Bioconjugate Chem. 1995, 6, 367-372
367
Cobra Venom Factor Immunoconjugates: Effects of Carbohydrate-Directedversus Amino Group-DirectedConjugation+ Jane Zara,S,B Nicholas Pomato,” Richard P. McCabe,’IJ Reinhard Bredehorst,t,v a n d Carl-Wilhelm Vogel*Jsv Departments of Biochemistry and Molecular Biology and Medicine, Georgetown University, Washington, D.C. 20007, and Organon Teknika CorporatiodBiotechnology Research Institute, Rockville, Maryland 20850. Received November 29, 1994@
Human IgM monoclonal antibody 16-88, derived from patients immunized with autologous colon carcinoma cells, was derivatized with two different cross-linkers, S-(2-thiopyridyl)-~-cysteine hydrazide (TPCH), which is carbohydrate-directed, and N-succinimidyl-3-(2-pyridyldithio)propionate(SPDP), which is amino group-directed. Two antibody functions, antigen binding and complement activation, were assayed upon derivatization with TPCH and SPDP. TPCH allowed for extensive modification (up to 17 TPCH molecules per antibody) without impairment of antigen binding activity, while this function was significantly compromised upon derivatization with SPDP. Antibody molecules derivatized with 16 SPDP residues showed almost complete loss of their antigen binding function. The complement activating ability of antibody 16-88 was significantly decreased after derivatization with TPCH or SPDP. In the case of SPDP derivatization, this decrease of the complement activating ability is predominantly a consequence of the impaired binding function. Upon conjugation of cobra venom factor (CVF), a nontoxic 137-kDa glycoprotein which is capable of activating the alternative pathway of complement, the antigen binding activity of SPDP-derivatized antibody was further compromised, whereas that of TPCH-derivatized antibody remained unaffected even after attachment of three or four CVF molecules per antibody. In both conjugates CVF retained good functional activity. CVF was slightly more active when attached to SPDP-derivatized antibody, suggesting a better accessibility of amino group-coupled CVF for its interaction with other complement proteins. These results indicate that carbohydrate-directed conjugation compromises the antibody function of complement activation, but allows for the generation of immunoconjugates with unimpaired antigen binding capability. Accordingly, carbohydrate-directed cross-linkers may contribute to improve the efficacy of immunoconjugates in cancer therapy.
The vast majority of immunoconjugates consisting of two proteins are synthesized by chemical methods with heterobifunctional cross-linking reagents consisting of two differently reactive groups (for reviews, see 1-4).The €-amino groups of lysine residues, present on the surface of most proteins including antibodies, are especially suitable for the attachment of cross-linkers in that they react with a number of reagents under conditions that leave other groups in the protein unmodified. As a result, heterobifunctional cross-linking reagents containing a primary amine-reactive group have gained wide popularity. Lysine residues, however, are randomly distributed on the surface of proteins, and therefore, derivatization of their €-amino groups with heterobifunctional crossPreliminary accounts of this work were presented at the 4th International Conference on Monoclonal Antibody Immunoconjugates for Cancer, San Diego, CA, MarcWApril, 1989. This work was supported by National Institutes of Health Grants CA 35525, CA 45800, and CA 01039 to C.-W.V. * To whom correspondence should be addressed: University of Hamburg, Department of Biochemistry and Molecular Biology, Martin-Luther-King-P1. 6, 20146 Hamburg, Germany. Georgetown University. 9 Current address: Department of Pharmacology, University of Texas Southwestern Medical Center, Dallas, TX 75235. l 1 Organon Teknika Corporatiofiiotechnology Research Institute. Current address: Centocor Corporation, Malvern, PA 19355. Current address: Department of Biochemistry and Molecular Biology, University of Hamburg, Martin-Luther-King-P1. 6, 20146 Hamburg, Germany. Abstract published in Advance ACS Abstracts, May 15, 1995. +
