Sage Journals: Discover world-class research

Abstract

One of the main reasons for the lack of drug efficacy in late-stage clinical trials is the lack of specific and selective target engagement. To increase the likelihood of success of new therapeutics, one approach is to conduct proximal target engagement testing during the early phases of preclinical drug discovery. To identify and optimize selective IRAK4 inhibitors, a kinase that has been implicated in multiple inflammatory and autoimmune diseases, we established an electrochemiluminescence (ECL)-based cellular endogenous IRAK1 activation assay as the most proximal functional evaluation of IRAK4 engagement to support structure–activity relationship (SAR) studies. Since IRAK1 activation is dependent on both the IRAK4 scaffolding function in Myddosome formation and IRAK4 kinase activity for signal transduction, this assay potentially captures inhibitors with different mechanisms of action. Data from this IRAK1 assay with compounds representing different structural classes showed statistically significant correlations when compared with results from both IRAK4 biochemical kinase activity and functional peripheral blood mononuclear cell (PBMC)-derived tumor necrosis factor α (TNFα) secretion assays, validating the biological relevancy of the IRAK1 target engagement as a biomarker of the IRAK4 activity. Plate uniformity and potency reproducibility evaluations demonstrated that this assay is amenable to high throughput. Using Bland–Altman assay agreement analysis, we demonstrated that incorporating such proximal pharmacological assessment of cellular target engagement to an in vitro screening funnel for SAR studies can prevent compound optimization toward off-target activity.

Keywords

target engagement IRAK1 IRAK4 high throughput assay agreement

Introduction

Interleukin-1 receptor-associated kinase 4 (IRAK4) is a serine/threonine protein kinase that plays an essential role in transducing signals from pathogen recognition toll-like receptors (TLRs) and the interleukin-1 (IL1) receptor superfamily to bridge innate and adaptive immune responses against microbial infection.¹ Dysregulation of the IRAK4 signaling pathway triggers aberrant expression of inflammatory cytokines/chemokines and hyperactivation of B cells, leading to multiple chronic inflammatory and autoimmune diseases such as multiple sclerosis (MS), systemic lupus erythematosus (SLE), rheumatoid arthritis (RA), and inflammatory bowel disease.^2–5 IRAK4 signaling has also been implicated in the initiation, development, and therapeutic responses of various cancers.^6–9 Abrogation of IRAK4 activity using either selective IRAK4 inhibitors or kinase-inactive mutants has demonstrated attenuation of disease activity in mouse models of RA, SLE, and experimental autoimmune encephalomyelitis (EAE).^4,10–12 In contrast to Irak4^–/– mice that suffer severe defects in response to broad bacterial and viral challenges,¹ human IRAK4 deficiency predisposes patients to a narrow range of pyogenic bacterial infection without severe disease manifestation to viral, parasitic, or fungal infection. Although life-threatening, such infections occur only in infancy and early childhood, which can be prevented with prophylactic antibacterial treatments.¹³ As patients grow into adolescence and adulthood, IRAK4 becomes dispensable for host defense against most pathogens.^13,14 These observations suggested that IRAK4 could be an attractive target for treating inflammatory, autoimmune diseases or cancer without increasing the risk of microbial infections.

To identify and optimize IRAK4 inhibitors, we have implemented a biochemical IRAK4 IMAP fluorescence polarization (FP) screening assay to assess compound binding and inhibition of IRAK4 kinase activity, followed by a phenotypic human peripheral blood mononuclear cell (PBMC) tumor necrosis factor α (TNFα) secretion assay to demonstrate functional perturbation of the IRAK4 pathway.¹⁵ Although biochemical and biophysical assays using purified recombinant protein to evaluate target engagement provided a robust, high-throughput, first-pass screening, the results did not always translate into cellular efficacy due to the lack of a physiological cellular context for the target.¹⁶ In addition, relying on distal phenotypic cellular readouts while omitting cellular proximal target engagement during SAR studies may derail the optimization of the molecule toward the off-target activity, as multiple signaling pathways could contribute to the regulation of a functional phenotype.

One approach to improve the probability of success is to measure direct drug–target interaction in living cells. Multiple methods have been developed, such as proximity detection using either labeled ligands,¹⁷ resonance energy transfer like fluorescence resonance energy transfer–fluorescence lifetime imaging (FRET-FLIM) and bioluminescence resonance energy transfer (BRET),^18,19 or protein stability-based enzyme fragment complementation (EFC). Although these approaches become pivotal in target identification and target validation, they require modifications of the ligand and/or target. Assays that depend on protein modification require generation of stable cell lines that may alter the physiological cellular context due to overexpression of the tagged target. Recent adaptation of thermal shift assay to cells (CETSA) allows label-free investigation of drug binding to native target in the physiological cellular environment.²⁰ Coupled with immunodetection, robust CETSA can be performed in high throughput, as demonstrated through library screening to identify B-Raf and poly (ADP-ribose) polymerase 1 (PARP1) binders.²¹ However, because CETSA measures changes in thermostability of the protein upon drug interaction, such changes do not always translate into functional modulation of the target in a cellular environment.²¹ It is also possible that small-molecule binding that results in a functional modulation of the target may not induce detectable changes in thermostability. Therefore, evaluation of functional modulation on the most proximal event following target activation is of paramount importance during lead optimization.

To complement the IRAK4 recombinant kinase assay and the functional PBMC-derived TNFα secretion assay, we considered a proximal cellular approach to demonstrate IRAK4-mediated drug–target engagement under physiological conditions. In the TLR/IL1R-initiated signaling cascade, multiple myeloid differentiation primary response factors (MyD88) bind to the receptor and assemble into a scaffold to promote the recruitment of other adaptor proteins and kinases, including IRAK-1 and IRAK-4, forming the Myddosome complex.^22,23 Within this complex, dimerized IRAK4 undergoes activation through autophosphorylation,^24,25 leading to subsequent IRAK1 phosphorylation and activation.²⁶ Hyperphosphorylated IRAK1 attracts and activates the E3 ubiquitin ligase tumor necrosis factor–associated factor 6 (TRAF6), which further cascades the signal to nuclear factor κB (NF-κB) and mitogen-activated protein kinase (MAPK) pathways following K63-mediated polyubiquitination.^27,28 Activated IRAK1 subsequently undergoes K48-linked polyubiquitination that directs it to proteasome-mediated degradation.^28–30 Although in mouse cells IRAK4 catalytic activity is critical for maximal production of inflammatory cytokines,^31–33 pharmacological inhibition with IRAK4 inhibitors or rescue of IRAK4-deficient cells with kinase-inactive IRAK 4 mutants illustrated that an IRAK4-mediated scaffold function was critical for Myddosome assembly and NF-κB or MAPK pathway activation, while human IRAK4 kinase activity was dispensable for cytokine production.^35–36 Hence, an assay that measures IRAK 4 phosphorylation may not fully capture different mechanisms of IRAK4 inhibitors. However, the recruitment of IRAK1 and its subsequent activation depends on both the scaffold functionality and kinase activity of IRAK4.^30,37,38 Evaluating IRAK1 activation, which is the most proximal event mediated by IRAK4, would potentially capture multiple mechanisms of inhibitors, including the disruption of IRAK4 catalytic activity and the Myddosome assembly.

Immunodetection has been used for quantitative measurement of protein markers in various complex biological matrices such as serum, tissues, or cellular lysates. However, conventional methods like enzyme-linked immunosorbent assay (ELISA) utilizing absorbance readouts require multiple time-consuming steps, a major disadvantage when adapting to automation and throughput testing, which are the key considerations when developing an assay for drug screening and development. The electrochemiluminescence (ECL)-based mesoscale assay (MSD) utilizes a streptavidin-coated microplate with built-in carbon electrodes at the base in close proximity to the analytes captured by analyte-specific biotin-labeled antibodies. Using a ruthenium-labeled detection antibody, luminescence signals are generated via multiple rounds of oxidative-reductive electrochemical reactions of ruthenium with tripropylamine provided in the buffer. High sensitivity and a wide dynamic range from the luminescence signal make the assay amenable to miniaturization. Here we describe the development of a high-throughput MSD-based human cellular endogenous IRAK1 activation assay. We validated the assay through an orthogonal method and demonstrated the biological relevance, as well as the potential value for inclusion in the SAR screening funnel for selective IRAK4 inhibitor optimization.

