Advertisement

EGFR-Mutant SCLC Exhibits Heterogeneous Phenotypes and Resistance to Common Antineoplastic Drugs

Open ArchivePublished:December 03, 2018DOI:https://doi.org/10.1016/j.jtho.2018.11.021

      Abstract

      Introduction

      Approximately 5% of patients with EGFR-activating mutations acquire EGFR tyrosine kinase inhibitor (TKI) resistance through SCLC transformation. However, the reason for the poor outcome and the molecular basis of EGFR-mutant SCLC that has transformed from adenocarcinoma remain unclear.

      Methods

      In this study, we established two EGFR-mutant SCLC cell lines from lung adenocarcinoma patients after failed EGFR-TKI treatment to investigate their molecular basis and potential therapeutic strategies in the hope of improving patient outcome.

      Results

      These two EGFR-mutant SCLC cell lines displayed two different phenotypes: suspensive and adherent. Both phenotypes shared the same genomic alterations analyzed by array-based comparative genomic hybridization assay. Increased expression of EGFR and mesenchymal markers and decreased expression of neuroendocrine markers were observed in adherent cells. Principal component analysis and hierarchical clustering analysis of RNA microarray revealed that these two cell lines displayed a unique gene expression pattern that was distinctly different from that in NSCLC and classical SCLC cells. Combined treatment using an EGFR-TKI and an AKT inhibitor attenuated cell viabilities in our two cell lines. Moreover, the use of a histone deacetylase inhibitor significantly inhibited the cell viabilities of both cell lines in vitro and in vivo.

      Conclusion

      Our findings suggest that EGFR-mutant SCLC may be a distinct subclass of SCLC that exhibits epithelial-mesenchymal transition phenotypes, and adding an AKT or histone deacetylase inhibitor to pre-existing therapies may be one of the therapeutic choices for transformed EGFR-mutant SCLC.

      Keywords

      Introduction

      Targeted therapies designed for specific genetic alternations, known as cancer driver mutations, have changed the treatment paradigm in advanced NSCLC, and especially for adenocarcinoma. EGFR-TKIs are effective in treating NSCLC harboring an activating mutation in the EGFR. Although patients may achieve robust responses to tyrosine kinase inhibitor (TKI) treatment, with tumor shrinkage and symptomatic relief, acquired drug resistance eventually develops in the majority of patients. SCLC transformation has been reported as one of the mechanisms of acquired resistance, and its incidence varied from 2%∼14% in different cohorts.
      • Ohashi K.
      • Maruvka Y.E.
      • Michor F.
      • et al.
      Epidermal growth factor receptor tyrosine kinase inhibitor-resistant disease.
      • Yu H.A.
      • Arcila M.E.
      • Rekhtman N.
      • et al.
      Analysis of tumor specimens at the time of acquired resistance to EGFR-TKI therapy in 155 patients with EGFR-mutant lung cancers.
      • Sequist L.V.
      • Waltman B.A.
      • Dias-Santagata D.
      • et al.
      Genotypic and histological evolution of lung cancers acquiring resistance to EGFR inhibitors.
      • Ahn S.
      • Hwang S.H.
      • Han J.
      • et al.
      Transformation to small cell lung cancer of pulmonary adenocarcinoma: clinicopathologic analysis of six cases.
      SCLC is an aggressive malignancy characterized by early metastases and an initial response to chemotherapy. However, the prognosis is poor because of early relapse and resistance to subsequent chemotherapy. SCLC and NSCLC are two major subtypes of lung cancer.
      • Davidson M.R.
      • Gazdar A.F.
      • Clarke B.E.
      The pivotal role of pathology in the management of lung cancer.
      • Oser M.G.
      • Niederst M.J.
      • Sequist L.V.
      • et al.
      Transformation from non-–small-cell lung cancer to small-cell lung cancer: molecular drivers and cells of origin.
      Although there are differences in biology and genomic abnormality, the hypothesis that NSCLC and SCLC might share common cells of origin has been proposed.
      • Oser M.G.
      • Niederst M.J.
      • Sequist L.V.
      • et al.
      Transformation from non-–small-cell lung cancer to small-cell lung cancer: molecular drivers and cells of origin.
      SCLC is characterized by neuroendocrine (NE) differentiation and small cell morphology, but not all SCLC cancer cells have the same growth pattern. Classical SCLC cells grow as suspension and aggregation, in contrast to the adherent monolayers in most NSCLC cancer cells.
      • Pettengill O.S.
      • Sorenson G.D.
      • Wurster-Hill D.H.
      • et al.
      Isolation and growth characteristics of continuous cell lines from small-cell carcinoma of the lung.
      However, some SCLC cells have adherent populations with a diminution of several NE properties, and are labeled as variant SCLC cells.
      • Doyle L.A.
      • Borges M.
      • Hussain A.
      • et al.
      An adherent subline of a unique small-cell lung cancer cell line downregulates antigens of the neural cell adhesion molecule.
      Calbo et al. found that SCLC cells were sometimes composed of two different phenotypes, with NE and mesenchymal (non-NE) profiles, in an SCLC mouse model. NE cells grew as classic SCLC cells, whereas non-NE cells grew as adherent monolayers with visible cytoplasm and spindle-like membrane extensions. Crosstalk between NE and non-NE can change tumor cell behavior through paracrine signaling.
      • Sutherland K.D.
      • Proost N.
      • Brouns I.
      • et al.
      Cell of origin of small cell lung cancer: inactivation of Trp53 and Rb1 in distinct cell types of adult mouse lung.
      • Calbo J.
      • van Montfort E.
      • Proost N.
      • et al.
      A functional role for tumor cell heterogeneity in a mouse model of small cell lung cancer.
      For this study, we established two EGFR-mutant SCLC cell lines from patients who were resistant to EGFR-TKIs after being initially diagnosed as having EGFR-mutant adenocarcinoma. Both cells presented two distinct phenotypes, a suspensive of spheroid aggregates and an adherent monolayer. Suspensive cells revealed the classic characteristics of NE cells, whereas adherent cells expressed epithelial-mesenchymal transition (EMT) markers. Our results implied that crosstalk exists between these cells, and that both cells are highly sensitive to histone deacetylase (HDAC) 1/2 inhibitors.

      Materials and Methods

      Patient Samples Collection

      Malignant pleural effusion (MPE) samples were consecutively collected from lung cancer patients at National Taiwan University Hospital (NTUH). This study was approved by the Institutional Review Board of the NTUH Research Ethics Committee (approval number: 200804019R).

      Primary Culture From MPE of Lung Cancer

      Primary cancer cells were established as previously described.
      • Ho C.C.
      • Liao W.Y.
      • Lin C.A.
      • et al.
      Acquired BRAF V600E Mutation as resistant mechanism after treatment with osimertinib.
      In brief, MPE cells from lung cancer patients were collected by centrifugation, and excess red blood cells (RBCs) were removed using RBC lysis buffer (155 mmol/L of NH4Cl, 10 mmol/L of KHCO3, and 0.1 mmol/L of ethylenediaminetetraacetic acid). Cancer cells and other normal cells were separated by differential trypsinization during cell passage. Cancer cells were maintained initially in F medium, and after several passages, pure cancer cells were obtained and adapted to Roswell Park Memorial Institute (RPMI-1640) medium containing 10% fetal bovine solution (FBS).
      • Liu X.
      • Ory V.
      • Chapman S.
      • et al.
      ROCK inhibitor and feeder cells induce the conditional reprogramming of epithelial cells.

