Supplemental Information. LKB1 Inactivation Elicits a Redox Imbalance. to Modulate Non-small Cell Lung Cancer Plasticity. and Therapeutic Response

Transcription

1 Cancer Cell, Volume 27 Supplemental Information LKB1 Inactivation Elicits a Redox Imbalance to Modulate Non-small Cell Lung Cancer Plasticity and Therapeutic Response Fuming Li, Xiangkun Han, Fei Li, Rui Wang, Hui Wang, Yijun Gao, Xujun Wang, Zhaoyuan Fang, Wenjing Zhang, Shun Yao, Xinyuan Tong, Yuetong Wang, Yan Feng, Yihua Sun, Yuan Li, Kwok-Kin Wong, Qiwei Zhai, Haiquan Chen, and Hongbin Ji

2 Supplemental Data 1

3 Figure S1, related to Figure 1. (A) Representative H&E staining, IHC staining of TTF1 and p63 in mixed Ad-SCC and ADC in KL mice at 8 weeks post Ad-Cre infection. AC: adenomatous component, SC: squamous component. Black scale bar: 50 µm, red scale bar: 10 µm. (B) Representative H&E staining, TTF1 and p63 IHC staining of KL SCC at 8 weeks, 10 weeks and weeks post Ad-Cre infection. Black scale bar: 50 µm, red scale bar: 10 µm. (C) Statistical analysis of the incidence of KL SCC with indicated p63 and TTF1 expression patterns at 8 weeks (n=70), 10 weeks (n=108) and weeks (n=220) post Ad-Cre infection. (D) Quantitative analysis of the levels (expressed as nm/g tissue) for individual metabolites or parameters in KL ADC 2

4 (n=6) and SCC (n=4) by LC-MS analysis. Data are shown as mean ± SEM. **p < 0.01, *p < (E) Representative IHC staining for HK2 and LDHA in KL ADC and SCC. Scale bar: 50 µm. (F) Western blotting analysis and relative quantification of G6PD and LDHA protein levels in KL ADC and SCC. (G) GSEA for GSH metabolism in KP ADC as compared to KL ADC. (H) IHC staining and statistical analysis of 8-oxo-dGuo + cells in KL ADC and KP ADC. Scale bar: 50 µm. Data are shown as mean ± SEM. ***p < (I) Representative IHC staining for NRF2 and NQO1 in KL ADC, KL SCC and KP ADC. Black scale bar: 50 µm, red scale bar: 10 µm. 3

5 Table S1, related to Figure 1. NAME SIZE ES NES P value FDR, q value FWER, p KEGG_DNA_REPLICATION KEGG_RIBOSOME KEGG_NUCLEOTIDE_EXCISION_REPAIR KEGG_MISMATCH_REPAIR KEGG_SPLICEOSOME KEGG_CELL_CYCLE KEGG_PROTEASOME KEGG_HOMOLOGOUS_RECOMBINATION KEGG_BASAL_TRANSCRIPTION_FACTORS KEGG_PYRIMIDINE_METABOLISM KEGG_BASE_EXCISION_REPAIR KEGG_STEROID_BIOSYNTHESIS KEGG_P53_SIGNALING_PATHWAY KEGG_AMINOACYL_TRNA_BIOSYNTHESIS KEGG_OOCYTE_MEIOSIS KEGG_GLUTATHIONE_METABOLISM KEGG_RNA_DEGRADATION KEGG_PURINE_METABOLISM Significantly enriched gene sets in KL SCC (n=5) as Compared to KL ADC (n=7). Datasets were described as previously (GSE6135) (Ji et al., 2007). Details of statistical methods are described in FDR, false-discovery rate; FWER, family-wise error rate; ES, enrichment score; NES, normalized enrichment score; Size refers to gene set size. 4

