Mutant LKB1 Confers Enhanced Radiosensitization in Combination with Trametinib in KRAS-Mutant Non–Small Cell Lung Cancer

Lung cancer is the leading cause of death from cancer in the United States (1) and is commonly managed with various combinations of surgery, chemotherapy, and radiotherapy. The 5-year overall survival rate for patients with lung cancer of any stage is 17%; for advanced disease, that rate is only 4% (1), indicating the need for improved therapeutic strategies. About 85% of lung cancers are non–small cell lung cancer (NSCLC), and KRAS is the most often mutated oncogene in NSCLC (2). KRAS-mutant NSCLC is a genetically and functionally complex disease with several subtypes, mainly defined by frequent comutations in p53 and LKB1 (STK11; ref. 3). p53 plays important roles in genome stability, DNA damage response, senescence, apoptosis, and cell-cycle regulation (4). The p53 signaling pathway is activated by several kinds of stress signals. The protein level of p53 is strictly regulated by MDM2, an E3 ubiquitin ligase. In cells without stress, MDM2 monoubiquitinates p53. Once cells are under stress, the interaction between p53 and MDM2 will be disrupted, stabilizing the p53 protein to trigger downstream signals (5). Liver Kinase B1 (LKB1), also named Serine/threonine kinase 11 (STK11), is a tumor-suppressor gene that is mutated in 15% to 30% of NSCLCs (6). A germline mutation in LKB1 is responsible for Peutz–Jeghers syndrome (7). LKB1 directly regulates the AMPK-mTOR pathway to control cell growth and metabolism, rendering LKB1-mutant cells resistant to metabolic stress (8). Mutations in LKB1 in patients with lung cancer are strongly associated with KRAS mutations and a history of smoking (9). Several studies have illustrated the relationship between LKB1 mutation and response to therapies, including chemotherapy, radiotherapy, targeted therapy, and immunotherapy (10–14).

In KRAS-mutant NSCLC, p53 and LKB1 mutations define the two major subtypes of lung cancer having distinct gene expression profiles and responses to therapies (3). In one study of genetically modified mice, Kras-Lkb1–mutated cells produced squamous tumors, whereas Kras or Kras-p53–mutated cells produced adenocarcinomas (15). Chen and colleagues showed that the effect of docetaxel could be enhanced by the MEK inhibitor selumetinib in Kras and Kras-p53 mouse lung cancer models but not in the Kras-Lkb1 model (16). Herter-Sprie and colleagues induced single lung tumor nodules in genetically modified mice implanted with Kras, Kras-p53, or Kras-Lkb1-mutated cells by injecting Adeno-Cre and treating the tumors by radiotherapy. They found that radiotherapy effectively controlled tumor growth in the Kras and Kras-p53 tumor models, but Kras-Lkb1 tumors had grown rapidly at 4 weeks after treatment (14). Collectively, these findings suggest that KRAS-LKB1 subtype may be more resistant to conventional therapies; therefore, specific therapeutic strategies should be developed.

Radiotherapy is a part of standard treatments for lung cancer (17). However, KRAS-mutant lung cancers can be radioresistant (18–20). The ability to sensitize tumor cells with molecular targeted drugs can increase the probability of tumor control in response to similar or even lower radiotherapy doses. In a previous high-content drug screen for radiosensitizers, we identified several MEK inhibitors as being radiosensitizers in KRAS-mutant lung cancer (21). Further preclinical studies indicated that one of those MEK inhibitors, trametinib, sensitized KRAS-mutant NSCLC cell lines to radiotherapy both in vitro and in vivo (21). Trametinib is an MEK1/2 inhibitor that has been approved for the treatment of BRAF V600E–mutant metastatic melanoma and is currently in phase I clinical trials in combination with chemoradiation for stage III NSCLC (NCT01912625). However, the effects of trametinib vary considerably among different cell lines, suggesting that additional mechanisms contribute to the radiosensitization effect. We undertook this study to identify a specific subgroup of KRAS-mutant NSCLC that can be radiosensitized by trametinib and to investigate mechanisms mediating resistance to trametinib sensitization, with the ultimate goal of generating novel strategies to overcome such resistance.

