Cholesteatoma: A molecular and cellular puzzle

Hypothesis: There are at least three possible molecular models of cholesteatoma pathogenesis. Cholesteatoma may arise as a result of 1) the induction of a preneoplastic or neoplastic transformation event; 2) a defective wound-healing process; and/or 3) a pathologic collision of the host inflammatory response, normal middle ear epithelium, and a bacterial infection. Background: There have been a number of speculations concerning the factors that foster the development of cholesteatoma. Before resolving the molecular basis for the pathogenesis of cholesteatomas, it is important to present and test plausible models that could explain how a cholesteatoma becomes invasive, migratory, hyperproliferative, aggressive, and recidivistic. Methods: The authors evaluated by various techniques (e.g., immunohistochemistry, flow cytometry, and image analysis) a large number of cholesteatomas of all types (e.g., primary and secondary acquired, recurrent, and congenital) and a range of normal tissues (tympanic membrane, canal wall skin, and postauricular skin) for the expression of various proteins (e.g., p53, ectopeptidases, tryptase) and for the presence of DNA aneuploidy. Results and Conclusions: The authors' published and unpublished studies to date support several suppositions concerning the pathology of cholesteatomas. First, cholesteatoma epithelium behaves more like a wound-healing process than a neoplasm. The available evidence to date does not indicate that cholesteatomas have inherent genetic instability, a critical feature of all malignant lesions. Second, the induction of hyperproliferative cells in all layers of the cholesteatoma epidermis implicates a potential idiopathic response to both internal events as well as external stimuli in the form of cytokines released by infiltrating inflammatory cells. Third, the presence of bacteria may provide a critical link between the cholesteatoma and the host, which prevents the cholesteatoma epithelium from terminating specific differentiation programs and returning to a quiescent state in which it becomes minimally proliferative, nonmigratory, and noninvasive. Fourth, none of our data suggest that there are any obvious molecular or cellular differences among the various types of cholesteatomas (e.g., primary and secondary acquired, recidivistic, and congenital). Continued research should delineate the precise molecular and cellular dysfunction involved in the pathogenesis of cholesteatomas and how this knowledge can be useful in the clinical management of cholesteatomas.