Calcium-induced alteration of mitochondrial morphology and mitochondrial-endoplasmic reticulum contacts in rat brown adipocytes.

Mitochondria are key organelles maintaining cellular bioenergetics and integrity, and their regulation of [Ca2+]i homeostasis has been investigated in many cell types. We investigated the short-term Ca-SANDOZ® treatment on brown adipocyte mitochondria, using imaging and molecular biology techniques. Two-month-old male Wistar rats were divided into two groups: Ca-SANDOZ® drinking or tap water (control) drinking for three days. Alizarin Red S staining showed increased Ca2+ level in the brown adipocytes of treated rats, and potassium pyroantimonate staining localized electron-dense regions in the cytoplasm, mitochondria and around lipid droplets. Ca-SANDOZ® decreased mitochondrial number, but increased their size and mitochondrial cristae volume. Transmission electron microscopy revealed numerous enlarged and fusioned-like mitochondria in the Ca-SANDOZ® treated group compared to the control, and megamitochondria in some brown adipocytes. The Ca2+ diet affected mitochondrial fusion as mitofusin 1 (MFN1) and mitofusin 2 (MFN2) were increased, and mitochondrial fission as dynamin related protein 1 (DRP1) was decreased. Confocal microscopy showed a higher colocalization rate between functional mitochondria and endoplasmic reticulum (ER). The level of uncoupling protein-1 (UCP1) was elevated, which was confirmed by immunohistochemistry and Western blot analysis. These results suggest that Ca-SANDOZ® stimulates mitochondrial fusion, increases mitochondrial-ER contacts and the thermogenic capacity of brown adipocytes.

Introduction

Brown adipose tissue (BAT) has the unique feature of non-shivering heat production, an essential physiological function which maintains body temperature and dissipates excess energy to control obesity.[1-3] BAT is mainly found in small rodents, but is also found in human newborns and adults.[4] Exposure to cold stimulates the proliferation and differentiation of brown adipocytes via adrenergic receptors, and this physiological response involves mitochondrial biogenesis and activation of genes encoding several mitochondrial proteins, including uncoupling protein-1 (UCP1).[5, 6] UCP1 utilizes proton-motive force for heat production instead of ATP production.[7]

Adrenergic stimulation also increases cytosolic concentrations of Ca2+.[8, 9] Uptake of Ca2+ by brown adipocytes is controlled by the cooperation of 1 and receptors stimulation, as 1-adrenergic stimulation leads to intracellular Ca2+ mobilization,[10] and -stimulation leads to inhibition of Ca2+ uptake into the mitochondria.[11] Following noradrenaline treatment, cultured brown adipocytes exhibited an increase in cytosolic Ca2+ levels[12] as a result of a flux of both intracellular and extracellular Ca2+ sources.[13] Furthermore, in experiments on cold-acclimated brown adipocytes, Nakagaki et al.[14] suggested that noradrenaline and forskolin evoked an elevation in cytosolic Ca2+ level followed by mitochondrial Ca2+ level, and that Ca2+ accumulation in mitochondria depends on the H+ gradient produced by electron transport in the respiratory chain. This is in line with earlier studies which indicated that mitochondria can accumulate a large amount of Ca2+ in dependence of mitochondrial energy metabolism.[15, 16] Mitochondria are physically linked to the endoplasmic reticulum (ER) in many cell types, forming a structure known as the mitochondria-associated ER membrane (MAM).[17] A recent study revealed that MAM enables mitochondria and ER to exchange calcium and lipids and that ER-mitochondria contacts are involved in regulating mitochondrial energy metabolism.[18]

Mitochondria are dynamic organelles constantly undergoing fission and fusion. During fusion, mitochondria form a mitochondrial network or reticulum, and mitofusins [mitofusin 1 (MFN1) and mitofusin 2 (MFN2)] are the proteins required for fusion of the outer mitochondrial membrane,[19, 20] and optic atrophy 1 protein (OPA1) is required for inner membrane fusion.[21] Under normal conditions, fusion of the outer membrane is coordinated with fusion of the inner membrane, however, some reports have shown that the fusion machinery of the outer membrane can operate in the absence of inner membrane fusion.[22, 23] The master regulator of mitochondrial fission is dynamin-related protein 1 (DRP1). Cells lacking DRP1 show highly interconnected mitochondrial reticulum formed by fusion in the absence of fission activity.[24]

De Meis et al.[25] performed in vitro experiments with brown adipocytes and showed that mitochondria fusion and influx of calcium into these organelles activate thermogenesis. In this study, we examined the in vivo effect of short-term calcium treatment on mitochondrial dynamics in brown adipocytes, and investigated how Ca2+ affects: i) the balance of mitochondrial fusion and fission; ii) mitochondrial-ER contacts; and iii) the thermogenic capacity of brown adipocytes.

