The Endosomal Protein CEMIP Links WNT Signaling to MEK1–ERK1/2

Published OnlineFirst June 18, 2018; DOI: 10.1158/0008-5472.CAN-17-3149

Cancer Molecular Cell Biology Research

The Endosomal Protein CEMIP Links WNT Signaling to MEK1–ERK1/2 Activation in Selumetinib-Resistant Intestinal Organoids Hong Quan Duong1,2,3,4, Ivan Nemazanyy5, Florian Rambow6, Seng Chuan Tang1,2, Sylvain Delaunay1,7, Lars Tharun8, Alexandra Florin8, Reinhard Buttner€ 8, Daniel Vandaele9, Pierre Close1,7, Jean-Christophe Marine6, Kateryna Shostak1,2, and Alain Chariot1,2,10

Abstract

MAPK signaling pathways are constitutively active in decrease of both ERK1/2 signaling and c-Myc. Together, our colon cancer and also promote acquired resistance to MEK1 data identify a cross-talk between Wnt and MAPK signaling inhibition. Here, we demonstrate that BRAFV600E-mutated cascades, which involves CEMIP. Activation of this pathway colorectal cancers acquire resistance to MEK1 inhibition by promotes survival by potentially regulating levels of speciﬁc inducing expression of the scaffold protein CEMIP through amino acids via a Myc-associated cascade. Targeting this a b-catenin– and FRA-1–dependent pathway. CEMIP was node may provide a promising avenue for treatment of found in endosomes and bound MEK1 to sustain ERK1/2 colon cancers that have acquired resistance to targeted activation in MEK1 inhibitor–resistant BRAFV600E-mutated therapies. colorectal cancers. The CEMIP-dependent pathway main- tained c-Myc protein levels through ERK1/2 and provided Signiﬁcance: MEK1 inhibitor–resistant colorectal cancer metabolic advantage in resistant cells, potentially by sus- relies on the scaffold and endosomal protein CEMIP to main- taining amino acids synthesis. CEMIP silencing circum- tain ERK1/2 signaling and Myc-driven transcription. Cancer Res; vented resistance to MEK1 inhibition, partly, through a 78(16); 4533–48. 2018 AACR.

Introduction underlying genetic alterations are loss-of-function mutations of the adenomatous polyposis coli (APC) gene, which leads to Colorectal cancer is the second leading cause of death from b-catenin activation and constitutive Wnt signaling, followed by cancer in Western countries and arises from a variety of genetic gain-of-function mutations in KRAS or BRAF proto-oncogenes alterations that result in the constitutive activation of both Wnt- (1). RAS signals though the RAF Ser/Thr kinase family and triggers and ErbB-dependent oncogenic signaling pathways. Among the the subsequent activation of the mitogen-activated protein/extracellular signal–regulated kinase 1 and 2 (MEK1/2) as well as the extracellular signal–regulated kinase 1 and 2 (ERK1/2). This 1Interdisciplinary Cluster for Applied Genoproteomics (GIGA), GIGA-Molecular signaling cascade gained signiﬁcant attention due to the high Biology of Diseases, University of Liege, CHU, Sart-Tilman, Liege, Belgium. frequency of KRAS and BRAF mutations found in human cancers 2 Laboratory of Medical Chemistry, University of Liege, CHU, Sart-Tilman, Liège, (2, 3). Indeed, activating mutations of KRAS are found in 40% of Belgium. 3Institute of Research and Development, Duy Tan University, Quang 4 advanced colorectal cancer (4). In addition, the BRAF valine 600 Trung, Danang, Vietnam. Department of Cancer Research, Vinmec Research V600E Institute of Stem Cell and Gene Technology, Hanoi, Vietnam. 5Paris Descartes (BRAF ) mutation, which leads to constitutive activation of University, Sorbonne Paris Cite, Paris, France. 6Laboratory for Molecular Cancer BRAF, is found in approximately 11% of colorectal cancers and Biology, VIB Center for Cancer Biology and KULeuven Department of Oncology, confers poor prognosis (5–7). As the pharmacologic inhibition of Leuven, Belgium. 7Institute for Pathology, University Hospital Cologne, Cologne, KRAS remains challenging, alternative approaches targeting 8 Germany. Laboratory of Cancer Signaling, University of Liege, Liege, Belgium. downstream RAS effectors (RAF and MEK1) have been proposed 9 Gastroenterology Department, University of Liege, CHU, Sart-Tilman, Liege, but were poorly effective in monotherapy for the treatment of Belgium. 10Walloon Excellence in Life Sciences and Biotechnology (WELBIO), Wallonia, Belgium. colorectal cancer, largely because of a feedback reactivation of MAPK signaling (8, 9). This reactivation occurs through the Note: Supplementary data for this article are available at Cancer Research ampliﬁcation of the driving oncogene KRAS or BRAF in colorectal Online (http://cancerres.aacrjournals.org/). cells treated with MEK1 inhibitors (10, 11). Other mechanisms K. Shostak and A. Chariot contributed equally to this article. involve the EGFR/HER1–dependent reactivation of MAPK in Corresponding Author: Alain Chariot, Laboratory of Medical Chemistry, GIGA BRAFV600E-mutated colorectal cancer cells treated with a BRAF Molecular Biology of Diseases, Tour GIGA, þ2 B34, Sart-Tilman, University of inhibitor (12, 13). Similarly, MAPK reactivation in KRAS-mutated Liege, Liege 4000, Belgium. Phone: 32-043662472; Fax: 32-043664534; colorectal cancer cells subjected to MEK1 inhibition also results E-mail: [email protected]. from the induction of HER3 (14). Clinical trials in which RAF and doi: 10.1158/0008-5472.CAN-17-3149 EGFR or RAF and MEK are cotargeted to suppress the feedback 2018 American Association for Cancer Research. reactivation of MAPK signaling were carried out but patients

www.aacrjournals.org 4533

Duong et al.

showing initial benefit nevertheless developed resistance and (20 ng/mL), Noggin (100 ng/mL), and R-Spondin (500 ng/mL) recurrence in disease progression (15). Here again, resistant was added every 2 days. Apc-mutated organoids were cultured colorectal cancer cells had KRAS or BRAF amplification as well in DMEM/F12 supplemented with EGF (20 ng/mL), Noggin þ as an activating MEK1 mutation (16). (100 ng/mL) without R-Spondin. The enrichment of Lgr5 stem The RAS–RAF–MEK1–ERK1/2 cascade is critically reliant on cells in ex vivo organoids generated with intestinal crypts from scaffold proteins, which assemble pathway molecules to regulate C57BL/6 mice was carried out by treating them with a combina- signaling. Among them are Ras GTPase-activating-like protein tion of valproic acid (1 mmol/L) and CHIR999021 (3 mmol/L), a (IQGAP1) as well as kinase suppressor of RAS (KSR; refs. 17–20). GSK3 inhibitor. Another scaffold protein is KIAA1199, now referred to as CEMIP ("Cell Migration-inducing and hyaluronan-binding protein"), Generation of selumetinib-resistant colorectal cancer cell lines whose expression is enhanced in cervical, breast, and colorectal (HT-29/SR, COLO-205/SR, SW480/SR and HCT116/SR) and cancer (21–25). CEMIP promotes cell survival and invasion, at selumetinib-resistant ex vivo organoids least through EGFR-dependent MEK1 and ERK1/2 activation in Four colorectal cancer cell lines (HT-29, COLO-205, SW480, cervical and breast cancer cells (23). It remains unclear which and HCT116) were used as parental cell lines (HT-29/P, COLO- scaffold proteins, if any, are specifically involved in MAPK reac- 205/P, SW480/P, and HCT116/P), from which were generated the tivation in colorectal cells showing intrinsic or acquired resistance selumetinib-resistant cell lines (HT-29/SR, COLO-205/SR, to BRAF or MEK1 inhibitors. SW480/SR, and HCT116/SR). These cell lines were generated by It is intuitive that both Wnt- and MAPK-dependent signaling repeated subculturing cells in the presence of incrementally in- pathways are interconnected in promoting resistance to targeted creasing concentrations of selumetinib (from 0.05 to 1.5 mmol/L therapies. Here we define CEMIP as a MEK1-binding protein for HT-29/P and SW480/P cells; from 0.05 to 2 mmol/L for induced by Wnt signaling. CEMIP promotes the acquired resis- HCT116/P cells; from 0.005 to 0.3 mmol/L for COLO-205/P) for tance to MEK1 inhibition in BRAFV600E-mutated colorectal cancer 6 months. For the maintenance of selumetinib-resistant colorectal cells, at least through ERK1/2 signaling and Myc. This CEMIP- cancer cell lines, the maximum concentration of selumetinib, dependent cascade is essential for amino acid synthesis in resis- namely 1.5 mmol/L (HT-29/SR and SW480/SR cells), 2 mmol/L tant cells. Collectively, our data define CEMIP as a key driver of (HCT116/SR cells), and 0.3 mmol/L selumetinib (COLO-205/SR resistance to MEK1 inhibition in BRAFV600E-mutated colorectal cells) was added into the normal medium. cancer that acts upstream of ERK1/2 and Myc cascade. For the generation of selumetinib-resistant ex vivo organoids, þ organoids generated from Apc /Min mice were first cultured with Materials and Methods 1 mmol/L of selumetinib for two weeks. The concentration was then increased by 0.5 mmol/L every two weeks to reach a final Cell culture and reagents concentration of 5 mmol/L. Colorectal cancer cell lines (HT-29, HCT116, SW480 and COLO-205) were purchased from ATCC in 2009. These cells were Lentiviral cell infection characterized by ATCC, using a comprehensive database of short Control shRNA, CEMIP, Myc, TAK1 and FRA-1 shRNA lentiviral fi tandem repeat (STR) DNA pro les. Frozen aliquots of freshly pLKO1-puro plasmid constructs were purchased from Sigma. cultured cells were generated and experiments were done with Control shRNA and CEMIP shRNA lentiviral pLKO1-puro- resuscitated cells cultured for less than 6 months. All cell lines IPTG-inducible plasmid constructs were also purchased from were mycoplasma tested. HT-29 and HCT116 cells were cultured Sigma. Lentiviral infections were carried out as described previ- in McCoy 5A supplemented with 10% heat-inactivated FBS ously (23). (HI-FBS; Gibco, Life Technologies) and 100 U/mL penicillin/ For lentiviral infections of ex vivo organoids, they were manu- streptomycin. SW480 cells were cultured in DMEM supplemented ally disrupted, washed with PBS to eliminate debris, and subse- with 10% HI-FBS, 1% glutamine, and 100 U/mL penicillin/ quently trypsinized for 30 minutes at 37C. After washing with streptomycin. COLO-205 cells were cultured in RPMI1640 sup- PBS, cells were washed via strainers (70 mmol/L) with 20 mL of plemented with 10% HI-FBS, 1% glutamine, and 100 U/mL PBS and centrifuged (200 g) for 5 minutes at 4C. They were penicillin/streptomycin. Selumetinib (AZD6244), vemurafenib then diluted in 500 mL of full media for organoid growth and (PLX4032, RG7204), PD98059, and PNU-74654 were from 500 mL of infectious supernatants were added, mixed, and incu- Selleck Chemicals. bated in a CO2 incubator for 12 hours. Organoids were subsequently centrifuged for 5 minutes at 4C, washed once with 1 mL Intestinal epithelial cell extraction and ex vivo organoid of PBS, and plated as usual. 24 hours later, full media containing cultures 2 mg/mL of puromycin was added. Intestines and colons were extracted from C57BL/6 (Wnt OFF) þ or Apc /Min (Wnt ON) mice. All our studies were approved by the Quantitative real-time PCR Institutional Animal Care and Use Committee of the University of Total RNAs were extracted using the E.Z.N.A Total RNA kit Liege (Liege, Belgium). Bowels were washed for 10 minutes at (Promega). cDNAs were synthesized using the Revert aid H 37C in a PBS-DTT (1 mmol/L) buffer and then incubated for 15 minus reverse transcriptase kit (Thermo Scientific) and real- minutes at 37C in a HBSS-EDTA buffer (30 mmol/L). Cells were time PCR analyses were performed as described previously harvested, washed twice in PBS, and flashed frozen. For the (23). mRNA levels in control organoids or cells were set to 1 generation of ex vivo organoid cultures, small pieces of intestine and mRNA levels in other experimental conditions were rela- were incubated in 2 mmol/L EDTA-PBS for 30 minutes at 4C. tive to the control after normalization with b-actin. Data from Crypts were extracted, washed twice in PBS, and cultured in at least two independent experiments performed in triplicates Matrigel (BD Biosciences). DMEM/F12 supplemented with EGF are shown.

