bioRxiv Science⌕ Search

Biology subjects

Fon, E. A.

Publications and source records attributed to Fon, E. A..

At least 19 recordsLinked to original sources

USP15 REGULATES NEUROINFLAMMATION AND DRIVES PATHOGENESIS IN SYNUCLEINOPATHIES

Neuroinflammation strongly contributes to the pathogenesis of neurological and neurodegenerative diseases, including Parkinson's disease. We show that ablation of Usp15 in astrocytes and in microglia protects against lethal neuroinflammation in vivo. In a mouse model of synucleinopathy, Usp15 deletion diminishes alpha-syn deposits in the brain, slows disease progression, and increases survival time. The neuroprotective effect of Usp15 is associated with differential expression of inflammatory pathways in situ including interferon stimulated genes. These USP15-dependent effects in vivo are recapitulated in vitro in primary human microglia and astrocytes. We detect high USP15 expression in microglia from Parkinson's patients with strong co-expression with LRRK2 and SNCA. In humans, we detect a strong cis-acting eQTL directing high USP15 expression in CD14+ myeloid cells. The allele driving this eQTL is itself associated with increased disease risk, linking myeloid USP15 expression, elevated USP15 plasma levels in Parkinson's patients, to genetic susceptibility.

neuroscience↗

WDR44 drives de novo α-synuclein aggregation at the lysosomal membrane and promotes neuronal dysfunction in Parkinson's Disease

The aggregation of -synuclein (-SYN) into Lewy bodies (LBs) is a central event in the pathogenesis of Parkinsons disease (PD) and related synucleinopathies1,2. Despite significant advances in understanding -SYN self-assembly, the precise sequence of early aggregation steps has not been directly visualized in living neurons. Here, we use an optogenetic-induced protein aggregation system with a high temporal resolution to monitor the onset of -SYN assembly in neurons. We found that the initiation and accumulation of -SYN aggregates occur predominantly at the lysosomal membrane, an event driven by the -SYN N-terminus and modulated by the membrane-associated adaptor protein WD repeat-containing protein 44 (WDR44). Remarkably, we demonstrate that WDR44 knockdown markedly reduced de novo -SYN aggregation in both neuronal cultures and in vivo, whereas WDR44 overexpression enhances -SYN aggregation in PD patient-derived iPSC neurons. Consistent with its potential pathogenic involvement, WDR44 aberrantly accumulates in vivo and in the brains of PD patients, where it colocalizes with LB inclusions. Finally, we show that lysosome-associated -SYN aggregates compromised lysosomal structure and function, leading to neuronal impairment, a phenotype worsened by WDR44 overexpression, linking early aggregation events to downstream toxicity. Together, these findings reveal the earliest dynamic stages of -SYN oligomerization in living neurons and identify the WDR44--SYN interaction as a promising therapeutic target for reducing -SYN pathology and enabling early intervention in PD.

neuroscience↗

Integration of iPSC-derived microglia into human midbrain organoids enhances microglial maturation and inflammatory signaling

Microglia are the resident immune cells of the central nervous system and play key roles in the healthy brain during development and adulthood, as well as during neurodegenerative diseases - including Parkinsons disease (PD). Yet the role of microglia in PD pathogenesis has not been fully elucidated. Limitations of 2D cell culture and animal models in simulating human microglia in the brain parenchyma have contributed to this knowledge gap. Human midbrain organoids (hMOs) provide a promising model that can recapitulate elements of PD pathology but lack microglial cells. Here we adapt protocols for the differentiation of hMOs and human iPSC-derived microglia (iMG) to generate iMG-hMO assembloids. Within assembloids, integrated iMG (intMG) express canonical microglia markers and induce the release of cytokines and chemokines. Transcriptomic profiling by single cell RNA sequencing reveals that intMG adopt a more mature and inflammation-responsive state compared to 2D iMG. The integration of microglia results in increased signaling through inflammatory and trophic pathways that drive altered transcriptional signatures of dopaminergic neurons and astrocytes within assembloids. Overall, iMG-hMO assembloids have the potential to more faithfully model the role of microglia and neuroinflammation in PD pathogenesis.

neuroscience↗

A high-content imaging workflow to screen for molecules that reduce cellular uptake of α-synuclein preformed fibrils

