Differentiation of nuclei during keratinization in middle ear cholesteatoma. DNA cytophotometry completed by computerized image analysis

Nuclear staining and relative distance for quantifying epidermal differentiation in biomarker expression profiling

BACKGROUND

The epidermal physiology results from a complex regulated homeostasis of keratinocyte proliferation, differentiation and death and is tightly regulated by a specific protein expression during cellular maturation. Cellular in silico models are considered a promising and inevitable tool for the understanding of this complex system. Hence, we need to incorporate the information of the differentiation dependent protein expression in cell based systems biological models of tissue homeostasis. Such methods require measuring tissue differentiation quantitatively while correlating it with biomarker expression intensities.

RESULTS

Differentiation of a keratinocyte is characterized by its continuously changing morphology concomitant with its movement from the basal layer to the surface, leading to a decreased average nuclei density throughout the tissue. Based thereon, we designed and evaluated three different mathematical measures (nuclei based, distance based, and joint approach) for quantifying differentiation in epidermal keratinocytes. We integrated them with an immunofluorescent staining and image analysis method for tissue sections, automatically quantifying epidermal differentiation and measuring the corresponding expression of biomarkers. When studying five well-known differentiation related biomarkers in an epidermal neck sample only the resulting biomarker profiles incorporating the relative distance information of cells to the tissue borders (distance based and joint approach) provided a high-resolution view on the whole process of keratinocyte differentiation. By contrast, the inverse nuclei density approach led to an increased resolution at early but heavily decreased resolution at late differentiation. This effect results from the heavy non-linear decay of DAPI intensity per area, probably caused by cytoplasmic growth and chromatin decondensation. In the joint approach this effect could be compensated again by incorporating distance information.

CONCLUSION

We suppose that key mechanisms regulating tissue homeostasis probably depend more on distance information rather than on nuclei reorganization. Concluding, the distance approach appears well suited for comprehensively observing keratinocyte differentiation.

Retinoic acid depletion induces keratinizing squamous differentiation in human middle ear epithelial cell cultures

OBJECTIVE

The pathogenesis of cholesteatoma behind an intact tympanic membrane remains controversial. Squamous metaplasia of the middle ear mucosa is thought to be a possible mechanism in such cases. However, to date, no definitive experimental results have proved this association. This study was undertaken to investigate whether normal human middle ear epithelial (NHMEE) cells undergo keratinizing squamous differentiation in a retinoic acid (RA)-deficient culture.

MATERIAL AND METHODS

We examined the morphological differences between RA-deficient and -sufficient cultures, and determined the expressions of the mucin gene and cornifin-alpha mRNAs as indicators of mucous and squamous differentiation, respectively.

RESULTS

Histomorphologically, the NHMEE cells differentiated into a keratinizing squamous epithelium in RA-deficient cultures. In addition, the expressions of mucin gene 5AC (MUCSAC) and MUC8 mRNAs were suppressed, and the expression of cornifin-alpha mRNA increased progressively as a function of differentiation in RA-deficient cultures.

CONCLUSIONS

This study shows that RA depletion induces keratinizing squamous differentiation in NHMEE cell cultures. This finding supports the hypothesis that middle ear cholesteatoma originates from metaplastic middle ear mucosa.

Immunohistochemical study of cell proliferation using BrdU labelling on tympanic membrane, external auditory canal and induced cholesteatoma in Mongolian gerbils

The Mongolian gerbil is a well-known animal model for induction of aural cholesteatomas. This animal model is useful for studying changes in the keratinizing epithelium. It is not known whether keratin accumulation can increase the proliferative activity of the keratinizing epithelium in tympanic membrane and meatal skin. In this study, we investigated the proliferative activity of the epidermis in induced aural cholesteatoma at various stages and in different areas of the tympanic membrane and meatal skin in normal gerbils. Anti-5-bromo-2- deoxyuridine (BrdU) was injected intraperitoneally to detect the proliferative activity of keratinizing epithelium. Immunohistochemistry with monoclonal BrdU antibody in the normal gerbil showed intense immunolabelled keratinocytes at the handle of malleus, and the superior parts of pars tensa and pars flaccida. Also, mitotic activity in the deep meatal skin was more active than in the lateral meatal skin. The induced aural cholesteatoma showed more active proliferation centre of the epithelial cell than eardrum and external ear canal of the normal gerbil. These observations suggest that the accumulation of the keratin debris might induce changes of the cellular proliferation in the external auditory meatus.

