Designing a Small Fluorescent Inhibitor to Investigate Soluble Epoxide Hydrolase Engagement in Living Cells

Abstract

Soluble epoxide hydrolase (sEH) is a promising target for a number of inflammation-related diseases. In addition, inhibition of sEH has been shown to reduce neuroinflammation, which plays a critical role in the development of central nervous system (CNS) diseases such as Alzheimer’s disease. In this study, we present the rational design of a small fluorescent sEH inhibitor. Starting from the clinical candidate GSK2256294A, we replaced the triazine moiety with the 4-chloro-7-nitrobenzo-2-oxa-1,3-diazole (NBD-Cl) fluorophore. The resulting fluorescent sEH inhibitor displayed excellent potency in an in vitro enzyme activity assay (IC₅₀ < 2 nM). The developed inhibitor is applicable in a NanoBRET-based assay system suitable for studying sEH target engagement in living cells. Furthermore, the inhibitor can be used to visualize sEH in sEH-transfected HEK293 cells and in primary mouse astrocytes by fluorescence microscopy.

Soluble epoxide hydrolase (sEH) is a bifunctional enzyme that is widely expressed in many peripheral tissues as well as in the central nervous system.¹ The sEH forms homodimers and is mainly located in the cell cytoplasm.² The C-terminal domain is an epoxide hydrolase (sEH-H) while the N-terminal domain exhibits lipid phosphatase activity. Over the past decades, the hydrolase activity has been intensively investigated, resulting in the development of various potent sEH-H inhibitors.³⁻⁶ In contrast, the function of the phosphatase domain is still under investigation.⁷ The main function of the sEH-H is the conversion of the epoxide moiety of polyunsaturated epoxy-fatty acids (EpFAs) to the corresponding diols.⁸ Due to its broad substrate scope and tissue expression profile, the sEH is involved in various pathophysiological processes such as diabetes, ischemic injury and nociception.⁹ In recent years, the role of sEH in the pathophysiology of various central nervous system (CNS) diseases has become a novel research focus.¹⁰ One interesting discovery is the contribution of sEH in neuroinflammation, which plays a key role in the onset and pathogenesis of Alzheimer′s disease and Parkinson′s disease.¹¹ In 2020, Griñán-Ferré et al.¹² as well as Ghosh et al.¹³ demonstrated that inhibition of sEH-H reduces neuroinflammation in mouse models of age-related cognitive decline and Alzheimer′s disease, suggesting sEH as a promising target in CNS disorders. To gain deeper insights into the signaling pathways triggered by sEH and to study its pathophysiological role, a fluorescent inhibitor would be a useful tool to study target engagement in vitro and in vivo. Clinical trials often fail due to the lack of efficacy of the clinical candidate. One of the reasons for this could be the use of assays that rely solely on recombinantly expressed proteins to detect compound activity.¹⁴ Evaluating the efficacy in living cells and correlating biochemical activity with target engagement (TE) should be a key step in drug development or target validation. Durham and Blanco showed in their review the importance of TE evaluation in early drug discovery and highlighted several cases where TE has been employed.¹⁵

Therefore, methods such as bioluminescence resonance energy transfer (BRET), fluorescence anisotropy, and cellular thermal shift assays were developed in order to study drug–target engagement. The readout of fluorescence polarization (FP) and BRET relies on binding of a fluorescent tracer. Consequently, there is a need for fluorescent ligands that target proteins of interest (POI) to examine the engagement in living cells. To develop a fluorescent sEH ligand, we used the scaffold of clinical candidate 1(17,18) developed by GlaxoSmithKline (GSK) and replaced the triazine moiety with the fluorophore NBD (7-nitrobenzo-2-oxa-1,3-diazole) to generate compound 2. Computational modeling of the binding mode of 2 suggested the whole compound fitting into the L-shaped binding pocket of the sEH-H (Figure 1). The synthesis was accomplished in three steps (Scheme 1). First, the benzylic amine 4 was coupled with the unsaturated carboxylic acid 3 using the coupling reagents PyBOP and HOBt. Next, the Boc-protected amine 5 was deprotected with trifluoroacetic acid (TFA). Finally, aromatic substitution reaction with NBD-Cl 7 yielded compound 2.

Figure 1.

Left: Predicted binding mode of 2 in the binding pocket of human sEH-H (PDB: 4JNC). Right: Design of the fluorescent ligand 2 derived from the clinical candidate GSK2256294 (1).¹⁶

Scheme 1. Synthetic Route for the Fluorescent Inhibitor 2.

