Inhibition of Folate Metabolism Drives Autophagy-Dependent Differentiation and Reduces Survival of Therapy-Resistant Leukaemic Stem Cells

Metabolic rewiring is an important hallmark of cancer. The folate metabolism pathway, also known as one-carbon (1C) metabolism, allows for transfer of 1C units through folate intermediates for biosynthetic processes, including precursors for DNA synthesis. Recent studies have shown that enzymes involved in the mitochondrial arm of 1C metabolism are overexpressed in a subset of aggressive cancers and that their expression affects responses to anti-metabolite drug treatments. However, the role of 1C metabolism in therapy resistant leukemic stem cells (LSCs) is currently unknown. Therefore, we aimed to investigate the activity and the impact of genetic and pharmacological inhibition of folate enzymes in primitive chronic myeloid leukaemia (CML) cells.

We initially performed transcriptomic analysis of CD34+38- cells, from individuals with chronic phase CML (E-MTAB-2581). This revealed a significant upregulation of folate metabolism genes in CML LSCs, including serine hydroxymethyltransferase (SHMT2; p≤0.05), a key mitochondrial enzyme. To assess the activity of 1C metabolism in primitive cells we performed gas chromatography-mass spectrometry-mediated secretomic analysis using patient-derived, c-KIT enriched CML cells, which revealed a significant increase in the exchange rate of formate (folate intermediate necessary for purine synthesis) in CML cells, when compared to the secretome of normal counterparts (p<0.05). This reinforced the idea that 1C metabolism may be a metabolic dependency in CML.

Following CRISPR-Cas9-mediated SHMT2 knockout (KO) in CML cell line, we observed a significant decrease in growth rate, together with a decrease in glycolytic capacity and oxygen consumption rate (p<0.01), suggesting impairment in proliferation and central carbon metabolism. Further metabolic characterisation of CML SHMT2 KO cells using liquid chromatography-mass spectrometry demonstrated a significant increase in AICAR, a purine biosynthesis intermediate and an AMP activated kinase (AMPK) activator. This prompted us to investigate the effect of 1C metabolism inhibition on AMPK. We found that AMPK phosphorylation on the conserved Thr 172 (a site that is phosphorylated under energy stress) was increased in SHMT2 KO cells, with similar effect seen following pharmacological inhibition of both SHMT2 and its cytosolic counterpart SHMT1 using SHIN1,which also promoted AMPK-dependent phosphorylation of the autophagy-inducing kinase ULK1 and downstream ULK1 target ATG13. Moreover, analysis of mitochondrial fraction revealed accumulation of mitochondrial fission related protein DRP1 and the mitophagy receptor NIX on mitochondria, hinting towards cellular interplay between 1C metabolism and mitochondrial homeostasis.

Phenotypically, both pharmacological and genetic inhibition of SHMT1/2 induced the expression of erythropoiesis markers CD71 and Glycophorin A, which was reversed following formate supplementation. CRISPR-Cas9-mediated double AMPKα1/α2 KO revealed that the increased expression of these erythropoiesis markers following SHMT1/2 inhibition was independent of AMPK activity. Conversely, while NIX KO had no effect, pharmacological inhibition of ULK1 kinase activity, or genetic inhibition of ULK1 and ATG7 (protein important for autophagosome formation), prevented increased expression of CD71/Glycophorin A following SHMT1/2 inhibition.

We next investigated the effect of 1C metabolism inhibition on differentiation and survival of primary CML cells. Of clinical relevance, pharmacological inhibition of SHMT1/2 promoted erythroid maturation of CD34+ CML cells (measured by expression of CD71, CD44, CD36 and Glycophorin A) when challenged with erythropoietin, which sensitises primitive cells to erythroid lineage commitment. Lastly, pharmacological inhibition of 1C metabolism decreased the colony formation capacity of CD34+ CML by 50%, with minimum effect on normal CD34+ cells. Moreover, combination treatment of SHIN1 with imatinib, a frontline treatment for CML patients, further increased the sensitivity of primary CML cells to imatinib by 40%.

Overall, our novel findings indicate that disruption of the folate metabolism pathway inhibits central carbon metabolism in CML cells, promotes autophagy dependent, but AMPK independent maturation phenotype and has detrimental effect on the survival of primitive CML cells.