(P 156) PROFILING SYNOVIAL SARCOMA EXTRACELLULAR VESICLES FOR DIAGNOSTIC RNA BIOMARKERS

Objective: Synovial sarcomas (SS) are an aggressive class of cancers, with 24-64% of patients dying within 5 years of diagnosis. Poor survival rates are largely due to diagnostic delays resulting from a lack of non-invasive diagnostic tools and biomarkers. The SS18-SSX fusion is a hallmark of SS and an important biomarker for diagnosis from biopsy samples. However, there are no non-invasive methods that can detect the fusion due to the limited sensitivity of PCR-based methods, the degradation of SS18-SSX fusion molecules circulating in the blood, and the heterogeneity of the fusion breakpoints and isoforms. Extracellular vesicles are secreted by all cells, including cancer cells, into bodily fluids, such as blood. Extracellular vesicles are exciting substrates for non-invasive cancer detection, whereby the presence and type of cancer could be detected from the patient’s blood’s circulating nucleic acid, avoiding invasive biopsy and complicated imaging procedures. Key advantages of secreted tumor RNAs as liquid biopsy substrates are that these RNA species are protected from degradation by vesicle membranes and are representative of the dynamic tumor states and tumor heterogeneity. They are also present in higher abundance than circulating tumor DNAs, potentially allowing for better tumor detection sensitivity. We hypothesized that profiling tumor RNA molecules secreted in extracellular vesicles by synovial sarcoma cells could identify RNA biomarkers for non-invasive diagnostics.

Methods: To capture secreted RNAs comprehensively, we utilized an innovative combination of long- and short-read sequencing to obtain the first complete profile of secreted RNAs in SS cells. We used a panel of four fusion-positive SS cell lines obtained from the Hoang Lab at Montefiore Einstein: ASKA, SYO-1, HSSY-II, and YAMATO. The SS cells were grown in culture with both cells and media collected for the extraction of cellular and extracellular vesicle RNA (evRNA), respectively. This method uses cells grown in media as a proxy for body fluids with tumor molecules circulating. All RNA underwent quality control (QC) checks and were used to prepare short-read Illumina and long-read nanopore libraries using different capture methods (polyA selection and ribosomal depletion) for sequencing. To assess the quality of RNA sequencing (RNA-Seq) data, all RNA-Seq data underwent QC checks quantifying the number of mapped, exonic, non-duplicate (MEND) reads. We consider RNA-Seq samples with at least 10 million MEND reads to be of sufficient quality to conduct robust gene expression quantification. To determine if SS18-SSX fusion transcripts were packaged into extracellular vesicles, we ran fusion caller pipelines and characterized the different isoforms expressed.

Results: We successfully extracted evRNA from all the SS cell lines profiled. For total cellular RNA extracted from each cell line, the majority of the RNA fragments fell between 2000-5000 nucleotides; for the total evRNA extracted from each cell line, we saw that the majority of RNA fragments were between 25-200 nucleotides in length. This is consistent with the literature, which shows that the majority of RNAs that get packaged into extracellular vesicles are small non-coding RNAs and fragmented tRNAs and rRNAs. However, for the evRNA, we detected that each cell line's RNA population had 25-44% of fragments larger than 200 nucleotides, which may contain mRNA biomarkers for detection. We created polyA Illumina, ribo-deplete Illumina, and cDNA nanopore sequencing libraries for each cell line's cellular RNA and ribo-deplete Illumina libraries for evRNA from each cell line. All cellular RNA-Seq samples passed QC checks with greater than 10 million MEND RNA-Seq reads and could be used for robust gene expression quantification and comparisons. The resulting libraries made from evRNA were of poorer quality compared to cellular RNA libraries because the large proportion of fragmented RNA made it difficult to construct sequencing libraries. We confirmed that SS18-SSX fusion transcripts are present in the cellular RNA of all four cell lines and packaged into extracellular vesicles released by three of the four cell lines with 12, 17, and 3 fusion junction reads, respectively. Follow-up experiments will be done to construct new ribo-deplete Illumina libraries for evRNA as well as polyA Illumina and long-read nanopore libraries for evRNA samples.

Conclusion: Using synovial sarcoma cancer cell lines, we have confirmed that SS18-SSX fusion transcripts are packaged into extracellular vesicles and can serve as potential biomarkers for detection. This is a significant discovery that sets the stage for the first non-invasive SS diagnostic assay from patient blood.