4534 Cancer Res; 78(16) August 15, 2018 Cancer Research

CEMIP Links Wnt and MEK1 Signaling Pathways

MTS assay plus 1% Triton X-100 to give 35%, 30%, 20%, and 5% (w/v) Cells counted using the TC20 Automated Cell Counter iodixanol. A total of 0.6 mL of each sample as well as the four (Bio-Rad) were plated in 96-well flat-bottom plates at a density gradient solutions were layered in tubes for the swinging- of 2,000 cells per well in triplicate and then treated with various bucket rotor. Samples were centrifuged at 30,000 rpm for 16 concentrations of selumetinib or vemurafenib for 72 hours. Cell hours. Fractions of equal volume were collected for subsequent viability was determined using the MTS assay reagent (CellTiter Western blot analyses. 96 AQueous One Solution Cell Proliferation Assay; Promega) For the second fractionation experiment, the Optiprep Density according to the manufacturer's protocol. The absorbance was Gradient centrifugation kit was purchased from Sigma Aldrich. measured at 490 nm using a Wallac Victor2 1420 Multilabel Briefly, about 300 million HT-29 cells showing some acquired counter (Perkin Elmer). Absorbance of untreated cells was des- resistance to selumetinib were trypsinized, washed in PBS, and ignated as 100% and the number of viable cells in other exper- centrifuged at 600 g for 5 minutes. Cells were then lysed with the imental conditions was relative to the untreated cells. extraction buffer, homogenized using the Dounce homogenizer, and centrifuged at 1,000 g for 10 minutes. The supernatant was Clonogenic assay collected and centrifuged at 20,000 g for 30 minutes. The pellet Cells were seeded in 60-cm dishes at a density of 3,000 cells per (which includes ER, lysosomes, peroxisomes, mitochondria, and dish in duplicate. Twenty-four hours after plating, various con- endosomes) was diluted to a 19% Optiprep Density gradient centrations of selumetinib or vemurafenib were added to each solution and centrifuged on an OptiPrep Density Gradient at dish. After treatment for 24 hours, cells were washed with PBS and 100,000 g for 8 hours. Fractions of equal volume were collected further incubated for 15 days. Cells were subsequently stained for subsequent Western blot analyses. with 0.5% crystal violet in 25% methanol-containing PBS. Col- onies were examined under a light microscope and counted after Immunoprecipitation capturing images. Anti-MEK1, -BRAF and -IgG (negative control) antibodies were first coupled noncovalently to a mixture of Protein A/G- Western blot analysis Sepharose. The antibody-Protein A/G-Sepharose conjugates Cells were lysed in a buffer containing 20 mmol/L Tris-HCl, were then pelleted by centrifugation at 5,000 rpm for 2 min- 0.5 mol/L NaCl, Triton X-100, 1 mmol/L EDTA, 1 mmol/L EGTA, utes, the supernatant removed, and the beads washed with 0.1 10 mmol/L b-glycophosphate, 10 mmol/L NaF, 300 mmol/L mol/L sodium borate pH 9.3. This was repeated four times, Na3VO4, 1 mmol/L benzamidine, 2 mmol/L PMSF, and 1 mmol/L after which the beads were resuspended in 20 mmol/L dimethyl DTT. Western blots were carried out as described, using antibodies pimelimidate dihydrochloride (DMP) freshly made in 0.1 mol/ listed in Supplementary Table S1 (23). L sodium borate pH 9.3 and gently mixed on a rotating wheel for 30 minutes at room temperature. Following centrifugation Caspase-3/7 activity assay at 5,000 rpm for 2 minutes, supernatant was removed and fresh Caspase-3/7 activity was quantified using the Caspase-3/7 Glo 20 mmol/L DMP/0.1 mol/L sodium borate pH 9.3 solution was Assay (Promega). Cells were treated with selumetinib or vemur- added to the beads, which were then gently mixed for a further afenib for the indicated periods of time and caspase-3/7 activity 20 minutes. The beads were then spun down at 5,000 rpm for was quantified from cell lysates. Luminescence was measured at 2 minutes, the supernatant removed, and four washes with 490 nm using Wallac Victor2 1420 Multilabel counter (Perkin 50 mmol/L glycine pH 2.5 carried out to remove any antibody Elmer). Luminescence values in vehicle-treated control samples coupled noncovalently. Afterwards, the beads were washed were set to 1 and values obtained in other experimental condi- twice with 0.2 mol/L Tris-HCl pH 8.0 (neutralization step) tions were relative to the vehicle control. and then resuspended in the same solution and mixed gently on a rotating wheel at room temperature for 2 hours. The beads Extraction of cytoplasmic and nuclear proteins were then used immediately for immunoprecipitation analyses Cells were incubated on ice for 10 minutes in cytoplasmic lysis as described previously (23). buffer (10 mmol/L HEPES pH 7.9, 10 mmol/L KCl, 0.1 mmol/L EDTA, NP-40 0.3%, and Protease inhibitor). After centrifugation Kinase assay at 3,000 rpm for 5 minutes at 4C, the supernatant fraction Control or CEMIP-depleted HT29 cells showing acquired resis- (cytoplasmic extract) was harvested and the pellet was resus- tance to selumetinib were subjected to anti-MEK1 or -IgG (neg- pended in nuclear lysis buffer (20 mmol/L HEPES pH 7.9, ative control) immunoprecipitation. Selumetinib was added as 0.4 mol/L NaCl, 1 mmol/L EDTA, Glycerol 25%, and Protease control in some experimental conditions to inhibit MEK1 activity. inhibitor), incubated on ice for 10 minutes, and centrifuged at The kinase assay was conducted at 30C for 30 minutes with 1 mg 14,000 rpm for 5 minutes at 4C and supernatant containing the of GST-ERK2 substrate (Thermo Fisher Scientific), 10 mCi of [g32P] nuclear fraction was retained. ATP in 20 mL of kinase buffer (25 mmol/L HEPES, pH 7.5, 10 mmol/L MgCl2, 25 mmol/L b-glycerophosphate, 1 mmol/L Biochemical fractionation Na3VO4, and 1 mmol/L dithiothreitol). ERK2 phosphorylation Cells were resuspended and homogenized in a Dounce was revealed by autoradiography. homogenizer with the lysis buffer (150 mmol/L NaCl, 5 mmol/L DTT, 5 mmol/L EDTA, 25 mmol/L Tris HCl pH7.4, Proximal ligation and chromatin immunoprecipitation assays protease inhibitors) and centrifuged at 1,000 g for 10 min- Parental and resistant HT-29 cells or control and CEMIP- utes at 4C. The supernatant was adjusted to 1% Triton X-100 depleted resistant HT-29 cells were plated in 8-chamber cell and left on ice for 30 minutes. 4 vol of OptiPrep were added to culture dishes and proximal ligation assays were performed as 2 vol of supernatant. OptiPrep was diluted with the lysis buffer described previously (23).

www.aacrjournals.org Cancer Res; 78(16) August 15, 2018 4535

Duong et al.

Chromatin immunoprecipitation (ChIP) was performed minutes at 4C. They were then diluted in 500 mLofPBSand using anti-TCF4 or IgG antibody as a negative control. A incubated for 40 minutes on ice with anti-mouse CD24-PE TCF4-binding site (site #1 located 839 bp upstream of the eBioscience and anti-mouse CD133-APC eBioscience antibo- transcriptional start site on the CEMIP promoter) was identified dies, washed once in PBS, and analyzed on the FACSCantoII. through in silico analysis (MatInspector, Genomatix). TCF- DAPI staining was used for selection of live cells. binding sites #2, #3, and #4 located 26417, 75344, and 79348 bp downstream of the transcriptional start site within Tumor xenograft experiments intron 1 respectively, were described previously (26). A nega- HT-29/SR cells were infected with an IPTG-inducible CEMIP tive binding site was randomly chosen in exon 1 of the CEMIP shRNA (HT-29/SR-iCEMIP shRNA) or control shRNA (HT-29/SR- sequence. Primer sequences used are available upon request. iControl shRNA). HT-29/SR-iCEMIP shRNA or HT-29/SR-iCon- Extracts from selumetinib-resistant HT-29 cells, left untreated trol shRNA cells (1.5 106) mixed with Matrigel at a ratio of or treated with 3 mmol/L selumetinib for 24 hours were pre- 1:1 were injected subcutaneously into the right or left flank of cleared through an incubation step with protein A/BSA/Herring 6- to 8-week-old NOD/SCID male mice, respectively. Mice were sperm DNA for 18 hours and immunoprecipitations were monitored and tumors tracked via caliper measurements. Tumor performed overnight at 4 C with the relevant antibody, fol- volume was determined using the following formula: length lowed by 1-hour incubation with protein A/BSA/Herring sperm width height 0.5236 (n ¼ 4 mice/group). Mice were treated – DNA. Protein DNA complexes were washed three times with with a combination of 10 mmol/L IPTG in their drinking water high salt buffer (1% Triton X-100, 0.1% SDS, 500 mmol/L and 25 mmol/L IPTG (200 mL) via intraperitoneal injection and NaCl, 2 mmol/L EDTA pH 8.0, 20 mmol/L Tris HCl pH 8.0, either untreated or treated with selumetinib (20 mg/kg) via with protease inhibitors) and once with LiCl buffer (20 mmol/L intraperitoneal injection five days a week for two weeks. Mice Tris HCl pH 8.0, 1 mmol/L EDTA, 250 mmol/L LiCl, 0.5% NP- from all treatment groups were euthanized and tumors were 40, 0.5% sodium deoxycholate, 0.5 mmol/L PMSF, and pro- excised and tissue archived for immunofluorescence and real- tease inhibitors). After elution, proteinase K treatment, and time PCR analysis. reversal of crosslinks, DNA fragments were analyzed by real- time PCR with SYBR Green detection. Values were calculated as fi ratios between ChIP signals obtained with the anti-TCF4 (spe- Establishment of metabolomic pro les by targeted cific) or IgG (nonspecific) antibodies. Input DNA was analyzed metabolomics simultaneously and used for normalization purposes. For targeted metabolomics analysis of ex vivo organoids, each sample was washed three times with cold PBS, collected into an Immunofluorescence Eppendorf tube, frozen in liquid nitrogen, and stored at 80 C Immunofluorescence on cells was carried out as described until extraction. The extraction solution used was 50% meth- previously (23). For immunofluorescence on ex vivo organoids, anol, 30% ACN, and 20% water. The volume of extraction 6 they were grown in 8-well chamber slides (Thermo Fisher Scien- solution added was calculated from the cell count (2 10 cells fi fi per mL). After addition of extraction solution, samples were ti c, Lab-TekTM), xed in 4% paraformaldehyde for 15 minutes, and washed twice in PBS. They were then incubated in permea- vortexed for 5 minutes at 4 C, and immediately centrifuged at bilization solution (PBS with 0.5% Triton X-100) for 15 minutes, 16,000 g for 15 minutes at 4 C. The supernatants were – washed in PBS, and incubated for 60 minutes in PBS containing collected and analyzed by liquid chromatography mass spec- 0.2% Triton X-100, 0.05% Tween, and 1% BSA. Organoids were trometry using SeQuant ZIC-pHilic column (Merck) for the incubated overnight at 4C with primary antibody in the same liquid chromatography separation. Mobile phase A consisted of solution without Triton. After incubation, organoids were washed 20 mmol/L ammonium carbonate plus 0.1% ammonia hydrox- fl three times in PBS and incubated for 40 minutes with secondary ide in water. Mobile phase B consisted of ACN. The ow rate antibody, washed three times with PBS and incubated for 5 was kept at 100 mL/minute, and the gradient was 0 minutes, minutes with DAPI solution, and mounted with the ProLong 80% of B; 30 minutes, 20% of B; 31 minutes, 80% of B; and Gold Antifade Mountant from Invitrogen. The anti-Ki-67 mouse 45 minutes, 80% of B. The mass spectrometer (QExactive fi mAb was from BD Biosciences. Orbitrap, Thermo Fisher Scienti c) was operated in a polarity fi For immunofluorescence on cells expressing the SNAP- switching mode and metabolites were identi ed using Trace- fi CEMIP construct, the CEMIP coding sequence was subcloned Finder Software (Thermo Fisher Scienti c). To obtain a robust statistical analysis, metabolomics data were normalized using into the pSNAPf Vector (New England BioLabs) in which the SNAP tag is localized at the C-terminal end. HCT116 cells were the median normalization method (27). The data were further transfected with controls (empty vector and pSNAPf -Cox8A) preprocessed with a log transformation. MetaboAnalyst 3.0 and CEMIP-SNAP plasmids and labeled with SNAP-Cell software (28) was used to conduct statistical analysis and TMR-Star (New England BioLabs) 24 hours after the transfec- heatmap generation, and anunpairedtwo-samplet test was tion. Cell were fixed with 4% PAF and costained with endo- chosen to perform the comparisons. The algorithm for heatmap somal or ER markers, as described previously (23). DAPI was clustering was based on the Pearson distance measure for used for nuclei staining. similarity and the Ward linkage method for biotype clustering. Metabolites with similar abundance patterns were positioned FACS analysis closer together. Organoids were manually disrupted, washed with PBS to eliminate debris, and subsequently trypsinized for 30 minutes Statistical analysis at 37C. After washing with PBS, cells were washed via strainers The two-tailed Student t test was applied for statistical analysis (70 meters) with 20 mL of PBS and centrifuged (200 g)for5 when only two groups of interest were compared. Results were

