Hermanson G. Bioconjugate Techniques, Second Edition

Подождите немного. Документ загружается.

820 20. Antibody Modiﬁ cation and Conjugation

Reagent options and protocols for the radioiodination of antibodies and other molecules

may be found in Chapter 12.

Another method of adding a radioactive tag to antibodies is to use a chelating compound

capable of complexing metal isotopes. One of the most frequently used chelating reagents is

diethylenetriamine pentaacetic acid (DTPA) (Chapter 10, Section 1). The reagent contains two

anhydride groups that can be used to modify primary amines in proteins and other molecules.

The reaction process involves ring opening and the formation of an amide bond. Ring open-

ing also creates up to four free carboxylate groups which, combined with the three nitrogen

atoms in the chelator, are able to form str820ong coordination complexes with metals such as

indium-111 ( Figure 20.18 ). Monoclonals labeled with chelated isotopes can be used in target-

ing tumor cells in vivo. The detection sensitivity of radiolabeled antibodies has led to effec-

tive diagnostic procedures to monitor primary and secondary cancer growths. In addition, the

intensity of radioactivity at the tumor site when labeled monoclonals are targeted therapeuti-

cally can be great enough to cause tumor cell death and remission.

Figure 20.17 Bolton–Hunter reagent may be used to add radioactive iodine labels to antibody molecules by

modiﬁ cation of amines.

2.3. Biotinylated Antibodies

Another popular tag for use with immunoglobulins is biotin. Modiﬁ cation reagents that can

add a functional biotin group to proteins, nucleic acids, and other molecules now come in many

shapes and reactivities (Chapter 11). Depending on the reactive group present on the biotinylation

Figure 20.18 The bifunctional chelating reagent DTPA may be used to modify amine groups on antibody mol-

ecules, forming amide bond linkages. Indium-111 then may be complexed to the chelator group to create a

radiolabeled-targeting reagent.

2. Preparation of Labeled Antibodies 821

822 20. Antibody Modiﬁ cation and Conjugation

reagent, speciﬁ c functional groups on antibodies may be modiﬁ ed to create an afﬁ nity tag

capable of binding avidin (or streptavidin) derivatives. Amines, carboxylates, sulfhydryls,

and carbohydrate groups can be speciﬁ cally targeted for biotinylation through the appropriate

choice of reactive biotin compound. Figure 20.19 shows the biotinylation of an antibody with

NHS-LC-biotin, one of the most common biotinylation reagents.

The presence of biotin labels on an antibody molecule provides multiple sites for the binding

of avidin or streptavidin. If the biotin binding protein is in turn labeled with an enzyme, ﬂ uoro-

phore, etc., then a very sensitive detection system is created. The potential for more than one

labeled (strept)avidin to become attached to each antibody through multiple biotinylation sites

provides an increase in detectability over antibodies directly labeled with a detectable tag.

Several assay designs that use the enhanced sensitivity afforded through biotinylated anti-

bodies have been developed. Most of these systems use conjugates of avidin or streptavidin

Figure 20.19 Biotinylated antibodies can be formed by reacting NHS-LC-biotin with available amine groups to

create amide bonds.

with enzymes (such as HRP or AP), although other labels (such as ﬂ uorophores) can be used as

well. In the simplest assay design, called the labeled avidin–biotin (LAB) system, a biotinylated

antibody is allowed to incubate and bind with its target antigen. Next, an avidin–enzyme con-

jugate is introduced and allowed to interact with the available biotinylation sites on the bound

antibody. Substrate development then provides the detectability necessary to quantify the

antigen.

In a slightly more complex design, the bridged avidin–biotin system (BRAB) uses (strept)

avidin’s multiple biotin binding sites to create an assay potentially of higher sensitivity than

that of the LAB assay. Again, the biotinylated antibody is allowed to bind its target, but next

unmodiﬁ ed (strept)avidin is introduced to bind with the biotin binding sites on the anti-

body. Finally, a biotinylated enzyme is added to provide a detection vehicle. Since the bound

(strept)avidin still has additional biotin binding sites available, the potential exists for more

than one biotinylated enzyme to interact with each bound (strept)avidin molecule. In some

cases, sensitivity can be increased over that of the LAB technique by using the bridging ability

of avidin or streptavidin (Chapter 23, Section 2).

