ASAP1 and ARF1 Regulate Myogenic Differentiation in Rhabdomyosarcoma by Modulating TAZ Activity

Rhabdomyosarcoma (RMS) is the most common pediatric soft-tissue sarcoma, with approximately 350 cases annually in the United States (1). It is genetically classified by fusion status. Fusion-positive RMS (FP-RMS), which frequently results from a fusion between PAX3/7 and FOXO1, has a poorer prognosis than fusion-negative RMS (FN-RMS), which is commonly driven by alterations in the receptor tyrosine kinase/RAS/MAPK (RTK/RAS/MAPK) pathway (2, 3). Molecular markers such as TP53 alteration and MYOD1^L122R mutations also affect outcome (4, 5). FN-RMS, associated with embryonal histology, accounts for approximately 70% of RMS cases and has a 5-year survival rate of approximately 70% to 90% (6–8). However, the 5-year survival rate drops to 10% to 30% upon metastasis (9). Patients with metastatic disease undergo aggressive, multimodal therapeutic approaches that are unfortunately associated with high toxicity and morbidity. Despite this aggressive therapeutic strategy, survival for these patients has not improved in more than 40 years (10). A better understanding of the mechanisms regulating FN-RMS progression is necessary to identify novel therapeutic strategies for these patients.

One strategy being actively explored is differentiation therapy. RMS resembles developing skeletal muscle and is partly characterized by a defect in skeletal muscle precursor cell differentiation (11–13). In FN-RMS, despite the expression of the key transcription factor regulating myogenic differentiation MYOD1, cells fail to differentiate past the myoblast stage. We showed that in RAS-mutant FN-RMS, ERK1/2 activates pro-proliferative RAS-dependent superenhancers and stalls transcription of MYOG, the gene that transcribes the myogenic transcription factor myogenin (14). Suppression of oncogenic RAS signaling through inhibition of MEK1/2 induced differentiation and slowed xenograft tumor growth in models of RAS-mutant FN-RMS (14). However, the tumors quickly gained resistance to single-agent MEK inhibition (MEKi).

Dual targeting of the RTK/RAS/MAPK pathway improves efficacy and delays resistance (14–18). We have shown that combining MEK1/2 inhibitors with inhibition of the upstream RTK and insulin-like growth factor receptor 1 (IGF1R) effectively enhances response and duration of response in a subset of RAS-mutant FN-RMS models (14, 16). However, opportunities for translation are limited by the availability of the IGF1R-targeting agent. Additionally, the effect of this combination on differentiation has yet to be discovered. A better understanding of the proteins and pathways regulating the differentiation defects caused by oncogenic RAS signaling is needed to discover promising combination therapies for patients with RMS with RAS alterations. In this study, we investigate the role of ADP ribosylation factor (ARF) GTPase–activating protein (GAP) with SH3 domain, ankyrin repeat, and PH domain (ASAP1), an ADP ribosylation factor GAP (ARF GAP; ref. 19), in FN-RMS differentiation to better understand the mechanisms controlling FN-RMS.

ASAP1 is a multidomain protein composed of a BAR domain, a PH domain, an ARF GAP domain, ankyrin repeats, E/DLPPKP repeats, a proline-rich domain, and an SH3 domain. ASAP1 is involved in several biological processes, including differentiation (20, 21), proliferation (22, 23), and cytoskeletal organization (24–31). Its expression is upregulated in numerous cancers and is associated with progression (22, 32–38). ASAP1 is a known YAP1/TAZ target gene (39), regulates β1-integrin recycling (34), and binds to focal adhesion kinase (FAK; ref. 26). These factors regulate the proliferation/differentiation switch in myoblasts and are dysregulated in RMS (40–42). Cytoskeletal organization, a well-characterized function of ASAP1, is necessary for metastasis in a mouse model of FN-RMS (43, 44).

In this study, we aim to elucidate ASAP1’s contribution to FN-RMS differentiation to understand better the mechanisms controlling FN-RMS. We identify a paradoxical role for ASAP1 as a protumorigenic, prometastatic protein that is necessary for trametinib-induced differentiation in FN-RMS. We show that components of the ARF GTPase pathway regulate myogenic differentiation by modulating the activity of WW domain–containing transcriptional regulator 1 (WWTR1, referred to by its common nomenclature, TAZ, from here on), a transcriptional co-activator and homolog of yes-associated protein 1 (YAP1) that has been implicated in FP-RMS tumorigenesis (45). This study improves our understanding of differentiation in RAS-mutant FN-RMS and provides a rationale for exploring RAS pathway and YAP1/TAZ inhibitor combinations in FN-RMS.

