Abstract
Neuronal mitochondria display distinct morphologies across compartments, with dendritic mitochondria being elongated and axonal ones shorter, and their morphologies are dynamically changed via fusion and fission machineries. Mitochondrial structural abnormalities are common in neurodegenerative diseases such as Alzheimer’s disease and Parkinson’s disease, yet systematic evaluation of therapeutic targets remains limited. Here, we tested key mitochondrial shape regulators, mitofusin 1/2 for fusion and Mff/Fis1 for fission, in an α-synucleinopathy model. Using MitoVis, a deep learning-based neuronal mitochondrial image analysis tool, we achieved rapid, compartment-specific analysis of mitochondrial morphologies. Among all interventions, Fis1 knockdown most effectively protected mitochondrial structure to control levels without inducing over-elongation of axonal mitochondria, which was linked to abnormal Ca2+ dynamics. While all manipulations preserved dendritic spine loss, Fis1 optimally maintained axonal mitochondrial function. These findings demonstrate a high-throughput screening approach for mitochondrial regulators and highlight Fis1 as a promising preventive/therapeutic target. Our results support targeting mitochondrial morphology as a viable strategy for treating α-synucleinopathy and potentially other mitochondria-related neurodegenerative diseases.
Similar content being viewed by others
Introduction
Mitochondria play essential roles in various neuronal properties throughout their lifespan, including neurodevelopment, synaptic function, and neuronal aging. Interestingly, recent studies have demonstrated that dendritic and axonal mitochondria possess distinct shapes, each contributing to the compartment-specific functions1,2. Elongated and tubular dendritic mitochondria regulate protein synthesis at postsynaptic sites and modulate synaptic plasticity through fusion and fission dynamics upon stimulation3,4, while short axonal mitochondria generate ATP and control Ca2+ homeostasis, influencing presynaptic maintenance and release properties5,6,7,8,9.
Alterations of mitochondrial function and structure are hallmarks in various neurodegenerative diseases, such as Alzheimer’s disease (AD), Huntington’s disease (HD), and Parkinson’s disease (PD)10,11. Notably, neuronal mitochondria exhibiting shortened, swollen, or the mitochondria-on-a-string structure have been observed in human patients and disease models11,12,13. Strikingly, a recent study demonstrated that restoration of fragmented dendritic mitochondria, achieved by the knockdown of the mitochondrial fission factor (Mff) in hippocampal neurons in an AD model, successfully rescued synaptic loss14. However, Mff-deficiency also led to elongated axonal mitochondria, resulting in reduced presynaptic release and impaired axon branching7. Therefore, restoring mitochondria in both dendritic and axonal compartments to their wild-type level is critical for maintaining proper neuronal function under pathological conditions.
Synucleinopathy, which displays oligomeric forms of α-synuclein (α-Syn), is also observed in PD, dementia with Lewy bodies (DLB), and multiple system atrophy (MSA). More interestingly, α-Syn pathology also has been reported to have mitochondrial dysfunction and structural changes although α-Syn aggregates are not fully confirmed in many model systems15,16,17,18.
α-Syn pathological phenotypes can be reproduced in cultured neurons with the addition of preformed fibrils (PFFs) synthesized from α-Syn, which induces aggregates of endogenous α-Syn within 7 days19. α-Syn PFFs treatment progressively affects synaptic structure and function, ultimately resulting in partial neuronal death after 2weeks20,21. However, the systematic testing of mitochondrial manipulation for the recovery of synucleinopathy phenotypes has yet to be conducted.
Here we establish a screening method to identify suitable regulators of mitochondrial dynamics using a deep learning-based program, which analyzes mitochondrial shapes in a neuronal compartment-specific manner22. Subsequently, we validate the efficacy of genetic regulation for the protection of postsynaptic features in synucleinopathy. This study will provide a novel target searching system for mitochondrial structural manipulators against various neurodegenerative diseases.
Results
α-Syn PFFs treatment induced mitochondrial fragmentation in cortical neurons
To investigate the mitochondrial shapes in a synucleinopathy model, we transfected cortical pyramidal neurons using ex utero electroporation with plasmids encoding HA-mCherry and mitochondria-targeting YFP (mito-YFP) to label neuronal processes and mitochondria (Fig. 1a). Then, α-Syn PFFs were added on cultured pyramidal neurons at 10 DIV and incubated for 9 days. Cells were fixed at 19 DIV and accumulation of α-Syn fibril was confirmed with phosphorylated α-Syn staining, which is a marker for fibrilized α-Syn in cells (Fig. 1c). Both axonal and dendritic mitochondrial morphologies were analyzed with MitoVis, which automatically recognizes mitochondrial structural characteristics in a compartment-specific way (Fig. 1a). Using MitoVis, various characteristics including length, circularity, and area of over 500 mitochondria per image were analyzed with manual correction, which is around 10 times faster than fully manual measurement22. Compared with DPBS control, α-Syn PFFs treated neurons showed significantly reduced mitochondrial lengths and areas, and increased circularity in dendrites (DPBS: 3.576 ± 0.049 µm, α-Syn: 3.066 ± 0.0312 µm in length, Fig. 1a, b, d, Supplementary Fig. 1). Furthermore, we observed that axonal mitochondria were also similarly affected by α-synucleinopathy condition (DPBS: 1.354 ± 0.025 µm, 1.197 ± 0.018 µm in length, Fig.1a, b, e, Supplementary Fig 1).
a, b Images of mitochondria in dendrites and axons. Both dendritic and axonal mitochondria showed reduced length in α-Syn PFFs-treated neurons. c) Images of p-α-Syn accumulation. d, e Quantification of dendritic and axonal mitochondrial length using MitoVis. nControl = 22 neurons, nα-Syn = 23 neurons. Samples were obtained from 4 independent experiments from different pups. Number of mitochondria is annotated at the bottom of bar graphs. Unpaired t-test, ****p < 0.0001, a, c Scale bar = 20 μm, b Scale bar = 5 μm.
