August 1, 2018 (Vol. 38, No. 14)
Before Launching an Immune Intervention, Get the Lay of the Land
Epitope mapping, which identifies antigen regions that serve as binding sites for antibodies, is starting to inform the discovery and development of new therapeutics, vaccines, and diagnostics for a myriad of afflictions including cancer, infectious diseases, and allergic diseases.
But it is still an emerging field and not without its technological and biological challenges. Industry experts regularly meet at annual scientific conferences to discuss epitope mapping and how its translational potential may be realized. For example they consider how mutations accrue in dividing tumor cells and result in modified peptide sequences, tumor-specific neoepitopes that can be the unique tumor-specific antigens by which the adaptive immune system can selectively target cancer cells.
Historically, epitope mapping experiments were very labor intensive, instrumentation lacked the needed sensitivity, and material requirements were high. As data patterns emerged, predictive algorithms became the principle way to identify new epitopes. Although good at predicting the biophysics of the peptide–receptor interaction, these algorithms do a poor job of predicting cellular processing and regulation.
From Prediction to Observation
“If you cannot predict what will be presented, the biophysics is irrelevant. A more effective approach is to capture the major histocompatibility complex receptors, elute the peptides, and measure what has actually been presented. Current improvements in mass spectrometers, mass accuracy, and bioinformatics have made looking directly into patient samples practical. This shift toward observation is also the challenge,” stated Eustache Paramithiotis, Ph.D., vice president, discovery, Caprion Biosciences.
Quality of mass spectrometry data varies widely; strong expertise in informatics is also required to define acceptable data. Overall, epitope mapping is a multidisciplinary activity that requires commitment and infrastructure. Caprion Biosciences has developed an industrialized, robust, reproducible process, and expertise in numerous sample types.
In the past, neoepitopes were assumed to be point mutations. Yet Caprion Biosciences’ data consistently demonstrate multiple mutation types presented on the tumor cells: inherited and somatic point mutations at a rate of about 25%, frame shifts at a similar proportion, and other splicing errors and fusion peptides.
Afterward, identification targets must be prioritized. The company uses a normal tissue atlas screen to ascertain if the epitope is being presented elsewhere to help determine the possibility of off-target effects.
Approximately 70–80% of peptides are already detected, so only incremental improvements are expected as instruments evolve. The big advance will occur once current clinical trials data are synthesized. The first generation of cancer immunotherapies was generated primarily through predictive algorithms. The next generation will have more direct evidence along with a greater understanding of the synergies that are required for a complete immune response to occur.
“Once we sort this out in oncology, the same approach can be applied to other areas in which the immune system is involved, including diseases that do receive as much investment. Cancer may enable the better use of the host immune system across the board. It is a golden age for an immunologist,” added Dr. Paramithiotis.
Bispecific Architectures
Combining binding specificities toward two distinct epitopes into a single molecule can greatly enhance the immunotherapeutic properties of monoclonal antibodies (mAbs). Most cancer targets currently pursued by antibody therapy developers are tumor-associated, not tumor-specific, antigens.
According to Changshou Gao, Ph.D., senior director and fellow, Yariv Mazor, Ph.D., senior scientist, and Nazzareno Dimasi, Ph.D., associate director, MedImmune, binding epitopes and affinities are two critical parameters determining antibody functions. Different bispecific architectures allow maximization of the additive functions of bispecific antibodies (bsAbs) while maintaining the potency of individual antibodies; no one bispecific architecture format fits all. The mechanism of action, capability to concurrently engage two different targets or two different epitopes on the same target, and the synergistic potential to improve efficacy should drive architecture choice.
Manipulation of affinity, avidity, valency, and architecture of bispecific therapeutics will provide opportunities that not only target double-antigen positive tumor cells over single-antigen positive normal tissues, but also selectively engage the highly upregulated antigens in tumor cells. This could lead to improved target selectivity, restricted escape mechanisms, reduced normal tissue toxicity, and improved therapeutic index.
The interplay of factors influencing target selectivity is not well understood and often overlooked. MedImmune recently showed in vivo that dual targeting alone was not sufficient to endow selective tumor targeting abilities, and reported the pivotal roles played by the affinity of the individual arms, overall avidity, and format valence in the capacity of a bsAb to promote selective tumor targeting.
In the study, the collective role of affinity, avidity, and format valence in the capacity of monovalent and bivalent bsAbs targeting the clinically validated EGFR and HER2 receptors was systemically interrogated to promote selective tumor targeting under physiological conditions.
The study used a dual-flank tumor xenograft mouse model system that carried a double-positive target tumor that was positive for EGFR and HER2 antigens on one flank and an isogenic HER2-knockout, single-positive tumor as normal tissue on the opposite flank. This system allowed the study to demonstrate that tumor-targeting selectivity is clearly influenced by the intrinsic affinity of the individual binding arms.
The affinity modulation of monovalent bsAb’s individual arms could significantly limit normal tissue targeting without impairing the bsAb’s potency against the targeted tumor. An increase in binding site valence has a detrimental effect on the ability of bsAbs to endow selective tumor targeting abilities.