A modiﬁ cation on this theme can be used to produce one of the most sensitive enzyme-

linked assay systems known. The ABC system (for avidin–biotin complex) increases antigen

detectability beyond that possible with either the LAB or BRAB designs by forming a polymer

of biotinylated enzyme and (strept)avidin before addition to an antigen-bound, biotinylated

antibody. When avidin and a biotinylated enzyme are mixed together in solution in the proper

proportion, the multiple binding sites on (strept)avidin create a high molecular weight, mul-

timeric complex. If the biotinylated enzyme is not in large enough excess to block the binding

sites on all the (strept)avidin molecules, then additional sites will still be available on this com-

plex to bind a biotinylated antibody bound to its complementary antigen. The large complex

provides multiple enzyme molecules to enhance dramatically the sensitivity of detecting anti-

gen. Thus, the ABC procedure is currently among the highest-sensitivity methods available for

immunoassay work.

More detailed discussions of avidin–biotin systems as well as the process of adding a biotin

afﬁ nity group to proteins, nucleic acids, and other biomolecules can be found in Chapters 11

and 23.

2. Preparation of Labeled Antibodies 823

Monoclonal antibodies directed against tumor antigens may be used as targeting agents for

conducting certain cytotoxic substances to malignant cells for selective killing. Numerous cell-

surface markers are known to proliferate in human solid tumors (Boyer et al., 1988; Carter

et al., 2004). The ability to raise mono-speciﬁ c antibodies to these markers creates the capac-

ity to target discretely the tumor, causing cell death while leaving healthy cells alone (Salmon,

1989; McCarron et al ., 2005).

The design of such cytotoxic antibodies is conceptually simple: attach a toxic substance or

a mediator of toxicity to the appropriate monoclonal and you have a “magic bullet ” that can

ﬁ nd and eliminate the one-in-a-billion cells that have the requisite marker ( Figure 21.1 ). The

antibody provides the recognition and binding capacity, while the associated toxic component

effects cellular alterations leading to cell death (Pastan et al ., 2006).

The approach to constructing antibody immunoconjugates for cancer has taken a number

of forms (Vogel, 1987). One of the ﬁ rst designs used conjugates of monoclonal antibodies with

toxins that were able to block protein synthesis at the ribosome level inside the cell (Lord et al .,

1988). Other conjugate forms used radioactive labels that killed cells by over-exposure to radi-

ation in proximity to where antibody docking occurred (Order, 1989). Drug conjugates also

were constructed that combined the known beneﬁ ts of chemotherapy with the targeting capa-

bility of monoclonals (Reisfeld et al., 1989; Willner et al., 1993). The expected result was the

effective concentration of the chemotherapeutic agent at the location of the tumor—hopefully

eliminating or minimizing the side-effects of traditional chemotherapy. In addition, conjugates

of monoclonals with certain biological modulators such as lymphokines or growth factors were

tried to affect malignant cell viability (Obrist et al ., 1988).

Some immunoconjugates utilize intermediate carrier systems consisting of polymeric mol-

ecules such as polysaccharides, particularly dextran (Hurwitz et al., 1978, 1980, 1983a, b,

1985; Manabe et al., 1983; Sela and Hurwitz, 1987). The activated dextran is crosslinked to

both the monoclonal and the cytotoxic agent, providing multivalent conjugation sites to create

larger complexes (Section 2.3, this chapter and Chapter 25, Section 2). Liposomes can be used

in similar fashion by anchoring the antibody to its outer surface and charging the vesicle with

cytotoxic compounds (Gregoriadis, 1984; Matthay et al., 1986; Singhal and Gupta, 1986; Ho

et al., 1986). See Chapter 22 for a survey of liposome conjugation techniques. Also, dendrimer

Immunotoxin Conjugation Techniques

824

conjugates with antibody-targeting molecules and cytotoxic components have been used to cre-

ate multivalent immunotoxin conjugates (Chapter 7).

Other systems were designed to use two-stage approaches where an antibody was conjugated

with an intermediate agent, which, when combined with another factor, could elicit cytotoxic-

ity. For instance, enzyme conjugates with monoclonals have been used that could transform an

inactive pro-drug into a chemotactic agent in vivo (Senter et al., 1988). Monoclonals also could

be tagged with a biotin label and a secondary, avidin–toxin conjugate used to target the anti-

body once it bound the tumor cell. Two-stage radiative treatment also has been tried through

the use of boron neutron capture therapy (Holmberg and Meurling, 1993). In this case, a carrier

molecule is modiﬁ ed to contain

B. After administration in vivo to target tumor cells, neutron

bombardment yields an unstable intermediate

B which immediately undergoes a ﬁ ssion reac-

tion to yield

Li and

He. The induced radiation then kills the cells.

