doesn't actually discover new drugs very effectively phenotypic screening by contrast is actually much better at finding brand new drugs but it is very slow and it doesn't work very well for high throughput screening that's because it's really difficult to measure the behavior of a million mice compared to measuring the behavior of a million cells but all this is changing because we've got better computer algorithms for measuring behavior we got new molecular toolkits and the evidence is that between 1999 and 2008 28 first in class drugs actually came from phenotypic screening and only 17 came from target-based screening let me give you a real example of how phenotypic screening works this is to find drugs for spinal muscular atrophy the animal model here is actually c elegans and c elegans is a one millimeter long worm which is actually really easy to work with because you can make mutants in just a matter of days in fact you can buy a mutant of every gene over the male and get it delivered to you in the post this is what the mutants look like they move very slowly and so what was done i say we because i was involved in this research you take these worms you put them into 96 well plates and you count automatically how often they swim from side to side sma worms and the black bars swim more slowly than the wild type so we're looking for a drug where this distance disappears and then what we found was that when you screen one particular library called the ninz library we found one which is four amino pyridine which actually rescued the effect so it stopped the worms from having muscular spinal muscular astrophy