Toxicon 56 (2010) 1198–1222

Contents lists available at ScienceDirect

Toxicon
journal homepage: www.elsevier.com/locate/toxicon

Review

Cobra venom factor: Structure, function, and humanization for

therapeutic complement depletion
Carl-Wilhelm Vogel*, David C. Fritzinger
Cancer Research Center of Hawaii, University of Hawaii at Manoa, 1236 Lauhala Street, Honolulu, HI 96813, USA

a r t i c l e i n f o a b s t r a c t

Article history: Cobra venom factor (CVF) is the complement-activating protein in cobra venom. This
Received 14 November 2009 manuscript reviews the structure and function of CVF, how it interacts with the
Received in revised form 7 April 2010 complement system, the structural and functional homology to complement compo-
Accepted 8 April 2010
nent C3, and the use of CVF as an experimental tool to decomplement laboratory
Available online 22 April 2010
animals to study the functions of complement in host defense and immune response
as well as in the pathogenesis of diseases. This manuscript also reviews the recent
Keywords:
progress in using the homology between CVF and C3 to study C3 structure and
Complement
C3 function, and to develop human C3 derivatives with the complement-depleting
Cobra venom factor function of CVF. These human C3 derivatives represent humanized CVF, and are
CVF a conceptually different concept for pharmacological intervention of the complement
Convertase system, therapeutic complement depletion. The use of humanized CVF for therapeutic
Protein humanization complement depletion in several pre-clinical models of human diseases is also
Therapeutic complement depletion reviewed.
Ó 2010 Elsevier Ltd. All rights reserved.

1. The complement system in health and the disease One activation product of the complement system is the
macromolecular membrane attack complex (MAC)1, which
The complement system is an intrinsic part of the inserts itself into target membranes, leading to the impair-
immune system, with important functions in both innate ment of membrane function and physical destruction of
and adaptive immunity. The complement system consists targets including viruses, bacteria, parasites, and tumor cells.
of approximately twenty plasma proteins as well as Another activation product is C3b, derived from complement
numerous receptors and regulatory proteins on the protein C3, which attaches itself covalently to cells, immune
membranes of host cells. Activation of the complement complexes, and other particles, a process called opsonization.
system can occur through three pathways, the classical Opsonization tags cells and particles for removal by phago-
pathway, the lectin pathway, and the alternative pathway, cytes. Further products of proteolytic degradation of C3b
all of which share a similar molecular architecture in that (iC3b, C3dg, C3d) are recognized by specific receptors on
an initial recognition event is amplified by a succession of various immune cells, and have important functions in the
proteolytic enzymes in a cascade-like fashion, resulting in
the generation of numerous biologically active comple-
ment activation products (Fig. 1). 1
CVF, cobra venom factor; rCVF, recombinant CVF; CVF-His, rCVF with
a histidine tag; MAC, membrane attack complex; ELISA, enzyme-linked
immunosorbant assay; FITC, fluorescein isothiocyanate. LPS, lipopoly-
saccharide; ADCC, antibody-dependent cellular cytotoxicity; MI/R,
myocardial infarction reperfusion; GPI, glycosylphospatidylinositol; NK
* Corresponding author. Tel.: þ1 808 564 5979; fax: þ1 808 586 2970. cells, natural killer cells; PNH, paroxysmal nocturnal hemoglobinuria;
E-mail addresses: [email protected] (C.-W. Vogel), dfritzin@ AMD, age-related macular degeneration. The acronyms for the domains
crch.hawaii.edu (D.C. Fritzinger). of C3 and CVF are explained in Section.

0041-0101/$ – see front matter Ó 2010 Elsevier Ltd. All rights reserved.
doi:10.1016/j.toxicon.2010.04.007
 C.-W. Vogel, D.C. Fritzinger / Toxicon 56 (2010) 1198–1222 1199

Fig. 1. Schematic representation of the three pathways of complement activation and the final pathway of membrane attack.

regulation of the adaptive immune response. Last but not three pathways, and its various activation products are
least, the anaphylatoxins C3a and C5a, biologically active crucial for mediating, directly or indirectly, virtually all
peptides derived from complement components C3 and C5, biological functions of complement. It has a molecular mass
respectively, exhibit important pro-inflammatory activities, Mr w 190,000 Da consisting of a Mr w 115,000 Da a-chain
such as degranulation of mast cells, increased vascular and an Mr w 75,000 Da b-beta chain which are linked by
permeability, and chemotaxis and activation of neutrophils. a single disulphide bond (Fig. 2) (Janatova, 1986). With
Both anaphylatoxins are rapidly inactivated by the removal of a concentration of approximately 1.2 mg/ml (6 mM) in
a C-terminal arginine residue through the action of human plasma, C3 is one of the more abundant plasma
carboxypeptidase N. However, the inactivated form of C5a, proteins in humans and all vertebrates. With the exception
C5a-des-Arg, still exhibits its activity for neutrophils, making of cobra C3, C3 from all other vertebrate species that have
C5a a particularly powerful pro-inflammatory peptide. been studied are glycoproteins, although the number, size,
This short description of the complement system and locations of the N-linked carbohydrate chains vary
summarizes its important physiological functions in host (Alsenz et al., 1992; Fritzinger et al., 1992). The major site
defense and immune response. Excellent and comprehensive for the synthesis of C3 in plasma is the liver although
reviews of the complement system describe the biochem- various cells from the immune system have been shown to
istry and genetics of the complement proteins, the three synthesize C3 locally (Botto, 2000). C3 is synthesized as
activation pathways, the structure and function of receptors, a single-chain pre-pro-protein. After removal of the signal
the regulation, and the biological functions (Campbell et al., peptide, the mature two-chain C3 protein is generated by
1988; Fujita, 2002; Gros et al., 2008; Law and Reid, 1988; a furin-type protease, removing a stretch of four arginine
Lutz and Jelezarova, 2006; Mollnes and Lachmann, 1988;
Morgan and Harris, 1999; Müller-Eberhard, 1988; Rother
Table 1
et al., 1998; Turnberg and Botto, 2003; Walport, 2001a,b).
Disease with complement pathogenesis.
Inappropriate complement activation can occur through
multiple mechanisms, leading to damage of host cells and Disease Reference

tissues. As such, the complement system serves as the sole Rheumatoid arthritis (Kerwar et al., 1981; Linton and
Morgan, 1999; Wang et al., 1995)
or an important pathogenetic factor in numerous diseases
Lupus erythematosus (Buyon et al., 1992)
(Cochrane, 1984; Holers, 2008; Holers and Thurman, 2004; Myasthenia gravis (Lennon et al., 1978)
Klos et al., 2009; Markiewski and Lambris, 2007; Petersen Hyperacute rejection after (Dalmasso, 1992; Leventhal
et al., 2001; Turner, 2003). Table 1 represents a list of xenotransplantation et al., 1993)
diseases in which an important role of complement in the Age-related macular (Edwards et al., 2005; Hageman
degeneration et al., 2005; Haines et al., 2005)
pathogenesis has being established.
Ischemia/reperfusion injury (Arumugam et al., 2004;
Riedemann and Ward, 2003;
2. The third component of complement (C3) and Zhou et al., 2000)
Paroxysmal nocturnal (Nishimura et al., 2001;
cobra venom factor (CVF)
hemoglobinuria (PNH) Parker et al., 1985)
Bullous pemphigoid (Jordon et al., 1973)
2.1. Structure and function of C3 Asthma (Arroyave et al., 1977)
Anti-phospholipid syndrome (Holers et al., 2002)
Complement component C3 is the most important Atypical hemolytic uremic (Ying et al., 1999)
syndrome
protein of the complement system as it is activated by all
1200 C.-W. Vogel, D.C. Fritzinger / Toxicon 56 (2010) 1198–1222

Furin protease
s s bimolecular complex C3b,Bb, the C3 convertase of the
alternative pathway of complement (Götze and Müller-
Eberhard, 1971). C3b,Bb goes on to cleave more C3 into
β-chain s s α-chain C3a and C3b, thereby amplifying the activation signal.
pro-C3 C3 contains a rare structural feature, an intra-molecular
thioester between a glutamic acid residue (encoded as
s s glutamine) and a cysteine residue, separated by two amino
acids in the a-chain (Tack, 1983; Tack et al., 1980). Upon
C3 convertase

α-chain activation of C3 by removal of the C3a anaphylatoxin from

C3
s s the N-terminus of the a-chain, the intra-molecular thio-
β-chain ester in nascent C3b becomes highly reactive and is rapidly
hydrolyzed by nucleophiles such as amino or hydroxyl
s s
groups on cell surface structures. This is the molecular basis
Factor I
C3a
for the covalent attachment of C3b to target cells or other
α particles. The intra-molecular thioester in circulating C3
C3b
s s also serves a second biological function. The thioester is
β-chain buried in the native C3 protein but subject to slow hydro-
lysis by water, resulting in a form of C3 that has the C3a
domain still attached but functionally resembles C3b. This
s s
Factor I
C3b-like C3 is referred to as C3(H2O) or iC3 and can bind
C3f
α factor B and form a convertase just like C3b (Pangburn and
iC3b Muller-Eberhard, 1983; Pangburn et al., 1981). The bimo-
s s
lecular C3(H2O),Bb convertase results in low-grade but
β-chain
constant activation of C3 and thereby serves as the
continuous trigger for the activation of the alternative
Protease
s s pathway of complement (Pangburn and Muller-Eberhard,
C3dg
1983).
α
C3c
s s
2.2. Structure and function of CVF and its homology to C3
β-chain

Fig. 2. Schematic representation of the chain structures of C3 and its

2.2.1. Structure of CVF
physiological degradation products. Single-chain pro-C3 is converted into Cobra venom factor (CVF) is the complement-activating
the mature C3 protein by removal of four arginine residues by a furin protein in cobra venom. Mature CVF has a molecular mass
protease, generating the a- and b-chains of C3. Removal of the C3a ana- of Mr w 149,000 Da (mostly isolated from the Asiatic cobra
phylatoxin by a C3 convertase results in the formation of C3b, consisting of
species Naja naja or Naja kaouthia; see 2.2.7. below). It is
the a0 -chain and the unaltered b-chain. Removal of the small C3f fragment
by factor I causes formation of iC3b. A third cleavage by factor I generates the a three-chain protein consisting of an Mr w 68,500 Da a-
C3dg fragment and C3c. C3dg is subsequently cleaved into the smaller C3d chain, an Mr w 48,500 Da b-chain, and an Mr w 32,000 Da
fragment. g-chain. The g-chain can exhibit size heterogeneity at its C-
terminus. CVF is highly homologous to complement
residues that separates the N-terminal b-beta chain from component C3 (Vogel and Müller-Eberhard, 1984). The
the C-terminal a-chain in pro-C3 (Fig. 2) (de Bruijn and Fey, extensive structural homology between CVF and C3 was
1985; Misumi et al., 1991). Human C3 mRNA is 5101 initially discovered by a series of biochemical studies (Alper
nucleotides long, encoding a 1663 amino acid residue pre- and Balavitch, 1976; Eggertsen et al., 1981; Lundwall et al.,
pro-C3, consisting of a 22-residue signal peptide, the 645 1984; Vogel et al., 1984) as well as immunological cross
amino acids of the b-beta chain, the four arginine residues, reactivity (Alper and Balavitch, 1976; Eggertsen et al., 1983;
and the 992 residues of the a-chain (de Bruijn and Fey, Vogel et al., 1984), subsequently confirmed by the molec-
1985; Vik et al., 1991). The C3 mRNA has 63 nucleotides ular cloning of CVF (Fritzinger et al., 1994, 1992), and,
of untranslated sequence at the 50 -end, and 49 nucleotides recently, by the determination of its crystal structure
of untranslated sequence at the 30 -end of the coding (Janssen et al., 2006, 2009; Krishnan et al., 2009). Like C3,
sequence. The human C3 gene is approximately 41 kb in CVF is synthesized in as a single-chain pre-pro-protein
size and contains 41 exons (Vik et al., 1991). which is subsequently processed into the mature three-
The activation of the complement system through any of chain protein. The a-chain of CVF is homologous to the b-
the three pathways leads to the formation of a C3 con- chain of C3, and the b- and g-chains of CVF are homologous
vertase. These enzymes, containing either C3b or C4b, cleave to the C-terminal and N-terminal portions of the C3 a0 -
C3 into the small C3a anaphylatoxin (Mr w 10,000 Da) and chain, respectively (Fig. 3) (Eggertsen et al., 1981; Fritzinger
the large C3b component (Mr w 180,000 Da) consisting of et al., 1994). After removal of the signal peptide, processing
the a0 -chain (Mr w 105,000 Da) and an unaltered b-chain of pro-CVF involves the removal of four arginine residues
(Fig. 2). C3b binds factor B in the presence of Mg2þ ions. The separating the CVF a-chain from a CVF g/b-precursor chain,
C3b,B complex, called the pro-convertase, is the substrate which is homologous to the C3 a-chain. Further processing
for factor D which cleaves factor B, resulting in the release of includes the removal of the C3a anaphylatoxin-like
the Ba activation peptide and the generation of the sequence as well as the central portion (C3dg-like
 C.-W. Vogel, D.C. Fritzinger / Toxicon 56 (2010) 1198–1222 1201

C3b

CVF
β-chain
γ-chain β-chain

C3c
α-chain
α kD α kD

β-chain

Fig. 3. Schematic representation of the chain structures of C3b, C3c, and CVF. Shown are the chain structures of three proteins as they relate to mature C3. Please
note the somewhat larger sizes of the CVF g- and b-chains compared to the corresponding a0 -chain fragments of C3c.

