J. L. Teeling, Pathophysiology of plasma proteins, Central Laboratory of The Netherlands' Red Cross Blood Transfusion Service, Plesmanlaan 125, 1066 CX Amsterdam, The Netherlands. E-mail: J_Teeling@clb.nl
Despite widespread use in various immune disorders, the in vivo mechanisms of action of intravenous immunoglobulin (IVIG) preparations are not well known. We previously reported that human neutrophils degranulate after incubation with IVIG in vitro as a result of interaction with FcγRII. The purpose of this study was to determine whether IVIG might stimulate neutrophils in vivo. Anaesthetized rats received a bolus intravenous injection of IVIG preparations, containing either high (aged IVIG) or low (fresh IVIG) amounts or IgG dimers at a dose of 250 mg/kg. Administration of aged IVIG induced neutrophil activation in vivo, whereas no effect was observed after infusion of fresh IVIG. Histological examination of lung tissue demonstrated mild influx of neutrophils into the pulmonary tissue after aged IVIG administration, though gross damage did not occur. Macrophage-depleted rats no longer showed activation of neutrophils after infusion of aged IVIG, suggesting that neutrophils become activated via an indirect macrophage dependent way. We conclude that IVIG induces a mild activation of neutrophils in vivo via triggering of macrophages depending on the amount of IgG dimers. For this reason, IVIG preparations with a high content of dimers may not always be as harmless as generally believed and may be responsible for some of the side-effects observed during IVIG infusions.
Although intravenous immunoglobulin (IVIG) preparations were initially introduced as replacement therapy in primary antibody deficiency disorders, they were found to have beneficial effects in patients with autoimmune thrombocytopenic purpura and other autoimmune diseases (Dwyer, 1992; Wolf, 1996). Nowadays, IVIG is used in a broad range of autoimmune and systemic inflammatory disorders. Despite its widespread use, the precise mechanisms of action of IVIG are still largely unknown. Different mechanisms have been proposed, such as Fcγ-receptor blockade, inhibition of complement deposition, neutralization of superantigens, neutralization of cytokines and manipulation of the idiotypic network (Imbach et al, 1981; Sultan et al, 1984; Aukrust et al, 1994; 1997; Mouton et al, 1994; Andersson et al, 1996; Basta, 1996; Skansén-Saphir et al, 1997).
We previously reported that IVIG interacts with neutrophils in vitro through binding of IgG dimers and polymers to their Fcγ-receptors (Teeling et al, 1998). In particular, binding of IVIG to FcγRIIa was shown to induce activation and, subsequently, degranulation of neutrophils.
Neutrophils are endowed with potent mechanisms to kill microbes and thus contribute to the host's defence against microbicidal infections. However, they have also long been recognized as important effectors of the adverse effects of inflammatory reactions because their excessive accumulation in tissues may contribute to injury of the host as a result of release of a variety of toxic products. The lungs are particularly susceptible to such insults and constitute a primary site of damage induced by host-defence mechanisms (Chignard & Renesto, 1995; Kubo et al, 1998). Hence, interaction of IVIG with neutrophils in vivo may, among others, contribute to the development of clinical (side) effects of IVIG.
Because IVIG is able to directly activate neutrophils in vitro, an animal model was developed to study if this also occurs in vivo. IVIG preparations, differing in their IgG dimer content, were administered and activation of neutrophils was determined in the circulation as well as at tissue level. As we considered a bolus infusion of IVIG, rather than continuous infusion, to be more robust and hence more suitable to detect potential effects, IVIG was administered as a bolus infusion in most experiments. However, to mimic the clinical situation, we also evaluated the effect of continuous infusion. Independent of the mode of infusion, our results showed that IVIG activates neutrophils in vivo, albeit not directly. Instead, macrophages are involved, because macrophage depletion completely abolished all parameters for neutrophil activation.
