To the editor:

With the advent of human immunoglobulin transgenic animals, where megabase-sized human DNA has been introduced into the mouse germline, the traditional approach to producing monoclonal antibodies has come into vogue again. The reason is that well-established technologies, developed since the introduction of B cell hybridomas in 1975, can be utilized to generate human monoclonal antibodies. Human monoclonal antibodies have for the last 20 years been the goal of every scientist interested in immunotherapy of humans, since these molecules evoke a minimal immune response, a severe problem associated with the use of mouse monoclonal antibodies in the clinic. However, since the first mouse monoclonal antibodies were introduced, a completely new approach was published in the beginning of the 1990s, namely phage display of human antibodies, where the diversity of the entire humoral immune response could be created in a small test tube. Overnight the monoclonal antibody technology was declared obsolete, since phage display of antibodies was put forward as a much faster and more robust alternative. When the arguments are put forward in favor of today's new kid on the block—the transgenic mouse—history repeats itself but in a reverse manner, i.e., now monoclonal antibodies are the fast track to the clinic1. This analysis is being pushed on the scientific community by analysts to investors of the biotech industry, probably not around when the phage antibodies were raised to the skies2. What also appears to be forgotten is that no matter how human the immune response of the transgenic mouse is, the problem of selecting a fusion partner to immortalize the B cell repertoire still remains the same. When "totally" human monoclonal antibodies are produced from human immunoglobulin, transgenic mice myelomas of mouse origin are used as fusion partners3,4. The main problem is then the glycosylation pattern the human antibody ends up with, which is far from human and contains the Galα1-3Gal residue. It is well known that human individuals contain a human anti-Galαl-3Gal IgG antibody titer of up to 100 μg/ml (ref. 5), which immediately will complex an in vivo administered human antibody derived from a transgenic mouse. This concern was published by us in 1993, based on the observation that the serum half-life of mouse monoclonal antibodies was inversely proportional to the antibody content of Galαl-3Gal residues6. It was quite clear that if an antibody contained Galαl-3Gal residues it would not survive for long in the human circulation. On the other hand, human monoclonal antibodies derived from EBV cell lines showed no evidence of the Galαl-3Gal residue and did consequently not react with the human anti-Galαl-3Gal IgG antibodies. Does then the transgenic animal approach offer any advantages over e.g. phage displayed antibodies? The major advantage is not the one put forward by biotech investors but is rather the access to the affinity maturation machinery of the mouse resulting in truly high affinities—albeit a feature also found in highly functional antibody libraries. In summary, the scientific achievement in constructing the human immunoglobulin transgenic mice is on the highest level but the commercial arguments, fueled by the biotech investors, are controversial and ill founded, where one major limiting factor is simply forgotten.