Inhibitors of the Hippo Pathway Kinases STK3/MST2 and STK4/MST1 Have Utility for the Treatment of Acute Myeloid Leukemia

Abstract

Serine/threonine-protein kinases 3 and 4 (STK3 and STK4, respectively) are key components of the Hippo signaling pathway, which regulates cell proliferation and death and provides a potential therapeutic target for acute myeloid leukemia (AML). Herein, we report the structure-based design of a series of pyrrolopyrimidine derivatives as STK3 and STK4 inhibitors. In an initial screen, the compounds exhibited low nanomolar potency against both STK3 and STK4. Crystallization of compound 6 with STK4 revealed two-point hinge binding in the ATP-binding pocket. Further characterization and analysis demonstrated that compound 20 (SBP-3264) specifically inhibited the Hippo signaling pathway in cultured mammalian cells and possessed favorable pharmacokinetic and pharmacodynamic properties in mice. We show that genetic knockdown and pharmacological inhibition of STK3 and STK4 suppress the proliferation of AML cells in vitro. Thus, SBP-3264 is a valuable chemical probe for understanding the roles of STK3 and STK4 in AML and is a promising candidate for further advancement as a potential therapy.

Figure 1.. Biochemical characterization of STK3/4 inhibitors.

(A) Inhibition of STK3 by 6 using the PhosphoSens CSox-based sensor assay (AssayQuant AQT0135), with fluorescence intensity as a readout (excitation/emission 360/485 nm). (B) The initial reaction velocity in (A) was plotted against the log compound concentration, and IC₅₀ values were calculated using log [inhibitor] vs response four-parameter variable slope equation in Prism 8 software. (C) Fluorescent Thermal melt curves for STK3 and STK4 in the presence of 6, 11, and 32 (20 μM) showing protein thermal shifts (DT_m) >8°C compared with DMSO. (D) Binding affinities and thermodynamic signatures from isothermal titration calorimetry (ITC) for compounds 6 and 32. (E) Representative ITC binding data for the interactions between 6 and STK3 and STK4. The normalized heat of binding with the single-site binding fits (red line) is shown, and the raw isotherms of titration heat are displayed as inserts.

Figure 2.. Crystal structure of the kinase domain of STK4 bound to compound 6.

(A) Compound 6 binds to the ATP-binding site of STK4 (2.58 Å, PDB ID: 6YAT) Surface representation (gray) demonstrates occupancy of the interlobe region. (B) Compound 6 forms “two-point” polar contacts with the hinge backbone residues Glu103 and Cys105 of STK4. (C) Schematic representation (LigPlus) of hydrophobic and polar interactions of 6 and STK4. The pyrrole NH forms a H-bond with the carbonyl of Glu103, the N1 maintains an H-bond from Cys105, and a water-bridged contact is maintained to N3. (D) Overall structure of STK4 (residues 18–308) with compound 6 bound in a Type I conformation. The N- and C- lobe domains and the activation loop are indicated.

Figure 3.. In vitro and in vivo target engagement of STK3/4 inhibitors.

(A) Western blot analysis of phosphorylated (p)-MOB1 levels in HEK293 cells treated with DMSO (−) or 10 μM 6, 7, 11, 15, 16, 19, 20, 24, 26, 27 or 32 for 4 h followed by treatment with H₂O₂ (50 μM) for an additional 2 h. Protein levels were quantified by densitometry using ImageJ software. Data are presented as pMOB1 levels relative to the DMSO-treated cells (n = 3). (B) Luciferase reporter assay of HEK293 cells transfected with 8xGTIIC-firefly luciferase plasmid (YAP/TAZ-TEAD-responsive promoter) and control Renilla luciferase plasmid. Cells were treated with DMSO (−) or 10 μM 6, 7, 11, 15, 16, or 32 for 48 h. Data are presented as normalized (firefly:Renilla) luciferase activity relative to the DMSO-treated cells (n = 3). (C and D) Western blot analysis and quantification of p-MOB1, MOB1, and GAPDH levels in liver lysates from adult female C57BL/6 mice injected intraperitoneally with vehicle, compound 11, 16 or 20 (20 mg/kg) for 1 h. Data are expressed as the mean ± SEM pMOB1 levels relative to the vehicle control (n = 3). For A, B, D, ns > 0.033, * p ≤ 0.033, ** p ≤ 0.002, *** p ≤ 0.001 by one-way ANOVA with Fisher’s LSD multiple comparison test.

Figure 4.. Effects of STK3 and STK4 shRNA knockdown in human AML cell lines.

(A) Western blot analysis of STK3, STK4, and GAPDH in MOLM 13 and MV4:11 cells transfected with doxycycline(dox)-inducible control (shScr) or STK3/4-specific shRNAs. Cells were analyzed 24 h after induction. (B) Proliferation assay of MOLM 13 and MV4:11 cells stably expressing dox-inducible control or STK3/4-specific shRNAs. Cell viability was measured on day 0, 2, 4, 6, 8 (RLU) using the CellTiter-Glo assay (Promega). For every group, the average RLU value for day 0 was used to calculate fold change. To determine the relative decrease in cell proliferation for shSTK4, the average fold change for the shScr was subtracted from both groups for each day. *** p ≤ 0.001 by two-way ANOVA with Fisher’s LSD multiple comparison test. (n = 3)

Figure 5.. Synergy between venetoclax and STK3/4 inhibitors in AML cell line MOLM 13.

Synergy graph of MOLM 13 cell viability (CTG) after treatment for 72 h with venetoclax and (A) compound 20 or (B) compound 16. Plot was generated with SynergyFinder web application. (C) CellTiter-Glo assay of MOLM 13 cell viability after treatment for 72 h with venetoclax alone or with 10 μM compound 20 or 16. Data were normalized to cells treated with DMSO ± compound 20 or 16. Results were plotted against the log compound concentration, and IC₅₀ values were calculated from the log [inhibitor] vs response curve using the four-parameter variable slope equation in Prism 8 software. CTG cell viability assay, (mean ± SEM of n=3)