Complement activation induced by air bubbles in rabbit and human sera was studied by measuring the generation of anaphylatoxin des-Arg-C5a. des-Arg-CSa ...
Variability over time of complement activation induced by air bubbles in human and rabbit sera KBiRE BERGH, ASTRID HJELDE, OLE-JAN IVERSEN, AND ALF 0. BRUBAKK Section for Extreme Work Environments and Department of Microbiology, Sintef-Unimed; and Department of Biomedical Engineering, University of Trondheim, N- 7006 Trondheim, Norway BERGH,
KARE,ASTRIDHJELDE,OLE-JANIVERSEN,AND ALF
0. BRUBAKK.Variability ouer time of complement activation induced by air bubbles in human and rabbit sera. J. Appl. Physiol. 74(4): 1811-1815, 1993.-Complement activation induced by
air bubblesin rabbit and human serawas studiedby measuring the generationof anaphylatoxin des-Arg-C5a.des-Arg-CSawas quantified by sandwichenzyme-linked immunosorbent assays basedon neoepitope-specificanti-des-Arg-C5a monoclonalantibodies. Air bubbles were continuously introduced to serum via a calibrated microflowmeter, and the serumwas incubated at 37°C for 30 min. Air bubbles clearly generated increased amounts of des-Arg-CSa compared with corresponding levels in control serum,and a dose-dependenteffect was also noted. Strong positive correlations between des-Arg-C5a concentrations in control sera and sera incubated with air bubbles at a flow of 0.5 ml/min were found. To study variation over time, serum was obtained at regular intervals from six rabbits and from six healthy humans during 66- and 196-day periods, respectively. A pronounced intraindividual variability over time was thus observed. The reason for the large variability is at present unknown. We conclude that the sensitivity of complement to activation by air bubblesis not an inherent, static feature of the complementsystem of an individual. Therefore single-point analysis of complement activation by air bubblesappears to be an inappropriate parameter by which to differentiate a “sensitive” or “insensitive” complement system between individuals. des-Arg-C5a; decompressionsickness
THE MECHANISMS of decompression
sickness (DCS) are still not completely understood. Also, the large variability in susceptibility to DCS both between individuals and in an individual undergoing repetitive diving (i.e., “acclimatization”) has not been plausibly explained. During decompression, air bubbles may be demonstrated intravascularly by ultrasonic monitoring. The prognostic significance of this phenomenon, however, is doubtful because air bubbles frequently occur without symptoms of DCS (1). Interactions at blood-bubble interfaces may result in several hematologic alterations. It has been shown that air bubbles become coated with proteins in plasma (lo), and this protein-coated bubble has been suggested to represent a foreign surface that may activate plasma complement along the alternative pathway (7,13). Activation of complement leads to an array of cellular and humoral reactions. The release of anaphylatoxins C3a and, in particular, C5a mediates a series of inflammatory responses
(6). Ward et al. (14) recently postulated that activation of plasma complement may be causally related to DCS and provided evidence that decomplementation of rabbits was associated with lower DCS incidence. By using granulocyte aggregometry as an index of anaphylatoxin (in particular C5a) activity in rabbit plasma and by measuring des-Arg-C5a in human plasma, Ward et al. were led to the proposal that individual variation in DCS susceptibility may be due to whether the complement system of an individual is capable of being activated by air bubbles or zymosan. According to Ward et al., both humans and rabbits could be divided into “sensitive” and “insensitive” subjects on the basis of sensitivity of their complement system to activation by air bubbles or zymosan (12, 13). By using another aspect of complement activation, i.e., erythrocyte-bound C3d, Zhang et al. (16) found that baseline sensitivity of air bubble-induced complement activation appeared to be normally distributed. We have produced two very