Upregulation of E3 Ubiquitin Ligase CBLC Enhances EGFR

Published OnlineFirst June 26, 2018; DOI: 10.1158/0008-5472.CAN-17-3858

Cancer Tumor Biology and Immunology Research

Upregulation of E3 Ubiquitin Ligase CBLC Enhances EGFR Dysregulation and Signaling in Lung Adenocarcinoma Shiao-Ya Hong1, Yu-Rung Kao1, Te-Chang Lee1, and Cheng-Wen Wu1,2,3,4

Abstract

CBLC (CBL proto-oncogene c) belongs to the CBL protein ubiquitinated and positively regulated aEGFR stability family, which has E3 ubiquitin ligase activity toward activated through the conjugation of polyubiquitin by K6 and K11 receptor tyrosine kinases. CBLC is frequently upregulated in linkages. This CBLC-mediated polyubiquitination promoted non–small cell lung cancer (NSCLC), yet very little is known either preferential recycling of aEGFR back to the plasma about the functions of CBLC in tumorigenesis. Here we show membrane or trafficking to the cell nucleus. IHC analyses that CBLC is an epigenetically demethylated target and its revealed a positive correlation between phospho-EGFR and expression can be upregulated in NSCLC after treatment with CBLC in lung adenocarcinoma. In summary, we demonstrate a the DNA methylation inhibitor 50-azacytidine. Depletion novel mechanism by which aEGFR escapes lysosomal degra- of CBLC significantly inhibited cell viability and clonogenicity dation in a CBLC/ubiquitin-dependent manner to sustain its in vitro and reduced tumor growth in a xenograft model. CBLC activation. Our work identifies CBLC as a potential diagnostic silencing further sensitized EGFR-mutated NSCLC cells to biomarker and also points to its utilization as a novel thera- treatment with tyrosine kinase inhibitors. Conversely, ectopic peutic target for NSCLC therapy. expression of CBLC enhanced the activation of EGFR and downstream ERK1/2 signaling after ligand stimulation by Significance: This work demonstrates the role of CBLC expres- competing with CBL for EGFR binding. Analysis of ubiquitin sion as a diagnostic biomarker and potential therapeutic target in linkages on activated EGFR (aEGFR) revealed that CBLC lung adenocarcinoma. Cancer Res; 78(17); 4984–96. 2018 AACR.

Introduction a linker region, and a RING finger motif. CBLC differs from CBL and CBLB in that the latter two both have a proline-rich region, The activation of tyrosine kinase receptors (RTK) engages a wide several regulatory tyrosine phosphorylation sites, and an ubiqui- range of signaling pathways that regulate cellular processes and tin-associated (UBA) domain in their C-termini (2). While CBL functions (1, 2). Aberrant activation of RTKs can be triggered by and CBLB are ubiquitously expressed, the expression of CBLC is genetic alterations (e.g., amplification, somatic mutation, and specifically restricted to normal epithelial cells (5). deletion/insertion) or evading lysosomal degradation. The EGFR, The ubiquitination mediated by CBL can negatively regulate an important member of RTKs, is frequently hyperactivated in a EGFR signaling through lysosomal degradation. Upon ligand variety of cancers (3). Dysregulation of EGFR signaling can lead to stimulation, CBL binds to activated EGFR (aEGFR) through its uncontrolled cell proliferation,tumorigenesis, and cancer progres- TKB domain (6). Subsequently, CBL-interacting protein of 85 kDa sion (3). (CIN85) and endophilin are recruited to EGFR complexes by The CBL family of E3 ligases has emerged as a key negative tyrosine phosphorylation of CBL. This process has been found to regulator of activated RTKs. CBL family proteins belong to the be required for EGFR internalization into endosomal sorting (7). RING finger class of E3 ligases, which catalyze the transfer of The ubiquitin modification mediated by CBL then can serve as a ubiquitin from E2 to substrates (4). The CBL family proteins (CBL, signal for directing aEGFR toward lysosomal degradation (8). CBLB, and CBLC in mammals) share a highly conserved N- However, accumulating evidence also links CBL protein family to terminus consisting of a tyrosine-kinase binding (TKB) domain, a potential role of oncogenic driver in cancer progression. Muta- tions in CBL are associated with approximately 5% of myeloid 1Institute of Biomedical Sciences, Academia Sinica, Taipei, Taiwan. 2Institute of neoplasms. The mutant proteins fail to ubiquitinate RTKs but Biochemistry and Molecular Biology, National Yang-Ming University, Taipei, retain the adaptor function to activate downstream PI3K pathway. 3 Taiwan. Institute of Microbiology and Immunology, National Yang-Ming CBL mutations in the RING finger motif lost its E3 activity, 4 University, Taipei, Taiwan. Institute of Clinical Medicine, National Yang-Ming assuming an oncogenic role in cancer progression (9). University, Taipei, Taiwan. In comparison with CBL, the genetic variations of CBLC are less Note: Supplementary data for this article are available at Cancer Research common and little is known about its functions in tumorigenesis. Online (http://cancerres.aacrjournals.org/). Although previous reports suggest that CBLC may as well serve as a Corresponding Author: Cheng-Wen Wu, Institute of Biomedical Sciences, negative regulator of aEGFR (1, 5), it has been revealed that CBLC Academia Sinica, 128 Section 2, Academia Road, Nankang, Taipei 11529, Taiwan. neither interacts nor colocalizes with CIN85 in mammalian cells Phone: 8862-2652-3015; Fax: 8862-2652-3075; E-mail: [email protected] due to the absence of the distal part of C-terminus found in CBL or doi: 10.1158/0008-5472.CAN-17-3858 CBLB proteins (10). Because CBL–CIN85–endophilin interaction 2018 American Association for Cancer Research. is required to initiate the endocytosis and degradation of EGFR,

