INFECTION AND IMMUNITY, Apr. 1981, p. 295-299 0019-9567/81/040295-05$02.00/0

Vol. 32, No. 1

Heparin Inhibits Phagocytosis by Polymorphonuclear Leukocytes MICHAEL VICTOR, JERROLD WEISS, AND PETER ELSBACH* Department ofMedicine, New York University School ofMedicine, New York, New York 10016

Phagocytosis of unopsonized Salmonella typhimurium 395 MR-10, opsonized Salmonella typhimurium 395 MS, and Staphylococcus epidermnidis by rabbit polymorphonuclear leukocytes was inhibited by heparin at concentrations as low as 0.5 U/ml. Inhibition was dose dependent and nearly complete at 20 U/ml. Provided that heparin concentrations did not exceed 100 U/ml, inhibition could be largely reversed by washing. Heparin also reversibly inhibited the adherence of polymorphonuclear leukocytes to glass. In contrast, hexose monophosphate shunt activity of polymorphonuclear leukocytes stimulated by noningested S. typhimurium MR-10 or Streptococcus pyogenes B14 was not inhibited by heparin at concentrations as high as 100 U/ml. The incidental observation that the presence of heparin in the flask used for collection of polymorphonuclear leukocytes (PMN) from peritoneal exudates elicited in rabbits was associated with reduced bactericidal activity prompted us to examine the effect of heparin on PMN function more systematically. We found that treatment of PMN with heparin inhibited, in a dose-dependent fashion, phagocytosis and adherence to glass surfaces. MATERIALS AND METHODS PMN. PMN were obtained from sterile peritoneal exudates elicited in rabbits by injection of 0.1% glycogen in physiological saline. Cells were collected from 12 to 16 h later. Differential counts showed that more than 95% of the cells were PMN. The cells were sedimented at 50 x g for 10 min and suspended at a concentration of 5 x 107/ml in 90% Hanks balanced salt solution (HBSS, without phenol red; Microbiological Associates, Bethesda, Md.)-10% 0.4 M tris(hydroxymethyl)aminomethane (Tris)-hydrochloride (pH 7.5). Growth and labeling of bacteria. Two strains of Salmonella typhimurium 395, rough MR-10 and smooth MS (kindly donated by Olle Stendahl of the Department of Medical Microbiology, University of Linkoping, Sweden), were grown in minimal medium buffered with triethanolamine at pH 7.75 to 7.90. Staphylococcus epidermidis and Streptococcus pyogenes B14 were grown in brain heart infusion broth (Difco Laboratories, Detroit, Mich.). All bacteria, after overnight culture, were transferred, diluted 1:10 with fresh medium, and subcultured for 2 to 3 h at 37°C. At this time bacterial concentration was estimated by measuring the optical density of the cultures at 550 nm in a Junior Spectrophotometer (Coleman Instruments, Maywood, Ill.). The bacteria were sedimented by centrifugation at 10,000 x g for 10 min and suspended in sterile physiological saline to a concentra295

tion of 2 x 109/ml. Bacterial ribonucleic acid was labeled during growth in subcultures supplemented with 0.2 uCi of ['4C]uracil (40 to 60 mCi/mmol; New England Nuclear Corp., Boston, Mass.) per ml. After incubation for 3 h, the bacteria were washed once, suspended in fresh growth medium, and reincubated for 30 min to permit incorporation of the remaining unincorporated radioactive precursor. After sedimentation and suspension in isotonic saline, suspensions of 107 bacteria contained 4,000 to 8,000 cpm, >98% of which were precipitable in cold 5% trichloroacetic acid. Heparin treatment. Heparin sodium (Liquaemin Sodium; 5,000 U/ml; Organon Inc., W. Orange, N.J.) at the desired concentration was added to PMN suspensions (4 x 107 cells per ml) either at 0 time or during a preincubation period of up to 45 min. The amounts of the preservative benzyl alcohol in Liquaemin Sodium (1%), present in 100 U of heparin per ml, have no effect on phagocytosis by PMN. S. typhimurium MS must be opsonized to be ingested by PMN. Opsonization was carried out during incubation at 37°C for 20 min in a batch consisting of 2.5% anti-MS serum, 10% HBSS, 10% 0.4 M Trishydrochloride (pH 7.5) buffer, 0.1% Casamino Acids (Difco Laboratories), and 109 S. typhirnurium MS organisms brought up to a final volume of 1 ml with sterile saline. The opsonized bacteria were then sedimented by centrifugation at 10,000 x g for 10 min and resuspended to a concentration of 2 x 109/ml in sterile isotonic saline. Assay for viability. The effect of granulocytes on the ability of various bacterial strains to multiply was assayed at a ratio of 20 microorganisms to 1 PMN for S. typhimurium MR-10 and a ratio of 10:1 for S. typhimurium MS, S. epidermidis, and S. pyogenes B14 in an incubation mixture containing 107 PMN, 225 pI of HBSS, 25 mM Tris-hydrochloride (pH 7.5), and physiological saline brought to a final volume of 0.4 ml. After 30 min of incubation at 37°C, 10-pl samples were serially diluted in sterile isotonic saline and plated on nutrient agar in the case of S. typhimurium