Materials and Methods

Cell Culture and Stimulation

Human mantle cell lymphoma JEKO-1 cells (ATCC, Manassas, VA) and diffuse large B-cell lymphoma line OCI-LY3 (DSMZ, Germany) were maintained in RPMI-1640 (Gibco, Gaithersburg, MD) with 20% fetal bovine serum (FBS; Fisher Scientific, Pittsburgh, PA). Human myeloma cell line RPMI-8226 (ATCC) was maintained in RPMI-1640 supplemented with 10% FBS. To stimulate the cells with TLR7 agonist R848 (Tocris, Minneapolis, MN), cells were cultured in 384-well plates in RPMI-1640 supplemented with 1× NEAA (nonessential amino acids; Gibco), 1 mM sodium pyruvate, 10 mM HEPES, and 0.5% charcoal/dextran-treated FBS (Fisher Scientific) at a density of 8 × 10⁵/mL for 16 h. Cells were treated with IRAK4 inhibitors for 1 h before stimulation with R848. Human lung fibroblast MRC5 cells (ATCC) were maintained in MEM (minimum essential medium; Gibco) with 10% FBS. To stimulate MRC5 cells with IL1-β, cells were plated in 384-well plates at a density of 50,000/well in MEM (Gibco) supplemented with 1× NEAA, 1 mM sodium pyruvate, 10 mM HEPES, and 0.5% charcoal/dextran-treated FBS overnight. Cells were treated with IRAK4 inhibitors for 1 h before stimulation with recombinant human IL1-β (R&D, Minneapolis, MN).

Immunoprecipitation

JEKO-1 cells were lysed in lysis buffer containing 150 mM NaCl, 20 mM Tris-HCl, pH 7.5, 1 mM EGTA, 1 mM EDTA, 1% Triton X-100, EDTA-free protease inhibitor cocktail (Roche, St. Louis, MO), phosphatase inhibitor II (Sigma-Aldrich, St. Louis, MO), phosphatase inhibitor III (Sigma-Aldrich), and 1 mM NaF (Sigma-Aldrich). Cells were lysed on ice for 30 min, followed by centrifugation at 4 °C, 14,000g for 10 min. Supernatants were collected and stored at –80 °C. Total protein concentration was quantified using Pierce’s BCA reagent (Thermo Fisher, Waltham, MA). IRAK1 in the cell lysate was pulled down with anti-IRAK1 antibody using the Dynabead Protein G immunoprecipitation kit protocol (Invitrogen, Carlsbad, CA). Briefly, 50 µL of Dynabeads Protein G was resuspended in 200 µL of antibody binding and washing solution containing 3 µg of anti-IRAK1 antibody. Antibody binding to the beads was facilitated by incubation for 10 min at room temperature on a rotator. Excess antibody was removed after placing the tube on the magnet and washing with antibody binding and washing buffer. Cell lysate with 500 µg of total protein was added to the beads/antibody complex and incubated overnight at 4 °C on a rotator. After washing with washing buffer three times, the Dynabeads/IRAK1/anti-IRAK1 complex was resuspended in 15 µL of elution buffer. The complex was separated from the beads after addition of 15 µL of premixed 2× NuPAGE LDS Sample Buffer (Invitrogen) and NuPAGE Sample Reducing Agent (Invitrogen), followed by heating for 10 min at 70 °C. Twenty-five microliters of eluants was loaded to sodium dodecyl sulfate–polyacrylamide gel electrophoresis (SDS-PAGE) gel for electrophoresis and Western blotting probing with corresponding antibodies.

Western Blot Analysis

Cell lysates were heat denatured at 70 °C for 10 min with 1× NuPAGE LDS sample buffer. Twenty micrograms of total proteins was loaded per lane of a 4%–12% NuPAGE Bis-TRIS mini gel (Invitrogen). Proteins on the gel were then electrophoretically transferred to 0.45 µm nitrocellulose membrane in 1× NuPAGE transfer buffer (Invitrogen) with 20% methanol for 1 h. The membrane was rinsed in 1× phosphate-buffered saline (PBS) and blocked with Odyssey blocking buffer (LI-COR Biosciences, Lincoln, NE) for 1 h at room temperature, followed by incubation with primary antibody diluted in Odyssey blocking buffer with 0.2% Tween-20 overnight at 4 °C on a shaker. The antibodies used were anti-human IRAK1 (1:1000; SC7883, Santa Cruz Biotechnology, Dallas, TX), anti-human IRAK4 (1:1000; ADI-KAP-ST206-E, Enzo Life Sciences, Farmingdale, NY), and anti-human phospho-IRAK4 (1:350; 11927, Cell Signaling Technology, Danvers, MA). Antibody reactive bands were detected after incubation with IRDye 800CW goat anti-rabbit IgG secondary antibody diluted in Odyssey blocking buffer with 0.2% Tween-20 at 1:15,000 dilution for 1 h at room temperature and scanned on an Odyssey CLx infrared imaging system (LI-COR Biosciences). The membrane was washed for 5 min in 1× PBS-T (PBS supplemented with 0.1% [v/v] Tween-20) three times after each antibody incubation. For reprobing, the blot was stripped with NewBlot IR stripping buffer (LI-COR Biosciences) for 20 min at room temperature. For capillary-based simple Western assays, samples were run on Peggy Sue system using the anti-rabbit detection module (ProteinSimple, San Jose, CA) according to the manufacturer’s protocol. The antibodies used were anti-IRAK1 SC7883 (1:50) and anti-β-actin (1:50; 4967, Cell Signaling Technology). Protein expression was quantified as the area under the curve of chemiluminescence signal using Compass software (ProteinSimple).

ECL Mesoscale Assay

Antibodies were labeled with biotin using the ChromaLink Biotin Antibody Labeling Kit (TriLink BioTechnologies, San Diego, CA) according to the manufacturer’s protocol. Briefly, bovine serum albumin (BSA)-free antibodies were buffer exchanged into 1× modification buffer through a Zeba 7K MWCO column (Thermo Fisher Scientific), adjusted to 1 mg/mL concentration following protein quantification using BCA reagent. Antibodies were incubated with sulfo-chromaLink biotin for 1 h at room temperature. Excess labeling reagents were removed by passing the reaction mixture through a Zeba 40K MWCO column (Thermo Fisher). Labeling antibodies with ruthenium (II) tris-bipyridine, N-hydroxysuccinimide followed the MSD Gold Sulfo-TAG-NHS-Ester labeling protocol provided by the vendor (Meso Scale Discovery, Rockville, MD). BSA-free antibodies were buffer exchanged into 1× PBS (pH 7.9) through a Zeba 7K MWCO column. Antibody concentrations were adjusted to 2 mg/mL following protein quantification with BCA reagent. Sulfo-TAG was dissolved in distilled H₂O. Based on the total amount of antibody being labeled, an appropriate amount of the Sulfo-TAG was added to the antibody solution, followed by an incubation at 23 °C for 2 h protected from light. Unlabeled Sulfo-TAG was removed by passing the solution through a 40K MWCO Zeba column. Labeled antibodies were stored at 4 °C in 1× PBS (pH 7.4 with 0.01% sodium azide).

