Jinqiu Yin (1), L.B.S. Aulin (1), P.J. Upadhyay (1), C. Moser (2), O. Ciofu (2), P.H. van der Graaf (1), N. Høiby (2), W. Hengzhuang (2), J. G. C. van Hasselt (1)
1.Leiden Academic Centre for Drug Research, Leiden University, Leiden, Netherlands.2.Department of Clinical Microbiology, University Hospital, Rigshospitalet, Copenhagen, Denmark
Objectives:
Biofilms have been implicated in a wide array of chronic infections. Biofilm-associated infections show reduced antibiotic sensitivity [1], which is mediated by several mechanisms including the production of a protective extracellular matrix that reduces antibiotic penetration, and the development of additional tolerance mechanisms in biofilm-associated bacteria [2]. Characterization of pharmacokinetic-pharmacodynamic (PKPD) relationships represents a crucial step to rationalize dose regimens of antibiotics[3]. Experimental and mathematical characterization of antibiotic PKPD has been extensively used to study planktonic (non-biofilm) infections. Much less attention has been attributed to characterize the PKPD of biofilm-associated infections to investigate opportunities for dose regimen optimization. Several experimental models have been proposed to study the effect of antibiotics on biofilm-associated infections. One example is the in vitro alginate bead model [4], which mimics biofilm aggregates found in in vivo infections. The alginate bead model is thus highly suited to quantify bacterial kill kinetics and is therefore of relevance for the characterization of antibacterial PKPD of biofilm-associated infections.
In the current analysis, we develop a pharmacodynamic model to characterize the effect of imipenem on an in vitro Pseudomonas aeruginosa biofilm model. The developed model may be used to guide dose regimen optimization for chronic biofilm infections.
Methods:
Experiments: We obtained experimental data from a previously conducted experiment studying the effect of imipenem on P. aeruginosa cultured in planktonic culture and the alginate bead model [5]. Planktonic and alginate beads with immobilized P. aeruginosa were exposed to a range of concentrations of imipenem for a duration of 24 hours (1-32 mg/L for planktonic bacteria, 0.5-2048 mg/L for alginate beads). At different time points, bacteria were quantified as colony forming units (CFU) by plating in triplicate.
Model development: Different model structures formulated assets of ordinary differential equations were used to capture the observed kill kinetics. First, a natural growth model was developed. We then developed separate models for both the planktonic and biofilm data, which were ultimately integrated. Antibiotic drug effects were tested using both linear and Hill-type relationships.
Results:
Complex kinetics of the bacterial time-kill data, with up to four different sub-populations present. A model was developed capturing the kill kinetics of imipenem in planktonic and biofilm experiments simultaneously. Planktonic bacterial kill kinetics was described using two sub-populations: drug-sensitive (S) and resistant (R), whereas for the biofilm experiments four sub-populations were needed: drug-sensitive (S), and resistant (R), and tolerance (T), and persistent (P). At the start of the experiment for planktonic and biofilm, we assumed that the total baseline bacterial count (B0,log10 CFU/mL) was 5.69(RSE 1%). In the planktonic part, the B0 was only including the S state. In the biofilm part, B0 consisted of the S state, T state, and P state. Among them, baseline tolerance bacterial count (T0,log10 CFU/mL) was estimated as 4.07(RSE 14%) and the baseline persister bacterial count (P0,log10 CFU/mL) was fixed as 0.5. Both Capacity-limited growth (BMAX, log10 CFU/mL) were estimated at 9.32(RSE 1%). Common growth rates (Kg) were identified for biofilm or planktonic bacteria for the S, R and T subpopulations at 0.913 h-1 (RSE 10%), 1.79 h-1 (RSE 4%), and 0.305 h-1(RSE 29%). A first-order transfer rate (Ksr, -Log10 h−1) from the S to R was estimated at 4.07(RSE 14%). Three separate Emax models were used to described concentration-effect relationships, with a single Emax 5.19 h-1(RSE 10%) parameter and three EC50 parameters for the S, R, and T sub-populations 0.652 (RSE 25%),28.4 (RSE 12%), and 68.2(RSE 43%) mg/L. The drug diffusion delay rate in alginate beads (Delay mg/L/h)of imipenem was estimated at 0.298 (RSE 10%), which used in a 3-transit compartments model.
Conclusions:
We successfully developed a pharmacodynamic model to quantify the complex growth and kill kinetics of imipenem on P. aeruginosa PAO1. As a next step, we will use this model to investigate the efficacy of clinical dose regimens of imipenem for biofilm-associated infections [6].
References:
[1] Bjarnsholt T. APMIS 2013(136):1-51.
[2] Bjarnsholt T et al. Nature Reviews Drug Discovery 12 791-808 (2013)
[3] Ciofu et al. Adv Drug Deliv Rev. 85:7-23. (2015)
[4] Pedersen et al. APMIS, 98 (3), 203-11 (1990)
[5] Wang et al. Antimicrobial Agents and Chemotherapy 55(9):4469-74 (2010).
[6] van Hasselt et al. Br J Clin Pharmacol. 2016 Jun; 81(6): 1113–1123.
Reference: PAGE () Abstr 9517 [www.page-meeting.org/?abstract=9517]
Poster: Drug/Disease Modelling - Infection