Improving Drug–Receptor Interactions To Make Medicines Work Better

Technology Networks had the pleasure of speaking with Laura Heitman, professor of molecular pharmacology within the Division of Drug Discovery and Safety at the Leiden Academic Centre for Drug Research (LACDR), Leiden University, to learn more about her research focused on drug–target kinetics. Heitman discusses why it is important to determine the length of time a drug stays bound to its target, explains how you can assess a drug’s target-binding kinetics and touches on how kinetic computational studies are helping to advance the field.

Laura Lansdowne (LL): Could you tell us about your work focused on understanding and improving drug–receptor interactions?

Laura Heitman (LH): In general terms the research in my group focuses on the theme “novel receptor concepts to target membrane proteins” with the ultimate aim to make medicines work better. I have selected membrane-bound proteins, such as G protein-coupled receptors (GPCRs), as many drugs act via these and they play a pivotal role in disease. Of note, the concepts that I work on are in principle “disease-agnostic” and can be applied to many targets and disease areas.

At the start of my tenure back in January of 2009, one of such concepts, i.e., “drug-target residence time” or “drug–target binding kinetics” had not received much attention, if any at all. Since then, it is slowly being realized that the time a drug remains bound to its target may be of greater importance than affinity, in terms of its effect in the patient. More papers are being published that describe the importance of optimizing a drug’s binding kinetics. However, few still report on this novel parameter as a prospective tool, i.e., designing compounds to have optimal kinetics, rather than “stumbling upon” a compound with an interesting kinetic profile. In the last years, my group has developed several robust and accessible kinetic assays and started to publish such data.

Specifically, we were able to show for the first time that the binding kinetics of a drug on its target can be tuned by a medicinal chemistry approach, next to their affinity. This might have great clinical value, as retrospective analysis proves that some marketed drugs have clinical efficacy due to a long target residence time. For example, using one of our in-house designed and synthesized long residence time (RT) CCR2 antagonists, we have shown that high receptor occupancy in an atherosclerosis mouse model was key for high efficacy. Notably, this high (or extended) receptor occupancy results in so called insurmountable antagonism, i.e., antagonists that cannot be disrupted/counteracted by high local concentrations of the endogenous receptor agonist that is often causal to the disease state. As a logical extension to “long” target residence time, my group is currently also working on covalent ligands. As these molecules stay bound to their target infinitely (limited by the protein’s life cycle), antagonists will be insurmountable.

LL: How are the binding kinetics of a drug to its target characterized?

LH: This can be done quantitatively and qualitatively depending on the method used. We tend to use radioligand binding assays to qualitatively assess a drug’s target-binding kinetics. This can either be done directly by radio- or fluorescently labeling the drug of interest, or indirectly by using a so-called competition association assay where one reference labeled ligand is used that then competes with an unlabeled ligand of interest. In both cases, data analysis by non-linear regression models will provide you with the kinetic rate (k_off and k_on) values.

With regard to quantitative analysis, one can consider using washout assays where wash-resistance of binding or a certain functional effect can be observed. Moreover, in functional assays one can also assess an antagonist’s level of “insurmountability”, which is basically a phenomenon that occurs when an antagonist occupies the target for an extended amount of time, resulting in a dampening of the maximum agonist response in that functional system.

LL: Is there one experimental technique that you feel has been most impactful?

LH: The introduction of the “surface plasmon resonance” technology has really helped to generate kinetic parameters in early drug discovery. Although some developments are being made, this technique is still not readily available or easily amenable to membrane-bound proteins.

LL: Why are the kinetics of association and dissociation of a target–ligand complex so important?

LH: Despite the efforts (and successes) in finding high-affinity and selective candidate drugs, attrition rates in clinical trials are disappointingly high. Novel concepts such as drug–target binding kinetics are seen as increasingly important for in vivo efficacy and safety. This is most likely true because dynamic flow and metabolism in the human body often prevent drug molecules from reaching equilibrium conditions that are otherwise readily attained in the test tube (i.e., equilibrium parameters are still the standard in early drug discovery). Moreover, in a disease-state often different conditions arise at the target site, i.e., increased levels of endogenous agonist, as mentioned above. The compounds’ kinetic behavior (association velocity to the target and to metabolic enzymes, dissociation from the target, etc.) might in fact be the guiding principle to obtain a desired and durable effect in vivo. Hence, it is important to get a better understanding of the drug–target interaction that is needed and to optimize this at a molecular level in vitro. Thus, providing the prospect of better chances for kinetically-optimized candidate drugs in later phases of the drug development process.

LL: How are advances in computational models influencing our ability to explore binding kinetics?

LH: This is not my area of expertise, but I would say that slowly more progress is being made in the field of kinetic computational studies. There are two computational techniques that can aid in understanding and optimizing drug–target binding kinetics – molecular dynamics (MD) and machine learning (ML). For both ligand–protein structures are needed, accompanied by computing power (MD) and kinetic data (ML). Depending on the type of protein (i.e., membrane-bound or cytosolic) structural data is limited, and kinetic data is currently also still scarce due to its underappreciation. Once the limitations are lifted, these techniques can be used to visualize molecular mode of target interaction, dissociation and maybe even association (MD), and aid in binding kinetics prediction for hit–lead optimization (ML).

Laura Heitman was speaking with Laura Elizabeth Lansdowne, Managing Editor for Technology Networks.