Errata: Corrigenda Volume 18, Issue 8, 614, Article first published online: 21 July 2011
Maria V Yusenko phd, Laboratory of Molecular Oncology, Medical Faculty, Ruprecht-Karls-University, Im Neuenheimer Feld 325, D-69120 Heidelberg, Germany. Email: firstname.lastname@example.org
Differentiating renal oncocytoma from its renal cell carcinoma (RCC) mimics, particularly chromophobe RCC, can be difficult, especially when limited tissue is available for evaluation and requires sophisticated microscopic, ultrastructural and immunohistochemical evaluation. In this review, the relevant literature has been reviewed, and supporting data obtained by applying modern microarray-based technologies are discussed with a focus on molecular pathology of renal oncocytoma. The high resolution whole-genome DNA-microarray based analyses excluded with all certainty the occurrence of small specific alterations. Renal oncocytomas are characterized by variable chromosomal patterns. The number of genes selected by global gene expression analyses and their usefulness in the diagnostic pathology based on immunohistochemical evaluation is far below the expectations. The conflicting staining patterns, together with the poor specificity of proposed antibodies, leads us to believe that these candidate immunomarkers might not help in the separation of these tumors. Applying DNA based tools might help in the diagnosis of renal oncocytoma with uncertain histology. However, only the combination of all available techniques could give reliable information.
Prior to 1975, there had been just six individual case reports of typical renal oncocytomas first described in 1942.1 Later, Klein and Valensi2 identified another 13 cases as a distinctive clinical pathological entity, which they introduced in the English literature as “proximal tubular adenoma with oncocytic features”. This specific prodrome, by convention now called “renal oncocytoma”, comprises approximately 5% of all neoplasms of renal cells in surgical series. The ages of patients with typical renal oncocytoma have varied considerably. Patients have ranged in age from 10 years to 94 years.3–6 The peak incidence of detection is in the seventh decade of life. This finding suggests that renal oncocytomas occur in a relatively older age group than in those patients with typical renal cell carcinoma (RCC), for whom the median age in large series is always approximately 55 years. However, taking into account that renal oncocytoma grows very slowly, it might start developing at an earlier age, around 30–50 years. For the cases of renal oncocytoma with reported sex, men are affected nearly twice as often as women.
Clinical symptoms, radiological findings and treatment
The majority of patients with renal oncocytoma are usually asymptomatic.3,5 Most tumor cases were discovered incidentally during the course of clinical check-up of a urinary tract infection or prostate disease or at autopsy.7 Few patients described adequately in published clinical series had renal symptoms: microscopic or gross hematuria, abnormal or flank pain, or a palpable mass.6 Because renal oncocytomas are well encapsulated and rarely invade the renal pelvis, hematuria is mostly unusual. In general, it is not possible to differentiate preoperatively renal oncocytomas from the other renal carcinomas by radiological imaging techniques. Such resemblance might lead to radical nephrectomy, as until now almost all renal oncocytomas were considered to be typical RCCs and therefore most patients were treated by total nephrectomy. Nowadays, new procedures, such as partial or heminephrectomy or tumor excision, are used for bi- and unilateral renal oncocytomas and the diagnosis of small renal oncocytoma might be suggested by computed tomography scanning, ultrasound examination and magnetic resonance imaging, especially large sized oncocytomas with a prominent central stellate scar.8
Current options for the treatment of renal oncocytomas include radical nephrectomy, nephron-sparing surgery and minimally invasive approaches, such as cryo- and radiofrequency ablation. In the context of a nephron-sparing approach, a preoperative diagnosis by needle core biopsy, which is being increasingly carried out, is needed to guide management decisions.9 However, in these small biopsies, the entire range of cytoarchitectural features that are generally necessary to make a diagnosis might not be fully evaluable.10 The overlapping morphology and topographic variability of cellular and growth pattern of renal cell tumors (RCT) makes the preoperative diagnosis of oncocytoma or chromophobe RCC uncertain in these small specimens. Properly carried out fine-needle aspiration biopsy is a more reliable means to avoid the sampling error preoperatively than needle core biopsy.11 However, in all cases a biomarker support is necessary to make an appropriate diagnosis.
Renal oncocytomas often have a typical macroscopic appearance, which differs significantly from the typical gross appearance of RCCs. They are solitary, in rare instances multiple, generally well-encapsulated with a thick, well-defined, fibrous capsule.12 The surface on sectioning is usually mahogany-brown and less often tan in color. Large oncocytomas occasionally have a characteristic central stellate fibrous scar, which is composed of loose or dense hyalinized connective tissue containing occasional entrapped tumor cells and represents a sign of slow growth of this renal epithelial neoplasm.13 On the cut section, renal oncocytomas generally have a homogeneous appearance with no hemorrhage or necrosis within the tumor. Cases of histologically typical renal oncocytoma with calcification in the tumor center,14 necrosis,14,15 a liquid center,16 or a large central cyst,17 and numerous cysts18 have been described.
