MIR17HG Expression Is Transcriptionally Regulated by PAX3::FOXO1 and MYCN and is Necessary for Oncogenic Activity in Fusion-Positive Rhabdomyosarcoma

Cell Culture and Mycoplasma Testing

The Duchenne muscular dystrophy myoblast cell line (Dbt) and human FP-RMS cell lines were cultured under standard conditions. The sources of the cell lines are as follows: Dbt (RRID:CVCL_C7HJ) - D. Trono, 2004; Rh30 (RRID: CVCL_0041) - ATCC, 1998; Rh30 Low P3F - E. Douglass, 1992; Rh28 (RRID: CVCL_8752) - B. Emanuel, 1992; Rh5 (RRID: CVCL_5917) - J. Khan, 2013 and Rh41 (RRID: CVCL_2176) - C. Linardic, 2014. Rh30, Rh41, Rh28 and Rh30 Low P3F were maintained in Dulbecco’s Modified Eagle Medium (DMEM) (Gibco) supplemented with 10% fetal bovine serum (FBS) (HyClone). Inducible Cas9-expressing Rh30 and Rh41 subclones targeting P3F were previously generated as described (14). Rh5 cells were cultured in RPMI-1640 (Gibco) with 10% FBS, while Dbt cells were maintained in a 1:1 mixture of F-10 (Gibco) and DMEM supplemented with 15% FBS. All cultures were supplemented with penicillin-streptomycin (Gibco). Cell line identity was verified using short tandem repeat genotyping analysis with the AmpFLSTR Profiler Plus PCR Amplification Kit (Applied Biosystems). To rule out Mycoplasma contamination, cell lines were routinely tested using a PCR-based Mycoplasma detection kit (ATCC).

Plasmid Constructs

To generate a constitutive MYCN expression construct, the MYCN open reading frame (ORF) was amplified from a previously cloned retroviral vector and inserted into the lentiviral backbone LeGO-iT (RRID:Addgene_27361). Directional primers were used for amplification, and the final construct was confirmed by Sanger sequencing. The miR-17-92 expression construct, containing the endogenous cluster of 6 miRNAs, was generated by amplifying the intronic region of MIR17HG from genomic DNA and cloning into a lentiviral expression plasmid (pLenti) (RRID:Addgene_65010) containing a blasticidin resistance gene. For inducible expression of P3F and MYCN, we used the doxycycline (Dox)-inducible lentiviral vector pInducer10b (RRID:Addgene_164935). The P3F ORF with a C-terminal 3×FLAG tag and the MYCN ORF with an N-terminal 3×FLAG tag were amplified from previously validated constructs. These inserts were cloned into the pInducer10b backbone using In-Fusion cloning (Takara Bio) and NEBuilder HiFi DNA Assembly (NEB). The resulting constructs, pInducer10b-P3F-3xFLAG-Puro and pInducer10b-3xFLAG-MYCN-BSD, were verified by sequencing.

Design and construction of miRNA-Sponges

miRNA-sponges that functionally inhibit miRNAs of the miR-17-92 cluster were designed using the miRNAsong online tool (https://www.med.muni.cz/histology/miRNAsong/) (15). The binding sites in the sponge were reverse complements of the mature miRNA sequences with 3-4 mismatches outside the seed region. The sponge constructs were synthesized (GenScript) with six bulged binding sites for each targeted miRNA. Binding specificity was validated using the RNAhybrid tool (16). Two sponge sets targeting multiple miRNAs were constructed (17). SET1 targets miR-17, miR-20a and miR-92a and SET2 targets miR-18a, miR-19a and miR-19b1. SET1 and individual sponges for miR-17 and miR-92a were cloned into pInducer10b with a puromycin resistance gene. SET2 and the miR-20a sponge were cloned into the same backbone with the puromycin resistance gene replaced by a blasticidin resistance cassette. All constructs were verified by sequencing.

