High Mobility Group Box Protein 1 Boosts Endothelial Albumin Transcytosis through the RAGE/Src/Caveolin-1 Pathway

High-mobility group box protein 1 (HMGB1), an inflammatory mediator, has been reported to destroy cell-cell junctions, resulting in vascular endothelial hyperpermeability. Here, we report that HMGB1 increases the endothelial transcytosis of albumin. In mouse lung vascular endothelial cells (MLVECs), HMGB1 at a concentration of 500 ng/ml or less did not harm cell-cell junctions but rapidly induced endothelial hyperpermeability to 125I-albumin. HMGB1 induced an increase in 125I-albumin and AlexaFluor 488-labeled albumin internalization in endocytosis assays. Depletion of receptor for advanced glycation end products (RAGE), but not TLR2 or TLR4, suppressed HMGB1-induced albumin transcytosis and endocytosis. Genetic and pharmacological destruction of lipid rafts significantly inhibited HMGB1-induced albumin endocytosis and transcytosis. HMGB1 induced the rapid phosphorylation of caveolin (Cav)-1 and Src. Either RAGE gene silencing or soluble RAGE suppressed Cav-1 Tyr14 phosphorylation and Src Tyr418 phosphorylation. The Src inhibitor 4-amino-5-(4-chlorophenyl)-7-(t-butyl) pyrazolo[3,4-d] pyrimidine (PP2) blocked HMGB1-induced Cav-1 Tyr14 phosphorylation. PP2 and overexpression of Cav-1 with a T14F mutation significantly inhibited HMGB1-induced transcytosis and albumin endocytosis. Our findings suggest that HMGB1 induces the transcytosis of albumin via RAGE-dependent Src phosphorylation and Cav-1 phosphorylation. These studies revealed a new mechanism of HMGB1-induced endothelial hyperpermeability.

A severe systemic inflammatory response can cause endothelial injury and high capillary permeability, in turn resulting in severe capillary leakage during the initial stages of extensive burns, severe acute pancreatitis, and severe sepsis 11,12 . High capillary permeability is associated with water, electrolyte, and albumin leakage into the tissue interspace. Albumin leakage occurs through two pathways, i.e., the paracellular pathway and transcytosis 13 . Albumin leakage via the paracellular pathway is a result of damage to cell-cell junctions 13 , whereas albumin transcytosis results from endocytosis after binding with the receptor gp60 and subsequent basolateral excretion 14 . Albumin endocytosis occurs via the caveolin (Cav)-1-dependent endocytic pathway 14 . In the absence of Cav-1, vascular endothelial cells can transfer albumin through the clathrin-dependent endocytic pathway 15 .
It has been reported that high concentrations of HMGB1 (≥ 5 μ g/mL), resulting in vascular barrier damage, quickly destroy the junctions between pulmonary vascular endothelial cells 16 , whereas lower concentrations of HMGB1 (200 ng/mL) for less than 12 h caused no destruction of the cell-cell junctions 16,17 . Several studies showed that although the HMGB1 level was elevated in patients with severe acute pancreatitis or sepsis, it remained far less than 200 ng/mL [18][19][20] . It was observed that exposure to lower concentrations (< 200 ng/ml) of HMGB1 for a shorter time period could not affect the integrity of endothelial cell-cell junctions. However, the effect of HMGB1 on albumin transcytosis remains unclear.
In vascular endothelial cells, Cav-1 phosphorylation by Src (the signalling protein associated with Cav-1) can lead to the increased transcellular transport of albumin 21 . In addition, after binding with its ligand, RAGE activates Src and downstream signals 22 . We speculated that HMGB1 induces endothelial albumin transcytosis through Src activation. In the present study, the results showed that a lower concentration (100 ng/ml) of HMGB1 did not damage the integrity of the junctions between mouse lung vascular endothelial cells (MLVECs) but boosted albumin transcytosis via RAGE-dependent Src and Cav-1 phosphorylation, resulting in vascular endothelial hyperpermeability to albumin.

