Elsevier

European Urology Focus

Volume 4, Issue 5, September 2018, Pages 740-748
European Urology Focus

Kidney Cancer
Integrated Phenotypic/Genotypic Analysis of Papillary Renal Cell Carcinoma Subtypes: Identification of Prognostic Markers, Cancer-related Pathways, and Implications for Therapy

https://doi.org/10.1016/j.euf.2016.09.002Get rights and content

Abstract

Background

Two histologic subtypes are recognized for papillary renal cell carcinoma (PRCC). Studies have shown that the subtypes differ in characteristic genetic alterations and clinical behavior. Clinically, the subtypes are managed similarly.

Objectives

To analyze the biological differences between the two PRCC histological subtypes, in order to further guide their clinical management.

Design, setting, and participants

PRCC cohort consisting of 317 patients from the Cancer Genome Atlas database and our institution. Patients were stratified according to histologic criteria as type 1, type 2, or not otherwise specified (NOS). Gene and miRNA expression data for the cohort were examined via unsupervised and supervised clustering.

Outcome measurements and statistical analysis

Significant molecular signatures for each subtype were used to unravel the implicated molecular pathways via bioinformatics analysis. Survival was compared between the subtypes. Newly discovered biomarkers were used to further stratify survival of patients in the NOS category.

Results and limitations

Tumor genotyping revealed two distinct PRCC subtypes. The top molecular pathways enriched in PRCC1 were WNT, Hedgehog, and Notch signaling (p = 0.001–0.01); highlighting an embryonic developmental theme to the pathogenesis of this subtype. PRCC2 showed enrichment in the mTOR, VEGF (p = 7.49E−09) and HIF (p = 7.63E−05) signaling pathways. Overall survival and disease-free survival significantly differed between the types. ABCC2 expression was identified as a significant prognostic biomarker for the NOS group in univariate (log rank p < 0.0001; hazard ratio [HR] >11.63) and multivariate analysis (p = 0.003; HR >2.12). ABCC2 expression and its effect on survival should be further validated at the protein level.

Conclusions

The classical PRCC types 1 and 2 have two distinct genotypes. We unraveled pathways that indicate that the two types could potentially respond differently to current therapies. We also identified biomarkers that stratify tumors within the PRCC NOS category into prognostic subgroups. Our findings highlight the need for molecular markers to accurately subtype PRCC and guide clinical management.

Patient summary

The two types of papillary renal cancer are treated similarly. We show that the two types have a different genetic makeup, and hence they should be considered two different tumors. There is a different biology underlying each tumor type that can potentially affect the way they respond to treatment. We uncovered genes that can be tested for to guide therapy in some problematic cases for which it hard to define the tumor type.

Introduction

Papillary renal cell carcinoma (PRCC) is the second most common adult renal malignancy [1], [2]. Although PRCC morphology has been described since 1976, it was recognized in 1997 as having two histologic subtypes. Type 1 is characterized by small, pale cells with low-grade nuclei. Conversely, type 2 has abundant eosinophilic cytoplasm and higher-grade nuclei with prominent nucleoli, in a pseudo-stratified arrangement [2], [3].

It has been shown that PRCC2 is associated with a higher grade and stage than PRCC1 [2], [4]. Studies found that type 2 is associated with poor survival [2], [4]. However, most studies have repeatedly shown that histologic grade is associated with survival in PRCC [5], [6]. By morphologic definition, PRCC2 is a higher-grade tumor and seems to be associated with worse prognosis.

Studies have opted to distinguish between the two subtypes at the molecular level. Chromosomal changes commonly accompanying PRCC1 as gains of chromosomes 7, 17, and loss of Y are not recapitulated in PRCC2. Type 2 chromosomal changes are less consistent, including losses of 3p [7], [8], 1p [1], [8], or 9p [9]. Gene expression profiling for the two types revealed two distinct tumors [7]. The Cancer Genome Atlas (TCGA) PRCC study also revealed unique molecular signatures, with type 2 being further divided into three clusters [9]. Altered MET gene status is present in 80% of PRCC1, while PRCC2 is reported to have overexpression of the NRF2-ARE pathway [9], [10].

Some PRCCs are difficult to classify as they do not fulfill all criteria for either subtype or they have mixed morphology [1], [7], [8], [11]. The percentage of these PRCCs not otherwise specified (NOS) can be as high as 47% [1]. This highlights the need for a more accurate molecular classification that relies on biology to complement the morphology for PRCC classification. Another subtype of PRCC was described in which the cells are oncocytic and have a non-overlapping histology. The limited number of cases described for this category seemed to have excellent clinical outcome [11], [12].

Current targeted therapies for metastatic RCCs include mTOR and VEGF inhibitors. The response rate of PRCC to these modalities is lower than that of clear-cell renal cell carcinoma (CCRCC) [13], [14], [15], [16]. However, studies analyzing the PRCC response to targeted therapies did not stratify patients into types 1 and 2 [14], [15], [16].

