The mammalian SPCA1 and SPCA2 ATPases localize in membranes of the secretory pathway and transport ions of Ca2+ and Mn2+. The role of tissue-specific SPCA2 isoform, highly expressed in lungs, mammary gland and gastrointestinal tract, is poorly understood. To elucidate the function of SPCA2, we studied human colon cancer HCT116 cells, grown under ambient and decreased O2 levels. We found that in contrast with other Ca2+-ATPase isoforms the expression of SPCA2 was up-regulated under hypoxia (3% O2), in both adherent (2D) and spheroid (3D) cultures. In spheroids, experiencing lowest O2 levels (30–50 μM, measured by phosphorescence lifetime imaging microscopy), we observed lower staining with reactive oxygen species (ROS)-specific fluorescent probe, which correlated with increased SPCA2. However, SPCA2 expression was up-regulated in cells exposed to reactive oxygen and nitrogen species donors, and when grown at higher density. We noticed that the culture exposed to hypoxia showed overall increase in S phase-positive cells and hypothesized that SPCA2 up-regulation under hypoxia can be linked to Mn2+-dependent cell cycle arrest. Consequently, we found that SPCA2-transfected cells display a higher number of cells entering S phase. Altogether, our results point at the important role of SPCA2 in regulation of cell cycle in cancer cells.
The life of multi-cellular organisms strongly relies on the dynamic regulation of various metal and non-metal ions, made possible through the function of numerous passive and active membrane transport systems [1–4]. Secretory pathway Ca2+/Mn2+-ATPases SPCA1 and SPCA2 are active ion transport systems located in membranes of the secretory pathway: Golgi sub-compartments, endosomes and some others [5–8]. Ubiquitous SPCA1 and tissue-specific SPCA2 isoforms have an affinity for Ca2+ comparable to that of sarco(endo)plasmic reticulum Ca2+-ATPase SERCA2 (one of the main Ca2+ transporting systems), however they can also transport Mn2+ with comparable affinity .
The regulation of intracellular Ca2+ in eukaryotic cells [5,10] is achieved by several main transporters (SERCA, Na+/Ca2+ exchanger NCX, and plasma membrane Ca2+-ATPase, PMCA) which keep its cytosolic levels at approximately 100 nM at rest. The role of another divalent cation, Mn2+, is less understood. This trace element is a cofactor of Mn2+-dependent enzymes/metalloenzymes such as oxidases, kinases, N-linked glycosylases, DNA/RNA-polymerases and the mitochondrial superoxide dismutase (MnSOD) [11,12]. All intracellular compartments have at least one enzyme that requires Mn2+ . Mn2+ itself displays antioxidant effects, is able to scavenge free radicals and regulate reactive oxygen species (ROS) generation under hypoxia . Hyperoxia and increased ROS generation up-regulate Mn2+-dependent enzymes, such as MnSOD . MnSOD is involved in destruction of superoxide radicals in human colon cancer cells, and its overexpression in HCT116 cells leads to senescence-associated growth arrest . From a clinical point of view, Mn2+ deficiency is rare, however its increased levels are neurotoxic and can lead to Parkinson's disease-like symptoms [16,17]. Still, the physiologically ‘normal’ Mn2+ concentrations in intracellular compartments are unknown and require investigation. Exposure of cells to 10–800 μM Mn2+ in the medium led to toxic effects, however the amounts of Mn2+ diffused through the cell membrane are not known. Oxidized Mn can also enter the cell via transferrin and become reduced . Recent reports also linked Mn2+ homoeostasis with the control and progression of the cell cycle: exposure to MnCl2 led to the arrest of A549 cells in G0/G1 phases . In addition, MnSOD can induce cell cycle arrest and display anti-proliferative function .
Collectively, the function of SPCA1 and SPCA2 can be viewed as Ca2+ deposition in ER, Golgi, secretory vesicles and recycling endosomes (or extrusion out of the cell) and Mn2+ detoxification, in tissue-specific manner. SPCA1−/− knockout mice die in utero before gestation day 10.5 and display numerous defects of Golgi structure, whereas heterozygous mice display pre-disposition to squamous cell tumours of epithelial cells from skin and esophagus . SPCA2, highly expressed in such ‘normal’ tissues as lungs, gastrointestinal tract and in mammary gland during lactation, as well as in some cancer cell lines [6,20–24], is expected to play a more specialized role.
