The endoplasmic reticulum Ca2+-ATPase SERCA2b is upregulated in activated microglia and its inhibition causes opposite effects on migration and phagocytosis

1 The endoplasmic reticulum Ca2+-ATPase SERCA2b is upregulated in activated microglia and its inhibition causes opposite effects on migration and phagocytosis Running title: SERCA2b in microglial functions Juan M. Morales-Ropero1, Sandra Arroyo-Urea1, Veronika E. Neubrand1, David Martín-Oliva1, José L. Marín-Teva1, Miguel A. Cuadros1, Peter Vangheluwe2, Julio Navascués1, Ana M. Mata3, M. Rosario Sepúlveda1#. 1 Department of Cell Biology, Faculty of Sciences, University of Granada, Granada, Spain 2 Laboratory of Cellular Transport Systems, Department of Cellular and Molecular Medicine, KU Leuven, Leuven, Belgium 3 Department of Biochemistry and Molecular Biology and Genetics, Faculty of Sciences, University of Extremadura, Badajoz, Spain # Corresponding author: M. Rosario Sepúlveda. Department of Cell Biology, Faculty of Sciences, University of Granada. Avda. Fuentenueva s/n, 18151, Granada, Spain. Tel. +34 958 246334, [email protected] 2 Abstract Activation of microglia is an early immune response to damage in the brain. Although a key role for Ca2+ as trigger of microglial activation has been considered, little is known about the molecular scenario for regulating Ca2+ homeostasis in these cells. Taking into account the importance of the endoplasmic reticulum as a cellular Ca2+ store, the sarco(endo)plasmic reticulum Ca2+-ATPase (SERCA2b) is an interesting target to modulate intracellular Ca2+ dynamics. We found upregulation of SERCA2b in activated microglia of human brain with Alzheimer´s disease and we further studied the participation of SERCA2b in microglial functions by using the BV2 murine microglial cell line and primary microglia isolated from mouse brain. To trigger microglia activation, we used the bacterial lipopolysaccharide (LPS), which is known to induce an increase of cytosolic Ca2+. Our results showed an upregulated expression of SERCA2b in LPS-induced activated microglia likely associated to an attempt to restore the increased cytosolic Ca2+ concentration. We analyzed SERCA2b contribution in microglial migration by using the specific SERCA inhibitor thapsigargin in scratch assays. Microglial migration was strongly stimulated with thapsigargin, even more than with LPS-induction, but delayed in time. However, phagocytic capacity of microglia was blocked in the presence of the SERCA inhibitor, indicating the importance of a tight control of cytosolic Ca2+ in these processes. All together, these results provide for the first time compelling evidence for SERCA2b as a major player regulating microglial functions, affecting migration and phagocytosis in an opposite manner. Keywords: Calcium pump, endoplasmic reticulum, microglia, brain, migration, phagocytosis, Alzheimer´s disease. 3 Main points: • SERCA2b is upregulated in activated microglia in vitro and in Alzheimer´s disease-affected brains. • SERCA2b inhibition stimulates migration but prevents phagocytosis in microglia. Table of Contents Image: 4 1. INTRODUCTION Calcium (Ca2+) signaling is crucial in the physiology of cells regulating functions as development, muscle contraction, neurotransmission, secretion and even cell death (Berridge, Lipp, & Bootman, 2000; Brini, Cali, Ottolini, & Carafoli, 2014). In fact, cellular processes can be specifically switched on or off depending on cytosolic Ca2+ concentration (Bootman & Bultynck, 2020). Different cell types use a plethora of molecules involved in Ca2+ signaling in order to control and tightly modulate different cellular functions. In fact, Ca2+ dysregulation is involved in the etiology of many disorders, especially in the central nervous system where it has been associated with neurodegenerative diseases, including Alzheimer´s disease (AD), Parkinson´s disease, bipolar disorders, schizophrenia and autism (Bezprozvanny & Mattson, 2008). Among all cell types in adult brain, microglial cells constitute the first and main line of defense against injury. This requires a cellular activation that implies morphological transformation with changes in gene expression, release of factors into the environment, proliferation, migration and phagocytic activity (Kettenmann, Hanisch, Noda, & Verkhratsky, 2011; Wolf, Boddeke, & Kettenmann, 2017). Although many triggers of microglial activation are known, the underlying subcellular mechanisms are not clear yet. Their description is crucial to understand the presumed dual role found for activated microglia in different neurodegenerative pathologies. Microglia can play a neuroprotective or beneficial role on one hand, but on the other hand they can also contribute to neuroinflammation with detrimental effects (Arcuri, Mecca, Bianchi, Giambanco, & Donato, 2017; Du, Wang, & Geng, 2018; Martin-Estebane & GomezNicola, 2020; Tang & Le, 2016). 