of the TN-C-gene are born live with no apparent defects (Saga et al., 1992). .... (Amersham Int. plc, Buckinghamshire, England) and were incubated for various ...
2069
Journal of Cell Science 109, 2069-2077 (1996) Printed in Great Britain © The Company of Biologists Limited 1996 JCS1212
Tenascin-X expression in tumor cells and fibroblasts: glucocorticoids as negative regulators in fibroblasts Takao Sakai1,*, Yusuke Furukawa1, Ruth Chiquet-Ehrismann2, Mitsuru Nakamura1, Seiichi Kitagawa1, Toshimichi Ikemura3 and Ken-ichi Matsumoto3 1Division of Hemopoiesis, Institute of Hematology, Jichi Medical School, 3311-1, Minamikawachi, Tochigi 2Friedrich Miescher Institute, PO Box 2543, 4002 Basel, Switzerland 3Department of Evolutionary Genetics, National Institute of Genetics, Mishima, Shizuoka 411, Japan
329-04, Japan
*Author for correspondence at present address: Department of Medicine and Biomolecular Chemistry, University of Wisconsin-Madison, 4468 Medical Science Center, 1300 University Avenue, Madison, WI 53706, USA
SUMMARY Tenascin-X has recently been shown to be a novel member of the tenascin family and its distribution is often reciprocal to that of tenascin-C in the developing mouse embryo. We have investigated the expression of tenascin-X in fibroblasts and carcinoma cells in culture. Tenascin-X protein was secreted in vitro in the conditioned media at an apparent molecular mass of ~450 kDa. In addition fibroblasts contained a major tenascin-X isoform of 220 kDa. On northern blots, a single major transcript with a size of ~13 kb was detected. No overexpression of tenascinX protein was found in primary fibroblasts of the tenascinC-gene knockout mice. Steroid hormone glucocorticoids, were found to downregulate tenascin-X mRNA levels and protein synthesis in fibroblasts but not carcinoma cells at physiological concentrations. None of the growth factors or
cytokines examined affected the expression level of tenascin-X. As in vivo study, carcinoma cells were transplanted into nude mice. In contrast to the ubiquitous presence of tenascin-X in adult skin, expression of tenascinX protein during tumorigenesis was found to be downregulated considerably not only in tumor cells themselves but also in tumor stroma. These findings provide evidence that the expression of tenascin-X can be influenced by stromal-epithelial interactions. We have identified glucocorticoids as physiological inhibitors of tenascin-X and suggest that glucocorticoids may in part participate in the downregulation of tenascin-X in fibroblasts in vivo.
INTRODUCTION
TN-X was initially identified as an adrenal transcript overlapping the human 21-hydroxylase B gene (CYP21) in the class III region of the human major histocompatibility complex locus on the opposite strand (Morel et al., 1989). Extensive analysis revealed that human TN-X was composed of an NH2-terminal domain, followed by four heptad repeats, 18.5 TN-type epidermal growth factor (EGF)-like repeats, at least 29 fibronectin type III domains and a carboxy-terminal fibrinogen domain (Matsumoto et al., 1992a,b; Bristow et al., 1993). We have recently demonstrated that TN-X is conserved in the mouse, and that the tissue distribution of TN-X during organogenesis is distinct and often reciprocal to that of TN-C (Matsumoto et al., 1994). TN-X is shown to be expressed in peripheral nerves (Geffrotin et al., 1995), and detailed observation in heart and muscle suggests that TN-X plays a role in morphogenesis and cell migration in connective tissues (Burch et al., 1995). However, the role of TN-X in cancer tissues is as yet unknown; the tissue distribution of TNX during tumorigenesis has not been analyzed. It has been demonstrated that homozygous null mutant mice of the TN-C-gene are born live with no apparent defects (Saga et al., 1992). It has been argued that other TN-like molecules could make up for the loss of TN-C in these knockout mice (Bristow et al., 1993; Erickson, 1993). We therefore decided
The tenascins (TNs) are a growing family of extracellular matrix (ECM) glycoproteins. To date three members of the family have been identified in mammals: tenascin-C (TN-C) (also called cytotactin), tenascin-R (TN-R) (for restrictin), and tenascin-X (TN-X) (for reviews, see Erickson, 1993; Chiquet-Ehrismann, 1995). The first member, termed TN-C (Chiquet-Ehrismann et al., 1986), was shown to be associated with the development of various organs and tumors (for reviews, see Erickson and Bourdon, 1989; Chiquet-Ehrismann, 1990). In particular, a large number of studies demonstrated that TN-C was expressed in a temporally and spatially restricted pattern during embryogenesis and carcinogenesis in association with stromal-epithelial interactions. In many adult tissues that normally lack TN-C, induction of expression occurs during neoplastic changes. The difference between an in vitro and an in vivo environment can influence the expression of TN-C, and some human carcinoma cell lines which do not normally synthesize TN-C in vitro are induced to synthesize TNC after transplantation into nude mice (Sakai et al., 1993b, 1994). These specific distribution patterns sugggest that TN-C expression can be stimulated by local autocrine or paracrine growth factors released from proliferating cells during neoplasia.
