Subjects

We included a male patient carrying the p.D620N mutation in VPS35, described previously by Follet et al ¹⁰ and two healthy male controls from Tübingen Biobank Control 1 and Control 2. Skin biopsies were taken from each individual aged 73, 72, and 77 years, respectively.

Ethical approval for the generation and functional characterization of patient‐derived iPSCs have been provided by informed consent.

Cell Culture and Treatments

For mitochondrial morphology, membrane potential, and ROS assessment, neurons were cultivated in neuronal medium without B27 and ascorbic acid (without antioxidants supplementation) 4 hours prior to the experiment. All treatments were performed in the neuronal medium without antioxidants. To assess the mitophagic clearance capacity, the edited neurons with the Rosella construct were treated with 10 μM CCCP (carbonyl cyanide 3‐chlorophenylhydrazone) (Abcam, Cambridge, UK) for 24 hours. For the autophagy experiment, neurons were treated with 100 nM Bafilomycin A1 (Enzo Life Sciences, Bruxelles, Belgium) for 24 hours. For autophagy enhancement, neurons were treated with 25 or 50 nM rapamycin (Enzo Life Sciences) for 24 hours. All experiments were repeated on three to six independent neuronal differentiations.

Live Cell Imaging and Analysis

Mitochondria were visualized using 100 nM MitoTracker Green FM (Invitrogen, Gent, Belgium) in neuronal medium without antioxidants and lysosomes with 100 nM LysoTracker Deep Red (Invitrogen) in neuronal medium with or without antioxidants. At least five Z‐stack images per well were acquired using a Zeiss spinning disk confocal microscope. All the raw image datasets used in this study are deposited online in our R3 lab of the University of Luxembourg (https://webdav-r3lab.uni.lu/public/MitoNetworks/VPS35Neurons/).

To segment mitochondria, the mitochondrial channel was pre‐processed with a difference of Gaussians where the foreground image was convolved with a Gaussian of size 11 and standard deviation 1 and the subtracted background image with a Gaussian of size 11 and standard deviation 3 (Mito_DoG). Only pixels above threshold 3000 in Mito_DoG and an intensity above 5000 in the raw mitochondrial channel were considered as foreground pixels. The mitochondrial mask was defined by removing connected components with less than 10 pixels. Mitochondrial morphometrics were quantified as previously described. ¹⁹ Additional data on mitochondrial network can be consulted in Zanin et al. ²⁰

Mitorosella Sensor, Generation of the Lines, Image Acquisition and Analysis

The generation of the lines carrying the Rosella reporter was performed as previously described. ²¹ Briefly, the tandem fluorescent proteins consisting of pH sensor fluorescent protein pHluorin (F64L, S65T, V193G, and H231Q) and DsRed were fused to the entire open reading frame of ATP5C1 serving as a mitochondrial targeting sequence, and placed in between the homology arms targeting the AAV1 safe harbour ²² (Addgene plasmid #22075). A double‐strand break for triggering homologous recombination was performed with the px330 ²³ (Addgene plasmid #42230) carrying the sgRNA targeting sequence for the safe harbor as described by Mali et al. ²⁴ SmNPC from patients (VPS35 1_2) and control individuals (Control 1) were nucleofected (P3 Primary Cell 4D‐Nucleofector, V4XP‐3024; Lonza, Basel, Switzerland) with both constructs and expanded before purification by fluorescence‐activated cell sorting (FACS) (Aria III; Beckton Dickinson, Franklin Lakes, NJ).

Images were obtained on an Opera QEHS confocal spinning disk microscope (Perkin Elmer, Waltham, MA) with a 60× water immersion objective (NA = 1.2). pHluorin was excited with a 488 nm laser and detected on camera 1 behind a 520/35 bandpass filter, while DsRed was excited with a 561 nm laser and detected on camera 2 behind a 600/40 bandpass filter. A 568 dichroic mirror split the light towards the corresponding cameras. Both fluorescent channels were acquired simultaneously with a binning setting of 2. One plane and 15 fields per well were acquired. One pixel corresponds to 0.2152 μm. Differentiated neurons were maintained under normal incubation conditions (37°C, 5% CO₂, and 80% humidity) within the microscope in between and during the different acquisition time points.

