Samuel Singer, MD

Soft tissue sarcoma is a heterogeneous disease with more than 50 histological subtypes, which have diverse biological behavior and – in many cases – unique genetics. Many sarcoma subtypes respond poorly to chemotherapy, so there is an urgent need for better treatment options for patients whose disease cannot be cured surgically. The major focus of my research is the discovery of genetic and epigenetic alterations in sarcoma to gain insight into sarcoma biology and find new targets for therapy.

To create a comprehensive picture of alterations in sarcoma, we have undertaken an integrated, genome-wide sequencing and microarray analysis, focusing on a few common sarcoma subtypes: myxofibrosarcoma, pleomorphic malignant fibrous histiocytoma, myxoid/round cell liposarcoma, and well-differentiated and dedifferentiated liposarcoma. The sequencing and microarray analyses are capable of detecting many types of alterations, including aberrant DNA methylation, mutations, gene fusions, novel transcripts, aberrant splice forms, and altered expression of mRNAs and microRNAs.

The tumor-specific alterations we discover are analyzed computationally to identify individual genes and pathways that are commonly affected in the tumors. The alterations are also tested for association with outcomes such as tumor recurrence and patient survival. The alterations that show associations can be used to build models to better predict individual patients’ outcomes. Finally, to assess the functional significance of individual alterations, the genes and microRNAs of interest are tested for effects on cell proliferation, apoptosis, and differentiation in vitro and tumor growth in vivo.

Through these methods, we have assembled the largest dataset of genomic alterations in sarcoma to date, and we have shown that some of these alterations may be useful therapeutic targets. In myxoid/round cell liposarcoma, these alterations include PIK3CA mutations, which are found in 18 percent of tumors and may be targeted with PI3 kinase inhibitors. In myxofibrosarcoma, we have identified NF1 mutations, which might be possible to target with MEK or mTOR inhibitors. In dedifferentiated liposarcoma, our results implicate several microRNAs and several effectors of cytokinesis (PRC1, PLK1), as well as CEBP-alpha, a regulator of differentiation. Interestingly, some of these genes, including those encoding CEBP-alpha and one of the microRNAs, appear to be downregulated by aberrant methylation. This suggests that demethylating agents may have a role in the treatment of dedifferentiated liposarcoma. Consistent with this, treating DDLS cells with a demethylating agent and the histone deacetylase inhibitor SAHA increased CEBPa expression, decreased proliferation, induced apoptosis, and reduced tumor growth in DDLS mice by 50 to 70 percent.

The results from our research will ultimately enable the identification of new targets based on a better understanding of the molecular genetics of solid tumors.

Structural rearrangements and DNA copy number alterations in retroperitoneal liposarcomas from two patients. A, a primary tumor; B, a local recurrence. Chromosomes are plotted in the outer ring with the centromeres indicated in red. DNA copy number data inferred from whole-genome sequencing is indicated in the inner ring with genomic amplifications highlighted (red). Structural rearrangements are edges between two indicated loci, either intra-chromosomal (light blue) or inter-chromosomal (dark blue). ⁽¹⁾ Epigenetic regulation of a microRNA through altered DNA methylation in dedifferentiated liposarcoma. A, global methylation patterns in dedifferentiated liposarcoma, which appear to differ little from those in normal adipose tissue. The panel shows the density of methylation on chromosome 16 in a tumor sample (blue) and its matched normal adipose tissue (gray). B, increased methylation near the miR-193b gene. Methylation in the tumor (DLSP1; solid blue, positive strands are dotted, and negative strands are dashed) and matched normal adipose tissue (gray) in the region of the putative promoter of miR-193b (as assessed by the position of enrichment of histone H3K4me3 in nine human cell types, indicated by horizontal lines). A CpG island is shown as a green bar. C, expression of miR-193b as a function of percentage methylation in the putative miR-193b promoter in well-differentiated liposarcomas (WD, green) and dedifferentiated liposarcomas (DD, blue; in red is DLPS1). The red curve is the regression, and the diagonal is indicated by a dotted line. Other analyses showed that miR-193b expression is significantly lower in dedifferentiated liposarcoma than in well-differentiated liposarcoma or normal adipose tissue (P<10^-9) and that many predicted miR-193b targets are overexpressed in dedifferentiated liposarcoma.⁽¹⁾