Adrenomedullin suppresses tumour necrosis factor alpha-induced CXC chemokine ligand 10 production by human gingival fibroblasts

Authors

  • I. Hosokawa,

    1. Department of Conservative Dentistry, Institute of Health Biosciences, The University of Tokushima Graduate School, Tokushima, and
    Search for more papers by this author
  • Y. Hosokawa,

    Corresponding author
    1. Department of Conservative Dentistry, Institute of Health Biosciences, The University of Tokushima Graduate School, Tokushima, and
    Search for more papers by this author
  • K. Ozaki,

    1. Department of Oral Health Care Promotion, School of Oral Health and Welfare, Faculty of Dentistry, The University of Tokushima, Tokushima, Tokushima, Japan
    Search for more papers by this author
  • H. Nakae,

    1. Department of Conservative Dentistry, Institute of Health Biosciences, The University of Tokushima Graduate School, Tokushima, and
    Search for more papers by this author
  • T. Matsuo

    1. Department of Conservative Dentistry, Institute of Health Biosciences, The University of Tokushima Graduate School, Tokushima, and
    Search for more papers by this author

Dr Yoshitaka Hosokawa, Department of Conservative Dentistry, Institute of Health Biosciences, The University of Tokushima Graduate School, 3-18-15 Kuramoto-cho, Tokushima, 770-8504, Japan.
E-mail: hosokawa@dent.tokushima-u.ac.jp

Summary

Periodontal disease is an inflammatory disorder characterized by the involvement of chemokines that are important for the recruitment of leucocytes. Several cytokines, including tumour necrosis factor alpha (TNF-α), are involved in regulating levels of chemokines in periodontal disease. CXC chemokine ligand 10 (CXCL10) is a chemokine related to the migration of T helper 1 cells. In this study, we examined CXCL10 expression in human gingival fibroblasts (HGFs). Moreover, we investigated the effects of adrenomedullin (AM), which is a multi-functional regulatory peptide, on the production of CXCL10 by HGFs. We revealed that TNF-α stimulation induced CXCL10 production by HGFs. HGFs expressed AM and AM receptors, calcitonin-receptor-like receptor (CRLR) and receptor-activity-modifying protein (RAMP) 2, mRNAs constitutively. AM treatment supressed CXCL10 production by TNF-α-stimulated HGFs. Moreover, we elucidated that AM produced by HGFs inhibited CXCL10 production by HGFs, because AM antagonist enhanced CXCL10 production by HGFs. TNF-α treatment enhanced CRLR and RAMP2 mRNA expression in HGFs. Furthermore, AM is expressed in human periodontal tissues, including both inflamed and clinically healthy tissues. These results suggest that the CXCL10 produced by HGFs may be involved in the migration of leucocytes into inflamed tissues and related to exacerbation of periodontal disease. AM might be a therapeutic target of periodontal disease, because AM can inhibit CXCL10 production by HGFs.

Introduction

Periodontitis is a chronic bacterial infection of tooth-supporting structures. It causes destruction of periodontal connective tissues and bone. Oral plaque bacteria including Porphyromonas gingivalis, Actinobacillus actinomycetemcomitans and Tanerella forsythia are recognized as aetiological agents in periodontitis. The disease initiation and progression result from host response to plaque bacteria. Immunohistochemical studies reveal dense inflammatory cells infiltration, including T and B cells and macrophages in the periodontitis region. In addition, high levels of cytokines and chemokines such as interleukin (IL)-1β, tumour necrosis factor (TNF)-α, interferon (IFN)-γ, IL-4, IL-17A, CCL5, CCL20, fractalkine et al. were detected in inflamed gingival tissues and gingival crevicular fluid [1–6].

Human gingival fibroblasts (HGFs), the major cell type in periodontal connective tissues, provide tissue framework for tooth anchorage. Until recently, they were presumed to be immunological inserts. Currently, however, researchers recognize their active role in host defence. Upon stimulation with cytokines as well as with bacterial pathogens, HGFs secrete various soluble mediators of inflammation such as IL-1β, IL-6 and IL-8 [7–10] and up-regulate expression of human leucocyte antigen D-related, intercellular adhesion molecule 1 and vascular cell adhesion molecule 1 [11]. These fibroblast-derived mediators and surface antigens are thought to play an important role in periodontal inflammatory response.

