The impact of systemic inflammation and bacterial infection on the blood-brain barrier transport of colistin

The spread of multidrug-resistant (MDR) infections caused by Gram-negative bacteria such as Pseudomonas aeruginosa, Acinetobacter baumannii, and Klebsiella pneumoniae, has led to increasing use of colistin, an old antibiotic from the polymyxin family. Colistin (also known as polymyxin E) is a cationic polypeptide antibiotic whose clinical use waned in the 1970s due to concerns related to its neurotoxicity. Though this toxicity may have been exaggerated in the past due to a lack of understanding of colistin pharmacology and the use of inappropriate doses, it has nonetheless attracted the attention of clinicians and scientists. Whether this neurotoxicity is centrally or peripherally mediated remains unclear, however, if colistin were to exert any centrally mediated toxicity following systemic administration, it would require transport across the blood-brain barrier (BBB). Therefore, this thesis investigates the brain uptake of colistin under healthy conditions, in addition to assessing the impact of systemic inflammation and bacterial infection on the brain uptake of this polypeptide antibiotic. In order to undertake this project, it was necessary to develop a liquid chromatography method to quantify colistin concentrations in mouse brain homogenate. Our method consisted of protein precipitation of mouse brain homogenates with trichloroacetic acid (TCA), solid-phase extraction (SPE) of colistin, and derivatization with fluorenylmethyloxycarbonyl chloride (FMOC-Cl), followed by liquid chromatography with fluorescence detection. A linear correlation between peak area and colistin concentration was observed over the concentration range from 93.8 to 3,000 ng/g in brain tissue (R2 > 0.994). Intra- and inter-day coefficients of variation were 5.1 to 8.3% and 5.8 to 8.5%, respectively, suggesting the method was reliable and sensitive enough to quantify the amount of colistin that had penetrated the BBB following systemic administration to mice. Following a single intravenous dose of colistin sulfate at 5 mg/kg in healthy Swiss outbred mice, brain-to-plasma (B:P) ratios of colistin were very low ranging from 0.03 to 0.06. To determine whether higher plasma concentrations would lead to higher brain uptake of colistin, colistin sulfate (40 mg/kg) was administered subcutaneously to Swiss outbred mice as single and multiple doses. The brain uptake of colistin with these dosing regimens was still low with average B:P ratios of 0.03 and 0.02. To assess whether P-glycoprotein (P-gp) at the BBB was responsible for the limited brain uptake of colistin, colistin sulfate (5 mg/kg) was intravenously co-administered with P-gp inhibitors (i.e. PSC833 or GF120918) (10 mg/kg). However, the BBB transport of colistin was not significantly enhanced in the presence of these inhibitors, suggesting that the negligible BBB transport of colistin in healthy mice was not a result of the P-gp transporter function at the BBB. Given colistin is not administered to healthy patients, but rather to patients with infections, it was decided to investigate whether the brain uptake of colistin altered during disease state. Systemic inflammation was induced in mice by three intraperitoneal injections of lipopolysaccharide (LPS; Salmonella enterica 3 mg/kg). Colistin concentration in brain homogenate was measured following subcutaneous administration of colistin sulfate and following transcardiac perfusion of colistin sulfate using a modified in situ brain perfusion technique. Significantly increased brain uptake of colistin was observed in LPS-treated animals compared with that in saline-treated animals. Following subcutaneous administration of colistin sulfate, the value of the area under the brain colistin concentration versus time curve (AUCbrain) from 0 to 4 h was 11.7 ± 2.7 µg•h/g and 4.0 ± 0.3 µg•h/g in LPS- and saline-treated mice, even though plasma exposure of colistin was no different between these two treatments. Similarly, in situ perfusion of colistin led to higher colistin brain concentrations in LPS-treated animals than in saline-treated animals, with colistin brain-to-perfusate concentration ratios of 0.019 ± 0.001 and 0.014 ± 0.001, respectively. The B:P ratio of a BBB integrity marker (i.e. 14C-sucrose) was also 2-fold higher in LPS-treated mice following a single intravenous dose (2 µCi), suggesting the enhanced brain uptake of colistin during systemic inflammation was likely due to disruption of the BB paracellular route. To further assess the impact of systemic bacterial infection on BBB integrity and colistin