2009 Innovative Research Grant 42-Month Progress Reports
Modeling Ewing Tumor Initiation in Human Neural Crest Stem Cells
Elizabeth R. Lawlor, M.D., Ph.D., University of Michigan
Many pediatric tumors develop when stem cells or other fetal cells develop alterations in their DNA leading them to differentiate into cancerous rather than normal tissues. Alterations can be induced by mutations that disrupt the DNA or by so-called epigenetic processes that alter gene expression without changing DNA sequence. In this project we used novel stem cell models and innovative technologies to study how changes in DNA methylation contribute to the origin and progression of Ewing sarcoma (ES).
ES is believed to arise from neural crest stem cells (NCSC). In Aim 1 we defined what DNA methylation should look like before, during and after these stem cells differentiate. Although changes in methylation occurred during differentiation, most did not occur in the regulatory regions of genes, so-called gene promoters. Instead, most changes to DNA methylation occurred in the regions that lie between genes. In contrast, the pattern of DNA methylation in ES cells differed dramatically and gene promoters were found to be selectively hypermethylated. Thus, DNA methylation in ES deviates from both normal NCSC as well as their normal progeny.
ES arises from cells that acquire a genetic mutation, termed EWS-FLI1. Although EWS-FLI1 can initiate the process of malignant transformation it is not sufficient to create a tumor. In Aim 2 we investigated whether NCSC that express EWS-FLI1 acquire changes to DNA methylation. Our studies showed that the methylation profiles of EWS-FLI1+ stem cells were abnormal and evolved over time. In particular, the promoters of genes that control embryonic development became progressively more methylated over time and the global profiles of EWS-FLI1+ cells evolved to resemble those of fully malignant tumor cells. Evaluation of DNA methylation in ES tumor biopsies confirmed that the patterns of abnormal methylation that we detected in EWS-FLI1+ stem cells and ES cell lines were present in patient tumors and were not simply a consequence of cell culture. Analysis of gene expression data confirmed that abnormal methylation was associated with abnormal gene expression. In particular, the normal regulation of genes involved in early embryonic development was disrupted in EWS-FLI1+ stem cells and tumor cells. Thus, our studies in Aim 2 confirm that abnormal epigenetic regulation of gene expression, via abnormal DNA methylation, is a critical event in the creation of ES and is driven by the EWS-FLI1 oncogene.
In Aim 3 we tested whether the growth of ES tumors in mice could be blocked by a drug that inhibits DNA methylation or by turning off the EWS-FLI1 oncogene. Significantly, both interventions slowed tumor growth demonstrating that the oncogene and abnormal DNA methylation both serve as key drivers of ES tumor progression.
In summary, our studies together confirm that abnormal DNA methylation and disrupted epigenetic regulation of embryonic gene expression is a key driver of ES initiation and progression. These studies support introduction of epigenetic modifying agents into ES treatment protocols.
Cancer Cell-Specific Self-delivering Prodrugs
Matthew Levy, Ph.D., Albert Einstein College of Medicine of Yeshiva University
We sought to develop a novel system for targeted delivery of drugs that localize specifically to tumor cells in whole animals. To do this, we made use of an alternate class of targeting molecules composed of nucleic acids called aptamers. These cell type specific aptamers were composed of or linked to cytotoxic nucleoside analog drugs. Following internalization within cancer cells, breakdown of the aptamers should result in release of the drug molecules. In this way, drugs will preferentially accumulate in cancer cells, which we believe will translate into minimization of off-target adverse side effects on healthy tissue while maintaining or even enhancing the efficacy of these drugs on diseased cells.
