© Copyright 2004 by Humana Press Inc. All rights of any nature whatsoever reserved. 1085-9195/04/40/289–303/$25.00

ORIGINAL ARTICLE

Comparative Analysis of Vascular Endothelial Cell Activation by TNF-α and LPS in Humans and Baboons Qiang Shi,1 Jian Wang,1,3 Xing Li Wang,1,3,* and John L. VandeBerg1,2 1Department

of Genetics and 2Southwest National Primate Research Center, Southwest Foundation for Biomedical Research, San Antonio, 3Michael E. DeBakey Department of Surgery, Baylor College of Medicine, Houston, TX

Abstract As an Old World nonhuman primate, baboons have been extensively used for research on dyslipidemia and atherogenesis. With increasing knowledge about the endothelium’s role in the initiation and progression of atherosclerosis, the value of the baboon model can be increased by developing it for research on the role of dysfunctional endothelium in atherogenesis. Toward that goal, we have established and validated methods of isolating and culturing baboon femoral artery endothelial cells (BFAECs) and compared baboon endothelial cellular characteristics with those of humans. Our results indicated that baboon and human endothelial cells share similar growth and culture behaviors. As was the case for human endothelial cells, BFAECs responded to tumor necrosis factor (TNF)-α stimulation with increased expression of adhesion molecules (maximum increase for intracellular adhesion molecule (ICAM): 1.76 ± 0.26-fold; vascular cell adhesion molecule (VCAM): 1.65 ± 0.25-fold; E-selectin: 2.86 ± 0.57-fold). However, BFAECs were hyporesponsive to lipopolysaccharide (LPS) (range, 0.25–20 µg/mL) in adhesion molecule expression, whereas 1 µg/mL LPS induced 2.14- to 3.71-fold increases in human endothelial cells. The differential responses to LPS were not related to TLR-2 and toll-like receptor (TLR)-4 expression on the cell surface. And baboon microvascular endothelial cells had similar features as BFAECs. We observed constitutive expression of interleukin (IL)-6, IL-8, granulocyte macrophage colony-stimulating factor (GM-CSF), and monocyte chemoattractant protein (MCP)-1 in both human and baboon endothelial cells, and these cytokines were further induced by TNF-α and LPS. We also demonstrated that the responses to TNF-α or LPS varied among baboons maintained under the same dietary and environmental conditions, suggesting that response may be controlled by genetic factors. Index Entries: Baboon; endothelial cells; adhesion molecules; cytokines; femoral artery.

*Author to whom all correspondence and reprint requests should be addressed. E-mail: [email protected].

Cell Biochemistry and Biophysics

289

Volume 40, 2004

290

INTRODUCTION As the interior lining of the vascular wall, the endothelium plays a crucial role in the initiation and progression of atherosclerosis. Numerous predisposing factors contribute to the diverse and complex processes of atherosclerosis, most of which ultimately converge to a common outcome: dysfunctional endothelium (1,2). The perturbation of endothelium is associated with a series of changes (3,4). Endothelial cell activation by factors such as tumor necrosis factor (TNF)-α, interleukin (IL)-β, interferon-γ, or lipopolysaccharide (LPS) forms an inflammatory process in the surrounding area. Dysfunctional endothelium loses vascular integrity, expresses surface adhesion molecules (CAMs), converts from antithrombotic to prothrombotic status, and overproduces cytokines (5). Cell adhesion molecules, such as ICAM-1, VCAM-1, and E-selectin, participate in the recruitment of leukocytes (5–7). Endothelial cell activation also results in a significant release of cytokines, including IL-6, IL-8, MCP-1, and GM-CSF, which facilitate monocytes to penetrate, reside, mature, proliferate and function in sub-endothelial space (8,9). The baboon is an important animal model for research on arteriosclerosis. It has a close phylogenetic relationship with humans and exhibits similar pathogenesis such as aortic fatty streaks and fibrous plaques that occur in human subjects (10). Investigators at the Southwest Foundation for Biomedical Research/Southwest National Primate Center have developed well-characterized pedigrees of baboons over several decades. A genetic linkage map of the baboons also has been developed, and the extended families have been studied extensively for genetic determinants of blood pressure and of dietary effects on plasma lipoproteins in relation to atherogenesis (11–13). However, little is known about the roles of endothelial cells in response to atherogenic lipoprotein profiles or other risk factors of atherosclerosis in baboons. Our long-term objective is to establish the baboon as a practical primate model for research on endothelial function in relation to

Cell Biochemistry and Biophysics

Shi et al. atherogenesis. In particular, we want to identify genes that may affect individual responses of endothelial cells to atherogenic stimuli in baboons. The aims of the present study were to establish and validate the methods of isolating and culturing endothelial cells from large vessels of baboons, to compare their cellular characteristics with those of endothelial cells isolated from human umbilical vein and artery vessels, and to determine the extent of variability among baboons in endothelial cell activation capacity.

MATERIALS AND METHODS Baboons The baboons used for this study were among the nearly 4000 maintained by the Southwest National Primate Research Center based at the Southwest Foundation for Biomedical Research. They were maintained in outdoor group caging and fed chow ad libitium (Harlan Teklad SWF Primate Diet, Madison, WI), supplemented with seeds and corn.

