A new method for the estimation of C3d

Journal of Immunological Methods, 120 (1989) 207-214

207

Elsevier JIM 05196

A new m e t h o d for the estimation of C3d Affinity clearance of C-determinant-bearing C3 molecules and fragments followed by estimation of C3d by ELISA Syed S. Asghar 1, Bart Schraag a, Antonius H.L. Backhaus a, Ina Z o r n 1, Gerardus T. Venneker 1 and Andr6 J. H a n n e m a 2 Department of Dermatology, Academisch Medisch Centrum, Universityof Amsterdam, Meibergdreef9, 1105 AZ Amsterdam, The Netherlands, and 2 Division of lmmunochemistry, CentralLaboratory of Red Cross Blood Transfusion Services, Plesmanlaan 125, Amsterdam, The Netherlands

(Received19 August1988, revisedreceived23 January1989, accepted30 January1989)

A method is described to quantitate human complement fragment C3d. Test samples were treated with a predetermined excess of anti-C3c-Sepharose beads in the presence of EDTA to remove all the C-determinant-bearing C3 molecules or fragments. C3d left in the supernatant was then estimated by ELISA. Using this method, C3d could be estimated accurately in normal plasma samples. A good correlation (r = 0.93) was observed between C3d values obtained by this method and values obtained by the widely used method of Perrin and coworkers. The average C3d plasma concentration was 2.8 mg/1 (SD = 0.7 mg/1, n = 21). The interassay coefficient of variation using a normal plasma pool (C3d 2.7 mg/1) was 8.3% and using normal plasma pools in which the C3d concentrations were raised to 10.3 and 17.4 mg/1 by the addition of aged normal serum the levels were 8.0 and 7.5% respectively. Intra-assay coefficients of variation with these samples were 4.6, 3.0 and 2.8%, respectively. 16 patients with renal dysfunction had C3d levels in the range of 4.3-10.0 mg/1 and 15 patients undergoing continued ambulant peritoneal dialysis had levels of 3.3-12.2 mg/1. The C3d content in peritoneal dialysate of patients undergoing dialysis varied from 9.3 to 383/tg/1. Key words: C3d estimation;ELISA;Renal dysfunction;Continuedambulantperitonealdialysis;Biologicalfluid

Introduction

Activation of the complement (C) system is observed in many pathological situations. Monitoting C activation on the basis of a fall in hemolytic titer does not usually give an accurate assessment of C activation. Immunochemical assays of

Correspondence to: S. Asghar,Departmentof Dermatology, Academisch Medisch Centrum, University of Amsterdam, Meibergdreef9, 1105 AZ Amsterdam,The Netherlands.

the C fragments C3a, C5a, C3d, C4a and factor Ba are believed to be relatively more sensitive (Perrin et al., 1975; Hugli and Chenoweth, 1980). When specific information is needed as to which pathway is activated, assays for the fragments of C4 and factor B may be helpful but a general assessment of C activation may be obtained by measuring C3a or C3d. C5a measurements do not provide such information since the C5a binds firmly to neutrophils (Chenoweth and Hugli, 1978; Webster et al., 1982) and is not found free in plasma until this capacity is exceeded. C3a and

0022-1759/89/$03.50 © 1989 ElsevierSciencePublishersB.V.(BiomedicalDivision)

208

C3d levels are better indicators of C activation. Since C3d has a relatively long half life (Charlesworth et al., 1974) and is present in the plasma in higher quantities than C3a, C3d levels in the plasma is generally regarded as the activation indicator of choice. Various methods have been used for the quantitation of C3d (Perrin et al., 1975; Brandslund et al., 1981; Bourke et al., 1982; Vergani et al., 1983; Bhakdi et al., 1984; Adelsberg, 1985; Mollnes, 1985; Holmskov-Nielsen et al., 1986). The most commonly used method is that described by Perrin et al. (1975). It is based on the estimation of C3d by radial immunodiffusion after separation from C3 molecules and fragments expressing the C epitope by precipitation with polyethylene glycol (PEG). This and other methods involving PEG precipitation suffer from the disadvantage that small amounts of C3, C3b and C3bi are found in PEG supernatants and in certain instances this influences the results (Hack et al., 1986). Most of the other methods are technically more demanding; for example, the estimation of C3d after separation from C3, C3b and C3bi by crossed electrophoresis (Teisberg, 1975) and by in situ precipitation in rocket electrophoresis with intermediate gel containing anti-C3c (Brandslund et al., 1981) or concanavalin A (Bhakdi et al., 1984). Holmskov-Nilsen et al. (1986) have developed a relatively simple method for C3d estimation by an ELISA technique using antisera against the neodeterminants of C3d. However, this method, estimates neo-determinants present not only on C3d but also on C3b. We have developed an assay involving the removal of C3, C3b and C3bi by a short incubation with anti-C3c-Sepharose in the presence of EDTA followed by an ELISA estimation of the residual C3d in the supernatant.