All of the human NSCLC cell lines were obtained from the American Type Culture Collection or provided by Dr. John Minna (University of Texas Southwestern Medical Center). Mouse lung adenocarcinoma cell lines were obtained from Dr. Tyler Jacks (Massachusetts Institute of Technology) and Dr. Jonathan Kurie (MD Anderson Cancer Center). All cell lines were maintained in RPMI1640 + glutamine medium supplemented with 10% FBS in a 37°C incubator with 5% carbon dioxide. Cell lines were authenticated every 6 months of use at the Characterized Cell Line Core Facility at MD Anderson Cancer Center with the short tandem repeat method. All cell lines were regularly tested for mycoplasma.

Trametinib, AICAR (Acadesine), and hydroxychloroquine sulfate (HCQ) were purchased from Selleck Chemicals. Cells and mice were irradiated with a Shepherd mark I-68A Cesium irradiator. Dosimetry and experimental set-up were done as previously described (22).

The lentiCRISPRv2 plasmid was provided by Dr. Feng Zhang (Massachusetts Institute of Technology) through Addgene (#52961). Guide DNA sequences were synthesized by Integrated DNA Technologies and cloned to the plasmid according to the methods described elsewhere (23). Oligo sequences are listed in Supplementary Table S1. The constructed lentiCRISPRv2, pRSV-Rev, pMDLg/pRRE, and pMD2G plasmids were transfected to 293FT cells by TransIT-LT1 (Mirus) to generate lentivirus. The 293FT cell medium was changed at 24 hours after transfection, and at 48 hours, the supernatant was used to infect the target cells. Single-cell–derived stable clones were obtained after 2 to 3 weeks of puromycin selection (2–5 μg/μL). Sanger sequencing of the targeted DNA region and Western blotting of targeted proteins were done to validate successful knockouts.

The siRNA knockdown experiments were done as described previously (24). siRNAs were transfected into cells at 24 hours before treatment with radiotherapy, trametinib, or both, and cell lysates were collected at 24 hours after treatment. Western blotting was done to test the knockdown efficiency. The siRNAs used in the study were SignalSilence Atg5 siRNA I (Cell Signaling Technology; #6345), SignalSilence Control siRNA (unconjugated; Cell Signaling Technology; #6568), and AMPK α1 siRNA (m) (Santa Cruz Biotechnology; sc-29674).

Whole-length LKB1 was cloned by PCR from human cDNA and then ligated to SFB (S-protein/Flag/SBP)–tagged plasmid driven by a CMV promoter. Cells were transfected by using the TransIT-LT1 (Mirus) according to the manufacturer's instructions. At 48 hours after transfection, the cells were treated with trametinib followed 4 hours later by radiotherapy or single treatment alone. Transfection efficiency was tested by Western blotting.

The clonogenic survival assay was carried out as we previously described (22, 24). Briefly, cells were trypsinized to single-cell suspension and then seeded in 6-well plates 16 hours before treatment. Each treatment condition had three to six replicates. After treating by vehicle (dimethylsulfoxide) or trametinib (10–60 nmol/L) for 4 to 5 hours, cells were treated with 0, 2, 4, or 6 Gy of radiation. Drugs were washed out 24 hours later, and the cells maintained for 2 to 3 weeks. Clonogenic survival after chemoradiation was assayed as reported previously (22). Briefly, the chemotherapy agents (7.02 nmol/L paclitaxel and 48.46 nmol/L carboplatin) were mixed and used to treat cells together with trametinib followed 4 hours later with radiotherapy. Medium was changed after 24 hours. After the colonies were fixed and stained by 0.1% crystal violet in 20% methanol, the colonies consisting of at least 50 cells were counted. Data were analyzed by using SigmaPlot 10.0 (Systat Software Inc.) or GraphPad Prism 7.0 (GraphPad Software, Inc.). The dose enhancement ratio 0.5 (DER_0.5) was calculated as follows: DER_0.5 = (Radiation dose correlated with 50% survival fraction in DMSO-treated cells)/(Radiation dose correlated with 50% survival fraction in drug-treated cells). A drug or drug combination that generated a DER_0.5 of ≥1.2 was considered a radiosensitizer.

LKR13-scrambled and LKR13-Lkb1-KO cells were seeded in 6-well plates at a density of 5 × 10⁴ cells per well. Cells were trypsinized and counted with a hemocytometer after 1, 2, and 3 days of culture. Three wells of cells were counted at each time point for each cell type. The relative cell numbers were calculated and plotted by using GraphPad Prism 7.0 (GraphPad Software, Inc.).

