Precision genetic cellular models identify therapies protective against endoplasmic reticulum stress

Abstract

Congenital disorders of glycosylation (CDG) and deglycosylation (CDDG) are a collection of rare pediatric disorders with symptoms that range from mild to life threatening. They typically affect multiple organ systems and usually present with neurological abnormalities including hypotonia, cognitive impairment, and intractable seizures. Several genes have been implicated in the thirty-six types of CDG, but currently NGLY1 is the only known CDDG gene. A common biological mechanism among CDG types and in CDDG is endoplasmic reticulum (ER) stress. Here, we develop two isogenic human cellular models of CDG (PMM2, the most prevalent type of CDG, and DPAGT1) and of the only CDDG (NGYL1) in an effort to identify drugs that can alleviate ER stress. Systematic phenotyping identified elevated ER stress and autophagy levels among other cellular and morphological phenotypes in each of the cellular models. We screened a complex drug library for compounds able to correct aberrant morphological phenotypes in each of the models using an agnostic phenotypic cell painting assay based on >300 cellular features. The image-based screen identified multiple candidate compounds able to correct aberrant morphology, and we show a subset of these are able to correct cellular and molecular defects in each of the models. These results provide new directions for the treatment of rare diseases of glycosylation and deglycosylation and a framework for new drug screening paradigms for more common neurodegenerative diseases characterized by ER stress.Here we utilize a morphological profiling and screening paradigm to identify agents that protect against the cellular stresses resulting from CDG and CDDG causal mutations. We focus specifically on mutations in PMM2 and DPAGT1, and in NGLY1, which causes the only reported CDDG (6). We used genetic engineering to generate clinically relevant CDG and CDDG genotypes in a karyotypically normal human cell line (hTERT RPE-1) in order to create cellular models amenable to mutation-specific phenotype identification. The hTERT RPE-1 line was selected to allow for morphology-based high content imaging screens to identify phenotypes that are consequences of ER stress. These CDG and CDDG cell lines were used in high-content small molecule screens to identify compounds that revert the imaging phenotypes caused by these mutations. Specifically, we selected 1049 annotated compounds for screening representing a broad chemical space and multiple target classes. In order to validate the performance of the screen, we selected 16 compounds that were ranked amongst the best at phenotype reversion in the screen (protective compounds) and 10 compounds that did not affect aberrant phenotypes (non-active negative control compounds). We then evaluated these compounds in assays designed to test how well they revert mutational phenotypes in our three cellular models. Results Establishment of precise human cellular models of CDG and CDDGGenome editing was used to generate hTERT RPE-1 cell lines that mimic genotypes associated with...