Sophie Marolleau 1, Jérémy Moreau 1, Nicolas Grégoire 1,2, Julien Buyck 1, Vincent Aranzana-Climent 1
1 INSERM U1070, Université de Poitiers (Poitiers, France), 2 Laboratoire de toxicologie et de pharmacocinétique, CHU de Poitiers (Poitiers, France)
1. Objectives
Antimicrobial resistance is a priority for the World Health Organization and was directly responsible for 1.27 million deaths worldwide in 2019 [1]. In this context, innovative strategies to counteract resistance and restore antibiotic efficacy are urgently needed. Targeted-antibacterial-plasmids (TAPs) were developed to combat antibiotic resistance by delivering a CRISPR/Cas9 system targeting resistance genes via bacterial conjugation [2], [3]. The TAPs studied here carry a CRISPR/dCas9 system programmed to target and thereby inhibit the expression of the blaOXA-48 gene on plasmid pOXA-48a, which confers resistance to ampicillin.
TAP plasmids are not self-transmissible; their dissemination depends on a mobilizable helper plasmid encoding the conjugation machinery. Importantly, the helper plasmid can be transmitted independently, in which case the TAP can no longer be transferred.
The objective of this study was to quantify the resensitization efficacy of TAPs in target bacteria. To this end, we developed a mechanistic nonlinear modelling framework to characterize TAP conjugation kinetics and evaluate their ability to resensitize resistant bacterial populations. This approach enables the identification of key factors limiting TAP-mediated resensitization and provides quantitative guidance for the design of resistance-reversal strategies.
2. Methods
2.1. Conjugation experiments
Time-course conjugation experiments were conducted by mixing donor bacteria carrying both TAPs and helper plasmids with ampicillin-resistant recipient bacteria harboring pOXA-48a. Bacterial counts were quantified over 24 hours (0, 1, 2, 3, 4, 6, 8, and 24 h) using selective agar plates, allowing discrimination of five subpopulations:
Donors [D]: bacteria initially carrying both TAP and helper plasmids; resistant to chloramphenicol.
Original recipients [R0]: recipient bacteria initially harboring the pOXA-48a resistance plasmid (target of the TAP system); resistant to streptomycin and ampicillin.
Helper-positive recipients (recusants) [Rh]: bacteria that acquired only the helper plasmid, refractory to further conjugation and remaining ampicillin-resistant; resistant to streptomycin, ampicillin, and tetracycline.
Transconjugants [T]: recipient bacteria that acquired both TAP and helper plasmids and became resensitized to ampicillin; these bacteria can subsequently act as secondary donors; resistant to streptomycin and tetracycline.
Escapers [TS]: transconjugants that escaped CRISPR/dCas9-mediated inhibition; they remained ampicillin-resistant despite carrying the TAP; resistant to streptomycin, ampicillin, and tetracycline.
Experiments were performed across a broad range of initial inocula of donors (10²–10⁸ CFU.mL-1), donor-to-recipient ratios (D:R from 1:100 to 100:1), and bacterial growth phases.
In addition, all bacterial subpopulations generated by conjugation were isolated and grown independently for 24 hours under conditions preventing further conjugation, allowing the characterization of their intrinsic growth dynamics.
2.2. Modeling
A system of nonlinear ordinary differential equations (ODEs) was developed to describe bacterial growth, plasmid transfer dynamics, helper-dependent mobilization of TAP, plasmid loss, and CRISPR/dCas9-mediated inhibition of resistance expression. All bacterial subpopulations were modelled simultaneously.
Model parameters were estimated using Monolix (2024R1) under a deterministic nonlinear modelling framework, without inclusion of inter-individual variability given the controlled in vitro experimental conditions, using the Stochastic Approximation Expectation–Maximization (SAEM) algorithm.
Competing mechanistic hypotheses regarding TAP and helper plasmid transfer, as well as plasmid loss, were evaluated based on goodness-of-fit diagnostics and biological plausibility. This modelling framework enabled quantitative characterization of the factors limiting TAP-mediated resensitization and may inform the design of optimized resistance reversal strategies in preclinical development.
2.3. Efficiencies
To evaluate the performance of the TAP system, three efficiencies were calculated based on recipient-derived populations (excluding donors): conjugation efficiency, TAP efficiency, and resensitization efficiency. Recipient-derived populations included all bacteria targeted by TAP: transconjugants, escapers, and recipients with or without the helper plasmid.
Conjugation efficiency quantifies the success of TAP transfer within recipient bacteria (Equation 1):
Eff_conj = (T + TS) / (T + TS + R0 + Rh)
TAP efficiency quantifies the intrinsic ability of TAP to resensitize bacteria once transferred (Equation 2):
Eff_TAP = T / (T + TS)
Resensitization efficiency measures the net biological outcome of the TAP system. This metric (Equation 3) is equivalent to the product of conjugation and TAP efficiencies and provides the most biologically relevant assessment, while the Conjugation and TAP efficiencies enable identification of limiting steps and potential optimization targets within the system.
Eff_tot= T / (T+TS+R0+Rh)
3. Results
3.1. Modelling
The mechanistic model successfully captured the observed bacterial counts on selective agar plates, enabling estimation of the abundance of the different subpopulations and simulation of alternative scenarios, including varying donor inocula and donor-to-recipient (D:R) ratios. The dynamics of donors, recipients, recusants, transconjugants and escapers could be quantified at each time point.
3.2. Efficiencies
Regarding system performance, TAP efficiency approached 100% over 24 hours across all simulated scenarios, indicating that nearly all transconjugants carrying TAP were effectively resensitized. Consequently, resensitization efficiency was equivalent to conjugation efficiency. However, in all simulated scenarios, conjugation efficiency reached a plateau ranging from approximately 65% to nearly 80%.
4. Conclusion
Overall, our results demonstrate that resensitization efficiency critically depends on conjugation dynamics and plasmid-transfer parameters. By providing a quantitative framework to predict TAP dissemination and activity, this mechanistic modelling approach offers valuable guidance for the rational optimization of TAP-based resistance-reversal strategies in preclinical settings.
References:
[1] C. J. L. Murray et al., « Global burden of bacterial antimicrobial resistance in 2019: a systematic analysis », The Lancet, vol. 399, no 10325, p. 629‑655, févr. 2022, doi: 10.1016/S0140-6736(21)02724-0.
[2] S. Djermoun, A. Reuter, E. Derollez, C. Lesterlin, et S. Bigot, « Reprogramming Targeted-Antibacterial-Plasmids (TAPs) to achieve broad-host range antibacterial activity », Plasmid, vol. 126, p. 102680, mai 2023, doi: 10.1016/j.plasmid.2023.102680.
[3] A. Reuter et al., « Targeted-antibacterial-plasmids (TAPs) combining conjugation and CRISPR/Cas systems achieve strain-specific antibacterial activity », Nucleic Acids Res., vol. 49, no 6, p. 3584‑3598, avr. 2021, doi: 10.1093/nar/gkab126.
Reference: PAGE 34 (2026) Abstr 12036 [www.page-meeting.org/?abstract=12036]
Poster: Drug/Disease Modelling - Other Topics