Emerging Role of Long Non-Coding RNA SOX2OT in SOX2 Regulation in Breast Cancer

The transcription factor SOX2 is essential for maintaining pluripotency in a variety of stem cells. It has important functions during embryonic development, is involved in cancer stem cell maintenance, and is often deregulated in cancer. The mechanism of SOX2 regulation has yet to be clarified, but the SOX2 gene lies in an intron of a long multi-exon non-coding RNA called SOX2 overlapping transcript (SOX2OT). Here, we show that the expression of SOX2 and SOX2OT is concordant in breast cancer, differentially expressed in estrogen receptor positive and negative breast cancer samples and that both are up-regulated in suspension culture conditions that favor growth of stem cell phenotypes. Importantly, ectopic expression of SOX2OT led to an almost 20-fold increase in SOX2 expression, together with a reduced proliferation and increased breast cancer cell anchorage-independent growth. We propose that SOX2OT plays a key role in the induction and/or maintenance of SOX2 expression in breast cancer.


Introduction
The SOX2 gene family (SRY-related HMG-box) encodes a group of transcription factors that are each characterized by the presence of a highly conserved high-mobility group (HMG) domain [1,2]. SOX2 genes are also highly conserved [3] and have been extensively studied in embryonic stem cells, especially in early foregut and neural development. They have been found to be expressed in a restricted spatial-temporal pattern and to play a critical role in stem cell biology, organogenesis, and animal development [3]. SOX2 was also shown to participate in reprogramming of adult somatic cells to a pluripotent stem cell state and has been implicated in tumorigenesis in various organs [4,5,6,7]. Differential expression of SOX2 is reported in human cancers [8,9,10,11,12,13,14,15,16,17,18], including breast cancer from cancer patients and breast cancer cell lines [10,13,14,19,20]. SOX2 expression has been observed in 43% of basal cell-like breast carcinomas and has been found to be strongly correlated with CK5/6, EGFR, and vimentin immunoreactivity, and to be inversely associated with estrogen and progesterone receptor status, suggesting that SOX2 plays a role in conferring a less differentiated phenotype in these tumors [21,22]. Other groups have reported SOX2 expression in a variety of early stage postmenopausal breast carcinomas and lymph nodes metastases, suggesting that SOX2 may play an early role in breast carcino-genesis and that high expression may promote metastatic potential [14]. Expression of SOX2 is up-regulated in a breast cancer cell line grown in a three-dimensional collagen scaffold which partially simulates in vivo conditions [23].
Stem cell-like features may be functionally demonstrated in vitro by the ability of cancer stem cells to grow as mammospheres in non-adherent/serum-free stem cell conditions [24]. Over-expression of SOX2 increased mammosphere formation, and the effect was dependent on continuous SOX2 expression; furthermore, SOX2 knockdown prevented mammosphere formation and delayed tumor formation in xenograft tumor initiation models [25].
Regulation of SOX2 expression is poorly understood. An area located between positions 2528 and +238 from the transcription start site is considered as the core proximal promoter region [26]. Additionally, an upstream enhancer centered between 23444 and 23833 of the transcription start site has an active role in controlling expression of SOX2 in the reprogramming of oligodendrocyte precursors [27] ( Figure 1A) and of pluripotent stem cells [28]. It has been shown the enhancer is activated upon sphere formation [25] in breast carcinoma cells, suggesting that reactivation of SOX2 expression upon sphere formation may be controlled at the promoter level, in the same way as it is activated in pluripotent stem cells.
Genome-wide studies indicate that most of the human genome is transcribed, although only 1-2% of transcripts have proteincoding capacity [29,30,31]. Many of these transcripts are classified as long non-coding RNAs (lncRNAs) [29,32]. Most studies have shown that lncRNAs genes are poorly conserved compared to protein-coding genes [33,34]. Conservation analysis of lncRNAs based on 50-nt window size revealed that many lncRNAs retain patches of higher conservation, possibly representing interaction A) The genomic context of SOX2 and its upstream region. The gray boxes show the proposed promoter and enhancer regions [26,27]. B) The genomic context of SOX2 and SOX2OT, derived from the UCSC browser. SOX2 is located in an intron of SOX2OT. The triangles above each gene show the location of primers used in qRT-PCR. The SOX2 region is enlarged, and the direction of gene transcription shown with arrows. The phylogenetic conservation of each region is shown below the gene diagram (Mammal Cons). Vertical rectangular shading shows the conserved exonic regions of SOX2OT. The locations of these genes are adopted from UCSC genome browser March 2006 (NCBI36/hg18). C) Heat map showing the expression of SOX2 and SOX2OT in breast cancer samples analyzed by TCGA. It covers the expression of SOX2OT; chromosome 3: 182810845-182941699. SOX2 region is shown by an arrow. The samples are classified based on the estrogen receptor status. D and E) Expression of SOX2 and SOX2OT respectively in ER+ (n = 595) and ER2 (n = 176) breast cancer samples (data derived from TCGA). Mann-Whitney rank sum test showed that these genes were expressed differently according to estrogen status: p values of ,0.05 and ,0.005 were calculated for D and E respectively. doi:10.1371/journal.pone.0102140.g001 sites with RNA-binding proteins [34]. Additionally, some lncRNAs appear to be highly conserved among eukaryotes. The SOX2 gene lies in an intron of a long multi-exon non-coding RNA gene that is transcribed in the same orientation ( Figure 1B). This transcript in humans is named SOX2OT or ''SOX2 overlapping transcript'' and shares 88% identity with orthologous sequences in mouse [35]. SOX2OT has been postulated to participate in SOX2 regulation [36]. It has been proposed that SOX2OT has a role in processes related to SOX2 transcription, acting as an enhancer [36]. Recently, two novel splice variants of SOX2OT have been identified in esophageal squamous cell carcinoma and concordant expression of SOX2 and SOX2OT was observed [37]. Currently, very little is known regarding SOX2OT expression in breast cancer and its role in tumor initiation or progression. In this report we show that SOX2OT positively regulates SOX2 expression, and the ectopic expression of SOX2OT in the breast cancer cell line MDA-MB-231 reduces the proliferation rate and increases anchorage independent growth.

