ATG5 is required for the differentiation of human eosinophilic leukemic cells, and ATG5^low-expressing human eosinophils are more susceptible to degranulation. (A) Immunoblotting. EoL-1 cells were transduced with lentiviral constructs containing Cas9 and guide RNA (gRNA) targeting ATG5. Whole-cell lysates of control and ATG5-knockout EoL-1 cells were collected and analyzed for the presence of ATG5 protein expression. Glyceraldehyde-3-phosphate dehydrogenase (GAPDH) served as loading control. Representative immunoblots of 2 independent experiments are shown. (B) Flow cytometry. Differentiation of control and ATG5-knockout EoL-1 cells was induced by the addition of 0.5 mM butyric acid, and the surface expression of eosinophil markers CD11b, Siglec-8, and CCR3 was assessed (n = 3). (C) Confocal microscopy. Formalin-fixed, paraffin-embedded human tissue sections of angiolymphoid hyperplasia (upper image), EoE (middle image), and sebaceous gland carcinoma (lower image) were stained with monoclonal mouse anti-EPX (green) and monoclonal rabbit anti-ATG5 (red) antibodies. Hoechst 33342 (blue) was used for nuclear DNA visualization. Representative confocal microscopy images are shown. Scale bars, 10 µm. Right: correlation analysis between the fluorescent intensity sums of EPX and ATG5, measured in all eosinophils, was performed for each tissue section, and the Pearson coefficient calculated in each case. (D) Correlation analysis between eosinophil-derived neurotoxin (EDN) levels in plasma and mRNA expression of ATG5 or ATG7 in blood eosinophils isolated from patients with hypereosinophilic syndrome (n = 19). Values are means ± standard error of the mean, or single data are presented in scatter dot plots in which the linear regression is indicated as blue lines. *P < .05; **P < .01; ***P < .001. AU, arbitrary unit; MFI, mean fluorescent intensity; n.s., not significant.