α-Syn PFFs-mediated mitochondrial fragmentation is prevented by manipulating mitochondrial dynamics
Mitochondria are highly dynamic organelles and their shape is controlled by the balance between fusion and fission. These processes are regulated by several proteins, including Mitofusin1 (Mfn1) and Mitofusin2 (Mfn2) for mitochondrial fusion, and Mff and Fis1 for mitochondrial fission23. In neurodegenerative diseases, however, the balance between fusion and fission is disrupted, resulting in mitochondrial fragmentation24,25.
To investigate whether manipulating the genes related to mitochondrial dynamics could prevent α-Syn PFFs-mediated mitochondrial fragmentation, we first examined the overexpression effect of mitochondrial fusion genes. Neurons were transfected with Mfn1- or Mfn2-expressing plasmids together with mitochondrial targeting fluorescent protein (mito-YFP) using ex utero cortical electroporation (Fig. 2a, Supplementary Fig. 2a). Then, α-Syn PFFs were treated for 9 days and mitochondrial lengths in dendrites and axons were analyzed with MitoVis after imaging. Our results showed increased relative frequency of short dendritic mitochondria in an α-Syn-treated group. Although the Mfn1 or 2 overexpressing neurons exhibit longer mitochondria within a 1–3 µm bins compared to the control, the overall average length was similar (Control: 4.051 ± 0.055 µm, α-Syn: 3.224 ± 0.038 µm, α-Syn + Mfn1: 3.998 ± 0.062 µm, α-Syn + Mfn2: 3.862 ± 0.101 µm, Fig. 2a–c). They could also prevent the shortening of mitochondria in axon, but resulted in slightly longer than the control (Control: 1.428 ± 0.021 µm, α-Syn: 1.282 ± 0.017 µm, α-Syn + Mfn1: 1.765 ± 0.040 µm, α-Syn + Mfn2: 1.583 ± 0.043 µm, Fig. 2a–c).
a Images of mitochondria in dendrites and axons. Both dendritic and axonal mitochondria showed reduced length in α-Syn PFFs-treated neurons. Overexpressing Mfn1 and Mfn2 could protect their morphologies. b Quantification of dendritic mitochondrial length using MitoVis. Mitochondrial lengths were binned in 1 μm intervals: <1 μm, 1–2 μm, 2–3 μm, and >15 μm. Intermediate bins between 3 and 15 μm were omitted from the graph as they represented a very small fraction of the population and showed no statistically significant differences across groups. c Quantification of axonal mitochondrial length using MitoVis. Mitochondrial lengths were binned in 1 μm intervals: <0.4 μm, 0.4–0.8 μm, 0.8–1.2 μm, 1.2–1.6 μm, and 1.6–2.0 μm. The y-axis indicates these length ranges, ordered from shortest (bottom) to longest (top). The x-axis shows the relative frequency (percentage) of mitochondria falling into each length category. nControl = 34 neurons, nα-Syn = 33 neurons, nα-Syn+Mfn1 = 15 neurons, nα-Syn+Mfn2 = 15 neurons. Samples were obtained from 5 independent experiments from different pups. Number of mitochondria is annotated at the bottom of bar graphs. Two-way ANOVA, One-way ANOVA ****p < 0.0001, ####p < 0.0001. Scale bar = 5 μm.
Next, we tested if inhibition of mitochondrial fission could show similar protective results as enhancement of fusion, thereby maintaining mitochondrial morphologies. After validating knockdown efficiency of shRNAs against mff and fis1 via Western blot (Supplementary Fig. 2b, ~84% knockdown efficiency for shFis1 and ~86% knockdown efficiency for shMff), the shRNA plasmids were transfected via ex utero electroporation with HA-mCherry and mito-YFP-encoding constructs at E15.5. Then, as described above, after α-Syn PFFs incubation, mitochondrial lengths were analyzed using MitoVis. Although knockdown of Mff resulted in longer mitochondria than the α-Syn PFFs treated neurons, they showed more elongated mitochondria both in dendrites and axons compared to the control (Control: 3.532 ± 0.060 µm, α-Syn: 2.666 ± 0.033 µm, α-Syn + shMff: 4.471 ± 0.076 µm in dendrite, Control: 1.457 ± 0.027 µm, α-Syn: 1.184 ± 0.023 µm, α-Syn + shMff: 2.033 ± 0.044 µm in axon, Fig. 3a–c). In contrast to the Mff, Fis1 knockdown prevented fragmentation and maintained mitochondria at control level both in dendrites and axons (3.406 ± 0.0.048 µm in dendrite, 1.385 ± 0.022 µm in axon, Fig. 3a–c). Because the previous study reported that knockdown of Fis1 could induce cellular death26,27, we checked whether neuronal survival was affected by Fis1 downregulation; however, in our experimental condition, neuronal death was not observed (Supplementary Fig. 3). Taken together, Fis1 knockdown could be a more ideal therapeutic tool for the α-synucleinopathy model.
