Binding of interleukin-8 to heparan sulphate enhances cervical maturation in rabbits

Cervical ripening is a cytokine-trigged process with substantial remodelling of the cervical extracellular matrix. Interleukin-8 (IL-8) is an important cytokine in cervical maturation. Glycosaminoglycans are also included in this process, but their role in not clearly understood. The effects of heparan sulphate (HS), hyaluronic acid (HA), IL-8, HS (cid:70) IL-8 and HA (cid:70) IL-8 on biochemical properties of the cervix were examined in non-pregnant rabbits. The changes in vascular pattern with collagen structure of the cervices and immunohistochemical studies, together with the relative collagen concentrations, were determined. A reduction in relative collagen concentration was signiﬁcant after HS (cid:70) IL-8 , IL-8 and HA (cid:70) IL-8 treatment (all P < 0.0001). Gel electrophoresis analysis showed that IL-8 bound preferentially to HS than to HA. Neutrophils were signiﬁcantly increased in number (P < 0.0001) and located predominantly beneath the glandular epithelium and around the blood vessels after HS (cid:70) IL-8 treatment. HS (cid:70) IL-8 treatment caused cervices to increase their water content and become oedematous. The collagen ﬁbres were considerably dissociated, the interﬁbrillar spaces markedly dilated, and the blood vessels notably increased and dilated. We conclude that binding to HS enhances the activity of IL-8 in inducing cervical maturation. computer analyser system; ARGUS-100, Hamamatsu Photonics, Japan). The principle of picrosirius red staining is the greater the collagen concentration, the greater the birefrin-gence and hence the greater the percentage of light


Introduction
The non-pregnant cervix is a fibrous structure which undergoes modifications during pregnancy to allow sufficient softening and growth for the passage of the fetus at birth (Harkness and Harkness, 1959;Friedman, 1980;Calder, 1981;Fitzpatrick, 1981). It is well established that the transformation of the uterine cervix to a soft and compliant structure during pregnancy is crucial for a normal delivery. This ripening process can be explained mainly in terms of connective tissue biology (Huszar and Walsh, 1991). Investigations of the structural alterations of cervical connective tissue during pregnancy and labour have focused predominantly on quantitative changes of the collagen content and the role of collagen-degraded enzymes (Ito et al., 1979;Danforth, 1980;Uldbjerg et al., 1983). However, there are strong indications that other macromolecular components of the cervix, such as proteoglycans (PG), may also be involved (von Maillot et al., 1979;Kitamura et al., 1980). PG are core proteins with covalently attached glycosaminoglycan (GAG) chains (Ruoslahti, 1988). In addition to heparin/ heparan sulphate (HS), chondroitin sulphate/dermatan sulphate, keratan sulphate and hyaluronic acid (HA) have also been isolated from the human cervix (von Maillot et al., 1979;Cabrol et al., 1980;Kitamura et al., 1980;Uldbjerg et al., 1983). A marked increase in the GAG content of the cervix during late pregnancy has been demonstrated in rats, rabbits, sheep and humans (Danforth et al., 1974;Golichowski et al., 1980;Fosang et al., 1984;Maradny et al., 1997). In particular, the cervical content of HS increases markedly during late pregnancy. It has been shown that HS is associated with the cell surface, together with various glycoproteins, and thought to be located predominantly in vessel walls (Caplan and Hascall, 1980;Kitamura et al., 1980;Scott, 1986;Isemura et al., 1987). Consequently, an increase in this GAG may reflect improved cervical vascularization in the fully ripened cervix during labour (Rath et al., 1988).
Recently, several reports have emphasized the role of interleukin-8 (IL-8), which is a proinflammatory cytokine (also known as a chemokine) that is produced by a variety of cells, mainly monocytes/macrophages, but also by fibroblasts and choriodecidual cells (Uchiyama et al., 1992). IL-8 attracts and activates neutrophils and lymphocytes and stimulates the release of storage enzymes and toxic metabolites from neutrophils (Peveri et al., 1988;Kelly et al., 1994). IL-8 has been detected in increasing amounts from the amniotic fluid in term pregnancies, while its production has also been demonstrated in vitro in the human cervix (Romero et al., 1991;Barclay et al., 1993;Axemo et al., 1996). IL-8 is one of the factors which can ripen the cervix in a similar manner to the physiological process at term (Maradny et al., 1994). There is also evidence that IL-8 plays an important role in promoting cervical modification in late pregnancy (Barclay et al., 1993).
One class of molecules likely to play a role in mediating tissue-specific associations of chemokines is the GAG. Chemokines such as IL-8 are known to bind to GAG, especially heparin (Talpas et al., 1993), and it has been suggested that heparin-related (heparan sulphate) GAG on endothelial cell surfaces localize and present chemokines to selectin-bound leukocytes (Butcher, 1991). A recent study has demonstrated that chemokines may be bound to cell-surface heparan sulphate in vitro (Huber et al., 1991), while an earlier report suggested that HS-which is present on the endothelial cell surface and in the basement membrane-may promote IL-8-dependent transmigration of neutrophils and thus enhance the activity of IL-8 (Webb et al., 1993). Based on these reports it seems important to evaluate the binding effect of HS to IL-8 in the cervical ripening process. The purpose of this study was to determine the effect of HS as an enhancer of IL-8 activity on the cervical ripening.