Tumor location and size
Renal oncocytomas occur in equal numbers in the right and left kidney. Few cases of bilateral oncocytomass, oncocytomas on one kidney with multiple oncocytomas in the contralateral kidney, and multiple oncocytomas in both kidneys have been reported.8,19 Tumors found in autopsy series have generally been small, but for the patients presented in clinical series, the median size has been approximately 6 cm. Accordingly, the old clinical axiom that detection of a solid renal mass in a clinical setting larger than 3 cm in diameter identified a malignant neoplasm, specifically a RCC, and that a 3 cm diameter provided a firm separation between adenoma and carcinoma8 has been totally disrupted by the simple existence of renal oncocytomas as well as papillary RCT.
Histopathological examination carried out for hematoxylin and eosin (HE) stained sections of tumor tissue is the first and the most important step in the diagnostic approach to epithelial renal neoplasms. It gives the information about cellular, nuclear, cytoplasmic, stromal and vascular network features as well as about growth pattern.
Renal oncocytomas are histopatologically characterized by “oncocytes”, which by definition are large neoplastic cells with intensely eosinophilic granular cytoplasm that results from the large number of mitochondria. This term was coined by Hamperl from Greek “onkousthai” and “cyte”, to swell and cell or swelling cytoplasm.20 Oncocytomas are found in a number of organs and have been described previously in the thyroid gland, salivary gland, parathyroid gland, adrenal gland and other anatomical sites.8 The exact cell type that gives rise to an oncocytoma in the kidney or other organs is unknown. Most pathologists suggest a distal tubular origin for renal oncocytomas,21,22 although a proximal tubular origin was first proposed.2,23
In general, renal oncocytoma has a variable morphological appearance. The cells are usually arranged in solid compact nests/alveoli (acinar growth exemplar) and/or in cords, tubules and sheets of trabeculae (tubulocystic arrangement) that are separated by loose edematous fibrous or hyalinised stroma (Fig. 1a). However, papillary and cystic architecture might occur in renal oncocytomas, which might be composed of regularly large oncocytes or small “basophilic” cells (Fig. 1b). Recently, another aspect of two cell types of renal oncocytoma has been discussed.24 The predominant classic form of cells (so-called “oncocyte”) corresponds to round-to-polygonal cells with densely granular eosinophilic cytoplasm, round and uniform nuclei with finely distributed chromatin, and a centrally placed prominent nucleolus. A smaller population of cells (called oncoblasts) has less conspicuous, paler, scanty, granular cytoplasm, a high nuclear/cytoplasmic ratio, and dense dark hyperchromatic, markedly pleomorphic nuclei. Bizarre, polyploidy, enlarged nuclei, which are characteristic for endocrine adenomas, might be scattered throughout the renal oncocytomas, but mitoses are absent.
Summarizing the histological data, although the diagnosis of renal oncocytoma is relatively easy in experienced hands, difficulties might arise when this neoplasm has atypical morphology. Additional studies, such as electron microscopy, chromosomal analysis and immunohistochemistry, might help in achieving a correct diagnosis.
Ultrastructural studies have shown that the eosinophilic granularity of oncocytes (Fig. 1c) results from the fact that the cytoplasm of oncocytes has scant lipid and the absence of glycogen, and is packed with mitochondria25,26 showing lamellar or focally stacked cristae, rare matrix densities, but not tubulovesicular cristae or other abnormalities.27,28 Their mean size was slightly larger than those of chromophobe RCC.26 The total absence or sparse occurrence of vesicles predominantly between lysosomes at the cell periphery, but not between mitochondria, has also been frequently observed in renal oncocytoma.28,29
Microvesicles, present in both renal oncocytoma and chromophobe RCC albeit in a different quantity, are also normally present in intercalated cells, which comprise up to 50% of the cortical collecting duct epithelium, decrease towards the medullary region and play an important role in regulating acid/base balance.22,30 It was suggested that renal oncocytoma and chromophobe RCC arise from normal α- and β- intercalated cells, respectively, with subsequent divergent differentiation, and therefore they are closely related.
