INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

SYSTEMIC INFLAMMATORY RESPONSE SYNDROME (SIRS) is a severe, often life-threatening, consequence of infections and trauma (33). A critical manifestation of SIRS is a generalized vascular leak caused, in part, by disruption of endothelial barrier function (28). The occurrence of such a leak in the pulmonary vasculature underlies adult respiratory distress syndrome (ARDS), an important cause of mortality in critically ill patients (35). The pulmonary vascular endothelium is uniquely situated at the interface between circulating blood and alveoli and presents an essential barrier against the hydrostatic forces driving intravascular fluid into the alveoli. Once thought of as an inert barrier, the vascular endothelium is now recognized to play an active role in 1) generating immunoregulatory cytokines, 2) regulating extravasation of circulating macromolecules, and 3) recruiting and activating inflammatory cells (9).

The proinflammatory cytokine tumor necrosis factor (TNF)- plays a central role in the pathogenesis of ARDS (18, 30), in part, through the mobilization and activation of neutrophils (25) and by inducing pulmonary vascular endothelial cell (EC) injury and dysfunction (18). In experimental human E. coli endotoxemia, TNF- levels in hepatic venous blood were twofold higher than those in systemic arterial blood, suggesting that the liver is a major source of circulating TNF- (15). In mice treated with E. coli endotoxin lipopolysaccharide (LPS), the increase in plasma TNF- is almost eliminated by depleting hepatic Kupffer cells before LPS challenge (24). These studies suggest that hepatic Kupffer cells are the major source of circulating TNF- in individuals with sepsis. Because the pulmonary vasculature is positioned immediately downstream of the hepatic venous effluent, it is exposed to the highest concentrations of Kupffer cell-generated TNF-.

Approximately 90% of patients with sepsis are febrile (1, 5, 10). Fever has the potential to both benefit and harm septic individuals. When the core temperature of rabbits infected with Streptococcus pneumoniae is increased 1.5°C by external warming, their survival time is reduced from 56 to 47 h, despite a fivefold decrease in bacteremia (11). In a Klebsiella pneumoniae peritonitis mouse model, survival is improved, and bacterial load is reduced when core temperature is maintained in the febrile (39.2-39.7°C) rather than the basal (36.5°-37.5°C) range (22). However, the tissue pathogen load at death is considerably lower in the warmer mice, suggesting that death occurs despite successful pathogen clearance in the warmer animals. In mice challenged with a sublethal dose of LPS, early TNF- secretion by Kupffer cells is increased, and circulating TNF- peaks earlier and at 2.7-fold higher levels when the animal's core temperature is maintained at 40°C rather than 37°C (24). Thus the pulmonary vasculature appears to be exposed to higher concentrations of TNF- when sepsis is accompanied by fever.

The influence of fever on TNF--mediated lung injury might also be mediated by its effects on TNF- activity in the lungs. Warming L929 fibroblasts to 40.5°C during treatment with TNF- enhances DNA fragmentation and cell death compared with cells incubated at 37°C with TNF- (37). Treating HT-29 colon cancer cells with TNF- at 42°C rather than 37°C enhances cytotoxicity (27). It is important to note that the biological response to temperatures in the febrile range and those in the heat-shock range (>42°C) are distinct. For example, our laboratory showed that the major stress-induced transcription factor, heat shock factor (HSF)-1 is fully activated in macrophages exposed to 43°C but only partially activated in macrophages exposed to 39.5°C (36). How exposure to febrile range temperatures modifies the sensitivity of EC to TNF--induced cytotoxicity or activation is not well understood.

The objective of this study was to determine whether exposing human pulmonary artery EC to febrile range temperatures modifies TNF--induced changes in EC viability, adhesiveness for neutrophils, barrier function, or cytokine secretion. We found that incubating primary cultured human pulmonary artery EC at 39.5°C rather than 37°C enhanced TNF--induced increases in permeability, reduced generation of interleukin (IL)-6 while increasing expression of granulocyte-macrophage colony stimulating factor (GM-CSF) and IL-8, and induced submaximal expression of heat shock protein (HSP)70.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Reagents. Recombinant human TNF- was generously provided by Knoll Pharmaceutical/BASF K & K (Whippany, NJ). The specific activity was determined to be 2.5 U/ng by using a standard L929 bioassay, as previously described (13). cDNAs for human IL-6, IL-8, GM-CSF, glyceraldehyde phosphate dehydrogenase (GAPDH), and hamster HSP72 were provided by Dr. Lester May (Rockefeller University, New York, NY), Dr. Joost Oppenheim (National Cancer Institute, Frederick, MD), Dr. Stephen Clark (Genetics Institute, Cambridge, MA), Dr. Mitch Olman (University of Alabama Birmingham, Birmingham, AL), and Dr. A. Fornace (National Cancer Institute, Bethesda, MD).

