Pathogenic germline variation in TP53, usually accompanied by familial Li-Fraumeni syndrome (LFS), is known to underly a small proportion of pediatric osteosarcoma (OS). However, the extent to which de novo mutations and rare but mildly deleterious variation in TP53 are found in the genomes of OS patients is only now coming into focus with next-generation sequencing. We recruited 285 patients with OS diagnosed at <20 years of age from North American Children's Oncology Group institutions between December, 2007 and March, 2010. Oragene saliva collection kits (DNA Genotek, Ottawa, Canada) or buccal swabs for DNA extraction were collected from patients and their parents, resulting in 229 complete triads and 59 dyads. From these, 95 triads were selected for whole exome sequencing (WES) based on the availability of both patient and parent samples and sufficient DNA mass. Libraries were created using the Agilent SureSelect All Exon V5+UTRs kit and were sequenced using a HiSeq 2000 that generated 100bp paired end reads with an average insert size of 188.8bp. We implemented the best practices as delineated in the GATK pipeline, including using BWA-MEM for alignment, GATK for quality recalibration and indel realignment, and GATK HaplotypeCaller for genotyping. On average, 35.0 million reads were delivered per sample, and the average read depth was 36.4x coverage. The average proportion of TP53 exons covered at 20x coverage was 76.2% (compared to 78.7% for all genes). Exons 1 and 9 were the only exons with poor coverage (mean coverage <20x). All putative TP53 variants were verified via Sanger sequencing; only verified variants are reported hereafter. Sequence variants were annotated using SNPEFF and ANNOVAR. Five patients exhibited rare, synonymous variation in exons 4 (Pro36Pro, 2 patients) and 6 (Arg213Arg, 3 patients). One patient exhibited rare, non-synonymous variation in exon 4 (Pro47Ser); this variant has a frequency of 0.005 across all populations in 1000 Genomes data and was predicted to be deleterious in 2 of 8 functional prediction algorithms. An additional two patients exhibited rare, non-synonymous variation in exons 7 (Cys238Tyr) and 8 (Gly279Glu). Neither variant was detected in parents or reported in 1000 Genomes data, and both were predicted to be deleterious by all functional prediction algorithms. The prevalence of pathogenic TP53 variation in our data (3/95 or 3.2%) is roughly similar to that reported previously, but importantly 2 of the 3 rare, non-synonymous variants were de novo. These results underscore the contribution of new mutations to OS risk and suggest that genetic counseling and TP53 mutation testing should be offered to patients even in the absence of a family history suggestive of LFS. This research was funded by NIH R01CA122371 and grants from the Hyundai Hope on Wheels foundation and the Zach Sobiech Osteosarcoma Fund of the Children's Cancer Research Fund, Minneapolis, MN.

Citation Format: Logan Spector, Ian Lock, John Lane, Aaron Sarver, Mark Krailo, Rajaram Nagarajan, Nathan Pankratz. De novo and transmitted germline TP53 variation in pediatric osteosarcoma: A report from the Children's Oncology Group. [abstract]. In: Proceedings of the AACR Special Conference on Advances in Pediatric Cancer Research: From Mechanisms and Models to Treatment and Survivorship; 2015 Nov 9-12; Fort Lauderdale, FL. Philadelphia (PA): AACR; Cancer Res 2016;76(5 Suppl):Abstract nr A37.