Materials and methods

Reagents and buffers Polyoxyethylene sorbitan mono-oleate (Tween 80), 3,3',5,5'-tetramethylbenzidine (TMB), bovine serum albumin (BSA) and horseradish peroxidase (type VI) were obtained from Sigma Chemical Co. (St. Louis, MO, U.S.A.), hydrogen peroxide from Merck (Darmstadt, F.R.G.), dimethylsulfoxide

(DMSO) from BDH (Poole, U.K.) and polyethylene glycol (PEG, M r 6000) from Koch-Light Laboratories (Colnbrook, U.K.). All other reagents used were of Analar quality. Carbonate buffer, pH 9.6, containing 14.7 mM Na2CO3, 35 mM NaHCO 3 and 3.0 mM NaN 3 was used for the coating step of the ELISA. Tween 80 (0.05%) in water was used as a washing solution. EDTA buffer contained 10 mM EDTA and 20 g/1 BSA in PBS, pH 7.4. EDTA-Tween buffer contained 0.05% Tween 80 in EDTA buffer, pH 7.4. A solution consisting of 10 ml of 1.1 M sodium acetate (pH adjusted to 5.5 with saturated citric acid), 1.67 ml of 0.6% TMB in DMSO, 0.01 ml of 30% H202 and 90 ml distilled water was used as substrate of horseradish peroxidase. Blood samples For preparing normal serum pool (NSP) blood was obtained from healthy donors, allowed to clot at room temperature for 1 h and centrifuged (800 x g, 10 min, room temperature). Equal volumes of 17 of the normal serum samples were pooled, aliquoted and stored at - 70 ° C. For preparing plasma, blood from healthy donors and patients was collected in EDTA-containing tubes (10 mM final concentration) and centrifuged immediately (10 min, 800 x g). For the preparation of a normal plasma pool (NPP), equal volumes of 17 plasma samples were pooled, aliquoted and frozen under liquid nitrogen and then immediately stored at - 70 ° C. Individual plasma samples obtained from normal individuals and patients were also quickly aliquoted, frozen and stored as described above. All the patients were from the Department of Nephrology of our Medical Center. 16 of them had renal dysfunction and 15 were undergoing continued ambulant peritoneal dialysis. Peritoneal dialysate samples from patients undergoing dialysis were collected in EDTA- and BSA-containing tubes (final concentration of EDTA 10 mM and BSA 2%) and centrifuged for 10 min at 800 x g. The supernatants were frozen and stored essentially as described above for the plasma samples. Aged normal serum (ANS) was prepared by incubating NSP at 37 ° C for 7 days in the presence of NaN 3 (0.02% w/v) and used as a C3d standard reference. Another standard refer-

209 ence serum with a known C3d content in terms of mg/1 was obtained commercially from Boehringer Mannheim (Mannheim, F.R.G.). The NPP contained 2.8 mg/1 C3d. A reference plasma having a higher C3d content was prepared by adding a small known volume of ANS (250 mg/1 C3d) to NPP to increase its C3d content to 10.3 mg/1. This reference plasma was referred to as the medium level control (MC). Another reference plasma of still higher C3d content was prepared by adding known volumes of ANS to NPP to increase its C3d content to 17.4 mg/1.' This reference plasma was referred to as the high C3d level control (HC).