On the whole, ultrastructural studies might help to differentiate renal oncocytoma from chromophobe RCC, but this technique is not always available, and in formalin-fixed material, such as paraffin-embedded tissue, the typical microvesicular structures are frequently disrupted by dehydrating agents during paraffin embedding.31 Alternatively, a glutaraldehyde or formalin fixed tissue can be used; however, a side-effect associated with formalin fixation is the formation of membrane blisters arising from both surface and cytoplasmic membranes.32
The tinctorial characteristics of different RCT appear to be dependent on cytoplasmic contents, including various organelles. As noted earlier, the distinction between renal oncocytoma and the eosinophilic variant of chromophobe RCC using routine light microscopy of HE-stained tissue sections remains problematic in some cases. The type of positivity to the proposed histochemical marker, Hale's colloidal iron, differs between the tumors, producing a microvesicular cytoplasmic pattern in chromophobe RCC and apical positivity in oncocytoma.33 Such apical positivity in oncocytomas and technical difficulty as well are the main limitations of a wider use of Hale's colloidal iron.
Recently, immunostaining patterns with anti-mitochondrial antibody have been reported to be a useful discriminatory adjunction in the complex differential diagnosis of granular RCT.25 Distinctive staining patterns were observed among the tumor groups, with chromophobe RCC showing characteristic peripheral accentuation of coarse cytoplasmic granules and oncocytoma with diffuse and fine granularity (Fig. 2). Therefore, the distribution of the immunoreactivity correlates with known electron microscopic findings in these tumors. The paradoxical finding is in the differences in size of the granular immunoreactivity between the tumor subtypes. As stated by ultrastructural examination findings, renal oncocytomas have the largest mitochondria and also diffuse but fine cytoplasmic granules as compared with the coarser granules in chromophobe RCC. One explanation might lie in the origin and nature of the microvesicles in chromophobe RCC. As suggested, the vesicles might be derived from the outer membranes of mitochondria.34,35 However, it would not explain the coarse granules that were observed in the granular variant of conventional RCC and the eosinophilic variant of papillary RCC.
On the whole, it seems difficult for general pathologists to accurately evaluate the staining patterns with colloidal iron staining and anti-mitochondrial antibody.
Locus-specific and whole-genome analysis
The ploidy analysis, generally carried out by flow cytometry, has been reported for many cases of renal oncocytoma and chromophobe RCC. A diploid pattern and, rarely, near-diploid aneuploidy has been shown in oncocytomas,36 whereas chromophobe RCC has been shown as a hypodiploid tumor.35,37 The latter did not provide any help in the clinical or pathological study of oncocytoma and its separation from other types of RCC with overlapping phenotype.8
Cytogenetic analysis on renal oncocytoma has shown a striking tendency to random telomere shortening and several telomeric associations of chromosomes, with the smaller chromosomes or the short arm of larger chromosomes being preferentially involved in this phenomenon.38,39
Cytogenetic studies of renal oncocytomas have shown a variety of chromosomal patterns, suggesting the existence of genetically distinct subsets,40–44 although most tumor cases display a mixed population of cells with normal and abnormal karyotypes.45
The first group consists of tumors showing the combined loss of chromosomes 1 and X/Y,46–48 which has been regarded as the most frequent chromosome abnormality in renal oncocytomas. These cytogenetic findings were further supported by the FISH-analysis showing the loss of chromosome 1 and Y,49 chromosomal CGH showing the loss of genetic material from chromosome 1,50 and single nucleotide polymorphism (SNP)-based oligoarrays showing complete or partial loss of chromosome 1.51,52 On the whole, partial or complete loss of chromosome 1 is the most common alteration in both sporadic and familial cases.53 Recent findings show that, in particular, the loss of 1p is important in renal oncocytoma, suggesting the presence of a tumor suppressor gene at this chromosomal region.54 The role of -Y, however, is debatable because oncocytomas also occur in females and the loss of chromosome Y is a very common phenomenon in solid tumors.
