Activation of the Classical Complement Pathway by a Polysaccharide from Sugar Cane Xiaoy
Li I and Walther Vogt
Abstract: The effects of an immunostimulating polysaccharide, Bo, from sugar cane, on the complement system have been investigated. Bo, a glucan of about 10,000 too/ wt, was found to activate complement in whole human and guinea pig serum in vitro by the classical pathway. Complement consumption was also demonstrated in guinea pigs upon intravenous injection. Specifically, C1 is activated, and C4 and C2, as well as C3, are consumed. The activation is prevented when Ca + + ions are chelated by ethyleneglycoltetraacetic acid, and when C1q is lacking. Hence, it does not rest on direct activation of C1s. Supplementation of Clq-deficient human serum with purified C l q restores the ability to be activated by Bo. The alternative pathway of complement is little if at all affected by the polysaccharide. The activation of C1 seems to be mediated by immune complex formation between Bo and naturally occurring immunoglobulins. Complement in sera from two severely hypogammaglobulinemic patients was not activated by Bo, but was made reactive by addition of ourified human immunoglobulin G.
INTRODUCTION
Various fungi contain polysaccharides that have immunopotentiating and tumor-inhibiting effects, e.g., "PS-K", a protein-bound polysaccharide extracted from Coriolus versicola (Usui et al., 1976; Ehrke et al., 1980). While attempting to isolate the active principles from other cultured fungi, it was found that some immunologically active polysaccharides originated from the growth medium rather than from the mycelia proper (Jin et al., 1981). The culture medium used in these cases was bagasse, the residue from sugar cane left after extraction of the sugar. Therefore, bagasse was taken directly as the source for the purification of immunopotentiating polysaccharides. By extraction, precipitation, and chromatographic procedures a highly purified compound (Bo) was obtained as well as three other polysaccharides (BI. B2. 133). Bo is a water-soluble glucan with a mol wt of about 10,000. It stimulates phagocytosis of mice recituloendothelial system (RES) macrophages in viva, causes a marked growth of the spleen in mice, antagonizes immunosuppressive effects of prednisolone and cyclophosphamide, and prolongs the survival time of mice irradiated by X-ray or poisoned with CCI 4 (Jin et al., 1981).
ISupported by a grant from the Alexander van Humboldt-Stiflung,on leave from Shanghai Institute of Materia Medica, Chinese Academy of Sciences. Received December 1i, 1981; revised February 23, 1982. From the Max-Planck-lnstitutf~r experimentelle Medizin, Department of Biochemical Pharmacology, 3400 G6ttingen, Germany. Address request for reprints to: Dr. Walther Vogt, Max-Planck-lnstitutffir experimentelle Medizin, Department of Biochemical Pharmacology, Hermann-Rein-Slrasse 3, 3400 G6ttingen, Germany. © ElsevierScience Publishing Co., Inc., 1982 52 Vanderbilt Ave., New York, N . Y . lmmunopharmacology 5, 31-38
31 0162-3109/82/0503100502.75
32
X. Li and W. Vogt
In the present investigation the effect of Bo on the complement system was studied. Rather unexpectedly it was found to activate CI and, hence, the classical pathway, whereas it has little, if any, direct effect on the alternative pathway of complement activation.
MATERIALS AND METHODS Glucan Bo The purification of Bo is described in detail elsewhere (Zhu et al., 1982). In brief, bagasse was heated in a water suspension to 100°C, the supernatant was then deproteinated with trichloroacetic acid, concentrated by ultzafil~ation, and dialyzed. Precipitation with ethyl alcohol yielded a crude polysaccharide fraction that was chromatographed on diethylaminoethyl (DEAE) cellulose and DEAE Sephadex. The nonabsorbed breakthrough fraction represents Bo. Its molecular weight is 10,700_+ I000 (according to osmotic pressure), only glucose was found as a constituent. Composition: C 41.88%, H 6.93%, N 0%; in good accordance with a carbohydrate molecule (C~H201) .. The optical rotation was [od34o + 134.3 °. Purification and analyses were performed by Dr. G.P. Zhu, Shanghai Institute of Materia Medica, Chinese Academy of Sciences. For assays, Bo was dissolved in veronal-buffered saline or other appropriate buffers. Concentrations are given on a molar basis, taking i0,000 as the molecular weight.