Chemicals and reagents
MammoCult medium was purchased from Stem Cell Technologies. Anti-SOX2 antibody was purchased from Cell Signaling Technology (catalogue number 2748). Goat anti-Rabbit IgG2a antibody was from Santa Cruz Biotechnology (catalogue number SC-2054).
SOX2OT and control vectors (EX-hLUC-M90) were purchased from GeneCopoeia. The SOX2OT construct was custom made using SOX2OT splice variant with accession number NR_004053.

Cell culture
Breast cancer cell lines as well as normal breast epithelial cell line MCF10A were cultured according to ATCC recommendations or as described previously [38]. Derivatives of MCF-7 cells were cultured as previously described [39]. MammoCult medium were used for suspension culture.

Cell suspension culture
Cells of the MCF-7 parental line and derivative sub-lines as well as MDA-MB-231 were trypsinized from monolayer culture, and suspensions containing 2610 6 cell seeded in non-tissue culture treated T-25 flasks (Nunc catalogue number 169900). Cells were grown in MammoCult medium, at 37uC in 5% CO 2 in air. All experiments were performed at least two times.

Mouse mammary fat pad model
The orthotopic human breast cancer xenograft model was generated as previously described [40]. Basically, 2610 6 MDA-MB-231-luc-D3H2LN cells (from Caliper Life Sciences) in 50 ml of 50% Matrigel (BD Biosciences), 50% serum free alpha-MEM medium were inoculated into the 4 th right mammary fat pad of female NIH III nude mice (bred at the University of Auckland). The primary tumors were collected 8 weeks after inoculation for quantitative measurement of gene expression by quantitative real time PCR (qRT-PCR). Animal studies were performed in accordance with the approval (CR830) from the Animal Ethics Committee of the University of Auckland.

