Acta path. microbiol. scand. Sect. B, 85: 374-380,
RABBIT POLYMORPHONUCLEAR LEUKOCYTE MIGRATION IN VZTRO IN RESPONSE TO LIPOPOLYSACCHARIDES FROM BAG T E ROZDES, FUSOBAC T E RIUM AND V E I L L O N E L L A KJELL SVEEN The Gade Institute, Department of Microbiology, Laboratory of Oral Microbiology, University of Bergen, Bergen, Norway
Sveen, K. Rabbit polymorphonuclear leukocyte migration in nitro in response to lipopolysaccharides from Buctcroidcs, Fusobuctcrium and Vcillonclla. Acta path. microbiol. scand. Sect. By 85: 374-380, 1977. Purified lipopolysaccharides (LPS) from strains of Buctcroidcs, Fusobuctcriurn and Vcilloncllu incubated with guinea pig serum, were tested for chemotactic activity against rabbit polymorphonuclear leukocytes ( PMNs) in modified Boyden chambers. Comparisons were made to a Salmonella LPS ( S . enteriditis S-795). Submicrogram amounts of LPS induced positive chemotaxis, and a typical dose-response relationship u p to certain dose levels was observed. The difference in chemotactic activity between the Vcillonclla, LPS and LPS-S-795 was not statistically significant. The Fusobuctcrium LPS showed either a non-significant or a highly significantly lower chemotactic capacity than LPS-S-795. The Bucteroidcs LPS were also clearly chemotactic, but considerable less when compared to the Sulmonclla LPS. When serum was not added, the LPS preparations showed no chemotactic activity. Key words: Bucteroidcs; Fusobuctcrium; Vcilloncllu; lipopolysaccharides; leukocyte chemotaxis; Boyden chambers. Kjell Sveen, Mikrobiologisk avdeling, MFH-bygget, N-50 16 Haukeland sykehus, Norway.
Activation of the complement system by bacterial lipopolysaccharides (LPS) when incubated with serum from different species leads to the generation of a factor chemotactic for polymorphonuclear leukocytes (PMNs) (9).During this activation process there is a consumption of the terminal components of C (C3 through C9) and a relative sparing of the earlier acting components by LPS (3, 4, 8 ) .
In examination of the activity of cytotaxins and cytotaxigens on PMNs, the Boyden chamber device has been extensively used. The aim of this study was, by use of a modified Boyden chamber device, to examine if LPS purified from the anaerobic Gram-negative organisms of Bacteroides fragilis, Bacteroides melaninogenicus, Fusobacterium nucleatum and Veillonella were able to elaborate a factor chemotactic for PMNs when incubated in serum from non-immunized
guinea pigs. If so,i t was of further interest to examine if any difference in this activity could be found between the various LPS preparations, and to what extent this activity was comparable to that of a Salmonella LPS.
(BSA) (Armour Pharmaceutical Company Ltd., Eastbourne, England), giving a final number of approximately 107 leukocytes per ml. Ninety-two to ninety-four per cent of the leukocytes resisted staining with trypan blue. All media had a p H 7.2-7.4, and were sterilized before use by passage through a filter with a pore size of 0.45 + (Millipore Filter Corp., Bedford, Mass., USA).