Materials and methods
Forty sexually mature Japanese White female non-pregnant rabbits (SLC, Hamamatsu, Japan) of bodyweight 3.0-3.5 kg were used. Rabbits were caged under controlled light and temperature and were given Purina rabbit chow (Clea Japan Inc., Tokyo, Japan) and water ad libitum. The rabbits were comparable in terms of age and bodyweight. Vaginal suppositories were prepared with 500 µl Witepsol-50 bases (WT-50, Adeps solidus: Mitsuba Co., Tokyo, Japan), the products of cocoa butter. The suppositories thus prepared were conical-shaped, quick to melt at 37°C, and showed poor attachment to the tissues. Release of drug from the suppository, when measured by the recommended method (Shintani, 1975), was found to occur rapidly. All animal experiments described were approved by the Research Committee of Laboratory Animals of Hamamatsu University School of Medicine.

Experimental design
Three approaches were used in this protocol. In the first approach, experiments were performed to determine the binding affinity of HS or HA to IL-8 using gel electrophoresis. In the second, the rabbits were treated either with placebo or recombinant IL-8, HS, HA, HS ϩ IL-8 and HA ϩ IL-8 to investigate the effects on cervical maturation. In the third approach, experiments were conducted to assess the dose-dependent effect of HS binding to IL-8 on cervical ripening.

Binding effects of HS or HA to IL-8 on cervical maturation
A series of five experiments was performed in rabbits (n ϭ 5 per group). Suppositories were administered vaginally once daily for 3 days to each animal to examine the effect of: (i) IL-8 on cervical maturation (suppositories contained 200 ng IL-8); (ii) HS on cervical growth (suppositories contained 1 mg HS); (iii) HS ϩ IL-8 on cervical growth (suppositories contained 1 mg HS ϩ 200 ng IL-8); (iv) HA on cervical dilatation (suppositories contained 1 mg HA); and (v) HA ϩ IL-8 on cervical ripening (suppositories contained 1 mg HA ϩ 200 ng IL-8). A control experiment was performed in five rabbits given suppositories containing 500 µl-WT base given once daily for 3 days.

Dose-dependent effect of HS binding to IL-8 on cervical maturation
Rabbits were allocated to two groups (n ϭ 5 per group) and treated once daily for 3 days with suppositories containing either 0.1 mg HS ϩ 200 ng IL-8 or 1 mg HS ϩ 200 ng IL-8. Animals were killed 24 h after receipt of the last suppository. The reproductive tract of each animal was immediately located and the cervices excised. One cervix from each animal was used to measure water content, followed by measurement of collagenase, elastase and gelatinase activities; the other cervix was fixed in 10% buffered formalin for 24 h. Following paraffin embedding (Technovit 7100; Kulzer GmbH, Germany), blocks were serially sectioned at a thickness of 5 µm using a Reidhert Jung 2050 motorized automatic microtome. Sections were stained with either haematoxylin followed by a 1% eosin counterstain (H/E), or with picrosirius red. Measurement of water content, collagenase, elastase and gelatinase activity, neutrophil infiltration in the cervical tissue and histological changes were performed to assess cervical maturation.