To test the inhibition of IRAK1 activation by IRAK4 inhibitors, compounds were serially diluted in DMSO and predispensed into each well of a 384-well tissue culture plate using Echo Acoustic Liquid Handlers (Labcyte, Indianapolis, IN). Cells were seeded in corresponding culture media at 35 µL/well, and the plate was incubated for 1 h in a humidified 37 °C incubator with 5% CO₂. Cells were stimulated by the addition of 4 µL of 10× stimulants and incubated under the same conditions for the required amount of time. To lyse the cells, 35 µL of media was removed, followed by the addition of 25 µL of lysis buffer. After incubation on a shaker for 30 min at 4 °C, the plate was centrifuged at 1000g for 5 min. Twenty microliters of cell lysate was transferred to a 384-well MSD streptavidin-coated plate (Meso Scale Discovery), which was preblocked with 3% BSA for 1 h at room temperature. Biotin-labeled capture antibody and ruthenium-labeled detection antibody were diluted in PBS, 0.2% Tween-20, and 2% BSA at 3× final concentration. After adding 10 µL of antibody mixture to each well, binding was facilitated by incubation at 4 °C on a plate shaker overnight and protected from light. The plate was washed three times in buffer containing 50 mM Tris-HCl, pH 7.5, 150 mM NaCl, and 0.1% Tween-20 and read on a MESO SECTOR S600 plate reader (Meso Scale Discovery) to record the ECL signal following addition of 35 µL of 2× MSD Read Buffer T (Meso Scale Discovery).

IRAK4 Fluorescence Polarization Kinase Assay

IRAK4 kinase activity was measured as described by Lim et al.¹⁵ Briefly, in 20 µL of reaction mixture, 0.5 nM GST-tagged human recombinant IRAK4 (SignalChem, Richmond, BC, Canada) was incubated with 100 nM 5-FAM-labeled peptide substrate (Molecular Devices, San Jose, CA) in a buffer containing 10 mM Tris-HCl, pH 7.2, 1 mM DTT, 1 mM MgCl₂, and 0.01% Tween-20. The kinase reaction was initiated by the addition of 100 µM ATP, and after 30 min of incubation at 25 °C, the reaction was stopped with 60 µL of IMAP Progressive Binding Reagent (Molecular Devices). Phosphorylated peptide binding to metal Mn(^III) immobilized on nanoparticles was conducted by incubation for 1 h, and the change of FP of the peptide was quantified on Analyst HT (Molecular Devices).

Human PBMC TNFα Release in Response to IL1β Stimulation

Frozen human PBMCs were recovered overnight in RPMI culture medium supplemented with 0.5% FBS, 2 mM Glutamax, 1× MEM NEAA, and 1 mM sodium pyruvate. Twenty thousand PBMCs in 36 µL of culture medium were seeded per well of a 384-well plate and preincubated with test compounds for half an hour in 5% CO₂ incubator at 37 °C. Cells were then stimulated with 4 µL of IL1β at a final concentration of 2 ng/mL for 5 h in the incubator, followed by the addition of 50 µL of medium and mixing. After centrifugation at 1000g for 5 min, 15 µL of medium was transferred to a human TNFα 384-well plate that was precoated with TNFα capture antibody (Meso Scale Discovery) and blocked with 3% BSA. TNFα was detected followed the manufacturer’s instructions. Briefly, after overnight at 4 °C, ruthenium (II)-labeled detection antibody was added and incubated for 2 h at room temperature, followed by three washes with 1× PBS–0.05% Tween-20. Data were acquired using a MESO SECTOR S600 plate reader.

Kinase Profiling

Kinase profiling was performed by Z′-LYTE Screening using Thermo Fisher’s SelectScreen Kinase Profiling Services. Two hundred fifty-six kinases ( Suppl. Table S2 ) were tested against three concentrations of IRAK4 inhibitors: 10, 1, and 0.1 µM.

Statistical Analysis

All plots and statistical analysis were performed in Prism (GraphPad Software, San Diego, CA). Test compound potencies (IC₅₀) were calculated from four-parameter logistic regression curve fitting of normalized concentration response data. For normalization, 100% inhibition was defined by signals from no stimulation control, while 0% inhibition was defined by signals from agonist stimulation in the absence of compound treatment. Both controls received 0.1% DMSO as in the compound treatment. The Pearson correlation coefficient was calculated from active compounds in both assays analyzed. For kinase profiling, the efficient IC₅₀ of each kinase was derived from variable slope logistic regression curve fitting with a defined top of 100% inhibition and bottom of 0% inhibition.

For assay performance analysis, Z′ was calculated as described by Zhang et al.³⁹

Z^{'} = 1 - \frac{3 SD n o s t i m u l a t i o n + 3 SD s t i m u l a t e d}{Mean of n o s t i m u l a t i o n - Mean of s t i m u l a t e d}

(1)

To evaluate reproducibility of the assay, 24 compounds were subject to two independent concentration response runs. The minimum significant ratio (MSR) from this test–retest experiment was calculated as per the Assay Guidance Manual,⁴⁰ where the difference of paired log potency (IC₅₀) from each compound was first calculated to derive the standard deviation (SD) for MSR calculation:

MSR = 10^{2 S D}

(2)

The limit of agreement (LsA), defined by Bland–Altman,⁴¹ was computed to determine statistical prediction of the potency ratio limits and calculated as described by Eastwood et al.⁴²

MeanRatio (MR) = 10^{\bar{d}}

(3)

LsA = (M R / MSR, MR \times MSR)

(4)

RatioLimits (RLs) = 10^(\bar{d} \pm \frac{2 S D}{\sqrt{N}})

(5)

$\bar{d}$ is the mean difference of log potency between the two runs for each compound, and N is the number of compounds tested.

Bland–Altman analysis was used to quantify the agreement of data sets by plotting each tested compound’s mean log potency from the two assays against their difference. Acceptable limits of agreement were defined by 2SD_agree that took into consideration the combined variation from both assays. To obtain SD_agree, each assay’s precision level was first determined through repeated testing of reference compounds and calculated as SD_pooled as described by Sun et al.:⁴³

{SD}_{pooled} = \frac{\sum_{i = 1}^{k} (n_{i} - 1) {SD}_{i^{2}}}{\sqrt{\sum_{i = 1}^{k} (n_{i} - k)}}

(6)

{SD}_{agree} = \sqrt{{SD}_{pooled 1}^{2} {+ SD}_{pooled 2}^{2}}

(7)

where k is the number of reference compounds tested, n is the number of repeats for each compound, and SD is the standard deviation of log potency.

Results

Development and Optimization of a High-Throughput IRAK1 Activation Assay

Signal transduction through IRAK4 is stimulated upon ligand activation of the TLR/IL1 receptor superfamily. To identify human cells that are responsive to IRAK4 pathway activation, we treated several cell lines that express either TLR7 or IL1R with selective agonists, including IL1β, R848, and lipopolysaccharide (LPS) to stimulate IL1R, TLR7, and TLR4, respectively. The experimental hallmark of IRAK1 activation is the disappearance of an unphosphorylated IRAK1 band on immunoblot due to activation through hyperphosphorylation, ubiquitination, and subsequent proteasomal degradation.³¹ Using Western blotting, we evaluated unphosphorylated IRAK1 levels in cell lysates following receptor activation, along with induction of phospho-IRAK4. Although we did not observe R848-dependent phospho-IRAK4 upregulation over time in mantle cell lymphoma JEKO-1 cells, progressive disappearance of unphosphorylated IRAK1 was detected as early as 20 min and sustained for up to 3 h poststimulation ( Fig. 1A ). In contrast, RPMI-8226 and OCL-Ly3 cells did not respond to R848 stimulation-induced IRAK1 activation (Suppl. Fig. S1). We also evaluated the responses of MRC5, a human lung fibroblast line, to various IL1/TLR stimuli. Figure 1B showed that IRAK1 activation was detected following IL1β treatment; this was not observed on R848 or LPS stimulation (Suppl. Fig. S2). The kinetics of the loss of unphosphorylated IRAK1 in this cell line was much faster compared with that in JEKO-1 cells mediated through TLR7, as IRAK1 was barely detectable after 15 min of treatment ( Fig. 1B ). A transient upregulation of phospho-IRAK4 was also observed in IL1-β-induced MRC5 activation (Fig. 1B). JEKO-1 and MRC5 cell lines were chosen for developing a high-throughput endogenous IRAK1 activation assay as our cell-based proximal functional assessment of IRAK4 inhibitors, considering their sustained IRAK1 activation following receptor stimulation.