*
@
linking reagents may lead to impairment of protein function, e.g., antigen binding function in the case of antibodies. For a variety of antitumor monoclonal antibodies significant compromises in antigen binding activity have been observed upon derivatization with aminereactive heterobifunctional cross-linking reagents such (SPDP)l as N-succinimidyl-3-(2-pyridyldithio)propionate (5-7). In addition, the impairment of the antigen binding function is further increased for steric reasons when another protein is coupled to amine-attached cross-linker molecules a t or in proximity to the antigen binding site. Recently, we developed a novel heterobifunctional cross-linking reagent, S-(2-thiopyridyl)-~-cysteinehydrazide (TPCH), which contains a hydrazide moiety for coupling to aldehyde groups generated in the carbohydrate residues of antibodies by mild periodate oxidation, and a pyridyl disulfide group for coupling of effector molecules with a free sulfhydryl group (8). Since the carbohydrate moieties are distal to the antigen binding region of antibodies (91,derivatization with this crosslinker and subsequent coupling of effector molecules can be expected to minimize impairment of the antigen binding function. In the present study, we evaluated the effects of carbohydrate-directed and amino group-directed conjugation procedures using the human monoclonal IgM antibody 16-88 raised against human colon carcinoma Abbreviations: C W , cobra venom factor; TPCH, S ( 2 thiopyridy1)-L-cysteine hydrazide; SPDP, N-succinimidyl-342pyridy1dithio)propionate; DTT, dithiothreitol; PBS, 10 mM Naphosphate and 100 mM NaC1, pH 7.2; GPBS2+,PBS containing 0.1% (w/v) gelatin, 0.5 mM MgC12, and 0.9 mM CaC12.
1043-1802/95/2906-0367$09.00/00 1995 American Chemical Society
Zara et al.
368 Bioconjugate Chem., Vol. 6, No. 4, 1995
cells (10) and cobra venom factor (CVF), a nontoxic 137kDa glycoprotein which is capable of activating the alternative pathway of complement (for a review, see 11), as a model system. The results of this comparative analysis demonstrate that IgM-CVF conjugates synthesized with SPDP are severely impaired in their antigen binding activities. In contrast, the novel carbohydratedirected cross-linker TPCH allowed for the synthesis of IgM-CVF conjugates with no measurable impairment of the antigen binding function even when the antibody was derivatized with up to 17 TPCH molecules and three or four CVF molecules. Therefore, coupling of effector molecules to the carbohydrate region of antibodies may provide immunoconjugates with improved efficacy for the treatment of cancer. MATERIALS AND METHODS
Antibody Derivatization with TPCH. Human monoclonal IgM antibody 16-88 was obtained from hollow fiber culture and purified by gel filtration and ion exchange chromatography as described previously (10, 12). The antibody (2 mg, 2.2 nmol) in 1.0 mL of 0.1 M sodium acetate, pH 5.5, was oxidized a t 0 "C with 1mM sodium metaperiodate for 15 min in the presence of 15 mM TPCH. The reaction mixture was subjected to sizeexclusion chromatography on Sephadex G-25 (Pharmacia, Piscataway, NJ) equilibrated in 10 mM sodium phosphate and 100 mM sodium chloride (PBS), pH 7.2, and the derivatized antibody was stored a t 4 "C. The extent of TPCH incorporation was determined by the release of pyridine-2-thione a t 343 nm upon reduction with dithiothreitol (DTT) (13). Antibody Derivatization with SPDP. The 16-88 antibody (2 mg, 2.2 nmol) was incubated with either 7, 13, or 26 nmol of SPDP (Sigma, St. Louis, MO) in a total volume of 1.0 mL of PBS, pH 7.2. After 30 min a t 25 "C the pyridyldithio-derivatized antibody was purified by gel filtration on a G-25 Sephadex column in PBS, pH 7.2. The extent of cross-linker incorporation was determined as described above for TPCH. Conjugation of CVF to Cross-Linker-Derivatized Antibody. CVF was purified from lyophilized cobra venom (Naja naja kaouthia, Serpentarium Laboratories, Salt Lake City, UT) and derivatized