Papillary renal cell carcinoma (RCC) represents 10% to 15% of adult renal neoplasms; however, the molecular genetic events that are associated with the development and progression of sporadic papillary RCC remain largely unclear. Papillary RCCs can be divided into two subtypes based on histologic, cytogenetic, and gene expression differences. Type 1 tumors (∼60–70%) are generally low grade with favorable outcome, whereas type 2 tumors (∼30–40%) are associated with increased cytogenetic complexity, high tumor grade, and poor prognosis. In this study, computational analysis of gene expression data derived from papillary RCC revealed that a transcriptional signature indicative of MYC pathway activation is present in high-grade type 2 papillary RCC. The MYC signature is associated with amplification of chromosome 8q and overexpression of MYC that maps to chromosome 8q24. The importance of MYC activation was confirmed by both pharmacologic and short interfering RNA–mediated inhibition of active Myc signaling in a cell line model of type 2 papillary RCC. These results provide both computational and genetic evidence that activation of Myc is associated with the aggressiveness of papillary type 2 RCC. Therefore, it will be useful to consider inhibition of components of the MYC signaling pathway as avenues for therapeutic intervention in high-grade papillary RCC. [Cancer Res 2007;67(7):3171–6]

Papillary renal cell carcinoma (RCC) is histologically characterized by the tumor cells arranged in a papillary configuration with associated fibrovascular cores. Subclassification of papillary RCC into two subtypes, type 1 and type 2, is supported by histologic criteria (cell size and single layer or stratified cell arrangement) and more recently by gene expression profiling data (1, 2). Although the stratification of papillary RCC into two subtypes is in part supported by cytogenetic studies, these studies also suggest that type 1 tumors progress into type 2 tumors (3, 4). Although the specific molecular genetic events that lead to the development of sporadic type 2 papillary RCC remain largely unclear, these tumors have increased cytogenetically complexity and are associated with poor long-term patient survival as compared with type 1 tumors.

Recent studies have shown that several types of biological information, in addition to relative transcript abundance, could be derived from the high-density gene expression profiling data. Changes in DNA copy number result in dramatic changes in gene expression within the abnormal region and are detectable through examination of the population of mRNA generated from the genes that map to each chromosome (5, 6). Additionally, the activation of certain oncogenes produce context-independent gene signatures that can be detected in other gene expression profiles (79). For example, genes that are up-regulated by the overexpression of RAS in breast epithelial cells also tend to be overexpressed in other samples containing activated RAS signaling, such as lung tumors that contain activating RAS mutations (10).

In this study, parallel analysis of gene sets that correspond to chromosomal regions and gene sets that reflect oncogenic pathway activation reveals that high-grade papillary RCC is associated with frequent chromosomal abnormalities of chromosome 8q, overexpression of the MYC gene that maps to chromosome 8q24, and activation of the MYC signaling pathway. In addition, a cell line model of papillary RCC type 2 was identified based on gene expression, cytogenetic, and histologic evaluation. The cell line model was used to validate the role of Myc activation in papillary RCC. These results suggest that activation of Myc leads to the development of aggressive papillary RCC.

Gene expression analysis. Gene expression profiles from 12 nondiseased kidney samples and 35 papillary RCC tumor samples (22 type 1 and 13 type 2) were produced using the Affymetrix HG-U133 Plus 2.0 GeneChip platform as previously described (2). This data has been deposited at the Gene Expression Omnibus (GDS1344). Gene expression profiles from 31 clear cell RCC samples [18 low stage (I and II) and 13 high stage (III and IV)] were generated similar to the papillary samples. A report describing the details and initial characterization of these samples has been submitted. These kidney samples were obtained from the Cooperative Human Tissue Network. Internal review board approval was obtained from the Van Andel Research Institute to study these samples. Gene expression values were preprocessed using the RMA method as implemented in the BioConductor affy package for the R environment (11, 12). Prior to data preprocessing, probe set mappings were updated (13). Chromosomal abnormalities were identified using a previously described parametric gene set enrichment method originally termed comparative genomic microarray analysis (14, 15). For pathway analysis, MYC, E2F, SRC, RAS, CTNNB1, synergistic HGF/VEGF, VHL, and hypoxia gene signature components were obtained from the literature (10, 1618). The HGF and VEGF signatures were generated using data from the Gene Expression Omnibus (GDS406 and GDS495, respectively). In these cases, control cells were compared with treated cells after 24 h. For consistent presentation, all “up” and “down” gene lists reflect the gene signatures of mutant cells verses the nearest approximation of wild-type cells. Pathway analysis was done identical to the cytogenetic analysis with the exception that chromosome-based gene sets were replaced with pathway-based gene sets. Software to perform the cytogenetic analysis is implemented in the BioConductor reb package and a BioConductor package to perform the pathway analysis can be found at http://www.vai.org/upload/departments/computationalbiology/PGSEA_0.9.0.tar.gz.

