Fig 2: KDM5A sustains ASCL1 levels in SCLC. (A,B) Immunoblot analysis of GLC16 cells (A) and NCI-H1876 cells (B) 14 d after infection with lentiviruses expressing Cas9 and the indicated sgRNAs. (C) Immunoblot analysis of histone extracts from GLC16, NCI-H1876, and NCI-H69 SCLC cells treated with the indicated concentrations of KDM5-C70 for 3 d. (D–I) Immunoblot analysis (top) and proliferation assays (bottom) of GLC16 cells (D,G), NCI-H1876 cells (E,H), and NCI-H69 cells (F,I) treated with the indicated concentrations of KDM5-C70. (J) Representative phase-contrast images of adherent NCI-H69 cells after treatment with the indicated concentrations of KDM5-C70 for 24 d. Scale bar, 10 µm. (K) Immunoblot analysis of NCI-H69 cells infected with a doxycycline (DOX)-inducible lentivirus that expresses ASCL1 or EV. The cells were treated with KDM5-C70 at the indicated concentrations for 12 d. DOX (2 µg/mL) was added to cells 48 h prior to the addition of KDM5-C70 where indicated by the “+”. (LE) Long exposure, (SE) short exposure. (L) Number of viable cells, relative to untreated controls, of cells treated as in (K). n = 5 biological replicates. (**) P < 0.01 (two-way ANOVA comparing either ASCL1- vs. ASCL1+ or EV+ vs. ASCL1+).

Fig 3: Kdm5a represses Notch2 and Notch signaling in SCLC in vivo. (A) Representative H&E and IHC for Kdm5a and Notch2 in SCLC lung tumors from LSL-Cas9 mice injected with either sgKdm5a RPP or sgControl RPP adenoviruses. Scale bar, 100 µm. (B) Quantification of percent Notch2-positive tumor cells by IHC from A. n = 11 lung tumors (sgKdm5a RPP) or 19 lung tumors (sgControl RPP). (C) Gene set enrichment analysis (GSEA) using the hallmarks gene sets of NOTCH signaling (left), epithelial mesenchymal transition (middle), and KRAS signaling up (right), from the RNA-seq data obtained from lung tumors arising in the LSL-Cas9 mice injected with sgKdm5a RPP compared with sgControl RPP adenoviruses. n = 8 lung tumors for each group. q-values are indicated. (D) Immunoblot analysis of cell lines maintained on J2 fibroblast feeder layers (see Materials and Methods) from SCLC tumors that grew in LSL-Cas9 mice injected with sgKdm5a RPP or sgControl RPP adenoviruses. (E) Volcano plot of the RNA-seq data showing the log2fold change in gene expression in sgKdm5a RPP tumors compared with sgControl RPP tumors. Dlk1 is a canonical Notch inhibitory ligand. (F,G) Dlk1 (F) and Nell1 (G) mRNA levels, as determined by qRT-PCR, in samples used for the RNA-seq experiment above. n = 8 for sgKdm5a RPP, n = 8 for sgControl RPP. For all experiments, data are represented as ±SEM. (*) P < 0.05. (**) P < 0.01.

Fig 4: Downregulation of KDM5A reduced CP-induced apoptosis in HEI-OC1 cells. A–C Effects of KDM5A on early (Annexin V-FITC positive, PI negative) and late (Annexin V-FITC/PI double positive) cell apoptosis were (A) determined via flow cytometry and (B, C) quantified. The data are presented as three independent experiments' mean ± SEM values. ****P < 0.0001 compared with the control; ####P < 0.0001 compared with CP only. D–F Representative images and histogram analysis of CellROX™ green in HEI-OC1 cells. Scale bar = 10 μm. The data are presented as three independent experiments' mean ± SEM values. ***P < 0.001, ****P < 0.0001 compared with the control; ####P < 0.0001 compared with CP only. G–I Representative images and histogram analysis of Rhodamine 123 in HEI-OC1 cells. Scale bar 10 μm. The data are presented as three independent experiments' mean ± SEM values. ****P < 0.0001 compared with the control; ###P < 0.001 compared with CP only. J–N OCRs and bioenergetics parameters in HEI-OC1 cells from each group. The data are presented as mean ± SEM values. **P < 0.01, ***P < 0.001, ****P < 0.0001 compared with the control; ##P < 0.01, ####P < 0.0001 compared with CP only. The experiments were performed in triplicate

Fig 5: High glutamine decreases H3K4me3 methylation.a Tumours from the control (Ctrl) or glutamine-supplemented (High Gln) groups were used for histone extraction and assessed by western blotting using the indicated histone antibodies. b Schematic of the modified cell-culturing conditions in vitro to mimic tumour microenvironment. c M229 cells cultured in 96-well plate as in b. Cell viability was assessed using CellTiter Glo assay (n = 3, independent experiments). d M229 cells, cultured as in b, were lysed and used for histone extraction and immunoblotting with the indicated antibodies. e M249 cells, cultured as in b, were assessed using Trypan blue exclusion (n = 3, independent cell cultures). f M229 cells, cultured as in b, were used for qPCR analysis (n = 3, independent cell cultures). g M229 cells cultured as in b with or without 3 mM DM-αKG. Cell viability was assessed using CellTiter Glo assay (n = 3, independent cell cultures). h, i M229 cells cultured as in b and treated with DMSO, PBIT (5 μM), JIB04 (1 μM) or GSKJ4 (9 μM). Cells were assessed for viability by CellTiter Glo assay in h (n = 3, independent experiments) or collected for western blotting using the indicated antibodies in i. j M249 cells cultured as in h. Live cell numbers were assessed using Trypan blue exclusion (n = 3, independent experiments). k, l M229 cells were transfected with siRNA against scramble control (siCtrl), JARID1A, JARID1B or JARID1C and cultured as in b. Live cell numbers were assessed using Trypan blue exclusion in k (n = 3, independent cell cultures) or collected for immunoblotting as indicated in l. m siCtrl, siJARID1A, siJARID1B or siJARID1C were collected for histone extraction and immunoblotting. Immunoblots were quantified using ImageJ. n siCtrl, siJARID1A, siJARID1B or siJARID1C cells cultured as in b were collected for qPCR analysis (n = 3, independent cell cultures). All immunoblots are representative of three independent experiments. Data represent means and error bars are s.d. p value calculated by t test (unpaired, two tailed). *p < 0.05, **p < 0.01, ***p < 0.001. Source data are provided as a Source data file.