with SPDP as described previously (14).The SPDP-derivatized CVF (2.8 mol of SPDP/mol of CVF) was incubated in the presence of 50 mM DTT for 20 min at 25 "C to reduce the pyridyl disulfide residues, subjected to gel filtration chromatography on Sephadex G-25 equilibrated with deaerated PBS, pH 7.2, and then used immediately for conjugation to either TPCH-modified or SPDP-modified antibody. The reaction mixture containing cross-linkerderivatized antibody (1.3 mg) and sulfhydryl-modified CVF (1.5 molar excess over the number of antibodyattached cross-linker molecules) in a total volume of 1.0 mL of PBS, pH 7.2, was flushed with nitrogen and incubated for 15 h a t 25 "C and then for 24 h a t 4 "C. After purification of the antibody conjugates by sizeexclusion chromatography using a Fractogel HW 65F column (1.5 x 115 cm) (EM Reagents, Gibbstown, NJ) equilibrated in PBS, pH 7.2, the fractions were pooled, concentrated by ultrafiltration (Amicon, Danvers, MA) and stored a t 4 "C. To determine the ratio of CVF per antibody molecule, lz5I-labeledCVF (1.5 x lo6 c p d m g ) was used for conjugation. Stoichiometries were determined from the difference in specific radioactivity before and after conjugation. Determination of Antigen Binding Activity. The antigen, obtained from human WIDR colon carcinoma cells by ammonium sulfate precipitation as described
previously (151, was coated onto Immulon-2 removable wells (Dynatech, Alexandria, VA) a t a concentration of 10 pg/mL protein in PBS, pH 7.2, for 1 h a t 37 "C. The wells were then blocked with 3% (v/v) fish gelatin (Norland Products, Inc., New Brunswick, NJ) in PBS, pH 7.2, for 1 h at 25 "C and washed three times with an aqueous solution containing 5%(v/v) glycerol and 0.05% (v/v) Tween-20. Varying amounts (8-2000 ng) of unmodified antibody 16-88, cross-linker-derivatized antibody, and conjugates (amounts are based on the antibody moiety) in 50 pL of PBS, pH 7.2, were mixed with 30 ng of lz51-labeledantibody 16-88 (specific radioactivity: 2 x lo6 cpdpg) in 50pL of PBS, pH 7.2, added to the antigencoated wells, and incubated for 15 h at 4 "C. After three washes with PBS, pH 7.2, containing 1%(w/v) bovine serum albumin, the wells were counted for radioactivity. The percent inhibition of [12511antibodybinding was calculated from the formula [l - (bound cpm in the presence of non-iodinated antibodyhound cpm in the absence of non-iodinated antibody)] x 100. Determination of the Hemolytic Activity of CVF. The complement activating ability of CVF was determined in a bystander lysis assay a s described previously (14). Briefly, 9 x lo6 guinea pig erythrocytes were incubated with 20 pL of guinea pig serum (diluted 1:2 with PBS) a t 37 "C for 30 min in the presence of varying concentrations of SPDP-derivatized or antibody-conjugated CVF in a total volume of 60 pL. Hemolysis was determined by measuring the release of hemoglobin from the lysed erythrocytes spectrophotometrically a t 412 nm. Determination of the Complement Activating Ability of Antibody 16-88. The ability of unmodified and cross-linker-derivatized antibody 16-88 to activate complement in the presence of antigen immobilized onto microtiter plates was measured in a complement depletion assay. The assay was based on a protocol described for a similar method using soluble antigen (16). Briefly, Costar 48-well polystyrene plates (Costar, Cambridge, MA) were coated with 500 pL of a crude antigen extract (5 pg) obtained from WIDR cells (15) for 15 h a t 4 "C. After the wells were washed three times with GPBS2', 250 pL of antibody (at either 5, 2.5, or 0.5 pg/mL) in GPBS2+and 125 pL of diluted human serum as complement source were added. The serum used was diluted in GPBS2- to a concentration which caused 90% hemolysis of sheep erythrocytes sensitized with hemolysin as described (16). After overnight incubation of the plates a t 4 "C, the reaction mixtures were transferred to glass tubes and 125 pL of sensitized sheep erythrocytes (6.3 x lo6 cells) was added. The solutions were incubated for 30 min a t 37 "C, diluted with 0.5 mL of cold GPBS2-, and centrifuged for 10 min a t 4 "C (800g). Released hemoglobin was measured spectrophotometrically at 412 nm. Percent consumption of complement hemolytic activity was calculated from these data. Other Methods. CVF was labeled with Nalz51using immobilized chloramine-?' (Iodo-Beads, Pierce, Rockford, IL) (17).Protein concentrations were determined by the Lowry method (18). RESULTS