A classical pathological hallmark of many neurodegenerative diseases is the formation of protein-rich aggregates and inclusions. In Parkinsons disease (PD), -synuclein (-syn) constitutes a major protein component of pathological inclusions, termed Lewy bodies. These -syn aggregates are hypothesized to spread throughout the nervous system by cell-to-cell transmission acting as templates to amplify aggregate formation. In vitro generated -syn aggregates, commonly called preformed fibrils (PFFs), have been used to investigate a number of aspects related to -syn mediated pathology across different model systems. Here we describe a semi-automated assay to screen for small molecules that interfere with the cellular uptake and accumulation of PFFs. The assay uses dopaminergic progenitor cells (DPCs), derived from human induced pluripotent stem cells (hiPSCs). In an initial screen, we tested 1520 small molecules and identified several molecules that strongly reduce intracellular PFF load in DPCs. From these hits, candidate compounds were validated in dopaminergic neurons (DNs) to demonstrate the utility of the assay. This assay provides a robust, scalable and adaptable tool to screen for molecules that affect PFF uptake in hiPSC-derived cell models. Within the scope of this screen, it led to the identification of a set of compounds with diverse annotated targets that effectively reduce the uptake of synuclein aggregates in DPCs and DNs.

neuroscience↗

An automated workflow for quantifying the formation of synuclein aggregates in human dopaminergic neurons

Parkinsons disease (PD) is a neurodegenerative disorder characterized by alpha-synuclein (-syn) aggregates termed Lewy bodies. To model PD pathology in vitro, preformed fibrils of -syn (PFFs), which can be taken up by cells, provide a seed that drives misfolding and aggregation of endogenous -syn, with new aggregates amplifying this process. External application of PFFs to dopaminergic neurons (DNs) increases aggregate formation, marked by -syn phosphorylation at serine 129 (pS129-syn), a pathological PD marker. Building on this, we developed an automated synuclein seeding assay to quantify new -syn aggregates in iPSC-derived DNs. Using pS129-syn as a readout, we show that PFFs elicit a time- and dose-dependent increase in pS129-syn aggregates. Our high-throughput assay further revealed that aggregate formation depends on endogenous -syn levels. Treatment with PFFs produced a greater increase in pS129-syn aggregates in iPSC DNs derived from a PD patient with a triplication in the SNCA gene, which encodes the -syn protein and which elevates total -syn levels, relative to DNs from an isogenic iPSC line from the same individual, in which the SNCA gene mutation had been corrected by CRISPR/Cas9. In contrast, no pS129-syn signal was detected in neurons in which all copies of the SNCA gene had been knocked out (KO). This high-content imaging assay for synuclein seeding offers a platform for assessing compounds and therapeutics that may impede -syn aggregate formation.

neuroscience↗

N-terminal acetylation reduces α-synuclein pathology in models of Parkinson's disease

The -synuclein protein, encoded by SNCA gene, is a major constituent of pathological intracellular inclusions such as Lewy bodies found in the brains of patients with Parkinsons disease and other synucleinopathies. Whereas -synuclein phosphorylation has been much studied, comparatively less work has been devoted to other post-translational modifications such as acetylation, especially given that N-terminally acetylated -synuclein is the most abundant endogenous form of the protein in the brain. In this study, using multiple in vitro and in vivo models, we sought to better understand the role of N-terminal acetylation in the pathogenesis of synucleinopathies. We found that N-terminal acetylation slowed aggregation of both -synuclein monomers and pre-formed fibrils in vitro. Uptake of acetylated -synuclein pre-formed fibrils into both immortalized cell lines and iPSC-derived dopamine neurons was also slowed compared non-acetylated fibrils. In addition, exposure to acetylated pre-formed fibrils induced less seeding of endogenous -synuclein, as measured by the accumulation of Serine129-phosphorylated -synuclein inclusions in both iPSC-derived dopamine neurons and mouse brain. Finally, mice injected with N-terminally acetylated -synuclein pre-formed fibrils survived significantly longer than mice injected with non-acetylated fibrils. Taken together, our study indicates that N-terminal acetylation reduces -synuclein aggregation, uptake into cells, seeding of endogenous -synuclein, and toxicity in vivo, suggesting that this prevalent post-translational modification represents a potent, physiologically relevant protective mechanism, which has thus far largely not been taken into consideration in most experimental paradigms of Parkinsons disease and synucleinopathies.

molecular biology↗

LRRK2 Kinase Mediates Increased GCase Activity in Microglia in Response to Proinflammatory Stimuli

