Introduction

Ameloblastomas (AME) are odontogenic tumours of epithelial origin. Although classified as a form of benign tumour, AME is characterised by aggressiveness, an infiltrative nature, and a high recurrence tendency. They can grow to large dimensions and invade adjacent structures, causing significant morbidity [1,2,3].

According to the most recent classification by the World Health Organisation (WHO), AME can be categorised into three types: conventional, unicystic, and extraosseous/peripheral. Among these, conventional AME is the most common and most aggressive [4].

The current therapeutic options include both conservative and radical approaches [5]. Conservative treatments involve enucleation and curettage [1], which avoid relevant facial deformities [5] but have recurrence rates of up to 55% [6]. More invasive treatments involve marginal resection, which is the method of choice, considering the high recurrence rates reported for AME [1]. However, this method can result in relapse rates of up to 15% in more aggressive AME types as well as significant facial deformities [7].

Although the causes of the aggressive biological behaviour and high rate of recurrence of this benign neoplasm are not fully understood, studies involving stem cells and their possible relationship with the aetiopathogenesis of neoplasms have been relevant in elucidating this behaviour [8,9,10].

Stem cells can perpetuate themselves through self-renewal mechanisms and differentiate into cells in specific tissues. These mechanisms are like those that occur in tumour cells and involve similar signalling pathways. Thus, both normal stem cells and tumorigenic cells have proliferative potential and the ability to give rise to new tissues called cancer stem cells [8].

Cancer stem cells proliferate uncontrollably, driving the formation and growth of tumors, generating heterogeneous malignant cells associated with recurrence and metastasis. It is believed that cancer cells can appropriate the self-renewal machinery that is normally expressed in normal stem cells [8].

It has been reported that AME cells originate from odontogenic stem cells located in the dental lamina [11] and that tumours are likely initiated in normal stem cells that contain a perpetual minority of cancer stem cells [12, https://imagej.nih.gov/ij/plugins/ihc-toolbox/index.html) using ImageJ (public domain software developed by Wayne Rasband (NIMH, National Institutes of Health, Bethesda, MD, USA; http://rsbweb.nih.gov/ij/). After segmenting the images, the area and diaminobenzidine (DAB) staining fraction were measured, and the immunostaining differences found in AME, DC, and DF were analysed.

Indirect immunofluorescence

AME-hTERT cells were cultured on glass coverslips in 24-well plates and subjected to indirect immunofluorescence to observe the expression of SALL4, LIN28A, and KLF4. The technique involved the following steps: cell fixation in 2% paraformaldehyde for 10 min, washing with phosphate-buffered saline (PBS), permeabilisation of the membrane with a 0.5% Triton X-100 solution (Sigma®) for 15 min, wash with PBS, incubation in PBS/1% BSA for 30 min, and incubation with primary monoclonal antibodies diluted in PBS/BSA at 1% for a minimum of 12 h and a maximum of 18 h in a humid chamber at 4ºC. The primary antibodies used were Anti-Sall4 (1:50, mouse, Santa Cruz Biotechnology®), Anti-LIN-28 (1:50, mouse, Santa Cruz Biotechnology®), and anti-GKLF (1:50, mouse, Santa Cruz Biotechnology®). To detect the primary antibody, incubation was performed in a solution containing the secondary antibody conjugated to Alexa Fluor 488 (Invitrogen, Carlsbad, CA, USA) for 1 h in a humid and dark chamber at room temperature. Nuclei were stained with Hoechst 33258 (1: 2,000, Sigma) and cytoskeletons were stained with Alexa Fluor 568 phalloidin (Life Technologies, Carlsbad, CA, USA). The coverslips were immersed in PBS and distilled water and mounted using the ProLong® Gold antifade reagent (Invitrogen®). Next, the cells were analysed using a fluorescence microscope (AxioScope.A1, Zeiss) equipped with a digital camera (AxiocamMRc, Zeiss). Images of the slides were obtained for the registration of immunoexpression using a 40x objective. As a negative control, the primary antibody was replaced with a non-immune serum.

