The gut mucosal epithelium separates the host from the microbiota, but enteropathogens such as Salmonella Typhimurium (S.Tm) can invade and breach this barrier. Defenses against such acute insults remain incompletely understood. Using a murine model of Salmonella enterocolitis, we analyzed mechanisms limiting pathogen loads in the epithelium during early infection. Although the epithelium-invading S.Tm replicate initially, this intraepithelial replicative niche is restricted by expulsion of infected enterocytes into the lumen. This mechanism is compromised if inflammasome components (NAIP1-6, NLRC4, caspase-1/-11) are deleted, or ablated specifically in the epithelium, resulting in ∼100-fold higher intraepithelial loads and accelerated lymph node colonization. Interestingly, the cytokines downstream of inflammasome activation, interleukin (IL)-1α/β and IL-18, appear dispensable for epithelial restriction of early infection. These data establish the role of an epithelium-intrinsic inflammasome, which drives expulsion of infected cells to restrict the pathogen's intraepithelial proliferation. This may represent a general defense mechanism against mucosal infections.

Vaccine-induced high-avidity IgA can protect against bacterial enteropathogens by directly neutralizing virulence factors or by poorly defined mechanisms that physically impede bacterial interactions with the gut tissues ('immune exclusion'). IgA-mediated cross-linking clumps bacteria in the gut lumen and is critical for protection against infection by non-typhoidal Salmonella enterica subspecies enterica serovar Typhimurium (S. Typhimurium). However, classical agglutination, which was thought to drive this process, is efficient only at high pathogen densities (≥10 non-motile bacteria per gram). In typical infections, much lower densities (10-10 colony-forming units per gram) of rapidly dividing bacteria are present in the gut lumen. Here we show that a different physical process drives formation of clumps in vivo: IgA-mediated cross-linking enchains daughter cells, preventing their separation after division, and clumping is therefore dependent on growth. Enchained growth is effective at all realistic pathogen densities, and accelerates pathogen clearance from the gut lumen. Furthermore, IgA enchains plasmid-donor and -recipient clones into separate clumps, impeding conjugative plasmid transfer in vivo. Enchained growth is therefore a mechanism by which IgA can disarm and clear potentially invasive species from the intestinal lumen without requiring high pathogen densities, inflammation or bacterial killing. Furthermore, our results reveal an untapped potential for oral vaccines in combating the spread of antimicrobial resistance.

The gut epithelium is a critical protective barrier. Its NAIP/NLRC4 inflammasome senses infection by Gram-negative bacteria, including Salmonella Typhimurium (S.Tm) and promotes expulsion of infected enterocytes. During the first ~12–24 h, this reduces mucosal S.Tm loads at the price of moderate enteropathy. It remained unknown how this NAIP/NLRC4-dependent tradeoff would develop during subsequent infection stages. In NAIP/NLRC4-deficient mice, S.Tm elicited severe enteropathy within 72 h, characterized by elevated mucosal TNF (>20 pg/mg) production from bone marrow-derived cells, reduced regeneration, excessive enterocyte loss, and a collapse of the epithelial barrier. TNF-depleting antibodies prevented this destructive pathology. In hosts proficient for epithelial NAIP/NLRC4, a heterogeneous enterocyte death response with both apoptotic and pyroptotic features kept S.Tm loads persistently in check, thereby preventing this dire outcome altogether. Our results demonstrate that immediate and selective removal of infected enterocytes, by locally acting epithelium-autonomous NAIP/NLRC4, is required to avoid a TNF-driven inflammatory hyper-reaction that otherwise destroys the epithelial barrier.

Host control of infections crucially depends on the capability to kill pathogens with reactive oxygen species (ROS). However, these toxic molecules can also readily damage host components and cause severe immunopathology. Here, we show that neutrophils use their most abundant granule protein, myeloperoxidase, to target ROS specifically to pathogens while minimizing collateral tissue damage. A computational model predicted that myeloperoxidase efficiently scavenges diffusible HO at the surface of phagosomal Salmonella and converts it into highly reactive HOCl (bleach), which rapidly damages biomolecules within a radius of less than 0.1 μm. Myeloperoxidase-deficient neutrophils were predicted to accumulate large quantities of HO that still effectively kill Salmonella, but most HO would leak from the phagosome. Salmonella stimulation of neutrophils from normal and myeloperoxidase-deficient human donors experimentally confirmed an inverse relationship between myeloperoxidase activity and extracellular HO release. Myeloperoxidase-deficient mice infected with Salmonella had elevated hydrogen peroxide tissue levels and exacerbated oxidative damage of host lipids and DNA, despite almost normal Salmonella control. These data show that myeloperoxidase has a major function in mitigating collateral tissue damage during antimicrobial oxidative bursts, by converting diffusible long-lived HO into highly reactive, microbicidal and locally confined HOCl at pathogen surfaces.

In Salmonella infection, the Salmonella pathogenicity island-2 (SPI-2)-encoded type three secretion system (T3SS2) is of key importance for systemic disease and survival in host cells. For instance, in the streptomycin-pretreated mouse model SPI-2-dependent Salmonella replication in lamina propria CD11c(-)CXCR1(-) monocytic phagocytes/macrophages (MΦ) is required for the development of colitis. In addition, containment of intracellular Salmonella in the gut critically depends on the antimicrobial effects of the phagocyte NADPH oxidase (PHOX), and possibly type 2 nitric oxide synthase (NOS2). For both antimicrobial enzyme complexes, oxygen is an essential substrate. However, the amount of available oxygen upon enteroinvasive Salmonella infection in the gut tissue and its impact on Salmonella-MΦ interactions was unknown. Therefore, we measured the gut tissue oxygen levels in a model of Salmonella enterocolitis using luminescence two-dimensional in vivo oxygen imaging. We found that gut tissue oxygen levels dropped from ∼78 Torr (∼11% O2) to values of ∼16 Torr (∼2% O2) during infection. Because in vivo virulence of Salmonella depends on the Salmonella survival in MΦ, Salmonella-MΦ interaction was analysed under such low oxygen values. These experiments revealed an increased intracellular replication and survival of wild-type and t3ss2 non-expressing Salmonella. These findings were paralleled by blunted nitric oxide and reactive oxygen species (ROS) production and reduced Salmonella ROS perception. In addition, hypoxia enhanced SPI-2 transcription and translocation of SPI-2-encoded virulence protein. Neither pharmacological blockade of PHOX and NOS2 nor impairment of T3SS2 virulence function alone mimicked the effect of hypoxia on Salmonella replication under normoxic conditions. However, if t3ss2 non-expressing Salmonella were used, hypoxia did not further enhance Salmonella recovery in a PHOX and NOS2-deficient situation. Hence, these data suggest that hypoxia-induced impairment of antimicrobial activity and Salmonella virulence cooperate to allow for enhanced Salmonella replication in MΦ.