Search bioRxiv⌕ Search

Biology subjects

Gandhi, A. K.

Publications and source records attributed to Gandhi, A. K..

2 recordsLinked to original sources

Dysregulation of chromatin via H3K27 methylation underpins differentiation arrest in Isocitrate dehydrogenase-mutant Acute Myeloid Leukaemia.

Dysregulation of cellular differentiation is a hallmark of cancer. Isocitrate dehydrogenase (IDH) is commonly mutated multiple cancers including glioma, cholangiocarcinoma, lymphoma and Acute Myeloid Leukaemia (AML). Mutant IDH generates d-2-hydroxyglutarate that inhibits enzymes including Jumonji histone demethylases and TET2. Using primary human IDH2-mutant AML cells as a model, single cell RNA-seq and ATAC-seq, we demonstrated the continuum of cell states during restoration of neutrophilic differentiation to leukaemic progenitors by Enasidenib, a mutant IDH2 inhibitor. In cells which ultimately differentiate, there is co-expression of competing GATA2/RUNX3/SOX4-driven stem-progenitor and pro-differentiation EGR1/JUN/FOS programmes, followed by expression of cell cycle and terminal neutrophil programmes involving CEBP family and SPI1/PU.1. Genes upregulated during differentiation display loss of H3K27me3 in bivalent chromatin but not of H3K4me3, while downregulated genes are enriched for PRC2/EZH2 targets. In contrast to previous reports of a TET2-dependent mechanism for IDH-mutations, we observed only a modest link between promoter DNA CpG methylation and gene expression. For the first time in primary AML, we describe the lifting of differentiation block by de-repression of pro-differentiation genes through modulation of H3K27 demethylation in bivalent chromatin, and thus highlight a novel and important mechanism in how IDH mutations disrupt cell fates in cancer.

cancer biology↗

SpatialSort: A Bayesian Model for Clustering and CellPopulation Annotation of Spatial Proteomics Data

Emerging spatial proteomics technologies have created new opportunities to move beyond quantifying the composition of cell types in tissue and begin probing spatial structure. However, current methods for analysing such data are designed for non-spatial data and ignore spatial information. We present SpatialSort, a spatially aware Bayesian clustering approach that allows for the incorporation of prior biological knowledge. SpatialSort clusters cells by accounting for affinities of cells of different types to neighbours in space. Additionally, by incorporating prior information about cell types, SpatialSort outperforms current methods and can perform automated annotation of clusters.

bioinformatics↗