Cancer predisposition in mice deficient for the metastasis-associated Mts1(S100A4) gene

Cancer predisposition in mice deficient for the metastasis-associated

Mts1(S100A4) gene

Christina EL Naaman1,5, Birgitte Grum-Schwensen1,5, Ahmed Mansouri2, Mariam Grigorian1,Eric Santoni-Rugiu3, Thomas Hansen1, Marina Kriajevska1, Beat W Schafer4, Claus WHeizmann4, Eugene Lukanidin1 and Noona Ambartsumian1,*

1Department of Molecular Cancer Biology, Danish Cancer Society, Strandboulevarden 49, DK2100 Copenhagen, Denmark;2Department of Molecular Cell Biology, Max-Planck Institute for Biophysical Chemistry, Am Fassberg 11, D37077 Gottingen,Germany; 3Department of Pathology, Rigshospitalet, Blegdamsvej 9, DK2100 Copenhagen, Denmark; 4Division of Clinical Chemistryand Biochemistry, Department of Pediatrics, University of Zurich, Steinwiesstrasse 75, CH-8032 Zurich, Switzerland

Metastasis-promoting Mts1(S100A4) protein belongs tothe S100 family of Ca2þ -binding proteins. A mouse strainwith a germ-line inactivation of the S100A4 gene wasgenerated. The mice were viable and did not displaydevelopmental abnormalities in the postnatal period.However, an abnormal sex ratio was observed in thelitters with the S100A4�/� genotype, raising thepossibility of a certain level of embryonic lethality in thisstrain. In all, 10% of 10–14-month-old S100A4-nullanimals developed tumors. This is a characteristic featureof mouse strains with inactivated tumor suppressor genes.Spontaneous tumors of S100A4�/� mice were p53positive. Recently, we have shown that S100A4 interactswith p53 tumor suppressor protein and induces apoptosis.We proposed that impairment of this interaction couldaffect the apoptosis-promoting function of p53 that isinvolved in its tumor suppressor activity. The frequency ofapoptosis in the spleen of S100A4�/� animals afterwhole-body c-irradiation was reduced compared to thewild-type animals. The same was true for the transcrip-tional activation of the p53 target genes – waf/p21/cip1and bax. Taken together, these observations indicate thatspontaneous tumors in S100A4�/� mice are a result offunctional destabilization of p53 tumor suppressor gene.Oncogene (2004) 23, 3670–3680. doi:10.1038/sj.onc.1207420

Keywords: animal models; Mts1(S100A4); p53; tumordevelopment; tumor suppressor activity

Introduction

Metastasis-inducing S100A4 protein belongs to the S100family of Ca2þ -binding proteins. Members of this familyfunction both intra- and extracellularly. Inside a cell,they are implicated in a variety of activities such as cellproliferation and differentiation, cytoskeleton dynamics

and apoptosis. When released into extracellular space,S100 proteins stimulate neuronal differentiation, astro-cyte proliferation, modulate activity of inflammatorycells and stimulate angiogenesis (for a review, seeDonato, 2001; Heizmann et al., 2002).Some of the S100 family members are involved in

cancer development (Weterman et al., 1993; Pedrocchiet al., 1994; Al-Haddad et al., 1999; Guerreiro Da Silvaet al., 2000; Feng et al., 2001). In this respect, muchinterest has been focused on S100A4 since its expressionwas strongly associated with stimulation of metastasis.The S100A4 gene was isolated as a gene differentiallyexpressed in metastatic mouse mammary adenocarcino-ma cells (Ebralidze et al., 1989). Introduction of theS100A4 gene into a number of nonmetastatic tumor celllines and suppression of its activity in metastatic onesproved its involvement in metastasis formation (Maz-zucchelli, 2002). Stimulation of tumor metastatic activ-ity was demonstrated in two transgenic mouse modelswith overexpression of the S100A4 gene (Ambartsumianet al., 1996; Davies et al., 1996). In a number of humancancers, the enhanced expression of S100A4 wasassociated with poor prognosis (Platt-Higgins et al.,2000; Yonemura et al., 2000; Davies et al., 2002; Rostyet al., 2002). The exact mechanism by which S100A4stimulates metastasis formation is poorly understood.When applied extracellularly, S100A4 stimulates

neurite outgrowth (Novitskaya et al., 2000) and exhibitsangiogenic activity (Ambartsumian et al., 2001). Thisimplies that S100A4 could contribute to tumor progres-sion by stimulating neovascularization of a tumor. Onthe other hand, a number of intracellular interactingpartners, such as heavy chain of nonmuscle myosin(Kriajevska et al., 1994; Ford and Zain, 1995), liprinb-1(Kriajevska et al., 2002), p53 (Grigorian et al., 2001) andrecently methionine aminopeptidase (Endo et al., 2002)were described for the S100A4 protein, suggesting itsparticipation in cell motility, adhesion and proliferation.This raised the possibility that S100A4 participates inmetastasis by interfering with any of these processes.One of the interacting partners of S100A4 in the cell is

p53 tumor suppressor protein (Chen et al., 2001;Grigorian et al., 2001). The wild-type p53 plays acentral role in reducing the probability of a cell

Received 26 August 2003; revised 3 November 2003; accepted 28November 2003

*Correspondence: N Ambartsumian; E-mail: [email protected] authors contributed equally to the work

