Antibodies as drugs-a Keystone Symposia report.

Therapeutic antibodies have broad indications across diverse disease states, such as oncology, autoimmune diseases, and infectious diseases. New research continues to identify antibodies with therapeutic potential as well as methods to improve upon endogenous antibodies and to design antibodies de novo. On April 27-30, 2022, experts in antibody research across academia and industry met for the Keystone symposium "Antibodies as Drugs" to present the state-of-the-art in antibody therapeutics, repertoires and deep learning, bispecific antibodies, and engineering.

Design and characterization of mouse IgG1 and IgG2a bispecific antibodies for use in syngeneic models.

The development of antibody therapeutics relies on animal models that accurately recapitulate disease biology. Syngeneic mouse models are increasingly used with new molecules to capture the biology of complex cancers and disease states, and to provide insight into the role of the immune system. The establishment of syngeneic mouse models requires the ability to generate surrogate mouse counterparts to antibodies designed for humans. In the field of bispecific antibodies, there remains a dearth of technologies available to generate native IgG-like mouse bispecific antibodies. Thus, we engineered a simple co-expression system for one-step purification of intact mouse IgG1 and IgG2a bispecific antibodies from any antibody pair. We demonstrated proof of concept with CD3/CD20 bispecific antibodies, which highlighted both the quality and efficacy of materials generated by this technology.

IL-2/CD25: A Long-Acting Fusion Protein That Promotes Immune Tolerance by Selectively Targeting the IL-2 Receptor on Regulatory T Cells.

Low-dose IL-2 represents an immunotherapy to selectively expand regulatory T cells (Tregs) to promote tolerance in patients with autoimmunity. In this article, we show that a fusion protein (FP) of mouse IL-2 and mouse IL-2Rα (CD25), joined by a noncleavable linker, has greater in vivo efficacy than rIL-2 at Treg expansion and control of autoimmunity. Biochemical and functional studies support a model in which IL-2 interacts with CD25 in the context of this FP in trans to form inactive head-to-tail dimers that slowly dissociate into an active monomer. In vitro, IL-2/CD25 has low sp. act. However, in vivo IL-2/CD25 is long lived to persistently and selectively stimulate Tregs. In female NOD mice, IL-2/CD25 administration increased Tregs within the pancreas and reduced the instance of spontaneous diabetes. Thus, IL-2/CD25 represents a distinct class of IL-2 FPs with the potential for clinical development for use in autoimmunity or other disorders of an overactive immune response.

Probing Sub-GeV Mass Strongly Interacting Dark Matter with a Low-Threshold Surface Experiment.

Using data from the ν-cleus detector, based on the surface of Earth, we place constraints on dark matter in the form of strongly interacting massive particles (SIMPs) which interact with nucleons via nuclear-scale cross sections. For large SIMP-nucleon cross sections, the sensitivity of traditional direct dark matter searches using underground experiments is limited by the energy loss experienced by SIMPs, due to scattering with the rock overburden and experimental shielding on their way to the detector apparatus. Hence, a surface-based experiment is ideal for a SIMP search, despite the much larger background resulting from the lack of shielding. We show using data from a recent surface run of a low-threshold cryogenic detector that values of the SIMP-nucleon cross section up to approximately 10^{-27} cm^{2} can be excluded for SIMPs with masses above 100 MeV.

Projections for Measuring the Size of the Solar Core with Neutrino-Electron Scattering.

We quantify the amount of data needed in order to measure the size and position of the ^{8}B neutrino production region within the solar core, for experiments looking at elastic scattering between electrons and solar neutrinos. The directions of the electrons immediately after scattering are strongly correlated with the incident directions of the neutrinos; however, this is degraded significantly by the subsequent scattering of these electrons in the detector medium. We generate distributions of such electrons for different neutrino production profiles, and use a maximum likelihood analysis to make projections for future experimental sensitivity. We find that with approximately 20 years worth of data the Super Kamiokande experiment could constrain the central radius of the shell in which ^{8}B neutrinos are produced to be less than 0.22 of the total solar radius at 95% confidence.

Glow in the dark matter: observing galactic halos with scattered light.

We consider the observation of diffuse halos of light around the discs of spiral galaxies, as a probe of the interaction cross section between dark matter (DM) and photons. Using the galaxy M101 as an example, we show that for a scattering cross section at the level of 10(-23)(m/GeV) cm(2) or greater dark matter in the halo will scatter light out from the more luminous center of the disc to larger radii, contributing to an effective increased surface brightness at the edges of the observed area on the sky. This allows us to set an upper limit on the DM-photon cross section using data from the Dragonfly instrument. We then show how to improve this constraint, and the potential for discovery, by combining the radial profile of DM-photon scattering with measurements at multiple wavelengths. Observation of diffuse light presents a new and potentially powerful way to probe the interactions of dark matter with photons, a way that is complementary to existing searches.

