Mitochondrial fusion suppresses tau pathology–induced neurodegeneration and cognitive decline

Abstract

Background:

Abnormalities of mitochondrial fission and fusion, dynamic processes known to be essential for various aspects of mitochondrial function, have repeatedly been reported to be altered in Alzheimer’s disease (AD). Neurofibrillary tangles are known as a hallmark feature of AD and are commonly considered a likely cause of neurodegeneration in this devastating disease.

Objective:

To understand the pathological role of mitochondrial dynamics in the context of tauopathy.

Methods:

The widely used P301S transgenic mice of tauopathy (P301S mice) were crossed with transgenic TMFN mice with the forced expression of Mfn2 specifically in neurons to obtain double transgenic P301S/TMFN mice. Brain tissues from 11-months old non-transgenic (NTG), TMFN, P301S, and P301S/TMFN mice were analyzed by electron microscopy, confocal microscopy, immunoblot, histological staining, and immunostaining for mitochondria, tau pathology, and tau pathology-induced neurodegeneration and gliosis. The cognitive function was assessed by the Barnes maze.

Results:

P301S mice exhibited mitochondrial fragmentation and a consistent decrease in Mfn2 compared to age-matched NTG mice. When P301S mice were crossed with TMFN mice (P301S/TMFN mice), neuronal loss, as well as mitochondria fragmentation were significantly attenuated. Greatly alleviated tau hyperphosphorylation, filamentous aggregates, and thioflavin-S positive tangles were also noted in P301S/TMFN mice. Furthermore, P301S/TMFN mice showed marked suppression of neuroinflammation and improved cognitive performance in contrast to P301S mice.

Conclusion:

These in vivo findings suggest that promoted mitochondrial fusion suppresses toxic tau accumulation and associated neurodegeneration, which may protect against the progression of AD and related tauopathies.

INTRODUCTION

Alzheimer’s disease (AD) is the most prevalent form of dementia in the elderly characterized by two major pathological hallmarks: neurofibrillary tangles (NFTs) and senile plaques (SPs) [1]. NFTs are intracellular aggregates composed of the hyperphosphorylated form of the microtubule-associated protein tau (MAPT), while SPs are extracellular lesions made up of bundles of amyloid-β (Aβ) peptide fibrils [1]. NFTs have a closer correlation with the severity of cognitive impairment than SPs in AD [2]. Tau has been implicated to mediate Aβ toxicity [3]. Although no mutation in the tau gene, MAPT, has been identified in AD, mutations in MAPT account for the vast majority of frontotemporal dementia and parkinsonism linked to chromosome 17 (FTDP-17) [4], establishing the direct link between tau dysfunction and dementia. Along this line, transgenic mice expressing FTDP-associated tau are widely used animal models for AD research, as they exhibit many AD features such as tau aggregation, tau hyperphosphorylation, NFT-formation, synapse loss, inflammation, and robust neurodegeneration [5]. However, although considerable efforts have been devoted to tau-based AD drug development [6], the mechanism underlying tau toxicity remains elusive.

AD is multifactorial and involves other pathogenic mechanisms such as glutamate excitotoxicity, oxidative stress, neuroinflammation, and mitochondrial dysfunction in addition to the widely studied Aβ and tau pathology [7]. Among them, mitochondrial dysfunction has been extensively studied in the past decade. Mitofusin2 (Mfn2) is a conserved dynamin-like GTPase protein predominantly localized in the mitochondrial outer membrane regulating mitochondrial fusion [8], processes reported to be essential for various aspects of mitochondrial function including respiratory complex assembly [9], ATP production [10], Ca²⁺ homeostasis [11, 12], and ROS production [13]. Mfn2 has also been reported to be present in the endoplasmic reticulum (ER) or mitochondria-associated membranes (MAMs) to regulate ER and mitochondria tethering [14–16], autophagosome formation [17], autophagosome-lysosome fusion [18], mitophagy [19], and axonal transport of calpastatin (an endogenous specific inhibitor of the calpain system) to maintain neuromuscular synapses based on our most recent study [20]. Altered mitochondrial fission and fusion dynamics [21], distribution [22], function [23], transport [24], ER and mitochondria association [25], autophagy [26], and calpain [27], all Mfn2-related pathways, have been observed in tau transgenic animal models. To investigate the unknown pathologic role of Mfn2 in tauopathy, this study generated P301S transgenic mice with forced overexpression of Mfn2 specifically in CNS neurons and assessed the protective effect of neuronal Mfn2 against tau toxicity in vivo.