MATERIALS AND METHODS Orginisms The bacterial strains used for isolation of LPS (endotoxin) were: Bacteroides fragilis subspecies fragilis strain NCTC 9343, and E 323, Bacteroides melaninogenicus subspecies intermedius strain B 10, Fusobacterium nucleatum strain F1 and Fev 1, Veillonella alcaIe~cemstrain Ve 5 and Veillonella parvula strain Ve 9. In a previous paper details about their isolation and cultivation have been described ( 11) LPS was isolated from whole or disintegrated ( 1 1) cells by the phenol-water extraction procedure (12) and further purified by ultracentrifugation. As reference endotoxin LPS from Salmonella enteritidis strain S-795 was used (generously provided by K . C. Milner, Rocky Mountain Laboratory, Hamilton, Mont., USA). The chemical and electronmicroscopic characterization of the LPS preparations have been reported ( 11)
Rabbit Polymorphonuclear Leukocytes Polymorphonuclear leukocytes ( PMNs) were obtained from the peritoneal cavity of adult New Zealand White rabbits 8-10 h after intraperitoneal (i.p.) injection of 200 ml of a solution containing 0.1 per cent glycogen (E. Merck AG, Darmstadt, W-Germany) in sterile isotonic saline (0.85 per cent NaCI), to which 50 +g garamycinQ3 (Schering Corporation, Bloomfield, N. J., USA) and 2.5 +g [email protected]
(Squibb, Flow Laboratories, Irvine, Scotland) per ml solution had been added. Before harvesting the cells, the rabbits were bled thoroughly. T h e i.p. cavity was opened under aseptic conditions, and the exudate collected by rinsing the cavity with sterile isotonic saline containing 10 U of heparin (4/S Apothekernes Laboratorium for Specialpraeparater, Oslo) per ml. The leukocyte-rich exudate was kept in an ice-bath until the isolation of the cells. After centrifugation of the exudate a t 225 x g for 5 min (IEC, Universal Model UV), the cells were suspended in heparinized (10 U/ml) chilled Gey's balanced salt solution (Gey's medium) containing 5 0 +g garamycina and 2.5 +g [email protected]
per ml. Erythrocytes present in the sediment were hemolyzed by adding a solution of 0.83 per cent ammoniumchloride. After centrifugation, the PMNs were resuspended in Gey's medium containing antibiotics and adjusted to a 2 per cent solution with bovine serum albumin 24.
Measurement of Chemotactic Activity As some of the LPS preparations were difficult to suspend in Gey's medium, the stock solutions containing 2 m u m 1 were ultrasonicated (MSE/ Mullard, 60 W, 2 0 kc/s) at 0" C for 2 min. If necessary, the p H was adjusted to 7.2 0.2 with triethylamine. The two compartments of a modified Boyden chamber used (10) (Neuroprobe, Bethesda, Md., USA) were separated by Millipore filter (Millipore Filter Corp.) of pore size 3 c. The lower compartment, having a volume of 0.2 ml, was filled with the chemotactic medium, viz. medium containing LPS and activated complement. Before placing the filter in the chamber, 0.4 ml of the cell suspension was spun down (58 x g for 1.5 min) on the filter in a cytocentrifuge (Shandon Elliot, Cytospin SCA-0030) giving approximately 4 x 10s PMNs on the upper surface. The upper compartment, thus containing the leukocytes, was filled with Gey's medium containing antibiotics and BSA. The chambers were then placed in a moist incubator for 3 h a t 37" C. Thereafter, the filters were removed, stained and cleared according to Boyden (1962) before mounting them bottom side up with diatex (Diatex AB, Wilh. Becker, Stockholm) on glass slides. PMNs which had completely crossed the intercompartmental filter during the incubation period were counted, using a 40 x objective and a n 8 x ocular containing a microgrid (Leitz Wetzlar, lox 10 mm). The average number of cells in 5 randomly selected high-power fields were counted, and the chemotactic activity expressed as the mean number of PMNs calculated from these fields. Fresh pooled guinea pig serum stored a t -25" C in 5 ml aliquots was used as the complement source. The serum was thawed at 4" C immediately before each experiment was performed. Aliquots of 1.5 ml from each stock solution of LPS containing 200 p d 0 . 1 ml were diluted two-fold. Dilutions thus containing 6.25 and 3.12 +g per 0.1 ml were tested for chemotactic activity. Five @per cent casein (E. Merck AG) was used as the control for positive chemotaxis. To examine if contaminating glucans of some of the LPS preparations used might be responsible for the chemotactic activity exerted by these preparations, the chemotactic activity of glycogen was tested. Whether a true chemotaxis was taking place or not was controlled by placing the same amount of LPS BUS-
pension (3.12 ~g/O.lml) in equal volume of undiluted serum in both compartments of the chamber. Aliquots of 1.5 ml of the LPS mixtures, containing equal volumes of LPS suspension and undiluted serum, were preincubated in water bath at 37°C for 30 min and thereafter at 56" C for 30 min. The dame was done with the controls, with the exception of casein. Controls were also included where the LPS-serum mixture was not preincubated or was preincubated at 56" C for 30 min. only. Statistical Methods Standard deviation was calculated as previously reported ( 11). The statistical evaluation of differences of chemotactic activities was performed according to the Wilcoxon rank test for two samples (two-tailed) (2). The dose X on the dose-response curve was transformed to X T according to the formula: X T = log (1 + X ) and the Pearson corFlation coefficient (7) was calculated according to the relationship between the transformed dose X T and the response Y.