Other statistical analyses. Differences in the proportion of chromosomal and pathway abnormalities between the subtypes were determined using Pearson's χ2 test. To identify genes that are associated with the MYC signature, the samples were partitioned into two groups, those that contained the strong active MYC signature (n = 11) and those that did not (n = 25), and discriminate gene analysis was done using the limma package using the default settings (19). For cell viability experiments, results are represented as mean ± SE. Comparisons of values were made using Student's t test. Significance was assumed at P < 0.05.

Immunohistochemistry/histology. Immunohistochemical staining was done on 4 μm formalin-fixed, paraffin-embedded tissue sections. Endogenous peroxidase activity was blocked with 3% hydrogen peroxide. Antigen retrieval was carried out in citrate buffer (10 mmol/L; pH 6) for 15 min at 100°C in a microwave oven. The slides were incubated with a primary rabbit monoclonal anti-Myc antibody at 1:100 dilution for 1 h at room temperature. Sections were then incubated with secondary anti-mouse IgG for 30 min. A subsequent reaction was done with biotin-free horseradish peroxidase enzyme–labeled polymer from EnVision plus detection system (Dako, Carpinteria, CA). A positive reaction was visualized with diaminobenzidine solution followed by counterstaining with hematoxylin. Prostatic adenocarcinoma was used as a positive control. Appropriate negative controls for immunostaining were prepared by using nonimmune rabbit IgG.

Cell culture experiments. The RCC cell lines Caki-2, A-498, and ACHN were obtained from American Type Culture Collection (Manassas, VA). For Affymetrix microarray and reverse transcription-PCR (RT-PCR) analysis, total RNA was extracted by using TRIzol reagents (Invitrogen, Carlsbad, CA) and further purified using RNeasy Mini Kit (QIAGEN Sciences, Germantown, MD). For Western blot analysis, protein was extracted using radioimmunoprecipitation assay lysis buffer and detected using antibodies specific for c-Myc (Santa Cruz Biotechnology, Santa Cruz, CA) and β-actin (Abcam, Cambridge, MA). Short interfering RNA (siRNA) oligonucleotides with two thymidine residues (dTdT) at the 3′ end of the sequence were designed for c-myc (sense, 5′-ACAGAAAUGUCCUGAGCAA-3′; antisense, 5′-UUGCUCAGGACAUUUCUGU-3′) and dissolved in RNase/DNase-free water at 20 μmol/L. Approximately 105 cells were plated per well in six-well plates and siRNAs were transfected using Transfectamine-2000 (Invitrogen) resulting in a final concentration of 100 nmol/L. The Myc-Max interaction inhibitor, 10058-F4 [(Z,E)-5-(4-ethylbenzylidine)-2-thioxothiazolidin-4-one], was purchased from EMD Biosciences (San Diego, CA) and dissolved in DMSO at a concentration of 50 mmol/L. Five hundred cells were seeded in each well of a 96-well plate and incubated at 37°C for 24 h. Viable cells were determined using CellTiter 96AQueous Nonradioactive Cell Proliferation Assay (Promega, Madison, WI).

Quantitative RT-PCR. Total RNA (3 μg) from siRNA-transfected cells was reverse-transcribed to cDNA using random hexamers and Superscript II RT (Invitrogen). PCR was carried out in a 20 μL volume with c-myc primers at 95°C for 5 min, followed by 95°C for 30 s, 58°C for 30 s, and 72°C for 45 s, for 10 cycles, followed by the addition of internal control primers (GAPD) and 18 additional cycles were done. The sequences of c-myc and GAPD PCR primers were: c-myc forward primer, 5′-CCAACAGGAACTATGACCTC-3′, reverse primer, 5′-CTCGGTCACCATCTCCAGCT-3′; GAPD forward primer, 5′-GAGTCCACTGGCGTCTTCAC-3′, and reverse primer, 5′-TGGTGCTCAGTGTAGCCCA-3′.

