Background
Sex-determining region Y-box 2 (Sox2) is a transcription factor that plays an important role in maintaining pluripotency in embryonic stem cells [
1] and in the generation of inducible pluripotent stem cells [
2]. In embryonic stem cells, Sox2 binds to the promoters of many genes, thereby transactivating or suppressing their expressions [
3]. In normal adult tissues, the expression of Sox2 is largely restricted to somatic stem cells [
4]. In recent years, Sox2 has been found to be aberrantly expressed in cancers, including those of the lungs, brain, ovaries, bone, colon, skin and breasts [
5‐
12]. In many of these tumor types, Sox2 was found in the cancer stem-like cell population [
8,
13‐
16]. Studies using various experimental models have demonstrated that Sox2 promotes key tumorigenic properties in cancer cells, including proliferation, invasion, migration, colony formation, non-adherent stem cell-associated sphere formations
in vitro and tumorigenicity in vivo [
8,
12‐
17]. Further, Sox2 expression has been found to correlate with a worse clinical outcome in cancer patients [
11,
18‐
20]. In breast cancer (BC), aberrant expression of Sox2 has been found in up to 30% of tumors [
11,
15], and
in vitro studies have provided evidence that Sox2 contributes to cell proliferation and mammosphere formation in BC cell lines [
12,
15].
Similar to Sox2, Y-box binding protein-1 (YB-1) is a transcription factor that has been found in embryonic stem cells, mammary progenitor cells and BC cells [
21‐
23]. Found in 40% of BC tumors [
24], YB-1 is believed to promote the tumorigenesis of BC, since it has been shown to enhance mammosphere formation
in vitro, and transcriptionally up-regulate the expressions of a large cassette of stem cell-associated proteins including CD44, CD49f (α6 integrin), c-Met, EGFR, Her-2, Cyclin D1, MDR-1, and p110α [
22,
25‐
28]. ln other cell types, it has been shown that YB-1 can also suppress gene expression, such as those encoding Fas and granulocyte macrophage colony-stimulating factor [
29,
30]. To mediate gene regulation, YB-1 translocates into the nucleus and interacts with the proximal promoter regions of its target genes [
31,
32]. YB-1 is activated by phosphorylation in its DNA binding domain, and this biochemical modification confers YB-1 the properties of nuclear translocation and DNA interactions; Akt, RSK1/2 and GSK3ß kinases have been shown to phosphorylate/activate YB-1 [
31‐
33].
Since both YB-1 and Sox2 are important embryonic stem cell proteins that appear to exert similar biological effects in BC [
1,
23], we hypothesized that they may have important interactions in BC cells. In this regard, we recently found that total YB-1 and phospho-YB-1
Ser102 (a commonly used surrogate marker of YB-1 activation) are expressed in MCF7 and ZR751, two Sox2-expressing BC cell lines. In this study, we found evidence that YB-1 regulates the expression of Sox2. Importantly, YB-1 exerted different biological effects in the distinct phenotypically distinct cell subsets, separated based on their differential responsiveness to a Sox2 reporter, namely reporter responsive (RR) cells and reporter unresponsive (RU) cells [
16]. The biological implications of our findings will be discussed, particularly in the context that RR cells have been previously shown to be more tumorigenic and stem-like as compared to RU cells [
16].
Discussion
While the functional importance of Sox2 in embryonic stem cells is well characterized, the biological significance of Sox2 in cancer has not been extensively studied. Sox2 expression is well characterized in embryonic stem cells [
1], where Sox2 is regulated by the Wnt, BMP, and JAK-STAT signaling pathways [
1]. Aberrant expression of Sox2 in cancer cells has been well documented in a number of tumor types, but the mechanisms underlying this biochemical aberrancy is largely unexplored. In BC, there is evidence that Sox2 carries biological significance [
11,
12,
15], although how Sox2 overexpression is regulated in this cell type is unknown. We hypothesized that YB-1, another transcription factor important in stem cell biology and the pathogenesis of BC [
22,
25,
37], regulates Sox2 in BC cells.
Our findings led us to conclude that YB-1 regulates the expression of Sox2 in BC, likely at the transcriptional level. This conclusion is supported by our observations that knockdown of YB-1 substantially up-regulated the
SOX2 transcripts and protein expression in MCF7 and ZR751. Furthermore, using the ChIP assay, we also found evidence that YB-1 interacts with the proximal promoter of
SOX2, which contains multiple YB-1 binding consensus sequences. We were surprised with the finding that YB-1 negatively regulates Sox2, since a previous report has described a positive correlation between YB-1 and Sox2 in glioma cells [
37]. It is likely that the regulatory relationship between these two important stem cell transcription factors is complex and the discrepancy between positive or negative regulation is cell type-specific.
