A High-Throughput BRET Cellular Target Engagement Assay Links Biochemical to Cellular Activity for Bruton’s Tyrosine Kinase

Abstract

Protein kinases are intensely studied mediators of cellular signaling. While traditional biochemical screens are capable of identifying compounds that modulate kinase activity, these assays are limited in their capability of predicting compound behavior in a cellular environment. Here, we aim to bridge target engagement and compound-cellular phenotypic behavior by utilizing a bioluminescence resonance energy transfer (BRET) assay to characterize target occupancy within living cells for Bruton’s tyrosine kinase (BTK). Using a diverse chemical set of BTK inhibitors, we determine intracellular engagement affinity profiles and successfully correlate these measurements with BTK cellular functional readouts. In addition, we leveraged the kinetic capability of this technology to gain insight into in-cell target residence time and the duration of target engagement, and to explore a structural hypothesis.

Introduction

Kinases are among the most successful target classes in drug discovery, being targeted with small-molecule inhibitors in diverse therapeutic areas. ¹ Bruton’s tyrosine kinase (BTK) is a member of the Tec family of nonreceptor tyrosine kinases and is validated as a key pathological driver in autoimmune diseases such as rheumatoid arthritis (RA) and multiple sclerosis. ² Drug discovery efforts have resulted in at least one marketed kinase drug (ibrutinib), which binds irreversibly to BTK via cysteine 481 in the kinase domain ( Fig. 1A ), as well as several other reversible and irreversible inhibitors currently in clinical trials, including the BMS candidates BMS-986142 (reversible) in phase II and branebrutinib (irreversible) in phase I.^3–5 Preclinical pharmacokinetic/pharmacodynamic studies in animals have suggested that high levels of target occupancy, that is, coverage as high as IC₉₀, is required for in vivo activity. ⁵ Thus, identifying cell-active and potent small-molecule inhibitors of BTK has been of great interest.^2,4,5 In addition to target occupancy, drug-target residence time is an important factor that may underly the efficacy and selectivity of BTK inhibitors in B-cell-mediated diseases. For example, a number of covalent but reversible BTK inhibitors with long residence times have been developed with promising preclinical results. ⁶ Thus, the ability to measure direct target engagement and residence time at BTK in cells, and ultimately in tissues and patient samples, is of critical importance. Some notable studies measuring BTK occupancy in living cells have been reported, for example, leveraging an antibody-based time-resolved fluorescence method to measure free and total BTK in peripheral blood mononuclear cells. ⁷ Other in vitro and cellular approaches characterizing various important mechanistic parameters, such as time dependence of inhibition, substrate competition, and recovery of function after inhibitor removal, have also been pursued. ⁸ However, higher-throughput methods capable of assessing cellular target engagement and binding kinetics in real time in living cells are limited.^9–11

Figure 1.

BTK structure and signaling pathway. (A) Domains in the BTK structure are shown alongside amino acid positions. Ibrutinib binds covalently with cysteine 481. (B) BTK activation and signaling ultimately leads to calcium release and gene expression changes that regulate cell proliferation and survival.

In the absence of cellular target occupancy readouts, cellular BTK activity is traditionally measured via pathway analysis methods. The BTK-mediated signaling events that link B-cell receptor (BCR) ligation to downstream signaling have been well characterized ( Fig. 1B ).^12,13 Upon BCR engagement, Src kinases associated with the receptor become activated and phosphorylate tyrosine residues on the BCR. These residues become docking sites for other kinases including phosphatidylinositol 3-kinase (PI3K). Once recruited, PI3K phosphorylates phosphatidylinositol 4,5-bisphosphate to form phosphatidylinositol 3,4,5-triphosphate (PIP3). PIP3 then recruits BTK to the plasma membrane via interaction with its pleckstrin homology (PH) domain, resulting in BTK activation via phosphorylation. Activated BTK subsequently phosphorylates several downstream substrates, including phospholipase Cγ2, which regulates release of Ca²⁺ from intracellular stores. Recruitment and activation of BTK results in elevation of intracellular Ca²⁺ and downstream alterations in gene expression. Therefore, measurements of Ca²⁺ mobilization can serve as a key pathway analysis tool for functional characterization of BTK inhibitors in cells.

