PMCCPMCCPMCC

Search tips
Search criteria 

Advanced

 
Logo of nihpaAbout Author manuscriptsSubmit a manuscriptHHS Public Access; Author Manuscript; Accepted for publication in peer reviewed journal;
 
Cell Stem Cell. Author manuscript; available in PMC 2014 March 7.
Published in final edited form as:
PMCID: PMC3594795
NIHMSID: NIHMS435989

Sox2 Cooperates with Inflammation-Mediated Stat3 Activation in the Malignant Transformation of Foregut Basal Progenitor Cells

Summary

Sox2 regulates the self-renewal of multiple types of stem cells. Recent studies suggest it also plays oncogenic roles in the formation of squamous carcinoma in several organs, including the esophagus where Sox2 is predominantly expressed in the basal progenitor cells of the stratified epithelium. Here, we use mouse genetic models to reveal a novel mechanism by which Sox2 cooperates with microenvironmental signals to malignantly transform epithelial progenitor cells. Conditional overexpression of Sox2 in basal cells expands the progenitor population in both the esophagus and forestomach. Significantly, carcinoma only develops in the forestomach where pathological progression correlates with inflammation and nuclear localization of Stat3 in progenitor cells. Importantly, co-overexpression of Sox2 and activated Stat3 (Stat3C) also transforms esophageal basal cells but not the differentiated suprabasal cells. These findings indicate basal stem/progenitor cells are the cells-of-origin of squamous carcinoma and that cooperation between Sox2 and microenvironment-activated Stat3 is required for Sox2-driven tumorigenesis.

Introduction

Stem/progenitor cells are important for maintaining tissue homeostasis, and their aberrant regulation contributes to tumor initiation and cancer progression (Barker et al., 2009; Schepers et al., 2012). This suggests that signaling molecules and transcription factors such as Sox2 that are important for stem cell maintenance need to be strictly regulated. Recently, genomic studies have shown that abnormal levels of Sox2 correlate with squamous cell carcinoma (SCC) in the lung and esophagus (Bass et al., 2009; Gen et al., 2010). However, the mechanisms underlying this association remain largely unexplored.

Sox2 plays a critical role in maintaining embryonic stem cells as well as adult stem cells in multiple tissues (Arnold et al., 2011; Avilion et al., 2003; Masui et al., 2007; Que et al., 2009; Sarkar and Hochedlinger, 2013). Sox2 is also required for the self-renewal of cancer stem cells (also known as tumor initiating cells) in several malignancies, including glioblastoma and breast cancer (Gangemi et al., 2009; Leis et al., 2012). Moreover, recently Sox2 has been identified as a direct target of Myeloid Elf-1 like factor (MEF, also known as ELF4) in glioblastoma cancer stem cells, and Sox2 overexpression could rescue the decrease in neurosphere formation seen in cells lacking MEF (Bazzoli et al., 2012). We previously demonstrated that Sox2 regulates the proliferation and differentiation of epithelial progenitor cells in the developing mouse esophagus and forestomach, which are both lined by a similar stratified keratinized epithelium (Que et al., 2007). In the adult, Sox2 is predominantly expressed in all of the basal progenitor cells in these tissues [this study and (Arnold et al., 2011)]. Intriguingly, recent clinical studies have revealed that SOX2 gene amplification and protein overexpression frequently occur in SCC of human foregut-derived tissues including the lung and esophagus (Bass et al., 2009; Gen et al., 2010). Conditional Sox2 overexpression in adult mouse lung epithelium leads to tumor formation in one study (Lu et al., 2010). In another study Sox2 overexpression in the same cell population results in hyperplasia but not tumor formation, and the reason for this discrepancy remains undetermined (Tompkins et al., 2011). In other important studies using human immortalized airway epithelial cells, SOX2 overexpression alone is insufficient to drive transformation and this outcome requires co-overexpression of additional genes such as FoxE1 or FgfR2 IIIb (Bass et al., 2009). Therefore, synergistic cooperation between multiple genes/pathways appears to be required for SOX2 overexpression to drive tumor initiation. However, how the cooperation is executed in an in vivo setting and whether the oncogenic role of Sox2 is specific for stem/progenitor cells have yet to be determined.

Inflammation is frequently observed in human esophageal SCC biopsies and facilitates tumor formation in the esophagus and forestomach of animal models (Stairs et al., 2011; Taccioli et al., 2011). However, the mechanism by which inflammation promotes tumor initiation in these tissues remains elusive. Tissue specific overexpression of the inflammatory factor IL-1β in the glandular mouse hindstomach induces severe inflammation, with increased levels of IL-6, and promotes adenocarcinoma in this region through the activation of both the Stat3 and NF-κB pathways (Tu et al., 2008). In addition, deletion of the intercellular adhesion molecule p120-catenin disrupts epithelial integrity and leads to SCC in the forestomach. The pathological progression of the SCC is also accompanied by the accumulation of inflammatory cells and increased nuclear localization of phosphorylated Stat3 (p-Stat3) in tumor cells (Stairs et al., 2011), but how this increased Stat3 activation is involved in SCC formation has not been determined.

Here, we use mouse models in combination with in vitro assays to investigate the mechanism by which Sox2 overexpression drives SCC formation. We show that conditional Sox2 overexpression increases proliferation and inhibits differentiation of basal progenitor cells in the stratified epithelium. Nevertheless, Sox2 overexpression alone is insufficient for driving SCC formation. Rather, this outcome is associated with microenvironment-activated Stat3, which cooperates with Sox2 to drive malignant transformation of progenitor cells.

