MAST1 Drives Cisplatin Resistance in Human Cancers by Rewiring cRaf-Independent MEK Activation

Abstract

Platinum-based chemotherapeutics represent a mainstay of cancer therapy, but resistance limits their curative potential. Through a kinome RNAi screen, we identified microtubule-associated serine/threonine kinase 1 (MAST1) as a main driver of cisplatin resistance in human cancers. Mechanistically, cisplatin but no other DNA-damaging agents inhibit the MAPK pathway by dissociating cRaf from MEK1, while MAST1 replaces cRaf to reactivate the MAPK pathway in a cRaf-independent manner. We show clinical evidence that expression of MAST1, both initial and cisplatin-induced, contributes to platinum resistance and worse clinical outcome. Targeting MAST1 with lestaurtinib, a recently identified MAST1 inhibitor, restores cisplatin sensitivity, leading to the synergistic attenuation of cancer cell proliferation and tumor growth in human cancer cells and patient-derived xenograft models.

Keywords: MAPK signaling; cisplatin resistance; dual-kinase inhibitor; lestaurtinib; microtubule-associated serine/threonine kinase 1; platinum-based cancer therapy.

Figure 1.. Targeting MAST1 sensitizes cisplatin treatment in vitro and in vivo.

(A) Primary screen testing 781 genes was carried out using sublethal dose (5 μg/ml) of cisplatin (left). Candidates with low virus infection efficiency (<25%; grey), and shRNAs that alone induced cell death (>15%; blue) were excluded (right). (B) Secondary screen used the top 50 candidates from the primary screen in 4 cisplatin-resistant cancer cell lines (KB-3-1^cisR, A549^cisR, A2780^cisR and PCI-15A^cisR) (left). 30 leads showing >10% cell death upon shRNA and cisplatin treatment are shown (right). (C-D) Cell viability and cisplatin sensitivity of KB-3-1^cisR and A549^cisR cells (C) and platinum-refractory (cis^R) TKO-002 cells (D) with MAST1 knockdown. Cells were transduced with 3 different MAST1 shRNA clones followed by sublethal dose of cisplatin (5 μg/ml for KB-3-1^cisR and 2 μg/ml for A549^cisR) or vehicle treatment. Cisplatin IC₅₀ was assessed after 72 hr. IP: immunoprecipitation. (E) Colony formation assays were performed using cancer cells with MAST1 knockdown and cisplatin treatment. (F-H) Effect of cisplatin treatment and MAST1 knockdown using two shRNA clones on tumor growth of KB-3-1^cisR xenograft mice. Mice were treated with vehicle-control or cisplatin 3 days after xenograft and tumor size was monitored (F). Tumor weight (G) and MAST1 expression in tumor lysates are shown (H; top). Representative images of Ki-67 IHC staining in harvested tumors from each group are shown (H; bottom). Scale bars represent 10 mm for (F) and 50 μm for (H). Data are mean ± SD from three technical replicates of each sample for (C-E). Error bars represent SEM for (F) and SD for (G). p values were determined by two-tailed Student’s t test (ns: not significant; **: p < 0.01). See also Figure S1.

Figure 2.. MAST1 directly phosphorylates MEK1 and activates anti-apoptotic signaling upon cisplatin treatment.