Materials and Methods

Experimental design

This study was approved by the Ethics Committee for the Treatment of Experimental Animals (Faculty of Biology, University of Belgrade, Serbia). Two-month-old male Wistar rats (160-240 g) were fed standard pelleted food ad libitum. They were divided into two groups each consisting of six animals - Ca-SANDOZ® (Novartis, Basel, Switzerland) drinking (480 mg/L Ca2+) and tap water (control) drinking for three days. Animals were sacrificed using a decapitator (Harvard Apparatus, Holliston, MA, USA). The inter-scapular portion of BAT was dissected out and weighed. One portion of BAT was fixed in 2.5% glutaraldehyde (v/v) in 0.1 M Sørensen phosphate buffer (PB, pH 7.2). This portion was postfixed in 2% osmium tetroxide in the same buffer, then routinely dehydrated using increasing concentrations of ethanol and embedded in Araldite (Fluka, Seeize, Germany). One half of another portion of BAT was immediately frozen for protein analyses and other half was cut using a tissue chopper for confocal microscopy analyses.

Transmission electron microscopy

Ultra-thin sections were obtained using a Leica UC6 ultramicrotome (Leica Microsystems, Wetzlar, Germany), mounted on copper grids and contrasted in uranyl acetate and lead citrate using Leica EM STAIN (Leica Microsystems). Sections were examined on a Philips CM12 transmission electron microscope (Philips/FEI, Eindhoven, The Netherlands) equipped with the digital camera SIS MegaView III (Olympus Soft Imaging Solutions, Münster, Germany). The obtained electron micrographs were used for ultrastructure analysis (mitochondrial morphology) and stereology. The number of mitochondria was obtained by counting the total number of mitochondria per 100 µm[2] of cell surface. For this analysis, we used 50 randomly selected micro-graphs per group. Mitochondrial mean diameter was calculated as the average diameter per mitochondrion, 200 mitochondria per group were randomly selected. Mitochondrial volume density was calculated using a derivation of the Deless equation: Vv=Pm/Ptotal, where Pm/Ptotal is the point fraction or the total points hitting the mitochondria divided by the total points hitting the adipocyte.[26] Mitochondrial cristae volume density was calculated using a previously described equation with minor modifications, where Pc/Pmt is the point fraction or the total points hitting the cristae divided by the total points hitting the mitochondria. All these measurements and analyses were performed using iTEM software (Olympus Soft Imaging Solutions).

Stereological analyses of brown adipocytes and lipid droplets

Semi-thin sections (2 µm) were mounted on glass slides, and stained with 0.1% toluidine blue in 1% sodium borate buffer. Stereological values were obtained using the equation: Vv=Px/Ptotal, where Px/Ptotal is the point fraction or the total points hitting the brown adipocytes or lipid droplets, respectively, divided by the total points hitting BAT. All these analyses were performed on a Leica DMLB microscope (Leica Microsystems) using a Weibel 2 graticule.[26]

Immunohistochemistry

After removal of Araldite with 1% sodium hydroxide in absolute ethanol (30 min, 37ºC), semi-thin BAT sections (2 µm), were rehydrated with decreasing concentrations of ethanol, subsequently incubated in citrate buffer for 3 min at 600W to retrieve antigens and washed in phosphate buffered saline (PBS, pH 7.4). After blocking endogenous peroxidase by incubating with 3% hydrogen peroxide in methanol and three sequential washings in PBS, the sections were incubated with rabbit polyclonal UCP-1 antibody (Abcam, Cambridge, UK) overnight at 4ºC. Immunodetection was performed using the Dako LSAB Universal Kit (Dako Scientific, Glostrup, Denmark). The sections which were washed three times were incubated with 0.012% hydrogen peroxide and 0.05% diaminobenzidine (Sigma-Aldrich, Seeize, Germany) in PBS for 10 min in the dark. Finally, after rinsing in distilled water, the sections were counterstained with hematoxylin, mounted and examined with a Leica DMLB microscope (Leica Microsystems).

