Acquisition of an oncogenic fusion protein serves as an initial driving mutation by inducing aneuploidy and overriding proliferative defects

While many solid tumors are defined by the presence of a particular oncogene, the role that this oncogene plays in driving transformation through the acquisition of aneuploidy and overcoming growth arrest are often not known. Further, although aneuploidy is present in many solid tumors, it is not clear whether it is the cause or effect of malignant transformation. The childhood sarcoma, Alveolar Rhabdomyosarcoma (ARMS), is primarily defined by the t(2;13)(q35;q14) translocation, creating the PAX3-FOXO1 fusion protein. It is unclear what role PAX3-FOXO1 plays in the initial stages of tumor development through the acquisition and persistence of aneuploidy. In this study we demonstrate that PAX3-FOXO1 serves as a driver mutation to initiate a cascade of mRNA and miRNA changes that ultimately reprogram proliferating myoblasts to induce the formation of ARMS. We present evidence that cells containing PAX3-FOXO1 have changes in the expression of mRNA and miRNA essential for maintaining proper chromosome number and structure thereby promoting aneuploidy. Further, we demonstrate that the presence of PAX3-FOXO1 alters the expression of growth factor related mRNA and miRNA, thereby overriding aneuploid-dependent growth arrest. Finally, we present evidence that phosphorylation of PAX3-FOXO1 contributes to these changes. This is one of the first studies describing how an oncogene and post-translational modifications drive the development of a tumor through the acquisition and persistence of aneuploidy. This mechanism has implications for other solid tumors where large-scale genomics studies may elucidate how global alterations contribute to tumor phenotypes allowing the development of much needed multi-faceted tumor-specific therapeutic regimens.


INTRODUCTION
Aneuploidy is common in solid tumors, with nearly 90% of such tumors being aneuploid [1].Despite the prevalence of aneuploidy in these cancers, it has not been firmly established if aneuploidy is an early step that drives malignant transformation, results as a consequence of transformation, or what role, if any, oncogenes Priority Research Paper www.impactjournals.com/oncotargetplay in this process.Although many solid tumors are characterized by a specific oncogenic mutation, very few reports examined the role that the resulting oncoprotein plays in driving the initial stages of tumor development, possibly through the acquisition of aneuploidy and in overcoming aneuploidy-dependent growth arrest.In many of these reports, oncogene-induced aneuploidy results from the overexpression of a protein that is central to the chromosome segregation machinery or regulates the process of chromosome segregation [2][3][4][5][6].Although one report demonstrates that an oncogenic transcription factor, TLX1, is sufficient to induce aneuploidy in T-cell progenitors, the process involves a limited mechanism in which specific genes important for maintaining mitotic checkpoint control are affected [7].
Upon the acquisition of aneuploidy, a cell's natural response is to attenuate proliferation.Previous work demonstrated that the induction of aneuploidy in diploid or near diploid cell lines resulted in significant proliferative defects [8,9], raising an apparent contradiction for tumor progression: rapidly proliferating tumor cells are often aneuploid, a state that under normal circumstances is not conducive to proliferation [10].Therefore, tumor cells must develop a mechanism to overcome aneuploiddependent proliferative defects and develop an enhanced proliferation rate.However, the mechanism by which an oncogene promotes aneuploidy in pre-cancerous cells, how these cells overcome the adverse proliferative effects associated with this state, and whether post-translational modifications can regulate this process are important, yet unanswered, questions in the development of solid tumors.
Like other solid tumors, aneuploidy is common in Rhabdomyosarcoma (RMS) [11], which accounts for nearly half of childhood soft tissue sarcomas.RMS is comprised of two main subtypes: embryonal (ERMS) and alveolar (ARMS), each defined by its unique histology, clinical presentation, therapy, and prognosis [12].ARMS, the more aggressive subtype, is primarily defined by the t(2;13)(q35;q14) translocation, which creates the oncogenic fusion protein PAX3-FOXO1 [13,14].In addition to the defining cytogenetic abnormality, ARMS tumor cells contain cytogenetic evidence of polyploidy, having cells ranging in chromosome number from hypodiploid to hypertetraploid, with a wide and heterogeneous range of chromosome number in cells within a single patient sample [11,[15][16][17][18].There were also instances of chromosome amplification as double minutes and heterogeneously staining regions, gains and losses through insertions and deletions, derivative chromosomes and additional translocations, and breakpoints around the centromere [11,[15][16][17][18].Despite extensive work understanding the altered molecular characteristics of PAX3-FOXO1 relative to PAX3 [19][20][21][22][23][24][25] and the knowledge that phosphorylation of the fusion protein contributes to ARMS tumor phenotypes [26], it is not known what role, if any, the fusion protein plays in the promotion of aneuploidy and chromosomal structural abnormalities, overcoming aneuploidy-dependent proliferative defects, and whether phosphorylation of the fusion protein contributes to this process.
In this study we examine how the expression of PAX3-FOXO1 affects global mRNA and miRNA expression.We are the first to show that the presence of PAX3-FOXO1 is sufficient to alter mRNA and miRNA expression, either through direct regulation of genes or indirectly through downstream effects, subsequently altering protein levels important for multiple aspects of chromosome number and structure, thereby promoting the aneuploid state.Further, we are the first to demonstrate that the presence of an oncogene is sufficient to alter, either directly or indirectly, the expression of multiple growth related mRNA and miRNA to overcome aneuploidy-induced proliferative defects.In addition, inhibition of phosphorylation at Ser201 or Ser205 on PAX3-FOXO1 reversed oncogene dependent mRNA and miRNA changes, and inhibited the proliferation of cells containing chromosomal abnormalities.Our results allow us to propose a model by which the expression and phosphorylation of an oncogene, PAX3-FOXO1, serves as the driving molecular event to initiate the development of a solid tumor by reprograming cells to induce aneuploidy and override aneuploidy-induced growth arrest.