4984 Cancer Res; 78(17) September 1, 2018

CBLC Dysregulates EGFR Signaling

the exact role of CBLC in regulating aEGFR turnover remains facturer's protocol. UPL probe-based qPCR (Roche) was carried elusive. Here, we identified that CBLC is an epigenetically out on a LightCycler 480 and analyzed using the manufacturer's demethylated target and its expression can be upregulated after software. The following primers and UPL probes were used for treatment with a DNA methylation inhibitor, 50-azacytidine in amplification: EGFR forward: 50-CATGTCGATGGACTTCCAGA- non–small cell lung cancer (NSCLC) cells. Our data revealed that 30, reverse: 50-GGGACAGCTTGGATCACACT-30, probe: 44; CBL CBLC knockdown renders EGFR-mutant NSCLC cells more sen- forward 50-TGACGTATGACGAAGTGAAAGC-30, reverse: 50- sitive to tyrosine kinase inhibitor (TKI) treatments, probably CAGCTCAGCCGGAAGATATAA-30, probe: 50; CBLB forward: through reducing the protein stability of aEGFR. We further found 50- GTGCACCTCTTGCCTTACG-30, reverse: 50-CCTTTTATTTCA- that CBLC overexpression reduces CBL-mediated ubiquitination CAACGACAGAAA, probe: 51; CBLC forward: 50-TGCTGAGAG- and degradation of aEGFR through the conjugation of ubiquitin CAACAAGGATG-30, reverse: 50-TCTGGCTGTCCGAGTGCT-30, K6 and K11 linkages. The CBLC-mediated stabilization of aEGFR probe: 71; b-actin forward: 50-AGAGCTACGAGCTGCCTGAC- is preferentially destined for the recycling program, leading to 30, reverse: 50-CGTGGATGCCACAGGACT-30, probe: 9. sustained EGFR activation in cancer cells. Immunoblotting Cultured cells were lysed with RIPA buffer (50 mmol/L Tris Materials and Methods HCl, pH 7.4, 150 mmol/L NaCl, 2 mmol/L EDTA, 1% NP-40, 1% Database and statistical analyses sodium deoxycholate, 0.1% SDS) supplemented with protease The level of CBLC promoter methylation was extracted from and phosphatase inhibitors (Roche). The protein concentration TCGA and analyzed by GraphPad Prism. Other statistical analyses was determined using the Bio-Rad protein assay Kit (Bio-Rad). were also performed using GraphPad Prism. P values less than Lysates were separated in a 8%–12% SDS-PAGE, transferred to 0.05 were considered significant. nitrocellulose membranes, incubated with specific antibodies overnight, and developed by enhanced chemiluminescence. Cell lines, plasmids, and antibodies Images were acquired with the LAS-3000 system (Fujifilm). The human lung normal cells (BEAS2B, NL-20), lung cancer cells (A549, H358, H1437, H3255, H1975, HCC827, H2170, Genomic DNA preparation and bisulfite conversion PCR H226, H520), and HEK293T were from ATCC. All cell lines were Genomic DNA from cultured cells was isolated by using a authenticatedbycytogeneticanalysisfromthisestablishedprovider. commercially available DNA Extraction kit (Macherey-Nagel) They were maintained according to ATCC's instructions and used followed by the corresponding manufacturer's protocols. DNA within 6 months after resuscitation. H928 lung adenocarcinoma samples were bisulfite converted using EpiTect Bisulfite Conver- cells were authenticated and provided by the National Health sion Kit (Qiagen) following the manufacturer's instructions. The Research Institute (Taipei, Taiwan), and cultured in RPMI1640 resulting modified DNA was amplified by PCR using methylation- medium supplemented with 10% FBS and 1% penicillin/strepto- specific and unmethylation-specific primer sets. The primer mycin.AllcellswereroutinelydetectedasMycoplasma-freebytheEZ- sequences were as follows: methylated forward primer: 50- PCR Mycoplasma Test Kit (Biological Industries). CBLC cDNA was TTTTTTTCGGTTTTTTAGGATTCGT-30, methylated reverse primer: cloned into pCDNA3/HA or pEGFP-C2 vector. Rab family (RAB5, 50-TAAACCACCTCTCGAAACAACTACG-30; unmethylated for- RAB7, and RAB11) cDNA was cloned into pLAS3w.RFP-C.Ppuro ward primer: 50-TTTTTTTTGGTTTTTTAGGATTTGT-30, unmethy- vector, individually. pCW7 containing human ubiquitin with His- lated reverse primer: 50-AAACCACCTCTCAAAACAACTACACT-30. Myc tag and pLXSN containing human EGFR cDNA with HA-tag constructs were kind gifts from Dr. Hui-Kuan Lin (Wake Forest Cell proliferation, colony formation, and drug sensitivity assay School of Medicine, Winston-Salem, NC) and Dr. Albert Wong For cell proliferation assay, cells were cultured at a density of (Thomas Jefferson University, Philadelphia, PA). Plasmids were 2 103 cells per well in flat-bottomed 96-well plates. AlamarBlue mutated using the site-directed mutagenesis method. All plasmids was applied to assess the cell viability by measuring the fluores- were identified by Sequencing Core Facility of the ScientificInstru- cence at 544 nm (excitation) and 590 nm (emission). For colony ment Center at Academia Sinica (Taipei, Taiwan). TRC lentiviral formation assay, 1 103 cells were seeded in 6-well plates shRNAs targeting human CBLC (TRCN 4273 and 4274) and CBL and colonies were fixed with formaldehyde and stained with (TRCN 288695)was obtainedfromthe National RNAiCoreFacility crystal violet after 2–3 weeks. For drug sensitivity assay, cells at Academia Sinica (Taipei, Taiwan). Plasmids were transfected by (2 103 cells/well) were treated with indicated concentrations PolyJet (Signagen) according to the manufacturer's manual. Mono- of TKI (gefitinib or osimertinib) for 72 hours and viable cells were clonal anti-EGFR (immunoprecipitation: Ab-3, Oncogene Science; measured by AlamarBlue assay. All assays were performed in or immunoblotting/staining: EP38Y, Abcam), phosphoY1068- triplicate and independently repeated two to three times. EGFR (Y38, Abcam), CBLC (10F4.2, Millipore), CBL (D4E10, Cell Signaling Technology), pTyr (P-Tyr-100, Cell Signaling Technolo- Xenograft models gy), pAKT (EP2109Y, Abcam), HA (16B12, Covance), and poly- For tumor growth studies, HCC827 cells (1 106) were clonal anti-pERK1/2 (GTX59568, GeneTex), GFP (sc-8334, Santa suspended in 200 mL of 50% Matrigel in PBS and injected Cruz Biotechnology), GAPDH (GTX100118, GeneTex) antibodies subcutaneously into the both flanks of male Balb/c nude mice. were purchased commercially. For pharmacodynamic studies, 1 106 of shCtrl cells and 2 106 of shCBLC cells were injected subcutaneously into mice. Treat- RNA extraction and qRT-PCR ment with gefitinib or osimertinib (formulation 0.5% methyl The RNA was extracted from cell lines using GENEzol reagent cellulose, 0.1% Tween 80) started 20–25 days after implant when (Geneaid) and then reverse transcribed using PrimeScript RT the average volume of tumors reached approximately 350–450 reagent Kit (Perfect Real Time; Takara) according to the manu- mm3. Animals were administered once daily with vehicle,

www.aacrjournals.org Cancer Res; 78(17) September 1, 2018 4985

Hong et al.

