Metabolic modeling for Microbial Antibiotic-Resistant studies

Nidhi Grover

Learning objectives:

1. Describe what Genomic Scale models (GEMs) is and how they are built.

2. Explain How pangenome metabolic models could help fight the microbial antibiotic resistance

3. Describe how metabolic models can help to fight the microbial antibiotic resistance

Figure 1: Graphic Abstract Applications of GEMs for microbial antibiotic Resistance studies. The figure in the gray area explains the process to build GEMs (Genome-Scale Models), A. after the new generation sequences are processed into Kbase, algorithms lookout to associate gene-protein-phenotype using databases as NCBI, GenBank, and KEEG. B. From those databases is also extracted information as energy requirements for each of the metabolic reactions, stoichiometry yields, localization in cell compartments, and define some reactions that are essential or constrain microbial growth function. C. Each of the metabolic reactions listed are represented in two matrices substrates and fluxes, afterward D. It became a mathematical linear problem where an objective function can be optimized (FBA) in this case is optimal growth. E. the FBA contains the entire list of reactions of the microorganism or the tissue, analysis of the merged FBA would allow for the simulation of the microenvironment of tissue and the mechanism (biochemical reactions) that are involved in the process of infection. F. Part of the simulation of the microenvironment in antibiotic resistance is to evaluate the effect of the carbon source of the host in the lethality of the antibiotic. G. Shows the mechanism of how a microorganism can resist the effect of certain antibiotics: Limiting uptake of the antibiotic, modification of the molecule, or efflux transport. (Note: Graphical abstract was made in Powerpoint, Figure E, F and G from Biorender and icons are available online). H GEMs help to identify metabolic pathways in the resistant pathogen that can be targeted for the development of new antibiotics.

The accessibility of new generation and high throughput technologies sequences has made available the genomic data of thousands of microorganisms. Then, is necessary to ask how this genomic data can be used to change the way how we perform microbiological and clinical studies. The idea of modeling the biochemical potential of microorganisms mathematically is not new, for instance back in time there were two types of metabolic models: Kinetic models and Stoichiometric models. Kinetic models considered the velocity of reactions of microorganisms, growth rates; consumption of substrate, one of the most known kinetic models is Monod’s Model, which is constructed based on the measurement of experimental parameters. The stoichiometric model described the metabolism in of microorganisms s a system of linear equations and fluxes, and is based on the assumption of steady-state, a known example of a stoichiometric model is the FBA (Flux Balance Analysis) (Krömer et al., 2014). Nevertheless, beyond a single repository database listing enzymes and reactions codified for a particular gene, the application of Flux Balance Analysis (FBA) allows for evaluating the in-silico response of a microorganism to a particular drug. It has allowed evaluation of the effect of type and concentration of carbon source in growth rate on E. coli models on the lethality of the antibiotic (Wareth et al., 2021) and the influence of metabolic mutations can increase resistance to the antibiotic.

Genome-Scale metabolic model (GEM) represents a microorganism by the collection of all metabolites, gene-reaction- protein network, the metabolites that are compiled in the matrix S, for the metabolites and the reaction matrix. GEMs can represent an entire microorganism and even a whole tissue (Zhang and Hua, 2016).

Finding alternatives to threat microbes that are resistant to known antibiotics might be accelerated by using Genome-Scale models. Different pathogens have been already evaluated using GEMs among them Mycobacterium tuberculosis, Acetinobacter baumannii, Pseudomona aeruginosa, E. coli, Helicobacter pylori have already reconstructed GEM (WHO) because its study is of global interest (Ag and Dale, 2021).

More than a million people died from tuberculosis in 2020 (WHO, and it is the second more deadly infectious disease after COVID-19. Despite tuberculosis is a treatable disease Multidrug-resistant TB (MDR-TB) has emerged by inappropriate use of antibiotic treatments of tuberculosis, therefore, the bacteria do not respond to regular treatment leading to the use of stronger drugs and longer treatments. In certain cases, TB patients do not respond either to second-line TB drugs condemning patients to live without treatment. In particular, having a model to study the MDR-TB is relevant because the efficacy of the TB treatment is affected by the heterogeneity of both host and bacteria(WHO, 2020). There are at least six different TB strains, and the mechanism of resistance is not fully understood due the wide number of mutations in resistant strains isolated from clinical samples suggesting its relationship with high-level of resistance. GEMs might help to identify strain-specific metabolic pathways linked to pathogenicity and resistance (Pearcy et al., 2021) and might also help to understand how a strain becomes resistant to multiple drugs.

How do metabolic models help develop new drugs or find treatments for pathogens that are resistant to antibiotics?

When a pathogen’s genome is sequenced and annotated, the majority of enzymes and reactions from the core metabolism are represented as well as the pathways from the pangenome used for each particular strain to infect, develop resistance and “interact” with its host. The sequencing and curation of the pangenome of bacterial strains is helping to disclose the antibiotic resistance due to small differences within species that can change the response of strain to antibiotics and that a large number of metabolic reactions might influence the susceptibility to the drug (Ferretti et al., 2018).