PAX3-FOXO1 is sufficient to alter the expression of mRNA and miRNA to promote aneuploidy
To understand how PAX3-FOXO1 affects global mRNA and microRNA (miRNA) expression, we stably transduced passage-matched mouse primary myoblasts with the MSCV-IRES-puromycin retroviral vector (negative control), or the same retroviral vector expressing FLAG-epitope tagged PAX3 (FLAG-PAX3), or FLAG-PAX3-FOXO1 (Figure 1A), a tag previously shown to not affect Pax3 or Pax3-FOXO1 function [24,26].The transduced cells were selected with puromycin; selected cells were harvested from three independent transductions and pooled, resulting in a single mixed population for each individual construct.By utilizing a population of transduced cells, we remove the potential for variability that may occur from clonal effects.The level of PAX3-FOXO1 expression we observed is equivalent to the level of expression of the fusion protein in ARMS tumor cell lines (Supplementary Figure 1 and [27,28]) and is therefore directly relevant to the role of the oncogenic fusion protein in ARMS.This model allows us to use a physiologically relevant cell system to determine how

SIRT2
Deacetylase required for mitotic exit +2.44 Gene name, function and fold change for each gene are listed; downregulated genes indicated by negative numbers, upregulated genes indicated by positive numbers.*Comparisons are made between empty vector and PAX3-FOXO1 (V vs. PF), PAX3 and PAX3-FOXO1 (P3 vs. PF) or empty vector and PAX3 (V vs. P3), ^indicates genes demonstrated to have PAX3-FOXO1 binding sites present in their proximal promoter [29], and #indicates genes that demonstrate similar trends in gene expression changes in human tumor samples [30][31][32][33].