The CXC chemokine ligand (CXCL10) was discovered as an IFN-γ-inducible protein of 10 kDa in the monocytic U937 cells [12]. CXCL10 attracts activated T helper 1 (Th1) cells through interaction with CXC chemokine receptor 3 (CXCR3) [13,14]. CXCL10 shares this receptor and hence biological activity with two more recently identified CXC chemokines, CXCL9 and CXCL11 [15–17]. In vivo, enhanced levels of CXCL10 have been reported in several inflammatory diseases that are associated predominantly with a Th1 phenotype. It is reported that CXCL10 and CXCR3 are detected in inflamed gingival tissues [18,19]. However, it is unknown whether or not HGFs are related to CXCL10 production in inflamed gingival tissues.

Adrenomedullin (AM) is a potent vasorelaxant peptide isolated originally from extracts of human pheochromocytoma [20]. AM has multiple regulatory functions, the most distinctive of which arises from its vasodilatory and hypotensive effect. Recently, it has been reported that AM has an anti-inflammatory effect. Isumi and the colleagues reported that AM inhibited IL-1β-induced TNF-α production in Swiss 3T3 cells [21]. In periodontal disease, Lundy and the colleagues reported that AM was present in human gingival crevicular fluid from the samples with periodontal diseased sites. However, the effect of AM on cytokine production by HGFs is uncertain [22].

In this study, we aimed first to study the effects of TNF-α on the CXCL10 production by HGFs. Moreover, we investigated the expression of AM receptors, which are calcitonin-receptor-like receptor (CRLR), receptor-activity-modifying protein (RAMP) 2 and RAMP3, in HGFs, and the affects of AM on CXCL10 production by TNF-α-stimulated HGFs.

Materials and methods

Cells and culture condition

Tissue samples were obtained at surgery from the inflamed gingiva of patients diagnosed with chronic periodontitis, or from the gingiva of clinically healthy subjects. All gingival biopsy sites in the chronic periodontitis group exhibited radiographic evidence of bone destruction, as well as having clinical probing depths greater than 4 mm, with sulcular bleeding on probing, otherwise the patients were systemically healthy. After basic periodontal therapy, we collected samples such as scaling. Samples of gingival tissues were obtained from nine chronic periodontitis patients (four males and five females; average age: 61·0 ± 9·8 years; average probing depth: 6·33 ± 2·06 mm; average attachment loss: 7·02 ± 2·26 mm) and five healthy control subjects (five females; average age: 31·2 ± 9·8 years; average probing depth: 2·4 ± 0·54 mm; average attachment loss: 2·7 ± 0·57 mm). HGFs were prepared from the explants of clinically normal gingiva from patients (three females, aged 26–40 years) during routine distal wedge surgical procedure with informed consent. Explants were cut into pieces and culture in 100-mm diameter tissue culture dishes in Dulbecco's modified Eagle's medium (Gibco, Grand Island, NY, USA) supplemented with 10% fetal bovine serum (JRH Bioscience, Lenexa, KA, USA), penicillin 50 IU/ml and streptomycin 50 μg/ml, with a medium change every 3 days for 10–15 days until confluent cell monolayers were formed. The cells were detached with 0·25% trypsin–ethylenediamine tetraacetic acid, washed with phosphate-buffered saline (PBS) and subcultured in plastic flasks. After three to four subcultures by trypsinization, homogeneous, slim spindle-shaped cells grown in characteristic swirls were obtained. The cells were used as confluent monolayers at subculture levels 5–15. Informed consent was obtained from all subjects participating in this study. The study was performed with the approval and compliance of the Tokushima University Ethical Committee.

