Antibodies are part of the bodies natural defence system against virus and bacterial infections. They bind to the pathogenic antigens and flag them for destruction by complement and cells of the immune system. Recent years has seen the development of antibodies to treat cancers such as breast and bowel cancer and immune disorders such as rheumatoid arthritis. Scientists at the MRC Laboratory of Molecular Biology in Cambridge have made key contributions to the science and its translation to the clinic, and to founding of three UK biotechnology companies (Celltech, Cambridge Antibody Technology and Domantis).
Background. Antibodies were discovered more than 100 years ago and horse antiserum was used in the 1890s to treat tetanus and diphtheria, and is still used to treat snake bites. However horse antiserum is seen as foreign by the human immune system, which reacts by producing antibodies against the horse antiserum, especially on repeat doses. Not only does this neutralise the beneficial effect of the horse antibodies but provokes adverse reactions, including fever and sometimes life-threatening anaphylactic shock.
During most of the 20th C, the adverse effect of animal antibodies prompted the use of human antiserum from donors who had recovered from disease, typically for prophylaxis of respiratory and hepatitis B infections. However the development of antibodies to treat cancer and auto-immune disease proved much more difficult. Firstly it was necessary to identify and isolate the disease target, which was typically a minor component in a complex mixture of human proteins. Secondly it was difficult to raise a human immune response against the human disease targets due to immunological tolerance.
Monoclonal antibodies (mAbs). A first step towards in solving these problems came from the invention of monoclonal antibodies by George Kohler and Cesar Milstein at the LMB in 1975. By immunising rodents, and fusing the antibody-producing cells from the spleen with a myeloma, Kohler and Milstein made hybrid cells (hybridomas) that were immortal and secreted rodent monoclonal antibodies (mAbs), and these cells could be grown in large scale culture in fermenters without any further requirement for animals. The technology allowed an antibody response directed against a complex mixture of antigens (such as on the surface of a cancer cell) to be dissected into its components. It became possible to distinguish between normal and cancer cells on the basis of reactivity to individual mAbs, and thereby to identify molecular targets.
The work provided a wealth of reagents for research and for diagnostic kits, including the Clearblue pregnancy testing kit. Kohler and Milstein were awarded the Nobel Prize in Physiology or Medicine in 1984. However, the application to therapy proved more problematic: the rodent mAbs were often poor in triggering human effector functions, and also proved antigenic in humans. Furthermore it proved difficult to make human hybridomas from immunised humans (in good measure due to lack of a suitable human myeloma fusion partner), and in any case the problem of raising an immune response to human self-antigens remained.
Chimeric and humanised antibodies. These difficulties prompted the use of genetic engineering to convert rodent mAbs into human-like mAbs. The antibody molecule is comprised of two chains (heavy and light) that associate with each other. At one end of each chain are the “variable” regions that bind antigen; the heavy and light chain variable regions make extensive contacts with each other and jointly form the antigen binding site. The rest of the chain comprises the constant regions that are responsible for other functions, including the triggering of host effector functions.
The first approach (pioneered by Michael Neuberger and Terrence Rabbitts from the LMB among several groups world-wide in 1984-5) involved the transplantation of the variable domains from the rodent antibodies in place of the corresponding domains of a human antibody. These antibodies (two-thirds human) were termed human-mouse chimeric antibodies, and had the binding activity of the rodent antibody and the effector functions of human antibodies.
The second approach (pioneered by Greg Winter at the LMB in 1986) involved the transplantation of just the antigen contact surfaces (the complementarity determining regions or CDRs) from the rodent mAb; these antibodies (about 95% human) became known as humanised antibodies.
The first demonstration of the therapeutic effect of a genetically engineered antibody came in 1988 from the collaborative work of the groups of Winter at the LMB and Herman Waldmann at the Cambridge University Department of Pathology. They humanized a rat mAb against a human lymphocyte marker developed by Waldmann; when this antibody (Campath-1H) used to treat patients with non-Hodgkin’s lymphoma, the antibody destroyed a huge mass of tumour in the spleen over a period of 30 days. Furthermore the antibody was not antigenic. Subsequently a range of chimeric and humanized antibodies have been developed against human self-antigens, including those against lymphomas, breast and bowel cancer, rheumatoid arthritis, Crohn’s disease, psoriasis and asthma. More than half of all therapeutic antibodies approved by the US Food and Drug Administration (FDA) are humanised antibodies.
Human antibodies. Scientists at the LMB also played key roles in the development of technology for making completely human antibodies against human self-antigens.
One approach (from the group of Greg Winter in collaboration with scientists from Cambridge Antibody Technology; also from the group of Richard Lerner at the Scripps Research Institute in collaboration with scientists from Stratagene) can be viewed as mimicking the strategy of the immune system using bacterial viruses. The work led to the creation of huge “combinatorial” repertoires of human heavy and light chain variable regions on the surface of filamentous bacteriophage so as to display the binding activity of the encoded antibodies on the surface of the phage. Each phage carried a single antibody fragment co-packaged with its encoding genes, and “phage antibodies” against target antigens isolated by binding of the phage to immobilized antigen. From a single tube of phage, it was possible to isolate phage antibodies of many different binding activities, including against self-antigens. The antibody genes could be harvested from the phage, rebuilt into complete human antibodies and produced in large scale tissue culture. This approach completely avoided the use of animals at any stage, and led to the development by Cambridge Antibody Technology of several human antibodies, including Humira, the first human antibody to be approved for therapeutic use by the FDA.