Effect of Cross-Linker Derivatization on Antibody Binding. Human monoclonal IgM antibody 1688, directed against a cytoplasmic antigen from human colon carcinoma cells, was derivatized with the carbohydrate-directed cross-linker TPCH or the amino groupdirected cross-linker SPDP. The effect of cross-linker attachment on two antibody functions, antigen binding and complement activation, was evaluated. Figure 1 shows that the antigen binding ability decreased with
Bioconjugafe Chem., Vol. 6, No. 4,1995 369
Random versus Site-Directed Conjugation 100 I
,
80
5 E
TPCH
SPDP
n
c
60
-E
I
80
c
P
0
40
cc
0 C 0
.-
CI
5cn
20
t
s
0
3
-r 5.1 8.6 2.4
0
1.8 7.8 9.2 16.6
Antibody Derivatization
40.00
10.00
2.50
0.63
0.16
0.04
Antibody [pglml) Figure 1. Effect of SPDP derivatization on the antigen binding activity of antibody 16-88. Shown is the ability of unmodified and modified antibody to bind antigen in a competition assay and with 1251-labeledantibody 16-88. Unmodified antibody (0) SPDP-modified antibody a t 2.0:l (O), 8.6:l (W), and 16.0:l (A) mol of SPDP/mol of antibody.
[Mole Crosslinker/Mole Antibody] Figure 3. Effect of SPDP (left panel) and TPCH derivatization (right panel) on the complement activating ability of antibody 16-88. The complement activating ability was determined using a complement depletion assay as described in Materials and Methods.
6
100
I PDP
80 n
g .-.-
SPDP
TPCH
60
U
0 E
-c
40
TPCH
20 0 5.0 0
5.2
8.6
8.0
Antibody Derivatitation 40.00
10.00
2.50
0.63
0.16
Antibody [pg/ml] Figure 2. Effect of TPCH derivatization on the antigen binding activity of antibody 16-88. Shown is the ability of unmodified and modified antibody to bind antigen in a competition assay and with 1251-labeledantibody 16-88. Unmodified antibody (0) TPCH-modified antibody a t 1.8:l (01,9.2:l (W), and 16.6:l (A) mol of TPCWmol of antibody.
increasing SPDP modification. Antibodies derivatized with 16 SPDP residues exhibited almost complete loss of antigen binding function. In contrast, modification of antibody 16-88 with TPCH led to no measurable change in the antigen binding capability, even when 16 or 17 TPCH residues were attached per antibody molecule (Figure 2). Effect of Cross-Linker Derivatization on the Complement-Activating Ability of the Antibody. Figure 3 shows that derivatization of the 16-88 antibody with TPCH or SPDP caused a similar decrease in complement consumption. Since the assay requires first the binding of the antibody to its antigen, the observed
[Mole Crosslinker/Mole Antibody]
Figure 4. Effect of cross-linker attachment site on CVF coupling efficiency. The conjugation was performed with a 1.5fold molar excess of sulfhydryl-derivatized C W (2.8 mol of SH groups/mol of C W ) over the number of antibody-attached crosslinker molecules. SPDP-derivatized antibody, open bars; TPCHderivatized antibody, hatched bars.
decrease in complement consumption with SPDP-derivatized antibody appears to be a consequence of the compromised antigen binding rather than impairment of the complement activating ability (compare Figure 1).In contrast, the observed decrease in complement consumption with TPCH-derivatized antibody demonstrates that this cross-linker impairs the complement activating ability of the antibody since the antigen binding ability of TPCH-derivatized antibody is indistinguishable from that of unmodified antibody (compare Figure 2). Effect of Cross-LinkerAttachment Site on CVF Coupling Efficiency. Figure 4 shows the coupling efficiency of sulfhydryl-derivatized CVF to either SPDPor TPCH-derivatized antibody 16-88. Two antibody preparations containing approximately 5 or 8 mol of
Zara et al.
