Studying kinase target engagement through a traditional biochemical method requires countless assumptions about how the assay represents kinase inhibition in vivo. Instead, Promega scientists are putting the purified enzymes away and looking inside a real, living cell. 

Problems related to kinase activity have been implicated in many cancers and autoimmune diseases. The kinome represents one of the broadest sources for therapeutic targets, and research into kinase inhibitors has been a hot topic for several years. When developing a drug targeting kinases, studying target engagement is crucial for understanding how the molecule behaves in relation to not only its target, but other intracellular biomolecules. 

The biochemical approach to studying target engagement fails to reflect the complexity of the full-length kinase and how it is affected by other components of the cell, including metabolites and regulatory circuits. It is unable to measure compound permeability and drugs that bind outside of the ATP-binding pocket often can’t be detected. As a result, there has been a growing need for a target engagement system that works within the natural cell environment. Promega research scientists Matt Robers and Jim Vasta address these problems in their recently published paper, “Quantitative, Wide-Spectrum Kinase Profiling in Live Cells for Assessing the Effect of Cellular ATP on Target Engagement.”

In the paper, Matt and Jim use NanoBRET™ to analyze inhibitor occupancy in living cells without disturbing the intracellular equilibrium or the cell membrane. The technique involves tagging the target kinase with NanoLuc and adding a specific energy transfer probe, resulting in a BRET (bioluminescence resonance energy transfer) signal. When the inhibitor molecule displaces the probe, the BRET signal disappears. The BRET can be quantified over time to measure kinase inhibitor occupancy, selectivity, and affinity. This system solves many of the problems of the biochemical approach because it happens within the typical cellular environment, instead of within defined assay conditions that might not fully recapitulate the target's behavior.

Binding of test compound results in loss of NanoBRET signal

Figure 1. Illustration of NanoBRET™ in live cells. When the drug binds to the target kinase, the energy transfer probe is displaced and the BRET signal disappears. 

“There are a lot of things in the realm of target engagement that people consider true from a theoretical standpoint,” Jim explains. “What would happen if we put our target in a cellular context? What do we expect to happen? There are very few situations where those theoreticals were measured empirically. Our method actually enables you to measure many of those parameters empirically because it’s literally happening inside the cell.”

“I like to use the word nexus—it’s the nexus between biochemistry and cell phenotype.” Matt jokes. “Really, though, when you purify a kinase protein and you test your potential drug molecules against it, what you’ve learned when you do that experiment doesn’t predict what you observe when you do those experiments in cells. Our technique bridges the gap in a way that captures the same quantitative capabilities as the biochemical approaches but in a way that’s more predictive of what happens in a live cell.” 

A clearer picture of selectivity

Most kinase inhibitors bind to a highly conserved ATP pocket, so targeting one specific kinase can prove difficult. Since promiscuous binding is a major cause of side effects for some drugs, one of the early stages of drug development involves generating selectivity profiles of the molecule to determine if the drug is specific enough for its targets. 

“When you purify a kinase protein and you test your potential drug molecules against it, what you’ve learned when you do that experiment doesn’t predict what you learn when you do those experiments in cells.”

The paper describes how the authors used six energy transfer probes to study 178 different protein kinases. To illustrate the improvements over biochemical methods, they generated selectivity profiles for several inhibitors and compared the results to the profiles found in published literature. The NanoBRET technique found that crizotinib, for example, bound 13 kinases above 50% occupancy, whereas the biochemical approaches had predicted 69 would reach that mark. Dasatinib produced similar results – 22 kinases were engaged above 50%, compared to 44 listed in previous literature. The authors explain that this discrepancy is because the biochemical methods failed to account for the role the intracellular milieu plays in the thermodynamics of inhibitor binding. 

"With a biochemical approach, you may come to a selectivity profile that doesn't reflect what's happening in disease-relevant cells," Matt says. "It may allow you to come to the conclusion that you have a highly promiscuous molecule, when in fact the selectivity profile may be vastly improved."

Dissecting the complexity of the cell

Since almost all studied kinase inhibitors are ATP-competitive, it is crucial to understand how intracellular ATP will affect the potency of the drug in question. In a biochemical approach, the assay will be performed either under assumed conditions—the expected ATP concentration around the target—or with the ATP concentration optimized for the kinase enzyme. Neither of these strategies are adequate, though, because interference from ATP is context-dependent. 

“In the paper, we actually modulated intracellular metabolites,” Jim explains. “We modulated ATP to ask the question, ‘How does that affect the behavior of my target?’ Or, more broadly, ‘What does target engagement look like once I start to deconstruct the cell?’ That hints at the idea that we can play with what’s going on in the cell and expect to see differences in the behavior of the inhibitor and target.” 

Jim designed an experiment that used rotenone, an environmental pesticide, to deplete cells of ATP. He then generated engagement profiles for crizotinib in three different kinases to analyze how ATP interferes with the inhibitor potency. Two of the kinases showed enhanced engagement profiles in the ATP-depleted cells, and the Kd calculated from the ATP-depleted cells was in close agreement with the value reported by previous biochemical analyses. This puts a spotlight on how much error can be introduced in biochemical methods that fail to account for intracellular factors. 

14845ma-w

Figure 2. Example of differences in engagement of EPHB6 between ATP-depleted cells and undisturbed cells. The enhanced, left-shifted potency curve for crizotinib in the ATP-depleted and permeabilized cells demonstrates the role of ATP in inhibitor competitive binding. 

“I think this will appeal to a lot of the reductionists out there,” Matt says. “With our technique, you can take something with a billion different interactions going on and start to dissect that. Imagine—the ability to analyze live-cell binding characteristics of your drug molecule with different little bits of the cell removed. But you’re still in an intact cell. And that has never been shown before.”

The future of NanoBRET™ Kinase Target Engagement

The NanoBRET™ Target Engagement technique currently works for 178 protein kinases, out of the 518 total in the human kinome. Matt and Jim say their primary goal right now is to make the technique compatible with as much of the kinome as possible. They are working with researchers at the Structural Genomics Consortium to characterize drug molecules and chemical probes against large sets of kinases and then use those scaffolds to create new energy transfer probes. 

Looking past the immediate goals, Matt has big plans for NanoBRET™ kinase target engagement. “The untapped territory for us is looking at the dynamics of kinase activation in the cell. We should be able to influence those activation states by how we handle and treat our cells, allowing us to bias our analysis to one or more of those states. Drugs might behave differently at those different states. We know that to be true for certain drug molecules.” 

That capability is heavily anticipated, since we know that kM values for ATP are dependent on the activation state of the kinase. Given the broad range of possible intracellular scenarios the drug could face, it’s difficult to address potential potency shifts without knowing the details and effects of activation and conformational change. Matt and Jim are confident that their technique can soon be used to begin elucidating these difficult areas, once they get the chance to dive in.

“The challenge is that we’re asking questions that have never been asked before,” Matt says. “And with that will come a lot of excitement, but also a lot of skepticism. We’ll probably get a lot of people with questions or concerns, or they’ll challenge us. But we believe the technology can address some important fundamental questions about target engagement, and I think that’s really where we want to operate in the chemical biology field.” 

 

Read the paper:

1. Vasta, J., et al. (2017), Quantitative Wide-Spectrum Kinase Profiling in Live Cells for Assessing the Effect of Cellular ATP on Target Engagement. Cell Chemical Biology