Molecular Human Reproduction, Vol. 7, No. 5, 463-474,
May 2001
© 2001 European Society of Human Reproduction and Embryology
Implantation and pregnancy |
Regional and cellular localization of osteonectin/SPARC expression in connective tissue and cytotrophoblastic layers of human fetal membranes at term
Preterm Birth Research Group, 1 Departments of Obstetrics and Gynaecology and 2 Department of Pathology, Leicester Warwick Medical School, Faculty of Medicine and Biological Sciences, University of Leicester, Leicester LE2 7LX, UK
Abstract
Fetal membranes overlying the cervix in patients prior to and during labour, and within the rupture tear after spontaneous delivery at term, exhibit altered morphology. In this study we report that in comparison to mid-zone fetal membranes biopsies, these regions are characterized by increased expression of the matricellular protein osteonectin or SPARC (Secreted Protein Acidic and Rich in Cysteine). In the reticular layer, the percentage of vimentin positive mesenchymal cells immunoreactive for osteonectin increased in these regions from 3–4% to 25–33% and represented a fraction of the 
-smooth muscle actin positive myofibroblasts elevated in the same regions. In the fibroblastic layer, the percentage of osteonectin positive cells increased from 1–5% to 8–13%; however, these did not exhibit the same relationship to the -smooth muscle actin positive myofibroblasts in this layer. In the cytotrophoblastic layer the percentage of cytotrophoblastic cells immunoreactive for osteonectin increased from 1% to 6–12%. Elevation of in-situ detectable mRNA was also observed in the same cellular populations in this region. The incidence of cells positive for osteonectin mRNA or protein in the reticular layer correlated with morphological changes. Osteonectin has been implicated in the regulation of extracellular matrix turnover, and its pattern of expression suggests a role in the regional connective tissue and cytotrophoblastic changes proposed to be involved in the cleavage and rupture of fetal membranes.
amnion/chorion/cytotrophoblast/myofibroblast/osteonectin
Introduction
In ~10% of term pregnancies and up to 60% of preterm deliveries, the fetal membranes, which encapsulate the fetus and amniotic fluid, rupture prior to labour. In the latter it is a direct antecedent of preterm birth (Alger and Pupkin, 1986
; Keirse et al., 1989; Kelly, 1995; French and McGregor, 1996; Romero et al., 1999). Although infection has been implicated in the aetiopathology of a proportion of cases (Gibbs et al., 1992; French and McGregor, 1996) the mechanisms of this pre-labour rupture in the absence of infection, and indeed the mechanisms underlying their spontaneous rupture during term labour, are essentially unknown (Kelly, 1995; French and McGregor, 1996; Major and Garite, 1997; Parry and Strauss, 1998; Romero et al., 1999; Bell, 2000).
Previous studies have identified unusual morphological features of the fetal membranes within a restricted part of the rupture tear after labour and delivery at term (Bou-Resli et al., 1981; Malak and Bell, 1993, 1994). These features include marked swelling of the connective tissue layers of the amnion and chorion (layers rich in fibrillar collagen types I and III and which bestow the structural strength of the membranes) and thinning of the cellular layers, the cytotrophoblastic and decidual layers (Malak and Bell, 1994). Subsequent studies have demonstrated a similar altered morphology in single biopsies obtained from fetal membranes overlying the cervix in the lower uterine pole prior to, and during, labour at term, compared to biopsies distal to this site (McLaren et al., 1999; McParland et al., 2000). The rupture tear after delivery invariably involves the fetal membranes overlying the cervix in the lower uterine pole, suggesting that this area represents the site of initial membrane rupture. It has been proposed that the morphological features detected prior to, and during, labour at this site reflect a structural weakness, therefore indicating a localized site of increased susceptibility to rupture during the increased intra-amniotic pressures experienced during labour (Malak and Bell, 1993, 1994, 1996; Bell and Malak, 1997).
We recently reported that the morphological features at these anatomical sites, over the cervix prior to rupture and within the rupture line after delivery, were associated with an alteration in the phenotype of mesenchymal cells of the connective tissue layers. This principally involved the reticular layer of the chorion, where a proportion of cells possessed -smooth muscle actin (-SMA) but not smooth muscle myosin (McParland et al., 2000). This phenotype is considered the hallmark of the `activated' or `differentiated' myofibroblast (Sappino et al., 1990; Schmitt-Graff et al., 1994). A number of functions have been ascribed to these cells, including that of extracellular matrix synthesis and degradation, and that of contraction, and reflect their proposed central role in wound healing, fibro-contractive diseases and desmoplastic reactions to cancer (Schurch et al., 1998; Powell et al., 1999; Serini and Gabbiani, 1999).
Recently, -SMA positive myofibroblasts arising during human liver fibrogenesis have been demonstrated to express the matricellular protein osteonectin (Blazejewski et al., 1997). Osteonectin, also known as SPARC (Secreted Protein Acidic and Rich in Cysteine), BM-40 and 43K protein, is a 43 kDa Ca2+ and collagen binding glycoprotein (Termine et al., 1981) considered to be a prototype of the matricellular proteins which include tenascin-C and thrombospondin (Sage and Bornstein, 1991; Lane and Sage, 1994; Motamed, 1999; Yan and Sage, 1999). In the adult its expression is limited but it has been identified in a variety of tissues undergoing cellular remodelling or renewal, and in sites exhibiting high cell proliferation and migration, extracellular matrix remodelling and active epithelial-mesenchymal interactions (Holland et al., 1987; Sage et al., 1989; Mundlos et al., 1992; Reed et al., 1993; Porter et al., 1995). Osteonectin exhibits a diverse range of properties in vitro, including those that would support its suggested key role in wound repair and extracellular matrix turnover such as the regulation of the synthesis of plasminogen activator inhibitor-1 (Hasselaar et al., 1991; Lane et al., 1992) and of the matrix metalloproteinases (MMP)-1, -2, -3 and -9 (Tremble et al., 1993; Shankavaram et al., 1997; Gilles et al., 1998).
In the present study we have therefore determined whether the anatomically defined regional alterations in fetal membrane morphology and myofibroblastic differentiation reported prior to, and during, labour and after membrane rupture at term, are associated with osteonectin expression.
