Establishing causal links between bacterial metabolites and human intestinal disease is a significant challenge. This study reveals the molecular basis of antibiotic-associated hemorrhagic colitis (AAHC) caused by intestinal resident Klebsiella oxytoca. Colitogenic strains produce the nonribosomal peptides tilivalline and tilimycin. Here, we verify that these enterotoxins are present in the human intestine during active colitis and determine their concentrations in a murine disease model. Although both toxins share a pyrrolobenzodiazepine structure, they have distinct molecular targets. Tilimycin acts as a genotoxin. Its interaction with DNA activates damage repair mechanisms in cultured cells and causes DNA strand breakage and an increased lesion burden in cecal enterocytes of colonized mice. In contrast, tilivalline binds tubulin and stabilizes microtubules leading to mitotic arrest. To our knowledge, this activity is unique for microbiota-derived metabolites of the human intestine. The capacity of both toxins to induce apoptosis in intestinal epithelial cells—a hallmark feature of AAHC—by independent modes of action, strengthens our proposal that these metabolites act collectively in the pathogenicity of colitis.

No abstract

Objectives: Klebsiella oxytoca is a gastrointestinal pathobiont with the potential to produce the toxins tilivalline and tilimycin, which cause antibiotic-associated hemorrhagic colitis. Overgrowth of toxigenic K oxytoca has recently been implicated in necrotizing enterocolitis. K oxytoca colonizes 2-9% of healthy adults, however, there is no systematic data on colonization in healthy children. We investigated K oxytoca colonization and its toxigenic properties in healthy infants. Methods: We sampled stool of healthy infants and determined K oxytoca colonization using stool culture and PCR ( pehX). Toxin in stool was measured with HPLC/high-resolution mass spectrometry. K oxytoca isolates were typed using multi-locus sequence typing (MLST) and Koxytoca toxin PCR (npsA/B). Cytotoxin production of isolates was analyzed by MTT assay. Results: K oxytoca was detected in 30 of 61 infants (49%) using stool culture and in 45 of 61 (73%) using PCR ( pehX). Toxin marker PCR (npsA/B) was positive in 66% of stool samples positive for K oxytoca PCR. Stool toxin levels were too low for quantitation but traces of tilivalline were detected. Contrarily, 49% of K oxytoca isolates demonstrated toxicity in the MTT assay. MLST revealed 36 distinct sequence types affiliated with all known K oxytoca sequence type clusters (A, B1 and B2). Conclusions: More than 70% of healthy infants were colonized with K oxytoca. Toxin quantities in stool of colonized healthy infants were below detection level, yet half of the isolates produced toxin in vitro demonstrating their pathobiont potential. The high occurrence of toxigenic K oxytoca in healthy infants has to be considered for future disease association studies.

Klebsiella spp. that secrete the DNA-alkylating enterotoxin tilimycin colonize the human intestinal tract. Numbers of toxigenic bacteria increase during antibiotic use, and the resulting accumulation of tilimycin in the intestinal lumen damages the epithelium via genetic instability and apoptosis. Here we examine the impact of this genotoxin on the gut ecosystem. 16S rRNA sequencing of faecal samples from mice colonized with Klebsiella oxytoca strains and mechanistic analyses show that tilimycin is a pro-mutagenic antibiotic affecting multiple phyla. Transient synthesis of tilimycin in the murine gut antagonized niche competitors, reduced microbial richness and altered taxonomic composition of the microbiota both during and following exposure. Moreover, tilimycin secretion increased rates of mutagenesis in co-resident opportunistic pathogens such as Klebsiella pneumoniae and Escherichia coli, as shown by de novo acquisition of antibiotic resistance. We conclude that tilimycin is a bacterial mutagen, and flares of genotoxic Klebsiella have the potential to drive the emergence of resistance, destabilize the gut microbiota and shape its evolutionary trajectory.

Different humanized mouse models have been developed to study human diseases such as autoimmune illnesses, cancer and viral infections. These models are based on the use of immunodeficient mouse strains that are transplanted with human tissues or human immune cells. Among the latter, mice transplanted with hematopoietic stem cells have been widely used to study human infectious diseases. However, mouse models built upon the transplantation of donor-specific mature immune cells are still under development, especially in the field of viral infections. These models can retain the unique immune memory of the donor, making them suitable for the study of correlates of protection upon natural infection or vaccination. Here, we will review some of these models and how they have been applied to virology research. Moreover, the future applications and the potential of these models to design therapies against human viral infections are discussed.