Too Much PLEKHM1 Protein Disrupts Cell Cleanup and Worsens Brain Cell Damage in Mice with Alpha-Synuclein

The AAV transfer plasmid hPlekhm1-dsRedM, which provides the DNA sequence to be packaged into the rAAV capsid, was a gift from Paul Odgren and obtained from Addgene (Plasmid #73592). The dsRedM ORF at the C-terminus of PLEKHM1 was replaced with a hemagglutinin epitope tag (HA-tag). Incorporating a HA-tag enables straightforward detection of the expressed PLEKHM1 protein. The high-fidelity restriction enzymes NotI-HF (cat. no. R3189, New England Biolabs, Ipswich, MA, USA) and KpnI-HF (cat. no. R3142, New England Biolabs) were used. Restriction fragments were separated by agarose gel electrophoresis and excised from the gel using the QIAquick Gel Extraction Kit (Qiagen, cat. no. 28704, Hilden, Germany). Annealed oligonucleotides encoding the HA tag were phosphorylated at the 5′ end using T4 polynucleotide kinase (cat. no. M0201SVIAL, New England Biolabs). Ligation was performed utilizing T4 DNA ligase (New England Biolabs) in '10x 4 Ligation NEB Buffer' (New England Biolabs). The resulting plasmid was amplified in DH5-alpha bacteria (cat. no. 18263012, Invitrogen, Waltham, MA, USA) under ampicillin selection. Next, the PLEKHM1-HA-tag sequence was cut out from the newly generated plasmid using Nhel-HF (cat. no. 3131, New England Biolabs) and Notl-HF (cat. no. R3189, New England Biolabs). The fragment was inserted into the AAV-backbone plasmid using the Quick Ligation Kit (cat. no. M2200S, New England Biolabs). The final rAAV transfer plasmid contains a CMVie-enhanced synapsin1 promoter, a WPRE sequence and bovine growth hormone polyadenylation sequence flanked by inverted terminal repeats.

rAAV-encoding A53T-αSyn was generated by the Leuven Viral Vector Core, as described previously [21,22]. rAAVs encoding PLEKHM1 and EGFP were generated at the Viral Core Facility at Charité Universitätsmedizin Berlin. The pseudotyped rAAV2/7 vector was selected for its high efficiency in neuronal transduction. For each rAAV used, genomic copies (GC) were determined as technical titers by real-time PCR analysis. Each rAAV was diluted to a concentration of 3 × 10¹² GC/mL, and aliquoted and stored at −80 °C until further use as described previously [21,22]. Two microliters of a 1:1 mixture of the viral vectors were stereotactically injected. The rAAV-EGFP control group received only rAAV-EGFP, but the total number of GC matched that used in the other groups.

All animal experiments were conducted in compliance with EU Directive 2010/63 on the protection of animals used for scientific purposes and were approved by the Animal Research Committee of Landesdirektion Sachsen, Dresden, Germany (25-5131/49615) on 5 May 2020.

All mice were housed under a 12 h light and dark cycle with free access to pellet food and water in the Experimental Center, TU Dresden, Dresden, Germany.

For the first experiment, twelve-week-old C57BLl/6 male mice were obtained from Janvier Labs (Le Genest Saint Isle, France). We included five animals per group, except for the aSyn + PLEKHM1 group, which comprised six animals. The one extra mouse was initially included as a precaution against potential loss but was ultimately not required. Sample sizes were determined based on our prior studies with the goal of minimizing animal use while ensuring sufficient statistical power. In the second experiment involving transgenic mice, six male RFP-EGFP-LC3 transgenic mice were used; initially developed by Li et al. [23] and obtained from the Jackson Laboratories (RRID: IMSR_JAX:027139; Jackson Laboratories, Bar Harbor, ME, USA). In total, 27 mice were included in this study. All animals designated for the experiments were included in the analysis; thus, no specific inclusion or exclusion criteria were applied.

Cages were randomly assigned to receive one of the rAAV treatments. To reduce potential confounding variables, rAAV injections were performed on four consecutive days, thereby minimizing the time difference between key experimental groups. Perfusions were completed within a single day to further control for variability. Staining protocols were standardized, and animals were randomly selected for staining since processing could not be completed in a single session. Imaging parameters were defined once per epitope and remained constant throughout the study. One investigator (D.G.-L.) as well as lab technicians were fully blinded to group assignments during key steps, including the sectioning, staining, imaging, and quantification of dopaminergic neurons. Further image analyses were conducted using automated and independent analysis tools to ensure objectivity (see paragraphs below).

