INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Albumin is an unglycosylated protein of 66.3 kD (1, 2). As the most abundant protein in plasma and extracellular compartments, albumin regulates vascular oncotic pressure, binds several drugs, and transports various endogenous molecules (3). However the extent of its physiological role, particularly in inflamed tissues, is still not fully understood. One of the cardinal features of inflammation is an increase in vascular permeability allowing large amounts of albumin to reach the interstitium and epithelial surfaces of various organs including the lung (4, 5). The presence of high concentrations of albumin in the interstitium during inflammation affects the transvascular oncotic gradient thus contributing to the formation of interstitial edema (6). An acute increase in tissue albumin content is also likely to have beneficial effects, such as providing antioxidant protection to tissues (7).

One of the striking features of human serum albumin is the presence of 34 cystine residues forming 17 disulfide bonds, and one free thiol at the Cys-34 position (1). Under normal circumstances, one-third of the serum albumin molecules form mixed disulfides with either glutathione (GSH) or half-cystine. However the remaining sulfhydryl groups at the Cys-34 residue of albumin constitute the major extracellular source of reactive free thiol (8). In this context, it has been suggested that albumin constitutes an important extracellular antioxidant in plasma (9, 10).

Although albumin is an extracellular molecule, it is taken up by most cell types either through receptor-mediated endocytosis or pinocytosis, and it is catabolized at a very high rate through lysosomal degradation (1). Such a metabolic pathway has specifically been demonstrated in lung epithelial cells (11), and likely occurs in most cell types (12). Cellular uptake and catabolism of albumin could therefore potentially represent a source of sulfur-containing amino acids for cells in the synthesis of thiol-containing molecules, particularly GSH. In addition, each albumin molecule contains six glutamylcysteine residues. The -glutamylcysteine dipeptide is particularly important in GSH synthesis since its formation through the enzyme -glutamylcysteine synthetase represents the rate-limiting step in GSH synthesis (13). Finally, albumin has been shown to play an essential role in preventing apoptosis in endothelial cells (14). Since an increase in cellular GSH is one of the most important mechanisms by which apoptosis can be prevented, it is possible that albumin may prevent apoptosis by regulating cellular GSH homeostasis.

In the context of the above we hypothesized that serum albumin may play a key role in preserving cellular GSH, thus protecting cells against oxidant-mediated injury and possibly down-regulating the activation of the oxidant-sensitive transcription protein complex nuclear factor-kappa B (NF-B).

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Cell Culture

All cells studied were of human origin. The human pulmonary epithelial cell line A549 (ATCC CCL 185) and the human lung fibroblast cell line HFL1 (ATCC CCL 153) were obtained from the American Type Culture Collection (Rockville, MD) and cultured as previously described (15, 16). Peripheral blood lymphocytes were isolated from healthy donors using Ficoll-Hypaque (Winthrop Laboratories, Aurora, ON, Canada) gradient centrifugation followed by adherence of monocytes to plastic.

Sources of Human Serum Albumin

Human serum albumin (HSA) (cat. no.: A1653) was purchased from Sigma Chemical Co. (St. Louis, MO) and tested for endotoxin contamination using the E-Toxate Limulus Amebocyte Lysate assay kit (Sigma Chemical Co.). Endotoxin levels were consistantly less than the lower limit of detection of 0.03 EU/ml. Since contaminating proteins may be present even in the highly purified commercial preparations of albumin derived from blood, purified recombinant human serum albumin (rHSA) prepared in the yeast Pichia pastoris was obtained from New Century Pharmaceuticals (Huntsville, AL), and compared with HSA obtained from Sigma for its GSH-enhancing properties.

Effect of Serum and HSA on Total Cellular Glutathione (GSH + GSSG)