Materials and methods
Immunoglobulin preparations Human γ-globulin for intravenous use was obtained from our institute [Immunoglobulin i.v., 6% (w/v), licensed for human use]. This is a freeze-dried product prepared from pooled plasma from at least 1000 donors by Cohn fractionation followed by pH 4/pepsin incubation. Fresh IVIG was immediately frozen after reconstitution and kept at −80°C until further use. An aged IVIG preparation was prepared by storing this reconstituted preparation at 4°C for 5–10 months, resulting in an increased IgG dimer content. This aged IVIG-preparation is an experimental product, not licensed for human use. IVIG preparations used in the experiments were analysed for actual monomer, dimer and polymer content by size exclusion chromatography using a calibrated Superose12 gel-filtration column connected to a FPLC system (Pharmacia, Uppsala, Sweden). A computer program (Ezchrom Chromatography Data System version 6·5) was applied to determine the peak areas of the chromatograms.
Flow cytometric analysis To determine whether human IVIG activates rat neutrophils in vitro, whole rat blood was diluted in endotoxin-free Iscove's modified Dulbecco's medium (IMDM; Biowithaker, Verviers, Belgium) supplemented with 0·1% (v/v) endotoxin-free fetal calf serum (FCS; Biowithaker), 100 U/ml penicillin and 100 µg/ml streptomycin and incubated with various concentrations of IVIG (0–30 mg/ml). After 2 h at 37°C, rat cells were washed and incubated in incubation medium (10 mmol/l HEPES, 150 mmol/l NaCl, 5 mmol/l KCl, 1·5 mmol/l CaCl2, 2 mmol/l MgCl2, pH 7·4, supplemented with 0·02% (w/v) NaN3 and 0·5% bovine serum albumin (BSA) with fluorescein isothiocyanate (FITC)-conjugated mouse anti-rat CD11b mAb (ED8, IgG1, Instruchemie B.V., Hilversum, The Netherlands) or mouse anti-rat L-selectin mAb (OX85, IgG1, Instruchemie B.V.) followed by phycoerythrin (PE)-conjugated F(ab)2 goat anti-mouse Ab (DAKO). Subsequently, erythrocytes were lysed using FACS lysing solution according to the manufacturer's instructions (Becton Dickinson, San Jose, CA). Human venous blood was collected from healthy volunteers by venepuncture in vacutainer tubes containing heparin in a final concentration of 15 U/ml (Becton Dickinson). Whole blood was incubated with various concentrations of IVIG as described above. Thereafter, blood cells were incubated with FITC-conjugated mouse-anti-human CD11b (CLB-mon-gran/1, B2, IgM, CLB, Amsterdam, the Netherlands) or mouse-anti-human L-selectin mAb (Leu-8, IgG2a, Becton Dickinson) followed by PE-conjugated F(ab)2 goat anti-mouse Ab. Subsequently, erythrocytes were lysed using FACS lysing solution. Leucocytes were resuspended in incubation medium and kept in the dark at 4°C until analysis with a FACScan flow cytometer (Becton Dickinson). Neutrophils were identified by their characteristic forward and sideward scatter pattern, which was confirmed by staining a rat neutrophil marker (HIS48, IgM, Instruchemie B.V.).
To assess the contribution of FcγRIIa to the activation of human neutrophils, cells were incubated with IVIG in the presence of F(ab)2 fragments of AT10, a mAb that inhibits the binding of IgG to FcγRIIa (Greenman et al, 1991; Tuijnman et al, 1992) mAb AT10, kindly provided by Prof. Dr J. V. D. Winkel.
Human IVIG binding to rat neutrophils was studied using biotinylated IVIG (Teeling et al, 1998). Briefly, cells from whole blood were washed to remove plasma proteins, and incubated with biotinylated IVIG for 30 min at 4°C. The cells were then washed and incubated with FITC-labelled streptavidin (Becton Dickinson) for 30 min at 4°C.
