- Email: [email protected]
Influence of Escherichia coli and Arcanobacterium pyogenes isolated from bovine puerperal uteri on phenotypic and functional properties of neutrophils
Veterinary Microbiology 79 (2001) 351±365
In¯uence of Escherichia coli and Arcanobacterium pyogenes isolated from bovine puerperal uteri on phenotypic and functional properties of neutrophils H. Zerbea,*, C. Oûadnika, W. Leiboldb, H.J. Schuberthb a
Clinic for Bovine Obstetrics and Gynaecology, School of Veterinary Medicine Hannover, Bischofsholer Damm 15, D-30173 Hannover, Germany b Immunology Unit, School of Veterinary Medicine Hannover, Bischofsholer Damm 15, D-30173 Hannover, Germany
Received 22 June 2000; received in revised form 17 October 2000; accepted 17 October 2000
Abstract When cows develop endometritis after birth, Escherichia coli and Arcanobacterium pyogenes are usually the most prominent bacteria present in bovine uterine lochial secretions. A. pyogenes alone is rarely found in the course of a disturbed puerperium. This was con®rmed in this study, since average and high-grade uterine contaminations were always associated with the presence of both bacteria. The contamination grade was positively correlated with uterine polymorphonuclear granulocyte (PMN) numbers and negatively correlated with blood PMN numbers. Whether E. coli and A. pyogenes affect the phenotype and function of bovine PMN in a similar or differential way was subject to in vitro studies. PMN were tested in the presence of washed bacterial fragments or culture supernatants taken as a source for soluble and/or secreted bacterial products. Fragments and soluble products differed only quantitatively in their effects on PMN. Usually, long-time exposure (24 h) of PMN to fragments induced the strongest effects. Accelerated death of granulocytes was only moderately induced by both E. coli and A. pyogenes products. Both E. coli and A. pyogenes products induced the enhanced expression of a membrane molecule detected by mAb IL-A110 and of CD11b. Expression of other surface structures remained largely unchanged (MHC class I, CD11c). Functional parameters of PMN (phagocytosis; generation of reactive oxygen species, ROS; antibody-independent cellular cytotoxicity, AICC) generally declined after pre-incubation for 24 h with products of E. coli or A. pyogenes. Interestingly, soluble products of A. pyogenes stimulated the phagocytosis of PMN. However, co-incubation with E. coli products abrogated this
*
Corresponding author. Tel.: 49-511-856-7611; fax: 49-511-856-7691. E-mail address: [email protected] (H. Zerbe). 0378-1135/01/$ ± see front matter # 2001 Elsevier Science B.V. All rights reserved. PII: S 0 3 7 8 - 1 1 3 5 ( 0 0 ) 0 0 3 6 8 - 0
352
H. Zerbe et al. / Veterinary Microbiology 79 (2001) 351±365
stimulatory effect. The results supply evidence for similar modes of action of the gram-negative E. coli and the gram-positive A. pyogenes on bovine PMN. Alterations in PMN function and phenotype are mainly triggered by direct contact between bacterial fragments and PMN. Inhibition experiments with polymyxin B demonstrated that E. coli-mediated effects were not solely due to the action of lipopolysaccharide. The dominant functional depression of neutrophils by E. coli products strengthens the suggestion that the earlier appearance of E. coli in the uterus may support the coinfection of this organ by A. pyogenes at later times. # 2001 Elsevier Science B.V. All rights reserved. Keywords: Cattle; Bacteria; Uterus; Escherichia coli; Arcanobacterium pyogenes; Neutrophils
1. Introduction Most health problems in high producing cows occur in the peripartum period. Together with vaginitis, endometritis causes economic losses accounting for 50% of all reproduction disorders in Germany (Lotthammer and Wittkowski, 1994). Lochial secretions taken as early as 1 day after birth contain mostly Escherichia coli and streptococci (Ahlers and Grunert, 1993). Streptococci appear to be members of the normal uterine bacterial ¯ora (Huszenicza et al., 1999). In some cases also staphylococci and Proteus spp. can be isolated (Vandeplassche, 1981; Olson et al., 1984; Noakes et al., 1991; Huszenicza et al., 1999). Later in the early puerperium Arcanobacterium pyogenes especially appears in lochial secretions (Olson et al., 1984; Noakes et al., 1991; Ahlers and Grunert, 1993). In cows with normal puerperium intrauterine bacteria are usually eliminated until the 10th day postpartum (Boitor et al., 1976). In summary, gram-negative bacteria, especially E. coli, seem to dominate in the uterus within the ®rst days after calving (Hussain et al., 1990; Huszenicza et al., 1999). Later after birth, A. pyogenes and gram-negative anaerobic bacteria are found in uterine secretions of animals suffering from severe puerperal endometritis, especially in combination with retention of the fetal membranes (Bekana et al., 1994). The appearance of A. pyogenes seems to be an important indicator of a puerperal disorder resulting in a disturbed fertility (Olson et al., 1984; Huszenicza et al., 1999). Since the local activity of uterine PMN is of major importance for the antimicrobial defence in this organ after birth (Vandeplassche and Bouters, 1976; Hussain, 1989), it is intriguing to examine possible inhibitory effects of products of the dominating bacterial species. However, little is known about local effects of bacterial contamination on granulocytes. Lipopolysaccharide (LPS) from E. coli was able to upregulate adhesion molecules (ICAM-1, ELAM-1) on human endothelial cells in vitro, which facilitated adhesion of granulocytes (Dobrina et al., 1995). Upregulation of CD11b (complement receptor 3) was also seen on lung-immigrating PMN after intratracheal administration of LPS in rats (Kermarrec et al., 1998). In cattle, the increased surface expression of CD11b on PMN seemed to be insuf®cient for a sustained phagocytosis during E. coli-induced mastitis in early postpartum dairy cows. Dosogne et al. (1997) observed a biphasic upregulation of CD11b expression on PMN but phagocytosis was decreased 12 h
H. Zerbe et al. / Veterinary Microbiology 79 (2001) 351±365
353
after infection. It is speculated that LPS-induced activation of PMN and macrophages results in a hyper-reactive state of the cells causing tissue damage in addition to the development of septicemic shock (Martich et al., 1991; Goya et al., 1994; Dalrymple et al., 1996). In¯uences of A. pyogenes on granulocytes have been rarely described. One strain of A. pyogenes (formerly Actinomyces pyogenes, ATCC 8164) secreted a hemolysin which was cytotoxic for PMN and for pTK2 kidney cells (Ding and LaÈmmler, 1996). It is unclear, however, whether A. pyogenes isolated from uterine secretions do secrete such leukotoxic factors. The reasons for an increased susceptibility for infections of cows after birth are not fully understood. Functional capacities (e.g. phagocytosis, generation of ROS) of PMN compared between pre- and postpartum periods were lower in the period after calving. This correlated with a diminished expression of functionally important PMN surface molecules, e.g. CD11b in the blood (Cai et al., 1994; Zerbe et al., 1998). Whether these changes are the cause of liver function disorders (Morrow et al., 1979; Kaneene et al., 1997; Zerbe et al., 2000), hormonal imbalances (Roth et al., 1983; Chacin et al., 1990) or due to the contact of PMN with products of intrauterine bacteria remains an open question. Here, we report on investigations dealing with the latter hypothesis. It was of special interest to determine whether secreted products and/or membrane-bound structures of different bacteria isolated from bovine uterine secretions exert common or even contrary effects on function and phenotype of bovine polymorphs. 2. Materials and methods 2.1. Animals and cells For collection of uterine ¯uid Holstein-Friesian postpartum cows 3 and 6 years of age (n 33, 550±650 kg) without retained placenta, but with disturbed early puerperium (delayed uterus involution and acute endometritis) were supplied by the Clinic for Bovine Obstetrics and Gynaecology of the Hannover School of Veterinary Medicine, Germany. Animals were kept in stalls under similar conditions. Diagnosis of delayed uterus involution and acute endometritis was based on the evaluation of lochia (presence of foul-smelling, malodorous cervical discharge) and rectal palpation (enlarged, atonic uterus). Neutrophilic granulocytes were separated from venous blood drawn from the V. jugularis externa into vacutainer tubes containing EDTA (Becton Dickinson). Blood donors were three clinically healthy ovariectomised German Blackpied cows (4±5 years). Blood samples were kept at room temperature and separation of PMN was performed within 2 h. Blood was centrifuged (1500g, 5 min, 48C) and contaminating erythrocytes from buffy coat leukocytes were removed by two successive hypotonic lysing steps. The cell suspension was then separated over a discontinuous density gradient (Percoll1, Pharmacia) to obtain pure granulocyte populations (Zerbe et al., 1996). The purity was consistently 95% as determined ¯ow cytometrically.
354
H. Zerbe et al. / Veterinary Microbiology 79 (2001) 351±365
2.2. Collection and identi®cation of bacteria Uterine secretions were taken manually from the cervix uteri or from the corpus uteri under aseptical conditions at days 4 or 9 postpartum. The outer genital was washed with water and soap, and sterile gloves were used. E. coli and A. pyogenes were isolated, identi®ed and subcultivated for experiments by means of standard bacteriological techniques (Bisping and Amtsberg, 1988). Isolates of E. coli obtained from 11 animals were typed serologically. Since virulence factors of E. coli isolated from uteri cannot be determined at present, typing characteristics for enteropathogenic E. coli were utilised. Three (of 14) isolates could not be recultivated after storage. Eight samples were untypable. Serotype O101:K28 which is considered to be pathogenic for cattle was present in two out of three samples (Selbitz, 1992). In one animal this was the only serotype isolated, the other animal harboured two serotypes in addition (O101:K30 and O101:K32). From the third animal, the serotype O141:H25 was isolated. Due to the relative dominance of O101:K28, this serotype was chosen for the subsequent preparation of E. coli products. All strains of A. pyogenes isolates were û-hemolytic. Biochemical typing was not performed. Puri®ed and lyophilised bacteria were reconstituted in 2 ml Ampuwa medium (Fresenius) and cultured for 24 h at 378C in 100 ml culture medium: RPMI 1640 supplemented with 2 mol/l L-glutamine, 15 mmol/l HEPES and 18 mmol/l NaHCO3 (medium and all supplements, Biochrom) without addition of serum. All buffers and media were free of contaminating LPS as determined with the limulus assay (Endotoxintest ``mini'', Company Schulz, Bad Krozingen, Germany). Numbers of colony forming units (CFU) were determined by standard techniques. 2.3. Preparations of bacterial products and co-incubation with PMN in vitro After centrifugation of cultured bacteria (15 000g, 20 min, 48C), the supernatant was removed, sterile-®ltered (0.5 mm) and stored in aliquots at ÿ1008C. The sedimented bacteria were washed twice in sterile PBS (15 000g, 20 min, 48C) and were ®nally disrupted into fragments (French Pressure Cells and Press1, American Instruments). Complete killing of the bacteria was judged by cultivation on blood agar (24 h, 378C, 5% CO2 in air). The concentrations of supernatants and bacterial fragments are expressed as equivalents of CFU present in the suspensions before preparation. E. coli preparations contained 3 mg/ml LPS (equivalent to 2 108 CFU=ml). LPS contents in A. pyogenes preparations were below the detection limit (<0.001 mg/ml). Co-incubations with PMN 2 105 per well were made in 96 well round-bottom microtiter plates for 3 or 24 h at 378C in a humidi®ed atmosphere with 5% CO2 in air. Bacterial fragments or solubles were added at ®nal ratios between 1:1 and 1:1000 (PMN:CFU equivalents). These ratios corresponded to the situation found ex vivo (up to 1:2000; data not shown). Parallel PMN preparations were co-incubated with puri®ed LPS (serotype 0111:B4, Sigma; 3 mg/ml) instead of bacterial fragments or solubles. All set-ups were made in triplicate. Sterile culture medium served as controls.
