BGPT: Paper Review: Heterogeneous Persister Cells Formation in Acinetobacter baumannii

Explore by Goal

Quick Explanation Copied

Key finding

Across 37 clinical Acinetobacter baumannii isolates (polymyxin B–susceptible), persister fractions after 6 h of exposure were highly heterogeneous (polymyxin B: 0.0007%–10.1%; tobramycin: 0.0003%–11.84%), and persistence magnitude was not correlated between the two unrelated antibiotics—suggesting tolerance mechanisms that are not explained by a single dormancy-like state alone.

Long Explanation

Paper review (science-first, skeptical, evidence-based)

Title: Heterogeneous Persister Cells Formation in Acinetobacter baumannii DOI: 10.1371/journal.pone.0084361 Year: 2013

1) Visual synthesis (from the paper’s quantitative ranges)

The paper reports persister-fraction ranges across isolates under two antibiotic exposures (6 h). Because per-isolate raw persister fractions are not included in the provided text excerpt, the plots below visualize only the reported minima/maxima (range envelopes), not the full distribution.

Source: polymyxin B persister fractions 0.0007%–10.1% (6 h, 15 mg/mL) and two isolates with no detectable persisters under these conditions.

Source: tobramycin persister fractions 0.0003%–11.84% (6 h, 160 mg/mL).

Source: 35/37 (94.6%) had detectable persister formation after polymyxin B exposure; two isolates showed none.

2) What the paper claims (and what the evidence supports)

2.1 Experimental premise

The authors treat “persisters” as survivors after a short, high-dose antibiotic challenge in susceptible strains (i.e., not classic resistance). Their core goal is to quantify persister fractions in A. baumannii clinical isolates and assess variation across strains and across two mechanistically unrelated antibiotics: polymyxin B (15 mg/mL) and tobramycin (160 mg/mL), each for 6 h in stationary-phase cultures.

2.2 Main quantitative result: heterogeneity

Under polymyxin B, persister fractions varied over orders of magnitude (0.0007%–10.1%). Under tobramycin, persister fractions also varied widely (0.0003%–11.84%). This supports the paper’s emphasis on strain-level heterogeneity in tolerance formation.

2.3 Key comparative result: lack of correlation between antibiotics

The authors report no significant correlation between persister formation intensity for polymyxin B vs tobramycin among isolates, which they interpret as evidence that distinct internal controls regulate tolerance for these unrelated antibiotics (and that a single dormancy state is likely insufficient to explain multidrug tolerance patterns).

2.4 Biological interpretation (what is supported vs speculative)

Supported by data: Extreme between-strain variability and lack of correlation between two antibiotic exposures, both quantified as survivor fractions after washout and CFU plating.
Mechanistic “leaps”: The paper argues that polymyxin B’s inability to eradicate persisters (despite its theoretical action on non-dividing cells) suggests a disconnection between persistence and dormancy, and that different mechanisms may regulate tolerance to unrelated antibiotics. That mechanistic inference depends critically on assumptions about what CFU-survival after washout represents (and whether antibiotic removal fully eliminates residual killing), which the paper itself partly acknowledges when discussing two “non-producer” isolates potentially affected by carryover drug during early dilutions.

3) Methodological critique (skeptical accounting)

3.1 Persister quantification depends on assay constraints

This work uses CFU-based survivor quantification after antibiotic exposure and washout. That approach is standard for persistence studies, but the paper’s own discussion highlights an important sensitivity: if residual antibiotic remains and suppresses the few survivors during initial plating/dilution steps, the assay can classify a strain as “non-persister producer” even if persisters exist at extremely low levels.

3.2 Correlation tests cannot prove mechanism

Absence of statistical correlation between persister fractions after polymyxin B vs tobramycin is consistent with “distinct internal controls,” but it does not uniquely determine the underlying molecular architecture. Different effects could also arise from (i) different stress response dynamics within the same general tolerance program, (ii) different residual-drug carryover effects by antibiotic class, or (iii) differences in killing kinetics that alter effective exposure state in a way that breaks correlation. The study reports the correlation result but does not (in the provided text) include direct mechanistic biomarkers (e.g., dormancy depth measures, TA activity, ROS states) for the two antibiotic conditions.

3.3 Population sampling bias

Strains are described as nosocomial clinical isolates from a single hospital over Jan–Sep 2012. That makes the dataset clinically relevant but also limits generalization across global lineages and ecological contexts (e.g., non-hospital reservoirs). The paper itself frames this as an initial report, which increases the need for follow-up across broader isolate collections.

4) What would change the conclusion (disproof targets)

Assay-artifact disproof: If improved antibiotic removal (validated by antibiotic carryover measurements) eliminates apparent “non-correlation” or eliminates “zero-producer” strains, then the cross-antibiotic heterogeneity could partly be experimental rather than biological. This is motivated by the paper’s carryover concern for undetectable polymyxin B persisters.
Mechanistic state disproof: If direct molecular-state measures (dormancy depth / TA module activity / stress-response markers) show that the same physiological program is active across both antibiotics in the high-persister isolates, then “distinct mechanisms” would be weakened even if statistical correlation of CFU fractions remains low due to nonlinear mapping between state and survival. (The excerpt does not show such biomarkers, so this remains an open falsifier.)

5) Author-review links (bespoke follow-ups)

Open a dedicated BGPT Author Review for each author to cross-check interpretation, technical choices, and any downstream citations or follow-up experiments.

Feedback:

Updated: April 28, 2026