Toward enhancement of antibody thermostability and affinity by computational design in the absence of antigen.

Over the past two decades, therapeutic antibodies have emerged as a rapidly expanding domain within the field of biologics. In silico tools that can streamline the process of antibody discovery and optimization are critical to support a pipeline that is growing more numerous and complex every year. High-quality structural information remains critical for the antibody optimization process, but antibody-antigen complex structures are often unavailable and in silico antibody docking methods are still unreliable. In this study, DeepAb, a deep learning model for predicting antibody Fv structure directly from sequence, was used in conjunction with single-point experimental deep mutational scanning (DMS) enrichment data to design 200 potentially optimized variants of an anti-hen egg lysozyme (HEL) antibody. We sought to determine whether DeepAb-designed variants containing combinations of beneficial mutations from the DMS exhibit enhanced thermostability and whether this optimization affected their developability profile. The 200 variants were produced through a robust high-throughput method and tested for thermal and colloidal stability (Tonset, Tm, Tagg), affinity (KD) relative to the parental antibody, and for developability parameters (nonspecific binding, aggregation propensity, self-association). Of the designed clones, 91% and 94% exhibited increased thermal and colloidal stability and affinity, respectively. Of these, 10% showed a significantly increased affinity for HEL (5- to 21-fold increase) and thermostability (>2.5C increase in Tm1), with most clones retaining the favorable developability profile of the parental antibody. Additional in silico tests suggest that these methods would enrich for binding affinity even without first collecting experimental DMS measurements. These data open the possibility of in silico antibody optimization without the need to predict the antibody-antigen interface, which is notoriously difficult in the absence of crystal structures.

Developability profiling of a panel of Fc engineered SARS-CoV-2 neutralizing antibodies.

To combat the COVID-19 pandemic, potential therapies have been developed and moved into clinical trials at an unprecedented pace. Some of the most promising therapies are neutralizing antibodies against SARS-CoV-2. In order to maximize the therapeutic effectiveness of such neutralizing antibodies, Fc engineering to modulate effector functions and to extend half-life is desirable. However, it is critical that Fc engineering does not negatively impact the developability properties of the antibodies, as these properties play a key role in ensuring rapid development, successful manufacturing, and improved overall chances of clinical success. In this study, we describe the biophysical characterization of a panel of Fc engineered ("TM-YTE") SARS-CoV-2 neutralizing antibodies, the same Fc modifications as those found in AstraZeneca's Evusheld (AZD7442; tixagevimab and cilgavimab), in which the TM modification (L234F/L235E/P331S) reduce binding to FcγR and C1q and the YTE modification (M252Y/S254T/T256E) extends serum half-life. We have previously shown that combining both the TM and YTE Fc modifications can reduce the thermal stability of the CH2 domain and possibly lead to developability challenges. Here we show, using a diverse panel of TM-YTE SARS-CoV-2 neutralizing antibodies, that despite lowering the thermal stability of the Fc CH2 domain, the TM-YTE platform does not have any inherent developability liabilities and shows an in vivo pharmacokinetic profile in human FcRn transgenic mice similar to the well-characterized YTE platform. The TM-YTE is therefore a developable, effector function reduced, half-life extended antibody platform.

Qualification of a Biolayer Interferometry Assay to Support AZD7442 Resistance Monitoring.

AZD7442, a combination of two long-acting monoclonal antibodies (tixagevimab [AZD8895] and cilgavimab [AZD1061]), has been authorized for the prevention and treatment of coronavirus disease 2019 (COVID-19). The rapid emergence of severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) variants requires methods capable of quickly characterizing resistance to AZD7442. To support AZD7442 resistance monitoring, a biolayer interferometry (BLI) assay was developed to screen the binding of tixagevimab and cilgavimab to SARS-CoV-2 spike proteins to reduce the number of viral variants for neutralization susceptibility verification. Six spike variants were chosen to assess the assay's performance: four with decreased affinity for tixagevimab (F486S:D614G and F486W:D614G proteins) or cilgavimab (S494L:D614G and K444R:D614G proteins) and two reference proteins (wild-type HexaPro and D614G protein). Equilibrium dissociation constant (KD) values from each spike protein were used to determine shifts in binding affinity. The assay's precision, range, linearity, and limits of quantitation were established. Qualification acceptance criteria determined whether the assay was fit for purpose. By bypassing protein purification, the BLI assay provided increased screening throughput. Although limited correlation between pseudotype neutralization and BLI data (50% inhibitory concentration versus KD) was observed for full immunoglobulins (IgGs), the correlations for antibody fragments (Fabs) were stronger and reflected a better comparison of antibody binding kinetics with neutralization potency. Therefore, despite strong assay performance characteristics, the use of full IgGs limited the screening utility of the assay; however, the Fab approach warrants further exploration as a rapid, high-throughput variant-screening method for future resistance-monitoring programs. IMPORTANCE SARS-CoV-2 variants harbor multiple substitutions in their spike trimers, potentially leading to breakthrough infections and clinical resistance to immune therapies. For this reason, a BLI assay was developed and qualified to evaluate the reliability of screening SARS-CoV-2 spike trimer variants against anti-SARS-CoV-2 monoclonal antibodies (MAbs) tixagevimab and cilgavimab, the components of AZD7442, prior to in vitro pseudovirus neutralization susceptibility verification testing. The assay bypasses protein purification with rapid assessment of the binding affinity of each MAb for each recombinant protein, potentially providing an efficient preliminary selection step, thus allowing a reduced testing burden in the more technically complex viral neutralization assays. Despite precise and specific measures, an avidity effect associated with MAb binding to the trimer confounded correlation with neutralization potency, negating the assay's utility as a surrogate for neutralizing antibody potency. Improved correlation with Fabs suggests that assay optimization could overcome any avidity limitation, warranting further exploration to support future resistance-monitoring programs.

