Enhanced mitochondrial buffering prevents Ca²⁺ overload in naked mole-rat brain

Abstract

Deleterious Ca²⁺ accumulation is central to hypoxic cell death in the brain of most mammals. Conversely, hypoxia-mediated increases in cytosolic Ca²⁺ are retarded in hypoxia-tolerant naked mole-rat brain. We hypothesized that naked mole-rat brain mitochondria have an enhanced capacity to buffer exogenous Ca²⁺ and examined Ca²⁺ handling in naked mole-rat cortical tissue. We report that naked mole-rat brain mitochondria buffer >2-fold more exogenous Ca²⁺ than mouse brain mitochondria, and that the half-maximal inhibitory concentration (IC₅₀) at which Ca²⁺ inhibits aerobic oxidative phosphorylation is > 2-fold higher in naked mole-rat brain. The primary driving force of Ca²⁺ uptake is the mitochondrial membrane potential (Δψ_m), and the IC₅₀ at which Ca²⁺ decreases Δψ_m is ~ 4-fold higher in naked mole-rat than mouse brain. The ability of naked mole-rat brain mitochondria to safely retain large volumes of Ca²⁺ may be due to ultrastructural differences that support the uptake and physical storage of Ca²⁺ in mitochondria. Specifically, and relative to mouse brain, naked mole-rat brain mitochondria are larger and have higher crista density and increased physical interactions between adjacent mitochondrial membranes, all of which are associated with improved energetic homeostasis and Ca²⁺ management. We propose that excessive Ca²⁺ influx into naked mole-rat brain is buffered by physical storage in large mitochondria, which would reduce deleterious Ca²⁺ overload and may thus contribute to the hypoxia and ischemia-tolerance of naked mole-rat brain.

Graphical Abstract

Introduction

Cellular Ca²⁺ is a key regulator of mitochondrial bioenergetics and cellular signalling. For example, Ca²⁺ accumulation in the mitochondrial matrix both stimulates aerobic metabolism (oxidative phosphorylation: OXPHOS; through the modulation of tricarboxylic acid (TCA) cycle and other enzymes; Griffiths & Rutter, 2009; Gellerich et al., 2010), and activates and/or amplifies numerous cellular cascades that impact mitochondrial function and fate (Polster & Fiskum, 2004; Pandya et al., 2013; Pamenter, 2014; Raffaello et al., 2016). Thus, tight regulation of mitochondrial and cytosolic Ca²⁺ concentrations ([Ca²⁺]_m and [Ca²⁺]_c, respectively) is important for maintaining neuronal homeostasis. In addition to being highly sensitive to Ca²⁺, mitochondria are also natural sponges of cytosolic Ca²⁺, which is rapidly pumped into the matrix by the mitochondrial calcium uniporter (MCU) complex, driven by the large negative mitochondrial membrane potential (−150 to −180 mV with regard to the extracellular milieu) (Finkel et al., 2015). However, during periods of low oxygen stress (e.g., hypoxia or ischemia), neuronal [Ca²⁺]_c becomes deleteriously elevated in the brain of hypoxia-intolerant mammals, and this Ca²⁺ overload is central to hypoxic/ischemic brain cell death.

Briefly, a rapid and large elevation (from 10⁻⁷ to 10⁻⁴ mole) of neuronal [Ca²⁺]_c occurs within ~ 8 min of hypoxia or ischemia onset (Silver & Erecinska, 1990). Most of this Ca²⁺ is sequestered by mitochondria in the form of calcium phosphate (Shen & Jennings, 1972; Kristian & Siesjo, 1998), but mitochondria quickly become overloaded, which leads to mitochondrial membrane depolarization, activation of the mitochondrial permeability transition pore (mPTP), and ultimately the catastrophic release of mitochondrial Ca²⁺ and other molecules from mitochondria, including cytochrome c, apoptosis-inducing factor, and serine protease HTRA2 (Htra2/Omi), which promote caspase-dependent and caspase-independent cell death pathways (Polster & Fiskum, 2004; Kroemer et al., 2007). Conversely, cytosolic Ca²⁺ overload is delayed or even prevented in the brain of hypoxia-tolerant species, such as naked mole-rats (NMRs; Heterocephalus glaber). Indeed, NMR brain tolerates acute and chronic in vivo hypoxia and in vitro ischemia (Nathaniel et al., 2009; Pamenter et al., 2018), such that mitochondrial membrane integrity is preserved and the electron transport system (ETS) remains tightly coupled (Pamenter et al., 2018; Cheng & Pamenter, 2021). This is partially due to blunted Ca²⁺ uptake by NMR neurons: hypoxia-mediated Ca²⁺ uptake is considerably slower and lower in magnitude in NMR brain than mouse (Peterson et al., 2012). However, mechanisms of mitochondrial Ca²⁺ handling that underlie this key difference have not been specifically evaluated in NMR brain.

We hypothesized that, relative to mouse brain, 1) NMR brain mitochondria have enhanced capacity to buffer otherwise deleterious [Ca²⁺]_c changes, and 2) this enhanced capacity both raises the [Ca²⁺]_m threshold for mPTP opening and preserves OXPHOS capacity and mitochondrial membrane potential (Δψ_m)_. To test this, we evaluated the ability of NMR and mouse brain mitochondria to take up and retain Ca²⁺, and the impact of exogenous Ca²⁺ addition on mitochondrial OXPHOS capacity and Δψ_m following in vitro Ca²⁺ titration. Furthermore, we recently observed enhanced mitochondrial connectivity and crista density in NMR brown fat tissue (Cheng et al., 2021), which has also been observed in hypoxia-tolerant ground squirrel brain (Popov et al., 2005), but not in hypoxia-intolerant rodent brain (Picard et al., 2015). Greater mitochondrial connectivity and crista density enhance mitochondrial bioenergetics and improve mitochondrial Ca²⁺ buffering capacity (Perkins et al., 2010). Moreover, mitochondrial Ca²⁺ management is tightly connected to mitochondrial dynamics in multiple tissues (Kazak et al., 2017; Guan et al., 2019). Therefore, the ultrastructure of may underlie improved mitochondrial Ca²⁺ management in NMR brain.

