MAPK-driven MYC/p21 axis controls senescence bypass and BCL-XL dependence in vincristine-resistant rhabdomyosarcoma

Summary

Therapy resistance remains the primary cause of treatment failure in rhabdomyosarcoma (RMS), the most common pediatric soft tissue sarcoma, yet underlying mechanisms are poorly understood. Using vincristine-resistant fusion-negative RMS models, RNA sequencing, genetic perturbations, and a syngeneic zebrafish tumor model, we demonstrate that resistance is driven by enhanced MAPK signaling, validated in post-chemotherapy patient specimens. MEK inhibition induced senescence through an MYC-p21 axis; MYC represses CDKN1A transcription, and its overexpression bypasses senescence. Resistant cells exhibited BCL-XL dependence, and combined MEK/BCL-XL inhibition eradicated resistant tumors in vitro and in vivo. These findings reveal that enhanced MAPK signaling and MYC-mediated senescence bypass are candidate resistance-associated mechanisms in fusion-negative RMS with implications across cancer types, and identify targetable vulnerabilities (MEK and BCL-XL), providing preclinical rationale for treatment-refractory RMS.

Graphical abstract

Highlights

• •MAPK pathway activation in vincristine-resistant fusion-negative rhabdomyosarcoma
• •MEK inhibition induces senescence in vincristine-resistant fusion-negative RMS
• •The MYC-p21 axis functions as a gatekeeper of senescence bypass
• •Dual MEK/BCL-XL inhibition targets multi-hit senescence bypass in resistant RMS

Therapy; Molecular mechanism of gene regulation; Cancer

Introduction

Rhabdomyosarcoma (RMS) is a malignant soft tissue tumor arising from impaired muscle differentiation. It is the most common pediatric soft tissue sarcoma, with approximately 300 cases annually in the United States.^1,2 RMS is classified into two major molecular subtypes: fusion-negative (FN) RMS and fusion-positive (FP) RMS. FN RMS, the most prevalent subtype, is driven by mutations in the receptor tyrosine kinase/RAS/PIK3CA signaling axis in over 90% of cases.³ In contrast, FP RMS is defined by a translocation involving PAX3 or PAX7 fused to FOXO1.^4,5

Despite aggressive treatment with surgery, chemotherapy, and radiation, outcomes remain poor for patients with relapsed or metastatic disease, with 3-year survival rates of less than 30%.⁶ Moreover, standard therapies are associated with significant long-term morbidity and secondary malignancies. No new effective treatments have emerged for relapsed RMS over the past three decades, underscoring the urgent need to elucidate the biological mechanisms underlying therapy resistance to develop more effective and less toxic treatment options.

However, the current understanding of chemotherapy resistance in RMS remains limited. Prior studies have shown that RMS cells treated with vincristine or doxorubicin retain an MYOD1-positive population,⁷ which may contribute to disease persistence. MYOD1, a master regulator of skeletal muscle differentiation, is aberrantly regulated in RMS, leading to repression of myogenic genes and promotion of proliferation.^8,9 Other studies have identified elevated anti-apoptotic proteins and activation of survival pathways, such as GP130/STAT and Hedgehog, in therapy-resistant RMS.^10,11 These data suggest that therapy resistance is accompanied by dynamic shifts in transcriptional and signaling programs, but the mechanisms enabling the long-term survival of therapy-resistant RMS cells remain incompletely characterized.

We have previously shown that vincristine treatment induces a primitive stem-like phenotype in RMS cells.¹² Building on these findings, we investigated the underlying biological mechanisms essential for maintaining long-term survival of therapy-resistant FN RMS cells. Our study identified upregulated MAPK signaling as a key driver of senescence bypass in vincristine-resistant FN RMS. We further demonstrated that the MEK/MYC/p21 axis and its dependence on BCL-xL-mediated survival represent targetable vulnerabilities. These findings suggest that co-targeting MAPK signaling and senescence-survival pathways is a promising therapeutic strategy for treatment-refractory FN RMS.

Results

Increased activity of the MAPK pathway in vincristine-resistant FN RMS cells

To investigate the underlying biological features of vincristine-resistant FN RMS cells, we generated two therapy-resistant FN RMS lines (RD and SMS-CTR) through gradual vincristine dose escalation over 9–12 months, achieving at least a 100-fold increase in IC50 (see schematic in Figure 1A; dose-response curves in Figures S1A and S1B). Vincristine-resistant RD cells were collected for RNA sequencing (RNA-seq). Pathway and oncogenic signature analyses revealed upregulation of MAPK signaling components, including RAS and RAF (Figure 1B). Similar upregulation of RAS/MAPK signaling components was also observed in a vincristine-resistant FP RMS line, Rh30 (Figure S1C), suggesting that this alteration is not subtype-specific.

Figure 1.

Increased MAPK signaling activity in therapy-resistant FN RMS cells

(A) Schematic of the generation of vincristine-resistant FN RMS cells in culture via gradual dose escalation over time. The y axis represents IC50 (nM) of RMS cells treated with vincristine; the x axis represents duration of culturing in days.

(B) Oncogenic signature analysis using the Broad Institute Molecular Signatures Database, based on gene expression profiles of vincristine-resistant RD cells from RNA sequencing.

(C) Representative immunofluorescence (IF) images for phospho-ERK in vincristine-sensitive parental (P) and vincristine-resistant RD and SMS-CTR cells (Vinc R) from one of three independent platings. Green, phospho-ERK; blue, DAPI. Scale bars, 300 μm.

(D) Representative immunohistochemistry (IHC) images for phospho-ERK on formalin-fixed paraffin-embedded (FFPE) tissue sections from an FN RMS patient: primary tumor (top) and matched residual tumor post-chemotherapy (bottom). Scale bars, 100 μm.

(E) Quantitative summary of phospho-ERK IHC in primary and post-chemotherapy FN RMS FFPE patient samples (n = 10 FN RMS patients). Matched samples are color-coded and connected by dashed lines. Untr, untreated; Tre Res, treated residual. ∗∗∗p < 0.005 by unpaired t test. Horizontal line represents the mean.

To validate MAPK pathway activation, we performed immunofluorescence analysis, which showed elevated phosphorylated ERK (pERK) expression in both vincristine-resistant RD and SMS-CTR cells compared with the parental lines (Figure 1C). Similarly, in archival formalin-fixed, paraffin-embedded (FFPE) tumor samples from 10 FN RMS patients, including 8 with matched pretreatment samples, residual post-chemotherapy tumors were enriched for pERK-positive cells in 7 of 8 matched pairs (Figures 1D and 1E). Next-generation sequencing data were available for three primary tumors with increased pERK positivity. Two harbored NRAS mutations, and one had wild-type RAS but carried NF1 and PTEN deletions. These findings suggest that MAPK signaling plays a key role in driving therapy resistance in FN RMS.

Therapeutic efficacy of MEK inhibition as monotherapy and in combination with vincristine

Given the activation of MAPK signaling, we assessed whether MEK inhibition could suppress the growth of vincristine-resistant FN RMS cells. Treatment of vincristine-resistant RD and SMS-CTR cells with trametinib, a MEK inhibitor, significantly reduced cell proliferation in a dose-dependent manner (Figure 2A). In contrast, vincristine-resistant Rh30 (FP RMS) cells showed lower sensitivity to trametinib (∼3- to 5-fold increase in IC50) (Figure S1B). Similarly, MEK1 knockdown using two independent shRNAs (Figure 2B) impaired the growth of resistant RD and SMS-CTR cells (Figure 2C). Notably, vincristine-resistant RD cells demonstrated greater sensitivity to MEK1 inhibition than parental cells, while this differential effect was not observed in SMS-CTR cells (Figure 2B).