Figure 1.

Analysis of IRAK1 activation and IRAK4 phosphorylation to identify cell lines responsive to IRAK4 pathway stimulation. (A) IRAK1 activation in response to TLR7 agonist R848 stimulation in JEKO-1 cells, indicated by the loss of detectable IRAK1. Cells were stimulated with R848 at 10 µg/mL for 2–180 min. Cell lysates were subject to 12% SDS-PAGE and then immunoblotted with anti-pIRAK4, anti-IRAK4, and anti-IRAK1 sequentially. (B) Sustained IRAK1 activation and transient IRAK4 phosphorylation in MRC5 cells in response to IL1β treatment. Cells were stimulated with IL1β (100 ng/mL) for the indicated periods of time. Cell lysates were electrophoresed through 12% SDS-PAGE, followed by immunoblots with anti-IRAK1, anti-phospho-IRAK4, or anti-IRAK4.

To identify antibodies that recognize native IRAK1 in cell lysate for developing the MSD-based immunoassay, we screened 18 commercially available antibodies by immunoprecipitation of JEKO-1 cell lysate ( Suppl. Table S1 ). The cells were lysed in a buffer containing mild nonionic detergent 1% Triton X-100 to keep IRAK1 in its native form and preserve the protein complex. Coupled with Western blot, we found that seven antibodies were able to pull down IRAK1 ( Fig. 2A ). To further develop ECL-based detection of total IRAK1, three antibodies were labeled with biotin and all seven antibodies were labeled with ruthenium. A grid-based screening of antibody pairs identified five pairs generating detectable ECL signals statistically above the background ( Fig. 2B–D ). Among the five pairs of antibodies that generated a positive stimulation assay window, biotin-744/Ru-M01 displayed the lowest cumulative background contributed by either biotin-744 or Ru-M01, with a signal ratio from nonstimulated cell lysate to cumulative background of 14 RLU and an assay window to assess IRAK1 activation, calculated from the ratio of nonstimulated to stimulated signals, above 3 (Fig. 2E). Cross-titration of biotin-744 and Ru-M01 was performed to optimize antibody concentrations, along with testing of both stimulated and nonstimulated cell lysates to evaluate the assay window. While Ru-M01 showed a dose-dependent increase of signals up to 4.3 µg/mL final concentration, biotin-744 displayed a bell-shaped curve, with a plateau around 0.16 µg/mL ( Fig. 2F ). The optimal antibody concentrations were determined to be 0.16 µg/mL biotin-744 and 0.08 µg/mL Ru-MO1, where a maximal assay window of approximately 9 was obtained. To determine the detection range of the assay, JEKO-1 cells were seeded at various densities and their subsequent lysates were subject to ECL detection using the optimal antibody concentrations. As shown in Supplemental Figure S3 , signal intensity exhibited a linear relationship with increased cell numbers.

Figure 2.

Antibody selection and optimization for the high-throughput IRAK1 activation assay. (A) Immunoprecipitation of IRAK1 from JEKO cell lysates using anti-IRAK1 antibodies ( Suppl. Table S1 ) to screen antibodies that recognize endogenous native IRAK1. Immunoprecipitated proteins were separated on SDS-PAGE (12%) and IRAK1 was detected following immunoblot with anti-IRAK1, SC7833 (1:50). (B–D) Screening of antibody pair for mesoscale ECL detection of IRAK1 in JEKO-1 cells. Cells were cultured for 2 h in the presence (Sti.) or absence (Non.) of R848 stimulation (10 µg/mL). Cell lysates were incubated with Ru-labeled anti-IRAK1 detection antibody and biotin-labeled anti-IRAK1 capture antibody in 384-well mesoscale streptavidin-coated plates. (B) Ru-labeled antibodies paired with the biotin-labeled antibody 743. (C) Ru-labeled antibodies paired with the biotin-labeled antibody 744. (D) Ru-labeled antibodies paired with the biotin-labeled antibody 709. (E) Background (BK) and assay window assessments from selected antibody pairs in mesoscale ECL detection of IRAK1. Background signals were evaluated by incubating cell lysates with either Ru-labeled antibody or Biotin-labeled antibody. Signal/BK, ratio of signal from nonstimulated samples to total background signals from both Ru-labeled and Biotin-labeled antibodies; assay window, ratio of signals from nonstimulated to R848 stimulated samples. Among the five antibody pairs that generated a positive assay window, MO1_Ru and 744_Biotin pair showed the lowest cumulative background with an assay window above 3. (F) Cross-titration of biotin-labeled antibody_744 and Ru labeled antibody_MO1 to optimize antibody concentrations for mesoscale ECL detection of IRAK1. JEKO-1 cells were cultured for 2 h in the absence (Non.) or presence (Sti.) of R848 (10 µg/mL) stimulation. Cell lysates were incubated with various concentrations of MO1_Ru and 744_Biotin labeled antibodies in 384-well mesoscale streptavidin-coated plates. The optimal antibody concentrations for IRAK1 detection are identified as 0.16 µg/mL 744_Biotin and 0.08 µg/mL MO1_Ru to achieve an assay window of 9.

Discordance between ECL and Western Blot Quantification to Assess IRAK4 Inhibition of IRAK1 Activation in JEKO-1 Cells Stimulated with TLR7 Agonist R848

With the optimized assay conditions established for quantifying IRAK1 activation stimulated by R848 in JEKO cells, we next titrated R848 in order to determine the maximal stimulation concentration. Figure 3A shows that 50% stimulation (EC₅₀) was achieved with 0.32 µg/mL R848, while 10 µg/mL induced a maximal activation of IRAK1. To evaluate whether inhibition of IRAK4 activity can prevent IRAK1 activation, cells were pretreated with IRAK4 inhibitor A ( Fig. 3B ) at 10 µM for 1 h before stimulation with R848 for 2 h. IRAK4 inhibitor was chosen for its potency on inhibiting IRAK4 kinase activity as evaluated by an in vitro IMAP FP kinase assay using recombinant IRAK4 and its substrate peptide (IC₅₀ = 4.5 ± 1.6 nM). This inhibitor also showed high selectivity over other kinases in a counterscreen panel including 265 kinases (96% >100-fold selectivity) (Suppl. Table. S2). Three doses of R848 corresponding to 1, 3, and 10 µg/mL were applied to reach EC₈₀, EC₉₅, and maximal stimulation, respectively. ECL signals showed that although IRAK4 inhibitor was able to prevent the disappearance of IRAK1, the effect was incomplete across all three doses of R848 stimulation ( Fig. 3C ). This phenomenon of partial rescue on unphosphorylated IRAK1 by IRAK4 inhibitor was not due to the loss of cell mass following the compound treatment, as total protein amounts were not affected by the compound treatment (Suppl. Fig. S4). To validate this observation, we used a simple Western detection method (Peggy Sue), a quantitative matrix–capillary-based protein separation, and immunodetection under denatured conditions. JEKO-1 cells were treated with R848, and cells were lysed in the ECL detection lysis buffer. One hundred nanograms of total denatured protein, which was predetermined to be within the linear detection range (Suppl. Fig. S5), was loaded for separation through the matrix and immunodetection with either anti-IRAK1 antibody or anti-β-actin antibody. Following quantification using the peak area of the signals and normalization to β-actin expression, a four-parameter concentration–response curve fitting determined the EC₅₀ of R848-induced IRAK1 activation to be 0.58 µg/mL ( Fig. 3D ), in agreement with the value obtained by ECL detection (EC₅₀ = 0.32 ± 0.016 µg/mL) ( Fig. 3A ). In contrast to ECL detection, we observed a full rescue of unphosphorylated IRAK1 following IRAK4 inhibitor treatment in JEKO-1 cells. The time course of R848 stimulation showed a complete loss of IRAK1 after 2 h of stimulation, which was fully recovered when the cells were pretreated with 10 µM IRAK4 inhibitor for 1 h ( Fig. 3E ). To overcome partial detection of IRAK1 rescue in response to R848-mediated activation by the ECL method, we explored additional IRAK4-mediated activation using IL1β in the MRC5 cell line.

