264

Biochimica et Biophysica Acta, 1052 (1990) 264-272 Elsevier

BBAMCR 12681

Leukotriene B4 generation by human monocytes and neutrophils stimulated by uropathogenic strains of Escherichia coli R o b e r t S t e a d m a n , N i c h o l a s T o p l e y , J a n i c e K n o w l d e n , B e r n t Spur, J o h n W i l l i a m s Institute of Nephrology, Kidney Research Unit Foundation, University of Wales College of Medicine, Cardiff Royal Infirmary, Cardiff (U. K.) (Received 14 September 1989)

Key words: Leukotriene B4; Hemolytic activity; Lipoxygenase; (Human neutrophil); (Human monocyte); (E. coli)

The generation of the 5-1ipoxygenase product, leukotriene B4 (LTB4) by human mononuclear phagocytes (monocytes) following incubation with 25 different uropathogenic strains of Escherichia coli correlated with the haemolytic activity of the strains (r--0.572, P < 0.01). LTB4 generation by human neutrophils (PMN), however, was unrelated to this haemolytic potential ( r = 0.164). In contrast, both prelabelled monocytes and PMN were stimulated by haemolytic strains of E. coli and by haemolytic culture supernatants to release significant amounts of [3Hlarachidonic acid. There was a significant correlation between haemolytic activity and [3H]arachidonic acid release generated by individual strains from monocytes (r-- 0.804, P < 0.001) and PMN (r-- 0.888, P < 0.001). In addition, nonhaemolytic strains but not their culture supernatants were capable of causing slow release of both [3H]arachidonic acid and L T B 4 from PMN and mononuclear cells. These results suggest that both the possession of haemolytic activity, and the direct interaction of bacteria with the leukocyte surface are mechanisms by which uropathogenic strains of E. coli may cause the release and metabolism of araehidonic acid. In addition, there was synergistic augmentation by nonhaemolytic bacteria of the PMN LTB4 response to haemolytic culture supernatants or to low doses of the calcium ionophore A23187. These results support an ionophore-like mechanism for the activation of the cell by haemolysin. L I B 4 generation by PMN incubated with haemolytic supernatants was also augmented by particulate zymosan in a manner dependent on the dose of zymosan, suggesting that the direct interaction of E. coli with PMN may involve an activation mechanism similar to that for zymosan. These results demonstrate differing responses of peripheral mononuclear cells and PMN from the same donors to identical strains of E. coli and suggest that the generation of the potent chemotactic agent L I B 4 in response to E. coli infection in vivo need not depend solely on the elaboration of cytotoxic haemolysins by individual strains.

Introduction

The interaction of Escherichia coli and host cells occurs through a variety of bacterial surface structures and secretory products. Proteinaceous fimbriae [1] and flagella [2,3] are believed to increase adherence whilst other structures such as the glycocalyx or polysaccharide capsule may decrease the susceptibility of the organism to phagocytosis [4,5]. Type 1 fimbriae have

Abbreviations: PBS, phosphate-buffered saline; LTB4, leukotriene B4; PMN neutrophils; KRPG, Krebs-Ringer phosphate buffer including 11 mM glucose; SRBC, sheep erythrocytes; BSA, bovine serum albumin; LDH, lactate dehydrogenase; HERE, hydroxy Correspondence: J. Williams, Kidney Research Unit Foundation, University of Wales College of Medicine, Cardiff Royal Infirmary, Cardiff, CF2 1SZ, U.K.

been shown to enhance the binding of E. coli to leukocytes and to promote phagocytosis [6,7]. In addition, recent studies have demonstrated that the possession of Type 1 fimbriae by uropathogenic strains of E. coli correlates significantly with the activation of the human neutrophil (PMN) respiratory burst measured as luminol dependent chemiluminescence [1,8,9] and with the release of lysosomal granule enzymes [5]. In contrast, however, it has been suggested that it is the capacity of an organism to generate a-haemolysin, the only truly secreted protein of E. coli having a molecular weight of 107000, which is responsible for the activation of the 5-1ipoxygenase pathway of human neutrophils leading to leukotriene generation [10-12]. In animal models there is a marked increase in the pathogenicity of strains elaborating a-haemolysin [1315] which, in addition to its haemolytic ability, may also have a less specific cytotoxic role [16-18] through the

0167-4889/90/$03.50 © 1990 Elsevier Science Pubfishers B.V. (Biomedical Division)

265 generation of transmembrane pores facilitating the influx of extracellular ions [19,20]. Previous work used cloned E. coli strains and plasmid transformed strains to assess the influence of haemolysin production on human and rat leukocyte stimulation [10-12]. The present work was undertaken to assess the effect on human peripheral neutrophils and monocytes of a variety of wild-type uropathogenic strains of E. coli defined by their pattern of fimbriation and ability to produce haemolysin but with no experimentally induced changes in their genome. Materials

