The Pathophysiology Of Cholesteatoma

Otolaryngol Clin N Am 39 (2006) 1143–1159

The Pathophysiology of Cholesteatoma Maroun T. Semaan, MD, Cliff A. Megerian, MD* Department of Otolaryngology and Head and Neck Surgery, University Hospitals of Cleveland, Case Western Reserve University, LKS 4500, 11100 Euclid Avenue, Cleveland, OH 44106, USA

Cholesteatoma is a cystic lesion formed from keratinizing stratified squamous epithelium, the matrix of which is composed of epithelium that rests on a stroma of varying thickness, the perimatrix. The resulting hyperkeratosis and shedding of keratin debris usually results in a cystic mass with a surrounding inflammatory reaction. It may present extradurally and intradurally. Extradurally, cholesteatoma most commonly involves the middle ear cleft but can occur in all portions of the petrous bone including the mastoid, petrous apex, and external auditory canal. Intradurally, cholesteatoma, also known as epidermoid, have been described in a variety of anatomic locations, the most common being the cerebellopontine angle. The history of cholesteatoma has been reviewed recently [1] and is summarized briefly. In 1683, Duverney [2] published the first description of what might correspond to a cholesteatoma. He described an abscess of the bone originating from the auditory canal that opened behind the auricle, forming a fistula above the mastoid process, shedding the small sheets composed of what he describes as scales. The abscess described was accompanied by a bad odor and gave rise to what was described as grave accidents. He also mentioned that the same process easily enters the middle ear cleft through the auditory canal, destroying its contents and resulting in deafness. Nearly a century and a half after Duverney’s original description, Cruveilhier [3] provided in 1829 a detailed description of what he thought was an avascular tumor originating from the cells of the subarachnoid space. Independently, Mu¨ller [4] in 1838 used the term cholesteatoma as he became aware of the presence of cholesterin and fat in what he believed to be a tumor. Although, he noted the resemblance between the squamae of cholesteatoma and the cells of the stratum corneum he did not postulate the epidermal origin of these lesions. In 1855, Virchow [5] classified * Corresponding author. E-mail address: [email protected] (C.A. Megerian). 0030-6665/06/$ - see front matter Ó 2006 Elsevier Inc. All rights reserved. doi:10.1016/j.otc.2006.08.003

oto.theclinics.com

1144

SEMAAN & MEGERIAN

cholesteatoma among squamous cell carcinomas and atheromas. However, because these lesions grew in bone, where epidermis does not exist, he considered them as heteroplastic tumors arising from mesenchymal cells that undergo dedifferentiation and then redifferentiation into epithelial cells. This postulation represents the first theory suggesting that cholesteatoma arise from mesenchymal cells undergoing metaplasia. Despite being a misnomer, the term cholesteatoma is still used today. Von Troeltsch [6,7] was the first to consider the epidermal origin of cholesteatoma. He theorized that epidermal debris accumulating in the external meatus are able to cause pressure-induced osteolysis of the bony wall of the meatus and thus invasion of the mastoid and the middle ear with extension if unchecked into the transverse sinus and brain. Gruber [8], Wendt [9] and Rokitansky [10] considered that middle ear mucosa rather than bone underwent malpighian metaplasia in response to chronic inflammation. The desquamated cells developed into cholesteatoma as the passage for squamae elimination became narrower. The theory of metaplasia became well accepted among otologists in the 19th century. At the end of the century, by studying two different pathologic entities, Bezold [11] and Habermann [12] proved that cholesteatoma could originate from the skin of the external auditory meatus, which migrates into the middle ear under the influence of chronic inflammation. Similar to normal skin, the migrated skin desquamated, and as the drainage passages became too narrow to enable migration, cholesteatoma forms. Habermann based his findings on the studying patients with marginal tympanic membrane perforation after acute necrotizing otitis; Bezold, however, studied cholesteatoma formation in patients with attic or posterosuperior retraction pockets secondary to eustachian tube dysfunction. Middle ear cholesteatoma occurs as two principle different entities that share many pathological resemblances: congenital and acquired. The latter is divided further into the more common primary acquired or attic retraction pocket cholesteatoma and the secondary acquired cholesteatoma as it occurs secondary to epithelial migration into the middle ear at the site of a tympanic membrane perforation or iatrogenically implanted during an otologic procedure. In this review, we limit our discussion to middle ear cholesteatoma and provide an updated literature review on the pathophysiology of congenital and acquired cholesteatoma. Emphasis will be placed on the pathophysiology of congenital and primary acquired cholesteatoma, cytokine-mediated inflammation and bony destruction. Congenital cholesteatoma The first published description of a congenital cholesteatoma appeared in 1885, by Lucae [13]. Ko¨rner’s initial criteria [14] to distinguish acquired from congenital cholesteatoma were revived half a century later by Derlacki

