|Home | About | Journals | Submit | Contact Us | Français|
The transcription factor NFκB is a major mediator of lipopolysaccharide (LPS) signaling. We determined the role of NFκB activation in regulatory changes of the P-glycoprotein (Pgp) drug efflux transporter at the blood-brain barrier (BBB) and proinflammatory cytokine receptors.
We treated NFκB knockout and wildtype mice with LPS or vehicle, obtained enriched cerebral microvessels, and determined target mRNA by qPCR for MDR1a/b, IL15Rα, IL2 Rα, IL2Rγ, LIFR, gp130, and TNFR1/2, and protein expression by western blotting for P-gp, IL15Rα, IL2Rγ, LIFR, and gp130.
The effects of LPS on the transporters and cytokine receptors showed differences between wildtype and NFκB knockout mice, and between mRNA and protein changes. NFκB not only mediated the LPS-induced increase of MDR1b, IL2Rγ, and TNFR2 mRNA in the wildtype mice, but it showed opposite effects by elevating IL15Rα and TNFR1 mRNA and decreasing IL2Ra in the knockout mice. Although basal vinblastine uptake was unchanged in the NFκB knockout mice, LPS induced an increase of the uptake (depressed efflux transport) greater than that seen in the wildtype mice, indicating that NFκB helps to maintain Pgp efflux transporter function.
The results show differential involvement of NFκB signaling in response to LPS at the BBB.
The multidrug resistance (MDR) gene encodes a class of proteins that belong to the ATP-binding cassette (ABC) transporter family. At the level of the blood-brain barrier (BBB), the MDR1b product P-glycoprotein (Pgp) is distributed on the apical membrane of cerebral endothelia, the main constituent of the BBB. Pgp hinders the blood-to-CNS influx of a variety of substrates, including several classes of agents used in the treatment of infection, neurodegeneration, epilepsy, and brain tumors . In cultured RBE4 cerebral endothelial cells treated with tumor necrosis factor a (TNF), we have shown that the MDR1b gene promoter is exclusively regulated by transcriptional activation and nuclear translocation of nuclear factor (NF)-κB [2, 3]. If this is a universal phenomenon during inflammation in-vivo, the targeting of NFκB activation would have therapeutic potential by increasing central nervous system (CNS) drug delivery. In this study, we determined the regulatory changes of MDR1b in response to lipopolysaccharide (LPS), an endotoxin produced by Gram negative bacteria and commonly used as a proinflammatory stimulus in animal models. The effect of NFκB was tested by use of knockout (KO) mice deficient in the p50 subunit of NFκB that is essential for the activation of the NFκB complex [4, 5].
Besides the drug efflux transporters, neuroinflammation also modulates the level of cytokine receptor expression at the BBB level. This in turn affects the permeation of specific cytokines across the BBB by specific transport systems, or their signaling in endothelia with modulation of CNS function by release of secondary mediators. LPS has been shown to regulate the transport of TNF , leukemia inhibitory factor (LIF) , and interleukin (IL)-15 . TNF, which can be induced by LPS, shows differential upregulation of the two receptors for LIF, with decreased LIFR (gp190)  and increased gp130 in cerebral endothelial cells . TNF and LIF cross the BBB by their respective receptor-mediated transport systems [10, 11, 12, 13]. IL15 permeation across the normal BBB is low but it can be activated by LPS . Thus, expression of the probable transporting receptors for these cytokines for blood-to-brain influx will provide a sharp contrast to the efflux transport by Pgp. The differential expression of cytokine receptors in response to LPS, as well as the efflux drug transporters, represents a complex regulatory network at the level of the BBB. Therefore, the involvement of NFκB signaling in the expression of cytokine receptors was studied in parallel.
The studies followed a protocol approved by the Institutional Animal Care and Use Committee. Homozygous p50 NFκB knockout mice (B6; 129P2-Nfkb1tm1bal.J, 002849, abbreviated as PKO in this study)  were used with their control B6.129PF2/J mice (stock number 100903). All were purchased from the Jackson Laboratory (Bar Harbor, ME), and maintained in group housing (3 – 4 mice/cage) with food and water ad lib under a 12 h light-dark cycle at 23 °C, 7:00 – 19:00 being the light period. Food intake, activity level, and body temperature were monitored throughout the study. At the time of tissue harvesting, the mice receiving LPS were in the recovery phase as shown by all of the parameters.