Antisera and conjugate The IgG fraction of rabbit anti-human C3d and the Ig fraction of rabbit anti-human C3c were obtained from Dako, Copenhagen, Denmark. Horseradish peroxidase (HRP)-conjugated antisera were prepared according to the method of Nakane and Kawaoi (1974). 100 gl of 1% 2,4-dinitrofluorobenzene in ethanol were added to 1.0 ml of HRP solution (5.0 mg in 1.0 ml of 0.3 M sodium carbonate) and the tube containing the mixture was rotated at room temperature for 1 h. 1 ml of 80 mM sodium periodate was then added and the mixture rotated for 30 min at room temperature. It was then dialysed at 4°C overnight against 10 mM carbonate buffer, pH 9.5.5 mg of Ig or the IgG fraction of the antiserum in 1.0 ml of 10 mM carbonate buffer, pH 9.5, were then added to the dialysed peroxidase-aldehyde preparation and the tube was rotated at room temperature for 4 h. The contents of the tube were then dialysed overnight at 4°C against PBS containing 0.01% thiomersal and passed through a column of AcA 44 Ultragel (1.6 × 95 cm) in the same buffer. The absorbance at 280 and 403 nm was measured. Fractions containing conjugate were pooled aliquoted and stored at - 2 0 °C in the presence of 0.01% thiomersal. Anti-C3c-Sepharose was prepared by treating 10 mg Ig fraction protein of anti-C3c with each milliliter of CNBr-activated Sepharose gel in coupling buffer (0.1 M NaHCO 3 containing 0.5 M NaC1, pH 8.3) essentially according to the instructions of the manufacturers (Pharmacia Fine Chemicals, Uppsala, Sweden).

Specificity testing by immunoblotting The specificity of both the HRP-anti-C3d and the HRP-anti-C3c was tested by performing agarose gel electrophoresis of NSP and ANS, transferring the bands developed in agarose gel to a nitrocellulose membrane and observing the reactivity of the above conjugates with these bands. In brief, agarose gel electrophoresis of NSP and ANS was carded out using an SPE-II agarose electrophoresis kit essentially as recommended by the manufacturer (Beckman Instruments, Fullerton, CA, U.S.A.). The gels were then blotted onto nitrocellulose membranes (Bio-Rad, Richmond, CA, U.S.A.) which were incubated with the EDTA buffer (containing BSA, see above) for 60 min at room temperature with constant shaking. The membranes were then cut into strips as required and incubated with 1/250 diluted HRP-anti C3c or 1/100 HRP-anti-C3d with constant shaking at room temperature for 60 min. After several washes with EDTA buffer, the HRP substrate was added. When protein bands became dearly visible, the strips were washed with water.

C3d estimation by radial immunodiffusion technique C3d in plasma was determined according to the method described by Perrin et al. (1975). 100 gl of plasma were incubated with 100 gl PEG 6000 (22%, w / v in 10 mM EDTA in 0.1 M borate buffer, pH 8.3) for 3 h at 0 ° C. The mixtures were then centrifuged at 4 ° C for 30 min at 1300 × g. 4 gl samples of the supernatants were tested for the presence of C3d by RID. Plates were incubated for 3 days at room temperature, washed with saline and stained with Coomassie brilliant blue. ANS and a reference serum of known C3d content were used as standards.

C3d estimation by affinity clearance of C-determinant-bearing molecules followed by ELISA (a) Treatment of samples with anti-C3c-Sepharose. 10 #1 of plasma were diluted with diluent buffer to 500 gl. 25, 50 and 100 gl of this diluted sample were then incubated with a predetermined excess of anti-C3c-Sepharose (usually 60 gl suspension) in a final volume of 500 /~1 of diluent buffer. The tubes were incubated with constant rolling at room temperature for 1 h and then

210

centrifuged for 5 min at 3000 × g. 400 #1 of the supernatant were removed and added to 20/zl of a 1% solution of Tween 80 in PBS. After mixing, the samples were analyzed by the ELISA procedure. When plasma was found to contain too low or too high an amount of C3d, the plasma dilution was changed accordingly and the assay repeated. For the determination of C3d in the peritoneal dialysate of patients undergoing peritoneal dialysis, 100, 50 and 25 /zl of undiluted sample were treated with a predetermined excess of anti-C3cSepharose (usually 50 #1) in a final volume of 500 /zl of diluent buffer. The rest of the procedure was the same as described above for plasma. When dialysates were found to contain too little C3d to be measured at the above dilutions, measurements were repeated in a modified incubation mixture containing 400 #1 of dialysate and 50 /~1 of antiC3c-Sepharose in a final volume of 500 /~1 of diluent buffer. (b) ELISA assay. Microtiter plates (Greiner, Alphen a / d Rijn, The Netherlands, cat. no. 655061 gamma-rayed) coated with anti-C3d were prepared by incubating 150 /zl of IgG fraction (5 /~g/ml in carbonate buffer, p H 9.6) in each well for 16 h at 20°C. The wells were then washed vigorously with washing solution. They were then incubated with EDTA buffer for at least 30 min and again washed with washing solution. The plates were then ready for use. C3d was quantitatively measured by incubating triplicate samples of each of the dilutions of the incubation mixture in the anti-C3d-coated wells for 2 h at 37 o C. The wells were washed four times with washing solution and 100 /~1 of a 1 / 1 0 0 0 0 dilution of peroxidase-labelled anti-human C3d in EDTA-Tween buffer were added to each well. After an incubation period of 1 h at 37 o C, the plates were again washed four times with washing solution. Freshly prepared substrate (100 /zl) was then added to each well. The reaction was stopped after exactly 10 min by the addition of 100/~1 of 2 M H 2 S O 4 to all the wells. The plates were then read in an automated microtiter plate reader at a wavelength of 405 nm. Blanks consisted of wells having first antibody, labelled antibody and substrate but no test sample. Values were recorded as the mean of the three test samples minus the mean of three blanks.