Figure 2.

Trametinib reduces growth and relapse in vincristine-resistant FN RMS

(A) Dose-response curves of parental and vincristine-resistant RD and SMS-CTR cells treated with trametinib, normalized to DMSO (vehicle control) cells. P, parental; Vinc R, vincristine-resistant. Each dose was tested in 3 replicate wells. Shown are results from one of at least three independent experiments. Error bars, standard deviation.

(B) Summary of CellTiter-Glo viability assays in RD and SMS-CTR cells (parental and vincristine-resistant) transduced with two independent MEK1-targeting shRNAs. Scrambled shRNA was used as a control. Cells were plated for the assay 6 days post-transduction, and relative fold changes in luminescence were measured over a 5-day period. Data are from one of three independent experiments with 3–4 samples per condition in each experiment. ns, not significant; ∗∗∗p < 0.001; ∗∗∗∗p < 0.0001 by two-way ANOVA with multiple comparisons correction.

(C) Western blot analysis of MEK1 protein levels at 5 days post-shRNA transduction. GAPDH served as a loading control. Normalized band intensity for MEK1 is indicated below each lane. The experiment was repeated twice. Each experiment consisted of 3 samples (n = 3, control and 2 samples transduced with 2 independent shRNAs, per cell line).

(D) Quantification of relative tumor volume changes from zebrafish drug treatment studies. Each symbol represents one animal (DMSO: n = 4; trametinib: n = 5; vincristine: n = 6; trametinib + vincristine: n = 7). Two-way ANOVA with multiple comparisons correction; ns, not significant; ∗p < 0.05; ∗∗∗∗p < 0.0001.

(E) Representative images of zebrafish engrafted with KRAS (G12D)-induced FN RMS tumors (co-expressing mCherry) treated with DMSO (vehicle), vincristine (0.1 mg/kg), trametinib (10 mg/kg), or the combination. Tumor-bearing zebrafish received intraperitoneal injections on days 0 and 4 over a 7-day period. Scale bars, 2 mm. The experiment was repeated twice, and for each experiment, DMSO: n = 4; trametinib: n = 5; vincristine: n = 6; trametinib + vincristine: n = 7.

(F) Quantification of cleaved Caspase-3 (CC3) IHC. The percentage of CC3-positive cells per whole tumor cross-section was measured using QuPath (n = 3 samples per condition). Two-way ANOVA with multiple comparisons correction; ns, not significant; ∗∗∗p < 0.001; ∗∗∗∗p < 0.0001.

(G) Representative tumor sections stained with H&E (top) and CC3 IHC (bottom). Scale bars, 100 μm. The experiment was repeated 3 times, and for each replicate, n = 3 per condition.

(H) Tumor volume measurements of relapsed tumors over time post-treatment. Each dot represents an animal (DMSO: n = 4; trametinib: n = 5; vincristine: n = 6; trametinib + vincristine: n = 7). Two-way ANOVA with multiple comparisons correction; ∗p < 0.05. In (B), (D), (F), and (H), each horizontal line represents the mean for each condition.

To evaluate whether MEK inhibition enhanced vincristine efficacy in vivo, we used a KRAS (G12D)-induced zebrafish FN RMS model. Primary RMS cells expressing the mylz:mCherry reporter were transplanted into syngeneic CG1 zebrafish, followed by treatment with vehicle (DMSO), trametinib (10 mg/kg), vincristine (0.1 mg/kg), or both drugs in combination for 7 days. The drug doses were titrated to IC50 values in vivo. Combination treatment significantly reduced tumor growth compared with either monotherapy (n = 4–7 per group, Figures 2D and 2E). Apoptosis, assessed via cleaved Caspase-3 immunohistochemistry (IHC), was significantly increased in the combination group compared with that in the single-agent arm (Figures 2F and 2G), supporting enhanced tumor cell death.

To test whether this regimen also prevented relapse, we monitored post-treatment tumor regrowth. By day 17, zebrafish receiving combined vincristine and trametinib showed a significantly lower relapse frequency and slower growth of recurrent tumors than those treated with either drug alone (p < 0.05, Figure 2H). These results indicate that dual targeting of MEK and vincristine may improve the treatment durability in FN RMS.

Trametinib treatment alters cell cycle progression, differentiation, and induces senescence in FN RMS cells

To investigate the cellular mechanisms underlying growth inhibition following MEK1 knockdown or trametinib treatment, we assessed alterations in cell cycle, apoptosis, and myogenic differentiation. DAPI-based flow cytometry revealed that vincristine-resistant RD and SMS-CTR cells exhibited a prolonged G0/G1 phase compared with the parental cells (representative flow cytometry plots for RD in Figure 3A). Trametinib treatment further exacerbated G0/G1 arrest in both parental and vincristine-resistant cells (Figures 3A and 3B).

Figure 3.

Trametinib induces cell-cycle arrest in vincristine-resistant FN RMS cells

(A) Representative cell cycle histograms by DAPI-based flow cytometry. Black, DMSO; red, trametinib (100 nM).

(B) Summary of cell cycle distribution from one of three independent experiments. Error bars, standard deviation of three samples within the experiment.

(C) Representative IF images of RD cells stained with MF20 antibody, day 3 of drug treatment. Green, MF20; blue, DAPI. Scale bars, 300 μm.

(D) Quantification of MF20 IF in RD and SMS-CTR cells (parental vs. vincristine-resistant) from one of three independent experiments. For each experiment, the percentage of MF20-positive cells per field was calculated from 5 to 6 independent fields over 3 replicate wells at 200× magnification (minimum ∼50 cells counted per field). Data are presented as mean ± SD across all fields within each condition.

(E) Representative IF images of RD cells stained with antibody for MYOD1, day 3 of drug treatment. Green, MYOD1; blue, DAPI. Scale bars, 300 μm.

(F) Quantification of MYOD1 IF in RD and SMS-CTR cells (vehicle control and trametinib-treated). Representative data from one of three independent experiments; all experiments demonstrated consistent trends. For each experiment, the percentage of MYOD1-positive nuclei per field was calculated from 5 to 6 independent fields over 3 replicate wells at 200× magnification (minimum ∼50 cells counted per field). Data are presented as mean ± SD across all fields within each condition.

(G) Quantitative RT-PCR of MYOD1 and MYOG expression parental and vincristine-resistant RD and SMS-CTR cells treated with vehicle control or trametinib (10 nM) for 24 h. Representative data from one of three independent biological replicates; all experiments showed consistent trends. Gene expression was normalized to GAPDH and presented as fold-change relative to vehicle-treated cells using the ΔΔCt method. Each condition includes three technical replicates, and data are displayed as mean ± SD from three independent replicates. In (B), ∗∗∗∗p < 0.0001 by unpaired Student’s t test. In (D), (F), and (G), ns, not significant; ∗p < 0.05; ∗∗p < 0.01; ∗∗∗p < 0.001; ∗∗∗∗p < 0.0001 by two-way ANOVA with multiple comparisons correction. In (D), (F), and (G), horizontal line represents the mean for each condition.

Next, we evaluated apoptosis using live-cell imaging with NucView 488 Caspase-3 substrate 6 days after shRNA transduction or 3 days of trametinib treatment. Trametinib did not significantly increase apoptosis in either parental or vincristine-resistant RD or SMS-CTR cells compared with DMSO controls (Figures S2A and S2B). While MEK1 shRNA induced a statistically significant increase in apoptosis in vincristine-resistant RD cells compared with parental cells, the effect was modest (mean 1.2% vs. 0.3%, Figures S2C and S2D), suggesting that apoptosis contributes minimally to the reduced cell viability.

