Reciprocal negative regulation between S100A7/psoriasin ... - Nature

Oncogene (2008) 27, 3527–3538

& 2008 Nature Publishing Group All rights reserved 0950-9232/08 $30.00 www.nature.com/onc

ORIGINAL ARTICLE

Reciprocal negative regulation between S100A7/psoriasin and b-catenin signaling plays an important role in tumor progression of squamous cell carcinoma of oral cavity G Zhou1, T-X Xie1, M Zhao1, SA Jasser1, MN Younes1, D Sano1, J Lin1, ME Kupferman1, AA Santillan1, V Patel2, JS Gutkind2, AK EI-Naggar3, ED Emberley4, PH Watson4, S-I Matsuzawa5, JC Reed5 and JN Myers1,6 1 Department of Head and Neck Surgery, The University of Texas MD Anderson Cancer Center, Houston, TX, USA; 2Oral and Pharyngeal Cancer Branch, National Institute of Dental and Craniofacial Research, National Institutes of Health, Bethesda, MD, USA; 3Department of Pathology, The University of Texas MD Anderson Cancer Center, Houston, TX, USA; 4Deeley Research Center, BC Cancer Agency, Victoria, British Columbia, Canada; 5Burnham Institute for Medical Research, La Jolla, CA, USA and 6 Department of Cancer Biology, The University of Texas MD Anderson Cancer Center, Houston, TX, USA

Overexpression of S100A7 (psoriasin), a small calciumbinding protein, has been associated with the development of psoriasis and carcinomas in different types of epithelia, but its precise functions are still unknown. Using human tissue specimens, cultured cell lines, and a mouse model, we found that S100A7 is highly expressed in preinvasive, well-differentiated and early staged human squamous cell carcinoma of the oral cavity (SCCOC), but little or no expression was found in poorly differentiated, later-staged invasive tumors. Interestingly, our results showed that S100A7 inhibits both SCCOC cell proliferation in vitro and tumor growth/invasion in vivo. Furthermore, we demonstrated that S100A7 is associated with the b-catenin complex, and inhibits b-catenin signaling by targeting b-catenin degradation via a noncanonical mechanism that is independent of GSK3b-mediated phosphorylation. More importantly, our results also indicated that b-catenin signaling negatively regulates S100A7 expression. Thus, this reciprocal negative regulation between S100A7 and b-catenin signaling implies their important roles in tumor development and progression. Despite its high levels of expression in early stage SCCOC tumorigenesis, S100A7 actually inhibits SCCOC tumor growth/invasion as well as tumor progression. Downregulation of S100A7 in later stages of tumorigenesis increases b-catenin signaling, leading to promotion of tumor growth and tumor progression. Oncogene (2008) 27, 3527–3538; doi:10.1038/sj.onc.1211015; published online 28 January 2008 Keywords: S100A7; psoriasin; b-catenin; oral squamous cell carcinoma; head and neck squamous cell carcinoma

Correspondence: Professor JN Myers, Department of Head and Neck Surgery, University of Texas MD Anderson Cancer Center, 1515 Holcombe Boulevard, Unit 441, Houston, TX 77030-4009, USA. E-mail: [email protected] Received 17 July 2007; revised 12 November 2007; accepted 3 December 2007; published online 28 January 2008

Introduction S100A7 (psoriasin), a low-molecular weight (11.4 kDa) Ca2 þ -binding protein, is a member of the S100 family of proteins and was originally identified through its brisk overexpression in the skin of patients suffering from psoriasis, a hyperproliferative skin disease characterized by abnormal differentiation. Like many other members of the S100 family, S100A7 gene is located within the ‘epidermal differentiation complex’ on human chromosome 1q21 and its protein is believed to be implicated in Ca2 þ -dependent and -independent regulation of a variety of intracellular and extracellular activities (Donato, 2001, 2003; Heizmann et al., 2002). It has also been found that S100A7 is highly overexpressed in wound healing, and many inflammatory and hyperproliferative epidermal diseases, including psoriasis, atopic dermatitis, mycosis fungoides, Darier’s disease and lichen sclerosus et atrophicus as well as skin cancers and breast cancers. However, the exact role of S100A7 in these diseases is still unclear. S100A7 overexpression during skin and breast tumorigenesis suggests that S100A7 might have an ‘oncogenic’ role in tumor development. However, a close examination of the expression profile of S100A7 during tumor progression suggests a more complex picture of functional roles for S100A7 in tumor development. Though its expression in skin is highly induced in keratinocytes under specific pathological circumstances, including hyperplasia associated with inflammatory skin lesions and dysplasia associated with neoplastic progression (Alowami et al., 2003; Emberley et al., 2003a), S100A7 actually first emerged in breast cancer when identified as a cDNA downregulated in a nodal metastasis relative to a primary invasive breast tumor (Moog-Lutz et al., 1995). Furthermore, analysis in breast cancers showed that S100A7 expression is relatively low or undetectable in normal, benign and atypical hyperplastic proliferative ductal lesions, but is often highly induced in the ductal epithelial cells of premalignant, preinvasive ductal carcinoma in situ