Rat model Female Wistar HsbCpb:WU rats (Harlan/CPB, Zeist, The Netherlands) weighing 250–300 g were anaesthetized by intraperitoneal injection of Hypnorm/Nembutal. The rat was fixed (back downward) on a heated support (38°C) to prevent decrease in body temperature. Cannulas (Silastic®) were introduced into the right jugular vein and left carotid artery. Cannulas were flushed at 2 ml/hour in to ensure patency. The arterial cannula was connected to a pressure transducer for continuous recording of the mean arterial blood pressure. Test solutions were administered via the venous cannula. The IVIG preparations were administered intravenously within 10 s at a dose of 250 mg/kg. To mimic the clinical method of administration, in some experiments IVIG was also given as a continuous infusion at a dose of 1 g/kg/hour or 2·5 g/kg/hour. Saline was used as a negative control. Lipopolysaccharide (LPS, Sigma Chemical, St. Louis, MO, USA) given at a dose of 0·5 mg/kg as a 5-min infusion was used as a positive control for neutrophil activation. Experiments were terminated 4 h after infusion of IVIG or LPS by administration of an overdose pentobarbital. Blood samples collected in EDTA (10 mmol/l, final concentration) were used to determine cell counts with a Coulter Counter. FACS analysis was performed for the differentiation of the white blood cells using their forward/sideward scatter characteristics.
Administration of rPAF-AH recombinant platelet-activating factor acetylhydrolase (rPAF-AH, kindly provided by Dr Dietsch, ICOS, Washington, DC, USA) was used to evaluate the role of platelet activating factor (PAF) in the effects observed. It was given intravenously as a bolus, at a dose of 5 mg/kg, 5 min before the IVIG preparation.
Activation of PMN in vivo To determine activation of circulating neutrophils after infusion of IVIG, EDTA blood samples were obtained and kept on melting ice until the end of the experiment. Cells were washed and incubated with anti-CD11b mAb or L-selectin mAb as described above. Whole blood samples were used to minimize in vitro activation of leucocytes.
Macrophage depletion by Cl2MBP-liposomes Intravenous administration of dichloromethylene bisphosphonate (Cl2MBP) liposomes results in depletion of Kupffer cells in the liver and macrophages in the spleen, within one day after a single administration of the liposomes (van Rooijen & Sanders, 1997), without affecting neutrophils. (van Rooijen & Sanders, 1990). Hence, to study the effect of macrophages in IVIG-induced alterations of neutrophils, liposomes that contain either Cl2MBP (a gift of Roche diagnostics GmbH, Mannheim, Germany) or PBS. These liposomes were prepared as described previously (van Rooijen & Sanders, 1990). The liposomes were resuspended in PBS before use. To deplete macrophages, rats were treated with 2 ml Cl2MBP liposomes by intravenous administration 1 or 3 d before administration of aged IVIG. Control rats received 2 ml of PBS-liposomes. In each experiment, three rats were used for each treatment.
Measurements of tumour necrosis factor-α (TNF-α) TNF was measured using a commercial enzyme-linked immunosorbent assay (ELISA) for rat TNF-α (Biosource International, Camarillo, CA, USA) according to the manufacturer's instructions. The concentrations of sample TNF-α was determined with reference to serially diluted recombinant rat TNF-α. The lower limit of detection was 2·3 pg/ml.
Measurement of myeloperoxidase in the lungs To assess sequestration of neutrophils in the lungs, myeloperoxidase (MPO) activity was measured as previously described (Seekamp et al, 1993). One lobe of the left lung was homogenized (Ultra turrax, IKA Labortechnik, Staufen, Germany) in 50 mmol/l potassium phosphate buffer (pH 6·0) and 0·5% (w/v) hexadecylmethyl ammonium bromide. The suspension was sonicated and centrifuged at 10 000 r.p.m. for 15 min at 4°C. Aliquots were added to 100 µl of citrate buffer (0·2 mol/l dibasic sodium phosphate/0·1 mol/l citric acid) supplemented with 0·03% hydrogen peroxidase and 1% (w/v) O-dianisidine hydrochloride. MPO activity was assayed by recording the absorbance after 10 min at 450 nm. Purified human MPO (Sigma) was used as standard.