Determination of water content
The water content of each cervix was measured using an IM-3SCV device (Fuji Technica Co., Osaka, Japan), which is an infrared spectrophotometric technique that compares absorbance at a wavelength of 1.45 nm with that at reference wavelengths of 1.3 µm and 1.6 µm (Sumimoto and Terau, 1993). Five different points on the cervix were measured and the mean was calculated.

Measurement of collagenase, elastase and gelatinase activity
Cervices were homogenized in 1 ml ice-cold phosphate-buffered solution (PBS), pH 7.6. After extraction twice by freeze-thawing, samples were sonicated (30 W, 120 pulses, 30% duty, W-220 type; Heat Systems Ultrasonics, NY, USA) and centrifuged at 10 000 g for 20 min at 4°C. The supernatant was used to measure collagenase elastase and gelatinase activity as described previously (Maradny et al., 1997). Collagenase activity was estimated using a highly specific kit (collagenase type 1 activity measurement; Yagai Co., Cosmo-Bio, Tokyo, Japan), whereas elastase activity was determined by a specific chromogenic substrate for granulocyte elastase S-2484 (L-pyroglutamyl-L-prolyl-L-valine-p-nitranilide; KABI Diagnostic, Mölndal, Sweden). Gelatinase activity was measured using highly specialized kits (gelatinase activity measurement; Yagai Co.). One unit of activity was defined as the quantity of enzyme that digested 1 mg in 1 min.

Immunohistochemistry of the cervix
Paraffin-embedded tissue was deparaffinized in xylene baths, rehydrated through graded 95% alcohol, and finally rinsed in PBS, pH 7.2. Endogenous peroxidase was blocked by fixation in 3% hydrogen peroxide in methanol for 20 min at 23°C. Bovine serum albumin (2%) dissolved in PBS (BSA-PBS) was applied to the slides for 1 h. Anti-rabbit RT2 (1:100) monoclonal antibodies (Cedarlane Laboratories Ltd, Hornby, Canada) were added to the sections and incubation continued overnight at 4°C, followed by PBS rinsing (Ponsard et al., 1986). The secondary antibody (goat anti-rabbit; DAKO, California, USA) was applied for 2 h at room temperature followed by a PBS rinse. Avidin-biotin-peroxidase complex (DAKO) was added for 1 h, followed by washing with PBS. The sections were counterstained in haematoxylin, dehydrated and then examined under light microscopy (ϫ200 magnification). Negative controls received the same treatment, but using non-immune mouse serum instead of primary antibody. Neutrophils in the tissue were used as a positive control, the number of neutrophils being estimated by counting the number of stained extravascular cells within a lined grid (10ϫ10 squares) occupying an area on the section of 0.125 mm 2 using a ϫ20 objective and ϫ10 eyepiece. Neutrophils were counted in one cervix from each animal (five to seven randomly chosen areas) and the mean was calculated.

Assay of relative collagen concentration
The relative collagen concentration was assessed histologically after staining with picrosirius red (Sirius red F3BA; Chroma-Gesellschaft Schmid GmbH, Köngen, Germany) as described previously and validated as a histological method to determine the polymerized collagen concentration of tissues, including the cervix (Junqueira et al., 1979). The histological analysis was performed by measuring the optical density (percentage polarized light transmission) from five random fields of the connective tissue of each biopsy and the mean optical density calculated. An image analyser was employed for all histological measurements (microscope; Olympus IMT-2, Videocamera SIT C2400-80: computer analyser system; ARGUS-100, Hamamatsu Photonics, Japan). The principle of picrosirius red staining is that, the greater the collagen concentration, the greater the birefringence and hence the greater the percentage of light transmission.

