Why is Antibody Humanization still important in the 21st century?

Monoclonal antibodies, or mAbs, have become a crucial tool and are used for both diagnostic and medical applications. This article will explain the importance of mAbs and antibody humanization techniques involved.

If you are interested in Antibody Humanization services, check out our main service page here.

The technology of monoclonal antibodies was pioneered by César Milstein and Georges Köhler, who in 1975 successfully fused immortal myeloma cell lines with antibody producing B-cells to produce hybridomas. In 1986 the first monoclonal antibody therapy was approved for use in humans (Muromonab-CD3). Muromonab is a complete, unmodified mouse antibody and is one of only four fully murine therapeutic mAbs to be approved for clinical use. Murine antibodies have a short therapeutic half-life because they are recognised by the patient immune system as foreign proteins resulting in a human anti-mouse (HAMA) response.

Since 1986 there has been a variety of antibody engineering techniques employed to reduce immunogenicity and the HAMA response. A chimeric antibody is one which has mouse variable domains, but human constant domains. As chimeric antibodies are about 70% human, they are not as readily cleared by the patient’s immune system. There have been seven chimeric antibodies approved, the most recent being Novartis’ Cosentyx® (Secukinumab) which is used to treat moderate to severe plaque psoriasis (raised, silvery flaking of the skin) in adults.

However, despite some specific use cases, chimeric antibodies are generally still too immunogenic to be used as therapeutic antibodies and further modification of the antibody sequence is required to reduce patient immune response. Humanization is a process by which xenogeneic antibody sequences are modified to reduce this immunogenicity and several approaches have been developed since the first approved humanized antibody in 1997 (Daclizumab). Since the 2002 approval of Adalimumab generated by phage display technology and the 2006 approval of Panitumumab there have also been two technologies capable of producing fully human antibodies.

In the context of this article we will discuss these three techniques; CDR grafting, phage display and approaches using transgenic mice to attempt to answer the question:

Why do we still need to humanize murine antibodies in the 21st Century?

Popular Humanization techniques

CDR Grafting

The complementarity determining regions (CDRs) are the hypervariable ‘ends’ of an antibody which are responsible for where antibodies bind to a specific antigen. CDR grafting is a humanization technique whereby humanized antibody sequences are generated by carefully selecting the CDRs of the parental antibody (typically murine, but increasingly other species are being humanized, including rabbits) and grafting them into a human framework.

Some of the first humanization strategies used a limited subset of well-characterised human mAbs and did not consider sequence similarity to the parental murine antibody (fixed frameworks approach). Modern approaches now use human variable regions with high amino acid similarity to the murine variable regions (homology matching or best-fit).

Although the molecular biology processes involved in CDR grafting is relatively straight forward, simply cutting and pasting CDR sequences from murine antibodies to a human backbone is not always sufficient to retain the binding strength and specificity of the parental antibody. Design is critical and in recent years has become an art form practiced by a few key individuals around the world. Design includes various choices such as the boundaries of the CDRs, which human frameworks to use and which residues (if any) to substitute from the murine mAb into the human framework regions (back mutations). CDR identification is critical to the humanization process and for the same antibody sequence different numbering systems may differently define the CDR boundaries. In extreme cases there could be as much as a 10 amino acid difference between two CDR definitions. Companies like Fusion Antibodies, have optimized this process and use a combination of CDR definitions to maintain only residues critical to binding.

Humanized antibodies developed by CDR grafting techniques cannot be classed as “human” in origin because they are derived from a combination of several antibodies (Murine antibody CDRs with human constant domains). In fact, it is theoretically possible for an antibody derived from humanization technologies to have the exact same sequence as a “human” antibody from phage display or transgenic mice but due to their origin they cannot be classed as “human”.

Phage display technologies

Originally described by George P. Smith in 1985, generating antibodies by phage display is based on the process of genetically engineering bacteriophage and repeated rounds of antigen-guided selection. In 2002 Adalimumab (HUMIRA®) was the first therapeutic antibody generated by phage display technology to be approved for therapeutic. It is also regarded as the world’s first fully human therapeutic antibody.

A second phage display generated therapeutic antibody (Belimumab, GlaxoSmithKline) was approved in 2011, but with the exception of the pending anti-EGFR therapeutic from Lilly Oncology (in phase three trials at the time of writing) these have been the only two successfully approved therapeutics coming from a phage display approach.

A human phage display library is constructed by first isolating antibody RNA from a given source (e.g. sequencing from human peripheral blood mononuclear cells), followed by ligation into a phage display vector. These vectors can then be used for expression of human IgG on bacteriophage hosts to represent the entire immune repertoire from which the RNA was isolated. One can then screen (or “pan”) a phage library for those which bind to a particular antigen and isolate the original IgG sequence.