370 Bioconjugafe Chem., Vol. 6,No. 4, 1995 100
1
40.00
10.00
2.50
0.63
100
10.00
40.00
0.16
Antibody [pg/ml] Figure 5. Effect of CVF conjugation to SPDP-derivatized antibody 16-88 on its antigen binding ability. Shown is the and antibody-CVF conjugates ability of unmodified antibody (0) [O, 1.2 mol of CVF/mol of antibody (derivatized with 2.4 mol of SPDP/mol of antibody); A, 3.0 mol of CVF/mol of antibody (derivatized with 5.0 mol of SPDP/mol of antibody); W, 5.6 mol of CVF/mol of antibody (derivatized with 8.6 mol of SPDP/mol of antibody)] to bind antigen in a competition assay with lZ5Ilabeled antibody 16-88.
either cross-linker per mole of antibody were tested in these experiments. With SPDP-derivatized antibody 6570% of the available pyridyl disulfide groups were used for coupling of CVF, compared to only 15-40% with TPCH-derivatized antibody. These data suggest that the carbohydrate-bound pyridyl disulfide groups introduced by TPCH are less accessible for coupling of CVF than the amino-bound pyridyl disulfide groups introduced by SPDP. Effect of CVF Attachment Site on the Antigen Binding Activity. As shown in Figure 5, CVF conjugates prepared from SPDP-derivatized antibody were significantly compromised in their antigen binding capability. The compromise in binding activity increased with higher coupling ratios of CVF per antibody. On the basis of the data in Figures 1 and 5, both SPDP derivatization and CVF coupling contributed to the compromise in antigen binding ability. At a low coupling ratio (1-3 CVF molecules/antibody) the additional compromise due to CVF coupling was moderate, but it increased to more than 80% a t a coupling ratio of 5 or 6 CVF molecules per antibody (Figure 5). In contrast, virtually no compromise in antigen binding ability was observed when CVF was coupled to TPCH-derivatized antibody (Figure 6). Even those conjugates containing 3 or 4 CVF molecules per antibody were almost indistinguishable in their antigen binding ability from unmodified antibody. These data demonstrate that the carbohydrate moieties can serve as attachment sites for large effector molecules such as CVF without impairment of antigen binding function. Effect of Attachment Site on the CVF Hemolytic Activity. Figure 7 shows the effect of conjugation on the hemolytic activity of CVF. Compared to SPDPderivatized CVF, antibody-conjugated CVF exhibited a decreased hemolytic activity whether coupled to SPDPor TPCH-derivatized antibody. The compromise in activity was greater for carbohydrate-attached CVF than for amino-attached CVF, suggesting a lower accessibility of
2.50
0.63
0.16
Antibody [pglml] Figure 6. Effect of CVF conjugation to TPCH-derivatized antibody 16-88 on its antigen binding ability. Shown is the and antibody-CVF conjugates ability of unmodified antibody (0) containing, per mole of antibody (derivatized with 8.0 mol of TPCWmol of antibody) 0.5 (01,2.0 (W) or 3.2 mol of CVF (A)t o bind antigen in a competition assay with 125I-labeled antibody 16-88.
0
'
I
I
I
I
2500 1250 625 313
I
I
1
I
I
I
156
78
40
20
10
5
CVF [ngl Figure 7. Effect of attachment site on the hemolytic activity of conjugated CVF. The hemolytic activity of CVF (derivatized with 2.8 mol of SPDP/mol of CVF) before and after conjugation to SPDP-derivatized (- -1 or TPCH-derivatized (-1 antibody 16-88 was measured in a bystander lysis assay as described in Materials and Methods; 0, SPDP-derivatized CVF; V, antibody, antibodyconjugated CVF (2.0 mol of CVF/mol of antibody); . conjugated CVF (0.4 mol of CVF/mol of antibody); 0 , antibodyconjugated CVF (3.2 mol of CVF/mol of antibody); A, antibodyconjugated CVF (0.5 mol of CVF/mol of antibody). Due to the comparatively low derivatization of CVF with 2.8 mol of SPDP/ mol of CVF, the hemolytic activity of the derivatized CVF was virtually indistinguishable from that of unmodified CVF (not shown) (6).