Key words: Tenascin-X, Tenascin family, Glucocorticoid, TenascinC-knockout mice, Stromal-epithelial interaction
2070 T. Sakai and others to investigate the expression of TN-X protein in cells of the TN-C-gene knockout mice. Jones et al. (1990) have identified an array of putative cisacting regulatory elements flanking the chicken TN-C gene, including sequences that confer responsiveness to soluble growth and differentiation factors. The induction of TN-C in vitro by various growth factors and cytokines is tightly regulated in a cell-type-specific manner (Rettig et al., 1989, 1994; Yavin et al., 1991; McCachren and Lightner, 1992; Sharifi et al., 1992; Meiners et al., 1993; Tucker et al., 1993; Lafleur et al., 1994; Sakai et al., 1994, 1995b,c). On the other hand, glucocorticoids could act as potent inhibitors of TN-C production in both fibroblasts and carcinoma cells (Ekblom et al., 1993; Sakai et al., 1995a,c; Talts et al., 1995). It is thus likely that the level of TNC expression in vivo could partly be determined by an interplay between extrinsic positive and negative regulating factors. However, the physiological factors that participate in the regulation of TN-X expression are still unknown. Since very little is known about the cellular source of TN-X protein and its possible isoforms, we decided to investigate the expression of TN-X in both in vitro and in vivo studies. In this paper, we investigated the expression of TN-X in vitro in both carcinoma- and fibroblast-derived cells in order to clarify the cellular source of TN-X. We propose a mechanism by which the expression level of TN-X protein could be altered in response to soluble factors in vitro. We focused on alterations in TN-X production when the cellular environment was changed from in vitro to in vivo. As an in vivo study, carcinoma cells were transplanted into nude mice and the tissue distribution of TN-X during tumorigenesis was studied by immunohistochemistry and compared with that of TN-C. The results of these studies allow us to propose a mechanism by which stromal fibroblasts downregulate TN-X expression in vivo in the vicinity of carcinoma cells transplanted into nude mice. MATERIALS AND METHODS Antibodies The following primary antibody preparations were used: (1) polyclonal chick antibody (pAb800) to mouse TN-X. Antibody pAb800 (raised against recombinant tenascin-X fusion protein GST-GE800) was affinity purified using β-gal-GE800 coupled to CNBr-activated Sepharose 4B. We confirmed that pAb800 was absorbed by this fusion protein and specific for tenascin-X (Matsumoto et al., 1994). (2) Polyclonal antibody (TN-3) (Sakai et al., 1993b) to TN-C purified from human fibroblast line HUCF-p2. (3) Polyclonal antibody to TN-C purified from human glioma line U-251 (Telios Pharm. Inc., San Diego, CA). (4) Monoclonal rat antibody to mouse TN-C (Sigma Chem. Co., St Louis, MO). (5) Polyclonal antibodies to mouse fibronectin and to rat fibronectin (Telios Pharm. Inc.). (6) Polyclonal antibody to rat laminin (Telios Pharm. Inc.). Growth factors, cytokines and analytical reagents Transforming growth factor (TGF)-β1, basic fibroblast growth factor (bFGF), and platelet-derived growth factor (PDGF) were obtained from R&D Systems (Minneapolis, MN). Recombinant TGF-α was from Gibco BRL (Gaithersburg, MD), hepatocyte growth factor (HGF) was from Collaborative Biomedical Products (Bedford, MA), EGF was from Takara Shuzo Co., Ltd (Kyoto, Japan). Recombinant tumor necrosis factor-α (TNF-α) was from Upstate Biotechnology Inc. (Lake Placid, NY), recombinant leukemia inhibitory factor (LIF) was from Pepro Tech Inc. (Rocky Hill, NJ), recombinant interleukin
1 receptor antagonist (IL-1ra) was from R&D Systems, and interferon-α (IF-α) was from Sumitomo Pharma., Co., Ltd (Osaka, Japan). Agarose-immobilized fibronectin was generously provided by Dr S. Asakura (Division of Hemostasis and Thrombosis Research, Jichi Medical School, Tochigi, Japan). Hydrocortisone (Sigma Chemical Co., Ltd) was dissolved in ethanol and we confirmed that ethanol at the final concentration (less than 0.003%) had no toxic effect. Fetal bovine serum (FBS) (Gibco Lab., Grand Island, NY) was dialyzed against distilled and deionized water before experimental use. Cells The mouse cell lines, 203Glioma and Ren-Ca (renal carcinoma), were generously provided by Dr H. Yagita (Department of Immunology, Juntendou University School of Medicine, Tokyo, Japan). The rat normal (immortalized) fibroblast cell line NRK and its v-src transfectants (transformed) line 77-NI, were kindly provided by Dr H. Kawakatsu (Biotechnology Laboratories, Nippon Shinyaku Co. Ltd, Kyoto, Japan). Primary cultures of fetal-skin, splenic and thymic fibroblasts from C3H/HeJ mouse 16 day embryos and from 16 day embryos of mice homozygous for a null mutation in the TN-C gene (TN-C-gene-knockout mice) were prepared as described elsewhere (Sakai et al., 1994, 1995b). Metabolic labeling and immunoprecipitation The ECM proteins were immunoprecipitated from metabolically labeled conditioned media essentially as described by Sakai et al. (1993b). The cells received L-[35S]methionine (100 µCi/ml medium) (Amersham Int. plc, Buckinghamshire, England) and were incubated for various time periods. After incubation the conditioned media were collected, then the same volumes or the same amounts of cpm were immunoprecipitated with each antibody (the amount of antibodies used was sufficient to precipitate all of the available proteins in each separate aliquot of the metabolically labeled conditioned medium). For the immunoprecipitation of TN-X, affinity-purified rabbit antichick IgG (Organon Teknika Corp., Durham, NC) was added as secondary antibody. Then Protein A-Sepharose 4 Fast-Flow suspensions (Pharmacia LKB Biotech., Sweden) were added to each sample. The proteins were eluted from the resins by incubation with SDS sample buffer (Laemmli, 1970). Samples were then subjected to SDSPAGE on 4-20% or 2-15% (w/v) gradient gels, followed by fluorography. In order to quantitate the radiolabeled bands, the radioactivities of TN and fibronectin bands were determined by means of a BAS 2000 Bio-Imaging Analyzer (Fuji Photo Film Co., Ltd, Tokyo, Japan). Western blotting Immunoblotting analysis was performed as previously described elsewhere (Sakai et al., 1993a), with a slight modification. Briefly, reduced samples were electrophoresed on SDS-PAGE gradient gels and transferred onto Immobilon-P membranes (Millipore Corp., Bedford, MA). The membrane was incubated with 5% (w/v) skim milk (Difco Lab., Detroit, MI) in Tris-buffered saline (TBS) in order to block nonspecific protein binding, and incubated with the primary antibody, then with an alkaline phosphatase-conjugated secondary antibody (Zymed Lab. Inc., San Fransisco, CA). The color reaction was developed with nitro blue tetrazolium and 5-bromo-4-chloro-3indolyl phosphate (Promega, Madison, WI). RNA isolation and northern blotting RNA isolation and northern blotting were performed essentially as described previously (Sakai et al., 1994). Total RNA (15 µg) was electrophoresed in a 1.0% (w/v) agarose/1% (v/v) formaldehyde-denaturing gel and transferred to Hybond-N+ membrane (Amersham Int. plc). Two different regions of mouse TN-X DNA were used as probes: one was obtained from the cDNA encoding the fibrinogen-like domain (about 0.6 kb, pG40) (Matsumoto et al., 1994), and the other consisted of the EGF-like domain of the genomic TN-X DNA (about 1.0 kb). The genomic DNA library was made from TT2 cells (Gibco BRL).