5 Emerging evidence indicates that intracellular Ca2+ signaling could play a critical role in the activation of microglia (Brawek & Garaschuk, 2013; Hoffmann, Kann, Ohlemeyer, Hanisch, & Kettenmann, 2003). In vivo studies have shown that the variations of cytosolic Ca2+ concentration ([Ca2+]c) are minimal in resting microglia, whereas great fluctuations occur during microglial activation (Eichhoff, Brawek, & Garaschuk, 2011; Olmedillas Del Moral, Asavapanumas, Uzcategui, & Garaschuk, 2019; Reddish, Miller, Gorkhali, & Yang, 2017; Tvrdik & Kalani, 2017). Thus, an increase of [Ca2+]c occurs in the presence of microglial activators such as the bacterial-endotoxin lipopolysaccharide (LPS) (abd-el-Basset & Fedoroff, 1995), extracellular nucleotides released during brain injury and involving P2X/P2Y receptors (James & Butt, 2002; Koizumi et al., 2007; Langfelder, Okonji, Deca, Wei, & Glitsch, 2015; Ohsawa et al., 2007) or even the amyloid β-peptide (Aβ), which is neurotoxic in AD (Alawieyah Syed Mortadza, Sim, Neubrand, & Jiang, 2018; Sarlus & Heneka, 2017), among many others. This triggers the signaling cascades that induce a functional response in activated microglia (Hoffmann et al., 2003). However, while Ca2+ influx in the cell is well described, the mechanisms involved in Ca2+ clearance remain poorly understood. In this respect, the endoplasmic reticulum (ER) is the major Ca2+ store in the cell and can participate as fine modulator of Ca2+ dynamics. ER membranes contain the sarco(endo)plasmic reticulum Ca2+-ATPase (SERCA), a P-type ion-motive ATPase that reduces [Ca2+]c by pumping 2 Ca2+ ions into the ER lumen at the expense of hydrolysis of 1 molecule of ATP (reviewed in (J. Chen, Sitsel, Benoy, Sepulveda, & Vangheluwe, 2020). Three SERCA genes have been identified in mammals, each with several splice variants. Of these, SERCA2b is the isoform with the highest expression in brain (Baba-Aissa et al., 1996; Baba-Aissa, Raeymaekers, Wuytack, Dode, & Casteels, 1998; Mata & Sepulveda, 2005; Miller, Verma, Snyder, & Ross, 1991). Although the importance of 6 the ER in microglial Ca2+ signaling has been recognized (Brawek & Garaschuk, 2013), the functional relevance of SERCA for microglia modulation remains unknown. Here, we aimed to explore the expression of SERCA2b in activated microglia and its participation in essential microglial functions, such as migration and phagocytosis. 2. MATERIAL AND METHODS 2.1 Immunohistochemistry in control and AD-human brain samples Paraffin-embedded tissue of medium frontal gyrus from human cerebral cortex of agematched controls (Braak stage II) and AD-diagnosed patients (Braak stages V-VI) were obtained from The Netherlands Brain Bank (NBB). Procedures, information, and consent forms of the NBB have been approved by the Medical Ethics Committee of the Vrije Universiteit Amsterdam Medical Centre on April 30, 2009. Microtome sections of 8 μm thickness were permeabilized by immersion in 0.05% (v/v) Triton X-100 in phosphate-buffered saline (PBS-T) for 15 min and endogenous peroxidase activity was quenched with PBS-0.5% (v/v) H2O2 for 45 min. After blocking with 0.2% (w/v) gelatin, 0.25% (v/v) Triton X-100 in PBS (PBS-G-T) and 0.1 M lysine for 1 h, sections were incubated overnight at room temperature in a humidified chamber with the following primary antibodies: the monoclonal anti-SERCA2 (clone IID8, Sigma, diluted 1:500 in PBS-G-T), the polyclonal microglial marker Iba1 (1:500, Wako) or the monoclonal anti-Aβ (1:500, STAB VIDA) diluted in PBS-G-T. Sections were subsequently washed in PBS-T and incubated with biotinylated goat anti-mouse or - rabbit antibodies (1:200, Sigma), respectively, and then with ExtrAvidin-peroxidase (1:200, Sigma). The immunodetection was carried out using 0.03% (w/v) 3,3- 7 diaminobenzidine tetrahydrochloride and 0.2% H2O2. Sections were dehydrated and mounted with Eukitt (Panreac) for observation under the microscope. Alternatively, double immunofluorescence was performed with both primary antibodies and the corresponding Alexa594 and Alexa488 secondary antibodies (1:2000, Molecular Probes). For visualization of nuclei, 3 µM 4´,6-diamidino-2-phenylindole (DAPI, Sigma) as a DNA-specific dye was used. Slides were mounted in FluorSave (Calbiochem) and analyzed with a Zeiss Axiophot fluorescence microscope with Axiocam 506 color camera and the ZEN 2.3 Lite Software. Negative controls were performed for every set of experiments by omitting the primary antibodies from the procedure. 2.2 BV2-cell line and microglial primary cultures BV2 cells were cultured in RPMI medium containing 10% (v/v) fetal bovine serum (GIBCO), supplemented with 2 mM L-glutamine, 100 U/ml penicillin, 100 µg/ml streptomycin and incubated at 37°C in a humidified atmosphere of 5% CO2. Primary cultures of mouse brain microglia were prepared as modifications of established procedures (Bolos et al., 2016; Neubrand et al., 2014). Briefly, newborn (1day old) C57BL/6 mice were obtained from the animal facility service of the Scientific Instrument Centre of the University of Granada (UGR), and experiments were performed with approval of the UGR Ethical Committee. Briefly, meninges-free cerebral cortex were dissected and collected in DMEM with 4.5 g/l D-Glucose, 4 mM glutamine, 10% (v/v) fetal bovine serum), 10% (v/v) horse serum and 100 U/ml penicillin, 100 µg/ml streptomycin (all reagents from GIBCO). After disaggregation and homogenization, cells were seeded and incubated at 37°C with 5% CO2 for 10-12 days. Then, cultures were softly shaken at 37ºC for 2 h and the primary microglia-enriched 8 supernatant was subcultured in the same medium for 2 days before experiments. Cultures of microglia showed >90% of microglial marker Iba1-positive cells by immunocytochemistry. Cell activation was induced by 100 ng/ml bacterial lipopolysaccharide (LPS, serotype 0111:B4, Sigma), according to (Henn et al., 2009). 