296

VICTOR, WEISS, AND ELSBACH

MS and MR-10 and on brain heart infusion broth in the case of S. epidermidis and S. pyogenes B14. Colonies were counted after growth overnight at 37°C (3, 9). [1-'4CJglucose metabolism. Hexose monophosphate shunt (HMS) activity of the PMN was measured by the conversion of [1-_4C]glucose to 14C02 as previously described (10). Briefly, suspensions of PMN (2 x 107 cells) in 90% HBSS-10% 0.4 M Tris-hydrochloride (pH 7.5) were placed in 10-ml Erlenmeyer flasks which contained 0.03 ,iCi of [1-'4C]glucose. Microorganisms and saline (4 x 108) were added to the flasks to provide a total volume of 0.4 ml at the beginning of incubation at 37°C. Evolved 14CO2 was collected in polyethylene cups suspended from rubber stoppers. After 30 min, the reaction was stopped by injecting 0.2 ml of 10 N H2S04 through the rubber stopper into the flask. After 15 min, hydroxide of hyamine (0.4 ml) was added to the collection cups, and incubation at 37°C was continued for 1 h. At this time the cups were removed, placed in counting vials, and vigorously shaken with 8 ml of toluene-(2,5-tert-butylbenzoxazolyl)-thiophene; Packard Instrument Co., Downers Grove, Ill.) scintillation mixture. Counting was performed in an LS 100C liquid scintillation counter (Beckman Instruments Co., Inc., Irvine, Calif.). Measurement of bacterial uptake by PMN. Incubation suspensions contained 107 PMN and 108 [14C]uracil-labeled S. typhimurium 395 MR-10 in 225 Al of.HBSS-25 mM Tris-hydrochloride (pH 7.5)-physiological saline brought to a final volume of 0.4 ml. After 30 min of incubation at 37°C, the PMN were sedimented by centrifugation in an International clinical centrifuge (speed setting no. 3; Intemnational Equipment Co., Div. of Damon Corp., Needham Heights, Mass.) for 90 s at room temperature. Less than 5% of the PMN were recovered in the supernatant as determined by cell counting with a hemocytometer. Recovery of ['4C]uracil-labeled bacteria in the supernatant was determined by measuring portions of the supernatant in a liquid scintillation counter. In the absence of PMN, >90% of the radiolabeled bacteria was recovered in the supernatant. No trichloroacetic acid-soluble radioactivity accumulated in the supernatant. Adherence assay. Rabbit PMN (5 x 105/ml) were incubated with heparin in a 20-ml scintillation vial (disposable, glass; Kimble, Owens, Ill.). Incubation was carried out with 90% HBSS-10% Tris-hydrochloride (pH 7.5) with various concentrations of heparin sodium for 30 min at 37°C in a shaking water bath. The cells remaining suspended in the medium were counted after incubation to determine the number of granulocytes which had become attached.