a Images of mitochondria in dendrites and axons. Both dendritic and axonal mitochondria showed reduced length in α-Syn PFFs-treated neurons. Knockdown of Mff and Fis1 could protect their morphology. b Quantification of dendritic mitochondrial length using MitoVis. Mitochondrial lengths were binned in 1 μm intervals: <1 μm, 1–2 μm, 2–3 μm, and >15 μm. Intermediate bins between 3 and 15 μm were omitted from the graph as they represented a very small fraction of the population and showed no statistically significant differences across groups. c Quantification of axonal mitochondrial length using MitoVis. Mitochondrial lengths were binned in 1 μm intervals: <0.4 μm, 0.4–0.8 μm, 0.8–1.2 μm, 1.2–1.6 μm, and >4 μm. Intermediate bins between 2 and 4 μm were omitted from the graph as they represented a very small fraction of the population and showed no statistically significant differences across groups. The y-axis indicates these length ranges, ordered from shortest (bottom) to longest (top). The x-axis shows the relative frequency (percentage) of mitochondria falling into each length category. nControl = 16 neurons, nα-Syn = 17 neurons, nα-Syn+shMff = 18 neurons, nα-Syn+shFis1 = 18 neurons. Samples were obtained from 3 independent experiments from different pups. Number of mitochondria is annotated at the bottom of bar graphs. Two-way ANOVA, One-way ANOVA ****p < 0.0001, ####p < 0.0001, n.s. not significant. Scale bar = 5 μm.
Preventing mitochondrial fragmentation in α-synucleinopathy neurons preserves dendritic spine density
α-Syn PFFs aggregates are known to impair synaptic functions by affecting the formation and maintenance of dendritic spines21. Besides, previous studies have shown that dendritic mitochondria play a role in regulating dendritic spine structures, and in a neurodegenerative model like AD, spine loss linked to mitochondrial fragmentation, thereby recovering their shapes returned spine loss to WT level4,14. Based on these studies, we tested if the protection of mitochondrial morphologies could reverse the synaptic impairment in α-synucleinopathy model. To test our hypothesis, we measured changes in spine density in control and α-Syn PFFs-treated neurons using the same images used in mitochondrial analyses. Consistent with previous reports, α-Syn PFFs treatment led to a significant reduction in spine density compared to the control (Control: 11.28 ± 0.506 spines/50 µm, α-Syn: 8.19 ± 0.552 spines/50 µm, Fig. 4a). However, both induction of fusion with Mfn1/2 overexpression and inhibition of fission using shRNA targeting Mff and Fis1 preserved spine density at control level (α-Syn + Mfn1: 11.60 ± 0.484 spines/50 µm, α-Syn + Mfn2: 11.72 ± 0.494 spines/50 µm, α-Syn + shMff: 11.85 ± 0.586 spines/50 µm, α-Syn + shFis1: 11.98 ± 0.502 spines/50 µm, Fig. 4a–c). Altogether, these results suggest that manipulating mitochondrial morphologies could be a preventive or therapeutic target for neurodegenerative diseases.
a Secondary dendritic segments of primary cortical neurons. Spine density was reduced in α-Syn PFFs-treated neurons. b Quantification of dendritic spine density. Mfn1/2 overexpression maintained spine density. One-way ANOVA, ***p = 0.0004, #p = 0.0235 (for Mfn1), #p = 0.0139 (for Mfn2). c Quantification of dendritic spine density. Knockdown of Mff and Fis1 maintained spine density. nControl_Fusion = 34 neurons, nα-Syn_Fusion = 33 neurons, nα-Syn+Mfn1 = 15 neurons, nα-Syn+Mfn2 = 15 neurons, nControl_Fission = 16 neurons, nα-Syn_Fission = 17 neurons, nα-Syn+shMff = 18 neurons, nα-Syn+shFis1 = 18 neurons. 2 dendrites/neuron. Samples were obtained from 5 independent experiments for Mfn/2 and 3 independent experiments for shMff/shFis1 from different pups. One-way ANOVA, ***p = 0.0005, ####p < 0.0001, n.s. not significant. Scale bar = 5 μm.
Over-elongation of presynaptic mitochondria by genetic manipulation causes abnormal Ca2+ dynamics in α-synucleinopathy neurons
Previous study reported that over-elongation of axonal mitochondria, leading to altered mitochondrial Ca2+ dynamics, presynaptic dysfunction, and abnormal axon branching development7. To check the potential side effect of extended mitochondrial length, we measured presynaptic mitochondrial Ca2+ levels after manipulating mitochondrial dynamics in an α-synucleinopathy model. Ca2+ dynamics in presynaptic mitochondria were monitored using a genetically encoded Ca2+ sensor targeted to the mitochondrial matrix (mito-jGCaMP8m), in combination with vGlut1-mScarlet and mito-mTagBFP2 to visualize presynaptic boutons in mouse cortical pyramidal neurons. At 19 DIV, mitochondrial Ca2+ responses were imaged following stimulation with 10 action potentials (at 10 Hz) delivered via a glass capillary electrode.
To quantify mitochondrial Ca2+ accumulation, we calculated the integrated intensity of mito-jGCaMP8m across the entire length of mitochondria co-localized with vGlut1 signals. In the α-synucleinopathy condition, marginally shortened axonal mitochondria showed no significant alteration in Ca2+ regulation (Fig. 5, Supplementary Fig. 4). Mfn1 overexpression and Mff knockdown, both of which produced markedly over-elongated axonal mitochondria, resulted in increased mitochondrial Ca2+ uptake following stimulation compared to both disease and control groups, whereas Mfn2 overexpression, which caused only slight elongation, showed a trend toward increased Ca2+ uptake, although the difference was not statistically significant. In contrast, Fis1 knockdown shows similar mitochondrial Ca2+ transients with control (Fig. 5, Supplementary Fig. 4). Although all genetic manipulation for mitochondrial fusion and fission maintained postsynaptic densities, these results suggest that a side effect of over-extended axonal mitochondria for treating α-synucleinopathy neurons.
a Representative images of axonal mito-jGCaMP8m fluorescence at peak intensity following stimulation (10AP at 10 Hz) in control, α-Syn, α-Syn + shMff and α-Syn + shFis1 groups. b Integrated intensity of mito-jGCaMP8m signals from full-length mitochondria associated with single presynaptic sites. c Quantification of area under the curve (AUC) shows enhanced mitochondrial Ca2+ uptake in the Mff knockdown group. One-way ANOVA, **p = 0.004, ###p < 0.0001, #p = 0.012 (for α-Syn + shMff), #p = 0.045 (for α-Syn + shFis1), n.s. not significant. nControl = 21 images, nα-Syn = 23 images, nα-Syn+shMff = 16 images, nα-Syn+shFis1 = 20 images. Samples were obtained from 6 independent experiments from different pups. Number of mitochondria are annotated at the bottom of bar graphs. Scale bar = 2 μm.
