The AMPK-Related Kinases SIK1 and SIK3 Mediate Key Tumor-Suppressive Effects of LKB1 in NSCLC

Abstract

Mutations in the LKB1 (also known as STK11) tumor suppressor are the third most frequent genetic alteration in non-small cell lung cancer (NSCLC). LKB1 encodes a serine/threonine kinase that directly phosphorylates and activates 14 AMPK family kinases ("AMPKRs"). The function of many of the AMPKRs remains obscure, and which are most critical to the tumor-suppressive function of LKB1 remains unknown. Here, we combine CRISPR and genetic analysis of the AMPKR family in NSCLC cell lines and mouse models, revealing a surprising critical role for the SIK subfamily. Conditional genetic loss of Sik1 revealed increased tumor growth in mouse models of Kras-dependent lung cancer, which was further enhanced by loss of the related kinase Sik3. As most known substrates of the SIKs control transcription, gene-expression analysis was performed, revealing upregulation of AP1 and IL6 signaling in common between LKB1- and SIK1/3-deficient tumors. The SIK substrate CRTC2 was required for this effect, as well as for proliferation benefits from SIK loss. SIGNIFICANCE: The tumor suppressor LKB1/STK11 encodes a serine/threonine kinase frequently inactivated in NSCLC. LKB1 activates 14 downstream kinases in the AMPK family controlling growth and metabolism, although which kinases are critical for LKB1 tumor-suppressor function has remained an enigma. Here we unexpectedly found that two understudied kinases, SIK1 and SIK3, are critical targets in lung cancer.This article is highlighted in the In This Issue feature, p. 1469.

Figure 1.. A lentiviral CRISPR-based screen targeting the AMPK-Related (AMPKR) kinase family in human NSCLC cells reveals the SIK subfamily mediates tumor suppressive phenotypes of LKB1.

(A) Left: schematic of the AMPK-Related (AMPKR) kinase family tree branch. Right: RPKM (Reads per kilobase million) plot showing the relative mRNA expression of the AMPKR kinase family in human A549 (LKB1-null) and LBK1 WT-reconstituted cells. (B-D) Western blots on lysates from A549-LKB1 WT cells treated with lentiviral sgRNAs targeting the AMPKR family. Control sgRNA-treated A549 (LKB1-null) and A549-LKB1 WT cells also shown. (B) Western blots of lysates from A549-LKB1 WT cells targeted to co-delete NUAK1+NUAK2 (left), and SNRK (right). (C) Western blot of lysates from A549-LKB1 WT cells targeted to co-delete MARK1+MARK4, MARK2+MARK3, and BRSK1+BRSK2. (D) Left: Western blots of lysates from A549-LKB1 WT cells targeted to delete AMPK catalytic subunits α1/2. Right: Western blots of lysates from A549-LKB1 WT cells targeted to delete the SIK subfamily, either with individual guides against SIK1, SIK2, and SIK3, or using combinations of guides to co-delete SIK1+SIK2, SIK1+SIK3, and SIK1+2+3. (E) Western blots on lysates from A549-pBabe and A549-LKB1 WT cells targeted to co-delete SIK1 + SIK3 or SIK1+SIK2+SIK3. SIK kinases are only phosphorylated in the presence of LKB1 and cell lines targeted for SIK deletion show loss of phosphorylation only on the targeted SIK proteins, validating deletion in these lines. (F) Soft agar colony formation of AMPKR-deficient cells. Average colony number after 3 weeks of growth in soft agar is shown. Assays were done two times in triplicate wells and the values represent the average of independently generated cell lines for a given AMPKR as described in Figure S1E. *P<0.05, **** P<0.0001 by ANOVA and post hoc student’s t-test. (G) Representative images of soft agar colony growth of A549 (LKB1-null) cells, A549 LKB1-WT (“L”) cells and AMPKR-deficient A549 LKB1-WT cells.

Figure 2.. CRISPR-based and GEMM-based mouse models of SIK family kinase inactivation in Kras-driven NSCLC demonstrate acceleration of tumor growth upon SIK1 loss.