Here, we have applied a quantitative method to measure BTK engagement in cells, and to evaluate the relationship between cellular target occupancy and modulation of second messenger responses downstream of BTK. Toward these goals, we applied the NanoBRET^14,15 cellular target engagement technology to assess direct BTK target occupancy in real time in a cellular context. Accordingly, we successfully miniaturized the BTK NanoBRET assay to 384-well plate format and determined the optimal conditions for the real-time analysis of BTK residence time. Using our adapted NanoBRET assay system, we were able to triage BTK inhibitors based on intracellular binding affinity. We observed a quantitative correlation between target occupancy and cellular functional activity across a diverse chemical set of BTK inhibitors. We further extended the analysis to a non-steady-state system, where the duration of BTK occupancy can be quantified for select inhibitors. Among the tested compounds, we characterized a reversible inhibitor with durable and potent occupancy of BTK in living cells, providing proof of concept that an extended intracellular engagement of BTK can be achieved via noncovalent and covalent mechanisms of action.

Materials and Methods

Cell Transfection and Cryopreservation

Cryo-banked cells were used in the bioluminescence resonance energy transfer (BRET) target engagement experiment. HEK293 cells (ATCC, Manassas, VA) were transfected with BTK-NanoLuc fusion plasmids (pF31K [CMV/Neo], Promega, Madison, WI) using FuGENE HD transfection reagent (Promega) according to the manufacturer’s protocol. First, BTK-NanoLuc DNA was mixed with transfection carrier DNA (pGEM3ZF vector, Promega) in a 1:10 mass ratio for a final total DNA concentration of 10 μg/mL in phenol red free Opti-MEM (Gibco, ThermoFisher Scientific, Gaithersburg, MD). FuGENE was added into the DNA solution for a final 3% (v/v) concentration, after which the resultant solution was incubated at room temperature for 15 min to form the transfection complex. HEK293 cells were resuspended at 200,000 cells/mL in DMEM (Gibco) supplemented with 10% fetal bovine serum (HyClone, GE Healthcare, Chicago, IL) and 1% HEPES (Gibco). The cells were transfected with the transfection complex solution at a volume ratio of 1:20 (complex–cell suspension), after which cells were plated into a flask and incubated overnight at 37 °C with 5% CO₂ supplementation.

Following overnight transfection, cells were harvested with 0.05% trypsin/EDTA (Life Technologies, ThermoFisher Scientific, Carlsbad, CA) and resuspended to a density of 1 million cells/mL in Recovery cell freezing media (Gibco). Cells were aliquoted to cryo-vials (Corning, Corning, NY) and were frozen first at −80 °C in a CoolCell storage unit (Corning) before being transferred to long-term storage in liquid nitrogen.

Steady-State (Endpoint) BRET Measurements

The NanoBRET assay was miniaturized from a 96-well format ¹⁶ by employing automated liquid dispensing. BRET measurements were made in white 384-well nonbinding surface plates (Corning). Cells were thawed and resuspended at 200,000 cells/mL in Opti-MEM without phenol red (Gibco) before seeding in a 384-well plate for a final count of 10,000 cells/well. Chemical compounds were resuspended in DMSO (Sigma-Aldrich, St. Louis, MO) and transferred to assay plates via acoustic dispensing (EDC Biosystems, Fremont, CA).

For the live-cell assay conditions, the plates were incubated for 2 h at 37 °C with 5% CO₂ supplementation. NanoBRET NanoGlo Substrate and Extracellular NanoLuc Inhibitor (Promega) were added according to the manufacturer’s guidelines. Assay plates were incubated at room temperature for 2 h before adding NanoBRET NanoGlo Substrate (Promega).