PATHOPHYSIOLOGY OF CHOLESTEATOMA

1145

and Clemis [15] who reintroduced the concept of congenital cholesteatoma in 1965. They proposed that congenital cholesteatoma be defined as a pearly white mass behind an intact tympanic membrane in the absence of history of otitis or otorrhea, tympanic membrane perforation, or previous otologic procedures. In 1986, Levenson and coworkers [16] suggested that the presence of prior bouts of otitis media does not necessarily exclude the presence of congenital cholesteatoma, because this inflammatory condition is very common among children. The incidence of congenital cholesteatoma is 0.12 per 100,000 [17]. There has been a recent increase in the reported incidence of this disease likely secondary to an increased awareness among pediatricians and otolaryngologists along with improvement in office based tools used for otologic examination (ie, otomicroscopy, halogen lightening, and photodocumentation). The pathogenesis of congenital cholesteatoma sparked an active debate that continues to this day. In 1936 Teed [18] described the presence of epithelial rests in fetal temporal bones that disappeared by 33 weeks of gestation. He postulated that the persistence of these cells leads to formation of congenital cholesteatoma. These rests were localized in the lateral wall of the eustachian tube in proximity of the tympanic ring in the anterosuperior quadrant of the middle ear. These findings were confirmed later by Michaels in 1986 [19] but failed to prove their persistence after 33 weeks of gestation. In 1998, Karmody and colleagues [20] described histologic findings of squamous epithelial rest in the temporal bones of two postpartum patients. This was the first description of these epithelial rests persisting beyond 33 weeks of gestation. In their first patient, they described the presence of a cupshaped elevation of squamous epithelium with a keratin cap noted in the anterosuperior quadrant of the middle ear. In their second patient, a small mass of squamous epithelium was seen embedded in the mucosa of the anterosuperior quadrant of the middle ear at the junction of the columnar and cuboidal epithelia. In their clinical study of a series of 160 congenital cholesteatoma, Potsic and coauthors [21] found that in cases of isolated quadrant involvement, 77% were anterosuperior and 22% were posterosuperior. The number of quadrants involved increased with age. The incidence of isolated posterosuperior quadrant involvement appears to be higher than initially thought. Many theories have been proposed to explain the origin of congenital cholesteatoma. The Teed-Michaels’ epithelial rest theory has been well accepted among otologists. Ru¨edi [22,23] speculated that inflammatory injury to an intact tympanic membrane results in microperforations in the basal layer that lead to invasion of the squamous epithelium by proliferating epithelial cones through a macroscopically intact but microscopically injured tympanic membrane. These epithelial cones fuse and expand forming a middle ear cholesteatoma. Tos [17] recently questioned the epithelial rest theory and proposed a different explanation for the pathogenesis of this disease. He observed that anterosuperior cholesteatoma had a frequent attachment to

1146

SEMAAN & MEGERIAN

the anterior aspect of the malleus handle or neck and that posterosuperior cholesteatoma had an attachment to the posterior aspect of the malleus handle and to the incudostapedial joint. This location was far from the anterior tympanic annulus and the lateral wall of the eustachian tube where epithelial rests are usually found. Furthermore, he speculated that if the site of origin was the lateral eustachian tube wall and the area anterior to the tympanic annulus, cholesteatoma would block the eustachian tube before extending into the tympanic cavity and the area of the malleus handle, a finding that has not been described previously. Therefore, he argued against the epithelial rest theory and explained the pathogenesis of congenital cholesteatoma by the acquired inclusion theory (Fig. 1). This theory speculates that keratinized squamous epithelium may be implanted or included into the tympanic cavity during one of many pathological events affecting the tympanic membrane and middle ear in childhood. According to Tos, viable

Fig. 1. ‘‘Acquired’’ inclusion theory suggested by Tos. (A1, 2) The tympanic membrane retracted and adherent to the malleus handle, malleus neck, or long process of the incus is loosened and torn leaving a small cuff of viable keratinized epithelium adherent to the ossicles with a small residual tear in the tympanic membrane. As the tear heals, the included epithelium leads to formation of an inclusion cholesteatoma. (B1, 2) A tangential tear is created as the retracted and adherent tympanic membrane is loosened from the underlying structure resulting in a remnant of epithelial cells without a perforation of the tympanic membrane that results in an inclusion cholesteatoma. (C1, 2) Microperforations of the traumatized retracted tympanic membrane result in invasion of the basal membrane by epithelial cones. As the ear drum is suddenly loosened, these cones are left behind and included in the tympanic cavity. (D1, 2) Similar to the previous mechanism, repeated inflammation of the tympanic membrane result in proliferating epithelial cones that penetrate the basal membrane and proliferate into the subepithelial space. These cones are included in the tympanic cavity as the drum is loosened and detached from the underlying bony structures.

PATHOPHYSIOLOGY OF CHOLESTEATOMA

1147

keratinized epithelial cells of the retracted and adherent tympanic membrane to the malleus handle, malleus neck, or the long process of the incus are left behind after loosening of the drum and are included into the tympanic cavity. Four mechanisms are thought to account for the inclusion of epithelial cells into the tympanic cavity. 1. The tympanic membrane retracted and adherent to the malleus handle, malleus neck, or long process of the incus is loosened and torn leaving a small cuff of viable keratinized epithelium adherent to the ossicles with a small residual tear in the tympanic membrane. As the tear heals, the included epithelium leads to formation of an inclusion cholesteatoma. 2. A tangential tear is created as the retracted and adherent tympanic membrane is loosened from the underlying structure resulting in a remnant of epithelial cells without a perforation of the tympanic membrane that results in an inclusion cholesteatoma. 3. Microperforations of the traumatized retracted tympanic membrane result in invasion of the basal membrane by epithelial cones. As the ear drum is suddenly loosened these cones are left behind and included in the tympanic cavity. 4. Similar to the previous mechanism, repeated inflammation of the tympanic membrane results in proliferating epithelial cones that penetrate the basal membrane and proliferate into the subepithelial space. These cones are included in the tympanic cavity as the drum is loosened and detached from the underlying bony structures. In response to Tos’ observations, Liang and coauthors [24] performed an immunohistochemical analysis of 36 temporal bones of 19 fetuses aged between 6 gestational weeks to 15 months postpartum. The investigators observed in each of the 22 temporal bones aged 16 gestational weeks to 8 months postpartum at least one epidermoid formation with a total of 116. The majority were found in the middle ear epithelium in the anterosuperior annular region of the tympanic cavity with a small number of epidermoid formations seen in the posterosuperior, anteroinferior, and posteroinferior region of the lateral wall in the vicinity of the annular zone. In addition, Liang and colleagues [24] examined the differential expression of 34bE12, a cytokeratin antigen expressed by the external ear epidermis and the pseudostratified columnar epithelium at all gestational ages, and 35bH11, a cytokeratin antigen expressed by pseudostratified columnar and simple cuboidal epithelium used to characterize the epidermoid formation seen in temporal bones histological sections. In addition, they used antibodies to antilymphoid enhancing factor-1 (LEF-1), a marker expressed by embryonic epidermis, to analyze the epidermoid formation precursor previously described by Michaels [19,25]. All epidermoid formations seen in their study stained positive for epidermal cytokeratin. The epidermoid formation precursor found in both temporal bones of an embryo aged 6 gestational weeks