Human serum was isolated from a healthy donor and either added directly to cells or, to remove albumin, 2 ml was passed over a 1.5 × 6-cm column of Cibacron Blue 3G (Blue Sepharose CL-6B; Pharmacia Canada Inc., Baie d'Urfe, PC, Canada) equilibrated in 0.05 M Tris, 0.05 M NaCl, pH 8.0. The column was washed with the same buffer, and recovered proteins (referred to as the albumin-depleted fraction) were pooled. Proteins bound to the gel (albumin-enriched fraction) were eluted from the gel in 0.05 M Tris, 0.5 M MgCl₂, pH 8.0. Both fractions were concentrated to a volume of 2 ml in RPMI medium using Centriprep-10 membranes (Amicon Inc., Beverly, MA), and each fraction was assayed for albumin using radial immunodiffusion LC-Partigen plates (Behring Diagnostics Inc., Westwood, MA). The untreated serum and the albumin-enriched fraction were each adjusted in RPMI to a final albumin concentration of 2%. The albumin-depleted fraction, in which albumin was undetectable (< 25 µg/ml), was adjusted to the same volume in RPMI as the albumin-enriched fraction. Each fraction (20 µg protein) was further analyzed by sodium dodecyl sulfate (SDS) 7.5% polyacrylamide gel electrophoresis stained with Coomasie blue (17). To determine the effect of serum proteins and HSA on cellular GSH, A549 cells were plated at 1.5 × 10⁵ cells/well in 24-well culture plates in complete medium for 24 h. The medium was then removed and replaced with fresh medium containing either serum, Cibacron Blue-treated serum (albumin-depleted fraction), or serum proteins eluted from Cibacron Blue sepharose (albumin-enriched fraction). The cells were incubated for 18 h before being lysed in 1% Triton X-100 for glutathione measurements as previously described (18).

To determine the effects of purified human serum albumin on lung cellular glutathione, human serum albumin purified from blood or produced by recombinant technology in Pichia pastoris was added to either A549 or HFL1 cells at the specified concentrations (0-2%) for different times (4-24 h). The cells were incubated for 18 h before being lysed for glutathione measurements.

Since the glutathione levels of human peripheral lymphocytes are much lower than in HFL1 or A549 cells, similar experiments were performed using 2 × 10⁶ cells/well and the cells were incubated 48 h in the presence of 0-2% HSA. At the end of each incubation period, cell numbers were counted by hemacytometer, and GSH + GSSG was determined in the cell lysates. The abbreviation GSH is used throughout the manuscript since the cellular levels of GSSG accounted for < 5% of total glutathione.

Specificity of HSA-mediated GSH Regulation

To determine whether HSA-mediated increase in GSH was a nonspecific effect of protein in the extracellular milieu, we purified the second most abundant serum protein, alpha₁-proteinase inhibitor (₁-PI) as previously described and determined its effect on cellular GSH as described above. Furthermore, since albumin is rich in cystine residues and disulfide bonds we replaced HSA by either thiols (1 mM cysteine, 1 mM GSH) or disulfide molecules (0.5 mM cystine, 0.5 mM GSSG). After 18 h incubation, cell number and glutathione concentrations were determined in the cell lysates.

In Vivo Effects of Albumin on Lung GSH

To determine whether the GSH-enhancing effects of HSA could be observed in vivo, C57BL/6 mice (n = 17 per group) were anesthetized with inhaled halothane and instilled intranasally with either 50 µl saline solution (control) or 50 µl saline solution containing 7.5 mg HSA, once daily for 3 d. On Day 4, the animals were sacrificed, and the lungs were cleared of blood by infusing 10 ml saline solution through the canulated right ventricule. Lung tissues were homogenized in a polytron homogenizer in phosphate-buffered saline (PBS), centrifuged at 15,000 × g for 15 min, and GSH measured in the supernatant.

Effect of Albumin Reduction on GSH Regulation

The effect of reduction on the GSH-enhancing properties of HSA was studied by incubating a 4% HSA solution with various amounts of dithiothreitol (DTT) in a range of molar ratios of DTT:HSA from 0:1 to 16:1 for 5 h in PBS at pH 6.9. To prepare HSA in which all free cysteine groups were blocked, 4% HSA was first incubated with DTT for 5 h to reduce the thiol residue, Cys-34, on all HSA molecules and subsequently incubated with N-ethylmaleimide in a molar ratio of 5:1 for 5 h. At the end of the incubation period, each sample was divided in two, one for incubation with cells and the other to determine albumin and thiol contents. For cell culture samples, a 2.5-ml aliquot was passed over a PD₁₀ column (Sepharose G-25 M; Pharmacia Biotech AB, Uppsala, Sweden) equilibrated in RPMI, and for HSA and thiol determinations, a 300-µl sample was diluted to 2.7 ml in PBS and passed over a PD₁₀ column to remove residual DTT. The HSA concentration of the eluate was determined in a spectrophotometer at 280 nm, and the number of thiol groups per albumin molecule was determined by adding 50 µl of each solution to 550 µl of a 140 µM dithiobis(2-nitrobenzoic acid) solution, assuming a molar extinction coefficient of 13,600 (19). Once the HSA and thiol concentrations had been determined, the final HSA concentration was adjusted to 2% in RPMI before being added to the cells for 18 h followed by GSH determinations as described above.