The automated image analysis was performed through a series of pre‐processing and thresholds in MATLAB (The MathWorks, Inc., Natick, MA) as previously described. ²¹ Briefly, a difference of Gaussian of convoluted foreground and background images was used for detecting all the events in the field. For classifying the events either as a mitochondrial or mitophagic event, a combination of green to red fluorescence ratio analysis and morphological filtering based on difference of Gaussians thresholding was used. Those presenting a mean ratio value below 0.6 were classified as mitophagic events.

Flow Cytometry

Neurons were detached with Accutase (Sigma, Bornem, Begium) and centrifuged at 300 g for 3 minutes. Batches of 200,000 cells were then incubated in the dye or in the buffer (unstained). MMP was assessed by staining the single‐cell suspension with 200 nM tetramethylrhodamine ethyl ester (TMRE; Invitrogen) for 30 minutes at 37°C. To correct for mitochondrial mass, MitoTracker Green FM (Invitrogen) was used as a counterstaining. For mitochondrial ROS, the single‐cell suspension was stained with 2 μM MitoSOX Red (Invitrogen) for 15 minutes at 37°C without CO₂. Cells were analyzed with the BD LSRFortessa flow cytometry analyser and the mean fluorescence intensity of each dye was assessed on at least 20,000 single cells by using FlowJo LLC software. Mean fluorescence of the unstained cells was subtracted to account for autofluorescence.

Oxygen Consumption Rate Measurement

Oxygen consumption rate (OCR) was measured in whole cells using the Seahorse XFe96 Cell Metabolism Analyser (Agilent, Diegem, Belgium). Neurons were plated in the Seahorse XFe96 well plates 24 hours prior to measuring at a density of 80,000 cells per well. The concentrations of mitochondrial toxins used were optimized for neurons according to the manufacturer's recommendations. The final concentrations of toxin used were: oligomycin (oligo) ‐ 2 μM; FCCP ‐ 250 nM; antimycin A (AA), and rotenone (rot) ‐ 5 μM. The cells of each well were lysed with radioimmunoprecipitation assay (RIPA) buffer after the experiment and the OCR of each well was corrected for protein amount.Statistics

Statistical analyses were performed with GraphPad Prism. The statistical analyses performed and the P value of each experiment can be found in the legend of the figures.

Data Availability

The authors confirm that the data supporting the findings of this study are available within the article and its supplementary material.

Clinical Phenotype of p.D620N VPS35 Patient

The male patient donor case #2610 comes from a multi‐incident family #445 reported previously from the Queensland Parkinson's Project, ²⁵ that immigrated from Western Europe. Besides the index patient, the diagnosis of PD was made in his mother and maternal grandfather but no additional family members are known to be affected. The diagnosis of PD was made by a movement disorders neurologist after a 1‐year history of muscular rigidity and tremor. The PD‐related motor symptoms responded well to levodopa therapy. The index patient underwent deep brain stimulation 6 years after diagnosis after experiencing motor fluctuations with good treatment response.

VPS35 and VPS29 Levels are Unchanged in p.D620N VPS35 Patient‐Derived Neurons

Previous studies have reported that VPS35 protein levels did not change in cells carrying mutant VPS35. ⁸ The p.D620N mutation in VPS35 has been shown not to impair its binding to the other components of the retromer. ¹⁰ To investigate the levels of retromer components we differentiated iPSC from the index patient (VPS35 1_1 and 1_2) ²⁶ and two age‐ and gender‐matched controls (Control 1 and 2) into smNPC. ¹⁸ All iPSC and smNPC clones were fully characterized in this study (Figs. S1–S3) or elsewhere. ²⁶ We successfully differentiated these smNPC into physiologically active neurons expressing the neuronal marker Tuj1 (tubulin β3), enriched in dopaminergic neurons expressing tyrosine hydroxylase (TH) (Figs. 1A and S4A‐H). ¹⁸ No difference was observed in terms of neuronal differentiation efficiency (Figs. 1B and S4A,B) or neuronal network complexity between control‐ and patient‐derived lines (Figs. 1C and S4D). We identified by western blotting that protein levels of both retromer components VPS35 and VPS29 were unchanged between control‐ and patient‐derived neurons (Fig. 1D,E).