C2C12 mouse myoblast cells (RRID: CVCL_0188) were obtained from the ATCC; RD (female, RRID: CVCL_1649), SMS-CTR (male, RRID: CVCL_A770), and JR1 (female, RRID: CVCL_J063) cells were obtained from J. Khan (NCI). Cells were maintained in DMEM (Quality Biological, # 112-013-101) supplemented with 10% FBS, 1× penicillin/streptomycin, and 1× glutamine, and they were cultured at 37°C in 5% CO₂. All cell lines were confirmed to be Mycoplasma negative using the MycoAlert kit (Lonza, # LT07-418), and their identity was confirmed by short tandem repeat fingerprinting (Genetica/Lapcorp) before experimental use. Cell lines are maintained in culture for fewer than 15 passages, approximately 6 weeks.

Transient transfections were performed using DharmaFECT 1 (1:2 ratio, Horizon Discovery, # T-2001-03) or FuGENE HD transfection (1:2 ratio, Promega Corporation, # E2311) reagents with 50 ng siRNA or 2 μg plasmid, respectively, in Opti-MEM Reduced Serum Media (Thermo Fisher Scientific, # 31985070) supplemented with 5% FBS.

For knockdown experiments, cells were plated on day 0, siRNA was transfected on day 1, and differentiation was initiated on day 2 (72-hour trametinib treatment for FN-RMS cell lines; 96-hour 2% horse serum treatment for C2C12 cells, detailed in “Differentiation assays”). For ARF1 rescue experiments, cells were plated on day 0, siRNA was transfected on day 1, rescue vectors were transfected on day 3, and differentiation was initiated on day 4 (72-hour treatment with 10 nmol/L trametinib). For ASAP1 rescue experiments, cells were plated on day 0, siRNA was transfected on day 1, rescue vectors were transfected on day 3, and differentiation was initiated on day 5 (48-hour treatment with 10 nmol/L trametinib). These timelines were rigorously tested to optimize effective knockdown throughout the experiment, efficient expression of the rescue vector, and limit off-target toxicity.

Knockdown was verified via immunoblot or qPCR. siRNA oligonucleotides are detailed in Supplementary Table S1. Mouse ASAP1a and ASAP1b used in overexpression experiment were generated by Carissa Grose (Protein Expression Laboratory, Cancer Research Technology Program, Frederick National Laboratory for Cancer Research). ARF1 and ARF1^I46D constructs were described previously (46). Human ASAP1a, ASAP1a^R485K, ASAP1b, and ASAP1b^R485K used in rescue experiment were synthesized by GenScript Biotech Corporation.

C2C12 (RRID: CVCL_0188) cells were switched from complete medium to DMEM containing 2% horse serum [differentiation medium (DM)] 16 to 22 hours following transfection. The cells were incubated in DM for 96 hours. DM was replenished every 48 hours. To induce differentiation of the FN-RMS cell lines, cultures were treated with 10 nmol/L trametinib (Selleck Chemicals, # S2673) or 0.1% vehicle control (DMSO) for 72 hours, without replenishment.

To observe differentiation, after 72 (FN-RMS cells) or 96 (C2C12 cells) hours, cells were fixed with 4% paraformaldehyde, permeabilized with 0.1% Triton X-100, blocked with 1% BSA in PBS, and probed with primary anti–myosin heavy chain 4 (MHC) antibody diluted in blocking buffer overnight. Antibodies are detailed in Supplementary Table S2. Nuclei were visualized with Hoechst 33342 Nuclear Stain (Thermo Fisher Scientific, # 62249). At least five random 10× fields per condition were imaged on a ZEISS Axio Vert.A1 wide-field fluorescent microscope (ZEN software, RRID: SCR_013672) or a Nikon TiE wide-field fluorescent microscope (NIS-Elements software, RRID: SCR_014329). Although investigators were not blinded to the experimental group, random fields were selected in the Hoechst channel to avoid bias associated with MHC staining levels. The differentiation index, defined by the total number of nuclei in MHC-positive cells divided by the total number of nuclei multiplied by 100, was calculated for each field. Briefly, total nuclei were calculated automatically by image thresholding, watershed segmentation, and particle counting on Fiji (RRID: SCR_002285). Code available upon request. Nuclei within MHC-positive cells were identified manually and quantified using Cell Counter in Fiji (RRID: SCR_002285).

Cells were dissociated with 0.5% trypsin with EDTA and pelleted from medium for 5 minutes at 500g at 4°C. The cell pellet was washed with ice-cold PBS. Nuclear and cytoplasmic fractions were separated using NE-PER Nuclear and Cytoplasmic Extraction Reagents (Thermo Fisher Scientific, # 78833) following the manufacturer’s protocol. The following modifications were used: (i) Nuclear pellets were washed once with ice-cold PBS supplemented with protease and phosphatase inhibitors (Cell Signaling Technology, # 5872) and pelleted before lysis; (ii) nuclear lysates were homogenized by sonication rather than vigorous vortexing.

