Oncocytic lesions of the ophthalmic region: a clinicopathological study with emphasis on cytokeratin expression


  • Jens Østergaard,

    1. Eye Pathology Section, Department of Neuroscience and Pharmacology, University of Copenhagen, Copenhagen, Denmark
    2. Department of Ophthalmology, Roskilde Hospital, Roskilde, Denmark
    Search for more papers by this author
  • Jan U. Prause,

    1. Eye Pathology Section, Department of Neuroscience and Pharmacology, University of Copenhagen, Copenhagen, Denmark
    Search for more papers by this author
  • Steffen Heegaard

    1. Eye Pathology Section, Department of Neuroscience and Pharmacology, University of Copenhagen, Copenhagen, Denmark
    Search for more papers by this author

Steffen Heegaard MD
Eye Pathology Section
Department of Neuroscience and Pharmacology
University of Copenhagen
Frederik V’s Vej 11
2100 Copenhagen
Tel: + 45 35 32 60 70
Fax: + 45 35 32 60 80
Email: sthe@sund.ku.dk


Purpose:  This study aimed to make a nationwide clinicopathological study of oncocytic lesions in the ophthalmic region and to characterize their cytokeratin (CK) expression.

Methods:  All histologically diagnosed oncocytic lesions in the ophthalmic region registered in Denmark over a 25-year period were collected and re-evaluated using a monoclonal antimitochondrial antibody (MU213-UC). Clinical data were registered. Immunohistochemical characterization was performed with a panel of anti-CK antibodies.

Results:  A total of 34 oncocytic lesions were identified and reviewed. The incidence that required surgical intervention in the Danish population could be approximated to 0.3 lesions per million capita per year. Patient ages ranged from 45 years to 89 years, with a peak incidence in the eighth decade. Female patients were twice as common as male. Lesions were typically described as red–brown, cystic and slow-growing. The antimitochondrial antibody MU213-UC produced a distinct and intense immunostaining of all oncocytic lesions and was found to be useful in substantiating oncocytic differentiation. Twenty-six of the lesions originated in the caruncle, three in the conjunctiva, two in the lacrimal sac, one at the semilunar plica, one on the eyelid margin and one peripunctally. Lesions were histologically classified as adenoma (oncocytoma) (26), hyperplasia (4) and metaplasia (4). Fourteen oncocytic lesions representing different locations and differentiation were further evaluated for CK expression. Basal-type oncocytic cells reacted with antibodies against CK 5/6, CK 7, CK 8, CK 13, CK 14, CK 17, CK 18 and CK 19, and suprabasal cells with CK 4, CK 7, CK 8, CK 18 and CK 19. Antibodies against CK 1+10 and CK 20 showed no reaction.

Conclusions:  Oncocytic lesions of the ophthalmic region most frequently present as caruncular oncocytomas. The CK profile is similar to the lacrimal- and accessory lacrimal gland duct elements and supports the theory that these lesions originate in the lacrimal- and accessory lacrimal glands.


An oncocyte is a characteristic cell phenotype defined by its swollen eosinophilic appearance and excessive mitochondria content (Tallini 1998). Oncocytic metaplastic transformation is a degenerative condition found in secretory epithelium (Tallini 1998). Oncocytic neoplasia has most frequently been reported in kidneys, thyroid and salivary glands (Tallini 1998).

In the ocular region the most frequently diagnosed oncocytic lesion is an oncocytoma (oncocytic adenoma) of the caruncle (Biggs & Font 1977). Oncocytomas account for about 3% of lesions in this location (Østergaard et al. 2006). Less frequently, lesions have been described arising in the lacrimal gland, conjunctiva, eyelid margin and lacrimal sac (Biggs & Font 1977).

The histogenesis of periocular oncocytic lesions has been addressed previously (Rennie 1980). However, whether oncocytic neoplasia in the ocular region arises de novo or develops from metaplastic epithelium is unknown. In the caruncle most authors point at accessory lacrimal gland tissue as the most likely site of origin (Biggs & Font 1977; Rennie 1980). Others have suggested the caruncular surface epithelium and the conjunctiva as the site of origin (Greer 1969). Notions of origin have primarily been based on histological appearance and relations.