the carbohydrate-attached CVF molecules for factor B (92 kDa) andor complement component C5 (191 kDa) required for its hemolytic activity. DISCUSSION
The results of this study demonstrate that amino group-directed and carbohydrate-directed attachment of
Random versus Site-Directed Conjugation
cross-linker molecules have different effects on two antibody functions, complement activation and antigen binding. Derivatization of the carbohydrate moieties of human monoclonal IgM 16-88 with as many as 16 or 17 TPCH molecules did not affect its binding activity, whereas attachment of the same number of molecules of the amino group-directed crosslinker SPDP caused virtually complete loss of the antibody’s binding function. These data show that carbohydrate-attached TPCH molecules are located distal to the antigen binding region, whereas a t least some of the amine-attached SPDP molecules are located within the antigen binding region. When SPDP-modified antibodies were used for the conjugation of CVF, the antigen binding function was further compromised. In contrast, coupling of CVF to TPCH-modified antibodies did not cause any significant impairment of antigen binding function. Even those conjugates carrying 3 or 4 CVF molecules were indistinguishable in their antigen binding ability from unmodified antibody molecules. The ability to activate complement was also tested, since the carbohydrate moieties are proposed to be involved in complement activation and, in particular, Clq binding. Several lines of evidence for this have been reported: The complement activating ability of a murine monoclonal IgGzb antibody was abolished in the corresponding aglycosyl derivative (19);deglycosylation of rabbit IgG by P-aspartyl-N-acetylglycosamido hydrolase eliminated the antibody’s ability to bind Clq (20);human IgG treated with P-galactosidase showed a significant reduction in Clq binding (21). The carbohydrates, which are proposed to act as a bridge between constant domains of a n immunoglobulin molecule (221, are thought to provide a higher order structure which is necessary for the constant region of the antibody to participate in C l q binding and to function as a specific ligand for receptor recognition and binding. The results of our study support the view that carbohydrates are involved in complement activation. It remains to be determined, however, whether the observed compromise in the complement activating ability upon TPCH derivatization is due to a direct alteration of the carbohydrates or to subsequent structural alteration(s) of the protein as a result of cross-linker attachment. When TPCH- and SPDP-modified antibodies were used to couple CVF, different coupling efficiencies were observed. The amine-attached SPDP residues were more accessible to coupling of CVF than the carbohydrateattached TPCH residues. This may be due to the localization of the carbohydrates in a sterically hindered region of the antibody, between the constant domains of the immunoglobulin. In addition, the discrete area in which the carbohydrates are localized may limit the number of CVF molecules that can be attached. The limited accessibility of the carbohydrate region is also suggested by the greater compromise observed in the hemolytic activity of CVF coupled to TPCH-derivatized antibody compared to SPDP-derivatized antibody. Factor B, which must bind to CVF in order to activate complement through the alternative pathway, has a molecular mass of 92 kDa and may for sterical reasons have impaired access to CVF when it is coupled to the carbohydrate region. The same applies t o complement component C5 (191 kDa), the binding of which is also required for the hemolytic activity of CVF. CVF coupled to the antibody through the more exposed amines via the cross-linker SPDP appears to have increased accessibility for interaction with factor B and C5, providing for immunoconjugates with better retention of CVF hemolytic activity.
