INTRODUCTION

Top Abstract Introduction Materials & Methods Results Discussion References

IN SETTINGS OF inflammation, tissue hypoxia often occurs. For instance, the hallmark of septic shock is vascular hypotension, leading to widespread cell and tissue hypoxia (reviewed in Ref. 24, 33). Endothelial cells that line blood vessels bear activable receptors for many inflammatory stimuli, including lipopolysaccharides (LPS). Because endothelia are anatomically positioned at the interface of the blood and tissue exchange, endothelia are especially influenced by conditions of hypoxemia (22). A number of studies have examined the influence of hypoxia on endothelial cell structure/function (recently reviewed in Ref. 30). Hypoxia-induced changes are complex and include changes in energy metabolism, alterations in gene expression, and induction of specific cell surface proteins (reviewed in Ref. 30).

Tissue injury resulting from reperfusion of hypoxic tissue has been shown to be mediated, at least in part, by recruitment of leukocytes (32). Leukocyte recruitment across the intact endothelium occurs through a concerted series of adhesion and deadhesion events involving a number of cell surface adhesion proteins (28). Previous studies have demonstrated that short-term hypoxia/anoxia induces increased adhesion of leukocytes to vascular endothelial cells, and such adhesion can be blocked, at least in part, by antibodies directed at intercellular adhesion molecule 1 (ICAM-1, CD54) (26, 35), a 90- to 110-kDa glycoprotein found on the surface of endothelial cells that mediates the firm adherence of leukocytes to the endothelium (28). ICAM-1 expression is readily induced by stimulation of cell surface receptors for a number of inflammatory mediators, including LPS (28), and is regulated by the nuclear transcription factor (NF)-B, a pleiotropic activator of a number of genes related to inflammation (9). Previous studies have defined a role for the ubiquitin-mediated proteasome pathway in NF-B induction, specifically through degradation of p50 and inhibitory (I)-B (20). At present, it is unclear whether molecules such as ICAM-1 are differentially regulated by a combination of hypoxia and mediators found within the microenvironment of inflamed tissue.

Thus we hypothesized that hypoxia may augment cell surface expression of leukocyte adhesion receptors, such as ICAM-1. We report here that, whereas hypoxia alone does not induce ICAM-1 expression over baseline, the combination of hypoxia and LPS enhances expression of ICAM-1. Such enhancement was not universal for other extracellular membrane proteins and correlated with hypoxia-elicited increases in nuclear content of the NF-B p65 subunit. We speculate that in the setting of hypoxia, inflammatory mediators such as LPS may perpetuate expression of molecules important in leukocyte trafficking, such as ICAM-1.

	MATERIALS AND METHODS

Top Abstract Introduction Materials & Methods Results Discussion References

Cell culture. Human umbilical vein endothelial cells (HUVECs) were obtained and harvested as described elsewhere (13) using 0.1% collagenase (Worthington Biochemical, Freehold, NJ). Human pulmonary microvascular endothelial cells (HPMVECs) were purchased from Clonetics (San Diego, CA) as first passage cells and, when used, were cultured under similar conditions as for HUVECs. Endothelial monolayers were established, maintained, and subcultured with Dulbecco's modified Eagle's medium (GIBCO, Grand Island, NY) containing 10% heat-inactivated fetal calf serum, glucose, pyruvate, glutamine, penicillin, and streptomycin (13). Confluent endothelial monolayers exhibited typical cobblestone appearance and uptake of acetylated low-density lipoprotein (Biomedical Technology, Stoughton, MA; data not shown). Where indicated, endothelial monolayers were exposed to LPS (from Escherichia coli; Sigma, St. Louis, MO) at indicated concentrations.

