Colistin is an antibiotic of last resort, but has poor efficacy and resistance is a growing problem. Whilst it is well established that colistin disrupts the bacterial outer membrane by selectively targeting lipopolysaccharide (LPS), it was unclear how this led to bacterial killing. We discovered that MCR-1 mediated colistin resistance in Escherichia coli is due to modified LPS at the cytoplasmic rather than outer membrane. In doing so, we also demonstrated that colistin exerts bactericidal activity by targeting LPS in the cytoplasmic membrane. We then exploited this information to devise a new therapeutic approach. Using the LPS transport inhibitor murepavadin, we were able to cause LPS accumulation in the cytoplasmic membrane of Pseudomonas aeruginosa, which resulted in increased susceptibility to colistin in vitro and improved treatment efficacy in vivo. These findings reveal new insight into the mechanism by which colistin kills bacteria, providing the foundations for novel approaches to enhance therapeutic outcomes.

Hepatocellular carcinoma (HCC) is the third most frequent cause of cancer-related death. The immune-rich contexture of the HCC microenvironment makes this tumour an appealing target for immune-based therapies. Here, we discuss how the functional characteristics of the liver microenvironment can potentially be harnessed for the treatment of HCC. We will review the evidence supporting a therapeutic role for vaccines, cell-based therapies and immune-checkpoint inhibitors and discuss the potential for patient stratification in an attempt to overcome the series of failures that has characterised drug development in this disease area.

Vibrio pathogenicity island-2 (VPI-2) is a 57-kb region integrated at a transfer RNA (tRNA)-serine locus that encompasses VC1758 to VC1809 on the V. cholerae N16961 genome and is present in pandemic isolates. VPI-2 encodes a P4-like integrase, a restriction modification system, a Mu phage-like region, and a sialic acid metabolism region, as well as neuraminidase (VC1784), which is a glycosylhydrolase known to release sialic acid from sialoglycoconjugates to unmask GM1 gangliosides, the receptor for cholera toxin. We examined the tRNA-serine locus among the sequenced V. cholerae genomes and identified five variant VPI-2 regions, four of which retained the sialometabolism region. Three variant VPI-2 regions contained a type three secretion system. By using an inverse nested PCR approach, we found that the VPI-2 region can form an extrachromosomal circular intermediate (CI) molecule after precise excision from its tRNA-serine attachment site. We constructed a knockout mutant of VC1758 (int) with V. cholerae strain N16961 and found that no excision PCR product was produced, indicating that a functional cognate, VPI-2 integrase, is required for excision. The Vibrio seventh pandemic island-I (VSP-I) and VSP-II regions are present in V. cholerae O1 El Tor and O139 serogroup isolates. Novel regions are present at the VSP-I insertion site in strain MZO-3 and at the VSP-II insertion site in strain 623-39. VSP-II is a 27-kb region that integrates at a tRNA-methionine locus, is flanked by direct repeats, and encodes a P4-like integrase. We show that VSP-II can excise and form a CI and that the cognate VSP-II integrase is required for excision. Interestingly, VSP-I is not inserted at a tRNA locus and does encode a XerDC-like recombinase, but similar to VPI-2 and VSP-II, VSP-I does excise from the genome to form a CI. These results show that all three pathogenicity islands can excise from the chromosome, which is likely a first step in their horizontal transfer.