Manipulating virulence factor availability can have complex consequences for infections.

Given the rise of bacterial resistance against antibiotics, we urgently need alternative strategies to fight infections. Some propose we should disarm rather than kill bacteria, through targeted disruption of their virulence factors. It is assumed that this approach (i) induces weak selection for resistance because it should only minimally impact bacterial fitness, and (ii) is specific, only interfering with the virulence factor in question. Given that pathogenicity emerges from complex interactions between pathogens, hosts and their environment, such assumptions may be unrealistic. To address this issue in a test case, we conducted experiments with the opportunistic human pathogen , where we manipulated the availability of a virulence factor, the iron-scavenging pyoverdine, within the insect host . We observed that pyoverdine availability was not stringently predictive of virulence and affected bacterial fitness in nonlinear ways. We show that this complexity could partly arise because pyoverdine availability affects host responses and alters the expression of regulatorily linked virulence factors. Our results reveal that virulence factor manipulation feeds back on pathogen and host behaviour, which in turn affects virulence. Our findings highlight that realizing effective and evolutionarily robust antivirulence therapies will ultimately require deeper engagement with the intrinsic complexity of host-pathogen systems.