RNA extraction and reverse transcription–polymerase chain reaction analysis

Total RNA was prepared from biopsied gingival tissue or HGFs using the Rneasy total RNA isolation Kit (Qiagen, Hilden, Germany). Single-stranded cDNA for a polymerase chain reaction (PCR) template was synthesized from 48 ng of total RNA using a primer, oligo(dT)12−18 (Invitrogen, Carlsbad, CA, USA), and superscript 3 reverse transcriptase (Invitrogen) under the conditions indicated by the manufacturer. Specific primers were designed from the cDNA sequences for CXCL10, AM, CRLR, RAMP2, RAMP3 and glyceraldehyde-3-phosphate dehydrogenase (GAPDH). Each cDNA was amplified by PCR using Hot star Taq DNA polymerase (Qiagen). The sequences of the primers were as follows: CXCL10-F (5′-TGACTCTAAGTGGCATTCAAGG-3′), CXCL10-R (5′-AGTTCAGACATCTCTCTTCTCACCC-3′), AM-F (5′-GAAGACAGCAGTCCGGATGC-3′), AM-R (5′-CGTTGTCCTTGTCCTTATCTGTGA-3′), CRLR-F (5′-CTGTACATGAAAGCTGTGAGAGCTACT-3′), CRLR-R (5′-TGGAAGTGCATAAGGATGTGCATGATG-3′), RAMP2-F (5′-GCAGAGAGGATCATCTTTGAGACTC-3′), RAMP2-R (5′-CCTCCATACTACAAGAGTGATGAGGAAG-3′), RAMP3-F (5′-CCGAGTTCATCGTGTACTATGAGAG-3′), RAMP3-R (5′-CTGTGGATGCCGGTGATGAAGC-3′), TNF-α-F (5′-GAGTGACAAGCCTGTAGCCCATGTTGTAGCA-3′), TNF-α-R (5′-GCAATGATCCCAAAGTAGACCTGCCCAGACT-3′), GAPDH-F (5′-TGAAGGTCGGAGTCA ACGGATTTGGT-3′) and GAPDH-R (5′-CATGTGGGCCATGAGGTCCACCAC-3′). The conditions for PCR were 1× (95°C, 15 min), 35× (94°C, 1 min, 59°C, 1 min, 72°C, 1 min) and 1× (72°C, 10 min). The products were analysed on a 1·5% agarose gel containing ethidium bromide.

Cytokine determination

Human gingival fibroblasts were stimulated with TNF-α (Peprotech, Rocky Hill, NJ, USA), AM (Peptide Institute, Osaka, Japan) and AM antagonist (Peptide Institute) for 24 h. Supernatants from the cells were collected and the concentration of CXCL10 was measured in triplicate by enzyme-linked immunosorbent assay (ELISA). A CXCL10 Duoset ELISA development system (R&D systems) was used for the determination. The assay was performed according to the manufacturer's instructions. The data were determined by using a standard curve prepared for each assay.

Immunohistochemistry

Gingival tissue samples were embedded immediately in the optical cutting temperature compound (Miles Laboratories Inc., Elkhart, IN, USA) and quenched and stored in liquid nitrogen. The specimens were cut into 6 μm sections using a cryostat (SFS; Bright Instrumental Company, Huntingdon, UK) and collected on poly l-lysine-coated slides. AM expression was analysed with specific antibodies; rabbit anti-human AM antibody (Phoenix Pharmaceuticals, Inc., Belmont, CA, USA). An isotype-matched control antibody (Dako, Kyoto, Japan) was used as a negative control. The sections were reacted with specific antibodies overnight at 4°C. After being washed with PBS, the sections were incubated with biotinylated anti-mouse and rabbit immunoglobulins (Dako) for 20 min at room temperature and washed with PBS to remove any unreacted antibodies. The sections were then treated with peroxidase-conjugated streptavidin (Dako) for 10 min, and washed and reacted with 3,3-diamino-benzidine tetrahydrochrolide (Dako) in the presence of 3% H2O2 to develop colour. The sections were counterstained with haematoxylin and mounted with glycerol.

Statistical analysis

Statistical significance was analysed by Student's t-test. P-values <0·05 were considered significant.