gefitinib (Iressa, AstraZeneca; 5 mg/kg, oral), or osimertinib revealed that CBLC is expressed in a variety of lung cancer cells (Tagrisso, AstraZeneca; 5 mg/kg, oral) for the duration of the (50%), including four lung adenocarcinoma (LUAD) cell lines study. Tumor diameters were measured with digital calipers every (H358, H1437, H1975, and HCC827) and one lung squamous week, and the tumor volume in mm3 was calculated by the cell carcinoma (LUSC) cell line (H2170;Supplementary formula, volume ¼ (width)2 length/2. All animal procedures Fig. S1A). We next analyzed public cancer datasets from The were approved by the Institutional Animal Care and Use Com- Cancer Genome Atlas (TCGA) to investigate the mRNA expres- mittee at Academia Sinica (Taipei, Taiwan). sion of CBL family proteins in two major subtypes of NSCLC, LUAD and LUSC. In concord with the results from in vitro IHC assay analyses, the expression of CBLC was specifically upregulated in fi Xenograft tumors were resected on day 42 and xed in 4% LUAD and LUSC datasets, while the expression of CBL and fi paraformaldehyde overnight. Tumors were paraf n-embedded, CBLB was not significantly changed (Supplementary Fig. S1B). sectioned, and dewaxed by the Pathology Core Laboratory, IBMS, We then further analyzed the clinical prognoses in the two Academia Sinica (Taipei, Taiwan). Tissue microarray slides TCGA cohorts of patients with lung cancer to examine the ware purchased from Shanghai Outdo Biotech Company correlation between high CBLC expression and overall survival. (LugA090Ly02). Antigen retrieval was performed by autoclave Interestingly, high expression of CBLC was significantly corre- heating in 10 mmol/L citrate buffer (pH 6.0) for 20 minutes. lated with poor overall survival in patients with LUAD (Sup- Slides were IHC stained using Novolink Polymer Detection System plementary Fig. S1C), whereas no significant difference was according to the manufacturer's instructions. The sections were found in LUSC (Supplementary Fig. S1D). Because the TCGA then counterstained with hematoxylin, dehydrated, and mounted data suggested very few genetic modifications of CBLC (0.43% by the Pathology Core Laboratory. The primary antibodies used for of mutation and 4.35% of amplification found in LUAD; 3.37% IHC assay were against Ki-67 (Abcam ab92742, 1:500), cleaved of amplification found in LUSC), we investigated whether the caspase-3 (Cell Signaling Technology 9964, 1:1,000), CD31 (Cell enhanced expression of CBLC is a function of reduced meth- Signaling Technology 77699, 1:100), CBLC (Rockland 600-401- ylation in the CBLC promoter of cancer cells. By employing 889, 1:1,000), and pEGFR (Abcam Y38, 1:50). MethPrimer software analysis (12), we found that the CBLC promoter is rich in guanine-cytosine (GC) content and contains Immunoprecipitation, coimmunoprecipitation, and Ni-NTA a single CpG island of 851 bp (Fig. 1A). The methylation status pull down of CBLC promoter was next examined by performing bisulfite The lysis buffer for immunoprecipitation and coimmunopre- conversion PCR analyses. The CpG sites in the CBLC promoter cipitation was 1% and 0.1% NP-40 lysis buffer (50 mmol/L Tris- region were less methylated in five CBLC-positive cell lines HCl, pH 7.4, 150 mmol/L NaCl, 2 mmol/L EDTA), while the (H1437, H1975, HCC827, H358, and H2170), compared with buffer for Ni-NTA pull down was the Triton X-100/imidazole lysis those in two normal (BEAS-2B and NL20) and five CBLC- buffer (50 mmol/L Tris-HCl, pH 7.4, 150 mmol/L NaCl, 2 mmol/L negative cancer cell lines (Fig. 1B). We subsequently treated EDTA, 1% Triton X-100, and 10 mmol/L imidazole). Briefly, cell five CBLC-negative cells (A549, H928, H3255, H226, and lysates of each 60-mm dish were prepared in 500 mL lysis buffer H520) with 50-azacytidine (aza), a widely used DNA methyl- supplemented with 1 protease and phosphatase inhibitor ation inhibitor, for 24 hours. Interestingly, the mRNA expres- cocktail (Roche Applied Science). The cell lysates were then incu- sion of CBLC was significantly upregulated (1.6- to 5-fold) by bated on ice for 20 minutes and centrifuged at 13,000 rpm for 10 0 5 -aza treatment (Fig. 1C). In addition, we analyzed the methy- minutes at 4 C. The supernatants were collected and incubated lome and RNASeq in LUAD and LUSC datasets from TCGA with specific antibody conjugated or Ni-NTA beads for 1 hour at database. The overall mean percentage of methylation, calcu- 4 C with agitation. Beads were washed three times by using lysis lated from the average of nine probes in the CBLC promoter, buffer supplemented with 300 mmol/L NaCl, and pellets were was 34.16% 8.877% in LUAD tumor samples (n ¼ 435), boiled for 5 minutes in 2 SDS sample buffer for the analysis. while the value was 41.53% 4.152% in adjacent normal Immunofluorescence and confocal microscopy samples (n ¼ 29). The value was 29.11% 8.274% in LUSC Cells were grown on Lab-Tek chamber slides (Nunc), fixed with tumor samples (n ¼ 461), compared with 40.58% 3.417% in 4% paraformaldehyde in PBS for 10 minutes, and permeabilized adjacent normal samples (n ¼ 41). These results revealed that with 0.1% Triton X-100 in PBS for 5 minutes. After blocking with the percentage of methylation in the CBLC promoter region was 5% goat serum for 30 minutes, slides were incubated overnight at significantly lower in LUAD and LUSC tumor samples, com- 4C with rabbit anti-EGFR (EP38Y) antibody (1:250; Abcam) pared with that in adjacent normal samples (P < 0.0001; Fig. followed by incubation for 1 hour with secondary antibodies 1D). We further analyzed the methylation of CBLC promoter in conjugated with Alexa Fluor 594 or 670 (1:500; Jackson Immu- the 29 tumors and their paired normal samples from patients noresearch). Nuclei were counterstained with 40, 6-diamidino-2- with LUAD and the 41 paired samples from patients with phenylindole (DAPI). The slides were mounted with ProLong LUSC. The results also showed that the majority of tumor Diamond (Invitrogen) overnight at room temperature and samples (27/29 in LUAD and 39/41 in LUSC) exhibited a observed using a Zeiss LSM 700 laser confocal microscope. lower methylation level in the CBLC promoter region, com- paredwiththatintheCBLC promoter region of their paired normal tissues (P < 0.0001; Fig. 1E). By analyzing the probe Results (cg23635599) located in the TSS200 of CBLC gene, we CBLC is an epigenetically demethylated target in lung cancer observed a significant negative correlation between mRNA Given that CBLC is overexpressed in NSCLC (11), we exam- expression and the level of methylation in TCGA LUAD (n ¼ ined the expression of CBL family proteins in NSCLC cell 477) and LUSC (n ¼ 378) cohorts, further confirming that the lines. Our qRT-PCR experiments and immunoblotting analyses hypomethylationintheCBLC promoter region corresponds to

4986 Cancer Res; 78(17) September 1, 2018 Cancer Research

CBLC Dysregulates EGFR Signaling

Figure 1. CBLC expression is upregulated through promoter hypomethylation in NSCLC. A, The prediction of CpG islands in CBLC promoter using MethPrimer software. Criteria used for prediction were island size >100 bp, GC percentage >50%, and observed/expected CpG ratio > 0.6. Location of CpG island (light gray), promoter regions (gray), and methylated (84þ194) and unmethylated (83þ193) specific-amplified regions are shown. Indicated positions are relative to the transcription start site (þ1). B, Methylation-specific PCR analysis of CBLC promoter. Bisulfite-converted genomic DNA extracted from lung normal and cancer cells was amplified with methylated DNA-specific primers (M) or with unmethylated DNA-specific primers (U). The methylated and unmethylated DNA- specific product (278 bp) was separated on 2% TAE agarose gels and visualized by ethidium bromide staining. C, Relative CBLC mRNA expression in five CBLC- negative cells treated by 50-aza. After treating cells with 10 mmol/L of 50-aza for 24 hours, the changes in CBLC expression levels were measured by qPCR (n ¼ 3). D, The hypomethylation of CBLC in LUAD or LUSC tumor and normal samples. DNA methylation is shown as the ratio of methylated DNA to total DNA. E, CBLC methylation analyses in tumor and normal paired LUAD (n ¼ 29) and LUSC (n ¼ 41) clinical samples. Tumor tissues were significantly hypomethylated when compared with their matched normal tissues (, P < 0.05; , P < 0.01; , P < 0.001). its mRNA upregulation in LUAD and LUSC (r ¼0.470 and r ¼ significant difference was observed (Supplementary Fig. S2). 0.447, P < 0.0001; Supplementary Fig. S2). CBLC hypomethyla- These findings indicated that hypomethylation in the CBLC tion was also analyzed based on age, gender, lymphocyte infil- promoter region may contribute to the overexpression of CBLC tration, pathologic stage, and tobacco smoking history, but no in LUAD and LUSC.

www.aacrjournals.org Cancer Res; 78(17) September 1, 2018 4987

Hong et al.