Statistical analysis
Data are expressed as mean Ϯ SD. Analysis of variance (ANOVA) for repeated measurement followed by Scheff's F analysis was used for multiple comparisons. A P value of Ͻ 0.05 was considered significant for all comparisons. Figure 1a shows the agarose electrophoretic pattern of IL-8 (band 1), HS (band 2), HA (band 3), HS (1 mg) ϩ IL-8 (band 4), HA (1 mg) ϩ IL-8 (band 5); Figure 1b shows patterns for HS (0.05 mg) ϩ IL-8 (band 1), HS (0.1 mg) ϩ IL-8 (band 2) and HS (1 mg) ϩ IL-8 (band 3). The molecular weights of HS, HA and IL-8 are 400, 800 and 8 kDa, respectively. Therefore, 1 nmol of HS has bound~0.5 nmol of IL-8 under our experimental conditions. The binding of HA to IL-8 was weak compared with that of HS. Moreover, 0.05 mg and 0.1 mg of HS and HA respectively were insufficient to bind 10 µg IL-8 (Figure 1b, bands 1 and 2). The migration of HSϩIL-8 was delayed compared with that of HA ϩ IL-8, and demonstrates that IL-8 exhibits substantial selectivity in its binding to HS.

Cervix water content
The influence of various treatments on cervix water content is shown in Figure 2. When compared with controls, the water The affinity of 1 mg HS to IL-8 was markedly higher than that of 1 mg HA. The movement of HS (1 mg) ϩ IL-8 (band 4) was delayed compared with HA (1 mg) ϩ IL-8 (band 5), and demonstrates that IL-8 exhibits substantial selectivity in its binding to HS. Low concentrations (0.05 and 0.1 mg) of HS were insufficient to bind IL-8 (10 µg).

Figure 2.
Water content (%) in rabbit cervices treated with HS ϩ IL-8, HA ϩ IL-8, IL-8, HS, HA, and in controls. Compared with controls, the water content was significantly increased with HS ϩ IL-8, HA ϩ IL-8 and IL-8 (P Ͻ 0.001, Ͻ 0.003 and Ͻ 0.01, respectively). The content in cervices treated with HS (P Ͻ 0.05) and HA (P Ͻ 0.05) was significantly increased compared with controls. Percent hydration was higher in cervices treated with HS ϩ IL-8 compared with IL-8 and HA ϩ IL-8. *, significantly different compared with controls; †, significantly different compared with HA treatment (n ϭ 5, P Ͻ 0.05). content was significantly increased in rabbits treated with IL-8, HS and HA (P Ͻ 0.01, Ͻ 0.05 and Ͻ 0.05 respectively), but was more significantly increased when the cervix was treated with HS ϩ lL-8 (P Ͻ 0.001) and HA ϩ IL-8 (P Ͻ 0.003). Notably, the cervix water content was higher in HS ϩ IL-8-than HA ϩ IL-8-treated rabbits.
The cervix water content of rabbits treated with 200 ng IL-8, 0.1 mg HS ϩ 200 ng IL-8 and 1 mg HS ϩ 200 ng IL-8 is shown in Figure 3. Compared with controls, water content was significantly increased in animals given 200 ng IL-8 (P Ͻ 0.009), 0.1 mg HS ϩ 200 ng IL-8 (P Ͻ 0.009) and 1 mg HS ϩ 200 ng IL-8 (P Ͻ 0.002). Cervical water content was greater in rabbits given 1 mg HS ϩ 200 ng IL-8 than in those given either 0.1 mg HS ϩ 200 ng IL-8 (P Ͻ 0.01) or 200 ng IL-8 alone (P Ͻ 0.05), though no significant difference was found between the latter two groups.

Number and localization of neutrophils
Increased numbers of neutrophils were found in cervical biopsies from rabbits treated with either HS or HA (P Ͻ 0.004  Compared with controls, the number of neutrophils was significantly increased in cervices treated with HS ϩ IL-8, HA ϩ IL-8 and IL-8 (P Ͻ 0.0001, Ͻ 0.0001 and Ͻ 0.0001, respectively). Increased numbers of neutrophils were found to be more significant in cervices treated with HS ϩ IL-8 (P Ͻ 0.05) than with HA ϩ IL-8. Cervices treated with HS and HA also showed significant increases in neutrophils (P Ͻ 0.004 and Ͻ 0.01) compared with controls. *, significantly different compared with control; †, significantly different compared with HA treatment (n ϭ 5, P Ͻ 0.05).
or Ͻ 0.01) compared with controls ( Figure 4). A remarkable increase in neutrophil number was seen in rabbits that had received HS ϩ IL-8, HA ϩ IL-8 and IL-8 by suppository (P Ͻ 0.0001, Ͻ 0.0001 and Ͻ 0.0001) compared with controls. Neutrophils were also increased in IL-8-treated cervices compared with HS (P Ͻ 0.05) or HA (P Ͻ 0.02). HS ϩ IL-8 treatment resulted in a greater increase in neutrophils than did HA ϩ IL-8 treatment (P Ͻ 0.05). Figure 5 shows changes in the number of neutrophils identified by immunohistochemical staining. Stained neutrophils were located predominantly around the blood vessels and beneath the glandular epithelium (not shown in figure) in cervices treated with HS ϩ IL-8 and IL-8 (Figure 5a and b) and HA ϩ IL-8 (not shown), compared with controls ( Figure   264 5d). Neutrophils were abundantly located in cervices treated with HS ϩ IL-8 compared with IL-8 alone. The overall number of neutrophils was considerably reduced with both HS ( Figure  5c) and HA treatment (not shown).

