Background

Unraveling the molecular regulation of cancer cells is critical to understand their insurgence, functioning, and potential therapeutic vulnerabilities. A dysregulated cellular metabolism is now widely accepted as a hallmark of many cancers (DeBerardinis and Chandel, 2016). Cancer cells display considerably different metabolic requirements compared to most normal differentiated cells and reprogram pathways controlling nutrient acquisition and metabolism to meet their specific bioenergetic, biosynthetic, and redox demands. Emerging evidence, mostly from solid tumors, shows that several metabolic characteristics of cancer cells can also be linked to therapeutic resistance (Zhao et al., 2013). Metabolic heterogeneity between cells, driven by genetic, cell state, or microenvironmental diversity, may further allow subpopulations of cells to escape therapy-related cell death (Martinez-Outschoorn et al., 2016).

The importance of examining the metabolism of cells in their native environment is underscored by several studies that have shown that artificial in vitro conditions can create metabolic dependencies that do not exist in vivo (Davidson et al., 2016; Cantor et al., 2017; Muir et al., 2017). However, studying metabolism in an in vivo setting brings several challenges. Gene or protein levels of metabolic enzymes may not always adequately reflect the metabolic state of a cell (Metallo and Vander Heiden, 2013). Measuring single metabolites in cells in vivo or on tissue sections is feasible, as is metabolic profiling of freshly-isolated cells through metabolomics analysis (Arnold et al., 2015; Faubert and DeBerardinis, 2017). These approaches have proven extremely valuable, but they provide a snapshot of metabolite levels and no dynamic information or insight into the flow of metabolites. Metabolic tracer analysis is a complementary method that uses stable isotope-labeled metabolites, which are taken up and metabolized by the cells, after which the fate of the labeled atoms (usually carbon, nitrogen, or hydrogen) can be analyzed by mass spectrometry. In this way, the relative activity of one or more selected metabolic pathways can be studied. Metabolic tracer analysis, as well as metabolic flux analysis, which combines metabolic tracer analysis with tracer uptake data and computational modeling, has become a standard technique in the study of cancer cell metabolism, offering insights into the relations between cell biology and biochemistry (Antoniewicz, 2018). While metabolic tracing is mainly performed with cells in culture, it is also more and more used in in vivo settings, both in animal models and in humans (Faubert and DeBerardinis, 2017; Fernández -García et al., 2020). Most in vivo metabolic tracing studies have focused on solid tumors, analyzing the tumor as a whole, without separating the cancer cells from other cells of the tumor microenvironment, such as immune cells, stromal cells, or blood vessel cells. The advantage of working with liquid tumors is that the cancer cells can be readily obtained from the bone marrow without the need for enzymatic tissue digestion and can be separated from the other cells by fluorescence-activated cell sorting (FACS), thus allowing specific analysis of cancer cell metabolism in vivo.

In our original study, we investigated the metabolic adaptations of acute myeloid leukemia (AML) cells in vivo during the course of induction chemotherapy treatment and identified transient changes in glutamine levels (van Gastel et al., 2020). We performed in vivo stable isotope tracing, according to the protocol described here, to understand the metabolic fate of glutamine in control AML cells or cells persisting after chemotherapy. We used a transplant-based mouse model of human AML, driven by the MLL-AF9 fusion protein (Corral et al., 1996). The AML cells also expressed GFP, which greatly facilitates cell sorting. The protocol described below should be compatible with any mouse leukemia model in which the leukemia cells express a fluorescent marker. It is also possible to perform metabolic tracing using our protocol in other hematopoietic cell populations, such as normal hematopoietic progenitor cells (Lineage^-cKit⁺), as long as they can be readily isolated from bone marrow and separated from other cells by FACS.