Elia, S. N. | bioRxiv Science

bioRxiv2025

Subtle Cellular Phenotypes Inform Pathological and Benign Genetic Mutants in the Iduronidase-2 Sulfatase Gene

As with most enzyme-storage disorders, DNA testing at an early age is critical for identifying genetic variants and their impact on disease burden. Still, most variants in genes such as Iduronate-2-sulfatase, IDS, are Variants of Uncertain Significance (VUS). In patients presenting with Hunter Syndrome, clinical testing for IDS enzyme activity has been the mainstay to determine whether a variant is likely damaging. However, these enzyme assays are unable to predict disease severity or neuronal toxicity and may be missing many aspects of IDS pathology. In this study, we have developed an image-based assay using genome-engineered cells with IDS mutations to identify if a specific mutation causes lysosomal and membrane disruptions that characterize the disease. Specifically, we generate twelve mutant cell lines and document both the biochemical changes in IDS activity and the reproducible phenotypic differences therein. Next, we examine two patient-derived cell lines and find the same phenotypic differences in these lines compared to parental controls. The phenotypic changes are measured on a specific scale, which we term the PathScoreLC. To determine whether the observed changes are specific to IDS, we reintroduce a recombinant version of the IDS enzyme (rhIDS) to rescue both the biochemical and phenotypic changes of these cells. Interestingly, even after 120 hours, rescue is present, but not all the cells return to normal, which may also reflect the pathological state. Finally, we examine the gene expression differences and find that recombinant enzyme is not sufficient to induce transcriptional changes in the mutant lines at the time points studied. Overall, these cell-based lysosomal and membrane phenotypes may be key to quickly and accurately profiling clinical variants in the IDS gene.

cell biology↗

bioRxiv2023

Classification of iPSC-Derived Cultures Using Convolutional Neural Networks to Identify Single Differentiated Neurons for Isolation or Measurement

Understanding neurodegenerative disease pathology depends on a close examination of neurons and their processes. However, image-based single-cell analyses of neurons often require laborious and time-consuming manual classification tasks. Here, we present a machine learning approach leveraging convolutional neural network (CNN) classifiers that have the capability to accurately identify various classes of neuronal images, including single neurons. We developed the Single Neuron Identification Model 20-Class (SNIM20) which was trained on a dataset of induced pluripotent stem cell (iPSC)-derived motor neurons, containing over 12,000 images from 20 distinct classes. SNIM20 is built in TensorFlow and trained on images of differentiated iPSC cultures stained for nuclei and microtubules. This classifier demonstrated high predictive accuracy (AUC = 0.99) for distinguishing single neurons. Additionally, the 2-stage training framework can be used more broadly for cellular classification tasks. A variation was successfully trained on images of a human osteosarcoma cell line (U2OS) for single-cell classification (AUC = 0.99). While this framework was primarily designed for single-cell microraft-based identification and capture, it also works with cells in standard plate formats. We additionally explore the impact of specific fluorescent channels and brightfield images, class groupings, and transfer learning on the quality of the classification. This framework can both assist in high throughput neuronal or cellular identification and be used to train a custom classifier for the users needs.

neuroscience↗