Confocal imaging and colocalization

Immediately after dissection, BAT was minced into small tissue fragments using a McIlwain tissue chopper (Mickle Laboratory Engineering, Guildford, UK), then incubated in PBS with 100 nM MitoTracker Orange CMTMRos (Invitrogen, Darmstadt, Germany) and 1 µM ER-Tracker Green (Invitrogen) for 45 min at 37ºC. After incubation, the samples were washed twice in PBS, and fixed with 4% paraformaldehyde in PBS. After a final washing in PBS, the samples were mounted with Mowiol and analyzed by a Leica TCS SP5 II confocal microscope (Leica Microsystems). The colocalization rate was determined using LAS AF software (Leica Microsystems) under the following parameters and formulas: threshold 30%, background 20%, colocalization rate [%] = colocalization area/area foreground, and area foreground = area image – area background.

Relative fluorescence intensity of MitoTracker Orange CMTMRos was determined at several regions of interest (ROI) using LAS AF software. ROI consisting of randomly localized intracellular regions were drawn by free hand and applied to images.

Alizarin Red S staining

After incubation in 1% sodium hydroxide in absolute ethanol for 30 min at 37ºC and rehydration with decreasing concentrations of ethanol, semi-thin BAT sections were incubated in Alizarin Red S solution for 2 min, followed by acetone, acetone/xylene and xylene dehydration. Sections were mounted, and analyzed using a Leica DMLB microscope (Leica Microsystems, Germany). Alizarin Red S shows the presence and localization of calcium by forming insoluble chelates with calcium ions from their deposits which gives an orange to red reaction.

Potassium pyroantimonate staining

BAT was fixed in 2.5% glutaraldehyde (v/v) in 0.1 M phosphate buffer (pH 7.2) containing 2% potassium pyroantimonate (w/v). Samples were washed, postfixed in 2% osmium tetroxide in phosphate-buffered antimonate. After rinsing with the same buffer without antimonate, the samples were routinely dehydrated using ethanol and embedded in Araldite. For the control samples, buffer without potassium pyroantimonate was used in the fixation and postfixation steps. Antimonate forms a precipitate with calcium ions which demonstrates the in situ localization of calcium.

SDS-PAGE and Western blotting

Protein content was estimated by the method of Lowry et al.[27] using bovine serum albumin as a reference. Immunoblot analyses were performed on isolated BAT as described previously.[28] Primary antibodies against MFN1 (5 µg/mL), MFN2 (5 µg/mL), DRP1 (1:1000), voltage-dependent anion-selective channel (VDAC) (1:1000), calnexin (1 µg/ml) and UCP1 (1:1000) were purchased from Abcam (Cambridge, UK). Quantitative analyses of immunoreactive bands were carried out using ImageJ software (NIH, USA). The volume represents the sum of all pixel intensities within a band, and 1 pixel = 0.007744 mm[2]. The density of each protein signal was normalized to that of the internal control, β-actin (Abcam, UK). The mean values obtained from the control group were taken as 100% and those from the Ca-SANDOZ® group were expressed as percentages with respect to the control.

Statistics

The Student’s t-test was performed for analyses of the data using GraphPad Prism for Mac OS X. Statistical significance was set at P<0.05.