Materials and Methodology

Ethics approval.

NMRs (adult subordinates of mixed sex, 1–3 years-old) were group-housed in interconnected multi-cage systems at 30°C and 21% O₂ in 50% humidity with a 12L:12D light cycle. Animals were fed fresh tubers, vegetables, and fruit, and Pronutro cereal supplement ad libitum. 12–14-week-old female mice were obtained from Charles River and were housed at room temperature under a 12L:12D light cycle and fed rodent chow ad libitum. Animals were not provided with food during the experimental treatment periods and were not fasted prior to experimentation. All experimental procedures were approved by the University of Ottawa Animal Care Committee (protocol #2535) in accordance with the Animals for Research Act and the relevant guidelines of the Canadian Council on Animal Care.

Calcium uptake assay.

Animals were killed by acute cervical dislocation followed by immediate decapitation. Brains were rapidly dissected on ice. Ca²⁺ uptake by permeabilized cortical tissue was measured with the Ca²⁺ sensitive Calcium green 5-N fluorescent dye using the Oroboros O2k-Fluo module (Oroboros Instruments, Innsbruck, Austria) modified from (Spinazzi et al., 2019) at physiological temperatures (mice: 37°C; NMRs: 32°C). To minimize Ca²⁺ contamination, all solutions were made using deionized water, and glass syringes and the Oroboros chambers were soaked in 6 M HCl overnight and then rinsed repeatedly using deionized water before experiments. Ca²⁺ titration was measured in an EGTA free medium (in mM: KCl 120, HEPES 20, KH₂PO₄ 20, MgCl₂ 2.5, NaCl 5, and BSA 0.03; pH 7.2) containing 1 mg/ml brain tissue (100 mg wet weight/ml homogenized in ice-cold Tris solution (in mM: KCl 105, Tris-HCl 50, pH 7.1)), with 50 μg/ml saponin 1 μM Calcium green 5-N, and substrates combinations including 10 mM glutamate, 5 mM malate, 1 mM ADP, 1 mM ATP, 1 μM Ru360, and carbonyl cyanide 3-chlorophenylhydrazone (CCCP: 1 μM for mice and 1.5 μM for NMRs). Following 5–10 mins equilibration, permeabilized brain tissue (2 mg/ml) was exposed to Ca²⁺ titrations (one 50 μM Ca²⁺ injection and then five minutes later a series of 20 μM Ca²⁺ injections at two min intervals). Decreases in fluorescence represented mitochondrial Ca²⁺ uptake and Ca²⁺ release was detected as a fluorescence increase. The end point of each experiment was the point at which fluorescence increased, indicating release of Ca²⁺ from mitochondria. The total amount of titrated Ca²⁺ in the cumulative steps preceding the end point was then used to determine the maximal Ca²⁺ resistance capacity. A final bolus of Ca²⁺ was given after the end point to confirm that no further Ca²⁺ uptake occurred. Ca²⁺ uptake rates (indicated in this study as the decay rate (K_off)) were calculated for the first five min of decreasing fluorescence using the “one phase decay-equation” in GraphPad Prism 9 (GraphPad Prism, La Jolla, CA, USA). Individual experiments were performed for each different substrate combination.

High resolution respirometry.

The impact of Ca²⁺ titrations on mitochondrial respiration was measured using an Oroboros O2k high-resolution respirometer at physiological temperatures in an EGTA free medium (in mM: D-sucrose 110, lactobionic acid 60, taurine 20, KH₂PO₄ 10, MgCl₂ 2, HEPES 20, and BSA 0.015; pH 7.2), containing 1 mg/ml brain tissue (100 mg wet weight/ml homogenized in ice-cold Tris solution), with 10 mM glutamate, 5 mM malate, 50 μg/ml saponin, and titrations of CaCl₂ (0, 0.2, 10, 20, 50, 100, and 150 μM in separate experiments). Tissues were incubated for ~10 min in the Oroboros chamber and then 1 mM ADP, 10 μM cytochrome c, 10 mM succinate, and CCCP (in 0.5 μM steps until reaching the maximal response) were added sequentially. O₂ consumption rates (i.e., the O₂ slope in Oroboros) were analyzed by selecting the mean rate during the first two mins following each substrate injection.

Mitochondrial membrane potential.

The impact of Ca²⁺ titrations on mitochondrial membrane potential was measured using an Oroboros O2k fluo module at physiological temperatures in an EGTA free medium (in mM: KCl 120, HEPES 20, KH₂PO₄ 20, MgCl₂ 2.5, NaCl 5, and BSA 0.03; pH 7.2). 1 mg/ml brain tissue (100 mg wet weight/ml homogenized in ice-cold Tris solution (in mM: KCl 105, Tris-HCl 50, pH 7.1), with 50 μg/ml saponin, and 2 μM Rhodamine 123) was incubated for ~ 10 min in the Oroboros chamber, and then 10 mM glutamate, 5 mM malate, and serial titrations of CaCl₂ (5, 5, 10, 30, 50, and 50 μM consecutively, alternatively expressed as 0, 5, 10, 20, 50, 100 and 150 μM, respectively) and CCCP (in 0.5 μM steps until reaching the maximal response) was added sequentially.

Transmission electron microscopy (TEM).