Discussion
In the present study, we demonstrated that the preservation of the mitochondrial lengths in both dendrites and axons to WT levels is sufficient to maintain the number of spines while avoiding mitochondrial Ca2+ over-uptake in α-synucleinopathy model neurons. To achieve efficient screening, we established a sequential workflow; initiating neuronal synucleinopathy by applying α-Syn PFF, imaging neuronal mitochondria, and performing compartment-specific analysis using a deep learning-based program (Supplementary Fig. 5). Through these steps, we unveiled that downregulation of Fis1 is a promising way to protect mitochondrial morphologies in both dendrites and axons of cultured neurons, highlighting its potential as a preventive or therapeutic target for α-synucleinopathy.
Manipulating other mitochondrial dynamics related genes, Mfn1 and Mff, led to excessive mitochondrial elongation, mainly in axons, through overexpression or downregulation, respectively. Such extension may impair presynaptic function by increasing mitochondrial Ca2+ uptake during stimulation7. α-Syn PFF treatment shortened mitochondria; however, their Ca2+ uptake ability remained comparable to controls. Although previous studies have shown that increased ER-mitochondria contacts in various neurodegenerative conditions can promote mitochondrial fission and excessive Ca2+ transfer from ER to mitochondria, thereby inducing cellular toxicity 28. However, these phenotypes observed mainly in somatodendritic regions, and axonal mitochondria are already relatively small (~1 μm), so further shortening would likely have a milder effect. Moreover, other studies demonstrated that axonal mitochondria take up Ca2+ independently of ER Ca2+ release, in contrast to dendritic mitochondria5,29. Therefore, fission-related impact on axonal mitochondrial Ca2+ uptake may be limited in this model.
Mfn2 overexpression induced mild over-elongation of axonal mitochondria, though without a statistically significant increase in mitochondrial Ca2+ uptake, despite a rising trend (Supplementary Fig. 4). Beyond its role in mitochondrial fusion, Mfn2 also regulates ER-mitochondria contacts by localizing to the ER membrane23,30, suggesting that Mfn2 upregulation could have additional side effects. Under non-disease conditions, Fis1 knockdown did not significantly elongate mitochondria, although a slight shortening was observed in axons (Supplementary Fig. 6a). In addition, spine density remained unaffected by manipulating both mitochondrial fusion and fission regulators (Supplementary Fig. 6b). Furthermore, the ability of Fis1 downregulation to protect against α-synucleinopathy-associated phenotypes may be attributed to its elevated expression under pathological conditions, whereas Mfn1/2 and Mff levels remain relatively unchanged (Supplementary Fig. 7). Taken together, these findings suggest that Fis1 is the optimal target for preventing α-synucleinopathy, as its downregulation preserves mitochondrial length and function at WT levels without inducing toxicity.
AI-assisted image analysis has been widely applied to recent studies, including automated estimation of mitochondrial morphology31,32,33,34,35. While a deep learning-based program was previously employed to examine neuronal mitochondrial lengths and circularities for rapid drug screening35, it primarily distinguished dendritic and axonal mitochondria based on their lengths rather than actual shapes of neuronal compartments. MitoVis, which is employed in the current study, initially segments pyramidal neuronal dendrites and axons, then performs analysis of mitochondrial structures22. In addition, fully automated analysis yielded similar results to manually corrected data, albeit with slightly lower significance (Figs. 2 and 3, Supplementary Fig. 8). Therefore, for assessing the impact of drug candidates on compartment-specific mitochondria, MitoVis appears to be a more suitable software.
Previously, multiple studies explored mitochondrial shapes in various synucleinopathy models15,16,17,18. Human α-Syn expressing C. elegans motor neurons or Drosophila larval neurons showed fragmented mitochondria36,37,38. In addition, cells expressing α-Syn WT- or PD-related mutants, as well as those seeded with α-Syn PFF, including SH-SY5Y, HeLa, COS, and neurons from rodents and human have shortened mitochondria16,36,39,40. Transgenic mice with A53T α-Syn mutation, found in PD patients, also exhibited reduced mitochondrial area and a higher prevalence of circular shapes41. However, some studies reported contrasting mitochondrial phenotypes in α-synucleinopathy models. The somatic region of mammalian neurons in A53T human α-Syn transgenic mice and Kenyon neurons in α-Syn transgenic flies showed enlarged mitochondria42,43. Nevertheless, these studies did not investigate correlation between dendritic mitochondrial shape and synapse number in the α-synucleinopathy condition.
Recent studies have revealed the substantial influence of mitochondria on dendritic spine number and structural dynamics. In AD mouse models, spine loss in hippocampal CA1 neurons is closely linked to mitophagy, potentially reversible by restoring mitochondria from fragmented shapes and decreased density to normal level by Mff knockdown14. In addition, inhibition of mitochondrial function impairs protein synthesis-dependent spine plasticity by dampening energy supply4,44. This study also highlights that the dendritic mitochondria tether to cytoskeleton, such as actin and microtubule, and destabilization of them decreases mitochondrial length significantly4. Previously, structural and functional reduction of excitatory synapses were observed in α-Syn PFF treated conditions21; however, the underlying mechanisms are not well studied yet. Like the AD model, mitochondrial fragmentation might be one of the potential mechanisms, inducing spine loss through mitophagy.