The apoptotic and nonapoptotic nature of the terminal differentiation of erythroid cells

The morphology of erythroid cells changes dramatically during the course of their terminal differentiation. According to calculations made with cytospin preparations obtained from Syrian hamster yolk-sac-derived erythroid cells, the area of nuclei at day 10 of gestation ranges from 25 to 85 micron 2 and is reduced to 15-25 micron 2 on day 13 [K. Morioka and R. Minamikawa-Tachino, Dev. Growth Differ. 35, 569-582, 1993]. The DNA and protein contents of each nucleus also decrease during this period. Nonspecific fragmentation of DNA was detected by agarose gel electrophoresis in all samples obtained from day 10 to day 13 of gestation, while distinct ladders of DNA fragments were not detected. DNA fragmentation was also detected by an in situ DNA-end labeling (TUNEL) assay. As the terminal differentiation proceeded, gradual decreases in levels of both histone H1 and most nonhistone proteins were observed by SDS-polyacrylamide gel electrophoresis, while levels of core histones appeared to be constant. In particular, lamin B2 was almost completely lost from the nuclear matrix fraction on day 11. These results suggest that the terminal differentiation of erythroid cells and apoptosis might have common mechanisms. However, expansion of the cytoplasm during the terminal differentiation distinguishes these processes. In addition, in the erythroid terminal differentiation, nuclei never form lobules or become fragmented; no apoptotic bodies are formed, occurrence of the apoptosis-like cellular change is not sporadic but rather synchronous, and the process is slow, with at least several days being required for cell death. These characteristics are different from those of typical apoptosis. Thus, the terminal differentiation of nucleated embryonic erythroid cells exhibits both apoptotic and nonapoptotic features.

Antigen expression of epithelial markers, collagen IV and Ki67 in middle ear cholesteatoma. An immunohistochemical Study

The expression of epithelial markers (cytokeratins, Filaggrin, BerEp4 and EMA), collagen IV and Ki67 was studied immunohistochemically in cholesteatoma and compared with that in epidermis of meatal skin, squamous epithelium of eardrum and simple epithelium of middle ear mucosa. MNF116 (cytokeratin 10, 17, 18) stained the full layer of normal epithelium and all cholesteatoma specimens. CK10 and Filaggrin were expressed in the upper layer of epidermis but more diffusely in cholesteatoma. BerEp4 was found in the basal layer of normal epithelium but was detected in most epithelial cells in cholesteatoma matrix. Variability was observed in EMA and CK14 immunostaining. Collagen IV was localized in the basement membrane of normal epithelium with a continuously staining pattern, an observation also made in the cholesteatomas studied. However, in one of these small areas the basement membrane was not stained with collagen IV. Ki67 was expressed in nuclei of the cells in the basal layer of normal epithelium but extended to epithelial cells in the upper layers of cholesteatoma matrix. The results of the present study indicate that the expression pattern of epithelial markers in cholesteatoma corresponds to that in normal epidermis. The increasing expression of BerEp4 and Ki67 confirms the hyperproliferative nature of cholesteatoma. Whether or not the lack of expression of collagen IV in one of the cholesteatomas reflects a true degradation of the basement membrane needs further investigation in extended materials.