Histological analysis At the end of the experiment, lungs were taken out immediately. The weight of the left lung was determined, before and after incubation for 48 h at 60°C to calculate the wet/dry weight ratio. After ligation of the main bronchus to the left lung, 2 ml of phosphate-buffered formalin was infused to fill the two lobes of the right lung. After fixation, lung tissue was embedded in resin, 3–5 µm thick sections were stained with haematoxilin and eosin and examined by light microscopy for the presence of neutrophils.
Plasma clearance of IgG dimers An aged IVIG-preparation was separated on a calibrated Superose12 gel-filtration column connected to a fast-performance liquid chromatography (FPLC) system (Pharmacia) to obtain an IgG dimer fraction. This IgG dimer fraction, consisting of approximately 60% dimers, was administered to rats, and blood samples were taken at different time-points. To determine the amount of circulating IgG dimers at different time-points after administration, plasma samples were separated by gel filtration as described above, and fractions corresponding to IgG dimers were evaluated in a human IgG ELISA. Briefly, plates were coated overnight with a monoclonal antibody against human IgG (MH16, CLB). Plasma samples were appropriately diluted in PBS/0·02% milk and incubated for 1 h at room temperature. After washing, plates were incubated with alkaline phosphatase-labelled antibodies to human IgG (Sigma). After addition of p-nitrophenyl phosphatate as substrate, the optical density was measured with a Titertek Multiscan (Flow Labs. McLean, VA, USA). Serial dilution of aged IVIG was used as a standard.
Statistical analysis Analyses of variance (ANOVA) were used to compare the neutrophil counts over time and the accumulation of neutrophils in the lungs. Individual group means were then compared with the Tukey multiple comparison tests. All values were expressed as mean ± SEM. A P-value < 0·05 was considered to indicate significant difference.
Activation of rat neutrophils in vitro
Forward/sideward scatter analysis, as well as flowcytometric analysis with mAb HIS48 of peripheral blood cells, revealed that neutrophils only constitute a minor population of total leucocytes in the rats used for the experiments. Binding studies with biotinylated IVIG showed that IVIG binds to these neutrophils, as well as to mononuclear cells, in a similar fashion to human cells (Fig 1). These experiments also showed binding of IVIG to a HIS48-negative population, which could be identified as lymphocytes presumably NK cells, based on their forward/sideward scatter characteristics.
Activation of neutrophils was determined by flowcytometric analysis of CD11b and L-selectin surface expression. Incubation of whole blood from a healthy donor with aged IVIG led to up regulation of CD11b and shedding of L-selectin (Fig 2). Incubation of rat whole blood with aged IVIG showed a similar effect on the expression of CD11b and L-selectin. Thus, these studies indicated that rat neutrophils interacted with human IVIG preparations as has been shown for human neutrophils (Teeling et al, 1998). To determine whether the activation of human neutrophils was mediated via direct interaction with Fcγ-receptors, blocking antibodies against FcγRIIa were added to the human whole-blood cultures. Up regulation of CD11b and shedding of L-selectin was completely abrogated, indicating that the activation of human neutrophils occurred via triggering of FcγRII, similarly as for degranulation (Teeling et al, 1998). Because antibodies against rat Fcγ-receptors are not available, we were not able to do similar studies with rat neutrophils.
Activation of circulating neutrophils in rats
The effect of IVIG in vivo was assessed using a well-defined rat model (Bleeker et al, 1989). Table I shows the actual amount of IgG monomer, dimer and polymers present in the IVIG preparations used in the experiments. After obtaining baseline blood samples from the rats, the IVIG preparations and saline were administered intravenously as a bolus injection, whereas the positive control LPS was administered as a 5-min-infusion. The effects of aged IVIG (250 mg/kg), fresh IVIG (250 mg/kg), saline (4 ml/kg), and LPS (0·5 mg/kg) on circulating neutrophil numbers in rats are shown in Fig 3.
Table I. Actual amount of IgG monomer, dimer and polymer contents in IVIG preparations used in the experiments.
