SIRT6 promotes mitochondrial fission and subsequent cellular invasion in ovarian cancer

Abstract

Ovarian cancer ranks fifth in terms of cancer mortality in women due to lack of early diagnosis and poor clinical management. Characteristics like high cellular proliferation, EMT and metabolic alterations contribute to oncogenicity. Cancer, being a "metabolic disorder," is governed by various key regulatory factors like metabolic enzymes, oncogenes, and tumor suppressors. Sirtuins (SIRT1-SIRT7) belong to the group of NAD⁺ deacetylase and ADP-ribosylation enzymes that function as NAD⁺ sensors and metabolic regulators. Among sirtuin orthologs, SIRT6 emerges as an important oncogenic player, although its possible mechanistic involvement in ovarian cancer advancement is still elusive. Our data indicated a higher expression of SIRT6 in ovarian cancer tissues compared with the non-malignant ovarian tissue. Further, we observed that overexpression of SIRT6 enhances glycolysis and oxidative phosphorylation in ovarian cancer cells. The energy derived from these processes facilitates migration and invasion through invadopodia formation by reorganization of actin fibers. Mechanistically, SIRT6 has been shown to promote ERK1/2-driven activatory phosphorylation of DRP1 at serine-616, which has an obligatory role in inducing mitochondrial fission. These fragmented mitochondria facilitate cell movement important for metastases. siRNA-mediated downregulation of SIRT6 was found to decrease cellular invasion through compromised mitochondrial fragmentation and subsequent reduction in stress fiber formation in ovarian cancer cells. Thus, the present report establishes the impact of SIRT6 in the regulation of morphological and functional aspects of mitochondria that modulates invasion in ovarian cancer cells.

Keywords: SIRT6; actin polymerization; aerobic glycolysis; mitochondrial fragmentation.

Fig. 1

SIRT6 protein expression is high in ovarian cancer, which regulates the metastatic phenotype (a) Immunohistochemical staining shows high SIRT6 levels in high‐grade carcinoma patient tissue sections with corresponding bar diagram for expression levels. Scale bar = 10 μm. n = 3. Paired two‐tailed t‐test was done to calculate the P‐value. (B) Human Protein Atlas data show high SIRT6 expression in ovarian cancer tissue. (C) qPCR data show elevated SIRT6 levels in PA1 cell lines compared with IOSE‐364. n = 3, Paired two‐tailed t‐test was done to calculate the P‐value. (D) Endogenous SIRT6 expression in IOSE‐364 vs. PA1 and PA1 vs. SKOV3 cell lines. (E) Ki‐67 staining in control (untransfected), EV‐transfected, SIRT6‐overexpressed, and SIRT6 siRNA transfected PA1 cells and relative fluorescence intensity have been depicted as bar graphs. Scale bar = 10 μm, n = 3 (F) qPCR data showed non‐significant change in MMP2 and an increase in MMP9 in EV‐transfected vs. pcDNA 3.1‐SIRT6‐transfected PA1 cells. n = 3, Paired two‐tailed t‐test was done to calculate the P‐value. (G) Matrigel invasion studies show the number of invaded cells in IOSE‐364, untransfected PA1 (control), EV‐transfected, SIRT6‐overexpressed, scrambled siRNA, and SIRT6‐silenced PA1 cells. Corresponding bar diagram to show the number of invaded cells per field. Scale bar = 200 μm. n = 3, ANOVA was done to calculate the P‐value. Error bars represent standard error of mean (SEM) from three independent experiments. ns—non‐significant, *P‐value < 0.05, **P‐value < 0.01, ***P‐value < 0.001. [Colour figure can be viewed at wileyonlinelibrary.com]

Fig. 2

Transcriptomics data indicate the tumorigenic role of SIRT6. (A) Volcano plot depicting a total of 16 558 genes among which 16256 genes were at the baseline level, 161 showed upregulation (red dots) and 141 genes were downregulated (green dots). (B) Heat map was generated according to −log₁₀P values with lower ‐log₁₀P having high significance levels. (C) Data obtained were grouped into different GO pathways. (D) Venn diagrams were produced considering 3 heads with maximum genes involved that are altered by SIRT6 viz. mitochondria, metabolism, and cancer. [Colour figure can be viewed at wileyonlinelibrary.com]

