Main

While chromophobe renal cell carcinoma was first recognized as a distinct type of renal cell carcinoma in 19851 and was given its name because the cytoplasm of the cells was pale, in 1988 Thoenes et al2 recognized that it has an eosinophilic variant. Cells with eosinophilic cytoplasm commonly are present in classic chromophobe renal cell carcinomas and the designation of the eosinophilic variant has been reserved for tumors composed either exclusively or predominantly of cells with eosinophilic cytoplasm. In sections stained with hematoxylin and eosin, these tumors closely resemble renal oncocytomas, posing a diagnostic challenge.

It has been known since 1992 that chromophobe renal cell carcinomas often have multiple losses of whole chromosomes, most frequently Y, 1, 2, 6, 10, 13, 17, and 21.3 These genetic losses have been found by classical cytogenetic methods,3, 4, 5, 6 comparative genomic hybridization,7, 8 and restriction fragment length polymorphism analysis.9 The losses are sometimes so substantial that they can be detected by DNA flow cytometry.10, 11, 12 Fluorescence in situ hybridization methods have been applied to only a few chromophobe renal cell carcinomas. Iqbal et al13 used noncentromeric probes on touch imprint smears from six chromophobe renal cell carcinomas and found variable percentages of nuclei giving single signals. The chromophobe renal cell carcinomas studied by fluorescence in situ hybridization in other papers gave mixed results.14, 15, 16 Renal oncocytomas have two common patterns of genetic abnormality: rearrangements involving 11q 1317, 18 and loss of chromosomes Y and 1.19, 20, 21 Morphologic similarities and these common losses of chromosomes have prompted some authors to suggest that there is a close relationship between chromophobe renal cell carcinoma and renal oncocytoma, particularly in cases in which there are large numbers of tumors and in the Birt–Hogg–Dubé syndrome.22, 23, 24 However, no genetic study has compared eosinophilic chromophobe renal cell carcinoma with typical chromophobe renal cell carcinoma, or with renal oncocytoma. The present study was undertaken to make this comparison and to test the hypothesis that the eosinophilic variant of chromophobe renal cell carcinoma is an entity linking classic chromophobe renal cell carcinoma with renal oncocytoma.

Materials and methods

Tissue Samples

Paraffin blocks from nine solitary sporadic eosinophilic chromophobe renal cell carcinomas, 10 solitary sporadic classic chromophobe renal cell carcinomas, and 10 solitary sporadic oncocytomas were selected from the files of the Department of Pathology of the University of Verona, Verona, Italy, the Department of Pathology and Molecular Medicine, Wellington Medical School, Wellington, New Zealand, and the consultation files of one of the authors (JNE). All of the tumors were in nephrectomy specimens. All available slides were reviewed by four of the authors (MB, BD, JNE, GM). Classic chromophobe renal cell carcinomas had the appearance typical of chromophobe renal cell carcinoma in sections stained with hematoxylin and eosin and had a positive reaction with the Hale's colloidal iron stain (Figure 1a and b). Eosinophilic chromophobe renal cell carcinomas were defined as tumors that had diffuse positive reactions with the Hale's colloidal iron stain and that, in the opinions of the authors, could be mistaken for renal oncocytoma in sections stained with hematoxylin and eosin (Figure 1c and d). Renal oncocytomas had the appearance typical of renal oncocytoma and negative reactions with Hale's colloidal iron stain (Figure 1e and f). The examples of classic chromophobe renal cell carcinoma and renal oncocytoma were selected to be typical examples of those entities. The eosinophilic chromophobe renal cell carcinomas were drawn from a large series of chromophobe renal cell carcinomas, of which they made up approximately 10%.25

Figure 1
figure 1

(a) Classic chromophobe renal cell carcinoma, (b) Classic chromophobe renal cell carcinoma with diffusely positive cytoplasmic reaction with Hale's colloidal iron stain. (c) Eosinophilic chromophobe renal cell carcinoma. (d) Eosinophilic chromophobe renal cell carcinoma with diffusely positive cytoplasmic reaction with Hale's colloidal iron stain. (e) Oncocytoma. (f) Oncocytoma with diffusely negative cytoplasmic reaction with Hale's colloidal iron stain.