H. Zerbe et al. / Veterinary Microbiology 79 (2001) 351±365
355
2.4. Cell counts, differentiation, determination of cellular viability Total blood leukocyte concentrations were determined in a cell counting chamber after staining with TuÈrk's solution (Merck). Differential counts (granulocytes, lymphocytic cells and monocytes) were determined by ¯ow cytometry (FACScan1, Becton Dickinson) after acquisition of at least 5000 leukocytes. Cellular suspensions contained propidium iodide (2 mg/ml, Calbiochem) for the assessment of cellular viability. Total numbers of viable PMN after co-culture with bacterial products were determined with a quantitative ¯ow cytometric method (SCDA, single cell dilution assay, Pechhold et al., 1994) which was adapted for the bovine system (Hendricks et al., 2000). 2.5. Phenotyping The procedure of staining cells with primary and secondary antibodies has been described in detail elsewhere (Zerbe et al., 1996). The phenotypic analysis was performed ¯ow cytometrically. The relative (mean) ¯uorescence intensity served to describe the relative expression density of the studied cell surface molecules on isolated PMN. Primary, bovine-speci®c monoclonal antibodies (mAbs) used were: mAb Bo1 (anti-MHC class I, Schuberth et al., 1992), mAb IL-A15 (anti-boCD11b, Naessens et al., 1993), mAb IL-A16 (anti-boCD11c, Howard and Morisson, 1991), mAb IL-A99 (anti-boCD11a, Naessens et al., 1993), and mAb IL-A110 (granulocyte-speci®c, Naessens et al., 1993). 2.6. Functional assays for neutrophilic granulocytes Phagocytosis assay. Nonviable streptococci in suspension (Omnisorb1, Calbiochem) were labelled with ¯uoresceinisothiocyanat (FITC, Sigma): FITC buffer (0.7 ml) (3 mg FITC in 7 ml 0.1 mol/l Na2CO3 and 0.1 mol/l NaCl, pH 9.2) and 5 ml NaN3 (10% (w/v) in H2O) were mixed with 0.3 ml Omnisorb1 suspension for 20 h at RT in the dark under constant rotation. Bacteria were washed three times with PBS (15 000g, 1 min) and ®nally adjusted to 2 108 cells=ml. Aliquots were opsonised by addition of bovine serum and stored at ÿ208C. Bacteria were mixed with PMN at ratios between 1:1 and 1000:1 and incubated for 60 min at 378C. By means of ¯ow cytometry, PMN with ingested bacteria were identi®ed due to their increased green ¯uorescence (Zerbe et al., 1996). Generation of ROS. The general procedure followed that described by EmmendoÈrffer et al. (1990). PMN were stimulated with phorbol myristate acetate (PMA, 300 nmol/l; Sigma). Thereafter, the cells were loaded with the non¯uorescent dye dihydrorhodamin 123 (DHR 123, 15 mg/ml; Molecular Probes) which, in case of ROS generation after stimulation, was converted into the green ¯uorescent rhodamin 123 via oxidation catalysed by cellular myeloperoxidase. The amount of ROS generation was measured ¯ow cytometrically by determination of the relative (mean) ¯uorescence intensity. Antibody-independent cellular cytotoxicity (AICC). Bovine lymphoblastoid cells (``Anna TA1'', Schuberth et al., 2000) served as target cells. Targets were washed twice in PBS (80g, 8 min) and adjusted to 1 106 cells=ml culture medium (RPMI 1640, with
356
H. Zerbe et al. / Veterinary Microbiology 79 (2001) 351±365
10% (v/v) fetal calf serum). After labelling with the dye D275 (Mobitec; 1.25 mg/ml cell suspension, 30 min, RT), the targets were washed twice in PBS (80g, 8 min). Finally, targets were adjusted to 1 106 cells=ml culture medium and mixed with effector PMN at a ratio of 30:1 (PMN:targets). Co-incubation was carried out for 4 h at 378C in the presence of 30 nmol/l PMA. Cell viability of target cells was evaluated by ¯ow cytometry after addition of propidium iodide (see above). 2.7. Statistical analysis All assays were set up in triplicate. All statistical procedures (means, SD and student's t-test) were performed using SAS software. 3. Results 3.1. Bacteria and leukocyte numbers Based on the grade of contamination of uterine secretions, 33 samples were grouped in four categories (Table 1). It turned out that middle grade (<100 bacteria per examination) and high-grade contamination exclusively was associated with the presence of E. coli and A. pyogenes. In nine of these 21 samples both E. coli and A. pyogenes were detectable; ®ve contained only E. coli and seven only A. pyogenes. With the exception of one case all A. pyogenes-positive isolates (mixed with E. coli or pure) were found at day 9 postpartum, all E. coli-pure samples at day 4 postpartum. Mean PMN concentrations in middle- and high-contaminated uterine secretions (table groups 3 and 4) were signi®cantly higher compared to not- or low-contaminated samples. Conversely, PMN numbers in the peripheral blood were lower in the middle- and high-contaminated animals compared to bacteriologically negative or low-contaminated cows.
Table 1 Bacterial contamination of uterine secretions and its in¯uence on PMN concentrations in uterus and blood of puerperal disordered cows without retained placenta Uterine bacteriology (group)
Bacteriaa
Cattle (No.)
PMN (106 per ml)b Uterus
Negative (1) Staphylococcus spp., Streptococcus spp., Proteus spp., Klebsiella spp. (2) E. coli and A. pyogenes (3) E. coli and A. pyogenes (4)
0 1±10 11±100 >100
4 8 8 13
1.40.7 a 5.86.9 c 7.85.3 b,e 30.930.2 d,f
Blood 4:3 1:2 g,i 2:9 1:5 2:7 1:0 h 1:9 1:1 k
a Bacteria were counted microscopically (oil immersion, 1250) under standardised conditions (numbers in 10 different areas). b PMN concentrations were determined in uterine secretions and in venous blood of cows in the early puerperium. Significances: a vs. b, c vs. d and e vs. f were different at P < 0:05; g vs. h were different at P < 0:05 and i vs. k at P < 0:001.
H. Zerbe et al. / Veterinary Microbiology 79 (2001) 351±365
357
Fig. 1. Viability of PMN after incubation with different bacterial preparations. Separated PMN of three ovariectomised cattle 2 105 per well were incubated for 24 h with fragments or soluble products of E. coli or A. pyogenes. Percentage of viable PMN were determined ¯ow cytometrically with the SCDA assay.
3.2. E. coli and A. pyogenes negatively affect viability of PMN Products of both bacteria dose-dependently accelerated the death of neutrophils in vitro. Compared to medium controls about 70% of the neutrophils remained vital after 24 h (Fig. 1). Whereas bacterial fragments did not differ in their potency between E. coli and A. pyogenes, soluble constituents of E. coli were slightly more potent compared to those obtained from A. pyogenes (Fig. 1). 3.3. Phenotypic changes of PMN induced by E. coli and A. pyogenes Differences in expression densities were regarded as biologically signi®cant, if the MFI of a given membrane molecules differed at least by 50% compared to the appropriate controls. CD11b and a structure de®ned by mAb IL-A110 were signi®cantly upregulated. CD11b expression was up to 3.5 times higher expressed after incubation of PMN with A. pyogenes fragments (Fig. 2A); the mAb IL-A110-de®ned molecule was expressed 4.7 times higher after incubation with E. coli fragments (Fig. 2A). Bacterial fragments seemed to be more potent than soluble components. An interesting exception from this rule was seen with CD11c (Fig. 2A): only E. coli and A. pyogenes solubles induced divers reactions, whereas fragments had no effect. Neither product modulated the expression of MHC class I molecules (data not shown). Enhanced expression of CD11b and the mAb IL-A110-de®ned molecule was also induced by pure LPS (Fig. 2B). This effect was sensitive to polymyxin B treatment of the cells which also abrogated the expression-enhancing effect of E. coli solubles. However, polymyxin B was less potent in the reduction of CD11b expression induced by E. coli fragments (Fig. 2B) and the enhanced expression of the IL-A110-de®ned molecule
358
H. Zerbe et al. / Veterinary Microbiology 79 (2001) 351±365
Fig. 2. Modulation of PMN surface expression by bacterial products. Separated PMN of three ovariectomised cattle 2 105 per well were incubated for 24 h with fragments or soluble products of E. coli or A. pyogenes. Expression densities (mean ¯uorescence intensities, MFI) of different surface structures were determined ¯ow cytometrically after indirect immuno¯uorescence. (A) Relative changes of the MFI (in percent to the controls, mean and standard deviation of three animals). Positive or negative changes >50% compared to controls were considered signi®cant (dashed lines). Standard deviations were omitted from the graph (coef®cient of variance 6%). (B) MFIs of CD11b and IL-A110 (granulocyte-speci®c molecule) after incubation of PMN with LPS (3 mg/ml), E. coli-derived fragments or E. coli-derived soluble products. Fragments and solubles corresponded to 2 108 CFU E. coli/ml containing 3 mg/ml LPS. Parallel cultures contained polymyxin B (3 mg/ml) to inhibit LPS induced changes of expression.