FIG. 1

No difference in neuronal morphology and network nor in levels of VPS35 and VPS29 proteins between patient and control induced pluripotent stem cell (IPSC)‐derived neurons. (A) Representative images of immunofluorescence staining show expression of tyrosine hydroxylase (TH), class III β‐tubulin (Tuj1), and nuclear DAPI in IPSC‐derived neurons. (B) Fluorescence‐activated cell sorting (FACS) analysis of neuronal culture revealed no difference in terms of percentage of dopaminergic neurons (TH+) among all neurons (Tuj1+) in control‐ and patient‐derived neuronal cultures after 30 days of maturation (n = 4). (C) Analysis of the neuronal network by comparison of neurite length (skeleton) and number of branchpoints shows similar complexity between control‐ and patient‐derived neurons. Values represent pixel count of skeleton or branchpoint normalized to the pixel count of the respective neuronal mask (n = 4). (D,E) Western blot analysis of VPS35 (D), VPS29 (E), and β‐actin (loading control) show no difference in protein levels between control (Control 1 and 2) and VPS35 D620N mutant (VPS35 clones 1_1 and 1_2) neurons under basal culture condition. Values normalized to Control 1 (n = 4–6). [Color figure can be viewed at wileyonlinelibrary.com]

p.D620N VPS35 Patient‐Derived Neurons Display Mitochondrial Dysfunction

To evaluate the mitochondrial network, we analyzed Z‐stack images from two controls and two clones of patient‐derived neurons stained with MitoTracker Green FM (Fig. 2A). Computational analyses revealed a decrease in mitochondrial size representative of fragmentation (Fig. 2B). Moreover, mitochondrial branching as a readout for connectivity within the mitochondrial network was impaired with a decreased average number of links (Fig. 2C) and nodes (Fig. 2D).

FIG. 2

Mitochondrial dysfunction in VPS35 mutant induced pluripotent stem cell (IPSC)‐derived neurons. (A) Representative image of mitochondria stained with MitoTracker Green FM and evaluation of mitochondrial size (B), average links per mitochondrion (C), and average nodes per mitochondrion (D) in control (Control 1 and 2) and VPS35 D620N mutant (VPS35 clones 1_1 and 1_2) neurons in culture medium without antioxidants (without B27 and ascorbic acid) for 24 hours (n = 4). (E) Mitochondrial membrane potential measured by tetramethylrhodamine ethyl ester (TMRE) mean fluorescence intensity and (F) mitochondrial reactive oxygen species measured by MitoSOX mean fluorescence intensity by flow cytometry in control (Control 1 and 2) and VPS35 D620N mutant (VPS35 clones 1_1 and 1_2) neurons in culture medium without antioxidants (without B27 and ascorbic acid) for 4 hours (n = 4). (G) Mean average oxygen consumption rate (OCR) of control (Control 1 and 2) and VPS35 D620N mutant (VPS35 clones 1_1 and 1_2) neurons over a time course. Measurement of basal OCR is followed by the addition of oligomycin (oligo) 2 μM final concentration, FCCP 250 nM final concentration, and antimycin A (AA) 5 μM final concentration and rotenone (rot) 5 μM final concentration (n = 4). (H) Calculated basal respiration, maximal respiration, spare respiratory capacity, proton leak, ATP production, and non‐mitochondrial oxygen consumption (n = 4). All statistical tests were Mann–Whitney tests to compare groups. Error bars show standard deviation and ns P > 0.05; *P < 0.05; **P < 0.01; ***P < 0.001. [Color figure can be viewed at wileyonlinelibrary.com]