Results

CXC chemokine ligand expression by TNF-α-stimulated HGFs

We examined the effects of TNF-α on CXCL10 production by HGFs. Reverse transcription (RT)–PCR analysis revealed that CXCL10 production was enhanced by TNF-α stimulation from HGFs in a dose-dependent manner (Fig. 1a). ELISA analysis revealed that TNF-α stimulation induced CXCL10 protein release from HGFs in a dose-dependent manner (Fig. 1b).

Figure 1.

CXC chemokine ligand 10 (CXCL10) expression by human gingival fibroblasts (HGFs). (a) HGFs were treated with tumour necrosis factor (TNF)-α (0·1–100 ng/ml) for 4 h. Total RNA was isolated, and reverse transcription–polymerase chain reaction was carried out for CXCL10 and glyceraldehyde-3-phosphate dehydrogenase. Similar results were obtained in three experiments. (b) HGFs were incubated for 24 h with TNF-α (0·1–100 ng/ml) at 37°C. Medium was removed and assayed for CXCL10 release by enzyme-linked immunosorbent assay. Data are presented as the mean ± standard deviation (n = 3; by Student's t-test, **P < 0·01, versus the medium). Similar results were obtained in three repeated experiments.

The expressions of AM and AM receptors in non-stimulated HGFs

It has been reported that CRLR, together with either RAMP2 or RAMP3, would function as a receptor for AM [23,24]. Therefore, we examined CRLR, RAMP2 and RAMP3 expression in HGFs. RT–PCR analysis showed that CRLR and RAMP2 expressed in HGFs constitutively. However, RAMP3 mRNA expression was not detected. At the same time, AM mRNA expression was detected in non-stimulated HGFs (Fig. 2).

Figure 2.

Adrenomedullin (AM) and AM receptors mRNA expression in non-stimulated human gingival fibroblasts (HGFs). Total RNA was isolated from non-stimulated HGFs, and reverse transcription–polymerase chain reaction was carried out for AM, calcitonin-receptor-like receptor, receptor-activity-modifying protein (RAMP) 2 and RAMP3. Similar results were obtained in three experiments.

The effects of AM on CXCL10 production by HGFs stimulated with TNF-α

Next, we examined the effects of AM on CXCL10 production by HGFs. Single stimuli of AM did not induce CXCL10 production by HGFs (data not shown). TNF-α stimulation induced CXCL10 mRNA and protein expression by HGFs. The enhanced CXCL10 mRNA and protein expression by HGFs was inhibited by AM treatment in a dose-dependent manner (Fig. 3).

Figure 3.

Adrenomedullin (AM) inhibited CXCL10 expression in tumour necrosis factor (TNF)-α-stimulated human gingival fibroblasts (HGFs). (a) HGFs were treated with TNF-α (10 ng/ml) with or without AM (1–100 nm) for 4 h. Total RNA was isolated, and reverse transcription–polymerase chain reaction was carried out for CXCL10 and glyceraldehyde-3-phosphate dehydrogenase. Similar results were obtained in three experiments. (b) HGFs were treated with TNF-α (10 ng/ml) with or without AM (1–100 nm) for 24 h. Medium was removed and assayed for CXCL10 release by enzyme-linked immunosorbent assay. Data are presented as the mean ± standard deviation (n = 3; by Student's t-test, *P < 0·05, versus TNF-α single-stimulated). Similar results were obtained in three repeated experiments. GAPDH, glyceraldehyde-3-phosphate dehydrogenase.

The effects of AM antagonist on CXCL10 production by HGFs

Figure 2 shows that HGFs produced AM constitutively. Therefore, we examined the effects of AM produced by HGFs on CXCL10 production to antagonize AM. AM antagonist enhanced CXCL10 mRNA and protein expression by HGFs in a dose-dependent manner (Fig. 4). This result means that AM produced by HGFs suppresses CXCL10 production.

Figure 4.