Fig. 3

Overexpression of SIRT6 leads to a glycolytic phenotype in ovarian cancer. (A, B) Seahorse Extracellular flux analysis shows elevated ECAR rate in EV‐transfected vs. pcDNA 3.1‐SIRT6‐transfected PA1 cells and decreased ECAR rate in scrambled siRNA vs. SIRT6‐silenced sample sets in PA1 cells. (C) Cell invasion assay in Control (untransfected), EV‐transfected, SIRT6‐transfected, MMP9 inhibitor‐treated, SIRT6 + MMP9 inhibitor‐treated, 2DG‐treated, and SIRT6 + 2DG‐treated PA1 cells. Number of invaded cells is plotted as bar graphs. n = 3, Scale bar = 200 μm. Paired two‐tailed t‐test was done to calculate the P‐value. Error bars represent standard error of mean (SEM) from three independent experiments. ns—non‐significant *P‐value < 0.05, **P‐value < 0.01, ***P‐value < 0.001. [Colour figure can be viewed at wileyonlinelibrary.com]

Fig. 4

SIRT6 affects mitochondrial morphology through DRP1 phosphorylation. (A) Seahorse extracellular flux analysis depicts higher ATP production and Basal respiration in EV and SIRT6‐transfected PA1 cells. n = 3, Paired two‐tailed t‐test was done to calculate P‐value. (B) Western blot analysis shows p‐DRP1^ser616, p‐ERK1/2, and PCNA levels in Control (untransfected), EV‐transfected, and pcDNA 3.1‐SIRT6 in PA1 cell line. α‐tubulin was used as gel loading control. (C) Western blot analysis of phospho‐DRP1^ser616, total DRP1 and α‐tubulin in EV, pcDNA 3.1‐SIRT6, EV + ERK inhibitor, and pcDNA 3.1‐SIRT6 + ERK inhibitor‐treated cells. α‐tubulin was used as gel loading control. (D) Confocal microscopy with Mitotracker red (mitochondria) and DAPI (nuclei) in Control (untransfected), EV, pcDNA 3.1‐SIRT6, and SIRT6‐silenced PA1 cell line. Mitochondrial length distribution over the sample sets was measured using Leica software. (E) Live videography with Mitotracker Green AF488 in Control (untransfected) PA1 and SIRT6‐transfected PA1 cells shown as time‐dependent images. Scale bar = 10 μm. n = 3, ANOVA was done to calculate the P‐value. Error bars represent standard error of mean (SEM) from three independent experiments. ns—non‐significant, *P‐value < 0.05, **P‐value < 0.01, ***P‐value < 0.001. [Colour figure can be viewed at wileyonlinelibrary.com]

Fig. 5

SIRT6 directs mitochondrial fission by augmenting actin polymerization. (A, B) Confocal microscopy in Control (untransfected), EV and pcDNA 3.1‐SIRT6‐transfected IOSE‐364 and PA1 cells with Phalloidin Green AF488 (actin), Mitotracker Red (mitochondria), and DAPI (nuclei). Bar graphs depicted for calculating mitochondrial length using Leica software. Scale bar = 10 μm. n = 3, ANOVA was done to calculate the P‐value. Error bars represent standard error of mean (SEM) from three independent experiments. ns—non‐significant *P‐value < 0.05, **P‐value < 0.01, ***P‐value < 0.001. [Colour figure can be viewed at wileyonlinelibrary.com]

Fig. 6

SIRT6 enhances ovarian cancer cell migration and invasion through mitochondrial fission. (A) Phalloidin Green AF488 + Mitotracker Red staining in Control (untransfected) and SIRT6‐transfected PA1 cells was carried out. Scale bar = 10 μm, Line‐scan analysis between Phalloidin Green AF488 and Mitotracker Red stained images was done using ImageJ software. (B) Matrigel invasion studies to show the number of invaded cells in sample sets of EV, pcDNA 3.1‐SIRT6, EV+ Oligomycin‐A, pcDNA 3.1‐SIRT6 + Oligomycin‐A, EV + mdivi‐1, pcDNA 3.1‐SIRT6 + mdivi‐1, scrambled siRNA, and SIRT6 siRNA. Scale bar = 200 μm. (C) Wound healing assay to show the distance after wound healing in sample sets EV, pcDNA 3.1‐SIRT6, EV+ Oligomycin‐A, pcDNA 3.1‐SIRT6 + Oligomycin‐A, EV + mdivi‐1, pcDNA 3.1‐SIRT6 + mdivi‐1, scrambled siRNA, and SIRT6 siRNA. Scale bar = 400 μm. n = 3, ANOVA was done to calculate the P‐value. Error bars represent standard error of mean (SEM) from three independent experiments. ns—non‐significant *P‐value < 0.05, **P‐value < 0.01, ***P‐value < 0.001. [Colour figure can be viewed at wileyonlinelibrary.com]