Regulation of the stemness factor, SOX2, by cytokine stimuli controls self-renewal and differentiation in cells. Activating mutations in EGFR are proven therapeutic targets for tyrosine kinase inhibitors (TKI) in lung adenocarcinoma, but acquired resistance to TKIs inevitably occurs. The mechanism by which stemness and differentiation signaling emerge in lung cancers to affect TKI tolerance and lung cancer dissemination has yet to be elucidated. Here, we report that cross-talk between SOX2 and TGFβ signaling affects lung cancer cell plasticity and TKI tolerance. TKI treatment favored selection of lung cancer cells displaying mesenchymal morphology with deficient SOX2 expression, whereas SOX2 expression promoted TKI sensitivity and inhibited the mesenchymal phenotype. Preselection of EGFR-mutant lung cancer cells with the mesenchymal phenotype diminished SOX2 expression and TKI sensitivity, whereas SOX2 silencing induced vimentin, but suppressed BCL2L11, expression and promoted TKI tolerance. TGFβ stimulation downregulated SOX2 and induced epithelial-to-mesenchymal transdifferentiation accompanied by increased TKI tolerance, which can interfere with ectopic SOX2 expression. SOX2-positive lung cancer cells exhibited a lower dissemination capacity than their SOX2-negative counterparts. Tumors expressing low SOX2 and high vimentin signature were associated with worse survival outcomes in patients with EGFR mutations. These findings provide insights into how cancer cell plasticity regulated by SOX2 and TGFβ signaling affects EGFR-TKI tolerance and lung cancer dissemination.

Significance:

These findings suggest the potential of SOX2 as a prognostic marker in EGFR-mutant lung cancer, as SOX2-mediated cell plasticity regulated by TGFβ stimulation and epigenetic control affects EGFR-TKI tolerance and cancer dissemination.

SOX2 belongs to the SOX (Sry-related HMG Box) family of proteins and is an important transcription factor that regulates self-renewal in embryonic stem cells (ESC). Its downregulation in response to cytokine stimulation determines the timing and degree of differentiation (1). Responding to respiratory tract injuries, SOX2 signaling initiates the proliferation and differentiation of lung progenitor cells to maintain tissue homeostasis (2, 3). SOX2, in conjunction with OCT4, KLF4, and MYC, can reverse the mesenchymal morphology of fibroblasts and reprogram them into induced pluripotent stem cells (iPSC; refs. 4, 5). Moreover, inhibition of TGFβ signaling facilitates the SOX2-mediated reprogramming process of fibroblasts, whereas activation of TGFβ blocks the reprogramming process (6, 7).

Cancer cell plasticity is characterized as a phenotypic switch between the mesenchymal and epithelial states. This switch is accompanied by signaling pathway alterations and has been proposed to both generate heterogeneity and mediate tumor progression (8, 9). SOX2 serves as a nodal epigenetic regulator in determining the mesenchymal-to-epithelial transdifferentiation (MET) of lung cancer cells (9). In addition, SOX2 promotes EGFR expression via a positive feedback manner in lung cancer (10).

Activating mutations (exon 19 deletions and L858R mutation) in EGFR are predictive markers and therapeutic targets of tyrosine kinase inhibitors (TKI) in the treatment of patients with lung cancer bearing mutations in EGFR (11–16). Despite initial positive responses to treatment, patients eventually develop resistance to EGFR–TKIs. The mutations, EGFR-T790M and EGFR–C797S, confer acquired resistance to gefitinib and osimertinib, respectively (17–21). However, the mutation EGFR-T790M correlates negatively with distant metastasis and confers better patient survival (22–25), suggesting that an EGFR mutation–independent TKI resistance mechanism promotes a worse survival outcome. Epithelial-to-mesenchymal transition (EMT), a reverse process of MET and preinvasive status in cancer plasticity, has been linked to EGFR–TKI resistance (26–29). Neuroendocrine transformation of adenocarcinoma has also been detected in lung tumors after EGFR–TKI treatment (19). A drug-tolerant state is essential for the development of acquired resistance to EGFR–TKI (30), and deficient expression of BCL2L11 has been linked to EGFR–TKI tolerance in lung cancer (31).

Because SOX2 signaling is highly expressed in both stem cells and lung cancer cells, this study aimed to test whether the mechanism behind SOX2-regulated stem cell differentiation and reprogramming was shared by lung cancer cells to affect TKI tolerance and cancer dissemination. We further characterized how cross-talk between SOX2 stemness and TGFβ cytokine signaling generates lung cancer cell plasticity with distinct TKI tolerance and dissemination patterns.

Cell culture

HCC827 and H1975 cells were obtained from Dr. Jeff Wang (Development Center for Biotechnology) and Dr. Wayne Chang (National Health Research Institutes, Zhunan, Miaoli, Taiwan), respectively (9, 10). Human fibroblasts and fibroblast-derived iPSCs were described previously (32). HCC827GR and H1975AZDR cells were established in our laboratory by exposure of HCC827 and H1975 cells to stepwise increased concentrations of gefitinib and osimertinib, respectively (29). All lung cancer cells were tested positive for human origin, carrying specific EGFR mutations, by EGFR-sequencing analysis (Supplementary Fig. S1A–S1C). All lung cancer cells were cultured in RPMI1640 medium containing l-glutamine (4 mmol/L), sodium pyruvate (1 mmol/L), HEPES (10 mmol/L), and FBS (10%).

qPCR and chromatin immunoprecipitation-qPCR

The qPCR and chromatin immunoprecipitation (ChIP)-qPCR assays were performed as described previously (9). Primer sequences and probes used in qPCR are listed in Supplementary Table S1. Detailed materials for the ChIP-qPCR are described in Supplementary Tables S2 and S3.

Chemicals and reagents

EGFR–TKIs (gefitinib, erlotinib, afatinib, and osimertinib) were purchased from Cayman Chem. Recombinant human TGFβ was obtained from Sino Biological. Trichostatin A was obtained from Sigma. Romidepsin and SB-431542 were purchased from MedChemExpress. Short hairpin RNA (shRNA) clones were ordered from the National RNAi Core Facility, Academia Sinica (Nangang, Taipei, Taiwan). Detailed information of shRNA clones is listed in Supplementary Table S4.

Electrical cell-substrate impedance sensing assay

The electric cell-substrate impedance sensing (ECIS) assay was performed as described previously (29).

Tissue samples and public domain data analysis

The specimens used in IHC were obtained from surgery or biopsy at Taipei Medical University Hospital (Xinyi, Taipei, Taiwan). Approval for this study was granted by the institutional review board protocol number CRC-04-11-05. The public gene expression profiling datasets used in this study were analyzed as described previously (9). The gene expression profiling data of the xenograft mouse model were derived from the report of Bivona and colleagues (33). The sources of these gene expression profiling datasets are listed in Supplementary Table S5.

IHC

Scores of immunoreactivity pattern of all tissues from patients were examined at the Department of Pathology, Taipei Medical University Hospital (Xinyi, Taipei, Taiwan). The Allred scoring system was used to give the staining scores for the expression of SOX2 or vimentin based on the intensity of the staining (on a scale from 0 to 3; ref. 11). The characteristics of patients are listed in Supplementary Table S6.

Statistical analysis

Overall survival curves were estimated by the Kaplan–Meier method. The differences of overall survival between the high- and low-gene–expressing groups were compared by log-rank test. All statistical analyses were performed using SPSS software, version 16 (SPSS, Inc.). P < 0.05 was considered to reach a statistically significant difference.

