Fig 2: MARK3 is activated by metabolic stress inducers.a Immunoblotting of chemical screening for metabolic stress inducers. Among tested compounds, anisomycin and thapsigargin increase the kinase-activated phosphorylation of doxycycline (DOX)-inducible MARK3. b Immunoblotting of chemical screening for DNA damage inducers with anisomycin as a positive control. DNA damage does not directly increase the phosphorylation of DOX-inducible MARK3. c Immunoblotting of inhibition experiments for the upstream regulators of MARK3. LKB1 knockdown interferes with DOX-inducible MARK3 activation in response to anisomycin treatment. d CRISPR–Cas9 experiment for DOX-inducible MARK3 knockout in TYK-nu cells. TYK-nu is a high-grade serous ovarian carcinoma (HGSOC)-like cell line, harboring the oncogenic TP53 R175H mutation. The copy numbers of LKB1, TAOK1, PIM1, and MARK3 are duplicates, suggesting that MARK3 can become active in TYK-nu. The genomic profile of TYK-nu is sourced from the Cancer Cell Line Encyclopedia (CCLE). e, f Cell viability assay for DOX-inducible MARK3 knockout in TYK-nu. The cytotoxic effects of anisomycin (e) and thapsigargin (f) are amplified upon MARK3 knockout. Relative cell number shows the relative ratio between the number of cells in the absence of the inhibitors (0 nM = 1.0). and the number of cells in the presence of various concentrations of anisomycin or thapsigargin. Error bars represent mean ± standard deviation (SD) of three biological replicates. Statistical analysis was performed using unpaired Student’s t-test (Anisomycin 250 nM; Thapsigargin 100 nM; *P < 0.01).

Fig 3: Lkb1 restoration in established lung tumors dramatically decreases lung tumor burden.a Schematic of the XTR cassette inserted within the first intron of Lkb1 in the eXpressed, Trapped, and Restored conformations. The XTR cassette is composed of an inverted gene trap consisting of an Ad40 splice acceptor upstream of eGFP. The gene trap is flanked by heterotypic loxP sites to allow for stable Cre-mediated inversion, which results in the truncation of wild-type Lkb1 transcripts. The gene trap is nested between two FRT sites to enable FLPo-ERT2 -mediated deletion in the presence of tamoxifen, which results in the restoration of wild-type Lkb1 transcripts and LKB1 protein. b Assessment of the impact of Lkb1 restoration on tumor burden. Lung tumors were initiated in KT, KT;Lkb1XTR/XTR, and KT;Lkb1XTR/XTR;FLPo-ERT2 mice. At 6 weeks post-initiation, tumor-bearing mice were treated for 6 weeks with corn oil vehicle or tamoxifen prior to analysis. IFU infectious units, VEH vehicle, TAM tamoxifen. c Representative fluorescence (top) and hematoxylin–eosin (H&E) staining (bottom) images of tumor-bearing lungs from KT, KT;Lkb1XTR/XTR, and KT;Lkb1XTR/XTR;FLPo-ERT2 mice treated with vehicle or tamoxifen. Lung lobes within fluorescent images are outlined in white. Top scale bars = 5 mm. Bottom scale bars = 2 mm. d, e Tumor area (d) and tumor size (e) as assessed by histology for tumor-bearing KT, KT;Lkb1XTR/XTR, and KT;Lkb1XTR/XTR;FLPo-ERT2 mice at 12 weeks after tumor initiation, following 6 weeks of treatment with vehicle or tamoxifen. In d, each dot represents a mouse, while each dot in e corresponds to a tumor. Red crossbars indicate the mean. In d, KT-vehicle, n = 3 mice; KT-tamoxifen, n = 3 mice; KT;Lkb1XTR/XTR-vehicle, n = 5 mice; KT;Lkb1XTR/XTR-tamoxifen, n = 5 mice; KT;Lkb1XTR/XTR;FLPo-ERT2-vehicle, n = 4 mice; KT;Lkb1XTR/XTR;FLPo-ERT2-tamoxifen, n = 4 mice. In e, KT-vehicle, n = 11 tumors; KT-tamoxifen, n = 14 tumors; KT;Lkb1XTR/XTR-vehicle, n = 147 tumors; KT;Lkb1XTR/XTR-tamoxifen, n = 135 tumors; KT;Lkb1XTR/XTR;FLPo-ERT2-vehicle, n = 167 tumors; KT;Lkb1XTR/XTR;FLPo-ERT2-tamoxifen, n = 19 tumors. One tissue section per mouse was analyzed. P values calculated by two-sided unpaired t test. VEH vehicle, TAM tamoxifen.

