J.Wahlgren, T. Salo, O.Teronen, H. Luoto, T. Sorsa & L.Tjaderhane
Faculty of Medicine and Biomedicum, Oral Pathology Unit/Laboratory Diagnostics, Helsinki University Central Hospital (HUCH), University of Helsinki, Helsinki.
Institute of Dentistry, University of Oulu, Oulu University Hospital, Oulu, Medix Biochemica, Kauniainen, Finland.
Department of Endodontics, Faculty of Dentistry, University of Toronto, Toronto, Canada.

Introduction.
Early pulpal and periapical responses to bacteria or their metabolites include the influx and recruitment of polymorphonuclear leucocytes (PMNs) and monocytes/macrophages. In the more chronic states, proliferation of fibroblasts and vascular elements, and the infiltration of macrophages, lymphocytes and plasma cells are characteristic features (Simon 1998). The role of all the factors participating in the spread of inflammation and necrosis in pulp tissue is not completely understood. Since matrix metalloproteinases (MMPs), the host enzymes responsible for the extracellular matrix degradation, have been suggested to be important in other inflammatory conditions such as periodontitis, it is possible that they also participate in the pathogenesis of pulp and periapical inflammation.
MMPs forma family of structurally related but genetically distinct endopeptidases expressed at low levels in normal tissues, but upregulated during inflammation (Birkedal-Hansen 1995). Collagenases (MMP-1 and MMP-8) and gelatinases (MMP-2 and MMP-9) are present in jaw cyst wall extracts and cyst fluids (Teronen et al. 1995a,b), but no data exists of their role in pulpal or periapical inflammation. Previously, MMP-8 has been thought to be uniquely produced by developing PMN cells in bone marrow, and that the MMP-8 activity is dependent on the release of the enzyme from PMN cells by degranulation (reviewed by Mainardi et al. 1991). However, recent findings demonstrated the expression of MMP-8 in mesenchyme derived, non-PMN lineage cells, including dental pulp fibroblasts, odontoblasts, inflamed epithelial cells and plasma cells (Hanemaaijer et al. 1997, Palosaari et al. 2000, Tervahartiala et al. 2000, Wahlgren et al. 2001). The exact role of MMP-8 synthesized by the cells in the pulpo-dentinal complex is not clear (Tjaderhane et al.2001).Therefore, a hypothesis was proposed that MMP-8 could be present in the inflamed or necrotic pulp, and the enzyme level could be related to the activity of tissue destruction of the pulp and periapical tissue. The aim of this study was to evaluate the presence and cellular sources, as well as the levels and molecular forms of MMP-8 in pulpal and periapical inflammation and during the root-canal treatment using immunofluorometric assay (IFMA), Western blot and immunohistochemical staining.

Materials and methods.

Patients.
Ten patients, with noncontributory medical histories and diagnosed to have periapical periodontitis were selected for the study. The diagnosis of apical periodontitis was based mainly on radiographic examination, demonstrating clear bone loss and disappearance of the periodontal ligament space in the periapical region. Occasionally, the patients reported mild spontaneous pain, and percussion sensitivity was observed upon clinical examination. Patients with primary or secondary acute periodontitis were excluded from the study. All teeth (n = 11) had one root canal only, and all were confirmed to be nonvital upon access. The root-canal treatments and the sample collection for the root-canal exudates MMP-8 analysis were performed by one of the authors ( J.Wahlgren) at a private dental office with the approval of Ethical Committee of Helsinki University Faculty of Medicine, and with the patient’s informed consent. The samples were collected during three consecutive appointments within 2-week intervals, during which the root-canal treatment was completed. For MMP-8 IHC stainings, the pulp tissue from vital teeth treated for irreversible pulpitis (n =10) were collected by the same operator.

