Jump to ContentJump to Main Navigation
The Lock and Key of MedicineMonoclonal Antibodies and the Transformation of Healthcare$

Lara V Marks

Print publication date: 2015

Print ISBN-13: 9780300167733

Published to Yale Scholarship Online: January 2016

DOI: 10.12987/yale/9780300167733.001.0001

Show Summary Details
Page of

PRINTED FROM YALE SCHOLARSHIP ONLINE (www.yale.universitypressscholarship.com). (c) Copyright Yale University Press, 2018. All Rights Reserved. Under the terms of the licence agreement, an individual user may print out a PDF of a single chapter of a monograph in YSO for personal use (for details see www.yale.universitypressscholarship.com/page/privacy-policy). Subscriber: null; date: 16 January 2019

Antibody Engineering

Antibody Engineering

A Renaissance for Mab Therapeutics

Chapter:
(p.159) Chapter Eight Antibody Engineering
Source:
The Lock and Key of Medicine
Author(s):

Lara V. Marks

Publisher:
Yale University Press
DOI:10.12987/yale/9780300167733.003.0008

Abstract and Keywords

This chapter discusses the emergence of engineered monoclonal antibodies (Mabs) in the 1990s. The failure of Centoxin to gain FDA approval prompted many in the biotechnology and pharmaceutical industry as well as the financial world to view Mab drugs as a lost cause. However, things began to change in 1994 with the approval of ReoPro. The drug signified a major engineering revolution in Mabs after the 1980s, when various competitive and complementary engineering methods began to be developed in the academic and corporate worlds. The first Mabs—known as murine Mabs—were produced by fusing myeloma tumor cells taken from mice with spleen cells derived from other mice or rats previously immunized with an antigen. While these could be made to target almost any antigen and in vast quantities, they were considered foreign by the human body, causing patients treated with murine Mabs to experience immune reactions. This also led to the rapid destruction and clearance of the Mab from the body before it could have its full therapeutic effect. Faced with these obstacles, scientists turned to genetic engineering to transform animal antibodies into human ones.

Keywords:   Mabs, monoclonal antibodies, Mab therapeutics, Mab drugs, drug development, ReoPro, antibody engineering, animal antibodies, murine Mabs, genetic engineering

AFTER THE CENTOXIN DISASTER, many in the biotechnology and pharmaceutical industry as well as the financial world viewed Mab drugs as a lost cause. This was to change in 1994 with the approval of ReoPro. Half-mouse and half-human in structure, its approval signified a major engineering revolution in Mabs since the 1980s, when various competitive and complementary engineering methods began to be developed in the academic and corporate worlds. During this era, some researchers distrusted Mabs whereas others believed that their therapeutic power and safety could be enhanced by improving their specificity and binding capabilities.

From the time of Köhler and Milstein’s breakthrough, scientists had been seeking to improve the technique. Milstein captured its limitations: “All that we seem to have acquired is the potential ability to select from an animal any of the antibodies of his repertoire. It is somewhat like selecting individual dishes out of a very elaborate menu: antibodies ‘à la carte’…. A gastronome worth his salt … wants to experiment with new ingredients, new combinations. His dream is to invent new dishes and not only to taste what others are doing. I am sure our next step will be to move from the dining table, where we order and consume our antibodies ‘à la carte’ to the kitchen, where we will attempt to mess them up.”1

(p.160) The first Mabs were produced by fusing myeloma tumor cells taken from mice with spleen cells derived from other mice or rats previously immunized with an antigen. These Mabs were known as murine Mabs, denoting their rodent origin. While they could be made to target almost any antigen and in vast quantities, they were considered foreign by the human body. As many as half of patients treated with murine Mabs experienced immune reactions, which not only could endanger the patients, but also led to the rapid destruction and clearance of the Mab from the body before it could have its full therapeutic effect.2

Producing human Mabs was a major challenge. A human’s immune system is intrinsically tolerant of most human antigens so will not produce many antibodies of use for therapy. Moreover, humans cannot legally be immunized and manipulated like animals and the process of immunization carries risks. One solution was to create hybridomas from antibodies isolated from the blood of individuals already exposed to particular antigens like cancer. Yet such hybridomas were unstable and stopped secreting Mabs after some time. Another approach was to fuse the Epstein-Barr virus with immortal human lymphocyte B cells taken from healthy volunteers, but this required humans to be immunized and only provided low yields of Mabs.3

Faced with these obstacles, scientists began wondering if genetic engineering could transform animal antibodies into human ones. One of the first people to suggest this was Milstein, who believed such a technique would free scientists from merely immortalizing naturally occurring antibodies, in the way that his and Köhler’s technique did, and allow them instead to design tailor-made antibodies, including human ones. Milstein made relatively modest efforts in this area himself, but inspired others—many of them based in the Laboratory of Molecular Biology (LMB) in Cambridge, England—to take up the gauntlet.4

Those pursuing the task were helped by the antibody’s basic uniform structure (see Figure 1.2). An antibody is a Y-shaped molecule consisting of two large, identical polypeptides, called heavy chains, and two smaller identical polypeptides, called light chains. The chains have both a variable region, located at the tip of the Y arm, which is responsible for binding different antigens, and a constant region, at the base of the Y, which is responsible for recruiting the body’s other resources (including natural killer cells, macrophages, and other “effector cells”) to destroy an (p.161) antigen. This modular structure means that genes present on the variable region of one antibody can be cut and pasted on to the constant region of another.5

For those familiar with DNA recombinant technology and steeped in antibody research and immunology, shuffling genes was merely a way of mimicking the immune system’s natural process, which was first described by Nobumichi Hozumi and Susumu Tonegawa at the Basel Institute of Immunology in 1976. Also known as immunoglobulins, antibodies are divided into five major classes, the main ones being IgM and IgG. Each class has a unique chemical structure and specific function. The IgM class is the first antibody produced by the immune system when it encounters a new antigen. Over time the immune system converts these into IgG antibodies. Such switching confers long-term immunity and is achieved by the rearrangement of genes between the variable and constant regions. Only three or four separate DNA segments are responsible for the vast array of antibodies that the body generates against multiple foreign invaders. Consequently, scientists could potentially combine segments of one antibody gene with the segments of another to create new antibodies.6

To get the ball rolling, scientists needed to find a way of engineering the appropriate rDNA for an antibody and then inserting it into a myeloma cell to facilitate transcription and translation of that gene (gene expression). The task was a major undertaking. Michael Neuberger, an immunologist and biochemist at the LMB inspired by Milstein, struggled for three years from 1980 to amalgamate the genes of the constant and variable regions. For Neuberger it was a “labor of love.” He could not just pluck a standard DNA fragment from a laboratory shelf. First he needed to identify a variable region of interest, create a library of DNA fragments taken from this region, and then clone the DNA of interest. Next he had to put the DNA into a plasmid—an indepen dent, self-replicating DNA molecule—which was inserted into a myeloma cell to express the recombinant gene and produce the antibody. Introducing DNA into myeloma cells was a challenge. Neuberger attempted various approaches, ranging from packaging the plasmid DNA into a virus, and then infecting the myeloma cell, to using different chemical treatments. He finally achieved the goal by stripping the bacterium wall that contained the plasmid and fusing the resultant “spheroplast” with a myeloma cell by using (p.162) polyethylene glycol—the same strategy deployed for creating hybridomas. Introducing a plasmid containing just the DNA for the amalgamated variable and constant regions proved insufficient. Neuberger discovered he also needed to introduce DNA segments into the plasma to direct the myeloma cell to transcribe the engineered antibody gene efficiently. With little known about the DNA segments regulating antibody gene expression, this step was far from simple. Each segment had to be identified before it was inserted.7

Unaware of Neuberger’s efforts, scientists across the Atlantic were pursuing a similar mission. They included David Baltimore, a Nobel Prize winner, and Douglas Rice, a postdoctoral fellow, both based in the Massachusetts Institute of Technology’s department of biology, as well as Sherie Morrison, who was attached to Columbia University’s genetics department and working with Vernon Oi, who was initially based at Stanford University, and then at Becton Dickinson Monoclonal Center in Mountain View, California. By 1984 both Neuberger and the American scientists had demonstrated the feasibility of making recombinant genes for both the light and heavy chains of an antibody and introducing them into a myeloma cell.8