SOX2 downregulation and vimentin upregulation during stem cell differentiation and EGFR–TKI tolerance development

To study the effect of SOX2 expression on the mesenchymal phenotype during stem cell differentiation and reprogramming, we monitored SOX2 and vimentin (a mesenchymal marker) expression in ESCs, fibroblasts, and iPSCs. We observed that during the differentiation of ESCs to fibroblasts, SOX2 was downregulated and vimentin was upregulated (Fig. 1A). TGFβ signaling antagonizes the SOX2-mediated reprogramming process of fibroblasts (6, 7). Gene expression profiling analysis revealed that TGFβ receptors (TGFBR1 and TGFBR2) and ligands (TGFB1, TGFB2, and TGFB3) were upregulated during ESC differentiation to fibroblasts (Supplementary Fig. S2A and S2B). In contrast, during the reprogramming of fibroblasts toward iPSCs, SOX2 and vimentin expression levels were reversed, accompanied by decreased expression of TGFβ receptors and ligands (Fig. 1A; Supplementary Fig. S2A and S2B). We previously demonstrated that SOX2 regulates EGFR signaling and maintains the epithelial feature in lung cancer cells (9, 10). To verify whether the interplay between SOX2 and vimentin expression also exists during the development of EGFR–TKI tolerance in lung cancer cells, we treated EGFR-mutant HCC827 (delE746_A750) and H1975 (L858R/T790M) lung adenocarcinoma cells with incremental concentrations of gefitinib and osimertinib, respectively, for 6 months. The surviving cells were pooled, propagated, and then further named HCC827GR and H1975AZDR. According to phase-contrast imaging, HCC827GR and H1975AZDR cells displayed a spindle-like phenotype, which was significantly different from that of the parental HCC827 and H1975 cells (Fig. 1B). Clonogenic analysis confirmed that HCC827GR cells were more tolerant to gefitinib than their parental cells (Fig. 1C, top). The same results were obtained regarding H1975AZDR cells, which were more tolerant to osimertinib than their parental cells (Fig. 1C, bottom). HCC827GR and H1975AZDR cells, although tolerant to EGFR–TKIs, did not acquire the EGFR mutations, T790M and C797S, respectively (Supplementary Fig. S1A–S1C). We examined SOX2 expression in the paired EGFR–TKI-sensitive (HCC827 and H1975) and -tolerant (HCC827GR and H1975AZDR) cells. qPCR and immunoblotting assays revealed that SOX2 expression was downregulated in both HCC827GR and H1975AZDR cells (Fig. 1D). Gene set enrichment analysis revealed that the EMT pathway was enriched in EGFR–TKI-tolerant cells (Supplementary Fig. S3A and S3B). We confirmed that vimentin was upregulated and E-cadherin was downregulated in HCC827GR and H1975AZDR as compared with their parental cells (Fig. 1E; Supplementary Fig. S4A and S4B). qPCR assays revealed that TGFBR1/2 and TGFB1/2/3 levels were upregulated in both HCC827GR and H1975AZDR (Supplementary Fig. S5A and S5B). Epigenetic modification of H3K27ac and H3K4me3, which marks the active enhancer and promoter, respectively, is highly involved in stem cell differentiation. ChIP-sequencing (ChIP-seq) analysis indicated that the SOX2 locus in both ESCs and iPSCs exhibited strong H3K27ac and H3K4me3 signals, which were absent in fibroblasts (Supplementary Fig. S6A and S6B). ChIP-qPCR revealed that H3K27ac and H3K4me3 signals were remarkably higher along the SOX2 locus in HCC827 compared with its TKI-tolerant counterpart, HCC827GR (Fig. 1F; Supplementary Fig. S6C). In contrast, the TGFBR1 and TGFBR2 loci in fibroblasts exhibited strong H3K27ac signals, which were absent in ESCs and iPSCs (Supplementary Fig. S6D). These data suggest the involvement of epigenetic silencing of SOX2 during stem cell differentiation and EGFR–TKI tolerance development.

Figure 1.

SOX2 is silenced in EGFR–TKI-selected lung cancer cells. A, RNA-seq (GSE73211; top) and qPCR (bottom) analysis to assess SOX2 and vimentin (VIM) expression in ESCs (HUES), fibroblasts, and reprogramed iPSCs. B, Representative phase contrast images of HCC827 versus HCC827GR and H1975 versus H1975AZDR cells. Scale bar, 100 μm. C, Clonogenic analysis of HCC827 versus HCC827GR (top) and H1975 versus H1975AZDR (bottom) cells treated with indicated concentrations of gefitinib and osimertinib, respectively, for 10 days. Photographs represent the growth of cells stained by crystal violet. D, qPCR (top) and immunoblotting (bottom) assays to assess SOX2 expression in HCC827 versus HCC827GR (left) and H1975 versus H1975AZDR (right). E, qPCR analysis to assess E-cadherin (E-cad) and vimentin expression in HCC827 versus HCC827GR (left) and H1975 versus H1975AZDR (right) cells. F, ChIP-qPCR analysis to assess the H3K27ac signal at the SOX2 locus in HCC827 and HCC827GR cells. *, P < 0.05; **, P < 0.01; ***, P < 0.001.

Figure 1.

SOX2 is silenced in EGFR–TKI-selected lung cancer cells. A, RNA-seq (GSE73211; top) and qPCR (bottom) analysis to assess SOX2 and vimentin (VIM) expression in ESCs (HUES), fibroblasts, and reprogramed iPSCs. B, Representative phase contrast images of HCC827 versus HCC827GR and H1975 versus H1975AZDR cells. Scale bar, 100 μm. C, Clonogenic analysis of HCC827 versus HCC827GR (top) and H1975 versus H1975AZDR (bottom) cells treated with indicated concentrations of gefitinib and osimertinib, respectively, for 10 days. Photographs represent the growth of cells stained by crystal violet. D, qPCR (top) and immunoblotting (bottom) assays to assess SOX2 expression in HCC827 versus HCC827GR (left) and H1975 versus H1975AZDR (right). E, qPCR analysis to assess E-cadherin (E-cad) and vimentin expression in HCC827 versus HCC827GR (left) and H1975 versus H1975AZDR (right) cells. F, ChIP-qPCR analysis to assess the H3K27ac signal at the SOX2 locus in HCC827 and HCC827GR cells. *, P < 0.05; **, P < 0.01; ***, P < 0.001.

Close modal

SOX2 downregulation in lung cancer cells with intrinsic EGFR–TKI tolerance

To verify whether EGFR–TKI-tolerant cells were present within the heterogeneous population of cancer cells, we selected single clones exhibiting the mesenchymal phenotype from EGFR–TKI treatment–naïve lung cancer cells. Two single-cell clones (M-1 and M-2) that displayed the mesenchymal phenotype were isolated from two 96-well plates of HCC827 cells (Fig. 2A). qPCR analysis revealed that SOX2 and E-cadherin were downregulated in M-1 and M-2, whereas vimentin, TGFBR1/2, and TGFB1/2/3 levels were upregulated relative to HCC827 (Fig. 2B; Supplementary Fig. S7A and S7B). The clonogenic assay revealed that M-1 and M-2 were more tolerant to EGFR–TKIs, including gefitinib, erlotinib, afatinib, and osimertinib, in comparison with parental HCC827 (Fig. 2C; Supplementary Fig. S8). As mentioned previously, H3K27ac and H3K4me3 signals at the SOX2 locus were lower in fibroblasts and EGFR–TKI-tolerant lung cancer cells compared with ESCs and EGFR–TKI-sensitive lung cancer cells, respectively. It was also observed that H3K27ac and H3K4me3 signals at the SOX2 locus were lower in EGFR–TKI treatment–naïve M-1 and M-2 cells than in the parental HCC827 (Fig. 2D; Supplementary Fig. S9). Trichostatin A, a pan-histone deacetylase (HDAC) inhibitor, and romidepsin, a HDAC1/2-specific inhibitor, induce stem cell differentiation, cell-cycle arrest, and cancer heterogeneity (9, 34, 35). Accordingly, the effect of Trichostatin A and romidepsin on SOX2 and vimentin expression was examined in EGFR-mutant HCC827 cells. We observed that treatment with trichostatin A or romidepsin induced SOX2 downregulation and increased vimentin expression in HCC827 cells (Fig. 2E). Cells pretreated with trichostatin A or romidepsin were further subjected to the gefitinib tolerance assay, and it was found that trichostatin A or romidepsin pretreatment enhanced gefitinib tolerance in HCC827 (Fig. 2F). Moreover, overexpression of SOX2 rendered romidepsin-pretreated cells more sensitive to EGFR–TKI (Supplementary Fig. S10A and S10B). ChIP-seq analysis indicated that the TGFB1, TGFB2, TGFB3, TGFBR1, and TGFBR2 loci possess potential HDAC1-binding sites (Supplementary Fig. S11A). We observed that both romidepsin treatment and HDAC1 silencing could induce TGFBR1/2 and TGFB1/2/3 expression in lung cancer cells (Supplementary Figs. S11B and S12A and S12B). These results indicate the potential involvement of SOX2 downregulation and TGFβ signaling in EGFR–TKI tolerance.

