Therapeutic Applications of Monoclonal
`Antibodies
`
`MITCHELL BERGER, MPH; VIDYA SHANKAR, PHD; ABBAS VAFAI, PHD
`
`ABSTRACT: Researchers have sought therapeutic appli-
`cations for monoclonal antibodies since their develop-
`ment in 1975. However, murine-derived monoclonal
`antibodies may cause an immunogenic response in hu-
`man patients, reducing their therapeutic efficacy. Chi-
`meric and humanized antibodies have been developed
`that are less likely to provoke an immune reaction in
`human patients than are murine-derived antibodies. An-
`tibody fragments, bispecific antibodies, and antibodies
`produced through the use of phage display systems and
`
`genetically modified plants and animals may aid
`researchers in developing new uses for monoclonal
`antibodies in the treatment of disease. Monoclonal an-
`tibodies may have a number of promising potential
`therapeutic applications in the treatment of asthma,
`autoimmune diseases, cancer, poisoning, septicemia,
`substance abuse, viral infections, and other diseases.
`KEY INDEXING TERMS: Antibodies; Monoclonal; Ther-
`apeutic. [Am J Med Sci 2002;324(1):14–30.]
`
`In 1975, Kohler and Milstein revolutionized the field
`
`of immunology by developing monoclonal antibod-
`ies (MAbs). Since that time, many MAbs have been
`developed for use in diagnostic procedures and in im-
`munotherapy. Ever since it was observed that the
`therapeutic use of heterologous MAbs elicited immu-
`nogenic responses in humans, significant research ef-
`forts have been devoted toward creating chimeric and
`humanized antibodies for use in human patients. Ma-
`jor achievements have been in the production of MAbs
`in transgenic plants and animals. The use of phage
`display libraries has created customized antibodies
`with defined affinity and specificity. This review de-
`scribes how rodent, chimeric, and humanized antibod-
`ies have each been used, with varying degrees of
`success to treat cancer, septicemia, autoimmune dis-
`orders, and infectious diseases. We also describe here
`recent applications of antibody engineering, such as
`the use of bispecific antibodies and antibody fragments
`in immunotherapy.
`
`History of MAb Development
`Von Behring and Kitasato discovered in 1890 that
`the serum of vaccinated persons contained certain
`
`From the Emory University School of Public Health, Atlanta,
`Georgia (MB); Biologics Branch, Scientific Resources Program,
`Centers for Disease Control and Prevention, Public Health Ser-
`vice, United States Department of Health and Human Services,
`Atlanta, Georgia (VS, AV).
`Submitted August 2, 2001; accepted December 21, 2001.
`Correspondence: Abbas Vafai, Ph.D., Biologics Branch, Scien-
`tific Resources Program, Centers for Disease Control and Preven-
`tion, Public Health Service, US Department of Health and Human
`Services, Atlanta, GA 30333 (E-mail: abv4@cdc.gov).
`
`14
`
`substances, which they termed “antibodies.” In
`1895, they treated diphtheria with an antiserum
`raised against the toxin. On the basis of research on
`tetanus toxin and trypanosome parasites, Ehrlich
`proposed in 1900 the “side-chain theory” of antibody
`formation, which hypothesized that physiologically
`active substances, including toxins, attach to cell
`surface receptors that are produced in response to
`toxin-cell interactions and then ejected from the
`cells into the bloodstream, leading to circulating
`antibodies.1,2
`Not until the 1950s, however, did scientists’ un-
`derstanding of antibodies become sufficient to lay
`the foundation for the development of MAbs. Jerne3
`postulated in 1955 a theory of natural selection for
`antibody formation. Animals vaccinated with an an-
`tigen were expected to produce several distinct an-
`tibodies against several epitopes of the antigen.