H. Zerbe et al. / Veterinary Microbiology 79 (2001) 351±365
359
Fig. 3. Products of E. coli and A. pyogenes depress functional capacities of bovine PMN. Blood PMN 2 105 per well of three ovariectomised cattle were incubated for 24 h in the presence of fragments or soluble products of E. coli or A. pyogenes in different concentrations. AICC, the capacity to phagocytose or to generate ROS was determined ¯ow cytometrically. Positive or negative changes >50% compared to controls (cell incubated without bacterial products) were considered signi®cant (dashed lines). Standard deviations were omitted from the graph (coef®cient of variance 6%).
remained nearly unaffected. A. pyogenes-induced changes of surface expression were insensitive to polymyxin B (data not shown). 3.4. Functional changes of PMN induced by E. coli and A. pyogenes Generation of ROS was downregulated most prominently by E. coli products (Fig. 3) in a dose-dependent manner. Again, E. coli fragments were more potent than solubles. Interestingly, A. pyogenes products were less potent and nearly failed to modulate this function. Only higher concentrations of E. coli and A. pyogenes fragments decreased the ability of PMN to ingest bacteria signi®cantly (Fig. 3). The effects of soluble bacterial preparations were contradictory. Whereas E. coli products slightly inhibited the phagocytosis, it was slightly enhanced by soluble constituents of A. pyogenes. The
360
H. Zerbe et al. / Veterinary Microbiology 79 (2001) 351±365
Fig. 4. Effects of bacterial fragments, and solubles and mixtures of both on functional parameters of bovine PMN. PMN of three ovariectomised animals were pre-cultivated for 24 h in the presence of fragments (F), soluble constituents (S), combination of both F S or combinations of fragments and solubles of E. coli and A. pyogenes (last bar). Fragments and solubles were adjusted to equivalents of 2 108 CFU. Cells were also preincubated with pure LPS (serotype O111:B4, 3 mg/ml). Cells incubated in medium alone served as controls (dashed line). PMN functions (AICC, generation of ROS, phagocytosis) were determined ¯ow cytometrically (means standard deviations).
effects of E. coli or A. pyogenes products on the AICC of PMN did not reach signi®cance (Fig. 3). All tested functional properties of PMN were nearly unaffected in the presence of pure LPS (3 mg/ml) (Fig. 4). In experiments where fragments and solubles of bacteria were mixed (Fig. 4), E. coli products acted additively for phagocytosis and ROS
H. Zerbe et al. / Veterinary Microbiology 79 (2001) 351±365
361
generation, whereas mixtures of A. pyogenes products showed the same effect as fragments alone. Notably, mixtures of fragments and solubles from both bacteria inhibited all tested functional parameters of PMN (Fig. 4). 4. Discussion E. coli and A. pyogenes are the dominating bacteria in cases of bovine endometritis without retained placenta. This picture was seen in 21 out of 33 lochial secretions (taken in the ®rst 10 days after parturition), which con®rms previous ®ndings (Huszenicza et al., 1999). Uterine infections with E. coli and A. pyogenes also resulted in high concentrations of viable neutrophils in the uterine secretions (Table 1). This argues against a defect in the recruitment of neutrophils into the uterine lumen, however, these large numbers of neutrophils obviously were not able to eliminate the bacteria. This problem (already noted by Boitor et al., 1976) raised the question whether E. coli and A. pyogenes actively affect the defence capacity of immigrated neutrophils. The relative high number of viable intrauterine neutrophils suggested that direct or indirect interaction of bacteria and/or their released products do not affect the longevity of these cells. This was con®rmed by co-incubation experiments where fragments and solubles reduced the PMN viability only about 30% in the highest ratio between bacterial products and PMN after 24 h in vitro (Fig. 1). Although only products of one isolate (E. coli or A. pyogenes) was used, the general observation of high numbers of viable PMN in uterine secretions argues against the infection with strains which produce potent leukotoxins (Ding and LaÈmmler, 1996). LPS of E. coli was even reported to inhibit apoptosis of neutrophils (Watson et al., 1997). However, our results with E. coli products did not support this ®nding. This also makes it unlikely that E. coli (appearing nearly always earlier in the uteri than A. pyogenes) induces resistance in neutrophils towards putative leukotoxic effects of A. pyogenes. CD11b (complement receptor 3) and the mAb IL-A110-de®ned structure were signi®cantly upregulated (Fig. 2) by both bacteria. Most notably, the effect was more prominent if bacterial fragments were used and if PMN were co-incubated for 24 h. Whereas A. pyogenes has not been characterised so far, an enhanced expression of CD11b on human PMN was reported after infection with LPS-producing bacteria (Kermarrec et al., 1998) or after incubation of PMN with LPS in vitro (Lynn et al., 1991; Simms and D'Amico, 1995; Haugen et al., 1999). In cattle, Roets et al. (1999) reported on an enhanced expression of the beta chain (CD18) of the CD11/CD18 heterodimer after intramammary infusion of E. coli in cattle. It is not clear whether this was restricted to the CD11b/CD18 molecule, since our data provide no convincing evidence for an upregulation of CD11a (LFA-1) (Fig. 2A). Most likely, the observed upregulation of CD11b re¯ects an adaptation response of the PMN favouring their ability to bind and phagocytose opsonised bacteria. This may also explain why expression of CD11a/CD18 (LFA-1) is largely unchanged at this point of the in¯ammatory reaction, since an upregulation of this molecule is only needed for an ef®cient binding to the endothelium before extravasation. However, the selective upregulation of CD11b (and not CD11a) was
362
H. Zerbe et al. / Veterinary Microbiology 79 (2001) 351±365
also seen by Dosogne et al. (1997) on blood PMN 24 h after an arti®cial udder infection with E. coli. CD11b upregulation on PMN is not only induced by gram-negative bacteria. It could be induced on bovine PMN after an experimental infection of udders with Streptococcus uberis (Smits et al., 1998). On human PMN, intact Staphylococcus epidermidis upregulated both CD11b and CD11c (complement receptor 4) expression (Sapatnekar et al., 1997). Enhanced CD11b expression can also be induced by pure LPS (Klein et al., 1990; Lynn et al., 1991). However, polymyxin B, an inhibitor of LPS effects (Marra et al., 1990; Cornelissen et al., 1991) was only partially able to counteract the effects of E. coli solubles and fragments on expression of CD11b and the IL-A110-de®ned structure (Fig. 2B). Thus, beside LPS, E. coli contain and secrete factors which selectively modulate the phenotype of bovine granulocytes. Notably, E. coli shares these features with A. pyogenes since most of the tested surface molecules were modulated by products of both bacteria in the same sense (Fig. 2A). Interestingly, both bacteria induced also nearly the same effects on functional capacities of bovine PMN. At least the fragments downregulated the AICC, the phagocytosis and the ROS generation in a comparable manner (Fig. 3). Similar ®ndings, although hard to compare, were obtained by Dosogne et al. (1997), who observed diminished phagocytosis and ROS generation of bovine blood PMN after an E. coliinduced mastitis. Oostveldt van et al. (1999) reported on a biphasic modulation of the respiratory burst activity of blood PMN: generation of ROS increased up to 6 h after intramammary inoculation of E. coli and dropped signi®cantly after 18 h. We saw only minor changes in the phenotype and functional parameters after short-time incubations (3 h, data not shown) with E. coli or A. pyogenes products, however, the longtime (24 h) results seem to be compatible. Thus, after prolonged exposition to bacterial products in vitro or in vivo signs of cellular fatigue seem to dominate over initial activation. In short-time assays (40 min), bovine PMN were able to kill 80% of A. pyogenes bacteria in vitro (Watson, 1989). This is in part re¯ected by the enhancing effect of A. pyogenes solubles on phagocytosis (Fig. 3). However, this may be of minor importance for the in vivo situation, since this effect disappeared in the simultaneous presence of E. coli fragments (Fig. 4) which are also present in vivo before A. pyogenes infects the uteri. The signi®cant downregulation of ROS generation (especially by E. coli) may also explain the high numbers of vital neutrophils found in infected uteri, since ROS production has been shown to accelerate the death of neutrophils in vitro (Watson et al., 1996). 5. Conclusions Prolonged contact between bacterial fragments or solubles with bovine PMN results in a functional depression of the cells. An induced upregulation of distinct, function-related surface structures cannot counteract this general phenomenon. Thus, even high numbers of viable neutrophils found in the uteri of infected cows may be unable to eliminate the bacteria. The more pronounced functional depression of neutrophils by E. coli products
H. Zerbe et al. / Veterinary Microbiology 79 (2001) 351±365
363
strengthens the hypothesis (Ahlers and Grunert, 1993) that the ®rst appearance of E. coli in the uterus may support the co-infection of this organ by A. pyogenes at later times. Acknowledgements This investigation was supported by the Hans Wilhelm Schaumann Foundation. Isolates of E. coli were typed serologically by the Bundesintitut fuÈr Gesundheitlichen Verbraucherschutz und VeterinaÈrmedizin, Dessau, Germany. Dr. B. Franz supported the handling of the bacteria. References Ahlers, D., Grunert, E., 1993. Zur Problematik der Behandlung des in®zierten Uterus beim Rind im Puerperium. Prakt. Tierarzt Sonderh. Coll. Vet. XXIV, 57±62. Bekana, M., Johnson, P., Ekman, T., Kindahl, H., 1994. Intrauterine bacterial ®ndings in postpartum cows with retained fetal membranes. J. Vet. Med. A 4, 663±670. Bisping, W., Amtsberg, G., 1988. Colour Atlas for the Diagnosis of Bacterial Pathogens in Animals. Paul Parey, Berlin, pp. 45±47 and 160±169. Boitor, I., Draghici, C., Gaboreanu, M., Cristea, E., Boitor, M., 1976. Die Dynamik der Leukozyten und der Mikro¯ora des Uterusinhaltes im normalen und pathologischen Puerperium und bei chronischen endometritiden des Rindes. Dtsch. TieraÈrztl. Wochenschr. 83, 24±25. Cai, T.Q., Weston, P.G., Lund, L.A., Brodie, B., McKenna, D.J., Wagner, W.C., 1994. Association between neutrophil functions and periparturient disorders in cows. Am. J. Vet. Res. 55, 934±943. Chacin, M., Hansen, P.J., Drost, M., 1990. Effects of stage of the estrous cycle and steroid treatment on uterine immunoglobulin content and polymorphonuclear leukocytes in cattle. Theriogenology 34, 1169±1184. Cornelissen, J.J., Kessel van, C.P.M., Brouwer, E., Kraaijeveld, C.A., Verhoff, J., 1991. Exposure to lipopolysaccharide (LPS) primes polymorphonuclear leucocytes for enhanced release of oxygen metabolites after subsequent stimulation. J. Med. Microbiol. 34, 233±238. Dalrymple, S.A., Slattery, R., Aud, D.M., Krishna, M., Lucian, L.A., Murray, R., 1996. These data clarify the role of IL-6 as protective in virulent E. coli infection and suggest that the protective effect may be least partially mediated through neutrophils. Infect. Immunol. 64, 3231±3235. Ding, H., LaÈmmler, C., 1996. Puri®cation and further characterization of a haemolysin of Actinomyces pyogenes. Zentralbl. VeterinaÈrmed. B 43, 179±188. Dobrina, A., Nardon, E., Vecile, E., Cinco, M., Patriarca, P., 1995. Leptospira icterohemorrhagiae and leptospire peptidolglycans induce endothelial cell adhesiveness for polymorphonuclear leukocytes. Infect. Immunol. 63, 2995±2999. Dosogne, H., Burvenich, C., Werven van, T., Roets, E., Noordhuizen-Stassen, E.N., Goddeeris, B., 1997. Increased surface expression of CD11b receptors on polymorphonuclear leukocytes is not suf®cient to sustain phagocytosis during Escherichia coli mastitis in early postpartum dairy cows. Vet. Immunol. Immunopathol. 60, 47±59. EmmendoÈrffer, A., Hecht, M., Lohmann-Matthes, M.L., Roesler, J., 1990. A fast and easy method to determine the production of reactive oxygen intermediates by human and murine phagocytes using dihydrorhodamine 123. J. Immunol. Meth. 131, 269±275. Goya, T., Morisaki, T., Torisu, M., 1994. Immunologic assessment of host defence impairment in patients with septic multiple organ failure: relationship between complement activation and changes in neutrophil function. Surgery 115, 145±155. Haugen, T.S., Skjonsberg, O.H., Nakstad, B., Lyberg, T., 1999. Modulation of adhesion molecule pro®les on alveolar macrophages and blood leukocytes. Respiration 66, 528±537.
