Acta path. microbiol. immunol. scand. Sect. C, 92: 161-165, 1984.

INTERACTION OF HUMAN IgG PREPARATIONS WITH POLYMORPHONUCLEAR LEUKOCYTES IN VITRO

CHRISTIAN KOCH, NIELS HENRIK VALERIUS and VAGN ANDERSEN

Statens Seruminstitut. Department of Clinical Microbiology and Department of Pediatrics TG and Internal Medicine TA, Rigshospitalet, University of Copenhagen, Denmark

Koch, C., Valerius, N. H. & Andersen V. Interaction of human IgG preparations with polymorphonuclear leukocytes in vifro. Acta path. microbiol. immunol. scand. Sect. C , 92: 161-165, 1984.

The effect of purified human immunoglobulin G , prepared for intravenous administration by polyethylene glycol precipitation and ion exchange chromatography, on polymorphonuclear leukocytes (PMN) in v i m was studied. IgG induced a dose-dependent stimulation ofthe oxidative metabolism ofthe cells, as measured by release of superoxide anion and by chemiluminescence. The lowest concentration giving a detectable effect was 0.001 mg/ml. Two commercially available preparations of IgG for clinical intravenous use caused similar stimulation of the PMN although they differ in methods of isolation and purification. The addition of 10% maltose markedly reduced the stimulation by all three preparations of IgG. We suggest that I . v . infusion of purified IgG may lead to activation in vivo of host phagocytic cells without participation of complement, and that this interaction may be responsible for the hitherto unexplained side-effects caused by i .v. infusion of such preparations.

Key words: Intravenous immunoglobulin; neutrophilocyte activation.

C. Koch, Department of Pediatrics, Rigshospitalet, Blegdamsvej 9, DK-2 100 Copenhagen 0, Denmark.

Received 14.vii.83 Accepted 19.i.84 Patients suffering from antibody deficiency can

be treated with exogenous standard immunoglobu- lin, prepared from normal human plasma by alco- hol fractionation (Cohn fraction 11). If standard immunoglobulin is given intravenously it leads to severe systemic reactions that are believed to result from in vivo activation of complement by ag- gregates of IgG molecules. It is therefore given intramuscularly. However, the injections are pain- ful and the IgG is incompletely absorbed, some being locally degraded; the levels of IgG attained in plasma are therefore usually far below the normal range. Great efforts have been made to produce preparations of IgG for i.v. use that do not form aggregates and do not activate complement ( 10, 1 1 ). Several now available preparations are reported to be essentially free of aggregates and of complement activating properties, but they still cause systemic reactions in a minority of patients, especially those with low levels of immunoglobulin in plasma (l,8). The reactions are clinically characterized by chills

and fever, malaise, joint and muscle pain, and sometimes flushing, anxiety and chest tightness (9).

We speculated that these adverse reactions might be caused by direct interaction of the infused IgG with phagocytic cells, without participation of com- plement, leading to activation of the cells and release of mediator substance(s). Since activation of phagocytic cells among other events leads to stimu- lation of their oxidative metabolism, we decided to measure the changes in this function in polymor- phonuclear leukocytes (PMN) during exposure to preparations of IgG for i.v. use.

MATERIALS AND METHODS

IgCprepurutions. A test preparation of i.v. immunoglobu- lin (IgG (I)) was obtained from Nordisk Insulin, Copenha- gen. The IgG was prepared by precipitation with polye- thylene glycol (PEG) and ion-exchange chromatography. The content of IgG polymers in the final lyophilized and reconstituted preparation was determined by gel filtration on an AcA 34 column. The buffer contained 0.05 mol/l Na-phosphate and 0.4 mol/l NaCI, pH 7.0 (5). The

161

maximal content of polymers was 0.5%. The lyophilized preparation was dissolved in distilled water and further diluted in Gey's solution or Krebs Ringer phosphate buffer (KRPB), adjusted to pH 7.30. In vitro, there was no evidence that the IgG preparation activated complement, since no formation of C3c and C3d was found when the IgG preparation wasadded to normal plasma (Rasmussen, J . M.. Brandslund. I . & Svehag, S.-E., in prep.).