`Frank Macfarlane Burnet subsequently refined and
`expanded Jerne’s theory.4 Burnet’s “clonal selection
`theory,” as it is generally known, postulates that
`cells specific for synthesizing 1 type of antibody are
`spontaneously generated due to random somatic
`mutations during the maturation of the immune
`system and that these cells proliferate when exposed
`to an antigen. At about the same time, Porter iso-
`lated fragment antigen binding (Fab) and fragment
`crystalline (Fc) from proteolytically cleaved rabbit
`␥-globulin.5
`Until the 1960s, antibody-producing cells were
`difficult to maintain in culture, because they died
`after a few days. In addition, only polyclonal anti-
`bodies could be obtained. In 1964, Littlefield6 devel-
`oped a way to isolate hybrid cells from 2 parent cell
`lines using the hypoxanthine-aminopterin-thymi-
`
`July 2002 Volume 324 Number 1
`
`1
`
`EX2094
`Eli Lilly & Co. v. Teva Pharms. Int'l GMBH
`IPR2018-01427
`
`

`

`dine (HAT) selection media. In 1975, K␱hler and
`Milstein7 provided the most outstanding proof of the
`clonal selection theory from results of heterokary-
`ons—cell hybrids formed by the fusion of normal
`and malignant cells. Twenty-five years after Kohler
`and Milstein produced the first monoclonal antibod-
`ies, dramatic progress has been made in using anti-
`bodies for diagnostic purposes, but the uses of MAbs
`to treat disease have—until recently—remained
`somewhat limited.8,9 However, as this review indi-
`cates, the potential of MAbs to aid in the treatment
`of a wide range of diseases is now beginning to be
`realized.
`
`Antibody Structure
`Antibodies are Y-shaped proteins composed of
`peptides called heavy and light chains, the ends of
`which vary from antibody to antibody. Each combi-
`nation of heavy and light chains binds to a particu-
`lar antigenic site. These glycoprotein chains are
`folded into domains of about 110 amino acids that
`become twisted into the immunoglobulin (Ig) fold
`and are stabilized by disulfide bonds. Structurally,
`each Ig molecule consists of 2 50-kD heavy chains
`and 2 25-kd light chains, linked by disulfide bonds.
`Human immunoglobulins are divided into 5
`classes or isotypes based on the amino acid compo-
`sition of their heavy chains: ␣, ␦, ⑀, ␥, and ␮, denoted
`IgA, IgD, IgE, IgG, and IgM, respectively. There are
`2 kinds of light chains, ␬ (k) and ␭ (l), which are
`common to all 5 classes. Four subclasses of IgG
`(IgG1, IgG2, IgG3, and IgG4) and 2 subclasses of IgA
`(IgA1 and IgA2) exist, each with a distinct function.
`Secretory IgA exists in dimeric form held together
`by a J chain and is associated with a secretory
`component that helps it pass through the cell
`membrane.10,11
`Each chain has a constant domain to bind host
`effector molecule and a variable domain to bind to
`the target antigen. Light chains have 1 variable (VL)
`and 1 constant domain (CL), whereas heavy chains
`have 1 variable (VH) and either 3 (␣, ␦, and ␥chains)
`or 4 constant domains (⑀ and ␮ heavy chains) de-
`pending upon the isotype class. Each variable do-
`main contains 3 regions known as “hypervariable
`loops,” also known as complementarity-determining
`regions (CDRs), that identify the antigen. The other
`amino acids in the variable (Fv) domain are known
`as framework residues and act as a scaffold to sup-
`port the loops. The VL and the VH, the CH1 and the
`CL, and the 2 CH3 domains are paired; the 2 CH2
`domains have carbohydrate side chains attached to
`them and are not paired. The folded constant do-
`mains may be homologous among different species,
`allowing hybrid domains (eg, mouse-human) to be
`produced. The variable domain confers specificity
`and affinity. The CDR amino acid sequences are
`extremely variable and play a large role in interac-
`
`Berger, Shankar, and Vafai
`
`tion with the targeted antigen. The chain type, re-
`gion, and distance from the amino terminus charac-
`terize the domains. Thus, CH2 domain refers to the
`second constant domain of the heavy chain.