Results

Sox2 and Krt5 Positive Basal Progenitor Cells both Self-renew and Differentiate In Vitro and In Vivo

We demonstrated previously that Sox2 is required for the development of the stratified epithelium of the mouse esophagus and forestomach (Jacobs et al., 2012; Que et al., 2007). In the adult, Sox2 is expressed prominently in the basal layer of these two tissues (Figure 1A, ,2E2E)(Arnold et al., 2011). All of these cells also express the intermediate filaments Keratin 5 (Krt5) and Keratin 14 (Krt14) (not shown), as well as p63 (Figure 1B), a transcription factor important for epithelial stratification (Wang et al., 2011). Ki67 immunostaining indicates that these basal cells are highly proliferative compared to non-dividing suprabasal cells (Figure 1C). When purified by FACS with the basal cell surface marker p75 (Nerve Growth Factor Receptor-NgfR, Figure 1C) and cultured in the presence of 20ng/ml Fgf2 and Egf, Sox2+ve basal cells maintain self-renewal capability and form large colonies with a colony forming efficiency of 4% (Figure 1D). In addition, isolated single basal cells proliferate to form organoids (esophageospheres) when embedded in extracellular matrix in a 3-D culture system (Figure 1E). Consistent with these findings, only FACS-sorted GFP positive epithelial cells from the esophagus of a Sox2GFP “knock-in” mouse line form esophageospheres (Figure S1A). In the absence of serum, basal cells self-renew to form solid spheres in which ~95% of the cells maintain high levels of Sox2 protein (Figure 1F). When dissociated and reseeded under the same culture conditions, single progenitor cells can reform spheres for four passages (data not shown). By contrast, the addition of 5% serum to the sphere culture induces the differentiation of cells in the center towards a Loricrin+ve squamous cell fate, while Sox2+ve cells are confined to the periphery (Figure 1G). This demonstrates that adult Sox2+ve basal cells are progenitors that both self-renew and differentiate in vitro.

Figure 1
Sox2 Positive Basal Progenitor Cells Self-renew and Differentiate in vitro and in vivo
Figure 2
Sox2 Overexpression Promotes Self-renewal and Inhibits the Differentiation of Basal Progenitor Cells

To test whether Sox2+ve basal cells can self-renew and differentiate in vivo, we exploited a KRT5-CreER transgenic mouse line to perform lineage-tracing in combination with the Rosa26–lacZ reporter allele (Jovov et al., 2011; Rock et al., 2009). By using a low dose of Tamoxifen to label individual basal cells in the esophagus and following their behavior for 3 days, we found that single cells can divide to duplicate themselves and also differentiate to form columns of squamous epithelium (Figure 1H). Similar results were observed in the forestomach (Figure 1I). These findings are consistent with previous lineage-tracing results using the Sox2-CreER mouse line (Arnold et al., 2011) and recent in vivo lineage-tracing data (Doupe et al., 2012). In addition, three injections with a higher dose of Tamoxifen gave genetic recombination in ~75% and 56% of basal cells in the esophagus and forestomach, respectively (n=3, Figure 1J), showing that the KRT5-CreER allele is an efficient tool for genetic manipulation of the basal progenitor cell population in these tissues.

Conditional Overexpression of Sox2 Results in Basal Progenitor Cell Hyperplasia in both the Esophagus and SCC in the Forestomach

SOX2 overexpression has been implicated in the formation of esophageal SCC (Bass et al., 2009; Gen et al., 2010). To directly test the oncogenic potential of Sox2 in vivo, we generated KRT5-CreER; Rosa26 CAG-loxp-stop-loxp-Sox2-IRES-Egfp compound mutants (hereafter referred to as KRT5-CreER;Rosa26Sox2) in which Sox2 is overexpressed in basal progenitor cells and their derivatives following Tamoxifen injection (Figure 2A). The conditional Rosa26Sox2 allele contains IRES-eGFP, which serves as an indicator for Sox2 overexpression at the cellular level (Figure 2A,B). Three injections with Tamoxifen (0.25mg/gram body weight) result in Sox2 overexpression in 71% and 63% of basal cells in the esophagus and forestomach, respectively with an insignificant difference between the esophagus and forestomach (p>0.05). The Sox2 protein levels in the Sox2-overexpressing progenitor cells (GFP+ve) isolated from both the esophagus and forestomach are comparable but apparently higher than wildtype basal progenitor cells (Figure S1B and data not shown). While no apparent phenotypic changes were observed in Rosa26Sox2/Sox2 or KRT5-CreER;Rosa26Sox2/+ (heterozygous) compound mutants six months after three Tamoxifen injections (data not shown), Sox2 overexpression in KRT5-CreER;Rosa26Sox2/Sox2 (homozygous) mutants disrupts epithelial structure within six weeks (Figure 2C). Basal cell populations (Sox2 and p63 double positive) in both the esophagus and forestomach are expanded (Figure 2D) and proliferating cells (phosphorylated Histone H3 positive) are present in the upper layers of the epithelium (Figure 2E). The proliferation index for the esophageal epithelium increases 2.3 fold after Sox2 overexpression (Figure S1C), and this correlates with increased expression of the cell cycle protein, Cyclin D1 (Figure S1D). When cultured in the 3D system, these Sox2 overexpressing progenitor cells form larger spheres than wildtype controls (Figure S1E). Moreover, Sox2 overexpression inhibits the differentiation of basal progenitor cells into squamous epithelium in the esophagus and forestomach, leading to the presence of patches of undifferentiated cells (Figure 2F, S1F).

Although epithelial hyperplasia is initially seen in both the esophagus and forestomach after Sox2 overexpression, further pathological progression to SCC was observed only in the forestomach (n=34 Figure 3A, B) (Table S1). Severe skin lesions were also noticed in the KRT5-CreER;Rosa26Sox2/Sox2 mutants and necessitated euthanasia of the animals. When Sox2 overexpression was induced by only two doses of Tamoxifen, the skin lesions were less severe. While no tumors were found in the esophagus over 30 weeks, multiple SCC nodules (n≥4) containing GFP+ve and p63+ve cells developed in the forestomach within 13 weeks (Figure 3A-H) (Table S1). Significantly, the cells invade into the muscularis externa in 14 out of 23 mice that have tumors at week 13 (Figure 3E, F) and 3 out of 3 at week 30 (Table S1), confirming their tumorigenic phenotype. Multiple invasive SCC nodules also develop in the skin at week 30 (Figure S2A).

Figure 3
Sox2 Overexpression Leads to the Development of Invasive Squamous Cancer in the Forestomach

Taken together, these results reveal that Sox2 overexpression in basal progenitor cells promotes proliferation and inhibits differentiation in both the esophagus and forestomach. Significantly, Sox2 overexpression induces invasive SCC formation only in the forestomach, suggesting that either region-specific intrinsic differences in tumor susceptibility or microenvironmental factors are important for the malignant transformation of Sox2 overexpressing progenitor cells.

