Targeting Sumoylation with Subasumstat (TAK-981), a Selective Inhibitor of SUMO-Activating Enzyme (SAE), Induces Mitochondrial Dysfunction in Preclinical Models of Non-Hodgkin Lymphoma (NHL)

Abstract

SUMOylation is a post-translational modification which leads to covalent attachment of small ubiquitin-like modifiers (SUMO) proteins to lysine residues of a target protein. SUMOylation is essential for the regulation of genomic integrity, gene expression and intracellular signaling. Many tumors exhibit deregulation of SUMOylation pathway. Subasumstat, (TAK-981), a novel selective inhibitor of SUMO-activating enzyme (SAEi), has entered early phase II clinical trials. Here we studied SAEi in pre-clinical models of B-cell non-Hodgkin's lymphoma (NHL). TAK-981 was provided by Takeda Development Center Americas, Inc. (Lexington, MA). Diffuse large B-cell lymphoma (DLBCL) cell lines (OCI-LY3, VAL, U-2932, SU-DHL10), and primary mantle cell lymphoma (MCL) cells were assayed for survival, mitochondrial function, and metabolic phenotype. In vivo experiments were performed using DLBCL cell line xenografts and MCL PDX models. Treatment with TAK-981 restricted growth of DLBCL cell lines with an IC50~10 nM. SAEi led to rapid (within 2 h) protein desumoylation in a dose dependent manner, prominently in the ~90 kDa region. TAK-981 induced pronounced double strand DNA breaks in DLBCL cells (comet assay) and cell cycle arrest (propidium iodine). RNA-Seq and ATAC-Seq analysis of TAK-981-treated DLBCL cells revealed upregulation of the DNA damage, NF-κB, and OXPHOS signaling pathways. We noted that in addition to cytoplasmic proteins, mitochondrial proteins were heavily de-SUMOylated upon treatment of NHL cells with TAK-981. In vitro exposure to TAK-981 resulted in mitochondrial membrane depolarization, loss of mitochondrial integrity (by electron microscopy) and a change of mitochondrial dynamics as assessed by expression of mitochondrial proteins (eg, Drp1). Seahorse respirometry analysis revealed dramatic downmodulation of OXPHOS following SAEi. Furthermore, metabolomic profiling of DLBCL cells treated with TAK-981 showed downmodulation of TCA substrates, while glycolysis substrates remained unchanged. The mitochondrial disruption observed following TAK-981 treatment was accompanied by rapid accumulation of reactive oxygen species, a mechanism that has been previously shown to lead to DNA damage in cancer cells. Consistent with cell line data, exposure of primary MCL cells to TAK-981 in stromal conditions (CD40L- or BAFF expressing stroma) induced apoptosis and reduced OXPHOS. Next, we conducted genome-wide CRISPR-Cas9 loss-of-function library screens and identified that loss of genes in the NFκB, TP53, DNA damage and centromere/telomere gene pathways contributed to resistance to SAEi in DLBCL cells. By means of genetic knockout we demonstrated that TP53 and its transcriptional target BAX contributed to TAK-981-induced G2 arrest and apoptosis, further implicating mitochondrial dysfunction in sensitivity to SAEi. For in vivo experiments, we used DLBCL xenografts. OCI-LY3 and U-2932 cells were inoculated subcutaneously in NSG mice. Once tumors reached 100 mm3, mice received 10 twice-weekly IV doses of 7.5 mg/kg TAK-981 (over 5 weeks), leading to resolution of flank tumors and improved survival compared with control. Finally, we used an MCL PDX model where MCL cells were inoculated intravenously. Once cells became detectable in the blood, mice were treated as above. Treatment with TAK-981 resulted in delayed tumor expansion in blood and spleen and extended animal survival. TAK-981-exposed splenocytes exhibited a reduction of maximal respiration by Seahorse analysis. Conclusions. Pharmacologic SAEi with TAK-981 demonstrated pre-clinical activity in NHL models in vitro and in vivo, accompanied by DNA damage, mitochondrial dysfunction, and metabolic reprogramming.