Abstract
Background/Aim: It has been suggested that eosinophilic variant of chromophobe renal cell carcinoma (chRCC) with low chromosome number or lack of genomic alteration has an excellent prognosis in comparison to classic chRCC. The aim of our study was to analyse the phenotypical variations of 77 chRCCs, including 7 eosinophilic ones, each diagnosed unequivocally by genetic means. Materials and Methods: DNA isolated from chRCCs was subjected to array comparative genomic hybridisation (CGH) for establishing the chromosome alteration. Original histological slides were evaluated for cellular phenotype and growth pattern and compared to the genetic alterations. Results: Loss of the entire chromosome 1, 2, 6, 10, 13, 17 and 21 occurred in 95%, 94%, 86% 90% 82% 90% and 66% of the cases, respectively. The number of chromosome alterations in eosinophilic forms of chRCC corresponded to those found in classic chRCC with pale-reticular cytoplasm or mixed cellular characteristics. Three of seven eosinophilic variants with loss of 4, 10 and 11 chromosomes showed metastasis at the time of diagnosis whereas only 3 metastatic tumors were noticed among the 70 classic chRCC. We did not find discriminating difference in number of chromosome alteration between classic and eosinophilic forms of chRCC. Conclusion: Eosinophilic chRCC has a more aggressive biology than the classic form. To avoid diagnostic pitfall of eosinophilic renal cell tumors with uncertain diagnosis, a genetic analysis should be carried out.
- Chromophobe renal cell carcinoma
- eosinophilic variant
- array comparative genomic hybridisation
- chromosome alteration
In 1974, based on histological and electron-microscopic characteristics, Bannasch and colleagues (1) established a classification of nitrosomorpholine-induced experimental rat tumors and described for the first time chromophobe renal cell carcinoma (chRCC). In 1985, Thoenes et al. adapted this classification and recognized chRCC among human kidney cancers (2). The typical chRCC is composed of cells with voluminous pale reticular cytoplasm of diagnostic importance (3). Later, Thoenes and colleagues described an “eosinophilic variant” of chRCC and noticed immediately that kidney cancers displaying eosinophilic cells may mimic renal oncocytoma (RO), eosinophilic chRCC or eosinophilic “clear” cell RCC (4). It is well known that kidney cancers display heterogeneous cytoplasmic and growth pattern characteristics, even within the same tumor (2, 4).
In 1993, there was a paradigm change in the renal cell tumor classification. The four main renal cancer entities, conventional, papillary and chromophobe RCC and RO, were unequivocally characterized by robust genetic alterations (5, 6). The loss of chromosomes 1, 2, 6, 10, 13, 17 and 21 was found to be highly characteristic of chRCC (7-10). In RO, loss of chromosome 1, or 14q or balanced translocation involving 11q13 region, random chromosomal changes or no changes have been described (11, 12). Based on the specific chromosomal alterations it became possible to differentiate chRCC including the “eosinophilic variant” from RO with overlapping phenotype (13-15).
During the last years, using the cellular characteristics for case selection, low chromosome number or no chromosomal losses has been described in eosinophilic chRCC (16, 17). Moreover, a new tumor type or variant of chRCC and/or RO, a “hybrid oncocytic-chromophobe tumors” (HOCT) has been described (18). Herein, the aim was to analyse the phenotypic variations of chRCCs diagnosed by genetic means to obtain information on the correlation of phenotype and genotype. Therefore, we selected genetically characterized chRCCs, according to the original classification system. Subsequently, we evaluated the histological variations, cellular and growth pattern diversity and compared to the number of genetic changes.
Materials and Methods
Tissue samples. Tumors were collected from patients who undervent a nephrectomy between 1994 and 2009 at the Department of Urology, Medical Faculty, Ruprecht-Karls-University, Heidelberg, Germany. We have included cases obtained for consultation from different countries to establish the correct diagnosis by genetic means. The histological diagnosis was made according to the Heidelberg classification system and morphological variations were recognised by a pathologist (GK) (19).
DNA extraction. The areas of interest were marked on Hematoxilin-Eosin (H-E)-stained slides, which were manually dissected or scraped from 20-μm-thick paraffin sections. The tissue fragments were placed into a 1.5 ml tube and washed with 1 ml of xylene, to ensure that paraffin was removed. After rehydration the tissue samples were briefly air-dried. DNA was extracted using the Dneasy Blood and Tissue kit (#69504, Qiagen, Hilden, Germany) according to the manufacturer’s protocol with 70% ethanol replacing the AW2 washing step. If it was necessary, a fresh aliquot of proteinase K was added until complete tissue digestion. Finally, DNA was eluted with nuclease-free water and evaluated on NanoDrop®ND-1000 spectrophotometer (Thermo Fisher, Budapest, Hungary). All samples used in this study consistently had a A260/A280 ratio of more than 1.8. To visualise size distribution, each DNA sample was analysed on 1.5% agarose gels with ethidium bromide and documented by AlphaEasy®FC software (version 5.0.1, Alpha Innotech Corporation, San Leandro, CA, USA). DNA from a pool of individuals was used as a sex-matched reference.
