An orally available small molecule that targets soluble TNF to deliver anti-TNF biologic-like efficacy in rheumatoid arthritis

Abstract

Tumor necrosis factor (TNF) is a pleiotropic cytokine belonging to a family of trimeric proteins with both proinflammatory and immunoregulatory functions. TNF is a key mediator in autoimmune diseases and during the last couple of decades several biologic drugs have delivered new therapeutic options for patients suffering from chronic autoimmune diseases such as rheumatoid arthritis and chronic inflammatory bowel disease. Attempts to design small molecule therapies directed to this cytokine have not led to approved products yet. Here we report the discovery and development of a potent small molecule inhibitor of TNF that was recently moved into phase 1 clinical trials. The molecule, SAR441566, stabilizes an asymmetrical form of the soluble TNF trimer, compromises downstream signaling and inhibits the functions of TNF in vitro and in vivo. With SAR441566 being studied in healthy volunteers we hope to deliver a more convenient orally bioavailable and effective treatment option for patients suffering with chronic autoimmune diseases compared to established biologic drugs targeting TNF.

Keywords: TNF; TNF trimer; immunogenicity; rheumatoid arthritis; small molecule.

FIGURE 1

In vitro analysis of SAR441566: (A) shows a representative Biacore T200 curve for SAR441566 binding to TNF immobilized on a CM5 sensor chip. The KD was determined to be 15.1 nM (n = 4) (Table 1). (B) inhibition of zymosan stimulated expression of CD11b on granulocytes in human whole blood by total (free and protein bound) SAR441566 with an IC50 of 35 nM (n = 50) (Supplementary Figures S1A,B show a representative CD11b inhibition curve and corresponding flow cytometry analysis. (A) TNF occupancy in plasma samples from zymosan-stimulated whole blood, was determined from the ratio of compound-bound to total TNF using the conformation specific and total TNF ELISAs (seven individual blood donors assessed in seven independent experiments). Blood from four donors was spiked with SAR441566 concentrations ranging from 1.6–1,000 ng/mL, presented as open circles on the graph. Blood from one donor was spiked with an extended curve including the additional concentrations of 0.66 and 2500 ng/mL (closed circles). Mean occupancy +/− standard deviation plotted for SAR441566 in the range of 1.6–1000 ng/mL for the total of these five donors. Blood from two further donors was spiked with SAR441566 in the range of 2.0–1250 ng/mL (open squares). TNF occupancy was proportional to SAR441566 concentration and plateaued at around 100%, indicating full occupancy.

FIGURE 2

Binding of SAR441566 to human TNF on cells: The binding of SAR441566 to human TNF on the cell surface and intracellularly. An in-house NSO cell line expressing a non-cleavable form of membrane TNF was generated (NSO-tmTNFΔ_1-12 Alexopoulou et al., 1997). An Alexa Fluor 488 labelled conformation specific antibody (CA1974) that only binds TNF when in complex with SAR441566 was used to determine binding on live cells (A) and intracellularly after cells were fixed and permeabilized (B). Images were acquired using a Cellomics Arrayscan (Life Technologies) with a 20× Objective, ORCA-ER camera.

FIGURE 3

In vivo efficacy of SAR441566 in the CIA Model: The dose dependent in vivo efficacy of SAR441566 in the CIA model when dosed therapeutically. SAR441566 was dosed orally BID at 1, 3, 10 and 30 mg/kg. Mouse anti-TNF antibody (Ab501) was dosed at 100 mg/kg. Efficacy assessed by the ability of SAR441566 to reduce the mean area under the curve (AUC) change from baseline (cfb) arthritis score compared to vehicle treated mice. Percent inhibition of each group is indicated on the graph. Data analyzed using a linear model suitable for the randomized design. Red dotted line represents a 50% reduction in the AUC of the disease score. Probability (Pr) statements of scientific interest were made using Bayesian analysis. Graph shows combined data from 3 independent CIA studies. The data demonstrates that the in vivo efficacy of SAR441566 at 10 and 30 mg/kg is comparable to an anti-TNF biologic (Ab501). Vehicle treated group n = 50, SAR441566 1 mg/kg n = 17, 3 mg/kg n = 26, 10 mg/kg n = 31, 30 mg/kg n = 34 and Ab501 treated group n = 16.

FIGURE 4

Effect of SAR441566 on bone in the CIA model as assessed by 3D volumetric micro-computed tomography: (A) shows a representative 3D rendering view of hind paws from animals treated therapeutically with SAR441566. A dose-dependent trend for bone improvement with SAR441566 is observed and at a dose of 10 and 30 mg/kg these effects are comparable to those seen in the Ab501 treated group. (B) 3D volumetric analysis of whole bone microstructure, using the trabecular bone analysis methods and parameters, of the hind paws with a dose-dependent trend for SAR441566 on bone volume fraction (BV/TV), bone surface density (BS/BV), bone mineral density (BMD), and bone fragment thickness (Tb.Th) reaching anti-TNF biologic-like effects at 10 and 30 mg/kg, respectively. Statistical analysis was performed using the one-way ANOVA and Dunnett’s multiple comparisons test, * for p < 0.05; ** for p < 0.005; *** for p < 0.001; and **** for p < 0.0001.

FIGURE 5

QSP model predictions: A computer model and simulations for synovial TNF biology. (A) Reduced model scheme of TNF life cycle and drug binding. (B) data on mTNF-SAR441566 complex appearance in comparison to 3 analogues from the same chemical family on cell surface (circles) and model prediction (lines) using molecule specific binding kinetics and intracellular accumulation factors. (C) Model prediction on free soluble TNF inhibition in synovial fluid for Infliximab (IFX), Etanercept (ETA) and SAR441566.