Cytokeratin (CK) is a family of intermediate filaments found in the cytoskeleton of epithelial tissue (Chu & Weiss 2002). To date, 20 distinct subtypes of CK have been identified (Chu & Weiss 2002). The expression of CK varies with cell origin and differentiation (Chu & Weiss 2002). Therefore, CK pattern generally reflects origin and is widely used in the differential diagnosis of epithelial neoplasms (Chu & Weiss 2002). An extended characterization with CK markers has not previously been performed on oncocytic lesions in the ophthalmic region.

The aim of the present study was to make a nationwide clinicopathological study of oncocytic lesions in the ophthalmic region and to characterize CK expression in order to differentiate the site of origin.

Materials and Methods

The archives of the Eye Pathology Institute, University of Copenhagen, and the national Danish pathology database, Patobank, were searched for oncocytic lesions using SNOMED (Systematized Nomenclature of MEDicine) codes. All histologically diagnosed oncocytic lesions in the ocular region registered from 1981 to 2005 in Denmark were included. Patient files were obtained and reviewed.

Paraffin-embedded specimens from the identified lesions were collected. All sections were stained with haematoxylin and eosin (H&E), periodic acid-Schiff (PAS) and phosphotungstic acid-haematoxylin (PTAH) and re-evaluated. Histological diagnosis and classification were based on the Armed Forces Institute of Pathology classification of salivary glands (Ellis & Auclair 1996). Immunohistochemical staining was performed using a monoclonal anti-human mitochondrial antibody MU213-UC (clone no. 113-1, lot. no. MU2130506; BioGenex Laboratories, Inc., San Ramon, California, USA) (Tickoo et al. 1997). Control sections were incubated omitting the primary antibody. Human liver was used as positive control.

A total of 14 oncocytic lesions representing different locations and differentiation were further examined for CK expression. The lesions comprised six oncocytomas (four caruncular, one from the semilunar plica and one peripunctal), four hyperplasias (two from the conjunctiva, one caruncular and one from the eyelid margin) and four metaplasias (two caruncular and two from the lacrimal sac).

Antibodies against the following CKs were applied: CK 1+10 (clone LH1, cat. no. GTX40121; GeneTex, Inc., San Antonio, Texas, USA), CK 4 (clone no. 6B10, code no. NCL-CK4; Novocastra Laboratories Ltd, Newcastle upon Tyne, UK), CK 5/6 (clone no. D5/16 B4, code no. M 7237; Dako A/S, Glostrup, Denmark), CK 7 (clone no. OV-TL 12/30, code no. M 7018; Dako A/S), CK 8 (clone no. TS1, code no. NCL-L-CK8-TS1; Novocastra Laboratories Ltd), CK 13 (clone KS-1A3, code no. NCL-CK13; Novocastra Laboratories Ltd), CK 14 (clone LL002, code no. NCL-LL002; Novocastra Laboratories Ltd), CK 17 (clone E3, code no. M 7046; Dako A/S), CK 18 (clone DC-10, NCL-CK18; Novocastra Laboratories Ltd), CK 19 (clone RCK 108, code no. M 0888; Dako A/S) and CK 20 (clone Ks 20.8, code no. M 7019; Dako A/S).

Immunohistochemical reactions were performed using a biotin-free method (Envision Flex + Dako) (Dako A/S) to avoid a false positive reaction caused by endogenous biotin in the mitochondrial-rich tissue. Heat-induced epitope retrieval (HIER) in an alkaline buffer (pH 9.0) using a Pre-Treatment Module (Dako A/S) for 20 mins at 97 ° was used in all cases. Positive controls for each of the cytokeratins were performed. A multiblock containing normal skin, pancreas, liver and tonsil tissue was used for CK 1+10, CK 4, CK 5/6, CK 7, CK 8, CK 13, CK 14, CK 18 and CK 19. Duodenal tissue was used for CK 20 and breast tissue for CK 17. Negative tissue controls were performed for all the antibodies. Negative reagent controls were performed on sections incubated without the primary antibodies.

Staining intensity was graded as − (no staining), + (moderate staining) or ++ (intense staining). The location of the immunoreaction was noted and in case only a subset of oncocytic cells stained, the fraction of positive cells was estimated.

The study was approved by the Danish Data Protection Agency and the Danish Scientific Ethical Committee.


