High levels of Mn2+ inhibit secretory pathway Ca2+/Mn2+-ATPase (SPCA) activity and cause Golgi fragmentation in neurons and glia


  • M. Rosario Sepúlveda,

    1. Departamento de Bioquímica y Biología Molecular y Genética, Facultad de Ciencias, Universidad de Extremadura, Badajoz, Spain
    Search for more papers by this author
  • Frank Wuytack,

    1. Laboratory of Cellular Transport Systems, Department of Molecular Cell Biology, Faculty of Medicine, Katholieke Universiteit Leuven, Leuven, Belgium
    Search for more papers by this author
  • Ana M. Mata

    Corresponding author
    • Departamento de Bioquímica y Biología Molecular y Genética, Facultad de Ciencias, Universidad de Extremadura, Badajoz, Spain
    Search for more papers by this author

Address correspondence and reprint requests to Ana M. Mata, Departamento de Bioquímica y Biología Molecular y Genética, Facultad de Ciencias, Universidad de Extremadura, 06006 Badajoz, Spain. E-mail: anam@unex.es


Excess Mn2+ in humans causes a neurological disorder known as manganism, which shares symptoms with Parkinson's disease. However, the cellular mechanisms underlying Mn2+-neurotoxicity and the involvement of Mn2+-transporters in cellular homeostasis and repair are poorly understood and require further investigation. In this work, we have analyzed the effect of Mn2+ on neurons and glia from mice in primary cultures. Mn2+ overload compromised survival of both cell types, specifically affecting cellular integrity and Golgi organization, where the secretory pathway Ca2+/Mn2+-ATPase is localized. This ATP-driven Mn2+ transporter might take part in Mn2+ accumulation/detoxification at low loads of Mn2+, but its ATPase activity is inhibited at high concentration of Mn2+. Glial cells appear to be significantly more resistant to this toxicity than neurons and their presence in cocultures provided some protection to neurons against degeneration induced by Mn2+. Interestingly, the Mn2+ toxicity was partially reversed upon Mn2+ removal by wash out or by the addition of EDTA as a chelating agent, in particular in glial cells. These studies provide data on Mn2+ neurotoxicity and may contribute to explore new therapeutic approaches for reducing Mn2+ poisoning.

Abbreviations used



dimethyl sulphoxide


ethylenediamine-tetraacetic acid


endoplasmic reticulum


3-(4,5-dimethylthiazol-2-yl-2,5-diphenyltetrazolium bromide


phosphate-buffered saline


polyvinylidene difluoride


secretory pathway Ca2+-ATPase


Tris-buffered saline

Manganese, which is found in cells mostly as Mn2+, is an essential trace element for mammals which acts as a cofactor for a number of enzymes. It is crucial for brain development and metabolism, but in excess can be neurotoxic. Although alimentary exposure to high manganese is usually not the problem, parenteral uptake of this element represents the main route for manganese intoxication. Indeed, chronic Mn2+ inhalation by occupational or environmental exposition (Nelson et al. 1993; Sierra et al. 1995) or increase of Mn2+ concentration by long-lasting total parenteral nutrition (Ono et al. 1995) or in cirrhotic patients (Krieger et al. 1995; Tuschl et al. 2008) results in Mn2+ deposition in brain, that can be detected as increased signal on T1-weighted magnetic resonance imaging mainly in the globus pallidus (Silva et al. 2004; Massaad and Pautler 2011). In view of its abundance in the environment and since Mn2+ can passively enter the cells in an unspecific way via a multitude of other bivalent ion channels and carriers, cells seldom experience a shortage of Mn2+. The problem that cells face is rather to remove the ion from the cells. Mn2+ can enter the brain via the blood capillaries and/or the cerebrospinal fluid, and its accumulation produces a severe and debilitating neurological disorder known as manganism (Olanow 2004). A generalized bradykinesia and widespread rigidity, with among other symptoms basal ganglia disturbances, make manganism resemble Parkinson′s disease (Calne et al. 1994). Cellular mechanisms of Mn2+ toxicity in brain are poorly understood, but seem to involve damage in different cellular types. Besides neurons acting as major functional elements in the brain, also the more numerous glial cells, which provide support for the neurons, are vulnerable to toxicity. In this study, we have prepared mouse primary cultures of neurons and glia to address the cellular impact of high extracellular Mn2+ concentration. We examined cell morphology and focused in particular on the Golgi apparatus, an organelle particularly affected in neurodegenerative diseases (Gonatas et al. 2006). Furthermore, the Golgi houses the Ca2+/Mn2+-ATPases of the secretory pathway (SPCAs), a group of ion-motive ATPases with two isoforms, SPCA1 and -2, identified in higher vertebrates (reviewed in Vangheluwe et al. 2009). SPCA1 represents a house-keeping isoform, highly expressed in the nervous system (Wootton et al. 2004; Sepulveda et al. 2007, 2008), whereas SPCA2 shows a more restricted distribution (Vanoevelen et al. 2005; Xiang et al. 2005; Pestov et al. 2012). Particularly, SPCAs differ from the other ATPase homologues by their high-affinity Mn2+ transport capacity (Ton et al. 2002) and currently represent the only known way for cellular Mn2+ detoxification by translocating the ion into the secretory pathway for disposal (Mandal et al. 2000; Leitch et al. 2011). Therefore, we analyzed the impact of cellular Mn2+ overload on SPCA function. In addition, we have investigated the reversibility of the Mn2+-induced neurotoxic damage to assess the potential of possible therapeutic strategies.

