Effect of apoptosis on phagocytosis, respiratory burst and CD18 adhesion receptor expression of bovine neutrophils

Domestic Animal Endocrinology 22 (2002) 37–50

Effect of apoptosis on phagocytosis, respiratory burst and CD18 adhesion receptor expression of bovine neutrophils K. Van Oostveldt a , M.J. Paape b , H. Dosogne a , C. Burvenich a,∗ a

b

Department of Physiology, Biochemistry and Biometrics, Faculty of Veterinary Medicine, Ghent University, Salisburylaan 133, 9820 Merelbeke, Belgium Immunology and Disease Resistance Laboratory, United States Department of Agriculture, Beltsville, MD, USA Received 22 March 2001; accepted 12 August 2001

Abstract Polymorphonuclear neutrophil leukocytes (PMN) play an important role in intramammary defense against infections by Escherichia coli. During mastitis, PMN are confronted with various inflammatory mediators that can modulate their function. In severely diseased cows, increased concentrations of lipopolysaccharide (LPS) and tumor necrosis factor (TNF)-␣ (TNF-␣) are detected in plasma. Binding of LPS to membrane bound CD14 molecules on monocytes cause release of inflammatory mediators such as TNF-␣. Because apoptosis of PMN promotes resolution of inflammation and because the LPS and TNF-␣ response in milk and blood is related to the severity of E. coli mastitis, the effect on apoptosis of bovine PMN of increased concentrations LPS and TNF-␣ was studied together with the functionality of apoptotic PMN. Bovine PMN apoptosis, as determined with annexin-V, was induced with high concentrations of either LPS (1000 and 10,000 ng/mL) or TNF-␣ (10,000 ng/mL) in whole blood following a 6 h incubation at 37◦ C. The apoptosis inducing effect of LPS on PMN was not inhibited following coculture with either anti-bovine TNF-␣ or anti-ovine CD14 monoclonal antibodies. When compared to controls, apoptotic PMN had a similar level of CD18 expression but lacked phagocytic and respiratory burst activity. This is the first study reporting the effects of apoptosis on bovine PMN function. These functional impairments in apoptotic PMN could be important in contributing to the establishment of intramammary infection. Well functioning PMN could finally determine the severity of mastitis following an invasion of bacteria in the mammary gland. © 2002 Elsevier Science Inc. All rights reserved.

∗

Corresponding author. Tel.: +32-9-2647321; fax: +32-9-2647499. E-mail address: [email protected] (C. Burvenich).

0739-7240/02/$ – see front matter © 2002 Elsevier Science Inc. All rights reserved. PII: S 0 7 3 9 - 7 2 4 0 ( 0 1 ) 0 0 1 1 5 - 1

38

K. Van Oostveldt et al. / Domestic Animal Endocrinology 22 (2002) 37–50

1. Introduction The polymorphonuclear neutrophilic leukocytes (PMN) are the major defense against bacterial infection of the bovine mammary gland [1]. In severe Escherichia coli infections of the mammary gland, the bacteria release sufficient lipopolysaccharide (LPS, endotoxin) to be detected in both milk and blood [2]. Lipopolysaccharide binds to LPS binding protein (LBP) which then binds to membrane bound CD14 receptors on PMN and monocytes/macrophages [3,4]. In response to this binding, inflammatory mediators such as TNF-␣ are released into the surrounding environment. The occurrence of LPS in plasma during coliform mastitis has been associated with an excessive plasma TNF-␣ response (up to 10 ng/mL) and milk TNF-␣ response (up to 200 ng/mL) in severely diseased cows [2,3]. Another well-characterized LPS receptor is the leukocyte integrin CD18 [5]. The CD18 integrin is also involved in nonopsonic recognition of LPS as a constituent of E. coli, as well as other Gram-negative bacterial cell membranes by phagocytes [6]. In response to the production and release of TNF-␣, interferons and interleukins-induced migration (diapedesis) of PMN from blood through the blood milk barrier into the milk [7]. In the mammary gland, PMN phagocytose the invading bacteria and kill them through both oxygen-independent and oxygen-dependent mechanisms [8,9]. During phagocytosis, nonmitochondrial oxygen consumption increases (respiratory burst) with the formation of a variety of reactive oxygen species [10,11]. Because of the systemic impact of inflammation, interest is drawn toward the participation and possible role of the nervous and endocrine system within this complex network of immunologic events during the local inflammatory response. Endotoxin and cytokines induce fever and alter neurotransmitter activity in the brain and hormone secretion by the pituitary and other endocrine glands [12]. During E. coli mastitis increases in the concentration of insulin-like growth factor-1 (IGF-1) were observed in milk. This release was preceded by an increase in TNF-␣ [12]. During mastitis, the respiratory burst capacity of PMN activated with phorbol myristate acetate (PMA) is strongly decreased [13] whereas circulating PMN that were not activated with PMA showed a spontaneous increase of the respiratory burst capacity during mastitis [14]. The capacity of PMN to diapedese though an in vitro blood milk barrier is also decreased during E. coli mastitis [15]. Presently, it is unclear whether these effects on PMN function during mastitis can be attributed to LPS or secondarily to the release of TNF-␣. In addition, dose-related responses of PMN to both LPS and TNF-␣ are poorly characterized. Studies have shown that recombinant human TNF-␣ stimulates respiratory burst activity of bovine PMN as measured by luminol-dependent chemiluminescence, superoxide anion generation and hydrogen peroxide production [16]. Further, diapedesis of human PMN in response to C5a is decreased following a preincubation of human PMN with TNF-␣ [16]. Resolution of an acute inflammatory response requires removal of PMN and their toxic products from the site of infection through a process called apoptosis. Failure to achieve this, results in chronic, persistent inflammation and severe tissue damage. Incubation of human PMN with increasing concentrations of recombinant human TNF-␣ was shown to accelerate apoptosis [17,18]. In contrast, LPS slowed the rate of apoptosis of isolated human PMN [19].