TUBB4B
Major component of microtubules -2.07 Gene name, function and fold change for each gene are listed; downregulated genes indicated by negative numbers, upregulated genes indicated by positive numbers.*Comparisons are made between empty vector and PAX3-FOXO1 (V vs. PF), PAX3 and PAX3-FOXO1 (P3 vs. PF) or empty vector and PAX3 (V vs. P3), ^indicates genes demonstrated to have PAX3-FOXO1 binding sites present in their proximal promoter [29], and #indicates genes that demonstrate similar trends in gene expression changes in human tumor samples [30][31][32][33].Gene name, function and fold change for each gene are listed; downregulated genes indicated by negative numbers, upregulated genes indicated by positive numbers.*Comparisons are made between empty vector and PAX3-FOXO1 (V vs. PF), PAX3 and PAX3-FOXO1 (P3 vs. PF) or empty vector and PAX3 (V vs. P3), ^indicates genes demonstrated to have PAX3-FOXO1 binding sites present in their proximal promoter [29], and # indicates genes that demonstrate similar trends in gene expression changes in human tumor samples [30][31][32][33].Gene name, function and fold change for each gene are listed; downregulated genes indicated by negative numbers, upregulated genes indicated by positive numbers.*Comparisons are made between empty vector and PAX3-FOXO1 (V vs. PF), PAX3 and PAX3-FOXO1 (P3 vs. PF) or empty vector and PAX3 (V vs. P3), ^indicates genes demonstrated to have PAX3-FOXO1 binding sites present in their proximal promoter [29], and # indicates genes that demonstrate similar trends in gene expression changes in human tumor samples [30][31][32][33].Gene name, function and fold change for each gene are listed; downregulated genes indicated by negative numbers, upregulated genes indicated by positive numbers.*Comparisons are made between empty vector and PAX3-FOXO1 (V vs. PF), PAX3 and PAX3-FOXO1 (P3 vs. PF) or empty vector and PAX3 (V vs. P3), ^indicates genes demonstrated to have PAX3-FOXO1 binding sites present in their proximal promoter [29], and # indicates genes that demonstrate similar trends in gene expression changes in human tumor samples [30][31][32][33].
the sole expression of the oncogenic PAX3-FOXO1 can initiate a cascade of events, both direct and downstream, over a series of proliferation events, to contribute to the initial stages of tumor development with respect to aneuploidy and overcoming aneuploidy-dependent proliferative defects.This model is in direct contrast to tumor cell lines or primary tumor samples in which it is difficult to determine whether these processes result as a "byproduct" of the final oncogenic state.We performed mRNA and miRNA deep sequencing on total RNA isolated from three independent growths of stably transduced primary myoblasts and utilized the resulting sequencing data in large-scale comparative transcriptomic analyses, using the on-line Galaxy program or miRNAKey, respectively (see Materials and Methods).Since all stable lines were generated simultaneously and with identical methods, any artifactual effects of the process on mRNA and miRNA expression would be expected to be present in all samples and would therefore by corrected for by our differential analysis.The data  Comparisons are made between empty vector and PAX3-FOXO1 (V vs. PF), PAX3 and PAX3-FOXO1 (P3 vs. PF) or empty vector and PAX3 (V vs. P3) and the *indicates miRNA that target genes involved with the promotion of aneuploidy and in proliferative control.www.impactjournals.com/oncotargetused for subsequent studies were limited to 1) mRNA or miRNA displaying statistically significant differences (p < 0.05), 2) mRNA or miRNA present in both data sets being analyzed to rule out artifactual differences resulting from depth of read, and 3) mRNA or miRNA that exhibited at least 2-fold difference in expression either up-or downregulated.Finally, the biological functions of the differentially expressed mRNA or target genes of miRNA were categorized using the Onto-Express gene ontology software (http://vortex.cs.wayne.edu/projects.htm).We found that in total PAX3-FOXO1 altered the expression of 846 genes, with 562 being upregulated and 284 being downregulated, relative to cells expressing wild-type PAX3 (data not shown).Fifty-nine of these differentially expressed genes have biological functions directly involved with multiple processes important for maintaining chromosome number and structure (Table 1) with 48/59 of these genes being downregulated.The affected genes include kinases that regulate chromosome separation and mitotic progression, cohesins, centromeric proteins, regulators of cytokinesis, components or regulators of the condensin complex, proteins involved with chromosomal segregation, and components of the mitotic spindle.Finally, six of the 59 genes that specifically contribute to promoting aneuploidy contain PAX3-FOXO1 binding sites in their proximal promoters, as previously described [29] (Table 1, carat) and six of these genes have altered gene expression levels consistent with changes seen in human tumor samples [30][31][32][33] (Table 1, pound sign).