and Methods

Reagents. Unless otherwise stated, all cell and bacterial washing stages were carried out in calcium and magnesium free phosphate-buffered saline (PBS) (Oxoid, Basingstoke, U.K.), pH 7.3. Leukocyte stimulations were carried out in a Krebs-Ringer phosphate buffer (pH 7.4) containing 0.54 mM Ca 2+, 1.2 mM Mg 2+ and 11.0 mM D-glucose (KRPG). A23187 was from Cambridge Bioscience, Cambridge, U.K. and stored (10 mM) at - 70 ° C in DMSO until diluted immediately before use. Zymosan (Sigma, Poole, U.K.) was prepared by boiling for 30 min and washing in normal saline prior to sonication to disrupt clumps. It was stored at 4°C until counted in a Coulter counter ZM (Coulter Electronics, Luton, U.K.) and used. Unless otherwise stated, all reagents were purchased from BDH, Poole, U.K., and were of analytical grade. All radiochemicals were purchased from New England Nuclear Dupont U.K., Stevenage, U.K. unless stated otherwise. Bacterial strains. 25 uropathogenic strains of E. coli were subcultured at least three times overnight either in Oxoid nutrient broth No. 2 or on Oxoid DST nutrient agar. Broth cultures were harvested by centrifugation and washed in PBS twice before use. Agar cultures were harvested by washing off the plate in PBS followed by centrifugation and washing twice in the same buffer. Bacteria were resuspended to the required optical absorbance (A) at 560 nm (A = 1.0 = 5- 108 colony forming units/ml (cfu/ml)). The 25 strains were characterised by their capacity for mannose-sensitive haemagglutination of guinea-pig or human erythrocytes and by their haemolytic potential (Table I). The haemolytic capacity of individual strains was measured as follows. (a) By plating onto DST agar plates containing 10% (v/v) sheep erythrocytes. Zones of fl-haemolysis were scored as complete (+ + ), patchy ( + ) or absent ( - ) . (b) Either live bacteria or their culture supernatants made cell free by filtration through a 0.22 #m filter (Millipore U.K., Harrow, U.K.) or by centrifugation at 11 000 × g for 5 min were added in a volume of 200 #1 to 200 #1 of a 10% (v/v) suspension of sheep erythrocytes and the suspension diluted with 600 #1 of KRPG.

TABLE I

Strains used in this study Strain

Serotype

Genotypic

Haemolysis d

fimbriation c

of SRBC plates

O a

K b

Type 1

P

DW

02

Ku

+

+

+ +

103B DB

AA 025

K5 Ku

+

+

+ -

88 DF

NT 011

Ku Ku

-

+ -

+ +

439

06

K5

+

-

-

57 63

NT NT

Ku Ku

+ +

+ +

+ +

ER 2 T1 Tre

04 075 NT

Ku Ku Ku

+ -

+ + -

+ + + +

49

08

Ku

+

+

+ +

SC NK 1 GG 2 KV 436

01 04 04 NT 04

K1 Ku Ku Ku Ku

+ + + +

+ + + + +

+ + + + -

Ks71,B 504 56

06 06 04

Ku Ku Ku

+ + +

+ + +

+ + + +

7 297 12

018 018 0.4

K5 Ku Ku

+ +

+ -

+ + + +

417 168

01 04

Ku Ku

+ +

-

+ -

a A A - a u t o a g g l u t i n a b l e ; N T - s m o o t h b u t n o t 01, 02, 04, 05, 06, 07, 08, 09, 011, 015, 017, 018, 025 or 075. b K u - n o t K1 o r K5. c T y p e 1 f i m b r i a t i o n c o n f i n e d b y m a n n o s e sensitive h a e m a g g l u t i n a tion of g u i n e a - p i g T B C . P - f i m b r i a t i o n c o n f i r m e d following m a n n o s e resistant h a e m a g g l u t i n a t i o n o f h u m a n R B C b y P F test latex agg l u t i n a t i o n kit. d H a e m o l y s i s scored as + + ( c o m p l e t e ) , + ( p a t c h y ) o r -

(negative).

The suspensions were incubated at 37°C for 60 min. The cells were then pelleted by centrifugation at 11 000 × g for 1 min. The absorbance of the supernatants (at 415 nm) was measured (with dilution where necessary) and the results expressed as a percentage of total lysis (200 #1 of erythrocytes lysed in water at the same dilution). All readings were corrected for controls consisting of erythrocytes incubated alone in KRPG at 37 °C for 60 min. One haemolytic unit was arbitrarily defined as that concentration of culture supernatant causing 1% lysis of the SRBC after 60 min. Measurement of bacterial numbers by A TP extraction. 100 #1 of a bacterial suspension was mixed for 10 s with 100 ~tl of a bacterial extraction reagent (Bac-Extract) prepared in this laboratory, (British Patent Application No. 851300). 800 #1 of 25 mM Hepes buffer (pH 7.5) containing 2 mM EDTA, was then added. 100 #1 of this dilution was added to the firefly bioluminescence reagent (ATP bioluminescence reagent, Boehringer Corporation (London), East Sussex, U.K.) and the lumines-

