Expression of epithelial growth factors and of apoptosis-regulating proteins, and presence of CD57+ cells in the development of inflammatory periapical lesions

Abstract The mechanisms that stimulate the proliferation of epithelial cells in inflammatory periapical lesions are not completely understood and the literature suggests that changes in the balance between apoptosis and immunity regulation appear to influence this process. Objective: To evaluate the expression of the epidermal growth factor (EGF), its receptor (EGFR) and of the keratinocyte growth factor (KGF), the presence of CD57+ cells, the epithelial cell proliferation index, and the expression of the Bcl-2 protein in inflammatory periapical lesions (IPL) at different stages of development. Methodology: Our sample was composed of 52 IPLs (22 periapical granulomas - PG - and 30 periapical cysts - PC), divided into three groups: PGs, small PCs, and large PCs. Specimens were processed for histopathologic and immunohistochemical analyses. Sections were evaluated according to the amount of positive staining for each antibody. Results: We found no significant differences among the groups regarding Bcl-2 (p=0.328) and Ki-67 (p>0.05) expression or the presence of CD57+ cells (p=0.748). EGF (p=0.0001) and KGF (p=0.0001) expression was more frequent in PCs than in PGs, and CD57+ cells were more frequent in IPLs with intense inflammatory infiltrates (p=0.0001). We found no significant differences in KGF (p=0.423), Bcl-2 (p=0.943), and EGF (p=0.53) expression in relation to inflammatory infiltrates or to the type of PC epithelial lining, but observed greater KGF expression (p=0.0001) in initial PCs. EGFR expression was similar among the groups (p>0.05). Conclusion: More frequent EGF and KGF expression in PCs and the greater presence of CD57+ cells in lesions with intense inflammatory infiltrates suggest that these factors influence IPL development. The greater KGF expression in initial PCs suggests its importance for the initial stages of PC formation.


Introduction
Infections in the root canal system can cause a group of inflammatory conditions known as inflammatory periradicular lesions (IPLs). The most common IPLs are periapical granulomas (PG) and periapical cysts (PC); both appear as unilocular radiolucent lesions with well-defined borders associated with necrotic teeth. 1,2 The histological characterization of PGs involve connective tissue infiltrated by a predominantlty chronic inflammatory infiltrate, small vessels, and, sometimes, proliferating epithelial islands that still fail to produce an inner cavity. [1][2][3] In turn, PCs consist of cystic cavities lined by stratified squamous epithelia of irregular thickness showing espongiosis and containing a fibrous capsule of connective tissue which holds the same components found in PG. [1][2][3] The literature claims PCs originate from pre-existing PGs; however, we still lack the full understanding of the exact mechanisms modulating this progression. 3 The literature shows that, in response to an inflammatory stimulus, the proliferation of quiescent epithelial cell rests of Malassez within the periodontal ligament originate the non-keratinized stratified squamous epithelium that lines PCs. 3 Various phenomena stimulating the proliferation of epithelial cells appear to influence this process, such as the higher expression of growth factors and their receptors, and changes in the balance between apoptosis and immunity regulation. [4][5][6][7][8] The CD57 protein relates to cytotoxic activity and mediate immunosuppression, 9 whereas Bcl-2, to cell division regulation, inhibiting apoptosis in several cell types. 10 The epidermal growth factor (EGF) stimulates the proliferation of keratinocytes and fibroblasts, and relates to cell proliferation and wound healing. 5 Its receptor, EGFR, is found in epithelial cells with proliferative potential. 4 The keratinocyte growth factor (KGF) also correlates with epithelial cell growth and differentiation, which inflammatory stimuli influence. 11 The literature has previously shown the individual importance of these cell types, growth factors (and their receptors), and anti-apoptotic proteins for IPLs. [4][5][6][7][8]  Specimens were subjected to conventional histological processing and embedded in paraffin wax.
The clinical (lesion site, and clinical signs and symptoms), demographical (patients' age and gender), and radiographical information (lesion size and border outline) of all specimens were obtained from laboratory files. Specimens from patients with immunological disorders such as diabetes, HIV infection, and autoimmune diseases were excluded, as were cases in which the surgical specimens were insufficient for adequate histological analyses or obtained by incisional biopsy.
The IPLs were divided into three groups: PGs (22 cases, 42%), small PCs (nine cases, 17%radiographical diameter smaller than 10 mm) and large PCs (21 cases, 41% -diameter greater than 10 mm); the last two groups were classified based on the greater diameter observed in high-quality periapical radiographs.
Paraffin blocks containing the selected specimens were cut into 5 µm-thick sections, stained with hematoxylin and eosin, and observed under an optical microscope (Leica DM500, Heerbrugg, Switzerland). The two fields selected for observation under high magnification were Ki-67 expression "hot spots". All slides were analyzed by two previously calibrated examiners. In case of disagreement, the final decision was based on a group discussion. The methods used for the histological and immunohistochemical analyses have been validated and previously used in other studies performed by the same group. [12][13][14][15] Clinical, demographic, radiographic, histological, and immunohistochemical data were recorded and