2.3 Immunocytochemistry BV2 cells or primary microglia (25 000 cells) were seeded onto 12 mm (diameter) round glass coverslips coated with 0.1 mg/ml poly-D-lysine. After treatment, cells were fixed with 4% paraformaldehyde in PBS for 20 min. Then, cells were permeabilized with 0.2% (v/v) Triton X-100 in PBS, and blocked with 3% bovine serum albumin (BSA) in PBS for 1 h. Subcellular co-localization of proteins was determined by incubation with the polyclonal rabbit anti-SERCA2b (1:500, (Wuytack, Eggermont, Raeymaekers, Plessers, & Casteels, 1989), or the monoclonal anti-β tubulin (1:1000, Sigma) primary antibodies diluted in the blocking solution for 1 h. Fluorescence labelling was obtained by the corresponding secondary antibodies Alexa594 goat antirabbit and Alexa488 goat anti-mouse (1:2000, Molecular Probes), and DAPI for nuclei staining. Images were taken with the Zeiss Axiophot fluorescent microscope. 2.4 Cell analysis Images were analyzed by Image J software (version 1.50i, NIH). The cell morphology was characterized by analysis of form factor and aspect ratio of individual cells, according to (Alawieyah Syed Mortadza et al., 2018). The form factor values range between 0 and 1, where values closer to 0 indicate elongated cell shapes and values closer to 1 correspond to a circle. Values of aspect ratio start at 1.0, which indicates a circle, while ascending values indicate enhanced cell ramification and elongation. 9 Quantification of immunostaining intensities were determined in terms of Mean Gray Value (average gray value within the selection) or Integrated Density (the product of Area and Mean Gray Value). The number of mitotic cells was quantified according to an established method (Tarnowski et al., 1993). Briefly, DAPI nuclear staining was used to identify condensed chromosomes of mitotic cells compared to a more uniform ovally shaped nucleus in interphase cells. 2.5 Preparation of protein extracts and Western blotting Cells were seeded at a density of 1.5 x 106 cell per p100 plate or 200 000 cells per well in 6-well plates. After treatments, cells were scraped and pelleted. Protein extracts were prepared using 50 mM Tris/HCl pH 8.0, 0.1 mM EDTA, 0.5 % Triton X-100 and 12.5 mM β-mercaptoethanol as lysis buffer, and vortexing every 5 min for 45 min on ice. After centrifugation at 16 000 x g for 15 min, protein extracts were obtained in the supernantant. The protein concentration was measured by the Bio-Rad Protein Assay using bovine serum albumin as protein standard. Twenty micrograms of extracts were electrophoresed in 7.5% (w/v) SDSpolyacrylamide gels and transferred to polyvinylidene difluoride (PVDF) membranes using a Trans-Blot SD semidry system (Bio-Rad). After blocking in Tris-buffered saline (TBS) containing 5% (w/v) of non-fat dry milk and 0.3% Tween 20 for 1 h, immunostaining reactions were performed by incubating the membranes for 3 h at room temperature with the following primary antibodies diluted in TBS-1% (v/v) Tween 20: anti-SERCA2b (1:1000), Iba1 (1:500), anti-β actin (1:3000, Sigma), anti-PCNA (1:3000, Sigma) and anti β-tubulin (1:3000). Afterwards, membranes were incubated for 1 h at room temperature with peroxidase-conjugated secondary antibodies (1:5000, Bio- 16 cells treated with thapsigargin exhibited an amoeboid morphology without filopodia or lamellipodia. 3.4 SERCA2b is also upregulated in activated primary microglia, and its inhibition by thapsigargin stimulates microglial migration In order to corroborate the results obtained in BV2 cells, we also analyzed SERCA2b expression in microglia from primary cultures prepared from mouse brain. Immunocytochemistry showed the SERCA2b expression in microglial cells with a subcellular localization compatible with its presence in the ER (Fig. 4A), but also the expected change in cell morphology after induction of activation with LPS (Fig. 4A, B). Similar to BV2 cells, LPS-activated microglia showed an increase in the integrated density corresponding to the SERCA2b staining (Fig. 4C). This was correlated with increased levels of SERCA2b protein after LPS-mediated activation, as detected by Western blot (Fig. 4D). We also tested Iba1 expression in these cultures (Fig. 4E). A small increase of Iba1 expression was observed in LPS-activated primary microglia. This observation slightly differs from the results obtained with LPS-activated BV2 cells (Fig. 2E) or microglia of control and AD-brains (Fig 1E), but it is not unexpected considering that some variations in Iba1 expression level between studies and brain regions have been reported (Hopperton, Mohammad, Trepanier, Giuliano, & Bazinet, 2018). We also analyzed the effect of SERCA inhibition on primary microglia migration in scratch assays (Fig. 5). Cells from primary culture migrated to cover the scratch, although they migrated more slowly comparing with the BV2 cells. After specific inhibition of SERCA by thapsigargin for 48 h, cell migration was increased compared with control and LPS-treated cells (Fig. 5A), similar to the results observed in BV2 17 cells. Cellular adherence in the scratch was also better preserved after washing in thapsigargin-treated cells. Analysis of migration dynamics in the presence of thapsigargin also showed a change in the slope of the curves, in this case after 24 h, reaching the highest number of cells in the scratch region at the end of the assay (Fig. 5B), as shown above for BV2 cells (Fig. 3B). 