Mitochondrial dysfunction–induced H3K27 hyperacetylation perturbs enhancers in Parkinson’s disease

Mitochondrial dysfunction is a major pathophysiological contributor to the progression of Parkinson’s disease (PD); however, whether it contributes to epigenetic dysregulation remains unknown. Here, we show that both chemically and genetically driven mitochondrial dysfunctions share a common mechanism of epigenetic dysregulation. Under both scenarios, lysine 27 acetylation of likely variant H3.3 (H3.3K27ac) increased in dopaminergic neuronal models of PD, thereby opening that region to active enhancer activity via H3K27ac. These vulnerable epigenomic loci represent potential transcription factor motifs for PD pathogenesis. We further confirmed that mitochondrial dysfunction induces H3K27ac in ex vivo and in vivo (MitoPark) neurodegenerative models of PD. Notably, the significantly increased H3K27ac in postmortem PD brains highlights the clinical relevance to the human PD population. Our results reveal an exciting mitochondrial dysfunction-metabolism-H3K27ac-transcriptome axis for PD pathogenesis. Collectively, the mechanistic insights link mitochondrial dysfunction to epigenetic dysregulation in dopaminergic degeneration and offer potential new epigenetic intervention strategies for PD.


Mitochondrial membrane potential, morphology, and superoxide production
The JC-1 mitochondrial potential sensor was purchased from Invitrogen and the standard commercial protocol was followed. Briefly, 2.0 μg/mL of JC-1 diluted in serum-free N27 media was added to N27 cell cultures and incubated at 37°C for 20 min. Following gentle, triple washes with PBS, images were immediately captured on the Keyence microscope (Itasca, IL) before cells dried out. The ratio of red to green was calculated. For MitoTracker staining, procedures were similar. Following the treatment paradigm, 300 μL of 166 nM CMXROS MitoTracker red dye diluted in serum-free N27 media was added and incubated at 37°C for 12 min. After triple-washing with PBS, cells were fixed in a 4% solution of paraformaldehyde (PFA) for 30 min prior to following the steps for ICC and 3D imaging. For MitoSox, live N27 cells were stained following the manufacturer's instructions to detect superoxide production.

ICC
For ICC, 4% PFA was used to fix N27 DAergic neuronal cells. After a double-wash, fixed cells were blocked and then incubated in primary antibodies following manufacturers' protocols. Following primary antibody (H3K27ac 1:500, Cell Signaling Technology) incubation, cells were washed with PBS and incubated in Alexa dye-conjugated secondary antibody. Next, cells were washed and mounted on slides using Fluoromount aqueous mounting medium (Sigma). Cells were visualized using an inverted fluorescence microscope (Nikon TE-2000U).

Acute midbrain slice
Organotypic slices were prepared as previously described with several steps improved to adapt to a PD disease model (86). The slice culture buffer comprised Gey's balanced salt solution supplemented with the excitotoxic antagonist kynurenic acid (GBSS). WT (C57BL/6) postnatal mouse pups (7-to 12-days old) were used for live brain slices. The microtome's (Compresstome VF-300, Precisionary Instruments) slicing reservoir was prefilled with ice-cold sterile GBSS. The dissected midbrain was oriented in the sagittal plane in the Compresstome's specimen tube with 2% liquid agarose inside. After quickly solidifying the agar, the specimen tube was inserted into the slicing reservoir. Nigrostriatal slices (200 μm) were transferred to 6-well plates with oxygenated GBSS. The plates were preheated to 37°C in a humidified atmosphere of 5% CO2. After a 1-to 2-h recovery, acute slices were treated with 25 µM rotenone for 3 h. Afterward, slices were gently double-washed in ice-cold PBS, temporarily stained with propidium iodide to ensure viability and homogenized to prepare lysates for Western blotting or RNA for qRT-PCR. Slices were exposed to rotenone treatment in 6-well plates in an incubator maintained at 5% CO2 and 37°C. Treatments were performed in serum-free RPMI media supplemented with penicillin (100 U/mL), streptomycin (100 ug/mL), and 2 mM L-glutamine. After a 3-h treatment, slices were washed with ice-cold PBS and harvested.  Approximately 11.5% of H3K27ac peaks were significantly changed after rotenone treatment with 2.79% and 14.36% of them located at annotated genes (proximal promoters) and distal promoters, respectively. Upon TFAM knockout, 8.0% of H3K27ac peaks were significantly changed with 2.82% and 9.63% of them at annotated genes (proximal promoters) and distal promoters, respectively.

Supplementary Figure 6. Elevated H3K27ac in substantia nigra (SN) of MitoPark mice (MP) and
in midbrain slices exposed to rotenone ex-vivo. (A) Representative immunoblots for H3K27ac in rotenone-exposed midbrain slices coupled with their densitometric analyses for the (B) SN and (C) striatum. Each data point is the average of three replicates. At least four independent experiments were measured for SN and two for striatum. (D) Representative immunoblot for H3K27ac in the SN (n=11-13) and striatum (n=7) from 16-to 20-week-old littermate controls (LCs) and MPs; β-actin is the internal control. The immunoblot quantification of H3K27ac is shown for (E) the SN and (F) striatum.
(G) IHC showing TH degradation for the entire SN region from 16-to 20-week-old LCs and MPs with H3K27ac in green, nucleus stained blue with Hoechst and tyrosine hydroxylase (TH) in red. Scale bar, 50 µm. Independent experiments were repeated four times. Bar graphs show mean ± SEM of unpaired two-tailed t-tests. ns, not significant; *p≤0.05; **p<0.01.

Supplementary Figure 7.
All littermate (LC) and MitoPark (MP) mice used in this study were genotyped and further characterized by behavioral phenotyping assays. Motor deficits in MP mice depicted as (A) representative track plots and heat maps generated in VersaPlot, (B) group horizontal activity, vertical activity and distance traveled, (C) representative horizontal movement velocity, and (D) group RotaRod activity. Bar graphs show mean ± SEM of unpaired two-tailed t-tests. ns, not significant; *p≤0.05; **p<0.01; ***p<0.001.