KEAP1 loss modulates sensitivity to kinase targeted therapy in lung cancer

Elsa B Krall; Belinda Wang; Diana M Munoz; Nina Ilic; Srivatsan Raghavan; Matthew J Niederst; Kristine Yu; David A Ruddy; Andrew J Aguirre; Jong Wook Kim; Amanda J Redig; Justin F Gainor; Juliet A Williams; John M Asara; John G Doench; Pasi A Janne; Alice T Shaw; Robert E McDonald III; Jeffrey A Engelman; Frank Stegmeier; Michael R Schlabach; William C Hahn

doi:10.7554/eLife.18970

. 2017 Feb 1;6:e18970. doi: 10.7554/eLife.18970

KEAP1 loss modulates sensitivity to kinase targeted therapy in lung cancer

Elsa B Krall ^1,^2,^3,^†,^‡, Belinda Wang ^1,^2,^3,^†, Diana M Munoz ^4,^†, Nina Ilic ^1,^2,³, Srivatsan Raghavan ^1,^2,³, Matthew J Niederst ⁴, Kristine Yu ⁴, David A Ruddy ⁴, Andrew J Aguirre ^1,^2,³, Jong Wook Kim ^1,^2,³, Amanda J Redig ^1,³, Justin F Gainor ⁵, Juliet A Williams ⁴, John M Asara ^6,⁷, John G Doench ², Pasi A Janne ^1,³, Alice T Shaw ⁵, Robert E McDonald III ⁴, Jeffrey A Engelman ⁴, Frank Stegmeier ^4,^‡, Michael R Schlabach ^4,^‡, William C Hahn ^1,^2,^3,^*

Editor: Martin McMahon⁸

PMCID: PMC5305212 PMID: 28145866

Abstract

Inhibitors that target the receptor tyrosine kinase (RTK)/Ras/mitogen-activated protein kinase (MAPK) pathway have led to clinical responses in lung and other cancers, but some patients fail to respond and in those that do resistance inevitably occurs (Balak et al., 2006; Kosaka et al., 2006; Rudin et al., 2013; Wagle et al., 2011). To understand intrinsic and acquired resistance to inhibition of MAPK signaling, we performed CRISPR-Cas9 gene deletion screens in the setting of BRAF, MEK, EGFR, and ALK inhibition. Loss of KEAP1, a negative regulator of NFE2L2/NRF2, modulated the response to BRAF, MEK, EGFR, and ALK inhibition in BRAF-, NRAS-, KRAS-, EGFR-, and ALK-mutant lung cancer cells. Treatment with inhibitors targeting the RTK/MAPK pathway increased reactive oxygen species (ROS) in cells with intact KEAP1, and loss of KEAP1 abrogated this increase. In addition, loss of KEAP1 altered cell metabolism to allow cells to proliferate in the absence of MAPK signaling. These observations suggest that alterations in the KEAP1/NRF2 pathway may promote survival in the presence of multiple inhibitors targeting the RTK/Ras/MAPK pathway.

DOI: http://dx.doi.org/10.7554/eLife.18970.001

Research Organism: Human

Introduction

The receptor tyrosine kinase (RTK)/ mitogen-activated protein kinase (MAPK) pathway plays an important role in the development of lung and other cancers, with the frequent occurrence of mutations or copy number alterations in multiple nodes of this pathway (Ding et al., 2008; Imielinski et al., 2012). However, single-agent therapy targeting this pathway has had limited clinical success. While BRAF and EGFR inhibitors can produce dramatic responses temporarily, acquired resistance inevitably occurs in lung and other cancers (Balak et al., 2006; Kosaka et al., 2006; Rudin et al., 2013; Wagle et al., 2011). In addition to this acquired resistance, many tumors also exhibit intrinsic resistance to these inhibitors (Corcoran et al., 2012; Prahallad et al., 2012), as well as to MEK inhibitors (Sun et al., 2014).

Several studies have now shown that a general theme of resistance to these targeted therapies is activation of the RTK/MAPK pathway by alternative mechanisms. For example feedback activation of EGFR has been shown to cause intrinsic resistance to BRAF inhibition in colon cancer (Corcoran et al., 2012; Prahallad et al., 2012), and resistance to BRAF inhibition in melanoma can be caused by reactivation of the MAPK pathway by RTKs, NRAS, or COT (Johannessen et al., 2010; Nazarian et al., 2010). In lung cancer, transcriptional induction of ERBB3 causes intrinsic resistance to MEK inhibition in KRAS-mutant cancers (Sun et al., 2014), and acquired resistance to EGFR inhibitors was found to result from amplification of MET (Engelman et al., 2007). These findings highlight the importance of maintaining RTK/MAPK signaling in lung and other cancers and also suggest redundancy among different genetic alterations in this pathway. Due to the many ways that cancers can acquire resistance to single therapies targeting the RTK/MAPK pathway, combination therapy may hold more promise for treating tumors with alterations in this pathway.

While patients with EGFR- and ALK-mutant lung cancer do often respond to EGFR and ALK inhibitors, these responses are temporary, as acquired resistance inevitably develops (Tang et al., 2013; Wilson et al., 2015). Likewise, in on-going clinical trials testing MEK and BRAF inhibitors in KRAS- and BRAF-mutant lung cancer, responses have been reported in some patients (Blumenschein et al., 2015; Hyman et al., 2015). However, it is clear that both intrinsic and acquired resistance limits efficacy of these inhibitors. To prospectively identify mechanisms of resistance to EGFR, ALK, BRAF, and MEK inhibition in lung cancer, we performed CRISPR-Cas9 drug resistance screens in five cancer cell lines with different alterations in the RTK/Ras/MAPK pathway. We identified a number of genes whose deletion confers cell survival in these contexts, and focused on KEAP1, as loss of this gene modulated the response to targeted therapies in multiple contexts.

Results

CRISPR-Cas9 screens identify genes whose loss confers resistance to MEK and BRAF inhibition

To identify genes that modulate the response to RTK/Ras/MAPK inhibition in lung cancer, we performed two sets of CRISPR-Cas9 knockout screens. In the first set (Figure 1A), we performed three genome scale screens with the MEK inhibitor trametinib, in the NRAS-mutant lung cancer cell line H1299 (NRAS^Q61K), the BRAF-mutant lung cancer cell line HCC364 (BRAF^V600E), and the KRAS-mutant lung cancer cell line CALU1 (KRAS^G12C). One additional screen was performed in HCC364 cells treated with the BRAF inhibitor vemurafenib. To perform genome scale screens, we introduced the GeCKO v2 library (Shalem et al., 2014) into Cas9-expressing cells, selected cells that incorporated the sgRNAs and allowed genome editing to occur over one week. Cells were then harvested for the Day 0 time point or passaged in the presence of trametinib or vemurafenib (Figure 1A). We used the lowest concentration of drug that inhibited ERK phosphorylation and resulted in proliferative arrest or death (Figure 1—figure supplement 1). Genomic DNA was isolated on Days 14 and 21, and sgRNAs were quantified by sequencing. sgRNAs that were enriched in the Day 14 and Day 21 samples compared to the Day 0 samples were then identified, using two methods: (1) a cutoff of log2 fold change of at least 2, or (2) a STARS score (Doench et al., 2016) of at least 5 (Supplementary file 1). Several of the genes that scored in these screens also scored in a previous vemurafenib resistance screen in BRAF-mutant melanoma (Shalem et al., 2014).

Figure 1. — (A,C) Pathway schematics and screening timelines. (B) Heatmap showing the STARS score for each gene with a score of at least five in at least one screen. (D–E) Crystal violet colony formation assays. (D) 5000 CALU1 cells were treated with 50 nM trametinib for 17 days. 2000 HCC364 cells were treated with 25 nM trametinib or 6.25 uM vemurafenib for 21 days. (E) 10,000 HCC827 cells were treated with 100 nM erlotinib for 10 days. 10,000 MGH065 cells were treated with 800 nM LDK378 or 800 nM crizotinib for 10 days. Error bars represent the SD of the mean of triplicate wells. Experiments were performed two independent times and one representative experiment is shown. ***p=0.0001, **p<0.005.

**DOI:** http://dx.doi.org/10.7554/eLife.18970.002

Figure 1—figure supplement 1. — (A,C) Pathway schematics and screening timelines. (B) Heatmap showing the STARS score for each gene with a score of at least five in at least one screen. (D–E) Crystal violet colony formation assays. (D) 5000 CALU1 cells were treated with 50 nM trametinib for 17 days. 2000 HCC364 cells were treated with 25 nM trametinib or 6.25 uM vemurafenib for 21 days. (E) 10,000 HCC827 cells were treated with 100 nM erlotinib for 10 days. 10,000 MGH065 cells were treated with 800 nM LDK378 or 800 nM crizotinib for 10 days. Error bars represent the SD of the mean of triplicate wells. Experiments were performed two independent times and one representative experiment is shown. ***p=0.0001, **p<0.005.

**DOI:** http://dx.doi.org/10.7554/eLife.18970.002

We annotated the functions of each of the genes that scored in these screens to determine if particular functional categories scored repeatedly (Figure 1B). As expected, several genes in the MAPK pathway scored, including NF1, a negative regulator of Ras/MAPK signaling, and DUSP1, a dual-specificity phosphatase that inhibits ERK. We also found several positive regulators of p38/JNK MAPK signaling, suggesting that these other MAPK pathways may play a pro-apoptotic or anti-proliferative role in these cells. PTEN, a negative regulator of PI3K/AKT signaling, and TSC1 and TSC2, negative regulators of mTOR signaling, also scored, suggesting that increased signaling through the PI3K/AKT/mTOR pathway compensates for loss of Ras/MAPK signaling. In addition to these expected pathways, several of the genes that scored are components of histone acetyltransferase (HAT) complexes or of the Mediator complex. There were also several genes whose products are components of E3 ubiquitin ligase complexes. Multiple transcription factors scored, as well as general transcription machinery genes. Other functional categories for which multiple genes scored include Rho signaling and histidine post-translational modifications. We noted that KEAP1, a substrate adaptor protein that targets NFE2L2/NRF2 for ubiquitination and proteasomal degradation, was the only gene that scored highly in all four genome scale screens (Figure 1B and Supplementary file 1).

The second set of screens (Figure 1C) were run independently using a more focused sgRNA library comprised of druggable targets, including kinases, proteases, phosphatases, and ligases (Munoz et al., 2016). This library was used to screen HCC827 cells (EGFR^Δ746-750) treated with the EGFR inhibitor erlotinib, and the patient-derived ALK-mutant line MGH-065 treated with the ALK inhibitors crizotinib or LDK-378 (Figure 1—figure supplement 2). Similar to the genome scale screens described above, Cas9-expressing cells were infected with the sgRNA library, selected for 4 days and cultured for an additional 10 days in the presence of DMSO or drug. SgRNAs that were enriched in the Day 10 drug-treated samples compared to the DMSO-treated samples were then identified. Notably, loss of KEAP1 also promoted survival in the presence of EGFR and ALK inhibitors (Figure 1—figure supplement 3 and Supplementary file 2). Thus, loss of KEAP1 was found to modulate the response to multiple targeted therapeutics in different genetic contexts using two independent sgRNA libraries.

KEAP1^KO confers resistance through increased NRF2 activity

To validate that KEAP1 loss modulates sensitivity to multiple inhibitors, we infected HCC364 (BRAF^V600E), CALU1 (KRAS^G12C), HCC827 (EGFR ^Δ746-750), and MGH-065 (EML4-ALK) cells with sgRNAs targeting KEAP1 or GFP (Figure 1—figure supplement 4). Cells were then seeded at low density in 12- or 24-well plates and treated with drug. Cell viability was assessed by crystal violet staining. Deletion of KEAP1 (KEAP1^KO) decreased sensitivity to each of the inhibitors that were tested (Figure 1D–E). In addition to the lines used in the CRISPR-Cas9 screens, we also tested the ability of KEAP1 loss to modulate sensitivity to EGFR or ALK inhibition in two additional cell lines. KEAP1^KO decreased sensitivity to afatinib treatment in NCI-H1975 (EGFR^L858R/T790M) cells and to crizotinib treatment in NCI-H3122 (EML4-ALK) cells (Figure 1—figure supplement 5). In short-term dose response experiments, loss of KEAP1 also caused a shift in the IC50 of erlotinib in HCC827 cells (Figure 1—figure supplement 6).

Unlike many other reported mechanisms of resistance to inhibitors of the RTK/Ras/MAPK pathway (Rudin et al., 2013; Wagle et al., 2011; Corcoran et al., 2012; Prahallad et al., 2012; Sun et al., 2014; Johannessen et al., 2010; Nazarian et al., 2010), we found that KEAP1^KO did not restore ERK activation (Figure 2A and Figure 2—figure supplement 1A), indicating that KEAP1^KO does not decrease sensitivity by reactivating the MAPK pathway. Loss of KEAP1 also did not affect expression of the pro-apoptotic protein BIM (Figure 2—figure supplement 1B). We concluded that loss of KEAP1 did not re-activate MAPK signaling or generally promote survival by inhibiting apoptosis.

Figure 2. — (A) KEAP1-KO does not affect p-ERK. Whole cell lysates of HCC364-Cas9 cells with the indicated sgRNAs treated with DMSO or trametinib for 48 hr. (B) KEAP1-KO increases NRF2 levels. Nuclear and cytoplasmic fractions of HCC364 cells. (C) Immunoblot showing NRF2 expression in CALU1 and HCC364. (D) Crystal violet colony formation assays. 10,000 CALU1 cells expressing the indicated ORFs were treated with DMSO for 8 days or trametinib for 10 Days. 10,000 HCC364 cells expressing the indicated ORFs were treated with DMSO for 10 days or trametinib/vemurafenib for 21 days. Error bars represent the SD of the mean of triplicate wells. (E) Immunoblot showing expression of wildtype KEAP1 or KEAP1 G333C in A549 cells. (F) Expression of wildtype KEAP1 resensitized A549 cells to trametinib. 5000 cells were treated with 25 nM trametinib for 12 days. Error bars represent the SD of six wells. (A–F) Experiments were performed two independent times, and one representative experiment is shown. ***p=0.0001, **p=0.0005, *p<0.002.

**DOI:** http://dx.doi.org/10.7554/eLife.18970.009

Figure 2—figure supplement 1. — (A) KEAP1-KO does not affect p-ERK. Whole cell lysates of HCC364-Cas9 cells with the indicated sgRNAs treated with DMSO or trametinib for 48 hr. (B) KEAP1-KO increases NRF2 levels. Nuclear and cytoplasmic fractions of HCC364 cells. (C) Immunoblot showing NRF2 expression in CALU1 and HCC364. (D) Crystal violet colony formation assays. 10,000 CALU1 cells expressing the indicated ORFs were treated with DMSO for 8 days or trametinib for 10 Days. 10,000 HCC364 cells expressing the indicated ORFs were treated with DMSO for 10 days or trametinib/vemurafenib for 21 days. Error bars represent the SD of the mean of triplicate wells. (E) Immunoblot showing expression of wildtype KEAP1 or KEAP1 G333C in A549 cells. (F) Expression of wildtype KEAP1 resensitized A549 cells to trametinib. 5000 cells were treated with 25 nM trametinib for 12 days. Error bars represent the SD of six wells. (A–F) Experiments were performed two independent times, and one representative experiment is shown. ***p=0.0001, **p=0.0005, *p<0.002.

**DOI:** http://dx.doi.org/10.7554/eLife.18970.009

KEAP1 serves as a substrate adaptor protein that recruits the CUL3 ubiquitin ligase to NRF2, targeting it for proteasomal degradation (Kobayashi et al., 2004). As expected, we found that KEAP1^KO led to increased NRF2 protein levels (Figure 2B and Figure 2—figure supplement 1C). To determine if increased NRF2 levels were necessary for KEAP1 loss to modulate drug sensitivity, we knocked out both KEAP1 and NRF2 in HCC827 cells. KEAP1 loss no longer promoted survival when NRF2 was also lost (Figure 2—figure supplement 2A). To determine if increased NRF2 levels were sufficient to promote survival, we overexpressed wildtype NRF2 or two different NRF2 mutants (G31R or D29H), which contain mutations in the KEAP1 binding domain. Overexpression of wildtype NRF2 led to an increase in cell survival in the presence of multiple drugs. Overexpression of NRF2^G31R or NRF2^D29H resulted in higher NRF2 protein levels and increased resistance to trametinib, vemurafenib, and erlotinib (Figure 2D and Figure 2—figure supplement 2B), suggesting that elevated NRF2 levels in KEAP1^KO cells mediates resistance. Although CALU1 cells harbor a KEAP1^P128L mutation, this mutation has not been reported in the cBioPortal or COSMIC (Cerami et al., 2012; Gao et al., 2013; Forbes et al., 2015), and NRF2 levels increased upon KEAP1 knockout (Figure 2—figure supplement 1), suggesting that the regulation of NRF2 by KEAP1 is intact in these cells. We also found that restoring wildtype KEAP1 expression in A549 cells, which are KRAS-mutant and KEAP1-null, increased their sensitivity to trametinib. In contrast, expression of the KEAP1^G333C mutant, which does not regulate NRF2, failed to alter trametinib sensitivity (Figure 2E,F). Together these observations suggest that increased NRF2 levels upon loss of KEAP1 modulates sensitivity to therapies targeting the RTK/MAPK pathway.