Unilateral stereotactic surgeries were conducted under aseptic conditions. Mice were anesthetized with intraperitoneal injections of Ketamine (100 mg/kg) and Xylazine (10 mg/kg), and cooling of the body temperature was prevented using a heated cushion. The animals were positioned in a flat skull orientation within a stereotactic head frame, and a bore-hole craniotomy was performed using the following stereotactic coordinates relative to Bregma: anterior–posterior −3 mm, lateral −1.2 mm and dorso-ventral −4.1 mm. Following a manual incision of the dura, a Hamilton glass syringe (Hamilton Bonaduz AG, Bonaduz, Switzerland) was used to inject 2 µL of the vector solution. The needle was advanced at a rate of 300 µm/min until reaching the final coordinates, where it was paused for five minutes both before and after the vector injection. Two hundred mL viral vector solutions were injected per minute. The syringe was slowly withdrawn and the skin was sutured. Lidocaine (Aspen Pharmacare, Durban, South Africa) was topically applied and Metamizole (WDT, Garbsen, Germany) was added to the drinking water.

Eight weeks after virus injection, animals were sacrificed with an overdose of isoflurane (Baxter, Lessines, Belgium). Subsequently, transcardial perfusion was performed using 4% paraformaldehyde (PFA) which was diluted in Tris-buffered saline (TBS, pH 7.6). For further fixation, the brains were immersed in 4% PFA for 48 h at 4 °C, followed by cryoprotection in TBS containing 30% sucrose. Next, the tissue was frozen in isopentane at −55 °C and stored at −80 °C until it was sectioned into 30 µm thick coronal brain slices using a cryostat (Leica Biosystems, Nussloch, Germany).

Coronal brain sections were stained using primary antibodies, fluorophore-linked or enzyme-linked secondary antibodies. The following antigens were visualized: tyrosine hydroxylase (TH), human αSyn (H-αSyn), phosphorylated αSyn at Serine 129 (P-αSyn), glial fibrillary acidic protein (GFAP), ionized calcium-binding adapter molecule 1 (Iba1), lysosomal-associated membrane protein 1 (LAMP1), p62 (SQSTM1), and the HA-tag added to pleckstrin homology and RUN domain-containing protein 1 (PLEKHM1). For immunofluorescence staining, sections were rinsed three times for 10 min in TBS, followed by incubation in a blocking solution for 1 h at room temperature. The blocking solution contained 10% donkey serum (BIOZOL Diagnostica Vertrieb GmbH, Eching, Germany) and 0.2% Triton X-100 (Thermo Scientific, Waltham, MA, USA) in TBS. Subsequently, the sections were incubated for 20 h at 4 °C with combinations of the following primary antibodies: rabbit anti-TH (1:1000, P40101, Pel-Freez, Rogers, AR, USA), rabbit anti-P-αSyn (1:1000, ab51253, Abcam, Cambridge, UK), rat anti-H-αSyn (1:1000, ALX-804-258-L001, ENZO Life Sciences, Farmingdale, NY, USA), chicken anti-GFAP (1:500, ab4674, Abcam), guinea pig anti-Iba1 (1:2000, HS-234308, Histo Sure, Göttingen, Germany), rabbit anti-p62 (1:500, ab56416, Abcam), and LAMP1 (1:600, ab208943, Abcam). After another three 10 min rinses, the sections were treated with fluorophore-conjugated secondary antibodies for 1 h at 21 °C: CF555-conjugated donkey anti-guinea pig (1:1000, SAB4600298, Sigma-Aldrich, St. Louis, MO, USA), Rhodamine red-conjugated donkey anti-rabbit (1:1000, 711-295-152, Jackson ImmunoResearch, Ely, UK), Alexa 647 donkey anti-rat (1:1000, ab150155, Abcam), Alexa 647-conjugated donkey anti-rabbit (1:1000, A31573, Molecular Probes, Eugene, OR, USA), and Alexa 405-conjugated goat anti-chicken (1:500, A48260, Invitrogen). Nuclei were counterstained with Hoechst (1:2000 for 5 min; Invitrogen, Waltham, MA, USA) and the sections were mounted in Fluoromount-G (Invitrogen).