`Upon digestion with papain, the antibody mole-
`cule is cleaved on the amino-terminal side of the
`disulfide bridges into 2 identical Fabs and an Fc
`fragment, whereas pepsin cleaves the antibody on
`the carboxy-terminal side of the disulfide bridges
`into 1 F(ab')2 fragment containing both the arms of
`the antibody, and many small pieces of the Fc frag-
`ment. Currently, the following are available: whole
`antibodies, enzymatically produced 50-kd Fab frag-
`ments, engineered 25-kd single-chain Fv (scFv) an-
`tibodies consisting of the VH and VL connected by a
`flexible peptide linker, diabodies
`(noncovalent
`dimers of scFv), minibodies (scFv-CH3 dimers), and
`heavy chain IgGs found in species of the Camelidae
`family, which are devoid of light chains and are
`referred to as VHH.
`
`Conventional MAb Production
`The first MAb described by Kohler and Milstein
`was created by the fusion of murine myeloma cells
`with murine-antibody–secreting
`lymphocytes.7,8
`Myeloma cells are immortalized B lymphocytes ca-
`pable of secreting homogeneous antibodies. The im-
`mortal myeloma cell lacks the enzyme hypoxanthine
`guanosine phosphoribosyl transferase (HGPRT) and
`is sensitive to the HAT media. However, a hybrid
`cell known as a hybridoma, generated by the fusion
`of myeloma cell and an antibody-producing B cell,
`can survive in the HAT media. The spleen B lym-
`phocytes contribute the HGPRT gene to the hybrid
`cell and, hence, unfused myeloma cells and spleen
`cells die in the HAT media.
`The conventional method of generating MAbs is
`the hybridoma technology in which spleen cells from
`immunized mice are fused with murine myeloma
`cells. Whereas the myeloma cell imparts immortal-
`ity to the hybridoma allowing cells to be cultivated
`indefinitely, the immune spleen B-cell confers anti-
`gen specificity. Because each hybridoma is derived
`from a single cell, the cells within a hybridoma cell
`line are identical and make the same antibody mol-
`ecule with same antigen-binding site and isotype,
`hence it is called a MAb. Among several excellent
`reviews that detail MAb production with hybrid-
`omas is a recent review by Dean and Shepherd.12
`Initial attempts to bypass the mouse to make hu-
`man MAbs involved fusion of human immune spleen
`lymphocytes with nonsecreting human myeloma
`partners to obtain hybrid cells that continually se-
`crete a specific antibody. However, poor fusion of
`human myelomas, unsatisfactory performance of
`the hybrid cells, and the difficulty in accessing im-
`mune lymphocytes have prevented success. Al-
`though heteromyelomas—which are fusions of hu-
`
`THE AMERICAN JOURNAL OF THE MEDICAL SCIENCES
`
`15
`
`2
`
`

`

`Therapeutic Applications of Monoclonal Antibodies
`
`man and mouse myelomas—work better, these
`hybrids are usually unstable. Attempts to use mouse
`myeloma cells to create hybrids and derive human
`MAbs led to the loss of human chromosomes and the
`inability to make human Igs.13
`Unfortunately, in vitro immunization is limited by
`its inability to produce a secondary response and by
`the absence of the affinity maturation process that
`occurs in vivo.13 Affinity maturation process is a
`complex phenomenon, a consequence of intense B-
`cell proliferation, somatic hypermutation of Ig vari-
`able domain genes, and selection for B cells with
`high-affinity antigen binding, all occurring in the
`specialized microenvironment of the germinal cen-
`ter within the lymphoid tissue. Thus, the search for
`an ideal fusion partner for generating human MAbs
`has been difficult, which is why the Epstein-Barr
`virus (EBV) technique for immortalization of B lym-
`phocytes is preferred. But immortalized B cells do
`not always replicate exactly the in vivo antibody
`response, because the tissues from which these cells
`were selected and the manipulations to which they
`are subjected to in the laboratory may alter the
`antibody specificity.14,15 Protocols for the prepara-
`tion of EBV virus, B cells, and cell fusion have been