Full size image

Fluorescence In Situ Hybridization

Chromosomes 1, 2, 6, 10, and 17 were selected for this study because they are the chromosomes most frequently found to be lost in previous studies of chromophobe renal cell carcinoma and because of the availability of centromeric probes specific for each of them. The Y chromosome was not used because of its absence from tumors arising in women and because its normal copy number of 1 introduces different analytical considerations from the others. As previously described,26, 27 five μm sections were cut from paraffin-embedded blocks of tissue. The paraffin was removed from the sections with two 10-min washes in xylene. After hydrating in 100, 85, and 70% ethanol solutions (10 min), rinsing in distilled water (10 min), and twice in phosphate buffer solution (pH 7, 10 min each), the slides were fixed in methanol–acetic acid 3:1 for 10 min and air dried. Next, the sections were treated in a 2 × standard saline citrate solution for 15 min at 37°C, and then dehydrated in consecutive 70, 85, and 100% ethanol solutions for 1 min each and then dried. Next, the sections were bathed in 0.1 mM citric acid (pH 6) solution at 85°C for 1 h. Then they were again dehydrated in a series of ethanol solutions and dried. The tissue was digested by applying 0.75 ml of pepsin (Sigma, St Louis, MO, USA) solution (4 mg/ml in 0.9% NaCl, pH 1.5) to each slide and incubating the slides in a humidified box for 30 min at 37°C. Next, the slides were rinsed with distilled water for a few seconds, dehydrated again in graded ethanol solutions and dried. Centromeric probes for chromosomes 1, 2, 6, 10, and 17 (Vysis, Downers Grove, IL, USA) were used. Each probe was diluted 1:50 in tDenHyb1 buffer (Insitus, Albuquerque, NM, USA). In all, 10 μl of diluted probe was applied to each slide and cover slips were placed over the slides. Denaturation was achieved by incubating the slides at 80°C for 10 min in a humidified box; then hybridization was carried out at 37°C for 3 h. The coverslips were then removed and the slides were immersed at room temperature in 0.5 × SSC for 2 min, in 50% formamide/1 × SSC for 5 min, and in 2 × SSC for 2 min. The slides were air dried and counterstained with 10 l DAPI/Antifade (DAPI in Fluorguard, 0.5 g/ml, Insitus, Albuquerque, NM, USA).

The slides were examined using an Olympus IX–50 microscope (Olympus, Tokyo, Japan) with the following filters: blue filter cube (Arcturus, Mountain View, CA, USA) for SpectrumGreen (centromeric 6, 10, and 17) Vysis probes, SpectrumOrange Filter (Chroma, Brattleboro, VT, USA) for SpectrumOrange (centromeric 1 and 2) Vysis probes, and the UV Filter (Chroma, Brattleboro, VT, USA) for the DAPI nuclear counterstain. The signals were recorded with a CCD camera.

In Situ Hybridization Analysis

The criteria for evaluating the fluorescence in situ hybridization signals were adapted from Hopman et al.28, 29 From 100 to 200 nuclei were scored for alpha-satellite signals observed with the fluorescence microscope at × 400 magnification. As much as possible, signals from solitary nuclei were counted; however, groups of two or three adjacent (but not overlapping) nuclei were occasionally included in the counts. Nuclei were counted when the entire nuclear circumference had a round-to-oval contour and showed no evidence of fragmentation. Two signals of the same size in close proximity, not connected by a link, were counted as two signals. A diffuse signal was regarded as a signal if it was contiguous and within an acceptable boundary. Two small signals connected by a visible link were counted as one signal. Overlapping nuclei and nuclei with uncertain signals were not counted. There was no significant variation in hybridization efficiency when different areas of the slides were examined. The number of signals visualized in the nuclei were tabulated from randomly selected areas of the slides in which nuclear overlap was minimal.

Statistical Analysis

Fisher's exact test was used to compare the proportion of chromosomal loss among the different types of tumors. Pairwise comparison was made using the Sidak adjustment method to control the overall significance level of the tests.