3.5 Inhibition of SERCA by thapsigargin decreases phagocytosis in BV2 and primary microglial cells We analyzed the involvement of the SERCA pump in the phagocytic capacity of BV2 cells and primary microglia by using fluorescent red-latex beads (Fig. 6). Before fixation, cells were extensively washed to remove beads adhering to the cell surface, and the effective internalization of beads into the cells was verified in orthogonal projections of confocal microscopy (Fig. 6A). Then, we quantified cells that did not phagocytose (0), or phagocytosed few (1-4) or many (>5) beads. In the presence of LPS, we detected an increase in the percentage of cells that phagocytosed few (in BV2 cells) and many (in primary microglia) beads as compared to their controls. However, the treatment with thapsigargin led to a significant reduction of the phagocytic capacity in both BV2 and primary microglial cells. To analyze if SERCA was also involved in receptor-triggered phagocytosis, we determined microglial phagocytic activity after addition of the Aβ peptide, a hallmark in AD. The Aβ peptide is recognized in microglia by Toll-like receptors, such as TLR4, which also recognize LPS, causing Ca2+ influx to the cytosol and triggering a signaling cascade that leads to microglial activation (Molteni, Gemma, & Rossetti, 2016; Walter et al., 2007). We specifically used oligomeric Aβ42, since an effect of the aggregation state of the peptide on the microglial phagocytic efficiency has been described 18 (Gouwens, Makoni, Rogers, & Nichols, 2016; Paranjape, Gouwens, Osborn, & Nichols, 2012; Sun, Chen, & Wang, 2015). In these experiments, we also extensively washed the cells before fixation and verified that Aβ was inside the cells by confocal microscopy (Fig. 7A). In both cultures of BV2 cells and primary microglia, control cells showed a considerable phagocytosis of Aβ peptide (Fig. 7), proving the efficiency of these cells in removing toxic Aβ. This was not increased by addition of LPS, which binds to the same receptor. However, similarly to the experiments with latex beads, a significant reduction of Aβ phagocytosis was found in BV2 and primary microglial cells treated with the SERCA inhibitor thapsigargin. These results demonstrated that inhibition of SERCA activity also strongly prevented receptor-induced phagocytosis in microglial cells. Furthermore, we examined CD68 expression, as a marker for the microglial capacity of phagocytosis (Walker & Lue, 2015; Zotova et al., 2013). Although there was some CD68 expression in resting BV2 cells, cells phagocytosing Aβ significantly increased CD68 expression in comparison with cells in the absence of Aβ or in the presence of both Aβ and thapsigargin (Fig. 8). Considering that our results point to opposite effects of SERCA inhibition on phagocytosis and migration, we analyzed both processes following the application of Aβ in the scratch assays (Fig. 9). The presence of Aβ induced a significant reduction in the coverage of the scratch in both control and thapsigargin conditions. On the other hand, under control conditions, the most migrating cells found in the middle of the scratch showed a significant reduction in Aβ phagocytosis, while the nonor lessmigrating cells closer to the scratch exhibited a slightly higher phagocytosis but not reaching the phagocytic capacity found in cultures without scratch stimulation (Fig. 9D). These results on Aβ phagocytosis support the view that microglia have a low capacity of phagocytosis while they are involved in active migration, and vice versa. In 19 the presence of thapsigargin, the phagocytosis was impaired in both migrating and nonmigrating cells (Fig. 9C-D), as shown before (Figs. 6 and 7). 4. DISCUSSION This work analyzes the protein expression and subcellular localization of the ER Ca2+- ATPase SERCA2b in microglial cells as well as its contribution to microglial functions. We showed that SERCA2b is highly expressed in activated microglia from AD-brains as well as after activation in vitro with LPS in BV2 murine microglial cell line and primary microglia isolated from mouse brain. By exploring a gene expression database for microglia (Friedman et al., 2018), we have found our results at protein expression levels in good agreement with RNA seq data showing a significant higher expression of SERCA2 transcript in microglia extracted from LPS-treated mice (Srinivasan et al., 2016), and in embryonic and perinatal microglia, which are more activated than the branched adult microglia of healthy brains (Matcovitch-Natan et al., 2016). This enhanced SERCA2b expression can be understood as a cellular feedback aimed to restore the maintained high [Ca2+]c caused by cell activation. Additionally, this upregulation of SERCA2b protein seems to be specific for microglial cells, since changes in its expression were not observed in neurons of AD-affected brain regions in respect to controls either by immunohistochemistry (see Fig. 1) or by Western blot (Berrocal et al., 2015). In order to study the contribution of SERCA2b to Ca2+ signaling in microglia, we used the specific SERCA-inhibitor thapsigargin, which inhibits specifically the pump to an irreversible state (Sagara & Inesi, 1991). In fact, we have used 100 nM thapsigargin in 20 order to specifically block the SERCA pump, since micromolar concentrations of thapsigargin can inhibit other Ca2+ pumps as the SPCA (J. Chen et al., 2017). The SERCA inhibition prevents Ca2+ refilling