(A) MAST1 in vitro kinase assay using MAST1 wild type (WT) or kinase-dead mutant (DA; D497A). GST-MAST1 variants were enriched from 293T and kinase activity was assessed by ADP-Glo Kinase assay using myelin basic protein (MBP) as a substrate. (B) Cell viability of parental cancer cells with WT or DA MAST1 overexpression in the presence of cisplatin. (C) Cell viability of cisplatin-resistant (cis^R) cells with endogenous MAST1 knockdown and expression of shRNA-resistant WT or DA MAST1. Relative viability was obtained by normalizing values to cisplatin untreated samples for (B) and (C). (D) Phospho-antibody array results using 1,318 antibodies in KB-3-1^cisR lysates. Top seven protein factors whose phosphorylation states decreased in MAST1 knockdown and cisplatin treatment are shown. (E) Parental and cis^R pairs of KB-3-1 and A549 cells with MAST1 knockdown and cisplatin treatment (5 μg/ml) were assayed for MEK1 phosphorylation by immunoblotting. (F) MAST1 in vitro kinase assay using recombinant inactive MEK1 (rMEK1 K79A) as a substrate. (G) Kinase activities of MEK1, cRaf and MAST1 in cells with MAST1 knockdown and cisplatin treatment. MEK1, cRaf and MAST1 were immunoprecipitated from KB-3-1 cells and kinase activities were determined by ADP-Glo kinase assay using recombinant inactive ERK2 or MEK1 as substrates. MEK inhibitor U0126 (10 μM) and Raf inhibitor L779450 (5 μM) were used as controls. (H) Western blot analysis of apoptosis-related factors. Cells with or without MAST1 shRNA were treated with cisplatin (5 μg/ml) for 24 hr. (I) Apoptosis assay using parental and cis^R cells with or without MAST1 knockdown. Cells were treated with sublethal dose of cisplatin (2 μg/ml: parental, 5 μg/ml: cis^R) for 48 hr and apoptotic cells were assayed by Annexin V staining. Error bars represent ± SD from three technical replicates. p values were determined by two-tailed Student’s t test (ns: not significant; *: 0.01 < p < 0.05; **: p < 0.01). See also Figure S2.

Figure 3.. Cisplatin but no other DNA damaging agents dissociates cRaf from MEK1 and reactivates MEK1 through MAST1.

(A) Interaction between cRaf and MEK1 upon cisplatin treatment in diverse cancer types. Cells were treated with 5 μg/ml cisplatin for 24 hr prior to MEK1 immunoprecipitation. (B) Purified MEK1-cRaf complex from KB-3-1 cells were incubated with increasing concentrations of cisplatin (0-5 μg/ml) for 2 hr and applied to Western blotting. (C) Biacore SPR (left) and thermal shift assay (right) were performed using purified recombinant MEK1 or cRaf with increasing concentrations of cisplatin. (D) MAST1 interacts with cRaf and MEK1 in cancer cells. ov: overexpressed. (E) Dissociation constant (K_d) values for MEK1-cRaf or MEK1-MAST1 interaction in the presence and absence of cisplatin were determined by Biacore SPR analysis. (F) Interactions between MEK1 variants and cisplatin or cRaf in the absence and presence of cisplatin were determined by SPR and shown as K_d values. N.D. (not determined). (G) Binding affinity of WT and C142A MEK1 to cRaf in the presence of cisplatin. MEK1 shRNA-resistant flag-MEK1 WT and C142A were expressed in MEK1 knockdown KB-3-1 cells. Cells were treated with cisplatin (5 μg/ml) for the indicated time and MEK1-cRaf interaction was determined by flag pull down. (H) Effect of cRaf (top) or MAST1 (bottom) downregulation on MEK1 activation in the presence and absence of cisplatin. cRaf and MAST1 knockdown cells were treated with cisplatin for 48 hr and MEK1 activity was determined by p-MEK S217/S221 Western blotting. (I) Effect of MAST1 rescue expression on MEK1 activation in MAST1 knockdown cells with cisplatin treatment. (J) Effect of WT or DA MAST1 expression on MEK1 activation in cRaf knockdown cells in the presence or absence of cisplatin. (K) Purified MEK1-cRaf complex from KB-3-1 cells were incubated with 5 μg/ml cisplatin. Samples were collected at different time points and subjected to immunoblotting. Density analysis of relative amount of cRaf bound to MEK1 from three biological replicates are shown. (L) Accumulation of cisplatin-induced DNA damage and repair in KB-3-1 cells with or without MAST1 shRNA was determined by flow cytometry analysis of phospho-γH2AX (upper) and phospho-53BP1 (lower). For DNA repair, cells were treated with cisplatin (5 μg/ml) for 2 hr before cisplatin washed off and cells were incubated in fresh medium. (M) cRaf-MEK1 dissociation induced by DNA damaging agents and cytotoxic drugs. KB-3-1 cells were treated with different compounds as indicated, followed by MEK1 immunoprecipitation and immunoblotting. (N) Cells with or without MAST1 knockdown were treated with different concentrations of DNA damaging agents and cytotoxic drugs and IC₅₀ values were obtained. Error bars represent ± SD from three biological (K) or technical (L) replicates. p values were determined by two-tailed Student’s t test (ns: not significant). See also Figure S3.

