Expression of Tumor Necrosis Factor-α in Cultured Human Endothelial Cells Stimulated With Lipopolysaccharide or Interleukin-1α
Abstract—Tumor-necrosis factor-α (TNF-α) is a proinflammatory cytokine with a wide variety of biological effects. The most important source of this cytokine is monocytes/macrophages. It is a potent agonist in the activation of endothelial cells; however, the precise role of endothelial cells as a source of TNF-α is not known. In the present study, we addressed the possibility that TNF-α is produced by cultured human umbilical vein endothelial cells (HUVEC) stimulated with factors such as lipopolysaccharide (LPS) or interleukin-1α (IL-1α). LPS and IL-1α induced expression of TNF-α mRNA in HUVEC. IL-1α induced expression and secretion of TNF-α protein, but LPS did not induce production of TNF-α protein. Most of the TNF-α protein in cell lysate was found in the membrane fraction. The mRNA for TNF-α–converting enzyme (TACE) was expressed in unstimulated HUVEC, and its level was not altered by treatment with LPS or IL-1α. Transfection of HUVEC with full-length cDNA encoding the precursor TNF-α enhanced secretion of TNF-α protein by these cells, and treatment of the cells with a TACE inhibitor reduced the secretion. These results suggest that HUVEC produce TNF-α and have TACE activity. Secreted TNF-α may be involved in autocrine activation of endothelial cells, and TNF-α retained in cell membrane may serve as a juxtacrine system to activate target cells on the endothelial surface.
- Received May 17, 1999.
- Accepted July 30, 1999.
Tumor necrosis factor-α (TNF-α) is a cytokine that contributes to a variety of inflammatory responses.1 The human TNF-α gene is located on chromosome 6,2 and the most important source of this cytokine is monocytes/macrophages.1 When cells are stimulated with appropriate agonists, mRNA for TNF-α is induced and a membrane-bound precursor protein with a relative molecular mass of 26 K is produced. This precursor protein is processed by a membrane-bound metalloproteinase, TNF-α-converting enzyme (TACE),3 4 to generate secreted 17-K mature TNF-α.
TNF-α is a potent agonist for vascular endothelial cell activation, along with bacterial lipopolysaccharide (LPS) and interleukin-1 (IL-1). The vascular endothelial cells play an important role in the process of inflammatory responses.5 6 7 When endothelial cells are stimulated with TNF-α, LPS, or IL-1, blood cell tethering molecules such as E-selectin, intercellular adhesion molecule-1 (ICAM-1), and vascular cell adhesion molecule-1,8 are expressed along with signaling factors such as interleukin-8 and epithelial neutrophil activating peptide-78.9
Characteristics of endothelial cells as a target for TNF-α have been well described. On the other hand, the role of endothelial cells as a source of TNF-α is not known. Therefore, we have conducted the present study to examine whether endothelial cells stimulated with LPS or IL-1α may express TNF-α.
Collagenase and cycloheximide (CHX) were obtained from Wako. Cell culture medium Humedia EB-2 and its supplements were obtained from Kurabo. Recombinant human IL-1α, the ELISA kit for TNF-α, and monoclonal anti-human TNF-α antibody were obtained from R&D Systems. Anti-mouse IgG labeled with biotin was obtained from Zymed Laboratories and streptavidin-FITC from Biomeda. LPS from Escherichia coli serotype 0III:B4, human serum albumin, and protease inhibitor cocktail were from Sigma. M199, DMEM, FBS, primer oligo(dT)12–18, and M-Mulv reverse-transcriptase were from Gibco-BRL. Ficoll-Paque Plus was from Amersham Pharmacia Biotech. RNeasy total RNA isolation kit, Taq DNA polymerase, plasmid kit, and Superfect transfection reagent were from Qiagen. Plasmid vector pcDNA3 and a chloramphenicol acetyl transferase (CAT) construct in pcDNA3 were from InVitrogen. Oligotex-dT30<Super> was from Takara. Positively charged nylon membranes, digoxigenin (DIG) RNA labeling kit, DIG nucleic acid detection kit, DIG-labeled actin RNA probe, and CAT ELISA kit were from Boeringer Mannheim. Northern Max kit and Lig’nScribe kit were from Ambion. HO-NH-CO-CH2-CH(CH2-CH[CH3]2)-CO-Nal-Ala-NH-CH2-CH2-NH2, a TACE inhibitor, was from Peptides International. Superblock blocking buffer in PBS was from Pierce. Oligonucleotide primers for polymerase chain reaction (PCR) were synthesized by the University of Utah DNA/peptide user facility or Greiner Japan.