The HA-tagged PLEKHM1 was visualized using an enzyme-linked staining protocol based on the horseradish peroxidase reaction using Vector VIP Peroxidase Substrate (SK-4600; Vector Laboratories, Newark, CA, USA). Three sections of each animal, that was transduced with rAAV-PLEKHM1, were rinsed in TBS for ten minutes. The quenching of endogenous peroxidase was performed in H₂O₂ (3%) for 15 min. After additional rinses, the sections were incubated in 10% donkey serum for 1 h. The primary antibody, rabbit anti-HA (1:1000, C29F4, Cell Signaling Technology, Danvers, MA, USA), was added for 18 h at 4 °C. Next, sections were rinsed three times and incubated in HRP-conjugated donkey anti-rabbit secondary antibody (1:1000, 711-035-152, Jackson Immunoresearch). The sections were rinsed again. The reagent solution was prepared according to the manufacturer's protocol applied onto the sections for 7.5 min. The tissue underwent one more rising cycle, was dehydrated in ethanol and finally mounted in Euparal (Waldek GmbH & Co. KG, Münster, Germany).

Every fourth section of the SN was stained for TH (approximately 10 sections per animal). We acquired the slide scans of each coronal brain section using a Zeiss Axio Observer.Z1 Inverted Microscope (Zeiss, Oberkochen, Germany) supported by a Yokogawa CSU-X1 unit (Yokogawa Life Science, Musashino-shi, Tokyo, Japan), enabling confocal imaging. It employed a 20×/0.8 NA objective lens for imaging. Seven Z-levels were imaged at 1 µm intervals and subsequently processed as maximum intensity projections using Zeiss Zen 3.1 software (Zeiss). An investigator blinded to the treatment groups manually quantified the number of TH-positive neurons in the SNc using the Zeiss Zen 3.1 software. The total number of TH-positive neurons per hemisphere was counted, and the sum in the treated hemisphere was expressed relative to that of the untreated hemisphere.

The SNr contains dopaminergic dendrites that belong to dopaminergic neurons in the SNc. We quantified the density of dopaminergic dendrites in the SNr using the above-mentioned slide scans. Image analysis was performed using the open-source software QuPath (version 0.5.1) [24]. The SNr was encircled in both hemispheres as an area of interest. To train the neural network-based pixel classifier of QuPath, TH-positive structures within the SNr were manually annotated in a subset of randomly selected sections. Next, the TH-positive area for each area of interest was determined utilizing the pixel classifier. The mean TH-positive area in the SNr per hemisphere was calculated for each mouse. The results from the treated hemisphere were normalized to the corresponding values from the non-treated hemisphere.

Dopaminergic neurons extend their axons into the striatum. To assess the integrity of dopaminergic axon terminals within the striatum, we acquired 10 to 15 high-magnification images (40×/0.95 objective) from two to three dorsal striatum-containing sections per mouse using a spinning disk confocal microscope. The imaging system consisted of a Zeiss Axio Observer.Z1 Inverted Microscope (Zeiss) equipped with a Yokogawa CSU-X1 unit (Yokogawa Life Science, Musashino-shi, Tokyo). Image analysis was performed in CellProfiler (version 4.2.4; Stirling et al., 2021 [25]). First, we used the "Enhance Neurites" function to define the borders of linear structures (e.g., axons) more clearly and applied a threshold to the image using the "robust background" algorithm. For each image, we calculated the ratio of positive pixels to the total image size. We calculated the mean value for each hemisphere of each animal. The results for each animal were then expressed as the ratio of the injected hemisphere to the contralateral hemisphere.

Every fourth section spanning the entire SN (approx. 10 sections per animal) was stained for either H-αSyn in combination with P-αSyn and a nuclei counterstain, or for TH, Iba1and GFAP. Slide scans were acquired as described above. The SNc in both hemispheres was encircled as a region of interest in QuPath. The cell detection function was applied using default parameters, enabling the identification of nuclei based on Hoechst staining. To train the built-in object classifier algorithm for detecting cells positive for H-αSyn or P-αSyn, we manually annotated cells in a subset of slices. The entire SNc was then analyzed using the QuPath algorithm. The number of H-αSyn- and P-αSyn-positive cells per animal was summed for each hemisphere and then multiplied by four to account for the fact that every fourth section was analyzed. For the analysis of microgliosis and astrogliosis, we could not use nuclear staining as a basis for further analysis because all channels were used for the analysis of Iba1, GFAP, TH, and rAAV-EGFP. We therefore needed to modify our approach. The Iba1-positive area and the GFAP-positive area in the SNc were quantified by QuPath's pixel classifier after training as outlined for αSyn. The relative Iba1-positive area and GFAP-positive area was quantified for each section. The values were summarized by calculating the mean for each hemisphere of each animal. Finally, for each animal, the mean value from the treated hemisphere was normalized to that of the corresponding non-treated hemisphere.

