Aims: Although alcohol drinking impairs the blood–brain barrier (BBB), the underlying mechanism is not fully understood. Thus, the effects of chronic ethanol drinking on the BBB were studied in vivo . Methods: Alcohol-preferring rats were given for 70 days free choice water and 15% ethanol. Then, they received LPS by i.p. injection. Efflux of [ 14 C]sucrose or [ 14 C]dextran was measured by their microinjection into the brain. Endothelial cells and neurons were isolated from the brain and analysed for mitogen-activated protein kinase (MAPK) and the tight-junction (TJ) protein phosphorylation, NFκB activation, mRNA levels of TJ proteins, inducible nitric oxide synthase, tumour necrosis factor α, interleukin-1 β (IL-1β), IL-10, CASPASE-8, and DNA damage. Results: LPS transiently increased [ 14 C]sucrose efflux in water drinking, while it caused a lasting increase in [ 14 C]sucrose and [ 14 C]dextran efflux in ethanol-drinking rats. The time-course of changes in the TJ correlated with (i) an increase in extracellular signal-regulated kinase (ERK), p38 mapk Jun-N-terminal Kinase (JNK), and TJ protein phosphorylation, (ii) RelA-p50 and p50-p50 activation, and (iii) a decrease in the TJ proteins' mRNA levels in endothelial cells and neurons. Apoptotic cells were detected in water drinking and LPS (WC-LPS) neurons at 24 h after LPS exposure. Neurons from Et-LPS rats did not exhibit apoptosis. Conclusions: LPS injection in WC-LPS rats transiently disrupted the BBB. Lack of JNK activation and CASPASE-8 may be responsible for lack of apoptosis in endothelial cells and vice versa in neurons. Chronic alcohol drinking in ethanol drinking and LPS (Et-LPS) rats augmented and dysregulated the LPS-induced BBB abnormalities but suppressed apoptosis in neurons.
Alcohol is most commonly used drug in the present society. Although short-term alcohol drinking causes euphoric and stress-relieving effects, numerous clinical and experimental studies have shown that alcohol use is a major risk factor for neurobehavioural diseases, inflammation disorders and enhanced susceptibility to bacterial infection (Bruno, 2003 ; Frank et al ., 2004 ; Singh and Jiang, 2004; Gauthier et al ., 2005 ; Gullo et al ., 2005 ). Alcohol abuse is also associated with the blood–brain barrier (BBB) impairment (Cornford et al ., 1982 ; Elmas et al ., 2001 ; Haorah et al ., 2005a , 2005b ) that may be a major contributing factor to various forms of neurodegenerative disorders. Despite seriousness of the alcohol-related disorders, mechanisms underlying the effects of alcohol on BBB are not fully understood.
The BBB plays an important role in maintaining a constant chemical environment into the brain including regulation of passage of ions, nutrients, and peptides into the brain. The BBB consists of endothelial cells interconnected by tight junctions (TJ) that effectively separate the brain from peripheral blood. The TJ are composed of transmembrane proteins including occluding (OCL), claudins (CLA), and junctional adhesion molecules (JAMs) that are linked to the f-actin cytoskeleton by zonula occludens (ZO) proteins (Hirase et al ., 1997 ; Morita et al ., 1998 ; Kneisel and Wolburg, 2000 ). The TJ proteins form rows of extensive, overlapping occlusions that block the entry of hydrophilic substances through intercellular routes. TJ integrity depends on high levels of TJ proteins in proper architecture that is determined by their phosphorylation rate (Albert et al ., 1997 ; Staddon et al ., 1997 ; Bazzoni et al ., 2000 ; Brown and Davis, 2002 ). Thus, the BBB endothelial cells, along with the barrier function, also participate in immune competent organ, release of inflammation mediators, and regulator of the CNS neurons (Arthur et al ., 1987 ; Persidsky et al ., 1999 ). Recent in vitro studies have shown that ethanol activated myosin light chain (MLC) kinase (MLCK) leading to phosphorylation of a number of TJ proteins that result in BBB impairment (Haorah et al ., 2005a , 2005b ). Formation of free radicals may also participate in the ethanol-induced BBB damage (Haorah et al ., 2005b ). However, further research is needed to decipher the effects of ethanol on TJ proteins and ensuing toxicity.
Therefore, the aim of this study was to investigate possible effects of ethanol on functional integrity of the BBB and the ensuing toxicity on the brain. We hypothesized that ethanol reduced expression and increased phosphorylation of the TJ proteins that disrupts the barrier. This allows toxins such as LPS to cross the BBB and induce inflammatory response in the brain. The BBB integrity was determined by measuring the Brain Efflux Index (BEI) of [ 14 C]sucrose or [ 14 C]dextran, mRNA levels and phosphorylation of tight junction proteins, expression of inflammation and apoptotic proteins, and activation of extracellular signal-regulated kinase (ERK), p38 mapk , Jun-N-terminal Kinase (JNK), and NFκB in neurons and endothelial cells.
Materials and Methods
All protocols for the animal experiments described in this study were carried out according to the Ethical Guidelines on Animal Experimentation and were approved by the Animal Usage Committee of the university. Male alcohol-preferring rats, weighing 100–150 g, were divided into three groups.
