Prioritization of mechanisms and potential biomarkers in inflammatory and malignant diseases

Despite significant advancements in biomedical research, diseases such as inflammatory and malignant conditions remain challenging to diagnose and treat effectively. These diseases exhibit shared underlying complexities, including the involvement of thousands of genes across various organs and cell types. This complexity hinders the identification and prioritization of reliable biomarkers and drug targets. Single-cell RNA sequencing (scRNA-seq) and multi-omics analyses have emerged as powerful tools to address these challenges. Specifically, analyzing cell-cell communication (CCC) using scRNA-seq data enables prioritization of upstream regulators (URs), which represent higher-order molecular mechanisms and may serve as key drivers in various diseases. However, systematic approaches are needed to identify and validate these URs as clinically relevant biomarkers.

The overarching aim of this PhD thesis was to systematically prioritize regulatory mechanisms in inflammatory and malignant diseases. By analyzing multi-omics data, including scRNA-seq, spatial transcriptomics, proteomics, and clinical datasets, we aimed to: 1) characterize complex gene expression changes using transcriptional programs and URs in inflammatory diseases; 2) extend this approach to prioritize biomarkers in pancreatic cancer; 3) identify URs shared across multiple cancer types; and 4) evaluate the clinical utility of protein biomarkers in tissue and blood for cancer diagnosis.

In this PhD thesis, we demonstrated that higher-order structures (transcriptional programs and URs) could be identified in immune-mediated inflammatory diseases (IMIDs), revealing a graded on/off system in immune regulation across multiple organs (Study I). These higher-order structures could be extended to pancreatic cancer, uncovering URs associated with two transcriptional programs linked to cancer risk and severity (Study II). Further analysis in Study III identified shared URs (shared-URs) across various cancer types that were associated with survival outcomes. This significant association of the prioritized URs with clinical traits including risk of disease and survival in large independent cohorts supported the general relevance of this approach. Moreover, Study IV revealed high diagnostic accuracy for tissue-derived proteins but limited predictive value for blood-based biomarkers using a machine learning approach.

In conclusion, this thesis offers insights into disease mechanisms and potential biomarkers for diagnosis and prognosis, utilizing both systematic computational analyses based on biological knowledge and machine-learning (ML) approaches. These approaches may contribute to developing a scalable framework to identify biomarkers and drug targets in inflammatory and malignant diseases.

List of scientific papers

I. Sandra Lilja*, Xinxiu Li*, Martin Smelik*, Eun Jung Lee, Joseph Loscalzo, Pratheek Bellur Marthanda, Lang Hu, Mattias Magnusson, Oleg Sysoev, Huan Zhang, Yelin Zhao, Christopher Sjöwall, Danuta Gawel, Hui Wang, Mikael Benson. Multi-organ single-cell analysis reveals an on/off switch system with potential for personalized treatment of immunological diseases. Cell Rep Med. 21;4(3):100956 (2023). https://doi.org/10.1016/j.xcrm.2023.100956

II. Yelin Zhao, Martin Smelik, Xiaojing An, Daniel Diaz-Roncero Gonzalez, Xinxiu Li, AKM Firoj Mahmud, Oleg Sysoev, Dina Mansour Aly, Mikael Benson. Single-cell and Spatial Transcriptomics Reveal Fibroblast Related Program Linked to Pancreatic Cancer Risk in the UK Biobank Cohort. [Manuscript]

III. Yelin Zhao, Xinxiu Li, Joseph Loscalzo, Martin Smelik, Oleg Sysoev, Yunzhang Wang, A. K. M. Firoj Mahmud, Dina Mansour Aly, Mikael Benson. Transcript and protein signatures derived from shared molecular interactions across cancers are associated with mortality. J Transl Med 22, 444 (2024). https://doi.org/10.1186/s12967-024-05268-7

IV. Martin Smelik, Yelin Zhao, Dina Mansour Aly, AKM Firoj Mahmud, Oleg Sysoev, Xinxiu Li, Mikael Benson. Multiomics biomarkers were not superior to clinical variables for pan-cancer screening. Communications Medicine. [Accepted]

*Shared first-author