Protein extraction, SDS-PAGE and Western Blotting

Cells were lysed in 1× SDS lysis buffer (Bio-Rad), followed by sonication using a cycle of 3 seconds ON / 3 seconds OFF on ice for a total of 8 pulses, or until the lysates appeared clear. Protein concentrations were determined using the Pierce BCA Protein Assay Kit (Thermo Scientific) according to the manufacturer’s instructions. It should be noted that β-mercaptoethanol was not added prior to protein estimation and was only included at the time that samples were prepared for gel loading. Unless otherwise stated, 30 μg of total protein per sample were loaded onto pre-cast TGX SDS-PAGE gels (Bio-Rad). These pre-stained gels allowed real-time visualization of protein migration during electrophoresis. Proteins were then transferred to PVDF membranes (Bio-Rad), which were subsequently blocked in 3% non-fat dry milk in Tris-buffered saline for 1 hour at room temperature. Membranes were incubated with antibodies against FOXO1 (1:1,000, Cell Signaling Technology Cat# 2880, RRID:AB_2106495 ), MYCN (B8.4.B) (1:1,000, Santa Cruz Biotechnology Cat# sc-53993, RRID:AB_831602 ), GAPDH (1:2,000 Santa Cruz Biotechnology Cat# sc-25778, RRID:AB_10167668) and Vinculin (1:5,000, Sigma-Aldrich Cat# V9131, RRID:AB_477629 ).

Transfection and Production of Lentiviral Particles

For the generation of lentiviral particles, plasmid DNA transfections were carried out using linear polyethyleneimine (PEI) (Polysciences) in HEK293T cells (RRID:CVCL_0063). A 3:1 PEI to DNA ratio was optimized for high-efficiency transfection in HEK293T. Lentivirus particles were produced in 10 cm² dishes with the following components: 6.0 μg pPax2 (RRID:Addgene_12260), 3.0 μg pMD2 (RRID:Addgene_12259) and 10.0 μg expression plasmid along with 60 μg PEI. Lentiviral transductions were performed in the presence of 1 μg/ml polybrene.

RNA Extraction, cDNA Synthesis and Real-Time PCR

Total RNA was isolated using TRIzol™ Reagent (Invitrogen) according to the manufacturer’s instructions. For mRNA expression analysis, cDNA synthesis was performed using random hexamer primers and Superscript IV reverse transcriptase (Life Technologies). The MIR17HG transcript levels were quantified using exon-exon junction primers from the MIR17HG ORF whereas pri-miR-17-92 levels were measured using primers from MIR17HG intron 3; both of these transcripts were normalized to 18S rRNA. miRNA expression was analyzed via stem-loop reverse transcription (18) followed by quantitative PCR (qPCR) using miRNA-specific forward and stem-loop specific universal primers. qPCR analysis of miRNA was performed using SYBR Green chemistry (Applied Biosciences), and data were quantified by 2^{^-ΔΔCT} methods (18). miRNA expression was normalized to RNU6 snRNA. All qPCR reactions were performed with 3 technical replicates. Primer sequences are listed in Table S1.

Targeted genomic deletion

Targeted deletions were generated using a dual-guide CRISPR/Cas9 approach in which two sgRNAs flanking the region of interest were cloned into a Cas9-GFP expression plasmid. Following transfection, GFP-positive cells were isolated by fluorescence-activated cell sorting and expanded as single-cell clones. Clones were screened by PCR and deletions were confirmed by Sanger sequencing. sgRNA and primer sequences are provided in Table S1.

Chromatin Immunoprecipitation (ChIP) Analysis

Cells were washed with cold PBS and crosslinked with formaldehyde, followed by quenching with glycine. Chromatin was sheared by sonication (30% amplitude; 10 cycles of 30s ON/30s OFF). The sheared chromatin was immunoprecipitated using either a C-terminal FOXO1-specific antibody (to capture P3F) or IgG as a negative control. Following immunoprecipitation, enrichment of target binding regions was assessed by endpoint PCR. Input DNA was included as a positive control. PCR products were separated on a 2% agarose gel containing GelRed (Biotium) and visualized using a gel imaging system (Bio-Rad).