Monomeric IgG % of total
Dimeric IgG % of total
Polymeric IgG % of total
Analysis was performed by size-exclusion chromatography using a calibrated Superose12 gel-filtration column connected to a FPLC system.
*Indicate P < 0·0001 as determined using Student's t-tests comparing aged IVIG with fresh IVIG. All values are expressed as mean ± SEM.
Baseline neutrophil numbers in the Wistar rats were well detectable but low, i.e. 0·53 ± 0·28 × 106 cells per ml blood. After infusion of aged IVIG, a transient decrease in the number of circulating neutrophils was observed, immediately followed by an increase in circulating neutrophils that reached maximal numbers 2 h after administration (Fig 3). Monocyte counts also increased after initial monocytopenia, reaching maximal levels 2 h after administration (data not shown). The neutropenia induced by LPS was more prolonged than that by aged IVIG but eventually also the LPS-induced neutropenia was followed by an increase in the number of circulating neutrophils. In contrast, the number of circulating monocytes did not increase during this period. Infusion of fresh IVIG or saline also induced an increase in the number of circulating neutrophils, although not as extensive as with aged IVIG. The increase in neutrophil numbers in saline-treated rats can be explained by the effects of the surgical procedure of cannulation of the carotid artery and/or the prolonged anaesthesia because control infusions resulted in similar increases. The possible contribution of endogenous adrenal hormones (potentially increased as a result of a stress reaction in the animals) to changes in circulating leucocyte subsets after infusion of IVIG was investigated using infusion of 2 µg/kg/h adrenaline. However, adrenaline induced less granulocytosis than IVIG, i.e. 1·4 × 106 neutrophils per ml of blood, ruling out that IVIG induced granulocytosis merely through an indirect effect mediated by adrenal hormones.
In the results presented so far, the rats were given a bolus injection of IVIG at a dose of 250 mg/kg. The administration of IVIG as a bolus over 10 s was much more rapid than the protocol used for clinical administration. Patients are usually treated by a continuous infusion of IVIG over a 1–3-h period at a dose of 0·4 g/kg. To simulate the clinical situation more closely, we also tested the effect of continuous infusion of IVIG in rats. In this protocol, aged IVIG was given at a dose of 1 g/kg or 2·5 g/kg over a 1-h period. A marked increase of neutrophils was observed after continuous infusion of aged IVIG (Fig 3), although not as extensive as observed after bolus injection of aged IVIG.
Both aged IVIG administered as a bolus injection as well as LPS induced extensive neutropenia (35·5% and 39·4% decrease respectively) within 5 min after its administration, which was accompanied by a decrease in the number of circulating neutrophils (Fig 4). Continuous infusion of aged IVIG induced a similar neutropenia as a bolus injection: the number of circulating neutrophils decreased by 35·7% compared with baseline.
The expression of CD11b and L-selectin on the surface of circulating neutrophils monitored throughout the 4-h period after administration of the test solutions is shown (Fig 5A,B). LPS induced an extensive increase in expression of CD11b (mean increase of 500%), reaching the highest expression at 2 h after injection. A concomitant decrease in L-selectin expression was observed, which started within 5 min after injection and lasted for the entire observation period. L-selectin did not change after administration of aged IVIG until 2 h after injection, after which a moderate decrease was observed. In contrast, the expression of CD11b did not change during the observation period. Fresh IVIG and saline did not induce significant changes CD11b or L-selectin, although a moderate increase of the latter was consistently observed 2 h after administration, coinciding with the granulocytosis. Finally, continuous infusion of aged IVIG induced changes in the expression of L-selectin, however, CD11b on the surface of neutrophils was also not altered.
