NCAM (CD56) Expression in keratin-producing odontogenic cysts: aberrant expression in KCOT

Keratin-producing odontogenic cysts (KPOCs) form a heterogeneous group of cystic lesions that are often aggressive in character, with high rates of recurrence and multifocality [1]. The lesional spectrum of KPOCs includes odontogenic keratocysts, that according to last World Health Organization (WHO) guidelines [2] are also referred to as keratocystic odontogenic tumors (KCOTs), in accordance with the fact that KCOTs are true tumoral growths. Effective management of these cystic lesions is subject to frequent discussion [3] and malignant transformation is possible, albeit very rare [4].

NCAM (neural cell adhesion molecule), also known as CD56, was originally characterized as a cell-surface glycoprotein member of the immunoglobulin superfamily, implicated in calcium-independent intercellular adhesion [5] and expressed in a wide variety of cells [6]. However, in the past decade the traditional view of NCAM has been challenged, such that it is now also considered a signaling receptor that impacts upon cellular migration, proliferation, differentiation, and survival, thereby playing a role in different models of tumor growth [7].

NCAM immunoreactivity in ameloblastomas within odontogenic tumors has been reported with very high frequency [8-11], and has also been observed in certain ameloblastic carcinomas [9]. In contrast, NCAM reactivity is indicated as relatively infrequent in KCOTs [10,11].

The aim of the present study was to assess the immunohistochemical expression of NCAM in a series of keratin-producing odontogenic cysts (KPOCs), and to evaluate the possible significance of adhesion molecule reactivity in this heterogeneous group of odontogenic cysts.

Of 58 KPCOs, only 16 cases (27,58%) dysplayed NCAM expresión at the of epithelial level and inmunoreactivity was in positive cases of focal character (<10% of cells expressing NCAM) with a weak (1+) to moderate (2+) intensity, and always demostrating a heterogeneous distribution within the structure of the cystic lesion.

NCAM expression was absent in all 12 OOC cases, with the epithelial lining of the cysts entirely negative for CD56. In contrast, focal NCAM reactivity was observed at the epithelial level in 16 of 46 KCOT cases (36.95%). There was a significant difference in NCAM expression in cases of OOC vs. KCOT (p = 0.012).

NCAM immunostaining was differentially expressed in NS- and S-KCOTs: four of the six cases of S-KCOT (66.66%) exhibited focal NCAM reactivity at the epithelial level, in contrast to only 12 of the 40 (30%) cases of NS-KCOT (Table 1). OOC did not recur in any case during clinical/radiological follow-up at 39.2 ± 26.01 m. In contrast, all cases of S-KCOT recurred at a mean follow-up of 112 ± 76.1 m. In NS-KCOT0 cases, the recurrence rate was 35% at a mean follow-up of 29.6 ± 31.04 m. When analyses were performed independent of lesional subtype, there was no association between recurrence and NCAM expression (p = 0.52).

In positive cases, NCAM expression was always epithelial, with weak (1+) to moderate (2+) intensity and a heterogeneous and focal distribution (Figures  1 A and 1 B). NCAM positivity (Table 2) was most frequent in the basal layer of cystic epithelium, followed by areas of basal budding (Figures  2 A and 3 A-D) and the basal portion of daughter cysts or satellite epithelial nests (Figures  2 B and 4 A-D). In NCAM-positive S-KCOT cases (four of six), NCAM positivity was discontinuous, weak or moderate, in the locations indicated above, although only 50% of satellite epithelial nests or cysts demonstrated basal positivity. Similarly, in NS-KCOTs, the most frequent NCAM-positive location was the basal cell layer of the cystic lining (observed in all cases exhibiting NCAM positivity), followed by areas of basal budding (n = 4) and basal portions of daughter cysts (n = 3).

Neural cell adhesion molecule (NCAM), a member of the immunoglobulin superfamily adhesion molecule, is expressed by a wide variety of neuroectodermal and mesenchymal cells [6]. Originally, NCAM was exclusively characterized as a mediator of cell-cell adhesion, but is now also considered to be a signaling receptor that impacts upon cellular adhesion, migration, proliferation, apoptosis, differentiation, and survival. NCAM is involved in various models of tumorigenesis [7], including certain types of odontogenic tumor: NCAM expression has been reported in certain ameloblastic carcinomas [9], and especially in the outer columnar cells of ameloblastoma (AB) [8-11]. This expression could be indicative of neuroectodermal differentiation in AB [10], especially considering the fact that neural crest cells are associated with tooth development, particularly ectomesenchymal differentiation in tooth germs [15].