Role of Lysosomal Acidification on GSH

To determine the role of endosome/lysosome acidification in the HSA-mediated GSH increase, A549 cells were incubated in the presence of 2% HSA and either 1 µM bafilomycin A₁, 1 µM concanamycin, 1 µM monensin, or 20 mM NH₄Cl for 18 h. The cells were then washed three times with PBS and the cellular GSH was determined as described above.

RNA Extraction and Northern Blot Analysis

A549 cells were seeded at a density of 1.5 × 10⁶ cells per 100-mm cell culture dish in RPMI, 10% FBS, and incubated at 37° C in 5% CO₂. All cells were harvested at 72 h. The effect of HSA on -glutamylcysteine synthetase heavy subunit (-GCS) mRNA expression was determined by adding 2% HSA at 3, 6, 12, 14, or 24 h before harvesting the cells for mRNA extraction. Total cell RNA was isolated by a one-step guanidium-phenol chloroform extraction procedure (20). RNA was separated by electrophoresis in 1% agarose and transferred onto a hybond-N+ membrane (Amersham, Oakville, ON, Canada) for analysis. Membranes were prehybridized for 4 h in a mixture containing 120 mM Tris, 600 mM NaCl, 0.1% Na₄P₂O₇, 8 mM EDTA, 0.2% SDS, 625 µg/ml heparin, and 10% dextran sulfate at pH 7.4. Hybridization was performed overnight at 68° C in the same buffer. The human -GCS probe was obtained from the American Type Culture Collection (GenBank/EMBL: M90656; ATCC, Rockville, MD) (21) and labeled with the multiprime DNA labeling system (Amersham Life Science, ON, Canada), using [-³²P]dCTP (specific activity > 3,000 Ci/mmol; Amersham Life Science). The membrane was then washed once at room temperature (RT) for 20 min in 2× SSC, 1 h at 68° C in 0.1% SDS, 0.1× SSC, and rinsed at RT in 0.1× SSC. The membrane was exposed to Kodak X-OMAT film (Eastman Kodak Co., Rochester NY) with an intensifying screen at 80° C. As a control for RNA integrity, the blot was hybridized with a 1 kb PstI cDNA probe (ATCC) of the housekeeping gene glyceraldehyde phosphate dehydrogenase (GAPDH). Signal intensity was quantitated by densitometry with a Pharmacia LKB Ultroscan XL (Pharmacia Biotech, Uppsala, Sweden). Densitometric values are expressed as the ratio of -GCShs/GAPDH densitometric quantifications (n = 5).

Cytotoxicity Assay

To determine the effect of HSA on A549 cell susceptibility to H₂O₂-mediated injury, a cytotoxicity assay was utilized as previously described (16). The A549 cells were plated at 1.5 × 10⁵ per well in 24-well culture plates with RPMI, 10% fetal bovine serum (FBS) for 24 h, in 5% CO₂ at 37° C. The cells were washed three times, and RPMI was supplemented with or without 2% HSA for 18 h before being washed three times in Earle's balanced salt solution (EBSS). The cells were labeled with 0.05 µCi/well [8-¹⁴C]adenine (specific activity: 1.96 GBq/mmol, 53 mCi/mmol; Amersham Life Science). After three washes, ¹⁴C-labeled A549 cells were incubated in the presence or absence of different concentrations of H₂O₂ (0-2 mM) in 0.5 ml of EBSS for 7 h in 5% CO₂ at 37° C. The amount of ¹⁴C released in the supernatant was then quantitated. Results are expressed as a cytotoxicity index (CIX) determined with the formula: CIX = 100 × (A  B)/ (C  B), where A = dpm of test sample, B = dpm of spontaneous release in EBSS alone, and C = dpm of 1% Triton X-100-treated cells as previously described (16).