Group 1: Ethanol-drinking rats with (Et-LPS rats) or without LPS (Et rats)
Animals of this group received free choice water and 15% ethanol in water as described previously (Benlhabib et al ., 2004a , 2004b ). After 65 days of alcohol drinking, each rat was anaesthetized with a mixture of ketamine.HCl/diazepam (58:4.5 mg/kg). A cannula with a replaceable inner guide (CMA/Microdialysis, IN, USA) was implanted stereotaxically in the hippocampus, 3 mm above the final probe membrane position. The coordinates were 4.6 mm lateral and 5.6 mm posterior to bregma and 4.6 mm ventral starting from the dura. Post-operative analgesia was ensured by a single injection of ketoprofen (4 mg/kg, i.p.). Rats were allowed to recover from surgery for 5 days during which they continued to receive free choice ethanol and water. At Day 70 of alcohol drinking, each rat was injected 0.2 ml of 2% LPS (Et-LPS rats) or the vehicle alone (Et rats) into the peritoneal cavity. At different time intervals (0.5, 1, 6, 12, 24, and 48 h) after LPS (Sigma Chemicals, St Louis, MO) injection, the guide cannula was removed and replaced with a microinjection needle that was inserted into the cannula to a depth of 3 mm. A 0.5 μl of a mixture consisting of either [ 14 C]sucrose (0.05 μCi/rat) and [ 3 H]carboxyl-inulin (0.005 μCi/rat) or [ 14 C]dextran (0.05 μCi/rat) and [ 3 H]carboxyl-inulin (0.005 μCi/rat) prepared in HEPES physiological buffer [HEPES 10 mM (pH 7.4), NaCl 122 mM, KCl 3 mM, CaCl 2 1.4 mM, NaHCO 3 25 mM, MgSO 4 1.2 mM, KH 2 PO 4 0.4 mM, d -glucose 10 mM] was injected into the hippocampus region. At 1, 3, 5, and 7 min after the injection, rats were anaesthetized as above and an aliquot of CSF was taken from the cisterna magna. Then the rat was decapitated, the brain was removed and individually dissolved in 2 N NaOH (1:2 w/v). The 14 C and 3 H radioactivity was determined using scintillation counting. The elimination rate was calculated using the equation ‘100-BEI (%) = [(S b or D b /(CIn b )]/[(S i or D i )/(CIn i ] × 100’, where S b is amount of sucrose in the brain, D b is dextran amount in the brain, CIn is carboxyl inulin, CIn b is the reference amount in the brain, S i is sucrose injected, D i is dextran injected, and CIn i is the reference injected.
Neuronal toxicity study
Et rats and Et-LPS rats were given free choice ethanol and water as above and, after 70 days, they were individually injected with either LPS (Et-LPS rats) or the matrix alone (Et rats). Rats were anaesthetized at different time intervals after the injection and blood samples were collected using standard procedures. Blood was centrifuged and plasma was separated. CSF and plasma were either analysed immediately or stored frozen at −70°C. The rats were decapitated and the brain was removed from the skull. The two cerebral cortices were dissected out, washed with the HEPES physiological buffer, pooled and processed for isolation of endothelial cells and neurons (Singh, 1993 ).
Group 2: Water drinking with LPS (WC-LPS rats) or without LPS (WC-rats)
These rats received pure water ad lib throughout the experimental period. The remaining BBB and toxicity experiments were comparable to that for Group 1 rats.
Isolation of brain endothelial cells and neurons
Endothelial cells were isolated from rat brain as described previously (Gerhart et al ., 1988 ; Tontsch et al ., 1998). The two cortices were minced and processed as described previously (Singh, 1993 ).
RT–PCR and real-time fluorescence PCR assays
Total RNA was extracted from microvessels and neurons using a method described previously (Chomczynski and Sacchi, 1987 ; Singh, 1993 ). RT–PCR and real-time fluorescence PCR analysis was performed using the primers and probes listed below. The primers and probes were either from earlier studies (Youakim and Ahdieh, 1999 ; Li et al ., 2000 ; Miller et al ., 2003 ; Guillemot et al ., 2004 ; Singh et al ., 2006 ) or designed in-house. For RT–PCR, 1.0 μg of the total RNA was reverse transcribed into cDNA using 200 U of SuperScript II reverse transcriptase (Life Technologies, Inc.) in the presence of 5 μM random hexamers, 500 μM each deoxynucleotide triphosphate, 20 U of RNasin (Promega N2115), 50 mM Tris–HCl, 10 mM dithiothreitol, 75 mM KCl, and 3 mM MgCl 2 in a 20 μl volume at 42°C for 90 min. After cDNA synthesis, samples were treated with RNase H for 90 min at 37°C to remove residual template RNA. PCR used ∼10% (v/v) of the first strand (reverse transcriptase) reaction as a DNA template in the presence of 200 μM deoxynucleotide triphosphates, 2 mM MgCl 2 , and 2 U of AmpliTaq polymerase (Perkin Elmer Corp.) in a final volume of 50 μl. This reaction included primer sets of 100 pM of each forward and reverse primers listed above. [γ- 32 P]dATP (>5000 Ci/mmol; Amersham Redivue AA0018) 5′-end-labelled primers were used in these reactions. Amplification was performed for 30 cycles after 5 min of denaturation at 94°C, 1 min of annealing at 60°C, 1 min of primer extension at 72°C, and 7 min of final extension at 72°C after 30 cycles using a GeneAmp 9700 PCR System (Perkin Elmer Corp.). Then 15 μl of each PCR was separated on 5 or 10% (acrylamide:bisacrylamide, 30:1) Mini Protean II ready gels (Bio-Rad 161-1101 and 161-1109) in 0.5× TBE buffer. Gels were dried onto Whatman No. 1 filter paper (Bio-Rad gel dryer model 583) for 90 min at 80°C and quantified via signal transfer onto PhosphorImager storage screens.