Figure 2. Silencing CBLC leads to reduced cell viability, clonogenicity, and tumor growth. A, The effect of CBLC expression on cell viability of HCC827 and H1975 cells. HCC827 or H1975 cells were infected with lentiviral control shRNA or two different shRNA against CBLC. After selection for three days with 2 mg/mL puromycin, cell growth was monitored by employing AlamarBlue assay at day 1, 3, and 5, using day 1 as reference. The corresponding level of CBLC protein in the control and CBLC knockdown cells was also analyzed by immunoblotting. B, Colony formation (left) and quantification (right) in CBLC knockdown HCC827 or H1975 cells. A–B, The data shown are the means of three independent experiments. Error bars, SEMs. C, Tumor growth of HCC827 cells expressing control and two independent CBLC shRNAs in xenograft models (each group n ¼ 8). Left, growth curve of subcutaneous tumors from HCC827 cells expressing indicated shRNAs in the first 6 weeks after inoculation. Right, image of the dissected tumors at day 42. The harvested tumors were subjected to immunoblotting (D) and IHC staining (E). D, Representative blots showing the efficiency of CBLC knockdown in xenograft tumors at day 42. E, Representative (left) and quantitative (right) graphs of Ki67, cleaved caspase-3 (cCasp3), and CD31-stained tumor section. Scale bar, 100 mm. Ki-67 and cleaved caspase-3 are expressed as percentage of positive cells per field and CD31 is expressed as positive vessels per field (400) from eight fields derived from four sections per group (, P < 0.05; , P < 0.01; , P < 0.001). ns, nonsignificant.

Targeting CBLC E3 ubiquitin ligase may have therapeutic blotting (Fig. 2D). Histologic examination of xenografts also potential in lung cancer treatment revealed a decrease (50%) in cell proliferation (Ki-67 staining) To determine whether CBLC could be a passenger or oncogenic and an increase (4 fold) in apoptosis (cleaved caspase-3 stain- driver involved in promoting lung cancer initiation and/or pro- ing) in shCBLC tumors from animals, but no marked difference in gression, we examined the growth of tumor cells using shRNA tumor vasculature (CD31 staining) was observed (Fig. 2E). These knockdown system. Knockdown of CBLC by employing two results suggested that inhibition of CBLC by shRNA was associ- independent shRNAs against CBLC resulted in a modest decrease ated with reduced cell proliferation and survival in HCC827 (33%–48%) in cell viability (Fig. 2A) and clonogenicity (Fig. 2B) xenografts. Together, these findings highlighted the role of CBLC in HCC827 and H1975 cells. Of note, these effects were not in promoting tumorigenesis and the potential of targeting CBLC limited to EGFR-mutant LUAD cells, as CBLC knockdown also in lung cancer treatment. decreased the viability of EGFR wild-type LUAD (H358 and H1437) and LUSC (H2170) cells (data not shown). In contrast, Depletion of CBLC renders EGFR-mutant LUAD cells more in two EGFR wild-type A549 and H520 cells, ectopic expression of sensitive to TKIs CBLC significantly promoted the colony formation in soft agar As the CBL family of E3 ubiquitin ligases is critical for the (2-fold; data not shown). Next, we further examined the effects regulation of aEGFR signaling, we tested whether CBLC silencing of silencing CBLC on tumor growth in vivo. CBLC knockdown affects the responses to TKIs in EGFR-mutant cells. HCC827 cells, showed approximately 60% delayed tumor growth in the harboring EGFR exon19 deletion, were sensitive to gefitinib with HCC827 xenografts compared with the control group (n ¼ 8; the IC50 value of 0.0337 0.0084 (mmol/L). Unexpectedly, CBLC Fig. 2C). Knockdown efficiency of HCC827 xenografts at the end silencing further sensitized cells to gefitinib treatment (IC50 of of the study was confirmed on CBLC protein level by immuno- 0.0032 0.0007 mmol/L, P < 0.05; Fig. 3A). Similar effects were

4988 Cancer Res; 78(17) September 1, 2018 Cancer Research

CBLC Dysregulates EGFR Signaling

Figure 3. CBLC knockdown increased the TKI sensitivity in LUAD cells with EGFR mutation. A and B, Cytotoxicity of EGFR-TKIs in control and CBLC knockdown cells. HCC827 (A) and H1975 (B) cells were treated with the indicated concentrations of gefitinib or osimertinib for 72 hours. Cell viability was determined using the AlamarBlue assay (n ¼ 3). C, The effect of CBLC deficiency on tumor growth in HCC827 xenograft models with gefitinib treatment. Mice bearing HCC827 xenografts expressing control or CBLC shRNAs were orally administrated daily with 5 mg/kg gefitinib (n ¼ 10) or vehicle (n ¼ 6) for the indicated time frame. D, The combined activity of CBLC knockdown and osimertinib treatment in H1975 xenograft models. Mice bearing H1975 xenografts expressing control or CBLC shRNAs were orally administrated daily with 5 mg/kg osimertinib (n ¼ 6 each group) for the indicated time frame. E, Total and phosphorylated levels of EGFR protein in HCC827 or H1975 control and CBLC knockdown cells. Three independent experiments were conducted. F, The mRNA level of EGFR in HCC827 and H1975 after CBLC knockdown (n ¼ 3). Data are represented as mean SEM (, P < 0.01; , P < 0.001). ns, nonsignificant. also observed in osimertinib-treated H1975 cells, which harbor in two EGFR wild-type H358 and H1437 cells did not show a EGFR-L858R/T790M mutation. While H1975 cells treated with significant difference in response to TKI treatments (data not osimertinib showed the IC50 of 0.0101 0.0010 mmol/L, knock- shown). Seeing that knockdown of CBLC increased the sensitivity down of CBLC resulted in an even lower IC50 at 0.0038 0.0009 of TKIs in LUAD cells with EGFR mutation in vitro, we next mmol/L (P < 0.05, ANOVA; Fig. 3B). However, CBLC knockdown investigated whether the promising results could translate into

www.aacrjournals.org Cancer Res; 78(17) September 1, 2018 4989

Hong et al.