Oncogene (2004) 23, 3670–3680& 2004 Nature Publishing Group All rights reserved 0950-9232/04 $25.00

www.nature.com/onc

becoming cancerous. It controls the induction ofapoptosis and growth arrest at cell cycle checkpoints.This function of p53 leads to elimination of damagedcells from a population and is believed to be a basis of itstumor suppressor function (Vousden, 2002). Transacti-vation of p53 target genes is regarded as one of themechanisms of regulation of p53 response (Bourdonet al., 1997; El-Deiry, 1998). In addition, wild-type p53regulates expression of several metastasis-related genes,such as several matrix metalloproteinases, cathepsin D,thrombospondin-1 and KAI1 (Dameron et al., 1994;Bian and Sun, 1997; Mashimo et al., 1998; Wu et al.,1998; Sun et al., 1999, 2000). By interacting with p53,S100A4 enhances p53-dependent apoptosis and differ-entially modulates p53 target gene expression (Grigorianet al., 2001).Although S100 genes are structurally and evolutiona-

rily related, the intracellular location, expression patternand localization of each individual S100 protein isdistinct (Mandinova et al., 1998; Hsieh et al., 2002).Expression of the S100A4 in an adult organism isrestricted mainly to the cells of the lymphoid system,presumably T lymphocytes and macrophages (Lukani-din and Georgiev, 1996). In embryonic development,S100A4 also demonstrates a very distinct pattern ofexpression in several mesenchymal tissues and in fetalmacrophages (Klingelhofer et al., 1997). The importantrole of S100A4 in embryogenesis could also bepostulated from its high level of expression in mousetrophoblast (Ford and Zain, 1995).S100A4 gene is located in a cluster of S100 genes on

human chromosome 1q21 and in a syntenic region onmouse chromosome 3 (Ridinger et al., 1998). Despitecluster organization, there are no indications of acoordinated regulation of transcription of S100 genes,each member of the family being under a complextranscriptional regulation (Chen and Barraclough, 1996;Cohn et al., 2001). Cell- and tissue-specific expression ofthe S100A4 protein is supported by a complex regula-tion on a post-transcriptional level (Ambartsumian et al.,1998–1999). Moreover, the S100A4 gene is differentlyexpressed in the tissues of related species, such as rat andmouse (Davies et al., 1995).To address the questions of the role of S100A4 in

normal mouse organism and in cancer development, wegenerated a mouse strain with a germ-line inactivationof the S100A4 gene.Here we show that deletion of S100A4 in the mouse

does not affect embryonic development. However, apart of S100A4-deficient mice was prone to spontaneoustumor development, suggesting the participation ofS100A4 in tumor suppressor function. Whole-body g-irradiation of mice revealed impairment of p53 functionin the S100A4�/� animals, which results in decreasedapoptosis and suppression of p53 target gene transcrip-tion. We suggest that disrupted interaction of S100A4with p53 affects its function as a guardian of thegenome raising the probability of tumor formation inS100A4�/� animals.Analysis of the expression of the neighboring S100

genes – S100A3 and S100A5 – in S100A4-null animals

demonstrated alterations in the expression of S100A5raising the possibility of a compensatory mechanism ofthe S100A4 activity in mice.

Results

Generation of S100A4-deficient mice

The murine S100A4 gene was inactivated by homo-logous recombination in ES cells, using the strategydepicted in Figure 1a. The entire coding sequence of theS100A4 gene from exons 2 and 3 was replaced by acassette containing selective markers flanked by loxPsites of bacteriophage P1. Next, the cassette was excisedby means of the Cre/Lox recombination system ofbacteriophage P1 (Sauer, 1993). Selected clones of EScells were aggregated with the morulae and transferredinto a foster mother. Mice bearing the deletion ofS100A4 were identified by Southern blot analysis(Figure 1b). The S100A4�/� mice were born viable,demonstrating that S100A4 is not required for embryo-nic development. Histological analysis of the tissuesobtained from 12- and 24-week-old animals did notreveal any abnormalities as compared to the controlmice (data not shown). S100A4 RNA and proteinwere not detected in the tissues of these mice (Figure 1cand d).Breeding data from the strain carrying deletion of the

S100A4 gene revealed a deficiency of S100A4�/�female progeny (Table 1). The reason for abnormalsex ratio is not clear and is currently under investigation.

Spontaneous tumor development in S100A4-null mice

We followed a cohort of animals bearing homozygousdeletion of S100A4 gene for an extended period of timefor a possible appearance of pathological manifesta-tions. Quite unexpectedly, 12% of the 47–59-week-oldS100A4-null animals developed tumors. Control mice ofmatching genetic background were tumor free at thisage (Table 2). Moreover, a tumor was also detected inone of the S100A4�/� animals of the 21–33-week-oldgroup. Two of the knockout animals developed twoindependent tumors. A total of eight tumors weredetected and subjected to histopathological analysis.The tumors were predominantly of epithelial origin,most frequently carcinomas in the lung (Table 3). Thelatter were represented by bronchioalveolar carcinomasof solid type (Turusov and Mohr, 1994), composed ofdense tumor masses infiltrating the alveolar spaces withundefined cellular borders and relatively mild nuclearpleomorphism (Figure 2a). The tumor in the mammarygland was diagnosed as an adenocarcinoma type B(Turusov and Mohr, 1994) composed of solid undiffer-entiated structures with a minor tubular component(Figure 2b) similar to invasive ductular carcinomas inhumans. The other tumors detected were a hepatocel-lular carcinoma surrounded by severe liver celldysplasia, a fibrosarcoma in the uterus and lymphomawith metastases to the liver. We did not observe

Spontaneous tumors in the S100A4-null miceC El-Naaman et al

3671

Oncogene

metastases in the animals bearing tumors of epithelialorigin.A high rate of spontaneous tumor incidence is a

characteristic feature of knockout mouse strains withinactivated tumor suppressor genes (Hakem and Mak,2001). To our knowledge, S100A4 does not exhibittumor suppressor function. Previously we have shownthat S100A4 interacts with p53 tumor suppressor

protein, stimulating p53-dependent apoptosis and af-fecting the transactivation of p53-dependent genes(Grigorian et al., 2001). Development of the tumors inS100A4�/�mice only in some of the animals and with asubstantial delay in the life together with the previouslyestablished interaction of S100A4 with p53 proteinallowed us to speculate that spontaneous tumor devel-opment in the S100A4�/� mice is a result of modula-tion of p53 tumor suppressor function.To test this hypothesis, we first determined the status

of p53 in spontaneous tumors developed in the S100A4-null animals.Under physiological circumstances, the wild-type p53

protein is detectable by immunohistochemical methodsonly in solitary cells in the tissues (Greenblatt et al.,