Chemicals Ethylenediaminetetraacetic acid (EDTA) was from Merck (Darmstadt) and ethyleneglycol-bis (2-aminoethyl ether) tetraacetic acid (EGTA) was from Serva (Heidelberg); stock solutions of these chelating agents were titrated to pH 7.4 with NaOH. Mg2*-EGTA was prepared by mixing equimolar amounts of MgCI2 and EGTA solution. Veronal-buffered (2.5 mM; pH 7.4) saline (80 mM) containing 2.5% glucose and 0.1% gelatine (VBS) was used as a diluent in complement assays. Zymosan was prepared as described (Pillemer and Ecker, 1941). Purified endotoxic lipopolysaccharide from S. minnesota was kindly provided by Dr. C. Galanos. Freiburg. It was dissolved with a trace amount of ethanol and further diluted with buffer.
Complement Components For assay of single complement components in sera, functionally pure preparations from Cordis Corp. (Miami) were used. CI, C6, C7, C8, and C9 were from guinea pig serum, all other components were of human origin. A C l q preparation was obtained from 5 ml of human serum. The serum was supplemented with 0.01 M EDTA and dialyzed against 0.01 M acetate buffer, pH 5.5, containing 0.I M NaCI and 0.01 M EDTA. The sample was then passed through a column of carboxymethyl cellulose (CMC) (CM 32; 1.2 × 15 cm) with the same buffer. The absorbed C l q was eluted with 0.01 M phosphate buffer, pH 7.3, containing 0.01 M EDTA and 0.15 M NaCI. The eluate was dialyzed against VBS lacking divalent cations and containing 0.001 M EDTA; it was then concentrated by ultrafiltzation to I ml. In order to block any contaminating Cls, the C1q sample was then treated with 2.5 × 10-2M diisopropylfluorophosphate (DFP). The nonabsorbed protein fractions were collected, passed once more through the regener-
Abbreviations. DFP: diisopropylfluorophosphate; EDTA: ethylendiaminetetraacetic acid; EGTA: ethyleneglycol-bis-(2-aminoethyl ether) tetraacetic acid; VBS: veronal-buffered saline; RES: reticuloendothelial system; DEAE: diethylaminoethyl; CMC: carboxymethyl cellulose; iv; intravenously.
Sugar Cane Polysaccharide Activates Complement Pathway
ated CMC column, dialyzed against the phosphate-EDTA buffer as above, and concentrated to 4 ml; this sample was used as Clq-deflcient serum. A sample of serum from a patient lacking C l q and some highly purified C l q were kindly provided by Dr. M. Loos, Mainz. Sera
Blood was taken from healthy donors by venipuncture and allowed to clot spontaneously. For most experiments individual sera were used, some were performed with pooled serum. Guinea pig sera were obtained from blood taken by cardiac puncture.
Human Immunoglobulin (Ig) G Pooled human serum ( 1 0 - 12 ml) was passed through a column of 10 ml protein A-Sepharose (Pharmacia) in 0.01 M phosphate buffer, pH 7.3. with 0.3 M NaCI. The absorbed IgG was eluted with 0.I M glycine buffer, pH 2.8, containing 0.4 M NaCI. The eluted material was collected in 2 ml fractions, in tubes containing 0.25 ml of I M phosphate, pH 8.0, to immediately neutralize the IgG solutions. The IgG fractions were then dialyzed against VBS, containing 0.04% NaN 3, and concentrated to appropriate volumes. A monoclonal human IgG preparation was kindly provided by Prof. N. Hilschmann, G6ttingen.
Complement Assays Whole complement activity of human and guinea pig sera was assessed by immune hemolysis of sensitized sheep red cells (107 EA in I ml of serum dilutions). Activities are given as CH50 or in percent of controls. Single components were assayed by incubation of appropriate cell intermediates with varying dilutions of the component to be tested and excess of the later reacting components, as described earlier (Vogt et al., 1979). Cleavage of factor B was observed qualitatively by immunoelectzophoresis, and quantitatively by rocket electrophoresis of human serum samples using anti-C3 activator serum (Behringwerke, Marburg). RESULTS
Effect on Whole Complement Activity Bo showed no direct lyric activity on sheep red blood cells when tested in concentrations of up to I mM. Human serum to which Bo was added did not change its complement activity when diluted and assayed immediately after the addition, i.e.. Bo had no significant direct inhibiting or enhancing effect on complement reactions. When the serum was incubated with Bo, for 30 min at 30°C prior to the assay, its hemolytic complement activity fell considerably. There was no simple linear concentration-effect relationship, but a plateau of highest activity in complement consumption by Bo was reached at about 0.05 to 0. I mM concentrations. Higher concentrations (0.25 to 0.5 mM) caused lesser effects (Figure I). In diluted serum the anticomplementary effect of Bo was reduced. Guinea pig serum complement was also activated by Bo. At concentrations between 0.05 and 0.5 m M slightly more than 50% of the total activity were consumed, in 30 min at 30°C.