K. Van Oostveldt et al. / Domestic Animal Endocrinology 22 (2002) 37–50

39

In this study, effects of LPS and TNF-␣ on PMN apoptosis was studied using a whole blood model as described by Van Oostveldt et al. [20]. Because functional impairment function of aging human PMN was associated with apoptosis [21], a second study was conducted to examine expression of CD18 cell surface adhesion molecules, respiratory burst activity and phagocytosis by apoptotic bovine PMN. Because of the importance of apoptosis in the resolution of an inflammation and because the LPS and TNF-␣ response in milk and blood is related to the severity of E. coli mastitis [2], we wanted to know how increased concentrations of LPS and TNF-␣ can affect apoptosis of bovine PMN and how this can affect the functionality of these PMN. Well functioning PMN will also finally determine the severity of mastitis following an invasion of bacteria in the mammary gland. 2. Materials and methods 2.1. Reagents Annexin-V-FITC (Boehringer Mannheim GmbH, Mannheim, Germany), phycoerythrinlabeled annexin-V (annexin-V-PE) (Alexis Biochemicals, San Diego, CA), anti-ovine CD14 monoclonal antibody (mAb) (Serotec Inc., Raleigh, NC, cross-reactivity with bovine cells is shown by Brodersen et al. [23]), anti-human CD18 mAb (Serotec, cross-reactivity with bovine cells is shown by Brodersen et al. [23]), anti-bovine TNF-␣ mAb (M.J. Paape, USDA, Beltsville, MD, USA), dihydrorhodamine 123 (DHR) (Molecular Probes, Junction City, OR, USA), E. coli strain P4, serotype 032:H37, recovered from a clinical case of bovine mastitis and kindly provided by Dr. J. Bramley, Burlington, VT, E. coli endotoxin (0111:B4 Sigma, St. Louis, MO), human recombinant TNF-␣ (rTNF-␣) (Serotec, high similarities in activity of human rTNF-␣ and bovine rTNF-␣ were detected by Adams and Czuprynski [22]), N-ethyl maleimide (NEM) (Sigma), PI (Sigma), phorbol myristate acetate (PMA) (Sigma), 0.01 M phosphate buffered 0.85% saline, pH 7.3 (PBS). 2.2. Animals Holstein dairy cows in their second to sixth lactation were selected from the Ghent University dairy herd (Biocentrum Agri-Vet, Melle, Belgium) and the USDA dairy herd, Beltsville, MD. All cows were free from intramammary infection by mastitis pathogens. Mean milk somatic cell count from all quarters were determined to be <100,000 cells/mL. 2.3. Blood sampling Blood was collected from the jugular vein by venipuncture in vacutainer tubes containing 125 IU heparin as anticoagulant (Laboratoire EGA, Nogent-Le-Roi, France). Smears were prepared from whole blood and stained with Hemacolor (Merck, D-64293 Darmstadt 1, Germany). Differential microscopic counts were determined by counting 100 cells. In this study, no distinction was made between eosinophils and neutrophils when apoptosis is studied.