Downregulation of tenascin-X by glucocorticoids 2071 cDNA and genomic DNA screenings were performed using standard protocols (Matsumoto et al., 1992a), and both regions were confirmed by sequencing. The mouse TN-C cDNA probe extended from the 5′ end to the FN2 constitutive repeat (about 2.5 kb) (Saga et al., 1991). For glyceraldehyde-3-phosphate dehydrogenase (GAPDH), a 0.6 kb cDNA fragment corresponding to nucleotides 146-743 (Tso et al., 1985) was generated by RT-PCR, purified and used as the probe. Probes were labeled with random primers by DNA polymerase Klenow fragment (Amersham Int. plc). After hybridization, the membrane was washed in 2× standard saline citrate (SSC) containing 0.1% (w/v) SDS, then in 1× SSC containing 0.1% (w/v) SDS at 65°C (high-stringency wash), and subjected to autoradiography. In order to quantitate the radiolabeled bands, the radioactivities of TN-X and GAPDH bands were determined by means of a BAS 2000. Fig. 1. Specificity of polyclonal antibody (pAb800) against mouse TN-X in combined immunoprecipitation/western blotting analysis. Conditioned media of mouse-derived cell line Ren-Ca (a) and 203Glioma (b) were immunoprecipitated with pAb800 (lane 1) or TN-C-specific antibody TN-3 (lane 2), then immune complexes were subjected to SDS-PAGE for western blotting by pAb800 (lanes 1 and 2). The positions of molecular mass markers (kDa) are indicated to the left. The low molecular mass protein (~68 kDa) in lane 1 corresponds to the heavy chains of the precipitating antibody. Major TN-X isoforms are indicated to the right (arrows).
Transplantation of the cell lines into nude mice Cells (1.0×107) were transplanted into male BALB/c nude mice (CLEA Japan Inc., Tokyo, Japan) by dorsal subcutaneous injection. One and two weeks after transplantation, the mice were killed and the tumorigenicity was investigated. Xenografts of the cell lines were used for histopathological examinations, and parts of them were directly embedded in Tissue-Tek OCT compound (Miles Inc., Elkhart, IN), then immediately frozen, and prepared for cryosectioning.
Fig. 2. (A) Secretion of TN-X and TN-C in fibroblasts. Fluorographs of SDS-PAGE of immunoprecipitates from the conditioned media of metabolically labeled primary embryonic skin fibroblasts from C3H mice (a) and TN-C-geneknockout mice (b), primary embryonic thymic (c) and splenic (d) fibroblasts from TN-C-gene-knockout mice, and rat-derived normal fibroblast cell line NRK (e) and its transformed line 77-NI (f). Proteins were precipitated by polyclonal antibody (TN-3) to TN-C (lane 3), monoclonal antibody to TN-C (lane 4), or polyclonal antibody (pAb800) to TN-X (lane 5). Controls were precipitated by normal rabbit IgG (lane 1) or normal chick IgG (lane 2). (B) Secretion of TNX, TN-C, and other ECM glycoproteins fibronectin and laminin in caricnoma cells: fluorographs of SDSPAGE of immunoprecipitates from the conditioned media of metabolically labeled Ren-Ca (a) and 203Glioma (b) cells. Proteins were precipitated by polyclonal antibody TN-3 (lane 3), polyclonal antibody to human TN-C (lane 4), polyclonal antibody pAb800 (lane 5), polyclonal antibody to mouse fibronectin (lane 6), or polyclonal antibody to rat laminin (lane 7). Controls were precipitated by normal rabbit IgG (lane 1) or normal chick IgG (lane 2). The positions of molecular mass markers (kDa) are indicated to the left. TN-C glycoprotein is recognized by both polyclonal and monoclonal antibodies in rat-derived fibroblast lines NRK and 77NI. No overexpression of TN-X is found in primary fibroblasts of TN-C-gene-knockout mice.
2072 T. Sakai and others Immunofluorescence microscopy Cryostat sections were fixed in cold acetone (unfixed sections were dried and also used for the analysis). After washing with TBS, they were incubated in TBS containing 1% (w/v) BSA to block nonspecific protein binding. Slides were incubated with polyclonal antiTN-X (pAb800) or monoclonal anti-TN-C antibody (normal chick IgG or rat IgG was used as a negative-control primary antibody) for 2 hours at room temperature, then treated with fluorescein-labeled, affinity-purified rabbit anti-chick IgG (Chemicon International Inc., Temecula, CA) or fluorescein-labeled, affinity-purified goat anti-rat IgG (MBL, Nagoya, Japan). Slides were briefly washed with TBS and mounted with glass coverslips, then viewed with a BX-60 fluorescence microscope (Olympus Optical Co. Ltd, Tokyo, Japan).