Dependence on Divalent Cations. In the presence of EDTA (0.025 M) or EGTA (0.025 M) Bo did not affect at all the overall complement activity of human serum during 30 min at 30°C. By recalcification, the blocking effect of the chelating agents could be overcome. Mg 2~ ions were not sufficient to correct the defect, for serum supplied with 0.01 M Mg 2÷ -EGTA did not lose significant complement activity on incubation with Bo. Under the same conditions zymosan led to a consumption of more than 50% of the whole complement activity (Table 1).
34
X. Li and W. Vogt % 100
x X
50
x
I
I
I
•0 0 5
.01
.025
i
.05
I
I
I
.1
.25
.5
Figure I Effect of Bo on total complement activity of human serum. Abscissa: Concentration of Bo in serum (raM). Ordinate: Complement consumption (%) after30 rain incubation at 30°C. (X) Individual determinations in sera from three different donors, (.) mean values.
Differentiation Between Bo and Endotoxic Lipopolysaccharides. The source of Bo, bagasse, might be contaminated with endotoxin-producing bacteria. To ascertain whether or not the Bo preparations were anticomplementary as a result of containing lipopolysaccharide (LPS) as an active principle, solutions of Bo were ultracentrifuged at I00,000 × g for 18 hr. A visible pellet did not form, the complement consuming aclivity in the upper four fifths of the centrifuged sample (per unit volume) was the same as that of the solution at the bottom. In contrast, after centrifugation of a LPS solution under the same conditions the complementconsuming activity had entirely accumulated in the bottom layer (Table 2). Thus, unlike the active principle in the LPS solution, the active constituent of Bo did not sediment during the run. Hence the activity of Bo does not reside in contaminating LPS. Effect of Bo on Individual Steps of Complement Activation Classical Pathway. Human serum was incubated with Bo for 30 rain at 30°C, then the residual activity of individual complement components was estimated by hemolytic assay. The results, presented in Table 3, show that Bo reduced C1, C4, and C2 drastically. More than 50% of the C3 and some C5 were also consumed. These results suggested that the effect of Bo
Table 1
Effect of selective Ca z* chelation by Mg-EGTA on complement consumption in human serum by Bo and, for comparison, by zymosan
Serum additions°
Complement activity (CH50/ml)
-
1275
Bo (0.1 raM)
Bo + Mg2~-EGTA(25 mM) Zymosan (5 mg/ml) + Mg2+-EGTA(25 mM) °The serum was incubated with the addi•ons indicated, for 30 rain at 30°C.
185
1099 505
35
Sugar Cane Polysaccharide Activates C o m p l e m e n t Pathway
Table 2
Comparative ultracentrifugation of Bo and LPS; localization of anticomplementary activity in the centrifugates a
Fraction added to serum
CH50/ml of incubated serum
Bo, upper layer Bo, lower layer LPS. upper layer LPS. lower layer
1333 400 370 1177 434
aThe upper four fifths of the centrifuge tubes were taken as upper layer, and the remaining fifth as lower layer. 50 p,l of the respective fraction (or saline, in the control) were added to 50 p,l serum, and incubated for 30 rain at 30°C.
consisted primarily in an activation of Cl(s). which then led to C4, C2. and C3 activation and consumption. Alternative P a t h w a y . To check activation of the alternative pathway the fate of factor B in human serum during incubation with Bo (0.05 mM, 30 min, 30°C) was observed, lmmunoelectrophoresis revealed hardly any cleavage of B, sometimes traces were cleaved in incubated control sera as well. Quantitative measurements by rocket electrophoresis showed up to 20% loss of intact factor B after incubation of serum with 0.05 mM Bo.
