A Potential Off-Target Effect of the Wnt/β-Catenin Inhibitor KYA1797K: PD-L1 Binding and Checkpoint Inhibition

Abstract

Introduction: The quest for small molecule inhibitors of the PD-1/PD-L1 checkpoint continues in parallel to the extensive development of monoclonal antibodies directed against this immune checkpoint. Drug screening strategies are being set up to identify novel PD-L1 inhibitors.

Methods: A virtual screening based on molecular docking with the PD-L1 protein dimer has been performed to identify a new binder. Binding of the identified ligand to PD-L1 has been validated experimentally using a microscale thermophoresis (MST) assay. The cellular effect of the compound was evidenced using a fluorescence resonance energy transfer (FRET) assay based on activation of tyrosine phosphatase SHP-2.

Results: We have identified the potent Wnt/β-catenin inhibitor KYA1797K as a weak PD-L1 binder. Molecular docking suggested that the compound can bind to the interface of a PD-L1 dimer, with a geometry superimposable to that of the reference PD-L1 inhibitor BMS-202. The atypical 2-thioxo-4-thiazolidinone motif of KYA1797K, derived from the natural product rhodanine, plays a major role in the interaction with PD-L1. Binding of KYA1797K to recombinant hPD-L1 was validated experimentally, using MST. The drug was found to bind modestly but effectively to hPD-L1. The FRET assay confirmed the weak capacity of KYA1797K to interfere with the activation of SHP-2 upon its interaction with human PD-1.

Discussion: Collectively, the data show that KYA1797K could function as a weak modulator of the PD-1/PD-L1 checkpoint. This effect may contribute, at least partially, to the reported capacity of the β-catenin inhibitor to downregulate PD-L1 in cancer cells. The work also underlines the interest to further consider the rhodanine moiety as a chemical motif for the design of new PD-L1 binders.

Keywords: Cancer; Drug design; Immune checkpoint; KYA1797K; PD-L1 inhibitor.

Fig. 1

Structure KYA1797K (C₁₇H₁₁KN₂O₆S₂, compound identity number [CID]: 119057297), which derives from the natural product rhodanine, also shown. KYA1797K is a synthetic product, with the following molecular characteristics: Mw = 404.4 Da; total solvent accessible surface area (SASA) = 649.6 Ų (hydrophobic SASA = 69.9 Ų and hydrophilic SASA = 252.6 Ų); molecular volume 1,134.7 ų; log p (octanol/water) = 3.3 and log S (aqueous solubility) = −5.2. SASAs were calculated with a probe of 1.4 Å radius. Drug properties were calculated with the BOSS 4.9 software.

Fig. 2

Molecular model of KYA1797K bound to the PD-L1 dimer. View (a) shows the drug located at the interface of the protein dimer. The two protein units (in cyan) sandwiches the drug ligand. The closer view (b) shows the compound in the binding pocket, closed by four tyrosine residues (two Y56 and Y123 on each side of the cavity). View (c) shows the KYA1797K molecule sandwiched by the two PD-L1 monomers, using a ribbon model with the α-helices (red) and β-sheets (cyan).

Fig. 3

Detailed views of KYA1797K bound to PD-L1 dimer. The compound inserted into the binding site with (a) the hydrogen bond donor (pink) and acceptor (green) sites, and (b) the solvent accessible surface (SAS, color code indicated). c Binding map contacts for KYA1797K bound to PD-L1 (color code indicated). The calculated potential energy of interaction (ΔE) and free energy of hydration (ΔG) of the KYA1797K/PD-L1 A complex are −91.96 kcal/mol and −29.740 kcal/mol, respectively.

Fig. 4

A superimposed model of KYA1797K and BMS-202 bound to PD-L1 dimer. The rhodanine moiety of KYA1797K occupies exactly the position of the pyridine ring of BMS-202, positioning the phenyl-furan unit in the pocket occupied by the biphenyl moiety of BMS-202. The propionic acid side chain of KYA1797K mimics the terminal acetamido chain of BMS-202.

Fig. 5

Dose-response curve for the interaction between His-PD-L1 and KYA1797K. The concentration of His-PD-L1 protein is kept constant at 10 nM, while the ligand concentration varies from 250 μM to 10 nM. Experiments were performed at 22°C with 30 min incubation with high MST and 20% LED power. Signal-to-noise: 25 (noise is defined as standard deviation of difference between responses of experimental data and fitted data. A signal-to-noise ratio of more than 5 is desirable, while more than 12 reflects an excellent assay).

Fig. 6

Dose response of LNZ on FRET signal. The curve shows the effect of increasing LNZ concentrations on PD-1/SHP-2 interaction measured by the FRET assay, from three independent experiments.