EC culture. Human pulmonary artery EC (Clonetics, San Diego, CA) were cultured at 37°C under 5% CO₂ in endothelial basal medium supplemented with 5% fetal bovine serum (FBS), hydrocortisone, human fibroblast growth factor-B with heparin, vascular endothelial growth factor, R3-insulin growth factor-1, ascorbic acid, human epidermal growth factor, gentamicin sulfate, and amphotericin-B, as previously described (7). Only cells from passages 2-7 were studied.

Assay of transendothelial albumin flux. Transendothelial [¹⁴C]BSA flux was assayed as described previously (4). EC (1 × 10⁵ cells per chamber) were cultured at 37°C for 72 h on gelatin-impregnated polycarbonate filters (13-mm diameter, 0.4-mm pore size) (Nucleopore, Pleasanton, CA) mounted in polystyrene chemotactic chambers (ADAPS, Dedham, MA) inserted into the wells of 24-well plates. The baseline barrier function of each monolayer was determined by applying an equivalent and reproducible amount of [¹⁴C]BSA tracer (1.1 pmol/0.5 ml; Sigma, St. Louis, MO), to each upper compartment for 1 h at 37°C, after which the lower compartment was counted for carbon-14 activity. Only EC monolayers retaining 97% of the tracer were used for experiments. Less than 5% of EC monolayers failed to meet this criterion. The monolayers were then treated and again assayed for transendothelial [¹⁴C]BSA flux.

Polymorphonuclear leukocyte preparation and fluorescent labeling. Whole peripheral blood from healthy human volunteers was collected into acid citrate dextran (Sigma) solution, and polymorphonuclear leukocytes (PMNs) were isolated by dextran erythrocyte sedimentation and density gradient centrifugation through Fcoll-Hypaque (Sigma), as previously described (19). PMNs were fluorescently labeled using a modification of the procedure described by Akeson and Woods (3). Briefly, PMNs were resuspended in Hanks balanced salt solution without divalent cations (HBSS) (Life Technologies, Gaithersburg, MD) at 1 × 10⁷ PMNs/ml and were incubated with 5 µM calcein AM (Molecular Probes, Eugene, OR) in the dark for 40 min with gentle agitation (3). PMNs were washed three times with HBSS, after which their purity was >95% and viability was >98% by trypan dye exclusion.

Assay of PMN adhesion. EC were seeded into the wells of 24-well culture plates (1 × 10⁵ EC/well; Costar, Cambridge, MA) in 1 ml of medium and cultured to confluence (72 h, 37°C, 5% CO₂). EC monolayers were treated, gently washed with PBS, and co-incubated with calcein AM-labeled PMNs (5 × 10⁵ PMNs/well) for 30 min. After gentle washing to remove nonadherent PMNs, the attached PMNs were fluorometrically assayed (excitation 485 nm, emission 530 nm) in a Cytofluor II multi-well fluorescence plate reader (PerSeptive Biosystems, Framingham, MA). For each experiment, serial dilutions of labeled PMNs (from 1 × 10³ to 1 × 10⁵ cells/ml) were used to generate a standard curve from which PMN numbers could be interpolated from fluorescence units. PMN-EC adhesion was expressed as a percentage of PMN adherence.

Assay of PMN transendothelial migration. EC cultured to confluence on gelatin-impregnated polycarbonate filters (13- mm diameter, 3-µm pore size) mounted in polystyrene chemotactic chambers were inserted into the wells of 24-well plates. To establish functional integrity for each monolayer, baseline barrier function was assayed as above. Only EC monolayers retaining 97% of the [¹⁴C]BSA tracer were studied. The EC monolayers were treated and then inserted into wells containing f-Met-Leu-Phe (1 × 10⁹ M). Calcein AM-labeled PMNs (5 × 10⁵ cells/well) were introduced into the upper compartments of assay chambers and incubated for 2 h at 37°C, followed by sampling and fluorometric assay of the lower compartment. After transendothelial migration (TEM) through the EC monolayers, >99% of fluorescence remained PMN associated (data not shown). A standard curve was established for each experiment, from which PMN numbers could be interpolated from fluorescence units. PMN TEM was expressed as a percentage of PMN migration.