Figure 1 Generation of S100A4 knockout mice. (a) Targeting strategy at the S100A4 locus. Exons 2 and 3 are replaced by theneomycin- and thymidin kinase-containing cassette in a recombinant allele generated by homologous recombination. Probe I was usedto identify homologous recombinant ES cells. Selection cassette was excised by transfection of selected clone with Cre-recombinasebearing plasmid. Probe II was used to identify ES clones bearing deletion of the S100A4 gene sequences. (b) Southern blot analysis ofgenomic DNA. A restriction digest with ScaI when hybridized with probe II shows two bands, 8 kb represents the knockout allelewhereas 6kb represents the wild type. (c) Northern blot analysis of total RNA isolated from different organs of þ /þ , þ /� and �/�animals. Hybridization with S100A4 probe. (d) Immunoprecipitation of S100A4 protein from primary embryonic fibroblasts isolatedfrom þ /þ and �/� animals. CSML100 cell lysate expressing S100A4 was used as a positive control

Table 1 Numbers of S100A4�/� progeny at weaning

Genotype No. oflitters

No. atweaning

No. offemales

No. ofmales

No. females/no. males

+/+ 13 139 72 67 1.09�/� 18 160 60 101 0.59


3672

Oncogene

1994). An alteration of p53 stabilizes the protein andmakes it detectable by immunohistochemical analysis.Detection of p53 protein is regarded as a prognosticmarker in a number of human cancers (Greenblatt et al.,1994). Such alterations of p53 are a rare event inspontaneous and chemically induced cancers in rodents(Buzard, 1996; Cazorla et al., 1998; Duddy et al., 1999).Immunohistochemical staining of the spontaneous

tumors of the S100A4-null mice (n¼ 5) revealed p53-positive nuclei present in all the tumors analysed (seeFigure 2c and d), indicating stabilization of p53 protein.p53-positive nuclei were also revealed with anti-p53antibodies that specifically recognize a mutant form ofp53 protein (clone 240, data not shown). Moreover,spontaneous tumors of the S100A4-null mice expressedhigh levels of proliferating cell nuclear antigen (PCNA),a negatively regulated target of wild-type p53 (Merceret al., 1991) (Figure 2e and f). In contrast, expression ofp21Waf, a gene that is positively regulated by p53 was notdetected in these tumors (Figure 2g and h). Therefore,assay of the activity of p53 responsive genes togetherwith enhanced expression of the p53 protein itselfsuggests that p53 function is impaired in these tumors.The functional status of p53 was tested by measuring

the level of apoptosis in the thymuses of 2- and 7-month-old knockout animals and compared it with the basallevel of apoptosis in the wild-type animals. The level ofapoptosis was measured by TUNEL assay. We haveobserved a certain decrease in the level of apoptosisin the thymuses of 2-month-old S100A4�/� micecompared to the control, but this difference was notstatistically significant (Table 4).

Alteration of p53-dependent functions in S100A4�/�mice after whole-body g-irradiation

Following g-irradiation in vivo, a number of p53-regulated genes are strongly activated in ap53-dependent manner (Komarova et al., 2000); p53-dependent apoptosis is induced shortly after irradiation

in radiation-sensitive organs (Bouvard et al., 2000).Different responses to whole-body irradiation couldreflect changes in the functional status of p53 and can beused as a test for p53 activity in vivo (Komarova et al.,2000).We studied the effect of the whole-body g-irradiation

in tissues of S100A4�/� and wild-type mice. Normally,Waf/p21/cip1 and bax, p53-regulated genes are inducedin response to whole-body g-irradiation in radiation-sensitive organs, such as the spleen and thymus(Bouvard et al., 2000). The expression of these genesin mouse organs from irradiated wild-type and null micewas analysed by Northern blot hybridization(Figure 3a). The waf1/p21/cip1 gene was highly inducedin the thymus and spleen of both wild-type andknockout animals in response to ionizing radiation;bax was also induced, but to a lesser extent. Notably, theinduction of transcription of both genes was less in thespleen of S100A4 knockout animals. Western blotanalysis of spleen lysates with anti-p21Waf antibodiesconfirmed this observation (Figure 3b). The averagelevel of induction of waf1/p21/cip1 and bax wascomparable in the thymus, whereas in the spleen ofS100A4�/� mice it was two times less than in thecontrol (Figure 3c).We also compared the level of apoptosis in the spleen

of the animals after irradiation by TUNEL assay.Strong apoptosis was detected in the white pulp of bothknockout and the wild-type animals. Virtually all thecells contain apoptotic nuclei making the quantificationimpossible. Apoptosis was also observed in the red pulpof the spleen and in this case the amount of apoptoticnuclei in the TUNEL assay was quantified (Figure 4a).The amount of apoptotic nuclei in the spleen wasapproximately two times less in the g-irradiatedS100A4�/� animals compared to the wild-typecontrols.In conclusion, evaluation of the results obtained by

Northern blot hybridization analysis of p53 target genesand assessment of the level of apoptosis in the spleen ofS100A4�/� g-irradiated animals supports the hypoth-esis that p53-dependent functions are altered in thesemice.

Expression of S100A4 adjacent genes in the S100A4-nullanimals

We studied the possibility of alterations in the expres-sion of neighboring S100 genes in response to germ-linedeletion of S100A4-specific sequences.

Table 2 Tumor incidences in the S100A4�/� mice

Age (weeks) S100A4(�/�) S100A4(+/+) P-value*

No. of animals No. with tumors No. of animals No. with tumors

8–20 10 0 10 0 —21–33 12 1 10 0 0.1734–46 12 0 12 0 —47–59 40 5 34 0 0.015

*P value was determined using the Wilcoxon rank sum test

Table 3 Spectrum of spontaneous tumors in S100A4-null mice

Type of the tumor No. of tumors Relative incidence (%)

Bronchioalveolar carcinoma 4 50Mammary gland carcinoma 1 12.5Hepatocellular carcinoma 1 12.5Fibrosarcoma 1 12.5Lymphoma 1 12.5


3673

Oncogene

Northern blot hybridization analysis of RNA isolatedfrom different organs of wild-type (þ /þ ), heterozy-gous (þ /�) and S100A4-null (�/�) mice with probesspecific for the S100 genes lying in the close vicinity ofthe S100A4 gene revealed that the expression profile ofupstream S100A3 was not changed. The expression of

downstream gene, S100A5, although was altered (Figure5a and b). In the wild-type mice, S100A5 RNA is mostlyexpressed in certain areas of the brain, but not in theother organs (Schafer et al., 2000). In the RNAextracted from the organs of S100A4-null animal, theS100A5 gene expression was detected in the thymus,