In contrast to the painstaking and time-consuming work required to insert rDNA into a myeloma cell, the next stage proved relatively simple and quick. It involved shuffling genes from the variable to the constant regions of the antibody. This was achieved by Neuberger together with Terence Rabbitts and other colleagues at the LMB; Morrison and Oi, together with Leonard Herzenberg at Stanford University; and Gabrielle Boulianne, Marc Shulman, and Nobumichi Hozumi at Toronto University.9 Working independently of each other, each of these academic groups took a broadly similar approach. Although they used different mammalian cell expression systems, they each linked a gene segment from a variable region of a mouse antibody to a corresponding gene segment on the constant region of a human antibody. The result, called a “chimeric antibody,” possessed genes that were half-human and half-mouse, taking its name from chimera, a mythological creature made up of parts of multiple animals.10

In parallel with such academic endeavors, two commercial groups, the first at Genentech and the Beckman Research Institute and the second at Celltech, also developed a means of producing recombinant Mabs (p.163) using bacteria expression systems. Over time these methods would bring the companies considerable revenue in royalties from the patents awarded to them. Yet the technique did not produce full and properly folded Mabs and provided a poor yield of Mabs.11

One advantage of the new recombinant Mabs was that their specific effector function could be selected and tailored. A Mab could thus be designed to stimulate the killing of a tumor cell by harnessing the body’s natural complement system. Recombinant Mabs were also less liable to provoke immune reactions in patients. In contrast to murine antibodies, for example, chimeric antibodies were made up of only 35 percent rather than 100 percent mouse proteins. Nonetheless, such antibodies were not fully human, so scientists continued hunting for ways to make them more so.12

One technique that proved crucial in taking the engineering further was site-directed mutagenesis (SDM), a method first conceived of in 1971 by Clyde Hutchison III, an American biochemist and microbiologist based at the University of North Carolina at Chapel Hill and further elaborated by Herbert Scott and Hans Kössel at the University of Freiberg in 1973. The technique had been opened up for wider adoption in 1978 by Hutchinson, together with Michael Smith, a British chemist based at the University of British Columbia. SDM allowed scientists to cause a specific mutation by changing, in a very precise and specific way, part of an organism’s DNA. Before SDM, the only way of achieving such a mutation was the time-consuming process of exposing organisms to radiation or chemicals and then selecting the desired mutant.13

Spearheading the use of SDM in antibody engineering was the LMB-based protein chemist Gregory Winter, who had learned the technique from Mark Zoller, a postdoctoral researcher in Smith’s laboratory (Figure 8.1). For Winter, SDM provided an opportunity to expand the expertise he had acquired in DNA sequencing during his research to understand the pathogenic mechanism of the influenza virus. His aim was to construct a new protein by altering a gene, an idea first floated in 1978. This was a major challenge. Winter explained, “The biggest problem with building a new protein was designing it so that it would fold, and assuming it had done so and not aggregated, that it would fold in such a manner as to be functional. At a lower level was the problem of building the novel gene. At that time only small oligonucleotides were available and (p.164)

Antibody EngineeringA Renaissance for Mab Therapeutics

Figure 8.1. Left to right: César Milstein with Michael Neuberger and Gregory Winter. The colleagues helped reengineer Mabs to make them safer and more clinically effective.

(MRC Laboratory of Molecular Biology)

they had then to be assembled into larger blocks which were then put together.”14

Winter believed the task would be easier if he used a protein with a basic scaffold amenable to manipulation. By an ironic stroke of fate, his identification of such a protein was helped by a terrible act of violence: he was injured when an assailant smashed an iron bar over his head and twisted his arm out of its socket. After losing the use of his right arm, Winter immersed himself in learning a computer-graphics system to continue his protein chemistry research at a desk. This allowed him to investigate the 3D structure of different proteins captured with x-ray crystallography. From this, he determined that an antibody had the best scaffold. It had three loops on its heavy and light chains that could be used to hang new loops specifying new binding or catalytic activities.15

In 1983 Winter, encouraged by Milstein and the fact that considerable data on the DNA sequences of antibodies already existed in the public domain, began planning his experiment. Many around him were skeptical about his venture. While scientists were beginning to play around with whole variable regions of antibodies, as in the case of chimerics, few (p.165) imagined that a functional antibody could be created by taking just parts of the variable region. Winter had an important advantage, however: by 1984 he was sharing an office with Neuberger, who was willing to share his newly expressed recombinant antibody genes and constructed chimeric antibodies.16

Winter first needed to demonstrate that the loops of one antibody’s variable region could be taken off and put on to another without affecting the antibody antigen-binding function. If this worked, Winter could construct a new protein by “simply mucking around with the loops” and not worrying about the rest of the molecule. To get the ball rolling, Winter looked to create a human antibody “by stealing only the antigen-binding site (rather than the entire variable domains) from the mouse antibody.” If this could be achieved, Winter would not only prove his theory, but also create a means to make rodent antibodies more humanlike. The major challenge was identifying which residues in the sequence of the mouse antibody were required for antigen binding. This task was daunting because it necessitated synthesizing DNA from the entire variable region of an antibody, and no sequence that long had been synthesized before. Winter would have to first determine the DNA sequence of the mouse antibody’s variable region, then synthesize a new gene that encoded the loops from the mouse antibody and the scaffold from a human antibody. Because he did not have access to an automated DNA synthesizer, he and his technician were forced to sit for many hours at a time, turning a switch every few minutes to keep a machine running. In the end, the work took a couple of years to complete, with two people working full-time and most of Winter’s laboratory resources devoted to it.17

By 1986 Winter and his team had demonstrated the feasibility of building a new antibody by grafting the antigen-binding loops, or complementarity-determining regions (CDRs; see Figure 1.2), from a mouse antibody into a human antibody. Their Mab was directed against a hapten, a nonprotein molecule that elicits an immune response when attached to a large carrier such as a protein. Winter’s Mab had been achieved by capitalizing on Neuberger’s previous development of a chimeric Mab targeting a hapten. The team filed for a patent, explaining that as the first “humanized” Mab it represented a technical breakthrough in antibody engineering. The technique, dubbed CDR grafting, reduced the mouse component of a Mab to just 5 percent.18

(p.166) It remained to be seen, however, whether CDR grafting could be used to develop a Mab of clinical importance. Winter did not have to look far to start answering this question: just across the road, Waldmann’s group was then trying to find a way to reduce the foreignness of Campath-1G, which was thought to be causing side-effects in patients. This they were doing by exploring the immunogenic effects of different therapeutic antibodies in animals and the chimerization of Campath-1G. The work was being done by the biochemist Michael Clark; Marianne Brüggemann, a geneticist and immunologist; and Mark Frewin, a technician; in collaboration with Neuberger’s team. Upon learning of Winter’s new humanizing approach, the team soon established a collaboration with Winter and his postdoctoral fellow Lutz Riechmann, a molecular biologist, to humanize Campath-1G. The work began in 1986 and involved first sequencing and cloning the variable region of the rat Campath-1G, then making some changes to the human antibody’s basic framework in order to produce a Mab with the desired binding activity. Proceeding in painstaking steps, with each intermediate Mab being expressed and tested for activity, in 1988 the collaborators finally produced Campath-1H, the first humanized rodent antibody with therapeutic potential.19

When used a few months later to treat two patients with non-Hodgkin’s lymphoma, Campath-1H proved a remarkably good treatment, going well beyond expectations. Within forty-three days of treatment, Campath-1H had destroyed a large part of the tumor mass in the treated patients. Just as important, it did not seem to provoke any negative immune reactions. Although the lack of immune response could have been due to the immunosuppressive nature of non-Hodgkin’s lymphoma as opposed to the humanization of the Mab, the testing did demonstrate that humanization did not undermine a Mab’s efficacy.20