S100A7 inhibits SCCOC progression G Zhou et al

3528

(DCIS) (Leygue et al., 1996; Enerback et al., 2002) and then its expression is diminished in invasive carcinomas (Leygue et al., 1996; Emberley et al., 2003b). Such ‘biphasic expression’ suggests that its overexpression is associated with altered differentiation in glandular epithelium of DCIS, that S100A7 may only have a role in early breast tumor progression (Watson et al., 1998), and that the concomitant loss of S100A7 expression in the later stage of invasive carcinomas may be required for tumor progression. In other words, its overexpression may actually ‘suppress’ the tumor progression to later stages of invasive carcinomas. Consistent with the putative ‘tumor suppressive’ role are the observations that the expression of S100A7 is relatively low or undetectable in most highly proliferating keratinocytes and tumor cells in vitro, but it can be induced by agents that stimulate cell differentiation such as retinoic acid (Tavakkol et al., 1994), calcium (Hoffmann et al., 1994), UV light (Di Nuzzo et al., 2000), DNA damaging agents (Kennedy et al., 2005), loss of cell attachment to extracellular matrix, growth factor starvation and cell confluent culture conditions (Enerback et al., 2002; Martinsson et al., 2005). In addition, in normal epithelium, S100A7 has been shown to be expressed in the superficial, differentiated region of the epithelium rather than in the highly proliferative basal region, and its expression correlates with the degree of keratinocyte differentiation (Broome et al., 2003). In skin tumors, it is absent in undifferentiated basalioma and strongly expressed in carcinoma in situ, as well as in keratoacanthoma and differentiated squamous cell carcinoma. These findings support the hypothesis that S100A7 is involved in epithelial differentiation (Martinsson et al., 2005) rather than in cell proliferation. So far very little is known about molecular mechanisms of S100A7’s function. To study the potential biological role of S100A7 in oral squamous epithelial tumor progression, we first evaluated S100A7 expression in both human and mouse specimens of human squamous cell carcinoma of the oral cavity (SCCOC) at different stages of tumor development. We then used both adenoviraland lentiviral-mediated gene overexpression and underexpression strategies in human SCCOC cell lines, which identified a critical role of S100A7 in inhibiting tumor growth and progression of SCCOC. We further established that S100A7 is an important negative modulator of b-catenin signaling through targeting b-catenin for degradation, and that b-catenin in turn can inhibit the expression of S100A7. Our results suggest that although S100A7 is overexpressed in early stages of tumorigenesis, this protein actually plays an inhibitory role in tumor progression of SCCOC, and loss of S100A7 expression is associated with more advanced disease. Results S100A7 is overexpressed in early stages of preinvasive, well-differentiated SCCOC but is lost in later stages of tumor development To investigate the role of S100A7 in the development of human SCCOC, we first evaluated the expression Oncogene

patterns of S100A7 in 27 archival human tongue tissue specimens, including those of histologically normal and dysplastic squamous mucosa and invasive SCCOC. In normal oral epithelia, no expression of S100A7 was observed in dividing basal layer keratinocytes, whereas some low-level, scattered expression of S100A7 was seen in the differentiated suprabasal layers (Figure 1a). S100A7 expression was increased in dysplastic lesions and in well-differentiated SCCOC, but in the later stages of poorly differentiated invasive SCCOC, expression was greatly reduced or absent (Figure 1a). Of the 27 SCCOC specimens we evaluated, 74% (20/27) were well differentiated or moderately differentiated carcinomas and they were all positive for S100A7 expression, whereas none of the seven poorly differentiated invasive carcinomas expressed S100A7 (Figure 1b). In addition, S100A7 expression level was inversely correlated with histological grade, with higher expression seen in the well-differentiated and preinvasive carcinomas, less in the moderately differentiated carcinomas, and none in the poorly differentiated invasive carcinomas (Figures 1a and b). We also examined the expression of S100A7 in orthotopic SCCOC tumor models in nude mice generated from three isogenic human SCCOC cell lines (Tu167, JMAR and DM14) with varying tumor growth and metastatic potentials (Supplementary Materials). As shown in Figure 1c, the Tu167-derived tongue tumors (less malignant) exhibited a more differentiated morphology and a higher level of S100A7 expression, whereas JMAR-derived tongue tumors (moderately malignant) demonstrated reduced S100A7 expression with only scattered staining detected. In the most malignant DM14 tumors, however, S100A7 expression was almost entirely lost. Together, these results demonstrate an inverse relationship between levels of S100A7 expression and tumor progression in both human SCCOC and orthotopic tumors in nude mice. Expression of S100A7 inhibits SCCOC cell proliferation in vitro and tumor growth and tumor progression in vivo While in vivo expression of S100A7 is very high in early staged, well-differentiated SCCOC, its expression in vitro SCCOC cells (for example, Tu167, JMAR and DM14) is very low or undetectable (Supplementary Figure S1B), but can be induced by cell confluence or cell detachment (Figure 6a) as previously reported in other cell lines (Enerback et al., 2002). To elucidate the role of S100A7 in cell growth and tumorigenesis of SCCOC, we generated an adenoviral mediated expression system to overexpress S100A7 (Figure 2a). As shown in Figure 2b and Supplementary Figure S1A, when S100A7 and control green fluorescent protein (GFP) virus were used to infect Tu167 and JMAR cells, cell proliferation was inhibited by S100A7 overexpression in a dose-dependent manner, Similarly, when Tu167 or immortalized human keratinocytes HOK16B were infected with the virus, then subjected to a colony formation and cell migration assays, the number of colonies and cell migration were greatly suppressed by S100A7 overexpression (Figures 2c–e). To further extrapolate these findings in vivo, we used an orthotopic


3529

Cornified layer Suprabasal layer Basal layer Dermis

Tu167 Tumor

JMAR Tumor

DM14 Tumor

Tumor malignancy Figure 1 S100A7 is overexpressed in the early stages of well-differentiated, preinvasive squamous cell carcinoma of the oral cavity (SCCOC). (a) Representative results from immunohistochemical (IHC) staining of human SCCOC specimens with S100A7 antibody. (b) Summary of S100A7 IHC staining from 27 of human SCCOC specimens. *Po0.05, Wilcoxon rank-sum test. (c) Representative results from IHC staining with S100A7 antibody of orthotopic tongue SCCOC tumor specimens in nude mice derived from three human SCCOC cell lines.

tongue tumor model, in which JMAR cells were infected with the adenovirus and then orthotopically inoculated into tongues of nude mice. Consistent with our observation in vitro, JMAR tumor growth was suppressed by S100A7 overexpression by B5-fold as measured by tumor volume (Figure 2f). In addition, histological analysis revealed that tumors derived from control cells and S100A7 cells had different tumor morphologies (Figure 2g). In the control group (JMAR-GFP-Ad), tumors consisted of sheets of poorly differentiated and highly infiltrative cells. In contrast, S100A7-infected JMAR tumor cells (JMAR-S100A7Ad) grew as well-defined, organized islands that were well demarcated from surrounding mesenchymal tissue and well differentiated, as shown by the presence of keratin pearls. An immunofluorescent assay using Flag tag antibody further demonstrated that exogenous S100A7 was indeed expressed in JMAR-S100A7-Ad tumors (Figure 2h). In addition, in control tumors, immunohistochemical (IHC) staining demonstrated a high cellular proliferative index, as determined by extensive proliferating cell nuclear antigen (PCNA) staining throughout the tumor (Figure 2g). However,

in JMAR-S100A7-Ad tumors, PCNA staining was restricted to cells at the periphery of the tumor islands that did not express S100A7, while tumor cells expressing S100A7 exhibited no PCNA staining as shown in consecutive sections (Figure 2g). This inverse correlation in vivo between S100A7 and PCNA staining further highlights that overexpression of S100A7 inhibits tumor cell proliferation, tumor cell dissemination and invasion. Overexpression of S100A7 inhibits b-catenin signaling To study the possible mechanisms involved in S100A7mediated tumor growth inhibition, we examined the effects of overexpression of S100A7 on multiple signaling pathways including b-catenin signaling. We first noticed that the tyrosine phosphorylation of b-catenin induced by epidermal growth factor (EGF) stimulation was concomitantly inhibited by adenovirus-mediated expression of S100A7 in JMAR cells (Figure 3a). Further analysis indicated that the reduced levels of phosphorylated b-catenin could be attributed to an overall decrease in the level of b-catenin protein secondary to S100A7 overexpression (Figures 3b and c). Oncogene