Figure 3.

Effect of IRAK4 inhibitor on IRAK1 activation in JEKO-1 cells stimulated with TLR7 agonist R848. (A) R848 titration to determine EC₅₀ of IRAK1 activation (n = 4). Cells were stimulated with R848 for 2 h. IRAK1 in cell lysates was subjected to mesoscale ECL detection. (B) Structure of IRAK4 inhibitor A. (C) Partial rescue of IRAK1 with IRAK4 inhibitor in JEKO-1 cells stimulated with R848. Cells were pretreated with IRAK4 inhibitor at 10 µM for 1 h and then stimulated with R848 for 2 h to induce IRAK1 activation, as shown by the reduction of detectable IRAK1. Cells were lysed and subject to mesoscale ECL detection (n = 3). (D) R848 titration to determine EC₅₀ of IRAK1 activation measured with simple Western detection. JEKO-1 cells were stimulated with R848 for 2 h. Cells lysates were run on Peggy Sue. IRAK1 was detected with anti-IRAK1 antibody SC7883 (1:50), and β-actin was probed with anti-β-actin antibody CST 4967 (1:50). Both proteins were quantified by the peak area of chemiluminescence signal. Data normalized to β-actin expression. (E) Full rescue of IRAK1 by IRAK4 inhibitor with simple Western detection. JEKO-1 cells were pretreated with 10 µM IRAK4 inhibitor for 1 h, followed by stimulation with 10 µg/mL R848 for 15–120 min. Cells lysates were run on Peggy Sue. IRAK1 was detected with anti-IRAK1 antibody SC7883 (1:50), and β-actin was probed with anti-β-actin antibody CST 4967 (1:50). IRAK1 was quantified by the peak area of chemiluminescence signal. Data normalized to β-actin expression. ND, not detected.

Effect of IRAK4 Inhibition on IRAK1 Activation Mediated by IL1β in MRC5 Cells

To evaluate how IRAK4 inhibition rescues IRAK1 in MRC5 cells, we pretreated MRC5 with IRAK4 inhibitor at 10 µM for 1 h, followed by IL1β (100 ng/mL) stimulation; R848 was omitted as MRC5 cells were unresponsive to this treatment. A time course of stimulation was performed and unphosphorylated IRAK1 was quantified with simple Western detection run as shown in Figure 4A ; more than 80% of IRAK1 was lost at 10 min of stimulation, which was sustained up to 3 h ( Fig. 1B ). IRAK4 inhibitor effectively blocked IRAK1 activation, with a recovery of unphosphorylated IRAK1 to 70% of the prestimulation level at 10 min ( Fig. 4A ). Although titration of IL1β resulted in similar EC₅₀ values in stimulating IRAK1 activation throughout the time course study ( Fig. 4B ), a decreased level of IRAK1 rescue by IRAK4 inhibitor at maximal stimulation was observed with longer IL1β treatment ( Fig. 4A ). We then performed a stimulation time course study by reducing the IL1β to 30 ng/mL at the EC₉₀ level and reevaluated IRAK4 inhibition on IRAK1 using ECL detection. In contrast to JEKO-1 cells stimulated with R848, where there was a disconnect between simple Western detection and ECL; we observed a full rescue of unphosphorylated IRAK1 following IRAK4 inhibition after 10 min of stimulation with IL1β ( Fig. 4C ). Compound titrations of IRAK4 inhibitor were carried out concomitantly with both simple Western detection and ECL. To calculate IC₅₀ for the inhibitor from the concentration–response curve quantified by simple Western detection, we normalized IRAK1 signal to total protein loading and obtained an average IC₅₀ of 300 ± 92 nM ( Fig. 4D ), which was similar to the IC₅₀ obtained by ECL detection (averaged 439 ± 9.9 nM) ( Fig. 4E ). The concordance between the two detection methods on MRC5 cells validated the ECL approach for a high-throughput assay.

Figure 4.

Effect of IRAK4 inhibitor on IRAK1 activation in MRC5 cells stimulated with IL1β. (A) Retention of IRAK1 by IRAK4 inhibitor with simple Western detection. MRC5 cells were pretreated with IRAK4 inhibitor at 10 µM for 1 h before stimulation with 100 ng/mL IL1β for 2–30 min. Cell lysates were run on Peggy Sue. IRAK1 was probed with anti-IRAK1 SC7883 (1:50), and GAPDH was probed with anti-GAPDH CST 2118 (1:100). Proteins were quantified by a peak area of chemiluminescence signal. IRAK1 expression was normalized to GAPDH. Black bars, no IRAK4 inhibitor treatment; red bars, IRAK4 inhibitor treatment. (B) Time course and dose titration of IL1β to stimulate IRAK1 activation in MRC5 cells. Cells were treated with various concentrations of IL1β for 10–30 min. IRAK1 in cell lysates was measured with the mesoscale ECL method (n = 4). Dose titration resulted in similar EC₅₀ values at different time points of stimulation, indicating that the duration of stimulation from 10 to 30 min had minimal effect on the level of IRAK1 activation. (C) Rescue of IRAK1 by IRAK4 inhibitor in IL1β-stimulated MRC5 cells. Cells were seeded at 50,000/well of a 384-well plate and pretreated with IRAK4 inhibitor at 10 µM for 1 h. Cells were lysed after stimulation with IL1β at 30 ng/mL (EC90) for the times indicated. IRAK1 was evaluated with the mesoscale ECL method (n = 4). (D,E) Concentration response of IRAK4 inhibitor on IRAK1 activation in MRC5. Cells were pretreated with IRAK4 inhibitor for 1 h, followed by stimulation with IL1β at 30 ng/mL for 10 min. (D) IRAK1 was detected by simple Western assay and quantified with the peak area normalized to total protein loading. (E) IRAK1 was detected using mesoscale ECL. The two detection methods resulted in comparable IC₅₀ values of IRAK4 inhibitor.

Determination of Performance of IRAK1 Activation ECL Assay in High-Throughput Format

To assess the robustness of the ECL assay in high-throughput format, MRC5 cells were seeded into every well of a 384-well plate except the edge wells. Alternate columns of the cells were stimulated with either IL1β at 30 ng/mL corresponding to EC₉₀ or buffer control for 10 min, and the total IRAK1 was detected in cell lysates by the ECL method. With manual operation of the assay, the run resulted in a mean maximum signal of 7974 RLU from no IL1β stimulation with a coefficient of variation (CV) of 10%, and a minimum signal of 1795 RLU from IL1β stimulation with a CV of 14% ( Fig. 5A ). Excluding one obvious outlier from the high signal, the Z′ calculated from eq 1 was 0.47. The maximal column-to-column or row-to-row drifts from high signals were 15.3% and 19%, respectively ( Fig. 5A ). This plate uniformity and signal variability assessment demonstrated that the assay was acceptable for high-throughput screening with a CV <20% for both high and low signals, Z′ ≥0.4, and no significant material drift (<20%) according to the Assay Guidance Manual.⁴⁰ To test the reproducibility of IC_50, 24 compounds went through a 10-point half-log concentration response in two independent runs. A potency plot showed agreement between run 1 and run 2 with a correlation coefficient of log potency (Pearson r) of 0.94 (p < 0.0001) ( Fig. 5B ). MSR and LsA were also calculated according to the Assay Guidance Manual using eqs 2–5:⁴⁶ MSR = 2.5 and LsA = 0.45–2.83 ( Fig. 5C ). The reproducibility test illustrated that the assay passed both test criteria of MSR <3.0 and LsA between 0.33 and 3.0, per the Assay Guidance Manual.⁴⁰

Figure 5.

