The role of molecular markers in the staging of renal cell carcinoma
Arie S. Belldegrun, Department of Urology, David Geffen School of Medicine at UCLA, 66–118 CHS, Box 951738, Los Angeles, CA 90095–1738, USA.
UCLA Integrated Staging System
carbonic anhydrase IX
Molecular Integrated Staging System
vascular endothelial growth factor
Cancers of the kidney are estimated to account for 38 890 new cases and 12 840 deaths in the USA in 2006 . RCC is a highly aggressive tumour; a third of patients will have evidence of metastasis at the time of diagnosis  and > 40% of patients with RCC will die from their disease . Advances in imaging, staging and the treatment have led to a significant and progressive increase in relative 5-year survival rates for patients with RCC . Despite these advances, the heterogeneous nature of the disease remains a clinical challenge.
RCC comprises a family of epithelial tumours arising from within the kidney. Each subtype of RCC presents a unique clinical picture, with varied tumour biology, patient prognosis and response to treatment. Furthermore, there is a wide disparity in clinical outcomes for patients even within specific RCC subtypes. These challenges underscore the critical importance of accurate indicators of prognosis for this patient population.
Historically, staging paradigms have been refined to move ever closer to a fixed goal; accurate prognostic information for individual patients and clinicians. Recent breakthroughs have led to greater knowledge of the molecular genetics and oncogenic pathways of kidney cancer. These advances have led directly to novel and exciting therapies for RCC. The overall prognosis for patients with localized and advanced RCC is likely to improve in step with these emerging therapies, necessitating a rapid reassessment of RCC staging constructs.
Protein and gene expression analysis promise to aid in recapitulating and, in the near future, to further refine the current histological classifications of RCC. Molecular markers will also advance current staging systems and improve prognostic information for patients and clinicians. Finally, the molecular signature of RCC tumours will allow for sophisticated application of systemic and targeted therapies, improving patient response and minimizing unnecessary exposure of patients to treatment toxicities. We review the significance of molecular markers in the understanding of tumour biology, and the development of future staging paradigms and treatment strategies for RCC.
THE DEVELOPMENT OF STAGING SYSTEMS: FROM MACRO TO MICRO TO MOLECULAR
Staging systems for kidney cancer serve to provide: (i) a description of the size and spread of a tumour; (ii) aid in selecting therapeutic options for each patient; (iii) accurate prognostic information to stratify risk of disease recurrence or cancer-related death; (iv) criteria to identify patient populations for specific adjuvant therapies; and (v) inclusion criteria for clinical trials .
RCC staging systems have developed in parallel with the rapid increase in understanding of renal tumour biology. The first formal staging systems, proposed by Flocks and Kadesky  and later modified by Robson et al., used the anatomical information available to clinicians at the time. Numerous refinements have led to the current TNM system proposed by the Union Internationale Contre le Cancer . Integrated staging systems have been revised to include a myriad of pathological, histological and clinical characteristics that have been shown to be prognostic indicators for RCC. For example, the UCLA Integrated Staging System (UISS) supplements the anatomical TNM staging with the Eastern Cooperative Oncology Group performance status and the Fuhrman grade of the tumour . The UISS, as well as other integrated staging systems such as the Kattan postoperative nomogram , have been shown to be powerful tools with improved prognostic ability compared to anatomical staging alone [10,11]. As research advances our understanding of the molecular and genetic biology of kidney cancer, including important genetic and protein molecular markers represents the next logical advance in RCC staging systems.
MOLECULAR STAGING SYSTEMS
The expression of gene/protein markers might provide prognostic information as: (i) surrogate markers of tumour aggressiveness; (ii) a mechanism directly involved in tumour growth or metastasis; (iii) antigens recognized by the host immune system; (iv) markers that serve as the targets of targeted therapies, or that predict response to treatments.