Despite Campath-1H’s success, researchers elsewhere initially struggled to apply CDR grafting, in part because the original scaffold of the antibody was not as static as Winter envisaged. Investigators soon discovered that inserting novel sequences of DNA within the CDR domain of an antibody could undermine the molecule’s stability. One solution was developed by Cary Queen, who in 1986 helped found Protein Design Labs, a biotechnology company located in Palo Alto, California. In 1989, Queen and his colleagues reported having stabilized a Mab by restoring some of the residue sequences found outside of the CDR region. Following (p.167) this breakthrough, they developed daclizumab, the first humanized Mab to win FDA approval. Designed to control the acute rejection of kidney transplants, it was marketed with the help of F. Hoffmann–La Roche. Within a short time other researchers were humanizing other potentially therapeutic Mabs, using and advancing Winter’s and Queen’s techniques.21

In 1987, Winter, not content with just humanizing rodent antibodies, launched a project to create an artificial human immune system to produce fully human antibodies. The first stage involved the generation of a large library of human antibody fragments. This was done using polymerase chain reaction (PCR), a laboratory technique first devised in 1983 that allowed for the multiple reproduction of very small samples of DNA to produce billions of copies. By 1988 Winter had adopted PCR to amplify and clone the genes of the heavy and light chain variable regions of murine antibodies, for which he filed a patent. In addition to facilitating the creation of a library of fragments, the technique eliminated the laborious and time-consuming steps involved in the sequencing and cloning of genes—steps then impeding the production of chimeric and humanized Mabs.22

In addition to PCR, Winter looked to phage display for his project. Originally developed to display small peptides in 1985, the technique deploys phages, viruses that infect bacteria, to connect proteins with genetic information. Upon replication, such modified viruses produce enormous, diverse libraries of proteins. Promising the rapid identification and isolation of proteins specific to any target of interest from a library of millions of different proteins bound to phages, phage display offered Winter a means to generate human antibodies. He soon discovered, however, that he had a major competitor: a large group of scientists headed by Richard Lerner and Carlos Barbas at the Scripps Research Institute, San Diego, who had been given a pre-print of Winter’s work by Lutz Riechmann, who was now working at the institute. Significantly, the Scripps team had the backing of Stratagene, an American biotechnology company set up in 1984 to exploit new antibody engineering techniques.23

What alarmed Winter most was that the Scripps team reported that they were ready to file for a patent. Desperately needing a way to boost his resources, Winter decided to create a new company, Cambridge Antibody Technology (CAT). This he did in 1989 with the support of (p.168) Aaron Klug, then the LMB director. The company was founded not for commercial reasons, but as a means to attract financial investment in Winter’s research. Joining Winter in the venture was David Chiswell, former product development and research manager at Amersham International. CAT quickly attracted investment from Peptide Technology Australia (Peptech), a biotechnology company whose founder, Gregory Grigg, had had spent time in Fred Sanger’s Laboratory at the LMB. Peptech agreed to provide £750,000 in exchange for a 40 percent equity stake in CAT. Additional funds were secured through equity investments from private individuals and investment funds. Both the MRC and Winter were given a stake in the company.24

By 1990, Winter and his CAT colleagues had developed a platform that enabled scientists to generate enormous, diverse libraries of randomly shaped human antibodies. How the system worked was explained as follows: “A phage antibody is … analogous to a B lymphocyte in that it displays an antibody on its surface and carries the genes encoding that antibody. It can be selected for its ability to bind a specific antigen and subsequently amplified by reinfection of bacteria. The antibody genes can then be rescued and used to produce soluble antibody fragments or even complete antibodies.”25

Now scientists could pour a “library” of phage antibodies through a column to which a target had been fixed that bound to specific phages. The target could be a protein expressed on a cancer cell or a molecule known to cause inflammation. Mimicking the immune system, this platform provided a means to select the most appropriate antibody fragments for a particular antigen, which could then be refined and made into a Mab. It marked a major turning point in the engineering of Mabs. Gone were the days of relying on an animal or human’s natural immune system, with all their limitations. The platform provided much greater power for tailoring the specificity and affinity of the Mab and offered the possibility of building synthetic Mabs with less immunogenicity. Moreover, Mabs of interest could be generated in two weeks and the process was amenable to automation.26

In parallel with the development of the phage display technique, another method was emerging in Cambridge to produce human antibodies based on the research of Neuberger and Brüggemann into the regulatory mechanism underlying the generation of different antibodies. (p.169) This work involved genetically engineering a transgenic mouse to automatically produce human antibodies when immunized. Begun in 1986, this project was helped by Azim Surani from the Babraham Institute, Cambridge, a pioneer in animal genetic engineering. Scientists had been making transgenic mice successfully since the mid-1970s. In general terms, the technique entailed introducing foreign DNA segments into the germ line of early mouse embryos, which would then pass this recombinant DNA on to their offspring. The technique inactivated a mouse’s own genes or introduced new genes.27

To start the ball rolling, the Cambridge team set about introducing gene segments from a human antibody into the DNA of early mouse embryos to create mice that could produce human antibodies after immunization. Such an idea was not new, having been floated as a possibility by Columbia University scientists in 1985. Nobody, however, had yet succeeded in such an endeavor. The major challenge was that human antibody loci are very large, containing many gene segments scattered over a few million base pairs of DNA.28

To make the task more manageable Neuberger and Brüggemann, with the help of Gareth Williams, set about assembling a miniature version, or minilocus, of the gene segment that codes for the heavy-chain domain from human antibodies. Making a minilocus was complex. The assembly of such genes in 1986 was much more challenging than that involved in making chimeric antibodies. As Neuberger recalls, “It meant obtaining all the various bits (multiple variable regions and joining segments, synthesizing diversity segments, adding a constant region as well as what we guessed would be sufficient gene regulatory sequences) and assembling them together.” The process did not require any new inventions, but it “was laborious and long-winded.” Using existing methods for gene cloning and assembly technology, the mission took about a year to complete. The result was a hybrid mouse and human-heavy-chain minigene construct. Once made, the minilocus had to be micro-injected into mouse eggs. This was not easy because it was a large DNA molecule and needed to go down a very fine needle. To prevent the DNA from being sheared, Surani attempted various solutions, such as adding spermine to compact the DNA.29

The team was uncertain whether the minilocus, once inserted, could rearrange itself to generate the repertoire of different human antibody (p.170) heavy chains in the mouse. Two French groups had proved in 1987 that rabbit or chicken genes for the light chain part of the antibody could assemble when injected into mice. Yet this was a far cry from what the Cambridge scientists were trying to achieve. Importantly, the minilocus for the heavy chain of an antibody was much larger, contained many more segments, and underwent more complex rearrangements than the light chain. There was no guarantee that the miniaturized version contained all the segments necessary to get successful rearrangement and expression. And moreover, once reassembled in the mouse, no one could be sure that the genes would interact with the necessary signaling components in the mouse’s lymphocytes so that, following immunization with an antigen, the mouse would produce human antibodies specific to that antigen. To the team’s relief, the inserted human antibody genes did indeed rearrange appropriately, enabling the expression of antibodies with human heavy chains. The group filed for a patent on the technique in 1988. Much more work was needed, however, before mice would readily produce the range of human Mabs with both heavy and light chains of human origin that would be suitable for therapeutic use.30

While Neuberger and Brüggemann continued developing their technique, they realized that its commercial development required far more resources than were then at their disposal. Unlike Winter, who set up his own company to advance his development of phage display Mabs, Neuberger and Brüggemann opted to provide broad nonexclusive licenses to companies prepared to take the technology further. This, they believed, would help accelerate the development of their transgenic mice and provide returns should any successful commercial products be produced. Yet few companies were prepared to invest in the technology. Even CAT, on whose scientific advisory board Neuberger served, decided not to pursue this course so as not to dilute its development of phage display. In the end only GenPharm and Cell Genesys, two Californian startups already developing their own transgenic mice, licensed the technology.31

By the end of 1980s it seemed that scientists would soon have transgenic mice to produce human antibodies. This was due not only to the progress in Cambridge, but also to the development of “gene-targeting technology” at the University of Utah, MIT, and Columbia University between 1988 and 1989, which allowed specific modifications to be made (p.171)

Antibody EngineeringA Renaissance for Mab Therapeutics

Figure 8.2. Nils Lonberg, n.d.