Results
HMGB1 induced pulmonary vascular endothelial hyperpermeability to albumin. We examined the effects of HMGB1 on the permeability of the pulmonary vascular endothelium using a Transwell chamber assay. After a 1-h treatment of MLVEC monolayers with 0, 50, 100, and 500 ng/mL HMGB1 and 125 I-albumin, the radioactivity of the liquid in the lower chamber was tested. We found that the permeability of MLVEC monolayers to 125 I-albumin increased with increasing concentrations of HMGB1 (Fig. 1A). To verify the role of HMGB1 at the organ level, we performed the experiment with mouse lungs that were treated via pulmonary vascular perfusion with HMGB1 (100 ng/mL) and 125 I-albumin for 30 min. We found that the lung W/D weight ratio (Fig. 1B) and the extravascular 125 I-albumin level (Fig. 1C) were significantly higher in the HMGB1 group than in the control group. These results suggested that lower concentrations of HMGB1 resulted in the hyperpermeability of the pulmonary vascular endothelium to albumin as well as pulmonary oedema.
Lower concentrations of HMGB1 did not damage the endothelial barrier. Albumin leakage can occur through the paracellular pathway or through transcytosis 13 . Studies have shown that HMGB1 stimulation destroys the integrity of endothelial cell junctions either at high concentrations or when exposed to cells for a longer period of time 16,17 .
The integrity of the cell-cell junctions was measured to further clarify the mechanism through which lower concentrations of HMGB1 affect the albumin hyperpermeability of the vascular endothelium. MLVEC monolayers were treated using 0, 50, 100 and 500 ng/mL HMGB1, with thrombin and 15 μ g/mL HMGB1 as positive controls. The resulting changes in TER were evaluated, and the results showed that the TER in MLVEC monolayers rapidly decreased after thrombin administration but remained unchanged after HMGB1 stimulation ( Fig. 2A).
Indirect immunofluorescence staining of VE-Cadherin was used to confirm the changes in cell-cell junctions, and 15 μ g/mL HMGB1 and Thrombin was used as a positive control. Addition of 0-500 ng/mL HMGB1 caused no increase in the intercellular gaps between MLVECs (Fig. 2B). Collectively, these results indicated that 0-500 ng/mL HMGB1 caused no rapid destruction of the cell-to-cell junctions of MLVECs. HMGB1 increased the endothelial transcytosis of albumin. Albumin endocytosis was assayed to evaluate whether HMGB1 results in increased transcytosis of albumin by vascular endothelial cells. Endothelial cells were stimulated with HMGB1 (0, 50, 100, and 500 ng/mL) and 125 I-albumin. 125 I-albumin endocytosis by MLVEC monolayers increased with increasing concentrations of HMGB1 (Fig. 3A). AlexaFluor 488-labeled albumin was also used in the endocytosis assay. Under a confocal microscope, green fluorescence in MLVECs increased with the increase of HMGB1 concentration (Fig. 3B). Combined with the results of the Transwell-based 125 I-albumin permeability assay (Fig. 1A), these data suggested that increasing concentrations of HMGB1 caused albumin hyperpermeability by inducing the transcytosis of albumin in vascular endothelial cells.

HMGB1-induced endothelial transcytosis of albumin was dependent on RAGE but not on TLR2/4.