266 cence was measured in a Lumac Biocounter M2010 (Lumac, Landgraaf, The Netherlands). Neutrophil preparation. Normal human leukocytes were isolated from citrated peripheral blood by dextran sedimentation, rendered plasma-free and platelet-poor by centrifugation at 300 x g, and washed with PBS three times. Neutrophils (PMN) and mononuclear cells were separated by density gradient centrifugation at 400 x g for 35 min at 23°C on cushions of Ficoll-Hypaque (Pharmacia, Milton Keynes, U.K.). Following hypotonic lysis of erythrocytes, the PMN were counted in a Coulter counter and resuspended to a concentration of 5 . 1 0 6 cells/ml PBS. The cells were over 98% PMN by morphology after centrifugation (Cytospin II, Shandon Southern Products, Runcorn, U.K.) and Neat staining (Guest Medical, Sevenoaks, U.K.). Mononuclear cells were decanted from the Ficoll-Hypaque interface and washed in PBS (twice) prior to resuspension (2.106 cells/ml) in RPMI 1640 (Gibco, Paisley, U.K.) containing 2% ( w / v ) BSA (ICN Biomedicals, High Wycombe, U.K.) (RPMI-BSA). P M N stimulation. 200 /~1 of PMN (1 - 1 0 6 cells) suspended in PBS were incubated with 200/~1 of bacterial suspension of culture supernatant and 600 ~1 K R P G at 37 ° C for up to 3 h. All incubations were performed in duplicate. In some experiments zymosan particles replaced E. coli as the particulate ligand. At the end of the incubation period samples from the P M N / b a c t e r i a l suspension were taken for haemagglutination and the remaining cells and bacteria were pelleted by centrifugation for 1 min at 11 000 x g. Duplicate samples were taken to assay for lactate dehydrogenase activity, immunoreactive LTB4 and haemolytic activity. In every experiment control incubations were included consisting of neutrophils incubated in K R P G alone, bacteria incubated in K R P G alone and neutrophils incubated with the calcium ionophore A23187 at 5 ffM concentration. Monocyte stimulation. Mononuclear cells from the Ficoll-Hypaque interface (2.106 cells) suspended in 1 ml RPMI-BSA were plated onto 35 mm tissue culture dishes (Gibco) and incubated at 3 7 ° C in 95% air/5% CO 2 for 30 min. Non-adherent cells were removed by washing and a further aliquot of 2 . 1 0 6 cells was added and incubated for a further 30 min in 5% CO 2 at 37°C. The number of adherent cells was estimated by counting the non-adherent component. Adherent monocytes (less than 5% lymphocytes) were stimulated with bacterial suspensions or culture supernatants as described for PMN. Incubations were terminated by removing the supernatants which were centrifuged for 1 min at 11 000 x g to remove bacteria and detached cells. Samples were then taken for L D H and LTB4 assay. Lactate dehydrogenase assay. The percentage release of L D H from leukocytes was calculated, after subtrac-

tion of the appropriate blank values, as a percentage of that released from cells disrupted by sonication, for two periods of 2 min, at 8 ~m peak-to-peak distance at 4 ° C in an MSE 150 W ultrasonic disintegrator (MSE, Crawley, U.K.) or, for monocytes, by solubilisation in 1% ( v / v ) Triton X-100 [5]. 300 /~1 of PMN supernatant and 200 ffl of 10 mM reduced N A D H (Sigma) was added to 2.4 ml of 0.1 M phosphate buffer (pH 7.4). The reaction was started with 100 /~1 of potassium pyruvate (1 m g / m l ) and the rate of change in absorbance at 340 nm was measured over 4 min at room temperature on a Cecil CE 292 spectrophotometer (Cecil Instruments, Cambridge, U.K.). Culture supernatants. Supernatants from the 25 strains of E. coli grown for 5 h in K R P G were added to PMN or adherent monocytes and after 1 h of incubation the percentage of L D H released and the amount of [3H]arachidonic acid and immunoreactive LTB4 generated were measured. The absence of bacteria in the supernatants was confirmed using the bacterial ATP extraction assay. L T B 4 radioimmunoassay. The radioimmunoassay (RIA) for LTB 4 was a modified version of that described previously [21]. Briefly, 100 ffl of supernatant from the cell stimulation was assayed in duplicate by incubating at 37 ° C overnight with 100/zl of antiserum at a final dilution of 1 / 1 2 0 0 0 in RIA buffer (10 mM Tris-HC1 containing 0.9% ( w / v ) NaC1, 0.01% (w/v) sodium azide and 0.1% ( w / v ) gelatin (pH 7.3)) and with 100 ffl (5000 cpm) of tracer, [5,6,8,9,11,12,14,15(n)3H]LTB4 (200 C i / m m o l ) (Amersham International, Aylesbury, U.K.). Dilutions of LTB4 standard from 10 n g / m l to 0.04 n g / m l (1 ng to 4 pg per tube) were included in each assay together with tubes to estimate nonspecific binding and total added tracer concentrations. Free [3H]LTB4 was separated from that bound to antibody by adsorption of non-protein-bound tracer to dextran-coated charcoal (1% ( w / v ) dextran T70 (Pharmacia), 1% ( w / v ) GSX charcoal) at 4 ° C followed by centrifugation at 3500 x g and 4 ° C for 10 min. The supernatant was decanted into 4 ml Optiphase MP (LKB) and counted in an LKB rackbeta scintillation counter (Pharmacia LKB, Milton Keynes, U.K.). Characterisation of immunoreactive L T B 4 released from leukocytes. Supernatants or cell extracts containing immunoreactive LTB4 were injected in a total volume of 500/zl of 50 : 50 ( v / v ) m e t h a n o l / w a t e r onto a Nucleosil C18 5 ff reversed phase column (25.4 cm x 4.6 mm) (Hichrom, Reading, U.K.) protected by a 6 cm guard column of the same packing. The samples were eluted for 30 min in a m e t h a n o l / w a t e r / a c e t i c acid (65 : 35:01) mobile phase brought to pH 5.6 with ammonium hydroxide. The gradient was then increased to 75% methanol over 1 min and continued for a further 25 min before changing to 100% methanol for 22 min. The flow