      Flow Cytometry

      CD44 expression on cell surface was analyzed by Cytomics FC500 flow cytometry (Beckman Coulter Inc., Indianapolis, IN). Cells were stained by a specific fluorescein isothiocyanate (FITC)–mouse anti-human CD44 antibody (BD Pharmingen #555478) and relative isotype controls for 15 minutes on ice. After staining, the cells were washed and re-suspended in fluorescent-activated cell sorter buffer (phosphate-buffered saline [PBS] 0.5% bovine serum albumin, 0.05% NaN3).

      Genomic DNA and Total RNA Extraction

      Genomic DNA and total RNA were extracted from cells using the NucleoSpin Tissue kit (Macherey-Nagel Cat #740952, Bethlehem, PA) and NucleoSpin RNA kit (Macherey-Nagel Cat #740955), respectively. The concentrations of DNA and RNA were measured with a NanoDrop ND-1000 spectrophotometer (Thermo Scientific, Wilmington, Delaware).

      Comparative Genomic Hybridization Analysis

      DNA was extracted from CLH-A, CLH-S, CLH5-A and CLH5-S cells. Array-based comparative genomic hybridization (aCGH) was performed with a SurePrint G3 CGH Microarray platform, 4 × 180K (Agilent Technologies Inc., Santa Clara, California). The data were analyzed using Agilent CytoGenomics software v 2.0.

      Proliferation Assay

      Cells (1 × 105 cells per well) were seeded in 12-well plates overnight. At the indicated times (days 1 to 6), cells were trypsinized and stained using 0.4% trypan blue (Gibco, Cat# 152250061). The number of viable cells was counted in a counting chamber and a proliferation plot was performed using Graphpad Prism 5.0 software (GraphPad Software Inc.).

      Invasion Assay

      The invasion assay was performed using a membrane invasion culture system. A polycarbonate membrane containing 8-μm pores (Corning, Cat#354578) was coated with Matrigel matrix (Corning, Cat #354230). The 2 × 105 cells were seeded onto Matrigel matrix-coated membrane. After 16 to 24 hours of incubation, the invaded cells were fixed with methanol, stained with Giemsa solution (Sigma-Aldrich, Cat#GS500500ML) and counted at ×200 original magnification under a light microscope.

      Colony-Formation Assay

      In the anchorage-independent colony-formation assay, 12-well plates were first layered with 1 mL 0.7% low-melting-point agarose in PBS. In the second layer, CLH and CLH5 cells (2000 and 5000 cells, respectively) were suspended in 1mL RPMI containing 0.35% low-melting-point agarose. Next, 1 mL RPMI medium or conditioned medium was added to cover the second layer. After 2 to 4 weeks of growth, the wells were washed in PBS, fixed in 4% paraformaldehyde, and stained with 0.1% crystal violet. Finally, the staining solutions were disposed and PBS was added. The images of colonies were captured and counted by a MetaMorph Imaging System (Universal Imaging Corporation, West Chester, Pennsylvania).

      Conditioned Medium Collection

      The 2 × 106 cells of CLH (A and S) and CLH5 (A and S) were seeded in a 75T flask and maintained in RPMI 1640 supplemented with 10% FBS. After 24-hour incubation, the culture supernatants were collected as conditioned medium.

      Western Blotting

      Western blotting was performed as previously described. The primary antibodies used included Rb, Synaptophysin, p-EGFR (T1068), EGFR, p-AKT (S473), AKT, p-MEK1/2 (S217/ 221), MEK1/2, p-ERK1/2 (T202/Y204), ERK1/2 (Cell Signaling Technology, Danvers, Massachusetts), NeuroD1, Zeb1, Twist, Snail, Slug (Abcam, Cambridge, United Kingdom) and CD44, Vimentin, GADH, and α-tubulin (Genetex, Zeeland, Michigan). Secondary antibodies used included horseradish peroxidase–conjugated anti-mouse immunoglobulin G and horseradish peroxidase–conjugated anti-rabbit immunoglobulin G (Genetex) antibodies.

      RB1 and TP53 Sequencing

      For Rb1 sequencing, the regions of Rb1 codons were amplified using specific polymerase chain reaction (PCR) primers (Supplementary Table 1). PCR products were then sequenced using the same primer set. For p53 sequencing, cDNA was used as a template for PCR of the p53 CDS sequence. Primer: Forward: CCAAAACTCTGGGAATACTGGCAC
      Reverse: GTGCCAGTATTCCCAGAGTTTTGG

      PCA Analysis and Hierarchical Clustering

      The microarray data of our EGFR-mutant SCLC cells (CLH-A, S, CLH5-A, and CLH5-S), Niederst's EGFR-mutant SCLC cells (MGH131-1 and MGH131-2), and lung cancer cells (six SCLC and nine NSCLC) from the Cancer Cell Line Encyclopedia (CCLE) underwent principal component analysis (PCA) using JMP Software (SAS Inc., Cary, North Carolina), and hierarchical clustering analysis using GAP software.
      • Niederst M.J.
      • Sequist L.V.
      • Poirier J.T.
      • et al.
      RB loss in resistant EGFR mutant lung adenocarcinomas that transform to small-cell lung cancer.
      • Barretina J.
      • Caponigro G.
      • Stransky N.
      • et al.
      The Cancer Cell Line Encyclopedia enables predictive modelling of anticancer drug sensitivity.
      • Wu H.-M.
      • Tien Y.-J.
      • Chen C.-h.
      GAP:a graphical environment for matrix visualization and cluster analysis.
      Hierarchical clustering analysis was performed with 1002 of the most differentially regulated genes with the highest variance among 21 cells.

      Phosphokinase Assay

      Phosphokinase assay was performed with a human phosphokinase array kit (R&D Systems Catalog # ARY003B), which can detect 43 kinase phosphorylation sites and two related total proteins without performing numerous immunoprecipitations and Western blots. Cells were starved with serum-free medium overnight and recovered with 10% FBS medium for 15 minutes. Cell lysates were extracted and quantified for kinase assay. After performing all procedures, the signals were visualized with a chemiluminescence assay kit (Millipore, Billerica, Massachusetts) and detected using a FUJIFILM LAS-3000 ECL system.

      Growth Inhibition Assay

      Cells (1 × 104 cells) were seeded in a 96-well plate and were treated with the indicated drugs, and then were incubated for 5 days. For the cell viability assay, Alamar Blue reagent (Invitrogen, Cat#DAL1100) was added for 3 to 8 hours and absorbance was read by a Wallac Victor3 1420 Multi-label Counter (Perkin Elmer). Growth inhibition plotting and half maximal inhibitory concentration (IC50) were performed with Graphpad Prism 5.0 software (GraphPad Software Inc.).