The identification of these morphological alterations observed in patient‐derived neurons led us to assess the mitochondrial function. We measured the bioenergetic profile and found a decreased MMP in patient‐derived neurons compared to controls (Fig. 2E). The reduced MMP was accompanied by an increase of intra‐mitochondrial ROS compared to controls (Fig. 2F). Subsequently, mitochondrial respiration was assessed by recording the OCR while we applied mitochondrial stressors: oligomycin, FCCP, antimycin A, and rotenone to measure different respiratory parameters (Fig. 2G). We found that neurons carrying the p.D620N VPS35 mutation displayed a reduced basal and maximal respiration, reduced spare respiratory capacity, and non‐mitochondrial oxygen consumption (Fig. 2H). This was associated with a significantly reduced ATP production in patient‐derived neurons compared to controls.

VPS35 D620N Patient‐Derived Neurons Show an Impaired Mitochondrial Clearance

As patient neurons present morphologically and functionally altered mitochondria, we hypothesized that mitochondrial mass, biogenesis, and clearance (ie, mitophagy) might be dysregulated. We found no difference in mitochondrial mass between patient and control neurons as defined by western blotting against the mitochondrial proteins TOM20 (Fig. 3A) and VDAC1 (Fig. 3B). Moreover, protein expression levels of PGC1α (Fig. 3C), the master regulator of mitochondrial biogenesis, were unchanged in patient‐derived neurons under basal conditions.

FIG. 3

Mitophagy clearance impairment in VPS35 mutant induced pluripotent stem cell (IPSC)‐derived neurons after CCCP (carbonyl cyanide 3‐chlorophenylhydrazone) treatment. (A,B,C) Western blot analysis of TOM20 (A), VDAC (B), PGC1α (C), and β‐actin (loading control) of control (Control 1 and 2) and VPS35 D620N mutant (VPS35 clones 1_1 and 1_2) neurons under basal culture condition (n = 4). (D) Representative image of mitochondria and mitophagy events under CCCP treatment over a time course. The scale bar represents 20 μm. Calculated mitophagic events by mitochondria count in Control 1 (E) and mutant VPS35 1_2 (F) from three independent differentiations of control (Control 1) and VPS35 D620N mutant (VPS35 clone 1_2) neurons expressing ATP5C1‐RFP‐pHluorin protein in culture medium without antioxidants (without B27 and ascorbic acid) and treated with CCCP 10 μM for 0, 3, 8, and 24 hours. Each time point is compared with the previous one (n = 3). (G) Comparison of both lines. All statistical tests were Mann–Whitney tests or one‐way ANOVA followed by Sidak's multiple comparisons tests to compare groups and conditions. Error bars show standard deviation and ns P > 0.05; *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001. [Color figure can be viewed at wileyonlinelibrary.com]

In order to study mitophagy, Control 1 and VPS35 1_2 cell lines underwent CRISPR‐Cas9 gene engineering as iPSC‐derived smNPC to express a mitochondrial fusion protein: ATP5C1‐DsRed‐pHluorin. Briefly, when mitochondria are in the cytoplasm, both fluorophores are functional. Once mitochondria are exposed to an acidic environment inside the autophagosome (mitophagic event) the green fluorescence will be quenched (Fig. 3D). To induce mitophagic events, we treated the gene‐edited neurons with CCCP and acquired images from the same field of view at different time points: t = 0, 3, 8, and 24 hours. In control neurons (Fig. 3E), the number of mitochondria inside autophagosomes increased significantly after 3 hours CCCP treatment. After 8 hours and further after 24 hours of CCCP treatment, the number of mitochondria inside autophagosomes were decreasing, showing an efficient clearance. In the patient‐derived neurons harbouring the p.D620N VPS35 (Fig. 3F), the number of mitochondria inside autophagosome also increased significantly after 3 hours of CCCP treatment. After 8 hours and 24 hours, the number of mitochondria inside autophagosomes failed to decrease and stayed elevated, indicating a deficient clearance (Fig. 3G). Under basal conditions (t = 0), we see no significant differences in the number of mitochondria inside autophagosomes between patient and control neurons. In addition, after 3 hours of CCCP treatment there was no difference in the number of mitochondria inside autophagosomes between patients and controls (Fig. 3G), indicating that the induction of mitophagy was not impaired in patient neurons. However, the difference between patient and control cells becomes significant at 8 hours and 24 hours, as the number of mitochondria inside autophagosomes decreased in control‐derived neurons and stayed elevated in patient‐derived neurons.

