Introduction
Non-melanoma skin cancer (NMSC) is the most common cancer in the U.S., with over a million new cases of the two most common forms, squamous and basal cell carcinoma, anticipated in 2004 [
1]. The more clinically aggressive form, squamous cell carcinoma (SCC) [
2], has been increasing in incidence since the 1960s at annual rates from 4% to as much as 10% in recent years [
3]. About 95% of skin SCC cases are diagnosed at an early stage and are easily controlled. Unlike early stage SCC, advanced SCC is aggressive, often resistant to local therapy, requires repeated surgical resections and courses of radiotherapy, and accounts for approximately 2000 U.S. deaths each year [
1,
4]. Advanced disease- and treatment-related morbidity have a profound impact on patients' quality of life, frequently producing cosmetic deformity, loss of function, and psychosocial problems. Improved control of advanced skin SCC is clearly necessary and will rely on a thorough understanding of the molecular basis for skin SCC progression.
Signal transducers and activators of transcription (Stat) proteins, a family of latent cytoplasmic transcription factors, are expressed in many cell types and, in response to a wide variety of extracellular polypeptides, regulate the transcription of a broad spectrum of genes that are critically involved in cytokine signaling [
5], cell proliferation and development [
6], and tumorigenesis [
7‐
9]. Upon binding of extracellular ligands, cell surface receptors oligomerize and activate associated Janus kinases (JAKs), which in turn phosphorylate Stats on a single critical tyrosine residue located adjacent to an -SH2 (src homology domain 2) domain. The Stats then dimerize via reciprocal -SH2 domain phosphorylation site interactions and translocate to the nucleus where they regulate gene expression by direct DNA binding or by associating with other transcription factors [
10,
11]. The activity of Stats can be abolished by mutation of this critical tyrosine [
12,
13].
Among the seven known members of mammalian Stat family, Stat3 has been most strongly implicated in tumorigenesis [
7‐
9]. Elevated levels of Stat3 activity have been observed in a number of human cancers and cancer cell lines [
9]. In cancers of epithelial origin, Stat3 is constitutively activated in head and neck squamous cell carcinoma (HNSCC) [
14,
15], breast cancer cell lines [
16,
17], ovarian cancer cell lines [
18], and lung cancer cell lines [
19]. In particular, Stat3 plays a critical role in the development of skin cancer [
20]. In an experimental two-stage mouse skin chemical carcinogenesis model it has been shown that Stat3 is constitutively activated in skin tumors [
21], and that activated Stat3 is indispensable for both the initiation and the promotion stages of epithelial carcinogenesis [
22]. The critical role of Stat3 in skin tumor development was further supported by data obtained from a transgenic mouse model in which a constitutively active mutant of Stat3 called Stat3C (7), was expressed in skin under the control of the keratin-5 promoter [
23]. These mice have a skin phenotype closely resembling psoriasis in humans and, when subjected to the two-stage skin chemical carcinogenesis protocol, rapidly developed carcinomas, bypassing the papilloma stage that normally takes place in this model [[
23], Chan et al, submitted].
Apoptosis or programmed cell death, is mediated through two major pathways, the extrinsic and intrinsic [
24,
25]. The extrinsic pathway is primarily triggered by the binding of extra-cellular death ligands (e.g. TNFα, TRAIL and FasL) to their cognate membrane death receptors. The intrinsic pathway is often initiated by cellular stresses such as withdrawal of survival factors, direct DNA damage (e.g. UV exposure, cytotoxic drugs), and is characterized by the disruption of mitochondrial membrane integrity, an event regulated by Bcl-2 protein family members [
26,
27]. There are more than 20 known members of the Bcl-2 family which, based on their functions in regulating apoptosis, can be divided into an anti-apoptotic 'Bcl-2-like' group, (Bcl-2, Bcl-XL, Bcl-w, Bfl-1/A1 and Mcl-1 etc) and a pro-apoptotic group (Bax, Bak, Bok, Bcl-Xs, Bad, Bid, Bik/Nbk, Bim, Hrk, Bmf, Noxa and Puma etc). It has been reported that Stat3 can regulate transcription of several Bcl-2 family proteins, such as Bcl-2 and Bcl-xL [
28‐
31], Bax [
32], and Mcl-1 [
33‐
35].