MATERIALS AND METHODS

Transgenic mice

TMFN mice were generated by pronuclear injection of the murine Thy-1.2 genomic expression cassette (gift of Dr. Philip C. Wong, Johns Hopkins University) expressing human Mfn2 into C57BL/6 fertilized eggs. P301S transgenic mice (B6;C3-Tg:Prnp-MAPT*P301S PS19Vle/J, # 008169) were obtained from the Jackson Laboratory. Two independent lines of TMFN mice were crossed with P301S mice to generate double transgenic mice (P301S/TMFN mice). 11-month-old male and female littermates raised in pathogen-free facilities at Case Western Reserve University were used in this study. All mouse procedures were performed in accordance with NIH guidelines and the Institutional Animal Care and Use Committee (IACUC) at Case Western Reserve University.

Barnes maze testing

Barnes maze was used to assess spatial learning and memory. The Barnes maze apparatus consisted of a white acrylic circular disk −120 cm in diameter with 20 equally spaced holes 5cm in diameter located 2 cm from the edge of the disk. The maze was illuminated by two 60W lamps to provide an aversive, bright disk surface. The maze was raised 30 cm from the floor and rested on a pedestal. Trials were recorded using a webcam and analyzed by automated ANY-maze video tracking software. Each trial started with a bin positioned in the center of the maze with the mouse placed inside. For testing, mice were allowed to explore the maze freely for 2 minutes. After mice entered the escape hole, they were left in the escape box for 90 seconds before being returned to home cages. If mice did not enter the escape box within 120 seconds, they were re-placed over the target hole and allowed to enter the escape box. Five sessions consisting of two trials each were run on subsequent days and escape latencies were measured. After each trial, the maze and escape box were cleaned carefully with a 10% alcohol solution to dissipate odor cues and provide a standard olfactory context. Trials were recorded using a webcam and analyzed by ANY-Maze behavioral tracking software (Stoelting Co., Wood Dale, IL).

Tissue collection

Mice were anesthetized with 2.5% Avertin in sterile phosphate-buffered saline (PBS), and transcardially perfused with ice-cold PBS at pH 7.4 for 5 minutes. The brain was isolated and separated into two hemispheres. One half was fixed with 10% neutral buffered formalin, dehydrated through a series of increasing concentrations of alcohol (e.g., 50, 70, 95, and 100%), cleared by xylene, and embedded in paraffin blocks. The other half was dissected, frozen on dry ice, and stored at −80°C.

Immunoblotting

Brain tissues were homogenized in 1 x cell lysis buffer (Cell Signaling; #9803) with 1mM PMSF (Millipore-Sigma, #329-98-6), protease inhibitor cocktail (Roche, #4693116001), and phosphatase inhibitor cocktail (Roche, #4906845001). The protein concentration was determined by the Pierce BCA Protein Assay (Thermo Fisher Scientific, #23227). Equal amounts of total protein extract were resolved by sodium dodecyl sulfate Polyacrylamide gel electrophoresis (SDS-PAGE) and transferred to Immobilon-P (Millipore, IPVH00010). Following blocking with 10% non-fat dry milk in TBST (50 mM Tris and 150 mM NaCl, 0.1% Tween20, pH=7.6), blots were incubated with primary antibody overnight in 4°C and secondary antibodies at room temperature for 1 hour and finally developed with Immobilon Western Chemiluminescent HRP Substrate (Millipore). Blots were imaged by the ChemiDoc Imaging System (BioRad, Hercules, CA) or X-ray films (Laboratory Products Sales Inc, Rochester, NY).