Material and methods

Neural and glial primary cultures

Primary cultures were prepared as modifications of established procedures (Hernandez et al. 2004). Embryos from female Swiss mice at day 17 of pregnancy (for neural cultures) and 1-day-old post-natal pups (for glial cultures) were obtained from the University of Extremadura (UEx) animal house. Experiments were performed with approval of the UEx Ethics Committee. Briefly, specimens were killed, meninges-free cerebral cortex were dissected and tissues were collected in Hank's Balanced Salt Solution (HBSS; Gibco, Rockville, MD, USA), and digested with 0.25% trypsin in HBSS (Gibco) for 15 min at 37°C. After gentle homogenization and rinsing in HBSS, dissociated cells were counted and plated onto 24-well plates (60 000 cells/well) or glass coverslips (25 000 cells/coverslip), coated previously with 0.1 mg/mL poly-l-lysine, in neurobasal medium (Gibco) containing 10% (v/v) horse serum (HS; Gibco) or in Minimum Essential Medium (MEM; Gibco)-10% HS for neurons or glia, respectively. Cells were incubated at 37°C and 5% CO2–95% air. After 3 h, the medium was replaced with the corresponding fresh medium and incubated for 7 days in vitro before treatment. Half of the culture medium was replaced with fresh medium every 3–4 days. Glial cultures showed > 90% of astrocyte marker glial fibrillary acidic protein-positive cells (Figure S1). For the Mn2+ treatment, half of the corresponding medium was replaced with fresh medium containing enough MnCl2 to reach the desired final concentration. Then, cells were incubated at 37°C and 5% CO2–95% air for 6 h for further analysis.

Cell viability

A colorimetric mitochondrial dehydrogenase activity assay was used to score cell viability (Mosmann 1983). Briefly, cells were incubated with 150 μg/mL 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) in balanced-salt solution (BSS: 137 mM NaCl, 3.5 mM KCl, 0.4 mM KH2PO4, 0.33 mM Na2HPO4, 10 mM Glucose, 100 mM Tes/NaOH, pH 7.4) for 1 h at 37°C. Metabolically active cells reduced MTT to a formazan precipitate that was solubilised with DMSO and quantified spectrophotometrically at 490 nm, with a background subtraction at 650 nm.


Neurons or glial cells seeded onto cover slips were washed with phosphate-buffered saline (PBS) and fixed with 4% paraformaldehyde, 4% sucrose in PBS for 20 min. Then, cells were permeabilized with 0.2% Triton ×100 in PBS, and blocked for 1 h with 3% bovine serum albumin in PBS. Subcellular localization of proteins was done by incubation with the corresponding primary antibody diluted in the blocking solution for 2 h: polyclonal rabbit anti-SPCA1 (1 : 500; Van Baelen et al. 2001), anti-GM130 (1 : 250; BD Biosciences, Madrid, Spain), anti-glial fibrillary acidic protein (1 : 500; Sigma, Madrid, Spain), anti-β tubulin (1 : 1000; Sigma), anti-SERCA2b (1 : 100, (Wuytack et al. 1989). Fluorescence labeling was obtained using secondary antibodies Alexa594 goat anti-rabbit, Alexa488 goat anti-mouse (1 : 2000, Molecular Probes, Eugene, OR, USA), and counter-staining for visualization of nuclei was done using 3 μM 4′, 6-diamidino-2-phenylindole (DAPI; Sigma) as a DNA-specific dye. FluorSave (Calbiochem, San Diego, CA, USA)-mounted slides were analyzed using a Nikon E600 fluorescence microscope. Negative controls were performed for every set of experiments by omitting the primary antibodies from the procedure. For mitochondrial staining, live cells were incubated with MitoTracker probe (Molecular Probes) for 30 min at 37°C in culture medium under air with 5% CO2, and then fixed and mounted for fluorescence microscopy.

Determination of Reactive Oxygen Species production

The probe 2′,7′-dichlorodihydrofluorescein diacetate (H2DCFDA; Molecular Probes) was used as indicator for Reactive Oxygen Species (ROS) production. This probe is non-fluorescent until the acetate groups are removed by intracellular esterases and oxidation occurs within the cell. Briefly, cells were incubated with 10 μM H2DCFDA in BSS buffer for 30 min at 37°C under air supplemented with 5% CO2. Thereupon, cells were washed with BSS buffer, detached in 500 μL BSS buffer by pipetting and ROS production was measured as the change in fluorescence (excitation 498 nm and emission 598 nm) in a Varioskan Flash (Thermo Scientific, Madrid, Spain) fluorescence spectrophotometer.

