Hemolysis Via the Lectin Pathway As C4 and C2 in Complement

Requirement for the Alternative Pathway as Well as C4 and C2 in Complement-Dependent Hemolysis Via the Lectin Pathway

This information is current as Chusana Suankratay, Xiao-Hui Zhang, Yonghong Zhang, Thomas of September 29, 2021. F. Lint and Henry Gewurz J Immunol 1998; 160:3006-3013; ; http://www.jimmunol.org/content/160/6/3006 Downloaded from

References This article cites 47 articles, 21 of which you can access for free at: http://www.jimmunol.org/content/160/6/3006.full#ref-list-1

Why The JI? Submit online. http://www.jimmunol.org/

• Rapid Reviews! 30 days* from submission to initial decision

• No Triage! Every submission reviewed by practicing scientists

• Fast Publication! 4 weeks from acceptance to publication

*average by guest on September 29, 2021

Subscription Information about subscribing to The Journal of Immunology is online at: http://jimmunol.org/subscription Permissions Submit copyright permission requests at: http://www.aai.org/About/Publications/JI/copyright.html Email Alerts Receive free email-alerts when new articles cite this article. Sign up at: http://jimmunol.org/alerts

The Journal of Immunology is published twice each month by The American Association of Immunologists, Inc., 1451 Rockville Pike, Suite 650, Rockville, MD 20852 Copyright © 1998 by The American Association of Immunologists All rights reserved. Print ISSN: 0022-1767 Online ISSN: 1550-6606. Requirement for the Alternative Pathway as Well as C4 and C2 in Complement-Dependent Hemolysis Via the Lectin Pathway1

Chusana Suankratay, Xiao-Hui Zhang, Yonghong Zhang, Thomas F. Lint, and Henry Gewurz2

Mannan-binding lectin (MBL) is a C1q-like molecule opsonic for several micro-organisms. MBL can activate C4, C2, and later acting complement components in the presence of serine proteases similar to but distinct from C1r and C1s via the lectin pathway of complement activation. We report here that mannan-coated MBL-sensitized erythrocytes are lysed via the lectin pathway in human serum-Mg-EGTA. The surprising occurrence of MBL-initiated lysis in the absence of calcium contrasts with the calcium requirement for C1q-initiated activation of C4 and C2. C2 is required, and lysis is signiﬁcantly enhanced when indicator cells Downloaded from presensitized with C4 and then coated with mannan (EAC4-M) are used. The alternative pathway also is required, since lysis is lost when either factor D or factor B is removed and is restored upon reconstitution with the puriﬁed protein. Even though MBL is a C-type lectin, it is retained on mannan-coated erythrocytes in the absence of calcium. This contrasts with the absence of calcium-independent retention on mannan immobilized on polystyrene plates or beads, and helps explain the MBL-initiated hemolysis in Mg-EGTA. These investigations show that the alternative pathway as well as C4 and C2 of the classical pathway are http://www.jimmunol.org/ required for complement-dependent hemolysis via the lectin pathway and provide a method for assay of lectin pathway-mediated complement activity in human serum that should be useful in unraveling the molecular interactions of this pathway. The Journal of Immunology, 1998, 160: 3006–3013.

annan-binding lectin (MBL)3 was ﬁrst isolated from cosamine, mannose, fucose, and glucose (14). Human MBL has an rabbit liver (3) by its strong binding to yeast mannan apparent molecular mass of about 400 to 700 kDa, consisting of M and subsequently was found in the serum and liver of three to six identical subunits of about 96 kDa, each containing the rabbit (4), rat (5–7), human (8), bovine (9), chicken (10), three identical chains with a globular carbohydrate recognition do-

mouse (11), and pig (12). It is a C-type lectin (13) that can bind main and a collagen-like region that is stabilized by a cysteine-rich by guest on September 29, 2021 specifically to terminal nonreducing sugars, including N-acetylglu- NH2-terminal domain (13, 15, 16). It has been grouped with proteins of similar structure termed collectins (14, 17). MBL is an acute phase protein (18, 19) that is thought to be an Department of Immunology/Microbiology, Rush Medical College, Chicago, IL 60612 important constituent of the innate immune system (14, 17, 20, Received for publication October 7, 1997. Accepted for publication November 21). It plays a crucial role in host defense against certain pathogens 24, 1997. that contain ligands for MBL on their surfaces. It can function as The costs of publication of this article were defrayed in part by the payment of page an opsonin. Native and recombinant human MBL are able to di- charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. rectly bind to wild-type virulent Salmonella montevideo express- 1 Presented in part to the Annual Meetings of the Central Society for Clinical Re- ing a mannose-rich O-polysaccharide; this results in the attach- search in Chicago, IL, September 26, 1997 (1), the Infectious Diseases Society of ment, uptake, and killing of MBL-coated bacteria by phagocytes in America in San Francisco, September 14, 1997 (2), and IV International Workshop on the absence of serum (22). MBL deficiency, which is linked to C1 and Collectins, Mienz, Germany, October 5, 1997 (3). This work was supported by a Thai Royal Government Scholarship (to C.S.) and was presented (by C.S. and three allelic mutations in codons 52, 54, and 57 of the first exon of Y.Z.) in partial fulfillment of the requirements for the Ph.D. from Rush University. the MBL gene (23–25), is associated with a common opsonic de- H.G. holds the Thomas J. Coogan Chair in Immunology/Microbiology established by Marjorie Lindheimer Everett. fect that results in recurrent or persistent infection early in life (26, 27). It recently was shown that homozygous carriers of these vari- 2 Address correspondence and reprint requests to Dr. Henry Gewurz, Department of Immunology/Microbiology, Rush Medical College, Chicago, IL 60612. ant MBL alleles are at increased risk of HIV infection (28). 3 Abbreviations used in this paper: MBL, mannan-binding lectin; MASP-1 and MBL and C1q have a similar ultrastructural organization, even MASP-2, mannan-binding lectin-associated serine proteases structurally homologous though they do not share amino acid sequence homology (14–17, 2ϩ to C1r and C1s, respectively; GVB, gelatin-veronal-buffered saline; GVB , gelatin- 29, 30). MBL is able to activate the complement system via the veronal-buffered saline containing 2 mM CaCl2 and 0.5 mM MgCl2; GGVB, a mixture containing equal parts of 5% glucose and gelatin-veronal-buffered saline; classical pathway (31–36) and to bind to the surface receptor for 2ϩ GGVB , GGVB containing 2 mM CaCl2 and 0.5 mM MgCl2; EDTA-GVB, gelatin- C1q (30, 37, 38). However, the newly discovered serine proteases veronal-buffered saline containing 10 mM EDTA; NHS, normal human serum; VBS-T, veronal (5 mM)-buffered saline (pH 7.4) containing 0.05% Tween-20; EA, MASP-1 (35, 39) and MASP-2 (40), rather than the homologous antibody-sensitized sheep erythrocytes; E, sheep erythrocytes; E-M, mannan-coated C1r and C1s, were shown to associate with MBL and to be re- sheep erythrocytes; E-M-MBL, mannan-coated sheep erythrocytes sensitized with quired for MBL to activate C4, C2, and hence the rest of the mannan-binding lectin; EAC4, C4-coated indicator cells; EAC4-M, C4- and mannan- complement cascade; this series of interactions has been termed coated indicator cells; Tmax, time of maximal formation of enzymatically active C42 2ϩ sites on EA; CH50, titer of serum lysing 50% of EA in GVB via the classical the lectin pathway of complement activation. However, MBL also pathway; AH50, titer of serum lysing 50% of rabbit erythrocytes in Mg-EGTA via the has been reported to trigger the complement system through the alternative pathway; LH50, titer of serum lysing 50% of mannan-binding lectin-sensitized mannan-coated sheep erythrocytes in Mg-EGTA via the lectin pathway under alternative pathway. Thus, it was reported that MBL-mediated standardized conditions. deposition of C3 on mannose-rich Salmonella occurred via the