Two other preparations of i. v. immunoglobulin, kindly made available by the manufacturers, were also tested after dilution in KRPB: IgG (11) = Sandoglobulin (Sandoz, Basel, Switzerland) prepared by treatment at pH 4 with traces of pepsin, and IgG (111) = Intraglobin (Biotest, GmbH, Frankfurt a.M., FRG) prepared by beta-propio- lactone treatment (10). The content of IgG polymers in these preparations, as determined by AcA gel filtration, has been reported to be 0 and 2% respectively (10). All three preparations contain 5% IgG plus sugar in their final, ready-for-use form. IgG ( I ) contains 0.146 mol/l sucrose, IgG (11) 0.242 mol/l sucrose, and IgG (111) 0.158 mol/l glucose. Because of dilution, these sugars are not thought to be of importance in the in vifro assay.

Reagents. Cytochrome C (cyt C), superoxide dismutase (SOD), luminol and phorbol myristate acetate (PMA) were purchased from Sigma, St. Louis, Missouri. D-maltose and D-sucrose were from Merck, Darmstadt, FRG; D- glucose from Statens Seruminstitut, Copenhagen.

Po1.vmorphonuclear ieukocytes. PMN were obtained from normal healthy adults by dextran sedimentation of heparinized or ACD-treated venous blood. Following hypotonic lysis ofthe red cells, the leukocytes were washed twice and resuspended in Gey's or KRPB.

Release of superoxide anion. Superoxide production (0; liberation) was measured according to Babior et al. (2) with some modifications ( 1 2). Briefly, duplicate reaction mix- tures containing 2.4 mg cytochrome C, 5 x 10' PMN, and IgG at concentrations as indicated, in a final volume of2.4 ml of KRPB, were incubated for 30 min at 37 "C. After incubation the reaction mixtures were centrifuged in the cold at 2000 rev/min for 10 min, and the absorbance ofthe supernates at 550 nm was measured in a Gilford Stasar 111 spectrophotometer (Oberlin, Ohio) against a blank cont- aining cyt C in KRPB. The difference between the absorb- ance of these supernates and supernates from identical samples, in which SOD (200 IU) had been present during incubation, was used to calculate the superoxide-depen- dent reduction of cytochrome C. The absorbance differ- ence was converted to nanomoles cyt C reduced, by using E,,, = 2 I .2 nm (3). For control, reaction mixtures containing PMA (final concentration of1 pg/ml) as stimu- lant instead of IgG were included in each experiment. In some experiments sugars were added to give final concen- trations as indicated in results. The total volume of the reaction mixtures was always kept constant at 2.4 ml.

Chemiluminescence. Luminol-enhanced chemilumi- nescence by stimulated PMN was measured as described by Mills et al. (7) with some modifications. Reaction mixtures were made up of 4.5 ml of IgG in Gey's (to give final concentrations as indicated) and 0.05 ml ofa solution of lurninol in Gey's (0.001 'rng/ml) in glass scintillation vials. The vials were placed in the dark in a Beckman

cm

(Fullerton, California) LS 8000 beta scintillation counter without filter, adjusted out ofcoincidence. After recording background activity at ambient temperature, 1 .O ml of a suspension of PMN in Gey's (2 x IO'/ml) was added under red illumination and emission of visible light recorded at intervals for periods of 30 s and expressed as counts per min.