Variants in the LRRK2 and GBA1 genes are among the most common risk factors associated with Parkinsons disease (PD). Both patients carrying PD-associated variants in GBA1, encoding lysosomal enzyme glucocerebrosidase (GCase), and a subset of non-carrier patients have been shown to have reduced GCase enzymatic activity, suggesting that reduced GCase activity may be a feature of both genetic and a subset of sporadic PD. However, the effect of PD-associated variants in LRRK2, encoding a serine/threonine kinase, on GCase activity remains controversial, with conflicting results in various tissues and cell types. Moreover, rare patients carrying both GBA1 and LRRK2 risk alleles seem to have a more benign disease course than carriers of GBA1 variants alone, suggesting a complex interplay between these two genes in PD. Here we evaluate the effect of LRRK2 kinase activity on GCase activity in human induced pluripotent stem cell (iPSC)-derived microglia (iMGs), a PD-relevant brain cell type expressing high levels of LRRK2. Using CRISPR editing, isogenic control iPSC lines were generated to match PD patient-derived iPSC lines harbouring the LRRK2 p.G2019S, p.M1646T, or p.N551K-p.R1398H protective haplotype variants. Whereas iMGs harbouring the p.M1646T variant, and the protective haplotype, respectively increased and decreased phosphorylation of canonical LRRK2 substrate, Rab10, GCase protein levels and activity were not altered in any of the LRRK2 variant lines. Additionally, whereas pharmacological inhibition of LRRK2 kinase activity had no impact on GCase activity in iMGs under basal conditions, it attenuated the increase in GCase activity elicited in response to interferon {gamma} (IFN{gamma}) treatment. Moreover, GCase activity induced by IFN{gamma} was reduced in PD risk LRRK2 p.M1646T iMGs and increased in p.N551K-p.R1398H protective haplotype iMGs compared to their isogenic corrected controls, congruent with their respective effects on LRRK2 kinase activity and PD risk. Thus, our data suggests a role for LRRK2 kinase activity in regulation of GCase activity in response to neuroinflammation.

neuroscience↗

Cell type specific analysis of ALS associated proteins reveals immune-regulated targets

A substantial fraction of amyotrophic lateral sclerosis (ALS)-associated proteins remain poorly characterized, in part because of the limited availability of validated research antibodies. We established knockout (KO)-based antibody characterization workflows and demonstrated that widely used antibodies against the major ALS-associated protein C9orf72 lacked specificity (Laflamme et al., 2019). We subsequently scaled this framework to systematically benchmark research antibodies, revealing that up to 61% fail to perform as recommended by manufacturers (Ayoubi et al., 2023). Here, we extend this approach by establishing the ALS-Reproducible Antibody Platform (ALS-RAP), a comprehensive effort to generate a publicly available dataset of KO-validated antibodies targeting proteins encoded by ALS risk genes. In total, we characterized 303 antibodies against 33 ALS-associated proteins to identify high-quality reagents for use in western blot, immunoprecipitation, and immunofluorescence. Using these antibodies, we profiled protein levels across human induced pluripotent stem cell (iPSC)-derived and primary neurological cell types. These analyses revealed diverse cellular distributions and higher levels of several ALS-associated proteins in glial populations, consistent with emerging evidence for immune contributions to ALS. Together, ALS-RAP provides a validated antibody toolbox and protein expression resource to support the study of ALS-associated proteins.

neuroscience↗

In vivo and in silico alpha-synuclein propagation dynamics: The role of genotype, epicentre, and connectivity

Neurodegeneration observed in synucleinopathies, like Parkinsons disease (PD), are hypothesized to be a consequence of the progressive accumulation and spread of misfolded alpha-synuclein (aSyn) throughout the brain. Here, we study the generalizability of this hypothesis across multiple biologically-relevant factors including genotype, aSyn species, and regional vulnerability (i.e.: does the introduction of aSyn into different brain regions have a similar impact) using a detailed longitudinal brain (using magnetic resonance imaging; [MRI]) and behavioural approaches. We first examined wild-type and M83 A53T hemizygous transgenic (engineered to over-express human aSyn) C57BL/6 x C3H mice receiving striatal inoculation with human or mouse preformed fibrils (PFFs) (n=89 mice at the last time point; n=687 MRI). Longitudinal analyses demonstrated a time-dependent increase in network-like atrophy and motor deficits, generalized across genotype and PFF species. We further derived latent dimensions relating brain-behaviour relationships revealing a pattern of sex- and genotype-relevant atrophy. Overall, atrophy was most prominent in M83 mice with mouse PFFs, while human PFFs or wild-type hosts showed attenuated effects. Changing the inoculation site to the hippocampus, a major connectivity hub, revealed differential regional vulnerability in the form of localized atrophy. Computational models previously validated in clinical PD further indicated regional vulnerability with increased predictive ability for atrophy patterns associated with striatal compared to hippocampal inoculation. Our findings reveal atrophy resulting from aSyn spreading is generalizable across genotype and PFF species, but not disease epicentres, emphasizing the role of regional vulnerability in disease progression.