Immunocytochemistry, immunofluorescence, and histochemistry

6-μm-thick consecutive sections using formaldehyde-fixed paraffin-embedded mouse brain tissues were prepared, deparaffinized in xylene, rehydrated in graded ethanol, and incubated in Tris Buffered Saline (TBS, 50 mM Tris HCl and 150 mM NaCl, pH = 7.6) for 10 minutes before antigen retrieval using ImmunoRetriever Citrate (BioSB, #BSB0021). After antigen retrieval, the sections were rinsed gently in distilled water for 5 minutes, blocked with 10% normal goat serum (NGS) in 1 x TBS buffer or 1 x PBS (137mM NaCl, 2.7mM KCl, 10mM Na2HPO4, 17.6mM KH2PO4, pH= 7.4) for 30 minutes, and then incubated with primary antibodies in TBS or PBS containing 1% NGS overnight. On The second day, the sections were washed three times with TBS or PBS for 5 minutes each time, and then incubated in 10% NGS for 10 minutes. For immunocytochemistry, the sections were incubated in either goat anti-Mouse (Millipore Sigma, #AP124) or goat anti-rabbit IgG Antibody (Millipore Sigma, #AP132), followed by species-specific peroxidase anti-peroxidase complex (Jackson ImmunoResearch Laboratories Inc. #223-005-024 or #323-005-024). 3-3′-Diaminobenzidine (DAB) was used as a chromagen. The sections were dehydrated in increasing ethanol concentrations and xylene and mounted with Permount mounting medium (Fisher Scientific, #SP15-100). For immunofluorescence, the sections were incubated in Alexa Fluor-conjugated secondary antibody (Invitrogen, 1:300) for 2 hours at room temperature in the dark. Next, the sections were rinsed three times with 1 x PBS, stained with DAPI (Sigma, #D9542) for 15minutes, washed again with PBS three times for 5 minutes each time, and mounted with Fluoromount-G mounting medium (Southern Biotech, #0100-01). For Thioflavin-S staining, the sections were incubated with 500 μM of Thioflavin-S (Millipore-Sigma, #1326-12-1) for 15 minutes before mounting. H&E stain was performed using hematoxylin stain solution (RICCA, #3530-16). Taken briefly, rehydrated sections were first incubated in hematoxylin stain solution for 10 minutes followed by rinsing under running water for 5 minutes. Then the sections were differentiated briefly with 1% acid alcohol (1ml Conc HCl in 100ml ethanol) to remove excess stain, incubated in eosin solution for 30 seconds, and dehydrated with 95% and 100% ethanol. Finally, the sections were mounted with Permount mounting medium (Fisher Scientific, #SP15-100).

Optical, fluorescent, confocal, and microscope

All optical microscope images were captured at room temperature with Zeiss Imager, A2 microscope. All fluorescent microscope images were captured with a Zeiss automated microscope, Celldiscoverer 7 (controlled by Zen software, Zeiss). Confocal images were captured as serial sections along the Z-axis on Zeiss LSM710 confocal microscope with 63x/1.4NA Oil objective. For electron microscopy, hippocampal tissues were dissected and incubated in fresh EM fixative solution (quarter strength Karnovsky-1.25% DMSO mixture) for 1 hour at room temperature (RT), followed by additional fixation in fresh EM fixative for another 1 hour. After 3 times wash in 1 X PBS, tissue blocks were postfixed in 1% osmium-1.25% ferrocyanide mixture twice at room temperature for 1 hour. Then the blocks were rinsed and soaked overnight in acidified 0.5% uranyl acetate. After rinsing in distilled water, the tissue blocks were further dehydrated in ascending concentrations of ethanol, passed through propylene oxide, and embedded in Poly-Bed resin. 80nm thin sections were stained with 2% acidified uranyl acetate, and then by Sato’s triple lead staining [28]. The sections were examined in an FEI Tecnai T12 electron microscope. All chemicals used in fixing, embedding, and staining were purchased from Electron Microscopy Sciences, Hatfield, PA.

Image and statistical analysis

Image analysis, including Immunoblot analysis, cell number quantification, and mitochondria length measuring, was performed with the open-source image-analysis program ImageJ. Statistical analysis was performed by one-way ANOVA or Student’s t-test. p < 0.05 was considered to be statistically significant.