In order to study the effect of PLEKHM1 overexpression on intracellular vesicles, we transduced RFP-EGFP-LC3 mice with rAAV-PLEKHM1. The RFP-EGFP-tagged LC3 enables pH-sensitive discrimination between autophagosomes and autolysosomes. EGFP fluorescence is quenched at acidic pH, resulting in autophagosomes appearing EGFP-positive and RFP-positive, while autolysosomes lose EGFP fluorescence and retain RFP fluorescence. Mature lysosomes are LAMP1-positive. We transduced six autophagy reporter mice and analyzed four sections per animal using the spinning disk confocal microscope described above, equipped with a 40×/0.95 NA objective. For each section, we acquired 15 to 20 images of the SNc in both hemispheres. Individual cells were cropped using Fiji (v2.16.0; [26]) and channels were split for further analysis. Cropped images depicting individual cells (n = 1239 cells) were processed using CellProfiler's "enhance speckles" function and thresholded using the "robust background method". The resulting binary images were processed with one opening step to minimize salt-and-pepper noise.

The area that represents lysosomes (LAMP1-positive, RFP-negative) was determined by the subtraction of binary images: RFP-positive pixels were subtracted from LAMP1-positive pixels, thus resulting in a binary image showing only those pixels that are LAMP1-positive and RFP-negative. Using the same approach, autolysosomes were defined by the subtraction of EGFP-positive pixels from the RFP-positive area. Autophagosomes were defined as EGFP and RFP-positive pixels. For each image, the number of positive pixels was divided by the total image area.

Using the same imaging, processing, and analysis workflow described above, the area covered by p62-positive structures was quantified in wild-type mice transduced with rAAV-EGFP or rAAV-PLEKHM1 (n = 1604 images depicting cells). In addition, autophagy reporter mice (RFP-EGFP-LC3) were analyzed (n = 1482). Signal intensity was also measured in cropped images depicting individual cells.

Colocalization analysis was performed using CellProfiler's colocalization tool, employing default settings.

Data analysis and visualization were conducted in R (version 4.2.0) using RStudio ("Spotted Wakerobin" Release, version 2022.07.1+554) with the following packages: Tidyverse (version 2.0.0) for data wrangling and visualization, and Ggpubr (version 0.6.1) and FSA (version 0.10.0) for statistical testing [27,28,29]. For data that followed normal distribution, we used Welch's two-sample t-test or a two-way ANOVA, followed by post hoc analyses with Tukey's HSD or Dunn–Bonferroni tests. Pairwise comparisons that did not satisfy the aforementioned criteria were evaluated using the Wilcoxon rank sum exact test. For data involving more than two groups, the Kruskal–Wallis Chi-squared test was applied, followed by post hoc pairwise comparisons using Dunn's Kruskal–Wallis multiple-comparison test with Benjamini–Hochberg adjustment. Correlations were calculated using Pearson's product–moment correlation. Error bars represent the standard error of the mean. The p-values are listed in the figure legends. Graphical illustrations were created using Draw.io (v25.0.1., JGraph Ltd., Northampton, UK) and Microsoft PowerPoint 2021 (Microsoft Corporation, Redmond, WA, USA).