sequence) from the g/b-precursor chain, resulting in the identity to cobra C3 is over 93%. The CVF mRNA is greater
mature three-chain CVF molecule. The removal of the four than 5948 nucleotides in length. Its open reading frame
arginine residues likely involves a furin-type protease encodes for the 1642 amino acid residues of the CVF pre-
(Misumi et al., 1991). The proteases in the venom gland pro-protein, consisting of a 22-residue signal sequence, the
responsible for generating mature CVF from its C3-like or 627 residue a-chain, four arginine residues, and the 989
C3b-like form of pro-CVF are not known. However, our residues of the g/b-precursor chain from which the mature
laboratory described a Mr w 48,000 Da metalloprotease in g- and b-beta chains are derived. The CVF mRNA has a 50 -
cobra venom, termed cobrin, that can cleave human C3 or untranslated region of 16 nucleotides and a 30 -untranslated
C3b into a three-chain derivative, termed C3o, with struc- region of 999 nucleotides, and a poly-A tail of at least 20
tural and functional properties similar to C3b (see Section nucleotides (Bammert et al., 2002).
2.2.4. below) (Bambai et al., 1998; O’Keefe et al., 1988, 1984; The high degree of protein sequence homology is
Petrella et al., 1991). It appears likely that cobrin is involved corroborated by the recently solved crystal structures of
in the posttranslational processing of pro-CVF into mature human C3 and its C3b and C3c derivatives (Janssen et al.,
CVF in the venom gland. The three-chain structure of CVF 2006, 2005; Wiesmann et al., 2006) as well as CVF
resembles that of C3c, a physiological degradation of (Janssen et al., 2009; Krishnan et al., 2009). Mature CVF has
product of C3 (Fig. 3). However, the sizes of the two C3 a- eleven domains, sharing the eight macroglobulin (MG)
chain derived fragments of C3c (Mr w 29,000 Da, N- domains, the linker (LNK) domain, and the C-terminal
terminal; Mr w 42,000 Da, C-terminal) are smaller than the C345C domain with C3c (Fig. 4). In addition, CVF also has
sizes of the corresponding CVF g- and b-chains. the “complement C1r/C1s, Uegf, Bmp1” (CUB) domain,
The high degree of homology between CVF and C3 is which is absent in C3c but present in C3b. The functionally
underscored by the sequence homology of both protein and important CUB domain (see Section 2.2.4. below) is formed
DNA sequence. CVF exhibits a sequence identity to human by two non-contiguous stretches of amino acid sequence
C3 and other mammalian C3 species of approximately 50%, from the C-terminus of the CVF g-chain and the N-terminus
and of approximately 70% allowing for conservative amino of the CVF b-chain (Fig. 4). Because the Mr w 29,000 Da
acid changes (Table 2). The sequence identity of CVF to chain of C3c lacks a significant portion of sequence of the
cobra C3 is approximately 85% with a similarity of almost CUB domain at its C-terminus, and the Mr w 42,000 kD
92% after allowing for conservative replacements chain lacks a significant portion of sequence of the CUB
(Fritzinger et al., 1994, 1992; Vogel et al., 1996). All thirteen domain at its N-terminus, C3c has no functional CUB
disulfide bonds are conserved. At the DNA level, CVF shares domain. C3b not only has the CUB domain but also the
a sequence identity with human and other mammalian thioester-containing (TED) domain which represents the
species of approximately 57% whereas the sequence C3dg portion of the C3 a-chain, and which is flanked by the
two stretches of sequence that make up the CUB domain
Table 2 (Fig. 4). Like C3c, CVF lacks the TED domain (Fig. 4).
Homology of CVF to C3s from different vertebrate species. Similar to C3, CVF is a glycoprotein. It has three glyco-
sylation sites for N-linked oligosaccharide chains, two of
C3 species Protein DNA identity
which are in the a-chain and one is in the b-chain. The
Identity Similarity carbohydrate content of CVF is 7.4% (w/w), significantly
Cobra 85 92 93 higher than that of human C3 (1.7%) and other mammalian
Chicken 55 72 57
C3 species (Tomana et al., 1985; Vogel and Müller-Eberhard,
Human 50 69 56
Mouse 52 70 57 1984). The structures of the CVF carbohydrate chains have
Rat 51 70 58 been extensively characterized (Gowda et al., 2001, 1994,
Guinea pig 53 70 56 1992; Grier et al., 1987). The major oligosaccharide is
Xenopus 51 67 55 a symmetric fucosylated biantennary complex-type chain
Carp (C3-H1) 52 62 51
with an unusual a-galactosylated Lex structure at its
1202 C.-W. Vogel, D.C. Fritzinger / Toxicon 56 (2010) 1198–1222

Fig. 4. Structures of C3b, CVF, and C3c. The upper panel shows the three-dimensional domain structures of CVF (2.2 Å), C3b (4.0 Å), and C3c (2.4 Å). The lower
panel shows the schematic domain structures of these three proteins. Please note the absence of the CUB domain in C3c compared to CVF and C3b. Only C3b has
the TED domain. (Modified from: Janssen et al., 2006,, 2009,, 2005).

non-reducing end (Gowda et al., 2001). Glycosylation of CVF Müller-Eberhard, 1971; Hensley et al., 1986; Lesavre et al.,
is not required for its biological activity. Partial or complete 1979; Müller-Eberhard and Fjellström, 1971; Smith et al.,
deglycosylation as well as the introduction of charges into 1984; Vogel, 1991; Vogel and Müller-Eberhard, 1982; Vogt
the carbohydrate chains are without functional conse- et al., 1974). In addition, CVF,Bb, like C3b,Bb, can also
quence. However, there is evidence that the oligosaccharide cleave C5 into C5a and C5b (Vogt et al., 1978; Vogt and
chains contribute to the immunogenicity of CVF (Gowda Schmidt, 1978; DiScipio et al., 1983; von Zabern et al.,
et al., 2001, 1994). CVF with its natural oligosaccharide 1980). Accordingly, the enzyme is both a C3 convertase
chains could be crystallized (Sharma et al., 2001; Janssen et and a C5 convertase, and is often referred to as C3/C5 con-
al., 2009). However, deglycosylation was required to obtain vertase. As the enzymatically active site, cleaving both C3
crystals suitable for high-resolution structural analysis and C5, resides in the Bb subunit of the bimolecular complex
(Janssen et al., 2009, Krishnan et al., 2009). independent of whether Bb is bound to CVF, C3b, or C3(H2O),
The CVF gene is greater than 89 kb in length, more than the alternative pathway C3/C5 convertase is identified by
twice as large as the human C3 gene (Bammert et al., 2002; a single EC number (EC 3.4.21.47).
Vik et al., 1991). However, the exon structure of the CVF Although the three enzymes CVF,Bb, C3b,Bb, and C3
gene is extremely homologous to the human C3 gene. The (H2O),Bb share the molecular architecture (Janssen et al.,
total number of exons of the CVF gene is 40, which is one 2009; Rooijakkers et al., 2009; Smith et al., 1984, 1982),
exon less than the human C3 gene (Vik et al., 1991). This consisting of a structural subunit (CVF, C3b, or C3(H2O))
difference is the result of the loss of the intron between and the identical active site-bearing subunit Bb, and the
exons 31 and 32 (using human C3 gene numbering), so that
exons 31 and 32 are contiguous in the CVF gene (Bammert Factor D
Formation: C3b + Factor B C3b,Bb + Ba
et al., 2002). Mg2+

Factor D
2.2.2. The CVF-dependent C3/C5 convertase: similarities and C3(H2O) + Factor B C3(H2O),Bb + Ba
Mg2+
differences to the C3b-dependent convertase
Factor D
Given the enormous structural homology between CVF CVF + Factor B CVF,Bb + Ba
Mg2+
and C3 and its physiological breakdown products, it is not
surprising that CVF and C3 share functional homology.
Although the three-chain structure of CVF resembles C3c
C3b,Bb, C3(H2O),Bb
more than C3b, CVF is a functional homolog of C3b. Like C3b, Function: C3 + H2O or CVF,Bb
C3b + C3a
CVF binds factor B in the presence of Mg2þ ions, to form the
CVF,B pro-convertase. This is subsequently cleaved by factor C5 + H2O C3b,Bb, C3(H2O),Bb
C5b + C5a
or CVF,Bb
D into the activation peptide Ba and the bimolecular
complex CVF,Bb. Like C3b,Bb and C3(H2O),Bb, CVF,Bb, is a C3 Fig. 5. Formation and function of the alternative pathway C3/C5
convertase, cleaving C3 into C3a and C3b (Fig. 5) (Götze and convertases.
 C.-W. Vogel, D.C. Fritzinger / Toxicon 56 (2010) 1198–1222 1203

substrate specificity for C3 and C5, the three enzymes Vogt et al., 1978). The molecular basis of the increased
exhibit considerable functional differences: affinity for C5 and the requirement for additional C3b
molecules in the proximity to the C3,Bb enzyme is not
1) All three enzymes exhibit spontaneous decay–dissoci- fully understood. However, C5 binds with high affinity
ation into the respective subunits, which abolishes the to multimeric C3b complexes; and this binding may
enzymatic activity. The free Bb subunit exhibits residual involve at least two different binding sites on C5 for C3b
enzymatic activity for small synthetic peptide ester (Low et al., 1999; Rawal and Pangburn, 2001; Sandoval
substrates, but can no longer act on C3 or C5 (Caporale et al., 2000). Although no studies of the C5-cleaving
et al., 1981). The C3b,Bb and C3(H2O),Bb enzymes are activity have been reported for the fluid-phase con-
very short-lived at 37  C, exhibiting half-lives of the vertase C3(H2O),Bb, it can safely be assumed that the C3
decay–dissociation of approximately 1.5 min (Fritzinger (H2O),Bb enzyme does not exhibit significant C5-
et al., 2009; Medicus et al., 1976; Pangburn and Muller- cleaving activity under physiological conditions, similar
Eberhard, 1986). In contrast, the CVF,Bb enzyme is to fluid-phase C3b,Bb. In contrast, CVF,Bb exhibits fluid-
physico-chemically very stable with a half-life of phase C5-cleaving activity (DiScipio et al., 1983; Petrella
approximately 7 h (Vogel and Müller-Eberhard, 1982). et al., 1987; von Zabern et al., 1980), consistent with a Km
2) Both the C3b,Bb and C3(H2O),Bb enzymes are subject to value of 0.036 mM, well below the physiological C5
rapid and efficient inactivation by factors H and I, above concentration in plasma (Rawal and Pangburn, 2000).
and beyond the rapid intrinsic decay–dissociation. Factor Like C3b, CVF has a binding site for C5; and C5 needs to
H dissociates C3b,Bb and C3(H2O),Bb (Pangburn et al., be bound to CVF for its cleavage by CVF,Bb (Fritzinger
1977; Whaley and Ruddy, 1976) and serves as cofactor et al., 2009; von Zabern et al., 1980). It has been
for the proteolytic inactivation of C3b and C3(H2O) by suggested that the C5 substrate can be bound to the
factor I, generating the cleavage products of iC3b, and, same CVF molecule that carries the Bb subunit (Vogt
ultimately, C3c, C3dg, and C3d (Fig. 2) (Pangburn et al., et al., 1977); and this may be the molecular reason
1977; Whaley and Ruddy, 1976). In contrast, both the that CVF, Bb exhibits efficient fluid-phase C5-cleaving
CVF,Bb enzyme and CVF are completely resistant to the activity.
regulatory actions of factors H and I (Lachmann and
Halbwachs, 1975; Nagaki et al., 1978). Both iC3b and C3c 2.2.3. Use of CVF to deplete complement in vitro and in vivo
are unable to form a convertase with factor B. In contrast, The ability of CVF to form a fluid-phase enzyme that is
since CVF is not cleaved by factor I it can (re)form a con- physico-chemically stable and resistant to inactivation by
vertase (Alper and Balavitch, 1976; Lachmann and factors H and I is the molecular basis for the striking bio-
Halbwachs, 1975; Nagaki et al., 1978). logical activity of being able to deplete serum complement.
3) Another difference between the three convertases is the When CVF is added to serum, the CVF,Bb enzyme forms,
fact that the C3b,Bb enzyme is surface-bound on a cell or utilizing factor B and factor D. Once formed, the CVF,Bb
a particle whereas as both the CVF,Bb enzyme and the enzyme cleaves C3 and C5. The generated C5b will also
C3(H2O),Bb enzyme are fluid-phase enzymes. consume the complement components of the terminal
4) All three enzymes exhibit differences in the kinetics of C3 membrane attack pathway. As a consequence of continuous
hydrolysis. Michaelis constants (Km) and catalytic activation of C3 and C5, the action of the CVF,Bb enzyme
constants (kcat) were determined for the three enzymes will lead to the depletion of serum complement activity,
(Pangburn and Muller-Eberhard,1986; Vogel and Müller- although depletion is primarily a function of C3
Eberhard, 1982). The resulting kcat/Km values for C3b,Bb consumption (Van den Berg et al., 1991).
(31.1 104 M1  s1), C3(H2O),Bb (16.3  104 M1  s1), The anti-complementary effect as well as of the ana-
and CVF,Bb (4.05  104 M1  s1) reveal that the C3b,Bb phylatoxin-generating effect of cobra venom had been
enzyme exhibits the best catalytic efficiency of the three known since the early years of the last century (Flexner and
enzymes. Noguchi, 1903; Friedberger et al., 1913). But it was not until
5) Finally, the enzymes exhibit differences in the kinetics of the late 1960s that it was discovered that complement
C5 hydrolysis. Soluble fluid-phase C3b,Bb can cleave C5. activation leads to the release of the anaphylatoxins, and
However, the Km of the monomeric C3b,Bb enzyme that both the anti-complementary activity and anaphyla-
(24 mM) is well above the physiological C5 concentration toxin-generating activity of cobra venom are a consequence
in plasma (0.37 mM), implying that C3b,Bb exhibits no of the exhaustive complement activation by CVF (Cochrane
significant C5-cleaving activity under physiological et al., 1970; Maillard and Zarco, 1968; Nelson, 1966; Vogt
conditions (Rawal and Pangburn, 2001). On surfaces and Schmidt, 1964). Ever since it was demonstrated that
such as a target cell, the C3-cleaving activity of the CVF can be safely administrated to laboratory animals for
C3b,Bb enzyme results in the attachment of numerous temporary depletion of complement activity (Cochrane
C3b molecules in close proximity to the convertase. In et al., 1970), CVF has become a widely used experimental
this setting, the additional C3b will bind the substrate tool to study the biological functions of complement as well
C5, resulting in a gradual increase in the affinity of the as its involvement in the pathogenesis of many diseases by
convertase for C5 and a concomitant reduction of the Km comparing normal (i.e., complement sufficient) with
to as low as 0.016 mM, well below the normal plasma complement-depleted animals. Fig. 6A shows the effect of
concentration of C5, and essentially switching the C3- a single i.p. injection of 150 mg CVF (600 mg/kg) on the
cleaving activity to the C5-cleaving activity (Rawal and serum complement activity in guinea pigs. Injection of CVF
Pangburn, 1998, 2000, 2001; Vogt and Schmidt, 1978; leads to rapid depletion of complement, resulting in very
1204 C.-W. Vogel, D.C. Fritzinger / Toxicon 56 (2010) 1198–1222

A 100 CVF has been shown to be able to deplete comple-

ment in serum from all vertebrates tested except cobras
(Vogel and Müller-Eberhard, 1985a). Cobras have
80
a complement system like all other vertebrates (Vogel
% Hemolytic activity

and Müller-Eberhard, 1985a,b). In contrast to other acti-

60 vators of the alternative pathway such as zymosan,
inulin, and lipopolysaccharide (LPS), CVF when incubated
40 with cobra plasma does not affect the complement
hemolytic activity of cobra plasma and does not convert
20 cobra C3 into C3b (Fig. 7) (Vogel, 1986; Vogel and Müller-
Eberhard, 1985a). In cobra plasma, CVF undergoes
a change in its electrophoretic mobility (Fig. 7) (Vogel,
0
0 2 4 6 8 10 1986), suggesting that CVF is bound to a plasma protein
Days after injection in cobra plasma which inhibits CVF activity. It appears
that the binding protein for CVF is cobra factor B
B 100
(Grunwald et al., 1996). Apparently, cobra factor B when
bound to CVF is not cleaved by cobra factor D thereby
80 preventing convertase formation and complement acti-
% Hemolytic activity

vation in cobra plasma.