Figure 2 Tumors in the S100A4�/� mice. (a, b) Bronchioalveolar carcinoma in the lung (a) and mammary gland carcinoma (b) H&E(� 20 objective). (c, d) (Immunohistochemical staining of lung carcinoma (c) and mammary gland carcinoma (d) with anti-p53antibodies (� 40 objective). Note the positive nuclei indicated by arrows. (e, f) Immunohistochemical staining of lung carcinoma (e)and mammary gland carcinoma (f) with anti-PCNA antibodies (� 40 objective). Note the positive nuclei indicated by arrows. (g, h)Immunohistochemical staining of lung carcinoma (g) and mammary gland carcinoma (h) with anti-p21 antibodies (� 40 objective).Note the expression of p21Waf only in some of the stromal cells, indicated by arrows, but not in the tumor cells. Positive nuclei areindicated by arrows


3674

Oncogene

spleen, skin, but absent from the liver (Figure 3b). Theprofile of the S100A5 expression corresponded to theprofile of the S100A4 gene expression in adult wild-typeanimals (Figure 5a, compare with Figure 1c).This result suggests that inactivation of S100A4 leads

to activated expression of neighboring S100A5 gene insome of the cells, normally expressing S100A4 and couldpartially compensate the S100A4 activity.Activation of S100A5 in the S100A4�/� animals

raised the possibility that spontaneous tumors detectedin these animals might be due to putative oncogenicactivity of S100A5. To explore this possibility, we testedS100A5 expression in the spontaneous tumors of

S100A4-null animals. Immunohistochemical stainingof tumors of S100A4-null animals with anti-S100A5antibodies revealed that tumor cells did not expressS100A5. We observed single positive cells in the stromaof all the tumors tested (Figure 5b).We also tested whether the S100A5 expression is

activated in mouse tumor cell lines and compared it withthe expression of S100A3 and S100A4. S100A4 andS100A3 were expressed in a number of mouse tumorcells, but S100A5 was silent in the entire tumor cell linestested (Table 5). Similar results were obtained both withNorthern blot analysis and RT–PCR. These data do nottherefore support the assumption of the oncogenicfunction of S100A5.

Discussion

We have inactivated the S100A4 gene in order to clarifyits role in mouse development and promotion ofmetastatic disease.Knockout animal models were applied to study the

functions of S100 proteins (Heizmann et al., 2002).While S100B is not essential for life, glial cells fromS100B-null mice display an altered Ca2þ homeostasis(Xiong et al., 2000). S100A1-deficient mice did notdisplay any abnormalities in growth and developmentunder baseline conditions, but demonstrate reducedCa2þ -sensitivity and impaired cardiac contractility inresponse to hemodynamic stress (Du et al., 2002).

Table 4 Quantification of apoptosis in the thymuses of S100A4knockout and wild-type animals

Age of the mice 2-month-old 7-month-old

Wild type 15.874.15* 11.5571.79S100A4-null 10.772.88 9.874.29

*Amount of apoptotic nuclei per field7s.d. (Po0.094)

Figure 3 Expression of waf/p21/cip1waf and bax mRNA inS100A4�/�mouse organs after whole-body g-irradiation. (a)Representative Northern blots hybridized with p21/waf/cip andbax probes. (b) Western blot analysis of whole-cell lysates of spleenfrom wild-type and knockout mice with probed with anti-p21antibodies. Anti-a-tubulin antibodies were used as a loadingcontrol. (c) Comparison of normalized mRNA levels in the spleenand thymus of wild-type and S100A4�/� mice after irradiation.The mean expression level in the wild-type animals was taken as100% and compared to the mean value (7s.d.) of expression in theorgans of S100A4�/� animals

Figure 4 Comparison of levels of apoptosis (TUNEL assay) in thespleen of S100A4�/� and wild-type (þ /þ ) mice 6 h after g-irradiation. (a) TUNEL assay in the red pulp of the spleen.Magnification 400. (b) Quantification of the amount of apoptoticnuclei in the red pulp of wild-type and S100A4�/� mice. The meanamount of apoptotic nuclei per field (7s.d.) is compared for groupsof three animals (Po0.02)


3675

Oncogene

S100A9-deficient neutrofils demonstrated altered migra-tory behavior and did not affect myeloid cell function(Hobbs et al., 2003; Manitz et al., 2003). On the otherhand, germ-line inactivation of S100A8 indicates that itis required for embryonic development of the mouseembryo (Passey et al., 1999).Targeted deletion of S100A4 does not largely affect

embryonic and postnatal development of the mice.However, misbalance in the ratio of sexes in the progenyof S100A4-null crosses indicates that S100A4 deletioncauses some female-related embryonic lethality. Devel-opmental abnormalities leading to a disordered sex ratiowere observed in particular in the p53-null mice. Asubset of the p53�/� female embryos developed defectsin neural tube closure and exencephaly (Armstrong et al.,1995; Sah et al., 1995).Another interesting finding is the propensity of

S100A4�/� mice to develop tumors in different organs.A wide range of tumors of different origin werediagnosed in the S100A4-null animals, which is char-acteristic of germ-line inactivated tumor suppressorgenes (Hakem and Mak, 2001; Meuwissen et al.,2001). This raises the possibility that S100A4 couldcontribute to tumor suppression. Participation of S100proteins in tumor suppression was postulated forS100A2, given its downregulation in progressed formsof mammary and lung carcinomas (Feng et al., 2001).