Measurement of TNF--induced cytokine secretion. EC cultured to confluence in the wells of 24-well plates were treated, gently washed with PBS, and incubated in the presence or absence of human TNF- for 2-24 h. The concentration of cytokines in the culture supernatants was analyzed in the UMAB Cytokine Core Laboratory using standard two-antibody ELISA with commercial antibody pairs and recombinant standards [IL-1, IL-6, and GM-CSF from Endogen, Boston, MA; IL-8 from Biosource, Camarillo, CA, and transforming growth factor (TGF)-₁ from R&D Systems, Minneapolis, MN]. Before the TGF-₁ assay, cell culture supernatants were activated by incubating with 0.2 volumes 1 N HCl at room temperature for 10 min, then neutralized by adding 0.2 volumes 1.2 N NaOH/0.5 M HEPES. Polystyrene plates (Maxisorb, Nunc) were coated with capture antibody in PBS overnight at 25°C. The plates were washed four times with 50 mM Tris, 0.2% Tween 20 (pH 7.2) and then blocked for 90 min at 25°C with assay buffer (PBS containing 4% BSA and 0.0 1% Thimerosal, pH 7.2). The plates were washed and 50 µl of assay buffer were added to each well with 50 µl of test sample or cytokine standard in assay buffer and incubated at 37°C for 2 h. The plates were washed and incubated with 100 µl strepavidin-peroxidase polymer in casein buffer (Research Diagnostics, Mount Pleasant, NJ) (0.5 h, 25°C), followed by 100 µl substrate (TMB, Dako, Carpentaria, CA) for 20-30 min. The reaction was stopped with 100 µl 2 N HCl, and the A450 (A650) was read on a microplate reader (Molecular Devices, Sunnyvale, CA). The data was analyzed using a computer program (SoftPro, Molecular Devices). The IL-1, IL-6, IL-8, GM-CSF, and TGF-₁ assays had lower detection limits of 0.8, 6, 4, 9, and 30 pg/ml, respectively.

Northern blot analysis. EC monolayers cultured in T75 culture flasks were exposed to 250 U TNF-/ml or medium alone for 3 or 6 h at 37° or 39.5°C. The monolayers were washed with PBS at 4°C, and total RNA was extracted by using the acid phenol extraction method of Chomczynski and Sacchi (8). RNAgents kits were purchased from Promega, and acid phenol extraction was performed according to the manufacturer's instructions. Electrophoresis of total RNA (25 µg/lane) was performed in denaturing 1.2% formaldehyde agarose gels, as previously described (12). RNA was transferred to nitrocellulose by capillary action and cross linked to the membrane by ultraviolet irradiation (Stratalinker 2400, Stratagene). ³²P-labeled cDNA probes were prepared from isolated restriction fragments using reagents supplied by Pharmacia and according to the random primer method of Feinberg and Vogelstein (14). Equal loading was confirmed by ethidium bromide staining and by stripping and reprobing the blot with the cDNA for the housekeeping gene GAPDH, as previously described (12). Hybridization was performed at 45°C for 18 h in buffer containing 50% formamide, 5X SSC, 0.08% Ficoll, 0.08% polyvinylpyrrolidine, 0.08% BSA, 0.1% SDS, 0.1% sodium pyrophosphate, 100 µg/ml denatured salmon sperm DNA, and 20 mM sodium phosphate, pH 6.5. The filters were first washed for 30 min in 2 × SSC with 0.1% SDS at room temperature and then subjected to four 30-min washes at 50°C in 0.1% SSC with 0.1% SDS. The activity of each band was quantified by phosphorimaging (Molecular Dynamics, Sunnyvale, CA), and autoradiography was done using Kodak XAR film.

Cell viability. EC were seeded (5 × 10⁴ cells/well) in 96-well culture plates and incubated at 37° or 39.5°C in the presence or absence of rhTNF- for up to 48 h. Viability was quantified by measuring the reduction of 3-(4,5-dimethylthiazol-2-yl)-5-(3-carboxymethoxyphenyl)-2-(4-sulfophenyl)- 2H-tetrazolium (MTS; Promega) to a formazan product during a 2-h incubation at 37°C. Results were reported as A₄₉₀ (optical density at 490 nm).

