Researchers from the National Institute of Health’s Clinical Proteomic Tumor Analysis Consortium (CPTAC) and collaborating institutions have developed a novel protein expression-based stemness index (PROTsi) to evaluate oncogenic dedifferentiation in cancer. Their work, published in Cell Genomics, aims to quantify stem-like features in cancer which are closely linked to aggressive tumor behavior, resistance to therapies, and poor patient outcomes.
In order to quantify stem-like features at the protein level, the team trained a machine-learning model on proteomic data from human induced pluripotent stem cells. As a result, the model was able to identify molecular patterns associated with stemness and calculate a quantitative PROTsi score (ranging from zero to one) for any given sample.
The team used their model to analyze CPTAC pan-cancer proteogenomic data, including 1,300 samples spanning eleven tumor types. This allowed researchers to pinpoint specific proteins, post-translational modifications, and activated kinase pathways strongly correlated with high PROTsi scores.
Proteins strongly correlated with high PROTsi scores were evaluated for their potential as prognostic biomarkers, including survival analysis and independent IHC validation. Researchers found that expression of specific stemness-associated proteins as well as overall PROTsi score both had significant predictive value with respect to tumor aggression and clinical outcomes across cancer types.
Researchers also conducted a computational drug connectivity analysis for seven CPTAC tumor types, selecting sets of 100 proteins most strongly or negatively correlated with PROTsi score. Their analysis revealed a number of existing compounds that may repress stemness-associated pathways in specific cancers, suggesting avenues for drug development or repurposing.
Dr. Tathiane M. Malta, study leader and Assistant Professor from the University of São Paulo, Brazil, emphasized the significance of their findings, stating “Once validated, these biomarkers can support clinical decision-making and the management of cancer patients. Moreover, many of these proteins represent potential targets for new therapeutic strategies, contributing to more effective anti-tumor treatments.”
The development and application of PROTsi in this study represents a step forward in our understanding of cancer as well as a strong foundation for further study. Looking ahead, Dr. Malta commented, “In addition to identifying new biomarkers and potential therapeutic targets, including targets shared across different tumor types, this work also serves as a rich source of data that can be explored by the scientific community to investigate specific questions that go far beyond what we present here.”