To determine the regulatory changes of MDR1 and cytokine receptors at the mRNA level, four groups of young adult male mice were studied (n = 6/group). They are wildtype or PKO mice receiving LPS or saline vehicle. LPS (2.5 mg/kg body weight, from Salmonella enterica serotype typhimurium, purified by phenol extraction; catalogue no L6511, Sigma, was dissolved in pyrogen-free normal saline (NS) at 5 mg/ml, and delivered to mice by intraperitoneal (ip) injection. The dosing was adapted from various reports in the literature [15, 16, 17, 18] and controls were injected with NS in the same volume. The mice were observed at least twice daily for food intake, weight loss, and general malaise, and care was taken to avoid hypothermia. At 48 h after the single injection, mice were anesthetized with ketamine/xylazine ip and decapitated. Cerebral cortical tissue was obtained for the capillary depletion procedure to obtain enriched microvessels as described previously [5, 19]. The effectiveness of enrichment has been verified by γ-glutamyl transpeptidase activity that showed more than 40-fold increase .
Total RNA was extracted from the enriched microvessels with an Absolutely RNA RT-PCR kit (Stratagene, La Jolla, CA), reversely transcribed, and subjected to qPCR by use of gene-specific primers and fluorescent probes (Table (Table1)1) on an Applied Biosystems (ABI 7900) analyzer (Foster City, CA). The housekeeping gene glyceraldehyde 3-phosphate dehydrogenase (GAPDH) was amplified for each sample to serve as an internal control. Standard curves for the target genes were generated. The levels of expression of target genes were normalized to that of GAPDH. The effects of strain (wildtype vs KO) and treatment (NS vs LPS) were determined by two-way analysis of variance (ANOVA), followed by the Bonferroni test.
To determine the effect of LPS treatment and NFκB deletion on the level of Pgp and cytokine receptor protein expression, a separate set of four groups of mice were used (n = 3/group). One sample each in the LPS-treated wildtype and PKO mice was unfortunately lost during processing. The enriched microvessels were obtained as described above, and lysed in protein lysis buffer [radioimmunoprecipitation assay (RIPA) buffer] containing complete protease inhibitor cocktail. After homogenization, sonication, and ultracentrifuge clearance, the protein lysate was collected. Protein concentration was determined by the bicinchoninic acid (BCA) assay. Forty μg of protein was loaded onto sodium dodecyl sulfate-polyacrylamide gels (SDS-PAGE). The electrophoresed proteins were transferred to nitrocellulose membranes and probed with antibodies against Pgp, IL 15Rα, IL2Rα, IL2Rγ TNFR1, TNFR2, LIFR, and gp130. After thorough wash, the membranes were incubated with horseradish peroxidase (HRP)-conjugated secondary antibody and enchanced chemiluminescence reagent, and the signals were developed on X-ray film. The differences between groups was determined by the NIH Image J program for densitometric analysis, and expressed as the signal/β-actin ratio. There was a single lane defect in the last PKO mouse after NS treatment and the average β-actin signal of the other two mice was used as the substitute value of β-actin for this lane in the final analysis.
For efflux drug transport assays, two groups of mice received ip injection of LPS or NS 48 h earlier (n = 3/group). The mouse was anesthetized, and the chest cavity was exposed to ligate the descending aorta for perfusion of 3H-vinblastine through the left cardiac ventricle after lacerations were made in the right atrium and bilateral jugular veins. After 1 min perfusion, the mouse was decapitated to obtain the brain. Solvable was used to dissolve the weighed brain tissue at 55 °C which was further processed by incubation with hydrogen peroxide and Formula 989 scintillation cocktail as previously described . After measurement of the radioactivity of the brain sample and perfusate in a Beckman β-counter, the brain/perfusate ratio of 3H-vinblastine in each group was determined.
Two-way ANOVA was performed to determine the effects of treatment (LPS vs NS) and strain (PKO vs wildtype) on the level of expression of the efflux transporter and cytokine receptors. One-way ANOVA was performed to determine the difference of western blotting signals and 3H-vinblastine uptake among the four groups, followed by Tukey's post-hoc test.
The mRNA for MDR1a differed significantly between the wildtype and PKO mice (p < 0.005). However, there was a significant interaction between strain and treatment (p < 0.05). Post-hoc analysis showed that LPS induced a significant increase of MDR1a in the wildtype mice (p < 0.05) but not in the PKO mice (Fig. (Fig.1A).1A). Significant effects of strain and treatment were seen for the mRNA for MDR1b. There was also a significant interaction between the two variables of strain and treatment (p < 0.0001). Post-hoc analysis showed a significant effect of LPS on increasing MDR1b in the wildtype mice (p < 0.001), but not in the PKO mice (Fig. (Fig.1B1B).