      In Vivo Mouse Model

      Six- to 8-week-old non-obese diabetic (NOD)/severe combined immunodeficiency (SCID) mice were supplied by BioLASCO Taiwan, and housed in specific pathogen-free animal rooms. All animal experiments were approved by the National Taiwan University College of Medicine and College of Public Health Institutional Animal Care and Use Committee (approval number: 20140527). For the subcutaneous implantation xenograft model, CLH cells (2 × 106) and CLH5 cells (5 × 106) in 50 μL PBS were mixed with matrix (Corning, Cat #354230) at a 1:1 ratio and implanted into each NOD/SCID mouse. The antitumor activity of FK-228 (Selleckchem Cat #S3020) was evaluated in xenograft mice by intraperitoneal administration at a dose of 1 mg/kg twice per week for 4 weeks. Tumor size and body weight were measured every 4 days. After 5 weeks, the mice were sacrificed.

      Microarray and Pathway Analysis

      Five micrograms of mRNAs derived from cells were used with Human Genome U133 plus 2.0 GeneChip, following the manufacturer’s protocols (Affymetrix). The raw data were analyzed with GeneSpring GX software (Silicon Genetics, Redwood City, California). The GO biological processes of genes with greater than a two-fold change were analyzed using the MetaCore Gene Set Enrichment Analysis (GSEA) software program (GeneGo Inc., St Joseph, Michigan).

      Results

      EGFR-Mutant SCLC Cells Derived From Patients

      Patient-derived cancer cells are valuable tools for studying the molecular basis of lung cancer and for evaluating therapeutic strategies.
      • Garnett M.J.
      • Edelman E.J.
      • Heidorn S.J.
      • et al.
      Systematic identification of genomic markers of drug sensitivity in cancer cells.
      We established 80 primary cancer cells from lung cancer patients with MPE. Two of them (CLH and CLH5) were EGFR-mutant SCLC. CLH was derived from a 51-year-old nonsmoking woman who had advanced NSCLC adenocarcinoma with an EGFR L858R mutation. She had an initial response to EGFR-TKI, but became refractory to therapy 6 months later. New lung lesions and MPE developed after disease progression. CLH5 was derived from a 45-year-old man who had advanced NSCLC adenocarcinoma with an EGFR exon 19 deletion mutation and who had received EGFR-TKI treatment. Two months later, a left chest wall lesion and MPE developed. The transformation from adenocarcinoma (Fig. 1A, left and Fig. 1B, left) to SCLC (Fig. 1A, right and Fig. 1B, right) was found after failure of EGFR-TKI treatment, compared to before treatment, by histologic examination.
      Figure thumbnail gr1
      Figure 1Transformed SCLC cells derived from EGFR-mutation patients exhibited two distinct morphologies. (A) Patient CLH had a L858R mutation in EGFR. (B) Patient CLH5 had an exon 19 deletion in EGFR. Hematoxylin and eosin (HE) (upper) and immunohistochemistry (IHC) staining (lower) for TTF-1 revealed adenocarcinoma morphologies in both pretreatment specimens (A and B, left panels). The resistant biopsy specimens (A and B, right panels) uncovered SCLC morphology with positive synaptophysin staining for IHC. (C) Photomicrograph shows the morphologies of cells growing in adherent-type cells (upper) and suspension-type cells (lower). (D) Flow cytometry analysis shows CD44 highly expressed on adherent cells and weakly expressed on suspension cells. (E) CGH array analysis shows that the adherent cells and suspension cells of CLH (upper) and CLH5 (lower) underwent the same genetic alterations. Original magnification is 400×. IHC, immunohistochemistry; HE, hematoxylin-eosin; TTF-1, thyroid transcription factor 1; CGH, comparative genomic hybridization; TKI, tyrosine kinase inhibitor; FITC, fluorescein isothiocyanate; A&S, adherent and suspension cell.

      EGFR-Mutant SCLC Displaying Two Distinct Phenotypes

      Two morphological subtypes were observed with both CLH and CLH5 cells during culture. One was a classical NE morphology in a suspension of small aggregated cells and the other was an adhesion monolayer with a mesenchymal morphology (non-NE-like) (Fig. 1C); these two morphological subtype cells could switch back and forth. CD44, a mesenchymal cell marker, was strongly expressed in adherent cells but not in suspensive cells (Fig. 1D). To clarify whether these two morphological cells were derived from the same origin, we first checked the EGFR mutation status using a highly sensitive matrix-assisted laser desorption ionization-time of flight mass spectrometry detection system. Both phenotypes from CLH and CLH5 were found to have the same mutation frequency: L858R mutation ∼21% for CLH suspension cells (CLH-S) and adherent cells (CLH-A), and exon 19 deletion ∼80% for CLH5 suspension cells (CLH5-S) and adherent cells (CLH5-A) (Supplementary Fig. 1).
      • Su K.Y.
      • Chen H.Y.
      • Li K.C.
      • et al.
      Pretreatment epidermal growth factor receptor (EGFR) T790M mutation predicts shorter EGFR tyrosine kinase inhibitor response duration in patients with non-small-cell lung cancer.
      Genome instability is a feature of cancer. So, we used aCGH analysis to assess whether both phenotype cells shared the same genome aberrations. The aCGH data showed nearly identical genetic aberrations between the two phenotypes of CLH and CLH5 (Fig. 1E). To clarify whether the morphological switching was due to a mixed cell population, we established a single clone from CLH cells. The single clone derived from CLH still exhibited two morphological phenotypes, and CD44 was still expressed in the adherent cells (Supplementary Fig. 2A). This phenomenon was also observed in the primary culture from the CLH xenograft (Supplementary Fig. 2B). CD44-positive and -negative cells coexisted in EGFR-mutant SCLC xenograft (Supplementary Fig. 2C). These results suggested that the suspensive and adherent cells had the same origin.

      Functional Characterization of Suspensive and Adherent Cells

      Both CLH-A and CLH5-A showed a lower growth rate and more invasive activity (Figs. 2A and B). In contrast, CLH-S and CLH5-S showed better colony-forming ability (Fig. 2C). In previous studies using a mouse SCLC model, SCLC was composed of NE cells and a small mesenchymal subset (non-NE cells). The non-NE cells could provide a metastatic ability to the NE cells through paracrine signaling.
      • Calbo J.
      • van Montfort E.
      • Proost N.
      • et al.
      A functional role for tumor cell heterogeneity in a mouse model of small cell lung cancer.
      • Kwon M.C.
      • Proost N.
      • Song J.Y.
      • et al.
      Paracrine signaling between tumor subclones of mouse SCLC: a critical role of ETS transcription factor Pea3 in facilitating metastasis.
      We next assayed whether adherent cells can function as non-NE cells that promote the invasion of NE cells. Adherent cell-derived conditioned medium from adherent cells alone or co-cultured with suspensive cells (at a 1:1 ratio) could promote the invasive ability of suspensive cells (Fig. 2D), but suspensive cell-derived conditioned medium had no such effect. Adherent cell-derived conditioned medium alone also promoted the colony-forming ability of suspensive cells (Fig. 2E). These data imply that adherent cells of EGFR-mutant SCLC play a role in promoting tumor growth and metastasis.
      Figure thumbnail gr2
      Figure 2Characterization of adherent cells and suspension cells. (A) Cell growth assay indicates that CLH-A grew faster than CLH-S, and that the CLH5-A growth rate was similar to CLH5-S. (B) Matrigel invasion assay shows that adherent cells of both CLH and CLH5 had greater invasive ability than suspension cells. (C) Anchorage-independent colony formation assay shows that CLH-S had greater colony-forming ability than CLH-A. Both CLH5-A and CLH5-S had low colony-forming ability. Statistical significance was determined by Student’s t test (*p < 0.05). (D) Conditioned medium collected from adherent cells only and co-cultured with suspension cells could enhance the invasive ability of suspension cells more than conditioned medium collected from suspension cells and the RPMI control. Data are representative of at least three independent experiments. Statistical significance was determined by Tukey's multiple comparison test (*p < 0.05). (E) Conditioned medium collected from adherent cells only and co-cultured with suspension cells could enhance the colony-forming ability of suspension cells more than conditioned medium collected from suspension cells and the RPMI control. Data are representative of at least triplicate experiments. Statistical significance was determined by Tukey's multiple comparison test (*p < 0.05).

