Human fibrous epulis specimens were collected for investigating the histopathologic features of the hyperplastic gingival tissues with informed consent received from the lesion donors. Histological sections of fibrous epulis (PF and POF) were examined with hematoxylin and eosin staining and immunofluorescence staining. The fibrous epulides were characterized by the presence of a layer of hyperplastic epithelium overlying the fibrous connective tissue.35 Mineralized tissues, which were only present in POF, consisted of abundant mineralized bony tissue and dystrophic or cementum-like calcification (Fig. 1).
To identify whether fibrocytes were present in fibrous epulides, human fibrous epulis specimens were detected with immunofluorescence staining and confocal laser scanning microscopy (CLSM). Fibrocytes are characterized by the co-expression of CD45 and collagen type I (Col-1) [36–38]. Figure 2a shows the presence of scattered CD45+ Col-1+ fibrocytes within the human fibrous epulis specimens. The fibrocytes in the local tissue are mostly derived from the peripheral blood.
Mature fibrous epulides are highly vascularized. This enables migration of PBMCs into the lesions upon local inflammatory stimulation [35]. Previous studies reported that PBMCs possess the potential to rapidly differentiate into fibrocytes in the presence of inflammatory signals [21, 39, 40]. Accordingly, peripheral blood was collected from human volunteers to isolate and induce PBMCs to differentiate into fibrocytes in vitro (Fig. S1a). The number of spindle-like cells in the isolated PBMCs was found to increase with lengthening of the culture time (Fig. S1b). The ratio of spindle-like cells was as high as 80% after the PBMCs were incubated for 12 days (Fig. S1c). Those spindle-like cells were identified to be fibrocytes by immunofluorescence staining (Fig. S1d). 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) assay showed that the proliferation activity of fibrocytes cultured in vitro increased with time (Fig. S1e).
Myofibroblasts are responsible for the synthesis, deposition and remodelling of the ECM (especially for fibrillar collagen) in tissue fibrosis [41, 42]. Fibrocytes proliferate and differentiate into myofibroblasts in response to TGF-β1. Immunohistochemical staining showed that TGF-β1 was significantly increased in PF and POF compared with normal gingival tissues (Fig. S2a, S2b). The distribution of fibrocytes were co-incident with the expression of TGF-β1 in the fibrous epulides (Fig. S2c-S2f). This observation suggests that the fibrocytes in fibrous epulides are in an activated state for differentiation.
To determine whether the fibrocytes differentiate into myofibroblasts under the stimulation of TGF-β1 in vivo, we further detected the cellular constituents in human fibrous epulis specimens. Mature myofibroblasts are characterized by the expression of α-smooth muscle actin (α-SMA) which endows myofibroblasts with the ability to remodel fibrillar collagen [43, 44]. Immunofluorescence staining of the human fibrous epulis specimens showed that the α-SMA was considerably upregulated in those lesions, when compared with normal gingival tissues (Fig. 3a, 3b). Although CD45 and Col-1 were co-localized with α-SMA, the fluorescence intensity of α-SMA was significantly higher than those exhibited by CD45 or Col-1 (Fig. 3c, 3d). This feature may be attributed to the differentiation of fibrocytes within the into myofibroblasts, which is accompanied by the loss of fibrocyte markers.
Based on the experimental in vivo results, in vitro experiments were further conducted by stimulating isolated fibrocytes with different concentration of TGF-β1 (1ng/mL, 5ng/mL, 10ng/mL) for 3 days (Fig. 4a). Representative immunofluorescence images showed that myofibroblasts were detected at 3 days and the expression of α-SMA was dependent on TGF-β1 concentration (Fig. 4b, 4c). In addition, fluorescence intensity of CD45 decreased with increasing TGF-β1 concentration. This is suggestive of a gradual loss of fibrocyte phenotype and differentiation of the fibrocytes into myofibroblasts (Fig. 4b, 4c). In contrast, fibrocytes cultured in vitro without TGF-β1 stimulation did not express α-SMA until they were cultured for 12 days (Fig. S3). Results derived from reverse transcription-quantitative polymerase chain reaction (RT-qPCR) showed that gene expressions of α-SMA was significantly increased at day 7, 14 and 21, when compared with control cells that were devoid of TGF-β1 stimulation (p < 0.05). These results validated that TGF-β1 accelerates the differentiation of fibrocytes into myofibroblasts in vitro.
To determine if fibrocytes are also responsible to the formation of calcified tissues in POF, those tissues were examined using dispersive X-ray spectroscopy (Fig. S4a). The results indicated that the calcified POF tissue contain calcium and phosphorus. Micro-Fourier transform infrared spectroscopy determined that the calcified tissues contain calcium phosphate, based on identification of the υ3 P-O stretching vibration at ~ 1030 cm− 1 (Fig. S4b-f). Apart from calcium phosphate, the characteristic amide I (~ 1640 cm− 1) and amide II (~ 1545 cm− 1) peaks of type I collagen were identified. This finding indicates that collagen fibrils are present in the matrix of fibrous epulides.