Quality assessments of the high -hroughput mesoscale IRAK1 activation assay in MRC5 cells. (A) Plate uniformity test under optimized conditions and manual operation. Cells were seeded into every well of a 384-well plate except the edge wells. Alternate columns of cells received 30 ng/mL IL1β stimulation for 10 min. IRAK1 was evaluated with mesoscale ECL detection. The solid line displays the mean signal of the samples, and the dotted line shows 3SD of the mean. The test resulted in an acceptable Z′ of 0.47 for high-throughput screening. (B,C) Potency replicate experiment with two independent runs. (B) potency correlation between run 1 and run 2. Correlation coefficient of log potency (Pearson r) = 0.97. (C) Bland–Altman plot and statistical analysis to calculate MSR, LsA, MR, and RLs as described in the methods: MSR = 2.6, LsA = 0.41–2.91, MR = 1.09, RLs = 0.89–1.35.

IRAK1 Activation Assay Evaluation for Biological Relevancy

Although we have demonstrated that the performance of the cell-based IRAK1 activation assay meets the criteria of a high-throughput screening assay, the biological relevance of the measured signal to IRAK4 and associated signaling pathway also needed to be validated in order to deploy the assay in SAR studies for the discovery of novel IRAK4 inhibitors. Twenty-seven compounds were evaluated in 10-point half-log concentration titrations starting at 10 µM, to determine their potency for inhibiting IRAK4 kinase activity in a biochemical IMAP FP assay, cellular IRAK1 activation in MRC5 stimulated by IL1β, and TNFα secretion from IL1β-treated human PBMCs. The selected test set included developing IRAK4 inhibitors from several structural series whose potencies span a range of 3 logs, and two negative control compounds, as demonstrated by their activities in the biochemical IRAK4 kinase IMAP FP assay ( Fig. 6A ). A scatterplot of potencies between the IRAK4 kinase assay and IRAK1 activation showed a positive correlation with the Pearson correlation coefficient (r = 0.64, p = 0.0072) and a bias toward higher potency in the IRAK4 kinase assay ( Fig. 6A ); this is comparable to the association observed between the IRAK4 kinase assay and IL1β-induced human PBMC TNFα secretion (r = 0.68, p = 0.0004) ( Fig. 6C ). There is a stronger positive correlation between the potencies from IRAK1 activation and PBMC TNFα secretion assays (r = 0.81, p = 0.0002) ( Fig. 6E ), both of which measured cellular signaling events following IRAK4 activation, with IRAK1 activation being a proximal functional biomarker while TNFα secretion is a phenotypical outcome of IRAK4 engagement. In addition, both negative control compounds failed to rescue the loss of unphosphorylated IRAK1 stimulated by IL1β in MRC5 cells ( Fig. 6A ), further demonstrating that the measured signal from the IRAK1 activation assay depends on IRAK4 target engagement.

Figure 6.

Biological validation of IRAK1 activation assay using agreement analysis. Twenty-seven compounds were run through the IRAK4 in vitro kinase activity IMAP FP assay, IRAK1 activation in MRC5 cells was stimulated by IL1β, and human PBMC TNFα secretion was stimulated by IL1β. Each compound’s IC₅₀ was calculated from a 10-point concentration titration and four-parameter logistic regression curve fitting. (A) Correlation and (B) Bland–Altman agreement analysis between the IRAK1 activation assay and IRAK4 in vitro kinase assay. (C) Correlation and (D) Bland–Altman agreement analysis between the IRAK1 human PBMC TNFα secretion assay and IRAK4 kinase assay. (D) Correlation and (E) Bland–Altman agreement analysis between the IRAK1 activation assay and human PBMC TNFα secretion assay. Red dot, negative control compounds; blue dot, IRAK4 selective inhibitor compound A; green dot, nonselective compound B; correlation plot r, correlation coefficient of log potency (Pearson r); red dotted line, acceptable limits of agreement defined by 2SD_agree; black line, mean difference; black dotted line, corresponds to zero difference. The calculation of SD_agree is described in Materials and Methods.

To define the degree of agreements between the assays, we performed Bland–Altman analysis with the 16 compounds that showed activity in all three assays and calculated acceptable limits of agreement defined by 2SD_agree that considers combined variations of the two plotted assays.⁴³ We first determined each assay’s precision level using SD_pooled (eq 6) from two reference compounds that were run multiple times in either the IRAK4 kinase IMAP FP or human PBMC TNFα secretion assay, and the IC₅₀ of 24 compounds from the test–retest runs of the IRAK1 activation assay described above ( Suppl. Table S3 ). SD_agree was then calculated using eq 7, described in Materials and Methods. The Bland–Altman graph plotted the difference against the mean of log potencies from the two analyzed assays for each tested compound and quantified bias between the assays. Figure 6B,D showed a consistent bias toward the biochemical IRAK4 kinase assay from either the IRAK1 activation (bias = 2.33) or the human PBMC TNFα secretion (bias = 1.97), corresponding to potency shifts of 214-fold and 93-fold, respectively. About 50% of compounds were found to fall within the limits of agreement defined by 2SD_agree while comparing IRAK4 biochemical assay results to either cell-based assays. The potency shifts between the cell-based IRAK1 activation and human PBMC TNFα secretion were much smaller, with a consistent bias of 0.37 (2.3-fold) toward the PBMC TNFα secretion in response to IL1β stimulation ( Fig. 6F ). Among the 16 compounds tested, 56% fell within the limits of agreement, indicating that the two cellular assays were not in full agreement and therefore the importance of running both assays in SAR studies.

Discussion

A review of clinical trials found that a lack of efficacy accounted for more than 50% of phase II and III failures;⁴⁴ this is partly attributed to the absence of drug target engagement or target-specific pharmacological intervention in humans.^45,46 For example, several unsuccessful phase III clinical trials of iniparib, a compound initially developed as a non-NAD⁺ competitive PARP prodrug that showed tumoricidal activity in cancer cell lines, was later demonstrated not to be a bona fide PARP inhibitor, but rather it nonselectively modifies cellular cysteine-containing proteins,^47,48 highlighting the importance of confirming drug target engagement and mechanism of action in cells during early phase preclinical drug development. To support IRAK4 inhibitor discovery, we developed an ECL-based endogenous IRAK1 activation assay as a proximal functional target engagement assessment for IRAK4 inhibitors.

During the assay development, we observed a disconnect between ECL and simple Western results on evaluating IRAK4 inhibitor in JEKO-1 cells. This phenomenon could be attributed to the difference between the two assay conditions. In ECL, a mild nondenaturing lysis buffer containing nonionic detergent Triton X-100 likely preserved IRAK1 in three native conformations: free IRAK1, recruited to the Myddosome, or activated through phosphorylation and ubiquitination. In simple Western immunoblot analysis, addition of ionic detergent and sample denaturing disrupted all protein–protein interactions. Comparable EC₅₀ values of the agonist R848 stimulation obtained from both detection methods (0.58 µg/mL vs 0.32 µg/mL) indicated that the two antibodies used in ECL were able to capture both free IRAK1 and IRAK1 in the Myddosome complex. However, binding of the IRAK4 kinase inhibitor likely affected interactions among the proteins in the Myddosome, rendering the antibodies incapable of recognizing IRAK1 in the complex. Hence, a partial recovery of IRAK1 signal following complete inhibition of IRAK4 kinase activity was observed in ECL, while a complete recovery was detected by Western blot. Indeed, inactivation of IRAK4 kinase activity has been reported to change the binding affinities among the proteins and increase the stability of the Myddosome structure in LPS-stimulated TLR4 signaling.³⁴ The assembly of different adaptor proteins into Myddosomes during various TLR or ILR activations⁴⁹ could also affect antibodies’ ability to recognize IRAK1 in the complex and would explain the different results of IRAK4 inhibitor on IRAK1 detection by ECL in MRC5 cells. Although IRAK1 levels could be fully recovered by IRAK4 inhibitor under short-term IL1R stimulation in MRC5 cells, receptor endocytic regulation following persistent stimulation and lysosome degradation⁵⁰ could cause the observed progressive loss of full recovery on IRAK1 (Fig. 4C). Our results showed the limitation of assays to explore a complex biological process and emphasized the importance of validating the biological relevance of an assay using orthogonal methods to rule out artifacts.