Materials and methods
Patient details and tissue sampling
Fetal membranes were obtained prior to (n = 15), during (n = 5), and following (n = 5) spontaneous labour at term as previously described (McParland et al., 2000). Membranes were collected on an anonymous basis according to guidance from the local Ethics Committee. In all cases potentially infected membranes were excluded following identification of polymorphonuclear infiltration of the amniochorion on histological examination. Fetal membrane specimens of ~1x3 cm were regionally biopsied from the following areas of the fetal membrane: (i) `cervical' (lower uterine pole) membrane biopsies from pre-labour (with intact membranes) and labour Caesarean sections; (ii) rupture tear biopsies following term vaginal delivery; (iii) mid-zone biopsies, obtained halfway between the cervical area and the placental edge in the cases of Caesarean sections and between the rupture line and the placental edge in patients after delivery. In all cases these biopsies were taken at least 10–12 cm from the cervical areas or rupture lines. Sequential biopsies of ~1x3 cm were taken from the rupture line to the placental edge in a single patient following labour and delivery at term. Biopsies were rolled and fixed in 10% formal saline for 48 h before processing and mounting in paraffin wax. Tissue sections were cut (4 µm) and then stained with haematoxylin and eosin. For larger specimens to be used for protein and RNA extraction procedures, membranes were collected from an area within 5 cm of the Babcock, i.e. to include the `cervical' area and at least 12 cm distally from the Babcock tissue forceps i.e. `mid-zone'. The restricted area of tissue taken, to ensure accurate mapping, meant that in some cases different tissue samples were used for different techniques.
Immunocytochemistry
Immunocytochemistry was performed on formalin-fixed paraffin-embedded tissue sections, as previously described (McParland et al., 2000). The following murine monoclonal antibodies used were: anti-osteonectin (clone N50; Biodesign International, Kennebunk, USA, 7.5 µg/ml), anti-vimentin (clone V9; Sigma, Poole, UK; 3.87 µg/ml), anti--SMA (clone 1A4; Dako, Cambridge, UK; 1.9 µg/ml), anti-cytokeratin (clone MNF116; Dako; 0.93µg/ml), and anti-CD68 (clone PG-M1; Dako; 3.6 µg/ml). Rabbit polyclonal antiserum against bovine osteonectin, BON-I, was a generous gift from L.W.Fisher (NIDCR; NIH, Bethesda, MA, USA; Fisher et al., 1995) and was used at 1:4000. Tissue sections were microwaved in 10 mmol/l sodium citrate, pH 6.0, for 30 min prior to incubation with BON-I. Slides were incubated overnight at 4°C for anti-vimentin and BON-I or at 37°C for 1 h with other antibodies. Incubation with anti-osteonectin clone NSC was performed in the presence of 2 mmol/l CaCl2. Negative controls were primary antibody omission and inclusion of mouse IgG (Sigma) or rabbit serum (Dako) at concentrations that matched that of the primary antibodies. Umbilical cord was used as a positive control tissue to validate the antibodies and techniques.
Cell line and culture
Human melanoma cell line SK-MEL-28 was obtained from the American Type Culture Collection (Rockville, MD, USA). Cells were grown in -minimal essential medium containing 10% v/v fetal calf serum (Life Technologies Ltd, Paisley, UK) to 90% confluence. SK-MEL-28 cells were harvested with 0.25% trypsin-EDTA (Life Technologies) and lysed in 100 mmol/l Tris–HCl pH 8.0, 500 mmol/l LiCl, 10 mmol/l EDTA pH 8.0, 1% sodium dodecyl sulphate (SDS) and 5 mmol/l dithiothreitol.
mRNA isolation
mRNA from SK-MEL-28 cells was captured onto Oligo d(T)25 Dynabeads© (Dynal, Merseyside, UK), and washed in 10 mmol/l Tris–HCl pH 8.0, 0.15 mol/l LiCl, 1 mmol/l EDTA with and then without 0.1% SDS. Fresh fetal membrane samples were homogenized in Tri Reagent (Sigma), and stored at –80°C until required. Total RNA extraction was performed according to the Tri Reagent extraction protocol from Sigma (Technical Bulletin MB-205).
In-situ hybridization
A probe to detect osteonectin mRNA in situ was synthesized by reverse transcription-polymerase chain reaction (RT-PCR) amplification of a 325 bp region of the osteonectin gene, employing mRNA extracted from SK-MEL-28 cells. This template was then used in an `asymmetric' PCR reaction incorporating digoxigenin to produce a single stranded antisense DNA probe.
Oligonucleotide primers were designed for osteonectin (forward primer 5'-GCTCCACCTGGACTACATCG-3', reverse primer 5'-GGAGAGGTACCCGTCAATGG-3'), to amplify a 325 bp product spanning exons 6–9. Primers for ß-actin were used as positive controls (forward 5'-GGAGACAAGCTTGCTCATCACCATTGGCAATGAGCG-3', reverse 5'-GCGAATTCGAGCTCTAGAAGCATTTGCGGTGGACG-3'). Forward strand primers were synthesized with a 5' biotin group.
cDNA was prepared using 1 µl SK-MEL-28 mRNA in RT buffer (50 mmol/l Tris–HCl pH 8.3, 50 mmol/l KCl, 10 mmol/l MgCl2, 10 mmol/l dithiothreitol, 0.5 mmol/l spermidine (Promega, Southampton, Hants), 1 mmol/l dNTPs (Roche Molecular Biochemicals, Lewes, East Sussex, UK), 25 IU RNasin© ribonuclease inhibitor (Promega), and 5 IU AMV reverse transcriptase (Promega) in a volume of 25 µl. The reaction was incubated at 42°C for 1 h. cDNA was prepared from 0.5 µg fetal membrane mRNA in the same manner, including 15 pmol oligo d(T) 12–18 (Amersham Pharmacia Biotech, St Albans, UK) in the reaction. Control reactions were prepared in the absence of AMV reverse transcriptase.