ZO-1: S: 5′-CGGTCCTCTGAGCCTGTAAG-3′, SA: 5′-GGATCTACATGCGACGACAA-3′, P: 5′-FAM-TGGGTTTGCGACTAGTCGGTGGAAAT-TAMARA-3′
ZO-2: S: 5′-GCCAAAACCCAGAACAAAGA-3′, SA: 5′-ACTGCTCTCTCCCACCTCCT-3′, P: 5′-FAM-TGCCCGATACCGGGACACAGAATTAT-TAMARA-3′
ZO-3: S: 5′-ACCCTATGGCCTGGGCTTC-3′, AS: 5′-CCCGGGTACAACGTGTCC-3′, P: 5′-FAM-TGCAGAAGTTGACTGCTGAGATGCCT-TAMARA-3′,
OCL: S: 5′-AGTGGCTCAGGAGCTGCCATTGACTTCACC-3′, AS: 5′-GGTGGATATTCCCTGATCCAGTCGTCGTC-3′, P: 5′-FAM-TCTCTCGCCTGGATAAAGAATTGGATGA-TAMARA-3′,
JAM1: S: 5′-CTCCCCGCCGTCTGAATACAAGTGGT-3′, AS: 5′-CTTCGAGCATTGGGCTGGCTGTAAAT-3′, P: 5′-FAM-ACCGTGCCTAGCCAAGGATGAGATTT-TAMARA-3′,
CIN: S: 5′-CTAAACCGACTTCCTCGATTAA-3′, AS: 5′-TGTTRATGAGCGAGTCCACTG-3′, P: 5′-FAM-ACTCCAGACCTCCTTCGAGACCAGCA-TAMARA-3′,
Cad-2: S: 5′-TGGACCGAGAGAAAGTGCAACAGT-3′, AS: 5′-ACATCTGTCACTGTGATGACGGCT-3′, P: 5′-FAM-AGGCAATCCCACTTACGGCCTTTCAA-TAMARA-3′,
CLA-2: S: 5′-CCCAGGCCATGATGGTGA-3′, AS: 5′-TCATGCCCACCACAGAGATAAT-3′, P: 5′-FAM-AGGTTGAATTGCCAAGGATGCTCGCCA-TAMARA-3′,
CLA-6: S: 5′-ATGTGGAAGGTGACCGCCT-3′, AS: 5′-CCCTCCCACACCATCTGG-3′, P: 5′-FAM-TCACCCTTGGATGATGGAGCCAAAGA-TAMARA-3′,
CLA-7: S: 5′-CCTGGTGTTGGGCTTCTTAGC-3′, AS: 5′-CCCAACAGCGTGTGCACTTC-3′, P: 5′-FAM-CCAACATTAAGTATGAGTTTGGCCCTGC-TAMARA-3′,
IL-1β S: 5′-CGAACATGTCTTCCGTGATG-3′, AS: 5′-TCTCTGTCCTGGAGTTTGCAT-3′, P: 5′ FAM-TGTGATGCAGCCGTGCAGTCA-TAMARA-3′,
TNFα: S: 5′-CAC-GCT-CTT-CTG-CCT-GCT-G-3′, AS: 5′-GAT-GAT-CTG-ACT-GCC-TGG-GC-3′, P: 5′-FAM-CCA-GAG-GGA-AGA-GTT-CCC-CAG-GGA-TAMARA-3′,
IL-10: S: 5′-ACTTTAAGGGTTACCTGGGTTG-3′, AS: 5′-CGTGCTGTTTGATGTCTGG-3′, P: 5′-FAM-TTGGAGGAGGTGATGCCCCA-TAMARA-3′,
β -Actin: S: 5′-GCATGGAGTCCTGTGGCAT-3′, AS: 5′-GCCGATCCACACGGAGTACT-3′, P: 5′-FAM-AGATCAAGATCATTGCTCCTCCTGAGCGC-TAMARA-3′,
CASPASE-8: 5′-ATGGATCCATGGACTTCAGCAGAAATCTTTA-3′, 5′-CATCTAGATCAATCAGAAGGGAAGAC-3′, P: 5′-FAM-AACGGATGGGAAGGAAGCCTCTATCT-TAMARA-3′.
Real-time PCR was performed as described previously (Singh and Jiang, 2004). Comparative CT Method as described in User Bulletin #2 ABI Prism 7700 Sequence Detection System (Applied Biosystems; updated version 10/01), and standardizing levels to that of β-actin mRNA was used. Dissociation curves were examined to ensure the absence of secondary PCR products. Plasmid concentration was determined using the following equation: P mol DNA = μg DNA × (pmol/660 pg) × (10 6 pg/number of nucleotides). Using this equation, a 50 fmol/μl of 3735 pb DNA will represent 2.9E + 10 copies/μl. DNA copies in each sample was determine using a standard curve constructed by plotting number of copies on x -axis and threshold cycles on y -axis.
Proteins were extracted from brain endothelial cells in lysis buffer D under denaturing conditions (heating at 100°C for 5 min). The clear tissue extract containing 0.4 mg protein in 0.8 ml of immunoprecipitation buffer (10 mM Tris–HCl, pH 7.4, 150 mM NaCl, 1 mM EDTA, 1 mM EGTA, 1% Triton X-100, 0.5% Nonidet P-40, 0.33 mM PMSF, 1 mM vanadate, 2 μg/ml leupeptin, 10 μg/ml aprotinin, 10 μg/ml bestatin, 10 μg/ml pepstatin A) was incubated with 2 μg of biotin-conjugated anti-p-Tyr antibody. Immune complexes were isolated by precipitation using streptavidin-agarose (for 1 h at 4°C). Washed beads were suspended in 25 μl of Laemmli's sample buffer and heated at 100°C for 5 min. Extracts were immunoblotted using specific primary and horseradish peroxidase (HRP)-conjugated secondary antibodies.