In a separate experiment, the luminol-enhanced chemi- luminescence by PMN stimulated with opsonized Staphy- lococcus aureus (strain 502 A) was studied. Staphylococci aureus were opsonized by incubation with 50% AB serum from a single donor for 60 min at 37 "C, followed by centrifugation and resuspension in Gey'sat approximately 2 x lo7 organisms/ml. Reaction mixtures were made up of 4.0 ml Gey's, 0.5 ml of the suspension of Staphylococcus aureus, 0.05 ml luminol solution, and 1 .O ml of a suspen- sion of PMN in Gey's ( 5 x 104/ml) and treated as above.

RESULTS

Stimulation of PMN Superoxide Production by IgG Fig. 1 shows that exposure of PMN to IgC ( I )

caused a dose-dependent stimulation of 0; libe- ration. The combined data of Fig. l and 2 indicate that stimulation was maximal at around 1 mg

30 -

20

I I I I I 1 10

0

10

OOOOI 0001 001 0 1

mg IgG p e r mi

Fig. 1. Superoxide production (0;-liberation) by PMN stimulated in vitro by various concentrations ofa prepara- tion of human IgG for 1.v. use (IgG (1) - see text). Results of five separate experiments in slightly different concentra- tion ranges are shown. 0;-liberation is expressed as nano- moles cytochrome C reduced per 5 x 10' PMN.

162

TABLE I . Inhibitory Effect ofsugars on theStimulation of PMNbvHuman lgG(1) andLackofEfftofPreincubation

of the Cells

Preincubation Stimulation with liberation of the PMN”’ IgG (1)I mg/ml in

presence of:

- Maltose 10% w/v Maltose 10% w/v

- Glucose 5% w/v Glucose 5% w/v

- Maltose 10% w/v”

Maltose 10% w/v

Glucose 5% w/v”

Glucose 5% w/v

-

-

-

I 8 .Oh’ 0

20.4 0

18.3d’ 1.3

26.6 3.2

Stimulation with PMA 1 Fg/rnl in presence of:

- - 66.5d’ - Maltose 10% w/v 63.6 - Glucose 59/0 w/v 64.4

” Preincubation was done at + 4 O, + 20 ’, and + 37 O C with identical results. The data shown are those at + 20 C . Nanomoles cytochrome C reduced per 2 x 10‘ PMN / 30 min.

-

” Final concentration. Equivalent to 0.292 mol/l. d’ Nanomoles cytochrome C reduced per 4 x lo6 PMN / 30

min. Phorbol myristate acetate.

IgG/ml, corresponding to one tenth ofphysiological plasma levels of IgG. Stimulation was, however, clearly detectable at much lower concentrations, i.e. 0.00 1 mg/ml. Assuming rapid intravascular distri- bution after infusion, this concentration corre sponds to infusion of approximately 3 mg IgG in an adult person; however, the local concentration during infusion may be much higher. 0; liberation at maximal stimulation with IgG (I) was approxi- mately 50% of that induced by PMA, the most potent stimulator of PMN oxidative metabolism (Table I). However, the 0; production following PMA stimulation occursat a much higher rate (data not shown).

Fig. 2 shows that the two commercially available preparations of IgG for i.v. infusion caused dose- dependent stimulation similar to IgG (I), although stimulation was somewhat lower. As preliminary control, a preparation of normal human serum albumin for clinical i. v. infusion was studied. Such preparations rarely cause side-effects. There was no stimulation of PMN 0; production by concentra- tions ranging from 0.1 to 10 mg albumin/ml (data not shown).

30 t IgGIl l .P

mg IgG per rnl

Fig. 2. Superoxide production by PMN stimulated with increasing concentrations of three preparations of IgG for i.v. use: IgG (I) from Nordisk Insulin (see text), IgG (11) =

Sandoglobulin ’, and IgG (111) = lntraglobin ’. Data are mean ofthree experiments for each preparation. The open symbols represent corresponding values in presence of maltose, 10% w/v, equivalent to 0.292 mol/l (one experi- ment each preparation). 0;-liberation is expressed as nanomoles cytochrome C reduced per 5 x lo6 PMN.

