Upregulating sirtuin 6 ameliorates glycolysis, EMT and distant metastasis of pancreatic adenocarcinoma with krüppel-like factor 10 deficiency

Abstract

Krüppel-like factor 10 (KLF10) is a tumor suppressor in multiple cancers. In a murine model of spontaneous pancreatic adenocarcinoma (PDAC), additional KLF10 depletion accelerated distant metastasis. However, Klf10 knockout mice, which suffer from metabolic disorders, do not develop malignancy. The mechanisms of KLF10 in PDAC progression deserve further exploration. KLF10-depleted and KLF10-overexpressing PDAC cells were established to measure epithelial-mesenchymal transition (EMT), glycolysis, and migration ability. A murine model was established to evaluate the benefit of genetic or pharmacological manipulation in KLF10-depleted PDAC cells (PDACshKLF10). Correlations of KLF10 deficiency with rapid metastasis, elevated EMT, and glycolysis were demonstrated in resected PDAC tissues, in vitro assays, and murine models. We identified sirtuin 6 (SIRT6) as an essential mediator of KLF10 that modulates EMT and glucose homeostasis. Overexpressing SIRT6 reversed the migratory and glycolytic phenotypes of PDACshKLF10 cells. Linoleic acid, a polyunsaturated essential fatty acid, upregulated SIRT6 and prolonged the survival of mice injected with PDACshKLF10. Modulating HIF1α and NFκB revealed that EMT and glycolysis in PDAC cells were coordinately regulated upstream by KLF10/SIRT6 signaling. Our study demonstrated a novel KLF10/SIRT6 pathway that modulated EMT and glycolysis coordinately via NFκB and HIF1α. Activation of KLF10/SIRT6 signaling ameliorated the distant progression of PDAC.Clinical Trial Registration: ClinicalTrials.gov. identifier: NCT01666184.

Fig. 1. Loss of Krüppel-like factor 10 (KLF10) correlated with enhanced migratory ability, epithelial-mesenchymal transition phenotypes, and distant metastasis of pancreatic adenocarcinoma (PDAC).

a Representative H&E (upper panel) and immunohistochemistry of KLF10 (middle and lower panels) in resected normal pancreas (left panel) and PDAC tissues of high (middle panel) and low (right panel) extent-intensity (EI) score immunostaining. The photos in the upper, middle, and lower panels are enlarged by ×100, ×100, and ×400, respectively. b Distant metastasis-free survival (DMFS) curves of 105 patients with curatively resected PDAC. The median DMFS was 15.5 and 25.4 months for patients with low and high KLF10 expression, respectively (p = 0.083). c Cumulated cell trajectory assay of Panc-1pLKO (left upper panel) and Panc-1shKLF10 (right upper panel) cells from 20 cell lines. Representative morphology of Panc-1pLKO (left lower panel) and Panc-1shKLF10 cells (right lower panel) enlarged by ×400 under bright field. d Cumulative migration (left panel) and invasion (right panel) assays of Panc-1pLKO (CTL) and Panc-1shKLF10 cells. Data are presented as the mean ± standard error (SE; * signifies p < 0.05). The experiments were repeated independently three times. e Representative in vivo imaging system (IVIS) images of mice injected with ASPC-1pLKO (upper panel) and ASPC-1shKLF10 (lower panel) for 1–4 weeks. Representative liver specimens with tumor nodules are displayed on the right side of each panel. f Cumulative IVIS signals of at least six mice in each experimental group injected with ASPC-1pLKO (black) and ASPC-1shKLF10 (gray) are shown at the time after injection. Data are presented as the mean ± SE (*** signifies p < 0.005). g Cumulative migratory assay data of Panc-1pLKO and Panc-1shKLF10 cells with and without forced replacement of KLF10 (* and ** represent p < 0.05 and p < 0.01, respectively). The experiments were repeated three times. h Representative immunoblots of the indicated proteins in Panc-1pLKO and Panc-1shKLF10 cells with and without forced replacement of KLF10. Quantitative analysis of at least three independent experiments is shown in addition to immunoblots. β-Actin was used as the internal control.