Following treatment, mice or NMRs were perfusion fixed first with a Ringer’s solution at 37°C followed by primary fixative consisting of 2% paraformaldehyde + 2.5% glutaraldehyde in 0.15 M sodium cacodylate buffer (pH 7.4) also at 37°C. The brain was rapidly removed from the skull and placed in ice-cold primary fixative for 1 h. Using a Leica vibratome with a trough containing ice-cold 0.15 M cacodylate buffer, brain was sliced into 100 micron-thick coronal slices and those containing the hippocampus were collected. The slices were rinsed in 0.15 M cacodylate buffer and incubated in a mixture of 1% OsO₄, 0.8% potassium ferrocyanide, 3 μM calcium chloride in sodium cacodylate buffer for 1 h on ice. Then, the slices were washed 3×3 min with ice-cold double-distilled water (ddH₂O) and stained with 2% uranyl acetate for 1 h on ice. The slices were next incubated in increasing ethanol solutions: 20%, 50%, 70%, 90% on ice, followed by 3 × 100% at RT, each for 10 min. Subsequently, the slices were infiltrated with a mixture of 50% dry acetone and 50% Durcupan ACM resin (Fluka) for 2 h with agitation, and then incubated 3×12 h in 100% Durcupan with agitation. Durcupan ACM resin was made by mixing 11.4 g component A, 10 g component B, 0.3 g component C and 0.1 g component D. The slices were sandwiched between 2 Mould-release slides and polymerized for 48 h at 60°C in an oven. The glass slides were removed, and a hacksaw was used to cut out a block about 2 mm across and glued on a dummy block. Thin sections about 70 nm thick were cut using a Leica UCT ultramicrotome. The sections were placed on 200-mesh uncoated thin-bar copper grids. A Tecnai Spirit (FEI; Hillsboro, OR) electron microscope operated at 80 kV was used to record images with a Gatan 2Kx2K CCD camera at 2.9 nm/pixel.

For quantitative analysis, the mitochondrial profile area was measured with ImageJ (NIH). The number of mitochondria per unit cytoplasmic area was calculated by counting the number of mitochondria in an image and dividing by the profile area of the cytoplasm measured using ImageJ. The mitochondrial volume fraction, defined as the volume occupied by mitochondria divided by the volume occupied by the cytoplasm, and the crista density, defined as the cristae surface area per mitochondrial volume, were estimated using stereology with the “grid” plug-in of ImageJ.

EM Tomography.

Semi-thick sections of ~ 350 nm thickness were cut from the blocks of tissues prepared for TEM with a Leica ultramicrotome and placed on 200-mesh uncoated thin-bar copper grids. 20-nm colloidal gold particles were deposited on each side of the grid to serve as fiducial cues. A grid was placed in a Tecnai HiBase Titan (FEI; Hillsboro, OR) electron microscope operated at 300 kV. The grid was irradiated with electrons for about 20 min to limit anisotropic specimen thinning during image collection at the magnification used to collect the tilt series before initiating a dual-axis tilt series. During data collection, the illumination was held to near parallel beam conditions and the beam intensity was kept constant. Tilt series were captured using SerialEM software (University of Colorado, Boulder, CO) at 0.81 nm/pixel. Images were recorded with a Gatan 4Kx4K CCD camera. Each dual-axis tilt series consisted of first collecting 121 images taken at 1-degree increments over a range of −60 to +60 degrees followed by rotating the grid 90 degrees and collecting another 121 images with the same tilt increment. After collecting the orthogonal tilt series, to improve the signal-to-noise ratio, 2x binning was performed on each image by averaging a 2×2 x-y pixel box into 1 pixel using the newstack command in IMOD (University of Colorado, Boulder, CO). The IMOD package (https://en.wikipedia.org/wiki/IMOD_(software))) was used for tilt-series alignment, reconstruction, and volume segmentation. R-weighted back projection was used to generate the reconstructions. Volume segmentation of mitochondrial membranes was performed using IMOD by tracing in each of the 1.62 nm-thick x-y planes in which the object appeared, then creating stacks of contours with the Drawing Tools plug-in in IMOD. The traced contours were then surface-rendered by turning contours into meshes to generate a 3D model. The surface-rendered volumes were visualized using 3DMOD. Lengths, surface areas, and volumes were extracted using the program imodinfo. The number of crista junctions per mitochondrial slice was measured by simply counting the number of crista junctions per mitochondrion observed in a 1.6-nm thick central slice through an EM tomography volume. Because NMR brain mitochondria were larger than mouse brain mitochondria, to normalize the crista junction parameter, we divided the number of crista junctions by the mitochondrial outer membrane length (ImageJ), because the junctions are next to the outer membrane. This normalization removed the mitochondrial size variability between NMR and mouse.

Statistical analysis.

Significant differences between groups were determined using one- or two-way ANOVAs followed by a Dunnett, Tukey, or Sidak multiple comparisons test, as appropriate, or unpaired two-tailed t-tests. Values are reported as mean ± SD. All statistical analyses were performed using GraphPad Prism 9 (GraphPad Prism, La Jolla, CA, USA), with a significance level of p < 0.05.

Results

NMR brain mitochondria have higher Ca²⁺ affinity and retention capacity.

To compare the uptake rate and retention capacity of Ca²⁺ by brain mitochondria from NMRs and mice, boluses of CaCl₂ were added in serial titrations to equal amounts of permeabilized brain tissue from each species (Fig. 1A). Mitochondrial Ca²⁺ uptake was monitored by measuring the decrease of Ca²⁺ in the solution using a light-excitable Ca²⁺ indicator, Calcium green-5N (Rajdev & Reynolds, 1993), such that each Ca²⁺ bolus was observed as an increase in Calcium green-5N fluorescence in the solution due to Ca²⁺-binding. Boluses of Ca²⁺ were added until Calcium green-5N fluorescence increased, indicating activation of the MPTP and release of mitochondrial Ca²⁺, which is a hallmark of Ca²⁺ overload in hypoxic brain. Ru360, a specific antagonist of the mitochondrial calcium uniporter (MCU) that does not impact endoplasmic reticulum Ca²⁺ uptake or release (Matlib et al., 1998; Baughman et al., 2011), entirely prevented the influx of Ca²⁺ (Fig. 1A–C), which confirmed that mitochondria were the main source of intracellular Ca²⁺ uptake in both species and in the subsequent overload of Ca²⁺ that triggered the release of mitochondrial matrix Ca²⁺. Note, ATP did not have any significant effect on Ca²⁺ uptake in NMR brain but caused a small release of Ca²⁺ in mouse brain (Figs. 1B&C, S1).

