MRC2+ AML Stem Cells Reveal the Role of Proline Metabolism in Venetoclax Resistance

Acute myeloid leukemia (AML) is a highly aggressive and deadly blood cancer with less than 30% five-year survival. AML develops when hematopoietic stem cells (HSC) or hematopoietic stem and progenitor cells (HSPC) acquire key mutations resulting in leukemic stem cell (LSC) transformation. Because of the diverse spectrum of mutations, and the different cell types that can give rise to AML LSC, there is considerable LSC heterogeneity from patient to patient. Thus, finding novel therapeutic avenues that target common LSC functions, such as cellular metabolism, is vital for consistent disease elimination across the patient population.

As part of previous studies on LSC, we found that the cell surface protein mannose receptor C type 2 (MRC2), which binds and internalizes collagen and gelatin, is differentially expressed in AML subpopulations and can distinguish LSC from their non-LSC counterparts. This was shown by sorting MRC2+ or MRC2- cells from primary AML patients and finding that MRC2+ cells have increased stemness both in vitro (n=13 patients) and in vivo (n=3 patients) using colony formation assays and patient derived xenografts, respectively. In line with this observation, RNA-seq analysis (n=3 patients) showed that stemness gene programs are significantly enriched in MRC2+ cells from AML patients compared to MRC2- cells. Importantly, we found that MRC2 is functional on AML cells, with MRC2+ cells having significantly increased gelatin internalization over MRC2- cells. Further, we performed genetic deletion of MRC2 in AML cell lines (n=3) and found that gelatin internalization was completely abolished. Interestingly, using the seahorse extracellular flux assay we found that gelatin consumption led to increased glycolysis in MRC2-hi (>70% MRC2+) AML patient samples (n=6). Altogether, these data highlight that MRC2 can be used as a surface marker to enrich for AML LSC, and that MRC2 is used to access gelatin as a carbon source for glycolytic metabolism.

To further understand our newly identified LSC metabolic circuit, we compared our RNA-seq analysis from MRC2+ cells, to our previously published RNA-seq data-set on NKG2DL- AML patient samples (which enriches for LSC), and to the TCGA data on MRC2-high AML cells. Using this three-way comparison, we found 8 common genes including PRODH, which encodes an enzyme essential for proline metabolism. Importantly, proline is the third most abundant amino acid in gelatin, and gelatin treatment increased PRODH expression in AML cell lines in an MRC2 dependent manner (n=3). Strikingly, PRODH inhibition (PRODHi) significantly decreased cell survival and colony formation capacity in AML cells lines (n=3) and primary AML patient samples (n=12). Surprisingly, this was also true in samples with low MRC2 expression, highlighting that proline is a key metabolite for LSC function, and not only important in the context of gelatin metabolism. Of note, PRODHi did not reduce survival or colony formation capacity in healthy CD34+ cells (n=3). Excitingly, PRODHi increased the efficacy of the Bcl-2 inhibitor venetoclax to kill AML cell lines (n=6) and patient samples (n=17), and PRODHi was able to induce venetoclax sensitivity in venetoclax resistant cells. Thus, we were able to show that proline metabolism is essential for AML LSC function and survival, and for limiting the cytotoxic effects of venetoclax.

Overall, we identified proline metabolism as a new mechanism for venetoclax resistance in AML. These insights can be instrumental for the development of co-treatments that enhance the efficacy of venetoclax to fully eradicate LSC and thus cure AML in a larger proportion of patients, especially as venetoclax becomes standard therapy. This is particularly important for hard to treat AML such as p53 mutant which is commonly venetoclax resistant. Future directions will aim at further understanding the mechanisms underlying the relationship between venetoclax and proline metabolism, and how gelatin impacts the metabolism of MRC2+ LSC.

No relevant conflicts of interest to declare.