Human umbilical vein endothelial cells (HUVEC) were isolated by use of collagenase and cultured in 6-well plates as described,10 with slight modifications. Cells were cultured in Humedia EB-2 supplemented with 2% FBS, 10 ng/mL recombinant human epidermal growth factor, 1 μg/mL hydrocortisone, 5 ng/mL recombinant human basic fibroblast growth factor, and 10 μg/mL heparin. When cells reached ≈80% confluence, the medium containing growth factors was removed and cells were washed twice. Then the cells were cultured in Humedia EB-2 supplemented with 20% human serum (complete medium). Tightly confluent monolayers of first to third passages were used for the experiments. The primary culture showed <1% CD45+ cells, but no CD45+ cells were found after first passage. Peripheral venous blood was drawn from a healthy volunteer, and monocytes were isolated with Ficoll-Paque Plus according to Böyum.11 Monocytes were plated in 35-mm culture dishes and cultured in DMEM supplemented with 10% FBS to differentiate into macrophages. Macrophages cultured for 3 days were used for the experiments.
RNA Extraction, Reverse-Transcription-PCR, and Northern Blot
HUVEC were incubated in complete medium containing LPS or IL-1α for the indicated times. In the experiments with CHX, cells were pretreated with 500 ng/mL CHX for 1 hour before the addition of LPS or IL-1α. Total RNA was isolated from the cells by use of an RNeasy total RNA isolation kit. Single-strand cDNA for a PCR template was synthesized from 1 μg of total RNA by use of primer oligo(dT)12–18 and M-Mulv reverse-transcriptase under conditions indicated by the manufacturer. Specific primers were designed from cDNA sequences for TNF-α, TACE, and GAPDH, and each cDNA was amplified by PCR with Taq DNA polymerase. Sequences of the primers were as follows:
Conditions for reactions were 1× (94°C, 1 minute); 30× (94°C, 1 minute; 55°C, 1 minute; and 72°C, 1 minute); and 1× (72°C, 10 minutes). Products were analyzed on a 1.5% agarose gel that contained ethidium bromide. Expected size for the PCR products for TNF-α, TACE, and GAPDH were 850, 508, and 598 bp, respectively. Because all of these primer pairs were designed from different exons, the products with the expected size were amplified from single-strand cDNA but not from contaminating genomic DNA. PCR products were confirmed to be specific for TNF-α, TACE, and GAPDH by sequencing.
For Northern blot, HUVEC were incubated with 10 μg/mL LPS for 1 hour or with 1 ng/mL IL-1α for 8 hours and macrophages were incubated with 10 μg/mL LPS for 2 hours. Total RNA was extracted as described above, and poly(A)+RNA was isolated from total RNA using Oligotex-dT30<Super>. Poly(A)+RNA, 560 ng/lane for macrophages and 1885 ng/lane for HUVEC, was analyzed by electrophoresis on a 1% agarose gel containing formaldehyde. RNA was blotted to a positively charged nylon membrane by capillary transfer and probed with the DIG-labeled antisense RNA for TNF-α or β-actin. Synthesis of a DIG-labeled probe and its detection were performed according to specifications of the supplier. A T7 promoter adaptor was ligated to a 850-bp PCR product specific for TNF-α by use of a Lig’nScribe kit, and this was used as a template for the synthesis of a DIG-labeled RNA-probe. Hybridization was performed at 68°C for 16 hours with 0.5 nmol/L probes with a Northern Max kit.
ELISA for TNF-α
For the measurement of TNF-α concentration in the medium or cell lysate, HUVEC were stimulated with 10 μg/mL LPS for ≤8 hours or with 1 ng/mL IL-1α for ≤24 hours. After incubation, conditioned medium was collected. Cells were washed twice with cold PBS, pH 7.4, and lysed with the cell lysis buffer (PBS that contained 1% NP40, 0.5% sodium deoxycholate, 0.1% SDS, and a 0.01% protease inhibitor cocktail). After cells were passed through a 23-gauge needle, cell debris was pelleted by centrifugation and the supernatant was collected. Concentrations of TNF-α in medium and cell lysate were determined by use of ELISA.
