Defective Mitochondrial Dynamics Underlie Manganese-Induced Neurotoxicity

Abstract

Perturbations in mitochondrial dynamics have been observed in most neurodegenerative diseases. Here, we focus on manganese (Mn)-induced Parkinsonism-like neurodegeneration, a disorder associated with the preferential of Mn in the basal ganglia where the mitochondria are considered an early target. Despite the extensive characterization of the clinical presentation of manganism, the mechanism by which Mn mediated mitochondrial toxicity is unclear. In this study we hypothesized whether Mn exposure alters mitochondrial activity, including axonal transport of mitochondria and mitochondrial dynamics, morphology, and network. Using primary neuron cultures exposed to 100 μM Mn (which is considered the threshold of Mn toxicity in vitro) and intraperitoneal injections of MnCl₂ (25mg/kg) in rat, we observed that Mn increased mitochondrial fission mediated by phosphorylation of dynamin-related protein-1 at serine 616 (p-s616-DRP1) and decreased mitochondrial fusion proteins (MFN1 and MFN2) leading to mitochondrial fragmentation, defects in mitochondrial respiratory capacity, and mitochondrial ultrastructural damage in vivo and in vitro. Furthermore, Mn exposure impaired mitochondrial trafficking by decreasing dynactin (DCTN1) and kinesin-1 (KIF5B) motor proteins and increasing destabilization of the cytoskeleton at protein and gene levels. In addition, mitochondrial communication may also be altered by Mn exposure, increasing the length of nanotunnels to reach out distal mitochondria. These findings revealed an unrecognized role of Mn in dysregulation of mitochondrial dynamics providing a potential explanation of early hallmarks of the disorder, as well as a possible common pathway with neurological disorders arising upon chronic Mn exposure.

Introduction

The image of mitochondria as immobile bean-shaped organelles that produce ATP via oxidative phosphorylation belies a far more dramatic reality [1]. Mitochondria are dynamic organelles that coordinate continuous cycles of fission (the division of a single organelle into two), fusion (the joining of two organelles into one), motility (directed movement within a cell), and mitophagy (targeted destruction via the autophagic pathway), which together are referred to as mitochondrial dynamics [2, 3]. Mitochondrial fission-fusion dynamics are coupled processes that are mutually regulated by the dynamin-related protein 1 (Drp1), as a mitochondrial constriction and scission protein, and by mitofusins 1 and 2 (Mfn1 and Mfn2) and optic atrophy 1 (OPA1), which mediate fusion [4]. Besides, neurons depend on mitochondria, which, associated with motor proteins and microtubules, deliver ATP where they are most required. Therefore, some mitochondria need to be anterogradely trafficked along the axon to be concentrated at energy-requiring sites and return via retrograde transport to the soma where they are destined for degradation [5].

Given the constant energy demand and distances over which mitochondria must be mobilized in neurons, “mitodynamics” mechanisms play a pivotal role in mitochondrial function and neuronal health. Consequently, disturbances in the fidelity of these mechanisms by stressors have emerged in recent years as a central pathophysiological phenomenon in some neurodegenerative disorders such as Parkinson’s disease (PD) or amyotrophic lateral sclerosis (ALS) [6, 7]. Among these stressors, manganese (Mn) is an essential metal with neurotoxic properties that target mitochondria via the mitochondrial calcium uniporter (MCU) with 97% of the Mn sequestrated in the mitochondrial matrix, concomitant with slow export from the brain by a NCLX-independent mechanism [8–10]. These observations are consistent with multiple in vivo studies, which corroborate Mn accumulation in the mitochondrial matrix after exposure [11–13]. Mechanistically, Mn exerts its effects, at least in part, impairing mitochondrial oxidative energy metabolism under conditions of cytotoxicity to trigger free radical production and the decrease in ATP synthesis [14]. In addition, previous studies have suggested that mitochondrial processes of fission and fusion are disrupted by Mn in various cell lines [15, 16]; however, studies have yet to elucidate the mechanisms underlying the altered mitochondrial dynamics in the brain. Thus, it is plausible to suggest that Mn-induced mitochondrial toxicity is associated with impairment in mitochondrial dynamics as an early and causal event that precedes the pathogenesis of Mn neurotoxicity; however, there are no published reports that investigate the toxic effects of Mn in other “mitodynamics” aspects, such mitochondrial structural and functional alteration and mitochondrial transport or communication.

Exposure to toxic levels of Mn, often due to high levels of Mn in drinking water or to occupational exposure, leads to an extrapyramidal disorder referred to as manganism, which is characterized by muscle weakness, limb tremor, bent posture, salivation, and whispering speech [17, 18]. In the brain, Mn toxicity exerts cognitive and memory defects, neuropsychological problems, and parkinsonism and appears to play a role in several disorders, including PD and Alzheimer’s disease (AD) [19]. Therefore, understanding the process by which stressors such as Mn mediate mitochondrial dysfunction may unravel a common pathway with neurological disorders arising upon chronic Mn exposure.

In the present study, for the first time, we characterize the mechanism of Mn cytotoxicity on impaired mitochondrial function, as it pertains not only to defects in mitochondrial energy production but also to mitochondrial dynamics (transport, distribution, morphology, and network) using primary neuron cultures and a sub-acute Mn-exposed rat model. We found that mitochondrial dynamics shifted toward fission and mitochondrial fragmentation in Mn toxicity, which in turn led to mitochondrial dysfunction. In addition, Mn induced mitochondrial transport defects and altered mitochondrial communication. Our findings uncovered novel and uncharacterized mechanisms of Mn-induced mitochondrial toxicity that can be targeted for novel therapeutic strategies to prevent cognitive and memory impairment.

Materials and Methods

Antibodies

For immunostaining and Western blotting analysis, sources of antibodies are as follows: mouse anti-ACTB/β-actin (Sigma-Aldrich, A1978), mouse anti-DRP1 (Santa Cruz Biotechnology, sc-271583), rabbit anti-p-ser616-DRP1 (Cell Signaling Technology, 3455), rabbit anti-MFN1 (Thermo Fisher Scientific, PA5–38042), mouse anti-MFN2 (Santa Cruz, sc-100560), rabbit anti-VPS35 (Novus Biologicals, 13517), mouse anti-TUBA/α tubulin (Abcam, ab7291), mouse anti-acetyl-K40-TUBA/α tubulin (Abcam, ab24610), mouse anti-TU-20/ β-tubulin (Cell Signaling Technology, 4466), rabbit anti-DCTN1/dynactin p150Glued (Cell Signaling Technology, 69399), rabbit anti-KIF5B (Abcam, ab167429), rabbit anti-Tom20 (Santa Cruz Biotechnology, sc-11415), and Alexa Fluor 488 or 555-conjugated secondary antibodies (Thermo Fisher Scientific, Alexa Fluor 488 goat anti-rabbit, A-11008; Alexa Fluor 555 goat anti-mouse, A-21422).

