Pathologically enhanced apoptosis in the epithelium of OLP lesions has been reported in the literature [4, 29]. In this regard, p53-mediated apoptotic pathway is among the most studied mechanisms in the pathogenesis of the disease. The “Guardian of the genome” p53 plays a crucial role in cell cycle control, triggering either cell cycle arrest or apoptosis, thereby preventing the replication of damaged DNA. During the cell cycle arrest this marker induces transcription of proteins involved in DNA repair [18]. However, if the DNA damage is irreversible, p53 activates programmed cell death through several mechanisms: precedes the release of cytochrome C and activation of procaspase-3; suppresses anti-apoptotic Bcl‐2 and Bcl‐xL and up-regulates pro-apoptotic BAX [30]. The reported p53 immunopositivity in OLP lesions varies from 18 to 100% among different studies [31]. According to Basheer S et al. a possible explanation for this large variation in the data in the literature may be the inclusion of oral lichenoid reaction (OLR) and oral lichenoid dysplasia (OLD) in the OLP contingent due to misdiagnosis; as well as different p53 antibodies and antigen retrieval methods used [31]. Another issue is whether given levels of p53 should be considered as pathologically increased expression. The confusion stems from the lack of constant expression pattern for the marker in normal oral mucosa. For example, Hadzi-Mihailovic M et al. demonstrated that 80% of OLP specimens were p53-positive but did not interpret these results as significantly increased expression [18]. In contrast, in the study of Basheer S. et al. immunoreactivity for p53 was found in 40% of the OLP cases and was considered overexpression [31]. The rationale: in the mentioned studies, p53 was expressed in 53.85% [18] and in none [31] of the control sections of normal oral mucosa, respectively. The prevailing view is that in normal cells p53 staining is absent [31, 32] or weak [19, 33, 34] and thus any protein levels other than these in OLP are considered pathologically enhanced [19, 32, 34]. p53 overexpression in OLP, on the other hand, has different explanations given by the scientific literature. Gudkov AV et al. stated that the inflammatory infiltrate seen in OLP causes keratinocyte DNA damage and higher levels of p53 represent a protective response to maintain genomic stability by activating cycle arrest or programmed cell death [35]. Another concept is that p53 overexpression is a physiologic reaction to the reported hyperproliferative state in OLP and the goal is again DNA damage repair or elimination [16, 18]. However, most authors interpret the higher detection of p53 in OLP tissues as an indication of а neoplastic transformation of these lesions [18, 30, 31]. Overexpression of p53 has been reported not only in OSCC [18] but also in dysplastic lesions [19], suggesting that such an immunohistochemical finding may correspond to an early stage of oral cancerization of OLP lesions. Furthermore, according to Hadzi-Mihailovic M et al., the observed p53 in OLP samples represents mainly a mutated form of this protein [18]. Mutations or other inactivation mechanisms prevent the p53 system from working properly, allowing abnormal cells to proliferate, which results in cancer [18, 36, 37]. In this regard, Calenic B et al. demonstrated increased p53 levels with low expression of caspase-3 (the final executor enzyme of apoptotic pathways) and concluded that p53 apoptotic system was activated but did not execute [11]. These facts suggest that the p53 network as a barrier to tumor initiation may not function well in patients with OLP.

In the present study p53 was expressed in 45% of the OLP samples with staining intensity as follows: 15%(+++); 25%(++); 5%(+). These protein levels are significantly lower compared to p53 positivity rate reported by other authors − 69.9% Shailaja G [34]; 71.4% Leyva-Huerta E [32]; 80% Hadzi-Mihailovic M [18] and 100% Safadi R [19], but are in line with those of de Sousa F et al. (41.67%) [38] and Basheer S. et al. (40%) [31]. However, the applied Chi-square test in the present study showed no significant difference in either the percentage of p53 positive samples or the staining intensity between the OLP group and healthy controls (Table 1). The explanation is that we detected p53 immunoreactivity in 40% of cases with normal oral mucosa, of which 30% (n = 3) demonstrated strong intensity (+++). According to these results the expression of the pro-apoptotic p53 is not pathologically altered in OLP lesions.

As already mentioned above, p53 is not the only protein that mediates the apoptotic process. In this regard, reduced expression of the anti-apoptotic bcl-2 is an indicator for a high degree of apoptosis [30]. The detailed mechanism by which Bcl‐2 inhibits apoptosis is still uncertain, but it was suggested that this protein blocks mitochondrial cytochrome C translocation and simultaneously prevent caspase activation [30]. Therefore, the absence of this anti-apoptotic marker is associated with a loss of its anti-cell death function, which in turn promotes apoptosis. Data in the literature regarding bcl-2 expression in the epithelium of OLP lesions vary widely [13, 32, 39,40,41] - from significantly increased [30] to significantly decreased [6] compared to that in normal oral mucosa. In the present study, we found a marked (although not statistically significant) reduction in bcl-2 levels in the keratinocytes of OLP tissues compared with those in healthy epithelium (Table 1). This finding is consistent with the results of other studies [13, 29, 32, 40] and may predispose to the occurrence of apoptosis.