Figure 5 Expression of S100A3 and S100A5 in tissues ofS100A4�/� mice. (a) Northern blot of total RNA isolated fromdifferent organs of þ /þ , þ /� and �/� animals. Hybridizationwith S100A3and S100A5 probes. Hybridization of the same filterwith polyU was used as a loading control. Immunohistostaining ofmammary gland carcinoma (b) and bronchioalveolar carcinoma ofthe lung (c) with anti-S100A5 antibodies. Arrows point to S100A5-positive fibroblasts (� 400 objective). (d) Antibody blockingcontrol

Table

5ExpressionofS100A3,S100A4andS100A5inmousetumorcelllines

VMR-lymph

VMR-liver

CSML-0

CSML-100

Line-1

LL-met

10T1/2

MT1TC1

MT1TC3

RAC34E

RAC5E

RAC10P

RAC311C

NIH3T3

LL-G2

Tumortype

MCa

MC

MC

MC

LCb

LC

IFc

MC

MC

MC

MC

MC

MC

IFLC

S100A3

NBd

��

�+

++

++

++

��

��

�PCRe

+�

�+

++

++

++

++

++

+

S100A4

NB

+�

�+

++

+�

++

++

++

+PCR

++

�+

++

++

++

++

++

+

S100A5

NB

��

��

��

��

��

��

��

�PCR

��

��

��

��

��

��

��

�

aMammarycarcinoma

bLungcarcinoma

cImmortalizedfibroblastsdNorthernblothybridizationanalysiseRT–PCRanalysis


3676

Oncogene

Based on the following arguments, we propose herethat spontaneous tumor development in S100A4�/�mice is a secondary event that emerges because ofdisruption of S100A4–p53 interaction in the S100A4-null mice.Spontaneous tumors in the S100A4�/� mice devel-

oped with a long latency and only in a part of theanimals. This indicates that besides the S100A4inactivation another event, or events could occur inthe cells in order to reveal S100A4 tumor suppressoractivity.Staining of tumors developed in the S100A4-null mice

revealed accumulation of immunohistochemically de-tectable p53 protein in the tumor cells. Under physio-logical circumstances, the wild-type p53 protein is notdetected in the cells by immunohistochemistry. Detec-tion of p53 in the nuclei of tumor cells is regarded as anindicator of functionally altered or mutated p53 (Green-blatt et al., 1994). S100A4-null tumors express elevatedlevels of PCNA and suppressed levels of p21Waf proteins.PCNA is negatively regulated by wild-type p53, whereasthe expression of p21Waf is upregulated by active wild-type p53. Detection of elevated levels of p53 togetherwith upregulation of PCNA and downregulation ofp21Waf expression in the spontaneous tumors of theS100A4-null mice denotes a functional inactivation ofthe p53 protein. Immunochemical detection of p53-positive cells in human tumors is suggested as a markerfor advanced stage of a number of human cancers(Borresen-Dale, 2003; Iacopetta, 2003; Schuijer andBerns, 2003; Staib et al., 2003). In contrast, alteration ofp53 status is a rare event in spontaneous and chemicallyinduced cancers in rodents (Buzard, 1996; Cazorla et al.,1998; Duddy et al., 1999). We can propose that in thespontaneous tumors observed in the S100A4�/� mice,p53 gene is either mutated or functionally altered.Tumor suppressor protein p53�/� homozygous

deletion causes early development of lymphomas inmice (Donehower et al., 1992). The heterozygous p53þ /� animals also developed a wide range of spontaneoustumors with a longer average latency (Jacks et al., 1994;Kuperwasser et al., 2000). Analysis of the tumorspectrum in the p53þ /� mice revealed mostly sarco-mas; lung carcinomas were also among the tumorsgenerated in the p53þ /� mice (Jacks et al., 1994). Thepredominant type of tumor developed in the S100A4�/� mice was adenocarcinoma of the lung. Long latency,together with a wide spectrum of tumors observed in theS100-null animals, reflects the phenotype of p53þ /�animals. It is proposed that another mutation event(possibly inactivation of the wild-type allele of p53) isnecessary for triggering the development of tumors inthe p53þ /� mice (Jacks et al., 1994; Kuperwasser et al.,2000).The functional activity of p53 protein in the

S100A4�/� animals is altered. After whole-body g-irradiation of the mice, the level of apoptosis andtransactivation of p53 target genes in the spleen ofS100A4�/� mice was suppressed. It has been shownthat a number of p53 target genes, including waf/p21/cip1 and bax, are strongly activated in the spleen of

irradiated mice (Bouvard et al., 2000). Activation ofthese genes in the irradiated S100A4�/� mice wassubstantially downregulated. Moreover, the amount ofcells that undergo apoptosis in the spleen of knockoutmice after irradiation was significantly lower than in thecontrols. p53 plays a major role in determining theradiation response in mouse tissues in vivo (Bourdonet al., 1997; Komarova et al., 2000; Fei et al., 2002). P53function helps to maintain the genome integrity and isregarded as a basis of the tumor suppressor function ofp53 protein (Hanahan and Weinberg, 2000).Carcinogenesis is a complex multistep process, invol-

ving a number of events occurring at molecular, cellularand morphological levels. The same molecules mayexhibit, depending on the context, opposing functions.The most striking example of such a diverse behavior isE2F1, which can act as an oncogene and tumorsuppressor gene (Johnson, 2000). Interaction of wild-type p53 protein with S100A4 leads, as it has beenshown, to stimulation of apoptosis in tumor cells(Grigorian et al., 2001). When the S100A4 protein isabsent in the cells of knockout animals, the cellsdamaged by genotoxic stress are not eliminated byapoptosis. This leads to a selection of cells bearingdifferent mutations, and eventually to development ofcancer. In developed cancer, activity of S100A4either in the tumor cells or in the surrounding stromacould activate functions stimulating angiogenesis orinvasion of tumor cells thereby stimulating metastasis(Ambartsumian et al., 2001).The important point is that the spontaneous tumors

developed in the S100A4 knockout animals werenonmetastatic. The only exception was the lymphomadeveloped in one of the animals. This fits with the abilityof S100A4 to stimulate metastasis in epithelial tumors(Ambartsumian et al., 1996).Germ-line inactivation of the S100A4 gene led to

alterations in transcription of one of the neighboringgenes in S100 gene cluster – S100A5. Normally, theS100A5 protein is expressed in restricted areas of wild-type mouse brain (Schafer et al., 2000). Expression ofS100A5 was recently detected in epidermoid lesions andin the human kidney (Teratani et al., 2002; Pelc et al.,2003).Activation of S100A5 might be a compensatory

reaction for S100A4 inactivation in certain cell types.Alternatively, the activation of S100A5 transcriptionmight be a result of deletion of a section of the DNAthat either bears a repressor of S100A5, or physicallybrings the S100A5 gene sequences under the influence ofS100A4 enhancer.The spontaneous tumors developed in S100A4-null

animals may be a result of unknown oncogenic activityof S100A5. Two observations argue against thisproposition.Firstly, S100A5 is not expressed in the spontaneous

tumors developed in the S100A4-null mice. Secondly, inthe panel of mouse tumor cell lines, we were unable todetect S100A5-specific transcripts. In our opinion, theseobservations reduce the probability for S100A5 to act asan oncogene.