alternative pathway without classical pathway involvement (41), Sepharose 4B column (100 ml) that had been equilibrated with starting and that MASP is able to cleave C3 directly with subsequent ac- buffer. The column was washed with starting buffer, and the bound proteins tivation of the alternative pathway (42). were eluted with a buffer containing 50 mM Tris-HCl, 1 M NaCl, and 20 mM EDTA, pH 7.8. The eluate was brought to 50 mM CaCl2 and reapplied In the present paper we sought to further clarify the mechanism to a second, smaller (25 ml) but otherwise identical, mannan-Sepharose 4B by which the lectin pathway leads to complement activation in column, and the bound proteins were eluted in the same manner as in the human serum. We report here that the alternative pathway as well first column. as C4 and C2 are required for complement activation by the lectin Preparation of hemolytic intermediate cells pathway, and we present a new assay for quantifying lectin pathway activity in normal human serum. Ab-sensitized sheep erythrocytes (EA) (45), and erythrocytes (E) coated with mannan (E-M) and sensitized with MBL (E-M-MBL) were prepared as previously described (31, 46). Briefly, 0.5 ml of sheep E (1 ϫ 109 Materials and Methods cells/ml) were mixed with 0.5 ml of CrCl3 solution (0.5 mg/ml), 0.5 ml of Buffers mannan solution (200 ␮g/ml) was added, and the mixture was incubated with occasional shaking for 5 min at 25°C. The reaction was stopped by Gelatin-veronal-buffered saline consisting of 5 mM veronal (pH 7.4), 0.145 adding 1.5 ml of ice-cold GVB2ϩ, and the E-M were washed three times ϩ M NaCl, and 0.1% gelatin (GVB); GVB containing 2 mM CaCl2 and 0.5 ϫ 9 2 ϩ and resuspended to a final concentration of 1 10 cells/ml in GVB .An 2 ϩ mM MgCl2 (GVB ); GVB containing 0.5 mM MgCl2; GVB containing 2 2ϩ 2ϩ aliquot (0.1 ml) was added to 0.4 ml of MBL (1–4000 ng) in GVB , 0.5 mM CaCl2; a mixture of equal parts of GVB and 5% glucose to incubated with gentle shaking for 15 min at room temperature, washed, and ␮ 2ϩ make the ionic strength ( /2) equal to 0.075 NaCl (GGVB ); GVB con- resuspended to 1 ϫ 108 cells/ml in Mg-EGTA. taining 0.5 mM MgCl2, 100 mM ethylene glycoltetraacetic acid, and glu- C4-coated indicator cells (EAC4) were prepared as described by Borsos cose to make the ionic strength equal to 0.075 M NaCl (Mg-EGTA); and and Rapp (47). Briefly, EA were washed three times and resuspended to Downloaded from GVB containing 10 mM EDTA (EDTA-GVB) were prepared as previously 5 ϫ 108 cells/ml in GGVB2ϩ prewarmed to 30°C. An equal volume of described (31, 43). purified human C1 in GGVB2ϩ was added slowly with continuous shaking, and after 15-min incubation at 30°C, the mixture was washed, resuspended Reagents ϩ to the original volume in ice-cold GGVB2 , and cooled to 0°C in an ice- Saccharomyces cerevisiae mannan was purchased from Sigma Chemical water bath. An equal volume of NHS diluted 1⁄4 in ice-cold EDTA-GVB Co. (St. Louis, MO). Anti-MBL mAb was purchased from Statens Serum was added and incubated with shaking for 15 min at 0°C. The mixture was washed three times and incubated at 37°C in EDTA-GVB to remove C1 Institute (Copenhagen, Denmark). R-phycoerythrin-conjugated streptavi- http://www.jimmunol.org/ din for use in flow cytometry was purchased from Dako Co. (Glostrup, and decay any C2 incorporated into EAC42, and the resulting EAC4 sus- ϩ Denmark). pension was washed and resuspended to the original volume in GGVB2 . The Tmax was determined after addition of purified C1, as previously de- Sera and complement components scribed (48), to establish the optimal NHS concentration to use in preparation of EAC4 and to quantitate the amount of C4 on the cells. EAC4-M Aliquots of normal human serum (NHS) collected from healthy, normal were prepared by coating EAC4 with M as described above. EAC43 also donors and from patients congenitally lacking individual complement com- were prepared as previously described (45). ponents were stored at Ϫ70°C. Every complement-deficient serum used was shown to have Ͻ2% normal hemolytic activity, with normal levels Hemolytic assays restored upon addition of small amounts of the purified deficient component. Serum deficient in both C1q and factor D was prepared by column The CH and AH were determined as previously described (43, 45).