Templates for later synthesis of in-situ probes were produced by PCR, carried out in a Techne Genius© thermal cycler. The reaction was performed with 1 µl SK-MEL-28 cDNA in the following reagents: 45 mmol/l Tris–HCl pH 8.8, 11 mmol/l (NH4)2SO4, 4.5 mmol/l MgCl2, 800 µmol/l dNTPs, 110 µg/ml bovine serum albumin, 6.7 mmol/l ß-mercaptoethanol, 4.4 µmol/l EDTA pH 8.0 and 10 pmol of forward and reverse primers in a reaction volume of 50 µl. The DNA was denatured at 98°C for 5 min, and held at 62°C during the addition of 1 IU Taq polymerase (Promega), and heated to 72°C for 1 min. The following cycle profile was used: 98°C for 1 min, 62°C for 45 s, 72°C for 1 min, then the reaction held at 72°C at the end of the amplification. PCR for osteonectin was carried out for 25 cycles. PCR amplification from fetal membrane cDNA was carried out for 30 cycles. PCR products were loaded onto a 3% agarose gel containing 15 µg/100 ml ethidium bromide, with 100 bp ladder (Life Technologies) in one lane as a size marker, and run at 100 V for 2 h. Bands were visualized on a UV transilluminator. One µl of the resultant product was used as a template for production of a probe for in-situ hybridization in an asymmetric PCR reaction including 70 µmol/l digoxigenin-11-dUTP (Roche Molecular Biochemicals, Lewes, East Sussex, UK) and 100 pmol reverse primer in reaction buffer as described above. One IU Taq polymerase (Promega) was added after an initial denaturation to 98°C and amplification was carried out for 20 cycles. Double-stranded contaminant was removed by incubation with Streptavidin-Dynabeads© (Dynal), to give a single-stranded digoxigenin labelled probe. A positive control probe to ß-actin was synthesized in the same manner. Probe concentration was titrated to eliminate background staining, whilst retaining good signal strength.
In-situ hybridization was carried out on de-waxed rehydrated tissue sections. Proteinase K pretreatment (2-10 µg/ml) was optimized for each tissue section. Prehybridization, hybridization and post-hybridization washing were modified from the method of Pringle (1995) to include the presence of 50% formamide in these steps. Detection was carried out by incubation with anti-digoxigenin antibody (Roche Molecular Biochemicals) followed by BCIP/NBT (Sigma) containing 1 mmol/l levamisole (Sigma) (Pringle, 1995). Negative controls were carried out using sense probe to osteonectin, RNase pre-treatment of the tissue section, and by omission of the probe in the hybridization protocol. Umbilical cord was used as a positive control tissue.
Histological, immunocytochemical and in-situ assessment
Haematoxylin and eosin and immunocytochemical stained sections of membrane rolls from all biopsies were examined under light microscopy connected to an image capture system. This system comprised an Apple Macintosh Centris running NIH Image© (version 1.51). Each section was divided into quadrants. For measurements of constituent layer thickness and for counting of immunoreactive and in-situ positive cells, 10 fields per roll were examined along the intersections of 12 o'clock to 6 o'clock and 3 o'clock to 9 o'clock. Measurements were taken only from sections cut vertically, showing a single layer of amniotic epithelium. The numbers of immunoreactive or in-situ positive cells were counted in the whole thickness of each specified layer in a standard fetal membrane length as determined by the width of the computer screen at magnification of x40. The derived fetal membrane morphometric index (FMMI) is the ratio of the total connective tissue layer thickness, i.e. amnion and chorionic reticular layers, and the total cellular layer thickness, i.e. cytotrophoblast and decidua (Malak and Bell, 1994). Fetal membrane morphology data, and vimentin and -SMA cell counting data for the majority of the samples in this study have been described previously (McParland et al., 2000), and were used to derive osteonectin cell density and percentage data, and to relate osteonectin immunoreactivity to morphology and -SMA immunoreactivity.
Protein extraction and Western blotting
Fetal membrane samples were washed thoroughly in phosphate-buffered saline (PBS), and homogenized (50% w/v) in 0.2 mol/l 3-[cyclohexylamino]-1-propane-sulphonic acid, 150 mmol/l NaCl, 1 mmol/l EDTA with 1 mmol/l PMSF at pH 11.5. The homogenate was centrifuged at 100 000 g for 1 h and the supernatant stored at –80°C. Proteins were separated under reducing conditions on 10% polyacrylamide gels with 0.1% SDS (Bio-Rad, Hemel Hempstead, Herts, UK). They were transferred to nitrocellulose in 25 mmol/l Tris, 192 mmol/l glycine, 20% methanol for 1 h at 100 V. The blots were blocked in 5% non-fat milk powder for 30 min and incubated in 0.2 µg/ml mouse anti-osteonectin monoclonal antibody N50 or 1/2000 rabbit anti-bovine osteonectin, BON-I, overnight at 4°C. The blots were washed in 50 mmol/l Tris, 150 mmol/l NaCl, 0.1% Tween 20 (TBS-Tween), incubated in either horseradish peroxidase-conjugated sheep anti-mouse or donkey anti-rabbit antibody (Amersham Pharmacia Biotech), and washed in TBS-Tween. Detection was with enhanced chemiluminescence (Amersham Pharmacia Biotech). Bands were visualized by exposure to Kodak Biomax X-ray film. Films were scanned on a Umax Astra 1220P flatbed scanner and densitometry performed using Scion Image©. Control blots were incubated with equivalent concentration of mouse IgG or rabbit serum.
Statistical analysis
Two-way analysis of variance was used for the analysis of the number of immunoreactive cellular populations of the tissue layers and thickness measurements. Sampling from different subjects and from different zones of the fetal membranes were included as independent factors in the analysis. Scheffé's test was used for subsequent comparisons between different membrane zones and between patient groups. For comparison between cell counts for different zones for in-situ hybridization, and for protein densitometry results, Student's t-test was employed. Correlation coefficients were calculated using linear regression analysis. All statistical analysis was carried out using StatView© software, version 4.51 (Abacus Concepts, CA, USA).
Results
Localization of osteonectin immunoreactivity
Osteonectin immunoreactive cells were identified in the amniotic epithelium, and in the fibroblast, reticular and cytotrophoblast layers in biopsies obtained from all zones and all patient groups studied with both monoclonal and polyclonal antibodies. Immunoreactive cells were also seen within degenerate villi in the cytotrophoblast layer. No immunoreactivity was detected in the extracellular matrix. Serial sections stained with antibodies to CD68 or to cytokeratin and confirmed that the osteonectin positive cells seen within the fibroblast and reticular layers were fibroblasts, and not macrophages (Figure 1), and that those within the cellular layers of the membrane were cytotrophoblast cells (Figure 2). The latter cells were found predominantly within the superficial maternal aspect of the cytotrophoblast layer, although in membranes exhibiting significant thinning of the cytotrophoblast layer, immunoreactive cytotrophoblast cells were also seen adjacent to the pseudobasement membrane (Figure 2). Significant numbers of immunoreactive cells and regional differences in their incidence, were seen in the non-macrophage cells in the chorionic reticular and amniotic fibroblastic layers, and in the cytotrophoblast cells of the cytotrophoblastic layer (Table I).