(A) Schematic of the experimental design using pSECC lentivirus to deliver Cre-recombinase, Cas9 and a sgRNA of choice as a single payload to the lungs of KrasG12D-mice (K) and Kras-p53 floxed mice (KP) after intratracheal virus delivery. (B-E) Analysis of inactivating Sik1 or Lkb1 with pSECC-sgRNA viruses in the lungs of K mice. (B) Longitudinal bioluminescent imaging (BLI) data from induced tumors in K mice treated with pSECC control (sgTom, N=4 or sgLacZ in Figure 2C, N=3), sgLkb1 (N=10), and three independent sgSik1 viruses (N=11). Average bioluminescence (photon flux) at each imaging time point is shown for each cohort. The last point shown indicates study endpoint. (C) Quantitation of tumor burden. Tumor area was calculated as a percent of total lung area from H&E-stained sections for each mouse. The average tumor burden for each cohort is shown as a fraction of tumor area/total lung. (D) Quantitation of individual tumor size in lungs of K mice treated with pSECC-control (sgTom) and pSECC-sgSik1-viruses. The area of individual tumors was quantitated in mm² from H&E sections for each mouse and are shown in a dot-plot per cohort for each sgRNA guide. Each dot represents a single tumor, with a bar denoting the average tumor size per cohort. (E) Distribution in tumor size observed in pSECC-control and pSECC-sgSik1-treated mice. Individual tumors from all mice in each cohort were binned by size and are shown as a percent for each size bin in a pie chart. (F-I) Analysis of inactivating the SIK kinase family or LKB1 with pSECC-sgRNA viruses in the lungs of KP mice. (F) Longitudinal BLI data from KP mice treated with control pSECC virus or pSECC viruses harboring sgRNA guides to target mouse LKB1 (N=6), SIK1 (N=10), SIK2 (N=9) and SIK3 (N=7). Average bioluminescence (photon flux) at each imaging time point is shown for each cohort. The last point shown indicates study endpoint. (G) Tumor burden analysis for all pSECC-treated cohorts of KP mice. Tumor area was calculated as a percentage of total lung area. The average tumor burden for each cohort treated with individual sgRNA guides is shown as a fraction of tumor area/total lung. (H) Quantitation of individual tumor size (area in mm²) in pSECC control and pSECC-SIK1-treated KP mice. The number of tumors per cohort treated with individual sgRNA guides targeting SIK1 is shown. Each dot represents an individual tumor, with a bar denoting the average tumor size per cohort. (I) Tumor size distribution in KP mice treated with control and pSECC-SIK1 viruses. Individual tumors were binned by size and are represented as percentages of different size bins shown in a pie chart. All values in (B-I) are expressed as means ± s.e.m. *P<0.05, ** P<0.01, *** P< 0.001 determined by student’s t test. (J) Schematic of the experimental design using GEMM mice with floxed alleles of Sik1, Sik2, and Stk11 (LKB1) in a conditional K-RasG12D background (KSik1, KSik2, KL), and Sik1 and Stk11 in a conditional KP genetic background (KPSik1 and KPL). Genetic conditional activation of Kras and conditional knockout of the listed genes in the lung is generated by Cre-recombinase-mediated IoxP recombination after intratracheal delivery of pGK-CRE-virus. (K-N) Analysis of tumor growth, endpoint tumor burden, and tumor size distribution in floxed K, KSik1, KSik2 and KL mice. (K) Longitudinal bioluminescent imaging (BLI) data from induced lung tumors in floxed K (N=13), KSik1 (N=12), KSik2 (N=6), and KL (N=6) mice after PGK-Cre-mediated recombination. Average bioluminescence (photon flux) at each imaging time point is shown for each cohort. The last point shown indicates study endpoint. (L) Tumor burden analysis of floxed K, KSik1, KSik2 and KL mice. Tumor area was calculated as a percent of total lung area per mouse, and shown as the average fraction of tumor/lung for each cohort. (M) Quantitation of tumor size (area in mm²) in lung tumors from floxed K and and KSik1 mice. Each dot represents an individual tumor, with a bar denoting the average tumor size per cohort. (N) Tumor size distribution in floxed K and KSik1 mice. Individual tumors were binned by size and are represented as percentages of different size bins shown in a pie chart. (O-R) Analysis of tumor growth, endpoint tumor burden, and tumor size distribution in floxed KP and KPSik1 mice. (O) Longitudinal bioluminescent imaging (BLI) data from induced lung tumors in floxed KP (N=8), KPSik1 (N=5), and KPL (N=5) mice after PGK-Cre-mediated recombination. Average bioluminescence (photon flux) at each imaging time point is shown for each cohort. The last point shown indicates study endpoint. (P) Tumor burden analysis of floxed KP, KPSik1 and KPL mice. Tumor area was calculated as a percent of total lung area per mouse, and shown as the average fraction of tumor/lung for each cohort. (Q) Quantitation of tumor size (area in mm²) in individual lung tumors from GEMM KP and KPSik1 mice. Each dot represents an individual with a bar denoting the average tumor size per cohort. (R) Tumor size distribution in GEMM KP and KPSik1 mice. Individual tumors were binned by size and represented as percentages of different size bins shown in a pie chart. (S) Colony growth in soft agar of mouse tumor-derived KPSik1 and KP cells lines. Data represent the combined average of n=4 independent KP lines and n=3 independent KPSik1 lines from assays performed in triple wells two times. Graph shows average colony number at a threshold to detect large colonies (50-pixel range). All values in (K-S) are expressed as means ± s.e.m. *P<0.05, ** P<0.01, *** P< 0.001, ****P<0.001 determined by student’s t test.