NF-B EMSA and Transactivation Assays

Modulation of NF-B activation by albumin was assessed by electrophoretic mobility shift assays (EMSA). A549 cells were seeded at 10⁶ cells/ml per 100-mm dish in RPMI, 10% FBS, 5% CO₂, 37° C for 24 h, then washed three times in PBS before adding either RPMI alone, 2% HSA, 2% HSA + 200 µM buthionine sulfoximine (BSO), 5% HSA, or 5% HSA + 200 µM BSO for 18 h. During the last 4 h of incubation, 5 ng/ml recombinant human tumor necrosis factor (rhTNF-; Sigma) was added to half of all the culture plates. After incubation, cells were scraped with a rubber policeman in PBS at 4° C, centrifuged, and nuclear extracts prepared at 4° C in lysis buffer (10 mM HEPES, 1 mM EDTA, 60 mM KCl, 1 mM DTT, 0.5% Nonidet P-40, pH 8.0) containing 1 tablet/10 ml Mini Complete (protease inhibitor mix), 1 mM Pefabloc SC (all reagents from Boehringer Mannheim). After centrifugation, cell nuclei were resuspended in 250 mM Tris-HCl, 60 mM KCl, 1 mM DTT, 1 mM Pefabloc SC, containing 1 tablet/10 ml Mini Complete, pH 7.8, lysed by freeze-thawing, and centrifuged at 13,000 × g for 10 min at 4° C. The supernatant was supplemented with 20% glycerol and total protein content was determined using the Bio-Rad assay (Bio-Rad Laboratories, Hercules, CA). Fractionated nuclear extracts (3 µg total protein) were added at RT for 30 min to 10 mM Tris-HCl, 50 mM NaCl, 1 mM DTT, 1 mM EDTA, 5% glycerol containing 0.2 µg dIdC, and 5,000 cpm of a ³²P-end-labeled double-stranded olidonucleotide with a high-affinity binding matrix as follows (binding site underlined):

5'-AGTTGAGGGGACTTTCCCAGGC-3'

3'-TCAACTCCCCTGAAAGGGTCCG-5'

DNA-binding reactions were analyzed in a 5% polyacrylamide gel and dried gels exposed to Kodak XAR5 film (Eastman Kodak Co.)

To determine whether HSA affected NF-B-dependent gene transcription, we performed a transactivation assay using the pNF-B-Luc vector from Clontech Laboratories Inc. (Palo Alto, CA). A549 cells were plated in 24-well dishes at 10⁵ cells/well in RPMI, 10 FBS, 5% CO₂ for 24 h. Cotransfection with 0.1 µg pNF-B-Luc and 0.1 ng pRL-SV40 renilla luciferase control reporter vector (Promega, Madison, WI), as an indicator of transfection efficiency, in the presence of 1 µl Fugene 6 Transfection Reagent (Boehringer Mannheim, Laval, QC, Canada) was carried out in 300 µl Opti-MEM (GIBCO BRL, Grand Island, NY) for 7 h at 37° C, 5% CO₂. The cells were then washed and medium replaced by RPMI with or without 2% HSA for 18 h at 37° C, 5% CO₂. The cells were then washed in PBS and stimulated with 5 ng/ml tumor necrosis factor alpha (TNF-) for 4 h. The cells were again washed, lysed, and luciferase activity determined with the dual luciferase reporter assay system (Promega) in a luminometer (Lumat LB9507; EG&G Berthold, St-Laurent, PQ, Canada).

Statistics

Results are expressed as the mean ± SEM. Data were analyzed using analysis of variance with Fisher's PLSD post hoc test. p < 0.05 was considered significant.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Albumin Depletion of Serum

The Cibacron Blue 3G affinity chromatography treatment of serum removed most of the albumin as determined by SDS- PAGE (Figure 1A, arrow corresponds to the expected migration of albumin), and the albumin could be recovered after elution of the proteins from the column. Incubation of either untreated serum or serum proteins eluted from the Cibacron Blue gel (Figure 1B, lanes 1 and 3) at a final HSA concentration of 2% increased cellular GSH from 4.58 ± 0.04 to 14.3 ± 0.94 and 12.27 ± 0.25 nmol/10⁵ cells (p < 0.001 both compared with media alone). In contrast, incubation of cells with serum depleted of albumin (lane 2) induced only a very slight increase in cellular GSH, which was markedly lower than that observed with either serum or with the albumin-enriched fraction (cells incubated with albumin-depleted serum, GSH = 5.81 ± 0.15 nmol/10⁵ cells, p < 0.001 versus GSH in cells incubated with whole serum or albumin-enriched fraction).

View larger version (32K): [in this window] [in a new window]   Figure 1.   Importance of albumin among extracellular proteins in the preservation of lung cellular GSH. (A) SDS PAGE of fresh human serum obtained from a healthy subject (lane 1), passed over a Cibacron Blue 3G column to remove most of the albumin (arrow, lane 2) and of albumin eluted from the Cibacron Blue 3G column (lane 3). (B) The effects of RPMI culture media alone (n = 8), whole serum (lane 1, n = 3), albumin-depleted fraction (lane 2, n = 10), or albumin-enriched fraction (lane 3, n = 10) on A549 cell GSH levels after 18 h. (*p < 0.001 versus media; §p < 0.001 versus serum;  p < 0.001 versus serum + Cibacron Blue 3G.)