267 rate was maintained at 1 m l / m i n throughout and 1 ml fractions were collected. Each fraction was stored under argon at - 2 0 ° C until analysed. Ultraviolet absorption was monitored throughout each run at 270 nm for 35 min then 235 nm for a further 35 rain and peak absorption was integrated on a Gilson 620 DataMaster (Anachem, Luton, U.K.) 100 /~1 duplicates of each fraction were dried down under vacuum and resuspended in R I A buffer. The elution time of P M N generated immunoreactive LTB 4 and integrated ultraviolet absorption were compared to those of authentic LTB 4 and [3H]LTB4 in this system. In addition the ultraviolet spectrum was established after pooling and concentrating the fractions containing immunoreactive LTB4 from leukocytes stimulated by individual E. coli strains. The reversed-phase H P L C (RP-HPLC) system was validated using authentic standards and tritium-labelled leukotrienes and monoHETEs. Authentic 20-carboxyLTB4 eluted at 9.5 _+ 1.2 min (mean _+ S.D.) (n = 3); LTC 4 eluted at 16.3 _+ 1.3 rain (n = 3); the all-trans-isomers of LTB 4 at 19.3 _ 0.6 (n = 3) and 21.6 _ 1.8 min (n = 3); LTB4 at 25.3 _+ 1.7 (n = 3); (5S,12S)-LTB 4 at 26.8 + 1.5 min (n = 3); 15-HETE at 4 4 . 3 _ 2.1 min (n = 3); 5-HETE at 51.6 + 2.0 min (n = 3); [3H]arachidonic acid~at 65.2 _+ 1.9 rain (n = 5). [3H]LTC4, [3H]LTB4, 15-[3H]HETE and 5-[3H]HETE all co-eluted with their respective authentic standards.

Metabolism of [3H]LTB4 by stimulated leukocytes. 100 pl of P M N suspended in PBS ( 1 - 1 0 7 / m l ) or five plates containing approx. 2 . 1 0 6 m o n o c y t e s / p l a t e were incubated in K R P G with strains of E. coli as described above but in the presence of approx. 1-104 cpm of added [3H]LTB4 (200 C i / m m o l ) . Incubations were carded out for varying times at 3 7 ° C and stopped by centrifugation at 11000 x g for I rain. The supernatants were decanted and the cell pellets extracted with 1 ml methanol for 60 rain at - 2 0 ° C. The methanol extract and the supernatant were combined and left at - 2 0 ° C for 60 rain. Each sample was then cleared by centrifugation at 11 000 x g for 1 min prior to separation in the R P - H P L C system described above. 1 ml fractions were collected and added to 4 ml Optiphase MP for scintillation counting. Leukocyte labelling with [3H]arachidonic acid. Normal human neutrophils (PMN) or mononuclear cells were suspended to a concentration of 1 • 1 0 7 / / m l in PBS and 0.1 /~Ci of [3H]arachidonic acid (94.5 C i / m m o l ) was added in 1/xl of ethanol. The cells were incubated at 35 ° C for 60 rain with occasional mixing. The P M N were pelleted by centrifugation at 300 x g for 7 rain, washed twice in PBS and resuspended to 1 • 107/ml in PBS. The mononuclear cells were plated as normal, non-adherent cells being counted to estimate the number of a d h e r e n t m o n o c y t e s . I n c o r p o r a t i o n of [3H]arachidonic acid into P M N was 67.3% (range 52.7

to 77.4%) and into mononuclear cells 73.5 (range 61.6 to 85.9%).

Leukocyte stimulation for [3H]arachidonic acid release. Stimulation of pre-labelled 1 • 106 P M N or 2- 106 adhered monocytes with E. coli strains was carried out in 1 ml K R P G for 60 min as described above. The P M N incubation was terminated by centrifugation for 1 min at 11000 x g. The monocyte incubations were terminated by decanting the supernatants and centrifuging at 11 000 x g to remove bacteria and detached cells. The supernatants were decanted and processed for scintillation counting. The pelleted P M N were extracted at - 2 0 ° C with 1 ml methanol for 60 min and added directly to 4 ml Optiphase MP for scintillation counting. The pelleted monocytes were resuspended in 1 ml methanol and added to the original tissue culture plate for 1 h at - 2 0 ° C prior to scraping the plate and adding the extract to scintillant for counting. The proportion of [3H]arachidonic acid and metabolites released into the supernatant was calculated as a percentage of the total counts per minute in both supernatant and cell pellet and compared to that released from unstimulated cells. This percentage release was compared to that generated by 60-rain incubations with 5 / t M A23187. In addition, the cell extracts and the supernatants from cell stimulations were separated on the R P - H P L C system described above in order to identify the [3H]arachidonic acid metabolites released during stimulation. The supernatants were added to 2 vol. of methanol and incubated overnight at - 2 0 ° C under argon before cleating by centrifugation at 11 000 x g. The cell extracts were concentrated, diluted with 1 vol. of HPLC-grade water and injected onto the column. Fractions of 1 ml were collected very minute for 70 rain and added to 4 ml Optiphase MP for scintillation counting.

