Cytogenetic analysis of 130 renal oncocytomas identify three distinct and mutually exclusive diagnostic classes of chromosome aberrations

The cytogenetic alterations in renal oncocytoma (RO) are poorly understood. We analyzed 130 consecutive RO for karyotypic alterations. Clonal chromosome abnormalities were identified in 63 (49%) cases, which could be categorized into three classes of mutually exclusive cytogenetic categories. Class 1 (N = 20) RO had diploid karyotypes with characteristic 11q13 rearrangement in balanced translocations with 10 or more different chromosome partners in all cases. We identified recurrent translocation partners at 5q35, 6p21, 9p24, 11p13‐14, and 11q23, and confirmed that CCND1 gene rearrangement at 11q13 utilizing fluorescence in situ hybridization (FISH). Class 2 RO (N = 25) exhibited hypodiploid karyotypes with loss of chromosome 1 and/or losses of Y in males and X in females in all cases. The class 3 tumors comprising of 18 cases showed diverse types of abnormalities with the involvement of two or more chromosomes exclusive of abnormalities seen in classes 1 and 2 tumors. Furthermore, karyotypically uninformative cases were subjected to FISH analysis to identify classes 1 and 2 abnormalities. In this group, we found similar frequencies of CCND1 rearrangement, loss of chromosome 1 or Y as with karyotypically abnormal cases. We validated our results against 91 tumors from the Mitelman database. Correlation of clinical data with all the three classes of ROs showed no clear evidence of overall patient survival. Our findings support the hypothesis that RO exhibit three principal cytogenetic categories, which may have different roles in initiation and/or progression. These cytogenetic markers provide a key tool in the diagnostic evaluation of RO.

Clinically, RO are benign tumors typically occur as solitary lesions. Histological features of RO are usually distinctive but can overlap chromophobe renal cell tumors. 3 Because chromophobe renal cell carcinoma (RCC) is a malignant tumor, differentiating RO from chromophobe is clinically important.
Only about 100 cases of RO with conventional karyotype have been reported in the literature (https://cgap.nci.nih.gov/Chromosomes/ CytList). 4 Most of these studies were either single case reports or small case series, which identified loss of chromosomes 1, Y, and 14, and 11q13 rearrangements as recurrent chromosome aberrations. However, the frequency of recurrent cytogenetic changes and their relation with each other has not been systematically examined.
We sought to identify recurrent chromosomal changes in RO by analyzing 63 karyotypically abnormal cases. We identified three mutually exclusive karyotype classes of ROs. Frequency of chromosome changes in the order of their occurrence was 11q13 rearrangement, loss of chromosome 1, loss of Y in males and loss of X in females, trisomy 7, and loss of chromosomes 14, 21, and 22. The present findings can be useful in differential diagnosis of RO and in understanding the role of genetic mechanisms in their pathogenesis.

| Patient cohort
A total of 130 consecutive RO specimens were subjected to karyotype analysis for diagnostic purposes at Columbia University Medical Center, New York, between 1999 and 2016. All patients had a radical or partial nephrectomy for a renal mass. We performed chart review to obtain clinical and histologic information on all the patients.

| Karyotype and fluorescence in situ hybridization analyses
We routinely perform karyotype and fluorescence in situ hybridization (FISH) analyses of all surgically removed renal masses at our institution.
Touch imprints were made and fixed immediately in methanol-acetic acid in majority of cases for FISH testing. Chromosome analysis was performed using standard methods. Briefly, fresh tumor specimens collected after surgery were washed with antimycotic solution (0.5 μg/mL Amphotericin B) and antibiotics (200 IU/mL penicillin and 200 μg/mL streptomycin), minced with scalpel blade into fine pieces, and digested 2-12 hours with collagenase (200 U/mL) in complete Dulbecco's Modified Eagle's Medium (DMEM medium). Dissociated cells were washed twice and then cultured in complete DMEM medium supplemented with epidermal growth factor, insulin-transferrin-sodium selenate (Invitrogen).
The cultures were monitored daily and appropriately confluent cells were subjected to metaphase preparations using standard methods after addition of colcemid for 12 hours. Metaphase preparations were subjected to Giemsa (GTG) banding and analyzed by standard methods. Karyotypes were described by the Standard Cytogenetic Nomenclature. FISH was performed either on touch imprints or cells processed for karyotype analysis using CCND1/CEP 11, CCND1 break apart, chromosome enumeration probes CEP X, CEP Y, CEP 1, and CEP 7 (Vysis, Downers Grove, IL) by standard methods. The cyclin D1 locus at 11q13 is labeled in SpectrumOrange fluorochrome in CCND1/CEP 11 cocktail spanning a 378-kb physical distance equally covering on both 5 0 and 3 0 ends of the gene. When 11q13 rearrangement occurs, there is a spit in orange signal while SpectrumGreen labeled chromosome 11 serve as control to enumerate chromosome number. Dual color break apart CCND1 probe consists of two probes flanked by a centromeric 700-kb probe labeled with SpectrumGreen and a telomeric 500-kb probe labeled with SpectrumOrange. This probe shows as orange-green signals together due to their close proximity in normal chromosome 11 and orange signal separated from green signal when 11q13 (CCND1) rearranged.