Trametinib has previously been shown to induce terminal differentiation of FN RMS cells.¹³ In agreement with previous findings, trametinib-treated parental RD and SMS-CTR cells exhibited increased numbers of MF20-positive differentiated muscle cells (Figures 3C and 3D). However, trametinib-treated vincristine-resistant RD and SMS-CTR cells failed to undergo terminal differentiation to the same extent, as shown by the reduced MF20 staining (Figures 3C and 3D). Additionally, MYOD1 protein expression decreased in resistant cells following trametinib treatment (Figures 3E and 3F), and quantitative RT-PCR confirmed attenuated induction of MYOD1 and MYOG (20- to 30-fold lower) compared with parental cells (Figure 3G).

Vincristine is known to induce senescence in various cancers through therapy-induced senescence (TIS) mechanisms.^14,15 Consistent with this, vincristine-resistant RD and SMS-CTR cells displayed an enlarged, flattened morphology consistent with senescence (Figure 4A). We used senescence-associated β-galactosidase (SA-β-gal) staining, a common marker of senescence, and found that vincristine-resistant RD and SMS-CTR cells had more β-gal-positive cells than treatment-naive parental cells (Figures 4B, 4C, and S3A).

Figure 4.

Trametinib induces senescence in vincristine-resistant FN RMS cells

(A) Representative bright-field images of parental (top) and vincristine-resistant (bottom) RD and SMS-CTR cells. Vincristine-resistant cells display larger, irregular morphology. Scale bars, 200 μm.

(B) β-Gal staining of RD parental and vincristine-resistant cells after 3 days of treatment with DMSO or trametinib (100 nM). Scale bars, 300 μm.

(C) Quantification of β-Gal staining in RD cells (≥3 fields per condition over duplicate wells at 10× from one of three independent experiments). ns, not significant; ∗∗∗∗p < 0.0001 by two-way ANOVA with multiple comparisons correction.

(D) Western blots of total lysates from RD parental and vincristine-resistant cells treated with DMSO or 10 nM trametinib for 24 h of drug treatment. GAPDH, loading control. The experiment was repeated three times. Each replicate had four samples (n = 4, parental and vincristine-resistant RD cells treated with vehicle, DMSO, and trametinib).

(E) IF for p16 in RD cells after 72 h of drug treatment. Red, p16; blue, DAPI. Scale bars, 300 μm.

(F) Quantification of p16 IF in RD cells (≥3 fields per condition over triplicate wells at 10× from one of three independent experiments). ns, not significant; ∗∗∗∗p < 0.0001 by two-way ANOVA with multiple comparisons correction.

(G) IF for p21 in RD cells after 72 h of drug treatment. Red, p21; blue, DAPI. Scale bar: 300 μm.

(H) Quantification of p21 IF in RD cells (≥3 fields per condition over triplicate wells at 10× from one of three independent experiments). ns, not significant; ∗p < 0.05; ∗∗∗p < 0.001; ∗∗∗∗p < 0.0001 by two-way ANOVA with multiple comparisons correction.

(I) Quantitative RT-PCR of vincristine-resistant RD cells following MYC disruption using Cas9 and dual gRNAs. RNA was collected 6 days post-transduction. Gene expression was normalized to GAPDH and presented as fold-change relative to scrambled control guide RNA using the ΔΔCt method. Data are from 1 of 3 independent experiments and are displayed as mean ± SD from three technical replicates per experiment (n = 3). Error bars represent SD across the three replicates. Statistical significance was determined by unpaired two-tailed t test: ns, not significant; ∗p < 0.05.

(J) β-Gal staining of vincristine-resistant RD cells transduced with GFP control or MYC-overexpression lentivirus and treated with DMSO or trametinib (10 nM) for 3 days. Scale bars, 300 μm.

(K) Quantification of β-Gal staining (5–6 fields over duplicate wells per condition at 20× from one of three independent experiments). ns, not significant; ∗∗p < 0.01; ∗∗∗∗p < 0.0001 by two-way ANOVA with multiple comparisons correction.

(L) Western blots of total lysates from RD vincristine-resistant cells stably transduced with control GFP or MYC overexpression lentiviral constructs, treated with DMSO or trametinib for 24 h. GAPDH, loading control. Quantitative results (normalized to GAPDH) are shown below corresponding bands. The experiment was repeated twice. Each replicate had four samples (n = 4; RD cells harboring GFP overexpression control and MYC overexpression treated with vehicle DMSO or trametinib 10 nM).

(M) CUT&RUN quantitative PCR for assessing MYC protein binding to the CDKN1A (p21) promoter in vincristine-resistant RD and SMS-CTR cells treated with trametinib for 24 h. The results are the average of three independent experiments, each with three technical replicates per condition. ∗∗∗p < 0.001 by one-way ANOVA with multiple comparisons correction. In (C), (F), (H), (K), and (M), horizontal line represents the mean for each condition.

Given that trametinib alters the cell cycle without inducing apoptosis or robust differentiation, we investigated whether it also promotes senescence in vincristine-resistant cells. Indeed, trametinib-treated vincristine-resistant RD and SMS-CTR cells showed a significant increase in β-gal-positive senescent cells compared with DMSO-treated controls (Figures 4B, 4C, and S3A). Parental cells did not exhibit this increase, supporting a resistance-associated senescent phenotype.

Next, we investigated whether p21 and p16, which inhibit cell cycle progression, contribute to senescence-associated cell-cycle arrest. Western blotting and immunofluorescence analysis showed that trametinib-treated vincristine-resistant RD and SMS-CTR cells had increased p21 expression but reduced p16 levels (Figures 4D–4H and S3B–S3D). These findings suggest that p21, rather than p16, is the primary driver of trametinib-induced senescence in vincristine-resistant FN RMS cells.

Senescence-associated secretory phenotype (SASP) is associated with senescent cells secreting pro-inflammatory factors that promote cancer relapse and therapy resistance.^16,17 Because SASP is highly context-dependent and heterogeneous across cell types and senescence inducers, we evaluated a focused panel of well-characterized SASP markers including CCL2, MMP2, and IL6. We observed downregulation of CCL2 and MMP2, but no significant change in IL6 expression. The attenuated and variable SASP response we observed is consistent with TIS, which frequently exhibits reduced SASP secretion compared with replicative or oncogene-induced senescence (OIS).¹⁸ Importantly, these findings indicate that the senescence phenotype induced by MEK inhibition is driven primarily by cell-cycle arrest and sustained growth suppression rather than by a secretory program. A comprehensive SASP profiling including additional time points and a broader range of secreted factors could provide deeper mechanistic insights in future investigations.

MYC represses p21 and suppresses trametinib-induced senescence in FN RMS

MYC is a known downstream effector of MAPK signaling and is a critical regulator of proliferation. We previously showed that MYC promotes the survival of vincristine-resistant FN RMS cells. Based on our observation that trametinib induces senescence and increases p21, we hypothesized that MYC suppresses p21 expression to block senescence.