|
|
|
Variation in distribution of osteonectin immunoreactivity
The variation in the distribution of cells immunoreactive with the osteonectin monoclonal antibody was examined on sections of membrane biopsies obtained pre-labour, during labour and post-labour and delivery (n = 5 per group). The numbers of osteonectin immunoreactive cells in the reticular layer expressed per unit length of membrane were 6.5-fold higher in the `cervical' biopsies in the pre-labour group (P = 0.0073), 10-fold higher in the `cervical' biopsies in the labour-affected group (P = 0.0001) and 12-fold higher in the post-labour rupture line biopsies after labour and delivery (P = 0.013), compared to their respective mid-zone biopsies (Table I
, Figure 3). These changes were not affected by thickness of the connective tissue layers since, in the same sites, the mean densities of osteonectin immunoreactive cells were increased in the pre-labour, labour-affected and post-labour groups 6.1-fold (P = 0.0091), 5.3-fold (P < 0.0001) and 7.5-fold (P = 0.0067) respectively (data not shown). When expressed as a percentage of the total vimentin positive population present, in the reticular layer of the mid-zone biopsies the means ranged from only 3.51 to 4.03% (Table II). However, in the pre-labour `cervical' biopsies, the labour-affected `cervical' biopsies and in the post-labour rupture line biopsies, this was 7.1-fold higher at 24.85% (P = 0.002), 8.2-fold higher at 33.00% (P = 0.003), and 9.3-fold higher at 33.00% (P = 0.0007) respectively. There was no significant difference between the numbers of osteonectin immunoreactive cells, when expressed as a number or as a percentage of vimentin positive cells, in the reticular layers of the mid-zone biopsies between the patient groups, or between the `cervical' biopsies and post-labour rupture line biopsies (Tables I and II).
|
|
In the amniotic connective fibroblast layer, significantly higher numbers of osteonectin immunoreactive cells were noted in the `cervical' compared to the mid-zone biopsies in the labour-affected group (12.7-fold higher, P = 0.014, Table I
), and in the proportion of vimentin positive cells immunoreactive for osteonectin in the `cervical' compared to the mid-zone biopsies in the pre-labour and the labour-affected groups (7.2-fold, P = 0.038, and 16.6-fold higher, P = 0.012 respectively; Table II). The failure to detect a difference in the post-labour group appeared to be due to increased numbers of immunoreactive cells in the mid-zone biopsy from one patient (16.9%), bringing the average for the group to 4.46%, compared to 1.44 and 0.81% in the pre-labour and labour-affected groups (Table II). Exclusion of the data from this patient brings the average for the group to 1.34%, significantly lower than the average for the rupture line biopsies, 8.32% (P = 0.015).
In the cytotrophoblast layer, there were 5.8-, 4.6- and 6.8-fold higher numbers of immunoreactive cells in the `cervical' biopsies in the pre-labour, labour-affected, and in the post-labour rupture line biopsies compared to their respective mid-zone biopsies, although this only achieved significance in the former group (P = 0.024, 0.078 and 0.11, Table I). The percentage of cytotrophoblast cells immunoreactive for osteonectin was significantly higher in the pre-labour and post-labour rupture line biopsies, compared to their respective mid-zone biopsies (8.8-fold P = 0.039, and 8.5-fold P = 0.041 respectively, Table II).
When the data from all biopsies were combined, there was a significant positive correlation between the absolute numbers of osteonectin positive cells in the reticular layer and the thickness of the reticular layer (r2 = 0.434, P < 0.0001), the combined thickness of the amniotic and chorionic connective tissue layers (r2 = 0.445, P < 0.0001), and the FMMI (r2 = 0.868, P < 0.0001) and a negative correlation with the total thickness of the cellular layers (r2 = 0.347, P = 0.0006). There was also a significant positive correlation between the number of osteonectin positive cells in the fibroblast layer and in the reticular layer (r2 = 0.436, P < 0.0001). The number of immunoreactive cells in the cytotrophoblast layer showed a weak but significant correlation with the number of immunoreactive cells in the reticular layer (r2 = 0.137, P = 0.044), and a weak inverse correlation with the total thickness of the cellular layers (r2 = 0.199, P = 0.014).
Relationship to distribution of -SMA immunoreactive cells
Serial sections demonstrated that the majority of immunoreactive osteonectin cells within the reticular layer were also -SMA immunoreactive (Figure 4). However, not all -SMA immunoreactive cells contained immunoreactive osteonectin, e.g. in the reticular layer of post-labour rupture line biopsies 57.7 ± 11.4% (mean ± SEM) of -SMA positive cells were immunoreactive for osteonectin. However, in mid-zone biopsies, where the numbers of osteonectin positive cells in the reticular layers were low, these exceeded the numbers of -SMA positive cells. When the data from all biopsies were combined there was a significant positive correlation between the absolute numbers of osteonectin and -SMA positive cells in the reticular layer (r2 = 0.630, P < 0.0001). On serial sections in the fibroblast layer, there was no apparent direct and consistent relationship between osteonectin and -SMA immunoreactive cells, with individual biopsies from cervical and rupture line fetal membranes possessing cells immunoreactive for either osteonectin or -SMA and for both. However, when data from all biopsies were combined, there was a significant correlation between the numbers of osteonectin and -SMA positive cells in the fibroblast layer (r2 = 0.286, P = 0.0023). In a fetal membrane obtained from a patient after rupture and delivery, where multiple biopsies were obtained, there was a correlation between numbers of osteonectin immunoreactive cells in the reticular layer, the number of -SMA positive cells (r2 = 0.86, P < 0.0001) and the FMMI of the biopsies (r2 = 0.552, P = 0.001).
