Therapeutic antibodies are a key driving force in the biotechnology industry thanks to their ability to safely and effectively target a broad range of diseases, including cancer, autoimmune disorders, and infectious diseases. Global sales of therapeutic antibodies reached $38 billion in 2009, and the market continues to grow.
Transmembrane proteins, including multipass membrane proteins (MMP) like G-protein coupled receptors (GPCRs) and ion channels, are essential for cellular function and important targets for therapeutic monoclonal antibody (mAb) discovery and companion diagnostic assays. Greater than 40% of all drug targets are MMP but only a few have been successfully targeted by antibodies.
While antibodies to a relatively large number of MMP are offered for commercial sale, the performance of the majority of these antibodies is poor, and for those that work, their utility is generally restricted to research applications like Western blot. MMP mAbs useful in flow cytometry and functional assays with living cells are rare and difficult and expensive to make.
Membrane protein function is dictated by three-dimensional structure, and for an antibody to exert a physiologic effect on a cell it must bind to the structure of a protein as it exists in the intact membrane. In living cells, only the regions of membrane proteins that are extracellular are available for antibody binding. For MMP, the extracellular structure is composed of discontinuous regions, or loops, of the protein sequence that associate with each other on the surface of the cell. The complex 3D structure is membrane dependent. Without the membrane the native structure does not form.
X-ray crystallography studies have shown that the “footprint” of an antibody binding site (CDR) encompasses an area of approximately 20x30 Å. Figure 1 illustrates the relative sizes of a mouse mAb CDR and the extracellular region of the GPCR ADORA2A. The size of the antibody footprint is similar to the size of the entire extracellular region of ADORA2A (56 extracellular amino acids), and large enough to make contact with multiple extracellular loops. The total number of GPCR extracellular amino acids is frequently between 50 and 125. Thus, antibodies could interact with significant portions of the entire extracellular region of most GPCRs.
To develop high-performance antibodies to the extracellular regions of MMP it is necessary to immunize animals and select antibodies that recognize membrane-dependent protein antigen structures that are large enough to result in numerous contacts with antibody and high-affinity binding. Short linear peptide antigens do not fold into native 3D structures and do not represent significant portions of MMP extracellular structures. Purified, full-length native and recombinant MMP may contain all of the extracellular regions, but they are devoid of the membrane and are not folded into native structures useful for functional antibody development.
Virus-like particles (VLPs) and transfected cells expressing the target of interest are frequently used for MMP antibody development; however, a limitation of these approaches is the existence of other, nontarget proteins that create a needle in a haystack problem, and success using these approaches has been disappointingly low.
Due to the value of functional mAbs, the dearth of available antibodies, and the difficulties and high cost associated with making them, SDIX has established methods for development of high-performance antibodies to the extracellular regions of MMP.
At the heart of this process is a DNA immunization protocol that elicits specific antibody responses significantly higher than those elicited using VLP and cell-based protocols. DNA immunization results in the production of pure target MMP in the membranes of cells within the immunized animal, and, importantly, without introducing other foreign contaminants that could distract the immune system. The newly expressed protein is recognized as foreign, and antibodies are produced. Because the protein is expressed in vivo by mammalian cells, there is no purification or attendant protein denaturation. This is particularly critical for MMPs since most are highly fragile and easily denature. Following immunization, hybridomas are derived by fusion technology and supernatants are screened by multiplexed, high-throughput flow cytometry using cells expressing the target MMP.
To demonstrate the efficacy of this approach we undertook the development of mAbs to three MMPs (GPCRs CXCR4, ADORA2A, and CD20). An important aspect of the technology is the capacity to make significant numbers of different high-performance mAbs with diverse binding sites and potential effects. For all three targets, panels of mAbs were generated with low numbers of hybridoma fusions. Ninety-three CXCR4 mAbs were isolated from two hybridoma fusions, 15 ADORA2A mAbs were isolated from a single fusion, and 51 CD20 mAbs from three fusions.
Antibody gene sequences from SDIX mAbs revealed highly diverse antibody responses to all three MMPs. Seventy-five of the 93 CXCR4 mAbs were shown to have unique sequences, (i.e., derived from a unique VDJ recombination event), as well as 14 of the 15 ADORA2A mAbs and 34 of the 51 CD20 mAbs. In addition, the level of somatic hypermutation in these mAbs was shown to be comparable to a benchmark set of 29 therapeutic mAbs suggesting similar levels of affinity maturation.
Diversity was also demonstrated by epitope mapping with cells expressing mutant MMPs. Most currently known CXCR4 mAbs bind primarily to the second extracellular loop. CXCR4 mAbs from DNA-immunized animals, however, exhibited eight different binding patterns including previously unreported, novel specificities (Figure 2). Similarly, most existing CD20 mAbs fall into two classes (rituximab-like and ofatumumab-like). The CD20 mAbs isolated from DNA-immunized mice exhibit nine known and novel specificities. The intrinsically gentle antigen delivery process of DNA immunization may explain this surprisingly diverse epitope pattern as these novel epitopes may be destroyed with conventional protein-based immunizations.
The performance of mAbs isolated from DNA-immunized animals was compared to existing benchmark therapeutic antibodies. In flow cytometry, SDIX mAbs exhibit high titer binding and unique patterns of reactivity. All of the SDIX ADORA2A mAbs react positively in flow cytometry, and multiple CXCR4 and CD20 mAbs exhibit greater binding than existing benchmark antibodies.
Performance in functional assays including apoptosis and receptor modulation was also evaluated, and many of the mAbs functioned as well as, or better than, existing benchmark antibodies (Figure 3).
There is a great need for high-performance flow cytometry, functional, therapeutic, biomarker, and companion diagnostic antibodies with performance specifications exceedingly more rigorous than antibodies generated from peptide antigens for Western blots. This is especially true for MMP where target sites of antibody binding are composed of multiple discontinuous protein sequences that are dependent on the cell membrane for proper 3D structure. DNA immunization is key to eliciting antibodies that recognize this class of critical therapeutic and diagnostic targets and the technologies described here represent a significant step toward routine, low-cost development of mAbs to MMP.