When an oral bacterium turns immune cells into killers

Neutrophils in colorectal tumors have a bad reputation, and so does Fusobacterium nucleatum. A new study¹ puts the two together and finds something unexpected: the bacterium draws neutrophils into the tumor and switches them on to kill it. Whether that happens in your patient comes down to one receptor.

Not every gut bacterium recruits neutrophils the same way

Of the two bacteria most common in colorectal tumors, only F. nucleatum pushed tumor cells to release neutrophil-attracting signals, and drew far more neutrophils toward it in lab tests than Bacteroides fragilis did. In mice, F. nucleatum raised neutrophil levels inside the tumor; B. fragilis did not. Direct contact matters too: within five minutes, F. nucleatum latched onto up to 25% of circulating neutrophils and switched on an activation marker that peaked by three hours.

One receptor flips the switch, and it isn't TLR4

Neutrophils exposed to F. nucleatum released a cocktail of antimicrobial and tumor-killing compounds. Their conditioned media killed colon cancer cells in the lab and slowed tumor growth in mice when injected directly into it; B. fragilis-conditioned media did nothing. Blocking TLR4 stopped the activation marker from appearing but left the killing intact. Blocking Siglec-5/14 did the opposite: killing stopped. Siglec-14 is the one that binds the bacterium tightly, while its partner Siglec-5 barely does. Unlike most of its receptor family, which calms neutrophils down, Siglec-14 switches them on instead.

A common gene variant may explain the conflicting outcomes data

Patients carrying two copies of a common SIGLEC14 gene variant have neutrophils that still respond to F. nucleatum but can't kill: the more copies of the variant, the weaker the response. That variant is present in under 10% of Europeans but up to 70% of East Asians, which may explain the split literature: European and US cohorts link neutrophil infiltration to better survival, Chinese cohorts to worse. Mice lack the activating receptor entirely, which is likely why animal studies have painted F. nucleatum as purely harmful. In one 444-patient sample, high neutrophil markers predicted longer survival, but only in tumors that also carried F. nucleatum. In a large public gene-expression database (The Cancer Genome Atlas), the 31 patients whose tumors lacked SIGLEC14 did worse than the 351 who had it.

This is a biomarker question before a prescribing one. The next time the evidence on neutrophil infiltration in CRC looks irreconcilable, the missing variables may be which bacteria occupy the tumor and which Siglec-14 allele the patient inherited.