|
Localization and distribution of osteonectin mRNA positive cells
RT-PCR of mRNA from both fetal membranes and SK-MEL-28 cells demonstrated a single 325 bp band (Figure 5
). In-situ hybridization was performed on paired mid-zone and `cervical' fetal membrane biopsies from five pre-labour patients. mRNA was detected in significant numbers of cells in both the reticular and fibroblast layers, and in smaller numbers of cells in the amniotic epithelial and cytotrophoblast layers (Figure 6). In the reticular layer the number of vimentin positive cells expressing osteonectin mRNA was 10.7-fold higher in the `cervical' biopsies compared to the mid-zone biopsies (8.4 ± 2.37 compared to 0.78 ± 0.39; P = 0.013). In the fibroblast layer, the numbers of cells expressing osteonectin mRNA in the `cervical' biopsies was 8.2-fold higher compared to those in the mid-zone biopsies (0.82 ± 0.29 compared to 0.10 ± 0.05; P = 0.04). This represented 9.27 and 25.44% of vimentin positive cells in the fibroblast and reticular layers of the `cervical' biopsies respectively. As a group, `cervical' biopsies exhibiting highest numbers of immunoreactive cells in the reticular and fibroblast layers also exhibited the highest numbers of in-situ positive cells. However, when serial sections were examined, areas found to contain the highest numbers of in-situ positive cells tended to contain low numbers of immunoreactive positive cells and vice versa. Within those areas positive for both, cells were detected that were positive for both protein and mRNA (Figure 6). On serial sections of `cervical biopsies', in-situ positive cells in the reticular layer were immunoreactive for -SMA. In the `cervical' biopsies of this group of patients, the percentages of cells immunoreactive for osteonectin and immunoreactive for -SMA were 41.27 ± 7.19% and 68.28 ± 3.61% in the reticular layer and 14.73 ± 3.67% and 18.35 ± 11.13% in the fibroblastic layers respectively (all results expressed as mean ± SEM).
|
|
Characterization of osteonectin protein in fetal membranes
Western blotting of fetal membrane extracts obtained during and following labour (n = 9) with monoclonal antibody revealed a single immunoreactive band with an apparent mol. wt of 43 000, consistent with that of authentic osteonectin (Figure 5
). An identical band was detected with polyclonal antisera, with no evidence of any cleaved fragments (data not shown). Densitometry demonstrated significantly higher amounts of protein in the `cervical'/rupture line biopsies compared to the mid-zone samples (2.9-fold higher, P = 0.0096). The densitometry value correlated significantly with the total connective tissue thickness (r2 = 0.403, P = 0.0046), and inversely with the total cellular thickness (r2 = 0.381, P = 0.0063), and therefore also correlated with the derived FMMI of the membrane biopsies (r2 = 0.50, P = 0.001). The densitometry value also correlated with the number of osteonectin immunoreactive cells within the reticular layer (r2 = 0.394, P = 0.0053) and with the number of cells immunoreactive for -SMA within the reticular layer of the biopsies (r2 = 0.25, P = 0.037). Discussion
In this study we have demonstrated the expression of osteonectin by specific mesenchymal populations within the connective tissue layers of human fetal membranes and, most importantly, that this pattern of expression was primarily restricted to specific anatomical regions, i.e. the `cervical' membranes lying in the lower uterine pole before and during labour, and within the rupture line after delivery. This was particularly prominent within the chorionic connective tissue reticular layer where 25–33% of vimentin positive cells were immunoreactive for osteonectin at these sites in contrast to only 3–4% at distal sites. In the amniotic connective tissue fibroblastic layer obtained from the same sites, more mesenchymal cells expressed immunoreactive osteonectin compared to those at distal sites, albeit at a lower frequency, i.e. 8–13% and 1–4% respectively. A marked increase in the numbers of cells positive for osteonectin mRNA was detected in the same cellular populations in the same regions; however, a strict cellular co-localization between mRNA and protein was not observed, suggesting a transient increase in osteonectin transcription in these cells in these regions.
The anatomical distribution of expression of osteonectin in the connective tissue layers correlated with the reported distribution of -SMA positive cells (McParland et al., 2000). However, in the present study we have noted that all patients in the pre-labour group exhibited higher expression of -SMA positive cells in the lower uterine pole `cervical' biopsies. Mesenchymal fibroblastic cells expressing -SMA are considered to represent `activated'/`differentiated' myofibroblasts (Sappino et al., 1990; Schmitt-Graff et al., 1994) and the association of osteonectin expression with -SMA positive myofibroblasts has been previously reported in pathological conditions such as liver fibrosis (Blazejewski et al., 1997) and tubulo-interstitial inflammation and fibrosis in the kidney (Pichler et al., 1996). A similar association seems to apply to the -SMA positive `activated' myofibroblastic cells in the reticular layer in the lower uterine pole `cervical' biopsies. However, although its expression may be dependent upon this phenotype, the observation that only 58% of -SMA positive cells in the reticular layer expressed osteonectin suggests that other factors may be required to induce osteonectin in these cells. In the fibroblastic layer, however, although osteonectin and -SMA positive cells also increased in frequency in the cervical biopsies, these were not always co-localized indicating that osteonectin expression and the `activated'/`differentiated' myofibroblast phenotype might not be intimately linked in the cells of this layer. This differential expression and, indeed, increased expression of osteonectin in the reticular layer compared to the fibroblastic layer, may relate to the normal phenotype of the mesenchymal cells present in these layers. Critically, mesenchymal cells of the reticular layer express desmin throughout the fetal membrane irrespective of anatomical location, and therefore represent a form of myofibroblast, albeit not expressing -SMA and hence not `activated', whereas those in the fibroblastic layer are desmin negative and represent `true' fibroblasts (Khong et al., 1986). However, the fact that in both layers increased osteonectin expression and frequency of -SMA positive cells are detected in the same specific anatomical region indicates a common inducer(s) within this region prior to the onset of labour.
The factors and mechanisms involved in regulation of osteonectin have not been fully elucidated, and although there is evidence for positive regulation by transforming growth factor (TGF)-ß1, interleukin-1, colony stimulating factor-1, progesterone and glucocorticoids and negative regulation by NO, it has been proposed that other unidentified cytokines may act in vivo (Lane and Sage, 1994). Of these, TGF-ß has been demonstrated to increase osteonectin mRNA and protein in rabbit articular chondrocytes (Nakamura et al., 1996), rat clonal pre-osteoblast-like cells (Zhou et al., 1993), human gingival fibroblasts (Wrana et al., 1991), dermal fibroblasts (Reed et al., 1994) and pulp cells (Shiba et al., 1998). Of the three independent and interactive signals that have been implicated in the induction of differentiation of -SMA-expressing myofibroblasts, i.e. cytokines, the extracellular matrix and mechanical forces, TGF-ß1 appears to exert an essential role (Desmouliere, 1995; Schurch et al., 1998). In the connective tissues of the fetal membranes, high osteonectin expression and/or the appearance of the myofibroblastic phenotype are principally confined to a specific anatomical region, i.e. lower uterine pole, and are present prior to labour and delivery. Labour or membrane rupture or delivery appear to have no effect upon expression, suggesting that this pattern is programmed into the fetal membranes during pregnancy. It has been suggested that the long-term differential stretch in the fetal membranes in the lower uterine pole, associated with the development of the lower uterine segment and/or cervical effacement during late pregnancy, may provide the inducing stimulus for myofibroblastic activation (McParland et al., 2000). Since myofibroblasts provide traction, this possibly may represent a functional response to that stretch. Whether osteonectin gene transcription is also regulated by stretch is not known, but it did not appear after acute stretch of fetal membranes in vitro (Nemeth et al., 2000). A direct role in contraction has been indicated by the identification of a keratocyte `contraction-stimulating factor' as osteonectin (Mishima et al., 1998). Osteonectin also enhances collagen gel contraction by fibroblasts isolated from type I collagen-deficient mice (Iruela-Arispe et al., 1996).