Effects of adrenomedullin (AM) antagonist on CXCL10 production by tumour necrosis factor (TNF)-α-stimulated human gingival fibroblasts (HGFs). (a) HGFs were treated with TNF-α (10 ng/ml) with or without AM antagonist (0·01–1 μm) for 4 h. Total RNA was isolated, and reverse transcription–polymerase chain reaction was carried out for CXCL10 and glyceraldehyde-3-phosphate dehydrogenase. Similar results were obtained in three experiments. (b) HGFs were treated with TNF-α (10 ng/ml) with or without AM antagonist (0·01–1 μm) for 24 h. Medium was removed and assayed for CXCL10 release by enzyme-linked immunosorbent assay. Data are presented as the mean ± standard deviation (n = 3; by Student's t-test, *P < 0·05, versus TNF-α single-stimulated). Similar results were obtained in three repeated experiments. CXCL10, CXC chemokine ligand 10; GAPDH, glyceraldehyde-3-phosphate dehydrogenase.

The effects of TNF-α on AM receptors in HGFs

Next, we examined the effects of TNF-α on AM receptors in HGFs. TNF-α (1 ng/ml or 10 ng/ml) enhanced CRLR and RAMP2 mRNA in HGFs. However, TNF-α treatment could not induce RAMP3 mRNA expression in HGFs. Moreover, TNF-α stimulation did not modulate AM mRNA expression in HGFs (Fig. 5). This result means that TNF-α might enhance the influence of AM on HGFs because TNF-α up-regulated AM receptors in HGFs.

Figure 5.

Effects of tumour necrosis factor (TNF)-α on adrenomedullin (AM) and AM receptors expression in human gingival fibroblasts (HGFs). HGFs were treated with TNF-α (0·1–100 ng/ml) for 4 h. Total RNA was isolated, and reverse transcription–polymerase chain reaction was carried out for calcitonin-receptor-like receptor, receptor-activity-modifying protein (RAMP) 2, RAMP3, AM and glyceraldehyde-3-phosphate dehydrogenase. Similar results were obtained in three experiments. GAPDH, glyceraldehyde-3-phosphate dehydrogenase.

The expression of AM in periodontal tissues

Finally, we investigated that AM expression in periodontal tissues. Figure 6 shows AM, CXCL10 and TNF-α mRNA expression in periodontal tissues. AM was detected in all the samples we used in this study, including both healthy tissues and inflamed tissues. We could detect TNF-α mRNA expression in inflamed gingival tissues (in six of nine inflamed gingival samples). CXCL10 was detected clearly in three of nine inflamed gingival samples, although weak CXCL10 mRNA expression was detected in healthy gingival tissues. Most high AM mRNA detected samples (lanes 1, 3, 4 and 6) expressed low levels of CXCL10 mRNA. This result shows that AM might inhibit CXCL10 production in periodontal tissues. Figure 7 shows the expression of AM in healthy periodontal tissues. Epithelial cells and fibroblasts were stained strongly by AM antibody. AM was expressed in diseased periodontal tissues, and the staining pattern was almost the same as the healthy tissues (data not shown).

Figure 6.

Reverse transcription–polymerase chain reaction (RT–PCR) analysis of tumour necrosis factor (TNF)-α, CXCL10 and adrenomedullin (AM) mRNA expression in human periodontal tissues. Total RNA was prepared from two clinically healthy gingival samples (pocket depth, 2 mm) and nine diseased gingival samples (pocket depth, 4–10 mm). The expression of TNF-α, CXCL10, AM and glyceraldehyde-3-phosphate dehydrogenase mRNA in periodontal tissues was analysed by RT–PCR as described in the Methods. CXCL10, CXC chemokine ligand 10; GAPDH, glyceraldehyde-3-phosphate dehydrogenase.

Figure 7.

Adrenomedullin (AM) immunostaining in healthy periodontal tissues. Immunohistochemical staining of human clinically healthy periodontal tissues with anti-AM antibody (a) and negative control antibody (b) respectively. Figure 6a and b are serial sections. The original magnification for each photograph is ×200.