40

K. Van Oostveldt et al. / Domestic Animal Endocrinology 22 (2002) 37–50

Earlier studies have shown apoptosis rates of neutrophils sometimes differ from eosinophils [24]. Therefore, cows with eosinophil counts greater than 10% were excluded from this study. 2.4. Flow cytometric cell sorting of annexin-V positive PMN and PI positive PMN Following a 6 h incubation of 100 ␮L blood at 37◦ C, PMN were labeled with FITC-annexin-V (dilution 1/100) and cellular DNA was stained using PI (final concentration of 1 ␮g/mL) in the presence of Ca2+ -rich medium (10 mM Hepes/NaOH, pH 7.4, 140 mM NaCl, 2.5 mM CaCl2 ) as previously described [20]. Erythrocytes were lysed with 600 ␮L formic acid (1.4 mg/mL) to better visualize the leukocytes on the dot plot scatter pattern during flow cytometric analysis. Isotonicity was restored by addition of 265 ␮L of buffer (6.0 g sodium carbonate, 14.5 g sodium chloride, 31.3 g sodium sulfate/L). Using a FACScan sorter (Beckton Dickinson, Immunocytometry Systems, San Jose, CA, USA) a minimum of 10,000 PMN positive for annexin-V-FITC and negative for PI and 10,000 PMN positive for both stains were sorted out and collected in PBS containing 50% bovine serum. Cytospin smears (Shandon Belgique, 1060 Brussels) of the sorted PMN were prepared and stained with Hemacolor (Merck). Morphology of the sorted PMN was studied using light microscopy (magnification, 1000×) (Carl Zeiss Inc., Thornwood, NC). 2.5. Effect of LPS and TNF-α on apoptosis Blood (100 ␮L) of 10 animals (n = 10) was incubated with varying concentrations of either human recombinant TNF-␣ (0, 10, 25, 50, 100 and 10,000 ng/mL) for 6 h at 37◦ C and 5% CO2 or E. coli endotoxin (0, 1, 100, 1000 and 10,000 ng/mL). Annexin-V-FITC in combination with PI was used to determine apoptosis [20]. Because the apoptosis inducing effect of LPS may be induced indirectly by TNF-␣, 100 ␮L blood of 10 animals (n = 10) was preincubated for 0.5 h at 37◦ C with anti-bovine TNF-␣ mAb (dilution 1/100) followed by 6 h of incubation with LPS under the same conditions. In parallel, 100 ␮L blood of seven animals (n = 7) was preincubated for 30 min at 37◦ C with anti-ovine CD14 mAb (dilution 1/10) followed by a 6 h incubation with LPS under the same conditions as already described. A flow cytometric assay was performed using annexin-V and PI to determine apoptosis [20]. 2.6. CD18 expression in combination with annexin-V and PI labeling Following an incubation period of 6 h at 37◦ C, 100 ␮L blood of three animals (n = 3) was incubated for 30 min at room temperature with FITC-labeled anti-human CD18 mAb (dilution 1/10) to study expression of the CD18 receptor on bovine PMN. After washing twice with PBS, annexin-V-PE was added (dilution 1/50) in a Ca2+ -rich medium [20]. The percentage of annexin-V-PE positive PMN was determined by flow cytometry. A gate was set on the annexin-V-PE positive PMN to determine the percentage of CD18 positive PMN in the gated population together with the mean fluorescence intensity of the gated population. Viability was determined in an additional 100 ␮L of blood with PI (final concentration, 1 ␮g/mL).

K. Van Oostveldt et al. / Domestic Animal Endocrinology 22 (2002) 37–50

41

Erythrocytes were lysed and remaining leukocytes were stabilized before flow cytometric analysis as described previously. 2.7. Respiratory burst and phagocytosis in combination with annexin-V and PI labeling To evaluate respiratory burst activity of annexin-V positive PMN, a dual color flow cytometric assay was developed. Following incubation of 100 ␮L blood of five animals (n = 5) for 6 h at 37◦ C, dihydrorhodamine (5 ␮M) was added. After 5 min of incubation at 37◦ C in a shaking water bath, E. coli (PMN:E. coli ratio 1:25) or PMA (0.1 ␮M) was added and incubated for an additional 20 min. Annexin-V-PE was added and incubated for 10 min at room temperature. N-ethylmaleimide (NEM, 0.1 mM) was added to stop phagocytosis of E. coli and samples were stored on ice until assayed by flow cytometry. Just before analysis, erythrocytes were lysed and samples were stabilized. To evaluate the phagocytic capacity of annexin-V positive PMN, 100 ␮L heparinized blood of three animals (n = 3) was incubated for 6 h at 37◦ C, followed by incubation for 20 min at 37◦ C with PI-labeled E. coli (PMN:E. coli ratio 1:40). Annexin-V-FITC was then added and incubated for 10 min at room temperature. NEM (0.1 mM) was added to stop phagocytosis and samples were stored on ice until measured by flow cytometry. Viability of PMN was determined by flow cytometry with PI (1 ␮g/mL, final concentration) in a separate 100 ␮L sample of blood. 2.8. Flow cytometry A FACScan flow cytometer (Becton Dickinson) was used in all the experiments except for the study of CD18 expression on annexin-V positive PMN where a Coulter Profile II flow cytometer (Coulter Electronics Inc., Hialeah, FL) was used. Linear amplification of the forward scatter (FS) and side scatter (SS) light signals was set with logarithmic amplification of the fluorescence signals. The 488 nm excitation wavelength was used. For each sample, 5000 PMN were acquired. PMN were selected for analysis by gating on the FS and SS dot plot. FITC and PI with PE fluorescence was measured through 530/30 and 585/42 band pass filters, respectively. Compensation for FITC–% PI and PI–% FITC was 2.0 and 36.2%, respectively, and FITC–% PE and PE–% FITC was 2.0 and 37%. Data were acquired and processed using CELL Quest software (Becton Dickinson). 2.9. Statistical analysis All statistical procedures, means, standard deviations and standard errors of the means were computed using statistical software (Statistix® NH Analytical Software, Tallahassee, FL, USA) according to Snedecor and Cochram [25]. Results are expressed as means plus or minus standard errors of the means. Normality was tested using the Wilk-Shapiro test. Statistical analysis was carried out using the Kruskal Wallis nonparametric test for pairwise comparison of the means with cows as the randomized factor when the population was not normally distributed. If the population was normally distributed, results were performed using analysis of variance with cow as the randomized factor, concentration as the fixed factor and

