BACKGROUND Propofol, midazolam and ketamine are widely used in today's anaesthesia practice. Both neuroprotective and neurotoxic effects have been attributed to all three agents.

OBJECTIVE To establish whether propofol, midazolam and ketamine in the same neuronal injury model exert neuroprotective effects on injured neurones in vitro and in vivo by modulation of the Toll-like receptor 4-nuclear factor kappa-light-chain-enhancer of activated B cells (TLR-4-NF-κB) pathway.

DESIGN AND SETTING Cell-based laboratory (n= 6 repetitions per experiment) and animal (n= 6 per group) studies using a neuronal cell line (SH-SY5Y cells) and adult Sprague-Dawley rats.

INTERVENTIONS Cells were exposed to oxygen-glucose deprivation before or after treatment using escalating, clinically relevant doses of propofol, midazolam and ketamine. In animals, retinal ischaemia (60 min) was induced followed by reperfusion and randomised treatment with saline or propofol.

MAIN OUTCOME MEASURES Neuronal cell death was determined using flow-cytometry (mitochondrial membrane potential) and lactate dehydrogenase (LDH) release. Nuclear factor NF-κB and hypoxia-inducible factor 1 α-activity were analysed by DNA-binding ELISA, expression of NF-κB-dependent genes and TLR-4 by luciferase-assay and flow-cytometry, respectively. In animals, retinal ganglion cell density, caspase-3 activation and gene expression (TLR-4, NF-κB) were used to determine in vivo effects of propofol. Results were compared using ANOVA (Analysis of Variance) and t test. A P value less than 0.05 was considered statistically significant.

RESULTS Post-treatment with clinically relevant concentrations of propofol (1 to 10 μg ml− 1) preserved the mitochondrial membrane potential in oxygen-glucose deprivation-injured cells by 54% and reduced LDH release by 21%. Propofol diminished TLR-4 surface expression and preserved the DNA-binding activity of the protective hypoxia-inducible factor 1 α transcription factor. DNA-binding and transcriptional NF-κB-activity were inhibited by propofol. Neuronal protection and inhibition of TLR-4-NF-κB signalling were not consistently seen with midazolam or ketamine. In vivo, propofol treatment preserved rat retinal ganglion cell densities (cells mm− 2, saline 1504±251 vs propofol 2088±144, P= 0.0001), which was accompanied by reduced neuronal caspase-3, TLR-4 and NF-κB expression.

CONCLUSION Propofol, but neither midazolam nor ketamine, provides neuroprotection to injured neuronal cells via inhibition of TLR-4-NF-κB-dependent signalling.