Metabolic Modeling of Cystic Fibrosis Airway Communities Predicts Mechanisms of Pathogen Dominance

Abstract

5Cystic fibrosis (CF) is a fatal genetic disease characterized by chronic lung infections due to aberrant 1 6 mucus production and the inability to clear invading pathogens. The traditional view that CF infections 1 7 are caused by a single pathogen has been replaced by the realization that the CF lung usually is colonized 1 8 by a complex community of bacteria, fungi and viruses. To help unravel the complex interplay between 1 9 the CF lung environment and the infecting microbial community, we developed a community metabolic 2 0 CF lung metabolomics and 16S sequencing could provide important insights into disease progression and 3 2 treatment efficacy. 3 3 Importance 3 4Cystic fibrosis (CF) is a genetic disease in which chronic airway infections and lung inflammation result 3 5 in respiratory failure. CF airway infections are usually caused by bacterial communities that are difficult 3 6 to eradicate with available antibiotics. Using species abundance data for clinically stable adult CF patients 3 7 assimilated from three published studies, we developed a metabolic model of CF airway communities to 3 8 better understand the interactions between bacterial species and between the bacterial community and the 3 9 lung environment. Our model predicted that clinically-observed CF pathogens could establish dominance 4 0 over other community members across a range of lung nutrient conditions. Heterogeneity of species 4 1 abundances across 75 patient samples could be predicted by assuming that sample-to-sample 4 2 heterogeneity was attributable to random variations in the CF nutrient environment. Our model 4 3 predictions provide new insights into the metabolic determinants of pathogen dominance in the CF lung 4 4 and could facilitate the development of improved treatment strategies. 4 5 4 6shaping community composition and behavior, and the impact of community composition on the efficacy 7 1 of antibiotic treatment regimens. 7 2In silico metabolic modeling has emerged as a powerful approach for analyzing complex microbial 7 3 communities by integrating genome-scale reconstructions of single-species metabolism within 7 4 mathematical descriptions of metabolically interacting communities (11, 12). Modeled species 7 5 interactions typically include competition for host-derived nutrients and cross-feeding of secreted 7 6 byproducts such as organic acids, alcohols and amino acids between species (13, 14). Due to challenges 7 7 in developing manually curated reconstructions of poorly studied species, including those present in the 7 8 CF lung, most in silico community models have been restricted to ~5 microbial species (15-17) and fail to 7 9adequately cover the diversity of in vivo communities. This limitation can be overcome in bacterial 8 0 communities by using semi-curated reconstructions developed through computational pipelines such as 8 1 the ModelSeed (18), AGORA (19) and other methods (20). Given the availability of suitable single-strain 8 2 metabolic reconstructions, a number of alternative methods have been develo...

Community composition shapes microbial-specific phenotypes in a cystic fibrosis polymicrobial model system

“…1D , red bars). Interestingly, we could not detect P. melaninogenica in monoculture, a finding predicted in our previous metabolic modeling study (44), but ∼6 log 10 CFU/mL of this microorganism could be detected when cultivated in the presence of other microbial partners in both biofilm and planktonic fractions ( Fig. 1D , green bars).…”

“…Furthermore, the average abundance of Burkholderia was skewed by a relatively small number of patients in our cohort that were dominated by this pathogen (21). (vi) Finally, work published by our group using metabolic modeling indicates that P. aeruginosa, S. aureus, S. sanguinis and P. melaninogenica are top contributors of cross-fed metabolites in communities detected in the CF lung (44). That is, their abundance in the CF airway could be explained by predicted metabolic cross-feeding among these four microorganisms.…”

An Innovative Protocol for Metaproteomic Analyses of Microbial Pathogens in Cystic Fibrosis Sputum

“…The competition for nutrients within the CF airways is an important selective pressure influencing the composition of the CF community. For example Pseudomonads and Streptococci are able to efficiently utilize amino acids, organic acids and alcohols (partly produced by other community members) leading to high growth rates within the lung (Yang et al, 2008;Henson et al, 2019;La Rosa and Molin, 2019). Notably, our metabolome data revealed various amino acids (and organic acids) within all three sputum samples (Figure S6),which supports that amino acids are the major carbon-and nitrogen source in CF sputum (Palmer et al, 2005;La Rosa and Molin, 2019).…”

Integration of constraint-based modeling with fecal metabolomics reveals large deleterious effects of Fusobacterium spp. on community butyrate production

“…26,27 COBRA represents a scalable systems biology computational modeling approach that is widely applied in the field of microbiome research. [28][29][30] Its strengths of integrating genomic data with condition-specific constraints are specifically designed to deliver on the task of characterizing the metabolic functions of microbial communities. 31 Importantly, in contrast to pure statistical approaches, COBRA allows for the mechanistic examination of microbe-metabolite relations within one community.…”