3530

Figure 2 Overexpression of S100A7 inhibits cell proliferation in vitro and tumor growth in vivo. (a) Generation of recombinant adenovirus expressing Flag-tagged S100A7. (top) Schematic of adenoviral construct encoding S100A7. (bottom left) S100A7 adenovirus-infected JMAR cells (10 multiplicity of infection) marked by green fluorescent protein (GFP) fluorescence. (bottom right) Western blotting of JMAR cells infected with S100A7 and control GFP adenoviruses. (b) MTT assay for cell proliferation of Tu167 cells infected with S100A7 or GFP adenovirus. (c and d) Colony formation assay of Tu167 and HOK16B cells infected with different doses of S100A7 or GFP adenovirus. (e) Cell migration assay of Tu167 cells infected with the virus. (f) Orthotopic tongue tumor volumes derived from JMAR cells infected with or without virus. In all the above-mentioned cases, error bars represent standard deviation. *Po0.05. (g) Consecutive tumor sections of immunohistochemical (IHC) staining of tumors derived from control GFP (JMAR-GFP-Ad) and S100A7 virus-infected (JMAR-S100A7-Ad) tumors using antibodies as indicated (insets 200). (h) Immunofluorescent staining of frozen sections of control and JMAR-S100A7-Ad tumors using Flag antibody as indicated. In each group, 200 images on right hand side magnify box areas of the 50 images on left-hand side. Hematoxylin and eosin (H&E) section is the consecutive slide of fluorescent image on the top.

The finding that S100A7 decreases b-catenin expression led us to hypothesize that S100A7 expression could negatively regulate b-catenin signaling through nuclear translocation and activation of the T-cell factor (TCF)/ b-catenin transcription complex and subsequent activation of TCF-regulated genes including c-Myc (He et al., 1998). To test this hypothesis, a b-catenin/TCF-dependent luciferase reporter TopFlash and its b-catenin/TCFbinding site mutant control FopFlash (Korinek et al., Oncogene

1997) were co-transfected into 293T, Tu167 and JMAR cells together with b-catenin and S100A7 expression vectors. As shown in Figure 3d, the b-catenin-activated luciferase activity was inhibited in a dose-dependent manner by S100A7 overexpression in all three cell lines. Moreover, both basal and EGF-induced expressions of c-Myc were suppressed by S100A7 overexpression in JMAR cells (Figure 3e). Together, these findings support a role for S100A7 expression in inhibition of b-catenin signaling.


d -A 00 S1

IgG

A7

Ad

- + IB: P-Tyr

P-

+

GF

EGF -

S1 00 A7 -A

GF

PAd

d

3531

IP: β-Catenin

β-Catenin p63

S100A7 Actin

S100A7

Actin

β-Catenin

Merge

Relative Luc Activity

Lysate

FopFlash TopFlash β-Catenin S100A7

3

293T

5

Tu167

JMAR 3

GFP-Ad S100A7-Ad 0 0.5 1 2 0 0.5 1 2 EGF (h)

4 2

3 2

2

1

c-Myc S100A7

1

1

Actin JMAR 1 + -

2 + + -

3 + + +

4 + -

5 + + -

6 + +

7 + +

8 + +

9 10 11 + - - + + + - + -

12 13 - + + + +

14 + + -

15 + -

16 + +

17 + +

18 + +

-

Figure 3 Overexpression of S100A7 inhibits b-catenin signaling. (a) Immunoprecipitation (IP) with b-catenin antibody followed by immunoblotting (IB) with phosphotyrosine (P-Tyr) in the absence and presence of epidermal growth factor (EGF; 50 ng ml1) stimulation (30 min) in serum-starved JMAR cells infected with control green fluorescent protein (GFP) or S100A7 adenovirus. (b) Western blotting of b-catenin protein in JMAR cells after S100A7 and control GFP virus infection. (c) Representative confocal microscopic images of immunofluorescent staining of JMAR cells infected with S100A7 adenovirus. (d) Luciferase assay of b-catenindependent reporter TopFlash in 293T, Tu167 and JMAR cells. (e) Western blotting of c-Myc protein in serum-starved JMAR cells infected with virus after EGF (50 ng ml1) stimulation for different time (h).

S100A7 targets b-catenin degradation via a noncanonical mechanism that is independent of GSK3b-mediated phosphorylation To further investigate the mechanism involved in the inhibition of b-catenin signaling by S100A7, we examined the subcellular localization of S100A7. To visualize endogenous S100A7, we used breast cancer MDA-MB-468 cells because this cell line has high levels of S100A7 expression (Supplementary Figure S1B). As shown in Figure 4a, most of the S100A7 was localized in the cytoplasm as well as the cell cortex region. Reciprocal immunoprecipitation (IP) assays showed that S100A7 and b-catenin were co-precipitated from MDA-MB-468 and confluent JMAR cell lysates (Figure 4b), indicating that S100A7 interacts with the b-catenin complex in these cells. To understand whether the downregulation of bcatenin is due to transcriptional or post-transcriptional regulation, CMV promoter-driven Akt or b-catenin constructs were transfected with S100A7 into 293T cells. As shown in Figure 4c, the expression of both endogenous and transfected b-catenin was suppressed by S100A7 expression (compare lane 3 with 1, 2 and lane 6 with 5), whereas transfected Akt expression was not affected (compare lanes 3, 4 with 2). Moreover, the S100A7-mediated suppression of b-catenin was