Statistical analysis. All data are presented as means ± SE. Differences between two groups were tested using an unpaired Student's t-test. Differences among more than two groups were tested by a Fisher's protected least-squares difference (PLSD) test applied to a one-way ANOVA.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Influence of culture temperature on endothelial barrier function. Mean (±SE) pretreatment baseline BSA flux across EC monolayers was 0.008 ± 0.001 pmol/h (n = 104) and was 0.201 ± 0.006 pmol/h (n = 5) across naked filters without monolayers. After 6 h at 37°C, rhTNF-, at concentrations 250 U/ml, increased BSA flux compared with the simultaneous 37°C medium controls (0.035 ± 0.006 vs. 0.021 ± 0.004 pmol/h; n = 13) (Fig. 1). Albumin flux across EC monolayers incubated at 39.5°C for 6 h in the absence of TNF- (0.027 ± 0.004 pmol/h) was not significantly different from flux across 37°C control monolayers. Treating EC with only 25 U/ml TNF- at 39.5°C increased BSA flux (0.042 ± 0.005 pmol/h) to 122% of that attained in the 37°C monolayers treated with 250 U/ml TNF-. The response to 2.5 U/ml TNF- administered at 39.5°C (0.031 ± 0.005 pmol/h) was comparable to 250 U/ml TNF- administered at 37°C, but the difference compared with the TNF--free 39.5°C EC cultures failed to reach statistical significance (P = 0.053).

View larger version (17K): [in this window] [in a new window]   Fig. 1.   Influence of incubation temperature on endothelial cell (EC) barrier function. Human pulmonary artery EC were cultured for 72 h on gelatin-impregnated polycarbonate filters, mounted in polystyrene chemotactic chambers, and inserted into the wells of 24-well plates, and the baseline barrier function of each monolayer was determined by measuring the transendothelial flux of [14C]BSA for 1 h at 37°C. The monolayers were then incubated for 6 h at either 37° or 39.5°C with the indicated concentration of recombinant human tumor necrosis factor (TNF)-, transendothelial [14C]BSA flux was assayed again, and the increase from baseline flux was determined for each monolayer. Data are means ± SE for 13 experiments. * P < 0.04 and  P < 0.002 compared with EC incubated at 37°C without TNF-;  P < 0.05 compared with 37°C EC treated with the same concentration of TNF-.

Influence of temperature on neutrophil adhesion to and migration across EC monolayers. At 37°C, a 6-h preincubation of EC monolayers with TNF- at 2.5 U/ml increased their adhesiveness for fluoroprobe-labeled neutrophils three-fold compared with that observed for TNF--free control monolayers (Fig. 2A). At 39.5°C, neutrophil adhesion to monolayers preexposed to medium or 25 U/ml TNF- was comparable to that observed in 37°C TNF--treated cells. At 250 U/ml TNF-, neutrophil adhesion to 39.5°C monolayers was slightly reduced compared with the 37°C monolayers (31.5 ± 3.3 vs. 37.9 ± 3.9%). To study the effects of temperature and TNF- on neutrophil transendothelial migration (TEM), the percentage of fluoroprobe-labeled neutrophils traversing EC monolayers over 2 h was quantified (Fig. 2B). At 37°C, TNF- at 2.5 U/ml increased TEM of neutrophils twofold compared with medium controls without further dose-dependent increments. In the absence of TNF-, incubating EC at 39.5°C increased TEM by 30% compared with the 37°C medium controls (11.1 ± 0.7 vs. 8.5 ± 0.5%; n = 12); however, exposure to the higher temperature did not modify neutrophil TEM in the TNF--treated monolayers.