Reagents and reaction conditions: (a) PyBOP, HOBt·H₂O, DIPEA, THF, rt, 16 h, 85%. (b) TFA, DCM, r.t., 16 h 90%. (c) ACN, reflux, 4 h, 66%.

The spectroscopic properties and in vitro inhibitory activity of 2 were evaluated. Figure 2A shows the normalized absorption and emission spectra of 2 with multiple absorption maxima in black and a single emission peak in green upon excitation at 450 nm. The inhibitory potential of 2 was examined in a fluorescence-based enzyme activity assay with recombinant mouse and human sEH (mouse sEH-H IC₅₀ = 7.71 ± 1.17 nM, human sEH-H IC₅₀ = 1.43 ± 0.04 nM). The compound demonstrated a dose dependent inhibition of both species’ isoforms (Figure 2D) without inhibiting the phosphatase activity of sEH (data not shown). To confirm direct binding of 2 to sEH-H an orthogonal fluorescence polarization (FP) assay was established. In an FP assay, a fluorophore, excited by polarized light, emits light with a degree of polarization which is inversely related to its rate of molecular rotation.¹⁹ Due to this circumstance, FP can be used to distinguish between the situations of bound vs unbound fluorophore. First, we optimized the assay window of the FP assay by varying the human sEH-H concentrations, yielding a saturation curve at two different concentrations of 2 (6 nM and 20 nM) with an incubation time of 1 h (Figure 3A). The concentrations of 20 nM for 2 and 30 nM for sEH-H were chosen for all further experiments due to a better signal-to-noise ratio. Second, the three known inhibitors 1, 8, and 9 were chosen to investigate the displacement of ligand 2. To determine a suitable incubation time, we performed overnight measurements with one measurement per hour. The best results were obtained with an incubation time of 5 h (data just shown for 5 h in Figure 3B). The standard deviations of the triplicates are relatively high, which might be caused by the tight and potent binding of 2 combined with the weak brightness of NBD and its tendency to photobleach.²⁰

Figure 2.

Evaluation of the spectral properties of compound 2. (A) Normalized absorption (black) and emission (green) spectra of compound 2 in DPBS buffer. (B) Fluorescence emission of 2 in different solvents. (C) Fluorescent signal of 2 at different concentrations of human sEH-H. (D) inhibition curve of compound 2 in a fluorescence-based assay with murine sEH (black) and human sEH (blue).

Figure 3.

Binding studies with different assay systems. (A and B) Graphical results of the FP assay: (A) saturation binding curve of recombinant sEH-H with 2 (20 nM) and (B) dose-dependent inhibition curves of known inhibitors detected with FP assay. Middle: Structure of the tested inhibitors 8, 9 and 10. (C) NanoBRET saturation binding curves in HEK293T cells expressing NanoLuc-sEH with increasing doses of compound 2 and either 0 μM or 25 μM sEH inhibitor 1. (D) Dose-dependent inhibition curves of multiple known sEH inhibitors in HEK293T cells expressing NanoLuc-sEH.

Due to the observation of high standard deviations, we further analyzed the spectroscopic behavior of compound 2. To investigate the influence of the protein environment, we performed a fluorescence scan of 2 at different concentrations of human sEH-H (Figure 2C). The measurements show that the higher the sEH-H concentration, the less fluorescent the signal that can be detected. These findings are in contrast to published results of Hansen et al. where the interaction of NBD with the environment of the protein led to an increase in fluorescence.²¹ As known from literature, the brightness of a fluorophore depends on the solvent used.²² Hansen et al. also showed that the brightness of NBD in n-octanol is higher than that in PBS buffer. Therefore, we measured the fluorescence of 2 in four different solvents (Figure 2B) and the results are in line with reports in the literature.

In contrast to n-octanol, the signal of 2 in toluene is even weaker than in aqueous solutions. This observation suggests that aromatic environment of toluene decreases the fluorescence of 2 strongly. We closely examined the predicted binding mode of compound 2 in complex with sEH-H. Two aromatic amino acids Trp335 and Phe379 are engaged in interactions with the predicted bound conformation of 2 (Figure S1). In particular, aromatic π–π stacking interactions of the NBD moiety to Phe379 might explain the decrease in fluorescence with increasing protein concentrations. Although the intensity of the fluorescent signal of NBD is moderate in an aqueous environment, 2 can be used in a FP assay to evaluate sEH-H inhibitors in a displacement mode.

