Abstract PR003: NF1 loss disrupts a SMARCA2–CTCF–H1 chromatin-remodelling axis and creates DNA repair vulnerability in retinoic acid–resistant high-risk neuroblastoma

Abstract

Neuroblastoma (NB) is an aggressive childhood malignancy arising from the developing sympathetic nervous system. NF1 loss-of-function (LoF) and activating-RAS mutations both result in downstream activation of RAS-pathway signalling, are enriched at relapse, and associated with differentiation block. We show that failure to downregulate RAS-MEK-ERK signalling in both NF1 mutant and RAS mutant NB cells is associated with resistance to the standard of care differentiation therapy: retinoic acid (RA). Given the role of SWI/SNF and PRC2 chromatin-remodelling complexes in RA-mediated differentiation, we predicted SWI/SNF, PRC2, and RAS pathway functional interactors, using publicly available datasets and then evaluated RNA and protein expression of these key interactors across large panel of NB cell lines. These included patient-derived NF1 wild-type and NF1 mutant cell lines, CRISPR–Cas9-generated NF1-knockout cells and an inducible NRAS Q61K overexpression model. We identified consistent downregulation of the SWI/SNF ATPase subunit SMARCA2 at both transcriptomic and proteomic levels across all NF1-deficient models. However, SMARCA2 expression remained unchanged following induction of NRAS Q61K expression. NF1-knockout cells were then treated with the MEK inhibitors (MEKi) trametinib and cobimetinib to suppress downstream phospho-ERK, but this failed to restore SMARCA2 expression. Taken together this identifies that although RA resistance is a common phenotype of oncogenic RAS-pathway signalling in NB, SMARCA2 downregulation is specific to NF1 LoF and occurs via a mechanism that is independent of canonical RAS–MEK–ERK signalling. We then evaluated the chromatin bound proteome by mass spectrometry in 1. matched NF1 wild-type and knock-out cells and 2. matched RAS wild-type and NRAS Q61K expressing cells. This confirmed NF1 loss–specific downregulation of SMARCA2 on chromatin. NF1 knock-out cells also showed significant down-regulation of chromatin bound DNA replication and repair machinery proteins in addition to loss of CTCF and linker histone H1 variants. Immunoprecipitation for the core SWI/SNF subunit SMARCC2 confirmed reduced enrichment of CTCF and H1 proteins, together with SMARCA2 loss, in NF1-knockout cells. These findings support disruption of a SMARCA2–CTCF–H1 chromatin architecture axis in NF1 mutant NB. Given our findings, we hypothesised that the combination of PARPi and MEKi would be synergistic in NF1-deficient NB. In vitro drug synergy studies revealed the combination of the PARPi niraparib and the MEKi cobimetinib to be highly synergistic in NF1-mutant, but not NF1 wild-type or NRAS Q61K overexpression models. In-vivo studies are ongoing. In summary, we identify NF1 loss–specific chromatin-remodelling resulting in impaired DNA-damage repair and show that PARP–MEKi combinations are a potential therapeutic strategy for NF1 mutant high-risk NB. This study addresses a critical unmet need in NF1-mutant high-risk NB, particularly at relapse or progression, where treatment options are limited (supported by ICR HEFCE and CRUK CSF).