Phosphorylation of JNK ERK and p38 mapk
The neurons were processed using M-PER Mammalian Protein Extraction Reagent (Pierce, IL) for extraction of proteins that were separated by SDS-PAGE and transferred onto nitrocellulose membranes. The blots were then washed in TBS-Tween [TTBS, 20 mM Tris–HCl buffer, pH 7.6, containing 137 mM NaCl and 0.05% (vol/vol) Tween 20], blocked for 2 h with 5% (wt/vol) non-fat dry milk, and incubated with phospho-specific antibodies to p46-p54 JNK, p42/p44 ERK, and p38 mapk . The membranes were washed in TTBS and incubated with a horseradish peroxidase-conjugated secondary anti-rabbit or anti-mouse antibody. Membranes were washed and stained with an appropriate substrate.
Analysis of LPS using MALDI-TOF-MS
Plasma and the cells were analysed for the bacterial toxins as described previously (Singh and Jiang, 2004). In brief, plasma and the cell homogenates were mixed with a solvent containing 90% phenol in distilled water. The mixture was incubated for 30 min at room temperature (RT) and dialysed against tap water for 5 h, then overnight against distilled water. The dialysed samples were concentrated under reduced pressure and nitrogen stream. The samples were run in the linear, negative mode on a Biflex III instrument (Bruker) using dihydroxybenzoic acid as the matrix. The method is described in detail elsewhere (Singh and Jiang, 2004b ).
DNA ladder assay for apoptosis
Apoptosis, at advance stages, is characterized by a fragmentation of the genomic DNA. These DNA fragments have a length of about 180 bp or multiples thereof, the characteristic DNA length of a nucleosome (DNA–histone complex). Endonucleases selectively cleave DNA at sites located between nucleosomal units (linker DNA). In agarose gel electrophoresis these DNA fragments are resolved to a distinctive ladder pattern. To analyse DNA fragmentation, neurons were collected and suspended in 15 μl of PBS at 42°C. DNA was extracted using an equal volume of 1.5% low melting point agarose pre-heated at 42°C (in PBS buffer containing 5.0 mM EDTA). The cell suspension was placed into 24-well cell culture plate incubated for 5 min at 4°C. Agarose-embedded cells were incubated for 1 h at 37°C with 500 μl of lysis buffer (20 mM Tris–HCl, pH 8.0, 20 mM EDTA, 50 mM NaCl) containing 0.5% sodium sarcosyl, 50 μg/ml RNase A, 100 μg/ml proteinase K. Agarose plugs containing deproteinised DNA were washed three times with 0.5× TBE buffer (0.045 M Tris–HCl, 0.045 M boric acid, 1.0 mM EDTA, pH 8.0) and quantitatively loaded to wells of 1% agarose gel. The wells were covered with 1.5% low melting point agarose and the DNA fragments were analysed at 5.0 V/cm for 2–3 h, using 0.5× TBE buffer.
The data were analysed by one-way ANOVA and the pair-wise comparison was performed using Student–Newman–Keuls post hoc test using SigmaStat (SPSS, Chicago, IL). Statistical probability of P < 0.05 was considered significant. The values are reported as mean ± SD.
[ 14 C]sucrose and [ 14 C]dextran efflux
The time profile of the remaining fraction (100-BEI %) of [ 14 C]sucrose and [ 14 C]dextran corrected by recovery of [ 3 H]carboxyl-inulin in hippocampus after microinjection are shown in Figs 1 and 2 . Less than 5% of the administered dose of sucrose or dextran was eliminated in 5 min. The slopes of the curves that were inversely related to the remaining fraction for the two compounds are shown in Figs 3 and 4 , respectively. The brains from WC rats were mostly impermeable to [ 14 C]sucrose (Fig. 3 , plot 1) and [ 14 C]dextran (Fig. 4 , plot 1). Ethanol drinking (Et rats) did not affect their permeability (Fig. 3 , plot 2 for [ 14 C]sucrose and Fig. 4 , plot 2 for [ 14 C]dextran). LPS injection in WC-LPS rats caused a significant, but transient increased [ 14 C]sucrose efflux (Fig. 3 , plot 3) with maximal increase occurring at 12 h after LPS. Then sucrose permeability decreased and returned to the basal levels within 48 h after the toxin exposure. LPS exposure in Wc-LPS rats was poorly effective in increasing dextran permeability (Fig. 4 , plot 3). LPS exposure in Et-LPS rats increased efflux of both [ 14 C]sucrose (Fig. 3 , plot 4) and [ 14 C]dextran (Fig. 4 , plot 4) that did not return to the control level within 48 h.