(Nils Lonberg)

in the mammalian genome through the insertion of genes at precise points in an animal’s DNA.32

Despite the optimism, certain obstacles remained. Those taking up the challenge included the team in Cambridge as well as those based at GenPharm, led by Nils Lonberg, and Cell Genesys, led by Aya Jakobovits (Figure 8.2). All three groups pursued a fairly similar strategy of creating two different strains of mice for cross-breeding. The first mouse had its genes modified to inactivate its immune system. This blocked its capacity to produce its own antibodies following immunization, and encouraged production instead by the transgenic human antibody gene. In the second mouse larger segments of the human antibody gene loci were introduced to facilitate the production of a wider range of human Mabs. Once cross-bred, transgenic mice could generate antibodies identical to human antibodies.33

Lonberg’s experiences reveal some of the twists and turns behind the development of such transgenic mice. A chemist by training, with a Harvard University doctorate in molecular biology, Lonberg had first begun wondering about the possibilities of engineering recombinant proteins for therapeutics in 1984, when he was based at Biogen with his colleague Harry Meade. His interest was prompted by a contract that Biogen had (p.172) to supply large amounts of recombinant protein. Because it was a large molecule, the protein could be expressed only through recombinant mammalian cells in roller bottles that took up an enormous amount of space. As Lonberg recalled, “Every spare closet got turned into a warm room and got filled with these roller bottles just to try to make enough protein to fulfill the company’s contract.” Lonberg and Meade wondered if a more efficient production system could be created by genetically engineering a cow to produce recombinant protein in its milk. Most of their Biogen colleagues, however, laughed at the proposal.34

In 1985 Lonberg joined the Memorial Sloan-Kettering Cancer Center, New York, as a postdoctoral scientist. Surrounded by pioneers in the generation of transgenic mice, including Elizabeth Lacy and Frank Costantini, Lonberg succeeded in genetically engineering mice to produce human proteins in their milk, which led to his being awarded the third patent ever granted for a transgenic animal. Sloan-Kettering was not only an ideal location for gaining expertise in transgenic animals; it also exposed Lonberg to Mab therapeutics. He first came into contact with them through his oncologist brother, who was also based at Sloan-Kettering. While clinically testing a mouse antibody developed by Lloyd Old and Alan Houghton for treating melanoma, his brother noted how some patients’ immune responses to its mouse component were causing the rapid disappearance of the Mab in the body, thereby reducing its effectiveness. The patients with the best outcomes were those with the most delayed response to the Mab. Nils Lonberg wondered if transgenic mice could offer a way to offer this delayed response.35

In 1989, Nils Lonberg joined GenPharm, which provided the resources to begin developing transgenic mice to produce human Mabs. GenPharm allocated $1.3 million to the project and he secured additional NIH money. This allowed for a team of fourteen. The project involved creating two varieties of gene-modified mice. The first was to have some of its genes inactivated and the second was to have its genes modified with fully human antibody genes to facilitate production of fully IgG human antibodies. These two mice were then to be cross-bred to create a transgenic mouse with a human immune system that could generate antibodies identical to human antibodies. By late 1993, Lonberg’s team had generated a mouse with a human immune system capable of producing fully human IgG antibodies with high affinities for their targets. Shortly (p.173) afterward the scientists in Cambridge and at Cell Gensys reported similar success.36

The major advantage of transgenic mice was that scientists could now easily produce fully human Mabs with enhanced affinity for a target without spending hours humanizing or optimizing an antibody molecule. Everything could be achieved within the mouse. Following immunization, the humanized mice produced target-specific human Mabs without any subsequent manipulation. This meant that Mab therapeutics could be developed much more quickly than before.37

Advances in Mab engineering in the twenty years since Milstein and Köhler’s invention had radically transformed the antibody molecule, helping to reduce its mouse component to almost nothing and make antibodies more compatible with the human body (see Table 8.1 and Figure 8.3). Following these developments, scientists began to investigate the possibility of using just the active portion of an antibody to create miniature Mabs. This was highly desirable because large quantities of Mabs had to be injected to achieve clinical efficacy. The high volumes needed for therapy posed significant formatting and manufacturing complications. As explained in chapter 4, Mab production required not only very large cultures of mammalian cells, but also extensive purification. All of this added to the costs of production, which raised the market price of the final therapeutics and so constrained their clinical application. If smaller fragments of antibodies could be deployed, scientists would have a means to improve production yields and the possibility of developing new and more versatile antibody formats to boost therapeutic efficacy.38

One of the first researchers to investigate the development of miniature antibodies was Winter, who noticed in 1989 with LMB colleagues that fragments, dubbed domain antibodies (dAbs), sometimes bound independently to antigens. These came from variable domain and heavy chains of antibodies produced from immunized animals. They hypothesized that such dAbs might have greater power to penetrate tissues. Moreover, dAbs offered the possibility of better targeting of pathogenic viruses, which remained inaccessible to full-size Mabs due to the narrow cavities on their surface antigens.39

The team soon discovered that dAbs could be generated without the immunization of animals. Advancing such antibodies, however, proved slow, because the single domains tended to be “sticky” and clump together. (p.174)

Table 8.1 Percentage of mouse and human protein components in different forms of Mabs

Form of Mab

Mouse Protein

Human Protein

Murine

100%

0%

Chimeric

65%

35%

Humanized

5%

95%

“Fully human”

0%

100%

Source: W. R. Gombotz and S. J. Shire, “Introduction,” in S. J. Shire et al., eds., Current Trends in Monoclonal Antibody Development and Manufacturing (New York, 2009), 3.

Furthermore, they did not always retain the affinity of their parent antibody and were not very soluble. Based on these features, dAbs seemed doomed to remain just a laboratory curiosity. In 2000, however, Winter’s team finally found a way to select single-domain antibodies that did not stick together, laying the foundation for a new company, Domantis. The new antibodies offered both a more effective therapy and a means to bypass the dominant patent monopoly that by this point Genentech and other companies held on Mab engineering.40

While a number of different engineering methods had been developed for Mabs by the end of the twentieth century, a number of academic-based scientists remained skeptical about the advantages of one technology over another in terms of reducing immunogenicity. Clark, who had played a pivotal role in the humanization of Mabs, believed that in terms of reducing immunogenicity there was little difference among the first chimeric antibodies, humanized antibodies, or “fully human” antibodies. From his perspective the new innovations were little more than the “Emperor’s new clothes.” Similarly, Geoff Hale, who like Clark had helped develop the first humanized Mab for therapy, believed that the large intellectual investment in antibody engineering to eradicate the unwanted immunogenicity of Mabs had on the whole been “a distraction rather than a stimulus to progress.” Neuberger, who had helped develop both chimeric antibodies and transgenic mice to produce fully human antibodies, (p.175)

Antibody EngineeringA Renaissance for Mab Therapeutics

Figure 8.3. Different generations of Mabs showing murine and human protein components

claims that the key reduction in immunogenicity was achieved early on with chimeric antibodies, because these replaced the constant regions of the antibody. How far the new methods—such as CDR grafting, phage antibodies, and transgenic mice, which tinkered with the variable region—further reduced the immunogenicity of Mabs is, he claims, difficult to know. The difficulty is that a properly controlled experiment to answer this question cannot be carried out.41

Not everyone holds such views. Lonberg, who is deeply involved in pharmaceutical development, while agreeing with Neuberger that it is difficult to perform an experiment to test how far newer techniques have reduced unwanted immunogenicity, believes that the clinical trials conducted with more recently developed CDR antibodies and with transgenic mice have shown the new techniques to have achieved significant progress in this area.42

One difficulty in determining how far antibody engineering has helped reduce immunogenicity is that the genes of the variable region of any therapeutic Mab, whether it be chimeric, CDR-grafted, or derived using the phage system or humanized mice, will have multiple mutations. (p.176) Consequently these genes will differ from the antibodies already existing in a nonimmunized patient and so any therapeutic Mab will be seen as slightly foreign by a patient’s immune system.43