RESULTS Effect of heparin on phagocytosis and killing of bacteria by PMN. Heparin inhibited killing by rabbit PMN of both rough nonopsonized MR-10 and opsonized smooth MS strains of gram-negative S. typhimurium and of a strain of gram-positive S. epidermidis (Table 1). During incubation for 30 min with PMN in the absence of heparin the viability of each of the three

INFECT. IMMUN. TABLE 1. Dose-dependent inhibition by heparin of bactericidal activitiy of PMNa Heparin concn

S. typhimu-

rium MR-10

S. typhimurium MS

S. epidermidis

I (%)b I (%)b (U/ml) I (%)b 27 ± 7 (4) 31 ± 13 (6) 37 ± 31 (3) 0.5 2 70 ± 9 (4) 40 ± 30 (3) 69 ± 17 (3) 84 ± 12 (5) 61 ± 20 (6) 59 ± 13 (4) 5 20 93 ± 3 (6) 88± 5(6) 75 ± 11 (4) 97 ± 3 (6) 100 100 (2) 85 ± 8 (4) 500 99 (2) a Incubations of PMN and bacteria with heparin in the indicated concentrations were carried out as described in the text. b Bacterial viability was measured as colony-forming ability. In the absence of heparin, PMN produced the following loss of viability: S. typhimurium MR-10, 86 ± 5% (n = 7); S. typhimurium MS, 68 ± 6% (n = 5); and S. epidermidis, 70 ± 5% (n = 4). Heparin-induced inhibition (I) of bacterial killing by PMN was calculated by the following equation: I = 1 - [killing (+heparin)/killing (-heparin)] x 100. Values shown represent the mean ± standard error of the mean of results of the indicated number of experiments (in parenthe-

ses).

microorganisms was reduced by at least 80%. Heparin sodium at concentrations as low as 0.5 U/ml inhibited bacterial killing by PMN. Protection of the bacteria increased as the heparin concentrations were increased and was nearly complete at 20 U/ml. Experiments with ['4C]uracil-labeled S. typhimurium (Table 2) showed a dose-dependent reduction by heparin of phagocytosis (uptake) of the labeled bacteria that closely matched the effect on viability. For determining whether the heparin-induced protection included inhibition of intracellular killing as well as ingestion, the cell suspensions were sonicated after incubation for 30 min. Sonication does not affect bacterial viability, but does break up PMN, releasing intracellular bacteria. Thus, a comparison of bacterial viability in sonicated and unsonicated samples reveals intracellular bacterial survival. Table 2 shows that the prevention by heparin of bacterial killing was no greater in sonicated than in unsonicated samples. Hence, the heparin effect on killing of bacteria by PMN appears to be mainly attributable to the inhibition of ingestion. Time course and reversibility of heparin effect. The heparin effect on the ingestion of bacteria by PMN was immediate, i.e., the inhibition of killing was the same when heparin was added to the PMN suspension at 0 time or at various times before addition of the bacteria. The inhibitory effect of heparin on the phagocytosis of bacteria by PMN was reversed by washing. Table 3 shows that PMN pretreated with heparin regained their capacity to ingest

HEPARIN INHIBITION OF PHAGOCYTOSIS BY PMN

VOL. 32, 1981

TABLE 2. Effect of heparin on phagocytosis and intracellular killing by PMN of S. typhimurium

MR_1Oa % Inhibition' of: Heparin Bacterial killing concn Uptake of ['4C](U/ uracil-labeled No sonication bacteria ml) Sonication' 15 (2) 14.5 (2) 0.5 12.5 ± 8.3 (3) 88.3 + 6.1 (3) 84.5 ± 15.5 (4) 81 ± 19 (4) 5 85 ± 13.8 (3) 95 ± 5.0 (3) 20 85 (2)

aThe experiments and sonication of samples after incubation of cell suspensions were carried out as described in the text. b Inhibition was calculated as described in Table 1, footnote b. Radiolabeled bacteria were prepared as described in the text. Extracellular bacteria were separated from PMN by centrifugation (see text). In the absence of heparin, 84% of the radiolabeled bacteria were associated with the PMN pellet. The values shown represent the mean (for two experiments) and mean ± standard error of the mean (for more than two experiments) (no. of experiments is shown in

parentheses). c Carried out twice at 40 to 60 W for 15 s at 0 to 4°C (Sonifier Cell Disrupter, model W185; Heat SystemsUltrasonics, Inc., Plainview, N.Y.).