`described elsewhere.14,16,17
`Methods used for large-scale production of MAbs
`may include the generation of ascites tumors in mice
`or in vitro mammalian cell culture fermentation by
`using bioreactors and continuous perfusion culture
`systems.18,19 The key issues in scale-up productions
`are the growth media, fermenter size, fermentation
`time, and purification procedures.19 Purification or
`downstream processing is accomplished by chromatog-
`raphy,
`fragmentation, conjugation with chelating
`agents, ultrafiltration, and controlled precipitation.18
`
`New Approaches and Developments
`A more recent technique for producing antibody-
`like molecules uses what is known as the Phage
`Display Library. It involves the construction of VH
`and VL gene libraries and their expression on the
`surface of a filamentous bacteriophage. Developed in
`the 1990s, the phage display method requires re-
`peated “panning” or screening of different antibodies
`based on their affinity for a specific antigen. Anti-
`body genes are linked to bacteriophage coat protein
`genes and the bacteriophages with the fusion genes
`are used to infect bacteria to create the phage dis-
`play library. The resulting bacteriophages express
`the fusion proteins and display them on their sur-
`face, and the phage display library comprises recom-
`binant phages, each displaying a different antigen-
`binding site on its surface. The phage expressing an
`antigen-binding domain specific for a particular an-
`tigen can be detected and isolated by binding to the
`surface coated with that antigen. Libraries of VH
`and VL genes may be generated from nonimmunized
`
`donors, immunized donors who have an immune
`response against a particular antigen or from a
`synthetic
`library
`consisting
`of
`antibody
`fragments.20,21
`One promising way to increase antibody yield or
`develop new antibodies may be by using genetically
`altered animals and plants. Abgenix, a company in
`Fremont, CA, has developed the transgenic “Xeno-
`Mouse,” in which the mouse antibody-producing
`genes have been inactivated and functionally re-
`placed by approximately 90% of the human Ig gene
`loci in germline configuration, coding for the heavy
`and ␬light chains.22,23 Upon immunization with any
`specific human or nonhuman antigens, the “Xeno-
`Mouse” generates MAbs, which are fully human Igs,
`with high affinities and antigen-binding specifici-
`ties. “XenoMouse” strains producing specifically
`IgG1, IgG2, or IgG4 isotypes have also been created
`to generate panels of diverse and highly specific
`MAbs.
`Kirin Brewery Company, Japan, has developed
`another transgenic mouse known as the “Trans-
`Chromo” mouse. The endogenous IgH and IgG loci of
`the “Trans-Chromo” mouse were inactivated, but it
`harbors 2 individual human chromosome fragments,
`derived from human chromosomes 2 and 14, that
`contain whole human Ig light- and heavy-chain loci,
`respectively.24 These mice are capable of producing
`every subtype of fully human Ig, including IgA and
`IgM. In these transgenic mouse models, human an-
`tibodies with high affinity to an immunized antigen
`are naturally selected by the murine immune sys-
`tem via an affinity maturation process, and thereby
`show increased diversity of the MAbs.
`Transgenic mice may be a suitable alternative to
`chimeric or humanized antibody production or the
`use of phage display systems to create less immuno-
`genic or novel antibodies.25 For instance, transgenic
`mice that express MAbs to coronavirus in their milk
`have been developed.26,27 Because ruminant ani-
`mals, such as cows, goats, and sheep, produce rela-
`tively large amounts of milk, genetically-modified
`members of these species could also be used to pro-
`duce large quantities of therapeutic proteins, includ-
`ing MAbs.28
`Plants may be a potential source of recombinant
`proteins, including MAbs.29 –31 Plant virus vectors,
`such as the tobacco mosaic virus, may be used to