Figure 2.

Loss of SOX2 expression in EGFR-mutant lung cancer cells with intrinsic TKI resistance. A, Phase contrast images of HCC827 and its single-cell derivatives, M-1 and M-2. Scale bar, 100 μm. B, qPCR analysis of vimentin (VIM; left), E-cadherin (E-cad; middle), and SOX2 (right) expression in HCC827, M-1, and M-2 cells. C, Clonogenic analysis of HCC827, M-1, and M-2 cells treated with indicated concentrations of gefitinib (left) and osimertinib (right) for 7 days. D, ChIP-qPCR assays to assess the H3K27ac signal at SOX2 locus in HCC827, M-1, and M-2 cells. E, qPCR analysis to access SOX2 and vimentin expression in HCC827 cells treated with trichostatin A (TSA; 500 nmol/L; left) or romidepsin (1 nmol/L; right) for 2 and 7 days, respectively. F, Clonogenic analysis of HCC827 cells pretreated with trichostatin A (500 nmol/L; left) or romidepsin (1 nmol/L; right) for 2 and 7 days, respectively, followed by gefitinib treatment (1 μmol/L) for 14 days. *, P < 0.05; **, P < 0.01; ***, P < 0.001; N.S., not significant.

Figure 2.

Loss of SOX2 expression in EGFR-mutant lung cancer cells with intrinsic TKI resistance. A, Phase contrast images of HCC827 and its single-cell derivatives, M-1 and M-2. Scale bar, 100 μm. B, qPCR analysis of vimentin (VIM; left), E-cadherin (E-cad; middle), and SOX2 (right) expression in HCC827, M-1, and M-2 cells. C, Clonogenic analysis of HCC827, M-1, and M-2 cells treated with indicated concentrations of gefitinib (left) and osimertinib (right) for 7 days. D, ChIP-qPCR assays to assess the H3K27ac signal at SOX2 locus in HCC827, M-1, and M-2 cells. E, qPCR analysis to access SOX2 and vimentin expression in HCC827 cells treated with trichostatin A (TSA; 500 nmol/L; left) or romidepsin (1 nmol/L; right) for 2 and 7 days, respectively. F, Clonogenic analysis of HCC827 cells pretreated with trichostatin A (500 nmol/L; left) or romidepsin (1 nmol/L; right) for 2 and 7 days, respectively, followed by gefitinib treatment (1 μmol/L) for 14 days. *, P < 0.05; **, P < 0.01; ***, P < 0.001; N.S., not significant.

Close modal

SOX2 expression inhibits EMT and promotes EGFR–TKI sensitivity

Because BCL2L11 downregulation is associated with EGFR–TKI tolerance (31), we further evaluated BCL2L11 expression in TKI-sensitive and -tolerant cells. qPCR analysis demonstrated that compared with H1975 and HCC827 cells, BCL2L11 expression was reduced in H1975AZDR and HCC827GR cells (Supplementary Fig. S13A). Gene expression profiling and qPCR analysis showed that BCL2L11 expression was downregulated during ESCs differentiation to fibroblasts and upregulated during SOX2-mediated reprogramming of fibroblasts to iPSCs (Supplementary Fig. S13B). Knockdown of BCL2L11 enriched the culture for cells harboring lower SOX2 expression accompanied by higher tolerance to EGFR–TKIs compared with parental HCC827 cells, while these BCL2L11-silenced cells displayed a slow-growing phenotype (Supplementary Fig. S13C–S13F). Overexpression of BCL2L11 increased sensitivity to EGFR–TKIs in HCC827GR and H1975AZDR (Supplementary Fig. S13G and S13H). Kaplan–Meier survival analysis revealed that low BCL2L11 expression predicted a low recurrence-free survival rate in patients with non–small cell lung cancer (NSCLC; Supplementary Fig. S14A). To study the effect of SOX2 expression on EGFR–TKI tolerance, BCL2L11 expression, and EMT, the SOX2 gene was knocked down in HCC827. qPCR analysis revealed that SOX2 silencing decreased BCL2L11 expression (Supplementary Fig. S14B). ChIP-qPCR analysis revealed that the amount of SOX2 bound to the BCL2L11 promoter was decreased in EGFR–TKI-tolerant cells compared with their parental cells (Supplementary Fig. S14C). Moreover, SOX2 knockdown induced vimentin expression (Fig. 3A, left and middle). HCC827 cells, in which SOX2 had been silenced, were pooled and subjected to EGFR–TKI treatment. The clonogenic assay determined that SOX2 knockdown enriched EGFR–TKI-tolerant cells (Fig. 3A, right). To test whether SOX2 expression prevents EMT and affects EGFR–TKI sensitivity in EGFR-mutant lung cancer cells, we overexpressed SOX2 in HCC827, followed by EGFR–TKI treatment. qPCR assays showed that vimentin was downregulated upon SOX2 overexpression in cells (Fig. 3B, left). Clonogenic analysis revealed that SOX2 upregulation increased EGFR–TKI sensitivity in HCC827 (Fig. 3B, middle and right). In addition, to study the short-term effect of EGFR–TKI treatment on SOX2 expression, we monitored SOX2 expression in lung cancer cells with mutated EGFR under EGFR–TKI treatment for 1 day. qPCR analysis revealed that SOX2 expression was induced by osimertinib in HCC827, but not in SOX2-negative EGFR–TKI-tolerant HCC827GR (Supplementary Fig. S15A–S15C). This indicates the presence of cross-talk between the EGFR mutation and SOX2 signaling in TKI-sensitive cells. To gain further insight into the role of SOX2 in EMT and EGFR–TKI tolerance, HCC827 cells were exposed to gefitinib for 2 weeks. qPCR assays revealed that gefitinib treatment selected cells harboring low SOX2 and high vimentin expression (Fig. 3C). These cells were further pooled and named HCC827GRs. To validate the role of SOX2 expression in the prevention of EMT and EGFR–TKI tolerance, we expressed SOX2 ectopically in HCC827GRs. qPCR assays demonstrated that SOX2 expression inhibited vimentin, but induced E-cadherin expression in HCC827GRs (Fig. 3D, left; Supplementary Fig. S16A and S16B). Clonogenic assays revealed that SOX2 expression decreased the tolerance of HCC827GRs cells under gefitinib treatment (Fig. 3D, right). Altogether, these data support the conclusion that SOX2 expression inhibits the mesenchymal phenotype and decreases EGFR–TKI tolerance in lung cancer cells with mutated EGFR.

Figure 3.

Effect of SOX2 expression on EGFR–TKI tolerance. A, qPCR analysis to assess SOX2 (left) and vimentin (VIM) (middle) expression in HCC827 cells transduced with the lentiviral vector encoding shRNA against SOX2 or scrambled control (SC) shRNA. shSOX2#1 and shSOX2#2 target different regions in SOX2 mRNA. Right, clonogenic analysis of HCC827-SC and HCC827-shSOX2 under the treatment of gefitinib (1 μmol/L) for 14 days. B, Left, qPCR analysis to assess SOX2 and vimentin expression in HCC827 cells transduced with the lentiviral vector encoding SOX2 cDNA (HCC827-SOX2) or empty control (HCC827-Ctrl). Clonogenic analysis (middle and right) of HCC827-Ctrl and HCC827-SOX2 cells under the treatment of gefitinib (100 nmol/L) for 14 days. C, qPCR analysis to assess SOX2 (left) and vimentin (right) expression in survived HCC827 cells after gefitinib (100 nmol/L) treatment for 14 days. D, Left, qPCR analysis of SOX2 and vimentin expression in HCC827GRs transduced with the lentiviral vector encoding SOX2 cDNA (HCC827GRs-SOX2) or empty control vector (HCC827GRs-Ctrl). HCC827GRs cells were derived from survived HCC827 cells under gefitinib (100 nmol/L) treatment for 14 days. Right, clonogenic analysis of HCC827GRs-Ctrl and HCC827GRs-SOX2 cells under gefitinib (100 nmol/L) treatment for 14 days. *, P < 0.05; **, P < 0.01; ***, P < 0.001.

Figure 3.