CARBONIC ANHYDRASE IX (CAIX) AS AN EXAMPLE OF A PROGNOSTIC MOLECULAR MARKER
Significant attention has been paid to CAIX (also known as G250 or MN), a member of the CA family. CAIX is thought to assist in regulating intracellular and extracellular pH levels in response to tumoral hypoxia and subsequent anaerobic metabolism. In one study, CAIX was detected in 86% of RCC samples studied, but was found in only 9% of normal kidney tissue samples, suggesting that CAIX expression might be a useful diagnostic biomarker . A study at UCLA found that 94% of clear cell RCC tumour samples stained positively for CAIX, while expression was completely absent in benign tumours such as oncocytoma . Low levels of CAIX staining were an independent indicator of poor survival in patients with metastatic RCC (hazard ratio 3.10). For patients with localized, high-risk RCC lesions, low CAIX staining also implied a worse prognosis. CAIX represents the potential of a single molecular marker to stratify patient prognosis. Finally, CAIX also represents a potential therapeutic target. Clinical trials with radiolabelled monoclonal antibodies against CAIX in patients with RCC showed selective and specific delivery of monoclonal antibody to renal cancer sites, with both primary and metastatic RCC deposits being capable of being targeted and imaged , leading to a large phase III adjuvant clinical trial in high-risk non-metastatic patients. Thus, CAIX represents a unique molecular marker that has diagnostic, prognostic and potentially therapeutic implications.
PROTEIN EXPRESSION AND STAGING NOMOGRAMS
In addition to CAIX, there are innumerable molecular markets relevant to RCC biology . Recently, Kim et al.[16,17] combined a panel of these protein molecular markers and the UISS to create a Molecular Integrated Staging System (MISS). Immunohistochemical analysis of Ki-67, p53, gelsolin, CAIX, CAXII, PTEN, EpCAM, and vimentin was analysed from 318 clear cell RCC tumours. Increased staining for Ki-67, p53, vimentin and gelsolin correlated with worse survival, while the inverse was true for CAIX, PTEN, CAXII and EpCAM. In multivariate analysis, the presence of metastasis, gelsolin, p53 expression and CAIX remained significant predictors of survival and were used to create a prognostic model (marker model). This model, based entirely on molecular marker information, performed better than individual clinical variables, e.g. tumour grade and T stage, and was as powerful as the UISS model. Combining clinical variables (UISS) and the marker model resulted in a prognostic system for clear cell RCC (MISS) that out-performed exclusively molecular or clinical staging systems  (Fig. 1). This report shows several important principles. First, it serves as a proof of principle that molecular marker information can be a powerful tool in the staging of RCC. In addition, combining molecular marker information will complement existing staging systems to improve prognostic accuracy.
GENE EXPRESSION ANALYSIS IN STAGING AND PROGNOSTIC SYSTEMS
Gene expression profiling might logarithmically expand the scope of molecular marker analysis by offering the ability to analyse simultaneously thousands of candidate genes in high-throughput arrays. While not as mature as protein molecular marker data, gene expression-analysis studies have also shown an ability to define patient prognosis, linking genetic signatures with clinical outcome. Takahashi et al. studied 29 patients with clear cell RCC. When compared with normal tissue samples, clear cell RCC tumours had high expression of markers such as vascular endothelial growth factor (VEGF) and ceruloplasmin, and down-regulation of kininogen. Among these clear cell tumours, the authors identified 40 genes that differentiated patients with the best prognosis. Increased expression of SPROUTY, an angiogenesis inhibitor, was associated with a good prognosis while loss of TGF-β receptor II and metalloproteinase-3 were associated with poor outcomes. In a second study by Takahashi et al., a unique expression profile in clear cell tumours differentiated a poor-prognosis subcluster, including patients who died from their disease, from patients with no evidence of metastasis. Jones et al. also reported the ability of gene expression analysis within the primary tumour to identify a metastatic signature. Vasselli et al. used gene expression to evaluate prognosis for patients with metastatic RCC. These authors evaluated 58 metastatic RCC specimens for overall gene-expression patterns most correlated with patient survival. The authors identified 45 genes that could be used to separate patients into groups by prognosis. Increased expression of VCAM-1 was particularly powerful in selecting patients with improved prognosis. The International Kidney Cancer Study Group recently reported the ability of gene expression profiling to predict survival for patients with clear cell RCC . Evaluating 110 primary tumours resulted in gene predictors of patient survival independent of traditional clinicopathological variables. Kosari et al. were also able to identify 35 differentially expressed genes that allowed for clustering of non-aggressive clear cell RCC specimens separate from aggressive and metastatic tumours. That aggressive but localized RCC samples had a gene expression profile similar to metastatic tumours presents an opportunity to identify patients at the highest risk of disease progression and recurrence, and those most likely to benefit from adjuvant therapies. These early reports suggest that, within a short period, gene expression analysis will become an integral component of future kidney cancer staging systems.
