Ssd induces the mitochondrial fragmentation via inhibition of phospho-Ser⁶³⁷-Drp1 and activation of Opa1 processing in Oma1 and p53-independent manner

To determine if and how mitochondrial dynamics are involved in the regulation of chemosensitivity in OVCA cells, tubular and fragmented mitochondrial morphology was defined in both chemosensitive (A2780s) and chemoresistant (Hey) cells, in the absence of CDDP and Ssd treatment (Figure ​(Figure2A).2A). CDDP alone significantly increased the number of fragmented mitochondria in only the chemosensitive cells and in a time-dependent manner, while Ssd alone significantly induced this response in both chemosensitive and chemoresistant cells. Moreover, the number of both chemosensitive and chemoresistant cells with fragmented mitochondria were significantly higher in the presence of both Ssd and CDDP than of either agent alone (Figure ​(Figure2B).2B). Western blot (WB) analysis indicated that CDDP alone failed to alter phospho-Ser⁶³⁷-Drp1 content in both cell lines, suggesting phospho-Ser⁶³⁷-Drp1 may not be involved CDDP-induced mitochondrial fragmentation in chemosensitive cells. However, Ssd alone decreased phospho-Ser⁶³⁷-Drp1 content in chemoresistant cells. Moreover, phospho-Ser⁶³⁷-Drp1 content and the ratio of phospho-Ser⁶³⁷-Drp1 to Drp1 were down-regulated to a greater extent by Ssd in the presence of CDDP, suggesting Ssd-induced mitochondrial fragmentation may be mediated via decreased phospho-Ser⁶³⁷-Drp1 and the presence of CDDP sensitized Ssd-effect in chemoresistant cells (Figure ​(Figure2C).2C). Importantly, Ssd-induced mitochondrial fragmentation and apoptosis were attenuated by the Drp1 inhibitor mitochondrial division inhibitor-1 (mdivi-1) (Figure ​(Figure2D),2D), suggesting that Drp1 may be a pivotal participant in the Ssd-induced fragmentation and apoptosis in chemoresistant cells.

Open in a separate window

Figure 2

Ssd-induced mitochondria fragmentation via decrease of phospho-Ser637-Drp1 in chemoresistant OVCA cells

(A) OVCA cells showed two major mitochondrial phenotypes: tubular and fragmented. The inset is an enlarged image of box area. (B) Ssd induced mitochondrial fragmentation and sensitized chemoresistant OVCA cells to CDDP. A2780s and Hey cells were cultured with CDDP (10 μM) and/or Ssd (1 μM, 0-6 h) and mitochondrial phenotype was assessed (n=3). ^#P<0.05, ^##P<0.01 and ^###P<0.001 (vs respective Ssd) and ⁺⁺⁺P<0.001 (vs respective CDDP). (C) Ssd decreased phospho-Ser⁶³⁷-Drp1 content and the combination with CDDP decreased phospho-Ser⁶³⁷-Drp1 content and the ratio of phospho-Ser⁶³⁷-Drp1 to Drp1 more than Ssd alone. A2780s and Hey cells were cultured with CDDP (10 μM) and/or Ssd (1 μM, 0-24 h). Drp1, phospho-Ser⁶³⁷-Drp1 and GAPDH contents were assessed by WB (n=3).^*P<0.05, ^**P<0.01 and^***P<0.001 (vs respective CTL). (D) Ssd-induced mitochondrial fragmentation and apoptosis were attenuated by mdivi1. Hey cells were cultured with CDDP (10 μM), Ssd (1 μM) and mdivi1 (0-20 μM, 24 h). Mitochondrial phenotype and apoptosis were assessed (n=3). ^*P<0.05, ^**P<0.01 and ^***P<0.001 (vs respective mdivi1 = 0 μM).

Our previous studies have shown the proteolytic activation of the mitochondrial fusion protein Opa1 by Oma1 (evident by 2 intact bands and 3 cleaved products on WB) results in mitochondrial fragmentation and is also a determinant of CDDP sensitivity in OVCA [15]. We have examined the influence of CDDP, alone or in combination with Ssd, on the ratio of short form to long form as an indication of Opa1 activation. CDDP (alone or with Ssd) increased this ratio in chemosensitive cells, but Ssd alone was ineffective, suggesting CDDP-induced mitochondrial fragmentation in chemosensitive cells may be mediated via Opa1 processing. In contrast, while CDDP alone appeared to have no effect on the ratio of the Opa1 isoform in chemoresistant cells, its action was facilitated by the presence of Ssd, suggesting that Ssd may sensitize CDDP-induced Opa1 processing in chemoresistant cells. We have previously shown that CDDP induces the p53-depending activation of Oma1 (40kDa) and subsequent mitochondrial fragmentation in chemosensitive but not chemoresistant OVCA cells [15]. As expected, CDDP alone increased the Oma1 content in chemosensitive but not chemoresistant cells (Supplementary Figure 2). Oma1 activation was not observed in the presence of Ssd irrespective of the presence of CDDP, suggesting this combination induces Opa1 processing directly or via pathways independent of Oma1 and p53. Collectively, our findings suggest mitochondrial fragmentation in chemosensitive cells is mainly induced by CDDP via Oma1 activation/Opa1 processing in a p53-dependent manner while decreased phospho-Ser⁶³⁷-Drp1 is mainly involved in Ssd-induced mitochondrial fragmentation and the sensitization of chemoresistant cells to CDDP.

