Influence of Different Serum Treatments on the Prozone Effect in SAB Testing
If our hypothesis regarding complement C1 is correct, not only EDTA but also other conditions adversely acting on C1 stability should eliminate the prozone. Accordingly, we tested sera containing IgG antibodies with a prozone effect after pretreatment with DTT, C1 inhibitor (C1INH), and after heat inactivation. Native serum and serum with EDTA added served as controls. Both, addition of DTT and heat inactivation abolished the prozone effect. C1INH was somewhat less effective but clearly abolished the prozone effect too (Fig. 3).
The results show that both destruction of the C1 molecule (heat, DTT) and cleavage of the C1r-C1s subunit from C1q (EDTA, C1INH) are effective in abolishing the prozone effect. In accordance with our hypothesis, the addition of EDTA and heat inactivation enhanced the reactivity of the IgM antibodies (data not shown).
Reconstitution With Complement or C1q
To further test our hypothesis that the C1 molecule prevents the binding of the secondary antibodies and causes the prozone effect, we tested two sera, which had lost their prozone effect by heat inactivation, after adding freshly drawn serum as a source of C1. The reconstitution of these heat-inactivated sera restored the prozone effect (Fig. 3), showing that the prozone effect observed in our patients depended on a heat labile factor, which is present in normal native serum. In contrast to fresh serum, addition of purified C1q did not restore the prozone effect (Fig. 3), suggesting that only C1 as a whole but not stripped C1q can cause the prozone.
Binding of C1q in SAB Testing
Native serum and serum after pretreatment with EDTA or DTT were tested in the single-antigen bead (SAB) with anti-C1q as secondary antibody. Native serum did not allow the detection of C1q, neither on beads with a prozone effect nor on beads without a prozone effect (Fig. 4). In the tests with EDTA serum, nearly all beads with a prozone effect were covered with C1q. Beads without a prozone had bound markedly less or no C1q. As expected, no C1q was detectable in tests with serum after DTT pretreatment, because DTT adversely affects C1q. The data indicate that the complete C1 molecule, as present in native serum, is not detected by anti-C1q secondary antibody. In EDTA serum, in contrast, where only the stripped C1q molecule is present, C1q was readily detected.
HLA-specific IgM antibodies bind C1q in direct proportion to their density on the bead, whereas HLA-specific IgG antibodies started to bind C1q only when a certain density of IgG on the bead surface was exceeded (Fig. 5).
Solid-phase assays facilitate detection of and improve sensitivity for various analytes but may be impaired by a prozone effect observed in the presence of high analyte concentration. Thus, high concentrations of an analyte lead to false results, lower than expected or even negative. A widely used procedure to overcome the prozone caused by high concentrations of an analyte is retesting after dilution (5). The prozone effect also has been reported for antibody detection with antigen-coated beads in the Luminex technique (2–4). However, it is not understood how a high titer of specific antibody causes the prozone in bead-based systems, because the capture antigen (i.e., HLA antigen) is already fixed to the solid phase (i.e., bead) when the serum containing the analyte (i.e., HLA antibodies) is added. The number of HLA-specific antibodies binding to the beads should increase proportionally to the amount of antibody present in the serum, reaching a plateau as soon as all HLA antigens on the bead are covered with antibodies. Closely spaced HLA-specific antibodies on the bead surface, in theory, could cause steric hindrance for the anti-IgG detection antibody, which can be overcome by dilution of the HLA antibodies in the serum. However, steric hindrance by IgG cannot explain how addition of EDTA overcomes the prozone effect. An additional factor may cause the prozone phenomenon, a factor that is sensitive to EDTA and that is able to block the access for the secondary antibodies to the HLA antibodies bound to the beads.
Therefore, we put forward the hypothesis that the complement system component C1 could be the cause of the prozone phenomenon in SAB testing as follows: C1 binds to two or more closely adjacent IgG molecules and blocks the Fc portion of the HLA-specific antibodies for the approaching anti-IgG detection antibody. Two adjacent IgG molecules, bound in a proper distance, are needed for C1 to bind to IgG. When serum contains HLA antibodies with a lower titer, not all of the HLA antigens on the SAB are covered by IgG, leaving spaces between the IgG molecules, which prevent C1 from binding, but allow the anti-IgG detection antibodies to catch their targets. In contrast, when serum contains HLA-specific antibodies in abundance, all HLA antigens on the beads may be covered with IgG. This situation allows C1 to bind to most or even to all the antibodies on the bead and, thereby, competitively prevents binding of the anti-IgG detection antibodies. A similar mechanism is conceivable for IgM antibodies as suggested by preliminary data from our tests with HLA-specific IgM antibodies.