the HMS activity of the PMN (conversion of [14C]glucose to '4CO2) in the presence of bacteria. In the absence of heparin, the addition of S. typhimurium MR-10 resulted in a manyfold stimulation of '4CO2 production. Heparin inhibited this stimulation by about 50%, independent of dose over a range of 2 to 100 U/ml (Table 4). Because at this concentration range heparin nearly abolished phagocytosis, these findings suggest that stimulation of the HMS when S. typhimurium MR-10 was added to the PMN suspension was only partially linked to ingestion. This conclusion is supported by the observation that the addition of S. pyogenes B14, microorganisms that were not ingested and killed by PMN under the conditions used, also resulted in markedly increased 14CO2 production. Heparin in concentrations as high as 100 U/ml had no appreciable effect on this response of the PMN. Effect of heparn on adherence of rabbit PMN to a glass surface. Heparin reduced the adherence of PMN to glass in a dose-dependent fashion (Table 5). Dilution of the heparin conTABLE 4. Effect of heparin on HMS activity of PMN

TABLE 3. Effect of washing on inhibition ofkilling of S. tphimurium MR-10 by PMNHeparin concn (U! ml)

% Inhibition' with: No washes Two washes

78 91 92

3 10 32 91 81 aAfter preincubation with heparin, PMN were washed by sedimentation and resuspension in heparinfree medium as described in the text. b Inhibition was calculated as described in Table 1, footnote b. Data shown represent the mean of the results of two closely similar experiments. 5 20 100 800

bacteria after centrifugation and suspension twice in heparin-free medium. Reversal of the heparin effect was only partial when higher concentrations of the agent were present during the preincubation, presumably because heparin was incompletely removed during two washes. S. typhimurium MR-10 pretreated with 5 U of heparin per ml (a concentration that inhibited the killing by PMN of MR-10 by 80%) and then diluted 16-fold to a final heparin concentration of 0.3 U/ml exhibited normal sensitivity to killing by the PMN. Thus, at 5 U/ml, heparin did not irreversibly protect the bacteria. Effect of heparin on HMS activity of PMN. To determine whether heparin also affected metabolic responses that typically accompany phagocytosis, we examined the effect of the agent on the respiratory burst by measuring

297

HMS activity Heparin concn (U/ml)

S. typhimurium 4.6a

S.B14 pyogenes 16.0a

Effect of heparin (% of control)'

0 (Control) 2 49 5 94 20 92 40 43 100 48 95 a The increase in HMS activity in the presence of bacteria over resting values is shown as n-fold stimulation (mean of at least two experiments). b The effect of heparin on stimulation is expressed as the percentage of the values obtained in the absence of heparin in each experiment. The values shown represent the mean of the results of three closely similar experiments.

TABLE 5. Effect of heparin on the adherence of PMN to glass surfaces' Heparin concn (U/ml) 0.

% PMN adherent

757(4)

2.60±4(3) 10 .42 100 .22 ± 3 (4)

100-4 1 ............................ 84 aAdherence of PMN in the presence of increasing concentrations of heparin was measured as described in the text. Values shown represent the mean ± standard error of the mean of the results of the indicated number of experiments (in parentheses).

298

VICTOR, WEISS, AND ELSBACH

centration from 100 to 1 U/ml restored adherence to control levels.

DISCUSSION Heparin is commonly used as an anticoagulant when blood is collected for the procurement of leukocytes. Hence, the demonstration that heparin at concentrations as low as 0.5 U/ml inhibits phagocytosis and adherence must be given serious consideration in the design and interpretation of experiments employing peripheral blood leukocytes. The inhibitory effects of heparin on PMN described here can be reversed by dilution if the initial drug concentrations are not too high. Resuspension(s) ofthe cells in heparin-free medium should therefore reduce or overcome the consequences of heparin use in most instances. Heparin partially inhibited the stimulation of the conversion of [1-'4C]glucose to 1'CO2 in the presence of readily ingested S. typhimurium MR-10 but had no inhibitory effect on the increased oxidation of [1-14C]glucose in the presence of noningested S. pyogenes B14. These findings may be added to others indicating that the respiratory burst of the PMN can be triggered by microbial populations in the absence of phagocytosis (2, 8). It further appears that heparin only inhibited that portion of 14CO2 production that is linked to ingestion but had no effect on the respiratory burst per se. Recent evidence suggests that enhanced superoxide production, another concomitant of the respiratory burst, is a plasma membrane-associated event triggered by cell surface perturbations (1, 4, 7). If this view is correct, it follows that polyanionic heparin does not affect those surface properties of the PMN that are involved il1 respiratory activation. It also follows from these observations that the activation of the respiratory burst when phagocytosis is inhibited does not generate extracellular conditions that impair the viability of uningested bacteria (whether gram positive or gram negative). Thus, any contribution of the 02-dependent bactericidal system of the leukocytes toward killing of these bacterial species appears to require their intracellular sequestration. These experiments did not reveal how heparin inhibits phagocytosis and adherence. The agent may alter the surfaces of all three components of the system that we have examined, namely, the PMN, the microorganisms, and the vessel wall, thereby modifying the surface-surface interactions that underlie both phagocytosis and adherence. We found, however, that washing of PMN suspensions (pretreated with high heparin concentrations), which should have lowered the