RESULTS

Mitochondrial fragmentation and Mfn2 reduction in P301S mice

Tau P301S transgenic mice, also named PS19 mice, is one of the most widely used mouse models for tauopathies [29]. Although these mice have been well characterized for tau pathology and neuronal loss, mitochondrial dynamics and their role in disease progression haven’t been extensively investigated so far. We first performed electron microscopic analyses of mitochondrial morphology in P301S transgenic mice at 11-months old, when neuronal loss and tau lesions were evident in the hippocampus [29]. Neurons of age-matched non-transgenic (NTG) mice demonstrated predominantly tubular mitochondrial morphology in cell bodies and axon initial segments (Fig. 1A, B). However, mitochondria became fragmented, and there was a significant decrease in mitochondrial length in the neurons of P301S mice (Fig. 1A–C), indicating likely altered mitochondrial fission and fusion dynamics by mutant tau. Mitochondria fission and fusion are tightly regulated by mitochondrial fission and fusion regulators, i.e., dynamin-like protein 1 (DLP1/Drp1) and its recruiting factors on mitochondria such as Mff for fission [30]; and mitofusin 1 (Mfn1), mitofusin 2 (Mfn2) and optic atrophy protein 1(OPA1) for fusion [31]. We then investigated the expression pattern of these proteins in hippocampal and cortical tissues from P301S and age-matched NTG mice. Immunoblot analyses revealed significantly reduced levels of DLP1 and Mfn2 consistently in both hippocampal and cortical tissues of P301S mice as compared to NTG mice (Fig. 1D, E and Supplementary Fig. 1). OPA1 level was decreased only in P301S cortical tissues, while other fission and fusion regulators such as Mfn1 and Mff remained unchanged (Supplementary Fig. 1). No significant change in overall mitochondrial content was noted between P301S and NTG mice as evidenced by the constant expression levels of both COX-IV and VDAC1 (Fig. 1D, E and Supplementary Fig. 1), two commonly used mitochondrial markers.

Figure 1. Inhibition of mitochondrial fragmentation in P301S mice by neuronal Mfn2.

A) Representative immunofluorescent staining of mitochondrial marker VDAC1 (red) in hippocampus neurons of 11-month-old NTG, TMFN, P301S, and P301S/TMFN mice. B, C) Representative EM images (B) and quantification (C) of mitochondria length in hippocampal neurons of 11-month-old NTG, TMFN, P301S, and P301S/TMFN mice. D, E) Representative immunoblot (D) and quantification (E) of Mfn2, and total tau (Tau46) in brain tissue lysates of 11-month-old NTG, TMFN, P301S, and P301S/TMFN mice (n = 4 mice per group). * in panel D indicates exogenously expressed human mutant tau. The data are expressed as the means ± SEM and were analyzed by one-way ANOVA or Student’s t-test. *p < 0.05; **p < 0.01; ****p < 0.0001; ns, non-significant.

Increased expression of Mfn2 in neurons inhibits mitochondrial fragmentation in P301S mice

Neurons with both DLP1 and Mfn2 reduction demonstrate predominantly fragmented mitochondria [32]. And, Mfn2 deficiency caused mitochondrial dysfunction and neuronal death, whereas DLP1 reduction had no effect [32]. On top of these, Mfn2 reduction has been consistently reported in AD patients [33] and the other tau transgenic mice expressing mutant tau P301L [21]. So, mitochondrial fragmentation in P301S mice appeared to be caused by Mfn2 reduction. We then crossed P301S mice with transgenic mice with the forced expression of Mfn2 specifically in neurons under the neuron-specific promoter thymus cell antigen 1 (Thy1), i.e., TMFN mice [32], to obtain double transgenic P301S/TMFN mice. As Mfn2 was reduced in both hippocampus and cortex of P301S mice, further immunoblot analysis of Mfn2 was only performed in the total brain homogenate. Mfn2 was confirmed to be significantly upregulated in the brains of P301S/TMFN mice as compared to age-matched P301S mice at 11-months old (Fig. 1D, E). Both P301S and P301S/TMFN mice showed similar levels of human mutant tau (Fig. 1D, E), suggesting that neuronal Mfn2 upregulation did not affect transgenic expression. To determine whether neuronal Mfn2 prevents mitochondrial fragmentation, microscopic analyses of mitochondria were also conducted. As expected, TMFN mice showed elongated mitochondria whereas P301S mice displayed highly fragmented mitochondria in brain neurons (Fig. 1A–C). Mitochondrial morphology within neurons of P301S/TMFN mice was restored to the tubular form, consistent with significantly increased mitochondrial length (Fig. 1A–C). Therefore, these results further imply that reduction of Mfn2 might underlie mitochondrial fragmentation in P301S neurons.