Previous studies of osteonectin expression have emphasized its presence in tissues that exhibit high rates of proliferation and morphological changes involving extracellular matrix turnover such as that during adult wound repair (Reed et al., 1993). Amongst the in-vitro actions of this protein (Lane and Sage, 1994; Yan and Sage, 1999) many are indeed consistent with its role in rapid active turnover of the extracellular matrix. It inhibits the synthesis of certain components of the extracellular matrix such as laminin, fibronectin and thrombospondin (Kamihagi et al., 1994), and increases the secretion of matrix metalloproteinase-1, -2, -3, -7 and -9 (Tremble et al., 1993; Shankavaram et al., 1997; Gilles et al., 1998) and plasminogen activator inhibitor (PAI-1) (Hasselaar et al., 1991). The anatomical region of the fetal membrane exhibiting altered morphology, i.e. the lower uterine pole, has been suggested to represent a site of structural weakness present at term prior to clinical labour. The dispersal of the fibrillar collagen type I/III-rich matrices, reflected by the thickened connective tissue layer in this region (Bell and Malak, 1997), is consistent with their degradation and/or increased turnover. The association of high osteonectin expression within connective tissue layers of this region, particularly within the chorion, and the correlation between osteonectin expression and layer thickness, suggests that these regions may be sites of active extracellular matrix remodelling mediated, in part, by the local action of osteonectin. The implication of this interpretation would be that the low osteonectin expression in the connective tissue layers of the mid-zone would be reflected by low extracellular matrix turnover and membrane stability in the upper uterine segment. The levels of the major enzyme activities involved in the intracellular processing and extracellular cross-linking of collagens indicates that the major active phase of synthesis and turnover of collagens by the mesenchymal cells of these layers in whole membranes is restricted to the first half of pregnancy (Casey and MacDonald, 1996, 1997). It will be of interest to determine whether osteonectin expression is detectable in the mesenchymal layers of the amnion and/or chorion during this early phase of pregnancy and indeed whether up-regulation of these enzymes occurs in the lower uterine pole at term in parallel with osteonectin expression.
Osteonectin expression was also detected within cytotrophoblasts, and in the mid-zone biopsies positive cells were concentrated at the maternal aspect adjacent to the decidua. In the regions of altered morphology, 6-9-fold higher percentages of immunoreactive cells were recorded. Since a characteristic of these regions is a thin cytotrophoblastic layer, this could simply reflect the smaller numbers of osteonectin negative cells beneath the interface. However, absolute numbers of immunoreactive cells were higher in this region and positive cells were apparent adjacent to the pseudobasement membrane in the basal aspect of the layer suggesting that there was additional expression induced at these sites. Whether osteonectin expression is a reflection of a function of these cells during differential attachment of the cytotrophoblastic layer to the decidua during early pregnancy and/or of differential breakdown of the maternal-fetal interface at these sites remains to be investigated. In either role, its expression may again relate to aspects of extracellular matrix stability and turnover at the maternal-fetal interface.
Despite the intracellular immunoreactivity, no osteonectin was ever detected in the extracellular matrix. This was consistent with previous reports of its almost exclusive intracellular localization in other tissues (Sage et al., 1989; Porter et al., 1995) in spite of evidence supporting its secretion and its putative functions in the extracellular matrix. This paradox has been suggested to be due to either its susceptibility to extracellular degradation or to binding in the extracellular matrix resulting in epitope masking. Osteonectin is readily degraded by cathepsins, neutral metalloproteinases, elastases and serine proteases (Lane and Sage, 1994). If this occurs in vivo it could have significant functional consequences since the domains and sequences of the protein linked to specific activities can be conserved in proteolytically generated peptides, e.g. modulation of cell shape, modulation of proliferation metal binding, collagen binding etc. (Lane and Sage, 1994). Indeed in some instances the peptide may even be more active than the native protein, for example, endogenous protease- or MMP-cleaved osteonectin binds collagens at a higher affinity than does native osteonectin (Sasaki et al., 1997). Additionally they may possess activities not evidenced in the intact protein, as seen with the angiogenic KGHK motif (Motamed and Sage, 1997). Osteonectin may also be incorporated into the extracellular matrix. It has recently been reported that the immunoreactive negative basement membrane underlying the lens epithelium is a `virtual repository' for osteonectin, as revealed by immunoblotting of extracts (Yan and Sage, 1999). This may relate to the ability of osteonectin to bind collagens, especially type III, via its C-terminal extracellular calcium-binding module (Sage et al., 1989; Yan and Sage, 1999). Although both interpretations may apply to the connective tissues of the amnion and chorion, their rich fibrillar collagen type I and III matrices (Bell and Malak, 1997; Bryant-Greenwood, 1998) and the lack of evidence of any degradation of the readily extractable osteonectin may support an analogy to the situation in the eye (Yan and Sage, 1999).
Further studies are required to determine the nature of osteonectin produced by the mesenchymal cells at these anatomical areas of the fetal membranes, particularly its regulation and function. This will determine whether it has a role in the generation of the morphological features at the site linked with labour induced rupture, and whether its abnormal expression may be associated with situations where the fetal membrane rupture occurs prior to labour.
Acknowledgements
We wish to thank Dr L.W.Fisher (NIDCR, NIH, Bethesda, MA, USA) who generously donated the immunoreagent described in Materials and methods. We also thank Mrs S.Spurling and J.Ashmore for expert technical assistance and the clinical members of the staff of the University Hospitals of Leicester NHS Trust for assistance with the collection of specimens. P.C.McParland was supported by a grant from Tommy's Campaign.
Notes
3 To whom correspondence should be addressed. E-mail: scb{at}leicester.ac.uk 
References
Alger, L.S. and Pupkin, M.J. (1986) Etiology of preterm premature rupture of the membranes. Clin. Obstet. Gynecol., 29, 758–770.[Web of Science][Medline]
Bell, S.C. (2000) Mechanisms underlying prelabour rupture of the fetal membranes. In Kingdom, J., Jauniaux, E. and O'Brian, S. (eds), The Placenta: Basic Sciences and Clinical Practice. RCOG Press, London. pp. 187–204.
