Researchers have developed a high-throughput imaging system that tracks how quickly antibiotics kill individual bacteria, revealing that the speed of bacterial death – not just drug resistance – determines whether infections will clear in patients.
The technique, reported in Nature Microbiology, is called Antimicrobial Single-Cell Testing (ASCT). It uses automated microscopy to monitor millions of bacterial cells simultaneously as they're exposed to antibiotics. By tracking each bacterium over time, and measuring when its cell membrane fails, the system creates detailed "time-kill curves" showing how rapidly entire bacterial populations die.
Testing the approach on tuberculosis and Mycobacterium abscessus infections, the team found surprising disconnects between conventional susceptibility tests and treatment outcomes. For tuberculosis, drugs performed well in standard growth inhibition tests but failed to predict which regimens worked in mice and humans. The key difference emerged when bacteria were starved before antibiotic exposure – mimicking conditions inside infected tissue. Under starvation, drugs like bedaquiline and pretomanid killed bacteria more effectively than standard first-line drugs, accurately reflecting their superior performance in clinical trials.
The researchers then analyzed 405 clinical M. abscessus isolates from patients across Europe and Australia, generating nearly 20,000 time-kill curves. They discovered that bacterial strains killed slowly by certain antibiotics were linked to treatment failures in individual patients, even when conventional drug resistance tests showed the bacteria should be susceptible.
Crucially, this "drug tolerance" – the ability to temporarily survive antibiotic exposure – proved to be genetically determined and heritable. Whole-genome sequencing revealed that specific bacterial genes control how quickly cells die during antibiotic treatment. The team even identified one gene that, when deleted, made bacteria more tolerant to several antibiotics.
The findings suggest that current minimum inhibitory concentration (MIC) testing provides incomplete information. The research indicates that combining standard MIC results with measurements of killing speed improved prediction of patient outcomes from 69 percent to 78 percent accuracy. This could enable more personalized antibiotic selection, particularly for difficult-to-treat infections where standard regimens often fail despite apparent drug susceptibility.
The platform's ability to simultaneously test hundreds of bacterial strains against multiple antibiotics makes it practical for clinical implementation, potentially transforming how laboratories guide antibiotic therapy decisions.
Newsletters
Receive the latest pathologist news, personalities, education, and career development – weekly to your inbox.