Inhibition of Drp1-Fis1 interaction is one of the targeted mechanisms for treating PD model animals. Treatment of P110, a peptide inhibitor of Drp1-Fis1 binding, has been shown to restore shortened mitochondrial length and reduces cell death in the substantia nigra45,46,47. Additionally, several locomotor behaviors were improved by P110 treatment although striatal dopamine levels were not recovered45. Furthermore, the Drp1 inhibitor, Mdivi-1, also prevents cell death in substantia nigra of α-Syn-mutated PD rat models and partially rescues dopamine levels and locomotor function48, while also restoring mitochondrial shape and function48. However, these studies have not systematically compared multiple mitochondrial shape regulators or assessed their impact on synaptic recovery. In this study, we addressed this gap by examining mitochondrial structural restoration at an earlier disease stage, thereby emphasizing the importance of Fis1 inhibition in preserving synaptic structure and function in α-synucleinopathy.
Our findings suggest that targeting mitochondrial morphology in a compartment-specific manner offers a promising preventive or therapeutic strategy for neurodegenerative diseases that impair synapses. These insights may inform clinical application, including the development of antisense oligonucleotides (ASOs) against Fis149. Furthermore, integrating an automated system with the MitoVis analysis software will accelerate the discovery of drug candidates that regulate mitochondrial morphology.
Methods
Animal
Time-pregnant CD1 females were obtained from DaeHan Bio Link (Eumseong, Korea). All experiments were performed in accordance with protocols approved by the Institutional Animal Care and Use Committee (IACUC) at the Korea Institute of Science and Technology (KIST-2020-133).
Plasmids
pCAG-mito-YFP and pCAG-HA-mCherry were previously published6. pCAG-Mfn1-Myc and pCAG-Mfn2-Myc were created by placing the DNA encoding Mfn1 (Addgene plasmid #23212) and Mfn2 (Addgene plasmid #23213) 3’ to the CAG promoter50. pLKO.1-scrambled, Mff shRNA and Fis1 shRNA constructs were validated in a previous study7,51.
Sequences for Mff shRNA: 5′-CCGGCTTCATTAAGACGTCAGATAACTCGAGTTATCTGACGTCTTAATGAAGTTTTTTG-3′ and for Fis1 shRNA: 5′-CCGGGGTGCCTGGTTCGAAGCAAATCTCGAGATTTGCTTCGAACCAGGCACCTTTTTG-3′
Ex utero cortical electroporation and primary cortical neuron culture
Electroporated primary cortical neurons were obtained as previously described7. Briefly, 1–2 mg/ml of plasmid was injected into the lateral ventricle of E15.5 mouse embryo and electroporated with four pulses of 20 V for 100 ms interval. Then, cortices from the electroporated mouse head were dissected and incubated with Hank’s Balanced Salt Solution (HBSS) containing papain (Worthington, Columbus, OH) and DNaseI (Sigma, St. Louis, MO) for 15 min at 37 °C. Cells were dissociated by pipetting and plated on Poly-D-lysine coated coverslip (Corning, Corning, NY) in Neurobasal media (Sigma) with FBS (Sigma), B27 (Gibco, Carlsbad, CA), Glutamax (Gibco) and penicillin/streptomycin (Gibco). After 5 days, media were changed with Neurobasal media without FBS.
For Western blot experiments, dissociated cells (5.1 × 105) were plated on PDL-coated 6-well plates in Neurobasal media (Invitrogen, Waltham, MA) containing 2% B27, 1% Glutamax, and 1% penicillin/streptomycin. Neurons were maintained for 10–11 days in vitro (DIV) in a humidified 37 °C incubator with 5% CO2.
Lentiviral production and infection
Lentiviruses were produced from HEK293FT cells by co-transfection with shuttle vectors, LP1, LP2, and VSV-G. jetPrime (Polyplus, Strasbourg, France) was used as manufacturer’s protocol and 24 h after transfection, media were changed with Neurobasal media, and 48 h later, supernatants were harvested and centrifuged to remove cellular debris. 100–200 μl of viral supernatants were added to primary cortical neurons at 3–5 DIV and harvested 10 days after.
α-Syn PFFs treatment
Active human recombinant α-Syn preformed fibrils (PFFs) were purchased from StreeMarq (Victoria, BC). Before the treatment, α-Syn PFFs was diluted in PBS at 0.1 mg/ml and sonicated for 30 s. At 10 DIV, α-Synuclein was added to the culture media in a concentration of 1 μg/ml and incubated for 9 days.
Immunocytochemistry
Neurons were fixed at 19 DIV for 15 min at room temperature (RT) in 4% PFA with 4% Sucrose and then washed with PBS. After permeabilization with Triton X-100 in PBS, cells were incubated in blocking buffer (0.1% BSA and 2.5% goat serum in PBS) to block nonspecific signals. Primary and secondary antibodies were diluted in the blocking buffer described above and incubated at 4 °C overnight or at RT for 1 h, respectively. Coverslips were mounted on slides with VECTASHIELD® Vibrance™ Antifade Mounting Medium (Vector Laboratories, Newark, CA). Primary antibodies used in this experiment are mouse anti-HA (Biolegend, San Diego, CA, 1:500), rabbit anti-α-Synuclein(pS129) (Abcam, San Francisco, CA, 1:300), mouse anti-Myc (Invitrogen, 1:500), and all used secondary antibodies are Alexa-conjugated (Invitrogen) and used at 1:1000 dilution.