RNA was extracted using an RNeasy kit (Qiagen, # 74104), followed by cDNA synthesis with the Applied Biosystems High-Capacity RNA-to-CDNA Kit (Thermo Fisher Scientific, # 4387406). qPCR was performed using the manufacturer’s protocol for TaqMan Gene Expression Assays (Thermo Fisher Scientific) with TaqMan Fast Advanced Master Mix (Thermo Fisher Scientific, # 4444556). Probes are detailed in Supplementary Table S3.

Whole-cell lysates were extracted using RIPA Lysis and Extraction Buffer (Thermo Fisher Scientific, # J63306.AK), supplemented with 0.4% SDS, 5% glycerol, and 1% Triton-X 100. Protein concentrations were determined using the Pierce BCA Protein Assay Kit (Thermo Fisher Scientific, # 23227). Lysates were separated by gel electrophoresis, transferred to nitrocellulose or polyvinylidene difluoride membranes, and probed with the indicated antibodies.

RD cells transfected with siControl (DharmaFECT) or siASAP1 (DharmaFECT) were differentiated as described above. Following the manufacturer’s protocol, RNA was extracted from the cells at 48 or 72 hours using an RNeasy kit (Qiagen, # 74104). The RNA concentration was determined using NanoDrop, and quality control was performed using the Agilent 2200 TapeStation. mRNA was purified, cDNA synthesized, and sequencing libraries prepared using the SureSelect XT HS2 mRNA Library Preparation Kit (Agilent). Libraries were pooled and sequenced on a NextSeq 500 System (Illumina) with paired-end reads of 2 × 75 bp. Reads were normalized, aligned, and mapped using Partek Genomics Suite software (RRID: SCR_011860). Differentially expressed genes were determined using the DESeq2 package (RRID: SCR_015687) in RStudio (RRID: SCR_000432). Differentially enriched gene sets were determined from differentially expressed genes using GSEA software (RRID: SCR_003199; refs. 47, 48). Human skeletal muscle myoblast differentiation gene set is described in (14). YAP1/TAZ transcriptional activity gene set is described in (49). Gene Ontology (biological processes), Kyoto Encyclopedia of Genes and Genomes (RRID: SCR_012773) pathway, and HALLMARK gene sets were interrogated using WEB-based GEne SeT AnaLysis Toolkit (RRID: SCR_006786; ref. 50).

All statistical analyses were performed using GraphPad Prism 8 for macOS Mojave (GraphPad Software, Inc., RRID: SCR_002798). The statistical tests are described in the figure legends.

The data generated in this study are available within the article and its Supplementary Materials (Supplementary Tables S4–S6). Gene expression datasets are available through Gene Expression Omnibus accession number GSE217533.

We previously showed that co-treatment with the IGF1R-targeting antibody ganitumab and the MEK inhibitor trametinib synergistically inhibited FN-RMS cell proliferation and tumor growth in a subset of FN-RMS preclinical models (16). Mechanistically, ganitumab improved response to trametinib by inhibiting feedback activation of the PI3K/protein kinase B (AKT) pathway caused by single-agent MEKi. Activation of AKT is a critical step in myogenic differentiation, inducing expression of the key myogenic transcription factor myogenin (51). However, AKT activation does not induce myogenin expression in FN-RMS cell lines (52). Therefore, it is unclear how IGF1R inhibition would affect RMS differentiation when combined with trametinib. To investigate the effect of IGF1R inhibition on differentiation, we measured differentiation in cells treated with single-agent trametinib or ganitumab or the combination using immunofluorescence (IF) for the terminal differentiation marker MHC or qRT-PCR for the expression of genes associated with myogenic differentiation. Ganitumab treatment alone did not induce differentiation in the RAS-mutant FN-RMS cell lines, SMS-CTR and RD. The addition of ganitumab had no effect on trametinib-induced differentiation (Supplementary Fig. S1) at concentrations known to inhibit MEK and IGF1R in these lines (16). Together, these data indicate that IGF1R inhibition does not affect FN-RMS differentiation either alone or in combination with trametinib. Therefore, we aimed to identify other pathways that might modulate MEKi-induced differentiation in FN-RMS.

RMS is characterized in part by defective myogenic differentiation. FP-RMS tumors express regulatory transcription factors associated with the more differentiated myocyte stage, whereas FN-RMS more closely resembles myoblasts (53). In a dataset of extracranial pediatric solid tumors and normal tissue (54), we found that compared with skeletal muscle and FP-RMS, ASAP1 is overexpressed in FN-RMS, the less-differentiated subtype (Fig. 1A). Because ASAP1 regulates both differentiation (20, 21) and tumorigenesis (22, 55) in different contexts, we investigated the effect of ASAP1 on FN-RMS differentiation. We previously showed that treatment with the MEK1/2 inhibitor trametinib induced differentiation in cell culture models and improved survival in preclinical models of RAS-mutated FN-RMS (14). In three RAS-mutant FN-RMS cell lines treated with trametinib for 72 hours, IF for MHC demonstrated that differentiation was significantly reduced upon transient knockdown of ASAP1 (Fig. 1B; Supplementary Fig. S2A). To determine if this effect on myogenic differentiation was RMS specific, we knocked down ASAP1 in immortalized mouse myoblasts (C2C12). After incubation in DM (2% horse serum) for 96 hours, IF for MHC demonstrated that ASAP1 knockdown significantly reduced myogenic differentiation (Fig. 1C). These data indicate that ASAP1 is necessary for myogenic differentiation in both transformed and nontransformed cells.