observed benefits in tumor xenograft models. To assess the lead to changes in structural and signaling properties of substrates combined activity with CBLC knockdown, the TKIs were given (17), we compared the differences in ubiquitin linkages of aEGFR orally at a dose of 5 mg/day. As shown in Fig. 3C, treatment of assembled by CBL and CBLC in CBL knockdown cells. While mice with 5 mg/kg gefitinib strongly reduced tumor growth mutations of ubiquitin at K27, K33, or K63 residues interfered (relative to vehicle treatment). Importantly, compared with gefi- with the polyubiquitination of aEGFR by CBL and CBLC (Fig. 5C tinib monotherapy, expression of CBLC shRNAs in combination and D), only the replacement of ubiquitin at K6 and K11 residues with gefitinib treatment further inhibited tumor growth in selectively prohibited the formation of polyubiquitin conjugates HCC827 xenograft models (P < 0.001, ANOVA). Interestingly, on aEGFR mediated by CBLC (Fig. 5E). Indeed, the replacement of the combined activity of CBLC knockdown and gefitinib became ubiquitin at K6 and K11 residues impaired the ability of CBLC to more visible after 25-day treatment, probably when EGFR was stabilize aEGFR (Fig. 5F). These results demonstrated that CBLC ineffectively inhibited by gefitinib (13). The increased sensitivity reduces CBL-mediated ubiquitination and degradation of aEGFR to TKIs with CBLC knockdown was also observed using 5 mg/kg of through the conjugation of polyubiquitin by K6 and K11 linkages. osimertinib in H1975 xenograft models (Fig. 3D). To reveal the possible mechanism through which CBLC regulates the responses CBLC promotes aEGFR trafficking to the nucleus and recycling to TKI treatments, we examined the EGFR expression and its to the plasma membrane phosphorylation in CBLC knockdown cells. Surprisingly, the Because aEGFR escapes ubiquitin-dependent degradation via levels of total and phosphorylated EGFR were reduced simulta- fast recycling to plasma membrane in cancer cells (18), we neously in CBLC-deficient cells (Fig. 3E). However, no significant examined whether CBLC binds aEGFR and regulates its cellular difference was found in the expression of EGFR mRNA in the trafficking. In EGFR wild-type H1437 cells, ectopically expressed presence or absence of CBLC (Fig. 3F). These results suggested that CBLC was widely distributed in the cytoplasm and nucleus, CBLC deficiency might enhance TKI responses through reducing whereas inactive EGFR was prominently localized to the plasma EGFR protein stability in a posttranscriptional manner. membrane (Fig. 6A). Interestingly, the addition of EGF facilitated the colocalization of CBLC and aEGFR at the perinuclear regions CBLC sustains aEGFR signaling by competing with CBL for (within 3 minutes) and subsequently to the nucleus (30 minutes). aEGFR binding In contrast, EGF treatment in control cells induced EGFR endo- To determine whether CBLC contributes to the prohibited cytosis, thus explaining the intense cytoplasmic, perinuclear, and protein turnover of aEGFR, 293T cells were transfected with a nuclear punctate pattern observed. Note that cells ectopically HA-tagged CBLC construct and stimulated with EGF at several expressing CBLC showed more EGFR accumulation, regardless time points. The ectopic expression of CBLC not only prolonged of receiving EGF stimulation or not. The results implied that CBLC the half-life of total and activated EGFR, but also enhanced its attenuated the degradation of aEGFR and instead, shipped it to downstream pERK1/2 and pAKT (Fig. 4A and B). To elucidate the the nucleus and plasma membrane. Moreover, in EGFR-mutant molecular mechanisms underlying this observation, the interac- H1975 cells, CBLC depletion resulted in decreased accumulation tion of CBLC and aEGFR were confirmed by coimmunoprecipita- of aEGFR in the nucleus (Fig. 6B). Because RAB5, RAB7, and tion in 293T cells (Fig. 4C). Interestingly, the ectopic expression of RAB11 are the indicators of early, late, and recycling endosomes, CBLC not only increased the stability of pY1068-EGFR and respectively, we performed colocalization assays in EGFR-mutant amplified its downstream pERK1/2 signal, but also reduced the H1975 cells expressing RFP-tagged RAB proteins. In the cytosol, CBL binding to aEGFR. Collectively, these findings suggested that CBLC and aEGFR preferentially accumulated together in RAB11- CBLC stabilized aEGFR through binding competition with CBL, positive endosomes, but only partially in RAB5- or RAB7-positive which mediated the ubiquitination and degradation of aEGFR. endosomes (Fig. 6C–E).

CBLC stabilizes aEGFR through conjugating ubiquitin chains CBLC and pEGFR protein levels show a positive correlation in by K6 and K11 linkages lung adenocarcinoma samples CBL-mediated ubiquitination is an essential signal for targeting Because aforementioned findings suggested a role of CBLC in aEGFR to lysosomal degradation (14). Because CBLC competes aEGFR dysregulation, we further analyzed the correlation with CBL for binding to aEGFR, we investigated whether CBLC between CBLC and pEGFR protein levels in clinical samples by affects the ubiquitination of aEGFR mediated by CBL. As shown IHC staining. The histology score (H-score) was obtained by the in Fig. 5A, the ectopic expression of CBLC retarded the polyubi- Pannoramic density IHC quantification software, giving a range of quitination of aEGFR (Ub/tEGFR) induced by EGF within the first 0 to 300. As shown in Fig. 7A, CBLC was expressed at moderate or 5 minutes and lasted about 1 hour, after which the ubiquitinated high levels (having an H-score of 100–300) in 43.3% of LUAD aEGFR began to accumulate due to decreased degradation. It is tumors, while weak or no expression was detected (having an H- noteworthy that the ubiquitination level of aEGFR was slightly score of 0–100) in 56.7% of LUAD tumors (n ¼ 30 tumors). increased in the EGF-untreated group that ectopically expressed Meanwhile, pEGFR was upregulated (exhibiting an H-score of CBLC. To rule out the possibility that the inhibition of polyubi- 100–300) in 50.0% of LUAD tumors. Among them, 30% of quitination mediated by CBLC is caused by a dominant negative samples showed both elevated CBLC and pEGFR, whereas effect, we tested the E3 ligase activity of CBLC in CBL knockdown 36.7% of samples displayed both low CBLC and pEGFR (P ¼ cells. The CBLC expression indeed ubiquitinated and positively 0.0656). These data indicated that CBLC and pEGFR are coupre- regulated aEGFR stability, even in CBL-silenced cells (Fig. 5B, gulated in LUAD tumors and suggested a positive correlation right). Previous quantitative mass spectrometric analyses have between CBLC and pEGFR in LUAD cancer. shown that CBL-mediated polyubiquitination of aEGFR is pri- Overall, our results indicated that upregulation of CBLC marily linked through K63, and secondarily through K48 and K11 reduces CBL-mediated ubiquitination and degradation of aEGFR (14–16). Given that alternatively linked polyubiquitin chains through the conjugation of ubiquitin by K6 and K11 linkages.

