Identification of gene signature for treatment response to guide precision oncology in clear-cell renal cell carcinoma

Abstract

Clear-cell renal cell carcinoma (ccRCC) is a common therapy resistant disease with aberrant angiogenic and immunosuppressive features. Patients with metastatic disease are treated with targeted therapies based on clinical features: low-risk patients are usually treated with anti-angiogenic drugs and intermediate/high-risk patients with immune therapy. However, there are no biomarkers available to guide treatment choice for these patients. A recently published phase II clinical trial observed a correlation between ccRCC patients' clustering and their response to targeted therapy. However, the clustering of these groups was not distinct. Here, we analyzed the gene expression profile of 469 ccRCC patients, using featured selection technique, and have developed a refined 66-gene signature for improved sub-classification of patients. Moreover, we have identified a novel comprehensive expression profile to distinguish between migratory stromal and immune cells. Furthermore, the proposed 66-gene signature was validated using a different cohort of 64 ccRCC patients. These findings are foundational for the development of reliable biomarkers that may guide treatment decision-making and improve therapy response in ccRCC patients. Clear-cell renal cell carcinoma (ccRCC) tumors have been reported to be highly angiogenic and with immunosuppressive features 1,2. Recent publications show increased expression of the immune inhibitory ligand and receptors (PD-L1/CTLA4) on tumor cells and/or tumor-infiltrating immune cells 3,4. Currently, tumor mutation burden is considered a predictive biomarker for response to immune checkpoint inhibitors (ICIs). However, research studies have shown that ccRCC has low mutational burden but highest immune infiltration score compared to other cancer types 5,6. In a different study, a pan-cancer analysis found renal cell carcinomas (RCC) to have the highest proportion of indel mutations, which can increase tumor neoantigen abundance 7. These anomalies in ccRCC make it the perfect platform to study dynamic biomarkers. ccRCC patients with clinically localised tumor undergo partial or radical nephrectomy, but ~30% of the patients present with de novo metastatic disease 8. Metastatic patients are usually treated with systemic therapies based on the clinical features 9. The prognostic value of different risk stratification tools is limited to clinical and pathological features of the patients 10. In Canada, International Metastatic Renal Cell Carcinoma Database Consortium (IMDC) risk model is applied with six clinical and laboratory factors: Karnofsky performance status, time of first-line targeted therapy from diagnosis, haemoglobin concentration, serum calcium concentration, neutrophil and platelet counts 9. According to the IMDC risk stratification, low-risk metastatic RCC patients are usually treated with anti-angiogenic tyrosine kinase inhibitors (TKIs) and intermediate/high-risk patients with ICIs 11. Risk stratification models based on gene expression pattern (both messenger and long non-coding RN...

Subgroup-Independent Mapping of Renal Cell Carcinoma—Machine Learning Reveals Prognostic Mitochondrial Gene Signature Beyond Histopathologic Boundaries

“…Previous functional analyses on RCC mainly focused on isolated gene signatures, which were characteristic and prognostic for single histopathologic subgroups (4, 5, 20, 21)-e.g., ClearCode34 (22) for determining the individual risk of recurrence in localized ccRCC. Moreover, researchers aimed to identify biomarkers and gene networks predictive of future therapy response-especially for angiogenesis inhibition, tyrosine kinase inhibition (TKI) and immune checkpoint blockade (23)(24)(25)(26)(27). Interestingly, a recent study was able to discriminate ccRCC and pRCC samples originating from proximal tubules of the nephron from chRCC specimen originating from distal tubules based on the metabolic and lipidomic profile of the samples (28).…”

A Molecularly Characterized Preclinical Platform of Subcutaneous Renal Cell Carcinoma (RCC) Patient-Derived Xenograft Models to Evaluate Novel Treatment Strategies

“…RNAseq raw count data were transformed to gene length corrected trimmed mean of M-values (GeTMM) ( 31 ) to perform single-sample gene set enrichment analyses (ssGSEAs) ( 32 ) ( ) regarding cancer hallmarks and angiogenesis/hypoxia-related reactome pathways ( 33 ). Gene sets to indicate the angiogenic or T-effector tumor type were generated with a recently published 66-gene signature for RCC ( 34 ).…”

The genomic and transcriptomic landscape of advanced renal cell cancer for individualized treatment strategies

“…Track C shows unsupervised clustering on the rows (patients) with color coding indicating the RCC subtype (purple for ccRCC and pink for pRCC) and colored according to Z-scores. The x-axis has been cut into several gene groups related to angiogenesis, invasion, Ca 2+- flux and T-effector cells, as defined by D’Costa et al 25 and by their stated order.…”

“…RNA-Seq could be applied to distinguish ccRCC, pRCC and histologically undefined RCC based on the differential gene expression. The application of the 66-gene signature 25 on the RNA-Seq data, made it possible to sub-categorize ccRCC into immunogenic or angiogenic signatures, whereas classification in pRCC using these signatures was not feasible.…”