Sample collection.
Primary access to the pulp was completed with dental burs using rubber dam isolation. Instrumentation of the root canal commenced with size 20 or 25 files. A radiograph was taken to determine the canal length. A sterile paper point was then inserted into the root canal up to the preparation length for 2 min to absorb exudates for the sample. The root canals were then instrumented to a minimum of size 40 with frequent rinsing with 2.5%sodiumhypochlorite (NaOCl). For the final rinsing, 15%ethylenediamine tetraacetic acid (EDTA) and NaOCl were used. The root canals were dried, Ca(OH)2 paste (pH 12.5) (Biocalc,Orion, Helsinki, Finland)was inserted into the canal and the cavity was sealed with Cavit1 (ESPE, Seefeld, Germany) and IRM1 (Dentsply Caulk, Milford, DE, USA).
After 2 weeks, the teeth were isolated with rubber dam and the temporary fillings were removed. The root canals were instrumented gently to facilitate the removal of Ca(OH)2 with sterile saline solution rinsing. After drying of the canals, a sterile paper point was inserted into the canal for 2 min for periapical exudate sample collection. The canal was rinsed with NaOCl, dried and Ca(OH)2 paste was inserted into the canal. After 2 weeks, the procedure for the sample collection was repeated, and the root canals were filled with gutta-percha and sealer using lateral condensation.
After sample collection, the exudate from the absorbent paper points was immediately eluted with 50 mL of50 mm Tris-HCl buffer, pH 7.5, containing 0.15 m NaCl and 1mm CaCl2, placed on a shaker for 3 h and stored at _70 8C for further handling.

MMP-8 analyzes.
The molecular forms of MMP-8were analyzed using the Western blotting method using rabbit polyclonal antibody against MMP-8 in normal fashion, as described previously (Palosaari et al. 2000). Briefly, samples were run on10%SDS-PAGE gels and transferred to nitrocellulose filter. The filters were incubated overnight with MMP-8 antibody, and then with peroxidase-conjugated antirabbit immunoglobulin’s for 1 h. Diaminobenzidine tetrahydrochloride (DAB) was used for the detection. The intensity of molecular forms of MMP-8, both latent and active forms separately (Kiili et al. 2002), were evaluated by scanning the respective bands and analyzing the band intensity with Bio-Rad Model GS-700 Imaging Densitometer using Molecular Analyst1/PC program.
Concentration of MMP-8 in the samples was also determined from the elution buffer by a time-resolved IFMA as described previously (Hanemaaijer et al. 1997). Briefly, two monoclonal MMP-8-specific antibodies were used as catching antibody and tracer antibody, respectively. The tracer antibody was labeled using europium chelate. After adding the enhancement solution, fluorescence was measured using a 1234 Delfia Research Fluorometer (Wallac,Turku, Finland). The specificity of the monoclonal antibodies against MMP-8 corresponded to that of polyclonalMMP-8 (Hanemaaijer et al.1997).
Immunohistochemical staining of the periapical tissue samples Periapical granuloma samples (n ј10) for the immunohistochemical stainings were randomly selected from the files of Department of Oral Pathology, Institute of Dentistry, University of Oulu. Periapical tissue samples were all collected during periapical surgery, undertaken as a part of the normal treatment of patients. In each case a periapical radiolucency was present in an endodontically treated tooth, and the diagnosis of periapical granuloma was confirmed with the histological evaluation.
Immunohistological stainings were performed as described previously (Wahlgren et al. 2001) with polyclonal antibody (diluted 1 : 800 in 1% BSA/PBS) for MMP-8 and monoclonal CD/68Ab-3 (Clone KP1)macrophage- marker (NeoMarkers, Fremont, CA, USA), on formalin fixed paraffin embedded pulp and periapical granuloma tissue sections. Brieffy, sections were incubated with pepsin to reveal the epitopes, and with H2O2 to quench the endogenous peroxidase activity. Nonspecific binding was blocked with 2% normal blocking serum inVectastain1 Elite ABC Kit (Vector Laboratories, Burlingame, CA, USA) and incubated with the primary antibody overnight. The control sections were incubated with the buffer alone or nonimmune rabbit serum. For detection, the sections were incubated with biotinylated secondary antibody (antirabbit IgG or antimouse IgG) and with Vectastain1 Elite ABC reagent, and stained with 3-amino-9-ethylcarbazole (Sigma, St Louis, MO, USA), diluted in N,N-dimethylformamide (Merck, Darmstadt, Germany), and counter stained with Mayer’s haematoxylin.