Quantification of apoptotic cells

Cells were fixed and labeled with 3 μM DAPI for nuclear staining. Then cells were identified morphologically as non-apoptotic cells, characterized by a homogeneous distribution of DNA in a normal-sized nucleus, or as apoptotic cells showing condensed chromatin (Kerr et al. 1972; Schmelz et al. 2004).

RT-PCR assays

Total RNA was isolated from each culture by using Trizol reagent (Invitrogen, Madrid, Spain). Reverse transcription was performed using the Thermoscript RT-PCR system (Invitrogen) with 2.5 μg of each RNA. Multiplex PCR for SPCA1 and β actin was carried out with 2 μL of each cDNA in a final volume of 25 μL containing 1× PCR buffer, 0.8 mM dNTPs (Roche Applied Science, Madrid, Spain), 1.5 mM MgCl2, 0.5 μM of each primer (described in Sepulveda et al. 2008), and 1.25 U GoTaq polymerase (Promega, Madrid, Spain). The PCR reaction conditions were: denaturation at 95°C for 2 min, followed by 30 cycles (matching the linear range of amplification) of denaturation at 95°C for 1 min, annealing at 54°C for 1 min, and extension at 72°C for 1 min, followed by a final extension at 72°C for 5 min. PCR products were separated in ethidium-bromide-containing 2% agarose gel in TAE buffer, visualized using the Molecular Imager scanner (Bio-Rad Laboratories, Madrid, Spain) and quantified using the TINA software (Raytest, Straubenhardt, Germany).

Over-expression of SPCA1

COS cells, which are derived from kidney cells of the African green monkey, were obtained from the American Type Culture Collection (Rockville, MD, USA). Cells were seeded in 100-mm culture plates at a density of 2.5 × 106 cells per plate, and transfected with the full-length cDNA of human SPCA1a cloned in the pMT2 expression vector (Dode et al. 2005) using GenJuice transfection reagent (Novagen, Merck, Madrid, Spain). After incubation for 60 h at 37°C in the presence of 5% CO2, cells were harvested from plates for membrane vesicle preparation. SPCA1a was clearly expressed in COS cells at levels much higher than that of the endogenous SPCA1 in control cells transfected with empty vector (data not shown), as in (Dode et al. 2005).

Preparation of membrane vesicles

Cells were scraped and pelleted. Membrane vesicles (MV) were prepared following the protocol described by (Sepulveda et al. 2005). The protein content was evaluated by the Bradford method (Bradford 1976) using bovine serum albumin as a protein standard.

Western blotting

Protein electrophoresis was performed in 7.5% sodium dodecyl sulfate–polyacrylamide gel electrophoresis mini gels, and the separated proteins transferred onto a polyvinylidene difluoride (PVDF) membrane. After blocking with Tris-buffered saline containing 0.3% (w/w) Tween 20 (TBS-T) and 5% (w/v) non-fat dry milk for 1 h, the PVDF membrane was incubated at 25°C for 3 h with primary antibodies against SPCA1 (1 : 1000) and GAPDH (1 : 500; Santa Cruz Biotechnology, Santa Cruz, CA, USA) diluted in Tris-buffered saline (TBS)-T. The membrane was then incubated with the corresponding peroxidase-conjugated secondary antibodies (1 : 3000; Sigma) for 1 h at 25°C. The PVDF membrane was extensively washed with 1% milk in TBS-T between steps and the labeling was visualized by incubation with 4-chloro-1-metoxinaftol substrate.

SPCA activity

The enzymatic activity was measured in MV by using a coupled enzyme assay at 37°C in specific conditions to measure only the contribution of Ca2+-ATPase activity of SPCA as described in (Sepulveda et al. 2007). Briefly, MV were incubated for 4 min in a reaction mixture containing 50 mM Hepes/KOH pH 7.4, 100 mM KCl, 100 μM CaCl2, 100 μM 1,2-bis(2-aminophenoxy)ethane-N,N,N',N'-tetraacetic acid (BAPTA) to obtain 3.16 μM free Ca2+, 2 mM MgCl2, 5 mM NaN3, 0.22 mM NADH, 0.42 mM phosphoenolpyruvate, 10 IU of pyruvate kinase (Roche), 28 IU of lactate dehydrogenase (Roche), 0.01% saponin (w/w), 100 nM thapsigargin (Sigma), and 2 μM vanadate (Panreac, Barcelona, Spain) (1 mL final volume). The reaction was started with 1 mM ATP (final concentration) and the SPCA activity was calculated by subtracting the Mg2+-ATPase activity (after addition of 3 mM EGTA). The Mn2+-dependent ATPase activity of SPCA was also measured in the presence of Mn2+ and the absence of Ca2+ and Mg2+ (to eliminate the contribution of the Ca2+ and Mg2+-ATPase activities to the reaction). Under these conditions, Mn2+ not only binds at the transport sites and activates the ATPase, but it also complexes with ATP and substitutes for the role normally taken by Mg2+ in the catalytic process. Free Ca2+ and Mn2+ concentrations were calculated using the BAD4 software (Brooks and Storey 1992).