Fig. 2. Loss of KLF10 correlated with enhanced glycolysis in PDAC.

a Cumulated data of glucose uptake (left panel) and lactate production assays (right panel) in Panc-1pLKO and Panc-1shKLF10 cells. Data are presented as the mean ± SE (* signifies p < 0.05). The experiments were repeated three times. b Representative oxygen consumption rate in Panc-1pLKO (black) and Panc-1shKLF10 (gray) cells in response to oligomycin, carbonyl cyanide-p-trifluoromethoxyphenyl-hydrazon (FCCP), antimycin A (antiA), and rotenone (Rot). The results were normalized according to cell number and are presented as the mean ± SE (n = 3). c Basal and maximal respiration were determined in the same experimental setting for Panc-1pLKO and Panc-1shKLF10 cells in (b). d Representative immunoblots of KLF10 and glycolytic enzyme expression in Panc-1pLKO and Panc-1shKLF10 cells with and without KLF10 replacement. β-Actin was used as the internal control. Quantitative analysis of at least three independent experiments is shown beside the immunoblots. e Cumulated lactate production assay data of Panc-1pLKO and Panc-1shKLF10 cells with and without KLF10 replacement. Data are presented as the mean ± SE (* signifies p < 0.05). f Representative hematoxylin and eosin (H&E) staining and immunohistochemistry of KLF10, PKM2, and PDK1 in nontumor normal pancreas and pancreatic tissues (×100 and ×400 for the middle and right panels, respectively) from patients with curatively resected PDAC. g Correlation of immunolabeling of KLF10 with PKM2 from the pancreatic tissues of 10 patients from the cohort mentioned in the “Materials and methods” section of PDAC patients who underwent curative resection. The correlation coefficient = −0.5, p = 0.011. h Representative six databases from Oncomine with correlated levels of KLF10 and PKM2 transcripts.

Fig. 3. KLF10 bound to the promoter and transcriptionally upregulated sirtuin 6 (SIRT6).

a Primer maps of the SIRT6 promoter for the chromatin immunoprecipitation-polymerase chain reaction (PCR) assay (upper panel) and luciferase reporter assay (middle and lower panels). b Upper panel: DNA fragments of Panc-1 cells immunoprecipitated with KLF10 (KLF10), IgG (IgG), or positive control followed by PCR amplification of the SIRT6 promoter region that contains KLF10 binding sites (input), as described in the “Materials and methods” section. Lower panel: Quantitative PCR of the SIRT6 promoter region using samples from Fig. 3b, upper panel (* denotes p < 0.05). c Upper panel: SIRT6 promoter luciferase reporter activity in Panc-1 cells with or without KLF10 treatment by using wild-type or mutated SIRT6 promoter plasmids, as described in the “Materials and methods” section. Each bar represents the mean ± SE from at least two experiments (* and *** denote p < 0.05 and 0.005, respectively). Lower panel: SIRT6 promoter luciferase reporter activity in Panc-1 cells after various KLF10 treatments using wild-type (black) or mutated (gray) plasmids of the SIRT6 promoter as described in the “Materials and methods”. Each point represents the mean ± SE from at least two experiments (** denotes p < 0.01). d Representative immunoblots of KLF10 and SIRT6 expression in Panc-1 (left upper panel) and BXPC-3 (right upper panel) cells with and without KLF10 mRNA silencing. MiaPaCa (left lower panel) cells with and without doxycycline (Dox) to induce KLF10 mRNA overexpression. Right lower panel: RNA level of SIRT6 in Panc-1shKLF10. Glyceraldehyde 3-phosphate dehydrogenase (GAPDH) was used as the control. Quantitative analysis is shown below the blots. e Representative H&E staining (upper panel) and immunohistochemical staining of KLF10 (middle panel) and SIRT6 (lower panel) in the nontumor normal (left panel, ×100) and pancreatic cancer tissues (middle and right panels, ×100 and ×400, respectively) from patients with PDAC who underwent curative resection. f Correlation between the KLF10 expression level and SIRT6 in pancreatic tissue from 29 patients of the cohort mentioned in the “Materials and methods” with PDAC who underwent resection. Correlation coefficient = 0.54; p = 0.003. g Representative five databases from Oncomine display a positive correlation between KLF10 and SIRT6 transcripts.

Fig. 4. Overexpression of SIRT6 ameliorated the migration, EMT phenotype, glycolysis activity, and metastasis of PDAC.