Effects of HSA and rHSA on A549 Cell GSH

A549 cells incubated with 2% HSA had significantly higher cellular GSH levels as early as 8 h (11.19 ± 0.17 versus 7.15 ± 0.14, p < 0.001), and the difference between cells incubated with and without HSA was accentuated throughout the 24-h incubation period (Figure 2A). Between 4 and 12 h cells incubated without HSA showed a progressive decrease in GSH. In addition to the A549 cells (closed circles), HFL1 cells (closed triangles), and peripheral blood lymphocytes (open diamonds) showed a dose-dependent increase in GSH levels upon incubation with HSA (Figure 2B). All cell lines had significantly higher GSH levels when cultured in  0.5% HSA (p < 0.05 for all compared with no HSA). The GSH-enhancing effects of HSA were observed to the same extent with either HSA purified from blood or produced recombinantly (Table 1, Experiment 1). In addition, no GSH increase was observed with ₁-PI, the second most abundant protein in serum, nor with cysteine, GSH, or their corresponding disulfide forms. Incubation of cells with thiols, cysteine, and GSH did not increase cellular GSH and at the highest dose induced a significant decrease in cellular GSH.

View larger version (15K): [in this window] [in a new window]   Figure 2.   Albumin preserves and enhances lung cell GSH levels. (A) A549 cells were pretreated with either 0% (open circles) or 2% (closed circles) HSA for 4-24 h in RPMI and treated with 1% Triton X-100 to measure cell GSH using the enzymatic DTNB recycling assay. (B) Either A549 (closed circles), HFL1 cells (closed triangles), or peripheral blood lymphocytes (PBL) (open diamonds) were incubated with 0-2% human serum albumin for 18 h (A549 and HFL1 cells) or 48 h (lymphocytes) at 37° C, followed by treatment with 1% Triton X-100 to determine cell GSH. (A: n = 4 and results are representative of four separate experiments. B: n = 4 for A549 and PBL; n = 5 for HFL1 cells.)

Effect of HSA on Normal Lung Tissue In Vivo

Application of HSA at the nares, once daily for 3 d, increased tissue GSH content of the normal murine lung from 0.84 ± 0.07 to 1.43 ± 0.08 nmol/µg protein (p < 0.01) (Figure 3).

View larger version (30K): [in this window] [in a new window]   Figure 3.   Effect of airway instillation of HSA on lung parenchymal GSH content. C57BL/6 mice received nasal instillations of 50 µl of either saline or 7.5 mg HSA in saline, daily for 3 d. On Day 4, mice were sacrificed and lung vasculature washed thoroughly. The lungs were then homogenized in a Polytron tissue homogenizer for GSH and total protein determinations. n = 17 per group. *p < 0.01 between groups.

Reduction of HSA by Dithiothreitol

Human serum albumin in which free thiol on the Cys-34 residue was either reduced in a 1:1 molar ratio (thiol:HSA) or blocked with N-ethylmaleimide (NEM), increased cellular GSH in the A549 cell line as effectively as untreated HSA (A549 GSH levels for no HSA = 6.44 ± 0.19, untreated HSA = 10.22 ± 0.10, DTT-reduced HSA = 9.75 ± 0.11, NEM-blocked HSA = 9.60 ± 0.10 nmol/µg protein, p < 0.01 latter three conditions compared with no HSA, Table 2). However as HSA was progressively reduced with DTT beyond a 1:1 thiol: HSA ratio, its capacity to maintain or enhance cellular GSH was markedly decreased (Figure 4).

View larger version (19K): [in this window] [in a new window]   Figure 4.   Effects of thiol-mediated reduction of HSA on the capacity of HSA to modulate cellular GSH. HSA was treated with increasing concentrations of DTT, and extensively dialyzed. The number of thiol groups on HSA was determined spectrophotometrically using dithiobis(2-nitrobenzoic acid). The treated HSA added to RPMI medium and incubated with A549 cells at a final concentration of 2%, for 18 h, followed by the measurement of cellular GSH content. To obtain a thiol/ HSA ratio of 0, free thiol groups of HSA were blocked by treatment with N-ethylmaleimide as described in METHODS. The dotted line and shaded box represent the mean ± SEM of GSH levels in cells incubated without albumin. Each data point represents the mean ± SEM of four experiments. (*p < 0.05 versus thiol/HSA ratio = 0.)