Tumor-targeting conjugates which use biospeciﬁ c agents other than monoclonal antibodies

have been developed as well. The targeting component in these systems consists of any mol-

ecule that can function as a ligand having speciﬁ c afﬁ nity for some receptor molecule on the

surface of the tumor cells. Such an afﬁ nity molecule might be a hormone (typically called hor-

monotoxins; Singh et al., 1989, 1993), a growth factor, an antigen speciﬁ c for binding particu-

lar antibodies projecting from B cell surfaces, transferrin, 

-macroglobulin, or anything else

able to speciﬁ cally interact with the targeted tumor cells.

It became apparent very early in the development of such agents that their conception and

design was much easier to imagine than to successfully implement. Monoclonal antibodies used

Figure 21.1 The basic design of an immunotoxin conjugate consists of an antibody-targeting component

crosslinked to a toxin molecule. The complexation typically includes a disulﬁ de bond between the antibody

portion and the cytotoxic component of the conjugate to allow release of the toxin intracellularly. In this illus-

tration, an intact A–B toxin protein provides the requisite disulﬁ de, but the linkage also may be designed into

the crosslinker itself.

21. Immunotoxin Conjugation Techniques 825

in vitro could easily detect antigen molecules in complex mixtures with very little nonspeciﬁ city

or cross-reactivity. Very quickly following the invention of hybridoma technology, monoclonals

were employed in numerous diagnostic assays. Their use as therapeutic agents in vivo, how-

ever, was complicated by the body ’s natural immune response designed to prevent the invasion

of foreign substances. At the time of this writing, there only are six unconjugated antibody

therapeutics approved for use in the U.S.A along with two radioimmunoconjugates, which use

a monoclonal antibody labeled with a radioactive metal ion. The only immunotoxin conjugate

employing an antibody–phytotoxin conjugate is Mylotarg, a conjugate of an anti-CD33 mono-

clonal with a calicheamicin derivative (an anti-tumor antibiotic) (Bross et al ., 2001).

A major problem in the use of immunotoxins is that injection of conjugates prepared

from mouse monoclonals usually results in antibody production against the foreign protein.

Sometimes allergic reactions can further complicate the side-effects, making continued therapy

infeasible. Most often, however, induced host immunoglobulins will quickly bind the immu-

noconjugate and remove it from the circulatory system. Instead of ﬁ nding the targeted tumor

cells, the immunotoxin ends up sequestered and degraded in the liver or removed by the kid-

neys. Since the common culprit in this scenario often is the monoclonal, the acronym HAMA,

for human anti-mouse antibody , is given to the response.

In an attempt to overcome the HAMA problem, “humanized” mouse monoclonals were

designed where large portions of the murine antibody are substituted for their human coun-

terparts. For instance, replacing a mouse Fc portion with the corresponding human ones can

signiﬁ cantly decrease the immune response against such conjugates. Replacing everything

but the hypervariable regions which code for antigen binding has been accomplished, too.

Unfortunately, regardless of how much “humanization” is done, the remaining murine part still

has the potential of causing immunological reactions. The advent of recombinant antibodies of

completely human origin essentially has overcome the problem of immune system response to

the antibody component. However, there still is the potential for generating an immune response

from the toxic part of the immunoconjugate, but this is unavoidable unless immune modulators

are administered to reduce the overall immune system responsiveness.

Modiﬁ cation of immunotoxin conjugates with synthetic polymers has been used to mask

the complex from the host immune response. Particularly, polyethylene glycol (PEG; Chapter

25, Section 1 and Chapter 18) has been found to be quite successful in reducing or eliminating

immune system reactions (Rofﬂ er and Tseng, 1994). Modiﬁ cation of an antibody conjugate

with 2–4 PEG molecules increases the serum half-life and improves tumor localization of the

targeted reagent.

Other innovations in preparing targeted conjugates for cancer utilize recombinant DNA tech-

niques to create antibody molecules that are entirely of human origin (Huse et al., 1989; Orlandi

et al., 1989; Sastry et al., 1989; Hudson and Souriqu, 2007). A completely human antibody

molecule eliminates the immunological problems associated with mouse monoclonals. Intact

antibodies, Fab fragments, small Fv fragments held together by synthetically designed amino

acid segments, and short peptides representing the antigen binding site have all been developed

by recombinant means. Although frequently the word “antibody” is used to describe these engi-

neered proteins, many of the molecules are far removed from the traditional picture of an anti-

body molecule. The terms “single chain antibody ” or “single chain Fv protein ” are commonly

used and more closely describe these new targeting molecules.