Primary Neuronal Culture and Mn Treatments

Primary striatal neuronal cultures were established from Sprague Dawley rat embryos at E15 (Charles River) with approval by the Institutional Animal Care and Use Committees of Albert Einstein College of Medicine. Timed pregnant dams were euthanized by carbon dioxide inhalation, and embryos were removed. Around 12 embryos from different dams were pooled and were used for the preparation of the cultures. The embryonic ventral midbrain was quickly dissected from embryos in Hanks’ balanced salt solution (HBSS; Thermo Fisher Scientific, 14025092) as described previously [20]. For dissociation, the tissue was incubated in fresh Dulbecco’s Modified Eagle Medium (DMEM; Thermo Fisher Scientific, 11995073) containing 20 U/ml papain (Worthington Biochemical, LK003150) and 0.01% DNase I for 30 min at 37°C followed by pipetting and filtering through a 70- and 40-μm nylon cell strainer (Corning, 352340). Neurons were maintained in fresh Neurobasal media (Thermo Fisher Scientific, 21103049) supplemented with 2% B-27 (Thermo Fisher Scientific, A3582801), GlutaMax (Thermo Fisher Scientific, 35050061), and 50 U/ml penicillin/streptomycin (Thermo Fisher Scientific, 15140122). After 3 days, half media was replaced with new media. Neurons were used for respective experiments on DIV8. Immunostaining confirmed that >95% of the cell population was positive for the neuron-specific marker TU-20 (Fig. 1 d).

Fig. 1.

Mn induces mitochondrial fission but not fusion in primary neuron cultures. a Experimental workflow of primary neuron culture isolated from E15 embryo rats and exposed to Mn at DIV8. b Representative Western blot of DRP1 protein level from primary neuron cultures exposed to Mn (100, 500, and 1000 μM) for 1 h and 24 h. Bars expressed as mean ± SD show the protein quantification of DRP1 in control and primary neurons exposed to 100 μM Mn for 1 h (n=6). ACTB was used as a loading control for both control and Mn-exposed primary neurons. c Representative Western blot of p-616-DRP1 in control and primary neuron cultures exposed to 100 μM Mn for 1 h. Bars expressed as mean ± SD show the fold change of p-616-DRP1/DRP1 in control and Mn-exposed primary neurons (n=5). d Representative confocal microscopy images show immunostaining of p-616-DRP1 (red) together with neuron marker TU-20 (green) in control and primary neurons exposed to 100 μM Mn for 1 h. Scale bars 50 μm. e Quantification of fluorescent intensity of p-s616-DRP1 in control and Mn-exposed primary neurons exposed to 100 μM Mn for 1 h. Bars expressed as mean ± SD show fluorescent intensity per cells (n=3 independent experiments with ≥ 50 neurons each). f Representative Western blot of VPS35, MFN1, and MFN2 proteins in control and primary neurons exposed to 100 μM Mn for 1 h. Bars expressed as mean ± SD show fold change of VPS35, MFN1, and MFN2 proteins in control and Mn-exposed primary neurons (n=5 independent experiments). g Representative confocal microscopy images of control primary neurons exposed to 100 μM Mn for 1 h. Scale bars 10 μm. h Quantification of mitochondrial morphology in control and primary neurons after Mn treatment to overt fragmented or overt intermediate-to-fused mitochondria. Bars expressed as mean ± SD show the percentage of mitochondria (n=4 independent experiments with n≥ 50 neurons each). Statistical significance was analyzed by Student’s t-test (*p < 0.05; **p < 0.01; ***p < 0.001, **** p < 0.0001) compared to control. Filam, filamentous; Interm, intermediate; Frag, fragmented; DIV, days in vitro

Besides Supplemental Fig. S1, where the doses and time periods are indicated in the figure legend, all Mn treatments were for 1 h at 100 μM (MnCl₂ 4H₂O, Sigma-Aldrich, 203734).

In Vivo Mn Administration

All experiments involving animals were performed in accordance with ethical standards conformed to the National Institutes of Health Guidelines and with approval by the Institutional Animal Care and Use Committee of Albert Einstein College of Medicine under protocol # 20171008. Rats were under 12 h light-dark cycle with ad libitum access to regular food and water. Ten-week-old Sprague Dawley male rats (n=6 per group) were randomly assigned to receive intraperitoneal injections of either saline vehicle (0.9% NaCl) or MnCl₂^.4H₂O (25 mg/kg at pH=7) every 2 days for 15 days. At day 16, rats were euthanized via isoflurane inhalation, brains were excised, and the striatum is isolated for subsequent analysis.

Immunocytochemistry

Primary neurons cultured on poly-D-lysine/Laminin (Sigma-Aldrich, 127–2.5)–coated four-well glass chambers (Thermo Fisher Scientific, 154526) were fixed with 4% paraformaldehyde (PFA, Santa Cruz Biotechnology, sc-281692) for 10 min at room temperature (RT). Following fixation, neurons were permeabilized with PBS (Thermo Fisher Scientific, 10010–023) containing 0.3% Triton (Sigma-Aldrich, X100) and blocked for 1 h in PBS with 3% bovine serum albumin (BSA, Sigma-Aldrich, A3059) for 1 h RT. Fixed cultures were incubated with primary antibodies in PBS-based buffer containing 1% BSA at 4 °C overnight. Samples were washed four times with PBS at RT for 5 min each, incubated with secondary fluorescent antibodies for 1 h at RT, re-washed with PBS, and then mounted with ProLong Diamond Antifade mounting medium (Thermo Fisher Scientific, P36962) for imaging. The images were acquired with a Leica SP8 point scanning confocal microscope (40× objective), and several images were automatically stitched together using the Leica LASX Software. At least 3 coverslips for each biological replicate per group and multiple areas per chamber selected on a random basis were used for quantification analysis. The intensity for a given cell was quantified and normalized to the number of cells using Volocity software. To make comparisons between experiments involving different batches of neurons, we used the same immunochemistry procedure and included all control and treatment groups in each preparation.

MTT Assay

The MTT assay was performed to assess cell viability in neuron culture. After treatment, neurons were washed, and 200 μl/well of culture medium containing 1 mg/mL of MTT (3-(4,5- dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide; Sigma-Aldrich, 88417) was replaced. After 4 h of incubation at 37 °C, neurons were washed and lysed with 200 μL of DMSO, and the absorbance at 570 nm and 690 nm was determined. Cell viability was normalized to that of the control group (100%). The representative data shown in this study were reproducible in four independent experiments.

Potential Measurement Mitochondrial

Mitochondrial membrane potential was measured with a cationic fluorescent dye, tetramethylrhodamine methyl ester (TMRM) (Invitrogen, I34361). After Mn exposure, neuron cells were washed and incubated with TMRM at 10 nM in culture medium for 30 min at 37 °C. Next, neurons were washed twice in PBS and analyzed on a microplate reader (BMG FLUOstar Omega) at 544 nm. Results were expressed as TMRM fluorescence intensity.

Seahorse Bioenergetic Analysis

The oxygen consumption rate (OCR) of mitochondria during respiration was measured in the assay solution using a Seahorse Bioscience XF24 extracellular flux analyzer. Primary rat neuron cultures were seeded (3×10⁴ cell/well) in XF 24-well microplates previously coated with poly-D-lysine/Laminin. Before initiation of measurements, neurons were rinsed and incubated with XF base medium (Agilent, 102353–100) supplemented with 25 mM glucose, 2 mM L-glutamine (Thermo Fisher Scientific, 25030081), and 1 mM sodium pyruvate (Thermo Fisher Scientific, 11360070), pH 7.4. After 45 min incubation in a CO₂-free incubator at 37°C, OCR was measured at the basal state after neuron permeabilization with 25 μg/ml digitonin (Sigma-Aldrich, D141) and sequential injection of 2 mM oligomycin (ATP synthase inhibitor; Sigma-Aldrich, 75351), 1.5 mM carbonyl cyanide-p trifluoromethoxyphenylhydrazone (FCCP; uncoupler; Sigma-Aldrich, C2920), and 1 mM rotenone (complex I inhibitor; Sigma-Aldrich, R8875). The readouts were used to define bioenergetic parameters as following: basal OCR = OCRbaseline – OCRRot; maximal OCR = OCRFCCP – OCR_Rot; and ATP-linked OCR = OCR_baseline – OCR_oligomycin. The experiment was repeated at least 3 times. The Wave report generator (Agilent) was used for analysis.