sensitive sandwich enzyme-linked immunosorbent assays (ELISAs) based on monoclonal antibodies (MAbs) against rabbit des- ArgC5a (3,5) and human des-Arg-C5a (4). Because of reactivity toward a neoepitope on des-Arg-C5a, i.e., an epitope that is concealed in the native complement component C5 and exposed on the anaphylatoxin fragment only after the enzymatic cleavage during complement activation, des-Arg-C5a may be directly quantified in plasma or serum in the presence of native C5. This assay is far more sensitive than the commercially available radioimmunoassay (RIA) manufactured by Amersham and, more importantly, it obviates the need for precipitation of native C5, which is a prerequisite in the RIA method because the polyclonal antibody employed does not discriminate between des-Arg-C5a and native C5 (11). The aim of the present study was to investigate whether the activation of complement by air bubbles in rabbits and humans is an inherent and static feature of the complement system of an individual. MATERIALS AND METHODS
Specimens. Blood was obtained from each of six healthy volunteers (two females and four males) by venipuncture six times at regular intervals during a period of 196 days. During this period, no volunteer displayed any clinical signs of infection. Similarly, blood was obtained from the central ear artery of six male chinchilla rabbits (anesthetized with fluanisone/fentanyl) six times at regu-
0161-7567193 $2.00 Copyright 0 1993 the American Physiological Society
1811
1812
VARIABILITY
IN
COMPLEMENT
ACTIVATION
1. Effect of air bubbles in rabbit and human sera on generation of anaphylatoxin des-Arg-C5a TABLE
des-ArgGa, Stimulus
Rabbit serum
Control
216.9k17.4
0.5 ml/min 14 ml/min
300.31-28.9
690.0k66.4
rig/ml Human
serum
40.6t2.0 57.5t-2.5 81.5-c-4.4
Values are means 4 SE of 36 analyses of serum incubated at 37T for 30 min in absence (control) or in presence of calibrated air bubbles at a flow of 0.5 or 1.0 rnl/min.
lar intervals during a period of 66 days. The animals were housed at the Animal Care Facilities of University Hospital, Trondheim and were regularly examined by a veterinarian. No signs of illness were detected during this period. Blood was drawn into sterile, siliconized tubes (Becton-Dickinson, Meylan Cedex, France) and allowed to clot at 4°C. After clotting, serum was centrifuged at 600g for 10 min and frozen at -3OOC for subsequent experiments. Serum samples were sequentially subjected to complement activation by air bubbles within 2 wk of sampling. Experimental design. Sera were thawed and kept at O*C until start of the experiment. One milliliter of serum was put into a sterile siliconized tube (Be&on-Dickinson), which was placed in a water bath at 37OC while calibrated air bubbles (see below) were bubbled through the serum. A triplicate serum sample incubated at 37OC in the absence of gas bubbles served as control. After 30 min, any further complement activation was terminated by addition of 1 ml of 20 mM EDTA in 0.9% saline. For practical reasons, experiments in which air bubbles were bubbled through serum at two flow rates were not replicated. Before this study, however, we have performed experiments in which five different human sera were subjected to air bubbles in triplicate. Maximum deviation thus observed in single sera from the mean of triplicate experiments was 5.0%. Air bubbles (average size -2 mm) were introduced into serum by use of a calibrated microflowmeter (model 1780-1839, Gilmont Instruments, Barrington, IL). A gas flow of 03 or 1.0 ml/min corresponded to 220 or 490 bubbles/min, respectively. A gas flow >l.O mlfmin regularly produced foaming and was therefore omitted. Quantification of des-Arg-C5a. Quantification of rabbit and human des-Arg-C5a was performed by sandwich ELISAs using neoepitope-specific MAbs, described in detail by Bergh and Iversen (3,4). Briefly, microwells of Immunoplate Maxisorp (Nunc, Roskilde, Denmark) were coated with MAb 5B2C5 (rabbit C5a assay) or MAb 4A2ElOE2 (human C5a assay) in 0.1 M carbonate buffer, pH 9.6,. by overnight incubation at 20°C. Unoccupied binding sites were blocked by incubating the wells with 1% skimmed milk in phosphate-buffered saline (PBS) pH 7.2 for 30 min. Rabbit and human serum or plasma in triplicate diluted in PBS containing 0.05% Tween-20 (Sigma Chemical, St. Louis, MO) (PBS-T) were then incubated for 1.5 h, followed by washing with PBS-T. Bio-
