cNEK6 induces gemcitabine resistance by promoting glycolysis in pancreatic ductal adenocarcinoma via the SNRPA/PPA2c/mTORC1 axis

Abstract

Resistance to gemcitabine in pancreatic ductal adenocarcinoma (PDAC) leads to ineffective chemotherapy and, consequently, delayed treatment, thereby contributing to poor prognosis. Glycolysis is an important intrinsic reason for gemcitabine resistance as it competitively inhibits gemcitabine activity by promoting deoxycytidine triphosphate accumulation in PDAC. However, biomarkers are lacking to determine which patients can benefit significantly from glycolysis inhibition under the treatment of gemcitabine activity, and a comprehensive understanding of the molecular mechanisms that promote glycolysis in PDAC will contribute to the development of a strategy to sensitize gemcitabine chemotherapy. In this study, we aimed to identify a biomarker that can robustly indicate the intrinsic resistance of PDAC to gemcitabine and guide chemotherapy sensitization strategies. After establishing gemcitabine-resistant cell lines in our laboratory and collecting pancreatic cancer and adjacent normal tissues from gemcitabine-treated patients, we observed that circRNA hsa_circ_0008383 (namely cNEK6) was highly expressed in the peripheral blood and tumor tissues of patients and xenografts with gemcitabine-resistant PDAC. cNEK6 enhanced resistance to gemcitabine by promoting glycolysis in PDAC. Specifically, cNEK6 prevented K48 ubiquitination of small ribonucleoprotein peptide A from the BTRC, a ubiquitin E3 ligase; thus, the accumulated SNRPA stopped PP2Ac translation by binding to its G-quadruplexes in 5' UTR of mRNA. mTORC1 pathway was aberrantly phosphorylated and activated owing to the absence of PP2Ac. The expression level of cNEK6 in the peripheral blood and tumor tissues correlated significantly and positively with the activation of the mTORC1 pathway and degree of glycolysis. Hence, the therapeutic effect of gemcitabine is limited in patients with high cNEK6 levels, and in combination with the mTORC1 inhibitor, rapamycin, can enhance sensitivity to gemcitabine chemotherapy.

Fig. 1. cNEK6 promotes gemcitabine chemotherapy resistance in PDAC.

A Venn diagram showing the intersection of differentially expressed circRNAs among the three gene sets. B Relative RNA levels of cNEK6 and NEK6 after actinomycin D treatment at different time points (n = 3) (RT-qPCR). C RT-qPCR was used to analyze the expression of cNEK6 and NEK6 after RNase R treatment of PANC-1 and PaTu8988t cells (n = 3). D Subcellular localization of cNEK6 detected by fluorescence in situ hybridization (FISH). Scale bars = 5 μm. E. Expression levels of cNEK6 between gemcitabine-resistant groups (PANC-1 GR and PaTu8988t GR) and their control groups (PANC-1 WT and PaTu8988t WT) (n = 3). F–H Effect of cNEK6 on the proliferative ability (IC50 (F), cell viability (G), and colony formation (H)) of gemcitabine-resistant groups (PANC-1 GR and PaTu8988t GR) and their control groups (PANC-1 WT and PaTu8988t WT) treated with gemcitabine (n = 3). I–K Schematic representation of the treatment regimen for patient-derived tumor xenograft (PDX) models (I) (n = 5). Xenografts were isolated and measured after euthanasia (J, K). L. CircRNA in situ hybridization (ISH) for cNEK6 in PDAC tissues. Scale bars = 100 μm (n = 20). M Kaplan–Meier analysis of overall survival and progression-free survival of patients with PDAC based on the expression of cNEK6.

Fig. 2. cNEK6 enhances glycolysis in PDAC to resist gemcitabine.

A Bubble plot showing GO and KEGG analyses of differentially expressed genes between the cNEK6 over-expression and control PANC-1 cells. B Glucose uptake, pyruvate levels, lactate production, and ATP levels were compared between gemcitabine-resistant groups (PANC-1 GR and PaTu8988t GR) and their control groups (PANC-1 WT and PaTu8988t WT) (n = 3). C ECAR was determined between the PANC-1 GR and PANC-1 WT cell lines (n = 3). D Effect of cNEK6 on glucose uptake, pyruvate levels, lactate production, and ATP levels in PANC-1 GR cells and their control, PANC-1 WT cells (n = 3). E ECAR showing the effect of cNEK6 on glycolysis in PANC-1 GR cells and control PANC-1 WT cells (n = 3). F–H Effect of the glycolysis inhibitor, 2-DG, on cNEK6 induced changes on proliferative ability (IC50 (F), colony formation (G), and cell viability (H)) of PANC-1 WT and PaTu8988t WT cell lines after gemcitabine treatment (n = 3).

Fig. 3. cNEK6 promotes glycolysis in PDAC through SNRPA.

A, B Cell lysate from PNC-1 cells was incubated with the cNEK6 probe. Potential cNEK6-binding proteins were pulled down, followed by RAP assay, visualized by silver staining (A), and identified by mass spectrometry (B). C Binding of cNEK6 to AGO2 or SNRPA was determined using an RNA pull-down assay. D Binding of AGO2 or SNRPA to cNEK6 was determined using the RIP assay (n = 3). E Subcellular localization of cNEK6 and SNRPA detected by FISH and immunofluorescence, respectively. F Protein levels of SNRPA in PANC-1 GR cell lines and their control, PANC-1 WT cells, transfected with sh-cNEK6 and OE-cNEK6, respectively. G, H Effect of SNRPA knockdown on cNEK6 induced changes in glucose uptake, pyruvate levels, lactate production, ATP levels (G), and ECAR (H) in PANC-1 WT cell lines (n = 3). I, J. Effect of SNRPA supplementation on sh-cNEK6 induced changes in glucose uptake, pyruvate levels, lactate production, ATP levels (I), and ECAR (J) in PANC-1 GR cell lines (n = 3).