Effect of SOX2 expression on EGFR–TKI tolerance. A, qPCR analysis to assess SOX2 (left) and vimentin (VIM) (middle) expression in HCC827 cells transduced with the lentiviral vector encoding shRNA against SOX2 or scrambled control (SC) shRNA. shSOX2#1 and shSOX2#2 target different regions in SOX2 mRNA. Right, clonogenic analysis of HCC827-SC and HCC827-shSOX2 under the treatment of gefitinib (1 μmol/L) for 14 days. B, Left, qPCR analysis to assess SOX2 and vimentin expression in HCC827 cells transduced with the lentiviral vector encoding SOX2 cDNA (HCC827-SOX2) or empty control (HCC827-Ctrl). Clonogenic analysis (middle and right) of HCC827-Ctrl and HCC827-SOX2 cells under the treatment of gefitinib (100 nmol/L) for 14 days. C, qPCR analysis to assess SOX2 (left) and vimentin (right) expression in survived HCC827 cells after gefitinib (100 nmol/L) treatment for 14 days. D, Left, qPCR analysis of SOX2 and vimentin expression in HCC827GRs transduced with the lentiviral vector encoding SOX2 cDNA (HCC827GRs-SOX2) or empty control vector (HCC827GRs-Ctrl). HCC827GRs cells were derived from survived HCC827 cells under gefitinib (100 nmol/L) treatment for 14 days. Right, clonogenic analysis of HCC827GRs-Ctrl and HCC827GRs-SOX2 cells under gefitinib (100 nmol/L) treatment for 14 days. *, P < 0.05; **, P < 0.01; ***, P < 0.001.

Close modal

TGFβ stimulation downregulates SOX2 and induces EMT with increased EGFR–TKI tolerance in lung cancer cells

Knowing that TGFβ inhibits SOX2-mediated reprogramming of fibroblasts (36), the effect of TGFβ stimulation on SOX2 expression and EGFR–TKI tolerance was tested. qPCR analysis determined that SOX2 was downregulated by TGFβ in HCC827 and H1975 in a time-dependent manner (Fig. 4A and B, left). Moreover, we observed that TGFβ stimulation, while inducing a mesenchymal phenotype, inhibited BCL2L11 expression in HCC827 cells (Supplementary Figs. S17 and S18A and S18B). ChIP-qPCR analysis indicated that less SOX2 bound to the BCL2L11 promoter upon TGFβ stimulation (Supplementary Fig. S18C). The results of the clonogenic analysis revealed that pretreatment with TGFβ enhanced the colony growth of HCC827 and H1975 under EGFR–TKI treatment (Fig. 4A and B, right). Cotreatment of HCC827 with a TGFβ inhibitor blocked TGFβ-mediated downregulation of SOX2 and restored EGFR-TKI sensitivity (Supplementary Fig. S18B–S18D). ChIP-qPCR assays also established that the H3K27ac and H3K4me3 signals at the SOX2 locus were lower in TGFβ-treated HCC827 compared with the control cells (Fig. 4C; Supplementary Fig. S18E). Furthermore, the clonogenic assay revealed that SOX2 overexpression antagonized TGFβ-induced colony formation in HCC827 under gefitinib treatment (Fig. 4D). Correlation analysis showed that SOX2 was negatively associated with TGFBR1 and TGFBR2 expression in NSCLC (Supplementary Fig. S19A and S19B). Kaplan–Meier survival analysis further revealed that tumors harboring the SOX2-low/TGFBR1-high or SOX2-low/TGFBR2-high signature predicted a worse survival rate in patients with NSCLC (Supplementary Fig. S19C and S19D). These findings indicate that SOX2 downregulation by TGFβ stimulation promotes EMT and increases EGFR–TKI tolerance, and SOX2 overexpression can interfere with both events.

Figure 4.

The interplay between SOX2 expression and TGFβ stimulation affects EGFR–TKI tolerance. A, Left, qPCR analysis of SOX2 expression in HCC827 cells treated with TGFβ (1 ng/mL) for the indicated time periods. Right, clonogenic analysis to assess the effect of TGFβ on gefitinib resistance in HCC827 cells. HCC827 cells were pretreated with or without TGFβ (1 ng/mL) for 72 hours, followed by gefitinib treatment (1 μmol/L) for 14 days. B, Left, qPCR analysis of SOX2 expression in H1975 cells treated with TGFβ (1 ng/mL) for the indicated time periods. Right, clonogenic analysis of H1975 cells pretreated with TGFβ (1 ng/mL) for 72 hours, followed by osimertinib treatment (1 μmol/L) for 14 days. C, ChIP-qPCR assays to assess the H3K27ac signal at SOX2 locus in TGFβ-treated HCC827 or control HCC827 cells. D, Clonogenic analysis to assess the effects of SOX2 expression and TGFβ stimulation on gefitinib resistance. HCC827 cells were first transduced with the lentiviral vector encoding SOX2 cDNA (SOX2) or empty control vector. HCC827-SOX2 or control cells were pretreated with TGFβ (1 ng/mL) for 72 hours, followed by gefitinib treatment (1 μmol/L) for 14 days. *, P < 0.05; **, P < 0.01; ***, P < 0.001; N.S., not significant.

Figure 4.

The interplay between SOX2 expression and TGFβ stimulation affects EGFR–TKI tolerance. A, Left, qPCR analysis of SOX2 expression in HCC827 cells treated with TGFβ (1 ng/mL) for the indicated time periods. Right, clonogenic analysis to assess the effect of TGFβ on gefitinib resistance in HCC827 cells. HCC827 cells were pretreated with or without TGFβ (1 ng/mL) for 72 hours, followed by gefitinib treatment (1 μmol/L) for 14 days. B, Left, qPCR analysis of SOX2 expression in H1975 cells treated with TGFβ (1 ng/mL) for the indicated time periods. Right, clonogenic analysis of H1975 cells pretreated with TGFβ (1 ng/mL) for 72 hours, followed by osimertinib treatment (1 μmol/L) for 14 days. C, ChIP-qPCR assays to assess the H3K27ac signal at SOX2 locus in TGFβ-treated HCC827 or control HCC827 cells. D, Clonogenic analysis to assess the effects of SOX2 expression and TGFβ stimulation on gefitinib resistance. HCC827 cells were first transduced with the lentiviral vector encoding SOX2 cDNA (SOX2) or empty control vector. HCC827-SOX2 or control cells were pretreated with TGFβ (1 ng/mL) for 72 hours, followed by gefitinib treatment (1 μmol/L) for 14 days. *, P < 0.05; **, P < 0.01; ***, P < 0.001; N.S., not significant.

Close modal

Decreased barrier function and enhanced invasiveness in SOX2-downregulated/TKI-tolerant cells

Alteration in barrier function (Rb) is involved in cellular differentiation (37). To understand the effect of EGFR–TKI selection on cell plasticity, analysis of ECIS was used to determine cellular electric resistance (impedance) and cell–cell contact-mediated Rb in HCC827 and HCC827GR cells. We found that, after seeding, the levels of both impedance and Rb surged in HCC827, but not in HCC827GR cells (Fig. 5A). These data confirmed that the loss of cell–cell contact adhesion occurred in HCC827GR. Because the loss of cell–cell contact adhesion is the key event for cancer cell dissemination, we measured the migration and invasion capabilities of HCC827 and HCC827GR. Cell-tracking assays proved that HCC827GR cells exhibited better migration ability than HCC827 cells (Fig. 5B). Moreover, transwell migration and invasion assays revealed that HCC827GR and H1975AZDR were more migratory and invasive than their EGFR–TKI-sensitive counterparts, HCC827 and H1975 (Fig. 5C). Our findings indicate that EGFR–TKI treatment selects cells with the property of a decreased barrier, promoting a more invasive cellular behavior.

Figure 5.

Increased invasive ability in SOX2-low EGFR–TKI-tolerant lung cancer cells. A, ECIS analysis to measure the changes of impedance (left) and Rb (right) in HCC827 (gefitinib sensitive) and HCC827GR (gefitinib tolerant) cells. B, Cell-tracking analysis to measure the trajectory (left) and relative migration distance (right) of HCC827 and HCC827GR cells. C, Transwell migration and invasion assays of HCC827 versus HCC827GR (left) and H1975 versus H1975AZDR (right). **, P < 0.01; ***, P < 0.001.

Figure 5.