Results

To determine the performance of the system and establish the distribution of signals in nuclei from normal tissue, non-neoplastic kidney tissues from 25 of the blocks containing tumors were studied. This approach has the advantages that section thickness and tissue processing parameters were the same for the normal samples as for the tumors. The results are presented in Table 1. The data from the normal tissue are quite uniform. For chromosomes 1, 2, and 6 the mean percentage of nuclei with two signals was 83% and for chromosomes 10 and 17, it was 85 and 84%, respectively. The standard deviations of the mean numbers of nuclei with two signals for chromosomes 1, 2, 6, 10, and 17 were 3.39, 3.12, 3.71, 2.94, and 3.91, respectively. For chromosomes 1, 6, and 10, the mean percentage of nuclei with single signals was 12% and for chromosomes 2 and 17, it was 13%. The standard deviations of the mean numbers of nuclei with single signals for chromosomes 1, 2, 6, 10, and 17 were 3.52, 3.51, 2.51, 3.48, and 4.43, respectively. In the tumors, chromosomal loss was defined as a percentage of nuclei with single signals greater than the normal tissue mean for that chromosome+4 times the normal tissue standard deviation for that chromosome. Thus, for chromosomes 1, 2, 6, 10, and 17, percentages of single signals greater than 26, 27, 22, 26, and 30%, respectively, were considered to indicate chromosomal loss.

Table 1 Percentages of different signal numbers in nuclei from normal tissue
Full size table

The results from the tumors are presented in Tables 2 and 3. Inspection of the data in Table 1 shows that in the carcinomas when chromosomal losses are detected, they are almost always striking, with single signals detected in 75% or more of nuclei. Lesser losses, in the range of 45–60% of nuclei with single signals, were observed only 4 times and only with chromosome 17. In the oncocytomas, a few tumors had single signal counts of 20% or more; however, the oncocytoma with loss of chromosome 1 had as striking a loss as the carcinomas: 75% single signals. Losses are summarized in Table 3. Among the 10 classic chromophobe renal cell carcinomas, seven had loss of chromosome 1, nine had loss of chromosome 2 (Figure 2a), eight had loss of chromosome 6, six had loss of chromosome 10, and nine had loss of chromosome 17. Five of the 10 tumors had loss of all five of the chromosomes and one classic chromophobe renal cell carcinoma had no loss of any of the chromosomes. Among the nine eosinophilic chromophobe renal cell carcinomas, six had loss of chromosome 1; five had loss of chromosome 2; five had loss of chromosome 6; four had loss of chromosome 10; and seven had loss of chromosome 17. Four of the nine eosinophilic chromophobe renal cell carcinomas had loss of all five of the chromosomes and one had no loss of any of the chromosomes. All but one of the renal oncocytomas showed two signals for all of the probes (Figure 2b); the exception had loss of chromosome 1.

Table 2 Percentages of nuclei with different numbers of signals from neoplastic cells
Full size table
Table 3 Summary of losses of chromosomes 1, 2, 6, 10, and 17 in chromophobe renal C
Full size table
Figure 2
figure 2

(a) Classic chromophobe renal cell carcinoma showing nuclei with single fluorescent signals for chromosome 2. (b) Oncocytoma showing two nuclei with two fluorescent signals for chromosome 2.

Full size image

No significant difference in the proportion of tumors with chromosomal losses was found between classic and eosinophilic chromophobe renal cell carcinomas for chromosome 1 (P=1.000), chromosome 2 (P=0.3659), chromosome 6 (P=0.7252), chromosome 10 (P=0.9594), or chromosome 17 (P=0.9270).

Discussion

These data show that the eosinophilic variant of chromophobe renal cell carcinoma and classic chromophobe renal cell carcinoma have similarly highly frequent losses of chromosomes 1, 2, 6, 10, and 17 and that renal oncocytomas do not have these genetic abnormalities. Some studies have suggested, based upon morphology and some tumors with hybrid appearances in settings of oncocytomatosis and Birt–Hogg–Dubé syndrome, that chromophobe renal cell carcinoma and renal oncocytoma may be closely related.22, 23, 24 If this is so, the connection lies elsewhere than in the most frequent chromosomal losses of chromophobe renal cell carcinoma and the data do not support the hypothesis that the eosinophilic variant of chromophobe renal cell carcinoma is a transition form between classic chromophobe renal cell carcinoma and renal oncocytoma.