Data processing and statistical analysis

Data are represented as mean ± SE and significant differences determined by an unpaired Student t-test using the SigmaPlot v10 software (SPSS Inc, Chicago, IL, USA). A p ≤ 0.05 value was considered statistically significant.


Viability and cellular effects of high extracellular Mn2+ in neurons and glia

Mouse primary cultures of neurons and glia were prepared to explore the effects of Mn2+ toxicity on both cell types from the nervous system. Short-term (6 h) acute incubations with increasing concentrations of Mn2+ drastically reduced the survival of both cell types, as shown by the MTT assay (Fig. 1). Glial cells were found to be more resistant to Mn2+ toxicity than neurons, with IC50 values of 1.8 and 0.5 mM MnCl2, respectively.

Figure 1.

Effect of Mn2+ on survival of neurons and glia in primary culture. (a) Neurons or glia in primary cultures at 7 DIV were incubated with the indicated concentrations of MnCl2 for 6 h. The 3-(4,5-dimethylthiazol-2-yl)-2, 5-diphenyltetrazolium bromide(MTT) assay was used to score cellular viability and representative images with MTT crystals (in dark gray) are shown. (b) Percentage of cell survival in the presence of increasing concentrations of MnCl2. Data are mean ± SE of duplicate MTT experiments, in three different cultures.

Structural rearrangements at the cellular level elicited by Mn2+ exposure were visualized by β-tubulin immunostaining which labels the cytoskeletal scaffold (Fig. 2). It can be seen that in neurons, the neurites were severely disrupted in the presence of 0.1 mM MnCl2, whereas glial cells only presented structural changes at 10-fold higher Mn2+ levels. However, it is worthy to point out that neurons cocultured with glia were more resistant to Mn2+ toxicity and only showed clear signs of neurodegeneration at concentrations above 2 mM MnCl2.

Figure 2.

Effect of Mn2+ on the integrity of neurons and glia. Cells were incubated for 6 h without or with increasing MnCl2 concentrations (0.05, 0.1, 1, 2, and 5 mM). After fixation, immunofluorescence detection was performed with anti-β tubulin (green) to analyze cytoskeletal scaffold. DAPI staining was used to visualize nuclei (blue). Neurons (left panel) were very sensitive to Mn2+ treatment, and showed severely disrupted neural processes (arrow) already at low Mn2+ concentrations (0.1 mM). Glial cultures (right panel) showed few neurons (arrowheads) cocultured with glia cells. However, glia cells were more resistant than neurons (see the microtubule scaffold at 0.1 mM Mn2+ in neurons comparing to that in glia cells at the same Mn2+ level). Interestingly, neurons found in glial cultures were more resistant to Mn2+ toxicity than neurons cultured alone (see magnification of the boxed neurons at the bottom). Scale bar: 20 μm.

As Golgi fragmentation is a particular hallmark of neural degeneration (Gonatas et al. 2006), structural integrity of the Golgi apparatus was assessed by immunostaining with the Golgi marker GM130 (a matrix protein located in cis-Golgi (Lowe et al. 2000) and with the secretory pathway Ca2+-ATPase (SPCA1), an enzyme that can actively transport Mn2+, mainly located in the trans-Golgi region (Van Baelen et al. 2004). In control cultures, both neurons and glial cells exhibited a predominant juxtanuclear Golgi distribution organized in one or a few GM130- and SPCA1-positive ribbon complexes (Fig. 3 and Figure S2), in line with previous results in rat brain cell cultures (Murin et al. 2006). The merged images document a slightly shifted localization of both markers which might be ascribed to their respective cis- and trans-Golgi distribution. It should be noted that, relative to glia cells, neurons show more closely apposed Golgi stacks because they present a more reduced cytoplasmic volume. As a consequence of Mn2+ treatment, the Golgi apparatus appeared fragmented in both cell types (Fig. 3). Neurons showed a very important fragmentation into smaller Golgi-derived fragments, although no segregation of GM130 and SPCA1 to the same stacks was clearly detected. In glia, fragmentation produced smaller ministack structures dispersed over the entire cytoplasm, which mostly retained the appearance and compartmentalization of the original Golgi apparatus (Fig. 3). These stacked Golgi profiles are not so evident and lost integrity at high Mn2+ concentration, as shown by the GM130 and SPCA1 distribution. It is worthy to point out that the presence of healthy neurons in glia cultures may reveal a putative role of glia protecting neurons against the Mn2+-induced Golgi fragmentation, as the integrity of Golgi stacks in these neurons remains unaltered and clustered to the nucleus. This observation is in line with the above-mentioned resistance of neurons cocultured with glia cells to Mn2+ toxicity, monitored by staining the preservation of cytoskeletal components like assembly of β-tubulin filaments (Fig. 2, bottom panel).

