Antibody humanization technology: ‘Reshaping’ the future of antibody therapeutics

Our immune system has the ability to discriminate self from non-self substances or any material that is considered "foreign." These "foreign" substances are collectively called antigens, which trigger the immune system to produce antibodies against them. An antibody binds specifically to an antigen and then sends signals to the immune system’s cells to sequester and destroy the foreign entity. This is one of the ways in which antibodies are involved in defending the body from infection by foreign organisms and in fighting intrinsic diseases, such as cancer, that arise from the body’s own defects.

Antibodies can be isolated from the blood by a relatively simple purification process. These purified antibodies are sometimes used for crude therapy. However, this procedure produces a highly complex mixture of antibodies that bind to thousands upon thousands of different molecules. For treating specific therapeutic needs, such a mixture of antibodies does not suffice. Instead, scientists have found the need to develop individual monoclonal antibodies, antibodies sourced from a single cell, which bind specifically to the antigen of interest.

Today, monoclonal antibodies are produced routinely for both medical and research applications. In the medical world, they are used as diagnostics, to detect cancers or infection by certain bacteria or viruses; as therapeutics, to target foreign bacteria, viruses or cancerous cells; and possibly also as vaccines, to boost the body’s immune response. (The specificity of antibody binding to antigen is due to the complementarity between the structures of the binding surface of the antibody and the part of the antigen to which the antibody binds. An antibody directed against an antibody to an antigen could conceivably mimic the structure of the antigen and thus the anti-antibody could potentially serve as a vaccine in lieu of the antigen.)

Monoclonal antibodies are usually produced in mice. However, when used in human therapeutic applications, mouse monoclonals present serious obstacles. First, mouse antibodies themselves are recognized as foreign substances by the human immune system and thus they provoke an immune response termed the HAMA (Human Anti-Mouse Antibody) reaction. The HAMA reaction decreases the efficacy of the mouse monoclonal. HAMA could also induce fairly severe symptoms, such as allergic reactions, in the recipient. Furthermore, mouse antibodies are simply not as good as human antibodies in signaling the human immune system to destroy the malignant cells, bacteria or virus of interest.

There have been two major strategies that have emerged over the past few years in order to circumvent these limitations. These are: (i) producing fully human antibodies in mice or in the laboratory, and (ii) "humanizing" mouse or other non-human antibodies to make them compatible with the human immune system.

Humanization and other attempts to "engineer" protein molecules are now routinely done. This is because molecular biologists have mastered the trick of introducing foreign genes into practically any organism (there are even attempts to introduce genes into human patients to correct genetic defects). This allows the production of foreign proteins in more convenient hosts. Indeed, human proteins are now being produced in large amounts in bacteria, in "transgenic" (bearing foreign genes) goats, cattle and other animals, and even plants (in tobacco, for example). Before a gene is put into a host organism, the gene could be modified so that the protein product would have altered properties, e.g. greater stability, improved function, etc. The protein product could be made shorter by making the gene smaller, or two genes could be spliced together to produce two proteins in tandem. A single amino acid could be changed, or a whole segment of the protein could be replaced. Parts of one protein could be "grafted" or "transplanted" onto another. In other words, the structure of the protein product could be manipulated at will. Sometimes, not just a single gene is introduced into a host organism, but an entire gene locus, e.g. the family of genes that produces the antibody repertoire of an animal.

In fact, fully human antibodies can now be produced in transgenic mice. These mice are engineered to synthesize human antibodies instead of murine antibodies when they are injected with a stimulatory molecule or antigen. Such fully human antibodies can also be produced through bacterium-infecting viruses, known as bacteriophages. Human antibodies produced either way are potentially attractive since they contain no mouse or otherwise foreign protein component. The drawback with these methods, however, lies in slow production times and greater expense. Moreover, there has been no convincing demonstration that they are in any way superior to humanized antibodies.

On the other hand, the humanization of the murine antibodies is a more promising tool to overcome such limitations – by making the non-human antibody compatible with the human immune system while still retaining the same binding strength (i.e. affinity) and specificity of the original mouse monoclonal.