We first confirmed by Western blotting that trametinib treatment reduced MYC protein levels in RD and SMS-CTR cells (Figures 4D and S3B). MYC protein stability is tightly regulated by phosphorylation at two critical sites: phosphorylation at Ser62 suppresses MYC degradation and stabilizes the protein for transcriptional activity, while phosphorylation at Thr58 triggers ubiquitin-mediated degradation.^{19,20,21,22,23,24} To determine whether trametinib treatment affects MYC protein stability, we performed cycloheximide chase experiments, which demonstrated similar MYC degradation kinetics between DMSO-treated control and trametinib-treated RD cells over 90 min (Figures S4A and S4B). Furthermore, trametinib treatment did not alter the phosphorylation status of MYC at Ser62 or Thr58 relative to total MYC levels in both RD and SMS-CTR cell lines (Figure S4C). These experiments, demonstrating stable MYC degradation kinetics and unchanged MYC phosphorylation status, rule out post-translational mechanisms as the primary driver of MYC downregulation. Collectively, these findings establish that trametinib-mediated reduction in MYC protein levels occurs through transcriptional suppression rather than altered protein stability or post-translational modification.

Next, we disrupted MYC using lentiviral CRISPR-Cas9 with validated dual gRNA constructs, which we have previously shown to effectively deplete MYC protein in RMS cells.¹² MYC disruption led to upregulation of CDKN1A (p21) mRNA, while CDKN2A (p16) levels remained unchanged (Figure 4I), supporting a specific MYC-p21 axis.

To test whether MYC regulates senescence in this context, we overexpressed MYC in RD cells and found that MYC overexpression alleviated trametinib-induced senescence, as measured by the reduced β-gal-positive cell frequency (Figures 4J and 4K). Similarly, MYC overexpression partially abrogated increased expression of p21 in response to trametinib treatment (Figure 4L). MYC has been shown to repress gene expression by interacting with Sp1/Sp3 proteins, which bind to GC-rich sites, including the GGGCGG motif, in the promoter.²⁵ To determine whether MYC suppresses p21 expression in vincristine-resistant FN RMS cells, we performed quantitative CUT&RUN quantitative PCR and showed MYC binding to the promoter region of CDKN1A (encoding p21), and this binding was abrogated by trametinib treatment (Figure 4M). These findings demonstrate that MYC suppresses p21-mediated senescence through transcriptional repression and is required for resistance to trametinib-induced cell-cycle arrest in FN RMS.

BCL-xL supports survival of senescent, vincristine-resistant FN RMS cells and represents a targetable vulnerability

While a subset of vincristine-resistant cells exhibited senescence markers, such as altered morphology and increased β-galactosidase staining (Figures 5A and 5B), the continued proliferation of the population suggests that some cells escape or bypass senescence-associated arrest. Prior studies in colorectal cancer have shown that escape from p21-mediated senescence can result in dependence on anti-apoptotic proteins, including BCL-xL and MCL1.²⁶

Figure 5.

Combined MEK and BCL-xL inhibition reduces growth of vincristine-resistant FN RMS cells

(A) Western blots of total lysates from parental and resistant RD and SMS-CTR cells treated with DMSO or trametinib (10 nM) for 24 h. GAPDH, loading control. Quantitative results (normalized to Tubulin) are shown below corresponding bands.

(B) CellTiter-Glo assays assessing trametinib and A-1155463 dose responses in parental and vincristine-resistant RD (left) and SMS-CTR (right) cells. Luminescence normalized to DMSO controls. Error bars represent standard deviation of three replicates from one of three independent experiments. ∗∗p < 0.01 and ∗∗∗∗p < 0.0001 by unpaired t test.

(C) Heatmaps and Bliss scores from 6-day combination treatment assays using trametinib and A-1155463. Heatmaps and Bliss synergy scores were generated using Synergy Finder (www.synergyfinder.org). The experiment was repeated at least three times, and each experiment had three replicate wells per treatment dose per condition.

(D) Representative images of vincristine-resistant RD cells from live imaging using fluorogenic NucView488 Caspase-3 substrate for detecting apoptotic cells 5 h post-treatment with drugs. Green, cleaved Caspase 3-positive apoptotic cells. Scale bars, 300 μm. The experiment was repeated three times, n = 4 per condition per experiment.

(E) Quantitative summary of cleaved Caspase-3 labeling in vincristine-resistant RMS cells. Left, RD; right, SMS-CTR. Each dot represents a replicate well quantified under 200× magnification for each condition (4 wells per condition) from one of three independent experiments. ns, not significant; ∗∗∗p < 0.001; ∗∗∗∗p < 0.0001 by one-way ANOVA with multiple comparisons correction. Horizontal line represents the mean for each condition.

To test whether vincristine-resistant FN RMS cells developed a similar dependence on anti-apoptotic proteins, we examined protein expression of BCL-xL, BCL2, and MCL1 following trametinib treatment. Western blot analysis revealed that, in contrast to the parental cells, vincristine-resistant RD and SMS-CTR cells showed increased BCL-xL expression following trametinib exposure (Figure 5A), while there was a decrease in BCL2 or MCL1 expression. Notably, these resistant cells demonstrated increased sensitivity to the selective BCL-xL inhibitor A-1155463 compared with the parental cells, as shown by dose-response assays (Figure 5B).

Given the apparent co-dependencies on MEK signaling and BCL-xL-mediated survival, we hypothesized that the combination targeting of these pathways would enhance therapeutic efficacy. Indeed, combination treatment with trametinib (5 nM) and A-1155463 (300 nM) at multiple doses significantly inhibited the growth of vincristine-resistant RD and SMS-CTR cells relative to either agent alone by inducing a greater increase in apoptosis (Figures 5C–5E).

To evaluate this drug combination strategy in vivo, we treated engrafted primary FN RMS tumors in zebrafish with vincristine (0.25 mg/kg), which led to a ∼90% tumor volume reduction over 7 days. Zebrafish were maintained on weekly vincristine dosing until relapsed tumors emerged (approximately day 21 after starting vincristine treatment). Each relapsed tumor was expanded into syngeneic CG1 host animals under vincristine exposure to ensure continued resistance and was subjected to combination or single-agent treatment. Compared with monotherapy, combined treatment with trametinib and A-1155463 more effectively inhibited the growth of tumors derived from two independent vincristine-resistant tumors (Figures 6A and 6B; p < 0.01). This is supported by a greater extent of apoptosis in the tumors treated with the two-drug combination compared with single-agent treatment (Figures 6C and 6D).

Figure 6.

Combined MEK and BCL-XL inhibition reduces growth of vincristine-resistant tumors in an FN RMS zebrafish model

(A) Representative images of zebrafish bearing vincristine-resistant FN RMS tumors (co-expressing mCherry) treated with DMSO (vehicle), trametinib (2.5 mg/kg), A-1155463 (10 mg/kg), or combination of trametinib and A-1155463 at day 0 and day 7. Scale bars, 2 mm.

(B) Quantification of tumor volume changes over 7 days of treatment. Each dot represents a single animal. Results shown are from two independent experiments as represented by two different colors (black and red). Total animal numbers per group: DMSO: n = 10; trametinib: n = 12; A1155463: n = 11; trametinib and A1155463: n = 10. ns, not significant; ∗∗p < 0.01 by one-way ANOVA with multiple comparisons correction.

(C) Representative tumor sections stained with H&E (top) and CC3 IHC (bottom). Scale bars, 100 μm.

(D) Quantification of cleaved Caspase-3 (CC3) IHC. The percentage of CC3-positive cells per whole tumor cross-section was measured using QuPath (n = 3 samples per condition). ∗∗∗p < 0.001 by one-way ANOVA with multiple comparisons correction.

In (B) and (D), horizontal line represents the mean for each condition.

These findings indicate that senescence-associated survival mechanisms mediated by BCL-xL can be therapeutically exploited in vincristine-resistant FN RMS. Dual inhibition of MEK and BCL-xL represents a rational strategy to eradicate treatment-refractory tumor cells.