Figure 3.

Effect of Mn2+ on the structure of the Golgi apparatus in cultured neurons and glia. Cells were incubated for 6 h with the indicated MnCl2 concentration (in mM). After fixation, immunofluorescence detection was performed with anti- Secretory Pathway Ca2+-ATPase (SPCA)1 (red) and the Golgi marker GM130 (green), to assess the effect of Mn2+ on Golgi structure. DAPI staining was used to visualize nuclei (blue). Neurons (upper half of the figure) are very sensitive to Mn2+, with Golgi complexes fragmented at low Mn2+ concentrations. Glia cells (lower half of the figure) were more resistant than neurons. Again neurons (arrowheads) in coculture with glia were much more resistant towards Mn2+-induced toxicity compared with neurons cultured alone. Scale bar: 20 μm.

The impact of Mn2+ overexposure on the structural integrity of the endoplasmic reticulum (ER) and mitochondria was also assessed (Fig. 4). An antibody against the sarco(endo)plasmic reticulum Ca2+-ATPase isoform (SERCA2b) (Fig. 4a), highly expressed in the ER, was used to label the ER. It can be seen that incubation of the cells with 1 mM MnCl2 caused severe ER structural rearrangements. In contrast, mitochondrial morphology, as analyzed by the cell-permeant MitoTracker probe, appeared not to be affected. As the MTT assay, which assesses the mitochondrial dehydrogenase activity, pointed to mitochondrial dysfunction, we looked for mitochondrial ROS production (Fig. 4b), another reporter for mitochondrial stress. Addition of 1 mM MnCl2 to the medium for 6 h was found to increase ROS levels in neurons and glia by a factor of 4 and 4.5, respectively.

Figure 4.

Analysis of endoplasmic reticular, mitochondrial, and nuclear morphology after Mn2+ treatment. (a) Neurons and glial cells were incubated for 6 h without (0) or with 1 mM MnCl2 (1) and stained with the anti-SERCA2b antibody, as endoplasmic reticulum marker, or the mitochondrial Mitotracker probe. DAPI staining was used to visualize nuclei. Glia showed disturbances in endoplasmic reticulum structure after Mn2+ treatment (large arrow) whereas mitochondrial morphology was apparently not altered. However, neurons showed a decrease in the number of mitochondria located in neurites (short arrows). Remaining mitochondria were located close to or in the soma (arrowheads). Scale bar: 25 μm. (b) Effect of 6-h Mn2+ treatment on mitochondrial ROS production. Data are mean ± SE (*p ≤ 0.0001). (c) Analysis of apoptotic nuclei. The upper panel shows representative images after DAPI staining showing healthy nuclei next to apoptotic nuclei with condensed chromatin (arrows). The lower panel shows a quantification of apoptotic cells in each type of culture after incubation without or with 1 mM MnCl2 for 6 or 24 h. Data are mean ± SE of four images per coverslip, obtained from eight independent cultures.

Nuclear morphology was analyzed by DAPI staining (Fig. 4c). This allowed identifying apoptotic cells by their more condensed chromatin. After 6 h of incubation with 1 mM MnCl2, neural cultures showed a significant rise (36.7 ± 7.9% vs. 5 ± 1.6% in controls) in the number of apoptotic cells, with a further increase to 73 ± 10.4% after 24 h. In contrast, in glia cultures, an increase of apoptotic nuclei (51.4 ± 6.5% vs. 6.5 ± 1.4% in controls) was only seen after 24 h.

Manganese treatment does not alter SPCA levels but affects its function

To find out whether the Mn2+-elicited relocalization of the Golgi-associated Mn2+-transporter SPCA in neurons and glia cells involved changes in its expression level and/or in its enzymatic activity, the following experiments were addressed. First, SPCA expression was determined in both types of culture at the transcript level by multiplex PCR (Fig. 5a) and at the protein level by western blotting (Fig. 5b). In control conditions, SPCA1 expression was found to be slightly higher in glia than in neurons. Treatment with 1 mM MnCl2 only produced a non-significant reduction of SPCA1 mRNA and protein expression levels in neurons with respect to control, whereas no changes were observed in glia.

Figure 5.