For the beta-gal histology staining, a kit for staining senescent cells was purchased from Sigma (CS0030). Briefly, cells were plated in 6-well plates 1 day before treatment. The next day, cells were treated with trametinib with or without radiotherapy 4 to 5 hours later. After a 72-hour incubation, cells were fixed and stained according to the manufacturer's protocol. Five to six pictures of stained cells at random fields were taken with an inverted microscope, and then the blue-stained cells (senescence-positive) and nonstaining cells in each picture were counted on a computer. The positive percentages were calculated and presented by using GraphPad Prism 7.0 (GraphPad Software, Inc.). For the senescence detection by flow cytometry, a Senescence Detection Kit was purchased from BioVision (K991). Cells were treated for 48 hours before being tested following the manufacturer's protocol. Senescent-positive cells were detected on a BD C6 flow cytometer in the FL1 channel. Results were analyzed by FlowJo software (V10).

The CellROX Green Reagent was purchased from ThermoFisher (C10444). Cells were seeded in 6-well plates 1 day prior to the treatment. Next morning, cells were treated by trametinib (30 nmol/L) followed by radiotherapy (6 Gy) 4 to 5 hours later. Six hours after radiotherapy, cells were incubated with 5 μmol/L CellROX Green Reagent for 30 minutes and then trypsinized and washed by PBS. Cells were analyzed by flow cytometer according to the manufacturer's protocol.

Western blotting was done as previously described (22). Antibodies were purchased from Cell Signaling Technology and Abcam. For Western blot of tumor protein lysates, tumors were snap frozen by liquid nitrogen and then homogenized by a Brinkmann Polytron Homogenizer (PT 10-35) and lysed in RIPA buffer (Cell Signaling Technology) supplemented with Protease Inhibitor Cocktail (ThermoFisher; 78430). Antibodies used in this study were LC3B (CST, 3868 for human; Abcam, 51520 for mouse), GAPDH (CST, 2118), HSP90 (CST, 4874), Actin (Sigma, A2228), LKB1 (CST, 3047), p53 (CST, 2524 for mouse; Santa Cruz Biotechnology, 47698 for human), phospho-p53-Ser15 (CST, 9284), p21 (Abcam, 109199), AMPK (CST, 5832), phospho-AMPK-Thr172 (CST, 2535), ULK1 (CST, 8054), phospho-ULK1-Ser555 (CST, 5869), ERK1/2 (CST, 4695), phospho-ERK1/2 (CST, 4370), ATG5 (CST, 8540), MDM2 (CST, 86934), and phospho-MDM2-Ser166 (CST, 3521). Quantification of band intensity was carried out by the volume tools within the Image Lab software (Bio-Rad).

All experiments involving animals were approved by the Institutional Animal Care and Use Committee of The University of Texas MD Anderson Cancer Center. All animal experiments were carried out in accordance with the institutional guidelines and the approved protocol. Male 129S background mice were purchased from the Jackson Laboratory. When the animals were 8 to 10 weeks old, 5 × 10⁵ LKR13 or LKR13-Lkb1 KO cells were s.c. injected into the right thighs of the mice. When the tumors reached 10 mm in diameter, the mice were randomly assigned to the following treatment groups: control; chemoradiation; trametinib; chemoradiation + trametinib; HCQ; and HCQ + chemoradiation + trametinib. Chemotherapy (10 mg/kg paclitaxel and 30 mg/kg carboplatin) was administered on day 1 by i.p. injection. Radiotherapy was given as 2 Gy once daily for 5 days, with a jig used to confine the mice and expose the tumor to radiotherapy as previously reported (22). Trametinib was administered daily at a dose of 2 mg/kg for 5 days by oral gavage. The trametinib was dissolved in DMSO and diluted 1:9 in 1% carboxymethylcellulose (medium viscosity) and 0.4% Tween-80 (Sigma). HCQ was dissolved in PBS and administered at a daily dose of 60 mg/kg for 5 days by i.p. injection. Tumor diameters were measured 2 to 3 times each week, and tumor volumes were calculated as 0.5 × length × width × width.