Fig 4: C/EBP transcription factors constrain oncogenic KRAS-driven lung tumor growth.a Interrogation of the tumor-suppressive capacity of a series of LKB1-dependent genes and C/EBP transcription factors (targets listed in Supplementary Fig. 16b). Lenti-sgRNA/Cre vectors targeting each candidate gene were pooled prior to delivery into KT and KT;H11LSL-Cas9 mice (method outlined in Supplementary Fig. 16b). b Bulk tumor-bearing lungs were analyzed by Tuba-seq. Percentile plot depicting tumor size at several percentiles relative to the distribution of tumors initiated with Lenti-sgRNA/Cre vectors encoding inert sgRNAs (sgNeo1, sgNeo2, sgNeo3, sgNT1, sgNT2). Lkb1 and C/ebp family targeting vectors are shown. Each vector has a distinct fill color, and fill saturation indicates percentile. Colored fill indicates that tumor size at a given percentile is significantly different from inert sgRNAs, while grayscale indicates no significant difference. Error bars indicate 95% confidence intervals centered on the mean of relative tumor size at a given percentile. c Validation of C/EBP factors as suppressors of oncogenic KRAS-driven tumor growth in a non-multiplexed format. Tumors were initiated in KT and KT;H11LSL-Cas9 mice with either Lenti-sgInert/Cre (sgNeo1/sgNT1/sgNeo2; sgInert) or Lenti-sgCebpa/b/d/Cre (sgCebpa/sgCebpb/sgCebpd; sgCebpa/b/d). N = 5 mice per genotype-virus cohort. d Representative fluorescence (top) and H&E (bottom) images of tumor-bearing lungs from KT and KT;H11LSL-Cas9 mice transduced with either Lenti-sgInert/Cre or Lenti-sgCebpa/b/d/Cre. Lung lobes within fluorescent images are outlined in white. N = 5 mice per genotype-virus cohort. Top scale bars = 5 mm. Bottom scale bars = 2 mm. e, f Quantification of tumor area (e) and tumor size (f) by histological examination of tumor-bearing lungs from KT and KT;H11LSL-Cas9 mice transduced with either Lenti-sgInert/Cre or Lenti-sgCebpa/b/d/Cre. Each dot represents either individual mice (e) or individual tumors (f). Red crossbars indicate the mean. In e, n = 5 mice per genotype-virus cohort. One tissue section per mouse was analyzed. In f, KT-sgInert, n = 22 tumors; KT-sgCebpa/b/d, n = 23 tumors; KT;H11LSL-Cas9-sgInert, n = 58 tumors; KT;H11LSL-Cas9-sgCebpa/b/d, n = 167 tumors. P values were calculated by a two-sided unpaired t test.

Fig 5: Transcriptional screens identify LKB1 as a regulator of CCL2 APrincipal component analysis (PCA) of RNAseq expression levels from MDCK cells with inducible knockdown of Lkb1 (Lkb1‐i) after tetracycline incubation (PC1: −Tet vs. +Tet). n = 2 biological replicates (PC2).BPCA of microarray expression data from littermate control and Lkb1 ΔTub kidneys at 5 weeks (PC1). n = 5 mice each (PC2). Replicates are enclosed in convex hulls to guide the eye.CNetwork representation of significantly enriched Biological Processes GO terms (commonly upregulated genes) derived from a hypergeometric test (P < 0.05). Nodes are connected if they have a semantic similarity > 0.6, and clusters are named according to the most representative GO terms. See also Dataset EV4.D–H Ccl2 mRNA expression evaluated by qRT–PCR in MDCK cells expressing inducible shRNA after tetracycline induction (Tet) against Lkb1‐i2 (D, n = 4), Lkb1‐i1 (E, n = 3) after 5 h of 1 mM AICAR treatment, Anks3‐i2 (F, n = 5), Nek7‐i2 (G, n = 5), PKD1 KO1 Lkb1‐i1, and PKD1 KO2 Lkb1‐i1 (H, n = 3). See also Appendix Fig S3B, F, G and S.Data information: In (D–H), mean ± SD. Paired t‐test, *P < 0.05, **P < 0.01, ns: not statistically different.