Relative collagen concentrations
As shown in Figure 6, the relative collagen concentration in the HS ϩ IL-8 treatment group was reduced by 69.2% (P Ͻ 0.0001) compared with controls, whereas HA ϩ IL-8 treatment reduced collagen by 51.8% (P Ͻ 0.0001) and IL-8 treatment by 49.9% (P Ͻ 0.001). The collagen concentration in rabbits treated with HS ϩ IL-8 was significantly reduced compared with HA ϩ IL-8 treatment (P Ͻ 0.03), while concentrations in rabbits treated with HS and HA were each also significantly reduced compared with controls (P Ͻ 0.002 and Ͻ 0.01).

Collagenase, gelatinase and elastase activities
Changes in cervical collagenase, gelatinase and elastase activities are summarized in Table I. Collagenase activities were significantly increased in rabbits treated with HS ϩ IL-8, HA ϩ IL-8 and IL-8 (P Ͻ 0.0001, Ͻ 0.0001 and Ͻ 0.0001, respectively), compared with controls. Collagenase activity in HS ϩ IL-8-treated rabbits was markedly increased compared with that in rabbits treated with HA ϩ IL-8 (P Ͻ 0.04), as was activity in HS-and HA-treated rabbits (P Ͻ 0.005 and Ͻ 0.02). Gelatinase activities were significantly increased in cervices treated with HS ϩ IL-8, HA ϩ IL-8 and IL-8 (P Ͻ 0.0001, Ͻ 0.0001 and Ͻ 0.0001, respectively), compared with controls. However, gelatinase activities in HS ϩ IL-8treated cervices were markedly increased compared with HA ϩ IL-8-treated tissues (P Ͻ 0.04). The difference in activity of cervical gelatinase was also statistically significant when the HS-and HA-treated rabbits were compared with controls (P Ͻ 0.002 and Ͻ 0.03). Cervical granulocyte elastase activities in HS ϩ IL-8-, HA ϩ IL-8-and IL-8-treated rabbits were significantly increased (P Ͻ 0.0001, Ͻ 0.0001 and Ͻ 0.0001, respectively), compared with controls. Furthermore, this activity was significantly greater in animals treated with HS ϩ IL-8 than with HA ϩ IL-8 (P Ͻ 0.009). Elastase activities were increased in rabbits treated with either HS or HA (P Ͻ 0.002 and Ͻ 0.04) when compared with controls.

Collagen study with picrosirious red staining
Examination of picrosirious red-stained sections of cervices disclosed various features (Figure 8). In HS ϩ IL-8-treated cervices, the collagen fibres appeared to be much thinner and more spread out, but were irregular with less compact and low-density collagen. The collagen fibres were not oriented in Downloaded from https://academic.oup.com/molehr/article-abstract/5/3/261/1404588 by guest on 27 July 2018  Concentrations were significantly reduced with HS ϩ IL-8 (P Ͻ 0.0001), HA ϩ IL-8 (P Ͻ 0.0001) and IL-8 (P Ͻ 0.0001), compared with controls. Collagen concentrations in rabbits after HS ϩ IL-8 treatment was markedly reduced compared with HA ϩ IL-8-treated animals (P Ͻ 0.03). Cervices treated with HS and HA showed significant reductions in collagen concentration (P Ͻ 0.002 and Ͻ 0.01) compared with controls. *, significantly different compared with control; †, significantly different compared with HA treatment (n ϭ 5, P Ͻ 0.05).
an orderly fashion, were irregularly separated one from another, and the interfibrillar spaces were markedly dilated due to oedematous changes (Figure 8a). Comparatively contained orderly fibres were observed in cervices treated with IL-8 ( Figure 8b) and HA ϩ IL-8 (not shown). No significant changes in collagen structure were found in rabbits treated with HS (Figure 8c). In control cervices (Figure 8d), the collagen fibres visualized by picrosirius red staining were well organized and densely packed in bundles. The collagen fibres were regularly separated one from another and oriented in an orderly fashion.

