Heterogeneity of human metabolism in health and disease: a modelling perspective

Doctoral thesis, 2017

Metabolism is broadly defined as the sum of biochemical reactions within cells that are involved in maintaining the living state of the organism. Profound importance of metabolism comes from the fact that it is the sole source of energy that allows life to resist to be degraded into entropy. Human metabolism is a complex interactive network consisting of highly regulated functional pathways, impacting or been impacted by many other cellular process. Internal or external perturbations may cause dysfunction of some of these functional or regulatory pathways and may lead to the rise of abnormal phenotypes. Many human diseases associated with irregular metabolic transformations that perturb normal physiology and lead to phenotype dysfunction. Discovering how biological systems reorganize their activities to force specific phenotypic transformation, e.g., normal to cancer/diabetes/obesity, is a main challenge in life science. This thesis is dedicated to investigating genome-scale metabolic transformations from health to disease states, with specific focus on non-symmetric reprogramming in cancer metabolism. The human gut microbiome has been associated with a variety of human diseases, but to go beyond association studies and elucidate causalities is a major challenge. We developed a comprehensive computational platform, CASINO (Community and Systems-level Interactive Optimization), for simulation of the microbial communities using genome-scale metabolic modeling. We demonstrated the power of the toolbox in predicting metabolic interactions between gut microbiota and host, through a diet-intervention study of obese and overweight individuals. Our modeling platform could provide a quantitative description of the altered plasma and fecal amino acid levels in response to dietary intervention. Next, we proceed to investigate heterogeneity of cancer related metabolic transformations at the genome-scale. First, we reconstructed genome scale metabolic models (GEMs) for eleven human cancer cell lines based on RNA-Seq data. We used the generated models to investigate inter-cell line heterogeneity of metabolic reprogramming and also to identify potential anti-growth factors. This was followed by two consecutive studies on two main subtypes of the non-small cell lung cancer, lung adenocarcinoma (LAC) and lung squamous cell carcinoma (SCC), by generating RNA sequencing (RNAseq) data for cancer biopsies and for normal tissue samples. We followed a systemic approach to investigate the heterogeneity and direction of the metabolic transformation in lung cancer at three levels of biochemical organization: global metabolic network level, individual biochemical pathways level and at the level of specific enzymatic reactions. We observed large heterogeneity in the expression of enzymes involved in the majority of the metabolic pathways, and identified significant association between some of these variations and patient prognosis. Our findings provide mechanistic insights into complex metabolic behavior of tumors and may be used to develop more effective diagnostic and prognostic methods.