Preclinical Drug Development
Preclinical development encompasses all the steps up until you first test your candidate drug in patients or healthy volunteers. The general steps are (not necessarily all sequential):
Target identification - what are you going to drug?
Target validation - does modulating this target have a beneficial effect in your disease of choice?
Hit identification - The initial suite of candidate molecules to drug your target
Lead identification - The lead molecule from the suite of hits.
Lead optimization - The optimization of the lead into its final form to be used in humans
Non-GLP efficacy - Studies justifying your therapeutic hypothesis
GLP safety & tox - Safety and efficacy studies, including provisional dosing
Investigative New Drug application - the application to the FDA to allow you to test your drug in humans.
Target Identification & Validation (?)
Your target is the biological molecule you are hitting and modifying with your drug. Targets can include cell membrane receptors, intracellular proteins.
Hit identification (approx 3-6 months including assay development, $200k-$300k fully loaded)
A ‘hit’ is a molecule which binds your drug with reasonable affinity. You will get potentially hundreds of initial hits out of a high throughput screen, depending on the size and composition of the library. Your final hits are molecules with confirmed structure and desirable activity against your target.
The strength of a binder is quantified by its binding affinity. A lower number means a stronger binder. A strong binder will bind on the nanomolar (10^-9) range; an exceptionally strong binder would be in the picomolar (10^-12) range. Understanding the biology of your target is important here: how strong does it bind its cognate binder (receptor, ligand, another protein, etc) - how strong of a binder do you need to find to block your target’s action?
Binding does not necessarily mean inhibiting. Therefore, it is important your assay is well-designed to allow a differentiation between
A high throughput screen is a screen that is able to be conducted quickly to test a large number of molecules against a biological target. In comparison to a low-throughput screen, which may be done manually or on more complex assays,
A library is a set of starter molecules which are routinely screened against a target. Most libraries have on the order of 300,000 to 500,000 molecules; this number may be lower for specialized libraries (e.g., generics, specific classes of inhibitors, etc.) or higher for synthetic, diversity libraries, or libraries owned by large corporations. Generally speaking, a good library will have a diversity of molecular structures to ensure that the best chemical structure for inhibiting your target is able to be identified.
An assay in the context of your screen is the simple biological test you use to determine whether you have found a molecule which binds and has the desired effect on your target. Your screen may be functional (assays for the desired functional change of a binder on your target) or binding (assays simply for the binding of a drug to your target). Considerations in your assay include:
Pharmacological relevance of the assay
Reproducibility
Costs
Quality & predicted signal window
Z’ factor: value between 0 and 1; >0.4 considered ‘robust’
Includes variables such as # of steps, washes, stability of reagents
Effect of solvent on the assay
Types of screens
A high throughput screen is one of many types of screens one can use to identify hits, including:
Focused screen - specific biological area e.g. kinases
Fragment screen - building block molecules to combine into a stronger hit (individual hits often in the mM range) Requires a crystal structure
Structural-aided drug design - use of the protein crystal structure to intelligently design molecules. Often used as an adjunct
Virtual screen - docking models of virtual molecules from a library; often informed by known binders
Physiological screen - lower throughput, tests the whole cell/tissue MoA and often target agnostic
Phenotypic screen - can be high throughput, looks for specific phenotypic response and by definition target agnostic.
ML/AI - various methods here
Lead Identification (3-6 months, $200-$400k)
Hopefully after your screen, you will have a diversity of candidates which you can then send to your medicinal chemists to determine the best base structures off which to build your therapeutic.
The purpose of lead identification is to choose from across all your hits the best structure to build your lead. You should consider a number of factors:
The patentability of the structure and its chemical space
Known structure activity relationship (SAR) - the identification of the chemical group(s) which are responsible for the biological activity against your target. This is important for understanding how and where to modify your drug when you begin to optimize.
The dose-response (EC50) - the concentration of the drug that gives half of the maximal possible response
The removal of PAINS/frequent hitters “pan-assay interference compounds”, compounds which appear to bind but are non-specific, are messing with the assay, or otherwise appear to be hits but are not.
Lipophilicity of the compound is the “fat-liking” of the compound. Lipophilicity is correlated with other drug characteristics that are incredibly important, such as solubility (the ability of the compound to dissolve in solvent, important for creation formulations and the ability of the drug to dissolve in biological fluids and have a therapeutic effect), permeability (the ability of the compound to penetrate the cell membrane, which is a lipid bilayer), and metabolism (xxx)
Lipophilicity is measured by LogD7.4 and varies from <1 to >5. Generally speaking, <1 means that the drug is the most lipophobic, so it has high solubility but low permeability and metabolism; the highest end of the range is not soluble but easily crosses cell membranes.