Estimation of C-determinant-bearing molecules Assessment of the removal of C-determinantbearing molecules in N P P treated with differing volumes of anti-C3c-Sepharose suspension was performed as follows. 10/~1 of NPP were diluted with diluent buffer to 500 ~tl. 100 txl of the diluted samples were treated with differing volumes (0-100 /d, see Fig. 1) of anti-C3c-Sepharose in diluent buffer in a final volume of 500/~1. Following centrifugation, 400/~1 of the supernatant were removed and added to 2 0 / d of a 1% solution of Tween 80 in PBS. After mixing, the samples were ready for ELISA. Triplicate samples of 100/~1 of the supernatant were then incubated in anti-C3ccoated wells for 2 h at 3 7 ° C (coating was performed as described above except that the Ig fraction of anti-Hu-C3c was used instead of the IgG fraction of anti-Hu-C3d). The wells were then washed four times with washing solution and 100 #1 of a 1 / 1 6 0 0 0 dilution of peroxidase-labelled anti-Hu-C3c in EDTA. Tween buffer was added to each well. After an incubation period of 1 h at 37 o C, the plates were washed and treated with the peroxidase substrate and the values recorded as described above for the C3d estimation.

Results

Removal of C-determinant bearing molecules or fragments by affinity clearance A fixed volume of N P P was treated with increasing amounts of anti-C3c-Sepharose suspension under the conditions described in the materials and methods section and C-determinantbearing C3 molecules and C3d determined in the supernatant (Fig. 1). Increasing amounts of beads removed increasing amounts of C-determinants. Increasing the concentration of the beads also caused some decrease of molecules bearing D-determinants (probably due to the removal of C3, C3b and C3bi) but after a small initial decrease, a plateau was reached. The inability of increasing amounts of beads to decrease the C3d levels further suggested that the only D-determinant bearing molecules left in the supernatant were those which could not be removed by anti-C3c-Sepharose, namely C3d molecules. This also suggested that anti-C3c present on the beads was not cross-re-

211 1.4o

1.2t~

o

1.0-

0.8-

0.6- ~ .

~

=

.,

~,

8

0.4-

0.2-

0 20 40 60 80 100 200 anti C3c-sepharosa suspension/ml 1/250 diluted NPP

Fig. 1. Removal of C-determinant-bearing C3 molecules and fragments by treatment with anti-C3c-labeUed Sepharose. Experimental conditions are described in the text. (o o) C-determinant-bearing and (• •) D-determinantbearing moleculesin the supernatant following treatment with anti-C3c-Sepharose. (,x zx) C-determinant- and (A A) D-determinant-bearing molecules in the supernatant following treatment with another batch of anti-C3cSepharose. acting with C3d and that there was no non-specific adsorption of C3d molecules on the beads. The above experiment was then carried out with three plasma samples of known C3d content. These samples included NPP, MC, H C (see materials and methods section). Treatment with increasing amounts of anti-C3c-Sepharose beads caused an initial small decrease in D-determinant bearing molecules after which a plateau was reached with all three samples. When the C3d levels in the supernatants of these samples were calculated, they were found to be as expected. These results indicated the possibility of the development of a method for C3d estimation in biological fluids based on removal of C3, C3b, C3bi and C3c by anti-C3c-Sepharose treatment in the presence of E D T A followed by estimation of C3d in the supernatant.