Ssd increases [Ca²⁺]c and, in combination with CDDP, induces MMP loss and activates CaMKI in OVCA cells

Previous studies have shown that Ssd increases [Ca²⁺]c by direct inhibition of sarcoplasmic/endoplasmic reticulum calcium ATPase (SERCA), leading to apoptosis and autophagy [7]. CDDP can also increase [Ca²⁺]c in some cancer cells, leading to apoptosis [19–22]. In the present study, Ssd increased [Ca²⁺]c in both chemosensitive (A2780s) and chemoresistant (A2780cp) cells (Supplementary Figure 3A) while CDDP alone was ineffective and had no effect on Ssd-induced Ca²⁺ dynamics (Figure ​(Figure3A).3A). Although Ssd or CDDP alone showed no significant loss in MMP, a significant decrease in MMP was observed in the presence of both CDDP with Ssd (Figure ​(Figure3B3B and Supplementary Figure 3B), supporting the concept that Ssd increases [Ca²⁺]c which in turn is directed into mitochondria by CDDP.

Open in a separate window

Figure 3

Ssd-increased [Ca2⁺]c and subsequent MMP loss and activation of CaMKI in chemoresistant OVCA cells

(A) Ssd increased [Ca²⁺]c in OVCA cells, but not CDDP. A2780s and A2780cp cells were cultured with CDDP (10 μM) and/or Ssd (2 μM) and [Ca²⁺]c was measured using the FLIPR Calcium 6 Assay Kit. (B) CDDP with Ssd caused MMP loss. A2780cp cells were cultured with CDDP (10 μM) and/or Ssd (2 μM, 24 h) and immunostained with JC-1. (C) CDDP in the presence of Ssd increased phospho-Thr¹⁷⁷-CaMKI contents and the ratio of phospho-Thr¹⁷⁷-CaMKI to CaMKI. A2780s and Hey cells were cultured with CDDP (10 μM) and/or Ssd (1 μM, 0-24 h). CaMKI, phospho-Thr¹⁷⁷-CaMKI and GAPDH contents were assessed by WB.^*P<0.05, ^**P<0.01 and ^***P<0.001 (vs respective CTL). (n=3).

Ssd could induce autophagy through the increase of [Ca²⁺]c and the subsequent activation of Ca²⁺/calmodulin-dependent kinase kinase β (CaMKKβ)–adenosine monophosphate-activated protein kinase (AMPK) pathway [7]. CaMKI is a substrate of CaMKKβ and is also involved in the induction of mitochondrial fragmentation via the activation of Drp1 [13, 14]. Irrespective of the presence of Ssd, CDDP increased the contents of both phospho-Thr¹⁷⁷-CaMKI and CaMKI, but not the phospho-Thr¹⁷⁷-CaMKI/CaMKI ratio in chemosensitive cells (A2780s). Ssd alone was ineffective, suggesting CaMKI-Drp1 pathway may be less involved in CDDP-induced mitochondrial fragmentation than Opa1 processing. In contrast, while CDDP alone appeared to have no effect on the contents of phospho-Thr¹⁷⁷-CaMKI and the phospho-Thr¹⁷⁷-CaMKI/CaMKI ratio in chemoresistant cells, its action was facilitated by the presence of Ssd, suggesting Ssd may sensitize CDDP-induced CaMKI phosphorylation in chemoresistant cells (Figure ​(Figure3C).3C). In summary, the increase of [Ca²⁺]c by Ssd can induce the MMP loss and phosphorylation of CaMKI in the presence of CDDP, resulting in Drp1-mediated mitochondrial fragmentation in chemoresistant cells.