Figure 1. Ca²⁺ uptake by permeabilized cortical brain tissues.

(A) Fluorescent sensorgram of Ca²⁺ uptake by permeabilized cortical brain tissues from mice (pink traces), NMRs (purple traces), and Ru360-treated tissue either in mouse or NMRs (black traces). Black arrows indicate one 50 μM Ca²⁺ injection followed by multiple 20 μM Ca²⁺ injections in Ru360-untreated tissue. The “End point” was determined as the last dose that did not trigger an increase in the Ca²⁺ signal. Inset: magnified image of the end point period for both species. Colour-matched dashed lines are included to demonstrate the increasing data signal after the “end-point” bolus addition. Increasing fluorescence indicates a net Ca²⁺ release by mitochondria. (B&C) Fluorescent traces following 50 μM Ca²⁺ addition (100% normalized) of permeabilized cortical brain tissues from mice (B) and NMRs (C). Substrates are GMD+Ru360 (black traces), GMD+CCCP+Ru360 (purple traces), ATP (red traces), GM (green traces), GMD+CCCP (orange traces), and GMD (blue traces). (D) Koff of Ca²⁺-binding Calcium green-5N is proportional to the [Ca²⁺] that remains in the external solution calculated from (B&C). Data are presented as mean ± SD from n = 4 independent biological replicates. Significant differences were determined using two-way ANOVAs followed by a Dunnett multiple comparisons test (* vs. GMD+Ru360; GMD, p = 0.0002; GMD+CCCP, p = 0.0475, in mice; GM, p = 0.0001; GMD, p < 0.0001; GMD+CCCP, p < 0.0001, in NMRs), or multiple unpaired t-tests with False Discovery Rate († mice vs. NMRs; GM, p = 0.0008; GMD, p = 0.0032; GMD+CCCP, p = 0.0004;). Abbreviations: G, glutamate; M, malate; D, ADP; CCCP, carbonyl cyanide m-chlorophenyl hydrazine; Koff, Ca²⁺-binding Calcium green 5-N decay rate.

We observed that permeabilized NMR brain tissue has a faster rate of Ca²⁺ uptake than mouse brain during the first five min following the initial bolus of 50 μM Ca²⁺, with ADP (> 2-fold), without ADP (~ 10-fold), or following uncoupling of the mitochondrial proton gradient with carbonyl cyanide 3-chlorophenylhydrazone (CCCP; > 3-fold; Fig. 1B&C, F_1,28 = 102.8, p < 0.0001; Fig. 1D). ADP significantly increased the rate of Ca²⁺ uptake in both NMR and mouse brain (Fig. 1B–D). CCCP abolished the ability of mouse brain to take up Ca²⁺ within five min (Fig. 1B), but only slightly inhibited the rate of Ca²⁺ uptake (Fig. 1C&D) and Ca²⁺ retention capacity (Table 1) in NMR brain. Taken together, these data demonstrate that NMR brain mitochondria have a higher affinity and retention capacity for Ca²⁺ than mouse brain (Table 1).

Table 1.

Ca²⁺ resistance of permeabilized brain tissue.

Substrates	Max Ca2+ resistance (μM/mg/ml wet tissue)
	NMRs	Mice
GM	210	110
GMD	210	110
GMD+CCCP	150	< 50
GMD+RU360	no observed Ca2+ uptake
GMD+CCCP+Ru360

Values were determined as the amount of CaCl₂ given before triggering Ca²⁺ release. Data are presented as means, from n = 4 independent biological replicates. Abbreviations: G, glutamate; M, malate; D, ADP; CCCP, carbonyl cyanide m-chlorophenyl hydrazine.

OXPHOS is more strongly inhibited by extramitochondrial Ca²⁺ in mouse brain than NMR brain.

To assess the effects of exogenous Ca²⁺ on mitochondrial OXPHOS capacity, we next exposed permeabilized brain tissues to step-wise increases in Ca²⁺ concentration and then measured glutamate/malate-stimulated mitochondrial oxygen consumption. We report that Ca²⁺ generally reduces OXPHOS capacity in a dose-dependent manner in both NMR and mouse brain (Fig. 2B; F_6,105 = 145.1, p < 0.0001; Fig. 2A & F_6,105 = 171.4 p <0.0001). Additionally, significant increases of leak respiration were triggered at > 50 μM Ca²⁺ in NMR brain (from 6.15 ± 0.28 to 21.39 ± 2.17, p = 0.0006 to 20.14 ± 3.39, p = 0.0017) but not in mouse brain. Finally, respiratory control ratios (RCRs), calculated as the ratio of OXPHOS capacity/leak respiration, were reduced by exogenous Ca²⁺ in a dose-dependent fashion in both species (Fig. 2C; F_1,42 = 574.0, p < 0.0001), but this effect was delayed and reduced in NMR brain, indicating better retention of respiratory capacity with greater Ca²⁺ challenges.

Figure 2. Ca²⁺ effects mitochondrial bioenergetics of permeabilized cortical brain tissues.