Previous study reported that fibrocytes can differentiate into osteoblasts after they were incubated in an osteogenic induction medium in vitro [17]. However, there was no reported in vivo evidence of the osteogenic differentiation of fibrocytes. Immunofluorescence staining was utilized to examine the existence of CD45+/Col-1+ OCN (osteocalcin)+ cells in the fibrous epulis specimens (Fig. 5a, 5b). Both CD45 and Col-1 were co-localized with OCN; the fluorescence intensity of OCN was higher than that of CD45 or Col-1 (Fig. 5c, 5d). Differentiation of the fibrocytes into osteoblasts could have resulted in the loss of fibrocyte fluorescent markers. A previous study reported that the inflammatory microenvironment of heterotopic ossification triggered the release of endogenous factors that stimulated the infiltration of mesenchymal cells and osteoprogenitor cells [45]. The present work provided first-hand in vivo evidence of the differentiation of fibrocytes to osteoblasts. The latter are responsible for the formation of calcified tissues in POF. Semi-quantitative statistics of the immunofluorescence staining indicated that the expression of OCN in POF was higher than the OCN expression in PF (Fig. 5c, 5d). This explains why calcified nodules are present in POF but not in PF.
The expression of myofibroblast markers in the vicinity of the calcified POF tissues was subsequently examined. Expression of α-SMA in the calcified tissues was localized to the blood vessels, whereas, α-SMA was scattered within the PF lesions (Fig. S5). This suggests that fibrocytes in the POF lesions differentiate into osteoblasts, while those in the PF lesions differentiate into myofibroblasts. The factors responsible for the differential differentiation of fibrocytes in POF and PF lesions were examined by SEM-energy dispersive spectrometry (SEM-EDS) to determine the element compositions within the ECM of those lesions. Surprisingly, higher concentrations of Ca and P were identified in the uncalcified tissues in POF, compared with those tissues in PF (P < 0.0001) (Fig. 6a-c). This result suggests that the different microenvironments with the POF and PF lesions are responsible for the different differentiation tendencies of the fibrocytes.
The aforementioned results suggest that higher concentrations of Ca/P in POF promote the differentiation of fibrocytes into osteoblasts. This notion was examined by mimicking the high concentration Ca/P matrix environment of the POF in vitro. The PBMCs were isolated and induced to differentiate into fibrocytes, as previously mentioned. When the number of fibrocytes increased to 80% confluency, different conditioned media (blank medium, TGF-β1, Ca/P, TGF-β1 + Ca/P and osteogenic induction medium) were used to stimulate the PBMCs for 7 days (Fig. 7a). Cells cultured in blank medium were used as the negative control and those cultured in osteogenic medium were used as the positive control. Then, CLSM was used to observe the expression of fibrocytes (Col-1) and osteoblasts (OCN) among the cultured cells. As shown in Fig. 7b, the expression of OCN in cells cultured with a high concentration of Ca/P was significantly higher than those cultured in the blank medium or treated with TGF-β1 only. The expression of OCN was further increased in the group treated with Ca/P and TGF-β1 (P < 0.0001) (Fig. 7c, 7d). Alizarin red S staining for calcium ions confirmed that the high Ca/P concentration of drove the fibrocytes to differentiate to osteoblasts; such an effect was time-dependent (Fig. 7e, 7f). Quantitative real-time polymerase chain reaction was further used to evaluate the osteogenic potential of fibrocytes after they were cultured in those media for 7 days, 14 days and 21 days. The results showed that osteogenic differentiation-related genes, including alkaline phosphatase (ALP), runt-related transcription factor 2 (Runx2) and OCN were all significantly upregulated in the groups treated with high Ca/P concentration, especially in the TGF-β1 + Ca/P group (P < 0.05) (Fig. 7g). These in vitro results validated that a high Ca/P concentration causes the fibrocytes to differentiate into osteoblasts and secrete the calcified nodules in POF. That said, the reason why POF contain high concentrations of Ca/P in their ECM is unknown and requires further investigation.
The present study identified, for the first time, the existence of CD45+ and Col-1+ fibrocytes in human fibrous epulides. Cytokines such as TGF-β1 recruit PBMCs to the site of gingival inflammation. These monocytes further differentiate into fibrocytes to participate in tissue repair. In the presence of TGF-β1, the fibrocytes differentiate into myofibroblasts and trigger gingival fibrosis. Instead of differentiating into myofibroblasts, the fibrocytes differentiate into osteoblasts in the presence of high concentrations of Ca/P in the ECM. The osteoblasts secrete mineralized nodules in the calcified fibrous epulides. Clinically, incomplete elimination of these gingival lesions during surgical resection and locally persistent chronic inflammation may continue to aggravate the fibrocytes locally. The aggregated fibrocytes proliferate and differentiate in response to the anti-inflammatory cytokines and initiate recurrent fibrous epulides. Hence, elimination of the local inflammatory environment to control the aggregation of fibrocytes may be a viable strategy to inhibit the recurrence of fibrous epulides. New therapeutic strategies such as carbon dioxide laser, sclerotherapy and cryotherapy are available in clinical practice to supplement surgical resection in combating fibrous epulides [46–50]. Further research is necessary to identify a multidisciplinary strategy to regulate the inflammatory environment and inactivate the chemokines relevant to fibrocytes transformation and proliferation.