Genomic dissection of tumor heterogeneity and progression

Division Of Basic Sciences - Nci

Investigators

Linked publications, trials & patents

Abstract

Neurofibromatosis type 1 (NF1) is a common autosomal dominant disorder caused by an inactivating germline mutation in the tumor suppressor gene NF1, affecting ~1/3000 newborns6,7. NF1 is a cancer predisposition syndrome associated with highly elevated risk of multiple cancer types including tumors of the peripheral nerves8. NF1 nerve sheath tumors provide a unique model system for the study of how a tumor evolves over time and with therapy. Roughly one third of NF1 patients suffer from histologically benign but debilitating plexiform neurofibroma (PN) that are genetically driven by biallelic inactivation of NF1 and consequent activation of the MAP kinase pathway9. Work from the POB and others has highlighted an intermediate premalignant tumor broadly classified as an Atypical Neurofibroma (AN)10. The final step towards malignancy is a transformation to an MPNST that frequently arises from within PN or AN and are highly aggressive sarcomas with a dismal prognosis. These tumors readily metastasize and are highly resistant to radiotherapy and chemotherapy11. Beyond surgery for localized disease, there have been no successful therapies that cure MPNST to date. Therefore, the goal of project 1 is to detect and intercept MPNST at a point when surgery is still a curative option and to provide novel treatment options for those patients who progress to MPSNT. This project is framed into two aims, both designed with an eye towards translation of our findings into therapeutic or interventional clinical trials. Aim 1 seeks to leverage the rapid advances in sequencing technology and analysis of cell-free DNA (cfDNA) to develop methods to non-invasively assess disease status in NF1 patients. Here, we hypothesize that changes in cfDNA or circulating proteins can be used to detect the transformation of a benign precursor to MPNST. We advanced the goal of this project by building, validating, and patenting a cfDNA whole genome sequencing (WGS) assay capable of detecting MPNST in the blood of affected NF1 patients in some cases prior to radiographic evidence of a tumor. In addition, we have improved the assay by establishing the importance of DNA fragment size as a specific way to improve the assay and track regions of open chromatin in the blood from a corresponding tumor. Finally, we have changed the assay to now include whole genome methylation profiling using a method known as EMseq. This allows the same assay to in parallel have multiple features that can be derived and provides improved accuracy. We are currently exploring ways to move this test into the clinical space in the form of a laboratory developed test (LDT) so that we can test the usefulness of this assay in prospective patient studies. In parallel, we are leveraging the largest single cell atlas of NF1 nerve tumors assembled to date to examine genes and transcriptional programs that define cell populations unique to the tumor. This atlas was created from clinical specimens collected over multiple years on the NF1 natural history study and represents NF1 tumors across the benign to malignant state. This year, we have added a sensitive high-throughput proteomics assay of blood plasma that when informed by the single cell sequencing appears to be a highly accurate predictor of early malignant transformation. We have completed data analysis for this effort and have identified a 50 protein signature from the blood that with high accuracy predicts patients that have a MPNST versus those that have a benign plexiform neurofibroma. We are currently working on an integrated model using machine learning that will allow us to combine the cell free DNA assay with circulating proteomics and patient imaging studies. Aim 2 is focused on therapeutic targeting of a critical genomic transition point found in ~80% of MPNST, namely loss of function mutations in members of the transcriptional repressor Polycomb Repressive Complex 2 (PRC2). Here we hypothesize that the genetic loss of PRC2 results in a distinct epigenetic state that represents a therapeutic vulnerability. To accomplish this, we have used engineered model systems and cutting edge CRISPR-single cell functional transcriptional screens to elucidate the expression programs dysregulated by PRC2 and validated our findings using clinical samples. Interestingly, these studies have demonstrated that MPNST corrupts an early developmental program seen in Schwann cell precursors and discovered two potential therapeutic avenues for the treatment of MPNST. First, we are exploring the utility of targeting the Swi/Snf (BAF) transcriptional activation complex in PRC2 deficient MPNST. Interestingly, we have uncovered that SMARCA4 is a unique vulnerability in PRC2 deficient MPNST and that Swi/Snf is transcriptionally regulates a program responsible for DNA damage repair. A combination strategy with a Swi/Snf inhibitor in combination with the DNA damaging agent etoposide has been trialed in our preclinical model systems. A second strategy that we are exploring is targeting a cell surface protein, DLK1, that we have identified as upregulated as a consequence of PRC2 loss. We are currently testing an antibody drug conjugate and shown remarkable potency in our MPNST model systems. Currently, we are amending a first in-human study of this antibody drug conjugate to include patients with MPNST.