Inflammation with IL-6/Stat3 Pathway Activation Facilitates Tumor Formation in the Forestomach of the KRT5-CreER;Rosa26Sox2/Sox2 Mice

We next investigated why SCC only develops in the forestomach after Sox2 overexpression and not in the esophagus. When retrospectively examining pathological progression, we observed small numbers of neutrophils and macrophages in the mesenchyme of the forestomach two weeks after Sox2 overexpression. As hyperplasia develops 4-6 weeks after Sox2 induction the inflammatory cells progressively accumulate in both the mesenchyme and epithelium (Figure 4A). This appears to correlate with the loss of differentiated cells and exposure of progenitor cells to the bile acid rich environment in the stomach. Significantly, such inflammation never occurs in the esophagus at any time examined (Figure S2B). We then compared gene expression between the hyperplastic forestomach and the normal tissue of KRT5-CreER controls (n=3 for each) 6 weeks after Sox2 induction by microarray analysis. Transcript levels of the endogenous Rosa26 gene in the mutants decreased 14.2 fold, as expected from recombination (Table S2). This is consistent with the overexpression of Sox2 protein, as validated by immunostaining (Figure 2F, S1F). Of the 84 genes significantly upregulated (p<0.05, fold-change ≥1.5) upon Sox2 overexpression (Table S3), many are associated with tissue damage and inflammation, including S100A8 (227 fold), Ccl20 (21fold), Cxcr3 (17 fold) and IL-1β (9 fold) (Table S4). The increase of IL-1β transcript levels was validated with independently prepared RNAs using quantitative real-time RT-PCR (Figure 4B). Although the microarray data showed no significant change, qPCR detected a 2.8-fold increase in IL-6 transcript levels in the hyperplastic epithelia (Figure 4B) and this upregulation was confirmed by immunohistochemistry (Figure 4D middle panel). These findings corroborate previous observations that overexpression of IL-1β in glandular stomach tissue is accompanied by increased levels of IL-6 (Quante et al., 2012; Tu et al., 2008). Interestingly, when FACS-sorted Sox2 overexpressing forestomach and esophageal cells (GFP+ve) were cultured in the presence of 10ng/ml IL-6, the cells around the periphery of the esophageospheres become protrusive and irregular, correlated with decreased levels of E-Cadherin (Figure S2C and data not shown), suggesting that IL-6 mediated signaling promotes invasiveness of Sox2 overexpressing progenitor cells.

Figure 4
Tumor Initiation in the Forestomach Involves Inflammatory Signaling after Sox2 Overexpression

IL-6 is a potent activator of the Janus kinase (Jak) /Stat3 signaling pathway, which is involved in tumor formation in several tissues, including the forestomach (Stairs et al., 2011). Consistently, we found that p-Stat3 levels are increased in the hyperplastic forestomach epithelium 6 weeks after Tamoxifen injection and the total protein level of Stat3 is also moderately increased upon Sox2 overexpression (Figure S2D). Immunostaining further confirmed that p-Stat3 is enriched in the nuclei of hyperplastic forestomach epithelium (Figure 4C) and that this expression pattern is maintained in invasive SCC tumor cells (Figure 4C). Of note is that we did not observe a similar increase in nuclear p-Stat3 in the esophagi of the same mice in which either forestomach hyperplasia or tumor was found (data not shown).

To test whether inflammation following Sox2 overexpression is important for tumor initiation in the forestomach, we injected the immunosuppressant Dexamethasone following Sox2 induction (Figure 4D). Serial Dexamethasone treatment reduces the inflammatory reaction in the hyperplastic forestomach and also dampens the accumulation of IL-6 and nuclear localization of p-Stat3 in the epithelial layers (Figure 4D, S2E). More importantly, Dexamethasone treatment dramatically reduces tumor incidence (1/8 Vs 6/8, p<0.01) when examined at 13 weeks after Sox2 overexpression (Figure S2F). Similarly, when phosphorylation of Stat3 is inhibited with a specific inhibitor WP1066 (Iwamaru et al., 2007), tumor incidence is also decreased (1/9 Vs 8/11, p<0.01) (Figure 4E, S2G). These findings support a model in which some unique feature of the forestomach microenvironment, for example exposure to acid secreted by the hindstomach as previously shown (Quante et al., 2012), leads to inflammation when combined with the disruption of epithelial integrity caused by Sox2 overexpression. Moreover, this inflammation, mediated through the IL-1β/IL-6/Stat3 pathway, is required for the malignant transformation of basal progenitor cells.

Activated Stat3 and Sox2 Synergistically Transforms Mouse and Human Esophageal Basal Progenitor, but Not Differentiated Suprabasal Cells

Unlike in the mouse, the human esophagus is immediately adjacent to the acid producing stomach and the formation of human esophageal SCC has recently been linked to acid reflux from this organ (Pandeya et al., 2010; Uno et al., 2011). Although our findings demonstrate that microenvironment-activated inflammation and Stat3 are important for the malignant transformation of basal progenitor cells, it is also possible that intrinsic differences between esophagus and forestomach contribute to the contrasting tumor susceptibility after Sox2 overexpression. We therefore asked whether a combination of Sox2 and activated Stat3 is capable of transforming basal progenitor cells of the mouse esophagus. Sox2-overexpressing basal progenitor cells (GFP+ve) from the esophagi of KRT5-CreER;Rosa26Sox2/Sox2 mice were isolated and infected with a constitutively activated Stat3 (Stat3C) lentivirus (Figure 5A,B). These cells were then injected into the flanks of the immunodeficient NOD/SCID gamma (NSG) mice (1×106 cells/injection). No tumor was observed in mice injected with esophageal basal cells overexpressing either Sox2 or Stat3C alone (3 mice for each group). In contrast, progenitors co-overexpressing Sox2 and Stat3C give rise to SCCs (GFP+ve) within a 5-week observation period in all of the three mice (Figure 5C,D). The experiment was repeated once and the same results were obtained. The tumor cells express p63, indicating their basal progenitor cell identity, and proliferating cells (pH3+ve) are present in the tumor mass (Figure 5E). We have also isolated and co-overexpressed Sox2 and Stat3 in the differentiated suprabasal cells (p75 negative), but no tumor formation was observed from injecting these cells into three NSG mice (1×106 cells/injection) after a 5-week observation period.

