Molecular Features of Osteosarcoma

The number of structural variants observed for osteosarcoma is high, at more than 200 structural variants per genome.[1,2] Thus, osteosarcoma has the most chaotic genome among childhood cancers. The Circos plots shown in Figure 7 illustrate the exceptionally high number of intra- and inter-chromosomal translocations that typify osteosarcoma genomes.

The tumor mutational burden (TMB) for children and adolescents with osteosarcoma is approximately 2 mutations per megabase and is higher than that of some other childhood cancers (e.g., Ewing sarcoma and rhabdoid tumors).[1,2] However, this rate is well below that for adult cancers such as melanoma and non-small cell lung cancer, which are responsive to checkpoint inhibitors.

Rather than activating variants in oncogenes and inactivating variants in tumor suppressor genes, as observed in many cancer types, the genomic landscape for osteosarcoma is driven by copy number gain/amplification in chromosome regions that include oncogenes and copy number loss (deletions) in chromosome regions that include tumor suppressor genes. Recurring copy number gains and losses that affect known oncogenes and tumor suppressor genes, respectively, are described below.Estimates of the frequency of specific genomic alterations in osteosarcoma vary from report to report. This finding could be a result of different definitions being used to define copy number alterations, different methods being used for their detection, or differences in tumor biology across patient populations (e.g., newly diagnosed versus relapsed, localized versus metastatic, or pediatric versus adult).

Genomic alterations in TP53, leading to loss of TP53 function, are present in most osteosarcoma cases.[1] A distinctive form of TP53 inactivation occurs through structural variations in the first intron of TP53 that lead to disruption of the TP53 gene.[1] Other mechanisms of TP53 inactivation are also observed, including missense and nonsense variants and deletions of the TP53 gene.[1,2] The combination of these various mechanisms for loss of TP53 function leads to its biallelic inactivation in most cases of osteosarcoma. Because many of the structural variations leading to TP53 inactivation are best detected through whole-genome sequencing, results based on clinical genomic testing panels may show lower rates of TP53 alterations because they do not detect these changes.[3]

MDM2 amplification, which is another genomic alteration that leads to loss of TP53 function, is observed in a minority of osteosarcoma cases (approximately 5%).[1-4]

RB1 is commonly inactivated in osteosarcoma, sometimes by deleterious variants but more commonly by chromosomal deletion of the chromosome 13q14 region that includes RB1.[1,2,5]

Chromosomal deletions involving chromosome 9p21 lead to CDN2A deletion in approximately 20% of osteosarcoma cases.[1,2,5]

Among tumor oncogenes, MYC at chromosome 8q24 shows gain/amplification in approximately 10% of patients.[3,5,6] In one study, MYC gain/amplification appeared to be associated with inferior prognosis. In a second study, MYC gain/amplification was enriched in children, compared with adults.[6]

CCNE1 at chromosome 19q12 is another tumor oncogene that shows gain/amplification in approximately 10% of patients.[3,5,6] Other oncogene-containing chromosomal regions showing chromosomal gain/amplification in a minority of osteosarcoma cases include the CDK4-harboring region at chromosome 12q14,[4,5,7] the VEGFA- and CCND3-harboring regions at chromosome 6p12,[3-5,7] the CCND1-harboring region at chromosome 11q13,[4] and the PDGFRA-, KIT-, and KDR-harboring regions at chromosome 4q12.[3-5]

Alternative lengthening of telomeres (ALT) is the telomere maintenance mechanism employed by the majority of osteosarcoma tumors.[1,8,9] ATRX inactivating variants and gene deletions are associated with the ALT telomere maintenance mechanism. ATRX genomic alterations are present in a subset of osteosarcoma tumors that use this telomere maintenance mechanism.[1,3,9]

Many of the genomic alterations reported for osteosarcoma tumors at diagnosis do not provide obvious therapeutic targets, as they reflect loss of tumor suppressor genes (e.g., TP53, RB1, PTEN) rather than activation of targetable oncogenes. In addition, there has been limited success across cancer diagnoses in using gains/amplifications of the oncogenes relevant to osteosarcoma to identify patients that may benefit from targeted therapy.