Discussion

The inflamed gingival tissue of periodontal disease is characterized by an infiltration of inflammatory cells, including T cells. It is reported that Th1 cells are related to bone resorption in diseased periodontal tissues [25,26]. The interaction of CXCR3 and their ligands may be critical in perpetuating the local Th1 immune response. In the present study, we demonstrated that HGFs was able to secrete CXCR3-agonistic chemokine, CXCL10, when stimulated with TNF-α. Therefore, HGFs appear to be an important cellular source of this Th1-associated chemokine in diseased periodontal tissues. Moreover, it is known that TNF-α is the main inducer of CXCL10 by skin fibroblasts [27]. Our results agree with Villagomez et al. [27].

It has been reported that AM could modulate cytokine production. For example, Wong and colleagues reported that AM inhibited TNF-α production from lipopolysaccharide (LPS)-stimulated rat macrophages [28]. On the other hand, they also reported that AM enhanced IL-6 production by rat macrophages stimulated with LPS [28]. Isumi and colleagues reported that AM suppressed IL-1β-induced TNF-α production in Swiss 3T3 cells [21], whereas they also reported that, remarkably, AM potentiated stimulatory effects of TNF-α, IL-1β and LPS on IL-6 production by Swiss 3T3 cells [29]. Previous reports and our report might explain that AM regulate cytokine production differentially. Further investigation would be necessary to elucidate the mechanism.

We show that HGFs express CRLR and RAMP2 constitutively. Choksi and colleagues reported that Rat-2 fibroblast expressed CRLR and RAMP2 [30]. They found no RAMP3 expression in Rat-2 fibroblasts; their reports agree with ours. However, Uzan and colleagues reported that human fibroblast-like synoviocytes expressed CRLR, RAMP2 and RAMP3 constitutively [31]. Their reports and ours show that AM receptors expression is dependent upon the sources of cells.

Gonzalez-Rey and colleagues reported recently that treatment with AM reduced significantly the incidence and severity of collagen-induced arthritis, an experimental model of rheumatoid arthritis, abrogating completely joint swelling and destruction of cartilage and bone. They explained that the therapeutic effect of AM was associated with a reduction of the Th1-driven autoimmune responses [32]. Our results showed clearly that AM treatment reduced Th1-type chemokine CXCL10 production by HGFs. Therefore, decrease of CXCL10 production might be related to the reduction of Th1-driven responses in collagen-induced arthritis by AM treatment. Furthermore, Gonzalez-Rey and colleagues reported that AM treatment induced generation and activation of regulatory T cells in an experimental arthritis model [32]. We did not examine the effects of AM on regulatory T cells function in this study. Further investigation should be necessary.

Lundy and colleagues reported that AM was present in all the gingival crevicular fluid samples they collected, and AM was significantly higher in periodontitis sites than in control healthy sites [22]. Our results showed that AM mRNA is detected from all samples we used in this study, and AM mRNA expression level was not different between periodontitis sites and healthy sites. Epithelial cells are related mainly to the production of proteins to gingival crevicular fluids. Kapas and colleagues reported that AM gene expression in human oral keratinocytes was increased in response to P. gingivalis, Streptococcus mutans, Candida albicans and Eikenella corrodens[33]. These reports and our results explain that bacteria-stimulated epithelial cells enhanced AM production to protect microorganism invasion, and HGFs constitutively produced AM to reduce Th1-type inflammation. Our gingival samples include both epithelial cells and fibroblasts. Therefore, we detected AM mRNA expression in all samples. Further investigation is necessary to elucidate the mechanism related to AM production by HGFs and epithelial cells.

We examined CXCL10 production by HGFs in this study. However, epithelial cells might be a source of CXCL10 in periodontal tissues, because epithelial cells from other tissues could produce CXCL10 [34,35]. Therefore, we should examine an ability of gingival epithelial cells to produce CXCL10.

Finally, we discovered that HGFs are the source of CXCL10, and production induced by TNF-α is inhibited by AM treatment. Our results show that AM can be a therapeutic target of periodontal disease, because Th1 condition is related to bone resorption in periodontal tissues and AM could inhibit Th1-type chemokine production by HGFs.

Acknowledgements

This work was supported by a Grant-in-Aid for Scientific Research (B) (19791616) from the Japan Society for the Promotion of Science.

Ancillary