Loss of KEAP1 has previously been reported to confer resistance to several chemotherapeutics (Ohta et al., 2008; Shibata et al., 2008a; Wang et al., 2008; Zhang et al., 2010). However, it is also clear that KEAP1/NRF2 and the MAPK pathway are mechanistically linked (DeNicola et al., 2011; Sun et al., 2009). To further explore the mechanism by which KEAP1^KO promotes survival in the presence of RTK/MAPK inhibitors, we investigated whether drug treatment affected KEAP1/NRF2 signaling. A prior report demonstrated that Ras/MAPK/Jun signaling increased expression of NRF2 mRNA and NRF2 target genes (DeNicola et al., 2011), so we hypothesized that drug treatment would decrease expression of NRF2 mRNA and NRF2-regulated target genes. Surprisingly, we found that trametinib, erlotinib, and LDK-378 treatment increased rather than decreased expression of NRF2 mRNA and NRF2 target genes (Figure 3A and Figure 3—figure supplement 1) in HCC364, CALU1, MGH-065, and HCC827 cells. As expected, KEAP1^KO also increased NRF2 target gene expression. Trametinib treatment also increased NRF2 protein levels and induced a shift in the migration of NRF2 protein on SDS-PAGE, whereas KEAP1^KO maintained the expression of the higher molecular weight form of NRF2 (Figure 3B and Figure 3—figure supplement 2). Although trametinib treatment decreased KEAP1 expression and increased NRF2 expression, the combination of KEAP1 knockout with trametinib treatment caused a greater effect on KEAP1 levels, NRF2 levels, and NRF2 target gene expression. These observations suggest that treatment with these inhibitors partially activates NRF2, and that loss of KEAP1 leads to further NRF2 activation, which decreases drug sensitivity.

Figure 3. — (A) Expression of NFE2L2/NRF2 mRNA and NRF2 target genes in HCC364 treated with DMSO or trametinib for 72 hr. Error bars represent the SD of the mean of three biological replicates. (B) HCC364 cells treated with DMSO or trametinib for 72 hr. TRAM, trametinib. (A,B) Experiments were performed two independent times, and one representative experiment is shown.

**DOI:** http://dx.doi.org/10.7554/eLife.18970.012

Figure 3—figure supplement 1. — (A) Expression of NFE2L2/NRF2 mRNA and NRF2 target genes in HCC364 treated with DMSO or trametinib for 72 hr. Error bars represent the SD of the mean of three biological replicates. (B) HCC364 cells treated with DMSO or trametinib for 72 hr. TRAM, trametinib. (A,B) Experiments were performed two independent times, and one representative experiment is shown.

**DOI:** http://dx.doi.org/10.7554/eLife.18970.012

Loss of KEAP1 promotes survival by increasing glutathione synthesis and decreasing drug-induced ROS

The KEAP1/NRF2 axis responds to oxidative and electrophilic stress by regulating expression of drug efflux pumps, by scavenging reactive oxygen species (ROS), and by altering cell metabolism (Hayes and Dinkova-Kostova, 2014). We investigated whether each of these functions was involved in modulating sensitivity to RTK/MAPK inhibition. Since MAPK pathway inhibition is maintained in KEAP1^KO cells (Figure 2A), drug efflux likely does not explain resistance. To determine if loss of KEAP1 promotes survival through modulation of ROS, we treated control or KEAP1^KO cells with DMSO or drug and measured ROS. We found that treatment with trametinib, erlotinib, or LDK-378 induced ROS in KEAP1-intact cells and that ROS was dramatically decreased in KEAP1^KO cells or NRF2 overexpressing cells (Figure 4A and Figure 4—figure supplement 1), suggesting that KEAP1^KO may promote survival by reducing ROS levels. The ability of RTK/MAPK inhibitors to induce ROS in control cells may explain their ability to increase NRF2 levels, as increased ROS would lead to release of NRF2 from KEAP1 regulation (Kobayashi et al., 2004). Further activation of NRF2 through deletion of KEAP1 may then reduce ROS levels, leading to increased survival in the presence of drug.

Figure 4. — (A) HCC364 or CALU1 cells were treated with DMSO or trametinib for 72 hr. ROS was measured by DCFDA fluorescence. Error bars represent SD of the mean of two biological replicates. (B, D) CALU1 or HCC364 cells were treated with DMSO or trametinib for 72 hr and metabolites were measured by liquid chromatography tandem mass spectrometry. Error bars represent SD of eight replicates. (C) Glutathione synthesis pathway. (E) Expression of GPX1 and GSR in HCC364 and CALU1 cells treated with DMSO or trametinib for 72 hr. Error bars represent the SD of the mean of three biological replicates. ***p<0.005, **p<0.01, *p<0.02.

**DOI:** http://dx.doi.org/10.7554/eLife.18970.015

Figure 4—figure supplement 1. — (A) HCC364 or CALU1 cells were treated with DMSO or trametinib for 72 hr. ROS was measured by DCFDA fluorescence. Error bars represent SD of the mean of two biological replicates. (B, D) CALU1 or HCC364 cells were treated with DMSO or trametinib for 72 hr and metabolites were measured by liquid chromatography tandem mass spectrometry. Error bars represent SD of eight replicates. (C) Glutathione synthesis pathway. (E) Expression of GPX1 and GSR in HCC364 and CALU1 cells treated with DMSO or trametinib for 72 hr. Error bars represent the SD of the mean of three biological replicates. ***p<0.005, **p<0.01, *p<0.02.

**DOI:** http://dx.doi.org/10.7554/eLife.18970.015

To better understand how KEAP1^KO reduced ROS, we performed metabolite profiling in control or KEAP1^KO CALU1 and HCC364 cells treated with DMSO or trametinib for 72 hr. In both cell lines, the most highly increased metabolite in KEAP1^KO cells treated with trametinib compared to control cells treated with trametinib was reduced glutathione (GSH), an antioxidant (Figure 4B and Supplementary file 3). In addition to glutathione, we identified several other changes in the glutathione synthesis pathway (Figure 4C). Both the reduced (GSH) and oxidized (GSSG) form of glutathione were increased in KEAP1^KO cells, although the ratio of reduced to oxidized glutathione (GSH/GSSG) also increased (Figure 4—figure supplement 2), indicating a more reduced intracellular environment. We also found that cystine, the oxidized form of cysteine, was among the most decreased metabolites in KEAP1^KO cells compared to control cells (Figure 4D), suggesting that KEAP1^KO contributes to a change in oxidative state in which cystine is converted to cysteine, which is then used to make glutathione.

In addition to the changes in metabolite levels, we also observed several changes in gene expression that were consistent with increased glutathione synthesis in KEAP1^KO cells. Glutamate cysteine ligase (GCL) catalyzes the rate-limiting step in glutathione synthesis, and both subunits of this enzyme (GCLC and GCLM) are NRF2 target genes. As described above, expression of both of these genes was higher in drug-treated KEAP1^KO cells compared to drug-treated control cells (Figure 3A and Figure 3—figure supplement 1). Glycine is also a precursor to glutathione, and NRF2 has previously been shown to regulate expression of serine hydroxymethyltransferase (SHMT1,2) (DeNicola et al., 2015), which catalyzes the conversion of serine to glycine. We found that expression of SHMT1 and SHMT2 decreased upon trametinib treatment in CALU1 and HCC364 control cells but was partially maintained in KEAP1^KO cells, although this was not observed in MGH-065 and HCC827 cells (Figure 4—figure supplement 3). In addition, we found that expression of glutathione peroxidase (GPX1), which converts GSH to GSSG, increased with drug treatment in control cells, but this increase was abrogated in KEAP1^KO cells (Figure 4E and Figure 4—figure supplement 4). Conversely, expression of glutathione reductase (GSR), which converts GSSG to GSH, was higher in KEAP1^KO cells than in control cells (Figure 4E and Figure 4—figure supplement 4). These observations suggest that the conversion of oxidized to reduced glutathione may be involved in the modulation of drug sensitivity meditated by loss of KEAP1.

To determine whether the increased glutathione levels promoted survival in the presence of drug, we treated KEAP1-intact cells with trametinib and N-acetyl cysteine (NAC), an antioxidant and glutathione precursor. NAC reduced ROS and decreased sensitivity to trametinib (Figure 5A,B), indicating that ROS reduction through increased glutathione synthesis is important for KEAP1^KO-mediated survival. To further investigate whether glutathione synthesis and ROS reduction was important for survival, we treated cells with trametinib and buthionine sulfoximine (BSO), which inhibits GCL. The combination of BSO and trametinib increased ROS and greatly decreased viability in control cells expressing sgGFP, while KEAP1^KO prevented the BSO-induced decrease in viability (Figure 5C,D and Figure 5—figure supplement 1). Furthermore, combined treatment with BSO and trametinib dramatically increased ROS levels in A549 cells in which wildtype KEAP1 expression had been restored, but not in the parental cells or cells expressing KEAP1^G333C (Figure 5—figure supplement 2). These observations indicate that KEAP1^KO can decrease ROS even when GCL is inhibited, suggesting that the conversion of oxidized to reduced glutathione may be more important for promoting survival in the presence of drug. Together, these observations suggest that trametinib treatment induces ROS, which activates NRF2 to low levels. Loss of KEAP1 leads to further activation of NRF2, which decreases drug sensitivity in part by increasing glutathione and decreasing ROS.

Figure 5. — (A, B) CALU1 cells were treated with DMSO or 50 nM trametinib and the indicated concentration of N-acetyl cysteine (NAC) for 16 days. (A) ROS was measured by DCFDA assay. Error bars represent SD of two replicates. (B) Population doublings of trametinib-treated cells compared to DMSO-treated cells. Error bars represent SD of two replicates. (C) KEAP1-KO reduces trametinib- and BSO-induced ROS in CALU1 cells. Cells were treated for 72 hr, and ROS was measured by DCFDA assay. Error bars represent SD of two replicates. (D) KEAP1-KO increases cell viability in CALU1 cells treated with trametinib and BSO. 20,000 CALU1 cells were treated with DMSO plus BSO for 7 days or trametinib plus BSO for 12 days. Error bars represent SD of triplicate wells. (E) Model of how KEAP1-KO confers resistance to trametinib. MAPK pathway inhibitors turn off MAPK signaling and also induce ROS, leading to low level activation of NRF2. Upon KEAP1 loss, NRF2 activity further increases, altering expression of genes involved in the antioxidant response and metabolism, allowing for proliferation in the absence of MAPK signaling.

**DOI:** http://dx.doi.org/10.7554/eLife.18970.020

Figure 5—figure supplement 1. — (A, B) CALU1 cells were treated with DMSO or 50 nM trametinib and the indicated concentration of N-acetyl cysteine (NAC) for 16 days. (A) ROS was measured by DCFDA assay. Error bars represent SD of two replicates. (B) Population doublings of trametinib-treated cells compared to DMSO-treated cells. Error bars represent SD of two replicates. (C) KEAP1-KO reduces trametinib- and BSO-induced ROS in CALU1 cells. Cells were treated for 72 hr, and ROS was measured by DCFDA assay. Error bars represent SD of two replicates. (D) KEAP1-KO increases cell viability in CALU1 cells treated with trametinib and BSO. 20,000 CALU1 cells were treated with DMSO plus BSO for 7 days or trametinib plus BSO for 12 days. Error bars represent SD of triplicate wells. (E) Model of how KEAP1-KO confers resistance to trametinib. MAPK pathway inhibitors turn off MAPK signaling and also induce ROS, leading to low level activation of NRF2. Upon KEAP1 loss, NRF2 activity further increases, altering expression of genes involved in the antioxidant response and metabolism, allowing for proliferation in the absence of MAPK signaling.

**DOI:** http://dx.doi.org/10.7554/eLife.18970.020

In addition to regulating ROS, NRF2 has been reported to regulate the expression of metabolic genes (DeNicola et al., 2015; Mitsuishi et al., 2012). We found that glycolytic and pentose phosphate pathway intermediates were higher in trametinib-treated KEAP1^KO cells compared to trametinib-treated control cells (Supplementary file 3), indicating that KEAP1 knockout partially restores anabolic metabolism. We also found that expression of several NRF2 target genes involved in the pentose phosphate pathway, de novo nucleotide synthesis, and NADPH synthesis was higher in drug-treated KEAP1^KO cells than in drug-treated control cells (Figure 5—figure supplement 3). Together these observations support a model in which RTK/MAPK pathway inhibitors block MAPK signaling and induce ROS, which activates NRF2 to low levels. KEAP1 loss increases NRF2 activity, which reduces ROS through increased glutathione synthesis and alters cell metabolism, allowing cells to proliferate in the absence of MAPK signaling (Figure 5E).

To determine whether KEAP1 loss also conferred resistance in vivo, we performed a xenograft study, in which HCC827 KEAP1^KO or control cells were implanted subcutaneously. When the average tumor volume reached 100–200 mm³, mice were randomized to three groups: vehicle, 5 mg/kg erlotinib, or 12 mg/kg erlotinib. KEAP1^KO cells showed a proliferative advantage over control cells in both the presence and absence of erlotinib (Figure 6).

Figure 6. — Growth of HCC827 sgGFP and sgKEAP1 cells in vivo. HCC827 sgGFP or sgKEAP1 cells were implanted bilaterally and treatment was initiated when average tumor volume reached 100–200 mm3. Data represent mean ± SD (n = 12).

**DOI:** http://dx.doi.org/10.7554/eLife.18970.024

Discussion

Previous genome scale screens to identify mechanisms of resistance to targeted therapeutics have focused on a single genetic alteration (Shalem et al., 2014; Berns et al., 2007; Johannessen et al., 2013; Whittaker et al., 2013). Here we have expanded this approach to identify mechanisms of resistance that are shared across multiple targeted therapeutics in several genetic contexts. We have found that KEAP1 loss modulates sensitivity to inhibition of EGFR, ALK, BRAF, or MEK in lung cancer cell lines with EGFR, ALK, BRAF, KRAS, or NRAS mutations. In multiple independent CRISPR-Cas9 screens, we found that sgRNAs targeting KEAP1 were more highly enriched than sgRNAs targeting NF1, DUSP1, PTEN, or TSC1/2, known negative regulators of the MAPK and PI3K pathways. Previous studies using shRNA screens to identify potential combination therapies for Ras-mutant cancers found that RTK/MAPK pathway reactivation (Manchado et al., 2016) or induction of anti-apoptotic proteins (Corcoran et al., 2013) limits therapeutic efficacy. Importantly, loss of KEAP1 does not lead to reactivation of the MAPK pathway or decreased levels of the pro-apoptotic protein BIM, and instead reduces drug-induced ROS and alters cell metabolism.

KEAP1 loss or NRF2 overexpression is sufficient to restore cell proliferation in the absence of MAPK signaling. It has previously been shown that ROS detoxification and modulation of redox state by NRF2 contributes to tumorigenesis (DeNicola et al., 2011), and NRF2 has recently been found to be an oncogene (Kim et al., 2016). In addition to its role in regulating redox state, NRF2 has also previously been shown to regulate expression of many metabolic enzymes, redirecting glucose and glutamine into anabolic pathways that support proliferation (Mitsuishi et al., 2012) and controlling serine and glycine biosynthesis to support glutathione and nucleotide production (DeNicola et al., 2015). Thus, increased expression of NRF2 upon KEAP1 loss can confer resistance to RTK/MAPK pathway inhibition by reducing ROS and regulating metabolic pathways. Chio et al. recently reported that NRF2 supports pancreatic tumor maintenance, and that combined targeting of AKT and glutathione synthesis inhibits pancreatic cancer (Chio et al., 2016). While their study focused on the role of NRF2 in regulating mRNA translation in pancreatic cancer, their findings regarding glutathione synthesis being a key function of NRF2 are in concordance with our observations.

We recently found that the number of CRISPR/Cas9-induced DNA breaks dictates a gene-independent anti-proliferative response in cells, such that targeting amplified regions decreases viability (Munoz et al., 2016; Aguirre et al., 2016). This effect confounds the use of CRISPR/Cas9 negative selection screening to identify essential genes in amplified regions. We do not believe that this effect is relevant to this study, in which we have performed positive selection screens to identify genes whose loss promotes proliferation under drug treatment. Moreover, we directly compare the same cells under two conditions; thus, any genes that are affected by the gene-independent effect will score in both conditions.

A recent vemurafenib BRAF^V600E basket trial showed that 42% of lung cancers with the BRAF V600E mutation responded to vemurafenib (Hyman et al., 2015). As seen with vemurafenib treatment in melanoma or with EGFR inhibitors in lung cancer, acquired resistance will likely arise. Furthermore, while MEK inhibitors only elicit responses in a small number of lung cancer patients (Blumenschein et al., 2015), these responders are also likely to develop resistance. Predicting how resistance may arise in these patients will be important for developing more effective combination therapies. In addition, for those patients that do not initially respond, intrinsic resistance in a subset of these patients may be explained by the mechanisms we describe here. The KEAP1/NRF2 pathway is genetically altered in approximately 30% of lung squamous cell carcinomas and approximately 20% of lung adenocarcinomas. Alterations in this pathway can co-occur with alterations in the RTK/Ras pathway (Cerami et al., 2012; Gao et al., 2013; Cancer Genome Atlas Research Network, 2014), although KEAP1/NRF2 alterations are enriched in the ‘oncogene negative’ subset of lung cancers (Cancer Genome Atlas Research Network, 2014). BRAF and MEK inhibitors are currently being tested in clinical trials for RAS- and BRAF-mutant lung cancer. However, for most of these trials matched pre-treatment and post-relapse biopsy specimens are not available for molecular analysis of resistance mechanisms. Gainor et al. recently identified a NRF2 mutation in a patient with acquired resistance to an ALK inhibitor (Gainor et al., 2016). This mutation (E79Q) is in a mutational hotspot and has previously been shown to impair recognition of NRF2 by KEAP1, thus activating the pathway (Shibata et al., 2008b). This tumor also harbored a secondary ALK mutation of unknown function and became resistant to a second generation ALK inhibitor. Thus it is possible that the NRF2 mutation contributed to survival in the presence of crizotinib treatment and allowed the cells to acquire additional resistance mutations over time.

Although KEAP1/NRF2 alterations are known to confer resistance to chemotherapy, KEAP1/NRF2 mutation status is not used to make treatment decisions in lung cancer. As more patients are treated with RTK/MAPK inhibitors, analyzing KEAP1 and NRF2 status in pre-treatment and post-resistance tumor samples will determine if loss of KEAP1 or gain of NRF2 are clinically relevant mechanisms of acquired and intrinsic resistance to these therapies in lung cancer. Stratifying patients based on these findings will be important for evaluating the efficacy of these inhibitors in clinical trials and for choosing the appropriate treatment for patients.