Figure 5
Co-overexpression of Sox2 and Constitutively Activated Stat3 (Stat3C) Leads to Malignant Transformation of Mouse Esophageal Basal Progenitors and Human Esophageal Progenitor-like EPC2 Cells

We next asked whether the combined overexpression of Sox2 and activated Stat3 can similarly transform human esophageal epithelial cells. We used EPC2, an immortalized human esophageal epithelial cell line. EPC2 cells express moderate levels of Sox2 and p63 (Figure S3A) and can proliferate and differentiate to form stratified epithelium in 3-D organotypic culture (Andl et al., 2003; Okawa et al., 2007). When embedded in extracellular matrix, single EPC2 cells are also able to self-renew to form spheres (Figure S3B), suggesting EPC2 cells maintain some progenitor cell functions. EPC2 cells infected with lentivirus harboring Sox2-IRES-GFP and Stat3C have a higher sphere forming efficiency than control cells that overexpress either Sox2-IRES-GFP or Stat3C alone (9.2% for co-overexpression, 6.4% and 7.8% for Sox2 and Stat3C overexpression alone, respectively; Figure 5F, G). When injected 1×106 cells into NSG mice none of the control groups (three mice/group) induces tumor formation. In contrast, prominent tumor masses (GFP+ve) were observed in three out of three mice injected with EPC2 cells co-overexpressing Sox2 and Stat3C (Figure 5G-I). Histological analysis indicates that they are invasive squamous cell cancer (Figure 5J). As expected, the nuclei of the tumor cells express high levels of Sox2 and moderate levels of p63 (Figure 5K, L).

Together, these findings demonstrate that high levels of Sox2 and activated Stat3 are able to transform basal progenitor cells but not the differentiated suprabasal cells isolated from the mouse esophagus. Moreover, the combination of these two transcription factors can also malignantly transform human esophageal progenitor-like cells, leading to squamous cell carcinoma when implanted into immunodeficient mice.

shRNA-mediated Knockdown of STAT3 and SOX2 Reduces the Growth of Mouse and Human Squamous Cancer Cells

High levels of Sox2 are needed for maintaining breast and brain cancer stem cells (tumor initiating cells) and tumor growth (Gangemi et al., 2009; Leis et al., 2012). To test whether Sox2 is similarly required for the maintenance of SCC, we isolated tumor cells from the forestomach of the KRT5-CreER;Rosa26Sox2/Sox2 mice and performed Sox2 knockdown with a lentivirus-based system. Sox2 knockdown reduces the size of the tumor spheres formed in the 3D culture system and is accompanied by an increased number of apoptotic cells (Figure 6A, S4A). Similarly, Stat3 knockdown also reduces the size of the tumor spheres (Figure 6A, S4A). When injected into NSG mice (1 ×106 cells/injection), the tumors formed by the Sox2 or Stat3 knockdown cells are apparently smaller than those formed by the non-knockdown tumor cells [0.31±0.078g (Sox2 knockdown), 0.24±0.046g (Stat3 knockdown) and 0.66±0.070g (non-knockdown control), n=6 for each group, Figure 6B]. The weight difference of Sox2 knockdown Vs control or Stat3 knockdown Vs control is significant (p<0.05).

Figure 6
shRNA-mediated Knockdown of Sox2 and Stat3 Reduces the Proliferation of Mouse Forestomach and Human Esophageal Squamous Cancer Cells

A previous study has shown that SOX2 knockdown reduces the proliferation of the nontumorigenic esophageal SCC line TE10 (Bass et al., 2009). SOX2 knockdown using the same system inhibits proliferation in the tumorigenic esophageal SCC line KYSE450 and leads to small esophageospheres (Figure 6C, D, S4B). Sox2 knockdown also increases the number of apoptotic cells within the esophageospheres (Figure S4B). Similarly, shRNA-mediated STAT3 knockdown also reduces the proliferation and increases the apoptosis of KYSE450 cells (Figure 6C, D, S4B) and combined knockdown of SOX2 and STAT3 leads to a further decrease in cell proliferation (Figure 6C, D). Moreover, individual knockdown of SOX2 and STAT3 in KYSE450 cells leads to smaller tumors than in mock controls (Figure 6E) and combined knockdown of SOX2 and STAT3 further reduces xenograft weight in NSG mice (Figure 6E). In aggregate, individual or combined knockdown of SOX2 and STAT3 in a tumorigenic esophageal SCC cell line reduces cell proliferation and tumor growth when injected into NSG mice.

Increased Protein Levels of p-STAT3 and SOX2 Correlate with a Poor Prognosis for Human ESCC

Upregulated SOX2 expression is associated with poor differentiation of esophageal SCC in human patients (Gen et al., 2010). In the human esophagus, SOX2 protein is highly enriched in basal progenitor cells that express p75 (Figure 7A,B) and ~8% of basal cells are positive for p-STAT3 (Figure 7C, also see Figure S5). Through immunohistochemical examination of 83 primary human esophageal SCC samples (from 74 male and 9 females, refer to Table S5, S6 for other clinical parameters), we found that 45 (54.2%) and 38 (45.8%) samples are positive for SOX2 and p-STAT3 staining, respectively. Interestingly, 31 (37.3%) samples are positive for both SOX2 and p-STAT3 (Figure 7D-F). High levels of p-STAT3 and SOX2 in esophageal SCC correlate with a poor 5-year survival as compared to SOX2 positive only group (14.3% Vs 42.9%, P<0.01) (Figure 7G). Of note is that patients with high levels of STAT3, including the STAT3 and SOX2 double positive group, have a worse prognosis than STAT3-negative patients (11.58% Vs 37.1%, p<0.05). Taken together, these results suggest that a simultaneous increase in SOX2 and p-STAT3 protein levels serves as a marker of poor prognosis for patients with esophageal SCC.

Figure 7
High Levels of SOX2 and p-STAT3 are Associated with a Poor Prognosis of Human ESCC and Model Figure

Discussion

It has been proposed that mutations accumulated in stem/progenitor cells can have catastrophic consequences for tissue integrity and homeostasis over the long term compared with mutations in cell populations committed to terminal differentiation (Barker et al., 2009; Visvader and Lindeman, 2012). In this study, we provide the first genetic evidence that links basal progenitor cells with the etiology of squamous cell carcinoma in the foregut. We identify a novel mechanism by which the transcription factor Sox2 cooperates with inflammation to transform basal progenitor cells and initiate invasive squamous cell carcinoma (Figure 7I).

