Discovery of an eIF4A Inhibitor with a Novel Mechanism of Action

Abstract

Increased protein synthesis is a requirement for malignant growth, and as a result, translation has become a pharmaceutical target for cancer. The initiation of cap-dependent translation is enzymatically driven by the eukaryotic initiation factor (eIF)4A, an ATP-powered DEAD-box RNA-helicase that unwinds the messenger RNA secondary structure upstream of the start codon, enabling translation of downstream genes. A screen for inhibitors of eIF4A ATPase activity produced an intriguing hit that, surprisingly, was not ATP-competitive. A medicinal chemistry campaign produced the novel eIF4A inhibitor 28, which decreased BJAB Burkitt lymphoma cell viability. Biochemical and cellular studies, molecular docking, and functional assays uncovered that 28 is an RNA-competitive, ATP-uncompetitive inhibitor that engages a novel pocket in the RNA groove of eIF4A and inhibits unwinding activity by interfering with proper RNA binding and suppressing ATP hydrolysis. Inhibition of eIF4A through this unique mechanism may offer new strategies for targeting this promising intersection point of many oncogenic pathways.

Figure 1.

Discovery and initial biochemical testing of 1. (A) RNA-accelerated malachite green ATPase assay screening data. Each black dot represents one compound. The green line indicates 3 standard deviations above the mean, which was the threshold used for hit determination (18 compounds). Initial hit 1 is circled in red. (B) Compound 1. (C) Dose response of 1 against eIF4A. IC₅₀ = 26.6 ± 2.46 μM. n = 4. (D) Kinetic testing of 1 using the malachite green assay. ATP titration (left) used 250 μg/mL RNA. Data were fit using unbiased mixed-model inhibition in Prism 6.0. The α value for ATP is 0.424, indicating an ATP-uncompetitive inhibitor. The K_iapp for ATP is 10.3 μM. RNA titration (right) used 250 μM ATP. The α value for RNA is 4.53 × 10¹³, indicating an RNA-competitive inhibitor. The K_iapp for RNA is 9.59 μM. Data shown are average of three independent experiments. (E) Left: PDB: 5ZC9. RNA is shown as a red cartoon. AMP-PNP is shown as purple sticks. Center: 1 (cyan) docks to the RNA pocket of eIF4A. Right: zoomed surface view of the binding pocket.

Figure 2.

Mutagenesis studies support the docking pose of 28 binding to the eIF4A RNA groove. (A) Surface view of 28 in the RNA groove of eIF4A (PDB: 5ZC9). (B) 28 docked to eIF4A with close residues labeled and shown as pink sticks. Right: potential interactions between the substituted quinoline of 28 and the pocket are shown as orange dashes. (C) 2D view of right portion of B. Distances indicated are given in Angstrom. (D) Potency of 28 against eIF4A mutants in the malachite green ATPase assay. Amino acid substitution is indicated under each bar. * indicates IC₅₀ > 50 μM.

Figure 3.

Biological activity of 28. (A) Compounds 1 and 28, but not 29, decrease the BJAB cell viability. 1, 28, and 29 were tested using a CellTiter-Glo luminescent cell viability assay in a BJAB Burkitt lymphoma cell line; n = 4. (B) 28 inhibits cap-dependent translation. Top: schematic representation of the pSP/(CAG)₃₃/FF/HCV/Ren·pA₅₁ bicistronic reporter. Bottom: effect of 28 on cap-dependent and HCV IRES-mediated translation in a rabbit reticulocyte lysate programmed with the pSP/(CAG)₃₃/FF/HCV/Ren·pA₅₁ vector. Luciferase luminescence values are normalized to the signal obtained in control experiments with 1% DMSO. The cycloheximide (CHX) dose was 600 μM. The silvestrol dose was 100 nM. * indicates a statistically significant difference between the Renilla and the firefly luciferase signal in a t-test; p < 0.002; n = 3. (C) 28 engages eIF4A in cells. CETSA dose–response stabilization of eIF4A by 28 and silvestrol (100 nM) in A549 cells is seen. Values represent Western blot band intensities of eIF4A measured using densitometry. Intensities are normalized to GAPDH and subtracted from the intensity of the DMSO-treated sample (Figure S4).

Figure 4.

28 is an RNA-competitive, ATP-uncompetitive eIF4A inhibitor. (A) Kinetic testing of 28 using the malachite green assay. Left: ATP titration used 250 μg/mL RNA. Data were fit using unbiased mixed-model inhibition in Prism 6.0. The α value for ATP is 0.514, indicating an ATP-uncompetitive inhibitor. The K_iapp for ATP = 4.79 μM. Right: RNA titration used 250 μM ATP. The α value for RNA is 4.52 × 10, indicating an RNA-competitive inhibitor. The K_iapp for RNA = 4.27 μM. Data shown are the average of three independent experiments. (B) FP assay with MANT-ATP as the fluorophore; PEL = polarized excitation light; n = 5. (C) Left: FP assay with FAM-labeled RNA as the fluorophore. Center: 28 IC₅₀ = 58.3 ± 2.5 μM. 1 IC₅₀ = 161 ± 9.8 μM. The center experiment utilized FAM-A(CAA)₅ RNA. The right experiment utilized two different RNAs; n = 5. (D) TSA with 28 and eIF4A. Error bars are too small to see; n = 3.

Figure 5.

1 and 28 inhibit eIF4A duplex unwinding. Top-right: cartoon of the duplex used in the unwinding assay with Cy3-labeled RNA (orange) and DNA loading strand (black). Left: increasing concentrations of 1 and 28 were added to the unwinding reactions containing the duplex, eIF4A, and ATP. Products of unwinding reactions were resolved on native polyacrylamide gels (Figure S5). The Cy3 signal on gels was quantified by band densitometry. Percent inhibition for each reaction was calculated using (Cy3 duplex signal/total Cy3 signal); n = 3.

Figure 6.

Model of the mechanism of eIF4A inhibition by 28. Under normal conditions (top), eIF4A can bind ATP (purple) and RNA (red), unwind the RNA, hydrolyze the ATP, and release its substrates. Upon introduction of 28 (bottom), eIF4A can bind ATP and weakly bind RNA but in a conformation that hinders its ability to hydrolyze ATP and unwind RNA.

Scheme 1.

Synthesis of Substituted Quinoline Analogues

Scheme 2.

Esterification of Substituted Quinoline Analogues

Scheme 3.

Synthesis of 5-Aminoisatin (24i) and 5-Acetamidoisatin (25i)

Scheme 4.

Synthesis of Central Ring Analogues

Scheme 5.

Synthesis of Pyrrole Analogue Intermediate 32iv

Scheme 6.

Synthesis of Substituted Phenyl Analogues

Scheme 7.

Deprotection of N-Boc Amines