Effects of Lysosomotropic Agents on Albumin-mediated GSH Modulation

The potent and specific vacuolar (H⁺)-ATPase inhibitors, bafilomycin A₁ and concanamycin, completely inhibited the effects of 2% HSA on cellular GSH in the A549 cell line. Similarly the ionophore monensin, which dissipates pH gradients across all cell membranes, markedly suppressed GSH synthesis in the presence of 2% HSA (Table 3). In marked contrast, a 20 mM concentration of the lysosomotropic agent NH₄Cl, known to alkalinize lysosomal pH and prevent HSA degradation, did not affect the increase in cellular GSH observed with 2% HSA (p > 0.3).

-Glutamylcysteine Synthetase Gene Expression

Incubation of A549 cells with 2% HSA for 3 to 24 h did not affect mRNA expression of the gene encoding GAPDH. In contrast, 2% HSA induced a time-dependent increase in mRNA expression for -GCS, reaching twice the level of cells incubated without albumin at 24 h (Figure 5).

View larger version (19K): [in this window] [in a new window]   Figure 5.   Effect of HSA on mRNA expression of the gene coding for the rate-limiting enzyme, -glutamylcysteine synthetase, in GSH synthesis. A549 cells were incubated with 2% HSA for 3-24 h. mRNA isolated from the cells was analyzed by Northern blot using probes against -GCS and GAPDH (as a reference). Results represent the mean ± SEM of 4 experiments. (p < 0.01 at 24 h compared with control samples at t = 0 h.)

Protection of A549 Cells against H₂O₂-mediated Cytotoxicity

H₂O₂-mediated cytotoxicity was measured to determine whether the differences in cellular GSH between A549 cells preincubated for 18 h with and without 2% HSA were sufficient to alter cell resistance to an oxidant stress. H₂O₂-mediated cytotoxicity in cells incubated with RPMI media alone was observed with an H₂O₂ concentration as low as 63 µM (Figure 6; CIX = 3.7 ± 1.3%, p < 0.05 compared with no H₂O₂). In contrast a fourfold higher H₂O₂ burden was needed to induce detectable cytotoxicity in cells that had been preincubated with 2% HSA (250 µM H₂O₂: CIX = 4.1 ± 0.3%, p = 0.005). Increased oxidant resistance of cells preincubated with 2% HSA was observed at H₂O₂ concentrations of 63-500 µM (p = 0.04 at 62.5 µM, p = 0.003 at 125 µM, p = 0.005 at 250 and 500 µM, all comparisons to RPMI media alone), but not at 1 mM H₂O₂ (2% HSA CIX = 31 ± 2.4 versus media alone CIX = 35.6 ± 1.2, p = 0.15).

View larger version (16K): [in this window] [in a new window]   Figure 6.   Protective effects of HSA against H2O2-dependent cytotoxicity. A549 cells were pretreated with either 0% (closed circles) or 2% (open circles) HSA for 18 h in RPMI media and labeled with [14C]adenine. The cells were then washed three times in EBSS and incubated with 0-1 mM H2O2 for 7 h at 37° C. Radioactivity released in the supernatant was measured as an index of cytotoxicity. The cytotoxicity index (CIX) was determined with the formula: CIX = 100 × (A  B)/ (C  B), where A = dpm of test sample, B = dpm of spontaneous release in EBSS alone, and C = dpm of 1% Triton X-100-treated cells. n = 3 per point; data are representative of four separate experiments. (*p < 0.05 versus 0% HSA.)

HSA-mediated Modulation of A549 Cells NF-B Activation

As expected, TNF- clearly induced the activation of NF-B as determined by EMSA (Figure 7A, arrow). Suppression of GSH alone with 200 µM BSO was not sufficient to activate NF-B, and BSO did not increase TNF--mediated NF-B activation. However, preincubation of the cells with 2% HSA and 5% HSA for 18 h clearly reduced the degree of NF-B activation. In addition, this protection of HSA against TNF--mediated NF-B activation was lost in the cells treated with BSO, suggesting that the decrease in NF-B activation observed with HSA was mediated by an increase in GSH synthesis.