The ASAP1 gene encodes multiple splice isoforms, which generally fall into two groups: one that includes an exon encoding the 57-residue proline-rich domain and one that excludes this exon. To determine the impact of ASAP1 overexpression on RMS differentiation, we used representatives of both groups of isoforms, termed ASAP1a (proline-rich domain included) and ASAP1b (proline-rich domain excluded; ref. 24). Overexpression of both splice variants ASAP1a and ASAP1b (Supplementary Fig. S2B) increased differentiation upon trametinib treatment in RD cells (Fig. 1D), demonstrating that ASAP1 positively regulates MEKi-induced differentiation in FN-RMS cells. Neither loss nor overexpression of ASAP1 affected myogenic differentiation without trametinib (Fig. 1B and D, DMSO), indicating that ASAP1 may regulate myogenic differentiation downstream of changes induced by MEKi.

ASAP1 is one of three homologous family members in the ASAP family of ARF GAPs. Therefore, we investigated the impact of homologs ASAP2 and ASAP3 on myogenic differentiation in FN-RMS. Knockdown of ASAP2 and ASAP3 (Supplementary Fig. S2C) reduced MEKi-induced differentiation in RD and SMS-CTR cells to levels comparable with ASAP1 knockdown, showing that all three ASAP homologs affect myogenic differentiation similarly (Fig. 1E).

ASAP1 regulates adipogenic and osteogenic differentiation by activating AKT and FAK signaling (21). Myogenic differentiation depends on the activation of the PI3K/AKT and p38 MAPK signaling pathways, which FAK can activate (51, 52). MEKi-induced differentiation of FN-RMS is also dependent on p38 MAPK activation because FN-RMS cells treated with the combination of trametinib and the p38 MAPK inhibitor SB203580 did not differentiate (Supplementary Fig. S3A). Therefore, we examined the effect of ASAP homolog knockdown on AKT, FAK, and p38 MAPK activity in FN-RMS cells treated with trametinib. However, there were no consistent changes across ASAP homologs or FN-RMS cell lines in the phosphorylation of AKT at S308 or S473, FAK at Y397, or MAPKAPK2, the activating kinase of p38 MAPK upon ASAP homolog loss (Supplementary Fig. S3B and S3C).

We performed RNA sequencing on RD cells that were transfected with control siRNA or ASAP1-targeting siRNA and then treated with DMSO or trametinib to investigate the mechanism by which ASAP1 affects myogenic differentiation. Gene set enrichment analysis (GSEA) showed that several Gene Ontology gene sets associated with myogenic differentiation biological processes were negatively enriched in trametinib-treated cells with ASAP1 knockdown versus trametinib-treated cells with control siRNA [Fig. 2A (first panel)]. ASAP1 knockdown in trametinib-treated cells led to the enrichment of pathways such as the RAS, MAPK, and Hippo pathways and downregulation of a muscle contraction–associated pathway [Fig. 2A (second panel)]. We also found that the HALLMARK gene set for myogenesis was the most negatively enriched of the 50 HALLMARK gene sets on the Molecular Signatures Database [https://www.gsea-msigdb.org/gsea/msigdb/; Fig. 2A (third panel)]. We confirmed the effect of ASAP1 knockdown on MEKi-induced myogenic differentiation by GSEA with an additional gene set (described in ref. 14) and found that genes upregulated during myogenic differentiation of human skeletal muscle myoblasts are not enriched in cells with ASAP1 knockdown (normalized enrichment score = 6.0; P = 0.0; FDR q-value = 0.0; Fig. 2B).