Bioconjugafe Chem., Vol. 6,No. 4, 1995 371
The contention that the somewhat compromised activity of CVF when coupled via TPCH to the carbohydrate moieties in the hinge region of the antibody is due to steric hindrance because of the comparatively large molecular weight of CVF and its requirement to interact with two rather large proteins to exert activity is further supported by the observation that the activity of barley toxin, a significantly smaller effector molecule, when coupled via TPCH to the same antibody was indistinguishable from the activity of the derivatized but noncoupled barley toxin (8). Although a rather large number of heterobifunctional cross-linking reagents have been developed over the last 15 years, only very few reactive groups are employed and, consequently, very few conceptionally different crosslinking reagents exist (for a review, see 23). Although the successful derivatization of the carbohydrate moieties of antibody molecules without impairment of the binding function was demonstrated some time ago (241, carbohydrate-directed heterobifunctional cross-linking reagents such as TPCH have only been described more recently (8,251. In conclusion, immunoconjugates synthesized with the carbohydrate-directed cross-linker TPCH appear to be superior compared to those synthesized with amino group-directed cross-linker molecules. Although the use of TPCH causes a slight decrease in the coupling efficiency, even large effector molecules such as CVF can be coupled to TPCH-derivatized antibodies without impairment of the antigen binding function. LITERATURE CITED (1) Wawrzynczak, E. J., and Thorpe, P. E. (1987) Methods for preparing immunotoxins: effect of the linkage on activity and stability. In Zmmunoconjugates, Antibody Conjugates in Radioimaging and Therapy of Cancer ((2.-w.Vogel, Ed.) pp 2855, Oxford University Press, Oxford, New York. (2) Vogel, C.-W. (1987) Antibody conjugates without inherent toxicity: the targeting of cobra venom factor and other biological response modifiers. In Zmmunoconjugates, Antibody Conjugates in Radioimaging and Therapy of Cancer ((2.-W. Vogel, Ed.) pp 170-188, Oxford University Press, Oxford, New York. (3) FitzGerald, D. J. P. (1987) Construction of immunotoxins using Pseudomonas exotoxin A. Methods Enzymol. 151,139145. (4) Frankel, A. E., Welsh, P. C., Withers, D. I., and Schlossman, D. M. (1988) Immunotoxin preparation and testing in vitro. In Antibody-Mediated Delivery Systems (J. D. Rodwell, Ed.) pp 225-244, Marcel Dekker, Inc., New York, Basel. (5) Vogel, C.-W. (1988) Synthesis of antibody conjugates with cobra venom factor using heterobifunctional cross-linking ragents. In Antibody-Mediated Delivery Systems (J. D. Rodwell, Ed.) pp 191-224, Marcel Dekker, Inc., New York, Basel. (6) Petrella, E. C., Wilkie, S. D., Smith, C. A,, Morgan, A. C., Jr., and Vogel, C.-W. (1987) Antibody conjugates with cobra venom factor. Synthesis and biochemical characterization. J . Immunol. Methods 104, 159-172. (7) Juhl, H., Petrella, E. C., Cheung, N.-K. V., Bredehorst, R., and Vogel, C.-W. (1990) Complement killing of human neuroblastoma cells: A cytotoxic monoclonal antibody and its F(ab)’n-cobra venom factor conjugate are equally cytotoxic. Mol. Immunol. 27, 957-964. (8) Zara, J., Wood, R., Boon, P., Kim, C.-H., Pomato, N., Bredehorst, R., and Vogel, C.-W. (1991) A carbohydratedirected heterobifunctional cross-linking reagent for the synthesis of immunoconjugates. Anal. Biochem. 194, 156162. (9) Silverton, E. W., Navia, M. A., and Davies, D. R. (1977) Three-dimensional structure of an intact human immunoglobulin. Proc. Natl. Acad. Sci. U.S.A. 74, 5140-5144. (10) Haspel, M. V., McCabe, R. P., Pomato, N., Janesch, J. J., Knowlton, J. V., Peters, L. C., Hoover, H. C., Jr., and Hanna,
Zara et al.