Expression of Secretory Pathway Ca2+-ATPase (SPCA) at mRNA (a) and at protein (b) levels and examination of SPCA activity (c–e) in cultured neurons and glia. (a) Multiplex PCR was performed on cDNA from each culture after 6-h treatment with 1 mM MnCl2, and PCR products of 225 bp and 130 bp were obtained for SPCA1 and β actin, respectively. Quantifications of SPCA1 bands relative to β actin are mean ± SE values obtained from three assays with different preparations. (b) Membrane vesicles from neurons or glial cells (10 μg) or COS cells over-expressing hSPCA1 (5 μg) were electrophoresed, blotted onto polyvinylidene difluoride membranes and incubated with the anti-SPCA1 antibody. The GAPDH labeling was used as loading control. Quantifications of SPCA1 immunoreactions relative to GAPDH are shown as mean ± SE values obtained from three experiments with different preparations. (c) SPCA activity measurements in membrane vesicles (10 μg) prepared from neurons and glial cells controls and pre-treated with 1 mM MnCl2 for 6 h. (d) SPCA activity from membrane vesicles of 3-month-old mouse brain (20 μg), of control cultures and cultures pre-treated with 1 mM MnCl2 (10 μg), and of COS cells over-expressing hSPCA1 (1 μg) were measured as indicated in the Methods in the presence of increasing concentrations of MnCl2. The 100% activity corresponds to 0.136 ± 0.01 (brain), 0.073 ± 0.008 (control neurons), 0.077 ± 0.006 (control glia), 0.043 ± 0.003 (Mn2+-pre-treated neurons), 0.033 ± 0.008 (Mn2+-pre-treated glia), and 0.358 ± 0.008 (hSPCA1) μmol.min−1.mg−1. (e) SPCA-dependent ATPase activity of brain membrane vesicles was measured in the absence of Ca2+ and Mg2+, but in the presence of increasing concentrations of MnCl2. Data are mean ± SE values of experiments per triplicate of two different preparations (*p ≤ 0.05).

Thereupon, functional assays were also performed by measuring Ca2+-dependent ATPase activity catalyzed by SPCA in membranes obtained from cultures pre-treated with 1 mM MnCl2 and from controls (Fig. 5c). The Ca2+-ATPase activity of contaminating SERCA was inhibited by adding 100 nM thapsigargin, whereas the PMCA activity was blocked by adding 2 μM vanadate to the assay medium (Sepulveda et al. 2007). It can be seen that SPCA activity decreased to around 50–60% of control values in both cell types upon Mn2+ treatment. This drop corresponds to an equally large decrease in the number of viable cells according to cell viability tests (Fig. 1).

The effect of Mn2+ on Ca2+-dependent SPCA activity was also analyzed by cumulative additions of Mn2+ to the assay medium (Fig. 5d). Membranes prepared from cultured neurons or glia, from whole brain or from COS cells over-expressing hSPCA1 all showed a Mn2+-dependent decrease of the SPCA activity, reaching a maximum at 0.45 mM free Mn2+, with an IC50 value of 6 μM. This Mn2+ effect might be because of competition of Ca2+ and Mn2+ for the same ion-transport sites in SPCA, but allosteric binding to other sites cannot be excluded. These effects were also assayed for membranes from Mn2+-pre-treated cells, for which a similar activity decrease was obtained. This result indicates that those cells (about 50% of the cells) that survive the Mn2+ pre-treatment are equally susceptible to Mn2+ inhibition than those not pre-exposed to Mn2+.

In addition, the Mn2+-dependent ATPase activity was measured in the absence of both Ca2+ and Mg2+ in the same membrane preparations, to assess the performance of SPCA activity as a Mn2+-extrusion pump. Representative results with MV from brain are shown in Fig. 5e. As can be seen, the Mn2+-induced enzyme kinetics displayed the typical bell-shaped Mn2+-dependence of the ATPase activity with stimulation of ATPase activity below 0.01 mM free Mn2+, attributed to Mn2+ binding to ATP and to the ion-transport sites of SPCA, followed by activity inhibition at higher values. Addition of 3.16 μM free Ca2+ under conditions where Mn2+ stimulation was maximal (0.025 mM free Mn2+ see Table 1), resulted in a significant decrease in ATPase activity (~45–55%). A similar fractional inhibition of ATPase activity was obtained in membranes prepared from cultured neurons, from glia or from COS cells over-expressing hSPCA1, as already documented above in assays initially performed in the presence of Ca2+ and with subsequent Mn2+ additions (Fig. 5d).

Table 1. Effect of Mn2+ and Ca2+ on the SPCA-dependent ATPase activity
 ATPase activity in the presence of 3.16 μM free Ca2+ ATPase activity in the presence of 25 μM free Mn2+
  1. Two SPCA-dependent ATPase activities (μmol. min−1 mg−1) were assayed in MV from mouse brain (20 μg), 7 DIV-cultured neurons and glia (10 μg) and also in over-expressed hSPCA1 (1 μg). The first ATPase activity was assayed in the presence of 3.16 μM free Ca2+ before or after addition of 25 μM free Mn2+. The second ATPase activity was measured in the presence of 25 μM free Mn2+ before or after addition of 3.16 μM free Ca2+ to the assay medium. Data are mean ± SE of three experiments performed per triplicate.

