Methodological validation of Miro1 retention as a candidate Parkinson's disease biomarker

Abstract

Mitochondrial markers help stratify Parkinson's disease (PD) patients. We use high-throughput blotting to quantify Miro1, Mfn2, and VDAC levels in fibroblasts, blood cells, and iPSC-derived neurons. Miro1 is specifically retained in PD cells but degraded in healthy ones after mitochondrial depolarization. We correlate Miro1 retention scores with pathogenic mutations, genetic background, age, and clinical data. This scalable assay and quantifiable score for mitochondrial-PD support biomarker development and pharmacological screening.

Fig. 1. Significantly more retention of Miro1 in CCCP-treated fibroblasts from PD patients compared to healthy controls.

A–C Whole-cell lysates from fibroblasts derived from (A) young healthy controls (HC), (B) Parkinson´s disease (PD) patients from the National Institute of Neurological Disorders and Stroke (NINDS) cohort and (C) an independent cohort of healthy individuals and PD patients from the Hertie Institute biobank (Tübingen cohort) were analyzed using simple western blotting (JESS, Bio-Techne). The mean age at sampling (AAS) and age at onset (AAO) ± standard deviation (SD) for each group are indicated. Fibroblasts from each cohort were seeded and treated simultaneously with either DMSO (vehicle control, -) or 40 µM CCCP for 6 h to induce mitochondrial depolarization. Miro1 protein levels were assessed under depolarizing conditions alongside control proteins, including mitochondrial markers (Mfn2 and VDAC) and the cytosolic housekeeping protein GAPDH (loading control). Protein intensities were quantified using CompassForSW software and normalized to GAPDH within the same lane. The response to depolarization was calculated as the ratio of protein levels in CCCP-treated samples relative to vehicle controls (CCCP/DMSO). The average CCCP-response value for each protein was determined from at least three biological replicates. Scatter-plot graphs display the average CCCP-response value for each protein, based on at least three biological replicates. Error bars represent the standard deviation (SD) for each individual. D–F Protein average levels are calculated per each individual and represented by scatter plot in each group of the following: Healthy controls (HC) n = 16, non-PD (Ataxia, FTD, and ALS) n = 3, all Parkinson´s disease patients (PD) n = 23, idiopathic PD (iPD) n = 16, familial PD (FPD) n = 8, non-pathogenic variants (NPV) n = 11 and pathogenic variants n = 5. D Miro1 depolarization-response values. E Miro1 baseline levels F Mfn2 depolarization-response values. The average values are compared between different groups. Error bars indicate the standard deviation between different individuals within each group. Statistical significance of the CCCP-response values between different groups was assessed using one-way ANOVA test, with P values <0.05, stated numerically on the graph.

Fig. 2. Miro1 degradation is retarded in IPD but inhibited in PINK1/PRKN PD.

A Fibroblast cell lines from two healthy controls and five PD patients were treated with 40 µM CCCP for the indicated time points to assess Miro1 degradation over time. Miro1 protein levels were quantified using CompassForSW software, and its response to depolarization was expressed as the ratio of treated to vehicle control (time point 0). B Fibroblast cell lines from two PD patients with low mitochondria-specific polygenic risk score (mitoPRS) and two PD patients with high mitoPRS were treated with 40 µM CCCP for 6 h. Miro1 depolarization response was measured using CompassForSW software and compared to Mfn2 depolarization response as show in the graph. Statistical significance of the CCCP-response values between low Mito-PRS and high Mito-PRS was assessed using one-way ANOVA test, with P values stated numerically on the graph.

Fig. 3. Miro1 degradation is linked to aging and disease duration.