364
H. Zerbe et al. / Veterinary Microbiology 79 (2001) 351±365
Hendricks, A., Leibold, W., Kaever, V., Schuberth, H.-J., 2000. Prostaglandin E2 is variably induced by bacterial superantigens in bovine mononuclear cells and has a regulatory role for the T cell proliferative response. Immunobiology 201, 493±505. Howard, C.J., Morisson, W.J., 1991. Leukocyte antigen in cattle, sheep and goats. Proceedings of the ®rst international workshop on leukocyte antigens in cattle, sheep and goats. Vet. Immun. Immunopath. 27, 1±276. Hussain, A.M., 1989. Bovine uterine defence mechanisms: a review. Zentralbl. VeterinaÈrmed. B 36, 641±651. Hussain, A.M., Daniel, R.C.W., O'Boyle, D., 1990. Postpartum uterine ¯ora following normal and abnormal puerperium in cows. Theriogenology 34, 291±302. Huszenicza, G., Fodor, M., Gacs, M., Kulcsar, M., Dohmen, M.J.W., Vamos, M., Porkolab, L., Kegl, T., Bartyik, J., 1999. Uterine bacteriology, resumption of cyclic ovarian activity and fertility in postpartum cows kept in large-scale dairy herds. Reprod. Dom. Anim. 34, 237±245. Kaneene, J.B., Miller, R.A., Herdt, T.H., Gardiner, J.C., 1997. The association of serum nonesteri®ed fatty acids and cholesterol, management and feeding practices with peripartum disease in dairy cows. Prev. Vet. Med. 31, 59±72. Kermarrec, N., Chollet-Martin, S., Beloucif, S., Faivre, V., Gougerot-Pocidalo, M.A., Payen, D.M., 1998. Alveolar neutrophil oxidative burst and beta2 integrin expression in experimental acute pulmonary in¯ammation are not modi®ed by inhaled nitric oxide. Shock 10, 129±134. Klein, J.B., Payne, V., Schepers, T.M., McLeish, K.R., 1990. Bacterial lipopolysaccharide enhances polymorphonuclear leukocyte function independent of changes in intracellular calcium. In¯ammation 14, 599±611. Lotthammer, K., Wittkowski, G., 1994. Scheiden- und GebaÈrmuttererkrankungen. In: Lotthammer, K.H., Wittkowski, G. (Eds.), Fruchtbarkeit und Gesundheit der Rinder. Ulmer, Stuttgart, pp. 60±66. Lynn, W.A., Raetz, C.R., Qureshi, N., Golenbock, D.T., 1991. Lipopolysaccharide-induced stimulation of CD11b/CD18 expression on neutrophils. Evidence of speci®c receptor-based response and inhibition by lipid A-based antagonists. J. Immunol. 147, 3072±3079. Marra, M.N., Wilde, C.R., Grif®th, J.E., Snable, J.L., Scott, R.W., 1990. Bactericidal/permeability-increasing protein has endotoxin-neutralizing activity. J. Immunol. 144, 662±666. Martich, G.D., Danner, R.L., Ceska, M., Suffredini, S.A.F., 1991. Interleukin 8 (IL-8): a potent activator of neutrophils. J. Exp. Med. 173, 1021±1024. Morrow, D.A., Hillman, D., Dade, A.W., Kitchen, H., 1979. Clinical investigation of a dairy herd with the fat cow syndrome. J. Am. Vet. Med. Assoc. 174, 161±167. Naessens, J., Olubayo, R.O., Davis, W.C., Hopkins, J., 1993. Cross-reactivity of workshop antibodies with cells from domestic and wild ruminants. Vet. Immunol. Immunopathol. 39, 283±290. Noakes, D.E., Wallace, L., Smith, G.R., 1991. Bacterial ¯ora of the uterus of cows after calving on two hygienically contrasting farms. Vet. Rec. 128, 440±442. Olson, J.D., Ball, L., Mortimer, R.G., Farin, P.W., Adney, W.S., Huffman, E.M., 1984. Aspects of bacteriology and endocrinology of cows with pyometra and retained fetal membranes. Am. J. Vet. Res. 45, 2251±2255. Oostveldt van, K., Burvenich, C., Silva da, F.M., Massart-Leen, A.M., 1999. Respiratory burst activity in activated and unstimulated isolated bovine blood neutrophils during experimentally induced Escherichia coli mastitis. J. Dairy Res. 66, 375±383. Pechhold, K., Pohl, T., Kabelitz, D., 1994. Rapid quanti®cation of lymphocyte subsets in heterogeneous cell populations by ¯ow cytometry. Cytometry 16, 152±159. Roets, E., Burvenich, C., Diez-Fraile, A., Noordhuizen-Stassen, E.N., 1999. Evaluation of the role of endotoxin and cortisol on modulation of CD18 adhesion receptors in cows with mastitis caused by Escherichia coli. Am. J. Vet. Res. 60, 534±540. Roth, J.A., Kaeberle, M.L., Loren, L.H., 1983. Association of increased estradiol and progesterone blood values with altered bovine polymorphonuclear leukocyte function. Am. J. Vet. Res. 44, 247±253. Sapatnekar, S., Kao, W.J., Anderson, J.M., 1997. Leukocyte±biomaterial interactions in the presence of Staphylococcus epidermidis: ¯ow cytometric evaluation of leukocyte activation. J. Biomed. Mater. Res. 35, 409±420. Schuberth, H.J., Anders, I., Pape, U., Leibold, W., 1992. One-dimensional isoelectric focusing and immunoblotting of equine major histocompatibility complex class I antigens. Anim. Genet. 23, 87±95.
H. Zerbe et al. / Veterinary Microbiology 79 (2001) 351±365
365
Schuberth, H.J., Freigofas, R., Daugschies, A., Leibold, W., 2000. Assessment of antibody-independent cellular cytotoxicity (AICC) of porcine neutrophilic granulocytes by quantitative ¯ow cytometry. Lack of modulation by larvae products of Oesophagostomum dentatum. J. Vet. Med. B 47, 607±614. Selbitz, H.J., 1992. Enterobacteriaceae: Escherichia coli. In: Selbitz, H.J. (Ed.), Lehrbuch der VeterinaÈrmedizinischen Bakteriologie. Fischer, Stuttgart, pp. 76±82. Simms, H.H., D'Amico, R., 1995. Lipopolysaccharide induces intracytoplasmic migration of the polymorphonuclear leukocyte CD11b/CD18 receptor. Shock 3, 196±203. Smits, E., Burvenich, C., Guidry, A.J., Roets, E., 1998. In vitro expression of adhesion receptors and diapedesis by polymorphonuclear neutrophils during experimentally induced Streptococcus uberis mastitis. Infect. Immunol. 66, 2529±2534. Vandeplassche, M., Bouters, R., 1976. Puerperal metritis in the bovine. In: Proceedings of the Fourth International Congress on Animal Reproduction Art. Insem, Krakov, Poland, 1976, pp. 660±666. Vandeplassche, M., 1981. Neue vergleichende Aspekte der Involution und der puerperalen Metritis bei Stute, Kuh und Sau. Monatsh. VeterinaÈrmed. 36, 804±807. Watson, E.D., 1989. In vitro function of bovine neutrophils against Actinomyces pyogenes. Am. J. Vet. Res. 50, 455±458. Watson, R.W., Redmond, H.P., Wang, J.H., Condron, C., Bouchier-Hayes, D., 1996. Neutrophils undergo apoptosis following ingestion of Escherichia coli. J. Immunol. 156, 3986±3992. Watson, R.W., Rotstein, O.D., Nathens, A.B., Parodo, J., Marshall, J.C., 1997. Neutrophil apoptosis is modulated by endothelial transmigration and adhesion molecule engagement. J. Immunol. 158, 945±953. Zerbe, H., Schuberth, H.J., Hoedemaker, M., Grunert, E., Leibold, W., 1996. A new model system for endometritis: basic concepts and characterization of phenotypic and functional properties of bovine uterine neutrophils. Theriogenology 46, 1339±1356. Zerbe, H., Schneider, N., Ossadnik, C., Wensink, T., Kruip, T.A.M., Grunert, E., Leibold, W., 1998. Eigenschaften neutrophiler Granulozyten aus Blut und Uterus Peripartaler KuÈhe. Wien. TieraÈrztl. Monatsschr. 85, 304±309. Zerbe, H., Schneider, N., Leibold, W., Wensing, T., Kruip, T.A.M., Schuberth, H.J., 2000. Altered functional and immunophenotypical properties of neutrophilic granulocytes in postpartum cows associated with fatty liver. Theriogenology 54, 771±786.
