PCYT1A Is a Metabolic Vulnerability in Monocytic Acute Myeloid Leukemia

Acute myeloid leukemia (AML) is a highly heterogeneous blood cancer with unique morphologic and genomic characteristics. Several genetic aberrations, such as TP53 mutations (TP53-MT) and MLL (KMT2A) rearrangements (MLL-r), are associated with inferior outcomes. The cellular origin of AML, mainly reflected by morphologic features, also influences drug response. For example, monocytic AMLs have been shown to be relatively resistant to venetoclax-based therapies. The development of appropriate therapies for each AML subtype is therefore an important challenge to improve therapeutic efficacy. To identify novel therapeutic targets in specific AML subtypes, we developed a Python-based in silico analysis system using the Cancer Dependency Map (DepMap) datasets. With this system, we can rank all genes according to their specific essentiality in each AML subtype. We applied this method to TP53-MT AML and MLL-r AML and identified BCL2L1 and PCYT1A as potential therapeutic targets, respectively. BCL2L1 (also known as BCL-XL) is an anti-apoptotic protein, and our analysis suggested that it is specifically important in TP53-MT AML. BCL2L1 depletion inhibited the growth of TP53-MT AML cell lines (HEL and TF1), whereas it showed only a marginal effect in AML cell lines with wild-type TP53 (MOLM13 and OCIAML3). However, prior depletion of TP53 in MOLM13 cells did not sensitize them to BCL2L1 depletion, suggesting that the increased sensitivity of TP53-MT AMLs to BCL2L1 depletion is not the direct consequence of TP53 inactivation. Because TP53 mutations are common in acute erythroid leukemia (AEL), we speculated that the strong dependence of TP53-MT AML cells on BCL2L1 is not due to the TP53 mutation itself, but to their erythroid characteristics. Indeed, two TP53-MT non-AEL cell lines (THP1 and U937) were not sensitive to the BCL2L1-selective inhibitor (A-1155463) despite the presence of TP53 mutations, whereas TP53-MT erythroid leukemias (HEL and TF-1) were highly sensitive to A-1155463. Thus, consistent with the findings in a recent report (Blood 141(13): 1610-1625, 2023), the erythroid differentiation, but not the presence of TP53 mutations, confers BCL2L1 dependency in AML. The top-ranked gene for MLL-r AML was PCYT1A, which is a rate-limiting enzyme for the synthesis of CDP-choline in the phosphatidylcholine biosynthetic pathway (Kennedy pathway). Consistent with our prediction, PCYT1A depletion inhibited the growth of MLL-r AML cell lines (THP1 and MOLM13), but not MLL wild-type (MLL-WT) AML cell lines (KASUMI1 and HEL) in vitro. Human cord blood cells and mouse bone marrow cells transformed by MLL-AF9 were also dependent on PCYT1A for their growth, confirming the critical role of PCYT1A in MLL-r AML. However, we found that cSAM cells, a mouse non-MLL-r AML cell line transformed by SETBP1 and ASXL1 mutations, were also dependent on PCYT1A. Because MLL rearrangements are common in monocytic AML and cSAM cells also show monocytic features, we speculated that monocytic AML including MLL-r AML are dependent on PCYT1A to sustain their leukemic growth. Indeed, PCYT1A depletion inhibited the growth of monocytic AML cell lines without MLL-r (OCIAML3 and U937), confirming the indispensable role of PCYT1A in monocytic AML. Mechanistically, we found that the expression of PCYT1B, a paralog of PCYT1A, was low in monocytic AML. Furthermore, forced expression of PCYT1B attenuated the growth-inhibitory effect of PCYT1A depletion in monocytic AML cells, whereas simultaneous depletion of PCYT1A and PCYT1B inhibited the growth of non-monocytic AML cells. These results suggest that monocytic AMLs specifically require PCYT1A for their growth due to the lack of expression of its paralog gene, PCYT1B. Finally, we confirmed the essential role of PCYT1A for AML development in vivo using MOLM13 cells and patient-derived xenograft (PDX) cells derived from a monocytic AML patient. In summary, we have identified BCL2L1 and PCYT1A as specific therapeutic targets in AEL (including those with TP53 mutations) and monocytic AML (including MLL-r AML), respectively. PCYT1A is a novel, synthetic lethal metabolic susceptibility in monocytic AML with low PCYT1B expression. More broadly, our findings highlight the importance of the cell of origin, rather than the genetic aberrations alone, to identify subtype-specific vulnerabilities in AML.