Immunohistochemical analysis of the cytokeratin expression in middle ear cholesteatoma and related epithelial tissues

Immunohistochemical investigations were carried out to determine the pattern of cytokeratin (CK) expression in middle ear cholesteatoma and related epithelia. Using monoclonal antibodies specific for CK chains and the indirect immunoperoxidase technique, we examined 10 CK polypeptides for expression. The external stratified squamous epithelium of the tympanic membrane generally expressed CKs 5, 10, and 14. In addition, basal keratinocytes in the annular region of the pars tensa expressed CK 19 (a simple epithelium marker), while suprabasally the hyperproliferative marker CK 16 was expressed. These data reflect the unusual proliferative nature of this region. The unexpected appearance of CK 16 (known to have a limited distribution in healthy epidermis) clearly relates to its expression in the neighboring deep meatus. The medial simple epithelium of the eardrum revealed mucosal CKs 7, 8, 14, 18, and 19. Acquired cholesteatoma lesions, besides CKs 5, 10, and 14, consistently expressed CK 16 in suprabasal layers. These results constitute the first direct molecular evidence for the hyperproliferative nature of the cholesteatoma matrix. Overall, our CK data suggest that aural cholesteatoma lesions and epidermal tissue in this area are related. However, they do not explain the mechanism(s) by which the eardrum or meatal epithelia might invade the middle ear cavity. Congenital cholesteatomas expressed CKs 5, 10, 14, and 16 equally. These CK data do not support the idea of a metaplastic origin from middle ear mucosa; instead, they suggest activation of an ectodermal rest in the middle ear cavity.

The nature of the epithelium in acquired cholesteatoma. Part 2. Cell culture

The exact nature and role of the epithelial layer in cholesteatoma remains undetermined. The aim of this study was to investigate cholesteatoma epithelium and normal aural epithelia in common cell culture conditions. Samples of cholesteatoma, external meatal epidermis and middle ear mucosa were obtained, successfully grown in cell culture, and subcultured. No significant morphological differences were found between cholesteatoma and aural epidermis. The only differences noted were delayed onset of colony formation, and the need to subculture prior to the cultures becoming confluent in the cholesteatoma cultures. Further research is required to account for these differences in growth patterns.

Keratinization of middle ear cholesteatomas. I. A histochemical study of epidermal transglutaminase

A histochemical study was performed to determine the involvement of epidermal transglutaminase (ETgase) in the keratinization of middle ear cholesteatomatous lesions, and to compare it with its role in the middle ear mucosa and epidermis. In a first assay, we localized the (E)Tgase activity in situ. A second immunohistochemical assay revealed the distribution of the particulate form of ETgase, which is involved in cross-linked envelope formation. A remarkable difference between strongly keratinized epidermal tissues and the cholesteatoma matrix is the frequent observation in the latter of the remnants of (E)Tgase activity in cytosol, even in advanced stages of differentiation. As a consequence, the cell-membrane-associated ETgase activity, and thus the extent of cross-linking within the envelope, is at a lower level than expected. This aspect is reminiscent of the keratinization phenomenon manifested by thin epidermal tissues. In addition, our findings are the first to show that ETgase is a substantial marker of middle ear mucosa.

Keratinization of middle ear cholesteatomas. II. A histochemical study of epidermal transglutaminase substrates

A histochemical study was performed to clarify further the role played by epidermal transglutaminase (ETgase) in the keratinization of aural cholesteatoma. Weakly and strongly keratinized epidermal tissues and healthy middle ear mucosa were included as references. A first assay revealed the distribution of non-specified acyl donor substrates. In a second assay, the topography of involucrin was assessed immunohistochemically. In both epidermal and cholesteatoma matrix tissues, the presence of acyl donors was not restricted to the sites of (E)Tgase activity, but was almost uniformly extended throughout living layers. In reference tissues, residual acyl donors were poorly detected in horny layers, while they were more abundant in the stratum corneum of the cholesteatomas studied. The presence of involucrin along the cell membrane was observed at varying distances throughout the spinous and granular layers, depending upon the epidermal and matrix configurations. In thick epithelia, involucrin rapidly became concentrated at the cell periphery (in spinous keratinocytes), while in thin epithelia it was usually associated with cell flattening. This latter staining profile was observed more frequently in cholesteatomatous tissues. In addition, we regularly noticed an immediately suprabasal accumulation of involucrin, suggesting a locally hyperproliferative state of the matrix. An insufficient availability of acyl donors, especially involucrin, could not be used to explain the defective ETgase-mediated cross-linking of cholesteatoma cell membranes during terminal stages of differentiation. The present investigation may be the first to demonstrate the presence of involucrin in middle ear mucosa.