These findings provide new opportunities for the use of affinity-modulated monovalent bsAbs in cancer therapy, as they facilitate the recruitment of a broader set of target antigens, previously considered unattractive for therapeutic applications due to comparable densities to those found on normal healthy cells.
Viral Immunotherapy
The bispecific approach is also being used to tackle some of the inherent challenges in viral immunotherapy. mAbs have a good safety profile and are highly specific. In addition, a protein or another molecule that is part of the pathogen, which is present only during an active infection, is typically targeted, thus lowering the risk of off-target effects.
Development and production costs are current barriers. If a cold chain and intravenous injection are required, delivery can be an issue, too—especially for types of viral infections that tend to occur in resource-limited settings. These settings can also impact return on investment. Historically, the government has sponsored development projects with smaller biotech firms, government laboratories, or academic laboratories.
Viral immunotherapy is not as advanced as cancer immunotherapy. An emerging theme is targeting multiple epitopes simultaneously to provide more arms of attack and to mitigate against the risk of viral escape. mAbs can be mixed into a cocktail; alternatively, multiple variable domains can be engineered into a bsAb.
“We work on Ebola and want to increase efficacy with a single therapeutic agent. Bispecific antibody engineering allows multiepitope targeting. Many viral antibodies are limited in breadth to one or two members of a viral family; one goal is to provide broad protection against an entire families of viruses,” explained Jonathan R. Lai, Ph.D., professor of biochemistry, Albert Einstein College of Medicine.
“The concept of targeting the host and the virus is intriguing. All viruses tend to need some type of host receptor to activate the viral machinery that allows the virus to gain cellular entry. This can often be the virus’ Achilles’ heel. If the interaction between the host and the virus is interrupted at that particular site, then you can have a broadly active therapeutic. Some of these host–receptor interactions occur only in the intercellular compartment, so methods are being explored to deliver bsAbs to these sites. We are starting to extend our approaches for Ebola to other pathogens such as Chikungunya.”
Work in smaller animals has demonstrated a therapeutic effect, and future experiments will be extended to a nonhuman primate model.
Every virus is different, with its own unique attributes, challenges, and points of susceptibility. Effectiveness of an agent depends on the type of virus and the disease pathology. Dr. Lai believes that the general utility of the bispecific approach will continue to be demonstrated in viral immunotherapy and advance as more researchers get involved.
Allergic Diseases
Mapping epitopes quickly and accurately is challenging; they tend to be nonlinear on antigens. In the allergy field, peptide arrays have been applied using overlapping linear peptides derived from the primary sequence of the antigen. Phage-displayed peptide libraries provide an alternative, but they generate mimotopes that are difficult to consistently map back to the antibody-specific antigen. Another technique is to mutagenize antigens, determine if the antibody still binds, and then perform a structural overlay characterization.
“Multivalent interactions are the most efficient at driving IgE receptor signaling pathways. Higher valency not only lowers the threshold for the number of antigen-specific IgE molecules that must be on the surface of a mast cell or basophil, it also lowers the threshold required for the allergen dose. A highly multivalent allergen is a very efficient aggregation mechanism for crosslinking the IgE receptor, FcεRI, and that sends a very strong signal,” discussed Bridget S. Wilson, Ph.D., professor of pathology and director at the New Mexico Center for Spatiotemporal Modeling of Cell Signaling, University of New Mexico School of Medicine.
A multivalent situation is automatic if the epitope is repeated on the surface, or if the antigen is dimeric or oligomeric. In reality, humans have a repertoire of antibodies, and very complicated immune complexes can result from allergen-mediated crosslinking. An allergic individual may have multiple IgE antibodies that recognize the same or different epitopes on the allergen. No tests exist to easily map epitope recognition sites or measure the relative abundance of allergen-specific IgE species in a single patient.
Dr. Wilson’s lab has demonstrated that as few as several hundred to 2000 IgE-bound receptors that are specific for a certain allergen need to be engaged for a measurable mast cell or basophil response. Exact thresholds vary from patient to patient, depending on the individual’s IgE repertoire and the allergen’s valency.
The New Mexico Center for Spatiotemporal Modeling of Cell Signaling, a National Center for Systems Biology supported by the National Institute of General Medical Sciences, takes a multidisciplinary approach to investigate receptors and their signaling partners during signaling transduction. Complex stability and lifetime and the active recruitment of signaling partners are studied at the Center. Valency is related to a complex’ lifetime; residence time in the interaction is dependent on the number of bonds. This applies to signaling partners as well.
One of the Center’s key investigators, Diane Lidke, Ph.D., has developed innovative methods to record the rapid binding and release of individual signaling molecules in real time. These rich datasets are fed into mathematical models for in silico predictions about the effects of certain perturbations on signal transduction. Understanding the valency and reactivity of allergens can lead to the design of recombinant engineered allergens that may be safer during immunotherapeutic tolerization.
At the University of New Mexico, Bridget S. Wilson, Ph.D., an authority on cell signaling, leads a scientific team that brings state-of-the-art imaging, refined quantitative measures, and computational approaches to the study of authentic allergens and how they crosslink and activate human receptors.
