Small molecules that inhibit TNF signalling by stabilising an asymmetric form of the trimer

Abstract

Tumour necrosis factor (TNF) is a cytokine belonging to a family of trimeric proteins; it has been shown to be a key mediator in autoimmune diseases such as rheumatoid arthritis and Crohn's disease. While TNF is the target of several successful biologic drugs, attempts to design small molecule therapies directed to this cytokine have not led to approved products. Here we report the discovery of potent small molecule inhibitors of TNF that stabilise an asymmetrical form of the soluble TNF trimer, compromising signalling and inhibiting the functions of TNF in vitro and in vivo. This discovery paves the way for a class of small molecule drugs capable of modulating TNF function by stabilising a naturally sampled, receptor-incompetent conformation of TNF. Furthermore, this approach may prove to be a more general mechanism for inhibiting protein-protein interactions.

Fig. 1. Fragments bind to TNF with slow-binding kinetics.

a SPR depicts the binding of UCB-6786 to immobilised TNF. b The analogue, UCB-6876, specifically binds to immobilised TNF (solid black circles) but not to control proteins, capture antibody plus TNFR1 (solid black squares), capture antibody (solid black triangle pointed down) or TNFR1 (solid black triangle pointed up). c Kinetic analysis of UCB-6876 using the Biacore T100 showing slow association and dissociation rates.

Fig. 2. Crystal structure of human TNF with UCB-6876.

a Top and side views of TNF (green ribbons) with UCB-6876 bound (orange sticks). b Detail showing the electron density of UCB-6876 and MPD bound within the TNF homotrimer. Monomers B (light green) and C (green) are shown surface rendered. Contour level of the electron density is set at 1 sigma. MPD occupies a space next to UCB-6876. Subsequent molecules described in this paper were modified to have chemical groups occupying the space where MPD was bound. c Detail of the compound-binding pocket within the TNF homotrimer, with key residues involved in binding highlighted (sticks and labels). d Side view of apo-TNF (left image) and UCB-6876 bound TNF (right image) revealing the distorted AC receptor-binding site (selected residues involved in TNFR1 binding are highlighted red).

Fig. 3. TNF stabilisation by molecular dynamics and monomer exchange.

a Meta-dynamics molecular dynamics simulation was used to produce this free-energy surface. The simulation used a torsion describing the leaning angle of subunit A with respect to the other subunits and a distance describing how open the AC cleft is. b Mass spectrometry analysis of a mixture of human TNF (orange circles) and mouse TNF (blue circles) showing peaks corresponding to homotrimers and heterotrimers. c An equivalent analysis using human TNF pre-incubated with UCB-6876 (black triangle) showing that the compound blocks the exchange of mouse and human monomers.

Fig. 4. Effect of compounds on TNF:TNFR1 stoichiometry.

a Analytical size exclusion chromatography (AnSEC) of human (h) TNF alone (blue trace), hTNF + UCB-5307 (red trace), hTNF + 3.2-fold excess hTNFR1 (green trace) and hTNF + 3.2-fold excess hTNFR1 + 690 µM UCB-5307 (brown trace). Expected migration position of hTNF with two and three receptors bound are indicated by arrows. Standard trace shown in Supplementary Fig. 5. b AnSEC comparing effect of 690 µM UCB-5307 on preformed hTNF/hTNFR1 complex (3.2-fold excess hTNFR1) (red trace) vs. hTNF preloaded with UCB-5307 followed by addition of 3.2-fold excess hTNFR1 (blue trace) as in experiment (a). Source data are provided as a Source Data file.

Fig. 5. Inhibition of TNF signalling in cells.

a Western blots measuring RIP1 ubiquitination following TNFR1 immunoprecipitation from Jurkat cells treated with TNF ± UCB-9260 or etanercept. pNFkB expression was assessed, along with GAPDH for loading control. An isotype control antibody was included as an immunoprecipitation control (lane 5). b TNF (10 pM) or anti-TNFR1 agonist antibody (0.3 mg/mL)-driven NFkB activity was measured following UCB-9260 treatment in Hek-293 cells using a reporter gene assay system. Representative concentration-response curves are shown. Percentage inhibitions for compound dilutions were calculated between a DMSO control and maximum inhibition (by excess biologic anti-TNF, or NFkB inhibitor–TPCA-1). Untransformed data shown in Supplementary Fig. 10c. c UCB-9260 inhibition of TNF (human or mouse) dependent cytotoxicity was measured in mouse L929 cells. Representative concentration-response curves are shown. Percentage inhibition was calculated between unstimulated wells (maximum signal), and wells with TNF, DMSO and actinomycin D (minimum signal). Untransformed data shown in Supplementary Fig. 10d. Source data are provided as a Source Data file.

Fig. 6. Oral treatment with UCB-9260 inhibits TNF-driven functional effects in vivo.

a UCB-9260 (10–300 mg/kg po, PBS n = 4; vehicle n = 8; UCB-9260 10, 30 and 300 mg/kg n = 6 and 100 mg/kg n = 8; CDP571 n = 5 mice) dose-dependently inhibits human and (b) mouse TNF-induced neutrophil recruitment to the peritoneal compartment in mice (PBS n = 3; vehicle n = 8; UCB-9260 10 mg/kg n = 7, 30 mg/kg n = 6, 100 and 300 mg/kg n = 8; CDP571 n = 5 mice). c UCB-9260 (150 mg/kg po bid) significantly reduces arthritis clinical score in the CAIA model (vehicle n = 8; UCB-9260 n = 8 and AB501 n = 6 mice). Mean data ± standard error is shown. *p < 0.05, **p < 0.01, ***p < 0.001 determined using one-way ANOVA with Dunnett’s multiple comparison test. Exposure of UCB-9260 can be found in Supplementary Tables 4–6. Source data are provided as a Source Data file.