View larger version (24K): [in this window] [in a new window]   Figure 7.   Effect of HSA on the activation of nuclear factor kappa B (NF-B). (A) Electrophoretic mobility shift assays of NF-B were carried out in A549 cells treated with either 0, 2%, or 5% HSA in the presence or absence of the GSH synthesis inhibitor, buthionine sulfoximine (BSO, 200 µg/ml), for 18 h at 37° C, followed by incubation with 5 ng/ml TNF- for 4 h. The upper band (arrow) represents the activated heterodimeric complex. Shown is one representative of four different experiments. (B) Transactivation assays of A549 cells transiently transfected with the pNF-B-Luc vector (firefly luciferase) and a pRL-SV40 (renilla luciferase) control. Cells were incubated with and without 2% HSA for 18 h, washed, and stimulated with or without 5 ng/ml TNF- for 4 h. Luciferase activity was determined using a dual luciferase reporter assay system. Induction units represent the fold-increase of the firefly/renilla luciferase ratio in TNF--stimulated versus unstimulated cells. Shown is the mean ± SEM of n = 6. (*p < 0.001 versus control;  p < 0.001 versus RPMI media alone + TNF-.)

Transfected cells incubated in RPMI media alone responded to TNF- stimulation by markedly increasing NF-B-dependent luciferase activity (Figure 7B, TNF- = 33.14 ± 2.97 versus without TNF- 1.25 ± 0.28 NF-B/RL induction units, p < 0.001). In contrast, cells that had been preincubated for 18 h with 2% HSA were significantly less responsive to TNF- than cells preincubated with RPMI media alone (NF-B/RL = 17.89 ± 1.18, p < 0.001 versus media alone with TNF-).

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

We have demonstrated that cellular and tissue levels of glutathione are regulated by extracellular concentrations of albumin. This is not a nonspecific protein effect since serum proteins depleted of albumin were unable to maintain GSH levels in A549 cells, and since up to 2% ₁-proteinase inhibitor, a major serum protein unrelated to albumin, did not affect cellular GSH levels. The HSA-induced increase in cellular GSH was not caused by plasma-derived impurities in the HSA preparation since recombinant HSA purified from the yeast Pichia pastoris caused a similar increase in cellular GSH. We conclude from these studies that albumin plays an essential role in determining the levels of cellular GSH in the lung and likely in other tissues.

Several pieces of evidence provide insights into potential mechanisms by which albumin may affect cellular GSH. Our results clearly indicate that the HSA-mediated GSH increase in cells is not related to the redox status of the free cysteine residue at position Cys-34 since reduction with DTT and oxidation with NEM did not inhibit the GSH-enhancing properties of HSA. However, albumin is known to be taken up by cells through receptor-mediated and fluid-phase endocytic pathways as well as by transcytosis (11, 22). The turnover rate of albumin is very high, suggesting that large amounts of albumin are continuously taken up by tissues (25). One of the striking features of albumin is the presence in each molecule of 34 cystine residues and one free cysteine group, which can carry glutathione (2). Although exogenously added cystine, cysteine, GSH, and GSSG did not increase cellular GSH, it is conceivable that by entering cells through endocytosis albumin could provide a more efficient source of GSH precursor molecules. Albumin is degraded in lysosomes, a process inhibited by agents that disrupt lysosomal acidification (26). Bafilomycin A₁ and concanamycin are specific inhibitors of the vacuolar (H⁺)-ATPase pump capable of disrupting endosomal pH (29, 30). When we incubated A549 cells with either bafilomycin A₁ or concanamycin, the GSH-enhancing properties of albumin were completely suppressed. The suppression was not specific to vacuolar (H⁺)-ATPase inhibitors since a similar effect was observed with the ionophore monensin, a sodium/ proton exchanger known to dissipate pH gradients across all cell membranes. Although these data would be consistent with the concept that GSH modulation is mediated by pH-sensitive lysosomal degradation of albumin, a key experiment seems to contradict this conclusion. The lysosomotropic agent NH₄Cl, at concentrations of up to 20 mM did not have any effect on cellular GSH levels in the presence of HSA. Since NH₄Cl clearly disrupts lysosomal pH and inhibits lysosomal degradation of albumin (24, 26), other mechanisms must be considered to explain the link between albumin and cellular GSH.

Removal of serum proteins from cell culture medium is known to induce apoptosis (31). Furthermore, Zoellner and coworkers have demonstrated that among serum proteins, albumin plays an essential role in preventing endothelial cell apoptosis, an effect that is abolished when albumin is reduced with DTT (14, 32). Similarly, in the current study we observed that the capacity of HSA to preserve and increase GSH was progressively lost as HSA was reduced with DTT, an observation consistent with the concept that albumin-dependent GSH regulation and prevention of apoptosis are controlled by similar pathways.

When A549 cells were incubated in the presence of HSA we observed a consistent and specific increase in the mRNA of the gene encoding -glutamylcysteine synthetase (-GCS, the rate-limiting enzyme in GSH synthesis) without any change in GAPDH mRNA expression, indicating that the effect of albumin on -GCS mRNA was relatively specific. These results suggest that modulation of -GCS gene expression may be one of the mechanisms by which albumin regulated cellular GSH at the later (> 12 h) time points.