60 The only side effect from massive activation of
complement in vivo by CVF has been an acute and
40 fleeting inflammatory injury of the lungs (Mulligan et al.,
1996; Till et al., 1982, 1987). The lung inflammation is
mediated by the C5a or C5a-des-Arg anaphylatoxin which
20
activates neutrophils, leading to their subsequent
sequestration in the lungs. Intravascular injection of CVF
0
0 2 4 6 8 10
(400 mg/kg) into guinea pigs pretreated with an inhibitor
Days after injection of carboxypeptidase N, resulting in a 97% inhibition of the
enzyme, was lethal (Huey et al., 1983). Even under these
Fig. 6. In vivo complement depletion by CVF. Guinea pigs (Panel A) or mice conditions of near complete inhibition of carboxypepti-
(Panel B) were injected i.p. with a single dose of CVF as shown. At time
dase N, a lower dose of CVF (200 mg/kg), though still
intervals as indicated, the hemolytic complement activity in serum was
measured. causing acute respiratory distress, was no longer lethal,
and all animals recovered. Collectively, these data
demonstrate the effective protection of a host by car-
low or undetectable levels of complement hemolytic boxypetidase N from massive intravascular release of the
activity (Vogel and Müller-Eberhard, 1984). Complement anaphylatoxins, and that the C5a-mediated acute
activity remains low for another day or two until resyn- inflammatory lung injury is transitory, and that both the
thesis of complement components in the liver results in rate of C5a production and the total quantity generated
normal complement levels. Fig. 6B shows a dose response are important parameters in causing or preventing lung
of complement depletion after a single i.p. injection of CVF injury (Huey et al., 1983).
into mice (Vogel, 1991). The experiment demonstrates that
a sufficient quantity of CVF needs to be injected in order to 2.2.4. The importance of the CUB domain for convertase
effect complement depletion. Small doses of CVF will result formation
in only partial depletion of complement activity which is The recently solved structure of the bimolecular
relatively quickly restored by resynthesis of new comple- complex of CVF and factor B (CVF,B), which is also referred
ment components. At a dose of 0.1 mg (4 mg/kg) no reduc- to as the pro-convertase, revealed that factor B is loaded
tion of serum complement activity was measurable after onto CVF or C3b through its Ba domain (Janssen et al.,
24 h. As even a small amount of CVF would be expected to 2009). This observation explains the fact that free Bb is
form a CVF,Bb convertase which would subsequently cleave not able to form a convertase by re-associating with CVF
some C3, this result suggests there is no linear relationship or C3b. The Ba domain of factor B contacts several domains
between C3 concentration and overall complement of CVF or C3b. Complex formation depends on an intact
hemolytic activity, consistent with the much higher and properly oriented CUB domain, which is only present
concentration of C3 in plasma compared to other comple- in CVF and C3b, but not in C3, iC3b, and C3c. In C3, the
ment components. Conversely, increasingly higher dosages elements of the CUB domain are assembled properly, but
of CVF will result in only minor or immeasurable decreases the domain is not oriented correctly. The structural
in serum complement activity but may increase the time to changes that occur upon activation of C3 to C3b place the
complete restoration of normal complement levels. Even at CUB in C3b into its proper orientation (Janssen et al.,
very high levels of CVF, complete activation of C3 will not 2006). In the first inactivation step of C3b (from C3b to
be achieved, either in vitro or in vivo, presumably as iC3b), factor I cleaves the CUB domain of C3b in the
a consequence of a low turnover rate at low substrate (C3) binding site for Ba, while C3c lacks significant portions of
concentrations. the CUB domain (Janssen et al., 2006; Wiesmann et al.,
 C.-W. Vogel, D.C. Fritzinger / Toxicon 56 (2010) 1198–1222 1205

Fig. 7. Resistance of cobra complement to activation by CVF. Cobra plasma was incubated with 125I-CVF and subsequently subjected to immunoelectrophoresis
and development with an anti-CVF antiserum (which crossreacts with cobra C3) (upper panels). After drying, the immunoelectrophoresis gels were auto-
radiographed (lower panels). Please note the lack of activation of C3 to C3b, and the anodal shift of CVF.

2006). Accordingly, only C3b and CVF have an intact and expressed as the single-chain pro-CVF protein with
properly oriented CUB domain essential for factor B a Mr w 185,000 Da. In stably transfected Drosophila S2 cells,
binding and subsequent convertase formation. the single-chain form of pro-CVF is converted into a two-
The importance of the CUB domain for factor B binding chain C3-like form of pro-CVF by removal of the four
also explains an earlier observation that C3o, an alternate arginine residues between the CVF a-chain and the CVF g/
cleavage product of C3 generated by the metalloprotease b-precursor chain. A portion of this C3-like pro-CVF is
cobrin (Bambai et al., 1998; O’Keefe et al., 1984; Petrella further processed into a C3b-like form of CVF by removal of
et al., 1991), supports factor B binding and its activation the C3a anaphylatoxin-like sequence. Further processing of
by factor D (O’Keefe et al., 1988). C3o has a three-chain pro-CVF into the mature three-chain form as it occurs in
structure resembling C3c. Whereas C3o lacks the entire C- cobra venom does not occur in the insect cells used for
terminal portion of the CUB domain, its N-terminal expression, presumably because of a lack of the cobrin
portion of the CUB domain extends beyond the C-terminus metalloprotease, which we believe to be involved in the
of the corresponding 29,000 Da chain of C3c (Fig. 8). Since processing of pro-CVF in the venom gland into the three-
this is the only amino acid sequence that is present in C3o chain mature CVF protein. The recombinant forms of pro-
but absent in C3c, it suggests that this portion of the N- CVF produced in Sf9 and S2 cells exhibit insect-type
terminal half of the CUB domain is important for binding glycosylation patterns.
to the Ba domain of factor B, and that this binding site is Surprisingly, all three forms of pro-CVF exhibit functional
present in C3o although it lacks a complete CUB domain activity indistinguishable from mature CVF. Recombinant
(O’Keefe et al., 1987, 1988). CVF supports the activation of factor B in the presence of
factor D and Mg2þ ions, forms a stable bimolecular con-
2.2.5. Recombinant CVF vertase (pro-CVF,Bb) that exhibits cleaving activity for both
CVF can be produced recombinantly using eukaryotic C3 and C5, and that depletes serum complement activity,
expression systems (Kock et al., 2004; Vogel et al., 2004). thereby demonstrating its resistance to the degradation by
In baculovirus-infected Sf9 cells, recombinant CVF is factors H and I, just like mature CVF. Recombinant CVF lacks
1206 C.-W. Vogel, D.C. Fritzinger / Toxicon 56 (2010) 1198–1222

S S
Convertase Cobrin I Cobrin I Cobrin

S S

α
CUB Anchor
MG6 MG7 TED MG8 C345C

Fig. 8. Structural differences between C3o and C3c. The upper panel shows the schematic chain structures of C3o and C3c. White areas represent sequences
present in C3c but not in C3o. The black area identifies the only sequence present in C3o but not in C3c. Protease cleavage sites are indicated (C3 convertase: R748/
S749; cobrin: E758/E759, Q989/M990, K1353/D1354; factor I: R954/E955, R1320/S1321). The lower panel shows the portions of the CUB domain that are present
or absent in C3o compared to C3b and C3c

the intra-molecular thioester, consistent with its C3(H2O)- Asiatic cobra (N. naja) and several of its closely related
like or C3b-like structure and function. species2 (N. kaouthia and N. atra) (Eggertsen et al., 1981;
Müller-Eberhard and Fjellström, 1971; Takahashi and
2.2.6. CVF transgenic mice Hayashi, 1982; Vogel and Müller-Eberhard, 1984). Virtu-
Transgenic mice constitutively expressing CVF were ally all studies involving complement depletion of animals
generated using linearized DNA containing the full-length used CVF from N. naja or N. kaouthia. In addition to Asiatic
cDNA encoding single-chain pre-pro-CVF (Andrä et al., cobras, CVF has also been demonstrated to be present in the
2002). The transgene was expressed in liver and, to venom of other cobra species including N. melanoleuca
a lower extent, in kidneys. CVF protein could be detected in (White-lipped or forest cobra) (Osipov et al., 2005), N. nivea
small amounts in the serum. The serum complement (Cape or yellow cobra), N. nigricollis (black-necked or spit-
hemolytic activity of CVF transgenic mice was very low or ting cobra), and N. haje (Egyptian cobra).
virtually absent. The plasma C3 concentration was signifi-
cantly reduced and varied in individual transgenic animals
between below 10% to almost 40% of normal mice. The CVF
2
transgenic animals display no apparent abnormal pheno- The taxonomy of the Asiatic cobras is complex and has undergone
rather recent changes. Currently, there are eleven recognized species
type. The mice exhibit a normal life span and reproduce within the Asiatic cobra complex (Wüster, 1996), six of which occur on
normally. No tendency to develop infections was apparent; the Asian mainland (N. naja (Indian or spectacled cobra), N. kaouthia
and the mice do not require special housing as used for (monocellate or monocled cobra), N. atra (Chinese cobra), N. oxiana
other immunocompromised animals (Fritzinger et al., (Central Asian cobra), N. siamensis (Indochinese spitting cobra), N. man-
dalayensis (Burmese spitting cobra)), including two species that were
2010b). In contrast to C3 knockout mice, which lack both
only rather recently described or redescribed (N. siamensis, N. man-
detectable C3 levels and serum hemolytic complement dalayensis) (Slowinski and Wüster, 2000; Wüster et al., 1997), and five
activity, CVF transgenic mice have apparently sufficient that occur only on southeast Asian archipelagos of Indonesia, the
protection against infections from residual levels of C3 and Philippines, and the Andaman Islands (N. sputatrix (Southern Indonesian
the ability to synthesis C3 upon stimulation (Circolo et al., or Javan spitting cobra), N. philippinensis (Northern Philippine cobra), N.
sagittifera (Andaman cobra), N. samarensis (Southern Philippine cobra), N.
1999; Wessels et al., 1995). sumatrana (Equatorial spitting cobra)). Formerly, there was only one
recognized Asiatic cobra species (N. naja) with multiple subspecies (e.g.,
2.2.7. Occurrence of CVF in snake venoms N. n. naja, N. n. kaouthia, etc.) (Mehrtens, 1987; Phelps, 1981; Wüster,
The presence of CVF is limited to the Elapidae family of 1996). Accordingly, many earlier reports in the toxinological literature,
including reports on CVF, identify N. naja as the venom source (Cochrane
poisonous snakes. Based primarily on the presence of CVF
et al., 1970; Müller-Eberhard and Fjellström, 1971). Many studies denote
antigen, CVF had been shown to be present in the venoms the origin of the venom from N. n. siamensis (Eggertsen et al., 1981, 1983;
of the three genera Naja, Ophiophagus, and Hemachatus Petrella and Vogel, 1986; Vogel and Müller-Eberhard, 1984). However, N.
(Birdsey et al., 1971; Eggertsen et al., 1981; Takahashi and n. siamensis was not a recognized species or subspecies until recently;
Hayashi, 1982), while recent studies have also shown and the older designation by venom suppliers of N. n. siamensis refers to
the fact that the venom originated from Thailand. Most likely, the venom
a similar protein to be present in the venom of the was derived from N. kaouthia, although the precise origin will remain
Australian Elapid snake, Austrelaps superbus (Rehana and obscure, and it may even represent a mixture of more than one species
Manjunatha Kini, 2007). CVF has been isolated from the (Vogel, 1991; Wüster and Harvey, 1996; Wüster et al., 1997).
 C.-W. Vogel, D.C. Fritzinger / Toxicon 56 (2010) 1198–1222 1207

2.2.8. CVF lacking C5-cleaving activity without a histidine tag but lacks C5-cleaving activity (Fig. 9)
CVF from N. haje has been isolated and found to be (Kock, 1996). The location of the C5 binding site(s) on CVF
structurally very similar to CVF from N. naja (von Zabern (or C3b) is unknown, as is the molecular understanding of
et al., 1982). Interestingly, CVF from N. haje exhibits how a C-terminal histidine tag interferes with C5 binding
a significant functional difference to CVF from N. naja as the and/or its cleavage. Humanized CVF (see Section 3.2.2.
CVF,Bb convertase generated with CVF from N. haje displays below) also lacks C5-cleaving activity.
C3-cleaving activity but lacks C5-cleaving activity (Bauman,
1978). The virtual lack of C5-cleaving activity was shown to 2.2.9. Why is CVF in cobra venom?
be a consequence of a dramatically reduced binding affinity Although CVF is an abundant protein (approximately 2%
of CVF from N. haje for C5 (Rawal and Pangburn, 2000; von (w/w) of the dry (lyophilized) cobra venom (Vogel and
Zabern et al., 1980). Müller-Eberhard, 1984)), the biological function of CVF in
In addition to CVF from N. haje we generated a form of cobra venom has never been established. CVF certainly falls
recombinant CVF which similarly lacks C5-cleaving activity. into a different category of a venom component. In contrast to
Recombinant CVF with a C-terminal histidine tag (CVF-His) the truly toxic and lethal membrane toxins and neurotoxins
exhibits C3-cleaving activity just like recombinant CVF which make up the bulk of cobra venom (Iwanaga and Suzuki,

Fig. 9. Lack of C5-cleaving activity of recombinant CVF with a histidine tag (CVF-His). Panel A shows a dose response of the complement-depleting activity in
human serum of CVF, recombinant CVF, and CVF-His. Please note that all three forms of CVF exhibit indistinguishable complement-depleting activities. Panel B
shows the hemolytic activity of CVF, recombinant CVF, and CVF-His that is based on the fluid-phase cleavage of C5 in guinea pig serum. Please note the
indistinguishable C5-cleaving activities of CVF and recombinant CVF, and the complete absence of C5-cleaving activity of CVF-His. rCVF and CVF-His were single-
chain forms of recombinant pro-CVF produced in Sf9 cells.
1208 C.-W. Vogel, D.C. Fritzinger / Toxicon 56 (2010) 1198–1222