Ssd, in combination with CDDP, decreases mitosis and increases subsequent apoptosis via inhibition of Cdk1 and Cyclin B1 complex in OVCA cells

In addition to Drp1-mediated mitochondrial dynamics, the Cdk1/Cyclin B1 complex is a major protein kinase involved in mitosis [10]. We demonstrated that CDDP with Ssd significantly decreased the contents of Cdk1 and Cyclin B1 in chemoresistant cells while the combination also significantly decreased the contents of Cdk1 in chemosensitive cells (Figure ​(Figure4A).4A). Moreover, the combination increased the inactive contents of phospho-Tyr¹⁵-Cdk1 in cytoplasm in both chemosensitive and chemoresistant cells (Supplementary Figure 4A) while Ssd or CDDP alone was ineffective, suggesting this combination may inhibit Cdk1/Cyclin B1 complex and the translocation of this complex to nucleus. We then determined whether inhibition of Cdk1/Cyclin B1 complex in OVCA cells by Ssd (alone or in combination with CDDP) could lead to G2/M arrest and apoptosis by examining the changes in the amount of mitotic, non-mitotic and apoptotic cells. CDDP increased apoptosis and decreased mitosis in chemosensitive cells, both responses were further enhanced by the presence of Ssd. Although CDDP or Ssd alone elicited minimal or no apoptotic and mitotic response in chemoresistant cells, the responses to CDDP were markedly enhanced by the presence of Ssd (Supplementary Figure 5). In apoptotic cells, the signal of Cdk1 and Cyclin B1 were much weaker than in the cells at interphase, suggesting that Ssd could sensitize chemoresistant cells to CDDP by inhibiting the Cdk1/Cyclin B1 complex, leading to G2/M arrest and subsequent apoptosis. This notion was supported by the observation that while CDDP alone increased the Cdk1/Cyclin B1 interaction in the nucleus, but not in the cytoplasm in chemosensitive cells (A2780s; as determined by in Situ Proximity Ligation Assay (PLA)) and this response was significant enhanced by the presence of Ssd. While CDDP and Ssd alone failed to influence nuclear Cdk1/Cyclin B1 interaction in chemoresistant cells (Hey), significant up-regulation of this response was evident as early as 6 h in the presence of both agonists (Figure ​(Figure4B4B).

Open in a separate window

Figure 4

Ssd, in combination with CDDP, decreases mitosis and increases subsequent apoptosis via inhibition of Cdk1 and Cyclin B1 complex in OVCA cells

(A) CDDP with Ssd decreased Cdk1 and CyclinB1 contents. A2780s and Hey cells were cultured with CDDP (10μM) and/or Ssd (1μM, 0-24h). Cyclin B1, Cdk1 and GAPDH contents were assessed by WB. (B) CDDP with Ssd increased the interaction between Cdk1 and Cyclin B1 in chemoresistant OVCA cells. A2780s and Hey cells were cultured with CDDP (10 μM) and Ssd (1 μM, 0-12h) and the interaction between Cdk1 and Cyclin B1 was determined by PLA. ^*P<0.05, ^**P<0.01 and^***P<0.001 (vs respective CTL) and ⁺⁺P<0.01 and ⁺⁺⁺P<0.001 (vs respective CDDP). (n=3).

Although CDDP but not Ssd alone increased the cytoplasmic levels of phospho-Tyr¹⁵-Cdk1 (inactive form of Cdk1) in chemosensitive but not chemoresistant cells, this response was markedly enhanced in both cell lines in the presence of both CDDP and Ssd (Supplementary Figure 4B). Thus, together CDDP and Ssd decreased the protein contents of Cdk1 and Cyclin B1, increased the levels of phospho-Tyr¹⁵-Cdk1, and induced G2/M arrest and apoptosis.

Cell lines and culture

CDDP-sensitive [A2780s (wt-p53)] and -resistant [A2780cp (p53-mutant), Hey (wt-p53) and SKOV3 (p53-null)] human OVCA cell lines, gifts from Drs Rakesh Goel and Barbara Vanderhyden (Ottawa Regional Cancer Center, Ottawa, ON, Canada), were cultured as previously reported [40, 41]. CDDP-sensitive [SGC-7901] and -resistant [SGC-7901cp] gastric cancer cells (KeyGEN BioTECH, China) were cultured in RPMI-1640 medium. A2780s cells and their resistant counterparts A2780cp cells are epithelial OVCA cells. Hey cells are from moderately differentiated OVCA. SKOV3 cells are of clear cell carcinoma origin. The cells with the different p53-status (A2780s, A2780cp, Hey and SKOV3 cells) were used to explain that Ssd sensitized chemoresistant OVCA cells to CDDP regardless of p53-status. The same p53-status chemosesitive A2780s and chemoresistant Hey cells were used to explain the mechanisms of Ssd-induced sensitization.

Adenoviral infection

SKOV3 cells were infected with adenoviral wt-p53 (multiplicity of infection (MOI) = 10, 5 h) or LacZ (as control) as previously described [40, 41].

Apoptosis assay

Apoptosis was determined morphologically using the Hoechst 33258 nuclear stain, as previously described [40, 41]. At least 300 cells were assessed per experimental group.