The debate over the relative merits of chimeric versus humanized or human antibodies is also relevant in the case of antibody fragments. While fragments permit better penetration of tissue, they can be filtered out by the kidney more easily than full-size Mabs, resulting in more rapid clearance from the body, so that they have a reduced time to work. In some circumstances, such as when they are delivering cytotoxic radioisotopes to destroy a tumor, their rapid clearance from the body can actually be advantageous. Yet in other situations it can be a disadvantage, for example when it prevents the targeted site from being sufficiently affected. Another disadvantage with fragments is that they are unable to stimulate the natural effector functions of the immune system, which sends out its own antibodies or complement in response to threats. Within the context of production, fragments can also be more likely to form undesirable clumps and be less stable.44

The use of the different engineering approaches has largely been driven by the accessibility to patents for each approach. In contrast to Milstein and Köhler’s invention, which was never patented, the number of patents governing subsequent engineering methods is vast. Almost every particular of Mab creation and manufacture is now protected by a patent that has to be licensed if used. In 1986 it was estimated that there were only 830 patents relating to the hybridoma technology. This number grew exponentially thereafter, making the field difficult and expensive to navigate for anyone wishing to enter the space. Even for those deeply embedded in the science and the commercialization of Mabs, the abundance of patents can result in companies becoming embroiled in battles to maintain rights to their intellectual property.45

Some idea of the pain that patents can inflict is borne out by the example of GenPharm. In February 1994, just days before it was due to file for its initial public offering (IPO), the company was sued by its rival Cell Gensys on the grounds that it had stolen a trade secret for inactivating a mouse gene. Many within GenPharm initially did not take the legal action seriously, nor did other scientists such as Neuberger. Nonetheless, Cell Gensys was a major threat given that it had the financial backing of Japan Tobacco and had already gone public. GenPharm countered with (p.177) an antitrust lawsuit and two patent lawsuits. The litigation dragged on for three long years and cost GenPharm dearly: it prevented it from going forward with its IPO, thereby crippling it financially, and it undermined efforts to find a buyer to rescue the company. GenPharm’s ability to raise cash was not made any easier in the wake of the Centoxin disaster. Having raised over $75 million between 1988 and 1994, GenPharm soon found that only $15,000 remained. In order to survive, its executive was forced to get rid of all but a few of the 110 staff, as well as most of the furniture, laboratory equipment, and patents. Owing large sums of money to lawyers and banks, they thereafter turned to bartering and looking after the mouse cages of another biotechnology company in exchange for the use of a small amount of laboratory space. Eventually, in January 1997, the two companies negotiated a cross-licensing agreement. In the aftermath of the legal battle GenPharm was acquired by Medarex, an antibody company based in Annandale, New Jersey, and Cell Gensys was granted the approval for the first fully human antibody, panitumumab (Vectibix), which was produced from transgenic mice.46

Patents alone did not determine which new engineering methods were employed for the formulation of drugs. Also crucial was how well the Mab drugs they helped generate performed in clinical trials and navigated through the regulatory framework. Figure 8.4 highlights how quickly each method evolved from materialization of the full invention to the first marketable therapeutic. On average this process took ten to twelve years from the time when the invention first became practical. In the case of the transgenic mice the technique took much longer to mature than is suggested by the timeline. As highlighted earlier, transgenic technology was first proposed as a means to generate human antibodies as early as 1985, but it took another nine years before this was realized in practical terms and a further twelve years before a therapeutic arrived on the market. Part of this time lag is attributable to the considerable work involved in engineering the right mouse, and mouse-based experiments take time to complete. But while the transgenic technology took longer to develop than other engineering techniques, once achieved, it provided the means to produce Mab therapeutics more quickly. With murine-based Mabs, all that is needed is to immunize the transgenic animal, fuse its spleen cells with a myeloma, and then screen for the desired Mab. For this reason, transgenic mice are likely to be the preferred route for producing future Mab therapeutics.47

(p.178)

Antibody EngineeringA Renaissance for Mab Therapeutics

Figure 8.4. Timeline for Mab engineering techniques from first publication to first drug approval

(Adapted from N. Lonberg, “Human Monoclonal Antibodies from Transgenic Mice,” in Y. Chemojovsky and A. Nissim, eds. Therapeutic Antibodies [Berlin, 2008], 69–97, esp. 71, fig. 1)

Certain Mab types have dominated the therapeutic landscape. Figure 8.5 shows that the FDA initially approved only murine and chimeric Mabs, but from 1997 began also to approve humanized and human Mabs, which were first produced by CDR grafting and then by phage display and later transgenic mice. By 2009 chimeric and humanized Mabs were the predominant form of the twenty-two Mab therapeutics on the market and the twenty-eight that were then in the final stages of clinical development. Humanized and human Mab therapeutics were also the most prevalent form in phase III trials (Figure 8.6). As Figure 8.7 reveals, most of these drugs were created using murine hybridomas rather than humanized mouse hybridomas or phage display antibody libraries, although the last two methods were beginning to be used for drugs entering late-stage clinical trials. Antibody fragments were also beginning to be used in the earlier clinical development stages. Of the 450 Mabs that were in clinical development in 2009, fifty-four (12 percent) were antibody fragments.48

(p.179)

Antibody EngineeringA Renaissance for Mab Therapeutics

Figure 8.5. Number of Mab drugs approved annually by the FDA by antibody type, 1986–2011

(Adapted from N. Lonberg, “From Mouse Embryos to Antibody Therapies,” paper presented to IBC, 6 Dec 2011)

Much of the rapid growth in Mabs in recent years has been in human antibodies (Figure 8.8). Transgenic mice have helped to fuel this growth. In 2010, 34 (60 percent) of the 56 transgenic-produced Mabs were generated by mice created by GenPharm and Kirin Brewery and 18 (32 percent) from the one derived from Cell Gensys.49 Most are intended for cancer treatment. By 2011, five out of the seven approved transgenic-mice Mabs had been developed using the Lonberg and GenPharm platform.50

The success of human Mabs made with transgenic mice stands in contrast to those created with phage display. In 2010, fifty-six human antibodies being clinically tested originated from transgenic mice and thirty-five from phage display. Similarly, in terms of FDA approval, seven of the Mab therapeutics on the market by 2011 came from transgenic mice, while only two had been made using phage display (see Figure 8.5). The difference may be explained by the fact that many companies and academic groups are more familiar with hybridoma technology than phage display, and by CAT’s restrictive licensing policy for its phage technology. Strikingly, both of the phage-display Mabs approved for the market originated from CAT, reflecting the company’s dominance in this area.51

(p.180)

Antibody EngineeringA Renaissance for Mab Therapeutics

Figure 8.6. Number of different types of Mabs on the market or in phase III of clinical testing, 2009

(Adapted from W. R. Strohl, “Therapeutic Antibodies: Past, Present and Future,” in Z. An and W. R. Strohl, eds., Antibodies: From Bench to Clinic [2009], 25)

What is particularly noticeable about the different antibody-engineering techniques that emerged from the 1980s is how many originated and were nurtured by LMB scientists who had been mentored by Milstein. It is also striking how active the MRC has been in guarding the intellectual property of these innovations. This stands in marked contrast to the early days when the first Mabs were deemed unworthy of patenting. The change in attitude occurred in part due to lessons the organization learned from failing to patent Köhler and Milstein’s technique—a costly mistake with political fallout. It was also boosted by the Thatcher government, which encouraged the patenting of innovations arising from research funded by U.K. research bodies so as to enhance the commercial exploitation of British research and make it more competitive.52

In 1982, the MRC established an industrial liaison group to handle the commercial exploitation of scientific innovations. Initially, technology transfers were organized by the British Technology Group (BTG), a nonstatutory body formed in 1981 through the merger of the National Research Development Corporation and the National Enterprise Board, which had the right of first refusal to exploit research funded by research councils. In 1985, however, the British government ended the BTG’s right of first refusal and gave MRC units the means to exploit their own research. (p.181)