6 Table S2, related to Figure 1. NAME SIZE ES NES P value FDR, q value FWER, p KEGG_COMPLEMENT_AND_COAGULATION_CASCADES KEGG_LYSOSOME KEGG_ECM_RECEPTOR_INTERACTION KEGG_INSULIN_SIGNALING_PATHWAY KEGG_NEUROACTIVE_LIGAND_RECEPTOR_INTERACTION KEGG_CALCIUM_SIGNALING_PATHWAY KEGG_FOCAL_ADHESION KEGG_JAK_STAT_SIGNALING_PATHWAY KEGG_CYTOKINE_CYTOKINE_RECEPTOR_INTERACTION KEGG_FATTY_ACID_METABOLISM KEGG_TRYPTOPHAN_METABOLISM KEGG_GLYCOSPHINGOLIPID_BIOSYNTHESIS_GANGLIO_SERIES KEGG_VASCULAR_SMOOTH_MUSCLE_CONTRACTION KEGG_BASAL_CELL_CARCINOMA KEGG_AXON_GUIDANCE KEGG_DILATED_CARDIOMYOPATHY KEGG_VALINE_LEUCINE_AND_ISOLEUCINE_DEGRADATION KEGG_GAP_JUNCTION KEGG_CHEMOKINE_SIGNALING_PATHWAY KEGG_HYPERTROPHIC_CARDIOMYOPATHY_HCM KEGG_ADIPOCYTOKINE_SIGNALING_PATHWAY KEGG_PPAR_SIGNALING_PATHWAY KEGG_MELANOGENESIS KEGG_INTESTINAL_IMMUNE_NETWORK_FOR_IGA_PRODUCTION KEGG_HEDGEHOG_SIGNALING_PATHWAY KEGG_TYPE_II_DIABETES_MELLITUS KEGG_ETHER_LIPID_METABOLISM KEGG_LYSINE_DEGRADATION KEGG_CARDIAC_MUSCLE_CONTRACTION KEGG_PEROXISOME KEGG_HEMATOPOIETIC_CELL_LINEAGE KEGG_MATURITY_ONSET_DIABETES_OF_THE_YOUNG KEGG_REGULATION_OF_ACTIN_CYTOSKELETON KEGG_NATURAL_KILLER_CELL_MEDIATED_CYTOTOXICITY KEGG_GLYCOSAMINOGLYCAN_DEGRADATION KEGG_PATHWAYS_IN_CANCER KEGG_PYRUVATE_METABOLISM KEGG_FC_GAMMA_R_MEDIATED_PHAGOCYTOSIS Significantly Enriched Gene Sets in KL ADC (n=7) as Compared to KL SCC (n=5). Datasets were described as previously (GSE6135) (Ji et al., 2007). 5

7 Details of statistical methods are described in FDR, false-discovery rate; FWER, family-wise error rate; ES, enrichment score; NES, normalized enrichment score; Size refers to gene set size. 6

8 Table S3, related to Figure 1. NAME SIZE ES NES P value FDR, q value FWER, p KEGG_ALPHA_LINOLENIC_ACID_METABOLISM KEGG_NEUROACTIVE_LIGAND_RECEPTOR_INTERACTION KEGG_LINOLEIC_ACID_METABOLISM KEGG_ARACHIDONIC_ACID_METABOLISM KEGG_GLUTATHIONE_METABOLISM KEGG_GLYCOSPHINGOLIPID_BIOSYNTHESIS_GANGLIO_S ERIES KEGG_DRUG_METABOLISM_CYTOCHROME_P KEGG_METABOLISM_OF_XENOBIOTICS_BY_ CYTOCHROME_P KEGG_TRYPTOPHAN_METABOLISM KEGG_MATURITY_ONSET_DIABETES_OF_THE_YOUNG KEGG_AMINOACYL_TRNA_BIOSYNTHESIS KEGG_ETHER_LIPID_METABOLISM KEGG_LONG_TERM_DEPRESSION KEGG_PROXIMAL_TUBULE_BICARBONATE_RECLAMATION KEGG_RIBOSOME KEGG_PYRIMIDINE_METABOLISM KEGG_GLYCEROPHOSPHOLIPID_METABOLISM KEGG_VALINE_LEUCINE_AND_ISOLEUCINE_ DEGRADATION KEGG_ALANINE_ASPARTATE_AND_GLUTAMATE_METABOL ISM KEGG_VEGF_SIGNALING_PATHWAY Significantly Enriched Gene Sets in KP ADC (n=5) as Compared to KL ADC (n=7). Datasets were described as previously (GSE6135) (Ji et al., 2007). Details of statistical methods are described in FDR, false-discovery rate; FWER, family-wise error rate; ES, enrichment score; NES, normalized enrichment score; Size refers to gene set size. 7

9 8

10 Figure S2, related to Figure 2. (A) Different levels of LKB1 and 8-oxo-dGuo as detected by IHC staining in human lung ADC TMAs. Scale bar: 50 µm. (B) Summary and statistical analysis of LKB1 and 8-oxo-dGuo levels in human lung ADC TMAs. (C) Representative IHC staining for K5 and K14 in human lung ADC without LKB1 expression in TMAs. Scale bar: 50 µm. (D) Representative IHC staining for p63 and K5 on serial sections from human ADC with LKB1 mutation. Representative H&E staining and IHC staining for TTF1, p63 and K14 on these serial sections were shown in Figure 2B. Scale bar: 50 µm. (E) Representative IHC staining of TTF1, p63, 8-oxo-dGuo, NRF2 and NQO1 on serial sections from human ADC with LKB1 mutation. The red arrow indicates the weak IHC staining. Scale bar: 50 µm. (F) IHC staining for TTF1 in the adenomatous component and squamous component of human Ad-SCC without LKB1 expression. Scale bar: 50 µm. (G) Representative IHC staining of LKB1, p63 and TTF1 in human lung SCC TMAs. The SCC sample #1 was shown for negative TTF1 expression, while the SCC sample #2 was shown for positive TTF1 expression. Scale bar: 50 µm. 9