neuroscience↗

Large-scale bidirectional arrayed genetic screens identify OXR1 and EMC4 as modifiers of α-synuclein aggregation

In Parkinsons disease and other synucleinopathies, -synuclein (-Syn) misfolds and forms Ser129-phosphorylated aggregates (pSyn129). The factors controlling this process are largely unknown. Here, we used arrayed CRISPR-mediated gene activation and ablation to discover new pSyn129 modulators. Using quadruple-guide RNAs (qgRNAs) and Cas9, or an inactive Cas9 version fused to a synthetic transactivator, we ablated 2304 and activated 2428 human genes related to mitochondrial, trafficking and motility function in HEK293 cells. After exposure of cells to -Syn fibrils, pSyn129 signals were recorded by high-throughput fluorescence microscopy and aggregates were identified by image analysis. We found that pSyn129 was increased by activating the mitochondrial protein OXR1, which decreased ATP levels and altered the mitochondrial membrane potential. Instead, pSyn129 was reduced by ablation of the endoplasmic reticulum (ER)-associated protein EMC4, which enhanced ER-driven autophagic flux and lysosomal clearance. OXR1 activation preferentially modulated cellular reactions to fibrils derived from multiple system atrophy (MSA) patients, whereas EMC4 ablation broadly reduced pSyn129 across diverse -Syn polymorphs. These findings were confirmed in human iPSC-derived cortical and dopaminergic neurons, where OXR1 preferentially promoted somatic aggregation and EMC4 reduced both somatic and neuritic aggregates. These results uncover previously unrecognized roles for OXR1 and EMC4 in -Syn aggregation, thereby broadening our mechanistic understanding of synucleinopathies.

neuroscience↗

α-Synuclein purification significantly impacts seed amplification assay performance and consistency

-Synuclein seed amplification assays are a promising diagnostic tool for synucleinopathies such as Parkinsons disease and multiple system atrophy. Standardized conditions are required to ensure a high degree of inter- and intra-laboratory reproducibility when performing these assays. A significant issue that hinders the utility of seed amplification assays is the de novo aggregation propensity of the -synuclein substrate as well as inter-batch heterogeneity. While much work has focused on determining appropriate seed amplification assay buffer compositions as well as the type and amount of seed used, a robust comparison of -synuclein substrate purification methods has not been reported. We therefore compared the utility of recombinant -synuclein purified using four different methods as seed amplification assay substrates across two laboratories. Osmotic shock-purified -synuclein monomer substrate showed the lowest propensity for de novo aggregation, which translated into being the best substrate for seed amplification assay reactions seeded with -synuclein preformed fibrils or patient brain homogenates. Furthermore, osmotic shock -synuclein monomer showed the best inter-batch reproducibility compared to all other substrates tested. As -synuclein seed amplification assays continue to evolve and move towards adoption in the clinical realm, this work showcases the vital importance of standardizing the production and characterization of recombinant -synuclein substrate. We encourage the widespread adoption of osmotic shock -synuclein monomer as the universal substrate for seed amplification assays to maximize intra- and inter-laboratory reproducibility.

biochemistry↗

T cell responses towards PINK1 and α-synuclein are elevated in prodromal Parkinson's disease

A role of the immune system in Parkinsons disease (PD) progression has long been suspected due to the increased frequency of activated glial cells and infiltrating T cells into the substantia nigra. It was previously reported that PD donors have increased T cell responses towards PINK1 and -synuclein (-syn), two Lewy body-associated proteins. Further, T cell reactivity towards -syn was highest closer to disease onset, highlighting that autoreactive T cells might play a role in PD pathogenesis. However, whether T cell autoreactivity is present during prodromal PD is unknown. Here, we investigated T cell responses towards PINK1 and -syn in donors at high risk of developing PD (i.e. prodromal PD: genetic risk, hyposmia, and or REM sleep behavior disorder), in comparison to PD and healthy control donors. T cell reactivity to these two autoantigens was detected in prodromal PD at levels comparable to those detected in individuals with clinically diagnosed PD. Aligned with the increased incidence of PD in males, we found that males with PD, but not females, had elevated T cell reactivity compared to healthy controls. However, among prodromal PD donors, males and females had elevated T cell responses. These differing trends in reactivity highlights the need for further studies of the impact of biological sex on neuroinflammation and PD progression.

