Fig. 1

Autophagy promotes tumorigenesis in vivo. a A549 cells deficient in ATG5 protein expression were generated by CRISPR/Cas9-mediated genome editing. Each of the two wild-type control clones (WT and WT-2) and ΔATG5 clones (ΔATG5 and ΔATG5-2) were immunoblotted for the indicated proteins. b Genomic DNA from A549 ΔATG5 and ΔATG5-2 cells was PCR amplified. The amplicon encompassed the sequence to which Cas9 had been targeted (PAM in bold) and different indels were identified in both clones via Sanger sequencing. c A549 WT and ΔATG5 cells were plated at low confluency for time-lapse phase-contrast videomicroscopy using an Incucyte microscope and cell proliferation was monitored by automated confluency analysis at set intervals post plating (means, n = 9 wells, ±S.D.). d A pooled derivative of ΔATG5 cells was generated by stable transduction with GFP-ATG5 retrovirus (rescue). The indicated cell lines were immunoblotted as shown. e WT, ΔATG5 and rescue cells were subcutaneously injected into immunocompromised mice and tumour volume was measured longitudinally (means, n = 10 flanks, ±S.E.M., *P < 0.05 vs. WT, two-tailed t-test). f WT-2 and ΔATG5-2 cells were compared for tumour growth after subcutaneous injection into immunocompromised mice (means, n = 12 flanks, ±S.E.M., *P < 0.05 or **P < 0.01 vs. WT-2, two-tailed t-test). g At the end of tumour growth in e, control tumours and sufficiently large ΔATG5 tumours were fixed and stained for LC3B via immunohistochemistry (DAB stain, arrows indicate regions of LC3B puncta, scale bar = 10 µm). Representative images are shown here. Uncropped blots are available in Supplementary Fig. 10