BY
AIR BUBBLES
tinylated MAb 2BlA2 (rabbit C5a assay) or MAb 3G3C4 (human C5a assay) in PBS-T was then incubated for 1 h, followed by three washes with PBS-T. Peroxidase-conjugated streptavidin was then incubated for 1 h. After washing, substrate o-phenylenediamine (OPD) and H,O, were then added. The enzymatic reaction was terminated by adding 2 M H,SO,, and optical density (OD) at 492 nm was read. The concentration of des-Arg-C5a in plasma or serum was determined by relating the OD to a standard curve obtained by employing known concentrations of des-Arg-C5a. The intra- and interassay coefficients of variation of the human des-Arg-C5a ELISA are 4.5 and 5.9%, respectively (4). RESULTS
Complement activation induced by air bubbles. Data were compiled from six sera analyzed six times each (control sera and sera incubated with air bubbles at a flow of 0.5 or 1.0 ml/min). As shown in Table 1, the presence of air bubbles in rabbit or human serum during incubation at 37OC had a strong impact on the generation of the anaphylatoxin des-Arg-C5a. The difference between sera incubated in the absence of air bubbles (controls) and sera incubated with air bubbles at a flow of 0.5 ml/min was statistically significant both for rabbit serum and human serum (P = 0.016 and P < 0.001, respectively, two-tailed Student’s t test). The difference in sera incubated with air bubbles between flows of 1.0 and 0.5 ml/ min was also statistically significant (rabbit serum: P < 0.0001, and human serum: P < 0.0001, respectively). Inter- and intraindividual variation. Mean measurement values of des-Arg-C5a for the six different times for all subjects in this study are presented in Fig. 1 and suggest that some degree of interindividual variation may be present. Individual data from both rabbits and humans clearly show a considerable degree of variation over time, re2000 1800 TA 5600 p1400 -,1200 2 1000 g 800 TJ 600 g 400 200 0 1
2
3
4
5 Rabbi;
NO.
El f-i T T 160 140
920 a00 $
80
g
60
z
40
t3 20 1
q
control
2
q
3 OS ml/min
4
n
1 .O mJ/min
5
6
Individual
No.
FIG. 1. Values of sera (means t SD) from 6 individual rabbits (A) and 6 humans (B) obtained 6 times and incubated in absence (control) OFpresence of air bubbles at a flow of 0.5 or 1.0 ml/mine
VARIABILITY
IN COMPLEMENT Rabbit
z 2000 &
ACTIVATION
1813
BY AIR BUBBLES
No. 1
3500
800
h m 1000 45 sf 500 0
)4 E &
s 600 a$jP ma
0
q
Control
q
m
0.5 ml/min
Time No.
1 mt/min
FIG. 2. fntraindividual variation over time of complement activation by air bubbles in rabbit 1. Serum samples were obtained at regular intervals during a period of 66 days.
fleeted in the large standard deviation from the mean value for each individual presented in Fig. 1. This intraindividual variation over time is illustrated for two representative cases in Figs. 2 and 3. Rabbit I, which may be considered as an overall high responder (Fig. lA), displays a very prominent variation in its susceptibility to complement activation at different times (Fig. 2). Similarly, individual! 2, who may be regarded as an overall intermediate responder (Fig. lB), responds very differently at single point analyses (Fig. 3). Baseline state and responsiveness of the complement system. Analyses of the concentrations of des-Arg-C5a in sera incubated with air bubbles indicate that baseline levels of sera (i.e., incubated at 37*@ in the absence of air bubbles) to some extent may predict levels of des-ArgC5a obtained after subjecting the sera to activation by air bubbles (Figs. l-3). The results of correlation analyses between baseline levels of des-Arg-C5a and levels obtained after incubation with air bubbles at 0.5 ml/min are shown in Fig. 4. Thus in both rabbit and human sera very strong positive correlations were found: 0.93 (P < 0.0001) and 0.90 (P < O.OOOl),respectively.