inhibited by proteasome inhibitor MG132 (compare lane 4 with 3, and lane 7 with 6). Taken together, all these results demonstrated the specificity of S100A7mediated b-catenin inhibition and strongly suggested that it might be due to the post-transcriptional regulation, which is dependent on proteasome-mediated protein degradation. Because the degradation of b-catenin is controlled by a GSK3b-mediated phosphorylation-dependent (Hart et al., 1999; Kitagawa et al., 1999) (Moon et al., 2004; Willert and Jones, 2006) as well as Siah-1/Sip-mediated GSK3b phosphorylation-independent pathways (Liu et al., 2001; Matsuzawa and Reed, 2001; Fukushima et al., 2006), we used both luciferase assay and western blotting to test the possible mechanisms involved. As shown in Figure 4d, S100A7 did not appear to influence the activity of GSK3b, which regulates b-catenin levels by targeting it for degradation. However, when S100A7 was transfected together with Siah-1/Sip into 293T cells, it further decreased b-catenin protein levels and b-catenin-dependent luciferase activity (Figure 4d). Furthermore, when a constitutively active mutant b-catenin that is resistant to GSK3b-mediated phosphorylation and degradation (Bienz and Clevers, 2000; Matsuzawa and Reed, 2001) was transfected together with S100A7 into cells, it was also inhibited by S100A7 Oncogene


IP:

S1 00 A7

IB:

IgG

IP:

ate

Merge

-C

E-Cadherin

IgG

S100A7

nin

3532

IB: -Catenin

S100A7

IgG

IgG JMAR

-

+ -

+ + -

+ + + -

+

+ +

5

+ + + -Catenin HA (Akt) S100A7 Actin

1

2

3

4

5

6

7

Relative Luc Activity

MG132 S100A7 HA-Akt -Catenin

MDA-MB-468

4 3 2 1

Top Flash -Catenin S100A7 GSK3 Siah-1 Sip

S100A7 MT--Catenin WT--Catenin

+ -

+ + -

+ + + -

+ + + -

+ + + + -

+ + + -

+ + + + -

+ + + +

+ + + + + Flag--Catenin

Flag--Catenin S100A7 Actin Myc-GSK3 Flag-Siah-1 Myc-Sip Actin Figure 4 S100A7 targets b-catenin degradation. (a) Representative confocal microscopic images of fluorescent immunostaining with S100A7 (green) and E-cadherin (red) antibodies in MDA-MB-468 cells. (b) Immunoprecipitation (IP) assays using S100A7 or b-catenin antibodies as indicated. Cell lysates were from serum-starved (12 h) confluent JMAR or MDA-MB-468 cells as indicated. IB, immunoblotting. (c) Western blotting of 293T cells 48 h after transfection with indicated DNA. In lanes 4 and 7, transfected cells were treated with 10 mM MG132 for 3 h before harvest. (d) (top) Luciferase assay and (bottom) western blotting of 293T cells after transfection with indicated DNA in 12-well culture plate. (e) Western blotting of 293T cell transfected with wild-type (WT) or constitutive active mutant (MT) b-catenin in the presence or absence of increasing amounts of S100A7.

overexpression in a dose-dependent manner even though this mutant form of b-catenin is more stable than wildtype b-catenin (Figure 4e). Together, these results strongly indicate that S100A7 targets b-catenin for degradation via a noncanonical mechanism that is independent of GSK3b-mediated phosphorylation. b-Catenin negatively regulates S100A7 expression Since our data indicate that S100A7 inhibits b-catenin signaling, and Wnt/b-catenin signaling functions as a ‘master switch’ for proliferation versus differentiation (van de Wetering et al., 2002), we set out to determine whether b-catenin signaling could in turn regulate S100A7 expression in SCCOC cells. To assess this, we first treated both JMAR and MDA-MB-468 cells with LiCl, an inhibitor of GSK3b that negatively regulates b-catenin signaling (Hart et al., 1999; Kitagawa et al., Oncogene

1999; Moon et al., 2004; Willert and Jones, 2006). As shown in Figure 5a, upon treatment with LiCl, the levels of S100A7 protein were significantly reduced in both cell lines, and this inhibition appeared to be independent of proteasome-mediated protein degradation since proteasome inhibitor MG132 did not affect the downregulation of S100A7 by LiCl. Consistent with this, real-time RT-PCR analyses demonstrated that mRNA levels of S100A7 were inhibited by LiCl (Figure 5b). Since the inhibition of GSK3b by LiCl releases the suppression of b-catenin signaling, we hypothesized that b-catenin signaling may be directly implicated in the regulation of S100A7 expression. To test this, we overexpressed a constitutively active mutant of b-catenin in MDA-MB468 and JMAR cells, As shown in Figure 5c, retrovirusmediated overexpression of b-catenin in stable clones reduced the S100A7 expression in these cells, suggesting


MDA-MB-468 +

-

+ -

JMAR + +

-

+

+ -

+ +

LiCl MG132 S100A7 Actin

Relative mRNA levels of S100A7

3533 - LiCl + LiCl

1.2 1 0.8 0.6 0.4 0.2 0

MDA 468

RN AB sh RN AA

rs Ct

JMAR

sh

hR NA

-C at e- ab

e

pB

pB ab

e- -C at pB ab

pB ab

e

MDA-MB-468

-Catenin

-Catenin

-Catenin

S100A7

S100A7

S100A7

Actin

Actin JMAR

Actin JMAR

Figure 5 b-Catenin signaling inhibits S100A7 expression. (a) Western blotting of MDA-MB-468, and confluent JMAR cells in the absence () or presence ( þ ) of LiCl (40 mM) and/or MG132 (10 mM) for 6 h. (b) Relative levels of S100A7 mRNA measured by realtime RT-PCR. (c) Western blotting of stable MDA-MB-468, and confluent JMAR cells generated from control pBabe retrovirus and pBabe-b-cat retrovirus expressing a constitutive active mutant b-catenin. Stable cells are polyclonal pools established from selection with puromycin (2 mg ml1) after retroviral infection. (d) Western blotting of stable JMAR cells infected with control (Ctr) and b-catenin (A and B) shRNA-expressing lentivirus (see ‘Materials and methods’). Stable cells are pools of green fluorescent protein (GFP)-positive cells sorted by fluorescence-activated cell sorting (FACS).