In our miRNA deep sequencing analysis we found a total of 104 miRNAs whose expression changed in a PAX3-FOXO1-dependent manner, 61 of which are downregulated and 43 of which are upregulated (data not shown).Using miRTarBase [34] we found that out of these 104 total miRNA, 19 have validated target genes important for maintaining chromosome number and integrity (10 downregulated, 9 upregulated, Table 2), with validation Comparisons are made between empty vector and PAX3-FOXO1 (V vs. PF), PAX3 and PAX3-FOXO1 (P3 vs. PF) or empty vector and PAX3 (V vs. P3) and the *indicates miRNA that target genes involved with the promotion of aneuploidy and in proliferative control.on miRTarBase by at least two independent experimental methods.In addition to these 19 validated targets, we found using the TargetScan database that 22 additional miRNAs (17 downregulated, 5 upregulated, Table 2) are highly predicted to target genes important for chromosome number and integrity, based on their predicted efficacy of targeting (context score ≥ 85%) [35,36] or probability of conserved targeting (P CT ) [37] [P CT ≥ 0.8], as previously described [38], with several of the genes identified by TargetScan being validated targets in MirTarBase.
Combined, the differentially expressed miRNAs target 56 additional genes that promote aneuploidy with many of these genes being targeted by multiple differentially expressed miRNA.
To determine if the PAX3-FOXO1-dependent mRNA and miRNA changes translate into differences in chromosome number and structure, we performed a cytogenetic analysis of proliferating primary myoblasts (Table 3A and Figure 1C) and found that the majority (48/62 cells -77.4%) of cells transduced with empty vector contain the normal complement of 40 (2N) or 80 (4N) chromosomes, which is consistent with previous reports for the presence of diploidy and tetraploidy in proliferating myoblasts [39].In a similar manner, a majority of cells (46/50 cells -92%) ectopically expressing PAX3 contained the normal complement of chromosomes, demonstrating that the process of transduction and ectopic expression of protein does not affect chromosome number or structure.
Further, we observed an increased number of disrupted chromosomal structures in cells stably expressing PAX3-FOXO1 relative to the negative control cells (vector and PAX3).We noted 16 cells (15.5%) with sister chromatid dissociation (Figure 1C, right panel, solid arrows), 17 cells (16.5%) with telomere association, and 4 cells (3.9%) with double minutes (Table 1B) with only two of these cells having the presence of both sister chromatid dissociation and telomere association.Although some chromosomal disruptions were observed in the control cells, these events were minimal (incidence <4%) and did not include the presence of double minutes (Table 3B).
We previously published that PAX3-FOXO1 is phosphorylated at three independent sites [27,28,40] and phosphorylation contributes to ARMS oncogenic phenotypes in an in vitro cellular model [26].To determine how phosphorylation at these sites affects PAX3-FOXO1dependent changes in chromosome number and structure, we utilized mutants in which each individual site was mutated to a phospho-incompetent alanine (S201A, S205A, or S209A) [26] primary myoblasts stably transduced with these mutants demonstrates that all mutants were expressed to levels equivalent to that of wild-type PAX3-FOXO1 (Figure 1B).Cytogenetic analysis demonstrated that similar to wild-type PAX3-FOXO1, a majority of cells stably expressing S201A [57/64 cells -89.1%], S205A [62/69 cells -89.8%], and S209A [55/57 cells -96.5%] had chromosome numbers <2N or >2N to <4N with an increase in the number of hypertetraploid cells (>4N) for all three samples (Table 3A).We also observed similar chromosomal abnormalities for each of the phopshoincompetent mutants relative to wild-type PAX3-FOXO1 (Figure 1D); however, the individual phosphorylation events seem to impact these chromosomal aberrations differently (Table 3B).While all three mutants have a similar percentage of cells containing sister chromatid dissociation relative to wild-type PAX3-FOXO1, the expression of S209A increased the number of individual events within each cell, with some cells having over 60 dissociations.Although both S201A and S205A increased the percentage of cells with telomere association, loss of phosphorylation at S205 or S209 decreased the number of individual events within each cell relative to PAX3-FOXO1.Finally, double minutes were present in a similar percentage of cells expressing the fusion protein as either wild type or mutant.However, cells stably expressing the S205A mutant had an increased number of double minutes per cell relative to the wild-type fusion protein.Taken together, these results are the first to demonstrate that the presence of an oncogenic fusion protein, PAX3-FOXO1, is sufficient to promote aneuploidy and disrupt chromosome structure and that phosphorylation contributes to this state.