To determine whether part of the TNF-α protein found in cell lysate exists in the membrane fraction, HUVEC were scraped in PBS that contained a 0.01% protease inhibitor cocktail after incubation with 1 ng/mL IL-1α for 16 hours. After it was sonicated, the cell homogenate was centrifuged at 10 500g for 60 minutes. The supernatant was designated as the cytoplasm fraction. The insoluble pellet was lysed in the lysis buffer described above. Lysate was briefly sonicated and designated as the membrane fraction. Concentration of TNF-α protein in each fraction was measured by ELISA.
Transfection of a TNF-α Construct
A full-length human TNF-α cDNA clone in mammalian expression vector pcDNA3 was isolated from a cDNA library constructed from LPS-stimulated HUVEC. Plasmid was prepared with a Qiagen plasmid kit. When HUVEC reached 80% confluence in a 35-mm dish, 1 μg/well of the TNF-α plasmid and 0.2 μg/well of the CAT plasmid were cotransfected by use of a Superfect transfection reagent according to the supplier’s protocol. On the next day of transfection, cells were washed twice and incubated in M199-HSA containing 50 μmol/L HO-NH-CO-CH2-CH(CH2-CH[CH3]2)-CO-Nal-Ala-NH-CH2-CH2-NH2 (a TACE inhibitor) or vehicle (0.5% DMSO) for an additional 8 hours. Medium was collected and cells were lysed as described above. Concentrations of TNF-α in the medium and cell lysate were measured by ELISA. Cell lysate was also subjected to CAT ELISA. Value of TNF-α expression was normalized with CAT protein expression.
Immunofluorescence Staining for TNF-α in HUVEC
HUVEC were stimulated with 1 ng/mL IL-1α for 16 hours. Cells were fixed either with 5% paraformaldehyde or ethanol/methanol (1:1) and subjected to immunofluorescent staining for TNF-α as described previously.12 Cells were incubated with Superblock and then with a 1:100 dilution of monoclonal anti–TNF-α antibody. After they were washed with PBS, cells were incubated with anti-mouse IgG labeled with biotin followed by incubation with streptavidin-FITC. Cells were examined by laser confocal microscope (LSM 410; Carl Zeiss).
Expression of mRNAs for TNF-α and TACE in HUVEC
In unstimulated HUVEC, mRNA for TNF-α was not detected by RT-PCR or Northern blot analysis. Expression of TNF-α mRNA was induced by treatment of HUVEC with LPS or IL-1α (Figures 1 through 4⇓⇓⇓⇓). Expression of TNF-α mRNA reached maximal level 1 hour after stimulation with LPS and significantly decreased after 8 hours (Figure 1A⇓). On the other hand, in cells stimulated with IL-1α, expression of TNF-α mRNA reached maximal level 8 hours after stimulation and decreased after 16 to 24 hours (Figure 1B⇓). Levels of mRNA depended on concentration of the agonists, and maximal stimulation was observed at 10 μg/mL for LPS and 1 ng/mL for IL-1α (Figures 2A⇓ and 2B⇓). Pretreatment of cells with CHX enhanced the accumulation of TNF-α mRNA induced by these agonists (Figure 3⇓). Results of Northern blot analysis for TNF-α mRNA after LPS and IL-1α stimulation of cells are shown in Figure 4⇓. Size of TNF-α mRNA from HUVEC (1.7 kb) was the same as that from macrophages used as a positive control. The level of the TNF-α mRNA expression was much lower in HUVEC than in macrophages.
Expression of TNF-α Protein by HUVEC
TNF-α protein levels both in the medium and cell lysate were enhanced by stimulation with 1 ng/mL IL-1α and reached maximal level 16 hours after stimulation (Figures 5⇓ and 6A⇓). In HUVEC stimulated with LPS, accumulation of TNF-α protein in medium and cell lysate was not clear (Figures 5⇓ and 6A⇓). Next, we examined whether TNF-α protein found in the IL-1α–treated cells exists in the membrane fraction. Eighty-eight percent of the cell-associated TNF-α protein was found in the membrane fraction (10 500g precipitate) and 12% in the cytoplasm fraction (Figure 6B⇓).
Expression of TNF-α Protein in HUVEC Transfected With TNF-α Construct
A small amount of TNF-α protein was detected both in the medium and cell lysate from HUVEC transfected with TNF-α construct, and TACE inhibitor reduced the fraction secreted into the medium. The transfected cells produced 110.0 pg/106 cells of secreted TNF-α and 56.0 pg/106 cells of cell-associated TNF-α (n=3). Treatment of cells with a TACE inhibitor resulted in a decrease in TNF-α secretion of 73% (30.6 pg/106 cells) with a concomitant increase in cell-associated protein of 53% (87.2 pg/106 cells) (n=3). Values of TNF-α expression, shown in Figure 7⇓, were normalized with CAT protein expression to assess transfection efficiency.