We used stereotactic injections of rAAV under the neuronal synapsin1 promoter to express PLEKHM1 in the substantia nigra (SN; Figure 1A). Transgene expression was confirmed by immunohistochemical staining for the HA-tag of PLEKHM1 (Figure 1B), demonstrating robust expression within the targeted region. As a negative control, an rAAV-encoding EGFP was used, and EGFP expression was verified by fluorescence microscopy (Figure 1C). To induce αSyn pathology, we similarly injected rAAV expressing the A53T mutant form of αSyn under the same neuronal promoter. Human αSyn expression was detected via immunostaining and confirmed transgene expression in the cells of SN (Figure 1D,E). This approach allowed us to evaluate the effects of PLEKHM1 overexpression on αSyn pathology within the SN. Pathological intracellular αSyn aggregates are commonly hyperphosphorylated at serine 129 (P-αSyn; [30]). This post-translational modification is an established marker for αSyn pathology [31,32]. P-αSyn-positive cells were detected in both cohorts transduced with rAAV-αSyn (rAAV-αSyn + rAAV-PLEKHM1 and rAAV-αSyn + rAAV-PLEKHM1), confirming the occurrence of αSyn pathology (Figure 1D,E). Based on previous results, we hypothesized that the overexpression of PLEKHM1 alters autophagy and subsequently changes the clearance of aSyn aggregates. Nevertheless, the number of P-αSyn-positive cells was similar in animals transduced with PLEKHM1 (rAAV-αSyn + rAAV-PLEKHM1) as in animals transduced with EGFP (rAAV-αSyn + rAAV-EGFP) (Figure 1G). The number of cells transduced with αSyn was also similar, as evidenced by the number of H-αSyn-positive cells (Figure 1F).

To histologically assess the dopaminergic system, we performed staining for tyrosine hydroxylase (TH). When the SN was transduced with EGFP, the number of TH-positive neurons in the SNc (red in Figure 2A) was unchanged compared to the uninjected hemisphere (Figure 2B), suggesting that viral transduction alone did not cause neurodegeneration. Transduction with PLEKHM1 (rAAV-EGFP + rAAV-PLEKHM1) or αSyn (rAAV-EGFP + rAAV-αSyn) reduced the number of TH-positive neurons in some animals, although this effect was not statistically significant at the group level (Figure 2B and Figure S1). Mice transduced with both αSyn and PLEKHM1 (rAAV-αSyn + rAAV-PLEKHM1) showed significantly fewer TH-positive neurons in the SN than the controls (rAAV-EGFP), or mice transduced with αSyn and EGFP (rAAV-αSyn + rAAV-EGFP) (Figure 2B). This observation can be explained by the hypothesis that PLEKHM1 potentiates the toxicity of αSyn.

The dendrites of dopaminergic neurons extend long apical dendrites into the SNr (orange in Figure 2A). We, therefore, measured the density of TH-positive dendrites in the SNr. In contrast to the analysis of neuron number, the expression of αSyn alone was sufficient to cause a significant reduction in the density of TH-positive dendrites in the SNr (Figure 2C). With the transduction of αSyn and PLEKHM1, the density of TH-positive dendrites in the SNr was also reduced as compared to the controls, but it was not significantly lower than with αSyn and EGFP (Figure 2C). We, therefore, interpret this finding as mainly representing αSyn toxicity.

The integrity of dopaminergic axon terminals was assessed by measuring the density of TH-positive structures in the striatum using images acquired at high magnification (Figure 2D). Regarding the neuron numbers in the SNc (Figure 2B), the co-expression of αSyn and PLEKHM1 reduced the density of TH-positive axon terminals in the striatum, whereas the individual expression of either αSyn or PLEKHM1 did not (Figure 2E).

The degeneration of dopaminergic somata, axons, and dendrites is closely correlated (Figure 2F,G). The loss of dendrites and cell somata is more pronounced than axonal degeneration, which may be explained by axonal sprouting compensating for striatal degeneration.

Neuroinflammation is a key aspect of synucleinopathies. Microglial cells are the resident immune cells in the central nervous system and can convey protective and toxic effects [33,34,35,36]. In order to investigate microglia activation in our model, we stained for the established microglia marker ionized calcium-binding adapter molecule 1 (Iba1; Figure 3A).

The Iba1-positive area in the SNc was increased two-fold compared to the uninjected hemisphere in all groups (Figure 3B), suggesting a microglial response to viral transduction. In mice transduced with αSyn and PLEKHM1, the increase in Iba1-positive area was even greater than in control mice injected with the same number of viral particles (Figure 3B). Accordingly, we observed a negative correlation between the Iba1-positive area in the SNc and the density of TH-positive axon terminals in the striatum (Figure 3C). This correlation might be caused by the toxic effects of microglial activation, or by microglial activation downstream of neurodegeneration.

Activation of astroglia was quantified by staining for the intracellular intermediate filament protein 'glial fibrillary acidic protein' (GFAP; Figure 3D). The GFAP-positive area in the SNc was higher than in the uninjected hemisphere for all treatment groups; there was no significant difference in group-wise comparisons (Figure 3E). Nevertheless, we did observe a significant correlation between the GFAP-positive area in the SNc and the density of TH-positive dendrites in the SNr (Figure 3F).