Results

Immunoreactive LTB 4 release from monocytes incubated with E. coli strains In three experiments using monocytes from three different donors, 21 of the 25 E. coli strains consistently caused significant immunoreactive LTB 4 release (1.81 + 1.37 ng, mean + S.D.) from 1 . 1 0 6 cells following incubation at 3 7 ° C for 60 min with 1 . 1 0 8 cfu. The concentration of LTB 4 generated (up to 4.9 n g / 1 . 1 0 6 monocytes) was significantly correlated (r = 0.572, P < 0.01) with the haemolytic activity of both bacterial cells (Fig. la) or their K R P G culture supernatants. The calcium ionophore A23187 (5 /~M) caused the generation of 11.1 + 3.4 n g / 1 • 106 cells (mean + S.D.) of immunoreactive LTB 4 from monocytes after 60 min at 3 7 ° C (n = 5).

268 5

~"

a

3-

Z

:s

%

£L

4

% 3

"~

0 •

2

c v

2 o $

1

§

o

o 0

o %0o0oO

•

rn

0 o

m

0

i

i

0

1

i

QO

0

0

0

I

0 1 Log. Percentage haemolysis

Log. Percentage haemolysis

Fig. 1. The generation of immunoreactive LTB4 by (a) 1 • 106 monocytes and (b) 1.106 PMN following 60 min incubation in KRPG at 37 ° C with 1 • 108 cfu of each of 25 strains of E. coli compared to the log of the haemolytic activity of each strain in the same buffer. Each point represents the mean of three experiments each with cells from a different donor.

Immunoreactive L T B 4 release from P M N incubated with E. coli strains In three experiments using cells from the same three donors 20 of the 25 strains of E. coli tested consistently produced significant release of immunoreactive LTB 4 (0.76 +_ 0.45 ng, mean _+ S.D.) from 1 • 10 6 P M N following incubation with 1 • 108 cfu at 3 7 ° C for 60 min (Fig. lb). The generation of LTB 4 from PMN, however, did not correlate with either an increase in the haemolytic activity of the stimulating strain (r = 0.164) or its culture supernatant (r = 0.132). N o r was there a significant difference between the levels of LTB 4 released in response to type 1 fimbriate strains (0.61 _+ 0.57 n g / 1 • 10 6 cells), non-fimbriate strains (0.70 _+ 0.30 n g / 1 • 10 6 cells) or P fimbriate strains (0.57 + 0.38 n g / 1 • 10 6 cells). The calcium ionophore A23187 (5/~M) caused generation of 9.4 _+ 2.2 n g / 1 • 1 0 6 cells (mean + S.D.) of immunoreactive LTB 4 after 60 min incubation with P M N at 3 7 ° C (n = 5).

Time-dependent generation of immunoreactiue L T B 4 from leukocytes Immunoreactive LTB 4 generation from 1-106 human P M N or adherent monocytes by the haemolytic E. coli strain 504 was dose- and time-dependent, reaching a plateau of generation by 60 min at a bacteria:cell ratio of 100:1 (Fig. 2a). In contrast, the significant generation of LTB 4 by cells incubated with SC, which lacked haemolytic activity, was markedly slower for both monocytes and P M N (Fig. 2b).

Characterisation of immunoreactioe L T B 4 from leukocytes Following R P - H P L C , a peak of immunoreactive LTB 4 was detected in the supernatant from both P M N and monocytes stimulated by E. coli 504 and had retention times of 24.7 _+ 2.3 min (n = 10) and 23.2 +_ 2.6 min (n = 3) for each cell type, respectively. This immuno-

2a

%

%

rm I.-.

T

133 t'-,.J

0 0

20

40 Time (min)

60

•

0

i

10

•

i

20

•

i

'

30

i

40

'

'

50

'....

60

Time (min.)

Fig. 2. The generation of immunoreactive LTB4 from 1.106 monocytes (O) or PMN ( 0 ) following incubation with 1.108 cfu of (a) E. coil 504 or (b) E. coli SC for up to 60 rain. Each represents the mean of three experiments each with cells from a different donor.

269 25

20-

t~

o

t~

20

"9 .o_

16

0

-£_ t~ "1¢o

o

15

±

12,

0

0 O0

0

cb t~

a.

10

'

1

•

0

'

1

i

O_

2

8

.

o

Log. Percentage haemolysis

,

.

,

1

2

Log. Percentage haem01ysis

Fig. 3. The percentage release of incorporated [3H]arachidonic acid from (a) 1.106 monocytes or (b) 1.106 P M N following 60 min incubation in K R P G at 37 o C with 1.108 cfu of each of 25 strains of E. coli compared to the log of the haemolytic activity of the strains in the same buffer. Each point represents the mean of three experiments each with cells from a different donor.

reactive LTB4 co-eluted with both authentic LTB4 and [3H]LTB4. These immunoreactive LTB4 peaks from PMN or monocytes (measured in the RIA as 22 or 35 ng/ml, respectively) were detected by ultraviolet absorption at 270 nm and had 72.3% and 77.2% homology with this integrated peak of ultraviolet absorption. The material in the pooled and concentrated fractions containing immunoreactive LTB4 had a typical triene ultraviolet absorption spectrum with a peak absorbance at 269 nm and shoulders at 260 and 279 nm.