with HRP-anti-C3d followed by substrate, the optical density obtained was as low as that of the blanks. Similarly, when wells were coated with anti-C3d and HRP-anti-C3c was used as the second antibody, the same results were obtained. This showed that there was total lack of cross-reactivity between anti-C3c and anti-C3d. Agarose gel electrophoresis of N S P and ANS followed by nitrocellulose m e m b r a n e blotting showed that anti-C3d did not react with C3c and anti-C3c did not react with C3d. Further evidence of the high specificity of the anti-C3d came from the fact that the C3d values in A N S with and without treatment with anti-C3c-Sepharose were essentially the same (Fig. 2). Similarly, after an initial removal of C3, C3c, C3b, C3bi by anti-C3c-Sepharose, the inability of increasing amounts of anti-C3c to decrease the level of D-determinants clearly suggested that the anti-C3c used in these experiments did not cross-react with C3d (see Fig. 1).

Construction of the standard curve A standard curve for the estimation of C3d was constructed using 1/2000 to 1/128000 dilutions of ANS (Fig. 2). The standard curves obtained with and without treatment of ANS dilutions with anti-C3c-Sepharose beads were superimposable. This confirmed the previous observations that there was no non-specific absorption of C3d to the Sepharose beads and that anti-C3c linked cova-

E 1.4-

• 11.2 1.00.80.60.40"2t dUuUon of aged serum pool

Specificity When wells were coated with anti-C3c, washed, incubated with PBS instead of sample and treated

Fig. 2. Standard curve for C3d-estimation. Experimentalconditions are described in the text. • •, after treatment with anti-C3c-Sepharose, o o, no treatment with anti-C3cSepharose.

212

lently to Sepharose beads did not cross-react with C3d. This also showed that coated anti-C3d and HRP-anti-C3d did not react with molecules bearing C-determinants present in ANS. Using a commercial standard serum of known C3d content, the concentration of C3d in ANS was found to be 250 mg/1. C3d values in a sample expressed as a percentage of ANS could be converted into mg/1 by multiplying with a conversion factor of 2.5. C3d in normal plasma and patient material Repeated determinations (n = 14) of C3d by the proposed ELISA method in NPP, MC and HC showed coefficients of variation of 8.3, 8.0 and 7.5%, respectively. The intra-assay coefficients of variation (n = 9) in the above samples were 4.6, 3.0 and 2.8%, respectively. The interassay coefficients of variation (n = 27) in the same samples measured by RID were 26, 16 and 14%, respectively and the intra-assay variations were 25, 6.5 and 5.5%, respectively. The average C3d concentration in EDTA plasma from 21 healthy donors was found to be 2.8 mg/1 with a standard deviation of 0.7 mg/1 (Fig. 3). The observed range

C3d (mgll) 12.0 C3d (pg~)

to

10.0

8.0

|

o

o

•

!

500

m 12;':;10-

I

O

% e

.~,o oeo

te o

4-

• o • o

ee •e

- ..?'~ :

20

• •

• •

0 C3d (mg/I) radial immunodiffusion

Fig. 4. Correlation between the proposed ELISA method and the most widely used method of Perrin et al. (1975). n = 52, y = 1 . 8 4 x +0.38; r = 0.93.

was 1.8-4.6 mg/1. Using the RID method, the average C3d concentration in the same samples was 4.7 _ 1.1 rag/1 (range 2.5-6.9 mg/1). The range of C3d levels in 15 patients undergoing continued ambulant peritoneal dialysis was 3.2-12 mg/1 as estimated by the ELISA method (Fig. 3). 16 patients with renal dysfunction had C3d levels within the range 4.3-10 mg/1. Comparison of the results of the two methods with normal (n = 21) as well as patient plasma (n = 31) gave the regression equation y = 1.84x + 0.38 (r = 0.93) (Fig. 4). The C3d in EDTA-dialysate of patients undergoing continued ambulant peritoneal dialysis was readily estimated (Fig. 3). Usually samples were diluted 5, 10 and 20 times (compared to plasma which was usually diluted 250, 500 and 1000 times) during anti-C3c-Sepharose treatment. Estimations were sometimes repeated with higher or lower dilutions depending

400

I

200-

3OO

6.0

....-4

f

I,

o

l.