1979), the action of CVF is not lethal. Its “receptor” is factor B, (Table 3). Examples include low molecular weight inhibitors
leading to the formation of the stable CVF,Bb convertase with for complement enzymes such as the convertases or factor D
subsequent activation of complement. Massive complement (Szalai et al., 2000), receptor antagonists for the C3a and C5a
activation after intravascular injection of CVF has been shown receptors (Kohl, 2006), a recombinant fluid-phase form of
to cause activation of neutrophils and their sequestration to the complement receptor CR1 to inhibit the C3 convertase
the lungs (the first capillary system encountered) with (Li et al., 2006), monoclonal antibodies to C5 preventing its
subsequent injury to lung tissue. Those effects are temporary activation and downstream formation of the membrane
and fleeting (Mulligan et al., 1996; Till et al., 1982, 1987). The attack complex (MAC) (Hillmen et al., 2004; Wang et al.,
primary mediator of these effects is C5a or C5a-des-Arg. 1995), and the targeting of factor H in a fusion protein with
Protection from the potentially lethal massive intravascular complement receptor CR2 to the site of complement acti-
activation of complement is provided by the very efficient vation for C3 convertase inhibition (Banda et al., 2009).
serum protein carboxypeptidase N which rapidly inactivates Currently, the only therapy in clinical use is a monoclonal
C3a and most of the pro-inflammatory activities of C5a by antibody to C5 in patients with paroxysmal nocturnal
removing the C-terminal arginine residue (Huey et al., 1983; hemoglobinuria (PNH) (Hillmen et al., 2004). Several recent
Plummer and Hurwitz, 1978). Accordingly, massive comple- reviews describe the various approaches of developing
ment activation does not represent a direct toxic effect of the complement inhibitors for therapeutic application (Holers
venom component CVF, nor does the C5a-mediated lung and Thurman, 2004; Morgan and Harris, 2003; Ricklin and
injury as CVF from N. haje lacks C5-cleaving activity (see Lambris, 2007; Sahu and Lambris, 2000).
Section 2.2.8. above). It is therefore more likely that A different concept for therapeutic intervention of the
complement activation by CVF is of indirect benefit for the complement system, developed by nature itself in the form
cobra. Massive complement activation by CVF locally at the of CVF, is complement depletion. If an intact complement
site of venom injection into the prey animal releases system is not available during a pathological trigger for
the anaphylatoxins C3a and C5a which, in turn, will increase activation, tissue damage will be prevented, regardless of
the vascular permeability and blood flow at the site of how strong the activating signal may be. This pharmaco-
envenomation. These local consequences of complement logical concept for complement inhibition was derived
activation can be expected to aid the toxic venom compo- from the numerous studies in which CVF was used to
nents to enter the blood stream faster and to reach their site of deplete laboratory animals of their complement in order to
toxic action such as the neuromuscular end plates more study the role of complement in the pathogenesis of
quickly. For the cobra, this would have the benefit of a prey diseases. As a matter of fact, in many cases these studies
being paralyzed or killed much faster thereby reducing the served to establish the involvement of complement in the
struggle with the prey or the prey’s chances of escape. pathogenesis; and CVF has subsequently been used as the
gold standard for evaluating the complement inhibitory
activity of other drugs and biological (Hebell et al., 1991;
3. Humanized CVF for therapeutic complement
Morgan and Harris, 2003; Sahu and Lambris, 2000).
depletion
3.2. Humanization of CVF
3.1. Concepts of pharmacological complement inhibition
Conceptually, CVF itself could serve as a pharmacolog-
The growing realization that complement is involved in
ical agent for complement depletion. However, cobra
the pathogenesis of many diseases led to significant efforts in
venom is an impractical source of a biological for clinical
developing drugs and biologicals for therapeutic comple-
application. Furthermore, the Asiatic cobra is on the list of
ment inhibition over the past two decades. Most often, these
protected species. Even recombinant CVF, although lacking
approaches involve complement inhibition by either pre-
the unusual and immunogenic oligosaccharides of native
venting the activation of a complement component or
CVF, is likely to be too immunogenic, especially for repeat
blocking the activity of an activated complement component
applications, given the phylogenetic distance between
Table 3
humans and cobras. Given the very high degree of struc-
Concepts of pharmacological complement inhibition. tural homology between CVF and C3 – in the case of cobra
C3 the amino acid sequence similarity is approximately 92%
 Complement inhibition
(see above) – it appeared reasonable to create human C3
B Inhibition of activatable or activated complement components derivatives with CVF-like functions, recombinant biologi-
cals collectively referred to as humanized CVF.
- Protease inhibitors
- Receptor antagonists
3.2.1. Identification of the molecular region responsible for
- Soluble receptors
- Monoclonal antibodies forming a stable convertase
In order for a human C3 derivative to exhibit CVF-like
 Complement depletion functions, it must demonstrate two core properties: 1) it
must be able to form a physico-chemically stable con-
B Prevention of complement activation vertase displaying a relatively long half-life of its decay–
dissociation similar to that of CVF,Bb, and 2) it must be
- CVF, humanized CVF
resistant to the action of factor H or the combined action
of factors H and I. Only human C3 derivatives with these
 C.-W. Vogel, D.C. Fritzinger / Toxicon 56 (2010) 1198–1222 1209

two core properties would be able to deplete complement hybrid protein-containing convertases by factor H and
in serum and would therefore represent a humanized proteolytic cleavage by factor I of the hybrid proteins were
form of CVF. expected just as is the case for C3(H2O) and C3b. Accord-
We were able to locate the region of the CVF protein ingly, the CVF sequence in the C-terminal end of the hybrid
responsible for forming a physico-chemically stable con- protein a-chain must also confer, directly or indirectly,
vertase to the very C-terminal end of the CVF b-chain. resistance to the inactivating functions of factors H and I.
Whereas preliminary work involving limited proteolysis of Indeed, incubation of purified hybrid proteins with human
CVF with chymotrypsin provided an initial hint of the factors H and I revealed a partial resistance to degradation
importance of that region (Grunwald et al., 1993), conclu- by the two regulatory proteins (Fritzinger et al., 2009). The
sive confirmation was obtained by creating recombinant resistance of the hybrid proteins to the actions of factors of
loss-of-function derivatives of CVF by replacing portions of H and I is not as strong as that of CVF or recombinant CVF
CVF sequence with homologous portions from cobra C3 which exhibit complete resistance to the two regulatory
(Fig. 10) (Wehrhahn et al., 2000). proteins; however, it is strong enough to allow the hybrid
proteins as well as the convertases formed with the hybrid
3.2.2. Generation, structure, and function of humanized CVF proteins to be active in serum long enough to allow
After the discovery that the C-terminal portion of CVF complement depletion. Recently, the structure of the
harbors the crucial stability site responsible for forming complex of C3b with domains 1–4 of factor H has been
a stable convertase, recombinant gain-of-function deriva- described (Wu et al., 2009). Unfortunately, more structural
tives of human C3 were successfully created by replacing analyses will be required to understand the molecular basis
small stretches of human C3 at the very C-terminus of the C3 of the resistance of both recombinant CVF and humanized
a-chain with homologous stretches of sequence from the CVF to factors H and I, as the substitutions in our human-
CVF b-chain3 (Fig. 11, upper panel) (Fritzinger et al., 2009; ized CVF proteins are in domains that do not bind to
Hew et al., 2004; Kölln et al., 2005, 2004; Vogel and domains 1–4 of factor H.
Fritzinger, 2007). Using the Drosophila S2 system for Fig. 14 shows in vivo complement depletion with the
recombinant protein expression, both the human C3/CVF humanized CVF protein HC3-1496 in rats, mice, and cyn-
hybrid proteins and recombinant CVF are expressed as omologous monkeys (Fritzinger et al., 2008b,c; Fritzinger et
a mixture of a two-chain C3-like and a two-chain C3b-like al., 2010a). Please note that decomplementation is accom-
protein (Fig. 11, lower panel) (Fritzinger et al., 2009; Kock plished rapidly within minutes after injection of human-
et al., 2004). The hybrid proteins lack the intra-molecular ized CVF (Fig. 14A). There was little difference in the effect
thioester, consistent with their C3(H2O)-like or C3b-like of two doses of HC3-1496 used (280 mg/kg and 760 mg/kg,
structure and function (Fritzinger et al., 2009). These human rats; 250 mg/kg and 1000 mg/kg, monkeys). Please also note
C3/CVF hybrid proteins represent human C3 derivatives that complement levels increase much faster after deple-
which exhibit the CVF-like function of forming physico- tion with humanized CVF compared to CVF (Fig. 14B). This
chemically stable convertases displaying C3-cleaving is most likely due to the gradual inactivation of humanized
activity. Different human C3/CVF hybrid proteins, although CVF by factors H and I, and is corroborated by the rapid
sometimes exhibiting only minor sequence differences, removal of humanized CVF from the circulation (Fig. 14C).
show different rates of factor B cleavage/convertase forma- Differences in glycosylation may also be responsible (Fu et
tion, and the resulting hybrid protein-containing con- al., 1997).
vertases show different rates of C3 cleavage and different Another property of humanized CVF is the fact that
half-lives of their intrinsic decay–dissociation (Fritzinger convertases formed with these hybrid proteins exhibit no or
et al., 2008b, 2009; Vogel and Fritzinger, 2007). Signifi- only very low, residual C5-cleaving activity (Fritzinger et al.,
cantly, some of these humanized CVF proteins exhibit faster 2008c, 2009; Gorsuch et al., 2009; Vogel and Fritzinger,
rates of convertase formation, faster rates of C3 cleavage, and 2007). This lack of C5-cleaving activity is consistent with
longer half-lives than CVF or CVF,Bb, respectively (Fig. 12). a lack of C5 binding (Fritzinger et al., 2009), though puzzling
Humanized CVF is not only able to exhibit CVF-like as both parent proteins C3b and CVF exhibit C5 binding and
properties using purified human proteins for convertase cleavage. Fig. 15A shows that the convertases formed with
formation and C3 cleavage. Humanized CVF is also able to humanized CVF and purified human complement proteins
deplete serum complement activity. Fig. 13 shows in vitro do not cleave purified human C5 in vitro (Fritzinger et al.,
complement depletion by humanized CVF in human, cyn- 2009). Using either a hemolytic assay that is based on
omolgus monkey, and rat serum. The ability of humanized fluid-phase C5 cleavage (Vogel and Müller-Eberhard, 1984)
CVF proteins to deplete complement in serum, though or an ELISA to measure C5a production (Klos et al., 1988),
fortuitous, was entirely unexpected as they are human C3 undetectable or only very low levels of C5 activation were
derivatives that contain all known binding sites for factor H, observed in guinea pig or monkey serum after incubation
exhibit factor H binding (Fritzinger et al., 2009), and have with humanized CVF in vitro (Fig. 15B and C) (Fritzinger et al.,
all three cleavage sites for factor I. Rapid disassembly of the 2008c, 2009). Similar results were obtained in vivo after
decomplementation with the humanized CVF protein HC3-
1496, both in mice (Gorsuch et al., 2009) and monkeys (see
3
Fig. 20, below) (Fritzinger et al., 2008c).
Nomenclature of humanized CVF proteins: The number given in the
name of the protein (e.g., HC3-1496) is the last C3 amino acid residue
In terms of human application, a lack of C5-cleaving
(using C3 pre-pro-protein numbering) preceding the CVF sequence activity is indeed fortuitous. The C5a anaphylatoxin exhibits
insertion. the most powerful pro-inflammatory activity, and its effects
1210 C.-W. Vogel, D.C. Fritzinger / Toxicon 56 (2010) 1198–1222

α-chain γ-chain β-chain

CVF

H1

H2

H5
β-chain C3a α-chain
Cobra C3

100

80
% Complement depleted

60

40
rCVF
H1
H2
20
H5
cobra C3

0
0 50 100 150 200
Protein (ng)

Fig. 10. Structure and function of recombinant CVF/cobra C3 hybrid proteins (CVF loss-of-function derivatives). The upper panel shows a schematic represen-
tation of the chain structures of pro-CVF, cobra pro-C3, and three recombinant derivates of pro-CVF in which either portions of the CVF a-chain (CVF/cobra C3
hybrid proteins H1 and H2) or the C-terminal portion of the CVF b-chain (CVF/cobra hybrid protein H5) were replaced with the corresponding sequence from
cobra C3. Pro-CVF sequences (white) and pro-C3 sequences (gray) are indicated. The locations of homologous chains are shown. The lower panel shows the
complement-depleting activity of hybrid proteins H1, H2, and H5. Please note the significantly reduced complement-depleting activity of hybrid protein H5,
indicating the importance of the C-terminal region of the CVF b-chain for forming a stable convertase.

β−chain C3a α−chain Human C3

HC3-1550

HC3-1496

HC3-1348

α−chain γ−chain β−chain CVF

C3-like form

α-chain

S S

β-chain

C3b-like form

S S

β-chain

Fig. 11. Schematic representation of the chain structures of humanized CVF. The upper panel shows the chain structures of human pro-C3, pro-CVF, and three
humanized CVF proteins. Pro-C3 sequences (white) and pro-CVF sequences (gray) are indicated. The locations of homologous chains are shown. The lower panel
shows the chain structures of the C3-like and C3b-like forms of humanized CVF protein HC3-1496 as produced in Drosophila S2 cells. The gray areas indicate CVF
sequence at the C-terminus of the a-chain or a0 -chain, respectively.
 C.-W. Vogel, D.C. Fritzinger / Toxicon 56 (2010) 1198–1222 1211

A A
100 100

80 80

% Complement depleted
% Factor B cleaved

60
60
CVF
HC3-1550
40 HC3-1496
HC3-1348 40

20
20

0 CVF
0 5 10 15 20 25 HC3-1496
0
Time (hr)s 0 50 100 150 200
B 60 Protein (ng)
B
100
50

% Complement depleted
80
% C3 cleaved

40

30 60

20 40
CVF
10 HC3-1550
HC3-1496 20
HC3-1348
0 CVF
HC3-1496
0 0.2 0.4 0.6 0.8 1 0
Time (hrs) 0 50 100 150 200

C 100 HC3-1348, HC3-1496

C
Protein (ng)
CVF
100
% Convertase remaining

80
% Complement depleted

HC3-1550
80
60
60
40
C3b 40

20
20
CVF
HC3-1496
0
0
0 10 20 30 40 50 60 70 80 0 10 20 30 40 50
Time (min)
Protein (ng)
Fig. 12. Functional properties of humanized CVF and humanized CVF-con-
Fig. 13. In vitro complement depletion by humanized CVF in serum from
taining convertases. Panel A shows the activation of human factor B (as
different species. Shown are dose response curves in human (Panel A),
monitored by the appearance of the Bb fragment) in the presence of factor D
cynomolgus monkey (Panel B), and rat serum (Panel C).
and Mg2þ by several humanized CVF proteins or CVF. Panel B shows a time
course of cleavage of purified human C3 by convertases pre-formed with
several humanized CVF proteins or CVF with human factor B in the presence 3.3. Use of humanized CVF for therapeutic complement
of factor D and Mg2þ. Panel C shows the decay–dissociation of convertases
depletion in pre-clinical models of human disease
pre-formed with several humanized CVF proteins, CVF, or human C3b by
surface plasmon resonance tracing at 25  C. Please note that the slow decay–
dissociations of the stable convertases formed with HC3-1496, HC3-1348, The humanized CVF protein HC3-1496 has been used in
and CVF are indistinguishable in the time frame shown. several pre-clinical studies to determine the therapeutic
effect of complement depletion. HC3-1496 is a human C3
derivative in which the C-terminal 168 amino acid residues
on neutrophils are not abrogated by conversion to the C5a- of the C3 a-chain have been replaced with the correspond-
des-Arg form by the action of carboxypeptidase N. As the ing 168 amino acid residues from the b-chain of CVF. Please
parent proteins for humanized CVF (CVF, recombinant CVF, note that in this stretch of 168 amino acids from CVF, 73
and human C3b) are all able to form a convertase that exhibits amino acids (43.5%) are identical to the human C3 sequence,
C5-cleaving activity, the lack of the C5-cleaving activity of the and 34 more amino acids (20.2%) represent conservative
convertases formed with humanized CVF is surprising. The replacements. Accordingly, only 61 amino acids (36.3%)
lack of the C5-cleaving activity of convertases formed with represent true CVF sequence replacements in the HC3-1496
humanized CVF resembles that of the convertases formed hybrid protein, corresponding to 5.7% (inclusive of conser-
with CVF from N. haje and recombinant CVF with a C-terminal vative replacement) or 3.7% (exclusive of conservative of
His-tag (CVF-His) (see Section 2.2.8. above). replacements) of the total number of 1637 amino acids of
1212 C.-W. Vogel, D.C. Fritzinger / Toxicon 56 (2010) 1198–1222