Cell viability assay

Cell viability was determined using the 3-(4,5-di-methylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) and cell viability assays [3].

Assessment of mitochondrial phenotypes

Assessment of mitochondrial morphology was carried out as previously described [15]. Cells were seeded in 6-well plates, trypsinized, and washed with phosphate buffered saline and centrifuged (900 x g, 1 min). 100μl of the cell suspension was centrifuged (450 x g, 4 min; Shandon Cytospin 4, Thermo Fisher Scientific) using cytological funnels together with silane-coated glass slides [29]. For immunostaining, cells were fixed in 4% paraformaldehyde (1 h, room temperature (RT)) and blocked with 3 % bovine serum albumin. Mitochondria were incubated with anti-Tom20 antibody and appropriate secondary antibody. Confocal images were obtained on Zeiss LM510 inverted microscope (Carl Zeiss, Jena, Germany) with appropriate argon lasers (488nm) and 40X objective. Mitochondrial phenotype of each cell was categorized as tubular or fragmented, as previously described [29]. A cell with “fragmented mitochondria” was defined as the cell with 8 or more individual mitochondrial fragments that were each 3 μm in length across the longest axis. At least 300 cells were analyzed per experimental group.

Measurement of [Ca²⁺]c

Changes in [Ca²⁺]c were measured by a fluorescent dye, Fluo-3-AM as previously described [42, 43]. A2780s and A2780cp cells were treated with Ssd, cell suspensions were incubated with Fluo-3-AM (5 μM, 37°C, 30 min) and subjected to Fluorescence-activated cell sorting (FACS) analysis (BD Biosciences). At least 10,000 events were analyzed.

Measurement of cytoplasmic calcium dynamic

Intracellular cytosolic Ca²⁺ dynamic was measured using the FLIPR Calcium 6 Assay Kit (Molecular Devices) according to the manufacturer’s instructions. In brief, A2780s and A2780cp cells were plated at a density of 10,000 cells per well in black wall/clear bottom 96-multiwell plates from Costar (Tewksbury, MA, USA) and cultured for 24 h before treatment. After that, calcium 6 reagent was added directly to cells, and cells were incubated for an additional 2 h at 37 °C and 5% CO2. Ssd and/or CDDP were then added to the wells and immediately subjected to data acquisition on the FLIPR Tetra High-Throughput Cellular Screening System (Molecular Devices) at RT using a 1 s reading interval throughout the experiments.

MMP detection

Visualization of MMP loss was performed by JC-1 staining. A2780s and A2780cp cells were seeded on confocal disc, treated with Ssd and/or CDDP and stained with JC-1 dye (10 μg/ml, 30 min, 37°C). The retained MMP (Tetramethylrhodamine-isothiocyanate (TRITC) signal) and the loss (Fluoresceinisothiocyanate (FITC) signal) were real-time examined and captured under epifluorescence microscopy (60X objective; Applied Precision DeltaVision Elite, Applied Precision, Inc, USA). At least 500 cells were analyzed per experimental group.

Protein extraction and western blot analysis

The procedures were performed as previously described [40, 41] and band densities were analyzed (Scion Image software; Scion Corporation, Frederick, MD, USA).

Immunofluorescence and in situ proximity ligation assay (PLA)

Immunofluorescence experiments were performed as previously described [44]. At least 100 cells were analyzed per experimental group.

PLA was conducted according to manufacturer’s instructions (Duolink kit^®, Sigma-Aldrich). Cells were incubated with a pair of anti-Cyclin B1 and anti-Cdk1antibody, followed by secondary proximity probes (PLA probe PLUS and MINUS). Subsequently, they were incubated with the ligation solution containing ligase and amplification solution containing polymerase and a fluorophore with 594nm excitation and 624 nm emission. Then, cells were incubated with anti-β tubulin antibody and appropriate secondary antibody. Fluorescence signals were detected by Zeiss Axioplan 2 upright fluorescence microscope (40X objective; Carl Zeiss) and analyzed with AxioVision software (Carl Zeiss). PLA-positive signals were quantified using Duolink Image Tool (Sigma-Aldrich). At least 30 cells were analyzed per experimental group.

Flow cytometry

Cell cycle analysis was performed by flow cytometry, as previously described [17]. A2780s and Hey cells were treated with CDDP and/or Ssd. Cell cycle analysis was performed using BD LSRFortessa Flow Cytometer (BD Biosciences) and data were analyzed with Kaluza Analysis Software (Beckman Coulter Life Sciences, Indianapolis, IN, USA). At least 10,000 events were analyzed per experimental group.

We are thankful to Drs Rakesh Goel and Barbara Vanderhyden for providing the cell lines.