At the surface of vascular endothelial cells, the three main receptors of HMGB1 are RAGE, TLR2, and TLR4 5 . After silencing of these three genes (Fig. 4A), we observed that only RAGE depletion blocked HMGB1-induced endocytosis and albumin transcytosis (Fig. 4B,C). To verify the role of RAGE on HMGB1-induced albumin transcytosis at the organ level, we evaluated the lungs of RAGE −/− and RAGE +/+ mice after pulmonary vascular perfusion with HMGB1 and 125 I-albumin for 30 min. We found that the extravascular 125 I-albumin level (Fig. 4D) and the W/D weight ratio (Fig. 4E) was lower in RAGE −/− lungs. Thus, it can be concluded that HMGB1-induced albumin transcytosis was dependent on RAGE but not on TLR2/4. Scientific RepoRts | 6:32180 | DOI: 10.1038/srep32180 HMGB1 induced Cav-1-dependent endothelial transcytosis of albumin. Endocytosis and transcytosis of albumin in vascular endothelial cells are thought to be Cav-1-dependent 14 , although studies have shown that in the absence of Cav-1, transport of albumin in the vascular endothelial cells takes place via the clathrin-dependent endocytic pathway 15 . To further clarify whether the albumin endocytic pathway affected by HMGB1 is dependent on Cav-1 or clathrin, we used methyl-β -cyclodextrin (Mβ CD) and clathrin depletion to block the Cav-1-dependent and clathrin-dependent endocytic pathways, respectively 23 . It was found that Mβ CD, rather than clathrin siRNA, blocked HMGB1-induced endocytosis and albumin transcytosis ( Fig. 5A-D).
As described above, HMGB1 mediated the albumin hyperpermeability of endothelial cells via its receptor RAGE (Fig. 4A-E). We thus studied the positional relationship of RAGE and Cav-1 in the cell membrane. The results of co-immunoprecipitation and immunofluorescence staining showed that Cav-1 bound with RAGE on the membrane of MLVECs, and HMGB1 stimulation seemed not to alter their colocalization (Fig. 5E,F). Furthermore, we silenced the expression of Cav-1 in MLVECs and found that HMGB1-induced endocytosis and transcytosis of albumin was blocked (Fig. 5G,H). These data showed that HMGB1 induced the Cav-1-dependent endothelial transcytosis of albumin.

HMGB1 induced the endothelial transcytosis of albumin via Cav-1 phosphorylation. Induction
of Cav-1 expression 24 or phosphorylation 21 can increase the transendothelial transport of albumin. A western blot assay showed that HMGB1 did not affect the Cav-1 expression level in MLVECs (Fig. 6A).
Cav-1 tyrosine phosphorylation (Y14) initiates the endocytosis and transcytosis of albumin 14 . The next experiment was based on the effects of HMGB1 on Cav-1 phosphorylation. HMGB1 was found to induce Cav-1 Y14 phosphorylation in MLVECs in a dose-dependent manner (Fig. 6B,C). Furthermore, overexpression of a Cav-1 Y14 phosphorylation-defective mutant (Cav-1 Y14F) in MLVECs (Fig. 6D) resulted in significantly reduced endocytosis and transcytosis of albumin compared to the wild-type group (Fig. 6E,F). These results indicated that Y14 phosphorylation of Cav-1 mediated the HMGB1-induced endothelial transcytosis of albumin.