Morphology with haematoxylin and eosin staining
Remarkable changes cervical histology after haematoxylin and eosin staining were found in rabbits after various treatments (Figure 9). HS ϩ IL-8 treatment (Figure 9a) resulted in the cervical lumens being dilated and the diameters of the whole cervices increased due to massive dilation of the blood vessels. The cervix walls were thin and the density of collagenous Figure 8. Picrosirius red-stained sections of rabbit cervices treated with HS ϩ IL-8, IL-8, HS, and in controls. In HS ϩ IL-8-treated cervices (a), the collagen fibres appeared to be much thinner and less compact, and a low density of collagen was observed. Comparatively more defined fibres were apparent in cervices treated with . No significant changes were observed in cervices treated with HS (c). In control cervices (d), the collagen fibres were well organized and densely packed as bundles. (Picrosirius staining; scale bar ϭ 0.05 mm).

Figure 9.
Histological findings in rabbit cervices treated with HS ϩ IL-8, IL-8, HS, and in controls. In HS ϩ IL-8-treated cervices, the cervical lumens are dilated due to massive dilation of blood vessels (a). Cervices in control rabbits (d) were thick-walled, with the collagenous network lying close together. The connective tissues were compact and the blood vessels small and non-dilated. More subtle histological changes were seen in animals treated with IL-8 (b); no remarkable changes were seen in HS-treated cervices (c). (Haematoxylin and eosin staining; scale bar ϭ 0.05 mm). network was decreased and markedly loosened. The cervices of control rabbits (Figure 9d) showed dense and firmly closed cervical rings, with the collagenous networks in the lamina propria (as well as of the inner circular and outer longitudinal smooth muscle layers) close together. The connective tissues were compact and the blood vessels small and non-dilated. More subtle changes were observed after treatment with IL-8 ( Figure 9b) and HA ϩ IL-8 (not shown) than after HS ϩ IL-8. No significant changes were found after HS treatment alone (Figure 9c).