Array-comparative genomic hybridization (aCGH) and data analysis. DNA was labeled using the Universal Linkage System (ULS) and hybridized to 4×44K HG-CGH arrays (Amadid 014950, Agilent Technologies Deutschland GmbH, Böblingen, Germany) at the Genomics Core Facility, EMBL (Heidelberg, Germany) according to the protocol provided by the company. Briefly, about 500 ng of tumor and normal reference DNA was chemically labelled with Oligo aCGH Labelling Kit (#5190-0419, Agilent) using ULS-Cy5 and ULS-Cy3 dye incorporation, respectively. The differentially labelled samples were purified using Agilent-KREApure columns (#5190-0419) and the labelling efficiency was estimated by NanoDrop®ND-1000 spectrophotometer. Afterwards, tumor and reference samples were combined and hybridized at 65°C for 40 hours in a slide hybridisation chamber with a rotation at 20 rpm. The arrays were washed, air-dried and scanned immediately on the Agilent DNA Microarray Scanner using the Agilent Scanner Control software (version 7.0).
Data were further extracted from the scanned microarray image (.tif), filtered and normalized by the Agilent Feature Extraction software (version 9.5) annotated against NCBI build 30 (hg18, March 2006). Raw copy number ratio data were transferred to the Agilent CGH Analytics software (short-time free trial version 3.4) for further analysis. The ADM-2 algorithm and the sensitivity threshold at 6.0 were used to identify DNA copy number anomalies at the probe level. A copy number gain was defined as a log2 ratio >0.25 and a copy number loss was defined as a log2 ratio <-0.5.
Results
DNA alterations. Loss of entire chromosome 1, 2, 10 and 17 was the most frequently observed genomic alteration noticed in 95%, 94%, 90% an 90% of the 77 chRCCs, respectively. The loss of other chRCC-specific chromosomes such as chromosomes 6, 13 and 21 occurred in 86%, 82% and 66% of the cases, respectively (Figure 1). In addition, we found a loss of chromosome 3 in 21% and chromosome 9 in 26% of the tumors. To show the differences of genetic changes in chRCC and RO, we included the chromosomal alterations observed earlier in 42 RO (14). The loss of chromosome 1 was detected in 23% of RO and loss of chromosome 14 in 5% of the cases. Additionally, random loss of chromosome 3, 6, 8, 9, 11, 18, and 22 was seen each in one or two cases. We do not have any information on the frequency of balanced translocation of chromosome 11q13 because translocation cannot be detected by array technology. The data presented in Figure 1 clearly demonstrate the differences in genetic alterations between chRCC and RO.
Taking into account the number of chromosomal losses in each individual chRCC, loss of 3 to 12 chromosomes leading to low chromosome number between 34 and 43 has been detected. Loss of 6, 7 or 8 chromosomes were observed in 44 of 77 cases. Only 2 and 3 chRCCs displayed loss of 3 or 4 of the specific chromosomes, respectively. More importantly, the 7 cases with eosinophilic cytoplasm displayed monosomy of 4, 6, 10, and 11 chromosomes in 1, 2, 3 and 1 cases, respectively. The variation in number of chromosomal alteration detected in eosinophilic chRCC corresponded to those found in classic chRCC with pale-reticular cytoplasm or mixed cellular characteristics.
Morphological heterogeneity. We analyzed the original slides of 77 chRCCs used for histological diagnosis and distinguished 4 cytological variations. Only 4 tumors composed exclusively of large chromophobe cells with reticular cytoplasm (Figure 2A). A mixture of medium sized eosinophilic cells and large cells with pale or reticular cytoplasm was seen in 39 chRCC cases (Figure 2B). In these tumors the chromophobe cells lined a vascular stroma whereas the eosinophilic cells were located in the centrum of large epithelial sheets. Medium sized tumor cells with pale eosinophilic cytoplasm, discrete perinuclear halo and prominent cell membrane were seen in 27 chRCC cases (Figure 2C). Finally, we identified 7 chRCC samples composed exclusively of eosinophilic cells (Figure 2D). The growth pattern of tumors varied from solid nested towards tubular, micro-cystic or comedo-like forms.
Metastasis at the time of operation was observed in 6 out of 77 chRCCs. Three of metastatic tumors displayed exclusively eosinophilic cytoplasm. One of them had multiplex liver metastases and showed high mitotic activity (Figure 2E). Another case, which was published earlier as a metastatic RO (20), developed extensive skeletal metastases (Figure 2F). The third tumor displayed sarcomatous change leading to metastatic tumor growth (Figure 2G, H). Thus, 3 of the 7 eosinophilic chRCCs showed an aggressive growth resulting in metastasis whereas only 3 of the 70 classic chRCCs displayed metastasis and 2 other cases showed local invasive growth at the time of diagnosis.