Improving the In Vivo Efficacy of an Anti-Tac (CD25) Immunotoxin by Pseudomonas Exotoxin A Domain II Engineering.

Tac (CD25) is expressed on multiple hematologic malignancies and is a target for cancer therapies. LMB-2 is an extremely active anti-Tac recombinant immunotoxin composed of an Fv that binds to Tac and a 38-kDa fragment of Pseudomonas exotoxin A (PE38). Although LMB-2 has shown high cytotoxicity toward Tac-expressing cancer cells in clinical trials, its efficacy was hampered by the formation of anti-drug antibodies against the immunogenic bacterial toxin and by dose-limiting off-target toxicity. To reduce toxin immunogenicity and nonspecific toxicity, we introduced six point mutations into domain III that were previously shown to reduce T-cell immunogenicity and deleted domain II from the toxin, leaving only the 11aa furin cleavage site, which is required for cytotoxic activity. Although this strategy has been successfully implemented for mesothelin and CD22-targeting immunotoxins, we found that removal of domain II significantly lowered the cytotoxic activity of anti-Tac immunotoxins. To restore cytotoxic activity in the absence of PE domain II, we implemented a combined rational design and screening approach to isolate highly active domain II-deleted toxin variants. The domain II-deleted variant with the highest activity contained an engineered disulfide-bridged furin cleavage site designed to mimic its native conformation within domain II. We found that this approach restored 5-fold of the cytotoxic activity and dramatically improved the MTD. Both of these improvements led to significantly increased antitumor efficacy in vivo We conclude that the next-generation anti-Tac immunotoxin is an improved candidate for targeting Tac-expressing malignancies. Mol Cancer Ther; 17(7); 1486-93. ©2018 AACR.

Unique binding modes for the broad neutralizing activity of single-chain variable fragments (scFv) targeting CD4-induced epitopes.

BACKGROUND: The CD4-induced (CD4i) epitopes in gp120 includes the co-receptor binding site, which are formed and exposed after interaction with CD4. Monoclonal antibodies (mAbs) to the CD4i epitopes exhibit limited neutralizing activity because of restricted access to their epitopes. However, small fragment counterparts such as single-chain variable fragments (scFvs) have been reported to neutralize a broad range of viruses compared with the full-size IgG molecule. To identify the CD4i epitope site responsible for this broad neutralization we constructed three scFvs of anti-CD4i mAbs from a human immunodeficiency virus type 1 (HIV-1)-infected elite controller, and investigated the neutralization coverage and precise binding site in the CD4i epitopes. RESULTS: We constructed scFvs from the anti-CD4i mAbs, 916B2, 4E9C, and 25C4b and tested their neutralization activity against a panel of 66 viruses of multi-subtype. Coverage of neutralization by the scFvs against this panel of pseudoviruses was 89% (59/66) for 4E9C, 95% (63/66) for 25C4b and 100% (66/66) for 916B2. Analysis using a series of envelope glycoprotein mutants revealed that individual anti-CD4i mAbs showed various dependencies on the hairpin 1 (H1) and V3 base. The binding profiles of 25C4b were similar to those of 17b, and 25C4b bound the region spanning multiple domains of H1 and hairpin 2 (H2) of the bridging sheet and V3 base. For 4E9C, the V3-base dependent binding was apparent from no binding to mutants containing the ΔV3 truncation. In contrast, binding of 916B2 was dependent on the H1 region, which is composed of ß2 and ß3 strands, because mutants containing the H1 truncation did not show any reactivity to 916B2. Although the H1 region structure is affected by CD4 engagement, the results indicate the unique nature of the 916B2 epitope, which may be structurally conserved before and after conformational changes of gp120. CONCLUSIONS: Identification of a unique structure of the H1 region that can be targeted by 916B2 may have an important implication in the development of small molecules to inhibit infection by a broad range of HIV-1 for the purpose of HIV treatment and prevention.