Mechanism of Activation D e p e n d e n c e on C l q . The effect of Bo m a d e it seem likely that it primarily activates C1, which in turn then acts on C4 and C2, and the question arose whether Bo acts directly on C l s to convert it to an active form. Therefore, C l q deficient human serum was prepared and was incubated with Bo in the usual way. No C4 consumption occurred. When purified C l q was added back to the deficient serum, Bo did considerably decrease the C4 content upon incubation (Table 4). The same result was obtained with the serum of a patient naturally deficient in C l q : Bo did not activate its content of C4 unless purified C l q was added (Table 4). Thus. apparently C l s is not directly activated by Bo. but the presence of C l q is essential, suggesting that Bo triggers the natural sequence of events in C1 activation. D e p e n d e n c e on I m m u n o g l o b u l i n s . Sera from two patients with severe hypogammaglobulinemia were available for testing the participation of immunoglobulins in complement activation by Bo. Serum 1 had 27 mg lgG/100 ml; serum 2 had 162 mg lgG/100 ml. The average IgG content of 9 "normal" sera was 1377 rag/100 ml = 118 SD. The complement titers of the two patients were slightly higher than the average of normal sera. Incubation of either deficient serum with Bo (0.1 raM, 30 min. 30°C) did not lead to complement
Table3
Consumption (Cleavage) of Single Complement Components in Human Serum After Incubation with Bo (30 rain 30°C)
Bo (mM)
C1
C4
C2
0.5 0.05 0.005
2.2
O.1 0.2 3.1
5.6
Residual activity (%) (23 C5 25
68
B 80
36
X. Li a n d W. Vogt
Table 4
Dependence of S e r u m Complement Consumption by Bo on C1q °
Serum specimen
Addition(s)
Activity after incubation C4 % change
Clq-def C1q-def Clq-def C1q-def
I I I I
Bo Clq C l q + Bo
952380 1052631 909090 555555
C lq-clef C1q-def Clq-def Clq-def
2 2 2 2
Bo Clq C l q + Bo
161290 200000 151515 83333
* I0 - 39 + 19 - 45
C2
% change
3394 3427 3128 2734
t I - 13
4453 4135 4793 1320
-
7
-- 72
aC I q-def I is serum made free from C l q by chromatography (see Materials and Methods). C I q-def 2 is serum from a patient naturally deficient in C I q. The sara were incubated with or without 0. I mM Bo and 2 p. I C I q, as indicated, for 30 rain at 30°C. Then the residual activity of C4 and C2 was assessed by immune hemolytic assay. Activities are given as C4H50 and C2H50. and in % of corresponding value without Bo.
consumption. When the sara were supplied with a purified human |gG preparation, their complement titer was considerably reduced by incubation with Bo (Table 5). Some complement activity was lost in the IgG-containing samples even without Bo; however, an additional effect of the polysaccharide was clearly present. The anticomplementary effect of lgG was presumably due to the presence of some aggregated globu]ins. Addition of monoclonal IgG did not render the hypogammaglobulinemic sara sensitive to Bo. Of 24 different sara from healthy donors or ambulant patients with various diseases and with no signs of deficiency, none was nonreactive. The mean CH50 was 1882 _+ 580, it was lowered after incubation with 0.1 mM Bo by 61 _+ 24% (mean _+ SD). Effect on Complement In Vivo Bo was dissolved in 0.15 M NaCI and injected in~avenously (iv) into guinea pigs (25 mg/kg). Control animals received saline only. Blood for complement assays was drawn by heart
Table 5
Dependence on lmmunoglobulins of Complement Activation by Bo o CHSO after
Experiment no. I I I
Specimen
Additions
incubation
% of corresponding sample without Bo
Serum Serum Serum Serum
I I I 1
Bo IgG IgG + Bo
2136 2083 1081 571
II II II II
Serum Serum Serum Serum
2 2 2 2
-Bo IgG IgG + Bo
1613 1653 1282 758
- 41
III Ill III
Serum 2 Serum 2 Serum 2
Bo monoclonal IgG + Bo
1852 1754 1754
-
2 47 ~ 2
5 5
°Hypogammaglobulin sara 1 and 2 were incubated with/without Bo (0.1 raM), human lgG ( 1.3 mg/ml), and monoclonal |gG (1.2 mg/ml) for 30 rain at 30°C, as indicated.
37
Sugar Cane Polysaccharide Activates Complement Pathway
100 ,,
Control
Bo 50
I
30
t
60
I
90
I
120
min
Figure 2 Effect of 25/mg/kg Bo, injected iv into guinea pigs, on complement in vivo. Blood for complement assays was taken by heart puncture, at the times indicated in abscissa. Each value represents the mean of 2--4 samples. Control injection: saline.
puncture, before injection and 30, 60, and 120 min thereafter ( 2 - 3 punctures in each animal). As shown in Figure 2, there was a slight decrease of complement activity in the contTols (around 10%), whereas the complement level of the injected guinea pigs showed a sustained reduction of about 25%.