RESULTS Steady-state production level of TN-X, TN-C and other ECM proteins in vitro in fibroblast and carcinoma-derived cells To study the cross-reactivity of TN-X antibody pAb800 we performed a combined immunoprecipitation/western blotting analysis (Fig. 1). TN-X protein was found to be expressed in RenCa carcinoma and 203Glioma cells and the major molecular mass isoforms gave bands (~450 kDa, ~400 kDa and ~220 kDa) under reducing conditions. We concluded that the lower 220 kDa isoform was TN-X and not TN-C, because immunoprecipitates with TN-C antibody were not recognized by the TN-X antibody (Fig. 1, lane 2). Although the molecular mass of 220 kDa is similar to that of fibronectin, we have excluded the possibility that the 220 kDa band represents fibronectin. There was no change in the amount and immunoreactivity of this band when TN-X antibody pAb800 was pretreated with agarose-immobilized fibronectin (data not shown). These results demonstrate that pAb800 has not cross-reacted with TN-C and fibronectin. We next carried out immunoprecipitation of metabolically labeled conditioned medium in fibroblasts and carcinoma cells (Fig. 2). The highest molecular mass isoform of TN-X gave a single-band of apparent molecular mass 450 kDa in Ren-Ca carcinoma and a doublet in fibroblasts under reducing conditions. In addition fibroblasts contained a major TN-X isoform of 220 kDa. In 203Glioma, TN-X gave a band of 400 kDa which was a little lower than that of other cells. The major isoform of TN-C gave a band at an apparent molecular mass of 220 kDa in Ren-Ca and fibroblasts, or 180 kDa in 203Glioma cells. In order to investigate whether TN-X protein was overexpressed in fibroblasts of TN-C-gene-knockout mice in vitro, we compared its expression in primary embryonic fibroblasts from C3H mice and those from mice homozygous for a null mutation in the TN-C gene (Fig. 2A, a-d). No overexpression of TN-X was found in primary fibroblasts of TN-C-gene-knockout mice in all cases examined; however, there was a slight difference in the expression levels or isoform patterns. The secretion level of TNX protein was rather low when compared with that of TN-C. The rat-derived normal (immortalized) fibroblast cell line NRK and its transformed counterpart 77-NI, secreted a rather large amount of TN-X and there was no remarkable change in the expression level between them (Fig. 2A, e and f). In contrast to a previous report showing that only the higher molecular mass isoform of TN-C was detectable in transformed cell lines (Carnemolla et al., 1992), no change in the isoform pattern of TN-C was observed
between them. TN-C was recognized by both polyclonal and monoclonal antibodies. These fibroblasts secreted also fibronectin and laminin (data not shown). Murine carcinoma cell lines, Ren-Ca and 203Glioma, produced TN-X, TN-C, and other ECM proteins fibronectin or laminin (Fig. 2B). Characteristically, Ren-Ca cells showed no fibronectin synthesis. Expression pattern of TN-X mRNA compared with that of TN-C mRNA Northern blot analysis was performed to study the expression and size of TN-X (Fig. 3). We used two different probes of mouse TN-X, pG40 from the fibrinogen-like domain (Matsumoto et al., 1994) and part of the EGF-like domain. No cross-hybridization among members of the TN family was detected under the conditions examined. For comparison, the expression pattern of TNC transcripts was examined (Fig. 3C). A major transcript was found to have a size of 7 kb, which was described previously
Fig. 3. Expression of TN-X transcripts in comparison to TN-C. Total RNAs from rat-derived fibroblast cell line NRK (lane 1) and 77-NI (lane 2), mouse-derived cell line 2H6GR (lane 3), Ren-Ca (lane 4), 203Glioma (lane 5), and primary murine embryonic fibroblasts from skin (lane 6), spleen (lane 7) and thymus (lane 8), were subjected to RNA blotting. The blot was hybridized with TN-X probes from part of the EGF-like domain (A) or pG40 (B). After removal of the TN-X probe, the same membranes were rehybridized with the probe generated from murine TN-C cDNA (one of them is shown) (C). As a loading control, 28 S and 18 S rRNAs were used after ethidium bromide staining (D). The positions of 28 S and 18 S rRNA are indicated to the left. Cross-hybridization of the TN-X probes with rRNAs may be due to the G+C% contents of these probes, as described previously (Matsumoto et al., 1994).
Downregulation of tenascin-X by glucocorticoids 2073
Relative Activity (% of Control)
A. TN-X
B. TN-C
100
100
80
80
60
60
40
40
20
20
0
0 Control
Fig. 4. Reduction of tenascin-X protein biosynthesis in NRK fibroblasts by hydrocortisone: fluorographs of SDS-PAGE of immunoprecipitates from the conditioned media of metabolically labeled cells. Proteins were precipitated by TN-X antibody pAb800 (A) or anti-rat fibronectin antibody (B). The cells (105) were incubated for 24 hours in the absence of hydrocortisone (lane 2) or presence of 10 nM (lane 3), 100 nM (lane 4) or 1 µM (lane 5) hydrocortisone. Then they were replaced with the labeling medium containing added factors. Following a 12 hour labeling period, the culture media were subjected to the immunoprecipitation (samples containing the same amounts of cpm were used for analysis). Each lane 1 is a control (no addition) precipitated by normal chick IgG and normal rabbit IgG, respectively. The radioactivity of TN-X bands was determined by means of a BAS 2000 Bio-Imaging Analyzer and the control culture (no added factors) was set to 100%. The radioactivity of TN-X determined under the different culture conditions is given as percentage inhibition relative to the control value of 100% as follows: lane 3, 48%; lane 4, 53%; lane 5, 64%. The positions of molecular mass markers (kDa) are indicated to the left.
(Matsumoto et al., 1994; Sakai et al., 1995c). In contrast to the TN-X protein isoforms, a single major transcript of TN-X with a size of 13 kb in mouse and rat cell cultures was found (Fig. 3A,B). Primary murine embryonic fibroblasts, which produced fairly small amounts of TN-X protein, expressed very low levels of TN-X mRNA which was hardly detectable in all cases (Fig. 3A,B, lanes 6-8). There was no remarkable change in the expression patterns and levels of TN-X mRNA between two different probes. Rat fibroblast lines NRK and 77-NI, which showed TN-C-protein secretion with two different antibodies (Fig. 2A, c and d), expressed very low levels of TN-C mRNA which was undetectable (Fig. 3C, lanes 1 and 2). Interestingly, 203Glioma cells, whose major molecular mass isoforms of TNX and TN-C proteins (400 kDa and 180 kDa, respectively) were found to be a little lower than that of other cells, expressed lower mRNA isoforms with a size of 12 kb for TN-X and 5.5 kb for TN-C, respectively (Fig. 3A-C, lane 5). This implies alternative splicing of TN-X transcripts in 203Glioma cells. Downregulation of tenacin-X expression in fibroblasts by glucocorticoids in vitro We added steroid hormones or various cytokines and tested
HC
Control
HC
Fig. 5. Effect of glucocorticoids on the synthesis and secretion of TN-X (A) and TN-C (B) protein in NRK fibroblasts (h) and Ren-Ca carcinoma cells (j). The cells (105) were incubated for 24 hours in the absence or presence of 1 µM hydrocortisone (HC). After a 24 hour incubation, they were labeled, then the culture media were subjected to the immunoprecipitation by TN-X antibody pAb800 or TN-C antibody TN-3 (samples containing the same amounts of cpm were used for analysis). In order to quantitate the effects of hydrocortisone, the radioactivities of TN-X and TN-C bands were determined by means of a BAS 2000 and each control culture (no addition) was set to 100%. The radioactivity determined in the presence of hydrocortisone is shown relative to the control value 100%. Each value represents the mean ± s.e. of at least two independent experiments.