Transgenic mice

In 1994 two papers described the production of genetically engineered mice which were capable of expression of full human antibody repertoires, since then the field has taken off. The first human antibody from a transgenic source was approved in 2006 (Panitumumab) and since 2009 there have been a further seven including the 2014 approval of Nivolumab for treatment of melanoma and squamous cell carcinoma. Transgenic mice are generated by targeted modification of the endogenous mouse antibody genes to suppress their expression combined with introduction of human antibody heavy and light chain gene sequences. The result is a mouse which expresses fully human antibodies.

Once of the major advantages of the transgenic mouse approach is that antibodies are generated by the same techniques which were developed in 1975 by César Milstein and Georges Köhler. This means that techniques are well established, optimized and understood, which facilitates an easy path to clinical trials and market approval. This has likely contributed to the rapid growth in the number of antibodies with 10 out of 32 antibodies approved since 2002 being from either the Medarex, Abgenix or Regeneron transgenic platforms.

Discussion and comparison of techniques

So why is CDR grafting so popular, what are the issues with phage display technologies and despite their potential why have we not seen as many antibodies from transgenic mice as we might have expected?

Since 1985 we have seen four murine antibodies approved for clinical use with long gaps between cases. In the 90’s and early 2000’s we seen a rise of chimeric antibody approval but this was quickly superseded by development of humanized antibodies from 1997 onwards, the first phage display antibodies from 2002 and transgenic antibodies from 2006. Despite the development of new technologies, in the past five years we have seen almost equal number of “human” and “humanized” antibodies. With technologies available to produce human antibodies, why are we not seeing more?

Although the numbers of human antibodies from transgenic mice have increased rapidly the number of platforms has remained limited and exclusive. All eight approved antibodies are from three platforms; Abgenix (purchased by Amgen in 2005 for 2.2$B), Medarex (purchased by Bristol Myers Squibb in 2009 for 2.4$B) or the more recent VelociMouse® (Alirocumab) developed by Sanofi/Regeneron. These large pharma partnerships and acquisitions leave transgenic mouse platforms far out of reach of the average small-mid-sized biotech and even further from the likes of academics and virtual companies. Even before going behind closed doors and becoming Bristol Myers Squibb or Amgen technologies, generating antibodies with transgenic mice was not only incredibly expensive but also came loaded with significant royalties.

Furthermore, there has been some debate over the limited germline repertoire which was engineered into the mouse immune system and therefore the mouse’s ability to produce a diverse range of human antibodies. Additionally, there has been some concern that the human antibodies from transgenic mice are essentially hybrids of mouse and human components (e.g. human immunoglobulin sequences and mouse signalling molecules) and although they develop into what appear to be “normal” human antibodies some academics initially expressed concern. Further development of the transgenic mouse platforms has reduced these concerns in recent years.

So what about human antibodies from phage display technologies? The main difficulty with antibodies derived from phage display is expression. Proteins such as antibodies, or antibody fragments, that are derived from eukaryotic organisms are often difficult to express within a prokaryotic cell. The “unnatural” human sequences which result from phage display have proven difficult to develop and express in sufficient quantities and appear to be the main reason for only two antibodies being currently available on the market.

CDR grafting was the original process developed by Greg Winter in 1986 and still remains one of the most popular techniques for the production of therapeutic antibodies. It has been stated that humanized antibodies do suffer from one disadvantage when compared to human antibodies in that they are slightly more immunogenic. However, when the alternatives are considered (difficult to express phage display antibodies, or inaccessible/expensive transgenic mice technologies) the small amount of potential additional immunogenicity presented by humanization technologies has proved to be an acceptable compromise. Additionally, modernised versions of the CDR grafting technique now include T-cell epitope avoidance technologies which further reduces the potential immunogenicity of humanized antibodies. We have also seen a growth in companies using powerful technologies to screen the full B cell repertoire of a wide range of host species of including Rabbit, Avian and Llama to select the optimum mAbs with the correct characteristics. By this evidence we will see a lot more specific antibodies being developed to a wider range of precision targets in unmet medical needs going forward whereby transgenic animals and CDR grafting humanization will lead the approvals for the next 5 years.

–    Richard Buick, Chief Technical Officer, Fusion Antibodies Ltd. &  Andrew Glover, Business Development Executive, Fusion Antibodies

This article is featured in the January 2016 Edition of European Biopharmaceutical Review.