For all BRET experiments, the Envision 2015 (PerkinElmer, Waltham, MA) equipped with a BRET2 Enh mirror and BRET 647 nM/75 nM Bandwidth and BC703 460 nM/80 nM Bandwidth filters was used for readouts. Each read was integrated for 0.5 s.

Probe K_d Determination

Affinity for probe 4 or 5 ¹⁶ for BTK-NanoLuc in the HEK293 cells was determined by titration. Briefly, a twofold dilution series was prepared in DMSO, after which the entire dilution series was diluted with four parts probe dilution buffer (Promega) to prepare a 20× working probe reagent. This dilution series was titrated onto BTK-NanoLuc-expressing cells and equilibrated in the presence and absence of a saturating dose (10–20 µM) of corresponding unlabeled probe parent compound for each probe. BRET was measured as described above. BRET ratios were converted to milliBRET (mBRET) values by multiplying by 1000, after which the mBRET values were plotted as a function of probe concentration. The apparent affinity of the probe (EC₅₀) was determined by fitting to the four-parameter dose–response equation (eq 1) using GraphPad Prism software (GraphPad, La Jolla, CA).

$$Y=Bottom+\frac{Top-Bottom}{1+\frac{x^{HillSlope}}{xC{50}^{Hillslope}}}$$

(1)

Compound Target Engagement Measurements at Equilibrium

Chemical inhibitors were synthesized in-house as previously described.^4,5,17–20 Compounds were diluted threefold in DMSO for a total of 11 points. Final assay conditions used a high concentration of 20−50 μM for most compounds. Probe 4 was used for equilibrium measurements, employing a final concentration of approximately EC_50–80. For live-cell measurements, 100 μM probe 4 was added. Plates were incubated and BRET was measured as indicated above.

To determine compound potency (IC₅₀), BRET ratios for each dilution series were first normalized to a percent inhibition value via eq 2, where S represents sample BRET, B represents background BRET, and T represents total BRET. In this equation, the background control (B) consists of BTK-NanoLuc-expressing cells and no probe 4. The total BRET ratio (T) consists of BTK-NanoLuc-expressing cells, probe 4, and an amount of DMSO at a concentration equivalent to that added when testing compounds.

$$\%Inhibition=100\times \left(1-\frac{S-B}{T-B}\right)$$

(2)

Following normalization, the percent inhibition values are plotted as a function of compound concentration in GraphPad Prism and fit to eq 1.

Kinetic Probe Characterization and Cellular Residence Time Assay

All kinetic parameters were assessed using adherent HEK293 cells transfected as described above. Probe 5 was used for kinetic experiments due to its fast association kinetics in cells. HEK293 cells in Opti-MEM medium containing 1% FBS and transfected with BTK-NanoLuc were plated in 96-well tissue culture-treated plates (Corning) and incubated overnight.

The association kinetics of probes were characterized in kinetic dose–response experiments. A probe dilution series was added to the cells and BRET was measured kinetically on a Glomax Discover instrument (Promega) at 3 min intervals using detection reagents as described above. BRET data were plotted as a function of time and fit to the one-phase association equation in GraphPad Prism (eq 3) to obtain an observed rate constant (k_obs) for each compound.

$$Y=Y_0+\left(Plateau-Y_0\right)\times \left(1-e^{-k_{obs}x}\right)$$

(3)

The k_obs values were replotted as a function of probe concentration, after which the k_on and k_off values were determined by linear regression using the observed rate constant equation (eq 4), where k_off is apparent from the y intercept and k_on is apparent from the slope.

$$k_{obs}=k_{on}\left[Tracer\right]+k_{off}$$

(4)

Cellular residence time was assessed via cellular washout. After transfection, adherent HEK293 cells expressing BTK-NanoLuc were treated with a saturating dose of test compound, 20- to 100-fold above the IC₅₀ value measured via live-cell equilibrium titration. Cells were incubated with compounds for 2 h, after which the medium was removed and replaced with fresh Opti-MEM to induce an equilibrium shift. NanoGlo Substrate and extracellular NanoLuc Inhibitor were added to each well, after which probe 5 was added to a final concentration of 0.25 µM and BRET was measured kinetically at 3-min intervals using a Glomax Discover Instrument. BRET data were plotted as a function of time.

