Carlos Olivares (1), Charles Burdet (1), Ariane Amoura (1), Imane El Meouche (1), Emmanuelle Comets(1,2)
(1) Université Paris Cité, IAME, Inserm, 75018 Paris, France, (2) Univ Rennes, Inserm, EHESP, Irset - UMRS 1085, 35000 Rennes, France
Objectives: Urinary tract infections (UTIs) are the second most widespread infections globally and present high recurrence rates, with Escherichia coli (E. coli) the most prevalent cause of infection [1]. Understanding the survival mechanisms and prevalence of these pathogens is also of major importance to avoid the appearance and proliferation of resistant bacteria. The objective of this work is to estimate the dynamics of the colonisation of bacteria in the urinary tract in a murine UTI model in the presence and absence of antibiotic treatment.
Methods:
A total of 100 eight week-old, 20g, healthy mice were infected with either a commensal E. coli strain (UTI) or a strain coming from UTI in patients (UTI89, CFT and NILS69) on day 0 with an inoculum of ~10^8 CFU injected in bladder. All strains colonise digestive and urinary tract [2-4] . Each group was divided in one control arm (Ct) and two arms treated by ciprofloxacin (2.5 and 10 mg/kg twice a day for 2 days), starting on day 2 post-infection. In the PAS control arm, 13, 17, 18, 5, 3, 9, and 11 mice were sacrificed at days 1, 2, 4, 6, 10, 16, and 22 post-infection, respectively. The mice treated with antibiotics (17 and 7 respectively for the 2.5 and 10 mg/kg dose) were sacrificed on day 4. Number of colony forming units (CFU) were determined from Luria-Bertani plates. Bladder and kidney were weighted to derive bacterial concentrations as CFU/mL in these organs, and mice were assumed to produce 1mL volume of urine daily. Lower limit of detection (LOD) was established at 10 CFU/mL. CFU counts were then log-transformed for the analysis.
First, we developed a compartmental model describing bacteria colonisation of kidney, urine and bladder for the control group infected by the PAS strain, using nonlinear mixed effect models. Different structural models were tested considering direct connections between organs, with and without logistic growth. Interindividual variability was tested one parameter at a time from a model with only one variability, basing model selection on the corrected Bayesian criteria information information (BICc), reduction in relative standard errors and goodness-of-fit. Ciprofloxacin effect was modelled with an antibiotic killing factor in all compartments, proportional to dose. We then simulated 1000 mice with an initial inoculum of 10^8 CFUs and estimated infection clearance rate in the different organs as the time to reach the LOD. We used Monolix and Simulx v2021R2 to perform inference and simulation.
Results:
A structural model without logistic growth and linear elimination in all compartments was selected with variability on urine elimination and an additive error. In bladder and urine we found a net elimination rate of 0.29 and 65 CFU/day respectively. The kidney showed net proliferation of 9.7 CFU/day. The residual errors for kidney, bladder and urine compartments were estimated 1.3, 0.23 and 1.8 log10(CFU), showing high unexplained variability especially in urine and kidney.
Simulations from the final model showed that 80% of the subjects cleared the infection in 42, 50 and 60 days for bladder, kidney and urine respectively in arm without antibiotic treatment. For the antibiotic effect we found a faster clearance in urine and kidney compartment but not in bladder.
Conclusions:
The competition of bacteria elimination and proliferation are different in the three compartments. The relationship among bladder, kidney and urine exchange rates indicates that the urine compartment is the main path to elimination and the large proliferation in the kidney positions it as the primary source of a long lasting bacteria. The persistence of bacteria in the bladder despite the administration of ciprofloxacin even at higher concentrations has been reported before [5] In this study we presented a reduced comprehensive mathematical model of the colonisation in UTI organs for PAS E. coli under the administration of two doses of ciprofloxacin which is the starting point to develop better models with other antibiotics with different mechanism of action and different strains.
Acknowledgments:
The data used in this work was collected in the Anoruti project funded by ANR (Agence Nationale de la Recherche, France).
References:
[1] Rosen DA, Hooton TM, Stamm WE, Humphrey PA, Hultgren SJ. Detection of intracellular bacterial communities in human urinary tract infection. PLoS Med. 2007 Dec;4(12):e329. doi: 10.1371/journal.pmed.0040329. PMID: 18092884; PMCID: PMC2140087.
[2] Kotula, Jonathan W., et al. “Programmable bacteria detect and record an environmental signal in the mammalian gut.” Proceedings of the National Academy of Sciences 111.13 (2014): 4838-4843.
[3] Lloyd A, Rasko D, Mobley H. Defining Genomic Islands and Uropathogen-Specific Genes in Uropathogenic Escherichia coli. J Bacteriol 2007;189. https://doi.org/10.1128/jb.01744-06
[4] Bleibtreu A, Clermont O, Darlu P, Glodt J, Branger C, Picard B, Denamur E. The rpoS gene is predominantly inactivated during laboratory storage and undergoes source-sink evolution in Escherichia coli species. J Bacteriol. 2014;196:4276-84. doi: 10.1128/JB.01972-14.
[5] Jakobsen LLundberg CV, Frimodt-Møller N 2020. Ciprofloxacin Pharmacokinetics/Pharmacodynamics against Susceptible and Low-Level Resistant Escherichia coli Isolates in an Experimental Ascending Urinary Tract Infection Model in Mice. Antimicrob Agents Chemother 65:10.1128/aac.01804-20.
Reference: PAGE 32 (2024) Abstr 11186 [www.page-meeting.org/?abstract=11186]
Poster: Drug/Disease Modelling - Infection