A good solubility is > 5ug/mL, ideally higher than 50 ug/mL.
The PD/PK/ADME of your drug (detailed below)
The binding and inhibition of cytochrome P450, liver enzymes which generally are responsible for 75% of drug metabolism.
P450 inhibitors are drugs which do not use the P450 pathway for their metabolism; these drugs increase the concentration of other drugs
P450 inducers use the pathway for their metabolism; these drugs decrease the concentration of other drugs
Generally, you do not want your drug to inhibit the CYP at a concentration less that 10 uM.
The minimum target product profile specific to your drug. There are a number of considerations in your TPP, including a number of PK/PD/ADME
During your hit-to-lead process, you will also consider the drug metabolism and pharmacokinetics (DMPK), also referred to as the ADME (absorption, distribution, metabolism, and elimination) profile of your hits. A couple of important definitions:
Pharmacodynamics (PD) is the physiological response of the body/tissue to the drug
Pharmacokinetics (PK) is the movement of the drug within the body
Absorption is the process by which the drug goes from its delivery mechanism (often oral for small molecules) to the bloodstream/its first biological tissue
Distribution is the reversible transmission of the drug from one compartment of the body to another e.g. the blood to a specific tissue
Metabolism is the breakdown of the compound, usually by liver enzymes, from its original structure into metabolites
Elimination is the irreversible removal of the drug from the bloodstream, usually via the kidneys (urine)
The PK/PD/ADME needs of a drug vary widely on the indication, patient population, target, etc. Important variables here are:
Oral bioavailability (F) = (AUC after an oral dose)/(AUC after equivalent iv dose)
Half-life (T1/2)
Volume of distribution (Vd) = Amount of drug in the body (D)/drug concentration in plasma (C)
Clearance (Cl) volume of blood that the drug is removed from over a specific unit of time. Cl = dose/AUC
Area under the curve (AUC) the amount of drug that survives in systemic (blood) circulation
Cmax - max concentration achieved in plasma after dosing
IC50 - the concentration of the inhibitor that gives a half-reduction of the response/binding of its target
EC50 - the concentration of the drug that gives half of the maximal response
Lipinski’s Rule of Five
Lipinsky’s rule of five is a short-hand for describing the ‘drug-likeness’ of a compound
MW < 500 Da
LogP < 5
octanol-water partition coefficient, the solubility of the compound in two immiscible phases/solvents, one water one hydrophobic, measuring the hydrophilicity vs the hydrophobicity of the drug.
High ratio of octanol/water are lipophilic (found in lipid bilayers), low ratio are hydrophilic (often in blood)
< 5 H-bond donors (sum of OH and NH groups)
< 10 H-bond acceptors (sum of N and Os)
Other considerations and relevant assays
Caco-2 permeability - gut cell line to assay potential intestinal barrier crossing (for oral delivery of drugs)
MDR1-MDCK permeability - permeability testing transporters relevant to intensional and CNS permeability
BBB permeability - in vitro assays possible to determine likelihood of crossing the tight junctions characteristic of this tissue
Hep G2 tox - surrogate for liver tox
Ames test - mutagenicity (genomic)
Cell line tox - general cytotoxicity
Lead Optimization ($1-2M, 12-18 months)
Once you have done the above, you will select a specific hit to optimize into your lead. This is one of the longest phases of the drug development process, and is incredibly important, as once you have identified and optimized your final molecule you cannot change it without large cost and redoing the above processes. The aim of lead optimization is to retain the favorable properties of your candidate while optimizing away any undesirable effects or characteristics.
Lead optimization follows a design-test-make cycle, where you synthesize small variants of the drug are synthesized and then assayed across the primary assay, whole cell binding, cellular function, PK/PD, and then finally the disease model, often in vivo.
A strong consideration in lead optimization is the toxicity and potential AEs of the drug. You will also want to consider the therapeutic ratio/margin of safety for your drug - if the therapeutic and toxic dose are close, there may be a danger to patients due to the eventual likelihood of an accidental double dose, etc.
There are a number of standard tox tests:
Here, you will also conduct a number of animal studies to determine whether your drug has a beneficial disease affect in your model of choice, and also to better understand the dosing.