TABLE 1. Leukocyte Chcmotaxis in vitro of LPS Isolated from V. alcalescens Strain V e 5 Amount of LPS-Ve 5 ( f i g )
EXPERIMENTS AND RESULTS
Incubation of LPS with guinea pig serum rendered the LPS-serum mixture chemotactically active on rabbit PMNs. When tested in ten-fold and two-fold dilutions, submicmgram amounts of all LPS preparations induced a positive chemotactic activity for rabbit PMNs. As illustrated in Table 1, there was a considerable variation in the number of cells per high-power field from one experiment to another. In these experiments, serum from the same pool was used, while the PMNs were fmm different batches. When the dose-response relationship was examined using two-fold dilutions of the different LPS preparations and plotted on a semi-logarithmic paper, each LPS exhibited its own specific curve. U p to a certain dose-level, the most active endotoxins showed a rather steep ascending dose-response relationship (Fig. 1) A maximum was usually found beyond which no further increase occurred but rather a diminution of the chemotactic response appeared. The less active endotoxins, viz. LPSE 323 and LPS-Fev 1, gave increasing response in leukocyte migration also at high concentrations. The positive correlation coefficients were: 0.99 for LPS-S-795 in the dose
range 0.39 to 12.5 pg, 0.98 for LPS-Ve5 in the dose range 0.39 to 12.5 pg, and 0.94 and 0.98 for LPS-E 323 and LPS-Fev 1, respectively in the dose range 0.39 to 200 pg. However, the less active LPS preparations were also ,highly chemotactic to the PMNs in doses lying within the ascending part of the doseresponse curve from the most active ones. T o avoid comparing the leukochemotactic activity of amounts of LPS lying on the decreasing part of the most active endotoxins and in order to spend minimal amounts of the LPS preparations, doses of 3.12 and 6.25 pg were chosen for a comparison with the chemotactic effect of LPS-S-795.
100 10 1
0.1 0.01 0.001 Serum alone Gey's sol. '+ lOcCgLPs Gey's sol. + serum**
Cells per high-power field EXP. 1 n = 2
EXP.2 n = 2
EXP-3 n = 2
573-685 399-523 30646 80- 88 65- 75 35- 63 13- 3.9
692-768 6 6 712 519413 421-459 332-358 104-140 17- 31
6 12480 528-622 279-307 134-190 91-121 38- 58 9- 19
0 - 2
Reaction mixture: Equal volumes of LPS sonicated in Gey's medium and guinea pig serum. Equal volume of each.
Using 3.12 pg of LPS, the Veillonella LPS and LPS-F 1 displayed no statistically significant difference in chemotactic activity from LPS-S-795 (Fig. 2). The difference between the LPS-S-795 and the other endotoxins or the controls was, however, highly significant ( p < 0.005). LPS in Gey's solution gave significantly lower cell counts per high-power field than serum alone. No difference was observed in the response when the amount of LPS was increased, nor when no LPS was added to Gey's solution. Glycogen incubated with guinea pig serum induced only slight
Cells x 100 7r
1 0.39 Y
Fig. I . Leukocyte chemotaxis of lipopolysaccharides isolated from the strains S-795, Ve 5, E 323 and Fev 1 incubated in guinea pig serum and tested in a modified Boyden chamber device. The abscissa represents the number of cells per high-power field on the chemotactic side of the filter disc. Each dose-response curve is drawn through the mean values of the number of cells per high-power field from two filter discs. -0- S-795, - 0 - Ve 5, -0- E 323, Fev 1.
Cells x 100
NCTC E323 9343 A
Fig. 2. Leukocyte chemotaxis of lipopolysaccharides isolated from strains of Vcilloncllu, Fusobucterium and Bucteroides in the dose of 3.12 fig preincubated in guinea pig serum. The tests were performed with leukocytes from the same pool. For explanation of bacterial LPS and doses used see Materials and Methods. Ve 5 + serum A) not preincubated, B) preincubated at 56" C for 30 min, C ) Ve 5 + Gey's solution, D) undiluted serum, E) LPS-F1 + serum in both compartments, F) glycogen + serum, G) casein. Each column represent the mean number of cells per high-power field f standard deviation (vertical bar) from six filter discs. LPS-S-795 was used for comparison. For statistical analysis the Wilcoxon rank test for two samples (two-tailed) was used. Values of p 5 0.05 = statistically significant. +p 0.05, "p 0.005.