Myogenic differentiation is regulated at the transcriptional level by the sequential expression of myogenic regulatory factors (MRF) and myocyte enhancement factors (MEF; refs. 56, 57). The block in differentiation in RAS-mutated FN-RMS occurs after myoblast determination and expression of the MRF, MYOD1. Activated ERK1/2 translocates to the nucleus, binds POLII, and stalls transcription of MYOG, an MRF that is required for myocyte differentiation (14). Given the significant effect of ASAP1 loss on the expression profile of trametinib-treated cells (Fig. 2A and B), we hypothesized that ASAP1 may affect MRF and MEF levels. MYOG expression was unaffected by ASAP1 loss in RD cells [Fig. 2C (second panel)]. However, unlike control cells, transcript levels of the MEF family member MEF2C were not increased by MEKi in ASAP1-deficient RD cells [Fig. 2C (third panel)]. In the absence of trametinib, ASAP1 loss also resulted in the negative enrichment of gene sets associated with myogenic differentiation (Supplementary Fig. S4A). However, because MYOG, MEF2C, and MYH3 expression levels in DMSO-treated cells are already relatively low (Fig. 2C), it is unclear whether the further reduction caused by ASAP1 loss is biologically relevant.

We next investigated the impact of ASAP2 or ASAP3 loss (Supplementary Fig. S4B) on MYOG, MEF2C, and MYH3 expression by qRT-PCR. In trametinib-treated RD cells, ASAP1, ASAP2, or ASAP3 loss did not affect MYOG expression [Fig. 2D (first panel)]. However, MEF2C and MYH3 expression was suppressed by ASAP1, ASAP2, and ASAP3 loss compared with trametinib-treated control cells [Fig. 2D (second and third panels)]. Immunoblot analysis shows that MYOG protein was also unaffected by ASAP1/2 knockdown in all cell lines, but ASAP3 knockdown reduced MYOG protein in RD and SMS-CTR cell lines. MEF2C protein was reduced by varying amounts in RD, SMS-CTR, and JR1 cells induced to differentiate after knockdown of ASAP1/2/3 (Fig. 2E). Fold change reductions in MYOG, MEF2C, and MYH3 mRNA expression were again observed in DMSO-treated RD cells, but protein expression levels were very low compared with trametinib-treated cells (Fig. 2E; Supplementary Fig. S4C). Together, these data suggest that ASAP1 affects transcriptional regulation of myogenic differentiation downstream of MYOG, though there may be cell line–specific differences.

The ARF GTPases ARF1 and ARF5 are the preferred substrates for the enzymatic activity of the ASAP1 ARF GAP domain (58). We hypothesized that ASAP1 may be regulating differentiation by modulating activity of the ARF signaling pathway. Interestingly, knockdown of ARF1 and ARF5 blocked trametinib-induced differentiation in RD and SMS-CTR cells (Fig. 3A; Supplementary Fig. S5A). Knockdown of ARF6 slightly reduced MEKi-induced differentiation in RD cells (mean, 39.2% vs. 34.2% for siCtrl vs. siARF6, respectively) but did not affect differentiation in SMS-CTR cells (Fig. 3A). By qRT-PCR, knockdown of ARF1, ARF5, or ARF6 (Supplementary Fig. S5B) did not affect MYOG mRNA expression in trametinib-treated RD cells. MEF2C expression was only reduced by ARF1 and ARF5 knockdown. Consistent with observations from IF, MYH3 expression was reduced more by ARF1 or ARF5 knockdown than by ARF6 knockdown (Fig. 3B). In DMSO-treated cells, MYOG and MEF2C transcript levels were not affected by ARF1 or ARF5 loss, but low MYH3 expression was further reduced by ARF1 or ARF5 loss (Supplementary Fig. S5C). Together, these data suggest that ARF1 and ARF5, substrates of ASAP1 GAP activity, are required for trametinib-induced differentiation.

To further characterize the role of ARF signaling in FN-RMS differentiation, we rescued ARF1 knockdown with either wild-type ARF1 or ARF1^I46D, a mutant that is insensitive to ASAP1 GAP activity (46). Both wild-type ARF1 and ARF1^I46D rescued trametinib-induced differentiation in RD cells at 72 hours (Fig. 3C and D; Supplementary Fig. S5D). These results raise the possibility that ASAP1-induced ARF1 cycling is not required for myogenic differentiation or that an ASAP1-independent ARF pathway is being affected. To further investigate this, we rescued ASAP1 knockdown with ASAP1a, a GAP-dead mutant ASAP1a^R485K (corresponding to the mouse R497K mutation), ASAP1b, or ASAP1b^R485K (25). Overexpression of both ASAP1 splice variants and their respective GAP-dead mutants increased expression of MHC at 48 hours (Fig. 3E; trametinib-treated siCtrl group). However, exogenous ASAP1 expression at levels comparable with endogenous expression did not affect differentiation in ASAP1-deficient cells treated with trametinib for 48 hours (Fig. 3E; trametinib-treated siASAP1 group). Therefore, interpretation of the effects of the splice variants or mutation on rescue of trametinib-induced differentiation is not possible.