Three animals from each treatment group were sacrificed at day 7 of the experiment. Lungs were fixed in 10% formalin overnight and embedded in paraffin. Three sections representing different locations from each sample were stained by hematoxylin and eosin (H&E). Total number of metastatic nodules from each section was counted under a microscope.

Because comutations in KRAS can affect cells' response to therapies, we tested a panel of human KRAS-TP53–mutated (KP) and KRAS-LKB1–mutated (KL) NSCLC cell lines. DER_0.5 greater than 1.2 was considered sensitization. Clonogenic survival assays indicated that trametinib radiosensitized KL cells, but not KP cells (Fig. 1A–C).

We observed that for the sensitive KL cell line H460, combination trametinib plus radiotherapy produced enlarged, flat cells, which was not observed in the KP cell line H441 (Supplementary Fig. S1A and S1B). Because this morphologic change often represents senescence, we assessed the senescence marker beta-galactosidase and found that trametinib and radiotherapy were able to induce strong beta-galactosidase activity in H460 KL cells but not in H441 KP cells (Fig. 2A and B). The senescence marker p21 was also upregulated in the A549 and H460 KL cell lines at 48 hours after treatment, but was not upregulated in H441 KP cells. Moreover, treatment with combined radiotherapy plus trametinib led to higher p21 levels in A549 and H460 cells than did either treatment alone, suggesting that the combination robustly induced senescence (Fig. 2C). Next, to determine if p53 was required for this effect, we generated p53-knockout (KO) A549 and H460 cell lines by CRISPR-Cas9 targeting. The KO cells did not express p53 and could not upregulate p21 after irradiation (Fig. 2D). Clonogenic survival assays showed that the p53 KO cells were not radiosensitized by trametinib (Fig. 2E) and did not undergo senescence (Supplementary Fig. S1C). We also examined the H23 cell line harboring natural KRAS-p53-LKB1 triple mutation and found it could not be radiosensitized by trametinib either (Supplementary Fig. S1D). Next, we measured levels of MDM2 in A549 and H460 cells under various treatment conditions. As expected (25), MDM2 expression was upregulated by radiotherapy. Trametinib did not affect MDM2 expression but robustly inhibited MDM2 phosphorylation at Ser-166, which has previously been shown to increase the ubiquitin ligase activity of MDM2 (26). Both the activity and levels of p53 were acutely induced by radiotherapy, as indicated by the levels of p53 phosphorylation at Ser-15 and total p53 levels. As a result of these converging pathways on p53 levels and activity, p21 was strongly activated only by the combination of radiotherapy and trametinib (Fig. 2F). Together, the results suggest that trametinib inhibited phosphorylation of MDM2 to stabilize p53 after radiotherapy and induced p53-dependent senescence in KRAS-LKB1–mutant NSCLC cells.

To better understand the functions of comutated TP53 and LKB1 in KRAS-mutant lung cancer, we used a Kras-MUT, p53-WT, and Lkb1-WT mouse lung cancer cell line, LKR13, to generate p53 and Lkb1 KO cells (Supplementary Fig. S2). We first tested the response of the KP and KL cells to radiotherapy and trametinib separately. The p53 and Lkb1 KO in LKR13 cells induced radioresistance (Fig. 3A). Interestingly, however, KO of Lkb1 also increased resistance to trametinib (Fig. 3B). Next, in testing the responses to trametinib-induced radiosensitization, we found that trametinib radiosensitized the Lkb1 KO cells in a dose-dependent manner, but not the scrambled control or the p53 KO cells (Fig. 3C). To further explore the role of LKB1 in trametinib-induced radiosensitization, we overexpressed LKB1 in A549 and H460 cells (both KL). Clonogenic survival assays showed that the A549-LKB1 and H460-LKB1 cell lines were resistant to trametinib radiosensitization (Fig. 3D and E). Overexpression of LKB1 in A549 cells also reduced senescence after trametinib plus radiotherapy (Fig. 3F; Supplementary Fig. S3A). This is also observed in the LKR13 cells, as the combination of trametinib and radiotherapy induced senescence only in Lkb1-KO cells, but not in the scrambled control (Supplementary Fig. S3B and S3C). Thus, LKB1 deletion seems to be required for the trametinib-induced radiosensitization.

