Acinetobacter baumannii poses an enduring clinical threat in healthcare environments, where it frequently causes severe hospital-acquired conditions such as ventilator-associated pneumonia, catheter-related bloodstream infections, urinary tract infections, and secondary meningitis in critically ill patients. While stable antimicrobial resistance driven by genetic mutations presents a widely recognized barrier to clinical therapy, phenotypic antibiotic tolerance presents an equally formidable challenge. Unlike resistant bacteria that possess heritable genetic changes conferring elevated minimum inhibitory concentrations, tolerant bacterial populations survive extended antimicrobial exposure through reversible physiological adaptations without altering their underlying baseline susceptibility. In hospital settings, environmental pressures including nutrient deprivation, oxidative stress, and host immune defenses frequently compel bacteria into restricted metabolic and slow-growing states. These stress-adapted phenotypes significantly diminish the bactericidal efficacy of traditional antibiotic regimens, allowing persistent sub-populations to survive treatment and potentially fostering an environment where genetic resistance mutations can subsequently arise.
To systematically investigate how varying energetic states affect antimicrobial susceptibility in Acinetobacter baumannii, a research team based at Ningbo No. 2 Hospital examined the reference strain ATCC 19606T alongside five clinical isolates gathered from intensive care unit patients. The researchers engineered an in vitro nutrient-gradient model using Luria-Bertani broth diluted across five distinct concentrations: 100 percent, 10 percent, 1 percent, 0.1 percent, and 0 percent. Following an initial growth phase to an optical density between 0.1 and 0.2, bacterial cells were harvested by centrifugation at 4000 times gravity for 10 minutes at 4 degrees Celsius, washed three times with sterile phosphate-buffered saline, and resuspended in the respective nutrient media. The cultures were then allowed an adaptation period of two hours at 37 degrees Celsius with shaking at 200 revolutions per minute, establishing stable nutrient-associated physiological conditions before any antimicrobial challenge was introduced.
The experimental nutrient gradient produced pronounced alterations in bacterial growth dynamics and cellular energy reserves. Continuous monitoring of optical density at 600 nanometers over a 24-hour incubation period revealed that the maximum specific growth rate decreased progressively from 1.358 reciprocal hours in nutrient-rich 100 percent broth to 0.831 reciprocal hours in 10 percent broth. In cultures containing 1 percent, 0.1 percent, and 0 percent broth, bacterial replication was so severely restricted that the predefined linear regression criteria of four consecutive sliding-window points with a coefficient of determination of at least 0.95 could not be met, rendering the growth rate not reliably determinable. Concurrently, intracellular adenosine triphosphate levels were quantified immediately following the two-hour adaptation interval using a commercial microbial luminescence assay and normalized against colony-forming unit counts. These measurements demonstrated that cellular energetic content per viable cell declined sharply under nutrient-restricted conditions, directly mirroring the severe reduction observed in bacterial replication kinetics.
Having mapped these distinct energetic states, the investigators evaluated the bactericidal activity of six clinically relevant antibiotics representing diverse mechanistic classes: meropenem, amikacin, ceftazidime, ciprofloxacin, polymyxin B, and tigecycline. Each antimicrobial agent was tested across standardized concentrations ranging from 0.5 times to 64 times its experimentally established minimum inhibitory concentration. In six-hour time-kill assays, bacterial survival differed extensively according to both the specific drug applied and the prevailing nutrient state. Meropenem and tigecycline exhibited substantial drops in killing effectiveness under severe nutrient limitation, leaving elevated numbers of surviving cells compared to tests in nutrient-rich broth. In contrast, amikacin maintained relatively consistent bactericidal activity across the entire nutrient gradient, displaying minimal survival divergence between rich and depleted media. The remaining agents, including ceftazidime, ciprofloxacin, and polymyxin B, exhibited intermediate, concentration-dependent variations in bacterial killing.
To provide a rigorous quantitative basis for comparing these disparate responses, the researchers established the Metabolic Dependency Index, termed MDI. For each antibiotic at each standardized concentration multiple, linear regressions were fitted between intracellular adenosine triphosphate concentrations per colony-forming unit at the time of drug addition and subsequent logarithmic bacterial survival after six hours of exposure. The MDI was calculated as the arithmetic mean of the absolute values of these regression slopes across the eight tested concentrations, with 95 percent confidence intervals estimated through 10,000 bootstrap resampling iterations. Rather than serving as an indicator of absolute causal biochemical mechanisms, the index provided a continuous descriptive scale ranking how strongly pre-exposure cellular energy states correlated with bactericidal efficacy across the panel of antimicrobials under identical experimental parameters.
The calculated index values positioned the six antibiotics along a broad continuous spectrum of energy-linked activity. Meropenem registered the highest index value at 168,619 with a bootstrap confidence interval ranging from 147,546 to 233,242, closely followed by tigecycline at 150,260 with an interval of 121,869 to 166,651. Ceftazidime recorded a value of 106,668, ciprofloxacin scored 96,631, and polymyxin B yielded 90,142. Amikacin occupied the lowest point on the spectrum, displaying an index of 54,682 with a confidence interval between 51,576 and 61,452. Exploratory sensitivity analyses that incorporated alternative pseudocount assignments for samples falling below detection limits and filtered out individual regressions with coefficients of determination below 0.10 confirmed the overall relative stability of this drug ranking. Meropenem and amikacin, anchoring the two opposing poles of the index spectrum, were consequently selected to explore optimized combination and dosing schedules under nutrient starvation.
