An Empirical Approach Leveraging Tumorgrafts to Dissect the Tumor Microenvironment in Renal Cell Carcinoma Identifies Missing Link to Prognostic Inflammatory Factors

Abstract

By leveraging tumorgraft (patient-derived xenograft) RNA-sequencing data, we developed an empirical approach, DisHet, to dissect the tumor microenvironment (eTME). We found that 65% of previously defined immune signature genes are not abundantly expressed in renal cell carcinoma (RCC) and identified 610 novel immune/stromal transcripts. Using eTME, genomics, pathology, and medical record data involving >1,000 patients, we established an inflamed pan-RCC subtype (IS) enriched for regulatory T cells, natural killer cells, T1 cells, neutrophils, macrophages, B cells, and CD8 T cells. IS is enriched for aggressive RCCs, including -deficient clear-cell and type 2 papillary tumors. The IS subtype correlated with systemic manifestations of inflammation such as thrombocytosis and anemia, which are enigmatic predictors of poor prognosis. Furthermore, IS was a strong predictor of poor survival. Our analyses suggest that tumor cells drive the stromal immune response. These data provide a missing link between tumor cells, the TME, and systemic factors. We undertook a novel empirical approach to dissect the renal cell carcinoma TME by leveraging tumorgrafts. The dissection and downstream analyses uncovered missing links between tumor cells, the TME, systemic manifestations of inflammation, and poor prognosis. .

Pan-cancer analysis identified inflamed microenvironment associated multi-omics signatures

“…Second, TCGA doesn't provide direct status of the tumor microenvironment. Therefore, we had to indirectly infer the relative score of immune composition as described in previous studies [13] and validated by another algorithm [14]. Third, our study provided a comprehensive catalogue of molecular alternations, but we couldn't further investigate the role of identified molecule for the lack of funding support and experimental environment.…”

“…Then, these genetic signatures of immune cells were as input for GSVA algorithm [12] to calculate the immune score of tumor samples. This algorithm has been validated in previous reports as an efficient way to evaluate the status of tumor microenvironment [13]. Considering the complexity of tumor microenvironment comprising various types of immune cells, we classified tumor samples into high immune-score(inflamed)/low immune-score(non-inflamed) based on their unsupervised clustering pattern of immune score.…”

“…Adjusted p-value <0.05 and log fold change >0 Classification of TME phenotypes across different tumor types To investigate the molecular characteristics of inflamed or non-inflamed immunophenotypes, immune cell signatures identified in above single cell RNA sequencing data were used as input for GSVA algorithm [12] to calculate immune score for each immune cell type. GSVA algorithm has been proven as an efficient way to reveal characteristics of tumor microenvironment [13]. Then, tumor samples were classified into low-immune score(non-inflamed), intermediate immune score and high immune score(inflamed) TME groups based on unsupervised clustering pattern of immune score by optCluster [33].…”

Overcoming Genetic Drop-outs in Variants-based Lineage Tracing from Single-cell RNA Sequencing Data

“…1b, we showed the VAFs of three representative non-chrM variants across all cells, where peaks can be easily discerned, which correspond to homozygous wildtype, heterozygous mutant, and homozygous mutant genotypes. We analyzed datasets from a variety of sequencing protocols, including SMART-Seq, 10x Genomics, mCelSeq2, Slide-Seq, Spatial Transcriptomics (ST), and scATAC-Seq 5,9,[11][12][13][14][15] . In Fig.…”

“…We generated the SMART-Seq datasets of 969 CD45 + cells from the tumor microenvironment of 5 kidney cancer (RCC) patients 12 . The SMART-Seq datasets of CML cells were created by Giustacchini et al 11 .…”