Rational design of low immunogenic anti CD25 recombinant immunotoxin for T cell malignancies by elimination of T cell epitopes in PE38.

LMB-2, is a potent recombinant immunotoxin (RIT) that is composed of scFv antibody that targets CD25 (Tac) and a toxin fragment (PE38). It is used to treat T cell leukemias and lymphomas. To make LMB-2 less immunogenic, we introduced a large deletion in domain II and six point mutations in domain III that were previously shown to reduce T cell activation in other RITs. We found that unlike other RITs, deletion of domain II from LMB-2 severely compromised its activity. Rather than deletion, we identified T cell epitopes in domain II and used alanine substitutions to identify point mutations that diminished those epitopes. The novel RIT, LMB-142 contains a 38kDa toxin and nine point mutations that diminished T cell response to the corresponding peptides by an average of 75%. LMB-142 has good cytotoxic activity and has lower nonspecific toxicity in mice. LMB-142 should be more efficient in cancer therapy because more treatment cycles can be given.

Elimination of murine and human T-cell epitopes in recombinant immunotoxin eliminates neutralizing and anti-drug antibodies in vivo.

Antibodies against the toxin portion of recombinant immunotoxins (RIT) reduce their efficacy and pose a potential safety risk. To overcome this problem we mutated the very immunogenic immunotoxin SS1P to produce LMB-T20, a de-immunized RIT that has the eight human T-cell epitopes in SS1P modified or removed. To determine the effect of T-cell epitope removal in vivo we mapped the T-cell epitopes in immune-competent BALB/c mice and found that these mice recognize two epitopes. One corresponds to the human immunodominant T-cell epitope and the other to a human subdominant epitope; both were eliminated in LMB-T20. We found that mice immunized with LMB-T20 did not have T-cell activation and did not develop anti-drug antibodies (ADA), whereas mice immunized with SS1P, showed T-cell activation, and developed ADA detected by both ELISA and drug neutralizing assays. The ability of the mice treated with LMB-T20 to respond to other antigens was not compromised. We conclude that elimination of T-cell epitopes is sufficient to prevent formation of antibodies to an immunogenic foreign protein.

Protection of the Furin Cleavage Site in Low-Toxicity Immunotoxins Based on Pseudomonas Exotoxin A.

Recombinant immunotoxins (RITs) are fusions of an Fv-based targeting moiety and a toxin. Pseudomonas exotoxin A (PE) has been used to make several immunotoxins that have been evaluated in clinical trials. Immunogenicity of the bacterial toxin and off-target toxicity have limited the efficacy of these immunotoxins. To address these issues, we have previously made RITs in which the Fv is connected to domain III (PE24) by a furin cleavage site (FCS), thereby removing unneeded sequences of domain II. However, the PE24 containing RITs do not contain the naturally occurring disulfide bond around the furin cleavage sequence, because it was removed when domain II was deleted. This could potentially allow PE24 containing immunotoxins to be cleaved and inactivated before internalization by cell surface furin or other proteases in the blood stream or tumor microenvironment. Here, we describe five new RITs in which a disulfide bond is engineered to protect the FCS. The most active of these, SS1-Fab-DS3-PE24, shows a longer serum half-life than an RIT without the disulfide bond and has the same anti-tumor activity, despite being less cytotoxic in vitro. These results have significance for the production of de-immunized, low toxicity, PE24-based immunotoxins with a longer serum half-life.

Range of CD4-Bound Conformations of HIV-1 gp120, as Defined Using Conditional CD4-Induced Antibodies.

UNLABELLED: The HIV envelope binds cellular CD4 and undergoes a range of conformational changes that lead to membrane fusion and delivery of the viral nucleocapsid into the cellular cytoplasm. This binding to CD4 reveals cryptic and highly conserved epitopes, the molecular nature of which is still not fully understood. The atomic structures of CD4 complexed with gp120 core molecules (a form of gp120 in which the V1, V2, and V3 loops and N and C termini have been truncated) have indicated that a hallmark feature of the CD4-bound conformation is the bridging sheet minidomain. Variations in the orientation of the bridging sheet hairpins have been revealed when CD4-liganded gp120 was compared to CD4-unliganded trimeric envelope structures. Hence, there appears to be a number of conformational transitions possible in HIV-1 monomeric gp120 that are affected by CD4 binding. The spectrum of CD4-bound conformations has been interrogated in this study by using a well-characterized panel of conditional, CD4-induced (CD4i) monoclonal antibodies (MAbs) that bind HIV-1 gp120 and its mutations under various conditions. Two distinct CD4i epitopes of the outer domain were studied: the first comprises the bridging sheet, while the second contains elements of the V2 loop. Furthermore, we show that the unliganded extended monomeric core of gp120 (coree) assumes an intermediate CD4i conformation in solution that further undergoes detectable rearrangements upon association with CD4. These discoveries impact both accepted paradigms concerning gp120 structure and the field of HIV immunogen design. IMPORTANCE: Elucidation of the conformational transitions that the HIV-1 envelope protein undergoes during the course of entry into CD4(+)cells is fundamental to our understanding of HIV biology. The binding of CD4 triggers a range of gp120 structural rearrangements that could present targets for future drug design and development of preventive vaccines. Here we have systematically interrogated and scrutinized these conformational transitions using a panel of antibody probes that share a specific preference for the CD4i conformations. These have been employed to study a collection of gp120 mutations and truncations. Through these analyses, we propose 4 distinct sequential steps in CD4i transitions of gp120 conformations, each defined by antibody specificities and structural requirements of the HIV envelope monomer. As a result, we not only provide new insights into this dynamic process but also define probes to further investigate HIV infection.

