A 3D-printed multi-compartment device that enables dynamic PK/PD profiles of antibiotics.

Abstract

Pathogens develop resistance to various drugs while under the selective pressure of antibiotics resulting in the emergence of bacterial strains that are resistant to multiple treatment options. Unfortunately, the resistance to antibiotics has also been accompanied by a reduction in the development of novel antibiotics to combat various pathogens. Current diagnostic tools, which are used in parts of the early developmental process of antibiotics, primarily consist of static susceptibility tests that do not resemble the pharmacokinetics of the therapy in vivo. Here, we designed and 3D-printed cubical inserts with membranes on two of the cube faces that allow diffusion of a molecule across two planes. These inserts are used with a 3D-printed device to create a two-compartment model to mimic the pharmacokinetics of a molecule in humans from multiple types of administration. Fluorescein was used to characterize the device and the diffusion of molecules from a flowing channel, through a membrane in the first plane (representing the primary compartment in vivo, or plasma), followed by measurement in the second compartment (that represents the interstitial fluid). The dynamic, two-compartment model was tested using both gram-positive and gram-negative bacterial strains in the secondary compartment. The ATP/OD600 (a measure of antibiotic activity)ofa kanamycin-resistant E. coli strain challenged with the antibiotic levofloxacin increased after reaching an effective concentration of the antibiotic at 2 h, equating to a secondary compartment concentration of 3.5 ± 1.3 µM levofloxacin. The ATP/OD600 of achloramphenicol-resistant B. subtilis strain challenged withthe antibiotic levofloxacin remained steady or increased slightly after reaching an effective concentration of the antibiotic. The earliest statistical difference was detected 3 h after the start of the PK curve, which corresponds with a secondary compartment concentration of 4.8 ± 1.8 µM levofloxacin. Our results demonstrate that a fabricated 2-compartment model (1) provides realistic PK values to those published from in vivo studies and (2) can be used to determine antibiotic pharmacodynamics.