42

K. Van Oostveldt et al. / Domestic Animal Endocrinology 22 (2002) 37–50

cow × concentration as the interaction. Least significant differences were used to compare means. Statistical significance was accepted at P < 0.05, P < 0.01 and P < 0.001.

3. Results 3.1. Flow cytometric cell sorting of annexin-V positive PMN and PI positive PMN Fig. 1A shows normal living PMN with typical multilobed nuclei. PMN that were recognized by FITC-labeled annexin-V and sorted out by flow cytometry, are presented in Fig. 1B and C. Nuclei of the annexin-V positive but not of the PI negative PMN (Fig. 1B) were picnotic. However, some of the cells showed condensed nuclei suggesting a later stage of apoptosis (marked with Roman numerals I and II, respectively). Fig. 1C presents PMN that were recognized by FITC-labeled annexin-V and positive for PI. These PMN showed loss of membrane integrity. The PMN that were positive for annexin-V and negative for PI had decreased SS indicating decreased granularity. The PMN that were positive for annexin-V and also positive for PI showed decreased SS and decreased FS indicating diminishment in size (results not shown). 3.2. Effect of LPS and TNF-α on apoptosis Low concentrations of LPS (up to 100 ng/mL) did not induce apoptosis (Fig. 2A). However, high concentrations of LPS (1000 and 10,000 ng/mL) caused 6.90 ± 1.46 and 7.62 ± 1.66% of the PMN to become apoptotic, respectively, compared to 3.71 ± 0.91% apoptotic PMN in the control samples (P < 0.01). Incubation of whole blood with increasing concentrations of recombinant human TNF-␣ for 6 h at 37◦ C only induced an increase (P < 0.01) in the percentage of apoptotic PMN at the highest concentration of TNF-␣ (10,000 ng/mL) (6.01 ± 1.26 versus control 2.52 ± 0.56, P < 0.05). It was hypothesized that the apoptosis inducing effect of LPS was caused indirectly by TNF-␣. However, coincubating blood samples with LPS and anti-bovine TNF-␣ mAb did not negate the apoptosis inducing effect of LPS (Table 1). Furthermore, coincubating whole blood with LPS and anti-ovine CD14 mAb did not negate the apoptosis inducing effect of LPS on PMN (Table 2). Table 1 Percentage of annexin-V positive PMN in whole blood following an incubation with and without anti-bovine TNF-␣ mAb in combination with increasing concentrations of LPS (6 h, 37◦ C) LPS concentrations (ng/mL)

Control

Anti-bovine TNF-␣ mAb

0 100 10000

1.84 ± 0.28 2.92 ± 0.51 4.14 ± 0.58

2.61 ± 0.70 2.77 ± 0.53 4.06 ± 0.55

Data are means ± SEM (n = 10).

K. Van Oostveldt et al. / Domestic Animal Endocrinology 22 (2002) 37–50

43

Fig. 1. (A) Living PMN with normal morphology. (B) PMN that were positive for FITC-labeled annexin-V are the apoptotic cells with picnotic nuclei (I) and condensed nuclei, a later stage of apoptosis (II). (C) PMN that were positive for FITC-labeled annexin-V and positive for PI are necrotic PMN with a degraded plasma membrane. The bar represents 2 ␮m.

44

K. Van Oostveldt et al. / Domestic Animal Endocrinology 22 (2002) 37–50

Fig. 2. Percentage of apoptotic PMN in whole blood measured with annexin-V and PI following a 6 h incubation at 37◦ C with varying concentrations of LPS (A) or recombinant human TNF-␣ (B). Data are means ± SEM (n = 10).

3.3. CD18 expression in combination with annexin-V and PI labeling Incubation of whole blood for 6 h at 37◦ C increased the mean fluorescence intensity (MFI) of CD18 in the total PMN population indicating that the amount of CD18 receptors on PMN was increased. In freshly taken blood, 97.8±0.7% of the PMN expressed the CD18 receptor (MFI = 9.2 ± 1.3). Following a 6 h incubation of whole blood at 37◦ C the MFI of annexin-V positive PMN increased (MFI = 20.6 ± 5.0) but was not different when compared to nonapoptotic PMN after 6 h of incubation (MFI = 14.6±2.7) (Fig. 3). The percentage of annexin-V positive PMN increased following a 6 h incubation of whole blood when compared to 0 h (14.6 ± 7.1 Table 2 Percentage of annexin-V positive PMN in whole blood following an incubation with and without anti-ovine CD14 mAb in combination with increasing concentrations of LPS (6 h, 37◦ C) LPS concentrations (ng/mL)

Control

Anti-ovine CD14 mAb

0 1000 10000

4.9 ± 1.06 9.37 ± 2.07 13.37 ± 2.98

4.49 ± 1.06 9.02 ± 3.02 13.51 ± 4.25

Data are means ± SEM (n = 7).