Immunofluorescent Staining for TNF-α in HUVEC
Results of immunofluorescent staining for TNF-α are shown in Figure 8⇓. HUVEC stimulated with IL-1α were positively stained for TNF-α; however, the distribution of the fluorescence was different for the sample fixed with paraformaldehyde versus that fixed with ethanol/methanol. Fluorescence was observed in a diffuse pattern when cells were fixed with paraformaldehyde, whereas cells fixed with ethanol/methanol had fluorescence with perinuclear granular distribution.
TNF-α is produced by various cell types, including basophils,13 B lymphocytes,14 astrocytes,15 and keratinocytes,16 but the primary source of this cytokine is activated monocytes/macrophages.1 As to the production of TNF-α by endothelial cells, cross-linking of resting HUVEC with anti–E-selectin and anti–ICAM-1 antibodies is reported to trigger the release of TNF-α.17 18 High levels of TNF-α have been found in endothelial cells of human atheroma when an immunohistochemical method was used.19 However, only limited information is available about the production of TNF-α by endothelial cells.
In the present study, we examined whether HUVEC stimulated with LPS or IL-1α express TNF-α. mRNA for TNF-α was not detected in unstimulated HUVEC. When cells were stimulated with LPS or IL-1α, TNF-α mRNA was induced in a dose-dependent manner (Figure 2⇑). We found that the accumulation of TNF-α mRNA was induced over different time courses for these 2 agonists. Accumulation reached maximal level 1 hour after simulation with LPS but 8 hours after stimulation with IL-1α. This result suggests that expression of TNF-α mRNA in response to these agonists is, at least in part, differentially regulated. We next examined whether the induction of TNF-α mRNA expression was due to the direct effect of these agonists or secondary effects that requires de novo protein synthesis. To examine this, the protein synthesis inhibitor CHX was used. When cells were pretreated with CHX, the accumulation of TNF-α mRNA induced by LPS or IL-1α was enhanced. This suggests that both agonists directly stimulate accumulation of TNF-α mRNA. Although we have not examined the difference between these agonists in the signal transduction pathways, this difference will be clarified in the future studies.
RT-PCR analysis is a “semi-quantitative” method and is not effective for determination of the size of the mRNA. Therefore, we also performed Northern blot analysis and found that stimulated HUVEC express TNF-α mRNA at the same size as do stimulated macrophages. However, the amount of TNF-α mRNA expressed by HUVEC was much smaller than that by macrophages.
IL-1α induced the expression of TNF-α protein in HUVEC, but the induction of TNF-α protein by LPS was not observed, although LPS is one of the most potent agonists for TNF-α production in other cell types. The low level of TNF-α production in HUVEC could be due to generation of a transcriptional product, different from the macrophage transcript, that is not translated into the secreted mature protein. This type of an untranslatable transcript has been described for pro-opiomelanocortin.20 However, the size of TNF-α mRNA expressed in HUVEC was the same as that in macrophages, and HUVEC can generate the normal transcript that is translated into mature TNF-α.
We also examined the intracellular localization of TNF-α protein in cells stimulated with IL-1α. Approximately 90% of the TNF-α protein was detected in the membrane fraction. This result suggests that the TNF-α expressed on the cell membranes may act in a juxtacrine fashion to activate target cells on the endothelial surface. The transmembrane form of TNF-α is reported to be superior to soluble TNF-α in T-cell activation, thymocyte proliferation, and production of granulocyte-macrophage colony-stimulating factor.21 If TNF-α works as a juxtacrine system on the plasma membrane of the endothelial cells, the local concentration may be high enough to play an important physiological role, even if its production is low.
Results of immunofluorescent staining were shown in Figure 8⇑. Fluorescence was observed over the cell surface in a diffuse pattern when the cells were fixed with paraformaldehyde. This suggests that TNF-α may be anchored in the membrane or may be processed and bound back to the cell surface. In the cells fixed and permeabilized by the treatment with ethanol/methanol, faint fluorescence was observed with perinuclear distribution, which suggests loss of cell surface antigens with membrane delipidation.