Stem/progenitor cells are located in the basal layer of the stratified epithelium in the esophagus (Kalabis et al., 2008; Seery and Watt, 2000). A recent study combining lineage-tracing and mathematic modeling suggests that the stratified esophageal epithelium is maintained by a single type of basal stem/progenitor cell (Doupe et al., 2012).The progenitor marker we use here, p75, has previously been used to isolate mouse tracheal basal progenitor cells and human esophageal progenitor cells (Okumura et al., 2003; Rock et al., 2009). Our immunostaining results confirm that like Sox2 and Krt5, p75 is expressed in all basal progenitors in the mouse and human esophagus. Purified p75+ve mouse basal progenitor cells form large colonies of cells in monolayer culture that continue to express the basal cell markers Sox2 and p63. When grown in matrix, single progenitor cells can proliferate to form spheres and the undifferentiated cells can regenerate spheres when dissociated and re-embedded, a property that has also been observed in progenitor cells isolated from the trachea and neural tissue (Reynolds and Weiss, 1992; Rock et al., 2009). Upon addition of serum, progenitor cells in the center of the sphere differentiate into squamous cells. Our lineage-tracing results using KRT5-CreER are consistent with previous findings using the Sox2-CreER mouse line (Arnold et al., 2011), confirming that basal cells are progenitor cells for the maintenance of the esophageal and forestomach epithelium. We recently established a new Sox2-CreER “knock-in” mouse line to confirm that Sox2+ve basal progenitor cells self-renew and differentiate to maintain the epithelium (see Figure S1G for lineage-tracing results. The details of this mouse line will be described elsewhere). Similar to the findings by Arnold et al. (Arnold et al., 2011), ablation of basal cells disrupts tissue integrity of the esophagus and forestomach in the Sox2-CreER;Rosa26DTA mice (Figure S1H), confirming that basal cells are stem/progenitor cells required for tissue maintenance.

Genomic amplification and protein overexpression of SOX2 has been identified in SCCs of several human tissues including the lung, oral cavity and esophagus (Bass et al., 2009; Freier et al., 2010; Gen et al., 2010). We have provided direct evidence that Sox2 overexpression promotes the proliferation of basal progenitor cells in both the esophagus and forestomach. Notably, when Sox2 is overexpressed in the differentiated suprabasal cells of the esophagus and forestomach using the Krt10-Cre mouse line, the proliferation status in these cells remains unchanged and no tumor formation is observed (unpublished observations J.Q), suggesting that the ability to promote cell proliferation by Sox2 overexpression is specific to progenitor cells. This concept is further supported by our evidence that a combination of Sox2 and Stat3C overexpression can only transform esophageal basal progenitor cells but not the differentiated suprabasal cells (discussed below).

We found that increased levels of Sox2 are unable to transform progenitor cells in the esophagus while tumor formation in the forestomach requires cooperation with inflammatory signaling that can be suppressed by Dexamethasone. In the mouse, the forestomach is in direct continuity with the hindstomach which secretes acid, while the esophagus is protected by a sphincter. The acidic forestomach environment presumably injures the hyperplastic epithelium which contains regions of undifferentiated progenitor cells and reduced levels of Loricrin after Sox2 overexpression. This is followed by inflammation and increased levels of cytokines IL-1β and IL-6. In the human, there is no keratinized forestomach expressing Sox2 and the esophagus is also protected by a sphincter (Figure 7H). However, bile acid may reach this tissue from the bile duct or stomach as a result of gastro-esophageal reflux (Figure 7H), a disorder affecting 30% of normal populations. In either case, exposure to bile acid is associated with pathological changes and even malignancy. Although it is bile acid-mediated injury that is more commonly associated with the pathobiology of adenocarcinoma (Souza, 2010), recent studies suggest acid reflux also increases the risk of esophageal SCC (Pandeya et al., 2010; Uno et al., 2011). Our previous study showed that acid refluxate disrupts epithelial integrity accompanied by Adam10-mediated E-Cadherin cleavage in human biopsies. Conditional deletion of E-cadherin in the mouse forestomach leads to a leaky epithelium and basal cell hyperplasia (Jovov et al., 2011). In addition, bile acid reflux precipitates inflammation and promotes the incidence of SCC after carcinogen treatment in the esophagus and forestomach in animal models which have undergone esophageo-duodenal anastomosis surgery (Chen et al., 2007; Hao et al., 2009). Our current study provides further evidence that inflammatory signaling is required for malignant transformation of basal progenitor cells and that inhibition of inflammation with Dexamethasone reduces tumor incidence.

Stat3 is a key player in mediating inflammation-driven tumorigenesis in multiple gastrointestinal organs (Corcoran et al., 2011; Grivennikov et al., 2009; Quante et al., 2012; Stairs et al., 2011). Activation of Stat3 through inflammatory cytokines in mouse models has been implicated in the development of esophageal cancers, including SCC and adenocarcinoma (Quante et al., 2012; Stairs et al., 2011). However, there is no direct evidence that activated Stat3 drives esophageal tumor initiation. Overexpression of Stat3C with the KRT5 promoter induces hyperproliferation of epidermal epithelium but not tumor formation in KRT5-Stat3C transgenic mice (Chan et al., 2008). Although Stat3C overexpression transforms immortalized mouse fibroblasts (Bromberg et al., 1999), we found that Stat3C overexpression alone is unable to transform mouse esophageal progenitors or immortalized human EPC2 esophageal progenitor-like cell line and requires the cooperation of Sox2 overexpression to generate tumorigenic cells. Sox2 and Stat3 are essential components of the core circuitry that regulates self-renewal of mouse embryonic stem cells. Studies combining Chromatin Immunoprecipitation (ChIP) with genomic sequencing demonstrate that Stat3 and Sox2 along with other pluripotency transcription factors are likely clustered in multiple transcription factor-binding loci (MTL) where they co-regulate common targets (Chen et al., 2008). Our data also suggest that Sox2 and Stat3 cooperate in the transformation of basal progenitor cells, and that disruption of this cooperation with a specific Stat3 inhibitor reduces tumor incidence.

In summary, we provide multiple lines of evidence supporting the idea that Sox2/Krt5/p75 positive basal progenitor cells in the esophagus and forestomach are able to self-renew and differentiate, and that ectopic Sox2 overexpression promotes the expansion of the progenitor pool. Furthermore, Sox2 cooperates with inflammation signaling to transform progenitor cells and leads to the formation of SCC. In clinical samples, increased levels of SOX2 and p-STAT3 correlate with a poor 5-year survival rate. Perturbing SOX2 and STAT3 expression reduces the proliferation and survival of tumorigenic esophageal SCC cells both in vitro and in a xenograft model. Cooperative oncogenic lesions malignantly transform cells by modulating downstream gene expression and signaling circuitry (Hanahan and Weinberg, 2000; McMurray et al., 2008). Identifying these target genes in Sox2 and Stat3 co-overexpressing esophageal progenitor cells will provide further insights into the pathobiology and new therapeutic avenues for more effective treatment of esophageal cancer.