Lysosomal Clearance Dysfunction and α‐Synuclein Accumulation in p.D620N VPS35 Patient‐Derived Neurons

The link between the retromer and macroautophagy has been identified by the sorting of ATG9, an important protein for induction of autophagy with the retromer. ²⁷ Additionally, in cells overexpressing the mutant p.D620N VPS35 it was shown that ATG9 was missorted, which is thought to lead to impaired autophagy. ⁸ Here, we measured the steady‐state level of the autophagy protein p62 (Fig. 4A), which was not differing between patient and control‐derived neurons. Upon treatment with Bafilomycin A1, which blocks autophagy by inhibiting the lysosomal v‐ATPase, p62 was accumulating in the controls, as shown by an increase compared to the untreated state. However, p62 did not significantly increase in the patient neurons after Bafilomycin A1 treatment, showing an impaired autophagic flux. Moreover, compared to control‐derived neurons, we found reduced Lamp1 and Lamp2 steady‐state protein levels in the patient neurons (Fig. 4B), suggesting a lower late‐endosome/lysosome mass compared to controls. This decrease in Lamp1 and Lamp2 levels was accompanied by the presence of smaller lysosomes in patient‐derived neurons, while the number of lysosomes was comparable between groups (Fig. 4C). This reduction in content and size of lysosome is implying an impaired functionality13 which is also revealed by a decreased intensity of the acidotropic probe, LysoTracker, in patient‐derived neurons when exposed to mild oxidative stress (antioxidant removal) (Fig. 4D) as already described in other genetic models of PD. ²⁸

FIG. 4

Impaired lysosomal clearance and α‐synuclein accumulation in VPS35 mutant induced pluripotent stem cell (IPSC)‐derived neurons. (A) Western blot analysis of p62 and β‐actin (loading control) in control (Control 1 and 2) and VPS35 D620N mutant (VPS35 clones 1_1 and 1_2) neurons under basal culture condition and Bafilomycin A1 (BafA1) 100 nM treatment for 24 hours (n = 5). (B) Western blot analysis of Lamp1, Lamp2, and β‐actin (loading control) in control (Control 1 and 2) and VPS35 D620N mutant (VPS35 clones 1_1 and 1_2) neurons under basal culture condition (n = 4). (C) Left: evaluation of lysosomal size and number by LysoTracker staining in control (Control 1 and 2) and VPS35 D620N mutant (VPS35 clones 1_1 and 1_2) neurons under basal culture condition (n = 3). Right: representative images of lysosomes stained with LysoTracker Deep Red (white). Nucleus are stained with Hoechst. (D) Evaluation of LysoTracker Deep Red staining intensity in control (Control 1 and 2) and VPS35 D620N mutant (VPS35 clones 1_1 and 1_2) neurons under basal culture condition and mild stress (antioxidant removal) (n = 3). (E) Western blot analysis of α‐synuclein (α‐syn) and β‐actin (loading control) in control (Control 1 and 2) and VPS35 D620N mutant (VPS35 clones 1_1 and 1_2) neurons under basal culture condition (n = 4). (F) Time‐resolved fluorescence energy transfer (TR‐FRET) measurement of total α‐synuclein (α‐syn) amount in control (Control 1 and 2) and VPS35 D620N mutant (VPS35 clones 1_1 and 1_2) neurons under basal culture condition (n = 4). All statistical tests were Mann–Whitney tests or one‐way ANOVA followed by Sidak's multiple comparisons tests to compare groups and conditions. Error bars show standard deviation and *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001. [Color figure can be viewed at wileyonlinelibrary.com]