In¯uence of Escherichia coli and Arcanobacterium pyogenes isolated from bovine puerperal uteri on phenotypic and functional properties of neutrophils H. Zerbea,*, C. Oûadnika, W. Leiboldb, H.J. Schuberthb a
Clinic for Bovine Obstetrics and Gynaecology, School of Veterinary Medicine Hannover, Bischofsholer Damm 15, D-30173 Hannover, Germany b Immunology Unit, School of Veterinary Medicine Hannover, Bischofsholer Damm 15, D-30173 Hannover, Germany
Received 22 June 2000; received in revised form 17 October 2000; accepted 17 October 2000
Abstract When cows develop endometritis after birth, Escherichia coli and Arcanobacterium pyogenes are usually the most prominent bacteria present in bovine uterine lochial secretions. A. pyogenes alone is rarely found in the course of a disturbed puerperium. This was con®rmed in this study, since average and high-grade uterine contaminations were always associated with the presence of both bacteria. The contamination grade was positively correlated with uterine polymorphonuclear granulocyte (PMN) numbers and negatively correlated with blood PMN numbers. Whether E. coli and A. pyogenes affect the phenotype and function of bovine PMN in a similar or differential way was subject to in vitro studies. PMN were tested in the presence of washed bacterial fragments or culture supernatants taken as a source for soluble and/or secreted bacterial products. Fragments and soluble products differed only quantitatively in their effects on PMN. Usually, long-time exposure (24 h) of PMN to fragments induced the strongest effects. Accelerated death of granulocytes was only moderately induced by both E. coli and A. pyogenes products. Both E. coli and A. pyogenes products induced the enhanced expression of a membrane molecule detected by mAb IL-A110 and of CD11b. Expression of other surface structures remained largely unchanged (MHC class I, CD11c). Functional parameters of PMN (phagocytosis; generation of reactive oxygen species, ROS; antibody-independent cellular cytotoxicity, AICC) generally declined after pre-incubation for 24 h with products of E. coli or A. pyogenes. Interestingly, soluble products of A. pyogenes stimulated the phagocytosis of PMN. However, co-incubation with E. coli products abrogated this
*
Corresponding author. Tel.: 49-511-856-7611; fax: 49-511-856-7691. E-mail address: [email protected] (H. Zerbe). 0378-1135/01/$ ± see front matter # 2001 Elsevier Science B.V. All rights reserved. PII: S 0 3 7 8 - 1 1 3 5 ( 0 0 ) 0 0 3 6 8 - 0
352
H. Zerbe et al. / Veterinary Microbiology 79 (2001) 351±365
stimulatory effect. The results supply evidence for similar modes of action of the gram-negative E. coli and the gram-positive A. pyogenes on bovine PMN. Alterations in PMN function and phenotype are mainly triggered by direct contact between bacterial fragments and PMN. Inhibition experiments with polymyxin B demonstrated that E. coli-mediated effects were not solely due to the action of lipopolysaccharide. The dominant functional depression of neutrophils by E. coli products strengthens the suggestion that the earlier appearance of E. coli in the uterus may support the coinfection of this organ by A. pyogenes at later times. # 2001 Elsevier Science B.V. All rights reserved. Keywords: Cattle; Bacteria; Uterus; Escherichia coli; Arcanobacterium pyogenes; Neutrophils
1. Introduction Most health problems in high producing cows occur in the peripartum period. Together with vaginitis, endometritis causes economic losses accounting for 50% of all reproduction disorders in Germany (Lotthammer and Wittkowski, 1994). Lochial secretions taken as early as 1 day after birth contain mostly Escherichia coli and streptococci (Ahlers and Grunert, 1993). Streptococci appear to be members of the normal uterine bacterial ¯ora (Huszenicza et al., 1999). In some cases also staphylococci and Proteus spp. can be isolated (Vandeplassche, 1981; Olson et al., 1984; Noakes et al., 1991; Huszenicza et al., 1999). Later in the early puerperium Arcanobacterium pyogenes especially appears in lochial secretions (Olson et al., 1984; Noakes et al., 1991; Ahlers and Grunert, 1993). In cows with normal puerperium intrauterine bacteria are usually eliminated until the 10th day postpartum (Boitor et al., 1976). In summary, gram-negative bacteria, especially E. coli, seem to dominate in the uterus within the ®rst days after calving (Hussain et al., 1990; Huszenicza et al., 1999). Later after birth, A. pyogenes and gram-negative anaerobic bacteria are found in uterine secretions of animals suffering from severe puerperal endometritis, especially in combination with retention of the fetal membranes (Bekana et al., 1994). The appearance of A. pyogenes seems to be an important indicator of a puerperal disorder resulting in a disturbed fertility (Olson et al., 1984; Huszenicza et al., 1999). Since the local activity of uterine PMN is of major importance for the antimicrobial defence in this organ after birth (Vandeplassche and Bouters, 1976; Hussain, 1989), it is intriguing to examine possible inhibitory effects of products of the dominating bacterial species. However, little is known about local effects of bacterial contamination on granulocytes. Lipopolysaccharide (LPS) from E. coli was able to upregulate adhesion molecules (ICAM-1, ELAM-1) on human endothelial cells in vitro, which facilitated adhesion of granulocytes (Dobrina et al., 1995). Upregulation of CD11b (complement receptor 3) was also seen on lung-immigrating PMN after intratracheal administration of LPS in rats (Kermarrec et al., 1998). In cattle, the increased surface expression of CD11b on PMN seemed to be insuf®cient for a sustained phagocytosis during E. coli-induced mastitis in early postpartum dairy cows. Dosogne et al. (1997) observed a biphasic upregulation of CD11b expression on PMN but phagocytosis was decreased 12 h
H. Zerbe et al. / Veterinary Microbiology 79 (2001) 351±365
353
after infection. It is speculated that LPS-induced activation of PMN and macrophages results in a hyper-reactive state of the cells causing tissue damage in addition to the development of septicemic shock (Martich et al., 1991; Goya et al., 1994; Dalrymple et al., 1996). In¯uences of A. pyogenes on granulocytes have been rarely described. One strain of A. pyogenes (formerly Actinomyces pyogenes, ATCC 8164) secreted a hemolysin which was cytotoxic for PMN and for pTK2 kidney cells (Ding and LaÈmmler, 1996). It is unclear, however, whether A. pyogenes isolated from uterine secretions do secrete such leukotoxic factors. The reasons for an increased susceptibility for infections of cows after birth are not fully understood. Functional capacities (e.g. phagocytosis, generation of ROS) of PMN compared between pre- and postpartum periods were lower in the period after calving. This correlated with a diminished expression of functionally important PMN surface molecules, e.g. CD11b in the blood (Cai et al., 1994; Zerbe et al., 1998). Whether these changes are the cause of liver function disorders (Morrow et al., 1979; Kaneene et al., 1997; Zerbe et al., 2000), hormonal imbalances (Roth et al., 1983; Chacin et al., 1990) or due to the contact of PMN with products of intrauterine bacteria remains an open question. Here, we report on investigations dealing with the latter hypothesis. It was of special interest to determine whether secreted products and/or membrane-bound structures of different bacteria isolated from bovine uterine secretions exert common or even contrary effects on function and phenotype of bovine polymorphs. 2. Materials and methods 2.1. Animals and cells For collection of uterine ¯uid Holstein-Friesian postpartum cows 3 and 6 years of age (n 33, 550±650 kg) without retained placenta, but with disturbed early puerperium (delayed uterus involution and acute endometritis) were supplied by the Clinic for Bovine Obstetrics and Gynaecology of the Hannover School of Veterinary Medicine, Germany. Animals were kept in stalls under similar conditions. Diagnosis of delayed uterus involution and acute endometritis was based on the evaluation of lochia (presence of foul-smelling, malodorous cervical discharge) and rectal palpation (enlarged, atonic uterus). Neutrophilic granulocytes were separated from venous blood drawn from the V. jugularis externa into vacutainer tubes containing EDTA (Becton Dickinson). Blood donors were three clinically healthy ovariectomised German Blackpied cows (4±5 years). Blood samples were kept at room temperature and separation of PMN was performed within 2 h. Blood was centrifuged (1500g, 5 min, 48C) and contaminating erythrocytes from buffy coat leukocytes were removed by two successive hypotonic lysing steps. The cell suspension was then separated over a discontinuous density gradient (Percoll1, Pharmacia) to obtain pure granulocyte populations (Zerbe et al., 1996). The purity was consistently 95% as determined ¯ow cytometrically.
354
H. Zerbe et al. / Veterinary Microbiology 79 (2001) 351±365
2.2. Collection and identi®cation of bacteria Uterine secretions were taken manually from the cervix uteri or from the corpus uteri under aseptical conditions at days 4 or 9 postpartum. The outer genital was washed with water and soap, and sterile gloves were used. E. coli and A. pyogenes were isolated, identi®ed and subcultivated for experiments by means of standard bacteriological techniques (Bisping and Amtsberg, 1988). Isolates of E. coli obtained from 11 animals were typed serologically. Since virulence factors of E. coli isolated from uteri cannot be determined at present, typing characteristics for enteropathogenic E. coli were utilised. Three (of 14) isolates could not be recultivated after storage. Eight samples were untypable. Serotype O101:K28 which is considered to be pathogenic for cattle was present in two out of three samples (Selbitz, 1992). In one animal this was the only serotype isolated, the other animal harboured two serotypes in addition (O101:K30 and O101:K32). From the third animal, the serotype O141:H25 was isolated. Due to the relative dominance of O101:K28, this serotype was chosen for the subsequent preparation of E. coli products. All strains of A. pyogenes isolates were û-hemolytic. Biochemical typing was not performed. Puri®ed and lyophilised bacteria were reconstituted in 2 ml Ampuwa medium (Fresenius) and cultured for 24 h at 378C in 100 ml culture medium: RPMI 1640 supplemented with 2 mol/l L-glutamine, 15 mmol/l HEPES and 18 mmol/l NaHCO3 (medium and all supplements, Biochrom) without addition of serum. All buffers and media were free of contaminating LPS as determined with the limulus assay (Endotoxintest ``mini'', Company Schulz, Bad Krozingen, Germany). Numbers of colony forming units (CFU) were determined by standard techniques. 2.3. Preparations of bacterial products and co-incubation with PMN in vitro After centrifugation of cultured bacteria (15 000g, 20 min, 48C), the supernatant was removed, sterile-®ltered (0.5 mm) and stored in aliquots at ÿ1008C. The sedimented bacteria were washed twice in sterile PBS (15 000g, 20 min, 48C) and were ®nally disrupted into fragments (French Pressure Cells and Press1, American Instruments). Complete killing of the bacteria was judged by cultivation on blood agar (24 h, 378C, 5% CO2 in air). The concentrations of supernatants and bacterial fragments are expressed as equivalents of CFU present in the suspensions before preparation. E. coli preparations contained 3 mg/ml LPS (equivalent to 2 108 CFU=ml). LPS contents in A. pyogenes preparations were below the detection limit (<0.001 mg/ml). Co-incubations with PMN 2 105 per well were made in 96 well round-bottom microtiter plates for 3 or 24 h at 378C in a humidi®ed atmosphere with 5% CO2 in air. Bacterial fragments or solubles were added at ®nal ratios between 1:1 and 1:1000 (PMN:CFU equivalents). These ratios corresponded to the situation found ex vivo (up to 1:2000; data not shown). Parallel PMN preparations were co-incubated with puri®ed LPS (serotype 0111:B4, Sigma; 3 mg/ml) instead of bacterial fragments or solubles. All set-ups were made in triplicate. Sterile culture medium served as controls.