Experimental Procedures

Cell Isolation for the Mouse Esophagus and Forestomach Tumor and FACS Analysis

Muscle layers were stripped off the mouse esophagus and the remaining tissue incubated in Dispase (BD Biosciences, 16 U/ml) in PBS (30 min) at room temperature. Digestion was stopped by addition of DMEM with 5% FBS. Epithelium was peeled off with forceps, cut into small pieces and incubated in 0.1% trypsin, 1.6 mM EDTA 20 min at 37 °C. Digestion was stopped with 5%FBS-containing medium followed by gentle pipetting and passage through a 40-μm cell strainer. Similarly, forestomach tumor tissues were cut into small pieces and digested with Trypsin, but followed by passing through a syringe with 21G needle for 3 times and then by passing through a strainer. For FACS with p75 antibody, cells were diluted to 1 × 106 cells/ml in 2% FBS, 2% BSA in PBS and incubated in 18 μg/ml rabbit anti-mouse p75 antibody (1:100) or IgG isotype control followed by washing and incubation in allophycocyanin-conjugated or Alexa Fluor 488 donkey anti-rabbit antibody (1:500, Jackson ImmunoResearch Laboratories, Inc). Cells were then washed with DMEM + 2% BSA and propidium iodide added to a final concentration of 200 ng/mL before sorting. GFP positive cells were also sorted from the esophagus and forestomach of the Sox2-GFP or KRT5-CreER;Rosa26Sox2/Sox2 mice following the previously described protocol (Rock et al., 2009). Cell sorting was performed on FACS Vantage SE or FACS Aria II and data analyzed with FACS Diva (BD Biosciences). Cells were collected in DMEM with 2% BSA and cultured immediately.

Cell Culture and Sphere Culture

FACS-sorted mouse basal progenitor or forestomach tumor cells were plated onto rat type IV collagen (BD Biosciences)-coated tissue culture dishes in MTEC/Plus containing 20 ng/ml Egf and Fgf2 (Que et al., 2009). Human esophageal SCC cell line KYSE450 was kindly provided by Dr. Yutaka Shimada at Kyoto University and maintained in 10%FBS RPMI medium. The EPC2 cell line was hTERT-immortalized with functionally intact p53 and p16 and maintained as previously described (Okawa et al., 2007).

FACS-sorted mouse esophageal, human EPC2 or KYSE450 cells were resuspended in MTEC/Plus containing 20 ng/ml Egf and Fgf2, mixed 1:1 with growth factor–reduced Matrigel (BD Biosciences), and 200 μl/cm2 pipetted into a 12-well 0.4-μm Transwell insert (Falcon). MTEC/Plus medium (1 ml) was added to the lower chamber and refreshed every other day. For esophageosphere reforming assay, cell balls were collected and incubated in 0.25% Trypsine with EDTA (GIBCO) for 12min at 37°C to dissociate into single cells, and then replated in the matrix. For the formation of spheres using sorted KRT5-CreER;Rosa26Sox2/Sox2 forestomach tumor or KYSE450 cells, drug-selected scramble control, Sox2 or Stat3 knockdown cells were dispersed into single cells and cultured with same matrix and medium. All cultures were maintained at 37°C, 5% CO2. For inducing differentiation, 5% fetal bovine serum (FBS) was added to MTEC medium from day one in culture. To use IL-6 to treat the esophageospheres formed by FACS-sorted Sox2-overexpressing esophageal or forestomach epithelium, cells were initially cultured for 8 days to form spheres, and then the culture was maintained for another 5 days in the presence of 10 ng/ml IL-6. Wholemount fluorescent pictures were taken with Leica DMI3000 inverted microscope. Samples were fixed in 4% paraformaldehyde in PBS, embedded in 3% UltraPure low-melting point agarose, then paraffin, and sectioned.

Statistical Analysis

Data are expressed as mean ± SD or mean ± SEM (for tumor weight). Differences between two samples were analyzed by Student’s t test. P-values of 0.05 or less were considered statistically significant. Differences among three or more groups were analyzed using ANOVA (SAS 9.2 version, SAS Institute Inc., Cary, NC).

Highlights

  • [triangle]Sox2+ve basal stem/progenitor cells self-renew and differentiate in vitro and in vivo
  • [triangle]Conditional Sox2 overexpression promotes the expansion of basal progenitor cells
  • [triangle]Sox2 cooperates with an inflammatory microenvironment to induce squamous cancer
  • [triangle]Sox2 and Stat3 together transform mouse and human esophageal progenitor cells

Supplementary Material

01

Acknowledgments

This work was initiated in Dr. Brigid Hogan’s laboratory at the Department of Cell Biology, Duke University. We are grateful to Dr. Brigid Hogan and Drs. Mark Noble, Hartmut Land, Craig Jordan and Dirk Bohmann at the Department of Biomedical Genetics in the University of Rochester for helpful discussion and critical reading of the manuscript. We also wish to thank Ian Jacobs and other members of the Que laboratory for discussion and technical assistance. This work was supported by the following grants: NIDDK K99/R00 DK082650 (J.Q), NCI U54 CA156735 (X. C), NCI P01-CA098101 (and its Molecular Pathology/Imaging and Molecular Biology Core Facilities, A.R), NCI U01-CA143056 (A.R) and American Cancer Society (A.R).