Identifying Solid Tumor Kinase Fusions Via Exon Capture and 454 Sequencing
William Pao, M.D., Ph.D., Vanderbilt University Medical Center
Cells rely upon molecular switches to carry out normal functions. One class of switches is called kinases. In a highly simplified model, when kinases are ‘on’, cells divide; when kinases are ‘off’, cells stop dividing. In some cancers, kinases have become altered, leading to abnormal signaling. One major type of alteration is called a ‘fusion’. A fusion ‘fuses’ together a part of a cellular protein (that normally has another function in the cell) with the signaling portion of the kinase molecule. Instead of turning ‘off’ and ‘on’ in a tightly regulated manner, kinase fusions are ‘on’ all the time, tricking cells into constantly dividing. If one blocks the abnormal signals from a kinase fusion with a drug, cancer cells can die, and patients can enormously benefit. When we started this project, only a limited number of TK fusion proteins had been identified in solid tumors, because experimental procedures for their identification were inadequate.
As part of this grant, we developed a novel method to identify kinase fusion proteins using DNA from any type of tumor. We demonstrated the feasibility of this platform on tumor cell lines (Chmielecki et al., Nucleic Acids Res, 2010) and then used the results of the pilot study to create an improved platform with streamlined workflow and computational analysis. Using the newer kinase fusion design, we screened multiple tumor samples and identified a novel PDGFRbeta fusion from one patient sample (Chmielecki et al., Genes Chromosomes Cancer, 2011). Since then, commercial entities like Foundation Medicine have incorporated a similar approach to detecting kinase fusions from formalin-fixed paraffin-embedded tumor samples; their FoundationOne test is now used routinely for patient care.
During the course of this grant, next-generation sequencing platforms were constantly improving. Thus, we subsequently used multiple different sequencing platforms to identify novel actionable targets for therapy. These included projects in non-small cell lung cancer, small cell lung cancer, and melanoma.
Probing EBV-LMP-1’s Transmembrane Activation Domain with Synthetic Peptide
Hang (Hubert) Yin, Ph.D., University of Colorado at Boulder
The human herpesvirus Epstein-Barr virus (EBV) infects and immortalizes naïve B lymphocytes and is associated with a number of human malignancies. Understanding the signaling mechanisms by which EBV gene products hijack existing B cell signaling pathways to gain access to B cell memory is crucial for designing rational strategies for the prevention and treatment of EBV-dependent malignancies. One of the key players for EBV activation is LMP-1, a viral oncoprotein that activates B cell signaling pathways promoting survival and proliferation. LMP-1 functions as a ligand-independent Tumor Necrosis Factor Receptor (TNFR) superfamily “analog”, most closely resembling the TNFR family member CD40. Unlike CD40, LMP-1 is constitutively active. LMP-1’s activity depends on the homo-oligomerization of its hydrophobic multi-spanning transmembrane domain. We have developed antagonistic peptides (anti-TMD-5) specific for LMP1’s fifth transmembrane domain (TMD-5) to use as probes to study LMP-1 signaling. Currently, we attempt to optimize the anti-TMD-5 sequences by combining rational design and the directed evolution. Next, we report a small molecule agent, NSC 259242,to be a TMD-5 self-association disruptor. Both the positively charged acetimidamide functional groups and the stilbene backbone of NSC 259242 are essential for its inhibitory activity. Two dimensional heteronuclear multiple quantum coherence (2D-HMQC) NMR titrations showed that NSC 259242 recognizes the TMD-5 helix and disrupts TMD-5 homo-trimerization. Furthermore, cell-based assays revealed that NSC 259242 inhibits full-length LMP-1 signaling in EBV infected B cells. Most recently, we engineered an EBV-infected 721 B cell line that responds to LMP-1 activation by expressing green fluorescent protein (GFP) using the commercially available pGreenFireTM NF-κB reporter lentivectors. Reiterative screening by this method produced novel traptamer sequences that potently inhibit LMP-1 signaling. These studies demonstrated a new strategy for identifying novel anti-TMD peptide/small molecule inhibitors for investigating transmembrane protein-protein interactions and ultimately as therapeutic agents in the treatment of EBV-dependent lymphoproliferative disease.