| RESULTS
Of the 130 tumor specimens subjected to karyotype analysis, 63 (49%) cases showed clonal chromosome abnormalities which include 11 previously reported cases. 5 Normal karyotypes were found in 30 (23%) cases and 37 (29%) were failed to yield metaphases. The clinical, histologic, and cytogenetic features of karyotypically abnormal cases are shown in Table S1. Nearly all tumors (61 of 63) exhibited near-diploid range or pseudo-diploid karyotypes in their stem-line, while one had triploid karyotype and the other had 53 chromosomes.

| Chromosome aberrations identify three mutually exclusive karyotypic classes of ROs
The types of clonal chromosome abnormalities identified were allowed us to classify three karyotypically distinct classes of ROs (Table 2, Figure 1). Class 1 tumors showed 11q13 translocation involving 10 known and two unknown chromosome partners in all 20 tumors.
Both recurrent (6p21 four cases, 5q35 and 9p24 in three cases each, and 11q23 in two cases) and nonrecurrent (1p13, 4q27, 6q13, 7q11. The only recurrent numerical abnormalities were gain of chromosome 7 and loss of sex chromosomes in five cases each (Table 2, Figure 1).
Among the structural abnormalities, 2q37 rearrangements were seen in three cases and telomeric associations in three cases. No evidence of WGD was seen in this group.

| FISH analysis of recurrent chromosome changes in classes 1 and 2 ROs
We validated the karyotypic changes of loss of chromosomes 1, Y, X, and 11q13 (CCND1) rearrangements by FISH. We tested 42 of 63 karyotypically abnormal cases by FISH based on the availability of materials. CCND1 rearrangement was found in all 15 class 1 RO cases with 11q13 translocation. None of the 10 class 2 or 3 RO lacking 11q13 translocation were positive for CCND1 rearrangement. This high concordance between karyotype and FISH suggests that karyotype is reliable in identifying 11q13 translocations. Loss of Y chromosome was confirmed by FISH in all six cases and loss of X chromosome in a single case ( Table 1). All five cases with loss of chromosome 1 were confirmed by FISH. Four cases that were karyotypic negative for loss of Y was also negative by FISH. However, FISH analysis identified

| Clinical correlations
No clinical and histological correlations were identified between various classes of ROs, except age and WGD. Ages between the RO classes were statistically significant, where class 1 ROs occur in younger patients (class 1, 55.07 years; class 2, 64.6 years; class 3, 69.5 years) (P = .0014). Frequencies of WGD between RO classes are also statistically significant with class 2 exhibiting a higher frequency (P = .01). As CCND1/11q13 translocations with multiple genomic regions occur as a sole abnormality in RO, they are likely to act as driver mutations.

| Validation of the present data with Mitelman database
Loss of chromosome 1 in association with sex chromosome loss is characteristic feature of class 2 RO. This class of RO shows a male predominance (68% males vs 32% females) with more frequent loss of Y in males (71% cases) compared to loss of X in females (38%). The biological consequence of chromosome 1 loss is not well understood.
However, no significant differences in expression of genes located on chromosome 1 have been found in a small series of RO cases reported recently. 19 Interestingly, we found two cases showing a common deletion at 1p22-p34 (one with deletion distal to 1p21 and other with an interstitial deletion at 1p22-p34  28 as well as both before and after other copy number alterations in different tumor types. 29 It has also been shown that tetraploid cells exhibit tolerance of chromosomal instability compared to diploid clones and associate with poor prognosis in colorectal tumors. 26 As ROs are benign, whether the presence of WGD fuels further genomic changes accelerating tumor progression is unknown. We observed tas involving various chromosomes in four cases (three class 3 and one class 1 tumors) of ROs. Such telomeric associations have also been identified in three previously published cases of class 3 RO. 30,31 Telomeric associations caused by end-to-end fusions of chromosome ends have been shown to be frequently associated with cancer. 32,33 Defects in telomere function was identified as the cause for the formation of dicentric chromosomes created through end-to-end fusions, 34 specifically mediated by telomeric DNA-binding protein, TRF2. 35 However, the causes and consequences of telomeric associations seen in the present study of RO remain to be explored.
In the present study, we performed extensive karyotypic and FISH analyses of large number of human ROs. These analyses identified three karyotypically distinct classes of oncocytomas. Overall, our data emphasize the importance of cytogenetic analysis in characterization and differential diagnosis of oncocytomas from chromophobe renal cancer. The identification of distinct karyotypic classes of RO can provide a basis in further characterization of these tumors in exploring potential mechanisms of oncogenesis in these tumors.