Targeting ERK-MYD88 interaction leads to ERK dysregulation and immunogenic cancer cell death

Abstract

The quest for targeted therapies is critical in the battle against cancer. The RAS/MAP kinase pathway is frequently implicated in neoplasia, with ERK playing a crucial role as the most distal kinase in the RAS signaling cascade. Our previous research demonstrated that the interaction between ERK and MYD88, an adaptor protein in innate immunity, is crucial for RAS-dependent transformation and cancer cell survival. In this study, we examine the biological consequences of disrupting the ERK-MYD88 interaction through the ERK D-recruitment site (DRS), while preserving ERK's kinase activity. Our results indicate that EI-52, a small-molecule benzimidazole targeting ERK-MYD88 interaction induces an HRI-mediated integrated stress response (ISR), resulting in immunogenic apoptosis specific to cancer cells. Additionally, EI-52 exhibits anti-tumor efficacy in patient-derived tumors and induces an anti-tumor T cell response in mice in vivo. These findings suggest that inhibiting the ERK-MYD88 interaction may be a promising therapeutic approach in cancer treatment.

Fig. 1. EI-52 inhibits ERK-MYD88 protein-protein interaction and induces cancer cell death.

A Chemical structure of EI-52. B Binding of EI-52 to ERK1 and ERK2 by fluorescence quenching. Representative of two independent experiments. C ERK1/2-MYD88 interaction inhibition by HTRF. Mean ± SEM from three independent experiments. D EI-52 docking on ERK. Negatively charged residues are in red. Zone A includes the Common Docking (CD) domain (red dotted circle). MKP3 peptide in magenta. Key hydrophobic/aromatic residues in yellow. E MYD88-ERK1 interaction by BiFC and flow cytometry. Mean VENUS Fluorescence Intensity ± SEM at varying EI-52 concentrations (left panel) from three independent experiments, Tukey’s one-way ANOVA; 8 vs 0 h *p = 0.0446, 10- and 12 vs 0 h ****p < 0.0001 (95% confidence interval). F PLA of endogenous MYD88 and ERK in HCT116 cells treated with EI-52 (8 μM) or DMSO for 3 h using antibodies against MYD88 and ERK. The interactions are visualized as red dots and the nuclei are counterstained with DAPI (blue). The number of interactions per cell, quantified by ImageJ, is expressed as the fold difference relative to DMSO treatment. Mean ± SEM from three independent experiments, two-tailed Student’s t-test; *p = 0.0161 (95% confidence interval). G PLA in HeLa cells transfected with 6His-WT or 6His-D319N ERK2 and Flag-MYD88. Cells were treated with EI-52 (10 μM) or DMSO for 8 h using anti-His and -Flag. Detection was carried out as in 1 F. Mean ± SEM from three independent experiments, Tukey’s one-way ANOVA; *p = 0.0303 (95% confidence interval). H Cell recovery and cell death following EI-52 or ERK kinase inhibitors over 48 h at 8 μM. Mean ± SEM from three independent experiments, two-way Tukey’s ANOVA; Cell recovery at 48 h EI-52 vs DMSO *p = 0.0123, LY32149966 vs DMSO p = 0.0145; Death at 48 h EI-52 vs DMSO *p = 0.0456 (95% confidence interval). I Apoptosis (annexin V/PI) in HCT116 cells treated for 24 h with DMSO or EI-52 (8 μM) by flow cytometry. Mean percentage of annexin V⁺ cells ± SEM from three independent experiments, two-tailed Student’s t-test; **F = 0.0016 (95% confidence interval). Source data are provided as a Source Data file.

Fig. 2. EI-52 activity is mechanistically distinctive and tumor-type agnostic.

A Waterfall plot of IC50 values of 301 cell lines of the OncoPanel™ collection. Each bar represents one cell line, colored by relative drug response. B Classification of the response to EI-52 according to the origin of the cancer cell lines. Pie charts showing the distribution of cell lines classified as sensitive (red), relatively insensitive (blue), or intermediate (gray) within each cancer type represented in OncoPanel™. The number of cell lines for each tissue/tumor type is indicated in parentheses. C Dendrogram with EI-52 and 63 anti-tumor reference compounds clustered by drug sensitivity in 301 cell lines. Source data are provided as a Source Data file.

Fig. 3. EI-52 modifies ERK activity and localization, leading to integrated stress response and apoptosis in cancer cells.