1148

SEMAAN & MEGERIAN

did not stain for LEF-1. Thus, they concluded that the epidermoid formation precursor initially reported is likely the result of a tangential cut artifact of a thickened actively growing epithelial bud from the tip of the tubotympanic recess. Microscopically, they observed that as the anterosuperior tip of the meatal plate (precursor of the pars tensa) develops, by gestational week 12, the epidermal interface becomes jagged, and by gestational week 16, epidermal cells become encroached onto the fibroblasts of the bilaminar collagen layer. As the fibroblasts become more condensed, small clumps of epidermal cell become trapped within the condensed bilaminar collagen layer. Despite improvements in our understanding, the pathophysiology of congenital cholesteatoma continues to be controversial and actively debated. Furthermore, many questions remain unanswered. These questions pertain to the biological factors that predict aggressiveness, growth, and recidivism of middle ear congenital cholesteatoma (Fig. 2).

Acquired cholesteatoma Primary acquired cholesteatoma The pathophysiology of acquired cholesteatoma is similarly controversial. As previously eluded to, the precise pathogenesis of cholesteatoma has been debated for more than two centuries. Four predominant theories have fueled the debate: (1) invagination, (2) basal cell hyperplasia or papillary ingrowth, (3) metaplasia, and (4) epithelial invasion. The invagination theory is currently regarded as one of the primary mechanism of the formation of primary acquired attic cholesteatoma. Anatomic or pathological conditions that predispose to eustachian tube

Fig. 2. Site of origin and patterns of spread of congenital cholesteatoma according to (A) Tos ‘‘acquired’’ inclusion theory and (B) Teed-Michael’s epidermal rest theory.

PATHOPHYSIOLOGY OF CHOLESTEATOMA

1149

dysfunction result in barometric perturbation of the middle ear space. Impaired ventilation secondary to a dysfunctional eustachian tube leads to negative middle ear pressure. The negative pressure is the culprit for structural weakening of the tympanic membrane and development of retraction pockets. The pars flaccida, having the weaker structural support, is the most common site of formation of a retraction pocket. Sade [26] and Sade and Halevy [27] described four stages of tympanic membrane retraction: stage I, retracted membrane; stage II, retraction onto the incus; stage III middle ear atelectasis; and stage IV, adhesive otitis media. The geometrical changes attributed to progressive retraction lead to narrowing of the anatomic passages and impairment of the epithelial migration and cleaning of the keratin debris. As the pocket deepens and insinuates between mucosal folds and crevices, it becomes non–self cleaning and leads to accumulation of keratin debris (Fig. 3) Bacterial proliferation and super-infection of the accumulated debris form a biofilm that leads to chronic infection and epithelial proliferation. The latter appears to be influenced by the cytokinemediated inflammatory response. Chole and Faddis [28], analyzed the presence of biofilm matrix in cholesteatoma debris of 22 surgically induced Mongolian gerbils and 24 human specimens. The investigators detected the amorphous polysaccharide matrix suggestive of biofilm formation in 21 of 22 animals and 16 of 24 human cholesteatoma. Recently, Wang and coworkers [29] found that otopathogenic strains of pseudomonas aeruginosa are

Fig. 3. Mucosal compartmentalization of the middle ear. The mucosal folds of the middle ear cleft define the spaces that limit the boundaries of the retraction pockets. Knowledge of their anatomy helps understand the formation and extension of primary acquired cholesteatoma (black arrows). (1) superior incudal fold, (2) superior malleolar fold, (3) lateral incudal fold, (4) anterior malleolar fold, (5) lateral malleolar fold, (6) posterior malleolar fold. ET, eustachian tube orifice; HAC, hypotympanic air cells; RW, round window niche. Eustachian tube dysfunction results in formation of a retraction pocket. Often, a pars flaccida retraction pocket is formed (star). As the pocket deepens and insinuates between folds, the self-cleaning mechanism is altered and keratin accumulates.

1150

SEMAAN & MEGERIAN

capable of producing biofilm and become highly resistant to antimicrobial therapy. These findings strongly suggest a role of bacterial biofilm in the pathogenesis of cholesteatoma. The experimental model illustrating the implication of eustachian tube dysfunction in the formation of retraction pockets and later cholesteatoma was described by Kim and Chole [30]. By ligating the eustachian tube of Mongolian gerbils, the investigators succeeded in creating an induced, surgical model of primary acquired cholesteatoma (Fig. 4). The exact mechanism and triggers that lead to development of an active cholesteatoma in some patients with an attic retraction pocket while others continue to have a quiescent and self-cleaning pocket remain unclear. It has been shown recently that the combination of tympanic membrane retraction and basal cell proliferation is the hallmark for cholesteatoma formation and development. In a cohort of healthy children age 5 to 16 years, the prevalence of attic retractions was between 14% and 25% of ears [31]. In a separate cohort of children treated for secretory otitis with pressure equalization tube insertion with or without adenoidectomy and followed up to 18 years, the incidence of severe retractions (behind the scutum with some bone resorption) was 5% to 6% and attic cholesteatoma was 0.2% to 1.7%. Sudhoff and Tos [31] performed immunohistochemical analysis of surgical specimens obtained from 14 patients with middle ear cholesteatoma. In their clinical study, they compared the expression of MIB-1, a marker of cellular proliferation, between the cholesteatoma content and the normal external auditory canal skin. In addition, the investigators analyzed the integrity of the basement membrane by using avidin biotin complex peroxidase to stain collagen type IV. At the level of the basement membrane, interruption in the