Plate uniformity and potency reproducibility studies showed that the assay is acceptable for testing IRAK4 inhibitors in high throughput. Sixteen IRAK4 inhibitors from multiple structural series and with potencies covering a wide range were selected to perform assay agreement analysis. We observed a large systematic potency shift toward the recombinant IRAK4 biochemical kinase assay performed at an ATP Km of 100 µM. Such a bias could be caused by compound cellular permeability or a higher cellular ATP concentration (1–10 mM) that decreases the apparent potency of ATP competitive inhibitors when compared with the biochemical assay result that measured IRAK4 catalytic activity in the presence of 100 µM ATP. With the consideration of variability from both assays, about half of the tested compounds were found to fall outside the limits of agreement between the biochemical assay and either the cellular IRAK1 activation or PBMC TNFα secretion assay. However, Pearson correlation plots showed that in evaluating IRAK4 inhibitors, the results from the IRAK1 assay to either the biochemical or PBMC TNFα cellular phenotypic assay are positively correlated, validating the biological relevancy that IRAK1 activation depends on IRAK4 engagement.

The decision on incorporating the assay into SAR studies depends on whether additional information can be provided for compound assessment. Theoretically, when the two assays measure signals that are purely generated from the same biological target, the results should be in agreement so that one assay can replace the other without losing information.⁴¹ However, between the two cellular assays, only 56% of the compounds tested were within the limits of agreement. This could be due to various causes, including selectivity and less optimized physiochemical properties during the early stage of lead discovery. For example, two compounds (A and B) were selected for a broad kinase panel profiling. Compound A turned out to be a highly selective IRAK4 inhibitor, with IC50 >100-fold selective against 98% of the kinases in the 265-kinase panel, while compound B was >100-fold selective against 66% of the kinases tested. In assay agreement analysis, compound A was within the limits of agreement among all three assays. Interestingly, compound B was within the limits of agreement between biochemical and IRAK1 activation assays, yet fell out of the limit of agreement for the biochemical to PBMC TNFα phenotypic assay or IRAK1 activation to the phenotypic assay analyses, a result reflecting that the phenotypic measurement was influenced by off-target activities. Although the IRAK1 activation assay is less prone to false-positive inhibitors than the downstream phenotypic assay, we should be aware that compounds stimulating dephosphorylation or preventing ubiquitination and proteasome degradation will appear as false positives. Our results suggested that proximal pharmacological assessment of cellular target engagement will be part of the in vitro screening funnel for SAR studies during early preclinical drug development, along with biochemical/biophysical assays and cellular phenotypic assays, especially at the early stage of lead optimization when compound cellular permeability and off-targets are largely unknown.

Supplemental Material

sj-pdf-1-jbx-10.1177_24725552211021074 – Supplemental material for Establishing and Validating Cellular Functional Target Engagement Assay for Selective IRAK4 Inhibitor Discovery

Supplemental material, sj-pdf-1-jbx-10.1177_24725552211021074 for Establishing and Validating Cellular Functional Target Engagement Assay for Selective IRAK4 Inhibitor Discovery by Yiping Chen, Dongyu Sun, Ruojing Yang, Jongwon Lim, Christopher Sondey, Jeremy Presland, Larissa Rakhilina, George Addona, Ilona Kariv and Hongmin Chen in SLAS Discovery

Footnotes

Acknowledgements

We thank Bruce Beutel and Melanie Kleinschek for their support of the project and Merck’s compound management group for compound preparation.

Supplemental material is available online with this article.

Declaration of Conflicting Interests

The authors declared the following potential conflicts of interest with respect to the research, authorship, and/or publication of this article: All authors are current employees of Merck & Co., Inc. (Kenilworth, NJ) and may own stock or stock options in Merck & Co.

Funding

The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This work was supported by Merck & Co., Inc. (Kenilworth, NJ, USA).

References

Suzuki

Duncan

G. S.

; et al. Severe Impairment of Interleukin-1 and Toll-Like receptor Signaling in Mice Lacking IRAK-4. Nature 2002, 416, 750–754.

Zheng

Chen

Chu

; et al. Inflammatory Role of TLR-MyD88 Signaling in Multiple Sclerosis. Front. Mol. Neurosci. 2020, 12, 314–326.

Celhar

H. K.

Benso

; et al. TLR7 Protein Expression in Mild and Severe Lupus-Prone Models Is Regulated in a Leukocyte, Genetic, and IRAK4 Dependent Manner. Front. Immunol. 2019, 10, 1546–1561.

Umar

Palasiewicz

Raemdonck

K. V.

; et al. IRAK4 Inhibition: A Promising Strategy for Treating RA Joint Inflammation and Bone Erosion. Cell. Mol. Immunol. 2020. DOI: 10.1038/s41423-020-0433-8.

Heiseke

A. F.

Jeuk

B. H.

Markota

; et al. IRAK1 Drives Intestinal Inflammation by Promoting the Generation of Effector Th Cells with Optimal Gut-Homing Capacity. J. Immunol. 2015, 195, 5787–5794.

Chen

Zhang

; et al. IRAK4 Mediates Colitis-Induced Tumorigenesis and Chemoresistance in Colorectal Cancer. JCI Insight 2019, 4, e130867.

Srivastava

Geng

Liu

; et al. Augmentation of Therapeutic Responses in Melanoma by Inhibition of IRAK-1,-4. Cancer Res. 2012, 72, 6209–6216.

Giménez

Schulz

Higashi

; et al. Targeting IRAK4 Disrupts Inflammatory Pathways and Delays Tumor Development in Chronic Lymphocytic Leukemia. Leukemia 2019, 34, 100–114.

Yeyeodu

S. T.

Kidd

L. R.

Oprea-Ilies

G. M.

; et al. IRAK4 and TLR3 Sequence Variants May Alter Breast Cancer Risk among African-American Women. Front. Immunol. 2013, 4, 338–348.

10.

Koziczak-Holbro

Littlewood-Evans

Pöllinger

; et al. The Critical Role of Kinase Activity of Interleukin-1 Receptor-Associated Kinase 4 in Animal Models of Joint Inflammation. Arthritis Rheum. 2009, 60, 1661–1671.

11.

Dudhgaonkar

Ranade

Nagar

; et al. Selective IRAK4 Inhibition Attenuates Disease in Murine Lupus Models and Demonstrates Steroid Sparing Activity. J. Immunol. 2016, 198, 1308–1319.

12.

Staschke

K. A.

Dong

Saha

; et al. IRAK4 Kinase Activity Is Required for Th17 Differentiation and Th17-Mediated Disease. J. Immunol. 2009, 183, 568–577.

13.

Picard

Bernuth

H. V.

Ghandil

; et al. Clinical Features and Outcome of Patients with IRAK-4 and MyD88 Deficiency. Medicine (Baltimore) 2010, 89, 403–425.

14.

Bernuth

H. V.

Picard

Puel

; et al. Experimental and Natural Infections in MyD88- and IRAK-4-Deficient Mice and Humans. Eur. J. Immunol. 2012, 42, 3126–3135.

15.

Lim

Altman

M. D.

Baker

; et al. Discovery of 5-Amino-N-(1H-Pyrazol-4-yl)Pyrazolo[1,5-a]Pyrimidine-3-Carboxamide Inhibitors of IRAK4. ACS Med. Chem. Lett. 2015, 6, 683−688

16.

Smyth

L. A.

Collins

Measuring and Interpreting the Selectivity of Protein Kinase Inhibitors. J. Chem. Biol. 2009, 2, 131–151.

17.