A In vitro ERK1 kinase assay with kinase inhibitors K252a or FR180204, or EI-52. Mean ± SEM from three independent experiments. B Cell cycle of HCT116 cells treated with DMSO, U0126 10 μM, paclitaxel 25 nM, or EI-52 8 μM for 24 h was measured by flow cytometry. Mean ± SEM of subG0 cells from three independent experiments, Dunnett’s one-way ANOVA; ****p < 0.0001 (95% confidence interval) (left panel). Percentage of viable cells across cell cycle phases (right panel). C Starved HeLa cells were treated for 8 h with DMSO or EI-52 8 μM, with or without U0126 10 μM. Where indicated, cells were activated for 8 min with EGF 100 ng/ml. ERK targets were analyzed by western blot. Representative data of three independent experiments. D HCT116 cells were transfected with empty vector, 6His-WT, or 6His-D319N ERK2, and treated with DMSO or EI-52 8 μM for 18 h. Representative of three independent DUSP5, ERK, and His western blots. E Phosphorylated-ERK and total ERK immunofluorescence after treatment of serum starved-HeLa cells with EI-52 8 μM for 6 h. Representative of three independent experiments. F Immunofluorescence staining of phosphorylated-ERK. Serum-starved HeLa cells were pretreated for 1 h with DMSO or EI-52 8 μM then activated 10 min with 20% serum. Representative of three independent experiments. G HCT116 cells were treated for the indicated times with DMSO or 8 μM of EI-52, and ISR markers were analyzed by western blot. Representative data of three independent experiments.  H HCT116 cells were pre-incubated or not with ISRIB 400 nM for 1 h then treated with EI-52 8 μM for 16 h. Mean fold change in caspase 3/7 activity ± SEM from three independent experiments, Tukey’s one-way ANOVA; *p < 0.0233 (95% confidence interval). I HCT116 cells were transfected with siHRI, siPERK, siGCN2, siPKR, or siControl, then treated with EI-52 (8 μM) or DMSO. Expression of ATF4, HRI, PERK, CGN2, PKR, and tubulin was determined by western blot. Representative of three independent experiments. Source data are provided as a Source Data file.

Fig. 4. EI-52 induces a stress response and cell death specifically in transformed cells.

A Cell death/confluence ratio was measured using the IncuCyte™ Kinetic Live Cell Imaging System following treatment of transformed (HCT116) or untransformed (HCEC) cell lines with EI-52 or DMSO for 24 h. Results are expressed as mean of ratio ± SEM from three independent experiments. Statistical analysis was performed using Bonferroni’s Two-way ANOVA; HCT116 DMSO vs EI-52 ****p < 0.0001 (95% confidence interval). B Transformed or untransformed cells were treated for the indicated times with DMSO or EI-52 8 μM, and the ISR marker ATF4 was analyzed by western blot. Representative data of three independent experiments are shown. C Caspase 3/7 activity in HCT116 cells or HUVEC (3 donors) was assessed after 16 h treatment with EI-52 8 μM or DMSO. Results are expressed as the mean fold activation of caspase 3/7 in EI-52-treated samples relative to DMSO-treated controls ± SEM. Statistical significance was assessed using a two-tailed one-sample t-test; **p = 0.0037 (95% confidence interval). Source data are provided as a Source Data file.

Fig. 5. EI-52 inhibits tumor growth in mouse tumor models and induces cell death of patient-derived tumoroids and ex-vivo tumors.

A C57BL/6 mice (n = 9/group) injected subcutaneously with Lewis Lung Carcinoma cells were treated daily by intraperitoneal injection with vehicle (PBS, 40% PEG 400, 20% DMSO) or with 25 or 50 mg/kg of EI-52. Tumor volume was measured twice a week with an electronic caliper. B Five-week-old K-ras^LA2 mice (WT n = 5, K-Ras mutant n = 7) were treated intraperitoneally 5 times a week for 10 weeks with 25 mg/kg of EI-52. Tumor load was evaluated on a Perkin Elmer Quantum FX microCT scan. Left panel: example of image reconstitution of lungs from 2 mice treated for 10 weeks with vehicle or with EI-52. Right panel: tumor load in the lungs of vehicle- or EI-52-treated mice at 10 weeks. Statistical analysis was performed using a two-tailed Mann-Withney test; *p = 0.0177 (95% confidence interval). C Patient-derived cancer tumoroids from colon (male, age 77) and lung (female, age 64) were treated for 48 h with DMSO or indicated concentrations of EI-52. Cell death (PI-positive) and apoptosis (caspase 3/7 cleavage) were observed by OPERA imaging. Shown are two of three independent experiments. D Thick (250 μm) sections of surgical pieces from head and neck cancer patients (3 females and 8 males, age 54–82) were cultured in optimized medium in presence of DMSO or EI-52 (8 μM). 24 h later, thin sections were sliced and stained for the apoptotic marker cleaved-PARP (left panel). IHC scoring allows a semi-quantitative analysis of apoptosis (right panel). Mean score ± SEM from 11 samples, statistical analysis was performed using a two-tailed Wilcoxon t-test; *p < 0.0469 (95% confidence interval). Source data are provided as a Source Data file.