To clarify the physiological state induced by starvation, the team analyzed transcriptional responses across several canonical stress pathways using reverse-transcription quantitative polymerase chain reaction. Cells maintained in 0.1 percent broth exhibited widespread transcriptional remodeling compared to those grown in full-strength broth. Following antibiotic exposure under nutrient-restricted conditions, toxin-antitoxin system genes, including hth, gnat, relE, and relB, demonstrated significant transcriptional upregulation across all six tested antibiotics, though the degree of induction varied by drug. Conversely, stringent response mediators relA and spoT showed minimal expression shifts, aside from a notable reduction in relA following ciprofloxacin administration. The DNA damage repair marker recA displayed antibiotic-specific induction, showing its highest fold increase after ceftazidime challenge. The authors noted that while these transcriptional changes demonstrate active stress-response remodeling during nutrient depletion, they represent associative gene expression profiles rather than direct functional genetic proof of tolerance mechanisms.
Guided by the distinct metabolic dependency profiles of meropenem and amikacin, the investigators tested whether coordinating their delivery pattern could enhance bacterial killing in nutrient-limited 0.1 percent broth. In baseline checkerboard assays, simultaneous co-administration of both drugs generated greater bacterial reduction than either agent administered as a monotherapy across multiple concentration pairings. To determine whether temporal separation could further augment efficacy, the team implemented a staggered regimen wherein meropenem was introduced first for three hours, followed by the addition of amikacin for the remaining three hours without removing the initial drug. This meropenem-first staggered approach led to significantly greater bacterial reductions than simultaneous delivery across several concentration matrices, confirming that altering the timing of antibiotic introduction directly influenced clearance of tolerant populations.
The enhancement provided by staggered exposure was confirmed through prespecified statistical comparisons evaluating matched biological replicates across two standardized dosing pairs. When 16 micrograms per milliliter of meropenem was combined with 2 micrograms per milliliter of amikacin, meropenem-first staggered administration achieved an additional 1.21 log10 colony-forming unit reduction relative to simultaneous addition, supported by a 95 percent confidence interval of 0.48 to 1.94 and a paired p-value of 0.019. Similarly, at a pairing of 16 micrograms per milliliter of meropenem and 1 microgram per milliliter of amikacin, the staggered schedule reduced surviving counts by an additional 0.86 log10 colony-forming units, with a 95 percent confidence interval from 0.47 to 1.26 and a paired p-value of 0.011. In both prespecified comparisons, all biological replicates produced quantifiable colonies without requiring data imputation, confirming that temporal separation yielded measurable bactericidal advantages in nutrient-deprived cultures.
To explore whether this observed synergy depended strictly on applying the high-MDI beta-lactam prior to the low-MDI aminoglycoside, the researchers reversed the delivery sequence, administering amikacin at the onset of incubation and adding meropenem three hours later. Unexpectedly, amikacin-first staggered treatment produced bacterial killing that was comparable to, and under several concentration combinations greater than, that achieved by the meropenem-first schedule. This finding indicated that the therapeutic benefit of staggered exposure was not driven by a unique biological requirement for meropenem pretreatment. Instead, the authors suggested that the advantage stems from broader dynamic interactions brought about by temporal separation, such as transient alterations in bacterial outer membrane permeability, proton motive force dynamics, or intracellular target accessibility that modify cellular vulnerability before the second compound is introduced.
The investigators subsequently examined whether these scheduling effects extended beyond the ATCC reference strain by testing five de-identified clinical Acinetobacter baumannii isolates obtained from patient sputum and venous catheter samples. Tested in identical 0.1 percent nutrient-depleted broth, all five isolates demonstrated marked bacterial reductions under combination treatment, with staggered administration consistently outperforming simultaneous dosing at several concentration pairings. Although the absolute magnitude of bacterial reduction varied between individual clinical strains, the overall trend remained reproducible across the panel. Following antibiotic-free recovery culture, re-tested surviving populations retained their original minimum inhibitory concentrations of 1 microgram per milliliter for meropenem and 4 micrograms per milliliter for amikacin, verifying that the surviving cells represented a reversible tolerant phenotype rather than stable genetic mutants.
While these findings demonstrate a reproducible framework for combining antibiotics based on physiological state, several methodological boundaries must be recognized. Moreover, individual concentration-specific regressions used in compiling the MDI often exhibited weak correlation coefficients, particularly at high antibiotic concentrations where bactericidal activity reached saturation. The clinical isolates evaluated were all susceptible to carbapenems, meaning these observations cannot yet be generalized to carbapenem-resistant Acinetobacter baumannii strains. Determining whether staggered dosing principles translate into improved therapeutic clearance in living hosts will ultimately require thorough validation in animal infection models and pharmacokinetic-pharmacodynamic investigations.
Subject of Research: Bacterial physiological state and staggered antibiotic dosing against Acinetobacter baumannii
Article Title: Metabolic dependency profiling guides staggered antibiotic administration to enhance killing of antibiotic-tolerant Acinetobacter baumannii
Article References: Tuo, P., Ji, Y., Ruan, C., Zhang, S., Wang, F., Shen, J., Xu, X., & Zhu, P. (2026). Metabolic dependency profiling guides staggered antibiotic administration to enhance killing of antibiotic-tolerant Acinetobacter baumannii. Frontiers in Microbiology, 17, Article 1853821. https://doi.org/10.3389/fmicb.2026.1853821
Image Credits: AI Generated
DOI: 10.3389/fmicb.2026.1853821
Keywords: Acinetobacter baumannii, antibiotic tolerance, meropenem, amikacin, Metabolic Dependency Index, staggered administration, bacterial metabolism, antimicrobial resistance, Metabolic, dependency, profiling, guides
