ABSTRACT
Rationale: Renal cell carcinoma (RCC) accounts for about 2% of all adult cancers, and clear cell RCC (ccRCC) is the most common RCC histologic subtype. A hallmark of ccRCC is the loss of the primary cilium, a cellular antenna that senses a wide variety of signals. Loss of this key organelle in ccRCC is associated with the loss of the von Hippel-Lindau protein (VHL). However, not all mechanisms of ciliopathy have been clearly elucidated. Methods: By using RCC4 renal cancer cells and patient samples, we examined the regulation of ciliogenesis via the presence or absence of the hypoxic form of the voltage-dependent anion channel (VDAC1-ΔC) and its impact on tumor aggressiveness. Three independent cohorts were analyzed. Cohort A was from PREDIR and included 12 patients with hereditary pVHL mutations and 22 sporadic patients presenting tumors with wild-type pVHL or mutated pVHL; Cohort B included tissue samples from 43 patients with non-metastatic ccRCC who had undergone surgery; and Cohort C was composed of 375 non-metastatic ccRCC tumor samples from The Cancer Genome Atlas (TCGA) and was used for validation. The presence of VDAC1-ΔC and legumain was determined by immunoblot. Transcriptional regulation of IFT20/GLI1 expression was evaluated by qPCR. Ciliogenesis was detected using both mouse anti-acetylated α-tubulin and rabbit polyclonal ARL13B antibodies for immunofluorescence. Results: Our study defines, for the first time, a group of ccRCC patients in which the hypoxia-cleaved form of VDAC1 (VDAC1-ΔC) induces resorption of the primary cilium in a Hypoxia-Inducible Factor-1 (HIF-1)-dependent manner. An additional novel group, in which the primary cilium is re-expressed or maintained, lacked VDAC1-ΔC yet maintained glycolysis, a signature of epithelial-mesenchymal transition (EMT) and more aggressive tumor progression, but was independent to VHL. Moreover, these patients were less sensitive to sunitinib, the first-line treatment for ccRCC, but were potentially suitable for immunotherapy, as indicated by the immunophenoscore and the presence of PDL1 expression. Conclusion: This study provides a new way to classify ccRCC patients and proposes potential therapeutic targets linked to metabolism and immunotherapy.
Subject(s)
Carcinoma, Renal Cell , Cilia , Kidney Neoplasms , Voltage-Dependent Anion Channel 1/physiology , Von Hippel-Lindau Tumor Suppressor Protein/metabolism , Adult , Aged , Aged, 80 and over , Carcinoma, Renal Cell/metabolism , Carcinoma, Renal Cell/pathology , Cell Line, Tumor , Cilia/metabolism , Cilia/pathology , Cohort Studies , Epithelial-Mesenchymal Transition , Female , Humans , Kidney Neoplasms/metabolism , Kidney Neoplasms/pathology , Male , Middle Aged , Young AdultABSTRACT
BACKGROUND: There are no validated markers that predict outcome in metastatic renal cell cancer (mRCC) patients treated with sunitinib. Recently, single nucleotide polymorphism (SNP) rs9582036 in VEGFR1 has been proposed as a predictor of progression-free survival (PFS) and overall survival (OS) to bevacizumab in patients with pancreatic cancer and rs7993418 in VEGFR1 as predictor for PFS in mRCC-patients treated with bevacizumab. Here, we aim to study the impact of these SNPs in mRCC patients treated with sunitinib. METHODS: We included patients with mRCC treated in 15 institutions in France and Belgium. Patients received sunitinib as first-line targeted therapy. We assessed response, time-to-tumor progression (TTP), OS, and clinical and biochemical parameters associated with outcome. We genotyped rs9582036 and rs7993418 as well as three other surrounding SNPs in VEGFR1: rs9554320, rs9554316 and rs9513070. Association between SNPs and treatment outcome were studied by univariate analysis and by multivariate Cox regression using relevant clinical factors associated with TTP and OS as covariates. FINDINGS: Ninety-one patients were included. We found that mRCC patients with the CC-variant in rs9582036 in VEGFR1 have a poorer response rate (RR) (0% vs. 46%, p = 0.028), a poorer PFS (10 vs. 18 months, p = 0.033 on univariate and 0.06 on multivariate analysis) and a poorer OS (14 vs. 31 months, p = 0.019 on univariate and 0.008 on multivariate analysis) compared to patients with the AC- and AA-genotypes. mRCC patients with the AA-variant in rs9554320 in VEGFR1 have a poorer PFS (12 vs. 21 months, p = 0.0066 on univariate and 0.005 on multivariate analysis) and a poorer OS (22 vs. 34 months, p = 0.019 on univariate and 0.067 on multivariate analysis) compared to patients with the AC- and CC-genotypes. Interpretation. mRCC patients with the CC-genotype in VEGFR1 SNP rs9582036 have a poorer response rate, PFS and OS when treated with sunitinib. These findings are in agreement with the association of rs9582036 and outcome observed in bevacizumab treated pancreatic cancer patients. Prospective validation of this SNP is warranted.