Materials and methods

Cell lines and reagents

Cells were obtained from ATCC and fingerprinted as in Barretina et al (Barretina et al., 2012). Cells were maintained in RPMI-1640 (NCI-H1299, HCC364, NCI-H1975, and HCC827; Corning, Corning, NY), McCoy’s 5A (CALU1; Gibco, Waltham, MA) or DMEM (MGH-065; Invitrogen, Carlsbad, CA) supplemented with 2 mM glutamine, 50 U/mL penicillin, 50 U/mL of streptomycin (Gibco), and 10% fetal bovine serum (Sigma, St. Louis, MO), and incubated at 37°C in 5% CO₂. MGH-065 cells were derived as previously described (Crystal et al., 2014). Cell lines were tested for mycoplasma prior to screening. Trametinib, vemurafenib, erlotinib, and afatinib, were purchased from Selleck Chemicals (Houston, TX). LDK-378 and Crizotinib were synthesized by the Novartis Global Discovery Chemistry Department.

Screen optimization for genome scale screens

Blasticidin and puromycin concentrations were optimized for each cell line by treating with different concentrations of drug for 3 days (puromycin) or 7 days (blasticidin). The lowest concentration of drug that killed all cells was used in the screens.

To produce Cas9-expressing cell lines, 200,000–400,000 cells were seeded in one well of a 6-well plate. The following day, cells were infected with 3 mL of pLX311-Cas9 virus with a final concentration of 4 ug/mL polybrene. Cells were spun for 2 hr at 2000 rpm at 30°C. 24 hr after infection, cells were selected with blasticidin for 7 days.

To determine Cas9 activity, parental cell lines and Cas9-expressing cell lines were infected with pXPR_011, a Cas9 activity reporter which expresses eGFP as well as a guide RNA targeting eGFP (Doench et al., 2014). 200,000–400,000 cells were seeded in six wells of a 6-well plate and were infected with 25–100 µL virus with a final concentration of 4 µg/mL polybrene. Cells were spun 2 hr at 2000 rpm at 30°C. 24 hr after infection, each well was split into two wells, one of which was selected with puromycin. After 2–3 days of puromycin selection, cells were counted and those with 30–40% infection efficiency were kept for the Cas9 activity assay. After 7 days of puromycin selection, cells were analyzed on an LSRII flow cytometer to determine the amount of GFP-positive cells. Parental cells not expressing Cas9 or pXPR_011 were used as a negative control. Cells expressing pXPR_011 but not Cas9 were used as a positive control.

To optimize inhibitor concentrations, Cas9-expressing cells were infected with different amounts of emptyT virus (to mimic sgRNA infection) and were selected with puromycin. After 3 days of puromycin selection, cells were counted and those with 30–40% infection efficiency were used to optimize inhibitor concentration. Cells were kept in puromycin selection for one week prior to optimizing inhibitor concentration.

To determine the optimal drug concentration for the screens, cells expressing Cas9 and emptyT were treated with different concentrations of drug for three weeks. Cells were passaged or fresh drug-containing media was added every 3–4 days. Cells were counted at each passage. The lowest concentration of drug that resulted in death or proliferative arrest was used in the screen (Extended Data Figure 1) In parallel, cells were treated with different concentrations of inhibitor for 24 hr and then lysed in RIPA buffer. Immunoblots were performed with total and phospho-ERK antibodies to determine the concentration of inhibitor that blocked ERK phosphorylation.

To titer the GeCKO v2 library in Cas9-expressing cells, 3 × 10^∧6 cells were seeded per well in a 12-well plate and were infected with different amounts of virus (0, 50, 100, 150, 200, 400 uL), with a final concentration of 4–8 ug/mL polybrene. Cells were spun for 2 hr at 2000 rpm at 30°C. Approximately 6 hr after infection, cells were split into 6-well plates. For each amount of virus, 100,000 cells per well were plated in two wells. 24 hr after infection, one well was treated with puromycin and one with media alone. After 2–3 days of selection, cells were counted to determine the amount of virus that resulted in 30–40% infection efficiency, and this amount of virus was used in the screen.

GeCKO v2 library construction

See Sanjana et al (Sanjana et al., 2014).

Genome-scale CRISPR knockout drug resistance screens with GeCKO v2 library

For each screen, two infection replicates were performed. 150 × 10^∧6 cells were infected per replicate with 40% infection efficiency, in order to obtain 500 cells per sgRNA after selection (60 × 10^∧6 surviving cells containing 120,000 sgRNAs). 3 × 10^∧6 cells per well were seeded in 12-well plates and were infected with the amount of virus determined during optimization, with a final polybrene concentration of 4 µg/mL. Plates were spun for 2 hr at 2000 rpm at 30°C. Approximately 6 hr after infection, all wells within a replicate were pooled and were split into T225 flasks. 24 hr after infection, cells were selected in puromycin for one week and were passaged as necessary. After one week of puromycin selection, 60 × 10^∧6 cells were harvested for the Day 0 time point, and 60 × 10^∧6 cells were treated with drug. HCC364 cells were treated with 24 nM trametinib or 6.25 µM vemurafenib; H1299 cells were treated with 1.5 µM trametinib; and CALU1 cells were treated with 50 nM trametinib. Cells were passaged or fresh drug-containing media was added every 3–4 days. Drug-treated cells were harvested on Day 14 and Day 21 of drug treatment. To harvest cells, cells were trypsinized, spun down, washed with PBS, and the cell pellets were frozen at −80°C.

Genomic DNA was extracted using the Qiagen Blood and Cell Culture DNA Maxi Kit according to the manufacturer’s protocol.

Sequencing and analysis of genome-scale screens

See Doench et al (Doench et al., 2016).

Screen optimization for focused sgRNA library screen

Live cell time-lapse imaging was used to determine optimal drug concentrations in HCC827 and MGH065 cells. Photomicrographs were taken every 6 hr using an Incucyte live cell imager (Essen Biosciences) and confluence was measured over 130 hr in culture.

Focused sgRNA library construction

The sgRNA libraries were designed as previously described (Wang et al., 2014). A modified tracrRNA scaffold (Chen et al., 2013) for Cas9 loading was cloned into the sgRNA vectors before cloning of the guide RNAs. This sgRNA library targets ~2700 genes and is comprised of 20 sgRNAs per gene.

Focused sgRNA library screens

For the screens in HCC827 and MGH065, cells were infected with a lentiviral sgRNA pool at a representation of 1000 cells per sgRNA at an MOI of 0.5. Cells were selected for four days in the presence of puromycin, and a reference sample was collected 72 hr after selection to ensure adequate selection/representation. Cells were split into a DMSO control arm and a treatment arm. HCC827 cells were treated with 0.01 uM erlotinib and MGH-065 was treated with 0.5 uM Crizotinib or LDK378. Cells were propagated for a total of 14 days with an average sgRNA representation of ≥1000 maintained at each passage. 100 million cells were harvested for DNA extraction by QIAmp Blood Maxi kit (Qiagen), sgRNAs were PCR amplified from 100 ug of genomic DNA, and PCR fragments of 260–280 bp were purified using Agencourt AMpure XP beads (Beckman). The resulting fragments were sequenced on a Hiseq 2500 (Illumina) with a single end 50 bp run. Sequencing reads were aligned to the sgRNA library and the enrichment or loss of individual bar codes or sgRNA were quantified.

Vectors

Cas9 in the pLX311 backbone (pXPR_BRD111) and sgRNAs in the pXPR_BRD003 backbone were obtained from the Genetic Perturbation Platform at the Broad Institute.

sgKEAP1 arrayed infection

500,000 cells per well were seeded in 48-well plates in 250 µL media with 4 µg/mL polybrene. 25 µL virus (sgKEAP1 or sgEGFP) was added per well and plates were spun 2 hr at 2000 rpm at 30°C. 6 hr later, each well was split into a 6 cm dish. 24 hr after infection, cells were selected with puromycin for one week.

TIDE-PCR

We performed TIDE-PCR (Brinkman et al., 2014) on CALU1 and HCC364 sgKEAP1 cells (forward primer CTGGTACATGACAGCACCGT and reverse primer TGCTTCACCTACTTTGCAGGA for sgKEAP1–1 cells; forward primer CTCCAGTTTCCTGCCTTGACAT and reverse primer GAACCCCATGCAGCCAGAT for sgKEAP1–2 cells) and found the editing efficiency to be approximately 93%, of which approximately 70% is a one base pair insertion, which would lead to a frame shift.

qRT-PCR

RNA was harvested using a Qiagen RNeasy Kit and was reverse transcribed into cDNA using SuperScriptIII according to the manufacturer’s recommendations. KEAP1 mRNA expression was measured using primers downstream of the sgRNA targeting sites. KEAP1 mRNA levels in KEAP1^KO cells were ~35–40% of those in control cells.

Cytoplasmic/nuclear fractionation

5 × 10^∧5 cells were seeded in 10 cm dishes. The following day, cells were treated with trametinib (25 nM for HCC364 or 50 nM for CALU1) or DMSO. After 72 hr of drug treatment, cells were lysed and fractionated using NE-PER Nuclear and Cytoplasmic Extraction Reagents (Pierce Biotechnology) according to the manufacturer’s recommendations.

Immunoblotting

Cells were lysed in RIPA buffer containing protease and phosphatase inhibitors and were cleared by centrifugation. Protein was quantified using the Pierce BCA assay, and lysate concentrations were normalized. Lysates were run on SDS-PAGE gels and were transferred to nitrocellulose membranes using the Invitrogen iBlot system. Membranes were blocked for one hour in 5% milk in Tris-buffered saline with 0.1% Tween (TBS-T). Membranes were incubated overnight at 4°C with primary antibodies in 5% BSA in TBS-T. Membranes were washed three times in TBST-T then incubated 1 hr at room temperature with secondary antibodies in 5% BSA in TBS-T. Membranes were washed in TBS-T and imaged on a Li-Cor Odyssey Infrared Imaging System. Primary antibodies were total ERK (Cell Signaling #9102), phospho-ERK (Cell Signaling #4370), total AKT (Cell Signaling #9272), phospho-AKT (Cell Signaling #4060), GAPDH (Cell Signaling #5174), LAMIN A/C (Cell Signaling #4777), KEAP1 (Proteintech 10503–2-AP), and NRF2 (Santa Cruz Biotechnology sc-13032).

Crystal violet assays

1000–10,000 cells were seeded in 12-well or 24-well plates in the indicated drug conditions. Media containing fresh drug was replaced every 3–4 days. After the indicated number of days, cells were washed in PBS, fixed in 10% formalin for 15 min, and stained with 0.1% crystal violet in 10% ethanol for 20 min. After acquiring images, crystal violet was extracted in 10% acetic acid for 20 min. The absorbance at 565 nm was determined using a Spectramax plate reader.

ORF expression

293 T cells were seeded in DMEM + 10% FBS + 0.1% Pen/Strep in 6 cm dishes. 24 hr later, cells were transfected with 100 ng VSVG, 900 ng delta8.9, and 1 µg pLX317-ORF plasmid using OptiMEM and Mirus TransIT. 16 hr after transfection, media was changed to DMEM +30% FBS + 1% Pen/Strep. Virus was harvested 24 hr later. Cell lines were seeded in 6-well plates and were infected the following day with 1:5 dilution of virus containing 4 µg/mL polybrene. 24 hr after infection, cells were selected with puromycin.

DCFDA assays to measure ROS

Unless otherwise indicated, cells were treated with drug for 3 days. Cells were trypsinized and resuspended in media with 10 μM DCFDA (Sigma D6883) and incubated at 37°C for 90 min in the dark. For a positive control, parental cells were treated with 20 μM tert-butyl hydroperoxide (Sigma Aldrich 458139) during incubation. For a negative control, parental cells were incubated in media without DCFDA. DCFDA fluorescence was detected by flow cytometry, using the FITC channel on an LSRII flow cytometer (BD Biosciences).

Metabolite profiling

Metabolite extraction was performed as described previously (Yuan et al., 2012). In brief, cells were plated in 10 cm dishes and treated for 72 hr with either DMSO or trametinib (25 nM for HCC364 and 50 nM for CALU1) in quadruplicate. Four hours before metabolite extraction, cells were re-fed with fresh media containing drug or DMSO. Cells were then washed with PBS at 37°C, followed by metabolite extraction by scraping in 80% methanol (on dry ice), insoluble materials collected, two sequential extractions combined and stored at −80°C. Prior to mass spectrometry, the metabolite extracts were lyophilized, resolubilized in 20 µL of LC/MS grade water, and then analyzed by liquid chromatography tandem mass spectrometry (LC-MS/MS) using positive ion/negative ion polarity switching via selected reaction monitoring (SRM) on a 5500 QTRAP hybrid triple quadrupole mass spectrometer (AB/SCIEX), as previously described (Yuan et al., 2012). Peaks were integrated using MultiQuant 2.1 data analysis software (AB/SCIEX). Metabolite differences were analyzed by normalizing samples by total levels and comparing replicate samples in groups that satisfy significance of p<0.05 using an unpaired two-tailed t-test.

Xenotransplantation and in vivo drug treatments

Mice were maintained and handled in accordance with the Novartis Institutes for Biomedical Research (NIBR) Animal Care and Use Committee protocols and regulations. Nude mice were inoculated subcutaneously with 10 × 10⁶ HCC827 KEAP1^KO or control cells implanted bilaterally in 50%PBS + 50% matrigel. Once the average tumor volume reached 100–200 mm³ in size, mice were randomized to three groups: vehicle, 5 mg/kg erlotinib, or 12 mg/kg erlotinib. Erlotinib was administered by oral gavage once daily for 40 days and tumor volume was measured twice weekly.

qPCR primer sequences

Gene	Forward primer	Reverse primer
KEAP1	GATTGGCTGTGTGGAGTTGC	GCAGTGGGACAGGTTGAAGA
KEAP1	TCGATGGCCACATCTATGCC	CCGATCCTTCGTGTCAGCAT
NFE2L2	TCCAGTCAGAAACCAGTGGAT	GAATGTCTGCGCCAAAAGCTG
GCLC	GTGTTTCCTGGACTGATCCCA	TCCCTCATCCATCTGGCAAC
GCLM	CATTTACAGCCTTACTGGGAGG	ATGCAGTCAAATCTGGTGGCA
HO1	CTTTCAGAAGGGCCAGGTGA	GTAGACAGGGGCGAAGACTG
NQO1	CTCACCGAGAGCCTAGTTCC	CGTCCTCTCTGAGTGAGCCA
MRP1	CTCTATCTCTCCCGACATGACC	AGCAGACGATCCACAGCAAAA
TKT	GCTGAACCTGAGGAAGATCA	TGTCGAAGTATTTGCCGGTG
TALDO1	GTCATCAACCTGGGAAGGAA	CAACAAATGGGGAGATGAGG
PGD	ATATAGGGACACCACAAGACGG	GCATGAGCGATGGGCCATA
MTHFD2	TGTCCTCAACAAAACCAGGG	TTCCTCTGAAATTGAAGCTGG
ME1	CTGCCTGTCATTCTGGATGT	ACCTCTTACTCTTCTCTGCC
G6PD	TGACCTGGCCAAGAAGAAGA	CAAAGAAGTCCTCCAGCTTG
SHMT1	TGAACACTGCCATGTGGTGACC	TCTTTGCCAGTCTTGGGATCC
SHMT2	GCCTCATTGACTACAACCAGCTG	ATGTCTGCCAGCAGGTGTGCTT
ACTIN	CAACCGCGAGAAGATGACC	ATCACGATGCCAGTGGTACG

Open in a new tab

sgRNA sequences

Name	Target sequence
sgGFP	GGCGAGGGCGATGCCACCTA
sgKEAP1–1	CTTGTGGGCCATGAACTGGG
sgKEAP1–2	TGTGTCCTCCACGTCATGAA
sgKEAP1–3	GAGGACACACTTCTCGCCCA
sgKEAP1–4	ACTGGGCGGCCGGTGCATCC
sgLACZ-1	AACGGCGGATTGACCGTAAT
sgLACZ-2	CTAACGCCTGGGTCGAACGC

Open in a new tab

Acknowledgements

We thank Eejung Kim and Kristen Hurov for NRF2 and KEAP1 ORFs, Celina Keating for support and NGS data acquisition, Thales Papagiannakopoulos and Rodrigo Romero for helpful discussions and advice on ROS assays; and Ophir Shalem and Neville Sanjana for advice on screening. We also thank William Sellers and Nicholas Keen for critical discussions. This work was supported in part by US NIH grants: R01 CA130988 (WCH), U01 CA176058 (WCH) and U01 CA199253 (WCH). EBK was supported by postdoctoral fellowships from the Hope Funds for Cancer Research (HFCR-11-03–03) and NIH F32 (CA189306). NI was supported by postdoctoral fellowships from the Susan G Komen Foundation (PDF12230602) and the Terri Brodeur Breast Cancer Foundation. AJA was supported by the Pancreatic Cancer Action Network Samuel Stroum Fellowship, Hope Funds for Cancer Research Postdoctoral Fellowship, American Society of Clinical Oncology Young Investigator Award, Dana-Farber Cancer Institute Hale Center for Pancreatic Cancer, Perry S Levy Endowed Fellowship, and the Harvard Catalyst and Harvard Clinical and Translational Science Center (UL1 TR001102).

Funding Statement

The funders had no role in study design, data collection and interpretation, or the decision to submit the work for publication.