whether these factors could alter the synthesis and secretion of TN-X in NRK fibroblasts and Ren-Ca carcinoma cells. TN-X protein synthesis was significantly reduced in NRK fibroblasts with the addition of hydrocortisone (Figs 4, 5). The concentration analysis showed that hydrocortisone at the highest concentration used gave about 65% inhibition of TN-X protein synthesis (Fig. 4A). No reduction of the synthesis and secretion of fibronectin was noted (Fig. 4B). Hydrocortisone at the concentrations used had little effect on NRK cell proliferation (data not shown). In a time-course study performed with NRK
Fig. 6. Time-course of the effect of hydrocortisone on steady-state levels of TN-X mRNA. NRK fibroblasts were incubated with 1 µM hydrocortisone for 1 hour (lane 2), 3 hours (lane 3), 6 hours (lane 4), 12 hours (lane 5), 24 hours (lane 6) and 48 hours (lane 7). Lane 1 is a control (no addition). The blot was hybridized with TN-X probe pG40 (A). As a control, the blot was rehybridized with the GAPDH probe (B). The radioactivities of TN-X bands were determined by means of a BAS 2000 and they were normalized to GAPDH mRNA (1.4 kb) signal intensities. Each radioactivity of the TN-X bands is given as percentage inhibition relative to the control value of 100% as follows: lane 2, 15%; lane 3, 37%; lane 4, 48%; lane 5, 69%; lane 6, 72%; lane 7, 81%.
2074 T. Sakai and others Relative TN-X Activity (% of Control)
200
150
100
50
0 Control
10%FBS
EGF
TGF-α
TGF-β1
HGF
PDGF
bFGF
Growth Factors
Fig. 7. Effect of growth factors on the synthesis and secretion of TNX protein in NRK fibroblasts (h) and Ren-Ca carcinoma cells (j). The cells (105) were incubated for 24 hours with the medium containing 0.5% FBS in the absence or presence of growth factors (20 ng/ml). They were then replaced with the labeling medium containing added factors. Following a 12-hour labeling period, the culture media were subjected to the immunoprecipitation by TN-X antibody pAb800 (samples containing the same amounts of cpm were used for analysis). In order to quantitate the effects of various factors on the secretion of TN-X, the radioactivity of TN-X bands was determined by means of a BAS 2000 and the culture containing 0.5% FBS (as a control culture) was set to 100%. The radioactivity determined under the different culture conditions is shown relative to the control value of 100%. Each value represents the mean ± s.e. of two independent experiments.
fibroblasts, we confirmed that hydrocortisone inhibited the steady-state levels of TN-X mRNA by almost 80% after a 48 hour treatment (Fig. 6). Hydrocortisone also downregulated the de novo synthesis and secretion of TN-C in both NRK fibroblasts and Ren-Ca carcinoma cells (Fig. 5). Other cytokines
such as TNF-α, LIF, IF-α, and IL-1ra hardly affected the expression level of TN-X in either cell type (data not shown). In the case of TN-C, the possibility of culturing cells (such as primary chick embryo fibroblasts) in low serum permitted a study of the alteration of TN-C synthesis (Pearson et al., 1988). Therefore, we investigated the effect of growth factors on TNX synthesis in NRK fibroblasts and Ren-Ca carcinoma cells in serum-starved culture conditions (0.5% FBS). When FBS was added to the serum-sterved media of these cells, as expected, the synthesis and secretion of TN-X were stimulated: a 1.6fold induction in NRK cells and a 1.3-fold induction in RenCa cells were demonstrated. However, these induction levels of TN-X were rather small and did not differ significantly (P>0.01). In addition, growth factors such as TGF-β1, EGF, TGF-α, PDGF or bFGF, which have already been shown to alter the expression level of TN-C, did not induce significant changes in the expression level of TN-X (Fig. 7). Downregulation of TN-X protein expression in both carcinoma cell lines transplanted into nude mice and host-mouse stroma The distributions of TN-X and TN-C in adult skin were shown to be distinct and detected in a discontinuous distribution: TNX was abundant throughout the stroma of dermis and epidermis, whereas TN-C was most intensely stained in the upper dermis adjacent to the basal lamina (at the area of dermal-epidermal junction) and around the hair follicles (Fig. 8). These distribution patterns were nearly reciprocal like those of murine embryos, as described previously (Matsumoto et al., 1994). Both Ren-Ca and 203Glioma cells developed tumors in nude mice. TN-C was distributed with a restricted pattern in the hostmouse fibrous stroma: it was the most prominent adjacent to the tumor nest composed of transplanted tumor cells (Fig. 9C,F). Ren-Ca cells themselves became completely negative for TN-C
Fig. 8. Distribution of TN-X and TN-C in murine skin. Serial sections of adult skin were stained with hematoxylin and eosin (A), TN-Xspecific antibody pAb800 (B) or monoclonal anti-TN-C antibody (C). The distributions of TN-X and TN-C are distinct and show the reciprocal staining pattern: TN-X is abundant throughout the stroma of dermis and epidermis, whereas TN-C is most intensely stained at the dermalepidermal junction (stars) and around the hair follicles. d, dermis; e, epidermis; h, hair follicle. Bar in C, 85 µm. To demonstrate the expression of TN-X in the stroma, adult skin stained with pAb800 in higher magnification is shown (D). Bar in D, 30 µm.