Bell, S.C. and Malak, T.M. (1997) Structural and cellular biology of the fetal membranes. In Romero, R., Elder, M.G. and Lamont, R.F. (eds), Preterm Labor. Churchill Livingstone, New York and London, pp. 401–428.
Blazejewski, S., Le Bail, B., Boussarie, L. et al. (1997) Osteonectin (SPARC) expression in human liver and in cultured human liver myofibroblasts. Am. J. Pathol., 151, 651–657.[Abstract]
Bou-Resli, M.N., Al-Zaid, N.S. and Ibrahim, M.E.A. (1981) Full-term and prematurely ruptured fetal membranes. An ultrastructural study. Cell Tiss. Res., 220, 263–278.[Web of Science][Medline]
Bryant-Greenwood, G.D. (1998) The extracellular matrix of the human fetal membranes: structure and function. Placenta, 19, 1–11.[Web of Science][Medline]
Casey, M.L. and MacDonald, P.C. (1996) Interstitial collagen synthesis and processing in human amnion: a property of the mesenchymal cells. Biol. Reprod., 55, 1253–1260.[Abstract]
Casey, M.L. and MacDonald, P.C. (1997) Lysyl oxidase (ras recision gene) expression in the human amnion: ontogeny and cellular localization. J. Clin. Endocrinol. Metab., 82, 167–172.
Desmoulière, A. (1995) Factors influencing myofibroblast differentiation during wound healing and fibrosis. Cell Biol. Int., 19, 471–476.[Web of Science][Medline]
Fisher, L.W., Stubbs, J.T. and Young, M.F. (1995) Antisera and cDNA probes to human and certain animal model bone matrix non-collagenous proteins. Acta Orthop. Scand., 61, 61–65.
French, J.I. and McGregor, J.A. (1996) The pathobiology of premature rupture of membranes. Semin. Perinatol., 20, 344–368.[Web of Science][Medline]
Gibbs, R.S., Romero, R., Hillier, S.L. et al. (1992) A review of premature birth and subclinical infection. Am. J. Obstet. Gynecol., 166, 1515–1528.[Web of Science][Medline]
Gilles, C., Bassuk, J.A., Pulyaeva, H. et al. (1998) SPARC/osteonectin induces matrix metalloproteinase 2 activation in human breast cancer cell lines. Cancer Res., 58, 5529–5536.
Hasselaar, P., Loskutoff, D.J., Sawdey, M. et al. (1991) SPARC induces the expression of type 1 plasminogen activator inhibitor in cultured bovine aortic endothelial cells. J. Biol. Chem., 266, 13178–13184.
Holland, P.W., Harper, S.J., McVey, J.H. et al. (1987) In vivo expression of mRNA for the Ca++-binding protein SPARC (osteonectin) revealed by in situ hybridization. J. Cell Biol., 105, 473–482.
Iruela-Arispe, M.L., Vernon, R.B., Wu, H. et al. (1996) Type I collagen-deficient Mov-13 mice do not retain SPARC in the extracellular matrix: implications for fibroblast function. Dev. Dyn., 207, 171–183.[Web of Science][Medline]
Kamihagi, K., Katayama, M., Ouchi, R. et al. (1994) Osteonectin/SPARC regulates cellular secretion rates of fibronectin and laminin extracellular matrix proteins. Biochem. Biophys. Res. Commun., 200, 423–428.[Web of Science][Medline]
Keirse, M.J., Ohlsson, A., Treffers, P. et al. (1989) Prelabour rupture of the membranes at preterm. In Chalmers, I., Enkin, M. and Keirse, M.J. (eds), Effective Care in Pregnancy and Childbirth. Oxford University Press, Oxford, pp. 666–693.
Kelly, T. (1995) The pathophysiology of premature rupture of the membranes. Curr. Opin. Obstet. Gynecol., 7, 140–145.[Web of Science][Medline]
Khong, T.Y., Lane, E.B. and Robertson, W.B. (1986) An immunocytochemical study of fetal cells at the maternal-fetal interface using monoclonal antibodies to keratins, vimentin and desmin. Cell Tissue Res., 246, 189–195.[Web of Science][Medline]
Lane, T.F. and Sage, E.H. (1994) The biology of SPARC, a protein that modulates cell-matrix interactions. FASEB J., 8, 163–173.[Abstract]
Lane, T.F., Iruela-Arispe, M.L. and Sage, E.H. (1992) Regulation of gene expression by SPARC during angiogenesis in vitro. Changes in fibronectin, thrombospondin-1, and plasminogen activator inhibitor-1. J. Biol. Chem., 267, 16736–16745.
Major, C.A. and Garite, T.J. (1997) Preterm premature rupture of membranes. In Romero, R., Elder, M.G. and Lamont, R.F. (eds), Preterm Labor. Churchill Livingstone, New York and London, pp. 153–164.
Malak, T.M. and Bell, S.C. (1993) The structural characteristics of spontaneously ruptured fetal membranes: their relationship to membrane rupture and parturition. Contemp. Rev. Obstet. Gynecol., 5, 117–123.
Malak, T.M. and Bell, S.C. (1994) Structural characteristics of term human fetal membranes (amniochorion and decidua): a novel zone of extreme morphological alteration within the rupture site. Br. J. Obstet. Gynaecol., 101, 375–386.[Web of Science][Medline]
Malak, T.M. and Bell, S.C. (1996) Fetal membranes structure and prelabour rupture. Fetal Mat. Med. Rev., 8, 143–164.
McLaren, J., Malak, T.M. and Bell, S.C. (1999) Structural characteristics of term human fetal membranes prior to labour: identification of an area of altered morphology overlying the cervix. Hum. Reprod., 14, 237–241.
McParland, P.C., Taylor, D.J. and Bell, S.C. (2000) Myofibroblast differentiation in the connective tissues of the amnion and chorion of term human fetal membranes: implications for fetal membrane rupture and labour. Placenta, 21, 44–53.[Web of Science][Medline]
Mishima, H., Hibino, T., Hara, H. et al. (1998) SPARC from corneal epithelial cells modulates collagen contraction by keratocytes. Invest. Ophthalmol. Vis. Sci., 39, 2547–2553.