that activation of b-catenin signaling is sufficient to inhibit S100A7 expression in these cell lines. Conversely, when two independent lentivirus-mediated b-catenin shRNAs (that is, no. A and B) were introduced into JMAR cells to downregulate b-catenin, S100A7 expression was induced under confluent culture condition (Figure 5d). Moreover, levels of S100A7 expression were inversely correlated with levels of b-catenin expression (Figure 5d). Taken together, our results have demonstrated that S100A7 inhibits b-catenin levels, and that b-catenin signaling in turn, inhibits the expression of S100A7 in a negative feedback loop. Expression of S100A7 and b-catenin is inversely correlated with each other in both in vitro and in vivo To further evaluate the reciprocal regulation of S100A7 and b-catenin, we examined the expression pattern of S100A7 and b-catenin both in vitro and in vivo. Although Tu167, JMAR and DM14 cells exhibit different levels of b-catenin expression in vitro (Supplementary Figure S1B), b-catenin expression was found to be downregulated by detachment from the extracellular matrix, a condition which leads to elevated levels of S100A7 (Figure 6a). Similarly, in both naturally occurring human SCCOC and orthotopic tongue tumor samples, b-catenin staining in S100A7-expressing areas

was found to be markedly diminished and localized to the cell membrane when compared with the adjacent S100A7-negative areas that exhibited higher levels of PCNA and b-catenin staining both in cytoplasm and nuclei (Figure 6b). The inverse correlation between S100A7 and b-catenin staining in tissue specimens lends further support to the hypothesis that these proteins are reciprocally regulated. Decreased expression of S100A7 enhances b-catenin signaling and promotes tumor growth in an orthotopic SCCOC xenograft mouse model To further study the physiological role of S100A7, we generated recombinant lentivirus expressing three independent shRNAs targeting different regions of S100A7 gene (Supplementary Materials and methods). Since the MDA-MB-468 cell line has a high level of S100A7 expression (Supplementary Figure S1B), we first used it to establish stable polyclonal cell lines to test whether these shRNAs can suppress S100A7 expression. As shown in Figure 7a, in stable polyclonal cell populations generated from shRNAs nos. 2 and 3, S100A7 expression was inhibited by more than 80% when compared to that in cells transfected with control shRNAs. More importantly, siRNA-mediated Oncogene


3534 Cell Detachment

Tu167 - +

JMAR - +

DM14 + S100A7

β-Catenin Actin

β-Catenin

S100A7

β-Catenin

Human SCCOC

S100A7

X 50

X 200

Case 1 H&E

Case 2 S100A7

β-Catenin

PCNA

Tu167 Tumor JMAR Tumor

Orthotopic SCCOC

X 50

X 200

X 50

X 200

Figure 6 Expression of S100A7 and b-catenin is inversely correlated in both in vitro and in vivo. (a) Western blotting of Tu167, JMAR and DM14 cells treated with cell detachment. Confluent cells were treated with ( þ ) or without () cell detachment for 24 h. (b) Representative immunohistochemical (IHC) images of consecutive tumor sections of human squamous cell carcinoma of the oral cavity (SCCOC) and orthotopic Tu167, JMAR tongue tumors. In each group, consecutive tumor sections stained with different antibodies are shown as indicated, in which 200 images at the bottom magnify box areas of the 50 images on the top. Note that S100A7-expressing area has a reduced staining of b-catenin (mainly on the cell membranes) and proliferating cell nuclear antigen (PCNA), when compared with adjacent S100A7-negative areas that has a higher level of both b-catenin (in both cytoplasm and nuclei) and PCNA staining.

suppression of S100A7 in shRNA polyclonal cells (nos. 2 and 3) resulted in increased b-catenin as well as c-Myc expression when compared with the control cells (no. 5) (Figure 7b). In addition, our results indicate that decreased expression of S100A7 also induces snail expression and inhibits E-cadherin-mediated cell–cell adhesion (Supplementary Figure S2). These data are consistent with the notion that S100A7 negatively regulates levels of b-catenin as well as its downstream target genes in this cell line. As shown in Figure 1c, Tu167 and JMAR tumors exhibit different levels of differentiation, malignancy and S100A7 expression in vivo. However the level of S100A7 expression in these two cell lines was virtually undetectable in vitro under subconfluent culture condition (Supplementary Figure S1B) unless induced by cell confluence or cell detachment (Figure 6a). In spite of this, we infected these cells with S100A7 shRNA lentivirus and established stable shRNA clones of both JMAR and Tu167 cells. Consistent with results in Oncogene

MDA-MB468 cells, these cells had decreased levels of S100A7, but elevated levels of b-catenin expression relative to control stable cells when examined under confluent culture conditions (Figure 7c). After orthotopic injection of these cells into the tongues of nude mice, JMAR cells with S100A7 shRNA (nos. 2 and 3) grew faster than control cells as measured by the tumor volume (Figure 7d), despite that no obvious morphological differences between two groups were observed (Supplementary Figure S3). However, since Tu167 tumor exhibits a more differentiated phenotype and a higher level of S100A7 expression than JMAR tumor in vivo (Figures 1c and 6b), our result demonstrated that downregulated S100A7 by S100A7 shRNA had a more impact on Tu167 tumor morphology, and that S100A7 shRNA Tu167 tumors exhibit a less differentiated phenotype when compared with the control (Figure 7e). In addition, an inverse correlation between S100A7 and b-catenin, PCNA was also found in both control and S100A7 shRNA tumors (Figure 7e), further


