Reduced B7-H3 expression by PAX3-FOXO1 knockdown inhibits cellular motility and promotes myogenic differentiation in alveolar rhabdomyosarcoma

B7-H3 (also known as CD276) is associated with aggressive characteristics in various cancers. Meanwhile, in alveolar rhabdomyosarcoma (ARMS), PAX3-FOXO1 fusion protein is associated with increased aggressiveness and poor prognosis. In the present study, we explored the relationship between PAX3-FOXO1 and B7-H3 and the biological roles of B7-H3 in ARMS. Quantitative real time PCR and flow cytometry revealed that PAX3-FOXO1 knockdown downregulated B7-H3 expression in all the selected cell lines (Rh-30, Rh-41, and Rh-28), suggesting that PAX3-FOXO1 positively regulates B7-H3 expression. Gene expression analysis revealed that various genes and pathways involved in chemotaxis, INF-γ production, and myogenic differentiation were commonly affected by the knockdown of PAX3-FOXO1 and B7-H3. Wound healing and transwell migration assays revealed that both PAX3-FOXO1 and B7-H3 were associated with cell migration. Furthermore, knockdown of PAX3-FOXO1 or B7-H3 induced myogenin expression in all cell lines, although myosin heavy chain induction varied depending on the cellular context. Our results indicate that PAX3-FOXO1 regulates B7-H3 expression and that PAX3-FOXO1 and B7-H3 are commonly associated with multiple pathways related to an aggressive phenotype in ARMS, such as cell migration and myogenic differentiation block.

www.nature.com/scientificreports/ ( Fig. 1c,d). These results suggest that PAX3-FOXO1 positively regulates the expression of B7-H3 at both the mRNA and protein levels.
PAX3-FOXO1 and B7-H3 are associated with multiple pathways related to an aggressive rhabdomyosarcoma phenotype. To unravel the pathways affected by PAX3-FOXO1 and B7-H3 in ARMS, gene expression analysis was performed using microarray and gene set enrichment analysis (GSEA). The gene expression signature of siPF or siRNA targeting B7-H3 (siB7-H3) transfected Rh-30 cells (Rh-30 siPF or Rh-30 siB7-H3, respectively) was compared with that of control siRNA (siCont) transfected Rh-30 cells (Rh-30 siCont). Gene expression analysis revealed that in Rh-30 siPF and Rh-30 siB7-H3, the expression levels of 636 and 565 genes were decreased, respectively (log 2 ratio, ≤ 1.0), and those of 700 and 549 genes were increased, respectively (log 2 ratio, ≥ 1.0), compared with the respective levels in Rh-30 siCont (Fig. 2a,b). Among them, the expression levels of 130 genes were downregulated and those of 159 genes were upregulated in both the Rh-30 siPF and Rh-30 siB7-H3 compared to the respective levels in the Rh-30 siCont (Fig. 2a,b). A heat map of top 50 shared downregulated or upregulated genes is shown in Fig. 2c. GSEA revealed that several pathways related to chemotaxis, INF-γ production, and regulation of T cell immunity were inactivated in both Rh-30 siPF and Rh-30 siB7-H3 unlike in Rh30 siCont, suggesting that PAX3-FOXO1 and B7-H3 may contribute to tumor cell metastasis and immune evasion (Fig. 3a,b). In contrast, several pathways related to muscle cell differentiation were activated in both Rh-30 siPF and Rh-30 siB7-H3 but not in Rh30 siCont, suggesting that PAX3-FOXO1 and B7-H3 may block myoblast differentiation (Fig. 3c,d).
PAX3-FOXO1 and B7-H3 are associated with cell migration. To confirm that PAX3-FOXO1 and B7-H3 contribute to metastasis in ARMS, as suggested by GSEA, we performed wound healing and transwell assays. Wound healing assays revealed that cell migration in ARMS was significantly attenuated by the knockdown of PAX3-FOXO1 or B7-H3 in all the three cell lines (Fig. 4a,b). The transwell assays similarly demonstrated that cell migration in ARMS was significantly attenuated by the knockdown of PAX3-FOXO1 or B7-H3 in all the three cell lines (Fig. 4c).