Fitting the annual modulation in DAMA with neutrons from muons and neutrinos.

The DAMA/LIBRA experiment searches for evidence of dark matter scattering off nuclei. Data from DAMA show 9.2 σ evidence for an annual modulation, consistent with dark matter having a cross section around 2 × 10(-40) cm(2). However, this is excluded by other direct detection experiments. We propose an alternative source of annual modulation in the form of neutrons, which have been liberated from material surrounding the detector by a combination of (8)B solar neutrinos and atmospheric muons. The phase of the muon modulation lags 30 days behind the data; however, we show that adding the modulated neutrino component shifts the phase of the combined signal forward. In addition, we estimate that neutrinos and muons need ∼ 1000 m(3) of scattering material in order to generate enough neutrons to constitute the signal. With current data, our model gives as good a fit as dark matter, and we discuss prospects for future experiments to discriminate between the two.

Properties of 7ND-CCL2 are modulated upon fusion to Fc.

7ND, a truncated version of the chemokine MCP-1/CCL2 lacking amino acids 2-8, is a potent antagonist of CCR2. In contrast to CCL2, 7ND is an obligate monomer. Similar to other chemokines, the in vivo half-life of 7ND is very short and its use as an antagonist in disease models is thus limited. We therefore constructed a 7ND-Fc fusion protein to extend the half-life of 7ND and overcome its limitations as a potential therapeutic antagonist. When we tested the properties of the fusion molecule in vitro, we found to our surprise that 7ND-Fc, in contrast to 7ND, produced a distinct, albeit small, chemotactic response in THP-1 cells, and a robust chemotactic response in L1.2 cells stably transfected with CCR2. To test whether this unexpected observation might be due to the bivalency of 7ND-Fc stemming from the dimeric nature of Fc fusions, we produced a heterodimeric Fc fusion which displays only one 7ND moiety, using a technology called strand exchange of engineered CH3 domains (SEED). The monovalent construct had properties equivalent to the parent 7ND. Furthermore, partial agonist activity appears to depend on receptor density as well as the signaling pathway examined. However, we were able to show that 7ND-Fc, but not 7ND alone, has antagonistic activity in experimental autoimmune encephalomyelitis, a murine model of multiple sclerosis.

Structures of adnectin/protein complexes reveal an expanded binding footprint.

Adnectins are targeted biologics derived from the tenth type III domain of human fibronectin (¹°Fn3), a member of the immunoglobulin superfamily. Target-specific binders are selected from libraries generated by diversifying the three ¹°Fn3 loops that are analogous to the complementarity determining regions of antibodies. The crystal structures of two Adnectins were determined, each in complex with its therapeutic target, EGFR or IL-23. Both Adnectins bind different epitopes than those bound by known monoclonal antibodies. Molecular modeling suggests that some of these epitopes might not be accessible to antibodies because of the size and concave shape of the antibody combining site. In addition to interactions from the Adnectin diversified loops, residues from the N terminus and/or the ß strands interact with the target proteins in both complexes. Alanine-scanning mutagenesis confirmed the calculated binding energies of these ß strand interactions, indicating that these nonloop residues can expand the available binding footprint.

SEEDbodies: fusion proteins based on strand-exchange engineered domain (SEED) CH3 heterodimers in an Fc analogue platform for asymmetric binders or immunofusions and bispecific antibodies.

Bispecific antibodies and asymmetric Fc fusion proteins offer opportunities for important advances in therapeutics. Bivalent IgG depends upon in vivo dimerization of its heavy chains, mediated by homodimeric association of its C(H)3 domains. We have developed a heterodimeric Fc platform that supports the design of bispecific and asymmetric fusion proteins by devising strand-exchange engineered domain (SEED) C(H)3 heterodimers. These derivatives of human IgG and IgA C(H)3 domains create complementary human SEED C(H)3 heterodimers that are composed of alternating segments of human IgA and IgG C(H)3 sequences. The resulting pair of SEED C(H)3 domains preferentially associates to form heterodimers when expressed in mammalian cells. SEEDbody (Sb) fusion proteins consist of [IgG1 hinge]-C(H)2-[SEED C(H)3], that may be genetically linked to one or more fusion partners. This investigation reports on the generation of mono-Fab-Sb and Sb-IL2 monocytokine as models. They were expressed at high levels in NS/0 cells, purified on recombinant protein A resin and were well-behaved in solution. When administered intravenously to mice, Sb pharmacokinetics exhibited the long serum half-life extensions typical of comparable Fc-containing immunofusion and IgG1 controls.