Increased expression of Mfn2 in neurons attenuates neurodegeneration and synaptic loss in P301S mice

P301S transgenic mice displayed neurodegeneration in the hippocampus [29]. To determine whether there was attenuated neuronal loss after neuronal Mfn2 upregulation, we performed both hematoxylin and eosin staining (H&E) and immunostaining using the antibody specific to the most widely used neuronal marker nuclear protein (​NeuN) in the hippocampus from 11-month-old P301S, P301S/TMFN, and age-matched control mice (Fig. 2A, B). The number of neurons by either H&E or NeuN staining significantly decreased in the CA1 and CA3 of the hippocampus of P301S mice, when compared with age-matched NTG or TMFN mice (Fig. 2 A–C). By contrast, no significant reduction in hippocampal neurons was observed in P301S/TMFN mice (Fig. 2 A–C), indicating a striking attenuation of neurodegeneration by neuronal Mfn2. To validate the neuroprotective effect of neuronal Mfn2 in the context of tauopathy, immunoblot analyses were further used to assess the expression of the synaptic protein PSD95. Compared with NTG or TMFN mice, P301S mice showed significantly lower levels of PSD95, while P301S/TMFN mice displayed no significant change (Fig. 2D, E), implying that synaptic degeneration in P301S mice could also be alleviated by neuronal Mfn2 upregulation. Together, these findings demonstrate the effective protection of neurons by neuronal Mfn2 in P301S mice even at a late disease stage.

Figure 2. Prevention of neuronal and synaptic loss in P301S mice by neuronal Mfn2.

A) Representative H&E staining of hippocampal neurons in 11-month-old NTG, TMFN, P301S, and P301S/TMFN mice. B, C) Representative immunocytochemistry (B) and quantification (C) of the pan-neuronal marker NeuN in the hippocampus (both CA1 and CA3 area) of 11-month-old NTG, TMFN, P301S, and P301S/TMFN mice. n = 6 mice per group. D, E) Representative immunoblots (D) and quantification (E) of the synaptic marker PSD95 in total hippocampus lysates of 11-month-old NTG, TMFN, P301S, and P301S/TMFN mice (n = 4 mice per group). The data are expressed as the means ± SEM and were analyzed by one-way ANOVA test. *p < 0.05; ns, non-significant.

Increased expression of Mfn2 in neurons reduces tau hyperphosphorylation and tau pathology in P301S mice

In addition to neurodegeneration, P301S mice also recapitulate other pathological hallmarks of tauopathy, such as tau hyperphosphorylation and tau filament formation. Next, we investigated whether neuronal Mfn2 overexpression could decrease tau hyperphosphorylation and filament formation. At 11-months old, P301S mice exhibited strong immunoreactivity with AT8 antibodies against tau protein phosphorylated at serine 202 and threonine 205, and PHF1 antibodies against tau phosphorylated at serine residues 396 and 404 in the hippocampus, whereas both NTG and TMFN mice did not show considerable AT8 or PHF1 immunoreactivity (Fig. 3A). Intriguingly, the number of AT8 or PHF1 positive neurons were greatly reduced in P301S/TMFN mice (Fig. 3A), indicating the suppression of tau hyperphosphorylation by neuronal Mfn2 upregulation. Consistently, immunoblot analyses found a remarkable reduction in the levels of hyperphosphorylated tau recognized by AT8 and PHF1 in P301S/TMFN mice (Fig. 3B, C), compared with P301S mice. Under electron microscopy, P301S mice displayed abundant filaments resembling the paired helical filaments of AD in the hippocampus (Fig. 3D), which were absent in P301S/TMFN mice (Fig. 1B), further suggesting that the formation of mature tau pathology was also inhibited by neuronal Mfn2. Sections of P301S and P301S/TMFN mouse brain tissues were also stained with the widely used amyloid dye, thioflavin-S, to assess tau amyloid conformation. Consistent with immunostaining, immunoblot, and electron microscopic analysis, thioflavin-S-positive aggregated tau lesions were only detected in the brain of P301S mice, but not in NTG, TMFN, or P301S/TMFN mice (Fig. 3E). Thus, these results collectively reveal that neuronal Mfn2 upregulation also abrogates tau hyperphosphorylation and tau pathology formation in P301S mice.