FUTURE DIRECTIONS IN MOLECULAR STAGING
PREDICTING RESPONSE TO IMMUNOTHERAPY
Immunotherapy currently remains the only systemic treatment option with significant, although modest, complete response rates. About 20% of patients have a clinical response to high-dose interleukin-2 (IL-2) therapy, but the treatment is associated with significant morbidity and complexity of use . A model using histological and pathological features of primary tumour specimens to select patients most likely to respond to IL-2-based immunotherapy has been described. Clear cell histology and lack of granular features are strong predictors of IL-2 response , and increase response rates to >35%, while decreasing the unnecessary exposure of patients to treatment toxicity. Molecular markers, e.g. CAIX, have also been shown to predict the response to IL-2 immunotherapy. Studies of protein expression of primary tumours have confirmed the high expression of CAIX in patients with a complete response to IL-2-based therapies [13,26]. Gene array analysis has also been used to evaluate thousands of candidate markers of response to immunotherapy in RCC . Complete responders to immunotherapy could be discriminated from nonresponders, using a set of 73 genes. This set included genes involved in apoptosis, G protein signalling, the ubiquitin-proteasome proteolytic system, chemokines, heat shock proteins, MHC antigens, and the mitogen-activated protein kinase pathway, as well as the PTEN tumour suppressor and CAIX. If prospectively validated, these data suggest that gene expression profiling might be able to further refine the selection of patients for immunotherapy.
PREDICTING RESPONSE TO TARGETED THERAPIES
Targeted therapies are likely to develop into the mainstay of cancer treatment. The successful implementation of these promising therapies requires an understanding of the expression of the target by the tumour, as well as the ability to specifically interfere with its action. The importance of angiogenesis factors, e.g. the VEGF family, in RCC is becoming clear . Gene expression analysis has repeatedly shown the importance of angiogenesis factors, e.g. VEGF-A, in RCC specimens [18,19,29,30]. Bevacizumab (Avastin®, Genentech, South San Francisco, CA, USA), a neutralizing antibody to VEGF-A, prolongs time to progression in patients with advanced cytokine-refractory clear cell RCC [31,32]. The USA Food and Drug Administration recently approved the use of the tyrosine kinase inhibitors SU11248 (sunitinib, Sutent®, Pfizer Inc., USA) and AG013736 (axitinib, Pfizer) in the treatment of clear cell RCC. These tyrosine kinase inhibitors, and others in development, have varied affinity for the VEGF and TGF family of tyrosine kinase receptors, and have shown promising objective response rates in ongoing clinical trials. Emerging data suggest that the expression of tyrosine kinase receptors will identify patients most likely to respond to agents targeting this angiogenesis axis .
Recently, data were reported relating gene expression of the mTOR pathway surrogates to response to the mTOR pathway inhibitor CCI-779, from a parent study in patients with metastatic RCC . No patient who had low expression of either pS6K or pAkt (downstream effectors of the mTOR pathway) was a responder to CCI-779. These results provide evidence that there is a correlation between surrogates of mTOR activation (tumour cell expression of pAkt and pS6K) and response to CCI-779. Thus, patients who have high gene expression appear to be those most likely to benefit from agents targeting the mTOR pathway.
As the factors that define patient prognosis are better understood, kidney cancer staging systems will continue to develop. High-throughput tissue arrays have facilitated the rapid analysis of potential protein and genetic molecular markers. The application of molecular marker information will continue remain the next frontier in the advance of RCC staging constructs leading to more accurate predictions of an individual patient’s prognosis. The emergence of novel targeted therapies in RCC will further necessitate the application of molecular marker research. Understanding the molecular signature of RCC tumours will allow for sophisticated application of systemic and targeted therapies and ultimately improve patient outcomes.
CONFLICT OF INTEREST