Fig. 4. Patterns of spread of primary acquired cholesteatoma from an attic retraction pocket (D). (A) Antrum, most common; (B) posterior mesotympanum, second most common; and (C) anterior mesotympanum, least common.

PATHOPHYSIOLOGY OF CHOLESTEATOMA

1151

continuity was seen at the cholesteatoma–lamina propria interface, whereas the integrity was preserved in the adjacent normal auditory canal skin. They also showed an increased expression of MIB-1 in the keratinocytic population of the basal cell layer. This increased expressivity was consistent with proliferating keratinocytes localized primarily in small epithelial cones or pseudopods growing into the subepithelial stroma through interruptions of the basement membrane. Their observation provides experimental evidence that support the implication of both the retraction and basal cell hyperplasia theories. They postulated that in the initial retraction pocket stage, the epithelial migratory pattern is maintained until the pockets deepen and the drainage pathways become small leading to keratin debris accumulation. As the debris becomes infected, the bacterial proliferation and resultant inflammation leads to an influx of inflammatory cells and production of cytokines. This progression along with local release of collagenases created breaks in the basement membrane allowing the formation of epithelial cones that grow toward the stroma (papillary ingrowth theory). The combination of subepithelial invasion and keratinocytic proliferation in the form of microcholesteatoma is the hallmark of the precholesteatomatous stage of cholesteatoma. As the microcones expand and fuse together, an attic cholesteatoma is formed. Using the normal postauricular skin as control, Albino and coworkers [32] found a nine- to 20-fold increase in the expression of p53 in cholesteatoma tissue, trough all epithelial layers. The p53 proteins by activating downstream products (p21/WAF1, GADD45, and mdm2) appear to have a role in the down-regulation of cellular proliferation and promotion of apoptosis [33], a checkpoint control mechanism to protect the cell from genetic alterations. Similarly, they noted a two-fold increase in the expression of Ki-67, a marker of cellular proliferation, in cholesteatoma tissue compared with control normal postauricular skin. According to Albino and coauthors [34], the increased p53 expression was a feedback negative response to control an increased proliferative state as witnessed by the increase expression of Ki-67. Using immunohistochemistry, Kim and coworkers [35], analyzed the pattern of cellular proliferation and epithelial migration in the Mongolian gerbil animal model. They showed an increase in the expression of cytokeratin (CK) 13/16, markers of epidermal cell proliferation, in the expanding part of the cholesteatoma and to a lesser degree an increase in the expression of CK 5/6 and CK 1/10, markers of epithelial migration. They concluded that cellular migration (or invasion) and proliferation play a role in the expansion of cholesteatoma. On the other hand, Olszewska and coauthors [36], by studying the expression of five different cytokeratin (CK 10, CK 14, CK 18, CK 19 and 34bE12) concluded that congenital and acquired cholesteatoma exhibit a similar expression pattern. These findings suggested that the so-called ‘‘acquired’’ cholesteatoma in children may be an advanced congenital cholesteatoma that

1152

SEMAAN & MEGERIAN

resulted in destruction of the tympanic membrane, erosion of the ossicular chain, and invasion of the mastoid cavity. Epithelial invasion by cholesteatoma appears to be an important characteristic of this disease. Cholesteatoma expand by invading into surrounding middle ear soft tissue structures and bone. It remains unclear to what factors predict the biologic behavior of these lesions, such as recidivism and a more aggressive clinical course. Mallet and colleagues [37] found a correlation between the aggressiveness of the clinical behavior of cholesteatoma and the index of proliferation. In their analysis of surgical specimen from 91 ears with cholesteatoma, MIB-1 was detected in 23% of the ears with moderate bony destruction (single ossicle affected) versus 56% of ears with severe bony destruction (two or more ossicles, meningeal exposure, denudation of the facial nerve or sigmoid sinus, and erosion of the lateral semicircular canal). These findings were statistically significant. Young age was found to be a predictor of aggressiveness as witnessed by a higher proliferative index in children. Tokuriki and coworkers [38] performed gene expression analysis on human middle ear cholesteatoma using complementary DNA arrays. They compared the expression pattern of eight cholesteatoma to normal postauricular skin samples. They found an upregulation or induction in genes involved in cellular proliferation and differentiation (calgranulin A, calgranulin B, psoriasin, thymosin b-10) and cell invasion (cathepsin C, cathepsin D, cathepsin H, and matrix metalloproteinase 9 [MMP-9]). These results were confirmed using reverse transcriptase-polymerase chain reaction (RT-PCR) analysis. Immunohistochemical analysis showed increased expression of calgranulin A, calgranulin B, and calgranulin D in the cytoplasm of all cell layers of the cholesteatoma epithelium. Calgranulin proteins belong to the S100 protein family. In epithelial cells they may be involved in Ca2þ- dependent reorganization of cytoskeletal filaments [39]. Psoriasin, also a member of the S100 protein family, has been shown to be increased in hyperproliferative and inflammatory skin conditions and are believed to play a role in keratinocytic differentiation [40]. Upregulation and induction of these genes may reflect an alteration in keratinocyte differentiation and migration leading to keratin overproduction and accumulation as seen in cholesteatoma. The cathepsin family is a group of lysosomal proteases that play a key role in the degradation of intracellular and extracellular proteins in the epidermis and have been shown to contribute to the invasive properties of some neoplasms [41]. Cathepsin B has been shown to play a role in the osteolysis seen in cholesteatoma [42]. The increased keratinocyte proliferation is coupled with an increased cell death resulting in the production of larger amount of keratin debris responsible for the expansion and keratin accumulation seen in cholesteatoma. The implication of apoptotic cell death has been demonstrated recently [43]. Caspase-8 activation, a known effector of the extrinsic pathway of apoptosis, is