11 10

12 Figure S3, related to Figure 3. (A) Characterization of KL primary cells in culture and subcutaneous tumors in nude mice. (a) Representative epithelial morphology in culture; (b-d) representative H&E staining (b) and p63 IHC staining (c, d) of subcutaneous tumors in nude mice. Scale bar: 100 µm. (B) Cell death rate of mouse KP and KL cells under indicated culture conditions. NAC: 5 mm. Data are shown as mean ± SEM. ***p < (C) ROS levels in mouse KP and KL cells under indicated culture conditions. Results are expressed as the percentage change in the mean DCFH-DA values relative to the glucose-rich (25 mm) control. Data are shown as mean ± SEM. ***p < (D) LKB1 knockdown decreased pampk and pacc levels in KP cells. (E) Left: Comparison of glucose deprivation-elicited ROS accumulation in mouse KP cells with Luciferase (shluci) or LKB1 (shlkb1) knockdown. Results are expressed as the percentage change in the mean DCFH-DA values 11

13 relative to the glucose-rich (25 mm) control. Right: Comparison of glucose deprivation-induced cell death rate in KP cells KP cells with Luciferase control (shluci) or LKB1 (shlkb1) knockdown. Data are shown as mean ± SEM. *p < (F) Representative IHC staining of TTF1 and p63 in SCC of Ad-Cre-infected KL mice from saline and NAC treatment groups. Black scale bar: 50 µm, red scale bar: 5 µm. (G) Representative IHC staining of TTF1 and p63 in ADC, mixed Ad-SCC and SCC of KL mice at 15 weeks post Ad5-SPC-Cre infection. AC: adenomatous component, SC: squamous component. Scale bar: 50 µm. (H) Representative H&E staining of lung tumors of KL mice from saline and NAC treatment groups at 15 weeks post Ad5-SPC-Cre infection. Scale bar: 2 mm. (I) Representative IHC staining of TTF1 and p63 in SCC of Ad5-SPC-Cre-infected KL mice from saline and NAC treatment groups. Scale bar: 50 µm. (J) Nuclear location and statistical analysis of NRF2 expression in KL ADC and SCC. Black scale bar: 50 µm, red scale bar: 5 µm. (K) qrt-pcr analysis for the relative expression of indicated genes in KL cells with or without NRF2 overexpression. Representative data were shown from three independent experiments. Data are shown as mean ± SEM. ***p < 0.001, **p < 0.01, *p < (L) Left: comparison of glucose deprivation-elicited ROS accumulation in KL cells with or without NRF2 overexpression. Results are expressed as the percentage change in the mean DCFH-DA values relative to the glucose-rich (25 mm) control. Right: comparison of glucose deprivation-induced cell death rate in KL cells with or without NRF2 overexpression. Cell death rate was determined at 24 hr post glucose deprivation. Data are shown as mean ± SEM. ***p < (M) Statistical analysis of KL tumor burden and tumor area per mouse in Cre (n=15) and Nrf2-Cre (n=16) groups. Data are shown as mean ± SEM. (N) Representative IHC staining of NRF2 and NQO1 in KL ADC of the Cre and Nrf2-Cre groups. Scale bar: 50 µm. (O) Representative IHC staining of TTF1 and p63 in KL SCC of the Cre and Nrf2-Cre groups. Scale bar: 50 µm. 12

14 13

15 Figure S4, related to Figure 4. (A) GSEA plots for DNA replication and cell cycle in KL ADC as compared to KL SCC. (B) Relative mrna levels of PPP signature genes in KL ADC (n=4) and KL SCC (n=4). Data are shown as mean ± SEM. **p < 0.01, *p < (C) Representative IHC staining of G6PD in KL ADC, KL SCC and KP ADC. Scale bar: 50 µm. (D) GSEA plot for pyrimidine metabolism in KP ADC as compared to KL ADC. (E) Relative mrna levels (expressed as fold change) of PPP signature genes in KP ADC as compared to KL ADC. Representative data were shown from three independent 14