  −Mn2+ +25 μM free Mn2+ −Ca2+ +3.16 μM free Ca2+
Brain0.136 ± 0.0100.030 ± 0.0050.406 ± 0.0160.180 ± 0.005
Neurons 7 DIV 0.085 ± 0.0040.021 ± 0.0010.420 ± 0.0200.193 ± 0.020
Glia 7 DIV 0.093 ± 0.0050.032 ± 0.0010.410 ± 0.0200.242 ± 0.010
hSPCA10.358 ± 0.0080.076 ± 0.0020.549 ± 0.0390.252 ± 0.001

Partial cell recovery occurs after Mn2+ withdrawal

The recovery from a Mn2+ insult was investigated by removal of the added Mn2+ from in vitro cultures. Cells were thereto first exposed to 1 mM MnCl2 for 6 h, and then Mn2+ was cleared either by changing the whole culture medium to medium without Mn2+ harvested from parallel untreated cultures, or by adding 1.2 mM EDTA (to chelate all free Mn2+). The use of either washout or chelation treatments after 6-h incubation with Mn2+ (described in Fig. 6a) quickly halted Mn2+-induced cytotoxicity as shown by MTT analysis (Fig. 6b). Besides, immunofluorescence assays showed that both Mn2+-removal methods prompted a partial recovery of cellular structures after 24 h, with a concurrent reassembly of Golgi complexes and reestablishment of ER integrity (Fig. 6c). This result was more pronounced in glia than in neurons. Note that cultured neurons are very sensitive to full replacement of culture medium and that the addition of EDTA to control cultures (which did not receive Mn2+) would drastically affect survival of both neurons and glia because the chelator will deplete important minerals and ions. Thus, EDTA chelation therapy can only be applied after Mn2+ overexposure.

Figure 6.

Cell recovery from Mn2+ toxicity. (a) The scheme represents the series of experimental steps performed in neurons and glia cultures to induce and remove Mn2+ toxicity. Briefly, in a first series of experiments, cells were incubated in the absence (Ct) or presence of 1 mM MnCl2 for 6 h (Mn) or 30 h (Mn30). In another series of cultures, the Mn2+-free or -containing medium was changed after 6 h to medium from parallel control cultures pre-grown in the absence of MnCl2 (Ct+W and Mn+W, respectively; Mn2+ wash out) or alternatively 1.2 mM EDTA was added (Ct+E and Mn+E, respectively; EDTA chelation). Then, cells were further incubated for 24 h. (b) Cell survival was scored after each treatment. Data are mean ± SE (comparison among groups are indicated by segments: *p ≤ 0.05; **p ≤ 0.0001). (c) Labeling of β tubulin (green), Secretory Pathway Ca2+-ATPase (SPCA)1 (red), and nuclei with DAPI (blue), and of endoplasmic reticulum with anti-SERCA2b (red) after each treatment. Scale bar: 20 μm.

The effect of Mn2+ withdrawal on neurons and glia cells was further tested in terms of SPCA activity. As shown in Fig. 7, the SPCA activity decreased in membranes from both cell types after 6 h of Mn2+ treatment and this drop became more pronounced after 30 h. Subsequent washing or EDTA treatments to cell cultures previously incubated with Mn2+ for 6 h did not arrest the decrease in SPCA activity on neurons. However, both treatments blocked the toxic effect of Mn2+ in glia. These results are in agreement with the partial cellular recovery observed in these cells.

Figure 7.

Effect of Mn2+ withdrawal on Secretory Pathway Ca2+-ATPase (SPCA) activity of neurons and glia. Ca2+-dependent SPCA activity was measured in membrane vesicles (10 μg) prepared from control cultures (Ct) of neurons (upper panel) or glia (lower panel), and from cultures incubated with 1 mM MnCl2 for 6 h (Mn) or 30 h (Mn30). It is also shown the SPCA activity in membranes prepared from cultures pre-treated with 1 mM MnCl2 for 6 h followed by Mn2+ wash out (Mn+W) or Mn2+ chelation by EDTA (Mn+E) (see scheme in Fig. 6a). Data are mean ± SE values of two experiments in triplicate (*p ≤ 0.05).


This work shows the toxic effects of high extracellular Mn2+ concentration on neurons and glia in culture and illustrates its inhibitory effect on SPCA activity. Glial cells seem to be more resistant to Mn2+ poisoning than neurons, and even play a protective role against neurodegeneration. This protection may be related to the ability of astrocytes to accumulate these divalent metals and their power to detoxify free radicals (Tholey et al. 1987; Akai et al. 1990; Aschner et al. 1992; Mena et al. 2002). Besides, glial cells contain a large number of Mn2+-dependent enzymes involved in countering oxidative stress and glutamate excitotoxicity (Wedler and Denman 1984; Akai et al. 1990). Note that an abnormal increase in the number of astrocytes has often been observed in neurodegenerative areas, for which it might be considered as a histopathological marker (Landis 1994; Olanow et al. 1996; Verkhratsky et al. 1998; Hazell 2002). Thus, this study forwards new evidence supporting the importance of glia in promoting neuronal viability under pathological conditions.