Funding Information

This paper was supported by the following grants:

National Cancer Institute R01 CA130998 to William C Hahn.
National Cancer Institute U01 CA176058 to William C Hahn.
National Cancer Institute U01 CA199253 to William C Hahn.
Hope Funds for Cancer Research Postdoctoral Fellowship HFCR-11-03-03 to Elsa B Krall.
National Institutes of Health Postdoctoral Fellowship F32 CA189306 to Elsa B Krall.
Susan G. Komen Foundation Postdoctoral Fellowship PDF12230602 to Nina Ilic.
Terri Brodeur Breast Cancer Foundation Postdoctoral Fellowship to Nina Ilic.
Pancreatic Cancer Action Network Samuel Stroum Fellowship to Andrew J Aguirre.
American Society of Clinical Oncology Young Investigator Award to Andrew J Aguirre.
Dana-Farber Cancer Institute Hale Center for Pancreatic Cancer to Andrew J Aguirre.
Perry S. Levy Endowed Fellowship to Andrew J Aguirre.
Harvard Catalyst and Harvard Clinical and Translational Science Center UL1 TR001102 to Andrew J Aguirre.

Additional information

Competing interests

WCH: A consultant and receives research support from Novartis.

The other authors declare that no competing interests exist.

Author contributions

EBK, Conceptualization, Data curation, Formal analysis, Investigation, Methodology, Writing—original draft, Writing—review and editing, Designed, performed, and analyzed screens, Designed and performed in vitro experiments, Wrote the manuscript.

BW, Conceptualization, Data curation, Formal analysis, Investigation, Methodology, Writing—original draft, Writing—review and editing, Performed screens, Designed and performed in vitro experiments, Wrote the manuscript.

DMM, Conceptualization, Data curation, Formal analysis, Investigation, Writing—review and editing, Designed, performed, and analyzed screens, Designed and performed in vitro experiments, Designed and performed in vivo experiments, Edited the manuscript.

NI, Data curation, Formal analysis, Investigation, Methodology, Writing—review and editing, Designed and performed in vitro experiments, Edited the manuscript.

SR, Data curation, Formal analysis, Investigation, Writing—review and editing, Designed and performed in vitro experiments, Edited the manuscript.

MJN, Formal analysis, Analyzed patient data, Edited the manuscript.

KY, Data curation, Provided sequencing support and data acquisition.

DAR, Data curation, Provided sequencing support and data acquisition.

AJA, Methodology, Designed in vitro experiments, Edited the manuscript.

JWK, Formal analysis, Performed bioinformatics analyses, Edited the manuscript.

AJR, Data curation, Formal analysis, Analyzed patient data.

JFG, Data curation, Formal analysis, Obtained patient samples, Edited the manuscript.

JAW, Methodology, Designed in vivo experiments.

JMA, Data curation, Methodology, Performed mass spectrometry.

JGD, Conceptualization, Formal analysis, Methodology, Designed and analyzed screens, Designed in vitro experiments, Edited the manuscript.

PAJ, Data curation, Obtained patient samples, Edited the manuscript.

ATS, Resources, Obtained patient samples.

REMcDI, Supervision, Methodology, Supervised the study, Edited the manuscript.

JAE, Supervision, Supervised the study.

FS, Supervision, Supervised the study.

MRS, Conceptualization, Supervision, Methodology, Designed screens, Designed in vitro experiments, Supervised the study.

WCH, Conceptualization, Supervision, Funding acquisition, Writing—original draft, Writing—review and editing, Designed screens, Designed in vitro experiments, Supervised the study, Wrote the manscript.

Ethics

Animal experimentation: Mice were maintained and handled in accordance with the Novartis Institutes for Biomedical Research (NIBR) Animal Care and Use Committee protocols and regulations.

Additional files

Supplementary file 1. Genome scale CRISPR-Cas9 screens in CALU1, HCC364, and NCIH1299 cells.

For each sgRNA, the average log2 fold-change (LFC) between drug-treated samples and Day 0 samples is shown. The hits from each screen, as determined by LFC or by STARS score, are also summarized.

DOI: http://dx.doi.org/10.7554/eLife.18970.025

elife-18970-supp1.xlsx^{(30MB, xlsx)}

DOI: 10.7554/eLife.18970.025

Supplementary file 2. Focused sgRNA library screens in HCC827 and MGH065 cells.

Raw counts and Z-scores are shown for each sgRNA in DMSO- or drug-treated samples.

DOI: http://dx.doi.org/10.7554/eLife.18970.026

elife-18970-supp2.xlsx^{(9.1MB, xlsx)}

DOI: 10.7554/eLife.18970.026

Supplementary file 3. Metabolite Profiling of KEAP1 knockout cells.

Metabolite levels are shown for DMSO- or drug-treated samples. Metabolite differences were analyzed by normalizing samples by total levels and comparing replicate samples in groups that satisfy significance of p<0.05 using an unpaired two-tailed t-test.