C1 consists of one C1q subcomponent, Ca2+ ions, and two subcomponents C1r and C1s each. C1q was described as “bouquet of tulips,” consisting of six globular heads with stalk-like protein chains ending in a central fibril-like stem (7). Two subcomponents C1r and two subcomponents C1s are positioned between the stalks of C1q, forming a Ca2+-dependent tetramer. The heads of C1q bind closely to the hinge region of the immunoglobulin molecule (8), and it is possible that the C1r-C1s tetramer blocks that part of the Fc region, which the anti-IgG detection antibody binds to. Withdrawal of Ca2+ dissociates C1r and C1s, leaving a stripped C1q subcomponent, which still is able to bind between immunoglobulins but does no longer block the access of secondary antibodies.
Several serum treatments known to dissociate or to destroy C1 successfully abolished the prozone effect: EDTA chelates Ca2+ ions, dissociating the C1r-C1s subunit from C1q. C1INH is a regulator of activated C1 present in normal human plasma. C1INH binds to activated C1s and dissociates the C1r-C1s-tetramer from the C1q molecule (7), thereby exerting a similar effect on C1 as Ca2+ withdrawal. Both, EDTA or C1INH added to serum abolished the prozone effect. DTT or heat inactivation destroys several serum proteins including C1 (9). Both addition of DTT and heat inactivation of serum abolished the prozone effect. DTT, EDTA, and protein denaturation by heat are unspecific treatments, which also act on serum proteins other than C1. C1INH not only acts on C1 but also inactivates the coagulation factors FXIIa and FXIa, tissue plasminogen activator, and kallikrein (10). However, to the best of our knowledge, C1 is the only serum protein that is sensitive to all four treatments: it is inactivated by 56°C, has disulfide bonds making it sensitive to DTT, depends on the presence of Ca2+ ions, and is a substrate of the serine-protease-inhibitor C1INH. Further support for a possible role of C1 in the prozone effect was provided by experiments in which the prozone effect was abolished by heat inactivation but was easily restored by the addition of fresh serum from nonimmunized male donors. Experiments in the past (11) showed that sera lost their complement activity in the antiglobulin test after storage at room temperature for 72 hr. We also noticed that a serum lost its prozone effect after storage at room temperature for 3 days (data not shown). Storage at +4°C or −22°C markedly slowed the deterioration of the complement, and storage at −55°C retained all the complement activity. Further, repeated freezing and thawing of samples is known to have detrimental effects not only on the stability of complement but also on the stability of a few IgG and some IgM antibodies. Thus, the conditions of transport and storage (duration and temperature) may determine whether a prozone effect is present or no longer detectable, when a serum is finally tested in the laboratory.
To detect complement on the SAB, a phycoerythrin-labeled anti-C1q antibody was used. Only HLA-specific antibodies with a prozone effect bound marked amounts of C1q. However, C1q was only detectable when EDTA serum was tested. Tests with native serum showed negative results. The reason for this apparent discrepancy again may lie in the structure of the C1 molecule. As mentioned earlier, in native serum, C1q forms a complex with C1r and C1s. In EDTA serum, the withdrawal of Ca2+ dissociates the C1 complex, leaving a stripped C1q molecule. So far, the epitopes of C1q to which anti-C1q binds to are not yet defined. We propose that the C1r-C1s tetramer may not only block the anti-IgG detection antibody but also block the anti-C1q antibody, preventing the detection of C1q. Thus, in native serum, C1 may be bound by HLA-specific antibodies with a prozone effect but cannot be detected by the anti-C1q used in this study. Further studies are needed to elucidate whether different anti-C1q antisera are able to detect native C1 on SAB.
HLA-specific IgG antibodies in EDTA serum showed different C1q-binding characteristics compared with IgM antibodies. IgM antibodies bound C1q proportional to their density on the beads (i.e., proportional to their MFI). In contrast, only IgG antibodies showing the prozone effect with high MFI bound C1q to the beads. These observations fit to our hypothesis that a high density of IgG on the bead surface is required to bind C1, and they further support our hypothesis that C1 causes the prozone effect in high-titer antibody sera.