Footnotes

Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

References

  • Andl CD, Mizushima T, Nakagawa H, Oyama K, Harada H, Chruma K, Herlyn M, Rustgi AK. Epidermal Growth Factor Receptor Mediates Increased Cell Proliferation, Migration, and Aggregation in Esophageal Keratinocytes in Vitro and in Vivo. Journal of Biological Chemistry. 2003;278:1824–1830. [PubMed]
  • Arnold K, Sarkar A, Yram MA, Polo JM, Bronson R, Sengupta S, Seandel M, Geijsen N, Hochedlinger K. Sox2(+) adult stem and progenitor cells are important for tissue regeneration and survival of mice. Cell Stem Cell. 2011;9:317–329. [PMC free article] [PubMed]
  • Avilion AA, Nicolis SK, Pevny LH, Perez L, Vivian N, Lovell-Badge R. Multipotent cell lineages in early mouse development depend on SOX2 function. Genes Dev. 2003;17:126–140. [PubMed]
  • Barker N, Ridgway RA, van Es JH, van de Wetering M, Begthel H, van den Born M, Danenberg E, Clarke AR, Sansom OJ, Clevers H. Crypt stem cells as the cells-of-origin of intestinal cancer. Nature. 2009;457:608–611. [PubMed]
  • Bass AJ, Watanabe H, Mermel CH, Yu S, Perner S, Verhaak RG, Kim SY, Wardwell L, Tamayo P, Gat-Viks I, et al. SOX2 is an amplified lineage-survival oncogene in lung and esophageal squamous cell carcinomas. Nat Genet. 2009;41:1238–1242. [PMC free article] [PubMed]
  • Bazzoli E, Pulvirenti T, Oberstadt MC, Perna F, Wee B, Schultz N, Huse JT, Fomchenko EI, Voza F, Tabar V, et al. MEF Promotes Stemness in the Pathogenesis of Gliomas. Cell Stem Cell. 2012;11:836–844. [PMC free article] [PubMed]
  • Bromberg JF, Wrzeszczynska MH, Devgan G, Zhao Y, Pestell RG, Albanese C, Darnell JE., Jr. Stat3 as an oncogene. Cell. 1999;98:295–303. [PubMed]
  • Chan KS, Sano S, Kataoka K, Abel E, Carbajal S, Beltran L, Clifford J, Peavey M, Shen J, Digiovanni J. Forced expression of a constitutively active form of Stat3 in mouse epidermis enhances malignant progression of skin tumors induced by two-stage carcinogenesis. Oncogene. 2008;27:1087–1094. [PubMed]
  • Chen KH, Mukaisho K, Ling ZQ, Shimomura A, Sugihara H, Hattori T. Association between duodenal contents reflux and squamous cell carcinoma--establishment of an esophageal cancer cell line derived from the metastatic tumor in a rat reflux model. Anticancer Res. 2007;27:175–181. [PubMed]
  • Chen X, Xu H, Yuan P, Fang F, Huss M, Vega VB, Wong E, Orlov YL, Zhang W, Jiang J, et al. Integration of external signaling pathways with the core transcriptional network in embryonic stem cells. Cell. 2008;133:1106–1117. [PubMed]
  • Corcoran RB, Contino G, Deshpande V, Tzatsos A, Conrad C, Benes CH, Levy DE, Settleman J, Engelman JA, Bardeesy N. STAT3 plays a critical role in KRAS-induced pancreatic tumorigenesis. Cancer Res. 2011;71:5020–5029. [PMC free article] [PubMed]
  • Doupe DP, Alcolea MP, Roshan A, Zhang G, Klein AM, Simons BD, Jones PH. A Single Progenitor Population Switches Behavior to Maintain and Repair Esophageal Epithelium. Science. 2012 [PMC free article] [PubMed]
  • Freier K, Knoepfle K, Flechtenmacher C, Pungs S, Devens F, Toedt G, Hofele C, Joos S, Lichter P, Radlwimmer B. Recurrent copy number gain of transcription factor SOX2 and corresponding high protein expression in oral squamous cell carcinoma. Genes, chromosomes & cancer. 2010;49:9–16. [PubMed]
  • Gangemi RM, Griffero F, Marubbi D, Perera M, Capra MC, Malatesta P, Ravetti GL, Zona GL, Daga A, Corte G. SOX2 silencing in glioblastoma tumor-initiating cells causes stop of proliferation and loss of tumorigenicity. Stem Cells. 2009;27:40–48. [PubMed]
  • Gen Y, Yasui K, Zen Y, Zen K, Dohi O, Endo M, Tsuji K, Wakabayashi N, Itoh Y, Naito Y, et al. SOX2 identified as a target gene for the amplification at 3q26 that is frequently detected in esophageal squamous cell carcinoma. Cancer genetics and cytogenetics. 2010;202:82–93. [PubMed]
  • Grivennikov S, Karin E, Terzic J, Mucida D, Yu GY, Vallabhapurapu S, Scheller J, Rose-John S, Cheroutre H, Eckmann L, Karin M. IL-6 and Stat3 are required for survival of intestinal epithelial cells and development of colitis-associated cancer. Cancer Cell. 2009;15:103–113. [PMC free article] [PubMed]
  • Hanahan D, Weinberg RA. The hallmarks of cancer. Cell. 2000;100:57–70. [PubMed]
  • Hao J, Liu B, Yang CS, Chen X. Gastroesophageal reflux leads to esophageal cancer in a surgical model with mice. BMC gastroenterology. 2009;9:59. [PMC free article] [PubMed]
  • Iwamaru A, Szymanski S, Iwado E, Aoki H, Yokoyama T, Fokt I, Hess K, Conrad C, Madden T, Sawaya R, et al. A novel inhibitor of the STAT3 pathway induces apoptosis in malignant glioma cells both in vitro and in vivo. Oncogene. 2007;26:2435–2444. [PubMed]
  • Jacobs IJ, Ku WY, Que J. Genetic and cellular mechanisms regulating anterior foregut and esophageal development. Dev Biol. 2012;369:54–64. [PMC free article] [PubMed]
  • Jovov B, Que J, Tobey NA, Djukic Z, Hogan BL, Orlando RC. Role of E-cadherin in the pathogenesis of gastroesophageal reflux disease. Am J Gastroenterol. 2011;106:1039–1047. [PMC free article] [PubMed]
  • Kalabis J, Oyama K, Okawa T, Nakagawa H, Michaylira CZ, Stairs DB, Figueiredo JL, Mahmood U, Diehl JA, Herlyn M, Rustgi AK. A subpopulation of mouse esophageal basal cells has properties of stem cells with the capacity for self-renewal and lineage specification. J Clin Invest. 2008;118:3860–3869. [PubMed]