Cell surface immunoassay. ICAM-1 cell surface expression was quantified with a cell surface enzyme-linked immunosorbent assay (ELISA), as previously described (7). Endothelial cells were grown and assayed for antibody binding following exposure to normoxia or hypoxia in the presence or absence of LPS, as indicated. Endothelia were lightly fixed with paraformaldehyde (1% wt /vol in phosphate-buffered saline) to preserve cell surface protein. Cells were washed with Hanks' balanced salt solution (HBSS; Sigma) and blocked with medium for 30 min at 4°C. Anti-ICAM-1 monoclonal antibody (MAb) [clone P2A4 (12), obtained from the Developmental Studies Hybridoma Bank (Iowa City, IA) and used as undiluted cell culture supernatant] or R6.5 (27) [a kind gift from Dr. Robert Rothlein, Boehringer Ingelheim Pharmaceuticals (Ridgefield, CT), and used as purified MAb at 20 µg/ml] was added to fixed cells and allowed to incubate for 2 h at 4°C. Where indicated, MAb to major histocompatibility complex (MHC) class I (5) (clone W6/32, obtained from the American Type Culture Collection and used as 1:100 diluted ascitic fluid) was used as a control. After a wash with HBSS, a peroxidase-conjugated sheep anti-mouse secondary antibody (Cappel, West Chester, PA) was added. Secondary antibody (1:1,000 final dilution) was diluted in medium containing 10% fetal bovine serum (FBS). After washing, plates were developed by addition of peroxidase substrate [2,2'-azino-bis(3-ethylbenzthiazoline-6-sulfonic acid), 1 mM final concentration, Sigma] and read on a microtiter plate spectrophotometer at 405 nm (Molecular Devices). Controls consisted of medium only and secondary antibody only. Optical density data at 405 nm (OD₄₀₅, background subtracted) are presented as means ± SE.

Immunoprecipitation of biotinylated endothelial membranes. HUVECs were grown to confluence on six-well plates, exposed to experimental conditions, and washed with HBSS. Extracellular cell surface proteins were then labeled with biotin (ImmunoPure sulfo-NHS-biotin; 1 mM; Pierce, Rockford, IL) as described previously (18). Unbound biotin was quenched with NH₄Cl (50 mM) in HBSS. Labeled HUVECs were lysed [in buffer containing 150 mM NaCl, 25 mM tris(hydroxymethyl)aminomethane (Tris), 1 mM MgCl₂, 1% Triton X-100, 1% Nonidet P-40, 5 mM EDTA, 5 µg/ml chymostatin, 2 µg/ml aprotinin, and 1.25 mM phenylmethylsulfonyl fluoride (PMSF), all from Sigma]. Cell debris was removed by centrifugation (10,000 g, 5 min). HUVEC lysates were precleared with 50 µl preequilibrated protein G-Sepharose (Pharmacia, Uppsala, Sweden) for 2 h. For immunoprecipitation of ICAM-1, primary antibody (500 µl of P2A4 cell culture supernatant) was added (immunoprecipitated for 2 h), followed by addition of 50 µl of preequilibrated protein-G Sepharose (immunoprecipitated overnight on an end-over-end rotator). Washed immunoprecipitates were boiled in nonreducing sample buffer [2.5% sodium dodecyl sulfate (SDS), 0.38 M Tris (pH 6.8), 20% glycerol, and 0.1% bromphenol blue], separated by SDS-polyacrylamide gel electrophoresis (PAGE; 10% linear gel) under nonreducing conditions and transferred to nitrocellulose using standard protocols. Biotinylated proteins were labeled with streptavidin-peroxidase and visualized by enhanced chemiluminescence (ECL; Amersham, Arlington Heights, IL). Resulting ICAM-1 bands were quantified from scanned images using National Institutes of Health (NIH) Image software (Bethesda, MD).

Lymphocyte adhesion assay. Peripheral blood mononuclear cells (PBMC) were obtained by centrifugation through Ficoll-Hypaque (1.077). Lymphocyte-enriched fractions were obtained by incubating PBMC in 10% FBS-RPMI 1640 on tissue culture-treated plastic 24-well plates for 1 h at 22°C and collecting nonadherent cells. For studies of adhesion, enriched lymphocyte populations (>95%) were labeled for 30 min at 37°C with 2',7'-bis(carboxyethyl)-5(6)-carboxyfluorescein-acetoxymethyl ester (BCECF-AM, 5 µM final concentration; Calbiochem, San Diego, CA) and washed three times in HBSS. Labeled lymphocytes (1 × 10⁵/monolayer) were added to washed normoxic or hypoxic monolayers, plates were centrifuged at 150 g for 4 min to uniformly settle lymphocytes, and adhesion was allowed for 10 min at 37°C. Monolayers were gently washed three times with HBSS, and fluorescence intensity (485-nm excitation, 530-nm emission) was measured on a fluorescent plate reader (Cytofluor 2300, Millipore, Bedford, MA). Adherent cell numbers were determined from standard curves generated by serial dilution of known lymphocyte numbers diluted in HBSS. All data were normalized for background fluorescence by subtraction of fluorescence intensity of samples collected from monolayers incubated in buffer only, without addition of lymphocytes.