Figure 4.. MAST1 induces cisplatin-resistant cancercell proliferation and tumor growth through MEK1 phosphorylation.

(A) MAST1 and flag-tagged S221A and S221D MEK1 expression were detected by immunoblotting in KB-3-1^cisR, A549^cisR and KB-3-1 cells. (B-C) Cell viability (B) and apoptosis (C) in MAST1 knockdown cells with the expression of the indicated MEK1 variants in the presence and absence of sublethal dose of cisplatin (2 μg/ml: parental, 5 μg/ml: cis^R). (D) Tumor volume (left), tumor weight (middle), and Ki-67 expression (right) of KB-3-1^cisR cell xenograft mice. Representative dissected tumors are shown on right top panel. KB-3-1^cisR cells with MAST1 knockdown and MEK1 S221A or S221D expression were injected and cisplatin was administered by intraperitoneal injection when there were palpable tumors. Tumor volume and tumor weight were normalized to the corresponding cisplatin untreated group. Scale bars represent 5 mm for dissected tumors and 20 μm for Ki-67 IHC staining images. Data are mean ± SD from three technical replicates for (B) and (C). Error bars represent SEM for tumor growth and SD for tumor weight in panel (D). p values were determined by two-tailed Student’s t test (ns: not significant; *: 0.01 < p < 0.05; **: p < 0.01).

Figure 5.. MAST1 expression correlates with cisplatin resistance in cancer cell lines and patient-derived tumors.

(A) MAST1 expression in HNSCC, lung, and ovarian cancer cell lines. (B) Correlation between MAST1 expression and cisplatin IC₅₀ in cancer cell lines shown in (A). (C) MAST1 expression in PDX tumors. (D) Correlation between MAST1 expression and cisplatin IC₅₀ in PDX tumors shown in (C). r = Pearson’s correlation coefficient. (E) Effect of MAST1 knockdown or WT or DA MAST1 overexpression on cisplatin response in diverse cancer cell lines. Cisplatin response was determined using cisplatin IC₅₀. Error bars represent SD from three technical replicates. p values were determined by two-tailed Student’s t test (ns: not significant; *: 0.01 < p < 0.05; **: p < 0.01). See also Figure S4.

Figure 6.. MAST1 and MEK activation is associated with cisplatin resistance and poor clinical outcome in human HNSCC.

(A) Schematic summary of HNSCC patients (n=97 cases) and samples (n=116 tumors) analyzed by IHC. Open circle: pre-therapy tumors; closed circle: post-therapy tumors. Tumor numbers of before or after CT or non-CT groups are indicated on the left. Case numbers for sensitive or resistant CT or non-CT treated patients are indicated at the bottom. CT: platinum-based chemotherapy. (B) MAST1 expression and MEK phosphorylation levels and drug response in specimens before (top) and after (bottom) platinum treatment. (C) Correlation between MAST1 and phospho-MEK in samples before (top) and after (bottom) platinum treatment. (D-E) Comparison of MAST1 status between pre- and post-therapy in paired (D) and non-paired (E) samples. Top: Platinum therapy; Bottom: non-platinum therapy. (F-G) Kaplan-Meier survival analysis of platinum-treated (F) and non-platinum-treated (G) patient groups. Patients were dichotomized by MAST1 expression level at median. Error bars represent ± SD for (B) and (E). p values were determined by Pearson correlation for (C), paired Student’s t test for (D), low-rank test for (F-G), and unpaired Student’s t test for the rest (ns: not significant; *: 0.01 < p < 0.05; **: p < 0.01). See also Figure S5.

Figure 7.. Validation of lestaurtinib as a MAST1 inhibitor.