Baboon Endothelial Cell Collection and Culture Segments of femoral arteries approximately 1 to 3 cm in length were collected from baboons by sterile methods. Baboon femoral artery endothelial cells (BFAECs) were harvested no longer than 2 h after the vessels were excised. Cell isolation was based on previous reports with modifications (14–19). The blood vessel was placed in a large sterile dish, and the surface was wiped with 2% antibiotic/antimycotic solution (GIBCO-BRL, Grand Island, NY) in phosphate-buffered saline (PBS). Then the vessel was gently cannulated at one end with a short blunt needle and flushed with PBS to remove any blood. It was then injected with 0.1% collagenase before the other end was cannulated. The vessel was incubated at 37˚C for 10 to 15 min, depending on the size of artery. After completion of digestion, the vessel was massaged gently, flushed with medium, and Volume 40, 2004

Vascular Endothelial Cell Activation the released cells were collected by centrifugation and resuspended in medium. The cells were seeded immediately on 1.0% gelatin coated culture plates. The medium was F-12K supplemented with 20% FCS (GIBCO-BRL), 75µg/mL EGCS (Sigma, St. Louis, MO), 50µg/mL heparin (Sigma), 10 mM HEPES (GIBCO-BRL), 2 mM glutamine (GIBCO-BRL), and antibiotics (GIBCO-BRL). Cells were maintained in a 5% CO2 incubator at 37˚C until confluence when they were passaged by 0.05% trypsin and Versene solution (GIBCO-BRL). The isolation of baboon microvascular endothelial cells from baboon abdominal fat tissues was carried out by using DynaBeads CD31 (DYNAL, Oslo, Norway), according to the instructions of the manufacturer. The culture conditions for microvascular endothelial cells were the same as for BFAECs.

Human Umbilical Vein and Aortic Endothelial Cells Human umbilical vein endothelial cells (HUVECs) were isolated from three umbilical cords by similar procedures (20). Human aortic endothelial cells (HAECs) were purchased from CELL Application Inc. (San Diego, CA). The medium for HUVEC and HAEC culture was exactly the same as that used for BFAEC culture.

Identification of Isolated Endothelial Cells To confirm their endothelial cell origin, 200 µL of cell suspension was seeded on Lab-Tek

culture chambers (Nunc, NY) coated with 1.0% gelatin. Upon 80% to 90% confluence, cells were fixed in 2% formalin in PBS at room temperature for 20 min. After they were air-dried, cells were blocked with 10% normal serum from the same species in which the secondary antibody was raised. Cells were then incubated with primary antibodies at 4˚C overnight. Anti-human vWF (Sigma) at a dilution of 1:400 was used to detect von Willbrand factor (vWF) in both HUVECs and BFAECs. The primary antibody was detected by FITC-labeled secondary antibody (Santa Cruz, CA) and images were taken under a Nikon Eclipse E800 microscope. We

Cell Biochemistry and Biophysics

291 also demonstrated the absence of smooth muscle α-actin (Sigma) to confirm that smooth muscle cells did not contaminate the cultures. Direct immunofluorescent staining for anti-human smooth muscle–specific α-actin antibody conjugated with FITC was employed at a dilution of 1:200 (Sigma, cat. no. F 2016 Anti-goat IgG FITC) for 1-h incubation at room temperature.

Uptake of Dil-LDL by Endothelial Cells Cells were seeded in Lab-Tek culture chambers to 80% to 90% confluence. We incubated cells for 4 h at 37˚C in growth medium containing 10 mg/mL Dil-Ac-LDL (BTI, MA) without supplementing serum, but with 5% calf albumin. The slides were fixed in 2% formalin for 20 min at room temperature and mounted with SlowFade (Molecular Probes, OR).

Quantification of CAMs on Endothelial Cell Surface Our protocol is based on previous reports (14–19). Cells at the third or fourth passage were seeded in 100µL at a density of 50,000/mL in 96-well plates. After 24 h, the cells were exposed to 10 ng/mL TNF-α (Sigma) and 1 µg/mL LPS (Sigma) for 4 and 24 h. We fixed the cells with 2% formalin at room temperature for 20 min and then blocked with 2% bovine serum albumin in PBS containing 1% H2O2 and 0.05% (w/v) Tween-20 in a 37˚C water bath for 1 h. The plates were incubated at 4˚C overnight with goat polyclonal antibodies against human E-selectin (R&D Systems, MN, cat. no. BBA19), VCAM-1 (R&D Systems, MN, cat. no. BBA19) and ICAM-1 (R&D Systems, MN, cat. no. BBA17) at dilutions of 1:500. After incubation with secondary antibody conjugated with peroxidase (anti-goat IgG Fab2, Sigma, cat. no. A5420) at 1:5000 in a 37˚C water bath for 1 h, oPhenylenediamine dihydrochloride (Sigma, cat. no. P 9187) was used as a substrate. The solution was made according to manufacturer’s method, color was allowed to develop at room temperature for 20 min, and optical density was read at 492 nm. All experiments had blank controls of medium and cells with-

Volume 40, 2004

292 out treatment. Experiments were repeated at least three times.