Statistical analysis.
The correlation analysis was performed between the IFMA analysis and respective Western blot scanning results using the Pearson’s product-moment correlation coefficient, to ensure the comparativeness of the two methods and to confirm the use of IFMA results as the primary method for the statistical analysis.
Kruskall-Wallis one-way anova and Mann-Whitney U-test were used to analyze the degree of difference in the MMP-8 levels between the different appointments. The nonparametric tests were used because of the absence of the normal distribution within each sample site groups. Student’s t-test was used to compare the latent and active forms of enzyme in the root-canal exudate.

Pulpitis.
The Western blots of the teeth with acute pulpitis demonstrated large amounts ofMMP-8inthepulpal tissue (data not shown). Whilst most of the 70-80 kDa MMP-8 observed in the Western blot most likely to originated from the PMN cells numerously present in the inflamed tissue, the immunohistochemical staining, revealed that other cells can also express MMP-8 in the inflamed human pulpal tissue (Fig. 1).
In the pulp proper, macrophage-like cells (not shown), PMN and plasma cells showed a clear expression of MMP-8, and the most intense MMP-8-positive cell accumulation was observed surrounding the pulpal abscess (Fig. 1A-D). Also, some endothelial cells of the pulp vessels were positive in MMP-8 staining (Fig. 1E). The control sections, incubated without primary antibody or with normal serum, were not stained (not shown).

Figure 1. The immunohistochemical staining of the inflamed pulp tissue with MMP-8-specific antibody.
(A) MMP-8-expressing cells were evenly distributed throughout the whole pulp tissue, but most abundantly (arrows) surrounding the pulp abscess (PA).
(B) PMN cells expressed an intense staining (arrows).
(C) In connection with the dentine chip removed with the pulp tissue, the ?broblast-type cells at the close proximity of the dentine occasionally expressed MMP-8 protein (arrow).
(D) Plasma cells (arrows) with strong expression of MMP-8.
(E) Some endothelial cells in the pulpal vessels expressed fainter, but clearly observable staining with MMP-8.
Scale bars: 10 mm (B, D and E); 25 mm (C);40 mm (A).

Root-canal exudate samples.
MMP-8 levels markedly decreased during the root-canal treatment of necrotic teeth (Fig. 2A-C). However, even after removal of the pulpal tissue remnants and debris and a careful mechanical preparation of the root canals accompanied with the local medication with calcium hydroxide paste, MMP-8 was detected occasionally (Fig. 2A-C). Only after the renewal of the local medication for fortnight, MMP-8 was almost completely absent form the canals (Fig. 2A-C). The difference between the first and last visit samples was statistically significant (Kruskall-Wallis P = 0.02; Mann-Whitney U-test P = 0.0107).
The relative proportions of latent and active forms of MMP-8 in the samples were equal in the first visit samples (Fig. 2D). In cases with MMP-8 observed in the subsequent visits, the relative amount of active MMP-8 according to the band scanning intensities was higher (Fig. 2D). However, the differences were not statistically significant, most likely owing to the small number of samples with MMP-8 present.
Only one outof11specimens failed to show decrease in theMMP-8 level below 1000 ng mL in the final periapical exudate sampling (Fig.2C); that tooth was later diagnosed to have a vertical root fracture and was extracted.
Spearman product-moment correlation analysis revealed a statistically significant positive correlation between the Western blot scanning analysis and IFMA (0.4871, P = 0.006). Thus, the scanning analysis of the Western blots confirmed the results obtained with IFMA, and justified the use of IFMA results in statistical analysis.
Immunohistochemical staining of the periapical tissue samples PMN cells and macrophage-like cells expressed the staining with MMP-8-specific antibody, the PMN cells being the predominant cell type to express MMP-8 (Fig. 3A). Staining with CD68 macrophage marker antibody indeed confirmed the nature of the macrophage-like cells staining with the MMP-8 antibody (Fig.3B-C).