Metabolism of [3H]LTB4 by leukocytes Exogenous [3H]LTB4 was metabolised by unstimulated PMN in a time-dependent manner, reaching 45.1 ___5.6% (mean + S.D.) at 60 min. This metabolism was not affected by incubating PMN with haemolytic or non-haemolytic strains of E. coil and was not dependent on the dose of [3H]LTB4 but was increased with increasing numbers of PMN in the incubation medium. The added [3H]LTB4 was metabolised to a product co-eluting on RP-HPLC with the 0~-oxidative products of LTB4. There was no metabolism of [3H]LTB4 to ~o-metabolites by monocytes.

Release of [3H]arachidonic acid from leukocytes All of the haemolytic strains of E. coli tested caused significant release of [3H]arachidonic acid from monocytes or PMN pre-labelled at 35 ° C for 60 min (Fig. 3a, b). This release was highly correlated with the haemolytic potential of the stimulating strain (r = 0.804 and 0.888 for monocytes and PMN, respectively). The generation of immunoreactive LTB4 was significantly correlated with the release of [3H]arachidonic acid from monocytes but not PMN ( r = 0 . 4 9 5 and

0.201, respectively). In both cases, however, a peak of radioactivity co-eluting with [3H]LTB4 could be detected following RP-HPLC. The percentage of the total cpm, recovered in this peak following stimulation with the haemolytic strain 504 was 3.4 + 1.2% and 1.1 + 0.6% (mean + S.D.) for monocytes and PMN, respectively.

Lactic dehydrogenase release There was no significant release of cytoplasmic lactic dehydrogenase ( < 10%) from PMN or monocytes following incubation at 3 7 ° C with any of the E. coil strains or culture supernatants tested for times up to 180 rain.

4

[] []

Z n

- 0.5 HU + 0.5 HU

3

% 2

I

0

PMN C/qLY

504

SC

10 3

DF

88

12

Strain Fig. 4. The generation of immunoreactive LTB 4 from 1.106 P M N incubated at 37 ° C for 60 min in K R P G containing E. coil 504, SC 103, DF, 88 and 12 alone or in the presence of 0.5 H U of E. coil 504 culture supernatant. The results are the mean 5: S.D. of three experiments each with cells from a different donor.

270 1.2

z~ a_

8z; ~-

[]

- 0.5 HU

[]

+ 0.5 HU

~o

[] []

- A23187 + A23187

6

0.9

%

b

4

N

2

0.6

v

0.3

5

Zymosan particles ( ×108 ) Fig. 7. The generation of immunoreactive LTB4 from 1.106 PMN in response to doses of zymosan particles alone or in the present of 0.1 p~M A23187. The results are the mean_+S.D, of three experiments each with cells from a different donor.

0.0 1

10

Zymosan particles ( xl08 ) Fig. 5. The generation of immunoreactive LTB4 from 1-106 PMN in response to dose of zymosan particles alone or in the presence of 0.5 HU of E. coli 504 culture supernatant. The results are the mean + S.D. of three experiments each with cells from a different donor.

The haemolytic activity of E. coli strains 18 of the E. coil strains tested caused greater than 10% haemolysis of 2% SRBC following subculture in broth. The ability of E. coli strains to release soluble ~xhaemolysin during a 1 h incubation was confirmed by assay of bacterial culture filtrates against SRBC, and correlated closely with the haemolytic ability of the stimulating strain. Augmentation of L T B 4 generation from P M N Aliquots of haemolytic culture supernatants ( < 0.5 HU), which did not cause detectable immunoreactive

LTB4 generation from PMN, stimulated significant LTB4 release following co-incubation with live bacteria, whether or not these were haemolytic (Fig. 4). There was no augmentation in response to the coincubation of non-haemolytic supernatants with PMN in the presence of bacterial cells. Similarly, incubation of 0.5 HU of a haemolytic supernatant with PMN in the presence of unopsonized zymosan also resulted in a synergistic augmentation of LTB 4 generation (Fig. 5). The LTB4 response of PMN to 0.1 # M A23187 was also augmented by coincubation with bacterial cells of the same strains (Fig. 6). In addition, coincubation of PMN with substimulatory doses of the calcium ionophore A23187 (0.1 /aM) and the same doses of particulate zymosan also resulted in a synergistic augmentation of the LTB 4 response (Fig. 7). Discussion

[]

- A23187

[]

+ A23187

Z Q_

% 32

h

1' 0 PMN

504

SC

1 03

DF

88

12

ONLY Strain

Fig. 6. The generation of immunoreactive LTB 4 from 1-106 PMN incubated at 37 ° C for 60 rain in KRPG containing E. coli 504, SC, 103, DF, 88 and 12 alone or in the presence of 0.1 gM A23187. The results are the mean + S.D. of three experiments each with cells from a different donor.