Antibody EngineeringA Renaissance for Mab Therapeutics

Figure 8.7. Number of Mabs, by source, on the market or in phase III of clinical testing, 2009

(Adapted from W. R. Strohl, “Therapeutic Antibodies: Past, Present and Future,” in Z. An and W. R. Strohl, eds., Antibodies: From Bench to Clinic [2009])

This gave units like the LMB greater freedom to negotiate industrial agreements and receive greater shares in royalties. Scientists could also now expect to share in the royalties and started to found their own companies to exploit their inventions.53

The LMB moved quickly to patent and license out its scientists’ antibody engineering techniques. First in line was the method for humanizing rodent antibodies developed by Neuberger and Rabbitts. A patent application was filed in 1984 and assigned, under an umbrella agreement, to Celltech. Second in line was the CDR grafting method developed by Winter, for which a patent application was filed in 1986. Its patent process followed a very different course because of Winter’s reservations about giving away the rights to just one company. Based on the frustrations that he and other LMB colleagues had experienced with Celltech in licensing out Mabs for interferon and blood reagents, Winter opted for a nonexclusive licensing agreement. This he did with the support of the LMB and the MRC. Their model was the agreement that Stanford University had used for the Cohen-Bayer rDNA patent.54

The MRC established three principles for the nonexclusive license for the CDR grafting patent: first, the license should be given for the generation of a product that was to be of benefit to patients; second, no (p.182)

Antibody EngineeringA Renaissance for Mab Therapeutics

Figure 8.8. Percentage of four types of Mabs in clinical development, 1990–2008. Due to rounding, the total is greater than one hundred.

(Adapted from figure 1, A. L. Nelson, E. Dhimolea, and J. M. Reichert, “Development Trends for Human Monoclonal Antibody Therapeutics,” Nature Reviews Drug Discovery, 9 [Oct 2010]: 767–74)

commercial secrecy could be enforced; and third, those originally involved in the development of a method could retain the freedom to collaborate with other researchers and publish any future developments on the technology while taking care not to make needless disclosures. Overall the strategy was to license the patents nonexclusively to whomever was interested, with very little up-front payment. The rationale was that very small upfront payments would encourage cooperation. In return for licensing its patents, the MRC asked for royalties on final products. This was generally limited to 2 percent of sales. The MRC stood to benefit if the licensees achieved success.55

The nonexclusive licensing arrangement proved highly lucrative, signifying how widespread the LMB’s antibody engineering techniques had become. Within a short time more than sixty companies worldwide had been granted nonexclusive licenses to LMB patents, and the revenue the MRC gained from these licenses more than made up for the income potentially lost from not having patented Köhler’s and Milstein’s techniques. Some of the greatest revenue came from the CDR grafting, but in time substantial sums also flowed from transgenic mice. In the last years of the patent for the transgenic mice, the licensing revenue exceeded £10 (p.183) million per year. Much of this revenue came from sales of therapeutics based on humanized and human Mabs, which were surging. In 2006 alone, the MRC gained £127 million from the sale of human antibodies and £84 million from humanized antibodies.56

The income for 2006 was boosted by a £121 million deal that the MRC had made the year before with the U. S. pharmaceutical company Abbott Laboratories in lieu of royalties for Humira, the first humanized Mab to win approval. Directed toward treating arthritis, Humira was developed using the MRC’s patent for phage display. Considered one of the world’s largest ever intellectual property deals at the time, it marked how far the MRC had changed its patenting and licensing approach since Köhler and Milstein’s first Mabs. The MRC put the money from the deal into its overall funding pot, which was then £500 million a year. By June 2012 the MRC reported that its portfolio of antibody-engineering patents had generated more than $750 million (£486 million) in royalties, which it was plowing back into medical research.57

The MRC has not been the only beneficiary of royalty payments; so too has the LMB. Starting in the mid-1990s, the LMB began earning just over £1 million a year in royalty income. This rose steeply in 2000, reaching approximately £6 million a year. The largest proportion of this income came from patents taken out on Winter’s techniques. By 2005 the payments had reached £20 million per year, most still from Winter’s patents. This exceeded the total MRC block grant to the LMB. Overall, the LMB’s gross annual income was equal to that of the entire commercial income of all universities in the United Kingdom, and comparable to that of much larger U.S. academic institutions such as MIT. By 2008 the total annual royalties coming to the LMB had risen to £70 million. One indication of just how important these royalties are to the LMB is that in 2013 it was able to move into a new state-of-the-art building paid for mostly from the royalties earned from antibody-related engineering. The building was estimated to have cost over £200 million.58

The large royalties paid to the MRC and LMB show how far antibody engineering had transformed Mab therapeutics. With antibody engineering, scientists and clinicians now had a new form of Mab to play with. It was just the start of another part of the journey, however. Much more work lay ahead as scientists turned to the task of using Mabs to create drugs for clinical use.

Notes:

(1.) C. Milstein, “Monoclonal Antibodies from Hybrid Myelomas,” Proceedings of the Royal Society of London 211 (1981): 393–412, 409.

(2.) M. B. Khazaeli, R. M. Conry, and A. F. LoBuglio, “Human Immune Response to Monoclonal Antibodies,” Journal of Immunotherapy 15 (1994): 42–52.

(3.) K. M. Thompson et al., “The Efficient Production of Stable, Human Monoclonal Antibody-Secreting Hybridomas from EBV-Transformed Lymphocytes Using the Mouse Myeloma X63-Ag8.653 as a Fusion Partner,” J Immunol Methods 94, nos. 1–2 (Nov. 1986): 7–12; J. L. Marx, “Antibodies Made to Order,” Science 229 (2 Aug. 1985): 455–56; G. P. Winter, “Antibody Engineering,” Phil Trans R Soc Lond B 324 (1989): 537–47, esp. 539; N. Lonberg, “Transgenic Approaches to Human Monoclonal Antibodies,” in M. Rosenberg and G. P. Moore, eds., The Pharmacology of Monoclonal Antibodies (Berlin, 1994), 49–101, esp. 50; Emails from Michael Neuberger to author, 31 July 2012, and 5 Aug. 2012.

(4.) M. S. Neuberger and B. Askonas, “César Milstein,” Biographical Memoirs of the Fellows of the Royal Society 51 (2005): 267–89, 278.

(5.) Interview with Michael Neuberger. Marx, “Antibodies,” 455; Winter, “Antibody,” 540.

(6.) N. Hozumi and S. Tonegawa, “Evidence for Somatic Rearrangement of Immunoglobulin Genes Coding for Variable and Constant Regions,” PNAS 73, no. 10 (1976): 3628–32Marx, “Antibodies,” 455

(7.) Interview with Neuberger; Emails from Neuberger.

(8.) D. Rice and D. Baltimore, “Regulated Expression of an Immunoglobulin Kappa Gene Introduced into a Mouse Lymphoid Cell Line,” PNAS 79, no. 24 (1982): 7862–65; S. L. Morrison, V. T. Oi, “Transfer and Expression of Immunoglobulin Genes,” Ann Rev Immunol 2 (1984): 239–56.

(9.) M. S. Neuberger, G. T. Williams, and R. O. Fox, “Recombinant Antibodies Possessing Novel Effector Functions,” Nature 312, no. 5995 (1984): 604–608; M.S. Neuberger et al., “A Hapten-Specific Chimaeric IgE Antibody with Human Effector Function,” Nature 314 (1985): 268–70; S. L. Morrison et al., “Chimeric Human Antibody Molecules: Mouse Antigen-Binding (p.278) Domains with Human Constant Region Domains,” PNAS 81 (1984): 6851–55; G. Boulianne, N. Hozumi, and M. J. Shulman, “Production of Functional Chimaeric Mouse/Human Antibody,” Nature 312 (13 Dec. 1984): 643–46.

(10.) C. Zeller, “The Spatial Innovation Biography of a Successful Monoclonal Antibody,” Espace 5 (20 Jan. 2008): 1–42.