View larger version (13K): [in this window] [in a new window]   Fig. 2.   Influence of temperature on neutrophil adhesion to and migration across EC monolayers. A: to measure neutrophil adherence, confluent EC monolayers were treated for 6 h with the indicated concentration of recombinant TNF- and culture temperatures, then washed, and coincubated with calcein AM-labeled polymorphonuclear cells (PMNs) for 30 min at 37°C. Nonadherent PMNs were removed by washing, and adherent neutrophils were quantified by measuring the fluorescence of each monolayer (excitation 485 nm, emission 530 nm). The number of neutrophils adhering to each monolayer was determined from a standard curve and was expressed as the percentage of total PMNs added. B: to measure transendothelial migration of PMNs, confluent EC monolayers were established on gelatin-impregnated polycarbonate filters and mounted in chemotactic chambers. After 6-h treatment of the EC monolayers at the indicated temperature and TNF- concentration, calcein AM-labeled PMNs were added to the upper chamber and 1 × 109 M f-Met-Leu-Phe to the lower chamber, and the number of PMNs that migrated into the lower chamber after 2-h incubation at 37°C was quantified fluorometrically and expressed as percentage of total PMNs added. n = 12 experiments. * P < 0.05 compared with 37°C EC treated with the same concentration of TNF-.

Influence of temperature on EC cytokine expression. Culture supernatants from control EC monolayers incubated for 24 h at 37°C contained little IL-6, GM-CSF, or IL-8. Treating with 250 U/ml TNF- stimulated secretion of each of these cytokines (Fig. 3). In contrast, TGF-₁ was constitutively released by untreated EC and was not enhanced by stimulation with TNF-. Secretion of IL-6 was comparable in 37° and 39.5°C EC cultures for 10 h after addition of TNF-, but, by 24 h, the IL-6 levels were 30% lower in the 39.5°C cultures. In contrast, levels of GM-CSF and IL-8 were increased by 23 and 35%, respectively in the 39.5°C EC 24 h culture supernatants. To determine whether the lower levels of IL-6 in the 39.5°C culture supernatants were caused by decreased stability of cytokine protein at the higher temperature, we sequentially measured the concentration of exogenous recombinant IL-6, GM-CSF, and IL-8 in cell-free culture medium during 24-h incubation at 37° or 39.5°C. There was no detectable decrease in concentration of any of the three cytokines during incubation at either temperature (data not shown).

View larger version (24K): [in this window] [in a new window]   Fig. 3.   Influence of temperature on EC cytokine secretion. EC cultured to confluence were incubated for 24 h at the indicated temperature and TNF- concentration, and the cytokine levels in the culture supernatants were measured using two-antibody ELISA assays. Data are means ± SE; n = 6 experiments. IL, interleukin; GM-CSF, granulocyte-macrophage colony stimulating factor; TGF-, transforming growth factor-.* P < 0.05 compared with 37°C EC treated with the same concentration of TNF-.

View larger version (71K): [in this window] [in a new window]   Fig. 4.   Influence of temperature on EC levels of cytokine and heat shock protein (HSP)72 mRNA. EC cultured to confluence were incubated for 6 h at the indicated temperature in the presence or absence of 250 U/ml TNF-. Total RNA was isolated, and levels of cytokine (A) and HSP72 (B) mRNA levels were analyzed on separate Northern blots. Each blot was stripped and reprobed for the housekeeping gene glyceraldehyde phosphate dehydrogenase (GAPDH). In B, a positive heat shock control (lane 9) was included by incubating EC at 43°C for 30 min and then for an additional 2.5 h at 37°C.

Effects of TNF- and 39.5°C exposure on HSP72 expression and EC viability. HSP72 mRNA levels were low in 37°C EC cultures and were not increased after 6-h incubation with 250 U/ml TNF- at 37°C. HSP72 mRNA accumulated during 6-h incubation at 39.5°C, reaching levels that were one-third of those attained in EC heat-shocked at 43°C for 30 min (compare lanes 3 and 7 with 9). TNF- did not further modify HSP72 expression in the 39.5°C cells. To determine whether exposure to TNF- or to elevated temperatures affected EC proliferation or viability, EC were seeded at low density and cultured for 48 h in the absence or presence of 2.5, 25, or 250 U/ml TNF- at 37° or 39.5°C. The reduction of MTS during a subsequent 2-h incubation at 37°C was determined spectrophotometrically (Table 1). In the absence of exogenous TNF-, EC proliferation and viability were comparable in 37° and 39.5°C cultures. Treating EC with 2.5 U/ml TNF- at 37°C decreased MTS reduction by 12%, but neither higher concentrations of TNF- nor incubation at 39.5°C caused further increases in MTS reduction.

