Extensive genotype-phenotype heterogeneity in renal cell carcinoma - a proof-of-concept study

Abstract

Background: Renal cell carcinoma (RCC) is characterized by a high degree of genomic but also functional intratumoral heterogeneity (ITH). Mutations in VHL, chromatin remodeling genes such as SETD2 and genes that regulate the PI3K/AKT/mTOR pathway have been identified as recurrent drivers despite genomic ITH. Whether and to what extent these mutations shape functional ITH including the formation of spatial niches is incompletely understood. Herein, we analyze the correlation between mutational drivers and their functional proxies in a spatially defined manner.

Methods: A total of 23 RCCs were analyzed by panel next-generation sequencing followed by immunohistochemistry for five functional proxies for key genetic alterations including the expression of CD31, GLUT1, phospho-mTOR S2448, H3K36me3 and Ki-67. Antibody stainings were scored semiquantitatively in the tumor periphery and the tumor center.

Results: Unexpectedly, the presence of a VHL mutation was not found to correlate with its functional proxies including the expression of CD31/microvessel density or the expression of the glucose transporter GLUT1. Likewise, there was no correlation between the presence of activating mutations in genes of the PI3K/AKT/mTOR pathway and the expression of activated phospho-mTOR S2448. Furthermore, mutations in the methyltransferase gene SETD2 were not found to correlate with the expression level of its downstream target H3K36me3. Lastly, there was no correlation between the expression of the proliferation marker Ki-67 and the number of driver mutations.

Conclusion: This proof-of-concept study adds genotype-phenotype heterogeneity as additional layer of complexity to the known genomic and functional ITH in RCC.

Keywords: SETD2; VHL; genotype; mTOR; phenotype; renal cell carcinoma; tumor heterogeneity.

Figure 1

Overview of selected RCC mutational drivers and their downstream pathways. Functional proxies of mutational driver events relevant for the present study are shown in red.

Figure 2

Immunohistochemical detection of functional proxies of mutational driver events in RCC. Representative photomicrographs of immunohistochemical stainings for CD31, GLUT1, phospho-mTOR S2448, H3K36me3 and Ki-67 in the tumor periphery and the tumor center are shown. Positive controls include endothelial cells for CD31 (asterisk and arrow), red blood cells for GLUT1 and Caki-1 RCC cells grown as subcutaneous xenografts (Charles River Laboratories, Freiburg, Germany) for phospho-mTOR S2448, H3K26me3 and Ki-67. Scale bar for large panels = 100 µm.

Figure 3

Intratumoral spatial expression of CD31, phospho-mTOR S2448 and Ki-67 in RCC. (A–C) Quantification of the expression of CD31, phospho-mTOR S2448 and Ki-67 in a total of 23 primary tumors, local recurrences, and metastases. Tumor periphery and tumor center were assessed separately in primary tumors and local recurrences. Metastases were excluded from the spatial analysis since fundamentally different growth characteristics can be presumed. Each bar represents mean and standard error of at least four and up to a maximum of ten tumor areas. (D) Quantification of the average expression of the three functional proxies in 11 primary tumors with clear cell histology subdivided into tumor periphery and tumor center. HPF, high power field; CC, clear cell; PAP, papillary; CH, chromophobe; CDC, collecting duct carcinoma; IRS, immunoreactive score; FOV, field of view [1.23 mm²]. Asterisks indicate statistical significance: * p<0.05, **p<0.005, ***p<0.0005.

Figure 4

No spatial differences in the intratumoral expression of GLUT1 (A) and H3K36me3 (B). Graphic representation of the immunoreactive score (IRS) for GLUT1 and H3K36me3 per tumor. Because of the homogenous expression patterns, no subdivision into tumor periphery and tumor center was performed. CC, clear cell; PAP, papillary; CH, chromophobe; CDC, collecting duct carcinoma.

Figure 5

No correlation between mutational drivers and their functional proxies in RCC. Graphic representation of driver mutations and expression of five functional proxies in 23 RCCs. Metastases were excluded from the spatial analysis. Because of the non-heterogeneous expression patterns, no subdivision into tumor periphery and tumor center was performed for GLUT1 and H3K36me3. Color gradients reflect the protein expression levels from low to high. MVD, microvessel density; HPF, high power field, IRS, immunoreactive score; FOV, field of view [1.23 mm²]; CC, clear cell; PAP, papillary; CH, chromophobe; CDC, collecting duct carcinoma.