Stimulation of neutrophil functions by C5adesArg: an in vitro model of haemodialysis

Cuprophane membranes during haemodialysis significantly increase the plasma levels of C5adesArg (maximal 55 μg C5aadesArg/1 blood after 30 min) whereas Hemophane or Polysulphonemembranes induce only low plasma levels of C5adesArg. C5adesArg generated in vitro by yeast incubation of autologous plasma stimulates PMN chemotaxis and oxidative metabolism but has no effect on enzyme release. Preincubation of whole blood with C5adesArg causes aggregation and changed oxidative burst activity of the isolated PMN. These changes are similar to those found in cells from patients after haemodialysis with cuprophane membranes. So the elevated plasma levels of C5adesArg after haemodialysis explain some of the changes in PMN functions, but additional mechanisms have to be assumed.


Introduction
During haemodialysis (HD) with various membranes complement components are activated, which results in elevated levels of C5a and C3a. 1 '2 Additionally neutrophil granulocytes (PMN) are activated, which results in the release of lysosomal enzymes and an altered oxidative metabolism. 4 To explain the change in the PMN functions two effects have to be evaluated: a direct effect of the membrane material on PMN and an indirect effect of membrane activated plasma components on the cells. If the two effects are exerted, they may differ in their extent, in their kinetic, and in their dependence on the membrane material. They can be studied in two ways: by using membrane activated cells without the addition of plasma and by using normal cells with activated plasma.
In this study we set out to investigate the incubation of normal PMN with yeast activated plasma as an in vitro model of the plasma effects during haemodialysis.

Material and Methods
Isolation of human PMN: Blood was drawn by venipuncture after obtaining written consent from normal, healthy donors. PMN were isolated using Ficoll (Pharmacia-LKB, Bromma, Sweden) separation and sedimentation of erythrocytes. The  Chemotaxis assay: Chemotactic assays were performed in modified Boyden chambers. 7 Activated plasma (2.5-20%) in HBSS (Gibco, Germany) was put in the lower compartment and the PMN (106) in the upper compartment of the Boyden chambers. Cellulose ester filters (ME 29, Schleicher and Schtill, Dassel, Germany) were used. The chambers were incubated for 90 min at 37C, then the filters were removed, stained with hematoxylin, and mounted on slides. The PMN were counted in the filters at every 10/m interval by a computerized imaging method (Bausch and Lomb, Frankfurt, Germany) and the chemotactic index (CI) calculated, which reflects the mean distance travelled by the activated cells. 7 Cytochrome c test: Superoxide anion generation was measured by cytochrome c reduction. PMN (106) were incubated with 75/M cytochrome c and stimuli (Phorbol 12-myristate 13-acetate, PMN, 1 #g m1-1 or activated plasma 0.3-20% as indicated in results) in Hanks' balanced salt solution (HBSS) for 10 min at 37C. The reaction was stopped by cooling in ice and centrifugation (10 min, 200 x g'). The absorption of the supernatants was determined at 550nm, and the superoxide generation was calculated.
Elastase release: PMN (4 x 106) were incubated with activated plasma (2.5-20% as indicated in results) in HBSS for 15 min at 37C. Released elastase (HLE) was complexed with a l-proteinase inhibitor (alPI) by addition of plasma (25%) and further incubated for 2 min. The samples were centrifuged (10 min, 200 x g) and the HLE-alPI-complex was measured by ELISA (Merck, Darmstadt, Germany). Total HLE content of the PMN was measured after lysis with cetyltrimethylammonium bromide. 1 Preincubation experiments: Sixty ml of anticoagulated blood were centrifuged and the plasma separated. Ten ml of the plasma were activated and 10 ml inactivated as described above. The blood was then divided into three parts and each was treated in one of the following ways: (1) PMN were directly isolated from one part, (2) one part of the blood was mixed with activated plasma (final concentration 10%) and incubated for 10 min at 37C, then the PMN were isolated, (3) one part of the blood was mixed with inactivated plasma (final concentration 10%) and treated as part 2. Patient studies: Twenty-seven patients gave their informed consent to participate in the study. Haemodialysis (HD) was performed by an A 2008 C (Fresenius, Bad Homburg, Germany) or an AK-10 (Gambro, Hechingen, Germany) volume controlled equipment. As our dialyser we used cuprophane: SMAD 125/140 (SMAD, Lyon, France), hemophane: MO 450 (SMAD) and GFS 120/140 MCH (Gambro, Lund, Sweden), and polysulfone: F6 (Fresenius).
EDTA blood was taken from the arterial line before systemic heparinization prior to HD and 10, 30, 60, 120 and 180 min after the beginning of HD. The blood was immediately centrifuged and the plasma stored deep frozen until assessed for C5a by ELISA. Chemicals were if not otherwise stated by Sigma Chemie, Deisenhofen, Germany. Statistical analysis" Data are represented as the mean-t-1 standard deviation (SD). Statistical comparisons were performed by the paired Student's t-test.