Western blotting
Cells were harvested and lysed using RIPA buffer (50 mM Tris-HCl pH7.4, 150 mM NaCl, 1% NP40, 0.5% Sodium deoxycholate, 0.1% SDS) supplemented with the cocktail of protease inhibitor and phosphatase inhibitor at 4 °C for 1 h. 5 or 10 µg of total protein per well was loaded on 10–12% SDS-PAGE gels and transferred to nitrocellulose membrane (Merck, Darmstadt, Germany). After transfer, membranes were blocked for 30 min with 5% skim milk (LPS Solution), followed by incubation with primary antibodies overnight at 4 °C. The following primary antibodies were used for Western blotting in the study: Mff (Proteintech, Rosemont, IL, 1:5000), Fis1 (Proteintech, 1:5000), NeuN (Sigma, 1:1000), GAPDH (Proteintech, 1:10000), beta-actin (Abbkine, Wuhan, China, 1:10000).
The membranes were washed three times with 0.1% Tween20/TBS (TBST) and incubated with HRP-conjugated secondary antibodies (Abbkine) for 1 h at room temperature, followed by three washes with TBST. The bands were detected using ECL Western blotting substrate (Amersham, Marlborough, MA) and then scanned by iBright CL750 Imaging System (Thermofisher, Waltham, MA).
Image acquisition and mitochondrial morphology analysis using MitoVis
Fixed samples were imaged on a Nikon A1R confocal microscope using Nikon software NIS-Element. Images were acquired using three lasers (488 nm, 561 nm, and 647 nm) together with Nikon objective 60× (1.25 NA) oil lens.
Automated neuronal structure segmentation and mitochondrial detection were done in the acquired raw images using MitoVis. The program was pre-trained for this experiment and an optimized version was used for this study.
As previously described, MitoVis first segments the dendrites, axons, and cell bodies from neuron image using a pretrained deep learning model. Mitochondrial objects are detected from mitochondrial images using another pretrained deep learning model. Among detected mitochondrial objects, those that are not co-localized with segmented neuronal structure are excluded. The remaining mitochondrial objects are classified by compartment (dendritic or axonal), thereby enabling compartment-specific morphological analysis. After initial segmentation and detection are completed, the user can manually correct errors in neuronal structure or mitochondrial object classification. The program is then retrained based on the user corrections to enhance accuracy.
Morphological analysis was conducted using the segmented structures and compartment-classified mitochondria. Four features are extracted for further analysis: Area (S): Number of pixels in a mitochondria segment. Length: Length of the skeleton curve connecting the farthest two points on a mitochondria segment. Circularity: Describes how close a mitochondria shape is to a circle. Eccentricity: Describes how elongated a mitochondria shape is by calculating the ratio of the major axis and the minor axis created by principal component analysis (PCA).
Analysis of spine density
Dendritic spine densities were quantified on secondary dendritic branches for cultured neurons and in the depth of the z stack for slices, using a Fiji (Image J) software. The length of the dendritic segment was measured on the z projection and only stubby and mushroom shape spines were quantified. Spine density was defined as the number of quantified spines divided by the length over which the spines were quantified. The analysis was conducted using the images that were previously used to analyze mitochondrial morphology.
Live imaging and analysis
Electroporated cortical neurons were imaged at 19 DIV with Nikon microscope FN1 (60× objective 1.0 NA). Modified normal tyrode solution (NT) was used as a bath solution: 145 mM NaCl, 2.5 mM KCl, 10 mM HEPES pH 7.4, 2 mM CaCl2, 1 mM MgCl2, 10 mM glucose. or mitochondrial calcium imaging, we treated a small amount of tetrodotoxin (TTX, less than 50 nM) to avoid the spontaneous activity of cultured neurons. The electrical stimulation is triggered by a glass capillary electrode placed near the transfected axon. We applied 10 pulses (10 Hz) with 30 µA using a stimulator (Model 2100, A-M systems, Sequim, WA) and imaged with 500 ms interval (2 Hz) for mito-jGCaMP8m signals.
Images were analyzed using a Fiji (Image J) plug-in, Time Series Analyzer (v3.0). mito-jGCaMP8m puncta and nearby backgrounds were selected by circular region of interests and intensities were measured by plug-in. After intensities were corrected for background subtraction, ΔF values were calculated from F–F0. F0 values were defined by averaging frames before stimulation.
Statistics
Statistical analysis was performed in GraphPad’s Prism 8 and graphs were drawn with the same software. Statistical tests, (n) numbers and significance levels are presented in the figure legends. All analysis was performed on raw imaging data without any adjustments. Images in figures have been adjusted for brightness and contrast (identical for control and experimental conditions in groups compared).
Data availability
All data reported in this paper will be shared by the corresponding authors upon request. This paper does not report original code. Any additional information required to reanalyze the data reported in this work paper is available from the corresponding authors upon request.
References
Rangaraju, V. et al. Pleiotropic mitochondria: the influence of mitochondria on neuronal development and disease. J. Neurosci. 39, 8200–8208 (2019).
Lee, A., Hirabayashi, Y., Kwon, S. K., Lewis, T. L. Jr. & Polleux, F. Emerging roles of mitochondria in synaptic transmission and neurodegeneration. Curr. Opin. Physiol. 3, 82–93 (2018).
Divakaruni, S. S. et al. Long-term potentiation requires a rapid burst of dendritic mitochondrial fission during induction. Neuron 100, 860–875.e867 (2018).
Rangaraju, V., Lauterbach, M. & Schuman, E. M. Spatially stable mitochondrial compartments fuel local translation during plasticity. Cell 176, 73–84.e15 (2019).
