Blood RNA reveals tuberculosis rRNA shifts during treatment

Current clinical tools for monitoring tuberculosis treatment — sputum culture and chest X-ray — can be unreliable for people with low bacterial loads and often do not give rapid answers. There is therefore a clear need for biomarkers that provide earlier, more accurate information about whether the pathogen is alive and responding to therapy. To explore this, Patricio Lopez-Exposito analyzed publicly available whole-blood RNA-seq data from 79 TB patients. These samples were collected at four time points: diagnosis, and weeks 1, 4, and 24 of standard anti-TB treatment. The study first removed human reads from the sequencing data and then mapped the remaining reads to the Mycobacterium tuberculosis H37Rv genome. The focus was on ribosomal RNA subunit transcripts—16S, 23S, 5S, ITS1, and ITS2—looking for patterns that change with treatment. By working directly with pathogen-derived reads in routine whole-blood RNA-seq, the analysis tested whether a blood-based, transcriptomic signal could complement or improve on sputum and X-ray approaches for following treatment response.

The analysis combined read mapping to the Mycobacterium tuberculosis H37Rv genome with microbiome profiling tools Kraken2/Bracken and statistical approaches including alpha and beta diversity measures and differential abundance testing with ANCOM-BC2. Two major rRNA subunits, 16S and 23S, were consistently detected at all treatment time points. At diagnosis and in the early stages of treatment 23S reads were more abundant, but by week 24 there was a marked shift toward predominance of 16S reads. Smaller rRNA components showed weaker signals: 5S and ITS1 appeared inconsistently across samples, and ITS2 was not detected. Diversity analyses found that alpha diversity (measured by the Shannon index) increased during treatment with statistically significant increases at weeks 1 and 4, while beta diversity showed significant shifts over time. Notably, these compositional changes occurred without a significant change in total M. tuberculosis abundance, suggesting transcript-level dynamics rather than simple loss of bacterial material.

These findings suggest it may be feasible to detect M. tuberculosis rRNA signatures in whole-blood RNA-seq and that those signatures change dynamically during the course of standard anti-TB treatment. In particular, the relative levels of 16S versus 23S and the presence or absence of minor rRNA units like 5S and ITS1 could serve as complementary biomarkers alongside other transcriptomic-based markers. Such blood-based signals could help overcome limitations of sputum culture and chest X-ray, especially for patients who produce little sputum or have low bacterial loads, and could offer earlier readouts of treatment response. The study emphasizes that these results are preliminary: future work should validate the patterns in larger cohorts and use optimized RNA-seq protocols that are designed to capture pathogen transcripts more efficiently. If confirmed, these transcriptomic signatures could become part of a more responsive toolkit for monitoring TB therapy.