Stimulation of PMN Chemiluminescence by IgG The time course of PMN stimulation was studied

by following the intensity of light emission. Fig. 3 shows the luminol-enhanced chemiluminescence by PMN during exposure to increasing amounts of IgG (I). Again, IgG caused a dose-dependent stimu- lation of the oxidative metabolism, but chemilumi- nescence was maximal at 0.01 mg/ml. When the IgG concentration was increased above this level, a gradual decline was noted, and at 1 mg/ml barely any reaction was detectable. We believe that this finding may be due to an unspecific “quenching” effect of high concentrations of protein on the light emission from the system.

The time-lag before reaching maximal levels of light emission was largely independent of the con- centration of IgG, maximal activity being obtained after approximately 1 hour. It appears from Fig. 3 that the time course of light emission after stimu- lation with IgG .was markedly different from that following stimulation with an opsonized particle.

163

. Opsonired Staph. *’ ‘, I . ‘.

? ’, , I

,

,!! ‘.

minutes

Fig. 3. Chemiluminescence by PMNstimulated in vitro by various concentrations of IgC (1) (see text). Light emission is expressed as counts per minute x 10.’. The data are the results ofa single experiment, representative of 3 identical experiments, yielding similar results. The dotted line represents a typical experiment using opsonized Stuphylo- COCCIIS uurcws (strain 502 A) as stimulant instead of IgG.

Whereas opsonized Staphylococci caused a rapid (within minutes) increase in chemiluminescence; the initial rate of light emission, within the first 20 min, was slow with IgG, even at concentrations that gave maximal stimulation.

Inhibitory Effect of Sugurs on the Stimulation oJ PMN by IgG

Since maltose was reported to protect patients against systemic reactions following i. v. infusion of IgG (9), we studied the effect of maltose and other sugars on the in vitro stimulation of PMN by IgG. Fig. 2 shows that 10% maltose caused a marked reduction in the degree of stimulation of oxidative metabolism induced by IgG (I), as well as by the two other preparations. Ten per cent sucrose had a similar effect, and Table I shows that equimolar concentrations of glucose also protected the cells against stimulation by IgG (I) in this assay. It was considered that this protection could be due to the hyperosmolarity causing cellular damage. How- ever, pre-incubation with either maltose or glucose had no effect on subsequent response to stimulation

164

with IgG (I). Nor did the response to the synthetic stimulant PMA change by addition of sugar (Table 1 ).

DISCUSSION

In this study we demonstrate that a purified prepa- ration of IgG for intravenous administration causes stimulation of PMN in vitro leading to increased oxidative metabolism. We suggest that this stimu- lation may be related to the clinical side-effects of infusion of such preparations. Several ofthe present findings support this notion. First, IgG (I) caused a dose-dependent stimulation with doses as low as 0.001 mg/ml activating the PMN. The stimulation occurred more slowly than that observed with opsonized particles. Correspondingly, clinical symptoms usually occur with some delay, after infusion for approximately 1 hour, and they are not always avoided by lowering the rate of infusion. Secondly, similar stimulation curves were seen with two commercially available preparations although they differ as regards the methods employed for isolation and purification (10). Thirdly, the pres- ence of 10% maltose nearly completely protected the cells from stimulation induced by all three preparations of IgG. Correspondingly, it has been found that addition of 10% maltose to an IgG solution markedly reduced the clinical side-effects of i.v. infusion (9).

We therefore suggest that i.v. infusion of IgG preparations may lead to in vivo activation of PMN, and possibly other phagocytic cells as well. The clinical symptoms are likely to by caused be release of endogenous pyrogen or other products of acti- vated phagocytes.

It seems likely that interaction of IgG with PMN is mediated via the Fc-part of the molecule. Anti- gen-complexed, or heat-aggregated IgG binds to and stimulates PMN. However, the content of IgG polymers in IgG (1) was very low, equal to what has been reported for the two commercially available preparations (see Materials). Denaturation of IgG induces conformational changes in the molecule that are similar to those induced by combination with antigen, leading to increased affinity for Fc receptors in the PMN cell membrane (4, 6). It is therefore tempting to ascribe the stimulation by IgG preparations to the presence of contaminating denatured IgG molecules.