RNA extraction, reverse transcription, quantitative PCR and PCR
Total RNA from cultured cells or tumor tissues was purified using Trizol (Invitrogen) and treated with DNase I (Sigma) according to the manufacturer's instructions. Oligo-dT or random hexamer were used to reverse transcribe 1 mg of RNA with SuperScript III Reverse Transcriptase (Sigma) according to the manufacturer's instructions. cDNA preparation and qRT-PCR analysis were performed as described previously [41]. For nonquantitative expression analysis, cDNA was amplified by PCR for 35 cycles at 60uC as annealing temperature and amplified products were visualized after electrophoresis in a 1-2% agarose gel.
Quantitative PCR was performed with a final 1 in 20 dilution of cDNA, 8 mM primers and Sybr Green MasterMix (Invitrogen). For normalization of transcript expression levels, human glyceraldehyde 3-phosphate dehydrogenase (GAPDH) and hypoxanthine-guanine phosphoribosyltransferase (HPRT) transcripts were used as internal controls. The average expression of internal control was used to calculate the relative gene expression. Table S1 contains the list of primers used.

Western blotting
Total protein was collected from cultured breast cancer cells. Cells were lysed in buffer containing 60 mM Tris-HCl, pH 6.8, 2% SDS, and 20% glycerol, and protein quantitated by BCA assay. Cell lysates containing 25 mg of protein were separated by SDS-PAGE, and transferred to PVDF membranes (Millipore). Membranes were immunoblotted with antibodies against SOX2 (1:500 for anti-SOX2) and b-actin (1:2500) (Sigma), and antibody visualized using SuperSignal West Pico (Thermo Scientific, Waltham, MA for b -actin) or ECL plus (for SOX2) (Thermo Scientific). Bound antibody was visualized using the chemiluminescence detection system (Fujifilm Las-3000).

Ectopic expression of SOX2OT
Constructs overexpressing SOX2OT and control empty plasmid (vector) (Ex-NEG-M90) were purchased from GeneCopoeia. Both plasmids express GFP to enable detection of transfected cells. Breast cancer cells (MDA-MB-231) were transfected with 5 mg of DNA and Lipofectamine Plus (Invitrogen) according to the manufacturer's instructions. Two biological replicates for each construct were made, the transfected cells treated with puromycin and selected on the basis of GFP expression by fluorescenceactivated cell sorting (FACS) as previously described [42]. The sorted cells were maintained in the presence of puromycin.

Cell proliferation assays
The sulforhodamine B colorimetric (SRB) assay, which is based on the measurement of cellular protein content, was used to measure cell density [43]. Briefly, 3000 cells were plated in 96 well tissue culture plates, and were harvested after 1, 3, 5 and 7 days. Cell proliferation was also measured using a thymidine incorporation assay in which 3000 cells per well were seeded in 96 well plates that were tissue culture-treated for monolayer culture and incubated for 3 days. Alternatively, 10,000 cells per well were seeded in 96 well plates (Corning Costar Ultra-Low attachment) for 6 days suspension culture. Briefly, 0.04 mCi of 3 H-thymidine was added to each well and the cultures incubated for 5 h, after which the cells were harvested on glass fiber filters using an automated TomTec harvester. Filters were incubated with Betaplate Scint and thymidine incorporation was measured using a Trilux/Betaplate counter.
All experiments were done using at least triplicate wells, and performed on at least two separate occasions.