To determine the pathways contributing to cell migration that are affected by PAX3-FOXO1 and B7-H3, we performed qRT-PCR for CXCR4 and STYK1, as gene expression analysis indicated that both these genes were downregulated in Rh-30 siPF and Rh-30 siB7-H3. As expected, CXCR4 was downregulated by transfection of all the three cell lines with siPF or siB7-H3 (Fig. 5a). In addition, STYK1 was downregulated in Rh-30 and Rh-41 by transfection with siPF or siB7-H3, although in Rh28, STYK1 was downregulated by transfection with siPAX3-FOXO1, but not siB7-H3 (Fig. 5b). myoblast differentiation, we investigated the expression level of myogenin, a transcription factor required for myoblast differentiation, following knockdown of PAX3-FOXO1 or B7-H3. Both qRT-PCR and western blot analyses revealed that myogenin was upregulated in all the three cell lines by transfection with siPF or siB7-H3 ( Fig. 6a,b). To investigate whether upregulation of myogenin led to differentiation, we examined the expression level of myosin heavy chain (MyHC) in the ARMS cell lines. Immunofluorescence of MyHC revealed that in Rh-30 cells, transfection with siPF induced MyHC more strongly than the transfection with siB7-H3, whereas in Rh-41 cells, transfection with siB7-H3 induced MyHC more strongly than did siPF (Fig. 6c).

Discussion
In this study, we investigated the regulatory mechanism and function of B7-H3 in PAX3-FOXO1-positive ARMS. PAX3-FOXO1 regulates the expression of many target genes 25 , and we determined that PAX3-FOXO1 also regulates B7-H3 expression in ARMS. Our data indicate that although regulated by PAX3-FOXO1, B7-H3 regulates the expression of numerous genes, some of which are also regulated by PAX3-FOXO1. Therefore, B7-H3 is considered to be an important protein downstream of PAX3-FOXO1 that contributes to aggressive characteristics of PAX3-FOXO1-positive ARMS. Based on the GSEA results, we demonstrate that B7-H3 is associated with cell migration in ARMS, similar to that in various other cancers. However, the mechanisms by which B7-H3 regulates migration differ according to the type of cancer. For example, B7-H3 regulates CXCR4 in gastric cancer 16 , PI3K signaling in bladder and colorectal cancers 14,15 , JAK2-STAT3 signaling in hepatocellular carcinoma 28 , and MMP-2 in melanoma 29 . Our data show that B7-H3 regulates CXCR4 in ARMS. Notably, CXCR4 is also regulated by PAX3-FOXO1 30 . Previous report have shown that antibody to CXCR4 inhibit metastasis in ARMS 31 . In addition, we found that PAX3-FOXO1 regulates the expression of STYK1, which is also associated with tumor invasion and metastasis 32,33 , although molecular targeted therapy against STYK1 is not developed. B7-H3 also regulates STYK1 expression in Rh-30 and Rh-41 cells, but not in Rh-28 cells, indicating that the effect of B7-H3 on STYK1 varies depending on the ARMS cell line.
Furthermore, we demonstrate that knockdown of B7-H3, as well as of PAX3-FOXO1, upregulates the expression of myogenin, a transcription factor needed for myogenic differentiation 34 . Although PAX3-FOXO1 inhibits myogenic differentiation in ARMS 26,35 , the effect of B7-H3 on myogenic differentiation has not been determined. Among the various functions of B7-H3 in cancer, inhibition of differentiation (not limited to myogenic differentiation) has not been reported. However, despite the observation regarding the induction of myogenin by B7-H3 knockdown in all three cell lines, the induction of further differentiation as detected by the expression of MyHC was found to vary with the cell line. Knockdown of PAX3-FOXO1 induced the expression of MyHC more strongly than did the knockdown of B7-H3 in Rh-30 cells, whereas B7-H3 knockdown induced the expression of MyHC to a greater extent than did PAX3-FOXO1 in Rh-41 cells, indicating that myogenin expression, which leads directly to increased expression of MyHC, may be dependent on the cellular context. www.nature.com/scientificreports/ These results suggest that B7-H3 might be a potential therapeutic target in ARMS with PAX3-FOXO1, which itself is not yet targetable in the present clinical settings. In various cancers, B7-H3 is an attractive surface antigen for molecular targeted therapy 36 . Enoblituzumab (MGA271) is a monoclonal antibody for B7-H3 and causes antibody-dependent cell-mediated cytotoxicity (ADCC) 11 . Enoblituzumab was evaluated in a phase 1 clinical study and demonstrated acceptable tolerability in the treated patients with solid cancer (trial NCT01391143). Additionally, antibody drug conjugate (ADC) therapies and CAR T cells for B7-H3 also demonstrated outstanding results in preclinical studies and in a phase 1 clinical trial for childhood cancer 10,12,37,38 . These targeted therapies for B7-H3 might be used for treating PAX3-FOXO1 positive ARMS that exhibits a poor prognosis with conventional intensive therapy. A phase 1 clinical trial of enoblituzumab in children and young adults with  Our results demonstrate that B7-H3 affects cell migration and myogenic differentiation and is one of the cancer driver genes in PAX3-FOXO1 positive ARMS. As cancer antigen escape seems less likely to occur when the cancer driver gene is a target antigen, B7-H3-targeting immunotherapy might be an attractive treatment option for PAX3-FOXO1 positive ARMS, although we did not examine the anti-B7-H3 targeted therapy for ARMS in this study. The analysis of B7-H3 targeted therapy is warranted in the future.