Fig. 6. EI-52 induces immunogenic cancer cell death in vitro.

A RNAseq GSEA plot and (B) top upregulated genes in HCT116 cells treated for 18 h with EI-52 (6 μM) or DMSO. C Chemokine mRNA from HCT116 cells treated for 16 h with DMSO or EI-52 8 μM. Mean ± SEM from three independent experiments, two-tailed one-sample t-test. CXCL8 **p = 0.0029; CXCL1 *p = 0.0312; CXCL2 *p = 0.0439 (95% confidence interval). D CXCL8, CXCL1, and CXCL2 in supernatants of HCT116 cells treated with DMSO, MEK inhibitor U0126 10 μM, or EI-52 8 μM for 24 h. Mean ± SEM from four (CXCL8) and three (CXCL1, CXCL2) independent experiments, unpaired two-tailed Student’s t-test; CXCL8 **p = 0.0073; CXCL1 *p = 0.0479; CXCL2 *p = 0.0325 (95% confidence interval) (E) NF-κB, CXCL1, and CXCL8 luciferase activity 16 h post-treatment with EI-52 8 μM or DMSO. Mean of treatment /DMSO ratio ± SEM from four (CXCL8) and three (CXCL1, CXCL2) independent experiments, unpaired two-tailed Student’s t-test; NF-κB *p = 0.0395; CXCL1 *p = 0.0165; CXCL8 *p = 0.0108 (95% confidence interval). F IL8 secretion in supernatant of HCT116 cells transfected for 24 h with control or p65 siRNA, then treated 16 h with DMSO or EI-52 8 μM. Mean ± SEM from three independent experiments, unpaired two-tailed Student’s t-test; ***p = 0.0002 (95% confidence interval) (G) Co-immunoprecipitation of ERK1-HA with p65-Flag after treatment of transfected HEK293T cells with EI-52 8 μM. Representative data of three independent experiments. H Extracellular ATP and HMGB1 in supernatants of HCT116 cells treated for 24 h with DMSO, U0126 10 μM, or EI-52 8 μM. Results are expressed as mean of treatment/DMSO ratio ± SEM from four (e-ATP) or three (HMGB1) independent experiments, two-tailed one sample t-test; e-ATP: U0126 *p = 0.0171; EI-52 **p = 0.0034 (95% confidence interval). HMGB1: EI-52 *p = 0.0157 (95% confidence interval). I Chemotaxis of migrating THP1 cells was evaluated in Boyden chamber containing supernatant of HCT116 cells treated 24 h with control, U0126 10 μM, or EI-52 8 μM. Mean ± SEM from three independent experiments, Dunnett’s one-way ANOVA; **EI-52 p = 0.0013 (95% confidence interval). Source data are provided as a Source Data file.

Fig. 7. ERK-MYD88 interaction inhibition by EI-52 induces cancer cell death and immune T-cell response in vivo.

Following s.c. implantation of syngeneic CT26 cells, BALB/c mice were treated intraperitoneally for 24 h with vehicle or 50 mg/kg of EI-52 and tumors were collected. A Cleaved-PARP was observed by IHC staining, representative of four separate mice or (B) by western Blot of tumor cell lysates recovered from four mice per treatment. C Following s.c. implantation of CT26 cells, syngeneic wild-type and nude BALB/c mice (n = 10/group) were treated daily with vehicle or with 25 mg/kg of EI-52 intraperitoneally. Tumor volume was measured twice a week with an electronic caliper. Results are presented as Mean tumor volume ± SEM, Sidak’s two-way ANOVA, Balb/c day 14 Vehicle/vs EI-52 ****p < 0,0001 (95% confidence interval). D Following s.c. implantation of CT26 cells, mice (n = 10/group) were treated daily with vehicle or with 25 mg/kg of EI-52 intraperitoneally. At days 0, 3, and 5, mice received an intraperitoneal injection of 200 μg isotype control or anti-PD1 antibody. Tumor volume was measured twice a week with an electronic caliper. Results are presented as Mean tumor volume ± SEM, Tukey’s two-way ANOVA, day 14 Vehicle/isotype control vs EI-52/isotype control *p = 0.0115, EI-52/isotype control vs EI-52/anti-PD1 *p = 0.0335 (95% confidence interval). Source data are provided as a Source Data file.

Fig. 8. Illustration.

Biological outcomes of ERK-MYD88 PPI vs ERK enzymatic inhibition.