A Flux graph illustrating Miro1 levels in fibroblasts from healthy controls (HC-upper graph) and PD patients (lower graph) under basal (DMSO) and CCCP-treated conditions. Each line represents an individual, with the slope indicating the rate of Miro1 degradation over 6 h of CCCP treatment. Miro1 band intensities were normalized to GAPDH levels from the same lane. Data points represent mean values from at least three independent experiments (n ≥ 3). B Heatmap displaying the degradation levels for Miro1 and Mfn2 ( + CCCP/DMSO), in each individual across HC, PD and non-PD cohorts. CUpper graph- Receiver Operating Characteristic (ROC) curves based on our study dataset, where machine learning models were trained using our study´s Miro1 degradation values in the HC and PD groups. The trained models were then used to predict Healthy Controls (HC) and Parkinson’s disease (PD) individuals from Hsieh et al. 2019 study. Area Under the Curve (AUC-ROC) values indicate the discriminative ability of each model. Models compared: random forest (rf), K-nearest neighbors (knn), naïve Bayes (bayes), linear discriminant analysis (lda), and generalized linear model (glm). Lower graph Confusion matrix summarizing classification performance, showing the number of correctly classified cases (true positives and true negatives) and misclassified cases (false positives and false negatives). D Comparison of Miro1 degradation relative to Mfn2 degradation in response to CCCP treatment across individuals in HC and PD groups. Data are presented as scatter-plot with mean ± SEM. Statistical significance was assessed using Mann-Whitney test, with P-values stated numerically on the graph. E Correlation between Miro1 average retention levels ± SD, and age at sampling (AAS) across all individuals and cohorts. F Correlation between Miro1 retention levels ± SD and age at sampling, analyzed separately for HC (upper graph) and PD (lower graph) groups. G Correlation between Miro1 retention levels ± SD and age at onset (AAO) is represented across individuals in PD group. H Correlation between Miro1 retention levels ± SD and disease duration (years since PD diagnosis) across individuals in PD group. F–H Pearson correlation (r) and Linear regression analysis were performed.

Fig. 4. Miro1 is retained following depolarization of mitochondria in peripheral blood cells and dopaminergic neurons from PD patients that had previously donated a skin biopsy.

A Peripheral blood mononuclear cell lines (PBMCs) from a total of six donors, including three healthy controls (HC) and three Parkinson’s disease (PD) patients were included in our study. The PBMCs PD donors were the same individuals in the fibroblast cohort (PD-9, PD-10, PD-11). B PBMCs were isolated from blood samples collected using BD Vacutainer® CPT™ tubes. Whole blood was centrifuged to separate plasma from the PBMC layer, which was then carefully transferred, washed, and cryopreserved for downstream analyses. C Whole-cell lysates of PBMCs were analyzed using simple western blotting (JESS, Bio-Techne) to assess Miro1 response following mitochondrial depolarization with 10 µM oligomycin and 10 µM antimycin (OA) for 6 h. Protein intensities were quantified using CompassForSW software and normalized to GAPDH within the same lane. The response to depolarization was calculated as the ratio of protein levels in oligomycin and antimycin- treated samples to those in the vehicle control - ethanol treated samples ( + OA/EtOH). Distribution of Miro1 (left graph) and Mfn2 (right graph) depolarization responses across individuals in the HC group (gray, n = 3) and PD group (red, n = 3). D Dopaminergic neuron (DAN) cell lines were generated from fibroblasts of three donors, including one healthy control and two sporadic Parkinson’s disease (PD) patients. The fibroblasts were reprogrammed into induced pluripotent stem cells (iPSCs), which were subsequently differentiated into neural progenitor cells (NPCs) and further matured into dopaminergic neurons. The three DANs lines were generated from individuals in our fibroblast cohort (HC-5, PD-11, PD-17). E WCL of DANs were analyzed using simple western blotting (JESS, Bio-Techne) to assess Miro1 response following mitochondrial depolarization with 100 µM antimycin for different time points (2, 4 and 6 h). Protein intensities were quantified using CompassForSW software and normalized to GAPDH within the same lane. (F-G) Distribution of Miro1 depolarization-response in HC individuals and PD patients across different cell types.

Fig. 5. Compound #1915758622 only minimally improved the Miro1 phenotype observed in PD fibroblasts.

A Fibroblasts from two healthy controls and four PD patients were used to assess the effect of compound 3 (Mcule #1915758622) on Miro1 retention upon mitochondrial depolarization. Cells were pretreated with varying concentrations of the compound (1, 10, and 30 µM) for 24 h, followed by 6 h treatment with 40 µM CCCP. A compound-only control (30 µM, no CCCP) was included for each cell line. B Graph showing Miro1 response to depolarization upon compound 3 pretreatment across different cell lines. Miro1 degradation levels were quantified using CompassForSW software and expressed as the ratio of treated to vehicle control (treatment/DMSO). C Distribution of Miro1 degradation levels in PD cell lines under different compound pretreatment concentrations. Statistical significance of the compound´s reducing potential of Miro1 retention at different concentrations was assessed using Friedman test (not significant, P > 0.05).