Discussion
We describe here the genetic alterations of 77 chRCCs with variable cytological and growth characteristics. The vast majority of chRCCs were classic chRCCs, but 7 of the 77 chRCCs were composed exclusively of eosinophilic cells. In spite of cytological and growth pattern heterogeneity, the genetic analysis improved in all cases the diagnosis of chRCC. The usefulness of genetic analysis in diagnosis of chRCC has already been demonstrated by comprehensive studies applying microsatellites, aCGH or FISH (9, 10, 14, 15).
Only few genetic analyses have been carried out on eosinophilic variant of chRCC. Brunelli and colleagues found nearly identical alterations of chromosomes 1, 2, 6, 10 and 17 in 10 classic and 9 eosinophilic chRCC by applying FISH on histological slides (15). However, in one of the classic variant and at least in two of the eosinophilic variant no monosomy of the five chromosomes has been found. In a recent study of 66 chRCCs, no chromosomal losses were found in 7 of 13 “eosinophilic” variant chRCCs (16). Another study analyzed the copy number variations in 33 tumors diagnosed as chRCC (17). One of the classic chRCCs and three of the eosinophilic tumors showed no copy number losses, a characteristic of chRCC. Among the 99 tumor samples analyzed in these two studies (16, 17), 10 tumor samples with the only loss of chromosome 1 or without any alterations were diagnosed as “eosinophilic” chRCC. Therefore, it was suggested that “eosinophilic” chRCC displays substantially less or no any chromosomal losses, which would justify the separation of two genetically and phenotypically distinct chRCC subtypes (17). In the present study we did not find any association between the number of chromosomal losses and the cytological characteristics of chRCC.
Ohashi and colleagues also suggested that “eosinophilic” chRCC has an excellent prognosis in comparison to the classic form (17). However, in our series 43% of “eosinophilic” chRCCs had already metastasis whereas only 4% of the classic chRCCs showed metastatic growth at the time of diagnosis. Our findings suggest that eosinophilic variant of chRCC diagnosed by genetic analysis, has no indolent clinical behavior but rather more aggressive growth characteristics than the classical chRCC. Taking into account the results of genetic analysis presented in the two cohorts of the previous studies (16, 17), we can hypothesize that several RO cases misdiagnosed as eosinophilic chRCC were included. As Thoeness and colleagues (3) have already noticed 30 years ago, “eosinophilic variant” of chRCC can easily be confused with RO. Electron microscopic studies showed that RO and eosinophilic chRCC are characterized by a high number of mitochondria whereas the classic chRCC display a high number of cytoplasmic vesicles (3). The transition between the two cell types is common and may lead to diagnostic pitfall.
In the last years, so-called “hybrid oncocytic-chromophobe tumor” (HOCT) has been described in patients with Birth-Hogg-Dube (BHD) syndrome and also in the general population (18). In two recent genetic studies none of the 27 “HOCT” cases showed chromosomal losses that characterize chRCC (21, 22). In the third study 2 of 14 HOCT displayed loss of chromosomes 1, 2, 6, 10, 13 and 21 indicating that these cases correspond to chRCC (23). Taking into account the excellent outcome of patients with HOCT reported in the literature (24), and the results of the above mentioned three genetic analyses, the vast majority of HOCT cases should be considered as RO and only some of them as chRCCs. From a genetic point of view, there is no explanation of how can a “hybrid” tumor display chromophobe cells with loss of chromosome 1, 2, 6, 10, 13, 17 and 21 and RO cells with the only loss of chromosome 1or no any changes.
One limitation of this study is the small number of eosinophilic variant of chRCCs cases. Moreover, no follow-up data were available and the biological behaviour of chRCC cases was estimated at the time of diagnosis. However, even in this setting, it was clear that the eosinophilic variant of chRCC has no excellent prognosis as suggested by others and therefore should be analysed by genetic means to differentiate from the benign RO.
In conclusion, loss of chromosome 1, 2, 6, 10, 13, 17 and 21 unequivocally marks chRCCs, irrespective of reticular, mixed or eosinophilic cellular characteristics. Eosinophilic chRCCs diagnosed by genetic analysis showed a more aggressive biology than “classic” chRCCs. Therefore, we suggest that in cases with uncertain diagnosis of benign RO versus malignant chRCC, a genetic analysis should be carried out.
Acknowledgements
The Authors thank all pathologists who sent paraffin blocks from cases with uncertain diagnosis for genetic analysis. The part of genetic data used in this study was published earlier by Yusenko et al. (14). This work was supported by a Grant of the Medical Faculty, University of Pecs, Hungary (PTE-AOK-KA-2018/41). The Authors thank Ms. Zsuzsanna Halas and Barbara Kanyo for selecting paraffin blocks and preparing slides for immunohistochemistry.
Footnotes
Authors’ Contributions
AM and GK designed the research study, MVY carried out the aCGH and analyzed the data, GK and AM evaluated the histology, AM and DB wrote the manuscript and GK reviewed the manuscript. All Authors read and approved the final version of the manuscript.
Conflicts of Interest
The Authors have no conflicts of interest to declare.
- Received August 8, 2020.
- Revision received October 9, 2020.
- Accepted October 15, 2020.
- Copyright © 2020 International Institute of Anticancer Research (Dr. George J. Delinasios), All rights reserved.