Isogenica Ltd.

Next-generation sequencing (NGS) has revolutionized antibody discovery by enabling high-depth analysis of diverse libraries, surpassing the limitations of traditional methods like Sanger sequencing. This article explores NGS’s role in validating antibody libraries, enhancing screening processes, and integrating with artificial intelligence (AI) and machine learning (ML) to predict and engineer superior binders. Drawing from examples such as Isogenica’s use of NGS for codon-level validation and rare binder identification, we highlight how NGS measures enrichment across selection rounds, identifies crossover binders, and removes background noise. Comparisons of NGS platforms—Illumina for high-accuracy short reads, Oxford Nanopore for long reads, and PacBio HiFi for high-fidelity long reads—underscore their applications in antibody sequencing. Challenges include bioinformatics bottlenecks and the need for larger datasets to fuel AI models. Ultimately, combining NGS with technologies like CIS Display™ positions antibody discovery for faster, more precise outcomes, empowering biotechnology to address complex therapeutic targets. Next-generation sequencing (NGS) has become a cornerstone in modern biotechnology, providing unprecedented insights into genetic and proteomic diversity. In antibody discovery, NGS plays a critical role in validating and improving antibody libraries, which are essential for identifying high-affinity binders. Standard practices involve using high-depth sequencing to assess library quality, ensuring diversity and functionality before screening. Beyond validation, NGS facilitates the discovery process by analyzing enriched populations post-selection, revealing sequences that bind to targets with high specificity or more rare binders not identified during standard screening. Moreover, NGS is instrumental in linking sequence data with functional outcomes, a key challenge in the field. This integration is particularly valuable for feeding machine learning (ML) algorithms, where vast datasets of sequences can train models to predict binding affinities, epitopes, and other properties. However, one of the biggest hurdles remains correlating sequence variations with functional data, such as binding strength or stability, to enable predictive modeling. Traditional methods like Sanger sequencing have long been used in antibody discovery but come with significant limitations. Sanger sequencing offers low throughput, typically processing only a few hundred sequences per run, making it inadequate for analyzing large, diverse antibody libraries. The high cost associated with scaling up Sanger for comprehensive library coverage further restricts its utility. Additionally, Sanger lacks the resolution needed for quality control (QC) of highly diverse libraries, often missing rare variants or subtle biases in codon usage. These constraints have driven the shift toward more advanced technologies capable of handling the complexity of modern antibody repertoires. NGS is pivotal for library quality validation, as demonstrated by Frigotto et al. (2015) from Isogenica, who used NGS to perform codon-level validation of synthetic antibody fragment libraries before screening. Their approach involved automated hexamer codon additions to build precise libraries, validated through NGS to ensure diversity and accuracy. This pre-screening step helps identify any biases or errors in library construction. Isogenica employs NGS to cut through noise in challenging targets, such as cell-surface proteins or nanodisc-embedded membrane proteins, where conventional screens may fail. It excels at identifying rare binders that traditional methods overlook, enhancing the discovery of novel antibodies and keeping the funnel of potential valuable molecules wider for longer. Figure 1: Tracking sequence abundance between rounds highlights which binders are enriched under selective pressure. This data-driven view of enrichment informs refinement of panning conditions and accelerates the identification of more relevant binders. In practice, NGS measures enrichment between selection rounds to evaluate and refine panning conditions, ensuring optimal binder selection. It also identifies crossover binders from different panning arms or selection streams, providing insights into shared epitopes. Furthermore, NGS corrects incomplete deselection by removing background noise from deselection streams, improving the signal-to-noise ratio in datasets and enabling informed inclusion of families of binders. Figure 2: NGS enables direct comparison between different selection arms, revealing crossover sequences that are common across multiple streams. These insights help validate selection strategies and identify highly robust candidates that traditional methods might miss. Emerging trends in antibody discovery integrate NGS with AI and ML for advanced applications. For instance, Charlotte Deane’s group at the University of Oxford, UK, has made numerous databases and algorithms available to drive forward AI for antibody discovery in academia and the