Src phosphorylation mediated the HMGB1-induced Cav-1 phosphorylation and endothelial transcytosis of albumin. Cav-1 Y14 phosphorylation is catalysed by both Src, the major catalysing enzyme 14 and c-Abl 21,25 . We measured the phosphorylation-mediated activation of Src in MLVECs after HMGB1 induced albumin hyperpermeability of MLVEC monolayers. 125 I-albumin and different concentrations of HMGB1 (0, 50, 100, 500 ng/mL) were added into the upper chamber of a Transwell chamber containing MLVEC monolayers. After 1 h of incubation, 200 μ L of liquid sample was removed from the lower chamber for radioactivity detection (n = 4 to 6 for each group). (B,C) HMGB1 induced mouse pulmonary vascular hyperpermeability to albumin (B) and pulmonary oedema (C). The mouse lung samples were irrigated with a solution containing 0 or 100 ng/mL HMGB1 and 125 I-albumin via pulmonary artery perfusion. After 30 min, the lungs were perfused with Krebs solution, weighed and used for radioactivity detection (n = 6/each group). After oven-drying at 60 °C for 72 h, the final weight was measured to calculate the wet/dry (W/D) weight ratio. *Compared with the control group, p < 0.05. stimulation and found that Src was activated rapidly via phosphorylation in a dose-dependent manner by exposure to HMGB1 (Fig. 7A,B). However, the use of an inhibitor of Src activation, 4-amino-5-(4-chlorophenyl)-7-(t-butyl) pyrazolo [3,4-d] pyrimidine (PP2), blocked HMGB1-induced Cav-1 phosphorylation (Fig. 7D) and the transcytosis of albumin (Fig. 7E,F). In addition, after exposure to HMGB1, the activation of c-Abl via phosphorylation was not observed in MLVECs (Fig. 7C). Hence, HMGB1-induced Cav-1 phosphorylation and albumin transcytosis were dependent on Src phosphorylation. RAGE mediated HMGB1-induced Src phosphorylation and Cav-1 phosphorylation. We next sought to further confirm the role of RAGE in HMGB1-induced Src phosphorylation and Cav-1 phosphorylation was studied. The results were consistent with the previous findings and confirmed that RAGE gene silencing ( Fig. 8A,C) or sRAGE (Fig. 8B,D) blocked HMGB1-induced Src Y418 phosphorylation and Cav-1 Y14 phosphorylation.

All-thiol HMGB1, but not disulfide HMGB1, induced the endothelial transcytosis of albumin.
The redox state of HMGB1 modulates its extracellular functions 26 . All-thiol HMGB1 uses RAGE signalling, while disulfide HMGB1 recognizes TLR4. We found that all-thiol HMGB1, but not disulfide HMGB1, induced the endocytosis and transcytosis of albumin (Fig. 9A,B).

Discussion
Our study is the first report on vascular endothelial hyperpermeability induced by pathophysiological concentrations of HMGB1. The results showed that 100 ng/mL of HMGB1 increased albumin transcytosis. Huang et al. 17 found that there were no effects of HMGB1 (100 ng/mL) on the permeability of HUVEC monolayers because they did not study the effects of HMGB1 on the transcellular pathway. Similarly, Wolfson et al. 16 only focused on the effects of HMGB1 on cell-cell junctions. In our study, TER and VE-Cadherin staining results showed that 100 ng/mL HMGB1 did not destroy the cell-cell junctions of MLVECs but still induced 125 I-albumin hyperpermeability. As was confirmed in the endocytosis assays of 125 I-and AlexaFluor 488-labeled albumins, this endothelial hyperpermeability was achieved by increasing the endocytosis and transcellular transport of albumin.
The three receptors of HMGB1 are RAGE, TLR4, and TLR2, among which RAGE seems to be altered when HMGB1 performs different functions in various cells. Animal experiments have shown that knockout of RAGE had a protective effect on mice with sepsis induced by caecal ligation puncture, and this protective effect was eliminated by exogenous RAGE expression in the endothelial cells and bone marrow cells of RAGE knockout mice 22 . This result suggested that RAGE is involved in the progression of sepsis in mice. RAGE knockout is more effective in reducing the HMGB1-induced release of pro-inflammatory cytokines than TLR2 knockout in bone marrow-derived macrophages, suggesting that RAGE is the main receptor of HMGB1 recognized by  22 . RAGE seems to be more important for the function of HMGB1 in vascular endothelial cells. Activation of HUVECs by HMGB1 can be inhibited by more than 50% with RAGE-neutralizing antibodies 9 . Huang et al. 17 and Wolfson et al. 16 showed that the damage caused by HMGB1 to cell-cell junctions was dependent on RAGE but was independent of TLR4 and TLR2. The present study showed that the HMGB1-induced transcytosis of albumin was dependent on RAGE but not TLR4 and TLR2. These data are consistent with reports that RAGE KO protected mice from acute lung injury 27 .