Apart from the impressive work of Griñán-Ferré et al.¹² who developed a cellular thermal shift assay to monitor sEH engagement in vivo, it has not been possible to show sEH engagement in living cells. Using compound 2, we established a cellular assay to detect spatial interactions between compound 2 and sEH by NanoBRET technology.²³ NanoBRET relies on a bioluminescent donor transferring its energy to a fluorescent acceptor molecule. A POI is fused to luciferase enzyme NanoLuc which cleaves the substrate furimazine, thereby generating a luminescence signal. In a nonradiative transfer of energy from the luminescent donor to an acceptor, the acceptor is excited. This can only occur in close proximity (∼10 nm distance), which indicates an interaction of the fluorophore with the POI.²⁴ To establish NanoBRET in a cellular setting, HEK293T cells were transiently transfected with a sEH-NanoLuc-construct using Lipofectamine 3000. At 24 h post-transfection, the cells were incubated for 5 h with different concentrations of tracer 2 with either 0 or 25 μM of inhibitor 1 to proof specific binding of 2 to sEH. After addition of the NanoGlo substrate, the NanoBRET signal was calculated from the simultaneously detected emission of the donor and the acceptor. The signal of total, specific, and unspecific binding with varied concentrations of 2 is shown in the saturation binding curve (Figure 3C). This experiment allowed us to calculate an apparent K_d of 6 nM (95% CI 3–11 nM) for 2 in the NanoBRET setting. Adding an excess of 1 resulted in complete depletion of the NanoBRET signal, indicating specific binding of 2 to sEH. To demonstrate the applicability of this assay we evaluated the TE of previously published sEH inhibitors with different reported affinities. We investigated compound 1 as well as sEH inhibitors TPPU²⁵ (8) and AUDA²⁶ (10) and the recently published dual modulator of sEH and peroxisome proliferator-activated receptor γ²⁷9. Transfected HEK293T cells were exposed to increasing concentrations of the competing inhibitors while the concentration of compound 2 remained constant at 60 nM (Figure 3D). Using the Cheng–Prusoff equation, we could determine apparent K_i-values in cells (K_iapp).²⁸ Compounds 1 and 10 display the highest affinity to sEH, 8 binds tightly to sEH, while compound 9 shows moderate binding affinity (Table 1). The results correlate with the IC₅₀ values from the cell-free fluorescence-based assay, although it should be noted that changing the system from a cell-free to a cell-based assay may result in lower affinity values.

Table 1. Generated Inhibition Values from Different Assay Systems.

	1	8	9	10
Kiapp [nM] (95% CI)a	24 (21–28)	104 (65–156)	583 (347–1764)	16 (14–22)
IC50[nM] ± SDb	0.9 ± 0.2	12.8 ± 0.4	29.4 ± 2.8	2.6 ± 0.2

^aNanoBRET assay.

^bFluorescence-based enzyme activity assay.

The NanoBRET assay is suitable to detect and to quantify cellular TE of sEH inhibitors; however, there is a discrepancy in the generated K_iapp values and reported values of 10 and 1. GSK developed a cell-based activity assay that determines the concentration of the sEH metabolite 14,15-dihydroxyeicosatrienoic acid (14,15-DHET) from cell media. The cell media is separately evaluated with fluorophore-labeled 14,15-DHET and rabbit anti-14,15-DHET antibody in an FP assay. 10 exhibits an IC₅₀ of 34 nM, and compound 1 has a potency of 0.66 nM.^29,30 Our NanoBRET assay suggests that 1 and 10 are equally potent inhibitors in cells. The discrepancy might be caused by the different cell permeability or cellular stability of 1 and 10 or by the resolution limit of the NanoBRET assay.

Next, fluorescence microscopy was used to visualize target engagement of compound 2 in living cells. HEK293 cells were stably transfected with full-length human sEH using the Sleeping Beauty system.³¹ The open reading frame for sEH is under the control of the doxycycline inducible TCE promotor, allowing for a doxycycline-inducible sEH overexpression. After 24 h of doxycycline treatment (200 ng/μL), the cells were incubated with 300 nM of compound 2 for 5 h. The co-occurrence of sEH and 2 is shown by fluorescence microscopy (Figure 4). The nuclei were stained with Hoechst 33342 (blue), the sEH was detected by immunofluorescence staining (red), and tracer 2 was detected by its fluorescence (green). Upon induction of sEH expression, we could detect sEH. Incubation with doxycycline and 2 resulted in a signal pattern of 2 consistent with anti-sEH antibody labeling (specificity of the antibody see Figure S2). In contrast, no compound 2 signal was detected in cells where sEH expression has not been induced. Taken together, this data demonstrates the specificity of 2 for sEH in living cells and its applicability for sEH visualization by fluorescence microscopy. We then investigated the applicability of 2 to visualize the sEH in primary cells. Therefore, cultured primary mouse astrocytes were treated with compound 2 (10 μM) for 5 h and were subsequently used for immunocytochemistry (Figure 5). sEH specific antibody labeling (red) shows the abundance of the sEH, and nuclei were labeled with DAPI (blue). Incubation with 2 led to a fluorescence signal (green) that overlapped with the anti-sEH antibody labeling, indicating a spatial co-occurrence of the inhibitor with its target in primary cells.