50 50 by guest on September 29, 2021 chromatography on Bio-Rex 70 (Bio-Rad, Hercules, CA) (44). Removal of Lectin pathway activity (LH50) was assayed by lysis of E-M-MBL in Mg- C1q was ascertained by lysis of Ab-sensitized sheep erythrocytes in the EGTA as follows. E-M-MBL (100 ␮l containing 1 ϫ 108 cells/ml in Mg- presence and absence of added C1q; depletion of factor D was documented EGTA) were incubated with 100 ␮l of test human serum in Mg-EGTA at by lysis of rabbit erythrocytes in Mg-EGTA in the presence and absence of 37°C for 1 h and centrifuged, and the OD414 of the supernatant was de- added factor D. Purified human C1 and C1q and guinea pig C2 were pre- termined. The percent lysis and the amount of serum generating 50% lysis pared as previously described (44, 45). Purified human C4, C2, and factors were determined in the usual way (43, 45); units are expressed as the D and B as well as factor B-depleted human serum were purchased from dilution of serum added in 0.1 ml yielding 50% hemolysis. Calbiochem (La Jolla, CA). Flow cytometry assay of MBL on sheep E ELISA for MBL E-M-MBL (100 ␮l containing 1 ϫ 108 cells/ml) were mixed with 100 ␮l MBL concentrations were assayed using a minor modification of previ- of 1/50 diluted biotinylated anti-human MBL and incubated at 4°C for 1 h. ously described sandwich ELISA methodology (27). Briefly, microtiter After washing three times, 100 ␮l of 1/5 diluted R-phycoerythrin-conju- plates were coated with monoclonal anti-MBL (3 ␮g/ml, 100 ␮l/well) in gated streptavidin solution was added, and incubation proceeded for1hat

coating buffer (0.03 M Na2CO3 and 0.02 M NaHCO3, pH 9.6) and incu- 4°C. The washed cells were resuspended to 1 ml and assayed for MBL bated overnight at 4°C. The microtiter plates were washed three times with using an Ortho Cytoron flow cytometer (Ortho Diagnostic Systems, Rari- veronal (5 mM)-buffered saline (pH 7.4) containing 0.05% Tween 20 tan, NJ) to quantitate orange fluorescence. The fluorescence of cells is (VBS-T) and blocked for2hatroom temperature by adding 200 ␮l VBS given as the mean channel intensity of 1 ϫ 104 cells as measured using containing 1% BSA to each well. Samples and MBL standards were loaded logarithmic amplification, except as noted. In this configuration, a 22-chan- into the wells in duplicate, incubated at room temperature for 1 h, and nel increase represents an approximate doubling of fluorescence. washed as before. Biotinylated monoclonal anti-MBL diluted in VBS (100 ␮l, 1 ␮g/ml) was added to each well, followed by incubation at room Results temperature for 1 h. The plates were washed, streptavidin-horseradish per- MBL-initiated hemolysis mediated via the lectin pathway in oxidase conjugate (1/1000 in VBS, 100 ␮l/well) was added, incubation was continued at room temperature for 1 h, and after additional washes, 100 ␮l human serum-Mg-EGTA substrate was added. Incubation was continued for 30 min at room tem- In experiments on complement activation initiated by MBL, we ␮ perature, the reaction was stopped by addition of 2 N HCl (100 l/well), observed that E-M sensitized with MBL were lysed when added to and the absorbance at 450 nm was read on a microplate reader. NHS in the presence of Mg-EGTA. Lysis occurred only with E-M Mannan-binding lectin sensitized with MBL; neither E-M in the absence of MBL nor MBL was prepared by sequential affinity column chromatography by minor Ab-sensitized E was lysed under these conditions. Lysis was di- modification of the method of Kawasaki et al. (8). Briefly, human serum rectly proportional to the amount of MBL and complement offered; obtained by recalcification of pooled human citrated plasma of known 1000 ng of MBL/108 E-M was optimal (Fig. 1). We determined the MBL concentration was brought to 20 mM CaCl2 and allowed to clot for serum dilution required for 50% lysis of MBL-sensitized E-M in 1 h at 37°C and overnight at 4°C. After removal of the clot, the serum was dialyzed extensively against starting buffer containing 50 mM Tris-HCl, 1 Mg-EGTA (LH50 units) in NHS from 12 healthy individuals. As shown in Figure 2, the mean LH of E-M-MBL was 9 Ϯ 8 U/ml M NaCl, 20 mM CaCl2, and 0.05% (w/v) NaN3 (pH 7.8). After centrifu- 50 gation at 10,000 ϫ g for 10 min, the supernatant was applied to a mannan- (range, 1–25 U/ml). This surprising occurrence of MBL-mediated 3008 ALTERNATIVE PATHWAY AND C4/C2 IN LECTIN PATHWAY HEMOLYSIS