A 100
80
% Hemolytic activity

60

40

20

0
0 1 2 3 4 5 6
Time after injection (hrs)

B 100
80
% Hemolytic activity

60

40

20

0
0 10 20 30 40 50
Time after injection (hrs)

C 100 1600

1400
80
1200
HC3-1496 (ng/ml)
% Hemolytic activity

60 1000

800
40 600

400
20
Complement activity
200
HC3-1496 (ng/ml)
0 0 Fig. 15. Lack of C5-cleaving activity by convertases containing humanized CVF.
0 5 10 15 20 25
Panel A shows the lack of C5-cleaving activity of convertases pre-formed with
Time after injection (hrs)
several humanized CVF proteins when incubated with purified human C5 as
measured by the conversion of the C5 a-chain into the C5 a0 -chain. Panel B
Fig. 14. In vivo complement depletion by humanized CVF in different
shows the lack of C5-cleaving activity in guinea pig serum of several humanized
species. Animals were injected with humanized CVF protein HC3-1496 or
CVF proteins as measured by the fluid-phase cleavage of C5 and subsequent
CVF either intra-arterially (cynomolgus monkeys, Panel A) or i.p. (rats, Panel
bystander lysis of unsensitized guinea pig erythrocytes. Panel C shows the lack
B; mice, Panel C) at dosages as shown, and serum complement hemolytic
of C5-cleaving activity of humanized CVF protein HC3-1496 in three cyn-
activity was determined at time intervals as indicated. For the determination
omolgus monkey sera as measured by the generation of the C5a anaphylatoxin
of the plasma levels of HC3-1496 in mice, biotinylated HC3-1496 was used at
using an ELISA assay. Please note that HC3-1496 generates the C3a anaphyla-
a dose of 500 mg/kg (Panel C).
toxin in monkey sera (see Fig. 20A), consistent with the C3-cleaving activity of
the HC3-1496-containing convertase (see Fig. 20A).
human C3. The convertase formed with HC3-1496 is more
stable than CVF,Bb, exhibiting a half-life of 31.2 h at 25  C
compared to 19.2 h at 25  C for CVF,Bb (see Fig.12C). The rate humanized CVF protein HC3-1496 (250 mg/kg) prior to
of convertase formation is very similar to that observed for inducing myocardial ischemia had a protective effect on
CVF (see Fig. 12A), and the convertase HC3-1496,Bb exhibits reperfusion injury. The protective effect of complement
a C3-cleaving activity that is even higher than that of CVF,Bb depletion was demonstrated functionally (better ejection
(see Fig. 12b) (Vogel and Fritzinger, 2007). Because of these fraction), morphologically (smaller infarct size), and
very favorable properties, HC3-1496 has been the primary immunohistochemically (less deposition of C3b) (Fig. 16)
candidate of humanized CVF for use in pre-clinical studies. (Gorsuch et al., 2009).

3.3.1. Therapeutic complement depletion with humanized CVF 3.3.2. Complement depletion with humanized CVF reduces
decreases myocardial ischemia reperfusion injury (MI/R) tissue damage in age-related macular degeneration (AMD)
In a mouse model of myocardial ischemia reperfusion In a mouse model of age-related macular degeneration
injury, complement depletion by i.p. injection of the that utilizes laser-induced photocoagulation of the Bruch’s
 C.-W. Vogel, D.C. Fritzinger / Toxicon 56 (2010) 1198–1222 1213

Fig. 16. Effect of complement depletion with humanized CVF in a mouse model of myocardial ischemia reperfusion injury. Mice were decomplemented by
a single i.p. injection of 250 mg/kg humanized CVF protein HC3-1496 prior to ligation for 30 min of the left anterior descending coronary artery. The protective
effect of complement depletion is demonstrated by a reduced infarct size (as measured by infarct size as percentage the total left ventricle area) (Panel A), by
maintaining a better ejection fraction (Panel B), and by a reduced deposition of C3b (Panel C).
1214 C.-W. Vogel, D.C. Fritzinger / Toxicon 56 (2010) 1198–1222

Fig. 17. Effect of complement depletion with humanized CVF in a mouse model of age-related macular degeneration. Mice were complement depleted by i.p.
injection of a low dose (25 mg/kg) of humanized CVF protein HC3-1496 prior to and daily for 28 days after laser photocoagulation of Bruch’s membrane.
PBS-treated mice served as control. The upper panel shows fundunscopic images after i.v. injection of fluorescein isothiocyanate (FTIC)-dextran on days
0 and 8 of the study. Please note the smaller lesions in HC3-1496-treated mice. The lower panel shows the size of the lesions on day 28 as determined
by histopathological examination of fixated sections of the eyes. Please note the smaller lesions in mice treated with HC3-1496 compared to PBS-treated
mice.

membrane and choroidal neovascularization, complement 3.3.4. Complement depletion with humanized CVF protects
depletion by i.p. injection of HC3-1496 at a very low dose human paroxysmal nocturnal hemoglobinuria (PNH)
(25 mg/kg) prior to and daily for 28 days after laser surgery erythrocytes from lysis
resulted in smaller lesions (Fig. 17). The low dose of HC3- Paroxysmal nocturnal hemoglobinuria (PNH) is a rare
1496 used in this study resulted in an average reduction of but potentially life-threatening disease in which the
about 30% complement hemolytic activity on day 28, with patient’s red blood cells are subject to complement-
significant animal to animal variation (Fritzinger et al., induced lysis because of the absence of the protective
2010b). complement regulatory proteins MCP (CD55) and CD59 on
the red blood cell surface, the result of a defect in the gly-
3.3.3. Complement depletion with humanized CVF reduces cosylphosphatidylinositol (GPI) anchor that normally
pathological changes in arthritis anchors the complement regulatory proteins in the blood
In a murine model of collagen-induced arthritis (Kerwar cell membrane. Incubation of PNH red blood cells from
et al., 1981), complement, depletion with an i.p. loading human patients in the presence of a recombinant truncated
dose of 500 mg/kg of humanized CVF HC3-1469 either prior form of factor H (rH19-20) blocks the cell regulatory
to the booster immunization with collagen (prophylactic functions and makes the PNH cells highly susceptible to
regimen) or six days after the booster immunization complement lysis in normal serum. When the PNH cells
(therapeutic regimen), and subsequently with a mainte- were incubated with serum that had been either partially
nance dose of 250 mg/kg on 5 days/week for two or three or completely depleted by prior treatment with the
more weeks, respectively, resulted in a significant reduc- humanized CVF protein HC3-1496, lysis was completely
tion of the paw and ankle swelling in both regimens prevented (Table 4). A reduction of 63% of serum comple-
(Fig. 18) (Fritzinger et al., 2008b). ment activity was sufficient to completely inhibit lysis of
 C.-W. Vogel, D.C. Fritzinger / Toxicon 56 (2010) 1198–1222 1215

17

16

15
Sum of Diameters (mm)

14

13

12

Vehicle
11 HC3-1496 (Prophylactic)
HC3-1496 (Therapeutic)

10
20 25 30 35
Days after Immunization

Fig. 18. Effect of complement depletion with humanized CVF in a mouse model of collagen-induced arthritis. Mice were injected i.p. with a loading dose of
500 mg/kg of humanized CVF protein HC3-1496 on the day of the booster immunization with collagen (prophylactic regimen) or six days after the booster
immunization (therapeutic regimen), and subsequently with an i.p. maintenance dose of 250 mg/kg on 5 days/week for two or three more weeks, respectively.
Arthritic changes were monitored by measuring the swelling of hind paws, fore paws, and ankles. The data shown represent the sum of the three measurements.

the PNH cells, demonstrating that only a partial reduction complement activation may actually be limiting the ther-
of the C3 concentration in serum is sufficient to reduce the apeutic effect of rituximab, we investigated the effect of
number of C5 convertases on the surface of the PNH complement depletion on human NK cell activation by
erythrocytes to the point that no lysis occurs (Fritzinger rituximab-coated target cells in vitro and on the therapeutic
et al., 2008a). efficacy of a monoclonal antibody in a murine model of
lymphoma in vivo. Depletion of complement prevented the
3.3.5. Complement depletion with humanized CVF enhances the inhibitory effect on NK cell activation in vitro (Wang et al.,
activation of human NK cells by the anti-lymphoma monoclonal 2009). Most strikingly, mice treated three days after the
antibody rituximab and induces survival in a mouse model of inoculation of the 38C13 tumor cells and 4 h prior to the
monoclonal antibody therapy for lymphoma injection of the MS11G6 anti-lymphoma monoclonal anti-
The therapeutic effect of monoclonal antibodies in body with an i.p. dose of 400 mg/kg of HC3-1496 and
tumor therapy can be due to multiple cytotoxic mecha- a second dose of 400 mg/kg HC3-1496 two days later
nisms, including complement activation and antibody- exhibited an 80% survival rate. In contrast, all untreated
dependent cellular cytotoxicity (ADCC). The chimeric anti- mice and all mice treated with either the antibody or HC3-
CD20 monoclonal antibody rituximab is a widely used and 1496 alone died (Fig. 19) (Wang et al., 2009).
clinically effective monoclonal antibody in the treatment of
B cell lymphomas (Anderson et al., 1997). Whereas in pre-
clinical models both complement activation and ADCC have
been shown to contribute to the anti-tumor activity of the
rituximab antibody, there is increasing evidence that ADCC
plays a significant role in the clinical effectiveness of rit-
uximab (Wang et al., 2009, and references therein). It was
recently described that complement interferes with the
binding of NK cells to rituximab, preventing both activation
of NK cells and inhibiting NK cell-mediated lysis of ritux-
imab-coated target cells (Wang et al., 2008). As these
results suggest that ADCC is the more important mecha-
nism for the clinical activity of rituximab, and that

Fig. 19. Effect of complement depletion with humanized CVF on the ther-
Table 4
apeutic efficacy of monoclonal antibody therapy in a syngeneic mouse
Lysis of PNH cells with CVF or humanized CVF-treated human serum.
model of lymphoma. Mice were complement depleted with a dose of
Treatment % Hemolytic activity remaining % Cells Lysed 400 mg/kg of humanized CVF protein HC3-1496 injected i.p. three days after
the inoculation of the 38C13 tumor cells and 4 h prior to the injection of the
None 100% 84%
MS11G6 anti-lymphoma monoclonal antibody, and with a second dose of
HC3-1496 (high) 17% 5%
HC3-1496 two days later. Please note that 80% of complement-depleted
HC3-1496 (low) 37% 5%
mice treated with antibody survived whereas all control mice, including
CVF 14% 0%
those treated with HC3-1496 or the antibody alone, died.
1216 C.-W. Vogel, D.C. Fritzinger / Toxicon 56 (2010) 1198–1222

Collectively, complement depletion with humanized specificity of CVF. Accordingly, any possible harmful side
CVF demonstrated significant benefit in every pre-clinical effects could either be acute complications from the
model of human disease with complement pathogenesis so massive activation of complement during the rapid deple-
far tested. tion, or a consequence of a prolonged state of being
complement depleted.
3.4. Potential toxicity of complement depletion with As mentioned above (see Section 2.2.3), the only acute
humanized CVF side effect observed after intravascular activation of
complement by CVF is a C5a-mediated transitory inflam-
The overall safety of depleting complement in vivo by mation of the lung due to the sequestration of activated
injection of CVF is demonstrated by its use for over four neutrophils. This lung inflammation is transitory, likely due
decades as a tool to deplete complement in laboratory to the fact that there is no intrinsic pulmonary reason for
animals, from mice to monkeys, to study the role of neutrophil activation and sequestration. The occurrence
complement in host defense and pathogenesis of disease. and extent of this inflammatory lung injury is easily
All known effects of CVF are mediated by binding to its controlled by the rate as well as the amount of the C5a
“receptor”, factor B and the resulting activation of generated. In contrast to CVF, humanized CVF exhibits
complement. No off-target activity or side effects have been virtually no C5-cleaving activity (see Section 3.2.2. above).
observed, nor is it expected given the high degree of Accordingly, virtually no C5a is generated during

A 45,000
40,000

35,000

30,000
C3a (ng/ml)

25,000

20,000

15,000

10,000

5,000

0
0 50 100 150 200 250 300 350 400
Time after injection (min)

B 45

40

35

30
C5a (ng/ml)

25

20

15

10

0
0 50 100 150 200 250 300 350 400

Time after injection (min)

Fig. 20. Lack of in vivo C5-cleaving activity of convertases formed with humanized CVF. Shown is the generation of the C3a (panel A) and C5a (panel B) ana-
phylatoxins in cynomolgus monkeys after injection of humanized CVF protein HC3-1496 at 250 mg/kg or 1000 mg/kg into the arteria pulmonalis as measured by
ELISA. C5a levels were near the limit of detection of the assay (please note the 1000-fold difference in the scale of the Y-axis in the two panels).
 C.-W. Vogel, D.C. Fritzinger / Toxicon 56 (2010) 1198–1222 1217

Lung Compliance Mean Airway Pressure

10

3.9

3.7 8

Mean Airway Pressure (cm H O)

Cdyn (ml/cm H O/sec)

3.5
6
3.3

3.1 4

2.9
2
2.7

2.5 0
0 2 4 6 8 10 0 2 4 6 8 10
Time after injection (min) Time after injection (min)

Tidal Volume Lung Resistance

60
0.06
58

56
0.055

Resistance (cm H O/ml/sec)

54
Tidal Volume (ml)

52

50 0.05

48

46
0.045
44

42

40 0.04
0 2 4 6 8 10 0 2 4 6 8 10
Time after injection (min) Time after injection (min)

Fig. 21. Pulmonary function during complement depletion with humanized CVF in cynomolgus monkeys. Monkeys were injected intra-arterially into the arteria
pulmonaris with humanized CVF protein HC3-1496 at 250 mg/kg or 1000 mg/kg. Physiological lung parameters were continuously monitored. Note the absence of
any impairment of pulmonary function.