Cell Growth Analysis

Cell proliferation assays were conducted using the Incucyte live-cell imaging system (Thermo Fisher: Sartorius: IncuCyte S3 Live Cell Analysis System [RRID:SCR_023147]). Cells were seeded in appropriate densities in 24-well plates, and images were acquired every 6 hours over a period of 6 days (19). Confluence was calculated using Incucyte’s integrated software based on phase-contrast imaging. Growth curves were generated from confluence measurements to assess proliferation dynamics following various genetic perturbations. All experiments were performed with three wells for each condition and repeated in at least three independent experiments for validation. For all cell growth experiments involving inducible expression constructs, expression was induced in the presence of 500 ng/ml Dox (Sigma-Aldrich) (20).

Clonogenic and Focus Formation Assays

The clonogenic assay was performed to assess colony growth by seeding 400 cells in 6 cm dishes in conditioned medium, which was prepared by growing cells for 48 hours, collecting and filtering the medium through a 0.22 μm filter, and mixing 1:1 with fresh medium. After 20 days, the medium was removed and then cells were fixed with methanol for 10 minutes and stained with Giemsa (Sigma-Aldrich) for 10 minutes. Plates were imaged using Image Lab Software (RRID:SCR_014210). The focus formation assay of oncogenic transformation was performed by seeding 3.5 × 10⁵ NIH3T3 cells per plate along with 400 test cells. After 3 weeks, cells were fixed and stained with Giemsa, and the number and area of foci were quantified using ImageJ (RRID:SCR_003070) software (21). Both assays were conducted in triplicate with empty vector as control. For all clonogenic and focus formation experiments involving inducible constructs, expression was induced in the presence of 500 ng/ml Dox.

miRNA Sensor Assay

To evaluate the efficacy of the miRNA-sponges, we used the miR-sensor assay (18) in pmirGLO vector (RRID:Addgene_49379). Three binding sites with perfect complementarity for miR-17a and miR-92a were cloned into the 3′ UTR of the luciferase gene. Sponge-expressing cells were transfected with the sensor constructs using the Fugene reagent (QIAGEN). After 48 hours, luciferase expression was detected by western blot.

Statistical Analysis

Experiments were performed in triplicate, and data are presented as means ± standard deviation of the means. Differences between test and control groups were analyzed using GraphPad Prism (RRID:SCR_002798)). An unpaired t-test was used to compare two samples, while one-way analysis of variance (ANOVA) with multiple comparisons was applied when three or more samples were involved in a single experiment. P-values were calculated and are shown in the figures as described in the figure legends.

Data Availability Statement

All data supporting the findings of this study are provided within the article and its Supplementary Materials. RNA sequencing data from the dataset reported by Hoffman et al (manuscript submitted) have been deposited in dbGaP under accession number phs004381.v1.p1. Publicly available ChIP-seq, Hi-C, and RNA-seq datasets analyzed in this study were obtained from the Gene Expression Omnibus (GEO) and dbGaP repositories under accession numbers GSE83728 and GSE163068. All other data generated during this study are available from the corresponding author upon reasonable request.

P3F and MYCN Influence Expression of the miR-17-92 Cluster

Given that MYCN binds to sites near the MIR17HG promoter and transcriptionally activates MIR17HG in neuroblastoma (22) and medulloblastoma (23), we examined the expression levels of the miR-17-92 cluster across a panel of FP-RMS lines with variable expression of P3F and MYCN. We compared miR-17-92 expression in these FP-RMS lines to an immortalized human myoblast line (Dbt) and a spontaneous non-transformed subclone of Rh30 cells with very low P3F levels (Rh30 Low P3F). The FP-RMS cell lines (Rh5, Rh28, Rh30 and Rh41) with moderate to high expression of both P3F and MYCN, as confirmed by western blot analysis (Figure 1A), showed high expression of the miR-17-92 cluster (Figure 1B). In contrast, the Dbt and Rh30 Low P3F lines, which exhibit low expression of both P3F and MYCN, demonstrated markedly reduced levels of miR-17-92. These findings support the premise that the upregulation of miR-17-92 may be related to the presence of either P3F, MYCN or both.