PAX3-FOXO1 overrides cell cycle inhibition to enhance cellular proliferation
One of the initial cellular responses to aneuploidy is an attenuation of proliferation [41].An examination of our comparative transcriptomic analysis revealed that of the 846 differentially expressed mRNA, 93 (nearly 11%) have biological functions important for cellular proliferation (Table 1), twelve of which have previously described PAX3-FOXO1 binding sites in their proximal promoter (Table 1, carat) [29], and 30 had altered gene expression levels consistent with changes seen in human tumor samples [30][31][32][33] (Table 1, pound sign).We also found 13 of the downregulated and 9 of the upregulated miRNAs are experimentally validated on miRTarBase to target mRNA whose biological function is important for proliferative control.Further, 15 of the downregulated and 9 of the upregulated miRNAs are highly predicted to target growth regulatory genes (Table 2), based on their predicted efficacy of targeting or probability of conserved targeting, as described above.
Consistent with the predicted cellular response to the aneuploid state, we found that of the 93 alternatively expressed mRNA, 23 are cell cycle regulatory genes and include decreases in cyclins and their related cyclin dependent kinases and increases in the expression of cyclin dependent kinase inhibitors (Table 1).Further, we found 11 of the downregulated and 12 of the upregulated miRNAs also affect cell cycle regulatory proteins in a manner consistent with the attenuation of growth.To determine how these changes translate into effects on cellular growth, we determined doubling times of primary myoblasts stably transduced with empty vector, PAX3, or PAX3-FOXO1.We found the doubling time of primary myoblasts transduced with empty vector to be approximately 35 hours and that the stable expression of PAX3 enhanced the growth rate by reducing the doubling time to approximately 20 hours (Figure 2).Primary myoblasts stably expressing PAX3-FOXO1 also had a reduced doubling time of approximately 20 hours (Figure 2), a result that seems to be in contrast to the presence of aneuploidy in these cells and the observed changes in cell cycle regulatory mRNA and miRNA.Further, we determined proliferation rates on primary myoblasts stably expressing the PAX3-FOXO1 phospho-mutants.We found that although cells expressing S201A have an enhanced proliferation rate relative to the negative control, the rate is significantly slower than for cells stably expressing wildtype PAX3-FOXO1 (Figure 2).The stable expression of S205A and S209A resulted in proliferation rates that were The microRNA, target gene name, function, and fold change for each miRNA are listed; downregulated microRNA indicated by negative numbers, upregulated miRNA indicated by positive numbers.Genes and functions listed in normal font indicate validated targets; genes and functions listed in italics indicate genes highly predicted to be targets of the indicated microRNA.
Comparisons are made between empty vector and PAX3-FOXO1 (V vs. PF), PAX3 and PAX3-FOXO1 (P3 vs. PF) or empty vector and PAX3 (V vs. P3) and the *indicates miRNA that target genes involved with the promotion of aneuploidy and in proliferative control.www.impactjournals.com/oncotargetindistinguishable from the empty vector negative control.
To better understand this result, we examined our mRNA and miRNA comparative transcriptomic analyses for alterations in mRNA and miRNAs whose biological functions may affect growth independent of cell cycle regulatory proteins.We found that the expression of PAX3-FOXO1 is sufficient to alter the mRNA for multiple growth factors (14 genes) and growth factor receptors (18 genes) in a manner that is in direct alignment with our observed increased proliferation rate.These include the growth factor receptors c-MET [22], and IGF1R [20] (both direct targets of PAX3-FOXO1), FGFR4, Erbb3 (HER3), IL6ST, and the receptor NOTCH2, which in certain cell types enhances proliferation.We also found the increased expression of many of the associated growth factors including IGF2, IGFBP3, and IGFBP5 and several proliferative transcription factors (17 genes), including CREB3, FOXO4, HOXB9, and Myc (Table 1).Further, most of the miRNA have validated or predicted targets that are growth factor receptors, growth factors, or proliferative transcription factors (Table 2).Interestingly, 18 of the downregulated and 13 of the upregulated miRNAs target genes important for both aneuploidy and the regulation of proliferative control.Taken together, these results demonstrate that the expression of an oncogene is sufficient to override aneuploid-dependent attenuation of growth by globally altering the expression of growth factor related mRNA and miRNA.Metaphase chromosome analysis on cells stably transduced with empty vector (Vector), PAX3, PAX3-FOXO1, or the PAX3-FOXO1 phosphomutants (S201A, S205A, or S209A).The number of cells containing the indicated structural abnormality along with a range of the number of occurrences of the abnormality within each cell is listed.www.impactjournals.com/oncotarget