Cell cycle analysis
Cells were grown as monolayers in flasks and in the presence or absence of the mitotic poison paclitaxel (200 nM) for 24 h prior to cell cycle analysis. Non-adherent cells were discarded by aspiration and the attached cells were trypsinized and counted. 10 6 cells were washed with cold PBS, then with cold PBS containing 1% FCS (twice), and fixed with ice-cold methanol overnight at 220uC. The cells were centrifuged at 110 g for 5 min, washed in 1 mL cold PBS containing 3% FCS, and resuspended in the same buffer. The cells were treated with RNase (100 mg/mL) and propidium iodide (PI) (100 mg/mL) for 30 min at room temperature. DNA content was determined by PI fluorescence using a BD FACSVantage cytometer and a total of 30,000 acquired events. The results were analyzed using ModFit LT (Verity Software House).

Assay for anchorage-independent proliferation
The 96 well plates were pre-coated with 40 ml of preheated 0.15% agarose solution. Pre-mixed 26 medium and 0.4% agarose to give the final concentration of 0.2% were prepared. A total of 9000 cells were added in 16 medium. When the top layer of cells containing agarose had gelled, 50 ml of cell culture medium was added to each well. Cells were cultured for 2 weeks with a change of medium every three to four days. Cells were visualized using 0.5% crystal violet, and colonies of 5 cells or more were counted under phase contrast microscope (56 magnification). All experiments were done using at least 10 replicates, and performed at least three times.

Statistical analysis
Statistical analysis was performed using SigmaPlot. Data were analyzed using either one-way ANOVA coupled with multiple comparisons versus treatment control applying the Holm-Sidak method correction, where p,0.05 denotes a statistically significant difference. Correlation analysis was performed with Spearman's rank correlation coefficient (R) and statistical significance (P). Values of p,0.05 were considered to be statistically significant.

Results
Expression of SOX2 and SOX2OT in breast cancer samples from The Cancer Genome Atlas (TCGA) data set We analyzed the genome-wide RNA transcript profile from TCGA (breast invasive carcinoma gene expression) by RNAseq data set (TCGA_BRCA_exp_HiSeqV2-2013-12-18) including 1106 samples from breast cancer patients, and found distinct differences in the expression patterns of SOX2 and SOX2OT in breast cancer samples (Heat map, Figure 1C). This data confirms expression of SOX2 and SOX2OT are positively correlated (Spearman Rank Order Correlation coefficient, r = 0.219; p = 2.23610 213 ) showing higher expression of both SOX2 and SOX2OT in estrogen receptor positive (ER+) compared to ER2 tumors. Statistically significant differential expression of SOX2 (p = 0.001) and SOX2OT (p,0.001) was observed in 595 ER+ and 176 ER2 samples from the same data set using Mann-Whitney Rank Sum test ( Figure 1D, 1E).

Expression analysis of SOX2 and SOX2OT in breast cancer cell lines
We next examined the relative expression of SOX2 and SOX2OT in 18 breast cell lines by qRT-PCR (Figure 2A, and Figure S1). SOX2 was expressed in all of the ER+ cell lines, as compared to six out of twelve ER2 cell lines; widely different levels of expression were detected. SOXOT showed lower levels of expression than its protein-coding counterpart, as is the case for many lncRNAs [44], and a strong positive correlation between the expression of the two genes was observed by Spearman rank order correlation (r = 0.748; p = 12610 25 ) ( Figure 2A). As with TCGA data, significantly higher SOX2 expression was observed in ER+ cell lines (p = 0.03). However, no significant difference in SOX2OT expression was observed between ER+ and ER2 cell lines ( Figure 2B).

Tamoxifen-resistant breast cancer cells express similar levels of SOX2
It has been reported that SOX2 expression is associated with the promotion of tamoxifen resistance in breast cancer, and that ectopic expression of SOX2 contributes to tamoxifen resistance in the MCF-7 breast cancer cell line [45]. We therefore examined the expression of SOX2 and SOX2OT in a series of tamoxifen-resistant MCF-7 breast cancer sub-lines that had been cultured for prolonged periods either in the absence of estrogen or in the presence of the antiestrogen tamoxifen [39]. Expression analysis confirmed the slightly higher levels of expression of SOX2 in three tamoxifen resistant sub-lines TamC6, TamR6 and TamR7 ( Figure 2C). Unexpectedly, two other cell lines, TamC3 and TamR3, showed reduced SOX2 expression. Concordant expression of SOX2 and SOX2OT was not observed in these sub-lines.