4536 Cancer Res; 78(16) August 15, 2018 Cancer Research

CEMIP Links Wnt and MEK1 Signaling Pathways

plotted as mean SD and were significant in all experiments at ErbB2, KRAS, or MEK1 was lost upon CEMIP deficiency (Fig. 2C P < 0.001 (), P < 0.01 (), and P < 0.05 (). and D). Genes controlled by the transcription factor LEF1 were also identified to be regulated by CEMIP (Fig. 2C; Supplementary Results Fig. S2A). We next carried out an iRegulon analysis to identify all genes coexpressed with CEMIP in colon adenocarcinoma and Ex vivo organoids with some acquired resistance to MEK1 found 285 candidates (Fig. 2E). Interestingly, many of them are inhibition show enhanced CEMIP expression regulated by the Myc family of transcription factors (Fig. 2E). We Cancer stem cells contribute toresistancetotargetedther- also carried out an Ingenuity analysis (31) on these 285 coex- apies. To make the link between cancer stem cells and acquired pressed genes and found a significant enrichment of genes con- resistance to targeted therapies, we subjected ex vivo organoids trolled by p53, Myc, and b-catenin among others (Supplementary showing constitutive Wnt signaling (i.e., generated with þ Fig. S2B). Therefore, CEMIP expression is linked to ErbB-, Myc-, extracts of intestinal crypts from Apc /Min mice) to increasing and b-catenin–dependent pathways. concentrations of selumetinib to generate resistant organoids (Fig. 1A). The maintenance of these ex vivo organoids relies on CEMIP promotes Myc expression through ERK1/2 activation the self-renewal potential of cancer stem cells. Resistant orga- To explore how CEMIP and Myc are linked, we assessed the noids were larger in size but without changes in cell prolifer- þ consequences of CEMIP deficiency in ex vivo organoids. The ation (as judged by the percentage of Ki-67 cells) and were depletion of CEMIP in parental or selumetinib-resistant orga- protected from caspase-3–dependent cell death, compared noids severely impaired their maintenance (Supplementary Fig. with parental organoids (Figs. 1A and B). The scaffold protein þ S3; Fig. 3A, respectively). Consistently, the pool of CD24 / CEMIP may connect protumorigenic Wnt- and MAPK signaling þ CD133 cancer stem cells was impaired upon CEMIP deficiency pathways as it is the most robust Wnt-induced gene candidate (Fig. 3B). CEMIP deficiency downregulated HER3, Cyclin D1, and is also required for MAPK activation upon activation of Cyclin D2, Myc, SOX9, and phosphorylated ERK1/2 levels (Fig. ErbB signaling (23, 29). As such, CEMIP may actively contrib- 3C). To explore whether the link between CEMIP and Myc was ute to acquired resistance to selumetinib as a signaling protein found in other experimental systems, we generated BRAFV600E- involved in MAPK reactivation. Ex vivo organoids treated with a mutated COLO205 cells with some acquired resistance to MEK1 combination of valproic acid and CHIR999021, a GSK3 inhib- inhibition by subjecting parental cells to increasing concentra- itor,toinduceWntsignalingshowedelevatedmRNAlevelsof tions of selumetinib (Fig. 3D). Selumetinib-resistant COLO205 Wnt target genes such as Lgr5 and CEMIP while the level of cells showed elevated CEMIP mRNA and protein levels as well as Dclk1, a marker of differentiated Tuft cells, was downregulated increased levels of pMEK1/2, pERK1/2, and pRSK1 (Fig. 3E and F). (Fig. 1C). CEMIP induction by these drugs was also detected at Importantly, CEMIP deficiency in these cells also impaired MEK1 the protein level (Fig. 1C). Immunofluorescence confirmed and ERK1/2 activation and decreased protein levels of Myc that treatment with valproic acid and CHIR999021, which þ (Fig. 3G). Of note, effects on MEK1 activation were largely due enriches Lgr5 cells in ex vivo organoids, decreased the number þ to decreased total levels of MEK1 in CEMIP-deficient cells, which of Dclk1 Tuft cells (Fig. 1C). Therefore, CEMIP expression was not the case in ex vivo organoids (Fig. 3G and C, respectively). is transcriptionally induced by Wnt signaling. Importantly, These selumetinib-resistant COLO205 cells were also resistant to resistant organoids showed increased CEMIP, SOX9, HER3, PD98059, another MEK1 inhibitor, as pERK1/2 levels barely and BRAF expression as well as enhanced activation of MEK1, decreased at high concentrations of this inhibitor (Supplementary ERK1/2, and mTOR (as judged by 4EBP1 phosphorylation; Fig. Fig. S4). Here also, CEMIP deficiency in these cells decreased 1D and E, respectively). CEMIP was actually increased at the pMEK1, pERK1/2, as well as Myc protein levels (Supplementary mRNA level in resistant organoids (Supplementary Fig. S1). Fig. S4). Conversely, the ectopic expression of CEMIP alone in Moreover, Myc, which controls protein synthesis and organ DLD-1 cells enhanced both pERK1/2 and Myc protein levels size, was increased at the protein but not mRNA levels without impacting on Myc mRNA levels and also protected from (Fig. 1D; Supplementary Fig. S1, respectively). Of note, we cell death triggered by selumetinib (Supplementary Figs. S5A and did not detect any mutation on BRAF or on MEK1 in these S5B, respectively). Therefore, CEMIP maintains Myc protein levels resistant organoids. Resistant organoids were enriched in þ þ in multiple experimental models. CD24 /CD133 cancer stem cells (Fig. 1F), which fits with the upregulation of CD133 in colon cancer cells showing CEMIP expression is induced through BRAF, ERK1/2, and FRA-1 hyperactivation of the RAS–RAF–MEK1 cascade (30). There- upon acquired resistance in BRAFV600E- but not KRASG13D or fore, selumetinib-resistant organoids show all molecular fea- G12A-mutated colorectal cancer cells tures classically associated with the acquired resistance to CEMIP expression is increased in ex vivo organoids as well as in MEK1 inhibition. BRAFV600E-mutated COLO205 cells, both with resistance to MEK1 inhibition. To explore whether this also applies to other exper- CEMIP is connected to ErbB/MEK1-, LEF1-, and Myc-dependent imental models, we cultured parental HT-29 cells (HT-29/P) with pathways in colon adenocarcinoma increasing concentrations of selumetinib and generated highly To explore whether CEMIP links Wnt-dependent gene tran- resistant HT-29 cells (HT-29/SR; Fig. 4A). These cells showed scription to MEK1 signaling, we depleted CEMIP in BRAFV600E- decreased E-cadherin levels, suggesting that they underwent epi- mutated HT-29 colorectal cancer cells and carried out RNA-seq thelial–mesenchymal transition (EMT), a known feature of che- experiments combined with gene set enrichment analyses moresistance (Supplementary Fig. S6A). Importantly, CEMIP (GSEA; Fig. 2A–C; Supplementary Fig. S2A). In agreement with mRNA and protein levels were strongly induced in resistant our previous observations (23), CEMIP expression was linked to HT-29 cells (Fig. 4B). We next looked at the nuclear levels of ErbB/MEK1 signaling as a signature of genes induced through transcription factors that drive CEMIP gene transcription, namely

www.aacrjournals.org Cancer Res; 78(16) August 15, 2018 4537

Duong et al.

A Parental B Parental C 4.5 Cleaved caspase-3 Dapi *** 0.5 mm 4 Untreated Ki-67 3.5 Valproic acid/ Valproic acid/ kDa 3 CHIR999021 + CHIR999021 180 2.5 CEMIP 2 *** 130 1.5 *** 100 HSP90 1 12 0.5 Resistant 60 µm mRNA Levels (fold induction) 0 0.5 mm Resistant Lgr5 CEMIP Dclk1 SOX9 Cleaved caspase-3 Dapi Ki-67 Untreated organoids 60 µm 60 µm

14 *** DAPI DAPI 60 µm DCLK1 DCLK1 12 Organoids + inhibitors (6 days) 10 30 *** 60 µm 60 µm 8 25 Ki-67 6 Cleaved caspase-3

4 20

2 15

% of Organoids > 0.5 mm DAPI DAPI 0 DCLK1 DCLK1

Parental resistant 10 Percentage of cells of Percentage 5 70 F *** D Organoids 60 μ 0 Selumetinib ( mol/L, 24 h) cells 50 Parental Resistant + kDa 010.5 3 0 10.53 40

180 D133 30 /C CEMIP E + 20 130 kDa PR

24 CD 55 of Percentage pMEK1 170 HER3 0 43 Parental Resistant 55 90 MEK1 Parental Resistant BRAF 43 5 63.7 33.4 34.9 58.1 10 + - + + 5 17 CD24 /CD133 CD24 /CD133 10 CD24+/CD133- CD24+/CD133+ 55 p4EBP1 pMEK1 4 45 10 10 104 55 3 17 4EBP1 10 3 MEK1 10 45 CD24PE-A CD24PE-A