4990 Cancer Res; 78(17) September 1, 2018 Cancer Research

CBLC Dysregulates EGFR Signaling

Figure 4. CBLC sustained aEGFR signaling by competing with CBL for aEGFR binding. A and B, The downregulation of EGFR and its downstream pERK1/2, pAKT, pSTAT3 levels in the vector control or CBLC-overexpressing 293T cells. After transfection with empty vector or CBLC-expressing plasmid, 293T cells were starved overnight and then treated with 20 ng/mL EGF for the indicated times. Cell lysates were immunoblotted using antibodies against total and phospho-EGFR (tEGFR and pEGFR) or its downstream effectors, such as pERK1/2, pAKT, and pSTAT3. The representative immunoblots and densitometry (tEGFR, the intensity of the control group at time 0 was denoted as 1; pEGFR, pERK1/2, pAKT, and pSTAT3, the intensity of the control group at 5 minutes of stimulation was denoted as 1) are shown in A. The signals were quantiﬁed by densitometry, normalized for protein loading, and are represented as a ratio of maximal tEGFR level (the intensity of each group at time 0 was denoted as 1) or pEGFR, pERK1/2, pAKT, pSTAT3 activity (the intensity of each group at 5 minutes of stimulation was denoted as 1). Graphs that represent the half-life of tEGFR, pEGFR, pERK1/2, and pAKT (n ¼ 3) are shown in B. Data are represented as mean SEM (, P < 0.05; , P < 0.01). C, The binding competition of CBLC with CBL to aEGFR. The control or HA-CBLC–expressing 293T transfectants were starved overnight and then treated with or without 20 ng/mL EGF for 10 minutes. Cell lysates were subjected to immunoprecipitation using anti-HA or EGFR antibodies, and the immunoprecipitants were analyzed by immunoblotting using the indicated antibodies. Two independent experiments were conducted.

www.aacrjournals.org Cancer Res; 78(17) September 1, 2018 4991

Hong et al.

Figure 5. CBLC reduced CBL-mediated ubiquitination and degradation of aEGFR through conjugating ubiquitin chains by K6 and K11 linkages. A, The polyubiquitination of aEGFR in the presence of CBLC. HA-EGFR and His-Myc-tagged ubiquitin were cotransfected with vector control or CBLC in 293T cells. Transfected cells were starved overnight and then treated with 20 ng/mL EGF for the indicated times. Cell lysates were subjected to immunoblotting with anti-EGFR antibodies, either directly or after pull down with Ni-NTA beads. The intensity of tEGFR in the control group at time 0 was denoted as 1. The ratio of ubiquitin-modified EGFR to total EGFR (Ub/tEGFR) was also shown. Two independent experiments were conducted. B, The polyubiquitination of aEGFR affected by CBLC overexpression in 293T parental (left) or CBL-knockdown (right) cells. Procedures are the same as with A, except that cells were treated with 20 ng/mL EGF for 10 minutes. C and D, The heterotypic polyubiquitin chains of aEGFR assembled by CBL and CBLC. HA-EGFR and the single substitution of His–Myc–ubiquitin constructs (K6, K11, K27, K29, K33, K48, or K63) with arginine were cotransfected with CBL or CBLC in CBL-knockdown 293T cells. Ubiquitinated aEGFR was induced with 20 ng/mL EGF for 10 minutes. Whole-cell extracts or pull-down precipitates by Ni-NTA resins were analyzed by immunoblotting using anti-EGFR antibodies. E, The graph represents the ratio of ubiquitin-modified EGFR to total EGFR (Ub-EGFR/tEGFR) shown in B and C (n ¼ 3; , P < 0.001). F, The effect of CBLC on the downregulation of aEGFR with different ubiquitin mutations. Immunoblotting assays were performed using the same procedure as described in D, except that cells were stimulated with 20 ng/mL EGF for 1 hour and only whole-cell extracts were analyzed. The band intensity of EGFR was quantified by densitometry and normalized against that of GAPDH. Values are expressed relative to that in wild-type ubiquitin-transfected cells treated with EGF.

CBLC-mediated polyubiquitination of aEGFR might be destined promotes aEGFR trafﬁcking to the nucleus, or to the recycling for the recycling program and thus contributed to the sustained endosome and subsequently to the plasma membrane, leading to signaling in cancer cells. the increased stability and prolonged activation of aEGFR. Although CBLC expression is frequently detected in NSCLC tumors and several other cancer cell lines, such as pancreas, breast, Discussion and colorectal carcinomas cells (5, 11, 19), the association Dysregulated EGFR signaling is associated with the aggressive between CBLC upregulation and the risk of cancer has remained behavior of cancer cells. Herein, we reported a novel mechanism elusive. To the best of our knowledge, this is the ﬁrst report to through which CBLC competes with CBL for aEGFR binding and show that CBLC is an epigenetically demethylated target in lung guides aEGFR toward recycling for sustained activation (Fig. 7B). cancer. As global DNA hypomethylation is often observed early in In normal cells, the K63-linkaged ubiquitination mediated by tumorigenesis or abnormal hyperplasia (20), it is plausible that CBL tightly regulates aEGFR signaling through lysosomal degra- methylation of CBLC promoter may serve as an indicator of dation. In lung cancer cells, however, upregulation of CBLC patient survival. However, we were not able to observe a signif- competes with CBL for aEGFR binding and mediates the ubiqui- icant association between CBLC hypomethylation and poor sur- tination of aEGFR by K6 and K11 linkages. CBLC upregulation vival in patients with lung cancer (data not shown), due to the

4992 Cancer Res; 78(17) September 1, 2018 Cancer Research

CBLC Dysregulates EGFR Signaling

Figure 6. CBLC promotes aEGFR recycling by trafficking to the plasma membrane and the nucleus. A, Subcellular trafficking of aEGFR in the presence of CBLC. The H1437 cells were transfected with EGFP-tagged CBLC and stimulated with EGF (100 ng/mL) for the indicated times prior to overnight starvation. The cells were then fixed, permeabilized, and stained with anti-EGFR antibodies, followed by Alexa-594–coupled anti-rabbit IgG (red) for visualization by confocal laser microscopy. Arrowhead, CBLC-overexpressing cells. B, The distribution of aEGFR in H1975 cells infected with control shRNA or CBLC shRNA lentivirus. The 2D and 3D reconstruction of confocal Z-stack images of EGFR is shown in control and CBLC knockdown cells. The 3D confocal images of EGFR were obtained from a stack of 5 consecutive images (interval thickness ¼ 1 mm) centered on the middle of the nucleus. C–E, The colocalization of aEGFR and CBLC in RAB5- (C), RAB7- (D), and RAB11-positive (E) endosomes. H1975 cells stably expressing RFP-tagged RAB5, RAB7, or RAB11 were transfected with expression plasmids for EGFP- CBLC. aEGFR was stained with anti-EGFR antibodies, followed by Alexa-670–coupled anti-rabbit IgG (white). Corresponding reflected fluorescence intensities along the plotted arrow were analyzed by Zeiss Zen software. Scale bars, 20 mm.