Figure 3. Inhibition of tau hyperphosphorylation and aggregation in P301S mice by neuronal Mfn2.

A) Representative immunocytochemistry of phosphorylated tau recognized by PHF1 and AT8 in hippocampal regions of 11-month-old NTG, TMFN, P301S, and P301S/TMFN mice. B, C) Representative immunoblot (B) and quantification (C) of phosphorylated tau (recognized by both PHF1 and AT8) in total hippocampus lysates of 11-month-old NTG, TMFN, P301S, and P301S/TMFN mice (n = 4 mice per group). D) Representative EM images of paired helical filaments noted in hippocampal neurons of P301S mice. E) Representative fluorescent images for AT8 (red), Thioflavin-S (green), and DAPI (blue) staining in hippocampal neurons of 11-month-old NTG, TMFN, P301S, and P301S/TMFN mice. The data are expressed as the means ± SEM and were analyzed by Student’s t-test. *p < 0.05; **p < 0.01.

Increased expression of Mfn2 in neurons attenuated neuroinflammation in P301S mice

Microgliosis and astrogliosis are pathological hallmarks of AD and have been implicated in tau-associated neurodegeneration [34]. Ionizing calcium- binding adaptor molecule 1 (Iba1) is a widely used pan-microglial marker whose expression increases during microglial activation, whereas glial fibrillary acidic protein (GFAP) is the most commonly used astrocytic marker. To assess the impact of neuronal Mfn2 upregulation on microgliosis and astrogliosis in P301S mice, we performed immunostaining using specific antibodies against Iba1 and GFAP. Compared with NTG or TMFN mice, P301S mice showed marked Iba1 and GFAP staining in the hippocampus at 11-month-old (Fig. 4A, B), indicating the proliferation or activation of both microglia and astrocytes. When we measured microgliosis and astrogliosis in P301S/TMFN mice, there were greatly decreased Iba1 and GFAP staining in their hippocampus in comparison with P301S mice, suggesting that mutant tau induced microglia and astrocyte activation can also be attenuated by neuronal Mfn2 overexpression. Consistent with these results, we observed greatly elevated levels of GFAP and Iba1 protein in the hippocampus of P301S mice compared with NTG or TMFN mice, which was also significantly reduced in P301S/TMFN mice (Fig. 4C, D). Accompanied with activated microglia and astrocytes, induction and overproduction of proinflammatory cytokines could be noted in transgenic expressing tau P301S [35]. Among tau pathology-associated inflammatory cytokines, interleukin-1β (IL-1β) is a major player in the brain and has been repeatedly implicated in the progress of AD [36]. As previously observed in tau transgenic mice, IL-1β immunoreactivity was drastically increased in the hippocampus of P301S mice compared with NTG (Fig. 4E). In parallel to the marked reduction in microgliosis and astrogliosis, we noted substantially decreased IL-1β production in P301S/TMFN mice (Fig. 4E). Taken together, these results indicate that neuronal Mfn2 upregulation suppresses mutant tau induced neuroinflammation in P301S mice.

Figure 4. Suppression of neuroinflammation in P301S mice by neuronal Mfn2.

A, B) Representative immunocytochemistry (A) and quantification (B) of microglia marker Iba1 and astrocyte marker GFAP in the hippocampal region of 11-month-old NTG, TMFN, P301S, and P301S/TMFN mice. n = 5 or 6 mice per group. C, D) Representative immunoblot (C) and quantification (D) of Iba1 and GFAP in the total hippocampus lysates of 11-month-old NTG, TMFN, P301S, and P301S/TMFN mice. n = 3 or 6 mice per group. E) Representative immunocytochemistry for IL-1β in the hippocampus of 11-month-old NTG, TMFN, P301S, and P301S/TMFN mice. The data are expressed as the means ± SEM and were analyzed by One-way analysis of variance (ANOVA) followed by Tukey’s multiple comparison test. *p < 0.05; **p < 0.01; ****p < 0.0001; ns, non-significant.