4.0

O

2OO

L 1002.0

tO0

r

low

0.0

0 normal plasma

CAPD plasma

renal dysfunction

CAPD dialyzate

plasma

Fig. 3. C3d levels in normal plasma, patient plasma and in the dialysate of patients undergoing continued ambulant peritoneal dialysis. Experimental conditions are described in the text. Note that C3d can be easily estimated in samples having ng quantifies of C3d.

o i

b

4

8 12 time of incubation ( h o u r s )

16

Fig. 5. Kinetics of C3d generation in the serum following M g 2+ treatment. C3d in serum incubated with (© ©) 20 m M EGTA, p H 7.4 and (e I)) 5.0 m M M g 2+ in 20 m M E G T A , p H 7.4.

213

on the C3d content of the samples. There was wide variation in the C3d content of these samples and the range was 9.3-385 # g / l (mean value 101.1 #g/l).

Kinetics of C3d generation The kinetics of C3d generation in NSP when NSP was incubated with high concentrations of Mg 2+ is shown in Fig. 5. It is apparent that the method detects minor as well as major changes in C3d level, occurring as a result of activation of the complement system.

Discussion In many clinical situations with slow C activation, consumption of C components is difficult to assess by hemolytic and immunochemical techniques because of simultaneous overcompensatory synthesis of C components. For the direct demonstration of C activation in these and other situations new tests are being developed which are based on the measurement of products of C activation. These tests include assays of complexes (Hack et al., 1981; Meyes et al., 1984; Nilsson and B~ick, 1985; Asghar et al., 1987) as well as fragments (Perrin et al., 1975; Hugli and Chenoweth, 1980; Brandslund et al., 1981; Adelsberg, 1985; Mollnes, 1985; Holmskov-Nielson et al., 1986) which are formed during C activation. Assays of fragments have been particularly useful (Westaby et al., 1986; Hammerschmidt, 1986). Of the several fragments generated during C activation, C5a is quickly removed by C5a receptors present on various cell types (Chenoweth and Hugli, 1978; Webster et al., 1982) and virtually disappears from biological fluids. C3a is also found in rather low concentrations in blood and other biological fluids. In contrast, C3d persists in blood and other biological fluids at much higher concentrations and elevated levels may be interpreted as evidence of C activation. However, the existing methods for the estimation of C3d suffer from various disadvantages. The new method described here was much more sensitive than the most widely used method of Perrin and coworkers (1975). Anti-C3cSepharose beads removed all C3 molecules beating C-determinant from the samples and the super-

natants contained C3d without the contamination of other interfering molecules. In the method of Perrin and coworkers, on the other hand, the supernatants of PEG treated plasma or samples did not appear to be totally free of C3 molecules bearing C determinants. C3d levels obtained using the method of Perrin et al. were higher than the levels obtained here by a factor of two. However, this factor was abolished when an excess of antiC3c was added to the RID plates (as described by Hack et al. (1981)) suggesting that the higher C3d levels obtained using the method of Perrin were due to contamination of the PEG supernatants by C3 molecules bearing C-determinants. Interassay and intra-assay coefficients of variation in our assay were much smaller than those observed with the method of Perrin and coworkers, and were comparable to those reported for a radioimmunoassay of C3a (Hugh and Chenoweth, 1980; Hammerschmidt, 1986). Our method was able to measure C3d in the dialysate of patients undergoing continued peritoneal dialysis. In general, samples were diluted 5-20-fold during the assay. The levels in most of the dialysates were, however, below the detection limit of the method of Perrin and coworkers. The new method has also been successfully used in our laboratory to estimate C3d in the urine of patients with kidney disorders and in the cerebrospinal fluids of patients with diverse neurological diseases. It is hoped that the method could also be used to assess C activation in blister fluids of skin in various dermatological diseases. In short, the method of C3d estimation described here may have considerable clinical utility. It can be used to monitor C activation in a wide variety of biological fluids and to study the role of C activation in diseases in general. It may also be useful in the screening of biomaterials and implantable devices for their ability to activate C in vivo and in vitro.

Acknowledgements The authors are indebted to Dr. L. Arisz and Dr. R.T. Krediet, Department of Internal Medicine, for providing us with the patient material

214 used in this study and to Mrs. M. Van Meegen for skilful technical assistance. We are grateful to the Chanfleury van IJsselstein Stichting for subsidizing the cost of the preparation of this manuscript.