Suspension culture
Growth in suspension culture is commonly used to examine both normal and neoplastic cells for clonogenic growth potential. Sphere formation has been reported for many established breast cancer cell lines and there is some evidence to suggest that spheres may be enriched for cells with cancer stem cell phenotypes [38,46,47,48]. To examine the influence of culture conditions on the relative expression of SOX2 and SOX2OT, MCF7 and MDA-MB-231 cell lines were grown either as monolayers (M), or in suspension cultures (S). MCF-7 cells showed higher SOX2 expression levels than did MDA-MB-231 cells when cultured as monolayers ( Figure S1). We observed some significant increases in expression of SOX2 or SOXOT in both MCF-7 and MDA-MB-231 cell lines when cultured as spheres compared to adherent monolayers ( Figure 3A and B). SOX2OT expression was upregulated relative to SOX2 in the triple negative MDA-MB-231 cell line but down-regulated in the estrogen receptor positive MCF-7 cell line ( Figure 3A and B). To determine whether upregulation of SOX2 mRNA expression led to higher protein levels in these cells, immunoblot analysis was performed using the first passage of cells propagated in suspension culture, confirming the up-regulation of SOX2 protein in this culture in both MCF-7 and MDA-MB-231 cells ( Figure 3C).
We also compared SOX2 and SOX2OT expression in our MCF-7 tamoxifen-resistant sub-lines TamC3 and TamR3 [39] cultured as monolayers or in suspension. In contrast to the MCF-7 parental line, SOX2 expression was down-regulated in suspension culture while SOX2OT was up-regulated in the third passage of suspension cultured TamR3 ( Figure 3D).

Expression of SOX2 and SOX2OT in xenografts
To examine whether SOX2 and SOX2OT are up-regulated in xenografts, we evaluated their expression using a MDA-MB-231 variant (MDA-MB-231-luc-D3H2LN) [40]. We compared the level of expression of SOX2 and SOX2OT in MDA-MB-231-luc-D3H2LN cells grown in monolayer with those injected to the mouse fat pad and grown for 8 weeks. The expression of SOX2 and SOX2OT ( Figure 3E) was found to be significantly up-regulated in tumor xenografts.

Ectopic expression of SOX2OT up-regulates SOX2 expression
The MDA-MB-231 cell line was selected for study because its basal expression of SOX2OT was very low. We attempted to identify different isoforms of SOX2OT in MDA-MB-231 and MCF7 cells grown in suspension culture and identified one isoform in both cell lines, as shown in Figure S2.
The expression of SOX2 and SOX2OT was measured by qRT-PCR in MDA-MB-231 cells transfected with control vector or SOX2OT overexpression vector ( Figure 4A). The relative expression of SOX2OT was increased by about 8-fold in SOX2OT transfected cells in comparison to control vector, while 20-fold upregulation was observed in SOX2 expression. This result provides evidence for a positive regulatory role of SOX2OT in SOX2 transcription.
SOX2 protein was also induced in parallel with the upregulation of its mRNA in SOX2OT-transfected MDA-MB-231 cells ( Figure 4B). These data confirm that the up-regulation of SOX2OT transcript leads to higher abundance of SOX2 protein in ectopic SOX2OT expressing cells. However, the expression of two other stem cell marker genes, OCT4 and NANOG, was not significantly altered in the SOX2OT-transfected MDA-MB-231 cells ( Figure S3). Expression of NANOG in these cells was very low while OCT4 had a higher level of expression.