With the great diversity of targeted toxic agents being developed for cancer therapy, it

would be difﬁ cult to characterize this section strictly as antibody conjugation. While many,

826 21. Immunotoxin Conjugation Techniques

if not most, studies utilize monoclonal antibodies of one form or another as the biospeciﬁ c-

targeting component, the complete picture involves the crosslinking of a wide variety of mol-

ecules together to create the ﬁ nal conjugate. This section presents some of the most common

methods of immunotoxin preparation. For the preparation of other unique targeted toxin con-

jugates, the methods found throughout this book for linking one particular functional group to

another can be followed with an excellent probability for success.

1. Properties and Use of Immunotoxin Conjugates

Conjugates of monoclonal antibodies and protein toxins are undergoing extensive research for

their usefulness in the treatment of cancer. Toxins of many different types can be used to cre-

ate effective immunotoxin conjugates, including the proteins ricin from castor beans ( Ricinus

communis), abrin from Abrus precatorius, modeccin, gelonin from Gelonium multiﬂ orum

seeds, diphtheria toxin produced by Corynebacterium diphtheriae, pokeweed antiviral proteins

(PAPs; three types: PAP, PAP-II, and PAP-S) from Phytolacca americana seeds, cobra venom

factor (CVF), Pseudomonas exotoxin, restrictocin from Aspergillus Restrictus, momordin from

Momordica Charantia seeds, saporin from Saponaria ofﬁ cinalis seeds, as well as other ribosome-

inactivating proteins (RIPs).

By far the most popular choices for the toxin component of immunotoxins are ricin, abrin,

modeccin, and diphtheria toxin. The three plant toxins have lectin binding activity toward

terminal -galactosyl residues and they can be inhibited by the presence of simple sugars like

galactose and lactose. The toxin proteins bind to cell-surface polysaccharide receptors with high

afﬁ nity (k

in the range of 10

–10

). Ricin, abrin, and modeccin consist of two subunits with

remarkably similar structures and activities. The intact proteins have molecular weights (MW)

of approximately 63–65,000 with each subunit of about equal size. The subunits are joined by

disulﬁ de linkages that are important reversible bonds in the mechanism of cytotoxicity. The A

chain is called the effectomer and possesses ribosomal-inactivating properties. The B chain con-

tains the carbohydrate binding site and it is termed the haptomer. While the intact toxin mol-

ecules have potent cytotoxic effects on cells, they exhibit no ribosomal-inactivating activity on

ribosomes in a cell-free system. By contrast, reduction of the toxins with a disulﬁ de reducing

agent creates the opposite effects. Reduced, dissociated toxin subunits inhibit ribosomal activity

in cell-free systems, but they have no affect on intact cells.

The reason for these properties is due to the toxins ’ mode of action. Toxin molecules bind

through saccharide-recognition sites on the B chain to particular -galactosyl-containing glyc-

oprotein or glycolipid components on the surface of cell membranes. In animals sensitive to

these toxins, the necessary polysaccharide ligands are present in large quantities on virtually

all cell types (Cumber et al., 1985). Upon binding of the dimer to the cell, the A chain enters

the cell either by active transport into endocytic vesicles or through some mechanism of its

own. Once inside the cell membrane, the A chain enters the cytoplasmic space, binding to and

enzymatically inactivating the 60S subunit of ribosomes (Olsnes and Pihl, 1976, 1982). The

result is cessation of protein synthesis and eventual cell death. Because the A chain ’s action is

through enzymatic means, as little as one active toxin molecule is enough to seriously disrupt

protein synthesis operations and probably sufﬁ cient to kill a target cell (Eiklid et al., 1980).

The turnover rate of one A-chain molecule is about 1,500 ribosomes inactivated per minute

(Olsnes, 1978).

1. Properties and Use of Immunotoxin Conjugates 827

Diphtheria toxin also is a two-subunit protein, but it is initially synthesized by certain strains

of Corynebacterium diphtheriae as a single polypeptide chain of MW 63,000. Proteolytic

processing results in the formation of a “nicked toxin ” which is enzymatically inactive, but

consists of two subunits bonded together by an interchain cystine disulﬁ de. Upon reduction

of the disulﬁ de, the A chain (MW 24,000) is released and manifests enzymatic activity toward

ribosomal proteins (Sandvig and Olsnes, 1981; Collier and Cole, 1969). Its mode of action is

different than the plant toxins. The A-chain fragment of diphtheria toxin catalyzes the ADP-

ribosylation of eukaryotic aminoacyl-transferase II (EF2) using NAD



(Honjo et al., 1968; Gill

et al., 1969). The B chain, by contrast, possesses no enzymatic activity, but evidence points

to the fact that a binding site on it recognizes certain cell-surface receptors. As in the action

of ricin, abrin, and modeccin, the B chain of diphtheria toxin is necessary for cytotoxicity

(Colombatti et al., 1986). There also is a role for the C-terminus cysteine residue of the B chain

in cell penetration (Dell ’Arciprete et al ., 1988).