A comparative study of enzyme histochemical features in the gerbilline and human cholesteatoma

Spontaneous and experimentally induced cholesteatoma in the Mongolian gerbil has been found to exhibit histopathological similarities to human aural cholesteatoma and has been suggested as an experimental model for studies of the clinical situation. In an attempt to further characterize this model, we compared experimentally induced cholesteatomas in the external auditory canal from gerbils with those of the human ear by means of a correlated histopathologic and enzyme histochemical study. The human and gerbilline cholesteatomas revealed similar histopathologic features. Even enzyme histochemically, the human and experimentally induced cholesteatomas demonstrated similar features. Thus glucose-6-phosphate dehydrogenase activity, an indicator of oxidative metabolism, was demonstrated especially in the stratum granulosum cells of the heavily orthokeratinizing squamous epithelium adjacent to the cholesteatomas. The human ear canal skin also revealed enzyme histochemical characteristics similar to the squamous epithelium lining the human cholesteatoma. The hydrolytic enzyme activity (leucyl-aminopeptidase) was strong in the connective tissue surrounding human cholesteatoma when compared with that of ear canal skin. In the gerbilline cholesteatoma, this activity was demonstrated especially in the connective tissue adjacent to eroded bone, which possibly may facilitate cholesteatoma progression. We conclude that experimentally induced cholesteatoma has both histophatological and enzyme histochemical similarities to human aural cholesteatoma and therefore it is suggested that the gerbilline model may be used for studies on the development of human cholesteatoma. Our results support the view that cholesteatoma may originate from migrated hyperkeratinizing cells from the epidermis of the tympanic membrane or the meatus.

A comparative immunohistochemical study of cytokeratin and vimentin expression in middle ear mucosa and cholesteatoma, and in epidermis

Cytokeratin expression was studied in human middle ear cholesteatoma lesions, using a variety of immunohistological techniques and a wide range of polyclonal antisera and monoclonal antibodies against cytokeratin (CK) subgroups or individual CK polypeptides. The expression of the other cytoskeletal proteins, vimentin and desmin, was also investigated. Middle ear mucosa and epidermal tissues were used as reference tissues. Our investigations also included epithelial structures present in the cholesteatoma perimatrix and in dermal tissues. The results indicate that, compared with epidermal tissues, the expression profile of CKs in cholesteatoma matrix is representative of a hyperproliferative disease. Evaluating the presence of a marker of terminal keratinization - the 56.5 kD acidic CK n degrees 10 - we found supportive evidence of a pronounced retardation of its expression, which did not parallel histological differentiation. In epidermal tissues, the first prickle cell layers are CK10 positive whereas in many cholesteatomas this finding was observed near the stratum granulosum only. Probing the early stages of keratinization - the 58 kD basic CK n degrees 5 and the 50 kD acidic CK n degrees 14 - we regularly observed an extended staining area in the cholesteatoma matrix. In epidermal reference tissues, only the basal and nearest suprabasal layers were convincingly labeled. As a rule, non-epidermal CKs did not belong to the cholesteatoma CK set. However, exceptions to that rule were noticed as a focal or more extended expression of one or more non-epidermal CKs in about half of the cases. Together with the extended CK5 topography, this is further evidence that CK expression is seriously affected by the diseased state. CK expression in the perimatrix is limited to mucous glands, either normal, atrophic or hyperplastic. CKs n degrees 4, 5, 7, 14, 18 and 19, also displayed by middle ear mucosa, were consistently observed. Where ductal arrangements were present, CK10 was also detected, in analogy with the CK10 registration in ductal portions of mucous glands in the external ear canal skin. The absence of CK8 in mucous glands of the perimatrix, however, strongly differentiates these structures from the mucous gland acini and ducti in the external ear canal, where CK8 is systematically expressed. Vimentin staining was restricted to dendritic cells of the matrix (Langerhans cells) and to perimatrix fibroblasts, blood cells and vascular endothelium. Coexpression of CK and vimentin was not observed.