Portraits of clonal landscape and mutational signature in primary tumors and matched lung metastatic model of osteosarcoma

Objective: The development of lung metastasis following primary tumor diagnosis, resection and chemotherapy remain a significant hurdle in the treatment of osteosarcoma. Research indicates that the metastatic progression of osteosarcoma is driven by clonal evolution where selective pressure influences the emergence of distinct subpopulation of cells within the heterogenous tumor population. The absence of a robust in vivo model to accurately identify rare invasive subpopulations and determine how they evolved makes clonal characterization challenging. Methodology: To understand the clonal landscape and genomic architecture driving metastases, we injected barcoded OS17 PDX cells in 25 SCID mice intratibially to monitor primary tumor growth and clonally track the cells during metastatic progression. We hypothesize that tumor cells in the early stage undergo prolonged evolution in the primary tumors producing several distinct subpopulations that ultimately migrate, seed and becoming dominant in the lungs. Next, we performed barcode and deep-whole exome sequencing on 27 samples derived from the mice. We used the data to delineate clonal structure, assess somatic variants and identify mutational signatures and potential drivers of osteosarcoma. Results: In each primary tumor and matched metastases, we identified shared clonal drivers and somatic mutations. Clonal characterization revealed an increased clonal abundance that was followed by a significant reduction in clonal diversity in lung metastatic nodules. A high degree of similarity in the subclonal populations was observed between the lung nodules and the primary tumors. The frequency of the identified clonal drivers was higher in metastases compared to primary tumors. Copy number profile showed higher amplification peaks compared to deletion. C>T, T>C base substitution had the highest proportion compared to T>A. The most common mutational signatures were clock-like SBS1, SBS5 and SBS37. Metastatic nodules additionally harbored private SBS signatures absent in primary tumors. We identified 77 driver genes in the tumors. Shared clonal mutations ( KMT2C , THBS1 , SDHC ) dominated early stages and persisted through metastasis. Each group of primary tumor/metastasis shared both clonal(truncal) and subclonal mutation cluster that expanded in metastases. Shared subclonal variants ( ARAP3 , SIGLEC12 , FANCA , ERCC4 ) exhibited stage-specific diversification in metastasis and matched primary tumors. Primary-specific subclones ( RAD21L1 ) disappeared after the primary stage, while metastasis-specific subclones ( PTEN , BCOR ) emerged and dominated in metastases indicating a major evolutionary shift toward aggressive or invasive phenotypes. Conclusion: Clonal expansion of subpopulations in the late passages or metastases suggested that these clones can maintain tumorigenic potential in a favorable environment. The clonal expansion was probably due to mutation in the identified osteosarcoma driver genes. Multiple clones seeded metastasis and there was a direct relationship between the clonal and subclonal drivers in the primary tumors and metastatic lesions. Our model has important implications for the diagnosis and therapeutic treatment of osteosarcoma patients since multiple clones may need to be targeted to inhibit invasion.