CRISPR-Cas9 Screen Identifies Me2, a Mitochondrial Malic Enzyme, As a Molecule Relevant for MCL1 Inhibitor Resistance in AML

The BCL-2 protein family plays a key role in leukemogenesis via counteracting proapoptotic signals, thereby enhancing leukemia cell survival. Venetoclax, a selective BCL-2 inhibitor, has been proven effective against AML in combination with hypomethylating agents. Since MCL1, another anti-apoptotic protein, is reportedly more abundant than BCL-2 in AML cells and associated with resistance to chemotherapy, development of clinical-grade MCL1 inhibitors has been highly regarded. In fact, a series of MCL1 inhibitors are currently being tested in clinical trials. While relationships between metabolic condition in mitochondria and sensitivity to Venetoclax have been proposed in recent studies, molecular mechanisms underlying MCL1 inhibitor resistance are not well understood.

To identify genes/pathways whose loss induce MCL1 inhibitor resistance in AML cells, we performed genome-wide CRISPR-Cas9 screens using a mouse AML cell line in the presence or absence of s63845, a MCL1 inhibitor (Kotschy et al. Nature 2016). To establish AML cell lines with a relatively clean genetic background, we first established AML in mice by transducing the MLL/AF9 leukemia oncogene into mouse bone marrow stem/progenitor cells ex vivo, followed by serial transplants (Yamauchi et al. Cancer Cell 2018). We then harvested AML cells from leukemic mice and established AML cell lines with normal karyotype and intact Trp53 activity. We then performed genome-wide CRISPR-Cas9 screens in the presence or absence of s63845, followed by a second screen with a small-scale library targeting genes that were negatively- or positively-selected upon s63845 treatment in the primary screen. Genes relevant for S63845 resistance were identified using MAGeK MLE (Li et al. Genome Biology 2015) and DrugZ (Colic et al. BioRxiv 2019) programs.

Single-guide RNAs (sgRNAs) targeting intrinsic pro-apoptotic genes, such as Bak1, Casp9 and Apaf1, were enriched exclusively in the presence of s63845, attesting to the validity of our experiment. To identify genes relevant to s63845 resistance, we focused on the genes whose sgRNAs were enriched exclusively upon s63845 treatment. We found that loss of Me2, which encodes a mitochondrial malic enzyme that catalyzes the oxidative decarboxylation of malate to pyruvate, promotes AML cell survival only in the presence of s63845, but not in vehicle- or Venetoclax-treated cells. This finding was validated in the second screen, in which 8 independent sgRNAs targeting Me2 were tested. We next generated two independent Me2-null mouse AML cell lines (MLL/AF9 and CALM/AF10) using sgRNAs targeting Me2. Me2-null cells exhibited survival advantage over control cells upon s63845 treatment, revealed by cell proliferation assay.

To determine at which step of the apoptotic pathway Me2 deficiency exerts s63845 resistance, we assessed MOMP (mitochondrial outer membrane permeabilization) levels by FACS with or without s63845 treatment. Me2-null cells released less cytochrome C than Me2-wild type (WT) cells upon s63845 treatment. As expected, Me2-null cells exhibited less Annexin-positivity than WT cells upon MCL1 inhibition. Importantly, both mRNA and protein levels of BCL-2 family members, including Bcl-2, Mcl1, Bcl2L1, Bax and Bak1, were comparable regardless of Me2 status.

We next performed CRISPR-Cas9 saturation mutagenesis scan of all ME2 exons in the presence or absence of s63845 using Molm-13, a human AML cell line. sgRNAs targeting ME2 coding exons were enriched only in the presence of s63845, while those targeting ME2 3'UTR were unchanged/neutral regardless of experimental conditions.

In conclusion, using an unbiased, genome-wide CRISPR/Cas9 screens, we identified Me2, a mitochondrial metabolic enzyme, as a factor relevant for MCL1 inhibitor resistance. Our study may facilitate the understanding of molecular mechanisms underlying acquired resistance to MCL1 inhibitors in AML.