10 2 2 43 10 10 pERK1/2 100 HSP90 0 2.37 0.5 0 5.15 1.83

12 5 0 102 103 104 10 2 3 4 5 43 ERK1/2 0 10 10 10 10 CD133 APC-A CD133 APC-A 90 72 SOX9

72 Myc 55 100 HSP90

12 345 678

Parental Resistant

Figure 1. Selumetinib-resistant organoids show elevated levels of CEMIP and reactivation of the MEK1/ERK1/2 pathway. A, Selumetinib-resistant organoids are larger in size. Ex vivo organoids with intestinal crypts extracted from Apcþ/Min mice were treated with increasing concentrations of selumetinib. Quantification of their size is illustrated. Statistical analysis was performed as described in Materials and Methods. B, Resistant organoids do not proliferate more but show less apoptotic cells. Ki67 and activated caspase-3 staining were carried out to quantify the percentage of proliferative and apoptotic cells, respectively. Data from 20 organoids are illustrated. C, CEMIP expression is induced by Wnt signaling in ex vivo organoids generated from intestinal crypts of C57BL/6 mice. Ex vivo þ organoids were untreated or treated with valproic acid (1 mmol/L) and CHIR999021 (3 mmol/L) to enrich for Lgr5 stem cells. Top, mRNA levels of indicated candidates were quantified by real-time PCR analysis (see Materials and Methods for the quantification). Bottom, immunofluorescence analysis of the Tuft cell marker Dclk1. Anti-CEMIP Western blot analysis using extracts from untreated or valproic acid and CHIR999021-stimulated ex vivo organoids is shown. D and E, Selumetinib-resistant organoids show elevated levels of CEMIP, HER3, BRAF, SOX9, and c-Myc and enhanced activation of MEK1, ERK1/2, and mTOR. Protein extracts from parental and resistant ex vivo organoids were subjected to Western blot analysis. F, Selumetinib-resistant organoids were enriched in CD24þ/CD133þ cancer stem cells. FACS analysis was conducted to quantify the percentage of CD24þ/CD133þ cells. Data from four experiments are illustrated.

4538 Cancer Res; 78(16) August 15, 2018 Cancer Research

CEMIP Links Wnt and MEK1 Signaling Pathways

A B NF-kB and AP-1 family members (23, 32). Both p65 and FRA-1 HT-29 but not BCL-3 and c-JUN levels were increased in nuclear extracts from resistant HT-29 cells (Fig. 4C). Of note, cytoplasmic BRAF 1.2 1.0 showed elevated levels in resistant cells (Supplementary Fig. S6B),

expression 0.8 which reﬂects intrachromosomal ampliﬁcation (10). These

0.6 molecular changes persisted even in circumstances in which cells RNA e 0.4 *** were constantly cultured with selumetinib (Supplementary Fig. (fold induction) (fold 0.2 S6C). Therefore, CEMIP expression is induced in resistant HT-29 Relativ 0 V600E ShRNA CEMIP #2 cells in which mutated BRAF , nuclear p65, and FRA-1 protein Control levels are increased. FRA-1, but not p65 was actually driving CEMIP expression in these cells as FRA-1 but not p65 deficiency impaired CEMIP expression at both mRNA and protein levels C Enrichment plot: KRAS.DF.V1_UP Enrichment plot: ERB2_UP.V1_UP (Supplementary Figs. S7A and S7B, respectively). FRA-1 controls the expression of several candidates such as WNT10, DKK-1, and DVL-1 acting in the canonical Wnt pathway in colon cancer cells (33). As CEMIP transcription is robustly induced upon Wnt activation (29), we hypothesized that FRA-1 indirectly controls CEMIP expression through Wnt signaling. FRA-1 deficiency Enrichment score (ES) Enrichment score (ES) indeed impaired nuclear b-catenin levels and both WNT10 and Enrichment profile Hits Enrichment profile Hits DKK-1, two b-catenin target genes downregulated upon FRA-1 shCTR shCEMIP HT-29 HT-29 shCTR shCEMIP deficiency in HT-29/SR cells (Supplementary Figs. S7C and S7D). HT-29 HT-29 FRA-1 deficiency also triggered cell death of HT-29/SR cells, as Enrichment plot: LEF1_UP.V1_UP Enrichment plot: MEK1_UP.V1_UP judged by clonogenic assays, at least due to caspase-3/7 activation (Supplementary Figs. S7E and S7F). As CEMIP expression is induced in BRAFV600E-mutated cells, we reasoned that a BRAF inhibitor may decrease CEMIP expression. CEMIP mRNA and protein levels were severely decreased in HT-29/SR cells subjected to vemurafenib (Supplementary Figs. S8A and S8B). As a conse- Enrichment score (ES) Enrichment score (ES) Enrichment profile Hits Enrichment profile Hits quence, vemurafenib triggered some cell death and interfered

shCTR shCEMIP with the capacity of these cells to form colonies (Supplementary shCTR shCEMIP HT-29 HT-29 HT-29 HT-29 Figs. S8C and S8D, respectively). Taken together, our results deﬁne BRAFV600E, FRA-1, and b-catenin as upstream actors D V600E

MIR1914 VTRNA2-1 SNORA21 SNORA21 RPL17-C8orf44 HIST1H4H HIST1H4E C8orf44-SGK3.SGK3 HIST1H2Al HIST1H2AK LOC285419 HIST1H4l TERC C19orf81 FGF8 RAB4B-EGLN2 STX16-NPEPL1 UBE2F-SCLY AKR1C4 RDH16 PLEKHF1 CDA VAV1 APOBEC3G SYS1-DBNDD2 SGK3 LOC100129216 SCARNA9 DIAPH3-AS1 URGCP-MRPS24 RPS10-NUDT3 TGIF2-C20orf24 IFITM1 MIR941-2.MIR941-3 SNORD128 SNORD88C SNORD119 SNORD888 MIR4648 SNORD19 that drive CEMIP transcription in BRAF -mutated resistant colorectal cancer cells. As selumetinib decreased CEMIP expression in both selumetinib-resistant organoids and HT-29 cells (Fig. 3C and 4D, respectively), we looked at FRA-1 proteins E levels upon MEK1 inhibition in HT-29/SR cells. FRA-1 but also p65, c-JUN, b-catenin, and TCF4 were downregulated upon MEK1 inhibition while the epithelial marker E-cadherin was increased, (Fig. 4D). Moreover, TCF4 promotes CEMIP expression as CEMIP protein levels severely decreased upon TCF4 deﬁciency in selumetinib-resistant HT-29 cells (Supplementary Fig. S9A). Con- sistently, the treatment of two selumetinib-resistant colon cancer cell lines with PNU-74654, which inhibits the Wnt/b-catenin pathway by blocking the interaction between b-catenin and TCF4, decreased CEMIP mRNA and protein levels (Supplementary Fig. S9B). Both TCF4 and b-catenin were actually recruited at TCF-binding sites located on the CEMIP promoter as well as TF NES (significance) #Targets #Motifs/Tracks on intron 1 (Fig. 4E). Therefore, selumetinib decreases CEMIP MAX 6.564 199 9 MXI1 6.474 181 3 expression, at least by negatively regulating protein levels of both MYC 5.85 174 6 FRA-1 and TCF4. TFAP2C 5.124 130 18

Figure 2. cells is illustrated. C, Gene set enrichment analysis of RNA-seq expression CEMIP controls the expression of ErbB-, RAS-, MEK1-, and LEF1-dependent data obtained with total RNA from control and CEMIP-depleted HT-29 cells. signaling cascades. A, Generation of CEMIP-depleted colorectal cancer– Candidate genes up- or downregulated are illustrated in red and blue, derived HT-29 cells. Cells were transduced with the indicated shRNA respectively. D, Listing of the most robust up- or downregulated candidates lentiviral constructs. Real-time PCR analysis was performed to assess mRNA (red and blue rectangles, respectively) upon CEMIP deficiency in HT-29 cells. expression of CEMIP. B, CEMIP controls the expression of gene candidates E, Identification of CEMIP coexpressed genes in colon adenocarcinoma linked to ErbB, RAS, and LEF1 signaling. A scatter plot of RNA-seq data through iRegulon analysis. The transcription factors (TF) known to regulate obtained with RNA extracted from control versus CEMIP-deficient HT-29 these coexpressed genes (N ¼ 285) are listed.

www.aacrjournals.org Cancer Res; 78(16) August 15, 2018 4539

Duong et al.

A shRNA Control shRNA CEMIP#2 B Control ShRNAs CEMIP 105 9.0 76.0 105 27.0 - 49.8 CD24+/CD133- CD24+/CD133+ CD24+/CD133 CD24+/CD133+

104 104

200 µm 200 µm 103 103 shRNA CEMIP#5 shRNA CEMIP#3 CD24PE-A CD24PE-A 102 102 0 5.85 9.1 0 15.7 7.49

0 102 103 104 105 0 10 2 10 3 10 4 10 5 CD133 APC-A CD133 APC-A 200 µm 200 µm

80 ** + Control *** shRNAs *** + CEMIP

60 cells

*** + 60 180 CEMIP 50 130

40 133 CD 100 / HSP90 40 + 20

30 12

24 CD Percentage of Percentage 0 20 shRNAs Control CEMIP E Number of organoids Number 10 4 CEMIP *** 0 D COLO-205/P 3 Control CEMIP#3 120 COLO-205/SR(0.3) shRNAs CEMIP#2 CEMIP#5 ***

)lortnocfo%( ** 100 *** 2

C 80 *** (fold induction) 1

shRNAs mRNA Relative expression ControlCEMIP#1CEMIP#2ControlCEMIP#1CEMIP#2 lavivrusl 60 0 *** P SR(0.3) Selumetinib (5 μmol/L, 24 hours) +++ COLO-205 kDa 40 *** HER3 leC COL0-205 170 kDa 100 20 PSR(0.3) BRAF 180 CEMIP 55 pMEK1 0 130 40 000.01 .030.1 0.3 13 55 MEK1 Selumetinib (μmol/L) 55 α-Tubulin 40 12 pGSK3α COLO-205 43 F kDa G PSR(0.3) Selumetinib (μmol/L, 24 hours) GSK3α/β 0.3 1 0.3 1 0.3 1 43 100 BRAF 0 0.5 0 0.5 0 0.5 55 pMEK1 pERK1/2 55 pMEK1/2 40 40 40 ERK1/2 55 55 MEK1 40 MEK1/2 40 72 40 Myc 55 pERK1/2 pERK1/2 40 40 SOX9 72 ERK1/2 40 ERK1/2 40 Cyclin D1 pRSK1 34 70 72 34 Cyclin D2 Myc RSK1 55 70 180 180 40 CEMIP FRA-1 CEMIP 130 130 100 100 HSP90 55 α-Tubulin HSP90 123456 12 123456789101112

Control CEMIP #2 shRNAs CEMIP #1

Figure 3. CEMIP is required for the maintenance of selumetinib-resistant ex vivo organoid cultures and for Myc expression. A, CEMIP deficiency impairs the growth of selumetinib-resistant ex vivo organoids. Images of control and CEMIP-depleted ex vivo organoids are illustrated. The number of organoids per dish was calculated in each experimental condition. B, CEMIP deficiency impairs the pool of CD24þ/CD133þ cancer stem cells in selumetinib-resistant ex vivo organoids. FACS analysis was conducted to quantify the percentage of CD24þ/CD133þ cells in control and CEMIP-depleted resistant organoids. Data from two experiments are illustrated. C, CEMIP deficiency impairs HER3, Myc, SOX9, cyclin D1/2 levels, and ERK1/2 activation in selumetinib-resistant ex vivo organoids. (Continued on the following page.)