Fig. 4. cNEK6 stabilizes SNRPA by suppressing the K48 ubiquitination.

A, B After cycloheximide (CHX, 100 μg/mL) (A) or MG132 treatment (10 μM) (B), temporal changes in SNRPA protein level in gemcitabine-resistance (PANC-1 GR and PaTu8988t GR) and control groups (PANC-1 WT and PaTu8988t WT) transfected with sh-cNEK6 and OE-cNEK6, respectively. C Effect of Chloroquine (CQ, 25 μM) or MG132 (10 μM) on cNEK6 induced changes of SNRPA protein level. D, E. Ubiquitination of SNRPA in gemcitabine-resistant (PANC-1 GR and PaTu8988t GR) and control groups (PANC-1 WT and PaTu8988t WT) transfected with sh-cNEK6 and OE-cNEK6, respectively. Western blot analysis was performed using antibodies against Ub (D), and K48-Ub (E).

Fig. 5. cNEK6 competitively binds with ubiquitin E3 ligase BTRC to SNRPA to inhibit its K48 ubiquitination.

A The protein levels of deubiquitinases in PANC-1 GR cell lines and control PANC-1 WT cells transfected with sh-cNEK6 and OE-cNEK6. B Binding of FBXW11 or BTRC with SNRPA was conducted by COIP in PANC-1 cells. C After co-transfecting of Myc-tagged full-length BTRC and full-length or different truncations of SNRPA with FLAG, COIP experiments were performed. D After co-transfecting FLAG-tagged full-length SNRPA and full-length or different truncations of BTRC with Myc, COIP experiments were performed. E RNA pull-down assay was performed using the wild-type or mutated probe of cNEK6 and wild-type SNRPA. F RNA pull-down assay was performed using the wild-type probe of cNEK6 and full-length or different truncations of SNRPA with FLAG. G Binding of BTRC with SNRPA was conducted by COIP experiment in PANC-1 GR and PANC-1 WT cell lines transfected with sh-cNEK6 and OE-cNEK6, respectively. H Binding of BTRC to SNRPA was conducted by COIP experiment in PANC-1 cells transfected with wild-type or mutant cNEK6. I Ubiquitination activities of BTRC, UBE2D3, and SNRPA with wild-type or mutant cNEK6. J Ubiquitination of Flag-SNRPA (WT or mutant) in 293 T cells transfected with Myc-BTRC or HAUb-K48. K Effect of wild-type or mutant SNRPA on cNEK6 induced changes in ECAR in PANC-1 GR cell lines (n = 3).

Fig. 6. SNRPA activates the mTROC1 pathway by recognizing G-quadruplex structures in PP2Ac.

A Venn diagram showing the intersection of glycolysis-related genes and genes containing G-quadruplex structures. B Protein levels of PP2Ac in OE-Vector/OE-cNEK6 PANC-1 WT cells transfected with or without sh-SNRPA and in sh-NC/sh-cNEK6 PANC-1 GR cells transfected with or without SNRPA. C Protein levels of PP2Ac and several downstream genes in the mTORC1 pathway. D Protein levels of SNRPA and several downstream genes in the mTORC1 pathway in OE-Vector/OE-cNEK6 PANC-1 WT cells transfected with or without PP2Ac. E–H After cycloheximide treatment (CHX, 100 μg/mL) (E, F) or MG132 (10 μM) treatment (G, H), temporal changes in PP2Ac protein level in PANC-1 GR cell lines, and their control, PANC-1 WT cell lines, with the indicated treatment. I EMSA of recombinant SNRPA binding the G4 structure. J EMSA shows the effect of PDS at different concentrations on the binding of SNRPA to the G4 structures. K, L EMSA of different truncations of SNRPA binding to the G4 structure. M Protein levels of PP2Ac in PANC-1 cells transfected with wild-type or different truncations of SNRPA.

Fig. 7. cNEK6 can serve as a biomarker for applying mTORC1 inhibitor to sensitize gemcitabine chemotherapy.

A Representative ISH images of cNEK6 and immunohistochemistry (IHC) images of p-S6K1 in pancreatic cancer tissues (scale bar: 100 μm). The correlation between the expression levels of cNEK6 and p-S6K1 was calculated based on ISH and IHC scores, respectively (n = 30). B Correlation between the expression level of cNEK6 as measured by RT-qPCR and the IHC score of p-S6K1 in pancreatic cancer tissues (n = 30). C Correlation between the expression level of cNEK6 in peripheral blood (P-cNEK6) (RT-qPCR), cNEK6 in pancreatic cancer tissues (T-cNEK6) (RT-qPCR), and p-S6K1 in pancreatic cancer tissues (IHC score) (n = 30). D–F Effect of the mTORC1 inhibitor, rapamycin, on the cNEK6 enhanced proliferative ability (colony formation (D), IC50 (E), and cell viability (F)) of PANC-1 WT lines under gemcitabine treatment. G Tumor growth in orthotopic models of nude mice treated with the indicated treatments (n = 5). H Representative images and statistical analysis of in vivo bioluminescence in orthotopic tumor models subjected to the indicated treatments (n = 5). I Overall survival of nude mice subjected to the indicated treatments (n = 5). J Representative images of IHC and hematoxylin and eosin staining in nude mice administered the indicated treatments. Scale bar: 100 µm.