  • Leis O, Eguiara A, Lopez-Arribillaga E, Alberdi MJ, Hernandez-Garcia S, Elorriaga K, Pandiella A, Rezola R, Martin AG. Sox2 expression in breast tumours and activation in breast cancer stem cells. Oncogene. 2012;31:1354–1365. [PubMed]
  • Lu Y, Futtner C, Rock JR, Xu X, Whitworth W, Hogan BL, Onaitis MW. Evidence that SOX2 overexpression is oncogenic in the lung. PLoS One. 2010;5:e11022. [PMC free article] [PubMed]
  • Masui S, Nakatake Y, Toyooka Y, Shimosato D, Yagi R, Takahashi K, Okochi H, Okuda A, Matoba R, Sharov AA, et al. Pluripotency governed by Sox2 via regulation of Oct3/4 expression in mouse embryonic stem cells. Nat Cell Biol. 2007;9:625–635. [PubMed]
  • McMurray HR, Sampson ER, Compitello G, Kinsey C, Newman L, Smith B, Chen SR, Klebanov L, Salzman P, Yakovlev A, Land H. Synergistic response to oncogenic mutations defines gene class critical to cancer phenotype. Nature. 2008;453:1112–1116. [PMC free article] [PubMed]
  • Okawa T, Michaylira CZ, Kalabis J, Stairs DB, Nakagawa H, Andl CD, Johnstone CN, Klein-Szanto AJ, El-Deiry WS, Cukierman E, et al. The functional interplay between EGFR overexpression, hTERT activation, and p53 mutation in esophageal epithelial cells with activation of stromal fibroblasts induces tumor development, invasion, and differentiation. Genes Dev. 2007;21:2788–2803. [PubMed]
  • Okumura T, Shimada Y, Imamura M, Yasumoto S. Neurotrophin receptor p75(NTR) characterizes human esophageal keratinocyte stem cells in vitro. Oncogene. 2003;22:4017–4026. [PubMed]
  • Pandeya N, Webb PM, Sadeghi S, Green AC, Whiteman DC. Gastrooesophageal reflux symptoms and the risks of oesophageal cancer: are the effects modified by smoking, NSAIDs or acid suppressants? Gut. 2010;59:31–38. [PubMed]
  • Quante M, Bhagat G, Abrams JA, Marache F, Good P, Lee MD, Lee Y, Friedman R, Asfaha S, Dubeykovskaya Z, et al. Bile acid and inflammation activate gastric cardia stem cells in a mouse model of Barrett-like metaplasia. Cancer Cell. 2012;21:36–51. [PMC free article] [PubMed]
  • Que J, Luo X, Schwartz RJ, Hogan BL. Multiple roles for Sox2 in the developing and adult mouse trachea. Development. 2009;136:1899–1907. [PubMed]
  • Que J, Okubo T, Goldenring JR, Nam KT, Kurotani R, Morrisey EE, Taranova O, Pevny LH, Hogan BL. Multiple dose-dependent roles for Sox2 in the patterning and differentiation of anterior foregut endoderm. Development. 2007;134:2521–2531. [PMC free article] [PubMed]
  • Reynolds BA, Weiss S. Generation of neurons and astrocytes from isolated cells of the adult mammalian central nervous system. Science. 1992;255:1707–1710. [PubMed]
  • Rock JR, Onaitis MW, Rawlins EL, Lu Y, Clark CP, Xue Y, Randell SH, Hogan BL. Basal cells as stem cells of the mouse trachea and human airway epithelium. Proc Natl Acad Sci U S A. 2009;106:12771–12775. [PubMed]
  • Sarkar A, Hochedlinger K. The Sox Family of Transcription Factors: Versatile Regulators of Stem and Progenitor Cells. Cell Stem Cell. 2013;12 [PMC free article] [PubMed]
  • Schepers AG, Snippert HJ, Stange DE, van den Born M, van Es JH, van de Wetering M, Clevers H. Lineage tracing reveals Lgr5+ stem cell activity in mouse intestinal adenomas. Science. 2012;337:730–735. [PubMed]
  • Seery JP, Watt FM. Asymmetric stem-cell divisions define the architecture of human oesophageal epithelium. Curr Biol. 2000;10:1447–1450. [PubMed]
  • Souza RF. The role of acid and bile reflux in oesophagitis and Barrett’s metaplasia. Biochem Soc Trans. 2010;38:348–352. [PMC free article] [PubMed]
  • Stairs DB, Bayne LJ, Rhoades B, Vega ME, Waldron TJ, Kalabis J, Klein-Szanto A, Lee J-S, Katz JP, Diehl JA, et al. Deletion of p120-Catenin Results in a Tumor Microenvironment with Inflammation and Cancer that Establishes It as a Tumor Suppressor Gene. Cancer Cell. 2011;19:470–483. [PMC free article] [PubMed]
  • Taccioli C, Chen H, Jiang Y, Liu XP, Huang K, Smalley KJ, Farber JL, Croce CM, Fong LY. Dietary zinc deficiency fuels esophageal cancer development by inducing a distinct inflammatory signature. Oncogene. 2011 doi: 10.1038/onc.2011.1592. [PMC free article] [PubMed]
  • Tompkins DH, Besnard V, Lange AW, Keiser AR, Wert SE, Bruno MD, Whitsett JA. Sox2 activates cell proliferation and differentiation in the respiratory epithelium. American journal of respiratory cell and molecular biology. 2011;45:101–110. [PMC free article] [PubMed]
  • Tu S, Bhagat G, Cui G, Takaishi S, Kurt-Jones EA, Rickman B, Betz KS, Penz-Oesterreicher M, Bjorkdahl O, Fox JG, Wang TC. Overexpression of Interleukin-1β Induces Gastric Inflammation and Cancer and Mobilizes Myeloid-Derived Suppressor Cells in Mice. Cancer Cell. 2008;14:408–419. [PMC free article] [PubMed]
  • Uno K, Iijima K, Hatta W, Koike T, Abe Y, Asano N, Kusaka G, Shimosegawa T. Direct measurement of gastroesophageal reflux episodes in patients with squamous cell carcinoma by 24-h pH-impedance monitoring. Am J Gastroenterol. 2011;106:1923–1929. [PubMed]
  • Visvader JE, Lindeman GJ. Cancer stem cells: current status and evolving complexities. Cell Stem Cell. 2012;10:717–728. [PubMed]
  • Wang X, Ouyang H, Yamamoto Y, Kumar PA, Wei TS, Dagher R, Vincent M, Lu X, Bellizzi AM, Ho KY, et al. Residual embryonic cells as precursors of a Barrett’s-like metaplasia. Cell. 2011;145:1023–1035. [PMC free article] [PubMed]