3535

S100A7 shRNA

#1

#2

S100A7 Ctr shRNA shRNA #5 #3 #2

Control shRNA

#3

#4

Control shRNA

S100A7

#5

#4

S100A7 Control S100A7 shRNA shRNA shRNA

#2 #3

#5

#2 S100A7

β-Catenin

S100A7

β-Catenin

c-Myc

Actin

Actin

Actin

MDA-MB468 MDA-MB468

JMAR

90 80 70 60 50 40 30 20 10 0

shRNA #3 shRNA #2 ctrl shRNA #4

* *

0

7 9 12 14 Day Post Cell Inoculation

Tu167 S100A7 shRNA #2

PCNA

β-Catenin

S100A7

H&E

Tu167 control shRNA #5

Tu167

3 Tumor Volum (mm )

JMAR

X50

X200

X50

X200

Figure 7 Decreased expression of S100A7 leads to increased levels of b-catenin, c-Myc and promotes tumor growth in an orthotopic squamous cell carcinoma of the oral cavity (SCCOC) xenograft mouse model. (a) Western blotting of lentiviral shRNA stably transfected MDA-MB468 cell lines. Each stable cell line was generated from pools of green fluorescent protein (GFP)-positive cells from different S100A7 shRNAs (nos. 1, 2 and 3) and control shRNA (nos. 4 and 5) lentivirus. (b) Western blotting of lentiviral shRNA transfected MDA-MB468 cells with different antibodies as indicated. (c) Western blotting of confluent Tu167 and JMAR shRNA stable polyclones by indicated antibodies. (d) Tumor volumes derived from JMAR shRNA stable polyclones. *Po0.05. (e) Representative immunohistochemical (IHC) images of consecutive sections of Tu167 tumors derived from S100A7 shRNA (no. 2) and control (no. 5) stable polyclones. In each panel, consecutive sections stained with different antibodies are shown as indicated, in which 200 images on right hand side magnify box areas of the 50 images on left hand side. Note that S100A7-expressing area has a reduced staining of b-catenin (mainly on the cell membranes) and proliferating cell nuclear antigen (PCNA), when compared with adjacent S100A7-negative areas that has higher levels of both b-catenin (in both cytoplasm and nuclei) and PCNA staining.

supporting that S100A7 expression inhibits b-catenin signaling and tumor cell proliferation in vivo. Therefore, from our vitro and in vivo analyses of overexpression and downregulation of S100A7, we conclude that S100A7 negatively regulates tumor growth and progression by targeting b-catenin signaling. Discussion In this study, we have shown that the reciprocal negative regulation of S100A7 and b-catenin signaling plays an important role in tumor progression of SCCOC. On the

basis of our results, we have proposed a working model for S100A7 function, which is summarized in Figure 8. In this model, S100A7 functions as a putative ‘tumor suppressor’ inhibiting tumor growth and tumor progression, despite its overexpression in the early stage of tumor development. In support of this model are our results showing that S100A7 inhibits SCCOC cell proliferation in vitro, promotes tumor differentiation and suppresses tumor growth, tumor cell invasion in vivo in an orthotopic xenograft SCCOC tumor model. Moreover, we have demonstrated that S100A7 is associated with the b-catenin complex, and it negatively regulates b-catenin signaling by targeting its Oncogene


3536 Pre-malignancy carcinoma in situ

Malignancy Invasive carcinoma

S100A7

S100A7 β-Catenin

Tumor Progression

β-Catenin

EMT

Degradation

X TCF

Differentiation

β-Catenin TCF

c-myc cyclin D1 etc.

Proliferation Invasion Metastasis

Figure 8 Working model of S100A7 in oral cancer tumor progression. Despite its overexpression in premalignant tumors, S100A7 actually suppresses tumor progression by targeting bcatenin signaling. Conversely, its downregulation by b-catenin signaling and/or other unknown mechanisms promotes tumor progression.

degradation via a noncanonical mechanism that is independent of GSK3b-mediated phosphorylation. Conversely, downregulation of S100A7 expression by siRNA was shown to enhance b-catenin signaling, including increased b-catenin and c-Myc, promoting tumor growth and tumor progression in vivo. Finally, we demonstrated that not only does S100A7 inhibit b-catenin, but b-catenin signaling can negatively regulate S100A7 expression to form a negative feedback loop. Therefore, the reciprocal negative regulation between S100A7 and b-catenin signaling highlights their important roles in tumor development and tumor progression. In contrast to our findings are previous studies of breast cancer MDA-MB-231 and MDA-MB-468 cells that suggested that S100A7 expression promotes tumor growth in nude mice (Emberley et al., 2003a, 2005; Krop et al., 2005). This discrepancy may be attributed to the biological differences between squamous and secretory epithelial cells, which are derived from distinct stem cell lineages. It should be noted that in agreement with our current results, a prior study of MDA-MB-468 cells indicated that S100A7 expression inhibited cell anchorage-independent growth, motility and invasion in in vitro assays (Krop et al., 2005). However, in the same study, downregulation of S100A7 expression by shRNAs did not increase MDA-MB-468 cell tumorigenicity in nude mice. The different in vitro and in vivo phenotypic behaviors of MDA-MB-468 cells remain unexplained. Unlike most cultured tumor cell lines that do not normally demonstrate constitutive S100A7 expression unless induced by differentiation stimuli (Hoffmann et al., 1994; Tavakkol et al., 1994; Di Nuzzo et al., 2000; Enerback et al., 2002; Kennedy et al., 2005), MDA-MB-468 is a relatively differentiated cell line that Oncogene

constitutively expresses extremely high levels of S100A7 protein (Supplementary Figure S1B) (Enerback et al., 2002; Emberley et al., 2003a), and exhibits a poorly tumorigenic phenotype in subcutaneous mouse model (Krop et al., 2005) when compared with our orthotopic SCCOC model (10 versus 2 weeks). Although our results indicate that decreased expression of S100A7 in MDAMB-468 cells increased b-catenin signaling in vitro (Figures 7a and b), how this will impact tumor growth in vivo in this particular cell line needs to be further investigated. However, given its unusually high levels of S100A7 expression when compared with most in vitro tumor cell lines, it is quite possible that MDA-MB-468 cells have developed certain genetic or epigenetic modifications that render them resistant to growth inhibition by S100A7 overexpression. It has been shown that activation of Wnt/b-catenin signaling functions as a master switch for proliferation versus differentiation (van de Wetering et al., 2002), and this plays a pivotal role in the initiation and progression of cancers (Morin and Weeraratna, 2003; Taketo, 2004). In normal epithelium, Wnt/b-catenin signaling pathways have been shown to control the specification, maintenance, and activation of stem cells, and dysregulation of the pathway often results in the development of familial and/or sporadic epithelial cancers (Blanpain et al., 2007). Our finding that S100A7 negatively regulates b-cateninsignaling pathways suggests that S100A7 functions through these pathways to inhibit cell proliferation and promote cell differentiation. In further support of this hypothesis, we have shown that S100A7 overexpression inhibited expression of the b-catenin target gene, c-Myc (Figure 3e). c-Myc, an oncogenic protein, has the ability to drive unrestricted cell proliferation, inhibit cell differentiation, stimulate cell growth and vasculogenesis, reduce cell adhesion and promote metastasis and genomic instability (Patel et al., 2004; Adhikary and Eilers, 2005). Conversely, the loss of Myc proteins not only inhibits cell proliferation and growth but also accelerates differentiation and increases cell adhesion (Patel et al., 2004; Adhikary and Eilers, 2005). Therefore, the tumor suppressive phenotype of S100A7 overexpression shown in our study is consistent with the phenotype derived from the loss of Myc function, supporting the involvement of S100A7 in the negative regulation of the Wnt/b-catenin/ Myc pathways. It should be noted that despite S100A7 overexpression in skin cells of patients suffering from psoriasis, a hyperproliferative skin disease characterized by abnormal differentiation, its role in psoriasis is largely unknown. On the basis of our observation, however, we believed that it plays a role in ‘abnormal differentiation’ rather than ‘proliferation’ in this disease. Consistent with this, both decreased b-catenin and lack of activation of Wnt/b-catenin signaling were observed in psoriatic skin lesions (Reischl et al., 2007). While our preliminary results indicate that S100A7 might interact with the b-catenin complex and target bcatenin degradation in a phosphorylation-independent manner (Figure 4), the detailed mechanisms involved are still largely unknown. Previously, other members of S100A7 proteins such as S100A1, S100A6, S100A12,