The impaired lysosomal clearance was accompanied by an increase of the amount of α‐synuclein in patient‐derived neurons as demonstrated by western blotting for the monomeric form of α‐synuclein (Fig. 4E) and validated by time‐resolved fluorescence energy transfer (TR‐FRET) assessing total α‐synuclein amount (Fig. 4F).

α‐Synuclein Accumulation is not the Main Cause of Mitochondrial Impairment in pD620N VPS35 Neurons

In patient‐derived neurons carrying a triplication of the SNCA gene locus, extensive mitochondrial defects are found. ²⁹ , ³⁰ Consequently, we hypothesized that the mitochondrial impairment seen in our patient‐derived neurons could be due to both a general impairment of the autophagy machinery and to α‐synuclein accumulation. Therefore, we used a pharmacological approach to rescue autophagy with rapamycin. ³¹ Rapamycin has been shown to increase lysosomal biogenesis and to enhance mitophagy in cellular models of PD. ³² , ³³ After 24 hours of treatment, Lamp1 levels were increased in both control‐ and patient‐derived neurons (Fig. 5A) and this led to a moderate decrease of α‐synuclein levels in patient‐derived neurons (Fig. 5B). We evaluated the effect of this treatment on mitochondrial dysfunction but did not observe a rescue of the decreased MMP nor of the increased ROS levels in patient‐derived lines (Fig. 5C,D). Of note, mitochondrial mass was not affected (Fig. 5E), which shows that rapamycin did not enhance basal mitophagy in these experimental conditions. We hypothesized that a more efficient decrease of α‐synuclein levels might influence the mitochondrial impairment. We knocked‐down α‐synuclein in both control‐ and patient‐derived neurons with shRNA against SNCA. ³⁴ After transduction, we detected a reduction of the levels of α‐synuclein in patient‐derived neurons to the physiological levels of α‐synuclein (Fig. S5A,B). Using the same technique as previously, we found that the reduction of α‐synuclein protein levels did not rescue the loss of MMP (Fig. 5F) nor the increased ROS level (Fig. 5G).

FIG. 5

Autophagy enhancement and α‐synuclein knockdown are not sufficient to alleviate mitochondrial dysfunction. (A,B) Western blot analysis of Lamp1 (A) and α‐synuclein (B) of control and VPS35 D620N mutant neurons after 24 hours of treatment with DMSO (50 nM) or rapamycin (25 or 50 nM) (n = 5). Values are normalized to DMSO‐treated line respectively. (C) Mitochondrial membrane potential measured by tetramethylrhodamine ethyl ester (TMRE) mean fluorescence intensity, (D) mitochondrial reactive oxygen species measured by MitoSOX mean fluorescence intensity, and (E) mitochondrial mass measured by MitoTracker Green FM mean fluorescence intensity by flow cytometry in control and VPS35 D620N mutant neurons after 24 hours of treatment with DMSO (50 nM) or rapamycin (25 or 50 nM) (n = 5). Values are normalized to DMSO‐treated Control 1. (F) Mitochondrial membrane potential measured by TMRE mean fluorescence intensity and (G) mitochondrial reactive oxygen species measured by MitoSOX mean fluorescence intensity by flow cytometry in control and VPS35 D620N mutant neurons transduced with scramble shRNA and shRNA against α‐synuclein (n = 3). Values are normalized to scramble transduced line respectively. All statistical tests were Mann–Whitney tests or one‐way ANOVA followed by Sidak's multiple comparisons tests to compare groups and conditions. Error bars show standard deviation and *P < 0.05; **P < 0.01. [Color figure can be viewed at wileyonlinelibrary.com]