View this table: [in this window] [in a new window]   Table 1.   Influence of incubation temperature and TNF- treatment on EC proliferation and viability

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The increase in body temperature during fever regulates expression of the proinflammatory cytokine TNF- (12, 13, 22-24), restricting TNF- generation by hepatic Kupffer cells to a brief, early burst (22-24). The present study extends these observations by demonstrating that the response of a critical target cell, the pulmonary vascular EC, to TNF- is also modified by exposure to febrile temperatures. We previously reported that treating bovine pulmonary artery EC with recombinant human TNF- induces actin depolymerization and increased paracellular permeability, as reflected by flux of [¹⁴C]BSA tracer (17). In the present study, we showed that rhTNF- causes similar disruption in barrier function of postconfluent human pulmonary artery EC and that this effect was augmented at 39.5°C. However, in the absence of exogenous TNF-, EC monolayers maintained similar barrier function after 6-h incubations at 37° and 39.5°C. This suggests that exposure to the higher temperature does not directly alter EC barrier function but, rather, modifies the EC response to the TNF- stimulus.

We previously showed that TNF- treatment induces barrier dysfunction in bovine pulmonary artery EC monolayers at 37°C in the absence of detectable cytotoxicity (17). However, the combination of TNF- treatment followed by heat shock (42°C) caused apoptosis in transformed murine microvascular EC (2). To determine whether the increment in TNF--induced barrier dysfunction in human pulmonary artery EC cultured at 39.5°C is caused by enhanced cytotoxicity, cell viability was analyzed by measuring the reduction of MTS, a reaction that reflects the number of metabolically active cells. In preliminary experiments, the relationship between MTS reduction and the number of viable EC added to culture wells was linear (data not shown). This analysis demonstrated that even after 48 h of exposure to TNF-, EC cytotoxicity is low and is not enhanced at 39.5°C. These data confirm and extend our previous findings (17) to show that neither TNF--induced EC barrier dysfunction at 37°C nor the increment in barrier dysfunction at 39.5°C is caused by EC cytotoxicity. These data are consistent with previous studies that show that sheep pulmonary artery EC (41) and human EC (40) tolerate simultaneous exposure to heat shock and inflammatory stimuli.

In contrast to the effect of warming on TNF--induced barrier dysfunction, exposing EC to 39.5°C had relatively little effect on the capacity of TNF--treated EC for neutrophil binding and TEM. Neutrophil-EC interactions require molecular processes that are very different from those required for the opening of the endothelial paracellular pathway. Whereas barrier dysfunction requires actin disassembly (17) and modification of proteins (20) within the multiprotein complex at EC-EC adherens junctions, neutrophil binding to EC requires de novo synthesis and EC surface expression of adhesion molecules (e.g., E-selectin, ICAM-1) (32). Prior protein synthesis inhibition blocks TNF--induced surface expression of these counter-ligands (32) but not TNF--induced loss of barrier function (17). In the 37°C EC, TNF- influenced neutrophil binding and TEM at a 100-fold lower concentration than that required to induce barrier dysfunction, providing additional evidence that these effects are mediated through distinct pathways that differ in TNF- sensitivity and temperature dependence. Exposing EC to 39.5°C in the absence of TNF- modestly enhanced neutrophil TEM. Previous studies have shown that increasing microenvironmental temperature from 37° to 40°C directly increases neutrophil chemokinesis (31). In the present study, the EC were exposed to 39.5°C, but the binding and TEM assays were performed at 37°C, thereby avoiding the direct effects of higher temperature on neutrophil motility. Neutrophil binding to EC monolayers was comparable at 37° and 39.5°C, suggesting that the exposure of EC to the higher temperature enhanced TEM by modifying postbinding events. Gnant et al. (16) reported that incubating human umbilical vein EC (HUVEC) at 39°C increases expression of platelet/endothelial cell adhesion molecule-1 (PECAM-1), a protein required for neutrophil TEM in the lung (38). However, Gnant et al. (16) also reported that incubating HUVEC at 39°C enhances TNF--induced expression of ICAM-1, a process that might cause qualitative changes in neutrophil/EC binding that could not be detected by our binding assay.