Phosphorylation contributes to PAX3-FOXO1dependent differential gene expression
To determine how phosphorylation of PAX3-FOXO1 affects the expression of genes important for maintaining chromosome number and structure, we performed a qRT-PCR analysis on a subset of genes with fusion protein-dependent altered expression.We tested genes in chromosome segregation (AurkA), chromosome cohesion (BUB1, Cep250, and Sgol1), and cytokinesis (COROC1, Myh10, and Prc1).We observed changes in gene expression consistent with our mRNA deep sequencing results, in which the presence of PAX3-FOXO1 promotes a significant decrease in the expression of all seven genes (Figure 3A).Further, with the exception of Myh10 and Cep250, we found inhibiting the phosphorylation of PAX3-FOXO1 at Ser201 or Ser205 corrected these decreases, with gene expression levels equivalent to the empty vector or PAX3 controls (Figure 3A).However, loss of phosphorylation at Ser209 was unable to correct these decreases and in fact promoted a further 2-fold decrease in the expression of Prc1.
A qRT-PCR analysis on a subset of proliferation genes, including cell cycle regulatory genes (CDKN1C, CDK1, and Cyclin D1) and growth factor genes (Erbb3, IGF2, and IGFBP5) once again showed PAX3-FOXO1dependent changes consistent with our mRNA deep sequencing results (Figure 3B).We found the stable expression of the phospho-incompetent mutants also increased (CDKN1C) or decreased (CDK1 and Cyclin D1) cell cycle regulatory gene expression relative to the negative control.However, with the exception of CDK1 in cells expressing the S209A mutant, these changes were much less severe relative to wild-type PAX3-FOXO1.With respect to growth factor related genes, we found that the S209A mutant reduced the expression of all three genes to levels indistinguishable from the negative control.Further, the expression of S201A and S205A also altered the expression of IGF2 and IGFBP5 that trended toward background levels, but had little effect on the expression We performed a similar quantitative analysis on a select set of miRNAs, chosen for their ability to target genes important for the promotion of aneuploidy and for the regulation of proliferation.We found changes in miRNA expression consistent with our deep sequencing results (Figure 3C).As with our mRNA results, we found that inhibiting phosphorylation of PAX3-FOXO1 at Ser201 or Ser205 returned miRNA expression levels to that of the empty vector or PAX3 controls (Figure 3C).While loss of phosphorylation at Ser209 was able to correct the altered miRNA expression for miR30c-2-5p, S209A had no effect on the altered expression of miR130b-3p.Interestingly, the expression of S209A resulted in a 2-or 4-fold increase in the expression of miR-196a-5p and miR301a-3p, respectively, relative to wild-type PAX3-FOXO1 (Figure 3C).Taken together, our results demonstrate that phosphorylation contributes to the altered expression of genes and miRNAs important for the PAX3-FOXO1-dependent promotion of aneuploidy, aberrant chromosome structure, and regulation of growth.

