|Home | About | Journals | Submit | Contact Us | Français|
Inability to distinguish periapical cysts from granulomas prior to performing root canal treatment leads to uncertainty in treatment outcomes, because cysts have lower healing rates. Searching for differential expression of molecules within cysts or granulomas could provide information with regard to the identity of the lesion or suggest mechanistic differences that may form the basis for future therapeutic intervention. Thus, we investigated whether granulomas and cysts exhibit differential expression of extracellular matrix (ECM) molecules.
Human periapical granulomas, periapical cysts, and healthy periodontal ligament tissues were used to investigate the differential expression of ECM molecules by microarray analysis. Since matrix metalloproteinases (MMP) showed the highest differential expression in the microarray analysis, MMPs were further examined by in situ zymography and immunohistochemistry. Data were analyzed using one-way ANOVA followed by Tukey test.
We observed that cysts and granulomas differentially expressed several ECM molecules, especially those from the matrix metalloproteinase (MMP) family. Compared to cysts, granulomas exhibited higher MMP enzymatic activity in areas stained for MMP-9. These areas were composed of polymorphonuclear cells (PMNs), in contrast to cysts. Similarly, MMP-13 was expressed by a greater number of cells in granulomas compared to cysts.
Our findings indicate that high enzymatic MMP activity in PMNs together with MMP-9 and MMP-13 stained cells could be a molecular signature of granulomas, unlike periapical cysts.
Periapical disease represents the progression of a bacterial infection from the dental pulp to apical foramen that results in a localized inflammatory response concomitant with bone resorption (1–4). Periapical lesions include granulomas and cysts, and both are thought to represent different stages of the same inflammatory process (5–7).
Differential diagnosis of granulomas and cysts using radiographic analysis is problematic. Although preliminary studies have proposed that CT scans and ultrasound with power Doppler flowmetry can provide an additional diagnostic tool in Endodontics (8–10), it is widely accepted that histologic evaluation is necessary to confirm diagnosis (4, 11–14). However, microscopic examination can only be performed after the periapical disease is removed, which is a limitation, since in most cases root canal treatment is performed without removing the lesion. This inability to identify the status of periapical disease makes treatment outcomes unpredictable because cysts exhibit lower healing rates and generally require additional surgical treatment (6, 15–18).
Identifying extracellular matrix molecules (ECM) specific to human periapical cysts or granulomas can provide information to potentially discriminate between these lesions. Specific proteins present in the extracellular matrix and their respective receptors may offer the basis to develop novel approaches aiming to detect disease biomarkers and therefore to improve diagnosis for cysts and granulomas prior to performing root canal treatment. Possible likely candidates as biomarkers for periapical inflammation include proteases that are responsible for ECM degradation such as matrix metalloproteinases (MMPs). MMPs are a family of metal-dependent endopeptidases, which are secreted as inactive proenzymes (zymogens) and activated in tissues by cleavage of the propeptide (19–23). Although MMPs have been reported in periapical lesions (11, 24–30), a direct comparison of MMPs in cysts and granulomas has not been undertaken. Since MMPs must be activated to exert their function, it is also important to localize areas of proteinase activity within lesions. Thus, the aim of this study was to compare the expression of several ECM molecules and cell membrane receptors within different cellular components from human periapical granulomas and cysts. In addition, we sought to examine the presence and activity of MMPs in these tissues because of their reported involvement in these lesions, and because our initial data revealed differential expression of this class of proteins in these tissues.
Formalin-fixed paraffin-embedded sections from 10 periapical granulomas and 10 periapical cysts were obtained from the archives of the Oral Pathology Biopsy Service at the University of Michigan School of Dentistry after Institutional Review Board approval. Periodontal ligament (PDL) tissues obtained from the middle third of the dental root of extracted healthy teeth, after Institutional Review Board approval, were used as controls.
Histological examination was performed by a pathologist on hematoxylin-eosin stained slides of tissues from the apical regions of non vital teeth. Periapical cysts were selected based on the presence of granulation tissue with lining stratified squamous epithelium. Periapical granulomas were selected based on the presence of granulation tissue without lining epithelium (16, 17, 31). Fibroblast-like cells were characterized by their spindle shaped morphology, mononuclear inflammatory cells were identified as cells with a large single nucleus, and PMNs were identified by their multilobed nuclei.
