KRAS Inhibitor that Simultaneously Inhibits Nucleotide Exchange Activity and Effector Engagement

Abstract

We describe a small molecule ligand ACA-14 (2-hydroxy-5-{[(2-phenylcyclopropyl) carbonyl] amino} benzoic acid) as an initial lead for the development of direct inhibitors of KRAS, a notoriously difficult anticancer drug target. We show that the compound binds to KRAS near the switch regions with affinities in the low micromolar range and exerts different effects on KRAS interactions with binding partners. Specifically, ACA-14 impedes the interaction of KRAS with its effector Raf and reduces both intrinsic and SOS-mediated nucleotide exchange rates. Likely as a result of these effects, ACA-14 inhibits signal transduction through the MAPK pathway in cells expressing mutant KRAS and inhibits the growth of pancreatic and colon cancer cells harboring mutant KRAS. We thus propose compound ACA-14 as a useful initial lead for the development of broad-acting inhibitors that target multiple KRAS mutants and simultaneously deplete the fraction of GTP-loaded KRAS while abrogating the effector-binding ability of the already GTP-loaded fraction.

Figure 1

(A) Chemical structures of ACA-14 (2-hydroxy-5-{[(2-phenylcyclopropyl) carbonyl] amino} benzoic acid). (B) Structure of GTP-KRAS(G12D)-derived from MD simulations, depicting pocket p1 occupied by ACA-14 (green sticks and sphere). Structural regions near p1 involved in nucleotide binding and undergoing major conformational changes during GDP/GTP exchange are labeled: P-loop (orange, residues 10–17), switch I (purple, residues 25–40), and switch II (yellow, residues 60–74). Selected KRAS residues that are predicted to interact with the ligand are highlighted. (C,D) Detailed interactions of ACA-14 (green sticks) with KRAS residues are shown in surface representation colored by electrostatic potential (C) or in a stick model (D).

Figure 2

ITC profiles of ACA-14 with (A) GDP-KRAS(WT), (B) GDP-KRAS(G12C), (C) GDP-KRAS(Q61H), and (D) GDP-KRAS(G13D) obtained using Malvern’s MicroCal PEAQ-ITC.

Figure 3

Effect of ACA-14 on the intrinsic (A) and SOS-mediated (B) nucleotide exchange rates determined by monitoring the change in fluorescence intensity increase in BGTP upon binding to KRAS. Ligand concentrations: 0 (black circle), 1.56 (red square), 3.12 (green triangle), 6.25 (blue inverted triangle), 12 (purple left-sided triangle), 25 (gray right-sided triangle), and 50 μM (light-orange diamond). Fluorescence intensity (I) was normalized using the formula:I_normalized = (I_raw – I₀)/(I_max – I₀) where I_raw is the fluorescence intensity at a given time, I₀ is the intensity at 60 s, and I_max is the maximum fluorescence (of KRAS alone) at the end of the experiment. Linear (A) or single exponential (B) fits are superimposed as solid lines. Inset: calculated rates (×10^–4 s^–1) as a function of ligand concentrations; EC₅₀ values derived by fitting the curves are shown. KRAS, BGTP, and SOS concentrations were 0.5 μM each, and the buffer (pH 7.5) contained 25 mM Tris Cl, 50 mM NaCl, and 10 mM MgCl₂.

Figure 4

ACA-14 disrupts KRAS-Raf interaction (A–D) and inhibits MAPK signaling (E–H). (A) Effect of ACA-14 on the interaction of varying concentrations of GST-RBD of Raf with 0.5 μM BGTP-γ-S-KRAS in the absence (red; K_d = 6.14 ± 1.86 μM) and presence of 20 μM ACA-14 (1.15 ± 0.09). The dissociation constant (K_d) was derived from the quadratic ligand-binding equation^,, where P1 and P2 are polarization of free and RBD-bound KRAS, respectively, c is the total concentration of KRAS, and x is the total concentration of RBD. (B–D) Representative western blots, including loading controls total GFP-KRAS(G12D) and total endogenous RAS (B), level of GFP-tagged GTP-KRAS(G12D) (C), and endogenous RAS (D) pulled-down by GST-RBD after treatment of BHK cell lysates ectopically expressing GFP-tagged KRAS(G12D) with increasing concentrations of ACA-14. (E) Representative western blots of whole cell lysates from BHK cells expressing GFP-tagged KRAS(G12D) treated for 3 h with vehicle (DMSO) or the indicated concentrations of ACA-14 under serum-starved condition. Vinculin was used as a loading control. (F–H) Quantitation of the levels of phosphorylated c-Raf (p-cRaf) (F), ERK (p-ERK) (G), and AKT (p-AKT) (H). (C,D,F–H) Values are normalized against DMSO control and averaged from three independent experiments. Statistical significance was calculated between vehicle (DMSO) and compound-treated samples using unpaired Student’s t test.

Figure 5

Effect of ACA-14 on the proliferation of pancreatic cells and tumor growth. (A) Effect of ACA-14 on the proliferation of pancreatic (BxPC-3, HPNE, MIAPaCa-2, MOH, and Panc-1) and colon (CaCO-2, SW948, and SW1116) cancer cell lines. (B) Effect of ACA-14 on MIAPaCa-2 subcutaneous tumor growth. ACA-14 was administered intraperitoneally at 5 mg/kg, diluted in corn oil. Controls are an equal volume of diluent (DMSO) added to corn oil alone. Tumor sizes were measured using electronic calipers, and tumor volumes were calculated using the formula volume = (length × width²)/2. Shown are the average tumor volumes for control (n = 8) and ACA-14-treated mice (n = 7). Errors are standard error of the mean (SEM). (C) Volume of tumor on day 20 of the experiment. (D) Weights of DMSO and ACA-14-treated mice. Mice were weighed once per week. Errors are SEM. (E) Representative western blots for p-ERK, total ERK, and GAPDH in DMSO (lanes 1 to 4) and ACA-14 (lanes 5 to 7)-treated tumors. (F) Quantification of p-ERK western blots in DMSO (n = 8) and ACA-14 (n = 7)-treated tumors. Errors are SEM * = p<0.05.