It should be noted that most studies of SPCA1 and SPCA2 were performed with conventional (2D) cultures of normal and overexpressing cells , and the roles of these enzymes in regulation of Ca2+ and Mn2+ in cancer were relatively unstudied . Cells cultured under tissue culture conditions often experience oxidative stress, do not exhibit natural 3D architecture and are not exposed to physiological levels of O2 [26,27]. Oxidative stress can lead to increase in cytosolic Ca2+ , inducing cell apoptosis, increase in mitochondria-derived ROS/reactive nitrogen species (RNS) production and activation of enzymes including phospholipases and xanthine oxidases . Similarly, decreased O2 (‘hypoxia’ or ‘physiological normoxia’, broadly defined as range of 0–5% O2, depending on the tissue) leads to increased intracellular Ca2+ in pulmonary arterial myocytes  and endothelial cells [31,32]. Overall, disregulation of intracellular Ca2+ stores can induce apoptosis resistance, a hallmark of cancer  and increased metastatic potential [34,35]. Disregulation of Mn2+ transport can be also linked to oxidative stress and environmental hypoxia. Compartments of gastrointestinal tract, enriched in SPCA2, experience oxygen gradients and probably oxidative stress [20,21,36]. One approach for studying Ca2+/Mn2+ homoeostasis under hypoxia is to use 3D multicellular spheroids and confocal fluorescence imaging . Such tumour models grown in vitro allow establishment of in-depth gradients of nutrients and metabolites and facilitate the formation of heterogeneous cell layers. Thus, colorectal carcinoma spheroids have similar organization of epithelium at the periphery and mesenchymal-like cells in the centre . Large (>200–400 μm) spheroids also normally contain necrotic cores similar to tumours in vivo [37,39,40]. Spheroid cultures are gaining popularity in various areas of research  such as evaluation and development of anti-cancer drugs activity and drug-toxicity studies [42,43], the impact of combining different cell types on cellular interactions, gene expression, shape and architecture . This can be complemented by new imaging approaches such as confocal fluorescence lifetime imaging microscopy–phosphorescence lifetime imaging microscopy (FLIM–PLIM) allowing the analysis of composition, oxygenation, distribution of live/dead cells and other important ions in live spheroids [39,45,46].
In the present study, we analysed the effects of tumor micro-environment (low O2, cell density, 3D organization) on the expression of SPCA2 in HCT116 cells. We found that SPCA1 and 2 isoforms are affected by cell density and hypoxia (3% O2) in both adherent (2D) and spheroid (3D) cell cultures, and this contrasts with other Ca2+ ATPases such as PMCA and SERCA. We also provide evidence that SPCA2 is involved in maintaining Mn2+ in the Golgi in live cells. The up-regulation of SPCA2 under hypoxia is correlated with ROS generation, emphasizing its role in cancer cell survival. Increased SPCA2 in cells grown at high density and under hypoxia points to its role in cell cycle progression and tumour growth. Preliminary data from this work were reported in .
MATERIALS AND METHODS
MitoImage-NanO2 phosphorescent O2 probe was from Luxcel Biosciences , perylene bisimide-based pH probe NSP was prepared as described before . GPP130-GFP plasmid DNA was kindly provided by Prof. A. Linstedt and Dr S. Mukhopadhyay (Carnegie Mellon University). Plasmid DNAs encoding hSPCA2 and hSERCA2b and anti-SPCA2 polyclonal rabbit antibody were provided by Prof. F. Wuytack and S. Smaardijk (Laboratory of Cellular Transport Systems, KU Leuven). Plasmid DNAs for transfection were purified from Escherichia coli using Genopure Plasmid Midi Kit (Roche). Fluorescent NAD(P)H probe was synthesized as described previously . Rabbit polyclonal pan-SPCA antibody  was provided by Dr N. Pestov (Shemyakin-Ovchinnikov Institute of Bioorganic Chemistry). Mouse colon RNA samples isolated from normal and hypoxia-preconditioned mice  were provided by Dr I. Okkelman, Dr A. Zhdanov and Prof. K. O'Halloran (University College Cork).