FIGURE 1. Lysis of 0.1 ml E-M (1 ϫ 108/ml) presensitized with MBL during 1-h incubation with 0.1 ml of human serum-Mg-EGTA at 37°C. Lysis was quantiﬁed using increasing amounts of MBL in the presence of optimal (1/4 diluted) human serum in A and increasing amounts of serum in the presence of optimal (1000 ng) MBL in B. There was no lysis of E-M, EA, or EA-M that had not been presensitized with MBL upon addition to human serum-Mg-EGTA.

hemolysis in the absence of calcium contrasts with the calcium C2 requirement for MBL-initiated hemolysis in Mg-EGTA requirement for C1q-mediated hemolysis in the classical comple- We next examined the role of C2 in MBL-mediated hemolysis of ment pathway. However, these titers are significantly less than the EAC4-M, using sera from five different individuals with inborn Downloaded from ϭ titers of 315 CH50 U/ml for lysis of EA (SD 72; range, 160–400 deficiency of C2. As shown in Figure 5, virtually no lysis occurred ϭ U/ml) and 91 AH50 U/ml (SD 24; range, 59–140 U/ml) for lysis upon incubation of EAC4-M-MBL in any of these sera, while the of EAC43 when the classical and alternative pathways, respec- addition of purified human C2 (200 U/ml) restored lysis to normal tively, were assayed in the same serum samples using identical Ϯ or near-normal levels (62 23 LH50 U/ml). These results empha- reaction volumes and numbers of indicator cells. size that complement-dependent hemolysis of E-M-MBL in Mg-

C4 enhancement of MBL-initiated hemolysis EGTA is indeed complement dependent and proceeds through the http://www.jimmunol.org/ classical pathway. To further characterize and enhance this system for study of the lectin pathway in human serum, we tested the ability of EAC4-M Factor D and factor B requirement for MBL-initiated hemolysis sensitized with MBL for lysis in human serum diluted in Mg- of E-M-MBL EGTA. These cells were significantly more susceptible to lysis in Factor D and C1q were depleted from NHS by affinity column Mg-EGTA than were E-M. Again, lysis was proportional to the chromatography on Bio-Rex 70, and this depleted serum was amount of MBL used to sensitize EAC4-M (2000 ng MBL/108 tested for the capacity to mediate hemolysis via the lectin pathway. cells was optimal), and significantly smaller amounts of human No lysis occurred when EAC4-M-MBL were incubated in the de- serum were required for 50% hemolysis (Fig. 3). The mean MBL- by guest on September 29, 2021 pleted serum, even when it was reconstituted with C1q (Fig. 6A). mediated hemolysis of EAC4-M-MBL in Mg-EGTA was 64 Ϯ 15 However, normal lysis was restored upon reconstitution with fac- LH U/ml (range, 50–100 U/ml), more than sevenfold greater 50 tor D. Restoration of lysis was directly dependent upon the than lysis of E-M-MBL (Fig. 2). The dependence of C4 enhance- amounts of factor D added; 0.25 ␮g was optimal (Fig. 6B). To ment on MBL-mediated hemolysis was further characterized. Ly- further test whether the alternative pathway is required for com- sis was greater as greater amounts of C4 were used for preparation plement-dependent lysis via the lectin pathway, experiments were of EAC4 (Fig. 4); the T of optimally reactive cells prepared max performed with factor B-depleted human serum. As shown in Fig- with 1/4 diluted human serum was 5 min. These results indicated ure 7, only minimal lysis occurred when EAC4-M-MBL were in- that MBL-mediated hemolysis was enhanced by C4, and that cubated in factor B-depleted serum, while normal lysis restored EAC4-M could serve as a valuable reagent for further study of upon reconstitution with 300 ␮g/ml purified human factor B. The lectin pathway activity. loss of lysis was incomplete because of incomplete depletion of factor B, as shown when mixing this depleted serum with EAC43 to assay for the presence of alternative pathway activity. Collec- tively, these results indicate that MBL-mediated hemolysis re- quires the alternative as well as the classical pathway.

Lysis occurs independently of Igs and C1 We conﬁrmed that the MBL-initiated lytic activity was indeed mediated via the lectin pathway by performing hemolytic assays using several different control cells and sera. Thus, EAC14 in the presence or absence of MBL were not lysed in NHS-Mg-EGTA, indicating that even when Ig and C1 were known to be present on the indicator cells, they could not lead to lysis under the conditions of the assay used in Mg-EGTA. Similarly, lysis of EAC4-M-MBL proceeded comparably in C1q-depleted NHS and in agammaglobulinemic sera in Mg-EGTA as it had in NHS, further exclud- ing a requirement for C1 and Ig (Figs. 6A and 8). Further, E-M FIGURE 2. Lysis of E-M-MBL and EAC4-M-MBL in sera from incubated in agammaglobulinemic serum diluted in calcium-containing buffer (i.e., in GGVB2ϩ) lysed only when the cells were healthy individuals in Mg-EGTA. Lysis is expressed in LH50 units; the Ϯ Ϯ LH50 was 9 8 U/ml NHS (range, 1–25 U/ml) using E-M-MBL and 64 presensitized with MBL, and then to a similar degree (average 15 U/ml NHS using EAC4-M-MBL (range, 50–100 U/ml). titer, 8.3 CH50 U/ml) as lysis of E-M-MBL when these same sera The Journal of Immunology 3009

FIGURE 3. Lysis of 0.1 ml of EAC4-M (1 ϫ 108/ ml) presensitized with MBL during 1-h incubation with 0.1 ml of human serum-Mg-EGTA at 37°C. Lysis is quantiﬁed using increasing amounts of MBL in the presence of optimal (1/16 diluted) human serum in A, while lysis of decreasing amounts of serum in the presence of optimal (2000 ng) MBL is shown in B.