Mitochondrial Network in Primary Neurons

In order to monitor the mitochondrial network feature analysis, we used the mitochondria-specific dye, MitoTracker Red CMXRos (MTR) (Thermo Fisher Scientific, M7513). Primary neurons at a density of 1.2 × 10⁵ cells/well were imaged on poly-D-lysine/Laminin–coated 35-mm glass-bottom dishes (MatTek Corporation, P35G-1.5–14-C). After Mn exposure (100 μM), neurons were incubated with MTR at a final concentration of 400 nM at 37 °C for 30 min. The cells were then rinsed twice with HBSS and imaged on phenol red-free Neurobasal medium (Thermo Fisher Scientific, 12348017) supplemented with 12mM HEPES buffer at a pH of 7.4. Images were taken using a Leica SP8 point scanning confocal microscope (63× objective). Mitochondrial morphology was scored as follows: filamentous (long and spaghettilike shape), fragmented (completely dotted), and intermediate pattern (when both filamentous and fragmented mitochondria were found) [21]. The percentage of neurons with indicated mitochondrial morphologies was determined as a percentage of the total number of neurons counted. For a deep analysis of the mitochondrial network, we use the Mitochondria Network Analysis (MiNa) toolset plug-in for Fiji [22]. Briefly, MiNA defined the mitochondrial network and converted it to binary, following the conversion to a skeleton that represents the features in the original image using a wireframe of lines of one pixel wide. The parameters used in the study are (1) individuals, punctate, rods, and large/round mitochondrial structures; (2) networks, mitochondrial structures with at least a single node and three branches; (3) the mean number of branches per network; and (4) the average of length of rods/branches. Ten randomly chosen fields containing between 10 and 15 neurons were used to quantify the pattern of mitochondria.

Live Cell Time-Lapse Imaging

Primary neurons labeled with MTR were imaged in Neurobasal medium minus red phenol (Thermo Fisher Scientific, 12348017) supplemented with 12mM HEPES buffer at pH 7.4 (Sigma-Aldrich, 83264) and maintained in an environmental chamber at 37 °C/5% CO₂. Time-lapse image series were captured at a frame rate of 500 ms/frame for 15 min at 5-min intervals. Image cropping and global adjustments to brightness and contrast and time-lapse movie formatting were performed using Fiji (https://imagej.net/Fiji). To quantify mitochondrial displacement rates, we tracked the mitochondrial distance traveled (μm/s) from an initial point to a final point in a period of time using the mitochondrial particle analysis TrackMate toolset plug-in for Fiji (https://imagej.net/TrackMate#Downloadable_jars). The measure is expressed as the average of 40 mitochondrial displacement rates for each condition (control and Mn exposure).

Immunoblotting

To prepare whole-cell extracts, primary neurons were lysed in RIPA cell lysis buffer (Sigma-Aldrich, R0278) with protease inhibitor cocktail (Sigma-Aldrich, P8340) and phosphatase inhibitor cocktail (Sigma-Aldrich, P2850). The lysates were sonicated for 15 s and centrifuged at 12,000 g at 4°C for 10 min to collect whole-cell proteins in supernatant. For animal tissue samples, rat striatum was homogenized in RIPA supplemented with protease/phosphatase inhibitor cocktail, sonicated for 20 s, and centrifuged at 12,000 g at 4°C for 10 min to collect whole-cell proteins in the supernatant. For subcellular fraction samples, cytoplasmic and mitochondrial fractions from striatum were isolated using the Mitochondria Isolation Kit for Cultured Cells according to the option A of the manufacturer’s instructions (Thermo Scientific, 89874). The protein concentrations were measured using the Pierce BCA Protein Assay Kit (Thermo Fisher Scientific, 23225). Then, the protein (10 μg) was resolved on 4–12% Bis-Tris gels (Bio-Rad, 161–0375) and transferred to nitrocellulose membranes, followed by Western blots on the same membranes after stripping with Restore™ PLUS Western Blot Stripping Buffer (Thermo Scientific, 46430) for 10min at RT between each application of the antibody. For whole-cell protein sample, band densitometry was normalized to ACTB or TUBA as a loading control. For mitochondrial and cytoplasmic fractions, TOM20 and ACTB were used as loading controls, respectively. The densitometric values were quantified using ImageJ (https://imagej.nih.gov/ij/).

Electron Microscopy and Image Analysis

Neuron cell cultures were fixed with 2.5% glutaraldehyde (Sigma-Aldrich, G5882) in 0.1 M sodium cacodylate buffer (Thermo Fisher Scientific, 50980232), postfixed with 1% osmium tetroxide (Sigma-Aldrich, 75633) followed by 2% uranyl acetate (VWR International, 102092–284), and dehydrated through a graded series of ethanol. Then, cells were lifted from the monolayer with propylene oxide and embedded as a loose pellet in LX112 resin (LADD Research Industries, Burlington VT) in Eppendorf tubes. Striatum tissue was fixed with 2% paraformaldehyde plus 2.5% glutaraldehyde in 0.1 M sodium cacodylate buffer, postfixed with 1% osmium tetroxide followed by 2% uranyl acetate, dehydrated through a graded series of ethanol, and embedded in LX112 resin. Ultrathin sections (80nm) were cut on a Leica Ultracut UC7, stained with uranyl acetate followed by lead citrate, and viewed on a JEOL 1400Plus transmission electron microscope at 80kv.

For the analysis of striatal mitochondrial morphometry, ten random areas from neuropils and soma containing mitochondria were imaged at 2500× magnification and selected to analyze. Mitochondrial area (μm²), perimeter (μm), circularity (4π·surface area/perimeter²), and density (number of mitochondria per image area) were quantified using Adobe Photoshop from 4 rats in each condition. Taking into account that the space occupied by somatic mitochondria differs from each micrograph, the following formulas were used to compare computed parameters: (1) somatic mitochondrial area = total mitochondrial area/somatic area; (2) somatic mitochondrial perimeter = total mitochondrial perimeter/somatic perimeter; and (3) somatic mitochondrial density = total mitochondrial density/somatic area. Computed values were imported into Microsoft Excel and GraphPad Prism for data analysis.

Striatal Mn Levels

In order to quantify the Mn concentration in the striatum, the tissue was digested in a MARS 6 microwave digestion system (CEM, Kamp-Lintfort, Germany) using a 15 min ramp to 200 °C which was held at 200 °C for 20 min in closed vessels. Post digestion, the samples were diluted with ultrapure water to give final concentrations of 3.25% HNO₃ and 1 μg/L Rh. This solution was subjected to ICP-MS/MS (ICP-MS; Agilent 8800 ICP-QQQ, Agilent Technologies Deutschland GmbH, Böblingen, Germany) analysis with the following parameters: 1550 W plasma Rf power, Ni-cones, MicroMist nebulizer at 1.08 L Ar/min, and Scott-type spray chamber. The following mass-to-charge ratios and gas modes were used (Q1 ➔ Q2): He-mode, Mn (55 ➔ 55) and Rh (103 ➔ 103) (internal standard). Results were checked using the certified reference material ERM-BB 422 (fish muscle).

Array Tomography-Scanning Electron Microscopy (AT-SEM)

The prepared striatum samples were serially re-sectioned into 40-nm-thick ultrathin sections using Leica ARTOS 3D ultra-microtome, collected onto silicon wafers, and stained with uranyl acetate. Sections were examined in a Zeiss Supra 40 field emission scanning electron microscope (Carl Zeiss Microscopy, LLC North America), in backscatter mode using an accelerating voltage of 8.0 kv. Regions of interest were collected using ATLAS 5.0, with a pixel size of 6.0 and dwell time of 25 us. Stacks were aligned, and segmentation and 3D reconstructions were done using IMOD [23, 24].