Because ARF pathway members, including ASAP1, ARF1, and ARF5, are required for FN-RMS differentiation, we hypothesized that other ARF pathways may also regulate FN-RMS differentiation. Therefore, we determined the effect of ARF effectors, GGA1, GGA2, and GGA3, on trametinib-induced differentiation (59). Knockdown of these ARF effectors (Supplementary Fig. S5E) reduced trametinib-induced differentiation to similar levels as ASAP1 in RD and SMS-CTR cells (Fig. 3F), supporting the hypothesis that ARF effectors are required for trametinib-induced differentiation in FN-RMS cells.

Next, we investigated the mechanism by which ASAP1 and ARF1 regulate myogenic differentiation. Given the substantial effect on the expression of genes associated with myogenic differentiation, including the MEF transcription factors, upon ASAP1 knockdown (Fig. 2), we investigated mechanisms regulating transcriptional activity. ASAP1 is a transcriptional target of YAP1/TAZ, homologous transcriptional co-activators regulated by the Hippo pathway (39). YAP1/TAZ are known to promote FN-RMS tumorigenesis and block myogenic differentiation (41, 42), and genes associated with the Hippo pathway are enriched in ASAP1-deficient trametinib-treated RD cells compared with trametinib-treated control cells [Fig. 2A (middle panel)]. Additionally, another ARF GAP, GIT1, represses YAP1/TAZ activity by promoting the activation of LATS, a member of the Hippo pathway core kinase cascade (60). Therefore, we investigated the role of ASAP1 and ARF1 in YAP1/TAZ activity. We found that treatment with trametinib increased the inactivating phosphorylation of TAZ (S89) in RD, SMS-CTR, and JR1 cells (Fig. 4A). However, knockdown of ASAP1, ASAP2, and ASAP3 blocked trametinib-induced TAZ phosphorylation (Fig. 4B). This was also seen upon ARF1 knockdown, although the effect is muted in SMS-CTR cells (Fig. 4C) and exogenous ARF1 or ARF1 (I46D) rescued phosho-TAZ in RD cells (Fig. 4D). Trametinib-induced phosphorylation of TAZ was also abrogated by knockdown of GGA1, GGA2, or GGA3, supporting the hypothesis that the ARF pathway regulates this effect (Fig. 4E). The effect of trametinib treatment on inactivating YAP1 phosphorylation (S127) was inconsistent and not modulated by ASAP1, ARF1, or GGA knockdown.

Phosphorylation at S89 inactivates TAZ by promoting 14-3-3–mediated nuclear export (61). Therefore, we investigated the effect of trametinib treatment and ASAP knockdown on nuclear localization of TAZ by cellular fractionation. Upon trametinib treatment, we found that a large fraction of TAZ is retained in the cytoplasm. However, upon knockdown of ASAP1, the fraction of TAZ retained in the cytoplasm is reduced. This effect is even more pronounced upon ASAP2 and ASAP3 knockdown (Fig. 4F).

By GSEA, the YAP1/TAZ transcription signature was enriched in RD cells treated with trametinib upon ASAP1 knockdown compared with trametinib-treated control cells (normalized enrichment score = 3.03; P = 0.0; FDR q-value = 0.0; Fig. 4G). Additionally, YAP1/TAZ target genes CTGF and CYR61 transcript levels were increased by ASAP1 knockdown in RD cells (Fig. 4H; DMSO, siCtrl vs. siASAP1). Trametinib treatment in cells transfected with control siRNA decreased CTGF and CYR61 expression, consistent with the increase in TAZ S89 phosphorylation (Fig. 4H; siCtrl, DMSO vs. trametinib). However, upon ASAP1 knockdown, CTGF and CYR61 expression was rescued to the level of DMSO-treated siCtrl cells (Fig. 4H; trametinib, siCtrl vs. siASAP1).

Finally, we performed a double knockdown experiment to test the hypothesis that ASAP1 inactivates TAZ to facilitate trametinib-induced differentiation. By IF in RD cells, ASAP1 knockdown decreased differentiation, whereas TAZ (WWTR1) loss did not affect trametinib-induced differentiation. However, TAZ knockdown rescued the block in differentiation caused by ASAP1 loss (Fig. 4I). Together, these data suggest that ASAP1 and ARF1 regulate myogenic differentiation, promote the phosphorylation and cytoplasmic retention of TAZ, and repress YAP1/TAZ transcriptional activity.

In this study, we showed that several different components of the ARF pathway are necessary for MEKi-induced differentiation in FN-RMS. In particular, knockdown of ASAP1, its homologs ASAP2 and ASAP3, the targets of its GAP activity ARF1 and ARF5, and ARF effectors GGA1, GGA2, and GGA3 blocked myogenic differentiation in FN-RMS cells treated with the MEK inhibitor trametinib. Compared with control cells, mRNA and protein expression of MEF2C and MHC was reduced in ASAP1-deficient cells, and genes associated with myogenic differentiation were negatively enriched in ASAP1-deficient cells by GSEA. The loss of ARF1 and ARF5 also blocked induction of MEF2C and MYH3 mRNA and protein expression in MEK inhibitor–treated cells. Mechanistically, this study showed that in RAS-mutant FN-RMS cells, the MEK inhibitor trametinib inactivated TAZ by inducing phosphorylation at S89, but the effect on YAP1 phosphorylation at S127 was inconsistent. However, trametinib did not induce TAZ phosphorylation upon loss of ASAP1/2/3, ARF1, or GGA1/2/3, and targets of YAP1/TAZ transcriptional activity were enriched in ASAP1-deficient cells. Together, our data strongly suggest that ASAP1 and ARF1 regulate myogenic differentiation by promoting the phosphorylation and cytoplasmic retention of TAZ.