16 experiments. Data are shown as mean ± SEM. (F) GSEA plots for ECM-receptor interaction and focal adhesion in KP ADC as compared to KL ADC, and in KL ADC as compared to KL SCC. (G) Left: validation of Flag-G6PD overexpression in KL cells by western blotting; Right: comparison of ROS levels in KL cells with or without G6PD overexpression under normal culture condition. Data are shown as mean ± SEM. ***p < (H) Statistical analysis of KL tumor burden and tumor area per mouse in the Cre (n=12) and G6pd-Cre (n=15) groups. Data are shown as mean ± SEM. (I) Representative IHC staining (Ki67, 8-oxo-dGuo and G6PD), and statistical analysis of Ki67 IHC staining in KL mouse ADC of Cre and G6pd-Cre groups. Black scale bar: 50 µm, red scale bar: 5 µm. Data are shown as mean ± SEM. **p < (J) Representative IHC staining of TTF1 and p63 in KL SCC of Cre and G6pd-Cre groups. Scale bar: 50 µm. (K) Validation of shrna-mediated G6PD knockdown in KL cells and comparison of ROS levels in KL cells with or without G6PD knockdown under normal culture condition. Data are shown as mean ± SEM. ***p < (L) Comparison of cell growth by MTT assay in KL cells with or without G6PD knockdown. Data are shown as mean ± SEM. ***p < 0.001, *p < (M) Relative mrna levels (expressed as fold change) of SCC signature genes in KL cells with G6PD knockdown group as compared to that in Luciferase knockdown (shluci) control group. Representative data were shown from three independent experiments. Data are shown as mean ± SEM. (N) Representative IHC staining (G6PD, 8-oxo-dGuo and Caspase 3) and statistical analysis of 8-oxo-dGuo and Caspase 3 IHC staining in KL mouse lung tumor sections from shluci-cre and shg6pd-cre groups (8 weeks post viral infection). Black scale bar: 50 µm, red scale bar: 5 µm. Data are shown as mean ± SEM. ***p < 0.001, **p <

17 Table S4, related to Figure 4. NAME SIZE ES NES p value FDR q FWER,p KEGG_P53_SIGNALING_PATHWAY KEGG_COMPLEMENT_AND_COAGULATION_CA SCADES KEGG_FOCAL_ADHESION KEGG_ECM_RECEPTOR_INTERACTION KEGG_SYSTEMIC_LUPUS_ERYTHEMATOSUS KEGG_VIBRIO_CHOLERAE_INFECTION KEGG_LEISHMANIA_INFECTION Significantly Enriched Gene Sets in KL ADC (n=7) as Compared to KP ADC (n=5). Datasets were described as previously (GSE6135) (Ji et al., 2007). Details of statistical methods are described in FDR, false-discovery rate; FWER, family-wise error rate; ES, enrichment score; NES, normalized enrichment score; Size refers to gene set size. 16

18 17

19 Figure S5, related to Figure 5. (A) Left: Comparison of glucose deprivation-induced ROS accumulation (24 hr) in KL cells with or without caampk expression. Results are expressed as the percentage change in the mean DCFH-DA values relative to the glucose-rich (25 mm) control. Right: Comparison of glucose deprivation-induced cell death rate at indicated time in KL cells with or without caampk expression. Data are shown as mean ± SEM. ***p < 0.001, **p < 0.01, *p < (B) Statistical analysis of KL tumor burden and tumor area per mouse in the Cre (n=14) and caampk-cre (n=21) groups. Data are shown as mean ± SEM. (C) Representative IHC staining of TTF1 and p63 in KL SCC of Cre and caampk-cre groups. Scale bar: 50 µm. (D) 18

20 Western blotting analysis on Ad-Cre-infected KL mouse lung tumor lysates from saline and rapamycin treatment groups. (E) Representative H&E staining of KL mouse lung tumor sections from saline and rapamycin treatment groups at 10 weeks post Ad-Cre infection. Scale bar: 2 mm. (F) Statistical analysis of average tumor number in Ad-Cre-infected KL mice of saline (n=8) and rapamycin (n=7) treatment groups. Data are shown as mean ± SEM. **p < (G) Representative IHC staining of TTF1 and p63 in SCC of Ad-Cre-infected KL mice from the saline and rapamycin groups. Scale bar: 50 µm. (H) Statistical analysis of average tumor number in Ad5-SPC-Cre infected KL mice of saline (n=10) and rapamycin (n=11) groups at 15 weeks post viral infection. 5-week rapamycin treatment was initiated at 10 weeks post Ad5-SPC-Cre infection. Data are shown as mean ± SEM. *p < (I) Representative IHC staining of LC3 and p62 in KL and Kras G12D ;Lkb1 lox/lox ;Atg5 lox/lox (KLA) mouse lung tumors at 10 weeks post Ad-Cre infection. The red arrow indicates LC3 or p62 aggregates. Scale bar: 50 µm. (J) Statistical analysis of tumor burden and average tumor number in KL mice (n=8) and KLA (n=9) mice. Data are shown as mean ± SEM. (K) Representative IHC staining of TTF1 and p63 in of KLA ADC and SCC. Scale bar: 50 µm. (L) Representative IHC staining of CPT1A in KL ADC and SCC. Scale bar: 50 µm. (M) Relative mrna levels of FAO signature genes in KL ADC (n=4) and SCC (n=4). Data are shown as mean ± SEM. **p < 0.01, *p < (N) Representative Oil red O staining of lipids in KL ADC and SCC frozen sections. Black scale bar: 50 µm, red scale bar: 5 µm. (O) Representative IHC staining of CPT1A on KL ADC sections from Cre and Cpt1a-Cre groups. Scale bar: 50 µm. (P) Statistical analysis of KL tumor burden and tumor area per mouse in the Cre (n=12) and Cpt1a-Cre (n=14) groups. Data are shown as mean ± SEM. (Q) Representative IHC staining of TTF1 and p63 in KL SCC of the Cpt1a-Cre group. Scale bar: 50 µm. 19