Increased invasive ability in SOX2-low EGFR–TKI-tolerant lung cancer cells. A, ECIS analysis to measure the changes of impedance (left) and Rb (right) in HCC827 (gefitinib sensitive) and HCC827GR (gefitinib tolerant) cells. B, Cell-tracking analysis to measure the trajectory (left) and relative migration distance (right) of HCC827 and HCC827GR cells. C, Transwell migration and invasion assays of HCC827 versus HCC827GR (left) and H1975 versus H1975AZDR (right). **, P < 0.01; ***, P < 0.001.

Close modal

SOX2 mediates cell proliferation in lung cancer with mutated EGFR

To gain additional insight into the effect of EGFR–TKI selection on cell proliferation, we performed the alamarBlue cell proliferation assay in HCC827GR and HCC827 cells. We found that SOX2-positive HCC827 proliferated faster than SOX2-negative HCC827GR (Fig. 6A). Correspondingly, the long-term clonogenic assay showed that SOX2-positive HCC827 and H1975 grew faster than their SOX2-negative counterparts, HCC827GR and H1975AZDR (Fig. 6A). To study the role of SOX2 in cell proliferation of EGFR-mutant lung cancer cells, a knockdown procedure was performed on SOX2 in HCC827 cells, followed by clonogenic assays and cell-cycle analysis. Clonogenic assays displayed that SOX2 silencing attenuated cell growth (Fig. 6B). Cell-cycle analysis revealed that SOX2 knockdown decreased the S-phase of the cell cycle (Fig. 6C). Furthermore, we found that accompanied by loss of SOX2 expression, HCC827GR cells harbored lower EGFR expression than their EGFR–TKI-sensitive parental cells (Fig. 6D). ChIP-qPCR analysis indicated that the amount of SOX2 bound to the EGFR promoter was decreased in HCC827GR and TGFβ-treated HCC827 compared with the parental HCC827 cells (Fig. 6E; Supplementary Fig. S20A). SOX2 overexpression in HCC827GR promoted cell growth and enhanced EGFR expression (Fig. 6F). In addition, knockdown of SOX2 in HCC827 cells decreased EGFR expression (Supplementary Fig. S20B). These findings support the conclusion that SOX2 regulates cell growth in lung cancer cells with mutated EGFR.

Figure 6.

Effect of SOX2 on cell growth. A, Left, alamarBlue proliferation analysis of HCC827 (SOX2 positive) and HCC827GR (SOX2 negative) cells. Right, clonogenic assay of HCC827 versus HCC827GR and H1975 versus H1975AZDR. B, Left, qPCR analysis of SOX2 expression in HCC827 cells transduced with shSOX2 (shSOX2#1) or scrambled control (SC) shRNA. Right, clonogenic analysis of HCC827 cells transduced with shSOX2 (shSOX2#1) or scrambled control shRNA. C, Cell-cycle analysis of HCC827 cells transduced with lentiviral vectors encoding scramble control shRNA (left) or shRNA against SOX2 (shSOX2#1; right). D, qPCR analysis of SOX2 (left) and EGFR (right) expression in HCC827 and HCC827GR cells. E, ChIP-qPCR analysis to assess the SOX2 signal at the EGFR locus in HCC827 (827) and HCC827GR (GR) cells. F, qPCR analysis of SOX2 (left) and EGFR (middle) expression in HCC827GR-SOX2 and HCC827GR-Ctrl. Right, clonogenic analysis of HCC827GR cells transduced with the lentiviral vector encoding SOX2 cDNA (SOX2) or empty control vector (Ctrl). *, P < 0.05; **, P < 0.01; ***, P < 0.001.

Figure 6.

Effect of SOX2 on cell growth. A, Left, alamarBlue proliferation analysis of HCC827 (SOX2 positive) and HCC827GR (SOX2 negative) cells. Right, clonogenic assay of HCC827 versus HCC827GR and H1975 versus H1975AZDR. B, Left, qPCR analysis of SOX2 expression in HCC827 cells transduced with shSOX2 (shSOX2#1) or scrambled control (SC) shRNA. Right, clonogenic analysis of HCC827 cells transduced with shSOX2 (shSOX2#1) or scrambled control shRNA. C, Cell-cycle analysis of HCC827 cells transduced with lentiviral vectors encoding scramble control shRNA (left) or shRNA against SOX2 (shSOX2#1; right). D, qPCR analysis of SOX2 (left) and EGFR (right) expression in HCC827 and HCC827GR cells. E, ChIP-qPCR analysis to assess the SOX2 signal at the EGFR locus in HCC827 (827) and HCC827GR (GR) cells. F, qPCR analysis of SOX2 (left) and EGFR (middle) expression in HCC827GR-SOX2 and HCC827GR-Ctrl. Right, clonogenic analysis of HCC827GR cells transduced with the lentiviral vector encoding SOX2 cDNA (SOX2) or empty control vector (Ctrl). *, P < 0.05; **, P < 0.01; ***, P < 0.001.

Close modal

SOX2 and vimentin as prognostic markers in lung adenocarcinomas with mutated EGFR

The aforementioned data indicate that SOX2 downregulation induces EMT and promotes cancer cell dissemination. We examined SOX2 and vimentin expression in lung tumors derived from EGFR–TKI-resistant xenograft mouse models. In the HCC827 xenograft model, erlotinib selection enriched the cells harboring high vimentin and low SOX2 expression in relapsed tumors (Fig. 7A). Primary lung cancer cells derived from patients who developed EGFR–TKI-acquired resistance (MGH119-R and MGH126) were compared with cells from an EGFR–TKI treatment–naïve patient (MGH119). It was found that SOX2 was downregulated, while vimentin was upregulated, in cancer cells with EGFR–TKI-acquired resistance (Fig 7B). Correlation analysis showed that SOX2 was negatively associated with vimentin expression in NSCLC (Supplementary Fig. S21A). In addition, vimentin and SOX2 expression was analyzed in primary EGFR-mutant lung tumors through IHC staining. Because EMT contributes to cancer cell dissemination, we examined whether there was a correlation between vimentin expression and metastasis. We found that primary EGFR-mutant lung cancer harboring high vimentin expression tended to develop metastases to lymph nodes and distant organs (Fig. 7C). A Kaplan–Meier survival analysis of these patients was conducted to determine the prognostic significance of SOX2 and vimentin expression. The results of this analysis showed a positive correlation between SOX2 and a good survival rate in patients, whereas SOX2-low/vimentin-high signature was associated with a worse prognosis (Fig. 7D; Supplementary Fig. S21B). Subsequently, patients were stratified into treatment with or without EGFR-TKIs and a Kaplan–Meier survival analysis further revealed that tumors harboring the SOX2-low/vimentin-high signature predicted a worse survival rate in patients treated with EGFR–TKIs (Supplementary Fig. S22A). Moreover, in the xenograft model, not only SOX2 but also BCL2L11 was downregulated in TKI-resistant tumors (Supplementary Fig. S22B). IHC staining of two paired lung tumors before and after TKI treatment confirmed that SOX2 expression was downregulated in TKI-resistant patients (Supplementary Fig. S22C). These data support the critical role of SOX2 in lung tumor progression.

Figure 7.

Correlation analysis of SOX2 and EMT in primary lung tumors. A, Gene expression profiling analysis to access SOX2 and vimentin (VIM) expression levels in erlotinib (ERL)-resistant xenograft tumors versus vehicle-treated control tumors in mice injected with HCC827 cells. The data were analyzed from Bivona database. B, List of EGFR–TKI-resistant status (top), and gene expression analysis of SOX2 (bottom, left) and vimentin (bottom, right) in EGFR-mutant lung cancer cells derived from the EGFR–TKI-naïve tumor (MGH119) and acquired EGFR–TKI-resistant tumors (MGH119-R and MGH126) from GSE64766 database. MGH119 and MGH119-R were obtained from the same patient. C,χ2 analysis to assess the correlation between vimentin expression versus metastasis to lymph nodes and distant organs in EGFR-mutant NSCLC. D, Kaplan–Meier analysis to assess the correlation of SOX2 (left) and vimentin (middle) expression with the overall survival of patients with NSCLC harboring EGFR mutations. The overall survival analysis was stratified by SOX2-high/vimentin-low and SOX2-low/vimentin-high signatures for Kaplan−Meier analysis in patients (right). Different groups were compared using log-rank test.