Figure 2. MMP-8 protein in root canals in the different time points of the study.
(A) Western blot using MMP-8-specific antibody. Representative samples from the root canals collected in the first, second and third visit of one tooth. Molecular weight standards were used to determine the molecular weight of the proteins detected. The latent and active forms of MMP-8 (pro-MMP-8 and MMP-8, respectively), as well as complexed forms (CF) and low molecular weight truncated forms (TF) are present prior to the root-canal preparation (Visit1). After 2 weeks of Ca(OH)2 treatment,MMP-8 levels are markedly reduced, but still a distinct band corresponding to active form of MMP-8 is present (Visit 2). After another 2 weeks of medication, MMP-8 is virtually absent (Visit 3).
(B) The concentrations of MMP-8 (mean and standard deviation) detected with the IFMA analysis in root canals during the three visits. The decrease in the root-canal samplesMMP-8 is evident, and theMMP-8 level in the last visit is significantly lower when compared to the first visit (Kruskall-Wallis P = 0.02; Mann-Whitney U-test P = 0.0107).
(C) The concentrations of root canalMMP-8 in individual cases during the three visits for endodontic treatment, as detected with the IFMAanalysis.5/11cases presented levels above1000 ng mL_1during the first visit; four of them decrease below detection level after 2 weeks of Ca(OH)2 medication. In one case, MMP-8-levelmarkedly increased after 2 weeks of medication, reaching the value of 14145 ng mL_1 (marked with _ in the figure), and remained high (>1000 ng mL_1) even after another period of Ca(OH)2 medication; the tooth was later found to be fractured. In two cases with initial level of 0 ng mL_1MMP-8, the levels slightly increased in the second measurement, with subsequent decrease after another period of medication.
(D) The relative levels of latent pro-MMP-8 and active MMP-8 in the root-canal samples, as percentage of total amount ofMMP-8 in each sample. The relative proportion of activeMMP-8 in the second and third visit samples increased, but the differences did not reach statistical significance.



Figure 3. Immunohistochemical stainings of the periapical granulomas withMMP-8-specific antibody.(A) PMN cells (short arrow) and macrophage-like cells (long arrow) express distinct staining.
(B, C) The staining of the consecutive sections with anti-MMP-8.
(B) and macrophage-specific marker antibody CD68.
(C) Confirmed the identity of the cells to be macrophages.
Scale bars: 10 mm (A); 25 mm (B and C).