DOI: http://dx.doi.org/10.7554/eLife.18970.027

elife-18970-supp3.xlsx^{(563.1KB, xlsx)}

DOI: 10.7554/eLife.18970.027

References

Aguirre AJ, Meyers RM, Weir BA, Vazquez F, Zhang CZ, Ben-David U, Cook A, Ha G, Harrington WF, Doshi MB, Kost-Alimova M, Gill S, Xu H, Ali LD, Jiang G, Pantel S, Lee Y, Goodale A, Cherniack AD, Oh C, Kryukov G, Cowley GS, Garraway LA, Stegmaier K, Roberts CW, Golub TR, Meyerson M, Root DE, Tsherniak A, Hahn WC. Genomic copy number dictates a Gene-Independent cell response to CRISPR/Cas9 Targeting. Cancer Discovery. 2016;6:914–929. doi: 10.1158/2159-8290.CD-16-0154. [DOI] [PMC free article] [PubMed] [Google Scholar]
Balak MN, Gong Y, Riely GJ, Somwar R, Li AR, Zakowski MF, Chiang A, Yang G, Ouerfelli O, Kris MG, Ladanyi M, Miller VA, Pao W. Novel D761Y and common secondary T790M mutations in epidermal growth factor receptor-mutant lung adenocarcinomas with acquired resistance to kinase inhibitors. Clinical Cancer Research. 2006;12:6494–6501. doi: 10.1158/1078-0432.CCR-06-1570. [DOI] [PubMed] [Google Scholar]
Barretina J, Caponigro G, Stransky N, Venkatesan K, Margolin AA, Kim S, Wilson CJ, Lehár J, Kryukov GV, Sonkin D, Reddy A, Liu M, Murray L, Berger MF, Monahan JE, Morais P, Meltzer J, Korejwa A, Jané-Valbuena J, Mapa FA, Thibault J, Bric-Furlong E, Raman P, Shipway A, Engels IH, Cheng J, Yu GK, Yu J, Aspesi P, de Silva M, Jagtap K, Jones MD, Wang L, Hatton C, Palescandolo E, Gupta S, Mahan S, Sougnez C, Onofrio RC, Liefeld T, MacConaill L, Winckler W, Reich M, Li N, Mesirov JP, Gabriel SB, Getz G, Ardlie K, Chan V, Myer VE, Weber BL, Porter J, Warmuth M, Finan P, Harris JL, Meyerson M, Golub TR, Morrissey MP, Sellers WR, Schlegel R, Garraway LA. The cancer cell line Encyclopedia enables predictive modelling of anticancer drug sensitivity. Nature. 2012;483:603–607. doi: 10.1038/nature11003. [DOI] [PMC free article] [PubMed] [Google Scholar]
Berns K, Horlings HM, Hennessy BT, Madiredjo M, Hijmans EM, Beelen K, Linn SC, Gonzalez-Angulo AM, Stemke-Hale K, Hauptmann M, Beijersbergen RL, Mills GB, van de Vijver MJ, Bernards R. A functional genetic approach identifies the PI3K pathway as a major determinant of trastuzumab resistance in breast cancer. Cancer Cell. 2007;12:395–402. doi: 10.1016/j.ccr.2007.08.030. [DOI] [PubMed] [Google Scholar]
Blumenschein GR, Smit EF, Planchard D, Kim DW, Cadranel J, De Pas T, Dunphy F, Udud K, Ahn MJ, Hanna NH, Kim JH, Mazieres J, Kim SW, Baas P, Rappold E, Redhu S, Puski A, Wu FS, Jänne PA. A randomized phase II study of the MEK1/MEK2 inhibitor trametinib (GSK1120212) compared with docetaxel in KRAS-mutant advanced non-small-cell lung cancer (NSCLC)†. Annals of Oncology. 2015;26:894–901. doi: 10.1093/annonc/mdv072. [DOI] [PMC free article] [PubMed] [Google Scholar]
Brinkman EK, Chen T, Amendola M, van Steensel B. Easy quantitative assessment of genome editing by sequence trace decomposition. Nucleic Acids Research. 2014;42:e168. doi: 10.1093/nar/gku936. [DOI] [PMC free article] [PubMed] [Google Scholar]
Cancer Genome Atlas Research Network Comprehensive molecular profiling of lung adenocarcinoma. Nature. 2014;511:543–550. doi: 10.1038/nature13385. [DOI] [PMC free article] [PubMed] [Google Scholar]
Cerami E, Gao J, Dogrusoz U, Gross BE, Sumer SO, Aksoy BA, Jacobsen A, Byrne CJ, Heuer ML, Larsson E, Antipin Y, Reva B, Goldberg AP, Sander C, Schultz N. The cBio cancer genomics portal: an open platform for exploring multidimensional cancer genomics data. Cancer Discovery. 2012;2:401–404. doi: 10.1158/2159-8290.CD-12-0095. [DOI] [PMC free article] [PubMed] [Google Scholar]
Chen B, Gilbert LA, Cimini BA, Schnitzbauer J, Zhang W, Li GW, Park J, Blackburn EH, Weissman JS, Qi LS, Huang B. Dynamic imaging of genomic loci in living human cells by an optimized CRISPR/Cas system. Cell. 2013;155:1479–1491. doi: 10.1016/j.cell.2013.12.001. [DOI] [PMC free article] [PubMed] [Google Scholar]
Chio II, Jafarnejad SM, Ponz-Sarvise M, Park Y, Rivera K, Palm W, Wilson J, Sangar V, Hao Y, Öhlund D, Wright K, Filippini D, Lee EJ, Da Silva B, Schoepfer C, Wilkinson JE, Buscaglia JM, DeNicola GM, Tiriac H, Hammell M, Crawford HC, Schmidt EE, Thompson CB, Pappin DJ, Sonenberg N, Tuveson DA. NRF2 promotes tumor maintenance by modulating mRNA translation in pancreatic cancer. Cell. 2016;166:963–976. doi: 10.1016/j.cell.2016.06.056. [DOI] [PMC free article] [PubMed] [Google Scholar]
Corcoran RB, Ebi H, Turke AB, Coffee EM, Nishino M, Cogdill AP, Brown RD, Della Pelle P, Dias-Santagata D, Hung KE, Flaherty KT, Piris A, Wargo JA, Settleman J, Mino-Kenudson M, Engelman JA. EGFR-mediated re-activation of MAPK signaling contributes to insensitivity of BRAF mutant colorectal cancers to RAF inhibition with vemurafenib. Cancer Discovery. 2012;2:227–235. doi: 10.1158/2159-8290.CD-11-0341. [DOI] [PMC free article] [PubMed] [Google Scholar]
Corcoran RB, Cheng KA, Hata AN, Faber AC, Ebi H, Coffee EM, Greninger P, Brown RD, Godfrey JT, Cohoon TJ, Song Y, Lifshits E, Hung KE, Shioda T, Dias-Santagata D, Singh A, Settleman J, Benes CH, Mino-Kenudson M, Wong KK, Engelman JA. Synthetic lethal interaction of combined BCL-XL and MEK inhibition promotes tumor regressions in KRAS mutant cancer models. Cancer Cell. 2013;23:121–128. doi: 10.1016/j.ccr.2012.11.007. [DOI] [PMC free article] [PubMed] [Google Scholar]
Crystal AS, Shaw AT, Sequist LV, Friboulet L, Niederst MJ, Lockerman EL, Frias RL, Gainor JF, Amzallag A, Greninger P, Lee D, Kalsy A, Gomez-Caraballo M, Elamine L, Howe E, Hur W, Lifshits E, Robinson HE, Katayama R, Faber AC, Awad MM, Ramaswamy S, Mino-Kenudson M, Iafrate AJ, Benes CH, Engelman JA. Patient-derived models of acquired resistance can identify effective drug combinations for cancer. Science. 2014;346:1480–1486. doi: 10.1126/science.1254721. [DOI] [PMC free article] [PubMed] [Google Scholar]
DeNicola GM, Karreth FA, Humpton TJ, Gopinathan A, Wei C, Frese K, Mangal D, Yu KH, Yeo CJ, Calhoun ES, Scrimieri F, Winter JM, Hruban RH, Iacobuzio-Donahue C, Kern SE, Blair IA, Tuveson DA. Oncogene-induced Nrf2 transcription promotes ROS detoxification and tumorigenesis. Nature. 2011;475:106–109. doi: 10.1038/nature10189. [DOI] [PMC free article] [PubMed] [Google Scholar]
DeNicola GM, Chen PH, Mullarky E, Sudderth JA, Hu Z, Wu D, Tang H, Xie Y, Asara JM, Huffman KE, Wistuba II, Minna JD, DeBerardinis RJ, Cantley LC. NRF2 regulates serine biosynthesis in non-small cell lung cancer. Nature Genetics. 2015;47:1475–1481. doi: 10.1038/ng.3421. [DOI] [PMC free article] [PubMed] [Google Scholar]
Ding L, Getz G, Wheeler DA, Mardis ER, McLellan MD, Cibulskis K, Sougnez C, Greulich H, Muzny DM, Morgan MB, Fulton L, Fulton RS, Zhang Q, Wendl MC, Lawrence MS, Larson DE, Chen K, Dooling DJ, Sabo A, Hawes AC, Shen H, Jhangiani SN, Lewis LR, Hall O, Zhu Y, Mathew T, Ren Y, Yao J, Scherer SE, Clerc K, Metcalf GA, Ng B, Milosavljevic A, Gonzalez-Garay ML, Osborne JR, Meyer R, Shi X, Tang Y, Koboldt DC, Lin L, Abbott R, Miner TL, Pohl C, Fewell G, Haipek C, Schmidt H, Dunford-Shore BH, Kraja A, Crosby SD, Sawyer CS, Vickery T, Sander S, Robinson J, Winckler W, Baldwin J, Chirieac LR, Dutt A, Fennell T, Hanna M, Johnson BE, Onofrio RC, Thomas RK, Tonon G, Weir BA, Zhao X, Ziaugra L, Zody MC, Giordano T, Orringer MB, Roth JA, Spitz MR, Wistuba II, Ozenberger B, Good PJ, Chang AC, Beer DG, Watson MA, Ladanyi M, Broderick S, Yoshizawa A, Travis WD, Pao W, Province MA, Weinstock GM, Varmus HE, Gabriel SB, Lander ES, Gibbs RA, Meyerson M, Wilson RK. Somatic mutations affect key pathways in lung adenocarcinoma. Nature. 2008;455:1069–1075. doi: 10.1038/nature07423. [DOI] [PMC free article] [PubMed] [Google Scholar]
Doench JG, Hartenian E, Graham DB, Tothova Z, Hegde M, Smith I, Sullender M, Ebert BL, Xavier RJ, Root DE. Rational design of highly active sgRNAs for CRISPR-Cas9-mediated gene inactivation. Nature Biotechnology. 2014;32:1262–1267. doi: 10.1038/nbt.3026. [DOI] [PMC free article] [PubMed] [Google Scholar]
Doench JG, Fusi N, Sullender M, Hegde M, Vaimberg EW, Donovan KF, Smith I, Tothova Z, Wilen C, Orchard R, Virgin HW, Listgarten J, Root DE. Optimized sgRNA design to maximize activity and minimize off-target effects of CRISPR-Cas9. Nature Biotechnology. 2016;34:184–191. doi: 10.1038/nbt.3437. [DOI] [PMC free article] [PubMed] [Google Scholar]
Engelman JA, Zejnullahu K, Mitsudomi T, Song Y, Hyland C, Park JO, Lindeman N, Gale CM, Zhao X, Christensen J, Kosaka T, Holmes AJ, Rogers AM, Cappuzzo F, Mok T, Lee C, Johnson BE, Cantley LC, Jänne PA. MET amplification leads to gefitinib resistance in lung cancer by activating ERBB3 signaling. Science. 2007;316:1039–1043. doi: 10.1126/science.1141478. [DOI] [PubMed] [Google Scholar]
Forbes SA, Beare D, Gunasekaran P, Leung K, Bindal N, Boutselakis H, Ding M, Bamford S, Cole C, Ward S, Kok CY, Jia M, De T, Teague JW, Stratton MR, McDermott U, Campbell PJ. COSMIC: exploring the world's knowledge of somatic mutations in human cancer. Nucleic Acids Research. 2015;43:D805–D811. doi: 10.1093/nar/gku1075. [DOI] [PMC free article] [PubMed] [Google Scholar]
Gainor JF, Dardaei L, Yoda S, Friboulet L, Leshchiner I, Katayama R, Dagogo-Jack I, Gadgeel S, Schultz K, Singh M, Chin E, Parks M, Lee D, DiCecca RH, Lockerman E, Huynh T, Logan J, Ritterhouse LL, Le LP, Muniappan A, Digumarthy S, Channick C, Keyes C, Getz G, Dias-Santagata D, Heist RS, Lennerz J, Sequist LV, Benes CH, Iafrate AJ, Mino-Kenudson M, Engelman JA, Shaw AT. Molecular mechanisms of resistance to first- and Second-Generation ALK Inhibitors in ALK-Rearranged Lung Cancer. Cancer Discovery. 2016;6:1118–1133. doi: 10.1158/2159-8290.CD-16-0596. [DOI] [PMC free article] [PubMed] [Google Scholar]
Gao J, Aksoy BA, Dogrusoz U, Dresdner G, Gross B, Sumer SO, Sun Y, Jacobsen A, Sinha R, Larsson E, Cerami E, Sander C, Schultz N. Integrative analysis of complex Cancer genomics and clinical profiles using the cBioPortal. Science Signaling. 2013;6:pl1. doi: 10.1126/scisignal.2004088. [DOI] [PMC free article] [PubMed] [Google Scholar]
Hayes JD, Dinkova-Kostova AT. The Nrf2 regulatory network provides an interface between redox and intermediary metabolism. Trends in Biochemical Sciences. 2014;39:199–218. doi: 10.1016/j.tibs.2014.02.002. [DOI] [PubMed] [Google Scholar]
Hyman DM, Puzanov I, Subbiah V, Faris JE, Chau I, Blay JY, Wolf J, Raje NS, Diamond EL, Hollebecque A, Gervais R, Elez-Fernandez ME, Italiano A, Hofheinz RD, Hidalgo M, Chan E, Schuler M, Lasserre SF, Makrutzki M, Sirzen F, Veronese ML, Tabernero J, Baselga J. Vemurafenib in multiple nonmelanoma cancers with BRAF V600 mutations. New England Journal of Medicine. 2015;373:726–736. doi: 10.1056/NEJMoa1502309. [DOI] [PMC free article] [PubMed] [Google Scholar]
Imielinski M, Berger AH, Hammerman PS, Hernandez B, Pugh TJ, Hodis E, Cho J, Suh J, Capelletti M, Sivachenko A, Sougnez C, Auclair D, Lawrence MS, Stojanov P, Cibulskis K, Choi K, de Waal L, Sharifnia T, Brooks A, Greulich H, Banerji S, Zander T, Seidel D, Leenders F, Ansén S, Ludwig C, Engel-Riedel W, Stoelben E, Wolf J, Goparju C, Thompson K, Winckler W, Kwiatkowski D, Johnson BE, Jänne PA, Miller VA, Pao W, Travis WD, Pass HI, Gabriel SB, Lander ES, Thomas RK, Garraway LA, Getz G, Meyerson M. Mapping the hallmarks of lung adenocarcinoma with massively parallel sequencing. Cell. 2012;150:1107–1120. doi: 10.1016/j.cell.2012.08.029. [DOI] [PMC free article] [PubMed] [Google Scholar]
Johannessen CM, Boehm JS, Kim SY, Thomas SR, Wardwell L, Johnson LA, Emery CM, Stransky N, Cogdill AP, Barretina J, Caponigro G, Hieronymus H, Murray RR, Salehi-Ashtiani K, Hill DE, Vidal M, Zhao JJ, Yang X, Alkan O, Kim S, Harris JL, Wilson CJ, Myer VE, Finan PM, Root DE, Roberts TM, Golub T, Flaherty KT, Dummer R, Weber BL, Sellers WR, Schlegel R, Wargo JA, Hahn WC, Garraway LA. COT drives resistance to RAF inhibition through MAP kinase pathway reactivation. Nature. 2010;468:968–972. doi: 10.1038/nature09627. [DOI] [PMC free article] [PubMed] [Google Scholar]
Johannessen CM, Johnson LA, Piccioni F, Townes A, Frederick DT, Donahue MK, Narayan R, Flaherty KT, Wargo JA, Root DE, Garraway LA. A melanocyte lineage program confers resistance to MAP kinase pathway inhibition. Nature. 2013;504:138–142. doi: 10.1038/nature12688. [DOI] [PMC free article] [PubMed] [Google Scholar]
Kim E, Ilic N, Shrestha Y, Zou L, Kamburov A, Zhu C, Yang X, Lubonja R, Tran N, Nguyen C, Lawrence MS, Piccioni F, Bagul M, Doench JG, Chouinard CR, Wu X, Hogstrom L, Natoli T, Tamayo P, Horn H, Corsello SM, Lage K, Root DE, Subramanian A, Golub TR, Getz G, Boehm JS, Hahn WC. Systematic functional interrogation of rare cancer variants identifies oncogenic Alleles. Cancer Discovery. 2016;6:714–726. doi: 10.1158/2159-8290.CD-16-0160. [DOI] [PMC free article] [PubMed] [Google Scholar]
Kobayashi A, Kang MI, Okawa H, Ohtsuji M, Zenke Y, Chiba T, Igarashi K, Yamamoto M. Oxidative stress sensor Keap1 functions as an adaptor for Cul3-based E3 ligase to regulate proteasomal degradation of Nrf2. Molecular and Cellular Biology. 2004;24:7130–7139. doi: 10.1128/MCB.24.16.7130-7139.2004. [DOI] [PMC free article] [PubMed] [Google Scholar]
Kosaka T, Yatabe Y, Endoh H, Yoshida K, Hida T, Tsuboi M, Tada H, Kuwano H, Mitsudomi T. Analysis of epidermal growth factor receptor gene mutation in patients with non-small cell lung cancer and acquired resistance to gefitinib. Clinical Cancer Research. 2006;12:5764–5769. doi: 10.1158/1078-0432.CCR-06-0714. [DOI] [PubMed] [Google Scholar]
Manchado E, Weissmueller S, Morris JP, Chen CC, Wullenkord R, Lujambio A, de Stanchina E, Poirier JT, Gainor JF, Corcoran RB, Engelman JA, Rudin CM, Rosen N, Lowe SW. A combinatorial strategy for treating KRAS-mutant lung cancer. Nature. 2016;534:647–651. doi: 10.1038/nature18600. [DOI] [PMC free article] [PubMed] [Google Scholar]
Mitsuishi Y, Taguchi K, Kawatani Y, Shibata T, Nukiwa T, Aburatani H, Yamamoto M, Motohashi H. Nrf2 redirects glucose and glutamine into anabolic pathways in metabolic reprogramming. Cancer Cell. 2012;22:66–79. doi: 10.1016/j.ccr.2012.05.016. [DOI] [PubMed] [Google Scholar]
Munoz DM, Cassiani PJ, Li L, Billy E, Korn JM, Jones MD, Golji J, Ruddy DA, Yu K, McAllister G, DeWeck A, Abramowski D, Wan J, Shirley MD, Neshat SY, Rakiec D, de Beaumont R, Weber O, Kauffmann A, McDonald ER, Keen N, Hofmann F, Sellers WR, Schmelzle T, Stegmeier F, Schlabach MR. CRISPR screens provide a comprehensive assessment of cancer vulnerabilities but generate False-Positive hits for highly amplified genomic regions. Cancer Discovery. 2016;6:900–913. doi: 10.1158/2159-8290.CD-16-0178. [DOI] [PubMed] [Google Scholar]
Nazarian R, Shi H, Wang Q, Kong X, Koya RC, Lee H, Chen Z, Lee MK, Attar N, Sazegar H, Chodon T, Nelson SF, McArthur G, Sosman JA, Ribas A, Lo RS. Melanomas acquire resistance to B-RAF(V600E) inhibition by RTK or N-RAS upregulation. Nature. 2010;468:973–977. doi: 10.1038/nature09626. [DOI] [PMC free article] [PubMed] [Google Scholar]
Ohta T, Iijima K, Miyamoto M, Nakahara I, Tanaka H, Ohtsuji M, Suzuki T, Kobayashi A, Yokota J, Sakiyama T, Shibata T, Yamamoto M, Hirohashi S. Loss of Keap1 function activates Nrf2 and provides advantages for lung cancer cell growth. Cancer Research. 2008;68:1303–1309. doi: 10.1158/0008-5472.CAN-07-5003. [DOI] [PubMed] [Google Scholar]
Prahallad A, Sun C, Huang S, Di Nicolantonio F, Salazar R, Zecchin D, Beijersbergen RL, Bardelli A, Bernards R. Unresponsiveness of colon cancer to BRAF(V600E) inhibition through feedback activation of EGFR. Nature. 2012;483:100–103. doi: 10.1038/nature10868. [DOI] [PubMed] [Google Scholar]
Rudin CM, Hong K, Streit M. Molecular characterization of acquired resistance to the BRAF inhibitor dabrafenib in a patient with BRAF-mutant non-small-cell lung cancer. Journal of Thoracic Oncology. 2013;8:e41–e42. doi: 10.1097/JTO.0b013e31828bb1b3. [DOI] [PMC free article] [PubMed] [Google Scholar]
Sanjana NE, Shalem O, Zhang F. Improved vectors and genome-wide libraries for CRISPR screening. Nature Methods. 2014;11:783–784. doi: 10.1038/nmeth.3047. [DOI] [PMC free article] [PubMed] [Google Scholar]
Shalem O, Sanjana NE, Hartenian E, Shi X, Scott DA, Mikkelsen TS, Heckl D, Ebert BL, Root DE, Doench JG, Zhang F. Genome-scale CRISPR-Cas9 knockout screening in human cells. Science. 2014;343:84–87. doi: 10.1126/science.1247005. [DOI] [PMC free article] [PubMed] [Google Scholar]
Shibata T, Kokubu A, Gotoh M, Ojima H, Ohta T, Yamamoto M, Hirohashi S. Genetic alteration of Keap1 confers constitutive Nrf2 activation and resistance to chemotherapy in gallbladder cancer. Gastroenterology. 2008a;135:1358–1368. doi: 10.1053/j.gastro.2008.06.082. [DOI] [PubMed] [Google Scholar]
Shibata T, Ohta T, Tong KI, Kokubu A, Odogawa R, Tsuta K, Asamura H, Yamamoto M, Hirohashi S. Cancer related mutations in NRF2 impair its recognition by Keap1-Cul3 E3 ligase and promote malignancy. PNAS. 2008b;105:13568–13573. doi: 10.1073/pnas.0806268105. [DOI] [PMC free article] [PubMed] [Google Scholar]
Sun Z, Huang Z, Zhang DD. Phosphorylation of Nrf2 at multiple sites by MAP kinases has a limited contribution in modulating the Nrf2-dependent antioxidant response. PLoS One. 2009;4:e6588. doi: 10.1371/journal.pone.0006588. [DOI] [PMC free article] [PubMed] [Google Scholar]
Sun C, Hobor S, Bertotti A, Zecchin D, Huang S, Galimi F, Cottino F, Prahallad A, Grernrum W, Tzani A, Schlicker A, Wessels LF, Smit EF, Thunnissen E, Halonen P, Lieftink C, Beijersbergen RL, Di Nicolantonio F, Bardelli A, Trusolino L, Bernards R. Intrinsic resistance to MEK inhibition in KRAS mutant lung and colon cancer through transcriptional induction of ERBB3. Cell Reports. 2014;7:86–93. doi: 10.1016/j.celrep.2014.02.045. [DOI] [PubMed] [Google Scholar]
Tang J, Salama R, Gadgeel SM, Sarkar FH, Ahmad A. Erlotinib resistance in lung Cancer: current progress and future perspectives. Frontiers in Pharmacology. 2013;4:15. doi: 10.3389/fphar.2013.00015. [DOI] [PMC free article] [PubMed] [Google Scholar]
Wagle N, Emery C, Berger MF, Davis MJ, Sawyer A, Pochanard P, Kehoe SM, Johannessen CM, Macconaill LE, Hahn WC, Meyerson M, Garraway LA. Dissecting therapeutic resistance to RAF inhibition in melanoma by tumor genomic profiling. Journal of Clinical Oncology. 2011;29:3085–3096. doi: 10.1200/JCO.2010.33.2312. [DOI] [PMC free article] [PubMed] [Google Scholar]
Wang XJ, Sun Z, Villeneuve NF, Zhang S, Zhao F, Li Y, Chen W, Yi X, Zheng W, Wondrak GT, Wong PK, Zhang DD. Nrf2 enhances resistance of cancer cells to chemotherapeutic drugs, the dark side of Nrf2. Carcinogenesis. 2008;29:1235–1243. doi: 10.1093/carcin/bgn095. [DOI] [PMC free article] [PubMed] [Google Scholar]
Wang T, Wei JJ, Sabatini DM, Lander ES. Genetic screens in human cells using the CRISPR-Cas9 system. Science. 2014;343:80–84. doi: 10.1126/science.1246981. [DOI] [PMC free article] [PubMed] [Google Scholar]
Whittaker SR, Theurillat JP, Van Allen E, Wagle N, Hsiao J, Cowley GS, Schadendorf D, Root DE, Garraway LA. A genome-scale RNA interference screen implicates NF1 loss in resistance to RAF inhibition. Cancer Discovery. 2013;3:350–362. doi: 10.1158/2159-8290.CD-12-0470. [DOI] [PMC free article] [PubMed] [Google Scholar]
Wilson FH, Johannessen CM, Piccioni F, Tamayo P, Kim JW, Van Allen EM, Corsello SM, Capelletti M, Calles A, Butaney M, Sharifnia T, Gabriel SB, Mesirov JP, Hahn WC, Engelman JA, Meyerson M, Root DE, Jänne PA, Garraway LA. A functional landscape of resistance to ALK inhibition in lung cancer. Cancer Cell. 2015;27:397–408. doi: 10.1016/j.ccell.2015.02.005. [DOI] [PMC free article] [PubMed] [Google Scholar]
Yuan M, Breitkopf SB, Yang X, Asara JM. A positive/negative ion-switching, targeted mass spectrometry-based metabolomics platform for bodily fluids, cells, and fresh and fixed tissue. Nature Protocols. 2012;7:872–881. doi: 10.1038/nprot.2012.024. [DOI] [PMC free article] [PubMed] [Google Scholar]
Zhang P, Singh A, Yegnasubramanian S, Esopi D, Kombairaju P, Bodas M, Wu H, Bova SG, Biswal S. Loss of Kelch-like ECH-associated protein 1 function in prostate cancer cells causes chemoresistance and radioresistance and promotes tumor growth. Molecular Cancer Therapeutics. 2010;9:336–346. doi: 10.1158/1535-7163.MCT-09-0589. [DOI] [PMC free article] [PubMed] [Google Scholar]

eLife. 2017 Feb 1;6:e18970. doi: 10.7554/eLife.18970.028

Decision letter

Editor: Martin McMahon¹

In the interests of transparency, eLife includes the editorial decision letter and accompanying author responses. A lightly edited version of the letter sent to the authors after peer review is shown, indicating the most substantive concerns; minor comments are not usually included.

Thank you for submitting your article "KEAP1 loss promotes resistance to BRAF, MEK, and EGFR inhibition in lung cancer" for consideration by eLife. Your article has been favorably evaluated by Tony Hunter (Senior Editor) and three reviewers, one of whom is a member of our Board of Reviewing Editors. The reviewers have opted to remain anonymous.

The reviewers have discussed the reviews with one another and the Reviewing Editor has drafted this decision to advise you on the additional analysis considered essential to publish this work in eLife. At this point, we request that you respond to the comments with a proposed set of experiments that you believe would satisfy the concerns of the reviewers and an estimate of the time it would take to complete this work. The editor and reviewers will consider your response and provide a binding recommendation.

Summary:

The article by Krall et al. describes the results of CRISPR/CAS9 based genetic screening in lung cancer cells to identify mechanisms of resistance to EGFR, BRAF or MEK inhibitor treatment. The authors identify loss of KEAP1 as a mechanism of this resistance. Mechanistically, they link KEAP1 loss to the suppression of reactive oxygen species (ROS) through the de-repression of the KEAP1 target, the transcription factor NRF2, which promotes expression of certain anti-oxidant genes. The authors contend that the induction of ROS is essential for the anti-proliferative effects of EGFR, BRAF or MEK inhibitors. The authors conclude that loss of KEAP1 therefore uncouples MEK inhibition and ROS induction to promote drug resistance. Conceptually, the findings are interesting, novel and potentially provide an important advance in the field, particularly as it pertains to understanding the mechanism and effectiveness of MEK inhibitors in KRAS or BRAF mutated lung cancer. However, there are several important issues that must to be experimentally addressed and/or clarified in the manuscript before this manuscript would be suitable for publication in eLife.

Essential revisions:

1) The authors' title indicates that CRISPR/CAS9-mediated KEAP1 silencing promotes resistance to BRAF, MEK, and EGFR inhibition in lung cancer. However, it is not clear whether loss of KEAP1 actually promotes drug resistance, (i.e. the capacity of KEAP(Null) cells to proliferate continuously in the presence of drug in vitro or to sustain tumor growth when transplanted into drug treated mice), or whether KEAP1 silencing simply attenuates the effects of the various inhibitors over the 14-21 day assay period without promoting long-term growth and tumorigenesis per se. For example, the authors' measure of drug resistance documents the effects of either KEAP1 silencing or ectopic NRF2 expression on "Relative Cell Number". CRISPR/CAS9 silencing of KEAP1 leads to a 1.5 to ~4 fold increase in "Relative Cell Number" but it is not clear that this reflects true drug resistance. The authors fail to provide any clonogenicity assays or to provide dose response curves of cell proliferation vs. drug concentration such that GI50 and/or GI90 values may be calculated. Moreover, the authors have not tested whether KEAP1 silencing can promote the growth of xenotransplanted tumors in appropriately drug treated mice, a point that all three reviewers agreed was a key piece of missing data. Hence, because effects of the various manipulations on "Relative Cell Number" were often very modest (from 50% – ~4 fold), there was substantial concern raised by all three reviewers that the authors are potentially overstating their observations with regard to long-term drug resistance vs. short-term attenuation of drug anti-proliferative effects, which substantially undermined enthusiasm for the research as presented and which would have to be addressed directly in any revised manuscript.

2) The authors state that, "We used the lowest concentration of drug that inhibited ERK phosphorylation and resulted in proliferative arrest or death" such that the concentrations of trametinib vary from cell line to cell line. Do the authors believe that different cells have different sensitivity thresholds for trametinib? Moreover, it is not clear that this is the relevant concentration for selection for drug resistance. Indeed, in the trametinib-treated HCC-364 cells there is clearly residual pERK1/2 present in lanes 4-6 of Figure 2A. Hence there is concern that the sub-maximal inhibition of BRAF(V600E)->MEK->ERK signaling may allow sufficient signal through this pathway to continue to promote cell proliferation. To that end, to what extent are the sgKEAP1 cells sensitive to an ERK1/2 inhibitor such as SCH772984? An additional concern is that the authors assessed the effects of the various pathway-targeted agents 48 hours post-addition, but their assay for drug resistance was at 14-21 days post drug treatment. Since many cancer cells display adaptive responses leading to ERK1/2 reactivation at late times after drug addition, how did the levels of pERK1/2 compare at 48 hours vs. 14 days post drug-addition? Finally, the authors use A549 cells, which developed as a KRAS mutated, KEAP1-Null cell line. To what extent are A549 cells resistant to trametinib and why is a concentration of 25 nM used in A549 cell experiments but 1.5 microM trametinib was used in KEAP1 proficient H1299 cells?