Biochemical Enzymatic Activity Assay

The biochemical activity of compounds was measured at K_M,ATP using Caliper technology as previously described.^4,20 Compounds were diluted threefold in DMSO for a total of 11 points and were tested for BTK activity at the kinase’s ATP K_m in a reaction buffer (20 mM HEPES, pH 7.4, 10 mM MgCl₂, 0.015% Brij 35 surfactant, and 4 mM DTT in 1.6% DMSO). Final assay conditions included compound, 1 nM human recombinant BTK (Invitrogen, Carlsbad, CA), 1.5 μM fluorescein-labeled peptide, and 20 μM ATP. Reactions were incubated for 1 h before quenching with 35 mM EDTA. The reaction mixture was read using capillary electrophoresis on the Caliper LabChip 3000 (Caliper, Hopkinton, MA) at an excitation wavelength of 488 nm and an emission wavelength of 530 nm. Percent inhibition was calculated according to eq 2 and curves were fitted to eq 1 to determine the IC₅₀ for each inhibitor.

Calcium-Flux Ramos Cell Assay

Human Ramos B-cell assays were performed as previously described.^4,20 Human Ramos (RA1) B cells (ATCC CRL-1596) were suspended to 2 × 10⁶ cells/mL in phenol red free RPMI (Invitrogen 11835-030) supplemented with 50 mM HEPES (Invitrogen 15630-130) and 0.1% BSA (Sigma A8577). The cells were added 1:1 with calcium loading buffer (BD PBX calcium assay kit no. 640177; BD, Franklin Lakes, NJ) and incubated at room temperature in the dark for 1 h. Afterward, cells were exchanged to a phenol red free RPMI media with 50 mM HEPES and 10% FBS for a final concentration of 1 × 10⁶ cells/mL. A total of 150,000 cells/well were plated into 96-well poly-d-lysine-coated assay plates (BD 35 4640) and treated with compounds diluted in DMSO (final 0.13% DMSO) for 1 h at room temperature in the dark. The FLIPR1 (Molecular Devices, Downington, PA,) was used to add goat anti-human IgM (Invitrogen AHI0601) to 2.5 μg/mL and to monitor changes in intracellular calcium over 180 s. Percent inhibition was ratiometrically normalized to peak calcium levels measured in the stimulation-only condition. IC₅₀ values were obtained using eq 1.

Results

Miniaturization of the BTK NanoBRET Assay and Adaptation for Binding Kinetics Analysis in Living Cells

BRET is highly sensitive to molecular distance and orientation between the donor (luciferase) and acceptor (fluorescent probe) pair.^15,21 This property enables a bona fide measurement of direct molecular interactions inside intact cells. Figure 2 explains the cellular BRET target engagement assay principle. The assay utilizes an intracellular BTK genetically fused to a NanoLuc luciferase donor and a cell-permeable, ATP-competitive fluorescent acceptor (probe 4 or 5 herein). ²² We used this strategy to develop a simplified BTK ligand-displacement assay system, and this assay was miniaturized to both 384- and 1536-well formats ( Suppl. Fig. S1 ). BRET is achieved inside the intact cells by reversible binding of this fluorescent probe to the intracellular target. Compound binding affinity at BTK is determined by competition with the probe, where probe displacement results in a BRET signal decrease.

Figure 2.