Recently, data were published (4,6,12) stating that HLA-specific IgM antibodies may be responsible for the prozone effect, because addition of DTT (which destroys IgM molecules) abolished the effect. IgM antibodies competing with IgG for the HLA binding sites on the beads are a possible mechanism for the prozone effect. However, in the sera presented in this study, not only DTT but also addition of EDTA abolished the prozone effect. The structure of IgM antibodies does not depend on the presence of Ca2+, and EDTA does not destroy IgM (13). Furthermore, nearly all IgG and most IgM molecules are resistant to heat inactivation for 30 min at 56°C (14), but heat inactivation abolished the prozone effect. After heat inactivation, the prozone effect was restored by the addition of serum containing no HLA-specific IgM antibodies. Therefore, we are convinced that in the sera presented in this study, IgM antibodies did not play a role in the prozone effect.
In conclusion, the results of this study confirmed our hypothesis of the role of C1 in the prozone effect in SAB testing. First, the prozone effect was caused by a heat labile factor and was restored by the addition of normal serum containing no HLA-specific antibodies. Second, four different serum treatments abolished the prozone effect, all of them capable to dissociate or to destroy the complement component C1. Third, only high-titer IgG antibodies with a prozone effect were found to bind marked amounts of C1q.
In our opinion, the C1r-C1s tetramer of the native C1 molecule inhibits the binding of the anti-IgG detection antibody leading to the prozone effect. This view was supported by experiments, in which C1q was found on SAB but did not block the access of either anti-IgG or anti-C1q detection antibodies. Further, in contrast to the addition of fresh serum, addition of purified C1q to heat-inactivated serum did not restore the prozone effect but still allowed the detection of all specificities with a prozone effect.
Anyway, whether or not C1 alone or IgM, or both, cause the prozone effect, our data show that the use of EDTA plasma or addition of EDTA to serum is highly efficient in overcoming the prozone effect. This is in contrast to the manufacturers' recommendations to use blood collected without anticoagulant, stating that plasma may lead to lower MFI values as described by Norris et al. (15). When we compared the reactivity of antibodies not showing any prozone effect, sera with EDTA added gave comparable results to the native sera. Most notably, we never saw false-negative results because of addition of EDTA.
The prozone effect of high-titer HLA antibodies forces laboratories to test sera undiluted and diluted. The results of this study suggest that testing EDTA plasma instead of serum might avoid double work load and double costs. As long as the complement-dependent cytotoxicity (CDC) is performed routinely in pretransplant diagnostics, it is, however, not possible to change from serum to plasma samples. Requesting an additional EDTA sample may also not be suitable. Addition of EDTA to the patients' sera before SAB testing could be a practicable alternative.
Two benefits for laboratories are obvious: first, addition of EDTA may give more reliable results, and second, it will save costs, because retesting using diluted serum is not necessary. We hope that our report, based on few patient samples only, will encourage other diagnostic laboratories to test this easy approach, so that a larger number of samples tested in this way will lead to a general use of EDTA addition in HLA antibody specification by SAB testing.
MATERIALS AND METHODS
Serum samples were obtained from patients awaiting kidney transplantation at the University Hospital of Tübingen. According to the guidelines of Eurotransplant, sera of these patients are regularly tested with the CDC test for the presence of HLA antibodies. Sera showing a high percentage of panel reactive antibodies in CDC were selected for testing with SABs for further specification of antibodies directed to HLA class I and class II antigens. In addition, sera of highly HLA-immunized patients awaiting stem-cell transplantation were included. If results raised suspicion of a prozone effect, sera were retested at 1:10 dilution. In this study, a prozone effect was defined as a 2-fold increase of the MFI after 1:10 dilution (or other serum treatment). These investigations were initiated to optimize the antibody differentiation by SAB. Therefore, according to the ethics committee of the Medical Faculty, no ethical conflicts are concerned.
SAB testing was performed with Lifecodes Luminex single antigen class I and class II kits (Gen-Probe Transplant Diagnostics, Stamford, CT) or LABScreen HLA class I SAB (One Lambda Inc., Canoga/Los Angeles, CA). LABScreen Singles (One Lambda Inc.) were used for testing different serum treatments of sera containing only few HLA antibody specificities showing the prozone effect. The tests were performed according to the manufacturers' instructions. Evaluation was carried out with the software provided. The results were expressed as adjusted MFI. Tests for IgM antibodies were performed using LifeMatch anti-IgM secondary antibody (Gen-Probe Transplant Diagnostics). Testing for C1q was done with LABScreen SAB and C1qScreen (both from One Lambda Inc.), containing purified C1q and phycoerythrin-conjugated anti-C1q.