For IL15Rα mRNA, there was a significant effect of LPS (p < 0.001) and an interaction between treatment and strain (p < 0.005). Only the PKO mice showed a significant increase in response to LPS (p < 0.001, Fig. Fig.2A).2A). This suggests that NFκB activity provides tonic inhibition preventing upregulation of IL15Rα mRNA. By contrast, there were significant effects of strain (p < 0.05) and treatment (p < 0.05) on the cognate IL2Ra mRNA, and there was an absence of interactions of these two variables. LPS caused a significant reduction of IL2Ra mRNA in the PKO mice, but not in the wildtype mice (Fig. (Fig.2B).2B). Thus, mice without an active NFκB complex failed to maintain IL2Ra levels after LPS challenge. Since IL15Rα and IL2Ra mRNA showed opposite changes, we further determined the level of the co-receptor IL2Rγ. There were significant effects of strain and treatment (p < 0.001), and there were significant interactions between the two variables (p < 0.001). Only the wildtype mice responded to LPS by a significant elevation of IL2Rγ mRNA (p < 0.001; Fig. Fig.2C).2C). This indicates that LPS can only upregulate IL2Rγ in the presence of an intact NFκB complex.
In contrast to the robust effect of LPS on IL15/IL2 receptors and the mediatory role of NFκB, the gp190 specific receptor (LIFR, Fig. Fig.3A)3A) and the co-receptor gp130 for LIF (Fig. (Fig.3B)3B) did not show a significant change either by LPS treatment or PKO at the time point studied. This serves as an important negative control.
Both IL15 and LIF receptors can be regulated by TNF. To determine whether TNF receptors play mediatory roles on the effects of LPS, we measured the mRNA levels for TNFR1 (p55, or p60) and TNFR2 (p75, or p80). For TNFR1 mRNA, there was a significant overall effect of LPS treatment (p < 0.005) and PKO (p < 0.01). There was no significant interaction between the variables of treatment and strain. LPS induced a significant increase of TNFR1 mRNA only in the PKO mice (p < 0.01) but not in the wildtype mice (Fig. (Fig.4A4A).
By contrast, TNFR2 mRNA showed a significant overall increase in response to LPS (p < 0.0001), a strain effect (p < 0.05), and an interaction between treatment and strain (p < 0.05). LPS induced a significant increase of TNFR2 mRNA in both wildtype (p < 0.001) and PKO (p < 0.05) groups (Fig. (Fig.4B4B).
In the wildtype mice, LPS treatment appeared to induce an increase of Pgp, LIFR, IL15Rα, and IL2Rγ in the western blotting images (Fig. (Fig.5A).5A). The increase was significant for LIFR and IL2Rγ based on densitometric analysis (Fig. (Fig.5B).5B). Pgp showed a trend toward an increase (p = 0.09). In the PKO mice, there was a persistent, though non-significant increase of Pgp, but reduction of LIFR (p < 0.005) and non-significant decrease of gp130 and IL15Rα after LPS treatment. PKO alone (without LPS challenge) caused a decrease of IL2Rγ (p < 0.005). In the PKO mice, there was also a reduction of LIFR, gp130, and IL15Rα in response to LPS treatment.
The increased expression of MDR1a and 1b mRNA and elevation of Pgp in the wildtype mice in response to LPS suggests an accelerated efflux transport. Thus, we expected that the wildtype mice would show decreased brain uptake of vinblastine after LPS treatment. However, there was no significant change in the wildtype mice. Since the PKO mice also showed a tendency toward higher Pgp expression in cerebral microvessels, a lower vinblastine uptake after LPS treatment would also be expected from the PKO mice. Nonetheless, the results were opposite from the predictions. Vinblastine uptake was increased in the PKO mice after LPS treatment in comparison with the vehicle-treated group (p < 0.05). PKO mice also showed a trend (p = 0.08) toward a greater response to LPS than the wildtype mice with higher elevation of vinblastine uptake (Fig. (Fig.66).
NFκB is a ubiquitious transcriptional factor that plays an essential role in many biological functions, including reactions to neuroinflammation. It is well known than LPS induces NFκB activation after binding to its receptors such as CD14 and toll-like receptor (TLR)-4 [22, 23]. LPS also affects the cytokine network and efflux drug transporters [24, 25, 26]. However, the role of NFκB in mediating the effects of LPS on cytokine receptors and Pgp at the BBB level has not been fully eludicated. Here, by use of PKO mice deficient in the p50 subunit of NFκB, we showed that LPS increased the expression of Pgp and selective cytokine receptors in cerebral microvessels composing the BBB, a complex process with regulation at both mRNA and protein levels. We also showed differential involvement of NFκB that can induce either upregulation or downregulation of cytokine receptors and modulate the response to LPS. Furthermore, despite the higher level of expression, the efflux transport function of Pgp was depressed in LPS-treated PKO mice. These findings broaden our understanding of the active role of the BBB in response to LPS-induced neuroinflammation.