Figure 7.

Correlation analysis of SOX2 and EMT in primary lung tumors. A, Gene expression profiling analysis to access SOX2 and vimentin (VIM) expression levels in erlotinib (ERL)-resistant xenograft tumors versus vehicle-treated control tumors in mice injected with HCC827 cells. The data were analyzed from Bivona database. B, List of EGFR–TKI-resistant status (top), and gene expression analysis of SOX2 (bottom, left) and vimentin (bottom, right) in EGFR-mutant lung cancer cells derived from the EGFR–TKI-naïve tumor (MGH119) and acquired EGFR–TKI-resistant tumors (MGH119-R and MGH126) from GSE64766 database. MGH119 and MGH119-R were obtained from the same patient. C,χ2 analysis to assess the correlation between vimentin expression versus metastasis to lymph nodes and distant organs in EGFR-mutant NSCLC. D, Kaplan–Meier analysis to assess the correlation of SOX2 (left) and vimentin (middle) expression with the overall survival of patients with NSCLC harboring EGFR mutations. The overall survival analysis was stratified by SOX2-high/vimentin-low and SOX2-low/vimentin-high signatures for Kaplan−Meier analysis in patients (right). Different groups were compared using log-rank test.

Close modal

The role of SOX2-mediated cell plasticity in EGFR–TKI tolerance and cancer dissemination is poorly understood. In this study, it was observed that EGFR–TKI selection enriched the culture for cells harboring low SOX2 expression with decreased H3K27ac and H3K4me3 signals on their promotors, whereas the suppression of SOX2 expression by TGFβ stimulation or epigenetic modifiers promoted TKI tolerance and cancer dissemination. These findings provide important insights into how the cell fate factor, SOX2, is regulated by TGFβ cytokine stimulation and epigenetic modification to affect EGFR–TKI tolerance and the property of dissemination in lung cancer cells with EGFR mutations (Supplementary Fig. S23A and S23B).

The SOX family of proteins regulate fate specification and differentiation of stem/progenitor cells (38, 39). The loss of SOX10 in rare cells is associated with the development of BRAF inhibitor resistance and heterogeneity in melanoma (40, 41). Accumulating data indicate that SOX2 mediates self-renewal in ESCs and adult progenitor cells (1). The downregulation of SOX2 initiates differentiation with a morphologic change and pathway switch (1). We observed that SOX2 promoted proliferation of EGFR-mutant lung cancer cells, and SOX2 silencing initiated EMT with increased EGFR–TKI tolerance. Expression of SOX2 together with OCT4, KLF4, and MYC can reprogram fibroblasts into iPSCs, and this reprogramming event can be inhibited by TGFβ stimulation (6, 7). We observed that the stimulation of TGFβ downregulated SOX2 expression while increasing EGFR–TKI tolerance. Ectopic SOX2 expression can interfere with these processes. These data suggest that a cross-talk between SOX2 and TGFβ signaling not only regulates stem cell pluripotency, but also mediates cancer cell plasticity and EGFR–TKI tolerance in lung cancer.

Slow-growing persister cells tend to develop a drug-tolerant state with better survival ability in both bacteria and cancer cells (42, 43). Hata and colleagues reported that clones with acquired resistance, such as those with EGFR-T790M, are derived from slow-growing persister cells, which show high survival and partially resistant state under EGFR–TKI selection. We observed that HCC827GR and H1975AZDR, although exhibited the characteristic of slow-growing cells, did not harbor EGFR-T790M and -C797S mutations, respectively, but were tolerant to EGFR–TKIs. In this article, we found that these persister cells are regulated by SOX2 expression and that SOX2-deficient status provides the cells with the features of slow growth and EGFR–TKI tolerance. This SOX2-mediated inhibitor-tolerant state may contribute to the generation of acquired resistant clones, such as those with EGFR-T790M mutation or MET amplification.

It has been shown that the loss of BCL2L11 expression is associated with EMT and EGFR–TKI tolerance (31). We found that SOX2 regulates BCL2L11 and vimentin expression during EGFR–TKI tolerance development and ESC differentiation/iPSC reprogramming. Knockdown of BCL2L11 increased EGFR–TKI tolerance, but decreased the growth rate of lung cancer cells, indicating the dual roles of BCL2L11 in regulating survival and proliferation. In addition to affecting BCL2L11 expression, we observed that the knockdown of SOX2, while decreasing cell proliferation, induced the mesenchymal phenotype. In contrast, SOX2 expression inhibited mesenchymal marker expression, enhanced EGFR expression, promoted proliferation, and increased EGFR–TKI sensitivity in lung cancer cells. We found that SOX2 bound to BCL2L11 and EGFR promoters in EGFR–TKI-sensitive cells. Positive SOX2EGFR feedback is essential for self-renewal in neural progenitors, and for proliferation in lung cancer cells (10, 44). It has been observed that SOX2 is highly expressed in murine lung tumors driven by the activating EGFR mutation, and that the knockdown of SOX2 in HCC827 enhances cell death under EGFR–TKI treatment (45). This phenomenon could be due to the fact that SOX2 is essential for the growth of EGFR-mutant cells and the simultaneous inhibition of SOX2 and EGFR by RNAi and TKI treatment, respectively, caused more severe cell death. In this study, we observed that SOX2 silencing endowed cells with the characteristics of slow growth and TKI tolerance. The upregulation of SOX2 by acute EGFR–TKI treatment has been implicated in EGFR–TKI tolerance (46). It was also observed that a 4-hour treatment of sensitive, and not tolerant, cells with EGFR–TKI induced SOX2 expression (Supplementary Fig. S15C). However, EGFR–TKI selection enriched the cells harboring low SOX2 expression accompanied by the EMT feature. These data suggest the presence of cross-talk between EGFR and SOX2 signaling in TKI-sensitive cells and further purport that the loss of SOX2 expression by stimulation of differentiation factors, such as TGFβ, switches off SOX2EGFR signaling, but induces EMT, accompanied by decreased BCL2L11 proapoptotic signaling, thus increasing EGFR–TKI tolerance.

Cancer cell plasticity plays a crucial role in lung tumor progression, a phenomenon that is attributed, in part, to epigenetic regulation (9). H3K27ac and H3K4me3 histone modifications play critical roles in stem cell fate determination, and the breadth of H3K27ac and H3K4me3 has been linked to cell identity and transcriptional consistency (47, 48). We observed that H3K27ac and H3K4me3 signals at the SOX2 locus in ESCs and lung cancer cells were diminished during differentiation and EGFR–TKI selection, respectively. We observed that HDAC1 inhibition by inhibitors or shRNAs can induce TGFβ signaling and downregulate SOX2 expression accompanied by enhanced EGFR–TKI tolerance in lung cancer cells. In addition, we observed that TGFβ stimulation decreased H3K27ac and H3K4me3 signals at the SOX2 locus and induced the mesenchymal feature and EGFR–TKI tolerance. Treatment-naïve lung cancer cells exhibiting the mesenchymal feature displayed low H3K27ac and H3K4me3 signals at the SOX2 locus and contained intrinsic EGFR–TKI tolerance. These data indicate that cytokine stimulation and epigenetic modification on SOX2 determine the epithelial feature and EGFR–TKI tolerance in EGFR-mutant lung cancer cells. Because blocking TGFβ signaling with TGFβ inhibitors can prevent SOX2 downregulation and reduce the EGFR–TKI-tolerant state, cotreatment with TGFβ inhibitors may be beneficial to EGFR–TKI therapy. However, the dual function of TGFβ and SOX2 in cancer proliferation and invasion and their pleiotropic activities might pose a challenge for the development of TGFβ inhibitors as an antipersister therapy (49). Recently, it has been reported that the TGFβ/EMT-induced drug-tolerant state is dependent on a druggable GPX4 pathway, making antipersister therapy more promising (50).

Together, our data indicate that switching SOX2 on and off generates cancer cell plasticity. This cellular property is under cytokine stimulation and epigenetic control and endows the cells with divergent tumor dissemination and EGFR–TKI tolerance abilities. These results demonstrate that the interplay between SOX2 expression and TGFβ signaling affects EGFR–TKI treatment and cancer dissemination. Moreover, our findings suggest that SOX2 and EMT markers in EGFR-mutant lung tumors serve as prognostic markers of cancer progression.