      RB1 and TP53 Loss in EGFR-Mutant SCLC

      RB1 and TP53 mutation is a hallmark of SCLC.
      • van Meerbeeck J.P.
      • Fennell D.A.
      • De Ruysscher D.K.
      Small-cell lung cancer.
      • Peifer M.
      • Fernandez-Cuesta L.
      • Sos M.L.
      • et al.
      Integrative genome analyses identify key somatic driver mutations of small-cell lung cancer.
      • Rudin C.M.
      • Durinck S.
      • Stawiski E.W.
      • et al.
      Comprehensive genomic analysis identifies SOX2 as a frequently amplified gene in small-cell lung cancer.
      Whole-genome sequencing recently showed mutation frequencies of more than 90% in both RB1 and TP53 in SCLC.
      • George J.
      • Lim J.S.
      • Jang S.J.
      • et al.
      Comprehensive genomic profiles of small cell lung cancer.
      Of importance, RB1 loss was 100% in SCLC transformed from TKI-resistant adenocarcinomas.
      • Niederst M.J.
      • Sequist L.V.
      • Poirier J.T.
      • et al.
      RB loss in resistant EGFR mutant lung adenocarcinomas that transform to small-cell lung cancer.
      In our data, CLH cells harbored a nonsense mutation at codon51 (GAA>TAA), and CLH5 cells harbored two missense mutations at codon72 (CCC>CGC) and codon179 (CAT>TAT) in the TP53 gene (Supplementary Figs. 3A and B). However, RB1 protein lost expression in CLH cells, despite the lack of a point mutation or deletion in the chromosome (data shown), and RB1 was truncated in CLH5 cells (Supplementary Fig. 3C). Moreover, adherent cells expressed more mesenchymal markers, vimentin, and CD44, and fewer NE markers, synaptophysin, and NEUROD1, than suspensive cells (Supplementary Fig. 3D). Also, adherent cells had more activating phosphorylation of Erk (Supplementary Fig. 3D), and EGFR expression was markedly reduced in suspensive cells (Supplementary Fig. 3D). To our knowledge, this is the first study to describe the dynamic expression of EGFR in NE and non-NE phenotypes in EGFR-mutant SCLC.

      Molecular Basis of EGFR-Mutant SCLC

      To determine the genome-wide transcriptome profiles of two phenotypes of two of our EGFR-mutant SCLC cell lines, we performed microarray analysis on CLH-S, CLH-A, CLH5-S, and CLH5-A cells, and concomitantly compared their gene expression data with those of classic SCLC (H69, H146, H526, H209, H446, and H82) and NSCLC (H1299, H23, H460, A549, H226, HCC827, H1975, H322, and H522) cells from the CCLE database and those of Niederst's EGFR-TKI resistant transformed SCLC cells (MGH131-1 and MGH131-2).
      • Niederst M.J.
      • Sequist L.V.
      • Poirier J.T.
      • et al.
      RB loss in resistant EGFR mutant lung adenocarcinomas that transform to small-cell lung cancer.
      • Barretina J.
      • Caponigro G.
      • Stransky N.
      • et al.
      The Cancer Cell Line Encyclopedia enables predictive modelling of anticancer drug sensitivity.
      To assess similarity of these cell lines, PCA was used with the expression of 54,675 probes. The results showed that the cluster of EGFR-mutant SCLC was distinct from that of classic SCLC and NSCLC cells. The clustering of CLH and CLH5 cell lines, with either adherent or suspensive phenotypes, was close to that of MGH131-1 and MGH-131-2 cells (Fig. 3A). Furthermore, we identified 1002 differentially expressed probes with the highest variance among 21 cell lines and analyzed them using hierarchical clustering. The results indicated that NSCLC is distinctly separate from SCLC, and EGFR-mutant SCLC (CLH-S, CLH-A, CLH5-S, and CLH5-A) cells are clustered together with the exception of MGH131-1 and MGH-131-2. MGH131-1 even formed an independent cluster apart from NSCLC and SCLC using the 1002 probes clustering analysis (Fig. 3B). MGH131-1 and MGH-131-2 were derived from two different biopsies (taken several months apart) of an EGFR TKI-resistant patient, the pattern of differentially expressed genes represents heterogeneous when it compared to CLH-S/CLH-A and CLH5-S/CLH5-A which were derived from different patients.
      Figure thumbnail gr3
      Figure 3EGFR-mutant SCLC exhibited a unique gene expression cluster. The gene expression data of NSCLC and SCLC cells was obtained from the Cancer Cell Line Encyclopedia (CCLE) and the data for two EGFR-mutant SCLC cells was obtained from Niederst’s study. (A) Principal component analysis separated the clustering of EGFR-mutant SCLC from NSCLC and SCLC. (B) Hierarchical clustering analysis of the highest variance expression of 1002 genes among the NSCLC and SCLC cells and EGFR-mutant SCLC cells. The clustering revealed that EGFR-mutant SCLC has a unique gene expression cluster.