RNA Sequencing (RNA Seq) of Striatum

Total RNA (n=3) from striatum tissue was extracted using the RNeasy Plus Mini Kit (Qiagen, Cat. 74034) according to the manufacturer’s instructions. RNA quality control was done using TapeStation Analysis Software version A.02.02 (Agilent Technologies). Samples were sequenced at Genewiz (South Plainfield, NJ) using a HiSeq4000 system with 30M reads (Illumina). Trimmed paired-end reads were mapped to the Rattus norvegicus Rnor6.0 reference genome available on ENSEMBL using the STAR aligner v.2.5.2b. Differential expression (DE) gene analysis was done using the DESeq2 package in R [25]. Genes were considered as significantly differentially expressed if absolute log₂ fold change was ≥ 1 and false discovery rate (FDR) or the p-adjusted value was below the 0.05 threshold. Heatmaps were generated using the pheatmap package [26], and PCA plots and log₂ fold change bi-plots were generated by the use of the ggplot2 package [27]. Network analysis was investigated for the significantly regulated mRNAs using the STRING Database v. 10.5 (http://www.string-db.org) applying a confidence score of 0.150 [28].

Statistical Analysis

Results are expressed as mean ± SD unless otherwise stated. The statistical analysis was performed by the Student t-test using GraphPad Prism software v 7.0. Asterisks indicated significant differences (*p < 0.05; **p < 0.01; ***p < 0.001; **** p ≤ 0.0001; nd=no differences). MTT data were statistically analyzed by one-way ANOVA (p<0.05) followed by a post hoc Tukey test to determine differences among dosages.

Results

Mn Exposure Induces Mitochondrial Fission but Not Fusion in Primary Neuron Cells

To determine whether Mn exposure affects mitochondrial dynamics, we first assessed dynamin-related protein 1 (DRP1), a cytosolic GTPase crucial regulator of mitochondrial fission, in primary neurons treated with Mn at different concentrations and different time intervals (Fig. 1a). Western blot analysis revealed that primary neurons treated with 100 μM Mn after 1 h exhibited a significant increase in the levels of DRP1 at protein level compared to controls (Fig. 1b). However, neurons exposed to Mn for 24 h exhibited a significant decrease in DRP1 in a concentration- and time-dependent manner, likely associated with the significant inhibition in cell proliferation (Fig. 1b and Supplemental Fig. S1). Since 100 μM Mn was closed to the estimated threshold of Mn toxicity (60.1–158.4 μM) [29], and since this Mn concentration did not induce cell death at 1 h (Supplemental Fig. S1), we selected 100 μM Mn 1 h exposure time in the following experiments. In this way, we ensured we will be assaying for the effects of Mn that mimic “real-world” exposure scenarios.

DRP1 activation is regulated by posttranslational modifications. For example, phosphorylation of DRP1 in the serine residue s616 can regulate its ability to promote fission and is deemed an initial step in mitochondrial dysfunction in neurodegenerative diseases [30]. Thus, next, we sought to examine DRP1 phosphorylation after Mn exposure in primary neurons. Immunoblotting of the neurons exposed to 100 μM Mn revealed a significant increase in p-ser616-DRP1 protein levels (Fig. 1c) similar to those obtained using MPP+, a neurotoxin which causes PD [31]. Concomitantly, double immunofluorescence labeling showed that the fluorescence intensity in pser616 of DRP1 is significantly increased in neurons exposed to Mn (Fig. 1d, e).

Mitochondrial fusion proteins promote mitochondrial quality control by eliminating damaged mitochondria [2]. Therefore, next, we examined the abundance of several fusion-regulating proteins. Western blot analysis showed that Vps35, MFN1, and MFN2 protein levels were significantly diminished in neurons treated with Mn (Fig. 1f). Combined, these data suggest that Mn may mediate mitochondrial dynamics in primary neurons by enhancing mitochondrial fission via DRP1 phosphorylation at ser616 and inhibiting the fusion machinery.

Changes in mitochondrial morphology are tightly regulated by the balance between fusion and fission processes [32, 33]. We hypothesized that Mn-induced mitochondrial dynamics imbalance is causally related to mitochondrial fragmentation. Toward this end, we addressed the mitochondrial network morphology in primary neurons exposed to Mn employing MitoTracker Red CMXRos (MTR) staining (Fig. 1g). Healthy neurons exhibited a filamentous and intermediate mitochondrial morphology; in contrast, the mitochondrial morphology shifted to fragmentation, characterized by a large number of small round-shaped mitochondria, when neurons were treated with Mn (Fig. 1g, h). Additionally, we characterized in detail the mitochondrial network using the Mitochondrial Network Analysis (MiNa) toolset (Supplemental Fig. S2). Neurons treated with Mn significantly increased the number of individuals (puncta and rods) and number of networks (Supplemental Fig. S2 A–C). This latter observation was related to the fragmentation of larger networks with abundant branches into many smaller networks, as seen in the decrease in number of branches per network and the length of rods/branches (Supplemental Fig. S2 D and E). Overall, these results suggest that Mn plays a key role in mitochondrial fragmentation in primary neurons.

Mitochondrial Respiration Is Decreased by Mn Exposure in Primary Neurons

Because mitochondria provide most of the energy that is necessary to maintain neuronal function, it is important to consider the possible link between changes in mitochondrial dynamics and bioenergetic failure. Indeed, mitochondrial fragmentation is often linked to mitochondrial dysfunction as this morphological state predominates during elevated stress levels and cell death [34, 35]. To determine whether the excess fragmentation is associated with changes in mitochondrial bioenergetics, we measured the impact of Mn toxicity on mitochondrial bioenergetics in primary neurons. Following Mn exposure, cellular oxygen consumption rate (OCR) was measured by XF24 extracellular flux analyzer. Mn-treated neurons exhibited significant reduction of 15% in the basal OCR and ATP synthesis-linked OCR (Fig. 2a–c). In addition, Mn induced a moderate reduction in maximal respiration capacity, as assessed by FCCP-stimulated OCR (Fig. 2d). Importantly, the decrease in respiration was not sufficient to generate a bioenergetic imbalance since 60–80% of reduction is considered the threshold to induce a bioenergetic problem [36]. Notably, a recent paper has shown a modest decrease in mitochondrial respiration after exposure of 75–100 μM Mn in various neuronal cell types, which is not sufficient to exacerbate mitochondrial metabolic function [37]. In addition, mitochondrial membrane potential was preserved despite Mn toxicity in a manner that would further support adaptive protective changes rather than direct mitochondrial membrane depolarization by toxic Mn (Fig. 2e). Combined, these results suggest that Mn-induced mitochondrial fission may accelerate mitochondrial deficiency, although other mechanisms may also be involved.

Fig. 2.

