Improving drug discovery using image-based multiparametric analysis of epigenetic landscape

Abstract

With the advent of automatic cell imaging and machine learning, high-content phenotypic screening has become the approach of choice for drug discovery because it can extract drug-specific multi-layered data, which could be compared to known profiles. In the field of epigenetics, such screening approaches have suffered from a lack of tools sensitive to selective epigenetic perturbations. Here we describe a novel approach, Microscopic Imaging of Epigenetic Landscapes (MIEL), which captures the nuclear staining patterns of epigenetic marks (e.g., acetylated and methylated histones) and employs machine learning to accurately distinguish between such patterns. We validated the fidelity and robustness of the MIEL platform across multiple cells lines using dose-response curves. We employed MIEL to uncover the mechanism by which bromodomain inhibitors synergize with temozolomidemediated killing of human glioblastoma lines. To explore alternative, non-cytotoxic, glioblastoma treatment, we screen the Prestwick chemical library and documented the power of MIEL platform to identify epigenetically active drugs and accurately rank them according to their ability to produce epigenetic and transcriptional alterations consistent with the induction of glioblastoma differentiation.

Understanding the Mechanisms by Which Epigenetic Modifiers Avert Therapy Resistance in Cancer

“…Thereby, treatment with epigenetic drugs elicits a “specific” effect on cancerous cells by reverting these unique expression changes. Additionally, sensitivity to epigenetic modifiers can be genomic loci specific, possibly due in part to the three-dimensional chromatin structure ( 21 – 23 ). Thus, epigenetic modifiers possess the unique ability to be effective in a broad category of patients; albeit via altering the expression of a set of genes in a patient- specific manner ( 24 ).…”

High-content analysis and Kinetic Image Cytometry identify toxic and epigenotoxic effects of HIV antiretrovirals on human iPSC-neurons and primary neural precursor cells

“…The plates were then scanned using Vala's IC200 image cytometer, and the images for each channel were analyzed via digital image analysis algorithms that first identify the nuclei of each cell; the patterns of marks within each nucleus are then analyzed creating multivariate signatures or "finger-prints" of each perturbation. Texture features [46][47][48] are used rather than intensities and morphologies, thus reducing culturing and immunolabeling artifacts. The results from discriminant analysis, plotted in 2D, are shown in this report to illustrate shifts in the epigenetic signatures in response to test compounds.…”

“…We then scanned the cells using Vala's IC200 Image Cytometer and analyzed the images to identify the multivariate epigenetic signatures of each histone modification in each condition. We characterized the epigenetic signatures using texture features 56,65,66 rather than intensities and morphologies to reduce culturing and immunolabeling artifacts. Quadratic discriminant analysis reports changes in the epigenetic signatures in response to each treatment.…”

“…Confusion matrices report the ability of the multiparametric discriminant analysis to correctly classify each sample with other samples that received the same treatment compared to other treatments (Fig. 7D, E) 56 . The confusion matrix for both H3K27me3/H3K27ac and H3K9me3/H3K4me1 show that classification occurred correctly 100% of the time for samples treated with 10 µM GSK343, 3 µM SAHA, and 10 µM TDF.…”

“…We also developed an assay to test ARVs for potential effects on neurogenesis, a process by which human neural precursor cells (hNPCs) differentiate to neurons to support central nervous system development and cognitive function in adults 54,55 . For this, we used the MIEL assay, a multiparametric approach to identify changes in histone acetylation and methylation patterns using HCA of immunofluorescence images 56 .…”