      Akt Activation in EGFR-Mutant SCLC

      EGFR protein was expressed in the adherent-type cells of both CLH-A and CLH5-A, but the cell viability assay indicated that both CLH-A and CLH5-A were highly resistant to both gefitinib (Fig. 4A, upper) and the third-generation EGFR inhibitor, osimertinib (AZD9291) (Fig. 4A, lower). Not only were CLH-A and CLH5-A resistant to EGFR-TKIs, but CLH-S and CLH5-S were insensitive to EGFR-TKIs, so apparently the suspension-type cells were resistant to EGFR-TKI because of the low EGFR expression. This result is consistent with clinical observations of CLH and CLH5 that were derived from EGFR-TKIs-resistant patients. Despite gefitinib blocking EGFR phosphorylation in adherent cells, downstream phosphorylation of Akt and Erk was still active (Fig. 4B). To clarify which signaling pathway offered drug resistance to adherent cells that expressed EGFR, a human phosphokinase array was performed. The data indicated that the most significant axis was the AKT/mTOR pathway in adherent cells, as well as in suspension cells, including phosphorylation of AKT1/2/3, GSK-3α/β, CREB, and PRAS40 protein (Fig. 4C). PI3K/AKT/mTOR activating is one of the mechanisms involved in EGFR-TKI resistance in NSCLC and in chemotherapy resistance in SCLC.
      • Sordella R.
      • Bell D.W.
      • Haber D.A.
      • et al.
      Gefitinib-sensitizing EGFR mutations in lung cancer activate anti-apoptotic pathways.
      • Sos M.L.
      • Koker M.
      • Weir B.A.
      • et al.
      PTEN loss contributes to erlotinib resistance in EGFR-mutant lung cancer by activation of Akt and EGFR.
      • Huang L.
      • Fu L.
      Mechanisms of resistance to EGFR tyrosine kinase inhibitors.
      • Kraus A.C.
      • Ferber I.
      • Bachmann S.O.
      • et al.
      In vitro chemo- and radio-resistance in small cell lung cancer correlates with cell adhesion and constitutive activation of AKT and MAP kinase pathways.
      • Belyanskaya L.L.
      • Hopkins-Donaldson S.
      • Kurtz S.
      • et al.
      Cisplatin activates Akt in small cell lung cancer cells and attenuates apoptosis by survivin upregulation.
      A cell viability assay indicated that the AKT inhibitor (MK2206) exhibited marked efficacy in CLH-A (half maximal inhibitory concentration [IC50]: 0.46 μmol/L) and CLH-S (IC50: 0.22μmol/L), and minor efficacy in CLH5-A (IC50: 12.3 μmol/L) and CLH5-S (IC50: 5.2 μmol/L) (Fig. 4C). The cell viabilities of CLH-A (IC50: 0.13 μmol/L) were markedly reduced, by approximately four-fold, by combining the AKT inhibitor with gefitinib (Figs. 4D and E). However, a modest reduction was observed in CLH5-A and CLH5-S (Fig. 4E). These data suggested the PI3K/AKT/mTOR pathway may be a feasible chemotherapeutic target in EGFR-mutant SCLC patients.
      • Sequist L.V.
      • Waltman B.A.
      • Dias-Santagata D.
      • et al.
      Genotypic and histological evolution of lung cancers acquiring resistance to EGFR inhibitors.
      • Makinoshima H.
      • Umemura S.
      • Suzuki A.
      • et al.
      Metabolic determinants of sensitivity to phosphatidylinositol 3-kinase pathway inhibitor in small-cell lung carcinoma.
      Figure thumbnail gr4
      Figure 4TKI resistance of EGFR-mutant SCLCs is EGFR-independent and activated in the AKT/mTOR pathway. The survival curve of cells was analyzed against a first-generation TKI (gefitinib) (A, upper) and third-generation TKI (osimertinib) (B, lower). (B) Iressa inhibited the phosphorylation of EGFR, but did not block downstream signaling, including AKT and Erk. (C) Cellular phosphokinases array showed that both adherent and suspension cells exhibited an activating AKT/mTOR pathway, including phosphorylation of GSK-3a/b, AKT1/2/3, CREB, and PRAS40. CLH (C, left) and CLH5 cells (C, right). CLH-A was sensitive to MK2206 (D, left upper). CLH-S was highly sensitive to MK2206 (D, left lower). CLH5-A and CLH5-S were slightly sensitive to MK2206 (D, right). CLH-A and CLH-S were markedly reduced by combined treatment (E, left). But, modest reduction was observed in CLH5-A and CLH5-S (E, right).

      EGFR-Mutant SCLC is Highly Sensitive to HDAC 1/2 Inhibitor

      Epigenetic regulation, including DNA methylation and histone modifications, has emerged as an important target for cancer therapy.
      • Jones P.A.
      • Issa J.P.
      • Baylin S.
      Targeting the cancer epigenome for therapy.
      SCLC may also be sensitive to HDAC inhibitors (HDACi).
      • Platta C.S.
      • Greenblatt D.Y.
      • Kunnimalaiyaan M.
      • et al.
      The HDAC inhibitor trichostatin A inhibits growth of small cell lung cancer cells.
      • Crisanti M.C.
      • Wallace A.F.
      • Kapoor V.
      • et al.
      The HDAC inhibitor panobinostat (LBH589) inhibits mesothelioma and lung cancer cells in vitro and in vivo with particular efficacy for small cell lung cancer.
      • Hubaux R.
      • Vandermeers F.
      • Crisanti M.C.
      • et al.
      Preclinical evidence for a beneficial impact of valproate on the response of small cell lung cancer to first-line chemotherapy.
      • Pan C.H.
      • Chang Y.F.
      • Lee M.S.
      • et al.
      Vorinostat enhances the cisplatin-mediated anticancer effects in small cell lung cancer cells.
      • Sabari J.K.
      • Lok B.H.
      • Laird J.H.
      • et al.
      Unravelling the biology of SCLC: implications for therapy.
      Here, we analyzed 1600 genes that were significantly different among EGFR-mutant SCLC, classical SCLC, and NSCLC using MetaCore software. As 1 of the top 10 ranked pathways, the Sin3 and NuRD pathway was up-expressed in EGFR-mutant SCLC. Sin3 and NuRD are two major HDAC complexes that mediate transcriptional regulation through HDAC (Supplementary Figs. 4A and B). So, we compared the drug sensitivities of the pan-HDAC inhibitor, trichostatin A (TSA), between EGFR-mutant SCLC, classical SCLC, and NSCLC. The data showed the EGFR-mutant SCLC was more sensitive to TSA, although classic SCLCs were also sensitive to the pan-HDACi. However, NSCLC was less sensitive to this drug (Supplemental Fig. 4C) Hence, in addition to AKT inhibitors, HDACi are assumed to be therapeutic targets for EGFR-mutant SCLC. We compared the cell viability inhibitory efficacy of treatment with TSA and the chemotherapy drugs cisplatin and etoposide, which are commonly used in SCLC treatment, and the BCL-2 inhibitor, navitoclax (ABT-263), which was reported to markedly reduce the viability of SCLC-transforming cells.
      • Niederst M.J.
      • Sequist L.V.
      • Poirier J.T.
      • et al.
      RB loss in resistant EGFR mutant lung adenocarcinomas that transform to small-cell lung cancer.
      Here, we found that TSA had a better inhibitory efficacy than cisplatin-etoposide and navitoclax (Figs. 5A and F). Also, adherent cells revealed more drug resistance than suspensive cells (Figs. 5AF). Then, we dissected the inhibitory efficacy of different classes of HDACi (Supplementary Figs. 5AC), and found that the HDAC1/2 inhibitor, romidepsin (FK-228), was the most potent inhibitor, with IC50 less than 10 nmol/L in all cells (Fig. 6A). We wanted to confirm this observation in vivo, so the subcutaneous implantation xenograft model was used. Romidepsin treatment resulted in significantly more reduced tumor growth than the vehicle control (Fig. 6B). These data suggest that HDACi may provide another clinical application for EGFR-TKI-resistant patients with transformed SCLC.
      Figure thumbnail gr5
      Figure 5Adherent cells of EGFR-mutant SCLC exhibited higher drug tolerance than suspension cells. Cells were analyzed for their survival curve against chemotherapy drugs (etoposide plus 0.1 μmol/L cisplatin) (A and B), a bcl-2 inhibitor (ABT-263) (C and D) and pan-HDAC inhibitors (trichostatin A [TSA]) (E and F). Adherent cells exhibited higher drug tolerance to all treatments than suspension cells. TSA was a better inhibitory drug than the other two drugs.
      Figure thumbnail gr6
      Figure 6FK228 inhibited the tumor growth of EGFR-mutant SCLC in vitro and in vivo. Cells were analyzed for their survival curve against the HDAC inhibitor, FK228, in vitro. (A) CLH-A (upper, left), CLH-S (upper, right), CLH5-A (lower, left), and CLH5-S (lower, right). (B) The xenograft mice were generated by implanting CLH cells (for 2 × 106) and CLH5 cells (for 5 × 106). After 1 week, FK228 was intraperitoneally administered at a dose of 1 mg/kg twice per week. Tumor size and body weight was measured every 4 days. After 5 weeks, the mice were sacrificed. Data are represented as mean ± SEM (n ≥ 5). Statistical analysis was performed by unpaired Student’s t test (*p < 0.05), and revealed significant tumor growth inhibition with FK228 (p < 0.05) compared with the control. Tumor photographs (scale: 1 cm). Tissue morphology was examined by hematoxylin and eosin (HE) staining (scale: 50 μm). HDAC, histone deacetylase.