To investigate the role of P3F in regulating miR-17-92 expression, we utilized previously established Rh30 and Rh41 subclones that harbor a Dox-inducible Cas9 system targeting P3F (14). CRISPR-Cas9-mediated KO of P3F was induced by Dox treatment (2000 ng/ml), and efficient loss of P3F protein was confirmed by western blot analysis in Rh30 (Figure 1C) and Rh41 (Figure S1A). P3F depletion resulted in a marked reduction in MYCN protein levels, consistent with MYCN being a downstream transcriptional target of P3F (13). qPCR analysis demonstrated that P3F KO led to a significant reduction in MIR17HG transcript and pri-miR-17-92 levels in both Rh30 (Figure S1B) and Rh41 (Figure S1C) cells. Correspondingly, the expression of mature miR-17-92 miRNAs was also markedly decreased in Rh30 (Figure 1D) and Rh41 (Figure S1D).

Given that MYCN is a known regulator of MIR17HG in neuroblastoma and medulloblastoma, we sought to determine whether MYCN is necessary for the miR-17-92 expression in these FP-RMS cells. To dissect the contribution of MYCN, we performed targeted MYCN KO as well as targeted P3F KO using CRISPR-Cas9 in Rh30 and Rh41 cells. Western blotting confirmed efficient depletion of each protein in Rh30 (Figure 1E) and Rh41 (Figure S2A) and confirmed that MYCN expression is dependent on P3F in these FP-RMS cells with P3F KO. To systematically evaluate the impact on the miR-17-92 locus, we measured expression of MIR17HG transcript, pri-miR-17-92, and mature miRNAs. qPCR showed that loss of either P3F or MYCN significantly reduced mature miR-17-92 miRNA expression in Rh30 (Figure 1F) and Rh41 (Figure S2B), consistent with results obtained from the inducible Cas9 KO. Correspondingly, MIR17HG and pri-miR-17-92 transcript levels were also markedly decreased following P3F or MYCN depletion in both Rh30 and Rh41 (Figure S2C, S2D). These results indicate that MYCN is necessary for high-level miR-17-92 expression, and that the effect of P3F at least partially depends on its ability to upregulate MYCN.

P3F and MYCN Cooperatively Induce miR-17-92 Expression

To define the individual and cooperative contributions of P3F and MYCN to the regulation of miR-17-92 expression, we introduced inducible expression constructs for P3F and/or MYCN into two independent models: immortalized Dbt human myoblasts and Rh30 Low P3F cells. In our human myoblast system with inducible P3F expression (Dbt-iP3F) (24), P3F expression is controlled by a Dox-inducible promoter and is not sufficient to upregulate high-level MYCN expression in these cells (Figure 2A). We transduced Dbt control cells or Dbt-iP3F cells with a lentiviral construct that constitutively expresses MYCN to generate cells expressing MYCN alone or P3F and MYCN. qPCR analysis revealed that P3F alone or MYCN alone elicited only a modest increase in miR-17-92 expression (Figure 2B). In contrast, co-expression of MYCN and P3F led to a marked elevation of miR-17-92 levels. These results suggest a synergistic interaction between MYCN and P3F in activating miR-17-92 expression.