`make MAbs. Transgenic tobacco plants may also be
`used for large-scale production of recombinant IgA,
`which is used in passive mucosal immunothera-
`py.30,31 This MAb could be added to toothpaste to
`effectively protect against bacteria that cause tooth
`decay.31,32
`Recombinant antigens obtained from plants may
`also have therapeutic applications. For instance,
`attempts to immunize mice with Escherichia coli
`heat labile enterotoxin B produced in transgenic
`tobacco and potato plants have proved promising.33
`
`16
`
`July 2002 Volume 324 Number 1
`
`3
`
`

`

`Hepatitis B viral surface protein produced in trans-
`genic tobacco plants has been shown to be immuno-
`genic in mice.32 The genes coding for murine malig-
`nant B-cell specific markers are inserted into
`tobacco mosaic virus to cultivate an immunogenic
`protein in tobacco plants that may eventually be
`used in developing a vaccine against non-Hodgkin
`lymphoma.34,35 Other immunotherapeutic proteins
`under development include Norwalk virus capsid
`proteins produced in tobacco and potatoes, cholera
`toxin and CT-B produced in potatoes, hepatitis B
`antigen produced in tobacco, anti-human IgG used
`to detect nonagglutinating antibodies produced in
`alfalfa, and humanized anti-herpes simplex virus
`(HSV)-2 grown in soybeans.36 –39
`
`FDA Regulation of Therapeutic Monoclonal
`Antibodies
`The Food and Drug Administration (FDA) consid-
`ers antibodies to be “biopharmaceuticals”; as such,
`MAb applications are regulated by the agency’s Cen-
`ter for Biologics Evaluation and Research (CBER)
`and the Center for Drug Evaluation and Re-
`search.40,41 The FDA has created a “Points to Con-
`sider” document advising manufacturers of factors
`to consider in the production and testing of MAbs
`intended for human use and identified information
`that should appear in Investigational New Drug or
`biologics license applications.40 The “Points to Con-
`sider” document serves to “indicate the agency’s
`current thinking on MAb products for human use.”40
`The agency’s recommendations are aimed at protect-
`ing human health because viruses and cellular DNA
`from antibody-producing cells with malignant phe-
`notypes may be integrated into host cells after
`transformation. Accordingly, among the Points to
`Consider are several steps—such as taking care to
`ensure purity of immunoconjugates and demon-
`strating the ability of any purification scheme to
`remove adventitious agents—designed to prevent
`contamination of the final product by human patho-
`gens.40 Manufacturers must also adhere to animal
`care standards and detail steps to prevent contam-
`ination of cell culture.
`The Points to Consider document includes a list of
`normal human tissues used in cross-reactivity test-
`ing, tests for murine viruses, and organs to be con-
`sidered in dosimetry estimates. The FDA recognizes
`that because of species differences, animal models
`expressing the antigen of interest or cross-reactive
`epitopes are not always available.
`The agency has also published a guidance entitled
`S6 Preclinical Safety Evaluation of Biotechnology
`Derived Pharmaceuticals based on International
`Conference of Harmonization technical require-
`ments.41 Before beginning phase 1 clinical studies—
`conducted to assess the safety of the drug and its
`mechanism of action—the FDA recommends that
`
`Berger, Shankar, and Vafai
`
`researchers conduct in vivo and in vitro testing of
`the MAb to assess cross-reactivity with human tis-
`sues or non–target-tissue binding. Preclinical safety
`testing aims to evaluate the immunogenicity, cross-
`reactivity, stability and effector functions of MAbs.
`The metabolism, carcinogenicity, and genotoxicity of
`the product should also be evaluated.41
`Guidance has also been published for MAbs used
`as reagents in drug manufacturing.42 The guidance
`emphasizes biological safety, performance charac-
`teristics of the reagent, and potential presence of
`residual amounts of the reagent in the final product.