ICAM-1 antisense oligonucleotide treatment of HUVECs. Antisense oligonucleotide treatment of confluent HUVECs was done exactly as described previously using the oligonucleotides ISIS-1939 (targets the mRNA 3' untranslated region, sequence CCCCCACCACTTCCCCTCTC) or ISIS-3067 (targets the mRNA 5' untranslated region, sequence TCTGAGTAGCAGAGGAGCTC) (6). Oligonucleotides were a kind gift from Dr. C. F. Bennett, ISIS Pharmaceuticals (Carlsbad, CA). Confluent HUVECs were washed in serum-free Opti-MEM (GIBCO) and then in Opti-MEM containing 10 µg/ml N-[1-(2,3-dioleyloxy)propyl]-N,N,N-trimethylammonium chloride (Lipofectin solution at 10 µg/ml, GIBCO), followed by the indicated concentrations of filter-sterilized oligonucleotides. Cells were incubated for 4 h at 37°C and then replaced with normal growth medium in the presence or absence of LPS and incubated in normoxia or hypoxia as described above. ICAM-1 expression or MHC class I expression (control) was quantified with the methods described in Cell surface immunoassay. Methods for lymphocyte adhesion to such oligonucleotide-treated cells were carried exactly as described in Lymphocyte adhesion assay.

Proteasome inhibition experiments. Confluent HUVEC monolayers grown on 96-well plates were equilibrated to hypoxia or normoxia for 1 h in the presence of the proteasome inhibitors N-acetyl-Leu-Leu-methioninal (ALLM, calpain inhibitor II, purchased from Boehringer Mannheim, Indianapolis, IN) or N-acetyl-Leu-Leu-norleucinal (ALLN, calpain inhibitor I, purchased from Boehringer Mannheim) at indicated concentrations before addition of LPS. After 18 h in hypoxia or normoxia, cell surface expression of ICAM-1 and MHC class I was examined by cell surface ELISA as described in Cell surface immunoassay.

Measurement of intracellular pH. Intracellular pH was measured as described previously (1) using HUVECs grown to confluence on 96-well plates. Briefly, endothelia were exposed to combinations of hypoxia or normoxia with or without addition of LPS (range was 1-100 ng/ml) for 18 h. Within the setting of normoxia or hypoxia, monolayers were washed with preequilibrated HBSS and loaded with BCECF-AM (10 µM final concentration, Calbiochem) for 30 min at 37°C. Loaded monolayers were washed three times in preequilibrated HBSS, and fluorescence intensity (485-nm excitation, 530-nm emission) was immediately measured on a fluorescent plate reader (Cytofluor 2300, Millipore). Similarly loaded cells were also measured at a reference wavelength (360-nm excitation, 530-nm emission), and the fluorescence ratio of 485:360 was used to determine intracellular pH. Intracellular pH was calibrated using nigericin (10 µg/ml) to equilibrate intracellular and extracellular pH (1), and this calibration was linear in the pH range of 6.7-7.5. In subsets of experiments, pH of extracellular medium was adjusted with HCl in the range of 6.8-7.3 before addition to HUVEC monolayers.