Statistical analysis

Data were analysed using GraphPad Prism 8 software (GraphPad Software Inc., San Diego, CA, USA). When parametric distribution was evidenced by the Shapiro-Wilk test, differences between groups were evaluated by one-way analysis of variance (ANOVA) followed by Tukey’s post-hoc test. When a non-parametric distribution was evidenced by the Shapiro-Wilk test, the differences between groups were evaluated by the Kruskal-Wallis test, followed by Dunn’s post-test of multiple comparisons. A 95% confidence interval (CI) was assumed (α = 0.05).

Results

Demographical and clinical data, and histopathological ty** of patients with AME

In the studied sample, the mean age was 37 years. Males represented 57% of the samples. The region of greatest involvement was the mandible, totalling 95% of the cases. As for the histological types, eight cases were of the follicular type, eight of plexiform, three acanthomatous and two of granular cells (see Table 1).

Table 1 Demographical and clinical data, and histopathological ty** of patients with AME (n = 21)
Full size table

Immunohistochemical staining for SALL4, LIN28a, and KLF4 in ameloblastoma, dentigerous cyst, and dental follicle

IHC staining for SALL4, LIN28A, and KLF4 was predominantly observed in the epithelial cells of the tumour islands (see Fig. 2). For SALL4, intense staining was observed in the tumour parenchyma in both the nucleus and cytoplasm of the cells, in the high columnar cells of the periphery, and in the central cells of the tumour island. LIN28A showed strong immunostaining with nuclear and cytoplasmic localisation limited to the central cells of the tumour island. Intense immunostaining was observed for KLF4 nuclear localisation in epithelial cells. The labelling was predominantly located in the nuclei of tall columnar cells at the periphery of the tumour island. Nuclear and cytoplasmic markings of SALL4, LIN28A, and KLF4 were also observed in some cases of DC and DF; however, the expression of SALL4, LIN28A, and KLF4 was significantly higher in AME samples compared to DC (p < 0.001) and DF (p < 0.001), as demonstrated in the statistical analysis (see Fig. 2).

Fig. 2
figure 2

Immunohistochemical staining for SALL4, LIN28 and KLF4 in AME, DC and DF samples. Intense SALL4 immunostaining was observed in the parenchymal cells of the plexiform AME, with nuclear and cytoplasmic location (A). Strong LIN28A immunostaining was observed in follicular-type AME, with nuclear and cytoplasmic localization limited to the central cells of the tumor island (E). Intense KLF4 immunostaining in plexiform AME was observed with a nuclear location (I). There was a subtle labeling of the three proteins in the nucleus and cytoplasm of some epithelial cells in both DC (B, F and J) and DF (C, G and K) samples. Statistical analysis of the percentage of parenchymal marking area of the three markers (D, H and L) between AME, DC and DF. (***p < 0.001). Scale bar: 20 µm. AME = Ameloblastoma; DC = Dentigerous cyst; DF = Dental follicle

Full size image

AME-hTERT lineage expressed SALL4, LIN28A, and KLF4

The immunofluorescence assays revealed that the AME-hTERT strain expressed SALL4, LIN28A, and KLF4 (see Fig. 3). Neoplastic cells demonstrated nuclear and cytoplasmic expression of SALL4 (see Fig. 3A), nuclear and cytoplasmic expression, with nuclear predominance of LIN28A (see Fig. 3E) and predominantly nuclear expression of KLF4 (Fig. 3I). The immunoexpression of all proteins was granular (see Fig. 3).

