Inhibition of Topors Ubiquitin Ligase Augments the Efficacy of DNA Hypomethylating Agents through DNMT1 Stabilization

Abstract

DNA hypomethylating agents (HMAs) are commonly used to treat myeloid malignancies including myelodysplastic syndrome (MDS) and acute myeloid leukemia (AML). However, their therapeutic effects are not satisfactory. We performed CRISPR-Cas9 knockout (KO) screening to identify molecular targets that enhance the efficacy of HMAs. MDS-L and MOLM-13 cells expressing Cas9 were infected with an sgRNA lentiviral library containing 12,409 sgRNAs targeting 1,383 epigenetic factors and exposed to low-dose HMAs, decitabine (DAC) or azacitidine (AZA), for 14 days. The read counts of sgRNA against TOPORS which encodes a ubiquitin/SUMO E3 ligase, were reproducibly decreased during culture specifically in the presence of HMAs in both screening. To validate the screening results, we performed competitive growth assays by co-culturing parental cells and TOPORS-KO cells using MOLM-13, MDS-L, SKK-1, and SKM-1 cells. In all cell lines, GFP-positive TOPORS-KO cells decreased over time only in the presence of HMAs. We established single-cell clones of TOPORS-KO MOLM-13 and MDS-L cells using two different sgRNAs. The cell growth of all TOPORS-KO clones was dramatically suppressed only in the presence of DAC. Mice transplanted with TOPORS-KO cells showed significantly longer survival when treated with DAC. Next, we analyzed Topors-deficient mice and no significant changes were observed in the peripheral blood or in the percentage of bone marrow (BM) hematopoietic stem and progenitor cells. Topors−/− BM cells successfully restored hematopoiesis in lethally irradiated recipient mice at levels comparable to those of WT BM cells after transplantation, suggesting that Topors is dispensable for normal hematopoiesis. Topors−/−MLL/AF9 leukemic cells showed comparable growth in the absence of DAC but significantly inferior growth in the presence of DAC compared to WT MLL/AF9 leukemic cells. We transplanted them with support BM cells into lethally irradiated recipient mice and treated with DAC. Topors−/− leukemic cells significantly decreased during treatment compared with WT leukemic cells. Enhanced cytotoxic effects of TOPORS KO combined with HMAs was closely associated with increased apoptosis. Cell cycle analysis revealed that among TOPORS-KO cells, those in the G2/M phase accumulate and those in the S phase are reduced upon DAC treatment. RNA sequencing on cells treated with DAC showed that gene sets associated with chromosome segregation and mitosis are significantly downregulated in TOPORS-KO cells. Addback experiment revealed that TOPORS RING finger domain, which mediates ubiquitination, is responsible for HMA resistance. To identify TOPORS ubiquitination substrates, we performed second CRISPR-Cas9 screening with the same library used as the first screening, using TOPORS-KO MDS-L and MOLM-13 clone cells. Both screening showed that UHRF1-DNMT1 axis is implicated in the augmented sensitivity to DAC. Consistently, DNMT1 remained more stable in TOPORS-KO cells than WT cells after DAC exposure not only in human MDS/AML cell lines but also in murine leukemic cells. HMAs incorporated into DNA trap DNMT1 and form DNA-DNMT1 crosslinks, which causes apoptosis and mitotic defects. DNMT1-DNA crosslink formation triggers the prominent SUMOylation of DNMT1 followed by ubiquitination, which promotes the resolution of crosslinks via proteasomal degradation and cell fitness. Mass analysis of ubiquitinated peptides and immunoprecipitation with Nanobody against DNMT1 demonstrated that TOPORS loss causes accumulation of SUMOylated DNMT1 due to inefficient ubiquitination, suggesting that TOPORS works as SUMO-targeted ubiquitin ligase for DNMT1. In patient samples treated with AZA (n=23), treatment resistance significantly correlated with TOPORS expression levels. Furthermore, an ubiquitination inhibitor TAK-243 as well as a SUMOylation inhibitor TAK-981 showed synergistic effect with HMAs through DNMT1 stabilization. These results suggested that they are likely to be promising therapeutic agents in clinical practice. Our findings unveil a novel mechanism of resistance to HMAs and provide an attractive therapeutic strategy for myeloid malignancies that interferes with resolution of DNA-DNMT1 crosslinks by targeting DNMT1 post-translational modification.