Despite being overexpressed in poorly differentiated FN-RMS tumors, our data clearly demonstrate that ASAP1 promotes myogenic differentiation. We previously showed that oncogenic RAS mutations block differentiation by inhibiting MYOG transcription (14). Our work indicates that ASAP1 regulates myogenic differentiation downstream of MYOG expression (Fig. 2C–E), and neither ASAP1 loss nor ASAP1 overexpression affected differentiation in the absence of MEKi (Fig. 1B and D), despite global transcriptional changes with loss of ASAP1 in the absence of MEKi (Supplementary Fig. S4A). Together, these data suggest that the differentiation-inhibiting effects of oncogenic RAS mutations, which are present in up to half of FN-RMS (3, 4), block any differentiation-promoting effects upstream of ASAP1 overexpression.

Inhibition of trametinib-induced differentiation was similar between ASAP homologs despite differences in the degree of knockdown (Fig. 1E). ASAP2 knockdown was not as efficient as ASAP1 and ASAP3 knockdown, and ASAP3 knockdown induced the expression of ASAP1 and ASAP2 (Supplementary Fig. S2C). Because single knockdown of ASAP1, ASAP2, or ASAP3 blocks myogenic differentiation, compensation between homologs is unlikely. Although following the same trend, changes in MEF2C expression (Fig. 2C), TAZ phosphorylation (Fig. 4A), and TAZ localization (Fig. 4C) vary by homolog. Additionally, ASAP3 knockdown (independent of trametinib treatment) was more toxic than ASAP1 or ASAP2 knockdown (reflected in nucleus density in Fig. 1E). Together, these observations justify further investigation into independent roles for each homolog in myogenic differentiation and how they may affect response to differentiation therapy. Nonredundant roles for the ASAP homologs have been identified in other contexts (62, 63).

The finding that both ASAP1 and ARF1 are necessary for myogenic differentiation was unexpected (Figs. 1 and 3A). Cycling between GTP- and GDP-bound states is critical for the regulation of ARF GTPase signaling. However, the ARFs, unlike RAS GTPase proteins, have no intrinsic GTPase activity and are dependent on GAPs for hydrolysis of GTP (64). Although the GAP activity of ASAP1 is generally thought to suppress ARF1 signaling, previous studies also provide evidence that ASAP1 may function as an ARF effector (19, 24–26, 65–67). Our observation that loss of both ARF1 and ASAP1 blocks MEK inhibitor–induced differentiation raises the possibility that ASAP1 may function as an ARF1 effector in RMS differentiation. Our study finds that the loss of the ARF1 effectors GGA1, 2, and 3 also blocked trametinib-induced differentiation to a similar extent as ASAP1 knockdown, which is consistent with activation of the ARF pathway supporting myogenic differentiation (Fig. 3F). Further investigation is needed to determine whether ASAP1 is an ARF effector in RMS differentiation, including determining the preferred ARF substrate for ASAP1 GAP activity in RMS cells.

Rescue of differentiation upon ARF1 knockdown was successful (Fig. 3C and D). However, we were unable to rescue MEKi-induced differentiation upon ASAP1 knockdown despite achieving exogenous protein levels comparable with endogenous protein levels. Although exogenous protein expression was lower in siASAP-transfected cells than in siCtrl-transfected cells, it is unlikely that the 3′ untranslated region-targeting siRNA affected the expression of the exogenous construct (Fig. 3E). Instead, it is possible that exogenous expression of ASAP1 cannot regulate differentiation because of structural or localization differences due to differences in posttranslational modification (24) or mRNA localization (68). Given that overexpression of ASAP1 in cells transfected with control siRNA modestly increased differentiation (Fig. 3C), it is conceivable that endogenous ASAP1 facilitated the function of exogenous ASAP1. In other reports, exogenous ASAP1 successfully rescued cell spreading, circular dorsal ruffle, and cytoskeleton-binding functions, so this may be a context-specific dependency.