To validate this finding in a second model system related to FP-RMS, we used the Rh30 Low P3F cells to establish two Dox-inducible lines, one for MYCN (iMYCN) and another for P3F (iP3F). Upon Dox treatment, western blot confirmed robust induction of each transgene (Figure 2C). Interestingly, MYCN induction alone led to an increase in endogenous P3F expression, reaching levels similar to those observed in Rh30 FP RMS cells. In qPCR studies of miR-17-92 expression, induction of P3F alone did not upregulate miR-17-92 expression whereas MYCN induction (which also induces P3F expression) upregulated expression of the miR-17-92 cluster (Figure 2D). To dissect the requirement for P3F in MYCN-driven miR-17-92 expression, we employed a CRISPR/Cas9 approach to knockout P3F in the iMYCN cells (Figure 2C). In contrast to the significant miR-17-92 activation in the iMYCN cells (in which both MYCN and P3F are induced), MYCN induction in the iMYCN/CR-P3F cells failed to upregulate miR-17-92 (Figure 2D). These gain-of-function studies support a model in which P3F and MYCN act cooperatively and are sufficient to drive high level miR-17-92 expression in FP-RMS.

Inhibition of the miR-17-92 Cluster Suppresses Oncogenic Properties in FP-RMS

To investigate the functional importance of the miR-17-92 cluster in FP-RMS, we employed a miRNA-sponge-based approach (25) to inhibit the activity of mature miRNAs derived from the MIR17HG locus. These synthetic sponge constructs contain tandem binding sites with partial complementarity to specific miRNAs, enabling them to sequester endogenous miRNAs and relieve repression of target mRNAs (26) (Figure S3A). For these experiments, we divided the six miRNAs of the miR-17-92 cluster into two sets, each designed to target three miRNAs. SET1 targets miR-17a, miR-20a and miR-92a whereas SET2 targets miR-18a, miR-19a and miR-19b-1. In addition, we generated individual sponges for miR-17a, miR-20a and miR-92a to further dissect the contributions of these miRNAs (Figure S3B). Of note, all these sponge constructs were cloned into Dox-inducible expression vectors.

We validated the functional efficacy of two miRNA-sponge constructs using a luciferase-based miRNA sensor assay (18). Sensor constructs containing three tandem binding sites with perfect complementarity to either miR-17a or miR-92a were cloned into the 3′ UTR of a luciferase reporter gene (Figure S3C). These constructs were transfected into Rh30 cells stably expressing the corresponding miRNA-sponge or control (empty vector). In empty vector (EV) controls, endogenous miRNAs bind to the sensors and suppress luciferase expression. In contrast, the miRNA-sponges sequester the corresponding miRNAs and relieves this suppression, resulting in increased luciferase protein detected by western blot analysis (Figure S3D). These findings verify that the sponges effectively inhibit miRNA activity and support their utility for downstream functional studies. To determine whether miRNA sponges affect miRNA expression, quanitification of endogenous miRNA levels in Rh30 and Rh41 revealed that levels of targeted miRNAs are unchanged or reduced, but never increased, following sponge expression (Figure S4A and S4B). These findings indicate that, although miRNA sponges effectively inhibit miRNA activity, their effects on steady-state miRNA levels are consistent with miRNA sequestration and/or miRNA degradation (27).

To assess whether miR-17-92 activity is essential for oncogenic transformation, we employed the Dbt-MYCN-iP3F model, in which MYCN is constitutively expressed and P3F is Dox-inducible, and we transduced Dox-inducible miRNA-sponge constructs into these cells. Upon Dox induction of P3F, EV-transduced cells formed dense colonies and foci, indicative of oncogenic transformation (Figure 3A, 3B) In contrast, cells with Dox induction of both P3F and either SET1 or individual sponges targeting miR-17a, miR-20a, or miR-92a showed markedly reduced transformation, as evidenced by fewer and smaller colonies and foci. In the clonogenic assay of cells transduced with miRNA-sponge constructs, the finding of decreased colony formation without Dox induction is attributed to leaky expression of the miRNA-sponges. Notably, cells transduced with the SET2 sponge construct, which targets miR-18a, miR-19a and miR-19b-1, formed dense colonies and foci upon Dox induction, similar to EV controls (Figure 3C, 3D), indicating that miR-18a, miR-19a, and miR-19b-1 have little to no role in suppressing oncogenic transformation under these conditions. Live-cell imaging further demonstrated that Dox induction of sponges targeting SET1, miR-17a, and miR-20a significantly impaired proliferation, while the miR-92a sponge had a more modest effect (Figure 3E). These results indicate that miR-17a, miR-20a, and miR-92a are the principal effectors of this miRNA cluster that drive oncogenic phenotypes in this myoblast model system.