small sample size of samples. Nevertheless, CBLC gene expression minutes to >120 minutes), supporting a role for CBLC in reducing was indeed highly correlated with its promoter hypomethylation the degradation of aEGFR. and the poor survival in LUAD, which accounts for approximately Previous X-ray structural analyses have shown that all TKB 40% of all lung cancer cases, supporting the notion that CBLC may domains of CBL family proteins can bind the same Y1045 assume an oncogenic function. phosphopeptide of EGFR (21). This could explain the binding CBL family proteins are E3 ubiquitin ligases that play a signif- competition of CBLC with CBL to aEGFR. The CBL-mediated K63 icant role in regulating aEGFR signaling (1, 5). Although both CBL linkages of polyubiquitination directs aEGFR for endocytosis, and CBLB have been relatively well studied, the mechanism endosomal sorting, and subsequently for degradation. In con- through which CBLC affects EGFR turnover remains poorly trast, the lack of binding afﬁnity toward CIN85 and the UBA understood. Here, we observed that the half-life of pERK1/2 domain for 26S proteasome targeting in the C-terminus of CBLC induced by EGF was increased in CBLC-overexpressing cells (from may account for the prolonged activation of aEGFR. Moreover, 10 minutes to 60 minutes; Fig. 4B) after normalization to the compared with CBL, CBLC assembled different atypical ubiquitin maximal pERK1/2 level occurring at 5 minutes in control cells. chains on aEGFR. It has been recently revealed that the topology of Speciﬁcally, after EGF stimulation, the phosphorylation level of ubiquitin chains of an ubiquitinated substrate may affect its pEGFR, pAKT, and pERK1/2 was much intense in CBLC transfec- functional consequences, such as the rate of lysosomal degrada- tants compared with that in control cells. In addition, the half-life tion and the recognition of ubiquitinbinding proteins (17, 22). of total EGFR was increased in the presence of CBLC (from 45 Although we did not examine whether the chains assembled by

www.aacrjournals.org Cancer Res; 78(17) September 1, 2018 4993

Hong et al.

Figure 7. The correlation of upregulated CBLC and pEGFR in LUAD. A, The representative IHC staining images and the correlation of CBLC and pEGFR in LUAD. The correlation between CBLC and pEGFR expression in 30 cases of LUAD was evaluated through Fisher exact test (P ¼ 0.0656). Scale bars, 100 mm. B, The model of CBLC-mediated EGFR dysregulation in LUAD cells. The aEGFR is normally downregulated by CBL-mediated ubiquitination through lysosomal degradation. When aberrant demethylation of CBLC promoter takes place in lung cancer cells, the upregulated CBLC competes with CBL for aEGFR binding and mediates polyubiquitination of aEGFR through K6 and K11 linkages. CBLC promotes aEGFR preferentially trafﬁcking into the nucleus or recycling back to the cell membrane, leading to enhanced stability of aEGFR and sustained activation of its downstream signaling.

4994 Cancer Res; 78(17) September 1, 2018 Cancer Research

CBLC Dysregulates EGFR Signaling

CBLC were in a linear or branched pattern, we did observe that signaling. Dual inhibition of aEGFR by binding competition CBLC preferentially assembled K6- and K11-linked polyubiquitin (TKI) and recycling interference (CBLC knockdown), may min- chains on aEGFR, which correlated with the enhanced aEGFR imize the chance of adverse effects caused by the antagonism stability. Nevertheless, more studies are required to investigate between different drugs used in clinics. More importantly, target- whether the diversity of polyubiquitin chains assembled by CBL ing CBLC in combination with TKIs administration could poten- or CBLC is determined by their pairing ubiquitin-conjugating tially provide great benefits to patients with NSCLC harboring enzymes (E2s). EGFR mutation, who have already developed adverse effects and While much attention has been paid to the activation and could not take full dose of TKIs. turnover of EGFR, little is known about the machinery that In conclusion, our findings suggest that upregulation of CBLC regulates the balance between degradation and recycling of in lung cancer is not merely a passenger event but rather assumes aEGFR. Postendocytic sorting is reported to be an important an oncogenic role that contributes to multiple aspects of tumor process controlling the lysosomal degradation or recycled traf- progression, including aEGFR dysregulation. Considering that ficking of aEGFR (23). It has been shown that endocytosis TKIs are widely used for the treatment of lung cancer with EGFR and endosomal sorting of aEGFR is mediated by ubiquitina- mutation (28), our findings highlight a promising strategy to tion, which can be recognized by endocytic cargo proteins for enhance the efficacy of current therapeutics, by use in combina- efficient lysosomal targeting. Our confocal microscopy results tion with CBLC-targeting agents. showed that CBLC and aEGFR colocalized in the RAB11-positive endosomes, supporting that CBL and CBLC may direct Disclosure of Potential Conflicts of Interest aEGFR toward different endosomes. Whether specificpolyubi- No potential conflicts of interest were disclosed. quitin linkages assembled by CBLC playing a role in endocytic trafficking needs further investigation. The aEGFR trafficking Authors' Contributions also involves nuclear translocation. Although EGFR may enter Conception and design: S.-Y.Hong, T.-C. Lee, C.-W. Wu the nucleus though the retrograde pathway (24), the exact Development of methodology: S.-Y. Hong, Y.-R. Kao Acquisition of data (provided animals, acquired and managed patients, mechanisms of nuclear translocation remain controversial. Our provided facilities, etc.): S.-Y. Hong, Y.-R. Kao, C.-W. Wu results demonstrated that depletion of CBLC inhibited the Analysis and interpretation of data (e.g., statistical analysis, biostatistics, transport of aEGFR to the nucleus. As nuclear localization of computational analysis): S.-Y. Hong, Y.-R. Kao, C.-W. Wu EGFR is closely correlated with the drug resistance to a variety of Writing, review, and/or revision of the manuscript: S.-Y. Hong, T.-C. Lee, anticancer therapeutics (25), our data also suggested that the C.-W. Wu upregulation of CBLC not only promotes tumorigenesis but Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): S.-Y. Hong, Y.-R. Kao, C.-W. Wu confers insensitivity to therapies by attenuating aEGFR degra- Study supervision: T.-C. Lee, C.-W. Wu dation in lung cancer. Targeting EGFR by first-line TKIs in patients with common Acknowledgments activating mutations has shown initial and significant response in We would like to thank Shih-Hsin Hsiao from Taipei Medical University for the clinic (26). However, after treatment, stress from EGFR-TKIs kindly providing TKIs and clinical suggestions. This work was supported in part selection either leads to the enriched growth of minor resistant by a postdoctoral fellowship from Academia Sinica to S.Y. Hong (2016–2018), clones with the exon 20 T790M mutation or induces other signal and research grants from the Ministry of Science and Technology of Taiwan pathways with bypassing effects (27, 28). Recently, several pre- (MOST 107-2321-B-010-007) and National Yang-Ming University (Aim for top, 107AE- P902) to C.W. Wu. clinical and clinical studies have reported that administering EGFR-TKIs with other therapeutics further improve outcomes in The costs of publication of this article were defrayed in part by the payment of patients with NSCLC (29). The cooperative effect of CBLC knock- page charges. This article must therefore be hereby marked advertisement in down plus TKIs administration, as described herein, has not been accordance with 18 U.S.C. Section 1734 solely to indicate this fact. observed at higher doses (25 mg/kg) of TKIs in the preclinical models tested. One of the probable reasons is that CBLC knock- Received December 15, 2017; revised April 26, 2018; accepted June 22, 2018; down inhibits tumor growth, as do TKIs, by targeting the aEGFR published first June 26, 2018.