H. Zerbe et al. / Veterinary Microbiology 79 (2001) 351±365
355
2.4. Cell counts, differentiation, determination of cellular viability Total blood leukocyte concentrations were determined in a cell counting chamber after staining with TuÈrk's solution (Merck). Differential counts (granulocytes, lymphocytic cells and monocytes) were determined by ¯ow cytometry (FACScan1, Becton Dickinson) after acquisition of at least 5000 leukocytes. Cellular suspensions contained propidium iodide (2 mg/ml, Calbiochem) for the assessment of cellular viability. Total numbers of viable PMN after co-culture with bacterial products were determined with a quantitative ¯ow cytometric method (SCDA, single cell dilution assay, Pechhold et al., 1994) which was adapted for the bovine system (Hendricks et al., 2000). 2.5. Phenotyping The procedure of staining cells with primary and secondary antibodies has been described in detail elsewhere (Zerbe et al., 1996). The phenotypic analysis was performed ¯ow cytometrically. The relative (mean) ¯uorescence intensity served to describe the relative expression density of the studied cell surface molecules on isolated PMN. Primary, bovine-speci®c monoclonal antibodies (mAbs) used were: mAb Bo1 (anti-MHC class I, Schuberth et al., 1992), mAb IL-A15 (anti-boCD11b, Naessens et al., 1993), mAb IL-A16 (anti-boCD11c, Howard and Morisson, 1991), mAb IL-A99 (anti-boCD11a, Naessens et al., 1993), and mAb IL-A110 (granulocyte-speci®c, Naessens et al., 1993). 2.6. Functional assays for neutrophilic granulocytes Phagocytosis assay. Nonviable streptococci in suspension (Omnisorb1, Calbiochem) were labelled with ¯uoresceinisothiocyanat (FITC, Sigma): FITC buffer (0.7 ml) (3 mg FITC in 7 ml 0.1 mol/l Na2CO3 and 0.1 mol/l NaCl, pH 9.2) and 5 ml NaN3 (10% (w/v) in H2O) were mixed with 0.3 ml Omnisorb1 suspension for 20 h at RT in the dark under constant rotation. Bacteria were washed three times with PBS (15 000g, 1 min) and ®nally adjusted to 2 108 cells=ml. Aliquots were opsonised by addition of bovine serum and stored at ÿ208C. Bacteria were mixed with PMN at ratios between 1:1 and 1000:1 and incubated for 60 min at 378C. By means of ¯ow cytometry, PMN with ingested bacteria were identi®ed due to their increased green ¯uorescence (Zerbe et al., 1996). Generation of ROS. The general procedure followed that described by EmmendoÈrffer et al. (1990). PMN were stimulated with phorbol myristate acetate (PMA, 300 nmol/l; Sigma). Thereafter, the cells were loaded with the non¯uorescent dye dihydrorhodamin 123 (DHR 123, 15 mg/ml; Molecular Probes) which, in case of ROS generation after stimulation, was converted into the green ¯uorescent rhodamin 123 via oxidation catalysed by cellular myeloperoxidase. The amount of ROS generation was measured ¯ow cytometrically by determination of the relative (mean) ¯uorescence intensity. Antibody-independent cellular cytotoxicity (AICC). Bovine lymphoblastoid cells (``Anna TA1'', Schuberth et al., 2000) served as target cells. Targets were washed twice in PBS (80g, 8 min) and adjusted to 1 106 cells=ml culture medium (RPMI 1640, with
356
H. Zerbe et al. / Veterinary Microbiology 79 (2001) 351±365
10% (v/v) fetal calf serum). After labelling with the dye D275 (Mobitec; 1.25 mg/ml cell suspension, 30 min, RT), the targets were washed twice in PBS (80g, 8 min). Finally, targets were adjusted to 1 106 cells=ml culture medium and mixed with effector PMN at a ratio of 30:1 (PMN:targets). Co-incubation was carried out for 4 h at 378C in the presence of 30 nmol/l PMA. Cell viability of target cells was evaluated by ¯ow cytometry after addition of propidium iodide (see above). 2.7. Statistical analysis All assays were set up in triplicate. All statistical procedures (means, SD and student's t-test) were performed using SAS software. 3. Results 3.1. Bacteria and leukocyte numbers Based on the grade of contamination of uterine secretions, 33 samples were grouped in four categories (Table 1). It turned out that middle grade (<100 bacteria per examination) and high-grade contamination exclusively was associated with the presence of E. coli and A. pyogenes. In nine of these 21 samples both E. coli and A. pyogenes were detectable; ®ve contained only E. coli and seven only A. pyogenes. With the exception of one case all A. pyogenes-positive isolates (mixed with E. coli or pure) were found at day 9 postpartum, all E. coli-pure samples at day 4 postpartum. Mean PMN concentrations in middle- and high-contaminated uterine secretions (table groups 3 and 4) were signi®cantly higher compared to not- or low-contaminated samples. Conversely, PMN numbers in the peripheral blood were lower in the middle- and high-contaminated animals compared to bacteriologically negative or low-contaminated cows.
Table 1 Bacterial contamination of uterine secretions and its in¯uence on PMN concentrations in uterus and blood of puerperal disordered cows without retained placenta Uterine bacteriology (group)
Bacteriaa
Cattle (No.)
PMN (106 per ml)b Uterus
Negative (1) Staphylococcus spp., Streptococcus spp., Proteus spp., Klebsiella spp. (2) E. coli and A. pyogenes (3) E. coli and A. pyogenes (4)
0 1±10 11±100 >100
4 8 8 13
1.40.7 a 5.86.9 c 7.85.3 b,e 30.930.2 d,f
Blood 4:3 1:2 g,i 2:9 1:5 2:7 1:0 h 1:9 1:1 k
a Bacteria were counted microscopically (oil immersion, 1250) under standardised conditions (numbers in 10 different areas). b PMN concentrations were determined in uterine secretions and in venous blood of cows in the early puerperium. Significances: a vs. b, c vs. d and e vs. f were different at P < 0:05; g vs. h were different at P < 0:05 and i vs. k at P < 0:001.
H. Zerbe et al. / Veterinary Microbiology 79 (2001) 351±365
357
Fig. 1. Viability of PMN after incubation with different bacterial preparations. Separated PMN of three ovariectomised cattle 2 105 per well were incubated for 24 h with fragments or soluble products of E. coli or A. pyogenes. Percentage of viable PMN were determined ¯ow cytometrically with the SCDA assay.
3.2. E. coli and A. pyogenes negatively affect viability of PMN Products of both bacteria dose-dependently accelerated the death of neutrophils in vitro. Compared to medium controls about 70% of the neutrophils remained vital after 24 h (Fig. 1). Whereas bacterial fragments did not differ in their potency between E. coli and A. pyogenes, soluble constituents of E. coli were slightly more potent compared to those obtained from A. pyogenes (Fig. 1). 3.3. Phenotypic changes of PMN induced by E. coli and A. pyogenes Differences in expression densities were regarded as biologically signi®cant, if the MFI of a given membrane molecules differed at least by 50% compared to the appropriate controls. CD11b and a structure de®ned by mAb IL-A110 were signi®cantly upregulated. CD11b expression was up to 3.5 times higher expressed after incubation of PMN with A. pyogenes fragments (Fig. 2A); the mAb IL-A110-de®ned molecule was expressed 4.7 times higher after incubation with E. coli fragments (Fig. 2A). Bacterial fragments seemed to be more potent than soluble components. An interesting exception from this rule was seen with CD11c (Fig. 2A): only E. coli and A. pyogenes solubles induced divers reactions, whereas fragments had no effect. Neither product modulated the expression of MHC class I molecules (data not shown). Enhanced expression of CD11b and the mAb IL-A110-de®ned molecule was also induced by pure LPS (Fig. 2B). This effect was sensitive to polymyxin B treatment of the cells which also abrogated the expression-enhancing effect of E. coli solubles. However, polymyxin B was less potent in the reduction of CD11b expression induced by E. coli fragments (Fig. 2B) and the enhanced expression of the IL-A110-de®ned molecule
358
H. Zerbe et al. / Veterinary Microbiology 79 (2001) 351±365
Fig. 2. Modulation of PMN surface expression by bacterial products. Separated PMN of three ovariectomised cattle 2 105 per well were incubated for 24 h with fragments or soluble products of E. coli or A. pyogenes. Expression densities (mean ¯uorescence intensities, MFI) of different surface structures were determined ¯ow cytometrically after indirect immuno¯uorescence. (A) Relative changes of the MFI (in percent to the controls, mean and standard deviation of three animals). Positive or negative changes >50% compared to controls were considered signi®cant (dashed lines). Standard deviations were omitted from the graph (coef®cient of variance 6%). (B) MFIs of CD11b and IL-A110 (granulocyte-speci®c molecule) after incubation of PMN with LPS (3 mg/ml), E. coli-derived fragments or E. coli-derived soluble products. Fragments and solubles corresponded to 2 108 CFU E. coli/ml containing 3 mg/ml LPS. Parallel cultures contained polymyxin B (3 mg/ml) to inhibit LPS induced changes of expression.