To determine whether this requirement for miR-17-92 extends to FP-RMS cell lines, we introduced SET1, SET2, individual miRNA-sponges or empty vector into Rh30 and Rh41 cells. Live-cell imaging of Dox-treated cells revealed distinct responses: in Rh30 cells, SET1 and SET2 sponges as well as the three individual sponges (from SET1) similarly suppressed proliferation, indicating a broad dependency on the miR-17-92 cluster (Figure 4A). In Rh41 cells, both SET1 and SET2 decreased proliferation, and for the three individual sponges from SET1, proliferation was selectively reduced by inhibition of miR-20a and miR-92a, but not miR-17a (Figure 4D). These findings point to the presence of variable cell line-specific dependencies on individual miRNAs within the cluster.

Clonogenic and focus formation assays further supported these observations. In Rh30 cells, expression of SET1, its individual components or SET2 reduced colony and focus formation (Figure 4B, 4C). In Rh41 cells, similar reductions in colony and focus formation were observed with SET1 and sponges for miR-20a and miR-92a, while the sponges for miR-17a and SET2 only modestly impaired clonogenic growth, again indicating context-dependent roles for specific miRNAs (Figure 4E, 4F). The colonies and foci were counted and quantified in three independent biological replicates for each condition (Figure S5A, S5B, S5C and S5D). The distinct clonogenic responses underscore the differential dependence of Rh30 and Rh41 cells on the component miRNAs of the miR-17-92 cluster.

miR-17-92 Cluster is not Sufficient for Oncogenic Transformation

Using our human myoblast system, we next examined whether the miR-17-92 cluster alone can drive oncogenic transformation. A lentiviral expression construct was used to express the miR-17-92 cluster in Dbt myoblasts. Despite high expression of the pri-miR-17-92 compared to Rh30 FP-RMS cells (Figure S6A), these miRNAs did not enhance proliferation (Figure 5A); in fact, a modest reduction in cell growth was observed compared to empty vector-transduced control cells. Moreover, the Dbt myoblasts expressing the miR-17-92 cluster did not show an increase in the number or size of the colonies in a clonogenic assay (Figure 5B) and failed to exhibit focus formation, (Figure 5C). These findings indicate that miR-17-92 expression alone is insufficient to initiate oncogenic transformation.

Next, we assessed whether miR-17-92 could cooperate with the P3F oncoprotein in Dbt myoblasts with Dox-induced P3F expression. Transduction of the miR-17-92 construct resulted in high-level expression of pri-miR-17-92 and mature miR-17-92 miRNAs (Figure S6B, S6C) and modestly enhanced proliferation (Figure 5D) and clonogenic colony formation (Figure 5E) but did not induce focus formation (Figure 5F). These findings indicate that, while miR-17-92 is necessary for transformation when inhibited and can augment certain growth-related features when added, this miRNA cluster can neither substitute for a primary transformation signal nor can it function similar to MYCN in collaborating with P3F to transform cells. These findings underscore the requirement for additional oncogenic signals or cooperating genetic events to achieve full oncogenic transformation.