Results and Discussion

Increased Ca[2]+ level and its localization in brown adipocytes

Alizarin Red S histochemical staining revealed a higher level of Ca2+ salts in the treated group, visualized as groups of brown adipocytes with higher intensities of red color (Figure 1 A,B). This is known as the harlequin effect,[29] due to the different functional capacity of brown adipocytes. At the ultrastructure level, potassium pyroantimonate staining localized Ca2+ as electron-dense regions in the cytoplasm, mitochondria and around lipid droplets, with higher concentrations in the Ca-SANDOZ® treated group compared to the control group (Figure 1 C,D). The negative control revealed no pyroantimonate precipitates in brown adipocytes (Figure 1E). Ca2+ uptake from the extracellular medium is necessary to maintain elevated cytosolic Ca2+ concentration as a result of hormone response in many cell types.[30] Depletion of Ca2+ from intracellular stores, such as ER or mitochondria, causes Ca2+ intake via Ca2+-permeable channels in the plasma membrane.[31] This is in accordance with our results which clearly shows that even short Ca-SANDOZ® intake induce Ca2+ uptake by rat brown adipocytes and their accumulation in cytoplasm and mitochondria. Brown adipocytes exhibited an increase in intracellular Ca2+ in vitro caused by a flux of both intracellular and extracellular Ca2+ after noradrenaline application via the -adrenergic and/or -adrenergic pathway.[12, 13] The mitochondrial uncoupling viaUCP-1 by 3 -adrenergic activation causes Ca2+ release from mitochondria and subsequently from the ER via inositol 1,4,5-triphosphate (InsP3) receptors, and further activates plasmalemmal Ca2+ entries into rodent brown adipocytes.[32] Hence, it could be suggested that Ca-SANDOZ® supplementation mimics cold-exposure and alters mitochondrial morphology and thermogenic capacity of rat brown adipocyte.

Figure 1.

Alizarin Red S and potassium pyroantimonate staining. Alizarin Red S staining showed a higher level of calcium salts in the Ca-SANDOZ® treated group (B) compared to the control group (A). Potassium pyroantimonate staining localized Ca2+ electron-dense regions in the control (C) and Ca-SANDOZ® group (D). Negative control for pyroantimonate staining (E). Scale bars: A,B, 20 µm; C-E, 2 µm.

Ca-SANDOZ® alters mitochondrial morphology and dynamics

Mitochondria of brown adipocytes show many variations in internal structure. TEM showed that mitochondria from the control group were mainly round with few lamellar cristae (Figure 2A). In contrast, the treated group had larger, fusioned-like mitochondria with numerous well-developed lamellar cristae (Figure 2B) and may be relevant to subsequent swelling of mitochondria.[3] Megamitochondria were found in the Ca-SANDOZ® treated group in some brown adipocytes (harlequin effect) (Figure 3), which has been described in the literature for various pathological disorders or perturbed physiological conditions.[33, 34] Megamitochondria are also associated with various diseases, where liver and skeletal muscles are the major tissues in which these structures have been detected. They are frequently associated with alcoholic liver disease, hepatic steatosis and other diseases,[35-37] where pathological changes, e.g. paracrystalline inclusions are detected.[38] In our experiment, no megamitochondria were observed in the control group. There are several possible mechanisms of megamitochondria formation. Hypertrophy of an individual mitochondrion has been detected in the liver of experimental animals fed a high protein diet, due to increased nitrogen metabolism.[39] Chloramphenicol induces the formation of hepatic megamitochondria, probably due to protein synthesis inhibition in mitochondria.[40] Free radicals initiate the formation of megamitochondria when they exceed the enzymatic and non-enzymatic defense system against reactive oxygen species (ROS). The rate of oxygen consumption and phosphorylation capacity of megamitochondria is often lowered[41-43] or remained normal.[44] Since TEM did not show any pathological changes in mitochondria and brown adipocytes, it cannot be determined what caused the appearance of these specific mitochondrial structures due to controversies regarding the formation of megamitochondria. Further investigations are needed to clarify the mechanism of a high Ca2+ diet on megamitochondria formation in brown adipocytes. Intimate contact between mitochondria and the ER was observed (Figure 2).

Figure 2.

Electron microscopy of brown adipocyte mitochondria. A) Round mitochondria with few lamellar cristae showing intimate contact between mitochondria (black arrowheads) and between mitochondria and endoplasmic reticulum (ER; white arrowheads) in the control group. B) Larger mitochondria with numerous cristae in the Ca-SANDOZ® group. Inset: Detail of intimate contact between mitochondria and ER. Scale bars: 2 µm.

Figure 3.

Megamitochondria. A brown adipocyte from the Ca-SANDOZ® treated group shows a mitochondrial cluster with megamitochondria (star) and numerous intermitochondrial contacts (arrows). Scale bar: 1 µm.