INFECT. IMMUN.

heparin concentration in the medium to noninhibitory levels, did not restore normal phagocytosis (Table 3). This observation suggests that heparin has an effect on the PMN themselves, perhaps through a weak surface interaction. Our observations show that the inhibition of adherence requires much higher concentrations of heparin than does the inhibition of phagocytosis and also that, in contrast to phagocytosis, normal adherence is readily restored by a single dilution. This may reflect a different affinity of heparin for sites involved in adherence and sites engaged in the interaction between PMN and bacteria. Alternatively, differences in the surface-surface interactions measured in the two assays may account for the apparently greater sensitivity of phagocytosis to heparin-mediated inhibition. HAkansson et al. (5) have recently reported the stimulatory effect of very low concentrations (5 to 500 ng/ml) of hyaluronic acid on multiple responses of human PMN, including adherence, locomotion, chemiluminescence, phagocytosis of latex particles, and intracellular adenosine 5'triphosphate levels. High concentrations of hyaluronic acid (three to four orders of magnitude higher than the stimulatory concentrations) inhibit these functions. In light of these observations we examined the effects of very low concentrations of heparin on ingestion and killing of bacteria and on hexose monophosphate shunt activity; we found no effects (data not shown). It should be stressed that heparin and hyaluronic acid, although both polyanions, are not equivalent in their effects. For example, Hirsch has shown that the antibacterial effect of histone-like proteins is much more effectively inhibited by heparin than by hyaluronic acid (6). Whatever the mechanism, we feel that the inhibitory effects of heparin on PMN need to be brought to the attention of students of leukocyte function. Finally, heparin may be another useful agent for use in experiments that require reversible inhibition of phagocytosis. LITERATURE CITED 1. Babior, B. M. 1978. Oxygen-dependent microbial killing by phagocytes. N. Engl. J. Med. 298:659-668. 2. Densen, P., and G. L Mandell. 1978. Gonococcal interactions with polymorphonuclear leukocytes. J. Clin. Invest. 62:1161-1171. 3. Elsbach, P., P. Pettis, S. Beckerdite, and R. Franson. 1973. Effects of phagocytosis by rabbit granulocytes on macromolecular synthesis and degradation in different species of bacteria. J. Bacteriol. 115:490-497. 4. Goldstein, L. M., M. Cerqueira, S. Lind, and H. B. Kaplan. 1977. Evidence that the superoxide-generating system of human leukocytes is associated with the cell surface. J. Clin. Invest. 59:249-254. 5. H&kansson, L., R. Hillgren, and P. Venge. 1980. Reg-

VOL. 32, 1981

HEPARIN INHIBITION OF PHAGOCYTOSIS BY PMN

ulation of granulocyte function by hyaluronic acid. In vitro and in vivo effects on phagocytosis, locomotion and metabolism. J. Clin. Invest. 66:298-305. 6. Hirsch, J. G. 1958. Bactericidal action of histone. J. Exp. Med. 108:925-944. 7. Salin, M. L., and J. M. McCord. 1974. Superoxide dismutases in polymorphonuclear leukocytes. J. Clin. Invest. 54:1005-1009. 8. Simberkoff, M. S., and P. Elsbach. 1971. The interaction in vitro between polymorphonuclear leukocytes

299

and mycoplasma. J. Exp. Med. 134:1417-1430. 9. Weiss, J., S. Beckerdite-Quagliata, and P. Elsbach. 1980. Resistance of Gram-negative bacteria to purified bactericidal leukocyte proteins. Relation to binding and bacterial lipopolysaccharide structure. J. Clin. Invest. 65:619-628.

10. Wurster, N., P. Elsbach, E. J. Simon, P. Pettis, and S. Lebow. 1971. The effects of the morphine analogue levorphanol on leukocytes. Metabolic effects at rest and during phagocytosis. J. Clin. Invest. 50:1091-1099.