Figure 3.. Combined effect on tumor growth by genetic loss of SIK1 with CRISPR-mediated inactivation of SIK3 in a Kras-driven mouse model of NSCLC.

(A) Schematic of the experimental design using pSECC lentivirus to deliver Cre-recombinase, Cas9 and a Sik1 or Sik3 sgRNA as a single payload to the lungs of KrasG12D-mice (K) and Kras-Sik1 floxed mice (KSik1) after intratracheal virus delivery. (B) Longitudinal bioluminescent imaging (BLI) data from induced lung tumors in GEMM KSik1 mice treated with pSECC control (sgTom) (N=7) or pSECC-sgSik3-B (N=7), and K mice treated with pSECC control (sgTom) (N=7), sgLkb1 (N=3 at study endpoint), or two sgRNA viruses targeting Sik1: sgSik1-D, (N=7), sgSik1-E (N=5), or Sik3: sgSik3-B (N=9), sgSik3-C (N=6). Average bioluminescence (photon flux) at each imaging time point is shown for each cohort. The last point shown indicates study endpoint. (C) BLI data of lung tumors from study endpoint for K and KSik1 mice treated with control sgTom and sgSik guides. Each dot represents the endpoint BLI value for each mouse overlayed on a bar graph denoting the mean for each group. (D) Overlay of BLI images at study endpoint for K+sgTom, KSik1+sgTom and KSik1+sgSik3 tumor-bearing mice. Shown are N=6 images for mice with the highest amounts of luminescence emitted from lung tumors within each cohort. N=2 images from K+sgLkb1 mice are shown for comparison. Scale bar represents photons/sec/cm²/sr. (E) Tumor burden analysis of K+sgTom, KSIK+sgSik3, KSik1+sgTom, and K+sgSik1 and K+sgSik3 mice. Tumor area was calculated as a percent of total lung area per mouse, and shown as the average fraction of tumor/lung for each cohort. (F) Quantitation of tumor size (area in mm²) in individual lung tumors from K+sgLkb1, K+sgTom, KSik1+sgSik3, KSik1+sgTom, and K+sgSik1 and K+sgSik3 mice. Each dot represents an individual with a bar denoting the average tumor size per cohort. (G) BrdU positivity in lung tumors from K+sgLkb1, K+sgTom, KSik1+sgSik3 or sgTom, and K+sgSik1 or sgSik3 mice. Data is displayed as percent BrdU+ cells over total cell number for each tumor. Each dot represents a single tumor within each cohort. (H) Lung tumor size distribution in Kras (K) and KSik1 mice treated with control or sgSik3 pSECC viruses (KSik1 mice), or control and sgLkb1, sgSik1, and sgSik3 pSECC viruses (K mice). Individual tumors were binned by size and represented as percentages of different size bins shown as a pie chart. All values in (B-H) are expressed as means ± s.e.m. *P<0.05, ** P<0.01, *** P< 0.001, ****P<0.001 determined by student’s t test.

Figure 4.. Transcriptional analysis of lung tumors derived from KSik1+sgSik3 and K+sgLkb1 mice reveal a shared core gene signature enriched for upregulation of genes involved in IL6/JAK/STAT and AP-1 signaling pathways.