Lipidure™ plates were from Amsbio, Perfecta 3D Hanging drop plates were from 3D Biomatrix. B27 serum-free supplement, CellEvent Caspase-3/7 substrate, LysoTracker Red, Alexa Fluor 488 and 594-conjugated secondary antibodies and tetramethylrhodamine methyl ester (TMRM) were from Invitrogen (Bio-Sciences). Epidermal growth factor (EGF), fibroblast growth factor (FGF) and monoclonal anti-BrdU ‘BU-1’ antibody were from Millipore. Monoclonal mouse anti-HIF1α antibody was from R&D Systems. TransIt-X2 reagent was from Mirus (MyBio) and ROSstar™ 550 dye was from Li-Cor (P/N 926-20000). PCR master mix, CellTiter-Glo, CellTox Green, Improm-II reverse transcriptase, ribonuclease inhibitor and ReliaPrep Total RNA extraction kit were from Promega (MyBio). Nucleo-spin II Total RNA extraction kit was from Macherey-Nagel (Fisher Scientific). ECL Prime chemiluminescent substrate was from GE Healthcare. Anti-α-tubulin antibody, horseradish peroxidase-conjugated anti-mouse and rabbit secondary antibodies, MnCl2, 5-bromo-2′-deoxyuridine (BrdU), oligonucleotides and all the other reagents were from Sigma–Aldrich (Dun Laoghaire). The sterile plastic ware (cell culture grade) was from Sarstedt (Wexford). μ-slide 8/12-well microscopy chambers were from Ibidi.
Adherent cell culture
HCT116 cells  were handled as described before . Typically, cells were grown in McCoy 5A medium supplemented with 10% FBS (heat-inactivated), 10 mM HEPES, pH 7.2, 2 mM glutamine, penicillin–streptomycin in Cell+ flasks. Control and hypoxia (continuously exposed to 3% O2 for 7 days)-treated cells were normally harvested at 50–70% of confluence unless indicated otherwise. DNA transfection of adherent cells was performed using TransIt-X2 reagent (Mirus), as per manufacturer's instructions. Briefly, cells seeded at 70–75% confluence were transfected with GPP130-GFP, hSPCA2 and hSERCA2b-encoding plasmid DNAs in Opti-MEM I medium for 24–36 h and then analysed. For the treatment with MnCl2, cells were first transfected for 18 h and then exposed to 0.5 mM MnCl2 for 24 h. For immunofluorescence, cells were pre-stained with BrdU (100 μM, 30 min).
RNA extraction and PCR analysis
2–4×106 cells were washed with PBS and total RNA was immediately isolated using Nucleo-spin II RNA extraction kit (Macherey-Nagel), according to manufacturer's instructions. For spheroid cultures, RNA was extracted from pooled spheroids (72 for 1000 or 24 for 3000 initially seeded cells) using the ‘ReliaPrep’ extraction kit (Promega, MyBio) according to manufacturer's instructions. One to two micrograms of total RNA annealed with 500 ng oligo-(dT)15 (70°C, 5 min, then 4°C, 5 min), reverse transcribed using ImProm-II Reverse transcriptase (42°C, 2 h), heat-inactivated (70°C, 15 min) and stored at −18°C. cDNA was used for semi-quantitative PCR analysis (25–35 cycles) or real-time PCR (45 cycles) using qPCR reagent (Maxima SYBR green, Fisher Scientific), and the AB7300 Real-Time PCR machine, analysed using the 7300 system SDS software from Applied Bio-systems (Life technologies). The sequences of oligonucleotides are presented in Supplementary Table S1.
Production and analysis of spheroids
Three different methods of spheroid production were used: (1) in ‘free floating’ (FF) method cells were seeded at concentration 1×106 cells/ml in Phenol Red-free DMEM, 10 mM glucose, 1 mM sodium pyruvate, 2 mM glutamine, 10 mM HEPES-Na, pH 7.2, 1% penicillin–streptomycin, 2% B27, 20 ng/ml EGF and 10 ng/ml FGF (low adhesion flasks) . Spheroids were formed after 3 days of culture . ‘Hanging drop’ method used seeding of cells at various concentrations in 40 μl of McCoys 5A medium supplemented with 10% FBS (heat-inactivated), 10 mM HEPES, pH 7.2, 2 mM glutamine, 1% penicillin–streptomycin per well, followed by addition of PBS to the 96-well plate to prevent excessive evaporation . Spheroids of approximately 500–650 μm size were formed after 3 days . Lipidure method used seeding of various concentrations of cells in 200 μl growth medium and 3 day incubation in 96-well plates.
Spheroid staining with NanO2 (5 μg/ml), pH probe (5 μg/ml), ROSstar 550 (5 μM) and CellTox Green (0.2%) was performed by addition of the probe/dye at the start of spheroid formation and incubation for 3 days. Staining with TMRM (20 nM, 30 min) was achieved with pre-formed spheroids. To label proliferating cells, BrdU (200 μM) was added to pre-formed spheroids 16–18 h before the fixation. For microscopy, spheroids were re-seeded and allowed to attach (3–8 h) on MatTek 35 mm glass-bottom dishes (MatTek) pre-coated with mixture of Collagen IV and poly-D-lysine .