Endothelial TJ protein's mRNA levels
Figure 5 shows the mRNA bands in endothelial cells and Figure 6 shows the mRNA values obtained from real-time PCR analysis. According to the RT–PCR patterns, the TJ proteins were classified into two groups as follows: (1) proteins including ZO-1, ZO-2, OCL, CAD, and JAM whose expressions were altered, and (2) proteins including CIN and CLAs whose expressions were not altered by Et-LPS exposure. Figure 6A shows the time-course of change in ZO-1, one of the members of group-1 proteins whose mRNA levels were sensitive to Et-LPS exposure in endothelial cells isolated from the brain of WC, WC-LPS, Et and Et-LPS rats. The time-course of change in mRNA levels for other group-1 proteins were comparable to that of ZO-1 mRNA levels, thus they were not shown to avoid duplication. Figure 6B shows mRNA values for a group-2 TJ protein, CIN, whose expression did not change. This study showed that endothelial cells from WC rats had high levels of the TJ proteins (Fig. 6A , row 1, 0 h). Ethanol (Fig. 6A , row 2) exposure in Et rats poorly affected the TJ protein's mRNA levels. LPS exposure in WC-LPS rats decreased the ZO-1 (Figs 5 and 6A , row 3), ZO-2 (Fig. 5 , row 3), OCL (Fig. 5 , row 3), JAM (Fig. 5 , row 3), and CAD (Fig. 5 , row 3) mRNAs levels with lowest values occurring at 12 h after the exposure. Thereafter, the values increased gradually and returned to the basal levels within 48 h. Ethanol exposure augmented (Fig. 5 , row 4, and Fig. 6A , plot 4) the effects of LPS on the TJ protein's mRNA levels, although the levels did not return to the basal level in 48 h. Both RT–PCR and real-time PCR results showed that the mRNA levels for CIN and CLAs were not sensitive to ethanol and/or LPS exposure.
Phosphorylation of TJ proteins
Figure 7 (the phosphorylated bands) and Figure 8 (digitized band intensity) show the effects of ethanol and/or LPS on the TJ protein phosphorylation. Endothelial cells from the WC and Et rats exhibited minimal phosphorylation (bands or plots A and B) of all TJ proteins measured in this study (data are shown only for ZO-1 whose mRNA levels decreased and CIN whose mRNA levels did not change in response to LPS exposure to avoid duplication). Endothelial cells from WC-LPS rats exhibited transient increase in phosphorylation of TJ proteins that peaked at 12 h after and returned to the control level at 48 h after LPS exposure (band and plot C). Endothelial cells from Et-LPS rats exhibited sustained increase in phosphorylation of TJ proteins that peaked at 12 h and remained elevated thereafter (band and plot D).
LPS levels in neurons and endothelial cells
Figure 9A shows chromatographic separation of LPS extracted from spiked neurons. Figure 9B–D show analysis of neurons extracted from the brain of LPS injected ET, WC-LPS and Et-LPS rats, respectively. Figure 9E shows chromatographic separation of LPS extracted from plasma. Figure 9F–H show analysis of LPS in endothelial cells extracted from the brain of LPS injected ET, WC-LPS, and Et-LPS rats, respectively. LPS was present in endothelial cells but not detected in neurons.
Effects of ethanol drinking and LPS on endothelial cells and neurons
RT–PCR and real-time PCR results—endothelial cells from WC, WC-LPS, Et, and Et-LPS rats
Figure 10 shows expression of inducible nitric oxide synthase (iNOS), tumour necrosis factor α (TNFα), interleukin-1 β (IL-1β), IL-10, and CASPASE-8 mRNA using RT–PCR method and Fig. 11 shows the levels of mRNAs determined by the real-time PCR assay. At 1 h after LPS injection, LPS significantly increased iNOS ( Fig. 10 , row 1, and Fig. 11 , plot i), TNFα ( Fig. 10 , row 2, and Fig. 11 , plot i) and IL-1β ( Fig. 10 , row 3, and Fig. 11 , plot i) mRNA levels that gradually declined and returned to the basal level within 24 h after LPS injection. IL-10 ( Fig. 10 , row 4, and Fig. 11 , plot i) mRNA levels increased at 6 h after and peaked at 12–24 h after the toxin exposure. Then the levels decreased. LPS did not increase CASPASE-8 expression in endothelial cells. Ethanol drinking did not affect iNOS ( Fig. 10 , row 1, and Fig. 11 , plot ii), TNFα ( Fig. 10 , row 2, and Fig. 11 , plot ii) and IL-1β ( Fig. 10 , row 3, and Fig. 11 , plot ii) mRNA levels. However, ethanol drinking significantly augmented the effects of LPS on iNOS ( Fig. 10 , row 1, and Fig. 11 , plot iii), TNFα ( Fig. 10 , row 2, and Fig. 11 , plot iii) and IL-1β ( Fig. 10 , row 3, and Fig. 11 , plot iii) mRNA levels, but suppressed the effects of LPS on IL-10 ( Fig. 10 , row 4, and Fig. 11 , plot iii) and CASPASE-8 mRNA levels.
RT–PCR and real-time PCR results—neurons
In WC-LPS rats, LPS injection caused a transient increase in iNOS ( Fig. 10 , row 1, and Fig. 2 , plot 1), TNFα ( Fig. 10 , row 2, and Fig. 11 , plot i), and IL-1β ( Fig. 10 , row 3, and Fig. 11 , plot i) mRNA levels in the brain neurons. The values were highest at 1 and 6 h after and returned to the basal level at 48 h after LPS injection. IL-10 and CASPASE-8 mRNA levels began to increase at 6 h after and peaked at 12–24 h after the toxin exposure. Then the levels decreased. Ethanol drinking did not affect iNOS ( Fig. 10 , row 1, and Fig. 11 , plot ii) mRNA but increased TNFα ( Fig. 10 , row 2, and Fig. 11 , plot ii) and IL-1β ( Fig. 10 , row 3, and Fig. 11 , plot ii) mRNA levels in neurons. Ethanol drinking significantly augmented the effects of LPS on iNOS, TNFα, and IL-1β mRNA levels, but suppressed the effects of LPS on IL-10 and CASPASE-8 mRNA levels.