Fluorescence in situ hybridization. Fluorescence in situ hybridization (FISH) was done as described in ref. (20) using BAC clone CTD-3056O22 (Invitrogen) as a probe. FISH image acquisition was done with a COOL-1300 SpectraCube camera (Applied Spectral Imaging, Vista, CA) mounted on an Olympus BX51 epifluorescence microscope using the FISHView software (version 2.1; Applied Spectral Imaging). FISH signals were scored for 200 nuclei per slide.

Gene expression analysis of papillary RCC. A parametric method for gene set enrichment analysis (P-GSEA) was used to identify regions of likely DNA copy number change from the gene expression profiles derived from papillary RCC samples (14, 21). Using this method, based on the gene expression data specific for each chromosomal region, each sample receives a t-statistic (t) and significance value (P) that reflects whether a predicted amplification or deletion has occurred (Fig. 1A). Consistent with previous cytogenetic studies, the predicted gains of chromosome 7, 12, 16, and 17 were strongly reflected in the gene expression data derived from papillary type 1, low-grade type 2 (2a), and mixed type 1/2 samples (2). Like previous cytogenetic studies, high-grade papillary type 2 samples contained less frequent gains of chromosome 7, 16, and 17p, and more frequent losses of chromosome 9p. An additional transcriptional abnormality on chromosome 8q, a region not reported in other cytogenetic studies, was also identified as being enriched in the high-grade papillary tumors (4% versus 50%, P = 0.005). Because the amplification of chromosome 8q is not commonly reported in papillary RCC, as an additional control for this analysis, chromosomal abnormalities were also predicted from the gene expression data derived from a set of clear cell, or conventional RCCs. Cytogenetic abnormalities commonly found in clear cell RCC are distinct from those found in papillary tumors and include chromosome 3p loss in 70% to 80% of cases and gain of chromosome 5q in 50% to 60% of cases. Consistent with these established cytogenetic findings in RCC, predicted losses of chromosome 3p and predicted gains of 5p were prominent features of the clear cell RCC gene expression data, whereas amplification of 8q was not prominent in this subtype.

Figure 1.

Deregulation of MYC pathway genes and genes mapping to chromosome 8q in papillary RCC. Gene expression profiles derived from clear cell (CC, n = 31), papillary type 1 (P1, n = 14), mixed type 1/type 2a (P1.2A, n = 5), type 2a (P2a, n = 4), and type 2b (P2b, n = 12) RCC samples were compared with gene expression profiles derived from nondiseased kidney tumor tissue (n = 12). A, gene lists that contain the genes mapping to each chromosomal arm. B, gene lists that contain genes responsive to perturbations in the indicated pathways were analyzed using P-GSEA (see Results). For pathway analysis, the “up” signature component indicates the list of genes that show increased expression relative to control cells for each pathway. Likewise, the “dn” signature component indicates the set of genes that show decreased expression relative to control cells for each pathway. GOF, signature components that were associated with gain of function of each pathway; LOF, signature components that were associated with loss of function of each pathway. The resulting summary statistic (t-statistic) for each gene list was plotted; red, a significant number of genes in the list had increased expression in the tumor samples relative to the normal kidney; blue, a significant number of genes in each list had decreased expression. Only the most significant data are displayed (P < 0.005). C, pathway analysis was done on gene expression profiles derived from the indicated cell lines as described in (B).

Figure 1.

