Hit validation can save valuable drug discovery time and resource. Ronan O"Brien, applications r&d manager, Michael Rongner, software engineer, and Thomas Lundbaeck, science director, from GE Healthcare, explain how biophysical analyses can be used to confirm hits
Successful drug discovery programmes require a thorough understanding of disease pathophysiology, which can be obtained through target identification and validation activities. Various approaches are used to identify compounds (hits) with a potential to generate target-selective and bioavailable compounds. Once hits have been identified, their interactions with relevant biomolecular targets are studied to confirm and characterise the interactions quantitatively and provide information on binding mechanisms. This gives researchers confidence that medicinal chemistry efforts are focused on appropriate compounds, ensuring cost-effective use of resources.
For an effective complement to screening assays, the techniques employed for hit confirmation must be label-free, enabling studies of interactions between native biomolecules and unmodified compounds. As this process often involves many low molecular weight compounds with different properties, several techniques, differing in measuring principle, working range of affinities and concentrations are required. By choosing the right combination of orthogonal techniques, the dual purposes of adding confidence to decision-making and bringing complementary information to the analysis can be met.
Reliable hit confirmation can be achieved by combining data from the complementary biophysical techniques Surface Plasmon Resonance (SPR) and Isothermal Titration Calorimetry (ITC). SPR enables studies of interaction kinetics and affinities, and can also be used to determine the concentrations of a broad range of different biomolecules. Binding activity and dissociation kinetics can be measured, and the kinetic data (koff and kon rates) can then be used to determine the affinity of the interaction (KD= koff/kon).
ITC measures the thermodynamics of a reaction, i.e. the heat released or absorbed when a ligand interacts with a biomolecule. This enables the binding enthalpy (∆Hcal), affinity (KD), and stoichiometry of the interaction to be determined in a single experiment.
Kinetic & thermodynamic characterisation:
Kinetic characterisation provides information about the rate of complex formation and its stability, whereas studies of interaction thermodynamics are required to understand why the interactions occur (see box).
Kinetic rate constants are important for understanding target occupancy over time, or target residence time, a factor that may influence both efficacy and selectivity in functional assays and in vivo.1 Compounds that appear equipotent or non-selective (i.e., they have the same KD) may thus differ in functional advantages, and a thorough understanding of the rate constants is important from both an efficacy and a safety perspective (Figure 1).
Besides confirming direct interactions with the target of interest, thermodynamic measure-ments using ITC also give insight into the nature of the non-covalent forces responsible for binding. Polar interactions tend to contribute favourably to the enthalpic component, whereas entropically favoured interactions tend to be more hydrophobic. Figure 2 shows representative ITC binding isotherms for two interactions with the same affinity but with different mechanisms of binding.
Because of the strong directional dependency of polar interactions, these are generally more difficult to engineer. Disruption of existing hydrogen bonding networks or a loss of degrees of freedom leads to enthalpy-entropy compensation effects, which reduce the impact of newly introduced interactions. Thus affinity maturation during lead generation and optimisation is often achieved by adding hydrophobic interactions. This approach has the risk of increasing non-specific interactions and solubility issues.2 Thus, identifying molecules through ITC experiments that have a high enthalpic efficiency, and taking those molecules through to the next stage, could result in a reduction in these unfavourable properties while improving potency.3
Whereas ITC instruments can be used to ensure that both the enthalpic and entropic contributions to the free energy of binding are optimised simultaneously, maximising favourable drug-like properties, SPR systems can be used to distinguish between equipotent compounds based on binding kinetics. These methods are thus excellent complements in the search for drug candidates with improved selectivity and other properties. Use of two different measuring principles also serves to maximise confidence in measured affinities.
straightforward design
ITC experimental design is straightforward if knowledge about the approximate affinities is available.4 To enable a determination of KD a suitable ITC experiment will have a C value between 1 and 1000 where C = [Protein] / KD given a 1:1 binding event. When C values are too high, saturation of available binding sites occurs over a very narrow concentration range of added compound, preventing an accurate determination of KD. However, at low C values there is little binding or detectable heat.
We have developed a procedure by which KD values from SPR experiments are used to generate quality ITC data using as little protein as possible (Table 1). To maintain this balance, the "low C" method5 is recommended for the weaker interactions (last two rows in Table 1).
First, KD values determined using the 1:1 binding models with SPR instrument are imported into the Bia2iTC software, which uses the KD values to determine all sample concentrations and run parameters needed to perform optimised ITC experiments. The Bia2iTC software has been designed to work specifically with BiacoreTM SPR systems and the MicroCalTM Auto-iTC200 system. This data is subsequently exported as a setup file to the ITC instrument. Once the ITC experiments have been completed, the result file is imported by the Bia2iTC software, enabling a comparison of the ITC and SPR results in an Excel sheet.
Combining the data in this format allows a direct examination of possible correlations between the rate constants, affinities and the directly measured thermodynamic entities, and comparative plots can be achieved using Excel functionalities. Of particular interest is a comparison of the KD values obtained with both methods, as shown in Figure 3.
Together with selection criteria, such as biological specificity, chemical amenability and ADME properties, data on the kinetics and thermodynamics of binding are increasingly being used during hit confirmation, lead generation, and lead optimisation. Knowledge of these properties enables a rational approach that serves to maximise drug-like properties as well as to optimise the target residence time.1-3 As discussed here, the Bia2iTC software can be used to design these experiments.
In summary, in small molecule drug discovery there is a great need to confirm interactions with target proteins using label-free techniques with sufficient throughput. Implementing an approach uniting the use of SPR and ITC adds valuable information on the mechanisms of binding, providing label-free determinations of affinities and the important parameters of target residence time and enthalpic efficiency. The addition of the Bia2iTC software provides an additional tool for optimised design of ITC experiments, and enables easy comparison on SPR and ITC data. The combination of ITC and SPR can bring confidence to decision-making and lead to accelerated discovery of potentially safer drugs.
Kinetic and thermodynamic characterisation
For a one-step reversible interaction, the dissociation constant KD is equal to the ratio between the rate constants for dissociation (koff or kd) and association (kon or ka) of the complex:
KD=kd/ka
An infinite number of combinations of ka and kd can result in the same KD. Consequently, an understanding of the dynamics requires knowledge also of the rate constants. The reciprocal of kd is referred to as the target residence time (τ) which is used to distinguish between transient and more long-lived complexes1.
KD is also related to the free energy difference, ΔG, between associated and disassociated states of the interacting molecules:
ΔG° = RTlnKD
where R and T are the universal gas constant and absolute temperature, respectively. In order to reveal the factors driving the interaction, it is necessary to look separately at the enthalpic (ΔH°) and entropic (ΔS°) contributions to ΔG°:
ΔG° = ΔH° - TΔS°
ΔH° is a measure of the heat change accompanying the interaction and is due to the formation/disruption of hydrogen bonds and van der Waals interactions. This means that interactions that produce large exothermic enthalpies contain a more optimal network of hydrogen bonds. ΔS° is a measure of the change in the degree of freedom in the system.
The largest contributions are from hydrophobic interactions and conformational changes, which favour the interaction due to the release of bound water molecules, and loss of motional and conformational freedom, respectively.