3537

S100B and S100P, were shown to interact with SIP in a calcium-dependent manner (Filipek et al., 2002). It remains to be determined whether S100A7 can also interact with SIP and participate in Siah-SIP-mediated b-catenin degradation as previously reported (Liu et al., 2001; Matsuzawa and Reed, 2001; Fukushima et al., 2006). Alternatively, S100A7 may target additional components of b-catenin-signaling or other important cell signaling pathways, which remain to be identified. Furthermore, although our results unambiguously demonstrate that b-catenin signaling provides negative feedback to inhibit S100A7 expression (Figure 5), the mechanisms involved are still unknown. Interestingly, c-Myc, one of downstream targets of b-catenin signaling, was recently shown to directly inhibit S100A7 expression (Kennedy et al., 2005). However, whether c-Myc or/and other downstream targets are also implicated in the inhibition of S100A7 expression by b-catenin signaling still remain to be further investigated. The exact mechanisms by which the biphasic expression (up- and downregulation) of S100A7 during the SCCOC tumorigenesis is achieved and their biological significance also remain to be determined. However, through this study, we have identified a critical role of S100A7 in a negative regulation of b-catenin signaling, and vice versa, in head and neck cancer cells. On the basis of our results, we believe that the initial increase in S100A7 expression in premalignant/highly differentiated tumors may represent a natural defense mechanism by which the cells may attempt to promote differentiation and restrict cell migration. When increased levels of S100A7 are lost by activated b-catenin signaling, the cells are then relieved from S100A7’s pro-differentiating/growth inhibitory effect, thereby unleashing their growth and invasive capacity. Therefore, despite its high levels of expression in early stage SCCOC tumorigenesis, S100A7 appears to function as a tumor suppressor by inhibiting tumor growth and tumor progression. Together, our findings contribute to elucidating the mechanistic roles of S100A7 and b-catenin signaling during tumor development and progression of SCCOC, and may help us to eventually develop diagnostic and therapeutic strategies for SCCOC. Materials and methods Materials and Plasmids Materials and plasmids are described in Supplementary Materials.

Cell culture, transient DNA and siRNA transfections Human SCCOC cell lines Tu167 and JMAR were cultured in Dulbecco’s modified Eagle’s medium, and human mammary adenocarcinoma cell line MDA-MB468 was cultured in RPMI 1640 medium. The spontaneously immortalized human oral keratinocyte cell line HOK16B was maintained in keratinocyte serum-free medium (Life Technologies, Bethesda, MD, USA) supplemented with EGF and bovine pituitary extract. The background information of these cells and transfection methods are described in Supplementary Materials. Generation of recombinant adenovirus, retrovirus and lentivirus The detail for generation of adenovirus, retrovirus and lentivirus for overexpression or suppression of S100A7 and b-catenin is described in Supplementary Materials. Cell proliferation, cell migration and colony formation assays Cell proliferation was determined using the tetrazolium-based (that is, MTT) assay as described in Yigitbasi et al. (2004) or cell counting. Cell migration and colony formation assays are described in Supplementary Materials Luciferase reporter assay Luciferase reporter assay is described in Supplementary Materials. Western blotting, IP and quantitative real-time RT-PCR The details of western blotting, IP and quantitative real-time RT-PCR are described in Supplementary Materials. Orthotopic injections of oral tumors The procedures for cell injection, tumor measurements and subsequent analyses were conducted as described in Yigitbasi et al. (2004) and Supplementary Materials. Tumor specimens, IHC and immunofluorescence microscopy The detail of tumor specimens, IHC and immunofluorescence microscopy is described in Supplementary Materials. Acknowledgements This work was supported by the National Institute of Dental and Craniofacial Research—NIH grant RO1DE01461301 and the NIH Cancer Center support grant CA16672 and CA69381. We thank Mrs Carol Johnston for the excellent technical assistance. We thank Dr PD McCrea, Dr M-C Hung, Dr X-W Wu, Dr E Fearon and Dr GP Nolan for providing plasmids.

References Adhikary S, Eilers M. (2005). Transcriptional regulation and transformation by Myc proteins. Nat Rev Mol Cell Biol 6: 635–645. Alowami S, Qing G, Emberley E, Snell L, Watson PH. (2003). Psoriasin (S100A7) expression is altered during skin tumorigenesis. BMC Dermatol 3: 1. Bienz M, Clevers H. (2000). Linking colorectal cancer to Wnt signaling. Cell 103: 311–320. Blanpain C, Horsley V, Fuchs E. (2007). Epithelial stem cells: turning over new leaves. Cell 128: 445–458.

