Abstract
Feline inductive odontogenic tumour (FIOT) is a rare and interesting odontogenic neoplasm in which the odontogenic epithelium has inductive potential to form aggregated foci of dental pulp-like mesenchymal cells. Two male cats aged 11 and 10 months presented with nasal swelling and a left maxillary mass. Histopathologically, the masses consisted of non-encapsulated invasive neoplasms exhibiting proliferation of epithelial and mesenchymal components with local infiltration into the maxillary bone in both cases. The epithelial component formed islands, anastomosing strands, and solid sheets of polygonal epithelial cells. Occasionally, these cells formed circular aggregates, resembling the cap stage of odontogenesis. Type IV collagen and laminin were constantly positive around the foci of epithelial cells, and Ki-67 positive indices were extremely low; therefore, these findings consistent with the benign clinical presentation of FIOT.
Odontogenic tumours are uncommon in domestic animals (Head et al 2002). These tumours are classified based on their degree of differentiation; namely whether neoplastic ameloblasts have the inductive potential for mesenchymal cells into odontogenic components. Feline inductive odontogenic tumour (FIOT) is an interesting odontogenic neoplasm in which the odontogenic epithelium has inductive potential to form aggregated foci of dental pulp-like mesenchymal cells (Dubielzig et al 1979, Gardner and Dubielzig 1995, Beatty et al 2000). Because this neoplasm is rare, its immunohistochemical features, particularly the expression of proliferative markers, which may be associated with prognosis, are unclear. We have reported the morphological and immunohistochemical features of two FIOT cases in juvenile cats.
A 11-month-old male Japanese domestic shorthair (case 1) was admitted to a private animal hospital with swelling of the left maxilla caused by a protruding mass. Laboratory examination did not reveal any abnormality. Radiographically, the tumour was diagnosed as a cystic lesion appearing as a unilocular radiolucent area involving the maxillary left canine. The left deciduous canine remained in the oral cavity; the left permanent canine had been pushed deeper into the maxilla and did not erupt. The mass was partially excised for diagnosis. Unfortunately, we could not follow-up this case.
A 10-month-old male domestic shorthair (case 2) presented to the Gifu University Veterinary Teaching Hospital with a left maxillary mass. Complete blood count and blood chemistry did not reveal any abnormality. On computed tomography, the left maxillary bone was lysed by expansion of neoplasm at an area between the incisor and premolar, and the left permanent canine was embedded in the mass (Fig 1). The mass was completely excised, and no metastasis and recurrence was observed.
Left: Computed tomographic image of the area between the incisor and premolar of the cat in case 2. The left maxilla was lysed by proliferation of tumour tissue. Right: The epithelial neoplastic cells formed circular aggregates around clusters of highly cellular mesenchyme, resembling the cap stage of odontogenesis. Case 2, Haematoxylin and eosin staining, Bar = 100 μm.
Histopathologically, in both cases, the masses consisted of non-encapsulated tumours exhibiting proliferation of epithelial and mesenchymal components. The epithelial component formed islands, anastomosing strands, and solid sheets of polygonal epithelial cells. Occasionally, these cells formed circular aggregates around clusters of highly cellular mesenchyme; these aggregates were round, discrete, and poor in collagen, resembling the cap stage of odontogenesis (Fig 1). Some islands exhibited areas of stellate reticulum. In case 1, foci of mineralization were associated with the epithelium. In case 2, a thin rim of eosinophilic material surrounding mesenchymal foci was recognised as the initial dentin deposition. The eosinophilic material was considered as the neoplastic component because it was detected in a few similar structures within the tumour. Atypia was not observed in the neoplastic epithelium. Mitotic figures were rarely observed in the neoplastic epithelium and mesenchymal cells, particularly in the aggregated clusters. In both cases, the tumours locally infiltrated the bone; however, no evidence of metastasis such as lymphatic invasions was found. No amyloid production was observed with the direct fast scarlet stain.