Figure 4.

Co-occurrence of compound 2 and sEH in HEK293 cells after doxycycline treatment. HEK293 cells were stably transfected using the Sleeping Beauty system with plasmid sEH_aa1-aa555_pSBtet-bla, and sEH overexpression was induced by 24 h doxycycline (200 ng/mL) treatment followed by incubation of compound 2 (300 nM) for 5 h. As controls, cells were either left untreated, not exposed to doxycycline and treated with compound 2, or subjected to doxycycline but not treated with compound 2. Intracellular sEH was stained via immunocytochemistry (Alexa633, λex: 590–650 nm, λem: 662–738 nm). Cell nuclei were labeled with Hoechst 33342 (λex: 325–375 nm, λem: 435–485 nm). Cell nuclei, sEH, and 2 (λex: 460–500 nm, λem: 512–542 nm) were visualized by fluorescence microscopy at 63× magnification. Representative experiment results are shown. Scale bar, 50 μm. n = 3.

Figure 5.

Co-occurrence of compound 2 and sEH in primary mouse astrocytes. Cells were either exposed to compound 2 (10 μM) or vehicle for 5 h. Intracellular sEH was labeled via immunocytochemistry (Alexa Fluor 555 λex: 555 nm, λem: 570 nm). Nuclei were labeled using DAPI (λex: 350 nm, λem: 457 nm), compound 2 was excited at 488 nm, and the fluorescence was recorded at 520 nm. Representative data is shown. Scale bar, 100 μm. n = 3.

In the present study, we have demonstrated the design, synthesis, and biological evaluation of a fluorescent probe 2 for the sEH. Compound 2 was applicable in different assay formats such as FP assay, NanoBRET, and fluorescence microscopy. Furthermore, we demonstrated specific binding of 2 to the sEH in vitro using both cell-free as well as cellular systems. The novel fluorescent sEH inhibitor 2 opens up new possibilities to both visualize and manipulate sEH in vitro and potentially in vivo using preclinical models of neurodegenerative diseases. Furthermore, the basic scaffold used in this study can be conjugated to different fluorophores, widening the range of its applicability and offering fluorescent probes for sEHs that are tailored to the individual experimental needs.

Glossary

Abbreviations

AD

Alzheimer′s disease

BRET

bioluminescence resonance energy transfer

CI

confidence interval

CNS

central nervous system

14,15-DHET

14,15-dihydroxyeicosatrienoic acid

EpFAs

epoxy-fatty acids

FP

fluorescence polarization

HOBt

hydroxybenzotriazole

GSK

GlaxoSmithKline

NBD-Cl

4-chloro-7-nitrobenzo-2-oxa-1,3-diazole

POI

protein of interest

PyBOP

benzotriazol-1-yloxytripyrrolidinophosphonium hexafluorophosphate

sEH

soluble epoxide hydrolase

TE

target engagement

TFA

trifluoroacetic acid

Supporting Information Available

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acsmedchemlett.2c00073.

• Biological and chemical procedures, assay procedures, NMR spectra of the compounds, HPLC data of the tracer, and imaging procedures (PDF)

Author Contributions

^∥ S.B. and J.S. contributed equally. The manuscript was written through contributions of all authors. All authors have given approval to the final version of the manuscript.

M.D. is supported by Alzheimer Forschung Initiative project #20041 and Buchmann Fellowship. This research was supported by Deutsche Forschungsgemeinschaft (DFG, Benjamin-Walter Stelle HI 2351/1-1 to K.H., Heisenberg-Professur PR 1405/7-1, Sachbeihilfe PR 1405/8-1, and SFB 1039 TP A07 to E.P., and HE 6867/3-1 to P.B. and J.H.).

The authors declare no competing financial interest.