of the ER, producing a net passive leak from the ER store that results in ER emptying and increased [Ca2+]c (Thastrup et al., 1990). Besides the Ca2+ depletion in the ER, the SERCA arrest may also induce a subsequent Ca2+ influx from the extracellular medium via store-operated channel entry (SOCE), since microglia express Ca2+ release-activated Ca2+ (CRAC) channels, similar to other immune cells (Parekh, 2010). In fact, SOCE has been reported for BV2 cells and primary microglia (Bagur & Hajnoczky, 2017; Heo, Lim, Nam, Lee, & Kim, 2015; Michaelis, Nieswandt, Stegner, Eilers, & Kraft, 2015), and the activation of this capacitive Ca2+ pathway can sustain long-lasting signals (Toescu et al., 1998). Therefore, the Ca2+ depletion in the ER via SERCA inhibition may trigger long-term effects on [Ca2+]c homeostasis that could explain the change in the slope and the highest stimulation of cell migration showed in our scratch assays in the presence of thapsigargin compared with the LPS-treatment. Combined incubation with LPS and thapsigargin showed less cell migration than each separate compound in both BV2 and primary microglia (data not shown), suggesting a negative effect via cytosolic Ca2+ overload from both plasma membrane and ER. Moreover, the inhibition of SERCA by thapsigargin produced a significant effect on actin cytoskeleton dynamics, changing microglial morphology to an amoeboid shape with blebs and ruffles, in line with faster moving cells. Actin dynamics are critical in cell migration and many actin-binding and -severing proteins are activated by high levels of Ca2+ in the cytosol (Choe et al., 2002; Ito et al., 1998). The change in morphology caused by thapsigargin also reduced adhesion in resting cells, but less in migrating cells, maybe because the formation of focal contacts required for moving 21 forward serves as anchorage sites. On the other hand, microglial activation may lead to readjusting mechanisms for Ca2+ handling despite Ca2+ depletion in the ER in migrating cells. This is worthy to investigate further, as compensation of different Ca2+ stores have been seen in other experimental models (Sepulveda, Vanoevelen, Raeymaekers, Mata, & Wuytack, 2009). Since microglia are the primary phagocytes in the brain parenchyma, they are responsible for clearance of apoptotic or necrotic cells (Green, Oguin, & Martinez, 2016) and for the removal of pathogens (Nau, Ribes, Djukic, & Eiffert, 2014), neurotoxic peptides and protein aggregates (Hansen, Hanson, & Sheng, 2018; Hoffmann et al., 2016). Hence, microglia play a central protective role in neurodegeneration. However, excessive pruning of synapses (Z. Chen et al., 2014; Rajendran & Paolicelli, 2018; Sierra, Abiega, Shahraz, & Neumann, 2013) or even phagoptosis of neurons (Brown & Neher, 2012, 2014) by microglia can also be enhanced in pathological conditions, reflecting the still unraveled dual role of microglia in neurodegenerative diseases, e.g. defective or overactivated phagocytosis has been found in AD and other neurodegenerative diseases (Janda, Boi, & Carta, 2018; Krasemann et al., 2017; Neniskyte, Neher, & Brown, 2011). Here, we showed the critical importance of Ca2+ transport into the ER on the modulation of phagocytic activity, since SERCA inhibition led to impaired phagocytosis. In addition, this impairment could be due to the effect of thapsigargin on cell morphology, as we observed a reduction of filopodia involved in microglial surveillance. Our treatment during 4-6 h with 100 nM thapsigargin induced a change in morphology from branched to amoeboid cells but it did not drastically affect cell viability, as observed in the MTT assay. In fact, we did not visualize symptoms of cell damage by microscopy, such as cell blebbing (by phase-contrast), pyknosis or nuclear 22 fragmentation (by DAPI staining) or disruption of cytoskeletal elements (by actin and tubulin stainings). However, it has been reported that prolonged exposure to thapsigargin can cause stress in the ER and other Ca2+ stores, and subsequently may evoke the unfolded protein response, caspase activation, mitochondrial release of apoptotic factors and activation of Ca2+-dependent endonucleases, eventually resulting in cell death (Foufelle & Fromenty, 2016; Tabas & Ron, 2011). In fact, local SERCA2b inhibition by thapsigargin-derivates and its effect on Ca2+ signaling has been recently explored as a potential anti-cancer therapy (Akinboye, Brennen, Denmeade, & Isaacs, 2019; Doan et al., 2015; Mahalingam et al., 2016). Although SERCA2b is ubiquitously expressed, this strategy relies on local activation of an inactive thapsigargin prodrug in the tumor environment, which may cause death specifically in the cancer cells. One of the most remarkable results of our study was the opposite effect caused by SERCA inhibition on microglial migration and phagocytosis. Many cellular processes can be specifically switched on or off depending on [Ca2+]c (Bootman & Bultynck, 2020), and our result suggests an opposite effect of [Ca2+]c on the molecular mechanisms regulating both processes. Thus, different activation states may be obtained depending on the type or strength of the extracellular signal. In fact, different extracellular nucleotides released after brain injury regulate different functions of activated microglia. Extracellular ATP stimulates chemotactic microglial migration (Haynes et al., 2006), whereas UDP promotes