(A&B) Oxygen consumption by permeabilized cortical brain tissues exposed to incremental increases in external Ca²⁺ from mice (A) and NMRs (B). Substrates were added sequentially in the order indicated on the x-axis. (C) Respiration control ratios (RCRs) of permeabilized cortical brain tissue. (D) Oxygen consumption in the GM-fueled OXPHOS state of permeabilized cortical brain tissue normalized to 0 μM Ca²⁺. (E) IC₅₀ of Ca²⁺ for inhibiting OXPHOS. (F) Effect of cytochrome c on mitochondrial respiration. (G) Phosphorylation system control ratio of complexes I-II (1-P/E %). Data are presented as mean ± SD from n = 4 independent biological replicates. Significant differences were determined using two-way ANOVAs followed by a Dunnett (* vs. 0 μM Ca²⁺, (A) 0.2 μM, D, p = 0.0326; Cytc, p = 0.0394; 10 μM D, p = 0.0003; Cytc, p = 0.0015; S, p = 0.0001; CCCP, p = 0.0002; 20 to 150 μM, D, p < 0.0001; Cytc, p <0.0001; S, p < 0.0001; CCCP, p < 0.0001, in mice; (B) 0.2 μM, D, p = 0.0267; S, p = 0.0271; CCCP, p = 0.0298; 50 μM, D, p < 0.0001; Cytc, p < 0.0001; S, p = 0.0002; CCCP, p = 0.0016; 100 & 150 μM, D, p < 0.0001; Cytc, p <0.0001; S, p < 0.0001; CCCP, p < 0.0001, in NMRs; (C) 10 μM, p = 0.0002; 20 to 150 μM, p < 0.0001, in mice; 10 to 150 μM, p < 0.0001, in NMRs; vs. 0 μM Ca²⁺; (E) 50 to 150 μM, p < 0.0001, in mice; 100 μM, p = 0.0075; 150 μM, p = 0.0013, in NMRs; vs. 0 μM Ca²⁺; (G) 50 μM, p = 0.0150;150 μM, p = 0.0014, in NMRs), or two-way ANOVAs followed by a Sidak (†, mice vs. NMRs, (C) 0 to 20 μM Ca²⁺, p < 0.0001; 50 μM Ca²⁺, p = 0.0475; (D) 0.2 μM Ca²⁺, p = 0.0002; 10 μM Ca²⁺, p = 0.0015; 20 μM Ca²⁺, p = 0.0004; 50 μM Ca²⁺, p = 0.0037; (F) 20 μM Ca²⁺, p = 0.0471; 50 μM Ca²⁺, p < 0.0001; 100 μM Ca²⁺, p = 0.0029; 150 μM Ca²⁺, p = 0.0120; (G) 0 μM Ca²⁺, p < 0.0001; 0.2 μM Ca²⁺, p = 0.0006; 10 to 50 μM Ca²⁺, p < 0.0001), or unpaired t-test with two-tailed calculations (for E). Abbreviations: G, glutamate; M, malate; D, Cytc, Cytochrome C; ADP; S, succinate; CCCP, carbonyl cyanide m-chlorophenyl hydrazine.

Importantly, the inhibitory effects of Ca²⁺ on OXPHOS were greater in mouse brain than NMR brain. Specifically, OXPHOS capacity was significantly inhibited in all Ca²⁺ concentrations in mouse brain but NMR brain was not impacted by [Ca²⁺] < 50 μM (Fig. 2A&B). Moreover, the relative reductions of OXPHOS capacity (values normalized to 100% at 0 μM Ca²⁺) in mouse brain were significantly higher than in NMR brain (Fig. 2D; F_1,42 = 98.30, p < 0.0001). Correspondingly, we calculated the half-maximal inhibitory concentration (IC₅₀, μM) at which Ca²⁺ inhibits OXPHOS capacity, which was 34.13 ± 4.61 in mouse brain versus 81.65 ± 6.73 in NMR brain (Fig. 2E; p = 0.286).

The effect of Ca²⁺ on mitochondrial membrane permeability and the phosphorylation system control ratio was evaluated by the cytochrome c membrane integrity test and the ratio of OXPHOS to ETS capacity (1-P/E %), separately. The effect of cytochrome c became significant at a lower [Ca²⁺] in mouse brain (50 μM) than in NMR brain (100 μM), and overall values were higher in mouse brain than NMR brain (Fig. 2F; F_{1, 6} = 75.11, p = 0.0001). Interestingly, NMR brain was elastic with regard to its Ca²⁺-induced phosphorylation system control ratio, which increased with addition of Ca²⁺ until 50 μM (from 11.20 ± 2.73 to 17.44 ± 1.78%, p = 0.015), and then decreased at 100 μM and 150 μM Ca²⁺ (Fig. 2G; 7.33 ± 0.68, p = 0.2296 and 3.30 ± 0.71%, p = 0.0014). On the other hand, mouse brain did not exhibit any significant change these parameters in any [Ca²⁺], and the values were generally lower than in NMR brain (F_{1, 42} = 160.1, p < 0.0001), except with 150 μM Ca²⁺ (Fig. 2G; mice at 1.44 ± 0.31 vs NMRs at 3.30 ± 0.71, p = 0.9538).

Mitochondrial membrane potential is hyperpolarized in NMR brain and drives Ca²⁺ buffering.