We next investigated the mechanism of LKB1-induced resistance to sensitization by trametinib. One of the pathways directly regulated by LKB1 is AMPK (27), which is known to regulate senescence (28–31). Western blotting of LKB1-overexpressing A549 and H460 cells and LKR13 p53- or LKB1-KO cells at 24 hours after combination treatment showed that AMPK was robustly activated in the cells with WT LKB1 (Fig. 4A and B; Supplementary Fig. S4) and was activated to a lesser extent by either treatment alone. AMPK activation was completely absent in the A549 and LKR13 cells without LKB1, suggesting that the activation of AMPK by radiotherapy and trametinib could be LKB1-dependent. To further investigate the roles of AMPK in trametinib-induced radiosensitization, we treated A549 cells with an AMPK activator (AICAR) simultaneously with trametinib and radiotherapy. AICAR can lead to the direct activation of AMPK in A549 cells despite the loss of LKB1 (Fig. 4C). In clonogenic survival assays, activation of AMPK abrogated trametinib-induced radiosensitization (Fig. 4D). Next, we knocked down AMPK in LKR13 cells using siRNAs (Fig. 4E) and found that the LKR13-AMPK-KD cells were radiosensitized by trametinib (DER_0.5 = 1.38 vs. DER_0.5 = 0.95 in scrambled siRNA control) despite their also having WT LKB1 (Fig. 4F), suggesting that inactivation of LKB1-mediated AMPK pathway is required for the trametinib-induced radiosensitization.

AMPK is known to regulate autophagy (32, 33), a self-degradation program in starving and damaged cells (34) that may protect damaged cells from senescence (35–37). We next examined if AMPK activation resulted in upregulation of autophagy. Autophagy activity was indeed increased in LKB1-overexpressing A549 and H460 cells, as evidenced by the increased LC3B II/I ratio (Fig. 5A and B; Supplementary Fig. S5A). In LKR13-Lkb1-KO cells, the level of LC3B-II was much lower compared with the scrambled control after treatment of trametinib alone or trametinib and radiotherapy. Another marker of autophagy activation, phospho-ULK1-Ser555, which could be directly phosphorylated by AMPK (38), was strongly upregulated by trametinib and further intensified by radiotherapy in the LKR13-scrambled control cells. However, in the Lkb1-KO cells, phospho-ULK1-Ser555 was lower compared with scrambled control at baseline level and was not upregulated by treatment, most likely as the consequence of AMPK inactivation in the absence of LKB1 (Supplementary Fig. S5B). Because autophagy may rescue cells from senescence, we next tested if inhibition of autophagy with HCQ could increase senescence in LKB1 WT cells and found that A549-LKB1 cells underwent senescence after being treated with the combination of HCQ, trametinib, and radiotherapy (Fig. 5C). Furthermore, the A549-LKB1 and LKR13 cells were radiosensitized by HCQ + trametinib but not by trametinib alone (Fig. 5D; Supplementary Fig. S5C). Knockdown of ATG5, which is a critical mediator of autophagy (39, 40), in A549-LKB1 cells (Fig. 5E) converts its resistant phenotype in trametinib-induced radiosensitization (Fig. 5F) and increased senescence after the combined treatment of trametinib and radiotherapy (Supplementary Fig. S5D). Because autophagy is capable of reducing reactive oxygen species (ROS) in stressed cancer cells (41), we next examined if autophagy deficiency in Lkb1-KO cells results in elevated ROS levels. Indeed, the LKR13-Lkb1-KO cells had higher baseline ROS levels, but trametinib treatment did not increase ROS generation in both KO and scrambled control cells (Supplementary Fig. S5E). In sum, our results showed that autophagy induced by the LKB1-AMPK activation may induce the resistance to trametinib and radiotherapy combination treatment.