(A) Transcriptional profile heatmap denoting unbiased clustering of mouse tumors from the following genotypes: K+sgTom (KTom) (n=9), KSik1+sgSik3 (n=6), and KL-A (n=10). Tumor RNA was analyzed by high throughput whole-transcriptome sequencing (RNAseq). Samples clustered by genotype in an unbiased manner. (B) Differential expression analysis on tumors from (A) (using fold change +/− log2 0.59, FDR <0.05 cut-offs) reveals that 38% of upregulated genes and 41% of downregulated genes which are LKB1-dependent respond in the same, statistically significant manner to SIK1+SIK3 inactivation. (C) Cluster of the 971 genes commonly regulated by inactivation of LKB1 or SIK1+3 inactivation identified in Figure 4B. (D) GSEA plots from KSik1+sgSik3 vs. KTom data queried against the “Hallmark” group of gene sets. Six gene sets were independently identified in both the K+sgLkb1 and KSik1+3 dataset queries. Two of these gene sets, shown in (D), represent biological processes well-established to be induced upon LKB1 deletion. The EMT signature was the top enriched gene set, IL6_JAK_STAT3_Signaling was 8^th most enriched gene set. EMT heatmap shows FPKMs for all enriched genes in EMT GSEA plot. IL6 heatmap shows FPKMs for all enriched genes in IL6 GSEA plot combined with Metascape IL6 category genes (Fig S10B). (E-G) Analysis of the 344 upregulated genes common to both K+sgLkb1 and KSik1+sgSik3. (E) Enrichr analyses identify IL-6 pathway enrichment only within upregulated genes common to K+sgLkb1 and KSik1+sgSik3. (F) Homer de novo Motif Enrichment analysis on the upregulated genes common to K+sgLkb1 and KSik1+sgSik3. The only significantly enriched motif (P=1e-12) is shown. The enriched motif corresponds to the JUND/FOS/FOSL1/AP-1/FRA2 motif, JUND motif is pictured. (G) Heatmap of genes with AP1+CRE elements in the promoter (+/−2kb of TSS). (H) FPKM plots for genes from the IL-6/JAK/STAT3 signaling pathway. Cytokines (IL6,IL33), receptors (IL4ra), core components (Jak3, Stat3), co-activators (Bcl3, Osmr), and canonical downstream STAT3 transcriptional targets (Socs3) mRNAs were all upregulated in both LKB1-deficient and SIK1+SIK3-deficient tumors.

Figure 5.. Functional analysis of murine and human NSCLC cell lines bearing LKB1 or SIK kinase family inactivation reveals CRTC2 dependent-induction of cytokine signaling and control of cell proliferation and growth under anchorage-free conditions.

(A) ELISA analysis of IL-6 secretion in mouse KP lines. KP cells lines were transduced with a sgRNA guide targeting Lkb1 or with two independent sets of guides to target Sik1 and Sik3 simultaneously. 500,000 control or KO cells were seeded in 6 well plates and supernatant was collected at 48hrs to quantitate the amount of IL6 released by the cells into the media. (B) ELISA analysis of IL-6 secretion in human A549 NSCLC cells. Two independent cell lines bearing distinct guides to simultaneously target Sik1 and Sik3 were generated each in pBabe cells (LKB1-null) and in cells reconstituted with Lkb1 cDNA. Additionally, the LKB1-reconstituted sgSik1/3 cell line #2 was further targeted with one of two guides to target expression of Crtc2 or a control guide. 500,000 cells from each genotype were seeded and supernant was collected at 48 hrs and IL6 secretion was quantitated by ELISA. (C) Relative expression of genes commonly regulated by LKB1, SIK1+3 kinases and CRTC2. qPCR was performed on pBabe and LKB1-expressing A549s, ± sgSik1+3 and ± sgCrtc2 guides to assess modulation of SIK1+3 and CRTC2-dependent transcription. For all genes examined, mRNA expression was plotted relative to LKB1-expressing control line ( A549 LKB1 + sgCtI). (D) Immunoblots from sgSIK1+3 A549 (Control or Lkb1-reconstituted) and sgSIK1+3+sgCrtc2 cell lines were assessed for expression and phosphorylation status of SIK kinases (E) Proliferation of cells under anoikis (attachment-free) conditions. Control pBabe or Lkb1-expressing A549 cells expressing guides targeting Sik1 + Sik3 +/− sgCrtc2 guides were challenged with detachment conditions for 5 days. The number of surviving cells was quantitated with cyquant staining. Data are means of triplicate wells ± s.d. (F) Left: Fold change in colony growth in soft agar of LKB1-expressing A549 cells expressing guides targeting Sik1 + Sik3 +/− sgCrtc2 guides. Colonies were counted at four weeks and the data are shown as the average of two biological replicates of triplicate wells for each genotype ± s.e.m. Right: representative images of colony formation in soft agar at experimental endpoint. (G) Western immunoblots of lysates from Lkb1-null pBabe or Lkb1-expressing A549s, further expressing sgRNA guides targeting all three SIK kinases. Lysates were collected from cells under adherent and suspension conditions. Immunoblots depict phosphorylation status of the SIKs and additional AMPKRs (AMPK, MARKs), direct SIK kinase substrates (CRTCs, HDACs) and AMPKR targets or downstream transcriptionally regulated genes.