Protein extraction and Western blotting
This was performed as described previously, with minor modifications . Briefly, 2–4×106 cells were washed with PBS and lysed in RIPA buffer (150 mM NaCl, 1% NP-40, 0.5% sodium deoxycholate, 0.1% SDS, 50 mM Tris, pH 8.0) supplemented with protease inhibitor cocktail (Sigma P2714) for 10–15 min on ice with gentle rocking. Extracts were cleared by centrifugation (15000 g, 15 min, 4°C) and total protein was quantified by the BCA protein assay kit (Pierce). Samples normalized for total protein were mixed with 5× Sample Laemmli Buffer, incubated for 10 min at room temperature (no boiling) and analysed by Western blotting.
Total cellular ATP was measured as before , using CellTiter-Glo reagent (Promega). Briefly, spheroids were allowed to attach onto the collagen-coated 96-well plates in Phenol Red-free DMEM (3 h), at concentration of one spheroid per well (∼6000 cells per spheroid) (Lipidure or Hanging Drop methods) or approximately 12–18 spheroids/well (FF method). After attachment, equal amount of CellTiter-Glo reagent was added, lysates were mixed and the luminescence was further analysed on time-resolved fluorescence plate reader Victor2 (PerkinElmer).
Confocal FLIM–PLIM microscopy
This was performed on an upright Axio Examiner Z1 microscope (Carl Zeiss), equipped with 20×/1.0 W-Plan-Apochromat and 63×/1.0 W-Apochromat objectives, temperature control (37°C) with motorized Z-axis control, DCS-120 confocal scanner (Becker & Hickl GmbH), R10467U-40 photon counting detector (Hamamatsu Photonics K.K.) and dedicated TCSPC hardware (Becker & Hickl GmbH) as described previously . The fluorescent probes were excited with tunable picosecond supercontinuum laser SC400-4 (Fianium) or BDL-SMC 405 nm pulsed diode laser (Becker & Hickl GmbH) for NanO2. O2 imaging with NanO2 probe was performed as described before using following calibration function: [O2, μM] =9274.667e(−τ/7.27143), where τ is in μs . Alexa Fluor 488, CellEvent Caspase 3/7 and CellTox Green dyes were excited at 488 nm with emission collected at 512–536 nm. TMRM was imaged using 540 nm excitation, 565–605 nm emission; Alexa Fluor 594 was excited at 594 nm, with emission collected at 635–675 nm. Transmission light (TL) images were collected with a D5100 digital SLR camera (Nikon) attached to the microscope.
Widefield fluorescence microscopy
The live cells, transfected with GPP130-GFP plasmid DNA  and counter-stained with LysoTracker Red (0.1 μM, 30 min), NAD(P)H probe (10 μM, 1 h) and ROSStar 550-stained as well as fixed and immunostained cells were analysed on a widefield fluorescence inverted microscope Axiovert 200 (Carl Zeiss) equipped with and oil-immersion objectives 40×/1.3 EC Plan Neofluar and 100×/1.4 Plan Apochromat, 470 and 590 nm LED excitation module, time-gated CCD camera and emission filters DAPI (Semrock 5060B-ZHE), FITC (472/30 nm excitation, 535/55 nm emission) and TXRed (Semrock 4040B-ZHE), ImSpector software (LaVision BioTec) and integrated CO2/O2 climate control chamber (PeCon) as described before .
This was performed as described before  using cells grown on 12-well μ-slide (Ibidi) and probed with rabbit anti-SPCA2 (XIB, provided by Prof. F. Wuytack), mouse anti-SERCA2 (Millipore) or BU-1 (Millipore) antibodies, followed up by staining with Alexa Fluor 488 or 594-labelled secondary antibodies (Invitrogen) and staining of nuclei by DAPI (300 nM, 5 min). Samples were analysed by widefield fluorescence microscopy.
Data analysis and statistics
Statistics were carried out using the results of at least three independent experiments. Western blotting, qPCR data and localization experiments were evaluated for statistical differences using t test with confidence levels of P<0.05, 0.005 or 0.001 (marked by an asterisk) accepted as significant. Electrophoretic band intensities were quantified in Fiji software (http://fiji.sc/Fiji) and then averaged between three independent experiments with standard deviations shown as error bars. qPCR experiments were normalized to control β-actin signals and compared between three independent experiments.
Fitting of phosphorescence decays was performed using SPCImage software (Becker & Hickl GmbH) using single-exponential decay fitting function and pixel binning as appropriate. 3D projections of spheroids were produced from intensity images representing individual optical sections using Volume Viewer plugin in Fiji software, cells stained with CellTox Green, BrdU and DAPI were counted using ‘3D objects counter’ plugin in Fiji software. Localization of GPP130 to the lysosomes was carried out using the co-localization threshold plugin in Fiji software. Data are presented as average values with standard deviation shown as error bars.