SOX2OT expression reduces proliferation
MDA-MB-231 cells ectopically expressing SOX2OT showed significantly slower growth rate than cells transfected with the control vector, as measured by SRB assay on monolayer culture ( Figure 5A). Proliferation of SOX2OTand control vectortransfected MDA-MB-231 cells was also studied using thymidine incorporation assay in either monolayers or suspension cultures in Ultra-Low attachment 96-well plates. The SOX2OT transfected cells showed a significantly higher proliferation rate in suspension culture, suggesting that SOX2OT overexpression can lead to increased anchorage independent cell growth. (Figure 5B black bars).
To further investigate anchorage independent growth, SOX2OT or control vector transfected cells were examined using a soft agar colony formation assay. The number of colonies formed in SOX2OT transfected MDA-MB-231 cells significantly increased, as compared with control cells ( Figure 5B). SOX2OT overexpression not only increased the number of colonies, but also induced a significant increase in colony size with cells expressing higher levels of SOX2OT ( Figure 5B and C).

SOX2OT effects on the cell cycle
To investigate whether ectopic expression of SOX2OT altered cell cycle progression, control and SOX2OT-overexpressing cells were grown as monolayer adherent cultures for 24 h in the presence of paclitaxel to induce mitotic arrest and to prevent cell division. The cell cycle distribution was then determined by flow cytometry. The progression of cells over 24 h from G1-and Sphase to G2/M-phase cells was used as a measure of cell cycling. The progression to G2/M phase for control cells was 77% (8% to 85%) while that for SOX2OT overexpressing cells was 57% (9% to 66%), confirming that cell cycle progression was delayed by SOX2OT overexpression and consistent with the slower growth rate ( Figure S4).