References 1. Levkowitz G, Waterman H, Ettenberg SA, Katz M, Tsygankov AY, Alroy I, 7. Petrelli A, Gilestro GF, Lanzardo S, Comoglio PM, Migone N, Giordano S. et al. Ubiquitin ligase activity and tyrosine phosphorylation underlie The endophilin-CIN85-Cbl complex mediates ligand-dependent down- suppression of growth factor signaling by c-Cbl/Sli-1. Mol Cell regulation of c-Met. Nature 2002;416:187–90. 1999;4:1029–40. 8.LevkowitzG,WatermanH,ZamirE,KamZ,OvedS,LangdonWY, 2. Thien CB, Langdon WY. Cbl: many adaptations to regulate protein tyrosine et al. c-Cbl/Sli-1 regulates endocytic sorting and ubiquitination kinases. Nat Rev Mol Cell Biol 2001;2:294–307. of the epidermal growth factor receptor. Genes Dev 1998;12: 3. Peschard P, Park M. Escape from Cbl-mediated downregulation: a recurrent 3663–74. theme for oncogenic deregulation of receptor tyrosine kinases. Cancer Cell 9. Liyasova MS, Ma K, Lipkowitz S. Molecular pathways: cbl proteins in 2003;3:519–23. tumorigenesis and antitumor immunity-opportunities for cancer treat- 4. Weissman AM, Shabek N, Ciechanover A. The predator becomes the prey: ment. Clin Cancer Res 2015;21:1789–94. regulating the ubiquitin system by ubiquitylation and degradation. Nat 10. Szymkiewicz I, Kowanetz K, Soubeyran P, Dinarina A, Lipkowitz S, Rev Mol Cell Biol 2011;12:605–20. Dikic I. CIN85 participates in Cbl-b-mediated down-regulation of 5. Keane MM, Ettenberg SA, Nau MM, Banerjee P, Cuello M, Penninger J, et al. receptor tyrosine kinases. J Biol Chem 2002;277:39666–72. cbl-3: a new mammalian cbl family protein. Oncogene 1999;18:3365–75. 11. Kim B, Lee HJ, Choi HY, Shin Y, Nam S, Seo G, et al. Clinical validity 6. Schmidt MH, Dikic I. The Cbl interactome and its functions. Nat Rev Mol of the lung cancer biomarkers identiﬁed by bioinformatics analysis of Cell Biol 2005;6:907–18. public expression data. Cancer Res 2007;67:7431–8.

www.aacrjournals.org Cancer Res; 78(17) September 1, 2018 4995

Hong et al.

12. Li LC, Dahiya R. MethPrimer: designing primers for methylation PCRs. the tyrosine kinase binding domain of Cbl-c. J Biochem 2012; Bioinformatics 2002;18:1427–31. 152:487–95. 13. Regales L, Gong YX, Shen RL, de Stanchina E, Vivanco I, Goel A, et al. Dual 22. Kim HT, Kim KP, Lledias F, Kisselev AF, Scaglione KM, Skowyra D, et al. targeting of EGFR can overcome a major drug resistance mutation in Certain pairs of ubiquitin-conjugating enzymes (E2s) and ubiquitin-pro- mouse models of EGFR mutant lung cancer. J Clin Invest 2009;119: tein ligases (E3s) synthesize nondegradable forked ubiquitin chains con- 3000–10. taining all possible isopeptide linkages. J Biol Chem 2007;282:17375–86. 14. Huang F, Zeng X, Kim W, Balasubramani M, Fortian A, Gygi SP, et al. Lysine 23. Mosesson Y, Mills GB, Yarden Y. Derailed endocytosis: an emerging feature 63-linked polyubiquitination is required for EGF receptor degradation. of cancer. Nat Rev Cancer 2008;8:835–50. Proc Natl Acad Sci U S A 2013;110:15722–7. 24. Liao HJ, Carpenter G. Role of the Sec61 translocon in EGF receptor 15. Tong J, Taylor P, Moran MF. Proteomic analysis of the epidermal growth trafficking to the nucleus and gene expression. Mol Biol Cell 2007; factor receptor (EGFR) interactome and post-translational modifications 18:1064–72. associated with receptor endocytosis in response to EGF and stress. Mol 25. Tan X, Lambert PF, Rapraeger AC, Anderson RA. Stress-induced EGFR Cell Proteomics 2014;13:1644–58. trafficking: mechanisms, functions, and therapeutic implications. Trends 16. Huang FT, Kirkpatrick D, Jiang XJ, Gygi S, Sorkin A. Differential regulation Cell Biol 2016;26:352–66. of EGF receptor internalization and degradation by multiubiquitination 26. Pao W, Miller V, Zakowski M, Doherty J, Politi K, Sarkaria I, et al. EGF within the kinase domain. Mol Cell 2006;21:737–48. receptor gene mutations are common in lung cancers from "never smokers" 17. Ikeda F, Dikic I. Atypical ubiquitin chains: new molecular signals. `Protein and are associated with sensitivity of tumors to gefitinib and erlotinib. Proc Modifications: Beyond the Usual Suspects` review series. EMBO Rep Natl Acad Sci U S A 2004;101:13306–11. 2008;9:536–42. 27. Fujii A, Harada T, Iwama E, Ota K, Furuyama K, Ijichi K, et al. Hyper- 18. Tomas A, Futter CE, Eden ER. EGF receptor trafficking: consequences for methylation of the CpG dinucleotide in epidermal growth factor receptor signaling and cancer. Trends Cell Biol 2014;24:26–34. codon 790: implications for a mutational hotspot leading to the T790M 19. Kim M, Tezuka T, Suziki Y, Sugano S, Hirai M, Yamamoto T. Molecular mutation in non–small-cell lung cancer. Cancer Genet 2015;208:271–8. cloning and characterization of a novel cbl-family gene, cbl-c. Gene 28. Thomas A, Liu SV, Subramaniam DS, Giaccone G. Refining the treatment of 1999;239:145–54. NSCLC according to histological and molecular subtypes. Nat Rev Clin 20. Ehrlich M. DNA hypomethylation in cancer cells. Epigenomics 2009; Oncol 2015;12:511–26. 1:239–59. 29. Yang Z, Tam KY. Combination strategies using EGFR-TKi in NSCLC 21. Takeshita K, Tezuka T, Isozaki Y, Yamashita E, Suzuki M, Kim M, et al. therapy: learning from the gap between pre-clinical results and clinical Structural flexibility regulates phosphopeptide-binding activity of outcomes. Int J Biol Sci 2018;14:204–16.

4996 Cancer Res; 78(17) September 1, 2018 Cancer Research

Upregulation of E3 Ubiquitin Ligase CBLC Enhances EGFR Dysregulation and Signaling in Lung Adenocarcinoma

Shiao-Ya Hong, Yu-Rung Kao, Te-Chang Lee, et al.

Cancer Res 2018;78:4984-4996. Published OnlineFirst June 26, 2018.

Updated version Access the most recent version of this article at: doi:10.1158/0008-5472.CAN-17-3858

Supplementary Access the most recent supplemental material at: Material http://cancerres.aacrjournals.org/content/suppl/2018/06/26/0008-5472.CAN-17-3858.DC1

Cited articles This article cites 29 articles, 10 of which you can access for free at: http://cancerres.aacrjournals.org/content/78/17/4984.full#ref-list-1

Citing articles This article has been cited by 1 HighWire-hosted articles. Access the articles at: http://cancerres.aacrjournals.org/content/78/17/4984.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/78/17/4984. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.