Paraffin-embedded tissue sections were deparaffinized, exposed to protease digestion and immersed in Trizol® Reagent (Invitrogen Corporation, Carlsbad, CA) for RNA extraction. Extracted RNA was purified using the RNEasy Micro Kit (Qiagen, Valencia, CA). The quality and concentration of isolated RNA was evaluated using a capillary electrophoresis system in an RNA 6000 Pico LabChip (Agilent 2100 BioAnalyzer, Agilent Technologies Inc, Santa Clara, CA). Analysis of expression of ECM molecules (Table 1) in tissues was performed using focused cDNA microarrays (SuperArray, Bioscience Corporation, Frederick, MD). Ten micrograms of total RNA were used as a template to generate biotin-16-dUTP labeled cDNA probes. cDNA probes were denatured and hybridized at 60°C with the SuperArray membranes, which were developed using chemiluminescence. Gene expression was evaluated by densitometric analysis of the membrane spots. Comparison among groups was performed on a gene-by-gene basis after normalization by β-actin mRNA expression.
Five µm-thick tissue sections were immersed in sodium borohydride (1 mg / ml) followed by incubation with a fluorescein isothiocyanate (FITC)-bound gelatin substrate (DQ™ Gelatin, Molecular Probes, Eugene, OR) dissolved in agarose (0.1 mg / ml) for 3 hr at 37°C in a humidified light-protected chamber. Nuclei were counterstained by adding 4'-6-Diamidino-2-phenylindole (DAPI; 0.5 µg / ml) to the incubation medium. Control slides were preincubated in 20 mM ethylene diamine tetraacetic acid (EDTA, Sigma, St Louis, MO) for 1 hr, and then EDTA was added to the incubation medium. Quantification of gelatinolytic activity in the sections was assessed by counting the number of spots of fluorescence in representative areas (40× magnification) and expressed as number of spots of enzymatic activity per mm2.
Specific MMPs were detected by immunohistochemistry. Tissue sections were quenched in a 6% H2O2 methanol solution for 30 min and boiled in 10 mM sodium citrate (pH 6.0) at 93°C for 15 min for antigen retrieval. Nonspecific binding was blocked with 1% bovine serum albumin (Sigma) for 30 min and sections were incubated for 1 hr with primary antibodies for MMP-2 (5 µg / ml; MAB3308, Chemicon, Temecula, CA), MMP-9 (5 µg / ml; MAB3309, Chemicon), and MMP-13 (5 µg / ml; M4052, Sigma). Sections were incubated with secondary antibody followed by streptavidin horseradish peroxidase and 3,3’-diaminobenzidine (DAB500 Chromogen System, Biocare Medical). Tissues were counterstained with Mayer’s Hematoxylin and mounted using standard protocols. Negative controls consisted of replacing the primary antibody with mouse or rabbit IgG. The number of positive cells was calculated for each antibody in three representative fields-of-view (100× magnification).
Data were analyzed using one-way ANOVA followed by Tukey test (α = 0.05).
Mononuclear inflammatory cells were the most prevalent cells found in both cysts and granulomas (59.8% ± 19.8 versus 43.1% ± 3.8, respectively; p > 0.05). Similar quantities of fibroblast-like cells were also observed in both lesions (24.1% ± 8.4 in cysts versus 24.6% ± 4.0 in granulomas; p > 0.05). In contrast, PMNs were more prevalent in periapical granulomas compared to cysts (32.1 ± 5.8% versus 15.9% ± 7.5 respectively; p < 0.05). Inflammatory activity, assessed as a ratio of inflammatory cellular components (PMNs plus mononuclear cells) to fibroblast-like cell count, was similar for both cysts and granulomas (3.2 and 3.3 respectively), however the composition of inflammatory activity was different between the lesions, since PMNs predominated in granulomas. Periapical cysts were surrounded by a stratified squamous epithelium and fibroblast-like cells were observed in the stroma and surrounding the lesions. Similar to cysts, fibroblast-like cells were distributed throughout granuloma lesions but no epithelial cell layer was evident.