K. Van Oostveldt et al. / Domestic Animal Endocrinology 22 (2002) 37–50

45

Fig. 3. CD18 expression on unstained and annexin-V stained PMN population following a 6 h incubation at 37◦ C expressed as MFI.

versus 2.4 ± 1.1) (results not shown). From these results, it is important to recognize that the annexin-V positive PMN include apoptotic and necrotic PMN. This can imply that any change in the expression of the CD18 receptor of the annexin-V positive population can be attributed to changes in CD18 expression of the necrotic, the apoptotic PMN or both. However, the mean percentage of necrotic PMN was not changed following the 6 h incubation at 37◦ C, only the percentage of apoptotic PMN was increased. Therefore, any change in the expression of the CD18 receptor is suggested to be related to the apoptotic PMN population and not to the necrotic PMN population. 3.4. Phagocytosis and respiratory burst in combination with annexin-V and PI labeling Following a 6 h incubation at 37◦ C of 100 ␮L whole blood, 60.7 ± 3.4% of the PMN that were labeled with the FITC-labeled annexin-V (apoptotic + necrotic) showed a phagocytosis capacity toward E. coli compared to 59.8 ± 9.4% for nonlabeled PMN (Table 3). However, within the annexin-V positive population that includes apoptotic and necrotic PMN, the percentage of PMN that showed no phagocytosis capacity was significantly higher (P < 0.05) than the percentage of necrotic PMN determined in a separate blood sample under the same conditions (results not shown). Within the population of annexin-V negative PMN, 90.1 ± 2.1 and 54.2 ± 8.7% of the cells showed a respiratory burst activity, compared to 76.8 ± 2.3 and 26.9 ± 2.5% of the annexin-V positive PMN, in response to stimulation with PMA or E. coli, respectively (Table 4). Again, within the annexin-V positive population that includes apoptotic and necrotic PMN, Table 3 Percentage of phagocytosis within annexin-V negative and positive PMN populations Phagocytosis (%) Annexin-V negative PMN Annexin-V positive PMN (apoptotic + necrotic PMN) Data represent mean ± SEM (n = 3).

59.8 ± 9.4 60.7 ± 3.4

46

K. Van Oostveldt et al. / Domestic Animal Endocrinology 22 (2002) 37–50

Table 4 Percentage of respiratory burst capacity within annexin-V negative and positive PMN population following stimulation with PMA or E. coli Respiratory burst (%)

Annexin-V negative PMN Annexin-V positive PMN (apoptotic + necrotic PMN)

After PMA stimulation

After E. coli stimulation

90.1 ± 2.1 76.8 ± 2.3

54.2 ± 8.7 26.9 ± 2.5

Data represent mean ± SEM (n = 5). P < 0.05 compared to the annexin-V negative PMN population. P < 0.01 compared to the annexin-V negative PMN population.

the percentage of PMN that showed no respiratory burst activity in response to PMA or E. coli was higher than the percentage of necrotic PMN determined in a separate blood sample under the same conditions (P < 0.05 and P < 0.01, respectively).

4. Discussion The phagocytic cells of the bovine mammary gland are PMN and macrophages. During mastitis, PMN migrate to the site of infection in response to chemical messengers that are released by activated inflammatory cells within the mammary gland [11,26]. The clearance of activated PMN from the site of infection is central to the resolution of an acute inflammatory response [27,28]. Spontaneous apoptotic cell death of bovine PMN has been demonstrated in vitro [20]. This process of cell death is distinct from necrosis. Necrosis results in the release of toxic cell components that are damaging to tissue [29]. Many procedures have been developed to measure apoptosis with the majority being flow cytometric. Recently, a flow cytometric procedure was adapted to detect apoptosis of bovine PMN in whole blood using annexin-V in combination with PI [20]. Two distinct morphologic types of apoptotic PMN could be seen (Fig. 1B, numerals I and II) indicating that annexin-V is exposed on both early and late stages of apoptosis. A picnotic (Fig. 1B, numeral I) and condensed nucleus (Fig. 1B, numeral II) are common morphologic features of apoptosis [28,30]. Based on the features of these PMN recognized by annexin-V, their capacity to exclude PI and their accompanying morphology compared to living PMN, indicate the suitability of this flow cytometric procedure for determining apoptosis of PMN in whole blood. Attenuation of in vitro PMN apoptosis has been demonstrated to be mediated by a number of inflammatory proteins that are commonly found at the site of inflammation. LPS was found to delay apoptosis in isolated human PMN [18,19]. In our experiments using whole blood, LPS accelerated PMN apoptosis. It is known that LPS stimulates monocytes in whole blood to release cytokines such as TNF-␣ and that these cytokines can in turn prime or stimulate PMN [3,4]. TNF-␣ was previously shown to induce apoptosis in various cell types [31–33], including human PMN [18,34]. However, in our experiments, much higher concentrations of human recombinant TNF-␣ were necessary to induce apoptosis in whole blood compared to the experiments of Takeda et al. [34] and Watson et al. [18]. This may be related to the