The endothelium is a rich source of the cytokines responsible for the local recruitment and activation of leukocytes. In this study, we analyzed the influence of exposure to 39.5°C on the constitutive and TNF--induced generation and secretion of three such cytokines, as well as the counterregulatory mediator, TGF-₁. We found three patterns of temperature and TNF- responses. EC biosynthesis of IL-6 was not detectable in untreated EC, increased after TNF- treatment at 37°C, and was modestly suppressed at 39.5°C. EC expression of GM-CSF and IL-8 were also induced by TNF- treatment, but expression of these cytokines was higher in 39.5°C than in the corresponding 37°C EC cultures. TGF-₁ was constitutively expressed, and its expression was not influenced by either TNF- treatment or temperature. We demonstrated that IL-6, IL-8, and GM-CSF proteins were each comparably stable at 37° and 39.5°C, excluding enhanced degradation as the cause of the reduced IL-6 levels in the 39.5°C culture supernatants. Gnant et al. (16) reported that TNF--treated HUVEC incubated at 39°C for 3 h and then returned to 37°C released higher levels of IL-6 and IL-8 in the 24-h culture supernatants than EC incubated continuously at 37°C. In our study, EC were incubated at 37° or 39.5°C for 24 h. Whereas IL-8 secretion increased in agreement with the Gnant et al. study (16), the IL-6 secretion rate fell but only after >10 h of continuous exposure to the warmer temperature. In a previous study of human macrophages, conducted in our laboratory, incubating cells at 40°C for 18 h failed to reduce IL-6 mRNA or secreted protein levels (13), demonstrating that the effects of exposure to increased temperature are both gene- and cell-dependent. The changes in IL-6 and IL-8 secretion occurred despite little associated change in mRNA levels, suggesting that this effect was mediated through translational or posttranslational modulation rather than effects on transcription. Our laboratory reported that submaximal heat shock is induced and global translation is decreased when human macrophages are cultured at 40°C (13). This is consistent with reports from other groups that demonstrate that heat shock can be induced by febrile range temperatures (6) and that heat shock is associated with reduced efficiency of cap-dependent translation initiation (34). The reduced translational efficiency during heat shock is attributed to an enhanced interaction between the cap-binding protein eIF4E and its inhibitor 4E-BP1 (39). In the present study, EC HSP70 mRNA accumulated during 6 h of incubation at 39.5°C, albeit to lower levels than in cells exposed to 43°C for 30 min. This demonstrates that submaximal heat shock is induced in EC after exposure to 39.5°C. Therefore, the observed inhibition of late IL-6 translation in the 39.5°C EC might be part of a generalized heat shock-induced reduction in cap-dependent protein synthesis. However, the persistence of IL-8 and GM-CSF expression in 39.5°C EC cultures indicates that the effects of warming on cytokine expression are gene-specific. If this is true, then it may be IL-8 and GM-CSF that are uniquely regulated in the warmer cells. EC might maintain IL-8 and GM-CSF expression at 39.5°C by increasing mRNA levels to compensate for the loss of translational efficiency or by maintaining translational efficiency of the transcripts at the same levels as in 37°C EC. Northern analysis showed a 2.25-fold increase in GM-CSF mRNA levels in the 39.5°C EC cultures. Laroia et al. (26) reported that heat-shock treatment (44°C) of human umbilical vein EC causes increase stability of a chimeric mRNA containing the GM-CSF AU-rich element. Although this raises the possibility that warming EC stabilized GM-CSF mRNA, it does not exclude the possibility that the GM-CSF transcription rate might also be increased in the warmer cells. The increase in IL-8 secretion with little change in mRNA levels in the 39.5°C EC suggests that IL-8 translation might be enhanced rather than reduced in the warmer cells. However, examination of the primary sequences of IL-6, IL-8, and GM-CSF does not reveal an obvious mechanism to explain potential differences in their temperature-dependent translational efficiencies. Like IL-6 and GM-CSF, IL-8 contains a complex 5' UTR architecture and AU-rich sequences within its 3' untranslated regions that can contribute to transcript stability and may regulate translational efficiency (21). To our knowledge, there are no known internal ribosome entry sites (29) in the IL-8 5' UTR that might allow cap-independent translation in the face of heat shock-induced eIF4E inhibition. The mechanisms responsible for the differential effects of febrile temperature on EC cytokine expression are presently under investigation in our laboratory.

In summary, we have shown that exposing pulmonary artery EC to temperatures within the febrile range increases neutrophil TEM, enhances TNF--induced barrier dysfunction, and alters their profile of cytokine expression. TNF- is central to the pulmonary vascular EC dysfunction in acute injury. Beacuse fever usually accompanies the acute phase response to such injuries and can change both TNF- expression and target tissue responsiveness, it might be crucial to the development of ARDS.