Metabolic Landscapes Across Cancer Types
A computational framework for making metabolomics data from different laboratories play nice, and the discovery of common metabolic signatures of aggressive disease across many cancer types. With Augustin Luna, Chris Sander, and colleagues at MSKCC and DFCI.
Mitochondrial Genetics in Cancer
Mitochondrial DNA and RNA is prevalent in most tumor sequencing data, but most people treat it like garbage. We became sanitation workers and recaptured this data. We made a remarkably simple observation: with a few key exceptions, solid tumors are depleted of mitochondrial DNA (and in subsequent work, RNA) relative to adjacent normal tissue. We asked why, and what the consequences were for the rest of the molecular program of the cell. With Chris Sander and colleagues at MSKCC.
Architectures of Metabolic Regulation
Metabolism is regulated at many scales. While metabolites have been known to allosterically regulate enzymes for over a century, our technical ability to broadly detect small-molecule regulation of enzymes remains poor. We used an informatic, rather than an experimental, approach to reconstruct genome-scale networks of small-molecule regulation, and along the way discovered (and mathematically proved) a fundamental tradeoff between catalytic activity and regulatory control. With Dimitris Christodoulou, Elad Noor, Uwe Sauer, Daniel Segre, and colleagues at BU.
A Metabolic Atlas of Clear Cell Renal Cell Carcinoma
Clear cell renal cell carcinoma is the most common form of kidney cancer, and is almost always associated with loss of the HIF regulator VHL. As a result, ccRCC cells behave like they are in hypoxia by shifting their metabolism away from OXPHOS and towards a more glycolytic phenotype. We studied this phenotype using metabolomics and transcriptomics, and found that aggressive disease in ccRCC is associated with specific shifts in the regulation of oxidative stress, the production of amino acid precursors, and the accumulation of polyamines. Not surprisingly, we observed that inferences of metabolism from gene expression data shed little light on corresponding metabolomic changes. A clear deficit in this work, and one which we are passionately interested in amending, is the lack of good, predictive computational models to infer metabolic flux from bulk tumor metabolomics and transcriptomics data. With Ari Hakimi, James Hsieh, and colleagues at MSKCC.
Genome Scale Models of Nutrient Limitation
Which nutrients limit the cell's growth? We show that constraint-based models of metabolism capture a quantitative notion of the limitation of each intracellular metabolite. Predictions of metabolic limitation match experimentally measured values from chemostat experiments in S. cerevisae. To our surprise, we also find that we can interpret metabolic limitation from a dynamical perspective, and that strongly limiting intracellular metabolites have comparatively fast returns to steady following a perturbation. With Pankaj Mehta and Daniel Segre.
Variational Principles for Enzyme Kinetics
An old favorite. We use methods from the Calculus of Variations to prove that the century-old Michaelis-Menten equation contains a beautiful symmetry that can be used to relax a key assumption in its derivation: that the total amount of enzyme remains constant. This enables us to ask what happens when the enzyme changes over time, and prove both theoretically and experimentally that the only quantity that matters is the average amount of enzyme. With Stefan Yohe and Daniel Segre.