Sequestration of neutrophils in the lungs
To determine if infusion of IVIG may lead to sequestration of neutrophils in the tissues, lungs were also analysed after the experiment. The number of neutrophils in lungs, as well as MPO activity in lung homogenates and the wet/dry weight ratio of lungs, were examined to determine sequestration and lung injury. Figure 6 shows the mean ± SEM of four different experiments. The major morphological changes in animals receiving IVIG were related to an intrapulmonary, intracapillary sequestration of neutrophils. To obtain quantitative estimates of the amount of neutrophils sequestered in IVIG-treated rats compared with animals treated with LPS or saline, numbers of neutrophils per alveolus were counted using light microscopy. Two lung sections per animal were examined. The number of neutrophils in the control rats were 0·69 ± 0·18 per alveolus, whereas aged IVIG-treated rats showed a significant increase i.e. 2·10 ± 0·68 per alveolus. Administration of fresh IVIG as a bolus infusion and continuous infusion of aged IVIG tended to increase sequestration of neutrophils in the lungs (1·08 ± 0·28 and 1·00 ± 0·15 neutrophils per alveolus respectively) but these increased numbers of sequestered neutrophils were not significant. The activity of MPO was measured in homogenates of lung tissue as another parameter for neutrophil sequestration. Rats treated with aged IVIG showed an increased MPO activity compared with saline-treated rats. Also, the rats that received fresh IVIG tended to have more MPO in their lungs (Fig 6C). However, continuous infusion of aged IVIG did not result in a significantly increased MPO activity. To determine if the sequestration of neutrophils by IVIG led to oedema formation, lungs from treated rats were analysed for wet/dry weight ratios. Aged IVIG as well as fresh IVIG both induced an increased wet/dry weight ratio. In none of the experiments (including the ones with LPS) did we observe fibrin deposition within alveolar spaces or morphologic damage to the alveolar lining of epithelial cells. Pre-treatment with rPAF-AH did not affect the number of sequestered neutrophils in lung tissue after infusion of IVIG (data not shown).
Plasma clearance of IgG dimers
To determine how long IgG dimers may interact with circulating neutrophils after a bolus administration of IVIG, we determined the plasma clearance of IgG dimers purified from aged IVIG using gel chromatography. Measurement of dimer concentration in plasma revealed that after infusion of aged IVIG, IgG dimers were present in the circulation for the whole 4-h observation period. These results indicate that despite the presence of IgG dimers during the entire observation period, neutrophil activation occurred only shortly after administration of aged IVIG. To examine if the circulating neutrophils were refractory to activation, rats were given a second bolus injection of aged IVIG 2 h after the first injection. The first infusion of aged IVIG resulted in a decrease in circulating neutrophils of 58 ± 6% compared with baseline levels followed by a granulocytosis. After 2 h, the number of circulating neutrophils was increased to 6·6 ± 1·2 × 106 neutrophils/ml blood. A second dose of aged IVIG (250 mg/kg) was given and the number of neutrophils was determined 5 min after infusion. In contrast to the first infusion, a second dose of aged IVIG (250 mg/kg) resulted only in a decrease of about 10 ± 1%. However, incubation of these cells in vitro showed activation after stimulation with IVIG. These data further suggest that in vivo neutrophils do not become activated in a direct way by IVIG.
Role of macrophages, PAF and cytokines
The mechanism by which IVIG induces changes in neutrophil numbers was studied by two different approaches. The role of macrophages was evaluated by pre-treatment of the rats with liposomes containing Cl2MBP to eliminate macrophages. In these macrophage-depleted rats, the biphasic changes in neutrophil numbers were no longer observed after infusion of aged IVIG, whereas in rats pre-treated with liposomes containing PBS the changes were observed. (Fig 7). Depletion of macrophages by Cl2MBP might result in an inflammatory reaction that would blunt the phenomena studied (Laman et al, 1990). However, it has been shown that 2 d after administration of Cl2MBP-liposomes, cell remnants have been cleared from the spleen and serum concentration of various released molecules have reached normal values (van Rooijen et al, 1989). Rats that were given Cl2MBP liposomes 3 d prior to the infusion of aged IVIG showed similar results, ruling out the possibility that Cl2MBP released by leakage from liposomes or dying macrophages may also contribute to the suppression of inflammatory reactions.