Results
The IPLs included in this study affected 32 women (62%) with a mean age of 44.9 years, and 20 men (38%) with a mean age of 41.9 years, with no significant differences in age and gender distribution among the groups (Table 1). Symptoms were present in 36% of the IPLs, without differences among the three groups, but local swelling was more common in small PCs (Table 1). The mean diameter of the IPLs was 14.4 mm, based on the largest dimension observed radiographically, and we considered most case images as well-defined, with no group differences ( Table 1).
The mean size of the gross specimens was 336.71 mm 3 , larger in large PCs than in PGs and small PCs (Table 2). Most cases showed moderate/intense chronic inflammatory infiltrates, but no difference among our three groups ( Table 2). We observed no significant difference in the thickness of the epithelial lining between small and large PCs ( Table 2). We also found cholesterol crystals clefts in three PCs (10%).     or moderate/intense inflammation had, respectively, 3.1 (DP±4.1) and 3 (DP±2.8) mean scores (p=0.951).
We found CD57+ cells in 54% of the IPLs, with no difference on their frequency among the groups (Table 3), and Bcl-2 expression in 65% of the IPLs, with no significant difference on its expression among the groups (Table 3). We observed EGF and KGF expression in, respectively, 54% and 58% of the IPLs, both more common in PCs than in PGs (Table   3). However, we found no statistically significant difference in the frequency of CD57, KGF, EGF, and

EGFR immunopositive cells in PGs, and small and large
PCs. We observed no statistically significant difference in the presence of symptoms and local swelling according to our histological parameters, the presence of CD57+ cells, and Bcl-2, EGF, and KGF expression.
When comparing the frequency of CD57, KGF, EGF, and EGFR immunopositive cells to the presence of symptoms and local swelling, EGF-positive cells were more frequent in symptomatic (53.75%±23.058) than in asymptomatic (37.59±22.104) cases (p=0.03).
We found CD57+ cells more frequently in IPLs with moderate/intense inflammatory infiltrates (Table 4), and no difference in Bcl-2, EGF, and KGF expression according to the intensity of inflammatory infiltrates (

Discussion
The role of epithelial growth factors, apoptosisregulating proteins, and CD57+ cells (including natural killer cells and CD8+ T cells) in the pathogenesis and development of PGs and PCs has been previously investigated. [4][5][6][7][16][17][18] However, few studies evaluated the importance of these factors by comparing inflammatory periradicular lesions at different stages of development.
The CD57 protein is a marker expressed by natural killer cells, which possess a natural cytotoxic ability against tumor cells. 9 Moreover, CD8+ T cells expressing the CD57 protein play an important role in mediating immunosuppression. 9 In this study, we found CD57+ cells in approximately half of the lesions studied -a greater frequency than previously reported in the literature. 19 We observed a greater presence of CD57+ cells in IPLs with moderate/intense inflammatory infiltrates, which may relate to a greater inflammatory response in general. 8 We found no significant differences among the groups or between PCs with different epithelial linings, although previous studies have reported that PCs with atrophic epithelial lining show significantly higher numbers of CD57+ cells. 8,20 Our results suggest that the cytotoxic activity of CD57+ cells is not associated to the proliferation of epithelial cells in the lining of PCs.
The Bcl-2 protein is an important mediator of cell division regulation. It can inhibit apoptosis and extend the survival of several types of cells, which may stimulate the development of PCs and other types of odontogenic cysts. 7,10,17,21,22 In this study, comparing the groups showed no difference in Bcl- EGF stimulates the proliferation of epithelial cells, hepatocytes, and fibroblasts via its interaction with receptors in the membrane of target cells. 5,24,25 The literature associates this protein with various vital functions in normal and altered cells, such as migration, proliferation, mobility, and wound healing. 5,24,25 Our results show no difference in EGF expression regarding the type and intensity of inflammatory infiltrates and the type of epithelial lining in the cystic lesions. Still,  when comparing PGs and PCs, the latter showed greater EGF expression. These findings corroborate previous studies 5 and support the hypothesis that EGF expression plays a role in stimulating epithelial cell proliferation, thus taking part in cyst formation. We should mention, however, that we found no difference in EGF expression between small and large PCs.

Intensity of the inflammatory infiltrate
We normally observe EGFR expression in epithelial cells with proliferative potential. 4,16,[25][26][27] EGFR activation by EGF binding increases tyrosine kynase activity and stimulates signal transduction cascades, which are associated with epithelial cell proliferation. 4,16,[25][26][27] The literature has shown its expression in the epithelia of IPLs, especially in the lining of PCs, suggesting that EGFR expression may be important for maintaining these lesions. 4,16,28,29 Additionally, the presence of an infection seems to modulate EGFR expression in odontogenic cysts, and may influence the proliferative ability of odontogenic epithelia. 27 In our study, however, EGFR expression failed to show a correlation with any of the parameters evaluated, including between small and large PCs.
Fibroblasts produce KGF, a mitogen involved in epithelial cell growth and differentiation. KGF also contributes toward vital functions such as tissue repair. The presence of inflammation in some chronic oral conditions seems to influence its expression. 11,30 In this study, KGF showed greater expression in PCs than in PGs, and in small PCs than in large PCs. These differences suggest that KGF may play a role in the mechanism of PC formation and that its effect is more significant in the initial stages of cyst development, possibly similar to its epithelial proliferative effect during wound healing. 11,30 KGF expression was not significantly different to the intensity and type of inflammatory infiltrates, contrasting with the to what a previous study reports. 14

Conclusions
The results from this study indicated that the more frequent EGF and KGF expression in PCs and the greater presence of CD57+ cells in lesions with intense inflammatory infiltrates suggest these factors influence the development of IPLs. Moreover, the greater KGF expression in initial PCs suggests its importance in the initial stages of PC development.

Conflict of interest
The authors have stated explicitly that there are no conflicts of interest in connection with this article.