PLoS ONE (Dec 2010)

Analysis of interactions of Salmonella type three secretion mutants with 3-D intestinal epithelial cells.

  • Andrea L Radtke,
  • James W Wilson,
  • Shameema Sarker,
  • Cheryl A Nickerson

DOI
https://doi.org/10.1371/journal.pone.0015750
Journal volume & issue
Vol. 5, no. 12
p. e15750

Abstract

Read online

The prevailing paradigm of Salmonella enteropathogenesis based on monolayers asserts that Salmonella pathogenicity island-1 Type Three Secretion System (SPI-1 T3SS) is required for bacterial invasion into intestinal epithelium. However, little is known about the role of SPI-1 in mediating gastrointestinal disease in humans. Recently, SPI-1 deficient nontyphoidal Salmonella strains were isolated from infected humans and animals, indicating that SPI-1 is not required to cause enteropathogenesis and demonstrating the need for more in vivo-like models. Here, we utilized a previously characterized 3-D organotypic model of human intestinal epithelium to elucidate the role of all characterized Salmonella enterica T3SSs. Similar to in vivo reports, the Salmonella SPI-1 T3SS was not required to invade 3-D intestinal cells. Additionally, Salmonella strains carrying single (SPI-1 or SPI-2), double (SPI-1/2) and complete T3SS knockout (SPI-1/SPI-2: flhDC) also invaded 3-D intestinal cells to wildtype levels. Invasion of wildtype and TTSS mutants was a Salmonella active process, whereas non-invasive bacterial strains, bacterial size beads, and heat-killed Salmonella did not invade 3-D cells. Wildtype and T3SS mutants did not preferentially target different cell types identified within the 3-D intestinal aggregates, including M-cells/M-like cells, enterocytes, or Paneth cells. Moreover, each T3SS was necessary for substantial intracellular bacterial replication within 3-D cells. Collectively, these results indicate that T3SSs are dispensable for Salmonella invasion into highly differentiated 3-D models of human intestinal epithelial cells, but are required for intracellular bacterial growth, paralleling in vivo infection observations and demonstrating the utility of these models in predicting in vivo-like pathogenic mechanisms.