Discussion.
The Western immunoblot, together with the immunohistochemical stainings with MMP-8-specific antibody, demonstrated that MMP-8 (collagenase-2) is present in an inflamed pulp tissue in abundant levels. Most of the MMP-8 in pulpitis eventually originates not only from PMN leucocytes, but also from the other cells, such as macrophage-like cells and plasma cells. In this regard, pulp seems not to differ from other tissues such as periodontium (Kiili et al. 2002) affected by the inflammatory challenge. However, as odontoblasts and pulp fibroblasts can express MMP-8 (Palosaari et al.2000) as wellas other MMPs (Panagakos et al. 1996, Tamura et al. 1996, Llano et al. 1997, Tjaderhane et al. 1998b), the cells of the pulpo-dentinal complex may participate to the matrix remodeling and degradation during the inflammatory processes in the pulp. Indeed, interleukin-1a (IL-1a), tumor necrosis factor-a (TNF-a) and bacterial lipopolysaccharides can stimulate gelatinase expression in pulp fibroblasts in vitro (Panagakos et al. 1996, Tamura et al. 1996, Ueda & Matsushima 2001). Recent studies also show increased expression of collagenase-1 (MMP-1) in cultured human pulp fibroblasts after IL-1a and TNF-a stimulation, whilst prostaglandin E2 (PGE2) stimulates the expression of tissue inhibitor of metalloproteinases- 1 (TIMP-1) (Lin et al. 2001).
Macrophages express various MMPs, including collagenase- 1 (MMP-1), gelatinases (MMP-2, MMP-9), stromelysins (MMP-3, MMP-7, MMP-12) (reviewed by Goetzl et al. 1996) and collagenase-3 (MMP-13) (Imai et al.1998). This may be the first time that macrophage like cells at the site of inflammation have been shown to express MMP-8. The role of collagenases from macrophages may be related to the suggested capability of macrophage-derived MMPs to release growth factors from extracellular matrix to regulate inflammatory reaction (Goetzl et al.1996) or simply to degrade extracellular matrix at the site of inflammation. The immunohistochemical stainings with MMP-8-specific antibody also confirm and further extend the recent findings with the plasma cells present in the close vicinity of bone destructive jaw cysts and periodontitis-affected gingiva, capable of synthesizing MMP-8 (Wahlgren et al. 2001, Kiili et al. 2002).
During the root-canal treatment, a clear decrease in the root-canal exudate MMP-8 levels was observed. The presence of MMP-8 during the second visit, despite the pulp tissue being removed and the canals being cleaned thoroughly during the first appointment, strongly indicates that MMP-8 originated from the periapical inflammation site. This further implies that degradation of the extracellular matrix proteins indicates the active phase of periapical site inflammation to be still present at that time. The virtual absence of MMP-8 in the root-canal exudate during the third visit evidently reflects the disappearance of inflammation and at least partial onset of healing in the periapical environment. This is further supported by the finding that the one specimen failing to show lowMMP-8 level in third appointment was later diagnosed to be vertically fractured. Measuring the presence and levels of MMP-8 from the root-canal exudates during the endodontic treatment could thus serve as a biochemical indicator for monitoring the inflammatory status of the periapical tissue. Therefore, it might be used as a diagnostic guide in deciding the treatment procedures, modalities and medication.
Chlorhexidine, which is sometimes used in root-canal medication, can inhibit MMPs (Gendron et al.1999).Even though the effect of other root-canal medicaments is not known, it would be logical to speculate that Ca(OH)2, because of its high initial pH, may inactivate MMPs which are neutral proteases. However, IFMA measures the amount of total, not the catalytic activity MMP, and therefore, the inhibition of MMP activity by rootcanal medication is not reflected in the results. If the treatment is successful, the number of PMNs in the periapical area will recede and subsequently the MMP-8 levels will be reduced, indicating the success in treatment procedure. Also, the higher proportion of active MMP-8 during the second and third visit (Fig. 2D) indicates that intracanal Ca(OH)2 does not affect the activation of MMP-8 in cases with persistent inflammatory activity, supporting the usefulness of measuring the levels of this surrogate inflammatory mediator from the root canal.
Chair side tests indicating inflammatory activity in periapical lesions have been suggested as future diagnostic tools in endodontics (Takahashi 1998, Metzger 2000). Previous studies have assessed the use of inflammatory mediators such as IL-1b and IL-1a, PGE2 and TNF-a in the evaluation of inflammatory status of periapical lesions (Safavi & Rossomando 1991, Matsuo et al. 1994, Takayama et al.1996, Kuo et al.1998, Ataoglu et al. 2002), but the results have been conflicting. The problem with the above-mentioned inflammatory mediators in this respect may be their multifunctional role in the inflammation, depending on, for example, the concentration, combination and target cells of the mediators. The role of matrix metalloproteinases at the site of the inflammation is, however, clearly defined to the degradation of extracellular proteins (Birkedal-Hansen 1995), offering, therefore, a potential for more specific analysis of inflammatory activity at the periapical site. Therefore, their role in pulpal and periapical pathogenesis should be more thoroughly examined. A chair side MMP-8 test has been recently developed to diagnose and monitor the course and the success of treatment of marginal periodontitis and peri-implantitis (Sorsa et al. 1999, Chen et al. 2000, Ma et al. 2000, Mantylaet al. 2000). Further clinical studies to use MMP measurements from the root canals during the treatment as a diagnostic tool to evaluate the status of the periapical inflammation should be conducted. Finally, in animal experiments specific MMP inhibitors alone or in combinations have decreased oedema and inflammatory tissue damage significantly, suggesting possible therapeutic benefits (Goetzl et al. 1996, Llavernas et al. 2001). There is a wide selection of MMP inhibitors, some of them already in clinical use, suggested as adjunctive therapeutic agents in periodontitis (Sorsa et al.1994, Golub et al.1997, Ciancio &Ashley 1998, Llavernas et al. 2001). MMP inhibition has also been suggested to decrease bone resorption in pathological conditions (Vernillo & Rifkin1998) and dentinal caries progression (Tjaderhane et al. 1998a; 1999, Sulkala et al.2001).MMP inhibition in the root canals and periapical tissues may thus offer new opportunities to rootcanal treatment in the future.

References.