Ca-SANDOZ® treatment caused a significant decrease in mitochondria number per 100 µm[2] and shifted the mean mitochondrial diameter to a higher value (Figure 4 A,B), which is consistent with the observed fusion process. Stereological analyses also showed an increase in mitochondrial cristae volume (Figure 4C). These results are in line with reports on the increasing number and density of cristae in brown adipocytes after exposure to cold.[45, 46] Increased cristae fraction, i.e. remodeling may be in response to mitochondrial fusion to avoid an excessive membrane area and maintain the density of the matrix.[47] During mitochondrial remodeling, cells use several strategies to increase ATP production including an increase in mitochondrial size, an increase in the total number of mitochondria per cell, and an increase in the volume of mitochondria per cell and/or surface area of mitochondrial cristae membranes. A wide variety of cells use these strategies during transitions between the normal physiological milieu, pharmacological or toxicological treatments, and pathological or disease states.[48, 49] In apoptosis, mitochondrial remodeling is important for mobilization of cytochrome c stores in cristae.[50] The volume density of mitochondria was not changed (Figure 4D), which was confirmed by the unchanged level of relative protein expression of VDAC, a mitochondrial outer membrane marker (Figure 4E). As expected, the levels of calnexin, an ER membrane marker, were not changed (Figure 4F). These results indicate that short-term in vivo Ca2+ treatment does not stimulate mitochondrial and ER biogenesis, but induces their rearrangement and numerous contact sites.

Figure 4.

Mitochondrial stereology and analyses. Calnexin (endoplasmic reticulum membrane marker) protein level. A) Mitochondrial number per unit surface was decreased in adipocytes of the Ca-SANDOZ® treatment group. B) Distribution histogram of mitochondrial diameters demonstrating a shift to higher values after Ca-SANDOZ® treatment. C) Distribution histogram of cristae volume density; treatment stimulated cristae formation. D) Volume density of mitochondria was unchanged. E) The relative protein expression level of VDAC, a mitochondrial outer membrane marker, was unchanged. F) The relative protein expression level of calnexin, an ER membrane marker, was unchanged. Protein content is expressed relative to the control, which was standardized to 100%, as the mean ± SEM and represents three similar independent experiments with triplicate observations in each experiment. A representative blot is shown. Volume is the sum of all pixel intensities within a band (1 pixel = 0.007744 mm[2]).

Mitochondrial fusion players, MFN1 and MFN2, were significantly increased in the treated group (Figure 5 A,B), suggesting mitochondrial fusion.[19] On the other hand, DRP1, a mitochondrial fission player, was decreased (Figure 5C). Mitochondria are dynamic organelles constantly undergoing fission and fusion, thus the ER provides a constant and well-regulated supply of phospholipids for cristae and membrane integrity.[51-53] More MFN2 proteins were available for establishing mitochondria-ER contacts, which was elegantly demonstrated using mouse embryonic fibroblasts.[54] To ensure optimal bioenergetics and efficient mitochondrial respiration, ER-mitochondria contact is necessary for a constant Ca2+ supply viathe InsP3 receptor. This Ca2+ supply provides reducing equivalents to support oxidative phosphorylation.[55] In addition, Ca2+ can activate enzymes of the tricarboxylic acid cycle (TCA) including pyruvate dehydrogenase, -ketoglutarate and isocitrate dehydrogenase[17, 56] to increase flux throughout the TCA cycle. Furthermore, fusion of mitochondria may be protective against damage to mitochondrial DNA[57] due to elevated production of ROS as a result of increased respiratory activity[58, 59] and fatty acid oxidation.[60, 61]

Figure 5.

Markers of mitochondrial dynamics. A) Increased MFN1 relative protein expression compared to the control. B) Increased MFN2 relative protein expression compared to the control. C) Reduced DRP1 protein expression in BAT of treated rats. Protein content is expressed relative to the control, which was standardized to 100%, as the mean ± SEM and represents three similar independent experiments with triplicate observations in each experiment. A representative blot is shown. Volume is the sum of all pixel intensities within a band (1 pixel = 0.007744 mm[2]).

Confocal microscopy revealed a higher colocalization rate between active mitochondria and ER in the Ca-SANDOZ® treated group (Figure 6B), indicating increased mitochondria-ER contacts probably due to higher demands for Ca2+ inter-organelle crosstalk, as previously described.[62, 63] Higher mitochondrial membrane potential is observed in Ca-SANDOZ® treated group (Figure 6C), indicating elevated uptake of Ca2+ ions into mitochondria, which is in line with some studies revealing that Ca2+ uptake is driven by the mitochondrial membrane potential.[64-66] Our future study will be directed at this specific point to determine to what extent this communication is important for activation of brown adipocytes.