a Cumulated data of the Transwell migratory assay of Panc-1pLKO and Panc-1shKLF10 cells with and without SIRT6 overexpression. Data are presented as the mean ± SE (** signifies p < 0.01). The experiments were repeated three times. b Cumulated data of directional migration over time (or. dist.) in Panc-1pLKO and Panc-1shKLF10 cells with and without SIRT6 overexpression. Data are presented as the mean ± SE (** signifies p < 0.01). The cell trajectory experiments were repeated three times. c Representative IVIS images of mice at 1–4 weeks after injection with ASPC-1pLKO and ASPC-1shKLF10 cells with and without SIRT6 overexpression as indicated. d Cumulated IVIS signal of at least six mice in each experimental group injected with ASPC-1pLKO and ASPC-1shKLF10 with and without SIRT6 overexpression. Data are presented as the mean ± SE (* represents p < 0.05). e Representative immunoblots of E-cadherin and mesenchymal and glycolytic protein expression in Panc-1pLKO and Panc-1shKLF10 cells with and without SIRT6 overexpression. β-Actin was used as the internal control. Quantitative analysis of at least three experiments is shown below the immunoblots. f Cumulated lactate production assay data of Panc-1pLKO and Panc-1shKLF10 cells with and without SIRT6 overexpression. Data are presented as the mean ± SE (** and *** signify p < 0.01 and p < 0.005, respectively). The experiments were repeated three times.

Fig. 5. Linoleic acid (LA) enhanced SIRT6 expression and reversed migration, EMT, and the glycolytic phenotypes of PDAC with KLF10 deficiency.

a Cumulative migratory assay of Panc-1pLKO and Panc-1shKLF10 cells with and without 50 μM LA treatment for 16 h. Data are presented as the mean ± SE (* and ** represent p < 0.05 and p < 0.01, respectively). The experiments were repeated three times. b Representative immunoblots of E-cadherin and mesenchymal and glycolytic protein expression in Panc-1pLKO and Panc-1shKLF10 cells with and without 50 μM LA treatment for 16 h. β-Actin was used as the internal control. Quantitative analysis of at least three experiments is shown below the immunoblots. c Upper panel: Representative IVIS image of mice at 1–4 weeks after injection with ASPC-1pLKO and ASPC-1shKLF10 with and without 1% LA treatment in daily water for 8 weeks as described in the “Materials and methods”. Lower panel: Cumulated IVIS signal of at least six mice in each experimental group that were injected with ASPC-1pLKO and ASPC-1shKLF10 and received 1% LA treatment in daily water for 8 weeks or did not, as indicated. Each point represents the mean ± SE at least six mice (* signifies p < 0.05). d Cumulated lactate production assay of Panc-1pLKO and Panc-1shKLF10 cells with and without LA treatment at 50 μM for 16 h. Data are presented as the mean ± SE (* represents p < 0.05). The experiments were repeated three times. e The survival curves of mice injected with Panc-1pLKO and Panc-1shKLF10 that received 1% LA treatment in daily water for 8 weeks or did not, as indicated. Each experimental group contained at least six mice. The median survival of mice injected with Panc-1shKLF10 without and with LA treatment was 37 and 50 days (p = 0.038; * signifies p < 0.05).

Fig. 6. HIF1α and NFκB are two downstream mediators of KLF10/SIRT6 signaling in PDAC cells.

a Representative databases from Oncomine display the correlation between KLF10 (upper panel)/SIRT6 (lower panel) and HIF1α or NFκB transcripts. b Representative immunoblots of HIF1α and phospho-NFκBp65 in Panc-1pLKO and Panc-1shKLF10 cells with and without forced expression of KLF10. Quantitative analysis of cumulative data from at least three experiments is shown in addition to immunoblots. β-Actin was used as the internal control. c Representative immunoblots of HIF1α in Panc-1 cells treated with various dosages of 1,4-dihydrophenonthrolin-4-one-3-carboxylic acid (DPCA) for 8 h. Quantitative analysis from at least two experiments is shown below the immunoblots. d Left panel: Cumulated migratory assay of Panc-1pLKO and Panc-1shKLF10 cells with or without SIRT6 overexpression treated with and without 400 μM DPCA for 16 h. Right panel: Cumulated lactate production assay of Panc-1pLKO and Panc-1shKLF10 cells with and without SIRT6 overexpression that were treated with and without 400 μM DPCA for 16 h. Data are presented as the mean ± SE (n = 3) (*, ** and *** indicate p < 0.05, p < 0.01, and p < 0.005, respectively). The experiments were repeated three times. e Representative immunoblots of NFκΒ in Panc-1 cells treated with various dosages of phorbol ester (PMA) for 8 h. Quantitative analysis from at least two experiments is shown below the immunoblots. f Left panel: Cumulated migratory assay of cells treated with and without 200 ng/ml PMA for 16 h. Right panel: Cumulated lactate production assay with and without 400 ng/ml PMA for 16 h. Data are presented as the mean ± SE (n = 3; *, **, and *** indicate p < 0.05, p < 0.01 and p < 0.005, respectively). The experiments were repeated three times. g Graphic summary of the role of the KLF10/SIRT6 signaling pathway in modulating EMT and glycolysis in PDAC.