Discussion
We show here that the ectopic expression of a long non-coding RNA, SOX2OT, dramatically up-regulates expression of the developmental gene SOX2. The SOX2 gene encodes an HMG domain-containing transcription factor that is expressed in various phases of embryonic development in a manner linked to cell-type specification and cellular differentiation [1,49]. Like many other developmental regulatory factors, the dysregulated expression of SOX2 genes has been implicated in a number of human cancers [16,17,22,50,51]. Poorly differentiated tumors show higher levels of expression of embryonic stem cell markers including SOX2 [21]. SOX2 is also overexpressed in human breast cancer tissue and cell lines, and its levels are correlated with tumor grade [10]. Abnormal expression of SOX2 affects cell fate, proliferation, cancer progression and the control of key transcriptional processes in stem cells [52] but little is known about its regulation. Our results therefore provide an important clue as to how SOX2  expression might be controlled in cancer tissue. Previously Chen et al. [10], analyzed gene expression in a panel of breast cancer cell lines and primary tumors, and both gain-and loss-of-function experiments indicated that SOX2 contributes to breast cancer cell proliferation and tumorigenic properties in vitro and in vivo. They also showed that at the cellular level, SOX2 promotes cell cycle progression. In contrast, Cox et al. [11], reported significant dosedependent reductions in cell number upon overexpression of SOX2 in MCF-7 and MDA-MB-231 breast cancer cell lines. These results, in conjunction with our own showing reduced cell proliferation following overexpression of SOX2 (Figure 5A), suggest a model according to which SOX2 expression has a biphasic effect on cell proliferation, stimulating it at lower levels and inhibiting it at high levels. This result may also suggest that SOX2 expression affects proliferation differentially according to whether proliferation is on monolayer culture or in suspension. This result also suggests that SOX2OT can act in trans to induce SOX2 expression.
The induction of SOX2 by ectopic expression of SOX2OT establishes a possible signaling pathway but does not prove that this pathway is important in human breast cancer. To address this question, we compared SOX2 and SOX2OT expression in a large collection of breast cancer samples from TCGA. A weak (r = 0.219) but highly significant (p = 2.23610 213 ) correlation between expression of SOX2 and SOX2OT was found (Figure 1), emphasizing the potential importance of this signaling pathway in human breast cancer. A further positive correlation (r = 0.761) was found (Figure 2A) between the expression of the two genes in cell lines (p = 2610 27 ). This cell line data also showed higher expression of SOX2 and SOX2OT in ER+ than in ER2 samples ( Figure 2B). However, our data differ from a previous report that found an inverse association of SOX2 expression with ER and PR [22]. This may suggest more complex associations of SOX2 and ER in different breast cancer subtypes, since Rodriguez-Pinilla et al. (2007) examined tumors with BRCA1 mutations and sporadic basal-like breast carcinomas while the samples from TCGA data are more variable with respect to sub-type. We could not detect any association between SOX2 and SOX2OT expression in basal and luminal cell lines. Our data with breast cancer cell lines are in agreement with TCGA data, and add to broader evidence obtained during recent years that lncRNAs are major regulatory molecules in normal cellular development and disease progression.
It has been proposed that SOX2 expression is induced when cells are propagated in suspension culture [23]. We therefore investigated the relative expression of SOX2 and SOX2OT in suspension cultures. Interestingly, SOX2OT was up-regulated in both MCF7 and MDA-MB-231 cells grown in suspension, and the extent of SOX2OT up-regulation was more pronounced in the triple negative MDA-MB-231 cell line. The cells grown as orthotopic breast xenograft tumors also showed significant up-regulation of both SOX2 and SOX2OT in the primary xenograft site. These data confirm and extend previous results for suspension cultures.
Enhanced SOX2 expression was reported to confer tamoxifen resistance in the MCF-7 cell line and SOX2 expression is also higher in breast cancer tissue taken from patients after endocrine therapy failure [45], suggesting that the development of tamoxifen resistance is associated with activation of the pathway leading to SOX2-expression [45]. We have previously developed a series of tamoxifen-resistant sub-lines of the MCF-7 cell line using conditions that mimic the development of clinical tamoxifen resistance [39,42,53,54,55,56,57], and it was of interest to determine whether these also exhibited increased SOX2 expression. As shown in Figure 2C, three lines (TamC6, TamR6 and TamR7) did indeed show a tendency to increase SOX2 expression. However, two (TamC3 and TamR3) showed lower expression, indicating that SOX2 expression is not always coupled to tamoxifen resistance. All lines except TamC6 showed reduced progesterone receptor and increased estrogen receptor expression [54] suggesting that steroid receptor expression was not closely coupled to SOX2 expression. However, TamC3 and TamR3 exhibit strongly reduced mTOR signaling [39], together with decreased glycolysis rates and increased sensitivity to cytotoxic drugs (unpublished data), suggesting that SOX2 expression is coupled to mTOR pathway utilization. This is consistent with the results of Corominas-Faja et al suggesting that low utilization of the mTOR pathway is associated not only with low SOX2 expression but also with reduced stem cell properties [58], and inhibition of mTOR disrupts pluripotency and represses SOX2 expression [59]. It is possible that decreased mTOR pathway utilization in TamC3 and TamR3 has contributed to the repressed SOX2 and SOX2OT expression observed. Hence, we compared the ability of these lines to form mammospheres. The resistant line TamR3 was found to be less efficient than MCF-7 parental cells in mammosphere formation and they formed only compact spheroids (data not shown). This supports the hypothesis that SOX2 expression, mTOR pathway utilization and stem cell functions have mechanistic relationships.
In conclusion, an earlier observation that the SOX2OT locus has complex architecture and is important during vertebrate develop-ment in a number of systems [36] has been extended in this study to a series of human breast cancer lines. The previous suggestion that SOX2OT plays a role in regulating chromatin structure [36] has been extended, to demonstrate a specific effect of SOX2OT on SOX2 regulation, using both expression arrays from human breast cancers, a series of breast cancer cell lines and a single line exhibiting ectopic expression of SOX2OT. The results add to the growing picture that emphasizes the importance of non-coding RNA in tumor cell properties.