(P 242) LIPOSARCOMAS WITH 12Q AMPLIFICATION - CAN GENETIC CHANGES DISCLOSE THE DEVELOPMENT AND ORIGIN OF RING CHROMOSOMES?

Objective: Liposarcoma is the most common soft tissue sarcoma in adults, accounting for about 20 % of all mesenchymal malignancies. The most common subgroup comprises liposarcomas harbouring giant marker and ring chromosomes. In addition, amplification of chromosome arm 12q, including genes such as MDM2, HMGA2 and CDK4 is another characteristic of this tumour subgroup. The subgroup comprises well-differentiated liposarcomas (WDLS) occurring in the extremities or retroperitoneum with a low-grade behaviour and the high grade dedifferentiated liposarcomas (DDLS). Despite the disparate biological behaviour, both WDLS and DDLS are closely related pathogenetically. A subset of WDLS display the capacity to transform into the highly aggressive DDLS. It is well known that WDLS arising in the extremities rarely become aggressive, while those arising in the retroperitoneum often transform into higher-grade sarcomas. However, little is known about this process and the role and origin of the ring chromosomes.
The aim of this study is to establish whether the risk for malignant transformation and aggressive behaviour of a subset of well-differentiated liposarcomas can be discerned already at diagnosis. The focus here is on single nucleotide variants (SNVs) found in these tumours.

Methods: The cohort of this study consists of 26 samples from 24 patients including well-differentiated liposarcomas (WDLS) located in the limbs or retroperitoneum, dedifferentiated liposarcomas (DDLS) and tumours diagnosed as other sarcomas but showing the characteristic ring chromosomes and 12q amplification. The copy number profile was assessed by SNP array (Affymetrix Cytoscan HD) and whole genome sequencing (30x or 90x). Whole exome sequencing (100x) was performed to obtain information on SNVs.

Results: All cases showed amplification on chromosome 12. Co-amplification of other chromosomal segments was seen in most cases, notably from Chr1 (54 %) and Chr6 (38 %).
A total of 419 non-synonymous somatic, exonic SNVs were detected in 26 tumor samples from 24 patients. The median number of SNVs from all subtypes was similar, with 12 for WDLS in the limbs (range 6-25), 14.5 for WDLS in the retroperitoneum (range 11-18), 14 for DDLS (range 7-36) and 16 for the sarcomas diagnosed as other types (range 9-48). A total of 19 recurrently mutated genes were identified, such as SLAMF7, OTOGL, ATP2B1 and MUC16. None of these genes were mutated in more than two samples. However, three recurrently mutated genes located on chromosome arm 12q (NCKAP1L, AGAP2, LYZ) were detected in a primary and local recurrence sample from the same patient. In relation to the copy number status, 27 SNVs were located in amplified regions (copy number > 5); 14 of these were located on Chr12 and 10 on Chr1. None of the detected SNVs has any previously known role in tumorigenesis.

Conclusion: Whole exome sequencing did not identify any potential driver mutations. All subtypes showed a similar small median number of mutations implying that SNVs do not play a significant role in tumour progression. Furthermore, there was no indication that significant SNVs precede the amplifications associated with ring chromosome formation. Moreover, there was no obvious clustering of SNVs to amplified regions, suggesting that the amplicon process is not associated with concomitant mutations at the nucleotide level. Thus, further studies aiming at identifying the early steps in ring chromosome formation, and the pathways involved in tumour progression, should focus on the chromosomal rearrangements leading to ring chromosome formation as well as transcriptomic features associated with disease progression.