Data concerning the fine structure of the 12q13-15 amplicon which contains MDM2 and CDK4 in well-differentiated and dedifferentiated liposarcomas (WDLPS/DDLPS) are scarce. We investigated a series of 38 WDLPS/DDLPS using fluorescence in situ hybridization analysis with 17 probes encompassing the 12q13-15 region. In addition, using quantitative RT-PCR we studied the expression of MDM2, CDK4, DDIT3 (CHOP/GADD153), DYRK2, HMGA2, TSPAN31 and YEATS4 (GAS41) in 11 cases. We showed that CDK4 (12q14.1) belonged to a distinct amplicon than MDM2 (12q15). There was no continuity in the amplified sequences between MDM2 and CDK4. Moreover, while MDM2 was amplified and overexpressed in all cases, CDK4 was not amplified or overexpressed in 13% of cases. The centromeric border of the CDK4 amplicon was located immediately downstream the 5 0 end of DDIT3, a gene known for being involved in myxoid liposarcoma translocations. DDIT3 was amplified in 3 cases and overexpressed in 9 cases. The overexpression of DDIT3 was correlated to the CDK4 amplification and not to its own amplification status. This suggested that the CDK4 amplicon, as well as the overexpression of DDIT3, might be generated by the disruption of a fragile region in 5 0 DDIT3. HMGA2 was always amplified and rearranged indicating that it plays a central role in WDLPS/DDLPS. HMGA2 rearrangement frequently resulted in a loss of the 3 0 end region that is a binding site for let-7. We also found a frequent amplification and overexpression of YEATS4, an oncogene that inactivates P53, suggesting that YEATS4 might play an important role together with MDM2 in WDLPS/DDLPS oncogenesis.

Purpose: The MDM2 and HMGA2 genes are consistently amplified in well-differentiated/ dedifferentiated liposarcomas (WDLPS/DDLPS) whereas CDK4 is frequently but not always amplified in these tumors. Our goal was to determine whether the absence of CDK4 amplification was (a) correlated to a specific clinico-histopathologic profile; and (b) compensated by another genomic anomaly involving the CCND1/CDK4/P16INK4a/ RB1/E2F pathway. Experimental Design: We compared the clinical characteristics of a series of 143 WDLPS/ DDLPS with amplification of both MDM2 and CDK4 (MDM2+/CDK4+) to a series of 45 WDLPS/DDLPS with MDM2 amplification and no CDK4 amplification (MDM2+/CDK4-). We used fluorescence in situ hybridization, real time quantitative reverse transcription PCR, and immunohistochemistry to explore the status of CCND1, P16INK4a, P14ARF, and RB1. Results: We found that MDM2+/CDK4-WDLPS/DDLPS represent a distinct clinical subgroup with favorable prognostic features, including low-grade lipoma-like histology, peripheral location, and lower rate of recurrence. By using fluorescence in situ hybridization, we found that genomic aberrations expected to be alternative mechanisms for compensating the lack of CDK4 amplification, such as RB1 and CDKN2A deletions or CCND1 amplification, were very uncommon. In contrast, by using real time quantitative reverse transcription PCR and immunohistochemistry, we observed that overexpression of P16INK4a (and P14ARF) and CCND1 and reduced expression of RB1 were very frequent, independently of the CDK4 status. Conclusions: Our results underscore the complex coordinated regulation of the RB and p53 growth-control pathways in WDLPS/DDLPS. Because the absence of CDK4 amplification is not specifically counterbalanced by a genomic alteration of the CCND1/CDK4/ P16INK4a/RB1/E2F pathway, CDK4 amplification may only represent a "MDM2-HMGA2-helper" in WDLPS/DDLPS tumorigenesis. (Clin Cancer Res 2009;15(18):5696-703)