Activation of ERK, JNK, p38 mapk STAT, and NFκB dimers RelA-p50 and p50-p50 ( Fig. 12 )
In CW rats, LPS, at 1 h after its injection, increased phosphorylation of p42, p44, and p38, did not affect phosphorylation of p54 and p46, and increased binding of RelA-p50 to the probe in endothelial cells. LPS increased p50-p50 binding to the probe at 12 and 24 h after its injection. In Et rats, ethanol drinking increased phosphorylation of p44, did not affect phosphorylation of p42, p38, p54, and p46, and did not activate RelA-p50 or p50-p50. In Et-LPS rats, LPS injection upregulated activation of ERK, p38 mapk , and RelA-p50, but suppressed the activation of p50-p50.
In CW rats, LPS injection activated ERK p44 and p42, p38 mapk , and JNK p54 and p46. LPS also sequentially activated RelA-p50 followed by p50-p50. Neurons from Et rats activated p44 and p46 only. Neurons from Et-LPS rats exhibited upregulation of the effects of LPS on phosphorylation of ERK, p38 and JNK, and activation of RelA-p50, but blocks p50-p50 activation.
DNA ladder assay ( Fig. 13 )
LPS either alone or with ethanol did not cause apoptosis in endothelial cells. In control neurons, DNA fragmentation appeared at 6 h after and peaked at 24 h after LPS injection in CW-LPS rats. Chronic ethanol drinking in Et-LPS rats significantly suppressed the LPS-induced DNA fragmentation.
Long-term abuse of alcohol has been shown to cause profound functional and morphological changes in the BBB that is associated with a number of diseases and predisposition to infection (Banks, 1999 ). Despite seriousness of the alcohol-induced alterations in the BBB, possible mechanisms causing the BBB abnormalities are not fully known. Recently, a number of in vitro studies have shown that (i) a short-term (acute) and a long-term (chronic) ethanol exposure disrupted the endothelial TJ resulting in an increase in the BBB permeability (Haorah et al ., 2005a , 2005b ) and (ii) the effects of ethanol may be mediated through an alteration in the TJ protein architecture (Atkinson and Rao, 2001 ). This study further characterizes the effects of ethanol drinking on the BBB and the ensuing neuronal toxicity in control or LPS-injected P-rats.
Permeability of the BBB
The brain microvascular endothelial cells are interconnected by tight junctions that effectively separate the brain from peripheral blood. The BBB is relatively impermanent to hydrophilic molecules, leukocytes, and proteins for which there are no specific transporters. However, many disease causing bacteria are able to penetrate the BBB and infect the brain (Pialoux et al ., 1997 ; Badger and Kim, 1998 ; Badger et al ., 2000 ; Stins et al ., 2001 ; Zysk et al ., 2001 ; Strelow et al ., 2002 ; Deli et al ., 2005 ; Kim et al ., 2005 ). Although mechanisms underlying the BBB transport of pathogens are not understood, earlier studies have shown that pathogens consistently release high levels of peptidoglycan such as lipopolysaccharide, LPS (for Gram-negative), or lipoteichoic acid, LTA (Gram-positive), that are highly inflammatory and damage the endothelial TJ (Tunkel et al ., 1991 ; Banks et al ., 1999a, b; Veszelka et al ., 2003 ; Boveri et al ., 2006 ; Mairey et al ., 2006 ), thus progressively increasing the BBB permeability (Maslin et al ., 2005 ). This may allow pathogenic bacteria to cross the BBB and accelerate disease process. This study showed that LPS injection in CW-LPS rats caused a transient but significant efflux of [ 14 C]sucrose (340 Da), but not [ 14 C]dextran (50–100 kDa) across the BBB. Thus, in WC-LPS rats, the LPS-induced opening of the BBB may be too small to allow permeation of LPS or intact pathogens across the barrier. However, since many pathogens have been shown to cross the BBB, mechanisms other than the LPS-mediated BBB abnormalities may be involved.
Chronic ethanol drinking for 70 days did not increase, while LPS injection in ethanol-drinking rats increased permeability of [ 14 C] labelled sucrose and dextran across the BBB. Also, the BBB remained porous for up to 48 h. Thus, in the presence of ethanol, LPS opened the BBB TJ for molecules >100 kDa that may allow LPS and related molecules to cross the barrier leading to predisposition for bacterial infection including streptococci , Listeria monocytogenes , and tuberculosis (Siboni et al ., 1989 ; Aladro Benito et al ., 1995; Burns et al ., 2004 ; Kim et al ., 2005 ; Jain et al ., 2006 ; Schmidt et al ., 2006 ). Alcoholism is also associated with an increased risk of contracting HIV infection (Pfefferbaum et al ., 2005 ). The present observation that chronic alcohol drinking augmented the adverse effects of LPS on the BBB in vivo and earlier observations that an impaired BBB function is associated with increased sensitivity to the brain infection (Ichimi et al ., 1989 ; Lossinsky and Shivers, 2004 ; Sharshar et al ., 2005; van de Beek et al ., 2006 ) provide an aetiological basis for increased susceptibility of alcoholics to pathogen infection.