Clinicopathological findings

A total of 34 oncocytic lesions in the ocular region were identified during the 25-year period. The incidence of oncocytic lesions that required surgical intervention in the Danish population could be approximated to 0.3 per million capita per year. The mean patient age was 72 years (range 45–89 years) and incidence peaked in the eighth decade. Female gender (65%) was almost twice as common as male (35%). Duration of > 1 year was found in the majority of cases. The most frequently reported appearance was of a red–brown cystic lesion. Other colours noted were bluish (Fig. 1A), light and pigmented. Median two-dimensional lesion size was 6 × 5 mm (range 1 × 1 mm to 25 × 15 mm). A tumour (44%) or a cystic lesion (24%) were the two most frequently stated clinical diagnoses. No tumours recurred clinically.

Figure 1.

 (A) Oncocytoma of the caruncle (arrow) showing a bluish lesion in an 87-year-old woman. (B) Microscopic survey of the same lesion demonstrating an eosinophilic multicystic encapsulated lesion (arrowhead). (Haematoxylin and eosin [H&E].) (C) Oncocytic hyperplasia of the inferior fornix (arrowhead). No capsule is present. (H&E.) (D) Oncocytic metaplasia of chronically inflamed lacrimal sac. Note the distinct border between the oncocytic transformed epithelium and remaining epithelial lining (arrowhead) (anti-human mitochondrial antibody MU213-UC). (E) CK 7 immunoreaction in an oncocytoma. (F) CK 4 immunoreaction in an oncocytoma. Scale bars: (B, E) 200 μm; (C, D, F) 100 μm.

Location correlated to histological classification is presented in Table 1. Histologically, the lesions were classified as 26 oncocytomas (Fig. 1B), three nodular oncocytic hyperplasias (Fig. 1C), one diffuse oncocytic hyperplasia and four metaplastic oncocytic lesions (Fig. 1D).

Table 1.   Number of oncocytic lesions according to type and location.
LocationHistological classificationTotal
Lacrimal sac0022
Semilunar plica1001
Eyelid margin0101

All oncocytomas were partially or completely encapsulated (Fig. 1B). Cystic lumina containing eosinophilic material were present in all oncocytomas (Fig. 1B) and hyperplasias (Fig. 1C). Inflammation was present in 62% of cases (Fig. 1B, C). Tumour-free section margins were found in only 48%.


Phosphotungstic acid-haematoxylin (PTAH) was found to stain the cytoplasm in cells of oncocytic differentiation. The eosinophilic material in the cystic lumina stained with PAS.


Antimitochondrial antibody MU213-UC showed an intense granular cytoplasmatic staining of the oncocytic cells (Fig. 1D).

Fourteen oncocytic lesions were selected for further immunohistochemical characterization with the extended panel of CKs. The immunoreaction was located in the cytoplasm in all cases. Table 2 shows a summary of the CK antigenic profile of the basal and suprabasal epithelium. Strong reactivity was present for CK 5/6, CK 7, CK 8, CK 17, CK 18 and CK 19 (Fig. 1E). Moderate reactivity was found for CK 4 (Fig. 1F). For CK 13 only a subset of basal cells (5–50%) was stained. This was also the case for CK 14, where 5–75% of the basal-type cells reacted with a variable staining. No significant binding of CK 14 antibodies was found in two caruncular adenomas, two oncocytic hyperplasias of the conjunctiva and one oncocytic metaplasia of the lacrimal sac. Antibodies against CK 1+10 and CK 20 did not react with any lesion. A similar CK profile was observed in all the various oncocytic lesions, regardless of histological classification or location.

Table 2.   Cytokeratin profile of oncocytic lesions of the ophthalmic region including normal lacrimal gland ducts.
TissueCK 4CK 5/6CK 7CK 8CK 1 + 10CK 13CK 14CK 17CK 18CK 19CK 20
  1. * 5–50% of cells.

  2. † 0–75% of cells.

  3.  Intralobular-, interlobular- and excretory ducts (Kivelä 1992).

Oncocytic lesions
 Basal cells++++++++*−/+/++++++++
 Suprabasal cells+++++++++
Normal lacrimal gland ducts
 Basal cells++++++++ ++++
 Suprabasal cells+++++ ++++


Oncocytic lesions in the ophthalmic region are rare. The present nationwide study presents a large series of adnexal lesions of oncocytic differentiation. The clinicopathological findings are in accordance with previous literature (Biggs & Font 1977; Rennie 1980; Tallini 1998). Oncocytomas showed a strong predilection for the caruncle. However, other mucosal surfaces of the ophthalmic region were also affected (Table 1). The most frequent clinical presentation was a red–brown cystic tumour, most commonly found in elderly women. The low recurrence rate despite tumour-free section margins in only half of the cases stresses the slow growth rate of the tumour.