Fig. 3
figure 3

Immunofluorescence photomicrographs of SALL4, LIN28 and KLF4 in AME-hTERT strains. SALL4 showed immunoexpression with a granular pattern located in the nucleus and cytoplasm (A). LIN28A showed immunoexpression with a granular pattern in the nucleus and cytoplasm, with nuclear predominance (E). Granular expression of KLF4 was observed in the nuclei, showing faint staining in the cytoplasm (I). Control group (CT) (M). The cytoskeleton was stained with phalloidin (red), and the nuclei were stained with Hoechst 33258 (blue). Overlap** images of the expression of SALL4 (D), LIN28 (H), KLF4 (L) and control group (P). Scale bar: 20 µm

Full size image

Discussion

The studied proteins showed significantly higher levels of immunostaining in AME cells than in the DC and DF cells. SALL4, LIN28A, and KLF4 were expressed in the AME parenchyma, with slight staining observed in some cells of the odontogenic epithelium of DC and DF. Furthermore, immunoexpression of the studied proteins was observed in the AME-hTERT strain.

SALL4 is a transcription factor that plays a key role in maintaining pluripotency and self-renewal of embryonic stem cells [14]. It interacts with other important regulatory proteins of embryonic pluripotency — OCT4 (octamer-binding protein 4), SOX-2 (HMG-box gene 2 related to SRY), and NANOG (homeodomain protein) — forming an autoregulatory circuit in which each of these proteins regulates its own expression and that of others [33,34,35]. High expression of SOX-2, NANOG, and OCT4 has been demonstrated in AME [10], suggesting that these proteins may act together to maintain undifferentiated stem cells in this tumour.

In the present study, cells from the AME tumour islands and the AME-hTERT lineage showed nuclear and cytoplasmic expression of SALL4, corroborating the marking pattern found by other studies in oral squamous cell carcinoma that suggests that this protein plays an important role in the progression of oral cancer and may serve as a potential therapeutic target [36,37,38]. Nuclear labelling indicated the transcriptional activity of SALL4. Such activity is associated with transcriptional repression mechanisms that prevent stem cell differentiation and increase the proliferation of undifferentiated cells [35, 39]. SALL4 cytoplasmic marking has also been demonstrated in breast cancer cells and is considered a predictor of poor prognosis [40].

Studies have shown that SALL4 protein expression is negatively regulated by miRNAs (miRNAs) belonging to the Let-7 family, particularly by miR-98, which leads to a reduction in tumour cell proliferation, indicating that miR-98 acts as a tumour suppressor that inhibits SALL4 protein expression [41, 42]. It is important to emphasise that LIN28 protein can downregulate the Let-7 microRNA family through the activation of its isoforms LIN28A or LIN28B [25, 43]. This suggests that the upregulation of LIN28 leads to the inhibition of miR-98, which, in turn, leads to the upregulation of SALL4. Therefore, the co-expression of SALL4 and LIN28A in AME observed in this study may play a significant role in tumour pathogenesis.

LIN28A is a highly conserved RNA-binding protein that plays a significant role in development, glucose metabolism, and pluripotency [44,45,46]. It is highly expressed in mouse embryonic stem cells, which decreases after differentiation, and in human embryonic carcinoma cells [47]. Oral squamous cell carcinoma has been demonstrated to be associated with the regulation of the proliferative and invasive activities of this neoplasm [48]. Furthermore, LIN28A has been identified as one of the four factors that convert fibroblasts into induced pluripotent stem cells, corroborating the role of this protein in pluripotent stem cells [49].

In this study, LIN28A immunostaining showed nuclear and cytoplasmic localisation limited to the central cells of the AME tumour islands. The same expression pattern was observed for the AME-hTERT strain. LIN28A is predominantly cytoplasmic and located on ribosomes, P bodies, and cytoplasmic stress granules [50]. The cytoplasmic expression found in the present study may be associated with its performance in the recruitment of terminal uridylyl transferase (TUTase4) ZCCHC11, which inhibits Let-7 processing in the cytoplasm and selectively blocks the expression of Let-7 miRNAs and their functions tumour suppressors, acting as an oncogene and promoting tumorigenesis [25, 51]. This action has been demonstrated in embryonic stem cells, suggesting a significant role of LIN28A in inhibiting cell differentiation through miRNAs in stem cells and certain types of cancer [52].