Another approach (from the group of Michael Neuberger in collaboration with Marianna Bruggemann at the AFRC Babraham) involved the creation of transgenic mice with human antibody genes. After immunisation of the mice, the application of hybridoma technology led to creation of rodent hybridomas that secreted human antibodies. Several human antibodies have been developed by MRC licensees, including Vectibix, the second human antibody to be approved for therapeutic use by the FDA.
Single domain antibodies (dAbs). Natural antibodies are based on paired heavy and light chain variable domains, both of which were thought necessary for antigen binding. In 1989 at the LMB it was discovered that isolated antibody variable domains were sometimes capable of binding antigen alone. Not only did such single domains (dAbs) offer a means of creating smaller antibodies, allowing for example greater tissue penetration, but also new and versatile antibody formats (see below). However these single domains had poor biophysical properties, leading to stickiness and aggregation.
In 2000, work on dAbs was resumed. The groups of Greg Winter and Ian Tomlinson embarked on the use of phage display technology to isolate dAbs, and developed approaches to overcome their poor properties (including the use of evolution to make dAbs that were intrinsically aggregation-resistant). Simple dAbs do not have effector functions, but are able to block ligands or receptors; effector functions can be added by fusion to antibody constant domains (Fc) or by conjugation with cytotoxic drugs. Simple dAbs are also cleared quickly from circulation, but can be sustained there by fusion to a second dAb that binds to serum albumin, or by conjugation to polyethylene glycol.
The groups also used dAb technology to invent an entirely new type of antibody. By pairing heavy and light chain dAbs against different targets, they created antibodies with dual specificity, allowing the targeting of two disease markers with a single antibody, potentially extending its therapeutic scope. In preclinical models Domantis scientists have shown that dAbs can perform as well as classical antibodies, but dAbs have yet to enter clinical trials.
Patents, licensing, start-ups. Under the MRC’s royal charter, the MRC has several duties including supporting research to improve human health, and advancing knowledge and technology to improve the economic competitiveness of the UK. The primary purpose of the MRC’s translational strategies are to meet the duties imposed on it, and in this LMB scientists have played key roles as inventors, as industry consultants and as company directors.
It is well known that in 1975 the National Research and Development Corporation (which owned the rights to inventions arising from work in the MRC) did not take out a patent on hybridoma technology. However antibody reagents and know-how from LMB scientists helped the development of antibody programmes in Celltech, the UK’s first biotechnology company. With the benefit of hindsight the failure to file a patent on mAbs may have stimulated the field; rodent mAbs were not the promised magic bullets for cancer and a further step of genetic engineering proved necessary.
By the mid-1980s the MRC had acquired the right to take out patents on MRC inventions, and by the late 1980s had started to non-exclusively license the patents on humanized antibodies. The aim of the non-exclusive licence was to encourage the uptake of the technology, although some areas of exclusivity were reserved for Celltech. More than 40 companies were licensed, facilitated by packaging of the patent with another from Celltech. In 1989 the MRC’s Collaborative Centre for industry at Mill Hill, London initiated an antibody humanization service, leading to the therapeutic antibodies Tysabri (approved US) and Actemra (approved Japan). The patents on transgenic human antibodies were also licensed non-exclusively.
By contrast the MRC licensed the patents on the antibody repertoires exclusively (except for single domain products) to local start-up company Cambridge Antibody Technology (CAT). Inventors and associated co-workers also had a founding stake or option packages and at various stages senior LMB scientists acted as members of the Scientific Advisory Board and as non-Executive Directors. In return the MRC received funding for a programme of research, an equity stake and an agreed share of royalties. Seed funding for CAT (founded 1989) came from Peptech (Australia) and was mediated by former LMB visiting scientist Geoffrey Grigg, Peptech’s founder. A highlight of the CAT’s work was the development of Humira (with BASF), the first human antibody to be FDA-approved. In 1997, CAT was listed on the London Stock Exchange, and in 2005 was purchased by Astra-Zeneca with a valuation of about £700M.
Similarly the MRC licensed the use of antibody repertoires for single domain products, and other single domain intellectual property developed in the LMB to local start-up company Domantis. Again inventors and associated co-workers had founding stakes or option packages, and again senior LMB scientists acted as members of the Scientific Advisory Board and as company Directors. Seed funding for Domantis in 2000 came from the MRC’s associated venture fund (MVM), but in 2001, again mediated by Grigg, Peptech also took a large stake in Domantis, providing cash for the rapid development of the company at an early stage. In 2001 Ian Tomlinson left the LMB to head the research of Domantis. In late 2006, Domantis was purchased by GSK for £230M.
Although the MRC is not primarily concerned with making money from inventions, the scale of LMB’s contribution to the field of therapeutic antibodies has led to significant revenues back to the MRC. For example, from 2000 to 2006 the MRC received approximately £84M from humanized antibodies and £127M from human antibodies (share sales and royalties), of which the LMB and the inventors and associated co-workers have received a share through the MRC’s Awards to Inventors Scheme.