`The FDA general principles for testing and man-
`ufacturing are broadly applicable to many classes of
`antibodies, including those produced by phage dis-
`play systems or transgenic plants and animals. De-
`pending on the expression system being used, other
`FDA guidance documents, such as CBER’s Points to
`Consider in the Production and Testing of Therapeu-
`tic Products for Human Use Derived from Trans-
`genic Animals, Points to Consider in the Production
`and Testing of New Drugs and Biologicals Produced
`by Recombinant DNA Technology, and Points to
`Consider in the Characterization of Cell Lines Used
`should
`also
`be
`to
`Produce
`Biologicals
`referenced.43– 45
`
`Humanizing Monoclonal Antibodies
`Rodent MAbs with excellent affinities and speci-
`ficities have been generated using conventional hy-
`bridoma technology, but their use in clinical medi-
`cine is limited due to the immune responses they
`elicit in humans. For instance, the human anti-
`mouse antibody (HAMA) response can compromise
`the clinical effectiveness of murine MAbs. Although
`HAMA responses are directed against the murine
`constant regions, which represent the major anti-
`genic features of the mouse Ig, significant responses
`are also directed toward the murine variable re-
`gions. As a result, patients may mount an immune
`response against the injected murine antibodies,
`leading to allergic or immune complex hypersensi-
`tivities, rapid clearance of the antibody, and reduced
`clinical efficacy.13,46 –51
`Initially, Morrison and colleagues48,52 introduced
`chimeric MAbs in 1984, which showed several ad-
`vantages over unmodified rodent antibodies. Gener-
`ally, chimeras combine the human constant regions
`with the intact rodent variable regions, replicating
`the rodent antibody variable regions by PCR and
`then cloning them into eukaryotic expression vec-
`tors containing human constant regions.49 Ideally,
`this allows better interaction with human effector
`cells and the complement system. Because the Fc
`region has little influence on the structure of the Fv
`region, the chimeric constructs’ affinity and specific-
`ity are virtually unchanged, and both rodent and
`chimeric antibodies cause apoptosis at a similar rate
`
`THE AMERICAN JOURNAL OF THE MEDICAL SCIENCES
`
`17
`
`4
`
`

`

`Therapeutic Applications of Monoclonal Antibodies
`
`and intensity against target cells in vitro.50,51,55 Al-
`though chimeric antibodies have helped solve some
`of the problems associated with the use of rodent
`MAbs, they still show significant immunogenicity in
`humans; because of their approximately 30% mouse
`sequence, they cause an human-antichimeric anti-
`body response.
`Humanized antibodies containing only the CDRs
`of the rodent variable region grafted onto the human
`variable region framework have been introduced to
`overcome these deficiencies.53 The early work on
`recombinant, chimeric, and rodent/human antibod-
`ies happened during the mid 1980s and, by the late
`1980s, Greg Winters and his colleagues demon-
`strated that a functional human-like antibody could
`be created by grafting the antigen-binding CDRs
`from variable domains of rodent antibodies onto
`human variable domains. Numerous humanized an-
`tibodies have now been designed and constructed,
`and many are currently being evaluated in clinical
`trials.13,14,39,47,54
`Efficient procedures for constructing humanized
`antibodies have been developed.13,14,47,55 The first
`step is to clone and sequence the complementary
`DNAs (cDNAs), coding for the variable domains of
`the mouse antibody to be humanized. The mouse
`hybridoma cell line is grown in an appropriate cul-
`ture medium, and cells are harvested for RNA iso-
`lation. Polymerase chain reaction (PCR) primers
`that hybridize to the 5' ends of the mouse leader
`sequences and to the 5' ends of the mouse constant
`regions are designed for cloning ␬ light chain vari-
`able regions and heavy chain variable regions.
`cDNA is synthesized from total RNA, followed by
`PCR amplification with light and heavy chain spe-
`cific primers. Positive bacterial colonies containing
`mouse variable regions are then screened.
`Construction of a chimeric antibody involves mod-
`ifying the cloned mouse leader-variable regions at
`the 5'- and 3'- ends, using PCR primers to create
`restriction enzyme sites for convenient insertion
`into expression vectors, to incorporate sequences for
`efficient eukaryotic translation, and to incorporate
`splice-donor sites for RNA splicing of the variable
`and constant regions. The adapted mouse light and
`heavy chain leader-variable regions are inserted
`into vectors containing, for example, human cyto-
`megalovirus enhancer and promoter for transcrip-
`tion, a human light or heavy chain constant region,
`a neomycin gene for selection of transformed cells,
`and the simian virus 40 origin of replication in COS
`cells. These vectors are designed to express chimeric
`or reshaped human light and heavy chains in mam-
`malian cells.
