Hidden turnover: TB bacteria replicate and die more than thought

Tuberculosis remains a disease where the immune signs that truly protect people and animals are still unclear. Researchers commonly compare colony forming units (CFUs) — a measure of live bacteria — to chromosomal equivalents (CEQs), which count bacterial DNA, and use the ratio Z = CFUs/CEQs as a shorthand for how well the body is killing Mycobacterium tuberculosis ( Mtb ). But that single ratio treats the DNA signal as if it were immortal and doesn’t ask whether loss of DNA over time might change what Z tells us. A Friesen and colleagues set out to test that assumption by building alternative mathematical models that follow CFUs, CEQs, and Z through infection. Rather than accepting that Z alone reveals bacterial death rates, the team examined how coupled changes in live bacteria and DNA could be driving observed patterns. They used these models to re-analyze animal data from mice, rabbits, and macaques and to ask whether previous interpretations had missed ongoing bacterial turnover. The work reframes a common experimental metric and asks for more caution when using CFUs and CEQs to judge how well immunity clears Mtb.

The team developed mathematical models that explicitly track both CFUs and CEQs and explored how their dynamics interact. They show that the ratio Z = CFUs/CEQs cannot, by itself, reveal the bacterial death rate unless CFU and CEQ dynamics are entirely uncoupled — a biologically unreasonable assumption because CEQs are thought to reflect accumulated viable and non-viable bacteria. Using their models on data sets, they estimated a decay rate of 3.6% per day (about a 20-day half life) for Mtb H37Rv CEQs in B6 mice, remarkably similar to a previously reported 4% per day for Mtb Erdman in cynomolgus macaques. Although those DNA decay rates are small, the authors found that estimates of Mtb replication and death are highly sensitive even to slow DNA loss, because replication and death rates themselves are small, especially during chronic infection. Applying the models to data covering the first 3 weeks of infection in macaques, they found evidence for substantial killing of Mtb before adaptive immunity arrives, challenging the idea that bacteria do not die in the absence of T cell immunity and granuloma formation. The paper also proposes experiments to measure Mtb DNA loss more accurately so future work can quantify immunity’s impact more rigorously.

These results matter because many studies use Z = CFUs/CEQs to infer how well a host is killing Mtb, and if CEQs decay measurably over time that inference can be misleading. By showing CEQ decay rates in B6 mice similar to those seen in macaques, the models highlight that even slow loss of bacterial DNA changes conclusions about replication and death, particularly during long-term infection when bacterial turnover is low. The evidence for substantial bacterial killing in the first 3 weeks after infection — before the arrival of adaptive T cell responses and granuloma formation — suggests early innate or other processes may remove bacteria and that the field should not assume bacteria are entirely non-dying until adaptive immunity appears. Practically, the paper calls for direct experimental measurements of Mtb DNA loss in lesions and lungs; those measurements would let researchers separate DNA persistence from live-bacteria counts and so more accurately estimate how immune responses, vaccines, or treatments influence within-host Mtb dynamics. Overall, the work urges more nuanced interpretation of CFUs and CEQs and better data to pin down how Mtb truly replicates and dies in hosts.