At early stages of tumor development, tumor cells often have to face and survive harsh physiological micro-environments, such as lack of nutrition and/or blood supply, survival factor insufficiency, and hypoxia, which generally lead to apoptosis in normal cells [
36]. In fact, it has been well accepted that one of the six hallmarks of tumor cells is the reduced or complete loss of dependence on exogenous growth factor stimulation for survival and proliferation [
37]. Stats are the first family of transcription factors found to be directly activated upon growth factor receptor stimulation. Stat3 is thought to confer protection against apoptosis in many transformed or tumor cells. Several studies in which Stat3 activity is either blocked by anti-sense oligonucleotides, small interfering RNA or expression of dominant negative Stat3 isoforms [
14,
32,
38‐
40], or elevated by expression of Stat3C [
41], have shown an inverse correlation between Stat3 activity and induced apoptosis.
In this study we have examined the activity of Stat3 in several human skin-derived cell lines, ranging from non-transformed to highly malignant, and observed a positive correlation between malignancy and constitutive Stat3 phosphorylation. In addition, we have generated human skin SCC cell lines with reduced Stat3 activity by stably expressing a dominant negative acting form of Stat3β, hereafter referred to as S3DN [
42‐
44]. The S3DN cells, unlike the parental SRB12-p9 cells, undergo apoptosis in the conditions of exogenous growth factor deprivation produced by culture in serum free medium (SFM).
Discussion
Elevated Stat3 activity has been observed in numerous spontaneous and experimentally established mammalian cancers, demonstrating a critical role in tumorigenesis [
7‐
9,
17]. In this study we provide direct evidence that Stat3 activity, as indicated by phosphorylation at tyrosine 705, positively correlates with malignancy in human skin-derived cell lines. Suppression of Stat3 activity, through forced expression of the S3DN protein, in human skin SCC cells blocks their growth factor- and/or other serum factor-independence, a major characteristic of malignancy.
Recent studies have provided convincing evidence for a critical role for Stat3 in every stage of mouse skin cancer development, from promoting the survival of initiated cells to conferring late-stage malignant characteristics such as enhanced motility and invasiveness [
21,
22]. In parallel with these studies we sought to develop a human skin SCC model in which Stat3 activity is stably suppressed, in order to assess the contribution of activated Stat3 to the malignant phenotype in human disease. The SRB12-p9 cell line was originally derived from an aggressive skin SCC tumor. These cells were chosen for stable transfection of the S3DN protein because, in addition to having constitutive and IFN-α inducible Stat3 phosphorylation, they are highly tumorigenic upon subcutaneous injection into nude mice and can be readily transfected and selected for stable gene expression [
45].
The S3DN protein is a unique Stat3 blocking reagent. It consists of Stat3β, τhe short alternative splice product of the Stat3 gene, bearing a point mutation at the critical tyrosine 705 phosphorylation site [
44], that has been FLAG-tagged to allow distinguishing it from the endogenous Stat3β. Stat3β was previously shown to act in a dominant negative manner to suppress the transcriptional activity of Stat3α [
42]. However Stat3β can be transcriptionally active under conditions where Stat3α is not, through interaction with the N-terminal segment of c-jun, [
43]. Unlike Stat3β, the S3DN protein lacks detectable DNA binding or transcriptional activity in EMSA and transient transfection reporter assays, respectively [
44], and would therefore not be expected to induce Stat3β-specific effects. The S3DN protein should be able to form non-functional heterodimers with endogenous forms of Stat3α or β however, blocking their ability to enter the nucleus and/or bind DNA. Alternatively, S3DN may interfere with endogenous Stat3 activity at another level, such as the phosphorylation by JAK kinases, where S3DN may occupy the Stat3 docking sites on the cytoplasmic domains of growth factor and cytokine receptors, thereby blocking phosphorylation of endogenous Stat3α. This later possibility is less likely since we do not observe a reduced level of phospho-Stat3α in the S3DN cells compared to SRB12-p9 or Neo cells (data not shown). Also, recent evidence has emerged indicating that unphosphorylated Stat3α can drive expression of several genes, including some well known oncoproteins, through a novel mechanism that is distinct from that of phosphorylated Stat3α [
47]. It therefore cannot be formally ruled out that S3DN, even though it cannot be phosphorylated, could itself have effects not involving interaction with endogenous Stats.
Although the precise mechanism of suppression of Stat3 activity by S3DN is unclear, its expression in the SRB12-p9 cells reduced binding of Stat3α to DNA (Fig.