Ashrafi, G., de Juan-Sanz, J., Farrell, R. J. & Ryan, T. A. Molecular tuning of the axonal mitochondrial Ca(2+) uniporter ensures metabolic flexibility of neurotransmission. Neuron 105, 678–687.e675 (2020).
Kwon, S. K. et al. LKB1 regulates mitochondria-dependent presynaptic calcium clearance and neurotransmitter release properties at excitatory synapses along cortical axons. PLoS Biol. 14, e1002516 (2016).
Lewis, T. L. Jr, Kwon, S.-K., Lee, A., Shaw, R. & Polleux, F. MFF-dependent mitochondrial fission regulates presynaptic release and axon branching by limiting axonal mitochondria size. Nat. Commun. 9, 5008 (2018).
Rangaraju, V., Calloway, N. & Ryan, T. A. Activity-driven local ATP synthesis is required for synaptic function. Cell 156, 825–835 (2014).
Sun, T., Qiao, H., Pan, P.-Y., Chen, Y. & Sheng, Z.-H. Motile axonal mitochondria contribute to the variability of presynaptic strength. Cell Rep. 4, 413–419 (2013).
Chen, W., Zhao, H. & Li, Y. Mitochondrial dynamics in health and disease: mechanisms and potential targets. Signal Transduct. Target. Ther. 8, 333 (2023).
Giorgi, C., Marchi, S. & Pinton, P. The machineries, regulation and cellular functions of mitochondrial calcium. Nat. Rev. Mol. Cell Biol. 19, 713–730 (2018).
Geibl, F. F. et al. α-Synuclein pathology disrupts mitochondrial function in dopaminergic and cholinergic neurons at-risk in Parkinson’s disease. Mol. Neurodegeneration 19, 69 (2024).
Zhang, L. et al. Altered brain energetics induces mitochondrial fission arrest in Alzheimer’s Disease. Sci. Rep. 6, 18725 (2016).
Lee, A. et al. Aβ42 oligomers trigger synaptic loss through CAMKK2-AMPK-dependent effectors coordinating mitochondrial fission and mitophagy. Nat. Commun. 13, 4444 (2022).
Ganjam, G. K. et al. Mitochondrial damage by α-synuclein causes cell death in human dopaminergic neurons. Cell Death Dis. 10, 865 (2019).
Nakamura, K. et al. Direct membrane association drives mitochondrial fission by the Parkinson disease-associated protein alpha-synuclein. J. Biol. Chem. 286, 20710–20726 (2011).
Vicario, M., Cieri, D., Brini, M. & Calì, T. The close encounter between alpha-synuclein and mitochondria. Front. Neurosci. 12, 388 (2018).
Zambon, F. et al. Cellular α-synuclein pathology is associated with bioenergetic dysfunction in Parkinson’s iPSC-derived dopamine neurons. Hum. Mol. Genet. 28, 2001–2013 (2019).
Volpicelli-Daley, L. A. et al. Exogenous α-synuclein fibrils induce Lewy body pathology leading to synaptic dysfunction and neuron death. Neuron 72, 57–71 (2011).
Volpicelli-Daley, L. A., Luk, K. C. & Lee, V. M. Addition of exogenous α-synuclein preformed fibrils to primary neuronal cultures to seed recruitment of endogenous α-synuclein to Lewy body and Lewy neurite-like aggregates. Nat. Protoc. 9, 2135–2146 (2014).
Wu, Q. et al. α-Synuclein (αSyn) preformed fibrils induce endogenous αSyn aggregation, compromise synaptic activity and enhance synapse loss in cultured excitatory hippocampal neurons. J. Neurosci. 39, 5080–5094 (2019).
Choi, J. et al. MitoVis: a unified visual analytics system for end-to-end neuronal mitochondria analysis. IEEE Trans. Vis. Comput. Graph. 30, 3457–3473 (2024).
Tábara, L.-C., Segawa, M. & Prudent, J. Molecular mechanisms of mitochondrial dynamics. Nat. Rev. Mol. Cell Biol. 26, 123–146 (2025).
Wang, X. et al. Impaired balance of mitochondrial fission and fusion in Alzheimer’s disease. J. Neurosci. 29, 9090–9103 (2009).
Wang, X. et al. LRRK2 regulates mitochondrial dynamics and function through direct interaction with DLP1. Hum. Mol. Genet. 21, 1931–1944 (2012).
Cheng, W. C. et al. Fis1 deficiency selects for compensatory mutations responsible for cell death and growth control defects. Cell Death Differ. 15, 1838–1846 (2008).
Fannjiang, Y. et al. Mitochondrial fission proteins regulate programmed cell death in yeast. Genes Dev. 18, 2785–2797 (2004).
Liu, X., Li, T., Tu, X., Xu, M. & Wang, J. Mitochondrial fission and fusion in neurodegenerative diseases:Ca(2+) signalling. Mol. Cell Neurosci. 132, 103992 (2025).
Jang, D. C. et al. Asymmetric distribution of mitochondrial Ca2+ regulators specifies compartment-specific mitochondrial function and neuronal development. Preprint at https://www.biorxiv.org/content/10.1101/2025.06.23.660978v1 (2025).
de Brito, O. M. & Scorrano, L. Mitofusin 2 tethers endoplasmic reticulum to mitochondria. Nature 456, 605–610 (2008).
Zahedi, A. et al. Deep analysis of mitochondria and cell health using machine learning. Sci. Rep. 8, 16354 (2018).
Rohani, A., Kashatus, J. A., Sessions, D. T., Sharmin, S. & Kashatus, D. F. Mito Hacker: a set of tools to enable high-throughput analysis of mitochondrial network morphology. Sci. Rep. 10, 18941 (2020).
Fischer, C. A. et al. MitoSegNet: easy-to-use deep learning segmentation for analyzing mitochondrial morphology. iScience 23, 101601 (2020).