Deregulation of MYC pathway genes and genes mapping to chromosome 8q in papillary RCC. Gene expression profiles derived from clear cell (CC, n = 31), papillary type 1 (P1, n = 14), mixed type 1/type 2a (P1.2A, n = 5), type 2a (P2a, n = 4), and type 2b (P2b, n = 12) RCC samples were compared with gene expression profiles derived from nondiseased kidney tumor tissue (n = 12). A, gene lists that contain the genes mapping to each chromosomal arm. B, gene lists that contain genes responsive to perturbations in the indicated pathways were analyzed using P-GSEA (see Results). For pathway analysis, the “up” signature component indicates the list of genes that show increased expression relative to control cells for each pathway. Likewise, the “dn” signature component indicates the set of genes that show decreased expression relative to control cells for each pathway. GOF, signature components that were associated with gain of function of each pathway; LOF, signature components that were associated with loss of function of each pathway. The resulting summary statistic (t-statistic) for each gene list was plotted; red, a significant number of genes in the list had increased expression in the tumor samples relative to the normal kidney; blue, a significant number of genes in each list had decreased expression. Only the most significant data are displayed (P < 0.005). C, pathway analysis was done on gene expression profiles derived from the indicated cell lines as described in (B).

Close modal

In parallel with the chromosomal abnormalities, gene expression profiles derived from the papillary RCC tumors were scanned for the presence of gene signatures reflective of oncogene activation, tumor suppressor gene inactivation, and more general cell responses. Similar to the DNA copy number predictions, each gene signature component receives a t-statistic (t) and significance value (P) indicating the potential for pathway activation or repression in each sample. The majority of clear cell RCCs are associated with inactivating sequence mutations in one allele of the VHL gene located on 3p25 and loss of the remaining wild-type VHL allele through chromosome 3p deletion (22). Nonfunctional VHL activity, as a result of VHL inactivation, leads to inappropriate HIF-1 transcriptional activation and subsequent induction of a hypoxic cell response. Consistent with this model, a set of genes that are down-regulated by VHL loss of function in a cell line model are also significantly down-regulated in the clear cell RCC tumor samples, but not in the papillary subtypes (Fig. 1B, VHL dn). In addition, a set of genes that are up-regulated when cells in culture are exposed to hypoxic conditions are also significantly up-regulated in the clear cell RCC tumors, but not the papillary tumors (Fig. 1B, Hypoxia up).

In contrast to the clear cell RCCs, the papillary RCCs do not contain predominant VHL or hypoxia-related pathway signatures. However, genes that are up-regulated by MYC overexpression in a cell line model are also significantly up-regulated in the high-grade papillary type 2 tumors (Fig. 1B, MYC up). In the high-grade papillary type 2 samples, 8 of 12 (67%) samples contain a prominent MYC activation signature versus 3 of 24 (13%) samples in the other papillary subtypes (P = 0.003). Reflective of the association of an active MYC signature component with papillary type 2, the presence of this pathway signature component is also associated with poor overall survival (Fig. 2). Therefore, whereas other molecular mechanisms besides MYC activation may be required for papillary type 2 tumor development, these data suggest that activation of MYC signaling may be a factor that influences the aggressiveness of this tumor subtype.

Figure 2.

Predicted Myc pathway activation was associated with decreased survival in papillary RCC. Kaplan-Meier analysis was done where clinical information was available. Comparisons of survival curves were done using the log-rank test. A, survival differences between papillary tumors that contain the active MYC signature component (n = 9) and those that did not (n = 19). B, survival differences between the type 2 samples that either have predicted Myc pathway activation (n = 7) or have no predicted Myc pathway activation (n = 5).

Figure 2.

Predicted Myc pathway activation was associated with decreased survival in papillary RCC. Kaplan-Meier analysis was done where clinical information was available. Comparisons of survival curves were done using the log-rank test. A, survival differences between papillary tumors that contain the active MYC signature component (n = 9) and those that did not (n = 19). B, survival differences between the type 2 samples that either have predicted Myc pathway activation (n = 7) or have no predicted Myc pathway activation (n = 5).