Figure 6.

Fluorescence labeling of mitochondria using MitoTracker Orange CMTMRos and ER using ER-Tracker Green. A) Confocal images of mitochondria and ER from the control group and Ca-SANDOZ® treated group. B) Increased colocalization rate between mitochondria and ER in the Ca-SANDOZ® treated group compared to the control group. C) Increased relative fluorescence intensity of MitoTracker Orange CMTMRos in the Ca-SANDOZ® treated group compared to control. Scale bars: 5 µm.

Ca-SANDOZ® increases thermogenic capacity of BAT

Immunohistochemical analysis showed higher expression of UCP1 protein in the Ca-SANDOZ® treated group (Figure 7B), which was confirmed by a significant increase in the level of relative UCP1 protein expression (Figure 8). Increased UCP1 expression leads to increased thermogenesis.[5],[67],[68] More cristae, where oxidative phosphorylation (OXPHOS) machinery is located,[69] are developed as a result of enhanced activity of the respiratory chain compensating proton leaks via the UCP1 uncoupling mechanism. However, the precise details of this process are still unclear.[70-72]

Figure 7.

UCP1 immunohistochemistry. Immunohistochemical analysis showed higher expression of UCP1 protein in the Ca-SANDOZ® treated group. A) Control group. B) Ca-SANDOZ® group. C) Negative control. Scale bars: 20 µm.

Figure 8.

UCP1 protein level. Increased level of UCP1 protein in the Ca-SANDOZ® group compared to the control group. Protein content is expressed relative to the control, which was standardized to 100%, as the mean ± SEM and represents three similar independent experiments with triplicate observations in each experiment. A representative blot is shown. Volume is the sum of all pixel intensities within a band (1 pixel = 0.007744 mm[2]).

In this experiment, we also observed body mass loss in the Ca-SANDOZ® treated group (-18.46±0.56%), compared to control animals. Stereological analyses of brown adipocytes and lipid droplets (Table 1) revealed that high calcium diet triggers lipolysis. This observation is in line with the process of adaptive thermogenesis, where catecholamine-stimulated lipolysis provides fatty acids for heat production in response to cold or overfeeding.[73] As a result of short-term thermogenesis, BAT may shrink due to increased lipolysis. Possible mechanisms of calcium effect on body weight were investigated using animal models and human volunteers.[74] Also, Marotte et al. reported that low calcium intake may increase body weight, especially in rats predisposed to obesity.[75] In humans, numerous studies have shown that high calcium intake may have an impact on weight loss.[76-80] Several mechanisms of high calcium diet are proposed: i) inhibition of lipogenesis and stimulation of lipolysis via suppression of circulating 1, 25-dihydroxycholecalciferol;[81-85] and ii) increase in fetal excretion of fat via the formation of insoluble calcium fatty acid soaps.[86, 87] From our results, it seems that high calcium uptake has a beneficial effect on weight control or loss, as a result of increased thermogenic capacity, at least in rats. Here, we propose a third mechanism – high calcium diet affects body weight due to increasing thermogenic capacity of brown adipocytes and, consequently, increased lipolysis for elevated energy expenditure.

Table 1.

Effect of Ca-SANDOZ® treatment on brown adipocytes.

	Control	Ca-SANDOZ®
Relative BAT mass (mg/100 g of animal weight)	82.6±5.39	77.03±2.36
Vv of brown adipocytes (%)	83.51±0.69	57.89±3.00*
Vv of lipid droplets (%)	43.63±2.14	27.67±0.87*

BAT, brown adipose tissue. Data represents the mean ± SEM. *P<0.005.

In conclusion, short-term in vivo treatment with Ca-SANDOZ® stimulates mitochondrial fusion, increases mitochondrial-ER contacts and thermogenic capacity in brown adipocytes. Further studies addressing the molecular mechanism of Ca2+ action on mitochondria are needed to understand their role in the mitochondrial bioenergetics of brown adipocytes during thermogenesis.