neuroscience↗

Midbrain organoids with an SNCA gene triplication display dopamine-dependent alterations in network activity

Human midbrain organoids (hMOs) show promise as a patient-derived model for the study of Parkinsons disease (PD). Yet, much remains unknown about how accurately hMOs recapitulate key features of PD in the human brain. In both PD patients and animal models, disease progression leads to characteristic changes in neural activity throughout the basal ganglia. Here we demonstrate that patient-derived induced pluripotent stem cell (iPSC) hMOs harboring a triplication in the SNCA gene, encoding -synuclein, a key protein in PD pathogenesis, can recapitulate PD-associated changes in neural activity. Namely, we observe hyperactive network activity in SNCA triplication hMOs, but not in isogenic, CRISPR-corrected iPSC hMOs. These changes are characterized by an increase in the number of bursts and network-wide bursts. Moreover, SNCA triplication hMOs exhibit an increase in network synchrony and burst/network burst strength similar to observations in animal and human PD brains. Subsequently, we show that the observed changes in neuronal activity are attributed to dopamine D2 receptor hypoactivity due to dopamine depletion, which could be reversed by the D2 receptor agonist quinpirole. Thus, hMOs faithfully model network wide electrophysiological changes associated with PD progression and serve as a promising tool for PD research and personalized medicine.

neuroscience↗

STIP1/HOP Promotes the Formation of Cytotoxic α-Synuclein Oligomers

The accumulation of alpha-synuclein (a-Syn) as toxic oligomers, and subsequently in Lewy bodies, is a pathological hallmark of Parkinsons disease (PD) and other synucleinopathies. Molecular chaperones and co-chaperones are expected to act in concert to maintain physiological activities of proteins, including a-Syn, but in neurodegeneration this process can become mal-adaptive. Transcript levels of Stress inducible phosphoprotein 1 (STIP1), a co-chaperone of Hsp90/Hsp70, are elevated in brain samples from PD patients. In synucleinopathy mouse models, STIP1 has unexpected bidirectional effects on a-Syn, with overexpression of STIP1 aggravating a-Syn toxicity, whereas knockdown of STIP1 improves toxicity and behavioural phenotypes. However, it is unclear how STIP1 enhances the toxicity of a-Syn. Here we unravel the mechanisms by which the direct interaction between STIP1/HOP and a-Syn regulates the neurotoxicity of a-Syn. Specifically, two binding motifs in the C-terminus of a-Syn directly interact with the TPR2A domain of STIP1/HOP in a dynamic manner, competing for a shared interface on TPR2A. Binding of STIP1/HOP to a-Syn attenuates the formation of a-Syn fibrils while promoting the accumulation of high molecular weight amorphous a-Syn species. Samples of a-Syn aggregated in the presence of STIP1/HOP contain significantly more A11-positive oligomeric species and cause a greater reduction in cell viability than a-Syn aggregated in the absence of STIP1/HOP in neuronal cells. Our results provide a mechanism by which the direct interaction between STIP1/HOP and the C-terminus of a-Syn promotes the formation of cytotoxic, non-amyloidogenic, high molecular weight a-Syn species. Our model offers an explanation for the unexpected pathological link between STIP1 and a-Syn toxicity, thus opening new therapeutic avenues for the treatment of synucleinopathies.

biochemistry↗

Cell type transcriptomics reveal shared genetic mechanisms in Alzheimer's and Parkinson's disease

Historically, Alzheimers disease (AD) and Parkinsons disease (PD) have been investigated as two distinct disorders of the brain. However, a few similarities in neuropathology and clinical symptoms have been documented over the years. Traditional single gene-centric genetic studies, including GWAS and differential gene expression analyses, have struggled to unravel the molecular links between AD and PD. To address this, we tailor a pattern-learning framework to analyze synchronous gene co-expression at sub-cell-type resolution. Utilizing recently published single-nucleus AD (70,634 nuclei) and PD (340,902 nuclei) datasets from postmortem human brains, we systematically extract and juxtapose disease-critical gene modules. Our findings reveal extensive molecular similarities between AD and PD gene cliques. In neurons, disrupted cytoskeletal dynamics and mitochondrial stress highlight convergence in key processes; glial modules share roles in T-cell activation, myelin synthesis, and synapse pruning. This multi-module sub-cell-type approach offers insights into the molecular basis of shared neuropathology in AD and PD.