microglial phagocytosis (Inoue, 2007; Koizumi et al., 2007). In addition, high or low ATP concentrations induce oscillations in [Ca2+]c after activation of P2X7 or P2X4 receptors, respectively (Gilbert et al., 2016). Therefore, the same nucleotide can induce variable changes in Ca2+ dynamics in a concentration-dependent manner. These changes could promote, in turn, a shift from the migratory to the phagocytic state. It seems reasonable that after a first signal of brain 23 injury, microglial cells need to migrate but not phagocyte, and once that they are in the damaged area, microglia begin phagocytic activity for clearance of dying neurons or other toxic substances (Jonas et al., 2012). Accordingly, it has been reported in photodegenerative experiments that microglial cells migrate crossing all retinal layers until they reach the outer nuclear layer where they exhibit phagocytic activity of degenerating photoreceptors (Santos et al., 2010). In a similar way, during embryonic development, microglial cells actively migrating to colonize the retina do not phagocytose dead cell debris, which are engulfed by Müller cells (Marin-Teva, Cuadros, Calvente, Almendros, & Navascues, 1999). These in vivo observations are consistent with our present in vitro results, supporting the view that microglia have a low capacity of phagocytosis while they are involved in active migration. It is tempting to hypothesize that factors triggering activation of microglia are initially in concentrations that favor their migration, and are subsequently decreased to levels that promote phagocytosis. These different concentrations would differentially regulate microglial [Ca2+]c with the active participation of SERCA2b to alternatively promote molecular pathways responsible for migration or phagocytosis. In this sense, it has been recently reported that BV2 microglia silenced for Iba1, which binds Ca2+ and acts as actin-cross linking protein (Sasaki, Ohsawa, Kanazawa, Kohsaka, & Imai, 2001), migrated less, but exhibited increased phagocytic activity (Gheorghe et al., 2020), supporting the important role of Ca2+ signaling and cytoskeleton dynamics in these microglial functions. Recently a new ‘extremely active’ microglial phenotype has been described in pathological states such as AD, known as ‘dark microglia’ (Bisht et al., 2016). Among their ultrastructural characteristics, an extensive and dilated ER stands out. Since Ca2+ overload in the ER cause dilatation of cisternae, it is tempting to speculate that this 24 feature may correlate with the upregulation of SERCA2b we detected in vitro as well as in AD-affected brain´s microglia. Moreover, dark microglia present extreme phagocytic activity at the synapse (Bisht et al., 2016), where the modulation of SERCA activity could have a significant impact, opening a window to new therapeutic strategies. ACKNOWLEDGEMENTS We are grateful to Prof F. Wuytack (Katholieke Universiteit Leuven, Belgium) for critically reading of the manuscript and kind support, to Prof J. Ávila and Dr M. Bolós (Centre for Molecular Biology Severo Ochoa, CSIC-UAM, Spain) for helpful advice with the primary cultures of microglia, and to Prof. A. Osuna (Dept. of Parasitology, UGR) for his kind gift of phalloidin-FITC. This work was financially supported by grants mP_BS_35-2014 from CEI BioTic Granada, PP2016-PJI05 from University of Granada and A1-CTS-324-UGR18 from FEDER-Junta de Andalucía, Spain (to M.R.S.). We also acknowledge PP2016-PIP08 (to J.N.) from University of Granada, BFU2017-85723-P (to A.M.M.) from Spanish Ministry of Economy and Competitiveness co-financed with FEDER, and G044212N from Flanders Research Foundation (to P.V.). J.M.M.R. received a postgraduate scholarship provided by Ministry of Education, Culture and Sports, Spain. REFERENCES 25 abd-el-Basset, E., & Fedoroff, S. (1995). Effect of bacterial wall lipopolysaccharide (LPS) on morphology, motility, and cytoskeletal organization of microglia in cultures. J Neurosci Res, 41(2), 222-237. doi:10.1002/jnr.490410210 Ahmed, Z., Shaw, G., Sharma, V. P., Yang, C., McGowan, E., & Dickson, D. W. (2007). Actin-binding proteins coronin-1a and IBA-1 are effective microglial markers for immunohistochemistry. J Histochem Cytochem, 55(7), 687-700. doi:10.1369/jhc.6A7156.2007 Akinboye, E. S., Brennen, W. N., Denmeade, S. R., & Isaacs, J. T. (2019). Albuminlinked prostate-specific antigen-activated thapsigarginand niclosamide-based molecular grenades targeting the microenvironment in metastatic castrationresistant prostate cancer. Asian J Urol, 6(1), 99-108. doi:10.1016/j.ajur.2018.11.004 Alawieyah Syed Mortadza, S., Sim, J. A., Neubrand, V. E., & Jiang, L. H. (2018). A critical role of TRPM2 channel in Abeta42 -induced microglial activation and generation of tumor necrosis factor-alpha. Glia, 66(3), 562-575. doi:10.1002/glia.23265 Arcuri, C., Mecca, C., Bianchi, R., Giambanco, I., & Donato, R. (2017). The Pathophysiological Role of Microglia in Dynamic Surveillance, Phagocytosis and Structural Remodeling of the Developing CNS. Front Mol Neurosci, 10, 191. doi:10.3389/fnmol.2017.00191 Baba-Aissa, F., Raeymaekers, L., Wuytack, F., De Greef, C., Missiaen, L., & Casteels, R. (1996). Distribution of the organellar Ca2+ transport ATPase SERCA2 isoforms in the cat brain. Brain Res, 743(1-2), 141-153. doi:10.1016/s00068993(96)01037-2 Baba-Aissa, F., Raeymaekers, L., Wuytack, F., Dode, L., & Casteels, R. (1998). Distribution