Downregulation of tenascin-X by glucocorticoids 2075
Fig. 9. Immunoreactive localization of TN-X and TN-C in the xenografts of murine carcinoma cell line Ren-Ca (A-C) and 203Glioma (D-F) in nude mice (low magnification). Serial sections of the xenografts were stained with hematoxylin and eosin (A and D), TN-X antibody pAb800 (B and E) or anti-TN-C antibody (C and F). TN-C is prominent in the host-mouse fibrous stroma with a typical staining pattern outlining the demarcation of the tumor mass composed of transplanted tumor cells in contrast to almost completely negative immunoreactivity for TN-X. h, hair follicle; t, tumor nest. Bars: 170 µm (A-C); 85 µm (D-F).
expression in vivo and only host-mouse-derived TN-C was prominently induced in the fibrous stroma adjacent to them (Fig. 10C) (Talts et al., 1993). In contrast to the distinct and restricted expression pattern of TN-C in the ECM, expresion of TN-X was downregulated considerably not only in the transplanted tumor cells themselves but also in the host-mouse fibrous stroma. No distinct expression pattern of TN-X was demonstrated (Figs 9B,E, and 10B). In the case of 203Glioma cells, some of the transplanted tumor cells sometimes showed positive expression for TN-C in addition to inducing host-mouse-derived stromal TN-C (Fig. 10F). Furthermore, a weak positive expression of TN-X was sometimes found in the cytoplasms of 203Glioma
cells (Fig. 10E). There was no staining in the control slides in which normal chick IgG or normal rat IgG was substituted for the primary antibody. We also analyzed paraformaldehyde-fixed paraffin-embedded tissue sections of the murine embryo and normal adult skin for immunohistochemistry. However, no distinct localization of TN-X was demonstrated in any cases whether or not sections were pretreated with trypsin, pepsin or hyaluronidase (data not shown). DISCUSSION This paper represents the first study of possible physiological
Fig. 10. Immunoreactive localization of TN-X and TN-C in the xenografts of murine carcinoma cell line Ren-Ca (A-C) and 203Glioma (D-F) in nude mice (high magnification in the tumor nests). Sections of the xenografts were stained with hematoxylin and eosin (A and D), TN-X antibody pAb800 (B and E) or anti-TN-C antibody (C and F). No distinct expression pattern of TN-X is noted. Bar, 30 µm.
2076 T. Sakai and others regulators of TN-X expression. We have clearly demonstrated that glucocorticoids downregulate TN-X mRNA levels and protein synthesis at physiological concentrations. TN-X protein was found to be secreted in vitro in the conditioned media of fibroblasts and carcinoma cells, indicating that these cells could be the cellular souce of TN-X in vivo. We used both fibroblasts and carcinoma cells, since it was important to test whether the expression of TN-X by fibroblasts and carcinoma cells could be regulated by the same or different mechanisms. A large number of studies have shown that certain cytokines can modulate the compositions of ECM by stimulating protease activity or inhibiting ECM synthesis of the cells (Le and Vilcek, 1987; Herlyn and Malkowicz, 1991). In fact, several cytokines are produced and present in inflammatory and cancer tissues. There is some evidence that glucocorticoid hormones downregulate steady-state levels of TN-C glycoprotein and its mRNA in both carcinoma cells and fibroblasts (Ekblom et al., 1993; Sakai et al., 1995a,c; Talts et al., 1995). The recent observation suggested that glucocorticoids could act directly through cis regulatory elements of the TN-C gene (Ekblom et al., 1993). We found that the treatment of NRK fibroblasts with glucocorticoids downregulated the steady-state levels of TN-X mRNA and the de novo synthesis of TN-X protein. We have now identified glucocorticoids as a negative physiological regulator of TN-X expression in fibroblasts. The downregulation of TN-X and TNC caused by glucocorticoids apparently does not reflect a general reduction of ECM components. The synthesis of fibronectin is known to be induced by glucocorticoids (Schwarzbauer, 1991; Sakai et al., 1995a). We showed that glucocorticoids did not affect fibronectin synthesis in NRK fibroblasts. It is therefore likely that the mechanism for downregulation of TN-X may be in part the same as that of TN-C. In contrast glucocorticoids had no effect on TN-X expression in Ren-Ca carcinoma cells. Steroid hormones are known to regulate gene expression in the nucleus, either directly by binding to steroid response elements (receptors for steroid hormones bind as homodimers to the consensus DNA sequence), or indirectly by interfering with other transcription factors (Amero et al., 1992; Tsai and O’Malley, 1994). We performed a computer analysis to find out the binding sites for steroid hormones in the mouse TN-X promoter region, however, there was no significant consensus sequence for the binding of glucocorticoid receptors in the approximate 700 bp upstream from the initiation methionine for the mouse TN-X gene (data not shown). In chicken TN-C, the potential transcription factor binding site for glucocorticoid receptors is located at position −985 of the TN-C promoter sequence (ChiquetEhrismann et al., 1994), and mouse and chicken TN-C promoters have an extended sequence similarity and contain common regulatory motifs (Copertino et al., 1995). We are currently characterizing the more 5′ upstream region of the TN-X gene. The synthesis and secretion of TN-X were found to be stimulated by FBS in both fibroblasts and carcinoma cells. However, their induction level (about 1.6-fold induction in NRK fibroblasts) was by far smaller than that of TN-C, which was described previously (about 5-fold induction) in the primary chick embyo fibroblasts (Pearson et al., 1988). Furthermore, growth factors such as TGF-β1, EGF, TGF-α, PDGF and bFGF, all of which can alter TN-C expression, did not cause significant changes in the expression level of TN-X in either cell type even with prior downregulation of basal TN-X levels through serum starvation. It is therefore likely that the