DISCUSSION
Aneuploidy is common in solid tumors, which given the enhanced proliferative activity of transformed or malignant cells, suggests that these cells have acquired mechanisms to overcome proliferative defects associated with this state.In this report we demonstrate for the first time that the presence of an oncogenic protein, PAX3-FOXO1, is sufficient to serve as a driving mutation to initiate a cascade of changes in mRNA and miRNA whose biological functions are important for maintaining proper chromosome number and structure and proliferative control (Tables 1 and 2).We show that these changes translate into biological effects, since cells expressing PAX3-FOXO1 are primarily aneuploid (Table 3A), have altered chromosome structure (Table 3B), and have compensated to override the antiproliferative responses to aneuploidy in physiologically relevant primary myoblasts (Figure 2).Consistent with our previous work [26], we demonstrate that inhibiting PAX3-FOXO1 phosphorylation reverts many of the gene expression changes to background levels (Figure 3), which translate into biological effects by altering the extent of chromosome structural abnormalities (Figure 1 and Table 3B) and enhanced proliferation rates (Figure 2).Given that these changes result explicitly from the expression and phosphorylation of PAX3-FOXO1, we conclude that the acquisition of the oncogene is the initiating driver mutation in the development of ARMS.
Because the aneuploid state is detrimental to a cell, growth retardation is one of the first cellular responses [42,43].Therefore, in order for muscle cells to develop into ARMS, the cells containing the fusion protein must overcome this proliferative defect.Consistent with the cell's normal response to the aneuploid state, we saw changes in cell cycle regulatory genes that would be expected to inhibit cellular proliferation.However, we also found that the presence of the fusion protein enhances proliferation, which seems in direct contrast to altered cell cycle regulatory genes.This apparent contradiction can be explained by the fact that we also observed significant increases in the expression of growth factor related genes and proliferative transcription factors, several of which are known direct targets of PAX3-FOXO1 [20,44].Taken together, these results support the idea that the oncogenic fusion protein serves a second role as a driving mutation: it promotes the increased expression of growth factor related genes, either directly or indirectly through downstream effects, that are capable of negating the inhibitory changes in cell cycle regulatory proteins to override the growth retarding effects of the aneuploid state.
Based on the evidence presented here, we propose the following model by which PAX3-FOXO1 serves as the initial driving mutation in the development of ARMS (Figure 4).In this model myogenic cells in situ randomly and somatically acquire the t(2;13)(q35;q14) translocation, an event which is mimicked in our in vitro model system through the stable transduction of primary myoblasts.The subsequent expression of PAX3-FOXO1, maintained over repeated cellular duplications either in situ or in vitro, will bind to select promoters and regulatory sequences, thereby directly altering the expression of a subset of genes (Table 1, carats).These alterations could then initiate a cascade of indirect downstream events resulting in a global alteration of gene regulatory networks that ultimately reprogram proliferating myoblasts.This reprogramming results in the disruption of multiple aspects that lead to aneuploidy, including maintenance of the segregation machinery, regulation of chromosome segregation, promotion of chromosome cohesion and condensation, and insuring In response to the aneuploid state, the cells attempt to halt proliferation, which is evidenced through our observed changes in the expression of cell cycle regulatory genes and miRNAs, and which most likely result from the aneuploid state and not from direct involvement of the fusion protein.However, PAX3-FOXO1 directly induces the expression of multiple and redundant proliferative genes, including growth factors, growth factor receptors, and proliferative transcription factors (Table 2, carats), thereby overriding aneuploidy-induced proliferation defects.Finally, the cells must be able to acquire a tolerance to the aneuploid state, most likely by affecting genes important in p53 regulation, in order to promote ARMS progression.
We reported that PAX3-FOXO1 is phosphorylated at Ser201 by GSK3β and small molecule inhibitors of this enzyme not only reduce phosphorylation at this site but also attenuate ARMS tumor phenotypes [26].Consistent with this work, and adding to our present model, we found that phosphorylation contributes to the PAX3-FOXO1dependent changes in gene and miRNA expression leading to the acquisition and persistence of aneuploidy.We found that although inhibiting phosphorylation of PAX3-FOXO1 at Ser201 and Ser205 did not affect aneuploidy, it altered the severity of chromosomal structural abnormalities (Figure 1 and Table 3B) and blocked the enhanced proliferation of primary myoblasts (Figure 2).We found phosphorylation contributes to the altered expression of multiple genes and miRNAs, which was evidenced by the reversion of gene expression to levels equivalent to the negative control in cells stably expressing phosphoincompetent mutants (Figure 3).Therefore, our results provide additional evidence to support the role posttranslational modifications make in regulating the contributions of an oncogene to the development of a solid tumor.Further, these results identify this event, the specific phosphorylation of PAX3-FOXO1, as a viable biological target for therapy development (Figure 4).Interestingly, inhibition of phosphorylation at Ser209 enhanced the severity of sister chromatid dissociation (Figure 1D and Table 3B), and had proliferation rates indistinguishable from the negative control (Figure 1).Although we previously reported undetectable levels of phosphorylation of PAX3-FOXO1 at Ser209 in this system [28], the results presented here would suggest that this event does occur at greatly reduced levels and below the limits of detection.
The results presented here are important for directing the development of new therapies for the treatment of ARMS (Figure 4) and provide implications for how others may approach the development of therapies for solid tumors.Many present experimental therapies for ARMS, and other solid tumors, focus on inhibiting a single gene or a single pathway genetically located downstream of the oncogene.However, these therapies have not proven effective in Phase I or Phase II clinical trials for ARMS [45][46][47][48].These outcomes are not surprising given the global nature of mRNA and miRNA alterations seen in cells that express PAX3-FOXO1.As such, therapies targeting a single gene product or pathway would be expected to have limited efficacy since multiple genes with similar biological functions could easily compensate.
Therefore, we propose the development of a multi-faceted regimen that targets genetic or biological characteristics identified from our work.We published that small molecules such as LiCl attenuate ARMS tumor phenotypes [26].Others demonstrated that compounds that interfere with growth factor dependent proliferation, such as Cixutumumab [48], or the survival of aneuploid cells, such as chloroquine [49], might be effective as new treatments for human solid tumors.Therefore, we propose a combination therapy that minimally utilizes these three drugs for the treatment of ARMS.This regimen is expected to target both the driver mutation, with LiCl or other GSK3beta inhibitors preventing the phosphorylation of PAX3-FOXO1 at Ser201 [26] thereby inhibiting altered gene expression and the biological consequences of this driver mutation, with chloroquine killing aneuploid cells while Cixutumumab removes the growth-factor-dependent proliferative advantage.This regimen is expected to be specific for ARMS tumor cells, since all targets derive from the presence of the unique genetic mutation, have minimal negative effects since all proposed drugs are proven safe in human subjects, and would inhibit three molecular events essential for the viability of ARMS tumor cells.
Finally, the paradigm created here has implications that may be applied to other tumor models.Studies are being published with more frequency that utilize deepsequencing strategies to examine global gene and miRNA expression changes resulting from the presence of an oncogene or the process of malignant transformation for examples see [50][51][52][53].Often a single gene, miRNA, or miRNA locus is selected from these global studies for more in depth investigation to utilize as a specific target for therapy development.However, given that our results demonstrate that individual tumor phenotypes (e.g., induction of aneuploidy) arise from alterations in multiple biological aspects of this phenotype (e.g., mitotic progression, segregation machinery, etc.), it would be expected that a cell would be capable of compensating for the therapeutic inhibition of a single gene or miRNA.Therefore, as illustrated in the present study, deep sequencing studies should be utilized to discover how global gene regulatory alterations affect biological phenotypes and then use this knowledge to develop a multi-faceted therapeutic approach to target specific aspects of multiple different tumor phenotypes that are essential for the progression of the disease.www.impactjournals.com/oncotargetDiscovery Rate (FDR) of 5%.We used the default 'Median Normalization' in EBSeq to make the counts comparable across samples.