NanoBRET technology enables high-throughput target engagement and residence time measurements in live cells. (A) BRET is obtained by labeling a target protein of interest with a NanoLuc tag and introducing a cell-permeable fluorescently labeled probe. Upon substrate addition and turnover by NanoLuc, the light produced is capable of undergoing BRET with the target-bound fluorophore-labeled probe. (B) BRET can be used to measure target engagement by adding competing unlabeled compound, resulting in a loss of BRET signal. (C) Cellular compound residence times can be measured by saturating the target with the compound of interest, washing out excess compound, and then introducing a large excess of probe. As the compound dissociates and the probe engages the target, increased BRET signals are observed.

To establish the BTK cellular target engagement assay in-house, the affinity of the competitive probe was first determined by the titration of varying probe concentrations onto live HEK293 cells transfected with the optimal C-terminal genetic fusion of BTK-NanoLuc. ¹⁶ Using both fluorescent probes 4 and 5 (from Vasta et al. ¹⁶ ), strong, dose-dependent BRET signals were observed, shown in Figure 3A,D, respectively. These BRET signals were strongly competed away in the presence of a saturating dose of the respective unlabeled parent compounds, demonstrating specificity. To ensure physiological conditions and avoid ligand depletion, various expression levels of the BTK-NanoLuc plasmid were also tested (data not shown), but BRET and probe engagement was unaffected by expression level. As demonstrated in Figure 3B for probe 4, the K_i of the unlabeled parent compound can be calculated in live cells between the ATP-competitive probe and the unlabeled parent compound using the Cheng–Prusoff relationship, which describes the mathematical relationship for how the apparent IC₅₀ of a competitive inhibitor is affected by the presence of another competitive ligand (i.e., the fluorescent probe). The Cheng–Prusoff relationship can also be used to extract the apparent intrinsic affinity (K_d,app) of the unlabeled test compound in the live-cell environment by linear regression analysis. ²³ In this case, the linearized Cheng–Prusoff analysis yielded a K_d,app of 11 nM for the BTK-NanoLuc in live HEK293 cells, and an order of magnitude more potent than the derivative probe 4 (EC₅₀ = 250 nM). Although the pendant fluorophore appears to affect target affinity compared with the unlabeled parent compound from which the probe was developed, our results support general method validity to provide accurate target engagement profiles for unlabeled compounds by competitive displacement, so long as the probe is cell permeable and specific and used at an appropriate subsaturating concentration (~EC₅₀ to EC₈₀).

Figure 3.

K_d and kinetic characterization of the BTK target engagement assays. Binding affinity and/or kinetic parameters of Bodipy-labeled probes 4 and 5 to BTK-NanoLuc were characterized in live cells. (A) The EC₅₀ of probe 4 was measured to be 250 nM, and competition with the unlabeled parent compound demonstrated assay specificity. (B) The IC₅₀ of the probe 4 parent compound was evaluated as a function of increasing probe 4 concentration. (C) Linearized Cheng–Prusoff analysis of the data from panel B afforded an apparent intracellular K_i (K_i,app) of 11 nM (from the y intercept) for the probe 4 parent compound. (D) The EC₅₀ of probe 5 was measured to be 250 nM, and competition with the unlabeled parent compound demonstrated assay specificity. (E) The association kinetics of probe 5 were measured at increasing probe 5 concentrations. (F) Replot of the observed rate constant (k_obs) measured versus the probe 5 concentration yielded a linear relationship. The k_on and k_off values for the probe were determined by linear regression, where k_on is equal to the slope and k_off is equal to the y intercept. The K_d determined from the kinetic constants (k_off/k_on) was in agreement with the EC₅₀ of the probe determined by titration in panel D.