The proton gradient generated across the inner mitochondrial membrane (i.e., Δψ_m), was measured in permeabilized NMR and mouse brain tissue using fluorescence quenching of Rhodamine-123 (Rh-123) (Emaus et al., 1986). The impact of Ca²⁺ on Δψ_m was assessed by serial additions of CaCl₂ (Fig. 3A). Glutamate/malate generated a greater signal in NMR brain than mouse brain without CaCl₂ (mice at 0.72 ± 0.06 vs NMRs at 1.08 ± 0.05; Fig. 3B; p < 0.0001), indicating a more hyperpolarized Δψ_m in the former (see below). Serial Ca²⁺ titrations reduced Δψ_m in both NMR and mouse brain, but a greater change was observed in NMR brain than mouse at each Ca²⁺ concentration (Fig. 3B). To compare the extent of this Ca²⁺-mediated depolarization in NMR vs. mouse brain mitochondria, we normalized the Rh-123 signal to 100% at 0 μM Ca²⁺ (Fig. 3C). In this analysis, we calculated that a greater proportion of Δψ_m was retained in NMR than mouse brain following each Ca²⁺ addition (Fig. 3C). Additionally, we calculated the half-maximal inhibitory concentration (IC₅₀, μM) of Ca²⁺ on Δψ_m, which was 8.25 ± 0.20 in mouse brain versus 36.11 ± 0.38 in NMR brain (Fig. 3D; p = 0.028).

Figure 3. Ca²⁺ effects mitochondrial membrane potential (Δψ_m) of permeabilized cortical brain tissues.

(A) Sensorgram of Rhodamine 123 (Rh-123) fluorescence during Ca²⁺ titration of permeabilized cortical brain tissues. (B) Dose-dependent Ca²⁺ effects on Δψ_m. (C) Dose-dependent Ca²⁺ effects on Δψ_m normalized to 0 μM Ca²⁺. (D) IC₅₀ of Ca²⁺ for decreasing Δψ_m. Data are presented as mean ± SD from n = 4 independent biological replicates. Significant differences were determined using unpaired t-test with two-tailed calculations (p = 0.0286 mice vs. NMRs).

NMR brain mitochondria are larger, occupy more of the cytoplasmic volume, and have increased crista density relative to mouse brain.

Mitochondrial structure is associated with Ca²⁺ sequestration capacity (Hoitzing et al., 2015; Favaro et al., 2019); therefore, we compared the ultrastructure of NMR and mouse brain mitochondria using TEM and EM tomography. NMR mitochondria were distinct from mouse mitochondria in the neuronal somas in several structural parameters (Fig. 4). Generally, NMR brain mitochondria were larger (F_{1, 196} = 51.68, p < 0.0001; 0.35 ± 0.03 vs. 0.22 ± 0.02 μm², p = 0.0180, Fig. 4A–E), and also had more densely packed cristae (Fig. 4H; F_{1, 76} = 18.85, p < 0.0001; 13.00 ± 1.30 vs. 8.70 ± 1.40 μm²/μm³, p = 0.0246,), and more crista junctions per mitochondrion (Fig. 4I&J; 9.00 ± 0.69 vs. 2.47 ± 1.19 per mitochondrion, p < 0.0005,). There was no significant difference in the number of mitochondria per unit cytoplasmic area (F_{1, 36} = 0.4754, p < 0.4949; Fig. 4F); however, the mitochondrial volume fraction (also called volume density) was larger in NMR brain than mouse brain (Fig. 4G; F_{1, 36} = 9.972, p < 0.0.0032; 17.00 ± 3.40 vs. 9.90 ± 1.20 %, p = 0.0239). Tethering and zippering of adjacent mitochondria were commonly observed in NMR brain but not in mouse brain (Fig. 4A–D). Interestingly, mouse cristae were more often networked, and sometimes highly so.

Figure 4. Naked mole-rat (NMR) mitochondria are larger, occupy more of the cytoplasmic volume and have greater crista density than mouse mitochondria in neuronal somas.

(A) Low-magnification TEM images showing parts of 2 NMR neuronal cell somas. The nuclei are marked with an “N”. (B) Low-magnification TEM of a mouse neuronal cell soma in the hippocampus with smaller mitochondria. (C) A central slice through an EM tomography volume of NMR hippocampal neuronal soma showing large mitochondria in close proximity. Tethering that connects adjacent mitochondria is observable in NMRs (inset), but less so in mouse. Inside the inset red box are three clear tethers. More common in NMR are “zippered” mitochondria, with outer membranes so close together that there does not appear to be any space between them, hence the zippered look (three red boxes). However, no intermembrane junctions with cristae aligned across adjacent mitochondria were observed. (D) In contrast, EM tomography of mouse hippocampal neuronal soma showed mostly elongated, less densely packed mitochondria. Crista density was also lower in mouse and, in contrast to NMR cristae, mouse cristae were often highly networked. (E-J) Summaries of (E) mitochondrial profile area (size) measured from TEM images. Mouse mitochondria were significantly smaller than NMR mitochondria (n = 50 mitochondria from 2 animals), (F) number of mitochondria per cytoplasmic area (n = 10 TEM images from 2 animals), (G) mitochondrial volume fraction (n = 10 TEM images from 2 animals), (H) crista density (cristae surface area per mitochondrial volume (n = 20 mitochondria from 2 animals), (I) number of crista junctions per mitochondria, and (J) number of crista junctions per mitochondrial outer membrane length (n = 20 mitochondria from 2 animals). Data are presented as means ± SD. Significant differences (p <0.05) were determined using a two-way ANOVA with Tukey multiple comparisons test; p values shown above the bars.

Discussion

Mitochondria are primary regulators of cellular bioenergetics and are a hub for multiple aspects of cellular signalling, including by one of the predominant cellular messengers: Ca²⁺ (Szabadkai & Duchen, 2008; Giorgi et al., 2018; Anderson et al., 2019). However, Ca²⁺ signalling is a double-edged sword as Ca²⁺ overload induces neuronal cell death in hypoxia and ischemia. In the present study we explore Ca²⁺ handling in a hypoxia-tolerant mammal brain and reveal novel adaptations at the mitochondrial level that contribute to the tolerance of NMR brain to Ca²⁺ stress. Specifically, we report that increasing external Ca²⁺ induces mPTP opening and the release of mitochondrial matrix Ca²⁺ in both NMR and mouse brain. However, NMR brain mitochondria can take up and retain more Ca²⁺, and better preserves bioenergetic and ETS function following in vitro Ca²⁺ challenges. These capabilities are likely mediated by ultrastructural differences between these species that precondition NMR brain mitochondria to better tolerate severe Ca²⁺ challenges and avoid the activation of downstream cell death pathways by physically storing Ca²⁺ in larger organelles. The mechanisms reported in this study may explain the slower accumulation of cytosolic Ca²⁺ in hypoxic NMR brain in vitro (Peterson et al., 2012), and likely contribute to hypoxia/ischemia tolerance in NMR brain in vitro and in vivo (Nathaniel et al., 2009; Pamenter et al., 2018; Cheng & Pamenter, 2021).