Proliferation assay
The proliferation rate of cells stably expressing empty vector, FLAG-PAX3, FLAG-PAX3-FOXO1, or the phospho-incompetent mutants was assessed using the CCK-8 colorimetric assay kit, according to the manufacturer's specifications (Cell Counting Kit-8, Dojindo Molecular Technologies) and as previously described [26].

qRT-PCR
Total RNA was extracted from cells stably expressing empty vector, FLAG-PAX3, FLAG-PAX3-FOXO1, or the FLAG-PAX3-FOXO1 phospho-mutants.Equal amounts of isolated RNA (100ng) were used for cDNA synthesis using the iScript cDNA synthesis kit (Bio-Rad) for mRNA and the Taqman miRNA reverse transcription kit (Applied Biosystems) for miRNA.The qRT-PCR was performed on the resulting cDNA using the CFX96 Touch™ Real-Time PCR Detection System using commercially available primer/probe sets and the Applied Biosystems Universal Master Mix, according to the manufacturer's specifications.All results were normalized for GAPDH (mRNA) or the U6 small nuclear RNA for miRNA and reported as fold expression relative to the results obtained for cells stably transduced with the empty vector.

Figure 1 :
Figure 1: The expression of PAX3-FOXO1 promotes aneuploidy and chromosomal structural abnormalities.Expression of A. PAX3, PAX3-FOXO1 or B. PAX3-FOXO1 phosphomutants.Total extracts were made from stably transduced cells and protein was determined using an antibody specific for PAX3, as described in the Materials and Methods.Representative picture of a metaphase chromosome analysis for C. cells stably transduced with empty vector, PAX3 or PAX3-FOXO1 or D. individual PAX3-FOXO1 phosphomutants.The closed arrows indicate representative sister chromatid dissociation, the open arrow indicate representative telomere association, and the dotted arrow indicate a representative double minute.

Figure 2 :
Figure 2: Stably transduced primary myoblasts (vector, PAX3, wild-type PAX3-FOXO1, or PAX3-FOXO1 phosphomutant) were plated, grown for four days, growth determined using the CCK-8 cell counting kit and plotted as a function of time (A) from which doubling times were determined, as described in the Materials and Methods.Error bars represent the standard deviation from three independent determinations and P-values were computed using non-parametric two-way ANOVA analyses comparing each treatment condition to the empty vector negative control.(**P = 0.005, ****P < 0.0001).

Figure 3 :
Figure 3: Quantitative RT-PCR analysis for aneuploidy mRNA (A), proliferation mRNA (B) and miRNA (C).Total RNA was isolated from stably transduced cells [empty vector (white), PAX3 (stippled), PAX3-FOXO1 (black), or the phosphomutants S201A (light gray), S205A (medium gray), or S209A (hashed gray).Quantitative RT-PCR was performed using primers specific for the indicated mRNA A. and B. or microRNA C., as described in the Materials and Methods.Error bars represent the standard deviation from three independent determinations and P-values were computed using non-parametric two-way ANOVA analyses.The asterisk indicates statistical comparisons between empty vector and each sample (*P = 0.001, **P < 0.0001).The hash-tap indicates comparisons between wild-type PAX3-FOXO1 and the sample (#P = 0.001).

Figure 4 :
Figure 4: Model describing the role of PAX3-FOXO1 in the development of ARMS and how it informs potential therapy development.

Table 2a : Differentially expressed miRNA that target genes important for promoting aneuploidy -miRNA downregulated by PAX3-FOXO1 relative to empty vector.
The microRNA, target gene name, function, and fold change for each miRNA are listed; downregulated microRNA indicated by negative numbers, upregulated miRNA indicated by positive numbers.Genes and functions listed in normal font indicate validated targets; genes and functions listed in italics indicate genes highly predicted to be targets of the indicated microRNA.

Table 2b : Differentially expressed miRNA that target genes important for promoting aneuploidy -miRNA upregulated by PAX3-FOXO1 relative to empty vector.
The microRNA, target gene name, function, and fold change for each miRNA are listed; downregulated microRNA indicated by negative numbers, upregulated miRNA indicated by positive numbers.Genes and functions listed in normal font indicate validated targets; genes and functions listed in italics indicate genes highly predicted to be targets of the indicated microRNA.

Table 2d : Differentially expressed miRNA that target genes important for proliferative control -miRNA upregulated by PAX3-FOXO1 relative to empty vector.
The microRNA, target gene name, function, and fold change for each miRNA are listed; downregulated microRNA indicated by negative numbers, upregulated miRNA indicated by positive numbers.Genes and functions listed in normal font indicate validated targets; genes and functions listed in italics indicate genes highly predicted to be targets of the indicated microRNA.Comparisons are made between empty vector and PAX3-FOXO1 (V vs. PF), PAX3 and PAX3-FOXO1 (P3 vs. PF) or empty vector and PAX3 (V vs. P3) and the *indicates miRNA that target genes involved with the promotion of aneuploidy and in proliferative control.