Close modal

Myc activation in papillary type 2b RCC cells. Based on the transcriptional abnormalities detected on chromosome 8q and the predicted MYC activation signature, we hypothesized that abnormalities associated with the MYC gene, which maps to chromosome 8q24, would influence the high-grade papillary samples. However, a computational approach that was modeled after the recently reported stepwise linkage analysis of microarray signatures method (23) was used to objectively identify both chromosomal abnormalities and individual gene abnormalities that associated with the MYC signature. In this analysis, the predicted amplification of 8q was the chromosomal region most highly correlated with the predicted MYC activation (Fig. 3A). Also, discriminate gene analysis identified MYC as one of the most highly overexpressed genes on chromosome 8 in samples that contain the MYC activation signature (Fig. 3B). In recent studies of liver and breast cancer, the biological effects of chromosome 8q amplification have been linked to two factors, MYC and a member of the COP9 signalosome, CSN5 (23, 24). However, in papillary RCC, overexpression of CSN5 was not associated with the MYC signature (Fig. 3B). Extraction of the gene expression level as measured on the microarray also shows that MYC is overexpressed in high-grade papillary samples (Fig. 3C), whereas immunohistochemical examination shows that the Myc protein is overexpressed (3+) in 5 of 5 (100%) high-grade samples versus 6 of 23 (26%) low-grade papillary samples (P = 0.01). Examples of c-Myc immunostaining on normal kidney tissue and high-grade papillary RCC are shown in Fig. 4. To confirm that MYC overexpression is the result of a DNA copy number change, FISH using a probe directed at MYC identified 8q24.21 amplification in zero of seven type 1 papillary tumors and three of nine (33%) papillary type 2 samples in an independent set of tumors (Fig. 3D; Supplemental Table S1). Taken together, these data support the hypothesis that type 2 papillary RCC is associated with functional activation of the MYC pathway and that amplification of chromosome 8q and subsequent overexpression of MYC is a common mechanism that is associated with MYC pathway activation.

Figure 3.

Myc deregulation in papillary RCC. A, for each sample, the predicted MYC activation signature summary score (t-statistic) was plotted relative to the predicted chromosome 8q amplification summary score (t-statistic). Samples were color-coded to reflect tumor subtype. The significance of the association was computed using Spearman correlation. B, relative log2-transformed expression of genes on chromosome 8 derived from comparisons of papillary samples that contain an active MYC signature component versus those that lack the signature. Relative changes in expression of the MYC (red) and CSN5 (blue) are highlighted. C, MYC expression in nondiseased kidney tissue (NO), papillary type 1 (P1), and papillary type 2b (P2b) as measured on the gene expression microarray. D, the relative copy number of MYC was measured using dual-color interphase FISH on tumor touch preparations. Representative photomicrograph with the CTD-3056O22 MYC probe (red) and 4′,6-diamidino-2-phenylindole counterstaining is shown for one of the papillary type 2 samples.

Figure 3.

Myc deregulation in papillary RCC. A, for each sample, the predicted MYC activation signature summary score (t-statistic) was plotted relative to the predicted chromosome 8q amplification summary score (t-statistic). Samples were color-coded to reflect tumor subtype. The significance of the association was computed using Spearman correlation. B, relative log2-transformed expression of genes on chromosome 8 derived from comparisons of papillary samples that contain an active MYC signature component versus those that lack the signature. Relative changes in expression of the MYC (red) and CSN5 (blue) are highlighted. C, MYC expression in nondiseased kidney tissue (NO), papillary type 1 (P1), and papillary type 2b (P2b) as measured on the gene expression microarray. D, the relative copy number of MYC was measured using dual-color interphase FISH on tumor touch preparations. Representative photomicrograph with the CTD-3056O22 MYC probe (red) and 4′,6-diamidino-2-phenylindole counterstaining is shown for one of the papillary type 2 samples.

Close modal
Figure 4.

Myc protein expression in papillary RCC. Myc expression in nondiseased kidney tissue (A) and in papillary type 2b RCC (B) as determined by immunohistochemistry.

Figure 4.

Myc protein expression in papillary RCC. Myc expression in nondiseased kidney tissue (A) and in papillary type 2b RCC (B) as determined by immunohistochemistry.

Close modal

To guide additional functional studies examining the role of MYC activation in papillary RCC, a set of kidney cancer–derived cell lines were examined for gene signatures indicative of pathway activation or repression. Cell lines derived from clear cell tumors, ACHN and A498, contain signatures indicative of VHL pathway inactivation and hypoxic cell response (Fig. 1C). The Caki-2 cell line, however, lacks a VHL and hypoxia gene signature and instead contains a MYC activation signature. Histologic evaluation of s.c. implantations of these cells are also consistent with high-grade type 2 papillary RCC (see Supplemental Fig. S1; ref. 25). In addition, cytogenetic analyses of these cell lines shows that although the Caki-2 cells are overall hyperploid (5n), these cells contain one to two extra copies of chromosome 8q as measured using FISH (see Supplemental Table S2). The combination of the histologic, cytogenetic, and gene expression data show that Caki-2 cells are very similar to high-grade papillary RCC. Additionally, although the ACHN cell line is triploid (3n), these cells contain no extra copies of chromosome 8q (see Supplemental Table S2) and no prominent MYC activation signature. In contrast, the A498 cell lines contain a MYC activation signature and two extra copies of chromosome 8q. Based on this analysis, the ACHN cell line shows the most similarity with clear cell RCC tumors.