The second group is composed of tumors showing translocations involving chromosome 11, with a breakpoint at 11q12-13. Genes located on this chromosomal region might be involved in the development of this tumor. Sinke et al.55 mapped this oncocytoma-specific 11q13 breakpoint in two t(5;11)-positive renal oncocytomas within an interval of at maximum 400 kb flanked by the markers D11S443/D11S146 and the CCND/BCL1 locus. A few oncocytomas with a translocation of t(5;11)(q35;q13) have also been detected later.48,56 Other cases, showing a t(9;11)(p23;q12)57 and t(9;11)(p23;q13),58,59 were found to be located in the same region. Later, chromosomes 1, 6, 760 and 861 were also reported to be partners involved in structural rearrangements of 11q13. The FISH analysis showed that the 11q13.3 breakpoints were located near the CCND1 (previously PRAD1, BCL1) gene. The findings of CCND1 rearrangement with FISH and the correlation with cyclin D1 overexpression under immunohistochemical analysis suggest that alterations of cyclin D1, which is important in cell cycle regulation, play a role in the subset of renal oncocytomas with 11q translocations, although other genes might be involved while this upregulation is independent of the partner chromosome involved.60 The fusion of 11q13 with a region of another chromosome might result in a new fusion gene, the expression of which is responsible for uncontrolled growth. Direct activation of a putative oncogene by juxtaposition to active promoter/enhancer sequences is another possible mechanism. It is well established that 11q13 is frequently involved in hematological disorders and in a subset of carcinomas arising from various organs including breast, head and neck, esophagus, and bladder.50,62 Considering the unique cytomorphology of oncocytomas, marked by an abundance of mitochondria, it is of interest to note that the band 11q13 harbors a number of genes encoding mitochondrial proteins (UCP2, UCP3, NDUFC2, SDHD) located downstream of the breakpoint.63–65 However, some of the rearranged bands, such as 2q13, 19q13 or 22q13, also harbor genes encoding functional or structural mitochondrial proteins, suggesting that mitochondrial enzymes might play an important role in tumorigenesis of oncocytomas.66,67
Other rare chromosome rearrangements have been reported, such as t(1;12)(p36;q13),68 loss of chromosome 1444,50,69 and gain of chromosome 12.49
Because most rearrangements are balanced translocations without any net change in genetic material, CGH studies or microsatellite analysis fail to detect these anomalies. Perhaps, more cytogenetic analyses are necessary to confirm the idea of at least two different subgroups of oncocytoma and to find possible candidate genes in the chromosomal bands involved. It should be noted that no morphological differences have been observed between the genetically different subsets of renal oncocytoma. Chromosome changes are much rarer in familial than in sporadic oncocytomas.53 The distinct genetic changes observed might have a biological meaning, one of which could be the reported malignant potential of some of the oncocytomas published thus far. Nevertheless, the lack of genetic changes specific for other types of RCT (Table 1) combined with the histological characteristics might also be helpful in the diagnosis of renal oncocytoma.69,70 Recently, in the sole report on the DNA-array analysis of a large number of renal oncocytomas and chromophobe RCCs, some common and highly discriminating genetic changes have been found.52 The loss of heterozygosity at chromosome 1 was revealed in 10/42 (23%) of oncocytomas and 28/30 (93%) of chromophobe RCCs, whereas monosomy of chromosomes 2, 10, 13, 17 and 21 occured exclusively in chromophobe RCC (Fig. 3). Based on this finding, probes localized to chromosomes 2, 10, 13, 17 or 21 can be used in histopathological laboratories for the accurate differential diagnosis of chromophobe RCC and renal oncocytoma with uncertain histology.
Table 1. Comparison of predominant distinguishing features of renal oncocytoma and eosinophilic variants of renal cell carcinomas
†The most characteristic chromosomal lose (−), translocation (t) or gain (+). cRCC, conventional renal cell carcinoma; chRCC, chromophobe renal cell carcinoma; pRCC, papillary renal cell carcinoma; RO, renal oncocytoma.
From the studies on the chromosome level, multiple genes appear to be involved in the development of sporadic tumors. They can largely be divided into tumor suppressor genes, whose physical and functional loss will stimulate tumor development, and oncogenes, whose activation and amplification will do the same. Frequent loss of heterozygosity of all or part of a chromosome arm in renal oncocytoma might point to tumor suppressor gene inactivation and can thus show the location of a tumor suppressor gene. However, a recent study using 250K SNP-based oligoarrays has excluded with all certainty the occurrence of small specific alterations in the 42 renal oncocytomas being analysed.52
The genes affected by the specific chromosomal alterations in renal oncocytomas have not been identified yet. The germline mutation of the folliculin gene is associated with Birt–Hogg–Dube syndrome and the development of so-called hybrid chromophobe-oncocytic RCT,71 but the folliculin gene is not mutated in sporadic renal oncocytoma.72
Any abnormalities in the mitochondrial genome might lead to the alteration of function and/or morphology seen in the cells of renal oncocytoma. Since Welter et al. showed a unique HinfI restriction pattern of mitochondrial DNA in renal oncocytoma,73 the question has been raised whether alterations of mtDNA also might be involved in both the tumoro- and mitochondriogenesis. It has been suggested that such a mutation in mtDNA might lead to the increased rate of mitochondrial replication and probably to the proliferation of neoplastic cells.74 However, the involvement of nuclear genes controlling mitochondrial replication cannot be excluded and a drop in any respiratory chain complexes I–V might activate mitochondrial proliferation. Indeed, the deficiency of complex I activity and protein contents as well as a lack of its assembly has been observed in renal oncocytoma.75,76 Recently, using a proteomic approach based on two-dimensional gel electrophoresis and mass-spectrometry has allowed the downregulation of NDUFS3 (complex I) and upregulation of COX5A and COX5B (complex IV) as well as ATP5H (complex V) to be shown.77
Gene expression and immunoprofile
On the whole, histopathological examination is the first important step in the diagnostic approach to eosinophilic RCT, including renal oncocytoma and eosinophilic variants of chromophobe, conventional and papillary RCC. However, there remains a small but significant number of renal tumors with cells having eosinophilic cytoplasm that, even in the hands of experienced pathologists, might not be classified with certainty using microscopy alone. Immunohistochemistry is a useful ancillary tool for distinguishing one subtype of renal cell neoplasia from another; however, immunohistochemical interpretation can be subjective.