PATHOPHYSIOLOGY OF CHOLESTEATOMA

1153

triggered by activation of the cell surface death receptors (tumor necrosis factor [TNF] family, Fas-L/Fas-R). This activation results in the activation of an end product of apoptosis, caspase-3, that induces the nuclear translocation of effector molecules that result in apoptosis and programmed cell death. The transcription factor nuclear factor (NF)-kB is a known key mediator of the TNF-mediated cellular response. NF-kB proteins are intracytosolic and are inactivated by IkB-a (an inhibitory protein). The inactivation of IkB-a activates NF-kB and results in nuclear translocation of the transcription factor. The activation of NF-kB suppresses apoptosis induced by TNF-a. Miyao and coauthors [43] found an increased expression of caspase-3 localized to the granular and spinous layers of the cholesteatoma epithelium and an increased expression of caspase-8 confined to the granular layer. The retroauricular skin was used as a control. The NF-kB proteins were localized in the perinuclear region suggesting that the mechanism of negatively controlling apoptosis was inactivated leading to keratinocyte cell death and keratin accumulation. These findings strongly suggest differential properties inherent to cholesteatoma compared with normal epidermal keratinocytes that may explain their clinical aggressiveness and behavior responsible for the expansion, bony destruction and recidivism. Numerous studies have confirmed the implication of invagination, basal cell hyperplasia, and invasion in the pathogenesis of primary acquired cholesteatoma. The exact inciting events and factors responsible for the genesis and progression of middle ear cholesteatoma remain unclear, and further research is warranted to help elucidate these missing links. Secondary acquired cholesteatoma Secondary acquired cholesteatoma has been described to occur as the result of the migration of tympanic membrane epidermis into the middle ear at the site of a marginal perforation or as the result of the implantation of viable keratinocytes into the middle ear cleft. The implantation occurs during a blast injury to the tympanic membrane leaving keratinocytes behind a healed perforation, at the site of a temporal bone fracture, or as the result of an iatrogenic introduction of these cells. The latter have been described to occur in various otologic surgeries such as stapedectomy, tympanoplasty, pressure equalization tube placement, and middle ear exploration. Wolf and coauthors [44] described the otologic findings in 210 ears from 147 soldier-patients that sustained blast injuries with perforation of tympanic membrane localized to the pars tensa. These investigators reported an incidence of 4.8% of invasive cholesteatoma. Freeman [45] reported three cases of cholesteatoma secondary to temporal bone fracture. The keratinocytes appear to have invaded into the middle ear cleft through the fracture sites.

1154

SEMAAN & MEGERIAN

Golz and coauthors [46] performed a retrospective analysis of 2829 children who underwent a ventilation tube placement between 1978 and 1997. These investigators noted an incidence of 1.1% of middle ear cholesteatoma attributed to the insertion of the pressure equalization tube. The presence of cholesteatoma around the tube site was a prerequisite to incriminate the procedure as a cause of the cholesteatoma. They also noted a higher incidence in children aged less than 5 years, those with placement of Goode T-tubes, children with frequent reinsertions, patients with duration of placement exceeding 12 months, and ears with history of frequent postoperative otorrhea. Ferguson and coworkers [47] described the reasons for cholesteatoma formation after a stapedectomy. The investigators described four mechanisms: prosthesis extrusion independent of eustachian tube dysfunction, inadvertent implantation of keratinocytes with the oval window fat graft, malpositioned inverted tympanomeatal flap, and migration at the site of a marginal tympanic membrane perforation. Eavey and coworkers [48], and Camacho and colleagues [49] were able to produce viable keratinocytes in the bulla of gerbils and chinchilla, respectively, by implanting the mastoid space with autogenous keratinocytes obtained from the conchal surface of the pinna. Production of new keratin was observed up to 9 months postimplantation. Various histopathologic changes ranging from granulation tissue to cholesteatoma formation were described. The investigators concluded that neonatal aspiration of lanugo and viable keratinocytes can result in middle ear inflammation that in the chronic stage can lead to cholesteatoma formation. Bernal-Sprekelsen and coworkers [50], argued against this model of implantation of keratinocytes as an etiology for cholesteatoma and failed to find keratinizing epithelial cells in 31 temporal bones of infants who died before 1 year of age and 27 temporal bones of preterm fetuses that succumbed to various conditions. Despite the fact that the neonatal aspiration of viable keratinocytes may not fully account for the development of congenital cholesteatoma, it provides a valuable experimental platform that the implantation of viable keratinocytes can lead to formation of middle ear or mastoid cholesteatoma. This is observed frequently in revision middle ear surgery and described as a ‘‘cholesteatomatous pearl’’ formation that is the result of a trapped viable keratinocytic formation that leads to a small localized cholesteatoma. Another experimental model recently described by Massuda and Oliveira [51] provides physiopathologic evidence that supports epithelial migration at the edges of a tympanic membrane perforation as a possible cause for cholesteatoma development. By creating a tympanic membrane perforation and latex with 50% propylene glycol, the investigators succeeded in producing cholesteatoma in 90% and 80% of their animals, respectively. They concluded that latex provides a biomembrane that favors neoangiogenesis and forms a bridge for epithelial migration. This environment is enhanced further by a cytokine-producing acute or chronic inflammatory milieu created by the inciting material. This model may provide evidence that