and isoform diversity of the organellar Ca2+ pumps in the brain. Mol Chem Neuropathol, 33(3), 199-208. Bagur, R., & Hajnoczky, G. (2017). Intracellular Ca(2+) Sensing: Its Role in Calcium Homeostasis and Signaling. Mol Cell, 66(6), 780-788. doi:10.1016/j.molcel.2017.05.028 Berridge, M. J., Lipp, P., & Bootman, M. D. (2000). The versatility and universality of calcium signalling. Nat Rev Mol Cell Biol, 1(1), 11-21. doi:10.1038/35036035 Berrocal, M., Corbacho, I., Vazquez-Hernandez, M., Avila, J., Sepulveda, M. R., & Mata, A. M. (2015). Inhibition of PMCA activity by tau as a function of aging and Alzheimer's neuropathology. Biochim Biophys Acta, 1852(7), 1465-1476. doi:10.1016/j.bbadis.2015.04.007 Berrocal, M., Sepulveda, M. R., Vazquez-Hernandez, M., & Mata, A. M. (2012). Calmodulin antagonizes amyloid-beta peptides-mediated inhibition of brain plasma membrane Ca(2+)-ATPase. Biochim Biophys Acta, 1822(6), 961-969. doi:10.1016/j.bbadis.2012.02.013 Bezprozvanny, I., & Mattson, M. P. (2008). Neuronal calcium mishandling and the pathogenesis of Alzheimer's disease. Trends Neurosci, 31(9), 454-463. doi:10.1016/j.tins.2008.06.005 Bisht, K., Sharma, K. P., Lecours, C., Sanchez, M. G., El Hajj, H., Milior, G., . . . Tremblay, M. E. (2016). Dark microglia: A new phenotype predominantly associated with pathological states. Glia, 64(5), 826-839. doi:10.1002/glia.22966 Bolos, M., Llorens-Martin, M., Jurado-Arjona, J., Hernandez, F., Rabano, A., & Avila, J. (2016). Direct Evidence of Internalization of Tau by Microglia In Vitro and In Vivo. J Alzheimers Dis, 50(1), 77-87. doi:10.3233/JAD-150704 32 Fig. 2 SERCA2b is upregulated in LPS-activated BV2 cells. A. Double immunofluorescence of SERCA2b (red) and β tubulin (green) in BV2 cells cultured in control medium (Ctr) and 100 ng/ml LPS-containing (LPS) medium. Cell nuclei were visualized with DAPI (blue). B. Scatter plot showing the distribution of form factor and aspect ratio parameters of control and LPS-activated BV2 cells. C. Integrated density of SERCA2b immunostaining in controls and LPS-activated cells showing a significantly higher value in the latter condition (p<0.005, asterisk). D and E. Representative Western blots using the anti-SERCA2b (D) and the microglial marker Iba1 (E) antibodies. Protein extracts of adult mouse brain (Brain) were used as control of the SERCA2b immunoreaction, and the β-actin was used as loading control. Quantifications of band intensities showed a significant difference (p<0.005, asterisk) in SERCA2b immunostaining in LPS-treated cells compared with untreated cells, that was not observed in the Iba1 immunostaining. Data in C-E are means±SE of three experiments performed in triplicates. Scale bar: 20 µm. Fig. 3 Inhibition of SERCA stimulates migration in BV2 cells. A. Representative images of scratch assays at the beginning of experiments (0 h) and after 6 h of culture in control (Ctr), 100 ng/ml LPS-containing (LPS) and 100 nM thapsigargin-containing (Tg) medium, before (upper panel) and after washing (lower panel). B. Percentages of occupied areas outside (black bars) and inside (grey bars) the scratch region are represented at 0 h and after 6 h in control medium (LPSTg-) and in the presence of LPS (LPS+ Tg-) or thapsigargin (LPSTg+), without (6 h) or with washing (6 h + washing). C. Area occupied by cells in the scratch region at different time points (in percentage). D. SERCA2b integrated density of BV2 cells in the scratch assay in control 33 and thapsigargin conditions at 6 h, compared with control cultures at 0 h. E. Quantification of the number of mitosis per area unit (equivalent to the scratched area in the image), outside (black bars) and inside (grey bars) the scratch region after 6 h with the different treatments (Ctr, LPS, Tg). F. Western blot of the cell proliferation marker PCNA from BV2 extracts treated with LPS or thapsigargin, and adult mouse brain extracts. β-actin was used as loading control. G. Phalloidin staining (green) to visualize actin at different time points of the scratch assay. Cells showed thin filopodia (arrows) at the beginning of the experiment, lamellipodia (arrowheads) after 5 min in both control and thapsigargin conditions. Unlike controls, cells turned to an amoeboid morphology (arrows) after 2 h of thapsigargin treatment. Cell nuclei were visualized with DAPI staining (blue). H. Phalloidin (green) and DAPI (blue) staining of cells outside the scratch after 6 h. Control cells showed thin filopodia while cells treated with thapsigargin exhibited an amoeboid morphology. Data in B, C, D and F are means±SE of three experiments performed in duplicates and asterisks denote significant differences (p<0.05) of data compared with their respective controls. Scale bars: 100 µm (A), 5 µm (G, H). Fig. 4 SERCA2b is overexpressed in microglial primary cultures from mouse brain after LPS activation. A. Double immunofluorescence of SERCA2b (red) and β tubulin (green) in primary microglial cells cultured in control (Ctr) and 100 ng/ml LPScontaining (LPS) medium. Cell nuclei were visualized with DAPI (blue). Localization of SERCA2b was compatible with its presence in the ER and the expected change in cell morphology after microglial activation with LPS was clearly observed. B. Scatter plot showing the distribution of form factor and aspect ratio parameters of control and LPS-activated microglia. C. Integrated density of SERCA2b immunodetection