Albumin can be endocytosed through the clathrin-dependent and Cav-1-dependent pathways 15 . After Mβ CD or siRNA were used to destruct lipid rafts and to silence Cav-1 expression, respectively, we found that HMGB1-induced albumin endocytosis and transcytosis were significantly inhibited; however, clathrin knockdown did not block HMGB1-induced albumin endocytosis or transcytosis. These results indicated that HMGB1 increased vascular endothelial permeability to albumin by upregulating the Cav-1-dependent transcytosis of albumin.
Upregulated Cav-1 expression allows increased endocytosis and transcytosis of albumin 24 . Our data showed that HMGB1 did not increase the expression level, but raised the phosphorylation level, of Cav-1 in endothelial cells. Phosphorylation of Cav-1 Tyr14 is considered to be a key step to initiate the endocytosis of albumin. A study has shown that hydrogen peroxide increased the transcytosis of albumin by inducing Cav-1 Tyr14 phosphorylation 21 .
In the present study, HMGB1-induced Cav-1 Tyr14 phosphorylation was mediated by RAGE, as either RAGE gene silencing or sRAGE blocked Cav-1 Tyr14 phosphorylation. Overexpression of Cav-1 with a Tyr14 phosphorylation-defective mutant significantly inhibited the HMGB1-induced transcytosis and endocytosis of albumin, strongly suggesting that HMGB1 increases endothelial permeability by regulating Cav-1 Tyr14 phosphorylation. It has been reported that Cav-1 Tyr14 phosphorylation is regulated by Src and c-Abl 21,25 . We found that HMGB1 induced Src phosphorylation, but c-Abl was not activated by phosphorylation.
Redox modulates the extracellular functions of HMGB1 26 . All-thiol HMGB1 used RAGE signalling, while disulfide HMGB1 recognized TLR4. We found that all-thiol HMGB1, but not disulfide HMGB1, induced the endocytosis and transcytosis of albumin. This is consistent with our findings that RAGE mediates HMGB1-induced albumin transcytosis.
HMGB1 requires additional factors to exert its inflammatory activity, including LPS, IL-1, and nucleosomes 28 . In a real pathophysiological environment, HMGB1 co-exists with these inflammatory mediators. In the presence of these mediators, the effect of HMGB1 on vascular permeability may have additional features that need to be studied.

Materials. Recombinant HMGB1 was purchased from R&D Systems (Minneapolis, MN). All-thiol HMGB1
and disulfide HMGB1 were obtained from HMGBiotech. Protein A+ G Agarose beads, HRP-conjugated secondary Abs, and anti-GAPDH, TLR4 and TLR2 Abs were from Santa Cruz Biotechnology. Anti-Cav-1, p(Y14)Cav-1, p(Y418)Src, RAGE, c-Abl, and phospho-tyrosine Abs, normal rabbit IgG and normal mouse IgG were from Cell Signalling Technology. TLR2, TLR4, RAGE, clathrin and Cav-1 siRNA, sc siRNA and Dharmafector 1 transfection reagent were obtained from Thermo Scientific. AlexaFluor 488-conjugated albumin was from Molecular Probes. TLR2-and TLR4-neutralizing Abs were from Abcam. Recombinant soluble RAGE (sRAGE) was from Biotrend. 125 I-albumin was purchased from PerkinElmer, Inc. Amaxa basic nucleofector kits for primary endothelial cells were obtained from Lonza. RIPA buffer and ECL reagents were from Pierce Biotechnology. Bicinchoninic acid kits and sample buffer were from Bio-Rad. AlexaFluor 568-and AlexaFluor 488-conjugated secondary Abs were from Invitrogen. MEM D-Val medium and FBS were from Gibco. Biotin-conjugated rat anti-mouse CD31 (PECAM-1) monoclonal antibody and BD IMag Streptavidin Particles Plus-DM were from BD Pharmingen. All other reagents were obtained from Sigma unless otherwise specified.   