Our data also led us to conclude that activation of YB-1, as evidenced by the phosphorylation of YB-1 at serine-102, is required to regulate Sox2 expression. Specifically, down-regulation of YB-1 activation by serine phosphorylation at residue 102 by the treatment of pharmacologic inhibitors effectively increased Sox2 expression. The concept that activated YB-1 can suppress Sox2 expression correlates well with the previous observations, in which YB-1 phosphorylation at the serine-102 residue is a necessary condition for YB-1 to exert transcriptional control over its downstream gene targets, many of which are stem cell genes [
22,
25]. This concept also correlates with the previous findings that YB-1 activation, known to be mediated by kinases such as Akt, RSK1/2, and GSK3ß, confers its ability to translocate to the nucleus, bind to various gene promoters, and regulate their expression in both ER- positive and negative BC cells [
31‐
33]. While we understand that our experiments in this study involves the use of pharmacologic inhibitors that carry some degree of non-specificity, results of these studies are in parallel with those derived from studies using YB-1 siRNAs.
Based on our findings that YB-1 suppresses the expression of Sox2 in BC, and the previous observation that ER-negative BC cell lines generally have a higher level of YB-1 activation (phosphorylation of YB-1 at serine 102) than ER-positive BC cell lines [
22], we speculated that the expression level of Sox2 is higher in ER-positive BC cells. To this end, we reviewed published cDNA gene expression microarray data collected from a comprehensive panel of 50 BC cell lines [
34]. Indeed, we found that ER-positive BC cell lines expressed a significantly higher
SOX2 expression than ER-negative cell lines (Additional file
1: Figure S1). Similar observations were made in our previously published study using western blotting [
16]. Of note, other YB-1 positively regulated genes such as
CD44,
MET, and
EGFR are likewise previously reported to be highly expressed in ER-negative BC compared to ER-positive BC [
38]. Further, in support that YB-1 also suppresses Sox2 expression in ER-negative BC, we demonstrated that YB-1 knockdown increased Sox2 expression in MDA-MD-231, an ER-negative BC cell line (Additional file
1: Figure S3).
While Sox2 protein expression was negatively regulated by YB-1 in both ER-positive BC cell lines in our studies, it appears that YB-1 only regulates Sox2 transcription activity in a small subset of these cells. Specifically, we found that the siRNA knockdown of YB-1 resulted in increased luciferase and GFP expression in the RR cell subset (in monolayer and in mammosphere culture) but not the RU cell subset. Our observation that the marked increase in Sox2 protein level in RU cells after YB-1 knockdown failed to induce detectable Sox2 transcription activity is in keeping with our previous observation, in which enforced expression of Sox2 in MCF7 and ZR751 using an retroviral Sox2 expression vector also failed to induce detectable Sox2 transcription activity [
16]. Thus, it appears that RU and RR cells are inherently biologically different. The mechanisms underlying this phenotypic difference between these two cell subsets are currently under active investigation in our laboratory.
The drastic difference in the relationship between YB-1 and Sox2 in the two cell subsets of BC is expected to result in substantial biochemical differences, which likely underlie the phenotypic differences between RU and RR cells described previously by our group [
16]. In this regard, differential results from our gene expression analyses in the RU and RR cell populations after YB-1 knockdown support this view. Specifically, following YB-1 knockdown, stem cell genes
NANOG and
ITGA6 (CD49f) were unchanged or down-regulated in RU cells but up-regulated in RR cells. Our results strongly suggest that, with inhibition of YB-1, RU and RR cells will undergo dramatically different biochemical changes, with the stem cell-associated genes being unchanged or suppressed in RU cells whereas the expressions of these genes in RR cells are increased or sustained due to the compensatory increase in Sox2 expression and transcription activity.
Correlating with these observations, we observed that the efficiency of mammosphere formation remained relatively high in RR cells after YB-1 siRNA knockdown. Based on our findings, it is tempting to speculate that the higher efficiency of mammosphere formation in RR cells after YB-1 knockdown is due to the increased Sox2 transcription activity and the sustained expression of various Sox2 downstream target genes in these cells. In parallel with these findings, we showed that the mammospheres derived from Unsorted and parental MCF7 cells treated with YB-1 knockdown exhibited a similar gene expression pattern of RR cells treated with YB-1 knockdown, with high expression levels of SOX2, NANOG, CCND1 and ITGA6.
The existence of tumor heterogeneity, as highlighted by RU and RR cell subsets in our models, may provide explanations to tumor resistance to cancer treatments. Based on the concept generated from this current study, treatments that result in YB-1 inhibition in cancer cells may up-regulate Sox2 expression and transcription activity in the RR cell subset. As a result, stem cell-related genes and possibly the stem cell phenotype can be increased or sustained in this small cell subset, leading to their persistent survival during the course of cancer treatment. Directly relevant to our discussion, at the time of writing, we are aware of an on-going NIH/NCI clinical trial examining the efficacy of Akt inhibitor MK2206 in BC patients. MK2206, like LY294002 inhibits Akt phosphorylation/activation, and we hypothesize that the inhibition of phosphorylation of YB-1 at Ser-102 will up-regulate Sox2 expression.
Competing interests
The authors declare no conflicts of interest.
Authors’ contributions
KJ conceived and designed the research plan, performed experiments, analyzed data, and wrote the manuscript. FW, PW, and XY established the MCF7 and ZR751 Unsorted, RU, and RR lines. BSA provided intellectual input and critical reading of the manuscript. RL conceived and designed the research plan and wrote the manuscript. All authors read and approved the final manuscript.