      Discussion

      SCLC transformation is one of the mechanisms of EGFR-TKI resistance. Here, we reported two EGFR-mutant SCLC cell lines that presented suspensive and adherent morphologies. These two phenotypes shared the same genetic aberrations and genetic background. Adherent cells have the potential to promote the invasive ability of suspensive cells. This phenotypic change and functional heterogeneity was also observed in a SCLC mouse model, in which non-NE and NE cells had highly genetic similarities, and the presence of non-NE cells enhanced the metastatic ability of NE cells.
      • Calbo J.
      • van Montfort E.
      • Proost N.
      • et al.
      A functional role for tumor cell heterogeneity in a mouse model of small cell lung cancer.
      • Kwon M.C.
      • Proost N.
      • Song J.Y.
      • et al.
      Paracrine signaling between tumor subclones of mouse SCLC: a critical role of ETS transcription factor Pea3 in facilitating metastasis.
      It is interesting that the phenotypic change was maintained in sub-cultured cells derived from a xenograft or single-cell selection obtained from CLH. We also found that adherent cells can promote colony-forming ability and enhance drug-resistance to cisplatin-etoposide chemotherapy. Hence, we suggest that adherent cells might have functions similar to tumor stroma cells that promote cancer growth, metastasis, and drug resistance.
      • Orimo A.
      • Gupta P.B.
      • Sgroi D.C.
      • et al.
      Stromal fibroblasts present in invasive human breast carcinomas promote tumor growth and angiogenesis through elevated SDF-1/CXCL12 secretion.
      • Farmer P.
      • Bonnefoi H.
      • Anderle P.
      • et al.
      A stroma-related gene signature predicts resistance to neoadjuvant chemotherapy in breast cancer.
      • Kalluri R.
      The biology and function of fibroblasts in cancer.
      RB1 loss is observed in all SCLC transformed from acquired EGFR-TKI resistant lung adenocarcinoma.
      • Niederst M.J.
      • Sequist L.V.
      • Poirier J.T.
      • et al.
      RB loss in resistant EGFR mutant lung adenocarcinomas that transform to small-cell lung cancer.
      Its loss is frequently disrupted by genetic rearrangement in introns and cannot be detected by mutation of the RB1 gene.
      • George J.
      • Lim J.S.
      • Jang S.J.
      • et al.
      Comprehensive genomic profiles of small cell lung cancer.
      Protein expression of RB1 is a reliable alternative assay, in addition to RB1 mutation analysis.
      • Lee J.K.
      • Lee J.
      • Kim S.
      • et al.
      Clonal history and genetic predictors of transformation into small-cell carcinomas from lung adenocarcinomas.
      In our EGFR-mutant SCLCs, CLH5 had an exon 10-13 deletion of RB1 and CLH had a loss of RB1 protein expression, even though no nonsense mutation and deletion were found in the RB1 genome. Moreover, the difference in the expression of EGFR between NE and non-NE phenotypes in EGFR-mutant SCLC highlights that EGFR mutation is not the driver gene of SCLCs. In the phenotype of NE cells, EGFR expression was markedly reduced and also was insensitive to EGFR-TKI treatment. This result was similar to Niederst's finding.
      • Niederst M.J.
      • Sequist L.V.
      • Poirier J.T.
      • et al.
      RB loss in resistant EGFR mutant lung adenocarcinomas that transform to small-cell lung cancer.
      Our data also indicated that the downstream MEK/ERK pathway of EGFR signaling was activated in adherent cells, but not in suspensive cells of CLH and CLH5. In contract to the MEK/ERK pathway, the AKT/mTOR pathway was activated in both cell types. In general, EGFR/RTK downstream signaling, including PI3K/AKT/mTOR and Ras/Raf/MEK pathways, was inactivated in SCLC.
      • Byers L.A.
      • Wang J.
      • Nilsson M.B.
      • et al.
      Proteomic profiling identifies dysregulated pathways in small cell lung cancer and novel therapeutic targets including PARP1.
      Activation of the MEK/ERK and PI3K/AKT pathways played an important role after acquired resistance to chemotherapy or targeted therapy for SCLC.
      • Kraus A.C.
      • Ferber I.
      • Bachmann S.O.
      • et al.
      In vitro chemo- and radio-resistance in small cell lung cancer correlates with cell adhesion and constitutive activation of AKT and MAP kinase pathways.
      • Belyanskaya L.L.
      • Hopkins-Donaldson S.
      • Kurtz S.
      • et al.
      Cisplatin activates Akt in small cell lung cancer cells and attenuates apoptosis by survivin upregulation.
      • Tsurutani J.
      • West K.A.
      • Sayyah J.
      • et al.
      Inhibition of the phosphatidylinositol 3-kinase/Akt/mammalian target of rapamycin pathway but not the MEK/ERK pathway attenuates laminin-mediated small cell lung cancer cellular survival and resistance to imatinib mesylate or chemotherapy.
      • Marinov M.
      • Ziogas A.
      • Pardo O.E.
      • et al.
      AKT/mTOR pathway activation and BCL-2 family proteins modulate the sensitivity of human small cell lung cancer cells to RAD001.
      • Cardnell R.J.
      • Feng Y.
      • Mukherjee S.
      • et al.
      Activation of the PI3K/mTOR pathway following PARP inhibition in small cell lung cancer.
      Our results suggest that AKT/mTOR pathway activation not only conferred EGFR-TKI resistance but also accelerated the chemotherapy resistance of EGFR-mutant SCLC. PCA and hierarchical clustering analysis indicated that EGFR-mutant SCLC clustering differs from NSCLC and typical SCLC. MGH131-1 cells were more like CLH-A and CLH5-A, expressed more non-NE marker vimentin, and grew more adherent in culture. Although MGH131-1 exhibited more non-NE phenotype than MGH131-2, neither of these two cell lines expressed EGFR proteins.
      • Niederst M.J.
      • Sequist L.V.
      • Poirier J.T.
      • et al.
      RB loss in resistant EGFR mutant lung adenocarcinomas that transform to small-cell lung cancer.
      The morphologies of EGFR-mutant SCLCs are similar to those of typical SCLCs, but gene expression is different. These results may imply that EGFR-mutant SCLC has more phenotypic and functional plasticity than typical SCLC. It has been proposed that transformed EGFR-mutant SCLC comes from alveolar type II cells and typical SCLC originates from NE cells.
      • Oser M.G.
      • Niederst M.J.
      • Sequist L.V.
      • et al.
      Transformation from non-–small-cell lung cancer to small-cell lung cancer: molecular drivers and cells of origin.
      EMT is a critical process involved in cancer stemness, metastasis, and drug resistance.
      • Lamouille S.
      • Xu J.
      • Derynck R.
      Molecular mechanisms of epithelial-mesenchymal transition.
      • Nieto M.A.
      • Huang R.Y.
      • Jackson R.A.
      • et al.
      EMT: 2016.
      • Polyak K.
      • Weinberg R.A.
      Transitions between epithelial and mesenchymal states: acquisition of malignant and stem cell traits.
      • Shibue T.
      • Weinberg R.A.
      Metastatic colonization: settlement, adaptation and propagation of tumor cells in a foreign tissue environment.
      • Thiery J.P.
      • Sleeman J.P.
      Complex networks orchestrate epithelial-mesenchymal transitions.
      EMT was also described in SCLC cells.
      • Krohn A.
      • Ahrens T.
      • Yalcin A.
      • et al.
      Tumor cell heterogeneity in small cell lung cancer (SCLC): phenotypical and functional differences associated with epithelial-mesenchymal transition (EMT) and DNA methylation changes.
      In contrast to classic SCLC NCI-H69 cells, variant SCLC NCI-H69V cells exhibited EMT phenotypes and chemotherapy drug resistance. Both of our adherent cells expressed higher EMT-related transcription factors, such as Zeb1 and Slug, than suspended cells (Supplementary Fig. 6). Pathway analysis of gene expression also showed that cell adhesion, EMT, and extracellular matrix remodeling pathways are more activated in adherent cells (Supplementary Table 2). Our findings and those of previous studies all seem to show that lung adenocarcinoma patients with an EGFR mutation have poor clinical behavior and rapid relapse, and are more resistant to chemotherapies after SCLC transformation.
      • Jiang S.Y.
      • Zhao J.
      • Wang M.Z.
      • et al.
      Small-cell lung cancer transformation in patients with pulmonary adenocarcinoma: a case report and review of literature.
      Taken together, the phenotypic plasticity and chemotherapy resistance of EGFR-mutant SCLC may occur through EMT processes. In summary, this study reveals some molecular changes associated with EGFR-mutant SCLC that transformed from lung adenocarcinoma in patients that failed EGFR-TKI therapy. We suggest that the adherent cells of EGFR-mutant SCLCs function like tumor stroma cells and play important roles in tumor growth, metastasis, and drug resistance. The foundation treatment for SCLC is still chemotherapy but the response is far from satisfactory. Our finding may provide a clue for the potential application of the combination therapy with AKT inhibitors or HDACi for patients with EGFR-mutant SCLC.