DISCUSSION Many polysaccharides are known to activate complement. By far most of them act via the alternative pathway, e.g. zymosan, inulin, starch, and agarose. Quite different is the effect of a polysaccharide from ant venom isolated and analyzed by Schultz et al. (1979). It interacts with C1q, which leads to activation of the CI complex, and, in consequence, of the classical complement pathway without involving immune complexes. The active moiety is an oligosaccharide (Dieminger et al., 1979). The sugar cane polysaccharide, Bo, differs from both types of complement activators mentioned above. It activates complement by the classical pathway, with immunoglobulins being essential for the activation. Although a clear demonstration of anti-Bo antibodies has so far not yet been successful it seems most likely that Bo acts by forming immune complexes that, in turn, combine with and activate CI. There is no complement activation by Bo without Clq, or without gamma globulins. Since normal, heterogeneous human IgG did, but monoclonal IgG did not, support complement activation by Bo, it is reasonable to assume that specific antibodies are involved in the effect of Bo. The concentration-effect relationship for Bo supports this assumption; it resembles typical antigen-antibody precipitation curves. It should be pointed out, however, that a direct demonstration of specific antibodies directed against Bo is as yet lacking, and their presence is, so far, an assumption. Attempts at precipitation of Bo with Ig in double immunodiffusion have failed; in preliminary experiments precipitates have been obtained upon addition of Bo to human serum in the presence of polyethyleneglycol. The alternative pathway seems not to be activated by Bo. Some B cleavage occurred occasionally in incubated control sera as well. Dextrans, polysaccharides used clinically as plasma substitutes, may cause adverse reactions due to the presence of anti-dextran antibodies in the patient's blood, resulting in complement
38
X. Li and W. Vogt
activation. This is, however, a rare event, and anti-dextran antibodies are found only in a small proportion of humans, notably in patients with gastrointestinal diseases (Palosuo and Milgrom, 1981). In contrast, the supposed antibodies against Bo seem to be ubiquitous. Of more than 20 serum specimen from patients and healthy volunteers none was nonreactive (except for the cases of general hypogammaglobulinemia). Since a widespread specific sensitization to sugar cane is highly unlikely, it may be that the antibodies involved are directed primarily against and induced by surface polysaccharide chains of microbial organisms, similar in structure to Bo. Some glucans have recently gained interest because of their immunostimulatory and/or tumor-suppressing effects. One of these, the glucan from yeast, has been found to activate complement by as yet undetermined pathways (Haendchen et al., 1981). Bo. also a glucan, has likewise been found to have immunostimulatory activity (Jin et al., 1981). Whether there is a causal connection between these effects and complement activation remains to be investigated. The authors wish to thank Dr. Ch. Galanos; Freiburg, for the lipopolysaccharide, Prof. N. Hilschmann, G6ttingen. for monoclonal lgG, Prof. M. Loos, Mainz, for a Clq preparation and Clq deficient serum, Prof. U. Kaboth, G6ttingen, Dr. H. Wismann. G6ttingen, and Dr. E. Bartlau, G6ttingen, for providing hypogammaglobulinemic sera
REFERENCES Dieminger L. Schultz DR, Arnold PI (1979) Activation of the classical complement pathway in human serum by a small oligosaccharide. J Immunol 123:2201. Ehrke MJ, Reino JM, Eppolito C, Mihich E (1980) The effect of a protein bound polysaccharide (PS-K) on immune responses against allogeneic and minor histocompatibility antigens. Int J Immunopharrnacol 2:184. Haendchen LC, Glovsky MM, DiLuzio NR, Alenty A, Ghehiere L (1981) Complement activation by glucan. Fed Proc 40:1151. Jin YF, Liang HZ, Cao CY, Wang ZW, Shu RS, Li XY (1981) Immunological activity of bagasse polysaccharides. Acta Pharmacol Sinica 2:269. Palosuo T, Milgrom F (1981) Appearance of dextrans and antidextran antibodies in human sera. Int Arch Allergy Appl Immunol 65:153. Pillemer L, Ecker EE (1941) Anticomplementary factor in fresh yeast. J Biol Chem 137:139. Schultz DR, Arnold PI, Wu M-C, Lo TM, Volanakis JE, Loos M (1979) Isolation and partial characterization of a polysaccharide in ant venom (Pseudomyrmex sp.) that activates the classical complement pathway. Mol Immunol 16:253. Usui S, Urano M, Koike S, Kobayashi Y (1976) Effect of PS-K, a protein polysaccharide, on pulmonary metastases of a C3H mouse squamous cell carcinoma. JNatl Cancerlnst 56:185. Vogt W, Hinsch B, Schmidt G, von Zabem I (1979) Multiple effects of a diamidine (propamidine) on complement activation. Immunology 36:131. Zhu GP. Bao QZ, Chou TC, Hsieh RY, Liu HI (1982) Studies of bagasse polysaccharide. I. Isolation and identification of bagasse polysaccharide Bo. Acta Biochim Biophys Sinica, in press.