neuroscience↗

Ensemblex: an accuracy-weighted ensemble genetic demultiplexing framework for population-scale scRNAseq sample pooling

Multiplexing samples from distinct individuals prior to sequencing is a promising step toward achieving population-scale single-cell RNA sequencing by reducing the restrictive costs of the technology. Individual genetic demultiplexing tools resolve the donor-of-origin identity of pooled cells using natural genetic variation but present diminished accuracy on highly multiplexed experiments, impeding the analytic potential of the dataset. In response, we introduce Ensemblex: an accuracy-weighted, ensemble genetic demultiplexing framework that integrates four distinct algorithms to identify the most probable subject labels. Using computationally and experimentally pooled samples, we demonstrate Ensemblexs superior accuracy and illustrate the implications of robust demultiplexing on biological analyses.

genomics↗

Sex-focused analyses of M83 A53T hemizygous mouse model with recombinant human alpha-synuclein preformed fibril injection identifies female resilience to disease progression: A combined magnetic resonance imaging and behavioural study.

Alpha-synuclein (aSyn) pathology has been extensively studied in mouse models harbouring human mutations. In spite of the known sex differences in age of onset, prevalence and disease presentation in human synucleinopathies, the impact of sex on aSyn propagation has received very little attention. To address this need, we examined sex differences in whole brain signatures of neurodegeneration due to aSyn toxicity in the M83 mouse model using longitudinal magnetic resonance imaging (MRI; T1-weighted; 100 m3 isotropic voxel; acquired -7, 30, 90 and 120 days post-injection [dpi]; n[≥]8 mice/group/sex/time point). To initiate aSyn spreading, M83 mice were inoculated with recombinant human aSyn preformed fibrils (Hu-PFF) or phosphate buffered saline (PBS) injected in the right dorsal striatum. We observed more aggressive neurodegenerative profiles over time for male M83 Hu-PFF-injected mice when examining voxel-wise trajectories. However, at 90 dpi, we observed widespread patterns of neurodegeneration in the female Hu-PFF-injected mice. These differences were not accompanied with any differences in motor symptom onset between the male and female Hu-PFF-injected mice. However, male Hu-PFF-injected mice reached their humane endpoint sooner. These findings suggest that post-motor symptom onset, even though more accelerated disease trajectories were observed for male Hu-PFF-injected mice, neurodegeneration may appear sooner in female Hu-PFF-injected mice (prior to motor symptomatology). These findings suggest that sex-specific synucleinopathy phenotypes urgently need to be considered to improve our understanding of neuroprotective and neurodegenerative mechanisms.

neuroscience↗

A deep learning convolutional neural network distinguishes neuronal models of Parkinson's disease from matched controls

Parkinsons disease (PD) is a neurodegenerative disorder that results in the loss of dopaminergic neurons in the substantia nigra pars compacta. Despite advances in understanding PD, there is a critical need for novel therapeutics that can slow or halt its progression. Induced pluripotent stem cell (iPSC)-derived dopaminergic neurons have been used to model PD but measuring differences between PD and control cells in a robust, reproducible, and scalable manner remains a challenge. In this study, we developed a binary classifier convolutional neural network (CNN) to accurately classify microscopy images of PD models and matched control cells. We acquired images of iPSC-derived neural precursor cells (NPCs) and dopaminergic (DANs) and trained multiple CNN models comparing control cells to genetic and chemical models of PD. Our CNN accurately predicted whether control NPC cells were treated with the PD-inducing pesticide rotenone with 97.60% accuracy. We also compared control to a genetic model of PD (deletion of the Parkin gene) and found a predictive accuracy of 86.77% and 95.47% for NPC and DAN CNNs, respectively. Our cells were stained for nuclei, mitochondria, and plasma membrane, and we compared the contribution of each to the CNNs accuracy. Using all three features together produced the best accuracy, but nuclear staining alone produced a highly predictive CNN. Our study demonstrates the power of deep learning and computer vision for analyzing complex PD-related phenotypes in DANs and suggests that these tools hold promise for identifying new targets for therapy and improving our understanding of PD.

neuroscience↗