mechanism of induction of TN-X might have a different signalling pathway from that of TN-C. Fibroblasts which produce very little TN-C in vivo rapidly begin to produce large amounts of TN-C when cultured in vitro. However, this was not the case with TN-X; the primary embryonic fibroblasts secreted far lower amounts of TN-X in vitro when compared with TN-C. Furthermore, no overexpression of TN-X was found in the primary embryonic fibroblasts of the TN-C-gene knockout mice. In the adult skin, we showed a reciprocal tissue distribution pattern between TN-X and TN-C. Considering these findings together, it appears unlikely that TN-X could compensate for the loss of TN-C in the mice lacking a functional TN-C gene. However, we cannot exclude the possibility that TN-X in certain locations might replace TN-C. We have studied the hypothesis that the expression of TN-X protein depends upon a change in the cellular environment from in vitro to in vivo, as is the case with TN-C (Sakai et al., 1993b, 1994). As our in vivo model system, we used Ren-Ca- and 203Glioma-tumor formations in nude mice; two cell lines which both produced TN-X in vitro. The formation of a tumor stroma with a distinct ECM in some ways resembles the development of a mesenchymal ECM during embryogenesis (Sakai et al., 1994). Immunohistochemical analysis of the adult skin revealed that the positive staining for TN-X was demonstrated in the stroma of dermis whereas TN-C was mainly localized adjacent to the epidermis; the expression patterns of TN-X and TN-C showed a reciprocal tissue distribution. We therefore examined the tissue distribution of TN-X during tumorigenesis and compared it with TN-C. In contrast to the distinct localization of TN-C in the tumor stroma, the expression of TN-X was shown to be downregulated considerably not only in the transplanted tumor cells themselves but also in the host fibrous stroma. It could indicate an epithelialderived negative signal acting on the surrounding host-stromal cells. The lack of TN-X in the tumor stroma was noteworthy, since TN-X has been expressed during epithelial-mesenchymal interactions in the embryo (Matsumoto et al., 1994). It suggests that the downregulation of TN-X may be a critical event during tumorigenesis or invasion and metastasis of carcinoma cells. Our in vivo studies indicate that certain physiological regulators present in cancer tissues participated in the downregulation of TN-X and we suggest that glucocorticoids may in part participate in its downregulation of TN-X in stromal fibroblasts. However, it remains to be seen whether lack of TN-X is a general feature of tumor stroma. We have investigated in the present study the analysis of the expression of TN-X in fibroblasts and carcinoma cells. Our current findings could provide a clear basis for clarifying the regulatory mechanisms of TN-X expression. The role of TNX is not well defined, however, we have demonstrated that glucocorticoids downregulate TN-X expression in fibroblasts and that TN-X expression can be influenced by epithelial-stromal interactions. Since interactions between carcinoma cells and their ECM are of importance for the invasion or metastasis of carcinoma cells, the present study also provides evidence for a regulatory role of TN-X in tumorigenesis. We are grateful to Dr S. Asakura, Division of Hemostasis and Thrombosis Research, Jichi Medical School, for his kind gift of agarose-immobilized fibronectin and useful discussions. We are also grateful to Dr H. Kawakatsu, Nippon Shinyaku Co. Ltd; Dr M.K. Owada, Kyoto Pharmaceutical University; and Dr M. Obara,
Downregulation of tenascin-X by glucocorticoids 2077 Hiroshima University, for their valuable suggestions and encouragement. We thank Mr T. Ichimura for his technical assistance in the immunohistochemical procedures, and Mr M. Todoriki for his excellent photographic assistance. This work was supported in part by a Grant-in-Aid for Scientific Research, from the Ministry of Education, Science and Culture of Japan, and NIG Cooperative Research Program (’95-31).
REFERENCES Amero, S. A., Kretsinger, R. H., Moncrief, N. D., Yamamoto, K. R. and Pearson, W. R. (1992). The origin of nuclear receptor proteins: a single precursor distinct from other transcription factors. Mol. Endocrinol. 6, 3-7. Bristow, J., Kian Tee, M., Gitelman, S. E., Mellon, S. H. and Miller, W. L. (1993). Tenascin-X: a novel extracellular matrix protein encoded by the human XB gene overlapping P450c21B. J. Cell Biol. 122, 265-278. Burch, G. H., Bedolli, M. A., McDonough, S., Rosenthal, S. M. and Bristow, J. (1995). Embryonic expression of tenascin-X suggests a role in limb, muscle and heart development. Dev. Dynam. 203, 491-504. Carnemolla, B., Borsi, L., Bannikov, G., Troyanovsky, S. and Zardi, L. (1992). Comparison of human tenascin expression in normal, Simian-virus40-transformed and tumor-derived cell lines. Eur. J. Biochem. 205, 561-567. Chiquet-Ehrismann, R., Mackie, E. J., Pearson, C. A. and Sakakura, T. (1986). Tenascin: an extracelluar matrix protein involved in tissue interactions during fetal development and oncogenesis. Cell 47, 131-139. Chiquet-Ehrismann, R. (1990). What distinguishes tenascin from fibronectin? FASEB J. 4, 2598-2604. Chiquet-Ehrismann, R., Tannheimer, M., Koch, M., Brunner, A., Spring, J., Martin, D., Baumgartner, S. and Chiquet, M. (1994). Tenascin-C expression by fibroblasts is elevated in stressed collagen gels. J. Cell Biol. 127, 2093-2101. Chiquet-Ehrismann, R. (1995). Tenascins, a growing family of extracellular matrix proteins. Experientia 51, 853-862. Copertino, D. W., Jenkinson, S., Jones, F. S. and Edelman, G. M. (1995). Structural and functional similarities between the promoters for mouse tenascin and chicken cytotactin. Proc. Nat. Acad. Sci. USA 92, 2131-2135. Ekblom, M., Fassler, R., Tomasini-Johansson, B., Nilsson, K. and Ekblom, P. (1993). Downregulation of tenascin expression by glucocorticoids in bone marrow stromal cells and in fibroblasts. J. Cell Biol. 123, 1037-1045. Erickson, H. P. and Bourdon, M. A. (1989). Tenascin: An extracellular matrix protein prominent in specialized embryonic tissues and tumors. Annu. Rev. Cell Biol. 5, 71-92. Erickson, H. P. (1993). Tenascin-C, tenascin-R and tenascin-X: a family of talented proteins in search of functions. Curr. Opin. Cell Biol. 5, 869-876. Geffrotin, C., Garrido, J. J., Tremet, L. and Vaiman, M. (1995). Distinct tissue distribution in pigs of tenascin-X and tenascin-C transcript. Eur. J. Biochem. 