RESULTS AND DISCUSSION
SPCA1 and SPCA2 are the only Ca2+-transporting ATPases affected by hypoxia and cell density in adherent HCT116 cells
Previously, the increased expression of SPCA2 was demonstrated in a number of cell lines from colon and mammary tumours . We hypothesized that rapidly proliferating cancer cells can reside in deoxygenated (hypoxic) regions in vivo  thus affecting expression of SPCA2. To study the expression of SPCA and related Ca2+-ATPases, we chose two types of human colorectal carcinoma HCT116 cells: wild type (WT), with strong activity of oxidative phosphorylation (OXPHOS) and knockout HCT116 SCO2−/− cells, lacking synthesis of cytochrome c oxidase 2 enzyme, deficient in OXPHOS and experiencing strong oxidative stress and ROS production under normal growth conditions . Using previously described procedures [52,58], we cultured cells for 7 days at 3% O2 (‘hypoxia’), mimicking physiological oxygenation conditions. We chose 3% O2 as an averaged concentration within the range of physiological hypoxia experienced by the intestine in vivo . Colon tissues and tumours in vivo have highly heterogeneous oxygen concentrations ranging from approximately 0 to 5% [36,60]. Both wild-type and knockout cell lines retained similar viability under normoxia and hypoxia conditions. Using semi-quantitative reverse transcription polymerase chain reaction (RT-PCR), we analysed expression of various Ca2+-ATPase isoforms in these cells (Supplementary Figure S1): in contrast to PMCA and SERCA isoforms, we observed up-regulation of SPCA1 and SPCA2 expression in WT cells exposed to hypoxic conditions, which was next confirmed by real-time PCR (Figure 1). On the other hand, SCO2−/− cells showed decreased levels of SPCA2 under hypoxia (Figure 1). This is in agreement with the fact that at increased atmospheric O2 (21% O2) this cell type experiences strong oxidative stress and can use SPCA1/2 for scavenging of ROS.
Expression of genes encoding SPCA1 and SPCA2 in adherent HCT116 cells (wild-type, WT and SCO2−/−) grown under different oxygenation conditions and cell density
Next, we noticed that slight variations in cell density affected the expression of SPCA2: an almost 3-fold increase in SPCA2 was observed when the density changed from 50% to 100% in wild-type cells, but not in the case of SPCA1 isoform (Figure 1). As the density can affect peri- and intra-cellular O2 , we analysed O2 in cells grown at 50% and 100% confluence in similar types of plates; however, no significant difference between ambient and intracellular O2 was observed (Supplementary Figure S2). Thus, we identified cell density as a separate factor, affecting the SPCA2 but not SPCA1 isoform.
To confirm the up-regulation of SPCA2 at the protein level, we analysed cell extracts using available pan-SPCA rabbit antibody . The Western blotting confirmed 1.5–2-fold increase in SPCA in WT, and decrease in SCO2−/− cells, when normalized to α-tubulin levels. In contrast, SERCA2b protein remained unaffected by O2 (Supplementary Figure S3).
To further see if the hypoxia-driven up-regulation of SPCA2 can occur in vivo, we analysed colon tissue samples from mice exposed to continuous hypoxic conditions (10% O2) . Real-time PCR revealed a significant (2-fold) up-regulation of SPCA2, compared with the mice kept at normal atmospheric O2. In contrast with monolayer culture, SPCA1 in mouse colon tissue showed no significant alterations (not shown). Although the actual O2 in the mouse colon upon adaptation to hypoxia is unknown, this result is in agreement with our data obtained with cell cultures (Supplementary Figure S4).
Thus, we have identified SPCA2 as one of the Ca2+/Mn2+- ATPases affected by environmental O2, both in vitro and in vivo. Since only SPCA2 isoform showed increased expression depending on cell density and displayed stronger up-regulation by O2, we decided to focus on this isoform in further work.