The Mn2+-induced fragmentation of the cellular Golgi apparatus might indicate a specific role of this compartment in guarding Mn2+ homeostasis. Similar Mn2+-induced fragmentations have been observed in non-neural cell lines (Towler et al. 2000) and in neuroblastoma N2a cells (own unpublished results). Golgi fragmentations also accompany rat nerve injuries and are observed in patients with amyotrophic lateral sclerosis (Fujita et al. 2011). Increasing evidence supports the notion that Golgi fragmentation occurs as an early lesion in neurodegeneration, before other symptoms appear and that it is not a consequence, but rather a potential trigger of cell apoptosis (Mourelatos et al. 1996; Gonatas et al. 2006). In fact, in cells undergoing apoptotic death Golgi stacks are replaced by clusters of vesicles, as was also observed in this work. In view of the higher Mn2+ tolerance of glia versus neurons in culture, the fact that glia cells also showed a similar Golgi fragmentation into numerous small disconnected elements upon Mn2+ treatment deserves to be further explored.

Given the role of the Golgi apparatus in Mn2+ homeostasis, it is not surprising that its SPCA activity was strongly modulated by external Mn2+ exposure. This suggests its participation in the management of Mn2+-induced neurotoxicity. This is also supported by in vivo studies reporting that brain areas with high SPCA1 expression also show enhanced Mn2+ accumulation upon continuous systemic MnCl2 infusion in mice (Sepulveda et al. 2012) and by the observation that a gain-of-function mutation of SPCA1, which specifically enhances Golgi Mn2+ transport, improves survival of Mn2+-exposed cells (Mukhopadhyay and Linstedt 2011).

The acute inhibitory effects of Mn2+ on the Ca2+-ATPase activity of SPCA and the similar inhibitory action of Ca2+ on the Mn2+-ATPase activity of SPCA (Table 1) suggest a competition of both ions for SPCA binding. In SPCAs, the transport of Mn2+ and Ca2+ is mutually exclusive, suggesting that each ion can occupy the same ion-transport site (Wei et al. 2000; Van Baelen et al. 2001). Thus, Mn2+ toxicity may affect Ca2+ homeostasis and thereby cell signaling and other Golgi functions in which Ca2+ and Mn2+ are critical. Pivotal processes in this respect include protein glycosylation or sulfatation, proteolytic processing of precursor polypeptides and proteins or membrane trafficking along the secretory pathway (reviewed in Vangheluwe et al. 2009). An effect on Ca2+ homeostasis may be also responsible for the ER disturbances observed in this work. Ca2+ dyshomeostasis would affect normal ER functions, e.g. folding of secretory and membrane proteins. The key role of SPCA1 activity in all those processes is supported by the fact that Spca1−/− mouse embryonic cells show altered trafficking of vesicles between the ER and the Golgi, presenting more dense vesicles on the cis-Golgi (Okunade et al. 2007). Besides, knockdown of SPCA1 by RNA interference in cultured neurons causes disturbed cellular Ca2+ homeostasis and altered targeting of organellar proteins affecting neural polarity (Sepulveda et al. 2009). Although the importance of SPCA1 function in central nervous system is nowadays well recognized, only few studies regarding its direct involvement in other neuropathologies have been reported, among them a response of SPCA1 gene expression to brain ischemic insults and the effect of ischemic pre-conditioning in rat brain (Pavlikova et al. 2009, 2011).

On the other hand, failure of efficient Mn2+ detoxification by saturating the SPCA-mediated removal via the Golgi could entail enhanced Mn2+ accumulation into the mitochondria, thereby causing the mitochondrial impairment observed by the MTT assay and the stimulation of ROS production. Actually, ROS generation has been related to an increase of Mn2+-induced oxidative stress in mitochondria (Gunter et al. 2006). In fact, Mn2+ at toxic levels could affect Mn2+-superoxide dismutase, i.e. an enzyme essential in detoxification of superoxide free radicals (Akai et al. 1990), and/or interfere with energy production in mitochondria, contributing to cytotoxicity. Thus, Mn2+ overexposure might produce a cellular stress that can eventually trigger cell death if cellular dysfunction is severe or prolonged.

Withdrawal from Mn2+ exposure or its chelation with EDTA in primary cultures stopped the progression of Mn2+-induced cellular degeneration and prompted partial recovery of Golgi and ER integrities as well as SPCA activity, particularly in glia. Similar treatments are currently used for patients with manganism, but are still moderately effective (Calne et al. 1994; Ono et al. 2002). Further studies at the molecular level need to be performed to link progression of neuropathologies to chronic Mn2+ exposure in vivo.


We thank D. Marcos for initial support with primary cultures preparation, Dr E. Guzman (Dept. Bioquímica y Biología Molecular y Genética, UEx) for letting us use the fluorescence microscope, and Dr M Lopez-Fanarraga (Dept. Biología Molecular, Facultad de Medicina, IFIMAV-Universidad de Cantabria,) for the gift of MitoTracker probe. M.R.S. received a Postdoctoral Fellowship from Programa de Incorporación de Doctores, Junta de Extremadura (Spain). This work was supported by grants BFU2011-23313 from MICINN, Fundación Marcelino Botín, Junta de Extremadura, and FEDER. Authors certify that there are no conflicts of interest.