Results
CSadesArg generation during diasis: Figure   In vitro generation of CSadesArg" Generation of C5adesArg by yeast activation of autologous plasma samples results in a mean value of 1040 -+-450/_zg C5adesarg/1 (n 8). In parallel inactivated plasma samples contain 1.7 -t-0.9/g C5adesarg/1 (n 8). Stimulation of chemotaxis by activated plasma" Figure 2 shows the concentration dependent stimulation of chemotaxis by the autologous plasma samples. Maximal stimulation is achieved by 10% activated plasma which corresponds to about 100 C5adesarg/l. Stimulation of oxidative burst by activated plasma" Direct effect of activated plasma. The superoxide generation of PMN is already stimulated by 1% of activated plasma corresponding to about 10 g C5a&sarg/1 to 2 nmol O-/10 min/106 PMN (Fig. 3). The yield of PMN is reduced to 18% of the control after incubation with activated plasma (Table 1, p < 0.002). Incubation of the blood with inactivated plasma also reduces the yield (55% of control, not significant NS). If the superoxide generation of the isolated PMN is tested without the addition of a stimulus, the plasma preincubated cells produce higher amounts of superoxide than  during haemodialysis have to be made in venous blood samples which come directly from the dialyser, where C5a is supposed to be generated. '15 We found maximal plasma levels of 55 #g C5adesArg after 30 min of haemodialysis with cuprophan membranes (Fig. 1) (Fig. 2). Maximal stimulation was achieved by 10% activated plasma corresponding to 60 nM C5adesArg the same value as described by Yancey et al. 19 and Chenoweth and Hugli. 2 The native C5a fragment was about 20-fold more potent (ECs0 C5a 0.5 nM). 21 The stimulation of the oxidative burst with activated plasma (Fig. 3) gave no maximum but was increased in the whole range tested (final concentration 1-20% activated plasma) as already described by Goldstein (Table 1). The prestimulated PMN produced less superoxide after further stimulation with PMA than the control cells or the cells that were preincubated with inactivated plasma (Fig. 4). The extent of these effects differed from one blood donor to the other, reflected by the large standard deviations, but the ratio of the values was always the same. So the preincubation experiments imitate the in vivo results, which also show leukopenia after cuprophan dialysis 25'26 and diminished superoxide generation of the isolated PMN. 4 It seems probable that after stimulation a subpopulation of PMN with lower activity was isolated, 27'28 because priming of PMN with a stimulus normally resulted in enhanced activity towards a second stimulus. 29 However, if the prestimulated PMN are further stimulated with activated plasma, the superoxide generation is enhanced in comparison to untreated cells (Fig. 4). So the activated plasma seems to contain a factor which compensates the cellular effect. These apparently contradictory results are also produced during dialysis, where the superoxide generation of PMN in autologous plasma is enhanced at the same time as the superoxide generation of isolated PMN is reduced. 3'31 Our preincubation experiments were performed with high CSadesarg concentrations which correspond to the values after cuprophane dialysis. And they show similar results for PMN function as are produced during cuprophane dialysis. Preincubation with lower C5aae,Arg concentrations (5% activated plasma corresponding to a final concentration of about 50/g C5ades,rg/1) shows no effect. But other membrane materials such as hemophane also induce leukopenia and changes of PMN oxidative burst although they only generate very small amounts of C5adesArg. 4'16-18 So C5adesArg may be responsible for the complement induced effects after cuprophan dialysis but not for the effects after HD with the other membranes. This is confirmed by several other authors who find no correlation between the extent of complement activation and leukopenia.  Moreover complement activation by cuprophane also depends on additional serum factors which diflCer greatly between individuals. 6 The second cellular effect during dialysis, the enzyme release, 37'4 is not mimicked by activated plasma. Incubation of isolated PMN with up to 20% activated plasma results in no release of HLE (Fig.  5) Release of azurophil granules is only stimulated after cytochalasin b preincubation 8 while only specific granules are released by soluble stimuli. 9 So the HLE release of PMN during dialysis 4 can not be explained by the activated plasma compo-nents but additional stimulating effects of the dialysis membrane have to be assumed. CSadesarg generated during haemodialysis may influence some of PMN functions e.g. the oxidative burst. But for other functions e.g. the enzyme release an additional influence of the membrane material has to take effect.