The origins of seven types of childhood and adult kidney cancers, including a number of rare subtypes, have been identified by a research team headed by scientists at the Wellcome Sanger Institute.
The investigators harnessed computational methods to analyze existing datasets and pinpoint the “cellular signals” given off by different cancers as they emerge. The results confirmed that childhood cancers are developmental in origin, occurring after errors in a particular developmental cell type’s journey to maturity. In contrast—and questioning the idea that adult, epithelial-derived kidney cancers revert to a fetal state—the findings indicated that adult kidney cancers emerged from mature cell types and in the vast majority of cases do not revert to a developmental pattern of gene expression.
The Wellcome Sanger Institute’s Matthew Young, PhD, commented, “It has long been assumed that childhood tumors have ‘fetal’ origins. Now we can replace that loose definition with a precise, quantitative measurement of the cellular signals that different tumor types exhibit. Our analysis also refutes the theory that adult tumors revert to a developmental state, unless they are a highly lethal subtype of adult kidney cancer.”
The team said their method holds promise as a tool for diagnosing patients with rare cancers. In their reported study, which is published in Nature Communications, they describe how one patient’s cryptic kidney cancer was identified as a Wilms-like tumor by looking at its cellular signals. First author Young and colleagues at the Wellcome Sanger Institute, Great Ormond Street Hospital (GOSH), the Princess Máxima Center for Pediatric Oncology, and Oncode Institute, reported on their studies in a paper titled, “Single cell derived mRNA signals across human kidney cancers.”
All cancers are derived from normal cells that have started to multiply uncontrollably. By comparing patterns of gene expression in cancer and normal cells, it is possible to learn about aspects of each tumor’s origin and behavior. This type of analysis has been made possible by the advent of single-cell mRNA sequencing, a high-resolution technology that can identify different cell types present in a tissue according to the genes expressed by individual cells. In recent studies, researchers have identified the origins of individual childhood cancers, such as neuroblastoma, using mRNA single-cell sequencing on small numbers of tumors.
Previous studies have used these techniques to compare normal and diseased tissue in some of the most common kidney cancers, but to conduct single-cell sequencing on many hundreds of tumors would not be achievable. For their newly reported work, researchers at the Wellcome Sanger Institute and their collaborators applied computational analysis of existing data to determine the origin of a larger groups of childhood cancers.
They used their approach to mine Human Cell Atlas (HCA) reference data, and databases of tumor gene expression. They assessed mRNA signals in 1,300 childhood and adult renal tumors, spanning seven different tumor types, to investigate the origins of these cancers.
The seven tumor types included:
- Congenital mesoblastic nephroma (CMN)
- Nephroblastoma (also known as Wilms tumor)
- Clear cell sarcoma of the kidney (CCSK)
- Malignant rhabdoid tumor of the kidney (MRTK)
- Clear cell renal cell carcinoma (ccRCC)
- Papilliary renal cell carcinoma (pRCC), subtypes type 1 and type 2
- Chromophobe renal cell carcinoma (ChRCC), subtype “Metabolically divergent ChRCC”
The study results indicated that childhood tumors, but not adult tumors, exhibited a fetal transcriptome. “A significant developmental signal was absent from almost all adult tumors,” they wrote. This suggests that global “dedifferentiation” to a developmental state does not generally occur in adult kidney tumors. And importantly, when there was transcriptional evidence of dedifferentiation in adult tumors, “… it conferred a dismal prognosis.”
Each cancer type was found to exhibit unique ‘cellular signals,’ or patterns of gene expression, that could be used to classify them in future. “A further finding of our study was that within each category, the majority of tumors exhibited remarkably uniform cellular signals,” the team noted. “That is, despite a high diversity in clinical outcome, tumors of the same type almost universally had the same dominant cellular signal … Therefore, cellular signals of renal tumors may lend themselves as diagnostic adjuncts ….”
The findings also shed light on the behavior and origins of some kidney tumor subtypes—mesoblastic nephroma, clear cell sarcoma of the kidney, malignant rhabdoid tumor of the kidney, and chromophobe renal cell carcinoma—whose rarity would have made it difficult to examine otherwise. The method pioneered by the researchers in addition helped to classify one patient’s tumor, which clinicians had been unable to diagnose fully.
Karin Straathof, PhD, co-senior author of the study, at Great Ormond Street Hospital, said: “Sometimes it is not possible to fully diagnose childhood kidney cancers via the usual methods, which can impact our ability to adopt the best course of treatment. One of the samples used in this study was from a child with one of these undiagnosed tumors. But by analyzing the genes expressed by the tumor cells, we were able to recognize it as Wilms’ tumor. My hope is that this approach can be used in such cases in future.”
Sam Behjati, Wellcome Sanger Institute co-senior author of the study, added, “Not only does this computational approach using existing datasets validate previous results on the origins of childhood kidney cancers, it provides a new way of expanding this research to much larger numbers of tumors and rare cancer types. I believe that the success of this approach could act as a blueprint for investigating the behavior and origins of the entire spectrum of human cancer.” The investigators concluded in their paper, “Our findings provide a cellular definition of human renal tumors through an approach that is broadly applicable to human cancer.”
The findings could have therapeutic, as well as diagnostic applications, they suggested. “Moreover, the cellular transcriptome itself may represent a therapeutic target that transcends individual patients, if we had tools available to manipulate transcription in a predictable manner. This may be a particularly attractive approach for targeting transcriptional states of fetal cells retained in childhood cancer that are absent from normal postnatal tissues.”