NF-B p65 Western blotting. After experimental treatment of HUVECs, nuclear extracts were prepared as described previously (34). Confluent monolayers of HUVECs in 100-mm petri dishes were washed in ice-cold phosphate-buffered saline and lysed by incubation in 500 µl of buffer A [in mM: 10 N-2-hydroxyethylpiperazine-N'-2-ethanesulfonic acid (HEPES), pH 8.0, 1.5 MgCl₂, 10 KCl, 0.5 dithiothreitol (DTT), 200 sucrose, and 0.5 PMSF and 1 µg/ml of both leupeptin and aprotinin and 0.5% Nonidet P-40] for 5 min at 4°C. The crude nuclei released by lysis were collected by microcentrifugation (15 s). Nuclei were rinsed once in buffer A and resuspended in 100 µl of buffer C [in mM: 20 HEPES (pH 7.9), 1.5 MgCl₂, 420 NaCl, 0.2 EDTA, 0.5 PMSF, and 1.0 DTT and 1 µg/ml of both leupeptin and aprotinin]. Nuclei were incubated on a rocking platform at 4°C for 30 min and clarified by microcentrifugation for 5 min. Proteins were measured (detergent compatible protein assay, Bio-Rad, Hercules, CA). Samples (25 µg/lane, as indicated) of HUVEC lysates were separated by nonreducing SDS-PAGE, transferred to nitrocellulose, and blocked overnight in blocking buffer (250 mM NaCl, 0.02% Tween 20, 5% goat serum, and 3% bovine serum albumin). Primary antibody (rabbit polyclonal specific for p65 subunit of NF-B; Biomol Research Laboratories, Plymouth Meeting, PA) was added for 3 h, blots were washed, and species-matched peroxidase-conjugated secondary antibody was added. Labeled bands from washed blots were detected by ECL. Resulting 65-kDa NF-B bands were quantified from scanned images using NIH Image software. Such 65-kDa bands were specific for NF-B, since preincubation of rabbit polyclonal antibody with standard p65 antigen (provided by Biomol as a control) resulted in a diminution of the 65-kDa band by more than 70% (data not shown).

Data presentation. Data were compared by analysis of variance (ANOVA) or by Student's t-test. Values are expressed as means ± SE of n experiments.

	RESULTS

Top Abstract Introduction Materials & Methods Results Discussion References

Hypoxia enhances LPS-induced ICAM-1 expression. Endothelial monolayers tolerated exposure to hypoxia well (ambient O₂ as low as 1%, for up to 24 h). No changes in morphology were observed, and no evidence of cell death was apparent (based on soluble lactate dehydrogenase measurements from supernatants, data not shown). We first examined the induction of ICAM-1 surface expression induced by LPS under conditions of endothelial exposure to hypoxia or to normoxia. In Fig. 1, the dose responses for LPS induction of ICAM-1 on HUVEC and on HPMVEC by cell surface ELISA are shown. As can be seen, unstimulated expression of ICAM-1 was low, and hypoxia alone (1% ambient O₂) did not induce cell surface ICAM-1 on either HUVEC or HPMVEC (P = not significant compared with normoxic control). However, in the presence of LPS, more than twofold augmentation in ICAM-1 was evident on both HUVEC (two-way ANOVA, P < 0.025 compared with normoxia) and HPMVEC (two-way ANOVA, P < 0.01 compared with normoxia).

View larger version (11K): [in this window] [in a new window]   Fig. 1.   Induction of endothelial intercellular adhesion molecule 1 (ICAM-1) by lipopolysaccharide (LPS) and hypoxia. Confluent human umbilical vein endothelial cell (HUVEC; A) or human pulmonary microvascular endothelial cell (B) monolayers were exposed to hypoxia (ambient O2 of 1%, ) or normoxia (ambient O2 of 21%, ) in the presence or absence of indicated concentrations of LPS for 18 h. Monolayers were washed and lightly fixed (1% paraformaldehyde), followed by addition of monoclonal antibody (MAb) specific for ICAM-1 (clone P2A4), and analyzed for specific expression by enzyme-linked immunosorbent assay (ELISA) as described in MATERIALS AND METHODS. Results of optical density at 405 nm (OD405, background subtracted) are means ± SE of 12 monolayers in each condition from at least 3 experiments.