private sector alike. AI-informed affinity maturation, as explored by specialist providers such as Nabla Bio, uses computational models to optimize antibody binding. Similarly, several tools are available to predict binding-relevant antibody regions, such as paratopes and epitopes, with implications for intellectual property in therapeutic design. Companies like Sortera are expanding structure/function datasets through their Deep Screening technology, generating millions of sequence-function pairs to train AI models for biologics discovery. This addresses data scarcity, enabling predictive engineering of antibodies with desired properties. CIS Display™, a cell-free in vitro display technology, complements this by producing large datasets for AI-driven prediction and engineering. By displaying ultra-diverse libraries (up to 1014), it facilitates high-throughput screening and NGS analysis, powering ML algorithms to forecast binder performance. With this system, discrete changes in conditions and inputs can be accurately associated with both positive and negative date to support better algorithm training and predictions at the sequence level. Figure 3: NGS analysis groups the top 100 enriched sequences into distinct clusters, illustrating diversity within the most enriched binders. This clustering provides a foundation for AI-driven modelling by linking sequence families to potential functional similarities. Different NGS platforms offer unique advantages for antibody sequencing. Illumina provides short reads with high accuracy, ideal for VHH domains under 350 bp. This makes it suitable for high-throughput, error-sensitive applications. Oxford Nanopore excels in long reads, portability, and real-time sequencing, enabling analysis of full-length antibody genes without assembly artifacts, though with higher error rates compared to Illumina. PacBio HiFi combines high-fidelity with long reads, offering superior accuracy for complex repertoires, making it valuable for resolving structural variations in antibody libraries. Despite its advantages, NGS in antibody discovery faces bioinformatics bottlenecks in interpreting vast datasets, requiring advanced tools for error correction and diversity analysis. Integrating NGS with AI-driven engineering poses challenges in predicting multiple properties like affinity, specificity, and manufacturability simultaneously, not to mention synthesizing and testing large numbers of antibody sequences for functional validation. Building larger, high-quality datasets remains essential to power predictive AI/ML models, with concerns over data bias and accessibility if controlled by commercial entities. The future outlook is promising, with collaborative data-sharing initiatives and interpretable AI techniques expected to enhance rational antibody design. Synergies between NGS, high-throughput experimentation, and ML will accelerate discovery, addressing global demands for therapeutic antibodies. NGS empowers Isogenica to enhance library quality, discover rare binders, and continually refine selection strategies, leading to more effective antibody development. When combined with AI/ML and CIS Display™, it transforms workflows into faster, more precise, and data-rich processes, paving the way for innovative therapeutics. Clark, L. A., et al. (2023). Affinity maturation of antibodies by combinatorial codon mutagenesis versus error-prone PCR-derived mutagenesis. BMC Biotechnology, 23(1): 55. Frigotto, L., et al. (2015). Codon-Precise, Synthetic, Antibody Fragment Libraries Built Using Automated Hexamer Codon Additions and Validated through Next Gen- eration Sequencing. Antibodies, 4(2): 88-102. Makowski, E. K., et al. (2024). Optimization of therapeutic antibodies for reduced self-association in high-concentration formulations. Protein Engineering, Design and Selection, 37: gzae005. Rouet, R., et al. (2018). Next-generation sequencing of antibody display repertoires. Frontiers in Immunology, 9: 118. Biologic therapies have reshaped oncology, but they come with real limitations. Traditional monoclonal antibodies are large, complex molecules that can struggle with tumor penetration, slow development timelines, and costly manufacturing. In a field where time is critical and healthcare budgets are under increasing pressure, these constraints can hold back promising innovations. Read the article. Conventional immunization-based discovery techniques frequently result in overlapping sequences with intellectual property (IP) concerns, due to the high number of antibody sequences already patented. These limitations can restrict opportunities for development, increase legal uncertainty, and reduce commercial viability. This case study delves into the development of anti-LRP6 VHH inhibitors that overcome these challenges. Isogenica’s VHH technology offers hope for more effective therapies targeting Wnt-related cancers. This case study delves into the development of anti-LRP6 VHH inhibitors that overcome these challenges. Isogenica’s VHH technology offers hope for more effective therapies targeting Wnt-related cancers.