3677

Oncogene

Besides p53, S100A4 interacts with a number of targetproteins, such as nonmuscle myosin or liprinb-1.Disruption of the interaction with these target proteinsalso could lead to changes in the physiology of the mice,which are not yet revealed. We cannot exclude also thatthese interactions are compensated by activation ofother members of S100 family proteins. Complex studiesof S100 family proteins are required to determine thephenotypic outcome of inactivation of single member ofthis multigenic family.

Materials and methods

Generation of S100A4-deficient mice

Genomic clones of the S100A4 gene were obtained byscreening of a mouse 129/SvJ genomic library with a mouseS100A4 cDNA fragment (Ebralidze et al., 1989). A clonecarrying a 10 kb BamH1 fragment containing the entireS100A4 gene and 50 and 30 flanking sequences was selectedfor generation of a targeting vector. The entire coding partof the S100A4 gene was excised by PCR using primers:50-AACTCGAGTGCCATGGTAACCGTTGAGAC-30 and50-AACTCGAGAGACTCCTCAGATGAAGTGTT-30. Thisgenerated deletion of the S100A4-coding sequences andintroduced the Xho1 recognition sequence. The tk/neo(Mombaerts et al., 1996) cassette was inserted into thegenerated Xho1 site. The targeting vector was introduced byelectroporation into the embryonic stem cell line R1 (The EScells were kindly provided by Dr A Nagy, Mount SinaiHospital, Toronto) and selected as described (Mansouri,1998). Genomic DNA of the selected colonies was screenedfor homologous recombination by Southern blot hybridizationof EcoR1 digested DNA with a probe 1 (Figure 1a). Theselection cassette was excised from the clones by transienttransfection with PGK-Cre plasmid (kindly provided by KlausRajewsky, Institute for Genetics, University of Cologne,Cologne, Germany). Selected colonies were screened forexcision of the cassette by Southern blot hybridization ofEcoR1 with the probe II (Figure 1a).Clones containing deletion of the S100A4 gene were

aggregated with the eight-cell embryos of NMR1 micefollowing by implantation into pseudo-pregnant females. Highdegree chimeras were mated with NMR1 mice. Germ-linetransmission was identified by Southern blotting of DNAextracted from tail biopsies digested by Sca1. The S100A4�/�mice were maintained by intercrossing of littermates. The wild-type littermates of the þ /� intercrosses were maintained as amatching control group.

Gene expression analysis

S100A4, S100A3 and S100A5 expression was assayed byNorthern blot hybridization and RT–PCR. S100A3- andS100A5-specific probes were obtained by PCR amplificationof genomic sequences corresponding to the exon 2 of S100A3and exon 3 of S100A5. The primers were 50-CAGGTGGGCAGCTCCTTCT-30 and 50-ATGACCCGGCCCCTGGAGCA-30 for S100A3, and 50-GTAGCCTCTGATCCAAAG-30

and 50-GGTGGGTTAGCACATCAA-30 for S100A5.S100A4-specific probe was obtained as in Ebralidze et al.(1989).RNA extraction from mouse organs and Northern blot

hybridization were performed as described (Ambartsumianet al., 1996). The same filter was reprobed with probes specific

to S100A4, S100A3 and S100A5. The above-mentionedprimers were used for RT–PCR analysis of RNA extractedfrom different tumor cell lines. A description of the tumor celllines used could be found in Christensen et al. (1998).Filters containing RNA extracted from mouse organs after

g-irradiation was sequentially hybridized with waf/p21/cip1and bax-specific probes.The amounts of mRNA on the filters were calibrated by

hybridization with [g-32P]ATP- poly(U) probe. Membraneswere scanned using Fujifilm FLA-3000 computing densit-ometer with Image Gauge software.Western blotting and immunoprecipitation were performed

as described (Ambartsumian et al., 2001). Rabbit anti-S100A4antibodies, mouse anti-p21 antibodies (NeoMarkers, USA)and mouse anti-alpha-tubulin antibodies (Sigma, USA) wereused.

Animal handling

Groups of S100A4-null mice (129/SvJxNMR1 background)and genetically matching controls were observed for 12–14months after birth for possible pathological manifestations.For the whole-body irradiation experiments, S100A4�/� mice(A/Sn genetic background) and matching in age A/Sn controlmice were used. Animals were maintained and killed accordingto the FELACA guidelines.

Irradiation of mice

Mice (8 weeks old) were given whole-body irradiation (8Gy)from an HF160 Pantak X-ray unit at 1.77Gy/min (150 kV,15mA). Five animals were used per time point. The animalswere killed at 0 and 6 h postirradiation and their tissues wereanalysed for mRNA expression and in situ apoptosis.

Histological analyses and immunohistochemical staining

Animals killed at the indicated time periods were subjected tomacroscopic examination. Tissues were fixed in 10% bufferedformalin, embedded in paraffin and analysed by lightmicroscopy and immunohistochemistry. Tumors were classi-fied according to the morphological criteria described byTurusov and Mohr (1994).Tumor tissues were stained with p53 (CM-5, Novocastra

Laboratories, Newcastle, UK), PCNA (Dako A/S) and p21(Neo-Markers, AB9) antibodies followed by biotinylated goatanti-rabbit antibodies and amplification of the signal with theTCA-biotin system (NEN Life Sciences, Boston, USA)according to the protocol of the manufacturer.Immunohistochemical staining with anti-S100A5 rabbit

polyclonal antibodies was performed as described (Schaferet al., 2000). The antigen blocking control was used for testingthe specificity of the antibodies.TUNEL assay for in situ apoptosis was performed on

paraffin-embedded tissues of mice using the ApopTag Perox-idase in situ apoptosis detection kit (Intergen, USA). TUNELassay was quantified by counting the amount of apoptoticnuclei in 10 randomly selected fields by a blinded investigator.