in 34 control and LPS-activated cells. D and E. Representative Western blots using the antiSERCA2b (D) and the anti-Iba1 (E) antibodies, with β-tubulin as loading control (left). Quantifications are shown on the right. Data in C-E are means±SE of three experiments performed in triplicates and asterisks denote significant differences (p<0.005) of data compared with their respective controls. Scale bar: 20 µm. Fig. 5 SERCA inhibition stimulates migration in murine primary microglia. A. Representative images (left) of scratch assays at the beginning of experiments (0 h) and after 48 h in control (Ctr), 100 ng/ml LPS-containing (LPS) and 100 nM thapsigargincontaining (Tg) medium, without (upper panel) and after washing (lower panel). Scale bar: 75 µm. The number of cells per area unit (equivalent to the scratched area in the image) outside (black bars) and inside (grey bars) the scratch area are represented in the graph at 0 h and after 48 h in control medium (LPSTg-) and in the presence of LPS (LPS+ Tg-) or thapsigargin (LPSTg+) before (48 h) or after washing (48 h + washing). B. Numbers of cells per area unit (cells/area) in the scratch at different time points. Data in A and B are means±SE of three experiments performed in duplicates and asterisks denote significant differences (p<0.005) of data compared with their respective controls. Fig. 6 Inhibition of SERCA decreases phagocytosis of latex beads in BV2 cells and primary microglia from mouse brain. A. BV2 cells subjected to the phagocytosis assay using red-fluorescent latex beads. Brightfield image merged with red (beads) and blue (DAPI nuclear staining) channels is shown on the left. On the right, the orthogonal projection obtained by confocal microscopy showed that beads (red) were inside the cell. B. BV2 and primary microglial (Mic) cells subjected to the phagocytosis assay in control (Ctr), 100 ng/ml LPS-containing (LPS) or 100 nM thapsigargin-containing (Tg) 35 medium. Representative images show merged brightfield and red fluorescent channels and demonstrate moderate, high and scarce phagocytosis of beads in Ctr, LPS and Tg conditions, respectively. Scale bars: 15 µm (A), 20 µm (B, in Ctr and LPS), 50 µm (B, in Tg). C. Percentages of BV2 and microglial cells that did not phagocytose (0, white bars) or phagocytosed few (1-4, gray bars) or many (>5, black bars) beads. Data are means±SE of three experiments performed in triplicates. Asterisks denote significant differences (p<0.005) of data compared with their respective controls. Fig. 7 Inhibition of SERCA decreases phagocytosis of amyloid-β peptide in BV2 cells and primary microglia. A. BV2 cell incubated with oligomeric Aβ peptide whose phagocytosis was detected by immunostaining with anti-Aβ antibody (green). Cell morphology is visualized on the left in a brightfield image merged with images showing anti-β-tubulin (red) immunocytochemistry and DAPI (blue) nuclear staining. On the right, the orthogonal projection obtained by confocal microscopy showed that Aβ (green) was inside the cell. B. Phagocytosis of oligomeric Aβ (green) in BV2 (BV2) and primary microglial (Mic) cells in control (Ctr), 100 ng/ml LPS-containing (LPS) or 100 nM thapsigargin-containing (Tg) medium. Cell morphology and nuclei were visualized by anti-β-tubulin antibody (red) and DAPI (blue), respectively. Scale bars: 15 µm (A), 20 µm (B). C. Quantification of Aβ phagocytosis per cell in terms of integrated density. Data are means±SE of three experiments performed in triplicates. Asterisks denote significant differences (p<0.05) of data compared with their respective controls. Fig. 8 Immunodetection of CD68 in a phagocytosis assay with amyloid-β peptide in BV2 cells. A. BV2 cells were incubated with vehicle (-Aβ) or oligomeric Aβ peptide (+Aβ) in control (Ctr), 100 ng/ml LPS-containing (LPS) or 100 nM thapsigargin- 36 containing (Tg) medium. Aβ phagocytosis was detected by immunostaining with antiAβ antibody (red), CD68 expression with the anti-CD68 antibody (green) and nuclei were visualized by DAPI staining (blue). Scale bar: 25 µm. B. Quantification of CD68 integrated density. Data are means±SE of three experiments performed in triplicates. Asterisks denote significant differences (p<0.05) of data compared with their respective controls. Fig. 9 Presence of Aβ reduces migration in BV2 cells. A. Representative images of scratch assays after 6 h of culture in control (Ctr) or 100 nM thapsigargin-containing (Tg) medium, in the absence (-Aβ) or the presence of oligomeric Aβ peptide (+Aβ). Cells were visualized by DAPI staining. B. Quantification of cells in the scratch region after 6 h of Aβ application revealed less cells in the scratch region in both control and Tg conditions. C. Immunodetection of Aβ phagocytosis (red) in two areas: in the middle of the scratch and at the border of the scratch. The β-tubulin immunodetection (green) and DAPI staining (blue) were used to visualize cell morphology and nuclei, respectively. Migrating cells showed reduced phagocytosis compared with nonor lessmigrating cells close to the border of the scratch (arrows). D. Quantification of Aβphagocytic cells (in percentage). Data in B and D are means±SE of three experiments performed in triplicates. Asterisks denote significant differences (p<0.05) of data compared with their respective controls. Scale bars: 100 µm (A), 50 µm (B). 37 Figure 1 38 Figure 2 39 Figure 3 40 Figure 4 41 Figure 5