Macrophages are known to produce a wide range of activating substances. We focused on two. First TNF-α was measured in plasma at different time-points, obtained from rats treated with aged IVIG, fresh IVIG, LPS or saline (Fig 8). In our model, detectable TNF-α release in plasma did not occur in rats treated with IVIG or saline, although administration of LPS was associated with a marked release of TNF-α, reaching maximal value 1 h after infusion. Second, PAF was previously found to be involved in IVIG-induced hypotension, which could completely be abrogated by pre-treatment with rPAF-AH (Bleeker et al, 2000). However, the protective effects of rPAF-AH were not seen in the changes in neutrophil numbers (Fig 9). Thus, these results showed that the release of PAF did not explain the changes in neutrophils numbers after infusion of IVIG.
The present study demonstrates a mild activation of neutrophils after infusion of IVIG in a rat model. For several reasons, we tested two IVIG preparations, i.e. fresh IVIG and aged IVIG, which only differed in their IgG dimer content. First, polymers and dimers present in the preparations were previously found to be responsible for the activation of human neutrophils in vitro (Teeling et al, 1998). Second, in some commercially available IVIG preparations we found high amounts of IgG dimers, up to 15% (Bleeker et al, 2000). During the production of the different IVIG preparations, IgG dimers can be formed, depending on the temperature, ionic strength, pH, additions such as glucose and duration of the whole manufacturing process.
Rat neutrophils were activated to the same extent as human neutrophils in vitro, indicating that these cells are able to interact with human IVIG preparations as was confirmed in direct binding studies.
In general, systemic activation of neutrophils in vivo is associated with a rapid disappearance of neutrophils from the circulation followed by sequestration in the different tissues, especially the lungs (Carlos & Harlan, 1990;Butcher, 1991; Worthen et al, 1991; van Zee et al, 1992; Borregaard et al, 1994). This sequestration of neutrophils is caused by changes in the cytoskeleton, transient cell-stiffening and changes in expression of adhesion molecules (Worthen et al, 1991; Borregaard et al, 1994). Administration of LPS, a well-known agonist for neutrophils, caused a marked neutropenia within 5 min of infusion. Aged IVIG also caused a decrease in circulating neutrophils whereas fresh IVIG did not, suggesting that activation of neutrophils is caused by IgG dimers present in the IVIG preparation and not by antibodies against constituents on rat cells. Continuous infusion of aged IVIG further demonstrated that the effects observed were not caused by the rapid administration of IVIG.
Activation of neutrophils was phenotypically studied by assessment of the expression of CD11b and L-selectin on the surface of circulating neutrophils. Administration of IVIG (either aged or fresh) did not clearly exhibit signs of activation as was observed in the in vitro experiments. One reason for this discrepancy between in vitro and in vivo findings could be that the activated neutrophils became under represented in the blood samples by the coinciding granulocytosis, with neutrophils freshly released from the bone marrow exhibiting a higher expression of L-selectin (van Eeden et al, 1995). Administration of LPS did not induce a coinciding granulocytosis, thus explaining the marked changes in CD11b and L-selectin induced by this agonist. Alternatively, histological observations and the MPO contents of the lungs support the idea that activated neutrophils may leave the circulation.
Conversely, L-selectin expression on neutrophils increased after infusion of fresh IVIG. As explained above, recruitment of neutrophils from the bone marrow may explain the increase in mean fluorescence intensities of L-selectin on circulating neutrophils after infusion of fresh IVIG. The histological results suggest that despite minimal changes observed in peripheral blood, administration of fresh IVIG as a bolus infusion also leads to mild activation of neutrophils in vivo. Interestingly, continuous infusion of aged IVIG showed less sequestration of neutrophils into lung tissue, suggesting that the rate of infusion plays a role in the IVIG-induced neutrophil activation. This finding is very recrudescent of clinical practice, in which adverse reactions are associated with rapid infusion rates (Duhem et al, 1994; Bagdasarian et al, 1998). Our observations suggest possible involvement of neutrophil activation in the induction of these clinical side-effects.