The TJ of the BBB is maintained by a number of endothelial BBB proteins such as ZO-1, ZO-2, ZO-3, OCL, f-AC, CLA, CIN, JAM, and CAD (Fanning et al ., 1999 ; Brown and Davis, 2002 ; Harhaj and Antonetti, 2004 ; Blasig et al ., 2006 ). ZO-1 interacts with JAM and ZO-3 via its PDZ domain and with CLAs via the guanylate kinase (Haskins et al ., 1998 ; Fanning et al ., 2002 ; Blasig et al ., 2006 ). Molecular scaffold provided by f-actin filaments further stabilizes the endothelial TJ (Madara et al ., 1986 ). A decrease in expression of the BBB proteins and/or an increase in their phosphorylation have been shown to disrupt the protein–protein interaction resulting in disruption of the barrier (Susomboon et al ., 2006 ). Many conditions such as hypoxia or calcium chelating agents damage the barrier's architecture and increase its leaking (Chio et al ., 1992 ; Bazzoni et al ., 2000 ; Brown et al ., 2003 ; Krizbai et al ., 2005 ; Lai et al ., 2005a , 2005b ). This study showed that LPS injection in CW-LPS rats transiently decreased ZO-1, ZO-2, OCL, JAM, and CAD, but not CIN and CLAs mRNA levels, and increased their phosphorylation in the brain endothelial cells. The mRNA levels and protein phosphorylation both returned to the pre-LPS level at 48 h after LPS exposure in CW-LPS. However, LPS injection in Et-LPS rats caused lasting, for up to 48 h, decrease in ZO-1, ZO-2, OCL, JAM, and CAD mRNA levels and increase in their phosphorylation. The protein phosphorylation band intensity in endothelial cells from Et-LPS rats was significantly greater than that in WC-LPS rats. In both WC-LPS and Et-LPS groups, the time-course of LPS-induced increase in the permeability of the BBB correlated with the time-course of decrease in the BBB proteins' mRNA levels and/or an increase in their phosphorylation. An earlier study has shown that a decrease in expression of the BBB proteins suppressed polymerization of f-actin filaments with ZO-1 and OCL that compromised the tightness of the BBB (Madara et al ., 1986 ). Thus, chronic alcohol drinking may cause lasting BBB damage in P-rats.
This study showed that LPS injection in WC-LPS rats, sacrificed at 1 h after LPS injection, activated ERK (p44/p42) and p30 mapk , but not JNK, in the brain endothelial cells. Both ERK and p38 mapk activities returned to the basal level at 48 h after of LPS injection. Unlike LPS injection in WC-LPS rats, LPS injection in Et-LPS rats, sacrificed at 1 h after LPS injection, augmented ERK (p44/p42), p30 mapk , and JNK p54/p46 activation that did not return to the basal level at 48 h after LPS injection. Although mechanisms by which ethanol and/or LPS altered MAPK activation expression and phosphorylation of the BBB proteins are not yet known, earlier studies have shown that LPS bound to the toll-like receptors (TLRs), especially TLR-4 and activated MyD88, JNK p54/p46, ERK p44/p42, and p38 mapk pathways (Lien et al ., 2000 ; Guha and Mackman, 2001 ; Meroni et al ., 2001 ; Takeda and Ichijo, 2002 ; Matsuguchi et al ., 2003 ; Miller et al ., 2005 ; Chen et al ., 2006 ; Gutierrez-Venegas et al ., 2006 ; Kim et al ., 2006 ). An activation of ERK and p38 mapk has been associated with phosphorylation of the BBB proteins resulting in disruption of endothelial tight junction (Gomez et al ., 1999 ; Ward et al ., 2002 ; Wang et al ., 2004 ). Recent studies have shown that HIV-1 Tat protein activated ERK p44/p42 that altered ZO-1 mRNA and protein levels possibly through activation of downstream target transcription factors NFκb, AP-1, or Elk (Pu et al ., 2005 , 2003 ). ERK signalling pathways may also be involved in (i) degradation of ZO-1 through matrix metalloproteinase-9 that has been shown to degrade ZO-1 (Mori et al ., 2002 ) and (ii) rearrangement of the TJ's E-cadhetin into stress fibres (Chen et al ., 2000 ). Taken together, these observations indicate a close association between ERK activation and the BBB disruption: a transient ERK phosphorylation was accompanied with a transient opening of the BBB as in WC-LPS rats, while a lasting ERK phosphorylation was accompanied with lasting opening of the BBB as in Et-LPS rats.
As discussed earlier, activation of the MAP kinases may also activate downstream target transcription factor, NFκB that regulates a number of cellular functions. Our earlier studies have shown that LPS caused a sequential activation of NFκB dimers RelA-p50 (immediately after LPS injection) and p50-p50 (12–24 h after LPS injection) in the brain (Singh and Jiang, 2004b ). Activated RelA-p50 translocated into the nucleus and induced expression of iNOS, TNFα, and pro-inflammatory cytokines, and suppressed apoptosis signalling that initiated inflammation. Activation of the dimer p50-p50 blocked RelA-p50 activation and, possibly, induced expression of pro-apoptosis signalling that suppressed inflammation, induced apoptosis, and restored homeostasis (Singh and Jiang, 2004b ). In endothelial cells from WC-LPS rats, RelA-p50 got activated immediately after and p50-p50 got activated 12 h after LPS injection. Thus, activation of RelA-p50 may be associated with the opening and activation of p50-p50 may also be associated with the closing of the BBB in rats injected with LPS.
Although ethanol drinking did not activate NFκB's RelA-p50, ERK p44/p42, and p38 mapk , it augmented the LPS-induced activation RelA-p50 and suppressed the activation of p50-p50 in endothelial cells, resulting in dysregulation and prolongation of LPS-induced activation of Rel-A-p50. Ethanol exposure also (i) prolonged the activation of ERK and p38 mapk pathways, and (ii) further increased iNOS expression and, possibly, NO release that has been implicated in the TNFα-induced disruption of the endothelial tight junction (Kornelisse et al ., 1996 ; VanFurth et al ., 1996). Based on the present observation, we speculate that NFκB and ERK may be the primary target of alcohol's dysregulation of LPS response. The present study also showed that LPS, either in control or alcohol-exposed cells, did not activate JNK, a member of MAP kinase group that is activated in response to treatment of calls with pro-inflammatory cytokines (Gupta et al ., 1996 ). NFκB has been shown to activate as well as suppress JNK signalling depending on the stimulus (Papa et al ., 2004 , 2006 ). Thus, it is possible that RelA-p50 activation may suppress JNK activation in endothelial cells.