`The design and construction of an engineered hu-
`man antibody require an analysis of the primary
`amino acid sequences of the mouse variable regions
`to identify the residues most critical in forming the
`antigen-binding site. A structural model of the
`
`mouse variable region is built on the basis of homol-
`ogy to known antibody variable regions. The frame-
`work regions (FRs) of the new variable regions are
`modeled on FRs from structurally similar immuno-
`globulin variable regions. The design process in-
`volves selecting human light and heavy chain vari-
`able regions that will serve as templates for the
`construction of a reshaped human antibody. The
`mouse CDRs are then joined to the FRs from se-
`lected human variable regions. The primary amino
`acid sequences are then carefully analyzed to ascer-
`tain whether they would recreate an antigen-bind-
`ing site that mimics the original mouse antibody.
`Within the FRs, the amino acid differences be-
`tween the mouse and the human sequences are
`examined, and the relative importance of each
`amino acid in the formation of antigen-binding site
`is evaluated. Minimum changes in the FRs are de-
`sirable and should closely match the sequences from
`natural human antibodies. Any potential glycosyla-
`tion site in the FRs of either mouse or human se-
`quence needs to be identified and its influence on
`antigen binding considered.
`The DNA sequences coding for the reshaped hu-
`man variable regions, either made synthetically or
`based on an existing sequence that is very similar to
`the newly designed reshaped human variable re-
`gion, are modified by PCR with specially designed
`oligonucleotide primers. The human variable re-
`gions together with their leader sequences are then
`cloned into a mammalian expression vector that
`already contains human constant regions. Each hu-
`man variable region is linked to the desired constant
`region via an intron.
`Preliminary expression and analysis of the re-
`shaped human antibodies are done by cotransfection
`of mammalian cell-expression vectors, 1 coding for
`human light chain and 1 coding for human heavy
`chain. The vectors will replicate in the COS cells and
`transiently express and secrete reshaped human
`antibodies. The concentration of the antibody pro-
`duced can be analyzed by using an enzyme-linked
`immunosorbent assay. Specific changes in the
`amino acids of the framework region may also be
`required to preserve the orientation and structure of
`the rodent CDR required for binding. Computer
`modeling using databases containing human vari-
`able genes will identify sequences homologous to the
`rodent V regions.13,47 A computer model of the ro-
`dent Fv can identify the non-CDR residues that
`interact with the CDR sequences, and choices can be
`made regarding which residues need to be included
`in the variable region.
`Antibodies humanized in this way may have bind-
`ing affinities up to one-third greater than the corre-
`sponding murine antibodies.49 Allergenicity is also
`significantly reduced. About 20 to 40% of patients
`exhibit a HAMA reaction to murine antibodies,
`
`18
`
`July 2002 Volume 324 Number 1
`
`5
`
`

`

`whereas only about 7% have a similar reaction to
`humanized antibodies.47,53,56 –58
`Humanization of MAbs still has several practical
`difficulties. First, a detailed knowledge of the anti-
`body structure and function is required. Second,
`methods for efficient construction of humanized
`MAbs are limited. Third, unpredictable immunoge-
`nicity may result when a new amino acid sequence is
`introduced to balance affinity retention. Fourth, the
`antibody repertoire is limited to the animal in which
`the progenitor MAb originated. Fifth, the rodent
`MAb producing hybridoma must be isolated and
`thoroughly characterized.
`However, obstacles to humanization are gradually
`being surmounted. Karpas et al15 reported creating
`a HAT-sensitive and ouabain-resistant human my-
`eloma cell line that can fuse with human lympho-
`blast cells. A HAT-sensitive subline of the myeloma
`cells that secreted only ␭light chains was fused with
`EBV-transformed white blood cells that produced
`IgG MAbs to HIV-1 gp41. Eventually, a clone was
`isolated that was polyethylene glycol-resistant and
`would not revert when placed in HAT medium.
`Standard polyethylene glycol fusion protocol was
`followed to fuse the myeloma cells with both EBV-
`transformed white cells and fresh white cells ob-
`tained from the peripheral blood of adults and tonsil
`cells from 2 children. The authors reported promis-
`ing rates of hybridoma formation, stable Ig produc-
`tion, and high yield of secreted antibodies compared
`with antibody-producing cell lines from mouse my-
`eloma cells. The authors have now developed 40
`hybridomas that have been secreting cells for more
`than 5 months.