Inhibition of Myc signaling disrupts the growth of papillary RCC cells. To verify that the previously reported MYC signature genes are influenced by the level of signaling downstream of c-Myc in kidney cancer cells, siRNA was used to knock down MYC expression in Caki-2 cells. In the siRNA-expressing cells, MYC knock-down was evident by measurement of expression levels using RT-PCR and by measurement of protein levels using Western blotting (Fig. 5A and B). Extraction of expression levels measured on the microarray also showed a 40% decrease in MYC expression (data not shown). Importantly, the MYC signature genes as a whole are diminished in the Caki-2 control cells relative to the Caki-2 knock-down cells. In this case, 63 of 98 (64%) genes up-regulated by MYC overexpression were subsequently down-regulated in the MYC knock-down cells (t = −3.6, P = 0.0005). Conversely, 59 of 85 (69%) genes down-regulated by MYC overexpression are now re-expressed in the knock-down cells (t = 4.9, P = 0.00005; Fig. 5B, inset). These data confirm that the downstream signaling of Myc in kidney cells influences the expression levels of a substantial number of genes in the MYC signature. Evaluation of Caki-2 cells expressing either a control siRNA or a MYC-directed siRNA shows that Myc signaling facilitates the growth of these cells (Fig. 5C). In addition, kidney cancer–derived cell lines were also exposed to the Myc-Max dimerization inhibitor 10058-F4; a compound that disrupts Myc-dependent growth in vitro (26). The addition of the 10058-F4 compound inhibited the growth of the Caki-2 cell line (Fig. 5D) but did not disrupt the growth of the ACHN cells. Taken together, these data support the role of Myc pathway activation in the development/progression of papillary type 2 RCC.

Figure 5.

Myc signaling inhibitors prevent the growth of papillary type 2b tumor cells in vitro. Caki-2 cells transfected with either control siRNA or siRNA directed against MYC were analyzed by RT-PCR (A) and by Western blotting (B) using an anti-Myc antibody. C, pathway analysis was done on comparisons of gene expression profiles derived from cell lines transfected with either control siRNA or MYC-directed siRNA (inset). Growth characteristics of cell expressing control siRNA or MYC-directed siRNA. The fraction of viable cells relative to cells at time zero was plotted. *, times of significant viability differences. D, fraction of viable cells after exposure to increasing concentrations of compound 10058-F4 relative to mock-treated cells. *, concentrations of significant viability differences.

Figure 5.

Myc signaling inhibitors prevent the growth of papillary type 2b tumor cells in vitro. Caki-2 cells transfected with either control siRNA or siRNA directed against MYC were analyzed by RT-PCR (A) and by Western blotting (B) using an anti-Myc antibody. C, pathway analysis was done on comparisons of gene expression profiles derived from cell lines transfected with either control siRNA or MYC-directed siRNA (inset). Growth characteristics of cell expressing control siRNA or MYC-directed siRNA. The fraction of viable cells relative to cells at time zero was plotted. *, times of significant viability differences. D, fraction of viable cells after exposure to increasing concentrations of compound 10058-F4 relative to mock-treated cells. *, concentrations of significant viability differences.

Close modal

In this study, we provide both computational and molecular genetic evidence that MYC activation is associated with high-grade papillary type 2 RCC. Activation of MYC in sporadic papillary RCC is reminiscent of both primary and secondary pediatric renal adenocarcinomas (27, 28), which are associated with a translocation involving the TFE3 gene located on Xp11.2. TFE3 encodes a member of the MYC family of helix-loop-helix transcription factors and inappropriate expression of the translocation protein is likely involved in the development of these tumors. As such, it will be useful to examine if there are any similarities between the childhood translocation and adult high-grade papillary tumors and to determine if inappropriate activation of these and other members of the Myc family are involved in the tumorigenesis of other types of RCC.