3) The authors claim to have knocked out KEAP1. What is the significance of the residual KEAP1 bands in Figure 2A,B, and Figure 1—figure supplement 2? Do these represent nonspecific bands, residual expression, degradation products or a sub-population of cells without KEAP1 KO? Is the mRNA expression of KEAP1 undetectable in KO cells? If indeed there is residual KEAP1 expression rather than complete KO, this needs to be addressed in the text. Related to this, the authors demonstrate that trametinib treatment of HCC364 cells leads to decreased KEAP1 and increased NRF2 expression (Figure 3B). Indeed, the magnitude of the effects of trametinib on KEAP1 and NRF2 expression appear equivalent to the effects of CRISPR/CAS9-mediated silencing of KEAP1 (compare KEAP1 and NRF2 expression in sgGFP vs. sgKEAP1-1 vs. sgKEAP1-2 in trametinib treated HCC364 cells in Figure 3B, lanes 16-18). Hence, if trametinib can promote the expression/activity of NRF2 without the need for CRISPR/CAS9-mediated silencing of KEAP1, should not these cells display some measure of drug resistance?

4) The authors demonstrate that trametinib treatment of HCC364 or CALU1 cells leads to a convincing increase in reactive oxygen species (ROS), which is prevented by CRISPR/CAS9-mediated silencing of KEAP1. They also claim that, "NAC reduced ROS and conferred resistance to trametinib (Figure 5A,B)". The same concern is raised with regard to these experiments as in #1 above: To what extent does NAC influence long-term clonogenicity and influence the dose response of cells to trametinib? In addition, do other MEK inhibitors have a similar effect in KRAS or BRAF mutant models? Does vemurafenib or erlotinib have a similar effect in an appropriate cellular context? Does expression of any NRF2 target genes such as Glutathione Peroxidase (GPX) promote resistance to any of the pathway-targeted agents in the appropriate cellular context? Finally, the effects of KEAP1 knockdown on cell growth and ROS metabolism in the absence of treatment (RAF, MEK inhibitors) is not shown and should be described?

5) Overall, the experiments presented by Krall et al. appear most relevant to primary response/resistance and not to the emergence of acquired resistance. Without any experimental evidence that the KEAP1-NRF2 pathway plays a functional role in acquired resistance, this aspect of the manuscript should be edited to ensure the focus is clearly on the initial treatment context so as not to confuse readers.

6) Access to patient biopsies before or after treatment with pathway-targeted inhibitors (EGFR or MEKi) to determine the presence of KEAP1 alterations would greatly strengthen the impact and clinical relevance of the study. For example, according to the TCGA analysis of lung adenocarcinoma, a percentage of EGFR/ERBB1 mutated lung cancers also carry alterations in either KEAP1 or NRF2. Do the authors have evidence to suggest that patients whose EGFR mutated lung tumors also have alterations in KEAP1 or NRF2 display primary resistance to EGFR blockade? If these are not available, the authors may want to test whether KEAP1 KO attenuates the effect of trametinib (given at MTD) on tumor growth in vivo.

7) The data in the EGFR mutated NSCLC cell lines is underdeveloped and, as it stands, distracts from the bulk of the work in the RAS or BRAF mutated cells.

The full reviews are appended below, in case there are any additional points that you wish to respond to, but this is not mandatory:

Reviewer #1:

The authors' title indicates that KEAP1 loss promotes resistance to BRAF, MEK, and EGFR inhibition in lung cancer. However, it is not clear whether loss of KEAP1 actually promotes drug resistance, i.e. the capacity of KEAP(Null) cells to proliferate continuously in the presence of drug in vitro or to form tumors when transplanted into drug treated mice or whether KEAP1 silencing simply provides the cells a short-term survival benefit measured over 14-21 days (Figure 1). For example, the authors' measure of drug resistance in Figures 1 and 2 documents the effects of either KEAP1 silencing or ectopic NRF2 expression on "Relative Cell Number". CRISPR/CAS9 silencing of KEAP1 leads to a 1.5 to ~4 fold increase in Relative Cell Number but it is not clear that this reflects true drug resistance. The authors fail to provide any clonogenicity assays or to provide dose response curves of cell proliferation vs. drug concentration such that GI50 and GI90 values may be calculated. Moreover, the authors have not tested whether KEAP1 silencing can promote the growth of xenotransplanted tumors in drug treated mice. Hence, there is substantial concern that the authors are overstating their observations with regard to long-term drug resistance vs. a short-term survival advantage, which fundamentally undermines the research as presented. Finally, the authors may wish to contrast the fold change of Relative Cell Number following KEAP1 silencing to the effects of a positive control such as the expression of p61-BRAF (V600E), which supports long-term drug resistance to vemurafenib in HCC364 BRAF mutated NSCLC cells.

The effects of NRF2(G31R) over-expression on Relative Cell Number are no greater than 2-fold and in the case of HCC364 cells are ~50% (Figure 2), which is less than that observed with CRISPR/CAS9-mediated silencing of KEAP1. Although the error bars indicate statistical significance, it is not clear that such modest differences would be biologically significant in terms of actual drug resistance per the comment above.

The authors state that, "We used the lowest concentration of drug that inhibited ERK phosphorylation and resulted in proliferative arrest or death". However, it is not clear that this is the relevant concentration for selection for drug resistance. Indeed, in the trametinib treated HCC-364 cells there is clearly residual pERK1/2 present in lanes 4-6 of Figure 2A. Hence there is concern that the sub-maximal inhibition of BRAF(V600E)->MEK->ERK signaling may allow sufficient signal through this pathway to continue to promote cell proliferation. To what extent are the sgKEAP1 cells sensitive to an ERK1/2 inhibitor such as SCH772984?

The authors demonstrate that Trametinib treatment of HCC364 cells leads to decreased KEAP1 expression and increased NRF2 expression (Figure 3B). Indeed, the magnitude of the effects of trametinib on KEAP1 and NRF2 expression appear equivalent to the effects of CRISPR/CAS9-mediated silencing of KEAP1 (compare KEAP1 and NRF2 expression in sgGFP vs. sgKEAP1-1 vs. sgKEAP1-2 in Trametinib treated HCC364 cells in Figure 3B, lanes 16-18). Hence, if trametinib can promote the expression/activity of NRF2 without the need for CRISPR/CAS9-mediated silencing of KEAP1, why are the cells not directly drug resistant?

The authors demonstrate that trametinib treatment of HCC364 or CALU1 cells leads to a convincing increase in reactive oxygen species (ROS), which is prevented by CRISPR/CAS9-mediated silencing of KEAP1. They also claim that, "NAC reduced ROS and conferred resistance to trametinib (Figure 5A,B)". The demonstration of NAC-mediated trametinib resistance was through a purported effect on relative population doublings of trametinib treated vs. trametinib/NAC treated cells (Figure 5B). This is not an adequate way in which to document drug resistance, which out to be measured using clonogenicity assays or to provide dose response curves of cell proliferation vs. drug treatment such that effects of NAC on the GI50 and GI90 values may be calculated.

According to the TCGA analysis of lung adenocarcinoma, a percentage of EGFR/ERBB1 mutated lung cancers also carry alterations in either KEAP1 or NRF2. Hence, do the authors have any evidence to suggest that patients whose EGFR mutated lung tumors also have alterations in KEAP1 or NRF2 display primary resistance to EGFR blockade?

Reviewer #2:

In this study Krall et al. performed CRISPR/CAS9 screens to determine genes that are required for sensitivity to inhibitors of RTK/RAF/MEK/ERK pathway in lung cancer. They made a very interesting observation that KEAP1 sgRNAs were enriched in screens utilizing trametinib, vemurafenib or erlotinib. They went on to show that KEAP1 KO attenuates the effect of trametinib, vemurafenib and erlotinib in BRAF, KRAS, EGFR models, respectively. From a mechanistic perspective, they show that trametinib induces ROS in cells with wt KEAP1 and, when KEAP1 is lost, increased expression of NRF2 and up regulation of glutathione result in lowered ROS species and decreased viability in the presence of trametinib. Conceptually, the findings are novel and provide an important advance in the field, particularly as it pertains to understanding the effectiveness of MEK inhibitors in KRAS mutant lung cancer. However, there are several issues that need to be experimentally addressed and/or clarified in the manuscript.

1) The authors claim to have knocked out KEAP1. What is the significance of the residual KEAP1 bands in Figure 2A,B, and Figure 1—figure supplement 2? Do these represent nonspecific bands, residual expression, degradation products or subpopulation of cells without KO? Is the mRNA expression of KEAP1 undetectable in KO cells? If indeed there is residual KEAP1 expression rather than complete KO, this needs to be addressed in the text.

2) The authors claim that KEAP1 KO causes resistance to trametinib. How are the authors defining resistance in this study? The data seem to show that KEAP1 KO has an attenuating effect, given that there is a significant antiproliferative effect even in the setting of KEAP1 KO. Also, a single concentration of trametinib was used in most studies. Dose response viability experiments with several trametinib concentrations in the presence or absence of KEAP1 are needed to clarify this issue.

3)KEAP1 sgRNAs scored across different screens/models. One worry is that KEAP1 has a housekeeper function in lung cancer cells. What is the effect of KEAP1 KO on cell viability in the absence of trametinib, vemurafenib or elrotinib treatment? How does this compare to the effect in the presence of trametinib?

4) The concentrations of trametinib vary from cell line to cell line. Do the authors believe that different cells have different sensitivity thresholds for trametinib? The author use A549 cells, which have endogenous loss of KEAP1. Are A549 cells resistant to trametinib? Why is a concentration of 25nM used in A549 but 1.5 μm in H1299?

5) The attenuation of trametinib effect by KEAP1 KO occurs in the absence of ERK reactivation. However, the viability assay was conducted at 21 days, whereas the effect of trametinib on pERK was tested after 48h. What is the status of ERK 21 days after treatment with trametinib in KEAP1 WT vs KO cells? How does this compare to the effect of trametinib 1h after inhibition?

6) The authors claim that resistance occurs because of an increase in NRF2 activity. Overexpression of an NRF2 mutant with defective KEAP1 binding attenuates effect of trametinib and vemurafenib when overexpressed in KEAP1 wt cells. Also, the expression of KEAP1 or a mutant with defective binding to NRF2 in A549 cells results in increased sensitivity to trametinib. These are valid studies, which support the mechanistic conclusions of the manuscript. The magnitude of the effect however is small (e.g. a 10% difference in A549 cells). A more direct approach, i.e. restoring the expression of KEAP1 vs. KEAP1 G333C or knocking out/down NRF2 in KEAP1 KO cells may yield more striking differences.

7) Is the effect of the KEAP1 P128L mutation known? The authors state that KEAP1 KO in these cells increased NRF2 expression. Was the magnitude of such increase similar to the magnitude of the increase observed in other cell lines with WT KEAP1?

8) Trametinib decreases KEAP1 expression, increases NRF2 protein levels and NRF2 dependent gene expression. Why is there an additive effect when combining trametinib with KO (Figure 3A)? Could trametinib regulate NRF2 independently of KEAP1?

9) Trametinib induces ROS. Do other MEKis have a similar effect in KRAS or BRAF mutant models? Does vemurafenib or erlotinib have a similar effect in appropriate cellular context?

10) Trametinib induces glutathione in KEAP1 KO cells. Is there an effect on ROS by KEAP1 KO in the absence of treatment (not clear if data in Figure 4A are normalized; there seems to be a small decrease in the supplementary companion).

11) Modulation of ROS by NAC or BSO attenuates or enhances, respectively, the effect of trametinib. In my opinion these are key experiments that bring together the mechanistic insights of the study and could potentially offer hints to optimize the performance of MEK inhibitors in KRAS mutant patients.

Some clarification is needed in the text and figure legends in order to help the reader better understand the experiments/findings.

Why are different concentrations of trametinib used in Figure 5C and D (50 nM and 10 nM, respectively). Lowering the concentration of trametinib would increase the likelihood that an additive effect on viability is observed. It would be important to show the effect of a fixed concentration of NAC/BSO on the trametinib dose response proliferation curve.

Also, does treatment with NAC or BSO modulate the effect of the other drugs tested in the study?

12) Access to patient biopsies before or after treatment with trametinib to determine the presence of KEAP1 alterations would greatly strengthen the impact and clinical relevance of the study. If these are not available, the authors may want to demonstrate that KEAP1 KO attenuates the effect of trametinib (given at MTD) on tumor growth in vivo.

Reviewer #3:

This article describes the results of CRISPR/Cas9 based genetic screening in lung cancer cells to identify mechanisms of resistance to RAF and MEK inhibitor treatment. The authors identify loss of Keap1 as a mechanism of this resistance. Mechanistically, they link Keap1 loss to the suppression of reactive oxygen species (ROS) through the de-repression of the Keap1 target gene NRF2. NRF2 promotes expression of certain anti-oxidant genes. The authors contend that the induction of ROS is important for the ability of MEK inhibition to suppress cell growth. The authors conclude that loss of Keap1 therefore uncouples MEK inhibition and ROS induction to promote resistance.

The studies are interesting, particularly the use of the CRISPR/Cas9 screening in multiple cell lines and the mechanistic connection to ROS and the Keap1-NRF2 axis. However, there are several issues that should be addressed to strengthen the authors' case and fully support their claims. First, the effects of RAF and MEK inhibition on Keap1 and NRF2 levels and function in the cell lines of interest are not shown systematically. Second, the effects of Keap1 knockdown on cell growth and ROS metabolism in the absence of treatment (RAF, MEK inhibitors) is not shown. Third, the data in the EGFR mutant cells is underdeveloped and as it stands distracts from the bulk of the work in the RAS mutant and BRAF mutant cells. Fourth, a limited number of cell lines is used to show the functional importance of either increased or decreased Keap1 and NRF2 in modulating RAF and MEK inhibitor response. This should be expanded. Fifth, no in vivo experiments are shown to corroborate the strength and potential translational relevance of the in vitro findings that are presented. This is especially important to add because most of the in vitro effects on resistance appear relatively modest. Sixth, functional experiments to show that a key metabolism enzyme contributes to the Keap1 knockdown phenotype are absent. Does genetic overexpression of GPX1, for instance, rescue the Keap1 knockdown phenotypes? Finally, the experiments shown deal primarily with primary response/resistance and not acquired resistance. Without any experimental evidence that this Keap1-NRF2 pathway plays a functional role in acquired resistance, this aspect of the manuscript should be edited to ensure the focus is more squarely on the initial treatment context so as not to confuse readers.

eLife. 2017 Feb 1;6:e18970. doi: 10.7554/eLife.18970.029

Author response

Essential revisions:

We agree with the reviewers that the effect observed after deleting KEAP1 is more modest than is sometimes seen when a secondary alteration leads to re-activation of the MAPK pathway to confer drug resistance. However, we believe that our findings are particularly interesting because activation of an independent pathway, rather than the MAPK pathway, is involved.

At the same time, we have performed additional experimentation to support the conclusion that deleting KEAP1 permits cells and tumors to grow in the presence of MAPK pathway inhibitors. Specifically, we now include dose-response curves for HCC827 (EGFR-mutant) sgGFP or sgKEAP1 cells treated with erlotinib (Figure 1—figure supplement 6). We found that the IC50 is indeed shifted in the polyclonal pool of sgKEAP1 cells. Furthermore, when single cell clones with complete KEAP1 knockout were derived, the IC50 of these clones was shifted even further, suggesting that in the initial experiments KEAP1 was not deleted in all of the cells in the population.

In addition, we have now included a xenograft experiment with HCC827 EGFR-mutant cells (Figure 6). HCC827 sgGFP or sgKEAP1 cells were implanted bilaterally into mice, and erlotinib treatment was initiated when the average tumor volume reached 100-200 cm³. We show that erlotinib has a cytostatic effect on sgGFP cells, with tumor size remaining stable during drug treatment. In contrast, sgKEAP1 tumors continue to grow larger during erlotinib treatment. Interestingly, sgKEAP1 cells also form larger tumors with vehicle treatment, indicating that loss of KEAP1 also confers a proliferative advantage in the absence of drug in vivo.

Taken together, these new experiments bolster the conclusion that loss of KEAP1 permits cells to proliferate both in vitro and in vivo in the presence of several clinically approved targeted agents, and we thank the reviewers for making these suggestions that have improved the manuscript.

In our experiments, we found that different cell lines require different concentrations of trametinib to achieve suppression of ERK phosphorylation and proliferative arrest. This likely reflects the differences in driver mutations (NRAS, KRAS, or BRAF) as well as the combination of additional genetic alterations in these different cell lines.

We believe that it is difficult to determine the “relevant” drug concentration for resistance. It remains unclear whether the MAPK pathway is fully or only partially suppressed in human tumors treated with MAPK pathway inhibitors. While the pathway may not be fully inhibited in our experiments, we are comparing cell proliferation with and without KEAP1 loss, and the extent of MAPK pathway inhibition is comparable in these isogenic systems. Therefore, we believe that the differences in proliferation are due to the differences in KEAP1 status, rather than sub-maximal inhibition of the MAPK pathway promoting proliferation.

We have assessed pERK levels at 14 days following drug addition, and the pathway remains suppressed. We have added this experiment as Figure 2—figure supplement 1B.