The construction of Genomic-scale models using sequences from Whole- generation sequencing to reconstruct the entire metabolic pathways whether of a single microorganism or the entire microbiome present in a host or an environmental sample. (Cuevas et al., 2016). Most of the clinical and microbiological studies of antibiotics are based in test on the Minimum Inhibitory concentration (MIC) that are incomplete because rely in phenotype and lack the connection between genotype and phenotype. Identifying inhibitors of the essential metabolic reactions in the pathogen can be used to develop potential drug targets, to study the combined effect of multiple drugs, and finally it provides a better understanding of the infectious process from a metabolic point. For instance, some pathogens might change their infectious mechanism depending on the microenvironment of the host. GEMs will include the complete list of potential biosynthetic pathways involved in the pathogenicity or survival of the microorganism that can be targeted with new molecules that additionally has little or no harm to the host.

The construction of metabolic models GEMs helps researcher of Microbial Antibiotic Resistance (AMR) in different ways:

GEMs can be used with Essential Analysis (EA) through FBA that determine the knockout reactions that will be deleterious for the pathogen and also allows the combined effect in silico (testing the biomass objective function) of using multiple medicines (Zhang and Hua, 2016).

Determining the components of the microbial metabolism involved in antibiotic tolerance. A different way is to evaluate how the microenvironment of the host might have effects on susceptibility of the microorganism to antibiotic treatment. How is their interaction through metabolic pathways? For instance, the microorganism can acquire resistance through four generalized mechanisms: Modification of the drug target, inactivation of a drug, limiting uptake of a drug, and finally drug efflux, which is a mechanism of transport that can be activated by the type of carbon source available within the host. Studies demonstrated there is a link between antibiotic resistance and metabolism, this relationship should be considered during the process of antibiotic design.

Questions:

What are some limitations of the MEGs applied to microbial antibiotic resistance studies

It is advised to be cautious with the conclusions raised from FBA models for several reasons, as mentioned it assumes a steady state. Second, the solution of the model is a solution space, which need further experiments to narrow down the constrains of the model. The second reason to be cautious are with concerns about the quality of the model, whether build up by the user or the ones publically available. There are currently tools (MEMOTE) which along with manual curing with information from database would help to improve the quality of the model if there are real concerns.

How do you think further improvements or applications of MEGs to fight AMR

GEMs will play an essential role simulating interactions between the host, multiple pathogens or microorganism present in the gut of humans that might play a role in the mechanism of resistance as product of horizontal or vertical transfer. Thus, evaluation of resistance in a community context will be key to unveil the interaction between microbiota and human tissues in the antibiotic resistant and finding new active compounds. Recent studies also show how the integrations of different omics data to the GEMs models and machine learning might increase the spectra of response and

What is pangenome and why do you think metabolic modeling should consider pangenome analysis for a better understanding of microbial antibiotic resistance?.

Pangenome is the union or entire set of genes from all strains within a clade. The core pangenome comprehend the common genes within the clade. The shell pangenome that represent genes present in two or more strains and accessory pangenome genes that are present just in a single strain. Is important to consider small differences can lead to specific metabolic response to antibiotic resistance within a same clade, it might also uncovered unknown resistance mechanism that has been developed or acquired.

References

Ag, B., and Dale, G. (2021). innovation with similar or higher efficacy. 8–9.

Cuevas, D.A., Edirisinghe, J., Henry, C.S., Overbeek, R., O’Connell, T.G., and Edwards, R.A. (2016). From DNA to FBA: How to build your own genome-scale metabolic model. Front. Microbiol. 7, 1–12.

Ferretti, P., Pasolli, E., Tett, A., Asnicar, F., Gorfer, V., Fedi, S., Armanini, F., Truong, D.T., Manara, S., Zolfo, M., et al. (2018). Mother-to-Infant Microbial Transmission from Different Body Sites Shapes the Developing Infant Gut Microbiome. Cell Host Microbe 24, 133-145.e5.

Krömer, J.O., Nielsen, L.K., Editors, L.M.B., and Walker, J.M. (2014). Metabolic Flux Analysis IN Series Editor.

Pearcy, N., Hu, Y., Baker, M., Maciel-Guerra, A., Xue, N., Wang, W., Kaler, J., Peng, Z., Li, F., and Dottorini, T. (2021). Genome-Scale Metabolic Models and Machine Learning Reveal Genetic Determinants of Antibiotic Resistance in Escherichia coli and Unravel the Underlying Metabolic Adaptation Mechanisms. MSystems 6.

Wareth, G., Neubauer, H., and Sprague, L.D. (2021). A silent network’s resounding success: how mutations of core metabolic genes confer antibiotic resistance. Signal Transduct. Target. Ther. 6, 1–2.

WHO (2020). WHO consolidated guidelines on tuberculosis. Module 4: treatment - drug-resistant tuberculosis treatment. Online annexes.

Zhang, C., and Hua, Q. (2016). Applications of genome-scale metabolic models in biotechnology and systems medicine. Front. Physiol. 6, 1–8.