21 Figure S6, related to Figure 6. (A) Representative IHC staining of different levels (relative score: 0, 1, 2) of pampk, pacc and LKB1 in human lung ADC TMAs. Scale bar: 50 µm. (B) Representative IHC staining of pampk and pacc in human lung ADC without LKB1 expression. Scale bar: 50 µm. (C) Representative IHC staining of LKB1, pampk and pacc in human lung SCC TMAs. Scale bar: 50 µm. (D) Representative IHC staining of p63 and CPT1A on serial sections of human Ad-SCC without LKB1 expression. AC: adenomatous component, SC: squamous component. Scale bar: 50 µm. 20

22 Figure S7, related to Figure 7. (A) Comparison of relative cell viabilities in LKB1-deficient cancer cell lines after treatment with BSO (0-10 mm), PEITC or PL (0-10 µm). The relative cell viabilities of the drug-free groups were set to 1.0. Data are shown as mean ± SEM. (B) Comparison of ROS levels in KL cells 21

23 under indicated treatments for 5 hr. PL: 10 µm, NAC: 5 mm. Results are expressed as the percentage change in the mean DCFH-DA values relative to control. Data are shown as mean ± SEM, ***p < (C) Representative IHC staining for Ki67 and Caspase 3 in ADC and SCC of Ad-Cre-infected KL mice from vehicle and PL treatment groups. Black scale bar: 50 µm, red scale bar: 5 µm. (D) Representative IHC staining and statistical analysis of 8-oxo-dGuo + cells in ADC and SCC of Ad-Cre-infected KL mice from vehicle and PL treatment groups. Scale bar: 50 µm. (E) Representative H&E staining, IHC staining of p63 and TTF1 in SCC of Ad-Cre-infected KL mice from PL treatment group. Black scale bar: 50 µm, red scale bar: 5 µm. (F) Representative H&E staining of lung tumors from Ad5-SPC-Cre-infected KL mice in vehicle and PL treatment groups at 15 weeks post viral infection. 3-week vehicle or PL treatment was initiated at 12 weeks post Ad5-SPC-Cre infection. Scale bar: 2 mm. (G) Representative IHC staining for TTF1 and p63 in ADC and SCC of Ad5-SPC-Cre-infected KL mice from PL treatment group. Scale bar: 50 µm. 22

24 23

25 Figure S8, related to Figure 8. (A) Statistical analysis of ROS levels in LKB1-deficient cancer cell lines after 5-hr phenformin treatment at indicated concentrations. Results are expressed as the percentage change in the mean DCFH-DA values relative to control. Data are shown as mean ± SEM, ***p < (B) Representative IHC staining of Ki67 in the lung tumors of Ad-Cre-infected KL mice from control and phenformin treatment groups. Black scale bar: 50 µm, red scale bar: 5 µm. (C) Representative IHC staining of Caspase 3 in the lung tumors of Ad-Cre-infected KL mice from control and phenformin treatment groups. Scale bar: 50 µm. (D) Representative IHC staining and statistical analysis of 8-oxo-dGuo + cells in the lung tumors of 24

26 Ad-Cre-infected KL mice from control and phenformin treatment groups. Scale bar: 50 µm. Data are shown as mean ± SEM, ***p < (E) Representative IHC staining of K5 and K14 in ADC of Ad-Cre-infected KL mice from the phenformin treatment group. Scale bar: 50 µm. (F) Statistical analysis of SCC incidence and average ADC number in Ad-Cre-infected KL mice treated with phenformin (n=10) or phenformin plus NAC (n=13) for 2 weeks. Tumor analysis was performed at 9 weeks post Ad-Cre infection. Data are shown as mean ± SEM. *p<0.05. (G) Representative H&E staining of lung tumors from Ad5-SPC-Cre-infected KL mice in control and phenformin treatment groups at 15 weeks post viral infection. Phenformin treatment was initiated at 12 weeks post viral infection. Scale bar: 2 mm. (H) Representative IHC staining of TTF1 and p63 in ADC and SCC of Ad5-SPC-Cre-infected KL mice from phenformin treatment group. Scale bar: 50 µm. (I) Relative expression of indicated genes in mouse KL cells (upper panel) or A549 cells (lower panel) after treatment with phenformin (5 mm for KL cells, 2 mm for A549 cells) for 24 hr. Representative data were shown from three independent experiments. Data are shown as mean ± SEM, ***p < 0.001, *p <