Regarding differing drug sensitivities in different cell lines, H1299 did require a higher concentration of trametinib to cause arrest, although suppression of ERK phosphorylation was observed at a lower concentration. We did not use these cells in any of the validation experiments for this reason. We have analyzed drug sensitivities in a large panel of cell lines (Barretina et al., Nature 2012 and Seashore-Ludlow et al., Cancer Discovery 2015), and found no correlation between KEAP1 status and trametinib sensitivity. This is likely due to the many other genetic and epigenetic differences between these cell lines. We believe our approach of comparing isogenic cell lines with and without KEAP1 (both knocking it out in cells with intact KEAP1 and restoring it in cells that are KEAP1-null) is the best approach to assessing the effect of KEAP1 on drug sensitivity.

We believe that the residual bands in Figure 2A,B and Figure 1—figure supplement 4 represent nonspecific bands. We have performed TIDE-PCR (Brinkman et al., Nucleic Acids Res 2014) on these cells and found the editing efficiency to be approximately 93%, of which approximately 70% is a 1 base pair insertion, which would lead to a frame shift. We also assessed mRNA expression using primers downstream of the sgRNA targeting sites. KEAP1 mRNA levels in KEAP1-KO cells were ~35-45% of that in control cells. This reduction in mRNA levels may reflect nonsense mediated decay of transcripts encoding premature stop codons as a consequence of CRISPR-Cas9 cutting. From the TIDE-PCR results, we expect that most of the remaining transcripts contain frame shift mutations. These experiments have been added to the Methods section of the revised manuscript.

While trametinib decreases KEAP1 expression and increases NRF2 expression, the combination of KEAP1-KO with trametinib treatment causes greater effects on KEAP1 levels, NRF2 levels, and NRF2 target gene expression. We have revised the text to state this observation more clearly (subsection “KEAP1^KO confers resistance through increased NRF2 activity”, third paragraph).

We have found that addition of NAC increases proliferation in the presence of drug, but we do not believe that it fully recapitulates KEAP1 loss. Addition of exogenous NAC to the culture media may not completely mimic the concentration and localization of glutathione that occurs upon KEAP1 loss. We have also shown that KEAP1 loss affects other aspects of cell metabolism, which NAC would not affect.

We believe that KEAP1 loss or overexpression of NRF2 alters expression of many genes, which together affect glutathione synthesis and cellular metabolism. Given the amplitude of the effect that KEAP1 loss or NRF2 overexpression have on proliferation in the presence of drug, we think it is unlikely that overexpressing a single target gene will promote resistance.

As to whether other inhibitors have a similar effect on ROS levels, we have included additional experiments in the revised manuscript showing that the EGFR inhibitor erlotinib and the ALK inhibitor LDK378 induce ROS in EGFR- and ALK-mutant cell lines, respectively, and that this increase is prevented by KEAP1 loss (Figure 4—figure supplement 1A).

The effects of KEAP1 knockout on ROS levels in the absence of drug are shown in Figure 4A (compare DMSO-treated sgGFP to DMSO-treated sgKEAP1). KEAP1 loss does not affect cell proliferation in the absence of drug.

While this manuscript was under revision, Gainor et al. (Cancer Discovery 2016) identified a NRF2 mutation in a patient with acquired resistance to an ALK inhibitor. This mutation (E79Q) is in a mutational hotspot and has previously been shown to impair recognition of NRF2 by KEAP1 (Shibata et al., PNAS 2008), thus activating the pathway. This tumor also had a secondary ALK mutation of unknown function and eventually became resistant to a second generation ALK inhibitor. We believe that the NRF2 mutation may have provided a means for such tumor cells to survive the initial crizotinib treatment and acquire additional resistance mutations over time. We therefore believe that our findings may be relevant for both primary and acquired resistance. We have described this finding in the Discussion section of the revised manuscript and have also clarified that further studies will be necessary to establish the frequency of KEAP1 loss as a resistance mechanism in patients treated with RTK/MAPK pathway inhibitors. We have also modified the title and Abstract to better describe how KEAP1 loss or NRF2 mutations affect the response of cells to these targeted inhibitors.

We agree that it seems reasonable to determine the presence of NRF2/KEAP1 alterations in patient biopsies before and after treatment with targeted inhibitors. Unfortunately, we are limited by the number of available samples. We have looked at patient data with two clinical collaborators who are leaders in leading such trials and have collected and followed different cohorts of patients as well as with a major pharmaceutical company. Despite these joint studies, the number of patient samples available for these analyses are quite small and do not preclude that such mutations occur in a fraction of treated patients.

We note (see response to comment #5), that Gainor et al., recently identified a patient that harbors a NRF2 mutation in a drug-resistant tumor but not the pre-treatment tumor. As the reviewers commented above, we believe our results may also be relevant to primary/intrinsic resistance. Although RTK/Ras alterations can co-occur with NRF2/KEAP1 alterations, the number of such patients is small, as is the number of patients enrolled in clinical trials. Thus, there is not sufficient statistical power to assess for a correlation between NRF2/KEAP1 status and drug response. At this point, a sufficient number of patients have not yet been studied to assess this point directly. As more patients are treated with RTK/MAPK pathway inhibitors, it will be extremely interesting to look for such a correlation.

As described above, the revised manuscript now includes an in vivo tumor xenograft experiment (Figure 6), showing that EGFR-mutant tumors with intact KEAP1 fail to grow in the presence of erlotinib treatment, while KEAP1-KO tumors continue to grow larger during treatment.

7) The data in the EGFR mutated NSCLC cell lines is underdeveloped and, as it stands, distracts from the bulk of the work in the RAS or BRAF mutated cells.

While our original submitted manuscript focused on MEK inhibition in KRAS and BRAF mutant cells, we subsequently found that KEAP1 loss also conferred resistance to EGFR and ALK inhibition. We believe the observation that KEAP1 loss confers resistance to multiple targeted therapies in lung cancer is a very important finding. We have added additional CRISPR-Cas9 screens showing that KEAP1 loss confers resistance to EGFR and ALK inhibitors, and that it does so by a similar mechanism to the one identified in trametinib-treated KRAS and BRAF mutant cells.

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Supplementary file 1. Genome scale CRISPR-Cas9 screens in CALU1, HCC364, and NCIH1299 cells.

For each sgRNA, the average log2 fold-change (LFC) between drug-treated samples and Day 0 samples is shown. The hits from each screen, as determined by LFC or by STARS score, are also summarized.

DOI: http://dx.doi.org/10.7554/eLife.18970.025

elife-18970-supp1.xlsx^{(30MB, xlsx)}

DOI: 10.7554/eLife.18970.025

Supplementary file 2. Focused sgRNA library screens in HCC827 and MGH065 cells.

Raw counts and Z-scores are shown for each sgRNA in DMSO- or drug-treated samples.

DOI: http://dx.doi.org/10.7554/eLife.18970.026

elife-18970-supp2.xlsx^{(9.1MB, xlsx)}

DOI: 10.7554/eLife.18970.026

Supplementary file 3. Metabolite Profiling of KEAP1 knockout cells.

DOI: http://dx.doi.org/10.7554/eLife.18970.027

elife-18970-supp3.xlsx^{(563.1KB, xlsx)}

DOI: 10.7554/eLife.18970.027

[bib1] Aguirre AJ, Meyers RM, Weir BA, Vazquez F, Zhang CZ, Ben-David U, Cook A, Ha G, Harrington WF, Doshi MB, Kost-Alimova M, Gill S, Xu H, Ali LD, Jiang G, Pantel S, Lee Y, Goodale A, Cherniack AD, Oh C, Kryukov G, Cowley GS, Garraway LA, Stegmaier K, Roberts CW, Golub TR, Meyerson M, Root DE, Tsherniak A, Hahn WC. Genomic copy number dictates a Gene-Independent cell response to CRISPR/Cas9 Targeting. Cancer Discovery. 2016;6:914–929. doi: 10.1158/2159-8290.CD-16-0154. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib2] Balak MN, Gong Y, Riely GJ, Somwar R, Li AR, Zakowski MF, Chiang A, Yang G, Ouerfelli O, Kris MG, Ladanyi M, Miller VA, Pao W. Novel D761Y and common secondary T790M mutations in epidermal growth factor receptor-mutant lung adenocarcinomas with acquired resistance to kinase inhibitors. Clinical Cancer Research. 2006;12:6494–6501. doi: 10.1158/1078-0432.CCR-06-1570. [DOI] [PubMed] [Google Scholar]

[bib3] Barretina J, Caponigro G, Stransky N, Venkatesan K, Margolin AA, Kim S, Wilson CJ, Lehár J, Kryukov GV, Sonkin D, Reddy A, Liu M, Murray L, Berger MF, Monahan JE, Morais P, Meltzer J, Korejwa A, Jané-Valbuena J, Mapa FA, Thibault J, Bric-Furlong E, Raman P, Shipway A, Engels IH, Cheng J, Yu GK, Yu J, Aspesi P, de Silva M, Jagtap K, Jones MD, Wang L, Hatton C, Palescandolo E, Gupta S, Mahan S, Sougnez C, Onofrio RC, Liefeld T, MacConaill L, Winckler W, Reich M, Li N, Mesirov JP, Gabriel SB, Getz G, Ardlie K, Chan V, Myer VE, Weber BL, Porter J, Warmuth M, Finan P, Harris JL, Meyerson M, Golub TR, Morrissey MP, Sellers WR, Schlegel R, Garraway LA. The cancer cell line Encyclopedia enables predictive modelling of anticancer drug sensitivity. Nature. 2012;483:603–607. doi: 10.1038/nature11003. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib4] Berns K, Horlings HM, Hennessy BT, Madiredjo M, Hijmans EM, Beelen K, Linn SC, Gonzalez-Angulo AM, Stemke-Hale K, Hauptmann M, Beijersbergen RL, Mills GB, van de Vijver MJ, Bernards R. A functional genetic approach identifies the PI3K pathway as a major determinant of trastuzumab resistance in breast cancer. Cancer Cell. 2007;12:395–402. doi: 10.1016/j.ccr.2007.08.030. [DOI] [PubMed] [Google Scholar]

[bib5] Blumenschein GR, Smit EF, Planchard D, Kim DW, Cadranel J, De Pas T, Dunphy F, Udud K, Ahn MJ, Hanna NH, Kim JH, Mazieres J, Kim SW, Baas P, Rappold E, Redhu S, Puski A, Wu FS, Jänne PA. A randomized phase II study of the MEK1/MEK2 inhibitor trametinib (GSK1120212) compared with docetaxel in KRAS-mutant advanced non-small-cell lung cancer (NSCLC)†. Annals of Oncology. 2015;26:894–901. doi: 10.1093/annonc/mdv072. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib6] Brinkman EK, Chen T, Amendola M, van Steensel B. Easy quantitative assessment of genome editing by sequence trace decomposition. Nucleic Acids Research. 2014;42:e168. doi: 10.1093/nar/gku936. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib7] Cancer Genome Atlas Research Network Comprehensive molecular profiling of lung adenocarcinoma. Nature. 2014;511:543–550. doi: 10.1038/nature13385. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib8] Cerami E, Gao J, Dogrusoz U, Gross BE, Sumer SO, Aksoy BA, Jacobsen A, Byrne CJ, Heuer ML, Larsson E, Antipin Y, Reva B, Goldberg AP, Sander C, Schultz N. The cBio cancer genomics portal: an open platform for exploring multidimensional cancer genomics data. Cancer Discovery. 2012;2:401–404. doi: 10.1158/2159-8290.CD-12-0095. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib9] Chen B, Gilbert LA, Cimini BA, Schnitzbauer J, Zhang W, Li GW, Park J, Blackburn EH, Weissman JS, Qi LS, Huang B. Dynamic imaging of genomic loci in living human cells by an optimized CRISPR/Cas system. Cell. 2013;155:1479–1491. doi: 10.1016/j.cell.2013.12.001. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib10] Chio II, Jafarnejad SM, Ponz-Sarvise M, Park Y, Rivera K, Palm W, Wilson J, Sangar V, Hao Y, Öhlund D, Wright K, Filippini D, Lee EJ, Da Silva B, Schoepfer C, Wilkinson JE, Buscaglia JM, DeNicola GM, Tiriac H, Hammell M, Crawford HC, Schmidt EE, Thompson CB, Pappin DJ, Sonenberg N, Tuveson DA. NRF2 promotes tumor maintenance by modulating mRNA translation in pancreatic cancer. Cell. 2016;166:963–976. doi: 10.1016/j.cell.2016.06.056. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib11] Corcoran RB, Ebi H, Turke AB, Coffee EM, Nishino M, Cogdill AP, Brown RD, Della Pelle P, Dias-Santagata D, Hung KE, Flaherty KT, Piris A, Wargo JA, Settleman J, Mino-Kenudson M, Engelman JA. EGFR-mediated re-activation of MAPK signaling contributes to insensitivity of BRAF mutant colorectal cancers to RAF inhibition with vemurafenib. Cancer Discovery. 2012;2:227–235. doi: 10.1158/2159-8290.CD-11-0341. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib12] Corcoran RB, Cheng KA, Hata AN, Faber AC, Ebi H, Coffee EM, Greninger P, Brown RD, Godfrey JT, Cohoon TJ, Song Y, Lifshits E, Hung KE, Shioda T, Dias-Santagata D, Singh A, Settleman J, Benes CH, Mino-Kenudson M, Wong KK, Engelman JA. Synthetic lethal interaction of combined BCL-XL and MEK inhibition promotes tumor regressions in KRAS mutant cancer models. Cancer Cell. 2013;23:121–128. doi: 10.1016/j.ccr.2012.11.007. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib13] Crystal AS, Shaw AT, Sequist LV, Friboulet L, Niederst MJ, Lockerman EL, Frias RL, Gainor JF, Amzallag A, Greninger P, Lee D, Kalsy A, Gomez-Caraballo M, Elamine L, Howe E, Hur W, Lifshits E, Robinson HE, Katayama R, Faber AC, Awad MM, Ramaswamy S, Mino-Kenudson M, Iafrate AJ, Benes CH, Engelman JA. Patient-derived models of acquired resistance can identify effective drug combinations for cancer. Science. 2014;346:1480–1486. doi: 10.1126/science.1254721. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib14] DeNicola GM, Karreth FA, Humpton TJ, Gopinathan A, Wei C, Frese K, Mangal D, Yu KH, Yeo CJ, Calhoun ES, Scrimieri F, Winter JM, Hruban RH, Iacobuzio-Donahue C, Kern SE, Blair IA, Tuveson DA. Oncogene-induced Nrf2 transcription promotes ROS detoxification and tumorigenesis. Nature. 2011;475:106–109. doi: 10.1038/nature10189. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib15] DeNicola GM, Chen PH, Mullarky E, Sudderth JA, Hu Z, Wu D, Tang H, Xie Y, Asara JM, Huffman KE, Wistuba II, Minna JD, DeBerardinis RJ, Cantley LC. NRF2 regulates serine biosynthesis in non-small cell lung cancer. Nature Genetics. 2015;47:1475–1481. doi: 10.1038/ng.3421. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib16] Ding L, Getz G, Wheeler DA, Mardis ER, McLellan MD, Cibulskis K, Sougnez C, Greulich H, Muzny DM, Morgan MB, Fulton L, Fulton RS, Zhang Q, Wendl MC, Lawrence MS, Larson DE, Chen K, Dooling DJ, Sabo A, Hawes AC, Shen H, Jhangiani SN, Lewis LR, Hall O, Zhu Y, Mathew T, Ren Y, Yao J, Scherer SE, Clerc K, Metcalf GA, Ng B, Milosavljevic A, Gonzalez-Garay ML, Osborne JR, Meyer R, Shi X, Tang Y, Koboldt DC, Lin L, Abbott R, Miner TL, Pohl C, Fewell G, Haipek C, Schmidt H, Dunford-Shore BH, Kraja A, Crosby SD, Sawyer CS, Vickery T, Sander S, Robinson J, Winckler W, Baldwin J, Chirieac LR, Dutt A, Fennell T, Hanna M, Johnson BE, Onofrio RC, Thomas RK, Tonon G, Weir BA, Zhao X, Ziaugra L, Zody MC, Giordano T, Orringer MB, Roth JA, Spitz MR, Wistuba II, Ozenberger B, Good PJ, Chang AC, Beer DG, Watson MA, Ladanyi M, Broderick S, Yoshizawa A, Travis WD, Pao W, Province MA, Weinstock GM, Varmus HE, Gabriel SB, Lander ES, Gibbs RA, Meyerson M, Wilson RK. Somatic mutations affect key pathways in lung adenocarcinoma. Nature. 2008;455:1069–1075. doi: 10.1038/nature07423. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib17] Doench JG, Hartenian E, Graham DB, Tothova Z, Hegde M, Smith I, Sullender M, Ebert BL, Xavier RJ, Root DE. Rational design of highly active sgRNAs for CRISPR-Cas9-mediated gene inactivation. Nature Biotechnology. 2014;32:1262–1267. doi: 10.1038/nbt.3026. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib18] Doench JG, Fusi N, Sullender M, Hegde M, Vaimberg EW, Donovan KF, Smith I, Tothova Z, Wilen C, Orchard R, Virgin HW, Listgarten J, Root DE. Optimized sgRNA design to maximize activity and minimize off-target effects of CRISPR-Cas9. Nature Biotechnology. 2016;34:184–191. doi: 10.1038/nbt.3437. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib19] Engelman JA, Zejnullahu K, Mitsudomi T, Song Y, Hyland C, Park JO, Lindeman N, Gale CM, Zhao X, Christensen J, Kosaka T, Holmes AJ, Rogers AM, Cappuzzo F, Mok T, Lee C, Johnson BE, Cantley LC, Jänne PA. MET amplification leads to gefitinib resistance in lung cancer by activating ERBB3 signaling. Science. 2007;316:1039–1043. doi: 10.1126/science.1141478. [DOI] [PubMed] [Google Scholar]