      Acknowledgments

      Accession code of Gene Expression Omnibus (GEO): GSE122698. This study was supported by grants from the Ministry of Science and Technology of the Republic of China, Taiwan (NSC 103-2314-B-002-142, NSC 104-2314-B-002-180-MY3). The authors thank the Pharmacogenomics Core Laboratory of National Core Facility for Biopharmaceuticals for technical support.

      Supplementary Data

      References

        • Ohashi K.
        • Maruvka Y.E.
        • Michor F.
        • et al.
        Epidermal growth factor receptor tyrosine kinase inhibitor-resistant disease.
        J Clin Oncol. 2013; 31: 1070-1080
        • Yu H.A.
        • Arcila M.E.
        • Rekhtman N.
        • et al.
        Analysis of tumor specimens at the time of acquired resistance to EGFR-TKI therapy in 155 patients with EGFR-mutant lung cancers.
        Clini Cancer Res. 2013; 19: 2240-2247
        • Sequist L.V.
        • Waltman B.A.
        • Dias-Santagata D.
        • et al.
        Genotypic and histological evolution of lung cancers acquiring resistance to EGFR inhibitors.
        Sci Transl Med. 2011; 3: 75ra26
        • Ahn S.
        • Hwang S.H.
        • Han J.
        • et al.
        Transformation to small cell lung cancer of pulmonary adenocarcinoma: clinicopathologic analysis of six cases.
        J Pathol Transl Med. 2016; 50: 258-263
        • Davidson M.R.
        • Gazdar A.F.
        • Clarke B.E.
        The pivotal role of pathology in the management of lung cancer.
        J Thorac Dis. 2013; 5: S463-S478
        • Oser M.G.
        • Niederst M.J.
        • Sequist L.V.
        • et al.
        Transformation from non-–small-cell lung cancer to small-cell lung cancer: molecular drivers and cells of origin.
        Lancet Oncol. 2015; 16: e165-e172
        • Pettengill O.S.
        • Sorenson G.D.
        • Wurster-Hill D.H.
        • et al.
        Isolation and growth characteristics of continuous cell lines from small-cell carcinoma of the lung.
        Cancer. 1980; 45: 906-918
        • Doyle L.A.
        • Borges M.
        • Hussain A.
        • et al.
        An adherent subline of a unique small-cell lung cancer cell line downregulates antigens of the neural cell adhesion molecule.
        J Clin Invest. 1990; 86: 1848-1854
        • Sutherland K.D.
        • Proost N.
        • Brouns I.
        • et al.
        Cell of origin of small cell lung cancer: inactivation of Trp53 and Rb1 in distinct cell types of adult mouse lung.
        Cancer Cell. 2011; 19: 754-764
        • Calbo J.
        • van Montfort E.
        • Proost N.
        • et al.
        A functional role for tumor cell heterogeneity in a mouse model of small cell lung cancer.
        Cancer Cell. 2011; 19: 244-256
        • Ho C.C.
        • Liao W.Y.
        • Lin C.A.
        • et al.
        Acquired BRAF V600E Mutation as resistant mechanism after treatment with osimertinib.
        J Thorac Oncol. 2017; 12: 567-572
        • Liu X.
        • Ory V.
        • Chapman S.
        • et al.
        ROCK inhibitor and feeder cells induce the conditional reprogramming of epithelial cells.
        Am J Pathol. 2012; 180: 599-607
        • Niederst M.J.
        • Sequist L.V.
        • Poirier J.T.
        • et al.
        RB loss in resistant EGFR mutant lung adenocarcinomas that transform to small-cell lung cancer.
        Nat Comm. 2015; 6: 6377
        • Barretina J.
        • Caponigro G.
        • Stransky N.
        • et al.
        The Cancer Cell Line Encyclopedia enables predictive modelling of anticancer drug sensitivity.
        Nature. 2012; 483: 603-607
        • Wu H.-M.
        • Tien Y.-J.
        • Chen C.-h.
        GAP:a graphical environment for matrix visualization and cluster analysis.
        Comp Stat Data Analysis. 2010; 54: 767-778
        • Garnett M.J.
        • Edelman E.J.
        • Heidorn S.J.
        • et al.
        Systematic identification of genomic markers of drug sensitivity in cancer cells.
        Nature. 2012; 483: 570-575
        • Su K.Y.
        • Chen H.Y.
        • Li K.C.
        • et al.
        Pretreatment epidermal growth factor receptor (EGFR) T790M mutation predicts shorter EGFR tyrosine kinase inhibitor response duration in patients with non-small-cell lung cancer.
        J Clin Oncol. 2012; 30: 433-440
        • Kwon M.C.
        • Proost N.
        • Song J.Y.
        • et al.
        Paracrine signaling between tumor subclones of mouse SCLC: a critical role of ETS transcription factor Pea3 in facilitating metastasis.
        Genes Develop. 2015; 29: 1587-1592
        • van Meerbeeck J.P.
        • Fennell D.A.
        • De Ruysscher D.K.
        Small-cell lung cancer.
        Lancet. 2011; 378: 1741-1755
        • Peifer M.
        • Fernandez-Cuesta L.
        • Sos M.L.
        • et al.
        Integrative genome analyses identify key somatic driver mutations of small-cell lung cancer.
        Nat Gen. 2012; 44: 1104-1110
        • Rudin C.M.
        • Durinck S.
        • Stawiski E.W.
        • et al.
        Comprehensive genomic analysis identifies SOX2 as a frequently amplified gene in small-cell lung cancer.
        Nat Gen. 2012; 44: 1111-1116
        • George J.
        • Lim J.S.
        • Jang S.J.
        • et al.
        Comprehensive genomic profiles of small cell lung cancer.
        Nature. 2015; 524: 47-53
        • Sordella R.
        • Bell D.W.
        • Haber D.A.
        • et al.
        Gefitinib-sensitizing EGFR mutations in lung cancer activate anti-apoptotic pathways.
        Science. 2004; 305: 1163-1167
        • Sos M.L.
        • Koker M.
        • Weir B.A.
        • et al.