Discussion

The MAPK signaling paradox in cancer senescence

The relationship between oncogenic MAPK signaling and cellular senescence is fundamentally paradoxical. In normal human fibroblasts, MEK/MAPK activation induces both p53 and p16, which are required for RAS-induced senescence as a protective mechanism against transformation.^27,28,29 Similarly, constitutive MAPK activation induces oncogene-dependent senescence in cancer cells, a well-established tumor suppressive mechanism.¹⁷ Yet paradoxically, inhibition of oncogenic MAPK signaling also induces senescence in KRAS-mutant cancers, including lung cancer, pancreatic cancer, and pilocytic astrocytoma.^30,31,32 This contradiction suggests that MAPK signaling operates in a dosage-dependent manner: both excessive activation and targeted inhibition can trigger senescence, yet the mechanistic basis for this bidirectional senescence-inducing capacity in therapy-resistant cancers remains poorly understood.

To address this gap in therapy-resistant RMS, we demonstrated using a syngeneic zebrafish tumor model that MEK inhibition induces senescence in vincristine-resistant fusion-negative RMS cells. Critically, this senescence-inducing effect was independent of the specific upstream RAS mutation, as validated by two converging lines of evidence: (1) enhanced pERK was detected in 7 of 8 matched post-chemotherapy patient tumors regardless of RAS genotype and (2) human FN RMS cell lines with distinct genetic backgrounds (NRAS-mutant RD and HRAS-mutant SMS-CTR) both responded to MEK inhibition. These findings establish that enhanced MAPK signaling itself, rather than any single upstream driver, is the critical mechanism promoting vincristine resistance. Notably, the phenotypic response to MEK inhibition diverged dramatically between parental and resistant cells: treatment-naive parental cells underwent myogenic differentiation,¹³ whereas vincristine-resistant cells exhibited elevated MAPK activity and did not differentiate but instead underwent senescence-associated cell-cycle arrest. This striking divergence reveals that the elevated MAPK activity in resistant cells creates a senescence-prone state in which MEK inhibition paradoxically reinforces growth arrest. Our findings support a dosage-dependent model in which oncogene addiction (parental cells) permits lineage differentiation, whereas enhanced but intermediate MAPK elevation (resistant cells) creates a tipping point favoring senescence over proliferation or differentiation.

The MYC-p21 axis as a gatekeeper of senescence bypass

MYC has a multifaceted relationship with senescence, as it regulates the expression of numerous senescence-associated genes and exerts distinct context-dependent roles in senescence regulation depending on cell type. In normal cells, MYC overexpression can paradoxically trigger OIS through induction of DNA damage and activation of stress response pathways.^17,33,34 Conversely, in established tumors, MYC primarily functions as an anti-senescence oncogene by suppressing senescence checkpoints. For instance, in mouse lung tumors, MYC cooperates with mutant BRAF V600E to promote growth specifically through senescence suppression.³⁵ Similarly, in KRAS-mutant pancreatic cancers, targeted ERK inhibition induces therapeutic senescence through MYC degradation,³¹ linking MYC downregulation to treatment-induced senescence. Notably, how MYC sustains senescence bypass specifically in therapy-resistant cancers remains incompletely understood. In our study, we addressed this mechanistic gap by identifying the MYC-p21 axis as the critical effector mechanism driving senescence bypass in resistant cells. MEK inhibition suppressed MYC expression and elevated the cyclin-dependent kinase inhibitor p21 (CDKN1A), triggering irreversible cell-cycle arrest. Mechanistically, MYC directly repressed CDKN1A transcription by binding to regulatory elements within the CDKN1A promoter, as demonstrated by CUT&RUN chromatin analysis. Loss-of-function studies using CRISPR/Cas9-mediated MYC disruption confirmed that MYC is necessary for senescence bypass, as MYC depletion restored p21 expression and enforced growth arrest upon MAPK inhibition. Complementary gain-of-function studies demonstrated sufficiency: forced overexpression of MYC in trametinib-treated cells abrogated senescence state and p21 induction, providing direct evidence that MYC is sufficient to maintain proliferative capacity despite active MAPK inhibition. These complementary gain-of-function and loss-of-function studies establish MYC as both necessary and sufficient for senescence bypass in vincristine-resistant cells. In addition, the multi-layered control with transcriptional suppression of p21 by MYC, complemented by post-translational antagonism via BCL-XL explains the redundancy of the senescence evasion program and justifies dual targeting of MAPK and anti-apoptotic pathways in therapy-resistant disease.

BCL-XL dependence emerges as a vulnerability in senescence-resistant cells

Beyond the MAPK-MYC-p21 axis, vincristine-resistant cells exploit anti-apoptotic pathways to evade TIS. We demonstrated that resistant cells exhibit significantly increased dependence on the anti-apoptotic protein BCL-XL compared with treatment-naïve parental cells. BCL-XL antagonizes p21-mediated growth arrest by suppressing apoptosis,²⁶ thereby promoting escape from senescence-associated cell-cycle arrest. This heightened BCL-XL dependence reveals that resistant cells employ multiple convergent mechanisms to bypass senescence: elevated MYC suppresses p21 transcription, while elevated BCL-XL antagonizes p21 function. Combined MEK/BCL-XL inhibition demonstrated synergistic suppression of resistant tumor growth both in vitro and in our syngeneic zebrafish model, with combined treatment significantly outperforming either agent as monotherapy. These findings establish that dual targeting of the MAPK pathway and anti-apoptotic mechanisms is necessary to overcome the redundant pro-survival adaptations in treatment-refractory RMS.

Clinical translational considerations and next-generation approaches

The combination of trametinib (FDA-approved MEK inhibitor) and BCL-XL inhibitors represents a promising treatment strategy for treatment-refractory RMS. However, BCL-2/BCL-XL inhibitors have shown treatment-related adverse events in clinical trials, with thrombocytopenia (low platelet count) being the most common.³⁶ Dose optimization, particularly using lower concentrations of BCL-XL inhibitor in combination with trametinib, may reduce toxicity while maintaining efficacy. Additionally, next-generation PROTAC-based BCL-2 and BCL-XL degraders offer promising alternatives by providing improved target selectivity and potentially reduced systemic toxicity through tumor-microenvironment-preferential degradation.^37,38

An integrated model of therapy resistance and senescence bypass

Our data support a multi-hit model in which vincristine-resistant FN RMS cells achieve therapy resistance through three coordinated mechanisms: (1) enhanced MAPK signaling drives proliferation and suppresses differentiation, (2) elevated MYC activity suppresses p21 (CDKN1A) transcription, and (3) increased BCL-XL antagonizes p21 function. This explains why monotherapy fails—each addresses only one mechanism while the others sustain resistance. Dual MEK/BCL-XL inhibition is synergistic because it simultaneously disrupts both transcriptional (MYC-CDKN1A) and post-translational (BCL-XL-p21) suppression of the senescence checkpoint, thereby overcoming the redundancy of the multi-hit resistance program.

In conclusion, our findings establish the MEK-MYC-p21-BCL-XL axis as the central regulatory network driving therapy resistance in vincristine-resistant FN RMS. By identifying senescence bypass as a dominant resistance mechanism and defining the molecular vulnerabilities that enable it, we propose a rational therapeutic strategy: simultaneous targeting of MAPK and anti-apoptotic pathways using FDA-approved agents (trametinib and BCL-XL inhibitors) or next-generation degraders. This approach addresses the underlying biology of resistance by restoring senescence-dependent tumor suppression. Given the poor prognosis of treatment-refractory RMS and the lack of effective salvage therapies, these findings have immediate clinical relevance and warrant promising translation into clinical trials for this pediatric malignancy.