To validate our findings, we established s.c. tumors by implantation of the LKR13-Lkb1 KO cells or scrambled controls in mice and treated the mice according to the schedule depicted in Fig. 6A. Daily 2 Gy fractionated radiotherapy in combination with paclitaxel and carboplatin was used to reflect the clinical standard of care for lung cancer. To test whether the radiosensitization effect of trametinib was still valid after chemoradiation, we used an MTS assay and found that KO of p53 or Lkb1 did not affect cell sensitivity to carboplatin and paclitaxel doublet chemotherapy (Supplementary Fig. S6A). Clonogenic survival assays showed that trametinib sensitized the LKR13-Lkb1 KO cells to both radiotherapy and chemoradiation (Supplementary Fig. S6B). Tumors formed from implanted Lkb1 KO cells were much more aggressive than those from the scrambled control (Supplementary Fig. S7A), despite their having similar growth rates in vitro (Supplementary Fig. S7B). For the LKR13-Lkb1-KO tumors, neither trametinib nor chemoradiation inhibited tumor growth when given separately, but the combination of trametinib and chemoradiation led to significantly better tumor control (Fig. 6B). For both the control and chemoradiation-treated animals, we observed that the animals had significant morbidity and died rapidly. H&E staining of lung tissue collected at day 7 revealed numerous metastatic nodules in the control group and the chemoradiation group, which were significantly reduced by trametinib, and was further reduced by chemoradiation (P = 0.04; Fig. 6D and E). Although trametinib treatment reduced lung metastasis and prolonged survival compared with the controls (P = 0.026), those mice had to be sacrificed before the end of the experiment because of the excessive primary tumor burden. The combination of chemoradiation and trametinib resulted in tumor stabilization and long-term survival (P < 0.0001; Fig. 6B and C; Supplementary Fig. S8). Tumors collected on day 7 showed significant upregulation of p21 levels in the chemoradiation and trametinib group that were not seen in the other treatment groups (Fig. 6F). For the LKR13-scrambled tumors, both trametinib alone and chemoradiation alone significantly inhibited tumor growth relative to control. The combination of chemoradiation and trametinib appears better than either treatment alone, but this difference was not statistically significant (P = 0.060). The autophagy inhibitor HCQ alone had no effect on tumor growth, but the combination of HCQ, trametinib, and chemoradiation produced better tumor growth inhibition than trametinib and chemoradiation or either treatment alone (P < 0.001; Fig. 6G). Together, the results showed that the combination of trametinib and chemoradiation was more effective than either treatment alone in the Kras-Lkb1–mutant tumors. Our data also suggested that inhibition of autophagy may overcome the LKB1-induced resistance in the LKB1-wild-type tumors.

In the current study, we found that the KRAS-LKB1–mutant lung cancer is very sensitive to the combination of trametinib and radiotherapy. The synergy of trametinib and radiotherapy was not seen in KRAS-p53 or KRAS-mutant–alone lung cancer cells. Our data also suggested that the wild-type tumor-suppressor LKB1 could induce resistance to the combination therapy. Specifically, in KRAS-mutant NSCLC, LKB1 induced resistance to this combination through activating the AMPK-autophagy pathway to rescue cells from undergoing senescence. As a result, tumors with both KRAS and LKB1 mutations were very sensitive to this combination because of the absence of LKB1-driven resistance. The combination of trametinib and radiotherapy induced senescence in cells with KRAS and LKB1 mutations, but not in cells with loss-of-function p53 or with WT LKB1. These intriguing findings could reflect the complicated interplay between the MAPK pathway and tumor-suppressor pathways. Our findings suggest that trametinib, when given with radiotherapy, simultaneously affects two tumor-suppressor pathways: it stabilizes p53 by reducing MDM2-Ser166 phosphorylation and by activating AMPK depending on LKB1. The p53 pathway is well known to be activated by radiotherapy (42), but the level of activation is strictly regulated by the negative feedback loop of MDM2. MDM2 phosphorylation at Ser166 reportedly increases the ubiquitin ligase activity of MDM2 (26, 43); thus, trametinib could stabilize the p53 protein after radiotherapy to augment the radiotherapy-induced senescent effects. This radiosensitization mechanism requires functional p53 signaling, suggesting the tumors with KRAS-LKB1-p53 triple mutation may not be radiosensitized by trametinib. In the LKB1 WT tumors, however, trametinib also acted synergistically with radiotherapy to induce robust AMPK activation and eventually upregulated autophagy to rescue cells from senescence. Autophagy is one mechanism that prevents cells from undergoing senescence (35–37). In a recent study of breast cancer, the CDK 4/6 inhibitor palbociclib induced both autophagy and senescence, but combining palbociclib with the autophagy inhibitor HCQ led to inhibited tumor cell growth and increased senescence both in vitro and in vivo (44). In summary, although LKB1 mutation in our study led to resistance to radiotherapy or trametinib given separately, it rendered the KL subtype of NSCLC vulnerable to the combination of these two treatments because of the absence of LKB1-AMPK autophagy–driven resistance. Our proposed working model for these effects is illustrated in Supplementary Fig. S9.