Figure 21.2 illustrates the basic structure of these common two-subunit toxins, showing

schematically their major characteristics. The molecular model of ricin is from Rutenber et al .

(1991), RSCB structure No. 2aai.

Due to the extraordinary toxicity of intact ribosome-inactivating toxins like ricin, abrin, and

modeccin, puriﬁ cation and handling of these proteins must be done with extreme care. Even

dust from crude seed powders or lyophilized proteins should be considered dangerous. During

Figure 21.2 Conceptualized construction of an A–B subunit protein toxin (left). The B chain contains a binding

region for docking onto cell surfaces, while the A chain contains a catalytic site that produces cytotoxic affects

intracellularly. The two subunits are joined by a disulﬁ de bond that is reductively cleaved at the cellular level to

allow the A subunit to affect cell death. A molecular model of ricin is on the right.

828 21. Immunotoxin Conjugation Techniques

the height of the cold war days, a Soviet KGB agent killed a man from the West by injecting at

most only milligram quantities of ricin into his leg using a modiﬁ ed umbrella tip. There even

have been instances of worker deaths at companies that routinely purify these proteins. For

this reason, all handling operations of intact toxin dimers and puriﬁ ed subunits should be done

in fume or laminar-ﬂ ow hoods. Avoid, also, the use of laboratory tools that could lead to punc-

ture wounds causing contaminating toxin injection.

Some potentially cytotoxic proteins contain only a single polypeptide chain, such as gelo-

nin and PAPs. Such toxins manifest similar enzymatic ribosome-inactivating properties as

the multi-subunit proteins like ricin, but do not possess the cell-recognition capacity that the

B-chain subunit contains. The result is the inability of these toxins to bind or affect intact cells.

However, they do maintain the typical ribosome-inactivating properties in a cell-free system

that the A chain of two-subunit toxins possess. If these proteins are conjugated with a cell-

targeting agent, such as the B chain of ricin or a speciﬁ c antibody that recognizes cell-surface

epitopes, full cytotoxicity results.

Gelonin and PAPs are much more convenient to work with than ricin and the other two-

subunit toxins. Most importantly, they are relatively nontoxic to cells unless conjugated

with something that can facilitate cell binding and internalization (Stirpe et al., 1980; Irvin,

1983). The pI of these toxins is in the basic range, and they each have a MW of about 30,000

(Barbieri and Stirpe, 1982). These single-subunit proteins are very stable, especially to puriﬁ ca-

tion techniques, but also to most modiﬁ cation and crosslinking steps associated with preparing

immunotoxins. Studies have shown (Lambert et al., 1985) that modiﬁ cation of gelonin or PAPs

can be done with 2-iminothiolane (Chapter 1, Section 4.1) to create sulfhydryl groups with-

out loss of activity. Previous studies, however, have determined that the use of SPDP to mod-

ify gelonin resulted in a 90 percent inactivation (Thorpe et al., 1981). The difference in these

results may be due to the retention of positive charge characteristics on the modiﬁ ed amine

when using Traut ’s reagent, but neutralization of that charge when using SPDP.

Immunotoxin conjugates consist of an antibody covalently crosslinked to a toxin molecule

in a way that maintains the unique properties of both proteins. The antibody component con-

sists of a monoclonal having speciﬁ city for an antigenic determinant on the surface of a par-

ticular cell type. Most often, the targeted cells are tumors that express a unique cell-surface

marker which can be recognized by the monoclonal. The role of the antibody, therefore, is

to function as a passive taxi, carrying the toxin component to the targeted cells. Once at the

tumor location, the toxin component effects its intended ribosome-inhibiting action, ultimately

causing cell death and tumor destruction.

Since immunotoxin conjugates are destined to be used in vivo, their preparation involves

more critical consideration of crosslinking methods than most of the other conjugation proto-

cols described in this book. The following sections discuss the problems associated with toxin

conjugates and the main crosslinking methods for preparing them.

2. Preparation of Immunotoxin Conjugates

It has become apparent that the method of crosslinking can dramatically affect the activity of

an immunotoxin in vivo. This is true not only with regard to possible direct blocking by the

crosslinker of the enzymatic active site which is responsible for inactivation of ribosomes, but

the chemistry of conjugation also is an important factor in proper binding and entry of the

2. Preparation of Immunotoxin Conjugates 829