Molecular exploration of host-pathogen interactions in severe Pseudomonas aeruginosa infection through a multi-level data integration approach

Francesco Messina¹Claudia Rotondo¹Luiz Ladeira²Sara Crosetti¹Michele Properzi¹Valentina Dimartino¹Benedetta Riccitelli¹Bernard Staumont²Giovanni Chillemi³Liesbet Geris²Maria Grazia Bocci⁴Carla Fontana^1*

• ¹Laboratory of Microbiology and Biobank, Department of Epidemiology, Preclinical Research and Advanced Diagnostics, National Institute for Infectious Diseases Lazzaro Spallanzani (IRCCS), Rome, Italy
• ²Biomechanics Research Unit, GIGA-In silico Medicine, GIGA Institute, University of Liège, Liege, Liege, Belgium
• ³Department for Innovation in Biological, Agri-food and Forestry Systems, University of Tuscia, Viterbo, Lazio, Italy
• ⁴Intensive Care Unit, National Institute for Infectious Diseases “Lazzaro Spallanzani” IRCCS, 00149 Rome, Italy, Rome, Italy

Word count: 210 Understanding host-pathogen interactions is crucial for explaining the variability in sepsis outcomes, with Pseudomonas aeruginosa (PA) remaining a significant public health concern. In this work, we explored PA-human host interaction mechanisms through a data integration workflow, focusing on protein-protein and metabolite-protein interactions, along with pathway modulation in affected organs during severe infections. A scoping literature review enabled us to construct a domain-based infection network encompassing pathogenesis concepts, molecular interactions, and host response signatures, providing a wide view of the relevant mechanisms involved in severe bacterial infections. Our analysis yielded a literature-based comprehensive description of PA infection mechanisms and an annotated dataset of 189 PA-human interactions involving 151 proteins/molecules (109 human proteins, 3 human molecules, 34 PA proteins, and 5 PA molecules). This dataset was complemented with gene expression analysis from in vivo PA-infected lung samples. The results indicated a notable overexpression of proinflammatory pathways and PA-mediated modulation of host lung responses. Our comprehensive molecular network of PA infection represents a valuable tool for the understanding of severe bacterial infections and offers potential applications in predicting clinical phenotypes. Through this approach combining omics data, clinical information, and pathogen characteristics, we have provided a foundation for future research in host-pathogen interactions and the mechanistic grounds to build dynamic computational models for clinical phenotype predictions.