The poor prognosis of patients with soft-tissue sarcoma as not changed in the past several decades, highlighting the necessity for new therapeutic approaches. T-cell based immunotherapies are a promising alternative to traditional cancer treatments due to their ability to target only malignant cells, leaving benign cells unharmed. The development of successful immunotherapy requires the identification and characterization of targetable immunogenic tumor antigens. Cancer-testis antigens (CTA) are a group of highly immunogenic tumor-associated proteins that have emerged as potential targets for CD8+ T-cell recognition. In addition to identifying a targetable antigen, it is crucial to understand the tumor immune microenvironment. The level of immune infiltration and mechanisms of immune suppression within the tumor play important roles in the outcome of immunotherapy.

The goal of this study is to identify targetable immunogenic antigens for T-cell based immunotherapy and to characterize the tumor immune microenvironment in human dedifferentiated liposarcoma (DDLS) by Nanostring and IHC.

To assess the complexity of the human DDLS tumor immune microenvironment and to identify target antigens we used the nCounter NanoString platform to generate a gene expression profile for hundreds of genes from RNA obtained from 29 DDLS and 10 control fat FFPE samples. To classify inflammatory status of DDLS tumors, we performed hierarchical clustering based on expression levels of selected tumor inflammatory signature genes (CCL5, CD27, CD274, CD276, CD8A, CMKLR1, CXCL9, CXCR6, HLA-DQA1, HLA-E, IDO1, LAG3, PDCDILG2, PSMB10, STAT1, TIGIT). To confirm protein expression and distribution of identified antigens, we performed immunohistochemistry on human tissue micro-arrays encompassing DDLPS tumor tissues and matched normal control tissue from 63 patients. IHC for the cancer testis antigens PBK, SPA17, MAGE-A3, NY-ESO-1 and SSX2 was performed, and the staining results were scored by two authors based on maximal staining intensity on a scale of zero to three (absent=0, weak=1, moderate=2, or strong=3) and the percentage of tumor cells that stained.

Hierarchical clustering of DDLS tumors based on expression of tumor inflammation signature genes revealed two distinct groups, consisting of 15 inflamed tumor and 14 non-inflamed tumors, demonstrating tumor heterogeneity within the DDLS sarcoma subtype. All antigens were found to be expressed in DDLS at an mRNA level. SPA17 was expressed at the highest levels in DDLS, however, this antigen was expressed at high levels in normal fat. Notably, antigens PBK and TTK had the largest fold change increase in expression in DDLS compared to normal fat controls.

Immunohistochemical analysis of selected antigens revealed that PBK was found to be expressed in 96% (52/54) of DDLS samples at high levels. Other antigens were absent or expressed at low levels in DDLS; MAGEA3 in 15.87% (10/63) NY-ESO-1 in 6.35% (4/62) and SSX2 in 12.7% (8/63) and SPA17 in 5.5% (3/54).

This data shows considerable inter-tumoral heterogeneity of inflammation, which should be taken into consideration when designing an immunotherapy for DDLS. To date, these results show promising expression of PBK antigen in DDLS, which may be used as a target in the future development of an immunotherapy for sarcoma.

The poor prognosis of patients with advanced bone and soft-tissue sarcoma has highlighted the necessity for new therapeutic approaches. T-cell based immunotherapies are a promising alternative to traditional cancer treatments due to their ability to target only malignant cells, leaving benign cells unharmed. The development of successful immunotherapy requires the identification of targetable immunogenic tumor antigens. Cancer-testis antigens (CTA) are a group of highly immunogenic tumor-associated proteins that have emerged as potential targets for CD8+ T-cell recognition. The goal of this study is to screen for CTA expression, HLA expression, and tumor T-cell infiltration in human dedifferentiated liposarcoma (DDLPS) and osteosarcoma (OS) to establish their immune profile and to identify targetable immunogenic antigens for T-cell based immunotherapy.

Human tissue micro-arrays composed of 78 cores of OS and 50 cores of DDLPS were obtained, along with matched control tissues from the same patients. IHC for the cancer testis antigens NY-ESO-1, MAGE-A3, and SSX2, was performed, and the staining results were scored by two authors based on maximal staining intensity on a scale of zero to three (absent=0, weak=1, moderate=2, or strong=3) and the percentage of tumor cells that stained. IHC for CD8 and CD3 was also performed, and T-cell tumor infiltration was defined as either brisk, nonbrisk, or absent, as described in melanoma literature. Concurrently, evaluation of 38 human DDLPS specimens and 10 healthy human fat specimens by the Nanostring nCounter platform is underway for identification of novel antigen targets and to establish the immune profile of DDLPS.

Immunohistochemical analysis of CTA expression showed considerable inter- and intra-tumoral heterogeneity. DDLPS showed relatively low expression of all CTAs tested, with only 22% of samples exhibiting MAGE-A3 and one sample each (3.1%) showing expression of SSX2 and NY-ESO-1 in low percentages of tumor cells. By contrast, in osteosarcoma, 74% of samples expressed MAGE-A3 and 68% expressed SSX, both with >80% of positive cases showing moderate to high expression. NY-ESO-1 was expressed in 41% of OS samples, predominantly at low levels. Brisk infiltration of CD8+ T cells was observed in over 70% of both sarcoma types tested. Furthermore, all sarcoma samples tested were positive for HLA expression.

To date, these results show promising expression of CTAs MAGE-A3 and SSX in OS, which may be used as targets in the future development of an immunotherapy for sarcoma. DDLPS shows relatively low expression, highlighting the need for more exploratory study with NanoString. The data generated throughout this project will provide insight into the immune profile of DDLPS.