PATHOPHYSIOLOGY OF CHOLESTEATOMA

1155

epithelial migration of keratinizing epithelium at the site of a tympanic membrane perforation, in the setting of recurrent inflammatory events, may be the culprit for cholesteatoma formation. Mechanism of bone destruction The ongoing debate on the pathogenesis of cholesteatoma is paralleled by the ongoing research to help elucidate the mechanism of expansion, bone destruction and invasion seen in middle ear cholesteatoma. Two predominant mechanisms are believed to account for the osteolysis seen in middle ear cholesteatoma: pressure-induced bone resorption and enzymatic dissolution of bone by cytokine-mediated inflammation. Pressure necrosis initially described by Steinbru¨gge in 1879 and Walsh in 1951, and direct bone resorption as described by Chole and coworkers [52] in 1985 have been proposed as possible mechanisms of bone destruction. Chole and colleagues implanted silicone sheets in the middle ear of gerbils without cholesteatoma and noted bone resorption at the pressure sites. They estimated that pressures of 50 to 120 mm Hg resulted in osteoclastic-induced bone resorption. The interaction of osteoclasts and osteoblasts to extrinsic biomechanical factors is a well-documented biological response [53,54]. It is uncertain to what degree the pressure-induced activation of osteoclasts play a role in the osteolysis seen in cholesteatoma. Enzymatic-induced and cytokine-induced bone destruction has been studied in the last two decades. Matrix metalloproteinases (MMP), a family of zinc metalloenzymes that degrades unmineralized extracellular matrix, have been shown to be present in the cholesteatoma [55]. MMP-2 (72 kD collagenase) and MMP-9 (92 kD collagenase) were expressed in suprabasal epithelial layers of cholesteatoma. Other investigators found the increased expression of MMP-9 but not MMP-2 in cholesteatoma cells [56]. Schmidt and coworkers [56] analyzed the in vivo significance of MMP-9 activity in relation to the production of cytokines interleukin (IL)-1a, IL-1b, TNF-a, transforming growth factor (TGF)-b, and epidermal growth factor (EGF) in tissue homogenates of 37 cholesteatoma and nine external ear skin specimens. IL-1a production was found to be significantly elevated; however, no correlation was found between MMP-9 activity and cytokine production. IL-1 and IL-8, important intercellular mediators of osteoclastic activities have been shown to increase in cultured cholesteatoma cells compared with normal external auditory canal skin. The role of another important cytokine, TNF-a, has also been found. Yan and coauthors [57] found that by in vitro stimulating monocytes, they were able to produce multinucleated cells with osteoclastlike activity that produced acid phosphatase-induced bone demineralization. The amount of osteolysis was increased by adding osteoblasts to the TNF-a– treated osteoclasts containing medium, suggesting a cell to cell interaction mediated by TNF-a. In addition, the latter enhanced the production of

1156

SEMAAN & MEGERIAN

collagenases by macrophages and osteoblasts. However, by performing enzyme-linked immunosorbent assay on tissue samples from 23 patients with cholesteatoma and 16 patients with chronic otitis without cholesteatoma, the detection of IL-1a, TNF-a, and EGF was significantly higher in the cholesteatoma samples [58]. Recent histopathologic evidence was obtained from the temporal bone of two patients with ruptured cholesteatoma sac resulting in local inflammation and osteolysis [59]. These changes were associated with a small abscess formation at the site of the rupture. They noted a marked inflammatory cellular infiltrate surrounding the rupture site with evidence of epithelial proliferation at the lining of the perforation site. Recent work by Jung and coworkers [60] showed the possible role of nitric oxide as an important mediator of osteoclast function. Using in vivo analysis of a murine model of cholesteatoma-induced bone resorption and in vitro analysis of osteoclast culture, the investigators studied the gene expression of nitric oxide synthase (NOS) and the effect of aminoguanidine (an inhibitor of cytokine mediated nitrite production). They showed a selective upregulation of the inducible NOS or NOS II compared with NOS I and III and a dose-dependent stimulation of osteoclastic activity (not proliferation) using low concentration of nitric oxide donors (sodium nitroprusside and S-nitro-N-acetyl-D, L-penicillamine). In vitro, only interferon (IFN)-g (not IL-1b or TNF-a) was able to generate nitrite. This nitrite production was blocked in vitro by the addition of aminoguanidine (but not in vivo) and was synergistically enhanced in the presence of IFN-g, IL-1b, and TNF-a. These findings indicate a role for nitric oxide in the osteoclastic-mediated bone resorption in cholesteatoma and suggest the implication of additional cytokines in the in vivo osteoclastogenesis and bone resorption. In contrast to the increased osteoclastic activity without increase in the number of osteoclasts seen by Jung and colleagues [60], in a separate study, Hamzei and coauthors [61] found an increase in the number of the osteoclast precursor cells in the perimatrix of 21 cholesteatoma surgically obtained. These studies highlight the importance of osteolysis and its regulatory mechanisms in the bone destruction seen in middle ear cholesteatoma that results in significant morbidity and mortality. Summary The pathophysiology of cholesteatoma continues to be debated widely. Cholesteatoma is classified as congenital or acquired. Recent studies appear to favor a possible common origin and overlap in the pathophysiology between both entities. Despite the growing evidence that the genesis, expansion, and progression of cholesteatoma is a complex interaction between anatomic, inflammatory, and regulatory factors of cellular proliferation and differentiation, the exact mechanism responsible for the invasion, recidivism, and destruction seen in this disease remains unknown.