It is interesting that the stimulation was inhibited by maltose. However, other sugars were equally effective and we do not know whether this inhibi- tion was specific in terms of competition for carbo-

hydrate-binding moieties on either the IgG mole- cule or the PMN cell membrane.

More studies are obviously needed before any conclusions can be drawn as to the exact nature of the interaction ofthe IgG preparations with PMN in v i m and the protective effect of sugars in vitro as well as in vivo (9). While such studies are under way, it does appear of potential value that the present findings may offer a novel approach to the in vitro testing of IgG preparations for safety, prior to their application for clinical use.

This study was supported by the Danish Medical Research Council, and the Danish Hospital Foundation for Medical Research, Regions of Copenhagen, the Faroe Islands, and Greenland. We are grateful to Mrs. Hanne Tarnsdorfand Mrs. Susanne Overbye for their skillful technical assis- tance. Dr. Mirella Ezban kindly performed the gel filtra- tion studies on IgG (I).

I .

2.

3.

REFERENCES

Alving, B. M. & Finlayson, J. S:(Eds.): Immunoglobu- lins: characteristics and uses of intravenous prepara- tions. Washington D. C.: U.S. Government Printing Office, DHEW publications No. (FDA)-80-9005, 1980. Babior. B. M.. Kipnes, R. S. & Curnutte. J . T'.: Biological defence mechanisms. The production by leukocytes of superoxide, a potent bactericidal agent. J. Clin. Invest. 52: 741-744, 1973. van Gelder, B. F. & Slater, E. C.: The extinction coefficient of cytochrome C. Biochem. Biophys. Acta 58: 593-598, 1962.

4.

5 .

6.

7.

8.

9.

10.

1 1 .

12.

Henson, P. M., Johnsen. H. B. & Spiegelberg, H. L.: The release of granule enzymes from human neutro- phils stimulated by aggregated immunoglobulins of different classes and subclasses. J. Immunol. 109:

Kranz, T., Guthohrlein, G., Schmidt, K.-H. & Hofstaetter, T.: Gel chromatography applied to quantitation of components of IgG preparations. Develop. Biol. Standard. 44: 19-30, 1979. Metzger, H.: Effect of antigen binding on the proper- ties of antibody. 'Advances Immunol. 18: 169-207, 1974. Mills, E. L., Rholl, K . S. & Quie, P. G.: X-linked inheritance in females with chronic granulomatous disease. J . Clin. Invest. 66: 332-340, 1980. Nydegger, U. E. (Ed.): Immunochemotherapy. A guide to immunoglobulin prophylaxis and therapy. Academic Press, London, 1981. Ochs, H. D., Buckley, R. H., Pirofski, B., Fischer, S. H.. Rousell, R. H., Andersen. C. J. & Wedgwood, R. J.: Safety and acceptability of intravenous immune glo- bulin in 10% maltose. Lancet ii: 1158-1 159, 1980. Romer, J.. Morgenthaler, J.-J., Schertz, R. & Skvaril. F.: Characterization of various immunoglobulin preparations for intravenous application I. Protein composition and antibody content. Vox Sang. 42:

Romer, J., Spath, P. J., Skvaril, F. & Nydegger. U. E.: Characterization of various immunoglobulin prepa- rations for intravenous application II. Complement activation and binding to Staphylococcus protein A. Vox Sang. 42: 74-80, 1982. Valeriits. N . H.:Chemotaxis, spreading. and oxidative metabolism of neutrophils: influence of albumin in vitro. Acta path. microbiol. immunol. scand. Sect. C, 91: 43-49. 1983.

1182-1 192, 1972.

62-73, 1982.

165