27 Supplemental Experimental Procedures Mouse Lung Tumor Collection Lung tumors ( 2mm in diameter) dissected from KL mice at 10 weeks post Ad-Cre infection were cut into several small pieces, one piece was used for histological inspection, and the rest pieces were snap frozen separately for western blotting and qrt-pcr analyses. Hematoxylin and Eosin (H&E) Staining and Immunohistochemistry (IHC) H&E staining and IHC staining was performed as previously (Gao et al., 2010; Han et al., 2014). Briefly, mouse lung tissues were inflated with 1 ml 4% paraformaldehyde (PFA), fixed overnight and dehydrated in ethanol, embedded in paraffin, sectioned (5 μm) followed by staining with hematoxylin and eosin (Sigma). For IHC staining, slides were deparaffinized in xylene and ethanol, and rehydrated in water. Antigen retrieval was performed by heating slides in a microwave for 20 minutes in sodium citrate buffer (ph 6.0) or 1 mm EDTA buffer. Slides were quenched in hydrogen peroxide (3%) to block endogenous peroxidase activity and then washed in TBST buffer. The primary antibodies were incubated at 4 o C overnight followed by using the SuperPicture Polymer Detection kit (Life Technologies) according to the manufacturer s instructions as described (Gao et al., 2010; Han et al., 2014). To calculate the percentages of cells positive for Ki67, Caspase 3 or 8-oxo-dGuo in IHC, at least high power fields from each mice, and totally over 100 high power fields were randomly selected for statistical analysis as 26

28 described (Gao et al., 2010). The following antibodies were used: 8-oxo-dGuo (ab48508, Abcam), NQO1 (ab34173, Abcam), p63 (SC-8431, Santa Cruz), LKB1 (13031, IHC Formulated, Cell Signaling), NRF2 (ab31163, Abcam), G6PD (HPA000834, Sigma), TTF1 (5883-1, EPITOMICS), Ki67 (NCL-Ki67p, Leica Biosystems), cleaved Caspase 3 (9664, Cell Signaling), pho-ampk Thr172 (2535, Cell Signaling), phosphor-acc1/2 Ser79 (3661, Cell Signaling), S6 (2217, Cell Signaling), ps6 (4858, Cell Signaling), Akt (9272, Cell Signaling), pho-akt (S473, 4060, Cell Signaling), Cpt1a (ab128568, Abcam), HK2 (2867, Cell Signaling), LC-3 (PM036, MBL), p62 (PM045, MBL), LDHA (2012, Cell Signaling), K5 (BS1208, Bioworlde) and K14 (PRB-155P, Covance). Bioinformatics Analysis The gene expression datasets have been submitted to the GEO repository under accession GSE6135 (Ji et al., 2007). The data sets for KP ADC, KL ADC and SCC were extracted for GSEA. We apply t test to calculate the t statistics of each probe. Next, GSEA was used to rank the probes and analyze the enrichment based on t statistics ( et al., 2005). The heatmap of the signature genes was generated in R by the function heatmap.2 based on their expression levels. Cell Lines and In vitro Assays Human lung cancer cell lines A549 and H460 were purchased from ATCC and 27

29 cultured as previously (Gao et al., 2010). To derive KL cell line, dissected KL mouse lung ADC was cultured in DMEM (Hyclone) supplemented with 10% FBS (Biochrom) and penicillin/streptomycin (Life Technologies). The medium was changed every other day until cells outgrew and stable cell line was established. The established primary cells were subcutaneously injected into the nude mice (2 x 10 6 cells/mice) and harvested 4 weeks later for analysis. The KP cell line was described previously (Han et al., 2014) For glucose deprivation assays, cells were firstly cultured in 6 or 24 well-plate in glucose-free DMEM (Life Technologies) supplemented with 25 mm D-glucose (Sigma). 24 hr later, the medium was changed and cells were washed with PBS, then cultured in glucose-free DMEM for another 24 hr followed by analysis. Cell death was performed by trypan blue exclusion assay, and the cell death rate was calculated by dividing number of dead cells to the total cell number as previously (Jeon et al., 2012). For MTT 3-(4, 5-Dimethylthiazol-2-yl)-2, 5-diphenyltetrazolium bromide (MTT) assays, cells were seeded in quadruplication in 96-well plate (2,500 cells/well), the cell number was determined at indicated time by multiskan spectrum microplate spectrophotometer as described (Gao et al., 2010). To compare the effects of ROS-inducing agents (Figure S7A), cells were seeded in quadruplication in 96-well plate (2,500 cells/well); 24 hr later, BSO, PEITC or PL (all from Sigma) was added at indicated concentrations, then the cells were cultured for another 72 hr followed by MTT assays. 28