K. Van Oostveldt et al. / Domestic Animal Endocrinology 22 (2002) 37–50

47

fact that in our study human recombinant TNF-␣ was used to induce apoptosis in bovine PMN. Unfortunately, there are no sources of recombinant bovine TNF-␣. Interestingly, the apoptotic inducing effect of LPS was not inhibited by anti-bovine TNF-␣ mAb. This suggests that LPS may have a direct effect on PMN apoptosis. However, in response to LPS not only TNF-␣ but other inflammatory mediators such as interleukines (IL), interferon-␥ (IFN-␥) and granulocyte-macrophage colony stimulating factors (GM-CSF) are also released [3,4,35]. In vitro tests on human PMN have shown that IL-1, IFN-␥ [36] and GM-CSF [37] delay PMN apoptosis. TNF-␣ and IL-6 [17] induce PMN apoptosis. The net biologic effect of these competing compounds is unknown. Also, the apoptosis inducing effect of LPS was not inhibited by anti-bovine CD14 mAb. The LPS–LBP complex binds to CD14 receptor on monocytes causing release of inflammatory mediators which in turn prime or stimulate PMN. Only high concentrations of LPS (1000 and 10,000 ng/mL) could induce PMN apoptosis in our experiments. When monocytes are incubated with high concentrations of LPS, other LPS receptors on monocytes may be involved [38]. Yang et al. [39] showed that Toll-like receptor 2 is a signaling receptor that is activated by LPS and is dependent on LBP. From these experiments, we conclude that the LPS inducing effect on bovine PMN apoptosis remains unclear as to whether this effect is a direct or indirect effect via the release of inflammatory mediators by monocytes. In the second part of this study, functional characteristics of apoptotic PMN were examined. Many of the effector functions of human and bovine PMN such as phagocytosis and respiratory burst are modulated by the adhesive state of the cell [40]. Engagement of ␤2 integrins serves to amplify the destructive potential of PMN in response to triggering stimuli [41,42]. From our results, increased expression of the CD18 receptor in a 6 h incubated population was not ascribed to an increased expression of the CD18 receptor on apoptotic PMN and suggests that other factors of activation are involved that cause upregulation of CD18. According to Jones and Morgan [43] apoptosis was associated with a reduced expression of CD18. This is in contrast with the results of Dransfield et al. [40] who showed that the expression of the ␤2 integrin CD11b/CD18 was maintained on apoptotic PMN, similar to our results. In vitro studies on human PMN reported no binding of apoptotic PMN to fibrinogen via CD11b/CD18 in spite of maintained expression of CD11b/CD18 [40]. Such research has not yet been performed with bovine PMN. A maintained expression of the CD18 receptor on apoptotic PMN as seen in this study, could indicate that ␤2 integrins might be involved in the rapid and efficient recognition and phagocytosis of apoptotic PMN by macrophages. In this study, it was demonstrated that annexin-V negative and positive PMN had a similar capacity to phagocytose E. coli. The population of annexin-V positive PMN-included apoptotic and necrotic PMN. In our study, the percentage of annexin-V positive PMN that showed no phagocytosis capacity was significantly higher than the percentage of necrotic PMN determined in a separate 100 ␮L blood sample. From these results it can be suggested that only late apoptotic PMN showed no phagocytosis capacity. Within the annexin-V positive PMN population, the percentages of PMN that showed a respiratory burst activity after PMA or E. coli stimulation was significantly lower than the percentages of PMN that showed a respiratory burst activity within the annexin-V negative PMN population. Again, in this study, the percentage of the annexin-V positive PMN that showed no respiratory burst activity was higher than the percentage of necrotic PMN determined in a separate blood sample. From

48

K. Van Oostveldt et al. / Domestic Animal Endocrinology 22 (2002) 37–50

these results concerning the respiratory burst activity, it can be concluded that a decreased respiratory burst activity after stimulation with PMA or E. coli is ascribed to the annexin-V positive PMN population including early and late apoptotic PMN. In human PMN, a close relationship between apoptosis and functional loss was observed [19,21]. Whyte et al. [21] could not detect decreased respiratory burst after stimulation with PMA in an aged population of PMN where more than 80% of the PMN were apoptotic. The correlation between apoptosis and the loss of cell function may be the result of defects in the cytoskeleton. In human PMN, apoptosis was related to loss of spontaneous spreading and shape change [19,21]. Until the current study, there was no data about the effect of inflammatory mediators on apoptosis of bovine PMN and the consequences of apoptosis on PMN function. High concentrations of LPS and TNF-␣-induced apoptosis in bovine PMN. It could not be shown if the apoptosis inducing effect of LPS in bovine PMN was due to release of inflammatory mediators such as TNF-␣ by monocytes. Inflammatory mediators other than TNF-␣ could be involved in apoptosis of bovine PMN. Apoptotic bovine PMN, as determined with annexin-V, were able to maintain CD18 expression but not their phagocytic and respiratory burst functions. This functional impairment could be important in the establishment of intramammary infection.