23 . Mice were anesthetized, and then 50 units of heparin were injected via the jugular vein. An incision was made into the chest to insert a pulmonary artery catheter, and blood in the vasa publica was removed by perfusion with PBS. The edge of the lung tissues was excised and cut into 1-mm 3 small pieces, which were placed in 60-mm petri dishes containing culture medium (MEM D-Val medium containing 2 mM glutamine, 10% FBS, 5% human serum, 50 μ g/ml penicillin/streptomycin, 5 μ g/ml heparin, 1 μ g/ml hydrocortisone, 80 μ g/ml endothelial cell growth supplement from bovine brain, 5 μ g/ml amphotericin, and 5 μ g/ml mycoplasma removal agent). After 60 h of incubation at 37 °C in 5% CO 2 , tissue blocks were removed, and adherent cells were continuously cultured for another 3 d. Endothelial cells were then purified using biotin-conjugated rat anti-mouse CD31 (PECAM-1) monoclonal antibody, BD IMag Streptavidin Particles Plus-DM, and an immunomagnetic separation system (BD Biosciences Pharmingen, San Diego, CA). The purified 3D cell culture was identified by Immunoblotting and immunoprecipitation. Cells were lysed in RIPA buffer supplemented with 1 mMol PMSF, 1 mMol Na 4 VO 3 and protease inhibitor cocktail. The lysates were sonicated and centrifuged at 10,000 g for 10 min at 4 °C. The protein concentrations were measured with bicinchoninic acid kits. The samples were loaded equally for PAGE and then transferred onto nitrocellulose membranes. After blocking with 5% non-fat milk in TBST, the membranes were probed with primary antibodies for 2 h at room temperature or overnight at 4 °C and then incubated with HRP-conjugated secondary antibodies at room temperature for 1 h. The bands were determined using ECL reagent and quantified using ImageJ software (NIH).
For immunoprecipitation, the cell lysates were pre-cleared with 1 μ g of normal IgG and 20 μ l Protein A+ G Agarose beads for 2 h at 4 °C. After centrifugation at 1000 g for 5 min, the supernatants were transferred to new tubes and incubated with 40 μ l of Protein A+ G Agarose beads and anti-Cav-1 or anti-RAGE antibodies overnight at 4 °C. The beads were collected for immunoblotting after 3 washes with PBS.
Immunofluorescence. Cells plated on coverslips were fixed with 2% paraformaldehyde for 15 min and then washed 3 times with 100 mM glycine in HBSS for 10 min and once with HBSS for 10 min. After permeabilization with 0.1% Triton X-100 in HBSS for 30 min, the cells were incubated with primary antibodies (diluted with HBSS containing 5% horse serum and 0.2% BSA at a 1:200 dilution) overnight at 4 °C. Three washes with HBSS were followed by incubation with fluorescence-conjugated secondary Abs (1:200) for 1 h. After another 3 washes with   pre-equilibrated culture medium and incubated for 24-96 h at 37 °C. The viability of the cells was measured by vital dye exclusion. Successful transfection was confirmed by immunoblotting.
Endocytosis assay with 125 I-albumin. Endocytosis of 125 I-albumin was assayed based on protocols from the literature 14,30 . Six-well plates were used to culture MLVECs until the cells formed confluent monolayers. HBSS containing unlabelled albumin was added at final concentrations of 0.1 mg/mL (in four wells) and 100 mg/mL (in two wells). Then, 125 I-albumin (1 × 10 6 cpm) was added to each well, and the plates were incubated for 30 min at 37 °C. The cells were washed successively with ice-cold acetate buffer (0.5 M NaCl and 0.2 M acetate, pH 2.5) and HBSS thrice to remove cell surface-bound 125 I-albumin. The cells were lysed with 1 mL of Tris-HCl buffer (0.05 M Tris-HCl, 1% Triton X-100 and 0.5% SDS, pH 7.4), and the radioactivity of the cell lysates was detected using a γ counter (PerkinElmer, Inc.). The intake of specific 125 I-albumin was calculated as total radioactivity (detection value of a sample containing 0.1 mg/mL unlabelled albumin) minus non-specific cell-related activity (detection value of a sample containing 100 mg/mL unlabelled albumin) and then corrected by the total cellular protein and expressed as cpm/mg cellular protein.