        PTEN loss contributes to erlotinib resistance in EGFR-mutant lung cancer by activation of Akt and EGFR.
        Cancer Res. 2009; 69: 3256-3261
        • Huang L.
        • Fu L.
        Mechanisms of resistance to EGFR tyrosine kinase inhibitors.
        Acta Pharm Sinica B. 2015; 5: 390-401
        • Kraus A.C.
        • Ferber I.
        • Bachmann S.O.
        • et al.
        In vitro chemo- and radio-resistance in small cell lung cancer correlates with cell adhesion and constitutive activation of AKT and MAP kinase pathways.
        Oncogene. 2002; 21: 8683-8695
        • Belyanskaya L.L.
        • Hopkins-Donaldson S.
        • Kurtz S.
        • et al.
        Cisplatin activates Akt in small cell lung cancer cells and attenuates apoptosis by survivin upregulation.
        Int J Cancer. 2005; 117: 755-763
        • Makinoshima H.
        • Umemura S.
        • Suzuki A.
        • et al.
        Metabolic determinants of sensitivity to phosphatidylinositol 3-kinase pathway inhibitor in small-cell lung carcinoma.
        Cancer Res. 2018; 78: 2179-2190
        • Jones P.A.
        • Issa J.P.
        • Baylin S.
        Targeting the cancer epigenome for therapy.
        Nat Rev Gen. 2016; 17: 630-641
        • Platta C.S.
        • Greenblatt D.Y.
        • Kunnimalaiyaan M.
        • et al.
        The HDAC inhibitor trichostatin A inhibits growth of small cell lung cancer cells.
        J Surg Res. 2007; 142: 219-226
        • Crisanti M.C.
        • Wallace A.F.
        • Kapoor V.
        • et al.
        The HDAC inhibitor panobinostat (LBH589) inhibits mesothelioma and lung cancer cells in vitro and in vivo with particular efficacy for small cell lung cancer.
        Mol Cancer Ther. 2009; 8: 2221-2231
        • Hubaux R.
        • Vandermeers F.
        • Crisanti M.C.
        • et al.
        Preclinical evidence for a beneficial impact of valproate on the response of small cell lung cancer to first-line chemotherapy.
        Eur J Cancer. 2010; 46: 1724-1734
        • Pan C.H.
        • Chang Y.F.
        • Lee M.S.
        • et al.
        Vorinostat enhances the cisplatin-mediated anticancer effects in small cell lung cancer cells.
        BMC Cancer. 2016; 16: 857
        • Sabari J.K.
        • Lok B.H.
        • Laird J.H.
        • et al.
        Unravelling the biology of SCLC: implications for therapy.
        Nat Rev Clin Oncol. 2017; 14: 549-561
        • Orimo A.
        • Gupta P.B.
        • Sgroi D.C.
        • et al.
        Stromal fibroblasts present in invasive human breast carcinomas promote tumor growth and angiogenesis through elevated SDF-1/CXCL12 secretion.
        Cell. 2005; 121: 335-348
        • Farmer P.
        • Bonnefoi H.
        • Anderle P.
        • et al.
        A stroma-related gene signature predicts resistance to neoadjuvant chemotherapy in breast cancer.
        Nat Med. 2009; 15: 68-74
        • Kalluri R.
        The biology and function of fibroblasts in cancer.
        Nat Rev Cancer. 2016; 16: 582-598
        • Lee J.K.
        • Lee J.
        • Kim S.
        • et al.
        Clonal history and genetic predictors of transformation into small-cell carcinomas from lung adenocarcinomas.
        J Clin Oncol. 2017; 35: 3065-3074
        • Byers L.A.
        • Wang J.
        • Nilsson M.B.
        • et al.
        Proteomic profiling identifies dysregulated pathways in small cell lung cancer and novel therapeutic targets including PARP1.
        Cancer Discov. 2012; 2: 798-811
        • Tsurutani J.
        • West K.A.
        • Sayyah J.
        • et al.
        Inhibition of the phosphatidylinositol 3-kinase/Akt/mammalian target of rapamycin pathway but not the MEK/ERK pathway attenuates laminin-mediated small cell lung cancer cellular survival and resistance to imatinib mesylate or chemotherapy.
        Cancer Res. 2005; 65: 8423-8432
        • Marinov M.
        • Ziogas A.
        • Pardo O.E.
        • et al.
        AKT/mTOR pathway activation and BCL-2 family proteins modulate the sensitivity of human small cell lung cancer cells to RAD001.
        Clin Cancer Res. 2009; 15: 1277-1287
        • Cardnell R.J.
        • Feng Y.
        • Mukherjee S.
        • et al.
        Activation of the PI3K/mTOR pathway following PARP inhibition in small cell lung cancer.
        PloS One. 2016; 11: e0152584
        • Lamouille S.
        • Xu J.
        • Derynck R.
        Molecular mechanisms of epithelial-mesenchymal transition.
        Nat Rev Mol Cell Biol. 2014; 15: 178-196
        • Nieto M.A.
        • Huang R.Y.
        • Jackson R.A.
        • et al.
        EMT: 2016.
        Cell. 2016; 166: 21-45
        • Polyak K.
        • Weinberg R.A.
        Transitions between epithelial and mesenchymal states: acquisition of malignant and stem cell traits.
        Nat Rev Cancer. 2009; 9: 265-273
        • Shibue T.
        • Weinberg R.A.
        Metastatic colonization: settlement, adaptation and propagation of tumor cells in a foreign tissue environment.
        Semin. Cancer Biol. 2011; 21: 99-106
        • Thiery J.P.
        • Sleeman J.P.
        Complex networks orchestrate epithelial-mesenchymal transitions.
        Nat Rev Mol Cell Biol. 2006; 7: 131-142
        • Krohn A.
        • Ahrens T.
        • Yalcin A.
        • et al.
        Tumor cell heterogeneity in small cell lung cancer (SCLC): phenotypical and functional differences associated with epithelial-mesenchymal transition (EMT) and DNA methylation changes.
        PloS One. 2014; 9: e100249
        • Jiang S.Y.
        • Zhao J.
        • Wang M.Z.
        • et al.
        Small-cell lung cancer transformation in patients with pulmonary adenocarcinoma: a case report and review of literature.
        Medicine. 2016; 95: e2752