Increased expression of Mfn2 in neurons improves cognitive function in P301S mice

Cognitive deficits have been reported in young or aged P301S mice [37–39]. The Barnes maze is a well-established, dry-land- based rodent behavioral paradigm widely used for assessing spatial learning and memory in AD mouse models [40]. To determine whether neuronal Mfn2 upregulation alleviates tau pathology–induced cognitive dysfunction, 11-month-old P301S, P301S/TMFN, and age-matched control mice were tested in the Barnes maze. P301S and P301S/TMFN mice that were paralyzed were excluded from testing. All four groups showed decreases in time to enter the escape hole with increased training (Fig. 5), indicating that spatial learning occurred. However, unlike NTG mice exhibiting greatly decreased time to reach the target in all other trails compared to day 1 trail, P301S mice only showed a very mild decrease in time in the final trial (Fig. 5), indicating that their learning ability to locate the hidden target using the spatial cues was greatly impaired. In the Barnes maze test, P301S/TMFN mice began to improve in the day 3 trail and exhibited significantly shorter time in the final trail compared with P301S mice (Fig. 5). And, although the time of P301S/TMFN mice in the day 3 and 4 trail was still longer than those of NTG mice, there is no significant difference between NTG and P301S/TMFN mice in the day 5 trail. Our findings data therefore suggest that neuronal Mfn2 also exerts beneficial effects on cognitive function in P301S mice.

Figure 5. Prevention of cognitive impairment in P301S mice by neuronal Mfn2.

A) Representative Barnes maze performance of 11-month-old NTG, TMFN, P301S, and P301S/TMFN mice. Arrowheads show the target (escape) hole. B) Quantification of time spent on finding the target hole in the Barnes maze test. Over ten mice for each genotype were used in testing (NTG: n = 11, TMFN: n = 11, P301S: n = 13, P301S/TMFN: n = 13). The data are expressed as the means ± SEM, and were analyzed by one-way ANOVA analysis. **p < 0.01; ns, non-significant.

DISCUSSION

Mitochondrial abnormalities have been implicated in AD and many other tau-related neurodegenerative diseases [41]. There is growing experimental evidence showing the multiple roles of Mfn2 in the regulation of mitochondrial function and dynamics, and their importance in neurodegeneration. Our previous studies have reported decreased expression of Mfn2 in patients with AD [33] or Amyotrophic lateral sclerosis (ALS) [20]. Our results showed that Mfn2 was also reduced in concert with mitochondrial fragmentation in the widely used animal model of tauopathy, P301S mice. Importantly, Mfn2 overexpression could greatly prevent mitochondria fragmentation, neuronal loss, tau pathology (including NFT-like inclusions), neuroinflammation, and even cognitive deficits in aged tauopathy mice. These data corroborate our previous studies regarding the common protective function of Mfn2 in neurodegeneration and implicate a likely mechanism of tau pathology regulation by Mfn2 and its mediated mitochondrial pathways.

Although mitochondrial dynamic abnormalities have been investigated in the mouse model of tauopathy [22], their role in the progression of tau pathology is not well understood. As a key regulator of the mitochondrial fusion process, Mfn2 was found reduced in P301S mice. Neuronal Mfn2 upregulation is sufficient to greatly mitigate tau hyperphosphorylation and mature tau accumulation. This is quite surprising, especially considering that Mfn2 is predominantly anchored at the outer mitochondrial membrane, and human mutant tau remained unchanged after overexpression of Mfn2. The main mitochondrial pathways regulated by Mfn2 are largely related to mitochondrial fusion and ER/mitochondria association. However, it should be noted that tau transgenic animal models exhibit increased ER/mitochondria association [25]. The fact that mutant tau-induced mitochondrial fragmentation could be alleviated by Mfn2 indicates that Mfn2 should influence tau pathology via mitochondrial fission and fusion dynamics. It remains to be studied whether the upregulation of other mitochondrial fusion factors such as Mfn1 and OPA1, or the inhibition of mitochondrial fusion factors such as DLP1 or Mff, leads to the suppression of tau pathology formation. Nevertheless, further experiments are necessary to assess whether and how tau transport, clearance, and phosphorylation are perturbed by neuronal Mfn2 and its mediated mitochondrial fusion.