Solution structure of an informationally complex high-affinity RNA aptamer to GTP.

Higher-affinity RNA aptamers to GTP are more informationally complex than lower-affinity aptamers. Analog binding studies have shown that the additional information needed to improve affinity does not specify more interactions with the ligand. In light of those observations, we would like to understand the structural characteristics that enable complex aptamers to bind their ligands with higher affinity. Here we present the solution structure of the 41-nt Class I GTP aptamer (K(d) = 75 nM) as determined by NMR. The backbone of the aptamer forms a reverse-S that shapes the binding pocket. The ligand nucleobase stacks between purine platforms and makes hydrogen bonds with the edge of another base. Interestingly, the local modes of interaction for the Class I aptamer and an RNA aptamer that binds ATP with a K(d) of 6 microM are very much alike. The aptamers exhibit nearly identical levels of binding specificity and fraction of ligand sequestered from the solvent (81%-85%). However, the GTP aptamer is more informationally complex (approximately 45 vs. 35 bits) and has a larger recognition bulge (15 vs. 12 nucleotides) with many more stabilizing base-base interactions. Because the aptamers have similar modes of ligand binding, we conclude that the stabilizing structural elements in the Class I aptamer are responsible for much of the difference in K(d). These results are consistent with the hypothesis that increasing the number of intra-RNA interactions, rather than adding specific contacts to the ligand, is the simplest way to improve binding affinity.

Informational complexity and functional activity of RNA structures.

Very little is known about the distribution of functional DNA, RNA, and protein molecules in sequence space. The question of how the number and complexity of distinct solutions to a particular biochemical problem varies with activity is an important aspect of this general problem. Here we present a comparison of the structures and activities of eleven distinct GTP-binding RNAs (aptamers). By experimentally measuring the amount of information required to specify each optimal binding structure, we show that defining a structure capable of 10-fold tighter binding requires approximately 10 additional bits of information. This increase in information content is equivalent to specifying the identity of five additional nucleotide positions and corresponds to an approximately 1000-fold decrease in abundance in a sample of random sequences. We observe a similar relationship between structural complexity and activity in a comparison of two catalytic RNAs (ribozyme ligases), raising the possibility of a general relationship between the complexity of RNA structures and their functional activity. Describing how information varies with activity in other heteropolymers, both biological and synthetic, may lead to an objective means of comparing their functional properties. This approach could be useful in predicting the functional utility of novel heteropolymers.

Isolation of high-affinity GTP aptamers from partially structured RNA libraries.

Aptamers, RNA sequences that bind to target ligands, are typically isolated by in vitro selection from RNA libraries containing completely random sequences. To see whether higher-affinity aptamers can be isolated from partially structured RNA libraries, we selected for aptamers that bind GTP, starting from a mixture of fully random and partially structured libraries. Because stem-loops are common motifs in previously characterized aptamers, we designed the partially structured library to contain a centrally located stable stem-loop. We used an off-rate selection protocol designed to maximize the enrichment of high-affinity aptamers. The selection produced a surprisingly large number of distinct sequence motifs and secondary structures, including seven different aptamers with K(d)s ranging from 500 to 25 nanomolar. The engineered stem-loop was present in the three highest affinity aptamers, and in 12 of 13 independent isolates with a single consensus sequence, suggesting that its inclusion increased the abundance of high-affinity aptamers in the starting pool.

Energetic landscape of alpha-lytic protease optimizes longevity through kinetic stability.

During the evolution of proteins the pressure to optimize biological activity is moderated by a need for efficient folding. For most proteins, this is accomplished through spontaneous folding to a thermodynamically stable and active native state. However, in the extracellular bacterial alpha-lytic protease (alphaLP) these two processes have become decoupled. The native state of alphaLP is thermodynamically unstable, and when denatured, requires millennia (t1/2 approximately 1,800 years) to refold. Folding is made possible by an attached folding catalyst, the pro-region, which is degraded on completion of folding, leaving alphaLP trapped in its native state by a large kinetic unfolding barrier (t1/2 approximately 1.2 years). alphaLP faces two very different folding landscapes: one in the presence of the pro-region controlling folding, and one in its absence restricting unfolding. Here we demonstrate that this separation of folding and unfolding pathways has removed constraints placed on the folding of thermodynamically stable proteins, and allowed the evolution of a native state having markedly reduced dynamic fluctuations. This, in turn, has led to a significant extension of the functional lifetime of alphaLP by the optimal suppression of proteolytic sensitivity.