brain uptake, bacteremia was established 8 h after intramuscular administration of Pseudomonas aeruginosa to mice, at which time a single intravenous dose of 14C-sucrose (2 µCi) or subcutaneous dose of colistin (40 mg/kg) was administered. Plasma levels of the pro-inflammatory cytokines tumor necrosis factor-α (TNF-α), interleukin-1β (IL-1β) and interleukin-6 (IL-6) were also measured to assess whether there was any relationship between the degree of BBB disruption and plasma cytokine levels. Despite a substantial elevation of these three cytokines in plasma, the brain uptake of colistin was low with average B:P ratios ranging from 0.012 to 0.044 and the B:P ratio of 14C-sucrose was no different between infected and non-infected mice, suggesting a lack of effect of bacteremia on BBB integrity. The results demonstrated that the brain uptake of colistin was not increased during P. aeruginosa-induced systemic bacteremia and there did not appear to be a direct relationship between BBB disruption and the plasma levels of three pro-inflammatory cytokines. The above findings suggested that there might be some LPS-specific effects on the BBB, given that LPS from S. enterica induced BBB disruption but bacterial infection with P. aeruginosa had no effect. To address this, the brain uptake of 14C-sucrose or colistin was assessed in mice treated with LPS from P. aeruginosa using the same dosing regimen as LPS from S. enterica. No difference was found between P. aeruginosa LPS- or saline-treated mice regarding to the B:P ratios of colistin or 14C-sucrose (p < 0.05), confirming the possibility of LPS-dependent effects on BBB disruption. To further clarify this, immortalized human brain capillary endothelial cells (hCMEC/D3) were treated with LPS from P. aeruginosa or S. enterica at pathologically-relevant concentrations of 3.75 to 30 µg/mL. The claudin-5 or β-actin bands were visualized following western blotting and the ratios of claudin-5: β-actin were then quantified using Image J analysis software. LPS (S. enterica) was found to significantly decrease the claudin-5 expression by approximately 33.3% at a relative low concentration (i.e. 7.5 µg/mL) (p < 0.05) while LPS (P. aeruginosa) was found to affect the claudin-5 expression only at the highest experimental concentration (i.e. 30 µg/mL). These results indicated that BBB disruption induced by LPS is species and dose-dependent, suggesting that different bacterial infections may induce varying disruptive effects on the BBB. In addition to its bactericidal effect, colistin has also been shown to bind to bacterial LPS and prevent the pathological effects of bacterial endotoxin in the circulation. Hence, it was hypothesized that colistin could have the potential to reverse the LPS-induced BBB disruption observed in earlier studies from this thesis. In hCMEC/D3 cells treated with LPS (S. enterica) and colistin simultaneously, the claudin-5:β-actin ratio was 1.09 ± 0.07, which was not different from that of medium-treated cells (p > 0.05). The results suggested that colistin was able to ameliorate the disrupting effect of LPS on claudin-5 expression, potentially the BBB integrity. To confirm whether any reversal effect of colistin observed in vitro was reflective of the in vivo setting, the brain uptake of 14C-sucrose was measured in mice treated with LPS (S. enterica) in the presence and absence of colistin sulfate. While LPS caused an 1.3-fold increase in brain uptake of 14C-sucrose, the B:P ratio of 14C-sucrose in mice colistin and S. enterica LPS simultaneously was 0.026 ± 0.001. These results suggested that the disruptive effect of LPS on the BBB can be rescued by co-administration of colistin. Overall, this body of work demonstrates that (1) the brain uptake of colistin following systemic administration is low in healthy conditions regardless of the plasma concentration and the functional activity of P-gp; (2) the brain uptake of colistin following systemic administration is significantly increased during systemic inflammation due to disruption of the BBB paracellular route, however, but not in the presence of P. aeruginosa infection likely due to bacterial species-dependent effects on the BBB; (3) the BBB-disrupting effect of LPS is dose- and species-dependent; (4) colistin is able to ameliorate BBB disruption induced by LPS from S. enterica. This thesis provided significant insight into the mechanisms associated with the BBB transport of colistin. Important findings about BBB transport of colistin under disease state and potential mechanisms of BBB disruption during systemic inflammation induced by different bacterial species have been provided.