Uropathogenic strains of E. coli generated immunoreactive LTB4 from human neutrophils and monocytes in a dose- and time-dependent manner. The PMN response was not dependent on the fimbrial type, nor was it associated with an increasing haemolytic potential of the stimulating strains. In contrast, the monocyte response was strongly correlated with the haemolytic activity of these strains. Both the ability of E. coli strains to elaborate ahaemolysin and the enhanced adherence of E. coli to leukocytes through the expression of surface fimbriae have been previously linked with the activation of the 5-1ipoxygenase pathway [10-12]. a-Haemolysin is generated during the logarithmic phase of the E. coli growth cycle [22]. Fimbriae, however, are expressed optimally in the stationary phase [23]. In the present study optimum immunoreactive LTB4 generation was independent of the growth cycle phase, since both agar and broth grown organisms could cause activation of the 5-1ipoxygenase enzyme, even though there was a reduc-

271 tion in the ability of agar grown strains to produce a-haemolysin during a 60 min incubation compared to broth grown strains (data not shown). Activation of the 5-1ipoxygenase pathway has been demonstrated following PMN stimulation by a variety of soluble and particulate ligands [24-30]. Leukotrienes are potent lipid mediators of inflammation. In particular, LTB4 is highly chemotactic for human neutrophils and eosinophils and is formed by the oxidative metabolism of the arachidonic acid released f r o m cell membranes following the activation of phospholipases [31,32]. LTB4 is the enzymatic product of the epoxide leukotriene A 4 which is, itself, enzymatically converted from the 5-1ipoxygenase metabolite of arachidonic acid, 5-hydroperoxyeicosatetraenoic acid. The metabolism of LTB4 in PMN, but not in monocytes, proceeds through ~0-oxidation to 20-hydroxy- and 20-carboxyLTB4 [33,34]. Besides these enzymatic products, non-enzymatic degradation of LTA 4 can also occur, causing the generation of all-trans-stereoisomers of LTB4. Because current antibodies and RIA are incapable of allowing complete specificity in the measurement of LTB4 due to crossreaction with metabolites having similar structures and functional groups, it is essential to separate LTB4 from crossreacting isomers in order to measure specific enzymatic LTB4 generation reliably. We were able to demonstrate authentic LTB4 generation by human PMN in response to strains of E. coil using this RP-HPLC separation method and also to demonstrate that by 60 rain 50% of added [3H]LTB4 could be metabolised by 1 • 107 PMN in this system to a product with a retention time consistent with that of the 0~-metabolites of LTB4, irrespective of whether the PMN had been stimulated. These results suggest that the metabolism of PMN-generated LTB4 could not account for the differences between PMN and monocyte responses. The interaction of E. coil with phagocytes may be mediated through the structural polysaccharides of the bacteria surface [4] or through the specific binding of surface structures such as type 1 fimbriae, which bind to the mannose residues of cell-surface glycoproteins [35]. There was no relationship between the generation of LTB4 from human PMN and bacterial adherence due to fimbriae, or increasing haemolytic activity. Most strains of E. coli tested were capable of causing a significant LTB4 generation from PMN, given sufficient time to interact. In contrast, the interaction of bacteria with monocytes to generate LTB4 was correlated with increasing bacterial haemolytic potential. This interaction was independent of whether the strains had been grown on agar or in broth, suggesting that optimising growth conditions which suppress or enhance fimbrial expression made no difference to the ability of strains to activate 5-1ipoxygenase activity. Our results suggest that the haemolytic activity of a particular E. coli strain is quantitatively related to its ability to activate the 5-

lipoxygenase pathway of monocytes but not that of human PMN. Under the incubation conditions used in the present study, cell lysis ( < 10% L D H released) had not occurred, indicating that cell damage did not prevent LTB4 synthesis. Previous studies demonstrating an association between a-haemolysin and cell activation have also demonstrated an increased cell response with intact cells or culture supernatants of haemolytic strains [10-12]. The release of [3H]arachidonic acid in the present study, however, was not quantitatively correlated with immunoreactive LTB4 generated from PMN by the same E. coli strains despite the presence of radioactive peaks following RP-HPLC, which co-eluted with authentic LTB4. These results suggest that, while a-haemolysin may be involved in activating phospholipases within P M N and monocytes to release arachidonic acid, this arachidonic acid may originate from different intracellular pools in the two cell types and may not necessarily be available for metabolism to LTB4 in PMN. Since non-haemolytic strains, strains lacking type 1 fimbriae and strains in the stationary phase of growth could generate immunoreactive LTB4 from leukocytes it seemed likely that a mechanism independent of type 1 fimbrial adherence or ahaemolysin was also involved in the 5-1ipoxygenase activation by these strains of E. coli. P. fimbriate and non-fimbriate E. coil strains do not adhere strongly to PMN [1]. The simultaneous incubation of strain SC (P-fimbriate only) with haemolytic supernatants, however, synergistically increased LTB4 release from PMN. This increase was also apparent following co-incubation of a haemolytic supernatant with the yeast cell wall product, zymosan. Zymosan binds to a putative fl-glucan receptor on PMN which shares structural homology with the CR3 adhesion glycoprotein [29,30,36-38]. While type 1 fimbriae specifically adhere to mannose residues on surface glycoproteins [6,35] it is not known which ligands are specifically recognised by the binding sites on the CR3 molecule and the present results suggest that there is a similar mechanism involved in both the E. coil and zymosan activation of PMN. Further studies will evaluate the possible importance of receptor occupancy in the generation of inflammatory mediators by PMN stimulated by E. coil strains. It is now apparent, however, that an interaction with cytotoxic products such as a-haemolysin is not the only mechanism by which leukocyte lipoxygenase products are generated in response to pathogenic organisms.

Acknowledgements We wish to thank Professor T. Korhonen, Finland; Professor M. Sussman, Department of Microbiology, University of Newcastle-upon-Tyne, U.K., and Dr. A. Roberts, Department of Microbiology, Chafing Cross Hospital, London, U.K., for providing some of the E.