complement activation by humanized CVF. In order to Other adverse side effects could be a consequence of
assess any potential acute lung damage or any other acute being in a state of prolonged complement depletion. No
side effects of intravascular complement activation by such side effects have been reported in the numerous
humanized CVF, we investigated the effect of injection of studies involving complement depletion in laboratory
humanized CVF at 250 mg/kg and even 1000 mg/kg by intra- animals. Most of these studies involved complement
arterial injection into the arteria pulmonalis of cynomolgus depletion for a few days, up to one month (compare Figs. 17
monkeys (Fritzinger et al., 2008c). The injection of either and 18). In humans, genetic deficiencies of virtually all
dose resulted in rapid complement depletion within complement components have been described (Botto et al.,
minutes (see Fig. 14A above). Fig. 20A shows the generation 2009; Pettigrew et al., 2009; Sjoholm et al., 2006), with
of C3a and its rapid removal from the circulation. Note the homozygous C3 deficiency exhibiting some of the most
absence of any measurable C5a generation (Fig. 20B). severe pathology in the form of recurrent gram-positive
Fig. 21 shows that multiple physiological lung parameters infections, pointing to the important role of opsonization
were not affected by the rapid complement depletion by for host defense against these microorganisms (Botto et al.,
humanized CVF. At the very high dose of 1000 mg/kg, 2009; Botto and Walport, 1993). Similarly, C3 knockout
a transient increase in the heart rate and systolic blood mice showed an increased susceptibility to infections, but
pressure was observed, although the causative relationship no other overt pathology (Circolo et al., 1999; Sylvestre
to complement activation remains to be established et al., 1996). As described above (see Section 2.2.6) CVF
(Fig. 22). In summary, complement activation by human- transgenic mice reproduce normally, exhibiting a normal
ized CVF is well tolerated and appears to be safe. life span and displaying no abnormal phenotype (Andrä
The only potential risk would be in individuals with et al., 2002). During approximately one decade of
a deficiency of carboxypeptidase N. No complete carboxy- continued maintenance under normal animal housing
peptidase N deficiency has been described. Only two conditions, no tendency to develop infections has been
siblings with 21% of normal carboxypeptidase N activity observed (Fritzinger et al., 2010b). Residual C3 levels as well
and a single patient with very low levels (3%) have been as the ability to synthesize C3 locally by macrophages and
described (Mathews et al., 1980; Willemse et al., 2008). All in the liver are apparently sufficient to protect against
three patients presented with episodical angioedema, but infections. Collectively, these data suggest that at least
no overt phenotype. Therapeutic complement depletion short-term or even extended complement depletion as
would be contraindicated in these extremely rare well as episodical complement depletion may be safe in
individuals. humans.
1218 C.-W. Vogel, D.C. Fritzinger / Toxicon 56 (2010) 1198–1222

Systolic blood pressure

180

160
Systolic pressure (cm H2O)

140

120

100

80

60
0 2 4 6 8 10
Time after injection (min)

Heart rate

220

200
Heart rate

180

160

140
0 2 4 6 8 10
Time after injection (min)

Fig. 22. Cardiac function during complement depletion with humanized CVF in monkeys. Monkeys were injected intra-arterially into the arteria pulmonaris with
humanized CVF protein HC3-1496 at 250 mg/kg or 1000 mg/kg. Physiological cardiac parameters were continuously monitored. No changes in heart rate or systolic
blood pressure were observed at the 250 mg/kg dose. At the 1000 mg/kg dose, a transient increase in both heart rate and systolic pressure was observed.

A potential problem that would limit the use of human- structural changes in a protein after each single amino acid
ized CVF, particularly for extended and repeated applications, residue change, the three-dimensional structures of the
is the potential immunogenicity, the extent of which is C345C domain of C3c and humanized CVF were virtually
unpredictable. However, humanized CVF is a derivative superimposable (Fritzinger et al., 2008b).
human C3 with an overall sequence identity of approximately Immunogenicity may also result from carbohydrate
94%. Amino acid differences are limited to the very C- chains as a result of recombinant production. However, any
terminus of the a-chain; and even in this region, approxi- form of recombinantly produced CVF in insect cells,
mately 43% of the CVF sequence is identical to human C3 (see Chinese hamster ovary (CHO) cells, or human embryonic
Section 3.3. above). Moreover, the CVF-specific residues are kidney (HEK) cells will be devoid of the unusual and
not contiguous but interspersed throughout the exchanged immunogenic carbohydrate chains of native CVF.
sequence. Importantly, CVF is not a structurally unrelated
protein to C3; rather, it is structurally highly homologous. The 3.5. Outlook and challenges
exchanged region contains the C-terminal C345C domain
which exhibits the same three-dimensional structure in both Humanized CVF represents a promising biological for
C3 and CVF (see Figs. 4 and 11). Using a program that predicts therapeutic complement depletion with potential
 C.-W. Vogel, D.C. Fritzinger / Toxicon 56 (2010) 1198–1222 1219

application in many different diseases. The clinical Bauman, N., 1978. Lack of C5 convertase-generating activity in Naja haje
cobra factor. J. Immunol. 120, 1763–1764.
usefulness in any given disease with complement patho-
Birdsey, V., Lindorfer, J., Gewurz, H., 1971. Interaction of toxic venoms with
genesis is unpredictable; however, given the potentially the complement system. Immunology 21, 299–310.
large number of diseases and the very encouraging result Botto, M., 2000. C3. In: Morley, B.J., Walport, M.J. (Eds.), The Complement
from all pre-clinical studies, it appears likely that Facts Book. Academic Press, New York, pp. 87–94.
Botto, M., Kirschfink, M., Macor, P., Pickering, M.C., Wurzner, R., Tedesco, F.,
complement depletion with humanized CVF will be an 2009. Complement in human diseases: lessons from complement
effective clinical tool for many diseases. Challenges deficiencies. Mol. Immunol. 46, 2774–2783.
include the relatively short time of complement depletion Botto, M., Walport, M.J., 1993. Hereditary deficiency of C3 in animals and
humans. Int. Rev. Immunol. 10, 37–50.
after injection of humanized CVF. While this might not de Bruijn, M.H., Fey, G.H., 1985. Human complement component C3:
represent a problem - or may even be a beneficial prop- cDNA coding sequence and derived primary structure. Proc. Natl.
erty – for clinical applications requiring short-term Acad. Sci. U.S.A. 82, 708–712.
Buyon, J.P., Tamerius, J., Ordorica, S., Young, B., Abramson, S.B., 1992.
complement depletion (e.g., reperfusion injury), it limits Activation of the alternative complement pathway accompanies
applications requiring prolonged periods of complement disease flares in systemic lupus erythematosus during pregnancy.
depletion. These applications will require different Arthritis Rheum. 35, 55–61.
Campbell, R.D., Law, S.K., Reid, K.B., Sim, R.B.,1988. Structure, organization, and
formulations and/or routes of applications (e.g., slow regulation of the complement genes. Annu. Rev. Immunol. 6, 161–195.
release formulations) or novel forms of humanized CVF Caporale, L.H., Gaber, S.S., Kell, W., Götze, O., 1981. A fluorescent assay for
with extended plasma half-lives. complement activation. J. Immunol. 126, 1963–1965.
Circolo, A., Garnier, G., Fukuda, W., Wang, X., Hidvegi, T., Szalai, A.J.,
Briles, D.E., Volanakis, J.E., Wetsel, R.A., Colten, H.R., 1999. Genetic
Ethical statement disruption of the murine complement C3 promoter region generates
deficient mice with extrahepatic expression of C3 mRNA. Immuno-
pharmacology 42, 135–149.
The authors (Carl-Wilhelm Vogel and David C. Fritzinger) Cochrane, C.G., 1984. The role of complement in experimental disease
declare that any animals used in this work were treated as models. Springer Semin. Immunopathol. 7, 263–270.
required by the appropriate Institutional Animal Care and Use Cochrane, C.G., Müller-Eberhard, H.J., Aikin, B.S., 1970. Depletion of
plasma complement in vivo by a protein of cobra venom: its effect on
Committee (IACUC). All data shown are an accurate repre- various immunologic reactions. J. Immunol. 105, 55–69.
sentation of results obtained by us, or by our co-workers, and Dalmasso, A.P., 1992. The complement system in xenotransplantation.
their work is acknowledged in references provided. Immunopharmacology 24, 149–160.
DiScipio, R.G., Smith, C.A., Müller-Eberhard, H.J., Hugli, T.E., 1983. The
activation of human complement component C5 by a fluid phase C5
Conflict of interest statement convertase. J. Biol. Chem. 258, 10629–10636.
Edwards, A.O., Ritter 3rd, R., Abel, K.J., Manning, A., Panhuysen, C.,
Farrer, L.A., 2005. Complement factor H polymorphism and age-
The authors both own stock in Incode Biopharmaceutics, related macular degeneration. Science 308, 421–424.
Inc., which partially supported this work. Eggertsen, G., Lind, P., Sjöquist, J., 1981. Molecular characterization of the
complement activating protein in the venom of the Indian cobra
(Naja N. siamensis). Mol. Immunol. 18, 125–133.
Eggertsen, G., Lundwall, A., Hellman, U., Sjöquist, J., 1983. Antigenic
References relationships between human and cobra complement factors C3 and
cobra venom factor (CVF) from the Indian cobra (Naja naja). J.
Alper, C.A., Balavitch, D., 1976. Cobra venom factor: evidence for its being Immunol. 131, 1920–1923.
altered cobra C3 (the third component of complement). Science 191, Flexner, S., Noguchi, H., 1903. Snake venom in relation to haemolysis,
1275–1276. bacteriolysis, and toxicity. J. Exp. Med. 6, 277–301.
Alsenz, J., Avila, D., Huemer, H.P., Esparza, I., Becherer, J.D., Kinoshita, T., Friedberger, E., Mita, S., Kumagai, T., 1913. Die Bildung eines akut wir-
Wang, Y., Oppermann, S., Lambris, J.D., 1992. Phylogeny of the third kenden Giftes (Anaphylatoxin) aus Toxinen (Tetanus, Diphtherie,
component of complement, C3: analysis of the conservation of Schlangengift). Z. Immunitätsforsch. 17, 506–538.
human CR1, CR2, H, and B binding sites, concanavalin A binding sites, Fritzinger, D.C., Petrella, E.C., Connelly, M.B., Bredehorst, R., Vogel, C.-W.,
and thiolester bond in the C3 from different species. Dev. Comp. 1992. Primary structure of cobra complement component C3. J.
Immunol. 16, 63–76. Immunol. 149, 3554–3562.
Anderson, D.R., Grillo-Lopez, A., Varns, C., Chambers, K.S., Hanna, N., 1997. Fritzinger, D.C., Bredehorst, R., Vogel, C.-W., 1994. Molecular cloning and
Targeted anti-cancer therapy using rituximab, a chimaeric anti-CD20 derived primary structure of cobra venom factor. Proc. Natl. Acad. Sci.
antibody (IDEC-C2B8) in the treatment of non-Hodgkin’s B-cell U.S.A. 91, 12775–12779.
lymphoma. Biochem. Soc. Trans. 25, 705–708. Fritzinger, D.C., Ferreira, V.P., Hew, B.E., St. John, W., Pangburn, M.K.,
Andrä, J., Halter, R., Kock, M.A., Niemann, H., Vogel, C.-W., Paul, D., 2002. Vogel, C.-W., 2008a. A novel concept for the treatment of paroxysmal
Generation and characterization of transgenic mice expressing cobra nocturnal haemoglobinuria (PNH): complement depletion with
venom factor. Mol. Immunol. 39, 357–365. a human C3 derivative with cobra venom factor-like activity prevents
Arroyave, C.M., Stevenson, D.D., Vaughan, J.H., Tan, E.M., 1977. Plasma lysis of PNH erythrocytes. Mol. Immunol. 45, 4177.
complement changes during bronchospasm provoked in asthmatic Fritzinger, D.C., Hew, B.E., Lee, J.Q., Newhouse, J., Alam, M., Ciallella, J.R.,
patients. Clin. Allergy 7, 173–182. Bowers, M., Gorsuch, W.B., Guikema, B.J., Stahl, G.L., Vogel, C.-W.,
Arumugam, T.V., Shiels, I.A., Woodruff, T.M., Granger, D.N., Taylor, S.M., 2008b. Derivatives of human complement component C3 for thera-
2004. The role of the complement system in ischemia-reperfusion peutic complement depletion: a novel class of therapeutic agents.
injury. Shock 21, 401–409. Adv. Exp. Med. Biol. 632, 293–307.
Bambai, B., Teppke, M., Bredehorst, R., Vogel, C.-W., 1998. cDNA cloning Fritzinger, D.C., Hew, B.E., Lee, J.Q., St. John, W., Scaife, M., Wilson, S.,
and expression of cobrin, the C3-cleaving metalloprotease from cobra Vogel, C.-W., 2008c. Human C3/cobra venom factor hybrid proteins
venom. Mol. Immunol. 35, 408. for therapeutic complement depletion: in vivo activity and lack of
Bammert, H., Kunze, B., Li, Y., Fritzinger, D.C., Bredehorst, R., Vogel, C.-W., toxicity in primates. Mol. Immunol. 45, 4112.
2002. Structure of the cobra venom factor (CVF) gene and compar- Fritzinger, D.C., Hew, B.E., Thorne, M., Pangburn, M.K., Janssen, B.J., Gros, P.,
ison with the human C3 gene. Intern. Immunopharmacol. 2, Vogel, C.-W., 2009. Functional characterization of human C3/cobra
1275–1276. venom factor hybrid proteins for therapeutic complement depletion.
Banda, N.K., Levitt, B., Glogowska, M.J., Thurman, J.M., Takahashi, K., Dev. Comp. Immunol. 33, 105–116.
Stahl, G.L., Tomlinson, S., Arend, W.P., Holers, V.M., 2009. Targeted Fritzinger, D.C., Damaj, B.B., Wong, K., Vogel, C.-W., Finnegan, P.W., 2010a.
inhibition of the complement alternative pathway with complement Pharmacokinetics of humanized cobra venom factor in mice. Mol.
receptor 2 and factor h attenuates collagen antibody-induced arthritis Immunol. in press.
in mice. J. Immunol. 183, 5928–5937.
1220 C.-W. Vogel, D.C. Fritzinger / Toxicon 56 (2010) 1198–1222