4540 Cancer Res; 78(16) August 15, 2018 Cancer Research

CEMIP Links Wnt and MEK1 Signaling Pathways

To explore whether KRASG13D or G12A-mutated colorectal (Supplementary Fig. S11C). CEMIP-deficient HT-29/SR cells cancer-derived cell lines showing some acquired resistance to showed lower levels of HER3, even after treatment with selume- selumetinib also show elevated levels of CEMIP, parental tinib and failed to maintain levels of phosphorylated ERK1/2 and HCT116, and SW480 cells were also treated with increasing RSK1 when treated with selumetinib (Fig. 5E). Moreover, E- concentrations of selumetinib to generate resistant HCT116 cadherin levels increased upon CEMIP deficiency, indicating that and SW480 cells, respectively (HCT116/SR and SW480/SR; CEMIP is required for EMT maintenance, a process linked to Supplementary Fig. S10A). Both resistant HCT116 and SW480 chemoresistance (Fig. 5E). CEMIP expression also decreased at the cell lines did not upregulate CEMIP, in contrast to resistant HT- protein level in selumetinib-treated cells, similar to FRA-1 and 29 cells (Supplementary Fig. S10A). Yet, ERK1/2 reactivation also to BRAF, but not KRAS levels, suggesting again that BRAF and was seen in all resistant cells (Fig. 4F). RSK1 activity was also FRA-1 controls CEMIP expression (Fig. 5E). Therefore, CEMIP specifically induced in BRAFV600E-butnotinKRASG13D or G12A- contributes to the acquired resistance to selumetinib, at least by mutated resistant cells (Fig. 4F). Both p65 and FRA-1 were not promoting MEK1–ERK1/2 signaling. dramatically induced in both HCT116- and SW480-resistant cells (Fig. 4F; Supplementary Fig. S10B). Levels of another scaffold protein, IQGAP-1, remained unchanged in resistant CEMIP is an endosomal protein HT-29 cells (Fig. 4F). BRAFV600E-mutated resistant HT-29 cells, We next carried out biochemical fractionation to identify cell and to a less extent KRASG13D-mutated HCT116 cells, also compartments from which CEMIP contributes to MEK1 and showed enhanced HER3 and MET reactivations (Fig. 4G). ERK1/2 reactivation in resistant cells. CEMIP cofractionated with Finally, a somatic mutation in MEK1 exon 3 (H119R) was EEA1 and APPL1, two signaling endosome markers, and to a less found in resistant HT-29 cells (Fig. 4H). The H119R mutation, extent with lysosomal markers (Rab7 and LAMP2) and with among others in the same domain of MEK1, was demonstrated Rab11, a recycling endosome marker (Fig. 6A). CEMIP also to underlie resistance to the MEK1 inhibitor, PD184352 (34). cofractionated with PDI, an endoplasmic reticulum marker, as Taken together, our data indicate that BRAFV600E-butnot described previously (24). Phosphorylated forms of MEK1 were KRASG13D or G12A-mutated colorectal cancer cells reactivate also detected in CEMIP-positive fractions, suggesting that signal- MAPK signaling and potently induce CEMIP gene transcription ing endosomes are critical for MEK1 reactivation (Fig. 6A). In upon acquired resistance to MEK1 inhibition. contrast, CEMIP did not cofractionate with Caveolin-1 and Flo- tillin-1, two lipid raft markers (Fig. 6A). To explore in which endosomes CEMIP is mainly located, we conducted a second CEMIP deficiency sensitizes BRAFV600E-mutated resistant HT-29 fractionation experiment in which organelles of interest (ER, cells to MEK1 inhibition peroxisomes, mitochondria, and endosomes) were enriched from To explore whether CEMIP contributes to the resistance to cell extracts and separated on a gradient by ultracentrifugation. þ selumetinib in BRAFV600E-mutated colorectal cancer cells, we CEMIP mainly cofractionated with EEA1 endosomes and to a þ þ depleted CEMIP from HT-29/SR cells and subjected them to much less extent with Rab5/7 or APPL1 endosomes (Fig. 6B). A selumetinib. MEK1 inhibition decreased CEMIP mRNA levels SNAP-CEMIP construct expressed in HCT116 cells also partially þ (Fig. 5A). CEMIP deficiency enhanced DNA damage, as evidenced colocalized with EEA1 endosomes and with the ER, as assessed by increased pH2A.X levels (S139) as well as caspase-3/7–depen- by immunofluorescence (Fig. 6C). CEMIP was the only endoso- dent apoptotic cell death upon selumetinib treatment (Fig. 5B and mal protein to be upregulated in HT-29/SR cells, as both EEA1 and C, respectively). CEMIP-depleted HT-29/SR cells generated less APPL1 levels remained unchanged (Supplementary Fig. S12A). colonies when subjected to selumetinib (Supplementary Fig. CEMIP associated with MEK1 in HT-29/SR cells, as evidenced by S11A). To assess whether this was also relevant in vivo, we carried coimmunoprecipitation (Fig. 6D) and BRAF more weakly bound out xenograft experiments in immunodeficient mice with control MEK1 in resistant versus parental HT-29 cells, more likely due to and CEMIP-depleted HT-29/SR cells and subsequently treated disengagement (Fig. 6E). Of note, pERK1/2 levels were totally mice with selumetinib. CEMIP mRNA levels were expectedly abolished after 0.5 and 1 hour of treatment with selumetinib in decreased in cells infected with the inducible shRNA construct both parental and resistant HT-29 cells but pERK1/2 levels were (Supplementary Fig. S11B). Also, selumetinib failed to signifi- again detectable after 24 hours of MEK1 inhibition (Supplemen- cantly trigger tumor regression in vivo in mice injected with HT- tary Fig. S12B). CEMIP actually contributes to MEK1 activity as an 29/SR cells (Fig. 5D). Importantly, CEMIP deficiency, combined anti-MEK1 immunoprecipitate from CEMIP-depleted HT-29/SR with selumetinib, caused significant tumor regression (Fig. 5D). cells was less potent at phosphorylating ERK2 (Supplementary This was due to DNA damage and apoptosis, as evidenced by Fig. S12C). CEMIP failed to bind mutated BRAFV600E in resistant anti-pH2A.X (S139) and cleaved caspase-3 immunofluorescence HT-29 cells (Supplementary Fig. S13A). Although more

(Continued.) Control or CEMIP-depleted resistant ex vivo organoids were untreated or stimulated with selumetinib and cell extracts were subjected to Western blot analysis. D, Sensitivity of parental (COLO-205/P) and resistant (COLO-205/SR) COLO-205 cells to selumetinib. COLO-205/P and COLO-205/SR cell lines were untreated or treated with selumetinib at indicated concentrations for 72 hours. The percentage of viable cells in untreated COLO-205/P or COLO-205/SR cells was set to 100% and the percentage of viable cells upon treatment with selumetinib in COLO-205/P or COLO-205/SR cells was relative to the untreated cells. Data from two independent experiments performed in triplicate are shown. E, Resistant COLO-205 cells show enhanced CEMIP expression. Top, mRNA levels of CEMIP in both COLO-205/P and COLO-205/SR cells were assessed by real-time PCR. Bottom, CEMIP protein levels in both COLO-205/P and COLO-205/SR cells were assessed by Western blot analysis. F, Resistant COLO-205 cells show enhanced MEK1/2 and ERK1/2 activation. Extracts from parental and resistant COLO-205 cells were subjected to Western blot analysis. G, CEMIP deﬁciency in selumetinib-resistent COLO205 cells impairs MEK1 and ERK1/2 activation and Myc stability. Control or CEMIP-depleted COLO-205/SR (0.3) cells were unstimulated or treated with selumetinib for 24 hours at the indicated concentrations and cell extracts were subjected to Western blot analysis.

www.aacrjournals.org Cancer Res; 78(16) August 15, 2018 4541

Duong et al.

A )lortnocfo%( B 140 Cytoplasmic Nuclear HT-29/P CEMIP *** n 28 HT-29 HT-29 120 HT-29/SR(1.5) oi ** *** s kDa P SR(1.5) P SR(1.5) *** s 24

100 e )noitcu

*** rpx HT-29 70 kDa p65 80 *** 20 P SR(1.5)

e 180 55 lavivruslleC 60 ** ANR 16 CEMIP 55 dnidl 130 BCL-3 12 40 40

e 100 HSP90 vitaleR

o 8 40 20 f( c-JUN 12 0 4 00.330.1 1 10 30 40 FRA-1 Selumetinib (μmol/L) 0 PSR(1.5) 40 FRA-1 μ D Selumetinib ( mol/L, 24 hours) E (long exposure) kDa 0 0,1 1 3 0 0,1 1 3 α 180 Site #1 Site #2 Site #3 Site #4 55 -Tubulin CEMIP 130 100 NBS1 40 FRA-1 Exon 1 Exon 2 12 34 Promoter Intron 1 90 β-Catenin

100 IgG 70 p65 90 80 TCF4 0 hours selumetinib 70 TCF4 70 TCF4 24 hours selumetinib 60 50 40 c-JUN 40 30

130 enrichment Relative 20 E-Cadherin 10 0 70 Lamin B2 Site #1 Site #2 Site #3 Site #4 Negative exon1 100 HSP90

12 34 56 78 G HT-29 HCT116 H HT-29/SR Cytoplasm Nucleus SR(1.5) SR(2) P P F HT-29 HCT116 SW480 kDa kDa P SR(1.5) P SR(2) P SR(1.5) pHER3 180 180 CEMIP 130 HER3 180 HT-29/P 180 IQGAP-1 180 pMET 100 BRAF 130

25 180 KRAS MET 130 15 55 H119R pMEK1/2 100 HSP90 40 55 40 MEK1/2 1234 F133L 40 pERK1/2 Q56P K59del I99T I103N K104N I111N/C L115R/P E120D C121S P124L/S G128D F129L V211D V2115P