272 coli strains used in this study. Dr. Roberts also serotyped the strains for us. We are also grateful to Miss C. Patterson and Miss R. Carter for their help in preparing this manuscript.

References 1 Svanborg Ed6n, C., Bjorksten, L.M., Hull, R., Hull, S., Magnusson, K.E., Leffler, H. and Moldavano, Z. (1984) Infect. Immun. 44, 672-680. 2 Fletcher, M. (1977) Can. J. Microbiol. 23, 1-6. 3 Harber, M.J., Mackenzie, R.K. and Asscher, A.W. (1983) J. Gen. Micro. 129, 621-632. 4 0 h m a n , L., Hed, J. and Stendahl, O. (1982) J. Infect. Dis. 146, 751-757. 5 Steadman, R., Topley, N., Jenner, D.E., Davies, M. and Williams, J.D. (1988) Infect. Immun. 56, 815-822. 6 Mangan, D.F. and Snyder, I.S. (1979) Infect. Immun. 26, 520-527. 7 Silverblatt, F.J., Dreyer, J.S. and Shauer, S. (1979) Infect. Immun. 24, 218-223. 8 Topley, N., Steadman, R., Mackenzie, R.K., Knowlden, J. and Williams, J.D. (1989) Kid. Int. 36, 609-616. 9 Topley, N., Steadman, R., Mackenzie, R.K., Williams, J.D., Davies, M. and Asscher, A.W. (1990) Rev. Infect. Dis., in press. 10 Scheffer, J., K~nig, W., Hacker, J. and Goebel, W. (1985) Infect. Immun. 50, 271-278. 11 K~nig, B., KiSnig, W., Scheffer, J., Hacker, J. and Goebel, W. (1986) Infect. Immun. 54, 886-892. 12 Scheffer, J., Vosbeck, K. and KSnig, W. (1986) Immunology 59, 541-548. 13 Hacker, J., Hughes, C., Hof, H. and Goebel, W. (1983) Infect. Immun. 42, 57-63. 14 Harber, M.J., Topley, N. and Asscher, A.W. (1986) Clin. Sci. 70, 531-538. 15 Hughes, C., Hacker, J., Roberts, A. and Goebel, W. (1983) Infect. Immun. 39, 546-551. 16 Cavalieri, S.J. and Snyder, I.S. (1982) J. Med. Microbiol. 15, 11-21.

17 Cavalieri, S.J. and Snyder, I.S. (1982) Infect. Immun. 36, 455-461. 18 Cavalieri, S.J. and Snyder, I.S. (1982) Infect. Immun. 37, 966-974. 19 Jorgensen, S.E., Mulcahy, P.F., Wu, G.K. and Louis, C.F. (1983) Toxicon. 21,717-727. 20 Bhakdi, S., Mackman, N., Nicaud, J.M. and Holland, I.B. (1986) Infect. Immun. 52, 63-69. 21 Rokach, J., Hayes, E.C., Girard, Y., Lombardo, D.L., Maycock, A.L., Rosenthal, A.S., Young, R.N., Zambani, R. and Zweerink, H.J. (1984) Prostaglandins, Leukotrienes Med. 13, 21. 22 Springer, W. and Goebel, B. (1980) J. Bacteriol. 144, 53-59. 23 Eisenstein, B.I. (1981) Science 214, 337-339. 24 Borgeat, P. and Samuelsson, B. (1979) Proc. Natl. Acad. Sci. USA 76(6), 2184-2152. 25 Borgeat, P. and Samuelsson, B. (1979) Proc. Natl. Acad. Sci. USA 76(7), 3213-3217. 26 Borgeat, P. and Samuelsson, B. (1979) J. Biol. Chem. 254, 2643. 27 Palmer, R.M.J. and Salmon, J.A. (1983) Immunology 50, 65-71. 28 Sellmayer, A., Strasser, Th. and Weter, P.C. (1987) Biochim. Biophys. Acta 927, 417-422. 29 Topley, N., Steadman, R., Petersen, M.M., Spur, B. and Williams, J.D. (1987) Adv. Biosci. 66, 123-129. 30 Williams, J.D., Lee, T.H., Lewis, R.A. and Austen, K.F. (1985) J. Immunol. 134(4), 2624-2630. 31 Bauldrey, S.A., Wykle, R.L. and Bass, D.A. (1988) J. Biol. CHem. 263, 16787-16795. 32 Shibata, Y., Abiko, Y. and Takiguchi, H. (1988) Biochim. Biophys. Acta 971, 121-126. 33 Hanson, G., Lindgren, J.A., Dahlen, S.E., Hedquist, P. and Samuelsson, B. (1981) FEBS Lett. 130, 107. 34 Shak, S. and Goldstein, I.M. (1984) J. Biol. Chem. 259, 10181. 35 Rodrigues-Ortega, M., Ofek, I. and Sharon, N. (1987) Infect. Immun. 55, 968-973. 36 Czop, J.K., Puglisi, A.V., Miorandi, D.Z. and Austere K.F. (1988) J. Immunol. 141, 3170-3176. 37 Ross, D.G., Cain, J.A. and Lachmann, P.J. (1985) J. Immunol. 134, 3307. 38 Williams, J.D., Topley, N., Alobaidi, H.M. and Harber, M.J. (1986) Immunology 58, 117-124.