Next we examined whether HUVEC are able to translate TNF-α mRNA by transfection experiments. When TNF-α plasmid was transfected into HUVEC, significant amount of TNF-α protein was detected in both cell lysate and conditioned medium. This result shows that HUVEC can produce TNF-α protein if high levels of mRNA are expressed.
The precursor protein of TNF-α is processed by TACE3 4 to generate the secreted 17-K mature TNF-α. As to the expression of TACE in endothelial cells, Black et al3 showed that TACE protein is expressed in endothelial cells as well as monocytes, T cells, neutrophils, and smooth muscle cells. No information exists about the expression of mRNA and activity for TACE in endothelial cells. We found that TACE mRNA is expressed by unstimulated HUVEC, and the level of its expression was not altered by treatment with LPS or IL-1α. We also examined whether HUVEC have TACE activity by measuring TNF-α secretion from cells transfected with a full-length TNF-α construct. If cells do not have TACE activity, the precursor form expressed in cell lysate will not be processed and the mature form will not be secreted into the supernatant. We found that a substantial amount of TNF-α protein was secreted from HUVEC transfected with the TNF-α construct, and release of TNF-α was decreased by a TACE inhibitor. These results suggest that HUVEC have activity of TACE and an effective mechanism for secretion of the mature protein once the precursor was produced.
We conclude that TNF-α protein is produced by endothelial cells stimulated with IL-1α, but not with LPS. Although the level of its expression is much lower than that of macrophages, TNF-α expressed by endothelial cells may act in a juxtacrine fashion. The small part of TNF-α secreted may also be involved in the autocrine activation of endothelial cells. TNF-α may mediate, in part, the inflammatory responses in endothelial cells elicited by LPS or IL-1α.
The authors thank Drs Stephen M. Prescott, Thomas M. McIntyre, Guy A. Zimmerman, and Julie M. Kessel of the University of Utah for critical discussions and technical advice; Drs Hisako Fujimori and Yoshiaki Fujimori for help in collection of umbilical cords; and Kumiko Munakata for excellent technical assistance. The initial experiments were done with endothelial cells supplied by a Special Center of Research in ARDS (P50HL50153), and a part of this study was supported by the Karoji Memorial Fund for Medical Research at Hirosaki University.
Black RA, Rauch CT, Kozlosky CJ, Peschon JJ, Slack JL, Wolfson MF, Castner BJ, Stocking KL, Reddy P, Srinivasan S, Nelson N, Boiani N, Schooley KA, Gerhart M, Davis R, Fitzner JN, Johnson RS, Paxton RJ, March CJ, Cerretti DP. A metalloproteinase disintegrin that releases tumour-necrosis factor-α from cells. Nature. 1997;385:729–733.
Moss ML, Catherine Jin SL, Milla ME, Burkhart W, Luke Carter H, Chen WJ, Clay WC, Didsbury JR, Hassler D, Hoffman CR, Kost TA, Lambert MH, Leesnitzer MA, McCauley P, McGeehan G, Mitchell J, Moyer M, Pahel G, Rocque W, Overton LK, Schoenen F, Seaton T, Su JL, Warner J, Willard D, Becherer JD. Cloning of a disintegrin metalloproteinase that processes precursor tumour-necrosis factor-α. Nature. 1997;385:733–736.
Modur V, Zimmerman GA, Prescott SM, McIntyre TM. Endothelial cell inflammatory responses to tumor-necrosis factor-α: ceramide-dependent and -independent mitogen-activated protein kinase cascades. J Biol Chem. 1996;71:13094–13102.
Böyum A. Isolation of mononuclear cells and granulocytes from human blood. Scand J Clin Lab Invest. 1968;21(suppl 97):77–89.
Sung SSJ, Jung LKL, Walters JA, Chen W, Wang CY, Fu SM. Production of tumor necrosis factor/cachectin by human B cell lines and tonsillar B cells. J Exp Med. 1988;168:1539–1551.
Köck A, Schwarz T, Kirnbauer R, Urbanski A, Perry P, Ansel JC, Luger TA. Human keratinocytes are a source for tumor necrosis factor-α: evidence for synthesis and release upon stimulation with endotoxin or ultraviolet light. J Exp Med. 1990;172:1609–1614.
Schmid EF, Binder K, Grell M, Scheurich P, Pfizenmaier K. Both tumor necrosis factor receptors, TNFR60 and TNFR80, are involved in signaling endothelial tissue factor expression by juxtacrine tumor necrosis factor α. Blood. 1995;86:1836–1841.