The following serum treatments were used to investigate the prozone effect in SAB testing:
- 1:10 dilution in NaCl (0.9%)
- Addition of EDTA: 5 μL EDTA solution (6%)+95 μL serum
- Addition of an equal volume of 0.01 M DTT, 30 min at 37°C
- Heat inactivation: 30 min at 56°C
- Addition of C1INH: 10 μL C1INH (50 U/mL)+40 μL serum
- Addition of an equal volume of freshly drawn AB serum from male, nonimmunized blood donors as a source of complement
The authors thank Professor Gertrud Maria Hänsch, Institute for Immunology, University of Heidelberg, for the valuable information on complement. C1NH (Berinert) was generously provided by Dr. Rosenthal from CSL Behring. The C1q kit was a generous gift from One Lambda in cooperation with BmT, Meerbusch-Osterath. Finally, we are indebted to Hans-Georg Rammensee for critically reviewing the manuscript.
1. Tait BD. Solid phase assays for HLA antibody detection in clinical transplantation. Curr Opin Immunol
2009; 21: 573.
2. Stenzel A, Bringmann G, Koch K, et al. Prozone phenomenon of anti-HLA-A2 in Lifecodes Luminex single antigen class I, Tepnel (LSA Class I). Transfus Med Hemotherapy
2009; 36(suppl 1): 24.
3. Lowe D, Hathaway M, Briggs D. The high-dose hook effect in the detection and monitoring of HLA specific antibody by Luminex assay. Int J Immunogenetics
2007; 34: 288.
4. Kosmoliaptsis V, Bradley JA, Peacock S, et al. Detection of immunoglobulin G human leukocyte antigen-specific alloantibodies in renal transplant patients using single-antigen-beads is compromised by the presence of immunoglobulin M human leukocyte antigen-specific alloantibodies. Transplantation
2009; 87: 813.
5. Butch AW. Dilution protocols for detection of hook effects/prozone phenomenon. Clin Chem
2000; 46: 1719.
6. Kosmoliaptsis V, O'Rourke C, Bradley JA, et al. Improved Luminex-based human leukocyte antigen-specific antibody screening using dithiothreitol-treated sera. Hum Immunol
2010; 71: 45.
7. Loos M, Trinder PKE, Kaul M. Factors of the classical pathway. In: Rother K, Till GO, Hänsch GM, eds. The complement
system. Berlin, Heidelberg, New York, Springer 1998, p 1.
8. Gaboriaud C, Thielens NM, Gregory LA, et al. Structure and activation of the C1 complex of complement
: Unraveling the puzzle. Trends Immunol
2004; 25: 368.
9. Horowitz PM, Wasylewski Z, Kolb WP. Fluorometric detection of low temperature thermal transition in the C1q component of human complement
. Biochem Biophys Res Commun
1980; 96: 382.
10. Davies AE III, Mejia P, Lu F. Biological activities of C1 inhibitor. Mol Immunol
2008; 45: 4057.
11. Garratty G. The effects of storage and heparin on the activity of serum complement
, with particular reference to the detection of blood group antibodies. Am J Clin Pathol
1970; 54: 531.
12. Zachary AA, Lucas DP, Detrick B, et al. Naturally occurring interference in Luminex assays for HLA-specific antibodies: Characteristics and resolution. Hum Immunol
2009; 70: 496.
13. Arnold JN, Wormald MR, Suter DM, et al. Human serum IgM glycosylation: Identification of glycoforms that can bind to mannan-binding lectin. J Biol Chem
2005; 280: 29080.
14. Houssein I, Wilcox H, Barron J. Effect of heat treatment on results for biochemical analysis of plasma and serum. Clin Chem
1985; 31: 2028.
15. Norris PJ, Lee JH, Carrick DM, et al. Long-term in vitro reactivity for human leukocyte antigen antibodies and comparison of detection using serum versus plasma. Transfusion
2009; 49: 243.
Keywords:© 2011 Lippincott Williams & Wilkins, Inc.
Single-antigen bead assay; HLA antibodies; Prozone effect; Complement