The first attempt toward humanizing murine monoclonals was to graft a portion of the murine antibody, known as the Fv fragment that is responsible for antigen binding and recognition, to a human antibody scaffold. This lowers the murine content of the monoclonal antibody and reduces the previously described HAMA complications. Subsequently, a number of "chimeric" (part human, part mouse) antibodies (approximately 30 percent mouse, 70 percent human) had been constructed. The first chimeric antibody that has been approved for human therapy is Rituxan (for the treatment of non-Hodgkin’s lymphoma, developed by IDEC).

An improved humanization technique (known as "CDR-grafting") was developed in which the mouse CDRs (Complementarity Determining Regions), the segments which form the antigen binding surface of the antibody, are transplanted onto a human antibody framework. This time, the resulting antibody is approximately 90 percent human. However, simple application of CDR grafting methods, although they provide a good yield of humanized antibody, often results in the loss of antigen-binding affinity. A few amino-acid residues outside the CDRs are usually also transplanted in order to maintain the antigen-binding properties.

Other methods have been developed for the humanization of non-human antibodies. For example, it is known that there are only a few amino acids in the entire antibody molecule that are critically involved in antigen binding. Consequently, scientists are zeroing in on the amino acids that are in direct contact with the antigen. This method, termed SDR-transfer, transplants not the entire CDRs but only those residues that are known or are thought to be involved in antigen binding. Since there are only a few of those residues, the mouse residues retained in the humanized antibody represent only a very small percentage of the total (often no more than three percent). Another humanization technique, termed "veneering" (literally, changing the outside appearance of the molecule), is accomplished by replacing the amino acids on the surface of the non-human molecule (the residues to which the antibodies of the human patient bind) with amino acids that are commonly found in human antibodies. A veneered antibody is thus camouflaged and may not be visible to the human immune system. This is akin to the technique used by schistosomes (worm parasites) to evade our immune system – these parasites coat themselves with our own molecules so that our antibodies do not "see" them as foreign. (Editor’s note: These two humanization techniques that were just described were developed by a Filipino scientist, Dr. Eduardo A. Padlan.)

This technology that "reshapes" an antibody structure is sometimes called the rational or knowledge-based approach. In contrast, some scientists are also trying "Nature’s way" in which amino acids in the CDRs are randomly replaced with every possible amino acid and the resulting molecules then screened for the desired antigen affinity and specificity.

Humanized antibodies produced by either method, or by combination of both, are really promising. Successfully marketed examples of this class of antibodies include Zenapax (for the treatment of kidney-transplant rejection; developed by Roche), Raptiva (for the treatment of a severe skin disease known as psoriasis; developed by Genentech and Xoma), Mylotarg (for the treatment of acute myeloid leukemia; developed by Celltech/American Home Products), Herceptin (for the treatment of breast cancer; developed by Genentech), and Xolair (for the treatment of asthma; developed by Genentech), among others.

The technique of antibody humanization has now matured into a major antibody engineering technology. Experience over the last decades has resulted in the evolution of a number of different methods to convert the technology into reproducible practice. As the interest in the antibody field grows in both biotechnology and pharmaceutical sectors, so grows the demand to develop monoclonal antibodies that are surely effective and with minimal side effects. Humanization technology, indeed, plays a part in a renaissance in the use of antibodies as versatile therapeutic compounds.

* * *

Cheryl D. Agdaca, 23, is an M.S. student major in Genetics and minor in Biochemistry at UP Los Baños (UPLB). Cerrone S. Cabanos, 22, is an M.S. student major in Molecular Biology and Biotechnology(MBB) and minor in Biochemistry at UPLB. Mary Ann Torio, 30, is an assistant professor at the Institute of Chemistry at UPLB and a Ph.D. student major in Biochemistry and minor in MBB, also at UPLB. Babak Nakhoda is a researcher at the National Biotechnology Institute of Iran and a Ph.D. student major in Genetics and minor in MBB at UPLB. This article was excerpted from reports written by the authors for MBB291, a Special Topics course on Protein Structure and Engineering conducted by Dr. Eduardo A. Padlan. The authors can be reached at [email protected].