Activation of the programmed cell death is a protective mechanism to avoid proliferation of cells containing abnormal DNA. Therefore, if apoptosis occurs the proliferation index should be low. In support of this thesis, Pérez-Sayáns M et al. reported that the proliferative marker Ki-67 was positively correlated with the anti-apoptotic bcl-2 and negatively with the pro-apoptotic BAX [6]. However, a great number of studies have shown increased proliferation in OLP than in normal oral mucosa [13, 15, 16, 34, 42]. Ki-67 is crucial in the proliferation process. Its expression begins in the G1 phase, increases gradually in the subsequent–S and G2 phases, and reaches its maximum during cell division (M phase) [15]. According to the results of this study, Ki-67 levels did not differ significantly between the working (OLP contingent) and control groups (Table 1). Moreover, the expression intensity for the marker was slightly lower in OLP lesions.

General limitation of most of the prior studies, in our opinion, is that conclusions are made based on the immunohistochemical results total for the studied contingent, while the molecular effects depend on precise interplay between the different markers which is not the same among all patients. To determine the significance of the processes apoptosis and cell proliferation in the pathogenesis of the disease, we examined the relationship between p53, bcl-2 and Ki-67 for each of the twenty patients with OLP (Fig. 3). Considering all above a combination of elevated p53, absence of bcl-2 and low Ki-67 levels points in favor of an activated cell death program.

In the specimens of four of the OLP patients (n-7,13,14,16) a positive reaction was found only for Ki-67 with moderate (++) to strong (+++) intensity; in another two (n-15,17) this Ki-67 expression was also accompanied by staining for the anti-apoptotic bcl-2. In the absence of pro-apoptotic p53, we concluded for all of them that there was an intense process of epithelial renewal. Absence (-) or weak (+) expression of proliferative Ki-67, combined with absence of bcl-2 expression was found in cases n- 2, 4, 9, 11. However, in these cases there was also no p53 reaction, therefore it cannot be claimed that the reduced proliferative index is due to pathologically increased apoptosis. The pro-apoptotic p53 was expressed in nine of the twenty OLP tissue sections. However, in seven of the p53 positive cases there was an intense immunoreaction for Ki-67 (n-1, 3, 8, 10, 18, 19, 20) and additionally in three of them (n-10, 18, 20) bcl-2 was also expressed. This combination of proteins did not correspond to an enhanced apoptosis. Therefore, it can be speculated that for these cases p53 had a preventive function against the amplification of the genetic errors that have occurred, but did not lead to cell death. Only in two of the analyzed cases (n-5 and 6) the combination of markers pointed in favor of activated mechanisms of programmed cell death– there was mild to moderate expression of p53, accompanied by low Ki-67 levels and absence of bcl-2. Clinically, these patients demonstrated erosive and atrophic lesions.

From the above analysis, it can be concluded that for the presented sample of patients there was no evidence of an activated p53-dependent apoptosis pathway. This is not the first study to not associate apoptosis with p53 positivity. Hadzi-Mihailovic M et al. reported low percentage and weak intensity of p53-positive cells in OLP specimens with highly expressed Civatte bodies (CB) [18]. Since p53 is a pro-apoptotic protein, and CB are symbols of apoptosis, positive correlation between these parameters should be expected. Therefore, it could be assumed that apoptosis may occur, but other molecules may be implemented in its activation. In this regard, p63 is also an apoptotic marker [11]. In the present study we found significantly lower levels of p63 in the epithelium of OLP lesions compared to those in normal oral mucosa (Table 1). The established deficiency of this ectodermal factor in OLP seems to be an important molecular mechanism in the pathogenesis of the disease, as none of the other four markers analyzed demonstrated a significant expression change compared to the control group. Furthermore, data in the literature are relatively consistent regarding lower levels of the protein in this patients [10, 33]. A few sources indicate increased expression [11]. P63 is a transcription factor of P53 family, which determines lots of overlapping functions with p53. p63 is required for both proliferation and differentiation of the keratinocytes. It is also involved in cell cycle arrest, apoptosis and cell senescence and maintains epithelial integrity by regulating the expression of different adhesion markers. The mechanisms for triggering p63-dependent programmed cell death are not fully understood. It is considered that p63 is required for the epidermal apoptosis mediated by p53 [11]. Ebrahimi M et al. stated that a combination of low p63- and high p53 levels is needed to induce apoptosis [33]. The applied Spearmаn correlation test in the present study demonstrated a significant association between these two proteins in OLP lesions (р<0.02). However, p63 correlated positively with p53, as seven out of nine p53-positive cases highly expressed also p63(+++). Calenic B. et al. reported a p63/p53 correlation with strong immunoreactivity for both markers in patients with OLP, which was however accompanied by low levels of caspase 3 [11]. Therefore, with similarly high expression of p53 and p63, apoptosis does not occur. On the other hand, an in vitro study demonstrated that simultaneous blocking of both proteins did not result in cell death but, on the contrary, partially corrected the epithelial hypoplasia caused by p63 deficiency [43]. It could be concluded that in the implementation of the molecular mechanisms of the disease, p53 and p63 act in a coordinated manner, but do not lead to cell death.

