Method Profile
A 2,707-compound screen uses longitudinal cone imaging in one human iPSC background, followed by damaging/protective hit retesting, kinase profiling and selected genetic tests.
Identify photoreceptor-protective or damaging compounds while separating target annotation, causal evidence and delayed toxicity.
30-week retinal organoids from the human iPSC line 01F49i-N-B7; approximately 20,000 organoids overall and six independent productions in the primary screen.
Low-glucose seven-day discovery with same-organoid baseline/follow-up cone counts; normal-glucose damaging-hit screen and low-glucose protective-hit screen; selected 14-day, rod, oxidative-stress and mouse follow-up.
Primary design
2,707 known-target compounds at 10 μM, five replicate organoids each; approximately 15,000 organoids on 175 plates. Four normal- and four low-glucose 0.1% DMSO control wells per plate.
Reporter validation
Cone-specific reporter 97% ± 1% specificity and 64% ± 8% labeling efficiency, mean ± standard deviation, n = 5 each.
Secondary design
33 damaging candidates in normal glucose and 31 protective candidates in low glucose at 0.01, 0.1, 1 and 10 μM, with five replicates/condition.
Mechanistic comparisons
CS-KI-1/CS-KI-2 versus nonprotective structural analogs; 350-kinase profiling; additional CK1/MAPK11 inhibitors; cone-specific knockdown. 50 μM H2O2 supplies alternative injury.
Cell-type expression comparison
Bulk RNA sequencing of sorted cones and rods after seven days compares CS-KI-1/2 with their nonfunctional analogs and low-glucose controls, alongside HSP90 inhibition. Gene-set changes are interpreted as associations.
The source covariance model adjusts for initial cone counts, with multiple-testing correction. Adjusted screening scores can exceed 100%; the smaller follow-up experiments use raw survival. Seven- and 14-day comparisons identify delayed changes in initially protective responses.
Five-replicate discovery/secondary screens with orthogonal chemistry, reporter checks, genetic and alternative-injury follow-up.
One human iPSC line; approximately 15,000 primary-screen organoids across six independent productions, with five replicate organoids per compound and secondary condition.
Normal/low glucose, vehicle, nonprotective analogs, selected inhibitors, scrambled knockdown and hydrogen-peroxide injury.
The primary screen identified 146 significantly damaging compounds; the first Results count reports 29 confirmations among 33 secondary retests. Four of 31 protective retests were significant. HSP90 inhibition protected cones at day 7 but was detrimental at day 14. CS-KI-1/2 protected cones through day 14; only CS-KI-2 retained significant day 14 rod protection. CK1-family knockdown improved cone survival; MAPK11 knockdown did not. Mouse CS-KI-1 outer nuclear layer preservation was nonsignificant (P = 0.069).
A 2,707-compound screen found protective and damaging responses in human retinal organoids. Follow-up implicated casein kinase 1, while HSP90 inhibition switched from short-term cone protection to harm at day 14 and a MAPK11 knockdown failed to improve cone survival.