were assayed in Mg-EGTA (average titer, 7.5 CH50 U/ml). These and C1s (35, 39, 40). While investigating the effectiveness of this results strengthen the conclusion that the lytic activity is mediated pathway in acute phase human sera, we were surprised to observe via the lectin pathway, and that the degree of lysis observed in the that MBL-sensitized E-M readily lysed in NHS in the absence of assay systems in Mg-EGTA is reﬂective of the degree of lysis in calcium, i.e., in Mg-EGTA. We initially anticipated that this oc- Downloaded from calcium-containing buffers. curred via the classical pathway because of the C1-like activities of MASP-1 and MASP-2 (35, 39, 40) and the well-documented uti- Calcium-independent retention of MBL on E-M lization of C4 and C2 in complement activation by MBL (31–36); Because E-M-MBL were lysed via the lectin pathway in NHS- indeed, MBL had been reported to lyse sheep erythrocytes sensi- Mg-EGTA, we wondered whether MBL was retained on the E-M tized with MBL ligands through the classical pathway in calcium-

surface upon exposure to and washes with Mg-EGTA. We inves- containing buffers, and both C4 (31, 32, 35) and C2 (32, 35) were http://www.jimmunol.org/ tigated the effect of divalent cation on binding and retention of required. Accordingly, we found that C2 was required for, and that MBL on E-M by flow cytometry. It can be seen in Figure 9 that C4 enhanced, MBL-mediated hemolysis, indicating classical path- calcium was required for binding of MBL on E-M, but, surpris- way participation. However, we found that factors D and B of the ingly, MBL was retained on E-M independent of calcium, even in alternative pathway were required as well. MBL had been reported Mg-EGTA, EGTA, and EDTA. This retention of MBL was selec- to activate the alternative pathway in human serum in the absence tive and seemed to involve a ligand present on the erythrocyte of C2, even in Mg-EGTA, when C3 deposition on appropriate surface distinct from mannan, because comparable calcium-inde- bacteria was quantitated (41). More recently, MASP was reported pendent retention was not observed on mannan immobilized on to cleave C3 directly and to result in activation of the alternative polystyrene plates or beads (data not shown). pathway in systems using either purified components or C4-defi- by guest on September 29, 2021 cient serum (42). Thus, the requirement for both the classical and Discussion alternative pathways, as reported herein, contrasts with previous reports implying that either the classical pathway alone or the al- MBL is a C-type lectin that can activate complement by a series of ternative pathway alone is sufficient for lectin pathway activity interactions known as the lectin pathway, i.e., via two new serum following activation of the MASP enzymes. proteases, MASP-1 and MASP-2, which are homologous to C1r We were surprised by the lysis of E-M-MBL in NHS in the presence of a calcium chelator, since MBL is a C-type lectin, and indeed, most procedures for purification of MBL involve elution by calcium chelation from mannose- or mannan-conjugated Sepharose. We therefore carefully investigated retention as well as binding of MBL to E-M and observed that in contrast to binding,

FIGURE 4. Increasing enhancement of lectin pathway-mediated lysis of EAC4-M-MBL in normal human serum (diluted 1/16 in Mg-EGTA) FIGURE 5. Absence of lysis of EAC4-M-MBL in each of ﬁve C2D cells prepared with increasing concentrations of human serum as a source human sera in Mg-EGTA, with lysis reaching levels characteristic of nor- of C4; cells prepared with 1/4 diluted human serum had a Tmax of 5 min. mal human serum upon reconstitution with puriﬁed human C2. 3010 ALTERNATIVE PATHWAY AND C4/C2 IN LECTIN PATHWAY HEMOLYSIS

FIGURE 6. Requirement of factor D for lysis of EAC4-M in human serum-Mg-EGTA. The absence of lysis of normal human serum depleted of factor D and C1q (bar 2) even when C1q was added back (bar 3) and the restoration of lysis by addition of factor D alone (bar 4) or in the presence of C1q (bar 5) are shown in A; lysis by normal human serum is shown in bar 1. Dose-dependent restoration of lysis using small amounts of purified factor D in 1/25 diluted (ᮀ) and 1/50 diluted (ࡗ) depleted serum, respectively, is shown in B. short term retention of MBL occurred not only in the absence of consumption and binding of C3 and/or C5; or the effect of control calcium but also in the presence of both EGTA and EDTA. This proteins is responsible. It cannot be attributable to decreased C4 calcium-independent retention of MBL on E-M is not seen in MBL binding alone, since even when the indicator cells were maximally reactions with mannan coated on polystyrene plates or conjugated presensitized with C4, alternative pathway amplification still was to Sepharose beads, suggesting that it involves binding to a still needed for lysis to occur. Further, an increase in C4 binding via the undefined ligand present on the erythrocyte surface. This is con- lectin pathway to levels well in excess of the minimum associated Downloaded from sistent with an earlier study postulating an erythrocyte binding site with Ab-initiated hemolysis via the classical pathway failed to re- that becomes available on MBL only consequent to its reactivity sult in hemolysis in absence of alternative pathway amplification with mannan (31). Hemolysis of E-M-MBL in NHS-Mg-EGTA (our unpublished observations). In any case, the alternative path- implies that calcium is not required for the association of the way requirement is consistent with a previous report that MBL MASPs with MBL, consistent with the observation that activated activates the complement system via the alternative pathway (36).