View larger version (20K): [in this window] [in a new window]   Fig. 2.   Time and O2 dependence of hypoxia-elicited increase in ICAM-1 expression. Confluent HUVEC monolayers were exposed to hypoxia () or normoxia () in the presence of LPS (100 ng/ml) for indicated periods of time (A) or at indicated O2 concentrations (B). Monolayers were washed and lightly fixed (1% paraformaldehyde), followed by addition of MAb specific for ICAM-1 (clone P2A4), and analyzed for specific expression by ELISA as described in MATERIALS AND METHODS. OD405 (background subtracted) results are means ± SE of 6-8 monolayers in each condition from at least 3 experiments.

Hypoxia increases immunoprecipitable ICAM-1. To confirm the findings of our whole cell ELISA, we determined the influence of hypoxia on immunoprecipitable ICAM-1 from biotinylated HUVEC monolayers. As shown in Fig. 3, HUVEC exposure to LPS (5 or 50 ng/ml, 18 h) under normoxic conditions revealed immunoprecipitation of an ~95-kDa biotinylated protein consistent with endothelial ICAM-1. Some expression was apparent from control samples not exposed to LPS, consistent with our ELISA findings (Figs. 1 and 2). Confluent HUVECs exposed to a combination of LPS and hypoxia revealed an immunoprecipitable band of increased density over normoxia at LPS concentrations of 5 and 50 ng/ml (Fig. 3). Consistent with our ELISA findings, no observable increase in ICAM-1 was apparent with hypoxia alone (Fig. 3). Densitometric analysis of these bands (Fig. 3B) revealed a 330% increase at 5 ng/ml LPS and an 83% increase at 50 ng/ml LPS under conditions of hypoxia (Fig. 3B), confirming our ELISA results.

View larger version (22K): [in this window] [in a new window]   Fig. 3.   Immunoprecipitation of ICAM-1 from cell surface biotinylated HUVECs induced by LPS in normoxia and hypoxia. Confluent HUVEC monolayers were exposed to medium alone (lanes 2 and 5) or medium containing 5 ng/ml LPS (lanes 3 and 6) or 50 ng/ml LPS (lanes 4 and 7) under conditions of normoxia (ambient O2 of 21%, lanes 2-4) or hypoxia (ambient O2 of 1%, lanes 5-7) for 18 h. Cells were cooled to 4°C, and cell surface proteins were biotinylated, followed by immunoprecipitation of ICAM-1 (with the exception of lane 1, negative control, no primary antibody) as described in MATERIALS AND METHODS. Blots were probed with streptavidin peroxidase and developed by enhanced chemiluminescence (ECL). A: resulting blots from immunoprecipitation of normoxic and hypoxic HUVEC with and without addition of LPS. MW, molecular weight. B: densitometry tracings from immunoprecipitated bands shown in A; representative of 1 of 4 experiments.

Hypoxia enhances ICAM-1-dependent lymphocyte adhesion. We next examined whether hypoxic conditions demonstrated here to enhance ICAM-1 expression on HUVECs were apparent at the functional level. First, as shown in Fig. 4, a significant increase in lymphocyte binding was apparent in HUVECs exposed to a combination of hypoxia and LPS (100 ng/ml) over that of exposure to LPS alone [see conditions of the binding control (MAb W6/32), 2.1 ± 0.4-fold increase with hypoxia, P < 0.001]. Lymphocyte adhesion to LPS-preexposed HUVECs was significantly diminished by addition of anti-ICAM-1 MAb directed against the LFA-1-dependent ICAM-1 site (29) (MAb was RR1.1, adhesion decreased by 79 ± 9% and 70 ± 7% for normoxia and hypoxia in the presence of LPS compared with binding control W6/32, respectively, P < 0.001 for both) but not the Mac-1-dependent ICAM-1 site (11) (MAb was R6.5, adhesion decreased by 11 ± 7% and 13 ± 5% for normoxia and hypoxia in the presence of LPS compared with binding control W6/32, respectively, P = not significant for both), indicating that a primary contribution of lymphocyte binding is LFA-1 binding through ICAM-1. However, even in the presence of saturating concentrations of MAb RR1.1 (20 µg/ml, determined by dilution and cell surface ELISA, data not shown), a nearly twofold increase in lymphocyte binding was apparent in HUVECs exposed to both LPS and hypoxia (Fig. 4). No significant differences existed in lymphocyte binding to HUVECs exposed to hypoxia or normoxia alone (Fig. 4, P = not significant for all).

