Keratin 8 and 18 (K8/K18) mutations have been implicated in the aetiology of certain pathogenic processes of the liver and pancreas. While some K8 mutations (K8 G62C, K8 K464N) are also presumed susceptibility factors for inflammatory bowel disease (IBD), the only K18 mutation (K18 S230T) discovered so far in an IBD patient is thought to be a polymorphism. The aim of our study was to demonstrate that these mutations might also directly affect intestinal cell barrier function. Cell monolayers of genetically engineered human colonocytes expressing these mutations were tested for permeability, growth rate and resistance to heat-stress. We also calculated the change in dissociation constant (Kd, measure of affinity) each of these mutations introduces into the keratin protein, and present the first model of a keratin dimer L12 region with in silico clues to how the K18 S230T mutation may affect keratin function. Physiologically, these mutations cause up to 30% increase in paracellular permeability in vitro. Heat-stress induces little keratin clumping but instead cell monolayers peel off the surface suggesting a problem with cell junctions. K18 S230T has pronounced pathological effects in vitro marked by high Kd, low growth rate and increased permeability. The latter may be due to the altered distribution of tight junction components claudin-4 and ZO-1. This is the first time intestinal cells have been suggested also functionally impaired by K8/K18 mutations. Although an in vitro colonocyte model system does not completely mimic the epithelial lining of the intestine, nevertheless the data suggest that K8/K18 mutations may be also able to produce a phenotype in vivo.