40 ERK1/2

100 pRSK1 Helix C Activation loop 100 RSK1 Exon 2 Exon 3 Exon 6

70 pp65

70 p65

40 FRA-1

55 α-Tubulin 1234 5 6

4542 Cancer Res; 78(16) August 15, 2018 Cancer Research

CEMIP Links Wnt and MEK1 Signaling Pathways

BRAFV600E dimers were detected by the proximal ligation assay in Myc or CEMIP confirmed that both proteins control the produc- resistant versus parental HT-29 cells, CEMIP was dispensable for tion of multiple metabolites such as amino acids (methionine, BRAFV600E dimerization (Supplementary Fig. S13B and S13C). threonine, tryptophan, valine, proline, histidine, asparagine, phe- Therefore, CEMIP is localized in several cell compartments and nylalanine, isoleucine, leucine, glycine, and L-alanine) and lactate contributes to MEK1 reactivation from signaling endosomes in among other candidates (Fig. 7D). Therefore, CEMIP may pro- resistant BRAFV600E-mutated colorectal cancer cells as a MEK1- mote the acquired resistance to MEK1 inhibition, in part by binding protein. potentially regulating levels of specific metabolites via a Myc- associated signaling pathway. CEMIP promotes metabolic reprogramming potentially through Myc To explore the biology downstream of CEMIP, we established Discussion the metabolomic signature of both parental and resistant orga- We describe here the characterization of CEMIP as an endoso- noids. Severe metabolic reprogramming was detected in resistant mal protein that links Wnt-dependent gene transcription to organoids as they showed elevated levels of TCA intermediates MEK1–ERK1/2 signaling to promote acquired resistance to MEK1 (fumarate, malate, citrate, and succinate; Fig. 7A). Multiple inhibition in BRAFV600E-mutated colorectal cancer cells. CEMIP nucleotides, whose synthesis relies on Myc (35), were increased expression is induced in resistant cells through BRAFV600E, MEK1, in resistant organoids (Fig. 7A). Proline, whose degradation is RSK1, and FRA-1, which provides a mechanism by which these inhibited by Myc (36), was detected at higher levels in resistant signaling proteins promote resistance to inhibitors of RAS effec- organoids (Fig. 7A). Moreover, levels of arachidonic acid, which is tors. In addition, CEMIP regulates levels of multiple amino acids regulated by Myc in lung cancer (37), were also elevated in seen in resistant cells, at least through Myc. resistant organoids as were levels of other unsaturated fatty acids Multiple transcription factors govern CEMIP transcription, such as oleic acid, a candidate reported to be upregulated in colon including the NF-kB proteins BCL-3 and p65 in cervical cancer cancer (Fig. 7A; ref. 38). Finally, levels of cystathionine, which is cells (23). Functional NF-kB- and AP-1–binding sites were also generated by cystathionine b-synthase (CBS), an enzyme down- identified on the CEMIP promoter in breast cancer cells (32). We regulated in gastrointestinal and hepatocellular malignancies define here FRA-1, one member of the AP-1 family of transcription (39, 40), were decreased in resistant organoids (Fig. 7A). There- factors as well as TCF4 as key drivers of CEMIP expression in fore, metabolic reprogramming is seen in selumetinib-resistant resistant BRAFV600E-mutated colorectal cancer cells. The BRAF intestinal organoids. Importantly, CEMIP expression contributes inhibitor, which indirectly turns off ERK1/2 activity, also to this process as the production of lactate as well as levels of decreases FRA-1 protein levels. This observation fits with the fact multiple amino acids was impaired upon CEMIP deficiency in that ERK1/2 signaling stabilizes FRA-1 by preventing its protea- these ex vivo–resistant organoids (Fig. 7B). To better define CEMIP some-dependent degradation in colorectal cancer cells (41). as an upstream regulator of Myc, we reasoned that Myc deficiency Therefore, interfering with ERK1/2 signaling downregulates would mimic CEMIP deficiency in selumetinib-resistant ex vivo CEMIP transcription, at least through the destabilization of organoids. Indeed, the depletion of Myc impaired ERK1/2 acti- FRA-1 in resistant BRAFV600E-mutated colorectal cancer cells. This vation and also downregulated CEMIP protein but not mRNA signaling cascade critically drives CEMIP transcription to establish levels, suggesting that CEMIP and Myc mutually posttranscrip- a positive loop as CEMIP physically binds MEK1 (but not BRAF) tionally control their expression (Fig. 7C; Supplementary Fig. to sustain MEK1 activity. S14A). Myc was also critical for the maintenance of selumeti- Our resistant BRAFV600E-mutated colorectal cancer cells have nib-resistant ex vivo organoids, similar to CEMIP (Supplementary several features linked to acquired resistance. They show Fig. S14B). Moreover, Myc depletion had a profound effect on the enhanced phosphorylation of MET and HER3, elevated levels of levels of multiple metabolites as levels of lactate and numerous BRAFV600E as well as a MEK1 mutation, all events contributing to amino acids were significantly downregulated in Myc-depleted ERK1/2 reactivation. The upregulation of CEMIP has been cells (Supplementary Fig. S14C). A comparison of the metabolic detected in all tested BRAFV600E- but not KRASG13D or G12A-mutated signatures of selumetinib-resistant ex vivo organoids with depleted colorectal cancer cells showing some acquired resistance to MEK1

Figure 4. Selumetinib-resistant colorectal cancer cells mutated for BRAFV600E but not for KRASG13D or G12A show higher levels of CEMIP. A, Sensitivity of parental (HT-29/P) and resistant (HT-29/SR) cells to selumetinib. HT-29/P and HT-29/SR cell lines were untreated or treated with selumetinib at indicated concentrations for 72 hours. The percentage of viable cells in untreated HT-29/P or HT-29/SR cells was set to 100% and values in other experimental conditions were relative to the untreated cells. Data from two independent experiments performed in triplicate are shown. B, Acquisition of selumetinib resistance in HT-29/SR cells induces CEMIP expression. HT-29/P or HT-29/SR cell lines grew without selumetinib (1.5 mmol/L) for 48 hours. Left, real-time PCR analysis was conducted on total RNA to assess CEMIP expression. Right, protein extracts were subjected to Western blot analysis. C, Acquisition of selumetinib resistance in HT-29/SR cells induces nuclear expression of p65 and FRA-1. HT-29/P or HT-29/SR cell lines grew without selumetinib (1.5 mmol/L) for 48 hours. Cytoplasmic and nuclear protein extracts were subjected to Western blot analysis. D, Selumetinib decreases CEMIP expression in selumetinib-resistant and BRAFV600E-mutated colon cancer cells. HT-29/SR cells were untreated or stimulated with selumetinib for 24 hours at indicated concentrations and both nuclear and cytoplasmic extracts were subjected to Western blot analysis. E, TCF4 is recruited at TCF-binding sites on both the CEMIP promoter and intron 1 (see Materials and Methods for details), as judged by ChIP assays using HT-29/SR cells. Primers within exon 1 were randomly chosen and used as negative controls. Values for each primer pair were calculated as ratios between ChIP signals obtained with the anti-TCF4 (specific) and or IgG (nonspecific) antibodies. F, Resistant colorectal cancer cell lines show elevated levels of pERK1/2 independent of the mutational status of BRAF and KRAS. Parental and selumetinib-resistant BRAFV600E-mutated (HT-29) and KRASG13D or G12A-mutated (HCT116 and SW480, respectively) cell lines were grown without selumetinib for 48 hours. Cell extracts were subjected to Western blot analysis. G, Resistant BRAFV600E-mutated cells show reactivation of HER3 and MET, as demonstrated by Western blots. H, Resistant HT-29 cells have a MEK1 somatic mutation. Top, identification of the MEK1 H119R mutation found in HT-29/SR cells. Bottom, a representation of the MEK1 coding sequence with described somatic mutations of MEK1 in exons 2, 3, and 6.

www.aacrjournals.org Cancer Res; 78(16) August 15, 2018 4543

Duong et al.

inhibition. Yet, the induction of CEMIP upon acquired resistance was found. Therefore, the combination of both BRAFV600E and was more severe in resistant BRAFV600E-mutated colorectal cancer MEK1 mutations may be key genetic events to efﬁciently drive HT-29 cells in which the MEK1 mutation within exon 3 (H119R) CEMIP expression upon acquired resistance.

A KIAA1199 CEMIP#2 C 8 shRNA CEMIP#2 shRNA Control Control

B ytivit *** Control CEMIP#2 shRNA

n 1.4 *** )noi o 6 isserpxeANRm *** Selumetinib (mmol/L)

1.2 ca7,3-es )n ** kDa 0 0.1 1 3 00.11 3 tcudnidlo

oi 1.0 4

tc 17 pH2A.X **

u 0.8 a

d pDNA-PKcs

p ni

300 f 2 (

0.6 s aC

dl DNA-PKcs

o e f 300

( 0.4 v

ita 0 55 a -Tubulin 03

leR 0.2 12345678 Selumetinib (mmol/L) 0 010.1 3 Selumetinib (mmol/L)

D E HT-29/SR(1.5)

Control d

etaer Control CEMIP#2 shRNA

Selumetinib (mmol/L) tn

shRNA kDa 00.11300.11 3 CEMIP U 180 HER3 130 100

bin 100

Control BRAF i temuleS 25 KRAS

shRNA CEMIP 15 55 pMEK1/2 40 55 2.0 MEK1/2 * ns 40 40 pERK1/2 1.5 *** * 40 ERK1/2

1.0 100 pRSK1

100 RSK1 0.5 40 FRA-1 Tumor weight (g)

0.0 130 E-cadherin 180 Control Control CEMIP shRNA CEMIP CEMIP 130 ++ 100 HSP90 Selumetinib 12345678

Figure 5. CEMIP promotes acquired resistance to selumetinib in BRAFV600E-mutated colorectal cancer cells. A, Selumetinib downregulates CEMIP expression in resistant HT- 29 cells. Control or CEMIP-depleted HT-29/SR cells were untreated or stimulated with selumetinib at indicated concentrations for 12 hours and real-time PCR analysis was conducted on total RNA to assess CEMIP expression. B, CEMIP deficiency enhances selumetinib-induced DNA damage in resistant HT-29 cells. Control or CEMIP-depleted HT-29/SR cells were untreated or stimulated with selumetinib at indicated concentrations for 24 hours. Protein extracts were subjected to Western blot analysis. C, CEMIP deficiency induces apoptotic cell death by selumetinib in resistant HT-29 cells. Control or CEMIP-depleted HT-29/SR cells were untreated or treated with selumetinib (3 mmol/L) for 24 hours. Caspase-3/7 activity in control and unstimulated cells was set to 1 and caspase-3/7 activity in other experimental conditions that were relative to the control. Data from three independent experiments performed in duplicate are shown. D, CEMIP deficiency significantly enhances tumor regression induced by selumetinib. A stable IPTG-inducible control or CEMIP-depleted HT-29/SR cell line (2 106 cells) was injected subcutaneously into NOD/SCID male mice. Mice were treated with IPTG from day 3 postinjection and treated with selumetinib (20 mg/kg) from day 6 postinjection for two weeks, as described in Materials and Methods (n ¼ 4/group). Top, representative images of tumors excised at day 20 postinjection. Bottom, tumor weights were quantified. E, CEMIP deficiency increases E-cadherin expression and impairs selumetinib-dependent HER3 expression in resistant HT-29 cells. Control or CEMIP-depleted HT-29/SR cells were untreated or stimulated with selumetinib at indicated concentrations for 24 hours. Protein extracts were subjected to Western blot analysis.

4544 Cancer Res; 78(16) August 15, 2018 Cancer Research

CEMIP Links Wnt and MEK1 Signaling Pathways

A 1 2 34 5 6 7 8 9 10 1112 1314 Fractions kDa B kDa 123456789101112 180 CEMIP 180 130 CEMIP 130 170 EEA1 170 EEA1 100 APPL1 72 26 Rab5 26 Rab5 26 Rab7 26 Rab7 100 APPL1 26 Rab11 72 72 90 LAMP2 55 PDI 26 Caveolin-1 55 C Flotillin-1 44 PDI SNAP-CEMIP SNAP-CEMIP DAPI 72 PDI PDI 55 20 µm 130 N-Cadherin 90

170 EGFR 20 µm 20 µm Merge 170 MET EEA1 SNAP-CEMIP SNAP-CEMIP DAPI 130 EEA1

72 SRC 55 20 µm 100 B-RAF 26 20 µm 20 µm K-RAS Merge 17 55 pMEK1/2 44 44 pERK1/2

D E B-RAF B-RAF IP IgG IgG MEK1 MEK1 IP IgG IgG Selumetinib (hours) Selumetinib (hours) kDa kDa 010 0102 2 00.50 2 400 0.52 4 55 MEK1 180 CEMIP 44 130 IP IP 55 pMEK1/2 100 B-RAF 44 55 MEK1 100 44 B-RAF 55 55 pMEK1/2 pMEK1/2 44 44 55 55 MEK1/2 MEK1 44 44 WCE 44 pERK1/2 44 pERK1/2 WCE 44 ERK1/2 44 ERK1/2 180 CEMIP 170 130 CEMIP 130 100 HSP90 100 HSP90 12345678 123456789 10 HT-29 P HT-29 SR(1.5) HT-29 P HT-29 SR(1.5)

Figure 6. CEMIP is an endosomal protein that binds MEK1 in resistant BRAFV600E-mutated HT-29 cells. A, A pool of CEMIP is found in signaling endosomes. Protein extracts from HT-29/SR cells were biochemically fractionated on an OptiPrep gradient as described in Materials and Methods and the resulting fractions were subjected to Western blot analysis. B, CEMIP mainly cofractionates with EEA1þ endosomes. A fractionation experiment in which some pellets enriched with organelles of interest were further separated into fractions by ultracentrifugation (see Materials and Methods). Fractions were subjected to Western blot analysis. C, CEMIP partially colocalizes with ER and endosomal markers. HCT116 cells were transfected with the SNAP-CEMIP construct and immunoﬂuorescence was conducted on resulting cells. Arrows, colocalization of CEMIP with EEA1þ endosomes and with the ER using the PDI marker. D, CEMIP binds MEK1 in resistant HT-29 cells. Protein extracts from parental and resistant HT-29 cells untreated or treated with selumetinib for the indicated periods of time were subjected to immunoprecipitation (IP) with anti-IgG (negative control) or anti-MEK1 antibodies, followed by Western blot analysis. Whole-cell extracts (WCE) were also subjected to Western blot analysis. E, Enhanced MEK1 activation in resistant BRAFV600E-mutated HT-29 cells leads to disengagement from BRAF. Protein extracts from parental and resistant HT-29 cells treated or not with selumetinib for the indicated periods of time were subjected to immunoprecipitation with anti-BRAF or anti-IgG antibodies, followed by Western blot analysis carried out on the immunoprecipitates or on whole-cell extracts.

www.aacrjournals.org Cancer Res; 78(16) August 15, 2018 4545

Duong et al.