(A) In vitro MAST1 kinas eassay using ten compounds potentially bind to MAST1. Compounds (10 μM) were incubated with purified GST-MAST1 and applied to the in vitro kinase assay. (B) Effect of top six MAST1 inhibitors from (A) on cisplatin sensitivity in KB-3-1 cells. A dose of inhibitor that did not affect viability when given alone (100 nM of lestaurtinib, dovitinib, staurosporine; 1 μM of sunitinib, SU14813, bosutinib) was used with increasing concentrations of cisplatin for 48 hr. (C) Correlation between MAST1 activity inhibition and cisplatin sensitivity of the top six MAST1 inhibitors. Red: lestaurtinib. (D) Effect of lestaurtinib on MAST1 activity in vitro (left) and in vivo in KB-3-1 cells (right). Purified GST-MAST1 variants were treated with increasing concentrations of lestaurtinib. (E) Effect of lestaurtinib on MAST1 activity at a range of ATP concentrations. (F) Cellular thermal shift assay using KB-3-1 cells harboring WT or L504D MAST1 treated without (−) or with (+) lestaurtinib. (G) Kinase activity of MAST1 WT or L504D treated with lestaurtinib. (H) cisplatin IC₅₀ upon lestaurtinib treatment in KB-3-1 cells with endogenous MAST1 knockdown expressing MAST1 WT or L504D in KB-3-1 cells. (I) Effect of lestaurtinib and WT or L504D MAST1 overexpression on MEK and cRaf phosphorylation in KB-3-1 cells. (J) Cisplatin IC₅₀ upon lestaurtinib treatment in cells with or without MAST1 knockdown in KB-3-1 cells. (K) Tumor growth of lestaurtinib and cisplatin treated mice carrying KB-3-1^cisR cell xenografts. Error bars represent SEM. Representative tumors for each group are shown. Scale bar represents 5 mm. (L) Tumor weight (left) and Ki-67 IHC staining (right) at the experimental endpoint. Error bars represent SD and scale bars represent 50 μm. (M) MAST1 activities in tumor lysates are shown. Data are mean ± SD from three replicates of each sample except panels (K and L). p values were determined by two-tailed Student’s t test (ns: not significant; *: 0.01 < p < 0.05; **: p < 0.01). See also Figure S6.

Figure 8.. Lestaurtinib sensitizes cancer cells to cisplatin treatment in vitro and in vivo.

(A) Cell viability, colony formation, apoptosis induction, and MEK1 phosphorylation in KB-3-1 and 212LN with vehicle control, lestaurtinib, cisplatin and the combination. (B) Lestaurtinib effect on cell viability and cisplatin sensitivity of TKO-002 cells. Combination index (CI) values for synergistic effect are shown. (C) Effect of lestaurtinib, cisplatin and the combination effect on tumor growth of cis^R HNSCC, lung cancer, and ovarian cancer PDX mice. Pt-8 tumor and Pt-11 tumor in Figure 5C were used for lung cancer PDX and ovarian cancer PDX, respectively. Error bars represent SEM. Scale bars for the dissected tumors represent 5 mm. (D) Ki-67 expression was determined by IHC staining. Scale bars represent 50 μm. (E) MAST1 activity was assessed by MAST1 in vitro kinase assay using MBP as a substrate. (F) MEK1 inhibition by lestaurtinib and cisplatin combination in PDX tumor lysates. (G) Proposed model for the role of MAST1 in cisplatin resistance in human cancer. Left: Cancer cells rely on cRaf-dependent MEK1 activation to promote tumor growth in the absence of cisplatin. Right: Cisplatin treatment dissociates cRaf from MEK1, while MAST1 phosphorylates MEK1 to activate the MAPK pathway in cRaf-independent manner, inhibiting BIM and providing a proliferative advantage to cancer cells. Targeting MAST1 by lestaurtinib restores cisplatin sensitivity to cells. Data are mean ± SD from three technical replicates of each sample except panel (C). p values were determined by two-tailed Student’s t test (*: 0.01 < p < 0.05; **: p < 0.01). See also Figures S7 and S8.