Measurement of TLR-4 and TLR-2 by Enzyme Immunoassay (EIA) Confluent cells were seeded on 96-well plates and fixed with 2% formalin in PBS for 20 min at room temperature. The plates were washed with 0.05% Tween-PBS three times and blocked by 2% bovine serum albumin for 1 h at 37˚C. Polyclonal primary antibody from rabbit against TLR-4 was from eBioscience (www. ebioscience.com) at a dilution of 2 µg/mL; and polyclonal antibody from rabbit against TLR-2 was from Activemotif (www.activemotif.com) at a dilution of 2 µg/mL. Anti-rabbit HRPlinked immunoglobulin (Ig)G at 1:1000 dilutions was from Cell Signaling (Waltham, MA). The substrate for HRP was o-Phenylenediamine dihydrochloride from Sigma, and optical density was read at 492 nm.

Cytokine Production After Activation in Endothelial Cells For quantification of cytokine production, we inoculated the cells at the third or fourth passage into six-well plates (Corning) at a density of 5 × 104/mL in 1 mL. After the cells were treated for 24 h, we collected the supernatants and centrifuged them at 3000g at 4˚C for 20 min. The samples were stored at –70˚C before they were assayed. Commercial enzyme-linked immunosorbent assay kits from R&D Systems were used for the measurement of MCP-1. We used the Human Cytokine Lincoplex kit (LINCO Research Inc, MO) to measure GM-CSF, IL-1β, IL-4, IL-6, IL-8, IL-12, and interferon (IFN)-γ concentrations. Twenty-five microliters of supernatant were incubated with multiplexed labeled beads according to the manufacturer’s directions. Bead signals were analyzed on a Luminex 100TS instrument.

Statistical Analysis Data are expressed as the mean ± standard deviation. Statistical comparison of means was

Cell Biochemistry and Biophysics

Shi et al. performed by two-tailed unpaired student’s t-tests. The null hypothesis was rejected at p < 0.05. Analysis of variance model was used for multiple comparisons.

RESULTS Properties of Isolated Endothelial Cells From Baboons We isolated endothelial cells from femoral arteries of 16 baboons. Nearly all of the cultures had formed discrete colonies by 24 to 72 h after inoculation. Normally 3 d to 1 wk later, depending on the number of harvested cells, the cells appeared as a confluent monolayer, demonstrating typical cobblestone morphology. We compared the plating efficiencies and growth curves among human umbilical vein, human aortic, and baboon femoral artery endothelial cells and found no significant differences among the three cell types (data not shown).

Characterization of BFAEC and Microvascular Endothelial Cells (mEC) We characterized all cultures for the presence of vWF staining and absence of α-actin staining as markers to check the purity of cell isolation. Cells from 14 of the 16 isolations were 98% to 100% vWF-positive (Fig. 1, A1) and 100% α-actin–negative staining. Among 14 endothelial cell cultures, we randomly selected 3 to confirm their endothelium origin further by measuring their functional activities. Two parameters were used. One was endothelial expression of E-selectin when stimulated by TNF-α treatment. The other was uptake fluorescence Dil-labeled low-density lipoproteins (Dil-LDL) by endothelial cells, which have scavenger receptors for acetylated LDL. As shown in Fig. 1(B1,B2), baboon endothelial cells and HUVECs exhibited very strong fluorescence of E-selectin on their cell surfaces after stimulation by 10 ng/mL TNF-α. Although there were differential degrees of fluorescence intensities among E-selectin positive cells,

Volume 40, 2004

Fig. 1. Immunohistochemical characteristics and Dil-LDL uptake of endothelial cells from human and baboon. (A) von Willebrand factor expression in Weibel-Palade bodies in baboon femoral artery endothelial cells (BFAECs) (A1) and human umbilical vein endothelial cells (HUVECs) (A2). (B) E-selectin expression elicited by 10 ng/mL tumor necrosis factor-α treatment for 4 h in BFAECs (B1) and HUVECs (B2) by immunofluorescent staining. (C) Dil-LDL uptake by in BFAECs (C1) and HUVECs (C2). Results represent 3 isolations from human umbilical veins and 14 isolations of endothelial cells from baboon femoral arteries. Images were taken with a Nikon Eclipse 800 microscope at ×600 magnification.

294 baboon and human umbilical vein endothelial cells stained at comparable levels. Figure 1 also demonstrated endocytosis of Dil-LDL in baboon cells (C1) and HUVECs (C2). To determine whether prolonged culture in vitro alters cellular reaction to stimuli, we used the same cells and measured their responses to 10 ng/mL TNF-α, 10 ng/mL IL-β1, and 1.0 µg/mL LPS at the third, fourth, fifth, and seventh passages. There were no significant deviations among third, fourth, and fifth passages in relation to cellular adhesion molecule expression (Fig. 2). However, there were increases of ICAM-1 and VCAM-1 expression at the seventh passage in response to 10 ng/mL TNF-α and 10 ng/mL IL-β1, but not to 1.0 µg/mL LPS.

Expression of Adhesion Molecules by Endothelial Cells in Culture We observed increased expression of CAMs in response to TNF-α in dosage- and cell number-dependent ways. We seeded the BFAECs and HUVECs in 96-well plates at the same conditions and exposed them to TNF-α for 4 and 24 h. Figure 3 shows that increases in TNF-α concentration (A) and in cell number (B) resulted in elevated expression of cellular adhesion molecules in BFAECs and HUVECs after 4 h of exposure. After 24 h of exposure, the cells exhibited similar response patterns (data not shown). Our preliminary experiments showed that E-selectin was induced at its peak within 4 to 6 h after activation by stimuli such as TNF-α and LPS, and declined to baseline after 24 h, despite the continuous presence of stimuli. VCAM-1 reached its highest level at 12 h and then declined to a lower level at 24 h in the continuous presence of stimuli. ICAM-1 was constitutively expressed in all endothelial cells and increased continuously from 4 to 24 h. For this study, we chose 4 and 24 h of exposure to observe responses of BFAECs to 10 ng/mL TNF-α and 1 µg/mL LPS. As shown in Table 1, TNF-α induced significant expressions of all three adhesion molecules in all three types of endothelial cells, although the peak induction

Cell Biochemistry and Biophysics

Shi et al. varied from molecule to molecule. Whereas LPS induced significantly increased production of adhesion molecules in HAECs, it did not have significant effects on BFAECs or HUVECs. We also observed that similar responses to TNF-α and LPS by use of baboon microvascular endothelial cells (Table 2).

Cytokine Production After Activation by Baboon Endothelial Cells As shown in Table 3, all three types of endothelial cells constitutively express IL-6, IL-8, GM-CSF, and MCP-1. Stimulation by TNF-α or LPS induced significantly increased production of all constitutively expressed cytokines. In contrast to the lack of response of BFAECs to LPS in adhesion molecule production (Table 1), cytokine production by BFAECs was increased in response to LPS (Table 3); however, their responsive degrees were lower than in response to TNF-α stimulation.

Dosage–Response Curves in LPS-Stimulated Baboon Endothelial Cells To confirm that expression of adhesion molecules of baboon endothelial cells is resistant to activation by LPS, we exposed BFAECs, HAECs, and HUVECs to LPS at a range of concentrations from 0.25 µg/mL to 20 µg/mL and measured the expression of CAMs. The results indicated that increased LPS dosage had little effect on expression of CAMs in baboon endothelial cells as compared with human endothelial cells (Fig. 4). It should be noted that a significant proportion of cells died at increased concentrations of LPS, particularly at higher than 20 µg/mL. Prolonged incubation did not alter cell activation.

Measurement of TLR-2 and TLR-4 in BFAEC To explore the possible mechanism for lack of responses to LPS in BFAECs in relation to expression of adhesion molecules, we measured the expression of TLR-2 and TLR-4 on the cell surfaces by enzyme immunoassay. As shown in Table 4, there were no significant dif-

Volume 40, 2004

Fig. 2. Variations of cellular adhesion molecule (CAM) expression in baboon femoral artery endothelial cells among the third to the seventh passages after activation by 10 ng/mL tumor necrosis factor (TNF)-α, 10 ng/mL interleukin-β1, or 1 µg/mL lipopolysaccharides for 4 (A) and 24 (B) h. The y axis represents increase of CAM expression compared with corresponding controls in fold(s). The results represent means and standard error of the means of six measures. The results of analysis of variance were not significant (p < 0.05) for comparisons of the third to fifth passages after activation under the previously discussed conditions in relation to E-selectin, VCAM-1, and ICAM-1 expression for 4 (A) and 24 (B) h. However, E-selectin expression was significantly lower (p < 0.05) at the seventh passage after LPS treatment for 4 h than it was at the third to fifth passages (A); and ICAM-1 and VCAM-1 expression were significantly higher (p < 0.05) at the seventh passage after TNF-α treatment for 24 h than at the third to fifth passages.

296

Shi et al.

Fig. 3. Tumor necrosis factor (TNF)-α dosage-dependent (A) and cell number-dependent (B) induction of cellular adhesion molecule expression of baboon femoral artery endothelial cells (A1, B1) and human umbilical vein endothelial cells (A2, B2) after activated by TNF-α for 4 h. Bars representing TNF-α concentrations at 10, 2, 0.1, 0.01, and 0.0 ng/mL in (A) and seeding densities at 15000, 7500, 3500, 1800, and 900 cells/well in (B). The y axis represents CAM expression in optical density from EIA compared with corresponding controls. The results are data from one experiment but are representative of three experiments conducted with endothelial cells from three baboon subjects. Each experiment was conducted in duplicate. Means of the duplicates of one experiment are shown. Analysis of variance: p < 0.01 for results from TNF-α concentrations at 10, 2, 0.1, and 0.01 ng/mL vs controls in (A); p < 0.01 for different seeding densities at 15,000, 7500, 3500, 1800, and 900 cells/well vs controls in (B).

Cell Biochemistry and Biophysics

Volume 40, 2004

297 1.00±0.03 1.00±0.06 1.00±0.05

1.00±0.06 1.00±0.09 1.00±0.04

0h

24 h

1.67±0.41 2.41±0.86b 1.55±0.32 1.29±0.43 3.71±1.23ab 1.45±0.40

1.91±0.63b 2.86±0.67a 1.53±0.32 b 1.36±0.17 b 3.23±0.43 a,b 1.22±0.32

4h

HAECs (n = 6)

1.00±0.01 1.00±0.04 1.00±0.06

1.00±0.03 1.00±0.01 1.00±0.04

0h

24 h

1.01±0.08 1.05±0.10 1.06±0.13

1.04±0.07 1.10±0.09 0.99±0.06

1.36±0.23a 1.76±0.26a 1.25±0.15 a,b 1.65±0.25 a,b 2.86±0.57 a,b 1.04±0.09

4h

BFAECs (n = 14)

bp

< 0.05 for results from HUVEC, HAEC, and BFAEC exposure to TNF-α or LPS stimulation for 4 or 24 h vs 0 h. < 0.05 for results from HAEC and BFAEC exposure to TNF-α or LPS stimulation for 4 or 24 h vs results from HUVEC exposure. All data are represented as mean ± SD. The numbers listed in the table are expressed as CAM increase in fold(s). The means represent results from three independent experiments conducted with three HUVEC subjects, 1 pooled HAEC culture, and 14 baboon endothelial cell subjects. Each experiment was conducted in duplicate.

ap

1.17±0.09 1.96±0.26a 1.03±0.10 1.04±0.10 1.77±0.23a 1.06±0.62

LPS stimulation ICAM 1.00±0.04 VCAM 1.00±0.08 E-selectin 1.00±0.07

24 h

1.27±0.27 2.45±0.14a 0.96±0.02 1.09±0.71 2.54±0.28 a 1.03±0.82

4h

TNF-α stimulation ICAM 1.00±0.03 VCAM 1.00±0.09 E-selectin 1.00±0.09

0h

HUVECs (n = 6)

Table 1 Changes in Adhesion Molecules When Endothelial Cells Were Stimulated by TNF-α (10 ng/mL) or LPS (1 µg/mL)

298

4h

1.00±0.02 1.02±0.11 1.00±0.06 1.12±0.05 1.00±0.09 1.10±0.06

1.14±0.19 1.26±0.15 1.16±0.08

1.28±0.19 1.37±0.24 1.27±0.11

1.00±0.06 1.00±0.05 1.00±0.06

1.00±0.04 1.00±0.08 1.00±0.01

1.00±0.06 1.00±0.03 1.00±0.07

0h

24 h

0.96±0.03 1.06±0.16 1.01±0.04

0.93±0.02 1.17±0.06 1.04±0.07

1.12±0.11 1.29±0.08 1.28±0.14

1.17±0.11 1.15±0.13 1.31±0.12

1.92±0.12 a 1.74±0.11a 1.70±0.13 a 1.97±0.04a 2.29±0.09 a 1.84±0.08 a

4h

Pooled mEC (n = 6)

bp

< 0.05 for the results from HAEC, pooled BFAEC, and pooled baboon mEC exposure to TNF-α or LPS stimulation at 4 or 24 h vs 0 h. < 0.05 for results from HAEC or BFAEC exposures to TNF-α or LPS stimulation for 4 or 24 h vs results from mEC exposure. All data are represented as mean ± SD. The numbers listed in the table are expressed as CAM increase in fold(s). The means represent results from two independent experiments conducted with pooled HAEC and BFAEC and pooled microvascular endothelial cells. Each experiment was conducted in duplicate.

ap

24 h

1.00±0.02 1.21±0.08 a,b 1.78±0.11a 1.00±0.05 1.23±0.02 a,b 1.60±0.09 a 1.00±0.05 1.93±0.03 a 123±0.07a

0h

10 µg/mL LPS stimulation ICAM 1.00±0.08 1.62±0.11a,b 2.21±0.08 a,b VCAM 1.00±0.04 1.57±0.05 a,b 1.62±0.01a E-selectin 1.00±0.03 3.02±0.06b 2.63±0.06 a,b

2.20±0.04 a 1.50±0.03 a 1.37±0.06 a

24 h

1.00±0.03 1.08±0.02 1.00±0.03 1.06±0.02 1.00±0.09 1.14±0.01

1.78±0.02 a 1.55±0.07 a 2.63±0.01a

4h

Pooled BFAECs (n = 6)

1 µg/mL LPS stimulation ICAM 1.00±0.02 1.75±0.07 a,b 2.12±0.02 a,b VCAM 1.00±0.05 1.57±0.01 a,b 1.42±0.01 a E-selectin 1.00±0.05 3.03±0.03 a,b 1.82±0.06 a

TNF-α stimulation ICAM 1.00±0.03 VCAM 1.00±0.05 E-selectin 1.00±0.02

0h

HAECs (n = 6)

Table 2 Changes in Adhesion Molecules When Endothelial Cells Were Stimulated by TNF-α (10 ng/mL) or LPS (1 and 10 µg/mL)

Vascular Endothelial Cell Activation

299

Table 3 Endothelial Cytokine Production When Stimulated by TNF-α (10 ng/mL) or LPS (1 µg/mL) for 24 h Cytokine levels (mean ± SD), ng/mL HUVEC

HAEC

BFAEC

Basal expression IL-1β IL-4 IL-6 IL-8 IL-12 GM-CSF MCP-1 INF-γ

n=3 ND ND 0.61±0.34 1.28±0.60 ND 0.004±0.005 1.24±0.29 ND

n=3 ND ND 1.00±0.50 2.90±1.27 ND 0.003±0.002 2.65±0.45 ND

n = 14 ND ND 0.04±0.007 21.90±25.31 c,d ND 0.035±0.038 a,b 0.23±0.28 a,b ND

TNF-α stimulation IL-1β IL-4 IL-6 IL-8 IL-12 GM-CSF MCP-1 INF-γ

ND ND 1.57±0.25 7.74±4.88 ND 0.14±0.04 2.99±1.08 ND

ND ND 4.61±2.40 a 30.49±2.6a ND 0.20±0.14 4.02±1.11 ND

ND ND 0.07±0.01 c,d 60.90±24.89 b,c ND 1.41±0.98 3.83±0.84 ND

LPS stimulation IL-1β IL-4 IL-6 IL-8 IL-12 GM-CSF MCP-1 INF-γ TNF-α

ND ND 4.50±1.53 17.77±11.40 ND 0.26±0.03 3.83±1.04 ND ND

ND ND 3.33±1.15 72.62±0.49a ND 0.53±0.21 3.55±1.32 ND ND

ND ND 0.18±0.17 a,b 43.99±37.59 a,b ND 0.18±0.25 1.45±0.91 a,b ND ND

ND, not detectable in culture supernatants by Lincoplex kit. The means represent results from three independent experiments conducted with three HUVEC cultures, 1 pooled HAEC culture, and 14 baboon endothelial cell cultures. Each experiment was conducted in duplicate. ap < 0.05 for results of HAECs and BFAECs vs HUVEC controls. bp < 0.05 for results from HAECs and BFAECs. cp < 0.01 for results of HAECs and BFAECs vs HUVEC controls. dp < 0.01 for results of HAECs vs BFAECs.

Cell Biochemistry and Biophysics

Volume 40, 2004

300

Shi et al.

Fig. 4. Dose–response curves for lipopolysaccharides (LPS)-stimulated human umbilical vein endothelial cells (HUVECs) (diamond), human aortic endothelial cells (HAECs) (square), and baboon femoral artery endothelial cells (triangle). The x axis indicates the LPS concentrations in µg/mL; the y axis represents increase of E-selectin expression after 4 h of exposure (A), of ICAM-1 expression after 24 h (B), and of VCAM-1 after 4 h (C) compared with corresponding controls in fold(s). We selected these different time points as the peak levels for each adhesion molecule. The results represent the means from three independent experiments with the same cells at passages 4 to 5. Each experiment was conducted in duplicate. Analysis of variance: p > 0.05 for results from controls vs LPS concentrations at 0.25, 0.5, 1.0, 2.5, 5.0, 10, and HAECs in (A); p > 0.05 for results from controls vs LPS concentrations at 0.25, 0.5, 1.0, 2.5, 5.0, 10, and 20 µg/mL in both HUVECs and HAECs in (B); p > 0.05 for results from controls vs LPS concentrations at 0.25, 0.5, 1.0, 2.5, 5.0, 10, and 20 µg/mL in HAECs in (B); and p < 0.05, controls vs LPS in HAECs in (C). ferences in the expression of TLR-2 and TLR-4 on human and baboon endothelial cells.

Interindividual Variation Among Baboons in Endothelial Responses to Stimulation We observed major individual differences in cellular activation among 14 endothelial cell cultures derived from femoral artery vessels of 14 baboons. This variation was evident in

Cell Biochemistry and Biophysics

cytokine production (Table 5) and CAM expression (Table 1). Table 5 gives results of the MCP-1 and GM-CSF production ranges of the 14 baboon endothelial cell cultures after activation. All baboon endothelial cells exhibited a wider range of differences in GM-CSF and MCP-1 expression than CAM expression after they were activated by TNF-α and LPS. It should be noted that TNF-α was more potent in stimulating GM-CSF expression than was

Volume 40, 2004

Vascular Endothelial Cell Activation

301

Table 4 TLR-2 and TLR-4 Expression on HAECs and BFAECs

TLR-2 TLR-4

Blank (N = 4)

HAECs (pooled cells)

BFAECs (N = 5)

1.00±0.01 1.00±0.01

1.75±0.01 2.28±0.08

2.01±0.05a 2.24±0.06a

ap > 0.05 for results from HAECs vs BFAECs. All data are represented as mean ± SD. The numbers listed in the table express TLR-2 and TLR-4 increases in fold(s). There were no significant differences in TLR-2 and TLR-4 expression between HAECs and BFAECs.

Table 5 Distributions of Endothelial-Derived GM-CSF and MCP-1 Induced by TNF-α or LPS for 24 h Range (pg/mL)

Mean ± SD, n = 14

GM-CSF Control 10 ng/mL TNF-α 1 µg/mL LPS

14±1–191±231 461±106–3477±450 32±4–909±305

36±38 1411±978a 179±258b

105.6 69.3 144.1

MCP-1 Control 10 ng/mL TNF-α 1 µg/mL LPS

17±0–454±21 3038±619–4513±304 589 ±181–2600±877

227±280 3823±836a 1445±906a,b

123.3 21.9 62.7

ap bp

Coefficient of variation (%)

< 0.05 vs control. < 0.05 TNF-α vs LPS.

LPS in both HAECs and BFAECs. However, MCP-1 was more responsive than GM-CSF to both TNF-α and LPS.

DISCUSSION Baboon endothelial cell dysfunction is intimately related to atherogenesis (1,2,5); the determination of endothelial cell functional profiles in baboons will enhance the value of this species as a model of atherosclerosis in vivo. We have developed a method by which enough endothelial cells (3–10 × 103 cells) can be collected from a small section of femoral artery (1–3 cm) to establish cultures for determining functional profiles. Because only a short length of peripheral artery is required,

Cell Biochemistry and Biophysics

the method is feasible for application to biopsied baboon femoral arteries. Baboons normally reactivate the collateral circulation to compensate the ligated femoral artery within days. There are no long-term side effects from removal of 2 cm of femoral artery from baboon. With our optimized protocol, we are able to harvest pure endothelial cells from most of the harvested arteries. These endothelial cells behave similarly to human endothelial cells during culture. They have a cobblestone shape, express vWF, and uptake Dil-LDL, just as human endothelial cells do. Baboon endothelial cells also produce adhesion molecules constitutively. When stimulated by TNF-α, IL-β1, or LPS, the production of the adhesion molecules increases significantly. The time-dependent responses of baboon endothelial cells to

Volume 40, 2004

302 these cytokines are very similar to those of human venous or arterial endothelial cells, although the degree of response varies (Table 1). We have further demonstrated that the CAM expression profiles are consistent in culture for up to five passages (Fig. 2). This consistent behavior ensures that a sufficient number of endothelial cells can be derived from a small piece of biopsied femoral artery for numerous experiments. Our results have also established that baboon endothelial cells resemble human endothelial cells in constitutively expressing IL-6, IL-8, GM-CSF, and MCP-1 (Table 3). In addition, when the cells are stimulated by TNF-α or LPS, the production of these cytokines is increased dramatically in endothelial cells of both species. These similar behaviors suggest that baboon endothelial cells may also respond in the same way to atherogenic stimuli in vivo as human endothelial cells. With close genetic, physiologic, and anatomic resemblance to humans and easy access for invasive studies, baboons will make an excellent animal model to study endothelial behavior when challenged by other atherogenic risk factors or pharmaceutical agents. However, we did observe some differences in response to endotoxin stimulation. At doses of 1 to 20 µg/mL, LPS did not induce additional CAM production in baboon endothelial cells, whereas even a low dose at 1 µg/mL was sufficient to activate CAM production in human cells. There is no satisfactory explanation why CAM expression in baboons is resistant to LPS. Earlier investigations in vivo by Briscoe and Pober indicated that injection of TNF-α in baboons caused late (24–48 h) T-cell infiltration, whereas injection of LPS did not (21). The ability of TNF, but not LPS, to recruit T cells correlates with the ability of TNF to cause sustained endothelial cell adhesion molecule expression (21). Another observation by Zurovsky indicated that injection of LPS (0.1–100 µg/kg) into baboons did not elicit fever (22). Redl et al. pointed out that the dramatic differences in LPS responses existed in nonhuman primates. Although humans and chimpanzees are sensitive to LPS, baboons

Cell Biochemistry and Biophysics

Shi et al. appear highly resistant (23). This difference offers us a unique tool in dissecting genes or molecules that may be involved in infection- or inflammation-induced vascular disease. On the other hand, cytokine production in baboon endothelial cells does respond to LPS (Table 3), although the degree of response varies among HUVECs, HAECs, and BFAECs. It is well known that LPS mediates a cellular signaling pathway through extracellular TLR molecules in the human and mouse systems. However, we did not see differential expression by either immunofluorescence staining or EIA measurement in vitro. We are currently investigating the involvement of CD14 and other components that might be responsible for the different responses between humans and baboons to LPS challenge. We further observed that that baboon endothelial cells exhibit large interindividual variations in extent of CAM and cytokine expression and in response to cytokine or endotoxin stimulation. The large between-cell differences in cytokine production may largely be regulated by genetic variations in the baboon population. The genetic contribution to CAMs and cytokine expression may be even easier to recognize in our in vitro cell culture model in which environmental factors are uniform than they would be in vivo where circulating factors are variable. Because our method is applicable to biopsied femoral arteries, and it can be applied to our large pedigreed baboon population, we have a unique opportunity to locate individual genes that affect constitutive or response levels of CAM and cytokines through a whole-genome scan using the baboon gene map (24).