A Class NAD C Urate 1.5 kDa Oxypurinol shRNAs Xanthine 180 Serine 1 CEMIP αKG Dodecanoyl-carnitine Class 130 Myristoyl-carnitine 0.5 Palmitoyl-carnitine 72 Thymidine P Cystathionine 0 Myc Acetyl-glutamine R 55 O-Phosphoethanolamine -0.5 Guanine α β CDP 43 pGSK3 / GDP -1 AMP Adenosine UMP -1.5 CMP 43 GSK3α/β GMP ADP UDP-GlcNAc Coenzyme A Uridine diphosphate 43 pERK1/2 Aminoadipase Sedoheptulose-7-phosphate Fumarate Malate 43 PEP ERK1/2 Citrate GSH Fructose 1,6-diphosphate 100 Pantothenate HSP90 Proline Cystidine Lysine 12 34 Linolenic acid Top 50 T-test FAD 2 groups only Arachidonic acid Cis-aconitate Dihydroxyacetone p Glyceraldehyde 3-p Oleic acid Palmitoleic acid P – Parental S-Adenosyl-L-homoc Succinic acid R – Resistant Nicotinamide Ornithine GSSG P1 P2 P3 R1 R2 R3

CEMIP Control (CEMIP) shRNAs B CEMIP Control shRNAs D Myc#3 Myc#2 Control (Myc)

L-Sarcosine PEP Proline Butyryl-carnitine Aminoadipate Betaine Folate Ethanolamine Phosphate Glycerol 3-phosphate Eicosapentaenoic acid Lactate Arachidonic acid Oxypurinol 1.5 Aspartate Xanthine 1.0 Acetyl-carnitine Glutamine SuccinylCys 0.5 Myristoyl-carnitine Urate 0 Palmitoyl-carnitine L-Alanine Propionyl-carnitine -0.5 Citrulline Acetyl-carnitine Asparagine IsoLeucine -1.0 Carnitine Carnitine -1.5 Methionine Phenylalanine Threonine Leucine Tryptophane Acetyl-lysine Oxoadipate Oxoadipate Glycerol-3 phosphate Hypoxanthine Valine 2 Inosine Thymidine Proline Threonine Histidine 1 Argininosuccinate Asparagine Succinyladenosine Phenylalanine IMP IsoLeucine 0 Acetyl-glutamine Cystathionine Leucine aKG Glycine -1 UDP Acetyl-aspartate N CTP Acetyl-glutamine CDP Folate ADP L-Sarcosine -2 UTP ATP Aminoadipate Betaine Urate GTP Lactate NADP Thymidine N-carbamoyl-L-aspa Oxypurinol Pyruvate Xanthine PEP Citrulline NADH GSH Cystathionine Ethanolamine Phosphate Argininosuccinate O-Phosphoethanolamine SuccinylCys Creatine L-Alanine GDP Acetyl-lysine Aspartate Succinyladenosine Butyryl-carnitine Taurine Creatine GLN Serine Inosine IMP Arginine

Figure 7. Metabolomic reprograming regulated by CEMIP and Myc is linked to the acquired resistance to selumetinib. A, Resistant organoids show metabolic reprogramming. The metabolomic profile was established in parental and resistant organoids and is presented as heatmap visualization and hierarchical clustering analysis of the top 50 compounds with P 0.05, Student t test. Rows, metabolites; columns, samples; color key indicates metabolite expression value (blue, lowest; red, highest). Data with triplicates is presented. B, CEMIP promotes the production of lactate and the synthesis of amino acids in selumetinib-resistant ex vivo organoids. The metabolomic profile was established in control and CEMIP-depleted resistant ex vivo organoids and is presented as heatmap visualization and hierarchical clustering analysis of the top 50 compounds with P 0.05, Student t test. Rows, metabolites; columns, samples; color key indicates metabolite expression value (blue, lowest; red, highest). Experiments were carried out in triplicates. C, Myc deficiency impairs CEMIP and ERK1/2 activation in selumetinib-resistant ex vivo organoids. Extracts from control and Myc-depleted resistant ex vivo organoids were subjected to Western blot analysis. D, CEMIP and Myc control the production of specific metabolites. The metabolic signatures established in control, Myc-depleted, and CEMIP-depleted selumetinib-resistant ex vivo organoids were compared. Control (CEMIP) and Control (Myc) experimental conditions represent selumetinib-resistant ex vivo organoids infected with control shRNA and were used to compare CEMIP and Myc-depleted organoids, respectively. The data are presented as heatmap visualization and hierarchical clustering of the top 50 compounds with P 0.05 Student t test. Rows, metabolites; columns, samples; color key indicates metabolite expression value (blue, lowest; red, highest).

4546 Cancer Res; 78(16) August 15, 2018 Cancer Research

CEMIP Links Wnt and MEK1 Signaling Pathways

CEMIP is found in signaling endosomes and is essential for proteins to support proliferation and survival (49). Essential ERK1/2 reactivation in BRAFV600E- but not KRASG13D or G12A- amino acids such as leucine, tryptophan, and phenylalanine, mutated cells, at least through binding to MEK1. Whether or not whose levels are controlled by both CEMIP and Myc in selume- ERK1/2 activation downstream of tyrosine kinase receptors occurs tinib-resistant organoids, has been defined as signaling molecules from endosomes or from the cytoplasmic membrane has been the for mTOR activation (50). Therefore, CEMIP and its downstream subject of an intense debate. While some studies support the effector Myc may indirectly control mTOR signaling through the notion that MAPK scaffold complexes found in endosomes are production of specific essential amino acids to support acquired critical for signal transduction, other reports state that signaling resistance to MEK1 inhibition. from a tyrosine kinase receptor occurs from the cytoplasmic In conclusion, our study defined the scaffold and endosomal membrane (42, 43). In support with this later hypothesis, EGFR protein CEMIP as an upstream regulator of Myc that links Wnt- endocytosis in endosomes helps to terminate Ras-dependent and MEK1-dependent signaling pathways. As CEMIP is linked to signaling to ERK1/2 as endogenous Ras is primarily located at Myc and to specific metabolic reprogramming seen in resistant the cytoplasmic membrane in low EGFR-expressing cells (44). cells, this oncogenic pathway may hold therapeutic interest. BRAFV600E does not bind CEMIP, which fits with the hypothesis that CEMIP only binds signaling proteins such as EGFR or MEK1 Disclosure of Potential Conflicts of Interest found in endosomes but not candidates such as BRAF, which is No potential conflicts of interest were disclosed. activated at the cytoplasmic membrane (23, 45). The enhanced fi V600E CEMIP expression that we speci cally see in resistant BRAF - Authors' Contributions mutated colorectal cancer cells may help to recycle HER3 and MET Conception and design: H.-Q. Duong, J.C. Marine, K. Shostak, A. Chariot at the cytoplasmic membrane to sustain ERK1/2 signaling and/or Development of methodology: H.-Q. Duong, R. Buttner,€ K. Shostak, A. Chariot to favor the assembly of a specific endosomal signaling platform Acquisition of data (provided animals, acquired and managed patients, for ERK1/2 reactivation. provided facilities, etc.): S.C. Tang, L. Tharun, R. Buttner,€ K. Shostak, A. Chariot A previous study showed that a pool of CEMIP can be found Analysis and interpretation of data (e.g., statistical analysis, biostatistics, in the ER (24). These results, combined with our study reveal- computational analysis): H.-Q. Duong, I. Nemazanyy, F. Rambow, S. Delaunay, P. Close, K. Shostak, A. Chariot ing CEMIP in endosomes, raises some questions on mechan- Writing, review, and/or revision of the manuscript: H.-Q. Duong, R. Buttner,€ isms by which a scaffold protein localized in two distinct cell J.C. Marine, K. Shostak, A. Chariot compartments, promotes survival and chemoresistance. The Administrative, technical, or material support (i.e., reporting or organizing answer may come from the existence of membrane contacts data, constructing databases): L. Tharun, A. Florin, J.C. Marine between endosomes and the ER, a process that contributes to Study supervision: S. Delaunay, K. Shostak, A. Chariot EGFR-dependent signaling (46). These physical contacts may help CEMIP to bring signaling proteins together to sustain Acknowledgments ERK1/2 activation in resistant BRAFV600E-mutated colorectal We thank Ganna Panasyuk (INSERM, Paris) and Phillip Williams (GIGA- cancer cells. Cardiovascular Sciences, ULiege, Belgium) for their critical reading of the fi manuscript and the GIGA Imaging and Flow Cytometry Facility. This study One key mechanism through which CEMIP de ciency circum- fi V600E was supported by grants from the Belgian National Funds for Scienti c Research vents the acquired resistance to MEK1 inhibition in BRAF - (FNRS), from the Concerted Research Action Program (Bio-Acet) and Special mutated colorectal cancer cells may be through Myc, which is in Research Funds (FSR) at the University of Liege, the Belgian Foundation against agreement with the fact that the pharmacologic inhibition of Myc Cancer (FAF-F/2016/794), as well as from the Walloon Excellence in Life circumvents the acquired resistance to c-Met inhibition (47). Our Sciences and Biotechnology (WELBIO-CR-2015A-02) and the Max Planck correlative metabolic data show that Myc is a key effector down- Society. We are also grateful to the "Fonds Leon Fredericq" and the "Centre fi stream of CEMIP as CEMIP and Myc similarly control the pro- Anticancereux" of the CHU Liege for their nancial support. duction of multiple metabolites including lactate as well as amino The costs of publication of this article were defrayed in part by the payment of acids such as glycine, which has been defined, among others, as a page charges. This article must therefore be hereby marked advertisement in driver of cancer pathogenesis (48). We demonstrated that selu- accordance with 18 U.S.C. Section 1734 solely to indicate this fact. metinib-resistant ex vivo organoids show high levels of multiple amino acids, which can be metabolized as a source of carbon and Received October 13, 2017; revised March 2, 2018; accepted June 12, 2018; nitrogen for biosynthesis of fatty acids, lipids, nucleotides, and published first June 18, 2018.