A key observation of this study was that the brain endothelial cells from WC-LPS and Et-LPS rats did not show apoptotic damage that has previously been reported during the resolution phase of inflammatory activity when RelA-p50 is inactive and p50-p50 is activated (Singh and Jiang, 2004b ). Earlier studies have shown that ERK activation is associated with the survival signalling (Hetman and Kharebava, 2006 ; Loucks et al ., 2006 ), while p38 mapk , in association with JNK, is associated with apoptosis signalling (Singh et al ., 2005 ; Wang et al ., 2005 ; Naetzke et al ., 2006) in cells. Also, interleukin-10 (Bouton et al ., 2004 ; Bailey et al ., 2006 ) and CASPASE-8 (Fan et al ., 2005 ) both have been implicated in the development of apoptosis in different cell types. Brain endothelial cells from WC-LPS or Et-LPS rats showed p38 mapk activation but not JNK activation and increased IL-10 but not CASPASE-8 mRNA. Taken together, these observations showed that brain endothelial cells may lack complete apoptosis signals.
A recent in vivo study has shown that although peripheral LPS did not cross the BBB, it activated a number of pro- and anti-inflammatory cytokines in the brain neurons possibly through releasing TNFα and IL-1β into the brain of control rats (Singh and Jiang, 2004a ). This study showed that neurons isolated from the brain of WC-LPS rats exhibited activation of pro-inflammatory/anti-apoptotic pathways immediately after LPS injection followed by an activation of anti-inflammatory/apoptotic pathways 24–48 h after LPS injection. Neurons isolated from the brain of WC-LPS rats sacrificed immediately after LPS injection exhibited ERK, p38 mapk , and JNK phosphorylation, RelA-p50 activation, and iNOS, TNFα, and IL-1β mRNA levels. Neurons isolated from the brain of WC-LPS rats sacrificed 48 h after LPS injection exhibited p50-p50 activation and an increase in IL-10 and CASPASE-8 mRNA levels. Chronic ethanol drinking in Et rats phosphorylated ERK p42 and JNK p46 but did not activate other signalling proteins, thus did not induce apoptosis in neurons. LPS injection in Et-LPS rats (i) caused lasting activation of ERK p44, JNK p54, and RelA-p50 that lasted for up to 48 h, (ii) suppressed activation of p50-p50 and mRNA levels of IL-10 and CASPASE-8, and (iii) suppressed apoptosis. Taken together, these observations suggest that an activation of JNK and an increase in CASPASE-8 expression may be associated with the induction of apoptosis possibly by suppressing the anti-apoptotic effects of TNFα in neurons. This study also showed that ethanol exposure suppressed the LPS-induced expression of CASPASE-8 well as the LPS-induced apoptosis in Et-LPS rats. Since LPS permeated across the BBB in and LPS accumulated in brain neurons isolated from Et-LPS rats but not from WC-LPS rats, we propose that the neurons isolated from the brain of Et-LPS rats exhibited direct effects of LPS, while neurons from the brain of WC-LPS rats exhibited indirect effects of LPS. Similar to the present observation in neurons, earlier studies have also shown that ethanol feeding suppressed apoptotic pathways in pancreas possibly by inhibiting caspase expression (Wang et al ., 2006 ). However, contrary to the present observations, ethanol has been shown to induce apoptosis in cultured faetal cortical neurons (Ramachandran et al ., 2001 , 2003 ) and liver hepatocytes (Deaciuc et al ., 2001 ; Adachi et al ., 2004 ). These discrepancies may be due to the signalling differences in different tissues.
Conclusions and a unified hypothesis
Chronic alcohol drinking adversely affects the immune system and progression of infectious diseases including HIV-1 and tuberculosis infection (Herper, 1998; deOliveira et al ., 2002 ; Rothlind et al ., 2005 ). Infection is more frequent in alcoholics possibly due to breakdown of protective barriers, brain damage, alteration of host immune defense, and malnutrition (Cook, 1998 ; Grove, 1996 ). This study demonstrated that chronic alcohol drinking, although did not significantly alter the BBB integrity, dysregulated the LPS-induced opening of the BBB possibly by modulating the expression and phosphorylation of the BBB proteins. Alcohol drinking also prolonged the pro-inflammatory but suppressed the apoptotic effects of LPS injection. Thus, chronic alcohol drinking may cause higher predisposition to infection and inflammation-related diseases possibly by augmenting the effects of pathogens on the BBB. As shown in Fig. 14 , LPS (i) binds to TLR and activates p38 mapk and ERK that induce iNOS expression and NO release that phosphorylates the BBB proteins and increases its permeability, and (ii) activates NFκB that induces inflammation and suppresses apoptosis. RelA-p50 and/or NO induces TNFα and IL-1β that reaches the neurons and, through binding to their receptors, activates p38 mapk , ERK, and JNK, induces expression of TNFα and IL-1β followed by expression of IL-10 and CASPASE-8 that may activate p50-p50 resulting in induction of apoptosis. Chronic ethanol drinking suppresses the JNK pathway in neurons resulting in suppression of apoptosis.
This project is funded by a NIH sub contract, grants from the Graduate School and a food safety grant from the College of Veterinary Medicine, University of Minnesota.