Histological classification was based on the established criteria for salivary gland oncocytic lesions (Ellis & Auclair 1996). Metaplasia (oncocytic transformed epithelial lining) and hyperplasia (foci of oncocytic cells forming cystic nodules) were relatively more frequent in locations outside the caruncle (Table 1). Oncocytomas were, by definition, larger, more circumscribed tumours and by far the most common histological type (Table 1).

Conventional histological evaluation with H&E and PTAH is often sufficient in diagnosing oncocytic lesions (Ellis & Auclair 1996). However, these stainings are unspecific and have been found not to be consistently reliable (Ellis & Auclair 1996). Electron microscopic evidence of numerous and abnormal mitochondria can be used in diagnosing oncocytic tumours. Immunohistochemical staining with monoclonal antimitochondrial antibodies is now available for use on paraffin-embedded material and has demonstrated a high sensitivity in light microscopic demonstration of mitochondria in oncocytic lesions of the kidneys and salivary glands (Shintaku & Honda 1997; Tickoo et al. 1997). In the present study, the oncocytic lesions were all found to exhibit a distinct and intense immunostaining markedly different from that of the surrounding cellular structures (Fig. 1D).

The high sensitivity makes the antimitochondrial antibody MU213-UC a reliable method of substantiating oncocytic differentiation in the ophthalmic region.

The present study characterizes for the first time a series of oncocytic lesions of the ophthalmic region with an extended panel of CK markers. CK 7, CK 8, CK 18 and CK 19 are all common markers of simple secretory epithelium (Chu & Weiss 2002). In the present study these markers were found to be strongly expressed throughout the oncocytic epithelium. Positive reaction was also found for CK 4, CK 5/6, CK 13 and CK 14. These CKs are generally expressed in stratified squamous epithelium (Chu & Weiss 2002). CK 1+10 is a marker of keratinized epithelium and CK 20 a marker of simple epithelium (Chu & Weiss 2002). In these cases no reaction was found in the oncocytic lesions.

Immunohistochemical findings in two caruncular oncocytic tumours consistent with oncocytomas were recently reported (George et al. 2007). As in the present study, both lesions were positive for CK 5/6/18, CK 5/6 and CK 7, but were CK 20-negative.

All lesions demonstrated a similar CK expression pattern (Table 2) irrespective of histological differentiation, which suggests that oncocytic metaplasia, hyperplasia and oncocytoma may represent the same type of lesion in different stages of development. This is in line with the histological observation that the distinction between metaplasia/hyperplasia and hyperplasia/adenoma is not always clear-cut (Ellis & Auclair 1996).

Similar CK expression patterns were found, irrespective of site of origin. This finding is controversial as several different structures have been proposed as precursors of oncocytic lesions of the ophthalmic region, including the conjunctiva, the epithelial lining of the caruncle, accessory lacrimal glands, sweat glands and mucous glands (Rennie 1980). Most authors, however, point at accessory lacrimal gland tissue as the most likely site of origin (Biggs & Font 1977; Rennie 1980). This observation is supported by electron microscopic findings showing morphological similarities between oncocytomas and accessory lacrimal glands (Rennie 1980). The CK profile in oncocytic lesions in the present study was found to be similar to lacrimal gland and accessory lacrimal gland duct elements (Table 2). The profile does not indicate an origin from lacrimal gland acini or intercalated ducts as cells of these locations do not express CK 4 (Kivelä 1992).

Finally, the CK profile was found to be distinctly different from that of renal oncocytomas (Langner et al. 2004). By contrast with oncocytomas of the ophthalmic region, renal oncocytomas were found to be negative for CK 4, CK 5/6, CK 13, CK 14 and CK 17 (Langer et al. 2004). Positive reactions for CK 7, CK 8, CK 18 and CK 19 were found in both locations. Thus, the CK profile of oncocytomas may vary according to the cells from which the oncocytomas derive.

In conclusion, oncocytic lesions of the ophthalmic region demonstrate a CK profile which is similar to lacrimal and accessory lacrimal gland duct elements. This finding supports the theory that oncocytic lesions are ductal neoplasms originating from lacrimal- and accessory lacrimal glands.


This study was supported by Synoptik Fonden, Copenhagen.