(11.) Interview with Gregory Winter; Emails from Neuberger; S. Cabilly et al., “Generation of Antibody Activity from Immunoglobulin Polypeptide Chains Produced in Escherichia coli,” PNAS 81 (1984): 3273–77; M. A. Boss, J. H. Kenten, C. R. Wood, and J. S. Emtage, “Assembly of Functional Antibodies from Immunoglobulin Heavy and Light Chains Synthesised in E. coli,” Nucleic Acids Research 12, no. 9 (1984): 3791–806; Winter, “Antibody,” 540; E. Waltz, “Industry Waits for Fallout from Cabilly,” Nature 25, no. 7 (2007): 699–700; R. L. Teskin, “It Lives for 29 Years,” Legal Times 26, no. 44 (2004): 1–2.

(12.) W. R. Gombotz and S. J. Shire, “Introduction,” in S. J. Shire et al., eds., Current Trends in Monoclonal Antibody Development and Manufacturing (New York, 2009), 3.

(13.) C. A. Hutchison and M. H. Edgell, “Genetic Assay for Small Fragments of Bacteriophage φ‎X174 Deoxyribonucleic Acid”, Journal of Virology 8, no. 2 (1971): 181–89; H. Schott and H. Kössel, “Synthesis of Phage Specific Deoxyribonucleic Acid Fragments, I: Synthesis of Four Undecanucleotides Complementary to a Mutated Region of the Coat Protein Cistron of fd Phage Deoxyribonucleic Acid,” Journal of the American Chemical Society 95, no. 11 (1973): 3778–85; C. A. Hutchison, S. Phillips, M. H. Edgell, S. Gillham, P. Jahnke, and M. Smith, “Mutagenesis at a Specific Position in a DNA Sequence,” J Biol Chem 253, no. 18 (1978): 6551–60; N. Kresge, R. D. Simoni, and R. L. Hill, “The Development of Site-Directed Mutagenesis by Michael Smith,” J Biol Chem 281 (2006): e31–e33; Anon., “Michael Smith,” online biography available online at http://www.science.ca/scientists/scientistprofile.php?pID=18&pg=1 (accessed 20 Sept. 2014).

(14.) Marks, 20 Jan. 2013G. Winter et al., “Redesigning Enzyme Structure by Site-Directed Mutagenesis: Tyrosyl tRNA Synthetase and ATP Binding,” Nature 299 (1982): 756–58Hutchison et al., “Mutagenesis”Winter, “Antibody,” 537

(15.) Interview with Winter.

(16.) Ibid.; E. Kabat et al., Sequences of Proteins of Immunological Interest, 5th ed. (Bethesda, Md., 1991); G. Hale, “Therapeutic Antibodies—Delivering the Promise,” Advanced Drug Delivery Reviews 58 (2006): 633–39.

(17.) Interview with Winter; Winter, “Antibody,” 540.

(18.) P. T. Jones et al., “Replacing the Complementarity-Determining Regions in a Mouse Antibody with Those from a Mouse,” Nature 321 (29 May 1986): 522–25Gombotz and Shire, “Introduction,” 3

(p.279) (19.) Interview with Michael Clark; R. J. Benjamin, S. Cobbold, M. R. Clark, and H. Waldmann, “Tolerance to Rat Monoclonal Antibodies: Implications for Serotherapy,” JEM 163 (June 1986): 1539–52; L. Riechmann, M. Clark, H. Waldmann, and G. Winter, “Reshaping Human Antibodies for Therapy,” Nature 332 (24 Mar. 1988): 323–24; G. Hale and H. Waldmann, “From Laboratory to Clinic: The Story of Campath-1,” in A. J. Y. George and C. E. Urch, eds., Diagnostic and Therapeutic Antibodies (New York, 2000), 243–66, 250; M. Clark, “Mike’s Campath Story,” available online at http://www.path.cam.ac.uk/~mrc7/campath/campath.html (accessed 20 Sept. 2014).

(20.) G. Hale, “Remission Induction in Non-Hodgkin Lymphoma with Reshaped Human Monoclonal Antibody CAMPATH-1H,” Lancet 2 (1988): 1394–99; Email from Herman Waldmann to Marks, 16 July 2012; L. Marks, “The Life Story of a Biotechnology Drug: Alemtuzumab,” http://www.whatisbiotechnology.org/exhibitions/campath (accessed 20 Sept. 2014).

(21.) L. R. Helms and R. I. Wetzel, “Destabilizing Loop Swaps in the CDRs of an Immunoglobulin VL Domain,” Protein Science 4 (1995): 2073–81; C. Queen et al., “A Humanised Antibody That Binds to the Interleukin 2 Receptor,” PNAS 86 (1989): 10029–33; M. S. Co, M. Deschamps, R. J. Whitley, and C. Queen, “Humanised Antibodies for Antiviral Therapy,” PNAS 88 (1991): 2869–73; S. Dübel, Handbook of Therapeutic Antibodies, vol. 1 (Hoboken, N.J., 2007), 122; J. W. Saldanha, “Humanization of Recombinant Antibodies,” in M. Little, ed., Recombinant Antibodies for Immunotherapy (Cambridge, Eng.), 3–11.

(22.) R. Olandi, D. H. Güssow, P. T. Jones, and G. Winter, “Cloning Immunoglobulin Variable Domains for Expression by Polymerase Chain Reaction,” PNAS 86 (1989): 3833–37

(23.) S. de Chadarevian, “The Making of an Entrepreneurial Science: Biotechnology in Britain, 1975–1995,” ISIS 102, no. 4 (2011): 601–33, fn84, 629

(24.) J. Finch, A Nobel Fellow on Every Floor: A History of the Medical Research Council Laboratory of Molecular Biology (Cambridge, Eng., 2008), 242de Chadarevian, “Making,” 628

(25.) J. Mcafferty, A. D. Griffiths, G. Winter, and D. J. Chiswell, “Phage Antibodies: Filamentous Phage Displaying Antibody Variable Domains,” Nature 348 (1990): 552–54; S. J. Russell, M. B. Llewelyn, and R. E. Hawkins, “The Human Antibody Library,” BMJ 304 (1992): 585–86.

(26.) H. R. Hoogenboom and G. Winter, “By-Passing Immunisation: Human Antibodies from Synthetic Repertoires of Germline VH Gene Segments Rearranged in Vitro,” J Mol Biol 227 (1992): 381–88; A. Nissim et al., “Antibody Fragments from a ‘Single Pot’ of Phage Display Library as Immunochemical Reagents,” EMBO J 13, no. 3 (1994): 692–98; J. D. Marks et al., “By-Passing Immunization: Human Antibodies from V-Gene Libraries Displayed on Phage,” J Mol Biol 222 (1991): 581–97; A. D. Griffiths et al., “Human Anti-Self (p.280) Antibodies with High Specificity from Phage Display Libraries,” EMBO J 12, no. 2 (1993): 725–34; H. R. Hoogenboom, “Designing and Optimizing Library Selection Strategies for Generating High-Affinity Antibodies,” Tibtech 15 (1997): 62–70; Russell, Llewelyn, and Hawkins, “Human.”

(27.) Interview with Neuberger; Finch, Nobel, 249–50. The first transgenic mouse was created in 1974 by Rudolf Jaenisch, a biologist based at Massachusetts Institute of Technology, to study cancer and neurological disease.

(28.) F. W. Alt, T. K. Blackwell, and G. D. Yancopopulos, “Immunoglobulin Genes in Transgenic Mice,” Trends Genetics (Aug. 1985): 231–36Lonberg, “Transgenic,” 60

(29.) Emails from Neuberger; Lonberg, “Transgenic,” 68.

(30.) Emails from Neuberger; D. T. Burke, G. F. Carle, and M. V. Olson, “Cloning of Large Segments of Exogenous DNA into Yeast by Means of Artificial Chromosome Vectors,” Science 236 (1987): 806–12; M. Goodhardt et al., “Rearrangement and Expression of Rabbit Immunoglobulin κ‎ Light Chain Gene in Transgenic Mice,” PNAS 84 (1987): 4229–33; G. Buttin, “Exogenous Ig Gene Rearrangement in Transgenic Mice: A New Strategy for Human Monoclonal Antibody Production,” Trends in Genetics 3, no. 8 (1987): 205–6; M. Brüggeman et al., “A Repertoire of Monoclonal Antibodies with Human Heavy Chains from Transgenic Mice,” PNAS 86 (1989): 6709–13.

