Dataset Information

Deep learning predicts function of live retinal pigment epithelium from quantitative microscopy.

ABSTRACT: Increases in the number of cell therapies in the preclinical and clinical phases have prompted the need for reliable and noninvasive assays to validate transplant function in clinical biomanufacturing. We developed a robust characterization methodology composed of quantitative bright-field absorbance microscopy (QBAM) and deep neural networks (DNNs) to noninvasively predict tissue function and cellular donor identity. The methodology was validated using clinical-grade induced pluripotent stem cell-derived retinal pigment epithelial cells (iPSC-RPE). QBAM images of iPSC-RPE were used to train DNNs that predicted iPSC-RPE monolayer transepithelial resistance, predicted polarized vascular endothelial growth factor (VEGF) secretion, and matched iPSC-RPE monolayers to the stem cell donors. DNN predictions were supplemented with traditional machine-learning algorithms that identified shape and texture features of single cells that were used to predict tissue function and iPSC donor identity. These results demonstrate noninvasive cell therapy characterization can be achieved with QBAM and machine learning.

SUBMITTER: Schaub NJ

PROVIDER: S-EPMC6994191 | biostudies-literature | 2020 Feb

REPOSITORIES: biostudies-literature

ACCESS DATA

Publications

Deep learning predicts function of live retinal pigment epithelium from quantitative microscopy.

Schaub Nicholas J NJ Hotaling Nathan A NA Manescu Petre P Padi Sarala S Wan Qin Q Sharma Ruchi R George Aman A Chalfoun Joe J Simon Mylene M Ouladi Mohamed M Simon Carl G CG Bajcsy Peter P Bharti Kapil K

The Journal of clinical investigation 20200201 2

Increases in the number of cell therapies in the preclinical and clinical phases have prompted the need for reliable and noninvasive assays to validate transplant function in clinical biomanufacturing. We developed a robust characterization methodology composed of quantitative bright-field absorbance microscopy (QBAM) and deep neural networks (DNNs) to noninvasively predict tissue function and cellular donor identity. The methodology was validated using clinical-grade induced pluripotent stem ce ...[more]

PMID: 31714897

Similar Datasets

Project description:Lipofuscin accumulates with age in the retinal pigment epithelium (RPE) in discrete granular organelles and may contribute to age-related macular degeneration. Because previous studies suggest that lipofuscin contains protein that may impact pathogenic mechanisms, we pursued proteomics analysis of lipofuscin. The composition of RPE lipofuscin and its mechanisms of pathogenesis are poorly understood in part because of the heterogeneity of isolated preparations. We purified RPE lipofuscin granules by treatment with proteinase K or SDS and showed by light, confocal, and transmission electron microscopy that the purified granules are free of extragranular material and associated membranes. Crude and purified lipofuscin preparations were quantitatively compared by (i) LC MS/MS proteomics analyses, (ii) immunoanalyses of oxidative protein modifications, (iii) amino acid analysis, (iv) HPLC of bisretinoids, and (v) assaying phototoxicity to RPE cells. From crude lipofuscin preparations 186 proteins were identified, many of which appeared to be modified. In contrast, very little protein ( approximately 2% (w/w) by amino acid analysis) and no identifiable protein were found in the purified granules, which retained full phototoxicity to cultured RPE cells. Our analyses showed that granules in purified and crude lipofuscin preparations exhibit no statistically significant differences in diameter or circularity or in the content of the bisretinoids A2E, isoA2E, and all-trans-retinal dimer-phosphatidylethanolamine. The finding that the purified granules contain minimal protein yet retain phototoxic activity suggests that RPE lipofuscin pathogenesis is largely independent of associated protein. The purified granules also exhibited oxidative protein modifications, including nitrotyrosine generated from reactive nitrogen oxide species and carboxyethylpyrrole and iso[4]levuglandin E(2) adducts generated from reactive lipid fragments. This finding is consistent with previous studies demonstrating RPE lipofuscin to be a potent generator of reactive oxygen species and supports the hypothesis that such species, including reactive fragments from lipids and retinoids, contribute to the mechanisms of RPE lipofuscin pathogenesis.

Dataset Information

Deep learning predicts function of live retinal pigment epithelium from quantitative microscopy.

Publications

Deep learning predicts function of live retinal pigment epithelium from quantitative microscopy.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets