- Email: [email protected]
Regulation of bovine E-selectin expression by recombinant tumor necrosis factor alpha and lipopolysaccharide
Veterinary Immunology and Immunopathology 79 (2001) 151±165
Regulation of bovine E-selectin expression by recombinant tumor necrosis factor alpha and lipopolysaccharide Corinne Van Kampen*, B.A. Mallard Department of Pathobiology, University of Guelph, Guelph, Ont., Canada N1G 2W1 Received 5 September 2000; received in revised form 31 January 2001; accepted 11 February 2001
Abstract Induction of adhesion molecules by cytokines and LPS is an important mechanism of regulating leukocyte migration into tissue. Expression and regulation of E-selectin may be differentially in¯uenced by the stimuli involved with effects on mRNA or surface protein kinetics. Surface protein and mRNA expression kinetics of bovine E-selectin were measured and compared in primary cultures of bovine aortic endothelial cells (BAEC) stimulated for various periods of time with recombinant bovine tumor necrosis factor alpha (rbTNF-a) or Escherichia coli lipopolysaccharide (LPS). E-selectin mRNA expression was measured via quantitative reverse transcription polymerase chain reaction (Q-RT-PCR) using a construct that contained multiple synthetic oligonucleotides for several bovine adhesion molecules and cytokines. Surface expression of Eselectin was measured by ¯ow cytometry. Unstimulated BAECs expressed minimum or no Eselectin on the surface. A low number of endothelial cells expressed surface E-selectin as early as 1 h post-stimulation and surface expression was sustained after both stimuli for 24±72 h. Mean ¯uorescence intensity (MFI) indicated peak surface concentration of E-selectin at 6 h poststimulation after LPS followed by a gradual decrease to 72 h without returning to baseline values. Mean ¯uorescence intensity following stimulation with TNF-a increased slightly between 0 and 72 h. The pattern of mRNA expression differed between stimuli. LPS-stimulated BAECs expressed peak amounts of E-selectin mRNA at 6 h, followed by a decline to baseline by 24 h. Conversely, BAECs stimulated with rbTNF-a expressed signi®cantly (p£ 0.05) higher amounts of mRNA at 1 h than compared to unstimulated controls (0 h), but this decreased to below baseline levels by 6 h; followed by a gradual increase and eventually a sharp increase between 18 and 72 h. To account for the lack of correlation between mRNA and protein expression, it was hypothesized that shedding of surface E-selectin accounted at least in part, for the large increase in mRNA expression seen at 18± 72 h. Culture supernatants from rbTNF-a-treated BAECs were harvested, and tested for the presence of shed E-selectin using ELISA. Unstimulated culture supernatants contained little or no *
Corresponding author. Present address: Department of Medicine, McMaster University, 1200 Main St. West, Hsc. Rm. 3N5, Hamilton, Ont., Canada L8N 3Z5. Tel.: 1-905-525-9140/ext. 22212. 0165-2427/01/$ ± see front matter # 2001 Elsevier Science B.V. All rights reserved. PII: S 0 1 6 5 - 2 4 2 7 ( 0 1 ) 0 0 2 4 9 - 5
152
C. Van Kampen, B.A. Mallard / Veterinary Immunology and Immunopathology 79 (2001) 151±165
E-selectin. Between 6 and 48 h, the concentration of E-selectin in culture supernatants from rbTNFa-stimulated BAECs increased approximately two-fold, suggesting that the sharp increase in Eselectin mRNA expression around 18 h may be related to signi®cant loss of surface E-selectin during this period. # 2001 Elsevier Science B.V. All rights reserved. Keywords: Adhesion molecules; Regulation; Bovine
1. Introduction Adhesion molecules regulate the movement of cells into and within tissue. A basic understanding of the mechanisms involved in this process, derived from experiments with human and murine cells, has resulted in the elucidation of a multistep adhesion cascade which controls neutrophil and lymphocyte traf®cking (Springer, 1994, 1995). E-selectin is one of the several adhesion molecules responsible for regulating the early steps of tethering and rolling in the adhesion cascade (Bevilacqua and Nelson, 1993), thereby regulating the entry of PMNs and lymphocytes into sites of infection and in¯ammation. Leukocytes have crucial roles as effectors and regulators of the immune response, making their presence at sites of infection an important part of host defence. The relevance of this concept to other mammals is still unclear because regulation and expression of adhesion molecules has been investigated as a basis of leukocyte migration in a limited number of species. E-selectin is an inducible glycoprotein of the selectin family of adhesion molecules that use a C-type lectin-binding domain in their structure to mediate adhesion of leukocytes to endothelium. E-selectin is expressed on endothelial cells that have been stimulated with different types of immunomodulators including TNF-a and LPS (Bevilacqua et al., 1987), and mediates binding of PMNs and certain subsets of lymphocytes to endothelium in vitro (Graber et al., 1990; Shimizu et al., 1991). Jutila (1996) and coworkers (Walcheck et al., 1993; Jutila et al., 1994), have reported a role for bovine E-selectin in regulating adhesion of bovine gd T cells to human and bovine endothelial cells in vitro. Another form of E-selectin that may have a role in certain diseases, is shed E-selectin that has been detected in the plasma or circulation of patients with certain types of cancer (Benekli et al., 1998), severe infections (Kayal et al., 1998), and disorders such as Graves' disease (Wenisch et al., 1994), as well as in culture supernatants from stimulated human endothelial cells (Ohno et al., 1997). In domestic animals, soluble E-selectin has not been previously reported. Comparisons between surface expression and mRNA expression of human E-selectin has revealed that they follow similar patterns (Whelan et al., 1991; Doukas and Pober, 1990) and this may also be true in cattle. In addition, the mechanism of induction by TNF-a may differ from those of LPS in cattle, partly because bovine endothelial cells (BAEC) do not express CD14, the LPS binding receptor (Sopp et al., 1996), but do express a TNF-a receptor (Ferran et al., 1996). Since dairy cattle are particularly susceptible to mastitis, often caused by gram negative E. coli that releases LPS and induces TNF-a and other pro-in¯ammatory cytokines (Leitner et al., 1996), these two
C. Van Kampen, B.A. Mallard / Veterinary Immunology and Immunopathology 79 (2001) 151±165
153
stimuli are particularly relevant to cattle. Thus, this study seeks to establish the expression kinetics of bovine E-selectin in vitro and determine how rbTNF-a and LPS in¯uence amounts of surface protein and mRNA expression in BAECs. 2. Materials and methods 2.1. Endothelial cell isolation and culture Endothelial cells were obtained from aortas of recently slaughtered cattle and immersed in ice cold sterile PBS (pH 7.2), supplemented with fungizone (1%, Gibco). Sections of aorta (2.5±5.0 cm) were clamped at one end using a hemostat, washed thoroughly with PBS/fungizone, and the resulting ``pocket'' was ®lled with warmed (378C) dissociation buffer containing 10% collagenase (Sigma, St. Louis, MO), 1 ml of 7.5% bovine serum albumin (BSA, Gibco), and 90% PBS (pH 7.4). After incubation for 10 min at room temperature, the dissociation buffer was removed and the tissue was ¯ushed with 1 ml of bovine aortic endothelial cell (BAEC) culture medium (minimal essential medium MEM 10% FCS gentamicin (50 mg/ml, Gibco). The cell suspension was transferred to 35 mm petridishes (Corning, Fisher Scienti®c, Unionville, Ont.) and incubated for 2±5 days at 378C and 5% CO2 until a con¯uent monolayer had formed. During primary culture and the ®rst subculture, endothelial cell growth supplement (ECGS from bovine pituitary gland, 3 mg/100 ml; Sigma) was added to the culture medium to augment the growth conditions. Subculturing involved rinsing the monolayers with PBS and incubating the cells with 1±2 ml of trypsin±EDTA (0.025%, 0.53 mM EDTA-4Na; Gibco BRL) for 3±4 min at 378C/5% CO2. Ice cold culture medium was added to the ¯asks and after two washes with culture medium, the cells were resuspended and placed into 25 cm2 tissue culture ¯asks (Costar, Fisher Scienti®c, Unionville, Ont.) until needed. After reaching con¯uency, monolayers of BAECs were treated with recombinant bovine tumor necrosis factor alpha (rbTNF-a, 10 hg/ml; a gift from L. Babiuk, VIDO, Saskatoon, SK), lipopolysaccharide (LPS from E. coli O55:B5, 1 mg/ml; Sigma), or BAEC culture medium for 0, 1, 3, 6, 12, 18, 24, and 72 h at 378C/5% CO2. The viability of the BAEC cultures after stimulation was measured using tryptan blue exclusion. 2.2. Immunostaining for ¯ow cytometry Cells were harvested by treatment of the monolayers for 5 min with a solution of trypsin±EDTA (0.025%, Gibco BRL), after which the reaction was stopped with 10% FCS (Gibco BRL) in MEM (Gibco BRL) and the cells were collected into 5 ml of ice cold BAEC culture medium. After centrifugation, BAECs were resuspended in PBS containing 0.1% sodium azide and kept on ice. Single color (¯uorescein isothiocyanate) indirect immunostaining for expression of bovine E-selectin used the monoclonal antibody, EL-246 (1:1000 dilution) (a gift from M. Jutila, Montana State University, Bozeman, MT). The positive control was a polyclonal rabbit anti-human von Willebrand's factor (vWF) (Dimension Labs, Mississauga, Ont.) that cross-reacts with
154
C. Van Kampen, B.A. Mallard / Veterinary Immunology and Immunopathology 79 (2001) 151±165
bovine vWF and detects all endothelial cells. The negative controls included mouse IgG and anti-bovine L-selectin (CC32; 1:1000 dilution) (a gift from C. Howard, IAH, Compton, UK). Incubation with primary antibodies was for 30 min at 48C and after two washes with PBS/azide, the secondary antibody, ¯uorescein isothiocyanate (FITC)conjugated goat anti-mouse IgG (H L) (Cedarlane Laboratories, Hornby, Ont.) was added. The labeled with anti-vWF antibody were detected with biotinylated goat antirabbit IgG (Zymed, San Francisco, CA) as a secondary antibody. After 30 min at 48C, the cells were washed twice followed by a further incubation of those BAECs stained with anti-vWF antibody, with streptavidin-phycoerythrin (SA-PE) (Cedarlane Laboratories, Honiby, Ont.) for 12 min. After ®nal washes with PBS/azide, all cells were resuspended in 300 ml of PBS/azide and 300 ml of 1% paraformaldehyde, and brie¯y refrigerated until examined by ¯ow cytometry. 2.3. Flow cytometric analysis A FACScan ¯ow cytometer (Becton Dickinson) was used to acquire all data. LYSYS II software (Becton Dickinson) was used for data analysis. Ten thousand events were acquired from each sample and FL1 versus cell number histograms were generated for each sample. The percent of BAECs expressing bovine E-selectin was quanti®ed by using the negative controls to determine the placement of the markers as described previously (Van Kampen and Mallard, 1997). Mean ¯uorescence intensity (MFI) values were generated from software-generated statistics of the histograms. Data were compiled in this manner for samples from all time points and in four separate experiments. 2.4. Isolation of RNA from cultured BAECs Total RNA was extracted from all stimulated and unstimulated BAEC cultures using the acid guanidinium thiocyanate method (Chomczynski and Sacchi, 1987) and concentrations were measured by optical density (OD) at 260 nm (GeneQuant II, Pharmacia Biotech Inc, Bale d'Urfe, Quebec, Canada). 2.5. Primers for bovine E-selectin and b2-microglobulin genes A multiple internal control was constructed for use as an exogenous control in quantitative reverse transcription polymerase chain reaction (Q-RT-PCR) for bovine cytokines and adhesion molecules. It was synthesized in a manner similar to that described by Reddy et al. (1996) for use with pig cytokines. It consists of 50 and 30 primer sequences for several bovine cytokines and bovine E-selectin, VCAM-1 and ICAM-1, and includes primer sequences for b2-microglobulin (b2-m) to be used as the housekeeping gene during quanti®cation of mRNA. The primers were designed to amplify two products of different sizes during coampli®cation of control and target RNA via RT-PCR in a single reaction tube. The primer sequences for bovine E-selectin (Genbank accession #L12039) are: (50 ±30 ) forward TGAAGATGGCAGAGACAGAG; reverse CGACTGAGACATCAACGACA, and the product sizes are 259 and 386 bp for control and target, respectively. The primer sequences for b2-m (Genbank accession
C. Van Kampen, B.A. Mallard / Veterinary Immunology and Immunopathology 79 (2001) 151±165
155
#X69084) are: forward TGGCCTTGGTCCTTCTC; reverse CATTCGTCGTGGTAGCTC and the product sizes are 200 and 358 bp for control and target, respectively. These primers were designed based upon known bovine cDNA and mRNA sequences from Genbank. 2.6. Speci®city of primers The speci®city of the primers for bovine E-selectin and b2-m was con®rmed by `hot start' PCR in 25 ml containing 1 PCR buffer, 1.5 mM MgCl2, 200 mM of each dNTPs, 0.12 mM of each sequence-speci®c primer, 100 ng of plasmid DNA and 0.625 units of Taq DNA polymerase (Gibco BRL) for 35 cycles using a thermal cycler (Robocycler Gradient 96, Stratagene-PDI Bioscence). Each cycle included denaturation (948C/1 min), annealing (518C/1 min), extension (728C/1 min) and ®nal extension (728C/10 min) steps. Ampli®ed products were resolved by 2% agarose gel electrophoresis and the DNA bands were compared to the expected size of the target. All PCR reactions were performed under the above conditions unless otherwise stated. 2.7. cDNA synthesis In order to perform a reverse transcriptase reaction with both control and target RNA, the plasmid construct was ®rst subjected to reverse transcription to obtain construct-RNA (c-RNA) which could then be reverse transcribed back into cDNA in the same tube as the target RNA (Reddy et al., 1996). The plasmid construct was ®rst linearized with EcoRI, transcribed at 37±408C for 2 h using SP6 polymerase (Riboprobe II core system, Promega, Madison, WI, SA) to obtain c-RNA and then repeatedly digested with ampli®cation grade DNAse I (Gibco BRL) to remove any plasmid DNA templates. The RT reaction was performed in 10 ml containing 25 mM MgCl2, 1 PCR buffer II (50 mM KCl, 10 mM Tris±HCl, pH 8.3), 1 mM each of dNTPs, 1 U/ml of RNAse inhibitor, 2.5 U/ ml MuLV reverse transcriptase, 2.5 mM random hexamers and 25±50 hg of total RNA(target)/4 106 6 106 copies of c-RNA. For quanti®cation of E-selectin mRNA, the RT reaction was performed in 80 ml volumes containing 400 hg of total RNA and 2 105 copies of c-RNA. Reaction mixtures were sequentially incubated at room temperature for 10 min, 428C/60 min, 998C/8 min and 58C/5 min using a thermal cycler (Robocycler, Stratagene, La Jolla, CA). 2.8. Polymerase chain reaction The quantitative PCR method was modi®ed from that used by Wang et al. (1989) and was previously described (Reddy et al., 1996). An amount of 5 ml of cDNA mixture was serially diluted in 1:3 increments and ampli®cations were done using `hot start' PCR for 33 to 35 cycles using a thermal cycler. Products of PCR were analyzed by electrophoresis in 2% agarose gels stained with ethidium bromide and photographed using a transilluminator. Images of gels were scanned and analyzed using a densitometer (Molecular Analyst, Bio-Rad Laboratories, Mississauga, Ont.), and target and construct data were expressed as log10 of the density. The number of target copies was calculated
156
C. Van Kampen, B.A. Mallard / Veterinary Immunology and Immunopathology 79 (2001) 151±165
utilizing ``best ®t'' standard curves and extrapolated to 1 mg of total RNA as previously described (Reddy et al., 1996; Wang et al., 1989). 2.9. Measurement of shed E-selectin Due to the difference between protein and mRNA expression of E-selectin being most pronounced with rbTNF-a-stimulated BAEC cultures, the measurement of shed Eselectin was performed only with those cultures treated with rbTNF-a. Supernatants from BAEC cultures were harvested after treatment of BAECs cultured in 75 cm2 tissue culture ¯asks with 5 ml of rbTNF-a (10 hg/ml) for the same time periods as previously described. After centrifugation for 10 min at 400 g/48C to remove any cells or debris, supernatants were passed through an Ultrafree-4 ®lter column (Millipore Corporation, Bedford, MA) by adding 4 ml to each column and centrifuging all columns for 10±20 min at 600 g/ 48C until the volume in the columns was reduced by 50%. Concentrates were stored at 208C until used in the ELISA. A capture ELISA was performed with concentrated supernatants from all time points. An Immunolon II 96 well ¯at-bottomed plate (Dynatech, Fisher Scienti®c, Don Mills, Ont.) was coated with 10 mg/ml of EL-246 (anti-human E-selectin) for 3 days at 48C. After washing the plate with 0.05% Tween (Fisher Scienti®c)±PBS, blocking was done with 3% Tween±PBS for 1 h at room temperature, followed by another washing step. An amount of 100 ml of concentrated supernatant from all time points were added to each well and plates were incubated at room temperature for 2 h. After washing the plate using an EL403 plate washer (Biotek, Instruments, Highland Park, Vermont, USA), anti-bovine IgG(H L)-alkaline phosphatase (1/250 dilution, Jackson Labs, Biocan, Mississauga, Ont.) was added to each well and incubated for 2 h at room temperature. A solution of substrate, p-nitrophenyl phosphate disodium tablets (p-NPP, Sigma) was prepared in 10% diethanolamine substrate buffer and added at 100 ml per well. Color was developed for 60 min at room temperature in the dark and then the entire plate was read at 405 and 630 nm on an EL311 Microplate autoreader (Biotek Instruments, Highland Park, Vermont, USA). A 630 nm ®lter was used as a reference ®lter to correct for irregularities in the plastic. The negative control or blank was comprised of coating buffer with blocker and BAEC culture medium (instead of supernatant) plus conjugate. Positive control was a highly concentrated (75%) supernatant from 24 h stimulated cultures. Results were expressed as optical density values and mean OD values and standard deviations were determined, using Student's t-test, from three experiments with each sample tested in triplicate. 2.10. Data analysis The percentages of E-selectin positive BAECs at each time point were averaged from four separate experiments and analyzed for signi®cant differences between time points using Student's t-tests. All samples were analyzed for b2-m and E-selectin mRNA. b2-m values from three replicates at each time point and for untreated or rbTNF-a or LPS-treated BAECs, were averaged and Student's t-tests were performed. Since differences between time points and
C. Van Kampen, B.A. Mallard / Veterinary Immunology and Immunopathology 79 (2001) 151±165
157
between treatments were not signi®cant for b2-m mRNA expression, the mean values for b2-m of 21 samples from each stimulation experiment (n 4) were used as a standard expected b2-m value to normalize each target value using the following equation: expected b2-m/observed b2-m X. Then the expected value for target was obtained by: X observed target value. Values of three replicates of target E-selectin were generated in this manner for each time point and for each stimulus in three separate experiments, averaged, and the tabulated data were analyzed by Student's t-test to compare results between stimuli and over time. Signi®cance was reported at P 0:05. 3. Results 3.1. Surface expression of E-selectin The proportions of BAECs that expressed E-selectin after stimulation with rbTNF-a or LPS are summarized in Fig. 1A. Very few unstimulated BAECs expressed E-selectin, however, as early as 1 h post-stimulation, the proportions of E-selectin cells began to signi®cantly increase. Expression plateaued in rbTNF-a stimulated BAECs, with the ®rst plateau between 3 and 12 h and the second between 18 and 72 h (Fig. 1A). The same phenomenon did not occur with LPS-stimulated BAECs; instead there was a gradual increase from 1±6 h and then a constant proportion of BAECs expressed E-selectin between 6 and 24 h. The mean ¯uorescence intensity peaked at 6 h and then gradually declined until 24 h following stimulation with LPS (Fig. 1B). The MFI gradually but not signi®cantly increased following TNF-a treatment. The difference in proportions of Eselectin BAECs between the two stimuli was not signi®cant except at 6 h (P 0:05). At 72 h, E-selectin expression in the rbTNF-a treatment group remained elevated above baseline. Expression of E-selectin on supramammary arterial endothelial cells did not differ signi®cantly from the expression kinetics on BAECs (data not shown). Viability of the BAECs after stimulation with either LPS or rbTNF-a was greater than 90% at all time points after stimulation. 3.2. Sensitivity of Q-RT-PCR Sensitivity of the assay was determined using a method from Reddy (1998, Ph.D. thesis), involving the use of two experiments that extrapolated the copy number after QRT-PCR of known amounts of starting material at different ratios of control to target DNA molecules. First, the amount of control cDNA molecules was held constant and the target DNA concentrations were varied in ratios of 1:1 to 1:6 using b2-m primers. Extrapolated starting values ranged from 9:62 107 1:22 108 for an expected value or actual starting value of 1:57 106 /mg total RNA (Table 1A). Quanti®cation was possible at up to a six-fold difference between control and target DNA templates with 9.77 hg of total RNA estimated to contain 8:95 106 copies and 3:98 106 copies in 0.326 hg at the 1:6 ratio. Using these extrapolations, it is possible to detect 1100 copies from a 80 ml RT product containing 2 105 copies of control RNA and an unknown copy number of target RNA.
158
C. Van Kampen, B.A. Mallard / Veterinary Immunology and Immunopathology 79 (2001) 151±165
Fig. 1. (A) E-selectin surface expression on cultured bovine endothelial cells. Monolayers were stimulated with recombinant bovine tumor necrosis factor alpha (rbTNF-a) (solid bars) or lipopolysaccharide (LPS) (hatched bars) for 0, 1, 3, 6, 12, 18, 24, or 72 h. Cells were harvested and immunostained with anti-E-selectin (EL246) mAb for ¯ow cytometric analysis. Ten thousand events per time point were analyzed using a FACScan ¯ow cytometer and results were expressed as a mean percentage of E-selectin cells S:D: for each time point in four separate experiments. Signi®cant differences between time points are indicated by different lower (TNF-a) or upper (LPS) case letters. (B) Mean ¯uorescence intensity (MFI) values for E-selectin expression on cultured bovine aortic endothelial cells. Monolayers were stimulated with recombinant bovine tumor necrosis factor alpha (rbTNF-a) (solid bars) or lipopolysacharride (LPS) (hatched bars) for 0, 1, 3, 6, 12, 18, 24, or 72 h. Cells were harvested and immunostained with anti-E-selectin (EL246) mAb for ¯ow cytometric analysis. Ten thousand events per time point were analyzed using a FACScan ¯ow cytometer and results were expressed as an average of MFI S:D: for each time point in four separate experiments. Signi®cant differences between time points are indicated by different lower (TNF-a) or upper (LPS) case letters.
In the second experiment, the minimum copy number was determined by Q-RT-PCR using varying concentrations of control cDNA molecules and a constant amount of targetDNA (Table 1B). Observed values for b2-m ranged between 0:92 107 and 1:83 107 molecules/mg total RNA. Accuracy was determined by comparing the expected and observed copy numbers at different dilutions. Generally, the observed values were always less than expected regardless of whether the target DNA or the control DNA was constant or varied (Table 1).
C. Van Kampen, B.A. Mallard / Veterinary Immunology and Immunopathology 79 (2001) 151±165
159
Table 1 Sensitivity of the quantitative reverse transcriptase polymerase chain reaction (Q-RT-PCR) was determined by subjecting different ratios of standard and target cDNA templates to Q-RT-PCR using b2-m primers and extrapolating the copy number in two separate experimentsa Standard copies (a)
Target (Zg) (b)
Ratio of a:b
Expected value (Zg)
Observed value (Zg ) mean S.D.
Ratio of expected: observed (folds)
A 4.6 108 4.6 108 4.6 108
2.93 0.977 0.326
1:1 1:3 1:6
1.57 107 1.57 107 1.57 107
9.62 106 3.1 104 9.16 106 4.25 105 1.22 107 5.8 106
1.63 1.71 1.29
B 4.6 108 1.53 108 0.511 108
2.93 2.93 2.93
1:1 1:3 1:6
1.57 107 1.57 107 1.57 107
1.08 107 4.6 105 0.92 107 7.5 105 1.23 107 9.1 105
1.45 1.70 1.28
a (A) Copy number determined at a constant amount of control cDNA molecules with varying concentrations of target cDNA templates. (B) Copy number determined at varying concentrations of control cDNA molecules and a constant amount of target cDNA templates. Accuracy was determined by comparing expected and observed values in each experiment.
3.3. b2-microglobulin mRNA Message for b2-m was quanti®ed in BAECs at 0, 1, 3, 6, 12, 18, and 24 h after stimulation in vitro with either LPS, rbTNF-a, or culture medium (Fig. 2). No signi®cant differences were seen for the mean values for b2-m message between unstimulated and stimulated BAEC cultures (P 0:05). Mean values for b2-m mRNA in unstimulated and rbTNF-a stimulated cultures were 2:4 108 and 2:8 108 copies/mg total RNA, respectively. Cultures stimulated with LPS expressed a mean value of 3:31 108 copies/mg
Fig. 2. Beta 2-microglobulin (b2-m) mRNA expression in cultured bovine aortic endothelial cells. Monolayers were stimulated with recombinant bovine tumor necrosis factor alpha (rbTNF-a) (solid bars) or lipopolysaccharide (LPS) (hatched bars) for 0, 1, 3, 6, 12, 18, 24, or 72 h. Cells were harvested and total RNA was extracted from all samples. Expression of b2-m mRNA was analyzed by Q-RT-PCR and copy number was determined per mg of total RNA. Mean copy number S:D: was obtained for three replicates at each time point. No signi®cant differences were found over time or between stimuli.
160
C. Van Kampen, B.A. Mallard / Veterinary Immunology and Immunopathology 79 (2001) 151±165
total RNA of b2-m message. Thus, the expression of b2-m message did not vary signi®cantly over time or between stimuli (Fig. 2). There was some variation in b2-m message between experiments, however, it was not signi®cant, and it was factored into normalization of the E-selectin mRNA data. 3.4. E-selectin mRNA expression In vitro stimulation of bovine BAEC with rbTNF-a resulted in a different pattern of mRNA expression than that induced by LPS (Fig. 3). Unstimulated cultures expressed very small amounts of E-selectin mRNA although the amounts were different between rbTNF-a (2:26 107 copies/mg) and LPS (0:111 107 copies/mg). There was a signi®cant increase (p£ 0.05) in E-selectin mRNA expression 1 h after treatment with rbTNF-a but then the number of copies decreased between 1 and 6 h to values lower (0:865 0:064 107 copies/mg) than those measured in the unstimulated cultures (2:26 0:48 107 copies/mg). After 6 h, the copy number increased gradually until 12 h, then remained constant until 18 h after which E-selectin mRNA expression rapidly increased to very high amounts at 72 h (67 27 107 copies/mg). In cultures treated with LPS, mRNA expression peaked at 6 h (14 8:21 107 copies/mg) and by 24 h, had gradually decreased to amounts similar to those in unstimulated cultures (0:11 0:05 107 copies/mg). 3.5. Shedding of E-selectin Amounts of shed E-selectin were expressed as OD values and are summarized in Fig. 4. Unstimulated BAEC cultures produced very little soluble E-selectin, and OD values were
Fig. 3. E-selectin mRNA expression in cultured bovine aortic endothelial cells. Monolayers were stimulated with recombinant bovine tumor necrosis factor alpha (rbTNF-a) (solid line) or lipopolysaccharide (LPS) (dotted line) for 0, 1, 3, 6, 12, 18, 24, or 72 h. Cells were harvested and total RNA was extracted from all samples. Expression of E-selectin mRNA was analyzed by Q-RT-PCR and copy number was determined per mg of total RNA. Mean copy number S:D: was obtained for three replicates at each time point from three separate experiments. Signi®cant differences over time are indicated by different lower (TNF-a) or upper (LPS) case letters.
C. Van Kampen, B.A. Mallard / Veterinary Immunology and Immunopathology 79 (2001) 151±165
161
Fig. 4. Shed E-selectin in supernatants from cultured bovine aortic endothelial cells. Monolayers were stimulated with recombinant bovine tumor necrosis factor alpha (rbTNF-a) for 0, 1, 3, 6, 12, 18, 24, or 48 h. Supernatants were collected and concentrated by approximately 50% and used in an ELISA. The amount of soluble E-selectin was expressed as OD (405, 630 nm). Each bar represents the mean OD S:D: of triplicate values from three separate experiments. Signi®cant differences over time are indicated by different lower case letters.
not signi®cantly different between unstimulated supernatants and the negative control. There was a signi®cant increase in the amount of soluble E-selectin in stimulated cultures at 1±6 h and then again at 12±48 h, with the highest amount of shed E-selectin present in the 48 h supernatants. 4. Discussion Bovine E-selectin surface expression and mRNA expression were measured in BAECs stimulated with either rbTNF-a or LPS. The proportions of BAECs that expressed Eselectin after stimulation were never higher than 56%. This low percentage is likely to be due to an effect of the trypsin±EDTA solution used to dissociate the cells from the culture ¯ask, that may have removed some of the surface E-selectin molecules. This seems likely based upon the observation that close to 100% of the cells expressed E-selectin by 18 h (data not shown). None of the commercially available dissociation buffers were suitable for detaching the BAECs from the ¯ask. Thus, the total number of cells expressing Eselectin, at all time points are less than 100%, but the effect of each stimulus over time remained constant between experiments. The increased numbers of E-selectin BAEC between 3 and 12 h, supports previous work that showed the induction of bovine Eselectin by LPS and TNF-a between 4 and 12 h (Zund et al., 1996). Changes in MFI were different between LPS and TNF-a, indicating that among other possibilities, the mechanism of induction of E-selectin may differ between stimuli, or the effect may be dose-related. Furthermore, the signi®cant increase in MFI at 6 h after LPS stimulation,
162
C. Van Kampen, B.A. Mallard / Veterinary Immunology and Immunopathology 79 (2001) 151±165
followed by a gradual decrease in surface concentration, paralleled the mRNA kinetics induced by LPS. These ®ndings suggest a correlation between mRNA and the number of E-selectin surface molecules, as the proportion of cells expressing LPS remained constant between 6 and 24 h. However, the lack of variation in MFI in cells treated with TNF-a differed considerably from the mRNA expression kinetics, again indicating a possible difference in the mechanism of induction by LPS and TNF-a. Measurement of bovine E-selectin mRNA expression used a technique involving a plasmid that contained multiple primer sequences for bovine cytokines and adhesion molecules (Reddy et al., 1996). A previous study of bovine cytokine expression used multiple primer pairs (Covert and Splitter, 1995) but they were not all contained within a single construct that can measure several molecules in a single reaction tube. Similar methods have not been previously used to quantify AM expression. Results of the current study indicate a prolonged expression of E-selectin mRNA in cultures stimulated with TNF-a, with peak surface expression at 18±24 h post-stimulation that persisted until 72 h. In cultures stimulated with LPS, the peak mRNA expression occurs at 6 h which is before peak surface expression at 18±24 h. During the early hours after stimulation, fewer mRNA transcripts were required for LPS-induced E-selectin surface expression than for rbTNF-a-induced expression. Later, the proportions of E-selectin BAECs remained fairly constant while mRNA expression was low, in cultures treated with LPS, whereas during the same period, mRNA expression was substantially higher in BAECs treated with rbTNF-a. It is dif®cult to know the reason(s) for these observations and further studies into the mechanism(s) that each stimulus uses to induce AM are necessary. Previously, E-selectin mRNA was measured in human umbilical vein endothelial cells stimulated with TNF-a and skin biopsies of delayed hypersensitivity sites (Meagher et al., 1994), and results indicated the highest amount of E-selectin mRNA in stimulated HUVECs, at 2 h with only a modest decrease at 8 h and sustained mRNA expression at 24 h, whereas skin biopsies contained large amounts of E-selectin mRNA between 6 and 72 h, indicating the potential long term expression of E-selectin mRNA in vivo. This supports the current ®ndings at the earlier time points, to a certain degree, but the signi®cant increase in E-selectin mRNA between 24 and 72 h in this study has not been reported in humans. The choice to use aortic endothelial cells rather than capillary endothelial cells from the mammary gland, which would be most relevant to study Eselectin expression, was based upon the preliminary observation that expression of this adhesion molecule on both types of endothelial cells was similar, and the relative dif®culty of isolating aortic endothelial cells was much lower than that for the capillary endothelial cells. Previous studies report that human E-selectin mRNA and surface protein expression are correlated (Whelan et al., 1991; Bevilacqua et al., 1989). However, this has not been the case for all proteins; for example, the product of certain cytokine mRNA transcription may not always be detectable (Kaspar and Gehrke, 1994). Surface expression did not correlate with mRNA expression of E-selectin in BAEC cultures stimulated with rbTNFa, however, the patterns of expression were more similar in cultures stimulated with LPS. Surface expression and mRNA expression both peaked in LPS-stimulated BAECs at 6 h. The small increase in mRNA expression at 1 h, followed by a signi®cant decrease at 6 h after treatment with rbTNF-a did not correlate with the continued increase in E-selectin
C. Van Kampen, B.A. Mallard / Veterinary Immunology and Immunopathology 79 (2001) 151±165
163
surface expression during the same period. Most studies of bovine adhesion molecules, use a single period of stimulation for their experiments, even with different stimuli. Thus, it is dif®cult to compare these ®ndings with other studies of bovine E-selectin expression. Previous studies of human endothelial cells, showed that ¯uctuations in human E-selectin protein expression correspond to similar changes in mRNA (Whelan et al., 1991; Bevilacqua et al., 1989). However, another report, showed dissimilar patterns between IL1 protein accumulation and IL-1 mRNA expression in peripheral blood mononuclear cells stimulated with LPS or C5a (Kaspar and Gehrke, 1994). There are various possible reasons for the lack of correlation between surface and mRNA expression of E-selectin in the case of TNF-a stimulation. One reason that was suggested by Kaspar and Gehrke (1994) regarding IL-1, is that mRNA translation does not occur, or that the gene product is not detectable. Adhesion molecules, including Eselectin are shed from the cell surface after or during stimulation (Grif®n et al., 1990; Wyble et al., 1997). Thus, the difference between surface expression and mRNA expression of bovine E-selectin in rbTNF-a-stimulated BAECs, may be due, at least in part to E-selectin being shed from the surface of the BAECs during the later time points, and at a much higher rate than earlier time points. This may cause the amount of mRNA transcripts to continue to rise to keep up with the demand of the protein synthesis machinery. Therefore, the amount of shed E-selectin in the supernatants of rbTNF-astimulated BAECs was measured using ELISA. The amount of E-selectin in the culture supernatants increased steadily over time. Subsequent experiments could investigate if the turnover rate of mRNA vis a vis stability or degradation, has an in¯uence on the patterns observed in this study, or if detection of intracellular E-selectin in cells stimulated with rbTNF-a might also account for the large number of transcripts between 18 and 72 h. E-selectin protein is not, however, known to be stored intracellularly (Bevilacqua et al., 1987). Finally, measuring shed E-selectin in LPS-stimulated culture supernatants might help to account for the different patterns of mRNA expression between the stimuli. In conclusion, this study showed that both TNF-a and LPS induced expression of Eselectin on BAEC but that surface protein and mRNA expression were not well correlated. Furthermore, the expression of bovine E-selectin differs from human E-selectin in that expression is sustained for at least 72 h in TNF-a, and 24 h in LPS-stimulated BAEC cultures, whereas E-selectin expression in HUVECs peaks at 6 h and returns to baseline by 24 h (Meagher et al., 1994). E-selectin expression on BAEC is, however, closer to the expression kinetics documented in the pig in which expression of porcine E-selectin was prolonged up to 48 h post-stimulation with IL-1a or TNF-a (Tsang et al., 1995). With clear differences between bovine E-selectin mRNA expression in BAECs treated with different stimuli, it will be important to determine if this occurs in vivo. Acknowledgements We thank Dr. N.R. Reddy for constructing the bovine construct and for his instruction in the art of Q-RT-PCR. We also thank Sophia Lim for performing the ELISA experiments and Marion Wilke for her assistance with primary culturing of the BAECs.
164
C. Van Kampen, B.A. Mallard / Veterinary Immunology and Immunopathology 79 (2001) 151±165
We thank the staff of the Animal and Poultry Science abbattoir for assisting in the recovery of bovine aortas: We gratefully acknowledge the individuals that provided mAbs (M. Jutila and C. Howard) and rbTNF-a (Lorne Babiuk) as gifts for this research. This research was supported by grants to Bonnie Mallard from the Natural Sciences and Engineering Research Council of Canada, and Ontario Ministry of Agriculture and Rural Affairs (OMAFRA). References Benekli, M., Gullu, I.H., Tekuzman, G., Savas, M.C., Hayran, M., Hascelik, G., Firat, D., 1998. Circulating intercellular adhesion molecule-1 and E-selectin levels in gastric cancer. Br. J. Cancer 78, 267±271. Bevilacqua, M.P., Nelson, R.M., 1993. Selectins. J. Clin. Invest. 91, 379±387. Bevilacqua, M.P., Pober, J.S., Mendrick, D.L., Cotran, R.S., Gimbrone Jr., M.A., 1987. Identi®cation of an inducible endothelial-leukocyte adhesion molecule. Proc. Natl. Acad. Sci. 84, 9238±9242. Bevilacqua, M.P., Stengelin, S., Gimbrone Jr., M.A., Seed, B., 1989. Endothelial leukocyte adhesion molecule-1: an inducible receptor for neutrophils related to complement regulatory proteins and lectins. Science 243, 1160±1165. Chomczynski, P., Sacchi, N., 1987. Single-step method of RNA isolation by acid guanidinium thiocyanatephenol-chloroform extraction. Anal. Biochem. 162, 156±159. Covert, J., Splitter, G., 1995. Detection of cytokine transcriptional pro®les from bovine peripheral blood mononuclear cells and CD4 lymphocytes by reverse transcriptase polymerase chain reaction. Vet. Immunol. Immunopathol. 49, 39±50. Doukas, J., Pober, J.S., 1990. IFN-g enhances endothelial activation induced by tumor necrosis factor but not IL1. J. Immunol. 145, 1727±1731. Ferran, C., Cooper, J.T., Brostjan, C., Stroka, D.M., Millan, M.T., Goodman, D.J., Bach, F.H., 1996. Expression of a truncated form of the human p55 TNF-receptor in bovine aortic endothelial cells renders them resistant to human TNF. Transplant. Proc. 28, 618±619. Graber, N., Copal, T.V., Wilson, D., Beall, L.D., Polte, T., Newman, W., 1990. T cells bind to cytokine-activated endothelial cells via a novel, inducible sialoglycoprotein and endothelial leukocyte adhesion molecule-1. J. Immunol. 145, 819±830. Grif®n, J.D., Spertini, O., Ernst, T.J., Belvin, M.P., Levine, H.B., Kanakura, Y., Tedder, T.F., 1990. Granulocytemacrophage colony stimulating factor and other cytokines regulate surface expression of the leukocyte adhesion molecule-1 on human neutrophils, monocytes and their precursors. J. Immunol. 145, 576±579. Jutila, M.A., Bargatze, R.F., Kurk, S., Warnock, R.A., Ehsani, N., Watson, S.R., Walcheck, B., 1994. Cell surface P- and E-selectin support shear-dependent rolling of bovine gd T cells. J. Immunol. 153, 3917± 3928. Jutila, M.A., 1996. gd T cell/endothelial cell interactions. Vet. Immunol. Immunopathol. 54, 105±110. Kaspar, R.L., Gehrke, L., 1994. Peripheral blood mononuclear cells stimulated with C5a or lipopolysaccharide to synthesize equivalent levels of IL-1 beta mRNA show unequal IL-1 beta protein accumulation but similar polyribosome pro®les. J. Immunol. 153, 277±286. Kayal, S., Jais, J.P., Aguini, N., Chaudiere, J., Labrousse, J., 1998. Elevated circulating E-selectin, intercellular adhesion molecule 1, and von Willebrand factor in patients with severe infection. Am. J. Respir. Crit. Care Med. 157, 776±784. Leitner, G., Shoshani, E., Berman, A., Trinin, Z., Feingold, D., Einkler, M., Shpigel, N.Y., Saran, A., 1996. Immune response of the mammary gland in dairy cows after experimental infection with Escherichia coli or Staphylococcus aureus. In: 3rd international Mastitis Seminar Proceedings, May 28±June 1, Tel Aviv, Israel. Meagher, L., Mahiouz, D., Sugars, K., Burrows, N., Norris, P., Yarwood, H., Becker-Andre, M., Haskard, D.O., 1994. Measurement of mRNA for E-selectin, VCAM-1 and ICAM-1 by reverse transcription and the polymerase chain reaction. J. Immunol. Methods 175, 237±246. Ohno, M., Ichikawa, H., Coe, L., Kvietys, P.R., Granger, D.N., Alexander, J.S., 1997. Soluble selectins and ICAM-1 modulate neutrophil-endothelial adhesion and diapedesis in vitro. In¯ammation 21 (3), 313±324.
C. Van Kampen, B.A. Mallard / Veterinary Immunology and Immunopathology 79 (2001) 151±165
165
Reddy, N.R., Wilkie, B.N., Mallard, B.A., 1996. Construction of an internal control to quantitate multiple porcine cytokine mRNAs by RT-PCR. BioTechniques 21, 868±875. Shimizu, Y., Newman, W., Gopal, T.V., Horgan, K.J., Graber, N., Beall, L.D., Van Seventer, O.A., Shaw, S., 1991. Four molecular pathways of T cell adhesion to endothelial cells: roles of LFA-1, VCAM-1 and ELAM-1 and changes in pathway hierarchy under different activation conditions. J. Cell Biol. 113, 1203± 1212. Sopp, P., Kwong, L.S., Howard, C.J., 1996. Identi®cation of bovine CD14. Vet. Immunol. Immunopathol. 52, 323±328. Springer, T.A., 1994. Traf®c signals for lymphocyte recirculation and leukocyte emigration: the multistep paradigm. Cell 76, 301±314. Springer, T.A., 1995. Traf®c signals on endothelium for lymphocyte recirculation and leukocyte emigration. Ann. Rev. Physiol. 57, 827±872. Tsang, Y.T.M., Stephens, P.E., Licence, S.T., Haskard, D.O., Binns, R.M., 1995. Porcine E-selectin: cloning and functional characterization. Immunology 85, 140±145. Van Kampen, C., Mallard, B.A., 1997. Effects of peripartum stress and disease on bovine lymphocyte subsets. Vet. Immunol. Immunopathol. 59 (1/2), 79±92. Walcheck, B., Watts, G., Jutila, M.A., 1993. Bovine gd T cells bind E-selectin via a novel glycoprotein receptor: ®rst characterization of a lymphocyte/E-selectin interaction in an animal model. J. Exp. Med. 178, 853±863. Wang, A.M., Doyle, M.V., Mark, D.F., 1989. Quantitation of mRNA by the polymerase chain reaction. Proc. Natl. Acad. Sci U.S.A. 86, 9717±9721 [published erratum: 1989. Proc. Natl. Acad. Sci. U.S.A. 87, 2865]. Wenisch, C., Myskiw, D., Parschalk, B., Hartmann, T., Dam, K., Graninger, W., 1994. Soluble endotheliumassociated adhesion molecules in patients with Graves' disease. Clin. Exp. Immunol. 98, 240±244. Whelan, J., Ghersa, P., Hooft van Huijsduijen, R., Gray, J., Chandra, G., Talabot, F., Delamarter, J.F., 1991. An NFkB-like factor is essential but not suf®cient for cytokine induction of endothelial leukocyte adhesion molecule 1 (ELAM-1) gene transcription. Nucleic Acids Res. 19, 2645±2653. Wyble, C.W., Hynes, K.L., Kuchbhotla, J., Marcus, B.C., Hallahan, D., Oewertz, B.L., 1997. TNF-alpha and IL1 upregulate membrane-bound and soluble E-selectin through a common pathway. J. Surg. Res. 73, 107±112. Zund, G., Nelson, D.P., Neufeld, E.J., Dzus, A.L., Bischoff, J., Mayer, J.E., Colgan, S.P., 1996. Hypoxia enhances stimulus-dependent induction of E-selectin on aortic endothelial cells. Proc. Natl. Acad. Sci. 93, 7075±7080.
Regulation of bovine E-selectin expression by recombinant tumor necrosis factor alpha and lipopolysaccharide Corinne Van Kampen*, B.A. Mallard Department of Pathobiology, University of Guelph, Guelph, Ont., Canada N1G 2W1 Received 5 September 2000; received in revised form 31 January 2001; accepted 11 February 2001
Abstract Induction of adhesion molecules by cytokines and LPS is an important mechanism of regulating leukocyte migration into tissue. Expression and regulation of E-selectin may be differentially in¯uenced by the stimuli involved with effects on mRNA or surface protein kinetics. Surface protein and mRNA expression kinetics of bovine E-selectin were measured and compared in primary cultures of bovine aortic endothelial cells (BAEC) stimulated for various periods of time with recombinant bovine tumor necrosis factor alpha (rbTNF-a) or Escherichia coli lipopolysaccharide (LPS). E-selectin mRNA expression was measured via quantitative reverse transcription polymerase chain reaction (Q-RT-PCR) using a construct that contained multiple synthetic oligonucleotides for several bovine adhesion molecules and cytokines. Surface expression of Eselectin was measured by ¯ow cytometry. Unstimulated BAECs expressed minimum or no Eselectin on the surface. A low number of endothelial cells expressed surface E-selectin as early as 1 h post-stimulation and surface expression was sustained after both stimuli for 24±72 h. Mean ¯uorescence intensity (MFI) indicated peak surface concentration of E-selectin at 6 h poststimulation after LPS followed by a gradual decrease to 72 h without returning to baseline values. Mean ¯uorescence intensity following stimulation with TNF-a increased slightly between 0 and 72 h. The pattern of mRNA expression differed between stimuli. LPS-stimulated BAECs expressed peak amounts of E-selectin mRNA at 6 h, followed by a decline to baseline by 24 h. Conversely, BAECs stimulated with rbTNF-a expressed signi®cantly (p£ 0.05) higher amounts of mRNA at 1 h than compared to unstimulated controls (0 h), but this decreased to below baseline levels by 6 h; followed by a gradual increase and eventually a sharp increase between 18 and 72 h. To account for the lack of correlation between mRNA and protein expression, it was hypothesized that shedding of surface E-selectin accounted at least in part, for the large increase in mRNA expression seen at 18± 72 h. Culture supernatants from rbTNF-a-treated BAECs were harvested, and tested for the presence of shed E-selectin using ELISA. Unstimulated culture supernatants contained little or no *
Corresponding author. Present address: Department of Medicine, McMaster University, 1200 Main St. West, Hsc. Rm. 3N5, Hamilton, Ont., Canada L8N 3Z5. Tel.: 1-905-525-9140/ext. 22212. 0165-2427/01/$ ± see front matter # 2001 Elsevier Science B.V. All rights reserved. PII: S 0 1 6 5 - 2 4 2 7 ( 0 1 ) 0 0 2 4 9 - 5
152
C. Van Kampen, B.A. Mallard / Veterinary Immunology and Immunopathology 79 (2001) 151±165
E-selectin. Between 6 and 48 h, the concentration of E-selectin in culture supernatants from rbTNFa-stimulated BAECs increased approximately two-fold, suggesting that the sharp increase in Eselectin mRNA expression around 18 h may be related to signi®cant loss of surface E-selectin during this period. # 2001 Elsevier Science B.V. All rights reserved. Keywords: Adhesion molecules; Regulation; Bovine
1. Introduction Adhesion molecules regulate the movement of cells into and within tissue. A basic understanding of the mechanisms involved in this process, derived from experiments with human and murine cells, has resulted in the elucidation of a multistep adhesion cascade which controls neutrophil and lymphocyte traf®cking (Springer, 1994, 1995). E-selectin is one of the several adhesion molecules responsible for regulating the early steps of tethering and rolling in the adhesion cascade (Bevilacqua and Nelson, 1993), thereby regulating the entry of PMNs and lymphocytes into sites of infection and in¯ammation. Leukocytes have crucial roles as effectors and regulators of the immune response, making their presence at sites of infection an important part of host defence. The relevance of this concept to other mammals is still unclear because regulation and expression of adhesion molecules has been investigated as a basis of leukocyte migration in a limited number of species. E-selectin is an inducible glycoprotein of the selectin family of adhesion molecules that use a C-type lectin-binding domain in their structure to mediate adhesion of leukocytes to endothelium. E-selectin is expressed on endothelial cells that have been stimulated with different types of immunomodulators including TNF-a and LPS (Bevilacqua et al., 1987), and mediates binding of PMNs and certain subsets of lymphocytes to endothelium in vitro (Graber et al., 1990; Shimizu et al., 1991). Jutila (1996) and coworkers (Walcheck et al., 1993; Jutila et al., 1994), have reported a role for bovine E-selectin in regulating adhesion of bovine gd T cells to human and bovine endothelial cells in vitro. Another form of E-selectin that may have a role in certain diseases, is shed E-selectin that has been detected in the plasma or circulation of patients with certain types of cancer (Benekli et al., 1998), severe infections (Kayal et al., 1998), and disorders such as Graves' disease (Wenisch et al., 1994), as well as in culture supernatants from stimulated human endothelial cells (Ohno et al., 1997). In domestic animals, soluble E-selectin has not been previously reported. Comparisons between surface expression and mRNA expression of human E-selectin has revealed that they follow similar patterns (Whelan et al., 1991; Doukas and Pober, 1990) and this may also be true in cattle. In addition, the mechanism of induction by TNF-a may differ from those of LPS in cattle, partly because bovine endothelial cells (BAEC) do not express CD14, the LPS binding receptor (Sopp et al., 1996), but do express a TNF-a receptor (Ferran et al., 1996). Since dairy cattle are particularly susceptible to mastitis, often caused by gram negative E. coli that releases LPS and induces TNF-a and other pro-in¯ammatory cytokines (Leitner et al., 1996), these two
C. Van Kampen, B.A. Mallard / Veterinary Immunology and Immunopathology 79 (2001) 151±165
153
stimuli are particularly relevant to cattle. Thus, this study seeks to establish the expression kinetics of bovine E-selectin in vitro and determine how rbTNF-a and LPS in¯uence amounts of surface protein and mRNA expression in BAECs. 2. Materials and methods 2.1. Endothelial cell isolation and culture Endothelial cells were obtained from aortas of recently slaughtered cattle and immersed in ice cold sterile PBS (pH 7.2), supplemented with fungizone (1%, Gibco). Sections of aorta (2.5±5.0 cm) were clamped at one end using a hemostat, washed thoroughly with PBS/fungizone, and the resulting ``pocket'' was ®lled with warmed (378C) dissociation buffer containing 10% collagenase (Sigma, St. Louis, MO), 1 ml of 7.5% bovine serum albumin (BSA, Gibco), and 90% PBS (pH 7.4). After incubation for 10 min at room temperature, the dissociation buffer was removed and the tissue was ¯ushed with 1 ml of bovine aortic endothelial cell (BAEC) culture medium (minimal essential medium MEM 10% FCS gentamicin (50 mg/ml, Gibco). The cell suspension was transferred to 35 mm petridishes (Corning, Fisher Scienti®c, Unionville, Ont.) and incubated for 2±5 days at 378C and 5% CO2 until a con¯uent monolayer had formed. During primary culture and the ®rst subculture, endothelial cell growth supplement (ECGS from bovine pituitary gland, 3 mg/100 ml; Sigma) was added to the culture medium to augment the growth conditions. Subculturing involved rinsing the monolayers with PBS and incubating the cells with 1±2 ml of trypsin±EDTA (0.025%, 0.53 mM EDTA-4Na; Gibco BRL) for 3±4 min at 378C/5% CO2. Ice cold culture medium was added to the ¯asks and after two washes with culture medium, the cells were resuspended and placed into 25 cm2 tissue culture ¯asks (Costar, Fisher Scienti®c, Unionville, Ont.) until needed. After reaching con¯uency, monolayers of BAECs were treated with recombinant bovine tumor necrosis factor alpha (rbTNF-a, 10 hg/ml; a gift from L. Babiuk, VIDO, Saskatoon, SK), lipopolysaccharide (LPS from E. coli O55:B5, 1 mg/ml; Sigma), or BAEC culture medium for 0, 1, 3, 6, 12, 18, 24, and 72 h at 378C/5% CO2. The viability of the BAEC cultures after stimulation was measured using tryptan blue exclusion. 2.2. Immunostaining for ¯ow cytometry Cells were harvested by treatment of the monolayers for 5 min with a solution of trypsin±EDTA (0.025%, Gibco BRL), after which the reaction was stopped with 10% FCS (Gibco BRL) in MEM (Gibco BRL) and the cells were collected into 5 ml of ice cold BAEC culture medium. After centrifugation, BAECs were resuspended in PBS containing 0.1% sodium azide and kept on ice. Single color (¯uorescein isothiocyanate) indirect immunostaining for expression of bovine E-selectin used the monoclonal antibody, EL-246 (1:1000 dilution) (a gift from M. Jutila, Montana State University, Bozeman, MT). The positive control was a polyclonal rabbit anti-human von Willebrand's factor (vWF) (Dimension Labs, Mississauga, Ont.) that cross-reacts with
154
C. Van Kampen, B.A. Mallard / Veterinary Immunology and Immunopathology 79 (2001) 151±165
bovine vWF and detects all endothelial cells. The negative controls included mouse IgG and anti-bovine L-selectin (CC32; 1:1000 dilution) (a gift from C. Howard, IAH, Compton, UK). Incubation with primary antibodies was for 30 min at 48C and after two washes with PBS/azide, the secondary antibody, ¯uorescein isothiocyanate (FITC)conjugated goat anti-mouse IgG (H L) (Cedarlane Laboratories, Hornby, Ont.) was added. The labeled with anti-vWF antibody were detected with biotinylated goat antirabbit IgG (Zymed, San Francisco, CA) as a secondary antibody. After 30 min at 48C, the cells were washed twice followed by a further incubation of those BAECs stained with anti-vWF antibody, with streptavidin-phycoerythrin (SA-PE) (Cedarlane Laboratories, Honiby, Ont.) for 12 min. After ®nal washes with PBS/azide, all cells were resuspended in 300 ml of PBS/azide and 300 ml of 1% paraformaldehyde, and brie¯y refrigerated until examined by ¯ow cytometry. 2.3. Flow cytometric analysis A FACScan ¯ow cytometer (Becton Dickinson) was used to acquire all data. LYSYS II software (Becton Dickinson) was used for data analysis. Ten thousand events were acquired from each sample and FL1 versus cell number histograms were generated for each sample. The percent of BAECs expressing bovine E-selectin was quanti®ed by using the negative controls to determine the placement of the markers as described previously (Van Kampen and Mallard, 1997). Mean ¯uorescence intensity (MFI) values were generated from software-generated statistics of the histograms. Data were compiled in this manner for samples from all time points and in four separate experiments. 2.4. Isolation of RNA from cultured BAECs Total RNA was extracted from all stimulated and unstimulated BAEC cultures using the acid guanidinium thiocyanate method (Chomczynski and Sacchi, 1987) and concentrations were measured by optical density (OD) at 260 nm (GeneQuant II, Pharmacia Biotech Inc, Bale d'Urfe, Quebec, Canada). 2.5. Primers for bovine E-selectin and b2-microglobulin genes A multiple internal control was constructed for use as an exogenous control in quantitative reverse transcription polymerase chain reaction (Q-RT-PCR) for bovine cytokines and adhesion molecules. It was synthesized in a manner similar to that described by Reddy et al. (1996) for use with pig cytokines. It consists of 50 and 30 primer sequences for several bovine cytokines and bovine E-selectin, VCAM-1 and ICAM-1, and includes primer sequences for b2-microglobulin (b2-m) to be used as the housekeeping gene during quanti®cation of mRNA. The primers were designed to amplify two products of different sizes during coampli®cation of control and target RNA via RT-PCR in a single reaction tube. The primer sequences for bovine E-selectin (Genbank accession #L12039) are: (50 ±30 ) forward TGAAGATGGCAGAGACAGAG; reverse CGACTGAGACATCAACGACA, and the product sizes are 259 and 386 bp for control and target, respectively. The primer sequences for b2-m (Genbank accession
C. Van Kampen, B.A. Mallard / Veterinary Immunology and Immunopathology 79 (2001) 151±165
155
#X69084) are: forward TGGCCTTGGTCCTTCTC; reverse CATTCGTCGTGGTAGCTC and the product sizes are 200 and 358 bp for control and target, respectively. These primers were designed based upon known bovine cDNA and mRNA sequences from Genbank. 2.6. Speci®city of primers The speci®city of the primers for bovine E-selectin and b2-m was con®rmed by `hot start' PCR in 25 ml containing 1 PCR buffer, 1.5 mM MgCl2, 200 mM of each dNTPs, 0.12 mM of each sequence-speci®c primer, 100 ng of plasmid DNA and 0.625 units of Taq DNA polymerase (Gibco BRL) for 35 cycles using a thermal cycler (Robocycler Gradient 96, Stratagene-PDI Bioscence). Each cycle included denaturation (948C/1 min), annealing (518C/1 min), extension (728C/1 min) and ®nal extension (728C/10 min) steps. Ampli®ed products were resolved by 2% agarose gel electrophoresis and the DNA bands were compared to the expected size of the target. All PCR reactions were performed under the above conditions unless otherwise stated. 2.7. cDNA synthesis In order to perform a reverse transcriptase reaction with both control and target RNA, the plasmid construct was ®rst subjected to reverse transcription to obtain construct-RNA (c-RNA) which could then be reverse transcribed back into cDNA in the same tube as the target RNA (Reddy et al., 1996). The plasmid construct was ®rst linearized with EcoRI, transcribed at 37±408C for 2 h using SP6 polymerase (Riboprobe II core system, Promega, Madison, WI, SA) to obtain c-RNA and then repeatedly digested with ampli®cation grade DNAse I (Gibco BRL) to remove any plasmid DNA templates. The RT reaction was performed in 10 ml containing 25 mM MgCl2, 1 PCR buffer II (50 mM KCl, 10 mM Tris±HCl, pH 8.3), 1 mM each of dNTPs, 1 U/ml of RNAse inhibitor, 2.5 U/ ml MuLV reverse transcriptase, 2.5 mM random hexamers and 25±50 hg of total RNA(target)/4 106 6 106 copies of c-RNA. For quanti®cation of E-selectin mRNA, the RT reaction was performed in 80 ml volumes containing 400 hg of total RNA and 2 105 copies of c-RNA. Reaction mixtures were sequentially incubated at room temperature for 10 min, 428C/60 min, 998C/8 min and 58C/5 min using a thermal cycler (Robocycler, Stratagene, La Jolla, CA). 2.8. Polymerase chain reaction The quantitative PCR method was modi®ed from that used by Wang et al. (1989) and was previously described (Reddy et al., 1996). An amount of 5 ml of cDNA mixture was serially diluted in 1:3 increments and ampli®cations were done using `hot start' PCR for 33 to 35 cycles using a thermal cycler. Products of PCR were analyzed by electrophoresis in 2% agarose gels stained with ethidium bromide and photographed using a transilluminator. Images of gels were scanned and analyzed using a densitometer (Molecular Analyst, Bio-Rad Laboratories, Mississauga, Ont.), and target and construct data were expressed as log10 of the density. The number of target copies was calculated
156
C. Van Kampen, B.A. Mallard / Veterinary Immunology and Immunopathology 79 (2001) 151±165
utilizing ``best ®t'' standard curves and extrapolated to 1 mg of total RNA as previously described (Reddy et al., 1996; Wang et al., 1989). 2.9. Measurement of shed E-selectin Due to the difference between protein and mRNA expression of E-selectin being most pronounced with rbTNF-a-stimulated BAEC cultures, the measurement of shed Eselectin was performed only with those cultures treated with rbTNF-a. Supernatants from BAEC cultures were harvested after treatment of BAECs cultured in 75 cm2 tissue culture ¯asks with 5 ml of rbTNF-a (10 hg/ml) for the same time periods as previously described. After centrifugation for 10 min at 400 g/48C to remove any cells or debris, supernatants were passed through an Ultrafree-4 ®lter column (Millipore Corporation, Bedford, MA) by adding 4 ml to each column and centrifuging all columns for 10±20 min at 600 g/ 48C until the volume in the columns was reduced by 50%. Concentrates were stored at 208C until used in the ELISA. A capture ELISA was performed with concentrated supernatants from all time points. An Immunolon II 96 well ¯at-bottomed plate (Dynatech, Fisher Scienti®c, Don Mills, Ont.) was coated with 10 mg/ml of EL-246 (anti-human E-selectin) for 3 days at 48C. After washing the plate with 0.05% Tween (Fisher Scienti®c)±PBS, blocking was done with 3% Tween±PBS for 1 h at room temperature, followed by another washing step. An amount of 100 ml of concentrated supernatant from all time points were added to each well and plates were incubated at room temperature for 2 h. After washing the plate using an EL403 plate washer (Biotek, Instruments, Highland Park, Vermont, USA), anti-bovine IgG(H L)-alkaline phosphatase (1/250 dilution, Jackson Labs, Biocan, Mississauga, Ont.) was added to each well and incubated for 2 h at room temperature. A solution of substrate, p-nitrophenyl phosphate disodium tablets (p-NPP, Sigma) was prepared in 10% diethanolamine substrate buffer and added at 100 ml per well. Color was developed for 60 min at room temperature in the dark and then the entire plate was read at 405 and 630 nm on an EL311 Microplate autoreader (Biotek Instruments, Highland Park, Vermont, USA). A 630 nm ®lter was used as a reference ®lter to correct for irregularities in the plastic. The negative control or blank was comprised of coating buffer with blocker and BAEC culture medium (instead of supernatant) plus conjugate. Positive control was a highly concentrated (75%) supernatant from 24 h stimulated cultures. Results were expressed as optical density values and mean OD values and standard deviations were determined, using Student's t-test, from three experiments with each sample tested in triplicate. 2.10. Data analysis The percentages of E-selectin positive BAECs at each time point were averaged from four separate experiments and analyzed for signi®cant differences between time points using Student's t-tests. All samples were analyzed for b2-m and E-selectin mRNA. b2-m values from three replicates at each time point and for untreated or rbTNF-a or LPS-treated BAECs, were averaged and Student's t-tests were performed. Since differences between time points and
C. Van Kampen, B.A. Mallard / Veterinary Immunology and Immunopathology 79 (2001) 151±165
157
between treatments were not signi®cant for b2-m mRNA expression, the mean values for b2-m of 21 samples from each stimulation experiment (n 4) were used as a standard expected b2-m value to normalize each target value using the following equation: expected b2-m/observed b2-m X. Then the expected value for target was obtained by: X observed target value. Values of three replicates of target E-selectin were generated in this manner for each time point and for each stimulus in three separate experiments, averaged, and the tabulated data were analyzed by Student's t-test to compare results between stimuli and over time. Signi®cance was reported at P 0:05. 3. Results 3.1. Surface expression of E-selectin The proportions of BAECs that expressed E-selectin after stimulation with rbTNF-a or LPS are summarized in Fig. 1A. Very few unstimulated BAECs expressed E-selectin, however, as early as 1 h post-stimulation, the proportions of E-selectin cells began to signi®cantly increase. Expression plateaued in rbTNF-a stimulated BAECs, with the ®rst plateau between 3 and 12 h and the second between 18 and 72 h (Fig. 1A). The same phenomenon did not occur with LPS-stimulated BAECs; instead there was a gradual increase from 1±6 h and then a constant proportion of BAECs expressed E-selectin between 6 and 24 h. The mean ¯uorescence intensity peaked at 6 h and then gradually declined until 24 h following stimulation with LPS (Fig. 1B). The MFI gradually but not signi®cantly increased following TNF-a treatment. The difference in proportions of Eselectin BAECs between the two stimuli was not signi®cant except at 6 h (P 0:05). At 72 h, E-selectin expression in the rbTNF-a treatment group remained elevated above baseline. Expression of E-selectin on supramammary arterial endothelial cells did not differ signi®cantly from the expression kinetics on BAECs (data not shown). Viability of the BAECs after stimulation with either LPS or rbTNF-a was greater than 90% at all time points after stimulation. 3.2. Sensitivity of Q-RT-PCR Sensitivity of the assay was determined using a method from Reddy (1998, Ph.D. thesis), involving the use of two experiments that extrapolated the copy number after QRT-PCR of known amounts of starting material at different ratios of control to target DNA molecules. First, the amount of control cDNA molecules was held constant and the target DNA concentrations were varied in ratios of 1:1 to 1:6 using b2-m primers. Extrapolated starting values ranged from 9:62 107 1:22 108 for an expected value or actual starting value of 1:57 106 /mg total RNA (Table 1A). Quanti®cation was possible at up to a six-fold difference between control and target DNA templates with 9.77 hg of total RNA estimated to contain 8:95 106 copies and 3:98 106 copies in 0.326 hg at the 1:6 ratio. Using these extrapolations, it is possible to detect 1100 copies from a 80 ml RT product containing 2 105 copies of control RNA and an unknown copy number of target RNA.
158
C. Van Kampen, B.A. Mallard / Veterinary Immunology and Immunopathology 79 (2001) 151±165
Fig. 1. (A) E-selectin surface expression on cultured bovine endothelial cells. Monolayers were stimulated with recombinant bovine tumor necrosis factor alpha (rbTNF-a) (solid bars) or lipopolysaccharide (LPS) (hatched bars) for 0, 1, 3, 6, 12, 18, 24, or 72 h. Cells were harvested and immunostained with anti-E-selectin (EL246) mAb for ¯ow cytometric analysis. Ten thousand events per time point were analyzed using a FACScan ¯ow cytometer and results were expressed as a mean percentage of E-selectin cells S:D: for each time point in four separate experiments. Signi®cant differences between time points are indicated by different lower (TNF-a) or upper (LPS) case letters. (B) Mean ¯uorescence intensity (MFI) values for E-selectin expression on cultured bovine aortic endothelial cells. Monolayers were stimulated with recombinant bovine tumor necrosis factor alpha (rbTNF-a) (solid bars) or lipopolysacharride (LPS) (hatched bars) for 0, 1, 3, 6, 12, 18, 24, or 72 h. Cells were harvested and immunostained with anti-E-selectin (EL246) mAb for ¯ow cytometric analysis. Ten thousand events per time point were analyzed using a FACScan ¯ow cytometer and results were expressed as an average of MFI S:D: for each time point in four separate experiments. Signi®cant differences between time points are indicated by different lower (TNF-a) or upper (LPS) case letters.
In the second experiment, the minimum copy number was determined by Q-RT-PCR using varying concentrations of control cDNA molecules and a constant amount of targetDNA (Table 1B). Observed values for b2-m ranged between 0:92 107 and 1:83 107 molecules/mg total RNA. Accuracy was determined by comparing the expected and observed copy numbers at different dilutions. Generally, the observed values were always less than expected regardless of whether the target DNA or the control DNA was constant or varied (Table 1).
C. Van Kampen, B.A. Mallard / Veterinary Immunology and Immunopathology 79 (2001) 151±165
159
Table 1 Sensitivity of the quantitative reverse transcriptase polymerase chain reaction (Q-RT-PCR) was determined by subjecting different ratios of standard and target cDNA templates to Q-RT-PCR using b2-m primers and extrapolating the copy number in two separate experimentsa Standard copies (a)
Target (Zg) (b)
Ratio of a:b
Expected value (Zg)
Observed value (Zg ) mean S.D.
Ratio of expected: observed (folds)
A 4.6 108 4.6 108 4.6 108
2.93 0.977 0.326
1:1 1:3 1:6
1.57 107 1.57 107 1.57 107
9.62 106 3.1 104 9.16 106 4.25 105 1.22 107 5.8 106
1.63 1.71 1.29
B 4.6 108 1.53 108 0.511 108
2.93 2.93 2.93
1:1 1:3 1:6
1.57 107 1.57 107 1.57 107
1.08 107 4.6 105 0.92 107 7.5 105 1.23 107 9.1 105
1.45 1.70 1.28
a (A) Copy number determined at a constant amount of control cDNA molecules with varying concentrations of target cDNA templates. (B) Copy number determined at varying concentrations of control cDNA molecules and a constant amount of target cDNA templates. Accuracy was determined by comparing expected and observed values in each experiment.
3.3. b2-microglobulin mRNA Message for b2-m was quanti®ed in BAECs at 0, 1, 3, 6, 12, 18, and 24 h after stimulation in vitro with either LPS, rbTNF-a, or culture medium (Fig. 2). No signi®cant differences were seen for the mean values for b2-m message between unstimulated and stimulated BAEC cultures (P 0:05). Mean values for b2-m mRNA in unstimulated and rbTNF-a stimulated cultures were 2:4 108 and 2:8 108 copies/mg total RNA, respectively. Cultures stimulated with LPS expressed a mean value of 3:31 108 copies/mg
Fig. 2. Beta 2-microglobulin (b2-m) mRNA expression in cultured bovine aortic endothelial cells. Monolayers were stimulated with recombinant bovine tumor necrosis factor alpha (rbTNF-a) (solid bars) or lipopolysaccharide (LPS) (hatched bars) for 0, 1, 3, 6, 12, 18, 24, or 72 h. Cells were harvested and total RNA was extracted from all samples. Expression of b2-m mRNA was analyzed by Q-RT-PCR and copy number was determined per mg of total RNA. Mean copy number S:D: was obtained for three replicates at each time point. No signi®cant differences were found over time or between stimuli.
160
C. Van Kampen, B.A. Mallard / Veterinary Immunology and Immunopathology 79 (2001) 151±165
total RNA of b2-m message. Thus, the expression of b2-m message did not vary signi®cantly over time or between stimuli (Fig. 2). There was some variation in b2-m message between experiments, however, it was not signi®cant, and it was factored into normalization of the E-selectin mRNA data. 3.4. E-selectin mRNA expression In vitro stimulation of bovine BAEC with rbTNF-a resulted in a different pattern of mRNA expression than that induced by LPS (Fig. 3). Unstimulated cultures expressed very small amounts of E-selectin mRNA although the amounts were different between rbTNF-a (2:26 107 copies/mg) and LPS (0:111 107 copies/mg). There was a signi®cant increase (p£ 0.05) in E-selectin mRNA expression 1 h after treatment with rbTNF-a but then the number of copies decreased between 1 and 6 h to values lower (0:865 0:064 107 copies/mg) than those measured in the unstimulated cultures (2:26 0:48 107 copies/mg). After 6 h, the copy number increased gradually until 12 h, then remained constant until 18 h after which E-selectin mRNA expression rapidly increased to very high amounts at 72 h (67 27 107 copies/mg). In cultures treated with LPS, mRNA expression peaked at 6 h (14 8:21 107 copies/mg) and by 24 h, had gradually decreased to amounts similar to those in unstimulated cultures (0:11 0:05 107 copies/mg). 3.5. Shedding of E-selectin Amounts of shed E-selectin were expressed as OD values and are summarized in Fig. 4. Unstimulated BAEC cultures produced very little soluble E-selectin, and OD values were
Fig. 3. E-selectin mRNA expression in cultured bovine aortic endothelial cells. Monolayers were stimulated with recombinant bovine tumor necrosis factor alpha (rbTNF-a) (solid line) or lipopolysaccharide (LPS) (dotted line) for 0, 1, 3, 6, 12, 18, 24, or 72 h. Cells were harvested and total RNA was extracted from all samples. Expression of E-selectin mRNA was analyzed by Q-RT-PCR and copy number was determined per mg of total RNA. Mean copy number S:D: was obtained for three replicates at each time point from three separate experiments. Signi®cant differences over time are indicated by different lower (TNF-a) or upper (LPS) case letters.
C. Van Kampen, B.A. Mallard / Veterinary Immunology and Immunopathology 79 (2001) 151±165
161
Fig. 4. Shed E-selectin in supernatants from cultured bovine aortic endothelial cells. Monolayers were stimulated with recombinant bovine tumor necrosis factor alpha (rbTNF-a) for 0, 1, 3, 6, 12, 18, 24, or 48 h. Supernatants were collected and concentrated by approximately 50% and used in an ELISA. The amount of soluble E-selectin was expressed as OD (405, 630 nm). Each bar represents the mean OD S:D: of triplicate values from three separate experiments. Signi®cant differences over time are indicated by different lower case letters.
not signi®cantly different between unstimulated supernatants and the negative control. There was a signi®cant increase in the amount of soluble E-selectin in stimulated cultures at 1±6 h and then again at 12±48 h, with the highest amount of shed E-selectin present in the 48 h supernatants. 4. Discussion Bovine E-selectin surface expression and mRNA expression were measured in BAECs stimulated with either rbTNF-a or LPS. The proportions of BAECs that expressed Eselectin after stimulation were never higher than 56%. This low percentage is likely to be due to an effect of the trypsin±EDTA solution used to dissociate the cells from the culture ¯ask, that may have removed some of the surface E-selectin molecules. This seems likely based upon the observation that close to 100% of the cells expressed E-selectin by 18 h (data not shown). None of the commercially available dissociation buffers were suitable for detaching the BAECs from the ¯ask. Thus, the total number of cells expressing Eselectin, at all time points are less than 100%, but the effect of each stimulus over time remained constant between experiments. The increased numbers of E-selectin BAEC between 3 and 12 h, supports previous work that showed the induction of bovine Eselectin by LPS and TNF-a between 4 and 12 h (Zund et al., 1996). Changes in MFI were different between LPS and TNF-a, indicating that among other possibilities, the mechanism of induction of E-selectin may differ between stimuli, or the effect may be dose-related. Furthermore, the signi®cant increase in MFI at 6 h after LPS stimulation,
162
C. Van Kampen, B.A. Mallard / Veterinary Immunology and Immunopathology 79 (2001) 151±165
followed by a gradual decrease in surface concentration, paralleled the mRNA kinetics induced by LPS. These ®ndings suggest a correlation between mRNA and the number of E-selectin surface molecules, as the proportion of cells expressing LPS remained constant between 6 and 24 h. However, the lack of variation in MFI in cells treated with TNF-a differed considerably from the mRNA expression kinetics, again indicating a possible difference in the mechanism of induction by LPS and TNF-a. Measurement of bovine E-selectin mRNA expression used a technique involving a plasmid that contained multiple primer sequences for bovine cytokines and adhesion molecules (Reddy et al., 1996). A previous study of bovine cytokine expression used multiple primer pairs (Covert and Splitter, 1995) but they were not all contained within a single construct that can measure several molecules in a single reaction tube. Similar methods have not been previously used to quantify AM expression. Results of the current study indicate a prolonged expression of E-selectin mRNA in cultures stimulated with TNF-a, with peak surface expression at 18±24 h post-stimulation that persisted until 72 h. In cultures stimulated with LPS, the peak mRNA expression occurs at 6 h which is before peak surface expression at 18±24 h. During the early hours after stimulation, fewer mRNA transcripts were required for LPS-induced E-selectin surface expression than for rbTNF-a-induced expression. Later, the proportions of E-selectin BAECs remained fairly constant while mRNA expression was low, in cultures treated with LPS, whereas during the same period, mRNA expression was substantially higher in BAECs treated with rbTNF-a. It is dif®cult to know the reason(s) for these observations and further studies into the mechanism(s) that each stimulus uses to induce AM are necessary. Previously, E-selectin mRNA was measured in human umbilical vein endothelial cells stimulated with TNF-a and skin biopsies of delayed hypersensitivity sites (Meagher et al., 1994), and results indicated the highest amount of E-selectin mRNA in stimulated HUVECs, at 2 h with only a modest decrease at 8 h and sustained mRNA expression at 24 h, whereas skin biopsies contained large amounts of E-selectin mRNA between 6 and 72 h, indicating the potential long term expression of E-selectin mRNA in vivo. This supports the current ®ndings at the earlier time points, to a certain degree, but the signi®cant increase in E-selectin mRNA between 24 and 72 h in this study has not been reported in humans. The choice to use aortic endothelial cells rather than capillary endothelial cells from the mammary gland, which would be most relevant to study Eselectin expression, was based upon the preliminary observation that expression of this adhesion molecule on both types of endothelial cells was similar, and the relative dif®culty of isolating aortic endothelial cells was much lower than that for the capillary endothelial cells. Previous studies report that human E-selectin mRNA and surface protein expression are correlated (Whelan et al., 1991; Bevilacqua et al., 1989). However, this has not been the case for all proteins; for example, the product of certain cytokine mRNA transcription may not always be detectable (Kaspar and Gehrke, 1994). Surface expression did not correlate with mRNA expression of E-selectin in BAEC cultures stimulated with rbTNFa, however, the patterns of expression were more similar in cultures stimulated with LPS. Surface expression and mRNA expression both peaked in LPS-stimulated BAECs at 6 h. The small increase in mRNA expression at 1 h, followed by a signi®cant decrease at 6 h after treatment with rbTNF-a did not correlate with the continued increase in E-selectin
C. Van Kampen, B.A. Mallard / Veterinary Immunology and Immunopathology 79 (2001) 151±165
163
surface expression during the same period. Most studies of bovine adhesion molecules, use a single period of stimulation for their experiments, even with different stimuli. Thus, it is dif®cult to compare these ®ndings with other studies of bovine E-selectin expression. Previous studies of human endothelial cells, showed that ¯uctuations in human E-selectin protein expression correspond to similar changes in mRNA (Whelan et al., 1991; Bevilacqua et al., 1989). However, another report, showed dissimilar patterns between IL1 protein accumulation and IL-1 mRNA expression in peripheral blood mononuclear cells stimulated with LPS or C5a (Kaspar and Gehrke, 1994). There are various possible reasons for the lack of correlation between surface and mRNA expression of E-selectin in the case of TNF-a stimulation. One reason that was suggested by Kaspar and Gehrke (1994) regarding IL-1, is that mRNA translation does not occur, or that the gene product is not detectable. Adhesion molecules, including Eselectin are shed from the cell surface after or during stimulation (Grif®n et al., 1990; Wyble et al., 1997). Thus, the difference between surface expression and mRNA expression of bovine E-selectin in rbTNF-a-stimulated BAECs, may be due, at least in part to E-selectin being shed from the surface of the BAECs during the later time points, and at a much higher rate than earlier time points. This may cause the amount of mRNA transcripts to continue to rise to keep up with the demand of the protein synthesis machinery. Therefore, the amount of shed E-selectin in the supernatants of rbTNF-astimulated BAECs was measured using ELISA. The amount of E-selectin in the culture supernatants increased steadily over time. Subsequent experiments could investigate if the turnover rate of mRNA vis a vis stability or degradation, has an in¯uence on the patterns observed in this study, or if detection of intracellular E-selectin in cells stimulated with rbTNF-a might also account for the large number of transcripts between 18 and 72 h. E-selectin protein is not, however, known to be stored intracellularly (Bevilacqua et al., 1987). Finally, measuring shed E-selectin in LPS-stimulated culture supernatants might help to account for the different patterns of mRNA expression between the stimuli. In conclusion, this study showed that both TNF-a and LPS induced expression of Eselectin on BAEC but that surface protein and mRNA expression were not well correlated. Furthermore, the expression of bovine E-selectin differs from human E-selectin in that expression is sustained for at least 72 h in TNF-a, and 24 h in LPS-stimulated BAEC cultures, whereas E-selectin expression in HUVECs peaks at 6 h and returns to baseline by 24 h (Meagher et al., 1994). E-selectin expression on BAEC is, however, closer to the expression kinetics documented in the pig in which expression of porcine E-selectin was prolonged up to 48 h post-stimulation with IL-1a or TNF-a (Tsang et al., 1995). With clear differences between bovine E-selectin mRNA expression in BAECs treated with different stimuli, it will be important to determine if this occurs in vivo. Acknowledgements We thank Dr. N.R. Reddy for constructing the bovine construct and for his instruction in the art of Q-RT-PCR. We also thank Sophia Lim for performing the ELISA experiments and Marion Wilke for her assistance with primary culturing of the BAECs.
164
C. Van Kampen, B.A. Mallard / Veterinary Immunology and Immunopathology 79 (2001) 151±165
We thank the staff of the Animal and Poultry Science abbattoir for assisting in the recovery of bovine aortas: We gratefully acknowledge the individuals that provided mAbs (M. Jutila and C. Howard) and rbTNF-a (Lorne Babiuk) as gifts for this research. This research was supported by grants to Bonnie Mallard from the Natural Sciences and Engineering Research Council of Canada, and Ontario Ministry of Agriculture and Rural Affairs (OMAFRA). References Benekli, M., Gullu, I.H., Tekuzman, G., Savas, M.C., Hayran, M., Hascelik, G., Firat, D., 1998. Circulating intercellular adhesion molecule-1 and E-selectin levels in gastric cancer. Br. J. Cancer 78, 267±271. Bevilacqua, M.P., Nelson, R.M., 1993. Selectins. J. Clin. Invest. 91, 379±387. Bevilacqua, M.P., Pober, J.S., Mendrick, D.L., Cotran, R.S., Gimbrone Jr., M.A., 1987. Identi®cation of an inducible endothelial-leukocyte adhesion molecule. Proc. Natl. Acad. Sci. 84, 9238±9242. Bevilacqua, M.P., Stengelin, S., Gimbrone Jr., M.A., Seed, B., 1989. Endothelial leukocyte adhesion molecule-1: an inducible receptor for neutrophils related to complement regulatory proteins and lectins. Science 243, 1160±1165. Chomczynski, P., Sacchi, N., 1987. Single-step method of RNA isolation by acid guanidinium thiocyanatephenol-chloroform extraction. Anal. Biochem. 162, 156±159. Covert, J., Splitter, G., 1995. Detection of cytokine transcriptional pro®les from bovine peripheral blood mononuclear cells and CD4 lymphocytes by reverse transcriptase polymerase chain reaction. Vet. Immunol. Immunopathol. 49, 39±50. Doukas, J., Pober, J.S., 1990. IFN-g enhances endothelial activation induced by tumor necrosis factor but not IL1. J. Immunol. 145, 1727±1731. Ferran, C., Cooper, J.T., Brostjan, C., Stroka, D.M., Millan, M.T., Goodman, D.J., Bach, F.H., 1996. Expression of a truncated form of the human p55 TNF-receptor in bovine aortic endothelial cells renders them resistant to human TNF. Transplant. Proc. 28, 618±619. Graber, N., Copal, T.V., Wilson, D., Beall, L.D., Polte, T., Newman, W., 1990. T cells bind to cytokine-activated endothelial cells via a novel, inducible sialoglycoprotein and endothelial leukocyte adhesion molecule-1. J. Immunol. 145, 819±830. Grif®n, J.D., Spertini, O., Ernst, T.J., Belvin, M.P., Levine, H.B., Kanakura, Y., Tedder, T.F., 1990. Granulocytemacrophage colony stimulating factor and other cytokines regulate surface expression of the leukocyte adhesion molecule-1 on human neutrophils, monocytes and their precursors. J. Immunol. 145, 576±579. Jutila, M.A., Bargatze, R.F., Kurk, S., Warnock, R.A., Ehsani, N., Watson, S.R., Walcheck, B., 1994. Cell surface P- and E-selectin support shear-dependent rolling of bovine gd T cells. J. Immunol. 153, 3917± 3928. Jutila, M.A., 1996. gd T cell/endothelial cell interactions. Vet. Immunol. Immunopathol. 54, 105±110. Kaspar, R.L., Gehrke, L., 1994. Peripheral blood mononuclear cells stimulated with C5a or lipopolysaccharide to synthesize equivalent levels of IL-1 beta mRNA show unequal IL-1 beta protein accumulation but similar polyribosome pro®les. J. Immunol. 153, 277±286. Kayal, S., Jais, J.P., Aguini, N., Chaudiere, J., Labrousse, J., 1998. Elevated circulating E-selectin, intercellular adhesion molecule 1, and von Willebrand factor in patients with severe infection. Am. J. Respir. Crit. Care Med. 157, 776±784. Leitner, G., Shoshani, E., Berman, A., Trinin, Z., Feingold, D., Einkler, M., Shpigel, N.Y., Saran, A., 1996. Immune response of the mammary gland in dairy cows after experimental infection with Escherichia coli or Staphylococcus aureus. In: 3rd international Mastitis Seminar Proceedings, May 28±June 1, Tel Aviv, Israel. Meagher, L., Mahiouz, D., Sugars, K., Burrows, N., Norris, P., Yarwood, H., Becker-Andre, M., Haskard, D.O., 1994. Measurement of mRNA for E-selectin, VCAM-1 and ICAM-1 by reverse transcription and the polymerase chain reaction. J. Immunol. Methods 175, 237±246. Ohno, M., Ichikawa, H., Coe, L., Kvietys, P.R., Granger, D.N., Alexander, J.S., 1997. Soluble selectins and ICAM-1 modulate neutrophil-endothelial adhesion and diapedesis in vitro. In¯ammation 21 (3), 313±324.
C. Van Kampen, B.A. Mallard / Veterinary Immunology and Immunopathology 79 (2001) 151±165
165
Reddy, N.R., Wilkie, B.N., Mallard, B.A., 1996. Construction of an internal control to quantitate multiple porcine cytokine mRNAs by RT-PCR. BioTechniques 21, 868±875. Shimizu, Y., Newman, W., Gopal, T.V., Horgan, K.J., Graber, N., Beall, L.D., Van Seventer, O.A., Shaw, S., 1991. Four molecular pathways of T cell adhesion to endothelial cells: roles of LFA-1, VCAM-1 and ELAM-1 and changes in pathway hierarchy under different activation conditions. J. Cell Biol. 113, 1203± 1212. Sopp, P., Kwong, L.S., Howard, C.J., 1996. Identi®cation of bovine CD14. Vet. Immunol. Immunopathol. 52, 323±328. Springer, T.A., 1994. Traf®c signals for lymphocyte recirculation and leukocyte emigration: the multistep paradigm. Cell 76, 301±314. Springer, T.A., 1995. Traf®c signals on endothelium for lymphocyte recirculation and leukocyte emigration. Ann. Rev. Physiol. 57, 827±872. Tsang, Y.T.M., Stephens, P.E., Licence, S.T., Haskard, D.O., Binns, R.M., 1995. Porcine E-selectin: cloning and functional characterization. Immunology 85, 140±145. Van Kampen, C., Mallard, B.A., 1997. Effects of peripartum stress and disease on bovine lymphocyte subsets. Vet. Immunol. Immunopathol. 59 (1/2), 79±92. Walcheck, B., Watts, G., Jutila, M.A., 1993. Bovine gd T cells bind E-selectin via a novel glycoprotein receptor: ®rst characterization of a lymphocyte/E-selectin interaction in an animal model. J. Exp. Med. 178, 853±863. Wang, A.M., Doyle, M.V., Mark, D.F., 1989. Quantitation of mRNA by the polymerase chain reaction. Proc. Natl. Acad. Sci U.S.A. 86, 9717±9721 [published erratum: 1989. Proc. Natl. Acad. Sci. U.S.A. 87, 2865]. Wenisch, C., Myskiw, D., Parschalk, B., Hartmann, T., Dam, K., Graninger, W., 1994. Soluble endotheliumassociated adhesion molecules in patients with Graves' disease. Clin. Exp. Immunol. 98, 240±244. Whelan, J., Ghersa, P., Hooft van Huijsduijen, R., Gray, J., Chandra, G., Talabot, F., Delamarter, J.F., 1991. An NFkB-like factor is essential but not suf®cient for cytokine induction of endothelial leukocyte adhesion molecule 1 (ELAM-1) gene transcription. Nucleic Acids Res. 19, 2645±2653. Wyble, C.W., Hynes, K.L., Kuchbhotla, J., Marcus, B.C., Hallahan, D., Oewertz, B.L., 1997. TNF-alpha and IL1 upregulate membrane-bound and soluble E-selectin through a common pathway. J. Surg. Res. 73, 107±112. Zund, G., Nelson, D.P., Neufeld, E.J., Dzus, A.L., Bischoff, J., Mayer, J.E., Colgan, S.P., 1996. Hypoxia enhances stimulus-dependent induction of E-selectin on aortic endothelial cells. Proc. Natl. Acad. Sci. 93, 7075±7080.