Limitations of the study

Single-agent versus combination chemotherapy

Standard RMS therapy combines vincristine with actinomycin D and cyclophosphamide, but our study used single-agent vincristine to isolate drug-specific resistance mechanisms without confounding contributions from multiple agents. This approach enabled precise dissection of vincristine-driven pathways but does not capture the complexity of multi-agent therapy. Future studies examining vincristine combinations will clarify whether the identified MAPK-MYC-p21-BCL-XL axis remains relevant in standard-of-care multi-agent chemotherapy.

Association of MAPK activation with resistance

Our study demonstrates that enhanced MAPK signaling is required to maintain the vincristine-resistant phenotype in established resistant fusion-negative RMS cells (via loss-of-function with MEK inhibition). However, whether elevated MAPK signaling is sufficient to initiate de novo resistance in treatment-naïve parental cells remains to be determined. Gain-of-function studies, such as constitutive MAPK pathway activation in parental cells combined with vincristine treatment, would directly test whether elevated MAPK signaling is sufficient to establish or accelerate the development of therapy resistance de novo.

Lack of comprehensive senescence/SASP profiling in resistant fusion-negative RMS cells

Our characterization of senescence in resistant fusion-negative RMS cells employed a comprehensive panel of senescence markers (SA-β-gal staining, cell-cycle arrest, elevated p21/CDKN1A activity) and selective SASP component profiling (MMP2 and CCL2). This focused multi-parameter approach supported senescence induction because (1) functional senescence is defined primarily by irreversible growth arrest, SA-β-gal positivity, and p21/p16-mediated CDK inhibition; (2) senescence-associated secretory phenotype (SASP) is highly heterogeneous and context-dependent, varying substantially based on senescence trigger (replicative, oncogene-induced, and therapy-induced), cell type, and temporal kinetics.^14,39,40,41 Importantly, TIS—the senescence phenotype in our vincristine-resistant RMS model—frequently exhibits attenuated SASP compared with replicative or OIS. Our detection of downregulated MMP2 and CCL2 is consistent with this attenuated SASP pattern and validates our senescence characterization. Nevertheless, comprehensive longitudinal profiling of a broader SASP panel (including additional cytokines, chemokines, growth factors, extracellular proteases, and inflammatory mediators) at multiple time points post-MEK inhibition would provide deeper mechanistic insights into the full senescence phenotype and potentially reveal additional drivers of resistance or disease adaptation in this model.

Fusion status generalizability

Our study focused on fusion-negative RMS. Preliminary observations in one fusion-positive RMS line showed similar MAPK upregulation but reduced MEK inhibitor sensitivity. Additional fusion-positive RMS cell lines are needed to determine whether the mechanistic findings extend across RMS subtypes.

In vivo model limitations

Our study demonstrated that control (vehicle-treated) tumors reached endpoint at approximately 7 days, whereas trametinib combined with vincristine extended survival to day 17 (Figure 2H), representing a >2-fold extension in survival in this rapid-growth model. This accelerated tumor kinetics—approximately 2.5× faster than mammalian xenografts—permits rapid proof-of-principle testing but limits our ability to assess truly long-term sustained responses or directly extrapolate human post-relapse survival timelines. Mammalian xenograft studies would provide additional validation in a more physiologically relevant system. However, consistent results across in vitro, zebrafish, and patient samples support mechanistic conclusions.

Patient sample size

Enhanced phosphorylated ERK was validated in 7 of 8 post-chemotherapy patient pairs. While substantial, this small sample size limits statistical power and generalizability. Validation in a larger independent cohort is warranted.

Clinical drug availability

While trametinib is FDA-approved, the BCL-XL inhibitor used (A-1155463) is not clinically available. FDA-approved BCL-2 selective inhibitors, e.g., venetoclax, have shown toxicity in trials, particularly thrombocytopenia. Next-generation PROTAC-based BCL-XL degraders offer promise but remain in development, delaying clinical translation of the proposed combination therapy.

Resource availability

Lead contact

Requests for further information and resources should be directed to and will be fulfilled by the lead contact, Eleanor Y. Chen (eleanor2@uw.edu).

Materials availability

All plasmids and cell lines generated in this study are available from the lead contact without restriction.

Data and code availability

• •RNA sequencing raw data (FASTQ files) have been deposited at NCBI Gene Expression Omnibus (GEO) as GSE325716 and are publicly available as of the date of publication.
• •This paper does not report original code.
• •Any additional information required to reanalyze the data reported in this paper is available from the lead contact upon request.

Acknowledgments

This study was supported by the Department of Defense, Rare Cancers Research Program, Idea Development Award (HT9425-25-1-0493, RA240110).

Author contributions

E.Y.C., Y.W., and M.I. performed the study concept and design. E.Y.C., Y.W., D.L., A.D.L., P.M., L.W., A.B., and M.I. performed data acquisition, analysis, or interpretation. E.Y.C., Y.W., and A.D.L. wrote the manuscript. E.Y.C., Y.W., D.L., and A.D.L. revised the manuscript. All authors read the approved final manuscript.

Declaration of interests

The authors declare no competing interests.