This study has several limitations. First, we did not investigate the association between B7-H3 and tumor immunity in ARMS, although the GSEA indicated that gene sets related to INF-γ production and T cell-mediated cytotoxicity were inactivated by knockdown of PAX3-FOXO1 or B7-H3. As B7-H3 is thought to be involved in tumor evasion from host immunity in other cancers, the relationship between B7-H3 and tumor evasion should be a focus of future studies. Second, we used only one siRNA for each gene. Therefore, we could not exclude the possibility of off-target effect by siRNA. Third, we did not investigate the function of B7-H3 in ARMS by in vivo experiments. In vivo experiments using the siRNA-treated cells would provide more evidence that strengthen our study results.
In conclusion, our findings demonstrate that PAX3-FOXO1 regulates B7-H3 expression, and PAX3-FOXO1 and B7-H3 are commonly associated with multiple pathways related to an aggressive phenotype in ARMS. In particular, PAX3-FOXO1 and B7-H3 contribute to cell migration and inhibit myogenic differentiation. This study provides evidence that B7-H3 contributes to the malignant characteristics of PAX3-FOXO1 positive ARMS.

Materials and methods
Cell lines and cell culture. PAX3-FOXO1 positive human ARMS cell lines, Rh-30, Rh-41, and Rh-28 that were kindly provided by Peter J. Houghton M (The Greehey Children's Cancer Research Institute, San Antonio, TX), as well as PAX3-FOXO1 negative human ERMS cell line, RD that was obtained from JCRB (Japanese Collection of Research Bioresources) Cell Bank, were used in this study. The cells were maintained in RPMI 1640 medium (Nakarai Tesque, Kyoto, Japan), supplemented with 10% fetal bovine serum (FBS; Gibco, Carlsbad, CA), penicillin (100 U/mL), and streptomycin (100 µg/mL) at 37 °C in a humidified atmosphere of 5% CO 2 . Gene expression analysis. Total RNA was extracted from RH-30 cells transfected with siCont, siPF, or siB7-H3. The total RNA was used for cRNA target synthesis using a GeneChip 3′IVT PLUS Reagent Kit (Thermo Fisher Scientific, Waltham, CA). Biotin-labeled cRNA was hybridized to a GeneChip Human Genome U133 Plus 2.0 Array (Thermo Fisher Scientific). Hybridized arrays were stained and scanned with a Genechip 3000 7G Scanner (Thermo Fisher Scientific). Data were analyzed with the Affymetrix Expression Console Software 1.4.1. All microarray data are available at the Gene Expression Omnibus database under accession number GSE127703. GSEA was performed using GSEA 3.0 (Broad Institute, San Diego, CA). A group consisting of RH-30-siPF and RH-30-siB7-H3 was compared with RH-30-siCont, and gene sets at a nominal p value (NOM p-val) < 5% were considered to be significantly enriched.
Wound healing assay. After transfection of cells with siRNAs, the cell layers were scratched using a pipette tip. Immediately after scratching (0 h), the plates were photographed and the wound area was measured using the ImageJ software and defined as the original area (100%). After 24 h of scratching, the plates were photographed again. In addition, the area of the healed wound was measured and presented as a percentage of the original area. All assays were performed in triplicate.
Transwell migration assay. At 24 h of transfection of cells with the siRNAs, each cell line was resuspended in serum-free RPMI medium. Twenty-four well-sized transwell plates with an 8-µm pore size (Corning, Corning, NY) were used for this assay. First, 1.0 × 10 5 cells suspended in 100 µL of serum-free medium were seeded on each insert, and 650 µL of RPMI medium with 10% FBS was added to the receiver plates. The plates were incubated for 24 h, and cells migrating onto the receiver plates were quantified as relative light units (RLU) using the Cell Titer Glo luminescent cell viability assay (Promega, Heidelberg, Germany) according to the procedures recommended by the manufacturer. All the assays were performed in triplicate.
Immunofluorescence. At 24 h of transfection of cells with the siRNAs, Rh-30 and Rh-41 cells were resuspended in serum-free RPMI medium. After 7 days of resuspension, cells were fixed with 4% paraformaldehyde and permeabilized with 0.1% Triton-X100. Next, the cells were washed and incubated with anti-MyHC antibody (Sigma-Aldrich) overnight (1:350), rinsed with PBS, and incubated with Alexa Fluor 488 goat anti-mouse IgG (Life Technologies, Carlsbad, CA) for 1 h (1:200). Nucleic acids were stained with Hoechst, using the NucBlue