Mitochondrial respiration is decreased by Mn exposure in primary neurons. a Oxygen consumption rate (OCR) measurements in control and primary neurons exposed to Mn (100 μM, 1h) were obtained over time (min) using an extracellular flux analyzer (Seahorse Bioscience). OCR was measured at baseline and following sequential additions of 2 mM oligomycin, 1.5 mM FCCP, and 1 mM rotenone in both control and Mn-exposed primary neuron cultures. b–d Quantification of basal OCR, maximal OCR, and ATP-linked OCR in control and primary neurons exposed to Mn (100 μM, 1h). e Mitochondrial membrane potential was measured by fluorescence after labeling primary neurons with TMRE dye. Results are expressed as mean ± SD (n = 4 independent cultures). Statistical significance was analyzed by Student’s t-test (*p < 0.05;** p < 0.01; nd, no differences)

Intraperitoneal Injections of Mn-Induced DRP1 Phosphorylation Alters Mitochondrial Ultrastructure in Rat Striatum

In order to determine the effect of Mn in fission and fusion machinery in vivo, we employed a model of sub-acute Mn neurotoxicity. Sprague Dawley male rats received every 2 days intraperitoneal injections of either vehicle saline or MnCl₂·4H₂O (25 mg/kg) for 15 days (Fig. 3a). The Mn dose was based on previous studies, where the human-relevant dose of 25 mg/kg induced neurobehavioral deficits in rats that resembled early mental disorders in humans [12, 38] One day after the last injection, we collected and processed rat striatum, since it is one of the major target sites for Mn accumulation and toxicity [39]. Our first aim was to measure striatal Mn concentrations by inductively coupled plasma mass spectrometry (ICP-MS). As shown in Supplemental Fig. S3 A, the levels of striatal Mn were significantly increased in Mn-exposed rats compared to controls. Moreover, on day 15, rats exposed to Mn did not gain weight compared to controls (Supplemental Fig. S3 B), which is one of the symptoms consistent with the toxic effect of Mn [40].

Fig. 3.

Intraperitoneal administration of Mn intervenes in mitochondrial dynamics in rat striatum. a Schematic design of in vivo experiments. Rats received an intraperitoneal injection of vehicle (control) or MnCl₂ 4H₂O (at 25mg/kg) for 15 days, and the striatum was collected 24 h after the last Mn injection (day 16). b Representative Western blot of enriched mitochondrial and cytosolic fractions from striatum of control and Mn-exposed groups. Bars expressed as mean ± SD show the fold change of DRP1 from the mitochondrial fraction in the control and Mn-exposed groups (n=3). Cytoplasmic and mitochondrial fractions are normalized to ACTB and TOM20, respectively. c Representative Western blot of mitochondrial dynamic-related proteins (DRP1, p-ser616, VPS35, MFN1, and MFN2) in control and Mn-treated striatum. Bars expressed as mean ± SD show the fold change of p-ser616, VPS35, MFN1, and MFN2 proteins in control and Mn-exposed groups (n=6). Statistical significance was analyzed by Student’s t-test (*p < 0.05; **p < 0.01; **** p < 0.0001). d, e Representative transmission electron micrographs of mitochondria from striatum in control group showing typical tubular cristae and crista junctions. f Representative electron micrograph showing appearance of spaces without cristae in the mitochondrial matrix from striatum of Mn-exposed group. g Three-dimensional reconstruction of a donut-shaped mitochondrion from striatum in Mn-exposed group. Scale bars 500 nm, 200 nm, 200 nm, 300 nm, respectively

DRP1 is primarily found in the cytosol, but it translocates from the cytosol to the mitochondria in response to various cellular stimuli to form spirals around the organelle and to promote the constriction that culminates in mitochondrial scission [4]. To investigate how Mn regulates DRP1-mediated mitochondrial fission, we examined the effect of Mn on the translocation of DRP1. The cellular fractionation experiments showed a significant increase in the expression of mito-DRP1 protein levels after Mn treatment, indeed promoting the translocation of DRP1 from the cytosol to mitochondrial membrane (Fig. 3b). Similar results were obtained in rat astrocytoma cell lines exposed to Mn [15]. In addition, strengthened by the results in vitro, Mn administration led to increased DRP1 phosphorylation at s616 and a decrease in fusion protein levels in Mn-treated rat striatum (Fig. 3c). Taken together, these data suggested that Mn promotes the phosphorylation of DRP1 and causes its subsequent translocation to mitochondria, offering a mechanism of Mn-induced mitochondrial fission.

Fusion/fission perturbation by stressors gives rise to complex mitochondrial shapes. In this context, abnormal striatal mitochondrial features have been poorly defined in the context of Mn-induced neurotoxicity. Here, we characterized defects in striatal mitochondrial ultrastructure after Mn exposure using array tomography-scanning electron microscopy (AT-SEM) with 3D reconstruction. Electron micrograph revealed non-uniform arrangement of the mitochondrial cristae, with some mitochondrial regions having the normal density of cristae alternated with regions that appeared devoid of cristae in striatum exposed to Mn compared to control (Fig. 3d–f). These findings suggest that Mn exposure alters the maintenance of the mitochondrial cristae architecture, which could theoretically influence inter-mitochondrial communication and mitochondrial function. Moreover, 3D reconstructions showed donut-shaped mitochondria in Mn-treated striatum (Fig. 3g), which result from self-fusion or “splitting” of mitochondria and are considered a hallmark of high oxidative stress in Parkinson’s disease. Donut formation can be formed under conditions of moderate swelling, and the detachment of mitochondria from the cytoskeleton is necessary for its formation [41]. Combined, these data indicate that these features may be specific hallmarks of mitochondria in the pathogenesis of Mn neurotoxicity.

Alterations in Mitochondrial Morphology in the Soma and Neuropils of Striatum Underlie Mn Neurotoxicity

Since both the in vivo and in vitro data strongly indicated that Mn perturbates mitochondrial fission and fusion, we performed in vivo transmission electron microscopy (TEM) to delineate mitochondrial morphology parameters in neuropils and somas of striatal mitochondria treated or untreated to Mn. Quantitative morphological analysis revealed that compared with the control group, where mitochondria in the neuropils exhibited a broader variety of shapes ranging from ovoid to elongated structures, an increased fraction of smaller and swollen mitochondria were observed in Mn-exposed striatum (Fig. 4a), consistent with the role of Mn in promoting mitochondrial fission. Analysis of mitochondrial shape descriptors indicates that both the average mitochondrial area (control, 0.33±0.01 μm²; Mn, 0.25±0.01 μm²) and average mitochondrial perimeter (control, 2.6±0.05 μm; Mn, 2.1±0.03 μm) significantly decreased in Mn-treated striatum (Fig. 4b, c). On the other hand, the average circularity index (control, 0.6 ±0.04; Mn, 0.63±0.003) significantly increased probably as a result of a more fragmented mitochondrial network upon Mn exposure (Fig. 4d). In addition, there was about 64% of mitochondria in the Mn-treated striatum that display fragmentation phenotype (length ≤0.20 μm², circularity ≥ 0.6), however, represents only 43% of mitochondria in the control (Supplemental Fig. S4 A–C). On the other hand, the average circularity index (control, 0.6±0.04; Mn, 0.63±0.003) significantly increased probably as a result of a more fragmented mitochondrial network upon Mn exposure (control relative area, 0.10±0.003; Mn relative area, 0.07±0.007) (Fig. 4e–h and Supplemental Fig. S4 D–F). Interestingly, we did not observe significant differences in mitochondrial distribution in the neuropils or in the soma (Fig. 4I, j). These data indicate, for the first time, that Mn affects striatal mitochondrial basal length that may contribute to Mn-dependent neurotoxicity.

Fig. 4.