SPCA2 expression in 3D spheroid model
To confirm and further study our finding of O2 and cell density effects on SPCA2 expression in a more relevant cell-based environment, we chose a 3D model of tumour spheroids. Since the variety of methods proposed results in spheroids having different size, metabolic activity and viability , we first optimized the methodology and compared three conventional methods of spheroid production: ‘FF’ culture, Lipidure™ and ‘Hanging drop’ (see ‘Materials and methods’). We aimed to achieve uniform size, high viability, physiologically relevant O2 and pH gradients, and ease of culturing. ‘FF’ method showed the greatest variation in size, whereas both Lipidure™ and ‘Hanging drop’ produced spheroids with controlled size and high reproducibility (Supplementary Figure S5). Next, we assessed cell death and viability using a panel of commercially available live cell imaging dyes. Staining with CellTox Green informing on membrane integrity, revealed differences in spheroids produced by the three methods, with the lowest number of necrotic cells for Lipidure™ plates (Supplementary Figures S5C and S5D). We thus chose Lipidure™ method for further work, considering the ease of use, high cell viability and ability to produce spheroids of controlled size and amount of cells (Supplementary Figure S5F): using DAPI staining we estimated the cell number as 2500–3500 per spheroid of 200–300 μm in diameter (Supplementary Figure S5E).
Next, using the fluorescence and phosphorescence lifetime imaging microscopies (FLIM–PLIM), we analysed O2 and pH gradients in spheroids from wild-type cells (Figure 2). Significant gradients were observed between the periphery and core, but not for the different depths and sizes of the spheroids. Higher fluorescence lifetimes of pH-sensitive nanoparticles correspond to acidic regions in the core (pH <6.5)  (Figures 2A and 2B). Staining with O2-sensitive phosphorescent NanO2 probe and PLIM method  revealed deeper hypoxia in spheroids with larger amounts of cells (Figures 2C and 2D): 75–55 μM at the periphery (corresponds to 7.5–5.5% O2) and 55–35 μM (5.5–3.5% O2) in the core for wild-type cell spheroids (Figure 2D). Spheroids from SCO2−/− cells, lacking functional OXPHOS, showed higher O2 concentrations (60–50 μM) and lower magnitude of gradients, despite the fact that their size was similar to the WT spheroids. Their shape however was irregular, lacking the spherical form observed when using WT cells (Supplementary Figure S6).
Characterization of wild-type HCT116 spheroids produced by Lipidure™ method
We reasoned that pronounced gradient of O2 and 3D architecture of the cells in HCT116 spheroids influence expression of SPCA2. To verify this, we prepared spheroids from 1000 and 3000 cells (displaying 35 and 20 μM O2 in the core, respectively), extracted total RNA and analysed by real-time PCR (Figures 2E and 2F). We found almost 2-fold up-regulation for SPCA2 and less profound up-regulation for SPCA1, compared with control cells, grown in 2D culture at 21% O2. Thus, deeper de-oxygenation and cell–cell interactions also positively affected expression of SPCA2 in the 3D spheroid model.
SPCA2 expression correlates with MIST1 and is dependent on HIF-1α
The regulation of SPCA2-encoding gene ATP2C2 was not studied in detail. We thus were interested in whether hypoxia affects the expression of transcription factor MIST1, which has previously been shown to regulate SPCA2 expression (alternatively spliced mRNA encoding ∼20 kDa C-terminal fragment) in salivary gland and pancreatic acinar cells [62,63]. MIST1 is a transcription factor of the class B family of basic helix–loop–helix (bHLH) proteins which regulate cell differentiation . MIST1 forms homodimer complexes to activate gene transcription and is localized in the nucleus . MIST1 expression was observed in acinar cells of salivary glands and secreting cells of the stomach, prostate and seminal vesicles, i.e. tissues overlapping with the expression pattern of SPCA2. All MIST1-positive cells have regulated exocytosis, which involves specific signalling pathways through which external cues induce secretion . We analysed MIST1 expression in HCT116 cells and found that it exhibits a similar pattern to SPCA1 and SPCA2, under hypoxia, normoxia and in both wild-type and SCO2−/− cells as well as in spheroids. Strikingly, MIST1 expression also positively correlated with cell density (Figure 3).
Correlation of gene expression patterns of MIST1 and SPCA2
The increase in SPCA2 expression under hypoxia meant that SPCA2 gene can also be regulated through the hypoxia-inducible factor (HIF) pathway. To verify this, we exposed the cells to well-known inhibitors of HIF-1α degradation, cobalt chloride and dimethyloxalylglycine (DMOG) , and analysed the expression of SPCA2 and MIST1 (Figure 3 and Supplementary Figure S7). Although both drugs can display non-specific effects, unrelated to HIF-1α stabilization, they both positively affected expression of SPCA2. Nearly 1.5–2-fold increase in expression of both MIST1 and SPCA2 genes (Figure 3) indicates that their expression most likely is regulated via the HIF-1α pathway and these genes represent its novel downstream targets.