Unlike our previous study showing the protective role of neuronal Mfn2 in abolishing neuromuscular synaptic loss but not spinal cord motor neuron loss in SOD1 G93A mouse models for ALS [20], our findings in this study demonstrated that neuronal Mfn2 was able to prevent brain neuron loss in aged tauopathy mice. The inhibition of calpain by its endogenous inhibitor calpastatin suppressed mutant tau induced brain neuron degeneration [42, 43]. Moreover, Mfn2 has been shown to mediate the axonal transport of calpastatin [20]. Thus, it is possible that neuronal Mfn2 engages the calpain/calpastatin system to confer neuroprotection in tauopathy mice. However, the neuroprotective effects of Mfn2 can be indirect, and driven by its impact on tau transport, clearance, and phosphorylation. Noteworthily, our recent study has shown that Mfn2 can suppress the activation of microglia and related IL-1β overproduction via regulating CX3CL1 expression [44]. Mfn2-mediated CX3CL1 expression and microglia suppression may also indirectly contribute to its neuroprotection.

Mfn2 and DLP1 are significantly reduced in P301S mice, whereas other fission and fusion regulators remain unchanged. This pattern is very similar to that of SOD1 G93A mice [20] or mice with glutamate excitotoxicity [32]. Thus, the mechanism by which Mfn2 is decreased in P301S mice is unlikely specific to mutant tau challenge. The mRNA levels of Mfn2 in brains of P301S mice are comparable to NTG mice (data not shown), excluding the possibility of tau induced Mfn2 reduction at the transcriptional level. Mfn2 has been reported to be degraded by the ubiquitin-proteasome system [45] or calpain [32]. It will be interesting to explore whether the ubiquitin-proteasome pathway or calpain protein is involved in tau pathology-induced Mfn2 in the future. Noteworthily, previous and recent studies reported the increased expression of DLP1 in tau cell and animal models [46, 47]. This discrepancy may be due to different model systems, animals at different disease stages, or tissue type and preparation differences. Further studies are required to clarify the previously implicated pathological role of DLP1 in tauopathy.

P301S mice with neuronal Mfn2 overexpression exhibited improved cognitive function, consistent with the suppressed neuronal loss, tau pathology, and neuroinflammation in their hippocampus. Mfn2 is only upregulated specifically in CNS neurons of TMFN mice [44]. The Barnes maze test found no difference between NTG and TMFN mice. Therefore, even though P301S mice might express mutant tau in peripheral tissues under the prion protein promoter, alleviated cognitive deficits by neuronal Mfn2 should be a result of suppressed neuronal loss or dysfunction. In this study, due to the hyperactivity of P301S mice, the motor coordination and balance could not be reliably assessed by the rotarod performance test. However, based on the previous studies [42, 43], improved motor neurons and motor function could be anticipated in mutant tau mice with neuronal Mfn2 upregulation.

Our present work provides evidence that the restoration of mitochondrial morphology by Mfn2 alleviates tau pathology and associated neurodegeneration, neuroinflammation, and brain dysfunction. The role of mitochondria in the pathogenesis of tauopathy and many other neurodegenerative diseases has been increasingly recognized. Our findings of Mfn2-mediated neuroprotection in the context of tau pathology are consistent with previous studies demonstrating that the enhanced mitochondrial fusion or inhibited mitochondrial fission can effectively protect neurons in other experimental models of neurodegeneration. Therefore, it would be of great interest to further validate whether Mfn2 or mitochondrial dynamics could be developed as a feasible common therapy for various major neurodegenerative diseases. In this regard, the interpretation of our results is limited to the TMFN mice used, which overexpresses Mfn2 before the appearance of tau pathology. There is a great need for future studies employing inducible mouse models to investigate whether Mfn2 and its mediated mitochondrial fusion can ameliorate or reverse tau pathology and associated neurodegeneration after disease onset.