View larger version (20K): [in this window] [in a new window]   Fig. 4.   Influence of anti-ICAM-1 MAb on lymphocyte adhesion following endothelial exposure to LPS and hypoxia. Confluent HUVEC monolayers were exposed to hypoxia alone (shaded bars), normoxia alone (hatched bars), hypoxia with LPS (500 ng/ml, open bars), or normoxia with LPS (500 ng/ml, solid bars) for 18 h. Monolayers were preincubated with anti-ICAM-1 MAb R6.5 (20 µg/ml), anti-ICAM-1 MAb RR1/1 (20 µg/ml), or anti-major histocompatibility complex (MHC) class I MAb W6/32 (20 µg/ml) and analyzed for lymphocyte adhesion as described in MATERIALS AND METHODS. Results of percent lymphocyte bound are means ± SE of 12 monolayers in each condition from 3 experiments.

View larger version (13K): [in this window] [in a new window]   Fig. 5.   Influence of ICAM-1 mRNA antisense oligonucleotides on lymphocyte adhesion following endothelial exposure to LPS and hypoxia. Confluent HUVEC monolayers were pretreated with indicated concentrations of the ICAM-1 antisense oligonucleotides (oligo) ISIS-1939 (A) or ISIS-3067 (B) as described in MATERIALS AND METHODS. Monolayers were then exposed to exposed to hypoxia (filled symbols) or normoxia (open symbols) in the presence of LPS (100 ng/ml) for 18 h. Monolayers were washed and lightly fixed (1% paraformaldehyde), followed by addition of MAb specific for ICAM-1 (circles, clone P2A4) or MHC class I (squares, clone W6/32), and analyzed for specific expression by ELISA. C: lymphocyte adhesion from monolayers exposed to ISIS-1939 (5 µM) or ISIS-3067 (5 µM) followed by LPS or no LPS. AS, antisense. Results are means ± SE of 6-10 monolayers in each condition from at least 3 different experiments.

Evidence for increased proteasome activation in hypoxic endothelia. The NF-B/Rel A family of transcription factors are important for induction and expression of a number of cellular gene products, including endothelial ICAM-1 (8). In quiescent endothelia, NF-B is complexed to the inhibitor I-B. When activation occurs, proteasome-dependent I-B degradation (9) allows cytoplasmic-to-nuclear localization of NF-B. Thus, as a measure of proteasome activation, we examined cytoplasm-to-nuclear localization of NF-B (p65 subunit) in hypoxic and normoxic endothelia. As shown in Fig. 6, and similar to our findings with ICAM-1 surface expression (Fig. 1), endothelial exposure to hypoxia alone failed to induce p65 nuclear localization. However, in the presence of LPS (concentration range of 1-100 ng/ml), a four- to sixfold increase in nuclear-to-cytoplasmic p65 ratio was observed in cells exposed to a combination of hypoxia and LPS (ANOVA, P < 0.05 compared with normoxia). Under these conditions, no observable differences between hypoxia or normoxia were observed in the cytoplasmic fraction of p65 (Fig. 6).

View larger version (28K): [in this window] [in a new window]   Fig. 6.   Western blot analysis of HUVEC nuclear transcription factor (NF)-B p65 subunit in endothelial cytoplasmic and nuclear fractions. Confluent HUVEC monolayers were exposed to medium or medium containing LPS (50 ng/ml) under conditions of normoxia (ambient O2 of 21%) or hypoxia (ambient O2 of 1%) for 4 h. Cells were cooled to 4°C, and cytoplasmic (cyto) and nuclear fractions were isolated as described in MATERIALS AND METHODS. Samples (25 µg/lane) were separated by SDS-polyacrylamide gel electrophoresis under nonreducing conditions, and blots were probed with polyclonal rabbit NF-B p65, labeled with peroxidase-conjugated secondary antibody, and developed by ECL. A: representative blot from cytoplasmic and nuclear fractions in normoxia or hypoxia in the presence of indicated concentrations of LPS (in ng/ml). B: demonstration of ratio of nuclear to cytoplasmic p65 from scanning densitometry pooled from 3 separate experiments. Hatched bars, hypoxia; solid bars, normoxia.