Ataoglu T, Ungor M, Serpek B, Haliloglu S, Ataoglu H, Ari H (2002) Interleukin-1b and tumour necrosis factor-alpha levels in periapical exudates. International Endodontic Journal 35, 181-5.
Birkedal-Hansen H (1995) Proteolytic remodellation of extracellular matrix. Current Opinion in Cell Biology 7, 728-35.
Chen HY, Cox SW, Eley BM, Mantyla P, Ronka H, Sorsa T (2000) Matrix metalloproteinase-8   levels and elastase activities in gingival crevicular fluid from chronic adult   periodontitis patients. Journal of Clinical Periodontology 27, 366-9.
Ciancio S, Ashley R (1998) Safety and efficacy of sub-antimicrobial dose   doxycycline therapy in patients with adult periodontitis. Advances in Dental   Research 12, 27-31.
Gendron R, Grenier D, Sorsa T, Mayrand D (1999) Inhibition of the activities   of matrix metalloproteinases-2, -8, and -9 by chlorhexidine. Clinical and   Diagnostic Laboratory Immunology 6, 437-9.
Goetzl EJ, Banda MJ, Leppert D (1996) Matrix metalloproteinases in immunity. Journal of Immunology 156, 1-4.
Golub LM, Lee HM, Greenwald RA et al. (1997) A matrix metalloproteinase inhibitor   reduces bone-type collagen degradation fragments and specific collagenases in   gingival crevicular fluid during adult periodontitis. Inflammation Research   46, 310-9.
Hanemaaijer R, Sorsa T, Konttinen Y Tet al. (1997) Matrix metalloproteinase-   8 is expressed in rheumatoid synovial ?broblasts and endothelial cells. Regulation   by tumor necrosis factor-alpha and doxycycline. Journal of Biological Chemistry   272, 31504-9.
Imai S, Konttinen YT, Jumppanen M et al. (1998) High levels of expression   of collagenase-3 (MMP-13) in pathological conditions associated with a foreign   body reaction. Journal of Bone and Joint Surgery (BritishVolume) 80, 701-10.
Kiili M, Cox SW, Chen HY et al. (2002) Collagenase-2 (MMP-8) and collagenase-3(MMP-13)   in adult periodontitis: molecular forms and levels in gingival crevicular fluid   and immunolocalisation in gingival tissue. Journal of Clinical Periodontology   29, 224-32.
Kuo M-L, Lamster IB, Hasselgren G (1998) Host mediators in endodontic exudates.   Part I. Indicators of inflammation and humoral immunity. Journal of Endodontics   24, 598-603.
Lin S-K, Wang C-C, Huang S et al. (2001) Induction of dental pulp ?broblast   matrix metalloproteinases-1 and tissue inhibitor of metalloproteinases-1 gene   expression by interleukin-1a and tumor necrosis factor-a through a prostaglandin-dependent   pathway. Journal of Endodontics 27, 185-9.
Llano E, Pendas AM, Knauper Vet al. (1997) Identification and structural   and functional characterization of human enamelysin (MMP-20). Biochemistry   36, 15101-8.
Llavernas A, Ramamurthy N-S, Heikkila Pet al. (2001) A combination of chemically   modified doxycycline and bisphosphonate synergistically inhibits endotoxin-induced   periodontal breakdown in rats. Journal of Periodontology 72, 1069-77.
Ma J, Kitti U, Teronen O et al. (2000) Collagenases in different categories   of peri-implant vertical bone loss. Journal of Dental Research 79, 1870-3.
Mainardi CL, Pourmotabbed TF, Hasty KA (1991) Inflammatory phagocytes and   connective tissue degrading metalloproteinases. American Journal of Medical   Sciences 302, 171-5.
Mantyla P, Kinane D, Luoto H, Tikanoja S, Sorsa T (2000) MMP (matrix metalloproteinase)   -8 in monitoring periodontal health and disease from gingival crevicular fluid   (GCF). Journal of Clinical Periodontology 27 (Suppl.1), 317(Abstract).
MatsuoT, Ebisu S, Nakanishi T, Yonemura K, Harada Y, Okada H (1994) Interleukin-1a   and interleukin-1b in periapical exudates of infected root canals: correlation   with the clinical findings of the involved teeth. Journal of Endodontics   20, 432-5.
Metzger Z (2000) Macrophages in periapical lesions. Endodontics and Dental   Traumatology 16, 1-8.