No potential conflicts of interest were disclosed.

M.-H. Kuo: Data curation, formal analysis, validation, investigation, writing-original draft, writing-review and editing. A.-C. Lee: Data curation, formal analysis, validation, investigation, writing-original draft. S.-H. Hsiao: Conceptualization, resources, validation, investigation. S.-E. Lin: Resources, investigation. Y.-F. Chiu: Validation, investigation. L.-H. Yang: Validation, investigation. C.-C. Yu: Investigation, methodology. S.-H. Chiou: Conceptualization, resources, funding acquisition. H.-N. Huang: Validation, investigation. J.-C. Ko: Conceptualization, resources, supervision, funding acquisition, writing-review and editing. Y.-T. Chou: Conceptualization, data curation, supervision, funding acquisition, writing-original draft, writing-review and editing.

Y.T. Chou received MOST109-2320-B-007-003-MY3 and J.C. Ko received MOST109-2314-B-002-175. This work was supported, in part, by the National Tsing Hua University-National Taiwan University Hospital Hsin-Chu Branch Joint research grant.

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

1.
Fong
H
,
Hohenstein
KA
,
Donovan
PJ
. 
Regulation of self-renewal and pluripotency by Sox2 in human embryonic stem cells
.
Stem Cells
2008
;
26
:
1931
8
.
2.
Que
J
,
Luo
X
,
Schwartz
RJ
,
Hogan
BL
. 
Multiple roles for Sox2 in the developing and adult mouse trachea
.
Development
2009
;
136
:
1899
907
.
3.
Tompkins
DH
,
Besnard
V
,
Lange
AW
,
Wert
SE
,
Keiser
AR
,
Smith
AN
, et al
Sox2 is required for maintenance and differentiation of bronchiolar Clara, ciliated, and goblet cells
.
PLoS One
2009
;
4
:
e8248
.
4.
Takahashi
K
,
Yamanaka
S.
. 
Induction of pluripotent stem cells from mouse embryonic and adult fibroblast cultures by defined factors
.
Cell
2006
;
126
:
663
76
.
5.
Takahashi
K
,
Tanabe
K
,
Ohnuki
M
,
Narita
M
,
Ichisaka
T
,
Tomoda
K
, et al
Induction of pluripotent stem cells from adult human fibroblasts by defined factors
.
Cell
2007
;
131
:
861
72
.
6.
Ichida
JK
,
Blanchard
J
,
Lam
K
,
Son
EY
,
Chung
JE
,
Egli
D
, et al
A small-molecule inhibitor of TGF-beta signaling replaces SOX2 in reprogramming by inducing NANOG
.
Cell Stem Cell
2009
;
5
:
491
503
.
7.
Maherali
N
,
Hochedlinger
K
. 
TGFβ signal inhibition cooperates in the induction of iPSCs and replaces Sox2 and cMyc
.
Curr Biol
2009
;
19
:
1718
23
.
8.
Shipitsin
M
,
Campbell
LL
,
Argani
P
,
Weremowicz
S
,
Bloushtain-Qimron
N
,
Yao
J
, et al
Molecular definition of breast tumor heterogeneity
.
Cancer Cell
2007
;
11
:
259
73
.
9.
Lin
SC
,
Chou
YT
,
Jiang
SS
,
Chang
JL
,
Chung
CH
,
Kao
YR
, et al
Epigenetic switch between SOX2 and SOX9 regulates cancer cell plasticity
.
Cancer Res
2016
;
76
:
7036
48
.
10.
Chou
YT
,
Lee
CC
,
Hsiao
SH
,
Lin
SE
,
Lin
SC
,
Chung
CH
, et al
The emerging role of SOX2 in cell proliferation and survival and its crosstalk with oncogenic signaling in lung cancer
.
Stem Cells
2013
;
31
:
2607
19
.
11.
Mok
TS
,
Wu
YL
,
Thongprasert
S
,
Yang
CH
,
Chu
DT
,
Saijo
N
, et al
Gefitinib or carboplatin-paclitaxel in pulmonary adenocarcinoma
.
N Engl J Med
2009
;
361
:
947
57
.
12.
Maemondo
M
,
Inoue
A
,
Kobayashi
K
,
Sugawara
S
,
Oizumi
S
,
Isobe
H
, et al
Gefitinib or chemotherapy for non-small-cell lung cancer with mutated EGFR
.
N Engl J Med
2010
;
362
:
2380
8
.
13.
Zhou
C
,
Wu
YL
,
Chen
G
,
Feng
J
,
Liu
XQ
,
Wang
C
, et al
Erlotinib versus chemotherapy as first-line treatment for patients with advanced EGFR mutation-positive non-small-cell lung cancer (OPTIMAL, CTONG-0802): a multicentre, open-label, randomised, phase 3 study
.
Lancet Oncol
2011
;
12
:
735
42
.
14.
Sequist
LV
,
Yang
JC
,
Yamamoto
N
,
O'Byrne
K
,
Hirsh
V
,
Mok
T
, et al
Phase III study of afatinib or cisplatin plus pemetrexed in patients with metastatic lung adenocarcinoma with EGFR mutations
.
J Clin Oncol
2013
;
31
:
3327
34
.
15.
Wu
YL
,
Zhou
C
,
Hu
CP
,
Feng
J
,
Lu
S
,
Huang
Y
, et al
Afatinib versus cisplatin plus gemcitabine for first-line treatment of Asian patients with advanced non-small-cell lung cancer harbouring EGFR mutations (LUX-Lung 6): an open-label, randomised phase 3 trial
.
Lancet Oncol
2014
;
15
:
213
22
.
16.
Rosell
R
,
Carcereny
E
,
Gervais
R
,
Vergnenegre
A
,
Massuti
B
,
Felip
E
, et al
Erlotinib versus standard chemotherapy as first-line treatment for European patients with advanced EGFR mutation-positive non-small-cell lung cancer (EURTAC): a multicentre, open-label, randomised phase 3 trial
.
Lancet Oncol
2012
;
13
:
239
46
.
17.
Arcila
ME
,
Oxnard
GR
,
Nafa
K
,
Riely
GJ
,
Solomon
SB
,
Zakowski
MF
, et al
Rebiopsy of lung cancer patients with acquired resistance to EGFR inhibitors and enhanced detection of the T790M mutation using a locked nucleic acid-based assay
.
Clin Cancer Res
2011
;
17
:
1169
80
.
18.
Kobayashi
S
,
Boggon
TJ
,
Dayaram
T
,
Janne
PA
,
Kocher
O
,
Meyerson
M
, et al
EGFR mutation and resistance of non-small-cell lung cancer to gefitinib
.
N Engl J Med
2005
;
352
:
786
92
.
19.
Sequist
LV
,
Waltman
BA
,
Dias-Santagata
D
,
Digumarthy
S
,
Turke
AB
,
Fidias
P
, et al
Genotypic and histological evolution of lung cancers acquiring resistance to EGFR inhibitors
.
Sci Transl Med
2011
;
3
:
75ra26
.
20.
Thress
KS
,
Paweletz
CP
,
Felip
E
,
Cho
BC
,
Stetson
D
,
Dougherty
B
, et al
Acquired EGFR C797S mutation mediates resistance to AZD9291 in non-small cell lung cancer harboring EGFR T790M
.
Nat Med
2015
;
21
:
560
2
.
21.
Piotrowska
Z
,
Niederst
MJ
,
Karlovich
CA
,
Wakelee
HA
,
Neal
JW
,
Mino-Kenudson
M
, et al
Heterogeneity underlies the emergence of EGFRT790 wild-type clones following treatment of T790M-positive cancers with a third-generation EGFR inhibitor
.
Cancer Discov
2015
;
5
:
713
22
.
22.
Antonicelli
A
,
Cafarotti
S
,
Indini
A
,
Galli
A
,
Russo
A
,
Cesario
A
, et al
EGFR-targeted therapy for non-small cell lung cancer: focus on EGFR oncogenic mutation
.
Int J Med Sci
2013
;
10
:
320
30
.
23.
Uramoto
H
,
Yano
S
,
Tanaka
F
. 