Mn toxicity promotes mitochondrial fission in rat striatum. a Representative electron microscopy micrographs showing mitochondrial morphology in neuropils from striatum of control and Mn-exposed groups. Scale bars 2 μm. b–d Photoshop software was used for quantifying mitochondrial area (μm²) (b), perimeter (μm) (c), and circularity index (d) in the neuropils from striatum of control and Mn-exposed groups. e Electron microscopy micrographs showing mitochondrial morphology in the soma. f–h Bar graphs show quantification of mitochondrial area (ratio) (f), mitochondrial perimeter (ratio) (g), and circularity index (h) in the soma from striatum of control and Mn-exposed groups. i, j Mitochondrial density quantification (number of mitochondria per area) in the neuropils (i) and in the soma (j) from striatum of control and Mn-exposed groups. Results are expressed as mean ± SD (n=4). Statistical significance was analyzed by Student’s t-test (*p < 0.05; **p < 0.01; **** p < 0.0001; nd, no differences). Scale bars 1 μm

Mitochondrial Transport Failure Underlies Mn-Induced Neurotoxicity

Mitochondrial transport allows for the maintenance of mitochondrial quality, which is fundamental to neuronal homeostasis and functions. Efficient regulation of mitochondrial trafficking may be protective under chronic mitochondrial stress conditions; thus, healthy mitochondria are distributed to the axon and dendrites (anterograde transport), while damaged ones return to the soma (retrograde transport) for repair or autophagic clearance [42, 43]. Since Mn-induced mitochondrial injury both in vitro and in vivo, we speculated that damaged mitochondria return to the soma after Mn exposure. Importantly, TEM images revealed that the number of mitochondria per area did not increase in the soma or decrease in the neuropils of neurons treated to Mn as we expected (Fig. 4i, j). These findings suggest that Mn may be involved in blockade of both anterograde and retrograde mitochondrial transports contributing to the accumulation of defected mitochondria in the neuropils. To confirm this hypothesis, we sought to investigate the possible mechanism by which Mn can impair mitochondrial transport in neurons.

Among the possible causes of the alteration of mitochondrial transport, cytoskeleton proteins can have a fundamental role, being the railway along which mitochondria move. Particularly, acetylation of α-tubulin on lysine 40 alters the functionality of microtubules. Loss of a-tubulin K40 acetylation is associated with microtubule breakage, whereas acetylated tubulin makes microtubules more resistant to mechanical stress [44]. Since Mn binds to tubulin with high affinity [45], we posited that Mn may have a specific role in microtubule stabilization, contributing to mitochondrial transport inhibition. Western blot analysis revealed that Mn exposure elicited a significant decrease in acetylated tubulin protein in primary neurons exposed to Mn (Fig. 5a). Further, Mn affected slightly the global microtubule architecture in primary neurons treated with Mn in contrast to control (Fig. 5b). Similarly, EM images revealed that the cytoskeleton exhibited a disorganized and misaligned distribution of their components and an increase in the frequency of filament breakage in the soma and axon of primary neurons exposed to Mn. By contrast, untreated neurons showed a more organized cytoskeletal network characterized by spaced linear arrays with longer cytoskeleton elements, suggesting that Mn may affect the cytoskeletal assembly (Fig. 5c). Combined, these findings suggest that compromised mitochondrial traffic underlies, at least in part, cytoskeletal disruption in Mn toxicity.

Fig. 5.

Mitochondrial transport failure underlies Mn-induced neurotoxicity. a Representative Western blot of acetyl-K40-TUBA in primary neurons exposed to Mn (100 μM).Bar graphs show the protein quantification analysis (n=4). TUBA was used as a loading control for both control and Mn-exposed groups. b Representative confocal microscopy images of a primary neuron immunostained for TUBA (green). Scale bars 45 μm. c Representative electron microscopy micrographs showing the distribution of the cytoskeleton in the axon (left) and in the soma (right) of control and primary neurons treated with Mn. Scale bars 1 μm (left) and 0.5 μm (right), respectively. d Representative Western blot of KIF5B and DCTN1 levels obtained from primary neurons in control and Mn-exposed groups. Bars expressed as mean ± SD show fold change of KIF5B and DCTN1 proteins in control and Mn-exposed groups (n=4). TUBA was used as a loading control for both control and Mn-exposed groups. e Representative Western blot analysis of acetyl-K40-TUBA, KIF5B, and DCTN1 in rat striatum exposed to Mn. Bars expressed as mean ± SD show the fold change of acetyl-K40-TUBA and KIF5B proteins in control and Mn-exposed groups (n=6). TUBA and ACTB were used as a loading control for both control and Mn-exposed groups. f Analysis of mitochondrial displacement rates in control and primary neurons exposed to Mn. Results are expressed as mean ± SD (n=4 independent experiments with ≥ 15 neurons each). Statistical significance was analyzed by Student’s t-test (*p < 0.05; **p < 0.01; ***p < 0.001;**** p < 0.0001)

Mitochondrial trafficking and localization throughout the cytoplasm depend on interactions with the cytoskeleton and molecular motors. Previous studies have reported that disruption in the motor proteins impairs mitochondrial transport and alters mitochondrial distribution [46]. To test whether motor proteins are involved in Mn toxicity, we focused on KIF5B, which regulates the anterograde transport of neural mitochondria, and dynactin (DCTN1), a multisubunit complex necessary for dynein activity which is the major motor-driven MT-based retrograde transport in axons [47, 48]. Western blot analysis revealed that compared to control, Mn-treated primary neurons showed a significant decrease in the dynactin protein level (Fig. 5d). Mn exposure did not alter KIF5B protein in vitro, but our results in vivo showed a significant reduction in the kinesin protein and the rest of cytoskeleton proteins analyzed (Fig. 5e and Supplemental Fig. S5). Our results suggest that Mn may disrupt mitochondrial transport in both directions along microtubules by reducing the levels of motor proteins.

To evaluate the functional outcomes of Mn-mediated tubulin deacetylation and motor protein inhibition, we measured mitochondrial transport velocities in living primary neurons. Cell live imaging showed that mitochondrial transport velocity was 4-fold slower (mean= 0.001 μm/s) in neuron cells exposed to Mn compared to controls (mean=0.004 μm/s) (Fig. 5f, Movie 1, 2). Together, these results suggest that Mn may induce defects in mitochondrial motility in neurons by impairing cytoskeletal and motor protein stability.

Whole-Transcriptome Analysis Reveals Dysregulation of Cytoskeletal Genes in the Striatum of Mn-Exposed Rats

To identify gene expression patterns specific for Mn dose-response, we performed whole transcriptome profiling by RNA sequencing (RNA-seq) of striatum subjected to vehicle saline or MnCl₂^.4H₂O (25 mg/kg). Principal component analysis (PCA) shows the separation of subsets. Along PC1, with a 76.37% of variance, control and Mn groups were clearly separated indicating distinct transcriptional profiles for each subset (Fig. 6a). Analysis of the whole transcriptome identified global signatures of 94 differentially expressed genes (DEG) associated with Mn exposure (p-adj ≤ 0.05). The full list of DEG can be found in Supplemental Table S1. Of those, 49 were significantly upregulated, and 45 genes were significantly downregulated as compared to control (Fig. 6b). Strikingly, using Gene Ontology analysis and an extensive literature review, we observed that up to 23 genes were associated with cytoskeleton remodeling in Mn toxicity (Fig. 6c). Among them, genes upregulated in Mn-treated striatum include those associated to actin disassembly (Plxnc1), cytoskeleton damage (Sfrp1), and microtubule assembly (Dcx, Kif23, Gda) [49–53]. In contrast, downregulated DEG in Mn toxicity includes those related to actin remodeling functions (Arhgap9, Fblim, Eps8) and disrupts neuronal migration (Dab2ip) [54–57]. Network analysis showed most of cytoskeleton-related proteins are interacting with up to 25 edges within 23 nodes revealing a significant protein-to-protein interaction enrichment (p= 0.00404) (Fig. 6d). These results are consistent with previous studies in which Mn-specific altered pathways at transcriptome analyses included cytoskeletal regulation [58]. Altogether, these analyses demonstrate Mn-induced toxic responses may disrupt actin-cytoskeletal architecture at a genetic level, which can compromise mitochondrial motility defects and accelerates axonal degeneration.