PATHOPHYSIOLOGY OF CHOLESTEATOMA

1157

References [1] Soldati D, Mudry A. Knowledge about cholesteatoma, from the first description to the modern histopathology. Otol Neurotol 2001;22(6):723–30. [2] Duverney J. Traite de l’Organe de l’Ouie. Paris: E.Michaillet; 1683. [3] Cruveilhier J. Anatomie Pathologique du Corps Humain. Paris: Bailliere; 1829. [4] Mu¨ller J. Ueber den feineren Bau und die formen der krankhaften Geschwulste. Berlin: G.Reimer; 1838. [5] Virchow R. Ueber Perlgeschwulste. Arch Anat Physiol Klin Med 1855;8:371–418. [6] VonTroeltsch A. Lehrbuch der Ohrenheilkundle. Leipzig Germany: FCW Vogel; 1873. [7] VonTroeltsch A. Die Anatomie des Ohres in ihrer Anwendung auf die Praxis und die Kranskheiten des Gehorogans. Leipzig Germany: FCW Vogel; 1861. [8] Gruber J. Das Cholesteatome (Perlgeschwulst). Berlin: Cart Gerald’s Sohn; 1888. [9] Wendt H. Desquamativd Entzundung des Mittelohres. Arch Phys Heilkdt Wagner 1873;14: 428–46. [10] Rokitansky C. Neubildung von ausserer haut, Schleim und seroser Haut. Vienna: W Braumuller; 1855. [11] Bezold F. Cholesteatom, Perforation der Membrana Flaccida Schrapnelli und Tubenverschulus: eime atiologische studie. Vol 20: Z. Ohrenheilkd; 1889. [12] Habermann J. Cholesteatom des Mittelohres, seine Entstehung. Z. Ohrenheilkdt. 1889; 19:348. [13] Lucae. Politzer’s Textbook of the disease of the ear for students and practitioners. Philadelphia: Lea and Febiger; 1926. [14] Ko¨rner O. Die eitrigen Erkrankungen des Schlafenneins. Weisbaden Germany: JF Bergmann; 1899. [15] Derlacki EL, Clemis JD. Congenital cholesteatoma of the middle ear and mastoid. Ann Otol Rhinol Laryngol 1965;74(3):706–27. [16] Levenson MJ, Parisier SC, Chute P, et al. A review of twenty congenital cholesteatomas of the middle ear in children. Otolaryngol Head Neck Surg 1986;94(5):560–7. [17] Tos M. A new pathogenesis of mesotympanic (congenital) cholesteatoma. Laryngoscope 2000;110(11):1890–7. [18] Teed RW. Cholesteatoma verum tympani (its relationship to the first epibranchial placode). Arch Otolaryngol 1936;24:455–74. [19] Michaels L. An epidermoid formation in the developing middle ear: possible source of cholesteatoma. J Otolaryngol 1986;15(3):169–74. [20] Karmody CS, Byahatti SV, Blevins N, et al. The origin of congenital cholesteatoma. Am J Otol 1998;19(3):292–7. [21] Potsic WP, Korman SB, Samadi DS, et al. Congenital cholesteatoma: 20 years’ experience at The Children’s Hospital of Philadelphia. Otolaryngol Head Neck Surg 2002;126(4):409–14. [22] Ru¨edi L. Cholesteatoma formation in the middle ear in animal experiments. Acta Otolaryngol 1959;50(3–4):233–40 [discussion: 240–2]. [23] Ru¨edi L. Pathogenesis and treatment of cholesteatoma in chronic suppuration of the temporal bone. Ann Otol Rhinol Laryngol 1957;66(2):283–305. [24] Liang J, Michaels L, Wright A. Immunohistochemical characterization of the epidermoid formation in the middle ear. Laryngoscope 2003;113(6):1007–14. [25] Michaels L. Origin of congenital cholesteatoma from a normally occurring epidermoid rest in the developing middle ear. Int J Pediatr Otorhinolaryngol 1988;15(1):51–65. [26] Sade J. Retraction pockets and attic cholesteatomas. Acta Otorhinolaryngol Belg 1980; 34(1):62–84. [27] Sade J, Halevy A. The natural history of chronic otitis media. J Laryngol Otol 1976;90(8): 743–51. [28] Chole RA, Faddis BT. Evidence for microbial biofilms in cholesteatomas. Arch Otolaryngol Head Neck Surg 2002;128(10):1129–33.

