LAUSR

Dear Editor, Esophageal squamous cell carcinoma (ESCC) is the major histologic subtype of esophageal cancer with high incidence and mortality. However, few achievements have been made in the development of targeted drugs. Neddylation, a reversible post-translational modification, attaches ubiquitin-like molecule NEDD8 to substrates in a three-step enzymatic reaction catalyzed by NEDD8-activating enzyme E1 (NAE, NAE1 and UBA3 heterodimer), NEDD8-conjugating enzyme E2s (UBE2M/UBC12 or UBE2F) and substrate-specific NEDD8-E3 ligases. The best-characterized substrates of neddylation are cullin family proteins, the essential components of multiunit Cullin-RING ubiquitin ligases (CRLs). Currently, the inhibition of cullin neddylation by targeting overactivated neddylation pathway has emerged as an attractive approach for anticancer therapy. Our previous study reported that MLN4924, a specific inhibitor of NAE, significantly inhibited the tumor growth of ESCC by blocking cullin neddylation and inactivating CRLs activity. However, recent studies found that MLN4924 treatment-emergent NAE mutations would confer the drug resistance. Therefore, it is urgent to identify other neddylation enzymes (E2s or E3s) as alternative anticancer targets and develop novel anti-ESCC strategies. In the present study, with a label-free quantitative proteomic approach, NEDD8-conjugating enzyme UBC12 was identified as a potential anticancer target against ESCC. Gene ontology (GO) analysis of proteins with altered expression revealed that silencing UBC12 by CRISPR/Cas9 system significantly triggered a series of tumor-suppressive cellular responses of ESCC cells, as indicated by the up-regulated proteins involved in the regulation of apoptotic process, positive regulation of programmed cell death, cellular response to DNA damage stimulus, negative regulation of cell cycle process and negative regulation of growth (Fig. 1a), and the down-regulated proteins involved in the regulation of microtubule cytoskeleton organization, positive regulation of cell cycle, negative regulation of apoptotic process, positive regulation of cell growth and protein neddylation (Supplementary Fig. S1). These findings indicated that downregulation of UBC12 activated a series of tumor-suppressive cellular responses, providing the rationality for further evaluation of UBC12 as a potential anti-ESCC target. To verify the above mass spectra findings, we first systematically evaluated the effects of UBC12 knockdown on malignant phenotypes of ESCC cells. We found that UBC12 knockdown dramatically inhibited the cell proliferation, colony formation (Fig. 1b, c), as well as the transwell migration and invasion abilities of ESCC cells (Supplementary Fig. S2a–d). In mechanisms, the level of global protein neddylation was remarkably suppressed upon UBC12 knockdown. Moreover, UBC12 downregulation dramatically decreased the neddylation levels of cullin 1, 2, 3, 4A, and 4B, the substrates of UBC12, but not cullin 5, the substrate of UBE2F (Fig. 1d), that such, resulting in the inactivation of CRLs and the corresponding induction of abnormal cellular responses. GO analysis results of proteins with altered expression suggested that the cell cycle procession was remarkably disturbed upon UBC12 knockdown (Fig. 1a and Supplementary Fig. S1), as confirmed by the upregulation of cell cycle inhibitors p27, p21 and Wee1, and the downregulation of M phase marker phosphorylated-histone H3 (Fig. 1d). Consistently, cell cycle profile analysis proved that the cell populations in G2 phase were significantly increased in both two UBC12-knockdown ESCC cell lines (Fig. 1e). Since p27, p21 and Wee1 also served as the substrates of CRLs, in turn, the half-lives of these cell cycle inhibitors were found to be dramatically extended owing to CRLs inactivation when silencing UBC12 (Fig. 1f). Therefore, UBC12 downregulation effectively blocked cullin neddylation to inactivate CRLs, leading to G2 phase cell cycle arrest in ESCC cells. Consistent with aforementioned mass spectra findings, we showed that CRLs substrates CDT1 and ORC1, two DNA replication licensing proteins, were obviously accumulated upon UBC12 knockdown, and subsequently led to DNA damage response, as reflected by increased levels of phosphorylated H2AX (Fig. 1g). Further phenotypic analysis revealed that UBC12 knockdown triggered senescence or apoptosis of ESCC cells in a cell linedependent manner (Fig. 1h, i). UBC12-knockdown EC1 cells displayed classical senescence morphology with an enlarged and flattened cellular shape and the positive staining for senescenceassociated β-Galactosidase (Fig. 1h). Unlike, the notable apoptosis feature of shrunk morphology was presented in UBC12-knockdown Kyse450 cells, which was accompanied with the significant increase of Annexin V-positive cells (Fig. 1i). Furthermore, we found that UBC12 knockdown induced the accumulation of CRLs substrate activating transcription factor 4 (ATF4), and therefore transactivated death receptor 5 (DR5) to trigger extrinsic apoptosis of Kyse450 cells. On the other hand, the apoptotic protein NOXA, another downstream target of ATF4, was also activated and triggered the intrinsic apoptosis (Fig. 1j). Additionally, NOXA downregulation via siRNA silencing significantly suppressed apoptotic induction induced by UBC12 downregulation (Fig. 1k). Together, these results suggested that UBC12 downregulation significantly diminished the growth of Kyse450 cells through activation of both extrinsic apoptosis and intrinsic apoptosis pathways. To investigate the therapeutic potential of UBC12 silencing in vivo, we established the subcutaneous-transplantation tumor model using EC1 cells. We found that downregulation of UBC12 significantly inhibited tumor growth, as analyzed by tumor growth curve (P < 0.001, Fig. 1l) and tumor weight (P < 0.001, Fig. 1m). Mechanistic studies revealed that silencing UBC12 led to cullin neddylation inhibition and CRLs inactivation. As a result, CRLs substrates p27, p21, Wee1, CDT1 and ORC1 were obviously accumulated (Fig. 1n). These findings were further supported by