231, 83-92. Herlyn, M. and Malkowicz, S. B. (1991). Regulatory pathways in tumor growth and invasion. Lab. Invest. 65, 262-271. Jones, F. S., Crossin, K. L., Cunningham, B. A. and Edelman, G. M. (1990). Identification and characterization of the promotor for the cytotactin gene. Proc. Nat. Acad. Sci. USA 87, 6497-6501. Laemmli, U. K. (1970). Cleavage of structural proteins during assembly of the head of the bacteriophage T4. Nature 227, 680-685. LaFleur, D. W., Fagin, J. A., Forrester, J. S., Rubin, S. A. and Sharifi, B. G. (1994). Cloning and characterization of alternatively spliced isoforms of rat tenascin. J. Biol. Chem. 269, 20757-20763. Le, J. and Vilcek, J. (1987). Tumor necrosis factor and interleukin 1: cytokines with multiple overlapping biological activities. Lab. Invest. 56, 234-248. Matsumoto, K., Arai, M., Ishihara, N., Ando, A., Inoko, H. and Ikemura, T. (1992a). Cluster of fibronectin type III repeats found in the human major histocompatibility complex class III region shows the highest homology with the repeats in an extracellular matrix protein, tenascin. Genomics 12, 485-491. Matsumoto, K., Ishihara, N., Ando, A., Inoko, H. and Ikemura, T. (1992b). Extracellular matrix protein tenascin-like gene found in human MHC class III region. Immunogenetics 36, 400-403. Matsumoto, K., Saga, Y., Ikemura, T., Sakakura, T. and ChiquetEhrismann, R. (1994). The distribution of tenascin-X is distinct and often reciprocal to that of tenascin-C. J. Cell Biol. 125, 483-493. McCachren, S. S. and Lightner, V. A. (1992). Expression of human tenascin
in synovitis and its regulation by interleukin-1. Arthritis Rheumatol. 35, 1185-1192. Meiners, S., Marone, M., Rittenhouse, J. L. and Geller, H. M. (1993). Regulation of astrocytic tenascin by basic fibroblast growth factor. Dev. Biol. 160, 480-493. Morel, Y., Bristow, J., Gitelman, S. E. and Miller, W. L. (1989). Transcript encoded on the opposite strand of the human steroid 21hydroxylase/complement component C4 gene locus. Proc. Nat. Acad. Sci. USA 86, 6582-6586. Pearson, C. A., Pearson, D., Shibahara, S., Hofsteenge, J. and ChiquetEhrismann, R. (1988). Tenascin: cDNA cloning and induction by TGF-β. EMBO J. 7, 2977-2982. Rettig, W. J., Triche, T. and Garin-Chesa, P. (1989). Stimulation of human neuronectin secretion by brain-derived growth factors. Brain Res. 487, 171177. Rettig, W. J., Erickson, H. P., Albino, A. P. and Garin-Chesa, P. (1994). Induction of human tenascin (neuronectin) by growth factors and cytokines: cell type-specific signals and signalling pathways. J. Cell Sci. 107, 487-497. Saga, Y., Tsukamoto, T., Jing, N., Kusakabe, M. and Sakakura, T. (1991). Mouse tenascin: cDNA cloning, structure and temporal expression of isoforms. Gene 104, 177-185. Saga, Y., Yagi, T., Ikawa, Y., Sakakura, T. and Aizawa, S. (1992). Mice develop normally without tenascin. Genes Dev. 6, 1821-1831. Sakai, T., Kawakatsu, H., Hirota, N., Yokoyama, T., Sakakura, T. and Saito, M. (1993a). Specific expression of tenascin in human colonic neoplasms. Br. J. Cancer. 67, 1058-1065. Sakai, T., Kawakatsu, H., Hirota, N., Yokoyama, T., Takaoka, T., Sakakura, T. and Saito, M. (1993b). Tenascin expression in vitro and in vivo: comparison between epithelial and nonepithelial rat cell lines. Exp. Cell Res. 206, 244-254. Sakai, T., Kawakatsu, H., Ohta, M. and Saito, M. (1994). Tenascin induction in tenascin nonproducing carcinoma cell lines in vivo and by TGF-β1 in vitro. J. Cell. Physiol. 159, 561-572. Sakai, T., Kawakatsu, H., Furukawa, Y. and Saito, M. (1995a). Regulation of EGF-induced tenascin-C by steroids in tenascin-C non-producing human carcinoma cells. Int. J. Cancer. 63, 720-725. Sakai, T., Ohta, M., Furukawa, Y., Saga, Y., Aizawa, S., Kawakatsu, H. and Saito, M. (1995b). Tenascin-C induction by the diffusible factor epidermal growth factor in stromal-epithelial interactions. J. Cell. Physiol. 165, 18-29. Sakai, T., Ohta, M., Kawakatsu, H., Furukawa, Y. and Saito, M. (1995c). Tenascin-C induction in Whitlock-Witte culture: a relevant role of the thiol moiety in lymphoid-lineage differentiation. Exp. Cell Res. 217, 395-403. Schwarzbauer, J. E. (1991). Fibronectin: from gene to protein. Curr. Opin. Cell Biol. 3, 786-791. Sharifi, B. G., LaFleur, D. W., Pirola, C. J., Forrester, J. S. and Fagin, J. A. (1992). Angiotensin II regulates tenascin gene expression in vascular smooth muscle cells. J. Biol. Chem. 267, 23910-23915. Talts, J. F., Aufderheide, E., Sorokin, L., Ocklind, G., Mattsson, R. and Ekblom, P. (1993). Induction of mouse tenascin expression by a human sarcomatoid Wilms’ tumor cell line growing in nude mice. Int. J. Cancer 54, 868-874. Talts, J. F., Weller, A., Timple, R., Ekblom, M. and Ekblom, P. (1995). Regulation of mesenchymal extracellular matrix protein synthesis by transforming growth factor-β and glucocorticoids in tumor stroma. J. Cell Sci. 108, 2153-2162. Tsai, M.-J. and O’Malley, B. W. (1994). Molecular mechanisms of action of steroid/thyroid receptor superfamily members. Annu. Rev. Biochem. 63, 451486. Tso, J. Y., Sun, X.-H., Kao, T.-H., Reece, K. S. and Wu, R. (1985). Isolation and characterization of rat and human glyceraldehyde-3-phosphate dehydrogenase cDNAs: genomic complexity and molecular evolution of the gene. Nucl. Acids Res. 13, 2485-2502. Tucker, R. P., Hammarback, J. A., Jenrath, D. A., Mackie, E. J. and Xu, Y. (1993). Tenascin expression in the mouse: in situ localization and induction in vitro by bFGF. J. Cell Sci. 104, 69-76. Yavin, E., Gabai, A. and Gil, S. (1991). Nerve growth factor mediates monosialoganglioside-induced release of fibronectin and J1/tenascin from C6 glioma cells. J. Neurochem. 56, 105-112. (Received 12 February 1996 – Accepted 22 April 1996)