Out of 113 genes, eleven genes in periapical granulomas and 17 genes in cysts showed higher expression levels than in healthy PDL (Figure 1). Of those overexpressed, ADAM metallopeptidase-1 (ADAMTS1), integrin-β4 (ITGB4), integrin-β7 (ITGB7), laminin-α1 (LAMA1), MMP-2 and tissue inhibitor of metalloproteinase-1 (TIMP1) were similarly expressed in cysts and granulomas.
Extracellular matrix protein-1 (ECM1), integrin-α2B (ITGA2B), MMP-10, MMP-7, integrin-αM (ITGAM), and laminin-β2 (LAMB2) genes were more frequently expressed in cysts, followed by a lower expression in granulomas, and the lowest expression was detected in healthy PDL. Osteonectin (SPARC) gene expression was similar in all tissues.
Integrin-α3, integrin-α5, and integrin-β1 in cysts and transforming growth factor β-in granulomas were significantly elevated in lesions compared to healthy PDL. MMP-24 was detected in granulomas and cysts but not in healthy PDL, and ADAM metallopeptidase-13 and secreted phosphoprotein-1 were found exclusively in cysts but not in granulomas or healthy PDL. Catenin-α1, TIMP-3, and selectin-L were found exclusively in healthy PDL but not in periapical lesions, thus these genes were suppressed in the lesions (Table 2). We selected for presentation only those genes which were up or downregulated in cysts and granulomas. Relative expression for genes not modulated or genes not expressed were not reported.
Given the high prevalence of MMPs and higher MMP-2 mRNA expression in both cysts and granulomas compared to healthy PDL, the presence of MMP activity in tissue sections was further investigated. In situ zymography revealed gelatinolytic activity in both cysts and granulomas. Gelatinolytic activity, detected in the inflammatory cell area, was more intense in granulomas (Figure 2A, 2B) than in cysts (Figure 2C, 2D). Low enzymatic activity was detected in healthy PDL (Figure 2G).
The shape of the DAPI-stained nuclei, indicated that gelatinolytic activity was predominantly in areas with PMNs (Figure 2B). Enzymatic activity was sparse adjacent to mononuclear inflammatory cells and fibroblast-like cells, and absent in the epithelial layer in periapical cysts (data not shown). Sections incubated with EDTA confirmed that proteases responsible for gelatinolytic activity were MMPs (Figure 2E). Absence of autofluorescence was confirmed by incubating tissues with agarose lacking FITC-bound gelatin (Figure 2F).
In situ zymography revealed the presence of generic MMP activity, however it did not reveal the specific MMP profile present in these tissues. Since MMP-2, MMP-9 and MMP-13 use gelatin as a substrate (22, 32), immunostaining for these specific MMPs was performed.
Immunopositive cytoplasmic staining for MMP-2 was observed in a higher number of stromal and mononuclear inflammatory cells compared to PMNs (p < 0.01) in both granulomas and cysts, without differences between them (Figure 2H, 2I, Figure 4A, 4D). In contrast, MMP-9 and MMP-13 were expressed by a greater number of cells in granulomas compared to cysts (p < 0.01) (Figure 2J, 2L, 2M, 2N, Figure 4B, 4C).
MMP-9 was expressed more intensely by inflammatory cells versus stromal cells in both granulomas and cysts. Among these, PMNs were predominantly stained in granulomas (Figure 2J), whereas mononuclear cells were stained to a higher percentage in cysts (Figure 2L). Of all PMNs in periapical granulomas, 95.6% were MMP-9 positive, unlike cysts, which exhibited fewer MMP-9 positive PMNs (Figure 4E). MMP-9 expression in granulomas was observed in areas with high levels of enzymatic activity in PMNs (Figure 3).
MMP-13 was observed primarily in mononuclear cells in both lesions, and to a lesser degree in stromal cells (Figure 2M, 2N). A lower percentage of PMNs was stained in granulomas, however this was higher than that in cysts, where most PMNs were not stained (p < 0.01) (Figure 4C and 4F).
In periapical cysts, epithelial cells were positively stained for MMP-2, MMP-9, and MMP-13 but no enzymatic activity was detected (data not shown), indicating that these MMPs were inactive. IgG staining controls revealed low non-specific staining (Figure 2O and 2P).