Allosteric induction of the CD4-bound conformation of HIV-1 Gp120.

BACKGROUND: HIV-1 infection of target cells is mediated via the binding of the viral envelope protein, gp120, to the cell surface receptor CD4. This interaction leads to conformational rearrangements in gp120 forming or revealing CD4 induced (CD4i) epitopes which are critical for the subsequent recognition of the co-receptor required for viral entry. The CD4-bound state of gp120 has been considered a potential immunogen for HIV-1 vaccine development. Here we report on an alternative means to induce gp120 into the CD4i conformation. RESULTS: Combinatorial phage display peptide libraries were screened against HIV-1 gp120 and short (14aa) peptides were selected that bind the viral envelope and allosterically induce the CD4i conformation. The lead peptide was subsequently systematically optimized for higher affinity as well as more efficient inductive activity. The peptide:gp120 complex was scrutinized with a panel of neutralizing anti-gp120 monoclonal antibodies and CD4 itself, illustrating that peptide binding does not interfere with or obscure the CD4 binding site. CONCLUSIONS: Two surfaces of gp120 are considered targets for the development of cross neutralizing antibodies against HIV-1; the CD4 binding site and CD4i epitopes. By implementing novel peptides that allosterically induce the CD4i epitopes we have generated a viral envelope that presents both of these surfaces simultaneously.

Deep Panning: steps towards probing the IgOme.

BACKGROUND: Polyclonal serum consists of vast collections of antibodies, products of differentiated B-cells. The spectrum of antibody specificities is dynamic and varies with age, physiology, and exposure to pathological insults. The complete repertoire of antibody specificities in blood, the IgOme, is therefore an extraordinarily rich source of information-a molecular record of previous encounters as well as a status report of current immune activity. The ability to profile antibody specificities of polyclonal serum at exceptionally high resolution has been an important and serious challenge which can now be overcome. METHODOLOGY/PRINCIPAL FINDINGS: Here we illustrate the application of Deep Panning, a method that combines the flexibility of combinatorial phage display of random peptides with the power of high-throughput deep sequencing. Deep Panning is first applied to evaluate the quality and diversity of naïve random peptide libraries. The production of very large data sets, hundreds of thousands of peptides, has revealed unexpected properties of combinatorial random peptide libraries and indicates correctives to ensure the quality of the libraries generated. Next, Deep Panning is used to analyze a model monoclonal antibody in addition to allowing one to follow the dynamics of biopanning and peptide selection. Finally Deep Panning is applied to profile polyclonal sera derived from HIV infected individuals. CONCLUSIONS/SIGNIFICANCE: The ability to generate and characterize hundreds of thousands of affinity-selected peptides creates an effective means towards the interrogation of the IgOme and understanding of the humoral response to disease. Deep Panning should open the door to new possibilities for serological diagnostics, vaccine design and the discovery of the correlates of immunity to emerging infectious agents.

A general insert label for peptide display on chimeric filamentous bacteriophages.

The foreign insert intended to be displayed via recombinant phage proteins can have a negative effect on protein expression and phage assembly. A typical example is the case of display of peptides longer than 6 amino acid residues on the major coat protein, protein VIII of the filamentous bacteriophages M13 and fd. A solution to this problem has been the use of "two-gene systems" generating chimeric phages that concomitantly express wild-type protein VIII along with recombinant protein VIII. Although the two-gene systems are much more permissive in regard to insert length and composition, some cases can still adversely affect phage assembly. Although these phages genotypically contain the desired DNA of the insert, they appear to be phenotypically wild type. To avoid false-negative results when using chimeric phages in binding studies, it is necessary to confirm that the observed lack of phage recognition is not due to faulty assembly and display of the intended insert. Here we describe a strategy for generating antibodies that specifically recognize recombinant protein VIII regardless of the nature of its foreign insert. These antibodies can be used as a general monitor of the display of recombinant protein VIII into phage particles.