Nuclear expression of LIN28A was observed when both RNA-binding domains were mutated [50]. A model has been proposed in which LIN28A regulates the post-transcriptional processing of its mRNA targets by first binding to these targets in the nucleus and subsequently transporting them between ribosomes, P bodies, or stress granules for translation regulation, depending on the environmental conditions [50, 53]; however, more studies are needed to better understand this process.

The central region of the AME tumour islands, which exhibits greater LIN28A labelling, is more prone to hypoxia. As the tumour progresses, the concentration of oxygen in the microenvironment around the tumour cells decreases, leading to intratumoural hypoxia [54]. In response to this condition, hypoxia-induced factor-1 alpha (HIF-1α) regulates the expression of genes that help cells adapt to this environment [55]. Studies have indicated that HIF-1α is overexpressed in AME, suggesting that hypoxia is related to proliferation and invasion of the solid areas of this tumour [56,57,58]. It has been shown that HIF-1α binds directly to the LIN28A promoter and induces its transcription [59] and that hypoxia is capable of inducing the expression of stem cell markers in cancer cell lines, thereby contributing to the dedifferentiation and reprogramming process that induces the formation of cancer stem cells [59,60,61]. From this, we can infer that the expression of LIN28A in the central cells of the AME tumour island close to the high columnar cells in the periphery may be associated with the adaptive response of tumour cells to hypoxia, inducing the dedifferentiation of peripheral cells, and thus promoting greater proliferation and invasion.

KLF4 is an essential transcription factor in the regulation of cellular processes (e.g. development, differentiation, proliferation, and apoptosis) [16] and in the renewal of stem cells and maintenance of pluripotency [29, 30]. It has been used as a reprogramming factor for fibroblasts and odontoblasts in induced pluripotent stem cells along with LIN28A [62, 63]. In different cell types, KLF4 functions as both a tumour suppressor and an oncogene [16]. Increased expression has been reported in human head and neck squamous cell carcinoma and is associated with poor prognosis and aggressiveness, corroborating its oncogenic role [64, 65]. In contrast, Land et al. [66] found an association between high KLF4 expression and a favourable prognosis. Another study reported the role of KLF4 in oral squamous cell carcinoma, in which mechanisms of action were described as both tumour suppressors and oncogenes [67]. Some scholars believe that the function of KLF4 as an oncogene or tumour suppressor is modulated by its complex interactions with several tumour microenvironments [68].

In the present study, intense nuclear immunostaining for KLF4 was observed in AME, predominantly in the tall columnar cells located on the periphery of the tumour island. In the AME-hTERT strain, nuclear expression of KLF4 with mild cytoplasmic expression was observed. KLF4 is mainly located in the nucleus, but its cytoplasmic localisation has also been reported in prostate and oral cancers [65, 68, 69]. Increased KLF4 nuclear expression has been associated with poor prognosis in patients with breast and head and neck cancer [64, 70]. However, another study suggested that the downregulation of KLF4 is associated with the progression of oral carcinoma [71]. Considering the ambiguity of this transcription factor, further studies are required to assess the role of KLF4 in AME.

The findings of the present study indicate that SALL4 and LIN28A may play a significant role in the biological behaviour of AME, suggesting a possible role for stem cells in the genesis and progression of AME. The KLF4 transcription factor plays a context-dependent role in carcinogenesis and may be up or downregulated in distinct types of cancer. Therefore, its role in AME needs to be better understood. However, considering its expression together with that of other studied proteins, we suggest its participation and interaction as an oncogene. Although these results are promising, mechanistic and in vivo studies are required to confirm these hypotheses and elucidate the underlying molecular mechanisms. Understanding these mechanisms may have significant implications for the diagnosis, prognosis, and treatment of AME, thus opening up new possibilities for personalised and effective therapies.

Conclusions

To the best of our knowledge, this is the first study to evaluate the expression of SALL4, LIN28A, and KLF4 proteins in a benign odontogenic tumour. The study results verify the expression of these stem cell markers in AME neoplastic cells by IHC and in the AME-hTERT cell line by immunofluorescence, suggesting the possible participation of stem cells in the origin, progression, and recurrence of this tumour.