Mitochondrial uptake is the primary mediator of Ca²⁺-buffering in NMR brain.

Mitochondria play major roles both as regulators of [Ca²⁺]_c and mediators of cellular Ca²⁺ signalling (Giorgi et al., 2018). Mitochondria can accumulate large amounts of Ca²⁺, and thereby play an essential role in maintaining cellular Ca²⁺ homeostasis, particularly when extracellular Ca²⁺ influx into the cytosolic space is increased during acute hypoxia or ischemia (Shen & Jennings, 1972). In normal conditions, the rapid mode of uptake (RaM), mitochondrial ryanodine receptor (mRyR), and MCU work in concert to mediate mitochondrial Ca²⁺ uptake, with the MCU being the primary conduit of Ca²⁺ uptake when RaM and mRyR are inhibited (which occurs when [Ca²⁺]_c is > 40 μM) (Ryu et al., 2010). Consistent with a primarily role for the MCU in mitochondrial buffering, we report a near complete inhibition of Ca²⁺ uptake with the MCU-antagonist Ru360 in both NMR and mouse brain (Fig. 1A–D). We also demonstrate that mPTP opening (as indicated by increasing Ca²⁺ signal after end points), which is necessary for the initiation of Ca²⁺-mediated cell death in hypoxia/ischemic brain (Kroemer et al., 2007), is triggered at a higher [Ca²⁺]_c in NMR brain mitochondria than mouse (210 vs. 110 μM, Fig. 1A, Table 1).

In addition to having a greater capacity for Ca²⁺, NMR brain mitochondria take up Ca²⁺ at a higher rate than mouse (Fig. 1B–D). The rate of mitochondrial Ca²⁺ uptake depends on multiple factors, including the energetic state of mitochondria, Δψ_m, and external Ca²⁺ concentration (Gunter & Pfeiffer, 1990). We have previously reported that NMR brain maintains [ATP] in acute in vivo hypoxia (7%, 4 h; Pamenter et al., 2019), and this likely supports rapid uptake of higher volumes of Ca²⁺ (Fig. 1B–D, Table 1). Indeed, glutamate/malate-induced OXPHOS capacity is similar between NMR and mouse brain per tissue weight (Fig. 2A&B), but NMRs have a more tightly coupled ETS (Pamenter et al., 2018), which should result in higher O₂/ATP coupled respiration rates. In addition, the more hyperpolarized Δψ_m of NMR brain also contributes to faster Ca²⁺ uptake, as we report stronger glutamate/malate induced Rh-123 quenching in NMR brain than mouse (Fig. 3B). Note that there is a linear correlation between fluorescence quenching of Rh-123 and Δψ_m (Emaus et al., 1986). Finally, NMR brain mitochondria have a greater volume than mouse (Fig. 4, and see below), which should physically support a higher Δψ_m, and provide more space in which to sequester Ca²⁺ (Gunter & Pfeiffer, 1990).

Intriguingly, the decay rate (K_off, Fig. 1D) indicates that ATP-dependent Ca²⁺ uptake is the major mechanism of Ca²⁺ uptake in mouse brain mitochondria; however, both ATP- and Δψ_m-dependent mechanisms contribute to Ca²⁺ uptake in NMR brain mitochondria. Taken together, these results indicate that NMR brain mitochondria have higher sensitivity to changes in [Ca²⁺], and greater capacity for Ca²⁺ buffering, and are also able to maintain Ca²⁺ buffering under ATP-limited situations, such as occurs in hypoxia/ischemia.

Enhanced mitochondrial Ca²⁺ buffering supports mitochondrial bioenergetics in NMR brain.

Ca²⁺ uptake by mitochondria plays an important role in regulating mitochondrial energetics (Szabadkai & Duchen, 2008; Giorgi et al., 2018; Anderson et al., 2019). Conversely, excessive mitochondrial Ca²⁺ accumulation leads to mitochondrial depolarization, mPTP opening, and release of cytochrome c due to outer mitochondrial membrane damage (Polster & Fiskum, 2004; Pandya et al., 2013; Raffaello et al., 2016). We report that exogenous Ca²⁺ addition triggers dose-dependent mitochondrial respiratory dysfunction in both NMR and mouse (Fig. 2A&B), in line with previous studies using isolated mouse brain mitochondria (Pandya et al., 2013; Hamilton et al., 2018). However, consistent with our finding that NMR brain mitochondria can take up more Ca²⁺ before activation of the mPTP (Fig. 1A & Table 1), mitochondrial OXPHOS capacity (Fig. 2C–D), outer mitochondrial membrane integrity (Fig. 2F), and Δψ_m (Fig. 3B–D) are all better maintained with the same Ca²⁺ challenges in NMR brain than mouse. Interestingly, we observed a significant increase in the phosphorylation system control ratio (1-P/E %) in NMR brain mitochondria at 50 μM Ca²⁺, which then decreased again at 150 μM Ca²⁺. This might indicate that ~50 μM Ca²⁺ is a critical value in NMR brain mitochondria for balancing proton (H⁺) and Ca²⁺ in the mitochondria matrix, which may relate to the regulation of ion carriers, including the mitochondrial H⁺/Ca²⁺ exchanger (Giorgi et al., 2018). The overall lower effect of cytochrome c and relatively steady Δψ_m demonstrates that an intact mitochondrial membrane system is retained for longer during Ca²⁺ stress in NMR brain, which would in turn participate in Ca²⁺ resistance. Similar results have been reported in NMR brain during ischemic stress (Cheng & Pamenter, 2021).