30 To determine the intracellular ROS, cultured cells in 6 or 12 well-plate in FBS-free medium were treated with DCFH-DA (Beyotime) for 20 min, cells were then washed with PBS and collected as single-cell suspensions. Fluorescence was detected by FACS Calibur flow cytometer (Becton Dickinson). ShRNA, Plasmids, Lentivirus Production and Infection ShRNAs against mouse G6pd or Lkb1 were designed and built in lentiviral vector plko.1-puro and adapted vector plko.1-cre. ShRNA against luciferase gene was used as control. The shrna sequences were shown as follows: shluciferase: 5 -CTTACGCTGAGTACTTCGA-3, shg6pd-1: 5 -CAGTGCAAGCGTAATGAGC-3, shg6pd-2: 5 -CTGTCGAACCACATCTCCT-3, shlkb1-1: 5 -GGGTCACACTTTACAACATCA-3, shlkb1-2: 5 -GGGTCAGAATGGACAGAGCCA-3. The ORFs of mouse Nrf2, G6pd and Cpt1a were amplified from mouse SCC cdnas, and built in expression vectors pcdh-ef1-puro (Systems Biosciences) or pcdh-ef1-cre (adapted vector for in vivo expression in Lenti-Cre system). Lentivirus production and infection was performed as described (Gao et al., 2010; Han et al., 2014). Western Blotting The western blotting was performed as described (Gao et al., 2010; Han et al., 2014). Briefly, whole cell lysates were prepared in lysis buffer supplemented 29

31 with protease and phosphatase inhibitors (Roche). Equal amounts of protein were resolved by electrophoresis on gradient gels (Bio-Rad). The proteins were electro-transferred to PVDF membrane (Pall). Primary antibodies were incubated at 4 o C overnight and the HRP-conjugated secondary antibodies were incubated at room temperature for 1 hr. Enzymatic Activity Assays and Liquid Chromatography-tandem Mass Spectrometry (LC-MS) Analysis Individual tumors were dissected from KL mice at 10 weeks post Ad-Cre infection. The tumors were washed in cold PBS, weighed, snap frozen and immediately stored at -80 C before use. Tissues were homogenized in ice cold buffer containing 250 mm sucrose, 10 mm Hepes, 1 mm EDTA, appropriate protease inhibitor cocktail, ph 7.4. After centrifuged at 1000 g for 20 minutes, supernatants were collected for enzyme assays. The glucose-6-phosphate dehydrogenase (G6PDH) activity assay was performed at 37 in a 0.2 ml media containing 100 mm KCl, 25 mm Hepes, 5 mm MgCl2, 2 mm glucose-6-phospate (G6P), and 0.5 mm NADP, ph 7.4. The absorbance at 340 nm was monitored spectrophotometrically in 96 well microplates, and one mill-unit of enzyme activity was defined as 1 nm NADPH produced. The long chain acyl-coa synthetase (ACS) activity was assayed in 100 mm KCl, 25 mm Hepes, 5 mm MgCl2, 4 mm ATP, 2 mm oleate, 1 mm HSCoA, 0.4 mm NADH, 1 mm PEP, 2 U/ml myokinase, 2 U/ml PK and 2 U/ml LDH, ph

32 The absorbance at 340 nm was recorded and one mill-unit of enzyme activity was defined as 1 nm NADH reduction. The Carnitine palmitoyltransferase-1 (CPT-1) activity was measured in reaction buffer containing 100 mm KCl, 25 mm K2HPO4, 25 M palmitoyl-coa, 2 mm L-carnitine, 125 μm 4-PDS (4, 4 -dipyridyldisulfide), ph 7.4. The absorbance at 324 nm was monitored and one mill-unit of enzyme activity was expressed as 1 nm 4-thiopyridone produced. For LC-MS analysis, the metabolites were extracted and analyzed as previously described (Wu et al., 2011). Briefly, frozen tumor tissues (~20 mg) were homogenized in 80% acetonitrile by TissueLyser II (Qiagen). The denatured proteins were separated by centrifuging at g for 20 minutes. Metabolites were separated by an Xbridge Amide Column ( mm, 3.5 m, Waters), at a flow rate of 0.3 ml/min. Metabolites levels were determined by AB 4000Q Tripe coupled to Aligent 1200 HPLC system. The data analysis was performed with the Applied Biosystems Analyst software Oil red O Staining Fixed frozen tissue sections were air-dried and stained with prewarmed Oil red O solution (0.5% in propylene glycol) for 15 min in a 60 C oven, differentiated in 85% propylene glycol solution for 1 min, and rinsed with distilled water twice. Slides were then lightly stained with hematoxylin for 10 s, rinsed with distilled water, and mounted for microscopy. Reverse Transcription and Quantitative PCR Analysis 31