The purpose of the present study was to assess the presence and significance of NCAM reactivity in a series of KPOCs. Our results highlight a total absence of NCAM expression in OOCs, a form of KPOCs characterized by orthokeratotic keratinization, reduced proliferative activity and non-recurrence [16,17]. Recurrence was not observed for any case of OOC, but in KCOTs, a form of KPOC now considered as a true odontogenic tumor [1,2,4], we observed focal and heterogeneous NCAM expression in 36.95% of cases, a significantly higher rate than OOC. Prior to our investigation, only two studies assessed NCAM expression in odontogenic cysts. Cairns et al. [11] reported NCAM expression in only 5.26% of KCOTs, with focal NCAM reactivity occurring at the basal cell level; in contrast, AB exhibited high levels of NCAM reactivity. More recently, Kusafuka et al. reported that 50% of KCOTs demonstrated NCAM reactivity, also with exclusive expression at the basal cell level [10]. Our series therefore represented an intermediate rate of positive observations.

In our KCOT cases, NCAM expression was exclusively focal and heterogeneous, with weak-to-moderate reactivity. Despite this limited, focal reactivity, our results demonstrate NCAM expression in locations not described previously: earlier studies noted NCAM reactivity only in the basal cell layer of KCOTs [10,11]. Furthermore, in one of these studies, which analyzed three cases of S-KCOT, there was no relationship between NCAM expression and this syndromic form of KCOT [11]. In contrast, we not only observed discontinuous positivity in the basal portion of cystic epithelium but also in areas of basal budding and in the basal portion of daughter cysts, both of which are occasionally detectable in KCOTs, and particularly in S-KCOTs.

Epithelial budding, arising from the basal layer of cystic lining, is a peculiar morphological aspect of KCOT that has been suggested as the source of daughter cysts; [18-20] influences lesional recurrence; [13,14,21] and is most frequently observed in S-KCOTs [22]. Aberrant NCAM expression in epithelial buds and the basal portion of daughter cysts of KCOTs was observed in 50% of S-KCOTs and represents a possible influence on NCAM expression in instances of lesion recurrence.

In our series, all S-KCOT cases recurred; in NS-KCOTs the recurrence rate was 35%. NCAM expression was detected in 66.66% and 32.50% of S- and NS-KCOTs, respectively. Systematic review of cases of NS-KCOT indicated an average recurrence rate of 28% [23], although rates were heavily influenced by the duration of follow-up, and especially by the type of treatment administered [3], indicating that adequate management of NS-KCOTs may prevent recurrence [2,24].

When we analyzed the relationship between NCAM expression and recurrence independent of KPOC subtype, there was no significant association. However, there was an important limitation: observations were not treated homogeneously due to the retrospective nature of the series. Nevertheless, the fact that NCAM expression was observed in morphological locations involved in recurrence [13,14,21] suggests a possible relationship between NCAM expression and lesional recurrence. Further studies pertaining to NCAM expression, in larger series of KCOTs treated with homogenous procedures or subjected to prospective randomized analysis [25], are required to obtain definitive conclusions. Likewise should be verify a possible association between increased expression of CD56 and others markers [26,27] and genes [28] overexpressed in KCOT.

A final interesting observation concerned NCAM expression in AB, another frequently recurring odontogenic tumor in which NCAM reactivity, in contrast to KCOTs, is virtually certain [11]. AB and KCOTs have been considered two morphologically distinct odontogenic tumors [29]. However, certain morphological and evolutionary data indicate a degree of overlap between these tumors; [30] this similarity is clearly reflected in so-called keratoameloblastoma [31,32], as well as in the solid variant of KCOT [33], characterized by hybrid AB and KCOT morphological features. Accordingly, aberrant NCAM expression in KCOTs could be indicative of foci of ameloblastic differentiation. As suggested previously [27,34], collaborative studies (given the scarcity of observations) concerning NCAM in keratoameloblastomas are required to confirm if this is the case.

The present study investigated the immunomorphological pattern of NCAM expression in KPOCs, and highlighted that NCAM expression in OOCs is entirely absent. In relation to KCOT our immunohistochemical analysis indicated NCAM reactivity in locations not reported previously (e.g., areas of basal budding and basal cells of daughter cysts), especially with respect to S-KCOTs. This suggests a possible role for this signaling molecule in lesional recurrence in KCOTs. Futures studies using homogeneously treated series of KCOT cases are required to confirm the influence of NCAM on lesion recurrence, as well as its potential utility as a marker of ameloblastic differentiation.