SPCA2 is involved in regulation of the cytosolic Mn2+ but is regulated by ROS
The up-regulation of SPCA2 by hypoxia in wild-type cells pointed to its involvement in regulation of ROS and RNS, which can be directly linked to its Mn2+-transporting activity. Although Mn2+-dependent activity of SPCA2 was demonstrated before, we wished to confirm if it regulates cytosolic Mn2+ in cultured HCT116 cells. To do this, we used a previously described fluorescent biosensor construct based on GPP130 protein . Elevated cytosolic Mn2+ leads to translocation of GPP130-GFP to lysosomes and its subsequent degradation (Figure 4A). Therefore, lysosomal GFP-tagged GPP130 can be used to ‘sense’ the increased Mn2+ in cultured cells. We transfected cells with plasmid DNA expressing full-length SPCA2, with and without addition of external Mn2+, and quantified regions of interest for GPP130-GFP co-localized with a marker of lysosomes, LysoTracker Red. Although the addition of Mn2+ alone did not lead to statistically significant change in lysosomal translocation of GPP130, introduction of SPCA2 showed its decrease. Combination of SPCA2/Mn2+, again led to an increase in lysosomal translocation of GPP130 (Figure 4). In another experiment, we compared the effect of SPCA2 transfection to that of SERCA2b (control, Ca2+-ATPase not involved in Mn2+ transport): the SPCA2-overexpressing cells showed a significantly lower amount of lysosomal GPP130 than non-transfected and SERCA2-transfected cells (Supplementary Figure S7). Even being indirect, these data support the functional involvement of SPCA2 in Mn2+ transport from the cytosol to Golgi and other secretory vesicles and is in line with previous studies of ion-transporting properties of this enzyme [9,20].
Overexpression of SPCA2 affects lysosomal accumulation of Mn2+-sensitive GPP130-GFP protein
Mn2+ ions can act as an antioxidant  and we wished to test if SPCA2 overexpression in the hypoxic core of spheroids can be correlated with ROS levels. Thus, we measured ROS production in spheroid model using fluorescent probe ROSstar 550, which is specific to superoxide and hydroxyl radicals. Core regions of the wild-type spheroids showed lower amount of ROS than at the periphery. With larger size spheroids (3000 versus 1000 cells seeded) ROS production decreased even more (Figures 5A and 5B). In contrast, in SCO2−/− spheroids, which are expected to display lower SPCA2 expression, we observed more profound (∼3 times brighter) staining with ROSstar 550 (Figures 5A and 5B). Unlike the wild-type spheroids, there was no significant gradient of ROS across the spheroids. The decrease in ROS across spheroids positively correlates with the expression of SPCA2 and distribution of O2, across the SCO2−/− spheroids (uniform O2, 40–60 μM) (Figure 2 and Supplementary Figure S6). The higher ROS in SCO2−/− spheroids further supports our hypothesis as the OXPHOS-deficient cells must require more anti-oxidant Mn2+ or have more capability to regulate its concentration in the cytosol. In this case, SPCA2 can be down-regulated to prevent its removal.
SPCA2 expression correlates with production of ROS and RNS
We confirmed higher ROS levels in SCO2−/− cells, compared with wild-type cells. Cells displayed ROS staining higher than upon treatment with H2O2. In agreement with the effect of cell density on SPCA2 expression, cells grown at higher density displayed lower ROS (Figures 5C and 5D). Thus, SPCA2 can be involved in regulation of ROS or is co-regulated with them. To better understand the interdependence between ROS and SPCA2 expression, we tested, if ROS and RNS themselves affect the expression of SPCA2. We treated adherent cells with hydrogen peroxide (donor of ROS), 3,3-bis(aminoethyl)-1-hydroxy-2-oxo-1-triazene (DETA NONOate) (donor of RNS), antimycin A (inhibitor of complex III of respiratory chain, induces ROS) and ROS scavenger N-acetyl-cysteine (NAC) and analysed the expression of SPCA2 (Figure 5E). Rather unexpectedly, we found that both ROS and RNS positively induced expression of SPCA2. However, SCO2−/− cells showed much lower levels of its expression, whereas NAC treatment did not lead to down-regulation of SPCA2. To explain this, we also analysed redox status in our model using fluorescent probe specific to NAD(P)H (Supplementary Figure S10). We did not see significant changes of NAD(P)H in cells under above-mentioned treatments, especially in the case of NAC. Hence, we explained such slightly controversial effect of NAC by its low ROS-scavenging activity in our system or non-specific action. The differences between expression of SPCA2 in WT and SCO2−/− cells can be a result of large differences in morphology, size and cell growth rate of these cell lines and potentially incorrect normalization of PCR results by expression of β-actin. On the other hand, the positive effect of ROS and RNS on SPCA2 expression is consistent (observed with three different treatments) and provides means to better understand the mechanism of SPCA2 regulation. Since it is known that ROS affect HIF-1α stabilization  and as we have shown in Figure 3 and Supplementary Figure S7, it can be speculated that SPCA2 is regulated via the axis ‘mitochondrial dysfunction–oxidative stress–ROS/RNS–stabilization of HIF-1α–up-regulation of SPCA2’.