Consistent with previous studies (4), type 2 papillary RCCs have increased cytogenetic complexity when compared with type 1 tumors. In the papillary type 1 samples, chromosomal abnormalities on chromosomes 2, 7, 12, 16, and 17 were identified in >70% of samples; five additional regions had abnormalities at intermediate frequencies (30–70%). In contrast, chromosome 9q and 17q abnormalities were identified in >70% of papillary type 2 samples, whereas 25 additional regions had abnormalities at intermediate frequencies. The cytogenetic complexity is coincident with Myc overexpression as shown via immunohistochemical staining. As MYC activation is known to lead to chromosomal instability (CIN), our results support a model in which type 1 and type 2 have similar origins, but subsequent MYC activation results in increased CIN and tumor growth in the type 2 tumors. A gene expression signature that correlates with CIN was recently published (29). Interestingly, the genes in the CIN signature were dramatically increased in high-grade type 2 papillary tumors as compared with type 1 tumors (data not shown). These observations suggest that therapeutic avenues that target either CIN or Myc activation have the potential to be effective interventions for papillary type 2 RCC.

Finally, in this report, P-GSEA was used to gain biological insights into both DNA copy number changes and deregulated signal transduction pathways. Although we have shown that chromosome-centric analysis of transcriptional data has the potential to identify genetic abnormalities, this type of analysis also has the potential to identify epigenetic abnormalities that could result in large-scale transcriptional deregulation. Likewise, detection of a pathway activation signature is independent of the mechanism of activation. Therefore, the effects of a wide spectrum of molecular genetic events (DNA sequence mutations, cytogenetic abnormalities, epigenetic modifications) in any critical MYC pathway component would be reflected in the appearance of the pathway activation signature.

Note: Supplementary data for this article are available at Cancer Research Online (http://cancerres.aacrjournals.org/).

Current address for X.J. Yang: Department of Pathology and Laboratory Medicine, Weill Medical College, Cornell University, 525 East 68th Street, New York, NY 10021.

Grant support: NIH grant R33-CA10113-01 (K.A. Furge) and the Van Andel Research Institute. The Gerber Foundation, Amway Japan Limited, the Schregardus Family Foundation, Fisher Family Trust, the Hauenstein Foundation, the Michigan Economic Development Corporation, and the Michigan Technology Tri-Corridor (B.T. Teh).

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

We thank the Cooperative Human Tissue Network of the National Cancer Institute for providing the renal tumor cases, and Sabrina Noyes for manuscript preparation and submission.