H. Zerbe et al. / Veterinary Microbiology 79 (2001) 351±365
359
Fig. 3. Products of E. coli and A. pyogenes depress functional capacities of bovine PMN. Blood PMN 2 105 per well of three ovariectomised cattle were incubated for 24 h in the presence of fragments or soluble products of E. coli or A. pyogenes in different concentrations. AICC, the capacity to phagocytose or to generate ROS was determined ¯ow cytometrically. Positive or negative changes >50% compared to controls (cell incubated without bacterial products) were considered signi®cant (dashed lines). Standard deviations were omitted from the graph (coef®cient of variance 6%).
remained nearly unaffected. A. pyogenes-induced changes of surface expression were insensitive to polymyxin B (data not shown). 3.4. Functional changes of PMN induced by E. coli and A. pyogenes Generation of ROS was downregulated most prominently by E. coli products (Fig. 3) in a dose-dependent manner. Again, E. coli fragments were more potent than solubles. Interestingly, A. pyogenes products were less potent and nearly failed to modulate this function. Only higher concentrations of E. coli and A. pyogenes fragments decreased the ability of PMN to ingest bacteria signi®cantly (Fig. 3). The effects of soluble bacterial preparations were contradictory. Whereas E. coli products slightly inhibited the phagocytosis, it was slightly enhanced by soluble constituents of A. pyogenes. The
360
H. Zerbe et al. / Veterinary Microbiology 79 (2001) 351±365
Fig. 4. Effects of bacterial fragments, and solubles and mixtures of both on functional parameters of bovine PMN. PMN of three ovariectomised animals were pre-cultivated for 24 h in the presence of fragments (F), soluble constituents (S), combination of both F S or combinations of fragments and solubles of E. coli and A. pyogenes (last bar). Fragments and solubles were adjusted to equivalents of 2 108 CFU. Cells were also preincubated with pure LPS (serotype O111:B4, 3 mg/ml). Cells incubated in medium alone served as controls (dashed line). PMN functions (AICC, generation of ROS, phagocytosis) were determined ¯ow cytometrically (means standard deviations).
effects of E. coli or A. pyogenes products on the AICC of PMN did not reach signi®cance (Fig. 3). All tested functional properties of PMN were nearly unaffected in the presence of pure LPS (3 mg/ml) (Fig. 4). In experiments where fragments and solubles of bacteria were mixed (Fig. 4), E. coli products acted additively for phagocytosis and ROS
H. Zerbe et al. / Veterinary Microbiology 79 (2001) 351±365
361
generation, whereas mixtures of A. pyogenes products showed the same effect as fragments alone. Notably, mixtures of fragments and solubles from both bacteria inhibited all tested functional parameters of PMN (Fig. 4). 4. Discussion E. coli and A. pyogenes are the dominating bacteria in cases of bovine endometritis without retained placenta. This picture was seen in 21 out of 33 lochial secretions (taken in the ®rst 10 days after parturition), which con®rms previous ®ndings (Huszenicza et al., 1999). Uterine infections with E. coli and A. pyogenes also resulted in high concentrations of viable neutrophils in the uterine secretions (Table 1). This argues against a defect in the recruitment of neutrophils into the uterine lumen, however, these large numbers of neutrophils obviously were not able to eliminate the bacteria. This problem (already noted by Boitor et al., 1976) raised the question whether E. coli and A. pyogenes actively affect the defence capacity of immigrated neutrophils. The relative high number of viable intrauterine neutrophils suggested that direct or indirect interaction of bacteria and/or their released products do not affect the longevity of these cells. This was con®rmed by co-incubation experiments where fragments and solubles reduced the PMN viability only about 30% in the highest ratio between bacterial products and PMN after 24 h in vitro (Fig. 1). Although only products of one isolate (E. coli or A. pyogenes) was used, the general observation of high numbers of viable PMN in uterine secretions argues against the infection with strains which produce potent leukotoxins (Ding and LaÈmmler, 1996). LPS of E. coli was even reported to inhibit apoptosis of neutrophils (Watson et al., 1997). However, our results with E. coli products did not support this ®nding. This also makes it unlikely that E. coli (appearing nearly always earlier in the uteri than A. pyogenes) induces resistance in neutrophils towards putative leukotoxic effects of A. pyogenes. CD11b (complement receptor 3) and the mAb IL-A110-de®ned structure were signi®cantly upregulated (Fig. 2) by both bacteria. Most notably, the effect was more prominent if bacterial fragments were used and if PMN were co-incubated for 24 h. Whereas A. pyogenes has not been characterised so far, an enhanced expression of CD11b on human PMN was reported after infection with LPS-producing bacteria (Kermarrec et al., 1998) or after incubation of PMN with LPS in vitro (Lynn et al., 1991; Simms and D'Amico, 1995; Haugen et al., 1999). In cattle, Roets et al. (1999) reported on an enhanced expression of the beta chain (CD18) of the CD11/CD18 heterodimer after intramammary infusion of E. coli in cattle. It is not clear whether this was restricted to the CD11b/CD18 molecule, since our data provide no convincing evidence for an upregulation of CD11a (LFA-1) (Fig. 2A). Most likely, the observed upregulation of CD11b re¯ects an adaptation response of the PMN favouring their ability to bind and phagocytose opsonised bacteria. This may also explain why expression of CD11a/CD18 (LFA-1) is largely unchanged at this point of the in¯ammatory reaction, since an upregulation of this molecule is only needed for an ef®cient binding to the endothelium before extravasation. However, the selective upregulation of CD11b (and not CD11a) was
362
H. Zerbe et al. / Veterinary Microbiology 79 (2001) 351±365
also seen by Dosogne et al. (1997) on blood PMN 24 h after an arti®cial udder infection with E. coli. CD11b upregulation on PMN is not only induced by gram-negative bacteria. It could be induced on bovine PMN after an experimental infection of udders with Streptococcus uberis (Smits et al., 1998). On human PMN, intact Staphylococcus epidermidis upregulated both CD11b and CD11c (complement receptor 4) expression (Sapatnekar et al., 1997). Enhanced CD11b expression can also be induced by pure LPS (Klein et al., 1990; Lynn et al., 1991). However, polymyxin B, an inhibitor of LPS effects (Marra et al., 1990; Cornelissen et al., 1991) was only partially able to counteract the effects of E. coli solubles and fragments on expression of CD11b and the IL-A110-de®ned structure (Fig. 2B). Thus, beside LPS, E. coli contain and secrete factors which selectively modulate the phenotype of bovine granulocytes. Notably, E. coli shares these features with A. pyogenes since most of the tested surface molecules were modulated by products of both bacteria in the same sense (Fig. 2A). Interestingly, both bacteria induced also nearly the same effects on functional capacities of bovine PMN. At least the fragments downregulated the AICC, the phagocytosis and the ROS generation in a comparable manner (Fig. 3). Similar ®ndings, although hard to compare, were obtained by Dosogne et al. (1997), who observed diminished phagocytosis and ROS generation of bovine blood PMN after an E. coliinduced mastitis. Oostveldt van et al. (1999) reported on a biphasic modulation of the respiratory burst activity of blood PMN: generation of ROS increased up to 6 h after intramammary inoculation of E. coli and dropped signi®cantly after 18 h. We saw only minor changes in the phenotype and functional parameters after short-time incubations (3 h, data not shown) with E. coli or A. pyogenes products, however, the longtime (24 h) results seem to be compatible. Thus, after prolonged exposition to bacterial products in vitro or in vivo signs of cellular fatigue seem to dominate over initial activation. In short-time assays (40 min), bovine PMN were able to kill 80% of A. pyogenes bacteria in vitro (Watson, 1989). This is in part re¯ected by the enhancing effect of A. pyogenes solubles on phagocytosis (Fig. 3). However, this may be of minor importance for the in vivo situation, since this effect disappeared in the simultaneous presence of E. coli fragments (Fig. 4) which are also present in vivo before A. pyogenes infects the uteri. The signi®cant downregulation of ROS generation (especially by E. coli) may also explain the high numbers of vital neutrophils found in infected uteri, since ROS production has been shown to accelerate the death of neutrophils in vitro (Watson et al., 1996). 5. Conclusions Prolonged contact between bacterial fragments or solubles with bovine PMN results in a functional depression of the cells. An induced upregulation of distinct, function-related surface structures cannot counteract this general phenomenon. Thus, even high numbers of viable neutrophils found in the uteri of infected cows may be unable to eliminate the bacteria. The more pronounced functional depression of neutrophils by E. coli products
H. Zerbe et al. / Veterinary Microbiology 79 (2001) 351±365
363
strengthens the hypothesis (Ahlers and Grunert, 1993) that the ®rst appearance of E. coli in the uterus may support the co-infection of this organ by A. pyogenes at later times. Acknowledgements This investigation was supported by the Hans Wilhelm Schaumann Foundation. Isolates of E. coli were typed serologically by the Bundesintitut fuÈr Gesundheitlichen Verbraucherschutz und VeterinaÈrmedizin, Dessau, Germany. Dr. B. Franz supported the handling of the bacteria. References Ahlers, D., Grunert, E., 1993. Zur Problematik der Behandlung des in®zierten Uterus beim Rind im Puerperium. Prakt. Tierarzt Sonderh. Coll. Vet. XXIV, 57±62. Bekana, M., Johnson, P., Ekman, T., Kindahl, H., 1994. Intrauterine bacterial ®ndings in postpartum cows with retained fetal membranes. J. Vet. Med. A 4, 663±670. Bisping, W., Amtsberg, G., 1988. Colour Atlas for the Diagnosis of Bacterial Pathogens in Animals. Paul Parey, Berlin, pp. 45±47 and 160±169. Boitor, I., Draghici, C., Gaboreanu, M., Cristea, E., Boitor, M., 1976. Die Dynamik der Leukozyten und der Mikro¯ora des Uterusinhaltes im normalen und pathologischen Puerperium und bei chronischen endometritiden des Rindes. Dtsch. TieraÈrztl. Wochenschr. 83, 24±25. Cai, T.Q., Weston, P.G., Lund, L.A., Brodie, B., McKenna, D.J., Wagner, W.C., 1994. Association between neutrophil functions and periparturient disorders in cows. Am. J. Vet. Res. 55, 934±943. Chacin, M., Hansen, P.J., Drost, M., 1990. Effects of stage of the estrous cycle and steroid treatment on uterine immunoglobulin content and polymorphonuclear leukocytes in cattle. Theriogenology 34, 1169±1184. Cornelissen, J.J., Kessel van, C.P.M., Brouwer, E., Kraaijeveld, C.A., Verhoff, J., 1991. Exposure to lipopolysaccharide (LPS) primes polymorphonuclear leucocytes for enhanced release of oxygen metabolites after subsequent stimulation. J. Med. Microbiol. 34, 233±238. Dalrymple, S.A., Slattery, R., Aud, D.M., Krishna, M., Lucian, L.A., Murray, R., 1996. These data clarify the role of IL-6 as protective in virulent E. coli infection and suggest that the protective effect may be least partially mediated through neutrophils. Infect. Immunol. 64, 3231±3235. Ding, H., LaÈmmler, C., 1996. Puri®cation and further characterization of a haemolysin of Actinomyces pyogenes. Zentralbl. VeterinaÈrmed. B 43, 179±188. Dobrina, A., Nardon, E., Vecile, E., Cinco, M., Patriarca, P., 1995. Leptospira icterohemorrhagiae and leptospire peptidolglycans induce endothelial cell adhesiveness for polymorphonuclear leukocytes. Infect. Immunol. 63, 2995±2999. Dosogne, H., Burvenich, C., Werven van, T., Roets, E., Noordhuizen-Stassen, E.N., Goddeeris, B., 1997. Increased surface expression of CD11b receptors on polymorphonuclear leukocytes is not suf®cient to sustain phagocytosis during Escherichia coli mastitis in early postpartum dairy cows. Vet. Immunol. Immunopathol. 60, 47±59. EmmendoÈrffer, A., Hecht, M., Lohmann-Matthes, M.L., Roesler, J., 1990. A fast and easy method to determine the production of reactive oxygen intermediates by human and murine phagocytes using dihydrorhodamine 123. J. Immunol. Meth. 131, 269±275. Goya, T., Morisaki, T., Torisu, M., 1994. Immunologic assessment of host defence impairment in patients with septic multiple organ failure: relationship between complement activation and changes in neutrophil function. Surgery 115, 145±155. Haugen, T.S., Skjonsberg, O.H., Nakstad, B., Lyberg, T., 1999. Modulation of adhesion molecule pro®les on alveolar macrophages and blood leukocytes. Respiration 66, 528±537.