Broome AM, Ryan D, Eckert RL. (2003). S100 protein subcellular localization during epidermal differentiation and psoriasis. J Histochem Cytochem 51: 675–685. Di Nuzzo S, Sylva-Steenland RM, Koomen CW, De Rie MA, Das PK, Bos JD et al. (2000). Exposure to UVB induces accumulation of LFA-1+ T cells and enhanced expression of the chemokine psoriasin in normal human skin. Photochem Photobiol 72: 374–382. Donato R. (2001). S100: a multigenic family of calcium-modulated proteins of the EF-hand type with intracellular and extracellular functional roles. Int J Biochem Cell Biol 33: 637–668. Oncogene


3538 Donato R. (2003). Intracellular and extracellular roles of S100 proteins. Microsc Res Tech 60: 540–551. Emberley ED, Niu Y, Curtis L, Troup S, Mandal SK, Myers JN et al. (2005). The S100A7-c-Jun activation domain binding protein 1 pathway enhances prosurvival pathways in breast cancer. Cancer Res 65: 5696–5702. Emberley ED, Niu Y, Leygue E, Tomes L, Gietz RD, Murphy LC et al. (2003a). Psoriasin interacts with Jab1 and influences breast cancer progression. Cancer Res 63: 1954–1961. Emberley ED, Niu Y, Njue C, Kliewer EV, Murphy LC, Watson PH. (2003b). Psoriasin (S100A7) expression is associated with poor outcome in estrogen receptor-negative invasive breast cancer. Clin Cancer Res 9: 2627–2631. Enerback C, Porter DA, Seth P, Sgroi D, Gaudet J, Weremowicz S et al. (2002). Psoriasin expression in mammary epithelial cells in vitro and in vivo. Cancer Res 62: 43–47. Filipek A, Jastrzebska B, Nowotny M, Kuznicki J. (2002). CacyBP/SIP, a calcyclin and Siah-1-interacting protein, binds EF-hand proteins of the S100 family. J Biol Chem 277: 28848–28852. Fukushima T, Zapata JM, Singha NC, Thomas M, Kress CL, Krajewska M et al. (2006). Critical function for SIP, a ubiquitin E3 ligase component of the beta-catenin degradation pathway, for thymocyte development and G1 checkpoint. Immunity 24: 29–39. Hart M, Concordet JP, Lassot I, Albert I, del los Santos R, Durand H et al. (1999). The F-box protein beta-TrCP associates with phosphorylated beta-catenin and regulates its activity in the cell. Curr Biol 9: 207–210. He TC, Sparks AB, Rago C, Hermeking H, Zawel L, Da Costa LT et al. (1998). Identification of c-MYC as a target of the APC pathway. Science 281: 1509–1512. Heizmann CW, Fritz G, Schafer BW. (2002). S100 proteins: structure, functions and pathology. Front Biosci 7: d1356–d1368. Hoffmann HJ, Olsen E, Etzerodt M, Madsen P, Thogersen HC, Kruse T et al. (1994). Psoriasin binds calcium and is upregulated by calcium to levels that resemble those observed in normal skin. J Invest Dermatol 103: 370–375. Kennedy RD, Gorski JJ, Quinn JE, Stewart GE, James CR, Moore S et al. (2005). BRCA1 and c-Myc associate to transcriptionally repress psoriasin, a DNA damage-inducible gene. Cancer Res 65: 10265–10272. Kitagawa M, Hatakeyama S, Shirane M, Matsumoto M, Ishida N, Hattori K et al. (1999). An F-box protein, FWD1, mediates ubiquitin-dependent proteolysis of beta-catenin. EMBO J 18: 2401–2410. Korinek V, Barker N, Morin PJ, van Wichen D, De Weger R, Kinzler KW et al. (1997). Constitutive transcriptional activation by a beta-catenin-Tcf complex in APC/ colon carcinoma. Science 275: 1784–1787.

Krop I, Marz A, Carlsson H, Li X, Bloushtain-Qimron N, Hu M et al. (2005). A putative role for psoriasin in breast tumor progression. Cancer Res 65: 11326–11334. Leygue E, Snell L, Hiller T, Dotzlaw H, Hole K, Murphy LC et al. (1996). Differential expression of psoriasin messenger RNA between in situ and invasive human breast carcinoma. Cancer Res 56: 4606–4609. Liu J, Stevens J, Rote CA, Yost HJ, Hu Y, Neufeld KL et al. (2001). Siah-1 mediates a novel beta-catenin degradation pathway linking p53 to the adenomatous polyposis coli protein. Mol Cell 7: 927–936. Martinsson H, Yhr M, Enerback C. (2005). Expression patterns of S100A7 (psoriasin) and S100A9 (calgranulin-B) in keratinocyte differentiation. Exp Dermatol 14: 161–168. Matsuzawa SI, Reed JC. (2001). Siah-1, SIP, and Ebi collaborate in a novel pathway for beta-catenin degradation linked to p53 responses. Mol Cell 7: 915–926. Moog-Lutz C, Bouillet P, Regnier CH, Tomasetto C, Mattei MG, Chenard MP et al. (1995). Comparative expression of the psoriasin (S100A7) and S100C genes in breast carcinoma and co-localization to human chromosome 1q21-q22. Int J Cancer 63: 297–303. Moon RT, Kohn AD, De Ferrari GV, Kaykas A. (2004). WNT and beta-catenin signalling: diseases and therapies. Nat Rev Genet 5: 691–701. Morin PJ, Weeraratna AT. (2003). Wnt signaling in human cancer. Cancer Treat Res 115: 169–187. Patel JH, Loboda AP, Showe MK, Showe LC, McMahon SB. (2004). Analysis of genomic targets reveals complex functions of MYC. Nat Rev Cancer 4: 562–568. Reischl J, Schwenke S, Beekman JM, Mrowietz U, Sturzebecher S, Heubach JF. (2007). Increased expression of Wnt5a in psoriatic plaques. J Invest Dermatol 127: 163–169. Taketo MM. (2004). Shutting down Wnt signal-activated cancer. Nat Genet 36: 320–322. Tavakkol A, Zouboulis CC, Duell EA, Voorhees JJ. (1994). A retinoic acid-inducible skin-specific gene (RIS-1/psoriasin): molecular cloning and analysis of gene expression in human skin in vivo and cultured skin cells in vitro. Mol Biol Rep 20: 75–83. van de Wetering M, Sancho E, Verweij C, De Lau W, Oving I, Hurlstone A et al. (2002). The beta-catenin/TCF-4 complex imposes a crypt progenitor phenotype on colorectal cancer cells. Cell 111: 241–250. Watson PH, Leygue ER, Murphy LC. (1998). Psoriasin (S100A7). Int J Biochem Cell Biol 30: 567–571. Willert K, Jones KA. (2006). Wnt signaling: is the party in the nucleus? Genes Dev 20: 1394–1404. Yigitbasi OG, Younes MN, Doan D, Jasser SA, Schiff BA, Bucana CD et al. (2004). Tumor cell and endothelial cell therapy of oral cancer by dual tyrosine kinase receptor blockade. Cancer Res 64: 7977–7984.

Supplementary Information accompanies the paper on the Oncogene website (http://www.nature.com/onc).

Oncogene