The minimal concentration of albumin needed to maintain cellular GSH levels was 0.5% in all cells tested (A549 and HFL1 cell lines, and peripheral blood lymphocytes). The albumin concentration at the normal epithelial surface of the lung is estimated to be approximately 0.35% (33), a level at which relatively small changes in albumin were found to result in the largest variations of cellular GSH. The albumin content of epithelial and interstitial fluids is therefore considerably lower than that of plasma and may well be within the critical range at which we observed significant modulation of cellular GSH.

The GSH-enhancing effects of HSA were not specific to lung cells. HSA also caused a dose-dependent increase in peripheral blood lymphocyte GSH. This may be particularly relevant to patients with hypoalbuminemia who are known to have defective cellular immunity (34). Glutathione is essential for normal lymphocyte cellular and humoral immune responses (35). Based on the present study in which peripheral blood lymphocyte GSH levels were found to be modulated by HSA, we propose that the mechanism by which hypoalbuminemia impedes normal immune responses may be related, at least in part, to a decrease in lymphocyte GSH.

Acute inflammation is characterized by increased vascular permeability and subsequent albumin leakage into the interstitium. During the adult respiratory distress syndrome, alveolar epithelial albumin concentrations can markedly increase. We observed that increasing the HSA content at the murine lung epithelial surface significantly augmented the lung tissue GSH. These results indicate that not only is albumin capable of preserving levels of cellular GSH but it also can increase lung tissue GSH above normal levels. This may have physiological relevance in situations of increased oxidant stress. In the current study, we observed that the HSA-mediated increase in GSH levels of A549 cells was sufficient to provide significant protection against H₂O₂-dependent cytotoxicity. Other tissues may also utilize HSA to limit the extent of oxidant injury. Recently it has been demonstrated in rats that the extent of brain tissue necrosis following prolonged ischemia is markedly reduced by the perfusion of a solution containing HSA in the ischemic area (36). Albumin-mediated protection of brain tissue correlated with cellular uptake of albumin. Since oxidants contribute to necrosis following ischemia-reperfusion injury, it is possible that albumin protected the brain tissue by increasing GSH levels.

In addition to increased vascular permeability, acute inflammation is also characterized by the recruitment of activated inflammatory cells releasing oxidants that can cause tissue damage and signal inflammatory gene transcription, thus fuelling the inflammatory cascade (37). In the current study, we have observed that not only does the albumin-mediated increase in GSH protect cells against H₂O₂, but it also decreases the activation of the nuclear transcription factor kappa B (NF-B) following stimulation with TNF-. Both 2% and 5% HSA partially suppressed TNF--mediated NF-B activation. The protective effect of HSA on NF-B activation was abolished by BSO, a specific inhibitor of GSH synthesis, thus confirming that the albumin-dependent downregulation of NF-B activation was mediated by GSH. The capacity of albumin to decrease NF-B-dependent gene transcription was confirmed using transactivation assays in which a reporter gene coding for luciferase was driven by a nucleotide construct containing four NF-B recognition sites in the 5'-flanking region. Cells transfected with the NF-B promoter/luciferase reporter gene construct showed a significantly lower transcriptional response to TNF- when preincubated with 2% HSA. Our studies suggest that the increased leakage of albumin into the interstitial space during inflammation may play a role in controlling the inflammatory response.

In summary, the current study demonstrates that HSA plays a key role in regulating GSH levels in lung cells and peripheral blood lymphocytes. The inhibitors of vacuolar (H⁺)- ATPase bafilomycin A₁ and concanamycin, as well as the transmembrane pH-disrupting agent monensin each blocked the GSH increase. However, regulation of GSH by HSA is unlikely to be mediated solely through lysosomal degradation of albumin since the lysosomal pH-disrupting agent NH₄Cl did not prevent the albumin-mediated GSH increase. These results suggest that, as in apoptosis, the cytoplasmic rather than the vacuolar pH may be important in regulating the HSA-mediated changes in GSH. Regardless of the mechanisms, the levels to which albumin raises cellular GSH are sufficient to protect lung epithelial cells against H₂O₂-mediated cytotoxicity and to decrease activation of the transcription protein complex NF-B by TNF-. These results suggest that modulation of cellular GSH is a key mechanism by which human serum albumin may protect tissues against oxidant injury, control inflammatory gene transcription, and regulate apoptosis.