1
Delahunt B, Eble JN. Papillary renal cell carcinoma: a clinicopathologic and immunohistochemical study of 105 tumors.
Mod Pathol
1997
;
10
:
537
–44.
2
Yang XJ, Tan MH, Kim HL, et al. A molecular classification of papillary renal cell carcinoma.
Cancer Res
2005
;
65
:
5628
–37.
3
Jiang F, Richter J, Schraml P, et al. Chromosomal imbalances in papillary renal cell carcinoma: genetic differences between histological subtypes.
Am J Pathol
1998
;
153
:
1467
–73.
4
Gunawan B, von Heydebreck A, Fritsch T, et al. Cytogenetic and morphologic typing of 58 papillary renal cell carcinomas: evidence for a cytogenetic evolution of type 2 from type 1 tumors.
Cancer Res
2003
;
63
:
6200
–5.
5
Hughes TR, Roberts CJ, Dai H, et al. Widespread aneuploidy revealed by DNA microarray expression profiling.
Nat Genet
2000
;
25
:
333
–7.
6
Phillips JL, Hayward SW, Wang Y, et al. The consequences of chromosomal aneuploidy on gene expression profiles in a cell line model for prostate carcinogenesis.
Cancer Res
2001
;
61
:
8143
–9.
7
Desai KV, Xiao N, Wang W, et al. Initiating oncogenic event determines gene-expression patterns of human breast cancer models.
Proc Natl Acad Sci U S A
2002
;
99
:
6967
–72.
8
Ferrando AA, Neuberg DS, Staunton J, et al. Gene expression signatures define novel oncogenic pathways in T cell acute lymphoblastic leukemia.
Cancer Cell
2002
;
1
:
75
–87.
9
Huang E, Ishida S, Pittman J, et al. Gene expression phenotypic models that predict the activity of oncogenic pathways.
Nat Genet
2003
;
34
:
226
–30.
10
Bild AH, Yao G, Chang JT, et al. Oncogenic pathway signatures in human cancers as a guide to targeted therapies.
Nature
2005
;
439
:
353
–7.
11
Ihaka R, Gentleman R. R: A language for data analysis and graphics.
J Comput Graph Stat
1996
;
5
:
299
–314.
12
Gentleman RC, Carey VJ, Bates DM, et al. Bioconductor: open software development for computational biology and bioinformatics.
Genome Biol
2004
;
5
:
R80
.
13
Dai M, Wang P, Boyd AD, et al. Evolving gene/transcript definitions significantly alter the interpretation of GeneChip data.
Nucleic Acids Res
2005
;
33
:
e175
.
14
Lindvall C, Furge KA, Bjorkholm M, et al. Combined genetic- and transcriptional profiling of acute myeloid leukemia with complex and normal karyotypes.
Haematologia
2004
;
89
:
1072
–81.
15
Crawley JJ, Furge KA. Identification of frequent cytogenetic aberrations in hepatocellular carcinoma using gene expression data. Genome Biol 2002;3:RESEARCH0075.
16
Gerritsen ME, Tomlinson JE, Zlot C, Ziman M, Hwang S. Using gene expression profiling to identify the molecular basis of the synergistic actions of hepatocyte growth factor and vascular endothelial growth factor in human endothelial cells.
Br J Pharmacol
2003
;
140
:
595
–610.
17
Staller P, Sulitkova J, Lisztwan J, Moch H, Oakeley EJ, Krek W. Chemokine receptor CXCR4 downregulated by von Hippel-Lindau tumour suppressor pVHL.
Nature
2003
;
425
:
307
–11.
18
Chi JT, Wang Z, Nuyten DS, et al. Gene expression programs in response to hypoxia: cell type specificity and prognostic significance in human cancers.
PLoS Med
2006
;
3
:
e47
.
19
Smyth GK. Linear models and empirical Bayes methods for assessing differential expression in microarray experiments.
Stat Appl Genet Mol Biol
2004
;
3
:
Article3
.
20
Graveel C, Su Y, Koeman J, et al. Activating Met mutations produce unique tumor profiles in mice with selective duplication of the mutant allele.
Proc Natl Acad Sci U S A
2004
;
101
:
17198
–203.
21
Kim S, Volsky DJ. PAGE: parametric analysis of gene set enrichment.
BMC Bioinformatics
2005
;
6
:
144
.
22
Cohen HT, McGovern FJ. Renal cell carcinoma.
N Engl J Med
2005
;
353
:
2477
–90.
23
Adler AS, Lin M, Horlings H, Nuyten DS, van de Vijver MJ, Chang HY. Genetic regulators of large-scale transcriptional signatures in cancer.
Nat Genet
2006
;
38
:
421
–30.
24
Patil MA, Gutgemann I, Zhang J, et al. Array-based comparative genomic hybridization reveals recurrent chromosomal aberrations and JAB1 as a potential target for 8q gain in hepatocellular carcinoma.
Carcinogenesis
2005
;
26
:
2050
–7.
25
Pulkkanen KJ, Parkkinen JJ, Kettunen MI, et al. Characterization of a new animal model for human renal cell carcinoma.
In Vivo
2000
;
14
:
393
–400.
26
Yin X, Giap C, Lazo JS, Prochownik EV. Low molecular weight inhibitors of Myc-Max interaction and function.
Oncogene
2003
;
22
:
6151
–9.
27
Bruder E, Passera O, Harms D, et al. Morphologic and molecular characterization of renal cell carcinoma in children and young adults.
Am J Surg Pathol
2004
;
28
:
1117
–32.
28
Argani P, Lae M, Ballard ET, et al. Translocation carcinomas of the kidney after chemotherapy in childhood.
J Clin Oncol
2006
;
24
:
1529
–34.
29
Carter SL, Eklund AC, Kohane IS, Harris LN, Szallasi Z. A signature of chromosomal instability inferred from gene expression profiles predicts clinical outcome in multiple human cancers.
Nat Genet
2006
;
38
:
1043
–8.

Supplementary data