Discovering a predictive metabolic signature of drug-induced structural cardiotoxicity

Bowen, Tara Jane (2023). Discovering a predictive metabolic signature of drug-induced structural cardiotoxicity. University of Birmingham. Ph.D.

Bowen2023PhD.pdf
Text - Accepted Version
Restricted to Repository staff only until 1 December 2024.
Available under License All rights reserved.
Download (13MB) | Request a copy

Abstract

Drug-induced structural cardiotoxicity is a leading cause of attrition and project failure in drug discovery and development, limiting the delivery by the pharmaceutical industry of safe and efficacious medicines. Overcoming this costly challenge requires comprehensive mapping of toxicological mechanisms and discovery of predictive, translational biomarkers, such that cardiac safety assessment of lead compounds and preclinical to clinical translation may be improved. Untargeted metabolomics is an established approach in toxicology for characterising endogenous metabolic responses to xenobiotic exposure, facilitating toxicological mode of action and metabolic biomarker discovery. With efforts to overcome challenges faced in the analysis of more advanced in vitro models, e.g., cardiac microtissues, and in the biological interpretation of data due to often low coverage and low confidence in metabolite annotation, untargeted metabolomics may help to deliver the unmet need. The approach is also capable of providing deep insights into xenobiotic fate, although this capability is yet to be fully demonstrated.
The overarching aim of this thesis was to develop and apply untargeted metabolomics workflows to characterise the fate and effects of xenobiotics in model systems, ultimately, to discover a molecular signature predictive of drug-induced structural cardiotoxicity. First, a workflow to discover and analyse xenobiotic-related measurements from routine ultra-high performance liquid chromatography-mass spectrometry (UHPLC-MS) untargeted metabolomics datasets was assembled. Implementation of this workflow clearly demonstrated how untargeted metabolomics can discover extensive xenobiotic biotransformation maps, temporally changing relative systemic exposure, and direct associations of endogenous biochemical responses to the internal dose.
Next, the need for a workflow enabling untargeted metabolomics analysis of cardiac microtissues was addressed. A filtration-based approach for harvesting microtissues was evaluated and the sensitivity and reproducibility of nano-electrospray direct infusion mass spectrometry (nESI-DIMS) measurements of intracellular extracts were assessed, revealing samples consisting of 28-pooled microtissues, harvested by filtration, are suitable for profiling the intracellular metabolome and lipidome. Subsequently, an extensive workflow combining nESI-DIMS untargeted metabolomics and lipidomics of intracellular extracts with complementary UHPLC-MS analysis of spent culture medium was implemented. The acquired data enabled investigation of time-resolved metabolic and lipid perturbations in cardiac microtissues induced by a structural cardiotoxin, demonstrating the workflow can facilitate toxicological mode of action and biomarker discovery.
To address the annotation challenge of untargeted metabolomics data and support biological inference, a library of 315 polar metabolites and 2350 lipids present in cardiac microtissue intracellular extracts and detectable by UHPLC-MS was curated. When used in place of larger, non-specific public metabolite databases, e.g., the Human Metabolome Database, the curated library should enable higher confidence annotation of untargeted metabolomics data.
Finally, implementation of the developed workflows in an extensive untargeted metabolomics study applying intracellular nESI-DIMS and extracellular UHPLC-MS-based analyses to measure the fate and effects of twelve xenobiotics, eight structurally cardiotoxic and four non-structurally cardiotoxic, each at two concentrations, after 6-, 48- and 72-hour exposures, in cardiac microtissues, enabled discovery of metabolic signatures predictive of drug-induced structural cardiotoxicity. This included the responses of creatine, ceramides, and several purine metabolism intermediates. Future efforts to validate and mechanistically anchor the proposed structural cardiotoxicity metabolic biomarker panel and subsequently deploy as a high-throughput targeted assay are needed to translate the findings of this thesis into an improved cardiac safety screening capability for drug development.

Type of Work:

Thesis (Doctorates > Ph.D.)

Award Type:

Doctorates > Ph.D.

Supervisor(s):