The mechanism by which ASAP1 and ARF1 affect TAZ activity is yet to be determined. ASAP1, a YAP1/TAZ transcriptional activity target (39), may control a negative feedback loop to downregulate YAP1/TAZ activity. Similarly, the ARF GAP GIT1 has been shown to negatively regulate YAP1/TAZ transcription by inhibiting SRC-mediated repression of the Hippo pathway core kinase cascade (60, 69). Alternatively, ASAP1 and ARF1 could negatively regulate YAP1/TAZ transcription by directly sequestering TAZ in the cytoplasm, perhaps through endosomal regulation (70). For example, AMOT, a Hippo pathway–independent regulator of YAP1 (71), sequesters YAP1 to endosomes in confluent cells to suppress YAP1 activity and induce contact inhibition of cell growth (72). The changes in TAZ phosphorylation (Fig. 4A and B) suggest an upstream role for ASAP1 and ARF1 in regulating YAP1/TAZ transcription, but it is plausible that phosphorylation occurs secondary to cytoplasmic retention. YAP1/TAZ regulation through biomechanical cues is well established and converges on the actin cytoskeleton (73). It is possible that ASAP1, which regulates focal adhesion dynamics (25–27), F-actin bundling in stress fibers (31, 74), and podosome formation (28, 29), could regulate TAZ through the actin cytoskeleton. However, focal adhesion assembly and f-actin bundling typically promote YAP1/TAZ nuclear localization, whereas tight junction and adherens junction assembly normally suppress YAP1/TAZ transcription (73).

Because MEKi affected TAZ phosphorylation most strongly, our study focused on TAZ. Additional models and conditions will be necessary to determine if ASAP1 and ARF1 affect YAP1. Interestingly, although our work identified an inhibitory role for TAZ in myogenic differentiation, TAZ has previously been shown to bind to the MRF MYOD1 to promote the differentiation of myoblasts (75). However, evidence indicates that YAP1 hyperactivation blocks myogenic differentiation and induces RMS tumorigenesis (41, 42). TAZ has also been shown to contribute to tumorigenesis in FP-RMS (45). These inconsistencies justify a deeper, well-controlled investigation into the specific transcriptomes of YAP1 and TAZ to define their role in myogenic differentiation and RMS tumorigenesis.

In conclusion, our study defines a novel role for ASAP1 and ARF1 in regulating myogenic differentiation. The work outlined here suggests a dependency on both ASAP1 and ARF1 for MEKi-induced RMS differentiation, warranting future investigation into ASAP1 as an ARF1 effector in this cellular context. Additionally, this investigation improves our understanding of myogenic differentiation and identifies YAP1/TAZ activity as a potential target for combination therapy with MEKi in FN-RMS. Many inhibitors of TEAD family transcription factors, which YAP1/TAZ bind, are in clinical development, making this a translatable opportunity for advanced RMS.

K.E. Hebron reports grants from Summer’s Way Foundation and Friends of T.J. Foundation during the conduct of the study. No disclosures were reported by the other authors.

K.E. Hebron: Conceptualization, data curation, formal analysis, supervision, funding acquisition, validation, investigation, visualization, methodology, writing–original draft, writing–review and editing. O.L. Perkins: Data curation, formal analysis, validation, investigation, visualization, writing–review and editing. A. Kim: Data curation, formal analysis, validation, investigation, visualization, writing–review and editing. X. Jian: Data curation, investigation, writing–review and editing. S.A. Girald-Berlingeri: Data curation, investigation. H. Lei: Data curation, formal analysis. J.F. Shern: Resources, data curation, formal analysis, supervision. E.A. Conner: Resources, data curation, formal analysis, supervision. P.A. Randazzo: Conceptualization, resources, supervision, funding acquisition, project administration, writing–review and editing. M.E. Yohe: Conceptualization, resources, data curation, formal analysis, supervision, funding acquisition, visualization, methodology, writing–original draft, project administration, writing–review and editing.

The content of this publication does not necessarily reflect the views or policies of the Department of Health and Human Services nor does mention of trade names, commercial products, or organizations imply endorsement by the US government.

This project has been partially funded by federal funds from the NCI, NIH, under contract number HHSN261201500003I. This project was supported by intramural funding to P.A. Randazzo (ZIA BC007365, supporting P.A. Randazzo, S.A. Girald-Berlingeri, X. Jian, O.L. Perkins, and K.E. Hebron) and M.E. Yohe (ZIA BC011993 supporting M.E. Yohe, O.L. Perkins, K.E. Hebron, and A. Kim). This project was also supported in part by a Young Investigator Award (K.E. Hebron) from the Summer’s Way Foundation and Friends of T.J. Foundation. The authors are grateful to Drs. Corinne Linardic, Deborah Morrison, and Berkley Gryder for their insightful discussions and critical editing of this manuscript. We thank Carissa Grose (Protein Expression Laboratory, Frederick National Laboratory for Cancer Research) for vector cloning assistance and Valery Bliskovsky and Steve Shema (CCR Genomics Core, NCI, Bethesda, MD) for library preparation and sequencing.