364
H. Zerbe et al. / Veterinary Microbiology 79 (2001) 351±365
Hendricks, A., Leibold, W., Kaever, V., Schuberth, H.-J., 2000. Prostaglandin E2 is variably induced by bacterial superantigens in bovine mononuclear cells and has a regulatory role for the T cell proliferative response. Immunobiology 201, 493±505. Howard, C.J., Morisson, W.J., 1991. Leukocyte antigen in cattle, sheep and goats. Proceedings of the ®rst international workshop on leukocyte antigens in cattle, sheep and goats. Vet. Immun. Immunopath. 27, 1±276. Hussain, A.M., 1989. Bovine uterine defence mechanisms: a review. Zentralbl. VeterinaÈrmed. B 36, 641±651. Hussain, A.M., Daniel, R.C.W., O'Boyle, D., 1990. Postpartum uterine ¯ora following normal and abnormal puerperium in cows. Theriogenology 34, 291±302. Huszenicza, G., Fodor, M., Gacs, M., Kulcsar, M., Dohmen, M.J.W., Vamos, M., Porkolab, L., Kegl, T., Bartyik, J., 1999. Uterine bacteriology, resumption of cyclic ovarian activity and fertility in postpartum cows kept in large-scale dairy herds. Reprod. Dom. Anim. 34, 237±245. Kaneene, J.B., Miller, R.A., Herdt, T.H., Gardiner, J.C., 1997. The association of serum nonesteri®ed fatty acids and cholesterol, management and feeding practices with peripartum disease in dairy cows. Prev. Vet. Med. 31, 59±72. Kermarrec, N., Chollet-Martin, S., Beloucif, S., Faivre, V., Gougerot-Pocidalo, M.A., Payen, D.M., 1998. Alveolar neutrophil oxidative burst and beta2 integrin expression in experimental acute pulmonary in¯ammation are not modi®ed by inhaled nitric oxide. Shock 10, 129±134. Klein, J.B., Payne, V., Schepers, T.M., McLeish, K.R., 1990. Bacterial lipopolysaccharide enhances polymorphonuclear leukocyte function independent of changes in intracellular calcium. In¯ammation 14, 599±611. Lotthammer, K., Wittkowski, G., 1994. Scheiden- und GebaÈrmuttererkrankungen. In: Lotthammer, K.H., Wittkowski, G. (Eds.), Fruchtbarkeit und Gesundheit der Rinder. Ulmer, Stuttgart, pp. 60±66. Lynn, W.A., Raetz, C.R., Qureshi, N., Golenbock, D.T., 1991. Lipopolysaccharide-induced stimulation of CD11b/CD18 expression on neutrophils. Evidence of speci®c receptor-based response and inhibition by lipid A-based antagonists. J. Immunol. 147, 3072±3079. Marra, M.N., Wilde, C.R., Grif®th, J.E., Snable, J.L., Scott, R.W., 1990. Bactericidal/permeability-increasing protein has endotoxin-neutralizing activity. J. Immunol. 144, 662±666. Martich, G.D., Danner, R.L., Ceska, M., Suffredini, S.A.F., 1991. Interleukin 8 (IL-8): a potent activator of neutrophils. J. Exp. Med. 173, 1021±1024. Morrow, D.A., Hillman, D., Dade, A.W., Kitchen, H., 1979. Clinical investigation of a dairy herd with the fat cow syndrome. J. Am. Vet. Med. Assoc. 174, 161±167. Naessens, J., Olubayo, R.O., Davis, W.C., Hopkins, J., 1993. Cross-reactivity of workshop antibodies with cells from domestic and wild ruminants. Vet. Immunol. Immunopathol. 39, 283±290. Noakes, D.E., Wallace, L., Smith, G.R., 1991. Bacterial ¯ora of the uterus of cows after calving on two hygienically contrasting farms. Vet. Rec. 128, 440±442. Olson, J.D., Ball, L., Mortimer, R.G., Farin, P.W., Adney, W.S., Huffman, E.M., 1984. Aspects of bacteriology and endocrinology of cows with pyometra and retained fetal membranes. Am. J. Vet. Res. 45, 2251±2255. Oostveldt van, K., Burvenich, C., Silva da, F.M., Massart-Leen, A.M., 1999. Respiratory burst activity in activated and unstimulated isolated bovine blood neutrophils during experimentally induced Escherichia coli mastitis. J. Dairy Res. 66, 375±383. Pechhold, K., Pohl, T., Kabelitz, D., 1994. Rapid quanti®cation of lymphocyte subsets in heterogeneous cell populations by ¯ow cytometry. Cytometry 16, 152±159. Roets, E., Burvenich, C., Diez-Fraile, A., Noordhuizen-Stassen, E.N., 1999. Evaluation of the role of endotoxin and cortisol on modulation of CD18 adhesion receptors in cows with mastitis caused by Escherichia coli. Am. J. Vet. Res. 60, 534±540. Roth, J.A., Kaeberle, M.L., Loren, L.H., 1983. Association of increased estradiol and progesterone blood values with altered bovine polymorphonuclear leukocyte function. Am. J. Vet. Res. 44, 247±253. Sapatnekar, S., Kao, W.J., Anderson, J.M., 1997. Leukocyte±biomaterial interactions in the presence of Staphylococcus epidermidis: ¯ow cytometric evaluation of leukocyte activation. J. Biomed. Mater. Res. 35, 409±420. Schuberth, H.J., Anders, I., Pape, U., Leibold, W., 1992. One-dimensional isoelectric focusing and immunoblotting of equine major histocompatibility complex class I antigens. Anim. Genet. 23, 87±95.
H. Zerbe et al. / Veterinary Microbiology 79 (2001) 351±365
365
Schuberth, H.J., Freigofas, R., Daugschies, A., Leibold, W., 2000. Assessment of antibody-independent cellular cytotoxicity (AICC) of porcine neutrophilic granulocytes by quantitative ¯ow cytometry. Lack of modulation by larvae products of Oesophagostomum dentatum. J. Vet. Med. B 47, 607±614. Selbitz, H.J., 1992. Enterobacteriaceae: Escherichia coli. In: Selbitz, H.J. (Ed.), Lehrbuch der VeterinaÈrmedizinischen Bakteriologie. Fischer, Stuttgart, pp. 76±82. Simms, H.H., D'Amico, R., 1995. Lipopolysaccharide induces intracytoplasmic migration of the polymorphonuclear leukocyte CD11b/CD18 receptor. Shock 3, 196±203. Smits, E., Burvenich, C., Guidry, A.J., Roets, E., 1998. In vitro expression of adhesion receptors and diapedesis by polymorphonuclear neutrophils during experimentally induced Streptococcus uberis mastitis. Infect. Immunol. 66, 2529±2534. Vandeplassche, M., Bouters, R., 1976. Puerperal metritis in the bovine. In: Proceedings of the Fourth International Congress on Animal Reproduction Art. Insem, Krakov, Poland, 1976, pp. 660±666. Vandeplassche, M., 1981. Neue vergleichende Aspekte der Involution und der puerperalen Metritis bei Stute, Kuh und Sau. Monatsh. VeterinaÈrmed. 36, 804±807. Watson, E.D., 1989. In vitro function of bovine neutrophils against Actinomyces pyogenes. Am. J. Vet. Res. 50, 455±458. Watson, R.W., Redmond, H.P., Wang, J.H., Condron, C., Bouchier-Hayes, D., 1996. Neutrophils undergo apoptosis following ingestion of Escherichia coli. J. Immunol. 156, 3986±3992. Watson, R.W., Rotstein, O.D., Nathens, A.B., Parodo, J., Marshall, J.C., 1997. Neutrophil apoptosis is modulated by endothelial transmigration and adhesion molecule engagement. J. Immunol. 158, 945±953. Zerbe, H., Schuberth, H.J., Hoedemaker, M., Grunert, E., Leibold, W., 1996. A new model system for endometritis: basic concepts and characterization of phenotypic and functional properties of bovine uterine neutrophils. Theriogenology 46, 1339±1356. Zerbe, H., Schneider, N., Ossadnik, C., Wensink, T., Kruip, T.A.M., Grunert, E., Leibold, W., 1998. Eigenschaften neutrophiler Granulozyten aus Blut und Uterus Peripartaler KuÈhe. Wien. TieraÈrztl. Monatsschr. 85, 304±309. Zerbe, H., Schneider, N., Leibold, W., Wensing, T., Kruip, T.A.M., Schuberth, H.J., 2000. Altered functional and immunophenotypical properties of neutrophilic granulocytes in postpartum cows associated with fatty liver. Theriogenology 54, 771±786.