Acknowledgements

We thank Ingrid Fossar-Larsen, Birgitte Kaas, DorritLutzh�ft and Sharif Mahsur for excellent technical assistance.We also thank Anne-Marie Hansen for the help in thepreparation of the manuscript. This work was supported bygrants from Danish Cancer Society, Danish Research Counciland Dansk Kraeftforsknings Fond.


3678

Oncogene

References

Al-Haddad S, Zhang Z, Leygue E, Snell L, Huang A, Niu Y,Hiller-Hitchcock T, Hole K, Murphy LC and Watson PH.(1999). Am. J. Pathol., 155, 2057–2066.

Ambartsumian N, Klingelhofer J, Grigorian M, ChristensenC, Kriajevska M, Tulchinsky E, Georgiev G, Berezin V,Bock E, Rygaard J, Cao R, Cao Y and Lukanidin E. (2001).Oncogene, 20, 4685–4695.

Ambartsumian N, Klingelhofer J, Grigorian M, Karlstrom O,Sidenius N, Georgiev G and Lukanidin E. (1998–99). Invas.Metast., 18, 96–104.

Ambartsumian NS, Grigorian MS, Larsen IF, Karlstrom O,Sidenius N, Rygaard J, Georgiev G and Lukanidin E.(1996). Oncogene, 13, 1621–1630.

Armstrong JF, Kaufman MH, Harrison DJ and Clarke AR.(1995). Curr. Biol., 5, 931–936.

Bian J and Sun Y. (1997). Mol. Cell. Biol., 17, 6330–6338.Borresen-Dale AL. (2003). Hum. Mutat., 21, 292–300.Bourdon JC, Deguin-Chambon V, Lelong JC, Dessen P, MayP, Debuire B and May E. (1997). Oncogene, 14, 85–94.

Bouvard V, Zaitchouk T, Vacher M, Duthu A, Canivet M,Choisy-Rossi C, Nieruchalski M and May E. (2000).Oncogene, 9, 649–660.

Buzard GS. (1996). Mutat. Res., 365, 43–58.Cazorla M, Hernandez L, Fernandez PL, Fabra A, PeinadoMA, Dasenbrock C, Tillmann T, Kamino K, Campo E,Kohler M, Morawieltz G, Cardesa A, Tomatis L and MohrU. (1998). Mol. Carcinogen., 21, 251–260.

Chen DS and Barraclough R. (1996). Biochem. Soc. Trans., 24,352S.

Chen H, Fernig DG, Rudland PS, Sparks A, Wilkinson MCand Barraclough R. (2001). Biochem. Biophys. Res. Com-mun., 286, 1212–1217.

Christensen CR, Klingelhofer J, Tarabykina S, Hulgaard EF,Kramerov D and Lukanidin E. (1998). Cancer Res., 58,

1238–1244.Cohn MA, Hjelmso I, Wu LC, Guldberg P, Lukanidin EMand Tulchinsky EM. (2001). Nucleic Acids Res., 29,

3335–3346.Dameron KM, Volpert OV, Tainsky MA and Bouck N.(1994). Science, 265, 1582–1584.

Davies BR, O’Donnell M, Durkan GC, Rudland PS,Barraclough R, Neal DE and Mellon J. (2002). J. Pathol.,196, 292–299.

Davies M, Harris S, Rudland P and Barraclough R. (1995).DNA Cell. Biol., 14, 825–832.

Davies MP, Rudland PS, Robertson L, Parry EW, Jolicoeur Pand Barraclough R. (1996). Oncogene, 13, 1631–1637.

Donato R. (2001). Int. J. Biochem. Cell. Biol., 33, 637–668.Donehower LA, Harvey M, Slagle BL, McArthur MJ,Montgomery Jr CA, Butel JS and Bradley A. (1992). Nature,356, 215–221.

Du XJ, Cole TJ, Tenis N, Gao XM, Kontgen F, Kemp BE andHeierhorst J. (2002). Mol. Cell. Biol., 22, 2821–2829.

Duddy SK, Parker RF, Bleavins MR, Gough AW, Rowse PE,Gorospe S, Dethlof LA and de la Iglesia FA. (1999).Toxicol. Appl. Pharmacol., 156, 106–112.

Ebralidze A, Tulchinsky E, Grigorian M, Afanasyeva A, SeninV, Revazova E and Lukanidin E. (1989). Genes Dev., 3,1086–1093.

El-Deiry WS. (1998). Semin. Cancer Biol., 8, 345–357.Endo H, Takenaga K, Kanno T, Satoh H and Mori S. (2002).

J. Biol. Chem., 277, 26396–26402.Fei P, Bernhard EJ and El-Deiry WS. (2002). Cancer Res., 62,7316–7327.

Feng G, Xu X, Youssef EM and Lotan R. (2001). Cancer Res.,61, 7999–8004.

Ford HL and Zain SB. (1995). Oncogene, 10, 1597–1605.Greenblatt MS, Bennett WP, Hollstein M and Harris CC.(1994). Cancer Res., 54, 4855–4878.

Grigorian M, Andresen S, Tulchinsky E, Kriajevska M,Carlberg C, Kruse C, Cohn M, Ambartsumian N, Chris-tensen A, Selivanova G and Lukanidin E. (2001). J. Biol.Chem., 276, 22699–22708.

Guerreiro Da Silva ID, Hu YF, Russo IH, Ao X, SalicioniAM, Yang X and Russo J. (2000). Int. J. Oncol., 16,

231–240.Hakem R and Mak TW. (2001). Annu. Rev. Genet., 35,209–241.

Hanahan D and Weinberg RA. (2000). Cell, 100, 57–70.Heizmann CW, Fritz G and Schafer BW. (2002). Front.

Biosci., 7, d1356–d1368.Hobbs JA, May R, Tanousis K, McNeill E, Mathies M,Gebhardt C, Henderson R, Robinson MJ and Hogg N.(2003). Mol. Cell. Biol., 23, 2564–2567.

Hsieh H-L, Schafer BW, Cox JA and Heizmann CW. (2002).J. Cell. Sci., 115, 3149–3158.