`A human myeloma cell line may eventually prove
`useful in creating MAbs against certain autoim-
`mune diseases and cancers. This technique may also
`make it easier for other researchers to generate
`human MAbs for use in therapy.
`
`Antibody Function in Immunotherapy
`Antibodies may act directly when binding to a
`target molecule by inducing apoptosis, inhibiting
`cell growth, mimicking or blocking a ligand, or in-
`terfering with a key function.53,59 In addition, anti-
`bodies may modulate or potentiate drugs or other
`therapies. The antibody may itself act as an effec-
`tor—as in antibody-dependent cellular cytotoxicity
`(ADCC) or antibody-dependent complement-medi-
`ated cascade—or it may involve effector elements
`such as cytotoxins, enzymes, radioactive isotopes,
`signals for other parts of the immune system, and/or
`cytotoxic drugs.49,50,60 – 62 ADCC occurs if Fc regions
`on the antibodies are recognized by receptors
`present on cytotoxic cells, such as natural killer
`cells, macrophages, granulocytes, and monocytes.
`Complement-mediated cytotoxicity ensues if the an-
`tibody binding prompts a complement cascade to
`
`Berger, Shankar, and Vafai
`
`occur. Active immunotherapy can be accomplished if
`the antigen can provoke a long-lasting T-cell re-
`sponse, which may be achieved by administering
`whole cancer cell extracts or by using small anti-
`genic peptides isolated from tumors in experimental
`patients or animals.58 For instance, MAbs mimick-
`ing breast cancer-specific antigens elicit anti-idio-
`type antibodies; this is another way of creating ac-
`tive immunity, which can lead to a humoral auto
`antibody-like immune response.58
`The use of bispecific molecules that recognize an-
`tigens on both the target cell and effector cell can
`increase ADCC. The most effective mechanisms are
`blockade of a crucial ligand or growth factor or
`ADCC in which tumor cells are killed by Fc receptor-
`bearing cytotoxic effectors. Cell killing by ADCC is
`proportional to the amount of antibody bound to a
`cell, whereas the blockade of an essential growth
`factor may not show effects until most of its receptor
`is saturated. A higher antigen expression by the
`target cell will increase antibody binding and sub-
`sequent ADCC. But high receptor expression will
`also make it difficult to prevent the binding of a
`cytokine or ligand at minimum threshold.53,59,63
`With a bispecific antibody, the specificity of MAb
`is combined with the cytotoxicity of immune effector
`cells, for instance, to neutralize a tumor.9,45,63,64
`Bispecific antibodies link the tumor cell directly to
`the killer cell via cytotoxic trigger molecules, such as
`T-cell receptors or Fc receptors, leading to lysis
`and/or phagocytosis by the effector cells. Although
`bispecific antibodies can enrich effectors at the tu-
`mor site and activate tumor bound effector cells by
`enabling cross-linking between effector and target
`cells, they can cause system-wide immune activa-
`tion because of T-cell receptor cross-linking. Bispe-
`cific antibodies can also mediate cellular cytotoxicity
`via various effector cells, including phagocytes, nat-
`ural killer cells, and T-lymphocytes.
`Another way to use the binding properties of an-
`tibodies is by conjugating antibodies to cytotoxic
`drugs, radioisotopes, or toxins.49,58,63,65,66 Tech-
`niques for conjugation have been described in sev-
`eral recent reviews and articles.49,58,67,70 MAbs can
`be conjugated to chemotherapeutic drugs such as
`doxorubicin, mitomycin, and methotrexate. Chemo-
`therapeutic agents constitute cytotoxic or cytostatic
`drugs that can be conjugated to antibodies; thus far,
`however, these drugs have shown poor specificity for
`target cells and frequently lead to toxicity.55,62,65 Of-
`ten, antibodies may lose reactivity upon conjugation
`with such agents. Immunotoxins used in cancer ther