We analyzed the molecular attributes responsible for the pathogenesis of the two subtypes of PRCC using two independent cohorts. We identified clinically useful diagnostic and prognostic biomarkers to stratify the subtypes. Considering the dilemma of NOS tumors, our aim was to identify new molecular markers of diagnostic and prognostic validity.

Section snippets

Specimen collection and morphologic subtyping

As detailed in the Supplementary material, we analyzed 281 PRCC cases from the TCGA KIRP cohort, in addition to 35 cases of PRCC from our institution. Tumors were classified as PRCC1 or PRCC2 according to the original PRCC criteria [3]. Cases that did not meet all the criteria or showed mixed morphology were classified as PRCC NOS. Cases that had an oncocytic non-overlapping papillary morphology 11,12] were designated oncocytic PRCC NOS (Fig. 1A).

RNA extraction and miRNA expression analysis

Pure tumor areas were selected from

Histologic subtyping of PRCC

A total of 290 of the 317 histologic slides available from both cohorts combined had sufficient tissue for accurate subtyping. Using the classic morphologic criteria [3], [11], 72 cases were subtyped as PRCC1, 70 as PRCC2, 131 as NOS, and 17 as oncocytic non-overlapping NOS (Fig. 1A–C).

Molecular subtyping of PRCC

To determine if PRCC1 and PRCC2 are biologically distinct, we performed unsupervised clustering of the gene expression profiles for 80 cases of PRCC (40 cases with classic PRCC1 morphology and 40 cases with

Discussion

We provide strong evidence of what other studies have implied: PRCC types 1 and 2 represent two distinct biological cancers, and should be clinically managed as such [7], [9]. In our study, cases that lacked any of the original histologic criteria were labeled as PRCC NOS (45% of cases). Our ratio of NOS is similar to what Chevarie-Davis et al [1] found in their PRCC study. In all likelihood, our NOS category had cases belonging to PRCC1 or PRCC2, and perhaps cases belonging to a third,

Conclusions

PRCC1 and PRCC2 are two distinct tumors. We provide the first evidence of molecular pathways that could have significant implications for PRCC management. Our study demonstrates that ABCC2 has great potential as a prognostic and predictive biomarker in the PRCC patient population. Our findings highlight the need to augment morphologic criteria with molecular markers to guide clinical management.


Author contributions: George M. Yousef had full access to all the data in the study and takes

References (30)

  • H. Guo et al.

    The PI3K/AKT pathway and renal cell carcinoma

    J Genet Genomics

    (2015)
  • H.-S. Jeong et al.

    Regulation of the expression of renal drug transporters in KEAP1-knockdown human tubular cells

    Toxicol In Vitro

    (2015)
  • M. Chevarie-Davis et al.

    Morphologic and immunohistochemical spectrum of papillary renal cell carcinoma: study including 132 cases with pure type 1 and type 2 morphology as well

    Am J Surg Pathol

    (2014)
  • B. Delahunt et al.

    Papillary renal cell carcinoma: a clinicopathologic and immunohistochemical study of 105 tumors

    Mod Pathol

    (1997)
  • P. Gontero et al.

    Prognostic factors in a prospective series of papillary renal cell carcinoma

    BJU Int

    (2008)
  • Cited by (18)

    • Tumors masquerading as type 2 papillary renal cell carcinoma: pathologists’ ever-expanding differential diagnosis for a heterogeneous group of entities

      2022, Urologic Oncology: Seminars and Original Investigations
      Citation Excerpt :

      Additional changes to the classification of renal tumors are not a matter of “if,” but “when.” Even within the silo of PRCC, there are proposals for “type 3” and “type 4” PRCC based on tumor histomorphology that correlates with clustering in molecular and clinical features [143,144]. “Type 3” PRCC is morphologically similar to T2PRCC, albeit with smaller cells, less abundant cytoplasm, less frequent pseudostratification [138].

    • ABCC2 expression in papillary renal cell carcinoma provides better prognostic stratification than WHO/ISUP nucleolar grade

      2022, Human Pathology
      Citation Excerpt :

      We have recently proposed a new PRCC stratification system that combined morphology and molecular characteristics of PRCC. The classification was based on an initial analysis of The Cancer Genome Atlas (TCGA) PRCC cohort, in which we identified a number of differentially expressed genes between the classic PRCC type 1 and type 2 [21]. ATP-binding cassette subfamily C member 2 (ABCC2), also called MRP2, was the only marker that exhibited significant in silico separation of the PRCC Not otherwise specified (NOS) category (i.e., cases not fully meeting all criteria for either type 1 or type 2) for overall survival (OS) [21].

    View all citing articles on Scopus
    View full text