ACKNOWLEDGMENTS We thank Drs. K. Dee Carey and Gene Hubbard for their help in collecting baboon artery samples, and Ms. Cathy Jett for her assistance in analyzing the cytokine levels. The project is supported by NIH grants P01 HL28972, P51 RR13986, R01 HL66053

Volume 40, 2004

Vascular Endothelial Cell Activation

REFERENCES 1. 1 Lusis, A. L. (2000) Atherosclerosis. Nature 407, 233–241. 2. Ross, R. (1993) The pathogenesis of atheroscle2 rosis: a perspective for the 1990s. Nature 362, 801–809. 3. 3 Haskard, D. O. and Landis, R. C. (2002) Interactions between leukocytes and endothelial cells in gout: lessons from a self-limiting inflammatory response. Arthritis Res. 4 Suppl 3, S91–S97. 4. Hennig, B., Toborek, M., and McClain, C. J. (2001) 4 High-energy diets, fatty acids and endothelial cell function: implications for atherosclerosis. J. Am. Coll. Nutr. 2001;20(2 Suppl), 97–105. 5. 5 Libby, P. (2002) Inflammation in atherosclerosis. Nature 420, 868–874. 6. 6 Krieglstein, C. F. and Granger, D. N. (2001) Adhesion molecules and their role in vascular disease. Am. J. Hypertens. 14, 44S-54S. 7. 7 Tailor, A. and Granger, D. N. (2000) Role of adhesion molecules in vascular regulation and damage. Curr. Hypertens. Rep. 2, 78–83. 8. 8 Ikeda, U., Matsui, K., Murakami, Y., and Shimada, K. (2002) Monocyte chemoattractant protein-1 and coronary artery disease. Clin. Cardiol. 25, 143–147. 9. Gerszten, R. E., Garcia-Zepeda, E. A., Lim, Y. C., 9 Yoshida, M., Ding, H. A., Andrew, G., et al. (1999) MCP-1 and IL-8 trigger firm adhesion of monocytes to vascular endothelium under flow conditions Nature 398, 718–723. 10. 10 Kushwaha, R. S. and McGill, H. C., Jr. (1998) Diet, plasma lipoproteins and experimental atherosclerosis in baboons (Papio sp.). Hum Reprod Update 4, 420–429. 11. 11 Kammerer, C. M., Cox, L. A., Mahaney, M. C., Rogers, J., and Shade, R. E. (2001) Sodium-lithium countertransport activity is linked to chromosome 5 in baboons. Hypertension 37, 398–402. 12. 12 Rainwater, D. L., Kammerer, C. M., Cox, L. A., Rogers, J., Carey, K. D., Dyke, B., et al. (2002) A major gene influences variation in large HDL particles and their response to diet in baboons. Atherosclerosis 163, 241–248. 13. 13 Cox, L. A., Birnbaum, S., and VandeBerg, J. L. (2002) Identification of candidate genes regulating HDL cholesterol using a chromosomal region expression array. Genome Res. 12, 1693–1702.

Cell Biochemistry and Biophysics

303 14. 14 Amberger, A., Maczek, C., Jurgens, G., Michaelis, D., Schett, G., Trieb, K., et al. (1997) Co-expression of ICAM-1, VCAM-1, ELAM-1 and Hsp60 in human arterial and venous endothelial cells in response to cytokines and oxidized low-density lipoproteins. Cell Stress Chaperones 2, 94–103. 15. Hewett, P. W. and Murray, J. C. Isolation, culture and properties of microvessel endothelium from human breast adipose tissue. In: Endothelial cell culture. (Bicknell, R., ed.). Cambridge Press, New York, 1996, pp. 55–76. 16. 16 Klein, C. L. (1994) Comparative studies on vascular endothelium in vitro. Pathobiology 62, 199–208. 17. 17 Shen, J., Ham, R. G., and Karmiol, S. (1995) Expression of adhesion molecules in cultured human pulmonary microvascular endothelial cells. Microvasc. Res. 50, 360–372. 18. 18 Zhang, F., Yu, W., Hargrove, J. L., Greenspan, P., Dean, R. G., Taylor, E. W., et al. (2002) Inhibition of TNF-alpha induced ICAM-1, VCAM-1 and E-selectin expression by selenium. Atherosclerosis 161, 381–386. 19. 19 Schmidt, A., Goepfert, C., Feitsma, K., and Buddecke, E. (2002) Lovastatin-stimulated superinduction of E-selectin, ICAM-1 and VCAM-1 in TNF-alpha activated human vascular endothelial cells. Atherosclerosis 164, 57–64. 20. Protocols of harvesting endothelial cells. (2003) http://vrd.bwh.harvard.edu/core_facilities/ cell_bio_protocols.html. 21. 21 Briscoe, D. M., Cotran, R. S., and Pober, J. S. (1992) Effects of tumor necrosis factor, lipopolysaccharide, and IL-4 on the expression of vascular cell adhesion molecule-1 in vivo. Correlation with CD3+ T cell infiltration. J. Immunol. 149, 2954–2960. 22. 22 Zurovsky, Y., Laburn, H., Mitchell, D., and MacPhail, A. P. (1987) Responses of baboons to traditionally pyrogenic agents. Can. J. Physiol. Pharmacol. 65, 1402–1407. 23. Redl, H., Bahrami, S., Schlag, G., and Traber, D. L. (1993) Clinical detection of LPS and animal models of endotoxemia. Immunology 187, 330–345. 24. Rogers, J., Mahaney, M. C., Witte, S. M., et al. (2002) A genetic linkage map of the baboon (Papio hamadryas) genome based on human microsatellite polymorphisms. Genomics 67, 237–247.

Volume 40, 2004