For immunohistochemistry, the anti-Ki-67 antigen murine monoclonal antibody (clone MIB-1, Dakocytomation, Carpinteria, CA), anti-collagen type IV murine monoclonal antibody (Dako), and anti-laminin rabbit polyclonal antibody (Lab Vision, Fremont, CA) were used for primary antibodies. Before incubation of the primary antibody, boiling (121°C, 15 min) for Ki-67 antigen, enzyme digestion (0.05% trypsin in phosphate buffered saline for 15 min at room temperature) for collagen type IV and laminin were employed as antigen retrieval treatments. Subsequently, the section was incubated with peroxidase-labelled secondary antibody (Envision+; Dako), and visualised by diaminobenzidine. Counterstain was used with Mayer's haematoxylin. The primary antibody was omitted and replaced with phosphate buffered saline for negative control. We confirmed positive reaction in the basement membrane of blood vessels and oral mucosal layer for collagen type IV and laminin, and in basal cells of stratified squamous epithelial layer for Ki-67 antigen as the internal positive controls.
Immunohistochemistry for the Ki-67 antigen revealed that the number of positive basal cells was extremely high in the hyperplastic oral squamous epithelium; however, the neoplastic epithelial and mesenchymal cells were rarely positive (1.57% and 1.63% in the epithelial components and 0.81% and 1.49% in the mesenchymal components in cases 1 and 2, respectively) (Fig 2). A periodic acid-Schiff (PAS), collagen type IV, and laminin positive basement membrane was observed around the epithelial proliferation (Fig 2).
Left: Immunohistochemistry of the Ki-67 antigen in case 2. Hyperplastic oral squamous epithelium contained many positive cells; however, the neoplastic epithelial and mesenchymal cells were rarely positive (1.63% in neoplastic epithelial components and 1.49% in mesenchymal foci). Bar = 100 μm. Right: Immunohistochemistry of type IV collagen (upper) and laminin (under) in case 1. Both type IV collagen and laminin were constantly positive around the epithelial proliferation. Bar = 250 μm.
The first case of FIOT was reported by Dubielzig et al (1979), and several cases have been reported thereafter. However, the nomenclature of this tumour is controversial; in fact, feline odontogenic tumours with histological features similar to those of FIOT have been designated as different tumours, eg, ameloblastic fibroma. In humans and cattle, ameloblastic fibroma presents as a well-circumscribed, encapsulated neoplasm that can be surgically excised, and its mesenchymal component is not condensed into spherical nodules (Gardner and Dubielzig 1995, Nyska and Dayan 1995).
Complete resection of FIOT is required because of its locally infiltrative nature; however, there have been no reports of metastasis. Complete excision was effective in preventing recurrence in case 2; however, this could not be ascertained for case 1 as this case was lost to follow-up. No histological evidence of metastasis was observed in either of the cases. The ameloblastic tumour islands were circumferentially delineated by linear staining of PAS, anti-collagen type IV or anti-laminin antibodies corresponding to the basement membrane. Immunohistochemical study of five ameloblastomas and their malignant counterparts in humans indicated that collagen type IV and laminin expressions in malignant ameloblastomas appeared as discontinuous lines around the original and metastatic tumour clusters as opposed to the continuous circumferential expressions observed in benign ameloblastomas (Sauk 1985). Therefore, the presence of a basement membrane around neoplastic foci may serve as a barrier against vascular invasion by neoplastic cells. Additionally, proliferative activity expressed as the Ki-67 labelling index (LI) was low in the epithelial and mesenchymal components. The Ki-67 LIs in human ameloblastomas were 0–6.8% (Sandra et al 2001, Jaaskelainen et al 2002); in contrast, those of p53 immunopositive odontogenic carcinomas were higher (more than 100 positive cells in a high-power field) (Slootweg 1995). The Ki-67 LI of the sarcomatous component of human ameloblastic fibrosarcomas was 13.5%; however, that of ameloblastic fibromas with no recurrence was 2.1% (range: 1.5–2.9%) (Sano et al 1998). Therefore, the Ki-67 LI of both the epithelial and mesenchymal components were essentially the same as for their human counterparts which further supports a benign nature for this tumour. These findings are consistent with the benign clinical presentation of FIOT.