Myxoid and round-cell liposarcoma is a frequently encountered liposarcoma subtype. The mainstay of treatment remains surgical excision with or without chemoradiation. However, treatment options are limited in the setting of metastatic disease. Cancer-testis antigens are immunogenic antigens with the expression largely restricted to testicular germ cells and various malignancies, making them attractive targets for cancer immunotherapy. Gene expression studies have reported the expression of various cancer-testis antigens in liposarcoma, with mRNA expression of CTAG1B, CTAG2, MAGEA9, and PRAME described specifically in myxoid and round-cell liposarcoma. Herein, we further explore the expression of the cancer-testis antigens MAGEA1, ACRBP, PRAME, and SSX2 in myxoid and round-cell liposarcoma by immunohistochemistry in addition to determining mRNA levels of CTAG2 (LAGE-1), PRAME, and MAGEA3 by quantitative real-time PCR. Samples in formalin-fixed paraffin-embedded blocks (n=37) and frozen tissue (n=8) were obtained for immunohistochemistry and quantitative real-time PCR, respectively. Full sections were stained with antibodies to MAGEA1, ACRBP, PRAME, and SSX2 and staining was assessed for intensity (1–2+) and percent tumor positivity. The gene expression levels of CTAG2, PRAME, and MAGEA3 were measured by quantitative real-time PCR. In total, 37/37 (100%) of the samples showed predominantly strong, homogenous immunoreactivity for PRAME. There was a variable, focal expression of MAGEA1 (11%) and SSX2 (16%) and no expression of ACRBP. Quantitative real-time PCR demonstrated PRAME and CTAG2 transcripts in all eight samples: six tumors with high mRNA levels; two tumors with low mRNA levels. The gene expression of MAGEA3 was not detected in the majority of cases. In conclusion, myxoid and round-cell liposarcomas consistently express PRAME by immunohistochemistry as well as CTAG2 and PRAME by qualitative real-time PCR. This supports the use of cancer-testis antigen-targeted immunotherapy in the treatment of this malignancy.

Myxoid and round cell liposarcomas constitute approximately one-third of all liposarcomas, a relatively common group of fat-derived soft tissue sarcomas. The histomorphology is a continuum between highly differentiated myxoid and poorly differentiated round cell components. The gold standard of diagnosis is dependent on histomorphology and/or identification of t(12;16)(q13;p11) translocation by cytogenetics or demonstration of DDIT3 rearrangements by fluorescence in situ hybridization. There are currently no diagnostic immunohistochemical stains available. The broad range of myxoid neoplasms in the differential diagnosis includes a variety of sarcomas. Given the notable differences in disease biology among myxoid neoplasms, which range from benign to aggressive, an accurate diagnosis is imperative for proper treatment and prognostication. Prompted by our recent study showing frequent expression of the cancer testis antigen NY-ESO-1 in myxoid and round cell liposarcomas, we sought to evaluate the utility of NY-ESO-1 as an immunohistochemical marker for myxoid and round cell liposarcoma among mesenchymal myxoid neoplasms within the differential diagnosis. Formalin-fixed, paraffin-embedded blocks were obtained for the following mesenchymal myxoid neoplasms (n=138): myxoid and round cell liposarcoma (n=38); well-differentiated liposarcoma (n=12); lipoma (n=20; 4 with myxoid change); extra-cardiac soft tissue myxoma (n=39); extraskeletal myxoid chondrosarcoma (n=12); myxofibrosarcoma (n=10: 5 low grade, 2 intermediate grade, 3 high grade); and low-grade fibromyxoid sarcoma (n=7). Utilizing standard immunohistochemistry protocols, full sections were stained with NY-ESO-1 (clone E978), and staining was assessed for intensity (1–2+), percentage of tumor positivity, and location. In all, 36/38 (95%) of the myxoid and round cell liposarcomas demonstrated NY-ESO-1 immunoreactivity. The majority of the positive cases (34/36; 94%) showed strong, homogenous staining (>50% tumor positivity), and two cases (6%) showed weak (1+ intensity), patchy staining (20–30% tumor positivity). Immunoreactivity was predominantly cytoplasmic. All the other neoplasms evaluated were negative for NY-ESO-1. NY-ESO-1 appears to be a sensitive and a specific marker for myxoid and round cell liposarcoma among mesenchymal myxoid neoplasms. The assessment of NY-ESO-1 expression by immunohistochemistry in the appropriate setting provides a cheaper, faster, and more accessible confirmatory test.

MicroRNAs (miRNAs) regulate the expression of networks of genes and their dysregulation is well documented in human malignancies; however, limited information exists regarding the impact of miRNAs on the development and progression of osteosarcoma (OS). Canine OS exhibits clinical and molecular features that closely resemble the corresponding human disease and it is considered a well-established spontaneous animal model to study OS biology. The purpose of this study was to investigate miRNA dysregulation in canine OS. We evaluated miRNA expression in primary canine OS tumors and normal canine osteoblast cells using the nanoString nCounter system. Quantitative PCR was used to validate the nanoString findings and to assess miR-9 expression in canine OS tumors, OS cell lines, and normal osteoblasts. Canine osteoblasts and OS cell lines were stably transduced with pre-miR-9 or anti-miR-9 lentiviral constructs to determine the consequences of miR-9 on cell proliferation, apoptosis, invasion and migration. Proteomic and gene expression profiling of normal canine osteoblasts with enforced miR-9 expression was performed using 2D-DIGE/tandem mass spectrometry and RNA sequencing and changes in protein and mRNA expression were validated with Western blotting and quantitative PCR. OS cell lines were transduced with gelsolin (GSN) shRNAs to investigate the impact of GSN knockdown on OS cell invasion. We identified a unique miRNA signature associated with primary canine OS and identified miR-9 as being significantly overexpressed in canine OS tumors and cell lines compared to normal osteoblasts. Additionally, high miR-9 expression was demonstrated in tumor-specific tissue obtained from primary OS tumors. In normal osteoblasts and OS cell lines transduced with miR-9 lentivirus, enhanced invasion and migration were observed, but miR-9 did not affect cell proliferation or apoptosis. Proteomic and transcriptional profiling of normal canine osteoblasts overexpressing miR-9 identified alterations in numerous genes, including upregulation of GSN, an actin filament-severing protein involved in cytoskeletal remodeling. Lastly, stable downregulation of miR-9 in OS cell lines reduced GSN expression with a concomitant decrease in cell invasion and migration; concordantly, cells transduced with GSN shRNA demonstrated decreased invasive properties. Our findings demonstrate that miR-9 promotes a metastatic phenotype in normal canine osteoblasts and malignant OS cell lines, and that this is mediated in part by enhanced GSN expression. As such, miR-9 represents a novel target for therapeutic intervention in OS.