With more than 180 therapeutic proteins and peptides approved by the US Food and Drug Administration (FDA), the market for biological therapeutics, also called biologics, is booming. But many small biologics, including VHH-based therapies, have one significant challenge that must be overcome: their short half-life. The short half-life of many promising biologics is attributed to their small size of less than 70 kDa, which makes it easy for the kidney to filter them out. Because of this limitation, have gruelling dosing schedules that require in-person visits to the hospital/clinic so patients can continuously receive infusions. At Isogenica, half-life extension for our VHHs was always top of our list. Improving half-life is essential to realise the potential of smaller biotherapeutics, like VHH-based drugs, by reducing dosing frequency and healthcare costs and, ultimately, improving the quality of life for patients. In this article, we look at some ways in which the half-life of biological therapeutics can be extended to deliver on this promise, and why we chose an anti-albumin approach – ISOXTEND®. Over the years, several strategies have gained momentum in improving the half-life of biologics. While some of them improve the half-life by simply increasing the size of the therapeutic agent, others modify the structure of the drug to reduce the rate at which it is cleared. Many strategies also involve ‘tying’ the biologic to biomolecules that have a better half-life, such as antibodies, transferrin, or human serum albumin (HSA). Here are some of the common techniques that can be used to increase the half-life of small biologics. PEGylation was the first successful technology used to extend the half-life of biologics and has been applied to clinical medicine for more than 25 years. Through this process, polyethylene glycol (PEG) molecules are covalently attached to the biologic, increasing its molecular weight and effective size, which prevents the kidneys from quickly filtering the biologic based on size. Additionally, PEGs prevent access to proteases and peptidases that can degrade protein-based drugs, playing a further role in extending the half-life of these therapeutics. Despite the efficacy of PEGylation, there are several disadvantages to this technology. For instance, research suggests that PEGs are not 100% safe – they are immunogenic and cells treated with PEGylated biologics can develop PEG-containing vacuoles. Additionally, because PEGs change the physicochemical properties of the biological therapeutic, they can also greatly reduce its potency. Repeating amino acid chains, which work in similar ways to PEG, can also be attached to biologics. Like PEG, they increase the effective size of the biological therapeutic to reduce direct filtration by the kidneys, improving the overall half-life of the biologic. There are two leading amino acid half-life extension technologies – XTEN and ELPylation – although the use of other amino acid chains like PASylation and HAPylation is also being researched. XTEN is a protein polymer developed by Amunix. It comprises a library of amino acid sequences expressed by E. coli that coil randomly around a small biological therapeutic. Depending on the length of the XTEN sequence used, the half-life of a biologic can be tuned to fit its clinical indication. ELPylation, a technology developed by PhaseBio, also uses repeating amino acid chains to increase the size of biologics and thereby improve their half-life. Unlike PEGs, these repeating amino acid technologies are biodegradable and non-toxic. But despite this, they have similar drawbacks to PEGs, including the potential for immunogenicity of the polypeptide repeat units. Additionally, these kinds of half-life extenders can raise a series of potential manufacturing and dosing challenges, including developability, stability, solubility, immunogenicity and loss of activity in comparison with the unmodified drug. Small proteins in the blood are normally cleared out of the body by being broken down by cells in the liver or filtered out by the kidneys. But some proteins – such as human IgG antibodies and human serum albumin (HSA) – escape this fate by binding to cell surface receptors (FcRn) that enable them to be taken up by cells and ‘recycled’, prolonging their persistence in the body. Fusing biologics with HSA or fragments from IgG is a commonly used technique for improving the half-life of many small biologics that would otherwise be lost by filtration or metabolism. Fusing a biological therapeutic with IgG enables it to bind to the FcRn receptor and be recycled through cells, bypassing degradation or filtration. Although this is an effective way to improve the half-life of biologics, Fc domains on IgG can cause toxicity in the liver and have unwanted side effects through interactions with immune cells. There are also limitations in applying this across preclinical species. HSA is the most abundant plasma protein. It is highly soluble, extremely stable, and has a long circulatory half-life as a result of the FcRn-mediated recycling pathway. Direct or indirect binding of biologics to Human Serum Albumin (HSA) also recruits the FcRn-mediated recycling pathway, protecting biologics from degradation by recycling them through cells and significantly extending half-life. Early attempts to exploit the long