Acknowledgments The authors express their gratitude to Prof. Dr. Plum (Ghent University) for the use of the sorter and to Mrs. De Smet for her technical help with the sorter. This study was supported by Bijzonder Onderzoeksfonds (Grant no. 01111299), Belgian Ministry of Agriculture (Grant S/5871) and Fonds voor Wetenschappelijk Onderzoek (Grant no. 31504200). References [1] Paape MJ, Schultze WD, Guidry AJ. Development of natural defense mechanisms. Kiel Milchwirtsch Forschungber 1985;37:447–54. [2] Hoeben D, Burvenich C, Trevisi E, Bertoni G, Hamann J, Bruckmaier RM, Blum JW. Role of endotoxin and TNF-␣ in the pathogenesis of experimentally-induced coliform mastitis in periparturient cows. J Dairy Res 2000;67:503–14. [3] Semnani MJ, Kabbur MB, Jain NC. Activation of bovine neutrophil functions by interferon-␥, TNF alpha, and interleukin-1. Comp Heamatol Int 1993;3:81–8. [4] Wright SD, Ramos RA, Tobias PS, Ulevitch RJ, Mathison JC. CD14, a receptor for complexes of lipoploysaccharide (LPS) and LPS binding protein. Science 1990;249:1431–6. [5] Paape MJ, Lilius EM, Wiitanen PA, Kontio MP, Miller RH. Intramammary defense against infections induced by E. coli in cows. Am J Vet Res 1996;57:477–82. [6] Wright S, Jong M. Adhesion-promoting receptors on human macrophages recognize E. coli by binding to lipopolysaccharide. J Vet Res 1986;164:1876–88. [7] Van Oss KJ. Phagocytosis: an overview. Meth Enzymol 1986;132:3–8. [8] Burvenich C. Introduction to acute mastitis in ruminants: some general pathophysiological aspects of acute inflammation. Vl Dierg Tijdschr 1985;54:1–8. [9] Paape MJ, Guidry AJ, Jain NC, Miller RH. Leukocytic defense mechanism in the udder. Vl Dierg Tijdschr 1991;62(Suppl 1):95–109. [10] Bellavite P. The superoxide forming enzymatic system of phagocytes. Free Rad Biol Med 1988;4:225–61.