Iacopetta B. (2003). Hum. Mutat., 21, 271–276.Jacks T, Remington L, Williams BO, Schmitt EM,Halachmi S, Bronson RT and Weinberg RA. (1994). Curr.Biol., 4, 1–7.

Johnson DG. (2000). Mol. Carcinogen., 27, 151–157.Klingelhofer J, Ambartsumian NS and Lukanidin EM. (1997).

Dev. Dyn., 210, 87–95.Komarova EA, Christov K, Faerman AI and Gudkov AV.(2000). Oncogene, 19, 3791–3798.

Kriajevska M, Fischer-Larsen M, Moertz E, Vorm O,Tulchinsky E, Grigorian M, Ambartsumian N and Lukani-din E. (2002). J. Biol. Chem., 277, 5229–5235.

Kriajevska MV, Cardenas MN, Grigorian MS, AmbartsumianNS, Georgiev GP and Lukanidin EM. (1994). J. Biol. Chem.,269, 19679–19682.

Kuperwasser C, Hurlbut GD, Kittrell FS, Dickinson ES,Laucirica R, Medina D, Naber SP and Jerry DJ. (2000). Am.J. Pathol., 157, 2151–2159.

Lukanidin EM and Georgiev GP. (1996). Curr. Top. Micro-biol. Immunol., 213, 171–195.

Mandinova A, Atar D, Schafer BW, Spiess M, Aebi U andHeizmann CW. (1998). J. Cell. Sci., 111, 2043–2054.

Manitz MP, Horst B, Seeliger S, Strey A, Skryabin BV,Gunzer M, Frings W, Schonlau F, Roth J, Sorg C andNacken W. (2003). Mol. Cell. Biol., 23, 1034–1043.

Mansouri A. (1998). Cell Biology: A Laboratory Handbook,2nd edn, Vol. 3. Celis JE (ed). Academic Press: New York,pp. 478–486.

Mashimo T, Watabe M, Hirota S, Hosobe S, Miura K,Tegtmeyer PJ, Rinker-Shaeffer CW and Watabe K. (1998).Proc. Natl. Acad. Sci. USA, 95, 11307–11311.

Mazzucchelli L. (2002). Am. J. Pathol., 160, 7–13.Mercer WE, Shields MT, Lin D, Appella E and Ullrich SJ.(1991). Proc. Natl. Acad. Sci. USA, 88, 1958–1962.

Meuwissen R, Jonkers J and Berns A. (2001). Exp. Cell. Res.,264, 100–110.

Mombaerts P, Wang F, Dulac C, Chao SK, Nemes A,Mendelsohn M, Edmondson J and Axel R. (1996). Cell, 87,675–686.

Novitskaya V, Grigorian M, Kriajevska M, Tarabykina S,Bronstein I, Berezin V, Bock E and Lukanidin E. (2000).J. Biol. Chem., 275, 41278–41286.


3679

Oncogene

Passey RJ, Williams E, Lichanska AM, Wells C, Hu S, GeczyCL, Little MH and Hume DA. (1999). J. Immunol., 163,2209–2216.

Pedrocchi M, Schafer BW, Mueller H, Eppenberger U andHeizmann CW. (1994). Int. J. Cancer, 57, 684–690.

Pelc P, Vanmuylder N, Lefranc F, Heizmann CW, Hassid S,Salmon I, Kiss R, Louryan S and Decaestecker C. (2003).Histopathology, 42, 387–394.

Platt-Higgins AM, Renshaw CA, West CR, Winstanley JH,De Silva Rudland S, Barraclough R and Rudland PS.(2000). Int. J. Cancer, 89, 198–208.

Ridinger K, Ilg EC, Niggli FK, Heizmann CW and SchaferBW. (1998). Biochim. Biophys. Acta, 1448, 254–263.

Rosty C, Ueki T, Argani P, Jansen M, Yeo CJ, Cameron JL,Hruban RH and Goggins M. (2002). Am. J. Pathol., 160,45–50.

Sah VP, Attardi LD, Mulligan GJ, Williams BO, Bronson RTand Jacks T. (1995). Nat. Genet., 10, 175–180.

Sauer B. (1993). Methods Enzymol., 225, 890–900.Schafer BW, Fritschy JM, Murmann P, Troxler H, Durussel I,Heizmann CW and Cox JA. (2000). J. Biol. Chem., 275,30623–30630.

Schuijer M and Berns EM. (2003). Hum. Mutat., 21,

285–291.

Staib F, Hussain SP, Hofseth LJ, Wang XW and Harris CC.(2003). Hum. Mutat., 21, 201–216.

Sun Y, Cheung JM, Martel-Pelletier J, Pelletier JP, Wenger L,Altman RD, Howell DS and Cheung HS. (2000). J. Biol.Chem., 275, 11327–11332.

Sun Y, Sun Y, Wenger L, Rutter JL, Brinckerhoff CE andCheung HS. (1999). J. Biol. Chem., 274, 11535–11540.

Teratani T, Watanabe T, Yamahara K, Kumagai H, IshikawaA, Arai K and Nozawa R. (2002). Biochem. Biophys. Res.Commun., 291, 623–627.

Turusov V and Mohr U (eds). (1994). Pathology of Tumors inLaboratory Animals, Vol.2—Tumors of the Mouse, 2nd edn.IARS Scientific Publications: Lyon.

Vousden KH. (2002). Biochem. Biophys. Acta, 1602, 47–59.Weterman MA, van Muijen GN, Bloemers HP and Ruiter DJ.(1993). Cancer Res., 53, 6061–6066.

Wu GS, Saftig P, Peters C and El-Deiry WS. (1998). Oncogene,16, 2177–2183.

Xiong Z, O’Hanlon D, Becker LE, Roder J, MacDonald JFand Marks A. (2000). Exp. Cell. Res., 257, 281–289.

Yonemura Y, Endou Y, Kimura K, Fushida S, Bandou E,Taniguchi K, Kinoshita K, Ninomiya I, Sugiyama K,Heizmann CW, Schafer BW and Sasaki T. (2000). Clin.Cancer Res., 6, 4234–4242.


3680

Oncogene

Cancer predisposition in mice deficient for the metastasis-associated Mts1(S100A4) gene

Documents