MASP associates with MBL independent of calcium (49) and in This latter report, which involved C3 deposition on mannose-rich http://www.jimmunol.org/ striking contrast to an absolute calcium requirement for the asso- Salmonella in NHS-Mg-EGTA, stated that this occurred via the ciation of C1q with C1r and C1s; it is not yet clear whether mag- alternative pathway alone, since it proceeded in a C2-deficient se- nesium or another divalent cation other than calcium is required. rum, whereas in the present report MBL-initiated lysis did not We do not know whether MASP activity is provided by the MBL occur in any of five C2-deficient sera until C2 was reconstituted. source (and hence absorbed to E-M before addition of NHS-Mg- Perhaps this difference is attributable to the different assay systems EGTA), by NHS-Mg-EGTA, or by both, but in each of these pos- used, with some (if not maximal) C3 deposition occurring partic- sibilities, MASP persistence on E-M would be dependent on per- ularly on appropriate bacteria in the absence of C2. It also is not sistence on the erythrocyte of the MBL with which it is associated. yet clear whether the recently reported direct C3 cleavage by In any case, calcium-independent retention of MBL on E-M is a MBL-MASP complexes (42) is involved in lectin pathway-initi- by guest on September 29, 2021 fundamental characteristic of the lectin pathway activity described ated hemolysis, perhaps via the previously described enhancement in this report. of terminal attack complex activity on cells presensitized with C3b It is striking that the classical pathway is necessary, but not sufficient, for hemolysis initiated by MBL-MASP complexes; factors D and B, and hence the alternative pathway, also are required for this lytic activity to occur. It is not yet clear why the alternative pathway is required. Perhaps inefficient C4 and C2 consumption, binding, and/or convertase assembly by E-MBL-MASP; decreased

FIGURE 7. Requirement of factor B for lysis of EAC4-M-MBL in hu- FIGURE 8. Control experiments showed that EAC14 are not lysed in man serum-Mg-EGTA. Human serum Ͼ90% depleted of factor B failed to human serum-Mg-EGTA under the conditions used; normal levels of lysis lyse EAC4-M-MBL unless puriﬁed factor B was added; the small amount via the lectin pathway are seen when agammaglobulinemic or C1q-deﬁ- of background lysis observed in the absence of added factor B also was cient serum are used as the human complement source in Mg-EGTA. These observed when EAC43 were added to the depleted serum, indicating in- experiments further exclude a requirement for either Ab or C1 in the lectin completeness of depletion of factor B. pathway-mediated lysis described in this report. The Journal of Immunology 3011 Downloaded from

FIGURE 9. Calcium-dependent binding (left column), but calcium-independent retention (right column), in the reaction of MBL with E-M, as quantiﬁed by ﬂow cy-

tometry. Binding was measured by addition http://www.jimmunol.org/ of MBL to E-M under various test conditions using biotinglated anti-MBL and R-phycoerythrin-conjugated streptavidin. Retention was measured by exposing E-M presensitized with MBL in the presence of calcium under the same array of conditions for 30 min at 37°C. The test conditions were E-M in the absence of MBL, GVB2ϩ, GVB con- 2ϩ taining 0.5 mM CaCl2 (GVB-Ca ), GVB 2ϩ by guest on September 29, 2021 containing 0.5 mM MgCl2 (GVB-Mg ), Mg-EGTA, EGTA, and EDTA.

(50, 51). Whatever the mechanism involved, clearly both the clas- MBL were used, and further optimization of both systems may be sical and the alternative pathways are required for the MBL-initi- possible. The basis for the enhanced activity by cells precoated ated hemolysis described herein. with C4 is not yet clear, but may derive from relatively inefﬁcient Hemolysis in NHS-Mg-EGTA of E-M presensitized with MBL C4 binding via the lectin pathway, improvement of C3/C5 con- in the presence of calcium proved to be a sensitive, quantitative, vertase formation, counteracting the effects of a lectin pathway and convenient assay for lectin pathway activity, generating titers inhibitory mechanism, and/or favoring of alternative pathway re- signiﬁcantly greater than those previously described. Activity was cruitment. Whatever the mechanism, we anticipate that assays in greater still when EAC4 coated with mannan and sensitized with Mg-EGTA involving C4-coated E-M presensitized with MBL will 3012 ALTERNATIVE PATHWAY AND C4/C2 IN LECTIN PATHWAY HEMOLYSIS

FIGURE 10. Diagrammatic sketch of complement activation and complement-mediated hemolysis via the lectin pathway (thick solid line), indicating the requirement for C4 and C2 of the classical pathway as well as for the alternative pathway; it is not yet clear how MBL and the MASP enzymes (M1 and M2) interact or whether a divalent cation (e.g., Mg2ϩ) is required. The classical and alternative pathways are indicated by thin solid and dashed lines, respectively; arrows indicate enzymatic cleav- ages and complement activation pathways. Downloaded from

be of use in further studies of the lectin pathway in the same 12. Storgaard, P., E. Holm Nielsen, O. Andersen, E. Skriver, H. Mortensen, P. manner that the EAC43 cell proved to be invaluable in analysis of Højrup, G. Leslie, U. Holmskov, and S.-E. Svehag. 1996. Isolation and characterization of porcine mannan-binding proteins of different size and ultrastructure. the alternative pathway (52). Scand. J. Immunol. 43:289.