Discussion
This study demonstrated that the binding of IL-8 to HS enhances its activity in cervical connective tissue in rabbits.
The most striking finding of the present study was that the changes in biochemical components in the cervix of rabbits were markedly influenced by HS ϩ IL-8. In cervices treated with HS ϩ IL-8, neutrophils, water content and collagenase, gelatinase and elastase activities were significantly increased, whereas the relative collagen concentration in the cervical fluid was markedly reduced. Moreover, in binding to IL-8, HS also influenced the growth and vascular pattern of the cervices in rabbits. More subtle changes in cervical connective tissue components, together with histological findings, were observed in rabbits treated with HA ϩ IL-8. HS has an important role in the structural alterations of the cervical connective tissue during pregnancy and parturition. HS is thought to be located predominantly in vessel walls (Caplan and Hascall, 1980), and dramatically increases during dilatation of the cervix at labour (Kitamura et al., 1980;Osmers et al., 1995). The increased content of HS may also play a role in dispersal of collagen fibrils, since they have been shown to bind to collagen and may prevent fibril growth (Obrink, 1973;Scott and Orford, 1981).
There is growing evidence that a cytokine network is present in gestational tissues and plays an important role during preterm and term parturition. Various uterine and embryonic cells, including the decidua and chorion (Kelly et al., 1994) and cervix fibroblasts (Uchiyama et al., 1992), are capable of producing inflammatory cytokines such as IL-8, IL-1 and tumour necrosis factor-alpha (TNF-α). These cytokines were also shown to increase the production of both IL-8 mRNA and IL-8 in endometrial stromal cells in vitro (Arici et al., 1996). It is well known that IL-8 is a potent chemotactic 267 activator of neutrophils and induces the migration of neutrophils from the vessels to surrounding connective tissue (Peveri et al., 1988). Previous studies have shown that the application of inflammatory cytokines IL-8, IL-1β and TNF-α induce cervical ripening (Chwalisz et al., 1994;Maradny et al., 1994Maradny et al., , 1996, the effects of IL-8 having been shown to be more specific for the uterine cervix (Chwalisz et al., 1994).
HS did not induce random or directed neutrophil migration and did not influence neutrophil activity; however, enhanced responsiveness resulted from its selective interaction with IL-8 (Webb et al., 1993). HS has a function in the promotion of IL-8-dependent transmigration of neutrophils across the endothelial barrier (Webb et al., 1993). It has been shown previously that IL-8 binds endothelial cells, possibly via surface proteoglycans, and it has been suggested that this interaction is important for IL-8-dependent neutrophil migration (Rot, 1992a,b). The C terminus of IL-8 has been implicated in heparin binding (Talpas et al., 1993), and also contains several basic residues. An acidic residue within a HS-binding site might be expected to increase the affinity for IL-8 (Witt and Lander, 1994). Gallagher and colleagues have shown that HS molecules have a domain structure with a higher degree of sulphation that permits tighter binding to IL-8 (Gallagher et al., 1986;Turnbull and Gallagher, 1991). It has been found that the backbone structure, sulphation and the arrangement of the charges are important in determining the capacity of GAG to bind to IL-8 and to modulate its activity on the cells. The high heterogeneity in HS structure may allow a more refined tailoring of selective binding regions that may influence the biological activity and bioavailability of HS-binding chemokines (Presta et al., 1998).
Decreased collagen concentrations and increased collagenase, gelatinase and elastase activities were found in rabbits treated with HS ϩ IL-8. Furthermore, an increased water content in cervices after HS ϩ IL-8 treatment accounts for the soft, swollen and fragile consistency of the ripened cervix (Osmers et al., 1995). An increased number of neutrophils was identified in cervices treated with HS ϩ IL-8. In binding to IL-8, HS could play a key role in increasing the specificity to the activation step of neutrophil recruitment. Gel electrophoresis analysis indicated that IL-8 migrated as a coherent band at all protein concentrations. The failure of the HA front to form a tight band suggests binding heterogeneity; that is, a fraction of HS has substantially higher affinity to IL-8. This result demonstrates that IL-8 exhibits substantial selectivity in its binding to HS (Witt and Lander,1994). Moreover, IL-8 binds HA considerably less strongly than it binds HS.
Our findings demonstrate that HA is an important factor in the process of cervical ripening and regulates the biochemical changes occurring in cervical tissues at term (Maradny et al., 1997). It was shown recently that the interactions of decorin (a small, sulphated proteoglycan) and collagen may be important for cervical dilatation. It has been suggested that cervical ripening is mainly due to changes in the decorin/collagen ratio, while increased hydration is due to HA (Rechberger et al., 1996). Nitric oxide (NO), a potent uterine relaxant, represents a powerful autocrine and/or paracrine mediator of cervical ripening. It is therefore likely that NO may be a component in the pathway of cervical ripening which acts in concert with prostaglandins (mainly prostaglandin E 2 ) by activating metalloproteinases and other molecules involved in extracellular matrix remodelling and modulation of proteoglycan synthesis (Chwalisz et al., 1997). The present findings show that HS ϩ IL-8-induced changes in the biochemical composition and physical properties of the cervix tend to be much more dramatic. The binding effects of IL-8 to HS persists for a long time due to the long-chain structure of HS; moreover, the receptor expression of IL-8 to HS may be stronger. Therefore, HS ϩ IL-8 may be considered as complementary approaches to cervical ripening, and the action of HS ϩ IL-8 more pronounced in the uterine cervix.
In conclusion, this study indicates that, in rabbits, binding to HS enhances neutrophil responses to IL-8 and plays an important role in promoting a decrease in the collagen concentration and increases in neutrophil number, water content and collagenase, gelatinase and elastase activities through IL-8dependent neutrophil migration. Moreover, we suggest that the HS ϩ IL-8-induced changes in cervical connective tissue components may account-at least in part-for cervical maturation during a normal delivery.