K. Van Oostveldt et al. / Domestic Animal Endocrinology 22 (2002) 37–50

49

[11] Burvenich C, Paape MJ, Hill AW, Guidry A, Miller RH, Heyneman R, Kremer WDJ, Brand A. Role of the neutrophil leukocyte in the local and systemic reactions during experimentally-induced E. coli mastitis in cows immediately after calving. Vet Quarter 1994;16:45–50. [12] Burvenich C, Paape MJ, Hoeben D, Dosogne H, Massart-Leen AM, Blum J. Modulation of the inflammatory reaction and neutrophil defense of the bovine lactating mammary gland by growth hormone. Dom Anim Endocrinol 1999;17:149–59. [13] Heyneman R, Burvenich C, Vercauteren R. Interaction between the respiratory burst activity of neutrophil leukocytes and experimentally-induced E. coli mastitis in cows. J Dairy Sci 1990;73:985–94. [14] Van Oostveldt K, Burvenich C, Moreira da Silva F, Massart-Leen AM. Respiratory burst activity in activated and unstimulated isolated bovine blood neutrophils during experimentally-induced E. coli mastitis. J Dairy Res 1999;66:375–83. [15] Smits E. Neutrophil diapedesis in primary cell culture systems of the bovine mammary gland. PhD thesis, Ghent University, 1998. [16] Sample AK, Czuprynski CJ. Priming and stimulation of bovine neutrophils by recombinant human interleukin-1 (IL-1) alpha and TNF alpha. J Leukocute Biol 1991;49:107–15. [17] Afford SC, Pongracz J, Stockley RA, Crocker J, Burnett D. The induction by human IL-6 of apoptosis in the promonocytic cell line U967 and human neutrophils. J Biol Chem 1992;267:21612–6. [18] Watson RWG, Redmond HP, Wang JH, Bouchier-Hayes D. Bacterial ingestion, TNF-␣, and heat-induced programmed cell death in activated neutrophils. Shock 1996;5:47–51. [19] Lee A, Whyte MKB, Haslett C. Inhibition of apoptosis and prologation of neutrophil functional longevity by inflammatory mediators. J Leukocyte Biol 1993;54:283–8. [20] Van Oostveldt K, Dosogne H, Burvenich C, Paape MJ, Brochez V, Van den Eeckhout E. Flow cytometric procedure to detect apoptosis of bovine polymorphonuclear leukocytes in whole blood. Vet Immunol Immunopathol 1999;70:125–33. [21] Whyte MKB, Meagher LC, MacDermot J, Haslett C. Impairment of function in aging neutrophils is associated with apoptosis. J Immunol 1993;150:5124–34. [22] Adams JL, Czuprynski CJ. Bacterial lipopolysaccharide induces release of TNF-␣ from bovine peripheral blood monocytes and alveolar macrophages in vitro. J Leukocyte Biol 1990;48:549–56. [23] Brodersen R, Bijlsma F, Gori K, Jensen KT, Chen W, Dominguez J, Haverson K, Moore PF, Saalmuller A, Sachs D, Slierendrecht WJ, Stokes C, Vaino O, Zuckermann F, Aasted B. Analysis of the immunological cross-reactivities of 213 well characterized monoclonal antibodies with specificities against various leukocyte surface antigens of human and 11 animal species. Vet Immunol Immunopathol 1998;64:1–13. [24] Meagher LC, Cousin JM, Seckl JR, Haslett C. Opposing effects of glucocorticoids on the rate of apoptosis in neutrophilic and eosinophilic granulocytes. J Immunol 1996;156:4422–9. [25] Snedecor LM, Cochram WG. Stastical methods. Ames, IA: Iowa State University Press, 1968. [26] Paape MJ, Wergin WP, Guidry AJ, Pearson RE. Leukocytes—the second line of defense against invading mastitis pathogens. J Dairy Sci 1979;62:135–53. [27] Haslett C. Resolution of acute inflammation and the role of apoptosis in the tissue fate of granulocytes. Clin Sci 1991;83:639–48. [28] Savill J, Haslett C. Granulocyte clearance by apoptosis in the resolution of inflammation. Semin Cell Biol 1995;6:385–93. [29] Schwartzman BA, Cidlowski JA. Apoptosis: the biochemistry and molecular biology of programmed cell death. Endocrine Rev 1993;2:133–51. [30] Hebert MJ, Takano T, Holthofer H, Brady HR. Sequential morphologic events during apoptosis of human neutrophils. J Immunol 1996;157:3105–15. [31] Schulte-Hermann R, Bursch W, Low-Baselli A, Wagner A, Grasl-Kraupp B. Apoptosis of the liver and its role in hepatocarcinogenesis. Cell Biol Toxicol 1997;13:339–48. [32] Schulz A, Bauer G. Selective effect of TNF on transformed versus nontransformed cells: nonselective signal recognition but differential target cell response. Anticancer Res 2000;20:3435–42. [33] Sidoti-de Fraisse C, Rincheval V, Risler Y, Mignotte B, Vayssiere JL. TNF-␣ activates at least two apoptotic signaling cascades. Oncogene 1998;17:1639–51.

50

K. Van Oostveldt et al. / Domestic Animal Endocrinology 22 (2002) 37–50

[34] Takeda Y, Watanabe H, Yonehara S, Yamashita T, Saito S, Sendo F. Rapid acceleration of neutrophil apoptosis by TNF alpha. Int Immunol 1993;5:691–4. [35] Janeway CA, Travers P. Immunobiology: the immune system in health and disease. Oxford: Blackwell Scientific, 1994. [36] Collota F, Re F, Polentarutti N, Sozzani S, Montovani A. Modulation of granulocyte and programmed cell death by cytokines and bacterial products. Blood 1992;80:2012–20. [37] Cox G, Gauldie J, Jordana M. Brochial epithelial cell derived cytokines (G-CSF and GM-CSF) promote the survival of peripheral blood neutrophils in vitro. Am J Cell Mol Biol 1992;7:507–17. [38] Wright SD. Multiple receptors for endotoxin. Curr Opin Immunol 1991;3:83–90. [39] Yang RB, Mark MR, Gray A, Huang A, Xie MH, Zhang M, Goddard A, Wood WI, Gurney AL, Godowski PJ. Toll-like receptor-2 mediates lipopolysaccharide-induced cellular signaling. Nature 1998;395:284–8. [40] Dransfield I, Stocks SC, Haslett C. Regulation of cell adhesion molecule expression and function associated with neutrophil apoptosis. Blood 1995;85:3264–73. [41] Nathan CF. Neutrophil activation on biological surfaces. J Clin Invest 1987;80:1550–60. [42] Nathan C, Farber C, Sanchez E, Kabbash L, Asch A, Gailit J, Wright SD. Cytokine-induced respiratory burst of human neutrophils: dependence on extracellular matrix proteins and CD11/CD18 integrins. J Cell Biol 1989;109:1341–52. [43] Jones J, Morgan BP. Apoptosis is associated with reduced expression of complement regulatory molecules, adhesion molecules and other receptors on polymorphonuclear leukocytes: functional relevance and role in inflammation. Immunology 1995;86:651–60.