A schematic diagram summarizing the interactions of the lectin 13. Weis, W. I., R. Kahn, R. Fourme, K. Drickamer, and W. A. Hendrickson. 1991. http://www.jimmunol.org/ Structure of the calcium-dependent lectin domain from a rat mannose-binding pathway as indicated by the present report compared with those of protein determined by MAD phasing. Science 254:1608. the classical and alternative pathways is shown in Figure 10. It 14. Holmskov, U., R. Malhotra, R. B. Sim, and J. C. Jensenius. 1994. Collectins: remains to be established whether this interpretation also applies to collagenous C-type lectins of the innate immune defense system. Immunol. Today 15:67. lectin pathway activation by indicator particles other than E-M, 15. Drickamer, K., M. S. Dordal, and L. Reynolds. 1986. Mannose-binding proteins such as bacteria, species of serum other than human, and comple- isolated from rat liver contain carbohydrate-recognition domains linked to col- ment-dependent interactions other than hemolysis. By its interac- lagenous tails. J. Biol. Chem. 261:6878. 16. Ezekowitz, R. A. B., L. E. Day, and G. A. Herman. 1988. A human mannose- tions with mannose moieties and other ligands (14, 17, 21), MBL binding protein is an acute phase reactant that shares sequence homology with thus can initiate complement activation and lead to complement- other vertebrate lectins. J. Exp. Med. 167:1034. dependent phagocytosis and lysis in a distinctive manner even in 17. Hoppe, H., and K. B. M. Reid. 1994. Collectins: soluble proteins containing by guest on September 29, 2021 collagenous regions and lectin domains and their role in innate immunity. Protein the absence of Ab. Studies of patients with deficiency of MBL Sci. 3:1143. have indicated that the lack of this pathway is associated with the 18. Sastry, K., G. A. Herman, L. Day, E. Deignan, G. Bruns, C. C. Morton, and R. A. B. Ezekowitz. 1989. The human mannose-binding protein gene exon struc- loss of normal host defense (23, 24). Further investigation of this ture reveals its evolutionary relationship to a human pulmonary surfactant gene newly appreciated mode of complement activation should yield and localization to chromosome 10. J. Exp. Med. 170:1175. insight into the role of complement and acute phase proteins in 19. Taylor, M. E., P. M. Brickell, R. K. Craig, and J. A. Summerfield. 1989. Structure and evolutionary origin of the gene encoding a human serum mannose-binding innate immunity, inflammation and host defense. protein. Biochem. J. 262:763. 20. Drickamer, K., and M. E. Taylor. 1993. Biology of animal lectins. Annu. Rev. Cell. Biol. 9:237. References 21. Epstein, J., Q. Eichbaum, S. Sheriff, and R. A. B. Ezekowitz. 1996. The collectins 1. Zhang, X.-H., C. Suankratay, Y.-H. Zhang, T. F. Lint, and H. Gewurz. 1997. in innate immunity. Curr. Opin. Immunol. 8:29. Calcium-independent retention of mannan-binding lectin (MBL) on mannan- 22. Kuhlman, M., K. Joiner, and R. A. B. Ezekowitz. 1989. The human mannose- coated erythrocytes and hemolysis via the lectin pathway in the absence of cal- binding protein functions as an opsonin. J. Exp. Med. 169:1733. cium. J. Invest. Med. 45:327A (Abstr.). 23. Sumiya, M., M. Super, P. Tabona, R. J. Levinsky, T. Arai, M. W. Turner, and 2. Suankratay, C., Y.-H. Zhang, X.-H. Zhang, T. F. Lint, and H. Gewurz. 1997. J. A. Summerfield. 1992. Molecular basis of opsonic defect in immunodeficiency Requirement of alternative pathway for hemolysis via the lectin pathway of com- children. Lancet 337:1569. plement (C) activation. Clin. Infect. Dis. 25:386 (Abstr.). 24. Lipscombe, R. J., M. Sumiya, A. V. Hill, Y. L. Lau, R. J. Levinsky, 3. Kawasaki, T., R. Etoh, and I. Yamashina. 1978. Isolation and characterization of J. A. Summerfield, and M. W. Turner. 1992. High frequencies in African and a mannan-binding protein from rabbit liver. Biochem. Biophys. Res. Commun. non-African populations of independent mutations in the mannose-binding pro- 81:1018. tein gene. Hum. Mol. Genet. 1:709. 4. Kozutsumi, Y., T. Kawasaki, and I. Yamashina. 1980. Isolation and character- 25. Madsen, H. O., P. Garred, J. A. Kurtzhals, L. U. Lamm, L. P. Ryder, S. Thiel, and ization of a mannan-binding protein from rabbit serum. Biochem. Biophys. Res. A. Svejgaard. 1994. A new frequent allele is the missing link in the structural Commun. 95:658. polymorphism of the human mannan-binding protein. Immunogenetics 40:37. 5. Mizuno, Y., Y. Kozutsumi, T. Kawasaki, and I. Yamashina. 1981. Isolation and 26. Super, M., S. Thiel, J. Lu, R. J. Levinsky, and M. W. Turner. 1989. Association characterization of a mannan-binding protein from rat liver. J. Biol. Chem. 256: of low level of mannan-binding protein with a common defect of opsonization. 4247. Lancet ii:1236. 6. Townsend, R., and P. Stahl. 1981. Isolation and characterization of a mannose/ 27. Garred, P., H. O. Madsen, B. Hofmann, and A. Svejgaard. 1995. Increased fre- N-acetylglucosamine/fucose-binding protein from rat liver. Biochem. J. 194:209. quency of homozygosity of abnormal mannose-binding protein alleles in patients 7. Oka, S., K. Ikeda, T. Kawasaki, and I. Yamashina. 1985. Isolation and charac- with suspected immunodeficiency. Lancet 346:941. terization of two distinct mannan-binding proteins from rat serum. Arch. Bio- 28. Garred, P., H. O. Madsen, U. Balslev, B. Hofmann, C. Pedersen, J. Gerstoft, and chem. Biophys. 260:257. A. Svejgaard. 1997. Susceptibility to HIV infection and progression of AIDS in 8. Kawasaki, N., T. Kawasaki, and I. Yamashina. 1983. Isolation and characteriza- relation to variant alleles of MBL. Lancet 349:236. tion of a mannan-binding protein from human serum. J. Biochem. 94:937. 29. Drickamer, K. 1987. Membrane receptors that mediate glycoprotein endocytosis: 9. Kawasaki, N., T. Kawasaki, and I. Yamashina. 1985. Mannan-binding protein structure and biosynthesis. Kidney Int. 32(Suppl. 23):167. and conglutinin in bovine serum. J. Biochem. 98:1309. 30. Malhotra, R., S. Thiel, K. B. M. Reid, and R. B. Sim. 1990. Human leukocyte 10. Oka, S., T. Kawasaki, and I. Yamashina. 1985. Isolation and characterization of C1q receptor binds other soluble proteins with collagen domains. J. Exp. Med. mannan-binding proteins from chicken liver. Arch. Biochem. Biophys. 241:95. 172:955. 11. Holt, P., U. Holmskov, S. Thiel, B. Teisner, P. Højrup, and J. C. Jensenius. 1994. 31. Ikeda, K., T. Sannoh, N. Kawasaki, T. Kawasaki, and I. Yamashina. 1987. Serum Purification and characterization of mannan-binding protein from mouse serum. lectin with known structure activates complement through the classical pathway. Scand. J. Immunol. 39:202. J. Biol. Chem. 262:7451. The Journal of Immunology 3013