STAR★Methods

Key resources table

REAGENT or RESOURCE	SOURCE	IDENTIFIER
Antibodies
Rabbit polyclonal anti-Cleaved Caspase-3 (Asp175)	Cell Signaling Technology	Cat#9961; RRID: AB_2341188
Rabbit monoclonal anti-Phospho-p44/42 MAPK (Erk1/2) (Thr202/Tyr204) (clone 20G11)	Cell Signaling Technology	Cat#4370; RRID: AB_2315112
Rabbit polyclonal anti-Phospho-p44/42 MAPK (Erk1/2) (Thr202/Tyr204) (clone D13.14.4E)	Cell Signaling Technology	Cat#9101; RRID: AB_331646
Mouse monoclonal anti-MYH1E (MF 20)	DSHB	Cat#MF 20; RRID: AB_2147781
Rabbit polyclonal anti-p16-INK4A	Proteintech	Cat#10883-1-AP; RRID: AB_2078303
Rabbit monoclonal anti-p21 Waf1/Cip1 (clone 12D1)	Cell Signaling Technology	Cat#2947T; RRID: AB_823586
Rabbit monoclonal anti-MEK1 (clone D2R1O)	Cell Signaling Technology	Cat#12671T; RRID: AB_2737427
Rabbit monoclonal anti-BCL-xL (clone 54H6)	Cell Signaling Technology	Cat#2764T; RRID: AB_2228008
Rabbit monoclonal anti-MYC (clone E5Q6W)	Cell Signaling Technology	Cat# 18583S; RRID: AB_2895543
Rabbit monoclonal anti-Phospho-c-MYC-S62	ABclonal	Cat#AP0989; RRID: AB_2863883
Rabbit monoclonal anti-GAPDH (clone 14C10)	Cell Signaling Technology	Cat#2118S; RRID: AB_561053
Rabbit polyclonal anti-Tubulin	Abcam	Cat#Ab6046; RRID: AB_2210370
Rabbit monoclonal anti-MCL1 (clone D5V5L)	Cell Signaling Technology	Cat#39224; RRID: AB_2799149
Mouse monoclonal anti-BCL2 (clone 124)	Cell Signaling Technology	Cat#15071; RRID: AB_2744528
Bacterial and virus strains
Lentiviruses expressing safe harbor control	Phelps et al.8	N/A
Lentiviruses expressing gene-specific double gRNAs against MYC	Fritzke et al.12	N/A
Biological samples
Archival paraffin-embedded tissue blocks of human RMS tumor samples	Northwest Biospecimen Service, University of Washington Medical Center	https://depts.washington.edu/nwbios/
Chemicals, peptides, and recombinant proteins
Vincristine	Sigma-Aldrich	Cat#S1241
Trametinib	Selleck Chem	Cat#S2673
A-1155463	MedChemExpress	Cat#HY-19725
Cycloheximide	ThermoFisher Scientific	Cat#J66665
Critical commercial assays
Qiagen RNeasy Mini Kit	Qiagen	Cat#74104
Qubit 2.0 Fluorometer (Reagents)	Life Technologies	N/A
NEBNext Ultra RNA Library Prep Kit for Illumina	New England Biolabs	N/A
CellTiter-Glo Cell Viability Assay	Promega	Cat#G7572
NucView® 488 Caspase 3 Substrate	Biotium	Cat#30029
High-Capacity cDNA Reverse Transcription Kit	Thermo Fisher Scientific	Cat#4368814
EpiNext™ CUT&RUN Fast Kit	EpiGentek	Cat#P-2028-24
Deposited data
RNA sequencing raw and processed data	This paper	GEO: GSE325716
Experimental models: Cell lines
Human: RD cell line	ATCC	Cat#CCL-136
Human: SMS-CTR cell line	ATCC	Cat#CRL-3687
Human: Vincristine-resistant RD cell line	This paper	N/A
Human: Vincristine-resistant SMS-CTR cell line	This paper	N/A
Experimental models: Organisms/strains
Zebrafish: syngeneic adult CG1 strain	Smith et al.42	N/A
Zebrafish: rag2:KRASG12D mCherry primary RMS tumors	Ignatius et al.43	N/A
Oligonucleotides
RT-PCR primer: MYOD1 Forward: AGCACTACAGCGGCGACT	Fritzke et al.12	N/A
RT-PCR primer: MYOD1 Reverse: GCGACTCAGAAGGCACGTG	Fritzke et al.12	N/A
RT-PCR primer: MYOG Forward: CCTGCCGTGGGCGTGTAAGG	Fritzke et al.12	N/A
RT-PCR primer: MYOG Reverse: GGACTGCAGGAGGCGCTGTG	Fritzke et al.12	N/A
RT-PCR primer: CDKN2A (p16) Forward: GTGGACCTGGCTGAGGAG	Tani et al.44	N/A
RT-PCR primer: CDKN2A (p16) Reverse: CTTTCAATCGGGGATGTCTG	Tani et al.44	N/A
RT-PCR primer: CDKN1A (p21) Forward: CTGGGGATGTCCGTCAGAAC	Conte et al.45	N/A
RT-PCR primer: CDKN1A (p21) Reverse: GTGACAGGTCCACATGGTCT	Conte et al.45	N/A
CUT&RUN primer: CDKN1A (p21) Forward: CAGTTGCCTGCTCTCCCAGTGACCTTG	This paper	N/A
CUT&RUN primer: CDKN1A (p21) Reverse: AGGGTCTCTCTCACTCCTTCCAGCTGG	This paper	N/A
Recombinant DNA
MEK1 shRNA 1: GCTTCTATGGTGCGTTCTACA	Sigma-Aldrich	N/A
MEK1 shRNA 2: GAGGGAGAAGCACAAGATCAT	Sigma-Aldrich	N/A
Software and algorithms
bcl2fastq v2.17	Illumina	https://support.illumina.com/sequencing/sequencing_software/bcl2fastq-conversion-software.html
STAR v2.5.2b	Dobin et al.46	https://github.com/alexdobin/STAR
featureCounts (Subread v1.5.2)	Liao et al.47	http://subread.sourceforge.net
DESeq2	Love et al.48	https://bioconductor.org/packages/release/bioc/html/DESeq2.html
GeneSCF	Subhash et al.49	http://genescf.altervista.org
QuPath version 0.5.0	Bankhead et al.50	https://qupath.github.io

Experimental models and study participant details

Cell lines

RD cells (female, obtained from ATCC) and SMS-CTR cells (male, obtained from ATCC) were cultured in Dulbecco’s modified Eagle’s medium (DMEM) supplemented with 10% fetal bovine serum, 100 U/mL penicillin and 100 μg/mL streptomycin at 37°C in a 5% CO₂ incubator. All cell lines were authenticated by STR profiling and tested for mycoplasma contamination.

Zebrafish

Zebrafish experiments were conducted using the Danio rerio CG1 strain, an isogenic clonal line that enables syngeneic tumor cell transplantation without immunological rejection.⁴² All zebrafish were housed in the University of Washington Comparative Medicine facility under Institutional Animal Care and Use Committee protocol #4330-01, following standard protocols for temperature (28.5°C), light-dark cycle (14-h light:10-h dark), and tank maintenance. For all experiments, adult CG1 zebrafish aged 4–5 months with approximately equal male-to-female ratios were used. Sex was not a significant variable across experiments.

Method details

Generation of vincristine-resistant RMS cell lines

RD and SMS-CTR FN RMS cell lines were initially treated with vincristine at a dose of 1–1.5 nM which killed approximately 80–90% of the cells. The remaining cells were maintained in long-term culture with weekly replacement of media [10% fetal bovine serum (FBS) and 1% penicillin/streptomycin in Dulbecco’s Modified Eagle Medium (DMEM)] until they grew to >50% confluency, were split to allow regrowth, and then restarted on vincristine at a 10–20% increase in dosage. Incremental increases in drug dosage were continued over the course of 9 months to a year to reach at least a 100-fold increase in the IC50 (half-maximal inhibitory concentration). Vincristine-resistant lines were maintained in 25–35 nM vincristine prior to plating.

CRISPR/Cas9 and shRNA gene targeting in human RMS cell lines

CRISPR/Cas9 targeting was accomplished by transducing RMS cells with lentiviruses expressing safe harbor control⁸ or gene-specific double guide RNAs (gRNAs) against MYC¹² and Cas9. Gene-specific short hairpin RNAs (shRNAs) were obtained from Sigma-Aldrich (Table S1, shRNA sequences). Cells transduced with lentivirus were plated for cell counts and cell-based assays following antibiotic selection six days post transduction. Cloning of the Cas9 and gRNA expression vectors was performed as described previously.⁸

RNA sequencing

RNA was extracted from parental and vincristine-resistant RD cells using the Qiagen RNeasy Mini Kit (catalog number: 74104, Qiagen, Germantown, Maryland), according to the manufacturer’s instructions. RNA sequencing was performed by Azenta Life Sciences (South Plainfield, NJ, USA). RNA quantity and integrity were assessed using a Qubit 2.0 Fluorometer (Life Technologies) and an Agilent TapeStation 4200 (Agilent Technologies). Poly(A)-selected libraries were prepared with the NEBNext Ultra RNA Library Prep Kit for Illumina (New England Biolabs) following the manufacturer’s protocol. Briefly, mRNA was enriched with oligo(dT) beads, fragmented at 94 °C for 15 min, converted to double-stranded cDNA, end-repaired, adenylated, ligated to adapters, and PCR-amplified with indexed primers. Libraries were validated, quantified by qPCR, and sequenced (2 × 150 bp) on an Illumina HiSeq 4000 (or equivalent). FASTQ files were generated and demultiplexed using bcl2fastq v2.17, and reads were trimmed for adapter sequences and quality. Clean reads were aligned to the ENSEMBL reference genome using STAR v2.5.2b, and gene-level counts were obtained with featureCounts (Subread v1.5.2) from uniquely mapped exonic reads. Differential expression analysis was performed with DESeq2 using the Wald test, and genes with adjusted p < 0.05 and |log₂FC| > 1 were considered significant. Gene ontology enrichment was assessed with GeneSCF using the MGI GO database, and principal component analysis (PCA) was performed in DESeq2 (plotPCA) using the 500 most variable genes.