Fig. 6.

RNA-seq analysis reveals alterations in cytoskeletal genes in striatum treated with Mn. a Principal component analysis (PCA) corresponding to all differentially expressed genes (DEG) from the striatum exposed to Mn and control (n=3). b Volcano plot of upregulated (green) and downregulated (blue) genes in striatum exposed to Mn when compared to the control after DESeq2 analysis. The differences were considered significant when adjusted p value < 0.05 and absolute log₂ fold change ≥ 1. c Heatmap representation of DE cytoskeleton-related genes in Mn-treated striatum when compared to the control group. Results are expressed as z-score. d Network analysis of differentially regulated cytoskeleton genes in striatum exposed to Mn. PC, principal component

Mn Exposure Alters Network Connectivity of Mitochondria

Recent studies suggest that subcellular information can be transferred between neighboring mitochondria and whole organisms. Thus, a recent hypothesis has pointed out that mitochondria communicate “kissing” between adjacent mitochondria and nanotunneling of well-separated mitochondria in the heart muscle [59, 60]. Nanotunnels are thin double-membrane protrusions that connect the matrices of non-adjacent mitochondria which could be capable of transport proteins between connected organelles and repair of damaged mitochondria [61]. The finding that Mn exposure resulted in mitochondrial motility failure prompted us to test whether reduced mitochondrial transport compromised mitochondrial communication in Mn toxicity. TEM micrographs from 1 to 60 consecutive serial sections and 3D reconstruction revealed for the first time that striatal inter-mitochondrial communication involves physical proximity and contact between mitochondria, but also nanotunnel connections (Fig. 7a, b; Movie 3, 4). TEM serial sections showed that adjacent mitochondria align mitochondrial cristae which may help to enhance the transmission of the electrochemical gradient or electrochemical coupling, hence mitochondrial communication [62]. Strikingly, mitochondrial cristae alignment still occurs in rats exposed to Mn despite the loss of cristae organization and density (Fig. 7a). Moreover, TEM images showed tubular nanotunnel connections between mitochondria in control and Mn-exposed striatum analogous to the mitochondrial nanotunnels reported previously in the mouse brain cortex [63] (Fig. 7b). Interestingly, in healthy striatum, most of the mitochondria were interconnected by tubular shorter nanotunnels with an average length of 719±102 nm compared to longer and slightly deformed nanotunnels with a length of 1287±185 nm observed between swollen mitochondria from Mn-treated striatum (Fig. 7c, d). It is noteworthy that we did not observe any change in the average diameter of nanotunnels in Mn toxicity with respect to control (Fig. 7e). Moreover, the number of elongated interconnected mitochondria was higher in control striatum (up to seven). By contrast, in Mn-treated striatum, we observed a shorter mitochondria cord containing 2 or 3 mitochondria connected, with longer and more exaggerated nanotunnels or intimate contacts (Fig. 7f, Movie 3, 4). These observations suggest that Mn may modify continuous intermitochondrial communication networks that share content and transfer signals within the striatum, although further studies are needed to elucidate this statement.

Fig. 7.

Mn alters mitochondrial communication in rat striatum. a Representative electron microscopy micrographs of mitochondrial contact sites in adjacent mitochondria illustrating cristae alignment. Scale bars 300 nm. b Z-stack at EM resolution from serial block face scanning electron microscopy (AT-SEM) of mitochondrial nanotunnel formation from the striatum in control and Mn-exposed groups. Scale bars 300 nm. c Three-dimensional reconstructions of mitochondria connected via nanotunnels in the striatum of control and Mn-exposed groups. Scale bars 300 nm. d, e Measurement of nanotunnel length (nm) (d) and diameter (nm) (e) in control and Mn-exposed groups assessed by IMOD image analysis software. Results are expressed as mean ± SD (n=3 with ≥ 10 nanotunnels each). Statistical significance was analyzed by Student’s t-test (*p < 0.05; nd, no differences). f Image stack from AT-SEM of mitochondria (colored in pink) in neuropils from striatum of control and Mn-exposed groups. Scale bars 800 nm

Discussion

Abnormalities in mitochondrial dynamics have been observed in most major neurological disorders, including PD and AD; however, it remains unclear how stressors such as Mn can contribute to mitochondrial dysfunction in manganism-associated neuromotor and neurocognitive deficits. Thus, the present study represents a first approach to characterize alterations of mitochondrial dynamics in the context of Mn toxicity. Our results revealed that Mn caused an imbalance in mitochondrial fission-fusion dynamics and defects in mitochondrial respiratory capacity. Strikingly, Mn exposure impaired mitochondrial trafficking through a decrease in motor proteins and destabilization of the cytoskeleton. Additionally, mitochondrial communication may be also altered by Mn exposure. Altogether, our results suggest a common pathogenic phenotype of mitochondrial dysfunction in manganism disease and some Mn-related human disorders.

In neurons, mitochondria play an essential role in ATP generation by mitochondrial respiration for maintenance neuronal survival and function [64]. Since mitochondrial respiratory deficiencies have been observed in numerous neurodegenerative disorders, we aimed to investigate mitochondrial bioenergetics in Mn toxicity. Our results showed slight defects of mitochondrial respiratory machinery in primary neurons exposed to Mn, which pointed out the fact that other mechanisms may be involved in Mn-induced mitochondrial toxicity.

Apart from energy supply, mitochondria engage in several dynamic activities to maintain neuron heath: fusion, fission, transport, and mitophagy, which together are called mitochondrial dynamics. Focusing on mitochondrial reshaping, our results show that Mn impairs mitochondrial dynamics by enhancing p-DRP1-s616 activity and decreasing fusion proteins (VPS35, MFN1, and MFN2), leading to mitochondrial fragmentation. Interestingly, Mn activates protein kinase Cδ (PKCδ), which is a substrate of p-ser616-DRP1, leading to apoptosis in dopaminergic neurons [65]. In addition, Mn-induced reduction in fusion proteins may be due to its excessive binding to metal-binding sites, but more studies are required to confirm the exact mechanism. Alternatively, it has been postulated that dysfunction in VPS35 is a risk factor in late-onset autosomal-dominant Parkinson’s disease [66]. Indeed, dysregulation of VPS35/retromer (due to deficiency or VPS35 mutation) ameliorates mitochondrial ubiquitin ligase 1 (MUL1) and, thus, leads to MFN2 degradation and mitochondrial fragmentation in dopaminergic neurons [67]. Importantly, Mn exposure reduced VPS35 in microglial cells and increases MUL1-mediated MFN2 degradation [68], findings that are consistent with our data. Although how Mn treatment decreases VPS35 abundance has yet to be determined, it appears that VSP35 is linked to the recycling of the major Mn transport protein, DMT1 [69]. VPS35 binds exclusively to the cytoplasmic tail of the −IRE species of DMT1; however, the molecular mechanism underlying this process is yet to be deciphered.

Nonetheless, findings detailed herein from both in vivo and in vitro experimental platforms establish the existence of a regulatory feedback mechanism that controls mitochondrial fission and fusion in response to Mn toxicity.