372 Bioconjugate Chem., Vol. 6,No. 4, 1995 M. G., Jr., (1985) Generation of tumor cell-reactive human monoclonal antibodies using peripheral blood lymphocytes from actively immunized colorectal cancer patients. Cancer Res. 45, 3951-3960. (11) Vogel, C.-W. (1990) Cobra venom factor, the complementactivating protein of cobra venom. In Handbook of Natural Toxins, Vol. 5, Reptile andAmphibian Venoms (A. T. Tu, Ed.) pp 147-188, Marcel Dekker, Inc., New York, Basel. (12) McCabe, R. P., Peters, L. C., Haspel, M. V., Pomato, N., Carrasquillo, J. A., and Hanna, M. G., Jr., (1988) Preclinical studies on the pharmacokinetic properties of human monoclonal antibodies to colorectal cancer and their use for detection of tumors. Cancer Res. 48, 4348-4353. (13) Carlsson, J., Drevin, H., and Axen, R. (1978) Protein thiolation and reversible protein-protein conjugation. Nsuccinimidyl 3-(2-pyridyldithio) propionate, a new heterobifunctional reagent. Biochem. J . 173, 723-737. (14) Vogel, C.-W., and Muller-Eberhard, H. J. (1984) Cobra venom factor: Improved method for purification and biochemical characterization. J. Immunol. Methods 73,203-220. (15) Hanna, M. G., Jr., (1989) Human tumor antigens and specific tumor therapy. In U.C.L.A. Symposia on Molecular and Cellular Biology (R. S. Metzgar and M. S. Mitchell, Eds.), Vol. 99, pp 127-136, Alan R. Liss, Inc., New York. (16) Mayer, M. M. (1961)Complement and complement futation. In Experimental Immunochemistry (E. A. Kabat and M. M. Mayer, Eds.) pp 133-240, Charles C. Thomas, Springfield, IL. (17) Lee, D. S. C., and Grifiths, B. W. (1984) Comparative studies on Iodo-bead and Chloramine-T methods for radioiodination of human a-fetoprotein. J . Immunol. Methods 74, 181-189. (18) Lowry, 0. H., Rosebrough, N. J., Farr, A..L., and Randall, R. J. (1951) Protein measurement with the folin phenol reagent. J . Biol. Chem. 193, 265-275.
(19) Nose, M., and Wigzell, H. (1983) Biological significance of carbohydrate chains on monoclonal antibodies. Proc. Natl. Acad. Sci. U.S.A. 80, 6632-6636. (20) Winkelhake, J. L., Kunicki, T. J., Elcombe, B. M., and Aster, R. H. (1980) Effects of pH treatment and deglycosylation of rabbit IgG on the binding of Clq. J. Biol. Chem. 255, 2822-2828. (21) Tsuchiya, N., Eneo, T., Matsuta, K., Yoshinoya, S.,Aikawa, T., Kosuge, E., Takeuchi, F., Miyamoto, T., and Kobata, A. (1989) Effects of galactose depletion from oligosaccharide chains on immunological activities of human IgG. J . Rheumatol. 16, 285-290. (22) Parekh, R. B., Dwek, R. A., Sutton, B. J., Fernandes, d. L., Leung, A., Stanworth, D., Rademacher, T. W., Mizuochi, T., Taniguchi, T., Matsuta, K., Takeuchi, F., Nagano, Y., Miyamoto, T., and Kobata, A. (1985) Association of rheumatoid arthritis and primary osteoarthritis with changes in the glycosylation pattern of total serum IgG. Nature 316, 452457. (23) Mattson, G., Conklin, E., Desai, S., Nielander, G., Savage, M. D., and Morgensen, S. (1993) A practical approach to crosslinking. Mol. Biol. Rep. 17, 167-183. (24) Rodwell, J. D., Alvarez, V. L., Lee, C., Lopes, A. D., Goers, J. W. F., King, H. D., Powsner, H. J., and McKearn, T. J. (1986) Site-specific covalent modification of monoclonal antibodies: In vitro and in vivo evaluations. Proc. Natl. Acad. Sci. U.S.A. 83, 2632-2636. (25) Chamow, S. M., Kogan, T. P., Peers, D. H., Hastings, R. C., Byrn, R A., and Ashkenazi, A. (1992) Conjugation of soluble CD4 without loss of biological activity via a novel carbohydrate-directed cross-linking reagent. J . Biol. Chem. 267, 15916-15922. BC950021W