400
‘a $g 200
0 0
100
200
300
400
500 600 Control sera C5a des Arg (nghnl)
100 80
g t 60 $ p z 2 2 s 40 ) a $5 20
0 0
20
40
60
80
Control sera C5a des Arg (rig/ml)
DISCUSSION
In the present study it is clearly demonstrated that air bubbles are capable of activating complement. This was previously shown by Ward et al. (13) for human serum by measurement of anaphylatoxin des-Arg-C5a with a commercially available RIA. Anaphylatoxin immunoassays in which precipitation of native complement components are required, however, are hampered with some degree of uncertainty, mainly because the antibody employed shows cross-reactivity to both the anaphylatoxin fragIndividual
1 tZ3
=m # s
Control
3
2
q
0.5 ml/min
4 W
1 ml/min
5
No. 2
6 Time No.
FIG. 3. Intraindividual variation over time of complement activation by air bubbles in individual 2. Serum samples were obtained at regular intervals during a period of 196 days.
FIG. 4. Correlation between levels of des-Arg-C5a at baseline and after air bubbles at. 0.5 ml/min for rabbit sera (A) and human sera (B).
ment and the native complement component. A critical prerequisite in these assays, therefore, is the complete removal of the native complement component before the anaphylatoxin is assayed. This is usually performed by precipitation of high-molecular-weight proteins. Klos et al. (8) demonstrated two possible sources of error of these assays: the precipitating reagent may fail to remove native C5 completely, which may yield falsely high levels of des-Arg-C5a, and some des-Arg-C5a may be coprecipitated. We are thus led to suspect that the findings of Ward et al. (13) of C5a levels up to 4 mg/ml in zymosanactivated plasma, a figure far beyond the maximum concentration attainable (9), may be due to inherent methodological problems of precipitation-based assays. The method used for quantification of human des-Arg-C5a in the present report is uninfluenced by native C5 (4). In the human sera employed in this study, zymosan activation (5 mg/ml, 37OC) consistently yielded des-Arg-C5a levels t3 pglml. Ward et al. (12) also related complement activation by air bubbles in rabbit serum to the generation of
1814
VARIABILITY
IN COMPLEMENT
C5a by using a leukocyte aggregation assay. Here we report the direct measurement of des-Arg-CSa generated by incubation with air bubbles also in rabbit serum. This method is advantageous over bioassays for several reasons. First, it provides a specific measurement of the mediator in question (i.e., des-Arg-C5a) with a sensitivity superior to a bioassay (2). Second, the great variability often encountered in bioassays because of the variable leukocyte response to identical stimulus concentrations may represent a serious drawback in obtaining reproducible results. Ward et al. (l&14) were led to the conclusion that rabbits could be divided into sensitive or insensitive groups on the basis of whether their complement system could be activated by air bubbles or zymosan. Because the method employed by Ward et al. is based on leukocyte aggregation, the results obtained are founded both on the stimulus inflicted (i.e., activated complement) and the responsiveness of the leukocytes. In our opinion, one cannot conclude whether the complement system in plasma can be activated by air bubbles by measuring the responsiveness of leukocytes in a bioassay (in which activated complement is only one of several, often unknown, variables). Ward et al. also used zymosan as a stimulant to determine whether the complement system of an individual was sensitive or insensitive to activation. Again, because this distinction was based on leukocyte responsiveness, one cannot conclude whether or not the complement system is susceptible to activation by zymosan. To date, from analyses of several hundreds of rabbit plasma or sera, we have been unable to detect a single rabbit in which the generation of des-Arg-C5a was
