(P 411) A SPATIAL TRANSCRIPTOMIC "GOOGLE MAP" OF HUMAN OSTEOSARCOMA REVEALS COMMON CELL CLUSTERS AT THE PRIMARY SITE AND PULMONARY METASTASES.

Objective: Osteosarcoma (OS) is the most common primary malignant bone tumor with poor prognosis. As the survival rate for patients with metastases still remains at only 20%, new research approaches are needed to better understand the mechanisms of their progression. Through spatial transcriptomic analysis of two pairs of matching primary and lung metastases (LM) of osteosarcoma obtained from two patients, we aimed to describe the identity, localization, transcriptomic profiles, and developmental interplay of tumor and non-tumor cells and of cells residing at the tumor-non tumor interface.

Methods: Samples were obtained from a 19yo female (Patient 1-P1) and a 15yo male (Patient 2-P2) diagnosed with high-grade conventional osteosarcoma. Both patients received neoadjuvant chemotherapy: 20% tumor necrosis was achieved in P1, and 92% tumor necrosis was achieved in P2. Upon collection, samples were fixed for 28h in PFA 4%, embedded in paraffin, and sectioned at 5µm. H&E staining was performed to evaluate tissue integrity and to select target areas containing both tumoral and non-tumoral tissue. RNA degradation of the tissue samples was assessed by means of DV200 (DV200 ≥ 50%). Selected sections were processed with 10x Genomics Visium Spatial Gene Expression for FFPE samples® (10X Genomics, Inc. USA) according to the provided protocol.

Results: The unbiased cluster analysis of all samples identified 12 major clusters, including clusters of osteogenic, immune, epithelial, and endothelial cells, and adipocytes. Boundaries of the clusters overlapped with the edges of the histological regions identified by the pathologist (Fig. 1 A, B). Three clusters of malignant osteosarcoma cells (MOCs) presenting an osteogenic-like phenotype were detected in the tumoral areas (Fig. 1B, C): MOCs-1 was found in all primary sites and pulmonary metastases; MOCs-2 was observed in the primary OS of P1 and in LM of both patients; and MOCs-3 was found only in the primary OS of P2 (Fig. 1B, C). Pathway analysis of MOCs-1 revealed a significant decrease in apoptosis and cell death signaling along with a pronounced increase of cell invasion and migration signaling. Similar observations were made for MOCs-2. In addition, a comparison between MOCs-1 of the primary OS cells vs MOCs-1 of the LM revealed that MOCs-1 of the metastases express higher levels of genes typically present in the lung (SFTPB, SFTPC) (Fig. 2A) which are absent in primary OS. Similar results were obtained by analyzing MOCs-2.

Immune cell clusters were found mainly in the LM localized at the tumor-lung interface in both patients, while primary tumors did not harbor significant immune infiltrates. Low expression of CD8, cytotoxic T-lymphocyte antigen 4 (CTLA4), and programmed cell death 1 (PD-1) was detected in all samples, indicating a limited T-cell infiltration within the tissues. Also, PD-L1 was minimally expressed (Fig. 2B). However, by looking at the gene expression profile of MOCs-1 and MOCs-2, we found remarkable expression of genes involved in carcinogenesis, cell transformation, and cell adhesion (YAP1, GREM1, and CDH11).

Conclusion: Our data show that the MOCs-1 and MOCs-2 of LM display expression of lung-specific genes while maintaining a high level of similarity with MOCs-1 and MOCs-2 of primary OS, suggesting that these genes may be responsible for the metastases’ onset and development within the lungs. The lack of significant T-cell infiltrates in all samples may explain the limited efficacy observed in OS clinical trials targeting the PD-1/PD-L1 pathway. YAP1, GREM1, and CDH11 being potentially targetable, may identify novel therapeutic avenues.

This study, for the first time, presents a spatial transcriptomic analysis of primary and metastatic lesions of human OS, in the form of a consultable “Google Map”, offering the opportunity to retrieve useful information about the localization, characteristics, and developmental dynamics of the cells of human OS.