It has been previously shown that murine Kras-Lkb1–mutant lung tumors were more resistant to radiotherapy compared with Kras-p53 or Kras-mutant tumors (14). Our results also demonstrated that Lkb1 KO in LKR13 cell lines conferred resistance to radiotherapy in vitro and chemoradiation in vivo. However, the mechanism by which LKB1 deficiency induces radioresistance is still unknown. The overexpression of LKB1 in A549 and H460 increased radiosensitivity in those cells through the downregulation of DNA damage repair–related genes, such as RAD50 and XRCC1 (45), suggesting LKB1 may be involved in DNA damage repair. Interestingly, MEK inhibition has been reported to attenuate DNA damage repair in lung cancer cells after radiotherapy (46, 47). Whether LKB1 and MAPK pathway interconnections regulate DNA repair needs future investigation.

Certain limitations exist in this study. First, the subcutaneous tumor model may not reflect the microenvironment in the lung. However, attaining intraparenchymal single orthotopic Kras-Lkb1 tumor model using the LKR13 cell line for lung radiotherapy experiments was not feasible, due to the aggressive nature of such double mutant tumors, which led to rapid dissemination. Second, although CRISPR KO of p53 and Lkb1 in the LKR13 cells provided a simple model system for us to assess the relative functions of these two tumor suppressors in the controlled genetic context, they still may not represent their functions in tumor initiation and evolution in vivo. Validation studies using cell lines isolated from genetically modified mouse models are needed in the future. Despite these limitations, we believe our present study has unveiled an interesting complexity of tumor signaling which is dependent on the genetic background, especially when multiple treatment modalities are being considered. Usually, deficiencies in tumor suppressors are linked with tumor initiation and progression. Activation of the tumor suppressor, LKB1, and its downstream AMPK pathway is being increasingly recognized as a novel cancer therapeutic strategy. The AMPK activator metformin has long been used for the treatment of diabetes, but it also has antitumor effects (48–50). However, our study suggests that the tumor-suppressive pathways could also be activated by certain combinations to induce therapy resistance. In the KRAS-LKB1–mutant tumors, although the genetic aberrations could portend to greater resistance to individual conventional therapies, synthetic lethality was conferred when the various therapies are combined. This implies that identifying proper combination of therapies in specific genetic contexts will enable improvements in therapy outcomes. More importantly, our in vivo models also showed that the current standard of care, which is chemoradiation, could be potentially enhanced by trametinib in KRAS-LKB1–mutant tumors. Whether our findings are translatable to the clinic will require prospective validation on future clinical trials.

S.H. Lin is a consultant/advisory board member for AstraZeneca, Inc., and reports receiving commercial research grants from BeyondSpring Pharmaceuticals, Inc., Genentech, Hitachi Chemical Diagnostic. No potential conflicts of interest were disclosed by the other authors.

Conception and design: Y. Wang, N. Li, M. Zhang, M.-C. Hung, S.H. Lin

Development of methodology: Y. Wang, N. Li, C. Xu, S.H. Lin

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): Y. Wang, N. Li, R. Ye, Y. Qiao, A. Sharma, S.H. Lin

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): Y. Wang, W. Jiang, W. Deng, C. Xu, M.-C. Hung, S.H. Lin

Writing, review, and/or revision of the manuscript: Y. Wang, W. Jiang, C. Xu, A. Sharma, S.H. Lin

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): S.H. Lin

Study supervision: M.-C. Hung, S.H. Lin

Other (funding): S.H. Lin

This work was supported in part by the Mabuchi Program in Targeted Radiotherapy, Uniting Against Lung Cancer, NIH/NCI Lung Cancer SPORE (5P50CA070907), and Cancer Center Support (Core) Grant CA016672 from the NIH/NCI, to The University of Texas MD Anderson Cancer Center. Y. Wang is supported by a T.C. Hsu Memorial Scholarship from The University of Texas MD Anderson Cancer Center UTHealth Graduate School of Biomedical Sciences. The authors thank Christine Wogan of the Division of Radiation Oncology at MD Anderson for editorial assistance. The authors thank Jie Zhang of the Center for Radiation Oncology Research at MD Anderson for technical assistance in histology.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.