View larger version (13K): [in this window] [in a new window]   Fig. 7.   Influence of proteasome inhibitors on hypoxia-elicited augmentation in HUVEC ICAM-1 expression. Confluent HUVEC monolayers were preequilibrated to hypoxia (ambient O2 of 1%, filled symbols) or normoxia (ambient O2 of 21%, open symbols) for 1 h in the presence of indicated concentrations of N-acetyl-Leu-Leu-norleucinal (ALLN, calpain inhibitor I, A) or N-acetyl-Leu-Leu-methioninal (ALLM, calpain inhibitor II, B). Monolayers were then stimulated for 18 h with LPS (50 ng/ml) and analyzed for specific ICAM-1 expression by ELISA. OD405 (background subtracted) results are means ± SE of 8-10 monolayers in each condition from at least 3 experiments. C: influence of ALLN (100 µM) on LPS-stimulated (50 ng/ml) cytoplasmic localization of p65 by Western blot from endothelia exposed to hypoxia or normoxia.

Role for metabolic acidosis in hypoxia-elicited increases in ICAM-1. A number of recent in vivo and ex vivo studies indicate a role for acidosis in activation of the ubiquitin-proteasome pathway (2, 15, 19). Thus we determined whether cell culture acidosis, measured as pH of extracellular medium and intracellular pH, was associated with our observed increase in ICAM-1 expression. As shown in Fig. 8A, 24-h endothelial exposure to hypoxia alone resulted in a small decrease in medium pH (P < 0.05). The combination of hypoxia and LPS enhanced this decrease in medium pH (ANOVA, P < 0.01) in a concentration-dependent manner. A small but significant fall in medium pH was associated with the highest LPS concentration (100 ng/ml) in normoxia (P < 0.05). Similar to extracellular medium, parallel decreases in intracellular pH were observed with endothelial exposure to hypoxia and LPS (Fig. 8B, ANOVA, P < 0.025).

View larger version (11K): [in this window] [in a new window]   Fig. 8.   Hypoxia-elicited enhancement is dependent on medium acidification. A: confluent HUVEC monolayers were exposed to hypoxia () or normoxia () in the presence of indicated concentrations of LPS for 18 h. Medium was collected, and pH was immediately measured. pH results are means ± SE of 4 monolayers in each condition from at least 3 experiments. B: intracellular pH from similarly exposed cells. C: confluent HUVEC monolayers were exposed to combination of hypoxia and LPS (100 ng/ml) in the medium with indicated concentrations of bicarbonate for 18 h and analyzed for medium acidification () and specific expression of ICAM-1 by ELISA (). Results are means ± SE of 4 monolayers in each condition from at least 3 experiments.

	DISCUSSION

Top Abstract Introduction Materials & Methods Results Discussion References

In a variety of disease states, tissue hypoxia occurs in conjunction with other inflammatory events and is associated with accumulation of leukocytes. Leukocyte recruitment to such sites involves a concerted series of adhesion and deadhesion events between leukocytes and vascular endothelial cells. An important adhesion molecule in this leukocyte recruitment cascade is ICAM-1. For this reason, we hypothesized that hypoxia may differentially regulate stimulated endothelial ICAM-1 expression. Our findings indicate that, whereas hypoxia alone fails to induce endothelial ICAM-1, in the presence of LPS, hypoxia enhances both expression of ICAM-1 and lymphocyte adhesion. With the use of specific inhibitors and cytoplasmic-to-nuclear translocation of NF-B, these studies implicate a hypoxia-elicited activation of proteasomes. Finally, our data reveal that conditions that result in enhanced ICAM-1 expression are associated with metabolic acidosis, the reversal of which normalizes ICAM-1 expression.

