Recognizing the Continuous Nature of Expression Heterogeneity and Clinical Outcomes in Clear Cell Renal Cell Carcinoma

Abstract

Clear cell renal cell carcinoma (ccRCC) has been previously classified into putative discrete prognostic subtypes by gene expression profiling. To investigate the robustness of these proposed subtype classifications, we evaluated 12 public datasets, together with a new dataset of 265 ccRCC gene expression profiles. Consensus clustering showed unstable subtype and principal component analysis (PCA) showed a continuous spectrum both within and between datasets. Considering the lack of discrete delineation and continuous spectrum observed, we developed a continuous quantitative prognosis score (Continuous Linear Enhanced Assessment of RCC, or CLEAR score). Prognostic performance was evaluated in independent cohorts from The Cancer Genome Atlas (TCGA) (n = 414) and EMBL-EBI (n = 53), CLEAR score demonstrated both superior prognostic estimates and inverse correlation with anti-angiogenic tyrosine-kinase inhibition in comparison to previously proposed discrete subtyping classifications. Inverse correlation with high-dose interleukin-2 outcomes was also observed for the CLEAR score. Multiple somatic mutations (VHL, PBRM1, SETD2, KDM5C, TP53, BAP1, PTEN, MTOR) were associated with the CLEAR score. Application of the CLEAR score to independent expression profiling of intratumoral ccRCC regions demonstrated that average intertumoral heterogeneity exceeded intratumoral expression heterogeneity. Wider investigation of cancer biology using continuous approaches may yield insights into tumor heterogeneity; single cell analysis may provide a key foundation for this approach.

Figure 1

Validation of continuous linear enhanced assessment of ccRCC (CLEAR) by correlation and survival analysis. (A) Validation of CLEAR approach by correlation analysis with clinical variables. 265 ccRCC samples are ranked in ascending order based on CLEAR score. The top panel color bar represents the grade distribution profile of 265 ccRCC samples, expressed as colored vertical bars (white bars represent absent data). An association between high grade tumors and high CLEAR scores is observed (Fisher’s exact, p = 7.594e-07). A similar association between sarcomatoid RCC (SRCC) and high CLEAR scores is observed in the second horizontal bar (Mann-Whitney U test, p = 1.38e-06). To depict association with other key clinical variables, samples were evenly separated into 4 subgroups along the ranking queue with bar charts showing proportional breakdown of clinical variables. Correlation of tumor stage, tumor size with CLEAR score were observed (Fisher’s exact, p = 5.892e-05; Mann-Whitney U test, p = 1.10e-05, respectively). (B) Kaplan-Meier curves of cancer-specific survival for 265 ccRCC samples. We evenly grouped 265 ccRCC samples along the CLEAR scale into two (Q1 and Q2 groups), four (Q1–Q4 groups), six (Q1–Q6 groups) and eight (Q1–Q8 groups). In addition to inspection, Kaplan-Meier analysis was used to evaluate the association of these subgroups with survival, demonstrating a significant correlation with cancer-specific survival at all subgroupings.

Figure 2

Heatmaps of 18-transcript signature profiles from internal TCGA cohort. The top panel color bar represents the tumor grade distribution profile as determined by the 18-transcript signature. Heatmaps below the color bar are gene expression profiles from cohorts of internal 265 samples and 414 TCGA ccRCC samples. Samples are ranked in ascending order based on CLEAR score. We observe that signature genes show a consistent expression pattern in these two cohorts. A similar association between CLEAR score and tumor grade was observed.

Figure 3

Correlation of CLEAR score and mutation in TCGA cohort. (A) Individual correlation between tumor CLEAR score with tumor grade, stage and size distribution as well as mutation status are presented. The correlation of CLEAR score and TCGA grade, stage was investigated by ANOVA (p = 7.769e-09, p = 3.22e-08, respectively); the correlation of CLEAR score and TCGA sample size was estimated by spearman correlation (p = 4.252e-09). (B) Boxplots of CLEAR score distribution of samples with relevant gene mutations. The dashed line represents the median CLEAR score of the 414-sample set. Samples with mutated BAP1, TP53, PTEN and MTOR are associated with higher CLEAR scores (Mann–Whitney U test, *p ≤ 0.05, **p ≤ 0.001”).

Figure 4

Predictive values for CLEAR Score model versus prior models. The predictive value is represented by the likelihood ratio test and adequacy index. The adequacy index is a measure of the evaluated model (as a subset of the full model), represented as a % relative to the full model, which does include both subtype and CLEAR Score model). (A) Predictive value of CCA/B subtyping (110 transcripts) (Gulati et al.) and CLEAR score model for cancer-specific survival (CSS) using TCGA dataset. (B) Predictive value of CCA/B subtyping (34 transcripts) (Brooks et al.) and CLEAR score model for CSS using TCGA dataset. (C) Predictive value of sunitinib-responsive subtyping (Beuselinck et al.) and CLEAR Score model for progression free survival (PFS) using E-MTAB-3267 datasets. These three comparison result showed the addition of the CLEAR score model to one containing the CCA/B subtype model significantly improves the predictive value of the final model, but there is no significant difference when the subtype model is added to the CLEAR model.

Figure 5

Phylogenetic tree with driver mutation status and CLEAR score. The region tree provides the information of the connection of sample diversity, prognosis and the driver gene mutation. The region tree shows majority of sample regions are clustered together with their siblings with several exceptions caused by intra-tumor heterogeneity and PBRM1 is associated with regions with low prognosis score, while SETD2, BAP1 and KDM5C are associated with the poor prognosis sample regions.