Damaged mitochondria are thought to be transported back to the soma for degradation by mitophagy [2]. Instead, we observed inhibition in the mitochondrial motility mediated by instable cytoskeleton and depletion of the motor proteins (including dynactin and kinesin-1) after Mn exposure both in vitro and in vivo. Although the molecular mechanism underlying Mn-induced motility arrest has yet to be elucidated, we showed the loss of α-tubulin K40 acetylation, which is consistent with a study in primary cortical neurons exposed to Cd [70]. Acetylation of microtubules confers protection against mechanical stress by strengthening lateral interactions between protofilaments [44]. It is feasible that Mn could accomplish modification in the enzymes that regulate tubulin acetylation (αTAT1) or tubulin deacetylation (SIRT2 and HDAC6) rendering the microtubules more vulnerable to degradation [71]. However, we cannot rule out another possibility, which is Mn binding to tubulin, as Mn may cause a steric hindrance of residues important for the accessibility of αTAT1 to the lumen of the microtubule, where the acetylation occurs [72]. Consistent with this possibility, the neurotoxin MPP+ binds to the surface of the microtubules leading to microtubular destabilization, providing an explanation for the impairment of axonal transport in tau or MPTP models of neurodegeneration [73]. Moreover, we show that Mn decreases the motor proteins dynactin and kinesin-1 responsible for the anterograde and retrograde movement of mitochondria in neurons. Previous studies have shown that molecular motors preferentially run along acetylated microtubules [74, 75]. Therefore, Mn-induced deacetylated microtubules may decrease the association between tubulin and motor proteins and, consequently, impair mitochondrial transport. Alternatively, Mn binding to tubulin may block the association of motor proteins to microtubules. When tubulin is cleaved, motor processivity or run length of both kinesin and dynein is decreased [76, 77].

Additionally, RNA-seq results revealed that Mn exposure may affect actin-cytoskeletal architecture at the gene expression level. Importantly, microtubules serve as rails for long-range transport of mitochondria, while actin cables are mainly used for their docking and short-range movements, thus ensuring neuronal homeostasis [5]. Therefore, Mn-induced changes in the regulation of the cytoskeleton may lead to mitochondria mispositioning and damage, providing an alternative explanation for the impairment of mitochondrial transport in manganism-associated neurodegeneration, consistent with the impairment of axonal transport of dopaminergic neurons in a MPTP model as a result of changes in the polarity and orientation of the cytoskeleton [73]. These genetic alterations’ response to Mn may contribute to defining novel hallmarks, identifying targets for future therapeutic strategies. Collectively, our results suggest that mitochondrial motility failure may be controlled by the affinity of Mn to tubulin or to active motor proteins, which could be a causal or consequence of modifications of the cytoskeleton at the gene level.

Indeed, Mn²⁺ has been widely used as a positive control for integrin activation due to its binding to the integrin domain metal ion-dependent adhesion site (MIDAS) [78, 79], and Mn²⁺-mediated activation of focal adhesion kinase (FAK) leads to enhance axon outgrowth in dorsal root ganglia (DRGs) neurons [80]. In theory, FAK activation should counteract the drop in microtubule stability induced by Mn²⁺ exposure. However, a recent study has shown that mitochondrial Ca²⁺ homeostasis controls actin remodeling through modulation of protease calpain [81]. Calpain inhibits integrin signaling through cleavage of FAK [82]. So, taking into account that Mn²⁺ impairs Ca²⁺ homeostasis leading to the accumulation of intramitochondrial Ca²⁺ by inhibiting its efflux [10, 83], it is feasible that calpain is also activated upon Mn exposure triggering the cleavage of FAK and thus impairing microtubule stability [84]. Indeed, several lines of evidence indicate that calpain activation is involved in acute Mn neurotoxicity [85, 86], but establishing this point requires further studies.

In this study, we provide new insight into the characterization of the alterations in mitodynamics caused by Mn-induced mitochondrial toxicity. Nevertheless, there is a major point to consider: what is the physiological relevance of perturbation in mitochondrial dynamics in the pathogenesis of Mn neurotoxicity? Besides influencing mitochondrial morphology, mitochondrial fission has been implicated in multiple functions, including the facilitation of mitochondrial transport, mitophagy, and apoptosis [34]. Thus, it is not surprising that the acceleration of mitochondrial fission could, in fact, be a positive regulator of neuronal survival [87, 88]. Therefore, the stimulation of mitochondrial fission may underlie a bioenergetic adaptation to respond to metabolic demands under Mn toxicity, but establishing this point requires further studies. In contrast, mitochondrial fusion facilitates the content exchange between mitochondria and maintains mitochondrial homeostasis [34]. In our model, Mn toxicity diminished mitochondrial fusion proteins; however, we observed long connections through 200–400-nm-wide double-membrane tubules called mitochondrial nanotunnels. Nanotunnels are bacteria-like thin double-membrane connections between mitochondria which may be capable of transporting proteins between physically constrained mitochondria [61]. To our knowledge, this is the first report of nanotunnels in rat striatum suggesting that striatal mitochondria may communicate with each other via the nanotunneling and kissing mechanism. Our results showed that the morphology of the nanotunnel becomes more exaggerated in mitochondria exposed to Mn, consistent with observed elongated interconnected mitochondria with thin and enlarged nanotunnels (aka mitochondria-on-a-string, MOAS) described in patients and models of AD [89] suggesting that this may be an adaptive phenotype in the response of Mn toxicity. Although the discovery that a selective communication between mitochondria under Mn toxicity raises a new set of questions, we speculate that longer nanotunnels arise from mitochondria under Mn toxicity that “reach out for help.” In this scenario, we propose that long nanotunnels are generated as a result of immobilized mitochondria and transport proteins, which may guarantee an efficient exchange of matrix components. In fact, under starvation conditions, it has been shown that elongated mitochondrial networks are more efficient at energy generation and capable of distributing energy through long distances resembling that achieved by mitochondrial fusion [90]. In this line, the most substantial MOAS accumulation was observed in the case of AD and hypoxia in human and mice where axonal trafficking of mitochondrial is significantly inhibited [91–93].

Finally, we cannot rule out another point: “which came first the chicken or the egg?” Although the molecular players that affect mitochondrial dynamics have been identified in several systems, the cellular signals that regulate these events are largely elusive. Interestingly, recent reports suggest that mitochondria piggyback on dynamic microtubules to selectively undergo fission when microtubules depolymerize, which means that microtubule dysfunction precedes transport impairment and mitochondrial damage [94, 95]. On the other side, mitochondrial transport in axons is suppressed by the deletion or expression of disease mutants of MFN2 [96]. Although these mechanisms need to be elucidated in Mn toxicity, our study provides a window onto the pathogenesis.

Conclusions

The present study points to a new way of thinking about Mn-related human disorders and the disruption of mitochondrial dynamics in the brain. We hypothesize that Mn can alter the balance in levels of mitochondrial fission and fusion proteins leading to fragmentation, decrease in mitochondrial respiration, and mitochondrial ultrastructural defects. Moreover, for the first time, we propose that Mn may impair mitochondrial traffic brought about by decreased motor protein levels, deacetylated tubulin, and transcriptional changes of genes related to the cytoskeleton remodeling. Lastly, Mn toxicity could result in elongated interconnected organelles. Although the underlying mechanisms of Mn-induced mitochondrial toxicity require further in-depth investigation, we hypothesize that the pathogenesis of Mn toxicity may have common phenotypes with human disorders that encompass mitochondrial dysfunction. Understanding these new mechanisms will likely have far-reaching implications for the development of novel therapeutic strategies to mitigate Mn toxicity, as well as other disorders involved in Mn-mediated neurotoxicity.

Data Availability

The RNA-sequencing dataset generated during this study is available on NCBI GEO with the accession number GSE153017. All other data are available from the corresponding authors upon reasonable request.