(31.) C. Zeller, “The Expectations on Mice—Rivalry and Collaboration in an Emerging Technological Arena,” ESPACE, Economics in Space—Working Papers in Economic Geography (10 Jan. 2008): 1–31, 12.

(32.) S. L. Mansour, K. R. Thomas, and M. R. Capecchi, “Disruption of the Proto-Oncogene Int-2 in Mouse Embryo-Derived Stem Cells: A General Strategy for Targeting Mutations to Non-Selectable Genes,” Nature 336 (1988): 348–52; M. Zilstra et al., “Germ-Line Transmission of a Disrupted Beta 2-Microglobulin Gene Produced by Homologous Recombination in Embryonic Stem Cells,” Nature 342 (1989): 435–38; P. L. Schwartzberg S. P. Goff, and E. J. Robertson, “Germline Transmission of a c-abi Mutation Produced by Targeted Gene Disruption in ES Cells,” Science 246 (1989): 799–803.

(33.) Emails from Neuberger; Interview with Nils Lonberg; Schwartzberg, Goff, and Robertson, “Germline.”

(34.) A. Weintraub, “Crossing the Gene Barrier,” Business Week, 3967 (15 Jan. 2006): 72.

(35.) Ibid.; A. N. Houghton et al., “Mouse Monoclonal Antibody Detecting GD3 Ganglioside: A Phase I Trial in Patients with Malignant Melanoma,” PNAS 82 (1985): 1242–46; S. Vadhan-Raj et al., “Phase I Trial of a Mouse Monoclonal Antibody Against GD3 Ganglioside in Patients with Melanoma: Induction of Inflammatory Responses at Tumor Sites,” Journal of Clinical Oncology 6, no. 10 (1988): 1636–48.

(36.) GenPharm International v. Japan Tobacco, “Complaint for Monopolization,” case no. C96–0487 CW, Northern District Court of California, Oakland, 18, 21–22, 33; Interview with Lonberg; G. Stix, “The Mice That Warred,” Sci Am (p.281) (June 2001): 34–35; Zeller, “Expectations,” 13–14; S. D. Wagner et al., “Antibodies Generated from Human Immunoglobulin Miniloci in Transgenic Mice,” Nucleic Acids Research 22, no. 8 (1994): 1389–93; S. D. Wagner et al., “The Diversity of Antigen-Specific Monoclonal Antibodies from Transgenic Mice Bearing Human Immunoglobulin Gene Miniloci,” EJI 24, no. 11 (1994): 2672–81; N. Lonberg et al., “Antigen-Specific Human Antibodies from Mice Comprising Four Distinct Genetic Modifications,” Nature 368 (1994): 856–59; L. D. Taylor et al., “Human Immunoglobulin Transgenes Undergo Rearrangement, Somatic Mutation and Class Switching in Mice That Lack Endogenous IgM,” International Immunology 6, no. 4 (1994): 579–91; L. L. Green et al., “Antigen-Specific Human Antibodies from Mice Engineered with Human Ig Heavy and Light Chain YACs,” Nature Genetics 7 (1994): 13–21; S. L. Morrison, “Success in Specification,” Nature 368 (1994): 812–13.

(37.) Emails from Neuberger to Marks; Interview with Lonberg.

(38.) P. Chames et al., British Journal of Pharmacology 157 (2008): 220–33, 225; L. J. Holt et al., “Domain Antibodies: Proteins for Therapy,” Trends in Biotechnology 21, no. 11 (2003): 484–90.

(39.) E. S. Ward et al., “Binding Activities of a Repertoire of Single Immunoglobulin Variable Domains Secreted from Escherichia coli,” Nature 341 (1989): 544–46; R. Krelle, “Winter’s Way: Why Domantis Still Leads in Antibody Technology,” Australian Life Scientist (7 May 2004); P. Holliger and P. J. Hudson, “Engineered Antibody Fragments and the Rise of Single Domains,” Nat Biotechnol 23, no. 9 (2005): 1126–36, esp. 1127.

(40.) Holliger and Hudson, “Engineered,” 1127G. Winter, “Past, Present and Future of Antibody Therapeutics,” paper presented to EAPB and EUFEPS Workshop, Amsterdam, Apr. 2007Ward et al., “Binding;”Krelle, “Winter’s.”

(41.) M. Clark, “Antibody Humanization: A Case of the “Emperor’s New Clothes?” Immunology Today 21, no. 8 (2000): 397–402; G. Hale, “Therapeutic Antibodies—Delivering the Promise,” Advanced Drug Delivery Reviews 58 (2006): 633–39, esp. 636; Emails from Neuberger.

(42.) Interview with Lonberg.

(43.) Emails from Neuberger to Marks.

(44.) A. L. Nelson and J. M. Reichert, “Development Trends for Therapeutic Antibody Fragments,” Nat Biotechnol 27, no. 4 (2009): 331–37, esp. 332.

(45.) Hale, “Therapeutic”; K. Eichmann, Köhler’s Invention (Basel, 2005), 91; J. Van Brunt, “The Monoclonal Maze,” Signals Magazine (30 Nov. 2005); C. Zeller, “From the Gene to the Globe: Extracting Rents Based on Intellectual Property Monopolies,” Review of International Political Economy 15, no. 1 (2008): 86–115; Emails from Don Drakeman to Marks, 19 and 31 July 2012.

(46.) Zeller, “Expectations”Stix, “Mice”A. Jakobovits et al., “From XenoMouse Technology to (p.282) Panitumumab, the First Fully Human Antibody Product from Transgenic Mice,” Nat Biotechnol 25, no. 10 (2007): 1134–43.

(47.) Emails from Neuberger; Interview with Lonberg.

(48.) Nelson and Reichert, “Development,” 332.

(49.) In the GenPharm and Kirin Brewery case, the mice were created from crossbreeding a GenPharm transgenic mouse with another created by Kirin Brewery. Zeller, “Expectations.”

(50.) A. L. Nelson, E. Dhimolea, and J. M. Reichert, “Development Trends for Human Monoclonal Antibody Therapeutics,” Nat Rev Drug Discov 9 (2010): 767–74

(53.) C. Milstein, “Draft paper,” 2 June 1988 attached to “Awards to Inventors—MRC Scheme,” 26 Oct. 1989, MSTN, file F19; “Policy Statement by the Secretary of State for Education and Science: The Exploitation of Research Council Funded Inventions,” 25 Mar. 1985, MSTN, file F21; Finch, Nobel; de Chadarevian, “Making,” 623–24.

(54.) Finch, Nobel, 241; de Chadarevian, “Making,” 623–24. The principle of broadly licensing out the technology was also pursued by companies.

(56.) Emails from Neuberger; Winter, “Past,” slide 29; de Chadarevian, “Making,” 623–24. In 2006 BTG reported receiving £3.4 million in gross revenue from the MRC’s humanization IP portfolio. The company reported £4.1 million such revenue for 2007, £6.3 million for 2008, and then £5.1 million for 2009. BTG, Annual Reports (2008), (2011).

(57.) A. Fazackerley, “Royalties Coup Boosts Science,” Times Higher Education (28 Oct. 2005); House of Commons Science and Technology Committee, Research Council for Knowledge Transfer, Third Report of Session 2005–06, vol. 2 (15 Mar. 2006), 87; MRC News, “Sir Gregory Winter Wins Prestigious Prince of Asturias Award,” 1 June 2012, available online at http://www.mrc.ac.uk/news-events/news/sir-gregory-winter-wins-prestigious-prince-ofasturias-award (accessed 2 Oct. 2014).

(58.) De Chadarevian, “Making,” fig. 4, 630LMB, “New Building,” available online at http://www2.mrc-lmb.cam.ac.uk/about-lmb/new-building (accessed 20 Sept. 2014).