Gut bacteria and viral infections: the tryptophan link

Gut bacteria and viral infections may be connected by a molecule made from an ordinary amino acid: tryptophan. In a new study, researchers traced a full biochemical pathway in which certain commensal bacteria convert tryptophan into indole-3-lactic acid, or ILA, and that metabolite shifts host-cell physiology in a way that makes replication harder for two very different viruses.

The story began with 24 infant rhesus macaques. The animals received weekly oral challenges with SHIV, a hybrid virus used to model HIV transmission. Some became infected after only a few exposures, while others required as many as 30 challenges. Fecal samples collected before the challenges linked delayed infection with Lactobacillus gasseri and members of the bacterial family Lachnospiraceae.

That association alone could not establish causality. The bacteria might simply have been markers of some other protective feature. The investigators therefore moved from microbiome correlations to experiments in cells.

From a bacterium to a molecule

Human-derived strains of several candidate bacteria suppressed HIV replication in a standard laboratory cell model. Clostridium immunis also inhibited HIV in primary CD4+ T cells isolated from tonsillar tissue. The antiviral activity was present in bacterial culture supernatant, indicating that a secreted product rather than direct bacterial contact was responsible.

The critical step involved aromatic amino acid aminotransferase, or ArAT, a bacterial enzyme encoded by araT. When the researchers disrupted araT in C. immunis, its ability to inhibit HIV fell sharply. A similar experiment with Lactobacillus reuteri produced the same pattern: the wild-type strain inhibited infection, whereas an ArAT-deficient mutant largely lost that activity.

Untargeted metabolomics provided the next clue. Among 214 annotated metabolites, only one — indole-3-lactic acid — was significantly enriched in wild-type C. immunis and absent from the ArAT-deficient mutant. ILA on its own was sufficient to inhibit HIV in the cell model, while a related tryptophan metabolite did not show the same effect.

ILA is an agonist of the aryl hydrocarbon receptor, or AhR, an intracellular regulator that alters the expression of many genes. In the experiments, C. immunis reduced expression of the cell-cycle regulator Cdk1 in an ArAT-dependent manner. The authors propose that this can ultimately reduce the pool of DNA building blocks available for viral replication. That downstream mechanism is biologically plausible and consistent with the HIV experiments, but it remains a cell-based mechanism rather than demonstrated protection in an intact organism.

Beyond HIV

The researchers next asked whether the pathway might matter for a very different virus. They tested human cytomegalovirus, a DNA herpesvirus, and found that C. immunis inhibited CMV replication in cultured cells. Removing ArAT weakened that effect.

They then examined four human cohorts. In Kenyan children with perinatal HIV exposure, Lachnospiraceae abundance was associated with HIV acquisition status. The cohort included 28 children who acquired HIV, 16 exposed but uninfected children, and 91 HIV-unexposed controls.

A separate adult HIV cohort included 9 men sampled shortly before documented HIV seroconversion and 20 individuals from a similar high-incidence population who remained uninfected. The future HIV-positive group had fewer araT copies before seroconversion, but the difference did not reach the conventional threshold for statistical significance (P=0.0768). This is supportive evidence, not a stand-alone confirmation.

The CMV analysis was stronger. Among 72 children undergoing hematopoietic stem-cell transplantation, araT abundance fell after transplantation in the 13 who developed detectable CMV viremia compared with the 59 who did not. The difference later diminished as the post-transplant period progressed.

For SARS-CoV-2, the researchers saw no difference in araT abundance between people who became infected and those who did not. But among 233 infected participants — 67 asymptomatic and 166 symptomatic — araT abundance was higher in the asymptomatic group. Importantly, symptomatic cases in this cohort were mild, so the study does not establish protection from severe COVID-19.

What the study shows — and what it does not

The strength of the work is the way it connects multiple levels of evidence: an observation in primates, associations in humans, experiments in human cells, genetic disruption of a bacterial enzyme, and identification of a specific metabolite. That goes well beyond a simple report that one microbiome pattern correlates with disease.

The major missing step is in vivo validation of the proposed antiviral effect. The authors did not show that administering C. immunis, another ArAT-expressing bacterium, or ILA prevents viral infection in an animal or a person. The human studies mainly measured the abundance of araT in fecal metagenomic data rather than concentrations of ILA at the relevant site of infection. AhR signaling is also context-dependent: the same pathway can have different effects depending on the virus, cell type, and ligand.

There is also a declared commercial interest worth noting. Several authors are inventors on a Duke University patent application covering therapeutic use of ArAT-expressing bacteria and their metabolites in viral infections.

So the study does not show that a probiotic or a tryptophan-derived supplement can prevent viral disease. It shows something more specific and potentially more useful: some commensal bacteria possess a defined biochemical pathway that can alter how susceptible cells are to viral replication. If that pathway holds up in relevant animal models and eventually in human intervention studies, the microbiome could become a tractable target rather than merely a correlate of infection outcomes.