Fig. 5

Minimizing the formation of megaclusters caused by fusion reduces cell loss between S5D3 and S6D7. A Representative aggregate morphology between S4D4 and S7D8, scale bar = 500 µm. B Ki67 positive cells (n = 3–6 biological replicates, 3 donor human islet preps; one-way ANOVA with Tukey post-test) based on flow cytometry. C Stage-specific net growth rate (n = 6–18 biological replicates; one-way ANOVA with Tukey post-test), D doubling times n = 5–18 biological replicates; one-way ANOVA with Tukey post-test) and E viability quantification using NC-200 (n = 8–21 biological replicates; one-way ANOVA with Tukey post-test). F Aggregate size distribution after testing different strategies to decrease cluster fusion (plot shows individual aggregates with median and interquartile range). G The modified protocol, which results in smaller aggregates, had significantly higher cell recovery during S5D3 and S6D7 (n = 8- 18 biological replicates; Mixed-effects analysis with Sidak post-test). H Representative viability of S6D7, S7D8 aggregates and donor human islets using calcein-AM and ethidium bromide, scale bar = 500 µm. (All box and whisker plots show individual points with median and interquartile range. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001. S4D4 = stage 4 day 4, S5D3 = stage 5 day 3, S6D7 = stage 6 day 7, S7D8 = stage 7 day 8, S7D10 = stage 7 day 10