In the non-activated state, the two main subunits of NFκB - p50 and p65 - exist as a homodimer or heterodimer. The p50/p65, p50/p50, or p65/p65 subunits are located in the cytoplasm and bound by the inhibitory subunit IκB. Once activated, such as by LPS or cytokines, there is rapid phosphorylation, ubiquitination, and trafficking of IκB to proteasomes for degradation. Concurrently, the activated dimer is translocated to the nucleus where it binds to gene promoters and modulates transcriptional activation [27, 28]. Since the p65 subunit KO is lethal with intrautero death of mice , the commonly used embryonic PKO is the p50 subunit KO surviving to adulthood . These PKO mice have an altered basal metabolic rate, with lower body weight, higher food intake (/body weight), lower body temperature, more sleep time (including an increase of slow wave and rapid eye movement sleep), and less locomotor activity . Since a low dose of LPS causes a greater sickness response in the PKO mice than the wildtype mice, it can be concluded that NFκB protects mice against the detrimental consequences of LPS exposure , although compensatory changes after p50 deletion are also present .
We have shown that the proinflammatory cytokine TNF increases the expression of Pgp in cerebral endothelia  and activates MDR1b promoter activity exclusively by NFκB binding to the promoter region . To determine whether LPS acts through NFκB signaling at the BBB level, we treated wildtype and PKO mice with LPS before obtaining enriched cerebral microvessels 48 h later. It is evident that LPS increased both MDR1a and MDR1b mRNA, with a corresponding increase of Pgp protein in the wildtype mice. Although the activation of transcription of the efflux transporter was no longer present in the PKO mice, Pgp protein remained elevated in response to LPS. This suggests compensatory changes at the BBB in the absence of the p50 subunit, and is consistent with previous observations in sleep studies on these mice . Despite the elevation of mRNA and protein after LPS, the wildtype mice showed no major change in the efflux transport of vinblastine. The PKO mice even had an increase of vinblastine uptake by the brain, indicating depressed transport function regardless of the higher level of protein expression. This is supported by a recent study with a 3 mg/kg dose of LPS in CD1 mice by use of verapamil, another substrate for Pgp. The effect of LPS on increasing the uptake of 3H-verapamil was time-dependent, and not associated with disruption of the BBB shown by co-administered 14C-sucrose . The results also show differences of in-vivo BBB assays with cellular uptake measurements . This is an important finding; contrary to conventional beliefs, the efflux transport function can be depressed in severe neuroinflammation.
LPS increases the expression of TNF and IL15, among many other cytokines [24, 32, 33]. Therefore, we also determined the levels of their respective receptors in enriched cerebral microvessels. The changes revealed the expected complexity. For the IL 15 system in the wildtype mice, LPS failed to increase IL15Rα and its cognate receptor IL2Rα, but induced the co-receptor subunit IL2Rγ. This was more evident at the mRNA level than the protein level. In the PKO mice, LPS caused a major elevation of IL15Rα, decreased IL2Rα, and failed to affect the low basal level of IL2Rγ. This is also more apparent at the mRNA level. For the TNF system, LPS increased the mRNA of TNFR2 without affecting that of TNFR1 in the wildtype mice, but caused a significant increase in both receptor subtypes in the PKO mice. For all of these receptors, PKO did not significantly affect the basal level of expression. However, the changes resulting from p50 deletion were accentuated after LPS challenge. By contrast, the levels of LIFR and gp130, which showed distinctive changes in response to TNF [8, 9], were unchanged in either wildtype or PKO mice. The differential role of NFκB in the expression of these receptors reflects complex cellular mechanisms of regulation of different genes in the cytokine network.
Although the Pgp system is one of the major ATP-binding cassette transporters at the BBB , many other cytokines and receptors also undergo regulatory changes during inflammation. For instance, NFκB is essential in transcriptional activation of IL15 , but there was no change in IL15 mRNA at the BBB level in our preliminary study even though the BBB is activated during neuroinflammation, as shown by elevated γ-glutamyl transpeptidase activity after stimulation by LPS or TNF . Upregulation of TNF transport has also been shown in mice with experimental autoimmune encephalomyelitis . The upregulation of IL15 and TNF receptor subtypes that may mediate the blood-to-brain transport of these cytokines correlates with the increase of efflux transporter function despite the reduction of efflux transport of vinblastine. The cytokine cascades, secondary mediators, and altered transport of many substrates as well as cytokine themselves [37, 38, 39] all contribute to the sickness behavior and eventual functional outcome .
In summary, we showed that NFκB signaling at the BBB mediates the effect of LPS in upregulating Pgp efflux transporter expression and decreasing its transport function. We also showed that NFκB participates in the selective effects of LPS in regulating several receptor subtypes for IL15 and TNF without affecting those for LIF. This novel information provides further insight into the complexity of regulation at the BBB during neuroinflammation.
Grant support is provided by NIH (DK54880, NS45751, and NS62291).