Large and zippered mitochondria, higher crista density, and more crista junctions may help NMR brain tolerate high Ca²⁺ loads.

It is well established that biological structures determine cellular function (Shenouda et al., 2011; Ramírez et al., 2017), and Ca²⁺ uptake during hypoxia is linked to mitochondrial matrix swelling in mammal brain (Halestrap et al., 1986; Safiulina et al., 2006; Anastacio et al., 2013). Ca²⁺ uptake into the mitochondrial matrix causes mitochondrial depolarization and related mitochondrial volume increases in neurons (Safiulina et al., 2006). Consistent with this, Ca²⁺ uptake discharges glutamate/malate-induced Δψ_m in both mouse and NMR brain mitochondria (Fig. 3), and more depolarized mitochondria have also been reported in hypoxic NMR brain mitochondria in vitro (Pamenter et al., 2018). Therefore, we speculate that differences of mitochondrial size (Fig. 4I) may be a fundamental contributor to the greater Ca²⁺ uptake capacity of NMR brain relative to mice (Table 1).

Indeed, more closely packed NMR brain mitochondria (Fig. 4A–H), with stronger zippering, provide enhanced physical fundamentals for mitochondrial communication or transfer/exchange of molecules among mitochondria, which may be beneficial, since disconnection of damaged mitochondria limits the cellular impact of local dysfunction (Glancy et al., 2018). For example, crista junctions control Ca²⁺ uptake through a constituent protein, optic atrophy type 1 (OPA1), and a stabilizing protein, mitochondrial Ca²⁺uptake 1 (MICU1) (reviewed by Gottschalk et al., 2022). When Ca²⁺ binds to MICU1, crista junctions can open transiently by unblocking the OPA1 cap positioned over the crista junction opening. The increased number of MICU1-regulated crista junctions in NMR brain mitochondria may help to explain how they are able to take up and retain more Ca²⁺.

Furthermore, a higher Δψ_m is observed in healthy and larger mitochondria (Twig et al., 2008), whereas swollen mitochondria have lower Δψ_m, which is due to Ca²⁺ overload (Schneider et al., 2019). Consistent with this, we report a relatively higher Δψ_m (Fig. 3B) and larger size (Fig. 4I) of NMR brain mitochondria than mouse, which suggests that the larger mitochondrial size also contributes to higher Δψ_m, which in turn induces a higher rate of Ca²⁺ uptake in NMR brain (Fig. 1D).

Conclusion.

Ca²⁺ overload is a hallmark initiator of catastrophic cell death during hypoxia in brain. Remarkably, we report that NMR brain mitochondria have robust and sensitive Ca²⁺ uptake capabilities and resist mitochondrial defects following in vitro Ca²⁺ challenges. We propose that NMR brain mitochondrial ultrastructure might support this outstanding Ca²⁺ regulation capacity, and help to maintain energetic balance, Ca²⁺ homeostasis, and membrane potential. These findings suggest a potential mechanism via which mitochondria may contribute to hypoxia tolerance in NMR brain.

Supplementary Material

Supinfo3

Supplimenaty Figure 1. Normalized Ca²⁺ uptake by permeabilized cortical brain tissues. Fluorescent sensorgram of Ca²⁺ uptake by permeabilized cortical brain tissues from mice (A & C) and NMRs (B & D) following addition of 50 μM Ca²⁺ (100% normalized). Substrates are GMD+Ru360 (black traces), GMD+CCCP+Ru360 (purple traces), ATP (red traces), GM (green traces), GMD+CCCP (orange traces), and GMD (blue traces). Data are presented as mean ± SD from n = 4 independent biological replicates.

Key points.

• Unregulated Ca²⁺ influx is a hallmark of hypoxic brain death; however, hypoxia-mediated Ca²⁺ influx into naked mole-rat brain is markedly reduced relative to mice.

• This is important because naked mole-rat brain is robustly tolerant against in vitro hypoxia, and because Ca²⁺ is a key driver of hypoxic cell death in brain.

• We show that in hypoxic naked mole-rat brain, oxidative capacity and mitochondrial membrane integrity are better preserved following exogenous Ca²⁺ stress.

• This is due to mitochondrial buffering of exogenous Ca²⁺ and is driven by a mitochondrial membrane potential-dependant mechanism.

• The unique ultrastructure of naked mole-rat brain mitochondria may support increased Ca²⁺ buffering, and thus hypoxia-tolerance, as a large physical storage space.

Abbreviations:

[Ca²⁺]_c

cytosolic Ca²⁺ concentration

[Ca²⁺]_m

mitochondrial Ca²⁺ concentration

CCCP

carbonyl cyanide 3-chlorophenylhydrazone

Δψ_m

mitochondrial membrane potential

ETS

electron transport system

MCU

mitochondrial Ca²⁺ uniporter

MPTP

mitochondrial permeability transition pore

NMR

naked mole-rat

OXPHOS

oxidative phosphorylation

TEM

transmission electron microscopy

Biography

Following the completion of a MSc in Preventative Veterinary medicine and a successful career in therapeutic antibody development in the Chinese Biotech sector, Cheng Hang recently defended his Ph.D. in Dr. Pamenter’s lab and has begun a postdoctoral fellowship at Yale University. His current research focuses on novel models and mechanisms of cancer therapy.

Data Availability:

All data will be posted to Dryad upon acceptance and are available to the reviewers upon request.