(190k) Understanding the Role of Central Carbon Metabolism in Myeloid and Monocytic Hematopoietic Differentiation Programs in Patient Derived HL-60 Cells

A major challenge in cancer research is relieving the toxicity of conventional chemotherapy. Differentiation induction chemotherapy (DIC), using agents such as the vitamin A derivative all-trans retinoic acid (ATRA), is a promising treatment for many cancers. For example, ATRA treatment induces remission in 80 - 90% of promyelocytic leukemia (APL) PML-RARα-positive patients, thereby transforming a fatal diagnosis into a manageable disease. However, this remission is often not durable and relapsed cases exhibit emergent ATRA resistance. Moreover, the success of differentiation induction chemotherapy has been limited to acute promyelocytic leukemia; other myeloid leukemias are resistant to retinoic acid differentiation therapy. To understand this resistance, we must first understand ATRA-induced differentiation programs, and how these programs are transformed in resistant cells. Toward this challenge, lessons learned in model systems, such as the lineage-uncommitted human myeloblastic cell line HL-60, which closely resembles patient derived cells, informs our analysis of the differentiation programs occurring in patients. Patient derived HL-60 cells, a durable experimental model since the 1970s to study differentiation, can undergo myeloid or monocytic differentiation and G0 arrest, making it one of the few hematopoietic lineage uncommitted precursor cell lines. Recently, we developed and characterized two ATRA resistant HL-60 cell lines and showed that the ATRA-induced molecular profile of HL-60 was consistent with subsets of wt-NPM1 primary acute myeloid leukemia (AML) blasts which respond to retinoic acid treatment. Additionally, we showed that ATRA resistant HL-60 cells, as well as primary AML samples from N = 12 patients (natively ATRA-resistant), could be made ATRA-responsive by combinations of ATRA and agents such as bosutinib (B) or 6-formylindolo(3,2-b) carbazole (FICZ). We also showed differences in intracellular glucose consumption amongst primary AML samples, and the correlation/anti-correlation of glucose metabolism with the expression of molecular markers of ATRA-induced differentiation. Thus, our working hypothesis is that combination approaches could rescue ATRA resistant cells, and this rescue involves potentially altered signaling and metabolic programs. Toward this hypothesis, we developed computational and experimental tools to study the integration of ATRA-induced differentiation with metabolism. First, liquid chromatography-mass spectrometry (LC-MS) was used to measure the extracellular concentration of amino acids, organic acids, and glucose as a function of time in HL-60 cultures with and without ATRA. These measurements were then used, along with a core metabolic model derived from the human metabolic network reconstruction RECON 2, to estimate the metabolic flux distribution of control and ATRA treated HL-60 cells. Finally, we integrated signaling, gene expression, and metabolism together into an inclusive model to better understand how metabolism is integrated with ATRA-induced signaling and gene expression. The specific growth rate decreased in ATRA treated HL-60 cells, consistent with literature, while the specific lactate production rate nearly doubled upon ATRA treatment. However, the increased rate of lactate production was not matched by increased glucose consumption; the specific rate of glucose consumption was not statistically significantly different between the cases. The constraint-based model predicted an incomplete TCA cycle for treated and untreated cells; there was no flux through citrate synthase (the conversion of mitochondrial Acetyl-CoA to citrate) in either ATRA-treated or untreated cells. Additional carbon consumption was required to meet the demand for lactate precursors such as pyruvate in the presence of ATRA. Serine, arginine, tryptophan, and glutamine were heavily consumed for both ATRA-treated and untreated cases. However, ATRA treatment decreased the consumption of serine, glutamine and alanine, but increased the uptake of arginine and the production of glutamate. This suggested increased arginase activity. Arginase converts arginine to urea and ornithine; ornithine can then be further converted into fumarate, a TCA intermediate, which can finally be converted to pyruvate through malic enzyme. The working hypothesis of elevated arginase and malic enzyme activity in ATRA treated cells could be explored by enzyme activity assays in subsequent studies. Lastly, the gene regulatory network model captured the expression of phenotypic markers associated with differentiation, such as CD38, CD11b, and CD14, consistent with experimental measurements. Taken together, we developed experimental and computational tools to analyze ATRA-induced differentiation of HL-60 cells. These tools supported the development of a comprehensive integrated model of signal transduction, gene expression and metabolism in HL-60 cells. We believe an integrated modeling approach could be used to identify new agents that could be used in combination with ATRA to rescue ATRA resistant cells.