Cell viability assays

RMS cell growth was quantified by cell counting or the ATP-based CellTiter-Glo cell viability assay according to the manufacturer’s protocol (Promega, Madison, WI, USA). Cells were plated in 96-well plates at 1000–2000 cells per well, and luminescence was read using a microplate reader at days 1 and 6 post-plating (BioTek Synergy H1, Winooski, VT). Drug treatments were introduced 1 day after plating.

Immunohistochemistry (IHC) and immunofluorescence (IF)

Archival paraffin-embedded tissue blocks of human RMS tumor samples were retrieved through the Northwest Biospecimen Service at the University of Washington Medical Center using an approved IRB protocol (#9565). De-identified tissue sections of 4-micron thickness were subjected to IHC. Briefly, following deparaffinization and antigen retrieval with Tris-EDTA buffer for cleaved Caspase 3 antibody (Cell Signaling Technology, Danvers, MA) slides and phosphate-buffered saline (PBS) for the phospho-ERK antibody (Cell Signaling Technology, Danvers, MA) slides for 15 min while boiling, the slides were stained with appropriate antibodies using a standard IHC protocol. DAB (3,3-diaminobenzidine) HRP substrate (Cell Signaling Technology, Danvers, MA) was used for signal detection. IHC slides for determining % positive cells across each tumor section were reviewed by two independent board-certified pathologists (EC and AB) and quantified using QuPath version 0.5.0. Immunofluorescence staining was performed as previously described.⁵¹ The antibodies used for IHC and IF are listed in table S1.

Real-time apoptosis detection in cultured RMS cells

To detect apoptosis in cultured cells, adherent RMS cells were incubated with NucView 488 Caspase 3 substrate (Biotium, Fremont, CA) diluted to 2 μM in 500 μL medium for 30 min prior to imaging by microscopy.

Western blots

Human cell lysates were prepared in RIPA buffer with protease inhibitors (Santa Cruz Biotechnology, Dallas, TX) and 2x Laemmli buffer supplemented with β-mercaptoethanol (Bio-Rad, Hercules, CA), electrophoresed on a 4–15% gradient sodium dodecyl sulfate (SDS)-polyacrylamide gel (Bio-Rad, Hercules, CA), and transferred to a polyvinylidene difluoride (PVDF) membrane using TurboTrans-Blot (Bio-Rad, Hercules, CA). Blots were blocked in 5% milk (or 5% bovine serum albumin for the phospho-ERK antibody) in Tris-buffered saline with Tween 20 (TBST) and probed with antibodies, as listed in Table S1.

Flow cytometry for cell cycle analysis

One million human RMS cells were placed in Eppendorf tubes and resuspended in 100 μL of phosphate-buffered saline (PBS) or FACS buffer [PBS (Gibco), 1% bovine serum albumin (BSA, Thermo Fisher Scientific, Waltham, MA) and 0.5 mM EDTA (Thermo Fisher Scientific, 0.5M)] containing flow antibodies for 20 min at 4°C. Cells were then fixed in 4% paraformaldehyde solution (Thermo Fisher Scientific, Waltham, MA, USA) for 20 min at 4°C before staining with DAPI (BD Biosciences, Franklin Lakes, NJ) for 30 min at room temperature.

β-Gal staining

β-Gal staining was performed as previously described.⁵² Imaging was performed using a BioTek Cytation 7 Cell Imaging Multimode Reader (Agilent, Santa Clara, CA, USA).

Quantitative RT-PCR (qPCR)

Human or zebrafish cells were lysed in TRIzol reagent (Thermo Fisher Scientific, Waltham, MA, USA), and RNA was isolated according to the manufacturer’s protocol. Approximately 1 μg of RNA was used for cDNA synthesis using a High-Capacity cDNA Reverse Transcription Kit (Thermo Fisher Scientific, Waltham, MA, USA). SYBR Green-based quantitative PCR was subsequently performed using gene-specific primers (Table S1) in a CFX Connect Real-time PCR Detection System Light Cycler (Bio-Rad, Hercules, CA, USA).

CUT&RUN quantitative PCR

Cells (∼100,000 to 200,000) were harvested, snap-frozen and processed with the reagents in the EpiNext CUT&RUN Fast Kit per the manufacturer’s protocol (Farmingdale, NY, USA). MYC antibody and non-immune IgG (provided in the kit as a negative control) were used (diluted 1:50). SYBR Green-based quantitative PCR was performed using primers against the CDKN1A (p21) promoter (STAR Methods) in a CFX Connect Real-time PCR Detection System Light Cycler (Bio-Rad, Hercules, CA, USA).

Drug treatment studies in the zebrafish RMS model

Liquid nitrogen-preserved primary tumors with rag2:KRASG12D co-expressing mCherry were expanded by transplantation into 3–4 syngeneic adult CG1 fish. Tumor cells were harvested by dissection and dissociation and transplanted subcutaneously into the peritoneal cavity at 25,000 cells per fish. A drug treatment study was performed on engrafted tumor-bearing fish ∼10 days post-transplantation for 7 days. Drugs reconstituted in DMSO [vincristine (0.1 mg/kg, Sigma-Aldrich, St. Louis, MO) and trametinib (10 mg/kg, Selleck Chemicals LLC, Houston, TX)] and vehicle [dimethyl sulfoxide (DMSO), Thermo Fisher Scientific] were delivered on days 0 and 4 in a 5-mL volume using a Hamilton syringe needle via the intraperitoneal route. Imaging of tumor-bearing fish was performed on days 0 and 7 using a Nikon fluorescent dissecting microscope (Nikon Instruments Inc., Melville, NY, USA). The change in tumor volume was quantified using ImageJ software (National Institutes of Health, Bethesda, MD, USA) as previously described.⁵¹

Quantification and statistical analysis

All experiments were performed with at least three independent biological replicates unless otherwise specified. Data are presented as mean ± standard deviation (SD). Statistical analyses were performed using GraphPad Prism version 9.0 (GraphPad Software, San Diego, CA, USA). A p value of less than 0.05 was considered statistically significant.

For comparisons between two groups, unpaired two-tailed Student’s t test was used. For comparisons among three or more groups, one-way analysis of variance (ANOVA) was performed, followed by post hoc pairwise comparisons with Tukey’s multiple comparisons correction. For experiments with two independent variables (e.g., treatment and cell line), two-way ANOVA was performed with Tukey’s multiple comparisons test as appropriate. Specific statistical tests used for each experiment are indicated in the corresponding figure legends.

For in vivo zebrafish studies, sample sizes were determined based on prior experience with the model to achieve adequate statistical power while minimizing animal use. Tumor-bearing fish were randomly assigned to treatment groups. Tumor volume measurements were analyzed using one-way or two-way ANOVA as appropriate, with multiple comparisons correction.

For dose-response curves, IC50 values were calculated using nonlinear regression with a variable slope (four-parameter) model in GraphPad Prism. Comparisons of IC50 values between groups were performed using the extra sum-of-squares F test.

For drug combination studies, synergy was assessed using the Bliss independence model. Bliss synergy scores and contour plots were generated using SynergyFinder (version 2.0, www.synergyfinder.org).

For immunohistochemistry quantification, the percentage of positive cells was determined by analyzing whole tumor cross-sections using QuPath software (version 0.3.2). For matched patient samples, paired t test was used to compare pre-treatment and post-treatment specimens. For immunofluorescence quantification, at least three fields per condition were analyzed, and mean fluorescence intensity or percentage of positive cells was quantified using ImageJ software.