Machine learning-aided multidimensional phenotyping of Parkinson’s disease patient stem cell-derived midbrain dopaminergic neurons

Abstract

Combining multiple Parkinson's disease (PD) relevant cellular phenotypes might increase the accuracy of midbrain dopaminergic (mDA) in vitro models. We differentiated patient-derived induced pluripotent stem cells (iPSCs) with a LRRK2 G2019S mutation, isogenic control and genetically unrelated iPSCs into mDA neurons. Using automated fluorescence microscopy in 384-well plate format, we identified elevated levels of α-synuclein and Serine 129 phosphorylation (pS129), reduced dendritic complexity, and mitochondrial dysfunction. Next, we measured additional image-based phenotypes and used machine learning (ML) to accurately classify mDA neurons according to their genotype. Additionally, we show that chemical compound treatments, targeting LRRK2 kinase activity or α-synuclein levels, are detectable when using ML classification based on multiple image-based phenotypes. We validated our approach using a second isogenic patient derived SNCA gene triplication mDA neuronal model. This phenotyping and classification strategy improves the exploitability of mDA neurons for disease modelling and the identification of novel PD drug targets.

Revealing invisible cell phenotypes with conditional generative modeling

“…Cell pellets were resuspended in differentiation medium supplemented with a ROCK inhibitor. Media compositions are detailed in Vuidel et al, 2022. 384-well plates (Perkin Elmer, #6007558) were coated with 15 μg/ml Poly-L-Ornithin for 1 hour at 37 °C followed by 10μg/ml Laminin overnight at 4 °C.…”

“…Subsequently, we wondered if this approach could be used to investigate subtle morphological variation induced by a mutation. We used iPSCs reprogrammed from fibroblasts of a Parkinson's disease (PD) patient carrying the LRRK2 G2019S mutation along with an isogenic control, in which the mutation was genetically rescued to the wild-type sequence by gene editing and differentiated the iPSCs into dopaminergic neurons (Vuidel et al 2022). The LRRK2 G2019S mutation, localized in exon 41 of the LRRK2 (leucine-rich repeat kinase 2) gene, is causally linked to the development of Parkinson's Disease (Tolosa et al 2020).…”

“…Differentiation of dopaminergic neurons from iPSCs, immunohistochemistry and image acquisition Dopaminergic neurons were differentiated from iPSCs using a modified protocol based on Kriks et al which is described in more detail in Vuidel et al (Kriks et al, 2011;Ryan et al, 2013;Weykopf et al, 2019;Vuidel et al, 2022). Differentiated dopaminergic neurons were cryopreserved at 30 days in vitro (DIV30).…”

“…Thawed cells were incubated at 37 °C and 5 % CO2 for seven days until 37 DIV with differentiation medium changes every other day. Cell fixation and immunohistochemistry were performed as detailed in Vuidel et al, 2022. All imaging experiments were performed on a Yokogawa CV7000 microscope in scanning confocal mode using a dual Nipkow disk. 384-well plates (Perkin Elmer, #6007558) were mounted on a motorized stage and images were acquired in a row-wise "zig-zag" fashion at RT.…”