Periapical disease pathogenesis involves degradation of several ECM components including collagen, fibronectin, laminin, and small proteoglycans (33). MMPs are likely candidates in this process since these ECM components are known MMP substrates. Using in situ zymography, we demonstrated the presence and increased levels of active MMPs in the inflammatory cell zone in periapical granulomas compared to cysts. Although MMP expression has been immunolocalized in human periapical lesions (11, 24–28, 34), this is the first study to identify areas of MMP enzymatic activity.
Different MMP immunostaining patterns were observed in cysts and granulomas. The minimal epithelial layer in cysts was positively stained for MMP-2, MMP-9, and MMP-13, although it exhibited low MMP enzymatic activity. This may result from the reduced stroma present in this thin epithelial compartment (35). However, the larger stromal compartments of both lesions, where the inflammatory cells are located, exhibited higher MMP enzymatic activity. In lesional stroma, MMP-2 expression was diffuse throughout the ECM, and cells exhibited weak cytoplasmic staining. On the other hand, MMP-9 and MMP-13 staining were localized and highly expressed in the inflammatory cell area, with strong intracellular staining. Interestingly, higher MMP-13 expression and more evident MMP-9 positive staining in PMNs were observed in granulomas compared to cysts.
Simultaneous upregulation of MMP-9 and MMP-13 in granulomas suggests coordinate MMP regulation in the dynamics of periapical disease. Interaction between MMP-9 and MMP-13 could be explained by the ability of MMP-13 to activate MMP-9 (32), thus explaining the more intense enzymatic activity observed in granulomas.
The role of PMNs in periapical disease pathogenesis is well-established (36). Differential expression of MMP-9 in PMNs from granulomas compared to cysts represents an interesting finding, since PMNs uniquely regulate MMP expression. Synthesis of MMP-9 is complete by the time PMNs enter the vasculature, and regulation of this protease is mediated by granule release rather than by transcriptional events (37). Thus, it is not surprising that microarray data did not reveal changes in mRNA expression for MMP-9. Furthermore, MMP activity in general is regulated post-transcriptionally by proteolytic processing and by tissue inhibitors of metalloproteinases, so changes in MMP activity would not necessarily be reflected at the gene level or in the microarray data for any of the MMPs. However, the mechanism involved in the recruitment of a higher percentage of PMNs positively stained for MMP-9 and the higher enzymatic activity observed in granulomas compared to cysts remains to be investigated.
Previous reports showed that MMP-1 expression is restricted to the lining epithelium, subepithelial macrophages, fibroblasts, and endothelial cells in human periapical cysts (11, 29) and to areas showing active bone resorption during periapical disease development in rats (39). However, we observed MMP-9 and MMP-13 staining more intensely in inflammatory cells in these lesions. Taken together, these findings indicate that specific MMPs are produced in different cell types in periapical lesions, suggesting specific roles for these enzymes in periapical disease pathogenesis. Indeed, recent reports have demonstrated distinct roles for MMPs in periapical disease progression, either stimulating (40) or attenuating lesion development (41). These data and our current findings underscore the importance of characterizing the profile of MMPs involved in periapical lesions prior to using pan-MMP inhibitors therapeutically that can non-specifically inhibit MMPs (23).
Using ECM-focused microarrays, we demonstrated that some MMPs are more highly expressed in cysts and granulomas than in healthy PDL. In addition, high gelatinolytic activity was present in granulomas, especially in the inflammatory cell zone, and immunostaining confirmed the presence of MMP-9 and MMP-13 in this process. Recently, gingival crevicular fluid analysis (GCF) revealed that patients with periapical lesions exhibit higher MMP-2 and MMP-9 expression compared to control subjects (29), however the identity of the disease was not further investigated. Since differential MMP-9 and MMP-13 expression was observed between cysts and granulomas in this study, these molecules may serve as molecular markers of the status of periapical disease and could be further explored in vivo in combination with GCF analysis to clinically differentiate periapical cysts from granulomas.
This study was supported by Grant NIH-R01-DE013725 to YLK and a CAPES Foundation Fellowship to FWGPS (0668/07-9). No financial affiliation exists.
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.