We also evaluated the probe binding kinetics by titrating increasing concentrations of probe and monitoring BRET at BTK-NanoLuc kinetically over time (Fig. 3E,F). Probe 5 was used for these kinetic evaluations as it demonstrated faster association rates compared with those previously measured for probe 4, ²⁴ which is advantageous for use in downstream cellular residence time assays. To characterize probe 5, the observed rate constant (k_obs) was determined by fitting the kinetic curves to a nonlinear pseudo-first-order association model ( Fig. 3E ). The k_obs values replotted with respect to probe concentration yield a linear relationship, which allows for dissection of k_on and k_off of the probe ( Fig. 3F ). For BTK, both on- and off-rates were fast, also informing lengths of time required to achieve a steady state in the endpoint target engagement assay. Comparison of the kinetically derived probe K_d (k_off/k_on) with the previously determined K_d from endpoint titration ( Fig. 3D ) shows that the two values are in gratifying agreement. Thus, kinetic analysis suggests that the fast rates of probe 5 can enable steady-state readouts after 2 h of compound incubation. Additionally, the fast on-rate of probe 5 is also advantageous for compound residence time measurements in live cells. Our endpoint and kinetic results support a robust and scalable approach to enable steady-state and real-time measurements using the NanoBRET probe at BTK in living cells.

Characterization of Target Occupancy for Diverse Chemical Series of BTK Inhibitors in Living Cells

To evaluate the utility of the BRET method to profile intracellular BTK target engagement, we tested a panel of structurally diverse inhibitors in a 384-well plate format with a range of in vitro potencies spanning at least three orders of magnitude. The design, synthesis, and optimization of these inhibitors has been reported previously.^{4,5,17–20,25} At K_d concentrations of the probe, the BTK BRET target engagement assay serves as a measure of the apparent occupancy for BTK-NanoLuc in live cells. Following 2 h of treatment in the presence of probe 4, robust dose–response characteristics were observed for the various inhibitor treatments. Signals were normalized against vehicle control-treated samples such that maximal inhibition corresponded to 100% signal inhibition. Figure 4 shows examples of competitive binding displacement IC₅₀ curves for representative compounds across each chemotype, with varying potency from single-digit nanomolar to micromolar, underscoring that the BRET-based BTK assay can report on cellular binding potencies for compounds with sub-micromolar or greater in vitro affinity. Inhibitor cellular target engagement profiles demonstrate the translatability of the assay across multiple chemotypes of interest as well as both reversible and irreversible binding mechanisms. A key structural feature of the BTK kinase domain is the noncatalytic cysteine (Cys481) residue in the extended hinge region, which can be targeted to generate irreversible BTK inhibitors. Within this test set, the irreversible inhibitor ibrutinib targeting Cys481 displayed the most potent target engagement, followed by the carbazole and nicotinamide analogs, with the pyrazolopyridines showing the weakest affinity. Notably, ibrutinib also exhibited time-dependent inhibition, a hallmark of irreversible binding (data not shown).

Figure 4.

Cellular target engagement IC₅₀ measurement using BRET. (A) Representative compounds are depicted for each chemotype studied. Top, chemical structure. Bottom, IC₅₀ curve. (B) Crystallographically based structural models highlighting similar interactions to the hinge, and different interactions in the front pocket.

To better understand the potency differences across these chemotypes, we examined the crystal structures and molecular models of representative compounds in BTK ( Fig. 4 ). The chemotypes share a tridentate interaction with the BTK hinge through the backbone carbonyl of glutamate 475 and the backbone carbonyl and NH of tyrosine 476, but diverge in their interactions with BTK’s front pocket. The least potent pyrazolopyridine, compound 1, projects the phenyl ether moiety toward the gatekeeper residue, threonine 474, with the terminal phenyl filling the back pocket. In contrast, the more potent nicotinamide and carbazole chemotypes form direct and water-bridged interactions with the extended hinge. The quinazolinone group from the carbazole in compound 2 packs against side chains of cysteine 481 and leucine 483 with hydrogen bonding to the backbone NH of cysteine 481 and backbone carbonyl of leucine 409 via bridging waters. In contrast, the piperidine linker in compound 3 positions the urea carbonyl-to-hydrogen bond directly with the backbone NH of cysteine 481, and the thiophene is packed against the side chain of asparagine 484. The irreversible compound, ibrutinib, achieves its very high affinity through a covalent bond to cysteine 481. These differential interactions in the BTK front pocket provide the structural basis underlying the potency rank order of these chemotypes.