[bib20] Forbes SA, Beare D, Gunasekaran P, Leung K, Bindal N, Boutselakis H, Ding M, Bamford S, Cole C, Ward S, Kok CY, Jia M, De T, Teague JW, Stratton MR, McDermott U, Campbell PJ. COSMIC: exploring the world's knowledge of somatic mutations in human cancer. Nucleic Acids Research. 2015;43:D805–D811. doi: 10.1093/nar/gku1075. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib21] Gainor JF, Dardaei L, Yoda S, Friboulet L, Leshchiner I, Katayama R, Dagogo-Jack I, Gadgeel S, Schultz K, Singh M, Chin E, Parks M, Lee D, DiCecca RH, Lockerman E, Huynh T, Logan J, Ritterhouse LL, Le LP, Muniappan A, Digumarthy S, Channick C, Keyes C, Getz G, Dias-Santagata D, Heist RS, Lennerz J, Sequist LV, Benes CH, Iafrate AJ, Mino-Kenudson M, Engelman JA, Shaw AT. Molecular mechanisms of resistance to first- and Second-Generation ALK Inhibitors in ALK-Rearranged Lung Cancer. Cancer Discovery. 2016;6:1118–1133. doi: 10.1158/2159-8290.CD-16-0596. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib22] Gao J, Aksoy BA, Dogrusoz U, Dresdner G, Gross B, Sumer SO, Sun Y, Jacobsen A, Sinha R, Larsson E, Cerami E, Sander C, Schultz N. Integrative analysis of complex Cancer genomics and clinical profiles using the cBioPortal. Science Signaling. 2013;6:pl1. doi: 10.1126/scisignal.2004088. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib23] Hayes JD, Dinkova-Kostova AT. The Nrf2 regulatory network provides an interface between redox and intermediary metabolism. Trends in Biochemical Sciences. 2014;39:199–218. doi: 10.1016/j.tibs.2014.02.002. [DOI] [PubMed] [Google Scholar]

[bib24] Hyman DM, Puzanov I, Subbiah V, Faris JE, Chau I, Blay JY, Wolf J, Raje NS, Diamond EL, Hollebecque A, Gervais R, Elez-Fernandez ME, Italiano A, Hofheinz RD, Hidalgo M, Chan E, Schuler M, Lasserre SF, Makrutzki M, Sirzen F, Veronese ML, Tabernero J, Baselga J. Vemurafenib in multiple nonmelanoma cancers with BRAF V600 mutations. New England Journal of Medicine. 2015;373:726–736. doi: 10.1056/NEJMoa1502309. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib25] Imielinski M, Berger AH, Hammerman PS, Hernandez B, Pugh TJ, Hodis E, Cho J, Suh J, Capelletti M, Sivachenko A, Sougnez C, Auclair D, Lawrence MS, Stojanov P, Cibulskis K, Choi K, de Waal L, Sharifnia T, Brooks A, Greulich H, Banerji S, Zander T, Seidel D, Leenders F, Ansén S, Ludwig C, Engel-Riedel W, Stoelben E, Wolf J, Goparju C, Thompson K, Winckler W, Kwiatkowski D, Johnson BE, Jänne PA, Miller VA, Pao W, Travis WD, Pass HI, Gabriel SB, Lander ES, Thomas RK, Garraway LA, Getz G, Meyerson M. Mapping the hallmarks of lung adenocarcinoma with massively parallel sequencing. Cell. 2012;150:1107–1120. doi: 10.1016/j.cell.2012.08.029. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib26] Johannessen CM, Boehm JS, Kim SY, Thomas SR, Wardwell L, Johnson LA, Emery CM, Stransky N, Cogdill AP, Barretina J, Caponigro G, Hieronymus H, Murray RR, Salehi-Ashtiani K, Hill DE, Vidal M, Zhao JJ, Yang X, Alkan O, Kim S, Harris JL, Wilson CJ, Myer VE, Finan PM, Root DE, Roberts TM, Golub T, Flaherty KT, Dummer R, Weber BL, Sellers WR, Schlegel R, Wargo JA, Hahn WC, Garraway LA. COT drives resistance to RAF inhibition through MAP kinase pathway reactivation. Nature. 2010;468:968–972. doi: 10.1038/nature09627. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib27] Johannessen CM, Johnson LA, Piccioni F, Townes A, Frederick DT, Donahue MK, Narayan R, Flaherty KT, Wargo JA, Root DE, Garraway LA. A melanocyte lineage program confers resistance to MAP kinase pathway inhibition. Nature. 2013;504:138–142. doi: 10.1038/nature12688. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib28] Kim E, Ilic N, Shrestha Y, Zou L, Kamburov A, Zhu C, Yang X, Lubonja R, Tran N, Nguyen C, Lawrence MS, Piccioni F, Bagul M, Doench JG, Chouinard CR, Wu X, Hogstrom L, Natoli T, Tamayo P, Horn H, Corsello SM, Lage K, Root DE, Subramanian A, Golub TR, Getz G, Boehm JS, Hahn WC. Systematic functional interrogation of rare cancer variants identifies oncogenic Alleles. Cancer Discovery. 2016;6:714–726. doi: 10.1158/2159-8290.CD-16-0160. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib29] Kobayashi A, Kang MI, Okawa H, Ohtsuji M, Zenke Y, Chiba T, Igarashi K, Yamamoto M. Oxidative stress sensor Keap1 functions as an adaptor for Cul3-based E3 ligase to regulate proteasomal degradation of Nrf2. Molecular and Cellular Biology. 2004;24:7130–7139. doi: 10.1128/MCB.24.16.7130-7139.2004. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib30] Kosaka T, Yatabe Y, Endoh H, Yoshida K, Hida T, Tsuboi M, Tada H, Kuwano H, Mitsudomi T. Analysis of epidermal growth factor receptor gene mutation in patients with non-small cell lung cancer and acquired resistance to gefitinib. Clinical Cancer Research. 2006;12:5764–5769. doi: 10.1158/1078-0432.CCR-06-0714. [DOI] [PubMed] [Google Scholar]

[bib31] Manchado E, Weissmueller S, Morris JP, Chen CC, Wullenkord R, Lujambio A, de Stanchina E, Poirier JT, Gainor JF, Corcoran RB, Engelman JA, Rudin CM, Rosen N, Lowe SW. A combinatorial strategy for treating KRAS-mutant lung cancer. Nature. 2016;534:647–651. doi: 10.1038/nature18600. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib32] Mitsuishi Y, Taguchi K, Kawatani Y, Shibata T, Nukiwa T, Aburatani H, Yamamoto M, Motohashi H. Nrf2 redirects glucose and glutamine into anabolic pathways in metabolic reprogramming. Cancer Cell. 2012;22:66–79. doi: 10.1016/j.ccr.2012.05.016. [DOI] [PubMed] [Google Scholar]

[bib33] Munoz DM, Cassiani PJ, Li L, Billy E, Korn JM, Jones MD, Golji J, Ruddy DA, Yu K, McAllister G, DeWeck A, Abramowski D, Wan J, Shirley MD, Neshat SY, Rakiec D, de Beaumont R, Weber O, Kauffmann A, McDonald ER, Keen N, Hofmann F, Sellers WR, Schmelzle T, Stegmeier F, Schlabach MR. CRISPR screens provide a comprehensive assessment of cancer vulnerabilities but generate False-Positive hits for highly amplified genomic regions. Cancer Discovery. 2016;6:900–913. doi: 10.1158/2159-8290.CD-16-0178. [DOI] [PubMed] [Google Scholar]

[bib34] Nazarian R, Shi H, Wang Q, Kong X, Koya RC, Lee H, Chen Z, Lee MK, Attar N, Sazegar H, Chodon T, Nelson SF, McArthur G, Sosman JA, Ribas A, Lo RS. Melanomas acquire resistance to B-RAF(V600E) inhibition by RTK or N-RAS upregulation. Nature. 2010;468:973–977. doi: 10.1038/nature09626. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib35] Ohta T, Iijima K, Miyamoto M, Nakahara I, Tanaka H, Ohtsuji M, Suzuki T, Kobayashi A, Yokota J, Sakiyama T, Shibata T, Yamamoto M, Hirohashi S. Loss of Keap1 function activates Nrf2 and provides advantages for lung cancer cell growth. Cancer Research. 2008;68:1303–1309. doi: 10.1158/0008-5472.CAN-07-5003. [DOI] [PubMed] [Google Scholar]

[bib36] Prahallad A, Sun C, Huang S, Di Nicolantonio F, Salazar R, Zecchin D, Beijersbergen RL, Bardelli A, Bernards R. Unresponsiveness of colon cancer to BRAF(V600E) inhibition through feedback activation of EGFR. Nature. 2012;483:100–103. doi: 10.1038/nature10868. [DOI] [PubMed] [Google Scholar]

[bib37] Rudin CM, Hong K, Streit M. Molecular characterization of acquired resistance to the BRAF inhibitor dabrafenib in a patient with BRAF-mutant non-small-cell lung cancer. Journal of Thoracic Oncology. 2013;8:e41–e42. doi: 10.1097/JTO.0b013e31828bb1b3. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib38] Sanjana NE, Shalem O, Zhang F. Improved vectors and genome-wide libraries for CRISPR screening. Nature Methods. 2014;11:783–784. doi: 10.1038/nmeth.3047. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib39] Shalem O, Sanjana NE, Hartenian E, Shi X, Scott DA, Mikkelsen TS, Heckl D, Ebert BL, Root DE, Doench JG, Zhang F. Genome-scale CRISPR-Cas9 knockout screening in human cells. Science. 2014;343:84–87. doi: 10.1126/science.1247005. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib40] Shibata T, Kokubu A, Gotoh M, Ojima H, Ohta T, Yamamoto M, Hirohashi S. Genetic alteration of Keap1 confers constitutive Nrf2 activation and resistance to chemotherapy in gallbladder cancer. Gastroenterology. 2008a;135:1358–1368. doi: 10.1053/j.gastro.2008.06.082. [DOI] [PubMed] [Google Scholar]

[bib41] Shibata T, Ohta T, Tong KI, Kokubu A, Odogawa R, Tsuta K, Asamura H, Yamamoto M, Hirohashi S. Cancer related mutations in NRF2 impair its recognition by Keap1-Cul3 E3 ligase and promote malignancy. PNAS. 2008b;105:13568–13573. doi: 10.1073/pnas.0806268105. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib42] Sun Z, Huang Z, Zhang DD. Phosphorylation of Nrf2 at multiple sites by MAP kinases has a limited contribution in modulating the Nrf2-dependent antioxidant response. PLoS One. 2009;4:e6588. doi: 10.1371/journal.pone.0006588. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib43] Sun C, Hobor S, Bertotti A, Zecchin D, Huang S, Galimi F, Cottino F, Prahallad A, Grernrum W, Tzani A, Schlicker A, Wessels LF, Smit EF, Thunnissen E, Halonen P, Lieftink C, Beijersbergen RL, Di Nicolantonio F, Bardelli A, Trusolino L, Bernards R. Intrinsic resistance to MEK inhibition in KRAS mutant lung and colon cancer through transcriptional induction of ERBB3. Cell Reports. 2014;7:86–93. doi: 10.1016/j.celrep.2014.02.045. [DOI] [PubMed] [Google Scholar]

[bib44] Tang J, Salama R, Gadgeel SM, Sarkar FH, Ahmad A. Erlotinib resistance in lung Cancer: current progress and future perspectives. Frontiers in Pharmacology. 2013;4:15. doi: 10.3389/fphar.2013.00015. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib45] Wagle N, Emery C, Berger MF, Davis MJ, Sawyer A, Pochanard P, Kehoe SM, Johannessen CM, Macconaill LE, Hahn WC, Meyerson M, Garraway LA. Dissecting therapeutic resistance to RAF inhibition in melanoma by tumor genomic profiling. Journal of Clinical Oncology. 2011;29:3085–3096. doi: 10.1200/JCO.2010.33.2312. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib46] Wang XJ, Sun Z, Villeneuve NF, Zhang S, Zhao F, Li Y, Chen W, Yi X, Zheng W, Wondrak GT, Wong PK, Zhang DD. Nrf2 enhances resistance of cancer cells to chemotherapeutic drugs, the dark side of Nrf2. Carcinogenesis. 2008;29:1235–1243. doi: 10.1093/carcin/bgn095. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib47] Wang T, Wei JJ, Sabatini DM, Lander ES. Genetic screens in human cells using the CRISPR-Cas9 system. Science. 2014;343:80–84. doi: 10.1126/science.1246981. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib48] Whittaker SR, Theurillat JP, Van Allen E, Wagle N, Hsiao J, Cowley GS, Schadendorf D, Root DE, Garraway LA. A genome-scale RNA interference screen implicates NF1 loss in resistance to RAF inhibition. Cancer Discovery. 2013;3:350–362. doi: 10.1158/2159-8290.CD-12-0470. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib49] Wilson FH, Johannessen CM, Piccioni F, Tamayo P, Kim JW, Van Allen EM, Corsello SM, Capelletti M, Calles A, Butaney M, Sharifnia T, Gabriel SB, Mesirov JP, Hahn WC, Engelman JA, Meyerson M, Root DE, Jänne PA, Garraway LA. A functional landscape of resistance to ALK inhibition in lung cancer. Cancer Cell. 2015;27:397–408. doi: 10.1016/j.ccell.2015.02.005. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib50] Yuan M, Breitkopf SB, Yang X, Asara JM. A positive/negative ion-switching, targeted mass spectrometry-based metabolomics platform for bodily fluids, cells, and fresh and fixed tissue. Nature Protocols. 2012;7:872–881. doi: 10.1038/nprot.2012.024. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib51] Zhang P, Singh A, Yegnasubramanian S, Esopi D, Kombairaju P, Bodas M, Wu H, Bova SG, Biswal S. Loss of Kelch-like ECH-associated protein 1 function in prostate cancer cells causes chemoresistance and radioresistance and promotes tumor growth. Molecular Cancer Therapeutics. 2010;9:336–346. doi: 10.1158/1535-7163.MCT-09-0589. [DOI] [PMC free article] [PubMed] [Google Scholar]

PERMALINK

KEAP1 loss modulates sensitivity to kinase targeted therapy in lung cancer

Elsa B Krall

Belinda Wang

Diana M Munoz

Nina Ilic

Srivatsan Raghavan

Matthew J Niederst

Kristine Yu

David A Ruddy

Andrew J Aguirre

Jong Wook Kim

Amanda J Redig

Justin F Gainor

Juliet A Williams

John M Asara

John G Doench

Pasi A Janne

Alice T Shaw

Robert E McDonald III

Jeffrey A Engelman

Frank Stegmeier

Michael R Schlabach

William C Hahn

Roles

Abstract

Introduction

Results

CRISPR-Cas9 screens identify genes whose loss confers resistance to MEK and BRAF inhibition

Figure 1. CRISPR-Cas9 genome scale drug resistance screens and validation that KEAP1-KO confers resistance.

Figure 1—figure supplement 1. Optimization of genome scale screening conditions.

Figure 1—figure supplement 2. Optimization of focused sgRNA library screening conditions.

Figure 1—figure supplement 3. EGFR and ALK inhibitor resistance screens.

Figure 1—figure supplement 4. Immunoblots showing deletion of KEAP1 by sgRNAs in HCC364, CALU1, MGH-065, and HCC827.

Figure 1—figure supplement 5. Crystal violet assays showing that KEAP1 loss confers resistance to afatinib and crizotinib.

Figure 1—figure supplement 6. Erlotinib dose-response curves in HCC827.

KEAP1KO confers resistance through increased NRF2 activity

Figure 2. KEAP1-KO alters NRF2 levels.

Figure 2—figure supplement 1. KEAP1-KO alters NRF2 levels and does not activate ERK.

Figure 2—figure supplement 2. NRF2 is necessary and sufficient for resistance.

Figure 3. Trametinib treatment and KEAP1-KO increase NRF2 activity.

Figure 3—figure supplement 1. Expression of NFE2L2/NRF2 mRNA and NRF2 target genes in (A) CALU1 cells, (B) HCC827 cells, or (C) MGH-065 cells treated with DMSO or the indicated drug for 72 hr.

Figure 3—figure supplement 2. CALU1 cells treated with DMSO or trametinib for 72 hr.

Loss of KEAP1 promotes survival by increasing glutathione synthesis and decreasing drug-induced ROS

Figure 4. KEAP1-KO reduces trametinib-induced ROS through increased glutathione synthesis.

Figure 4—figure supplement 1. KEAP1-KO reduces drug-induced ROS.

Figure 4—figure supplement 2. Glutathione levels in HCC364 and CALU1 cells.

Figure 4—figure supplement 3. KEAP1-KO alters expression of SHMT1/2 in some cell lines.

Figure 4—figure supplement 4. KEAP1-KO alters expression of GPX1 and GSR.

Figure 5. Altering ROS levels affects cell survival.

Figure 5—figure supplement 1. KEAP1-KO reduces ROS and increases viability in the presence of BSO.

Figure 5—figure supplement 2. Expression of WT KEAP1 but not G333C in KEAP1-null A549 cells increases trametinib- and BSO-induced ROS.

Figure 5—figure supplement 3. KEAP1-KO alters cell metabolism.

Figure 6. Alterations in the KEAP1/NRF2 pathway alter tumor growth in vivo.

Discussion

Materials and methods

Cell lines and reagents

Screen optimization for genome scale screens

GeCKO v2 library construction

Genome-scale CRISPR knockout drug resistance screens with GeCKO v2 library

Sequencing and analysis of genome-scale screens

Screen optimization for focused sgRNA library screen

Focused sgRNA library construction

Focused sgRNA library screens

Vectors

sgKEAP1 arrayed infection

TIDE-PCR

qRT-PCR

Cytoplasmic/nuclear fractionation

Immunoblotting

Crystal violet assays

ORF expression

DCFDA assays to measure ROS

Metabolite profiling

Xenotransplantation and in vivo drug treatments

qPCR primer sequences

sgRNA sequences

Acknowledgements

Funding Statement

Funding Information

KEAP1^KO confers resistance through increased NRF2 activity