32. Ji, Y.-H., M. Matsushita, H. Okada, T. Fujita, and M. Kawakami. 1988. The C4 43. Gewurz, A. T., T. F. Lint, S. M. Imherr, S. S. Garber, and H. Gewurz. 1982. and C2 but not C1 components of complement are responsible for the comple- Detection and analysis of inborn and acquired complement abnormalities. Clin. ment activation triggered by the Ra-reactive factor. J. Immunol. 141:4271. Immunol. Immunopathol. 23:297. 33. Ohta, M., M. Okada, I. Yamashina, and T. Kawasaki. 1990. The mechanism of 44. Praz, F., M. C. Roque Antunes Barreira, and P. Lesavre. 1982. A one-step pro- carbohydrate-mediated complement activation by the serum mannan-binding pro- cedure for preparation of classical pathway (C1q) and alternative pathway (factor tein. J. Biol. Chem. 265:1980. D) depleted human serum. J. Immunol. Methods 50:227. 34. Lu, J. H., S. Thiel, H. Wiedemann, R. Timpl, and K. B. M. Reid. 1990. Binding of the pentamer/hexamer forms of mannan-binding protein to zymosan activates 45. Harrison, R. A., and P. J. Lachmann. 1986. Complement technology. In Hand- the proenzyme C1r2C1s2 complex, of the classical pathway of complement, with- book of Experimental Immunology. D. M. Weir, ed. Blackwell Scientific Publi- out involvement of C1q. J. Immunol. 144:2287. cations, Palo Alto, p. 39.1. 35. Matsushita, M., and T. Fujita. 1992. Activation of the classical complement path- 46. Perucca, P. J., W. P. Faulk, and H. H. Fudenberg. 1969. Passive immune lysis way by mannose-binding protein in association with a novel C1s-like serine with chromic chloride-treated erythrocytes. J. Immunol. 102:812. protease. J. Exp. Med. 176:1497. 36. Ji, Y.-H., T. Fujita, H. Hatsuse, A. Takahashi, M. Matsushita, and M. Kawakami. 47. Borsos, T., and H. J. Rapp. 1967. Immune hemolysis: a simplified method for the 1993. Activation of the C4 and C2 components of complement by a proteinase in preparation of EAC4 with guinea pig or with human complement. J. Immunol. serum bactericidal factor, Ra reactive factor. J. Immunol. 150:571. 99:263. 37. Guan, E., S. L. Robinson, E. B. Goodman, A. J. Tenner. 1994. Cell-surface 48. Borsos, T., H. J. Rapp, and M. M. Mayer. 1961. Studies on the second component protein identified on phagocytic cells modulates the C1q-mediated enhancement of complement. I. The reaction between EAC14 and C2: evidence on the single of phagocytosis. J. Immunol. 152:4005. site mechanism of immune hemolysis and determination of C2 on a molecular 38. Tenner, A. J., S. L. Robinson, R. A. B. Ezekowitz. 1995. Mannose-binding pro- basis. J. Immunol. 87:310. tein (MBP) enhances mononuclear phagocytic function via a receptor that con- 49. Tan, S. M., M. C. M. Chung, O. L. Kon, S. Thiel, S. H. Lee, and J. Lu. 1996. tains the 126,000 M(r) component of the C1q receptor. Immunity 3:485. 39. Sato, T., Y. Endo, M. Matsushita, and T. Fujita. 1994. Molecular characterization Improvements on the purification of mannan-binding lectin and demonstration of of a novel serine protease involved in activation of the complement system by its calcium-independent association with a C1s-like serine protease. Biochem. J. 319:329. mannose-binding protein. Int. Immunol. 6:665. Downloaded from 40. Thiel, S., T. Vorup-Jensen, C. M. Stover, W. Schwaeble, S. B. Laursen, 50. Hammer, C. H., A. S. Abramovitz, and M. M. Mayer. 1976. A new activity of K. Pousen, A. C. Willis, P. Eggleton, S. Hansen, U. Holmskov, K. B. M. Reid, complement component C3b: cell-bound C3b potentiates lysis of erythrocytes by and J. C. Jensenius. 1997. A second serine protease associated with mannan- C5b,6 and terminal components. J. Immunol. 117:830. binding lectin that activates complement. Nature 386:506. 51. Yamamoto, K., T. F. Lint, and H. Gewurz. 1977. Enhancement of C56-initiated 41. Schweinle, J. E., R. A. B. Ezekowitz, A. J. Tenner, M. Kuhlman, and lysis by cell-bound C3 fragments: evidence for a mechanism independent of the K. A. Joiner. 1989. Human mannose-binding protein activates the alternative prior binding of C56 to C3b. J. Immunol. 119:1346. complement pathway and enhances serum bactericidal activity on a mannose-rich isolate of salmonella. J. Clin. Invest. 84:1821. 52. Fearon, D. T., K. F. Austen, and S. Ruddy. 1973. Formation of a hemolyti- 42. Matsushita, M., and T. Fujita. 1995. Cleavage of the third complement (C3) by cally active cellular intermediate by the interaction between properdin factors http://www.jimmunol.org/ mannose-binding protein-associated serine protease (MASP) with subsequent B and D and the activated third component of complement. J. Exp. Med. complement activation. Immunobiology 194:443. 138:1305. by guest on September 29, 2021