Fritzinger, D.C., Dean R., Meschter C., Wong K., Halter R., Borlak J., St. John Janssen, B.J., Huizinga, E.G., Raaijmakers, H.C., Roos, A., Daha, M.R., Nils-
W.D., Vogel C.-W. 2010b. Complement depletion with humanized son-Ekdahl, K., Nilsson, B., Gros, P., 2005. Structures of complement
cobra venom factor in a mouse model of age-related macular component C3 provide insights into the function and evolution of
degeneration. Adv. Exp. Med. Biol. in press. immunity. Nature 437, 505–511.
Fu, Q, Satyaswaroop, P.G., Gowda, D.C., 1997. Tissue targeting and plasma Janssen, B.J., Christodoulidou, A., McCarthy, A., Lambris, J.D., Gros, P., 2006.
clearance of cobra venom factor in mice. Biochem. Biophys. Res. Structure of C3b reveals conformational changes that underlie
Commun 231, 316–320. complement activity. Nature 444, 213–216.
Fujita, T., 2002. Evolution of the lectin-complement pathway and its role Janssen, B.J., Gomes, L., Koning, R.I., Svergun, D.I., Koster, A.J., Fritzinger, D.
in innate immunity. Nat. Rev. Immunol. 2, 346–353. C., Vogel, C.-W., Gros, P., 2009. Insights into complement convertase
Gorsuch, W.B., Guikema, B.J., Fritzinger, D.C., Vogel, C.-W., Stahl, G.L., formation based on the structure of the factor B-cobra venom factor
2009. Humanized cobra venom factor decreases myocardial complex. EMBO J. 28, 2469–2478.
ischemia-reperfusion injury. Mol. Immunol. 47, 506–510. Jordon, R.E., Day, N.K., Sams Jr., W.M., Good, R.A., 1973. The complement
Götze, O., Müller-Eberhard, H.J., 1971. The C3-activator system: an alter- system in bullous pemphigoid. I. Complement and component levels
nate pathway of complement activation. J. Exp. Med. 134, 90s–108s. in sera and blister fluids. J. Clin. Invest. 52, 1207–1214.
Gowda, D.C., Schultz, M., Bredehorst, R., Vogel, C.-W., 1992. Structure of the Kerwar, S.S., Bauman, N., Oronsky, A.L., Sloboda, A.E., 1981. Studies on type
major oligosaccharide of cobra venom factor. Mol. Immunol. 29, 335–342. II collagen induced polyarthritis in rats. Effect of complement
Gowda, D.C., Petrella, E.C., Raj, T.T., Bredehorst, R., Vogel, C.-W., 1994. depletion. J. Immunopharmacol. 3, 323–337.
Immunoreactivity and function of oligosaccharides in cobra venom Klos, A., Ihrig, V., Messner, M., Grabbe, J., Bitter-Suermann, D., 1988.
factor. J. Immunol. 152, 2977–2986. Detection of native human complement components C3 and C5 and
Gowda, D.C., Glushka, J., Halbeek, H., Thotakura, R.N., Bredehorst, R., their primary activation peptides C3a and C5a (anaphylatoxic
Vogel, C.-W., 2001. N-linked oligosaccharides of cobra venom factor peptides) by ELISAs with monoclonal antibodies. J. Immunol.
contain novel alpha(1-3)galactosylated Lex structures. Glycobiology Methods 111, 241–252.
11, 195–208. Klos, A., Tenner, A.J., Johswich, K.O., Ager, R.R., Reis, E.S., Kohl, J., 2009. The
Grier, A.H., Schultz, M., Vogel, C.-W., 1987. Cobra venom factor and human role of the anaphylatoxins in health and disease. Mol. Immunol. 46,
C3 share carbohydrate antigenic determinants. J. Immunol. 139, 1245– 2753–2766.
1252. Kock, M.A., 1996. Expression and Characterization of Recombinant Cobra
Gros, P., Milder, F.J., Janssen, B.J., 2008. Complement driven by confor- Venom Factor. Ph.D. thesis, University of Hamburg.
mational changes. Nat. Rev. Immunol. 8, 48–58. Kock, M.A., Hew, B.E., Bammert, H., Fritzinger, D.C., Vogel, C.-W., 2004.
Grunwald, T., Ziegelmüller, P., Ollert, M.W., Bredehorst, R., Vogel, C.-W., Structure and function of recombinant cobra venom factor. J. Biol.
1993. Cobra venom factor: an intact b-chain is not required for Chem. 279, 30836–30843.
activity. Mol. Immunol. 30, 13. Kohl, J., 2006. Drug evaluation: the C5a receptor antagonist PMX-53. Curr.
Grunwald, T., Bredehorst, R., Vogel, C.-W., 1996. Cobra complement factor B: Opin. Mol. Ther. 8, 529–538.
evidence for additional function as CVF-inhibitor. Mol. Immunol. 33, 14. Kölln, J., Spillner, E., Andrä, J., Klensang, K., Bredehorst, R., 2004.
Hageman, G.S., Anderson, D.H., Johnson, L.V., Hancox, L.S., Taiber, A.J., Complement inactivation by recombinant human C3 derivatives. J.
Hardisty, L.I., Hageman, J.L., Stockman, H.A., Borchardt, J.D., Gehrs, K. Immunol. 173, 5540–5545.
M., Smith, R.J., Silvestri, G., Russell, S.R., Klaver, C.C., Barbazetto, I., Kölln, J., Bredehorst, R., Spillner, E., 2005. Engineering of human
Chang, S., Yannuzzi, L.A., Barile, G.R., Merriam, J.C., Smith, R.T., Olsh, A. complement component C3 for catalytic inhibition of complement.
K., Bergeron, J., Zernant, J., Merriam, J.E., Gold, B., Dean, M., Immunol. Lett. 98, 49–56.
Allikmets, R., 2005. A common haplotype in the complement regu- Krishnan, V., Ponnuraj, K., Xu, Y., Macon, K., Volanakis, J.E., Narayana, S.V.,
latory gene factor H (HF1/CFH) predisposes individuals to age-related 2009. The crystal structure of cobra venom factor, a cofactor for C3-
macular degeneration. Proc. Natl. Acad. Sci. U.S.A. 102, 7227–7232. and C5-convertase CVFBb. Structure 17, 611–619.
Haines, J.L., Hauser, M.A., Schmidt, S., Scott, W.K., Olson, L.M., Gallins, P., Lachmann, P.J., Halbwachs, L., 1975. The influence of C3b inactivator (KAF)
Spencer, K.L., Kwan, S.Y., Noureddine, M., Gilbert, J.R., Schnetz- concentration on the ability of serum to support complement acti-
Boutaud, N., Agarwal, A., Postel, E.A., Pericak-Vance, M.A., 2005. vation. Clin. Exp. Immunol. 21, 109–114.
Complement factor H variant increases the risk of age-related Law, S.K.A., Reid, K.B.M., 1988. Complement. IRL Press, Oxford, OX,
macular degeneration. Science 308, 419–421. England ; Washington, DC.
Hebell, T., Ahearn, J.M., Fearon, D.T., 1991. Suppression of the immune Lennon, V.A., Seybold, M.E., Lindstrom, J.M., Cochrane, C., Ulevitch, R.,
response by a soluble complement receptor of B lymphocytes. Science 1978. Role of complement in the pathogenesis of experimental
254, 102–105. autoimmune myasthenia gravis. J. Exp. Med. 147, 973–983.
Hensley, P., O’Keefe, M.C., Spangler, C.J., Osborne Jr., J.C., Vogel, C.-W., 1986. Lesavre, P.H., Hugli, T.E., Esser, A.F., Müller-Eberhard, H.J., 1979. The
The effects of metal ions and temperature on the interaction of cobra alternative pathway C3/C5 convertase: chemical basis of factor B
venom factor and human complement factor B. J. Biol. Chem. 261, activation. J. Immunol. 123, 529–534.
11038–11044. Leventhal, J.R., Dalmasso, A.P., Cromwell, J.W., Platt, J.L., Manivel, C.J.,
Hew, B.E., Thorne, M., Fritzinger, D.C., Vogel, C.-W., 2004. Humanized Bolman 3rd, R.M., Matas, A.J., 1993. Prolongation of cardiac xenograft
cobra venom factor (CVF): generation of Human C3 derivitives with survival by depletion of complement. Transplantation 55, 857–865.
CVF-like function. Mol. Immunol. 41, 244–245. discussion 65–6.
Hillmen, P., Hall, C., Marsh, J.C., Elebute, M., Bombara, M.P., Petro, B.E., Li, J.S., Jaggers, J., Anderson, P.A., 2006. The use of TP10, soluble comple-
Cullen, M.J., Richards, S.J., Rollins, S.A., Mojcik, C.F., Rother, R.P., 2004. ment receptor 1, in cardiopulmonary bypass. Expert Rev. Cardiovasc.
Effect of eculizumab on hemolysis and transfusion requirements in Ther. 4, 649–654.
patients with paroxysmal nocturnal hemoglobinuria. N. Engl. J. Med. Linton, S.M., Morgan, B.P., 1999. Complement activation and inhibition in
350, 552–559. experimental models of arthritis. Mol. Immunol. 36, 905–914.
Holers, V.M., 2008. The spectrum of complement alternative pathway- Low, P.J., Ai, R., Ogata, R.T., 1999. Active sites in complement components
mediated diseases. Immunol. Rev. 223, 300–316. C5 and C3 identified by proximity to indels in the C3/4/5 protein
Holers, V.M., Girardi, G., Mo, L., Guthridge, J.M., Molina, H., Pierangeli, S.S., family. J. Immunol. 162, 6580–6588.
Espinola, R., Xiaowei, L.E., Mao, D., Vialpando, C.G., Salmon, J.E., 2002. Lundwall, Å., Hellman, U., Eggertsen, G., Sjöquist, J., 1984. Chemical
Complement C3 activation is required for antiphospholipid antibody- characterization of cyanogen bromide fragments from the beta-chain
induced fetal loss. J. Exp. Med. 195, 211–220. of human complement factor C3. FEBS Lett. 169, 57–62.
Holers, V.M., Thurman, J.M., 2004. The alternative pathway of comple- Lutz, H.U., Jelezarova, E., 2006. Complement amplification revisited. Mol.
ment in disease: opportunities for therapeutic targeting. Mol. Immunol. 43, 2–12.
Immunol. 41, 147–152. Maillard, J.L., Zarco, R.M., 1968. Decomplementization by a factor
Huey, R., Bloor, C.M., Kawahara, M.S., Hugli, T.E., 1983. Potentiation of the extracted from cobra venom. Effect on several immune reactions of
anaphylatoxins in vivo using an inhibitor of serum carboxypeptidase the guinea pig and rat. Ann. Inst. Pasteur (Paris) 114, 756–774.
N (SCPN). I. Lethality and pathologic effects on pulmonary tissue. Am. Markiewski, M.M., Lambris, J.D., 2007. The role of complement in inflamma-
J. Pathol. 112, 48–60. tory diseases from behind the scenes into the spotlight. Am. J. Pathol. 171.
Iwanaga, S., Suzuki, T., 1979. Enzymes in snake venom. In: Lee, C.Y. (Ed.), Mathews, K.P., Pan, P.M., Gardner, N.J., Hugli, T.E., 1980. Familial
Snake Venoms. Springer-Verlag, Berlin, New York, pp. 61–158. carboxypeptidase N deficiency. Ann. Intern. Med. 93, 443–445.
Janatova, J., 1986. Detection of disulphide bonds and localization of Medicus, R.G., Götze, O., Müller-Eberhard, H.J., 1976. The serine protease
interchain linkages in the third (C3) and the fourth (C4) components nature of the C3 and C5 convertases of the classical and alternative
of human complement. Biochem. J. 233, 819–825. complement pathways. Scand. J. Immunol. 5, 1049–1055.
 C.-W. Vogel, D.C. Fritzinger / Toxicon 56 (2010) 1198–1222 1221