2B) and was predicted to inhibit the constitutive Stat3 activity, thereby suppressing proliferation and possibly de-repressing apoptotic signals. To our surprise, the initial characterization of the S3DN stable transfectants indicated no obvious effects on proliferation rate or viability compared to the parental SRB12-p9 cells (Table
1). Similarly, forced overexpression of the S3WT protein did not produce an increase in proliferation rate, but rather the opposite occurred, with an approximately 2 hour increase in cell doubling time observed for 2 of the S3WT clones (Table
1). While this latter result is difficult to explain, the very short doubling time for SRB12-p9 cells (15–16 hours) suggests that further increases in proliferation rate may be limited by other intrinsic factors such as nutrient and biomolecule availability. The lack of consistent reduction in cell proliferation rate for the S3DN cells could be explained by an insufficient amount of S3DN protein expression necessary to block endogenous Stat3α activity under these culture conditions. The highest expressing S3DN clone, S3DN5, shows approximately equal signal for the Stat3 α and β bands, indicating roughly equal amounts of both proteins (Fig.
2A, middle panel, lanes labeled S3DN2 and 5). While this is a substantial increase over the wild type amount of Stat3β protein (compare levels of Stat3β between S3DN and S3WT cells in Fig.
2A, middle panel), it appears insufficient to illicit the effects on cell proliferation and/or cell viability that would be consistent with a suppression of endogenous Stat3α activity.
In an attempt to determine whether expression of S3DN could affect any Stat3α regulated cellular processes, we grew the cells in SFM, an experimental stress condition known to suppress cell growth and induce apoptosis in many cell lines. Indeed, it was only after depriving the S3DN cells of FCS that a dramatic effect was observed. Several independently selected S3DN clones underwent cell death induction when grown in SFM, which was not observed with SRB12-p9, Neo or S3WT clones. This effect was characterized by cell rounding and detachment from the plate followed by disintegration into subcellular particles, all characteristics of apoptotic cell death. This effect was quantified using the MTT cell viability assay. A reduction in cell viability for the S3DN cell lines after 2 days in SFM indicated that cell death was induced, in addition to a possible reduction in proliferation rate. In contrast, the SRB12-p9 and Neo cells remained viable and continued to proliferate, with only an approximately 2–3 hour increase in doubling time.
The induction of cell death could entirely account for the reduced viability of S3DN cells grown in SFM. However, because Stat3 is also a key regulator of cell proliferation [
7‐
9], reduced proliferation may also contribute to this effect. Comparison of the cell cycle profiles of SRB12-p9, Neo, S3DN and S3WT cells indicated that, in FCS-containing media, the distribution of cells in the G1/0, S and G2 phases of the cell cycle was similar in all cases. One exception was the higher amount of sub-G1 DNA-containing particles for the S3DN2 cell line than for the other cell lines (12% compared to approximately 5% for the other cell lines, Fig.
4A upper row), suggesting a higher rate of apoptosis even in the presence of 10% FCS. Four days of culture in SFM resulted in an increase to 28.5% sub-G1 particles for the S3DN2 cells, but not the other cell lines. It should be noted that the data in the S3DN2 -FCS panel represents the relatively small population of S3DN2 cells remaining alive after 4 days in SFM (see corresponding panels in Fig.
3B). Surprisingly, these cells still exhibited a similar percentage of cells in the G1/0 phase to those growing in FCS-containing media, indicating that S3DN expression induces cell death, but does not cause an accumulation of cells in G1/0. However there was a reduction in the percent of cells in S phase with a proportional increase in the G2 fraction for the S3DN2 cells in SFM, indicating potential blocks at both the G1 to S phase transition and at mitosis. Thus the contribution of reduced proliferation to the overall effect of the S3DN expression on these cells is minor compared to the apoptotic effect. This conclusion is further supported by our finding that the pro-apoptotic Bcl-2 family member, Bax, is upregulated in the S3DN cells grown in SFM and this effect is accompanied by accumulation of c-PARP.
The results of the present study are consistent with an anti-apoptotic role for Stat3 in human skin SCC and are also in agreement with much of the predicted role for Stat3 derived from recent mouse skin tumorigenesis studies [
21‐
23]. In addition, it has been demonstrated that gene therapy with Stat3β was effective in suppressing tumor growth in an
in-vivo mouse melanoma model [
40]. This effect was associated with induction of the secreted death ligand TRAIL, which could induce apoptosis and cell cycle arrest of adjacent non-transfected cells [
48]. Other investigators, using a complimentary approach to assessing Stat3 function, have demonstrated that expression of the constitutively active Stat3C protein in fibroblasts can protect them from UV-induced apoptosis [
41].
Authors' contributions
WY carried out Western blotting, EMSA, cell viability and cell cycle analyses and contributed to the draft of the manuscript. SC and JR participated in the generation and characterization of the stably transfected cell lines. KS-C contributed to the initial study design and participated in the characterization of the stably transfected cells. JD contributed to the initial study design and to the draft of the manuscript. JC conceived of the study, coordinated the study, participated in the generation of the stably transfected cell lines and contributed to the draft of the manuscript.