Enzymatic disaggregation protocol for simultaneous isolation of primary keratinocytes and fibroblasts from cholesteatoma tissue has been developed. The biopsy samples were harvested from patients undergoing surgical treatment in the Department of Otolaryngology, Ajou University Hospital, Korea. An informed consent form approved by the Institutional Review Board of Ajou University Hospital was signed by each patient. Cholesteatoma keratinocytes were cultured in Defined Keratinocytes Serum Free Medium (GIBCO, BRL, Gaithersburg, MD, USA) supplied with growth supplement and 5 mg/mL Gentamicin (GIBCO, BRL) according to the manufacture instructions. Cholesteatoma primary fibroblasts were successfully cultured in Dulbeco Modified Eagle Medium (GIBCO, BRL) supplemented with 10% Fetal Bovine Serum (FBS, Clonetics, Walkersville, MD, USA) and 1% Penicillin-Streptomycin mixture (GIBCO, BRL). Primary keratinocytes (passage 1) and fibroblasts (between passages 1-5) were used for establishment of the co-cultured system. HaCaT cell line was kindly provided from the Department of Dermatology in Ajou University Hospital. Feeder layer from cholesteatoma fibroblasts was developed using pretreatment with Mitomycin C (Sigma, St. Louis, MO, USA) in concentration 0.25 g/mL. Fibroblasts were incubated overnight at 37℃, 5% CO₂ in humidified atmosphere. On the next day the media was exchanged and the cells were incubated additionally for another 24 hr. The feeder layer was subsequently trypsinized from the substrate and transferred on polyester membrane in Transwell (Costar, Cambridge, MA, USA) system to develop a subconfluent feeder layer. Feeder layer was incubated for 24-48 hr prior co-culturing. Cholesteatoma keratinocytes and HaCaT cells were seeded at density 2×10⁵ cells/mL over the feeder layer, co-cultured for 3-5 days and air-exposed for a period of 14 days. Two alternative models were elaborated using primary cholesteatoma keratinocytes and HaCaT cells (Fig. 1). The control group was prepared without feeder layer.

Transwell membranes were fixed in 4% formaldehyde overnight at 4℃, dehydrated in graded alcohols, cleared in HistoClear (National diagnostics, Atlanta, Georgia) and embedded in paraffin. Six µm-thick sections were prepared and subjected to routine hematoxylin-eosin staining.

Polyacrylamide gel electrophoresis has been used for protein separation and staining with Coomassie blue was performed to compare the protein profile between cholesteatoma keratinocytes and HaCaT cells.

After separation, the proteins were transferred to a nitrocellulose membrane. Blocking of nonspecific staining was accomplished with membrane incubation with 5% non-fat skim milk for 1 hr at room temperature (RT). Membranes were then incubated with primary mouse monoclonal anti-cytokeratin antibody clone K 8.12 (Sigma) which is immunospecific for cytokeratin 13 (CK 13) and cytokeratin 16 (CK 16), with dilution factor 1:100, for 1 hr at RT. After washing membranes were incubated with goat anti-mouse horseradish peroxidase conjugate (Zymed, San Francisco, CA, USA). The positive reaction was visualized with ECL Western Blotting Detection System (Amersham, Bucks, UK).

The localization of cytokeratin 13 and 16 in middle ear cholesteatoma tissue was accomplished by IHC study as described before (9). Five middle ear cholesteatoma samples were compared with the corresponding retroauricular skin samples obtained during surgery. Six µm thick formalin-fixed paraffin-embedded tissue sections were processed with ABC-method. Samples were deparaffinized in xylene and rehydrated in alcohol. Heat-induced antigen retrieval procedure was applied for antigen unmasking. Primary mouse monoclonal anticytokeratin antibody (Sigma) at concentration 1:20 was incubated overnight at RT. Vectorlab Elite ABC kit (Vector Laboratories, Burlingame, CA, USA ) was used for detection of positive reaction using diaminobenzidine substrate.

Enzymatic disaggregation protocol yield a high number of separated cells. The low selectivity and potential damages of cells in case of prolonged incubation were major disadvantages of the procedure. As an inflammatory tissue, cholesteatoma posed some difficulties in isolation of primary keratinocytes, because of the attendant inflammation and possible contamination, which requires an adequate decontamination. The first primary keratinocytes appeared usually 7-10 days after trypsinization and generally within the next 10 days keratinocytes reached a confluence. The individual variations in proliferative activity of cholesteatoma keratinocytes were related to cholesteatoma localization and sex of the patient. For instance, attic cholesteatoma frequently resulted in isolation of primary keratinocytes with bigger proliferative ability. Another interesting correlation was the fact that cholesteatoma samples obtained from young female patients were characterized by higher proliferative activity in vitro. The above-mentioned dependencies were not statistically significant. Primary cholesteatoma keratinocytes were uniform in shape and arranged in a typical monolayer pattern (Fig. 2A). In passage 2 an abrupt deceleration in proliferation was established, accompanied by morphological alterations such as enlargement and flattening of the cells, as well as vacuolization of the cytoplasm. In passage 3 cells were practically incapable of proliferation. Primary cholesteatoma fibroblasts were steadily maintained without significant changes of doubling time until passage 10. Primary cholesteatoma fibroblasts displayed typical spindle-shaped morphology upon confluence (Fig. 2B). No spontaneous transformation was detected in any of the cultures.

Bigger stratification and proliferation have been observed in the proposed 3D in vitro system when primary cholesteatoma keratinocytes were co-cultured in the presence of fibroblasts (Fig. 3B) in comparison with the control group without fibroblasts (Fig. 3A). Similar morphological landmarks were established in the control group of HaCaT cells (Fig. 4A). Disrupted differentiation in the suprabasal layers was revealed when cholesteatoma keratinocytes were air-exposed in the presence of fibroblasts mimicking the parakeratosis state, which is typical for this pathological tissue in vivo (Fig. 4B).

Coomassie blue staining of the protein extracts from primary cholesteatoma keratinocytes and HaCaT cells revealed significantly different protein profiles (Fig. 5A). Stronger expression of CK 16 (48 kDa) and CK 13 (51 kDa) was established in cholesteatoma keratinocytes in comparison with HaCaT cells (Fig. 5B).

Using mouse monoclonal antibody clone K 8.12, tissue localization of CK 13 and CK 16 was verified both in the basal and suprabasal layers of cholesteatoma (Fig. 6). The described staining patterns were established in all 5 cholesteatoma samples. The intensity of the CK 13, 16-positive staining varied in different areas of the same cholesteatoma sample. Specific positive reaction was not detected in retroauricular skin samples used as a control. The positive control revealed specific expression of CK 13 and 16 confined to the islands of neoplastic squamous cells in laryngeal carcinoma specimens.