As generally accepted, genome-wide gene expression profiling is a powerful technique for identifying diagnostic markers. Until now, mRNA expression profiles of more than 300 RCT, including 35 renal oncocytomas, have been studied by filter and microarray hybridization.78–87 Although renal oncocytoma and chromophobe RCC have a distinct genotype, their morphological features can overlap and some authors consequently underline the high similarities in mRNA expression profiles obtained. At the protein level, both types of RCT are characterized by a high expression of DEFB1,88 Ksp-cadherin,89–91 KIT,92–94 CLDN8,85,95 PVALB78,88,96-98 and PNPN11/SHP2,99 which are expressed normally by distal nephron epithelium, as well as by low content of VMT.88,100
When using different array platforms and algorithms for microarray data management, only a varying degree of overlap of the differentially regulated genes was identified. After further validation of the selected candidate genes by RT–PCR, the diagnostic use of antibodies against any protein of interest is usually tested immunohistochemically using a comprehensive tissue microarray (TMA) technique on a substantial number of cases. For example, in a recent study, the selective overexpression of AQP6 in renal oncocytoma versus other types of RCT was initially shown by microarray-based transcriptomic profiling using pooled and individual samples, and was further confirmed by RT–PCR and western blot analysis at the gene and protein level, respectively.87 Nevertheless, subsequent immunohistochemical staining showed a positivity in 77% (43/56) of renal oncocytomas as well as in 16% (7/43) of chromophobe RCC (Fig. 4). To date, several immunohistochemical markers have been proposed for the differential diagnosis of renal oncocytomas, however, with controversial results. Initially obtained positive staining in renal oncocytomas for antibody against MST1R/RON101 was not in accordance with the data obtained by the subsequent investigators.102 Of interest, positivity to ANKRD2/ARPP has been recently reported in 12 of 14 oncocytomas, compared with negative staining in all 11 chromophobe RCC cases.103 However, the clinical diagnostic usefulness of this marker for renal oncocytoma needs to be further validated by additional large-scale studies, as well as improving suggested negative staining for anti-CD74 antibody.78 The exact segment of renal epithelium from which renal oncocytomas are derived or are differentiating toward is still a matter of speculation. Recently, it was also suggested that high endogeneous avidin-biotin binding activity (EABA) might be used to differentiate renal oncocytoma from chromophobe RCCs. The staining results obtained using TMA slides and automated cellular imaging system104 showed the high EABA positivity in 97% of oncocytomas, whereas it was negative in 94% of chromophobe RCCs.
In summary, many published articles have emphasized the differential diagnosis of RCT; however, it is clear that there is no perfect expression marker that reliably separates renal oncocytoma from eosinophilic variants of RCT, including chromophobe RCC. On the whole, to use an immunohistochemical marker effectively in the diagnosis of any tumor, there should be absolute certainly that the cases to be used in the study are true examples of the tumor, because there is considerable variation in interobserver interpretation of the given tumor having eosinophilic cytoplasm.
Prognosis and predictive factors
Renal oncocytoma is regarded as a benign neoplasia. Although some malignant oncocytomas with local organ invasion and metastases have been described in early reports,20,8,105,4 those cases examined could be confused with eosinophilic variants of chromophobe RCC, because Hale's colloidal iron stain and electron microscopy were not applied.106 However, despite the strict histological criteria, two oncocytomas that caused liver metastasis confirmed by needle biopsy or metastatic (liver and bone) death have been reported.6 Unfortunately, no genetic material from such cases was available for analysis to confirm the original histological diagnosis.
To Professor Gyula Kovacs for his helpful discussions.