Motamed, K. (1999) SPARC (osteonectin/BM-40) Int. J. Biochem. Cell Biol., 31, 1363–1366.[Web of Science][Medline]
Motamed, K. and Sage, E.H. (1997) Regulation of vascular morphogenesis by the matricellular protein SPARC. Kidney Int., 51, 1383–1387.[Web of Science][Medline]
Mundlos, S., Schwahn, B., Reichert, T. et al. (1992) Distribution of osteonectin mRNA and protein during human embryonic and fetal development. J. Histochem. Cytochem., 40, 283–291.[Abstract]
Nakamura, S., Kamihagi, K., Satakeda, H. et al. (1996) Enhancement of SPARC (osteonectin) synthesis in arthritic cartilage. Arth. Rheum., 39, 539–551.[Web of Science][Medline]
Nemeth, E., Millar, L.K. and Bryant-Greenwood, G. (2000) Fetal membrane distension. II. Differentially expressed genes regulated by acute distension in vitro. Am. J. Obstet. Gynecol., 186, 60–67.
Parry, S. and Strauss, J.F. (1998) Premature rupture of the fetal membranes. New Eng. J. Med., 338, 663–670.
Pichler, R.H., Hugo, C., Shankland, S.J. et al. (1996) SPARC is expressed in renal interstitial fibrosis and in renal vascular injury. Kidney Int., 50, 1978–1989.[Web of Science][Medline]
Porter, P.L., Sage, E.H., Lane, T.F. et al. (1995) Distribution of SPARC in normal and neoplastic human tissue. J. Histochem. Cytochem., 43, 791–800.[Abstract]
Powell, D.W., Mifflin, R.C., Valentich, J.D. et al. (1999) Myofibroblasts I. Paracrine cells important in health and disease. Am. J. Physiol., 277, C1-C19.
Pringle, J.H. (1995) Non-isotopic detection of RNA in situ. In Levey E.R. and Herrington, C.S. (eds), Non-isotopic Methods in Molecular Biology: A Practical Approach. IRL Press, Oxford, UK, pp. 85–110.
Reed, M.J., Puolakkainen, P., Lane, T.F. et al. (1993) Differential expression of SPARC and thrombospondin 1 in wound repair: immunolocalization and in situ hybridization. J. Histochem. Cytochem., 41, 1467–1477.[Abstract]
Reed, M.J., Vernon, R.B., Abrass, I.B. et al. (1994) TGF-beta 1 induces the expression of type I collagen and SPARC, and enhances contraction of collagen gels, by fibroblasts from young and aged donors. J. Cell. Physiol., 158, 169–179.[Web of Science][Medline]
Romero, R., Athayde, N., Maymon, E. et al. (1999) Premature rupture of the membranes. In Reece, E.A. and Hobbins, J.C. (eds), Medicine of the Fetus and Mother, 2nd edn. Lippincott-Raven, Philadelphia, pp. 1581–1625.
Sage, E.H. and Bornstein, P. (1991) Extracellular proteins that modulate cell-matrix interactions. SPARC, tenascin, and thrombospondin. J. Biol. Chem., 266, 14831–14834.
Sage, H., Vernon, R.B., Decker, J. et al. (1989) Distribution of the calcium-binding protein SPARC in tissues of embryonic and adult mice. J. Histochem. Cytochem., 37, 819–829.[Abstract]
Sappino, A.P., Schurch, W. and Gabbiani, G. (1990) Differentiation repertoire of fibroblastic cells: expression of cytoskeletal proteins as marker of phenotypic modulations. Lab. Invest., 63, 144–161.[Web of Science][Medline]
Sasaki, T., Göhring, W., Mann, K. et al. (1997) Limited cleavage of extracellular matrix protein BM-40 by matrix metalloproteinases increases its affinity for collagens. J. Biol. Chem., 272, 9237–9243.
Schmitt-Gräff, A., Desmoulière, A. and Gabbiani, G. (1994) Heterogeneity of myofibroblast phenotypic features: an example of fibroblastic cell plasticity. Virch. Archiv, 425, 3–24.
Schurch, W., Seemayer, T. and Gabbiani, G. (1998) The myofibroblast: a quarter century after its discovery. Am. J. Surg. Pathol., 22, 141–147.[Web of Science][Medline]
Serini, G. and Gabbiani, G. (1999) Mechanisms of myofibroblast activity and phenotypic modulation. Exp. Cell Res., 250, 273–283.[Web of Science][Medline]
Shankavaram, U.T., DeWitt, D.L., Funk, S.E. et al. (1997) Regulation of human monocyte matrix metalloproteinases by SPARC. J. Cell. Physiol., 173, 327–334.[Web of Science][Medline]
Shiba, H., Fujita, T., Doi, N. et al. (1998) Differential effects of various growth factors and cytokines on the syntheses of DNA, type I collagen, laminin, fibronectin, osteonectin/secreted protein, acidic and rich in cysteine (SPARC), and alkaline phosphatase by human pulp cells in culture. J. Cell. Physiol., 174, 194–205.[Web of Science][Medline]
Termine, J.D., Kleinman, H.K., Whitson, S.W. et al. (1981) Osteonectin, a bone-specific protein linking mineral to collagen. Cell, 26, 99–105.[Web of Science][Medline]
Tremble, P.M., Lane, T.F., Sage, E.H. et al. (1993) SPARC, a secreted protein associated with morphogenesis and tissue remodeling, induces expression of metalloproteinases in fibroblasts through a novel extracellular matrix-dependent pathway. J. Cell Biol., 121, 1433–1444.
Wrana, J.L., Overall, C.M. and Sodek, J. (1991) Regulation of the expression of a secreted acidic protein rich in cysteine (SPARC) in human fibroblasts by transforming growth factor beta. Comparison of transcriptional and post-transcriptional control with fibronectin and type I collagen. Eur. J. Biochem., 197, 519–528.[Web of Science][Medline]
Yan, Q. and Sage, E.H. (1999) SPARC, a matricellular glycoprotein with important biological functions. J. Histochem. Cytochem., 47, 1495–1505.
Zhou, H., Hammonds, R.G.J., Findlay, D.M. et al. (1993) Differential effects of transforming growth factor-beta 1 and bone morphogenetic protein 4 on gene expression and differentiated function of preosteoblasts. J. Cell. Physiol., 155, 112–119.[Web of Science][Medline]
Submitted on December 29, 2000;
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  M. Adams, J. L. Jones, R. A. Walker, J. H. Pringle, and S. C. Bell Changes in Tenascin-C Isoform Expression in Invasive and Preinvasive Breast Disease Cancer Res., June 1, 2002; 62(11): 3289 - 3297. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||