half-life of HSA involved building fusion proteins combining biologics with serum albumin. However, HSA is a large protein and can exert significant effects on the biophysical properties of the drug, as well as its efficacy being limited to human studies. However, novel approaches sidestep this issue by linking therapeutics with HSA in a way that preserves the structure and function of the drug. Originally derived from camelids, small-format single-domain VHH antibodies have binding capabilities similar to those of conventional monoclonal antibodies across a wide pH range, but are less immunogenic and can penetrate into smaller spaces than conventional IgG antibodies with Fc domains. ISOXTEND® is a humanised VHH antibody that binds to albumin with nanomolar affinity maintained across a wide pH range. This enables the antibody to ‘piggyback’ through the endocytic recycling pathway, protecting it from clearance by the kidney while maintaining all other functional characteristics. ISOXTEND® is in vivo validated and has multi-species cross-reactivity, making it suitable for use in preclinical models including mouse and non-human primates. Our studies have shown that ISOXTEND® extends the half-life of VHH-based therapeutics up to 26 hours in mice – the approximate equivalent of 16-19 days in humans. The adaptability and flexibility of VHHs mean that ISOXTEND® can easily be chained together with other VHHs to create mono- or multi-specific biotherapeutics, or added to other biologics. Half-life extenders like ISOXTEND® could make a big difference to the quality of life for patients being treated with cutting-edge biological therapeutics, increasing the time between doses, reducing the time they spend in hospital, and realising the potential of these next-generation treatments. Interested to see how you can use ISOXTEND® to improve the half-life of your biological therapeutics? Get in touch to discover how we can help.

Small, simple, and robust – VHH antibodies make ideal building blocks for advanced therapies. With all the binding affinity of a full-length monoclonal in just 10% of the size, VHHs (sometimes known as nanobodies) have unique properties that are advantageous for developing antibody drug conjugates: – High solubility – naturally monomeric, VHH single domains are more soluble than the VH domains found in traditional IgGs, for improved drug developability. – Simple and robust – Isogenica’s VHH libraries have a fixed scaffold with just one internal disulphide bond. Together with the innate robustness of the molecules, they fold well across a wide pH range allowing buffer transitions needed for payload conjugation. – Low immunogenicity – With 4 VHH-based therapies already approved, patient studies show low immunogenicity risks for VHHs generally. To mitigate this further, Isogenica offers pre-humanised solutions to reduce requirements for lead engineering. – Unique CDR3 domains – VHHs have longer CDR3s than traditional IgGs, this means they can penetrate deep into binding pockets to hit difficult targets. – Ease of manufacture – Due to their simplicity, VHHs can be manufactured readily in microbial systems such as bacteria or yeast. This can be a big advantage when using non-canonical amino acid handles for payload conjugation. Knowing where your drug payload is conjugated on your targeting antibody is critical for making a consistent and reliable ADC drug. Even on small antibodies like VHHs, there are several options for site-specific drug conjugation, but they need to be built in. The most straightforward is to use a C-terminal handle. With a C-terminal handle present on your VHH antibody, a wide variety of payloads can be conjugated specifically and effectively, including: – small molecule drugs for ADCs – radionuclides for radiopharmaceutical or radiodiagnostic applications – oligonucleotides for DNA or RNA delivery as antibody oligonucleotide conjugates (AOCs) – viral particles for gene therapy delivery – LNPs for delivery of genetic and other payloads The C-terminal of the VHH is a low-risk choice for payload conjugation because it extends away in the opposite direction from the binding region of the VHH antibody. This limits any interference of the ADC payload in antibody binding. The two most popular choices of tag with our partners are: 1. C-terminal cysteine (Cys) for thiol conjugation to e.g. maleimide-linked payloads 2. C-terminal sortase for enzymatic conjugation to payloads with a polyglycine tag Potential handles for site-specific antibody drug conjugation on VHH antibodies. Additionally, synthetic “non-canonical” amino acids can be encoded into VHHs to create bio-orthogonal handles within the VHH molecule itself at appropriate sites in the frameworks. Due to unique reactive groups present on non-canonical amino acids, incredibly specific and efficient chemical reactions can be used to link the desired payload to the amino acid side chain, as described by Axup et al in their PNAS article from 2012. Thanks to the simplicity of VHHs, they can be produced in most biological systems, including microbial expression platforms like E. coli which are easily engineered to incorporate non-canonical amino acids. Our further partnerships include research collaborations to demonstrate the efficacy of VHHs conjugated to proprietary cytotoxic payloads on a variety of different cell types.