33 Equal amounts of total RNA was retrotranscribed into first-strand cdna using RevertAid First Strand cdna Synthesis Kit (Fermentas). The cdnas were then used for real time PCR on a 7500 Fast Real-Time PCR System (Applied Biosystems) using SYBR-Green Master PCR mix (Roche). Β-actin or GAPDH (for human genes) served as internal control. The primers for Q-PCR were summarized in the following table. Primer sequences used for qrt-pcr analysis. Gene Forward Primer (5-3 ) Reverse Primer (5-3 ) Actin TGAGCGCAAGTACTCTGTGTGGAT ACTCATCGTACTCCTGCTTGCTGA Krt5 GGCCCACAGAGACTGCTTCTTT AACATTTTGGGGTCTGGGTCAC Krt14 CAGCGGCCCACTGAGATCAAAG CTTGGTCCGGAAGTCATCGGCA Myc CAACGACAGCAGCTCGCCCA AGCCCGACTCCGACCTCTTGG Dlx5 TCTCTAGGACTGACGCAAACA GTTACACGCCATAGGGTCGC Klf5 CCTCCGTCCTATGCCGCTACAA GCTTCTCGCCCGTATGAGTCCT Sccro AGAGTGTAAAAGGATCGTTGGAC CTGGATCGAGGGCCAGATCA Perp ATCGCCTTCGACATCATCGC CCCCATGCGTACTCCATGAG Trp63 GTGTTGGTGTGGCACAGGGG TCTTCCCCACAGCTCTGGCT Gclm AGGAGCTTCGGGACTGTATCC GGGACATGGTGCATTCCAAAA Gsta4 AGCCATTTTGATGGTGGAAG ACCTTAAAGCACGCTGCACT Gstya1 CGCCACCAAATATGACCTCT CCTGTTGCCCACAAGGTAGT Gsto1 ATCCGGCACGAAGTCATCAAT TGACAGATTCGGTGACCAAGT Cat GAGACCTGGGCAATGTGAT GTTTACTGCGCAATCCCAAT Nqo1 TTCTCTGGCCGATTCAGAGT GGCTGCTTGGAGCAAAATAG Gpx2 GCCTCAAGTATGTCCGACCTG GGAGAACGGGTCATCATAAGGG Nrf2 CTCGCTGGAAAAAGAAGTG CCGTCCAGGAGTTCAGAGAG G6pd CACAGTGGACGACATCCGAAA AGCTACATAGGAATTACGGGCAA Pgd ATGGCCCAAGCTGACATTG GCACAGACCACAAATCCATGAT Gpi1 TCAAGCTGCGCGAACTTTTTG GGTTCTTGGAGTAGTCCACCAG Tkt ATGGAAGGTTACCATAAGCCAGA TGCAGCATGATGTGGGGTG Pfkl GGAGGCGAGAACATCAAGCC CGGCCTTCCCTCGTAGTGA Pfkp GAAACATGAGGCGTTCTGTGT CCCGGCACATTGTTGGAGA Dera TCGCAGAAGCGTGAAGAAGG CTGCCCGGATTGGATATTTGG Rpe GCACCTGGATGTAATGGACGG CCTGGCCTAGCTGCTTTCG Cpt1a GATCATGGTTAACAGCAACT AGGCATAAACGCAGAGCATTC 32

34 Cpt1b GCACACCAGGCAGTAGCTTT CAGGAGTTGATTCCAGACAGGTA Cpt1c TCTTCACTGAGTTCCGATGGG ACGCCAGAGATGCCTTTTCC Hmgcs2 GAAGAGAGCGATGCAGGAAAC GTCCACATATTGGGCTGGAAA Fatp TCTGTTCTGATTCGTGTTCGG CAGCATATACCACTACTGGCG Pdk4 AGGGAGGTCGAGCTGTTCTC GGAGTGTTCACTAAGCGGTCA Ucp3 CTGCACCGCCAGATGAGTTT ATCATGGCTTGAAATCGGACC Hadha TGCATTTGCCGCAGCTTTAC GTTGGCCCAGATTTCGTTCA GAPDH CAGGTGGTCTCCTCTGACTT CCAAATTCGTTGTCATACCA TP63 GGAAAACAATGCCCAGACTC GTGGAATACGTCCAGGTGGC KRT5 CGAGCAGTACATCAACAACCTCAG GGCAGCATCTACATCCTTCTTCAG KRT14 CTCTGGAGGAGGCCAACGCC TTGGCATTGTCCACTGTGGCTGTG KLF5 TCGGATGAGCTGACCCGCCA TCAGAGCGCGAGAAGCTGCG DLX5 AAGCGCCACCAACCAGCCAG CGTTCCGGCAAGGCGAGGTA 33

35 Supplemental References Subramanian, A., Tamayo, P., Mootha, V. K., Mukherjee, S., Ebert, B. L., Gillette, M. A., Paulovich, A., Pomeroy, S. L., Golub, T. R., Lander, E. S., and Mesirov, J. P. (2005). Gene set enrichment analysis: a knowledge-based approach for interpreting genome-wide expression profiles. Proc Natl Acad Sci U S A 102, Wu, J., Zhang, F., Yan, M., Wu, D., Yu, Q., Zhang, Y., Zhou, B., McBurney, M. W., and Zhai, Q. (2011). WldS enhances insulin transcription and secretion via a SIRT1-dependent pathway and improves glucose homeostasis. Diabetes 60,