In-Cell Target Engagement Correlates Quantitatively with Modulation of BTK Function

The ideal target engagement format provides a direct assessment of compound binding at the selected intracellular target, therefore providing a bridge between reductionist biochemical assays and cellular pathway analysis readouts. To what extent do BTK in vitro measurements actually correlate with cellular target engagement? To address this question, we analyzed in vitro BTK enzymatic activity data and compared it with cellular BRET target engagement data for the test set of compounds. Figure 5 plots the relationship between the BRET and biochemically derived IC₅₀ values. Notably, the in vitro biochemical assay utilizes full-length BTK and measures the phosphorylation activity of a peptide substrate by capillary electrophoresis at ATP concentrations corresponding to K_M. The cellular IC₅₀ for BTK inhibitors is, on average, 33-fold above the biochemical IC₅₀, and most compounds fall within a fivefold window of this value. Several factors can contribute to this potency offset, including cellular permeability, differential protein binding, and intracellular ATP concentrations, reported in the 1–10 mM range. Using the experimentally determined K_M,ATP for BTK of 100 µM and the Cheng–Prusoff equation to try to superimpose the biochemical and cellular BRET IC₅₀, we estimate that the ATP concentration impacting BTK engagement in the cellular milieu is approximately 1.5 mM (~15-fold over the K_M,ATP for BTK), well within the reported literature range. It is important to note that kinase K_M,ATP is also influenced by the activation state, which may be more dynamic in cells. ²⁵

Figure 5.

Comparison of BRET with in vitro enzymatic and cellular functional assays. (A) ATP-corrected and uncorrected biochemical IC₅₀ values plotted against cellular target engagement IC₅₀ values. (B) Correlation of ATP-corrected biochemical IC₅₀ values with cellular functional IC₅₀ values. (C) Correlation between cellular BRET and cellular functional IC₅₀ values. Pearson correlation is displayed for all plots. Compounds with measured IC₅₀ values outside the detection limit are not shown.

Next, we investigated whether the cellular target engagement method could establish a functional link between physical binding and modulation of endogenous cellular kinase activity. Accordingly, we performed a correlation analysis between binding affinity and cellular function. To detect a proximal functional readout of BTK inhibition, we measured Ca²⁺ flux in Ramos cells. Figure 5 depicts data across the broad test set of BTK inhibitors assayed according to their ability to block Ca²⁺ release. The results show a tight correlation (Pearson R = 0.8) between cellular BRET target engagement and Ramos Ca²⁺ flux functional measurements across different structural analogs, with chemotypes clustering together. Notably, the correlation is higher in comparison to that observed between the biochemically derived IC₅₀ values and the BRET values. It is not surprising that pharmacological inhibition of BTK should inhibit Ca²⁺ flux in this cell model, consistent with BTK signaling leading to activation of phospholipase C ( Fig. 1B ). Important to note is that while BTK inhibition is necessary, it is not sufficient to block BCR activation, and so, distal functional endpoints may also be affected by additional factors.

Extended Target Engagement of a Noncovalent Inhibitor at BTK in Living Cells

Because of the high cellular potency of several of these compounds, we further examined how these compounds behave kinetically and whether any of them offer protracted target binding. We utilized the NanoBRET technology to measure residence time comparing the potent, reversible compound 2 against the irreversible ibrutinib. Cells were pretreated with compound at a saturating concentration (5- to 10-fold IC₅₀) for extended times before washout and the addition of excess competing probe 5. Notably, the off-rate is an intrinsic compound property that follows first-order kinetics and is independent of compound concentration. Remarkably, limited dissociation of compound 2 was observed over the course of 120 min, resembling the kinetic profile of the irreversible ibrutinib inhibitor, thus supporting a cellular half-life (T_1/2) of >120 min for this reversible compound. The extended residence times can be explained from the structural model of compound 2, in which the quinazoline moiety is shown to make van der Waals and water-mediated hydrogen bonding interactions with the extended hinge ( Fig. 6 ). Watermap analysis ²⁶ of this site suggests that the long residence time is driven by the displacement of entropically unfavorable waters from the extended hinge ( Fig. 6B ) and stabilization of waters via hydrogen bonding. Because the quinazolinedione moiety lacks any specific interactions with the extended hinge, this represents a unique example of a surface-exposed site giving rise to enhanced potency and residence time via solvent-driven effects.

