Alternative non-antibody scaffolds for molecular recognition

Originally proposed one decade ago, the idea of engineering proteins outside the immunoglobulin family for novel binding functions has evolved as a powerful technology. Several classes of protein scaffolds proved to yield reagents with specificities and affinities in a range that was previously considered unique to antibodies. Such engineered protein scaffolds are usually obtained by designing a random library with mutagenesis focused at a loop region or at an otherwise permissible surface area and by selection of variants against a given target via phage display or related techniques. Whereas a plethora of protein scaffolds has meanwhile been proposed, only few of them were actually demonstrated to yield specificities towards different kinds of targets and to offer practical benefits such as robustness, smaller size, and ease of expression that justify their use as a true alternative to conventional antibodies or their recombinant fragments. Currently, the most promising scaffolds with broader applicability are protein A, the lipocalins, a fibronectin domain, an ankyrin consensus repeat domain, and thioredoxin. Corresponding binding proteins are not only of interest as research reagents or for separation in biotechnology but also as potential biopharmaceuticals, especially in the areas of cancer, autoimmune and infectious diseases as well as for in vivo diagnostics. The medical prospects have boosted high commercial expectations, and many of the promising scaffolds are under development by biotech start-up companies. Although some issues still have to be addressed, for example immunogenicity, effector functions, and plasma half-life in the context of therapeutic use or low-cost high-throughput selection for applications in proteomics research, it has become clear that scaffold-derived binding proteins will play an increasing role in biotechnology and medicine.

Introduction

Until recently, antibodies have almost uniquely served as a natural biomolecular scaffold for various applications in basic science and medicine. Their structurally variable antigen-binding site has evolved in the immune system of higher vertebrates to provide an effective humoral response against almost all kinds of foreign substances and invading pathogens [1]. Apart from their pivotal role in vivo, natural as well as engineered antibodies have yielded valuable reagents for virtually any type of antigen, and at least 10 000 antibodies are commercially available [2]. In addition, more than 20 antibody-based products have been approved as biopharmaceuticals, mainly for the treatment of cancer and autoimmune diseases, and several hundred drug candidates are currently under clinical development [3].

However, with increasing application in research, biotechnology, and medical therapy, it has become obvious that antibodies suffer from some fundamental disadvantages. Immunoglobulins (Igs) are rather large molecules, thus limiting tissue penetration. They are made of two different polypeptides, the light and heavy chains, which leads to unstable domain association when dealing with small Fv fragments and also requires complicated cloning steps for recombinant expression. The complex architecture of their antigen-binding site, which is formed by six hypervariable loops, three from each chain, is difficult to manipulate simultaneously if synthetic libraries are to be generated. The constant Fc region mediates immunological effector functions that are only crucial for few biopharmaceutical applications and often lead to undesired interactions. The manufacture of glycosylated full-size antibodies requires expensive eukaryotic cell culture, whose optimization and fermentation is time-consuming and faces limited capacities [4]. Finally, the complex intellectual property situation hampers the generation and production of recombinant antibodies [5]. Consequently, there is an emerging need for antibody alternatives with improved features.