Palosaari H, Wahlgren J, Larmas Met al. (2000) The expression of MMP-8 in   odontoblasts and dental pulp cells is down-regulated by TGF-b1. Journal   of Dental Research 79, 77-84.
Panagakos FS, O'Boskey JF, Rodriguez E (1996) Regulation of pulp cell matrix   metalloproteinase production by cytokines and lipopolysaccharides. Journal   of Endodontics 22, 358-61.
Safavi KE, Rossomando EF (1991) Tumor necrosis factor identified in periapical   tissue exudates of teeth with apical periodontitis. Journal of Endodontics17,   12-4.
Simon JHS (1998) Periapical pathology. In: Cohen S, Burns RC, eds. Pathways   of the Pulp, 7th edn. St. Louis, USA: Mosby, pp. 425-62.
Sorsa T, Ding Y, Salo T et al. (1994) Effects of tetracyclines on neutrophil,   gingival, and salivary collagenases. A functional and Western blot assessment   with special reference to their cellular sources in periodontal diseases. Annals of NewYork Academy of Science 732, 112-31.
Sorsa T, Mantyla P, Ronka H et al. (1999) Scientific basis of a matrix metalloproteinase-8   specific chair-side test for monitoring periodontal and peri-implant health   and disease. Annals of New York Academy of Science 878, 130-40.
Sulkala M, Wahlgren J, LarmasMet al. (2001) The effect of MMP inhibitors   on human salivary MMP activity and caries progression in rats. Journal of   Dental Research 80, 1545-9.
Takahashi K (1998) Microbiological, pathological, inflammatory, immunological   and molecular biological aspects of periradicular disease. International   Endodontic Journal 31, 311-25.
Takayama S, MikiY, Shimauchi H, Okada H (1996) Relationship between prostaglandin   E2 concentrations in periapical exudates from root canals and clinical findings   of periapical periodontitis. Journal of Endodontics 22, 677-80.
Tamura M, Nagaoka S, Kawagoe M (1996) Interleukin-1 alpha stimulates interstitial   collagenase gene expression in human dental pulp fibroblast. Journal of   Endodontics 22, 240-3.
Teronen O, SaloT, Konttinen YT et al. (1995a) Identification and characterization   of gelatinases/type IV collagenases in jaw cysts. Journal of Oral Pathology   and Medicine 24, 78-84.
Teronen O, Salo T, Laitinen J et al. (1995b) Characterization of interstitial   collagenases in jaw cyst wall. European Journal of Oral Sciences103, 141-7.
Tervahartiala T, Pirila E, Ceponis A et al. (2000) The in vivo expression   of the collagenolytic matrix metalloproteinases (MMP-2, -8, -13, and -14) and   matrilysin (MMP-7) in adult and localized juvenile periodontitis. Journal   of Dental Research 79, 1969-77.
Tjaderhane L, Larjava H, Sorsa T, Uitto VJ, Larmas M, Salo T (1998a) Matrix   metalloproteinases are responsible for the collagen matrix breakdown in tooth   decay. Journal of Dental Research 77, 1622-9.
Tjaderhane L, Salo T, Larjava H, Larmas M, Overall CM (1998b) A novel organ   culture method to study the function of the human odontoblasts in vitro: gelatinase   expression by odontoblasts is differentially regulated by TGF-b1. Journal   of Dental Research 77, 1488-98.
Tjaderhane L, Sulkala M, Sorsa T, Teronen O, Larmas M, Salo T (1999) The   effect of MMP inhibitor Metastat on fissure caries progression in rats. Annals of New York Academy of Science 878, 686-8.
Tjaderhane L, PalosaariH,WahlgrenJ, LarmasM, SorsaT, SaloT (2001) Human odontoblast   culture method: the expression of collagen and matrix metalloproteinases (MMPs). Advances in Dental Research15, 55-8.
Ueda L, Matsushima K (2001) Stimulation of plasminogen activator activity   and matrix metalloproteinases of human dental pulp-derived cells by tumor necrosis   factor-a. Journal of Endodontics 27, 175-9.
Vernillo AT, Rifkin B-R (1998) Effects of tetracyclines on bone metbolism. Advances in Dental Research12, 56-62.
Wahlgren J, Maisi P, Sutinen M et al. (2001) Expression and induction of   collagenases (MMP-8 and -13) in plasma cells associated with bone-destructive   lesions. Journal of Pathology 194, 217-24.