Figure 6.

Cellular residence time evaluation for reversible and irreversible BTK inhibitors. (A) Kinetic profiles for ibrutinib (an irreversible inhibitor) and compound 2 (a reversible inhibitor). Cells were equilibrated for 2 h with each compound at a saturating dose or with DMSO vehicle, after which the cellular residence time was assessed by compound washout and addition of probe 5 (0.25 µM). Compound 2 displayed prolonged inhibition of BRET similar to that of ibrutinib over the 2 h time course, suggestive of long cellular residence time. (B) Watermap analysis of the BTK extended the hinge region from the compound 2/BTK model. Water sites are colored as a gradient from red (unstable) to green (stable). Water sites that are displaced by the quinazolinedione moiety are indicated with arrows. Potential bridging waters are denoted with asterisks. (C) Predicted bridging water network between the quinazolinedione carbonyl and cysteine 481 and leucine 408.

Discussion

In summary, we have found that NanoBRET methods can provide quantitative analysis of BTK intracellular target engagement and residence time, and serve as a useful tool for the large-scale characterization of cell-active BTK inhibitors, offering a potentially valuable template to evaluate compound structure–activity relationships for kinase inhibitors in cells. We have demonstrated that intracellular binding can be correlated with functional potency for a broad chemical inhibitor set, establishing biochemical-to-cellular connectivity and a framework to explore these relationships. Notably, additional broadly applicable technologies have also emerged that leverage measurements of target stability, mass spectrometry, and chemical proteomics, as well as fluorescence imaging, to measure binding characteristics directly in living cells and correlate them with phenotypic results, utilizing in some cases label-free and proteome-wide assessments. A broader survey of these methodologies and their translational value in the context of drug discovery is nicely presented in Robers et al. ²⁷ and beyond the scope of this discussion. Notably, one of the unique capabilities of this approach versus others is that it can be used to assess in-cell binding kinetics via dynamic monitoring of the BRET signal in live cells. Such kinetic assays proved to be beneficial in further characterizing compounds that may have desirable engagement kinetics, highlighting open system behaviors that may not be evident from steady-state analysis. This opens the potential to study additional pharmacology, such as kinetic selectivity, where it may be possible to achieve on-target selectivity through extended residence time at the primary target and transient engagement to off-targets (such as LCK in the case of BTK). Thus, the cellular BRET assay technology provides a powerful and scalable approach for evaluating key engagement parameters, providing a profiling tool for target-oriented screening and mechanistic studies.^16,22 With the emergence of new genome editing and expression techniques, the BRET method will also be expandable to additional physiologically relevant cell types. Our results support that this method has the capacity to translate binding parameters into functional pharmacology in living cells.

Supplemental Material

Supplementary_Material – Supplemental material for A High-Throughput BRET Cellular Target Engagement Assay Links Biochemical to Cellular Activity for Bruton’s Tyrosine Kinase

Supplemental material, Supplementary_Material for A High-Throughput BRET Cellular Target Engagement Assay Links Biochemical to Cellular Activity for Bruton’s Tyrosine Kinase by L. L. Ong, J. D. Vasta, L. Monereau, G. Locke, H. Ribeiro, M. A. Pattoli, S. Skala, J. R. Burke, S. H. Watterson, J. A. Tino, P. L. Meisenheimer, B. Arey, J. Lippy, L. Zhang, M. B. Robers, A. Tebben and C. Chaudhry in SLAS Discovery