SPCA2 is involved in regulation of Mn2+-dependent cell cycle progression
The effect of cell density on the expression of SPCA2 (Figure 1) pointed to a possible link between the function of this pump and progression of cell cycle. To test this, we looked at the difference in proliferation in cell cultures exposed to hypoxia (3% O2). Staining with BrdU showed twice the amount of cells entering S phase compared with normoxia (Figures 6A and 6B).
SPCA2 restores cell cycle arrest induced by Mn2+
These data correlate with increase in SPCA2 in hypoxia and its involvement in Mn2+ transport (Figure 4 and Supplementary Figure S8). In addition, it was previously shown that Mn2+ can affect cell cycle progression via MnSOD  and arrest the cell cycle in G0/G1 and S phases . We confirmed this using BrdU staining (Figure 6C). Hence, we transfected cells with SPCA2 and SERCA2b (control), challenged them with high MnCl2 in the medium and compared the number of cells entering S phase. Similarly to the data of Sarsour et al. , treatment with 0.5 mM MnCl2 significantly decreased cell proliferation in non-transfected cells (Figure 6D), but increased in SPCA2-transfected cells (both Mn2+-treated and non-treated samples). In contrast, SERCA2b-expressing cells treated with high Mn2+ showed a similar decrease in proliferation as non-transfected cells (Figure 6). These data confirm that transport of cytosolic Mn2+ by SPCA2 prevents cell cycle arrest induced by this ion. The exact mechanisms of how SPCA2-mediated transport is involved in it are unclear. However this effect highlights an important role of SPCA2 in colon cancer cells in their adaption to hypoxia, preventing their cell death, increasing proliferation capacity and promoting tumour growth.
Here, we report that SPCA2 expression in HCT116 cells depends on both cell density and cell deoxygenation. Using wild-type and OXPHOS-deficient SCO2−/− cells, grown in 2D culture and 3D spheroid model, we correlated SPCA2 up-regulation with O2 gradients, expression of transcription factor MIST1, HIF-1α and production of reactive oxygen and nitrogen species. We provide evidence that SPCA2 contribution to Mn2+ transport from the cytosol is in agreement with the regulation of ROS. On the other hand, increased ROS can also affect SPCA2 expression, linking it with regulation via HIF-1α pathway. Changes in ROS production in cancer cells, caused by increased cell density and hypoxia can therefore (auto)regulate expression of SPCA2, which in turn is involved in potential removal of ROS. Another direct effect of SPCA2 on the physiology of HCT116 cells is the increase in their proliferation, possibly through the minimization of exposure to high cytosolic Mn2+. Thus, the function of Mn2+-transporting SPCA2, regulated by hypoxia and cell density, is linked to generation of ROS and regulation of cancer cell survival.
Ruslan Dmitriev, James Jenkins and Dmitri Papkovsky conceived the study and wrote the manuscript. James Jenkins and Ruslan Dmitriev designed and performed experiments, analysed the data and prepared the figures. All the authors reviewed the results and approved the final version of the manuscript.
We thank Dr I. Okkelman, Dr A. Zhdanov and L. Dunne (University College Cork) for technical help and useful discussions, Prof. K. O'Halloran (University College Cork) for sharing experimental samples, Prof. F. Wuytack and S. Smaardijk (KU Leuven), Dr N.B. Pestov (Shemyakin-Ovchinnikov Institute of Bioorganic Chemistry), Prof. A. Linstedt and Dr S. Mukhopadhyay (Carnegie Mellon University) for generous gifts of expression constructs and antibodies.
This work was supported by the Science Foundation Ireland [grant numbers 13/SIRG/2144 and 12/RC/2276].
- DETA NONOate
epidermal growth factor
fibroblast growth factor
fluorescence lifetime imaging microscopy
mitochondrial superoxide dismutase
phosphorescence lifetime imaging microscopy
plasma membrane Ca2+-ATPase
reactive nitrogen species
reactive oxygen species
sarco(endo)plasmic reticulum Ca2+-ATPase
secretory pathway Ca2+/Mn2+ -ATPase
tetramethylrhodamine methyl ester