Simon

G. M.

Niphakis

M. J.

Cravatt

B. F.

Determining Target Engagement in Living Systems. Nat. Chem. Biol. 2013, 9, 200–205.

18.

Sun

Day

R. N.

Periasamy

Investigating Protein-Protein Interactions in Living Cells Using Fluorescence Lifetime Imaging Microscopy. Nat. Protoc. 2011, 6, 1324–1340.

19.

Robers

M. B.

Dart

M. L.

Woodroofe

C. C.

; et al. Target Engagement and Drug Residence Time Can Be Observed in Living Cells with BRET. Nat. Commun. 2015, 6, 10091–10105.

20.

Molina

D. M.

Nordlund

The Cellular Thermal Shift Assay: A Novel Biophysical Assay for In Situ Drug Target Engagement and Mechanistic Biomarker Studies. Annu. Rev. Pharmacol. Toxicol. 2016, 56, 141–161.

21.

Shaw

Dale

Hemsley

; et al. Positioning High-Throughput CETSA in Early Drug Discovery through Screening against B-Raf and PARP1. SLAS Discov. 2018, 24, 121–132.

22.

Lin

S.-C.

Y.-C.

Helical Assembly in the MyD88–IRAK4–IRAK2 Complex in TLR/IL-1R Signalling. Nature 2010, 465, 885–890.

23.

Wesche

Henzel

W. J.

Shillinglaw

; et al. MyD88: An Adapter That Recruits IRAK to the IL-1 Receptor Complex. Immunity 1997, 7, 837–847.

24.

Ferrao

Zhou

Shan

; et al. IRAK4 Dimerization and Trans-Autophosphorylation Are Induced by Myddosome Assembly. Mol. Cell. 2014, 55, 891–903.

25.

Cushing

Stochaj

Siegel

; et al. Interleukin 1/Toll-Like Receptor-Induced Autophosphorylation Activates Interleukin 1 Receptor-Associated Kinase 4 and Controls Cytokine Induction in a Cell Type-Specific Manner. J. Biol. Chem. 2014, 289, 10865–10875.

26.

Cao

Henzel

W. J.

Gao

IRAK: A Kinase Associated with the Interleukin-1 Receptor. Science 1996, 271, 1128–1131.

27.

Cao

Xiong

Takeuchi

; et al. TRAF6 Is a Signal Transducer for Interleukin-1. Nature 1996, 383, 443–446.

28.

Conze

D. B.

C.-J.

Thomas

J. A.

; et al. Lys63-Linked Polyubiquitination of IRAK-1 Is Required for Interleukin-1 Receptor- and Toll-Like Receptor-Mediated NF-κB Activation. Mol. Cell. Biol. 2008, 28, 3538–3547.

29.

Cui

Xiao

; et al. β-TrCP-Mediated IRAK1 Degradation Releases TAK1-TRAF6 from the Membrane to the Cytosol for TAK1-Dependent NF- B Activation. Mol. Cell. Biol. 2012, 32, 3990–4000.

30.

Yamin

T.-T.

Miller

D. K.

The Interleukin-1 Receptor-Associated Kinase Is Degraded by Proteasomes Following Its Phosphorylation. J. Biol. Chem. 1997, 272, 21540–21547.

31.

Lye

Mirtsos

Suzuki

; et al. The Role of Interleukin 1 Receptor-Associated Kinase-4 (IRAK-4) Kinase Activity in IRAK-4-Mediated Signaling. J. Biol. Chem. 2004, 279, 40653–40658.

32.

Kim

T. W.

Staschke

Bulek

; et al. A Critical Role for IRAK4 Kinase Activity in Toll-Like Receptor–Mediated Innate Immunity. J. Exp. Med. 2007, 204, 1025–1036.

33.

Kawagoe

Sato

Jung

; et al. Essential Role of IRAK-4 Protein and Its Kinase Activity in Toll-Like Receptor–Mediated Immune Responses but Not in TCR Signaling. J. Exp. Med. 2007, 204, 1013–1024.

34.

De Nardo

Balka

Cardona Gloria

; et al. Interleukin-1 Receptor–Associated Kinase 4 (IRAK4) Plays a Dual Role in Myddosome Formation and Toll-Like Receptor Signaling. J. Biol. Chem. 2018, 293, 15195–15207.

35.

Karim

Kiessu

; et al. Mechanism of Dysfunction of Human Variants of the IRAK4 Kinase and a Role for Its Kinase Activity in Interleukin-1 Receptor Signaling. J. Biol. Chem. 2018, 293, 15208–15220.

36.

Cushing

Winkler

Jelinsky

S. A.

; et al. IRAK4 Kinase Activity Controls Toll-Like Receptor–Induced Inflammation through the Transcription Factor IRF5 in Primary Human Monocytes. J. Biol. Chem. 2017, 292, 18689–18698.

37.

Qin

Jiang

Qian

; et al. IRAK4 Kinase Activity Is Redundant for Interleukin-1 (IL-1) Receptor-Associated Kinase Phosphorylation and IL-1 Responsiveness. J. Biol. Chem. 2004, 279, 26748–26753.

38.

Strelow

Fontana

E. J.

; et al. IRAK-4: A Novel Member of the IRAK Family with the Properties of an IRAK-Kinase. Proc. Natl. Acad. Sci. U.S.A. 2002, 99, 5567–5572.

39.

Zhang

J.-H.

Chung

T. D. Y.

Oldenburg

K. R.

A Simple Statistical Parameter for Use in Evaluation and Validation of High Throughput Screening Assays. J. Biomol. Screen. 1999, 4, 67–73.

40.

Iversen

P. W.

Beck

Chen

Y. F.

; et al. HTS Assay Validation. In Assay Guidance Manual. Bethesda, MD: Eli Lilly & Company and the National Center for Advancing Translational Sciences; 2004.

41.

Bland

J. M.

Altman

D. G.

Statistical Methods for Assessing Agreement between Two Methods of Clinical Measurement. Lancet 1986, 1, 307–310.

42.

Eastwood

B. J.

Farmen

M. W.

Iversen

P. W.

; et al. The Minimum Significant Ratio: A Statistical Parameter to Characterize the Reproducibility of Potency Estimates from Concentration-Response Assays and Estimation by Replicate-Experiment Studies. J. Biomol. Screen. 2006, 11, 253–261.

43.

Sun

Whitty

Papadatos

; et al. Adopting a Practical Statistical Approach for Evaluating Assay Agreement in Drug Discovery. J. Biomol. Screen. 2005, 10, 508–516.

44.

Harrison

R. K.

Phase II and Phase III Failures: 2013–2015. Nat. Rev. Drug Discov. 2016, 15, 817–818.

45.

Bunnage

M. E.

Chekler

E. L. P.

Jones

L. H.

Target Validation Using Chemical Probes. Nat. Chem. Biol. 2013, 9, 195–199.

46.

Cook

Brown

Alexander

; et al. Lessons Learned from the Fate of AstraZeneca’s Drug Pipeline: A Five-Dimensional Framework. Nat. Rev. Drug Discov. 2014, 13, 419–431.

47.

Mendeleyev

Kirsten

Hakam

; et al. Potential Chemotherapeutic Activity of 4-Iodo-3-Nitrobenzamide. Biochem. Pharmacol. 1995, 50, 705–714.

48.

Liu

Shi

Maag

D. X.

; et al. Iniparib Nonselectively Modifies Cysteine-Containing Proteins in Tumor Cells and Is Not a Bona Fide PARP Inhibitor. Clin. Cancer Res. 2011, 18, 510–523.

49.

Oneill

L. A.

Fitzgerald

K. A.

Bowie

A. G.

The Toll–IL-1 Receptor Adaptor Family Grows to Five Members. Trends Immunol. 2003, 24, 286–289.

50.

Cendrowski

Mamińska

Miaczynska

Endocytic Regulation of Cytokine Receptor Signaling. Cytokine Growth Factor Rev. 2016, 32, 63–73.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

2.29 MB