Up to this point we have focused much of our attention on the development, structure, and applications of antibodies – after all the blog is called the “Biology of Antibodies.” In this blog post, we are going to switch perspectives and think a little bit about the antigen. An antigen is any structure that can be recognized by an antibody, however the binding of the antigen to an antigen receptor (membrane bound antibody) does not necessarily lead to the lymphocyte activation and the development of an immune response. A substance that has the ability to drive an immune response is referred to as an immunogen. An immunogen is also an antigen, but not all antigens are immunogens. As defined in an earlier blog, the specific contacts that mediate antibody-antigen are defined as paratope (antibody) and epitope (antigen). A large antigen is likely to have several epitopes, meaning that several different antibodies may have, or develop, the ability to bind to the same antigen. While all biological macromolecules, as well as many small chemical compounds, have the ability to be antigens, in this post we are going to focus on proteins as antigens and the ability to define and/or engineer good antigens.
II. What Makes a Good Antigen?
Immunogenicity is the measure of how well a substance or complex induces an immune response. Antigenicity is the specific measure of the antibody response toa specific target antigen. The potency of the antigen is defined by the composition and structures of the antigen that are exposed to the antibody, specifically the complementary determining regions that define the paratope of the antibody. Proteins larger than 10-20 kDa are generally immunogenic meaning they have the ability to develop immune responses by activating both B-cell responses and T-cell responses. Larger proteins generally have more epitopes than smaller proteins.
Epitopes have two general types: 1) linear or a continuous sequence of amino acids, or 2) conformational which is a structure comprised of discontinuous amino acids. As it turns out a vast majority, as much as 90%, of epitopes are conformational; indicating that the structure of the epitope is critical to the antibody-antigen interaction. Examination of defined epitopes has identified amino acid compositions that are commonly found in epitopes (Figure 1). In general, charged and polar amino acids are more prevalent than non-polar. This is not that surprising, since nearly all epitopes (linear or conformational) are found at the solvent exposed regions of the outer surface of proteins. In addition, the coiled regions are more common than alpha-helical or beta-strand regions. Therefore, the amino acid composition, structural presentation, and the geometric location of an epitope will define the antigenicity.
Even though proteins larger than 10-20 kDa are more likely to be immunogenic, smaller polypeptides, while not immunogenic, may be quite antigenic. In fact, it is always surprising to me the diversity of antibody responses that can be generated from a small peptide. Since a smaller polypeptide or other non-protein compounds are typically unable to elicit sufficient T-cell responses (due to the lack of T-cell epitopes; a topic for another blog), they must be linked; generally covalently conjugated to a carrier protein. The carrier protein provides the immunogenic stimuli (and will also elicit antibody responses), and through its association (conjugation) to the peptide or small compound, referred to as a hapten, will also generate antibody responses to the hapten. Therefore, we have the ability to generate substantial, high affinity, isotype-switched antibody responses to small peptides or organic compounds; not just larger complex proteins.
III. Epitope Prediction
The ability to predict linear and conformational epitopes is no simple task. The prediction of linear epitopes is generally focused on amino acid composition, hydrophilicity, and secondary structure analysis. The prediction of conformational epitopes is far more complicated and through the analysis of defined crystal structures of antibody-antigen interactions, the structural properties of epitopes are beginning to be understood.
There are a number of prediction models for both linear and conformational epitopes that are useful in predicting overall antigenicity, as well as regions that would be good targets for antibody recognition. Even though great progress has been made since the first predictive models of the late 1980’s, as well as substantially greater computational power, we are still a far cry from being able to predict all epitopes of an antigen.
IV. Antigen Design
Understanding what makes a good antigen or epitope is critical for the development of vaccines or high quality monoclonal antibodies. There are a number of considerations that are important when defining or engineering an antigen. First the homology of the protein to host proteins must be considered, as the host will be tolerant of its own proteins and therefore any homologous proteins from another species would likely not be antigenic or immunogenic. If there is substantial homology, the immune system can sometimes be tricked by conjugating to a carrier which will “present” the antigen uniquely to the host and, therefore, allow for the development of antibody responses. If that strategy is not successful, then identifying specific antigenic regions; as described above; and utilizing peptides conjugated to carrier proteins may be an alternative approach.
If homology is not an issue than it likely can be directly immunized into the host with the expectation that specific antibodies will be developed. However, in addition to homology biological function must also be considered. Many mammalian proteins function across species, therefore the possibility that an immunogen could have unintended biological consequences would need to be addressed, typically through conjugation to a carrier to prevent or minimize the native biological activity. Toxins, either bacterial or chemical, can also be problematic and need to be addressed through conjugation or inactivation, if possible.
Through the examination of protein structures and evaluating paratope-epitope interactions, the nature of antibody-antigen binding has become better understood, specifically the amino acid compositions and surface availability of protein regions that characterize eptiopes. The ability to design peptides as immunogens that generate high affinity antibodies can be approached through structural evaluation of the antigen. As a majority of epitopes are conformational or have some conformational element to them, the identification of regions of the protein that could be mimicked structurally by a peptide will allow for cross-reactivity with the native protein target. For, example, the identification of a loop or coiled structure at the surface of a protein that has an antigenic amino acid composition (charged or polar), will often be structurally similar between the peptide and the protein. Therefore, using the peptide as immunogen will likely yield an antibody that also recognizes the protein with high affinity (Figure 2). As the peptide target is usually not antigenic, due to the small size and likely lacking T-cell epitopes, they nearly always need to be conjugated to a carrier protein to be immunogenic. Using the peptide-carrier as immunogen has the benefit of reducing the concerns with homology and biological activity, however the one concern with conjugated peptides as immunogens is that sometimes the anti-peptide response becomes very specific for the peptide and will not cross-react (or bind) to the intended native protein target. This can, at times, be mitigated by using a combinations of peptide and protein (native or recombinant) immunizations during an antibody development project.
Depending on the availability, source, and nature of a given antigen, several approaches might be evaluated in the design of antigen, as well as strategy to develop high quality monoclonal antibodies. Often, upfront consultation will be of benefit to design an appropriate approach for the development of monoclonal antibodies.