In settings of tissue hypoperfusion in vivo, endothelial surfaces can be exposed to high concentrations of inflammatory stimuli. A primary example is the state of sepsis in which profound alterations in tissue perfusion are coincident with activation of high circulating levels of LPS and other inflammatory stimuli (33). Thus a relevant in vitro model that addresses focused questions regarding the role of individual perturbations should include combinations of cellular hypoxia and inflammatory stimuli. During development of this model, we consistently observed that hypoxia alone is unlikely to be a relevant signal for induction of ICAM-1 and, as we have recently demonstrated, does not induce endothelial E-selectin (36). Such findings are consistent with previous reports by others in which short-term (minutes) or relatively long-term (hours) exposure of endothelium to hypoxia resulted in no change in surface expression of ICAM-1 (21, 26, 35). The presence of LPS during hypoxia elicited a dose- and time-dependent augmentation of surface ICAM-1 (Figs. 1-3). Moreover, such enhancement by hypoxia was verified at the functional level using lymphocyte adhesion assays (Figs. 4 and 5), and the addition of functionally inhibitable MAb or ICAM-1 mRNA antisense oligonucleotides reduced such adhesion to nearly baseline levels. Interestingly, we consistently observed some enhanced ICAM-1-independent adhesion associated with endothelia exposed to hypoxia (see Figs. 4 and 5). The source of such residual adhesion is not known, although others have described similar ICAM-1-independent, LFA-1-dependent adhesion using endothelial cells exposed to hypoxia (14). Further work will be necessary to reveal the identity of this residual lymphocyte adhesion.

These studies demonstrate that hypoxia augments proteasome activation in the presence of LPS. First, whereas the calpain inhibitor I (ALLN) reduced LPS-activated ICAM-1 in both normoxic and hypoxic endothelia, hypoxic cells showed a greater than one log increase in sensitivity to this inhibitor (Fig. 7). Second, because proteasome activation is required for induction of ICAM-1 (9) and cytoplasmic-to-nuclear translocation of NF-B p65 directly reflects proteasome activity (25), Western blot analysis revealed that hypoxia significantly enhanced the LPS-stimulated nuclear p65 (Fig. 6). The mechanisms that elicit increased proteasome activation by hypoxia are not known at this time, but evidence is provided that metabolic acidification (reflected as a decrease in both extracellular and intracellular pH in our system) may play a role (Fig. 8). Namely, decreased medium and intracellular pH by hypoxia correlated with increases in LPS-stimulated ICAM-1, and such responses were normalized by increasing the buffering capacity of the extracellular medium; a similar ICAM-1 induction pattern was observed by adjusting medium pH in the presence of LPS (see RESULTS). Our results are reflective of recent ex vivo (16, 31) and in vitro (15) studies indicating a distinct role for both metabolic acidosis and proteasome activation during sepsis. Metabolic acidosis, as reflected by circulating lactate levels, has in fact been shown to be an excellent clinical indicator for outcome of septic patients (3, 4). In our model, it is possible that conditions of metabolic acidification activate proteasomes and result in persistent decreases in cytoplasmic I-B, since conditions that maintain low levels of cytoplasmic I-B- have been recently demonstrated to sustain increased surface expression of ICAM-1, vascular cell adhesion molecule 1 (CD106), and E-selectin (CD62E) (17).

Our recent studies demonstrated that endothelial exposure to hypoxia mediates increased induction of transcription and translation of E-selectin (36). Such observations correlated with decreased endothelial generation of adenosine 3',5'-cyclic monophosphate (cAMP) under conditions of hypoxia, and this effect is likely mediated by the cAMP-responsive element/activating transcription factor (CRE/ATF) within the E-selectin gene (10). It is unlikely that similar mechanisms explain our present observations with ICAM-1, since no CRE/ATF binding domain has been demonstrated in the ICAM-1 gene (9) and previous observations have shown that agents that influence intracellular cAMP levels modulate surface expression of endothelial E-selectin but not ICAM-1 (23).

At present, it is unclear how universal our findings might be with regard to other molecules, additional agonists, and ICAM-1 expression on different cell types. Because E-selectin induction occurs through activation of NF-B (9), it is likely that proteasome activation plays a role in our previous findings with hypoxia and E-selectin (36). Overall, these findings indicate that tissue hypoxia serves as a relevant pathophysiological condition during inflammation. Therapeutic strategies aimed at minimizing tissue hypoxia, and by association metabolic acidosis, may dampen such responses.