There are also less known mechanisms of p63-mediated apoptosis. Anoikis is a form of programmed cell death that occurs in anchorage-dependent cells when they lose contacts to their neighboring cells or extracellular matrix [20]. Decreased expression of adhesion proteins such as E-cadherin and b-catenin has been reported in OLP [10]. E-cadherin and b-catenin are p63-dependent adhesion molecules [10]. The reduced expression of p63, found in the present study, indirectly points to a lack of adhesion proteins. Therefore, it could be speculated that the observed tissue destruction in OLP is precisely due to anoikis.

Cyclooxygenase 2 (COX-2) is also involved in the process of apoptosis. COX-2 is an enzyme responsible for the synthesis of bioactive prostanoids from arachidonic acid and is implicated in the processes of inflammation, reproduction, and carcinogenesis. Lichen planus is a chronic inflammatory disease and predictably, COX-2 levels have been reported to be higher in these patients [41, 44]. Overexpression of COX-2 may represent a marker for inhibition of apoptosis [9]. Among the mechanisms by which this inflammatory marker blocks programmed cell death are: reduction of cytochrome C and caspase 3 activity and induction of increased expression of anti-apoptotic bcl-2 [45]. In the conducted study we found increased levels of COX-2, which however, did not differ significantly from those in normal oral mucosa (Table 1). Furthermore, no significant correlation was demonstrated between this inflammatory marker and pro-apoptotic p53 (p = 0.44). Additionally, in the present study bcl-2 levels were low, so it cannot be argued that COX-2 enhanced its expression. Therefore, there is no evidence to support COX-2 promoted cell survival.

In summary, for the presented OLP cohort, the immunohistochemical analysis, including five apoptotic markers, did not confirm the significance of programmed cell death in the pathogenesis of the disease. In making this conclusion, we still keep in mind that the expression pattern of pro- and anti-apoptotic proteins conditionally indicates whether apoptosis has occurred or not. In order to correctly determine the extent of apoptosis, additional diagnostic techniques are required, such as examination of caspase 3 expression; DNA fragmentation (TUNEL method); chromatin condensation (ssDNA); number and distribution of apoptotic bodies (Civatte bodies). Bascones C. et al. applied a TUNEL assay and Caspase-3 expression examination in their study and, like us, concluded that the apoptotic phenomenon was of little quantitative importance in OLP [46]. Instead, they reported basal expression of p21. The latter is a cyclin dependent kinase inhibitor, that arrests the cell cycle in G1, inducing DNA repair or senescence. Low rate of apoptosis with increased expression of p21 [30] have been demonstrated also in other studies [13]. p21 is known to be negatively regulated by p63 [43]. In this regard, the p63 deficiency found in the present study may indirectly lead to cell cycle arrest through the resulting intense expression of p21. Therefore, we share the position of Bascones C. et al. that epithelial cells under attack in OLP respond more frequently with arrest or senescence than with apoptosis [46].

The results of our study did not support the notion of pathologically increased proliferation in OLP, as Ki-67 expression in these lesions was even lower (although not statistically significant) compared to that in normal oral mucosa. In vitro studies have shown that lack of p63 resulted in severe epithelial hypoplasia [43]. We also found a significant positive correlation between p63 and Ki-67 expression. Therefore, the established reduced expression of p63 may result in suppression of cell proliferation, which predisposes to epithelial thinning.

Keratotic forms of OLP are characterized by thickening of the epithelium. Thus, increased expression of proliferative markers and decreased expression of apoptotic proteins can be expected. The opposite expression pattern is more likely to be detected in the more severe - atrophic-erosive forms of the disease. However, statistical analysis showed no significant difference in p53, p63, bcl-2, Ki-67, and COX-2 levels between these two groups, thus none of them can be used as prognostic marker for disease severity.

As stated above, increased expression levels of p53, p63, bcl-2, COX-2 and Ki-67 are considered indicative of dysplastic transformation of the tissue. Since the expression of p53, bcl-2, COX-2, and Ki-67 in the studied OLP lesions was not significantly different from that in normal oral mucosa and the levels of the epidermal factor p63 were found to be even lower in these patients we can conclude that for the study cohort presented, there was no evidence of carcinogenesis.

The main drawbacks of the studies focusing on immunohistochemistry in OLP (including the present one) are: (1) There are no generally accepted values ​​for the expression intensity of a given marker in normal oral mucosa. This leads to large discrepancies in the interpretation of the results in OLP: same values ​​are interpreted as overexpression by some authors and as downexpression by others; (2) Ideally, control biopsies and those taken from patients should be from the same anatomical area, whereas in most studies the control samples are from gingiva adjacent to the teeth planned for extraction, which is firstly a keratinized mucosa and secondly may show signs of low-grade inflammation; (3) The antibodies used to detect the markers cannot distinguish between the wild-type and mutated form of p53, nor between the TAp63 and ΔNp63 isoforms of p63, which have different, often opposite, effects. Further studies addressing the aforementioned issues are needed to improve the accuracy of immunohistochemical analysis as a diagnostic tool.