1158

SEMAAN & MEGERIAN

[29] Wang EW, Jung JY, Pashia ME, et al. Otopathogenic pseudomonas aeruginosa strains as competent biofilm formers. Arch Otolaryngol Head Neck Surg 2005;131(11):983–9. [30] Kim HJ, Chole RA. Experimental models of aural cholesteatomas in Mongolian gerbils. Ann Otol Rhinol Laryngol 1998;107(2):129–34. [31] Sudhoff H, Tos M. Pathogenesis of attic cholesteatoma: clinical and immunohistochemical support for combination of retraction theory and proliferation theory. Am J Otol 2000;21(6): 786–92. [32] Albino AP, Reed JA, Bogdany JK, et al. Expression of p53 protein in human middle ear cholesteatomas: pathogenetic implications. Am J Otol 1998;19(1):30–6. [33] Perry ME, Levine AJ. Tumor-suppressor p53 and the cell cycle. Curr Opin Genet Dev 1993; 3(1):50–4. [34] Albino AP, Kimmelman CP, Parisier SC. Cholesteatoma: a molecular and cellular puzzle. Am J Otol 1998;19(1):7–19. [35] Kim HJ, Tinling SP, Chole RA. Expression patterns of cytokeratins in cholesteatomas: evidence of increased migration and proliferation. J Korean Med Sci 2002;17(3):381–8. [36] Olszewska E, Lautermann J, Koc C, et al. Cytokeratin expression pattern in congenital and acquired pediatric cholesteatoma. Eur Arch Otorhinolaryngol 2005;262(9):731–6. [37] Mallet Y, Nouwen J, Lecomte-Houcke M, et al. Aggressiveness and quantification of epithelial proliferation of middle ear cholesteatoma by MIB1. Laryngoscope 2003;113(2):328–31. [38] Tokuriki M, Noda I, Saito T, et al. Gene expression analysis of human middle ear cholesteatoma using complementary DNA arrays. Laryngoscope 2003;113(5):808–14. [39] Goebeler M, Roth J, van den Bos C, et al. Increase of calcium levels in epithelial cells induces translocation of calcium-binding proteins migration inhibitory factor-related protein 8 (MRP8) and MRP14 to keratin intermediate filaments. Biochem J 1995;309(Pt 2):419–24. [40] Algermissen B, Sitzmann J, LeMotte P, et al. Differential expression of CRABP II, psoriasin and cytokeratin 1 mRNA in human skin diseases. Arch Dermatol Res 1996;288(8):426–30. [41] Ravdin PM. Evaluation of cathepsin D as a prognostic factor in breast cancer. Breast Cancer Res Treat 1993;24(3):219–26. [42] Amar MS, Wishahi HF, Zakhary MM. Clinical and biochemical studies of bone destruction in cholesteatoma. J Laryngol Otol 1996;110(6):534–9. [43] Miyao M, Shinoda H, Takahashi S. Caspase-3, caspase-8, and nuclear factor-kappaB expression in human cholesteatoma. Otol Neurotol 2006;27(1):8–13. [44] Wolf M, Kronenberg J, Ben-Shoshan J, et al. Blast injury of the ear. Mil Med 1991;156(12): 651–3. [45] Freeman J. Temporal bone fractures and cholesteatoma. Annals of Otology, Rhinology and Laryngology 1983;92(6 Pt 1):558–60. [46] Golz A, Goldenberg D, Netzer A, et al. Cholesteatomas associated with ventilation tube insertion. Arch Otolaryngol Head Neck Surg 1999;125(7):754–7. [47] Ferguson BJ, Gillespie CA, Kenan PD, et al. Mechanisms of cholesteatoma formation following stapedectomy. Am J Otol 1986;7(6):420–4. [48] Eavey RD, Camacho A, Northrop CC. Chronic ear pathology in a model of neonatal amniotic fluid ear inoculation. Arch Otolaryngol Head Neck Surg 1992;118(11):1198–203. [49] Camacho AE, Eavey RD, Northrop C. Juvenile keratin inoculation induces chronic ear pathology. Am J Otol 1997;18(6):773–9. [50] Bernal-Sprekelsen M, Sudhoff H, Hildmann H. Evidence against neonatal aspiration of keratinizing epithelium as a cause of congenital cholesteatoma. Laryngoscope 2003;113(3): 449–51. [51] Massuda ET, Oliveira JA. A new experimental model of acquired cholesteatoma. Laryngoscope 2005;115(3):481–5. [52] Chole RA, McGinn MD, Tinling SP. Pressure-induced bone resorption in the middle ear. Ann Otol Rhinol Laryngol 1985;94(2 Pt 1):165–70. [53] Burger EH, Klein-Nulen J. Responses of bone cells to biomechanical forces in vitro. Adv Dent Res 1999;13:93–8.

PATHOPHYSIOLOGY OF CHOLESTEATOMA

1159

[54] Klein-Nulend J, van der Plas A, Semeins CM, et al. Sensitivity of osteocytes to biomechanical stress in vitro. FASEB J 1995;9(5):441–5. [55] Schonermark M, Mester B, Kempf HG, et al. Expression of matrix-metalloproteinases and their inhibitors in human cholesteatomas. Acta Otolaryngol 1996;116(3):451–6. [56] Schmidt M, Grunsfelder P, Hoppe F. Up-regulation of matrix metalloprotease-9 in middle ear cholesteatoma–correlations with growth factor expression in vivo? Eur Arch Otorhinolaryngol 2001;258(9):472–6. [57] Yan SD, Huang CC. The role of tumor necrosis factor-alpha in bone resorption of cholesteatoma. Am J Otolaryngol 1991;12(2):83–9. [58] Yetiser S, Satar B, Aydin N. Expression of epidermal growth factor, tumor necrosis factoralpha, and interleukin-1alpha in chronic otitis media with or without cholesteatoma. Otol Neurotol 2002;23(5):647–52. [59] Suzuki C, Ohtani I. Bone destruction resulting from rupture of a cholesteatoma sac: temporal bone pathology. Otol Neurotol 2004;25(5):674–7. [60] Jung JY, Pashia ME, Nishimoto SY, et al. A possible role for nitric oxide in osteoclastogenesis associated with cholesteatoma. Otol Neurotol 2004;25(5):661–8. [61] Hamzei M, Ventriglia G, Hagnia M, et al. Osteoclast stimulating and differentiating factors in human cholesteatoma. Laryngoscope 2003;113(3):436–42.