Endocytosis assay of fluorescently labelled albumin. MLVECs were grown on cover slips until the cells formed confluent monolayers. AlexaFluor 488-labeled BSA (50 μ g/mL) and unlabelled albumin (500 μ g/mL) were added to the cell culture, and the culture was then incubated at 37 °C for 30 min. The cells were washed thrice with HBSS and examined under a confocal microscope (Zeiss LSM 510 Meta) to quantify the amount of fluorescence-labelled albumin endocytosed 14,30 . Transendothelial permeability assay with 125 I-albumin. MLVECs were grown on fibronectin-coated microporous polyester Transwell membranes (12 wells, 1 cm 2 growth area, 0.4 μ m pore size; Corning Costar, Cambridge, MA) until the cells formed confluent monolayers. The upper chamber was supplemented with 0.5 mL of HBSS containing 125 I-albumin (1 × 10 6 cpm) and unlabelled albumin (0.1 or 100 mg/mL). The lower chamber was supplemented with 1.5 mL of HBSS containing the same amount of unlabelled albumin used for the upper chamber. The reaction time was 1 h, after which 200 μ L of liquid sample was removed from the lower chamber and used for a radioactivity assay with a γ counter (PerkinElmer Inc.). The transendothelial permeability to 125 I-albumin was calculated and expressed as μ L/min.cm 2 in accordance with the literature 14,30 . Transepithelial electrical resistance (TER) detection. TER detection was performed in accordance to the previous literature 14,31 . MLVECs were grown in culture plates containing gold-plated microelectrodes until the cells formed confluent monolayers. The positive and negative poles were connected to a synchronous phase sensitive amplifier. Computerized recording of the data related to voltage changes in HMGB1-treated cells was performed, and the data were expressed as the relative value to the initial reading of zero.
Detection of permeability-surface area product (PS). Mouse lungs were prepared as described previously 32 . In brief, mice were anesthetized, intubated, and ventilated. For pulmonary artery catheterization, the chests of the mice were surgically dissected, followed by perfusion with a modified Krebs-Henseleit solution at flow rate of 2 mL/min, a venous pressure of 3 cm H 2 O, and a pulmonary artery pressure of 8 ± 2 cm H 2 O.
Albumin PS was determined as previously described 14 with slight modification. In brief, lung artery lavage was performed with a fluid containing 80000 counts/mL 125 I-albumin and 0 or 100 ng/mL HMGB1. After 30 min, the lungs were irrigated with Krebs solution (containing 5% unlabelled albumin) for 6 min to remove excess 125 I-albumin on the cell surface and within the cycling. Then, the attached tissues were quickly removed from the lung. The sample was weighed and used for γ radioactivity measurement. The PS value was calculated with the formula A/(Cp · t), where A and Cp are concentrations of tracer albumin in the tissue (in counts/g) and in the perfusate (in counts/mL), respectively, and t is the perfusion time for tracer albumin (30 min). The PS product was expressed as μ L/(min·g) dry lung.
Determination of lung tissue wet-dry weight ratio (W/D). The wet lungs were weighed and then oven-dried at 60 °C for 72 h. The final weight was measured to calculate the W/D weight ratio. Statistical Analysis. The data are expressed as the mean ± SEM. One-way ANOVA with post-tests for individual treatments and Student's Newman-Keuls test for post hoc comparisons were used to determine differences. Differences were considered significant when p < 0.05.