References 1. Vogelstein B, Kinzler KW. Cancer genes and the pathways they control. Nat 6. Pai RK, Jayachandran P, Koong AC, Chang DT, Kwok S, Ma L, et al. BRAF- Med 2004;10:789–99. mutated, microsatellite-stable adenocarcinoma of the proximal colon: an 2. Bos JL. ras oncogenes in human cancer: a review. Cancer Res 1989;49: aggressive adenocarcinoma with poor survival, mucinous differentiation, 4682–9. and adverse morphologic features. Am J Surg Pathol 2012;36:744–52. 3. Davies H, Bignell GR, Cox C, Stephens P, Edkins S, Clegg S, et al. Mutations 7. Popovici V, Budinska E, Tejpar S, Weinrich S, Estrella H, Hodgson G, et al. of the BRAF gene in human cancer. Nature 2002;417:949–54. Identiﬁcation of a poor-prognosis BRAF-mutant-like population of 4. Normanno N, Tejpar S, Morgillo F, De Luca A, Van Cutsem E, Ciardiello F. patients with colon cancer. J Clin Oncol 2012;30:1288–95. Implications for KRAS status and EGFR-targeted therapies in metastatic 8. Prahallad A, Bernards R. Opportunities and challenges provided by CRC. Nat Rev Clin Oncol 2009;6:519–27. crosstalk between signalling pathways in cancer. Oncogene 2016;35: 5. Clancy C, Burke JP, Kalady MF, Coffey JC. BRAF mutation is associated with 1073–9. distinct clinicopathological characteristics in colorectal cancer: a systematic 9. Bernards R. A missing link in genotype-directed cancer therapy. Cell review and meta-analysis. Colorectal Dis 2013;15:e711–8. 2012;151:465–8.

www.aacrjournals.org Cancer Res; 78(16) August 15, 2018 4547

Duong et al.

10. Little AS, Balmanno K, Sale MJ, Newman S, Dry JR, Hampson M, et al. 30. Kemper K, Versloot M, Cameron K, Colak S, de Sousa e Melo F, de Jong JH, Amplification of the driving oncogene, KRAS or BRAF, underpins acquired et al. Mutations in the Ras-Raf Axis underlie the prognostic value of CD133 resistance to MEK1/2 inhibitors in colorectal cancer cells. Sci Signal 2011;4: in colorectal cancer. Clin Cancer Res 2012;18:3132–41. ra17. 31. Janky R, Verfaillie A, Imrichova H, Van de Sande B, Standaert L, 11. Corcoran RB, Dias-Santagata D, Bergethon K, Iafrate AJ, Settleman J, Christiaens V, et al. iRegulon: from a gene list to a gene regulatory Engelman JA. BRAF gene amplification can promote acquired resistance network using large motif and track collections. PLoS Comput Biol to MEK inhibitors in cancer cells harboring the BRAF V600E mutation. Sci 2014;10:e1003731. Signal 2010;3:ra84. 32. Kuscu C, Evensen N, Kim D, Hu YJ, Zucker S, Cao J. Transcriptional and 12. Corcoran RB, Ebi H, Turke AB, Coffee EM, Nishino M, Cogdill AP, et al. epigenetic regulation of KIAA1199 gene expression in human breast EGFR-mediated re-activation of MAPK signaling contributes to insensitiv- cancer. PLoS One 2012;7:e44661. ity of BRAF mutant colorectal cancers to RAF inhibition with vemurafenib. 33. Iskit S, Schlicker A, Wessels L, Peeper DS. Fra-1 is a key driver of colon cancer Cancer Discov 2012;2:227–35. metastasis and a Fra-1 classifier predicts disease-free survival. Oncotarget 13. Prahallad A, Sun C, Huang S, Di Nicolantonio F, Salazar R, Zecchin D, et al. 2015;6:43146–61. Unresponsiveness of colon cancer to BRAF(V600E) inhibition through 34. Delaney AM, Printen JA, Chen H, Fauman EB, Dudley DT. Identification of feedback activation of EGFR. Nature 2012;483:100–3. a novel mitogen-activated protein kinase kinase activation domain recog- 14.SunC,HoborS,BertottiA,ZecchinD,HuangS,GalimiF,etal. nized by the inhibitor PD 184352. Mol Cell Biol 2002;22:7593–602. Intrinsic resistance to MEK inhibition in KRAS mutant lung and colon 35. Liu YC, Li F, Handler J, Huang CR, Xiang Y, Neretti N, et al. Global cancer through transcriptional induction of ERBB3. Cell Reports regulation of nucleotide biosynthetic genes by c-Myc. PLoS One 2008;3: 2014;7:86–93. e2722. 15. Corcoran RB, Atreya CE, Falchook GS, Kwak EL, Ryan DP, Bendell JC, et al. 36. Liu W, Le A, Hancock C, Lane AN, Dang CV, Fan TW, et al. Reprogramming Combined BRAF and MEK inhibition with dabrafenib and trametinib in of proline and glutamine metabolism contributes to the proliferative and BRAF V600-mutant colorectal cancer. J Clin Oncol 2015;33:4023–31. metabolic responses regulated by oncogenic transcription factor c-MYC. 16. Ahronian LG, Sennott EM, Van Allen EM, Wagle N, Kwak EL, Faris JE, et al. Proc Natl Acad Sci U S A 2012;109:8983–8. Clinical acquired resistance to RAF inhibitor combinations in BRAF- 37. Hall Z, Ament Z, Wilson CH, Burkhart DL, Ashmore T, Koulman A, et al. mutant colorectal cancer through MAPK pathway alterations. Cancer Myc expression drives aberrant lipid metabolism in lung cancer. Cancer Discov 2015;5:358–67. Res. 2016;76:4608–18. 17. Roy M, Li Z, Sacks DB. IQGAP1 is a scaffold for mitogen-activated protein 38. Szachowicz-Petelska B, Sulkowski S, Figaszewski ZA. Altered membrane kinase signaling. Mol Cell Biol 2005;25:7940–52. free unsaturated fatty acid composition in human colorectal cancer tissue. 18. Roy M, Li Z, Sacks DB. IQGAP1 binds ERK2 and modulates its activity. Mol Cell Biochem 2007;294:237–42. J Biol Chem 2004;279:17329–37. 39. Zhao H, Li Q, Wang J, Su X, Ng KM, Qiu T, et al. Frequent epigenetic 19. Smith JM, Hedman AC, Sacks DB. IQGAPs choreograph cellular signaling silencing of the folate-metabolising gene cystathionine-beta-synthase in from the membrane to the nucleus. Trends Cell Biol 2015;25:171–84. gastrointestinal cancer. PLoS One 2012;7:e49683. 20.MeisterM,TomasovicA,BanningA,TikkanenR.Mitogen-activated 40. Kim J, Hong SJ, Park JH, Park SY, Kim SW, Cho EY, et al. Expression of protein (MAP) kinase scaffolding proteins: a recount. Int J Mol Sci cystathionine beta-synthase is downregulated in hepatocellular carcinoma 2013;14:4854–84. and associated with poor prognosis. Oncol Rep 2009;21:1449–54. 21. Fink SP, Myeroff LL, Kariv R, Platzer P, Xin B, Mikkola D, et al. Induction of 41. Vial E, Marshall CJ. Elevated ERK-MAP kinase activity protects the FOS KIAA1199/CEMIP is associated with colon cancer phenotype and poor family member FRA-1 against proteasomal degradation in colon carcino- patient survival. Oncotarget 2015;6:30500–15. ma cells. J. Cell Sci. 2003;116:4957–63. 22. Xu J, Liu Y, Wang X, Huang J, Zhu H, Hu Z, et al. Association between 42. Teis D, Wunderlich W, Huber LA. Localization of the MP1-MAPK scaffold KIAA1199 overexpression and tumor invasion, TNM stage, and poor complex to endosomes is mediated by p14 and required for signal prognosis in colorectal cancer. Int J Clin Exp Pathol 2015;8:2909–18. transduction. Dev Cell 2002;3:803–14. 23. Shostak K, Zhang X, Hubert P, Goktuna SI, Jiang Z, Klevernic I, et al. NF- 43. Sousa LP, Lax I, Shen H, Ferguson SM, De Camilli P, Schlessinger J. kappaB-induced KIAA1199 promotes survival through EGFR signalling. Suppression of EGFR endocytosis by dynamin depletion reveals that EGFR Nat Commun 2014;5:5232. signaling occurs primarily at the plasma membrane. Proc Natl Acad Sci 24. Evensen NA, Kuscu C, Nguyen HL, Zarrabi K, Dufour A, Kadam P, et al. U S A 2012;109:4419–24. Unraveling the role of KIAA1199, a novel endoplasmic reticulum protein, 44. Pinilla-Macua I, Watkins SC, Sorkin A. Endocytosis separates EGF receptors in cancer cell migration. J Natl Cancer Inst 2013;105:1402–16. from endogenous fluorescently labeled HRas and diminishes receptor 25. Birkenkamp-Demtroder K, Maghnouj A, Mansilla F, Thorsen K, Ander- signaling to MAP kinases in endosomes. Proc Natl Acad Sci U S A sen CL, Oster B, et al. Repression of KIAA1199 attenuates Wnt-signalling 2016;113:2122–7. and decreases the proliferation of colon cancer cells. Br J Cancer 45. Lavoie H, Therrien M. Regulation of RAF protein kinases in ERK signalling. 2011;105:552–61. Nat Rev Mol Cell Biol 2015;16:281–98. 26. Hatzis P, van der Flier LG, van Driel MA, Guryev V, Nielsen F, Denissov S, 46. Eden ER, White IJ, Tsapara A, Futter CE. Membrane contacts between et al. Genome-wide pattern of TCF7L2/TCF4 chromatin occupancy in endosomes and ER provide sites for PTP1B-epidermal growth factor colorectal cancer cells. Mol Cell Biol 2008;28:2732–44. receptor interaction. Nat Cell Biol 2010;12:267–72. 27. Hendriks MM, Smit S, Akkermans WL, Reijmers TH, Eilers PH, Hoefsloot 47. Shen A, Wang L, Huang M, Sun J, Chen Y, Shen YY, et al. c-Myc alterations HC, et al. How to distinguish healthy from diseased? Classification strategy confer therapeutic response and acquired resistance to c-Met inhibitors in for mass spectrometry-based clinical proteomics. Proteomics 2007;7: MET-addicted cancers. Cancer Res 2015;75:4548–59. 3672–80. 48. Locasale JW.Serine, glycine and one-carbon units: cancer metabolism in 28. Xia J, Sinelnikov IV, Han B, Wishart DS. MetaboAnalyst 3.0–making full circle. Nat. Rev. Cancer 2013;13:572–83. metabolomics more meaningful. Nucleic Acids Res 2015;43:W251–7. 49. Tsun ZY, Possemato R. Amino acid management in cancer. Semin Cell Dev 29. Van der Flier LG, Sabates-Bellver J, Oving I, Haegebarth A, De Palo M, Anti Biol 2015;43:22–32. M, et al. The Intestinal Wnt/TCF Signature. Gastroenterology 2007;132: 50. Yuan HX, Xiong Y, Guan KL. Nutrient sensing, metabolism, and cell growth 628–32. control. Mol Cell 2013;49:379–87.

4548 Cancer Res; 78(16) August 15, 2018 Cancer Research

The Endosomal Protein CEMIP Links WNT Signaling to MEK1− ERK1/2 Activation in Selumetinib-Resistant Intestinal Organoids

Hong Quan Duong, Ivan Nemazanyy, Florian Rambow, et al.

Cancer Res 2018;78:4533-4548. Published OnlineFirst June 18, 2018.

Updated version Access the most recent version of this article at: doi:10.1158/0008-5472.CAN-17-3149

Supplementary Access the most recent supplemental material at: Material http://cancerres.aacrjournals.org/content/suppl/2018/06/16/0008-5472.CAN-17-3149.DC1

Cited articles This article cites 50 articles, 18 of which you can access for free at: http://cancerres.aacrjournals.org/content/78/16/4533.full#ref-list-1

Citing articles This article has been cited by 2 HighWire-hosted articles. Access the articles at: http://cancerres.aacrjournals.org/content/78/16/4533.full#related-urls

E-mail alerts Sign up to receive free email-alerts related to this article or journal.

Reprints and To order reprints of this article or to subscribe to the journal, contact the AACR Publications Department at Subscriptions [email protected].

Permissions To request permission to re-use all or part of this article, use this link http://cancerres.aacrjournals.org/content/78/16/4533. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.