Initially, a variety of genetic abnormalities, such as an interstitial deletion of the short arm of chromosome 11, trisomies of chromosomes 7, 12, 16, 18, and 20 and alteration of chromosome 11q, were reported in chromophobe renal cell carcinomas.19, 30 Schwerdtle et al9 studied 11 chromophobe renal cell carcinomas and found a high frequency of loss of heterozygosity (73–91%) at 1p, 2p, 6p, 13p, 17p, and 21q with a panel of nine loci and also a high frequency (63%) of loss at chromosome 14q that has not been confirmed by other studies. Speicher et al7 scanned metaphase spreads of 19 chromophobe renal cell carcinoma using comparative genomic hybridization analysis and found that losses of multiple chromosomes from among chromosomes 1, 2, 6, 10, 13, 17, and 21 is a genetic pattern occurring in most chromophobe renal cell carcinomas. These findings have been confirmed by karyotyping studies. Bugert et al31 analyzed 42 chromophobe renal cell carcinomas with a set of two to three microsatellite markers for each of chromosomes 1, 2, 6, 10, 13, 17, and 21 and found loss of chromosomes in up to 95% of tumors. In all but one tumor, losses of multiple chromosomes were present. Cytogenetic analysis also has shown variable losses of chromosomes in chromophobe renal cell carcinomas.3, 4, 5, 6 A DNA cytometric analysis of 43 chromophobe renal cell carcinomas revealed an aneuploid DNA index between 1.1 and 2.1 in 77% of the cases.11 An aneuploid DNA index was found in three of five chromophobe renal cell carcinomas in another study.10 Akhtar et al32 found a hypodiploid DNA index varying between 0.70 and 0.90 in 24 chromophobe renal cell carcinomas. These studies did not distinguish between results from classic chromophobe renal cell carcinomas and eosinophilic ones. Iqbal et al13 recognized the two types but it is unclear whether eosinophilic chromophobe renal cell carcinomas were included in the study. Our results are in agreement with previous data for chromophobe renal cell carcinomas in general and for the first time demonstrate that both the classic and eosinophilic variants of chromophobe renal cell carcinoma have the same frequent losses of chromosomes 1, 2, 6, 10, and 17.

Chromophobe renal cell carcinomas with normal complements of chromosomes are in the minority. We found two examples among 19 tumors (one classic and one eosinophilic). This is similar to what others have reported: Speicher et al7 found 2/19 cases, Bugert et al31 1/42, Iqbal et al13 2/8, and Verdorfer et al16 1/3. Four of the carcinomas had roughly equal populations of nuclei with single and double signals for chromosome 17.

We found heterogeneous populations with respect to chromosome 17 in four of the carcinomas. Heterogeneous populations in tumors are well recognized; in 1988, Hopman et al28 reported that an apparently homogeneous tumor cell population, at least with respect to DNA content, can contain fractions of cells containing chromosome numbers that deviate from the expected number. Some DNA cytometric and cytogenetic studies have observed mosaicism in chromophobe renal cell carcinomas4, 6, 33, 34, 35 Gunawan et al6 concluded that endoreduplication resulting in triploid-to-hypodiploid chromosome numbers occurs in chromophobe renal cell carcinoma.

All but one of the renal oncocytomas showed normal numbers of all the chromosomes studied and the exception showed loss only of chromosome 1. Loss of chromosome 1 in renal oncocytomas is infrequent, but has been reported before.20, 36, 37 We observed loss of chromosome 1 at a lower frequency (10%) than was described by Polascik et al38 (57% LOH at 1p and 30% LOH at 1q on 13 cases), Crotty et al19 (33%, karyotyping study on six cases), Presti et al37 (45%, CGH study on 13 cases), Thrash-Bingham et al39 (75%, LOH at 1p on four cases), and Brown et al36 (50%, FISH study on 20 cases), but exactly the same as Herbers et al40 (10%, LOH with two markers on 1p/1q in a study performed on 41 cases). The differences may be attributable to our use of centromeric probes to detect losses of whole chromosomes, while some of the other studies used methods that detect partial losses of chromosomes.

In summary, we found with interphase cytogenetics that eosinophilic and classic chromophobe renal cell carcinomas have similarly frequent losses of chromosomes 1, 2, 6, 10, and 17 and renal oncocytomas usually show normal complements of these chromosomes, with occasional loss of chromosome 1. This does not lend support to the hypothesis that chromophobe renal cell carcinoma and renal oncocytoma are closely related, or that eosinophilic chromophobe renal cell carcinoma is a transition form between chromophobe renal cell carcinoma and renal oncocytoma. In the differential diagnostic context of renal oncocytoma vs eosinophilic chromophobe renal cell carcinoma, loss of chromosomes 2, 6, 10, or 17 is a helpful marker of eosinophilic chromophobe renal cell carcinoma, supplementing Hale's colloidal iron stain, which some laboratories find difficult to perform. Unfortunately, the method is less to be helpful with needle biopsies because they may not afford the number of sections needed or may not contain enough nuclei for the FISH analysis.