Mehrtens, J.M., 1987. Living Snakes of the World in Color. Sterling Plummer Jr., T.H., Hurwitz, M.Y., 1978. Human plasma carboxypeptidase
Publishing Co., Inc., New York. N. Isolation and characterization. J. Biol. Chem. 253, 3907–3912.
Misumi, Y., Oda, K., Fujiwara, T., Takami, N., Tashiro, K., Ikehara, Y., 1991. Rawal, N., Pangburn, M.K., 1998. C5 convertase of the alternative pathway
Functional expression of furin demonstrating its intracellular locali- of complement. Kinetic analysis of the free and surface-bound forms
zation and endoprotease activity for processing of proalbumin and of the enzyme. J. Biol. Chem. 273, 16828–16835.
complement pro-C3. J. Biol. Chem. 266, 16954–16959. Rawal, N., Pangburn, M.K., 2000. Functional role of the noncatalytic
Mollnes, T.E., Lachmann, P.J., 1988. Regulation of complement. Scand. J. subunit of complement C5 convertase. J. Immunol. 164, 1379–1385.
Immunol. 27, 127–142. Rawal, N., Pangburn, M.K., 2001. Structure/function of C5 convertases of
Morgan, B.P., Harris, C.L., 1999. Complement Regulatory Proteins. complement. Int. Immunopharmacol. 1, 415–422.
Academic Press, San Diego. Rehana, S., Manjunatha Kini, R., 2007. Molecular isoforms of cobra venom
Morgan, B.P., Harris, C.L., 2003. Complement therapeutics; history and factor-like proteins in the venom of Austrelaps superbus. Toxicon 50,
current progress. Mol. Immunol. 40, 159–170. 32–52.
Müller-Eberhard, H.J., 1988. Molecular organization and function of the Ricklin, D., Lambris, J.D., 2007. Complement-targeted therapeutics. Nat.
complement system. Annu. Rev. Biochem. 57, 321–347. Biotechnol. 25, 1265–1275.
Müller-Eberhard, H.J., Fjellström, K.E., 1971. Isolation of the anti- Riedemann, N.C., Ward, P.A., 2003. Complement in ischemia reperfusion
complementary protein from cobra venom and its mode of action on injury. Am. J. Pathol. 162, 363–367.
C3. J. Immunol. 107, 1666–1672. Rooijakkers, S.H., Wu, J., Ruyken, M., van Domselaar, R., Planken, K.L.,
Mulligan, M.S., Schmid, E., Beck-Schimmer, B., Till, G.O., Friedl, H.P., Tzekou, A., Ricklin, D., Lambris, J.D., Janssen, B.J., van Strijp, J.A.,
Brauer, R.B., Hugli, T.E., Miyasaka, M., Warner, R.L., Johnson, K.J., Gros, P., 2009. Structural and functional implications of the alterna-
Ward, P.A., 1996. Requirement and role of C5a in acute lung inflam- tive complement pathway C3 convertase stabilized by a staphylo-
matory injury in rats. J. Clin. Invest. 98, 503–512. coccal inhibitor. Nat. Immunol. 10, 721–727.
Nagaki, K., Iida, K., Okubo, M., Inai, S., 1978. Reaction mechanisms of Rother, K., Till, G.O., Hänsch, G.M., 1998. The Complement System, second
beta1H globulin. Int. Arch. Allergy Appl. Immunol. 57, 221–232. ed. Springer, Berlin, New York.
Nelson Jr., R.A., 1966. A new concept of immunosuppression in hyper- Sahu, A., Lambris, J.D., 2000. Complement inhibitors: a resurgent concept
sensitivity reactions and in transplantation immunity. Surv. Oph- in anti-inflammatory therapeutics. Immunopharmacology 49, 133–
thalmol. 11, 498–505. 148.
Nishimura, J., Phillips, K.L., Ware, R.E., Hall, S., Wilson, L., Gentry, T.L., Sandoval, A., Ai, R., Ostresh, J.M., Ogata, R.T., 2000. Distal recognition site
Howard, T.A., Murakami, Y., Shibano, M., Machii, T., Gilboa, E., for classical pathway convertase located in the C345C/netrin module
Kanakura, Y., Takeda, J., Kinoshita, T., Rosse, W.F., Smith, C.A., 2001. of complement component C5. J. Immunol. 165, 1066–1073.
Efficient retrovirus-mediated PIG-A gene transfer and stable resto- Sharma, S., Jabeen, T., Kumar Singh, R., Bredehorst, R., Vogel, C.-W.,
ration of GPI-anchored protein expression in cells with the PNH Betzel, C., Singh, T.P., 2001. Structural studies on the cobra venom
phenotype. Blood 97, 3004–3010. factor: isolation, purification, crystallization and preliminary crystal-
O’Keefe, M.C., Vogel, C.-W., Caporale, L.H., 1984. Characterization of lographic analysis. Acta Cryst D57, 596–598.
a protease from cobra venom that cleaves the human complement Sjoholm, A.G., Jonsson, G., Braconier, J.H., Sturfelt, G., Truedsson, L., 2006.
component C3. Fed. Proc. 43, 1956. Complement deficiency and disease: an update. Mol. Immunol. 43,
O’Keefe, M.C., Caporale, L.H., Vogel, C.-W., 1987. Comparison of the alpha 78–85.
chain fragments of C3o, C3c, and CVF: implications for C3 convertase Slowinski, J., Wüster, W., 2000. A new cobra (Elapidae: naja) from
formation. Complement 4, 204. Myanmar (Burma). Herpetologica 56, 257–270.
O’Keefe, M.C., Caporale, L.H., Vogel, C.-W., 1988. A novel cleavage product Smith, C.A., Vogel, C.-W., Muller-Eberhard, H.J., 1982. Ultrastructure of
of human complement component C3 with structural and functional cobra venom factor-dependent C3/C5 convertase and its zymogen,
properties of cobra venom factor. J. Biol. Chem. 263, 12690–12697. factor B of human complement. J. Biol. Chem. 257, 9879–9882.
Osipov, A.V., Mordvintsev, D.Y., Starkov, V.G., Galebskaya, L.V., Ryumina, E. Smith, C.A., Vogel, C.-W., Müller-Eberhard, H.J., 1984. MHC Class III
V., Bel’tyukov, P.P., Kozlov, L.V., Romanov, S.V., Doljansky, Y., Tsetlin, V. products: an electron microscopic study of the C3 convertases of
I., Utkin, Y.N., 2005. Naja melanoleuca cobra venom contains two human complement. J. Exp. Med. 159, 324–329.
forms of complement-depleting factor (CVF). Toxicon 46, 394–403. Sylvestre, D., Clynes, R., Ma, M., Warren, H., Carroll, M.C., Ravetch, J.V.,
Pangburn, M.K., Schreiber, R.D., Müller-Eberhard, H.J., 1977. Human 1996. Immunoglobulin G-mediated inflammatory responses
complement C3b inactivator: isolation, characterization, and develop normally in complement-deficient mice. J. Exp. Med. 184,
demonstration of an absolute requirement for the serum protein 2385–2392.
beta1H for cleavage of C3b and C4b in solution. J. Exp. Med. 146, 257– Szalai, A.J., Digerness, S.B., Agrawal, A., Kearney, J.F., Bucy, R.P., Niwas, S.,
270. Kilpatrick, J.M., Babu, Y.S., Volanakis, J.E., 2000. The Arthus reaction in
Pangburn, M.K., Schreiber, R.D., Muller-Eberhard, H.J., 1981. Formation of rodents: species-specific requirement of complement. J. Immunol.
the initial C3 convertase of the alternative complement pathway. 164, 463–468.
Acquisition of C3b-like activities by spontaneous hydrolysis of the Tack, B.F., 1983. The beta-Cys-gamma-Glu thiolester bond in human C3,
putative thioester in native C3. J. Exp. Med. 154, 856–867. C4, and alpha 2-macroglobulin. Springer Semin. Immunopathol. 6,
Pangburn, M.K., Muller-Eberhard, H.J., 1983. Initiation of the alternative 259–282.
complement pathway due to spontaneous hydrolysis of the thioester Tack, B.F., Harrison, R.A., Janatova, J., Thomas, M.L., Prahl, J.W., 1980.
of C3. Ann. N.Y. Acad. Sci. 421, 291–298. Evidence for presence of an internal thiolester bond in third compo-
Pangburn, M.K., Muller-Eberhard, H.J., 1986. The C3 convertase of the nent of human complement. Proc. Natl. Acad. Sci. U.S.A. 77, 5764–5768.
alternative pathway of human complement. Enzymic properties of Takahashi, H., Hayashi, K., 1982. Purification and characterization of
the bimolecular proteinase. Biochem. J. 235, 723–730. anticomplement factor (cobra venom factor) from the Naja naja atra
Parker, C.J., Wiedmer, T., Sims, P.J., Rosse, W.F., 1985. Characterization of venom. Biochim. Biophys. Acta 701, 102–110.
the complement sensitivity of paroxysmal nocturnal hemoglobinuria Till, G.O., Johnson, K.J., Kunkel, R., Ward, P.A., 1982. Intravascular activa-
erythrocytes. J. Clin. Invest. 75, 2074–2084. tion of complement and acute lung injury. Dependency on neutro-
Petersen, S.V., Thiel, S., Jensenius, J.C., 2001. The mannan-binding lectin phils and toxic oxygen metabolites. J. Clin. Invest. 69, 1126–1135.
pathway of complement activation: biology and disease association. Till, G.O., Morganroth, M.L., Kunkel, R., Ward, P.A., 1987. Activation of C5
Mol. Immunol. 38, 133–149. by cobra venom factor is required in neutrophil-mediated lung injury
Petrella, E.C., Wilkie, S.D., Smith, C.A., Morgan Jr., A.C., Vogel, C.-W., 1987. in the rat. Am. J. Pathol. 129, 44–53.
Antibody conjugates with cobra venom factor. Synthesis and Tomana, M., Niemann, M., Garner, C., Volanakis, J.E., 1985. Carbohydrate
biochemical characterization. J. Immunol. Methods 104, 159–172. composition of the second, third and fifth components and factors B
Petrella, E.C., O’Keefe, M.C., Bredehorst, R., Vogel, C.-W., 1991. Purification and D of human complement. Mol. Immunol. 22, 107–111.
and characterization of two C3-cleaving proteases from cobra venom. Turnberg, D., Botto, M., 2003. The regulation of the complement system:
Complement. Inflamm. 9, 210. insights from genetically-engineered mice. Mol. Immunol. 40,145–153.
Petrella, E.C., Vogel, C.-W., 1986. Cobra venom factor from different Naja Turner, M.W., 2003. The role of mannose-binding lectin in health and
naja subspecies exhibits different structural and functional differ- disease. Mol. Immunol. 40, 423–429.
ences. Fed. Proc. 45, 1942. Van den Berg, C.W., Aerts, P.C., Van Dijk, H., 1991. In vivo anti-comple-
Pettigrew, H.D., Teuber, S.S., Gershwin, M.E., 2009. Clinical significance of mentary activities of the cobra venom factors from Naja naja and Naja
complement deficiencies. Ann. N.Y. Acad. Sci. 1173, 108–123. haje. J. Immunol. Methods 136, 287–294.
Phelps, T., 1981. Poisonous Snakes. Blandford Press, Distributed by Ster- Vik, D.P., Amiguet, P., Moffat, G.J., Fey, M., Amiguet-Barras, F., Wetsel, R.A.,
ling Pub. Co., Poole, Dorset, New York, N.Y. Tack, B.F., 1991. Structural features of the human C3 gene: intron/exon
1222 C.-W. Vogel, D.C. Fritzinger / Toxicon 56 (2010) 1198–1222

organization, transcriptional start site, and promoter region Wang, Y., Rollins, S.A., Madri, J.A., Matis, L.A., 1995. Anti-C5 monoclonal
sequence. Biochemistry 30, 1080–1085. antibody therapy prevents collagen-induced arthritis and ameliorates
Vogel, C.-W., 1986. Untersuchungen zur Strukturhomologie von Kobra- established disease. Proc. Natl. Acad. Sci. U.S.A. 92, 8955–8959.
faktor mit dem menschlichen Komplementprotein C3 sowie Synthese Wang, S.Y., Racila, E., Taylor, R.P., Weiner, G.J., 2008. NK-cell activation and
kovalenter Hybrodproteine aus Kobrafaktor und monoklonalen antibody-dependent cellular cytotoxicity induced by rituximab-
Antikoerpern als selektiv-zytotoxisches Prinzip. Ph.D. thesis, Univer- coated target cells is inhibited by the C3b component of complement.
sity of Hamburg. Blood 111, 1456–1463.
Vogel, C.-W., 1991. Cobra venom factor, the complement-activating Wang, S.Y., Veeramani, S., Racila, E., Cagley, J., Fritzinger, D., Vogel, C.-W.,
protein of cobra venom. In: Tu, A.T. (Ed.), Handbook of Natural Toxins: St John, W., Weiner, G.J., 2009. Depletion of the C3 component of
Reptile and Amphibian Venoms. Marcel Dekker, New York, pp. 147– complement enhances the ability of rituximab-coated target cells to
188. activate human NK cells and improves the efficacy of monoclonal
Vogel, C.-W., Smith, C.A., Müller-Eberhard, H.J., 1984. Cobra venom factor: antibody therapy in an in vivo model. Blood 114, 5322–5330.
structural homology with the third component of human comple- Wehrhahn, D., Meiling, K., Fritzinger, D.C., Bredehorst, R., Andrä, J.,
ment. J. Immunol. 133, 3235–3241. Vogel, C.-W., 2000. Analysis of the structure/function relationship of
Vogel, C.-W., Bredehorst, R., Fritzinger, D.C., Grunwald, T., Ziegelmüller, P., cobra venom factor (CVF) and C3: generation of CVF/cobra C3 hybrid
Kock, M.A., 1996. Structure and function of cobra venom factor, the proteins. Immunopharmacology 49, 94.
complement-activating protein in cobra venom. Adv. Exp. Med. Biol. Wessels, M.R., Butko, P., Ma, M., Warren, H.B., Lage, A.L., Carroll, M.C.,
391, 97–114. 1995. Studies of group B streptococcal infection in mice deficient in
Vogel, C.-W., Fritzinger, D.C., Hew, B.E., Thorne, M., Bammert, H., 2004. complement component C3 or C4 demonstrate an essential role for
Recombinant cobra venom factor. Mol. Immunol. 41, 191–199. complement in both innate and acquired immunity. Proc. Natl. Acad.
Vogel, C.-W., Fritzinger, D.C., 2007. Humanized cobra venom factor: Sci. U.S.A. 92, 11490–11494.
experimental therapeutics for targeted complement activation and Whaley, K., Ruddy, S., 1976. Modulation of the alternative complement
complement depletion. Curr. Pharm. Des. 13, 2916–2926. pathways by beta1H globulin. J. Exp. Med. 144, 1147–1163.
Vogel, C.-W., Müller-Eberhard, H.J., 1982. The cobra venom factor- Wiesmann, C., Katschke, K.J., Yin, J., Helmy, K.Y., Steffek, M.,
dependent C3 convertase of human complement. A kinetic and Fairbrother, W.J., McCallum, S.A., Embuscado, L., DeForge, L., Hass, P.E.,
thermodynamic analysis of a protease acting on its natural high van Lookeren Campagne, M., 2006. Structure of C3b in complex with
molecular weight substrate. J. Biol. Chem. 257, 8292–8299. CRIg gives insights into regulation of complement activation. Nature
Vogel, C.-W., Müller-Eberhard, H.J., 1984. Cobra venom factor: improved 444, 217–220.
method for purification and biochemical characterization. J. Immunol. Willemse, J.L., Chen, D., Hendriks, D.F., 2008. Major carboxypeptidase N
Methods 73, 203–220. deficiency. Clin. Chim. Acta 389, 181–182.
Vogel, C.-W., Müller-Eberhard, H.J., 1985a. The cobra complement system: I. Wu, J., Wu, Y.Q., Ricklin, D., Janssen, B.J., Lambris, J.D., Gros, P., 2009.
The alternative pathway of activation. Dev. Comp. Immunol. 9, 311–325. Structure of complement fragment C3b-factor H and implications for
Vogel, C.-W., Müller-Eberhard, H.J., 1985b. The cobra complement system: host protection by complement regulators. Nat. Immunol. 10, 728–733.
II. The membrane attack complex. Dev. Comp. Immunol. 9, 327–333. Wüster, W., 1996. Taxonomic changes and toxinology: systematic revi-
Vogt, W., Dieminger, L., Lynen, R., Schmidt, G., 1974. Alternative pathway sions of the Asiatic cobras (Naja naja species complex). Toxicon 34,
for the activation of complement in human serum. Formation and 399–406.
composition of the complex with cobra venom factor that cleaves the Wüster, W., Warrell, D.A., Cox, M.J., Jintakune, P., Nabhitabhata, J., 1997.
third component of complement. Hoppe Seyler’s. Z. Physiol. Chem. Redescription of Naja siamensis (Serpentes: Elapidae), a widely over-
355, 171–183. looked spitting cobra from S.E. Asia: geographic variation, medical
Vogt, W., Schmidt, G., Von Buttlar, B., 1977. Factors and conditions importance, and designation of a neotype. J. Zool. Lond. 243, 771–788.
essential for the activation of the fifth component of complement and Wüster, W., Harvey, A.L., 1996. Reviews of venomous snake systematics in
release of anaphylatoxin. Monogr. Allergy 12, 86–89. Toxicon. Toxicon 34, 397–398.
Vogt, W., Schmidt, G., Von Buttlar, B., Dieminger, L., 1978. A new function Ying, L., Katz, Y., Schlesinger, M., Carmi, R., Shalev, H., Haider, N., Beck, G.,
of the activated third component of complement: binding to C5, an Sheffield, V.C., Landau, D., 1999. Complement factor H gene mutation
essential step for C5 activation. Immunology 34, 29–40. associated with autosomal recessive atypical hemolytic uremic
Vogt, W., Schmidt, G., 1964. Separation of an anaphylatoxin-forming syndrome. Am. J. Hum. Genet. 65, 1538–1546.
principle from other venom components in cobra venom. Experientia von Zabern, I., Hinsch, B., Przyklenk, H., Schmidt, G., Vogt, W., 1980.
20, 207–208. Comparison of Naja n. naja and Naja h. haje cobra-venom factors:
Vogt, W., Schmidt, G., 1978. Hydrolytic activation of the fifth component correlation between binding affinity for the fifth component of
of complement (C5) by C3 convertases depends on its binding to complement and mediation of its cleavage. Immunobiology 157, 499–
surface-fixed C3b: specific interference with binding, by properdin, 514.
factor B, or propamidine, leads to inhibition of C5 cleavage and von Zabern, I., Przyklenk, H., Vogt, W., 1982. Chain structure of cobra
utilization. J. Immunol. 120, 1801. venom factor from Naja naja and Naja haje venom. Scand. J. Immunol.
Walport, M.J., 2001a. Complement. First of two parts. N. Engl. J. Med. 344, 15, 357–362.
1058–1066. Zhou, W., Farrar, C.A., Abe, K., Pratt, J.R., Marsh, J.E., Wang, Y., Stahl, G.L.,
Walport, M.J., 2001b. Complement. Second of two parts. N. Engl. J. Med. Sacks, S.H., 2000. Predominant role for C5b-9 in renal ischemia/
344, 1140–1144. reperfusion injury. J. Clin. Invest. 105, 1363–1371.