Mobilization of the marginated pool, as occurs during a stress reaction (Gordon, 1955), cannot explain the marked granulocytosis observed after infusion of aged IVIG. This is because demargination of neutrophils after administration of adrenaline, resulted in a maximum two- to threefold increase in the circulating numbers of neutrophils. Several inflammatory mediators are able to induce granulocytosis, such as C5a, cytokines, the growth factors, granulocyte colony-stimulating factor (G-CSF), granulocye macrophage CSF (GM-CSF), eicosanoids and glucocorticosteroids (van Zee et al, 1992; Jagels & Hugli, 1994; 1995; Ulich et al, 1987; Kubo et al, 1998; Opdenakker et al, 1998). In the present study, we tested whether TNF-α, one of the early response cytokines, was released and hence constitutes a mechanism by which IVIG induced granulocytosis. LPS caused a rise in TNF-α plasma levels, whereas levels of TNF-α remained undetectable after infusion of IVIG, making the involvement of TNF-α less probable. Another mediator, previously shown to be involved in IVIG-induced changes, may be PAF that is released from macrophages (Bleeker et al, 1989). Although rPAF-AH protected rats from hypotension, this inhibition could not prevent the quantitative changes in neutrophil counts. Yet, macrophages were major contributors herein: the biphasic effect of aged IVIG is completely abrogated in macrophage-depleted rats. Notably, the experiments with IVIG were carried out 2 d after depletion of macrophages by Cl2MBP, by which time an inflammatory reaction induced by the depletion would have been unlikely to blunt the phenomena studied (van Rooijen et al, 1989; van Rooijen & Sanders, 1990; Qian et al, 1994). At the moment, complement activation and subsequent formation of C5a as a cause of the granulocytosis cannot be ruled out but it seems less apparent knowing that macrophages play such an important role.
A more likely explanation for the discrepancy between in vivo and in vitro findings would be indirect activation of neutrophils via macrophages because macrophage-depleted rats no longer show activation of circulating neutrophils after infusion of a bolus injection of aged IVIG. IgG dimers were present in the circulation during the entire 4-h study period, making it less probable that direct activation of neutrophils by IgG dimers occurs in vivo. Furthermore, a second dose of aged IVIG did not induce neutropenia. However, the macrophage mediators involved in the activation of neutrophils still remains to be elucidated because potential mediators like TNF-α and PAF were not found to play a major role in our model.
The present study demonstrates that a mild activation of neutrophils occurs after infusion of IVIG in a rat model. To what extent this occurs in patients receiving IVIG remains to be established, although we have found increasing plasma levels of elastase and lactoferrin in a few patients experiencing clinical side-effects after IVIG infusion (unpublished observations). In addition, a recent clinical study in volunteers was performed in which different experimental IVIG-preparations were compared (Spycher et al, 1999). In this study, a correlation was found between IgG dimer content in the preparations and the occurrence of clinical adverse reactions in humans. These clinical side-effects in the humans were accompanied by a transient decrease in neutrophil and monocyte numbers in the peripheral blood and an increase in TNF serum concentrations.
Although, in general, the neutrophil activation will be mild and hence well tolerated by most patients, IVIG infusion may have serious consequences in patients with pre-activated neutrophils in their circulation. Recently, a healthy volunteer was described who developed adult respiratory distress syndrome (ARDS) after administration of IVIG (Dooren et al, 1998).
The beneficial effect of IVIG therapy in the treatment of several autoimmune diseases is thought to be mediated by autoantibody idiotypes present in the IVIG preparations. Formation of such idiotype–anti-idiotype dimers may therefore prevent autoantibodies from binding to antigens (Roux & Tankersley, 1990; Mouton et al, 1994; Vassilev et al, 1995). In this study, we have shown that the dimers present in IVIG preparations may also constitute the less favourable effects in the treatment of patients. For this reason, IVIG with a high content of dimers may be harmful in patients with pre-activated neutrophils in the circulation.