To better understand the impact of PLEKHM1 overexpression on the autophagy-lysosomal system, we used RFP-EGFP-LC3 reporter mice (Figure 4A). These mice express the autophagosome marker, microtubule-associated protein 1A/1B-light chain 3 (LC3), fused to tandem RFP-EGFP fluorescence under the control of a CAG promoter [23]. The RFP-EGFP-tagged LC3 enables pH-sensitive discrimination between autophagosomes and autolysosomes. EGFP fluorescence is quenched at acidic pH, resulting in autophagosomes appearing EGFP-positive and RFP-positive, while autolysosomes lose EGFP fluorescence and retain RFP fluorescence (Figure 4B). We used pH-sensitive tandem fluorescence previously in vitro [37].

Neurons in the SN of RFP-EGFP-LC3 reporter mice were transduced by stereotactic injection of rAAV-PLEKHM1. In addition to the RFP-EGFP-LC3 fluorescence, we stained for the lysosomal marker LAMP1 and the nuclear marker DAPI (Figure 4C). Transduced cells in the SNc were compared to cells in the non-treated hemisphere using confocal microscopy. The area positive for RFP and LAMP1 in the SNc was reduced following rAAV-PLEKHM1 transduction, whereas the area positive for EGFP remained unchanged (Figure 4D).

We then examined these findings in more detail. Autophagosomes were defined as being EGFP-positive and RFP-positive. We did not observe a relevant difference between cells with and without rAAV-PLEKHM1 transduction (Figure 4E). Autolysosomes were defined as being RFP-positive and EGFP-negative. The area of autolysosomes was significantly decreased in cells with rAVV-PLEKHM1 transduction compared to the uninjected hemisphere (Figure 4F). Lysosomes were defined as RFP-negative and LAMP1-positive. The area of lysosomes was lower in cells with rAVV-PLEKHM1 transduction as compared to the uninjected hemisphere (Figure 4G). Finally, we noted an increase in the colocalization coefficient for LAMP1 with RFP (Figure 4H) and a decrease in the ratio of autolysosomes to autophagosomes (Figure 4I). In summary, the area corresponding to acidic vesicles (lysosomes and autolysosomes) was decreased, but the colocalization of LAMP1 with RFP was increased.

In the next step, we investigated the functional consequences of the observed changes in wild-type mice, as well as the autophagy reporter mice. To this end, we utilized p62 (also known as SQSTM1) as a marker (Figure 5A,B) [38]. p62 is a selective autophagy adaptor protein that serves as a surrogate marker for autophagic flux. It is continuously degraded under normal autophagic conditions and accumulates when autophagy is impaired.

p62 signal intensity showed a mild increase in rAAV-EGFP-transduced hemispheres compared to the untreated, contralateral hemisphere, indicating that viral transduction can modestly elevate baseline p62 levels (Figure 5C). Crucially, rAAV-PLEKHM1 transduction resulted in both genotypes in a pronounced, up to threefold increase in p62 intensity (Figure 5C,D). Next, we quantified the area per image in the SNc that was positive for p62 staining across both genotypes (Figure 5E,F). This area was comparable between untreated hemispheres and those injected with the control vector (rAAV-EGFP), suggesting that baseline p62 distribution remains largely unaffected by rAAV transduction alone. However, in both wild-type and transgenic autophagy reporter mice, rAAV-PLEKHM1 transduction led to a significant reduction in the p62-positive area. These findings support our visual observations of a shift in p62 staining from a diffuse cytoplasmic pattern to a markedly brighter and more punctate, irregular distribution upon PLEKHM1 overexpression.

Furthermore, we analyzed the colocalization of p62 with LAMP1 in wild-type mice, as well as the colocalization of p62 with LC3 in transgenic mice (Figure 5G–I). These analyses revealed an increase in p62-LAMP1 colocalization in wild-type mice transduced with rAAV-EGFP compared with the untreated hemisphere, which was markedly reduced following Plekhm1 overexpression. In autophagy reporter mice, p62 colocalization with autophagosomes (LC3EGFP+/RFP+) was significantly decreased, whereas colocalization with autolysosomes (LC3EGFP−/RFP+) was increased.