T790M is associated with a favorable prognosis in Japanese patients treated with an EGFR-TKI
.
Lung Cancer
2012
;
76
:
129
30
.
24.
Kuiper
JL
,
Heideman
DA
,
Thunnissen
E
,
Paul
MA
,
van Wijk
AW
,
Postmus
PE
, et al
Incidence of T790M mutation in (sequential) rebiopsies in EGFR-mutated NSCLC-patients
.
Lung Cancer
2014
;
85
:
19
24
.
25.
Li
W
,
Ren
S
,
Li
J
,
Li
A
,
Fan
L
,
Li
X
, et al
T790M mutation is associated with better efficacy of treatment beyond progression with EGFR-TKI in advanced NSCLC patients
.
Lung Cancer
2014
;
84
:
295
300
.
26.
Thomson
S
,
Buck
E
,
Petti
F
,
Griffin
G
,
Brown
E
,
Ramnarine
N
, et al
Epithelial to mesenchymal transition is a determinant of sensitivity of non-small-cell lung carcinoma cell lines and xenografts to epidermal growth factor receptor inhibition
.
Cancer Res
2005
;
65
:
9455
62
.
27.
Yauch
RL
,
Januario
T
,
Eberhard
DA
,
Cavet
G
,
Zhu
W
,
Fu
L
, et al
Epithelial versus mesenchymal phenotype determines in vitro sensitivity and predicts clinical activity of erlotinib in lung cancer patients
.
Clin Cancer Res
2005
;
11
:
8686
98
.
28.
Witta
SE
,
Gemmill
RM
,
Hirsch
FR
,
Coldren
CD
,
Hedman
K
,
Ravdel
L
, et al
Restoring E-cadherin expression increases sensitivity to epidermal growth factor receptor inhibitors in lung cancer cell lines
.
Cancer Res
2006
;
66
:
944
50
.
29.
Hwang
W
,
Chiu
YF
,
Kuo
MH
,
Lee
KL
,
Lee
AC
,
Yu
CC
, et al
Expression of neuroendocrine factor VGF in lung cancer cells confers resistance to EGFR kinase inhibitors and triggers epithelial-to-mesenchymal transition
.
Cancer Res
2017
;
77
:
3013
26
.
30.
Hata
AN
,
Niederst
MJ
,
Archibald
HL
,
Gomez-Caraballo
M
,
Siddiqui
FM
,
Mulvey
HE
, et al
Tumor cells can follow distinct evolutionary paths to become resistant to epidermal growth factor receptor inhibition
.
Nat Med
2016
;
22
:
262
9
.
31.
Song
KA
,
Niederst
MJ
,
Lochmann
TL
,
Hata
AN
,
Kitai
H
,
Ham
J
, et al
Epithelial-to-mesenchymal transition antagonizes response to targeted therapies in lung cancer by suppressing BIM
.
Clin Cancer Res
2018
;
24
:
197
208
.
32.
Chou
SJ
,
Tseng
WL
,
Chen
CT
,
Lai
YF
,
Chien
CS
,
Chang
YL
, et al
Impaired ROS scavenging system in human induced pluripotent stem cells generated from patients with MERRF syndrome
.
Sci Rep
2016
;
6
:
23661
.
33.
Zhang
Z
,
Lee
JC
,
Lin
L
,
Olivas
V
,
Au
V
,
LaFramboise
T
, et al
Activation of the AXL kinase causes resistance to EGFR-targeted therapy in lung cancer
.
Nat Genet
2012
;
44
:
852
60
.
34.
Kaichi
S
,
Hasegawa
K
,
Takaya
T
,
Yokoo
N
,
Mima
T
,
Kawamura
T
, et al
Cell line-dependent differentiation of induced pluripotent stem cells into cardiomyocytes in mice
.
Cardiovasc Res
2010
;
88
:
314
23
.
35.
VanderMolen
KM
,
McCulloch
W
,
Pearce
CJ
,
Oberlies
NH
. 
Romidepsin (Istodax, NSC 630176, FR901228, FK228, depsipeptide): a natural product recently approved for cutaneous T-cell lymphoma
.
J Antibiot
2011
;
64
:
525
31
.
36.
Li
R
,
Liang
J
,
Ni
S
,
Zhou
T
,
Qing
X
,
Li
H
, et al
A mesenchymal-to-epithelial transition initiates and is required for the nuclear reprogramming of mouse fibroblasts
.
Cell Stem Cell
2010
;
7
:
51
63
.
37.
Bagnaninchi
PO
,
Drummond
N
. 
Real-time label-free monitoring of adipose-derived stem cell differentiation with electric cell-substrate impedance sensing
.
Proc Natl Acad Sci U S A
2011
;
108
:
6462
7
.
38.
Kamachi
Y
,
Kondoh
H
. 
Sox proteins: regulators of cell fate specification and differentiation
.
Development
2013
;
140
:
4129
44
.
39.
Sarkar
A
,
Hochedlinger
K
. 
The SOX family of transcription factors: versatile regulators of stem and progenitor cell fate
.
Cell Stem Cell
2013
;
12
:
15
30
.
40.
Shaffer
SM
,
Dunagin
MC
,
Torborg
SR
,
Torre
EA
,
Emert
B
,
Krepler
C
, et al
Rare cell variability and drug-induced reprogramming as a mode of cancer drug resistance
.
Nature
2017
;
546
:
431
35
.
41.
Ennen
M
,
Keime
C
,
Gambi
G
,
Kieny
A
,
Coassolo
S
,
Thibault-Carpentier
C
, et al
MITF-high and MITF-low cells and a novel subpopulation expressing genes of both cell states contribute to intra- and intertumoral heterogeneity of primary melanoma
.
Clin Cancer Res
2017
;
23
:
7097
107
.
42.
Vallette
FM
,
Olivier
C
,
Lezot
F
,
Oliver
L
,
Cochonneau
D
,
Lalier
L
, et al
Dormant, quiescent, tolerant and persister cells: four synonyms for the same target in cancer
.
Biochem Pharmacol
2019
;
162
:
169
76
.
43.
Kochanowski
K
,
Morinishi
L
,
Altschuler
S
,
Wu
L
. 
Drug persistence - from antibiotics to cancer therapies
.
Curr Opin Syst Biol
2018
;
10
:
1
8
.
44.
Hu
Q
,
Zhang
L
,
Wen
J
,
Wang
S
,
Li
M
,
Feng
R
, et al
The EGF receptor-sox2-EGF receptor feedback loop positively regulates the self-renewal of neural precursor cells
.
Stem Cells
2010
;
28
:
279
86
.
45.
Dogan
I
,
Kawabata
S
,
Bergbower
E
,
Gills
JJ
,
Ekmekci
A
,
Wilson
W
 III
, et al
SOX2 expression is an early event in a murine model of EGFR mutant lung cancer and promotes proliferation of a subset of EGFR mutant lung adenocarcinoma cell lines
.
Lung Cancer
2014
;
85
:
1
6
.
46.
Rothenberg
SM
,
Concannon
K
,
Cullen
S
,
Boulay
G
,
Turke
AB
,
Faber
AC
, et al
Inhibition of mutant EGFR in lung cancer cells triggers SOX2-FOXO6-dependent survival pathways
.
Elife
2015
;
4
:
06132
.
47.
Creyghton
MP
,
Cheng
AW
,
Welstead
GG
,
Kooistra
T
,
Carey
BW
,
Steine
EJ
, et al
Histone H3K27ac separates active from poised enhancers and predicts developmental state
.
Proc Natl Acad Sci U S A
2010
;
107
:
21931
6
.
48.
Benayoun
BA
,
Pollina
EA
,
Ucar
D
,
Mahmoudi
S
,
Karra
K
,
Wong
ED
, et al
H3K4me3 breadth is linked to cell identity and transcriptional consistency
.
Cell
2014
;
158
:
673
88
.
49.
Akhurst
RJ
. 
Targeting TGF-β signaling for therapeutic gain
.
Cold Spring Harb Perspect Biol
2017
;
9
:
a022301
.
DOI: 10.1101/cshperspect.a022301
.
50.
Viswanathan
VS
,
Ryan
MJ
,
Dhruv
HD
,
Gill
S
,
Eichhoff
OM
,
Seashore-Ludlow
B
, et al
Dependency of a therapy-resistant state of cancer cells on a lipid peroxidase pathway
.
Nature
2017
;
547
:
453
57
.