References Adelsberg, B.R. (1985) A new and rapid quantitative assay for complement activation: An ELISA for C3d. Diagn. Immunol. 3, 187. Asghar, S.S., Barendsen, C.Y. and Van der Helm, H.J. (1987) Reinvestigations into the formation and assay of C3bBbP complexes. Clin. Claim. Acta 165, 243. Bhakdi, S., Roth, M. and Ntirnberger, W. (1984) A simple method for quantitative measurement of C3d in human plasma. J. Immunol. Methods 74, 79. Bourke, B.E., Moss, I.K. and Maine, R.N. (1982) Measurement of the complement C3 breakdown product C3d by rocket immunoelectrophoresis. J. Immunol. Methods 74, 79. Brandslund, I., Siersted, H.C., Svehag, S.E. and Teisner, B. (1981) Double-decker rocket immunoelectrophoresis for direct quantitation of complement C3 splitproducts with C3d specificities. J. Immunol. Methods 44, 63. Brandslund, I., Peters, N.D., Ejstrup, L., Teisner, B. and Rasmussen, G.G. (1983) Steroids and complement activation in rheumatoid arthritis. Lancet ii, 346. Charlesworth, J.A., Williams, D.G., Sherington, E., Lachmann, P.J. and Peters, D.K. (1974) Metabolic studies of the third component of complement and the glycine rich beta glycoprotein in patients with hypocomplementemia. J. Clin. Invest. 53, 1578. Chenoweth, D.E. and Hugli, T.E. (1978) Demonstration of specific C5a receptor on intact human polymorphonuclear leucocytes. Proc. Natl. Acad. Sci. U.S.A. 75, 3943. Hack, C.E., Hannema, A.J., Eerenberg-Belmer, A.J.M., Out, T.A. and Aalberse, R.C. (1981) A Cl-inhibitor-complex assay (INCA): A method to detect C1 activation in vitro and in vivo. J. Immunol. 127, 1450. Hack, C.E., Paardekooper, J. and Hannema, A.J. (1986) Influence of C3 level on the determination of C3d in plasma and synovial fluid by radial immunodiffusion. J. Immunol. Methods 86, 191.

Hammerschmidt, D.E. (1986) Clinical utility of complement anaphylatoxin assays. Complement 3, 166. Holmskov-Nielsen, H., Jensenius, J.C., Teisner, B. and Erb, K. (1986) Measurement of C3 conversion by ELISA estimation of n e o determinant on the C3d moiety. J. Immunol. Methods 94, 1. Hugli, T.E. and Chenoweth, E.D. (1980) Biologically active peptides of complement: techniques and significance of C3a and C4a measurements. In: R.M. Nakamura, W.A. Dito and E.S. Tucker III (Eds.), Clinical Laboratory Techniques for the 1980's. Allen R. Liss, New York, p. 443. Meyes, J.T., Schreiber, R.D. and Cooper, N.R. (1984) Development and application of an enzyme-linked immunosorbent assay for the quantitation of alternative complement pathway. J. Clin. Invest. 73, 160. Molnes, T.E. (1985) Quantification of the C3d sprit products of human complement by a senstitive enzyme-linked immunoassay. Scand. J. Immunol. 21, 607. Nakane, P.K. and Kawaoi, A. (1974) Peroxidase-labelled antibody. A new method of conjugation. J. Histochem. Cytochem. 22, 1084. Nilsson, T. and B~ick, O. (1985) Determination of Cls-C1 inhibitor complexes in plasma by means of an enzyme linked immunosorbent assay. Clin. Exp. Immunol. 60, 178. Perrin, L.H., Lambert, P.H. and Miescher, P.A. (1975) Complement breakdown products in plasma from patients with systemic lupus erythematosus and patients with membranoproliferative or other glomerulonephritis. J. Clin. Invest. 56, 165. Teisberg, P. (1975) In vivo activation of C3 revealed by crossed immunoelectrophoresis as a parameter of immunological activity in disease. Clin. Chim. Acta 62, 35. Vergani, D., Bevis, L.L., Nasaruddin, B.A., Mieli-Vergani, G. and Tee, D.E.H. (1983) Clinical application of a new nephlometeric technique to measure complement activation. J. Clin. Pathol. 36, 793. Webster, R.O., Larsen, G.L. and Henson, P.M. (1982) In vivo clearance and tissue distribution of C5a and C5a desarginine complement fragments in rabbits. J. Clin. Invest. 70, 1177. Westaby, S., Dawson, P., Turner, M.W. and Pride, R.B. (1985) Angiography and complement activation. Evidence for generation of C3a anaphylatoxin by intravaseular contrast agents. Cardiovasc. Res. 19, 85.