A Distal P3F DNA-Binding Element Regulates MIR17HG Expression

To investigate the transcriptional regulation of MIR17HG by P3F in FP-RMS, we analyzed P3F chromatin immunoprecipitation sequencing (ChIP-seq) datasets from Rh4 FP-RMS cells (28) and Dbt-iP3F myoblasts (29). These analyses revealed a prominent P3F-bound DNA element in Rh4 cells located approximately 1.84 Mb upstream of the MIR17HG locus (Figure 6A, S7A). This same site was also detected 24 hours following Dox-induced P3F expression in Dbt-iP3F myoblasts. Consistent with prior studies showing that P3F frequently regulates transcription through distal intergenic enhancers, this region overlapped with a MED1 co-activator binding peak and exhibited strong H3K27ac enrichment (Figure S7B); these findings suggest the presence of active enhancer-associated chromatin (28-31). Notably, no other recurrent and consistent P3F-, MED1-, and H3K27ac-enriched sites were detected between this distal element and the MIR17HG locus in both cell lines (30). A similar study of MYCN ChIP-seq data revealed a MYCN binding site in the proximal 5’ region of the MIR17HG locus (Figure S7B) (28).

To assess the functional significance of this distal P3F binding site, we utilized a CRISPR/Cas9-mediated strategy and deleted a 550 bp region surrounding this binding site in Dbt-MYCN-iP3F cells. This deletion was identified in two subclones (C2 and C3) (Figure 6B) and was validated by Sanger sequencing of PCR products (Figure S7C). Importantly, P3F and MYCN protein levels remained unchanged in the deleted clones, as assessed by western blotting (Figure 6C), indicating that downstream effects were not due to altered expression of these oncogenic transcription factors. Expression analysis by qPCR demonstrated that P3F cannot upregulate MIR17HG, pri-miR-17-92 and mature miR-17-92 miRNA levels in the deleted subclones in contrast to the induction of these RNAs seen in the parental and undeleted subclones (Figure 6D, S7D). These findings suggest that this distal P3F-bound region is critical for transcriptional activation of the MIR17HG locus. Of note, analysis of published Rh4 Hi-C datasets (30) did not reveal evidence of a chromatin looping interaction between this P3F-bound element and the MIR17HG locus, suggesting that regulation may occur through an indirect chromatin mechanism rather than a direct loop (Figure S8A).

To confirm loss of P3F occupancy, we performed ChIP with a C-terminal FOXO1 antibody in parental and C2-KO cells cultured under Dox- or Dox+ conditions. We then used PCR to assay binding to a 410 bp region surrounding the P3F binding site in parental Dbt-MYCN-iP3F cells and a 440 bp region surrounding the deletion breakpoint in C2-KO cells (Figure S8B and S8C, Table S1). The target fragments were only amplified by PCR in input samples and in FOXO1 ChIP from parental Dox+ cells, whereas PCR amplification of the target fragments was not observed in IgG controls, parental Dox- cells, or deletion clones with or without Dox. These studies corroborate P3F binding to this region in parental Dox+ cells and loss of P3F binding following deletion of the P3F-binding region.

At the phenotypic level, the deleted subclones showed diminished focus-forming ability (Figure 6E), indicating loss of oncogenic transformation. However, the deleted subclones retained the capacity for cell growth, as evidenced by the presence of colonies in a clonogenic assay (Figure S9). In contrast to the transformed Dox-treated control cells, cells from the deleted subclones do not form dense, multilayered colonies and instead grow as flat sheet-like collections. These findings suggest that while this P3F regulatory element is dispensable for basal cell growth, it is required for transformation-associated phenotypes. We postulate that the residual MYCN-stimulated MIR17HG expression in the deleted subclones is sufficient for basal cell growth.

To determine whether this P3F-binding site regulates other nearby transcripts, we analyzed RNA-seq data (Hoffman et al, manuscript submitted) from parental Dbt-MYCN-iP3F cells with and without Dox for expression of genes (including long intergenic non-coding RNAs, miRNAs and protein-coding genes) within a 4 Mb region around this P3F binding site (Figure S7A). Surprisingly, none of these genes showed detectable expression in the RNA-seq dataset, except for MIR17HG (Table S2). The complete absence of expression of neighboring loci supports the specificity of this regulatory element in controlling MIR17HG transcription. These findings provide compelling evidence that the identified distal P3F-binding site serves as a regulatory element critical for MIR17HG expression in FP-RMS.