Vojtová, J. et al. A fully automated morphological analysis of yeast mitochondria from wide-field fluorescence images. Sci. Rep. 14, 30144 (2024).
Varkuti, B. H. et al. Neuron-based high-content assay and screen for CNS active mitotherapeutics. Sci. Adv. 6, eaaw8702 (2020).
Kamp, F. et al. Inhibition of mitochondrial fusion by α-synuclein is rescued by PINK1, Parkin and DJ-1. EMBO J. 29, 3571–3589 (2010).
Krzystek, T. J. et al. Differential mitochondrial roles for α-synuclein in DRP1-dependent fission and PINK1/Parkin-mediated oxidation. Cell Death Dis. 12, 796 (2021).
Furlong, R. M., O’Keeffe, G. W., O’Neill, C. & Sullivan, A. M. Alterations in α-synuclein and PINK1 expression reduce neurite length and induce mitochondrial fission and Golgi fragmentation in midbrain neurons. Neurosci. Lett. 720, 134777 (2020).
Grassi, D. et al. Identification of a highly neurotoxic α-synuclein species inducing mitochondrial damage and mitophagy in Parkinson’s disease. Proc. Natl. Acad. Sci. USA 115, E2634–e2643 (2018).
Butler, E. K. et al. The mitochondrial chaperone protein TRAP1 mitigates α-Synuclein toxicity. PLoS Genet. 8, e1002488 (2012).
Xie, W. & Chung, K. K. Alpha-synuclein impairs normal dynamics of mitochondria in cell and animal models of Parkinson’s disease. J. Neurochem. 122, 404–414 (2012).
Ordonez, D. G., Lee, M. K. & Feany, M. B. α-Synuclein induces mitochondrial dysfunction through spectrin and the actin cytoskeleton. Neuron 97, 108–124.e106 (2018).
Portz, P. & Lee, M. K. Changes in Drp1 function and mitochondrial morphology are associated with the α-synuclein pathology in a transgenic mouse model of Parkinson’s disease. Cells 10, 885 (2021).
Bapat, O. et al. VAP spatially stabilizes dendritic mitochondria to locally support synaptic plasticity. Nat. Commun. 15, 205 (2024).
Filichia, E., Hoffer, B., Qi, X. & Luo, Y. Inhibition of Drp1 mitochondrial translocation provides neural protection in dopaminergic system in a Parkinson’s disease model induced by MPTP. Sci. Rep. 6, 32656 (2016).
Qi, X., Qvit, N., Su, Y.-C. & Mochly-Rosen, D. A. novel Drp1 inhibitor diminishes aberrant mitochondrial fission and neurotoxicity. J. Cell Sci. 126, 789–802 (2013).
Su, Y.-C. & Qi, X. Inhibition of excessive mitochondrial fission reduced aberrant autophagy and neuronal damage caused by LRRK2 G2019S mutation. Hum. Mol. Genet. 22, 4545–4561 (2013).
Bido, S., Soria, F. N., Fan, R. Z., Bezard, E. & Tieu, K. Mitochondrial division inhibitor-1 is neuroprotective in the A53T-α-synuclein rat model of Parkinson’s disease. Sci. Rep. 7, 7495 (2017).
Rinaldi, C. & Wood, M. J. A. Antisense oligonucleotides: the next frontier for treatment of neurological disorders. Nat. Rev. Neurol. 14, 9–21 (2018).
Chen, H. et al. Mitofusins Mfn1 and Mfn2 coordinately regulate mitochondrial fusion and are essential for embryonic development. J. Cell Biol. 160, 189–200 (2003).
Sarbassov, D. D., Guertin, D. A., Ali, S. M. & Sabatini, D. M. Phosphorylation and regulation of Akt/PKB by the rictor-mTOR complex. Science 307, 1098–1101 (2005).
Acknowledgements
We also thank all members of the Kwon lab and collaborators for feedback and discussion along the way. This research was supported by the National Research Foundation of Korea (NRF) grant funded by the Korean government (MSIT) (2019M3E5D2A0106379412, 2020R1C1C1006386, RS-2022-NR067817, RS-2023-00264980 to S.-K.K.; RS-2021-NR061738 to D.C.J.) and KIST Program (26Z9001 and 26E0131 to S.-K.K.).
Author information
Authors and Affiliations
Contributions
S.Y.K., K.H., S-K.K. conceptualized the study. S.Y.K., J.Y.C., D.C.J., P.L., G.S.H., J.K., W-K.J., K.H., S-K.K. performed the study and S.Y.K., J.Y.C, D.C.J., P.L. analyzed data with a manual way or a software. S.Y.K. and S-K.K. originally wrote the paper, and all authors reviewed and edited it.
Corresponding author
Ethics declarations
Competing interests
S.Y.K. and S-K.K. are preparing the patent application related to the contents of this article, in which S.Y.K. and S-K.K. are listed as inventors. J.Y.C. and W-K.J. are co-founders of VIENCE Inc. Other authors declare that they have no competing interests.
Additional information
Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Supplementary information
Rights and permissions
Open Access This article is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License, which permits any non-commercial use, sharing, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if you modified the licensed material. You do not have permission under this licence to share adapted material derived from this article or parts of it. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by-nc-nd/4.0/.
About this article
Cite this article
Kim, S.Y., Choi, J., Jang, D.C. et al. Systematic evaluation of mitochondrial morphology regulators for amelioration of neuronal α-synucleinopathy. npj Parkinsons Dis. 12, 58 (2026). https://doi.org/10.1038/s41531-026-01277-z
Received:
Accepted:
Published:
Version of record:
DOI: https://doi.org/10.1038/s41531-026-01277-z
